diff options
Diffstat (limited to 'drivers/gpu/drm')
264 files changed, 12621 insertions, 3152 deletions
diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile index c62339b89d46..784229d4504d 100644 --- a/drivers/gpu/drm/Makefile +++ b/drivers/gpu/drm/Makefile @@ -129,7 +129,6 @@ obj-$(CONFIG_DRM_TTM_HELPER) += drm_ttm_helper.o drm_kms_helper-y := \ drm_atomic_helper.o \ drm_atomic_state_helper.o \ - drm_bridge_connector.o \ drm_crtc_helper.o \ drm_damage_helper.o \ drm_encoder_slave.o \ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 7fe41a3c2541..e095572458cd 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -1164,6 +1164,7 @@ struct amdgpu_device { bool debug_disable_soft_recovery; bool debug_use_vram_fw_buf; bool debug_enable_ras_aca; + bool debug_exp_resets; bool enforce_isolation[MAX_XCP]; /* Added this mutex for cleaner shader isolation between GFX and compute processes */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c index c63528a4e894..1254a43ec96b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c @@ -1151,6 +1151,10 @@ uint64_t kgd_gfx_v9_hqd_get_pq_addr(struct amdgpu_device *adev, uint32_t low, high; uint64_t queue_addr = 0; + if (!adev->debug_exp_resets && + !adev->gfx.num_gfx_rings) + return 0; + kgd_gfx_v9_acquire_queue(adev, pipe_id, queue_id, inst); amdgpu_gfx_rlc_enter_safe_mode(adev, inst); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dev_coredump.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_dev_coredump.c index cf2b4dd4d865..5ac59b62020c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dev_coredump.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dev_coredump.c @@ -28,8 +28,8 @@ #include "atom.h" #ifndef CONFIG_DEV_COREDUMP -void amdgpu_coredump(struct amdgpu_device *adev, bool vram_lost, - struct amdgpu_reset_context *reset_context) +void amdgpu_coredump(struct amdgpu_device *adev, bool skip_vram_check, + bool vram_lost, struct amdgpu_job *job) { } #else @@ -315,7 +315,9 @@ amdgpu_devcoredump_read(char *buffer, loff_t offset, size_t count, } } - if (coredump->reset_vram_lost) + if (coredump->skip_vram_check) + drm_printf(&p, "VRAM lost check is skipped!\n"); + else if (coredump->reset_vram_lost) drm_printf(&p, "VRAM is lost due to GPU reset!\n"); return count - iter.remain; @@ -326,12 +328,11 @@ static void amdgpu_devcoredump_free(void *data) kfree(data); } -void amdgpu_coredump(struct amdgpu_device *adev, bool vram_lost, - struct amdgpu_reset_context *reset_context) +void amdgpu_coredump(struct amdgpu_device *adev, bool skip_vram_check, + bool vram_lost, struct amdgpu_job *job) { - struct amdgpu_coredump_info *coredump; struct drm_device *dev = adev_to_drm(adev); - struct amdgpu_job *job = reset_context->job; + struct amdgpu_coredump_info *coredump; struct drm_sched_job *s_job; coredump = kzalloc(sizeof(*coredump), GFP_NOWAIT); @@ -341,11 +342,12 @@ void amdgpu_coredump(struct amdgpu_device *adev, bool vram_lost, return; } + coredump->skip_vram_check = skip_vram_check; coredump->reset_vram_lost = vram_lost; - if (reset_context->job && reset_context->job->vm) { + if (job && job->vm) { + struct amdgpu_vm *vm = job->vm; struct amdgpu_task_info *ti; - struct amdgpu_vm *vm = reset_context->job->vm; ti = amdgpu_vm_get_task_info_vm(vm); if (ti) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dev_coredump.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_dev_coredump.h index 52459512cb2b..ef9772c6bcc9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dev_coredump.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dev_coredump.h @@ -26,7 +26,6 @@ #define __AMDGPU_DEV_COREDUMP_H__ #include "amdgpu.h" -#include "amdgpu_reset.h" #ifdef CONFIG_DEV_COREDUMP @@ -36,12 +35,12 @@ struct amdgpu_coredump_info { struct amdgpu_device *adev; struct amdgpu_task_info reset_task_info; struct timespec64 reset_time; + bool skip_vram_check; bool reset_vram_lost; struct amdgpu_ring *ring; }; #endif -void amdgpu_coredump(struct amdgpu_device *adev, bool vram_lost, - struct amdgpu_reset_context *reset_context); - +void amdgpu_coredump(struct amdgpu_device *adev, bool skip_vram_check, + bool vram_lost, struct amdgpu_job *job); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 0aad05b4f32e..e479ff33a0a7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -4531,6 +4531,9 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev) { dev_info(adev->dev, "amdgpu: finishing device.\n"); flush_delayed_work(&adev->delayed_init_work); + + if (adev->mman.initialized) + drain_workqueue(adev->mman.bdev.wq); adev->shutdown = true; /* make sure IB test finished before entering exclusive mode @@ -4551,9 +4554,6 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev) } amdgpu_fence_driver_hw_fini(adev); - if (adev->mman.initialized) - drain_workqueue(adev->mman.bdev.wq); - if (adev->pm.sysfs_initialized) amdgpu_pm_sysfs_fini(adev); if (adev->ucode_sysfs_en) @@ -5489,7 +5489,7 @@ int amdgpu_do_asic_reset(struct list_head *device_list_handle, vram_lost = amdgpu_device_check_vram_lost(tmp_adev); if (!test_bit(AMDGPU_SKIP_COREDUMP, &reset_context->flags)) - amdgpu_coredump(tmp_adev, vram_lost, reset_context); + amdgpu_coredump(tmp_adev, false, vram_lost, reset_context->job); if (vram_lost) { DRM_INFO("VRAM is lost due to GPU reset!\n"); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index 5dd39e6c6223..8dee7c62c801 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -131,6 +131,7 @@ enum AMDGPU_DEBUG_MASK { AMDGPU_DEBUG_DISABLE_GPU_SOFT_RECOVERY = BIT(2), AMDGPU_DEBUG_USE_VRAM_FW_BUF = BIT(3), AMDGPU_DEBUG_ENABLE_RAS_ACA = BIT(4), + AMDGPU_DEBUG_ENABLE_EXP_RESETS = BIT(5), }; unsigned int amdgpu_vram_limit = UINT_MAX; @@ -2199,6 +2200,11 @@ static void amdgpu_init_debug_options(struct amdgpu_device *adev) pr_info("debug: enable RAS ACA\n"); adev->debug_enable_ras_aca = true; } + + if (amdgpu_debug_mask & AMDGPU_DEBUG_ENABLE_EXP_RESETS) { + pr_info("debug: enable experimental reset features\n"); + adev->debug_exp_resets = true; + } } static unsigned long amdgpu_fix_asic_type(struct pci_dev *pdev, unsigned long flags) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c index aad2027e5c7c..0e617dff8765 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c @@ -348,6 +348,9 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data, return -EINVAL; } + /* always clear VRAM */ + flags |= AMDGPU_GEM_CREATE_VRAM_CLEARED; + /* create a gem object to contain this object in */ if (args->in.domains & (AMDGPU_GEM_DOMAIN_GDS | AMDGPU_GEM_DOMAIN_GWS | AMDGPU_GEM_DOMAIN_OA)) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c index b4efeef848de..b779d47a546a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c @@ -660,7 +660,7 @@ int amdgpu_gfx_enable_kcq(struct amdgpu_device *adev, int xcc_id) uint64_t queue_mask = 0; int r, i, j; - if (adev->enable_mes) + if (adev->mes.enable_legacy_queue_map) return amdgpu_gfx_mes_enable_kcq(adev, xcc_id); if (!kiq->pmf || !kiq->pmf->kiq_map_queues || !kiq->pmf->kiq_set_resources) @@ -722,7 +722,7 @@ int amdgpu_gfx_enable_kgq(struct amdgpu_device *adev, int xcc_id) amdgpu_device_flush_hdp(adev, NULL); - if (adev->enable_mes) { + if (adev->mes.enable_legacy_queue_map) { for (i = 0; i < adev->gfx.num_gfx_rings; i++) { j = i + xcc_id * adev->gfx.num_gfx_rings; r = amdgpu_mes_map_legacy_queue(adev, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c index 3bb9d1ca74b8..570820edcfda 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c @@ -30,6 +30,60 @@ #include "amdgpu.h" #include "amdgpu_trace.h" #include "amdgpu_reset.h" +#include "amdgpu_dev_coredump.h" +#include "amdgpu_xgmi.h" + +static void amdgpu_job_do_core_dump(struct amdgpu_device *adev, + struct amdgpu_job *job) +{ + int i; + + dev_info(adev->dev, "Dumping IP State\n"); + for (i = 0; i < adev->num_ip_blocks; i++) + if (adev->ip_blocks[i].version->funcs->dump_ip_state) + adev->ip_blocks[i].version->funcs + ->dump_ip_state((void *)adev); + dev_info(adev->dev, "Dumping IP State Completed\n"); + + amdgpu_coredump(adev, true, false, job); +} + +static void amdgpu_job_core_dump(struct amdgpu_device *adev, + struct amdgpu_job *job) +{ + struct list_head device_list, *device_list_handle = NULL; + struct amdgpu_device *tmp_adev = NULL; + struct amdgpu_hive_info *hive = NULL; + + if (!amdgpu_sriov_vf(adev)) + hive = amdgpu_get_xgmi_hive(adev); + if (hive) + mutex_lock(&hive->hive_lock); + /* + * Reuse the logic in amdgpu_device_gpu_recover() to build list of + * devices for code dump + */ + INIT_LIST_HEAD(&device_list); + if (!amdgpu_sriov_vf(adev) && (adev->gmc.xgmi.num_physical_nodes > 1) && hive) { + list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) + list_add_tail(&tmp_adev->reset_list, &device_list); + if (!list_is_first(&adev->reset_list, &device_list)) + list_rotate_to_front(&adev->reset_list, &device_list); + device_list_handle = &device_list; + } else { + list_add_tail(&adev->reset_list, &device_list); + device_list_handle = &device_list; + } + + /* Do the coredump for each device */ + list_for_each_entry(tmp_adev, device_list_handle, reset_list) + amdgpu_job_do_core_dump(tmp_adev, job); + + if (hive) { + mutex_unlock(&hive->hive_lock); + amdgpu_put_xgmi_hive(hive); + } +} static enum drm_gpu_sched_stat amdgpu_job_timedout(struct drm_sched_job *s_job) { @@ -48,9 +102,14 @@ static enum drm_gpu_sched_stat amdgpu_job_timedout(struct drm_sched_job *s_job) return DRM_GPU_SCHED_STAT_ENODEV; } - adev->job_hang = true; + /* + * Do the coredump immediately after a job timeout to get a very + * close dump/snapshot/representation of GPU's current error status + */ + amdgpu_job_core_dump(adev, job); + if (amdgpu_gpu_recovery && amdgpu_ring_soft_recovery(ring, job->vmid, s_job->s_fence->parent)) { dev_err(adev->dev, "ring %s timeout, but soft recovered\n", @@ -101,6 +160,12 @@ static enum drm_gpu_sched_stat amdgpu_job_timedout(struct drm_sched_job *s_job) reset_context.src = AMDGPU_RESET_SRC_JOB; clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags); + /* + * To avoid an unnecessary extra coredump, as we have already + * got the very close representation of GPU's error status + */ + set_bit(AMDGPU_SKIP_COREDUMP, &reset_context.flags); + r = amdgpu_device_gpu_recover(ring->adev, job, &reset_context); if (r) dev_err(adev->dev, "GPU Recovery Failed: %d\n", r); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c index 44c74a08987d..f7d5d4f08a53 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c @@ -826,6 +826,24 @@ int amdgpu_mes_reset_hw_queue(struct amdgpu_device *adev, int queue_id) return 0; } +int amdgpu_mes_reset_hw_queue_mmio(struct amdgpu_device *adev, int queue_type, + int me_id, int pipe_id, int queue_id, int vmid) +{ + struct mes_reset_queue_input queue_input; + int r; + + queue_input.use_mmio = true; + queue_input.me_id = me_id; + queue_input.pipe_id = pipe_id; + queue_input.queue_id = queue_id; + queue_input.vmid = vmid; + r = adev->mes.funcs->reset_hw_queue(&adev->mes, &queue_input); + if (r) + DRM_ERROR("failed to reset hardware queue by mmio, queue id = %d\n", + queue_id); + return r; +} + int amdgpu_mes_map_legacy_queue(struct amdgpu_device *adev, struct amdgpu_ring *ring) { @@ -873,7 +891,8 @@ int amdgpu_mes_unmap_legacy_queue(struct amdgpu_device *adev, int amdgpu_mes_reset_legacy_queue(struct amdgpu_device *adev, struct amdgpu_ring *ring, - unsigned int vmid) + unsigned int vmid, + bool use_mmio) { struct mes_reset_legacy_queue_input queue_input; int r; @@ -882,11 +901,13 @@ int amdgpu_mes_reset_legacy_queue(struct amdgpu_device *adev, queue_input.queue_type = ring->funcs->type; queue_input.doorbell_offset = ring->doorbell_index; + queue_input.me_id = ring->me; queue_input.pipe_id = ring->pipe; queue_input.queue_id = ring->queue; queue_input.mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj); queue_input.wptr_addr = ring->wptr_gpu_addr; queue_input.vmid = vmid; + queue_input.use_mmio = use_mmio; r = adev->mes.funcs->reset_legacy_queue(&adev->mes, &queue_input); if (r) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h index a5b1ea60cac8..96788c0f42f1 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h @@ -75,6 +75,7 @@ struct amdgpu_mes { uint32_t sched_version; uint32_t kiq_version; + bool enable_legacy_queue_map; uint32_t total_max_queue; uint32_t max_doorbell_slices; @@ -251,6 +252,13 @@ struct mes_remove_queue_input { struct mes_reset_queue_input { uint32_t doorbell_offset; uint64_t gang_context_addr; + bool use_mmio; + uint32_t queue_type; + uint32_t me_id; + uint32_t pipe_id; + uint32_t queue_id; + uint32_t xcc_id; + uint32_t vmid; }; struct mes_map_legacy_queue_input { @@ -287,6 +295,8 @@ struct mes_resume_gang_input { struct mes_reset_legacy_queue_input { uint32_t queue_type; uint32_t doorbell_offset; + bool use_mmio; + uint32_t me_id; uint32_t pipe_id; uint32_t queue_id; uint64_t mqd_addr; @@ -396,6 +406,8 @@ int amdgpu_mes_add_hw_queue(struct amdgpu_device *adev, int gang_id, int *queue_id); int amdgpu_mes_remove_hw_queue(struct amdgpu_device *adev, int queue_id); int amdgpu_mes_reset_hw_queue(struct amdgpu_device *adev, int queue_id); +int amdgpu_mes_reset_hw_queue_mmio(struct amdgpu_device *adev, int queue_type, + int me_id, int pipe_id, int queue_id, int vmid); int amdgpu_mes_map_legacy_queue(struct amdgpu_device *adev, struct amdgpu_ring *ring); @@ -405,7 +417,8 @@ int amdgpu_mes_unmap_legacy_queue(struct amdgpu_device *adev, u64 gpu_addr, u64 seq); int amdgpu_mes_reset_legacy_queue(struct amdgpu_device *adev, struct amdgpu_ring *ring, - unsigned int vmid); + unsigned int vmid, + bool use_mmio); uint32_t amdgpu_mes_rreg(struct amdgpu_device *adev, uint32_t reg); int amdgpu_mes_wreg(struct amdgpu_device *adev, diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c index ca983a014ba0..45ed97038df0 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c @@ -6692,13 +6692,13 @@ static int gfx_v10_0_gfx_mqd_init(struct amdgpu_device *adev, void *m, return 0; } -static int gfx_v10_0_gfx_init_queue(struct amdgpu_ring *ring) +static int gfx_v10_0_kgq_init_queue(struct amdgpu_ring *ring, bool reset) { struct amdgpu_device *adev = ring->adev; struct v10_gfx_mqd *mqd = ring->mqd_ptr; int mqd_idx = ring - &adev->gfx.gfx_ring[0]; - if (!amdgpu_in_reset(adev) && !adev->in_suspend) { + if (!reset && !amdgpu_in_reset(adev) && !adev->in_suspend) { memset((void *)mqd, 0, sizeof(*mqd)); mutex_lock(&adev->srbm_mutex); nv_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0); @@ -6750,7 +6750,7 @@ static int gfx_v10_0_cp_async_gfx_ring_resume(struct amdgpu_device *adev) r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr); if (!r) { - r = gfx_v10_0_gfx_init_queue(ring); + r = gfx_v10_0_kgq_init_queue(ring, false); amdgpu_bo_kunmap(ring->mqd_obj); ring->mqd_ptr = NULL; } @@ -7030,13 +7030,13 @@ static int gfx_v10_0_kiq_init_queue(struct amdgpu_ring *ring) return 0; } -static int gfx_v10_0_kcq_init_queue(struct amdgpu_ring *ring) +static int gfx_v10_0_kcq_init_queue(struct amdgpu_ring *ring, bool restore) { struct amdgpu_device *adev = ring->adev; struct v10_compute_mqd *mqd = ring->mqd_ptr; int mqd_idx = ring - &adev->gfx.compute_ring[0]; - if (!amdgpu_in_reset(adev) && !adev->in_suspend) { + if (!restore && !amdgpu_in_reset(adev) && !adev->in_suspend) { memset((void *)mqd, 0, sizeof(*mqd)); mutex_lock(&adev->srbm_mutex); nv_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0); @@ -7098,7 +7098,7 @@ static int gfx_v10_0_kcq_resume(struct amdgpu_device *adev) goto done; r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr); if (!r) { - r = gfx_v10_0_kcq_init_queue(ring); + r = gfx_v10_0_kcq_init_queue(ring, false); amdgpu_bo_kunmap(ring->mqd_obj); ring->mqd_ptr = NULL; } @@ -8949,7 +8949,9 @@ static void gfx_v10_0_ring_soft_recovery(struct amdgpu_ring *ring, value = REG_SET_FIELD(value, SQ_CMD, MODE, 0x01); value = REG_SET_FIELD(value, SQ_CMD, CHECK_VMID, 1); value = REG_SET_FIELD(value, SQ_CMD, VM_ID, vmid); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); WREG32_SOC15(GC, 0, mmSQ_CMD, value); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); } static void @@ -9416,6 +9418,156 @@ static void gfx_v10_ring_insert_nop(struct amdgpu_ring *ring, uint32_t num_nop) amdgpu_ring_write(ring, ring->funcs->nop); } +static int gfx_v10_0_reset_kgq(struct amdgpu_ring *ring, unsigned int vmid) +{ + struct amdgpu_device *adev = ring->adev; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[0]; + struct amdgpu_ring *kiq_ring = &kiq->ring; + unsigned long flags; + u32 tmp; + u64 addr; + int r; + + if (amdgpu_sriov_vf(adev)) + return -EINVAL; + + if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues) + return -EINVAL; + + spin_lock_irqsave(&kiq->ring_lock, flags); + + if (amdgpu_ring_alloc(kiq_ring, 5 + 7 + 7 + kiq->pmf->map_queues_size)) { + spin_unlock_irqrestore(&kiq->ring_lock, flags); + return -ENOMEM; + } + + addr = amdgpu_bo_gpu_offset(ring->mqd_obj) + + offsetof(struct v10_gfx_mqd, cp_gfx_hqd_active); + tmp = REG_SET_FIELD(0, CP_VMID_RESET, RESET_REQUEST, 1 << vmid); + if (ring->pipe == 0) + tmp = REG_SET_FIELD(tmp, CP_VMID_RESET, PIPE0_QUEUES, 1 << ring->queue); + else + tmp = REG_SET_FIELD(tmp, CP_VMID_RESET, PIPE1_QUEUES, 1 << ring->queue); + + gfx_v10_0_ring_emit_wreg(kiq_ring, + SOC15_REG_OFFSET(GC, 0, mmCP_VMID_RESET), tmp); + gfx_v10_0_wait_reg_mem(kiq_ring, 0, 1, 0, + lower_32_bits(addr), upper_32_bits(addr), + 0, 1, 0x20); + gfx_v10_0_ring_emit_reg_wait(kiq_ring, + SOC15_REG_OFFSET(GC, 0, mmCP_VMID_RESET), 0, 0xffffffff); + kiq->pmf->kiq_map_queues(kiq_ring, ring); + amdgpu_ring_commit(kiq_ring); + + spin_unlock_irqrestore(&kiq->ring_lock, flags); + + r = amdgpu_ring_test_ring(kiq_ring); + if (r) + return r; + + r = amdgpu_bo_reserve(ring->mqd_obj, false); + if (unlikely(r != 0)) { + DRM_ERROR("fail to resv mqd_obj\n"); + return r; + } + r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr); + if (!r) { + r = gfx_v10_0_kgq_init_queue(ring, true); + amdgpu_bo_kunmap(ring->mqd_obj); + ring->mqd_ptr = NULL; + } + amdgpu_bo_unreserve(ring->mqd_obj); + if (r) { + DRM_ERROR("fail to unresv mqd_obj\n"); + return r; + } + + return amdgpu_ring_test_ring(ring); +} + +static int gfx_v10_0_reset_kcq(struct amdgpu_ring *ring, + unsigned int vmid) +{ + struct amdgpu_device *adev = ring->adev; + struct amdgpu_kiq *kiq = &adev->gfx.kiq[0]; + struct amdgpu_ring *kiq_ring = &kiq->ring; + unsigned long flags; + int i, r; + + if (amdgpu_sriov_vf(adev)) + return -EINVAL; + + if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues) + return -EINVAL; + + spin_lock_irqsave(&kiq->ring_lock, flags); + + if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->unmap_queues_size)) { + spin_unlock_irqrestore(&kiq->ring_lock, flags); + return -ENOMEM; + } + + kiq->pmf->kiq_unmap_queues(kiq_ring, ring, RESET_QUEUES, + 0, 0); + amdgpu_ring_commit(kiq_ring); + spin_unlock_irqrestore(&kiq->ring_lock, flags); + + r = amdgpu_ring_test_ring(kiq_ring); + if (r) + return r; + + /* make sure dequeue is complete*/ + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); + mutex_lock(&adev->srbm_mutex); + nv_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0); + for (i = 0; i < adev->usec_timeout; i++) { + if (!(RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE) & 1)) + break; + udelay(1); + } + if (i >= adev->usec_timeout) + r = -ETIMEDOUT; + nv_grbm_select(adev, 0, 0, 0, 0); + mutex_unlock(&adev->srbm_mutex); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); + if (r) { + dev_err(adev->dev, "fail to wait on hqd deactivate\n"); + return r; + } + + r = amdgpu_bo_reserve(ring->mqd_obj, false); + if (unlikely(r != 0)) { + dev_err(adev->dev, "fail to resv mqd_obj\n"); + return r; + } + r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr); + if (!r) { + r = gfx_v10_0_kcq_init_queue(ring, true); + amdgpu_bo_kunmap(ring->mqd_obj); + ring->mqd_ptr = NULL; + } + amdgpu_bo_unreserve(ring->mqd_obj); + if (r) { + dev_err(adev->dev, "fail to unresv mqd_obj\n"); + return r; + } + + spin_lock_irqsave(&kiq->ring_lock, flags); + if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->map_queues_size)) { + spin_unlock_irqrestore(&kiq->ring_lock, flags); + return -ENOMEM; + } + kiq->pmf->kiq_map_queues(kiq_ring, ring); + amdgpu_ring_commit(kiq_ring); + spin_unlock_irqrestore(&kiq->ring_lock, flags); + + r = amdgpu_ring_test_ring(kiq_ring); + if (r) + return r; + + return amdgpu_ring_test_ring(ring); +} + static void gfx_v10_ip_print(void *handle, struct drm_printer *p) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; @@ -9619,6 +9771,7 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_gfx = { .emit_reg_write_reg_wait = gfx_v10_0_ring_emit_reg_write_reg_wait, .soft_recovery = gfx_v10_0_ring_soft_recovery, .emit_mem_sync = gfx_v10_0_emit_mem_sync, + .reset = gfx_v10_0_reset_kgq, }; static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_compute = { @@ -9655,6 +9808,7 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_compute = { .emit_reg_write_reg_wait = gfx_v10_0_ring_emit_reg_write_reg_wait, .soft_recovery = gfx_v10_0_ring_soft_recovery, .emit_mem_sync = gfx_v10_0_emit_mem_sync, + .reset = gfx_v10_0_reset_kcq, }; static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_kiq = { diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c index ee8604722467..d3e8be82a172 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c @@ -3984,13 +3984,13 @@ static int gfx_v11_0_gfx_mqd_init(struct amdgpu_device *adev, void *m, return 0; } -static int gfx_v11_0_gfx_init_queue(struct amdgpu_ring *ring) +static int gfx_v11_0_kgq_init_queue(struct amdgpu_ring *ring, bool reset) { struct amdgpu_device *adev = ring->adev; struct v11_gfx_mqd *mqd = ring->mqd_ptr; int mqd_idx = ring - &adev->gfx.gfx_ring[0]; - if (!amdgpu_in_reset(adev) && !adev->in_suspend) { + if (!reset && !amdgpu_in_reset(adev) && !adev->in_suspend) { memset((void *)mqd, 0, sizeof(*mqd)); mutex_lock(&adev->srbm_mutex); soc21_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0); @@ -4026,7 +4026,7 @@ static int gfx_v11_0_cp_async_gfx_ring_resume(struct amdgpu_device *adev) r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr); if (!r) { - r = gfx_v11_0_gfx_init_queue(ring); + r = gfx_v11_0_kgq_init_queue(ring, false); amdgpu_bo_kunmap(ring->mqd_obj); ring->mqd_ptr = NULL; } @@ -4321,13 +4321,13 @@ static int gfx_v11_0_kiq_init_queue(struct amdgpu_ring *ring) return 0; } -static int gfx_v11_0_kcq_init_queue(struct amdgpu_ring *ring) +static int gfx_v11_0_kcq_init_queue(struct amdgpu_ring *ring, bool reset) { struct amdgpu_device *adev = ring->adev; struct v11_compute_mqd *mqd = ring->mqd_ptr; int mqd_idx = ring - &adev->gfx.compute_ring[0]; - if (!amdgpu_in_reset(adev) && !adev->in_suspend) { + if (!reset && !amdgpu_in_reset(adev) && !adev->in_suspend) { memset((void *)mqd, 0, sizeof(*mqd)); mutex_lock(&adev->srbm_mutex); soc21_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0); @@ -4391,7 +4391,7 @@ static int gfx_v11_0_kcq_resume(struct amdgpu_device *adev) goto done; r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr); if (!r) { - r = gfx_v11_0_kcq_init_queue(ring); + r = gfx_v11_0_kcq_init_queue(ring, false); amdgpu_bo_kunmap(ring->mqd_obj); ring->mqd_ptr = NULL; } @@ -4781,7 +4781,7 @@ static int gfx_v11_0_soft_reset(void *handle) int r, i, j, k; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - gfx_v11_0_set_safe_mode(adev, 0); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); tmp = RREG32_SOC15(GC, 0, regCP_INT_CNTL); tmp = REG_SET_FIELD(tmp, CP_INT_CNTL, CMP_BUSY_INT_ENABLE, 0); @@ -4900,7 +4900,7 @@ static int gfx_v11_0_soft_reset(void *handle) tmp = REG_SET_FIELD(tmp, CP_INT_CNTL, GFX_IDLE_INT_ENABLE, 1); WREG32_SOC15(GC, 0, regCP_INT_CNTL, tmp); - gfx_v11_0_unset_safe_mode(adev, 0); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); return gfx_v11_0_cp_resume(adev); } @@ -5923,6 +5923,9 @@ static int gfx_v11_0_ring_preempt_ib(struct amdgpu_ring *ring) struct amdgpu_ring *kiq_ring = &kiq->ring; unsigned long flags; + if (adev->enable_mes) + return -EINVAL; + if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues) return -EINVAL; @@ -6088,7 +6091,9 @@ static void gfx_v11_0_ring_soft_recovery(struct amdgpu_ring *ring, value = REG_SET_FIELD(value, SQ_CMD, MODE, 0x01); value = REG_SET_FIELD(value, SQ_CMD, CHECK_VMID, 1); value = REG_SET_FIELD(value, SQ_CMD, VM_ID, vmid); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); WREG32_SOC15(GC, 0, regSQ_CMD, value); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); } static void @@ -6541,6 +6546,99 @@ static void gfx_v11_0_emit_mem_sync(struct amdgpu_ring *ring) amdgpu_ring_write(ring, gcr_cntl); /* GCR_CNTL */ } +static int gfx_v11_0_reset_kgq(struct amdgpu_ring *ring, unsigned int vmid) +{ + struct amdgpu_device *adev = ring->adev; + int r; + + if (amdgpu_sriov_vf(adev)) + return -EINVAL; + + r = amdgpu_mes_reset_legacy_queue(ring->adev, ring, vmid, false); + if (r) + return r; + + r = amdgpu_bo_reserve(ring->mqd_obj, false); + if (unlikely(r != 0)) { + dev_err(adev->dev, "fail to resv mqd_obj\n"); + return r; + } + r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr); + if (!r) { + r = gfx_v11_0_kgq_init_queue(ring, true); + amdgpu_bo_kunmap(ring->mqd_obj); + ring->mqd_ptr = NULL; + } + amdgpu_bo_unreserve(ring->mqd_obj); + if (r) { + dev_err(adev->dev, "fail to unresv mqd_obj\n"); + return r; + } + + r = amdgpu_mes_map_legacy_queue(adev, ring); + if (r) { + dev_err(adev->dev, "failed to remap kgq\n"); + return r; + } + + return amdgpu_ring_test_ring(ring); +} + +static int gfx_v11_0_reset_kcq(struct amdgpu_ring *ring, unsigned int vmid) +{ + struct amdgpu_device *adev = ring->adev; + int i, r = 0; + + if (amdgpu_sriov_vf(adev)) + return -EINVAL; + + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); + mutex_lock(&adev->srbm_mutex); + soc21_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0); + WREG32_SOC15(GC, 0, regCP_HQD_DEQUEUE_REQUEST, 0x2); + WREG32_SOC15(GC, 0, regSPI_COMPUTE_QUEUE_RESET, 0x1); + + /* make sure dequeue is complete*/ + for (i = 0; i < adev->usec_timeout; i++) { + if (!(RREG32_SOC15(GC, 0, regCP_HQD_ACTIVE) & 1)) + break; + udelay(1); + } + if (i >= adev->usec_timeout) + r = -ETIMEDOUT; + soc21_grbm_select(adev, 0, 0, 0, 0); + mutex_unlock(&adev->srbm_mutex); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); + if (r) { + dev_err(adev->dev, "fail to wait on hqd deactivate\n"); + return r; + } + + r = amdgpu_bo_reserve(ring->mqd_obj, false); + if (unlikely(r != 0)) { + dev_err(adev->dev, "fail to resv mqd_obj\n"); + return r; + } + r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr); + if (!r) { + r = gfx_v11_0_kcq_init_queue(ring, true); + amdgpu_bo_kunmap(ring->mqd_obj); + ring->mqd_ptr = NULL; + } + amdgpu_bo_unreserve(ring->mqd_obj); + if (r) { + dev_err(adev->dev, "fail to unresv mqd_obj\n"); + return r; + } + r = amdgpu_mes_map_legacy_queue(adev, ring); + if (r) { + dev_err(adev->dev, "failed to remap kcq\n"); + return r; + } + + return amdgpu_ring_test_ring(ring); +} + static void gfx_v11_ip_print(void *handle, struct drm_printer *p) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; @@ -6742,6 +6840,7 @@ static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_gfx = { .emit_reg_write_reg_wait = gfx_v11_0_ring_emit_reg_write_reg_wait, .soft_recovery = gfx_v11_0_ring_soft_recovery, .emit_mem_sync = gfx_v11_0_emit_mem_sync, + .reset = gfx_v11_0_reset_kgq, }; static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_compute = { @@ -6779,6 +6878,7 @@ static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_compute = { .emit_reg_write_reg_wait = gfx_v11_0_ring_emit_reg_write_reg_wait, .soft_recovery = gfx_v11_0_ring_soft_recovery, .emit_mem_sync = gfx_v11_0_emit_mem_sync, + .reset = gfx_v11_0_reset_kcq, }; static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_kiq = { diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c index 54059cbcfc08..d1357c01eb39 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c @@ -2916,13 +2916,13 @@ static int gfx_v12_0_gfx_mqd_init(struct amdgpu_device *adev, void *m, return 0; } -static int gfx_v12_0_gfx_init_queue(struct amdgpu_ring *ring) +static int gfx_v12_0_kgq_init_queue(struct amdgpu_ring *ring, bool reset) { struct amdgpu_device *adev = ring->adev; struct v12_gfx_mqd *mqd = ring->mqd_ptr; int mqd_idx = ring - &adev->gfx.gfx_ring[0]; - if (!amdgpu_in_reset(adev) && !adev->in_suspend) { + if (!reset && !amdgpu_in_reset(adev) && !adev->in_suspend) { memset((void *)mqd, 0, sizeof(*mqd)); mutex_lock(&adev->srbm_mutex); soc24_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0); @@ -2958,7 +2958,7 @@ static int gfx_v12_0_cp_async_gfx_ring_resume(struct amdgpu_device *adev) r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr); if (!r) { - r = gfx_v12_0_gfx_init_queue(ring); + r = gfx_v12_0_kgq_init_queue(ring, false); amdgpu_bo_kunmap(ring->mqd_obj); ring->mqd_ptr = NULL; } @@ -3262,13 +3262,13 @@ static int gfx_v12_0_kiq_init_queue(struct amdgpu_ring *ring) return 0; } -static int gfx_v12_0_kcq_init_queue(struct amdgpu_ring *ring) +static int gfx_v12_0_kcq_init_queue(struct amdgpu_ring *ring, bool reset) { struct amdgpu_device *adev = ring->adev; struct v12_compute_mqd *mqd = ring->mqd_ptr; int mqd_idx = ring - &adev->gfx.compute_ring[0]; - if (!amdgpu_in_reset(adev) && !adev->in_suspend) { + if (!reset && !amdgpu_in_reset(adev) && !adev->in_suspend) { memset((void *)mqd, 0, sizeof(*mqd)); mutex_lock(&adev->srbm_mutex); soc24_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0); @@ -3332,7 +3332,7 @@ static int gfx_v12_0_kcq_resume(struct amdgpu_device *adev) goto done; r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr); if (!r) { - r = gfx_v12_0_kcq_init_queue(ring); + r = gfx_v12_0_kcq_init_queue(ring, false); amdgpu_bo_kunmap(ring->mqd_obj); ring->mqd_ptr = NULL; } @@ -4501,6 +4501,9 @@ static int gfx_v12_0_ring_preempt_ib(struct amdgpu_ring *ring) struct amdgpu_ring *kiq_ring = &kiq->ring; unsigned long flags; + if (adev->enable_mes) + return -EINVAL; + if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues) return -EINVAL; @@ -4617,7 +4620,9 @@ static void gfx_v12_0_ring_soft_recovery(struct amdgpu_ring *ring, value = REG_SET_FIELD(value, SQ_CMD, MODE, 0x01); value = REG_SET_FIELD(value, SQ_CMD, CHECK_VMID, 1); value = REG_SET_FIELD(value, SQ_CMD, VM_ID, vmid); + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); WREG32_SOC15(GC, 0, regSQ_CMD, value); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); } static void @@ -5155,6 +5160,93 @@ static void gfx_v12_ip_dump(void *handle) amdgpu_gfx_off_ctrl(adev, true); } +static int gfx_v12_0_reset_kgq(struct amdgpu_ring *ring, unsigned int vmid) +{ + struct amdgpu_device *adev = ring->adev; + int r; + + if (amdgpu_sriov_vf(adev)) + return -EINVAL; + + r = amdgpu_mes_reset_legacy_queue(ring->adev, ring, vmid, false); + if (r) { + dev_err(adev->dev, "reset via MES failed %d\n", r); + return r; + } + + r = amdgpu_bo_reserve(ring->mqd_obj, false); + if (unlikely(r != 0)) { + dev_err(adev->dev, "fail to resv mqd_obj\n"); + return r; + } + r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr); + if (!r) { + r = gfx_v12_0_kgq_init_queue(ring, true); + amdgpu_bo_kunmap(ring->mqd_obj); + ring->mqd_ptr = NULL; + } + amdgpu_bo_unreserve(ring->mqd_obj); + if (r) { + DRM_ERROR("fail to unresv mqd_obj\n"); + return r; + } + + r = amdgpu_mes_map_legacy_queue(adev, ring); + if (r) { + dev_err(adev->dev, "failed to remap kgq\n"); + return r; + } + + return amdgpu_ring_test_ring(ring); +} + +static int gfx_v12_0_reset_kcq(struct amdgpu_ring *ring, unsigned int vmid) +{ + struct amdgpu_device *adev = ring->adev; + int r, i; + + if (amdgpu_sriov_vf(adev)) + return -EINVAL; + + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); + mutex_lock(&adev->srbm_mutex); + soc24_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0); + WREG32_SOC15(GC, 0, regCP_HQD_DEQUEUE_REQUEST, 0x2); + WREG32_SOC15(GC, 0, regSPI_COMPUTE_QUEUE_RESET, 0x1); + for (i = 0; i < adev->usec_timeout; i++) { + if (!(RREG32_SOC15(GC, 0, regCP_HQD_ACTIVE) & 1)) + break; + udelay(1); + } + soc24_grbm_select(adev, 0, 0, 0, 0); + mutex_unlock(&adev->srbm_mutex); + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); + + r = amdgpu_bo_reserve(ring->mqd_obj, false); + if (unlikely(r != 0)) { + DRM_ERROR("fail to resv mqd_obj\n"); + return r; + } + r = amdgpu_bo_kmap(ring->mqd_obj, (void **)&ring->mqd_ptr); + if (!r) { + r = gfx_v12_0_kcq_init_queue(ring, true); + amdgpu_bo_kunmap(ring->mqd_obj); + ring->mqd_ptr = NULL; + } + amdgpu_bo_unreserve(ring->mqd_obj); + if (r) { + DRM_ERROR("fail to unresv mqd_obj\n"); + return r; + } + r = amdgpu_mes_map_legacy_queue(adev, ring); + if (r) { + dev_err(adev->dev, "failed to remap kcq\n"); + return r; + } + + return amdgpu_ring_test_ring(ring); +} + static const struct amd_ip_funcs gfx_v12_0_ip_funcs = { .name = "gfx_v12_0", .early_init = gfx_v12_0_early_init, @@ -5217,6 +5309,7 @@ static const struct amdgpu_ring_funcs gfx_v12_0_ring_funcs_gfx = { .emit_reg_write_reg_wait = gfx_v12_0_ring_emit_reg_write_reg_wait, .soft_recovery = gfx_v12_0_ring_soft_recovery, .emit_mem_sync = gfx_v12_0_emit_mem_sync, + .reset = gfx_v12_0_reset_kgq, }; static const struct amdgpu_ring_funcs gfx_v12_0_ring_funcs_compute = { @@ -5251,6 +5344,7 @@ static const struct amdgpu_ring_funcs gfx_v12_0_ring_funcs_compute = { .emit_reg_write_reg_wait = gfx_v12_0_ring_emit_reg_write_reg_wait, .soft_recovery = gfx_v12_0_ring_soft_recovery, .emit_mem_sync = gfx_v12_0_emit_mem_sync, + .reset = gfx_v12_0_reset_kcq, }; static const struct amdgpu_ring_funcs gfx_v12_0_ring_funcs_kiq = { diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c index 21089aadbb7b..8cf5d7925b51 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c @@ -7233,6 +7233,10 @@ static int gfx_v9_0_reset_kcq(struct amdgpu_ring *ring, unsigned long flags; int i, r; + if (!adev->debug_exp_resets && + !adev->gfx.num_gfx_rings) + return -EINVAL; + if (amdgpu_sriov_vf(adev)) return -EINVAL; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c index 2067f26d3a9d..408e5600bb61 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c @@ -3052,6 +3052,9 @@ static void gfx_v9_4_3_ring_soft_recovery(struct amdgpu_ring *ring, struct amdgpu_device *adev = ring->adev; uint32_t value = 0; + if (!adev->debug_exp_resets) + return; + value = REG_SET_FIELD(value, SQ_CMD, CMD, 0x03); value = REG_SET_FIELD(value, SQ_CMD, MODE, 0x01); value = REG_SET_FIELD(value, SQ_CMD, CHECK_VMID, 1); @@ -3466,6 +3469,98 @@ static void gfx_v9_4_3_emit_wave_limit(struct amdgpu_ring *ring, bool enable) } } +static int gfx_v9_4_3_unmap_done(struct amdgpu_device *adev, uint32_t me, + uint32_t pipe, uint32_t queue, + uint32_t xcc_id) +{ + int i, r; + /* make sure dequeue is complete*/ + gfx_v9_4_3_xcc_set_safe_mode(adev, xcc_id); + mutex_lock(&adev->srbm_mutex); + soc15_grbm_select(adev, me, pipe, queue, 0, GET_INST(GC, xcc_id)); + for (i = 0; i < adev->usec_timeout; i++) { + if (!(RREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_HQD_ACTIVE) & 1)) + break; + udelay(1); + } + if (i >= adev->usec_timeout) + r = -ETIMEDOUT; + else + r = 0; + soc15_grbm_select(adev, 0, 0, 0, 0, GET_INST(GC, xcc_id)); + mutex_unlock(&adev->srbm_mutex); + gfx_v9_4_3_xcc_unset_safe_mode(adev, xcc_id); + + return r; + +} + +static bool gfx_v9_4_3_pipe_reset_support(struct amdgpu_device *adev) +{ + /*TODO: Need check gfx9.4.4 mec fw whether supports pipe reset as well.*/ + if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) && + adev->gfx.mec_fw_version >= 0x0000009b) + return true; + else + dev_warn_once(adev->dev, "Please use the latest MEC version to see whether support pipe reset\n"); + + return false; +} + +static int gfx_v9_4_3_reset_hw_pipe(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + uint32_t reset_pipe, clean_pipe; + int r; + + if (!gfx_v9_4_3_pipe_reset_support(adev)) + return -EINVAL; + + gfx_v9_4_3_xcc_set_safe_mode(adev, ring->xcc_id); + mutex_lock(&adev->srbm_mutex); + + reset_pipe = RREG32_SOC15(GC, GET_INST(GC, ring->xcc_id), regCP_MEC_CNTL); + clean_pipe = reset_pipe; + + if (ring->me == 1) { + switch (ring->pipe) { + case 0: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL, + MEC_ME1_PIPE0_RESET, 1); + break; + case 1: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL, + MEC_ME1_PIPE1_RESET, 1); + break; + case 2: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL, + MEC_ME1_PIPE2_RESET, 1); + break; + case 3: + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL, + MEC_ME1_PIPE3_RESET, 1); + break; + default: + break; + } + } else { + if (ring->pipe) + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL, + MEC_ME2_PIPE1_RESET, 1); + else + reset_pipe = REG_SET_FIELD(reset_pipe, CP_MEC_CNTL, + MEC_ME2_PIPE0_RESET, 1); + } + + WREG32_SOC15(GC, GET_INST(GC, ring->xcc_id), regCP_MEC_CNTL, reset_pipe); + WREG32_SOC15(GC, GET_INST(GC, ring->xcc_id), regCP_MEC_CNTL, clean_pipe); + mutex_unlock(&adev->srbm_mutex); + gfx_v9_4_3_xcc_unset_safe_mode(adev, ring->xcc_id); + + r = gfx_v9_4_3_unmap_done(adev, ring->me, ring->pipe, ring->queue, ring->xcc_id); + return r; +} + static int gfx_v9_4_3_reset_kcq(struct amdgpu_ring *ring, unsigned int vmid) { @@ -3473,7 +3568,10 @@ static int gfx_v9_4_3_reset_kcq(struct amdgpu_ring *ring, struct amdgpu_kiq *kiq = &adev->gfx.kiq[ring->xcc_id]; struct amdgpu_ring *kiq_ring = &kiq->ring; unsigned long flags; - int r, i; + int r; + + if (!adev->debug_exp_resets) + return -EINVAL; if (amdgpu_sriov_vf(adev)) return -EINVAL; @@ -3495,26 +3593,23 @@ static int gfx_v9_4_3_reset_kcq(struct amdgpu_ring *ring, spin_unlock_irqrestore(&kiq->ring_lock, flags); r = amdgpu_ring_test_ring(kiq_ring); + if (r) { + dev_err(adev->dev, "kiq ring test failed after ring: %s queue reset\n", + ring->name); + goto pipe_reset; + } + + r = gfx_v9_4_3_unmap_done(adev, ring->me, ring->pipe, ring->queue, ring->xcc_id); if (r) - return r; + dev_err(adev->dev, "fail to wait on hqd deactive and will try pipe reset\n"); - /* make sure dequeue is complete*/ - amdgpu_gfx_rlc_enter_safe_mode(adev, ring->xcc_id); - mutex_lock(&adev->srbm_mutex); - soc15_grbm_select(adev, ring->me, ring->pipe, ring->queue, 0, GET_INST(GC, ring->xcc_id)); - for (i = 0; i < adev->usec_timeout; i++) { - if (!(RREG32_SOC15(GC, 0, regCP_HQD_ACTIVE) & 1)) - break; - udelay(1); - } - if (i >= adev->usec_timeout) - r = -ETIMEDOUT; - soc15_grbm_select(adev, 0, 0, 0, 0, GET_INST(GC, ring->xcc_id)); - mutex_unlock(&adev->srbm_mutex); - amdgpu_gfx_rlc_exit_safe_mode(adev, ring->xcc_id); - if (r) { - dev_err(adev->dev, "fail to wait on hqd deactive\n"); - return r; +pipe_reset: + if(r) { + r = gfx_v9_4_3_reset_hw_pipe(ring); + dev_info(adev->dev, "ring: %s pipe reset :%s\n", ring->name, + r ? "failed" : "successfully"); + if (r) + return r; } r = amdgpu_bo_reserve(ring->mqd_obj, false); diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c index 8edcd85a1261..0f055d1b1da6 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c @@ -26,6 +26,7 @@ #include "amdgpu.h" #include "soc15_common.h" #include "soc21.h" +#include "gfx_v11_0.h" #include "gc/gc_11_0_0_offset.h" #include "gc/gc_11_0_0_sh_mask.h" #include "gc/gc_11_0_0_default.h" @@ -360,9 +361,83 @@ static int mes_v11_0_remove_hw_queue(struct amdgpu_mes *mes, offsetof(union MESAPI__REMOVE_QUEUE, api_status)); } +static int mes_v11_0_reset_queue_mmio(struct amdgpu_mes *mes, uint32_t queue_type, + uint32_t me_id, uint32_t pipe_id, + uint32_t queue_id, uint32_t vmid) +{ + struct amdgpu_device *adev = mes->adev; + uint32_t value; + int i, r = 0; + + amdgpu_gfx_rlc_enter_safe_mode(adev, 0); + + if (queue_type == AMDGPU_RING_TYPE_GFX) { + dev_info(adev->dev, "reset gfx queue (%d:%d:%d: vmid:%d)\n", + me_id, pipe_id, queue_id, vmid); + + mutex_lock(&adev->gfx.reset_sem_mutex); + gfx_v11_0_request_gfx_index_mutex(adev, true); + /* all se allow writes */ + WREG32_SOC15(GC, 0, regGRBM_GFX_INDEX, + (uint32_t)(0x1 << GRBM_GFX_INDEX__SE_BROADCAST_WRITES__SHIFT)); + value = REG_SET_FIELD(0, CP_VMID_RESET, RESET_REQUEST, 1 << vmid); + if (pipe_id == 0) + value = REG_SET_FIELD(value, CP_VMID_RESET, PIPE0_QUEUES, 1 << queue_id); + else + value = REG_SET_FIELD(value, CP_VMID_RESET, PIPE1_QUEUES, 1 << queue_id); + WREG32_SOC15(GC, 0, regCP_VMID_RESET, value); + gfx_v11_0_request_gfx_index_mutex(adev, false); + mutex_unlock(&adev->gfx.reset_sem_mutex); + + mutex_lock(&adev->srbm_mutex); + soc21_grbm_select(adev, me_id, pipe_id, queue_id, 0); + /* wait till dequeue take effects */ + for (i = 0; i < adev->usec_timeout; i++) { + if (!(RREG32_SOC15(GC, 0, regCP_GFX_HQD_ACTIVE) & 1)) + break; + udelay(1); + } + if (i >= adev->usec_timeout) { + dev_err(adev->dev, "failed to wait on gfx hqd deactivate\n"); + r = -ETIMEDOUT; + } + + soc21_grbm_select(adev, 0, 0, 0, 0); + mutex_unlock(&adev->srbm_mutex); + } else if (queue_type == AMDGPU_RING_TYPE_COMPUTE) { + dev_info(adev->dev, "reset compute queue (%d:%d:%d)\n", + me_id, pipe_id, queue_id); + mutex_lock(&adev->srbm_mutex); + soc21_grbm_select(adev, me_id, pipe_id, queue_id, 0); + WREG32_SOC15(GC, 0, regCP_HQD_DEQUEUE_REQUEST, 0x2); + WREG32_SOC15(GC, 0, regSPI_COMPUTE_QUEUE_RESET, 0x1); + + /* wait till dequeue take effects */ + for (i = 0; i < adev->usec_timeout; i++) { + if (!(RREG32_SOC15(GC, 0, regCP_HQD_ACTIVE) & 1)) + break; + udelay(1); + } + if (i >= adev->usec_timeout) { + dev_err(adev->dev, "failed to wait on hqd deactivate\n"); + r = -ETIMEDOUT; + } + soc21_grbm_select(adev, 0, 0, 0, 0); + mutex_unlock(&adev->srbm_mutex); + } + + amdgpu_gfx_rlc_exit_safe_mode(adev, 0); + return r; +} + static int mes_v11_0_reset_hw_queue(struct amdgpu_mes *mes, struct mes_reset_queue_input *input) { + if (input->use_mmio) + return mes_v11_0_reset_queue_mmio(mes, input->queue_type, + input->me_id, input->pipe_id, + input->queue_id, input->vmid); + union MESAPI__RESET mes_reset_queue_pkt; memset(&mes_reset_queue_pkt, 0, sizeof(mes_reset_queue_pkt)); @@ -648,6 +723,11 @@ static int mes_v11_0_reset_legacy_queue(struct amdgpu_mes *mes, { union MESAPI__RESET mes_reset_queue_pkt; + if (input->use_mmio) + return mes_v11_0_reset_queue_mmio(mes, input->queue_type, + input->me_id, input->pipe_id, + input->queue_id, input->vmid); + memset(&mes_reset_queue_pkt, 0, sizeof(mes_reset_queue_pkt)); mes_reset_queue_pkt.header.type = MES_API_TYPE_SCHEDULER; @@ -775,6 +855,28 @@ static void mes_v11_0_free_ucode_buffers(struct amdgpu_device *adev, (void **)&adev->mes.ucode_fw_ptr[pipe]); } +static void mes_v11_0_get_fw_version(struct amdgpu_device *adev) +{ + int pipe; + + /* get MES scheduler/KIQ versions */ + mutex_lock(&adev->srbm_mutex); + + for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) { + soc21_grbm_select(adev, 3, pipe, 0, 0); + + if (pipe == AMDGPU_MES_SCHED_PIPE) + adev->mes.sched_version = + RREG32_SOC15(GC, 0, regCP_MES_GP3_LO); + else if (pipe == AMDGPU_MES_KIQ_PIPE && adev->enable_mes_kiq) + adev->mes.kiq_version = + RREG32_SOC15(GC, 0, regCP_MES_GP3_LO); + } + + soc21_grbm_select(adev, 0, 0, 0, 0); + mutex_unlock(&adev->srbm_mutex); +} + static void mes_v11_0_enable(struct amdgpu_device *adev, bool enable) { uint64_t ucode_addr; @@ -1144,18 +1246,6 @@ static int mes_v11_0_queue_init(struct amdgpu_device *adev, mes_v11_0_queue_init_register(ring); } - /* get MES scheduler/KIQ versions */ - mutex_lock(&adev->srbm_mutex); - soc21_grbm_select(adev, 3, pipe, 0, 0); - - if (pipe == AMDGPU_MES_SCHED_PIPE) - adev->mes.sched_version = RREG32_SOC15(GC, 0, regCP_MES_GP3_LO); - else if (pipe == AMDGPU_MES_KIQ_PIPE && adev->enable_mes_kiq) - adev->mes.kiq_version = RREG32_SOC15(GC, 0, regCP_MES_GP3_LO); - - soc21_grbm_select(adev, 0, 0, 0, 0); - mutex_unlock(&adev->srbm_mutex); - return 0; } @@ -1402,15 +1492,24 @@ static int mes_v11_0_kiq_hw_init(struct amdgpu_device *adev) mes_v11_0_enable(adev, true); + mes_v11_0_get_fw_version(adev); + mes_v11_0_kiq_setting(&adev->gfx.kiq[0].ring); r = mes_v11_0_queue_init(adev, AMDGPU_MES_KIQ_PIPE); if (r) goto failure; - r = mes_v11_0_hw_init(adev); - if (r) - goto failure; + if ((adev->mes.sched_version & AMDGPU_MES_VERSION_MASK) >= 0x47) + adev->mes.enable_legacy_queue_map = true; + else + adev->mes.enable_legacy_queue_map = false; + + if (adev->mes.enable_legacy_queue_map) { + r = mes_v11_0_hw_init(adev); + if (r) + goto failure; + } return r; diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c index 47a73f6ae4da..e499b2857a01 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c @@ -1332,6 +1332,7 @@ static int mes_v12_0_sw_init(void *handle) adev->mes.funcs = &mes_v12_0_funcs; adev->mes.kiq_hw_init = &mes_v12_0_kiq_hw_init; adev->mes.kiq_hw_fini = &mes_v12_0_kiq_hw_fini; + adev->mes.enable_legacy_queue_map = true; adev->mes.event_log_size = AMDGPU_MES_LOG_BUFFER_SIZE; @@ -1488,9 +1489,11 @@ static int mes_v12_0_kiq_hw_init(struct amdgpu_device *adev) mes_v12_0_set_hw_resources_1(&adev->mes, AMDGPU_MES_KIQ_PIPE); } - r = mes_v12_0_hw_init(adev); - if (r) - goto failure; + if (adev->mes.enable_legacy_queue_map) { + r = mes_v12_0_hw_init(adev); + if (r) + goto failure; + } return r; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c index ce2a5d9f90d3..40c94c4cdd96 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c @@ -2464,11 +2464,14 @@ svm_range_unmap_from_cpu(struct mm_struct *mm, struct svm_range *prange, adev = pdd->dev->adev; /* Check and drain ih1 ring if cam not available */ - ih = &adev->irq.ih1; - checkpoint_wptr = amdgpu_ih_get_wptr(adev, ih); - if (ih->rptr != checkpoint_wptr) { - svms->checkpoint_ts[i] = amdgpu_ih_decode_iv_ts(adev, ih, checkpoint_wptr, -1); - continue; + if (adev->irq.ih1.ring_size) { + ih = &adev->irq.ih1; + checkpoint_wptr = amdgpu_ih_get_wptr(adev, ih); + if (ih->rptr != checkpoint_wptr) { + svms->checkpoint_ts[i] = + amdgpu_ih_decode_iv_ts(adev, ih, checkpoint_wptr, -1); + continue; + } } /* check if dev->irq.ih_soft is not empty */ diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c index 5f7b178ba870..a8d0d1b71723 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -1754,6 +1754,30 @@ static struct dml2_soc_bb *dm_dmub_get_vbios_bounding_box(struct amdgpu_device * return bb; } +static enum dmub_ips_disable_type dm_get_default_ips_mode( + struct amdgpu_device *adev) +{ + /* + * On DCN35 systems with Z8 enabled, it's possible for IPS2 + Z8 to + * cause a hard hang. A fix exists for newer PMFW. + * + * As a workaround, for non-fixed PMFW, force IPS1+RCG as the deepest + * IPS state in all cases, except for s0ix and all displays off (DPMS), + * where IPS2 is allowed. + * + * When checking pmfw version, use the major and minor only. + */ + if (amdgpu_ip_version(adev, DCE_HWIP, 0) == IP_VERSION(3, 5, 0) && + (adev->pm.fw_version & 0x00FFFF00) < 0x005D6300) + return DMUB_IPS_RCG_IN_ACTIVE_IPS2_IN_OFF; + + if (amdgpu_ip_version(adev, DCE_HWIP, 0) >= IP_VERSION(3, 5, 0)) + return DMUB_IPS_ENABLE; + + /* ASICs older than DCN35 do not have IPSs */ + return DMUB_IPS_DISABLE_ALL; +} + static int amdgpu_dm_init(struct amdgpu_device *adev) { struct dc_init_data init_data; @@ -1864,8 +1888,14 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) if (amdgpu_dc_debug_mask & DC_DISABLE_IPS) init_data.flags.disable_ips = DMUB_IPS_DISABLE_ALL; - else + else if (amdgpu_dc_debug_mask & DC_DISABLE_IPS_DYNAMIC) + init_data.flags.disable_ips = DMUB_IPS_DISABLE_DYNAMIC; + else if (amdgpu_dc_debug_mask & DC_DISABLE_IPS2_DYNAMIC) + init_data.flags.disable_ips = DMUB_IPS_RCG_IN_ACTIVE_IPS2_IN_OFF; + else if (amdgpu_dc_debug_mask & DC_FORCE_IPS_ENABLE) init_data.flags.disable_ips = DMUB_IPS_ENABLE; + else + init_data.flags.disable_ips = dm_get_default_ips_mode(adev); init_data.flags.disable_ips_in_vpb = 0; @@ -4507,7 +4537,7 @@ static void amdgpu_dm_backlight_set_level(struct amdgpu_display_manager *dm, struct amdgpu_dm_backlight_caps caps; struct dc_link *link; u32 brightness; - bool rc; + bool rc, reallow_idle = false; amdgpu_dm_update_backlight_caps(dm, bl_idx); caps = dm->backlight_caps[bl_idx]; @@ -4520,6 +4550,12 @@ static void amdgpu_dm_backlight_set_level(struct amdgpu_display_manager *dm, link = (struct dc_link *)dm->backlight_link[bl_idx]; /* Change brightness based on AUX property */ + mutex_lock(&dm->dc_lock); + if (dm->dc->caps.ips_support && dm->dc->ctx->dmub_srv->idle_allowed) { + dc_allow_idle_optimizations(dm->dc, false); + reallow_idle = true; + } + if (caps.aux_support) { rc = dc_link_set_backlight_level_nits(link, true, brightness, AUX_BL_DEFAULT_TRANSITION_TIME_MS); @@ -4531,6 +4567,11 @@ static void amdgpu_dm_backlight_set_level(struct amdgpu_display_manager *dm, DRM_DEBUG("DM: Failed to update backlight on eDP[%d]\n", bl_idx); } + if (dm->dc->caps.ips_support && reallow_idle) + dc_allow_idle_optimizations(dm->dc, true); + + mutex_unlock(&dm->dc_lock); + if (rc) dm->actual_brightness[bl_idx] = user_brightness; } @@ -6441,7 +6482,8 @@ static void apply_dsc_policy_for_stream(struct amdgpu_dm_connector *aconnector, dc_link_get_highest_encoding_format(aconnector->dc_link), &stream->timing.dsc_cfg)) { stream->timing.flags.DSC = 1; - DRM_DEBUG_DRIVER("%s: [%s] DSC is selected from SST RX\n", __func__, drm_connector->name); + DRM_DEBUG_DRIVER("%s: SST_DSC [%s] DSC is selected from SST RX\n", + __func__, drm_connector->name); } } else if (sink->link->dpcd_caps.dongle_type == DISPLAY_DONGLE_DP_HDMI_CONVERTER) { timing_bw_in_kbps = dc_bandwidth_in_kbps_from_timing(&stream->timing, @@ -6460,7 +6502,7 @@ static void apply_dsc_policy_for_stream(struct amdgpu_dm_connector *aconnector, dc_link_get_highest_encoding_format(aconnector->dc_link), &stream->timing.dsc_cfg)) { stream->timing.flags.DSC = 1; - DRM_DEBUG_DRIVER("%s: [%s] DSC is selected from DP-HDMI PCON\n", + DRM_DEBUG_DRIVER("%s: SST_DSC [%s] DSC is selected from DP-HDMI PCON\n", __func__, drm_connector->name); } } @@ -11637,7 +11679,7 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev, if (dc_resource_is_dsc_encoding_supported(dc)) { ret = compute_mst_dsc_configs_for_state(state, dm_state->context, vars); if (ret) { - drm_dbg_atomic(dev, "compute_mst_dsc_configs_for_state() failed\n"); + drm_dbg_atomic(dev, "MST_DSC compute_mst_dsc_configs_for_state() failed\n"); ret = -EINVAL; goto fail; } @@ -11658,7 +11700,7 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev, */ ret = drm_dp_mst_atomic_check(state); if (ret) { - drm_dbg_atomic(dev, "drm_dp_mst_atomic_check() failed\n"); + drm_dbg_atomic(dev, "MST drm_dp_mst_atomic_check() failed\n"); goto fail; } status = dc_validate_global_state(dc, dm_state->context, true); diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c index 165e010fe69c..50109d13d967 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c @@ -759,7 +759,7 @@ static uint8_t write_dsc_enable_synaptics_non_virtual_dpcd_mst( uint8_t ret = 0; drm_dbg_dp(aux->drm_dev, - "Configure DSC to non-virtual dpcd synaptics\n"); + "MST_DSC Configure DSC to non-virtual dpcd synaptics\n"); if (enable) { /* When DSC is enabled on previous boot and reboot with the hub, @@ -772,7 +772,7 @@ static uint8_t write_dsc_enable_synaptics_non_virtual_dpcd_mst( apply_synaptics_fifo_reset_wa(aux); ret = drm_dp_dpcd_write(aux, DP_DSC_ENABLE, &enable, 1); - DRM_INFO("Send DSC enable to synaptics\n"); + DRM_INFO("MST_DSC Send DSC enable to synaptics\n"); } else { /* Synaptics hub not support virtual dpcd, @@ -781,7 +781,7 @@ static uint8_t write_dsc_enable_synaptics_non_virtual_dpcd_mst( */ if (!stream->link->link_status.link_active) { ret = drm_dp_dpcd_write(aux, DP_DSC_ENABLE, &enable, 1); - DRM_INFO("Send DSC disable to synaptics\n"); + DRM_INFO("MST_DSC Send DSC disable to synaptics\n"); } } @@ -823,14 +823,14 @@ bool dm_helpers_dp_write_dsc_enable( DP_DSC_ENABLE, &enable_passthrough, 1); drm_dbg_dp(dev, - "Sent DSC pass-through enable to virtual dpcd port, ret = %u\n", + "MST_DSC Sent DSC pass-through enable to virtual dpcd port, ret = %u\n", ret); } ret = drm_dp_dpcd_write(aconnector->dsc_aux, DP_DSC_ENABLE, &enable_dsc, 1); drm_dbg_dp(dev, - "Sent DSC decoding enable to %s port, ret = %u\n", + "MST_DSC Sent DSC decoding enable to %s port, ret = %u\n", (port->passthrough_aux) ? "remote RX" : "virtual dpcd", ret); @@ -838,7 +838,7 @@ bool dm_helpers_dp_write_dsc_enable( ret = drm_dp_dpcd_write(aconnector->dsc_aux, DP_DSC_ENABLE, &enable_dsc, 1); drm_dbg_dp(dev, - "Sent DSC decoding disable to %s port, ret = %u\n", + "MST_DSC Sent DSC decoding disable to %s port, ret = %u\n", (port->passthrough_aux) ? "remote RX" : "virtual dpcd", ret); @@ -848,7 +848,7 @@ bool dm_helpers_dp_write_dsc_enable( DP_DSC_ENABLE, &enable_passthrough, 1); drm_dbg_dp(dev, - "Sent DSC pass-through disable to virtual dpcd port, ret = %u\n", + "MST_DSC Sent DSC pass-through disable to virtual dpcd port, ret = %u\n", ret); } } @@ -858,12 +858,12 @@ bool dm_helpers_dp_write_dsc_enable( if (stream->sink->link->dpcd_caps.dongle_type == DISPLAY_DONGLE_NONE) { ret = dm_helpers_dp_write_dpcd(ctx, stream->link, DP_DSC_ENABLE, &enable_dsc, 1); drm_dbg_dp(dev, - "Send DSC %s to SST RX\n", + "SST_DSC Send DSC %s to SST RX\n", enable_dsc ? "enable" : "disable"); } else if (stream->sink->link->dpcd_caps.dongle_type == DISPLAY_DONGLE_DP_HDMI_CONVERTER) { ret = dm_helpers_dp_write_dpcd(ctx, stream->link, DP_DSC_ENABLE, &enable_dsc, 1); drm_dbg_dp(dev, - "Send DSC %s to DP-HDMI PCON\n", + "SST_DSC Send DSC %s to DP-HDMI PCON\n", enable_dsc ? "enable" : "disable"); } } diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c index 2e9f6da1acdc..6b5eed37532b 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c @@ -253,7 +253,7 @@ static bool validate_dsc_caps_on_connector(struct amdgpu_dm_connector *aconnecto aconnector->dsc_aux = &aconnector->mst_root->dm_dp_aux.aux; /* synaptics cascaded MST hub case */ - if (!aconnector->dsc_aux && is_synaptics_cascaded_panamera(aconnector->dc_link, port)) + if (is_synaptics_cascaded_panamera(aconnector->dc_link, port)) aconnector->dsc_aux = port->mgr->aux; if (!aconnector->dsc_aux) @@ -578,6 +578,8 @@ dm_dp_add_mst_connector(struct drm_dp_mst_topology_mgr *mgr, if (!aconnector) return NULL; + DRM_DEBUG_DRIVER("%s: Create aconnector 0x%p for port 0x%p\n", __func__, aconnector, port); + connector = &aconnector->base; aconnector->mst_output_port = port; aconnector->mst_root = master; @@ -872,11 +874,11 @@ static void set_dsc_configs_from_fairness_vars(struct dsc_mst_fairness_params *p if (params[i].sink) { if (params[i].sink->sink_signal != SIGNAL_TYPE_VIRTUAL && params[i].sink->sink_signal != SIGNAL_TYPE_NONE) - DRM_DEBUG_DRIVER("%s i=%d dispname=%s\n", __func__, i, + DRM_DEBUG_DRIVER("MST_DSC %s i=%d dispname=%s\n", __func__, i, params[i].sink->edid_caps.display_name); } - DRM_DEBUG_DRIVER("dsc=%d bits_per_pixel=%d pbn=%d\n", + DRM_DEBUG_DRIVER("MST_DSC dsc=%d bits_per_pixel=%d pbn=%d\n", params[i].timing->flags.DSC, params[i].timing->dsc_cfg.bits_per_pixel, vars[i + k].pbn); @@ -1054,6 +1056,7 @@ static int try_disable_dsc(struct drm_atomic_state *state, if (next_index == -1) break; + DRM_DEBUG_DRIVER("MST_DSC index #%d, try no compression\n", next_index); vars[next_index].pbn = kbps_to_peak_pbn(params[next_index].bw_range.stream_kbps, fec_overhead_multiplier_x1000); ret = drm_dp_atomic_find_time_slots(state, params[next_index].port->mgr, @@ -1064,10 +1067,12 @@ static int try_disable_dsc(struct drm_atomic_state *state, ret = drm_dp_mst_atomic_check(state); if (ret == 0) { + DRM_DEBUG_DRIVER("MST_DSC index #%d, greedily disable dsc\n", next_index); vars[next_index].dsc_enabled = false; vars[next_index].bpp_x16 = 0; } else { - vars[next_index].pbn = kbps_to_peak_pbn(params[next_index].bw_range.stream_kbps, fec_overhead_multiplier_x1000); + DRM_DEBUG_DRIVER("MST_DSC index #%d, restore minimum compression\n", next_index); + vars[next_index].pbn = kbps_to_peak_pbn(params[next_index].bw_range.max_kbps, fec_overhead_multiplier_x1000); ret = drm_dp_atomic_find_time_slots(state, params[next_index].port->mgr, params[next_index].port, @@ -1082,6 +1087,15 @@ static int try_disable_dsc(struct drm_atomic_state *state, return 0; } +static void log_dsc_params(int count, struct dsc_mst_fairness_vars *vars, int k) +{ + int i; + + for (i = 0; i < count; i++) + DRM_DEBUG_DRIVER("MST_DSC DSC params: stream #%d --- dsc_enabled = %d, bpp_x16 = %d, pbn = %d\n", + i, vars[i + k].dsc_enabled, vars[i + k].bpp_x16, vars[i + k].pbn); +} + static int compute_mst_dsc_configs_for_link(struct drm_atomic_state *state, struct dc_state *dc_state, struct dc_link *dc_link, @@ -1104,6 +1118,7 @@ static int compute_mst_dsc_configs_for_link(struct drm_atomic_state *state, return PTR_ERR(mst_state); /* Set up params */ + DRM_DEBUG_DRIVER("%s: MST_DSC Set up params for %d streams\n", __func__, dc_state->stream_count); for (i = 0; i < dc_state->stream_count; i++) { struct dc_dsc_policy dsc_policy = {0}; @@ -1145,6 +1160,9 @@ static int compute_mst_dsc_configs_for_link(struct drm_atomic_state *state, params[count].bw_range.stream_kbps = dc_bandwidth_in_kbps_from_timing(&stream->timing, dc_link_get_highest_encoding_format(dc_link)); + DRM_DEBUG_DRIVER("MST_DSC #%d stream 0x%p - max_kbps = %u, min_kbps = %u, uncompressed_kbps = %u\n", + count, stream, params[count].bw_range.max_kbps, params[count].bw_range.min_kbps, + params[count].bw_range.stream_kbps); count++; } @@ -1159,6 +1177,7 @@ static int compute_mst_dsc_configs_for_link(struct drm_atomic_state *state, *link_vars_start_index += count; /* Try no compression */ + DRM_DEBUG_DRIVER("MST_DSC Try no compression\n"); for (i = 0; i < count; i++) { vars[i + k].aconnector = params[i].aconnector; vars[i + k].pbn = kbps_to_peak_pbn(params[i].bw_range.stream_kbps, fec_overhead_multiplier_x1000); @@ -1177,7 +1196,10 @@ static int compute_mst_dsc_configs_for_link(struct drm_atomic_state *state, return ret; } + log_dsc_params(count, vars, k); + /* Try max compression */ + DRM_DEBUG_DRIVER("MST_DSC Try max compression\n"); for (i = 0; i < count; i++) { if (params[i].compression_possible && params[i].clock_force_enable != DSC_CLK_FORCE_DISABLE) { vars[i + k].pbn = kbps_to_peak_pbn(params[i].bw_range.min_kbps, fec_overhead_multiplier_x1000); @@ -1201,14 +1223,26 @@ static int compute_mst_dsc_configs_for_link(struct drm_atomic_state *state, if (ret != 0) return ret; + log_dsc_params(count, vars, k); + /* Optimize degree of compression */ + DRM_DEBUG_DRIVER("MST_DSC Try optimize compression\n"); ret = increase_dsc_bpp(state, mst_state, dc_link, params, vars, count, k); - if (ret < 0) + if (ret < 0) { + DRM_DEBUG_DRIVER("MST_DSC Failed to optimize compression\n"); return ret; + } + log_dsc_params(count, vars, k); + + DRM_DEBUG_DRIVER("MST_DSC Try disable compression\n"); ret = try_disable_dsc(state, dc_link, params, vars, count, k); - if (ret < 0) + if (ret < 0) { + DRM_DEBUG_DRIVER("MST_DSC Failed to disable compression\n"); return ret; + } + + log_dsc_params(count, vars, k); set_dsc_configs_from_fairness_vars(params, vars, count, k); @@ -1230,17 +1264,19 @@ static bool is_dsc_need_re_compute( /* only check phy used by dsc mst branch */ if (dc_link->type != dc_connection_mst_branch) - return false; + goto out; /* add a check for older MST DSC with no virtual DPCDs */ if (needs_dsc_aux_workaround(dc_link) && (!(dc_link->dpcd_caps.dsc_caps.dsc_basic_caps.fields.dsc_support.DSC_SUPPORT || dc_link->dpcd_caps.dsc_caps.dsc_basic_caps.fields.dsc_support.DSC_PASSTHROUGH_SUPPORT))) - return false; + goto out; for (i = 0; i < MAX_PIPES; i++) stream_on_link[i] = NULL; + DRM_DEBUG_DRIVER("%s: MST_DSC check on %d streams in new dc_state\n", __func__, dc_state->stream_count); + /* check if there is mode change in new request */ for (i = 0; i < dc_state->stream_count; i++) { struct drm_crtc_state *new_crtc_state; @@ -1250,6 +1286,8 @@ static bool is_dsc_need_re_compute( if (!stream) continue; + DRM_DEBUG_DRIVER("%s:%d MST_DSC checking #%d stream 0x%p\n", __func__, __LINE__, i, stream); + /* check if stream using the same link for mst */ if (stream->link != dc_link) continue; @@ -1262,8 +1300,11 @@ static bool is_dsc_need_re_compute( new_stream_on_link_num++; new_conn_state = drm_atomic_get_new_connector_state(state, &aconnector->base); - if (!new_conn_state) + if (!new_conn_state) { + DRM_DEBUG_DRIVER("%s:%d MST_DSC no new_conn_state for stream 0x%p, aconnector 0x%p\n", + __func__, __LINE__, stream, aconnector); continue; + } if (IS_ERR(new_conn_state)) continue; @@ -1272,21 +1313,36 @@ static bool is_dsc_need_re_compute( continue; new_crtc_state = drm_atomic_get_new_crtc_state(state, new_conn_state->crtc); - if (!new_crtc_state) + if (!new_crtc_state) { + DRM_DEBUG_DRIVER("%s:%d MST_DSC no new_crtc_state for crtc of stream 0x%p, aconnector 0x%p\n", + __func__, __LINE__, stream, aconnector); continue; + } if (IS_ERR(new_crtc_state)) continue; if (new_crtc_state->enable && new_crtc_state->active) { if (new_crtc_state->mode_changed || new_crtc_state->active_changed || - new_crtc_state->connectors_changed) - return true; + new_crtc_state->connectors_changed) { + DRM_DEBUG_DRIVER("%s:%d MST_DSC dsc recompte required." + "stream 0x%p in new dc_state\n", + __func__, __LINE__, stream); + is_dsc_need_re_compute = true; + goto out; + } } } - if (new_stream_on_link_num == 0) - return false; + if (new_stream_on_link_num == 0) { + DRM_DEBUG_DRIVER("%s:%d MST_DSC no mode change request for streams in new dc_state\n", + __func__, __LINE__); + is_dsc_need_re_compute = false; + goto out; + } + + DRM_DEBUG_DRIVER("%s: MST_DSC check on %d streams in current dc_state\n", + __func__, dc->current_state->stream_count); /* check current_state if there stream on link but it is not in * new request state @@ -1310,11 +1366,18 @@ static bool is_dsc_need_re_compute( if (j == new_stream_on_link_num) { /* not in new state */ + DRM_DEBUG_DRIVER("%s:%d MST_DSC dsc recompute required." + "stream 0x%p in current dc_state but not in new dc_state\n", + __func__, __LINE__, stream); is_dsc_need_re_compute = true; break; } } +out: + DRM_DEBUG_DRIVER("%s: MST_DSC dsc recompute %s\n", + __func__, is_dsc_need_re_compute ? "required" : "not required"); + return is_dsc_need_re_compute; } @@ -1343,6 +1406,9 @@ int compute_mst_dsc_configs_for_state(struct drm_atomic_state *state, aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context; + DRM_DEBUG_DRIVER("%s: MST_DSC compute mst dsc configs for stream 0x%p, aconnector 0x%p\n", + __func__, stream, aconnector); + if (!aconnector || !aconnector->dc_sink || !aconnector->mst_output_port) continue; @@ -1375,8 +1441,11 @@ int compute_mst_dsc_configs_for_state(struct drm_atomic_state *state, stream = dc_state->streams[i]; if (stream->timing.flags.DSC == 1) - if (dc_stream_add_dsc_to_resource(stream->ctx->dc, dc_state, stream) != DC_OK) + if (dc_stream_add_dsc_to_resource(stream->ctx->dc, dc_state, stream) != DC_OK) { + DRM_DEBUG_DRIVER("%s:%d MST_DSC Failed to request dsc hw resource for stream 0x%p\n", + __func__, __LINE__, stream); return -EINVAL; + } } return ret; @@ -1405,6 +1474,9 @@ static int pre_compute_mst_dsc_configs_for_state(struct drm_atomic_state *state, aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context; + DRM_DEBUG_DRIVER("MST_DSC pre compute mst dsc configs for #%d stream 0x%p, aconnector 0x%p\n", + i, stream, aconnector); + if (!aconnector || !aconnector->dc_sink || !aconnector->mst_output_port) continue; @@ -1494,12 +1566,12 @@ int pre_validate_dsc(struct drm_atomic_state *state, int ret = 0; if (!is_dsc_precompute_needed(state)) { - DRM_INFO_ONCE("DSC precompute is not needed.\n"); + DRM_INFO_ONCE("%s:%d MST_DSC dsc precompute is not needed\n", __func__, __LINE__); return 0; } ret = dm_atomic_get_state(state, dm_state_ptr); if (ret != 0) { - DRM_INFO_ONCE("dm_atomic_get_state() failed\n"); + DRM_INFO_ONCE("%s:%d MST_DSC dm_atomic_get_state() failed\n", __func__, __LINE__); return ret; } dm_state = *dm_state_ptr; @@ -1553,7 +1625,8 @@ int pre_validate_dsc(struct drm_atomic_state *state, ret = pre_compute_mst_dsc_configs_for_state(state, local_dc_state, vars); if (ret != 0) { - DRM_INFO_ONCE("pre_compute_mst_dsc_configs_for_state() failed\n"); + DRM_INFO_ONCE("%s:%d MST_DSC dsc pre_compute_mst_dsc_configs_for_state() failed\n", + __func__, __LINE__); ret = -EINVAL; goto clean_exit; } @@ -1567,12 +1640,15 @@ int pre_validate_dsc(struct drm_atomic_state *state, if (local_dc_state->streams[i] && dc_is_timing_changed(stream, local_dc_state->streams[i])) { - DRM_INFO_ONCE("crtc[%d] needs mode_changed\n", i); + DRM_INFO_ONCE("%s:%d MST_DSC crtc[%d] needs mode_change\n", __func__, __LINE__, i); } else { int ind = find_crtc_index_in_state_by_stream(state, stream); - if (ind >= 0) + if (ind >= 0) { + DRM_INFO_ONCE("%s:%d MST_DSC no mode changed for stream 0x%p\n", + __func__, __LINE__, stream); state->crtcs[ind].new_state->mode_changed = 0; + } } } clean_exit: @@ -1697,7 +1773,7 @@ enum dc_status dm_dp_mst_is_port_support_mode( end_to_end_bw_in_kbps = min(root_link_bw_in_kbps, virtual_channel_bw_in_kbps); if (stream_kbps <= end_to_end_bw_in_kbps) { - DRM_DEBUG_DRIVER("No DSC needed. End-to-end bw sufficient."); + DRM_DEBUG_DRIVER("MST_DSC no dsc required. End-to-end bw sufficient\n"); return DC_OK; } @@ -1710,7 +1786,8 @@ enum dc_status dm_dp_mst_is_port_support_mode( /*capable of dsc passthough. dsc bitstream along the entire path*/ if (aconnector->mst_output_port->passthrough_aux) { if (bw_range.min_kbps > end_to_end_bw_in_kbps) { - DRM_DEBUG_DRIVER("DSC passthrough. Max dsc compression can't fit into end-to-end bw\n"); + DRM_DEBUG_DRIVER("MST_DSC dsc passthrough and decode at endpoint" + "Max dsc compression bw can't fit into end-to-end bw\n"); return DC_FAIL_BANDWIDTH_VALIDATE; } } else { @@ -1721,7 +1798,8 @@ enum dc_status dm_dp_mst_is_port_support_mode( /*Get last DP link BW capability*/ if (dp_get_link_current_set_bw(&aconnector->mst_output_port->aux, &end_link_bw)) { if (stream_kbps > end_link_bw) { - DRM_DEBUG_DRIVER("DSC decode at last link. Mode required bw can't fit into available bw\n"); + DRM_DEBUG_DRIVER("MST_DSC dsc decode at last link." + "Mode required bw can't fit into last link\n"); return DC_FAIL_BANDWIDTH_VALIDATE; } } @@ -1734,7 +1812,8 @@ enum dc_status dm_dp_mst_is_port_support_mode( virtual_channel_bw_in_kbps = kbps_from_pbn(immediate_upstream_port->full_pbn); virtual_channel_bw_in_kbps = min(root_link_bw_in_kbps, virtual_channel_bw_in_kbps); if (bw_range.min_kbps > virtual_channel_bw_in_kbps) { - DRM_DEBUG_DRIVER("DSC decode at last link. Max dsc compression can't fit into MST available bw\n"); + DRM_DEBUG_DRIVER("MST_DSC dsc decode at last link." + "Max dsc compression can't fit into MST available bw\n"); return DC_FAIL_BANDWIDTH_VALIDATE; } } @@ -1751,9 +1830,9 @@ enum dc_status dm_dp_mst_is_port_support_mode( dc_link_get_highest_encoding_format(stream->link), &stream->timing.dsc_cfg)) { stream->timing.flags.DSC = 1; - DRM_DEBUG_DRIVER("Require dsc and dsc config found\n"); + DRM_DEBUG_DRIVER("MST_DSC require dsc and dsc config found\n"); } else { - DRM_DEBUG_DRIVER("Require dsc but can't find appropriate dsc config\n"); + DRM_DEBUG_DRIVER("MST_DSC require dsc but can't find appropriate dsc config\n"); return DC_FAIL_BANDWIDTH_VALIDATE; } @@ -1775,11 +1854,11 @@ enum dc_status dm_dp_mst_is_port_support_mode( if (branch_max_throughput_mps != 0 && ((stream->timing.pix_clk_100hz / 10) > branch_max_throughput_mps * 1000)) { - DRM_DEBUG_DRIVER("DSC is required but max throughput mps fails"); + DRM_DEBUG_DRIVER("MST_DSC require dsc but max throughput mps fails\n"); return DC_FAIL_BANDWIDTH_VALIDATE; } } else { - DRM_DEBUG_DRIVER("DSC is required but can't find common dsc config."); + DRM_DEBUG_DRIVER("MST_DSC require dsc but can't find common dsc config\n"); return DC_FAIL_BANDWIDTH_VALIDATE; } #endif diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_plane.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_plane.c index a573a6639898..25f63b2e7a8e 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_plane.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_plane.c @@ -1283,6 +1283,7 @@ int amdgpu_dm_plane_get_cursor_position(struct drm_plane *plane, struct drm_crtc struct dc_cursor_position *position) { struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); + struct amdgpu_device *adev = drm_to_adev(plane->dev); int x, y; int xorigin = 0, yorigin = 0; @@ -1314,12 +1315,14 @@ int amdgpu_dm_plane_get_cursor_position(struct drm_plane *plane, struct drm_crtc y = 0; } position->enable = true; - position->translate_by_source = true; position->x = x; position->y = y; position->x_hotspot = xorigin; position->y_hotspot = yorigin; + if (amdgpu_ip_version(adev, DCE_HWIP, 0) < IP_VERSION(4, 0, 1)) + position->translate_by_source = true; + return 0; } diff --git a/drivers/gpu/drm/amd/display/dc/bios/command_table2.c b/drivers/gpu/drm/amd/display/dc/bios/command_table2.c index 4254bdfefe38..7d18f372ce7a 100644 --- a/drivers/gpu/drm/amd/display/dc/bios/command_table2.c +++ b/drivers/gpu/drm/amd/display/dc/bios/command_table2.c @@ -227,7 +227,7 @@ static void init_transmitter_control(struct bios_parser *bp) uint8_t frev; uint8_t crev = 0; - if (!BIOS_CMD_TABLE_REVISION(dig1transmittercontrol, frev, crev)) + if (!BIOS_CMD_TABLE_REVISION(dig1transmittercontrol, frev, crev) && (bp->base.ctx->dc->ctx->dce_version <= DCN_VERSION_2_0)) BREAK_TO_DEBUGGER(); switch (crev) { diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c index 0ce9b40dfc68..97164b5585a8 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c @@ -305,9 +305,6 @@ void dcn35_update_clocks(struct clk_mgr *clk_mgr_base, if (new_clocks->dtbclk_en && !new_clocks->ref_dtbclk_khz) new_clocks->ref_dtbclk_khz = 600000; - if (dc->debug.min_disp_clk_khz > 0 && new_clocks->dispclk_khz < dc->debug.min_disp_clk_khz) - new_clocks->dispclk_khz = dc->debug.min_disp_clk_khz; - /* * if it is safe to lower, but we are already in the lower state, we don't have to do anything * also if safe to lower is false, we just go in the higher state @@ -385,6 +382,9 @@ void dcn35_update_clocks(struct clk_mgr *clk_mgr_base, if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr_base->clks.dispclk_khz)) { dcn35_disable_otg_wa(clk_mgr_base, context, safe_to_lower, true); + if (dc->debug.min_disp_clk_khz > 0 && new_clocks->dispclk_khz < dc->debug.min_disp_clk_khz) + new_clocks->dispclk_khz = dc->debug.min_disp_clk_khz; + clk_mgr_base->clks.dispclk_khz = new_clocks->dispclk_khz; dcn35_smu_set_dispclk(clk_mgr, clk_mgr_base->clks.dispclk_khz); dcn35_disable_otg_wa(clk_mgr_base, context, safe_to_lower, false); @@ -1100,7 +1100,7 @@ void dcn35_clk_mgr_construct( clk_mgr->smu_wm_set.wm_set = (struct dcn35_watermarks *)dm_helpers_allocate_gpu_mem( clk_mgr->base.base.ctx, - DC_MEM_ALLOC_TYPE_FRAME_BUFFER, + DC_MEM_ALLOC_TYPE_GART, sizeof(struct dcn35_watermarks), &clk_mgr->smu_wm_set.mc_address.quad_part); @@ -1112,7 +1112,7 @@ void dcn35_clk_mgr_construct( smu_dpm_clks.dpm_clks = (DpmClocks_t_dcn35 *)dm_helpers_allocate_gpu_mem( clk_mgr->base.base.ctx, - DC_MEM_ALLOC_TYPE_FRAME_BUFFER, + DC_MEM_ALLOC_TYPE_GART, sizeof(DpmClocks_t_dcn35), &smu_dpm_clks.mc_address.quad_part); @@ -1209,7 +1209,7 @@ void dcn35_clk_mgr_construct( } if (smu_dpm_clks.dpm_clks && smu_dpm_clks.mc_address.quad_part != 0) - dm_helpers_free_gpu_mem(clk_mgr->base.base.ctx, DC_MEM_ALLOC_TYPE_FRAME_BUFFER, + dm_helpers_free_gpu_mem(clk_mgr->base.base.ctx, DC_MEM_ALLOC_TYPE_GART, smu_dpm_clks.dpm_clks); if (ctx->dc->config.disable_ips != DMUB_IPS_DISABLE_ALL) { diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c index 01ea3a31e54d..8cfc5f435937 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c @@ -1366,9 +1366,6 @@ static void dcn401_update_clocks(struct clk_mgr *clk_mgr_base, unsigned int num_steps = 0; - if (dc->work_arounds.skip_clock_update) - return; - if (dc->debug.enable_legacy_clock_update) { dcn401_update_clocks_legacy(clk_mgr_base, context, safe_to_lower); return; diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c index c8dabb081b3d..ae788154896c 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc.c @@ -2690,6 +2690,9 @@ static enum surface_update_type check_update_surfaces_for_stream( stream_update->vrr_active_variable || stream_update->vrr_active_fixed)) su_flags->bits.fams_changed = 1; + if (stream_update->scaler_sharpener_update) + su_flags->bits.scaler_sharpener = 1; + if (su_flags->raw != 0) overall_type = UPDATE_TYPE_FULL; @@ -3022,6 +3025,8 @@ static void copy_stream_update_to_stream(struct dc *dc, update->dsc_config = NULL; } } + if (update->scaler_sharpener_update) + stream->scaler_sharpener_update = *update->scaler_sharpener_update; } static void backup_planes_and_stream_state( @@ -4713,7 +4718,8 @@ static bool full_update_required(struct dc *dc, stream_update->func_shaper || stream_update->lut3d_func || stream_update->pending_test_pattern || - stream_update->crtc_timing_adjust)) + stream_update->crtc_timing_adjust || + stream_update->scaler_sharpener_update)) return true; if (stream) { @@ -5161,6 +5167,8 @@ void dc_set_power_state(struct dc *dc, enum dc_acpi_cm_power_state power_state) dc_z10_restore(dc); + dc_dmub_srv_notify_fw_dc_power_state(dc->ctx->dmub_srv, power_state); + dc->hwss.init_hw(dc); if (dc->hwss.init_sys_ctx != NULL && @@ -5172,6 +5180,8 @@ void dc_set_power_state(struct dc *dc, enum dc_acpi_cm_power_state power_state) default: ASSERT(dc->current_state->stream_count == 0); + dc_dmub_srv_notify_fw_dc_power_state(dc->ctx->dmub_srv, power_state); + dc_state_destruct(dc->current_state); break; diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c index ef585a89847b..c7599c40d4be 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c @@ -1506,8 +1506,6 @@ bool resource_build_scaling_params(struct pipe_ctx *pipe_ctx) pipe_ctx->plane_res.scl_data.lb_params.depth = LB_PIXEL_DEPTH_30BPP; pipe_ctx->plane_res.scl_data.lb_params.alpha_en = plane_state->per_pixel_alpha; - spl_out->scl_data.h_active = pipe_ctx->plane_res.scl_data.h_active; - spl_out->scl_data.v_active = pipe_ctx->plane_res.scl_data.v_active; // Convert pipe_ctx to respective input params for SPL translate_SPL_in_params_from_pipe_ctx(pipe_ctx, spl_in); diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h index 5bbc7d2daca6..4c94dd38be4b 100644 --- a/drivers/gpu/drm/amd/display/dc/dc.h +++ b/drivers/gpu/drm/amd/display/dc/dc.h @@ -55,7 +55,7 @@ struct aux_payload; struct set_config_cmd_payload; struct dmub_notification; -#define DC_VER "3.2.297" +#define DC_VER "3.2.299" #define MAX_SURFACES 3 #define MAX_PLANES 6 @@ -761,7 +761,8 @@ union dpia_debug_options { uint32_t extend_aux_rd_interval:1; /* bit 2 */ uint32_t disable_mst_dsc_work_around:1; /* bit 3 */ uint32_t enable_force_tbt3_work_around:1; /* bit 4 */ - uint32_t reserved:27; + uint32_t disable_usb4_pm_support:1; /* bit 5 */ + uint32_t reserved:26; } bits; uint32_t raw; }; @@ -1051,6 +1052,7 @@ struct dc_debug_options { unsigned int disable_spl; unsigned int force_easf; unsigned int force_sharpness; + unsigned int force_sharpness_level; unsigned int force_lls; bool notify_dpia_hr_bw; bool enable_ips_visual_confirm; @@ -1347,7 +1349,7 @@ struct dc_plane_state { enum mpcc_movable_cm_location mcm_location; struct dc_csc_transform cursor_csc_color_matrix; bool adaptive_sharpness_en; - unsigned int sharpnessX1000; + int sharpness_level; enum linear_light_scaling linear_light_scaling; }; diff --git a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c index b1265124608b..1e7de0f03290 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c +++ b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c @@ -1476,7 +1476,7 @@ static void dc_dmub_srv_exit_low_power_state(const struct dc *dc) ips2_exit_count); } -void dc_dmub_srv_set_power_state(struct dc_dmub_srv *dc_dmub_srv, enum dc_acpi_cm_power_state powerState) +void dc_dmub_srv_set_power_state(struct dc_dmub_srv *dc_dmub_srv, enum dc_acpi_cm_power_state power_state) { struct dmub_srv *dmub; @@ -1485,12 +1485,38 @@ void dc_dmub_srv_set_power_state(struct dc_dmub_srv *dc_dmub_srv, enum dc_acpi_c dmub = dc_dmub_srv->dmub; - if (powerState == DC_ACPI_CM_POWER_STATE_D0) + if (power_state == DC_ACPI_CM_POWER_STATE_D0) dmub_srv_set_power_state(dmub, DMUB_POWER_STATE_D0); else dmub_srv_set_power_state(dmub, DMUB_POWER_STATE_D3); } +void dc_dmub_srv_notify_fw_dc_power_state(struct dc_dmub_srv *dc_dmub_srv, + enum dc_acpi_cm_power_state power_state) +{ + union dmub_rb_cmd cmd; + + if (!dc_dmub_srv) + return; + + memset(&cmd, 0, sizeof(cmd)); + + cmd.idle_opt_set_dc_power_state.header.type = DMUB_CMD__IDLE_OPT; + cmd.idle_opt_set_dc_power_state.header.sub_type = DMUB_CMD__IDLE_OPT_SET_DC_POWER_STATE; + cmd.idle_opt_set_dc_power_state.header.payload_bytes = + sizeof(cmd.idle_opt_set_dc_power_state) - sizeof(cmd.idle_opt_set_dc_power_state.header); + + if (power_state == DC_ACPI_CM_POWER_STATE_D0) { + cmd.idle_opt_set_dc_power_state.data.power_state = DMUB_IDLE_OPT_DC_POWER_STATE_D0; + } else if (power_state == DC_ACPI_CM_POWER_STATE_D3) { + cmd.idle_opt_set_dc_power_state.data.power_state = DMUB_IDLE_OPT_DC_POWER_STATE_D3; + } else { + cmd.idle_opt_set_dc_power_state.data.power_state = DMUB_IDLE_OPT_DC_POWER_STATE_UNKNOWN; + } + + dc_wake_and_execute_dmub_cmd(dc_dmub_srv->ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT); +} + bool dc_dmub_srv_should_detect(struct dc_dmub_srv *dc_dmub_srv) { volatile const struct dmub_shared_state_ips_fw *ips_fw; diff --git a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.h b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.h index 580940222777..42f0cb672d8b 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.h +++ b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.h @@ -109,7 +109,29 @@ bool dc_dmub_srv_is_hw_pwr_up(struct dc_dmub_srv *dc_dmub_srv, bool wait); void dc_dmub_srv_apply_idle_power_optimizations(const struct dc *dc, bool allow_idle); -void dc_dmub_srv_set_power_state(struct dc_dmub_srv *dc_dmub_srv, enum dc_acpi_cm_power_state powerState); +/** + * dc_dmub_srv_set_power_state() - Sets the power state for DMUB service. + * + * Controls whether messaging the DMCUB or interfacing with it via HW register + * interaction is permittable. + * + * @dc_dmub_srv - The DC DMUB service pointer + * @power_state - the DC power state + */ +void dc_dmub_srv_set_power_state(struct dc_dmub_srv *dc_dmub_srv, enum dc_acpi_cm_power_state power_state); + +/** + * dc_dmub_srv_notify_fw_dc_power_state() - Notifies firmware of the DC power state. + * + * Differs from dc_dmub_srv_set_power_state in that it needs to access HW in order + * to message DMCUB of the state transition. Should come after the D0 exit and + * before D3 set power state. + * + * @dc_dmub_srv - The DC DMUB service pointer + * @power_state - the DC power state + */ +void dc_dmub_srv_notify_fw_dc_power_state(struct dc_dmub_srv *dc_dmub_srv, + enum dc_acpi_cm_power_state power_state); /** * @dc_dmub_srv_should_detect() - Checks if link detection is required. diff --git a/drivers/gpu/drm/amd/display/dc/dc_spl_translate.c b/drivers/gpu/drm/amd/display/dc/dc_spl_translate.c index 8f85a1db5eba..cd6de93eb91c 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_spl_translate.c +++ b/drivers/gpu/drm/amd/display/dc/dc_spl_translate.c @@ -42,26 +42,26 @@ static void populate_spltaps_from_taps(struct spl_taps *spl_scaling_quality, static void populate_taps_from_spltaps(struct scaling_taps *scaling_quality, const struct spl_taps *spl_scaling_quality) { - scaling_quality->h_taps_c = spl_scaling_quality->h_taps_c; - scaling_quality->h_taps = spl_scaling_quality->h_taps; - scaling_quality->v_taps_c = spl_scaling_quality->v_taps_c; - scaling_quality->v_taps = spl_scaling_quality->v_taps; + scaling_quality->h_taps_c = spl_scaling_quality->h_taps_c + 1; + scaling_quality->h_taps = spl_scaling_quality->h_taps + 1; + scaling_quality->v_taps_c = spl_scaling_quality->v_taps_c + 1; + scaling_quality->v_taps = spl_scaling_quality->v_taps + 1; } static void populate_ratios_from_splratios(struct scaling_ratios *ratios, - const struct spl_ratios *spl_ratios) + const struct ratio *spl_ratios) { - ratios->horz = spl_ratios->horz; - ratios->vert = spl_ratios->vert; - ratios->horz_c = spl_ratios->horz_c; - ratios->vert_c = spl_ratios->vert_c; + ratios->horz = dc_fixpt_from_ux_dy(spl_ratios->h_scale_ratio >> 5, 3, 19); + ratios->vert = dc_fixpt_from_ux_dy(spl_ratios->v_scale_ratio >> 5, 3, 19); + ratios->horz_c = dc_fixpt_from_ux_dy(spl_ratios->h_scale_ratio_c >> 5, 3, 19); + ratios->vert_c = dc_fixpt_from_ux_dy(spl_ratios->v_scale_ratio_c >> 5, 3, 19); } static void populate_inits_from_splinits(struct scl_inits *inits, - const struct spl_inits *spl_inits) + const struct init *spl_inits) { - inits->h = spl_inits->h; - inits->v = spl_inits->v; - inits->h_c = spl_inits->h_c; - inits->v_c = spl_inits->v_c; + inits->h = dc_fixpt_from_int_dy(spl_inits->h_filter_init_int, spl_inits->h_filter_init_frac >> 5, 0, 19); + inits->v = dc_fixpt_from_int_dy(spl_inits->v_filter_init_int, spl_inits->v_filter_init_frac >> 5, 0, 19); + inits->h_c = dc_fixpt_from_int_dy(spl_inits->h_filter_init_int_c, spl_inits->h_filter_init_frac_c >> 5, 0, 19); + inits->v_c = dc_fixpt_from_int_dy(spl_inits->v_filter_init_int_c, spl_inits->v_filter_init_frac_c >> 5, 0, 19); } /// @brief Translate SPL input parameters from pipe context /// @param pipe_ctx @@ -139,24 +139,36 @@ void translate_SPL_in_params_from_pipe_ctx(struct pipe_ctx *pipe_ctx, struct spl else if (pipe_ctx->stream->ctx->dc->debug.force_easf == 2) spl_in->disable_easf = true; /* Translate adaptive sharpening preference */ - if (pipe_ctx->stream->ctx->dc->debug.force_sharpness > 0) { - spl_in->adaptive_sharpness.enable = (pipe_ctx->stream->ctx->dc->debug.force_sharpness > 1) ? true : false; - if (pipe_ctx->stream->ctx->dc->debug.force_sharpness == 2) - spl_in->adaptive_sharpness.sharpness = SHARPNESS_LOW; - else if (pipe_ctx->stream->ctx->dc->debug.force_sharpness == 3) - spl_in->adaptive_sharpness.sharpness = SHARPNESS_MID; - else if (pipe_ctx->stream->ctx->dc->debug.force_sharpness >= 4) - spl_in->adaptive_sharpness.sharpness = SHARPNESS_HIGH; - } else { - spl_in->adaptive_sharpness.enable = plane_state->adaptive_sharpness_en; - if (plane_state->sharpnessX1000 == 0) + unsigned int sharpness_setting = pipe_ctx->stream->ctx->dc->debug.force_sharpness; + unsigned int force_sharpness_level = pipe_ctx->stream->ctx->dc->debug.force_sharpness_level; + if (sharpness_setting == SHARPNESS_HW_OFF) + spl_in->adaptive_sharpness.enable = false; + else if (sharpness_setting == SHARPNESS_ZERO) { + spl_in->adaptive_sharpness.enable = true; + spl_in->adaptive_sharpness.sharpness_level = 0; + } else if (sharpness_setting == SHARPNESS_CUSTOM) { + spl_in->adaptive_sharpness.sharpness_range.sdr_rgb_min = 0; + spl_in->adaptive_sharpness.sharpness_range.sdr_rgb_max = 1750; + spl_in->adaptive_sharpness.sharpness_range.sdr_rgb_mid = 750; + spl_in->adaptive_sharpness.sharpness_range.sdr_yuv_min = 0; + spl_in->adaptive_sharpness.sharpness_range.sdr_yuv_max = 3500; + spl_in->adaptive_sharpness.sharpness_range.sdr_yuv_mid = 1500; + spl_in->adaptive_sharpness.sharpness_range.hdr_rgb_min = 0; + spl_in->adaptive_sharpness.sharpness_range.hdr_rgb_max = 2750; + spl_in->adaptive_sharpness.sharpness_range.hdr_rgb_mid = 1500; + + if (force_sharpness_level > 0) { + if (force_sharpness_level > 10) + force_sharpness_level = 10; + spl_in->adaptive_sharpness.enable = true; + spl_in->adaptive_sharpness.sharpness_level = force_sharpness_level; + } else if (!plane_state->adaptive_sharpness_en) { spl_in->adaptive_sharpness.enable = false; - else if (plane_state->sharpnessX1000 < 999) - spl_in->adaptive_sharpness.sharpness = SHARPNESS_LOW; - else if (plane_state->sharpnessX1000 < 1999) - spl_in->adaptive_sharpness.sharpness = SHARPNESS_MID; - else // Any other value is high sharpness - spl_in->adaptive_sharpness.sharpness = SHARPNESS_HIGH; + spl_in->adaptive_sharpness.sharpness_level = 0; + } else { + spl_in->adaptive_sharpness.enable = true; + spl_in->adaptive_sharpness.sharpness_level = plane_state->sharpness_level; + } } // Translate linear light scaling preference if (pipe_ctx->stream->ctx->dc->debug.force_lls > 0) @@ -171,6 +183,22 @@ void translate_SPL_in_params_from_pipe_ctx(struct pipe_ctx *pipe_ctx, struct spl /* Translate transfer function */ spl_in->basic_in.tf_type = (enum spl_transfer_func_type) plane_state->in_transfer_func.type; spl_in->basic_in.tf_predefined_type = (enum spl_transfer_func_predefined) plane_state->in_transfer_func.tf; + + spl_in->h_active = pipe_ctx->plane_res.scl_data.h_active; + spl_in->v_active = pipe_ctx->plane_res.scl_data.v_active; + /* Check if it is stream is in fullscreen and if its HDR. + * Use this to determine sharpness levels + */ + spl_in->is_fullscreen = dm_helpers_is_fullscreen(pipe_ctx->stream->ctx, pipe_ctx->stream); + spl_in->is_hdr_on = dm_helpers_is_hdr_on(pipe_ctx->stream->ctx, pipe_ctx->stream); + spl_in->hdr_multx100 = 0; + if (spl_in->is_hdr_on) { + spl_in->hdr_multx100 = (uint32_t)dc_fixpt_floor(dc_fixpt_mul(plane_state->hdr_mult, + dc_fixpt_from_int(100))); + /* Disable sharpness for HDR Mult > 6.0 */ + if (spl_in->hdr_multx100 > 600) + spl_in->adaptive_sharpness.enable = false; + } } /// @brief Translate SPL output parameters to pipe context @@ -179,15 +207,15 @@ void translate_SPL_in_params_from_pipe_ctx(struct pipe_ctx *pipe_ctx, struct spl void translate_SPL_out_params_to_pipe_ctx(struct pipe_ctx *pipe_ctx, struct spl_out *spl_out) { // Make scaler data recout point to spl output field recout - populate_rect_from_splrect(&pipe_ctx->plane_res.scl_data.recout, &spl_out->scl_data.recout); + populate_rect_from_splrect(&pipe_ctx->plane_res.scl_data.recout, &spl_out->dscl_prog_data->recout); // Make scaler data ratios point to spl output field ratios - populate_ratios_from_splratios(&pipe_ctx->plane_res.scl_data.ratios, &spl_out->scl_data.ratios); + populate_ratios_from_splratios(&pipe_ctx->plane_res.scl_data.ratios, &spl_out->dscl_prog_data->ratios); // Make scaler data viewport point to spl output field viewport - populate_rect_from_splrect(&pipe_ctx->plane_res.scl_data.viewport, &spl_out->scl_data.viewport); + populate_rect_from_splrect(&pipe_ctx->plane_res.scl_data.viewport, &spl_out->dscl_prog_data->viewport); // Make scaler data viewport_c point to spl output field viewport_c - populate_rect_from_splrect(&pipe_ctx->plane_res.scl_data.viewport_c, &spl_out->scl_data.viewport_c); + populate_rect_from_splrect(&pipe_ctx->plane_res.scl_data.viewport_c, &spl_out->dscl_prog_data->viewport_c); // Make scaler data taps point to spl output field scaling taps - populate_taps_from_spltaps(&pipe_ctx->plane_res.scl_data.taps, &spl_out->scl_data.taps); + populate_taps_from_spltaps(&pipe_ctx->plane_res.scl_data.taps, &spl_out->dscl_prog_data->taps); // Make scaler data init point to spl output field init - populate_inits_from_splinits(&pipe_ctx->plane_res.scl_data.inits, &spl_out->scl_data.inits); + populate_inits_from_splinits(&pipe_ctx->plane_res.scl_data.inits, &spl_out->dscl_prog_data->init); } diff --git a/drivers/gpu/drm/amd/display/dc/dc_spl_translate.h b/drivers/gpu/drm/amd/display/dc/dc_spl_translate.h index c73d640c3632..eaa5c5373b28 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_spl_translate.h +++ b/drivers/gpu/drm/amd/display/dc/dc_spl_translate.h @@ -6,6 +6,7 @@ #define __DC_SPL_TRANSLATE_H__ #include "dc.h" #include "resource.h" +#include "dm_helpers.h" /* Map SPL input parameters to pipe context * @pipe_ctx: pipe context diff --git a/drivers/gpu/drm/amd/display/dc/dc_stream.h b/drivers/gpu/drm/amd/display/dc/dc_stream.h index de9bd72ca514..14ea47eda0c8 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_stream.h +++ b/drivers/gpu/drm/amd/display/dc/dc_stream.h @@ -142,6 +142,7 @@ union stream_update_flags { uint32_t mst_bw : 1; uint32_t crtc_timing_adjust : 1; uint32_t fams_changed : 1; + uint32_t scaler_sharpener : 1; } bits; uint32_t raw; @@ -308,6 +309,7 @@ struct dc_stream_state { bool is_phantom; struct luminance_data lumin_data; + bool scaler_sharpener_update; }; #define ABM_LEVEL_IMMEDIATE_DISABLE 255 @@ -353,6 +355,7 @@ struct dc_stream_update { struct dc_cursor_attributes *cursor_attributes; struct dc_cursor_position *cursor_position; bool *hw_cursor_req; + bool *scaler_sharpener_update; }; bool dc_is_stream_unchanged( diff --git a/drivers/gpu/drm/amd/display/dc/dc_types.h b/drivers/gpu/drm/amd/display/dc/dc_types.h index 97279b080f3e..fd6dca735714 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_types.h +++ b/drivers/gpu/drm/amd/display/dc/dc_types.h @@ -1050,6 +1050,23 @@ union replay_error_status { unsigned char raw; }; +union replay_low_refresh_rate_enable_options { + struct { + //BIT[0-3]: Replay Low Hz Support control + unsigned int ENABLE_LOW_RR_SUPPORT :1; + unsigned int RESERVED_1_3 :3; + //BIT[4-15]: Replay Low Hz Enable Scenarios + unsigned int ENABLE_STATIC_SCREEN :1; + unsigned int ENABLE_FULL_SCREEN_VIDEO :1; + unsigned int ENABLE_GENERAL_UI :1; + unsigned int RESERVED_7_15 :9; + //BIT[16-31]: Replay Low Hz Enable Check + unsigned int ENABLE_STATIC_FLICKER_CHECK :1; + unsigned int RESERVED_17_31 :15; + } bits; + unsigned int raw; +}; + struct replay_config { /* Replay feature is supported */ bool replay_supported; @@ -1073,6 +1090,8 @@ struct replay_config { bool replay_support_fast_resync_in_ultra_sleep_mode; /* Replay error status */ union replay_error_status replay_error_status; + /* Replay Low Hz enable Options */ + union replay_low_refresh_rate_enable_options low_rr_enable_options; }; /* Replay feature flags*/ diff --git a/drivers/gpu/drm/amd/display/dc/dccg/dcn20/dcn20_dccg.h b/drivers/gpu/drm/amd/display/dc/dccg/dcn20/dcn20_dccg.h index 6ac2bd86c4db..160c299419b7 100644 --- a/drivers/gpu/drm/amd/display/dc/dccg/dcn20/dcn20_dccg.h +++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn20/dcn20_dccg.h @@ -328,6 +328,17 @@ type DPSTREAMCLK1_GATE_DISABLE;\ type DPSTREAMCLK2_GATE_DISABLE;\ type DPSTREAMCLK3_GATE_DISABLE;\ + type SYMCLKA_FE_GATE_DISABLE;\ + type SYMCLKB_FE_GATE_DISABLE;\ + type SYMCLKC_FE_GATE_DISABLE;\ + type SYMCLKD_FE_GATE_DISABLE;\ + type SYMCLKE_FE_GATE_DISABLE;\ + type SYMCLKA_GATE_DISABLE;\ + type SYMCLKB_GATE_DISABLE;\ + type SYMCLKC_GATE_DISABLE;\ + type SYMCLKD_GATE_DISABLE;\ + type SYMCLKE_GATE_DISABLE;\ + #define DCCG401_REG_FIELD_LIST(type) \ type OTG0_TMDS_PIXEL_RATE_DIV;\ diff --git a/drivers/gpu/drm/amd/display/dc/dccg/dcn35/dcn35_dccg.c b/drivers/gpu/drm/amd/display/dc/dccg/dcn35/dcn35_dccg.c index 7f91e48902e2..ee02b78e290f 100644 --- a/drivers/gpu/drm/amd/display/dc/dccg/dcn35/dcn35_dccg.c +++ b/drivers/gpu/drm/amd/display/dc/dccg/dcn35/dcn35_dccg.c @@ -24,6 +24,7 @@ #include "reg_helper.h" #include "core_types.h" +#include "resource.h" #include "dcn35_dccg.h" #define TO_DCN_DCCG(dccg)\ @@ -136,7 +137,7 @@ static void dccg35_set_dsc_clk_rcg(struct dccg *dccg, int inst, bool enable) { struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); - if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dsc) + if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dsc && enable) return; switch (inst) { @@ -165,7 +166,7 @@ static void dccg35_set_symclk32_se_rcg( { struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); - if (!dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se) + if (!dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se && enable) return; /* SYMCLK32_ROOT_SE#_GATE_DISABLE will clock gate in DCCG */ @@ -204,7 +205,7 @@ static void dccg35_set_symclk32_le_rcg( { struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); - if (!dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le) + if (!dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_le && enable) return; switch (inst) { @@ -231,7 +232,7 @@ static void dccg35_set_physymclk_rcg( { struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); - if (!dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk) + if (!dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk && enable) return; switch (inst) { @@ -262,35 +263,45 @@ static void dccg35_set_physymclk_rcg( } static void dccg35_set_symclk_fe_rcg( - struct dccg *dccg, - int inst, - bool enable) + struct dccg *dccg, + int inst, + bool enable) { struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); - if (!dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk) + if (!dccg->ctx->dc->debug.root_clock_optimization.bits.symclk_fe && enable) return; switch (inst) { case 0: + REG_UPDATE(DCCG_GATE_DISABLE_CNTL2, + SYMCLKA_FE_GATE_DISABLE, enable ? 0 : 1); REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, - SYMCLKA_FE_ROOT_GATE_DISABLE, enable ? 0 : 1); + SYMCLKA_FE_ROOT_GATE_DISABLE, enable ? 0 : 1); break; case 1: + REG_UPDATE(DCCG_GATE_DISABLE_CNTL2, + SYMCLKB_FE_GATE_DISABLE, enable ? 0 : 1); REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, - SYMCLKB_FE_ROOT_GATE_DISABLE, enable ? 0 : 1); + SYMCLKB_FE_ROOT_GATE_DISABLE, enable ? 0 : 1); break; case 2: + REG_UPDATE(DCCG_GATE_DISABLE_CNTL2, + SYMCLKC_FE_GATE_DISABLE, enable ? 0 : 1); REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, - SYMCLKC_FE_ROOT_GATE_DISABLE, enable ? 0 : 1); + SYMCLKC_FE_ROOT_GATE_DISABLE, enable ? 0 : 1); break; case 3: + REG_UPDATE(DCCG_GATE_DISABLE_CNTL2, + SYMCLKD_FE_GATE_DISABLE, enable ? 0 : 1); REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, - SYMCLKD_FE_ROOT_GATE_DISABLE, enable ? 0 : 1); + SYMCLKD_FE_ROOT_GATE_DISABLE, enable ? 0 : 1); break; case 4: + REG_UPDATE(DCCG_GATE_DISABLE_CNTL2, + SYMCLKE_FE_GATE_DISABLE, enable ? 0 : 1); REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, - SYMCLKE_FE_ROOT_GATE_DISABLE, enable ? 0 : 1); + SYMCLKE_FE_ROOT_GATE_DISABLE, enable ? 0 : 1); break; default: BREAK_TO_DEBUGGER(); @@ -307,27 +318,37 @@ static void dccg35_set_symclk_be_rcg( struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); /* TBD add symclk_be in rcg control bits */ - if (!dccg->ctx->dc->debug.root_clock_optimization.bits.physymclk) + if (!dccg->ctx->dc->debug.root_clock_optimization.bits.symclk_fe && enable) return; switch (inst) { case 0: + REG_UPDATE(DCCG_GATE_DISABLE_CNTL2, + SYMCLKA_GATE_DISABLE, enable ? 0 : 1); REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKA_ROOT_GATE_DISABLE, enable ? 0 : 1); break; case 1: + REG_UPDATE(DCCG_GATE_DISABLE_CNTL2, + SYMCLKB_GATE_DISABLE, enable ? 0 : 1); REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKB_ROOT_GATE_DISABLE, enable ? 0 : 1); break; case 2: + REG_UPDATE(DCCG_GATE_DISABLE_CNTL2, + SYMCLKC_GATE_DISABLE, enable ? 0 : 1); REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKC_ROOT_GATE_DISABLE, enable ? 0 : 1); break; case 3: + REG_UPDATE(DCCG_GATE_DISABLE_CNTL2, + SYMCLKD_GATE_DISABLE, enable ? 0 : 1); REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKD_ROOT_GATE_DISABLE, enable ? 0 : 1); break; case 4: + REG_UPDATE(DCCG_GATE_DISABLE_CNTL2, + SYMCLKE_GATE_DISABLE, enable ? 0 : 1); REG_UPDATE(DCCG_GATE_DISABLE_CNTL5, SYMCLKE_ROOT_GATE_DISABLE, enable ? 0 : 1); break; @@ -342,7 +363,7 @@ static void dccg35_set_dtbclk_p_rcg(struct dccg *dccg, int inst, bool enable) struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); - if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dpp) + if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dpp && enable) return; switch (inst) { @@ -370,7 +391,7 @@ static void dccg35_set_dppclk_rcg(struct dccg *dccg, struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); - if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dpp) + if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dpp && enable) return; switch (inst) { @@ -399,7 +420,7 @@ static void dccg35_set_dpstreamclk_rcg( { struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); - if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream) + if (!dccg->ctx->dc->debug.root_clock_optimization.bits.dpstream && enable) return; switch (inst) { @@ -436,7 +457,7 @@ static void dccg35_set_smclk32_se_rcg( { struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); - if (!dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se) + if (!dccg->ctx->dc->debug.root_clock_optimization.bits.symclk32_se && enable) return; switch (inst) { @@ -1082,7 +1103,8 @@ static void dccg35_trigger_dio_fifo_resync(struct dccg *dccg) uint32_t dispclk_rdivider_value = 0; REG_GET(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_RDIVIDER, &dispclk_rdivider_value); - REG_UPDATE(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_WDIVIDER, dispclk_rdivider_value); + if (dispclk_rdivider_value != 0) + REG_UPDATE(DENTIST_DISPCLK_CNTL, DENTIST_DISPCLK_WDIVIDER, dispclk_rdivider_value); } static void dcn35_set_dppclk_enable(struct dccg *dccg, @@ -1692,6 +1714,12 @@ static void dccg35_disable_symclk32_se( } } +static void dccg35_init_cb(struct dccg *dccg) +{ + (void)dccg; + /* Any RCG should be done when driver enter low power mode*/ +} + void dccg35_init(struct dccg *dccg) { int otg_inst; @@ -2042,8 +2070,6 @@ static void dccg35_set_dpstreamclk_cb( enum dtbclk_source dtb_clk_src; enum dp_stream_clk_source dp_stream_clk_src; - ASSERT(otg_inst >= DP_STREAM_DTBCLK_P5); - switch (src) { case REFCLK: dtb_clk_src = DTBCLK_REFCLK; @@ -2098,6 +2124,13 @@ static void dccg35_update_dpp_dto_cb(struct dccg *dccg, int dpp_inst, { struct dcn_dccg *dccg_dcn = TO_DCN_DCCG(dccg); + if (dccg->dpp_clock_gated[dpp_inst]) { + /* + * Do not update the DPPCLK DTO if the clock is stopped. + */ + return; + } + if (dccg->ref_dppclk && req_dppclk) { int ref_dppclk = dccg->ref_dppclk; int modulo, phase; @@ -2125,19 +2158,20 @@ static void dccg35_update_dpp_dto_cb(struct dccg *dccg, int dpp_inst, } static void dccg35_dpp_root_clock_control_cb( - struct dccg *dccg, - unsigned int dpp_inst, - bool power_on) + struct dccg *dccg, + unsigned int dpp_inst, + bool power_on) { + if (dccg->dpp_clock_gated[dpp_inst] == power_on) + return; /* power_on set indicates we need to ungate * Currently called from optimize_bandwidth and prepare_bandwidth calls * Since clock source is not passed restore to refclock on ungate * Redundant as gating when enabled is acheived through update_dpp_dto */ - if (power_on) - dccg35_enable_dpp_clk_new(dccg, dpp_inst, DPP_REFCLK); - else - dccg35_disable_dpp_clk_new(dccg, dpp_inst); + dccg35_set_dppclk_rcg(dccg, dpp_inst, !power_on); + + dccg->dpp_clock_gated[dpp_inst] = !power_on; } static void dccg35_enable_symclk32_se_cb( @@ -2321,7 +2355,7 @@ static const struct dccg_funcs dccg35_funcs_new = { .update_dpp_dto = dccg35_update_dpp_dto_cb, .dpp_root_clock_control = dccg35_dpp_root_clock_control_cb, .get_dccg_ref_freq = dccg31_get_dccg_ref_freq, - .dccg_init = dccg35_init, + .dccg_init = dccg35_init_cb, .set_dpstreamclk = dccg35_set_dpstreamclk_cb, .set_dpstreamclk_root_clock_gating = dccg35_set_dpstreamclk_root_clock_gating_cb, .enable_symclk32_se = dccg35_enable_symclk32_se_cb, diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_replay.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_replay.c index 14f935961672..c31e4f26a305 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dmub_replay.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_replay.c @@ -12,6 +12,8 @@ #define MAX_PIPES 6 +#define GPINT_RETRY_NUM 20 + static const uint8_t DP_SINK_DEVICE_STR_ID_1[] = {7, 1, 8, 7, 3}; static const uint8_t DP_SINK_DEVICE_STR_ID_2[] = {7, 1, 8, 7, 5}; @@ -222,6 +224,7 @@ static void dmub_replay_residency(struct dmub_replay *dmub, uint8_t panel_inst, uint32_t *residency, const bool is_start, enum pr_residency_mode mode) { uint16_t param = (uint16_t)(panel_inst << 8); + uint32_t i = 0; switch (mode) { case PR_RESIDENCY_MODE_PHY: @@ -249,10 +252,17 @@ static void dmub_replay_residency(struct dmub_replay *dmub, uint8_t panel_inst, if (is_start) param |= REPLAY_RESIDENCY_ENABLE; - // Send gpint command and wait for ack - if (!dc_wake_and_execute_gpint(dmub->ctx, DMUB_GPINT__REPLAY_RESIDENCY, param, - residency, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY)) - *residency = 0; + for (i = 0; i < GPINT_RETRY_NUM; i++) { + // Send gpint command and wait for ack + if (dc_wake_and_execute_gpint(dmub->ctx, DMUB_GPINT__REPLAY_RESIDENCY, param, + residency, DM_DMUB_WAIT_TYPE_WAIT_WITH_REPLY)) + return; + + udelay(100); + } + + // it means gpint retry many times + *residency = 0; } /* diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c index 710a25dcfef0..b0d9aed0f265 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c @@ -777,6 +777,14 @@ static void populate_dml21_plane_config_from_plane_state(struct dml2_context *dm * certain cases. Hence do corrective active and disable scaling. */ plane->composition.scaler_info.enabled = false; + } else if ((plane_state->ctx->dc->config.use_spl == true) && + (plane->composition.scaler_info.enabled == false)) { + /* To enable sharpener for 1:1, scaler must be enabled. If use_spl is set, then + * allow case where ratio is 1 but taps > 1 + */ + if ((scaler_data->taps.h_taps > 1) || (scaler_data->taps.v_taps > 1) || + (scaler_data->taps.h_taps_c > 1) || (scaler_data->taps.v_taps_c > 1)) + plane->composition.scaler_info.enabled = true; } /* always_scale is only used for debug purposes not used in production but has to be diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_types.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_types.h index 1c773bbb9992..eeb96c455658 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_types.h +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_types.h @@ -5,7 +5,6 @@ #ifndef __DML_TOP_TYPES_H__ #define __DML_TOP_TYPES_H__ -#include "dml_top_types.h" #include "dml_top_display_cfg_types.h" #include "dml_top_soc_parameter_types.h" #include "dml_top_policy_types.h" diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c index 805fd783131f..3ea54fd52e46 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c @@ -2085,7 +2085,11 @@ static void CalculateDCCConfiguration( unsigned int full_swath_bytes_vert_wc_l; unsigned int full_swath_bytes_vert_wc_c; - yuv420 = dml_is_420(SourcePixelFormat); + if (dml_is_420(SourcePixelFormat)) + yuv420 = 1; + else + yuv420 = 0; + horz_div_l = 1; horz_div_c = 1; vert_div_l = 1; @@ -2553,8 +2557,11 @@ static void calculate_mcache_setting( l->luma_time_factor = (double)l->mvmpg_width_c / l->mvmpg_width_l * 2; // The algorithm starts with computing a non-integer, avg_mcache_element_size_l/c: - l->avg_mcache_element_size_l = l->meta_row_width_l / *p->num_mcaches_l; - if (l->is_dual_plane) { + if (*p->num_mcaches_l) { + l->avg_mcache_element_size_l = l->meta_row_width_l / *p->num_mcaches_l; + } + + if (l->is_dual_plane && *p->num_mcaches_c) { l->avg_mcache_element_size_c = l->meta_row_width_c / *p->num_mcaches_c; if (!p->imall_enable || (*p->mall_comb_mcache_l == *p->mall_comb_mcache_c)) { @@ -2683,9 +2690,9 @@ static double dml_get_return_bandwidth_available( double ideal_fabric_bandwidth = fclk_mhz * (double)soc->fabric_datapath_to_dcn_data_return_bytes; double ideal_dram_bandwidth = dram_bw_mbps; //dram_speed_mts * soc->clk_table.dram_config.channel_count * soc->clk_table.dram_config.channel_width_bytes; - double derate_sdp_factor = 1; - double derate_fabric_factor = 1; - double derate_dram_factor = 1; + double derate_sdp_factor; + double derate_fabric_factor; + double derate_dram_factor; double derate_sdp_bandwidth; double derate_fabric_bandwidth; @@ -5056,6 +5063,8 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch s->trip_to_mem = 0.0; *p->Tvm_trips = 0.0; *p->Tr0_trips = 0.0; + s->Tvm_no_trip_oto = 0.0; + s->Tr0_no_trip_oto = 0.0; s->Tvm_trips_rounded = 0.0; s->Tr0_trips_rounded = 0.0; s->max_Tsw = 0.0; @@ -5293,31 +5302,38 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch s->Lsw_oto = math_ceil2(4.0 * math_max2(s->prefetch_sw_bytes / s->prefetch_bw_oto / s->LineTime, s->min_Lsw_oto), 1.0) / 4.0; if (p->display_cfg->gpuvm_enable == true) { - s->Tvm_oto = math_max3( - *p->Tvm_trips, + s->Tvm_no_trip_oto = math_max2( *p->Tno_bw + vm_bytes * p->HostVMInefficiencyFactor / s->prefetch_bw_oto, s->LineTime / 4.0); + s->Tvm_oto = math_max2( + *p->Tvm_trips, + s->Tvm_no_trip_oto); #ifdef __DML_VBA_DEBUG__ dml2_printf("DML::%s: Tvm_oto max0 = %f\n", __func__, *p->Tvm_trips); dml2_printf("DML::%s: Tvm_oto max1 = %f\n", __func__, *p->Tno_bw + vm_bytes * p->HostVMInefficiencyFactor / s->prefetch_bw_oto); dml2_printf("DML::%s: Tvm_oto max2 = %f\n", __func__, s->LineTime / 4.0); #endif } else { + s->Tvm_no_trip_oto = s->Tvm_trips_rounded; s->Tvm_oto = s->Tvm_trips_rounded; } if ((p->display_cfg->gpuvm_enable == true || p->setup_for_tdlut || dcc_mrq_enable)) { - s->Tr0_oto = math_max3( - *p->Tr0_trips, + s->Tr0_no_trip_oto = math_max2( (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes) / s->prefetch_bw_oto, s->LineTime / 4.0); + s->Tr0_oto = math_max2( + *p->Tr0_trips, + s->Tr0_no_trip_oto); #ifdef __DML_VBA_DEBUG__ dml2_printf("DML::%s: Tr0_oto max0 = %f\n", __func__, *p->Tr0_trips); dml2_printf("DML::%s: Tr0_oto max1 = %f\n", __func__, (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes) / s->prefetch_bw_oto); dml2_printf("DML::%s: Tr0_oto max2 = %f\n", __func__, s->LineTime / 4); #endif - } else - s->Tr0_oto = (s->LineTime - s->Tvm_oto) / 4.0; + } else { + s->Tr0_no_trip_oto = (s->LineTime - s->Tvm_oto) / 4.0; + s->Tr0_oto = s->Tr0_no_trip_oto; + } s->Tvm_oto_lines = math_ceil2(4.0 * s->Tvm_oto / s->LineTime, 1) / 4.0; s->Tr0_oto_lines = math_ceil2(4.0 * s->Tr0_oto / s->LineTime, 1) / 4.0; @@ -5595,6 +5611,9 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch dml2_printf("DML::%s: Tvm_equ = %f\n", __func__, s->Tvm_equ); dml2_printf("DML::%s: Tr0_equ = %f\n", __func__, s->Tr0_equ); #endif + // Lsw = dst_y_prefetch - (dst_y_per_vm_vblank + 2*dst_y_per_row_vblank) + s->Lsw_equ = s->dst_y_prefetch_equ - math_ceil2(4.0 * (s->Tvm_equ + 2 * s->Tr0_equ) / s->LineTime, 1.0) / 4.0; + // Use the more stressful prefetch schedule if (s->dst_y_prefetch_oto < s->dst_y_prefetch_equ) { *p->dst_y_prefetch = s->dst_y_prefetch_oto; @@ -5603,25 +5622,28 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch *p->dst_y_per_vm_vblank = math_ceil2(4.0 * s->TimeForFetchingVM / s->LineTime, 1.0) / 4.0; *p->dst_y_per_row_vblank = math_ceil2(4.0 * s->TimeForFetchingRowInVBlank / s->LineTime, 1.0) / 4.0; + s->dst_y_per_vm_no_trip_vblank = math_ceil2(4.0 * s->Tvm_no_trip_oto / s->LineTime, 1.0) / 4.0; + s->dst_y_per_row_no_trip_vblank = math_ceil2(4.0 * s->Tr0_no_trip_oto / s->LineTime, 1.0) / 4.0; #ifdef __DML_VBA_DEBUG__ dml2_printf("DML::%s: Using oto scheduling for prefetch\n", __func__); #endif - } else { *p->dst_y_prefetch = s->dst_y_prefetch_equ; s->TimeForFetchingVM = s->Tvm_equ; s->TimeForFetchingRowInVBlank = s->Tr0_equ; - *p->dst_y_per_vm_vblank = math_ceil2(4.0 * s->TimeForFetchingVM / s->LineTime, 1.0) / 4.0; - *p->dst_y_per_row_vblank = math_ceil2(4.0 * s->TimeForFetchingRowInVBlank / s->LineTime, 1.0) / 4.0; + *p->dst_y_per_vm_vblank = math_ceil2(4.0 * s->TimeForFetchingVM / s->LineTime, 1.0) / 4.0; + *p->dst_y_per_row_vblank = math_ceil2(4.0 * s->TimeForFetchingRowInVBlank / s->LineTime, 1.0) / 4.0; + s->dst_y_per_vm_no_trip_vblank = *p->dst_y_per_vm_vblank; + s->dst_y_per_row_no_trip_vblank = *p->dst_y_per_row_vblank; #ifdef __DML_VBA_DEBUG__ dml2_printf("DML::%s: Using equ bw scheduling for prefetch\n", __func__); #endif } - // Lsw = dst_y_prefetch - (dst_y_per_vm_vblank + 2*dst_y_per_row_vblank) - s->LinesToRequestPrefetchPixelData = *p->dst_y_prefetch - *p->dst_y_per_vm_vblank - 2 * *p->dst_y_per_row_vblank; // Lsw + /* take worst case Lsw to calculate bandwidth requirement regardless of schedule */ + s->LinesToRequestPrefetchPixelData = math_min2(s->Lsw_equ, s->Lsw_oto); // Lsw s->cursor_prefetch_bytes = (unsigned int)math_max2(p->cursor_bytes_per_chunk, 4 * p->cursor_bytes_per_line); *p->prefetch_cursor_bw = p->num_cursors * s->cursor_prefetch_bytes / (s->LinesToRequestPrefetchPixelData * s->LineTime); @@ -5741,13 +5763,13 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch if (vm_bytes == 0) { prefetch_vm_bw = 0; - } else if (*p->dst_y_per_vm_vblank > 0) { + } else if (s->dst_y_per_vm_no_trip_vblank > 0) { #ifdef __DML_VBA_DEBUG__ dml2_printf("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, p->HostVMInefficiencyFactor); dml2_printf("DML::%s: dst_y_per_vm_vblank = %f\n", __func__, *p->dst_y_per_vm_vblank); dml2_printf("DML::%s: LineTime = %f\n", __func__, s->LineTime); #endif - prefetch_vm_bw = vm_bytes * p->HostVMInefficiencyFactor / (*p->dst_y_per_vm_vblank * s->LineTime); + prefetch_vm_bw = vm_bytes * p->HostVMInefficiencyFactor / (s->dst_y_per_vm_no_trip_vblank * s->LineTime); #ifdef __DML_VBA_DEBUG__ dml2_printf("DML::%s: prefetch_vm_bw = %f\n", __func__, prefetch_vm_bw); #endif @@ -5759,8 +5781,8 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch if (p->PixelPTEBytesPerRow == 0 && tdlut_row_bytes == 0) { prefetch_row_bw = 0; - } else if (*p->dst_y_per_row_vblank > 0) { - prefetch_row_bw = (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + tdlut_row_bytes) / (*p->dst_y_per_row_vblank * s->LineTime); + } else if (s->dst_y_per_row_no_trip_vblank > 0) { + prefetch_row_bw = (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + tdlut_row_bytes) / (s->dst_y_per_row_no_trip_vblank * s->LineTime); #ifdef __DML_VBA_DEBUG__ dml2_printf("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, p->PixelPTEBytesPerRow); @@ -7194,7 +7216,7 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out mode_lib->ms.support.WritebackLatencySupport = true; for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) { if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.enable == true && - (mode_lib->ms.WriteBandwidth[k] > mode_lib->ip.writeback_interface_buffer_size_kbytes * 1024 / mode_lib->soc.qos_parameters.writeback.base_latency_us)) { + (mode_lib->ms.WriteBandwidth[k] > mode_lib->ip.writeback_interface_buffer_size_kbytes * 1024 / ((double)mode_lib->soc.qos_parameters.writeback.base_latency_us))) { mode_lib->ms.support.WritebackLatencySupport = false; } } @@ -10739,7 +10761,7 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex mode_lib->mp.UrgentLatency, mode_lib->mp.TripToMemory, !dml_is_phantom_pipe(&display_cfg->plane_descriptors[k]) && display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.drr_config.enabled ? - get_g6_temp_read_blackout_us(&mode_lib->soc, (unsigned int)(mode_lib->ms.uclk_freq_mhz * 1000), in_out_params->min_clk_index) : 0.0); + get_g6_temp_read_blackout_us(&mode_lib->soc, (unsigned int)(mode_lib->mp.uclk_freq_mhz * 1000), in_out_params->min_clk_index) : 0.0); myPipe->Dppclk = mode_lib->mp.Dppclk[k]; myPipe->Dispclk = mode_lib->mp.Dispclk; diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h index 13961c2eb634..cbdfbd5a0bde 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h @@ -1187,11 +1187,15 @@ struct dml2_core_calcs_CalculatePrefetchSchedule_locals { double prefetch_bw_oto; double Tvm_oto; double Tr0_oto; + double Tvm_no_trip_oto; + double Tr0_no_trip_oto; double Tvm_oto_lines; double Tr0_oto_lines; double dst_y_prefetch_oto; double TimeForFetchingVM; double TimeForFetchingRowInVBlank; + double dst_y_per_vm_no_trip_vblank; + double dst_y_per_row_no_trip_vblank; double LinesToRequestPrefetchPixelData; unsigned int HostVMDynamicLevelsTrips; double trip_to_mem; @@ -1199,6 +1203,7 @@ struct dml2_core_calcs_CalculatePrefetchSchedule_locals { double Tr0_trips_rounded; double max_Tsw; double Lsw_oto; + double Lsw_equ; double Tpre_rounded; double prefetch_bw_equ; double Tvm_equ; diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn3.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn3.c index 8e68a8094658..a31db5742675 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn3.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn3.c @@ -497,7 +497,6 @@ bool pmo_dcn3_optimize_dcc_mcache(struct dml2_pmo_optimize_dcc_mcache_in_out *in in_out->cfg_support_info->plane_support_info[i].dpps_used)) { result = false; } else { - free_pipes -= planes_on_stream; break; } } else { @@ -666,7 +665,7 @@ bool pmo_dcn3_optimize_for_pstate_support(struct dml2_pmo_optimize_for_pstate_su struct dml2_pmo_instance *pmo = in_out->instance; unsigned int stream_index; bool success = false; - bool reached_end = true; + bool reached_end; memcpy(in_out->optimized_display_config, in_out->base_display_config, sizeof(struct display_configuation_with_meta)); diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c index 30767f330fd4..d63558ee3135 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c @@ -334,7 +334,6 @@ bool pmo_dcn4_fams2_optimize_dcc_mcache(struct dml2_pmo_optimize_dcc_mcache_in_o in_out->cfg_support_info->plane_support_info[i].dpps_used)) { result = false; } else { - free_pipes -= planes_on_stream; break; } } else { @@ -672,8 +671,6 @@ bool pmo_dcn4_fams2_initialize(struct dml2_pmo_initialize_in_out *in_out) /* populate list */ expand_base_strategies(pmo, base_strategy_list_4_display, base_strategy_list_4_display_size, 4); break; - default: - break; } } @@ -944,7 +941,8 @@ static void build_synchronized_timing_groups( for (j = i + 1; j < display_config->display_config.num_streams; j++) { if (memcmp(master_timing, &display_config->display_config.stream_descriptors[j].timing, - sizeof(struct dml2_timing_cfg)) == 0) { + sizeof(struct dml2_timing_cfg)) == 0 && + display_config->display_config.stream_descriptors[i].output.output_encoder == display_config->display_config.stream_descriptors[j].output.output_encoder) { set_bit_in_bitfield(&pmo->scratch.pmo_dcn4.synchronized_timing_group_masks[timing_group_idx], j); set_bit_in_bitfield(&stream_mapped_mask, j); } diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_optimization.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_optimization.c index dc8af4dd0410..d0e026d981b5 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_optimization.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_optimization.c @@ -219,7 +219,6 @@ bool dml2_top_optimization_perform_optimization_phase_1(struct dml2_optimization copy_display_configuration_with_meta(&l->cur_candidate_display_cfg, params->display_config); highest_state = l->cur_candidate_display_cfg.stage1.min_clk_index_for_latency; lowest_state = 0; - cur_state = 0; while (highest_state > lowest_state) { cur_state = (highest_state + lowest_state) / 2; diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_internal_types.h b/drivers/gpu/drm/amd/display/dc/dml2/dml2_internal_types.h index 3ba184be25d3..140ec01545db 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_internal_types.h +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_internal_types.h @@ -101,7 +101,7 @@ struct dml2_wrapper_scratch { struct dml2_dml_to_dc_pipe_mapping dml_to_dc_pipe_mapping; bool enable_flexible_pipe_mapping; bool plane_duplicate_exists; - unsigned int dp2_mst_stream_count; + int hpo_stream_to_link_encoder_mapping[MAX_HPO_DP2_ENCODERS]; }; struct dml2_helper_det_policy_scratch { diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c index 7e39873832bf..bde4250853b1 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c @@ -733,8 +733,7 @@ static void populate_dml_timing_cfg_from_stream_state(struct dml_timing_cfg_st * } static void populate_dml_output_cfg_from_stream_state(struct dml_output_cfg_st *out, unsigned int location, - const struct dc_stream_state *in, const struct pipe_ctx *pipe, - unsigned int dp2_mst_stream_count) + const struct dc_stream_state *in, const struct pipe_ctx *pipe, struct dml2_context *dml2) { unsigned int output_bpc; @@ -747,8 +746,8 @@ static void populate_dml_output_cfg_from_stream_state(struct dml_output_cfg_st * case SIGNAL_TYPE_DISPLAY_PORT_MST: case SIGNAL_TYPE_DISPLAY_PORT: out->OutputEncoder[location] = dml_dp; - if (is_dp2p0_output_encoder(pipe, dp2_mst_stream_count)) - out->OutputEncoder[location] = dml_dp2p0; + if (dml2->v20.scratch.hpo_stream_to_link_encoder_mapping[location] != -1) + out->OutputEncoder[dml2->v20.scratch.hpo_stream_to_link_encoder_mapping[location]] = dml_dp2p0; break; case SIGNAL_TYPE_EDP: out->OutputEncoder[location] = dml_edp; @@ -1199,36 +1198,6 @@ static void dml2_populate_pipe_to_plane_index_mapping(struct dml2_context *dml2, } } -static unsigned int calculate_dp2_mst_stream_count(struct dc_state *context) -{ - int i, j; - unsigned int dp2_mst_stream_count = 0; - - for (i = 0; i < context->stream_count; i++) { - struct dc_stream_state *stream = context->streams[i]; - - if (!stream || stream->signal != SIGNAL_TYPE_DISPLAY_PORT_MST) - continue; - - for (j = 0; j < MAX_PIPES; j++) { - struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j]; - - if (!pipe_ctx || !pipe_ctx->stream) - continue; - - if (stream != pipe_ctx->stream) - continue; - - if (pipe_ctx->stream_res.hpo_dp_stream_enc && pipe_ctx->link_res.hpo_dp_link_enc) { - dp2_mst_stream_count++; - break; - } - } - } - - return dp2_mst_stream_count; -} - static void populate_dml_writeback_cfg_from_stream_state(struct dml_writeback_cfg_st *out, unsigned int location, const struct dc_stream_state *in) { @@ -1269,6 +1238,30 @@ static void populate_dml_writeback_cfg_from_stream_state(struct dml_writeback_cf } } } + +static void dml2_map_hpo_stream_encoder_to_hpo_link_encoder_index(struct dml2_context *dml2, struct dc_state *context) +{ + int i; + struct pipe_ctx *current_pipe_context; + + /* Scratch gets reset to zero in dml, but link encoder instance can be zero, so reset to -1 */ + for (i = 0; i < MAX_HPO_DP2_ENCODERS; i++) { + dml2->v20.scratch.hpo_stream_to_link_encoder_mapping[i] = -1; + } + + /* If an HPO stream encoder is allocated to a pipe, get the instance of it's allocated HPO Link encoder */ + for (i = 0; i < MAX_PIPES; i++) { + current_pipe_context = &context->res_ctx.pipe_ctx[i]; + if (current_pipe_context->stream && + current_pipe_context->stream_res.hpo_dp_stream_enc && + current_pipe_context->link_res.hpo_dp_link_enc && + dc_is_dp_signal(current_pipe_context->stream->signal)) { + dml2->v20.scratch.hpo_stream_to_link_encoder_mapping[current_pipe_context->stream_res.hpo_dp_stream_enc->inst] = + current_pipe_context->link_res.hpo_dp_link_enc->inst; + } + } +} + void map_dc_state_into_dml_display_cfg(struct dml2_context *dml2, struct dc_state *context, struct dml_display_cfg_st *dml_dispcfg) { int i = 0, j = 0, k = 0; @@ -1291,8 +1284,8 @@ void map_dc_state_into_dml_display_cfg(struct dml2_context *dml2, struct dc_stat if (dml2->v20.dml_core_ctx.ip.hostvm_enable) dml2->v20.dml_core_ctx.policy.AllowForPStateChangeOrStutterInVBlankFinal = dml_prefetch_support_uclk_fclk_and_stutter; - dml2->v20.scratch.dp2_mst_stream_count = calculate_dp2_mst_stream_count(context); dml2_populate_pipe_to_plane_index_mapping(dml2, context); + dml2_map_hpo_stream_encoder_to_hpo_link_encoder_index(dml2, context); for (i = 0; i < context->stream_count; i++) { current_pipe_context = NULL; @@ -1313,7 +1306,7 @@ void map_dc_state_into_dml_display_cfg(struct dml2_context *dml2, struct dc_stat ASSERT(disp_cfg_stream_location >= 0 && disp_cfg_stream_location <= __DML2_WRAPPER_MAX_STREAMS_PLANES__); populate_dml_timing_cfg_from_stream_state(&dml_dispcfg->timing, disp_cfg_stream_location, context->streams[i]); - populate_dml_output_cfg_from_stream_state(&dml_dispcfg->output, disp_cfg_stream_location, context->streams[i], current_pipe_context, dml2->v20.scratch.dp2_mst_stream_count); + populate_dml_output_cfg_from_stream_state(&dml_dispcfg->output, disp_cfg_stream_location, context->streams[i], current_pipe_context, dml2); /*Call site for populate_dml_writeback_cfg_from_stream_state*/ populate_dml_writeback_cfg_from_stream_state(&dml_dispcfg->writeback, disp_cfg_stream_location, context->streams[i]); @@ -1378,7 +1371,7 @@ void map_dc_state_into_dml_display_cfg(struct dml2_context *dml2, struct dc_stat if (j >= 1) { populate_dml_timing_cfg_from_stream_state(&dml_dispcfg->timing, disp_cfg_plane_location, context->streams[i]); - populate_dml_output_cfg_from_stream_state(&dml_dispcfg->output, disp_cfg_plane_location, context->streams[i], current_pipe_context, dml2->v20.scratch.dp2_mst_stream_count); + populate_dml_output_cfg_from_stream_state(&dml_dispcfg->output, disp_cfg_plane_location, context->streams[i], current_pipe_context, dml2); switch (context->streams[i]->debug.force_odm_combine_segments) { case 2: dml2->v20.dml_core_ctx.policy.ODMUse[disp_cfg_plane_location] = dml_odm_use_policy_combine_2to1; diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.h b/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.h index 55659b22d87f..d764773938f4 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.h +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.h @@ -36,6 +36,6 @@ void dml2_translate_socbb_params(const struct dc *in_dc, struct soc_bounding_box void dml2_translate_soc_states(const struct dc *in_dc, struct soc_states_st *out, int num_states); void map_dc_state_into_dml_display_cfg(struct dml2_context *dml2, struct dc_state *context, struct dml_display_cfg_st *dml_dispcfg); void dml2_update_pipe_ctx_dchub_regs(struct _vcs_dpi_dml_display_rq_regs_st *rq_regs, struct _vcs_dpi_dml_display_dlg_regs_st *disp_dlg_regs, struct _vcs_dpi_dml_display_ttu_regs_st *disp_ttu_regs, struct pipe_ctx *out); -bool is_dp2p0_output_encoder(const struct pipe_ctx *pipe, unsigned int dp2_mst_stream_count); +bool is_dp2p0_output_encoder(const struct pipe_ctx *pipe); #endif //__DML2_TRANSLATION_HELPER_H__ diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_utils.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_utils.c index 9e8ff3a9718e..9a33158b63bf 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_utils.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_utils.c @@ -153,7 +153,7 @@ unsigned int dml2_util_get_maximum_odm_combine_for_output(bool force_odm_4to1, e } } -bool is_dp2p0_output_encoder(const struct pipe_ctx *pipe_ctx, unsigned int dp2_mst_stream_count) +bool is_dp2p0_output_encoder(const struct pipe_ctx *pipe_ctx) { if (pipe_ctx == NULL || pipe_ctx->stream == NULL) return false; @@ -161,14 +161,6 @@ bool is_dp2p0_output_encoder(const struct pipe_ctx *pipe_ctx, unsigned int dp2_m /* If this assert is hit then we have a link encoder dynamic management issue */ ASSERT(pipe_ctx->stream_res.hpo_dp_stream_enc ? pipe_ctx->link_res.hpo_dp_link_enc != NULL : true); - /* Count MST hubs once by treating only 1st remote sink in topology as an encoder */ - if (pipe_ctx->stream->link && pipe_ctx->stream->link->remote_sinks[0] && dp2_mst_stream_count > 1) { - return (pipe_ctx->stream_res.hpo_dp_stream_enc && - pipe_ctx->link_res.hpo_dp_link_enc && - dc_is_dp_signal(pipe_ctx->stream->signal) && - (pipe_ctx->stream->link->remote_sinks[0]->sink_id == pipe_ctx->stream->sink->sink_id)); - } - return (pipe_ctx->stream_res.hpo_dp_stream_enc && pipe_ctx->link_res.hpo_dp_link_enc && dc_is_dp_signal(pipe_ctx->stream->signal)); @@ -181,7 +173,7 @@ bool is_dtbclk_required(const struct dc *dc, struct dc_state *context) for (i = 0; i < dc->res_pool->pipe_count; i++) { if (!context->res_ctx.pipe_ctx[i].stream) continue; - if (is_dp2p0_output_encoder(&context->res_ctx.pipe_ctx[i], context->bw_ctx.dml2->v20.scratch.dp2_mst_stream_count)) + if (is_dp2p0_output_encoder(&context->res_ctx.pipe_ctx[i])) return true; } return false; diff --git a/drivers/gpu/drm/amd/display/dc/dpp/dcn401/dcn401_dpp_dscl.c b/drivers/gpu/drm/amd/display/dc/dpp/dcn401/dcn401_dpp_dscl.c index 505929800426..01f98139292e 100644 --- a/drivers/gpu/drm/amd/display/dc/dpp/dcn401/dcn401_dpp_dscl.c +++ b/drivers/gpu/drm/amd/display/dc/dpp/dcn401/dcn401_dpp_dscl.c @@ -280,7 +280,8 @@ static void dpp401_dscl_set_scaler_filter( static void dpp401_dscl_set_scl_filter( struct dcn401_dpp *dpp, const struct scaler_data *scl_data, - bool chroma_coef_mode) + bool chroma_coef_mode, + bool force_coeffs_update) { bool h_2tap_hardcode_coef_en = false; bool v_2tap_hardcode_coef_en = false; @@ -343,7 +344,7 @@ static void dpp401_dscl_set_scl_filter( || (filter_v_c && (filter_v_c != dpp->filter_v_c)); } - if (filter_updated) { + if ((filter_updated) || (force_coeffs_update)) { uint32_t scl_mode = REG_READ(SCL_MODE); if (!h_2tap_hardcode_coef_en && filter_h) { @@ -656,274 +657,252 @@ static void dpp401_dscl_set_recout(struct dcn401_dpp *dpp, RECOUT_HEIGHT, recout->height); } /** - * dpp401_dscl_program_easf - Program EASF + * dpp401_dscl_program_easf_v - Program EASF_V * * @dpp_base: High level DPP struct * @scl_data: scalaer_data info * - * This is the primary function to program EASF + * This is the primary function to program vertical EASF registers * */ -static void dpp401_dscl_program_easf(struct dpp *dpp_base, const struct scaler_data *scl_data) +static void dpp401_dscl_program_easf_v(struct dpp *dpp_base, const struct scaler_data *scl_data) { struct dcn401_dpp *dpp = TO_DCN401_DPP(dpp_base); PERF_TRACE(); - REG_UPDATE(DSCL_SC_MODE, - SCL_SC_MATRIX_MODE, scl_data->dscl_prog_data.easf_matrix_mode); - REG_UPDATE(DSCL_SC_MODE, - SCL_SC_LTONL_EN, scl_data->dscl_prog_data.easf_ltonl_en); /* DSCL_EASF_V_MODE */ - REG_UPDATE(DSCL_EASF_V_MODE, - SCL_EASF_V_EN, scl_data->dscl_prog_data.easf_v_en); - REG_UPDATE(DSCL_EASF_V_MODE, - SCL_EASF_V_2TAP_SHARP_FACTOR, scl_data->dscl_prog_data.easf_v_sharp_factor); - REG_UPDATE(DSCL_EASF_V_MODE, + REG_SET_3(DSCL_EASF_V_MODE, 0, + SCL_EASF_V_EN, scl_data->dscl_prog_data.easf_v_en, + SCL_EASF_V_2TAP_SHARP_FACTOR, scl_data->dscl_prog_data.easf_v_sharp_factor, SCL_EASF_V_RINGEST_FORCE_EN, scl_data->dscl_prog_data.easf_v_ring); - REG_UPDATE(DSCL_EASF_V_BF_CNTL, - SCL_EASF_V_BF1_EN, scl_data->dscl_prog_data.easf_v_bf1_en); - REG_UPDATE(DSCL_EASF_V_BF_CNTL, - SCL_EASF_V_BF2_MODE, scl_data->dscl_prog_data.easf_v_bf2_mode); - REG_UPDATE(DSCL_EASF_V_BF_CNTL, - SCL_EASF_V_BF3_MODE, scl_data->dscl_prog_data.easf_v_bf3_mode); - REG_UPDATE(DSCL_EASF_V_BF_CNTL, - SCL_EASF_V_BF2_FLAT1_GAIN, scl_data->dscl_prog_data.easf_v_bf2_flat1_gain); - REG_UPDATE(DSCL_EASF_V_BF_CNTL, - SCL_EASF_V_BF2_FLAT2_GAIN, scl_data->dscl_prog_data.easf_v_bf2_flat2_gain); - REG_UPDATE(DSCL_EASF_V_BF_CNTL, + + if (!scl_data->dscl_prog_data.easf_v_en) { + PERF_TRACE(); + return; + } + + /* DSCL_EASF_V_BF_CNTL */ + REG_SET_6(DSCL_EASF_V_BF_CNTL, 0, + SCL_EASF_V_BF1_EN, scl_data->dscl_prog_data.easf_v_bf1_en, + SCL_EASF_V_BF2_MODE, scl_data->dscl_prog_data.easf_v_bf2_mode, + SCL_EASF_V_BF3_MODE, scl_data->dscl_prog_data.easf_v_bf3_mode, + SCL_EASF_V_BF2_FLAT1_GAIN, scl_data->dscl_prog_data.easf_v_bf2_flat1_gain, + SCL_EASF_V_BF2_FLAT2_GAIN, scl_data->dscl_prog_data.easf_v_bf2_flat2_gain, SCL_EASF_V_BF2_ROC_GAIN, scl_data->dscl_prog_data.easf_v_bf2_roc_gain); - REG_UPDATE(DSCL_EASF_V_RINGEST_3TAP_CNTL1, - SCL_EASF_V_RINGEST_3TAP_DNTILT_UPTILT, scl_data->dscl_prog_data.easf_v_ringest_3tap_dntilt_uptilt); - REG_UPDATE(DSCL_EASF_V_RINGEST_3TAP_CNTL1, + /* DSCL_EASF_V_RINGEST_3TAP_CNTLn */ + REG_SET_2(DSCL_EASF_V_RINGEST_3TAP_CNTL1, 0, + SCL_EASF_V_RINGEST_3TAP_DNTILT_UPTILT, scl_data->dscl_prog_data.easf_v_ringest_3tap_dntilt_uptilt, SCL_EASF_V_RINGEST_3TAP_UPTILT_MAXVAL, scl_data->dscl_prog_data.easf_v_ringest_3tap_uptilt_max); - REG_UPDATE(DSCL_EASF_V_RINGEST_3TAP_CNTL2, - SCL_EASF_V_RINGEST_3TAP_DNTILT_SLOPE, scl_data->dscl_prog_data.easf_v_ringest_3tap_dntilt_slope); - REG_UPDATE(DSCL_EASF_V_RINGEST_3TAP_CNTL2, + REG_SET_2(DSCL_EASF_V_RINGEST_3TAP_CNTL2, 0, + SCL_EASF_V_RINGEST_3TAP_DNTILT_SLOPE, scl_data->dscl_prog_data.easf_v_ringest_3tap_dntilt_slope, SCL_EASF_V_RINGEST_3TAP_UPTILT1_SLOPE, scl_data->dscl_prog_data.easf_v_ringest_3tap_uptilt1_slope); - REG_UPDATE(DSCL_EASF_V_RINGEST_3TAP_CNTL3, - SCL_EASF_V_RINGEST_3TAP_UPTILT2_SLOPE, scl_data->dscl_prog_data.easf_v_ringest_3tap_uptilt2_slope); - REG_UPDATE(DSCL_EASF_V_RINGEST_3TAP_CNTL3, + REG_SET_2(DSCL_EASF_V_RINGEST_3TAP_CNTL3, 0, + SCL_EASF_V_RINGEST_3TAP_UPTILT2_SLOPE, scl_data->dscl_prog_data.easf_v_ringest_3tap_uptilt2_slope, SCL_EASF_V_RINGEST_3TAP_UPTILT2_OFFSET, scl_data->dscl_prog_data.easf_v_ringest_3tap_uptilt2_offset); - REG_UPDATE(DSCL_EASF_V_RINGEST_EVENTAP_REDUCE, - SCL_EASF_V_RINGEST_EVENTAP_REDUCEG1, scl_data->dscl_prog_data.easf_v_ringest_eventap_reduceg1); - REG_UPDATE(DSCL_EASF_V_RINGEST_EVENTAP_REDUCE, + /* DSCL_EASF_V_RINGEST_EVENTAP_REDUCE */ + REG_SET_2(DSCL_EASF_V_RINGEST_EVENTAP_REDUCE, 0, + SCL_EASF_V_RINGEST_EVENTAP_REDUCEG1, scl_data->dscl_prog_data.easf_v_ringest_eventap_reduceg1, SCL_EASF_V_RINGEST_EVENTAP_REDUCEG2, scl_data->dscl_prog_data.easf_v_ringest_eventap_reduceg2); - REG_UPDATE(DSCL_EASF_V_RINGEST_EVENTAP_GAIN, - SCL_EASF_V_RINGEST_EVENTAP_GAIN1, scl_data->dscl_prog_data.easf_v_ringest_eventap_gain1); - REG_UPDATE(DSCL_EASF_V_RINGEST_EVENTAP_GAIN, + /* DSCL_EASF_V_RINGEST_EVENTAP_GAIN */ + REG_SET_2(DSCL_EASF_V_RINGEST_EVENTAP_GAIN, 0, + SCL_EASF_V_RINGEST_EVENTAP_GAIN1, scl_data->dscl_prog_data.easf_v_ringest_eventap_gain1, SCL_EASF_V_RINGEST_EVENTAP_GAIN2, scl_data->dscl_prog_data.easf_v_ringest_eventap_gain2); - REG_UPDATE(DSCL_EASF_V_BF_FINAL_MAX_MIN, - SCL_EASF_V_BF_MAXA, scl_data->dscl_prog_data.easf_v_bf_maxa); - REG_UPDATE(DSCL_EASF_V_BF_FINAL_MAX_MIN, - SCL_EASF_V_BF_MAXB, scl_data->dscl_prog_data.easf_v_bf_maxb); - REG_UPDATE(DSCL_EASF_V_BF_FINAL_MAX_MIN, - SCL_EASF_V_BF_MINA, scl_data->dscl_prog_data.easf_v_bf_mina); - REG_UPDATE(DSCL_EASF_V_BF_FINAL_MAX_MIN, + /* DSCL_EASF_V_BF_FINAL_MAX_MIN */ + REG_SET_4(DSCL_EASF_V_BF_FINAL_MAX_MIN, 0, + SCL_EASF_V_BF_MAXA, scl_data->dscl_prog_data.easf_v_bf_maxa, + SCL_EASF_V_BF_MAXB, scl_data->dscl_prog_data.easf_v_bf_maxb, + SCL_EASF_V_BF_MINA, scl_data->dscl_prog_data.easf_v_bf_mina, SCL_EASF_V_BF_MINB, scl_data->dscl_prog_data.easf_v_bf_minb); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG0, - SCL_EASF_V_BF1_PWL_IN_SEG0, scl_data->dscl_prog_data.easf_v_bf1_pwl_in_seg0); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG0, - SCL_EASF_V_BF1_PWL_BASE_SEG0, scl_data->dscl_prog_data.easf_v_bf1_pwl_base_seg0); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG0, + /* DSCL_EASF_V_BF1_PWL_SEGn */ + REG_SET_3(DSCL_EASF_V_BF1_PWL_SEG0, 0, + SCL_EASF_V_BF1_PWL_IN_SEG0, scl_data->dscl_prog_data.easf_v_bf1_pwl_in_seg0, + SCL_EASF_V_BF1_PWL_BASE_SEG0, scl_data->dscl_prog_data.easf_v_bf1_pwl_base_seg0, SCL_EASF_V_BF1_PWL_SLOPE_SEG0, scl_data->dscl_prog_data.easf_v_bf1_pwl_slope_seg0); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG1, - SCL_EASF_V_BF1_PWL_IN_SEG1, scl_data->dscl_prog_data.easf_v_bf1_pwl_in_seg1); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG1, - SCL_EASF_V_BF1_PWL_BASE_SEG1, scl_data->dscl_prog_data.easf_v_bf1_pwl_base_seg1); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG1, + REG_SET_3(DSCL_EASF_V_BF1_PWL_SEG1, 0, + SCL_EASF_V_BF1_PWL_IN_SEG1, scl_data->dscl_prog_data.easf_v_bf1_pwl_in_seg1, + SCL_EASF_V_BF1_PWL_BASE_SEG1, scl_data->dscl_prog_data.easf_v_bf1_pwl_base_seg1, SCL_EASF_V_BF1_PWL_SLOPE_SEG1, scl_data->dscl_prog_data.easf_v_bf1_pwl_slope_seg1); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG2, - SCL_EASF_V_BF1_PWL_IN_SEG2, scl_data->dscl_prog_data.easf_v_bf1_pwl_in_seg2); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG2, - SCL_EASF_V_BF1_PWL_BASE_SEG2, scl_data->dscl_prog_data.easf_v_bf1_pwl_base_seg2); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG2, + REG_SET_3(DSCL_EASF_V_BF1_PWL_SEG2, 0, + SCL_EASF_V_BF1_PWL_IN_SEG2, scl_data->dscl_prog_data.easf_v_bf1_pwl_in_seg2, + SCL_EASF_V_BF1_PWL_BASE_SEG2, scl_data->dscl_prog_data.easf_v_bf1_pwl_base_seg2, SCL_EASF_V_BF1_PWL_SLOPE_SEG2, scl_data->dscl_prog_data.easf_v_bf1_pwl_slope_seg2); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG3, - SCL_EASF_V_BF1_PWL_IN_SEG3, scl_data->dscl_prog_data.easf_v_bf1_pwl_in_seg3); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG3, - SCL_EASF_V_BF1_PWL_BASE_SEG3, scl_data->dscl_prog_data.easf_v_bf1_pwl_base_seg3); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG3, + REG_SET_3(DSCL_EASF_V_BF1_PWL_SEG3, 0, + SCL_EASF_V_BF1_PWL_IN_SEG3, scl_data->dscl_prog_data.easf_v_bf1_pwl_in_seg3, + SCL_EASF_V_BF1_PWL_BASE_SEG3, scl_data->dscl_prog_data.easf_v_bf1_pwl_base_seg3, SCL_EASF_V_BF1_PWL_SLOPE_SEG3, scl_data->dscl_prog_data.easf_v_bf1_pwl_slope_seg3); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG4, - SCL_EASF_V_BF1_PWL_IN_SEG4, scl_data->dscl_prog_data.easf_v_bf1_pwl_in_seg4); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG4, - SCL_EASF_V_BF1_PWL_BASE_SEG4, scl_data->dscl_prog_data.easf_v_bf1_pwl_base_seg4); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG4, + REG_SET_3(DSCL_EASF_V_BF1_PWL_SEG4, 0, + SCL_EASF_V_BF1_PWL_IN_SEG4, scl_data->dscl_prog_data.easf_v_bf1_pwl_in_seg4, + SCL_EASF_V_BF1_PWL_BASE_SEG4, scl_data->dscl_prog_data.easf_v_bf1_pwl_base_seg4, SCL_EASF_V_BF1_PWL_SLOPE_SEG4, scl_data->dscl_prog_data.easf_v_bf1_pwl_slope_seg4); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG5, - SCL_EASF_V_BF1_PWL_IN_SEG5, scl_data->dscl_prog_data.easf_v_bf1_pwl_in_seg5); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG5, - SCL_EASF_V_BF1_PWL_BASE_SEG5, scl_data->dscl_prog_data.easf_v_bf1_pwl_base_seg5); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG5, + REG_SET_3(DSCL_EASF_V_BF1_PWL_SEG5, 0, + SCL_EASF_V_BF1_PWL_IN_SEG5, scl_data->dscl_prog_data.easf_v_bf1_pwl_in_seg5, + SCL_EASF_V_BF1_PWL_BASE_SEG5, scl_data->dscl_prog_data.easf_v_bf1_pwl_base_seg5, SCL_EASF_V_BF1_PWL_SLOPE_SEG5, scl_data->dscl_prog_data.easf_v_bf1_pwl_slope_seg5); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG6, - SCL_EASF_V_BF1_PWL_IN_SEG6, scl_data->dscl_prog_data.easf_v_bf1_pwl_in_seg6); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG6, - SCL_EASF_V_BF1_PWL_BASE_SEG6, scl_data->dscl_prog_data.easf_v_bf1_pwl_base_seg6); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG6, + REG_SET_3(DSCL_EASF_V_BF1_PWL_SEG6, 0, + SCL_EASF_V_BF1_PWL_IN_SEG6, scl_data->dscl_prog_data.easf_v_bf1_pwl_in_seg6, + SCL_EASF_V_BF1_PWL_BASE_SEG6, scl_data->dscl_prog_data.easf_v_bf1_pwl_base_seg6, SCL_EASF_V_BF1_PWL_SLOPE_SEG6, scl_data->dscl_prog_data.easf_v_bf1_pwl_slope_seg6); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG7, - SCL_EASF_V_BF1_PWL_IN_SEG7, scl_data->dscl_prog_data.easf_v_bf1_pwl_in_seg7); - REG_UPDATE(DSCL_EASF_V_BF1_PWL_SEG7, + REG_SET_2(DSCL_EASF_V_BF1_PWL_SEG7, 0, + SCL_EASF_V_BF1_PWL_IN_SEG7, scl_data->dscl_prog_data.easf_v_bf1_pwl_in_seg7, SCL_EASF_V_BF1_PWL_BASE_SEG7, scl_data->dscl_prog_data.easf_v_bf1_pwl_base_seg7); - REG_UPDATE(DSCL_EASF_V_BF3_PWL_SEG0, - SCL_EASF_V_BF3_PWL_IN_SEG0, scl_data->dscl_prog_data.easf_v_bf3_pwl_in_set0); - REG_UPDATE(DSCL_EASF_V_BF3_PWL_SEG0, - SCL_EASF_V_BF3_PWL_BASE_SEG0, scl_data->dscl_prog_data.easf_v_bf3_pwl_base_set0); - REG_UPDATE(DSCL_EASF_V_BF3_PWL_SEG0, + /* DSCL_EASF_V_BF3_PWL_SEGn */ + REG_SET_3(DSCL_EASF_V_BF3_PWL_SEG0, 0, + SCL_EASF_V_BF3_PWL_IN_SEG0, scl_data->dscl_prog_data.easf_v_bf3_pwl_in_set0, + SCL_EASF_V_BF3_PWL_BASE_SEG0, scl_data->dscl_prog_data.easf_v_bf3_pwl_base_set0, SCL_EASF_V_BF3_PWL_SLOPE_SEG0, scl_data->dscl_prog_data.easf_v_bf3_pwl_slope_set0); - REG_UPDATE(DSCL_EASF_V_BF3_PWL_SEG1, - SCL_EASF_V_BF3_PWL_IN_SEG1, scl_data->dscl_prog_data.easf_v_bf3_pwl_in_set1); - REG_UPDATE(DSCL_EASF_V_BF3_PWL_SEG1, - SCL_EASF_V_BF3_PWL_BASE_SEG1, scl_data->dscl_prog_data.easf_v_bf3_pwl_base_set1); - REG_UPDATE(DSCL_EASF_V_BF3_PWL_SEG1, + REG_SET_3(DSCL_EASF_V_BF3_PWL_SEG1, 0, + SCL_EASF_V_BF3_PWL_IN_SEG1, scl_data->dscl_prog_data.easf_v_bf3_pwl_in_set1, + SCL_EASF_V_BF3_PWL_BASE_SEG1, scl_data->dscl_prog_data.easf_v_bf3_pwl_base_set1, SCL_EASF_V_BF3_PWL_SLOPE_SEG1, scl_data->dscl_prog_data.easf_v_bf3_pwl_slope_set1); - REG_UPDATE(DSCL_EASF_V_BF3_PWL_SEG2, - SCL_EASF_V_BF3_PWL_IN_SEG2, scl_data->dscl_prog_data.easf_v_bf3_pwl_in_set2); - REG_UPDATE(DSCL_EASF_V_BF3_PWL_SEG2, - SCL_EASF_V_BF3_PWL_BASE_SEG2, scl_data->dscl_prog_data.easf_v_bf3_pwl_base_set2); - REG_UPDATE(DSCL_EASF_V_BF3_PWL_SEG2, + REG_SET_3(DSCL_EASF_V_BF3_PWL_SEG2, 0, + SCL_EASF_V_BF3_PWL_IN_SEG2, scl_data->dscl_prog_data.easf_v_bf3_pwl_in_set2, + SCL_EASF_V_BF3_PWL_BASE_SEG2, scl_data->dscl_prog_data.easf_v_bf3_pwl_base_set2, SCL_EASF_V_BF3_PWL_SLOPE_SEG2, scl_data->dscl_prog_data.easf_v_bf3_pwl_slope_set2); - REG_UPDATE(DSCL_EASF_V_BF3_PWL_SEG3, - SCL_EASF_V_BF3_PWL_IN_SEG3, scl_data->dscl_prog_data.easf_v_bf3_pwl_in_set3); - REG_UPDATE(DSCL_EASF_V_BF3_PWL_SEG3, - SCL_EASF_V_BF3_PWL_BASE_SEG3, scl_data->dscl_prog_data.easf_v_bf3_pwl_base_set3); - REG_UPDATE(DSCL_EASF_V_BF3_PWL_SEG3, + REG_SET_3(DSCL_EASF_V_BF3_PWL_SEG3, 0, + SCL_EASF_V_BF3_PWL_IN_SEG3, scl_data->dscl_prog_data.easf_v_bf3_pwl_in_set3, + SCL_EASF_V_BF3_PWL_BASE_SEG3, scl_data->dscl_prog_data.easf_v_bf3_pwl_base_set3, SCL_EASF_V_BF3_PWL_SLOPE_SEG3, scl_data->dscl_prog_data.easf_v_bf3_pwl_slope_set3); - REG_UPDATE(DSCL_EASF_V_BF3_PWL_SEG4, - SCL_EASF_V_BF3_PWL_IN_SEG4, scl_data->dscl_prog_data.easf_v_bf3_pwl_in_set4); - REG_UPDATE(DSCL_EASF_V_BF3_PWL_SEG4, - SCL_EASF_V_BF3_PWL_BASE_SEG4, scl_data->dscl_prog_data.easf_v_bf3_pwl_base_set4); - REG_UPDATE(DSCL_EASF_V_BF3_PWL_SEG4, + REG_SET_3(DSCL_EASF_V_BF3_PWL_SEG4, 0, + SCL_EASF_V_BF3_PWL_IN_SEG4, scl_data->dscl_prog_data.easf_v_bf3_pwl_in_set4, + SCL_EASF_V_BF3_PWL_BASE_SEG4, scl_data->dscl_prog_data.easf_v_bf3_pwl_base_set4, SCL_EASF_V_BF3_PWL_SLOPE_SEG4, scl_data->dscl_prog_data.easf_v_bf3_pwl_slope_set4); - REG_UPDATE(DSCL_EASF_V_BF3_PWL_SEG5, - SCL_EASF_V_BF3_PWL_IN_SEG5, scl_data->dscl_prog_data.easf_v_bf3_pwl_in_set5); - REG_UPDATE(DSCL_EASF_V_BF3_PWL_SEG5, + REG_SET_2(DSCL_EASF_V_BF3_PWL_SEG5, 0, + SCL_EASF_V_BF3_PWL_IN_SEG5, scl_data->dscl_prog_data.easf_v_bf3_pwl_in_set5, SCL_EASF_V_BF3_PWL_BASE_SEG5, scl_data->dscl_prog_data.easf_v_bf3_pwl_base_set5); + PERF_TRACE(); +} +/** + * dpp401_dscl_program_easf_h - Program EASF_H + * + * @dpp_base: High level DPP struct + * @scl_data: scalaer_data info + * + * This is the primary function to program horizontal EASF registers + * + */ +static void dpp401_dscl_program_easf_h(struct dpp *dpp_base, const struct scaler_data *scl_data) +{ + struct dcn401_dpp *dpp = TO_DCN401_DPP(dpp_base); + + PERF_TRACE(); /* DSCL_EASF_H_MODE */ - REG_UPDATE(DSCL_EASF_H_MODE, - SCL_EASF_H_EN, scl_data->dscl_prog_data.easf_h_en); - REG_UPDATE(DSCL_EASF_H_MODE, - SCL_EASF_H_2TAP_SHARP_FACTOR, scl_data->dscl_prog_data.easf_h_sharp_factor); - REG_UPDATE(DSCL_EASF_H_MODE, + REG_SET_3(DSCL_EASF_H_MODE, 0, + SCL_EASF_H_EN, scl_data->dscl_prog_data.easf_h_en, + SCL_EASF_H_2TAP_SHARP_FACTOR, scl_data->dscl_prog_data.easf_h_sharp_factor, SCL_EASF_H_RINGEST_FORCE_EN, scl_data->dscl_prog_data.easf_h_ring); - REG_UPDATE(DSCL_EASF_H_BF_CNTL, - SCL_EASF_H_BF1_EN, scl_data->dscl_prog_data.easf_h_bf1_en); - REG_UPDATE(DSCL_EASF_H_BF_CNTL, - SCL_EASF_H_BF2_MODE, scl_data->dscl_prog_data.easf_h_bf2_mode); - REG_UPDATE(DSCL_EASF_H_BF_CNTL, - SCL_EASF_H_BF3_MODE, scl_data->dscl_prog_data.easf_h_bf3_mode); - REG_UPDATE(DSCL_EASF_H_BF_CNTL, - SCL_EASF_H_BF2_FLAT1_GAIN, scl_data->dscl_prog_data.easf_h_bf2_flat1_gain); - REG_UPDATE(DSCL_EASF_H_BF_CNTL, - SCL_EASF_H_BF2_FLAT2_GAIN, scl_data->dscl_prog_data.easf_h_bf2_flat2_gain); - REG_UPDATE(DSCL_EASF_H_BF_CNTL, + + if (!scl_data->dscl_prog_data.easf_h_en) { + PERF_TRACE(); + return; + } + + /* DSCL_EASF_H_BF_CNTL */ + REG_SET_6(DSCL_EASF_H_BF_CNTL, 0, + SCL_EASF_H_BF1_EN, scl_data->dscl_prog_data.easf_h_bf1_en, + SCL_EASF_H_BF2_MODE, scl_data->dscl_prog_data.easf_h_bf2_mode, + SCL_EASF_H_BF3_MODE, scl_data->dscl_prog_data.easf_h_bf3_mode, + SCL_EASF_H_BF2_FLAT1_GAIN, scl_data->dscl_prog_data.easf_h_bf2_flat1_gain, + SCL_EASF_H_BF2_FLAT2_GAIN, scl_data->dscl_prog_data.easf_h_bf2_flat2_gain, SCL_EASF_H_BF2_ROC_GAIN, scl_data->dscl_prog_data.easf_h_bf2_roc_gain); - REG_UPDATE(DSCL_EASF_H_RINGEST_EVENTAP_REDUCE, - SCL_EASF_H_RINGEST_EVENTAP_REDUCEG1, scl_data->dscl_prog_data.easf_h_ringest_eventap_reduceg1); - REG_UPDATE(DSCL_EASF_H_RINGEST_EVENTAP_REDUCE, + /* DSCL_EASF_H_RINGEST_EVENTAP_REDUCE */ + REG_SET_2(DSCL_EASF_H_RINGEST_EVENTAP_REDUCE, 0, + SCL_EASF_H_RINGEST_EVENTAP_REDUCEG1, scl_data->dscl_prog_data.easf_h_ringest_eventap_reduceg1, SCL_EASF_H_RINGEST_EVENTAP_REDUCEG2, scl_data->dscl_prog_data.easf_h_ringest_eventap_reduceg2); - REG_UPDATE(DSCL_EASF_H_RINGEST_EVENTAP_GAIN, - SCL_EASF_H_RINGEST_EVENTAP_GAIN1, scl_data->dscl_prog_data.easf_h_ringest_eventap_gain1); - REG_UPDATE(DSCL_EASF_H_RINGEST_EVENTAP_GAIN, + /* DSCL_EASF_H_RINGEST_EVENTAP_GAIN */ + REG_SET_2(DSCL_EASF_H_RINGEST_EVENTAP_GAIN, 0, + SCL_EASF_H_RINGEST_EVENTAP_GAIN1, scl_data->dscl_prog_data.easf_h_ringest_eventap_gain1, SCL_EASF_H_RINGEST_EVENTAP_GAIN2, scl_data->dscl_prog_data.easf_h_ringest_eventap_gain2); - REG_UPDATE(DSCL_EASF_H_BF_FINAL_MAX_MIN, - SCL_EASF_H_BF_MAXA, scl_data->dscl_prog_data.easf_h_bf_maxa); - REG_UPDATE(DSCL_EASF_H_BF_FINAL_MAX_MIN, - SCL_EASF_H_BF_MAXB, scl_data->dscl_prog_data.easf_h_bf_maxb); - REG_UPDATE(DSCL_EASF_H_BF_FINAL_MAX_MIN, - SCL_EASF_H_BF_MINA, scl_data->dscl_prog_data.easf_h_bf_mina); - REG_UPDATE(DSCL_EASF_H_BF_FINAL_MAX_MIN, + /* DSCL_EASF_H_BF_FINAL_MAX_MIN */ + REG_SET_4(DSCL_EASF_H_BF_FINAL_MAX_MIN, 0, + SCL_EASF_H_BF_MAXA, scl_data->dscl_prog_data.easf_h_bf_maxa, + SCL_EASF_H_BF_MAXB, scl_data->dscl_prog_data.easf_h_bf_maxb, + SCL_EASF_H_BF_MINA, scl_data->dscl_prog_data.easf_h_bf_mina, SCL_EASF_H_BF_MINB, scl_data->dscl_prog_data.easf_h_bf_minb); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG0, - SCL_EASF_H_BF1_PWL_IN_SEG0, scl_data->dscl_prog_data.easf_h_bf1_pwl_in_seg0); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG0, - SCL_EASF_H_BF1_PWL_BASE_SEG0, scl_data->dscl_prog_data.easf_h_bf1_pwl_base_seg0); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG0, + /* DSCL_EASF_H_BF1_PWL_SEGn */ + REG_SET_3(DSCL_EASF_H_BF1_PWL_SEG0, 0, + SCL_EASF_H_BF1_PWL_IN_SEG0, scl_data->dscl_prog_data.easf_h_bf1_pwl_in_seg0, + SCL_EASF_H_BF1_PWL_BASE_SEG0, scl_data->dscl_prog_data.easf_h_bf1_pwl_base_seg0, SCL_EASF_H_BF1_PWL_SLOPE_SEG0, scl_data->dscl_prog_data.easf_h_bf1_pwl_slope_seg0); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG1, - SCL_EASF_H_BF1_PWL_IN_SEG1, scl_data->dscl_prog_data.easf_h_bf1_pwl_in_seg1); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG1, - SCL_EASF_H_BF1_PWL_BASE_SEG1, scl_data->dscl_prog_data.easf_h_bf1_pwl_base_seg1); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG1, + REG_SET_3(DSCL_EASF_H_BF1_PWL_SEG1, 0, + SCL_EASF_H_BF1_PWL_IN_SEG1, scl_data->dscl_prog_data.easf_h_bf1_pwl_in_seg1, + SCL_EASF_H_BF1_PWL_BASE_SEG1, scl_data->dscl_prog_data.easf_h_bf1_pwl_base_seg1, SCL_EASF_H_BF1_PWL_SLOPE_SEG1, scl_data->dscl_prog_data.easf_h_bf1_pwl_slope_seg1); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG2, - SCL_EASF_H_BF1_PWL_IN_SEG2, scl_data->dscl_prog_data.easf_h_bf1_pwl_in_seg2); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG2, - SCL_EASF_H_BF1_PWL_BASE_SEG2, scl_data->dscl_prog_data.easf_h_bf1_pwl_base_seg2); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG2, + REG_SET_3(DSCL_EASF_H_BF1_PWL_SEG2, 0, + SCL_EASF_H_BF1_PWL_IN_SEG2, scl_data->dscl_prog_data.easf_h_bf1_pwl_in_seg2, + SCL_EASF_H_BF1_PWL_BASE_SEG2, scl_data->dscl_prog_data.easf_h_bf1_pwl_base_seg2, SCL_EASF_H_BF1_PWL_SLOPE_SEG2, scl_data->dscl_prog_data.easf_h_bf1_pwl_slope_seg2); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG3, - SCL_EASF_H_BF1_PWL_IN_SEG3, scl_data->dscl_prog_data.easf_h_bf1_pwl_in_seg3); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG3, - SCL_EASF_H_BF1_PWL_BASE_SEG3, scl_data->dscl_prog_data.easf_h_bf1_pwl_base_seg3); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG3, + REG_SET_3(DSCL_EASF_H_BF1_PWL_SEG3, 0, + SCL_EASF_H_BF1_PWL_IN_SEG3, scl_data->dscl_prog_data.easf_h_bf1_pwl_in_seg3, + SCL_EASF_H_BF1_PWL_BASE_SEG3, scl_data->dscl_prog_data.easf_h_bf1_pwl_base_seg3, SCL_EASF_H_BF1_PWL_SLOPE_SEG3, scl_data->dscl_prog_data.easf_h_bf1_pwl_slope_seg3); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG4, - SCL_EASF_H_BF1_PWL_IN_SEG4, scl_data->dscl_prog_data.easf_h_bf1_pwl_in_seg4); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG4, - SCL_EASF_H_BF1_PWL_BASE_SEG4, scl_data->dscl_prog_data.easf_h_bf1_pwl_base_seg4); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG4, + REG_SET_3(DSCL_EASF_H_BF1_PWL_SEG4, 0, + SCL_EASF_H_BF1_PWL_IN_SEG4, scl_data->dscl_prog_data.easf_h_bf1_pwl_in_seg4, + SCL_EASF_H_BF1_PWL_BASE_SEG4, scl_data->dscl_prog_data.easf_h_bf1_pwl_base_seg4, SCL_EASF_H_BF1_PWL_SLOPE_SEG4, scl_data->dscl_prog_data.easf_h_bf1_pwl_slope_seg4); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG5, - SCL_EASF_H_BF1_PWL_IN_SEG5, scl_data->dscl_prog_data.easf_h_bf1_pwl_in_seg5); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG5, - SCL_EASF_H_BF1_PWL_BASE_SEG5, scl_data->dscl_prog_data.easf_h_bf1_pwl_base_seg5); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG5, + REG_SET_3(DSCL_EASF_H_BF1_PWL_SEG5, 0, + SCL_EASF_H_BF1_PWL_IN_SEG5, scl_data->dscl_prog_data.easf_h_bf1_pwl_in_seg5, + SCL_EASF_H_BF1_PWL_BASE_SEG5, scl_data->dscl_prog_data.easf_h_bf1_pwl_base_seg5, SCL_EASF_H_BF1_PWL_SLOPE_SEG5, scl_data->dscl_prog_data.easf_h_bf1_pwl_slope_seg5); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG6, - SCL_EASF_H_BF1_PWL_IN_SEG6, scl_data->dscl_prog_data.easf_h_bf1_pwl_in_seg6); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG6, - SCL_EASF_H_BF1_PWL_BASE_SEG6, scl_data->dscl_prog_data.easf_h_bf1_pwl_base_seg6); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG6, + REG_SET_3(DSCL_EASF_H_BF1_PWL_SEG6, 0, + SCL_EASF_H_BF1_PWL_IN_SEG6, scl_data->dscl_prog_data.easf_h_bf1_pwl_in_seg6, + SCL_EASF_H_BF1_PWL_BASE_SEG6, scl_data->dscl_prog_data.easf_h_bf1_pwl_base_seg6, SCL_EASF_H_BF1_PWL_SLOPE_SEG6, scl_data->dscl_prog_data.easf_h_bf1_pwl_slope_seg6); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG7, - SCL_EASF_H_BF1_PWL_IN_SEG7, scl_data->dscl_prog_data.easf_h_bf1_pwl_in_seg7); - REG_UPDATE(DSCL_EASF_H_BF1_PWL_SEG7, + REG_SET_2(DSCL_EASF_H_BF1_PWL_SEG7, 0, + SCL_EASF_H_BF1_PWL_IN_SEG7, scl_data->dscl_prog_data.easf_h_bf1_pwl_in_seg7, SCL_EASF_H_BF1_PWL_BASE_SEG7, scl_data->dscl_prog_data.easf_h_bf1_pwl_base_seg7); - REG_UPDATE(DSCL_EASF_H_BF3_PWL_SEG0, - SCL_EASF_H_BF3_PWL_IN_SEG0, scl_data->dscl_prog_data.easf_h_bf3_pwl_in_set0); - REG_UPDATE(DSCL_EASF_H_BF3_PWL_SEG0, - SCL_EASF_H_BF3_PWL_BASE_SEG0, scl_data->dscl_prog_data.easf_h_bf3_pwl_base_set0); - REG_UPDATE(DSCL_EASF_H_BF3_PWL_SEG0, + /* DSCL_EASF_H_BF3_PWL_SEGn */ + REG_SET_3(DSCL_EASF_H_BF3_PWL_SEG0, 0, + SCL_EASF_H_BF3_PWL_IN_SEG0, scl_data->dscl_prog_data.easf_h_bf3_pwl_in_set0, + SCL_EASF_H_BF3_PWL_BASE_SEG0, scl_data->dscl_prog_data.easf_h_bf3_pwl_base_set0, SCL_EASF_H_BF3_PWL_SLOPE_SEG0, scl_data->dscl_prog_data.easf_h_bf3_pwl_slope_set0); - REG_UPDATE(DSCL_EASF_H_BF3_PWL_SEG1, - SCL_EASF_H_BF3_PWL_IN_SEG1, scl_data->dscl_prog_data.easf_h_bf3_pwl_in_set1); - REG_UPDATE(DSCL_EASF_H_BF3_PWL_SEG1, - SCL_EASF_H_BF3_PWL_BASE_SEG1, scl_data->dscl_prog_data.easf_h_bf3_pwl_base_set1); - REG_UPDATE(DSCL_EASF_H_BF3_PWL_SEG1, + REG_SET_3(DSCL_EASF_H_BF3_PWL_SEG1, 0, + SCL_EASF_H_BF3_PWL_IN_SEG1, scl_data->dscl_prog_data.easf_h_bf3_pwl_in_set1, + SCL_EASF_H_BF3_PWL_BASE_SEG1, scl_data->dscl_prog_data.easf_h_bf3_pwl_base_set1, SCL_EASF_H_BF3_PWL_SLOPE_SEG1, scl_data->dscl_prog_data.easf_h_bf3_pwl_slope_set1); - REG_UPDATE(DSCL_EASF_H_BF3_PWL_SEG2, - SCL_EASF_H_BF3_PWL_IN_SEG2, scl_data->dscl_prog_data.easf_h_bf3_pwl_in_set2); - REG_UPDATE(DSCL_EASF_H_BF3_PWL_SEG2, - SCL_EASF_H_BF3_PWL_BASE_SEG2, scl_data->dscl_prog_data.easf_h_bf3_pwl_base_set2); - REG_UPDATE(DSCL_EASF_H_BF3_PWL_SEG2, + REG_SET_3(DSCL_EASF_H_BF3_PWL_SEG2, 0, + SCL_EASF_H_BF3_PWL_IN_SEG2, scl_data->dscl_prog_data.easf_h_bf3_pwl_in_set2, + SCL_EASF_H_BF3_PWL_BASE_SEG2, scl_data->dscl_prog_data.easf_h_bf3_pwl_base_set2, SCL_EASF_H_BF3_PWL_SLOPE_SEG2, scl_data->dscl_prog_data.easf_h_bf3_pwl_slope_set2); - REG_UPDATE(DSCL_EASF_H_BF3_PWL_SEG3, - SCL_EASF_H_BF3_PWL_IN_SEG3, scl_data->dscl_prog_data.easf_h_bf3_pwl_in_set3); - REG_UPDATE(DSCL_EASF_H_BF3_PWL_SEG3, - SCL_EASF_H_BF3_PWL_BASE_SEG3, scl_data->dscl_prog_data.easf_h_bf3_pwl_base_set3); - REG_UPDATE(DSCL_EASF_H_BF3_PWL_SEG3, + REG_SET_3(DSCL_EASF_H_BF3_PWL_SEG3, 0, + SCL_EASF_H_BF3_PWL_IN_SEG3, scl_data->dscl_prog_data.easf_h_bf3_pwl_in_set3, + SCL_EASF_H_BF3_PWL_BASE_SEG3, scl_data->dscl_prog_data.easf_h_bf3_pwl_base_set3, SCL_EASF_H_BF3_PWL_SLOPE_SEG3, scl_data->dscl_prog_data.easf_h_bf3_pwl_slope_set3); - REG_UPDATE(DSCL_EASF_H_BF3_PWL_SEG4, - SCL_EASF_H_BF3_PWL_IN_SEG4, scl_data->dscl_prog_data.easf_h_bf3_pwl_in_set4); - REG_UPDATE(DSCL_EASF_H_BF3_PWL_SEG4, - SCL_EASF_H_BF3_PWL_BASE_SEG4, scl_data->dscl_prog_data.easf_h_bf3_pwl_base_set4); - REG_UPDATE(DSCL_EASF_H_BF3_PWL_SEG4, + REG_SET_3(DSCL_EASF_H_BF3_PWL_SEG4, 0, + SCL_EASF_H_BF3_PWL_IN_SEG4, scl_data->dscl_prog_data.easf_h_bf3_pwl_in_set4, + SCL_EASF_H_BF3_PWL_BASE_SEG4, scl_data->dscl_prog_data.easf_h_bf3_pwl_base_set4, SCL_EASF_H_BF3_PWL_SLOPE_SEG4, scl_data->dscl_prog_data.easf_h_bf3_pwl_slope_set4); - REG_UPDATE(DSCL_EASF_H_BF3_PWL_SEG5, - SCL_EASF_H_BF3_PWL_IN_SEG5, scl_data->dscl_prog_data.easf_h_bf3_pwl_in_set5); - REG_UPDATE(DSCL_EASF_H_BF3_PWL_SEG5, + REG_SET_2(DSCL_EASF_H_BF3_PWL_SEG5, 0, + SCL_EASF_H_BF3_PWL_IN_SEG5, scl_data->dscl_prog_data.easf_h_bf3_pwl_in_set5, SCL_EASF_H_BF3_PWL_BASE_SEG5, scl_data->dscl_prog_data.easf_h_bf3_pwl_base_set5); + PERF_TRACE(); +} +/** + * dpp401_dscl_program_easf - Program EASF + * + * @dpp_base: High level DPP struct + * @scl_data: scalaer_data info + * + * This is the primary function to program EASF + * + */ +static void dpp401_dscl_program_easf(struct dpp *dpp_base, const struct scaler_data *scl_data) +{ + struct dcn401_dpp *dpp = TO_DCN401_DPP(dpp_base); + + PERF_TRACE(); + /* DSCL_SC_MODE */ + REG_SET_2(DSCL_SC_MODE, 0, + SCL_SC_MATRIX_MODE, scl_data->dscl_prog_data.easf_matrix_mode, + SCL_SC_LTONL_EN, scl_data->dscl_prog_data.easf_ltonl_en); /* DSCL_EASF_SC_MATRIX_C0C1, DSCL_EASF_SC_MATRIX_C2C3 */ - REG_UPDATE(DSCL_SC_MATRIX_C0C1, - SCL_SC_MATRIX_C0, scl_data->dscl_prog_data.easf_matrix_c0); - REG_UPDATE(DSCL_SC_MATRIX_C0C1, + REG_SET_2(DSCL_SC_MATRIX_C0C1, 0, + SCL_SC_MATRIX_C0, scl_data->dscl_prog_data.easf_matrix_c0, SCL_SC_MATRIX_C1, scl_data->dscl_prog_data.easf_matrix_c1); - REG_UPDATE(DSCL_SC_MATRIX_C2C3, - SCL_SC_MATRIX_C2, scl_data->dscl_prog_data.easf_matrix_c2); - REG_UPDATE(DSCL_SC_MATRIX_C2C3, + REG_SET_2(DSCL_SC_MATRIX_C2C3, 0, + SCL_SC_MATRIX_C2, scl_data->dscl_prog_data.easf_matrix_c2, SCL_SC_MATRIX_C3, scl_data->dscl_prog_data.easf_matrix_c3); + dpp401_dscl_program_easf_v(dpp_base, scl_data); + dpp401_dscl_program_easf_h(dpp_base, scl_data); PERF_TRACE(); } /** @@ -958,10 +937,11 @@ static void dpp401_dscl_set_isharp_filter( REG_UPDATE(ISHARP_DELTA_CTRL, ISHARP_DELTA_LUT_HOST_SELECT, 0); + /* LUT data write is auto-indexed. Write index once */ + REG_SET(ISHARP_DELTA_INDEX, 0, + ISHARP_DELTA_INDEX, 0); for (level = 0; level < NUM_LEVELS; level++) { filter_data = filter[level]; - REG_SET(ISHARP_DELTA_INDEX, 0, - ISHARP_DELTA_INDEX, level); REG_SET(ISHARP_DELTA_DATA, 0, ISHARP_DELTA_DATA, filter_data); } @@ -976,107 +956,76 @@ static void dpp401_dscl_set_isharp_filter( * */ static void dpp401_dscl_program_isharp(struct dpp *dpp_base, - const struct scaler_data *scl_data) + const struct scaler_data *scl_data, + bool program_isharp_1dlut, + bool *bs_coeffs_updated) { struct dcn401_dpp *dpp = TO_DCN401_DPP(dpp_base); + *bs_coeffs_updated = false; PERF_TRACE(); - /* ISHARP_EN */ - REG_UPDATE(ISHARP_MODE, - ISHARP_EN, scl_data->dscl_prog_data.isharp_en); - /* ISHARP_NOISEDET_EN */ - REG_UPDATE(ISHARP_MODE, - ISHARP_NOISEDET_EN, scl_data->dscl_prog_data.isharp_noise_det.enable); - /* ISHARP_NOISEDET_MODE */ - REG_UPDATE(ISHARP_MODE, - ISHARP_NOISEDET_MODE, scl_data->dscl_prog_data.isharp_noise_det.mode); - /* ISHARP_NOISEDET_UTHRE */ - REG_UPDATE(ISHARP_NOISEDET_THRESHOLD, - ISHARP_NOISEDET_UTHRE, scl_data->dscl_prog_data.isharp_noise_det.uthreshold); - /* ISHARP_NOISEDET_DTHRE */ - REG_UPDATE(ISHARP_NOISEDET_THRESHOLD, - ISHARP_NOISEDET_DTHRE, scl_data->dscl_prog_data.isharp_noise_det.dthreshold); - REG_UPDATE(ISHARP_MODE, - ISHARP_NOISEDET_MODE, scl_data->dscl_prog_data.isharp_noise_det.mode); - /* ISHARP_NOISEDET_UTHRE */ - REG_UPDATE(ISHARP_NOISEDET_THRESHOLD, - ISHARP_NOISEDET_UTHRE, scl_data->dscl_prog_data.isharp_noise_det.uthreshold); - /* ISHARP_NOISEDET_DTHRE */ - REG_UPDATE(ISHARP_NOISEDET_THRESHOLD, + /* ISHARP_MODE */ + REG_SET_6(ISHARP_MODE, 0, + ISHARP_EN, scl_data->dscl_prog_data.isharp_en, + ISHARP_NOISEDET_EN, scl_data->dscl_prog_data.isharp_noise_det.enable, + ISHARP_NOISEDET_MODE, scl_data->dscl_prog_data.isharp_noise_det.mode, + ISHARP_LBA_MODE, scl_data->dscl_prog_data.isharp_lba.mode, + ISHARP_FMT_MODE, scl_data->dscl_prog_data.isharp_fmt.mode, + ISHARP_FMT_NORM, scl_data->dscl_prog_data.isharp_fmt.norm); + + /* Skip remaining register programming if ISHARP is disabled */ + if (!scl_data->dscl_prog_data.isharp_en) { + PERF_TRACE(); + return; + } + + /* ISHARP_NOISEDET_THRESHOLD */ + REG_SET_2(ISHARP_NOISEDET_THRESHOLD, 0, + ISHARP_NOISEDET_UTHRE, scl_data->dscl_prog_data.isharp_noise_det.uthreshold, ISHARP_NOISEDET_DTHRE, scl_data->dscl_prog_data.isharp_noise_det.dthreshold); - /* ISHARP_NOISEDET_PWL_START_IN */ - REG_UPDATE(ISHARP_NOISE_GAIN_PWL, - ISHARP_NOISEDET_PWL_START_IN, scl_data->dscl_prog_data.isharp_noise_det.pwl_start_in); - /* ISHARP_NOISEDET_PWL_END_IN */ - REG_UPDATE(ISHARP_NOISE_GAIN_PWL, - ISHARP_NOISEDET_PWL_END_IN, scl_data->dscl_prog_data.isharp_noise_det.pwl_end_in); - /* ISHARP_NOISEDET_PWL_SLOPE */ - REG_UPDATE(ISHARP_NOISE_GAIN_PWL, + + /* ISHARP_NOISE_GAIN_PWL */ + REG_SET_3(ISHARP_NOISE_GAIN_PWL, 0, + ISHARP_NOISEDET_PWL_START_IN, scl_data->dscl_prog_data.isharp_noise_det.pwl_start_in, + ISHARP_NOISEDET_PWL_END_IN, scl_data->dscl_prog_data.isharp_noise_det.pwl_end_in, ISHARP_NOISEDET_PWL_SLOPE, scl_data->dscl_prog_data.isharp_noise_det.pwl_slope); - /* ISHARP_LBA_MODE */ - REG_UPDATE(ISHARP_MODE, - ISHARP_LBA_MODE, scl_data->dscl_prog_data.isharp_lba.mode); + /* ISHARP_LBA: IN_SEG, BASE_SEG, SLOPE_SEG */ - REG_UPDATE(ISHARP_LBA_PWL_SEG0, - ISHARP_LBA_PWL_IN_SEG0, scl_data->dscl_prog_data.isharp_lba.in_seg[0]); - REG_UPDATE(ISHARP_LBA_PWL_SEG0, - ISHARP_LBA_PWL_BASE_SEG0, scl_data->dscl_prog_data.isharp_lba.base_seg[0]); - REG_UPDATE(ISHARP_LBA_PWL_SEG0, + REG_SET_3(ISHARP_LBA_PWL_SEG0, 0, + ISHARP_LBA_PWL_IN_SEG0, scl_data->dscl_prog_data.isharp_lba.in_seg[0], + ISHARP_LBA_PWL_BASE_SEG0, scl_data->dscl_prog_data.isharp_lba.base_seg[0], ISHARP_LBA_PWL_SLOPE_SEG0, scl_data->dscl_prog_data.isharp_lba.slope_seg[0]); - REG_UPDATE(ISHARP_LBA_PWL_SEG1, - ISHARP_LBA_PWL_IN_SEG1, scl_data->dscl_prog_data.isharp_lba.in_seg[1]); - REG_UPDATE(ISHARP_LBA_PWL_SEG1, - ISHARP_LBA_PWL_BASE_SEG1, scl_data->dscl_prog_data.isharp_lba.base_seg[1]); - REG_UPDATE(ISHARP_LBA_PWL_SEG1, + REG_SET_3(ISHARP_LBA_PWL_SEG1, 0, + ISHARP_LBA_PWL_IN_SEG1, scl_data->dscl_prog_data.isharp_lba.in_seg[1], + ISHARP_LBA_PWL_BASE_SEG1, scl_data->dscl_prog_data.isharp_lba.base_seg[1], ISHARP_LBA_PWL_SLOPE_SEG1, scl_data->dscl_prog_data.isharp_lba.slope_seg[1]); - REG_UPDATE(ISHARP_LBA_PWL_SEG2, - ISHARP_LBA_PWL_IN_SEG2, scl_data->dscl_prog_data.isharp_lba.in_seg[2]); - REG_UPDATE(ISHARP_LBA_PWL_SEG2, - ISHARP_LBA_PWL_BASE_SEG2, scl_data->dscl_prog_data.isharp_lba.base_seg[2]); - REG_UPDATE(ISHARP_LBA_PWL_SEG2, + REG_SET_3(ISHARP_LBA_PWL_SEG2, 0, + ISHARP_LBA_PWL_IN_SEG2, scl_data->dscl_prog_data.isharp_lba.in_seg[2], + ISHARP_LBA_PWL_BASE_SEG2, scl_data->dscl_prog_data.isharp_lba.base_seg[2], ISHARP_LBA_PWL_SLOPE_SEG2, scl_data->dscl_prog_data.isharp_lba.slope_seg[2]); - REG_UPDATE(ISHARP_LBA_PWL_SEG3, - ISHARP_LBA_PWL_IN_SEG3, scl_data->dscl_prog_data.isharp_lba.in_seg[3]); - REG_UPDATE(ISHARP_LBA_PWL_SEG3, - ISHARP_LBA_PWL_BASE_SEG3, scl_data->dscl_prog_data.isharp_lba.base_seg[3]); - REG_UPDATE(ISHARP_LBA_PWL_SEG3, + REG_SET_3(ISHARP_LBA_PWL_SEG3, 0, + ISHARP_LBA_PWL_IN_SEG3, scl_data->dscl_prog_data.isharp_lba.in_seg[3], + ISHARP_LBA_PWL_BASE_SEG3, scl_data->dscl_prog_data.isharp_lba.base_seg[3], ISHARP_LBA_PWL_SLOPE_SEG3, scl_data->dscl_prog_data.isharp_lba.slope_seg[3]); - REG_UPDATE(ISHARP_LBA_PWL_SEG4, - ISHARP_LBA_PWL_IN_SEG4, scl_data->dscl_prog_data.isharp_lba.in_seg[4]); - REG_UPDATE(ISHARP_LBA_PWL_SEG4, - ISHARP_LBA_PWL_BASE_SEG4, scl_data->dscl_prog_data.isharp_lba.base_seg[4]); - REG_UPDATE(ISHARP_LBA_PWL_SEG4, + REG_SET_3(ISHARP_LBA_PWL_SEG4, 0, + ISHARP_LBA_PWL_IN_SEG4, scl_data->dscl_prog_data.isharp_lba.in_seg[4], + ISHARP_LBA_PWL_BASE_SEG4, scl_data->dscl_prog_data.isharp_lba.base_seg[4], ISHARP_LBA_PWL_SLOPE_SEG4, scl_data->dscl_prog_data.isharp_lba.slope_seg[4]); - REG_UPDATE(ISHARP_LBA_PWL_SEG5, - ISHARP_LBA_PWL_IN_SEG5, scl_data->dscl_prog_data.isharp_lba.in_seg[5]); - REG_UPDATE(ISHARP_LBA_PWL_SEG5, + REG_SET_2(ISHARP_LBA_PWL_SEG5, 0, + ISHARP_LBA_PWL_IN_SEG5, scl_data->dscl_prog_data.isharp_lba.in_seg[5], ISHARP_LBA_PWL_BASE_SEG5, scl_data->dscl_prog_data.isharp_lba.base_seg[5]); - /* ISHARP_FMT_MODE */ - REG_UPDATE(ISHARP_MODE, - ISHARP_FMT_MODE, scl_data->dscl_prog_data.isharp_fmt.mode); - /* ISHARP_FMT_NORM */ - REG_UPDATE(ISHARP_MODE, - ISHARP_FMT_NORM, scl_data->dscl_prog_data.isharp_fmt.norm); /* ISHARP_DELTA_LUT */ - dpp401_dscl_set_isharp_filter(dpp, scl_data->dscl_prog_data.isharp_delta); - /* ISHARP_NLDELTA_SCLIP_EN_P */ - REG_UPDATE(ISHARP_NLDELTA_SOFT_CLIP, - ISHARP_NLDELTA_SCLIP_EN_P, scl_data->dscl_prog_data.isharp_nldelta_sclip.enable_p); - /* ISHARP_NLDELTA_SCLIP_PIVOT_P */ - REG_UPDATE(ISHARP_NLDELTA_SOFT_CLIP, - ISHARP_NLDELTA_SCLIP_PIVOT_P, scl_data->dscl_prog_data.isharp_nldelta_sclip.pivot_p); - /* ISHARP_NLDELTA_SCLIP_SLOPE_P */ - REG_UPDATE(ISHARP_NLDELTA_SOFT_CLIP, - ISHARP_NLDELTA_SCLIP_SLOPE_P, scl_data->dscl_prog_data.isharp_nldelta_sclip.slope_p); - /* ISHARP_NLDELTA_SCLIP_EN_N */ - REG_UPDATE(ISHARP_NLDELTA_SOFT_CLIP, - ISHARP_NLDELTA_SCLIP_EN_N, scl_data->dscl_prog_data.isharp_nldelta_sclip.enable_n); - /* ISHARP_NLDELTA_SCLIP_PIVOT_N */ - REG_UPDATE(ISHARP_NLDELTA_SOFT_CLIP, - ISHARP_NLDELTA_SCLIP_PIVOT_N, scl_data->dscl_prog_data.isharp_nldelta_sclip.pivot_n); - /* ISHARP_NLDELTA_SCLIP_SLOPE_N */ - REG_UPDATE(ISHARP_NLDELTA_SOFT_CLIP, + if (!program_isharp_1dlut) + dpp401_dscl_set_isharp_filter(dpp, scl_data->dscl_prog_data.isharp_delta); + + /* ISHARP_NLDELTA_SOFT_CLIP */ + REG_SET_6(ISHARP_NLDELTA_SOFT_CLIP, 0, + ISHARP_NLDELTA_SCLIP_EN_P, scl_data->dscl_prog_data.isharp_nldelta_sclip.enable_p, + ISHARP_NLDELTA_SCLIP_PIVOT_P, scl_data->dscl_prog_data.isharp_nldelta_sclip.pivot_p, + ISHARP_NLDELTA_SCLIP_SLOPE_P, scl_data->dscl_prog_data.isharp_nldelta_sclip.slope_p, + ISHARP_NLDELTA_SCLIP_EN_N, scl_data->dscl_prog_data.isharp_nldelta_sclip.enable_n, + ISHARP_NLDELTA_SCLIP_PIVOT_N, scl_data->dscl_prog_data.isharp_nldelta_sclip.pivot_n, ISHARP_NLDELTA_SCLIP_SLOPE_N, scl_data->dscl_prog_data.isharp_nldelta_sclip.slope_n); /* Blur and Scale Coefficients - SCL_COEF_RAM_TAP_SELECT */ @@ -1086,12 +1035,14 @@ static void dpp401_dscl_program_isharp(struct dpp *dpp_base, dpp, scl_data->taps.v_taps, SCL_COEF_VERTICAL_BLUR_SCALE, scl_data->dscl_prog_data.filter_blur_scale_v); + *bs_coeffs_updated = true; } if (scl_data->dscl_prog_data.filter_blur_scale_h) { dpp401_dscl_set_scaler_filter( dpp, scl_data->taps.h_taps, SCL_COEF_HORIZONTAL_BLUR_SCALE, scl_data->dscl_prog_data.filter_blur_scale_h); + *bs_coeffs_updated = true; } } PERF_TRACE(); @@ -1122,12 +1073,29 @@ void dpp401_dscl_set_scaler_manual_scale(struct dpp *dpp_base, dpp_base, scl_data, dpp_base->ctx->dc->debug.always_scale); bool ycbcr = scl_data->format >= PIXEL_FORMAT_VIDEO_BEGIN && scl_data->format <= PIXEL_FORMAT_VIDEO_END; + bool program_isharp_1dlut = false; + bool bs_coeffs_updated = false; + if (memcmp(&dpp->scl_data, scl_data, sizeof(*scl_data)) == 0) return; PERF_TRACE(); + /* If only sharpness has changed, then only update 1dlut, then return */ + if (scl_data->dscl_prog_data.isharp_en && + (dpp->scl_data.dscl_prog_data.sharpness_level + != scl_data->dscl_prog_data.sharpness_level)) { + /* ISHARP_DELTA_LUT */ + dpp401_dscl_set_isharp_filter(dpp, scl_data->dscl_prog_data.isharp_delta); + dpp->scl_data.dscl_prog_data.sharpness_level = scl_data->dscl_prog_data.sharpness_level; + dpp->scl_data.dscl_prog_data.isharp_delta = scl_data->dscl_prog_data.isharp_delta; + + if (memcmp(&dpp->scl_data, scl_data, sizeof(*scl_data)) == 0) + return; + program_isharp_1dlut = true; + } + dpp->scl_data = *scl_data; if ((dpp->base.ctx->dc->config.use_spl) && (!dpp->base.ctx->dc->debug.disable_spl)) { @@ -1181,7 +1149,7 @@ void dpp401_dscl_set_scaler_manual_scale(struct dpp *dpp_base, if (dscl_mode == DSCL_MODE_SCALING_444_BYPASS) { if (dpp->base.ctx->dc->config.prefer_easf) dpp401_dscl_disable_easf(dpp_base, scl_data); - dpp401_dscl_program_isharp(dpp_base, scl_data); + dpp401_dscl_program_isharp(dpp_base, scl_data, program_isharp_1dlut, &bs_coeffs_updated); return; } @@ -1208,12 +1176,18 @@ void dpp401_dscl_set_scaler_manual_scale(struct dpp *dpp_base, SCL_V_NUM_TAPS_C, v_num_taps_c, SCL_H_NUM_TAPS_C, h_num_taps_c); - dpp401_dscl_set_scl_filter(dpp, scl_data, ycbcr); + /* ISharp configuration + * - B&S coeffs are written to same coeff RAM as WB scaler coeffs + * - coeff RAM toggle is in EASF programming + * - if we are only programming B&S coeffs, then need to reprogram + * WB scaler coeffs and toggle coeff RAM together + */ + //if (dpp->base.ctx->dc->config.prefer_easf) + dpp401_dscl_program_isharp(dpp_base, scl_data, program_isharp_1dlut, &bs_coeffs_updated); + + dpp401_dscl_set_scl_filter(dpp, scl_data, ycbcr, bs_coeffs_updated); /* Edge adaptive scaler function configuration */ if (dpp->base.ctx->dc->config.prefer_easf) dpp401_dscl_program_easf(dpp_base, scl_data); - /* isharp configuration */ - //if (dpp->base.ctx->dc->config.prefer_easf) - dpp401_dscl_program_isharp(dpp_base, scl_data); PERF_TRACE(); } diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c index fbbb20b9dbee..a4c6decee0f8 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c @@ -147,37 +147,6 @@ void dcn35_init_hw(struct dc *dc) hws->funcs.bios_golden_init(dc); } - if (!dc->debug.disable_clock_gate) { - REG_WRITE(DCCG_GATE_DISABLE_CNTL, 0); - REG_WRITE(DCCG_GATE_DISABLE_CNTL2, 0); - - /* Disable gating for PHYASYMCLK. This will be enabled in dccg if needed */ - REG_UPDATE_5(DCCG_GATE_DISABLE_CNTL2, PHYASYMCLK_ROOT_GATE_DISABLE, 1, - PHYBSYMCLK_ROOT_GATE_DISABLE, 1, - PHYCSYMCLK_ROOT_GATE_DISABLE, 1, - PHYDSYMCLK_ROOT_GATE_DISABLE, 1, - PHYESYMCLK_ROOT_GATE_DISABLE, 1); - - REG_UPDATE_4(DCCG_GATE_DISABLE_CNTL4, - DPIASYMCLK0_GATE_DISABLE, 0, - DPIASYMCLK1_GATE_DISABLE, 0, - DPIASYMCLK2_GATE_DISABLE, 0, - DPIASYMCLK3_GATE_DISABLE, 0); - - REG_WRITE(DCCG_GATE_DISABLE_CNTL5, 0xFFFFFFFF); - REG_UPDATE_4(DCCG_GATE_DISABLE_CNTL5, - DTBCLK_P0_GATE_DISABLE, 0, - DTBCLK_P1_GATE_DISABLE, 0, - DTBCLK_P2_GATE_DISABLE, 0, - DTBCLK_P3_GATE_DISABLE, 0); - REG_UPDATE_4(DCCG_GATE_DISABLE_CNTL5, - DPSTREAMCLK0_GATE_DISABLE, 0, - DPSTREAMCLK1_GATE_DISABLE, 0, - DPSTREAMCLK2_GATE_DISABLE, 0, - DPSTREAMCLK3_GATE_DISABLE, 0); - - } - // Initialize the dccg if (res_pool->dccg->funcs->dccg_init) res_pool->dccg->funcs->dccg_init(res_pool->dccg); @@ -305,20 +274,6 @@ void dcn35_init_hw(struct dc *dc) if (!dc->debug.disable_clock_gate) { /* enable all DCN clock gating */ - REG_WRITE(DCCG_GATE_DISABLE_CNTL, 0); - - REG_UPDATE_5(DCCG_GATE_DISABLE_CNTL2, SYMCLKA_FE_GATE_DISABLE, 0, - SYMCLKB_FE_GATE_DISABLE, 0, - SYMCLKC_FE_GATE_DISABLE, 0, - SYMCLKD_FE_GATE_DISABLE, 0, - SYMCLKE_FE_GATE_DISABLE, 0); - REG_UPDATE(DCCG_GATE_DISABLE_CNTL2, HDMICHARCLK0_GATE_DISABLE, 0); - REG_UPDATE_5(DCCG_GATE_DISABLE_CNTL2, SYMCLKA_GATE_DISABLE, 0, - SYMCLKB_GATE_DISABLE, 0, - SYMCLKC_GATE_DISABLE, 0, - SYMCLKD_GATE_DISABLE, 0, - SYMCLKE_GATE_DISABLE, 0); - REG_UPDATE(DCFCLK_CNTL, DCFCLK_GATE_DIS, 0); } diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c index 02e63b95c36d..9d56fbdcd06a 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c @@ -76,6 +76,9 @@ #include "dml2/dml2_wrapper.h" +#include "spl/dc_spl_scl_easf_filters.h" +#include "spl/dc_spl_isharp_filters.h" + #define DC_LOGGER_INIT(logger) enum dcn401_clk_src_array_id { @@ -2126,6 +2129,10 @@ static bool dcn401_resource_construct( dc->dml2_options.max_segments_per_hubp = 20; dc->dml2_options.det_segment_size = DCN4_01_CRB_SEGMENT_SIZE_KB; + /* SPL */ + spl_init_easf_filter_coeffs(); + spl_init_blur_scale_coeffs(); + return true; create_fail: diff --git a/drivers/gpu/drm/amd/display/dc/spl/Makefile b/drivers/gpu/drm/amd/display/dc/spl/Makefile index f8df85ea4d32..5edf3c6cf3e2 100644 --- a/drivers/gpu/drm/amd/display/dc/spl/Makefile +++ b/drivers/gpu/drm/amd/display/dc/spl/Makefile @@ -23,7 +23,7 @@ # Makefile for the 'spl' sub-component of DAL. # It provides the scaling library interface. -SPL = dc_spl.o dc_spl_scl_filters.o dc_spl_scl_filters_old.o dc_spl_isharp_filters.o +SPL = dc_spl.o dc_spl_scl_filters.o dc_spl_scl_easf_filters.o dc_spl_isharp_filters.o dc_spl_filters.o spl_fixpt31_32.o spl_custom_float.o AMD_DAL_SPL = $(addprefix $(AMDDALPATH)/dc/spl/,$(SPL)) diff --git a/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c b/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c index 9eccdb38bed4..15f7eda903e6 100644 --- a/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c +++ b/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c @@ -4,9 +4,11 @@ #include "dc_spl.h" #include "dc_spl_scl_filters.h" +#include "dc_spl_scl_easf_filters.h" #include "dc_spl_isharp_filters.h" +#include "spl_debug.h" -#define IDENTITY_RATIO(ratio) (dc_fixpt_u2d19(ratio) == (1 << 19)) +#define IDENTITY_RATIO(ratio) (spl_fixpt_u2d19(ratio) == (1 << 19)) #define MIN_VIEWPORT_SIZE 12 static struct spl_rect intersect_rec(const struct spl_rect *r0, const struct spl_rect *r1) @@ -107,26 +109,26 @@ static struct spl_rect calculate_plane_rec_in_timing_active( const struct spl_rect *stream_src = &spl_in->basic_out.src_rect; const struct spl_rect *stream_dst = &spl_in->basic_out.dst_rect; struct spl_rect rec_out = {0}; - struct fixed31_32 temp; + struct spl_fixed31_32 temp; - temp = dc_fixpt_from_fraction(rec_in->x * (long long)stream_dst->width, + temp = spl_fixpt_from_fraction(rec_in->x * (long long)stream_dst->width, stream_src->width); - rec_out.x = stream_dst->x + dc_fixpt_round(temp); + rec_out.x = stream_dst->x + spl_fixpt_round(temp); - temp = dc_fixpt_from_fraction( + temp = spl_fixpt_from_fraction( (rec_in->x + rec_in->width) * (long long)stream_dst->width, stream_src->width); - rec_out.width = stream_dst->x + dc_fixpt_round(temp) - rec_out.x; + rec_out.width = stream_dst->x + spl_fixpt_round(temp) - rec_out.x; - temp = dc_fixpt_from_fraction(rec_in->y * (long long)stream_dst->height, + temp = spl_fixpt_from_fraction(rec_in->y * (long long)stream_dst->height, stream_src->height); - rec_out.y = stream_dst->y + dc_fixpt_round(temp); + rec_out.y = stream_dst->y + spl_fixpt_round(temp); - temp = dc_fixpt_from_fraction( + temp = spl_fixpt_from_fraction( (rec_in->y + rec_in->height) * (long long)stream_dst->height, stream_src->height); - rec_out.height = stream_dst->y + dc_fixpt_round(temp) - rec_out.y; + rec_out.height = stream_dst->y + spl_fixpt_round(temp) - rec_out.y; return rec_out; } @@ -144,7 +146,7 @@ static struct spl_rect calculate_mpc_slice_in_timing_active( mpc_rec.x = plane_clip_rec->x + mpc_rec.width * mpc_slice_idx; mpc_rec.height = plane_clip_rec->height; mpc_rec.y = plane_clip_rec->y; - ASSERT(mpc_slice_count == 1 || + SPL_ASSERT(mpc_slice_count == 1 || spl_in->basic_out.view_format != SPL_VIEW_3D_SIDE_BY_SIDE || mpc_rec.width % 2 == 0); @@ -157,7 +159,7 @@ static struct spl_rect calculate_mpc_slice_in_timing_active( } if (spl_in->basic_out.view_format == SPL_VIEW_3D_TOP_AND_BOTTOM) { - ASSERT(mpc_rec.height % 2 == 0); + SPL_ASSERT(mpc_rec.height % 2 == 0); mpc_rec.height /= 2; } return mpc_rec; @@ -197,7 +199,7 @@ static struct spl_rect calculate_odm_slice_in_timing_active(struct spl_in *spl_i return spl_in->basic_out.odm_slice_rect; } -static void spl_calculate_recout(struct spl_in *spl_in, struct spl_out *spl_out) +static void spl_calculate_recout(struct spl_in *spl_in, struct spl_scratch *spl_scratch, struct spl_out *spl_out) { /* * A plane clip represents the desired plane size and position in Stream @@ -340,20 +342,23 @@ static void spl_calculate_recout(struct spl_in *spl_in, struct spl_out *spl_out) /* shift the overlapping area so it is with respect to current * ODM slice's position */ - spl_out->scl_data.recout = shift_rec( + spl_scratch->scl_data.recout = shift_rec( &overlapping_area, -odm_slice.x, -odm_slice.y); - spl_out->scl_data.recout.height -= + spl_scratch->scl_data.recout.height -= spl_in->debug.visual_confirm_base_offset; - spl_out->scl_data.recout.height -= + spl_scratch->scl_data.recout.height -= spl_in->debug.visual_confirm_dpp_offset; } else /* if there is no overlap, zero recout */ - memset(&spl_out->scl_data.recout, 0, + memset(&spl_scratch->scl_data.recout, 0, sizeof(struct spl_rect)); } + /* Calculate scaling ratios */ -static void spl_calculate_scaling_ratios(struct spl_in *spl_in, struct spl_out *spl_out) +static void spl_calculate_scaling_ratios(struct spl_in *spl_in, + struct spl_scratch *spl_scratch, + struct spl_out *spl_out) { const int in_w = spl_in->basic_out.src_rect.width; const int in_h = spl_in->basic_out.src_rect.height; @@ -364,59 +369,75 @@ static void spl_calculate_scaling_ratios(struct spl_in *spl_in, struct spl_out * /*Swap surf_src height and width since scaling ratios are in recout rotation*/ if (spl_in->basic_in.rotation == SPL_ROTATION_ANGLE_90 || spl_in->basic_in.rotation == SPL_ROTATION_ANGLE_270) - swap(surf_src.height, surf_src.width); + spl_swap(surf_src.height, surf_src.width); - spl_out->scl_data.ratios.horz = dc_fixpt_from_fraction( + spl_scratch->scl_data.ratios.horz = spl_fixpt_from_fraction( surf_src.width, spl_in->basic_in.dst_rect.width); - spl_out->scl_data.ratios.vert = dc_fixpt_from_fraction( + spl_scratch->scl_data.ratios.vert = spl_fixpt_from_fraction( surf_src.height, spl_in->basic_in.dst_rect.height); if (spl_in->basic_out.view_format == SPL_VIEW_3D_SIDE_BY_SIDE) - spl_out->scl_data.ratios.horz.value *= 2; + spl_scratch->scl_data.ratios.horz.value *= 2; else if (spl_in->basic_out.view_format == SPL_VIEW_3D_TOP_AND_BOTTOM) - spl_out->scl_data.ratios.vert.value *= 2; + spl_scratch->scl_data.ratios.vert.value *= 2; - spl_out->scl_data.ratios.vert.value = div64_s64( - spl_out->scl_data.ratios.vert.value * in_h, out_h); - spl_out->scl_data.ratios.horz.value = div64_s64( - spl_out->scl_data.ratios.horz.value * in_w, out_w); + spl_scratch->scl_data.ratios.vert.value = spl_div64_s64( + spl_scratch->scl_data.ratios.vert.value * in_h, out_h); + spl_scratch->scl_data.ratios.horz.value = spl_div64_s64( + spl_scratch->scl_data.ratios.horz.value * in_w, out_w); - spl_out->scl_data.ratios.horz_c = spl_out->scl_data.ratios.horz; - spl_out->scl_data.ratios.vert_c = spl_out->scl_data.ratios.vert; + spl_scratch->scl_data.ratios.horz_c = spl_scratch->scl_data.ratios.horz; + spl_scratch->scl_data.ratios.vert_c = spl_scratch->scl_data.ratios.vert; if (spl_in->basic_in.format == SPL_PIXEL_FORMAT_420BPP8 || spl_in->basic_in.format == SPL_PIXEL_FORMAT_420BPP10) { - spl_out->scl_data.ratios.horz_c.value /= 2; - spl_out->scl_data.ratios.vert_c.value /= 2; + spl_scratch->scl_data.ratios.horz_c.value /= 2; + spl_scratch->scl_data.ratios.vert_c.value /= 2; } - spl_out->scl_data.ratios.horz = dc_fixpt_truncate( - spl_out->scl_data.ratios.horz, 19); - spl_out->scl_data.ratios.vert = dc_fixpt_truncate( - spl_out->scl_data.ratios.vert, 19); - spl_out->scl_data.ratios.horz_c = dc_fixpt_truncate( - spl_out->scl_data.ratios.horz_c, 19); - spl_out->scl_data.ratios.vert_c = dc_fixpt_truncate( - spl_out->scl_data.ratios.vert_c, 19); + spl_scratch->scl_data.ratios.horz = spl_fixpt_truncate( + spl_scratch->scl_data.ratios.horz, 19); + spl_scratch->scl_data.ratios.vert = spl_fixpt_truncate( + spl_scratch->scl_data.ratios.vert, 19); + spl_scratch->scl_data.ratios.horz_c = spl_fixpt_truncate( + spl_scratch->scl_data.ratios.horz_c, 19); + spl_scratch->scl_data.ratios.vert_c = spl_fixpt_truncate( + spl_scratch->scl_data.ratios.vert_c, 19); + + /* + * Coefficient table and some registers are different based on ratio + * that is output/input. Currently we calculate input/output + * Store 1/ratio in recip_ratio for those lookups + */ + spl_scratch->scl_data.recip_ratios.horz = spl_fixpt_recip( + spl_scratch->scl_data.ratios.horz); + spl_scratch->scl_data.recip_ratios.vert = spl_fixpt_recip( + spl_scratch->scl_data.ratios.vert); + spl_scratch->scl_data.recip_ratios.horz_c = spl_fixpt_recip( + spl_scratch->scl_data.ratios.horz_c); + spl_scratch->scl_data.recip_ratios.vert_c = spl_fixpt_recip( + spl_scratch->scl_data.ratios.vert_c); } + /* Calculate Viewport size */ -static void spl_calculate_viewport_size(struct spl_in *spl_in, struct spl_out *spl_out) +static void spl_calculate_viewport_size(struct spl_in *spl_in, struct spl_scratch *spl_scratch) { - spl_out->scl_data.viewport.width = dc_fixpt_ceil(dc_fixpt_mul_int(spl_out->scl_data.ratios.horz, - spl_out->scl_data.recout.width)); - spl_out->scl_data.viewport.height = dc_fixpt_ceil(dc_fixpt_mul_int(spl_out->scl_data.ratios.vert, - spl_out->scl_data.recout.height)); - spl_out->scl_data.viewport_c.width = dc_fixpt_ceil(dc_fixpt_mul_int(spl_out->scl_data.ratios.horz_c, - spl_out->scl_data.recout.width)); - spl_out->scl_data.viewport_c.height = dc_fixpt_ceil(dc_fixpt_mul_int(spl_out->scl_data.ratios.vert_c, - spl_out->scl_data.recout.height)); + spl_scratch->scl_data.viewport.width = spl_fixpt_ceil(spl_fixpt_mul_int(spl_scratch->scl_data.ratios.horz, + spl_scratch->scl_data.recout.width)); + spl_scratch->scl_data.viewport.height = spl_fixpt_ceil(spl_fixpt_mul_int(spl_scratch->scl_data.ratios.vert, + spl_scratch->scl_data.recout.height)); + spl_scratch->scl_data.viewport_c.width = spl_fixpt_ceil(spl_fixpt_mul_int(spl_scratch->scl_data.ratios.horz_c, + spl_scratch->scl_data.recout.width)); + spl_scratch->scl_data.viewport_c.height = spl_fixpt_ceil(spl_fixpt_mul_int(spl_scratch->scl_data.ratios.vert_c, + spl_scratch->scl_data.recout.height)); if (spl_in->basic_in.rotation == SPL_ROTATION_ANGLE_90 || spl_in->basic_in.rotation == SPL_ROTATION_ANGLE_270) { - swap(spl_out->scl_data.viewport.width, spl_out->scl_data.viewport.height); - swap(spl_out->scl_data.viewport_c.width, spl_out->scl_data.viewport_c.height); + spl_swap(spl_scratch->scl_data.viewport.width, spl_scratch->scl_data.viewport.height); + spl_swap(spl_scratch->scl_data.viewport_c.width, spl_scratch->scl_data.viewport_c.height); } } + static void spl_get_vp_scan_direction(enum spl_rotation_angle rotation, bool horizontal_mirror, bool *orthogonal_rotation, @@ -440,6 +461,7 @@ static void spl_get_vp_scan_direction(enum spl_rotation_angle rotation, if (horizontal_mirror) *flip_horz_scan_dir = !*flip_horz_scan_dir; } + /* * We completely calculate vp offset, size and inits here based entirely on scaling * ratios and recout for pixel perfect pipe combine. @@ -449,13 +471,13 @@ static void spl_calculate_init_and_vp(bool flip_scan_dir, int recout_size, int src_size, int taps, - struct fixed31_32 ratio, - struct fixed31_32 init_adj, - struct fixed31_32 *init, + struct spl_fixed31_32 ratio, + struct spl_fixed31_32 init_adj, + struct spl_fixed31_32 *init, int *vp_offset, int *vp_size) { - struct fixed31_32 temp; + struct spl_fixed31_32 temp; int int_part; /* @@ -468,33 +490,33 @@ static void spl_calculate_init_and_vp(bool flip_scan_dir, * init_bot = init + scaling_ratio * to get pixel perfect combine add the fraction from calculating vp offset */ - temp = dc_fixpt_mul_int(ratio, recout_offset_within_recout_full); - *vp_offset = dc_fixpt_floor(temp); + temp = spl_fixpt_mul_int(ratio, recout_offset_within_recout_full); + *vp_offset = spl_fixpt_floor(temp); temp.value &= 0xffffffff; - *init = dc_fixpt_add(dc_fixpt_div_int(dc_fixpt_add_int(ratio, taps + 1), 2), temp); - *init = dc_fixpt_add(*init, init_adj); - *init = dc_fixpt_truncate(*init, 19); + *init = spl_fixpt_add(spl_fixpt_div_int(spl_fixpt_add_int(ratio, taps + 1), 2), temp); + *init = spl_fixpt_add(*init, init_adj); + *init = spl_fixpt_truncate(*init, 19); /* * If viewport has non 0 offset and there are more taps than covered by init then * we should decrease the offset and increase init so we are never sampling * outside of viewport. */ - int_part = dc_fixpt_floor(*init); + int_part = spl_fixpt_floor(*init); if (int_part < taps) { int_part = taps - int_part; if (int_part > *vp_offset) int_part = *vp_offset; *vp_offset -= int_part; - *init = dc_fixpt_add_int(*init, int_part); + *init = spl_fixpt_add_int(*init, int_part); } /* * If taps are sampling outside of viewport at end of recout and there are more pixels * available in the surface we should increase the viewport size, regardless set vp to * only what is used. */ - temp = dc_fixpt_add(*init, dc_fixpt_mul_int(ratio, recout_size - 1)); - *vp_size = dc_fixpt_floor(temp); + temp = spl_fixpt_add(*init, spl_fixpt_mul_int(ratio, recout_size - 1)); + *vp_size = spl_fixpt_floor(temp); if (*vp_size + *vp_offset > src_size) *vp_size = src_size - *vp_offset; @@ -509,15 +531,24 @@ static void spl_calculate_init_and_vp(bool flip_scan_dir, static bool spl_is_yuv420(enum spl_pixel_format format) { - if ((format >= SPL_PIXEL_FORMAT_VIDEO_BEGIN) && - (format <= SPL_PIXEL_FORMAT_VIDEO_END)) + if ((format >= SPL_PIXEL_FORMAT_420BPP8) && + (format <= SPL_PIXEL_FORMAT_420BPP10)) + return true; + + return false; +} + +static bool spl_is_rgb8(enum spl_pixel_format format) +{ + if (format == SPL_PIXEL_FORMAT_ARGB8888) return true; return false; } /*Calculate inits and viewport */ -static void spl_calculate_inits_and_viewports(struct spl_in *spl_in, struct spl_out *spl_out) +static void spl_calculate_inits_and_viewports(struct spl_in *spl_in, + struct spl_scratch *spl_scratch) { struct spl_rect src = spl_in->basic_in.src_rect; struct spl_rect recout_dst_in_active_timing; @@ -528,11 +559,11 @@ static void spl_calculate_inits_and_viewports(struct spl_in *spl_in, struct spl_ int vpc_div = (spl_in->basic_in.format == SPL_PIXEL_FORMAT_420BPP8 || spl_in->basic_in.format == SPL_PIXEL_FORMAT_420BPP10) ? 2 : 1; bool orthogonal_rotation, flip_vert_scan_dir, flip_horz_scan_dir; - struct fixed31_32 init_adj_h = dc_fixpt_zero; - struct fixed31_32 init_adj_v = dc_fixpt_zero; + struct spl_fixed31_32 init_adj_h = spl_fixpt_zero; + struct spl_fixed31_32 init_adj_v = spl_fixpt_zero; recout_clip_in_active_timing = shift_rec( - &spl_out->scl_data.recout, odm_slice.x, odm_slice.y); + &spl_scratch->scl_data.recout, odm_slice.x, odm_slice.y); recout_dst_in_active_timing = calculate_plane_rec_in_timing_active( spl_in, &spl_in->basic_in.dst_rect); overlap_in_active_timing = intersect_rec(&recout_clip_in_active_timing, @@ -555,8 +586,8 @@ static void spl_calculate_inits_and_viewports(struct spl_in *spl_in, struct spl_ &flip_horz_scan_dir); if (orthogonal_rotation) { - swap(src.width, src.height); - swap(flip_vert_scan_dir, flip_horz_scan_dir); + spl_swap(src.width, src.height); + spl_swap(flip_vert_scan_dir, flip_horz_scan_dir); } if (spl_is_yuv420(spl_in->basic_in.format)) { @@ -568,17 +599,17 @@ static void spl_calculate_inits_and_viewports(struct spl_in *spl_in, struct spl_ switch (spl_in->basic_in.cositing) { case CHROMA_COSITING_LEFT: - init_adj_h = dc_fixpt_zero; - init_adj_v = dc_fixpt_from_fraction(sign, 2); + init_adj_h = spl_fixpt_zero; + init_adj_v = spl_fixpt_from_fraction(sign, 4); break; case CHROMA_COSITING_NONE: - init_adj_h = dc_fixpt_from_fraction(sign, 2); - init_adj_v = dc_fixpt_from_fraction(sign, 2); + init_adj_h = spl_fixpt_from_fraction(sign, 4); + init_adj_v = spl_fixpt_from_fraction(sign, 4); break; case CHROMA_COSITING_TOPLEFT: default: - init_adj_h = dc_fixpt_zero; - init_adj_v = dc_fixpt_zero; + init_adj_h = spl_fixpt_zero; + init_adj_v = spl_fixpt_zero; break; } } @@ -586,59 +617,60 @@ static void spl_calculate_inits_and_viewports(struct spl_in *spl_in, struct spl_ spl_calculate_init_and_vp( flip_horz_scan_dir, recout_clip_in_recout_dst.x, - spl_out->scl_data.recout.width, + spl_scratch->scl_data.recout.width, src.width, - spl_out->scl_data.taps.h_taps, - spl_out->scl_data.ratios.horz, - dc_fixpt_zero, - &spl_out->scl_data.inits.h, - &spl_out->scl_data.viewport.x, - &spl_out->scl_data.viewport.width); + spl_scratch->scl_data.taps.h_taps, + spl_scratch->scl_data.ratios.horz, + spl_fixpt_zero, + &spl_scratch->scl_data.inits.h, + &spl_scratch->scl_data.viewport.x, + &spl_scratch->scl_data.viewport.width); spl_calculate_init_and_vp( flip_horz_scan_dir, recout_clip_in_recout_dst.x, - spl_out->scl_data.recout.width, + spl_scratch->scl_data.recout.width, src.width / vpc_div, - spl_out->scl_data.taps.h_taps_c, - spl_out->scl_data.ratios.horz_c, + spl_scratch->scl_data.taps.h_taps_c, + spl_scratch->scl_data.ratios.horz_c, init_adj_h, - &spl_out->scl_data.inits.h_c, - &spl_out->scl_data.viewport_c.x, - &spl_out->scl_data.viewport_c.width); + &spl_scratch->scl_data.inits.h_c, + &spl_scratch->scl_data.viewport_c.x, + &spl_scratch->scl_data.viewport_c.width); spl_calculate_init_and_vp( flip_vert_scan_dir, recout_clip_in_recout_dst.y, - spl_out->scl_data.recout.height, + spl_scratch->scl_data.recout.height, src.height, - spl_out->scl_data.taps.v_taps, - spl_out->scl_data.ratios.vert, - dc_fixpt_zero, - &spl_out->scl_data.inits.v, - &spl_out->scl_data.viewport.y, - &spl_out->scl_data.viewport.height); + spl_scratch->scl_data.taps.v_taps, + spl_scratch->scl_data.ratios.vert, + spl_fixpt_zero, + &spl_scratch->scl_data.inits.v, + &spl_scratch->scl_data.viewport.y, + &spl_scratch->scl_data.viewport.height); spl_calculate_init_and_vp( flip_vert_scan_dir, recout_clip_in_recout_dst.y, - spl_out->scl_data.recout.height, + spl_scratch->scl_data.recout.height, src.height / vpc_div, - spl_out->scl_data.taps.v_taps_c, - spl_out->scl_data.ratios.vert_c, + spl_scratch->scl_data.taps.v_taps_c, + spl_scratch->scl_data.ratios.vert_c, init_adj_v, - &spl_out->scl_data.inits.v_c, - &spl_out->scl_data.viewport_c.y, - &spl_out->scl_data.viewport_c.height); + &spl_scratch->scl_data.inits.v_c, + &spl_scratch->scl_data.viewport_c.y, + &spl_scratch->scl_data.viewport_c.height); if (orthogonal_rotation) { - swap(spl_out->scl_data.viewport.x, spl_out->scl_data.viewport.y); - swap(spl_out->scl_data.viewport.width, spl_out->scl_data.viewport.height); - swap(spl_out->scl_data.viewport_c.x, spl_out->scl_data.viewport_c.y); - swap(spl_out->scl_data.viewport_c.width, spl_out->scl_data.viewport_c.height); + spl_swap(spl_scratch->scl_data.viewport.x, spl_scratch->scl_data.viewport.y); + spl_swap(spl_scratch->scl_data.viewport.width, spl_scratch->scl_data.viewport.height); + spl_swap(spl_scratch->scl_data.viewport_c.x, spl_scratch->scl_data.viewport_c.y); + spl_swap(spl_scratch->scl_data.viewport_c.width, spl_scratch->scl_data.viewport_c.height); } - spl_out->scl_data.viewport.x += src.x; - spl_out->scl_data.viewport.y += src.y; - ASSERT(src.x % vpc_div == 0 && src.y % vpc_div == 0); - spl_out->scl_data.viewport_c.x += src.x / vpc_div; - spl_out->scl_data.viewport_c.y += src.y / vpc_div; + spl_scratch->scl_data.viewport.x += src.x; + spl_scratch->scl_data.viewport.y += src.y; + SPL_ASSERT(src.x % vpc_div == 0 && src.y % vpc_div == 0); + spl_scratch->scl_data.viewport_c.x += src.x / vpc_div; + spl_scratch->scl_data.viewport_c.y += src.y / vpc_div; } + static void spl_handle_3d_recout(struct spl_in *spl_in, struct spl_rect *recout) { /* @@ -647,7 +679,7 @@ static void spl_handle_3d_recout(struct spl_in *spl_in, struct spl_rect *recout) * This may break with rotation, good thing we aren't mixing hw rotation and 3d */ if (spl_in->basic_in.mpc_combine_v) { - ASSERT(spl_in->basic_in.rotation == SPL_ROTATION_ANGLE_0 || + SPL_ASSERT(spl_in->basic_in.rotation == SPL_ROTATION_ANGLE_0 || (spl_in->basic_out.view_format != SPL_VIEW_3D_TOP_AND_BOTTOM && spl_in->basic_out.view_format != SPL_VIEW_3D_SIDE_BY_SIDE)); if (spl_in->basic_out.view_format == SPL_VIEW_3D_TOP_AND_BOTTOM) @@ -665,6 +697,7 @@ static void spl_clamp_viewport(struct spl_rect *viewport) if (viewport->width < MIN_VIEWPORT_SIZE) viewport->width = MIN_VIEWPORT_SIZE; } + static bool spl_dscl_is_420_format(enum spl_pixel_format format) { if (format == SPL_PIXEL_FORMAT_420BPP8 || @@ -673,6 +706,7 @@ static bool spl_dscl_is_420_format(enum spl_pixel_format format) else return false; } + static bool spl_dscl_is_video_format(enum spl_pixel_format format) { if (format >= SPL_PIXEL_FORMAT_VIDEO_BEGIN @@ -681,17 +715,21 @@ static bool spl_dscl_is_video_format(enum spl_pixel_format format) else return false; } + static enum scl_mode spl_get_dscl_mode(const struct spl_in *spl_in, - const struct spl_scaler_data *data) + const struct spl_scaler_data *data, + bool enable_isharp, bool enable_easf) { - const long long one = dc_fixpt_one.value; + const long long one = spl_fixpt_one.value; enum spl_pixel_format pixel_format = spl_in->basic_in.format; + /* Bypass if ratio is 1:1 with no ISHARP or force scale on */ if (data->ratios.horz.value == one && data->ratios.vert.value == one && data->ratios.horz_c.value == one && data->ratios.vert_c.value == one - && !spl_in->basic_out.always_scale) + && !spl_in->basic_out.always_scale + && !enable_isharp) return SCL_MODE_SCALING_444_BYPASS; if (!spl_dscl_is_420_format(pixel_format)) { @@ -700,69 +738,196 @@ static enum scl_mode spl_get_dscl_mode(const struct spl_in *spl_in, else return SCL_MODE_SCALING_444_RGB_ENABLE; } - if (data->ratios.horz.value == one && data->ratios.vert.value == one) - return SCL_MODE_SCALING_420_LUMA_BYPASS; - if (data->ratios.horz_c.value == one && data->ratios.vert_c.value == one) - return SCL_MODE_SCALING_420_CHROMA_BYPASS; + + /* Bypass YUV if at 1:1 with no ISHARP or if doing 2:1 YUV + * downscale without EASF + */ + if ((!enable_isharp) && (!enable_easf)) { + if (data->ratios.horz.value == one && data->ratios.vert.value == one) + return SCL_MODE_SCALING_420_LUMA_BYPASS; + if (data->ratios.horz_c.value == one && data->ratios.vert_c.value == one) + return SCL_MODE_SCALING_420_CHROMA_BYPASS; + } return SCL_MODE_SCALING_420_YCBCR_ENABLE; } + +static bool spl_choose_lls_policy(enum spl_pixel_format format, + enum spl_transfer_func_type tf_type, + enum spl_transfer_func_predefined tf_predefined_type, + enum linear_light_scaling *lls_pref) +{ + if (spl_is_yuv420(format)) { + *lls_pref = LLS_PREF_NO; + if ((tf_type == SPL_TF_TYPE_PREDEFINED) || + (tf_type == SPL_TF_TYPE_DISTRIBUTED_POINTS)) + return true; + } else { /* RGB or YUV444 */ + if ((tf_type == SPL_TF_TYPE_PREDEFINED) || + (tf_type == SPL_TF_TYPE_BYPASS)) { + *lls_pref = LLS_PREF_YES; + return true; + } + } + *lls_pref = LLS_PREF_NO; + return false; +} + +/* Enable EASF ?*/ +static bool enable_easf(struct spl_in *spl_in, struct spl_scratch *spl_scratch) +{ + int vratio = 0; + int hratio = 0; + bool skip_easf = false; + bool lls_enable_easf = true; + + if (spl_in->disable_easf) + skip_easf = true; + + vratio = spl_fixpt_ceil(spl_scratch->scl_data.ratios.vert); + hratio = spl_fixpt_ceil(spl_scratch->scl_data.ratios.horz); + + /* + * No EASF support for downscaling > 2:1 + * EASF support for upscaling or downscaling up to 2:1 + */ + if ((vratio > 2) || (hratio > 2)) + skip_easf = true; + + /* + * If lls_pref is LLS_PREF_DONT_CARE, then use pixel format and transfer + * function to determine whether to use LINEAR or NONLINEAR scaling + */ + if (spl_in->lls_pref == LLS_PREF_DONT_CARE) + lls_enable_easf = spl_choose_lls_policy(spl_in->basic_in.format, + spl_in->basic_in.tf_type, spl_in->basic_in.tf_predefined_type, + &spl_in->lls_pref); + + if (!lls_enable_easf) + skip_easf = true; + + /* Check for linear scaling or EASF preferred */ + if (spl_in->lls_pref != LLS_PREF_YES && !spl_in->prefer_easf) + skip_easf = true; + + return skip_easf; +} + +static bool spl_get_isharp_en(struct spl_in *spl_in, + struct spl_scratch *spl_scratch) +{ + bool enable_isharp = false; + int vratio = 0; + int hratio = 0; + struct spl_taps taps = spl_scratch->scl_data.taps; + + /* Return if adaptive sharpness is disabled */ + if (spl_in->adaptive_sharpness.enable == false) + return enable_isharp; + + vratio = spl_fixpt_ceil(spl_scratch->scl_data.ratios.vert); + hratio = spl_fixpt_ceil(spl_scratch->scl_data.ratios.horz); + + /* No iSHARP support for downscaling */ + if (vratio > 1 || hratio > 1) + return enable_isharp; + + // Scaling is up to 1:1 (no scaling) or upscaling + + /* + * Apply sharpness to all RGB surfaces and to + * NV12/P010 surfaces + */ + + /* + * Apply sharpness if supports horizontal taps 4,6 AND + * vertical taps 3, 4, 6 + */ + if ((taps.h_taps == 4 || taps.h_taps == 6) && + (taps.v_taps == 3 || taps.v_taps == 4 || taps.v_taps == 6)) + enable_isharp = true; + + return enable_isharp; +} + /* Calculate optimal number of taps */ static bool spl_get_optimal_number_of_taps( - int max_downscale_src_width, struct spl_in *spl_in, struct spl_out *spl_out, - const struct spl_taps *in_taps) + int max_downscale_src_width, struct spl_in *spl_in, struct spl_scratch *spl_scratch, + const struct spl_taps *in_taps, bool *enable_easf_v, bool *enable_easf_h, + bool *enable_isharp) { int num_part_y, num_part_c; int max_taps_y, max_taps_c; int min_taps_y, min_taps_c; enum lb_memory_config lb_config; + bool skip_easf = false; - if (spl_out->scl_data.viewport.width > spl_out->scl_data.h_active && + if (spl_scratch->scl_data.viewport.width > spl_scratch->scl_data.h_active && max_downscale_src_width != 0 && - spl_out->scl_data.viewport.width > max_downscale_src_width) + spl_scratch->scl_data.viewport.width > max_downscale_src_width) return false; + + /* Check if we are using EASF or not */ + skip_easf = enable_easf(spl_in, spl_scratch); + /* * Set default taps if none are provided * From programming guide: taps = min{ ceil(2*H_RATIO,1), 8} for downscaling * taps = 4 for upscaling */ - if (in_taps->h_taps == 0) { - if (dc_fixpt_ceil(spl_out->scl_data.ratios.horz) > 1) - spl_out->scl_data.taps.h_taps = min(2 * dc_fixpt_ceil(spl_out->scl_data.ratios.horz), 8); - else - spl_out->scl_data.taps.h_taps = 4; - } else - spl_out->scl_data.taps.h_taps = in_taps->h_taps; - if (in_taps->v_taps == 0) { - if (dc_fixpt_ceil(spl_out->scl_data.ratios.vert) > 1) - spl_out->scl_data.taps.v_taps = min(dc_fixpt_ceil(dc_fixpt_mul_int( - spl_out->scl_data.ratios.vert, 2)), 8); - else - spl_out->scl_data.taps.v_taps = 4; - } else - spl_out->scl_data.taps.v_taps = in_taps->v_taps; - if (in_taps->v_taps_c == 0) { - if (dc_fixpt_ceil(spl_out->scl_data.ratios.vert_c) > 1) - spl_out->scl_data.taps.v_taps_c = min(dc_fixpt_ceil(dc_fixpt_mul_int( - spl_out->scl_data.ratios.vert_c, 2)), 8); - else - spl_out->scl_data.taps.v_taps_c = 4; - } else - spl_out->scl_data.taps.v_taps_c = in_taps->v_taps_c; - if (in_taps->h_taps_c == 0) { - if (dc_fixpt_ceil(spl_out->scl_data.ratios.horz_c) > 1) - spl_out->scl_data.taps.h_taps_c = min(2 * dc_fixpt_ceil(spl_out->scl_data.ratios.horz_c), 8); + if (skip_easf) { + if (in_taps->h_taps == 0) { + if (spl_fixpt_ceil(spl_scratch->scl_data.ratios.horz) > 1) + spl_scratch->scl_data.taps.h_taps = spl_min(2 * spl_fixpt_ceil( + spl_scratch->scl_data.ratios.horz), 8); + else + spl_scratch->scl_data.taps.h_taps = 4; + } else + spl_scratch->scl_data.taps.h_taps = in_taps->h_taps; + if (in_taps->v_taps == 0) { + if (spl_fixpt_ceil(spl_scratch->scl_data.ratios.vert) > 1) + spl_scratch->scl_data.taps.v_taps = spl_min(spl_fixpt_ceil(spl_fixpt_mul_int( + spl_scratch->scl_data.ratios.vert, 2)), 8); + else + spl_scratch->scl_data.taps.v_taps = 4; + } else + spl_scratch->scl_data.taps.v_taps = in_taps->v_taps; + if (in_taps->v_taps_c == 0) { + if (spl_fixpt_ceil(spl_scratch->scl_data.ratios.vert_c) > 1) + spl_scratch->scl_data.taps.v_taps_c = spl_min(spl_fixpt_ceil(spl_fixpt_mul_int( + spl_scratch->scl_data.ratios.vert_c, 2)), 8); + else + spl_scratch->scl_data.taps.v_taps_c = 4; + } else + spl_scratch->scl_data.taps.v_taps_c = in_taps->v_taps_c; + if (in_taps->h_taps_c == 0) { + if (spl_fixpt_ceil(spl_scratch->scl_data.ratios.horz_c) > 1) + spl_scratch->scl_data.taps.h_taps_c = spl_min(2 * spl_fixpt_ceil( + spl_scratch->scl_data.ratios.horz_c), 8); + else + spl_scratch->scl_data.taps.h_taps_c = 4; + } else if ((in_taps->h_taps_c % 2) != 0 && in_taps->h_taps_c != 1) + /* Only 1 and even h_taps_c are supported by hw */ + spl_scratch->scl_data.taps.h_taps_c = in_taps->h_taps_c - 1; else - spl_out->scl_data.taps.h_taps_c = 4; - } else if ((in_taps->h_taps_c % 2) != 0 && in_taps->h_taps_c != 1) - /* Only 1 and even h_taps_c are supported by hw */ - spl_out->scl_data.taps.h_taps_c = in_taps->h_taps_c - 1; - else - spl_out->scl_data.taps.h_taps_c = in_taps->h_taps_c; + spl_scratch->scl_data.taps.h_taps_c = in_taps->h_taps_c; + } else { + if (spl_is_yuv420(spl_in->basic_in.format)) { + spl_scratch->scl_data.taps.h_taps = 6; + spl_scratch->scl_data.taps.v_taps = 6; + spl_scratch->scl_data.taps.h_taps_c = 4; + spl_scratch->scl_data.taps.v_taps_c = 4; + } else { /* RGB */ + spl_scratch->scl_data.taps.h_taps = 6; + spl_scratch->scl_data.taps.v_taps = 6; + spl_scratch->scl_data.taps.h_taps_c = 6; + spl_scratch->scl_data.taps.v_taps_c = 6; + } + } /*Ensure we can support the requested number of vtaps*/ - min_taps_y = dc_fixpt_ceil(spl_out->scl_data.ratios.vert); - min_taps_c = dc_fixpt_ceil(spl_out->scl_data.ratios.vert_c); + min_taps_y = spl_fixpt_ceil(spl_scratch->scl_data.ratios.vert); + min_taps_c = spl_fixpt_ceil(spl_scratch->scl_data.ratios.vert_c); /* Use LB_MEMORY_CONFIG_3 for 4:2:0 */ if ((spl_in->basic_in.format == SPL_PIXEL_FORMAT_420BPP8) @@ -771,16 +936,16 @@ static bool spl_get_optimal_number_of_taps( else lb_config = LB_MEMORY_CONFIG_0; // Determine max vtap support by calculating how much line buffer can fit - spl_in->funcs->spl_calc_lb_num_partitions(spl_in->basic_out.alpha_en, &spl_out->scl_data, + spl_in->funcs->spl_calc_lb_num_partitions(spl_in->basic_out.alpha_en, &spl_scratch->scl_data, lb_config, &num_part_y, &num_part_c); /* MAX_V_TAPS = MIN (NUM_LINES - MAX(CEILING(V_RATIO,1)-2, 0), 8) */ - if (dc_fixpt_ceil(spl_out->scl_data.ratios.vert) > 2) - max_taps_y = num_part_y - (dc_fixpt_ceil(spl_out->scl_data.ratios.vert) - 2); + if (spl_fixpt_ceil(spl_scratch->scl_data.ratios.vert) > 2) + max_taps_y = num_part_y - (spl_fixpt_ceil(spl_scratch->scl_data.ratios.vert) - 2); else max_taps_y = num_part_y; - if (dc_fixpt_ceil(spl_out->scl_data.ratios.vert_c) > 2) - max_taps_c = num_part_c - (dc_fixpt_ceil(spl_out->scl_data.ratios.vert_c) - 2); + if (spl_fixpt_ceil(spl_scratch->scl_data.ratios.vert_c) > 2) + max_taps_c = num_part_c - (spl_fixpt_ceil(spl_scratch->scl_data.ratios.vert_c) - 2); else max_taps_c = num_part_c; @@ -789,48 +954,108 @@ static bool spl_get_optimal_number_of_taps( else if (max_taps_c < min_taps_c) return false; - if (spl_out->scl_data.taps.v_taps > max_taps_y) - spl_out->scl_data.taps.v_taps = max_taps_y; - - if (spl_out->scl_data.taps.v_taps_c > max_taps_c) - spl_out->scl_data.taps.v_taps_c = max_taps_c; - if (spl_in->prefer_easf) { - // EASF can be enabled only for taps 3,4,6 - // If optimal no of taps is 5, then set it to 4 - // If optimal no of taps is 7 or 8, then set it to 6 - if (spl_out->scl_data.taps.v_taps == 5) - spl_out->scl_data.taps.v_taps = 4; - if (spl_out->scl_data.taps.v_taps == 7 || spl_out->scl_data.taps.v_taps == 8) - spl_out->scl_data.taps.v_taps = 6; - - if (spl_out->scl_data.taps.v_taps_c == 5) - spl_out->scl_data.taps.v_taps_c = 4; - if (spl_out->scl_data.taps.v_taps_c == 7 || spl_out->scl_data.taps.v_taps_c == 8) - spl_out->scl_data.taps.v_taps_c = 6; - - if (spl_out->scl_data.taps.h_taps == 5) - spl_out->scl_data.taps.h_taps = 4; - if (spl_out->scl_data.taps.h_taps == 7 || spl_out->scl_data.taps.h_taps == 8) - spl_out->scl_data.taps.h_taps = 6; - - if (spl_out->scl_data.taps.h_taps_c == 5) - spl_out->scl_data.taps.h_taps_c = 4; - if (spl_out->scl_data.taps.h_taps_c == 7 || spl_out->scl_data.taps.h_taps_c == 8) - spl_out->scl_data.taps.h_taps_c = 6; + if (spl_scratch->scl_data.taps.v_taps > max_taps_y) + spl_scratch->scl_data.taps.v_taps = max_taps_y; + if (spl_scratch->scl_data.taps.v_taps_c > max_taps_c) + spl_scratch->scl_data.taps.v_taps_c = max_taps_c; + + if (!skip_easf) { + /* + * RGB ( L + NL ) and Linear HDR support 6x6, 6x4, 6x3, 4x4, 4x3 + * NL YUV420 only supports 6x6, 6x4 for Y and 4x4 for UV + * + * If LB does not support 3, 4, or 6 taps, then disable EASF_V + * and only enable EASF_H. So for RGB, support 6x2, 4x2 + * and for NL YUV420, support 6x2 for Y and 4x2 for UV + * + * All other cases, have to disable EASF_V and EASF_H + * + * If optimal no of taps is 5, then set it to 4 + * If optimal no of taps is 7 or 8, then fine since max tap is 6 + * + */ + if (spl_scratch->scl_data.taps.v_taps == 5) + spl_scratch->scl_data.taps.v_taps = 4; + + if (spl_scratch->scl_data.taps.v_taps_c == 5) + spl_scratch->scl_data.taps.v_taps_c = 4; + + if (spl_scratch->scl_data.taps.h_taps == 5) + spl_scratch->scl_data.taps.h_taps = 4; + + if (spl_scratch->scl_data.taps.h_taps_c == 5) + spl_scratch->scl_data.taps.h_taps_c = 4; + + if (spl_is_yuv420(spl_in->basic_in.format)) { + if ((spl_scratch->scl_data.taps.h_taps <= 4) || + (spl_scratch->scl_data.taps.h_taps_c <= 3)) { + *enable_easf_v = false; + *enable_easf_h = false; + } else if ((spl_scratch->scl_data.taps.v_taps <= 3) || + (spl_scratch->scl_data.taps.v_taps_c <= 3)) { + *enable_easf_v = false; + *enable_easf_h = true; + } else { + *enable_easf_v = true; + *enable_easf_h = true; + } + SPL_ASSERT((spl_scratch->scl_data.taps.v_taps > 1) && + (spl_scratch->scl_data.taps.v_taps_c > 1)); + } else { /* RGB */ + if (spl_scratch->scl_data.taps.h_taps <= 3) { + *enable_easf_v = false; + *enable_easf_h = false; + } else if (spl_scratch->scl_data.taps.v_taps < 3) { + *enable_easf_v = false; + *enable_easf_h = true; + } else { + *enable_easf_v = true; + *enable_easf_h = true; + } + SPL_ASSERT(spl_scratch->scl_data.taps.v_taps > 1); + } + } else { + *enable_easf_v = false; + *enable_easf_h = false; } // end of if prefer_easf - if (!spl_in->basic_out.always_scale) { - if (IDENTITY_RATIO(spl_out->scl_data.ratios.horz)) - spl_out->scl_data.taps.h_taps = 1; - if (IDENTITY_RATIO(spl_out->scl_data.ratios.vert)) - spl_out->scl_data.taps.v_taps = 1; - if (IDENTITY_RATIO(spl_out->scl_data.ratios.horz_c)) - spl_out->scl_data.taps.h_taps_c = 1; - if (IDENTITY_RATIO(spl_out->scl_data.ratios.vert_c)) - spl_out->scl_data.taps.v_taps_c = 1; + + /* Sharpener requires scaler to be enabled, including for 1:1 + * Check if ISHARP can be enabled + * If ISHARP is not enabled, for 1:1, set taps to 1 and disable + * EASF + * For case of 2:1 YUV where chroma is 1:1, set taps to 1 if + * EASF is not enabled + */ + + *enable_isharp = spl_get_isharp_en(spl_in, spl_scratch); + if (!*enable_isharp && !spl_in->basic_out.always_scale) { + if ((IDENTITY_RATIO(spl_scratch->scl_data.ratios.horz)) && + (IDENTITY_RATIO(spl_scratch->scl_data.ratios.vert))) { + spl_scratch->scl_data.taps.h_taps = 1; + spl_scratch->scl_data.taps.v_taps = 1; + + if (IDENTITY_RATIO(spl_scratch->scl_data.ratios.horz_c)) + spl_scratch->scl_data.taps.h_taps_c = 1; + + if (IDENTITY_RATIO(spl_scratch->scl_data.ratios.vert_c)) + spl_scratch->scl_data.taps.v_taps_c = 1; + + *enable_easf_v = false; + *enable_easf_h = false; + } else { + if ((!*enable_easf_h) && + (IDENTITY_RATIO(spl_scratch->scl_data.ratios.horz_c))) + spl_scratch->scl_data.taps.h_taps_c = 1; + + if ((!*enable_easf_v) && + (IDENTITY_RATIO(spl_scratch->scl_data.ratios.vert_c))) + spl_scratch->scl_data.taps.v_taps_c = 1; + } } return true; } + static void spl_set_black_color_data(enum spl_pixel_format format, struct scl_black_color *scl_black_color) { @@ -848,38 +1073,38 @@ static void spl_set_black_color_data(enum spl_pixel_format format, static void spl_set_manual_ratio_init_data(struct dscl_prog_data *dscl_prog_data, const struct spl_scaler_data *scl_data) { - struct fixed31_32 bot; + struct spl_fixed31_32 bot; - dscl_prog_data->ratios.h_scale_ratio = dc_fixpt_u3d19(scl_data->ratios.horz) << 5; - dscl_prog_data->ratios.v_scale_ratio = dc_fixpt_u3d19(scl_data->ratios.vert) << 5; - dscl_prog_data->ratios.h_scale_ratio_c = dc_fixpt_u3d19(scl_data->ratios.horz_c) << 5; - dscl_prog_data->ratios.v_scale_ratio_c = dc_fixpt_u3d19(scl_data->ratios.vert_c) << 5; + dscl_prog_data->ratios.h_scale_ratio = spl_fixpt_u3d19(scl_data->ratios.horz) << 5; + dscl_prog_data->ratios.v_scale_ratio = spl_fixpt_u3d19(scl_data->ratios.vert) << 5; + dscl_prog_data->ratios.h_scale_ratio_c = spl_fixpt_u3d19(scl_data->ratios.horz_c) << 5; + dscl_prog_data->ratios.v_scale_ratio_c = spl_fixpt_u3d19(scl_data->ratios.vert_c) << 5; /* * 0.24 format for fraction, first five bits zeroed */ dscl_prog_data->init.h_filter_init_frac = - dc_fixpt_u0d19(scl_data->inits.h) << 5; + spl_fixpt_u0d19(scl_data->inits.h) << 5; dscl_prog_data->init.h_filter_init_int = - dc_fixpt_floor(scl_data->inits.h); + spl_fixpt_floor(scl_data->inits.h); dscl_prog_data->init.h_filter_init_frac_c = - dc_fixpt_u0d19(scl_data->inits.h_c) << 5; + spl_fixpt_u0d19(scl_data->inits.h_c) << 5; dscl_prog_data->init.h_filter_init_int_c = - dc_fixpt_floor(scl_data->inits.h_c); + spl_fixpt_floor(scl_data->inits.h_c); dscl_prog_data->init.v_filter_init_frac = - dc_fixpt_u0d19(scl_data->inits.v) << 5; + spl_fixpt_u0d19(scl_data->inits.v) << 5; dscl_prog_data->init.v_filter_init_int = - dc_fixpt_floor(scl_data->inits.v); + spl_fixpt_floor(scl_data->inits.v); dscl_prog_data->init.v_filter_init_frac_c = - dc_fixpt_u0d19(scl_data->inits.v_c) << 5; + spl_fixpt_u0d19(scl_data->inits.v_c) << 5; dscl_prog_data->init.v_filter_init_int_c = - dc_fixpt_floor(scl_data->inits.v_c); - - bot = dc_fixpt_add(scl_data->inits.v, scl_data->ratios.vert); - dscl_prog_data->init.v_filter_init_bot_frac = dc_fixpt_u0d19(bot) << 5; - dscl_prog_data->init.v_filter_init_bot_int = dc_fixpt_floor(bot); - bot = dc_fixpt_add(scl_data->inits.v_c, scl_data->ratios.vert_c); - dscl_prog_data->init.v_filter_init_bot_frac_c = dc_fixpt_u0d19(bot) << 5; - dscl_prog_data->init.v_filter_init_bot_int_c = dc_fixpt_floor(bot); + spl_fixpt_floor(scl_data->inits.v_c); + + bot = spl_fixpt_add(scl_data->inits.v, scl_data->ratios.vert); + dscl_prog_data->init.v_filter_init_bot_frac = spl_fixpt_u0d19(bot) << 5; + dscl_prog_data->init.v_filter_init_bot_int = spl_fixpt_floor(bot); + bot = spl_fixpt_add(scl_data->inits.v_c, scl_data->ratios.vert_c); + dscl_prog_data->init.v_filter_init_bot_frac_c = spl_fixpt_u0d19(bot) << 5; + dscl_prog_data->init.v_filter_init_bot_int_c = spl_fixpt_floor(bot); } static void spl_set_taps_data(struct dscl_prog_data *dscl_prog_data, @@ -890,79 +1115,28 @@ static void spl_set_taps_data(struct dscl_prog_data *dscl_prog_data, dscl_prog_data->taps.v_taps_c = scl_data->taps.v_taps_c - 1; dscl_prog_data->taps.h_taps_c = scl_data->taps.h_taps_c - 1; } -static const uint16_t *spl_dscl_get_filter_coeffs_64p(int taps, struct fixed31_32 ratio) -{ - if (taps == 8) - return spl_get_filter_8tap_64p(ratio); - else if (taps == 7) - return spl_get_filter_7tap_64p(ratio); - else if (taps == 6) - return spl_get_filter_6tap_64p(ratio); - else if (taps == 5) - return spl_get_filter_5tap_64p(ratio); - else if (taps == 4) - return spl_get_filter_4tap_64p(ratio); - else if (taps == 3) - return spl_get_filter_3tap_64p(ratio); - else if (taps == 2) - return spl_get_filter_2tap_64p(); - else if (taps == 1) - return NULL; - else { - /* should never happen, bug */ - BREAK_TO_DEBUGGER(); - return NULL; - } -} -static void spl_set_filters_data(struct dscl_prog_data *dscl_prog_data, - const struct spl_scaler_data *data) -{ - dscl_prog_data->filter_h = spl_dscl_get_filter_coeffs_64p( - data->taps.h_taps, data->ratios.horz); - dscl_prog_data->filter_v = spl_dscl_get_filter_coeffs_64p( - data->taps.v_taps, data->ratios.vert); - dscl_prog_data->filter_h_c = spl_dscl_get_filter_coeffs_64p( - data->taps.h_taps_c, data->ratios.horz_c); - dscl_prog_data->filter_v_c = spl_dscl_get_filter_coeffs_64p( - data->taps.v_taps_c, data->ratios.vert_c); -} -#ifdef CONFIG_DRM_AMD_DC_FP -static const uint16_t *spl_dscl_get_blur_scale_coeffs_64p(int taps) -{ - if ((taps == 3) || (taps == 4) || (taps == 6)) - return spl_get_filter_isharp_bs_4tap_64p(); - else { - /* should never happen, bug */ - BREAK_TO_DEBUGGER(); - return NULL; - } -} -static void spl_set_blur_scale_data(struct dscl_prog_data *dscl_prog_data, - const struct spl_scaler_data *data) -{ - dscl_prog_data->filter_blur_scale_h = spl_dscl_get_blur_scale_coeffs_64p( - data->taps.h_taps); - dscl_prog_data->filter_blur_scale_v = spl_dscl_get_blur_scale_coeffs_64p( - data->taps.v_taps); -} -#endif + /* Populate dscl prog data structure from scaler data calculated by SPL */ -static void spl_set_dscl_prog_data(struct spl_in *spl_in, struct spl_out *spl_out) +static void spl_set_dscl_prog_data(struct spl_in *spl_in, struct spl_scratch *spl_scratch, + struct spl_out *spl_out, bool enable_easf_v, bool enable_easf_h, bool enable_isharp) { struct dscl_prog_data *dscl_prog_data = spl_out->dscl_prog_data; - const struct spl_scaler_data *data = &spl_out->scl_data; + const struct spl_scaler_data *data = &spl_scratch->scl_data; struct scl_black_color *scl_black_color = &dscl_prog_data->scl_black_color; + bool enable_easf = enable_easf_v || enable_easf_h; + // Set values for recout - dscl_prog_data->recout = spl_out->scl_data.recout; + dscl_prog_data->recout = spl_scratch->scl_data.recout; // Set values for MPC Size - dscl_prog_data->mpc_size.width = spl_out->scl_data.h_active; - dscl_prog_data->mpc_size.height = spl_out->scl_data.v_active; + dscl_prog_data->mpc_size.width = spl_scratch->scl_data.h_active; + dscl_prog_data->mpc_size.height = spl_scratch->scl_data.v_active; // SCL_MODE - Set SCL_MODE data - dscl_prog_data->dscl_mode = spl_get_dscl_mode(spl_in, data); + dscl_prog_data->dscl_mode = spl_get_dscl_mode(spl_in, data, enable_isharp, + enable_easf); // SCL_BLACK_COLOR spl_set_black_color_data(spl_in->basic_in.format, scl_black_color); @@ -973,103 +1147,135 @@ static void spl_set_dscl_prog_data(struct spl_in *spl_in, struct spl_out *spl_ou // Set HTaps/VTaps spl_set_taps_data(dscl_prog_data, data); // Set viewport - dscl_prog_data->viewport = spl_out->scl_data.viewport; + dscl_prog_data->viewport = spl_scratch->scl_data.viewport; // Set viewport_c - dscl_prog_data->viewport_c = spl_out->scl_data.viewport_c; + dscl_prog_data->viewport_c = spl_scratch->scl_data.viewport_c; // Set filters data - spl_set_filters_data(dscl_prog_data, data); + spl_set_filters_data(dscl_prog_data, data, enable_easf_v, enable_easf_h); } -/* Enable EASF ?*/ -static bool enable_easf(int scale_ratio, int taps, - enum linear_light_scaling lls_pref, bool prefer_easf) + +/* Calculate C0-C3 coefficients based on HDR_mult */ +static void spl_calculate_c0_c3_hdr(struct dscl_prog_data *dscl_prog_data, uint32_t hdr_multx100) { - // Is downscaling > 6:1 ? - if (scale_ratio > 6) { - // END - No EASF support for downscaling > 6:1 - return false; - } - // Is upscaling or downscaling up to 2:1? - if (scale_ratio <= 2) { - // Is linear scaling or EASF preferred? - if (lls_pref == LLS_PREF_YES || prefer_easf) { - // LB support taps 3, 4, 6 - if (taps == 3 || taps == 4 || taps == 6) { - // END - EASF supported - return true; - } - } - } - // END - EASF not supported - return false; + struct spl_fixed31_32 hdr_mult, c0_mult, c1_mult, c2_mult; + struct spl_fixed31_32 c0_calc, c1_calc, c2_calc; + struct spl_custom_float_format fmt; + + SPL_ASSERT(hdr_multx100); + hdr_mult = spl_fixpt_from_fraction((long long)hdr_multx100, 100LL); + c0_mult = spl_fixpt_from_fraction(2126LL, 10000LL); + c1_mult = spl_fixpt_from_fraction(7152LL, 10000LL); + c2_mult = spl_fixpt_from_fraction(722LL, 10000LL); + + c0_calc = spl_fixpt_mul(hdr_mult, spl_fixpt_mul(c0_mult, spl_fixpt_from_fraction( + 16384LL, 125LL))); + c1_calc = spl_fixpt_mul(hdr_mult, spl_fixpt_mul(c1_mult, spl_fixpt_from_fraction( + 16384LL, 125LL))); + c2_calc = spl_fixpt_mul(hdr_mult, spl_fixpt_mul(c2_mult, spl_fixpt_from_fraction( + 16384LL, 125LL))); + + fmt.exponenta_bits = 5; + fmt.mantissa_bits = 10; + fmt.sign = true; + + // fp1.5.10, C0 coefficient (LN_rec709: HDR_MULT * 0.212600 * 2^14/125) + spl_convert_to_custom_float_format(c0_calc, &fmt, &dscl_prog_data->easf_matrix_c0); + // fp1.5.10, C1 coefficient (LN_rec709: HDR_MULT * 0.715200 * 2^14/125) + spl_convert_to_custom_float_format(c1_calc, &fmt, &dscl_prog_data->easf_matrix_c1); + // fp1.5.10, C2 coefficient (LN_rec709: HDR_MULT * 0.072200 * 2^14/125) + spl_convert_to_custom_float_format(c2_calc, &fmt, &dscl_prog_data->easf_matrix_c2); + dscl_prog_data->easf_matrix_c3 = 0x0; // fp1.5.10, C3 coefficient } + /* Set EASF data */ -static void spl_set_easf_data(struct dscl_prog_data *dscl_prog_data, - bool enable_easf_v, bool enable_easf_h, enum linear_light_scaling lls_pref, - enum spl_pixel_format format) +static void spl_set_easf_data(struct spl_scratch *spl_scratch, struct spl_out *spl_out, bool enable_easf_v, + bool enable_easf_h, enum linear_light_scaling lls_pref, + enum spl_pixel_format format, enum system_setup setup, + uint32_t hdr_multx100) { - if (spl_is_yuv420(format)) /* TODO: 0 = RGB, 1 = YUV */ - dscl_prog_data->easf_matrix_mode = 1; - else - dscl_prog_data->easf_matrix_mode = 0; - + struct dscl_prog_data *dscl_prog_data = spl_out->dscl_prog_data; if (enable_easf_v) { dscl_prog_data->easf_v_en = true; dscl_prog_data->easf_v_ring = 0; - dscl_prog_data->easf_v_sharp_factor = 1; + dscl_prog_data->easf_v_sharp_factor = 0; dscl_prog_data->easf_v_bf1_en = 1; // 1-bit, BF1 calculation enable, 0=disable, 1=enable dscl_prog_data->easf_v_bf2_mode = 0xF; // 4-bit, BF2 calculation mode - dscl_prog_data->easf_v_bf3_mode = 2; // 2-bit, BF3 chroma mode correction calculation mode - dscl_prog_data->easf_v_bf2_flat1_gain = 4; // U1.3, BF2 Flat1 Gain control - dscl_prog_data->easf_v_bf2_flat2_gain = 8; // U4.0, BF2 Flat2 Gain control - dscl_prog_data->easf_v_bf2_roc_gain = 4; // U2.2, Rate Of Change control + /* 2-bit, BF3 chroma mode correction calculation mode */ + dscl_prog_data->easf_v_bf3_mode = spl_get_v_bf3_mode( + spl_scratch->scl_data.recip_ratios.vert); + /* FP1.5.10 [ minCoef ]*/ dscl_prog_data->easf_v_ringest_3tap_dntilt_uptilt = - 0x9F00;// FP1.5.10 [minCoef] (-0.036109167214271) + spl_get_3tap_dntilt_uptilt_offset(spl_scratch->scl_data.taps.v_taps, + spl_scratch->scl_data.recip_ratios.vert); + /* FP1.5.10 [ upTiltMaxVal ]*/ dscl_prog_data->easf_v_ringest_3tap_uptilt_max = - 0x24FE; // FP1.5.10 [upTiltMaxVal] ( 0.904556445553545) + spl_get_3tap_uptilt_maxval(spl_scratch->scl_data.taps.v_taps, + spl_scratch->scl_data.recip_ratios.vert); + /* FP1.5.10 [ dnTiltSlope ]*/ dscl_prog_data->easf_v_ringest_3tap_dntilt_slope = - 0x3940; // FP1.5.10 [dnTiltSlope] ( 0.910488988173371) + spl_get_3tap_dntilt_slope(spl_scratch->scl_data.taps.v_taps, + spl_scratch->scl_data.recip_ratios.vert); + /* FP1.5.10 [ upTilt1Slope ]*/ dscl_prog_data->easf_v_ringest_3tap_uptilt1_slope = - 0x359C; // FP1.5.10 [upTilt1Slope] ( 0.125620179040899) + spl_get_3tap_uptilt1_slope(spl_scratch->scl_data.taps.v_taps, + spl_scratch->scl_data.recip_ratios.vert); + /* FP1.5.10 [ upTilt2Slope ]*/ dscl_prog_data->easf_v_ringest_3tap_uptilt2_slope = - 0x359C; // FP1.5.10 [upTilt2Slope] ( 0.006786817723568) + spl_get_3tap_uptilt2_slope(spl_scratch->scl_data.taps.v_taps, + spl_scratch->scl_data.recip_ratios.vert); + /* FP1.5.10 [ upTilt2Offset ]*/ dscl_prog_data->easf_v_ringest_3tap_uptilt2_offset = - 0x9F00; // FP1.5.10 [upTilt2Offset] (-0.006139059716651) + spl_get_3tap_uptilt2_offset(spl_scratch->scl_data.taps.v_taps, + spl_scratch->scl_data.recip_ratios.vert); + /* FP1.5.10; (2.0) Ring reducer gain for 4 or 6-tap mode [H_REDUCER_GAIN4] */ dscl_prog_data->easf_v_ringest_eventap_reduceg1 = - 0x4000; // FP1.5.10; (2.0) Ring reducer gain for 4 or 6-tap mode [H_REDUCER_GAIN4] + spl_get_reducer_gain4(spl_scratch->scl_data.taps.v_taps, + spl_scratch->scl_data.recip_ratios.vert); + /* FP1.5.10; (2.5) Ring reducer gain for 6-tap mode [V_REDUCER_GAIN6] */ dscl_prog_data->easf_v_ringest_eventap_reduceg2 = - 0x4100; // FP1.5.10; (2.5) Ring reducer gain for 6-tap mode [V_REDUCER_GAIN6] + spl_get_reducer_gain6(spl_scratch->scl_data.taps.v_taps, + spl_scratch->scl_data.recip_ratios.vert); + /* FP1.5.10; (-0.135742) Ring gain for 6-tap set to -139/1024 */ dscl_prog_data->easf_v_ringest_eventap_gain1 = - 0xB058; // FP1.5.10; (-0.135742) Ring gain for 6-tap set to -139/1024 + spl_get_gainRing4(spl_scratch->scl_data.taps.v_taps, + spl_scratch->scl_data.recip_ratios.vert); + /* FP1.5.10; (-0.024414) Ring gain for 6-tap set to -25/1024 */ dscl_prog_data->easf_v_ringest_eventap_gain2 = - 0xA640; // FP1.5.10; (-0.024414) Ring gain for 6-tap set to -25/1024 + spl_get_gainRing6(spl_scratch->scl_data.taps.v_taps, + spl_scratch->scl_data.recip_ratios.vert); dscl_prog_data->easf_v_bf_maxa = 63; //Vertical Max BF value A in U0.6 format.Selected if V_FCNTL == 0 dscl_prog_data->easf_v_bf_maxb = 63; //Vertical Max BF value A in U0.6 format.Selected if V_FCNTL == 1 dscl_prog_data->easf_v_bf_mina = 0; //Vertical Min BF value A in U0.6 format.Selected if V_FCNTL == 0 dscl_prog_data->easf_v_bf_minb = 0; //Vertical Min BF value A in U0.6 format.Selected if V_FCNTL == 1 - dscl_prog_data->easf_v_bf1_pwl_in_seg0 = -512; // S0.10, BF1 PWL Segment 0 - dscl_prog_data->easf_v_bf1_pwl_base_seg0 = 0; // U0.6, BF1 Base PWL Segment 0 - dscl_prog_data->easf_v_bf1_pwl_slope_seg0 = 3; // S7.3, BF1 Slope PWL Segment 0 - dscl_prog_data->easf_v_bf1_pwl_in_seg1 = -20; // S0.10, BF1 PWL Segment 1 - dscl_prog_data->easf_v_bf1_pwl_base_seg1 = 12; // U0.6, BF1 Base PWL Segment 1 - dscl_prog_data->easf_v_bf1_pwl_slope_seg1 = 326; // S7.3, BF1 Slope PWL Segment 1 - dscl_prog_data->easf_v_bf1_pwl_in_seg2 = 0; // S0.10, BF1 PWL Segment 2 - dscl_prog_data->easf_v_bf1_pwl_base_seg2 = 63; // U0.6, BF1 Base PWL Segment 2 - dscl_prog_data->easf_v_bf1_pwl_slope_seg2 = 0; // S7.3, BF1 Slope PWL Segment 2 - dscl_prog_data->easf_v_bf1_pwl_in_seg3 = 16; // S0.10, BF1 PWL Segment 3 - dscl_prog_data->easf_v_bf1_pwl_base_seg3 = 63; // U0.6, BF1 Base PWL Segment 3 - dscl_prog_data->easf_v_bf1_pwl_slope_seg3 = -56; // S7.3, BF1 Slope PWL Segment 3 - dscl_prog_data->easf_v_bf1_pwl_in_seg4 = 32; // S0.10, BF1 PWL Segment 4 - dscl_prog_data->easf_v_bf1_pwl_base_seg4 = 56; // U0.6, BF1 Base PWL Segment 4 - dscl_prog_data->easf_v_bf1_pwl_slope_seg4 = -48; // S7.3, BF1 Slope PWL Segment 4 - dscl_prog_data->easf_v_bf1_pwl_in_seg5 = 48; // S0.10, BF1 PWL Segment 5 - dscl_prog_data->easf_v_bf1_pwl_base_seg5 = 50; // U0.6, BF1 Base PWL Segment 5 - dscl_prog_data->easf_v_bf1_pwl_slope_seg5 = -240; // S7.3, BF1 Slope PWL Segment 5 - dscl_prog_data->easf_v_bf1_pwl_in_seg6 = 64; // S0.10, BF1 PWL Segment 6 - dscl_prog_data->easf_v_bf1_pwl_base_seg6 = 20; // U0.6, BF1 Base PWL Segment 6 - dscl_prog_data->easf_v_bf1_pwl_slope_seg6 = -160; // S7.3, BF1 Slope PWL Segment 6 - dscl_prog_data->easf_v_bf1_pwl_in_seg7 = 80; // S0.10, BF1 PWL Segment 7 - dscl_prog_data->easf_v_bf1_pwl_base_seg7 = 0; // U0.6, BF1 Base PWL Segment 7 if (lls_pref == LLS_PREF_YES) { + dscl_prog_data->easf_v_bf2_flat1_gain = 4; // U1.3, BF2 Flat1 Gain control + dscl_prog_data->easf_v_bf2_flat2_gain = 8; // U4.0, BF2 Flat2 Gain control + dscl_prog_data->easf_v_bf2_roc_gain = 4; // U2.2, Rate Of Change control + + dscl_prog_data->easf_v_bf1_pwl_in_seg0 = 0x600; // S0.10, BF1 PWL Segment 0 = -512 + dscl_prog_data->easf_v_bf1_pwl_base_seg0 = 0; // U0.6, BF1 Base PWL Segment 0 + dscl_prog_data->easf_v_bf1_pwl_slope_seg0 = 3; // S7.3, BF1 Slope PWL Segment 0 + dscl_prog_data->easf_v_bf1_pwl_in_seg1 = 0x7EC; // S0.10, BF1 PWL Segment 1 = -20 + dscl_prog_data->easf_v_bf1_pwl_base_seg1 = 12; // U0.6, BF1 Base PWL Segment 1 + dscl_prog_data->easf_v_bf1_pwl_slope_seg1 = 326; // S7.3, BF1 Slope PWL Segment 1 + dscl_prog_data->easf_v_bf1_pwl_in_seg2 = 0; // S0.10, BF1 PWL Segment 2 + dscl_prog_data->easf_v_bf1_pwl_base_seg2 = 63; // U0.6, BF1 Base PWL Segment 2 + dscl_prog_data->easf_v_bf1_pwl_slope_seg2 = 0; // S7.3, BF1 Slope PWL Segment 2 + dscl_prog_data->easf_v_bf1_pwl_in_seg3 = 16; // S0.10, BF1 PWL Segment 3 + dscl_prog_data->easf_v_bf1_pwl_base_seg3 = 63; // U0.6, BF1 Base PWL Segment 3 + dscl_prog_data->easf_v_bf1_pwl_slope_seg3 = 0x7C8; // S7.3, BF1 Slope PWL Segment 3 = -56 + dscl_prog_data->easf_v_bf1_pwl_in_seg4 = 32; // S0.10, BF1 PWL Segment 4 + dscl_prog_data->easf_v_bf1_pwl_base_seg4 = 56; // U0.6, BF1 Base PWL Segment 4 + dscl_prog_data->easf_v_bf1_pwl_slope_seg4 = 0x7D0; // S7.3, BF1 Slope PWL Segment 4 = -48 + dscl_prog_data->easf_v_bf1_pwl_in_seg5 = 48; // S0.10, BF1 PWL Segment 5 + dscl_prog_data->easf_v_bf1_pwl_base_seg5 = 50; // U0.6, BF1 Base PWL Segment 5 + dscl_prog_data->easf_v_bf1_pwl_slope_seg5 = 0x710; // S7.3, BF1 Slope PWL Segment 5 = -240 + dscl_prog_data->easf_v_bf1_pwl_in_seg6 = 64; // S0.10, BF1 PWL Segment 6 + dscl_prog_data->easf_v_bf1_pwl_base_seg6 = 20; // U0.6, BF1 Base PWL Segment 6 + dscl_prog_data->easf_v_bf1_pwl_slope_seg6 = 0x760; // S7.3, BF1 Slope PWL Segment 6 = -160 + dscl_prog_data->easf_v_bf1_pwl_in_seg7 = 80; // S0.10, BF1 PWL Segment 7 + dscl_prog_data->easf_v_bf1_pwl_base_seg7 = 0; // U0.6, BF1 Base PWL Segment 7 + dscl_prog_data->easf_v_bf3_pwl_in_set0 = 0x000; // FP0.6.6, BF3 Input value PWL Segment 0 dscl_prog_data->easf_v_bf3_pwl_base_set0 = 63; // S0.6, BF3 Base PWL Segment 0 dscl_prog_data->easf_v_bf3_pwl_slope_set0 = 0x12C5; // FP1.6.6, BF3 Slope PWL Segment 0 @@ -1090,13 +1296,41 @@ static void spl_set_easf_data(struct dscl_prog_data *dscl_prog_data, 0x136B; // FP1.6.6, BF3 Slope PWL Segment 3 dscl_prog_data->easf_v_bf3_pwl_in_set4 = 0x0C37; // FP0.6.6, BF3 Input value PWL Segment 4 (0.125 * 125^3) - dscl_prog_data->easf_v_bf3_pwl_base_set4 = -50; // S0.6, BF3 Base PWL Segment 4 + dscl_prog_data->easf_v_bf3_pwl_base_set4 = 0x4E; // S0.6, BF3 Base PWL Segment 4 = -50 dscl_prog_data->easf_v_bf3_pwl_slope_set4 = 0x1200; // FP1.6.6, BF3 Slope PWL Segment 4 dscl_prog_data->easf_v_bf3_pwl_in_set5 = 0x0CF7; // FP0.6.6, BF3 Input value PWL Segment 5 (1.0 * 125^3) - dscl_prog_data->easf_v_bf3_pwl_base_set5 = -63; // S0.6, BF3 Base PWL Segment 5 + dscl_prog_data->easf_v_bf3_pwl_base_set5 = 0x41; // S0.6, BF3 Base PWL Segment 5 = -63 } else { + dscl_prog_data->easf_v_bf2_flat1_gain = 13; // U1.3, BF2 Flat1 Gain control + dscl_prog_data->easf_v_bf2_flat2_gain = 15; // U4.0, BF2 Flat2 Gain control + dscl_prog_data->easf_v_bf2_roc_gain = 14; // U2.2, Rate Of Change control + + dscl_prog_data->easf_v_bf1_pwl_in_seg0 = 0x440; // S0.10, BF1 PWL Segment 0 = -960 + dscl_prog_data->easf_v_bf1_pwl_base_seg0 = 0; // U0.6, BF1 Base PWL Segment 0 + dscl_prog_data->easf_v_bf1_pwl_slope_seg0 = 2; // S7.3, BF1 Slope PWL Segment 0 + dscl_prog_data->easf_v_bf1_pwl_in_seg1 = 0x7C4; // S0.10, BF1 PWL Segment 1 = -60 + dscl_prog_data->easf_v_bf1_pwl_base_seg1 = 12; // U0.6, BF1 Base PWL Segment 1 + dscl_prog_data->easf_v_bf1_pwl_slope_seg1 = 109; // S7.3, BF1 Slope PWL Segment 1 + dscl_prog_data->easf_v_bf1_pwl_in_seg2 = 0; // S0.10, BF1 PWL Segment 2 + dscl_prog_data->easf_v_bf1_pwl_base_seg2 = 63; // U0.6, BF1 Base PWL Segment 2 + dscl_prog_data->easf_v_bf1_pwl_slope_seg2 = 0; // S7.3, BF1 Slope PWL Segment 2 + dscl_prog_data->easf_v_bf1_pwl_in_seg3 = 48; // S0.10, BF1 PWL Segment 3 + dscl_prog_data->easf_v_bf1_pwl_base_seg3 = 63; // U0.6, BF1 Base PWL Segment 3 + dscl_prog_data->easf_v_bf1_pwl_slope_seg3 = 0x7ED; // S7.3, BF1 Slope PWL Segment 3 = -19 + dscl_prog_data->easf_v_bf1_pwl_in_seg4 = 96; // S0.10, BF1 PWL Segment 4 + dscl_prog_data->easf_v_bf1_pwl_base_seg4 = 56; // U0.6, BF1 Base PWL Segment 4 + dscl_prog_data->easf_v_bf1_pwl_slope_seg4 = 0x7F0; // S7.3, BF1 Slope PWL Segment 4 = -16 + dscl_prog_data->easf_v_bf1_pwl_in_seg5 = 144; // S0.10, BF1 PWL Segment 5 + dscl_prog_data->easf_v_bf1_pwl_base_seg5 = 50; // U0.6, BF1 Base PWL Segment 5 + dscl_prog_data->easf_v_bf1_pwl_slope_seg5 = 0x7B0; // S7.3, BF1 Slope PWL Segment 5 = -80 + dscl_prog_data->easf_v_bf1_pwl_in_seg6 = 192; // S0.10, BF1 PWL Segment 6 + dscl_prog_data->easf_v_bf1_pwl_base_seg6 = 20; // U0.6, BF1 Base PWL Segment 6 + dscl_prog_data->easf_v_bf1_pwl_slope_seg6 = 0x7CB; // S7.3, BF1 Slope PWL Segment 6 = -53 + dscl_prog_data->easf_v_bf1_pwl_in_seg7 = 240; // S0.10, BF1 PWL Segment 7 + dscl_prog_data->easf_v_bf1_pwl_base_seg7 = 0; // U0.6, BF1 Base PWL Segment 7 + dscl_prog_data->easf_v_bf3_pwl_in_set0 = 0x000; // FP0.6.6, BF3 Input value PWL Segment 0 dscl_prog_data->easf_v_bf3_pwl_base_set0 = 63; // S0.6, BF3 Base PWL Segment 0 dscl_prog_data->easf_v_bf3_pwl_slope_set0 = 0x0000; // FP1.6.6, BF3 Slope PWL Segment 0 @@ -1115,11 +1349,11 @@ static void spl_set_easf_data(struct dscl_prog_data *dscl_prog_data, 0x1878; // FP1.6.6, BF3 Slope PWL Segment 3 dscl_prog_data->easf_v_bf3_pwl_in_set4 = 0x0761; // FP0.6.6, BF3 Input value PWL Segment 4 (0.375) - dscl_prog_data->easf_v_bf3_pwl_base_set4 = -60; // S0.6, BF3 Base PWL Segment 4 + dscl_prog_data->easf_v_bf3_pwl_base_set4 = 0x44; // S0.6, BF3 Base PWL Segment 4 = -60 dscl_prog_data->easf_v_bf3_pwl_slope_set4 = 0x1760; // FP1.6.6, BF3 Slope PWL Segment 4 dscl_prog_data->easf_v_bf3_pwl_in_set5 = 0x0780; // FP0.6.6, BF3 Input value PWL Segment 5 (0.5) - dscl_prog_data->easf_v_bf3_pwl_base_set5 = -63; // S0.6, BF3 Base PWL Segment 5 + dscl_prog_data->easf_v_bf3_pwl_base_set5 = 0x41; // S0.6, BF3 Base PWL Segment 5 = -63 } } else dscl_prog_data->easf_v_en = false; @@ -1127,52 +1361,63 @@ static void spl_set_easf_data(struct dscl_prog_data *dscl_prog_data, if (enable_easf_h) { dscl_prog_data->easf_h_en = true; dscl_prog_data->easf_h_ring = 0; - dscl_prog_data->easf_h_sharp_factor = 1; + dscl_prog_data->easf_h_sharp_factor = 0; dscl_prog_data->easf_h_bf1_en = 1; // 1-bit, BF1 calculation enable, 0=disable, 1=enable dscl_prog_data->easf_h_bf2_mode = 0xF; // 4-bit, BF2 calculation mode - dscl_prog_data->easf_h_bf3_mode = - 2; // 2-bit, BF3 chroma mode correction calculation mode - dscl_prog_data->easf_h_bf2_flat1_gain = 4; // U1.3, BF2 Flat1 Gain control - dscl_prog_data->easf_h_bf2_flat2_gain = 8; // U4.0, BF2 Flat2 Gain control - dscl_prog_data->easf_h_bf2_roc_gain = 4; // U2.2, Rate Of Change control + /* 2-bit, BF3 chroma mode correction calculation mode */ + dscl_prog_data->easf_h_bf3_mode = spl_get_h_bf3_mode( + spl_scratch->scl_data.recip_ratios.horz); + /* FP1.5.10; (2.0) Ring reducer gain for 4 or 6-tap mode [H_REDUCER_GAIN4] */ dscl_prog_data->easf_h_ringest_eventap_reduceg1 = - 0x4000; // FP1.5.10; (2.0) Ring reducer gain for 4 or 6-tap mode [H_REDUCER_GAIN4] + spl_get_reducer_gain4(spl_scratch->scl_data.taps.h_taps, + spl_scratch->scl_data.recip_ratios.horz); + /* FP1.5.10; (2.5) Ring reducer gain for 6-tap mode [V_REDUCER_GAIN6] */ dscl_prog_data->easf_h_ringest_eventap_reduceg2 = - 0x4100; // FP1.5.10; (2.5) Ring reducer gain for 6-tap mode [V_REDUCER_GAIN6] + spl_get_reducer_gain6(spl_scratch->scl_data.taps.h_taps, + spl_scratch->scl_data.recip_ratios.horz); + /* FP1.5.10; (-0.135742) Ring gain for 6-tap set to -139/1024 */ dscl_prog_data->easf_h_ringest_eventap_gain1 = - 0xB058; // FP1.5.10; (-0.135742) Ring gain for 6-tap set to -139/1024 + spl_get_gainRing4(spl_scratch->scl_data.taps.h_taps, + spl_scratch->scl_data.recip_ratios.horz); + /* FP1.5.10; (-0.024414) Ring gain for 6-tap set to -25/1024 */ dscl_prog_data->easf_h_ringest_eventap_gain2 = - 0xA640; // FP1.5.10; (-0.024414) Ring gain for 6-tap set to -25/1024 + spl_get_gainRing6(spl_scratch->scl_data.taps.h_taps, + spl_scratch->scl_data.recip_ratios.horz); dscl_prog_data->easf_h_bf_maxa = 63; //Horz Max BF value A in U0.6 format.Selected if H_FCNTL==0 dscl_prog_data->easf_h_bf_maxb = 63; //Horz Max BF value B in U0.6 format.Selected if H_FCNTL==1 dscl_prog_data->easf_h_bf_mina = 0; //Horz Min BF value B in U0.6 format.Selected if H_FCNTL==0 dscl_prog_data->easf_h_bf_minb = 0; //Horz Min BF value B in U0.6 format.Selected if H_FCNTL==1 - dscl_prog_data->easf_h_bf1_pwl_in_seg0 = -512; // S0.10, BF1 PWL Segment 0 - dscl_prog_data->easf_h_bf1_pwl_base_seg0 = 0; // U0.6, BF1 Base PWL Segment 0 - dscl_prog_data->easf_h_bf1_pwl_slope_seg0 = 3; // S7.3, BF1 Slope PWL Segment 0 - dscl_prog_data->easf_h_bf1_pwl_in_seg1 = -20; // S0.10, BF1 PWL Segment 1 - dscl_prog_data->easf_h_bf1_pwl_base_seg1 = 12; // U0.6, BF1 Base PWL Segment 1 - dscl_prog_data->easf_h_bf1_pwl_slope_seg1 = 326; // S7.3, BF1 Slope PWL Segment 1 - dscl_prog_data->easf_h_bf1_pwl_in_seg2 = 0; // S0.10, BF1 PWL Segment 2 - dscl_prog_data->easf_h_bf1_pwl_base_seg2 = 63; // U0.6, BF1 Base PWL Segment 2 - dscl_prog_data->easf_h_bf1_pwl_slope_seg2 = 0; // S7.3, BF1 Slope PWL Segment 2 - dscl_prog_data->easf_h_bf1_pwl_in_seg3 = 16; // S0.10, BF1 PWL Segment 3 - dscl_prog_data->easf_h_bf1_pwl_base_seg3 = 63; // U0.6, BF1 Base PWL Segment 3 - dscl_prog_data->easf_h_bf1_pwl_slope_seg3 = -56; // S7.3, BF1 Slope PWL Segment 3 - dscl_prog_data->easf_h_bf1_pwl_in_seg4 = 32; // S0.10, BF1 PWL Segment 4 - dscl_prog_data->easf_h_bf1_pwl_base_seg4 = 56; // U0.6, BF1 Base PWL Segment 4 - dscl_prog_data->easf_h_bf1_pwl_slope_seg4 = -48; // S7.3, BF1 Slope PWL Segment 4 - dscl_prog_data->easf_h_bf1_pwl_in_seg5 = 48; // S0.10, BF1 PWL Segment 5 - dscl_prog_data->easf_h_bf1_pwl_base_seg5 = 50; // U0.6, BF1 Base PWL Segment 5 - dscl_prog_data->easf_h_bf1_pwl_slope_seg5 = -240; // S7.3, BF1 Slope PWL Segment 5 - dscl_prog_data->easf_h_bf1_pwl_in_seg6 = 64; // S0.10, BF1 PWL Segment 6 - dscl_prog_data->easf_h_bf1_pwl_base_seg6 = 20; // U0.6, BF1 Base PWL Segment 6 - dscl_prog_data->easf_h_bf1_pwl_slope_seg6 = -160; // S7.3, BF1 Slope PWL Segment 6 - dscl_prog_data->easf_h_bf1_pwl_in_seg7 = 80; // S0.10, BF1 PWL Segment 7 - dscl_prog_data->easf_h_bf1_pwl_base_seg7 = 0; // U0.6, BF1 Base PWL Segment 7 if (lls_pref == LLS_PREF_YES) { + dscl_prog_data->easf_h_bf2_flat1_gain = 4; // U1.3, BF2 Flat1 Gain control + dscl_prog_data->easf_h_bf2_flat2_gain = 8; // U4.0, BF2 Flat2 Gain control + dscl_prog_data->easf_h_bf2_roc_gain = 4; // U2.2, Rate Of Change control + + dscl_prog_data->easf_h_bf1_pwl_in_seg0 = 0x600; // S0.10, BF1 PWL Segment 0 = -512 + dscl_prog_data->easf_h_bf1_pwl_base_seg0 = 0; // U0.6, BF1 Base PWL Segment 0 + dscl_prog_data->easf_h_bf1_pwl_slope_seg0 = 3; // S7.3, BF1 Slope PWL Segment 0 + dscl_prog_data->easf_h_bf1_pwl_in_seg1 = 0x7EC; // S0.10, BF1 PWL Segment 1 = -20 + dscl_prog_data->easf_h_bf1_pwl_base_seg1 = 12; // U0.6, BF1 Base PWL Segment 1 + dscl_prog_data->easf_h_bf1_pwl_slope_seg1 = 326; // S7.3, BF1 Slope PWL Segment 1 + dscl_prog_data->easf_h_bf1_pwl_in_seg2 = 0; // S0.10, BF1 PWL Segment 2 + dscl_prog_data->easf_h_bf1_pwl_base_seg2 = 63; // U0.6, BF1 Base PWL Segment 2 + dscl_prog_data->easf_h_bf1_pwl_slope_seg2 = 0; // S7.3, BF1 Slope PWL Segment 2 + dscl_prog_data->easf_h_bf1_pwl_in_seg3 = 16; // S0.10, BF1 PWL Segment 3 + dscl_prog_data->easf_h_bf1_pwl_base_seg3 = 63; // U0.6, BF1 Base PWL Segment 3 + dscl_prog_data->easf_h_bf1_pwl_slope_seg3 = 0x7C8; // S7.3, BF1 Slope PWL Segment 3 = -56 + dscl_prog_data->easf_h_bf1_pwl_in_seg4 = 32; // S0.10, BF1 PWL Segment 4 + dscl_prog_data->easf_h_bf1_pwl_base_seg4 = 56; // U0.6, BF1 Base PWL Segment 4 + dscl_prog_data->easf_h_bf1_pwl_slope_seg4 = 0x7D0; // S7.3, BF1 Slope PWL Segment 4 = -48 + dscl_prog_data->easf_h_bf1_pwl_in_seg5 = 48; // S0.10, BF1 PWL Segment 5 + dscl_prog_data->easf_h_bf1_pwl_base_seg5 = 50; // U0.6, BF1 Base PWL Segment 5 + dscl_prog_data->easf_h_bf1_pwl_slope_seg5 = 0x710; // S7.3, BF1 Slope PWL Segment 5 = -240 + dscl_prog_data->easf_h_bf1_pwl_in_seg6 = 64; // S0.10, BF1 PWL Segment 6 + dscl_prog_data->easf_h_bf1_pwl_base_seg6 = 20; // U0.6, BF1 Base PWL Segment 6 + dscl_prog_data->easf_h_bf1_pwl_slope_seg6 = 0x760; // S7.3, BF1 Slope PWL Segment 6 = -160 + dscl_prog_data->easf_h_bf1_pwl_in_seg7 = 80; // S0.10, BF1 PWL Segment 7 + dscl_prog_data->easf_h_bf1_pwl_base_seg7 = 0; // U0.6, BF1 Base PWL Segment 7 + dscl_prog_data->easf_h_bf3_pwl_in_set0 = 0x000; // FP0.6.6, BF3 Input value PWL Segment 0 dscl_prog_data->easf_h_bf3_pwl_base_set0 = 63; // S0.6, BF3 Base PWL Segment 0 dscl_prog_data->easf_h_bf3_pwl_slope_set0 = 0x12C5; // FP1.6.6, BF3 Slope PWL Segment 0 @@ -1190,12 +1435,40 @@ static void spl_set_easf_data(struct dscl_prog_data *dscl_prog_data, dscl_prog_data->easf_h_bf3_pwl_slope_set3 = 0x136B; // FP1.6.6, BF3 Slope PWL Segment 3 dscl_prog_data->easf_h_bf3_pwl_in_set4 = 0x0C37; // FP0.6.6, BF3 Input value PWL Segment 4 (0.125 * 125^3) - dscl_prog_data->easf_h_bf3_pwl_base_set4 = -50; // S0.6, BF3 Base PWL Segment 4 + dscl_prog_data->easf_h_bf3_pwl_base_set4 = 0x4E; // S0.6, BF3 Base PWL Segment 4 = -50 dscl_prog_data->easf_h_bf3_pwl_slope_set4 = 0x1200; // FP1.6.6, BF3 Slope PWL Segment 4 dscl_prog_data->easf_h_bf3_pwl_in_set5 = 0x0CF7; // FP0.6.6, BF3 Input value PWL Segment 5 (1.0 * 125^3) - dscl_prog_data->easf_h_bf3_pwl_base_set5 = -63; // S0.6, BF3 Base PWL Segment 5 + dscl_prog_data->easf_h_bf3_pwl_base_set5 = 0x41; // S0.6, BF3 Base PWL Segment 5 = -63 } else { + dscl_prog_data->easf_h_bf2_flat1_gain = 13; // U1.3, BF2 Flat1 Gain control + dscl_prog_data->easf_h_bf2_flat2_gain = 15; // U4.0, BF2 Flat2 Gain control + dscl_prog_data->easf_h_bf2_roc_gain = 14; // U2.2, Rate Of Change control + + dscl_prog_data->easf_h_bf1_pwl_in_seg0 = 0x440; // S0.10, BF1 PWL Segment 0 = -960 + dscl_prog_data->easf_h_bf1_pwl_base_seg0 = 0; // U0.6, BF1 Base PWL Segment 0 + dscl_prog_data->easf_h_bf1_pwl_slope_seg0 = 2; // S7.3, BF1 Slope PWL Segment 0 + dscl_prog_data->easf_h_bf1_pwl_in_seg1 = 0x7C4; // S0.10, BF1 PWL Segment 1 = -60 + dscl_prog_data->easf_h_bf1_pwl_base_seg1 = 12; // U0.6, BF1 Base PWL Segment 1 + dscl_prog_data->easf_h_bf1_pwl_slope_seg1 = 109; // S7.3, BF1 Slope PWL Segment 1 + dscl_prog_data->easf_h_bf1_pwl_in_seg2 = 0; // S0.10, BF1 PWL Segment 2 + dscl_prog_data->easf_h_bf1_pwl_base_seg2 = 63; // U0.6, BF1 Base PWL Segment 2 + dscl_prog_data->easf_h_bf1_pwl_slope_seg2 = 0; // S7.3, BF1 Slope PWL Segment 2 + dscl_prog_data->easf_h_bf1_pwl_in_seg3 = 48; // S0.10, BF1 PWL Segment 3 + dscl_prog_data->easf_h_bf1_pwl_base_seg3 = 63; // U0.6, BF1 Base PWL Segment 3 + dscl_prog_data->easf_h_bf1_pwl_slope_seg3 = 0x7ED; // S7.3, BF1 Slope PWL Segment 3 = -19 + dscl_prog_data->easf_h_bf1_pwl_in_seg4 = 96; // S0.10, BF1 PWL Segment 4 + dscl_prog_data->easf_h_bf1_pwl_base_seg4 = 56; // U0.6, BF1 Base PWL Segment 4 + dscl_prog_data->easf_h_bf1_pwl_slope_seg4 = 0x7F0; // S7.3, BF1 Slope PWL Segment 4 = -16 + dscl_prog_data->easf_h_bf1_pwl_in_seg5 = 144; // S0.10, BF1 PWL Segment 5 + dscl_prog_data->easf_h_bf1_pwl_base_seg5 = 50; // U0.6, BF1 Base PWL Segment 5 + dscl_prog_data->easf_h_bf1_pwl_slope_seg5 = 0x7B0; // S7.3, BF1 Slope PWL Segment 5 = -80 + dscl_prog_data->easf_h_bf1_pwl_in_seg6 = 192; // S0.10, BF1 PWL Segment 6 + dscl_prog_data->easf_h_bf1_pwl_base_seg6 = 20; // U0.6, BF1 Base PWL Segment 6 + dscl_prog_data->easf_h_bf1_pwl_slope_seg6 = 0x7CB; // S7.3, BF1 Slope PWL Segment 6 = -53 + dscl_prog_data->easf_h_bf1_pwl_in_seg7 = 240; // S0.10, BF1 PWL Segment 7 + dscl_prog_data->easf_h_bf1_pwl_base_seg7 = 0; // U0.6, BF1 Base PWL Segment 7 + dscl_prog_data->easf_h_bf3_pwl_in_set0 = 0x000; // FP0.6.6, BF3 Input value PWL Segment 0 dscl_prog_data->easf_h_bf3_pwl_base_set0 = 63; // S0.6, BF3 Base PWL Segment 0 dscl_prog_data->easf_h_bf3_pwl_slope_set0 = 0x0000; // FP1.6.6, BF3 Slope PWL Segment 0 @@ -1213,25 +1486,30 @@ static void spl_set_easf_data(struct dscl_prog_data *dscl_prog_data, dscl_prog_data->easf_h_bf3_pwl_slope_set3 = 0x1878; // FP1.6.6, BF3 Slope PWL Segment 3 dscl_prog_data->easf_h_bf3_pwl_in_set4 = 0x0761; // FP0.6.6, BF3 Input value PWL Segment 4 (0.375) - dscl_prog_data->easf_h_bf3_pwl_base_set4 = -60; // S0.6, BF3 Base PWL Segment 4 + dscl_prog_data->easf_h_bf3_pwl_base_set4 = 0x44; // S0.6, BF3 Base PWL Segment 4 = -60 dscl_prog_data->easf_h_bf3_pwl_slope_set4 = 0x1760; // FP1.6.6, BF3 Slope PWL Segment 4 dscl_prog_data->easf_h_bf3_pwl_in_set5 = 0x0780; // FP0.6.6, BF3 Input value PWL Segment 5 (0.5) - dscl_prog_data->easf_h_bf3_pwl_base_set5 = -63; // S0.6, BF3 Base PWL Segment 5 + dscl_prog_data->easf_h_bf3_pwl_base_set5 = 0x41; // S0.6, BF3 Base PWL Segment 5 = -63 } // if (lls_pref == LLS_PREF_YES) } else dscl_prog_data->easf_h_en = false; if (lls_pref == LLS_PREF_YES) { dscl_prog_data->easf_ltonl_en = 1; // Linear input - dscl_prog_data->easf_matrix_c0 = - 0x504E; // fp1.5.10, C0 coefficient (LN_BT2020: 0.2627 * (2^14)/125 = 34.43750000) - dscl_prog_data->easf_matrix_c1 = - 0x558E; // fp1.5.10, C1 coefficient (LN_BT2020: 0.6780 * (2^14)/125 = 88.87500000) - dscl_prog_data->easf_matrix_c2 = - 0x47C6; // fp1.5.10, C2 coefficient (LN_BT2020: 0.0593 * (2^14)/125 = 7.77343750) - dscl_prog_data->easf_matrix_c3 = - 0x0; // fp1.5.10, C3 coefficient + if ((setup == HDR_L) && (spl_is_rgb8(format))) { + /* Calculate C0-C3 coefficients based on HDR multiplier */ + spl_calculate_c0_c3_hdr(dscl_prog_data, hdr_multx100); + } else { // HDR_L ( DWM ) and SDR_L + dscl_prog_data->easf_matrix_c0 = + 0x4EF7; // fp1.5.10, C0 coefficient (LN_rec709: 0.2126 * (2^14)/125 = 27.86590720) + dscl_prog_data->easf_matrix_c1 = + 0x55DC; // fp1.5.10, C1 coefficient (LN_rec709: 0.7152 * (2^14)/125 = 93.74269440) + dscl_prog_data->easf_matrix_c2 = + 0x48BB; // fp1.5.10, C2 coefficient (LN_rec709: 0.0722 * (2^14)/125 = 9.46339840) + dscl_prog_data->easf_matrix_c3 = + 0x0; // fp1.5.10, C3 coefficient + } } else { dscl_prog_data->easf_ltonl_en = 0; // Non-Linear input dscl_prog_data->easf_matrix_c0 = @@ -1243,27 +1521,43 @@ static void spl_set_easf_data(struct dscl_prog_data *dscl_prog_data, dscl_prog_data->easf_matrix_c3 = 0x0; // fp1.5.10, C3 coefficient } + + if (spl_is_yuv420(format)) { /* TODO: 0 = RGB, 1 = YUV */ + dscl_prog_data->easf_matrix_mode = 1; + /* + * 2-bit, BF3 chroma mode correction calculation mode + * Needs to be disabled for YUV420 mode + * Override lookup value + */ + dscl_prog_data->easf_v_bf3_mode = 0; + dscl_prog_data->easf_h_bf3_mode = 0; + } else + dscl_prog_data->easf_matrix_mode = 0; + } + /*Set isharp noise detection */ -static void spl_set_isharp_noise_det_mode(struct dscl_prog_data *dscl_prog_data) +static void spl_set_isharp_noise_det_mode(struct dscl_prog_data *dscl_prog_data, + const struct spl_scaler_data *data) { // ISHARP_NOISEDET_MODE // 0: 3x5 as VxH // 1: 4x5 as VxH // 2: // 3: 5x5 as VxH - if (dscl_prog_data->taps.v_taps == 6) - dscl_prog_data->isharp_noise_det.mode = 3; // ISHARP_NOISEDET_MODE - else if (dscl_prog_data->taps.h_taps == 4) - dscl_prog_data->isharp_noise_det.mode = 1; // ISHARP_NOISEDET_MODE - else if (dscl_prog_data->taps.h_taps == 3) - dscl_prog_data->isharp_noise_det.mode = 0; // ISHARP_NOISEDET_MODE + if (data->taps.v_taps == 6) + dscl_prog_data->isharp_noise_det.mode = 3; + else if (data->taps.v_taps == 4) + dscl_prog_data->isharp_noise_det.mode = 1; + else if (data->taps.v_taps == 3) + dscl_prog_data->isharp_noise_det.mode = 0; }; /* Set Sharpener data */ static void spl_set_isharp_data(struct dscl_prog_data *dscl_prog_data, struct adaptive_sharpness adp_sharpness, bool enable_isharp, enum linear_light_scaling lls_pref, enum spl_pixel_format format, - const struct spl_scaler_data *data) + const struct spl_scaler_data *data, struct spl_fixed31_32 ratio, + enum system_setup setup) { /* Turn off sharpener if not required */ if (!enable_isharp) { @@ -1272,10 +1566,12 @@ static void spl_set_isharp_data(struct dscl_prog_data *dscl_prog_data, } dscl_prog_data->isharp_en = 1; // ISHARP_EN - dscl_prog_data->isharp_noise_det.enable = 1; // ISHARP_NOISEDET_EN // Set ISHARP_NOISEDET_MODE if htaps = 6-tap - if (dscl_prog_data->taps.h_taps == 6) - spl_set_isharp_noise_det_mode(dscl_prog_data); // ISHARP_NOISEDET_MODE + if (data->taps.h_taps == 6) { + dscl_prog_data->isharp_noise_det.enable = 1; /* ISHARP_NOISEDET_EN */ + spl_set_isharp_noise_det_mode(dscl_prog_data, data); /* ISHARP_NOISEDET_MODE */ + } else + dscl_prog_data->isharp_noise_det.enable = 0; // ISHARP_NOISEDET_EN // Program noise detection threshold dscl_prog_data->isharp_noise_det.uthreshold = 24; // ISHARP_NOISEDET_UTHRE dscl_prog_data->isharp_noise_det.dthreshold = 4; // ISHARP_NOISEDET_DTHRE @@ -1284,50 +1580,93 @@ static void spl_set_isharp_data(struct dscl_prog_data *dscl_prog_data, dscl_prog_data->isharp_noise_det.pwl_end_in = 13; // ISHARP_NOISEDET_PWL_END_IN dscl_prog_data->isharp_noise_det.pwl_slope = 1623; // ISHARP_NOISEDET_PWL_SLOPE - if ((lls_pref == LLS_PREF_NO) && !spl_is_yuv420(format)) /* ISHARP_FMT_MODE */ + if (lls_pref == LLS_PREF_NO) /* ISHARP_FMT_MODE */ dscl_prog_data->isharp_fmt.mode = 1; else dscl_prog_data->isharp_fmt.mode = 0; dscl_prog_data->isharp_fmt.norm = 0x3C00; // ISHARP_FMT_NORM dscl_prog_data->isharp_lba.mode = 0; // ISHARP_LBA_MODE - // ISHARP_LBA_PWL_SEG0: ISHARP Local Brightness Adjustment PWL Segment 0 - dscl_prog_data->isharp_lba.in_seg[0] = 0; // ISHARP LBA PWL for Seg 0. INPUT value in U0.10 format - dscl_prog_data->isharp_lba.base_seg[0] = 0; // ISHARP LBA PWL for Seg 0. BASE value in U0.6 format - dscl_prog_data->isharp_lba.slope_seg[0] = 32; // ISHARP LBA for Seg 0. SLOPE value in S5.3 format - // ISHARP_LBA_PWL_SEG1: ISHARP LBA PWL Segment 1 - dscl_prog_data->isharp_lba.in_seg[1] = 256; // ISHARP LBA PWL for Seg 1. INPUT value in U0.10 format - dscl_prog_data->isharp_lba.base_seg[1] = 63; // ISHARP LBA PWL for Seg 1. BASE value in U0.6 format - dscl_prog_data->isharp_lba.slope_seg[1] = 0; // ISHARP LBA for Seg 1. SLOPE value in S5.3 format - // ISHARP_LBA_PWL_SEG2: ISHARP LBA PWL Segment 2 - dscl_prog_data->isharp_lba.in_seg[2] = 614; // ISHARP LBA PWL for Seg 2. INPUT value in U0.10 format - dscl_prog_data->isharp_lba.base_seg[2] = 63; // ISHARP LBA PWL for Seg 2. BASE value in U0.6 format - dscl_prog_data->isharp_lba.slope_seg[2] = -20; // ISHARP LBA for Seg 2. SLOPE value in S5.3 format - // ISHARP_LBA_PWL_SEG3: ISHARP LBA PWL Segment 3 - dscl_prog_data->isharp_lba.in_seg[3] = 1023; // ISHARP LBA PWL for Seg 3.INPUT value in U0.10 format - dscl_prog_data->isharp_lba.base_seg[3] = 0; // ISHARP LBA PWL for Seg 3. BASE value in U0.6 format - dscl_prog_data->isharp_lba.slope_seg[3] = 0; // ISHARP LBA for Seg 3. SLOPE value in S5.3 format - // ISHARP_LBA_PWL_SEG4: ISHARP LBA PWL Segment 4 - dscl_prog_data->isharp_lba.in_seg[4] = 1023; // ISHARP LBA PWL for Seg 4.INPUT value in U0.10 format - dscl_prog_data->isharp_lba.base_seg[4] = 0; // ISHARP LBA PWL for Seg 4. BASE value in U0.6 format - dscl_prog_data->isharp_lba.slope_seg[4] = 0; // ISHARP LBA for Seg 4. SLOPE value in S5.3 format - // ISHARP_LBA_PWL_SEG5: ISHARP LBA PWL Segment 5 - dscl_prog_data->isharp_lba.in_seg[5] = 1023; // ISHARP LBA PWL for Seg 5.INPUT value in U0.10 format - dscl_prog_data->isharp_lba.base_seg[5] = 0; // ISHARP LBA PWL for Seg 5. BASE value in U0.6 format - switch (adp_sharpness.sharpness) { - case SHARPNESS_LOW: - dscl_prog_data->isharp_delta = spl_get_filter_isharp_1D_lut_0p5x(); - break; - case SHARPNESS_MID: - dscl_prog_data->isharp_delta = spl_get_filter_isharp_1D_lut_1p0x(); - break; - case SHARPNESS_HIGH: - dscl_prog_data->isharp_delta = spl_get_filter_isharp_1D_lut_2p0x(); - break; - default: - BREAK_TO_DEBUGGER(); + + if (setup == SDR_L) { + // ISHARP_LBA_PWL_SEG0: ISHARP Local Brightness Adjustment PWL Segment 0 + dscl_prog_data->isharp_lba.in_seg[0] = 0; // ISHARP LBA PWL for Seg 0. INPUT value in U0.10 format + dscl_prog_data->isharp_lba.base_seg[0] = 0; // ISHARP LBA PWL for Seg 0. BASE value in U0.6 format + dscl_prog_data->isharp_lba.slope_seg[0] = 62; // ISHARP LBA for Seg 0. SLOPE value in S5.3 format + // ISHARP_LBA_PWL_SEG1: ISHARP LBA PWL Segment 1 + dscl_prog_data->isharp_lba.in_seg[1] = 130; // ISHARP LBA PWL for Seg 1. INPUT value in U0.10 format + dscl_prog_data->isharp_lba.base_seg[1] = 63; // ISHARP LBA PWL for Seg 1. BASE value in U0.6 format + dscl_prog_data->isharp_lba.slope_seg[1] = 0; // ISHARP LBA for Seg 1. SLOPE value in S5.3 format + // ISHARP_LBA_PWL_SEG2: ISHARP LBA PWL Segment 2 + dscl_prog_data->isharp_lba.in_seg[2] = 450; // ISHARP LBA PWL for Seg 2. INPUT value in U0.10 format + dscl_prog_data->isharp_lba.base_seg[2] = 63; // ISHARP LBA PWL for Seg 2. BASE value in U0.6 format + dscl_prog_data->isharp_lba.slope_seg[2] = 0x18D; // ISHARP LBA for Seg 2. SLOPE value in S5.3 format = -115 + // ISHARP_LBA_PWL_SEG3: ISHARP LBA PWL Segment 3 + dscl_prog_data->isharp_lba.in_seg[3] = 520; // ISHARP LBA PWL for Seg 3.INPUT value in U0.10 format + dscl_prog_data->isharp_lba.base_seg[3] = 0; // ISHARP LBA PWL for Seg 3. BASE value in U0.6 format + dscl_prog_data->isharp_lba.slope_seg[3] = 0; // ISHARP LBA for Seg 3. SLOPE value in S5.3 format + // ISHARP_LBA_PWL_SEG4: ISHARP LBA PWL Segment 4 + dscl_prog_data->isharp_lba.in_seg[4] = 520; // ISHARP LBA PWL for Seg 4.INPUT value in U0.10 format + dscl_prog_data->isharp_lba.base_seg[4] = 0; // ISHARP LBA PWL for Seg 4. BASE value in U0.6 format + dscl_prog_data->isharp_lba.slope_seg[4] = 0; // ISHARP LBA for Seg 4. SLOPE value in S5.3 format + // ISHARP_LBA_PWL_SEG5: ISHARP LBA PWL Segment 5 + dscl_prog_data->isharp_lba.in_seg[5] = 520; // ISHARP LBA PWL for Seg 5.INPUT value in U0.10 format + dscl_prog_data->isharp_lba.base_seg[5] = 0; // ISHARP LBA PWL for Seg 5. BASE value in U0.6 format + } else if (setup == HDR_L) { + // ISHARP_LBA_PWL_SEG0: ISHARP Local Brightness Adjustment PWL Segment 0 + dscl_prog_data->isharp_lba.in_seg[0] = 0; // ISHARP LBA PWL for Seg 0. INPUT value in U0.10 format + dscl_prog_data->isharp_lba.base_seg[0] = 0; // ISHARP LBA PWL for Seg 0. BASE value in U0.6 format + dscl_prog_data->isharp_lba.slope_seg[0] = 32; // ISHARP LBA for Seg 0. SLOPE value in S5.3 format + // ISHARP_LBA_PWL_SEG1: ISHARP LBA PWL Segment 1 + dscl_prog_data->isharp_lba.in_seg[1] = 254; // ISHARP LBA PWL for Seg 1. INPUT value in U0.10 format + dscl_prog_data->isharp_lba.base_seg[1] = 63; // ISHARP LBA PWL for Seg 1. BASE value in U0.6 format + dscl_prog_data->isharp_lba.slope_seg[1] = 0; // ISHARP LBA for Seg 1. SLOPE value in S5.3 format + // ISHARP_LBA_PWL_SEG2: ISHARP LBA PWL Segment 2 + dscl_prog_data->isharp_lba.in_seg[2] = 559; // ISHARP LBA PWL for Seg 2. INPUT value in U0.10 format + dscl_prog_data->isharp_lba.base_seg[2] = 63; // ISHARP LBA PWL for Seg 2. BASE value in U0.6 format + dscl_prog_data->isharp_lba.slope_seg[2] = 0x10C; // ISHARP LBA for Seg 2. SLOPE value in S5.3 format = -244 + // ISHARP_LBA_PWL_SEG3: ISHARP LBA PWL Segment 3 + dscl_prog_data->isharp_lba.in_seg[3] = 592; // ISHARP LBA PWL for Seg 3.INPUT value in U0.10 format + dscl_prog_data->isharp_lba.base_seg[3] = 0; // ISHARP LBA PWL for Seg 3. BASE value in U0.6 format + dscl_prog_data->isharp_lba.slope_seg[3] = 0; // ISHARP LBA for Seg 3. SLOPE value in S5.3 format + // ISHARP_LBA_PWL_SEG4: ISHARP LBA PWL Segment 4 + dscl_prog_data->isharp_lba.in_seg[4] = 1023; // ISHARP LBA PWL for Seg 4.INPUT value in U0.10 format + dscl_prog_data->isharp_lba.base_seg[4] = 0; // ISHARP LBA PWL for Seg 4. BASE value in U0.6 format + dscl_prog_data->isharp_lba.slope_seg[4] = 0; // ISHARP LBA for Seg 4. SLOPE value in S5.3 format + // ISHARP_LBA_PWL_SEG5: ISHARP LBA PWL Segment 5 + dscl_prog_data->isharp_lba.in_seg[5] = 1023; // ISHARP LBA PWL for Seg 5.INPUT value in U0.10 format + dscl_prog_data->isharp_lba.base_seg[5] = 0; // ISHARP LBA PWL for Seg 5. BASE value in U0.6 format + } else { + // ISHARP_LBA_PWL_SEG0: ISHARP Local Brightness Adjustment PWL Segment 0 + dscl_prog_data->isharp_lba.in_seg[0] = 0; // ISHARP LBA PWL for Seg 0. INPUT value in U0.10 format + dscl_prog_data->isharp_lba.base_seg[0] = 0; // ISHARP LBA PWL for Seg 0. BASE value in U0.6 format + dscl_prog_data->isharp_lba.slope_seg[0] = 40; // ISHARP LBA for Seg 0. SLOPE value in S5.3 format + // ISHARP_LBA_PWL_SEG1: ISHARP LBA PWL Segment 1 + dscl_prog_data->isharp_lba.in_seg[1] = 204; // ISHARP LBA PWL for Seg 1. INPUT value in U0.10 format + dscl_prog_data->isharp_lba.base_seg[1] = 63; // ISHARP LBA PWL for Seg 1. BASE value in U0.6 format + dscl_prog_data->isharp_lba.slope_seg[1] = 0; // ISHARP LBA for Seg 1. SLOPE value in S5.3 format + // ISHARP_LBA_PWL_SEG2: ISHARP LBA PWL Segment 2 + dscl_prog_data->isharp_lba.in_seg[2] = 818; // ISHARP LBA PWL for Seg 2. INPUT value in U0.10 format + dscl_prog_data->isharp_lba.base_seg[2] = 63; // ISHARP LBA PWL for Seg 2. BASE value in U0.6 format + dscl_prog_data->isharp_lba.slope_seg[2] = 0x1D9; // ISHARP LBA for Seg 2. SLOPE value in S5.3 format = -39 + // ISHARP_LBA_PWL_SEG3: ISHARP LBA PWL Segment 3 + dscl_prog_data->isharp_lba.in_seg[3] = 1023; // ISHARP LBA PWL for Seg 3.INPUT value in U0.10 format + dscl_prog_data->isharp_lba.base_seg[3] = 0; // ISHARP LBA PWL for Seg 3. BASE value in U0.6 format + dscl_prog_data->isharp_lba.slope_seg[3] = 0; // ISHARP LBA for Seg 3. SLOPE value in S5.3 format + // ISHARP_LBA_PWL_SEG4: ISHARP LBA PWL Segment 4 + dscl_prog_data->isharp_lba.in_seg[4] = 1023; // ISHARP LBA PWL for Seg 4.INPUT value in U0.10 format + dscl_prog_data->isharp_lba.base_seg[4] = 0; // ISHARP LBA PWL for Seg 4. BASE value in U0.6 format + dscl_prog_data->isharp_lba.slope_seg[4] = 0; // ISHARP LBA for Seg 4. SLOPE value in S5.3 format + // ISHARP_LBA_PWL_SEG5: ISHARP LBA PWL Segment 5 + dscl_prog_data->isharp_lba.in_seg[5] = 1023; // ISHARP LBA PWL for Seg 5.INPUT value in U0.10 format + dscl_prog_data->isharp_lba.base_seg[5] = 0; // ISHARP LBA PWL for Seg 5. BASE value in U0.6 format } + + spl_build_isharp_1dlut_from_reference_curve(ratio, setup, adp_sharpness); + dscl_prog_data->isharp_delta = spl_get_pregen_filter_isharp_1D_lut(setup); + dscl_prog_data->sharpness_level = adp_sharpness.sharpness_level; + // Program the nldelta soft clip values if (lls_pref == LLS_PREF_YES) { dscl_prog_data->isharp_nldelta_sclip.enable_p = 0; /* ISHARP_NLDELTA_SCLIP_EN_P */ @@ -1346,62 +1685,7 @@ static void spl_set_isharp_data(struct dscl_prog_data *dscl_prog_data, } // Set the values as per lookup table -#ifdef CONFIG_DRM_AMD_DC_FP spl_set_blur_scale_data(dscl_prog_data, data); -#endif -} -static bool spl_get_isharp_en(struct adaptive_sharpness adp_sharpness, - int vscale_ratio, int hscale_ratio, struct spl_taps taps, - enum spl_pixel_format format) -{ - bool enable_isharp = false; - - if (adp_sharpness.enable == false) - return enable_isharp; // Return if adaptive sharpness is disabled - // Is downscaling ? - if (vscale_ratio > 1 || hscale_ratio > 1) { - // END - No iSHARP support for downscaling - return enable_isharp; - } - // Scaling is up to 1:1 (no scaling) or upscaling - - /* Only apply sharpness to NV12 and not P010 */ - if (format != SPL_PIXEL_FORMAT_420BPP8) - return enable_isharp; - - // LB support horizontal taps 4,6 or vertical taps 3, 4, 6 - if (taps.h_taps == 4 || taps.h_taps == 6 || - taps.v_taps == 3 || taps.v_taps == 4 || taps.v_taps == 6) { - // END - iSHARP supported - enable_isharp = true; - } - return enable_isharp; -} - -static bool spl_choose_lls_policy(enum spl_pixel_format format, - enum spl_transfer_func_type tf_type, - enum spl_transfer_func_predefined tf_predefined_type, - enum linear_light_scaling *lls_pref) -{ - if (spl_is_yuv420(format)) { - *lls_pref = LLS_PREF_NO; - if ((tf_type == SPL_TF_TYPE_PREDEFINED) || (tf_type == SPL_TF_TYPE_DISTRIBUTED_POINTS)) - return true; - } else { /* RGB or YUV444 */ - if (tf_type == SPL_TF_TYPE_PREDEFINED) { - if ((tf_predefined_type == SPL_TRANSFER_FUNCTION_HLG) || - (tf_predefined_type == SPL_TRANSFER_FUNCTION_HLG12)) - *lls_pref = LLS_PREF_NO; - else - *lls_pref = LLS_PREF_YES; - return true; - } else if (tf_type == SPL_TF_TYPE_BYPASS) { - *lls_pref = LLS_PREF_YES; - return true; - } - } - *lls_pref = LLS_PREF_NO; - return false; } /* Calculate scaler parameters */ @@ -1410,67 +1694,74 @@ bool spl_calculate_scaler_params(struct spl_in *spl_in, struct spl_out *spl_out) bool res = false; bool enable_easf_v = false; bool enable_easf_h = false; - bool lls_enable_easf = true; int vratio = 0; int hratio = 0; - const struct spl_scaler_data *data = &spl_out->scl_data; + struct spl_scratch spl_scratch; + struct spl_fixed31_32 isharp_scale_ratio; + enum system_setup setup; + bool enable_isharp = false; + const struct spl_scaler_data *data = &spl_scratch.scl_data; + + memset(&spl_scratch, 0, sizeof(struct spl_scratch)); + spl_scratch.scl_data.h_active = spl_in->h_active; + spl_scratch.scl_data.v_active = spl_in->v_active; + // All SPL calls /* recout calculation */ /* depends on h_active */ - spl_calculate_recout(spl_in, spl_out); + spl_calculate_recout(spl_in, &spl_scratch, spl_out); /* depends on pixel format */ - spl_calculate_scaling_ratios(spl_in, spl_out); + spl_calculate_scaling_ratios(spl_in, &spl_scratch, spl_out); /* depends on scaling ratios and recout, does not calculate offset yet */ - spl_calculate_viewport_size(spl_in, spl_out); + spl_calculate_viewport_size(spl_in, &spl_scratch); res = spl_get_optimal_number_of_taps( spl_in->basic_out.max_downscale_src_width, spl_in, - spl_out, &spl_in->scaling_quality); + &spl_scratch, &spl_in->scaling_quality, &enable_easf_v, + &enable_easf_h, &enable_isharp); /* * Depends on recout, scaling ratios, h_active and taps * May need to re-check lb size after this in some obscure scenario */ if (res) - spl_calculate_inits_and_viewports(spl_in, spl_out); + spl_calculate_inits_and_viewports(spl_in, &spl_scratch); // Handle 3d recout - spl_handle_3d_recout(spl_in, &spl_out->scl_data.recout); + spl_handle_3d_recout(spl_in, &spl_scratch.scl_data.recout); // Clamp - spl_clamp_viewport(&spl_out->scl_data.viewport); + spl_clamp_viewport(&spl_scratch.scl_data.viewport); if (!res) return res; - /* - * If lls_pref is LLS_PREF_DONT_CARE, then use pixel format and transfer - * function to determine whether to use LINEAR or NONLINEAR scaling - */ - if (spl_in->lls_pref == LLS_PREF_DONT_CARE) - lls_enable_easf = spl_choose_lls_policy(spl_in->basic_in.format, - spl_in->basic_in.tf_type, spl_in->basic_in.tf_predefined_type, - &spl_in->lls_pref); - // Save all calculated parameters in dscl_prog_data structure to program hw registers - spl_set_dscl_prog_data(spl_in, spl_out); + spl_set_dscl_prog_data(spl_in, &spl_scratch, spl_out, enable_easf_v, enable_easf_h, enable_isharp); - vratio = dc_fixpt_ceil(spl_out->scl_data.ratios.vert); - hratio = dc_fixpt_ceil(spl_out->scl_data.ratios.horz); - if (!lls_enable_easf || spl_in->disable_easf) { - enable_easf_v = false; - enable_easf_h = false; + if (spl_in->lls_pref == LLS_PREF_YES) { + if (spl_in->is_hdr_on) + setup = HDR_L; + else + setup = SDR_L; } else { - /* Enable EASF on vertical? */ - enable_easf_v = enable_easf(vratio, spl_out->scl_data.taps.v_taps, spl_in->lls_pref, spl_in->prefer_easf); - /* Enable EASF on horizontal? */ - enable_easf_h = enable_easf(hratio, spl_out->scl_data.taps.h_taps, spl_in->lls_pref, spl_in->prefer_easf); + if (spl_in->is_hdr_on) + setup = HDR_NL; + else + setup = SDR_NL; } + // Set EASF - spl_set_easf_data(spl_out->dscl_prog_data, enable_easf_v, enable_easf_h, spl_in->lls_pref, - spl_in->basic_in.format); + spl_set_easf_data(&spl_scratch, spl_out, enable_easf_v, enable_easf_h, spl_in->lls_pref, + spl_in->basic_in.format, setup, spl_in->hdr_multx100); + // Set iSHARP - bool enable_isharp = spl_get_isharp_en(spl_in->adaptive_sharpness, vratio, hratio, - spl_out->scl_data.taps, spl_in->basic_in.format); + vratio = spl_fixpt_ceil(spl_scratch.scl_data.ratios.vert); + hratio = spl_fixpt_ceil(spl_scratch.scl_data.ratios.horz); + if (vratio <= hratio) + isharp_scale_ratio = spl_scratch.scl_data.recip_ratios.vert; + else + isharp_scale_ratio = spl_scratch.scl_data.recip_ratios.horz; + spl_set_isharp_data(spl_out->dscl_prog_data, spl_in->adaptive_sharpness, enable_isharp, - spl_in->lls_pref, spl_in->basic_in.format, data); + spl_in->lls_pref, spl_in->basic_in.format, data, isharp_scale_ratio, setup); return res; } diff --git a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_filters.c b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_filters.c new file mode 100644 index 000000000000..99238644e0a1 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_filters.c @@ -0,0 +1,15 @@ +// SPDX-License-Identifier: MIT +// +// Copyright 2024 Advanced Micro Devices, Inc. + +#include "dc_spl_filters.h" + +void convert_filter_s1_10_to_s1_12(const uint16_t *s1_10_filter, + uint16_t *s1_12_filter, int num_taps) +{ + int num_entries = NUM_PHASES_COEFF * num_taps; + int i; + + for (i = 0; i < num_entries; i++) + *(s1_12_filter + i) = *(s1_10_filter + i) * 4; +} diff --git a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_filters.h b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_filters.h new file mode 100644 index 000000000000..20439cdbdb10 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_filters.h @@ -0,0 +1,15 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright 2024 Advanced Micro Devices, Inc. */ + +#ifndef __DC_SPL_FILTERS_H__ +#define __DC_SPL_FILTERS_H__ + +#include "dc_spl_types.h" + +#define NUM_PHASES_COEFF 33 + +void convert_filter_s1_10_to_s1_12(const uint16_t *s1_10_filter, + uint16_t *s1_12_filter, int num_taps); + +#endif /* __DC_SPL_FILTERS_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_isharp_filters.c b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_isharp_filters.c index a5d9a6223d06..33712f50d303 100644 --- a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_isharp_filters.c +++ b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_isharp_filters.c @@ -2,6 +2,8 @@ // // Copyright 2024 Advanced Micro Devices, Inc. +#include "spl_debug.h" +#include "dc_spl_filters.h" #include "dc_spl_isharp_filters.h" //======================================== @@ -15,7 +17,7 @@ // C_start = 40.000000 // C_end = 64.000000 //======================================== -static const uint32_t filter_isharp_1D_lut_0[32] = { +static const uint32_t filter_isharp_1D_lut_0[ISHARP_LUT_TABLE_SIZE] = { 0x02010000, 0x0A070503, 0x1614100D, @@ -61,7 +63,7 @@ static const uint32_t filter_isharp_1D_lut_0[32] = { // C_end = 127.000000 //======================================== -static const uint32_t filter_isharp_1D_lut_0p5x[32] = { +static const uint32_t filter_isharp_1D_lut_0p5x[ISHARP_LUT_TABLE_SIZE] = { 0x00000000, 0x02020101, 0x06050403, @@ -106,7 +108,7 @@ static const uint32_t filter_isharp_1D_lut_0p5x[32] = { // C_start = 96.000000 // C_end = 127.000000 //======================================== -static const uint32_t filter_isharp_1D_lut_1p0x[32] = { +static const uint32_t filter_isharp_1D_lut_1p0x[ISHARP_LUT_TABLE_SIZE] = { 0x01000000, 0x05040302, 0x0B0A0806, @@ -151,7 +153,7 @@ static const uint32_t filter_isharp_1D_lut_1p0x[32] = { // C_start = 96.000000 // C_end = 127.000000 //======================================== -static const uint32_t filter_isharp_1D_lut_1p5x[32] = { +static const uint32_t filter_isharp_1D_lut_1p5x[ISHARP_LUT_TABLE_SIZE] = { 0x01010000, 0x07050402, 0x110F0C0A, @@ -196,7 +198,7 @@ static const uint32_t filter_isharp_1D_lut_1p5x[32] = { // C_start = 40.000000 // C_end = 127.000000 //======================================== -static const uint32_t filter_isharp_1D_lut_2p0x[32] = { +static const uint32_t filter_isharp_1D_lut_2p0x[ISHARP_LUT_TABLE_SIZE] = { 0x02010000, 0x0A070503, 0x1614100D, @@ -230,6 +232,53 @@ static const uint32_t filter_isharp_1D_lut_2p0x[32] = { 0x080B0D0E, 0x00020406, }; +//======================================== +// Delta Gain 1DLUT +// LUT content is packed as 4-bytes into one DWORD/entry +// A_start = 0.000000 +// A_end = 10.000000 +// A_gain = 3.000000 +// B_start = 11.000000 +// B_end = 127.000000 +// C_start = 40.000000 +// C_end = 127.000000 +//======================================== +static const uint32_t filter_isharp_1D_lut_3p0x[ISHARP_LUT_TABLE_SIZE] = { +0x03010000, +0x0F0B0805, +0x211E1813, +0x2B292624, +0x3533302E, +0x3E3C3A37, +0x46444240, +0x4D4B4A48, +0x5352504F, +0x59575655, +0x5D5C5B5A, +0x61605F5E, +0x64646362, +0x66666565, +0x68686767, +0x68686868, +0x68686868, +0x67676868, +0x65656666, +0x62636464, +0x5E5F6061, +0x5A5B5C5D, +0x55565759, +0x4F505253, +0x484A4B4D, +0x40424446, +0x373A3C3E, +0x2E303335, +0x2426292B, +0x191B1E21, +0x0D101316, +0x0003060A, +}; + +//======================================== // Wide scaler coefficients //======================================================== // <using> gen_scaler_coeffs.m @@ -284,7 +333,7 @@ static const uint16_t filter_isharp_wide_6tap_64p[198] = { // <CoefType> Blur & Scale LPF // <CoefQuant> S1.10 //======================================================== -static const uint16_t filter_isharp_bs_4tap_64p[198] = { +static const uint16_t filter_isharp_bs_4tap_in_6_64p[198] = { 0x0000, 0x00E5, 0x0237, 0x00E4, 0x0000, 0x0000, 0x0000, 0x00DE, 0x0237, 0x00EB, 0x0000, 0x0000, 0x0000, 0x00D7, 0x0236, 0x00F2, 0x0001, 0x0000, @@ -319,6 +368,138 @@ static const uint16_t filter_isharp_bs_4tap_64p[198] = { 0x0000, 0x003B, 0x01CF, 0x01C2, 0x0034, 0x0000, 0x0000, 0x0037, 0x01C9, 0x01C9, 0x0037, 0x0000 }; +//======================================================== +// <using> gen_BlurScale_coeffs.m +// <date> 25-Apr-2022 +// <num_taps> 4 +// <num_phases> 64 +// <CoefType> Blur & Scale LPF +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t filter_isharp_bs_4tap_64p[132] = { +0x00E5, 0x0237, 0x00E4, 0x0000, +0x00DE, 0x0237, 0x00EB, 0x0000, +0x00D7, 0x0236, 0x00F2, 0x0001, +0x00D0, 0x0235, 0x00FA, 0x0001, +0x00C9, 0x0234, 0x0101, 0x0002, +0x00C2, 0x0233, 0x0108, 0x0003, +0x00BB, 0x0232, 0x0110, 0x0003, +0x00B5, 0x0230, 0x0117, 0x0004, +0x00AE, 0x022E, 0x011F, 0x0005, +0x00A8, 0x022C, 0x0126, 0x0006, +0x00A2, 0x022A, 0x012D, 0x0007, +0x009C, 0x0228, 0x0134, 0x0008, +0x0096, 0x0225, 0x013C, 0x0009, +0x0090, 0x0222, 0x0143, 0x000B, +0x008A, 0x021F, 0x014B, 0x000C, +0x0085, 0x021C, 0x0151, 0x000E, +0x007F, 0x0218, 0x015A, 0x000F, +0x007A, 0x0215, 0x0160, 0x0011, +0x0074, 0x0211, 0x0168, 0x0013, +0x006F, 0x020D, 0x016F, 0x0015, +0x006A, 0x0209, 0x0176, 0x0017, +0x0065, 0x0204, 0x017E, 0x0019, +0x0060, 0x0200, 0x0185, 0x001B, +0x005C, 0x01FB, 0x018C, 0x001D, +0x0057, 0x01F6, 0x0193, 0x0020, +0x0053, 0x01F1, 0x019A, 0x0022, +0x004E, 0x01EC, 0x01A1, 0x0025, +0x004A, 0x01E6, 0x01A8, 0x0028, +0x0046, 0x01E1, 0x01AF, 0x002A, +0x0042, 0x01DB, 0x01B6, 0x002D, +0x003F, 0x01D5, 0x01BB, 0x0031, +0x003B, 0x01CF, 0x01C2, 0x0034, +0x0037, 0x01C9, 0x01C9, 0x0037, +}; +//======================================================== +// <using> gen_BlurScale_coeffs.m +// <date> 09-Jun-2022 +// <num_taps> 3 +// <num_phases> 64 +// <CoefType> Blur & Scale LPF +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t filter_isharp_bs_3tap_64p[99] = { +0x0200, 0x0200, 0x0000, +0x01F6, 0x0206, 0x0004, +0x01EC, 0x020B, 0x0009, +0x01E2, 0x0211, 0x000D, +0x01D8, 0x0216, 0x0012, +0x01CE, 0x021C, 0x0016, +0x01C4, 0x0221, 0x001B, +0x01BA, 0x0226, 0x0020, +0x01B0, 0x022A, 0x0026, +0x01A6, 0x022F, 0x002B, +0x019C, 0x0233, 0x0031, +0x0192, 0x0238, 0x0036, +0x0188, 0x023C, 0x003C, +0x017E, 0x0240, 0x0042, +0x0174, 0x0244, 0x0048, +0x016A, 0x0248, 0x004E, +0x0161, 0x024A, 0x0055, +0x0157, 0x024E, 0x005B, +0x014D, 0x0251, 0x0062, +0x0144, 0x0253, 0x0069, +0x013A, 0x0256, 0x0070, +0x0131, 0x0258, 0x0077, +0x0127, 0x025B, 0x007E, +0x011E, 0x025C, 0x0086, +0x0115, 0x025E, 0x008D, +0x010B, 0x0260, 0x0095, +0x0102, 0x0262, 0x009C, +0x00F9, 0x0263, 0x00A4, +0x00F0, 0x0264, 0x00AC, +0x00E7, 0x0265, 0x00B4, +0x00DF, 0x0264, 0x00BD, +0x00D6, 0x0265, 0x00C5, +0x00CD, 0x0266, 0x00CD, +}; + +/* Converted Blur & Scale coeff tables from S1.10 to S1.12 */ +static uint16_t filter_isharp_bs_4tap_in_6_64p_s1_12[198]; +static uint16_t filter_isharp_bs_4tap_64p_s1_12[132]; +static uint16_t filter_isharp_bs_3tap_64p_s1_12[99]; + +/* Pre-generated 1DLUT for given setup and sharpness level */ +struct isharp_1D_lut_pregen filter_isharp_1D_lut_pregen[NUM_SHARPNESS_SETUPS] = { + { + 0, 0, + { + 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, + } + }, + { + 0, 0, + { + 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, + } + }, + { + 0, 0, + { + 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, + } + }, + { + 0, 0, + { + 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, + } + }, +}; + const uint32_t *spl_get_filter_isharp_1D_lut_0(void) { return filter_isharp_1D_lut_0; @@ -339,11 +520,165 @@ const uint32_t *spl_get_filter_isharp_1D_lut_2p0x(void) { return filter_isharp_1D_lut_2p0x; } +const uint32_t *spl_get_filter_isharp_1D_lut_3p0x(void) +{ + return filter_isharp_1D_lut_3p0x; +} const uint16_t *spl_get_filter_isharp_wide_6tap_64p(void) { return filter_isharp_wide_6tap_64p; } -const uint16_t *spl_get_filter_isharp_bs_4tap_64p(void) +uint16_t *spl_get_filter_isharp_bs_4tap_in_6_64p(void) { - return filter_isharp_bs_4tap_64p; + return filter_isharp_bs_4tap_in_6_64p_s1_12; } +uint16_t *spl_get_filter_isharp_bs_4tap_64p(void) +{ + return filter_isharp_bs_4tap_64p_s1_12; +} +uint16_t *spl_get_filter_isharp_bs_3tap_64p(void) +{ + return filter_isharp_bs_3tap_64p_s1_12; +} + +static unsigned int spl_calculate_sharpness_level(int discrete_sharpness_level, enum system_setup setup, + struct spl_sharpness_range sharpness_range) +{ + unsigned int sharpness_level = 0; + + int min_sharpness, max_sharpness, mid_sharpness; + + switch (setup) { + + case HDR_L: + min_sharpness = sharpness_range.hdr_rgb_min; + max_sharpness = sharpness_range.hdr_rgb_max; + mid_sharpness = sharpness_range.hdr_rgb_mid; + break; + case HDR_NL: + /* currently no use case, use Non-linear SDR values for now */ + case SDR_NL: + min_sharpness = sharpness_range.sdr_yuv_min; + max_sharpness = sharpness_range.sdr_yuv_max; + mid_sharpness = sharpness_range.sdr_yuv_mid; + break; + case SDR_L: + default: + min_sharpness = sharpness_range.sdr_rgb_min; + max_sharpness = sharpness_range.sdr_rgb_max; + mid_sharpness = sharpness_range.sdr_rgb_mid; + break; + } + + int lower_half_step_size = (mid_sharpness - min_sharpness) / 5; + int upper_half_step_size = (max_sharpness - mid_sharpness) / 5; + + // lower half linear approximation + if (discrete_sharpness_level < 5) + sharpness_level = min_sharpness + (lower_half_step_size * discrete_sharpness_level); + // upper half linear approximation + else + sharpness_level = mid_sharpness + (upper_half_step_size * (discrete_sharpness_level - 5)); + + return sharpness_level; +} + +void spl_build_isharp_1dlut_from_reference_curve(struct spl_fixed31_32 ratio, enum system_setup setup, + struct adaptive_sharpness sharpness) +{ + uint8_t *byte_ptr_1dlut_src, *byte_ptr_1dlut_dst; + struct spl_fixed31_32 sharp_base, sharp_calc, sharp_level; + int j; + int size_1dlut; + int sharp_calc_int; + uint32_t filter_pregen_store[ISHARP_LUT_TABLE_SIZE]; + + /* Custom sharpnessX1000 value */ + unsigned int sharpnessX1000 = spl_calculate_sharpness_level(sharpness.sharpness_level, + setup, sharpness.sharpness_range); + sharp_level = spl_fixpt_from_fraction(sharpnessX1000, 1000); + + /* + * Check if pregen 1dlut table is already precalculated + * If numer/denom is different, then recalculate + */ + if ((filter_isharp_1D_lut_pregen[setup].sharpness_numer == sharpnessX1000) && + (filter_isharp_1D_lut_pregen[setup].sharpness_denom == 1000)) + return; + + + /* + * Calculate LUT_128_gained with this equation: + * + * LUT_128_gained[i] = (uint8)(0.5 + min(255,(double)(LUT_128[i])*sharpLevel/iGain)) + * where LUT_128[i] is contents of 3p0x isharp 1dlut + * where sharpLevel is desired sharpness level + * where iGain is base sharpness level 3.0 + * where LUT_128_gained[i] is adjusted 1dlut value based on desired sharpness level + */ + byte_ptr_1dlut_src = (uint8_t *)filter_isharp_1D_lut_3p0x; + byte_ptr_1dlut_dst = (uint8_t *)filter_pregen_store; + size_1dlut = sizeof(filter_isharp_1D_lut_3p0x); + memset(byte_ptr_1dlut_dst, 0, size_1dlut); + for (j = 0; j < size_1dlut; j++) { + sharp_base = spl_fixpt_from_int((int)*byte_ptr_1dlut_src); + sharp_calc = spl_fixpt_mul(sharp_base, sharp_level); + sharp_calc = spl_fixpt_div(sharp_calc, spl_fixpt_from_int(3)); + sharp_calc = spl_fixpt_min(spl_fixpt_from_int(255), sharp_calc); + sharp_calc = spl_fixpt_add(sharp_calc, spl_fixpt_from_fraction(1, 2)); + sharp_calc_int = spl_fixpt_floor(sharp_calc); + /* Clamp it at 0x7F so it doesn't wrap */ + if (sharp_calc_int > 127) + sharp_calc_int = 127; + *byte_ptr_1dlut_dst = (uint8_t)sharp_calc_int; + + byte_ptr_1dlut_src++; + byte_ptr_1dlut_dst++; + } + + /* Update 1dlut table and sharpness level */ + memcpy((void *)filter_isharp_1D_lut_pregen[setup].value, (void *)filter_pregen_store, size_1dlut); + filter_isharp_1D_lut_pregen[setup].sharpness_numer = sharpnessX1000; + filter_isharp_1D_lut_pregen[setup].sharpness_denom = 1000; +} + +uint32_t *spl_get_pregen_filter_isharp_1D_lut(enum system_setup setup) +{ + return filter_isharp_1D_lut_pregen[setup].value; +} + +void spl_init_blur_scale_coeffs(void) +{ + convert_filter_s1_10_to_s1_12(filter_isharp_bs_3tap_64p, + filter_isharp_bs_3tap_64p_s1_12, 3); + convert_filter_s1_10_to_s1_12(filter_isharp_bs_4tap_64p, + filter_isharp_bs_4tap_64p_s1_12, 4); + convert_filter_s1_10_to_s1_12(filter_isharp_bs_4tap_in_6_64p, + filter_isharp_bs_4tap_in_6_64p_s1_12, 6); +} + +uint16_t *spl_dscl_get_blur_scale_coeffs_64p(int taps) +{ + if (taps == 3) + return spl_get_filter_isharp_bs_3tap_64p(); + else if (taps == 4) + return spl_get_filter_isharp_bs_4tap_64p(); + else if (taps == 6) + return spl_get_filter_isharp_bs_4tap_in_6_64p(); + else { + /* should never happen, bug */ + SPL_BREAK_TO_DEBUGGER(); + return NULL; + } +} + +void spl_set_blur_scale_data(struct dscl_prog_data *dscl_prog_data, + const struct spl_scaler_data *data) +{ + dscl_prog_data->filter_blur_scale_h = + spl_dscl_get_blur_scale_coeffs_64p(data->taps.h_taps); + + dscl_prog_data->filter_blur_scale_v = + spl_dscl_get_blur_scale_coeffs_64p(data->taps.v_taps); +} + diff --git a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_isharp_filters.h b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_isharp_filters.h index 1aaf4c50c1bc..fe0b12571f2c 100644 --- a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_isharp_filters.h +++ b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_isharp_filters.h @@ -7,11 +7,44 @@ #include "dc_spl_types.h" +#define ISHARP_LUT_TABLE_SIZE 32 const uint32_t *spl_get_filter_isharp_1D_lut_0(void); const uint32_t *spl_get_filter_isharp_1D_lut_0p5x(void); const uint32_t *spl_get_filter_isharp_1D_lut_1p0x(void); const uint32_t *spl_get_filter_isharp_1D_lut_1p5x(void); const uint32_t *spl_get_filter_isharp_1D_lut_2p0x(void); -const uint16_t *spl_get_filter_isharp_bs_4tap_64p(void); +const uint32_t *spl_get_filter_isharp_1D_lut_3p0x(void); +uint16_t *spl_get_filter_isharp_bs_4tap_in_6_64p(void); +uint16_t *spl_get_filter_isharp_bs_4tap_64p(void); +uint16_t *spl_get_filter_isharp_bs_3tap_64p(void); const uint16_t *spl_get_filter_isharp_wide_6tap_64p(void); +uint16_t *spl_dscl_get_blur_scale_coeffs_64p(int taps); + +struct scale_ratio_to_sharpness_level_lookup { + unsigned int ratio_numer; + unsigned int ratio_denom; + unsigned int sharpness_numer; + unsigned int sharpness_denom; +}; + +struct isharp_1D_lut_pregen { + unsigned int sharpness_numer; + unsigned int sharpness_denom; + uint32_t value[ISHARP_LUT_TABLE_SIZE]; +}; + +enum system_setup { + SDR_NL = 0, + SDR_L, + HDR_NL, + HDR_L, + NUM_SHARPNESS_SETUPS +}; + +void spl_init_blur_scale_coeffs(void); +void spl_set_blur_scale_data(struct dscl_prog_data *dscl_prog_data, + const struct spl_scaler_data *data); + +void spl_build_isharp_1dlut_from_reference_curve(struct spl_fixed31_32 ratio, enum system_setup setup, struct adaptive_sharpness sharpness); +uint32_t *spl_get_pregen_filter_isharp_1D_lut(enum system_setup setup); #endif /* __DC_SPL_ISHARP_FILTERS_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_scl_easf_filters.c b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_scl_easf_filters.c new file mode 100644 index 000000000000..09bf82f7d468 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_scl_easf_filters.c @@ -0,0 +1,1726 @@ +// SPDX-License-Identifier: MIT +// +// Copyright 2024 Advanced Micro Devices, Inc. + +#include "spl_debug.h" +#include "dc_spl_filters.h" +#include "dc_spl_scl_filters.h" +#include "dc_spl_scl_easf_filters.h" + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 03-Apr-2024 +// <coeffDescrip> 3t_64p_LanczosEd_p_0.3_p_10qb_ +// <num_taps> 3 +// <num_phases> 64 +// <scale_ratio> input/output = 0.300000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_3tap_64p_ratio_0_30[99] = { + 0x0200, 0x0200, 0x0000, + 0x01F6, 0x0206, 0x0004, + 0x01EC, 0x020B, 0x0009, + 0x01E2, 0x0211, 0x000D, + 0x01D8, 0x0216, 0x0012, + 0x01CE, 0x021C, 0x0016, + 0x01C4, 0x0221, 0x001B, + 0x01BA, 0x0226, 0x0020, + 0x01B0, 0x022A, 0x0026, + 0x01A6, 0x022F, 0x002B, + 0x019C, 0x0233, 0x0031, + 0x0192, 0x0238, 0x0036, + 0x0188, 0x023C, 0x003C, + 0x017E, 0x0240, 0x0042, + 0x0174, 0x0244, 0x0048, + 0x016A, 0x0248, 0x004E, + 0x0161, 0x024A, 0x0055, + 0x0157, 0x024E, 0x005B, + 0x014D, 0x0251, 0x0062, + 0x0144, 0x0253, 0x0069, + 0x013A, 0x0256, 0x0070, + 0x0131, 0x0258, 0x0077, + 0x0127, 0x025B, 0x007E, + 0x011E, 0x025C, 0x0086, + 0x0115, 0x025E, 0x008D, + 0x010B, 0x0260, 0x0095, + 0x0102, 0x0262, 0x009C, + 0x00F9, 0x0263, 0x00A4, + 0x00F0, 0x0264, 0x00AC, + 0x00E7, 0x0265, 0x00B4, + 0x00DF, 0x0264, 0x00BD, + 0x00D6, 0x0265, 0x00C5, + 0x00CD, 0x0266, 0x00CD, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 03-Apr-2024 +// <coeffDescrip> 3t_64p_LanczosEd_p_0.4_p_10qb_ +// <num_taps> 3 +// <num_phases> 64 +// <scale_ratio> input/output = 0.400000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_3tap_64p_ratio_0_40[99] = { + 0x0200, 0x0200, 0x0000, + 0x01F6, 0x0206, 0x0004, + 0x01EB, 0x020E, 0x0007, + 0x01E1, 0x0214, 0x000B, + 0x01D7, 0x021A, 0x000F, + 0x01CD, 0x0220, 0x0013, + 0x01C2, 0x0226, 0x0018, + 0x01B8, 0x022C, 0x001C, + 0x01AE, 0x0231, 0x0021, + 0x01A3, 0x0237, 0x0026, + 0x0199, 0x023C, 0x002B, + 0x018F, 0x0240, 0x0031, + 0x0185, 0x0245, 0x0036, + 0x017A, 0x024A, 0x003C, + 0x0170, 0x024F, 0x0041, + 0x0166, 0x0253, 0x0047, + 0x015C, 0x0257, 0x004D, + 0x0152, 0x025A, 0x0054, + 0x0148, 0x025E, 0x005A, + 0x013E, 0x0261, 0x0061, + 0x0134, 0x0264, 0x0068, + 0x012B, 0x0266, 0x006F, + 0x0121, 0x0269, 0x0076, + 0x0117, 0x026C, 0x007D, + 0x010E, 0x026E, 0x0084, + 0x0104, 0x0270, 0x008C, + 0x00FB, 0x0271, 0x0094, + 0x00F2, 0x0272, 0x009C, + 0x00E9, 0x0273, 0x00A4, + 0x00E0, 0x0274, 0x00AC, + 0x00D7, 0x0275, 0x00B4, + 0x00CE, 0x0275, 0x00BD, + 0x00C5, 0x0276, 0x00C5, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 03-Apr-2024 +// <coeffDescrip> 3t_64p_LanczosEd_p_0.5_p_10qb_ +// <num_taps> 3 +// <num_phases> 64 +// <scale_ratio> input/output = 0.500000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_3tap_64p_ratio_0_50[99] = { + 0x0200, 0x0200, 0x0000, + 0x01F5, 0x0209, 0x0002, + 0x01EA, 0x0211, 0x0005, + 0x01DF, 0x021A, 0x0007, + 0x01D4, 0x0222, 0x000A, + 0x01C9, 0x022A, 0x000D, + 0x01BE, 0x0232, 0x0010, + 0x01B3, 0x0239, 0x0014, + 0x01A8, 0x0241, 0x0017, + 0x019D, 0x0248, 0x001B, + 0x0192, 0x024F, 0x001F, + 0x0187, 0x0255, 0x0024, + 0x017C, 0x025C, 0x0028, + 0x0171, 0x0262, 0x002D, + 0x0166, 0x0268, 0x0032, + 0x015B, 0x026E, 0x0037, + 0x0150, 0x0273, 0x003D, + 0x0146, 0x0278, 0x0042, + 0x013B, 0x027D, 0x0048, + 0x0130, 0x0282, 0x004E, + 0x0126, 0x0286, 0x0054, + 0x011B, 0x028A, 0x005B, + 0x0111, 0x028D, 0x0062, + 0x0107, 0x0290, 0x0069, + 0x00FD, 0x0293, 0x0070, + 0x00F3, 0x0296, 0x0077, + 0x00E9, 0x0298, 0x007F, + 0x00DF, 0x029A, 0x0087, + 0x00D5, 0x029C, 0x008F, + 0x00CC, 0x029D, 0x0097, + 0x00C3, 0x029E, 0x009F, + 0x00BA, 0x029E, 0x00A8, + 0x00B1, 0x029E, 0x00B1, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 03-Apr-2024 +// <coeffDescrip> 3t_64p_LanczosEd_p_0.6_p_10qb_ +// <num_taps> 3 +// <num_phases> 64 +// <scale_ratio> input/output = 0.600000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_3tap_64p_ratio_0_60[99] = { + 0x0200, 0x0200, 0x0000, + 0x01F4, 0x020B, 0x0001, + 0x01E8, 0x0216, 0x0002, + 0x01DC, 0x0221, 0x0003, + 0x01D0, 0x022B, 0x0005, + 0x01C4, 0x0235, 0x0007, + 0x01B8, 0x0240, 0x0008, + 0x01AC, 0x0249, 0x000B, + 0x01A0, 0x0253, 0x000D, + 0x0194, 0x025C, 0x0010, + 0x0188, 0x0265, 0x0013, + 0x017C, 0x026E, 0x0016, + 0x0170, 0x0277, 0x0019, + 0x0164, 0x027F, 0x001D, + 0x0158, 0x0287, 0x0021, + 0x014C, 0x028F, 0x0025, + 0x0140, 0x0297, 0x0029, + 0x0135, 0x029D, 0x002E, + 0x0129, 0x02A4, 0x0033, + 0x011D, 0x02AB, 0x0038, + 0x0112, 0x02B0, 0x003E, + 0x0107, 0x02B5, 0x0044, + 0x00FC, 0x02BA, 0x004A, + 0x00F1, 0x02BF, 0x0050, + 0x00E6, 0x02C3, 0x0057, + 0x00DB, 0x02C7, 0x005E, + 0x00D1, 0x02CA, 0x0065, + 0x00C7, 0x02CC, 0x006D, + 0x00BD, 0x02CE, 0x0075, + 0x00B3, 0x02D0, 0x007D, + 0x00A9, 0x02D2, 0x0085, + 0x00A0, 0x02D2, 0x008E, + 0x0097, 0x02D2, 0x0097, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 03-Apr-2024 +// <coeffDescrip> 3t_64p_LanczosEd_p_0.7_p_10qb_ +// <num_taps> 3 +// <num_phases> 64 +// <scale_ratio> input/output = 0.700000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_3tap_64p_ratio_0_70[99] = { + 0x0200, 0x0200, 0x0000, + 0x01F3, 0x020D, 0x0000, + 0x01E5, 0x021B, 0x0000, + 0x01D8, 0x0228, 0x0000, + 0x01CB, 0x0235, 0x0000, + 0x01BD, 0x0243, 0x0000, + 0x01B0, 0x024F, 0x0001, + 0x01A2, 0x025C, 0x0002, + 0x0195, 0x0268, 0x0003, + 0x0187, 0x0275, 0x0004, + 0x017A, 0x0280, 0x0006, + 0x016D, 0x028C, 0x0007, + 0x015F, 0x0298, 0x0009, + 0x0152, 0x02A2, 0x000C, + 0x0145, 0x02AD, 0x000E, + 0x0138, 0x02B7, 0x0011, + 0x012B, 0x02C0, 0x0015, + 0x011E, 0x02CA, 0x0018, + 0x0111, 0x02D3, 0x001C, + 0x0105, 0x02DB, 0x0020, + 0x00F8, 0x02E3, 0x0025, + 0x00EC, 0x02EA, 0x002A, + 0x00E0, 0x02F1, 0x002F, + 0x00D5, 0x02F6, 0x0035, + 0x00C9, 0x02FC, 0x003B, + 0x00BE, 0x0301, 0x0041, + 0x00B3, 0x0305, 0x0048, + 0x00A8, 0x0309, 0x004F, + 0x009E, 0x030C, 0x0056, + 0x0094, 0x030E, 0x005E, + 0x008A, 0x0310, 0x0066, + 0x0081, 0x0310, 0x006F, + 0x0077, 0x0312, 0x0077, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 03-Apr-2024 +// <coeffDescrip> 3t_64p_LanczosEd_p_0.8_p_10qb_ +// <num_taps> 3 +// <num_phases> 64 +// <scale_ratio> input/output = 0.800000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_3tap_64p_ratio_0_80[99] = { + 0x0200, 0x0200, 0x0000, + 0x01F1, 0x0210, 0x0FFF, + 0x01E2, 0x0220, 0x0FFE, + 0x01D2, 0x0232, 0x0FFC, + 0x01C3, 0x0241, 0x0FFC, + 0x01B4, 0x0251, 0x0FFB, + 0x01A4, 0x0262, 0x0FFA, + 0x0195, 0x0271, 0x0FFA, + 0x0186, 0x0281, 0x0FF9, + 0x0176, 0x0291, 0x0FF9, + 0x0167, 0x02A0, 0x0FF9, + 0x0158, 0x02AE, 0x0FFA, + 0x0149, 0x02BD, 0x0FFA, + 0x013A, 0x02CB, 0x0FFB, + 0x012C, 0x02D7, 0x0FFD, + 0x011D, 0x02E5, 0x0FFE, + 0x010F, 0x02F1, 0x0000, + 0x0101, 0x02FD, 0x0002, + 0x00F3, 0x0308, 0x0005, + 0x00E5, 0x0313, 0x0008, + 0x00D8, 0x031D, 0x000B, + 0x00CB, 0x0326, 0x000F, + 0x00BE, 0x032F, 0x0013, + 0x00B2, 0x0337, 0x0017, + 0x00A6, 0x033E, 0x001C, + 0x009A, 0x0345, 0x0021, + 0x008F, 0x034A, 0x0027, + 0x0084, 0x034F, 0x002D, + 0x0079, 0x0353, 0x0034, + 0x006F, 0x0356, 0x003B, + 0x0065, 0x0358, 0x0043, + 0x005C, 0x0359, 0x004B, + 0x0053, 0x035A, 0x0053, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 03-Apr-2024 +// <coeffDescrip> 3t_64p_LanczosEd_p_0.9_p_10qb_ +// <num_taps> 3 +// <num_phases> 64 +// <scale_ratio> input/output = 0.900000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_3tap_64p_ratio_0_90[99] = { + 0x0200, 0x0200, 0x0000, + 0x01EE, 0x0214, 0x0FFE, + 0x01DC, 0x0228, 0x0FFC, + 0x01CA, 0x023C, 0x0FFA, + 0x01B9, 0x024F, 0x0FF8, + 0x01A7, 0x0262, 0x0FF7, + 0x0195, 0x0276, 0x0FF5, + 0x0183, 0x028A, 0x0FF3, + 0x0172, 0x029C, 0x0FF2, + 0x0160, 0x02AF, 0x0FF1, + 0x014F, 0x02C2, 0x0FEF, + 0x013E, 0x02D4, 0x0FEE, + 0x012D, 0x02E5, 0x0FEE, + 0x011C, 0x02F7, 0x0FED, + 0x010C, 0x0307, 0x0FED, + 0x00FB, 0x0318, 0x0FED, + 0x00EC, 0x0327, 0x0FED, + 0x00DC, 0x0336, 0x0FEE, + 0x00CD, 0x0344, 0x0FEF, + 0x00BE, 0x0352, 0x0FF0, + 0x00B0, 0x035E, 0x0FF2, + 0x00A2, 0x036A, 0x0FF4, + 0x0095, 0x0375, 0x0FF6, + 0x0088, 0x037F, 0x0FF9, + 0x007B, 0x0388, 0x0FFD, + 0x006F, 0x0391, 0x0000, + 0x0064, 0x0397, 0x0005, + 0x0059, 0x039D, 0x000A, + 0x004E, 0x03A3, 0x000F, + 0x0045, 0x03A6, 0x0015, + 0x003B, 0x03A9, 0x001C, + 0x0033, 0x03AA, 0x0023, + 0x002A, 0x03AC, 0x002A, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 03-Apr-2024 +// <coeffDescrip> 3t_64p_LanczosEd_p_1_p_10qb_ +// <num_taps> 3 +// <num_phases> 64 +// <scale_ratio> input/output = 1.000000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_3tap_64p_ratio_1_00[99] = { + 0x0200, 0x0200, 0x0000, + 0x01EB, 0x0217, 0x0FFE, + 0x01D5, 0x022F, 0x0FFC, + 0x01C0, 0x0247, 0x0FF9, + 0x01AB, 0x025E, 0x0FF7, + 0x0196, 0x0276, 0x0FF4, + 0x0181, 0x028D, 0x0FF2, + 0x016C, 0x02A5, 0x0FEF, + 0x0158, 0x02BB, 0x0FED, + 0x0144, 0x02D1, 0x0FEB, + 0x0130, 0x02E8, 0x0FE8, + 0x011C, 0x02FE, 0x0FE6, + 0x0109, 0x0313, 0x0FE4, + 0x00F6, 0x0328, 0x0FE2, + 0x00E4, 0x033C, 0x0FE0, + 0x00D2, 0x034F, 0x0FDF, + 0x00C0, 0x0363, 0x0FDD, + 0x00B0, 0x0374, 0x0FDC, + 0x009F, 0x0385, 0x0FDC, + 0x0090, 0x0395, 0x0FDB, + 0x0081, 0x03A4, 0x0FDB, + 0x0072, 0x03B3, 0x0FDB, + 0x0064, 0x03C0, 0x0FDC, + 0x0057, 0x03CC, 0x0FDD, + 0x004B, 0x03D6, 0x0FDF, + 0x003F, 0x03E0, 0x0FE1, + 0x0034, 0x03E8, 0x0FE4, + 0x002A, 0x03EF, 0x0FE7, + 0x0020, 0x03F5, 0x0FEB, + 0x0017, 0x03FA, 0x0FEF, + 0x000F, 0x03FD, 0x0FF4, + 0x0007, 0x03FF, 0x0FFA, + 0x0000, 0x0400, 0x0000, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 03-Apr-2024 +// <coeffDescrip> 4t_64p_LanczosEd_p_0.3_p_10qb_ +// <num_taps> 4 +// <num_phases> 64 +// <scale_ratio> input/output = 0.300000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_4tap_64p_ratio_0_30[132] = { + 0x0104, 0x01F8, 0x0104, 0x0000, + 0x00FE, 0x01F7, 0x010A, 0x0001, + 0x00F8, 0x01F6, 0x010F, 0x0003, + 0x00F2, 0x01F5, 0x0114, 0x0005, + 0x00EB, 0x01F4, 0x011B, 0x0006, + 0x00E5, 0x01F3, 0x0120, 0x0008, + 0x00DF, 0x01F2, 0x0125, 0x000A, + 0x00DA, 0x01F0, 0x012A, 0x000C, + 0x00D4, 0x01EE, 0x0130, 0x000E, + 0x00CE, 0x01ED, 0x0135, 0x0010, + 0x00C8, 0x01EB, 0x013A, 0x0013, + 0x00C2, 0x01E9, 0x0140, 0x0015, + 0x00BD, 0x01E7, 0x0145, 0x0017, + 0x00B7, 0x01E5, 0x014A, 0x001A, + 0x00B1, 0x01E2, 0x0151, 0x001C, + 0x00AC, 0x01E0, 0x0155, 0x001F, + 0x00A7, 0x01DD, 0x015A, 0x0022, + 0x00A1, 0x01DB, 0x015F, 0x0025, + 0x009C, 0x01D8, 0x0165, 0x0027, + 0x0097, 0x01D5, 0x016A, 0x002A, + 0x0092, 0x01D2, 0x016E, 0x002E, + 0x008C, 0x01CF, 0x0174, 0x0031, + 0x0087, 0x01CC, 0x0179, 0x0034, + 0x0083, 0x01C9, 0x017D, 0x0037, + 0x007E, 0x01C5, 0x0182, 0x003B, + 0x0079, 0x01C2, 0x0187, 0x003E, + 0x0074, 0x01BE, 0x018C, 0x0042, + 0x0070, 0x01BA, 0x0190, 0x0046, + 0x006B, 0x01B7, 0x0195, 0x0049, + 0x0066, 0x01B3, 0x019A, 0x004D, + 0x0062, 0x01AF, 0x019E, 0x0051, + 0x005E, 0x01AB, 0x01A2, 0x0055, + 0x005A, 0x01A6, 0x01A6, 0x005A, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 03-Apr-2024 +// <coeffDescrip> 4t_64p_LanczosEd_p_0.4_p_10qb_ +// <num_taps> 4 +// <num_phases> 64 +// <scale_ratio> input/output = 0.400000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_4tap_64p_ratio_0_40[132] = { + 0x00FB, 0x0209, 0x00FC, 0x0000, + 0x00F5, 0x0209, 0x0101, 0x0001, + 0x00EE, 0x0208, 0x0108, 0x0002, + 0x00E8, 0x0207, 0x010E, 0x0003, + 0x00E2, 0x0206, 0x0114, 0x0004, + 0x00DB, 0x0205, 0x011A, 0x0006, + 0x00D5, 0x0204, 0x0120, 0x0007, + 0x00CF, 0x0203, 0x0125, 0x0009, + 0x00C9, 0x0201, 0x012C, 0x000A, + 0x00C3, 0x01FF, 0x0132, 0x000C, + 0x00BD, 0x01FD, 0x0138, 0x000E, + 0x00B7, 0x01FB, 0x013E, 0x0010, + 0x00B1, 0x01F9, 0x0144, 0x0012, + 0x00AC, 0x01F7, 0x0149, 0x0014, + 0x00A6, 0x01F4, 0x0150, 0x0016, + 0x00A0, 0x01F2, 0x0156, 0x0018, + 0x009B, 0x01EF, 0x015C, 0x001A, + 0x0095, 0x01EC, 0x0162, 0x001D, + 0x0090, 0x01E9, 0x0168, 0x001F, + 0x008B, 0x01E6, 0x016D, 0x0022, + 0x0085, 0x01E3, 0x0173, 0x0025, + 0x0080, 0x01DF, 0x0179, 0x0028, + 0x007B, 0x01DC, 0x017E, 0x002B, + 0x0076, 0x01D8, 0x0184, 0x002E, + 0x0071, 0x01D4, 0x018A, 0x0031, + 0x006D, 0x01D1, 0x018E, 0x0034, + 0x0068, 0x01CD, 0x0193, 0x0038, + 0x0063, 0x01C8, 0x019A, 0x003B, + 0x005F, 0x01C4, 0x019E, 0x003F, + 0x005B, 0x01C0, 0x01A3, 0x0042, + 0x0056, 0x01BB, 0x01A9, 0x0046, + 0x0052, 0x01B7, 0x01AD, 0x004A, + 0x004E, 0x01B2, 0x01B2, 0x004E, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 03-Apr-2024 +// <coeffDescrip> 4t_64p_LanczosEd_p_0.5_p_10qb_ +// <num_taps> 4 +// <num_phases> 64 +// <scale_ratio> input/output = 0.500000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_4tap_64p_ratio_0_50[132] = { + 0x00E5, 0x0236, 0x00E5, 0x0000, + 0x00DE, 0x0235, 0x00ED, 0x0000, + 0x00D7, 0x0235, 0x00F4, 0x0000, + 0x00D0, 0x0235, 0x00FB, 0x0000, + 0x00C9, 0x0234, 0x0102, 0x0001, + 0x00C2, 0x0233, 0x010A, 0x0001, + 0x00BC, 0x0232, 0x0111, 0x0001, + 0x00B5, 0x0230, 0x0119, 0x0002, + 0x00AE, 0x022F, 0x0121, 0x0002, + 0x00A8, 0x022D, 0x0128, 0x0003, + 0x00A2, 0x022B, 0x012F, 0x0004, + 0x009B, 0x0229, 0x0137, 0x0005, + 0x0095, 0x0226, 0x013F, 0x0006, + 0x008F, 0x0224, 0x0146, 0x0007, + 0x0089, 0x0221, 0x014E, 0x0008, + 0x0083, 0x021E, 0x0155, 0x000A, + 0x007E, 0x021B, 0x015C, 0x000B, + 0x0078, 0x0217, 0x0164, 0x000D, + 0x0072, 0x0213, 0x016D, 0x000E, + 0x006D, 0x0210, 0x0173, 0x0010, + 0x0068, 0x020C, 0x017A, 0x0012, + 0x0063, 0x0207, 0x0182, 0x0014, + 0x005E, 0x0203, 0x0189, 0x0016, + 0x0059, 0x01FE, 0x0191, 0x0018, + 0x0054, 0x01F9, 0x0198, 0x001B, + 0x0050, 0x01F4, 0x019F, 0x001D, + 0x004B, 0x01EF, 0x01A6, 0x0020, + 0x0047, 0x01EA, 0x01AC, 0x0023, + 0x0043, 0x01E4, 0x01B3, 0x0026, + 0x003F, 0x01DF, 0x01B9, 0x0029, + 0x003B, 0x01D9, 0x01C0, 0x002C, + 0x0037, 0x01D3, 0x01C6, 0x0030, + 0x0033, 0x01CD, 0x01CD, 0x0033, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 03-Apr-2024 +// <coeffDescrip> 4t_64p_LanczosEd_p_0.6_p_10qb_ +// <num_taps> 4 +// <num_phases> 64 +// <scale_ratio> input/output = 0.600000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_4tap_64p_ratio_0_60[132] = { + 0x00C8, 0x026F, 0x00C9, 0x0000, + 0x00C0, 0x0270, 0x00D1, 0x0FFF, + 0x00B8, 0x0270, 0x00D9, 0x0FFF, + 0x00B1, 0x0270, 0x00E1, 0x0FFE, + 0x00A9, 0x026F, 0x00EB, 0x0FFD, + 0x00A2, 0x026E, 0x00F3, 0x0FFD, + 0x009A, 0x026D, 0x00FD, 0x0FFC, + 0x0093, 0x026C, 0x0105, 0x0FFC, + 0x008C, 0x026A, 0x010F, 0x0FFB, + 0x0085, 0x0268, 0x0118, 0x0FFB, + 0x007E, 0x0265, 0x0122, 0x0FFB, + 0x0078, 0x0263, 0x012A, 0x0FFB, + 0x0071, 0x0260, 0x0134, 0x0FFB, + 0x006B, 0x025C, 0x013E, 0x0FFB, + 0x0065, 0x0259, 0x0147, 0x0FFB, + 0x005F, 0x0255, 0x0151, 0x0FFB, + 0x0059, 0x0251, 0x015A, 0x0FFC, + 0x0054, 0x024D, 0x0163, 0x0FFC, + 0x004E, 0x0248, 0x016D, 0x0FFD, + 0x0049, 0x0243, 0x0176, 0x0FFE, + 0x0044, 0x023E, 0x017F, 0x0FFF, + 0x003F, 0x0238, 0x0189, 0x0000, + 0x003A, 0x0232, 0x0193, 0x0001, + 0x0036, 0x022C, 0x019C, 0x0002, + 0x0031, 0x0226, 0x01A5, 0x0004, + 0x002D, 0x021F, 0x01AF, 0x0005, + 0x0029, 0x0218, 0x01B8, 0x0007, + 0x0025, 0x0211, 0x01C1, 0x0009, + 0x0022, 0x020A, 0x01C9, 0x000B, + 0x001E, 0x0203, 0x01D2, 0x000D, + 0x001B, 0x01FB, 0x01DA, 0x0010, + 0x0018, 0x01F3, 0x01E3, 0x0012, + 0x0015, 0x01EB, 0x01EB, 0x0015, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 03-Apr-2024 +// <coeffDescrip> 4t_64p_LanczosEd_p_0.7_p_10qb_ +// <num_taps> 4 +// <num_phases> 64 +// <scale_ratio> input/output = 0.700000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_4tap_64p_ratio_0_70[132] = { + 0x00A3, 0x02B9, 0x00A4, 0x0000, + 0x009A, 0x02BA, 0x00AD, 0x0FFF, + 0x0092, 0x02BA, 0x00B6, 0x0FFE, + 0x0089, 0x02BA, 0x00C1, 0x0FFC, + 0x0081, 0x02B9, 0x00CB, 0x0FFB, + 0x0079, 0x02B8, 0x00D5, 0x0FFA, + 0x0071, 0x02B7, 0x00DF, 0x0FF9, + 0x0069, 0x02B5, 0x00EA, 0x0FF8, + 0x0062, 0x02B3, 0x00F4, 0x0FF7, + 0x005B, 0x02B0, 0x00FF, 0x0FF6, + 0x0054, 0x02AD, 0x010B, 0x0FF4, + 0x004D, 0x02A9, 0x0117, 0x0FF3, + 0x0046, 0x02A5, 0x0123, 0x0FF2, + 0x0040, 0x02A1, 0x012D, 0x0FF2, + 0x003A, 0x029C, 0x0139, 0x0FF1, + 0x0034, 0x0297, 0x0145, 0x0FF0, + 0x002F, 0x0292, 0x0150, 0x0FEF, + 0x0029, 0x028C, 0x015C, 0x0FEF, + 0x0024, 0x0285, 0x0169, 0x0FEE, + 0x001F, 0x027F, 0x0174, 0x0FEE, + 0x001B, 0x0278, 0x017F, 0x0FEE, + 0x0016, 0x0270, 0x018D, 0x0FED, + 0x0012, 0x0268, 0x0199, 0x0FED, + 0x000E, 0x0260, 0x01A4, 0x0FEE, + 0x000B, 0x0258, 0x01AF, 0x0FEE, + 0x0007, 0x024F, 0x01BC, 0x0FEE, + 0x0004, 0x0246, 0x01C7, 0x0FEF, + 0x0001, 0x023D, 0x01D3, 0x0FEF, + 0x0FFE, 0x0233, 0x01DF, 0x0FF0, + 0x0FFC, 0x0229, 0x01EA, 0x0FF1, + 0x0FFA, 0x021F, 0x01F4, 0x0FF3, + 0x0FF8, 0x0215, 0x01FF, 0x0FF4, + 0x0FF6, 0x020A, 0x020A, 0x0FF6, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 03-Apr-2024 +// <coeffDescrip> 4t_64p_LanczosEd_p_0.8_p_10qb_ +// <num_taps> 4 +// <num_phases> 64 +// <scale_ratio> input/output = 0.800000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_4tap_64p_ratio_0_80[132] = { + 0x0075, 0x0315, 0x0076, 0x0000, + 0x006C, 0x0316, 0x007F, 0x0FFF, + 0x0062, 0x0316, 0x008A, 0x0FFE, + 0x0059, 0x0315, 0x0096, 0x0FFC, + 0x0050, 0x0314, 0x00A1, 0x0FFB, + 0x0048, 0x0312, 0x00AD, 0x0FF9, + 0x0040, 0x0310, 0x00B8, 0x0FF8, + 0x0038, 0x030D, 0x00C5, 0x0FF6, + 0x0030, 0x030A, 0x00D1, 0x0FF5, + 0x0029, 0x0306, 0x00DE, 0x0FF3, + 0x0022, 0x0301, 0x00EB, 0x0FF2, + 0x001C, 0x02FC, 0x00F8, 0x0FF0, + 0x0015, 0x02F7, 0x0106, 0x0FEE, + 0x0010, 0x02F1, 0x0112, 0x0FED, + 0x000A, 0x02EA, 0x0121, 0x0FEB, + 0x0005, 0x02E3, 0x012F, 0x0FE9, + 0x0000, 0x02DB, 0x013D, 0x0FE8, + 0x0FFB, 0x02D3, 0x014C, 0x0FE6, + 0x0FF7, 0x02CA, 0x015A, 0x0FE5, + 0x0FF3, 0x02C1, 0x0169, 0x0FE3, + 0x0FF0, 0x02B7, 0x0177, 0x0FE2, + 0x0FEC, 0x02AD, 0x0186, 0x0FE1, + 0x0FE9, 0x02A2, 0x0196, 0x0FDF, + 0x0FE7, 0x0297, 0x01A4, 0x0FDE, + 0x0FE4, 0x028C, 0x01B3, 0x0FDD, + 0x0FE2, 0x0280, 0x01C2, 0x0FDC, + 0x0FE0, 0x0274, 0x01D0, 0x0FDC, + 0x0FDF, 0x0268, 0x01DE, 0x0FDB, + 0x0FDD, 0x025B, 0x01EE, 0x0FDA, + 0x0FDC, 0x024E, 0x01FC, 0x0FDA, + 0x0FDB, 0x0241, 0x020A, 0x0FDA, + 0x0FDB, 0x0233, 0x0218, 0x0FDA, + 0x0FDA, 0x0226, 0x0226, 0x0FDA, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 03-Apr-2024 +// <coeffDescrip> 4t_64p_LanczosEd_p_0.9_p_10qb_ +// <num_taps> 4 +// <num_phases> 64 +// <scale_ratio> input/output = 0.900000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_4tap_64p_ratio_0_90[132] = { + 0x003F, 0x0383, 0x003E, 0x0000, + 0x0034, 0x0383, 0x004A, 0x0FFF, + 0x002B, 0x0383, 0x0054, 0x0FFE, + 0x0021, 0x0381, 0x0061, 0x0FFD, + 0x0019, 0x037F, 0x006C, 0x0FFC, + 0x0010, 0x037C, 0x0079, 0x0FFB, + 0x0008, 0x0378, 0x0086, 0x0FFA, + 0x0001, 0x0374, 0x0093, 0x0FF8, + 0x0FFA, 0x036E, 0x00A1, 0x0FF7, + 0x0FF3, 0x0368, 0x00B0, 0x0FF5, + 0x0FED, 0x0361, 0x00BF, 0x0FF3, + 0x0FE8, 0x035A, 0x00CD, 0x0FF1, + 0x0FE2, 0x0352, 0x00DC, 0x0FF0, + 0x0FDE, 0x0349, 0x00EB, 0x0FEE, + 0x0FD9, 0x033F, 0x00FC, 0x0FEC, + 0x0FD5, 0x0335, 0x010D, 0x0FE9, + 0x0FD2, 0x032A, 0x011D, 0x0FE7, + 0x0FCF, 0x031E, 0x012E, 0x0FE5, + 0x0FCC, 0x0312, 0x013F, 0x0FE3, + 0x0FCA, 0x0305, 0x0150, 0x0FE1, + 0x0FC8, 0x02F8, 0x0162, 0x0FDE, + 0x0FC6, 0x02EA, 0x0174, 0x0FDC, + 0x0FC5, 0x02DC, 0x0185, 0x0FDA, + 0x0FC4, 0x02CD, 0x0197, 0x0FD8, + 0x0FC3, 0x02BE, 0x01AA, 0x0FD5, + 0x0FC3, 0x02AF, 0x01BB, 0x0FD3, + 0x0FC3, 0x029F, 0x01CD, 0x0FD1, + 0x0FC3, 0x028E, 0x01E0, 0x0FCF, + 0x0FC3, 0x027E, 0x01F2, 0x0FCD, + 0x0FC4, 0x026D, 0x0203, 0x0FCC, + 0x0FC5, 0x025C, 0x0215, 0x0FCA, + 0x0FC6, 0x024B, 0x0227, 0x0FC8, + 0x0FC7, 0x0239, 0x0239, 0x0FC7, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 03-Apr-2024 +// <coeffDescrip> 4t_64p_LanczosEd_p_1_p_10qb_ +// <num_taps> 4 +// <num_phases> 64 +// <scale_ratio> input/output = 1.000000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_4tap_64p_ratio_1_00[132] = { + 0x0000, 0x0400, 0x0000, 0x0000, + 0x0FF6, 0x03FF, 0x000B, 0x0000, + 0x0FED, 0x03FE, 0x0015, 0x0000, + 0x0FE4, 0x03FB, 0x0022, 0x0FFF, + 0x0FDC, 0x03F7, 0x002E, 0x0FFF, + 0x0FD5, 0x03F2, 0x003B, 0x0FFE, + 0x0FCE, 0x03EC, 0x0048, 0x0FFE, + 0x0FC8, 0x03E5, 0x0056, 0x0FFD, + 0x0FC3, 0x03DC, 0x0065, 0x0FFC, + 0x0FBE, 0x03D3, 0x0075, 0x0FFA, + 0x0FB9, 0x03C9, 0x0085, 0x0FF9, + 0x0FB6, 0x03BE, 0x0094, 0x0FF8, + 0x0FB2, 0x03B2, 0x00A6, 0x0FF6, + 0x0FB0, 0x03A5, 0x00B7, 0x0FF4, + 0x0FAD, 0x0397, 0x00CA, 0x0FF2, + 0x0FAB, 0x0389, 0x00DC, 0x0FF0, + 0x0FAA, 0x0379, 0x00EF, 0x0FEE, + 0x0FA9, 0x0369, 0x0102, 0x0FEC, + 0x0FA9, 0x0359, 0x0115, 0x0FE9, + 0x0FA9, 0x0348, 0x0129, 0x0FE6, + 0x0FA9, 0x0336, 0x013D, 0x0FE4, + 0x0FA9, 0x0323, 0x0153, 0x0FE1, + 0x0FAA, 0x0310, 0x0168, 0x0FDE, + 0x0FAC, 0x02FD, 0x017C, 0x0FDB, + 0x0FAD, 0x02E9, 0x0192, 0x0FD8, + 0x0FAF, 0x02D5, 0x01A7, 0x0FD5, + 0x0FB1, 0x02C0, 0x01BD, 0x0FD2, + 0x0FB3, 0x02AC, 0x01D2, 0x0FCF, + 0x0FB5, 0x0296, 0x01E9, 0x0FCC, + 0x0FB8, 0x0281, 0x01FE, 0x0FC9, + 0x0FBA, 0x026C, 0x0214, 0x0FC6, + 0x0FBD, 0x0256, 0x022A, 0x0FC3, + 0x0FC0, 0x0240, 0x0240, 0x0FC0, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 02-Apr-2024 +// <coeffDescrip> 6t_64p_LanczosEd_p_0.3_p_10qb_ +// <num_taps> 6 +// <num_phases> 64 +// <scale_ratio> input/output = 0.300000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_6tap_64p_ratio_0_30[198] = { + 0x004B, 0x0100, 0x0169, 0x0101, 0x004B, 0x0000, + 0x0049, 0x00FD, 0x0169, 0x0103, 0x004E, 0x0000, + 0x0047, 0x00FA, 0x0169, 0x0106, 0x0050, 0x0000, + 0x0045, 0x00F7, 0x0168, 0x0109, 0x0052, 0x0001, + 0x0043, 0x00F5, 0x0168, 0x010B, 0x0054, 0x0001, + 0x0040, 0x00F2, 0x0168, 0x010E, 0x0057, 0x0001, + 0x003E, 0x00EF, 0x0168, 0x0110, 0x0059, 0x0002, + 0x003C, 0x00EC, 0x0167, 0x0113, 0x005C, 0x0002, + 0x003A, 0x00E9, 0x0167, 0x0116, 0x005E, 0x0002, + 0x0038, 0x00E6, 0x0166, 0x0118, 0x0061, 0x0003, + 0x0036, 0x00E3, 0x0165, 0x011C, 0x0063, 0x0003, + 0x0034, 0x00E0, 0x0165, 0x011D, 0x0066, 0x0004, + 0x0033, 0x00DD, 0x0164, 0x0120, 0x0068, 0x0004, + 0x0031, 0x00DA, 0x0163, 0x0122, 0x006B, 0x0005, + 0x002F, 0x00D7, 0x0163, 0x0125, 0x006D, 0x0005, + 0x002D, 0x00D3, 0x0162, 0x0128, 0x0070, 0x0006, + 0x002B, 0x00D0, 0x0161, 0x012A, 0x0073, 0x0007, + 0x002A, 0x00CD, 0x0160, 0x012D, 0x0075, 0x0007, + 0x0028, 0x00CA, 0x015F, 0x012F, 0x0078, 0x0008, + 0x0026, 0x00C7, 0x015E, 0x0131, 0x007B, 0x0009, + 0x0025, 0x00C4, 0x015D, 0x0133, 0x007E, 0x0009, + 0x0023, 0x00C1, 0x015C, 0x0136, 0x0080, 0x000A, + 0x0022, 0x00BE, 0x015A, 0x0138, 0x0083, 0x000B, + 0x0020, 0x00BB, 0x0159, 0x013A, 0x0086, 0x000C, + 0x001F, 0x00B8, 0x0158, 0x013B, 0x0089, 0x000D, + 0x001E, 0x00B5, 0x0156, 0x013E, 0x008C, 0x000D, + 0x001C, 0x00B2, 0x0155, 0x0140, 0x008F, 0x000E, + 0x001B, 0x00AF, 0x0153, 0x0143, 0x0091, 0x000F, + 0x0019, 0x00AC, 0x0152, 0x0145, 0x0094, 0x0010, + 0x0018, 0x00A9, 0x0150, 0x0147, 0x0097, 0x0011, + 0x0017, 0x00A6, 0x014F, 0x0148, 0x009A, 0x0012, + 0x0016, 0x00A3, 0x014D, 0x0149, 0x009D, 0x0014, + 0x0015, 0x00A0, 0x014B, 0x014B, 0x00A0, 0x0015, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 02-Apr-2024 +// <coeffDescrip> 6t_64p_LanczosEd_p_0.4_p_10qb_ +// <num_taps> 6 +// <num_phases> 64 +// <scale_ratio> input/output = 0.400000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_6tap_64p_ratio_0_40[198] = { + 0x0028, 0x0106, 0x01A3, 0x0107, 0x0028, 0x0000, + 0x0026, 0x0102, 0x01A3, 0x010A, 0x002B, 0x0000, + 0x0024, 0x00FE, 0x01A3, 0x010F, 0x002D, 0x0FFF, + 0x0022, 0x00FA, 0x01A3, 0x0113, 0x002F, 0x0FFF, + 0x0021, 0x00F6, 0x01A3, 0x0116, 0x0031, 0x0FFF, + 0x001F, 0x00F2, 0x01A2, 0x011B, 0x0034, 0x0FFE, + 0x001D, 0x00EE, 0x01A2, 0x011F, 0x0036, 0x0FFE, + 0x001B, 0x00EA, 0x01A1, 0x0123, 0x0039, 0x0FFE, + 0x0019, 0x00E6, 0x01A1, 0x0127, 0x003B, 0x0FFE, + 0x0018, 0x00E2, 0x01A0, 0x012A, 0x003E, 0x0FFE, + 0x0016, 0x00DE, 0x01A0, 0x012E, 0x0041, 0x0FFD, + 0x0015, 0x00DA, 0x019F, 0x0132, 0x0043, 0x0FFD, + 0x0013, 0x00D6, 0x019E, 0x0136, 0x0046, 0x0FFD, + 0x0012, 0x00D2, 0x019D, 0x0139, 0x0049, 0x0FFD, + 0x0010, 0x00CE, 0x019C, 0x013D, 0x004C, 0x0FFD, + 0x000F, 0x00CA, 0x019A, 0x0141, 0x004F, 0x0FFD, + 0x000E, 0x00C6, 0x0199, 0x0144, 0x0052, 0x0FFD, + 0x000D, 0x00C2, 0x0197, 0x0148, 0x0055, 0x0FFD, + 0x000B, 0x00BE, 0x0196, 0x014C, 0x0058, 0x0FFD, + 0x000A, 0x00BA, 0x0195, 0x014F, 0x005B, 0x0FFD, + 0x0009, 0x00B6, 0x0193, 0x0153, 0x005E, 0x0FFD, + 0x0008, 0x00B2, 0x0191, 0x0157, 0x0061, 0x0FFD, + 0x0007, 0x00AE, 0x0190, 0x015A, 0x0064, 0x0FFD, + 0x0006, 0x00AA, 0x018E, 0x015D, 0x0068, 0x0FFD, + 0x0005, 0x00A6, 0x018C, 0x0161, 0x006B, 0x0FFD, + 0x0005, 0x00A2, 0x0189, 0x0164, 0x006F, 0x0FFD, + 0x0004, 0x009E, 0x0187, 0x0167, 0x0072, 0x0FFE, + 0x0003, 0x009A, 0x0185, 0x016B, 0x0075, 0x0FFE, + 0x0002, 0x0096, 0x0183, 0x016E, 0x0079, 0x0FFE, + 0x0002, 0x0093, 0x0180, 0x016F, 0x007D, 0x0FFF, + 0x0001, 0x008F, 0x017E, 0x0173, 0x0080, 0x0FFF, + 0x0001, 0x008B, 0x017B, 0x0175, 0x0084, 0x0000, + 0x0000, 0x0087, 0x0179, 0x0179, 0x0087, 0x0000, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 02-Apr-2024 +// <coeffDescrip> 6t_64p_LanczosEd_p_0.5_p_10qb_ +// <num_taps> 6 +// <num_phases> 64 +// <scale_ratio> input/output = 0.500000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_6tap_64p_ratio_0_50[198] = { + 0x0000, 0x0107, 0x01F3, 0x0106, 0x0000, 0x0000, + 0x0FFE, 0x0101, 0x01F3, 0x010D, 0x0002, 0x0FFF, + 0x0FFD, 0x00FB, 0x01F3, 0x0113, 0x0003, 0x0FFF, + 0x0FFC, 0x00F6, 0x01F3, 0x0118, 0x0005, 0x0FFE, + 0x0FFA, 0x00F0, 0x01F3, 0x011E, 0x0007, 0x0FFE, + 0x0FF9, 0x00EB, 0x01F2, 0x0124, 0x0009, 0x0FFD, + 0x0FF8, 0x00E5, 0x01F2, 0x0129, 0x000B, 0x0FFD, + 0x0FF7, 0x00E0, 0x01F1, 0x012F, 0x000D, 0x0FFC, + 0x0FF6, 0x00DA, 0x01F0, 0x0135, 0x0010, 0x0FFB, + 0x0FF5, 0x00D4, 0x01EF, 0x013B, 0x0012, 0x0FFB, + 0x0FF4, 0x00CF, 0x01EE, 0x0141, 0x0014, 0x0FFA, + 0x0FF3, 0x00C9, 0x01ED, 0x0147, 0x0017, 0x0FF9, + 0x0FF2, 0x00C4, 0x01EB, 0x014C, 0x001A, 0x0FF9, + 0x0FF1, 0x00BF, 0x01EA, 0x0152, 0x001C, 0x0FF8, + 0x0FF1, 0x00B9, 0x01E8, 0x0157, 0x001F, 0x0FF8, + 0x0FF0, 0x00B4, 0x01E6, 0x015D, 0x0022, 0x0FF7, + 0x0FF0, 0x00AE, 0x01E4, 0x0163, 0x0025, 0x0FF6, + 0x0FEF, 0x00A9, 0x01E2, 0x0168, 0x0028, 0x0FF6, + 0x0FEF, 0x00A4, 0x01DF, 0x016E, 0x002B, 0x0FF5, + 0x0FEF, 0x009F, 0x01DD, 0x0172, 0x002E, 0x0FF5, + 0x0FEE, 0x009A, 0x01DA, 0x0178, 0x0032, 0x0FF4, + 0x0FEE, 0x0094, 0x01D8, 0x017E, 0x0035, 0x0FF3, + 0x0FEE, 0x008F, 0x01D5, 0x0182, 0x0039, 0x0FF3, + 0x0FEE, 0x008A, 0x01D2, 0x0188, 0x003C, 0x0FF2, + 0x0FEE, 0x0085, 0x01CF, 0x018C, 0x0040, 0x0FF2, + 0x0FEE, 0x0081, 0x01CB, 0x0191, 0x0044, 0x0FF1, + 0x0FEE, 0x007C, 0x01C8, 0x0196, 0x0047, 0x0FF1, + 0x0FEE, 0x0077, 0x01C4, 0x019C, 0x004B, 0x0FF0, + 0x0FEE, 0x0072, 0x01C1, 0x01A0, 0x004F, 0x0FF0, + 0x0FEE, 0x006E, 0x01BD, 0x01A4, 0x0053, 0x0FF0, + 0x0FEE, 0x0069, 0x01B9, 0x01A9, 0x0058, 0x0FEF, + 0x0FEE, 0x0065, 0x01B5, 0x01AD, 0x005C, 0x0FEF, + 0x0FEF, 0x0060, 0x01B1, 0x01B1, 0x0060, 0x0FEF, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 02-Apr-2024 +// <coeffDescrip> 6t_64p_LanczosEd_p_0.6_p_10qb_ +// <num_taps> 6 +// <num_phases> 64 +// <scale_ratio> input/output = 0.600000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_6tap_64p_ratio_0_60[198] = { + 0x0FD9, 0x00FB, 0x0258, 0x00FB, 0x0FD9, 0x0000, + 0x0FD9, 0x00F3, 0x0258, 0x0102, 0x0FDA, 0x0000, + 0x0FD8, 0x00EB, 0x0258, 0x010B, 0x0FDB, 0x0FFF, + 0x0FD8, 0x00E3, 0x0258, 0x0112, 0x0FDC, 0x0FFF, + 0x0FD8, 0x00DC, 0x0257, 0x011B, 0x0FDC, 0x0FFE, + 0x0FD7, 0x00D4, 0x0256, 0x0123, 0x0FDE, 0x0FFE, + 0x0FD7, 0x00CD, 0x0255, 0x012B, 0x0FDF, 0x0FFD, + 0x0FD7, 0x00C5, 0x0254, 0x0133, 0x0FE0, 0x0FFD, + 0x0FD7, 0x00BE, 0x0252, 0x013C, 0x0FE1, 0x0FFC, + 0x0FD7, 0x00B6, 0x0251, 0x0143, 0x0FE3, 0x0FFC, + 0x0FD8, 0x00AF, 0x024F, 0x014B, 0x0FE4, 0x0FFB, + 0x0FD8, 0x00A8, 0x024C, 0x0154, 0x0FE6, 0x0FFA, + 0x0FD8, 0x00A1, 0x024A, 0x015B, 0x0FE8, 0x0FFA, + 0x0FD9, 0x009A, 0x0247, 0x0163, 0x0FEA, 0x0FF9, + 0x0FD9, 0x0093, 0x0244, 0x016C, 0x0FEC, 0x0FF8, + 0x0FD9, 0x008C, 0x0241, 0x0174, 0x0FEF, 0x0FF7, + 0x0FDA, 0x0085, 0x023E, 0x017B, 0x0FF1, 0x0FF7, + 0x0FDB, 0x007F, 0x023A, 0x0183, 0x0FF3, 0x0FF6, + 0x0FDB, 0x0078, 0x0237, 0x018B, 0x0FF6, 0x0FF5, + 0x0FDC, 0x0072, 0x0233, 0x0192, 0x0FF9, 0x0FF4, + 0x0FDD, 0x006C, 0x022F, 0x0199, 0x0FFC, 0x0FF3, + 0x0FDD, 0x0065, 0x022A, 0x01A3, 0x0FFF, 0x0FF2, + 0x0FDE, 0x005F, 0x0226, 0x01AA, 0x0002, 0x0FF1, + 0x0FDF, 0x005A, 0x0221, 0x01B0, 0x0006, 0x0FF0, + 0x0FE0, 0x0054, 0x021C, 0x01B7, 0x0009, 0x0FF0, + 0x0FE1, 0x004E, 0x0217, 0x01BE, 0x000D, 0x0FEF, + 0x0FE2, 0x0048, 0x0212, 0x01C6, 0x0010, 0x0FEE, + 0x0FE3, 0x0043, 0x020C, 0x01CD, 0x0014, 0x0FED, + 0x0FE4, 0x003E, 0x0207, 0x01D3, 0x0018, 0x0FEC, + 0x0FE5, 0x0039, 0x0200, 0x01DA, 0x001D, 0x0FEB, + 0x0FE6, 0x0034, 0x01FA, 0x01E1, 0x0021, 0x0FEA, + 0x0FE7, 0x002F, 0x01F5, 0x01E7, 0x0025, 0x0FE9, + 0x0FE8, 0x002A, 0x01EE, 0x01EE, 0x002A, 0x0FE8, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 02-Apr-2024 +// <coeffDescrip> 6t_64p_LanczosEd_p_0.7_p_10qb_ +// <num_taps> 6 +// <num_phases> 64 +// <scale_ratio> input/output = 0.700000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_6tap_64p_ratio_0_70[198] = { + 0x0FC0, 0x00DA, 0x02CC, 0x00DA, 0x0FC0, 0x0000, + 0x0FC1, 0x00D0, 0x02CC, 0x00E4, 0x0FBF, 0x0000, + 0x0FC2, 0x00C6, 0x02CB, 0x00EF, 0x0FBE, 0x0000, + 0x0FC3, 0x00BC, 0x02CA, 0x00F9, 0x0FBE, 0x0000, + 0x0FC4, 0x00B2, 0x02C9, 0x0104, 0x0FBD, 0x0000, + 0x0FC5, 0x00A8, 0x02C7, 0x010F, 0x0FBD, 0x0000, + 0x0FC7, 0x009F, 0x02C5, 0x0119, 0x0FBC, 0x0000, + 0x0FC8, 0x0095, 0x02C3, 0x0124, 0x0FBC, 0x0000, + 0x0FC9, 0x008C, 0x02C0, 0x012F, 0x0FBC, 0x0000, + 0x0FCB, 0x0083, 0x02BD, 0x0139, 0x0FBC, 0x0000, + 0x0FCC, 0x007A, 0x02BA, 0x0144, 0x0FBC, 0x0000, + 0x0FCE, 0x0072, 0x02B6, 0x014D, 0x0FBD, 0x0000, + 0x0FD0, 0x0069, 0x02B2, 0x0159, 0x0FBD, 0x0FFF, + 0x0FD1, 0x0061, 0x02AD, 0x0164, 0x0FBE, 0x0FFF, + 0x0FD3, 0x0059, 0x02A9, 0x016E, 0x0FBF, 0x0FFE, + 0x0FD4, 0x0051, 0x02A4, 0x017A, 0x0FBF, 0x0FFE, + 0x0FD6, 0x0049, 0x029E, 0x0184, 0x0FC1, 0x0FFE, + 0x0FD8, 0x0042, 0x0299, 0x018E, 0x0FC2, 0x0FFD, + 0x0FD9, 0x003A, 0x0293, 0x019B, 0x0FC3, 0x0FFC, + 0x0FDB, 0x0033, 0x028D, 0x01A4, 0x0FC5, 0x0FFC, + 0x0FDC, 0x002D, 0x0286, 0x01AF, 0x0FC7, 0x0FFB, + 0x0FDE, 0x0026, 0x0280, 0x01BA, 0x0FC8, 0x0FFA, + 0x0FE0, 0x001F, 0x0279, 0x01C4, 0x0FCB, 0x0FF9, + 0x0FE1, 0x0019, 0x0272, 0x01CE, 0x0FCD, 0x0FF9, + 0x0FE3, 0x0013, 0x026A, 0x01D9, 0x0FCF, 0x0FF8, + 0x0FE4, 0x000D, 0x0263, 0x01E3, 0x0FD2, 0x0FF7, + 0x0FE6, 0x0008, 0x025B, 0x01EC, 0x0FD5, 0x0FF6, + 0x0FE7, 0x0002, 0x0253, 0x01F7, 0x0FD8, 0x0FF5, + 0x0FE9, 0x0FFD, 0x024A, 0x0202, 0x0FDB, 0x0FF3, + 0x0FEA, 0x0FF8, 0x0242, 0x020B, 0x0FDF, 0x0FF2, + 0x0FEC, 0x0FF3, 0x0239, 0x0215, 0x0FE2, 0x0FF1, + 0x0FED, 0x0FEF, 0x0230, 0x021E, 0x0FE6, 0x0FF0, + 0x0FEF, 0x0FEB, 0x0226, 0x0226, 0x0FEB, 0x0FEF, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 02-Apr-2024 +// <coeffDescrip> 6t_64p_LanczosEd_p_0.8_p_10qb_ +// <num_taps> 6 +// <num_phases> 64 +// <scale_ratio> input/output = 0.800000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_6tap_64p_ratio_0_80[198] = { + 0x0FBF, 0x00A1, 0x0340, 0x00A1, 0x0FBF, 0x0000, + 0x0FC1, 0x0095, 0x0340, 0x00AD, 0x0FBC, 0x0001, + 0x0FC4, 0x0089, 0x033E, 0x00BA, 0x0FBA, 0x0001, + 0x0FC6, 0x007D, 0x033D, 0x00C6, 0x0FB8, 0x0002, + 0x0FC9, 0x0072, 0x033A, 0x00D3, 0x0FB6, 0x0002, + 0x0FCC, 0x0067, 0x0338, 0x00DF, 0x0FB3, 0x0003, + 0x0FCE, 0x005C, 0x0334, 0x00EE, 0x0FB1, 0x0003, + 0x0FD1, 0x0051, 0x0331, 0x00FA, 0x0FAF, 0x0004, + 0x0FD3, 0x0047, 0x032D, 0x0108, 0x0FAD, 0x0004, + 0x0FD6, 0x003D, 0x0328, 0x0116, 0x0FAB, 0x0004, + 0x0FD8, 0x0033, 0x0323, 0x0123, 0x0FAA, 0x0005, + 0x0FDB, 0x002A, 0x031D, 0x0131, 0x0FA8, 0x0005, + 0x0FDD, 0x0021, 0x0317, 0x013F, 0x0FA7, 0x0005, + 0x0FDF, 0x0018, 0x0311, 0x014D, 0x0FA5, 0x0006, + 0x0FE2, 0x0010, 0x030A, 0x015A, 0x0FA4, 0x0006, + 0x0FE4, 0x0008, 0x0302, 0x0169, 0x0FA3, 0x0006, + 0x0FE6, 0x0000, 0x02FB, 0x0177, 0x0FA2, 0x0006, + 0x0FE8, 0x0FF9, 0x02F3, 0x0185, 0x0FA1, 0x0006, + 0x0FEB, 0x0FF1, 0x02EA, 0x0193, 0x0FA1, 0x0006, + 0x0FED, 0x0FEB, 0x02E1, 0x01A1, 0x0FA0, 0x0006, + 0x0FEE, 0x0FE4, 0x02D8, 0x01B0, 0x0FA0, 0x0006, + 0x0FF0, 0x0FDE, 0x02CE, 0x01BE, 0x0FA0, 0x0006, + 0x0FF2, 0x0FD8, 0x02C5, 0x01CB, 0x0FA0, 0x0006, + 0x0FF4, 0x0FD3, 0x02BA, 0x01D8, 0x0FA1, 0x0006, + 0x0FF6, 0x0FCD, 0x02B0, 0x01E7, 0x0FA1, 0x0005, + 0x0FF7, 0x0FC8, 0x02A5, 0x01F5, 0x0FA2, 0x0005, + 0x0FF9, 0x0FC4, 0x029A, 0x0202, 0x0FA3, 0x0004, + 0x0FFA, 0x0FC0, 0x028E, 0x0210, 0x0FA4, 0x0004, + 0x0FFB, 0x0FBC, 0x0283, 0x021D, 0x0FA6, 0x0003, + 0x0FFD, 0x0FB8, 0x0276, 0x022A, 0x0FA8, 0x0003, + 0x0FFE, 0x0FB4, 0x026B, 0x0237, 0x0FAA, 0x0002, + 0x0FFF, 0x0FB1, 0x025E, 0x0245, 0x0FAC, 0x0001, + 0x0000, 0x0FAE, 0x0252, 0x0252, 0x0FAE, 0x0000, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 02-Apr-2024 +// <coeffDescrip> 6t_64p_LanczosEd_p_0.9_p_10qb_ +// <num_taps> 6 +// <num_phases> 64 +// <scale_ratio> input/output = 0.900000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_6tap_64p_ratio_0_90[198] = { + 0x0FD8, 0x0055, 0x03A7, 0x0054, 0x0FD8, 0x0000, + 0x0FDB, 0x0047, 0x03A7, 0x0063, 0x0FD4, 0x0000, + 0x0FDF, 0x003B, 0x03A5, 0x006F, 0x0FD1, 0x0001, + 0x0FE2, 0x002E, 0x03A3, 0x007E, 0x0FCD, 0x0002, + 0x0FE5, 0x0022, 0x03A0, 0x008D, 0x0FCA, 0x0002, + 0x0FE8, 0x0017, 0x039D, 0x009B, 0x0FC6, 0x0003, + 0x0FEB, 0x000C, 0x0398, 0x00AC, 0x0FC2, 0x0003, + 0x0FEE, 0x0001, 0x0394, 0x00BA, 0x0FBF, 0x0004, + 0x0FF1, 0x0FF7, 0x038E, 0x00CA, 0x0FBB, 0x0005, + 0x0FF4, 0x0FED, 0x0388, 0x00DA, 0x0FB8, 0x0005, + 0x0FF6, 0x0FE4, 0x0381, 0x00EB, 0x0FB4, 0x0006, + 0x0FF9, 0x0FDB, 0x037A, 0x00FA, 0x0FB1, 0x0007, + 0x0FFB, 0x0FD3, 0x0372, 0x010B, 0x0FAD, 0x0008, + 0x0FFD, 0x0FCB, 0x0369, 0x011D, 0x0FAA, 0x0008, + 0x0000, 0x0FC3, 0x0360, 0x012E, 0x0FA6, 0x0009, + 0x0002, 0x0FBC, 0x0356, 0x013F, 0x0FA3, 0x000A, + 0x0003, 0x0FB6, 0x034C, 0x0150, 0x0FA0, 0x000B, + 0x0005, 0x0FB0, 0x0341, 0x0162, 0x0F9D, 0x000B, + 0x0007, 0x0FAA, 0x0336, 0x0173, 0x0F9A, 0x000C, + 0x0008, 0x0FA5, 0x032A, 0x0185, 0x0F97, 0x000D, + 0x000A, 0x0FA0, 0x031E, 0x0197, 0x0F94, 0x000D, + 0x000B, 0x0F9B, 0x0311, 0x01A9, 0x0F92, 0x000E, + 0x000C, 0x0F97, 0x0303, 0x01BC, 0x0F8F, 0x000F, + 0x000D, 0x0F94, 0x02F6, 0x01CD, 0x0F8D, 0x000F, + 0x000E, 0x0F91, 0x02E8, 0x01DE, 0x0F8B, 0x0010, + 0x000F, 0x0F8E, 0x02D9, 0x01F1, 0x0F89, 0x0010, + 0x0010, 0x0F8B, 0x02CA, 0x0202, 0x0F88, 0x0011, + 0x0010, 0x0F89, 0x02BB, 0x0214, 0x0F87, 0x0011, + 0x0011, 0x0F87, 0x02AB, 0x0226, 0x0F86, 0x0011, + 0x0011, 0x0F86, 0x029C, 0x0236, 0x0F85, 0x0012, + 0x0011, 0x0F85, 0x028B, 0x0249, 0x0F84, 0x0012, + 0x0012, 0x0F84, 0x027B, 0x0259, 0x0F84, 0x0012, + 0x0012, 0x0F84, 0x026A, 0x026A, 0x0F84, 0x0012, +}; + +//======================================================== +// <using> gen_scaler_coeffs_cnf_file.m +// <using> make_test_script.m +// <date> 02-Apr-2024 +// <coeffDescrip> 6t_64p_LanczosEd_p_1_p_10qb_ +// <num_taps> 6 +// <num_phases> 64 +// <scale_ratio> input/output = 1.000000000000 +// <CoefType> LanczosEd +// <CoefQuant> S1.10 +//======================================================== +static const uint16_t easf_filter_6tap_64p_ratio_1_00[198] = { + 0x0000, 0x0000, 0x0400, 0x0000, 0x0000, 0x0000, + 0x0003, 0x0FF3, 0x0400, 0x000D, 0x0FFD, 0x0000, + 0x0006, 0x0FE7, 0x03FE, 0x001C, 0x0FF9, 0x0000, + 0x0009, 0x0FDB, 0x03FC, 0x002B, 0x0FF5, 0x0000, + 0x000C, 0x0FD0, 0x03F9, 0x003A, 0x0FF1, 0x0000, + 0x000E, 0x0FC5, 0x03F5, 0x004A, 0x0FED, 0x0001, + 0x0011, 0x0FBB, 0x03F0, 0x005A, 0x0FE9, 0x0001, + 0x0013, 0x0FB2, 0x03EB, 0x006A, 0x0FE5, 0x0001, + 0x0015, 0x0FA9, 0x03E4, 0x007B, 0x0FE1, 0x0002, + 0x0017, 0x0FA1, 0x03DD, 0x008D, 0x0FDC, 0x0002, + 0x0018, 0x0F99, 0x03D4, 0x00A0, 0x0FD8, 0x0003, + 0x001A, 0x0F92, 0x03CB, 0x00B2, 0x0FD3, 0x0004, + 0x001B, 0x0F8C, 0x03C1, 0x00C6, 0x0FCE, 0x0004, + 0x001C, 0x0F86, 0x03B7, 0x00D9, 0x0FC9, 0x0005, + 0x001D, 0x0F80, 0x03AB, 0x00EE, 0x0FC4, 0x0006, + 0x001E, 0x0F7C, 0x039F, 0x0101, 0x0FBF, 0x0007, + 0x001F, 0x0F78, 0x0392, 0x0115, 0x0FBA, 0x0008, + 0x001F, 0x0F74, 0x0385, 0x012B, 0x0FB5, 0x0008, + 0x0020, 0x0F71, 0x0376, 0x0140, 0x0FB0, 0x0009, + 0x0020, 0x0F6E, 0x0367, 0x0155, 0x0FAB, 0x000B, + 0x0020, 0x0F6C, 0x0357, 0x016B, 0x0FA6, 0x000C, + 0x0020, 0x0F6A, 0x0347, 0x0180, 0x0FA2, 0x000D, + 0x0020, 0x0F69, 0x0336, 0x0196, 0x0F9D, 0x000E, + 0x0020, 0x0F69, 0x0325, 0x01AB, 0x0F98, 0x000F, + 0x001F, 0x0F68, 0x0313, 0x01C3, 0x0F93, 0x0010, + 0x001F, 0x0F69, 0x0300, 0x01D8, 0x0F8F, 0x0011, + 0x001E, 0x0F69, 0x02ED, 0x01EF, 0x0F8B, 0x0012, + 0x001D, 0x0F6A, 0x02D9, 0x0205, 0x0F87, 0x0014, + 0x001D, 0x0F6C, 0x02C5, 0x021A, 0x0F83, 0x0015, + 0x001C, 0x0F6E, 0x02B1, 0x0230, 0x0F7F, 0x0016, + 0x001B, 0x0F70, 0x029C, 0x0247, 0x0F7B, 0x0017, + 0x001A, 0x0F72, 0x0287, 0x025D, 0x0F78, 0x0018, + 0x0019, 0x0F75, 0x0272, 0x0272, 0x0F75, 0x0019, +}; + +/* Converted scaler coeff tables from S1.10 to S1.12 */ +static uint16_t easf_filter_3tap_64p_ratio_0_30_s1_12[99]; +static uint16_t easf_filter_3tap_64p_ratio_0_40_s1_12[99]; +static uint16_t easf_filter_3tap_64p_ratio_0_50_s1_12[99]; +static uint16_t easf_filter_3tap_64p_ratio_0_60_s1_12[99]; +static uint16_t easf_filter_3tap_64p_ratio_0_70_s1_12[99]; +static uint16_t easf_filter_3tap_64p_ratio_0_80_s1_12[99]; +static uint16_t easf_filter_3tap_64p_ratio_0_90_s1_12[99]; +static uint16_t easf_filter_3tap_64p_ratio_1_00_s1_12[99]; +static uint16_t easf_filter_4tap_64p_ratio_0_30_s1_12[132]; +static uint16_t easf_filter_4tap_64p_ratio_0_40_s1_12[132]; +static uint16_t easf_filter_4tap_64p_ratio_0_50_s1_12[132]; +static uint16_t easf_filter_4tap_64p_ratio_0_60_s1_12[132]; +static uint16_t easf_filter_4tap_64p_ratio_0_70_s1_12[132]; +static uint16_t easf_filter_4tap_64p_ratio_0_80_s1_12[132]; +static uint16_t easf_filter_4tap_64p_ratio_0_90_s1_12[132]; +static uint16_t easf_filter_4tap_64p_ratio_1_00_s1_12[132]; +static uint16_t easf_filter_6tap_64p_ratio_0_30_s1_12[198]; +static uint16_t easf_filter_6tap_64p_ratio_0_40_s1_12[198]; +static uint16_t easf_filter_6tap_64p_ratio_0_50_s1_12[198]; +static uint16_t easf_filter_6tap_64p_ratio_0_60_s1_12[198]; +static uint16_t easf_filter_6tap_64p_ratio_0_70_s1_12[198]; +static uint16_t easf_filter_6tap_64p_ratio_0_80_s1_12[198]; +static uint16_t easf_filter_6tap_64p_ratio_0_90_s1_12[198]; +static uint16_t easf_filter_6tap_64p_ratio_1_00_s1_12[198]; + +struct scale_ratio_to_reg_value_lookup easf_v_bf3_mode_lookup[] = { + {3, 10, 0x0000}, + {4, 10, 0x0000}, + {5, 10, 0x0000}, + {6, 10, 0x0000}, + {7, 10, 0x0000}, + {8, 10, 0x0000}, + {9, 10, 0x0000}, + {1, 1, 0x0000}, + {-1, -1, 0x0002}, +}; + +struct scale_ratio_to_reg_value_lookup easf_h_bf3_mode_lookup[] = { + {3, 10, 0x0000}, + {4, 10, 0x0000}, + {5, 10, 0x0000}, + {6, 10, 0x0000}, + {7, 10, 0x0000}, + {8, 10, 0x0000}, + {9, 10, 0x0000}, + {1, 1, 0x0000}, + {-1, -1, 0x0002}, +}; + +struct scale_ratio_to_reg_value_lookup easf_reducer_gain6_6tap_lookup[] = { + {3, 10, 0x4100}, + {4, 10, 0x4100}, + {5, 10, 0x4100}, + {6, 10, 0x4100}, + {7, 10, 0x4100}, + {8, 10, 0x4100}, + {9, 10, 0x4100}, + {1, 1, 0x4100}, + {-1, -1, 0x4100}, +}; + +struct scale_ratio_to_reg_value_lookup easf_reducer_gain4_6tap_lookup[] = { + {3, 10, 0x4000}, + {4, 10, 0x4000}, + {5, 10, 0x4000}, + {6, 10, 0x4000}, + {7, 10, 0x4000}, + {8, 10, 0x4000}, + {9, 10, 0x4000}, + {1, 1, 0x4000}, + {-1, -1, 0x4000}, +}; + +struct scale_ratio_to_reg_value_lookup easf_gain_ring6_6tap_lookup[] = { + {3, 10, 0x0000}, + {4, 10, 0x251F}, + {5, 10, 0x291F}, + {6, 10, 0xA51F}, + {7, 10, 0xA51F}, + {8, 10, 0xAA66}, + {9, 10, 0xA51F}, + {1, 1, 0xA640}, + {-1, -1, 0xA640}, +}; + +struct scale_ratio_to_reg_value_lookup easf_gain_ring4_6tap_lookup[] = { + {3, 10, 0x0000}, + {4, 10, 0x9600}, + {5, 10, 0xA460}, + {6, 10, 0xA8E0}, + {7, 10, 0xAC00}, + {8, 10, 0xAD20}, + {9, 10, 0xAFC0}, + {1, 1, 0xB058}, + {-1, -1, 0xB058}, +}; + +struct scale_ratio_to_reg_value_lookup easf_reducer_gain6_4tap_lookup[] = { + {3, 10, 0x4100}, + {4, 10, 0x4100}, + {5, 10, 0x4100}, + {6, 10, 0x4100}, + {7, 10, 0x4100}, + {8, 10, 0x4100}, + {9, 10, 0x4100}, + {1, 1, 0x4100}, + {-1, -1, 0x4100}, +}; + +struct scale_ratio_to_reg_value_lookup easf_reducer_gain4_4tap_lookup[] = { + {3, 10, 0x4000}, + {4, 10, 0x4000}, + {5, 10, 0x4000}, + {6, 10, 0x4000}, + {7, 10, 0x4000}, + {8, 10, 0x4000}, + {9, 10, 0x4000}, + {1, 1, 0x4000}, + {-1, -1, 0x4000}, +}; + +struct scale_ratio_to_reg_value_lookup easf_gain_ring6_4tap_lookup[] = { + {3, 10, 0x0000}, + {4, 10, 0x0000}, + {5, 10, 0x0000}, + {6, 10, 0x0000}, + {7, 10, 0x0000}, + {8, 10, 0x0000}, + {9, 10, 0x0000}, + {1, 1, 0x0000}, + {-1, -1, 0x0000}, +}; + +struct scale_ratio_to_reg_value_lookup easf_gain_ring4_4tap_lookup[] = { + {3, 10, 0x0000}, + {4, 10, 0x0000}, + {5, 10, 0x0000}, + {6, 10, 0x9900}, + {7, 10, 0xA100}, + {8, 10, 0xA8C0}, + {9, 10, 0xAB20}, + {1, 1, 0xAC00}, + {-1, -1, 0xAC00}, +}; + +struct scale_ratio_to_reg_value_lookup easf_3tap_dntilt_uptilt_offset_lookup[] = { + {3, 10, 0x0000}, + {4, 10, 0x0000}, + {5, 10, 0x0000}, + {6, 10, 0x0000}, + {7, 10, 0x0000}, + {8, 10, 0x4100}, + {9, 10, 0x9F00}, + {1, 1, 0xA4C0}, + {-1, -1, 0xA8D8}, +}; + +struct scale_ratio_to_reg_value_lookup easf_3tap_uptilt_maxval_lookup[] = { + {3, 10, 0x0000}, + {4, 10, 0x0000}, + {5, 10, 0x0000}, + {6, 10, 0x0000}, + {7, 10, 0x0000}, + {8, 10, 0x4000}, + {9, 10, 0x24FE}, + {1, 1, 0x2D64}, + {-1, -1, 0x3ADB}, +}; + +struct scale_ratio_to_reg_value_lookup easf_3tap_dntilt_slope_lookup[] = { + {3, 10, 0x3800}, + {4, 10, 0x3800}, + {5, 10, 0x3800}, + {6, 10, 0x3800}, + {7, 10, 0x3800}, + {8, 10, 0x3886}, + {9, 10, 0x3940}, + {1, 1, 0x3A4E}, + {-1, -1, 0x3B66}, +}; + +struct scale_ratio_to_reg_value_lookup easf_3tap_uptilt1_slope_lookup[] = { + {3, 10, 0x3800}, + {4, 10, 0x3800}, + {5, 10, 0x3800}, + {6, 10, 0x3800}, + {7, 10, 0x3800}, + {8, 10, 0x36F4}, + {9, 10, 0x359C}, + {1, 1, 0x3360}, + {-1, -1, 0x2F20}, +}; + +struct scale_ratio_to_reg_value_lookup easf_3tap_uptilt2_slope_lookup[] = { + {3, 10, 0x0000}, + {4, 10, 0x0000}, + {5, 10, 0x0000}, + {6, 10, 0x0000}, + {7, 10, 0x0000}, + {8, 10, 0x0000}, + {9, 10, 0x359C}, + {1, 1, 0x31F0}, + {-1, -1, 0x1F00}, +}; + +struct scale_ratio_to_reg_value_lookup easf_3tap_uptilt2_offset_lookup[] = { + {3, 10, 0x0000}, + {4, 10, 0x0000}, + {5, 10, 0x0000}, + {6, 10, 0x0000}, + {7, 10, 0x0000}, + {8, 10, 0x0000}, + {9, 10, 0x9F00}, + {1, 1, 0xA400}, + {-1, -1, 0x9E00}, +}; + +void spl_init_easf_filter_coeffs(void) +{ + convert_filter_s1_10_to_s1_12(easf_filter_3tap_64p_ratio_0_30, + easf_filter_3tap_64p_ratio_0_30_s1_12, 3); + convert_filter_s1_10_to_s1_12(easf_filter_3tap_64p_ratio_0_40, + easf_filter_3tap_64p_ratio_0_40_s1_12, 3); + convert_filter_s1_10_to_s1_12(easf_filter_3tap_64p_ratio_0_50, + easf_filter_3tap_64p_ratio_0_50_s1_12, 3); + convert_filter_s1_10_to_s1_12(easf_filter_3tap_64p_ratio_0_60, + easf_filter_3tap_64p_ratio_0_60_s1_12, 3); + convert_filter_s1_10_to_s1_12(easf_filter_3tap_64p_ratio_0_70, + easf_filter_3tap_64p_ratio_0_70_s1_12, 3); + convert_filter_s1_10_to_s1_12(easf_filter_3tap_64p_ratio_0_80, + easf_filter_3tap_64p_ratio_0_80_s1_12, 3); + convert_filter_s1_10_to_s1_12(easf_filter_3tap_64p_ratio_0_90, + easf_filter_3tap_64p_ratio_0_90_s1_12, 3); + convert_filter_s1_10_to_s1_12(easf_filter_3tap_64p_ratio_1_00, + easf_filter_3tap_64p_ratio_1_00_s1_12, 3); + + convert_filter_s1_10_to_s1_12(easf_filter_4tap_64p_ratio_0_30, + easf_filter_4tap_64p_ratio_0_30_s1_12, 4); + convert_filter_s1_10_to_s1_12(easf_filter_4tap_64p_ratio_0_40, + easf_filter_4tap_64p_ratio_0_40_s1_12, 4); + convert_filter_s1_10_to_s1_12(easf_filter_4tap_64p_ratio_0_50, + easf_filter_4tap_64p_ratio_0_50_s1_12, 4); + convert_filter_s1_10_to_s1_12(easf_filter_4tap_64p_ratio_0_60, + easf_filter_4tap_64p_ratio_0_60_s1_12, 4); + convert_filter_s1_10_to_s1_12(easf_filter_4tap_64p_ratio_0_70, + easf_filter_4tap_64p_ratio_0_70_s1_12, 4); + convert_filter_s1_10_to_s1_12(easf_filter_4tap_64p_ratio_0_80, + easf_filter_4tap_64p_ratio_0_80_s1_12, 4); + convert_filter_s1_10_to_s1_12(easf_filter_4tap_64p_ratio_0_90, + easf_filter_4tap_64p_ratio_0_90_s1_12, 4); + convert_filter_s1_10_to_s1_12(easf_filter_4tap_64p_ratio_1_00, + easf_filter_4tap_64p_ratio_1_00_s1_12, 4); + + convert_filter_s1_10_to_s1_12(easf_filter_6tap_64p_ratio_0_30, + easf_filter_6tap_64p_ratio_0_30_s1_12, 6); + convert_filter_s1_10_to_s1_12(easf_filter_6tap_64p_ratio_0_40, + easf_filter_6tap_64p_ratio_0_40_s1_12, 6); + convert_filter_s1_10_to_s1_12(easf_filter_6tap_64p_ratio_0_50, + easf_filter_6tap_64p_ratio_0_50_s1_12, 6); + convert_filter_s1_10_to_s1_12(easf_filter_6tap_64p_ratio_0_60, + easf_filter_6tap_64p_ratio_0_60_s1_12, 6); + convert_filter_s1_10_to_s1_12(easf_filter_6tap_64p_ratio_0_70, + easf_filter_6tap_64p_ratio_0_70_s1_12, 6); + convert_filter_s1_10_to_s1_12(easf_filter_6tap_64p_ratio_0_80, + easf_filter_6tap_64p_ratio_0_80_s1_12, 6); + convert_filter_s1_10_to_s1_12(easf_filter_6tap_64p_ratio_0_90, + easf_filter_6tap_64p_ratio_0_90_s1_12, 6); + convert_filter_s1_10_to_s1_12(easf_filter_6tap_64p_ratio_1_00, + easf_filter_6tap_64p_ratio_1_00_s1_12, 6); +} + +uint16_t *spl_get_easf_filter_3tap_64p(struct spl_fixed31_32 ratio) +{ + if (ratio.value < spl_fixpt_from_fraction(3, 10).value) + return easf_filter_3tap_64p_ratio_0_30_s1_12; + else if (ratio.value < spl_fixpt_from_fraction(4, 10).value) + return easf_filter_3tap_64p_ratio_0_40_s1_12; + else if (ratio.value < spl_fixpt_from_fraction(5, 10).value) + return easf_filter_3tap_64p_ratio_0_50_s1_12; + else if (ratio.value < spl_fixpt_from_fraction(6, 10).value) + return easf_filter_3tap_64p_ratio_0_60_s1_12; + else if (ratio.value < spl_fixpt_from_fraction(7, 10).value) + return easf_filter_3tap_64p_ratio_0_70_s1_12; + else if (ratio.value < spl_fixpt_from_fraction(8, 10).value) + return easf_filter_3tap_64p_ratio_0_80_s1_12; + else if (ratio.value < spl_fixpt_from_fraction(9, 10).value) + return easf_filter_3tap_64p_ratio_0_90_s1_12; + else + return easf_filter_3tap_64p_ratio_1_00_s1_12; +} + +uint16_t *spl_get_easf_filter_4tap_64p(struct spl_fixed31_32 ratio) +{ + if (ratio.value < spl_fixpt_from_fraction(3, 10).value) + return easf_filter_4tap_64p_ratio_0_30_s1_12; + else if (ratio.value < spl_fixpt_from_fraction(4, 10).value) + return easf_filter_4tap_64p_ratio_0_40_s1_12; + else if (ratio.value < spl_fixpt_from_fraction(5, 10).value) + return easf_filter_4tap_64p_ratio_0_50_s1_12; + else if (ratio.value < spl_fixpt_from_fraction(6, 10).value) + return easf_filter_4tap_64p_ratio_0_60_s1_12; + else if (ratio.value < spl_fixpt_from_fraction(7, 10).value) + return easf_filter_4tap_64p_ratio_0_70_s1_12; + else if (ratio.value < spl_fixpt_from_fraction(8, 10).value) + return easf_filter_4tap_64p_ratio_0_80_s1_12; + else if (ratio.value < spl_fixpt_from_fraction(9, 10).value) + return easf_filter_4tap_64p_ratio_0_90_s1_12; + else + return easf_filter_4tap_64p_ratio_1_00_s1_12; +} + +uint16_t *spl_get_easf_filter_6tap_64p(struct spl_fixed31_32 ratio) +{ + if (ratio.value < spl_fixpt_from_fraction(3, 10).value) + return easf_filter_6tap_64p_ratio_0_30_s1_12; + else if (ratio.value < spl_fixpt_from_fraction(4, 10).value) + return easf_filter_6tap_64p_ratio_0_40_s1_12; + else if (ratio.value < spl_fixpt_from_fraction(5, 10).value) + return easf_filter_6tap_64p_ratio_0_50_s1_12; + else if (ratio.value < spl_fixpt_from_fraction(6, 10).value) + return easf_filter_6tap_64p_ratio_0_60_s1_12; + else if (ratio.value < spl_fixpt_from_fraction(7, 10).value) + return easf_filter_6tap_64p_ratio_0_70_s1_12; + else if (ratio.value < spl_fixpt_from_fraction(8, 10).value) + return easf_filter_6tap_64p_ratio_0_80_s1_12; + else if (ratio.value < spl_fixpt_from_fraction(9, 10).value) + return easf_filter_6tap_64p_ratio_0_90_s1_12; + else + return easf_filter_6tap_64p_ratio_1_00_s1_12; +} + +uint16_t *spl_dscl_get_easf_filter_coeffs_64p(int taps, struct spl_fixed31_32 ratio) +{ + if (taps == 6) + return spl_get_easf_filter_6tap_64p(ratio); + else if (taps == 4) + return spl_get_easf_filter_4tap_64p(ratio); + else if (taps == 3) + return spl_get_easf_filter_3tap_64p(ratio); + else { + /* should never happen, bug */ + SPL_BREAK_TO_DEBUGGER(); + return NULL; + } +} + +void spl_set_filters_data(struct dscl_prog_data *dscl_prog_data, + const struct spl_scaler_data *data, bool enable_easf_v, + bool enable_easf_h) +{ + /* + * Old coefficients calculated scaling ratio = input / output + * New coefficients are calculated based on = output / input + */ + if (enable_easf_h) { + dscl_prog_data->filter_h = spl_dscl_get_easf_filter_coeffs_64p( + data->taps.h_taps, data->recip_ratios.horz); + + dscl_prog_data->filter_h_c = spl_dscl_get_easf_filter_coeffs_64p( + data->taps.h_taps_c, data->recip_ratios.horz_c); + } else { + dscl_prog_data->filter_h = spl_dscl_get_filter_coeffs_64p( + data->taps.h_taps, data->ratios.horz); + + dscl_prog_data->filter_h_c = spl_dscl_get_filter_coeffs_64p( + data->taps.h_taps_c, data->ratios.horz_c); + } + if (enable_easf_v) { + dscl_prog_data->filter_v = spl_dscl_get_easf_filter_coeffs_64p( + data->taps.v_taps, data->recip_ratios.vert); + + dscl_prog_data->filter_v_c = spl_dscl_get_easf_filter_coeffs_64p( + data->taps.v_taps_c, data->recip_ratios.vert_c); + } else { + dscl_prog_data->filter_v = spl_dscl_get_filter_coeffs_64p( + data->taps.v_taps, data->ratios.vert); + + dscl_prog_data->filter_v_c = spl_dscl_get_filter_coeffs_64p( + data->taps.v_taps_c, data->ratios.vert_c); + } +} + +static uint32_t spl_easf_get_scale_ratio_to_reg_value(struct spl_fixed31_32 ratio, + struct scale_ratio_to_reg_value_lookup *lookup_table_base_ptr, + unsigned int num_entries) +{ + unsigned int count = 0; + uint32_t value = 0; + struct scale_ratio_to_reg_value_lookup *lookup_table_index_ptr; + + lookup_table_index_ptr = (lookup_table_base_ptr + num_entries - 1); + value = lookup_table_index_ptr->reg_value; + + while (count < num_entries) { + + lookup_table_index_ptr = (lookup_table_base_ptr + count); + if (lookup_table_index_ptr->numer < 0) + break; + + if (ratio.value < spl_fixpt_from_fraction( + lookup_table_index_ptr->numer, + lookup_table_index_ptr->denom).value) { + value = lookup_table_index_ptr->reg_value; + break; + } + + count++; + } + return value; +} +uint32_t spl_get_v_bf3_mode(struct spl_fixed31_32 ratio) +{ + uint32_t value; + unsigned int num_entries = sizeof(easf_v_bf3_mode_lookup) / + sizeof(struct scale_ratio_to_reg_value_lookup); + value = spl_easf_get_scale_ratio_to_reg_value(ratio, + easf_v_bf3_mode_lookup, num_entries); + return value; +} +uint32_t spl_get_h_bf3_mode(struct spl_fixed31_32 ratio) +{ + uint32_t value; + unsigned int num_entries = sizeof(easf_h_bf3_mode_lookup) / + sizeof(struct scale_ratio_to_reg_value_lookup); + value = spl_easf_get_scale_ratio_to_reg_value(ratio, + easf_h_bf3_mode_lookup, num_entries); + return value; +} +uint32_t spl_get_reducer_gain6(int taps, struct spl_fixed31_32 ratio) +{ + uint32_t value; + unsigned int num_entries; + + if (taps == 4) { + num_entries = sizeof(easf_reducer_gain6_4tap_lookup) / + sizeof(struct scale_ratio_to_reg_value_lookup); + value = spl_easf_get_scale_ratio_to_reg_value(ratio, + easf_reducer_gain6_4tap_lookup, num_entries); + } else if (taps == 6) { + num_entries = sizeof(easf_reducer_gain6_6tap_lookup) / + sizeof(struct scale_ratio_to_reg_value_lookup); + value = spl_easf_get_scale_ratio_to_reg_value(ratio, + easf_reducer_gain6_6tap_lookup, num_entries); + } else + value = 0; + return value; +} +uint32_t spl_get_reducer_gain4(int taps, struct spl_fixed31_32 ratio) +{ + uint32_t value; + unsigned int num_entries; + + if (taps == 4) { + num_entries = sizeof(easf_reducer_gain4_4tap_lookup) / + sizeof(struct scale_ratio_to_reg_value_lookup); + value = spl_easf_get_scale_ratio_to_reg_value(ratio, + easf_reducer_gain4_4tap_lookup, num_entries); + } else if (taps == 6) { + num_entries = sizeof(easf_reducer_gain4_6tap_lookup) / + sizeof(struct scale_ratio_to_reg_value_lookup); + value = spl_easf_get_scale_ratio_to_reg_value(ratio, + easf_reducer_gain4_6tap_lookup, num_entries); + } else + value = 0; + return value; +} +uint32_t spl_get_gainRing6(int taps, struct spl_fixed31_32 ratio) +{ + uint32_t value; + unsigned int num_entries; + + if (taps == 4) { + num_entries = sizeof(easf_gain_ring6_4tap_lookup) / + sizeof(struct scale_ratio_to_reg_value_lookup); + value = spl_easf_get_scale_ratio_to_reg_value(ratio, + easf_gain_ring6_4tap_lookup, num_entries); + } else if (taps == 6) { + num_entries = sizeof(easf_gain_ring6_6tap_lookup) / + sizeof(struct scale_ratio_to_reg_value_lookup); + value = spl_easf_get_scale_ratio_to_reg_value(ratio, + easf_gain_ring6_6tap_lookup, num_entries); + } else + value = 0; + return value; +} +uint32_t spl_get_gainRing4(int taps, struct spl_fixed31_32 ratio) +{ + uint32_t value; + unsigned int num_entries; + + if (taps == 4) { + num_entries = sizeof(easf_gain_ring4_4tap_lookup) / + sizeof(struct scale_ratio_to_reg_value_lookup); + value = spl_easf_get_scale_ratio_to_reg_value(ratio, + easf_gain_ring4_4tap_lookup, num_entries); + } else if (taps == 6) { + num_entries = sizeof(easf_gain_ring4_6tap_lookup) / + sizeof(struct scale_ratio_to_reg_value_lookup); + value = spl_easf_get_scale_ratio_to_reg_value(ratio, + easf_gain_ring4_6tap_lookup, num_entries); + } else + value = 0; + return value; +} +uint32_t spl_get_3tap_dntilt_uptilt_offset(int taps, struct spl_fixed31_32 ratio) +{ + uint32_t value; + unsigned int num_entries; + + if (taps == 3) { + num_entries = sizeof(easf_3tap_dntilt_uptilt_offset_lookup) / + sizeof(struct scale_ratio_to_reg_value_lookup); + value = spl_easf_get_scale_ratio_to_reg_value(ratio, + easf_3tap_dntilt_uptilt_offset_lookup, num_entries); + } else + value = 0; + return value; +} +uint32_t spl_get_3tap_uptilt_maxval(int taps, struct spl_fixed31_32 ratio) +{ + uint32_t value; + unsigned int num_entries; + + if (taps == 3) { + num_entries = sizeof(easf_3tap_uptilt_maxval_lookup) / + sizeof(struct scale_ratio_to_reg_value_lookup); + value = spl_easf_get_scale_ratio_to_reg_value(ratio, + easf_3tap_uptilt_maxval_lookup, num_entries); + } else + value = 0; + return value; +} +uint32_t spl_get_3tap_dntilt_slope(int taps, struct spl_fixed31_32 ratio) +{ + uint32_t value; + unsigned int num_entries; + + if (taps == 3) { + num_entries = sizeof(easf_3tap_dntilt_slope_lookup) / + sizeof(struct scale_ratio_to_reg_value_lookup); + value = spl_easf_get_scale_ratio_to_reg_value(ratio, + easf_3tap_dntilt_slope_lookup, num_entries); + } else + value = 0; + return value; +} +uint32_t spl_get_3tap_uptilt1_slope(int taps, struct spl_fixed31_32 ratio) +{ + uint32_t value; + unsigned int num_entries; + + if (taps == 3) { + num_entries = sizeof(easf_3tap_uptilt1_slope_lookup) / + sizeof(struct scale_ratio_to_reg_value_lookup); + value = spl_easf_get_scale_ratio_to_reg_value(ratio, + easf_3tap_uptilt1_slope_lookup, num_entries); + } else + value = 0; + return value; +} +uint32_t spl_get_3tap_uptilt2_slope(int taps, struct spl_fixed31_32 ratio) +{ + uint32_t value; + unsigned int num_entries; + + if (taps == 3) { + num_entries = sizeof(easf_3tap_uptilt2_slope_lookup) / + sizeof(struct scale_ratio_to_reg_value_lookup); + value = spl_easf_get_scale_ratio_to_reg_value(ratio, + easf_3tap_uptilt2_slope_lookup, num_entries); + } else + value = 0; + return value; +} +uint32_t spl_get_3tap_uptilt2_offset(int taps, struct spl_fixed31_32 ratio) +{ + uint32_t value; + unsigned int num_entries; + + if (taps == 3) { + num_entries = sizeof(easf_3tap_uptilt2_offset_lookup) / + sizeof(struct scale_ratio_to_reg_value_lookup); + value = spl_easf_get_scale_ratio_to_reg_value(ratio, + easf_3tap_uptilt2_offset_lookup, num_entries); + } else + value = 0; + return value; +} diff --git a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_scl_easf_filters.h b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_scl_easf_filters.h new file mode 100644 index 000000000000..8bb2b8108e38 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_scl_easf_filters.h @@ -0,0 +1,38 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright 2024 Advanced Micro Devices, Inc. */ + +#ifndef __DC_SPL_SCL_EASF_FILTERS_H__ +#define __DC_SPL_SCL_EASF_FILTERS_H__ + +#include "dc_spl_types.h" + +struct scale_ratio_to_reg_value_lookup { + int numer; + int denom; + const uint32_t reg_value; +}; + +void spl_init_easf_filter_coeffs(void); +uint16_t *spl_get_easf_filter_3tap_64p(struct spl_fixed31_32 ratio); +uint16_t *spl_get_easf_filter_4tap_64p(struct spl_fixed31_32 ratio); +uint16_t *spl_get_easf_filter_6tap_64p(struct spl_fixed31_32 ratio); +uint16_t *spl_dscl_get_easf_filter_coeffs_64p(int taps, struct spl_fixed31_32 ratio); +void spl_set_filters_data(struct dscl_prog_data *dscl_prog_data, + const struct spl_scaler_data *data, bool enable_easf_v, + bool enable_easf_h); + +uint32_t spl_get_v_bf3_mode(struct spl_fixed31_32 ratio); +uint32_t spl_get_h_bf3_mode(struct spl_fixed31_32 ratio); +uint32_t spl_get_reducer_gain6(int taps, struct spl_fixed31_32 ratio); +uint32_t spl_get_reducer_gain4(int taps, struct spl_fixed31_32 ratio); +uint32_t spl_get_gainRing6(int taps, struct spl_fixed31_32 ratio); +uint32_t spl_get_gainRing4(int taps, struct spl_fixed31_32 ratio); +uint32_t spl_get_3tap_dntilt_uptilt_offset(int taps, struct spl_fixed31_32 ratio); +uint32_t spl_get_3tap_uptilt_maxval(int taps, struct spl_fixed31_32 ratio); +uint32_t spl_get_3tap_dntilt_slope(int taps, struct spl_fixed31_32 ratio); +uint32_t spl_get_3tap_uptilt1_slope(int taps, struct spl_fixed31_32 ratio); +uint32_t spl_get_3tap_uptilt2_slope(int taps, struct spl_fixed31_32 ratio); +uint32_t spl_get_3tap_uptilt2_offset(int taps, struct spl_fixed31_32 ratio); + +#endif /* __DC_SPL_SCL_EASF_FILTERS_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_scl_filters.c b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_scl_filters.c index e2baaf584139..b02c7b0b262b 100644 --- a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_scl_filters.c +++ b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_scl_filters.c @@ -2,6 +2,7 @@ // // Copyright 2024 Advanced Micro Devices, Inc. +#include "spl_debug.h" #include "dc_spl_scl_filters.h" //========================================= // <num_taps> = 2 @@ -1317,97 +1318,97 @@ static const uint16_t filter_8tap_64p_183[264] = { 0x3FD4, 0x3F84, 0x0214, 0x0694, 0x0694, 0x0214, 0x3F84, 0x3FD4 }; -const uint16_t *spl_get_filter_3tap_16p(struct fixed31_32 ratio) +const uint16_t *spl_get_filter_3tap_16p(struct spl_fixed31_32 ratio) { - if (ratio.value < dc_fixpt_one.value) + if (ratio.value < spl_fixpt_one.value) return filter_3tap_16p_upscale; - else if (ratio.value < dc_fixpt_from_fraction(4, 3).value) + else if (ratio.value < spl_fixpt_from_fraction(4, 3).value) return filter_3tap_16p_116; - else if (ratio.value < dc_fixpt_from_fraction(5, 3).value) + else if (ratio.value < spl_fixpt_from_fraction(5, 3).value) return filter_3tap_16p_149; else return filter_3tap_16p_183; } -const uint16_t *spl_get_filter_3tap_64p(struct fixed31_32 ratio) +const uint16_t *spl_get_filter_3tap_64p(struct spl_fixed31_32 ratio) { - if (ratio.value < dc_fixpt_one.value) + if (ratio.value < spl_fixpt_one.value) return filter_3tap_64p_upscale; - else if (ratio.value < dc_fixpt_from_fraction(4, 3).value) + else if (ratio.value < spl_fixpt_from_fraction(4, 3).value) return filter_3tap_64p_116; - else if (ratio.value < dc_fixpt_from_fraction(5, 3).value) + else if (ratio.value < spl_fixpt_from_fraction(5, 3).value) return filter_3tap_64p_149; else return filter_3tap_64p_183; } -const uint16_t *spl_get_filter_4tap_16p(struct fixed31_32 ratio) +const uint16_t *spl_get_filter_4tap_16p(struct spl_fixed31_32 ratio) { - if (ratio.value < dc_fixpt_one.value) + if (ratio.value < spl_fixpt_one.value) return filter_4tap_16p_upscale; - else if (ratio.value < dc_fixpt_from_fraction(4, 3).value) + else if (ratio.value < spl_fixpt_from_fraction(4, 3).value) return filter_4tap_16p_116; - else if (ratio.value < dc_fixpt_from_fraction(5, 3).value) + else if (ratio.value < spl_fixpt_from_fraction(5, 3).value) return filter_4tap_16p_149; else return filter_4tap_16p_183; } -const uint16_t *spl_get_filter_4tap_64p(struct fixed31_32 ratio) +const uint16_t *spl_get_filter_4tap_64p(struct spl_fixed31_32 ratio) { - if (ratio.value < dc_fixpt_one.value) + if (ratio.value < spl_fixpt_one.value) return filter_4tap_64p_upscale; - else if (ratio.value < dc_fixpt_from_fraction(4, 3).value) + else if (ratio.value < spl_fixpt_from_fraction(4, 3).value) return filter_4tap_64p_116; - else if (ratio.value < dc_fixpt_from_fraction(5, 3).value) + else if (ratio.value < spl_fixpt_from_fraction(5, 3).value) return filter_4tap_64p_149; else return filter_4tap_64p_183; } -const uint16_t *spl_get_filter_5tap_64p(struct fixed31_32 ratio) +const uint16_t *spl_get_filter_5tap_64p(struct spl_fixed31_32 ratio) { - if (ratio.value < dc_fixpt_one.value) + if (ratio.value < spl_fixpt_one.value) return filter_5tap_64p_upscale; - else if (ratio.value < dc_fixpt_from_fraction(4, 3).value) + else if (ratio.value < spl_fixpt_from_fraction(4, 3).value) return filter_5tap_64p_116; - else if (ratio.value < dc_fixpt_from_fraction(5, 3).value) + else if (ratio.value < spl_fixpt_from_fraction(5, 3).value) return filter_5tap_64p_149; else return filter_5tap_64p_183; } -const uint16_t *spl_get_filter_6tap_64p(struct fixed31_32 ratio) +const uint16_t *spl_get_filter_6tap_64p(struct spl_fixed31_32 ratio) { - if (ratio.value < dc_fixpt_one.value) + if (ratio.value < spl_fixpt_one.value) return filter_6tap_64p_upscale; - else if (ratio.value < dc_fixpt_from_fraction(4, 3).value) + else if (ratio.value < spl_fixpt_from_fraction(4, 3).value) return filter_6tap_64p_116; - else if (ratio.value < dc_fixpt_from_fraction(5, 3).value) + else if (ratio.value < spl_fixpt_from_fraction(5, 3).value) return filter_6tap_64p_149; else return filter_6tap_64p_183; } -const uint16_t *spl_get_filter_7tap_64p(struct fixed31_32 ratio) +const uint16_t *spl_get_filter_7tap_64p(struct spl_fixed31_32 ratio) { - if (ratio.value < dc_fixpt_one.value) + if (ratio.value < spl_fixpt_one.value) return filter_7tap_64p_upscale; - else if (ratio.value < dc_fixpt_from_fraction(4, 3).value) + else if (ratio.value < spl_fixpt_from_fraction(4, 3).value) return filter_7tap_64p_116; - else if (ratio.value < dc_fixpt_from_fraction(5, 3).value) + else if (ratio.value < spl_fixpt_from_fraction(5, 3).value) return filter_7tap_64p_149; else return filter_7tap_64p_183; } -const uint16_t *spl_get_filter_8tap_64p(struct fixed31_32 ratio) +const uint16_t *spl_get_filter_8tap_64p(struct spl_fixed31_32 ratio) { - if (ratio.value < dc_fixpt_one.value) + if (ratio.value < spl_fixpt_one.value) return filter_8tap_64p_upscale; - else if (ratio.value < dc_fixpt_from_fraction(4, 3).value) + else if (ratio.value < spl_fixpt_from_fraction(4, 3).value) return filter_8tap_64p_116; - else if (ratio.value < dc_fixpt_from_fraction(5, 3).value) + else if (ratio.value < spl_fixpt_from_fraction(5, 3).value) return filter_8tap_64p_149; else return filter_8tap_64p_183; @@ -1422,3 +1423,29 @@ const uint16_t *spl_get_filter_2tap_64p(void) { return filter_2tap_64p; } + +const uint16_t *spl_dscl_get_filter_coeffs_64p(int taps, struct spl_fixed31_32 ratio) +{ + if (taps == 8) + return spl_get_filter_8tap_64p(ratio); + else if (taps == 7) + return spl_get_filter_7tap_64p(ratio); + else if (taps == 6) + return spl_get_filter_6tap_64p(ratio); + else if (taps == 5) + return spl_get_filter_5tap_64p(ratio); + else if (taps == 4) + return spl_get_filter_4tap_64p(ratio); + else if (taps == 3) + return spl_get_filter_3tap_64p(ratio); + else if (taps == 2) + return spl_get_filter_2tap_64p(); + else if (taps == 1) + return NULL; + else { + /* should never happen, bug */ + SPL_BREAK_TO_DEBUGGER(); + return NULL; + } +} + diff --git a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_scl_filters.h b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_scl_filters.h index 6d96aca53b24..48202bc4f81e 100644 --- a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_scl_filters.h +++ b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_scl_filters.h @@ -7,53 +7,16 @@ #include "dc_spl_types.h" -const uint16_t *spl_get_filter_3tap_16p(struct fixed31_32 ratio); -const uint16_t *spl_get_filter_3tap_64p(struct fixed31_32 ratio); -const uint16_t *spl_get_filter_4tap_16p(struct fixed31_32 ratio); -const uint16_t *spl_get_filter_4tap_64p(struct fixed31_32 ratio); -const uint16_t *spl_get_filter_5tap_64p(struct fixed31_32 ratio); -const uint16_t *spl_get_filter_6tap_64p(struct fixed31_32 ratio); -const uint16_t *spl_get_filter_7tap_64p(struct fixed31_32 ratio); -const uint16_t *spl_get_filter_8tap_64p(struct fixed31_32 ratio); +const uint16_t *spl_get_filter_3tap_16p(struct spl_fixed31_32 ratio); +const uint16_t *spl_get_filter_3tap_64p(struct spl_fixed31_32 ratio); +const uint16_t *spl_get_filter_4tap_16p(struct spl_fixed31_32 ratio); +const uint16_t *spl_get_filter_4tap_64p(struct spl_fixed31_32 ratio); +const uint16_t *spl_get_filter_5tap_64p(struct spl_fixed31_32 ratio); +const uint16_t *spl_get_filter_6tap_64p(struct spl_fixed31_32 ratio); +const uint16_t *spl_get_filter_7tap_64p(struct spl_fixed31_32 ratio); +const uint16_t *spl_get_filter_8tap_64p(struct spl_fixed31_32 ratio); const uint16_t *spl_get_filter_2tap_16p(void); const uint16_t *spl_get_filter_2tap_64p(void); -const uint16_t *spl_get_filter_3tap_16p_upscale(void); -const uint16_t *spl_get_filter_3tap_16p_116(void); -const uint16_t *spl_get_filter_3tap_16p_149(void); -const uint16_t *spl_get_filter_3tap_16p_183(void); +const uint16_t *spl_dscl_get_filter_coeffs_64p(int taps, struct spl_fixed31_32 ratio); -const uint16_t *spl_get_filter_4tap_16p_upscale(void); -const uint16_t *spl_get_filter_4tap_16p_116(void); -const uint16_t *spl_get_filter_4tap_16p_149(void); -const uint16_t *spl_get_filter_4tap_16p_183(void); - -const uint16_t *spl_get_filter_3tap_64p_upscale(void); -const uint16_t *spl_get_filter_3tap_64p_116(void); -const uint16_t *spl_get_filter_3tap_64p_149(void); -const uint16_t *spl_get_filter_3tap_64p_183(void); - -const uint16_t *spl_get_filter_4tap_64p_upscale(void); -const uint16_t *spl_get_filter_4tap_64p_116(void); -const uint16_t *spl_get_filter_4tap_64p_149(void); -const uint16_t *spl_get_filter_4tap_64p_183(void); - -const uint16_t *spl_get_filter_5tap_64p_upscale(void); -const uint16_t *spl_get_filter_5tap_64p_116(void); -const uint16_t *spl_get_filter_5tap_64p_149(void); -const uint16_t *spl_get_filter_5tap_64p_183(void); - -const uint16_t *spl_get_filter_6tap_64p_upscale(void); -const uint16_t *spl_get_filter_6tap_64p_116(void); -const uint16_t *spl_get_filter_6tap_64p_149(void); -const uint16_t *spl_get_filter_6tap_64p_183(void); - -const uint16_t *spl_get_filter_7tap_64p_upscale(void); -const uint16_t *spl_get_filter_7tap_64p_116(void); -const uint16_t *spl_get_filter_7tap_64p_149(void); -const uint16_t *spl_get_filter_7tap_64p_183(void); - -const uint16_t *spl_get_filter_8tap_64p_upscale(void); -const uint16_t *spl_get_filter_8tap_64p_116(void); -const uint16_t *spl_get_filter_8tap_64p_149(void); -const uint16_t *spl_get_filter_8tap_64p_183(void); #endif /* __DC_SPL_SCL_FILTERS_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_scl_filters_old.c b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_scl_filters_old.c deleted file mode 100644 index bb0e1b80ec3c..000000000000 --- a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_scl_filters_old.c +++ /dev/null @@ -1,25 +0,0 @@ -/* - * Copyright 2012-16 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - diff --git a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_types.h b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_types.h index 36d10b0f2eed..85b19ebe2c57 100644 --- a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_types.h +++ b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_types.h @@ -2,14 +2,16 @@ // // Copyright 2024 Advanced Micro Devices, Inc. -#include "os_types.h" // swap -#ifndef ASSERT -#define ASSERT(_bool) ((void *)0) -#endif -#include "include/fixed31_32.h" // fixed31_32 and related functions #ifndef __DC_SPL_TYPES_H__ #define __DC_SPL_TYPES_H__ +#include "spl_os_types.h" // swap +#ifndef SPL_ASSERT +#define SPL_ASSERT(_bool) ((void *)0) +#endif +#include "spl_fixpt31_32.h" // fixed31_32 and related functions +#include "spl_custom_float.h" // custom float and related functions + struct spl_size { uint32_t width; uint32_t height; @@ -22,16 +24,16 @@ struct spl_rect { }; struct spl_ratios { - struct fixed31_32 horz; - struct fixed31_32 vert; - struct fixed31_32 horz_c; - struct fixed31_32 vert_c; + struct spl_fixed31_32 horz; + struct spl_fixed31_32 vert; + struct spl_fixed31_32 horz_c; + struct spl_fixed31_32 vert_c; }; struct spl_inits { - struct fixed31_32 h; - struct fixed31_32 h_c; - struct fixed31_32 v; - struct fixed31_32 v_c; + struct spl_fixed31_32 h; + struct spl_fixed31_32 h_c; + struct spl_fixed31_32 v; + struct spl_fixed31_32 v_c; }; struct spl_taps { @@ -64,6 +66,8 @@ enum spl_pixel_format { SPL_PIXEL_FORMAT_420BPP10, /*end of pixel format definition*/ SPL_PIXEL_FORMAT_INVALID, + SPL_PIXEL_FORMAT_422BPP8, + SPL_PIXEL_FORMAT_422BPP10, SPL_PIXEL_FORMAT_GRPH_BEGIN = SPL_PIXEL_FORMAT_INDEX8, SPL_PIXEL_FORMAT_GRPH_END = SPL_PIXEL_FORMAT_FP16, SPL_PIXEL_FORMAT_VIDEO_BEGIN = SPL_PIXEL_FORMAT_420BPP8, @@ -135,6 +139,7 @@ struct spl_scaler_data { struct spl_rect viewport_c; struct spl_rect recout; struct spl_ratios ratios; + struct spl_ratios recip_ratios; struct spl_inits inits; }; @@ -402,13 +407,19 @@ struct dscl_prog_data { /* blur and scale filter */ const uint16_t *filter_blur_scale_v; const uint16_t *filter_blur_scale_h; + int sharpness_level; /* Track sharpness level */ }; /* SPL input and output definitions */ -// SPL outputs struct -struct spl_out { +// SPL scratch struct +struct spl_scratch { // Pack all SPL outputs in scl_data struct spl_scaler_data scl_data; +}; + +/* SPL input and output definitions */ +// SPL outputs struct +struct spl_out { // Pack all output need to program hw registers struct dscl_prog_data *dscl_prog_data; }; @@ -450,14 +461,26 @@ struct basic_out { bool alpha_en; bool use_two_pixels_per_container; }; -enum explicit_sharpness { - SHARPNESS_LOW = 0, - SHARPNESS_MID, - SHARPNESS_HIGH -}; -struct adaptive_sharpness { +enum sharpness_setting { + SHARPNESS_HW_OFF = 0, + SHARPNESS_ZERO, + SHARPNESS_CUSTOM +}; +struct spl_sharpness_range { + int sdr_rgb_min; + int sdr_rgb_max; + int sdr_rgb_mid; + int sdr_yuv_min; + int sdr_yuv_max; + int sdr_yuv_mid; + int hdr_rgb_min; + int hdr_rgb_max; + int hdr_rgb_mid; +}; +struct adaptive_sharpness { bool enable; - enum explicit_sharpness sharpness; + int sharpness_level; + struct spl_sharpness_range sharpness_range; }; enum linear_light_scaling { // convert it in translation logic LLS_PREF_DONT_CARE = 0, @@ -491,6 +514,11 @@ struct spl_in { bool prefer_easf; bool disable_easf; struct spl_debug debug; + bool is_fullscreen; + bool is_hdr_on; + int h_active; + int v_active; + int hdr_multx100; }; // end of SPL inputs diff --git a/drivers/gpu/drm/amd/display/dc/spl/spl_custom_float.c b/drivers/gpu/drm/amd/display/dc/spl/spl_custom_float.c new file mode 100644 index 000000000000..be2f34d034c5 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/spl/spl_custom_float.c @@ -0,0 +1,151 @@ +// SPDX-License-Identifier: MIT +// +// Copyright 2024 Advanced Micro Devices, Inc. + +#include "spl_debug.h" +#include "spl_custom_float.h" + +static bool spl_build_custom_float(struct spl_fixed31_32 value, + const struct spl_custom_float_format *format, + bool *negative, + uint32_t *mantissa, + uint32_t *exponenta) +{ + uint32_t exp_offset = (1 << (format->exponenta_bits - 1)) - 1; + + const struct spl_fixed31_32 mantissa_constant_plus_max_fraction = + spl_fixpt_from_fraction((1LL << (format->mantissa_bits + 1)) - 1, + 1LL << format->mantissa_bits); + + struct spl_fixed31_32 mantiss; + + if (spl_fixpt_eq(value, spl_fixpt_zero)) { + *negative = false; + *mantissa = 0; + *exponenta = 0; + return true; + } + + if (spl_fixpt_lt(value, spl_fixpt_zero)) { + *negative = format->sign; + value = spl_fixpt_neg(value); + } else { + *negative = false; + } + + if (spl_fixpt_lt(value, spl_fixpt_one)) { + uint32_t i = 1; + + do { + value = spl_fixpt_shl(value, 1); + ++i; + } while (spl_fixpt_lt(value, spl_fixpt_one)); + + --i; + + if (exp_offset <= i) { + *mantissa = 0; + *exponenta = 0; + return true; + } + + *exponenta = exp_offset - i; + } else if (spl_fixpt_le(mantissa_constant_plus_max_fraction, value)) { + uint32_t i = 1; + + do { + value = spl_fixpt_shr(value, 1); + ++i; + } while (spl_fixpt_lt(mantissa_constant_plus_max_fraction, value)); + + *exponenta = exp_offset + i - 1; + } else { + *exponenta = exp_offset; + } + + mantiss = spl_fixpt_sub(value, spl_fixpt_one); + + if (spl_fixpt_lt(mantiss, spl_fixpt_zero) || + spl_fixpt_lt(spl_fixpt_one, mantiss)) + mantiss = spl_fixpt_zero; + else + mantiss = spl_fixpt_shl(mantiss, format->mantissa_bits); + + *mantissa = spl_fixpt_floor(mantiss); + + return true; +} + +static bool spl_setup_custom_float(const struct spl_custom_float_format *format, + bool negative, + uint32_t mantissa, + uint32_t exponenta, + uint32_t *result) +{ + uint32_t i = 0; + uint32_t j = 0; + uint32_t value = 0; + + /* verification code: + * once calculation is ok we can remove it + */ + + const uint32_t mantissa_mask = + (1 << (format->mantissa_bits + 1)) - 1; + + const uint32_t exponenta_mask = + (1 << (format->exponenta_bits + 1)) - 1; + + if (mantissa & ~mantissa_mask) { + SPL_BREAK_TO_DEBUGGER(); + mantissa = mantissa_mask; + } + + if (exponenta & ~exponenta_mask) { + SPL_BREAK_TO_DEBUGGER(); + exponenta = exponenta_mask; + } + + /* end of verification code */ + + while (i < format->mantissa_bits) { + uint32_t mask = 1 << i; + + if (mantissa & mask) + value |= mask; + + ++i; + } + + while (j < format->exponenta_bits) { + uint32_t mask = 1 << j; + + if (exponenta & mask) + value |= mask << i; + + ++j; + } + + if (negative && format->sign) + value |= 1 << (i + j); + + *result = value; + + return true; +} + +bool spl_convert_to_custom_float_format(struct spl_fixed31_32 value, + const struct spl_custom_float_format *format, + uint32_t *result) +{ + uint32_t mantissa; + uint32_t exponenta; + bool negative; + + return spl_build_custom_float(value, format, &negative, &mantissa, &exponenta) && + spl_setup_custom_float(format, + negative, + mantissa, + exponenta, + result); +} diff --git a/drivers/gpu/drm/amd/display/dc/spl/spl_custom_float.h b/drivers/gpu/drm/amd/display/dc/spl/spl_custom_float.h new file mode 100644 index 000000000000..cdc4e107b9de --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/spl/spl_custom_float.h @@ -0,0 +1,29 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright 2024 Advanced Micro Devices, Inc. */ + +#ifndef SPL_CUSTOM_FLOAT_H_ +#define SPL_CUSTOM_FLOAT_H_ + +#include "spl_os_types.h" +#include "spl_fixpt31_32.h" + +struct spl_custom_float_format { + uint32_t mantissa_bits; + uint32_t exponenta_bits; + bool sign; +}; + +struct spl_custom_float_value { + uint32_t mantissa; + uint32_t exponenta; + uint32_t value; + bool negative; +}; + +bool spl_convert_to_custom_float_format( + struct spl_fixed31_32 value, + const struct spl_custom_float_format *format, + uint32_t *result); + +#endif //SPL_CUSTOM_FLOAT_H_ diff --git a/drivers/gpu/drm/amd/display/dc/spl/spl_debug.h b/drivers/gpu/drm/amd/display/dc/spl/spl_debug.h new file mode 100644 index 000000000000..5696dafd0894 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/spl/spl_debug.h @@ -0,0 +1,25 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright 2024 Advanced Micro Devices, Inc. */ + +#ifndef SPL_DEBUG_H +#define SPL_DEBUG_H + +#ifdef SPL_ASSERT +#undef SPL_ASSERT +#endif +#define SPL_ASSERT(b) + +#define SPL_ASSERT_CRITICAL(expr) do {if (expr)/* Do nothing */; } while (0) + +#ifdef SPL_DALMSG +#undef SPL_DALMSG +#endif +#define SPL_DALMSG(b) + +#ifdef SPL_DAL_ASSERT_MSG +#undef SPL_DAL_ASSERT_MSG +#endif +#define SPL_DAL_ASSERT_MSG(b, m) + +#endif // SPL_DEBUG_H diff --git a/drivers/gpu/drm/amd/display/dc/spl/spl_fixpt31_32.c b/drivers/gpu/drm/amd/display/dc/spl/spl_fixpt31_32.c new file mode 100644 index 000000000000..a95565df5487 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/spl/spl_fixpt31_32.c @@ -0,0 +1,497 @@ +// SPDX-License-Identifier: MIT +// +// Copyright 2024 Advanced Micro Devices, Inc. + +#include "spl_fixpt31_32.h" + +static const struct spl_fixed31_32 spl_fixpt_two_pi = { 26986075409LL }; +static const struct spl_fixed31_32 spl_fixpt_ln2 = { 2977044471LL }; +static const struct spl_fixed31_32 spl_fixpt_ln2_div_2 = { 1488522236LL }; + +static inline unsigned long long abs_i64( + long long arg) +{ + if (arg > 0) + return (unsigned long long)arg; + else + return (unsigned long long)(-arg); +} + +/* + * @brief + * result = dividend / divisor + * *remainder = dividend % divisor + */ +static inline unsigned long long complete_integer_division_u64( + unsigned long long dividend, + unsigned long long divisor, + unsigned long long *remainder) +{ + unsigned long long result; + + ASSERT(divisor); + + result = spl_div64_u64_rem(dividend, divisor, remainder); + + return result; +} + + +#define FRACTIONAL_PART_MASK \ + ((1ULL << FIXED31_32_BITS_PER_FRACTIONAL_PART) - 1) + +#define GET_INTEGER_PART(x) \ + ((x) >> FIXED31_32_BITS_PER_FRACTIONAL_PART) + +#define GET_FRACTIONAL_PART(x) \ + (FRACTIONAL_PART_MASK & (x)) + +struct spl_fixed31_32 spl_fixpt_from_fraction(long long numerator, long long denominator) +{ + struct spl_fixed31_32 res; + + bool arg1_negative = numerator < 0; + bool arg2_negative = denominator < 0; + + unsigned long long arg1_value = arg1_negative ? -numerator : numerator; + unsigned long long arg2_value = arg2_negative ? -denominator : denominator; + + unsigned long long remainder; + + /* determine integer part */ + + unsigned long long res_value = complete_integer_division_u64( + arg1_value, arg2_value, &remainder); + + ASSERT(res_value <= LONG_MAX); + + /* determine fractional part */ + { + unsigned int i = FIXED31_32_BITS_PER_FRACTIONAL_PART; + + do { + remainder <<= 1; + + res_value <<= 1; + + if (remainder >= arg2_value) { + res_value |= 1; + remainder -= arg2_value; + } + } while (--i != 0); + } + + /* round up LSB */ + { + unsigned long long summand = (remainder << 1) >= arg2_value; + + ASSERT(res_value <= LLONG_MAX - summand); + + res_value += summand; + } + + res.value = (long long)res_value; + + if (arg1_negative ^ arg2_negative) + res.value = -res.value; + + return res; +} + +struct spl_fixed31_32 spl_fixpt_mul(struct spl_fixed31_32 arg1, struct spl_fixed31_32 arg2) +{ + struct spl_fixed31_32 res; + + bool arg1_negative = arg1.value < 0; + bool arg2_negative = arg2.value < 0; + + unsigned long long arg1_value = arg1_negative ? -arg1.value : arg1.value; + unsigned long long arg2_value = arg2_negative ? -arg2.value : arg2.value; + + unsigned long long arg1_int = GET_INTEGER_PART(arg1_value); + unsigned long long arg2_int = GET_INTEGER_PART(arg2_value); + + unsigned long long arg1_fra = GET_FRACTIONAL_PART(arg1_value); + unsigned long long arg2_fra = GET_FRACTIONAL_PART(arg2_value); + + unsigned long long tmp; + + res.value = arg1_int * arg2_int; + + ASSERT(res.value <= (long long)LONG_MAX); + + res.value <<= FIXED31_32_BITS_PER_FRACTIONAL_PART; + + tmp = arg1_int * arg2_fra; + + ASSERT(tmp <= (unsigned long long)(LLONG_MAX - res.value)); + + res.value += tmp; + + tmp = arg2_int * arg1_fra; + + ASSERT(tmp <= (unsigned long long)(LLONG_MAX - res.value)); + + res.value += tmp; + + tmp = arg1_fra * arg2_fra; + + tmp = (tmp >> FIXED31_32_BITS_PER_FRACTIONAL_PART) + + (tmp >= (unsigned long long)spl_fixpt_half.value); + + ASSERT(tmp <= (unsigned long long)(LLONG_MAX - res.value)); + + res.value += tmp; + + if (arg1_negative ^ arg2_negative) + res.value = -res.value; + + return res; +} + +struct spl_fixed31_32 spl_fixpt_sqr(struct spl_fixed31_32 arg) +{ + struct spl_fixed31_32 res; + + unsigned long long arg_value = abs_i64(arg.value); + + unsigned long long arg_int = GET_INTEGER_PART(arg_value); + + unsigned long long arg_fra = GET_FRACTIONAL_PART(arg_value); + + unsigned long long tmp; + + res.value = arg_int * arg_int; + + ASSERT(res.value <= (long long)LONG_MAX); + + res.value <<= FIXED31_32_BITS_PER_FRACTIONAL_PART; + + tmp = arg_int * arg_fra; + + ASSERT(tmp <= (unsigned long long)(LLONG_MAX - res.value)); + + res.value += tmp; + + ASSERT(tmp <= (unsigned long long)(LLONG_MAX - res.value)); + + res.value += tmp; + + tmp = arg_fra * arg_fra; + + tmp = (tmp >> FIXED31_32_BITS_PER_FRACTIONAL_PART) + + (tmp >= (unsigned long long)spl_fixpt_half.value); + + ASSERT(tmp <= (unsigned long long)(LLONG_MAX - res.value)); + + res.value += tmp; + + return res; +} + +struct spl_fixed31_32 spl_fixpt_recip(struct spl_fixed31_32 arg) +{ + /* + * @note + * Good idea to use Newton's method + */ + + ASSERT(arg.value); + + return spl_fixpt_from_fraction( + spl_fixpt_one.value, + arg.value); +} + +struct spl_fixed31_32 spl_fixpt_sinc(struct spl_fixed31_32 arg) +{ + struct spl_fixed31_32 square; + + struct spl_fixed31_32 res = spl_fixpt_one; + + int n = 27; + + struct spl_fixed31_32 arg_norm = arg; + + if (spl_fixpt_le( + spl_fixpt_two_pi, + spl_fixpt_abs(arg))) { + arg_norm = spl_fixpt_sub( + arg_norm, + spl_fixpt_mul_int( + spl_fixpt_two_pi, + (int)spl_div64_s64( + arg_norm.value, + spl_fixpt_two_pi.value))); + } + + square = spl_fixpt_sqr(arg_norm); + + do { + res = spl_fixpt_sub( + spl_fixpt_one, + spl_fixpt_div_int( + spl_fixpt_mul( + square, + res), + n * (n - 1))); + + n -= 2; + } while (n > 2); + + if (arg.value != arg_norm.value) + res = spl_fixpt_div( + spl_fixpt_mul(res, arg_norm), + arg); + + return res; +} + +struct spl_fixed31_32 spl_fixpt_sin(struct spl_fixed31_32 arg) +{ + return spl_fixpt_mul( + arg, + spl_fixpt_sinc(arg)); +} + +struct spl_fixed31_32 spl_fixpt_cos(struct spl_fixed31_32 arg) +{ + /* TODO implement argument normalization */ + + const struct spl_fixed31_32 square = spl_fixpt_sqr(arg); + + struct spl_fixed31_32 res = spl_fixpt_one; + + int n = 26; + + do { + res = spl_fixpt_sub( + spl_fixpt_one, + spl_fixpt_div_int( + spl_fixpt_mul( + square, + res), + n * (n - 1))); + + n -= 2; + } while (n != 0); + + return res; +} + +/* + * @brief + * result = exp(arg), + * where abs(arg) < 1 + * + * Calculated as Taylor series. + */ +static struct spl_fixed31_32 fixed31_32_exp_from_taylor_series(struct spl_fixed31_32 arg) +{ + unsigned int n = 9; + + struct spl_fixed31_32 res = spl_fixpt_from_fraction( + n + 2, + n + 1); + /* TODO find correct res */ + + ASSERT(spl_fixpt_lt(arg, spl_fixpt_one)); + + do + res = spl_fixpt_add( + spl_fixpt_one, + spl_fixpt_div_int( + spl_fixpt_mul( + arg, + res), + n)); + while (--n != 1); + + return spl_fixpt_add( + spl_fixpt_one, + spl_fixpt_mul( + arg, + res)); +} + +struct spl_fixed31_32 spl_fixpt_exp(struct spl_fixed31_32 arg) +{ + /* + * @brief + * Main equation is: + * exp(x) = exp(r + m * ln(2)) = (1 << m) * exp(r), + * where m = round(x / ln(2)), r = x - m * ln(2) + */ + + if (spl_fixpt_le( + spl_fixpt_ln2_div_2, + spl_fixpt_abs(arg))) { + int m = spl_fixpt_round( + spl_fixpt_div( + arg, + spl_fixpt_ln2)); + + struct spl_fixed31_32 r = spl_fixpt_sub( + arg, + spl_fixpt_mul_int( + spl_fixpt_ln2, + m)); + + ASSERT(m != 0); + + ASSERT(spl_fixpt_lt( + spl_fixpt_abs(r), + spl_fixpt_one)); + + if (m > 0) + return spl_fixpt_shl( + fixed31_32_exp_from_taylor_series(r), + (unsigned char)m); + else + return spl_fixpt_div_int( + fixed31_32_exp_from_taylor_series(r), + 1LL << -m); + } else if (arg.value != 0) + return fixed31_32_exp_from_taylor_series(arg); + else + return spl_fixpt_one; +} + +struct spl_fixed31_32 spl_fixpt_log(struct spl_fixed31_32 arg) +{ + struct spl_fixed31_32 res = spl_fixpt_neg(spl_fixpt_one); + /* TODO improve 1st estimation */ + + struct spl_fixed31_32 error; + + ASSERT(arg.value > 0); + /* TODO if arg is negative, return NaN */ + /* TODO if arg is zero, return -INF */ + + do { + struct spl_fixed31_32 res1 = spl_fixpt_add( + spl_fixpt_sub( + res, + spl_fixpt_one), + spl_fixpt_div( + arg, + spl_fixpt_exp(res))); + + error = spl_fixpt_sub( + res, + res1); + + res = res1; + /* TODO determine max_allowed_error based on quality of exp() */ + } while (abs_i64(error.value) > 100ULL); + + return res; +} + + +/* this function is a generic helper to translate fixed point value to + * specified integer format that will consist of integer_bits integer part and + * fractional_bits fractional part. For example it is used in + * spl_fixpt_u2d19 to receive 2 bits integer part and 19 bits fractional + * part in 32 bits. It is used in hw programming (scaler) + */ + +static inline unsigned int ux_dy( + long long value, + unsigned int integer_bits, + unsigned int fractional_bits) +{ + /* 1. create mask of integer part */ + unsigned int result = (1 << integer_bits) - 1; + /* 2. mask out fractional part */ + unsigned int fractional_part = FRACTIONAL_PART_MASK & value; + /* 3. shrink fixed point integer part to be of integer_bits width*/ + result &= GET_INTEGER_PART(value); + /* 4. make space for fractional part to be filled in after integer */ + result <<= fractional_bits; + /* 5. shrink fixed point fractional part to of fractional_bits width*/ + fractional_part >>= FIXED31_32_BITS_PER_FRACTIONAL_PART - fractional_bits; + /* 6. merge the result */ + return result | fractional_part; +} + +static inline unsigned int clamp_ux_dy( + long long value, + unsigned int integer_bits, + unsigned int fractional_bits, + unsigned int min_clamp) +{ + unsigned int truncated_val = ux_dy(value, integer_bits, fractional_bits); + + if (value >= (1LL << (integer_bits + FIXED31_32_BITS_PER_FRACTIONAL_PART))) + return (1 << (integer_bits + fractional_bits)) - 1; + else if (truncated_val > min_clamp) + return truncated_val; + else + return min_clamp; +} + +unsigned int spl_fixpt_u4d19(struct spl_fixed31_32 arg) +{ + return ux_dy(arg.value, 4, 19); +} + +unsigned int spl_fixpt_u3d19(struct spl_fixed31_32 arg) +{ + return ux_dy(arg.value, 3, 19); +} + +unsigned int spl_fixpt_u2d19(struct spl_fixed31_32 arg) +{ + return ux_dy(arg.value, 2, 19); +} + +unsigned int spl_fixpt_u0d19(struct spl_fixed31_32 arg) +{ + return ux_dy(arg.value, 0, 19); +} + +unsigned int spl_fixpt_clamp_u0d14(struct spl_fixed31_32 arg) +{ + return clamp_ux_dy(arg.value, 0, 14, 1); +} + +unsigned int spl_fixpt_clamp_u0d10(struct spl_fixed31_32 arg) +{ + return clamp_ux_dy(arg.value, 0, 10, 1); +} + +int spl_fixpt_s4d19(struct spl_fixed31_32 arg) +{ + if (arg.value < 0) + return -(int)ux_dy(spl_fixpt_abs(arg).value, 4, 19); + else + return ux_dy(arg.value, 4, 19); +} + +struct spl_fixed31_32 spl_fixpt_from_ux_dy(unsigned int value, + unsigned int integer_bits, + unsigned int fractional_bits) +{ + struct spl_fixed31_32 fixpt_value = spl_fixpt_zero; + struct spl_fixed31_32 fixpt_int_value = spl_fixpt_zero; + long long frac_mask = ((long long)1 << (long long)integer_bits) - 1; + + fixpt_value.value = (long long)value << (FIXED31_32_BITS_PER_FRACTIONAL_PART - fractional_bits); + frac_mask = frac_mask << fractional_bits; + fixpt_int_value.value = value & frac_mask; + fixpt_int_value.value <<= (FIXED31_32_BITS_PER_FRACTIONAL_PART - fractional_bits); + fixpt_value.value |= fixpt_int_value.value; + return fixpt_value; +} + +struct spl_fixed31_32 spl_fixpt_from_int_dy(unsigned int int_value, + unsigned int frac_value, + unsigned int integer_bits, + unsigned int fractional_bits) +{ + struct spl_fixed31_32 fixpt_value = spl_fixpt_from_int(int_value); + + fixpt_value.value |= (long long)frac_value << (FIXED31_32_BITS_PER_FRACTIONAL_PART - fractional_bits); + return fixpt_value; +} diff --git a/drivers/gpu/drm/amd/display/dc/spl/spl_fixpt31_32.h b/drivers/gpu/drm/amd/display/dc/spl/spl_fixpt31_32.h new file mode 100644 index 000000000000..8a045e2f8699 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/spl/spl_fixpt31_32.h @@ -0,0 +1,525 @@ +/* SPDX-License-Identifier: MIT */ + +/* Copyright 2024 Advanced Micro Devices, Inc. */ + +#ifndef __SPL_FIXED31_32_H__ +#define __SPL_FIXED31_32_H__ + +#include "os_types.h" +#include "spl_os_types.h" // swap +#ifndef ASSERT +#define ASSERT(_bool) ((void *)0) +#endif + +#ifndef LLONG_MAX +#define LLONG_MAX 9223372036854775807ll +#endif +#ifndef LLONG_MIN +#define LLONG_MIN (-LLONG_MAX - 1ll) +#endif + +#define FIXED31_32_BITS_PER_FRACTIONAL_PART 32 +#ifndef LLONG_MIN +#define LLONG_MIN (1LL<<63) +#endif +#ifndef LLONG_MAX +#define LLONG_MAX (-1LL>>1) +#endif + +/* + * @brief + * Arithmetic operations on real numbers + * represented as fixed-point numbers. + * There are: 1 bit for sign, + * 31 bit for integer part, + * 32 bits for fractional part. + * + * @note + * Currently, overflows and underflows are asserted; + * no special result returned. + */ + +struct spl_fixed31_32 { + long long value; +}; + + +/* + * @brief + * Useful constants + */ + +static const struct spl_fixed31_32 spl_fixpt_zero = { 0 }; +static const struct spl_fixed31_32 spl_fixpt_epsilon = { 1LL }; +static const struct spl_fixed31_32 spl_fixpt_half = { 0x80000000LL }; +static const struct spl_fixed31_32 spl_fixpt_one = { 0x100000000LL }; + +/* + * @brief + * Initialization routines + */ + +/* + * @brief + * result = numerator / denominator + */ +struct spl_fixed31_32 spl_fixpt_from_fraction(long long numerator, long long denominator); + +/* + * @brief + * result = arg + */ +static inline struct spl_fixed31_32 spl_fixpt_from_int(int arg) +{ + struct spl_fixed31_32 res; + + res.value = (long long) arg << FIXED31_32_BITS_PER_FRACTIONAL_PART; + + return res; +} + +/* + * @brief + * Unary operators + */ + +/* + * @brief + * result = -arg + */ +static inline struct spl_fixed31_32 spl_fixpt_neg(struct spl_fixed31_32 arg) +{ + struct spl_fixed31_32 res; + + res.value = -arg.value; + + return res; +} + +/* + * @brief + * result = abs(arg) := (arg >= 0) ? arg : -arg + */ +static inline struct spl_fixed31_32 spl_fixpt_abs(struct spl_fixed31_32 arg) +{ + if (arg.value < 0) + return spl_fixpt_neg(arg); + else + return arg; +} + +/* + * @brief + * Binary relational operators + */ + +/* + * @brief + * result = arg1 < arg2 + */ +static inline bool spl_fixpt_lt(struct spl_fixed31_32 arg1, struct spl_fixed31_32 arg2) +{ + return arg1.value < arg2.value; +} + +/* + * @brief + * result = arg1 <= arg2 + */ +static inline bool spl_fixpt_le(struct spl_fixed31_32 arg1, struct spl_fixed31_32 arg2) +{ + return arg1.value <= arg2.value; +} + +/* + * @brief + * result = arg1 == arg2 + */ +static inline bool spl_fixpt_eq(struct spl_fixed31_32 arg1, struct spl_fixed31_32 arg2) +{ + return arg1.value == arg2.value; +} + +/* + * @brief + * result = min(arg1, arg2) := (arg1 <= arg2) ? arg1 : arg2 + */ +static inline struct spl_fixed31_32 spl_fixpt_min(struct spl_fixed31_32 arg1, struct spl_fixed31_32 arg2) +{ + if (arg1.value <= arg2.value) + return arg1; + else + return arg2; +} + +/* + * @brief + * result = max(arg1, arg2) := (arg1 <= arg2) ? arg2 : arg1 + */ +static inline struct spl_fixed31_32 spl_fixpt_max(struct spl_fixed31_32 arg1, struct spl_fixed31_32 arg2) +{ + if (arg1.value <= arg2.value) + return arg2; + else + return arg1; +} + +/* + * @brief + * | min_value, when arg <= min_value + * result = | arg, when min_value < arg < max_value + * | max_value, when arg >= max_value + */ +static inline struct spl_fixed31_32 spl_fixpt_clamp( + struct spl_fixed31_32 arg, + struct spl_fixed31_32 min_value, + struct spl_fixed31_32 max_value) +{ + if (spl_fixpt_le(arg, min_value)) + return min_value; + else if (spl_fixpt_le(max_value, arg)) + return max_value; + else + return arg; +} + +/* + * @brief + * Binary shift operators + */ + +/* + * @brief + * result = arg << shift + */ +static inline struct spl_fixed31_32 spl_fixpt_shl(struct spl_fixed31_32 arg, unsigned char shift) +{ + ASSERT(((arg.value >= 0) && (arg.value <= LLONG_MAX >> shift)) || + ((arg.value < 0) && (arg.value >= ~(LLONG_MAX >> shift)))); + + arg.value = arg.value << shift; + + return arg; +} + +/* + * @brief + * result = arg >> shift + */ +static inline struct spl_fixed31_32 spl_fixpt_shr(struct spl_fixed31_32 arg, unsigned char shift) +{ + bool negative = arg.value < 0; + + if (negative) + arg.value = -arg.value; + arg.value = arg.value >> shift; + if (negative) + arg.value = -arg.value; + return arg; +} + +/* + * @brief + * Binary additive operators + */ + +/* + * @brief + * result = arg1 + arg2 + */ +static inline struct spl_fixed31_32 spl_fixpt_add(struct spl_fixed31_32 arg1, struct spl_fixed31_32 arg2) +{ + struct spl_fixed31_32 res; + + ASSERT(((arg1.value >= 0) && (LLONG_MAX - arg1.value >= arg2.value)) || + ((arg1.value < 0) && (LLONG_MIN - arg1.value <= arg2.value))); + + res.value = arg1.value + arg2.value; + + return res; +} + +/* + * @brief + * result = arg1 + arg2 + */ +static inline struct spl_fixed31_32 spl_fixpt_add_int(struct spl_fixed31_32 arg1, int arg2) +{ + return spl_fixpt_add(arg1, spl_fixpt_from_int(arg2)); +} + +/* + * @brief + * result = arg1 - arg2 + */ +static inline struct spl_fixed31_32 spl_fixpt_sub(struct spl_fixed31_32 arg1, struct spl_fixed31_32 arg2) +{ + struct spl_fixed31_32 res; + + ASSERT(((arg2.value >= 0) && (LLONG_MIN + arg2.value <= arg1.value)) || + ((arg2.value < 0) && (LLONG_MAX + arg2.value >= arg1.value))); + + res.value = arg1.value - arg2.value; + + return res; +} + +/* + * @brief + * result = arg1 - arg2 + */ +static inline struct spl_fixed31_32 spl_fixpt_sub_int(struct spl_fixed31_32 arg1, int arg2) +{ + return spl_fixpt_sub(arg1, spl_fixpt_from_int(arg2)); +} + + +/* + * @brief + * Binary multiplicative operators + */ + +/* + * @brief + * result = arg1 * arg2 + */ +struct spl_fixed31_32 spl_fixpt_mul(struct spl_fixed31_32 arg1, struct spl_fixed31_32 arg2); + + +/* + * @brief + * result = arg1 * arg2 + */ +static inline struct spl_fixed31_32 spl_fixpt_mul_int(struct spl_fixed31_32 arg1, int arg2) +{ + return spl_fixpt_mul(arg1, spl_fixpt_from_int(arg2)); +} + +/* + * @brief + * result = square(arg) := arg * arg + */ +struct spl_fixed31_32 spl_fixpt_sqr(struct spl_fixed31_32 arg); + +/* + * @brief + * result = arg1 / arg2 + */ +static inline struct spl_fixed31_32 spl_fixpt_div_int(struct spl_fixed31_32 arg1, long long arg2) +{ + return spl_fixpt_from_fraction(arg1.value, spl_fixpt_from_int((int)arg2).value); +} + +/* + * @brief + * result = arg1 / arg2 + */ +static inline struct spl_fixed31_32 spl_fixpt_div(struct spl_fixed31_32 arg1, struct spl_fixed31_32 arg2) +{ + return spl_fixpt_from_fraction(arg1.value, arg2.value); +} + +/* + * @brief + * Reciprocal function + */ + +/* + * @brief + * result = reciprocal(arg) := 1 / arg + * + * @note + * No special actions taken in case argument is zero. + */ +struct spl_fixed31_32 spl_fixpt_recip(struct spl_fixed31_32 arg); + +/* + * @brief + * Trigonometric functions + */ + +/* + * @brief + * result = sinc(arg) := sin(arg) / arg + * + * @note + * Argument specified in radians, + * internally it's normalized to [-2pi...2pi] range. + */ +struct spl_fixed31_32 spl_fixpt_sinc(struct spl_fixed31_32 arg); + +/* + * @brief + * result = sin(arg) + * + * @note + * Argument specified in radians, + * internally it's normalized to [-2pi...2pi] range. + */ +struct spl_fixed31_32 spl_fixpt_sin(struct spl_fixed31_32 arg); + +/* + * @brief + * result = cos(arg) + * + * @note + * Argument specified in radians + * and should be in [-2pi...2pi] range - + * passing arguments outside that range + * will cause incorrect result! + */ +struct spl_fixed31_32 spl_fixpt_cos(struct spl_fixed31_32 arg); + +/* + * @brief + * Transcendent functions + */ + +/* + * @brief + * result = exp(arg) + * + * @note + * Currently, function is verified for abs(arg) <= 1. + */ +struct spl_fixed31_32 spl_fixpt_exp(struct spl_fixed31_32 arg); + +/* + * @brief + * result = log(arg) + * + * @note + * Currently, abs(arg) should be less than 1. + * No normalization is done. + * Currently, no special actions taken + * in case of invalid argument(s). Take care! + */ +struct spl_fixed31_32 spl_fixpt_log(struct spl_fixed31_32 arg); + +/* + * @brief + * Power function + */ + +/* + * @brief + * result = pow(arg1, arg2) + * + * @note + * Currently, abs(arg1) should be less than 1. Take care! + */ +static inline struct spl_fixed31_32 spl_fixpt_pow(struct spl_fixed31_32 arg1, struct spl_fixed31_32 arg2) +{ + if (arg1.value == 0) + return arg2.value == 0 ? spl_fixpt_one : spl_fixpt_zero; + + return spl_fixpt_exp( + spl_fixpt_mul( + spl_fixpt_log(arg1), + arg2)); +} + +/* + * @brief + * Rounding functions + */ + +/* + * @brief + * result = floor(arg) := greatest integer lower than or equal to arg + */ +static inline int spl_fixpt_floor(struct spl_fixed31_32 arg) +{ + unsigned long long arg_value = arg.value > 0 ? arg.value : -arg.value; + + if (arg.value >= 0) + return (int)(arg_value >> FIXED31_32_BITS_PER_FRACTIONAL_PART); + else + return -(int)(arg_value >> FIXED31_32_BITS_PER_FRACTIONAL_PART); +} + +/* + * @brief + * result = round(arg) := integer nearest to arg + */ +static inline int spl_fixpt_round(struct spl_fixed31_32 arg) +{ + unsigned long long arg_value = arg.value > 0 ? arg.value : -arg.value; + + const long long summand = spl_fixpt_half.value; + + ASSERT(LLONG_MAX - (long long)arg_value >= summand); + + arg_value += summand; + + if (arg.value >= 0) + return (int)(arg_value >> FIXED31_32_BITS_PER_FRACTIONAL_PART); + else + return -(int)(arg_value >> FIXED31_32_BITS_PER_FRACTIONAL_PART); +} + +/* + * @brief + * result = ceil(arg) := lowest integer greater than or equal to arg + */ +static inline int spl_fixpt_ceil(struct spl_fixed31_32 arg) +{ + unsigned long long arg_value = arg.value > 0 ? arg.value : -arg.value; + + const long long summand = spl_fixpt_one.value - + spl_fixpt_epsilon.value; + + ASSERT(LLONG_MAX - (long long)arg_value >= summand); + + arg_value += summand; + + if (arg.value >= 0) + return (int)(arg_value >> FIXED31_32_BITS_PER_FRACTIONAL_PART); + else + return -(int)(arg_value >> FIXED31_32_BITS_PER_FRACTIONAL_PART); +} + +/* the following two function are used in scaler hw programming to convert fixed + * point value to format 2 bits from integer part and 19 bits from fractional + * part. The same applies for u0d19, 0 bits from integer part and 19 bits from + * fractional + */ + +unsigned int spl_fixpt_u4d19(struct spl_fixed31_32 arg); + +unsigned int spl_fixpt_u3d19(struct spl_fixed31_32 arg); + +unsigned int spl_fixpt_u2d19(struct spl_fixed31_32 arg); + +unsigned int spl_fixpt_u0d19(struct spl_fixed31_32 arg); + +unsigned int spl_fixpt_clamp_u0d14(struct spl_fixed31_32 arg); + +unsigned int spl_fixpt_clamp_u0d10(struct spl_fixed31_32 arg); + +int spl_fixpt_s4d19(struct spl_fixed31_32 arg); + +static inline struct spl_fixed31_32 spl_fixpt_truncate(struct spl_fixed31_32 arg, unsigned int frac_bits) +{ + bool negative = arg.value < 0; + + if (frac_bits >= FIXED31_32_BITS_PER_FRACTIONAL_PART) { + ASSERT(frac_bits == FIXED31_32_BITS_PER_FRACTIONAL_PART); + return arg; + } + + if (negative) + arg.value = -arg.value; + arg.value &= (~0ULL) << (FIXED31_32_BITS_PER_FRACTIONAL_PART - frac_bits); + if (negative) + arg.value = -arg.value; + return arg; +} + +struct spl_fixed31_32 spl_fixpt_from_ux_dy(unsigned int value, unsigned int integer_bits, unsigned int fractional_bits); +struct spl_fixed31_32 spl_fixpt_from_int_dy(unsigned int int_value, + unsigned int frac_value, + unsigned int integer_bits, + unsigned int fractional_bits); + +#endif diff --git a/drivers/gpu/drm/amd/display/dc/spl/spl_os_types.h b/drivers/gpu/drm/amd/display/dc/spl/spl_os_types.h index 7ebea91c84f6..709706ed4f2c 100644 --- a/drivers/gpu/drm/amd/display/dc/spl/spl_os_types.h +++ b/drivers/gpu/drm/amd/display/dc/spl/spl_os_types.h @@ -1,28 +1,7 @@ -/* - * Copyright 2012-16 Advanced Micro Devices, Inc. - * Copyright 2019 Raptor Engineering, LLC - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ +/* SPDX-License-Identifier: MIT */ + +/* Copyright 2024 Advanced Micro Devices, Inc. */ +/* Copyright 2019 Raptor Engineering, LLC */ #ifndef _SPL_OS_TYPES_H_ #define _SPL_OS_TYPES_H_ @@ -39,7 +18,6 @@ * general debug capabilities * */ -// TODO: need backport #define SPL_BREAK_TO_DEBUGGER() ASSERT(0) static inline uint64_t spl_div_u64_rem(uint64_t dividend, uint32_t divisor, uint32_t *remainder) diff --git a/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h b/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h index c5f99cbff0b6..e20c220aa8b4 100644 --- a/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h +++ b/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h @@ -111,7 +111,7 @@ #define DMUB_MAX_PHANTOM_PLANES ((DMUB_MAX_PLANES) / 2) /* Trace buffer offset for entry */ -#define TRACE_BUFFER_ENTRY_OFFSET 16 +#define TRACE_BUFFER_ENTRY_OFFSET 16 /** * Maximum number of dirty rects supported by FW. @@ -1879,7 +1879,12 @@ enum dmub_cmd_idle_opt_type { /** * DCN hardware notify idle. */ - DMUB_CMD__IDLE_OPT_DCN_NOTIFY_IDLE = 2 + DMUB_CMD__IDLE_OPT_DCN_NOTIFY_IDLE = 2, + + /** + * DCN hardware notify power state. + */ + DMUB_CMD__IDLE_OPT_SET_DC_POWER_STATE = 3, }; /** @@ -1907,6 +1912,33 @@ struct dmub_rb_cmd_idle_opt_dcn_notify_idle { }; /** + * enum dmub_idle_opt_dc_power_state - DC power states. + */ +enum dmub_idle_opt_dc_power_state { + DMUB_IDLE_OPT_DC_POWER_STATE_UNKNOWN = 0, + DMUB_IDLE_OPT_DC_POWER_STATE_D0 = 1, + DMUB_IDLE_OPT_DC_POWER_STATE_D1 = 2, + DMUB_IDLE_OPT_DC_POWER_STATE_D2 = 4, + DMUB_IDLE_OPT_DC_POWER_STATE_D3 = 8, +}; + +/** + * struct dmub_idle_opt_set_dc_power_state_data - Data passed to FW in a DMUB_CMD__IDLE_OPT_SET_DC_POWER_STATE command. + */ +struct dmub_idle_opt_set_dc_power_state_data { + uint8_t power_state; /**< power state */ + uint8_t pad[3]; /**< padding */ +}; + +/** + * struct dmub_rb_cmd_idle_opt_set_dc_power_state - Data passed to FW in a DMUB_CMD__IDLE_OPT_SET_DC_POWER_STATE command. + */ +struct dmub_rb_cmd_idle_opt_set_dc_power_state { + struct dmub_cmd_header header; /**< header */ + struct dmub_idle_opt_set_dc_power_state_data data; +}; + +/** * struct dmub_clocks - Clock update notification. */ struct dmub_clocks { @@ -5298,6 +5330,10 @@ union dmub_rb_cmd { * Definition of a DMUB_CMD__IDLE_OPT_DCN_NOTIFY_IDLE command. */ struct dmub_rb_cmd_idle_opt_dcn_notify_idle idle_opt_notify_idle; + /** + * Definition of a DMUB_CMD__IDLE_OPT_SET_DC_POWER_STATE command. + */ + struct dmub_rb_cmd_idle_opt_set_dc_power_state idle_opt_set_dc_power_state; /* * Definition of a DMUB_CMD__REPLAY_COPY_SETTINGS command. */ diff --git a/drivers/gpu/drm/amd/include/amd_shared.h b/drivers/gpu/drm/amd/include/amd_shared.h index f5b725f10a7c..745fd052840d 100644 --- a/drivers/gpu/drm/amd/include/amd_shared.h +++ b/drivers/gpu/drm/amd/include/amd_shared.h @@ -61,7 +61,7 @@ enum amd_apu_flags { * acquires the list of IP blocks for the GPU in use on initialization. * It can then operate on this list to perform standard driver operations * such as: init, fini, suspend, resume, etc. -* +* * * IP block implementations are named using the following convention: * <functionality>_v<version> (E.g.: gfx_v6_0). @@ -251,19 +251,92 @@ enum DC_FEATURE_MASK { DC_REPLAY_MASK = (1 << 9), //0x200, disabled by default for dcn < 3.1.4 }; +/** + * enum DC_DEBUG_MASK - Bits that are useful for debugging the Display Core IP + */ enum DC_DEBUG_MASK { + /** + * @DC_DISABLE_PIPE_SPLIT: If set, disable pipe-splitting + */ DC_DISABLE_PIPE_SPLIT = 0x1, + + /** + * @DC_DISABLE_STUTTER: If set, disable memory stutter mode + */ DC_DISABLE_STUTTER = 0x2, + + /** + * @DC_DISABLE_DSC: If set, disable display stream compression + */ DC_DISABLE_DSC = 0x4, + + /** + * @DC_DISABLE_CLOCK_GATING: If set, disable clock gating optimizations + */ DC_DISABLE_CLOCK_GATING = 0x8, + + /** + * @DC_DISABLE_PSR: If set, disable Panel self refresh v1 and PSR-SU + */ DC_DISABLE_PSR = 0x10, + + /** + * @DC_FORCE_SUBVP_MCLK_SWITCH: If set, force mclk switch in subvp, even + * if mclk switch in vblank is possible + */ DC_FORCE_SUBVP_MCLK_SWITCH = 0x20, + + /** + * @DC_DISABLE_MPO: If set, disable multi-plane offloading + */ DC_DISABLE_MPO = 0x40, + + /** + * @DC_ENABLE_DPIA_TRACE: If set, enable trace logging for DPIA + */ DC_ENABLE_DPIA_TRACE = 0x80, + + /** + * @DC_ENABLE_DML2: If set, force usage of DML2, even if the DCN version + * does not default to it. + */ DC_ENABLE_DML2 = 0x100, + + /** + * @DC_DISABLE_PSR_SU: If set, disable PSR SU + */ DC_DISABLE_PSR_SU = 0x200, + + /** + * @DC_DISABLE_REPLAY: If set, disable Panel Replay + */ DC_DISABLE_REPLAY = 0x400, + + /** + * @DC_DISABLE_IPS: If set, disable all Idle Power States, all the time. + * If more than one IPS debug bit is set, the lowest bit takes + * precedence. For example, if DC_FORCE_IPS_ENABLE and + * DC_DISABLE_IPS_DYNAMIC are set, then DC_DISABLE_IPS_DYNAMIC takes + * precedence. + */ DC_DISABLE_IPS = 0x800, + + /** + * @DC_DISABLE_IPS_DYNAMIC: If set, disable all IPS, all the time, + * *except* when driver goes into suspend. + */ + DC_DISABLE_IPS_DYNAMIC = 0x1000, + + /** + * @DC_DISABLE_IPS2_DYNAMIC: If set, disable IPS2 (IPS1 allowed) if + * there is an enabled display. Otherwise, enable all IPS. + */ + DC_DISABLE_IPS2_DYNAMIC = 0x2000, + + /** + * @DC_FORCE_IPS_ENABLE: If set, force enable all IPS, all the time. + */ + DC_FORCE_IPS_ENABLE = 0x4000, }; enum amd_dpm_forced_level; diff --git a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c index 9d7454b3c314..bb3bc68dfc39 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c +++ b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c @@ -1257,7 +1257,6 @@ static int smu_sw_init(void *handle) atomic_set(&smu->smu_power.power_gate.vpe_gated, 1); atomic_set(&smu->smu_power.power_gate.umsch_mm_gated, 1); - smu->workload_mask = 1 << smu->workload_prority[PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT]; smu->workload_prority[PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT] = 0; smu->workload_prority[PP_SMC_POWER_PROFILE_FULLSCREEN3D] = 1; smu->workload_prority[PP_SMC_POWER_PROFILE_POWERSAVING] = 2; @@ -1265,6 +1264,7 @@ static int smu_sw_init(void *handle) smu->workload_prority[PP_SMC_POWER_PROFILE_VR] = 4; smu->workload_prority[PP_SMC_POWER_PROFILE_COMPUTE] = 5; smu->workload_prority[PP_SMC_POWER_PROFILE_CUSTOM] = 6; + smu->workload_mask = 1 << smu->workload_prority[PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT]; smu->workload_setting[0] = PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT; smu->workload_setting[1] = PP_SMC_POWER_PROFILE_FULLSCREEN3D; @@ -2224,8 +2224,9 @@ static int smu_bump_power_profile_mode(struct smu_context *smu, } static int smu_adjust_power_state_dynamic(struct smu_context *smu, - enum amd_dpm_forced_level level, - bool skip_display_settings) + enum amd_dpm_forced_level level, + bool skip_display_settings, + bool force_update) { int ret = 0; int index = 0; @@ -2254,7 +2255,7 @@ static int smu_adjust_power_state_dynamic(struct smu_context *smu, } } - if (smu_dpm_ctx->dpm_level != level) { + if (force_update || smu_dpm_ctx->dpm_level != level) { ret = smu_asic_set_performance_level(smu, level); if (ret) { dev_err(smu->adev->dev, "Failed to set performance level!"); @@ -2271,7 +2272,7 @@ static int smu_adjust_power_state_dynamic(struct smu_context *smu, index = index > 0 && index <= WORKLOAD_POLICY_MAX ? index - 1 : 0; workload[0] = smu->workload_setting[index]; - if (smu->power_profile_mode != workload[0]) + if (force_update || smu->power_profile_mode != workload[0]) smu_bump_power_profile_mode(smu, workload, 0); } @@ -2292,11 +2293,13 @@ static int smu_handle_task(struct smu_context *smu, ret = smu_pre_display_config_changed(smu); if (ret) return ret; - ret = smu_adjust_power_state_dynamic(smu, level, false); + ret = smu_adjust_power_state_dynamic(smu, level, false, false); break; case AMD_PP_TASK_COMPLETE_INIT: + ret = smu_adjust_power_state_dynamic(smu, level, true, true); + break; case AMD_PP_TASK_READJUST_POWER_STATE: - ret = smu_adjust_power_state_dynamic(smu, level, true); + ret = smu_adjust_power_state_dynamic(smu, level, true, false); break; default: break; diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c index a7d0231727e8..7bc95c404377 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c @@ -2378,7 +2378,7 @@ static int smu_v13_0_7_get_power_profile_mode(struct smu_context *smu, char *buf size += sysfs_emit_at(buf, size, " "); for (i = 0; i <= PP_SMC_POWER_PROFILE_WINDOW3D; i++) - size += sysfs_emit_at(buf, size, "%-14s%s", amdgpu_pp_profile_name[i], + size += sysfs_emit_at(buf, size, "%d %-14s%s", i, amdgpu_pp_profile_name[i], (i == smu->power_profile_mode) ? "* " : " "); size += sysfs_emit_at(buf, size, "\n"); @@ -2408,7 +2408,7 @@ static int smu_v13_0_7_get_power_profile_mode(struct smu_context *smu, char *buf do { \ size += sysfs_emit_at(buf, size, "%-30s", #field); \ for (j = 0; j <= PP_SMC_POWER_PROFILE_WINDOW3D; j++) \ - size += sysfs_emit_at(buf, size, "%-16d", activity_monitor_external[j].DpmActivityMonitorCoeffInt.field); \ + size += sysfs_emit_at(buf, size, "%-18d", activity_monitor_external[j].DpmActivityMonitorCoeffInt.field); \ size += sysfs_emit_at(buf, size, "\n"); \ } while (0) diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_kms.c b/drivers/gpu/drm/arm/display/komeda/komeda_kms.c index fe46b0ebefea..e5eb5d672bcd 100644 --- a/drivers/gpu/drm/arm/display/komeda/komeda_kms.c +++ b/drivers/gpu/drm/arm/display/komeda/komeda_kms.c @@ -160,6 +160,7 @@ static int komeda_crtc_normalize_zpos(struct drm_crtc *crtc, struct drm_plane *plane; struct list_head zorder_list; int order = 0, err; + u32 slave_zpos = 0; DRM_DEBUG_ATOMIC("[CRTC:%d:%s] calculating normalized zpos values\n", crtc->base.id, crtc->name); @@ -199,10 +200,13 @@ static int komeda_crtc_normalize_zpos(struct drm_crtc *crtc, plane_st->zpos, plane_st->normalized_zpos); /* calculate max slave zorder */ - if (has_bit(drm_plane_index(plane), kcrtc->slave_planes)) + if (has_bit(drm_plane_index(plane), kcrtc->slave_planes)) { + slave_zpos = plane_st->normalized_zpos; + if (to_kplane_st(plane_st)->layer_split) + slave_zpos++; kcrtc_st->max_slave_zorder = - max(plane_st->normalized_zpos, - kcrtc_st->max_slave_zorder); + max(slave_zpos, kcrtc_st->max_slave_zorder); + } } crtc_st->zpos_changed = true; diff --git a/drivers/gpu/drm/bridge/Kconfig b/drivers/gpu/drm/bridge/Kconfig index c0ab5b620b57..683cb33805b2 100644 --- a/drivers/gpu/drm/bridge/Kconfig +++ b/drivers/gpu/drm/bridge/Kconfig @@ -397,6 +397,7 @@ config DRM_TI_SN65DSI86 depends on OF select DRM_DISPLAY_DP_HELPER select DRM_DISPLAY_HELPER + select DRM_BRIDGE_CONNECTOR select DRM_KMS_HELPER select REGMAP_I2C select DRM_PANEL diff --git a/drivers/gpu/drm/display/Kconfig b/drivers/gpu/drm/display/Kconfig index a2e42014ffe0..0f07cf1483ff 100644 --- a/drivers/gpu/drm/display/Kconfig +++ b/drivers/gpu/drm/display/Kconfig @@ -1,19 +1,26 @@ # SPDX-License-Identifier: MIT +config DRM_DISPLAY_DP_AUX_BUS + tristate + depends on DRM + depends on OF || COMPILE_TEST + config DRM_DISPLAY_HELPER tristate depends on DRM help DRM helpers for display adapters. -config DRM_DISPLAY_DP_AUX_BUS - tristate - depends on DRM - depends on OF || COMPILE_TEST +if DRM_DISPLAY_HELPER + +config DRM_BRIDGE_CONNECTOR + bool + select DRM_DISPLAY_HDMI_STATE_HELPER + help + DRM connector implementation terminating DRM bridge chains. config DRM_DISPLAY_DP_AUX_CEC bool "Enable DisplayPort CEC-Tunneling-over-AUX HDMI support" - depends on DRM && DRM_DISPLAY_HELPER select DRM_DISPLAY_DP_HELPER select CEC_CORE help @@ -25,7 +32,6 @@ config DRM_DISPLAY_DP_AUX_CEC config DRM_DISPLAY_DP_AUX_CHARDEV bool "DRM DP AUX Interface" - depends on DRM && DRM_DISPLAY_HELPER select DRM_DISPLAY_DP_HELPER help Choose this option to enable a /dev/drm_dp_auxN node that allows to @@ -34,7 +40,6 @@ config DRM_DISPLAY_DP_AUX_CHARDEV config DRM_DISPLAY_DP_HELPER bool - depends on DRM_DISPLAY_HELPER help DRM display helpers for DisplayPort. @@ -67,19 +72,18 @@ config DRM_DISPLAY_DSC_HELPER config DRM_DISPLAY_HDCP_HELPER bool - depends on DRM_DISPLAY_HELPER help DRM display helpers for HDCP. config DRM_DISPLAY_HDMI_HELPER bool - depends on DRM_DISPLAY_HELPER help DRM display helpers for HDMI. config DRM_DISPLAY_HDMI_STATE_HELPER bool - depends on DRM_DISPLAY_HELPER select DRM_DISPLAY_HDMI_HELPER help DRM KMS state helpers for HDMI. + +endif # DRM_DISPLAY_HELPER diff --git a/drivers/gpu/drm/display/Makefile b/drivers/gpu/drm/display/Makefile index a023f72fa139..629c834c3192 100644 --- a/drivers/gpu/drm/display/Makefile +++ b/drivers/gpu/drm/display/Makefile @@ -3,6 +3,8 @@ obj-$(CONFIG_DRM_DISPLAY_DP_AUX_BUS) += drm_dp_aux_bus.o drm_display_helper-y := drm_display_helper_mod.o +drm_display_helper-$(CONFIG_DRM_BRIDGE_CONNECTOR) += \ + drm_bridge_connector.o drm_display_helper-$(CONFIG_DRM_DISPLAY_DP_HELPER) += \ drm_dp_dual_mode_helper.o \ drm_dp_helper.o \ diff --git a/drivers/gpu/drm/drm_bridge_connector.c b/drivers/gpu/drm/display/drm_bridge_connector.c index a4fbf1eb7ac5..3da5b8bf8259 100644 --- a/drivers/gpu/drm/drm_bridge_connector.c +++ b/drivers/gpu/drm/display/drm_bridge_connector.c @@ -216,8 +216,19 @@ static void drm_bridge_connector_debugfs_init(struct drm_connector *connector, } } +static void drm_bridge_connector_reset(struct drm_connector *connector) +{ + struct drm_bridge_connector *bridge_connector = + to_drm_bridge_connector(connector); + + drm_atomic_helper_connector_reset(connector); + if (bridge_connector->bridge_hdmi) + __drm_atomic_helper_connector_hdmi_reset(connector, + connector->state); +} + static const struct drm_connector_funcs drm_bridge_connector_funcs = { - .reset = drm_atomic_helper_connector_reset, + .reset = drm_bridge_connector_reset, .detect = drm_bridge_connector_detect, .fill_modes = drm_helper_probe_single_connector_modes, .atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state, diff --git a/drivers/gpu/drm/drm_fbdev_dma.c b/drivers/gpu/drm/drm_fbdev_dma.c index 7ef5a48c8029..b0602c4f3628 100644 --- a/drivers/gpu/drm/drm_fbdev_dma.c +++ b/drivers/gpu/drm/drm_fbdev_dma.c @@ -36,20 +36,11 @@ static int drm_fbdev_dma_fb_release(struct fb_info *info, int user) return 0; } -FB_GEN_DEFAULT_DEFERRED_DMAMEM_OPS(drm_fbdev_dma, - drm_fb_helper_damage_range, - drm_fb_helper_damage_area); - static int drm_fbdev_dma_fb_mmap(struct fb_info *info, struct vm_area_struct *vma) { struct drm_fb_helper *fb_helper = info->par; - struct drm_framebuffer *fb = fb_helper->fb; - struct drm_gem_dma_object *dma = drm_fb_dma_get_gem_obj(fb, 0); - if (!dma->map_noncoherent) - vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot); - - return fb_deferred_io_mmap(info, vma); + return drm_gem_prime_mmap(fb_helper->buffer->gem, vma); } static void drm_fbdev_dma_fb_destroy(struct fb_info *info) @@ -73,10 +64,37 @@ static const struct fb_ops drm_fbdev_dma_fb_ops = { .owner = THIS_MODULE, .fb_open = drm_fbdev_dma_fb_open, .fb_release = drm_fbdev_dma_fb_release, + __FB_DEFAULT_DMAMEM_OPS_RDWR, + DRM_FB_HELPER_DEFAULT_OPS, + __FB_DEFAULT_DMAMEM_OPS_DRAW, + .fb_mmap = drm_fbdev_dma_fb_mmap, + .fb_destroy = drm_fbdev_dma_fb_destroy, +}; + +FB_GEN_DEFAULT_DEFERRED_DMAMEM_OPS(drm_fbdev_dma, + drm_fb_helper_damage_range, + drm_fb_helper_damage_area); + +static int drm_fbdev_dma_deferred_fb_mmap(struct fb_info *info, struct vm_area_struct *vma) +{ + struct drm_fb_helper *fb_helper = info->par; + struct drm_framebuffer *fb = fb_helper->fb; + struct drm_gem_dma_object *dma = drm_fb_dma_get_gem_obj(fb, 0); + + if (!dma->map_noncoherent) + vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot); + + return fb_deferred_io_mmap(info, vma); +} + +static const struct fb_ops drm_fbdev_dma_deferred_fb_ops = { + .owner = THIS_MODULE, + .fb_open = drm_fbdev_dma_fb_open, + .fb_release = drm_fbdev_dma_fb_release, __FB_DEFAULT_DEFERRED_OPS_RDWR(drm_fbdev_dma), DRM_FB_HELPER_DEFAULT_OPS, __FB_DEFAULT_DEFERRED_OPS_DRAW(drm_fbdev_dma), - .fb_mmap = drm_fbdev_dma_fb_mmap, + .fb_mmap = drm_fbdev_dma_deferred_fb_mmap, .fb_destroy = drm_fbdev_dma_fb_destroy, }; @@ -89,6 +107,7 @@ static int drm_fbdev_dma_helper_fb_probe(struct drm_fb_helper *fb_helper, { struct drm_client_dev *client = &fb_helper->client; struct drm_device *dev = fb_helper->dev; + bool use_deferred_io = false; struct drm_client_buffer *buffer; struct drm_gem_dma_object *dma_obj; struct drm_framebuffer *fb; @@ -111,6 +130,15 @@ static int drm_fbdev_dma_helper_fb_probe(struct drm_fb_helper *fb_helper, fb = buffer->fb; + /* + * Deferred I/O requires struct page for framebuffer memory, + * which is not guaranteed for all DMA ranges. We thus only + * install deferred I/O if we have a framebuffer that requires + * it. + */ + if (fb->funcs->dirty) + use_deferred_io = true; + ret = drm_client_buffer_vmap(buffer, &map); if (ret) { goto err_drm_client_buffer_delete; @@ -130,7 +158,10 @@ static int drm_fbdev_dma_helper_fb_probe(struct drm_fb_helper *fb_helper, drm_fb_helper_fill_info(info, fb_helper, sizes); - info->fbops = &drm_fbdev_dma_fb_ops; + if (use_deferred_io) + info->fbops = &drm_fbdev_dma_deferred_fb_ops; + else + info->fbops = &drm_fbdev_dma_fb_ops; /* screen */ info->flags |= FBINFO_VIRTFB; /* system memory */ @@ -144,14 +175,28 @@ static int drm_fbdev_dma_helper_fb_probe(struct drm_fb_helper *fb_helper, } info->fix.smem_len = info->screen_size; - /* deferred I/O */ - fb_helper->fbdefio.delay = HZ / 20; - fb_helper->fbdefio.deferred_io = drm_fb_helper_deferred_io; + /* + * Only set up deferred I/O if the screen buffer supports + * it. If this disagrees with the previous test for ->dirty, + * mmap on the /dev/fb file might not work correctly. + */ + if (!is_vmalloc_addr(info->screen_buffer) && info->fix.smem_start) { + unsigned long pfn = info->fix.smem_start >> PAGE_SHIFT; - info->fbdefio = &fb_helper->fbdefio; - ret = fb_deferred_io_init(info); - if (ret) - goto err_drm_fb_helper_release_info; + if (drm_WARN_ON(dev, !pfn_to_page(pfn))) + use_deferred_io = false; + } + + /* deferred I/O */ + if (use_deferred_io) { + fb_helper->fbdefio.delay = HZ / 20; + fb_helper->fbdefio.deferred_io = drm_fb_helper_deferred_io; + + info->fbdefio = &fb_helper->fbdefio; + ret = fb_deferred_io_init(info); + if (ret) + goto err_drm_fb_helper_release_info; + } return 0; diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.h b/drivers/gpu/drm/exynos/exynos_drm_drv.h index 81d501efd013..23646e55f142 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_drv.h +++ b/drivers/gpu/drm/exynos/exynos_drm_drv.h @@ -254,10 +254,6 @@ static inline int exynos_drm_check_fimc_device(struct device *dev) } #endif -int exynos_atomic_commit(struct drm_device *dev, struct drm_atomic_state *state, - bool nonblock); - - extern struct platform_driver fimd_driver; extern struct platform_driver exynos5433_decon_driver; extern struct platform_driver decon_driver; diff --git a/drivers/gpu/drm/exynos/exynos_drm_fimc.c b/drivers/gpu/drm/exynos/exynos_drm_fimc.c index 142184c8c3bc..4d7ea65b7dd8 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_fimc.c +++ b/drivers/gpu/drm/exynos/exynos_drm_fimc.c @@ -1125,7 +1125,7 @@ static void fimc_abort(struct exynos_drm_ipp *ipp, } } -static struct exynos_drm_ipp_funcs ipp_funcs = { +static const struct exynos_drm_ipp_funcs ipp_funcs = { .commit = fimc_commit, .abort = fimc_abort, }; diff --git a/drivers/gpu/drm/exynos/exynos_drm_gsc.c b/drivers/gpu/drm/exynos/exynos_drm_gsc.c index 1b111e2c3347..59fa22050717 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_gsc.c +++ b/drivers/gpu/drm/exynos/exynos_drm_gsc.c @@ -1162,7 +1162,7 @@ static void gsc_abort(struct exynos_drm_ipp *ipp, } } -static struct exynos_drm_ipp_funcs ipp_funcs = { +static const struct exynos_drm_ipp_funcs ipp_funcs = { .commit = gsc_commit, .abort = gsc_abort, }; @@ -1174,7 +1174,7 @@ static int gsc_bind(struct device *dev, struct device *master, void *data) struct exynos_drm_ipp *ipp = &ctx->ipp; ctx->drm_dev = drm_dev; - ctx->drm_dev = drm_dev; + ipp->drm_dev = drm_dev; exynos_drm_register_dma(drm_dev, dev, &ctx->dma_priv); exynos_drm_ipp_register(dev, ipp, &ipp_funcs, diff --git a/drivers/gpu/drm/exynos/exynos_drm_scaler.c b/drivers/gpu/drm/exynos/exynos_drm_scaler.c index a9d469896824..2788105ac780 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_scaler.c +++ b/drivers/gpu/drm/exynos/exynos_drm_scaler.c @@ -403,7 +403,7 @@ static int scaler_commit(struct exynos_drm_ipp *ipp, return 0; } -static struct exynos_drm_ipp_funcs ipp_funcs = { +static const struct exynos_drm_ipp_funcs ipp_funcs = { .commit = scaler_commit, }; diff --git a/drivers/gpu/drm/i915/display/g4x_dp.c b/drivers/gpu/drm/i915/display/g4x_dp.c index c2128b46bdbd..526c8c4d7b53 100644 --- a/drivers/gpu/drm/i915/display/g4x_dp.c +++ b/drivers/gpu/drm/i915/display/g4x_dp.c @@ -89,6 +89,7 @@ void g4x_dp_set_clock(struct intel_encoder *encoder, static void intel_dp_prepare(struct intel_encoder *encoder, const struct intel_crtc_state *pipe_config) { + struct intel_display *display = to_intel_display(encoder); struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_dp *intel_dp = enc_to_intel_dp(encoder); enum port port = encoder->port; @@ -118,7 +119,7 @@ static void intel_dp_prepare(struct intel_encoder *encoder, /* Preserve the BIOS-computed detected bit. This is * supposed to be read-only. */ - intel_dp->DP = intel_de_read(dev_priv, intel_dp->output_reg) & DP_DETECTED; + intel_dp->DP = intel_de_read(display, intel_dp->output_reg) & DP_DETECTED; /* Handle DP bits in common between all three register formats */ intel_dp->DP |= DP_VOLTAGE_0_4 | DP_PRE_EMPHASIS_0; @@ -140,7 +141,7 @@ static void intel_dp_prepare(struct intel_encoder *encoder, } else if (HAS_PCH_CPT(dev_priv) && port != PORT_A) { intel_dp->DP |= DP_LINK_TRAIN_OFF_CPT; - intel_de_rmw(dev_priv, TRANS_DP_CTL(crtc->pipe), + intel_de_rmw(display, TRANS_DP_CTL(crtc->pipe), TRANS_DP_ENH_FRAMING, pipe_config->enhanced_framing ? TRANS_DP_ENH_FRAMING : 0); @@ -166,9 +167,10 @@ static void intel_dp_prepare(struct intel_encoder *encoder, static void assert_dp_port(struct intel_dp *intel_dp, bool state) { + struct intel_display *display = to_intel_display(intel_dp); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); struct drm_i915_private *dev_priv = to_i915(dig_port->base.base.dev); - bool cur_state = intel_de_read(dev_priv, intel_dp->output_reg) & DP_PORT_EN; + bool cur_state = intel_de_read(display, intel_dp->output_reg) & DP_PORT_EN; I915_STATE_WARN(dev_priv, cur_state != state, "[ENCODER:%d:%s] state assertion failure (expected %s, current %s)\n", @@ -179,7 +181,8 @@ static void assert_dp_port(struct intel_dp *intel_dp, bool state) static void assert_edp_pll(struct drm_i915_private *dev_priv, bool state) { - bool cur_state = intel_de_read(dev_priv, DP_A) & DP_PLL_ENABLE; + struct intel_display *display = &dev_priv->display; + bool cur_state = intel_de_read(display, DP_A) & DP_PLL_ENABLE; I915_STATE_WARN(dev_priv, cur_state != state, "eDP PLL state assertion failure (expected %s, current %s)\n", @@ -191,6 +194,7 @@ static void assert_edp_pll(struct drm_i915_private *dev_priv, bool state) static void ilk_edp_pll_on(struct intel_dp *intel_dp, const struct intel_crtc_state *pipe_config) { + struct intel_display *display = to_intel_display(intel_dp); struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc); struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); @@ -198,7 +202,7 @@ static void ilk_edp_pll_on(struct intel_dp *intel_dp, assert_dp_port_disabled(intel_dp); assert_edp_pll_disabled(dev_priv); - drm_dbg_kms(&dev_priv->drm, "enabling eDP PLL for clock %d\n", + drm_dbg_kms(display->drm, "enabling eDP PLL for clock %d\n", pipe_config->port_clock); intel_dp->DP &= ~DP_PLL_FREQ_MASK; @@ -208,8 +212,8 @@ static void ilk_edp_pll_on(struct intel_dp *intel_dp, else intel_dp->DP |= DP_PLL_FREQ_270MHZ; - intel_de_write(dev_priv, DP_A, intel_dp->DP); - intel_de_posting_read(dev_priv, DP_A); + intel_de_write(display, DP_A, intel_dp->DP); + intel_de_posting_read(display, DP_A); udelay(500); /* @@ -223,14 +227,15 @@ static void ilk_edp_pll_on(struct intel_dp *intel_dp, intel_dp->DP |= DP_PLL_ENABLE; - intel_de_write(dev_priv, DP_A, intel_dp->DP); - intel_de_posting_read(dev_priv, DP_A); + intel_de_write(display, DP_A, intel_dp->DP); + intel_de_posting_read(display, DP_A); udelay(200); } static void ilk_edp_pll_off(struct intel_dp *intel_dp, const struct intel_crtc_state *old_crtc_state) { + struct intel_display *display = to_intel_display(intel_dp); struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc); struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); @@ -238,22 +243,23 @@ static void ilk_edp_pll_off(struct intel_dp *intel_dp, assert_dp_port_disabled(intel_dp); assert_edp_pll_enabled(dev_priv); - drm_dbg_kms(&dev_priv->drm, "disabling eDP PLL\n"); + drm_dbg_kms(display->drm, "disabling eDP PLL\n"); intel_dp->DP &= ~DP_PLL_ENABLE; - intel_de_write(dev_priv, DP_A, intel_dp->DP); - intel_de_posting_read(dev_priv, DP_A); + intel_de_write(display, DP_A, intel_dp->DP); + intel_de_posting_read(display, DP_A); udelay(200); } static bool cpt_dp_port_selected(struct drm_i915_private *dev_priv, enum port port, enum pipe *pipe) { + struct intel_display *display = &dev_priv->display; enum pipe p; - for_each_pipe(dev_priv, p) { - u32 val = intel_de_read(dev_priv, TRANS_DP_CTL(p)); + for_each_pipe(display, p) { + u32 val = intel_de_read(display, TRANS_DP_CTL(p)); if ((val & TRANS_DP_PORT_SEL_MASK) == TRANS_DP_PORT_SEL(port)) { *pipe = p; @@ -261,7 +267,7 @@ static bool cpt_dp_port_selected(struct drm_i915_private *dev_priv, } } - drm_dbg_kms(&dev_priv->drm, "No pipe for DP port %c found\n", + drm_dbg_kms(display->drm, "No pipe for DP port %c found\n", port_name(port)); /* must initialize pipe to something for the asserts */ @@ -274,10 +280,11 @@ bool g4x_dp_port_enabled(struct drm_i915_private *dev_priv, i915_reg_t dp_reg, enum port port, enum pipe *pipe) { + struct intel_display *display = &dev_priv->display; bool ret; u32 val; - val = intel_de_read(dev_priv, dp_reg); + val = intel_de_read(display, dp_reg); ret = val & DP_PORT_EN; @@ -333,6 +340,7 @@ static void g4x_dp_get_m_n(struct intel_crtc_state *crtc_state) static void intel_dp_get_config(struct intel_encoder *encoder, struct intel_crtc_state *pipe_config) { + struct intel_display *display = to_intel_display(encoder); struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_dp *intel_dp = enc_to_intel_dp(encoder); u32 tmp, flags = 0; @@ -344,12 +352,12 @@ static void intel_dp_get_config(struct intel_encoder *encoder, else pipe_config->output_types |= BIT(INTEL_OUTPUT_DP); - tmp = intel_de_read(dev_priv, intel_dp->output_reg); + tmp = intel_de_read(display, intel_dp->output_reg); pipe_config->has_audio = tmp & DP_AUDIO_OUTPUT_ENABLE && port != PORT_A; if (HAS_PCH_CPT(dev_priv) && port != PORT_A) { - u32 trans_dp = intel_de_read(dev_priv, + u32 trans_dp = intel_de_read(display, TRANS_DP_CTL(crtc->pipe)); if (trans_dp & TRANS_DP_ENH_FRAMING) @@ -390,7 +398,7 @@ static void intel_dp_get_config(struct intel_encoder *encoder, g4x_dp_get_m_n(pipe_config); if (port == PORT_A) { - if ((intel_de_read(dev_priv, DP_A) & DP_PLL_FREQ_MASK) == DP_PLL_FREQ_162MHZ) + if ((intel_de_read(display, DP_A) & DP_PLL_FREQ_MASK) == DP_PLL_FREQ_162MHZ) pipe_config->port_clock = 162000; else pipe_config->port_clock = 270000; @@ -410,17 +418,18 @@ static void intel_dp_link_down(struct intel_encoder *encoder, const struct intel_crtc_state *old_crtc_state) { + struct intel_display *display = to_intel_display(encoder); struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_dp *intel_dp = enc_to_intel_dp(encoder); struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc); enum port port = encoder->port; - if (drm_WARN_ON(&dev_priv->drm, - (intel_de_read(dev_priv, intel_dp->output_reg) & + if (drm_WARN_ON(display->drm, + (intel_de_read(display, intel_dp->output_reg) & DP_PORT_EN) == 0)) return; - drm_dbg_kms(&dev_priv->drm, "\n"); + drm_dbg_kms(display->drm, "\n"); if ((IS_IVYBRIDGE(dev_priv) && port == PORT_A) || (HAS_PCH_CPT(dev_priv) && port != PORT_A)) { @@ -430,12 +439,12 @@ intel_dp_link_down(struct intel_encoder *encoder, intel_dp->DP &= ~DP_LINK_TRAIN_MASK; intel_dp->DP |= DP_LINK_TRAIN_PAT_IDLE; } - intel_de_write(dev_priv, intel_dp->output_reg, intel_dp->DP); - intel_de_posting_read(dev_priv, intel_dp->output_reg); + intel_de_write(display, intel_dp->output_reg, intel_dp->DP); + intel_de_posting_read(display, intel_dp->output_reg); intel_dp->DP &= ~DP_PORT_EN; - intel_de_write(dev_priv, intel_dp->output_reg, intel_dp->DP); - intel_de_posting_read(dev_priv, intel_dp->output_reg); + intel_de_write(display, intel_dp->output_reg, intel_dp->DP); + intel_de_posting_read(display, intel_dp->output_reg); /* * HW workaround for IBX, we need to move the port @@ -454,12 +463,12 @@ intel_dp_link_down(struct intel_encoder *encoder, intel_dp->DP &= ~(DP_PIPE_SEL_MASK | DP_LINK_TRAIN_MASK); intel_dp->DP |= DP_PORT_EN | DP_PIPE_SEL(PIPE_A) | DP_LINK_TRAIN_PAT_1; - intel_de_write(dev_priv, intel_dp->output_reg, intel_dp->DP); - intel_de_posting_read(dev_priv, intel_dp->output_reg); + intel_de_write(display, intel_dp->output_reg, intel_dp->DP); + intel_de_posting_read(display, intel_dp->output_reg); intel_dp->DP &= ~DP_PORT_EN; - intel_de_write(dev_priv, intel_dp->output_reg, intel_dp->DP); - intel_de_posting_read(dev_priv, intel_dp->output_reg); + intel_de_write(display, intel_dp->output_reg, intel_dp->DP); + intel_de_posting_read(display, intel_dp->output_reg); intel_wait_for_vblank_if_active(dev_priv, PIPE_A); intel_set_cpu_fifo_underrun_reporting(dev_priv, PIPE_A, true); @@ -480,7 +489,7 @@ static void g4x_dp_audio_enable(struct intel_encoder *encoder, const struct intel_crtc_state *crtc_state, const struct drm_connector_state *conn_state) { - struct drm_i915_private *i915 = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); struct intel_dp *intel_dp = enc_to_intel_dp(encoder); if (!crtc_state->has_audio) @@ -488,7 +497,7 @@ static void g4x_dp_audio_enable(struct intel_encoder *encoder, /* Enable audio presence detect */ intel_dp->DP |= DP_AUDIO_OUTPUT_ENABLE; - intel_de_write(i915, intel_dp->output_reg, intel_dp->DP); + intel_de_write(display, intel_dp->output_reg, intel_dp->DP); intel_audio_codec_enable(encoder, crtc_state, conn_state); } @@ -497,7 +506,7 @@ static void g4x_dp_audio_disable(struct intel_encoder *encoder, const struct intel_crtc_state *old_crtc_state, const struct drm_connector_state *old_conn_state) { - struct drm_i915_private *i915 = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); struct intel_dp *intel_dp = enc_to_intel_dp(encoder); if (!old_crtc_state->has_audio) @@ -507,7 +516,7 @@ static void g4x_dp_audio_disable(struct intel_encoder *encoder, /* Disable audio presence detect */ intel_dp->DP &= ~DP_AUDIO_OUTPUT_ENABLE; - intel_de_write(i915, intel_dp->output_reg, intel_dp->DP); + intel_de_write(display, intel_dp->output_reg, intel_dp->DP); } static void intel_disable_dp(struct intel_atomic_state *state, @@ -596,7 +605,7 @@ cpt_set_link_train(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state, u8 dp_train_pat) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); intel_dp->DP &= ~DP_LINK_TRAIN_MASK_CPT; @@ -615,8 +624,8 @@ cpt_set_link_train(struct intel_dp *intel_dp, return; } - intel_de_write(dev_priv, intel_dp->output_reg, intel_dp->DP); - intel_de_posting_read(dev_priv, intel_dp->output_reg); + intel_de_write(display, intel_dp->output_reg, intel_dp->DP); + intel_de_posting_read(display, intel_dp->output_reg); } static void @@ -624,7 +633,7 @@ g4x_set_link_train(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state, u8 dp_train_pat) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); intel_dp->DP &= ~DP_LINK_TRAIN_MASK; @@ -643,14 +652,14 @@ g4x_set_link_train(struct intel_dp *intel_dp, return; } - intel_de_write(dev_priv, intel_dp->output_reg, intel_dp->DP); - intel_de_posting_read(dev_priv, intel_dp->output_reg); + intel_de_write(display, intel_dp->output_reg, intel_dp->DP); + intel_de_posting_read(display, intel_dp->output_reg); } static void intel_dp_enable_port(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); /* enable with pattern 1 (as per spec) */ @@ -665,8 +674,8 @@ static void intel_dp_enable_port(struct intel_dp *intel_dp, */ intel_dp->DP |= DP_PORT_EN; - intel_de_write(dev_priv, intel_dp->output_reg, intel_dp->DP); - intel_de_posting_read(dev_priv, intel_dp->output_reg); + intel_de_write(display, intel_dp->output_reg, intel_dp->DP); + intel_de_posting_read(display, intel_dp->output_reg); } static void intel_enable_dp(struct intel_atomic_state *state, @@ -674,12 +683,13 @@ static void intel_enable_dp(struct intel_atomic_state *state, const struct intel_crtc_state *pipe_config, const struct drm_connector_state *conn_state) { + struct intel_display *display = to_intel_display(state); struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_dp *intel_dp = enc_to_intel_dp(encoder); - u32 dp_reg = intel_de_read(dev_priv, intel_dp->output_reg); + u32 dp_reg = intel_de_read(display, intel_dp->output_reg); intel_wakeref_t wakeref; - if (drm_WARN_ON(&dev_priv->drm, dp_reg & DP_PORT_EN)) + if (drm_WARN_ON(display->drm, dp_reg & DP_PORT_EN)) return; with_intel_pps_lock(intel_dp, wakeref) { @@ -1026,21 +1036,21 @@ static void g4x_set_signal_levels(struct intel_encoder *encoder, const struct intel_crtc_state *crtc_state) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); struct intel_dp *intel_dp = enc_to_intel_dp(encoder); u8 train_set = intel_dp->train_set[0]; u32 signal_levels; signal_levels = g4x_signal_levels(train_set); - drm_dbg_kms(&dev_priv->drm, "Using signal levels %08x\n", + drm_dbg_kms(display->drm, "Using signal levels %08x\n", signal_levels); intel_dp->DP &= ~(DP_VOLTAGE_MASK | DP_PRE_EMPHASIS_MASK); intel_dp->DP |= signal_levels; - intel_de_write(dev_priv, intel_dp->output_reg, intel_dp->DP); - intel_de_posting_read(dev_priv, intel_dp->output_reg); + intel_de_write(display, intel_dp->output_reg, intel_dp->DP); + intel_de_posting_read(display, intel_dp->output_reg); } /* SNB CPU eDP voltage swing and pre-emphasis control */ @@ -1074,21 +1084,21 @@ static void snb_cpu_edp_set_signal_levels(struct intel_encoder *encoder, const struct intel_crtc_state *crtc_state) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); struct intel_dp *intel_dp = enc_to_intel_dp(encoder); u8 train_set = intel_dp->train_set[0]; u32 signal_levels; signal_levels = snb_cpu_edp_signal_levels(train_set); - drm_dbg_kms(&dev_priv->drm, "Using signal levels %08x\n", + drm_dbg_kms(display->drm, "Using signal levels %08x\n", signal_levels); intel_dp->DP &= ~EDP_LINK_TRAIN_VOL_EMP_MASK_SNB; intel_dp->DP |= signal_levels; - intel_de_write(dev_priv, intel_dp->output_reg, intel_dp->DP); - intel_de_posting_read(dev_priv, intel_dp->output_reg); + intel_de_write(display, intel_dp->output_reg, intel_dp->DP); + intel_de_posting_read(display, intel_dp->output_reg); } /* IVB CPU eDP voltage swing and pre-emphasis control */ @@ -1126,21 +1136,21 @@ static void ivb_cpu_edp_set_signal_levels(struct intel_encoder *encoder, const struct intel_crtc_state *crtc_state) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); struct intel_dp *intel_dp = enc_to_intel_dp(encoder); u8 train_set = intel_dp->train_set[0]; u32 signal_levels; signal_levels = ivb_cpu_edp_signal_levels(train_set); - drm_dbg_kms(&dev_priv->drm, "Using signal levels %08x\n", + drm_dbg_kms(display->drm, "Using signal levels %08x\n", signal_levels); intel_dp->DP &= ~EDP_LINK_TRAIN_VOL_EMP_MASK_IVB; intel_dp->DP |= signal_levels; - intel_de_write(dev_priv, intel_dp->output_reg, intel_dp->DP); - intel_de_posting_read(dev_priv, intel_dp->output_reg); + intel_de_write(display, intel_dp->output_reg, intel_dp->DP); + intel_de_posting_read(display, intel_dp->output_reg); } /* @@ -1185,15 +1195,15 @@ intel_dp_hotplug(struct intel_encoder *encoder, static bool ibx_digital_port_connected(struct intel_encoder *encoder) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); - u32 bit = dev_priv->display.hotplug.pch_hpd[encoder->hpd_pin]; + struct intel_display *display = to_intel_display(encoder); + u32 bit = display->hotplug.pch_hpd[encoder->hpd_pin]; - return intel_de_read(dev_priv, SDEISR) & bit; + return intel_de_read(display, SDEISR) & bit; } static bool g4x_digital_port_connected(struct intel_encoder *encoder) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); u32 bit; switch (encoder->hpd_pin) { @@ -1211,15 +1221,15 @@ static bool g4x_digital_port_connected(struct intel_encoder *encoder) return false; } - return intel_de_read(dev_priv, PORT_HOTPLUG_STAT(dev_priv)) & bit; + return intel_de_read(display, PORT_HOTPLUG_STAT(display)) & bit; } static bool ilk_digital_port_connected(struct intel_encoder *encoder) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); - u32 bit = dev_priv->display.hotplug.hpd[encoder->hpd_pin]; + struct intel_display *display = to_intel_display(encoder); + u32 bit = display->hotplug.hpd[encoder->hpd_pin]; - return intel_de_read(dev_priv, DEISR) & bit; + return intel_de_read(display, DEISR) & bit; } static void g4x_dp_suspend_complete(struct intel_encoder *encoder) @@ -1241,7 +1251,8 @@ static void intel_dp_encoder_destroy(struct drm_encoder *encoder) enum pipe vlv_active_pipe(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base; enum pipe pipe; @@ -1254,10 +1265,11 @@ enum pipe vlv_active_pipe(struct intel_dp *intel_dp) static void intel_dp_encoder_reset(struct drm_encoder *encoder) { + struct intel_display *display = to_intel_display(encoder->dev); struct drm_i915_private *dev_priv = to_i915(encoder->dev); struct intel_dp *intel_dp = enc_to_intel_dp(to_intel_encoder(encoder)); - intel_dp->DP = intel_de_read(dev_priv, intel_dp->output_reg); + intel_dp->DP = intel_de_read(display, intel_dp->output_reg); intel_dp->reset_link_params = true; @@ -1293,7 +1305,7 @@ bool g4x_dp_init(struct drm_i915_private *dev_priv, /* FIXME bail? */ if (!devdata) - drm_dbg_kms(&dev_priv->drm, "No VBT child device for DP-%c\n", + drm_dbg_kms(display->drm, "No VBT child device for DP-%c\n", port_name(port)); dig_port = kzalloc(sizeof(*dig_port), GFP_KERNEL); @@ -1313,7 +1325,7 @@ bool g4x_dp_init(struct drm_i915_private *dev_priv, mutex_init(&dig_port->hdcp_mutex); - if (drm_encoder_init(&dev_priv->drm, &intel_encoder->base, + if (drm_encoder_init(display->drm, &intel_encoder->base, &intel_dp_enc_funcs, DRM_MODE_ENCODER_TMDS, "DP %c", port_name(port))) goto err_encoder_init; @@ -1397,7 +1409,7 @@ bool g4x_dp_init(struct drm_i915_private *dev_priv, dig_port->hpd_pulse = intel_dp_hpd_pulse; - if (HAS_GMCH(dev_priv)) { + if (HAS_GMCH(display)) { dig_port->connected = g4x_digital_port_connected; } else { if (port == PORT_A) diff --git a/drivers/gpu/drm/i915/display/intel_alpm.c b/drivers/gpu/drm/i915/display/intel_alpm.c index 82ee778b2efe..186cf4833f71 100644 --- a/drivers/gpu/drm/i915/display/intel_alpm.c +++ b/drivers/gpu/drm/i915/display/intel_alpm.c @@ -228,7 +228,7 @@ bool intel_alpm_compute_params(struct intel_dp *intel_dp, int tfw_exit_latency = 20; /* eDP spec */ int phy_wake = 4; /* eDP spec */ int preamble = 8; /* eDP spec */ - int precharge = intel_dp_aux_fw_sync_len() - preamble; + int precharge = intel_dp_aux_fw_sync_len(intel_dp) - preamble; u8 max_wake_lines; io_wake_time = max(precharge, io_buffer_wake_time(crtc_state)) + diff --git a/drivers/gpu/drm/i915/display/intel_audio.c b/drivers/gpu/drm/i915/display/intel_audio.c index 9b8508a503f7..f5e7eefab2f1 100644 --- a/drivers/gpu/drm/i915/display/intel_audio.c +++ b/drivers/gpu/drm/i915/display/intel_audio.c @@ -980,7 +980,8 @@ retry: static unsigned long i915_audio_component_get_power(struct device *kdev) { - struct drm_i915_private *i915 = kdev_to_i915(kdev); + struct intel_display *display = to_intel_display(kdev); + struct drm_i915_private *i915 = to_i915(display->drm); intel_wakeref_t ret; /* Catch potential impedance mismatches before they occur! */ @@ -1012,7 +1013,8 @@ static unsigned long i915_audio_component_get_power(struct device *kdev) static void i915_audio_component_put_power(struct device *kdev, unsigned long cookie) { - struct drm_i915_private *i915 = kdev_to_i915(kdev); + struct intel_display *display = to_intel_display(kdev); + struct drm_i915_private *i915 = to_i915(display->drm); /* Stop forcing CDCLK to 2*BCLK if no need for audio to be powered. */ if (--i915->display.audio.power_refcount == 0) @@ -1025,7 +1027,8 @@ static void i915_audio_component_put_power(struct device *kdev, static void i915_audio_component_codec_wake_override(struct device *kdev, bool enable) { - struct drm_i915_private *i915 = kdev_to_i915(kdev); + struct intel_display *display = to_intel_display(kdev); + struct drm_i915_private *i915 = to_i915(display->drm); unsigned long cookie; if (DISPLAY_VER(i915) < 9) @@ -1053,7 +1056,8 @@ static void i915_audio_component_codec_wake_override(struct device *kdev, /* Get CDCLK in kHz */ static int i915_audio_component_get_cdclk_freq(struct device *kdev) { - struct drm_i915_private *i915 = kdev_to_i915(kdev); + struct intel_display *display = to_intel_display(kdev); + struct drm_i915_private *i915 = to_i915(display->drm); if (drm_WARN_ON_ONCE(&i915->drm, !HAS_DDI(i915))) return -ENODEV; @@ -1112,7 +1116,8 @@ static struct intel_audio_state *find_audio_state(struct drm_i915_private *i915, static int i915_audio_component_sync_audio_rate(struct device *kdev, int port, int cpu_transcoder, int rate) { - struct drm_i915_private *i915 = kdev_to_i915(kdev); + struct intel_display *display = to_intel_display(kdev); + struct drm_i915_private *i915 = to_i915(display->drm); struct i915_audio_component *acomp = i915->display.audio.component; const struct intel_audio_state *audio_state; struct intel_encoder *encoder; @@ -1154,7 +1159,8 @@ static int i915_audio_component_get_eld(struct device *kdev, int port, int cpu_transcoder, bool *enabled, unsigned char *buf, int max_bytes) { - struct drm_i915_private *i915 = kdev_to_i915(kdev); + struct intel_display *display = to_intel_display(kdev); + struct drm_i915_private *i915 = to_i915(display->drm); const struct intel_audio_state *audio_state; int ret = 0; @@ -1189,24 +1195,25 @@ static const struct drm_audio_component_ops i915_audio_component_ops = { .get_eld = i915_audio_component_get_eld, }; -static int i915_audio_component_bind(struct device *i915_kdev, +static int i915_audio_component_bind(struct device *drv_kdev, struct device *hda_kdev, void *data) { + struct intel_display *display = to_intel_display(drv_kdev); + struct drm_i915_private *i915 = to_i915(display->drm); struct i915_audio_component *acomp = data; - struct drm_i915_private *i915 = kdev_to_i915(i915_kdev); int i; if (drm_WARN_ON(&i915->drm, acomp->base.ops || acomp->base.dev)) return -EEXIST; if (drm_WARN_ON(&i915->drm, - !device_link_add(hda_kdev, i915_kdev, + !device_link_add(hda_kdev, drv_kdev, DL_FLAG_STATELESS))) return -ENOMEM; drm_modeset_lock_all(&i915->drm); acomp->base.ops = &i915_audio_component_ops; - acomp->base.dev = i915_kdev; + acomp->base.dev = drv_kdev; BUILD_BUG_ON(MAX_PORTS != I915_MAX_PORTS); for (i = 0; i < ARRAY_SIZE(acomp->aud_sample_rate); i++) acomp->aud_sample_rate[i] = 0; @@ -1216,11 +1223,12 @@ static int i915_audio_component_bind(struct device *i915_kdev, return 0; } -static void i915_audio_component_unbind(struct device *i915_kdev, +static void i915_audio_component_unbind(struct device *drv_kdev, struct device *hda_kdev, void *data) { + struct intel_display *display = to_intel_display(drv_kdev); + struct drm_i915_private *i915 = to_i915(display->drm); struct i915_audio_component *acomp = data; - struct drm_i915_private *i915 = kdev_to_i915(i915_kdev); drm_modeset_lock_all(&i915->drm); acomp->base.ops = NULL; @@ -1228,7 +1236,7 @@ static void i915_audio_component_unbind(struct device *i915_kdev, i915->display.audio.component = NULL; drm_modeset_unlock_all(&i915->drm); - device_link_remove(hda_kdev, i915_kdev); + device_link_remove(hda_kdev, drv_kdev); if (i915->display.audio.power_refcount) drm_err(&i915->drm, "audio power refcount %d after unbind\n", diff --git a/drivers/gpu/drm/i915/display/intel_ddi.c b/drivers/gpu/drm/i915/display/intel_ddi.c index 25ff3ff0ab95..00fbe9f8c03a 100644 --- a/drivers/gpu/drm/i915/display/intel_ddi.c +++ b/drivers/gpu/drm/i915/display/intel_ddi.c @@ -1400,7 +1400,7 @@ static void tgl_dkl_phy_set_signal_levels(struct intel_encoder *encoder, static int translate_signal_level(struct intel_dp *intel_dp, u8 signal_levels) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); int i; for (i = 0; i < ARRAY_SIZE(index_to_dp_signal_levels); i++) { @@ -1408,7 +1408,7 @@ static int translate_signal_level(struct intel_dp *intel_dp, return i; } - drm_WARN(&i915->drm, 1, + drm_WARN(display->drm, 1, "Unsupported voltage swing/pre-emphasis level: 0x%x\n", signal_levels); @@ -2211,14 +2211,14 @@ static void intel_dp_sink_set_msa_timing_par_ignore_state(struct intel_dp *intel const struct intel_crtc_state *crtc_state, bool enable) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); if (!crtc_state->vrr.enable) return; if (drm_dp_dpcd_writeb(&intel_dp->aux, DP_DOWNSPREAD_CTRL, enable ? DP_MSA_TIMING_PAR_IGNORE_EN : 0) <= 0) - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "Failed to %s MSA_TIMING_PAR_IGNORE in the sink\n", str_enable_disable(enable)); } @@ -2227,20 +2227,20 @@ static void intel_dp_sink_set_fec_ready(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state, bool enable) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); if (!crtc_state->fec_enable) return; if (drm_dp_dpcd_writeb(&intel_dp->aux, DP_FEC_CONFIGURATION, enable ? DP_FEC_READY : 0) <= 0) - drm_dbg_kms(&i915->drm, "Failed to set FEC_READY to %s in the sink\n", + drm_dbg_kms(display->drm, "Failed to set FEC_READY to %s in the sink\n", enable ? "enabled" : "disabled"); if (enable && drm_dp_dpcd_writeb(&intel_dp->aux, DP_FEC_STATUS, DP_FEC_DECODE_EN_DETECTED | DP_FEC_DECODE_DIS_DETECTED) <= 0) - drm_dbg_kms(&i915->drm, "Failed to clear FEC detected flags\n"); + drm_dbg_kms(display->drm, "Failed to clear FEC detected flags\n"); } static int read_fec_detected_status(struct drm_dp_aux *aux) diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c index 78ce402a5cd0..b4ef4d59da1a 100644 --- a/drivers/gpu/drm/i915/display/intel_display.c +++ b/drivers/gpu/drm/i915/display/intel_display.c @@ -7793,7 +7793,7 @@ void intel_setup_outputs(struct drm_i915_private *dev_priv) struct intel_encoder *encoder; bool dpd_is_edp = false; - intel_pps_unlock_regs_wa(dev_priv); + intel_pps_unlock_regs_wa(display); if (!HAS_DISPLAY(dev_priv)) return; diff --git a/drivers/gpu/drm/i915/display/intel_display_debugfs.c b/drivers/gpu/drm/i915/display/intel_display_debugfs.c index 74f527647aa9..f5f618199d39 100644 --- a/drivers/gpu/drm/i915/display/intel_display_debugfs.c +++ b/drivers/gpu/drm/i915/display/intel_display_debugfs.c @@ -1071,7 +1071,7 @@ void intel_display_debugfs_register(struct drm_i915_private *i915) intel_fbc_debugfs_register(display); intel_hpd_debugfs_register(i915); intel_opregion_debugfs_register(display); - intel_psr_debugfs_register(i915); + intel_psr_debugfs_register(display); intel_wm_debugfs_register(i915); intel_display_debugfs_params(display); } diff --git a/drivers/gpu/drm/i915/display/intel_display_driver.c b/drivers/gpu/drm/i915/display/intel_display_driver.c index eced20d2ce6e..069426d9260b 100644 --- a/drivers/gpu/drm/i915/display/intel_display_driver.c +++ b/drivers/gpu/drm/i915/display/intel_display_driver.c @@ -428,7 +428,7 @@ int intel_display_driver_probe_nogem(struct drm_i915_private *i915) intel_panel_sanitize_ssc(i915); - intel_pps_setup(i915); + intel_pps_setup(display); intel_gmbus_setup(i915); @@ -459,7 +459,7 @@ int intel_display_driver_probe_nogem(struct drm_i915_private *i915) intel_vga_disable(i915); intel_setup_outputs(i915); - ret = intel_dp_tunnel_mgr_init(i915); + ret = intel_dp_tunnel_mgr_init(display); if (ret) goto err_hdcp; @@ -580,6 +580,8 @@ void intel_display_driver_remove(struct drm_i915_private *i915) /* part #2: call after irq uninstall */ void intel_display_driver_remove_noirq(struct drm_i915_private *i915) { + struct intel_display *display = &i915->display; + if (!HAS_DISPLAY(i915)) return; @@ -600,7 +602,7 @@ void intel_display_driver_remove_noirq(struct drm_i915_private *i915) intel_mode_config_cleanup(i915); - intel_dp_tunnel_mgr_cleanup(i915); + intel_dp_tunnel_mgr_cleanup(display); intel_overlay_cleanup(i915); diff --git a/drivers/gpu/drm/i915/display/intel_display_irq.c b/drivers/gpu/drm/i915/display/intel_display_irq.c index d85c33eabc47..73369847ed66 100644 --- a/drivers/gpu/drm/i915/display/intel_display_irq.c +++ b/drivers/gpu/drm/i915/display/intel_display_irq.c @@ -576,6 +576,7 @@ void valleyview_pipestat_irq_handler(struct drm_i915_private *dev_priv, static void ibx_irq_handler(struct drm_i915_private *dev_priv, u32 pch_iir) { + struct intel_display *display = &dev_priv->display; enum pipe pipe; u32 hotplug_trigger = pch_iir & SDE_HOTPLUG_MASK; @@ -589,7 +590,7 @@ static void ibx_irq_handler(struct drm_i915_private *dev_priv, u32 pch_iir) } if (pch_iir & SDE_AUX_MASK) - intel_dp_aux_irq_handler(dev_priv); + intel_dp_aux_irq_handler(display); if (pch_iir & SDE_GMBUS) intel_gmbus_irq_handler(dev_priv); @@ -664,6 +665,7 @@ static void cpt_serr_int_handler(struct drm_i915_private *dev_priv) static void cpt_irq_handler(struct drm_i915_private *dev_priv, u32 pch_iir) { + struct intel_display *display = &dev_priv->display; enum pipe pipe; u32 hotplug_trigger = pch_iir & SDE_HOTPLUG_MASK_CPT; @@ -677,7 +679,7 @@ static void cpt_irq_handler(struct drm_i915_private *dev_priv, u32 pch_iir) } if (pch_iir & SDE_AUX_MASK_CPT) - intel_dp_aux_irq_handler(dev_priv); + intel_dp_aux_irq_handler(display); if (pch_iir & SDE_GMBUS_CPT) intel_gmbus_irq_handler(dev_priv); @@ -709,7 +711,7 @@ void ilk_display_irq_handler(struct drm_i915_private *dev_priv, u32 de_iir) ilk_hpd_irq_handler(dev_priv, hotplug_trigger); if (de_iir & DE_AUX_CHANNEL_A) - intel_dp_aux_irq_handler(dev_priv); + intel_dp_aux_irq_handler(display); if (de_iir & DE_GSE) intel_opregion_asle_intr(display); @@ -775,7 +777,7 @@ void ivb_display_irq_handler(struct drm_i915_private *dev_priv, u32 de_iir) } if (de_iir & DE_AUX_CHANNEL_A_IVB) - intel_dp_aux_irq_handler(dev_priv); + intel_dp_aux_irq_handler(display); if (de_iir & DE_GSE_IVB) intel_opregion_asle_intr(display); @@ -1065,6 +1067,7 @@ static void gen8_read_and_ack_pch_irqs(struct drm_i915_private *i915, u32 *pch_i void gen8_de_irq_handler(struct drm_i915_private *dev_priv, u32 master_ctl) { + struct intel_display *display = &dev_priv->display; u32 iir; enum pipe pipe; @@ -1100,7 +1103,7 @@ void gen8_de_irq_handler(struct drm_i915_private *dev_priv, u32 master_ctl) intel_uncore_write(&dev_priv->uncore, GEN8_DE_PORT_IIR, iir); if (iir & gen8_de_port_aux_mask(dev_priv)) { - intel_dp_aux_irq_handler(dev_priv); + intel_dp_aux_irq_handler(display); found = true; } diff --git a/drivers/gpu/drm/i915/display/intel_display_power_well.c b/drivers/gpu/drm/i915/display/intel_display_power_well.c index adf5d1fbccb5..46e9eff12c23 100644 --- a/drivers/gpu/drm/i915/display/intel_display_power_well.c +++ b/drivers/gpu/drm/i915/display/intel_display_power_well.c @@ -861,6 +861,8 @@ void skl_enable_dc6(struct drm_i915_private *dev_priv) void bxt_enable_dc9(struct drm_i915_private *dev_priv) { + struct intel_display *display = &dev_priv->display; + assert_can_enable_dc9(dev_priv); drm_dbg_kms(&dev_priv->drm, "Enabling DC9\n"); @@ -870,19 +872,21 @@ void bxt_enable_dc9(struct drm_i915_private *dev_priv) * because PPS registers are always on. */ if (!HAS_PCH_SPLIT(dev_priv)) - intel_pps_reset_all(dev_priv); + intel_pps_reset_all(display); gen9_set_dc_state(dev_priv, DC_STATE_EN_DC9); } void bxt_disable_dc9(struct drm_i915_private *dev_priv) { + struct intel_display *display = &dev_priv->display; + assert_can_disable_dc9(dev_priv); drm_dbg_kms(&dev_priv->drm, "Disabling DC9\n"); gen9_set_dc_state(dev_priv, DC_STATE_DISABLE); - intel_pps_unlock_regs_wa(dev_priv); + intel_pps_unlock_regs_wa(display); } static void hsw_power_well_sync_hw(struct drm_i915_private *dev_priv, @@ -1184,6 +1188,7 @@ static void vlv_init_display_clock_gating(struct drm_i915_private *dev_priv) static void vlv_display_power_well_init(struct drm_i915_private *dev_priv) { + struct intel_display *display = &dev_priv->display; struct intel_encoder *encoder; enum pipe pipe; @@ -1229,11 +1234,13 @@ static void vlv_display_power_well_init(struct drm_i915_private *dev_priv) intel_vga_redisable_power_on(dev_priv); - intel_pps_unlock_regs_wa(dev_priv); + intel_pps_unlock_regs_wa(display); } static void vlv_display_power_well_deinit(struct drm_i915_private *dev_priv) { + struct intel_display *display = &dev_priv->display; + spin_lock_irq(&dev_priv->irq_lock); valleyview_disable_display_irqs(dev_priv); spin_unlock_irq(&dev_priv->irq_lock); @@ -1241,7 +1248,7 @@ static void vlv_display_power_well_deinit(struct drm_i915_private *dev_priv) /* make sure we're done processing display irqs */ intel_synchronize_irq(dev_priv); - intel_pps_reset_all(dev_priv); + intel_pps_reset_all(display); /* Prevent us from re-enabling polling on accident in late suspend */ if (!dev_priv->drm.dev->power.is_suspended) diff --git a/drivers/gpu/drm/i915/display/intel_display_reset.c b/drivers/gpu/drm/i915/display/intel_display_reset.c index c2c347b22448..49e2e650ebcd 100644 --- a/drivers/gpu/drm/i915/display/intel_display_reset.c +++ b/drivers/gpu/drm/i915/display/intel_display_reset.c @@ -83,7 +83,8 @@ void intel_display_reset_prepare(struct drm_i915_private *dev_priv) void intel_display_reset_finish(struct drm_i915_private *i915) { - struct drm_modeset_acquire_ctx *ctx = &i915->display.restore.reset_ctx; + struct intel_display *display = &i915->display; + struct drm_modeset_acquire_ctx *ctx = &display->restore.reset_ctx; struct drm_atomic_state *state; int ret; @@ -94,7 +95,7 @@ void intel_display_reset_finish(struct drm_i915_private *i915) if (!test_bit(I915_RESET_MODESET, &to_gt(i915)->reset.flags)) return; - state = fetch_and_zero(&i915->display.restore.modeset_state); + state = fetch_and_zero(&display->restore.modeset_state); if (!state) goto unlock; @@ -112,7 +113,7 @@ void intel_display_reset_finish(struct drm_i915_private *i915) * The display has been reset as well, * so need a full re-initialization. */ - intel_pps_unlock_regs_wa(i915); + intel_pps_unlock_regs_wa(display); intel_display_driver_init_hw(i915); intel_clock_gating_init(i915); intel_hpd_init(i915); diff --git a/drivers/gpu/drm/i915/display/intel_display_types.h b/drivers/gpu/drm/i915/display/intel_display_types.h index 868ff8976ed9..f29e5dc3db91 100644 --- a/drivers/gpu/drm/i915/display/intel_display_types.h +++ b/drivers/gpu/drm/i915/display/intel_display_types.h @@ -1907,6 +1907,10 @@ struct intel_dp { } alpm_parameters; u8 alpm_dpcd; + + struct { + unsigned long mask; + } quirks; }; enum lspcon_vendor { @@ -2081,8 +2085,6 @@ dp_to_lspcon(struct intel_dp *intel_dp) return &dp_to_dig_port(intel_dp)->lspcon; } -#define dp_to_i915(__intel_dp) to_i915(dp_to_dig_port(__intel_dp)->base.base.dev) - static inline struct intel_digital_port * hdmi_to_dig_port(struct intel_hdmi *intel_hdmi) { @@ -2205,7 +2207,11 @@ to_intel_frontbuffer(struct drm_framebuffer *fb) * intel_display pointer. */ #define __drm_device_to_intel_display(p) \ - (&to_i915(p)->display) + ((p) ? &to_i915(p)->display : NULL) +#define __device_to_intel_display(p) \ + __drm_device_to_intel_display(dev_get_drvdata(p)) +#define __pci_dev_to_intel_display(p) \ + __drm_device_to_intel_display(pci_get_drvdata(p)) #define __intel_atomic_state_to_intel_display(p) \ __drm_device_to_intel_display((p)->base.dev) #define __intel_connector_to_intel_display(p) \ @@ -2231,6 +2237,8 @@ to_intel_frontbuffer(struct drm_framebuffer *fb) #define to_intel_display(p) \ _Generic(*p, \ __assoc(drm_device, p), \ + __assoc(device, p), \ + __assoc(pci_dev, p), \ __assoc(intel_atomic_state, p), \ __assoc(intel_connector, p), \ __assoc(intel_crtc, p), \ diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c index 789c2f78826d..a1fcedfd404b 100644 --- a/drivers/gpu/drm/i915/display/intel_dp.c +++ b/drivers/gpu/drm/i915/display/intel_dp.c @@ -84,11 +84,14 @@ #include "intel_pch_display.h" #include "intel_pps.h" #include "intel_psr.h" +#include "intel_quirks.h" #include "intel_tc.h" #include "intel_vdsc.h" #include "intel_vrr.h" #include "intel_crtc_state_dump.h" +#define dp_to_i915(__intel_dp) to_i915(dp_to_dig_port(__intel_dp)->base.base.dev) + /* DP DSC throughput values used for slice count calculations KPixels/s */ #define DP_DSC_PEAK_PIXEL_RATE 2720000 #define DP_DSC_MAX_ENC_THROUGHPUT_0 340000 @@ -4053,6 +4056,7 @@ intel_edp_init_dpcd(struct intel_dp *intel_dp, struct intel_connector *connector drm_dp_read_desc(&intel_dp->aux, &intel_dp->desc, drm_dp_is_branch(intel_dp->dpcd)); + intel_init_dpcd_quirks(intel_dp, &intel_dp->desc.ident); /* * Read the eDP display control registers. @@ -4165,6 +4169,8 @@ intel_dp_get_dpcd(struct intel_dp *intel_dp) drm_dp_read_desc(&intel_dp->aux, &intel_dp->desc, drm_dp_is_branch(intel_dp->dpcd)); + intel_init_dpcd_quirks(intel_dp, &intel_dp->desc.ident); + intel_dp_update_sink_caps(intel_dp); } diff --git a/drivers/gpu/drm/i915/display/intel_dp_aux.c b/drivers/gpu/drm/i915/display/intel_dp_aux.c index cbc817bb0cc3..04a7acd7f73c 100644 --- a/drivers/gpu/drm/i915/display/intel_dp_aux.c +++ b/drivers/gpu/drm/i915/display/intel_dp_aux.c @@ -13,16 +13,17 @@ #include "intel_dp_aux.h" #include "intel_dp_aux_regs.h" #include "intel_pps.h" +#include "intel_quirks.h" #include "intel_tc.h" #define AUX_CH_NAME_BUFSIZE 6 -static const char *aux_ch_name(struct drm_i915_private *i915, +static const char *aux_ch_name(struct intel_display *display, char *buf, int size, enum aux_ch aux_ch) { - if (DISPLAY_VER(i915) >= 13 && aux_ch >= AUX_CH_D_XELPD) + if (DISPLAY_VER(display) >= 13 && aux_ch >= AUX_CH_D_XELPD) snprintf(buf, size, "%c", 'A' + aux_ch - AUX_CH_D_XELPD + AUX_CH_D); - else if (DISPLAY_VER(i915) >= 12 && aux_ch >= AUX_CH_USBC1) + else if (DISPLAY_VER(display) >= 12 && aux_ch >= AUX_CH_USBC1) snprintf(buf, size, "USBC%c", '1' + aux_ch - AUX_CH_USBC1); else snprintf(buf, size, "%c", 'A' + aux_ch); @@ -55,17 +56,18 @@ static void intel_dp_aux_unpack(u32 src, u8 *dst, int dst_bytes) static u32 intel_dp_aux_wait_done(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); i915_reg_t ch_ctl = intel_dp->aux_ch_ctl_reg(intel_dp); const unsigned int timeout_ms = 10; u32 status; int ret; - ret = intel_de_wait_custom(i915, ch_ctl, DP_AUX_CH_CTL_SEND_BUSY, 0, + ret = intel_de_wait_custom(display, ch_ctl, DP_AUX_CH_CTL_SEND_BUSY, + 0, 2, timeout_ms, &status); if (ret == -ETIMEDOUT) - drm_err(&i915->drm, + drm_err(display->drm, "%s: did not complete or timeout within %ums (status 0x%08x)\n", intel_dp->aux.name, timeout_ms, status); @@ -74,7 +76,7 @@ intel_dp_aux_wait_done(struct intel_dp *intel_dp) static u32 g4x_get_aux_clock_divider(struct intel_dp *intel_dp, int index) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); if (index) return 0; @@ -83,12 +85,12 @@ static u32 g4x_get_aux_clock_divider(struct intel_dp *intel_dp, int index) * The clock divider is based off the hrawclk, and would like to run at * 2MHz. So, take the hrawclk value and divide by 2000 and use that */ - return DIV_ROUND_CLOSEST(DISPLAY_RUNTIME_INFO(i915)->rawclk_freq, 2000); + return DIV_ROUND_CLOSEST(DISPLAY_RUNTIME_INFO(display)->rawclk_freq, 2000); } static u32 ilk_get_aux_clock_divider(struct intel_dp *intel_dp, int index) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); u32 freq; @@ -101,15 +103,16 @@ static u32 ilk_get_aux_clock_divider(struct intel_dp *intel_dp, int index) * divide by 2000 and use that */ if (dig_port->aux_ch == AUX_CH_A) - freq = i915->display.cdclk.hw.cdclk; + freq = display->cdclk.hw.cdclk; else - freq = DISPLAY_RUNTIME_INFO(i915)->rawclk_freq; + freq = DISPLAY_RUNTIME_INFO(display)->rawclk_freq; return DIV_ROUND_CLOSEST(freq, 2000); } static u32 hsw_get_aux_clock_divider(struct intel_dp *intel_dp, int index) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *i915 = to_i915(display->drm); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); if (dig_port->aux_ch != AUX_CH_A && HAS_PCH_LPT_H(i915)) { @@ -142,16 +145,21 @@ static int intel_dp_aux_sync_len(void) return precharge + preamble; } -int intel_dp_aux_fw_sync_len(void) +int intel_dp_aux_fw_sync_len(struct intel_dp *intel_dp) { + int precharge = 10; /* 10-16 */ + int preamble = 8; + /* * We faced some glitches on Dell Precision 5490 MTL laptop with panel: * "Manufacturer: AUO, Model: 63898" when using HW default 18. Using 20 * is fixing these problems with the panel. It is still within range - * mentioned in eDP specification. + * mentioned in eDP specification. Increasing Fast Wake sync length is + * causing problems with other panels: increase length as a quirk for + * this specific laptop. */ - int precharge = 12; /* 10-16 */ - int preamble = 8; + if (intel_has_dpcd_quirk(intel_dp, QUIRK_FW_SYNC_LEN)) + precharge += 2; return precharge + preamble; } @@ -195,8 +203,8 @@ static u32 skl_get_aux_send_ctl(struct intel_dp *intel_dp, int send_bytes, u32 unused) { + struct intel_display *display = to_intel_display(intel_dp); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); - struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev); u32 ret; /* @@ -211,7 +219,7 @@ static u32 skl_get_aux_send_ctl(struct intel_dp *intel_dp, DP_AUX_CH_CTL_TIME_OUT_MAX | DP_AUX_CH_CTL_RECEIVE_ERROR | DP_AUX_CH_CTL_MESSAGE_SIZE(send_bytes) | - DP_AUX_CH_CTL_FW_SYNC_PULSE_SKL(intel_dp_aux_fw_sync_len()) | + DP_AUX_CH_CTL_FW_SYNC_PULSE_SKL(intel_dp_aux_fw_sync_len(intel_dp)) | DP_AUX_CH_CTL_SYNC_PULSE_SKL(intel_dp_aux_sync_len()); if (intel_tc_port_in_tbt_alt_mode(dig_port)) @@ -221,7 +229,7 @@ static u32 skl_get_aux_send_ctl(struct intel_dp *intel_dp, * Power request bit is already set during aux power well enable. * Preserve the bit across aux transactions. */ - if (DISPLAY_VER(i915) >= 14) + if (DISPLAY_VER(display) >= 14) ret |= XELPDP_DP_AUX_CH_CTL_POWER_REQUEST; return ret; @@ -233,6 +241,7 @@ intel_dp_aux_xfer(struct intel_dp *intel_dp, u8 *recv, int recv_size, u32 aux_send_ctl_flags) { + struct intel_display *display = to_intel_display(intel_dp); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); struct intel_encoder *encoder = &dig_port->base; struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev); @@ -291,7 +300,7 @@ intel_dp_aux_xfer(struct intel_dp *intel_dp, /* Try to wait for any previous AUX channel activity */ for (try = 0; try < 3; try++) { - status = intel_de_read_notrace(i915, ch_ctl); + status = intel_de_read_notrace(display, ch_ctl); if ((status & DP_AUX_CH_CTL_SEND_BUSY) == 0) break; msleep(1); @@ -300,10 +309,10 @@ intel_dp_aux_xfer(struct intel_dp *intel_dp, trace_i915_reg_rw(false, ch_ctl, status, sizeof(status), true); if (try == 3) { - const u32 status = intel_de_read(i915, ch_ctl); + const u32 status = intel_de_read(display, ch_ctl); if (status != intel_dp->aux_busy_last_status) { - drm_WARN(&i915->drm, 1, + drm_WARN(display->drm, 1, "%s: not started (status 0x%08x)\n", intel_dp->aux.name, status); intel_dp->aux_busy_last_status = status; @@ -314,7 +323,7 @@ intel_dp_aux_xfer(struct intel_dp *intel_dp, } /* Only 5 data registers! */ - if (drm_WARN_ON(&i915->drm, send_bytes > 20 || recv_size > 20)) { + if (drm_WARN_ON(display->drm, send_bytes > 20 || recv_size > 20)) { ret = -E2BIG; goto out; } @@ -330,17 +339,17 @@ intel_dp_aux_xfer(struct intel_dp *intel_dp, for (try = 0; try < 5; try++) { /* Load the send data into the aux channel data registers */ for (i = 0; i < send_bytes; i += 4) - intel_de_write(i915, ch_data[i >> 2], + intel_de_write(display, ch_data[i >> 2], intel_dp_aux_pack(send + i, send_bytes - i)); /* Send the command and wait for it to complete */ - intel_de_write(i915, ch_ctl, send_ctl); + intel_de_write(display, ch_ctl, send_ctl); status = intel_dp_aux_wait_done(intel_dp); /* Clear done status and any errors */ - intel_de_write(i915, ch_ctl, + intel_de_write(display, ch_ctl, status | DP_AUX_CH_CTL_DONE | DP_AUX_CH_CTL_TIME_OUT_ERROR | DP_AUX_CH_CTL_RECEIVE_ERROR); @@ -364,7 +373,7 @@ intel_dp_aux_xfer(struct intel_dp *intel_dp, } if ((status & DP_AUX_CH_CTL_DONE) == 0) { - drm_err(&i915->drm, "%s: not done (status 0x%08x)\n", + drm_err(display->drm, "%s: not done (status 0x%08x)\n", intel_dp->aux.name, status); ret = -EBUSY; goto out; @@ -376,7 +385,7 @@ done: * not connected. */ if (status & DP_AUX_CH_CTL_RECEIVE_ERROR) { - drm_err(&i915->drm, "%s: receive error (status 0x%08x)\n", + drm_err(display->drm, "%s: receive error (status 0x%08x)\n", intel_dp->aux.name, status); ret = -EIO; goto out; @@ -387,7 +396,7 @@ done: * -- don't fill the kernel log with these */ if (status & DP_AUX_CH_CTL_TIME_OUT_ERROR) { - drm_dbg_kms(&i915->drm, "%s: timeout (status 0x%08x)\n", + drm_dbg_kms(display->drm, "%s: timeout (status 0x%08x)\n", intel_dp->aux.name, status); ret = -ETIMEDOUT; goto out; @@ -402,7 +411,7 @@ done: * drm layer takes care for the necessary retries. */ if (recv_bytes == 0 || recv_bytes > 20) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "%s: Forbidden recv_bytes = %d on aux transaction\n", intel_dp->aux.name, recv_bytes); ret = -EBUSY; @@ -413,7 +422,7 @@ done: recv_bytes = recv_size; for (i = 0; i < recv_bytes; i += 4) - intel_dp_aux_unpack(intel_de_read(i915, ch_data[i >> 2]), + intel_dp_aux_unpack(intel_de_read(display, ch_data[i >> 2]), recv + i, recv_bytes - i); ret = recv_bytes; @@ -462,7 +471,7 @@ static ssize_t intel_dp_aux_transfer(struct drm_dp_aux *aux, struct drm_dp_aux_msg *msg) { struct intel_dp *intel_dp = container_of(aux, struct intel_dp, aux); - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); u8 txbuf[20], rxbuf[20]; size_t txsize, rxsize; u32 flags = intel_dp_aux_xfer_flags(msg); @@ -477,10 +486,10 @@ intel_dp_aux_transfer(struct drm_dp_aux *aux, struct drm_dp_aux_msg *msg) txsize = msg->size ? HEADER_SIZE + msg->size : BARE_ADDRESS_SIZE; rxsize = 2; /* 0 or 1 data bytes */ - if (drm_WARN_ON(&i915->drm, txsize > 20)) + if (drm_WARN_ON(display->drm, txsize > 20)) return -E2BIG; - drm_WARN_ON(&i915->drm, !msg->buffer != !msg->size); + drm_WARN_ON(display->drm, !msg->buffer != !msg->size); if (msg->buffer) memcpy(txbuf + HEADER_SIZE, msg->buffer, msg->size); @@ -505,7 +514,7 @@ intel_dp_aux_transfer(struct drm_dp_aux *aux, struct drm_dp_aux_msg *msg) txsize = msg->size ? HEADER_SIZE : BARE_ADDRESS_SIZE; rxsize = msg->size + 1; - if (drm_WARN_ON(&i915->drm, rxsize > 20)) + if (drm_WARN_ON(display->drm, rxsize > 20)) return -E2BIG; ret = intel_dp_aux_xfer(intel_dp, txbuf, txsize, @@ -715,7 +724,7 @@ static i915_reg_t tgl_aux_data_reg(struct intel_dp *intel_dp, int index) static i915_reg_t xelpdp_aux_ctl_reg(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); enum aux_ch aux_ch = dig_port->aux_ch; @@ -726,16 +735,16 @@ static i915_reg_t xelpdp_aux_ctl_reg(struct intel_dp *intel_dp) case AUX_CH_USBC2: case AUX_CH_USBC3: case AUX_CH_USBC4: - return XELPDP_DP_AUX_CH_CTL(i915, aux_ch); + return XELPDP_DP_AUX_CH_CTL(display, aux_ch); default: MISSING_CASE(aux_ch); - return XELPDP_DP_AUX_CH_CTL(i915, AUX_CH_A); + return XELPDP_DP_AUX_CH_CTL(display, AUX_CH_A); } } static i915_reg_t xelpdp_aux_data_reg(struct intel_dp *intel_dp, int index) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); enum aux_ch aux_ch = dig_port->aux_ch; @@ -746,10 +755,10 @@ static i915_reg_t xelpdp_aux_data_reg(struct intel_dp *intel_dp, int index) case AUX_CH_USBC2: case AUX_CH_USBC3: case AUX_CH_USBC4: - return XELPDP_DP_AUX_CH_DATA(i915, aux_ch, index); + return XELPDP_DP_AUX_CH_DATA(display, aux_ch, index); default: MISSING_CASE(aux_ch); - return XELPDP_DP_AUX_CH_DATA(i915, AUX_CH_A, index); + return XELPDP_DP_AUX_CH_DATA(display, AUX_CH_A, index); } } @@ -763,19 +772,20 @@ void intel_dp_aux_fini(struct intel_dp *intel_dp) void intel_dp_aux_init(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *i915 = to_i915(display->drm); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); struct intel_encoder *encoder = &dig_port->base; enum aux_ch aux_ch = dig_port->aux_ch; char buf[AUX_CH_NAME_BUFSIZE]; - if (DISPLAY_VER(i915) >= 14) { + if (DISPLAY_VER(display) >= 14) { intel_dp->aux_ch_ctl_reg = xelpdp_aux_ctl_reg; intel_dp->aux_ch_data_reg = xelpdp_aux_data_reg; - } else if (DISPLAY_VER(i915) >= 12) { + } else if (DISPLAY_VER(display) >= 12) { intel_dp->aux_ch_ctl_reg = tgl_aux_ctl_reg; intel_dp->aux_ch_data_reg = tgl_aux_data_reg; - } else if (DISPLAY_VER(i915) >= 9) { + } else if (DISPLAY_VER(display) >= 9) { intel_dp->aux_ch_ctl_reg = skl_aux_ctl_reg; intel_dp->aux_ch_data_reg = skl_aux_data_reg; } else if (HAS_PCH_SPLIT(i915)) { @@ -789,7 +799,7 @@ void intel_dp_aux_init(struct intel_dp *intel_dp) intel_dp->aux_ch_data_reg = g4x_aux_data_reg; } - if (DISPLAY_VER(i915) >= 9) + if (DISPLAY_VER(display) >= 9) intel_dp->get_aux_clock_divider = skl_get_aux_clock_divider; else if (IS_BROADWELL(i915) || IS_HASWELL(i915)) intel_dp->get_aux_clock_divider = hsw_get_aux_clock_divider; @@ -798,17 +808,17 @@ void intel_dp_aux_init(struct intel_dp *intel_dp) else intel_dp->get_aux_clock_divider = g4x_get_aux_clock_divider; - if (DISPLAY_VER(i915) >= 9) + if (DISPLAY_VER(display) >= 9) intel_dp->get_aux_send_ctl = skl_get_aux_send_ctl; else intel_dp->get_aux_send_ctl = g4x_get_aux_send_ctl; - intel_dp->aux.drm_dev = &i915->drm; + intel_dp->aux.drm_dev = display->drm; drm_dp_aux_init(&intel_dp->aux); /* Failure to allocate our preferred name is not critical */ intel_dp->aux.name = kasprintf(GFP_KERNEL, "AUX %s/%s", - aux_ch_name(i915, buf, sizeof(buf), aux_ch), + aux_ch_name(display, buf, sizeof(buf), aux_ch), encoder->base.name); intel_dp->aux.transfer = intel_dp_aux_transfer; @@ -817,10 +827,10 @@ void intel_dp_aux_init(struct intel_dp *intel_dp) static enum aux_ch default_aux_ch(struct intel_encoder *encoder) { - struct drm_i915_private *i915 = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); /* SKL has DDI E but no AUX E */ - if (DISPLAY_VER(i915) == 9 && encoder->port == PORT_E) + if (DISPLAY_VER(display) == 9 && encoder->port == PORT_E) return AUX_CH_A; return (enum aux_ch)encoder->port; @@ -830,10 +840,10 @@ static struct intel_encoder * get_encoder_by_aux_ch(struct intel_encoder *encoder, enum aux_ch aux_ch) { - struct drm_i915_private *i915 = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); struct intel_encoder *other; - for_each_intel_encoder(&i915->drm, other) { + for_each_intel_encoder(display->drm, other) { if (other == encoder) continue; @@ -849,7 +859,7 @@ get_encoder_by_aux_ch(struct intel_encoder *encoder, enum aux_ch intel_dp_aux_ch(struct intel_encoder *encoder) { - struct drm_i915_private *i915 = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); struct intel_encoder *other; const char *source; enum aux_ch aux_ch; @@ -870,23 +880,23 @@ enum aux_ch intel_dp_aux_ch(struct intel_encoder *encoder) other = get_encoder_by_aux_ch(encoder, aux_ch); if (other) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[ENCODER:%d:%s] AUX CH %s already claimed by [ENCODER:%d:%s]\n", encoder->base.base.id, encoder->base.name, - aux_ch_name(i915, buf, sizeof(buf), aux_ch), + aux_ch_name(display, buf, sizeof(buf), aux_ch), other->base.base.id, other->base.name); return AUX_CH_NONE; } - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[ENCODER:%d:%s] Using AUX CH %s (%s)\n", encoder->base.base.id, encoder->base.name, - aux_ch_name(i915, buf, sizeof(buf), aux_ch), source); + aux_ch_name(display, buf, sizeof(buf), aux_ch), source); return aux_ch; } -void intel_dp_aux_irq_handler(struct drm_i915_private *i915) +void intel_dp_aux_irq_handler(struct intel_display *display) { - wake_up_all(&i915->display.gmbus.wait_queue); + wake_up_all(&display->gmbus.wait_queue); } diff --git a/drivers/gpu/drm/i915/display/intel_dp_aux.h b/drivers/gpu/drm/i915/display/intel_dp_aux.h index 76d1f2ed7c2f..90ee1c5fae28 100644 --- a/drivers/gpu/drm/i915/display/intel_dp_aux.h +++ b/drivers/gpu/drm/i915/display/intel_dp_aux.h @@ -9,7 +9,7 @@ #include <linux/types.h> enum aux_ch; -struct drm_i915_private; +struct intel_display; struct intel_dp; struct intel_encoder; @@ -18,8 +18,8 @@ void intel_dp_aux_init(struct intel_dp *intel_dp); enum aux_ch intel_dp_aux_ch(struct intel_encoder *encoder); -void intel_dp_aux_irq_handler(struct drm_i915_private *i915); +void intel_dp_aux_irq_handler(struct intel_display *display); u32 intel_dp_aux_pack(const u8 *src, int src_bytes); -int intel_dp_aux_fw_sync_len(void); +int intel_dp_aux_fw_sync_len(struct intel_dp *intel_dp); #endif /* __INTEL_DP_AUX_H__ */ diff --git a/drivers/gpu/drm/i915/display/intel_dp_link_training.c b/drivers/gpu/drm/i915/display/intel_dp_link_training.c index 9c8738295106..40bedc31d6bf 100644 --- a/drivers/gpu/drm/i915/display/intel_dp_link_training.c +++ b/drivers/gpu/drm/i915/display/intel_dp_link_training.c @@ -39,13 +39,13 @@ drm_dp_phy_name(_dp_phy) #define lt_dbg(_intel_dp, _dp_phy, _format, ...) \ - drm_dbg_kms(&dp_to_i915(_intel_dp)->drm, \ + drm_dbg_kms(to_intel_display(_intel_dp)->drm, \ LT_MSG_PREFIX _format, \ LT_MSG_ARGS(_intel_dp, _dp_phy), ## __VA_ARGS__) #define lt_err(_intel_dp, _dp_phy, _format, ...) do { \ if (intel_digital_port_connected(&dp_to_dig_port(_intel_dp)->base)) \ - drm_err(&dp_to_i915(_intel_dp)->drm, \ + drm_err(to_intel_display(_intel_dp)->drm, \ LT_MSG_PREFIX _format, \ LT_MSG_ARGS(_intel_dp, _dp_phy), ## __VA_ARGS__); \ else \ @@ -216,7 +216,8 @@ static int intel_dp_init_lttpr(struct intel_dp *intel_dp, const u8 dpcd[DP_RECEI int intel_dp_read_dprx_caps(struct intel_dp *intel_dp, u8 dpcd[DP_RECEIVER_CAP_SIZE]) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *i915 = to_i915(display->drm); if (intel_dp_is_edp(intel_dp)) return 0; @@ -225,7 +226,7 @@ int intel_dp_read_dprx_caps(struct intel_dp *intel_dp, u8 dpcd[DP_RECEIVER_CAP_S * Detecting LTTPRs must be avoided on platforms with an AUX timeout * period < 3.2ms. (see DP Standard v2.0, 2.11.2, 3.6.6.1). */ - if (DISPLAY_VER(i915) >= 10 && !IS_GEMINILAKE(i915)) + if (DISPLAY_VER(display) >= 10 && !IS_GEMINILAKE(i915)) if (drm_dp_dpcd_probe(&intel_dp->aux, DP_LT_TUNABLE_PHY_REPEATER_FIELD_DATA_STRUCTURE_REV)) return -EIO; @@ -256,7 +257,8 @@ int intel_dp_read_dprx_caps(struct intel_dp *intel_dp, u8 dpcd[DP_RECEIVER_CAP_S */ int intel_dp_init_lttpr_and_dprx_caps(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *i915 = to_i915(display->drm); int lttpr_count = 0; /* @@ -264,7 +266,7 @@ int intel_dp_init_lttpr_and_dprx_caps(struct intel_dp *intel_dp) * period < 3.2ms. (see DP Standard v2.0, 2.11.2, 3.6.6.1). */ if (!intel_dp_is_edp(intel_dp) && - (DISPLAY_VER(i915) >= 10 && !IS_GEMINILAKE(i915))) { + (DISPLAY_VER(display) >= 10 && !IS_GEMINILAKE(i915))) { u8 dpcd[DP_RECEIVER_CAP_SIZE]; int err = intel_dp_read_dprx_caps(intel_dp, dpcd); @@ -327,10 +329,11 @@ static bool intel_dp_phy_is_downstream_of_source(struct intel_dp *intel_dp, enum drm_dp_phy dp_phy) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); int lttpr_count = drm_dp_lttpr_count(intel_dp->lttpr_common_caps); - drm_WARN_ON_ONCE(&i915->drm, lttpr_count <= 0 && dp_phy != DP_PHY_DPRX); + drm_WARN_ON_ONCE(display->drm, + lttpr_count <= 0 && dp_phy != DP_PHY_DPRX); return lttpr_count <= 0 || dp_phy == DP_PHY_LTTPR(lttpr_count - 1); } @@ -339,7 +342,7 @@ static u8 intel_dp_phy_voltage_max(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state, enum drm_dp_phy dp_phy) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); u8 voltage_max; /* @@ -351,7 +354,7 @@ static u8 intel_dp_phy_voltage_max(struct intel_dp *intel_dp, else voltage_max = intel_dp_lttpr_voltage_max(intel_dp, dp_phy + 1); - drm_WARN_ON_ONCE(&i915->drm, + drm_WARN_ON_ONCE(display->drm, voltage_max != DP_TRAIN_VOLTAGE_SWING_LEVEL_2 && voltage_max != DP_TRAIN_VOLTAGE_SWING_LEVEL_3); @@ -361,7 +364,7 @@ static u8 intel_dp_phy_voltage_max(struct intel_dp *intel_dp, static u8 intel_dp_phy_preemph_max(struct intel_dp *intel_dp, enum drm_dp_phy dp_phy) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); u8 preemph_max; /* @@ -373,7 +376,7 @@ static u8 intel_dp_phy_preemph_max(struct intel_dp *intel_dp, else preemph_max = intel_dp_lttpr_preemph_max(intel_dp, dp_phy + 1); - drm_WARN_ON_ONCE(&i915->drm, + drm_WARN_ON_ONCE(display->drm, preemph_max != DP_TRAIN_PRE_EMPH_LEVEL_2 && preemph_max != DP_TRAIN_PRE_EMPH_LEVEL_3); @@ -383,10 +386,11 @@ static u8 intel_dp_phy_preemph_max(struct intel_dp *intel_dp, static bool has_per_lane_signal_levels(struct intel_dp *intel_dp, enum drm_dp_phy dp_phy) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *i915 = to_i915(display->drm); return !intel_dp_phy_is_downstream_of_source(intel_dp, dp_phy) || - DISPLAY_VER(i915) >= 10 || IS_BROXTON(i915); + DISPLAY_VER(display) >= 10 || IS_BROXTON(i915); } /* 128b/132b */ @@ -950,7 +954,8 @@ static u32 intel_dp_training_pattern(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state, enum drm_dp_phy dp_phy) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *i915 = to_i915(display->drm); bool source_tps3, sink_tps3, source_tps4, sink_tps4; /* UHBR+ use separate 128b/132b TPS2 */ @@ -1586,7 +1591,7 @@ void intel_dp_start_link_train(struct intel_atomic_state *state, struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(state); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); struct intel_encoder *encoder = &dig_port->base; bool passed; @@ -1631,7 +1636,7 @@ void intel_dp_start_link_train(struct intel_atomic_state *state, * For test cases which rely on the link training or processing of HPDs * ignore_long_hpd flag can unset from the testcase. */ - if (i915->display.hotplug.ignore_long_hpd) { + if (display->hotplug.ignore_long_hpd) { lt_dbg(intel_dp, DP_PHY_DPRX, "Ignore the link failure\n"); return; } @@ -1683,14 +1688,14 @@ static struct intel_dp *intel_connector_to_intel_dp(struct intel_connector *conn static int i915_dp_force_link_rate_show(struct seq_file *m, void *data) { struct intel_connector *connector = to_intel_connector(m->private); - struct drm_i915_private *i915 = to_i915(connector->base.dev); + struct intel_display *display = to_intel_display(connector); struct intel_dp *intel_dp = intel_connector_to_intel_dp(connector); int current_rate = -1; int force_rate; int err; int i; - err = drm_modeset_lock_single_interruptible(&i915->drm.mode_config.connection_mutex); + err = drm_modeset_lock_single_interruptible(&display->drm->mode_config.connection_mutex); if (err) return err; @@ -1698,7 +1703,7 @@ static int i915_dp_force_link_rate_show(struct seq_file *m, void *data) current_rate = intel_dp->link_rate; force_rate = intel_dp->link.force_rate; - drm_modeset_unlock(&i915->drm.mode_config.connection_mutex); + drm_modeset_unlock(&display->drm->mode_config.connection_mutex); seq_printf(m, "%sauto%s", force_rate == 0 ? "[" : "", @@ -1754,7 +1759,7 @@ static ssize_t i915_dp_force_link_rate_write(struct file *file, { struct seq_file *m = file->private_data; struct intel_connector *connector = to_intel_connector(m->private); - struct drm_i915_private *i915 = to_i915(connector->base.dev); + struct intel_display *display = to_intel_display(connector); struct intel_dp *intel_dp = intel_connector_to_intel_dp(connector); int rate; int err; @@ -1763,14 +1768,14 @@ static ssize_t i915_dp_force_link_rate_write(struct file *file, if (rate < 0) return rate; - err = drm_modeset_lock_single_interruptible(&i915->drm.mode_config.connection_mutex); + err = drm_modeset_lock_single_interruptible(&display->drm->mode_config.connection_mutex); if (err) return err; intel_dp_reset_link_params(intel_dp); intel_dp->link.force_rate = rate; - drm_modeset_unlock(&i915->drm.mode_config.connection_mutex); + drm_modeset_unlock(&display->drm->mode_config.connection_mutex); *offp += len; @@ -1781,14 +1786,14 @@ DEFINE_SHOW_STORE_ATTRIBUTE(i915_dp_force_link_rate); static int i915_dp_force_lane_count_show(struct seq_file *m, void *data) { struct intel_connector *connector = to_intel_connector(m->private); - struct drm_i915_private *i915 = to_i915(connector->base.dev); + struct intel_display *display = to_intel_display(connector); struct intel_dp *intel_dp = intel_connector_to_intel_dp(connector); int current_lane_count = -1; int force_lane_count; int err; int i; - err = drm_modeset_lock_single_interruptible(&i915->drm.mode_config.connection_mutex); + err = drm_modeset_lock_single_interruptible(&display->drm->mode_config.connection_mutex); if (err) return err; @@ -1796,7 +1801,7 @@ static int i915_dp_force_lane_count_show(struct seq_file *m, void *data) current_lane_count = intel_dp->lane_count; force_lane_count = intel_dp->link.force_lane_count; - drm_modeset_unlock(&i915->drm.mode_config.connection_mutex); + drm_modeset_unlock(&display->drm->mode_config.connection_mutex); seq_printf(m, "%sauto%s", force_lane_count == 0 ? "[" : "", @@ -1856,7 +1861,7 @@ static ssize_t i915_dp_force_lane_count_write(struct file *file, { struct seq_file *m = file->private_data; struct intel_connector *connector = to_intel_connector(m->private); - struct drm_i915_private *i915 = to_i915(connector->base.dev); + struct intel_display *display = to_intel_display(connector); struct intel_dp *intel_dp = intel_connector_to_intel_dp(connector); int lane_count; int err; @@ -1865,14 +1870,14 @@ static ssize_t i915_dp_force_lane_count_write(struct file *file, if (lane_count < 0) return lane_count; - err = drm_modeset_lock_single_interruptible(&i915->drm.mode_config.connection_mutex); + err = drm_modeset_lock_single_interruptible(&display->drm->mode_config.connection_mutex); if (err) return err; intel_dp_reset_link_params(intel_dp); intel_dp->link.force_lane_count = lane_count; - drm_modeset_unlock(&i915->drm.mode_config.connection_mutex); + drm_modeset_unlock(&display->drm->mode_config.connection_mutex); *offp += len; @@ -1883,17 +1888,17 @@ DEFINE_SHOW_STORE_ATTRIBUTE(i915_dp_force_lane_count); static int i915_dp_max_link_rate_show(void *data, u64 *val) { struct intel_connector *connector = to_intel_connector(data); - struct drm_i915_private *i915 = to_i915(connector->base.dev); + struct intel_display *display = to_intel_display(connector); struct intel_dp *intel_dp = intel_connector_to_intel_dp(connector); int err; - err = drm_modeset_lock_single_interruptible(&i915->drm.mode_config.connection_mutex); + err = drm_modeset_lock_single_interruptible(&display->drm->mode_config.connection_mutex); if (err) return err; *val = intel_dp->link.max_rate; - drm_modeset_unlock(&i915->drm.mode_config.connection_mutex); + drm_modeset_unlock(&display->drm->mode_config.connection_mutex); return 0; } @@ -1902,17 +1907,17 @@ DEFINE_DEBUGFS_ATTRIBUTE(i915_dp_max_link_rate_fops, i915_dp_max_link_rate_show, static int i915_dp_max_lane_count_show(void *data, u64 *val) { struct intel_connector *connector = to_intel_connector(data); - struct drm_i915_private *i915 = to_i915(connector->base.dev); + struct intel_display *display = to_intel_display(connector); struct intel_dp *intel_dp = intel_connector_to_intel_dp(connector); int err; - err = drm_modeset_lock_single_interruptible(&i915->drm.mode_config.connection_mutex); + err = drm_modeset_lock_single_interruptible(&display->drm->mode_config.connection_mutex); if (err) return err; *val = intel_dp->link.max_lane_count; - drm_modeset_unlock(&i915->drm.mode_config.connection_mutex); + drm_modeset_unlock(&display->drm->mode_config.connection_mutex); return 0; } @@ -1921,17 +1926,17 @@ DEFINE_DEBUGFS_ATTRIBUTE(i915_dp_max_lane_count_fops, i915_dp_max_lane_count_sho static int i915_dp_force_link_training_failure_show(void *data, u64 *val) { struct intel_connector *connector = to_intel_connector(data); - struct drm_i915_private *i915 = to_i915(connector->base.dev); + struct intel_display *display = to_intel_display(connector); struct intel_dp *intel_dp = intel_connector_to_intel_dp(connector); int err; - err = drm_modeset_lock_single_interruptible(&i915->drm.mode_config.connection_mutex); + err = drm_modeset_lock_single_interruptible(&display->drm->mode_config.connection_mutex); if (err) return err; *val = intel_dp->link.force_train_failure; - drm_modeset_unlock(&i915->drm.mode_config.connection_mutex); + drm_modeset_unlock(&display->drm->mode_config.connection_mutex); return 0; } @@ -1939,20 +1944,20 @@ static int i915_dp_force_link_training_failure_show(void *data, u64 *val) static int i915_dp_force_link_training_failure_write(void *data, u64 val) { struct intel_connector *connector = to_intel_connector(data); - struct drm_i915_private *i915 = to_i915(connector->base.dev); + struct intel_display *display = to_intel_display(connector); struct intel_dp *intel_dp = intel_connector_to_intel_dp(connector); int err; if (val > 2) return -EINVAL; - err = drm_modeset_lock_single_interruptible(&i915->drm.mode_config.connection_mutex); + err = drm_modeset_lock_single_interruptible(&display->drm->mode_config.connection_mutex); if (err) return err; intel_dp->link.force_train_failure = val; - drm_modeset_unlock(&i915->drm.mode_config.connection_mutex); + drm_modeset_unlock(&display->drm->mode_config.connection_mutex); return 0; } @@ -1963,17 +1968,17 @@ DEFINE_DEBUGFS_ATTRIBUTE(i915_dp_force_link_training_failure_fops, static int i915_dp_force_link_retrain_show(void *data, u64 *val) { struct intel_connector *connector = to_intel_connector(data); - struct drm_i915_private *i915 = to_i915(connector->base.dev); + struct intel_display *display = to_intel_display(connector); struct intel_dp *intel_dp = intel_connector_to_intel_dp(connector); int err; - err = drm_modeset_lock_single_interruptible(&i915->drm.mode_config.connection_mutex); + err = drm_modeset_lock_single_interruptible(&display->drm->mode_config.connection_mutex); if (err) return err; *val = intel_dp->link.force_retrain; - drm_modeset_unlock(&i915->drm.mode_config.connection_mutex); + drm_modeset_unlock(&display->drm->mode_config.connection_mutex); return 0; } @@ -1981,17 +1986,17 @@ static int i915_dp_force_link_retrain_show(void *data, u64 *val) static int i915_dp_force_link_retrain_write(void *data, u64 val) { struct intel_connector *connector = to_intel_connector(data); - struct drm_i915_private *i915 = to_i915(connector->base.dev); + struct intel_display *display = to_intel_display(connector); struct intel_dp *intel_dp = intel_connector_to_intel_dp(connector); int err; - err = drm_modeset_lock_single_interruptible(&i915->drm.mode_config.connection_mutex); + err = drm_modeset_lock_single_interruptible(&display->drm->mode_config.connection_mutex); if (err) return err; intel_dp->link.force_retrain = val; - drm_modeset_unlock(&i915->drm.mode_config.connection_mutex); + drm_modeset_unlock(&display->drm->mode_config.connection_mutex); intel_hpd_trigger_irq(dp_to_dig_port(intel_dp)); @@ -2004,17 +2009,17 @@ DEFINE_DEBUGFS_ATTRIBUTE(i915_dp_force_link_retrain_fops, static int i915_dp_link_retrain_disabled_show(struct seq_file *m, void *data) { struct intel_connector *connector = to_intel_connector(m->private); - struct drm_i915_private *i915 = to_i915(connector->base.dev); + struct intel_display *display = to_intel_display(connector); struct intel_dp *intel_dp = intel_connector_to_intel_dp(connector); int err; - err = drm_modeset_lock_single_interruptible(&i915->drm.mode_config.connection_mutex); + err = drm_modeset_lock_single_interruptible(&display->drm->mode_config.connection_mutex); if (err) return err; seq_printf(m, "%s\n", str_yes_no(intel_dp->link.retrain_disabled)); - drm_modeset_unlock(&i915->drm.mode_config.connection_mutex); + drm_modeset_unlock(&display->drm->mode_config.connection_mutex); return 0; } diff --git a/drivers/gpu/drm/i915/display/intel_dp_tunnel.c b/drivers/gpu/drm/i915/display/intel_dp_tunnel.c index 6503abdc2b98..94198bc04939 100644 --- a/drivers/gpu/drm/i915/display/intel_dp_tunnel.c +++ b/drivers/gpu/drm/i915/display/intel_dp_tunnel.c @@ -69,7 +69,7 @@ static int get_current_link_bw(struct intel_dp *intel_dp, static int update_tunnel_state(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base; bool old_bw_below_dprx; bool new_bw_below_dprx; @@ -81,7 +81,7 @@ static int update_tunnel_state(struct intel_dp *intel_dp) ret = drm_dp_tunnel_update_state(intel_dp->tunnel); if (ret < 0) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[DPTUN %s][ENCODER:%d:%s] State update failed (err %pe)\n", drm_dp_tunnel_name(intel_dp->tunnel), encoder->base.base.id, encoder->base.name, @@ -103,7 +103,7 @@ static int update_tunnel_state(struct intel_dp *intel_dp) !new_bw_below_dprx) return 0; - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[DPTUN %s][ENCODER:%d:%s] Notify users about BW change: %d -> %d\n", drm_dp_tunnel_name(intel_dp->tunnel), encoder->base.base.id, encoder->base.name, @@ -121,20 +121,20 @@ static int update_tunnel_state(struct intel_dp *intel_dp) */ static int allocate_initial_tunnel_bw_for_pipes(struct intel_dp *intel_dp, u8 pipe_mask) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base; struct intel_crtc *crtc; int tunnel_bw = 0; int err; - for_each_intel_crtc_in_pipe_mask(&i915->drm, crtc, pipe_mask) { + for_each_intel_crtc_in_pipe_mask(display->drm, crtc, pipe_mask) { const struct intel_crtc_state *crtc_state = to_intel_crtc_state(crtc->base.state); int stream_bw = intel_dp_config_required_rate(crtc_state); tunnel_bw += stream_bw; - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[DPTUN %s][ENCODER:%d:%s][CRTC:%d:%s] Initial BW for stream %d: %d/%d Mb/s\n", drm_dp_tunnel_name(intel_dp->tunnel), encoder->base.base.id, encoder->base.name, @@ -145,7 +145,7 @@ static int allocate_initial_tunnel_bw_for_pipes(struct intel_dp *intel_dp, u8 pi err = drm_dp_tunnel_alloc_bw(intel_dp->tunnel, tunnel_bw); if (err) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[DPTUN %s][ENCODER:%d:%s] Initial BW allocation failed (err %pe)\n", drm_dp_tunnel_name(intel_dp->tunnel), encoder->base.base.id, encoder->base.name, @@ -172,12 +172,12 @@ static int allocate_initial_tunnel_bw(struct intel_dp *intel_dp, static int detect_new_tunnel(struct intel_dp *intel_dp, struct drm_modeset_acquire_ctx *ctx) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base; struct drm_dp_tunnel *tunnel; int ret; - tunnel = drm_dp_tunnel_detect(i915->display.dp_tunnel_mgr, + tunnel = drm_dp_tunnel_detect(display->dp_tunnel_mgr, &intel_dp->aux); if (IS_ERR(tunnel)) return PTR_ERR(tunnel); @@ -189,7 +189,7 @@ static int detect_new_tunnel(struct intel_dp *intel_dp, struct drm_modeset_acqui if (ret == -EOPNOTSUPP) return 0; - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[DPTUN %s][ENCODER:%d:%s] Failed to enable BW allocation mode (ret %pe)\n", drm_dp_tunnel_name(intel_dp->tunnel), encoder->base.base.id, encoder->base.name, @@ -266,14 +266,15 @@ bool intel_dp_tunnel_bw_alloc_is_enabled(struct intel_dp *intel_dp) */ void intel_dp_tunnel_suspend(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct intel_connector *connector = intel_dp->attached_connector; struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base; if (!intel_dp_tunnel_bw_alloc_is_enabled(intel_dp)) return; - drm_dbg_kms(&i915->drm, "[DPTUN %s][CONNECTOR:%d:%s][ENCODER:%d:%s] Suspend\n", + drm_dbg_kms(display->drm, + "[DPTUN %s][CONNECTOR:%d:%s][ENCODER:%d:%s] Suspend\n", drm_dp_tunnel_name(intel_dp->tunnel), connector->base.base.id, connector->base.name, encoder->base.base.id, encoder->base.name); @@ -295,7 +296,7 @@ void intel_dp_tunnel_resume(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state, bool dpcd_updated) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct intel_connector *connector = intel_dp->attached_connector; struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base; u8 dpcd[DP_RECEIVER_CAP_SIZE]; @@ -307,7 +308,8 @@ void intel_dp_tunnel_resume(struct intel_dp *intel_dp, intel_dp->tunnel_suspended = false; - drm_dbg_kms(&i915->drm, "[DPTUN %s][CONNECTOR:%d:%s][ENCODER:%d:%s] Resume\n", + drm_dbg_kms(display->drm, + "[DPTUN %s][CONNECTOR:%d:%s][ENCODER:%d:%s] Resume\n", drm_dp_tunnel_name(intel_dp->tunnel), connector->base.base.id, connector->base.name, encoder->base.base.id, encoder->base.name); @@ -347,7 +349,7 @@ void intel_dp_tunnel_resume(struct intel_dp *intel_dp, return; out_err: - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[DPTUN %s][CONNECTOR:%d:%s][ENCODER:%d:%s] Tunnel can't be resumed, will drop and reject it (err %pe)\n", drm_dp_tunnel_name(intel_dp->tunnel), connector->base.base.id, connector->base.name, @@ -369,12 +371,12 @@ add_inherited_tunnel(struct intel_atomic_state *state, struct drm_dp_tunnel *tunnel, struct intel_crtc *crtc) { - struct drm_i915_private *i915 = to_i915(state->base.dev); + struct intel_display *display = to_intel_display(state); struct drm_dp_tunnel *old_tunnel; old_tunnel = get_inherited_tunnel(state, crtc); if (old_tunnel) { - drm_WARN_ON(&i915->drm, old_tunnel != tunnel); + drm_WARN_ON(display->drm, old_tunnel != tunnel); return 0; } @@ -394,7 +396,7 @@ static int check_inherited_tunnel_state(struct intel_atomic_state *state, struct intel_dp *intel_dp, const struct intel_digital_connector_state *old_conn_state) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(state); struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base; struct intel_connector *connector = to_intel_connector(old_conn_state->base.connector); @@ -422,7 +424,7 @@ static int check_inherited_tunnel_state(struct intel_atomic_state *state, old_crtc_state->dp_tunnel_ref.tunnel == intel_dp->tunnel) return 0; - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[DPTUN %s][CONNECTOR:%d:%s][ENCODER:%d:%s][CRTC:%d:%s] Adding state for inherited tunnel %p\n", drm_dp_tunnel_name(intel_dp->tunnel), connector->base.base.id, connector->base.name, @@ -441,12 +443,13 @@ static int check_inherited_tunnel_state(struct intel_atomic_state *state, */ void intel_dp_tunnel_atomic_cleanup_inherited_state(struct intel_atomic_state *state) { + struct intel_display *display = to_intel_display(state); enum pipe pipe; if (!state->inherited_dp_tunnels) return; - for_each_pipe(to_i915(state->base.dev), pipe) + for_each_pipe(display, pipe) if (state->inherited_dp_tunnels->ref[pipe].tunnel) drm_dp_tunnel_ref_put(&state->inherited_dp_tunnels->ref[pipe]); @@ -457,7 +460,7 @@ void intel_dp_tunnel_atomic_cleanup_inherited_state(struct intel_atomic_state *s static int intel_dp_tunnel_atomic_add_group_state(struct intel_atomic_state *state, struct drm_dp_tunnel *tunnel) { - struct drm_i915_private *i915 = to_i915(state->base.dev); + struct intel_display *display = to_intel_display(state); u32 pipe_mask; int err; @@ -466,7 +469,7 @@ static int intel_dp_tunnel_atomic_add_group_state(struct intel_atomic_state *sta if (err) return err; - drm_WARN_ON(&i915->drm, pipe_mask & ~((1 << I915_MAX_PIPES) - 1)); + drm_WARN_ON(display->drm, pipe_mask & ~((1 << I915_MAX_PIPES) - 1)); return intel_modeset_pipes_in_mask_early(state, "DPTUN", pipe_mask); } @@ -504,7 +507,7 @@ static int check_group_state(struct intel_atomic_state *state, struct intel_connector *connector, struct intel_crtc *crtc) { - struct drm_i915_private *i915 = to_i915(state->base.dev); + struct intel_display *display = to_intel_display(state); struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base; const struct intel_crtc_state *crtc_state = intel_atomic_get_new_crtc_state(state, crtc); @@ -512,7 +515,7 @@ static int check_group_state(struct intel_atomic_state *state, if (!crtc_state->dp_tunnel_ref.tunnel) return 0; - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[DPTUN %s][CONNECTOR:%d:%s][ENCODER:%d:%s][CRTC:%d:%s] Adding group state for tunnel %p\n", drm_dp_tunnel_name(intel_dp->tunnel), connector->base.base.id, connector->base.name, @@ -583,7 +586,7 @@ int intel_dp_tunnel_atomic_compute_stream_bw(struct intel_atomic_state *state, const struct intel_connector *connector, struct intel_crtc_state *crtc_state) { - struct drm_i915_private *i915 = to_i915(state->base.dev); + struct intel_display *display = to_intel_display(state); struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base; struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); int required_rate = intel_dp_config_required_rate(crtc_state); @@ -592,7 +595,7 @@ int intel_dp_tunnel_atomic_compute_stream_bw(struct intel_atomic_state *state, if (!intel_dp_tunnel_bw_alloc_is_enabled(intel_dp)) return 0; - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[DPTUN %s][CONNECTOR:%d:%s][ENCODER:%d:%s][CRTC:%d:%s] Stream %d required BW %d Mb/s\n", drm_dp_tunnel_name(intel_dp->tunnel), connector->base.base.id, connector->base.name, @@ -708,7 +711,7 @@ static void queue_retry_work(struct intel_atomic_state *state, struct drm_dp_tunnel *tunnel, const struct intel_crtc_state *crtc_state) { - struct drm_i915_private *i915 = to_i915(state->base.dev); + struct intel_display *display = to_intel_display(state); struct intel_encoder *encoder; encoder = intel_get_crtc_new_encoder(state, crtc_state); @@ -716,7 +719,7 @@ static void queue_retry_work(struct intel_atomic_state *state, if (!intel_digital_port_connected(encoder)) return; - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[DPTUN %s][ENCODER:%d:%s] BW allocation failed on a connected sink\n", drm_dp_tunnel_name(tunnel), encoder->base.base.id, @@ -765,7 +768,7 @@ void intel_dp_tunnel_atomic_alloc_bw(struct intel_atomic_state *state) /** * intel_dp_tunnel_mgr_init - Initialize the DP tunnel manager - * @i915: i915 device object + * @display: display device * * Initialize the DP tunnel manager. The tunnel manager will support the * detection/management of DP tunnels on all DP connectors, so the function @@ -773,14 +776,14 @@ void intel_dp_tunnel_atomic_alloc_bw(struct intel_atomic_state *state) * * Return 0 in case of success, a negative error code otherwise. */ -int intel_dp_tunnel_mgr_init(struct drm_i915_private *i915) +int intel_dp_tunnel_mgr_init(struct intel_display *display) { struct drm_dp_tunnel_mgr *tunnel_mgr; struct drm_connector_list_iter connector_list_iter; struct intel_connector *connector; int dp_connectors = 0; - drm_connector_list_iter_begin(&i915->drm, &connector_list_iter); + drm_connector_list_iter_begin(display->drm, &connector_list_iter); for_each_intel_connector_iter(connector, &connector_list_iter) { if (connector->base.connector_type != DRM_MODE_CONNECTOR_DisplayPort) continue; @@ -789,23 +792,23 @@ int intel_dp_tunnel_mgr_init(struct drm_i915_private *i915) } drm_connector_list_iter_end(&connector_list_iter); - tunnel_mgr = drm_dp_tunnel_mgr_create(&i915->drm, dp_connectors); + tunnel_mgr = drm_dp_tunnel_mgr_create(display->drm, dp_connectors); if (IS_ERR(tunnel_mgr)) return PTR_ERR(tunnel_mgr); - i915->display.dp_tunnel_mgr = tunnel_mgr; + display->dp_tunnel_mgr = tunnel_mgr; return 0; } /** * intel_dp_tunnel_mgr_cleanup - Clean up the DP tunnel manager state - * @i915: i915 device object + * @display: display device * * Clean up the DP tunnel manager state. */ -void intel_dp_tunnel_mgr_cleanup(struct drm_i915_private *i915) +void intel_dp_tunnel_mgr_cleanup(struct intel_display *display) { - drm_dp_tunnel_mgr_destroy(i915->display.dp_tunnel_mgr); - i915->display.dp_tunnel_mgr = NULL; + drm_dp_tunnel_mgr_destroy(display->dp_tunnel_mgr); + display->dp_tunnel_mgr = NULL; } diff --git a/drivers/gpu/drm/i915/display/intel_dp_tunnel.h b/drivers/gpu/drm/i915/display/intel_dp_tunnel.h index 08b2cba84af2..a0c00b7d3303 100644 --- a/drivers/gpu/drm/i915/display/intel_dp_tunnel.h +++ b/drivers/gpu/drm/i915/display/intel_dp_tunnel.h @@ -9,14 +9,13 @@ #include <linux/errno.h> #include <linux/types.h> -struct drm_i915_private; struct drm_connector_state; struct drm_modeset_acquire_ctx; - struct intel_atomic_state; struct intel_connector; struct intel_crtc; struct intel_crtc_state; +struct intel_display; struct intel_dp; struct intel_encoder; struct intel_link_bw_limits; @@ -53,8 +52,8 @@ int intel_dp_tunnel_atomic_check_state(struct intel_atomic_state *state, void intel_dp_tunnel_atomic_alloc_bw(struct intel_atomic_state *state); -int intel_dp_tunnel_mgr_init(struct drm_i915_private *i915); -void intel_dp_tunnel_mgr_cleanup(struct drm_i915_private *i915); +int intel_dp_tunnel_mgr_init(struct intel_display *display); +void intel_dp_tunnel_mgr_cleanup(struct intel_display *display); #else @@ -121,12 +120,12 @@ intel_dp_tunnel_atomic_alloc_bw(struct intel_atomic_state *state) } static inline int -intel_dp_tunnel_mgr_init(struct drm_i915_private *i915) +intel_dp_tunnel_mgr_init(struct intel_display *display) { return 0; } -static inline void intel_dp_tunnel_mgr_cleanup(struct drm_i915_private *i915) {} +static inline void intel_dp_tunnel_mgr_cleanup(struct intel_display *display) {} #endif /* CONFIG_DRM_I915_DP_TUNNEL */ diff --git a/drivers/gpu/drm/i915/display/intel_dpll.c b/drivers/gpu/drm/i915/display/intel_dpll.c index d67d5e2fd570..340dfce480b8 100644 --- a/drivers/gpu/drm/i915/display/intel_dpll.c +++ b/drivers/gpu/drm/i915/display/intel_dpll.c @@ -1823,6 +1823,7 @@ static bool i9xx_has_pps(struct drm_i915_private *dev_priv) void i9xx_enable_pll(const struct intel_crtc_state *crtc_state) { + struct intel_display *display = to_intel_display(crtc_state); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); const struct i9xx_dpll_hw_state *hw_state = &crtc_state->dpll_hw_state.i9xx; @@ -1833,7 +1834,7 @@ void i9xx_enable_pll(const struct intel_crtc_state *crtc_state) /* PLL is protected by panel, make sure we can write it */ if (i9xx_has_pps(dev_priv)) - assert_pps_unlocked(dev_priv, pipe); + assert_pps_unlocked(display, pipe); intel_de_write(dev_priv, FP0(pipe), hw_state->fp0); intel_de_write(dev_priv, FP1(pipe), hw_state->fp1); @@ -2004,6 +2005,7 @@ static void _vlv_enable_pll(const struct intel_crtc_state *crtc_state) void vlv_enable_pll(const struct intel_crtc_state *crtc_state) { + struct intel_display *display = to_intel_display(crtc_state); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); const struct i9xx_dpll_hw_state *hw_state = &crtc_state->dpll_hw_state.i9xx; @@ -2012,7 +2014,7 @@ void vlv_enable_pll(const struct intel_crtc_state *crtc_state) assert_transcoder_disabled(dev_priv, crtc_state->cpu_transcoder); /* PLL is protected by panel, make sure we can write it */ - assert_pps_unlocked(dev_priv, pipe); + assert_pps_unlocked(display, pipe); /* Enable Refclk */ intel_de_write(dev_priv, DPLL(dev_priv, pipe), @@ -2150,6 +2152,7 @@ static void _chv_enable_pll(const struct intel_crtc_state *crtc_state) void chv_enable_pll(const struct intel_crtc_state *crtc_state) { + struct intel_display *display = to_intel_display(crtc_state); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); const struct i9xx_dpll_hw_state *hw_state = &crtc_state->dpll_hw_state.i9xx; @@ -2158,7 +2161,7 @@ void chv_enable_pll(const struct intel_crtc_state *crtc_state) assert_transcoder_disabled(dev_priv, crtc_state->cpu_transcoder); /* PLL is protected by panel, make sure we can write it */ - assert_pps_unlocked(dev_priv, pipe); + assert_pps_unlocked(display, pipe); /* Enable Refclk and SSC */ intel_de_write(dev_priv, DPLL(dev_priv, pipe), diff --git a/drivers/gpu/drm/i915/display/intel_frontbuffer.c b/drivers/gpu/drm/i915/display/intel_frontbuffer.c index 4923c340a0b6..af4576dee92a 100644 --- a/drivers/gpu/drm/i915/display/intel_frontbuffer.c +++ b/drivers/gpu/drm/i915/display/intel_frontbuffer.c @@ -83,6 +83,8 @@ static void frontbuffer_flush(struct drm_i915_private *i915, unsigned int frontbuffer_bits, enum fb_op_origin origin) { + struct intel_display *display = &i915->display; + /* Delay flushing when rings are still busy.*/ spin_lock(&i915->display.fb_tracking.lock); frontbuffer_bits &= ~i915->display.fb_tracking.busy_bits; @@ -96,7 +98,7 @@ static void frontbuffer_flush(struct drm_i915_private *i915, might_sleep(); intel_td_flush(i915); intel_drrs_flush(i915, frontbuffer_bits); - intel_psr_flush(i915, frontbuffer_bits, origin); + intel_psr_flush(display, frontbuffer_bits, origin); intel_fbc_flush(i915, frontbuffer_bits, origin); } @@ -172,6 +174,7 @@ void __intel_fb_invalidate(struct intel_frontbuffer *front, unsigned int frontbuffer_bits) { struct drm_i915_private *i915 = intel_bo_to_i915(front->obj); + struct intel_display *display = &i915->display; if (origin == ORIGIN_CS) { spin_lock(&i915->display.fb_tracking.lock); @@ -183,7 +186,7 @@ void __intel_fb_invalidate(struct intel_frontbuffer *front, trace_intel_frontbuffer_invalidate(i915, frontbuffer_bits, origin); might_sleep(); - intel_psr_invalidate(i915, frontbuffer_bits, origin); + intel_psr_invalidate(display, frontbuffer_bits, origin); intel_drrs_invalidate(i915, frontbuffer_bits); intel_fbc_invalidate(i915, frontbuffer_bits, origin); } diff --git a/drivers/gpu/drm/i915/display/intel_hdcp.c b/drivers/gpu/drm/i915/display/intel_hdcp.c index 94418f218448..6980b98792c2 100644 --- a/drivers/gpu/drm/i915/display/intel_hdcp.c +++ b/drivers/gpu/drm/i915/display/intel_hdcp.c @@ -2181,10 +2181,11 @@ static void intel_hdcp_check_work(struct work_struct *work) DRM_HDCP_CHECK_PERIOD_MS); } -static int i915_hdcp_component_bind(struct device *i915_kdev, +static int i915_hdcp_component_bind(struct device *drv_kdev, struct device *mei_kdev, void *data) { - struct drm_i915_private *i915 = kdev_to_i915(i915_kdev); + struct intel_display *display = to_intel_display(drv_kdev); + struct drm_i915_private *i915 = to_i915(display->drm); drm_dbg(&i915->drm, "I915 HDCP comp bind\n"); mutex_lock(&i915->display.hdcp.hdcp_mutex); @@ -2195,10 +2196,11 @@ static int i915_hdcp_component_bind(struct device *i915_kdev, return 0; } -static void i915_hdcp_component_unbind(struct device *i915_kdev, +static void i915_hdcp_component_unbind(struct device *drv_kdev, struct device *mei_kdev, void *data) { - struct drm_i915_private *i915 = kdev_to_i915(i915_kdev); + struct intel_display *display = to_intel_display(drv_kdev); + struct drm_i915_private *i915 = to_i915(display->drm); drm_dbg(&i915->drm, "I915 HDCP comp unbind\n"); mutex_lock(&i915->display.hdcp.hdcp_mutex); diff --git a/drivers/gpu/drm/i915/display/intel_hdcp_gsc_message.c b/drivers/gpu/drm/i915/display/intel_hdcp_gsc_message.c index 6548e71b4c49..35bdb532bbb3 100644 --- a/drivers/gpu/drm/i915/display/intel_hdcp_gsc_message.c +++ b/drivers/gpu/drm/i915/display/intel_hdcp_gsc_message.c @@ -7,6 +7,7 @@ #include <drm/intel/i915_hdcp_interface.h> #include "i915_drv.h" +#include "intel_display_types.h" #include "intel_hdcp_gsc_message.h" int @@ -15,17 +16,19 @@ intel_hdcp_gsc_initiate_session(struct device *dev, struct hdcp_port_data *data, { struct wired_cmd_initiate_hdcp2_session_in session_init_in = {}; struct wired_cmd_initiate_hdcp2_session_out session_init_out = {}; + struct intel_display *display; struct drm_i915_private *i915; ssize_t byte; if (!dev || !data || !ake_data) return -EINVAL; - i915 = kdev_to_i915(dev); - if (!i915) { + display = to_intel_display(dev); + if (!display) { dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } + i915 = to_i915(display->drm); session_init_in.header.api_version = HDCP_API_VERSION; session_init_in.header.command_id = WIRED_INITIATE_HDCP2_SESSION; @@ -72,17 +75,19 @@ intel_hdcp_gsc_verify_receiver_cert_prepare_km(struct device *dev, { struct wired_cmd_verify_receiver_cert_in verify_rxcert_in = {}; struct wired_cmd_verify_receiver_cert_out verify_rxcert_out = {}; + struct intel_display *display; struct drm_i915_private *i915; ssize_t byte; if (!dev || !data || !rx_cert || !km_stored || !ek_pub_km || !msg_sz) return -EINVAL; - i915 = kdev_to_i915(dev); - if (!i915) { + display = to_intel_display(dev); + if (!display) { dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } + i915 = to_i915(display->drm); verify_rxcert_in.header.api_version = HDCP_API_VERSION; verify_rxcert_in.header.command_id = WIRED_VERIFY_RECEIVER_CERT; @@ -135,17 +140,19 @@ intel_hdcp_gsc_verify_hprime(struct device *dev, struct hdcp_port_data *data, { struct wired_cmd_ake_send_hprime_in send_hprime_in = {}; struct wired_cmd_ake_send_hprime_out send_hprime_out = {}; + struct intel_display *display; struct drm_i915_private *i915; ssize_t byte; if (!dev || !data || !rx_hprime) return -EINVAL; - i915 = kdev_to_i915(dev); - if (!i915) { + display = to_intel_display(dev); + if (!display) { dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } + i915 = to_i915(display->drm); send_hprime_in.header.api_version = HDCP_API_VERSION; send_hprime_in.header.command_id = WIRED_AKE_SEND_HPRIME; @@ -183,17 +190,19 @@ intel_hdcp_gsc_store_pairing_info(struct device *dev, struct hdcp_port_data *dat { struct wired_cmd_ake_send_pairing_info_in pairing_info_in = {}; struct wired_cmd_ake_send_pairing_info_out pairing_info_out = {}; + struct intel_display *display; struct drm_i915_private *i915; ssize_t byte; if (!dev || !data || !pairing_info) return -EINVAL; - i915 = kdev_to_i915(dev); - if (!i915) { + display = to_intel_display(dev); + if (!display) { dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } + i915 = to_i915(display->drm); pairing_info_in.header.api_version = HDCP_API_VERSION; pairing_info_in.header.command_id = WIRED_AKE_SEND_PAIRING_INFO; @@ -234,17 +243,19 @@ intel_hdcp_gsc_initiate_locality_check(struct device *dev, { struct wired_cmd_init_locality_check_in lc_init_in = {}; struct wired_cmd_init_locality_check_out lc_init_out = {}; + struct intel_display *display; struct drm_i915_private *i915; ssize_t byte; if (!dev || !data || !lc_init_data) return -EINVAL; - i915 = kdev_to_i915(dev); - if (!i915) { + display = to_intel_display(dev); + if (!display) { dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } + i915 = to_i915(display->drm); lc_init_in.header.api_version = HDCP_API_VERSION; lc_init_in.header.command_id = WIRED_INIT_LOCALITY_CHECK; @@ -280,17 +291,19 @@ intel_hdcp_gsc_verify_lprime(struct device *dev, struct hdcp_port_data *data, { struct wired_cmd_validate_locality_in verify_lprime_in = {}; struct wired_cmd_validate_locality_out verify_lprime_out = {}; + struct intel_display *display; struct drm_i915_private *i915; ssize_t byte; if (!dev || !data || !rx_lprime) return -EINVAL; - i915 = kdev_to_i915(dev); - if (!i915) { + display = to_intel_display(dev); + if (!display) { dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } + i915 = to_i915(display->drm); verify_lprime_in.header.api_version = HDCP_API_VERSION; verify_lprime_in.header.command_id = WIRED_VALIDATE_LOCALITY; @@ -330,17 +343,19 @@ int intel_hdcp_gsc_get_session_key(struct device *dev, { struct wired_cmd_get_session_key_in get_skey_in = {}; struct wired_cmd_get_session_key_out get_skey_out = {}; + struct intel_display *display; struct drm_i915_private *i915; ssize_t byte; if (!dev || !data || !ske_data) return -EINVAL; - i915 = kdev_to_i915(dev); - if (!i915) { + display = to_intel_display(dev); + if (!display) { dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } + i915 = to_i915(display->drm); get_skey_in.header.api_version = HDCP_API_VERSION; get_skey_in.header.command_id = WIRED_GET_SESSION_KEY; @@ -382,17 +397,19 @@ intel_hdcp_gsc_repeater_check_flow_prepare_ack(struct device *dev, { struct wired_cmd_verify_repeater_in verify_repeater_in = {}; struct wired_cmd_verify_repeater_out verify_repeater_out = {}; + struct intel_display *display; struct drm_i915_private *i915; ssize_t byte; if (!dev || !rep_topology || !rep_send_ack || !data) return -EINVAL; - i915 = kdev_to_i915(dev); - if (!i915) { + display = to_intel_display(dev); + if (!display) { dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } + i915 = to_i915(display->drm); verify_repeater_in.header.api_version = HDCP_API_VERSION; verify_repeater_in.header.command_id = WIRED_VERIFY_REPEATER; @@ -442,6 +459,7 @@ int intel_hdcp_gsc_verify_mprime(struct device *dev, { struct wired_cmd_repeater_auth_stream_req_in *verify_mprime_in; struct wired_cmd_repeater_auth_stream_req_out verify_mprime_out = {}; + struct intel_display *display; struct drm_i915_private *i915; ssize_t byte; size_t cmd_size; @@ -449,11 +467,12 @@ int intel_hdcp_gsc_verify_mprime(struct device *dev, if (!dev || !stream_ready || !data) return -EINVAL; - i915 = kdev_to_i915(dev); - if (!i915) { + display = to_intel_display(dev); + if (!display) { dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } + i915 = to_i915(display->drm); cmd_size = struct_size(verify_mprime_in, streams, data->k); if (cmd_size == SIZE_MAX) @@ -504,17 +523,19 @@ int intel_hdcp_gsc_enable_authentication(struct device *dev, { struct wired_cmd_enable_auth_in enable_auth_in = {}; struct wired_cmd_enable_auth_out enable_auth_out = {}; + struct intel_display *display; struct drm_i915_private *i915; ssize_t byte; if (!dev || !data) return -EINVAL; - i915 = kdev_to_i915(dev); - if (!i915) { + display = to_intel_display(dev); + if (!display) { dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } + i915 = to_i915(display->drm); enable_auth_in.header.api_version = HDCP_API_VERSION; enable_auth_in.header.command_id = WIRED_ENABLE_AUTH; @@ -549,17 +570,19 @@ intel_hdcp_gsc_close_session(struct device *dev, struct hdcp_port_data *data) { struct wired_cmd_close_session_in session_close_in = {}; struct wired_cmd_close_session_out session_close_out = {}; + struct intel_display *display; struct drm_i915_private *i915; ssize_t byte; if (!dev || !data) return -EINVAL; - i915 = kdev_to_i915(dev); - if (!i915) { + display = to_intel_display(dev); + if (!display) { dev_err(dev, "DRM not initialized, aborting HDCP.\n"); return -ENODEV; } + i915 = to_i915(display->drm); session_close_in.header.api_version = HDCP_API_VERSION; session_close_in.header.command_id = WIRED_CLOSE_SESSION; diff --git a/drivers/gpu/drm/i915/display/intel_hdmi.c b/drivers/gpu/drm/i915/display/intel_hdmi.c index 19498ee455fa..cd9ee171e0df 100644 --- a/drivers/gpu/drm/i915/display/intel_hdmi.c +++ b/drivers/gpu/drm/i915/display/intel_hdmi.c @@ -60,30 +60,25 @@ #include "intel_panel.h" #include "intel_snps_phy.h" -inline struct drm_i915_private *intel_hdmi_to_i915(struct intel_hdmi *intel_hdmi) -{ - return to_i915(hdmi_to_dig_port(intel_hdmi)->base.base.dev); -} - static void assert_hdmi_port_disabled(struct intel_hdmi *intel_hdmi) { - struct drm_i915_private *dev_priv = intel_hdmi_to_i915(intel_hdmi); + struct intel_display *display = to_intel_display(intel_hdmi); u32 enabled_bits; - enabled_bits = HAS_DDI(dev_priv) ? DDI_BUF_CTL_ENABLE : SDVO_ENABLE; + enabled_bits = HAS_DDI(display) ? DDI_BUF_CTL_ENABLE : SDVO_ENABLE; - drm_WARN(&dev_priv->drm, - intel_de_read(dev_priv, intel_hdmi->hdmi_reg) & enabled_bits, + drm_WARN(display->drm, + intel_de_read(display, intel_hdmi->hdmi_reg) & enabled_bits, "HDMI port enabled, expecting disabled\n"); } static void -assert_hdmi_transcoder_func_disabled(struct drm_i915_private *dev_priv, +assert_hdmi_transcoder_func_disabled(struct intel_display *display, enum transcoder cpu_transcoder) { - drm_WARN(&dev_priv->drm, - intel_de_read(dev_priv, TRANS_DDI_FUNC_CTL(dev_priv, cpu_transcoder)) & + drm_WARN(display->drm, + intel_de_read(display, TRANS_DDI_FUNC_CTL(display, cpu_transcoder)) & TRANS_DDI_FUNC_ENABLE, "HDMI transcoder function enabled, expecting disabled\n"); } @@ -158,35 +153,35 @@ static u32 hsw_infoframe_enable(unsigned int type) } static i915_reg_t -hsw_dip_data_reg(struct drm_i915_private *dev_priv, +hsw_dip_data_reg(struct intel_display *display, enum transcoder cpu_transcoder, unsigned int type, int i) { switch (type) { case HDMI_PACKET_TYPE_GAMUT_METADATA: - return HSW_TVIDEO_DIP_GMP_DATA(dev_priv, cpu_transcoder, i); + return HSW_TVIDEO_DIP_GMP_DATA(display, cpu_transcoder, i); case DP_SDP_VSC: - return HSW_TVIDEO_DIP_VSC_DATA(dev_priv, cpu_transcoder, i); + return HSW_TVIDEO_DIP_VSC_DATA(display, cpu_transcoder, i); case DP_SDP_ADAPTIVE_SYNC: - return ADL_TVIDEO_DIP_AS_SDP_DATA(dev_priv, cpu_transcoder, i); + return ADL_TVIDEO_DIP_AS_SDP_DATA(display, cpu_transcoder, i); case DP_SDP_PPS: - return ICL_VIDEO_DIP_PPS_DATA(dev_priv, cpu_transcoder, i); + return ICL_VIDEO_DIP_PPS_DATA(display, cpu_transcoder, i); case HDMI_INFOFRAME_TYPE_AVI: - return HSW_TVIDEO_DIP_AVI_DATA(dev_priv, cpu_transcoder, i); + return HSW_TVIDEO_DIP_AVI_DATA(display, cpu_transcoder, i); case HDMI_INFOFRAME_TYPE_SPD: - return HSW_TVIDEO_DIP_SPD_DATA(dev_priv, cpu_transcoder, i); + return HSW_TVIDEO_DIP_SPD_DATA(display, cpu_transcoder, i); case HDMI_INFOFRAME_TYPE_VENDOR: - return HSW_TVIDEO_DIP_VS_DATA(dev_priv, cpu_transcoder, i); + return HSW_TVIDEO_DIP_VS_DATA(display, cpu_transcoder, i); case HDMI_INFOFRAME_TYPE_DRM: - return GLK_TVIDEO_DIP_DRM_DATA(dev_priv, cpu_transcoder, i); + return GLK_TVIDEO_DIP_DRM_DATA(display, cpu_transcoder, i); default: MISSING_CASE(type); return INVALID_MMIO_REG; } } -static int hsw_dip_data_size(struct drm_i915_private *dev_priv, +static int hsw_dip_data_size(struct intel_display *display, unsigned int type) { switch (type) { @@ -197,7 +192,7 @@ static int hsw_dip_data_size(struct drm_i915_private *dev_priv, case DP_SDP_PPS: return VIDEO_DIP_PPS_DATA_SIZE; case HDMI_PACKET_TYPE_GAMUT_METADATA: - if (DISPLAY_VER(dev_priv) >= 11) + if (DISPLAY_VER(display) >= 11) return VIDEO_DIP_GMP_DATA_SIZE; else return VIDEO_DIP_DATA_SIZE; @@ -211,12 +206,12 @@ static void g4x_write_infoframe(struct intel_encoder *encoder, unsigned int type, const void *frame, ssize_t len) { + struct intel_display *display = to_intel_display(encoder); const u32 *data = frame; - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); - u32 val = intel_de_read(dev_priv, VIDEO_DIP_CTL); + u32 val = intel_de_read(display, VIDEO_DIP_CTL); int i; - drm_WARN(&dev_priv->drm, !(val & VIDEO_DIP_ENABLE), + drm_WARN(display->drm, !(val & VIDEO_DIP_ENABLE), "Writing DIP with CTL reg disabled\n"); val &= ~(VIDEO_DIP_SELECT_MASK | 0xf); /* clear DIP data offset */ @@ -224,22 +219,22 @@ static void g4x_write_infoframe(struct intel_encoder *encoder, val &= ~g4x_infoframe_enable(type); - intel_de_write(dev_priv, VIDEO_DIP_CTL, val); + intel_de_write(display, VIDEO_DIP_CTL, val); for (i = 0; i < len; i += 4) { - intel_de_write(dev_priv, VIDEO_DIP_DATA, *data); + intel_de_write(display, VIDEO_DIP_DATA, *data); data++; } /* Write every possible data byte to force correct ECC calculation. */ for (; i < VIDEO_DIP_DATA_SIZE; i += 4) - intel_de_write(dev_priv, VIDEO_DIP_DATA, 0); + intel_de_write(display, VIDEO_DIP_DATA, 0); val |= g4x_infoframe_enable(type); val &= ~VIDEO_DIP_FREQ_MASK; val |= VIDEO_DIP_FREQ_VSYNC; - intel_de_write(dev_priv, VIDEO_DIP_CTL, val); - intel_de_posting_read(dev_priv, VIDEO_DIP_CTL); + intel_de_write(display, VIDEO_DIP_CTL, val); + intel_de_posting_read(display, VIDEO_DIP_CTL); } static void g4x_read_infoframe(struct intel_encoder *encoder, @@ -247,22 +242,22 @@ static void g4x_read_infoframe(struct intel_encoder *encoder, unsigned int type, void *frame, ssize_t len) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); u32 *data = frame; int i; - intel_de_rmw(dev_priv, VIDEO_DIP_CTL, + intel_de_rmw(display, VIDEO_DIP_CTL, VIDEO_DIP_SELECT_MASK | 0xf, g4x_infoframe_index(type)); for (i = 0; i < len; i += 4) - *data++ = intel_de_read(dev_priv, VIDEO_DIP_DATA); + *data++ = intel_de_read(display, VIDEO_DIP_DATA); } static u32 g4x_infoframes_enabled(struct intel_encoder *encoder, const struct intel_crtc_state *pipe_config) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); - u32 val = intel_de_read(dev_priv, VIDEO_DIP_CTL); + struct intel_display *display = to_intel_display(encoder); + u32 val = intel_de_read(display, VIDEO_DIP_CTL); if ((val & VIDEO_DIP_ENABLE) == 0) return 0; @@ -279,14 +274,14 @@ static void ibx_write_infoframe(struct intel_encoder *encoder, unsigned int type, const void *frame, ssize_t len) { + struct intel_display *display = to_intel_display(encoder); const u32 *data = frame; - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); i915_reg_t reg = TVIDEO_DIP_CTL(crtc->pipe); - u32 val = intel_de_read(dev_priv, reg); + u32 val = intel_de_read(display, reg); int i; - drm_WARN(&dev_priv->drm, !(val & VIDEO_DIP_ENABLE), + drm_WARN(display->drm, !(val & VIDEO_DIP_ENABLE), "Writing DIP with CTL reg disabled\n"); val &= ~(VIDEO_DIP_SELECT_MASK | 0xf); /* clear DIP data offset */ @@ -294,23 +289,23 @@ static void ibx_write_infoframe(struct intel_encoder *encoder, val &= ~g4x_infoframe_enable(type); - intel_de_write(dev_priv, reg, val); + intel_de_write(display, reg, val); for (i = 0; i < len; i += 4) { - intel_de_write(dev_priv, TVIDEO_DIP_DATA(crtc->pipe), + intel_de_write(display, TVIDEO_DIP_DATA(crtc->pipe), *data); data++; } /* Write every possible data byte to force correct ECC calculation. */ for (; i < VIDEO_DIP_DATA_SIZE; i += 4) - intel_de_write(dev_priv, TVIDEO_DIP_DATA(crtc->pipe), 0); + intel_de_write(display, TVIDEO_DIP_DATA(crtc->pipe), 0); val |= g4x_infoframe_enable(type); val &= ~VIDEO_DIP_FREQ_MASK; val |= VIDEO_DIP_FREQ_VSYNC; - intel_de_write(dev_priv, reg, val); - intel_de_posting_read(dev_priv, reg); + intel_de_write(display, reg, val); + intel_de_posting_read(display, reg); } static void ibx_read_infoframe(struct intel_encoder *encoder, @@ -318,25 +313,25 @@ static void ibx_read_infoframe(struct intel_encoder *encoder, unsigned int type, void *frame, ssize_t len) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); u32 *data = frame; int i; - intel_de_rmw(dev_priv, TVIDEO_DIP_CTL(crtc->pipe), + intel_de_rmw(display, TVIDEO_DIP_CTL(crtc->pipe), VIDEO_DIP_SELECT_MASK | 0xf, g4x_infoframe_index(type)); for (i = 0; i < len; i += 4) - *data++ = intel_de_read(dev_priv, TVIDEO_DIP_DATA(crtc->pipe)); + *data++ = intel_de_read(display, TVIDEO_DIP_DATA(crtc->pipe)); } static u32 ibx_infoframes_enabled(struct intel_encoder *encoder, const struct intel_crtc_state *pipe_config) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); enum pipe pipe = to_intel_crtc(pipe_config->uapi.crtc)->pipe; i915_reg_t reg = TVIDEO_DIP_CTL(pipe); - u32 val = intel_de_read(dev_priv, reg); + u32 val = intel_de_read(display, reg); if ((val & VIDEO_DIP_ENABLE) == 0) return 0; @@ -354,14 +349,14 @@ static void cpt_write_infoframe(struct intel_encoder *encoder, unsigned int type, const void *frame, ssize_t len) { + struct intel_display *display = to_intel_display(encoder); const u32 *data = frame; - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); i915_reg_t reg = TVIDEO_DIP_CTL(crtc->pipe); - u32 val = intel_de_read(dev_priv, reg); + u32 val = intel_de_read(display, reg); int i; - drm_WARN(&dev_priv->drm, !(val & VIDEO_DIP_ENABLE), + drm_WARN(display->drm, !(val & VIDEO_DIP_ENABLE), "Writing DIP with CTL reg disabled\n"); val &= ~(VIDEO_DIP_SELECT_MASK | 0xf); /* clear DIP data offset */ @@ -372,23 +367,23 @@ static void cpt_write_infoframe(struct intel_encoder *encoder, if (type != HDMI_INFOFRAME_TYPE_AVI) val &= ~g4x_infoframe_enable(type); - intel_de_write(dev_priv, reg, val); + intel_de_write(display, reg, val); for (i = 0; i < len; i += 4) { - intel_de_write(dev_priv, TVIDEO_DIP_DATA(crtc->pipe), + intel_de_write(display, TVIDEO_DIP_DATA(crtc->pipe), *data); data++; } /* Write every possible data byte to force correct ECC calculation. */ for (; i < VIDEO_DIP_DATA_SIZE; i += 4) - intel_de_write(dev_priv, TVIDEO_DIP_DATA(crtc->pipe), 0); + intel_de_write(display, TVIDEO_DIP_DATA(crtc->pipe), 0); val |= g4x_infoframe_enable(type); val &= ~VIDEO_DIP_FREQ_MASK; val |= VIDEO_DIP_FREQ_VSYNC; - intel_de_write(dev_priv, reg, val); - intel_de_posting_read(dev_priv, reg); + intel_de_write(display, reg, val); + intel_de_posting_read(display, reg); } static void cpt_read_infoframe(struct intel_encoder *encoder, @@ -396,24 +391,24 @@ static void cpt_read_infoframe(struct intel_encoder *encoder, unsigned int type, void *frame, ssize_t len) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); u32 *data = frame; int i; - intel_de_rmw(dev_priv, TVIDEO_DIP_CTL(crtc->pipe), + intel_de_rmw(display, TVIDEO_DIP_CTL(crtc->pipe), VIDEO_DIP_SELECT_MASK | 0xf, g4x_infoframe_index(type)); for (i = 0; i < len; i += 4) - *data++ = intel_de_read(dev_priv, TVIDEO_DIP_DATA(crtc->pipe)); + *data++ = intel_de_read(display, TVIDEO_DIP_DATA(crtc->pipe)); } static u32 cpt_infoframes_enabled(struct intel_encoder *encoder, const struct intel_crtc_state *pipe_config) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); enum pipe pipe = to_intel_crtc(pipe_config->uapi.crtc)->pipe; - u32 val = intel_de_read(dev_priv, TVIDEO_DIP_CTL(pipe)); + u32 val = intel_de_read(display, TVIDEO_DIP_CTL(pipe)); if ((val & VIDEO_DIP_ENABLE) == 0) return 0; @@ -428,14 +423,14 @@ static void vlv_write_infoframe(struct intel_encoder *encoder, unsigned int type, const void *frame, ssize_t len) { + struct intel_display *display = to_intel_display(encoder); const u32 *data = frame; - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); i915_reg_t reg = VLV_TVIDEO_DIP_CTL(crtc->pipe); - u32 val = intel_de_read(dev_priv, reg); + u32 val = intel_de_read(display, reg); int i; - drm_WARN(&dev_priv->drm, !(val & VIDEO_DIP_ENABLE), + drm_WARN(display->drm, !(val & VIDEO_DIP_ENABLE), "Writing DIP with CTL reg disabled\n"); val &= ~(VIDEO_DIP_SELECT_MASK | 0xf); /* clear DIP data offset */ @@ -443,24 +438,24 @@ static void vlv_write_infoframe(struct intel_encoder *encoder, val &= ~g4x_infoframe_enable(type); - intel_de_write(dev_priv, reg, val); + intel_de_write(display, reg, val); for (i = 0; i < len; i += 4) { - intel_de_write(dev_priv, + intel_de_write(display, VLV_TVIDEO_DIP_DATA(crtc->pipe), *data); data++; } /* Write every possible data byte to force correct ECC calculation. */ for (; i < VIDEO_DIP_DATA_SIZE; i += 4) - intel_de_write(dev_priv, + intel_de_write(display, VLV_TVIDEO_DIP_DATA(crtc->pipe), 0); val |= g4x_infoframe_enable(type); val &= ~VIDEO_DIP_FREQ_MASK; val |= VIDEO_DIP_FREQ_VSYNC; - intel_de_write(dev_priv, reg, val); - intel_de_posting_read(dev_priv, reg); + intel_de_write(display, reg, val); + intel_de_posting_read(display, reg); } static void vlv_read_infoframe(struct intel_encoder *encoder, @@ -468,25 +463,25 @@ static void vlv_read_infoframe(struct intel_encoder *encoder, unsigned int type, void *frame, ssize_t len) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); u32 *data = frame; int i; - intel_de_rmw(dev_priv, VLV_TVIDEO_DIP_CTL(crtc->pipe), + intel_de_rmw(display, VLV_TVIDEO_DIP_CTL(crtc->pipe), VIDEO_DIP_SELECT_MASK | 0xf, g4x_infoframe_index(type)); for (i = 0; i < len; i += 4) - *data++ = intel_de_read(dev_priv, + *data++ = intel_de_read(display, VLV_TVIDEO_DIP_DATA(crtc->pipe)); } static u32 vlv_infoframes_enabled(struct intel_encoder *encoder, const struct intel_crtc_state *pipe_config) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); enum pipe pipe = to_intel_crtc(pipe_config->uapi.crtc)->pipe; - u32 val = intel_de_read(dev_priv, VLV_TVIDEO_DIP_CTL(pipe)); + u32 val = intel_de_read(display, VLV_TVIDEO_DIP_CTL(pipe)); if ((val & VIDEO_DIP_ENABLE) == 0) return 0; @@ -504,75 +499,75 @@ void hsw_write_infoframe(struct intel_encoder *encoder, unsigned int type, const void *frame, ssize_t len) { + struct intel_display *display = to_intel_display(encoder); const u32 *data = frame; - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); enum transcoder cpu_transcoder = crtc_state->cpu_transcoder; - i915_reg_t ctl_reg = HSW_TVIDEO_DIP_CTL(dev_priv, cpu_transcoder); + i915_reg_t ctl_reg = HSW_TVIDEO_DIP_CTL(display, cpu_transcoder); int data_size; int i; - u32 val = intel_de_read(dev_priv, ctl_reg); + u32 val = intel_de_read(display, ctl_reg); - data_size = hsw_dip_data_size(dev_priv, type); + data_size = hsw_dip_data_size(display, type); - drm_WARN_ON(&dev_priv->drm, len > data_size); + drm_WARN_ON(display->drm, len > data_size); val &= ~hsw_infoframe_enable(type); - intel_de_write(dev_priv, ctl_reg, val); + intel_de_write(display, ctl_reg, val); for (i = 0; i < len; i += 4) { - intel_de_write(dev_priv, - hsw_dip_data_reg(dev_priv, cpu_transcoder, type, i >> 2), + intel_de_write(display, + hsw_dip_data_reg(display, cpu_transcoder, type, i >> 2), *data); data++; } /* Write every possible data byte to force correct ECC calculation. */ for (; i < data_size; i += 4) - intel_de_write(dev_priv, - hsw_dip_data_reg(dev_priv, cpu_transcoder, type, i >> 2), + intel_de_write(display, + hsw_dip_data_reg(display, cpu_transcoder, type, i >> 2), 0); /* Wa_14013475917 */ - if (!(IS_DISPLAY_VER(dev_priv, 13, 14) && crtc_state->has_psr && + if (!(IS_DISPLAY_VER(display, 13, 14) && crtc_state->has_psr && !crtc_state->has_panel_replay && type == DP_SDP_VSC)) val |= hsw_infoframe_enable(type); if (type == DP_SDP_VSC) val |= VSC_DIP_HW_DATA_SW_HEA; - intel_de_write(dev_priv, ctl_reg, val); - intel_de_posting_read(dev_priv, ctl_reg); + intel_de_write(display, ctl_reg, val); + intel_de_posting_read(display, ctl_reg); } void hsw_read_infoframe(struct intel_encoder *encoder, const struct intel_crtc_state *crtc_state, unsigned int type, void *frame, ssize_t len) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); enum transcoder cpu_transcoder = crtc_state->cpu_transcoder; u32 *data = frame; int i; for (i = 0; i < len; i += 4) - *data++ = intel_de_read(dev_priv, - hsw_dip_data_reg(dev_priv, cpu_transcoder, type, i >> 2)); + *data++ = intel_de_read(display, + hsw_dip_data_reg(display, cpu_transcoder, type, i >> 2)); } static u32 hsw_infoframes_enabled(struct intel_encoder *encoder, const struct intel_crtc_state *pipe_config) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); - u32 val = intel_de_read(dev_priv, - HSW_TVIDEO_DIP_CTL(dev_priv, pipe_config->cpu_transcoder)); + struct intel_display *display = to_intel_display(encoder); + u32 val = intel_de_read(display, + HSW_TVIDEO_DIP_CTL(display, pipe_config->cpu_transcoder)); u32 mask; mask = (VIDEO_DIP_ENABLE_VSC_HSW | VIDEO_DIP_ENABLE_AVI_HSW | VIDEO_DIP_ENABLE_GCP_HSW | VIDEO_DIP_ENABLE_VS_HSW | VIDEO_DIP_ENABLE_GMP_HSW | VIDEO_DIP_ENABLE_SPD_HSW); - if (DISPLAY_VER(dev_priv) >= 10) + if (DISPLAY_VER(display) >= 10) mask |= VIDEO_DIP_ENABLE_DRM_GLK; - if (HAS_AS_SDP(dev_priv)) + if (HAS_AS_SDP(display)) mask |= VIDEO_DIP_ENABLE_AS_ADL; return val & mask; @@ -604,7 +599,7 @@ u32 intel_hdmi_infoframe_enable(unsigned int type) u32 intel_hdmi_infoframes_enabled(struct intel_encoder *encoder, const struct intel_crtc_state *crtc_state) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); struct intel_digital_port *dig_port = enc_to_dig_port(encoder); u32 val, ret = 0; int i; @@ -615,7 +610,7 @@ u32 intel_hdmi_infoframes_enabled(struct intel_encoder *encoder, for (i = 0; i < ARRAY_SIZE(infoframe_type_to_idx); i++) { unsigned int type = infoframe_type_to_idx[i]; - if (HAS_DDI(dev_priv)) { + if (HAS_DDI(display)) { if (val & hsw_infoframe_enable(type)) ret |= BIT(i); } else { @@ -830,11 +825,11 @@ intel_hdmi_compute_drm_infoframe(struct intel_encoder *encoder, struct intel_crtc_state *crtc_state, struct drm_connector_state *conn_state) { + struct intel_display *display = to_intel_display(encoder); struct hdmi_drm_infoframe *frame = &crtc_state->infoframes.drm.drm; - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); int ret; - if (DISPLAY_VER(dev_priv) < 10) + if (DISPLAY_VER(display) < 10) return true; if (!crtc_state->has_infoframe) @@ -848,13 +843,13 @@ intel_hdmi_compute_drm_infoframe(struct intel_encoder *encoder, ret = drm_hdmi_infoframe_set_hdr_metadata(frame, conn_state); if (ret < 0) { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "couldn't set HDR metadata in infoframe\n"); return false; } ret = hdmi_drm_infoframe_check(frame); - if (drm_WARN_ON(&dev_priv->drm, ret)) + if (drm_WARN_ON(display->drm, ret)) return false; return true; @@ -865,11 +860,11 @@ static void g4x_set_infoframes(struct intel_encoder *encoder, const struct intel_crtc_state *crtc_state, const struct drm_connector_state *conn_state) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); struct intel_digital_port *dig_port = enc_to_dig_port(encoder); struct intel_hdmi *intel_hdmi = &dig_port->hdmi; i915_reg_t reg = VIDEO_DIP_CTL; - u32 val = intel_de_read(dev_priv, reg); + u32 val = intel_de_read(display, reg); u32 port = VIDEO_DIP_PORT(encoder->port); assert_hdmi_port_disabled(intel_hdmi); @@ -889,21 +884,21 @@ static void g4x_set_infoframes(struct intel_encoder *encoder, if (!(val & VIDEO_DIP_ENABLE)) return; if (port != (val & VIDEO_DIP_PORT_MASK)) { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "video DIP still enabled on port %c\n", (val & VIDEO_DIP_PORT_MASK) >> 29); return; } val &= ~(VIDEO_DIP_ENABLE | VIDEO_DIP_ENABLE_AVI | VIDEO_DIP_ENABLE_VENDOR | VIDEO_DIP_ENABLE_SPD); - intel_de_write(dev_priv, reg, val); - intel_de_posting_read(dev_priv, reg); + intel_de_write(display, reg, val); + intel_de_posting_read(display, reg); return; } if (port != (val & VIDEO_DIP_PORT_MASK)) { if (val & VIDEO_DIP_ENABLE) { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "video DIP already enabled on port %c\n", (val & VIDEO_DIP_PORT_MASK) >> 29); return; @@ -916,8 +911,8 @@ static void g4x_set_infoframes(struct intel_encoder *encoder, val &= ~(VIDEO_DIP_ENABLE_AVI | VIDEO_DIP_ENABLE_VENDOR | VIDEO_DIP_ENABLE_SPD); - intel_de_write(dev_priv, reg, val); - intel_de_posting_read(dev_priv, reg); + intel_de_write(display, reg, val); + intel_de_posting_read(display, reg); intel_write_infoframe(encoder, crtc_state, HDMI_INFOFRAME_TYPE_AVI, @@ -977,6 +972,7 @@ static bool intel_hdmi_set_gcp_infoframe(struct intel_encoder *encoder, const struct intel_crtc_state *crtc_state, const struct drm_connector_state *conn_state) { + struct intel_display *display = to_intel_display(encoder); struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); i915_reg_t reg; @@ -985,8 +981,8 @@ static bool intel_hdmi_set_gcp_infoframe(struct intel_encoder *encoder, intel_hdmi_infoframe_enable(HDMI_PACKET_TYPE_GENERAL_CONTROL)) == 0) return false; - if (HAS_DDI(dev_priv)) - reg = HSW_TVIDEO_DIP_GCP(dev_priv, crtc_state->cpu_transcoder); + if (HAS_DDI(display)) + reg = HSW_TVIDEO_DIP_GCP(display, crtc_state->cpu_transcoder); else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) reg = VLV_TVIDEO_DIP_GCP(crtc->pipe); else if (HAS_PCH_SPLIT(dev_priv)) @@ -994,7 +990,7 @@ static bool intel_hdmi_set_gcp_infoframe(struct intel_encoder *encoder, else return false; - intel_de_write(dev_priv, reg, crtc_state->infoframes.gcp); + intel_de_write(display, reg, crtc_state->infoframes.gcp); return true; } @@ -1002,6 +998,7 @@ static bool intel_hdmi_set_gcp_infoframe(struct intel_encoder *encoder, void intel_hdmi_read_gcp_infoframe(struct intel_encoder *encoder, struct intel_crtc_state *crtc_state) { + struct intel_display *display = to_intel_display(encoder); struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); i915_reg_t reg; @@ -1010,8 +1007,8 @@ void intel_hdmi_read_gcp_infoframe(struct intel_encoder *encoder, intel_hdmi_infoframe_enable(HDMI_PACKET_TYPE_GENERAL_CONTROL)) == 0) return; - if (HAS_DDI(dev_priv)) - reg = HSW_TVIDEO_DIP_GCP(dev_priv, crtc_state->cpu_transcoder); + if (HAS_DDI(display)) + reg = HSW_TVIDEO_DIP_GCP(display, crtc_state->cpu_transcoder); else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) reg = VLV_TVIDEO_DIP_GCP(crtc->pipe); else if (HAS_PCH_SPLIT(dev_priv)) @@ -1019,7 +1016,7 @@ void intel_hdmi_read_gcp_infoframe(struct intel_encoder *encoder, else return; - crtc_state->infoframes.gcp = intel_de_read(dev_priv, reg); + crtc_state->infoframes.gcp = intel_de_read(display, reg); } static void intel_hdmi_compute_gcp_infoframe(struct intel_encoder *encoder, @@ -1049,12 +1046,12 @@ static void ibx_set_infoframes(struct intel_encoder *encoder, const struct intel_crtc_state *crtc_state, const struct drm_connector_state *conn_state) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); struct intel_digital_port *dig_port = enc_to_dig_port(encoder); struct intel_hdmi *intel_hdmi = &dig_port->hdmi; i915_reg_t reg = TVIDEO_DIP_CTL(crtc->pipe); - u32 val = intel_de_read(dev_priv, reg); + u32 val = intel_de_read(display, reg); u32 port = VIDEO_DIP_PORT(encoder->port); assert_hdmi_port_disabled(intel_hdmi); @@ -1068,13 +1065,13 @@ static void ibx_set_infoframes(struct intel_encoder *encoder, val &= ~(VIDEO_DIP_ENABLE | VIDEO_DIP_ENABLE_AVI | VIDEO_DIP_ENABLE_VENDOR | VIDEO_DIP_ENABLE_GAMUT | VIDEO_DIP_ENABLE_SPD | VIDEO_DIP_ENABLE_GCP); - intel_de_write(dev_priv, reg, val); - intel_de_posting_read(dev_priv, reg); + intel_de_write(display, reg, val); + intel_de_posting_read(display, reg); return; } if (port != (val & VIDEO_DIP_PORT_MASK)) { - drm_WARN(&dev_priv->drm, val & VIDEO_DIP_ENABLE, + drm_WARN(display->drm, val & VIDEO_DIP_ENABLE, "DIP already enabled on port %c\n", (val & VIDEO_DIP_PORT_MASK) >> 29); val &= ~VIDEO_DIP_PORT_MASK; @@ -1089,8 +1086,8 @@ static void ibx_set_infoframes(struct intel_encoder *encoder, if (intel_hdmi_set_gcp_infoframe(encoder, crtc_state, conn_state)) val |= VIDEO_DIP_ENABLE_GCP; - intel_de_write(dev_priv, reg, val); - intel_de_posting_read(dev_priv, reg); + intel_de_write(display, reg, val); + intel_de_posting_read(display, reg); intel_write_infoframe(encoder, crtc_state, HDMI_INFOFRAME_TYPE_AVI, @@ -1108,11 +1105,11 @@ static void cpt_set_infoframes(struct intel_encoder *encoder, const struct intel_crtc_state *crtc_state, const struct drm_connector_state *conn_state) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder); i915_reg_t reg = TVIDEO_DIP_CTL(crtc->pipe); - u32 val = intel_de_read(dev_priv, reg); + u32 val = intel_de_read(display, reg); assert_hdmi_port_disabled(intel_hdmi); @@ -1125,8 +1122,8 @@ static void cpt_set_infoframes(struct intel_encoder *encoder, val &= ~(VIDEO_DIP_ENABLE | VIDEO_DIP_ENABLE_AVI | VIDEO_DIP_ENABLE_VENDOR | VIDEO_DIP_ENABLE_GAMUT | VIDEO_DIP_ENABLE_SPD | VIDEO_DIP_ENABLE_GCP); - intel_de_write(dev_priv, reg, val); - intel_de_posting_read(dev_priv, reg); + intel_de_write(display, reg, val); + intel_de_posting_read(display, reg); return; } @@ -1138,8 +1135,8 @@ static void cpt_set_infoframes(struct intel_encoder *encoder, if (intel_hdmi_set_gcp_infoframe(encoder, crtc_state, conn_state)) val |= VIDEO_DIP_ENABLE_GCP; - intel_de_write(dev_priv, reg, val); - intel_de_posting_read(dev_priv, reg); + intel_de_write(display, reg, val); + intel_de_posting_read(display, reg); intel_write_infoframe(encoder, crtc_state, HDMI_INFOFRAME_TYPE_AVI, @@ -1157,11 +1154,11 @@ static void vlv_set_infoframes(struct intel_encoder *encoder, const struct intel_crtc_state *crtc_state, const struct drm_connector_state *conn_state) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder); i915_reg_t reg = VLV_TVIDEO_DIP_CTL(crtc->pipe); - u32 val = intel_de_read(dev_priv, reg); + u32 val = intel_de_read(display, reg); u32 port = VIDEO_DIP_PORT(encoder->port); assert_hdmi_port_disabled(intel_hdmi); @@ -1175,13 +1172,13 @@ static void vlv_set_infoframes(struct intel_encoder *encoder, val &= ~(VIDEO_DIP_ENABLE | VIDEO_DIP_ENABLE_AVI | VIDEO_DIP_ENABLE_VENDOR | VIDEO_DIP_ENABLE_GAMUT | VIDEO_DIP_ENABLE_SPD | VIDEO_DIP_ENABLE_GCP); - intel_de_write(dev_priv, reg, val); - intel_de_posting_read(dev_priv, reg); + intel_de_write(display, reg, val); + intel_de_posting_read(display, reg); return; } if (port != (val & VIDEO_DIP_PORT_MASK)) { - drm_WARN(&dev_priv->drm, val & VIDEO_DIP_ENABLE, + drm_WARN(display->drm, val & VIDEO_DIP_ENABLE, "DIP already enabled on port %c\n", (val & VIDEO_DIP_PORT_MASK) >> 29); val &= ~VIDEO_DIP_PORT_MASK; @@ -1196,8 +1193,8 @@ static void vlv_set_infoframes(struct intel_encoder *encoder, if (intel_hdmi_set_gcp_infoframe(encoder, crtc_state, conn_state)) val |= VIDEO_DIP_ENABLE_GCP; - intel_de_write(dev_priv, reg, val); - intel_de_posting_read(dev_priv, reg); + intel_de_write(display, reg, val); + intel_de_posting_read(display, reg); intel_write_infoframe(encoder, crtc_state, HDMI_INFOFRAME_TYPE_AVI, @@ -1215,12 +1212,12 @@ static void hsw_set_infoframes(struct intel_encoder *encoder, const struct intel_crtc_state *crtc_state, const struct drm_connector_state *conn_state) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); - i915_reg_t reg = HSW_TVIDEO_DIP_CTL(dev_priv, + struct intel_display *display = to_intel_display(encoder); + i915_reg_t reg = HSW_TVIDEO_DIP_CTL(display, crtc_state->cpu_transcoder); - u32 val = intel_de_read(dev_priv, reg); + u32 val = intel_de_read(display, reg); - assert_hdmi_transcoder_func_disabled(dev_priv, + assert_hdmi_transcoder_func_disabled(display, crtc_state->cpu_transcoder); val &= ~(VIDEO_DIP_ENABLE_VSC_HSW | VIDEO_DIP_ENABLE_AVI_HSW | @@ -1229,16 +1226,16 @@ static void hsw_set_infoframes(struct intel_encoder *encoder, VIDEO_DIP_ENABLE_DRM_GLK | VIDEO_DIP_ENABLE_AS_ADL); if (!enable) { - intel_de_write(dev_priv, reg, val); - intel_de_posting_read(dev_priv, reg); + intel_de_write(display, reg, val); + intel_de_posting_read(display, reg); return; } if (intel_hdmi_set_gcp_infoframe(encoder, crtc_state, conn_state)) val |= VIDEO_DIP_ENABLE_GCP_HSW; - intel_de_write(dev_priv, reg, val); - intel_de_posting_read(dev_priv, reg); + intel_de_write(display, reg, val); + intel_de_posting_read(display, reg); intel_write_infoframe(encoder, crtc_state, HDMI_INFOFRAME_TYPE_AVI, @@ -1256,16 +1253,16 @@ static void hsw_set_infoframes(struct intel_encoder *encoder, void intel_dp_dual_mode_set_tmds_output(struct intel_hdmi *hdmi, bool enable) { - struct drm_i915_private *dev_priv = intel_hdmi_to_i915(hdmi); + struct intel_display *display = to_intel_display(hdmi); struct i2c_adapter *ddc = hdmi->attached_connector->base.ddc; if (hdmi->dp_dual_mode.type < DRM_DP_DUAL_MODE_TYPE2_DVI) return; - drm_dbg_kms(&dev_priv->drm, "%s DP dual mode adaptor TMDS output\n", + drm_dbg_kms(display->drm, "%s DP dual mode adaptor TMDS output\n", enable ? "Enabling" : "Disabling"); - drm_dp_dual_mode_set_tmds_output(&dev_priv->drm, + drm_dp_dual_mode_set_tmds_output(display->drm, hdmi->dp_dual_mode.type, ddc, enable); } @@ -1331,7 +1328,7 @@ static int intel_hdmi_hdcp_write_an_aksv(struct intel_digital_port *dig_port, u8 *an) { - struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev); + struct intel_display *display = to_intel_display(dig_port); struct intel_hdmi *hdmi = &dig_port->hdmi; struct i2c_adapter *ddc = hdmi->attached_connector->base.ddc; int ret; @@ -1339,14 +1336,14 @@ int intel_hdmi_hdcp_write_an_aksv(struct intel_digital_port *dig_port, ret = intel_hdmi_hdcp_write(dig_port, DRM_HDCP_DDC_AN, an, DRM_HDCP_AN_LEN); if (ret) { - drm_dbg_kms(&i915->drm, "Write An over DDC failed (%d)\n", + drm_dbg_kms(display->drm, "Write An over DDC failed (%d)\n", ret); return ret; } ret = intel_gmbus_output_aksv(ddc); if (ret < 0) { - drm_dbg_kms(&i915->drm, "Failed to output aksv (%d)\n", ret); + drm_dbg_kms(display->drm, "Failed to output aksv (%d)\n", ret); return ret; } return 0; @@ -1355,13 +1352,13 @@ int intel_hdmi_hdcp_write_an_aksv(struct intel_digital_port *dig_port, static int intel_hdmi_hdcp_read_bksv(struct intel_digital_port *dig_port, u8 *bksv) { - struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev); + struct intel_display *display = to_intel_display(dig_port); int ret; ret = intel_hdmi_hdcp_read(dig_port, DRM_HDCP_DDC_BKSV, bksv, DRM_HDCP_KSV_LEN); if (ret) - drm_dbg_kms(&i915->drm, "Read Bksv over DDC failed (%d)\n", + drm_dbg_kms(display->drm, "Read Bksv over DDC failed (%d)\n", ret); return ret; } @@ -1370,13 +1367,14 @@ static int intel_hdmi_hdcp_read_bstatus(struct intel_digital_port *dig_port, u8 *bstatus) { - struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev); + struct intel_display *display = to_intel_display(dig_port); int ret; ret = intel_hdmi_hdcp_read(dig_port, DRM_HDCP_DDC_BSTATUS, bstatus, DRM_HDCP_BSTATUS_LEN); if (ret) - drm_dbg_kms(&i915->drm, "Read bstatus over DDC failed (%d)\n", + drm_dbg_kms(display->drm, + "Read bstatus over DDC failed (%d)\n", ret); return ret; } @@ -1385,13 +1383,13 @@ static int intel_hdmi_hdcp_repeater_present(struct intel_digital_port *dig_port, bool *repeater_present) { - struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev); + struct intel_display *display = to_intel_display(dig_port); int ret; u8 val; ret = intel_hdmi_hdcp_read(dig_port, DRM_HDCP_DDC_BCAPS, &val, 1); if (ret) { - drm_dbg_kms(&i915->drm, "Read bcaps over DDC failed (%d)\n", + drm_dbg_kms(display->drm, "Read bcaps over DDC failed (%d)\n", ret); return ret; } @@ -1403,13 +1401,13 @@ static int intel_hdmi_hdcp_read_ri_prime(struct intel_digital_port *dig_port, u8 *ri_prime) { - struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev); + struct intel_display *display = to_intel_display(dig_port); int ret; ret = intel_hdmi_hdcp_read(dig_port, DRM_HDCP_DDC_RI_PRIME, ri_prime, DRM_HDCP_RI_LEN); if (ret) - drm_dbg_kms(&i915->drm, "Read Ri' over DDC failed (%d)\n", + drm_dbg_kms(display->drm, "Read Ri' over DDC failed (%d)\n", ret); return ret; } @@ -1418,13 +1416,13 @@ static int intel_hdmi_hdcp_read_ksv_ready(struct intel_digital_port *dig_port, bool *ksv_ready) { - struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev); + struct intel_display *display = to_intel_display(dig_port); int ret; u8 val; ret = intel_hdmi_hdcp_read(dig_port, DRM_HDCP_DDC_BCAPS, &val, 1); if (ret) { - drm_dbg_kms(&i915->drm, "Read bcaps over DDC failed (%d)\n", + drm_dbg_kms(display->drm, "Read bcaps over DDC failed (%d)\n", ret); return ret; } @@ -1436,12 +1434,12 @@ static int intel_hdmi_hdcp_read_ksv_fifo(struct intel_digital_port *dig_port, int num_downstream, u8 *ksv_fifo) { - struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev); + struct intel_display *display = to_intel_display(dig_port); int ret; ret = intel_hdmi_hdcp_read(dig_port, DRM_HDCP_DDC_KSV_FIFO, ksv_fifo, num_downstream * DRM_HDCP_KSV_LEN); if (ret) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "Read ksv fifo over DDC failed (%d)\n", ret); return ret; } @@ -1452,7 +1450,7 @@ static int intel_hdmi_hdcp_read_v_prime_part(struct intel_digital_port *dig_port, int i, u32 *part) { - struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev); + struct intel_display *display = to_intel_display(dig_port); int ret; if (i >= DRM_HDCP_V_PRIME_NUM_PARTS) @@ -1461,7 +1459,8 @@ int intel_hdmi_hdcp_read_v_prime_part(struct intel_digital_port *dig_port, ret = intel_hdmi_hdcp_read(dig_port, DRM_HDCP_DDC_V_PRIME(i), part, DRM_HDCP_V_PRIME_PART_LEN); if (ret) - drm_dbg_kms(&i915->drm, "Read V'[%d] over DDC failed (%d)\n", + drm_dbg_kms(display->drm, + "Read V'[%d] over DDC failed (%d)\n", i, ret); return ret; } @@ -1469,15 +1468,15 @@ int intel_hdmi_hdcp_read_v_prime_part(struct intel_digital_port *dig_port, static int kbl_repositioning_enc_en_signal(struct intel_connector *connector, enum transcoder cpu_transcoder) { - struct drm_i915_private *dev_priv = to_i915(connector->base.dev); + struct intel_display *display = to_intel_display(connector); struct intel_digital_port *dig_port = intel_attached_dig_port(connector); struct intel_crtc *crtc = to_intel_crtc(connector->base.state->crtc); u32 scanline; int ret; for (;;) { - scanline = intel_de_read(dev_priv, - PIPEDSL(dev_priv, crtc->pipe)); + scanline = intel_de_read(display, + PIPEDSL(display, crtc->pipe)); if (scanline > 100 && scanline < 200) break; usleep_range(25, 50); @@ -1486,7 +1485,7 @@ static int kbl_repositioning_enc_en_signal(struct intel_connector *connector, ret = intel_ddi_toggle_hdcp_bits(&dig_port->base, cpu_transcoder, false, TRANS_DDI_HDCP_SIGNALLING); if (ret) { - drm_err(&dev_priv->drm, + drm_err(display->drm, "Disable HDCP signalling failed (%d)\n", ret); return ret; } @@ -1494,7 +1493,7 @@ static int kbl_repositioning_enc_en_signal(struct intel_connector *connector, ret = intel_ddi_toggle_hdcp_bits(&dig_port->base, cpu_transcoder, true, TRANS_DDI_HDCP_SIGNALLING); if (ret) { - drm_err(&dev_priv->drm, + drm_err(display->drm, "Enable HDCP signalling failed (%d)\n", ret); return ret; } @@ -1507,6 +1506,7 @@ int intel_hdmi_hdcp_toggle_signalling(struct intel_digital_port *dig_port, enum transcoder cpu_transcoder, bool enable) { + struct intel_display *display = to_intel_display(dig_port); struct intel_hdmi *hdmi = &dig_port->hdmi; struct intel_connector *connector = hdmi->attached_connector; struct drm_i915_private *dev_priv = to_i915(connector->base.dev); @@ -1519,7 +1519,7 @@ int intel_hdmi_hdcp_toggle_signalling(struct intel_digital_port *dig_port, cpu_transcoder, enable, TRANS_DDI_HDCP_SIGNALLING); if (ret) { - drm_err(&dev_priv->drm, "%s HDCP signalling failed (%d)\n", + drm_err(display->drm, "%s HDCP signalling failed (%d)\n", enable ? "Enable" : "Disable", ret); return ret; } @@ -1539,6 +1539,7 @@ static bool intel_hdmi_hdcp_check_link_once(struct intel_digital_port *dig_port, struct intel_connector *connector) { + struct intel_display *display = to_intel_display(dig_port); struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev); enum port port = dig_port->base.port; enum transcoder cpu_transcoder = connector->hdcp.cpu_transcoder; @@ -1558,9 +1559,9 @@ bool intel_hdmi_hdcp_check_link_once(struct intel_digital_port *dig_port, if (wait_for((intel_de_read(i915, HDCP_STATUS(i915, cpu_transcoder, port)) & (HDCP_STATUS_RI_MATCH | HDCP_STATUS_ENC)) == (HDCP_STATUS_RI_MATCH | HDCP_STATUS_ENC), 1)) { - drm_dbg_kms(&i915->drm, "Ri' mismatch detected (%x)\n", - intel_de_read(i915, HDCP_STATUS(i915, cpu_transcoder, - port))); + drm_dbg_kms(display->drm, "Ri' mismatch detected (%x)\n", + intel_de_read(i915, HDCP_STATUS(i915, cpu_transcoder, + port))); return false; } return true; @@ -1570,14 +1571,14 @@ static bool intel_hdmi_hdcp_check_link(struct intel_digital_port *dig_port, struct intel_connector *connector) { - struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev); + struct intel_display *display = to_intel_display(dig_port); int retry; for (retry = 0; retry < 3; retry++) if (intel_hdmi_hdcp_check_link_once(dig_port, connector)) return true; - drm_err(&i915->drm, "Link check failed\n"); + drm_err(display->drm, "Link check failed\n"); return false; } @@ -1628,13 +1629,13 @@ hdcp2_detect_msg_availability(struct intel_digital_port *dig_port, u8 msg_id, bool *msg_ready, ssize_t *msg_sz) { - struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev); + struct intel_display *display = to_intel_display(dig_port); u8 rx_status[HDCP_2_2_HDMI_RXSTATUS_LEN]; int ret; ret = intel_hdmi_hdcp2_read_rx_status(dig_port, rx_status); if (ret < 0) { - drm_dbg_kms(&i915->drm, "rx_status read failed. Err %d\n", + drm_dbg_kms(display->drm, "rx_status read failed. Err %d\n", ret); return ret; } @@ -1655,7 +1656,7 @@ static ssize_t intel_hdmi_hdcp2_wait_for_msg(struct intel_digital_port *dig_port, u8 msg_id, bool paired) { - struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev); + struct intel_display *display = to_intel_display(dig_port); bool msg_ready = false; int timeout, ret; ssize_t msg_sz = 0; @@ -1670,7 +1671,8 @@ intel_hdmi_hdcp2_wait_for_msg(struct intel_digital_port *dig_port, !ret && msg_ready && msg_sz, timeout * 1000, 1000, 5 * 1000); if (ret) - drm_dbg_kms(&i915->drm, "msg_id: %d, ret: %d, timeout: %d\n", + drm_dbg_kms(display->drm, + "msg_id: %d, ret: %d, timeout: %d\n", msg_id, ret, timeout); return ret ? ret : msg_sz; @@ -1691,8 +1693,8 @@ static int intel_hdmi_hdcp2_read_msg(struct intel_connector *connector, u8 msg_id, void *buf, size_t size) { + struct intel_display *display = to_intel_display(connector); struct intel_digital_port *dig_port = intel_attached_dig_port(connector); - struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev); struct intel_hdmi *hdmi = &dig_port->hdmi; struct intel_hdcp *hdcp = &hdmi->attached_connector->hdcp; unsigned int offset; @@ -1708,7 +1710,7 @@ int intel_hdmi_hdcp2_read_msg(struct intel_connector *connector, * available buffer. */ if (ret > size) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "msg_sz(%zd) is more than exp size(%zu)\n", ret, size); return -EINVAL; @@ -1717,7 +1719,7 @@ int intel_hdmi_hdcp2_read_msg(struct intel_connector *connector, offset = HDCP_2_2_HDMI_REG_RD_MSG_OFFSET; ret = intel_hdmi_hdcp_read(dig_port, offset, buf, ret); if (ret) - drm_dbg_kms(&i915->drm, "Failed to read msg_id: %d(%zd)\n", + drm_dbg_kms(display->drm, "Failed to read msg_id: %d(%zd)\n", msg_id, ret); return ret; @@ -1783,16 +1785,17 @@ static const struct intel_hdcp_shim intel_hdmi_hdcp_shim = { static int intel_hdmi_source_max_tmds_clock(struct intel_encoder *encoder) { + struct intel_display *display = to_intel_display(encoder); struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); int max_tmds_clock, vbt_max_tmds_clock; - if (DISPLAY_VER(dev_priv) >= 13 || IS_ALDERLAKE_S(dev_priv)) + if (DISPLAY_VER(display) >= 13 || IS_ALDERLAKE_S(dev_priv)) max_tmds_clock = 600000; - else if (DISPLAY_VER(dev_priv) >= 10) + else if (DISPLAY_VER(display) >= 10) max_tmds_clock = 594000; - else if (DISPLAY_VER(dev_priv) >= 8 || IS_HASWELL(dev_priv)) + else if (DISPLAY_VER(display) >= 8 || IS_HASWELL(dev_priv)) max_tmds_clock = 300000; - else if (DISPLAY_VER(dev_priv) >= 5) + else if (DISPLAY_VER(display) >= 5) max_tmds_clock = 225000; else max_tmds_clock = 165000; @@ -1848,7 +1851,8 @@ hdmi_port_clock_valid(struct intel_hdmi *hdmi, int clock, bool respect_downstream_limits, bool has_hdmi_sink) { - struct drm_i915_private *dev_priv = intel_hdmi_to_i915(hdmi); + struct intel_display *display = to_intel_display(hdmi); + struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_encoder *encoder = &hdmi_to_dig_port(hdmi)->base; if (clock < 25000) @@ -1885,7 +1889,7 @@ hdmi_port_clock_valid(struct intel_hdmi *hdmi, * FIXME: We will hopefully get an algorithmic way of programming * the MPLLB for HDMI in the future. */ - if (DISPLAY_VER(dev_priv) >= 14) + if (DISPLAY_VER(display) >= 14) return intel_cx0_phy_check_hdmi_link_rate(hdmi, clock); else if (IS_DG2(dev_priv)) return intel_snps_phy_check_hdmi_link_rate(clock); @@ -1908,13 +1912,13 @@ int intel_hdmi_tmds_clock(int clock, int bpc, return DIV_ROUND_CLOSEST(clock * bpc, 8); } -static bool intel_hdmi_source_bpc_possible(struct drm_i915_private *i915, int bpc) +static bool intel_hdmi_source_bpc_possible(struct intel_display *display, int bpc) { switch (bpc) { case 12: - return !HAS_GMCH(i915); + return !HAS_GMCH(display); case 10: - return DISPLAY_VER(i915) >= 11; + return DISPLAY_VER(display) >= 11; case 8: return true; default: @@ -1960,7 +1964,7 @@ intel_hdmi_mode_clock_valid(struct drm_connector *connector, int clock, bool has_hdmi_sink, enum intel_output_format sink_format) { - struct drm_i915_private *i915 = to_i915(connector->dev); + struct intel_display *display = to_intel_display(connector->dev); struct intel_hdmi *hdmi = intel_attached_hdmi(to_intel_connector(connector)); enum drm_mode_status status = MODE_OK; int bpc; @@ -1973,7 +1977,7 @@ intel_hdmi_mode_clock_valid(struct drm_connector *connector, int clock, for (bpc = 12; bpc >= 8; bpc -= 2) { int tmds_clock = intel_hdmi_tmds_clock(clock, bpc, sink_format); - if (!intel_hdmi_source_bpc_possible(i915, bpc)) + if (!intel_hdmi_source_bpc_possible(display, bpc)) continue; if (!intel_hdmi_sink_bpc_possible(connector, bpc, has_hdmi_sink, sink_format)) @@ -1985,7 +1989,7 @@ intel_hdmi_mode_clock_valid(struct drm_connector *connector, int clock, } /* can never happen */ - drm_WARN_ON(&i915->drm, status == MODE_OK); + drm_WARN_ON(display->drm, status == MODE_OK); return status; } @@ -1994,8 +1998,9 @@ static enum drm_mode_status intel_hdmi_mode_valid(struct drm_connector *connector, struct drm_display_mode *mode) { + struct intel_display *display = to_intel_display(connector->dev); struct intel_hdmi *hdmi = intel_attached_hdmi(to_intel_connector(connector)); - struct drm_i915_private *dev_priv = intel_hdmi_to_i915(hdmi); + struct drm_i915_private *dev_priv = to_i915(display->drm); enum drm_mode_status status; int clock = mode->clock; int max_dotclk = to_i915(connector->dev)->display.cdclk.max_dotclk_freq; @@ -2073,17 +2078,16 @@ bool intel_hdmi_bpc_possible(const struct intel_crtc_state *crtc_state, static bool hdmi_bpc_possible(const struct intel_crtc_state *crtc_state, int bpc) { - struct drm_i915_private *dev_priv = - to_i915(crtc_state->uapi.crtc->dev); + struct intel_display *display = to_intel_display(crtc_state); const struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode; - if (!intel_hdmi_source_bpc_possible(dev_priv, bpc)) + if (!intel_hdmi_source_bpc_possible(display, bpc)) return false; /* Display Wa_1405510057:icl,ehl */ if (intel_hdmi_is_ycbcr420(crtc_state) && - bpc == 10 && DISPLAY_VER(dev_priv) == 11 && + bpc == 10 && DISPLAY_VER(display) == 11 && (adjusted_mode->crtc_hblank_end - adjusted_mode->crtc_hblank_start) % 8 == 2) return false; @@ -2130,7 +2134,7 @@ static int intel_hdmi_compute_clock(struct intel_encoder *encoder, struct intel_crtc_state *crtc_state, bool respect_downstream_limits) { - struct drm_i915_private *i915 = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); const struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode; int bpc, clock = adjusted_mode->crtc_clock; @@ -2153,7 +2157,7 @@ static int intel_hdmi_compute_clock(struct intel_encoder *encoder, */ crtc_state->pipe_bpp = min(crtc_state->pipe_bpp, bpc * 3); - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "picking %d bpc for HDMI output (pipe bpp: %d)\n", bpc, crtc_state->pipe_bpp); @@ -2230,10 +2234,10 @@ static int intel_hdmi_compute_output_format(struct intel_encoder *encoder, const struct drm_connector_state *conn_state, bool respect_downstream_limits) { + struct intel_display *display = to_intel_display(encoder); struct intel_connector *connector = to_intel_connector(conn_state->connector); const struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode; const struct drm_display_info *info = &connector->base.display_info; - struct drm_i915_private *i915 = to_i915(connector->base.dev); bool ycbcr_420_only = drm_mode_is_420_only(info, adjusted_mode); int ret; @@ -2241,7 +2245,7 @@ static int intel_hdmi_compute_output_format(struct intel_encoder *encoder, intel_hdmi_sink_format(crtc_state, connector, ycbcr_420_only); if (ycbcr_420_only && crtc_state->sink_format != INTEL_OUTPUT_FORMAT_YCBCR420) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "YCbCr 4:2:0 mode but YCbCr 4:2:0 output not possible. Falling back to RGB.\n"); crtc_state->sink_format = INTEL_OUTPUT_FORMAT_RGB; } @@ -2302,7 +2306,7 @@ int intel_hdmi_compute_config(struct intel_encoder *encoder, struct intel_crtc_state *pipe_config, struct drm_connector_state *conn_state) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); struct drm_display_mode *adjusted_mode = &pipe_config->hw.adjusted_mode; struct drm_connector *connector = conn_state->connector; struct drm_scdc *scdc = &connector->display_info.hdmi.scdc; @@ -2335,7 +2339,7 @@ int intel_hdmi_compute_config(struct intel_encoder *encoder, if (ret) ret = intel_hdmi_compute_output_format(encoder, pipe_config, conn_state, false); if (ret) { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "unsupported HDMI clock (%d kHz), rejecting mode\n", pipe_config->hw.adjusted_mode.crtc_clock); return ret; @@ -2370,22 +2374,22 @@ int intel_hdmi_compute_config(struct intel_encoder *encoder, conn_state); if (!intel_hdmi_compute_avi_infoframe(encoder, pipe_config, conn_state)) { - drm_dbg_kms(&dev_priv->drm, "bad AVI infoframe\n"); + drm_dbg_kms(display->drm, "bad AVI infoframe\n"); return -EINVAL; } if (!intel_hdmi_compute_spd_infoframe(encoder, pipe_config, conn_state)) { - drm_dbg_kms(&dev_priv->drm, "bad SPD infoframe\n"); + drm_dbg_kms(display->drm, "bad SPD infoframe\n"); return -EINVAL; } if (!intel_hdmi_compute_hdmi_infoframe(encoder, pipe_config, conn_state)) { - drm_dbg_kms(&dev_priv->drm, "bad HDMI infoframe\n"); + drm_dbg_kms(display->drm, "bad HDMI infoframe\n"); return -EINVAL; } if (!intel_hdmi_compute_drm_infoframe(encoder, pipe_config, conn_state)) { - drm_dbg_kms(&dev_priv->drm, "bad DRM infoframe\n"); + drm_dbg_kms(display->drm, "bad DRM infoframe\n"); return -EINVAL; } @@ -2418,13 +2422,14 @@ intel_hdmi_unset_edid(struct drm_connector *connector) static void intel_hdmi_dp_dual_mode_detect(struct drm_connector *connector) { + struct intel_display *display = to_intel_display(connector->dev); struct drm_i915_private *dev_priv = to_i915(connector->dev); struct intel_hdmi *hdmi = intel_attached_hdmi(to_intel_connector(connector)); struct intel_encoder *encoder = &hdmi_to_dig_port(hdmi)->base; struct i2c_adapter *ddc = connector->ddc; enum drm_dp_dual_mode_type type; - type = drm_dp_dual_mode_detect(&dev_priv->drm, ddc); + type = drm_dp_dual_mode_detect(display->drm, ddc); /* * Type 1 DVI adaptors are not required to implement any @@ -2438,7 +2443,7 @@ intel_hdmi_dp_dual_mode_detect(struct drm_connector *connector) if (type == DRM_DP_DUAL_MODE_UNKNOWN) { if (!connector->force && intel_bios_encoder_supports_dp_dual_mode(encoder->devdata)) { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "Assuming DP dual mode adaptor presence based on VBT\n"); type = DRM_DP_DUAL_MODE_TYPE1_DVI; } else { @@ -2451,17 +2456,17 @@ intel_hdmi_dp_dual_mode_detect(struct drm_connector *connector) hdmi->dp_dual_mode.type = type; hdmi->dp_dual_mode.max_tmds_clock = - drm_dp_dual_mode_max_tmds_clock(&dev_priv->drm, type, ddc); + drm_dp_dual_mode_max_tmds_clock(display->drm, type, ddc); - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "DP dual mode adaptor (%s) detected (max TMDS clock: %d kHz)\n", drm_dp_get_dual_mode_type_name(type), hdmi->dp_dual_mode.max_tmds_clock); /* Older VBTs are often buggy and can't be trusted :( Play it safe. */ - if ((DISPLAY_VER(dev_priv) >= 8 || IS_HASWELL(dev_priv)) && + if ((DISPLAY_VER(display) >= 8 || IS_HASWELL(dev_priv)) && !intel_bios_encoder_supports_dp_dual_mode(encoder->devdata)) { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "Ignoring DP dual mode adaptor max TMDS clock for native HDMI port\n"); hdmi->dp_dual_mode.max_tmds_clock = 0; } @@ -2470,6 +2475,7 @@ intel_hdmi_dp_dual_mode_detect(struct drm_connector *connector) static bool intel_hdmi_set_edid(struct drm_connector *connector) { + struct intel_display *display = to_intel_display(connector->dev); struct drm_i915_private *dev_priv = to_i915(connector->dev); struct intel_hdmi *intel_hdmi = intel_attached_hdmi(to_intel_connector(connector)); struct i2c_adapter *ddc = connector->ddc; @@ -2482,7 +2488,7 @@ intel_hdmi_set_edid(struct drm_connector *connector) drm_edid = drm_edid_read_ddc(connector, ddc); if (!drm_edid && !intel_gmbus_is_forced_bit(ddc)) { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "HDMI GMBUS EDID read failed, retry using GPIO bit-banging\n"); intel_gmbus_force_bit(ddc, true); drm_edid = drm_edid_read_ddc(connector, ddc); @@ -2511,13 +2517,14 @@ intel_hdmi_set_edid(struct drm_connector *connector) static enum drm_connector_status intel_hdmi_detect(struct drm_connector *connector, bool force) { + struct intel_display *display = to_intel_display(connector->dev); enum drm_connector_status status = connector_status_disconnected; struct drm_i915_private *dev_priv = to_i915(connector->dev); struct intel_hdmi *intel_hdmi = intel_attached_hdmi(to_intel_connector(connector)); struct intel_encoder *encoder = &hdmi_to_dig_port(intel_hdmi)->base; intel_wakeref_t wakeref; - drm_dbg_kms(&dev_priv->drm, "[CONNECTOR:%d:%s]\n", + drm_dbg_kms(display->drm, "[CONNECTOR:%d:%s]\n", connector->base.id, connector->name); if (!intel_display_device_enabled(dev_priv)) @@ -2528,7 +2535,7 @@ intel_hdmi_detect(struct drm_connector *connector, bool force) wakeref = intel_display_power_get(dev_priv, POWER_DOMAIN_GMBUS); - if (DISPLAY_VER(dev_priv) >= 11 && + if (DISPLAY_VER(display) >= 11 && !intel_digital_port_connected(encoder)) goto out; @@ -2549,9 +2556,10 @@ out: static void intel_hdmi_force(struct drm_connector *connector) { + struct intel_display *display = to_intel_display(connector->dev); struct drm_i915_private *i915 = to_i915(connector->dev); - drm_dbg_kms(&i915->drm, "[CONNECTOR:%d:%s]\n", + drm_dbg_kms(display->drm, "[CONNECTOR:%d:%s]\n", connector->base.id, connector->name); if (!intel_display_driver_check_access(i915)) @@ -2608,9 +2616,9 @@ static const struct drm_connector_funcs intel_hdmi_connector_funcs = { static int intel_hdmi_connector_atomic_check(struct drm_connector *connector, struct drm_atomic_state *state) { - struct drm_i915_private *i915 = to_i915(state->dev); + struct intel_display *display = to_intel_display(connector->dev); - if (HAS_DDI(i915)) + if (HAS_DDI(display)) return intel_digital_connector_atomic_check(connector, state); else return g4x_hdmi_connector_atomic_check(connector, state); @@ -2625,7 +2633,7 @@ static const struct drm_connector_helper_funcs intel_hdmi_connector_helper_funcs static void intel_hdmi_add_properties(struct intel_hdmi *intel_hdmi, struct drm_connector *connector) { - struct drm_i915_private *dev_priv = to_i915(connector->dev); + struct intel_display *display = to_intel_display(intel_hdmi); intel_attach_force_audio_property(connector); intel_attach_broadcast_rgb_property(connector); @@ -2634,10 +2642,10 @@ intel_hdmi_add_properties(struct intel_hdmi *intel_hdmi, struct drm_connector *c intel_attach_hdmi_colorspace_property(connector); drm_connector_attach_content_type_property(connector); - if (DISPLAY_VER(dev_priv) >= 10) + if (DISPLAY_VER(display) >= 10) drm_connector_attach_hdr_output_metadata_property(connector); - if (!HAS_GMCH(dev_priv)) + if (!HAS_GMCH(display)) drm_connector_attach_max_bpc_property(connector, 8, 12); } @@ -2664,14 +2672,14 @@ bool intel_hdmi_handle_sink_scrambling(struct intel_encoder *encoder, bool high_tmds_clock_ratio, bool scrambling) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); struct drm_scrambling *sink_scrambling = &connector->display_info.hdmi.scdc.scrambling; if (!sink_scrambling->supported) return true; - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "[CONNECTOR:%d:%s] scrambling=%s, TMDS bit clock ratio=1/%d\n", connector->base.id, connector->name, str_yes_no(scrambling), high_tmds_clock_ratio ? 40 : 10); @@ -2752,7 +2760,7 @@ static u8 cnp_encoder_to_ddc_pin(struct intel_encoder *encoder) static u8 icl_encoder_to_ddc_pin(struct intel_encoder *encoder) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); enum port port = encoder->port; if (intel_encoder_is_combo(encoder)) @@ -2760,7 +2768,7 @@ static u8 icl_encoder_to_ddc_pin(struct intel_encoder *encoder) else if (intel_encoder_is_tc(encoder)) return GMBUS_PIN_9_TC1_ICP + intel_encoder_to_tc(encoder); - drm_WARN(&dev_priv->drm, 1, "Unknown port:%c\n", port_name(port)); + drm_WARN(display->drm, 1, "Unknown port:%c\n", port_name(port)); return GMBUS_PIN_2_BXT; } @@ -2808,10 +2816,11 @@ static u8 rkl_encoder_to_ddc_pin(struct intel_encoder *encoder) static u8 gen9bc_tgp_encoder_to_ddc_pin(struct intel_encoder *encoder) { + struct intel_display *display = to_intel_display(encoder); struct drm_i915_private *i915 = to_i915(encoder->base.dev); enum phy phy = intel_encoder_to_phy(encoder); - drm_WARN_ON(&i915->drm, encoder->port == PORT_A); + drm_WARN_ON(display->drm, encoder->port == PORT_A); /* * Pin mapping for GEN9 BC depends on which PCH is present. With TGP, @@ -2871,6 +2880,7 @@ static u8 g4x_encoder_to_ddc_pin(struct intel_encoder *encoder) static u8 intel_hdmi_default_ddc_pin(struct intel_encoder *encoder) { + struct intel_display *display = to_intel_display(encoder); struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); u8 ddc_pin; @@ -2880,7 +2890,7 @@ static u8 intel_hdmi_default_ddc_pin(struct intel_encoder *encoder) ddc_pin = dg1_encoder_to_ddc_pin(encoder); else if (IS_ROCKETLAKE(dev_priv)) ddc_pin = rkl_encoder_to_ddc_pin(encoder); - else if (DISPLAY_VER(dev_priv) == 9 && HAS_PCH_TGP(dev_priv)) + else if (DISPLAY_VER(display) == 9 && HAS_PCH_TGP(dev_priv)) ddc_pin = gen9bc_tgp_encoder_to_ddc_pin(encoder); else if ((IS_JASPERLAKE(dev_priv) || IS_ELKHARTLAKE(dev_priv)) && HAS_PCH_TGP(dev_priv)) @@ -2902,10 +2912,11 @@ static u8 intel_hdmi_default_ddc_pin(struct intel_encoder *encoder) static struct intel_encoder * get_encoder_by_ddc_pin(struct intel_encoder *encoder, u8 ddc_pin) { + struct intel_display *display = to_intel_display(encoder); struct drm_i915_private *i915 = to_i915(encoder->base.dev); struct intel_encoder *other; - for_each_intel_encoder(&i915->drm, other) { + for_each_intel_encoder(display->drm, other) { struct intel_connector *connector; if (other == encoder) @@ -2925,6 +2936,7 @@ get_encoder_by_ddc_pin(struct intel_encoder *encoder, u8 ddc_pin) static u8 intel_hdmi_ddc_pin(struct intel_encoder *encoder) { + struct intel_display *display = to_intel_display(encoder); struct drm_i915_private *i915 = to_i915(encoder->base.dev); struct intel_encoder *other; const char *source; @@ -2939,20 +2951,22 @@ static u8 intel_hdmi_ddc_pin(struct intel_encoder *encoder) } if (!intel_gmbus_is_valid_pin(i915, ddc_pin)) { - drm_dbg_kms(&i915->drm, "[ENCODER:%d:%s] Invalid DDC pin %d\n", + drm_dbg_kms(display->drm, + "[ENCODER:%d:%s] Invalid DDC pin %d\n", encoder->base.base.id, encoder->base.name, ddc_pin); return 0; } other = get_encoder_by_ddc_pin(encoder, ddc_pin); if (other) { - drm_dbg_kms(&i915->drm, "[ENCODER:%d:%s] DDC pin %d already claimed by [ENCODER:%d:%s]\n", + drm_dbg_kms(display->drm, + "[ENCODER:%d:%s] DDC pin %d already claimed by [ENCODER:%d:%s]\n", encoder->base.base.id, encoder->base.name, ddc_pin, other->base.base.id, other->base.name); return 0; } - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[ENCODER:%d:%s] Using DDC pin 0x%x (%s)\n", encoder->base.base.id, encoder->base.name, ddc_pin, source); @@ -2962,6 +2976,7 @@ static u8 intel_hdmi_ddc_pin(struct intel_encoder *encoder) void intel_infoframe_init(struct intel_digital_port *dig_port) { + struct intel_display *display = to_intel_display(dig_port); struct drm_i915_private *dev_priv = to_i915(dig_port->base.base.dev); @@ -2975,7 +2990,7 @@ void intel_infoframe_init(struct intel_digital_port *dig_port) dig_port->read_infoframe = g4x_read_infoframe; dig_port->set_infoframes = g4x_set_infoframes; dig_port->infoframes_enabled = g4x_infoframes_enabled; - } else if (HAS_DDI(dev_priv)) { + } else if (HAS_DDI(display)) { if (intel_bios_encoder_is_lspcon(dig_port->base.devdata)) { dig_port->write_infoframe = lspcon_write_infoframe; dig_port->read_infoframe = lspcon_read_infoframe; @@ -3003,6 +3018,7 @@ void intel_infoframe_init(struct intel_digital_port *dig_port) void intel_hdmi_init_connector(struct intel_digital_port *dig_port, struct intel_connector *intel_connector) { + struct intel_display *display = to_intel_display(dig_port); struct drm_connector *connector = &intel_connector->base; struct intel_hdmi *intel_hdmi = &dig_port->hdmi; struct intel_encoder *intel_encoder = &dig_port->base; @@ -3012,11 +3028,11 @@ void intel_hdmi_init_connector(struct intel_digital_port *dig_port, struct cec_connector_info conn_info; u8 ddc_pin; - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "Adding HDMI connector on [ENCODER:%d:%s]\n", intel_encoder->base.base.id, intel_encoder->base.name); - if (DISPLAY_VER(dev_priv) < 12 && drm_WARN_ON(dev, port == PORT_A)) + if (DISPLAY_VER(display) < 12 && drm_WARN_ON(dev, port == PORT_A)) return; if (drm_WARN(dev, dig_port->max_lanes < 4, @@ -3036,18 +3052,18 @@ void intel_hdmi_init_connector(struct intel_digital_port *dig_port, drm_connector_helper_add(connector, &intel_hdmi_connector_helper_funcs); - if (DISPLAY_VER(dev_priv) < 12) + if (DISPLAY_VER(display) < 12) connector->interlace_allowed = true; connector->stereo_allowed = true; - if (DISPLAY_VER(dev_priv) >= 10) + if (DISPLAY_VER(display) >= 10) connector->ycbcr_420_allowed = true; intel_connector->polled = DRM_CONNECTOR_POLL_HPD; intel_connector->base.polled = intel_connector->polled; - if (HAS_DDI(dev_priv)) + if (HAS_DDI(display)) intel_connector->get_hw_state = intel_ddi_connector_get_hw_state; else intel_connector->get_hw_state = intel_connector_get_hw_state; @@ -3061,7 +3077,7 @@ void intel_hdmi_init_connector(struct intel_digital_port *dig_port, int ret = intel_hdcp_init(intel_connector, dig_port, &intel_hdmi_hdcp_shim); if (ret) - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "HDCP init failed, skipping.\n"); } @@ -3071,7 +3087,7 @@ void intel_hdmi_init_connector(struct intel_digital_port *dig_port, cec_notifier_conn_register(dev->dev, port_identifier(port), &conn_info); if (!intel_hdmi->cec_notifier) - drm_dbg_kms(&dev_priv->drm, "CEC notifier get failed\n"); + drm_dbg_kms(display->drm, "CEC notifier get failed\n"); } /* diff --git a/drivers/gpu/drm/i915/display/intel_hdmi.h b/drivers/gpu/drm/i915/display/intel_hdmi.h index 6b39df38d57a..9b97623665c5 100644 --- a/drivers/gpu/drm/i915/display/intel_hdmi.h +++ b/drivers/gpu/drm/i915/display/intel_hdmi.h @@ -58,6 +58,5 @@ int intel_hdmi_dsc_get_num_slices(const struct intel_crtc_state *crtc_state, int src_max_slices, int src_max_slice_width, int hdmi_max_slices, int hdmi_throughput); int intel_hdmi_dsc_get_slice_height(int vactive); -struct drm_i915_private *intel_hdmi_to_i915(struct intel_hdmi *intel_hdmi); #endif /* __INTEL_HDMI_H__ */ diff --git a/drivers/gpu/drm/i915/display/intel_hotplug_irq.c b/drivers/gpu/drm/i915/display/intel_hotplug_irq.c index a1f07ee69a86..2c4e946d5575 100644 --- a/drivers/gpu/drm/i915/display/intel_hotplug_irq.c +++ b/drivers/gpu/drm/i915/display/intel_hotplug_irq.c @@ -456,6 +456,7 @@ u32 i9xx_hpd_irq_ack(struct drm_i915_private *dev_priv) void i9xx_hpd_irq_handler(struct drm_i915_private *dev_priv, u32 hotplug_status) { + struct intel_display *display = &dev_priv->display; u32 pin_mask = 0, long_mask = 0; u32 hotplug_trigger; @@ -477,7 +478,7 @@ void i9xx_hpd_irq_handler(struct drm_i915_private *dev_priv, u32 hotplug_status) if ((IS_G4X(dev_priv) || IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) && hotplug_status & DP_AUX_CHANNEL_MASK_INT_STATUS_G4X) - intel_dp_aux_irq_handler(dev_priv); + intel_dp_aux_irq_handler(display); } void ibx_hpd_irq_handler(struct drm_i915_private *dev_priv, u32 hotplug_trigger) @@ -513,6 +514,7 @@ void ibx_hpd_irq_handler(struct drm_i915_private *dev_priv, u32 hotplug_trigger) void xelpdp_pica_irq_handler(struct drm_i915_private *i915, u32 iir) { + struct intel_display *display = &i915->display; enum hpd_pin pin; u32 hotplug_trigger = iir & (XELPDP_DP_ALT_HOTPLUG_MASK | XELPDP_TBT_HOTPLUG_MASK); u32 trigger_aux = iir & XELPDP_AUX_TC_MASK; @@ -545,7 +547,7 @@ void xelpdp_pica_irq_handler(struct drm_i915_private *i915, u32 iir) } if (trigger_aux) - intel_dp_aux_irq_handler(i915); + intel_dp_aux_irq_handler(display); if (!pin_mask && !trigger_aux) drm_err(&i915->drm, diff --git a/drivers/gpu/drm/i915/display/intel_pch_display.c b/drivers/gpu/drm/i915/display/intel_pch_display.c index 0d48b9bec29c..f13ab680c2cf 100644 --- a/drivers/gpu/drm/i915/display/intel_pch_display.c +++ b/drivers/gpu/drm/i915/display/intel_pch_display.c @@ -358,6 +358,7 @@ void ilk_pch_pre_enable(struct intel_atomic_state *state, void ilk_pch_enable(struct intel_atomic_state *state, struct intel_crtc *crtc) { + struct intel_display *display = to_intel_display(state); struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); const struct intel_crtc_state *crtc_state = intel_atomic_get_new_crtc_state(state, crtc); @@ -399,7 +400,7 @@ void ilk_pch_enable(struct intel_atomic_state *state, intel_enable_shared_dpll(crtc_state); /* set transcoder timing, panel must allow it */ - assert_pps_unlocked(dev_priv, pipe); + assert_pps_unlocked(display, pipe); if (intel_crtc_has_dp_encoder(crtc_state)) { intel_pch_transcoder_set_m1_n1(crtc, &crtc_state->dp_m_n); intel_pch_transcoder_set_m2_n2(crtc, &crtc_state->dp_m2_n2); diff --git a/drivers/gpu/drm/i915/display/intel_pps.c b/drivers/gpu/drm/i915/display/intel_pps.c index 68141af4da54..feddc30e3375 100644 --- a/drivers/gpu/drm/i915/display/intel_pps.c +++ b/drivers/gpu/drm/i915/display/intel_pps.c @@ -18,15 +18,18 @@ #include "intel_pps_regs.h" #include "intel_quirks.h" -static void vlv_steal_power_sequencer(struct drm_i915_private *dev_priv, +static void vlv_steal_power_sequencer(struct intel_display *display, enum pipe pipe); static void pps_init_delays(struct intel_dp *intel_dp); static void pps_init_registers(struct intel_dp *intel_dp, bool force_disable_vdd); -static const char *pps_name(struct drm_i915_private *i915, - struct intel_pps *pps) +static const char *pps_name(struct intel_dp *intel_dp) { + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *i915 = to_i915(display->drm); + struct intel_pps *pps = &intel_dp->pps; + if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915)) { switch (pps->pps_pipe) { case INVALID_PIPE: @@ -60,14 +63,15 @@ static const char *pps_name(struct drm_i915_private *i915, intel_wakeref_t intel_pps_lock(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); intel_wakeref_t wakeref; /* * See intel_pps_reset_all() why we need a power domain reference here. */ wakeref = intel_display_power_get(dev_priv, POWER_DOMAIN_DISPLAY_CORE); - mutex_lock(&dev_priv->display.pps.mutex); + mutex_lock(&display->pps.mutex); return wakeref; } @@ -75,9 +79,10 @@ intel_wakeref_t intel_pps_lock(struct intel_dp *intel_dp) intel_wakeref_t intel_pps_unlock(struct intel_dp *intel_dp, intel_wakeref_t wakeref) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); - mutex_unlock(&dev_priv->display.pps.mutex); + mutex_unlock(&display->pps.mutex); intel_display_power_put(dev_priv, POWER_DOMAIN_DISPLAY_CORE, wakeref); return 0; @@ -86,7 +91,8 @@ intel_wakeref_t intel_pps_unlock(struct intel_dp *intel_dp, static void vlv_power_sequencer_kick(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); enum pipe pipe = intel_dp->pps.pps_pipe; bool pll_enabled, release_cl_override = false; @@ -94,22 +100,22 @@ vlv_power_sequencer_kick(struct intel_dp *intel_dp) enum dpio_channel ch = vlv_pipe_to_channel(pipe); u32 DP; - if (drm_WARN(&dev_priv->drm, - intel_de_read(dev_priv, intel_dp->output_reg) & DP_PORT_EN, + if (drm_WARN(display->drm, + intel_de_read(display, intel_dp->output_reg) & DP_PORT_EN, "skipping %s kick due to [ENCODER:%d:%s] being active\n", - pps_name(dev_priv, &intel_dp->pps), + pps_name(intel_dp), dig_port->base.base.base.id, dig_port->base.base.name)) return; - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "kicking %s for [ENCODER:%d:%s]\n", - pps_name(dev_priv, &intel_dp->pps), + pps_name(intel_dp), dig_port->base.base.base.id, dig_port->base.base.name); /* Preserve the BIOS-computed detected bit. This is * supposed to be read-only. */ - DP = intel_de_read(dev_priv, intel_dp->output_reg) & DP_DETECTED; + DP = intel_de_read(display, intel_dp->output_reg) & DP_DETECTED; DP |= DP_VOLTAGE_0_4 | DP_PRE_EMPHASIS_0; DP |= DP_PORT_WIDTH(1); DP |= DP_LINK_TRAIN_PAT_1; @@ -119,7 +125,7 @@ vlv_power_sequencer_kick(struct intel_dp *intel_dp) else DP |= DP_PIPE_SEL(pipe); - pll_enabled = intel_de_read(dev_priv, DPLL(dev_priv, pipe)) & DPLL_VCO_ENABLE; + pll_enabled = intel_de_read(display, DPLL(display, pipe)) & DPLL_VCO_ENABLE; /* * The DPLL for the pipe must be enabled for this to work. @@ -130,7 +136,7 @@ vlv_power_sequencer_kick(struct intel_dp *intel_dp) !chv_phy_powergate_ch(dev_priv, phy, ch, true); if (vlv_force_pll_on(dev_priv, pipe, vlv_get_dpll(dev_priv))) { - drm_err(&dev_priv->drm, + drm_err(display->drm, "Failed to force on PLL for pipe %c!\n", pipe_name(pipe)); return; @@ -143,14 +149,14 @@ vlv_power_sequencer_kick(struct intel_dp *intel_dp) * to make this power sequencer lock onto the port. * Otherwise even VDD force bit won't work. */ - intel_de_write(dev_priv, intel_dp->output_reg, DP); - intel_de_posting_read(dev_priv, intel_dp->output_reg); + intel_de_write(display, intel_dp->output_reg, DP); + intel_de_posting_read(display, intel_dp->output_reg); - intel_de_write(dev_priv, intel_dp->output_reg, DP | DP_PORT_EN); - intel_de_posting_read(dev_priv, intel_dp->output_reg); + intel_de_write(display, intel_dp->output_reg, DP | DP_PORT_EN); + intel_de_posting_read(display, intel_dp->output_reg); - intel_de_write(dev_priv, intel_dp->output_reg, DP & ~DP_PORT_EN); - intel_de_posting_read(dev_priv, intel_dp->output_reg); + intel_de_write(display, intel_dp->output_reg, DP & ~DP_PORT_EN); + intel_de_posting_read(display, intel_dp->output_reg); if (!pll_enabled) { vlv_force_pll_off(dev_priv, pipe); @@ -160,7 +166,7 @@ vlv_power_sequencer_kick(struct intel_dp *intel_dp) } } -static enum pipe vlv_find_free_pps(struct drm_i915_private *dev_priv) +static enum pipe vlv_find_free_pps(struct intel_display *display) { struct intel_encoder *encoder; unsigned int pipes = (1 << PIPE_A) | (1 << PIPE_B); @@ -169,11 +175,11 @@ static enum pipe vlv_find_free_pps(struct drm_i915_private *dev_priv) * We don't have power sequencer currently. * Pick one that's not used by other ports. */ - for_each_intel_dp(&dev_priv->drm, encoder) { + for_each_intel_dp(display->drm, encoder) { struct intel_dp *intel_dp = enc_to_intel_dp(encoder); if (encoder->type == INTEL_OUTPUT_EDP) { - drm_WARN_ON(&dev_priv->drm, + drm_WARN_ON(display->drm, intel_dp->pps.active_pipe != INVALID_PIPE && intel_dp->pps.active_pipe != intel_dp->pps.pps_pipe); @@ -181,7 +187,7 @@ static enum pipe vlv_find_free_pps(struct drm_i915_private *dev_priv) if (intel_dp->pps.pps_pipe != INVALID_PIPE) pipes &= ~(1 << intel_dp->pps.pps_pipe); } else { - drm_WARN_ON(&dev_priv->drm, + drm_WARN_ON(display->drm, intel_dp->pps.pps_pipe != INVALID_PIPE); if (intel_dp->pps.active_pipe != INVALID_PIPE) @@ -198,36 +204,36 @@ static enum pipe vlv_find_free_pps(struct drm_i915_private *dev_priv) static enum pipe vlv_power_sequencer_pipe(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); enum pipe pipe; - lockdep_assert_held(&dev_priv->display.pps.mutex); + lockdep_assert_held(&display->pps.mutex); /* We should never land here with regular DP ports */ - drm_WARN_ON(&dev_priv->drm, !intel_dp_is_edp(intel_dp)); + drm_WARN_ON(display->drm, !intel_dp_is_edp(intel_dp)); - drm_WARN_ON(&dev_priv->drm, intel_dp->pps.active_pipe != INVALID_PIPE && + drm_WARN_ON(display->drm, intel_dp->pps.active_pipe != INVALID_PIPE && intel_dp->pps.active_pipe != intel_dp->pps.pps_pipe); if (intel_dp->pps.pps_pipe != INVALID_PIPE) return intel_dp->pps.pps_pipe; - pipe = vlv_find_free_pps(dev_priv); + pipe = vlv_find_free_pps(display); /* * Didn't find one. This should not happen since there * are two power sequencers and up to two eDP ports. */ - if (drm_WARN_ON(&dev_priv->drm, pipe == INVALID_PIPE)) + if (drm_WARN_ON(display->drm, pipe == INVALID_PIPE)) pipe = PIPE_A; - vlv_steal_power_sequencer(dev_priv, pipe); + vlv_steal_power_sequencer(display, pipe); intel_dp->pps.pps_pipe = pipe; - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "picked %s for [ENCODER:%d:%s]\n", - pps_name(dev_priv, &intel_dp->pps), + pps_name(intel_dp), dig_port->base.base.base.id, dig_port->base.base.name); /* init power sequencer on this pipe and port */ @@ -246,13 +252,13 @@ vlv_power_sequencer_pipe(struct intel_dp *intel_dp) static int bxt_power_sequencer_idx(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); int pps_idx = intel_dp->pps.pps_idx; - lockdep_assert_held(&dev_priv->display.pps.mutex); + lockdep_assert_held(&display->pps.mutex); /* We should never land here with regular DP ports */ - drm_WARN_ON(&dev_priv->drm, !intel_dp_is_edp(intel_dp)); + drm_WARN_ON(display->drm, !intel_dp_is_edp(intel_dp)); if (!intel_dp->pps.pps_reset) return pps_idx; @@ -268,37 +274,38 @@ bxt_power_sequencer_idx(struct intel_dp *intel_dp) return pps_idx; } -typedef bool (*pps_check)(struct drm_i915_private *dev_priv, int pps_idx); +typedef bool (*pps_check)(struct intel_display *display, int pps_idx); -static bool pps_has_pp_on(struct drm_i915_private *dev_priv, int pps_idx) +static bool pps_has_pp_on(struct intel_display *display, int pps_idx) { - return intel_de_read(dev_priv, PP_STATUS(dev_priv, pps_idx)) & PP_ON; + return intel_de_read(display, PP_STATUS(display, pps_idx)) & PP_ON; } -static bool pps_has_vdd_on(struct drm_i915_private *dev_priv, int pps_idx) +static bool pps_has_vdd_on(struct intel_display *display, int pps_idx) { - return intel_de_read(dev_priv, PP_CONTROL(dev_priv, pps_idx)) & EDP_FORCE_VDD; + return intel_de_read(display, PP_CONTROL(display, pps_idx)) & EDP_FORCE_VDD; } -static bool pps_any(struct drm_i915_private *dev_priv, int pps_idx) +static bool pps_any(struct intel_display *display, int pps_idx) { return true; } static enum pipe -vlv_initial_pps_pipe(struct drm_i915_private *dev_priv, +vlv_initial_pps_pipe(struct intel_display *display, enum port port, pps_check check) { enum pipe pipe; for (pipe = PIPE_A; pipe <= PIPE_B; pipe++) { - u32 port_sel = intel_de_read(dev_priv, PP_ON_DELAYS(dev_priv, pipe)) & + u32 port_sel = intel_de_read(display, + PP_ON_DELAYS(display, pipe)) & PANEL_PORT_SELECT_MASK; if (port_sel != PANEL_PORT_SELECT_VLV(port)) continue; - if (!check(dev_priv, pipe)) + if (!check(display, pipe)) continue; return pipe; @@ -310,41 +317,43 @@ vlv_initial_pps_pipe(struct drm_i915_private *dev_priv, static void vlv_initial_power_sequencer_setup(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); enum port port = dig_port->base.port; - lockdep_assert_held(&dev_priv->display.pps.mutex); + lockdep_assert_held(&display->pps.mutex); /* try to find a pipe with this port selected */ /* first pick one where the panel is on */ - intel_dp->pps.pps_pipe = vlv_initial_pps_pipe(dev_priv, port, + intel_dp->pps.pps_pipe = vlv_initial_pps_pipe(display, port, pps_has_pp_on); /* didn't find one? pick one where vdd is on */ if (intel_dp->pps.pps_pipe == INVALID_PIPE) - intel_dp->pps.pps_pipe = vlv_initial_pps_pipe(dev_priv, port, + intel_dp->pps.pps_pipe = vlv_initial_pps_pipe(display, port, pps_has_vdd_on); /* didn't find one? pick one with just the correct port */ if (intel_dp->pps.pps_pipe == INVALID_PIPE) - intel_dp->pps.pps_pipe = vlv_initial_pps_pipe(dev_priv, port, + intel_dp->pps.pps_pipe = vlv_initial_pps_pipe(display, port, pps_any); /* didn't find one? just let vlv_power_sequencer_pipe() pick one when needed */ if (intel_dp->pps.pps_pipe == INVALID_PIPE) { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "[ENCODER:%d:%s] no initial power sequencer\n", dig_port->base.base.base.id, dig_port->base.base.name); return; } - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "[ENCODER:%d:%s] initial power sequencer: %s\n", dig_port->base.base.base.id, dig_port->base.base.name, - pps_name(dev_priv, &intel_dp->pps)); + pps_name(intel_dp)); } -static int intel_num_pps(struct drm_i915_private *i915) +static int intel_num_pps(struct intel_display *display) { + struct drm_i915_private *i915 = to_i915(display->drm); + if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915)) return 2; @@ -365,23 +374,24 @@ static int intel_num_pps(struct drm_i915_private *i915) static bool intel_pps_is_valid(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *i915 = to_i915(display->drm); if (intel_dp->pps.pps_idx == 1 && INTEL_PCH_TYPE(i915) >= PCH_ICP && INTEL_PCH_TYPE(i915) <= PCH_ADP) - return intel_de_read(i915, SOUTH_CHICKEN1) & ICP_SECOND_PPS_IO_SELECT; + return intel_de_read(display, SOUTH_CHICKEN1) & ICP_SECOND_PPS_IO_SELECT; return true; } static int -bxt_initial_pps_idx(struct drm_i915_private *i915, pps_check check) +bxt_initial_pps_idx(struct intel_display *display, pps_check check) { - int pps_idx, pps_num = intel_num_pps(i915); + int pps_idx, pps_num = intel_num_pps(display); for (pps_idx = 0; pps_idx < pps_num; pps_idx++) { - if (check(i915, pps_idx)) + if (check(display, pps_idx)) return pps_idx; } @@ -391,11 +401,12 @@ bxt_initial_pps_idx(struct drm_i915_private *i915, pps_check check) static bool pps_initial_setup(struct intel_dp *intel_dp) { + struct intel_display *display = to_intel_display(intel_dp); struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base; struct intel_connector *connector = intel_dp->attached_connector; struct drm_i915_private *i915 = to_i915(encoder->base.dev); - lockdep_assert_held(&i915->display.pps.mutex); + lockdep_assert_held(&display->pps.mutex); if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915)) { vlv_initial_power_sequencer_setup(intel_dp); @@ -403,46 +414,47 @@ pps_initial_setup(struct intel_dp *intel_dp) } /* first ask the VBT */ - if (intel_num_pps(i915) > 1) + if (intel_num_pps(display) > 1) intel_dp->pps.pps_idx = connector->panel.vbt.backlight.controller; else intel_dp->pps.pps_idx = 0; - if (drm_WARN_ON(&i915->drm, intel_dp->pps.pps_idx >= intel_num_pps(i915))) + if (drm_WARN_ON(display->drm, intel_dp->pps.pps_idx >= intel_num_pps(display))) intel_dp->pps.pps_idx = -1; /* VBT wasn't parsed yet? pick one where the panel is on */ if (intel_dp->pps.pps_idx < 0) - intel_dp->pps.pps_idx = bxt_initial_pps_idx(i915, pps_has_pp_on); + intel_dp->pps.pps_idx = bxt_initial_pps_idx(display, pps_has_pp_on); /* didn't find one? pick one where vdd is on */ if (intel_dp->pps.pps_idx < 0) - intel_dp->pps.pps_idx = bxt_initial_pps_idx(i915, pps_has_vdd_on); + intel_dp->pps.pps_idx = bxt_initial_pps_idx(display, pps_has_vdd_on); /* didn't find one? pick any */ if (intel_dp->pps.pps_idx < 0) { - intel_dp->pps.pps_idx = bxt_initial_pps_idx(i915, pps_any); + intel_dp->pps.pps_idx = bxt_initial_pps_idx(display, pps_any); - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[ENCODER:%d:%s] no initial power sequencer, assuming %s\n", encoder->base.base.id, encoder->base.name, - pps_name(i915, &intel_dp->pps)); + pps_name(intel_dp)); } else { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "[ENCODER:%d:%s] initial power sequencer: %s\n", encoder->base.base.id, encoder->base.name, - pps_name(i915, &intel_dp->pps)); + pps_name(intel_dp)); } return intel_pps_is_valid(intel_dp); } -void intel_pps_reset_all(struct drm_i915_private *dev_priv) +void intel_pps_reset_all(struct intel_display *display) { + struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_encoder *encoder; - if (drm_WARN_ON(&dev_priv->drm, !IS_LP(dev_priv))) + if (drm_WARN_ON(display->drm, !IS_LP(dev_priv))) return; - if (!HAS_DISPLAY(dev_priv)) + if (!HAS_DISPLAY(display)) return; /* @@ -455,16 +467,16 @@ void intel_pps_reset_all(struct drm_i915_private *dev_priv) * should use them always. */ - for_each_intel_dp(&dev_priv->drm, encoder) { + for_each_intel_dp(display->drm, encoder) { struct intel_dp *intel_dp = enc_to_intel_dp(encoder); - drm_WARN_ON(&dev_priv->drm, + drm_WARN_ON(display->drm, intel_dp->pps.active_pipe != INVALID_PIPE); if (encoder->type != INTEL_OUTPUT_EDP) continue; - if (DISPLAY_VER(dev_priv) >= 9) + if (DISPLAY_VER(display) >= 9) intel_dp->pps.pps_reset = true; else intel_dp->pps.pps_pipe = INVALID_PIPE; @@ -482,7 +494,8 @@ struct pps_registers { static void intel_pps_get_registers(struct intel_dp *intel_dp, struct pps_registers *regs) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); int pps_idx; memset(regs, 0, sizeof(*regs)); @@ -494,17 +507,17 @@ static void intel_pps_get_registers(struct intel_dp *intel_dp, else pps_idx = intel_dp->pps.pps_idx; - regs->pp_ctrl = PP_CONTROL(dev_priv, pps_idx); - regs->pp_stat = PP_STATUS(dev_priv, pps_idx); - regs->pp_on = PP_ON_DELAYS(dev_priv, pps_idx); - regs->pp_off = PP_OFF_DELAYS(dev_priv, pps_idx); + regs->pp_ctrl = PP_CONTROL(display, pps_idx); + regs->pp_stat = PP_STATUS(display, pps_idx); + regs->pp_on = PP_ON_DELAYS(display, pps_idx); + regs->pp_off = PP_OFF_DELAYS(display, pps_idx); /* Cycle delay moved from PP_DIVISOR to PP_CONTROL */ if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv) || INTEL_PCH_TYPE(dev_priv) >= PCH_CNP) regs->pp_div = INVALID_MMIO_REG; else - regs->pp_div = PP_DIVISOR(dev_priv, pps_idx); + regs->pp_div = PP_DIVISOR(display, pps_idx); } static i915_reg_t @@ -529,49 +542,51 @@ _pp_stat_reg(struct intel_dp *intel_dp) static bool edp_have_panel_power(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); - lockdep_assert_held(&dev_priv->display.pps.mutex); + lockdep_assert_held(&display->pps.mutex); if ((IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) && intel_dp->pps.pps_pipe == INVALID_PIPE) return false; - return (intel_de_read(dev_priv, _pp_stat_reg(intel_dp)) & PP_ON) != 0; + return (intel_de_read(display, _pp_stat_reg(intel_dp)) & PP_ON) != 0; } static bool edp_have_panel_vdd(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); - lockdep_assert_held(&dev_priv->display.pps.mutex); + lockdep_assert_held(&display->pps.mutex); if ((IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) && intel_dp->pps.pps_pipe == INVALID_PIPE) return false; - return intel_de_read(dev_priv, _pp_ctrl_reg(intel_dp)) & EDP_FORCE_VDD; + return intel_de_read(display, _pp_ctrl_reg(intel_dp)) & EDP_FORCE_VDD; } void intel_pps_check_power_unlocked(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); if (!intel_dp_is_edp(intel_dp)) return; if (!edp_have_panel_power(intel_dp) && !edp_have_panel_vdd(intel_dp)) { - drm_WARN(&dev_priv->drm, 1, + drm_WARN(display->drm, 1, "[ENCODER:%d:%s] %s powered off while attempting AUX CH communication.\n", dig_port->base.base.base.id, dig_port->base.base.name, - pps_name(dev_priv, &intel_dp->pps)); - drm_dbg_kms(&dev_priv->drm, + pps_name(intel_dp)); + drm_dbg_kms(display->drm, "[ENCODER:%d:%s] %s PP_STATUS: 0x%08x PP_CONTROL: 0x%08x\n", dig_port->base.base.base.id, dig_port->base.base.name, - pps_name(dev_priv, &intel_dp->pps), - intel_de_read(dev_priv, _pp_stat_reg(intel_dp)), - intel_de_read(dev_priv, _pp_ctrl_reg(intel_dp))); + pps_name(intel_dp), + intel_de_read(display, _pp_stat_reg(intel_dp)), + intel_de_read(display, _pp_ctrl_reg(intel_dp))); } } @@ -589,68 +604,71 @@ static void intel_pps_verify_state(struct intel_dp *intel_dp); static void wait_panel_status(struct intel_dp *intel_dp, u32 mask, u32 value) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); i915_reg_t pp_stat_reg, pp_ctrl_reg; - lockdep_assert_held(&dev_priv->display.pps.mutex); + lockdep_assert_held(&display->pps.mutex); intel_pps_verify_state(intel_dp); pp_stat_reg = _pp_stat_reg(intel_dp); pp_ctrl_reg = _pp_ctrl_reg(intel_dp); - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "[ENCODER:%d:%s] %s mask: 0x%08x value: 0x%08x PP_STATUS: 0x%08x PP_CONTROL: 0x%08x\n", dig_port->base.base.base.id, dig_port->base.base.name, - pps_name(dev_priv, &intel_dp->pps), + pps_name(intel_dp), mask, value, - intel_de_read(dev_priv, pp_stat_reg), - intel_de_read(dev_priv, pp_ctrl_reg)); + intel_de_read(display, pp_stat_reg), + intel_de_read(display, pp_ctrl_reg)); - if (intel_de_wait(dev_priv, pp_stat_reg, mask, value, 5000)) - drm_err(&dev_priv->drm, + if (intel_de_wait(display, pp_stat_reg, mask, value, 5000)) + drm_err(display->drm, "[ENCODER:%d:%s] %s panel status timeout: PP_STATUS: 0x%08x PP_CONTROL: 0x%08x\n", dig_port->base.base.base.id, dig_port->base.base.name, - pps_name(dev_priv, &intel_dp->pps), - intel_de_read(dev_priv, pp_stat_reg), - intel_de_read(dev_priv, pp_ctrl_reg)); + pps_name(intel_dp), + intel_de_read(display, pp_stat_reg), + intel_de_read(display, pp_ctrl_reg)); - drm_dbg_kms(&dev_priv->drm, "Wait complete\n"); + drm_dbg_kms(display->drm, "Wait complete\n"); } static void wait_panel_on(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); - drm_dbg_kms(&i915->drm, "[ENCODER:%d:%s] %s wait for panel power on\n", + drm_dbg_kms(display->drm, + "[ENCODER:%d:%s] %s wait for panel power on\n", dig_port->base.base.base.id, dig_port->base.base.name, - pps_name(i915, &intel_dp->pps)); + pps_name(intel_dp)); wait_panel_status(intel_dp, IDLE_ON_MASK, IDLE_ON_VALUE); } static void wait_panel_off(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); - drm_dbg_kms(&i915->drm, "[ENCODER:%d:%s] %s wait for panel power off time\n", + drm_dbg_kms(display->drm, + "[ENCODER:%d:%s] %s wait for panel power off time\n", dig_port->base.base.base.id, dig_port->base.base.name, - pps_name(i915, &intel_dp->pps)); + pps_name(intel_dp)); wait_panel_status(intel_dp, IDLE_OFF_MASK, IDLE_OFF_VALUE); } static void wait_panel_power_cycle(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); ktime_t panel_power_on_time; s64 panel_power_off_duration; - drm_dbg_kms(&i915->drm, "[ENCODER:%d:%s] %s wait for panel power cycle\n", + drm_dbg_kms(display->drm, + "[ENCODER:%d:%s] %s wait for panel power cycle\n", dig_port->base.base.base.id, dig_port->base.base.name, - pps_name(i915, &intel_dp->pps)); + pps_name(intel_dp)); /* take the difference of current time and panel power off time * and then make panel wait for t11_t12 if needed. */ @@ -695,13 +713,13 @@ static void edp_wait_backlight_off(struct intel_dp *intel_dp) static u32 ilk_get_pp_control(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); u32 control; - lockdep_assert_held(&dev_priv->display.pps.mutex); + lockdep_assert_held(&display->pps.mutex); - control = intel_de_read(dev_priv, _pp_ctrl_reg(intel_dp)); - if (drm_WARN_ON(&dev_priv->drm, !HAS_DDI(dev_priv) && + control = intel_de_read(display, _pp_ctrl_reg(intel_dp)); + if (drm_WARN_ON(display->drm, !HAS_DDI(display) && (control & PANEL_UNLOCK_MASK) != PANEL_UNLOCK_REGS)) { control &= ~PANEL_UNLOCK_MASK; control |= PANEL_UNLOCK_REGS; @@ -716,13 +734,14 @@ static u32 ilk_get_pp_control(struct intel_dp *intel_dp) */ bool intel_pps_vdd_on_unlocked(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); u32 pp; i915_reg_t pp_stat_reg, pp_ctrl_reg; bool need_to_disable = !intel_dp->pps.want_panel_vdd; - lockdep_assert_held(&dev_priv->display.pps.mutex); + lockdep_assert_held(&display->pps.mutex); if (!intel_dp_is_edp(intel_dp)) return false; @@ -733,16 +752,16 @@ bool intel_pps_vdd_on_unlocked(struct intel_dp *intel_dp) if (edp_have_panel_vdd(intel_dp)) return need_to_disable; - drm_WARN_ON(&dev_priv->drm, intel_dp->pps.vdd_wakeref); + drm_WARN_ON(display->drm, intel_dp->pps.vdd_wakeref); intel_dp->pps.vdd_wakeref = intel_display_power_get(dev_priv, intel_aux_power_domain(dig_port)); pp_stat_reg = _pp_stat_reg(intel_dp); pp_ctrl_reg = _pp_ctrl_reg(intel_dp); - drm_dbg_kms(&dev_priv->drm, "[ENCODER:%d:%s] %s turning VDD on\n", + drm_dbg_kms(display->drm, "[ENCODER:%d:%s] %s turning VDD on\n", dig_port->base.base.base.id, dig_port->base.base.name, - pps_name(dev_priv, &intel_dp->pps)); + pps_name(intel_dp)); if (!edp_have_panel_power(intel_dp)) wait_panel_power_cycle(intel_dp); @@ -750,21 +769,22 @@ bool intel_pps_vdd_on_unlocked(struct intel_dp *intel_dp) pp = ilk_get_pp_control(intel_dp); pp |= EDP_FORCE_VDD; - intel_de_write(dev_priv, pp_ctrl_reg, pp); - intel_de_posting_read(dev_priv, pp_ctrl_reg); - drm_dbg_kms(&dev_priv->drm, "[ENCODER:%d:%s] %s PP_STATUS: 0x%08x PP_CONTROL: 0x%08x\n", + intel_de_write(display, pp_ctrl_reg, pp); + intel_de_posting_read(display, pp_ctrl_reg); + drm_dbg_kms(display->drm, + "[ENCODER:%d:%s] %s PP_STATUS: 0x%08x PP_CONTROL: 0x%08x\n", dig_port->base.base.base.id, dig_port->base.base.name, - pps_name(dev_priv, &intel_dp->pps), - intel_de_read(dev_priv, pp_stat_reg), - intel_de_read(dev_priv, pp_ctrl_reg)); + pps_name(intel_dp), + intel_de_read(display, pp_stat_reg), + intel_de_read(display, pp_ctrl_reg)); /* * If the panel wasn't on, delay before accessing aux channel */ if (!edp_have_panel_power(intel_dp)) { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "[ENCODER:%d:%s] %s panel power wasn't enabled\n", dig_port->base.base.base.id, dig_port->base.base.name, - pps_name(dev_priv, &intel_dp->pps)); + pps_name(intel_dp)); msleep(intel_dp->pps.panel_power_up_delay); } @@ -779,7 +799,8 @@ bool intel_pps_vdd_on_unlocked(struct intel_dp *intel_dp) */ void intel_pps_vdd_on(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *i915 = to_i915(display->drm); intel_wakeref_t wakeref; bool vdd; @@ -792,27 +813,27 @@ void intel_pps_vdd_on(struct intel_dp *intel_dp) I915_STATE_WARN(i915, !vdd, "[ENCODER:%d:%s] %s VDD already requested on\n", dp_to_dig_port(intel_dp)->base.base.base.id, dp_to_dig_port(intel_dp)->base.base.name, - pps_name(i915, &intel_dp->pps)); + pps_name(intel_dp)); } static void intel_pps_vdd_off_sync_unlocked(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); - struct intel_digital_port *dig_port = - dp_to_dig_port(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); + struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); u32 pp; i915_reg_t pp_stat_reg, pp_ctrl_reg; - lockdep_assert_held(&dev_priv->display.pps.mutex); + lockdep_assert_held(&display->pps.mutex); - drm_WARN_ON(&dev_priv->drm, intel_dp->pps.want_panel_vdd); + drm_WARN_ON(display->drm, intel_dp->pps.want_panel_vdd); if (!edp_have_panel_vdd(intel_dp)) return; - drm_dbg_kms(&dev_priv->drm, "[ENCODER:%d:%s] %s turning VDD off\n", + drm_dbg_kms(display->drm, "[ENCODER:%d:%s] %s turning VDD off\n", dig_port->base.base.base.id, dig_port->base.base.name, - pps_name(dev_priv, &intel_dp->pps)); + pps_name(intel_dp)); pp = ilk_get_pp_control(intel_dp); pp &= ~EDP_FORCE_VDD; @@ -820,15 +841,16 @@ static void intel_pps_vdd_off_sync_unlocked(struct intel_dp *intel_dp) pp_ctrl_reg = _pp_ctrl_reg(intel_dp); pp_stat_reg = _pp_stat_reg(intel_dp); - intel_de_write(dev_priv, pp_ctrl_reg, pp); - intel_de_posting_read(dev_priv, pp_ctrl_reg); + intel_de_write(display, pp_ctrl_reg, pp); + intel_de_posting_read(display, pp_ctrl_reg); /* Make sure sequencer is idle before allowing subsequent activity */ - drm_dbg_kms(&dev_priv->drm, "[ENCODER:%d:%s] %s PP_STATUS: 0x%08x PP_CONTROL: 0x%08x\n", + drm_dbg_kms(display->drm, + "[ENCODER:%d:%s] %s PP_STATUS: 0x%08x PP_CONTROL: 0x%08x\n", dig_port->base.base.base.id, dig_port->base.base.name, - pps_name(dev_priv, &intel_dp->pps), - intel_de_read(dev_priv, pp_stat_reg), - intel_de_read(dev_priv, pp_ctrl_reg)); + pps_name(intel_dp), + intel_de_read(display, pp_stat_reg), + intel_de_read(display, pp_ctrl_reg)); if ((pp & PANEL_POWER_ON) == 0) intel_dp->pps.panel_power_off_time = ktime_get_boottime(); @@ -869,7 +891,8 @@ static void edp_panel_vdd_work(struct work_struct *__work) static void edp_panel_vdd_schedule_off(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *i915 = to_i915(display->drm); unsigned long delay; /* @@ -896,9 +919,10 @@ static void edp_panel_vdd_schedule_off(struct intel_dp *intel_dp) */ void intel_pps_vdd_off_unlocked(struct intel_dp *intel_dp, bool sync) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); - lockdep_assert_held(&dev_priv->display.pps.mutex); + lockdep_assert_held(&display->pps.mutex); if (!intel_dp_is_edp(intel_dp)) return; @@ -907,7 +931,7 @@ void intel_pps_vdd_off_unlocked(struct intel_dp *intel_dp, bool sync) "[ENCODER:%d:%s] %s VDD not forced on", dp_to_dig_port(intel_dp)->base.base.base.id, dp_to_dig_port(intel_dp)->base.base.name, - pps_name(dev_priv, &intel_dp->pps)); + pps_name(intel_dp)); intel_dp->pps.want_panel_vdd = false; @@ -919,25 +943,26 @@ void intel_pps_vdd_off_unlocked(struct intel_dp *intel_dp, bool sync) void intel_pps_on_unlocked(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); u32 pp; i915_reg_t pp_ctrl_reg; - lockdep_assert_held(&dev_priv->display.pps.mutex); + lockdep_assert_held(&display->pps.mutex); if (!intel_dp_is_edp(intel_dp)) return; - drm_dbg_kms(&dev_priv->drm, "[ENCODER:%d:%s] %s turn panel power on\n", + drm_dbg_kms(display->drm, "[ENCODER:%d:%s] %s turn panel power on\n", dp_to_dig_port(intel_dp)->base.base.base.id, dp_to_dig_port(intel_dp)->base.base.name, - pps_name(dev_priv, &intel_dp->pps)); + pps_name(intel_dp)); - if (drm_WARN(&dev_priv->drm, edp_have_panel_power(intel_dp), + if (drm_WARN(display->drm, edp_have_panel_power(intel_dp), "[ENCODER:%d:%s] %s panel power already on\n", dp_to_dig_port(intel_dp)->base.base.base.id, dp_to_dig_port(intel_dp)->base.base.name, - pps_name(dev_priv, &intel_dp->pps))) + pps_name(intel_dp))) return; wait_panel_power_cycle(intel_dp); @@ -947,36 +972,36 @@ void intel_pps_on_unlocked(struct intel_dp *intel_dp) if (IS_IRONLAKE(dev_priv)) { /* ILK workaround: disable reset around power sequence */ pp &= ~PANEL_POWER_RESET; - intel_de_write(dev_priv, pp_ctrl_reg, pp); - intel_de_posting_read(dev_priv, pp_ctrl_reg); + intel_de_write(display, pp_ctrl_reg, pp); + intel_de_posting_read(display, pp_ctrl_reg); } /* * WA: 22019252566 * Disable DPLS gating around power sequence. */ - if (IS_DISPLAY_VER(dev_priv, 13, 14)) - intel_de_rmw(dev_priv, SOUTH_DSPCLK_GATE_D, + if (IS_DISPLAY_VER(display, 13, 14)) + intel_de_rmw(display, SOUTH_DSPCLK_GATE_D, 0, PCH_DPLSUNIT_CLOCK_GATE_DISABLE); pp |= PANEL_POWER_ON; if (!IS_IRONLAKE(dev_priv)) pp |= PANEL_POWER_RESET; - intel_de_write(dev_priv, pp_ctrl_reg, pp); - intel_de_posting_read(dev_priv, pp_ctrl_reg); + intel_de_write(display, pp_ctrl_reg, pp); + intel_de_posting_read(display, pp_ctrl_reg); wait_panel_on(intel_dp); intel_dp->pps.last_power_on = jiffies; - if (IS_DISPLAY_VER(dev_priv, 13, 14)) - intel_de_rmw(dev_priv, SOUTH_DSPCLK_GATE_D, + if (IS_DISPLAY_VER(display, 13, 14)) + intel_de_rmw(display, SOUTH_DSPCLK_GATE_D, PCH_DPLSUNIT_CLOCK_GATE_DISABLE, 0); if (IS_IRONLAKE(dev_priv)) { pp |= PANEL_POWER_RESET; /* restore panel reset bit */ - intel_de_write(dev_priv, pp_ctrl_reg, pp); - intel_de_posting_read(dev_priv, pp_ctrl_reg); + intel_de_write(display, pp_ctrl_reg, pp); + intel_de_posting_read(display, pp_ctrl_reg); } } @@ -993,24 +1018,25 @@ void intel_pps_on(struct intel_dp *intel_dp) void intel_pps_off_unlocked(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); u32 pp; i915_reg_t pp_ctrl_reg; - lockdep_assert_held(&dev_priv->display.pps.mutex); + lockdep_assert_held(&display->pps.mutex); if (!intel_dp_is_edp(intel_dp)) return; - drm_dbg_kms(&dev_priv->drm, "[ENCODER:%d:%s] %s turn panel power off\n", + drm_dbg_kms(display->drm, "[ENCODER:%d:%s] %s turn panel power off\n", dig_port->base.base.base.id, dig_port->base.base.name, - pps_name(dev_priv, &intel_dp->pps)); + pps_name(intel_dp)); - drm_WARN(&dev_priv->drm, !intel_dp->pps.want_panel_vdd, + drm_WARN(display->drm, !intel_dp->pps.want_panel_vdd, "[ENCODER:%d:%s] %s need VDD to turn off panel\n", dig_port->base.base.base.id, dig_port->base.base.name, - pps_name(dev_priv, &intel_dp->pps)); + pps_name(intel_dp)); pp = ilk_get_pp_control(intel_dp); /* We need to switch off panel power _and_ force vdd, for otherwise some @@ -1022,8 +1048,8 @@ void intel_pps_off_unlocked(struct intel_dp *intel_dp) intel_dp->pps.want_panel_vdd = false; - intel_de_write(dev_priv, pp_ctrl_reg, pp); - intel_de_posting_read(dev_priv, pp_ctrl_reg); + intel_de_write(display, pp_ctrl_reg, pp); + intel_de_posting_read(display, pp_ctrl_reg); wait_panel_off(intel_dp); intel_dp->pps.panel_power_off_time = ktime_get_boottime(); @@ -1048,7 +1074,7 @@ void intel_pps_off(struct intel_dp *intel_dp) /* Enable backlight in the panel power control. */ void intel_pps_backlight_on(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); intel_wakeref_t wakeref; /* @@ -1066,15 +1092,15 @@ void intel_pps_backlight_on(struct intel_dp *intel_dp) pp = ilk_get_pp_control(intel_dp); pp |= EDP_BLC_ENABLE; - intel_de_write(dev_priv, pp_ctrl_reg, pp); - intel_de_posting_read(dev_priv, pp_ctrl_reg); + intel_de_write(display, pp_ctrl_reg, pp); + intel_de_posting_read(display, pp_ctrl_reg); } } /* Disable backlight in the panel power control. */ void intel_pps_backlight_off(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); intel_wakeref_t wakeref; if (!intel_dp_is_edp(intel_dp)) @@ -1087,8 +1113,8 @@ void intel_pps_backlight_off(struct intel_dp *intel_dp) pp = ilk_get_pp_control(intel_dp); pp &= ~EDP_BLC_ENABLE; - intel_de_write(dev_priv, pp_ctrl_reg, pp); - intel_de_posting_read(dev_priv, pp_ctrl_reg); + intel_de_write(display, pp_ctrl_reg, pp); + intel_de_posting_read(display, pp_ctrl_reg); } intel_dp->pps.last_backlight_off = jiffies; @@ -1101,7 +1127,7 @@ void intel_pps_backlight_off(struct intel_dp *intel_dp) */ void intel_pps_backlight_power(struct intel_connector *connector, bool enable) { - struct drm_i915_private *i915 = to_i915(connector->base.dev); + struct intel_display *display = to_intel_display(connector); struct intel_dp *intel_dp = intel_attached_dp(connector); intel_wakeref_t wakeref; bool is_enabled; @@ -1112,7 +1138,7 @@ void intel_pps_backlight_power(struct intel_connector *connector, bool enable) if (is_enabled == enable) return; - drm_dbg_kms(&i915->drm, "panel power control backlight %s\n", + drm_dbg_kms(display->drm, "panel power control backlight %s\n", enable ? "enable" : "disable"); if (enable) @@ -1123,14 +1149,14 @@ void intel_pps_backlight_power(struct intel_connector *connector, bool enable) static void vlv_detach_power_sequencer(struct intel_dp *intel_dp) { + struct intel_display *display = to_intel_display(intel_dp); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); - struct drm_i915_private *dev_priv = to_i915(dig_port->base.base.dev); enum pipe pipe = intel_dp->pps.pps_pipe; - i915_reg_t pp_on_reg = PP_ON_DELAYS(dev_priv, pipe); + i915_reg_t pp_on_reg = PP_ON_DELAYS(display, pipe); - drm_WARN_ON(&dev_priv->drm, intel_dp->pps.active_pipe != INVALID_PIPE); + drm_WARN_ON(display->drm, intel_dp->pps.active_pipe != INVALID_PIPE); - if (drm_WARN_ON(&dev_priv->drm, pipe != PIPE_A && pipe != PIPE_B)) + if (drm_WARN_ON(display->drm, pipe != PIPE_A && pipe != PIPE_B)) return; intel_pps_vdd_off_sync_unlocked(intel_dp); @@ -1144,27 +1170,27 @@ static void vlv_detach_power_sequencer(struct intel_dp *intel_dp) * port select always when logically disconnecting a power sequencer * from a port. */ - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "detaching %s from [ENCODER:%d:%s]\n", - pps_name(dev_priv, &intel_dp->pps), + pps_name(intel_dp), dig_port->base.base.base.id, dig_port->base.base.name); - intel_de_write(dev_priv, pp_on_reg, 0); - intel_de_posting_read(dev_priv, pp_on_reg); + intel_de_write(display, pp_on_reg, 0); + intel_de_posting_read(display, pp_on_reg); intel_dp->pps.pps_pipe = INVALID_PIPE; } -static void vlv_steal_power_sequencer(struct drm_i915_private *dev_priv, +static void vlv_steal_power_sequencer(struct intel_display *display, enum pipe pipe) { struct intel_encoder *encoder; - lockdep_assert_held(&dev_priv->display.pps.mutex); + lockdep_assert_held(&display->pps.mutex); - for_each_intel_dp(&dev_priv->drm, encoder) { + for_each_intel_dp(display->drm, encoder) { struct intel_dp *intel_dp = enc_to_intel_dp(encoder); - drm_WARN(&dev_priv->drm, intel_dp->pps.active_pipe == pipe, + drm_WARN(display->drm, intel_dp->pps.active_pipe == pipe, "stealing PPS %c from active [ENCODER:%d:%s]\n", pipe_name(pipe), encoder->base.base.id, encoder->base.name); @@ -1172,7 +1198,7 @@ static void vlv_steal_power_sequencer(struct drm_i915_private *dev_priv, if (intel_dp->pps.pps_pipe != pipe) continue; - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "stealing PPS %c from [ENCODER:%d:%s]\n", pipe_name(pipe), encoder->base.base.id, encoder->base.name); @@ -1185,13 +1211,13 @@ static void vlv_steal_power_sequencer(struct drm_i915_private *dev_priv, void vlv_pps_init(struct intel_encoder *encoder, const struct intel_crtc_state *crtc_state) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); struct intel_dp *intel_dp = enc_to_intel_dp(encoder); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); - lockdep_assert_held(&dev_priv->display.pps.mutex); + lockdep_assert_held(&display->pps.mutex); - drm_WARN_ON(&dev_priv->drm, intel_dp->pps.active_pipe != INVALID_PIPE); + drm_WARN_ON(display->drm, intel_dp->pps.active_pipe != INVALID_PIPE); if (intel_dp->pps.pps_pipe != INVALID_PIPE && intel_dp->pps.pps_pipe != crtc->pipe) { @@ -1207,7 +1233,7 @@ void vlv_pps_init(struct intel_encoder *encoder, * We may be stealing the power * sequencer from another port. */ - vlv_steal_power_sequencer(dev_priv, crtc->pipe); + vlv_steal_power_sequencer(display, crtc->pipe); intel_dp->pps.active_pipe = crtc->pipe; @@ -1217,9 +1243,9 @@ void vlv_pps_init(struct intel_encoder *encoder, /* now it's all ours */ intel_dp->pps.pps_pipe = crtc->pipe; - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "initializing %s for [ENCODER:%d:%s]\n", - pps_name(dev_priv, &intel_dp->pps), + pps_name(intel_dp), encoder->base.base.id, encoder->base.name); /* init power sequencer on this pipe and port */ @@ -1229,10 +1255,11 @@ void vlv_pps_init(struct intel_encoder *encoder, static void pps_vdd_init(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); - lockdep_assert_held(&dev_priv->display.pps.mutex); + lockdep_assert_held(&display->pps.mutex); if (!edp_have_panel_vdd(intel_dp)) return; @@ -1243,11 +1270,11 @@ static void pps_vdd_init(struct intel_dp *intel_dp) * schedule a vdd off, so we don't hold on to the reference * indefinitely. */ - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "[ENCODER:%d:%s] %s VDD left on by BIOS, adjusting state tracking\n", dig_port->base.base.base.id, dig_port->base.base.name, - pps_name(dev_priv, &intel_dp->pps)); - drm_WARN_ON(&dev_priv->drm, intel_dp->pps.vdd_wakeref); + pps_name(intel_dp)); + drm_WARN_ON(display->drm, intel_dp->pps.vdd_wakeref); intel_dp->pps.vdd_wakeref = intel_display_power_get(dev_priv, intel_aux_power_domain(dig_port)); } @@ -1281,7 +1308,7 @@ static void pps_init_timestamps(struct intel_dp *intel_dp) static void intel_pps_readout_hw_state(struct intel_dp *intel_dp, struct edp_power_seq *seq) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); u32 pp_on, pp_off, pp_ctl; struct pps_registers regs; @@ -1290,11 +1317,11 @@ intel_pps_readout_hw_state(struct intel_dp *intel_dp, struct edp_power_seq *seq) pp_ctl = ilk_get_pp_control(intel_dp); /* Ensure PPS is unlocked */ - if (!HAS_DDI(dev_priv)) - intel_de_write(dev_priv, regs.pp_ctrl, pp_ctl); + if (!HAS_DDI(display)) + intel_de_write(display, regs.pp_ctrl, pp_ctl); - pp_on = intel_de_read(dev_priv, regs.pp_on); - pp_off = intel_de_read(dev_priv, regs.pp_off); + pp_on = intel_de_read(display, regs.pp_on); + pp_off = intel_de_read(display, regs.pp_off); /* Pull timing values out of registers */ seq->t1_t3 = REG_FIELD_GET(PANEL_POWER_UP_DELAY_MASK, pp_on); @@ -1305,7 +1332,7 @@ intel_pps_readout_hw_state(struct intel_dp *intel_dp, struct edp_power_seq *seq) if (i915_mmio_reg_valid(regs.pp_div)) { u32 pp_div; - pp_div = intel_de_read(dev_priv, regs.pp_div); + pp_div = intel_de_read(display, regs.pp_div); seq->t11_t12 = REG_FIELD_GET(PANEL_POWER_CYCLE_DELAY_MASK, pp_div) * 1000; } else { @@ -1317,9 +1344,10 @@ static void intel_pps_dump_state(struct intel_dp *intel_dp, const char *state_name, const struct edp_power_seq *seq) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); - drm_dbg_kms(&i915->drm, "%s t1_t3 %d t8 %d t9 %d t10 %d t11_t12 %d\n", + drm_dbg_kms(display->drm, + "%s t1_t3 %d t8 %d t9 %d t10 %d t11_t12 %d\n", state_name, seq->t1_t3, seq->t8, seq->t9, seq->t10, seq->t11_t12); } @@ -1327,7 +1355,7 @@ intel_pps_dump_state(struct intel_dp *intel_dp, const char *state_name, static void intel_pps_verify_state(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct edp_power_seq hw; struct edp_power_seq *sw = &intel_dp->pps.pps_delays; @@ -1335,7 +1363,7 @@ intel_pps_verify_state(struct intel_dp *intel_dp) if (hw.t1_t3 != sw->t1_t3 || hw.t8 != sw->t8 || hw.t9 != sw->t9 || hw.t10 != sw->t10 || hw.t11_t12 != sw->t11_t12) { - drm_err(&i915->drm, "PPS state mismatch\n"); + drm_err(display->drm, "PPS state mismatch\n"); intel_pps_dump_state(intel_dp, "sw", sw); intel_pps_dump_state(intel_dp, "hw", &hw); } @@ -1350,9 +1378,9 @@ static bool pps_delays_valid(struct edp_power_seq *delays) static void pps_init_delays_bios(struct intel_dp *intel_dp, struct edp_power_seq *bios) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); - lockdep_assert_held(&dev_priv->display.pps.mutex); + lockdep_assert_held(&display->pps.mutex); if (!pps_delays_valid(&intel_dp->pps.bios_pps_delays)) intel_pps_readout_hw_state(intel_dp, &intel_dp->pps.bios_pps_delays); @@ -1397,9 +1425,9 @@ static void pps_init_delays_vbt(struct intel_dp *intel_dp, static void pps_init_delays_spec(struct intel_dp *intel_dp, struct edp_power_seq *spec) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); - lockdep_assert_held(&dev_priv->display.pps.mutex); + lockdep_assert_held(&display->pps.mutex); /* Upper limits from eDP 1.3 spec. Note that we use the clunky units of * our hw here, which are all in 100usec. */ @@ -1418,11 +1446,11 @@ static void pps_init_delays_spec(struct intel_dp *intel_dp, static void pps_init_delays(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct edp_power_seq cur, vbt, spec, *final = &intel_dp->pps.pps_delays; - lockdep_assert_held(&dev_priv->display.pps.mutex); + lockdep_assert_held(&display->pps.mutex); /* already initialized? */ if (pps_delays_valid(final)) @@ -1452,13 +1480,13 @@ static void pps_init_delays(struct intel_dp *intel_dp) intel_dp->pps.panel_power_cycle_delay = get_delay(t11_t12); #undef get_delay - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "panel power up delay %d, power down delay %d, power cycle delay %d\n", intel_dp->pps.panel_power_up_delay, intel_dp->pps.panel_power_down_delay, intel_dp->pps.panel_power_cycle_delay); - drm_dbg_kms(&dev_priv->drm, "backlight on delay %d, off delay %d\n", + drm_dbg_kms(display->drm, "backlight on delay %d, off delay %d\n", intel_dp->pps.backlight_on_delay, intel_dp->pps.backlight_off_delay); @@ -1481,14 +1509,15 @@ static void pps_init_delays(struct intel_dp *intel_dp) static void pps_init_registers(struct intel_dp *intel_dp, bool force_disable_vdd) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); u32 pp_on, pp_off, port_sel = 0; - int div = DISPLAY_RUNTIME_INFO(dev_priv)->rawclk_freq / 1000; + int div = DISPLAY_RUNTIME_INFO(display)->rawclk_freq / 1000; struct pps_registers regs; enum port port = dp_to_dig_port(intel_dp)->base.port; const struct edp_power_seq *seq = &intel_dp->pps.pps_delays; - lockdep_assert_held(&dev_priv->display.pps.mutex); + lockdep_assert_held(&display->pps.mutex); intel_pps_get_registers(intel_dp, ®s); @@ -1507,16 +1536,16 @@ static void pps_init_registers(struct intel_dp *intel_dp, bool force_disable_vdd if (force_disable_vdd) { u32 pp = ilk_get_pp_control(intel_dp); - drm_WARN(&dev_priv->drm, pp & PANEL_POWER_ON, + drm_WARN(display->drm, pp & PANEL_POWER_ON, "Panel power already on\n"); if (pp & EDP_FORCE_VDD) - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "VDD already on, disabling first\n"); pp &= ~EDP_FORCE_VDD; - intel_de_write(dev_priv, regs.pp_ctrl, pp); + intel_de_write(display, regs.pp_ctrl, pp); } pp_on = REG_FIELD_PREP(PANEL_POWER_UP_DELAY_MASK, seq->t1_t3) | @@ -1547,32 +1576,33 @@ static void pps_init_registers(struct intel_dp *intel_dp, bool force_disable_vdd pp_on |= port_sel; - intel_de_write(dev_priv, regs.pp_on, pp_on); - intel_de_write(dev_priv, regs.pp_off, pp_off); + intel_de_write(display, regs.pp_on, pp_on); + intel_de_write(display, regs.pp_off, pp_off); /* * Compute the divisor for the pp clock, simply match the Bspec formula. */ if (i915_mmio_reg_valid(regs.pp_div)) - intel_de_write(dev_priv, regs.pp_div, + intel_de_write(display, regs.pp_div, REG_FIELD_PREP(PP_REFERENCE_DIVIDER_MASK, (100 * div) / 2 - 1) | REG_FIELD_PREP(PANEL_POWER_CYCLE_DELAY_MASK, DIV_ROUND_UP(seq->t11_t12, 1000))); else - intel_de_rmw(dev_priv, regs.pp_ctrl, BXT_POWER_CYCLE_DELAY_MASK, + intel_de_rmw(display, regs.pp_ctrl, BXT_POWER_CYCLE_DELAY_MASK, REG_FIELD_PREP(BXT_POWER_CYCLE_DELAY_MASK, DIV_ROUND_UP(seq->t11_t12, 1000))); - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "panel power sequencer register settings: PP_ON %#x, PP_OFF %#x, PP_DIV %#x\n", - intel_de_read(dev_priv, regs.pp_on), - intel_de_read(dev_priv, regs.pp_off), + intel_de_read(display, regs.pp_on), + intel_de_read(display, regs.pp_off), i915_mmio_reg_valid(regs.pp_div) ? - intel_de_read(dev_priv, regs.pp_div) : - (intel_de_read(dev_priv, regs.pp_ctrl) & BXT_POWER_CYCLE_DELAY_MASK)); + intel_de_read(display, regs.pp_div) : + (intel_de_read(display, regs.pp_ctrl) & BXT_POWER_CYCLE_DELAY_MASK)); } void intel_pps_encoder_reset(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *i915 = to_i915(display->drm); intel_wakeref_t wakeref; if (!intel_dp_is_edp(intel_dp)) @@ -1618,17 +1648,19 @@ bool intel_pps_init(struct intel_dp *intel_dp) static void pps_init_late(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *i915 = to_i915(display->drm); struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base; struct intel_connector *connector = intel_dp->attached_connector; if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915)) return; - if (intel_num_pps(i915) < 2) + if (intel_num_pps(display) < 2) return; - drm_WARN(&i915->drm, connector->panel.vbt.backlight.controller >= 0 && + drm_WARN(display->drm, + connector->panel.vbt.backlight.controller >= 0 && intel_dp->pps.pps_idx != connector->panel.vbt.backlight.controller, "[ENCODER:%d:%s] power sequencer mismatch: %d (initial) vs. %d (VBT)\n", encoder->base.base.id, encoder->base.name, @@ -1657,32 +1689,34 @@ void intel_pps_init_late(struct intel_dp *intel_dp) } } -void intel_pps_unlock_regs_wa(struct drm_i915_private *dev_priv) +void intel_pps_unlock_regs_wa(struct intel_display *display) { int pps_num; int pps_idx; - if (!HAS_DISPLAY(dev_priv) || HAS_DDI(dev_priv)) + if (!HAS_DISPLAY(display) || HAS_DDI(display)) return; /* * This w/a is needed at least on CPT/PPT, but to be sure apply it * everywhere where registers can be write protected. */ - pps_num = intel_num_pps(dev_priv); + pps_num = intel_num_pps(display); for (pps_idx = 0; pps_idx < pps_num; pps_idx++) - intel_de_rmw(dev_priv, PP_CONTROL(dev_priv, pps_idx), + intel_de_rmw(display, PP_CONTROL(display, pps_idx), PANEL_UNLOCK_MASK, PANEL_UNLOCK_REGS); } -void intel_pps_setup(struct drm_i915_private *i915) +void intel_pps_setup(struct intel_display *display) { + struct drm_i915_private *i915 = to_i915(display->drm); + if (HAS_PCH_SPLIT(i915) || IS_GEMINILAKE(i915) || IS_BROXTON(i915)) - i915->display.pps.mmio_base = PCH_PPS_BASE; + display->pps.mmio_base = PCH_PPS_BASE; else if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915)) - i915->display.pps.mmio_base = VLV_PPS_BASE; + display->pps.mmio_base = VLV_PPS_BASE; else - i915->display.pps.mmio_base = PPS_BASE; + display->pps.mmio_base = PPS_BASE; } static int intel_pps_show(struct seq_file *m, void *data) @@ -1716,21 +1750,23 @@ void intel_pps_connector_debugfs_add(struct intel_connector *connector) connector, &intel_pps_fops); } -void assert_pps_unlocked(struct drm_i915_private *dev_priv, enum pipe pipe) +void assert_pps_unlocked(struct intel_display *display, enum pipe pipe) { + struct drm_i915_private *dev_priv = to_i915(display->drm); i915_reg_t pp_reg; u32 val; enum pipe panel_pipe = INVALID_PIPE; bool locked = true; - if (drm_WARN_ON(&dev_priv->drm, HAS_DDI(dev_priv))) + if (drm_WARN_ON(display->drm, HAS_DDI(display))) return; if (HAS_PCH_SPLIT(dev_priv)) { u32 port_sel; - pp_reg = PP_CONTROL(dev_priv, 0); - port_sel = intel_de_read(dev_priv, PP_ON_DELAYS(dev_priv, 0)) & PANEL_PORT_SELECT_MASK; + pp_reg = PP_CONTROL(display, 0); + port_sel = intel_de_read(display, PP_ON_DELAYS(display, 0)) & + PANEL_PORT_SELECT_MASK; switch (port_sel) { case PANEL_PORT_SELECT_LVDS: @@ -1751,20 +1787,21 @@ void assert_pps_unlocked(struct drm_i915_private *dev_priv, enum pipe pipe) } } else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) { /* presumably write lock depends on pipe, not port select */ - pp_reg = PP_CONTROL(dev_priv, pipe); + pp_reg = PP_CONTROL(display, pipe); panel_pipe = pipe; } else { u32 port_sel; - pp_reg = PP_CONTROL(dev_priv, 0); - port_sel = intel_de_read(dev_priv, PP_ON_DELAYS(dev_priv, 0)) & PANEL_PORT_SELECT_MASK; + pp_reg = PP_CONTROL(display, 0); + port_sel = intel_de_read(display, PP_ON_DELAYS(display, 0)) & + PANEL_PORT_SELECT_MASK; - drm_WARN_ON(&dev_priv->drm, + drm_WARN_ON(display->drm, port_sel != PANEL_PORT_SELECT_LVDS); intel_lvds_port_enabled(dev_priv, LVDS, &panel_pipe); } - val = intel_de_read(dev_priv, pp_reg); + val = intel_de_read(display, pp_reg); if (!(val & PANEL_POWER_ON) || ((val & PANEL_UNLOCK_MASK) == PANEL_UNLOCK_REGS)) locked = false; diff --git a/drivers/gpu/drm/i915/display/intel_pps.h b/drivers/gpu/drm/i915/display/intel_pps.h index 07ef96ca8da2..0c5da83a559e 100644 --- a/drivers/gpu/drm/i915/display/intel_pps.h +++ b/drivers/gpu/drm/i915/display/intel_pps.h @@ -11,9 +11,9 @@ #include "intel_wakeref.h" enum pipe; -struct drm_i915_private; struct intel_connector; struct intel_crtc_state; +struct intel_display; struct intel_dp; struct intel_encoder; @@ -43,16 +43,16 @@ void intel_pps_wait_power_cycle(struct intel_dp *intel_dp); bool intel_pps_init(struct intel_dp *intel_dp); void intel_pps_init_late(struct intel_dp *intel_dp); void intel_pps_encoder_reset(struct intel_dp *intel_dp); -void intel_pps_reset_all(struct drm_i915_private *i915); +void intel_pps_reset_all(struct intel_display *display); void vlv_pps_init(struct intel_encoder *encoder, const struct intel_crtc_state *crtc_state); -void intel_pps_unlock_regs_wa(struct drm_i915_private *i915); -void intel_pps_setup(struct drm_i915_private *i915); +void intel_pps_unlock_regs_wa(struct intel_display *display); +void intel_pps_setup(struct intel_display *display); void intel_pps_connector_debugfs_add(struct intel_connector *connector); -void assert_pps_unlocked(struct drm_i915_private *i915, enum pipe pipe); +void assert_pps_unlocked(struct intel_display *display, enum pipe pipe); #endif /* __INTEL_PPS_H__ */ diff --git a/drivers/gpu/drm/i915/display/intel_psr.c b/drivers/gpu/drm/i915/display/intel_psr.c index 257526362b39..1f83b3b67ea6 100644 --- a/drivers/gpu/drm/i915/display/intel_psr.c +++ b/drivers/gpu/drm/i915/display/intel_psr.c @@ -205,14 +205,14 @@ bool intel_encoder_can_psr(struct intel_encoder *encoder) static bool psr_global_enabled(struct intel_dp *intel_dp) { + struct intel_display *display = to_intel_display(intel_dp); struct intel_connector *connector = intel_dp->attached_connector; - struct drm_i915_private *i915 = dp_to_i915(intel_dp); switch (intel_dp->psr.debug & I915_PSR_DEBUG_MODE_MASK) { case I915_PSR_DEBUG_DEFAULT: - if (i915->display.params.enable_psr == -1) + if (display->params.enable_psr == -1) return connector->panel.vbt.psr.enable; - return i915->display.params.enable_psr; + return display->params.enable_psr; case I915_PSR_DEBUG_DISABLE: return false; default: @@ -222,14 +222,14 @@ static bool psr_global_enabled(struct intel_dp *intel_dp) static bool psr2_global_enabled(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); switch (intel_dp->psr.debug & I915_PSR_DEBUG_MODE_MASK) { case I915_PSR_DEBUG_DISABLE: case I915_PSR_DEBUG_FORCE_PSR1: return false; default: - if (i915->display.params.enable_psr == 1) + if (display->params.enable_psr == 1) return false; return true; } @@ -237,9 +237,9 @@ static bool psr2_global_enabled(struct intel_dp *intel_dp) static bool psr2_su_region_et_global_enabled(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); - if (i915->display.params.enable_psr != -1) + if (display->params.enable_psr != -1) return false; return true; @@ -247,9 +247,9 @@ static bool psr2_su_region_et_global_enabled(struct intel_dp *intel_dp) static bool panel_replay_global_enabled(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); - if ((i915->display.params.enable_psr != -1) || + if ((display->params.enable_psr != -1) || (intel_dp->psr.debug & I915_PSR_DEBUG_PANEL_REPLAY_DISABLE)) return false; return true; @@ -257,111 +257,111 @@ static bool panel_replay_global_enabled(struct intel_dp *intel_dp) static u32 psr_irq_psr_error_bit_get(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); - return DISPLAY_VER(dev_priv) >= 12 ? TGL_PSR_ERROR : + return DISPLAY_VER(display) >= 12 ? TGL_PSR_ERROR : EDP_PSR_ERROR(intel_dp->psr.transcoder); } static u32 psr_irq_post_exit_bit_get(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); - return DISPLAY_VER(dev_priv) >= 12 ? TGL_PSR_POST_EXIT : + return DISPLAY_VER(display) >= 12 ? TGL_PSR_POST_EXIT : EDP_PSR_POST_EXIT(intel_dp->psr.transcoder); } static u32 psr_irq_pre_entry_bit_get(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); - return DISPLAY_VER(dev_priv) >= 12 ? TGL_PSR_PRE_ENTRY : + return DISPLAY_VER(display) >= 12 ? TGL_PSR_PRE_ENTRY : EDP_PSR_PRE_ENTRY(intel_dp->psr.transcoder); } static u32 psr_irq_mask_get(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); - return DISPLAY_VER(dev_priv) >= 12 ? TGL_PSR_MASK : + return DISPLAY_VER(display) >= 12 ? TGL_PSR_MASK : EDP_PSR_MASK(intel_dp->psr.transcoder); } -static i915_reg_t psr_ctl_reg(struct drm_i915_private *dev_priv, +static i915_reg_t psr_ctl_reg(struct intel_display *display, enum transcoder cpu_transcoder) { - if (DISPLAY_VER(dev_priv) >= 8) - return EDP_PSR_CTL(dev_priv, cpu_transcoder); + if (DISPLAY_VER(display) >= 8) + return EDP_PSR_CTL(display, cpu_transcoder); else return HSW_SRD_CTL; } -static i915_reg_t psr_debug_reg(struct drm_i915_private *dev_priv, +static i915_reg_t psr_debug_reg(struct intel_display *display, enum transcoder cpu_transcoder) { - if (DISPLAY_VER(dev_priv) >= 8) - return EDP_PSR_DEBUG(dev_priv, cpu_transcoder); + if (DISPLAY_VER(display) >= 8) + return EDP_PSR_DEBUG(display, cpu_transcoder); else return HSW_SRD_DEBUG; } -static i915_reg_t psr_perf_cnt_reg(struct drm_i915_private *dev_priv, +static i915_reg_t psr_perf_cnt_reg(struct intel_display *display, enum transcoder cpu_transcoder) { - if (DISPLAY_VER(dev_priv) >= 8) - return EDP_PSR_PERF_CNT(dev_priv, cpu_transcoder); + if (DISPLAY_VER(display) >= 8) + return EDP_PSR_PERF_CNT(display, cpu_transcoder); else return HSW_SRD_PERF_CNT; } -static i915_reg_t psr_status_reg(struct drm_i915_private *dev_priv, +static i915_reg_t psr_status_reg(struct intel_display *display, enum transcoder cpu_transcoder) { - if (DISPLAY_VER(dev_priv) >= 8) - return EDP_PSR_STATUS(dev_priv, cpu_transcoder); + if (DISPLAY_VER(display) >= 8) + return EDP_PSR_STATUS(display, cpu_transcoder); else return HSW_SRD_STATUS; } -static i915_reg_t psr_imr_reg(struct drm_i915_private *dev_priv, +static i915_reg_t psr_imr_reg(struct intel_display *display, enum transcoder cpu_transcoder) { - if (DISPLAY_VER(dev_priv) >= 12) - return TRANS_PSR_IMR(dev_priv, cpu_transcoder); + if (DISPLAY_VER(display) >= 12) + return TRANS_PSR_IMR(display, cpu_transcoder); else return EDP_PSR_IMR; } -static i915_reg_t psr_iir_reg(struct drm_i915_private *dev_priv, +static i915_reg_t psr_iir_reg(struct intel_display *display, enum transcoder cpu_transcoder) { - if (DISPLAY_VER(dev_priv) >= 12) - return TRANS_PSR_IIR(dev_priv, cpu_transcoder); + if (DISPLAY_VER(display) >= 12) + return TRANS_PSR_IIR(display, cpu_transcoder); else return EDP_PSR_IIR; } -static i915_reg_t psr_aux_ctl_reg(struct drm_i915_private *dev_priv, +static i915_reg_t psr_aux_ctl_reg(struct intel_display *display, enum transcoder cpu_transcoder) { - if (DISPLAY_VER(dev_priv) >= 8) - return EDP_PSR_AUX_CTL(dev_priv, cpu_transcoder); + if (DISPLAY_VER(display) >= 8) + return EDP_PSR_AUX_CTL(display, cpu_transcoder); else return HSW_SRD_AUX_CTL; } -static i915_reg_t psr_aux_data_reg(struct drm_i915_private *dev_priv, +static i915_reg_t psr_aux_data_reg(struct intel_display *display, enum transcoder cpu_transcoder, int i) { - if (DISPLAY_VER(dev_priv) >= 8) - return EDP_PSR_AUX_DATA(dev_priv, cpu_transcoder, i); + if (DISPLAY_VER(display) >= 8) + return EDP_PSR_AUX_DATA(display, cpu_transcoder, i); else return HSW_SRD_AUX_DATA(i); } static void psr_irq_control(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); enum transcoder cpu_transcoder = intel_dp->psr.transcoder; u32 mask; @@ -373,80 +373,81 @@ static void psr_irq_control(struct intel_dp *intel_dp) mask |= psr_irq_post_exit_bit_get(intel_dp) | psr_irq_pre_entry_bit_get(intel_dp); - intel_de_rmw(dev_priv, psr_imr_reg(dev_priv, cpu_transcoder), + intel_de_rmw(display, psr_imr_reg(display, cpu_transcoder), psr_irq_mask_get(intel_dp), ~mask); } -static void psr_event_print(struct drm_i915_private *i915, +static void psr_event_print(struct intel_display *display, u32 val, bool sel_update_enabled) { - drm_dbg_kms(&i915->drm, "PSR exit events: 0x%x\n", val); + drm_dbg_kms(display->drm, "PSR exit events: 0x%x\n", val); if (val & PSR_EVENT_PSR2_WD_TIMER_EXPIRE) - drm_dbg_kms(&i915->drm, "\tPSR2 watchdog timer expired\n"); + drm_dbg_kms(display->drm, "\tPSR2 watchdog timer expired\n"); if ((val & PSR_EVENT_PSR2_DISABLED) && sel_update_enabled) - drm_dbg_kms(&i915->drm, "\tPSR2 disabled\n"); + drm_dbg_kms(display->drm, "\tPSR2 disabled\n"); if (val & PSR_EVENT_SU_DIRTY_FIFO_UNDERRUN) - drm_dbg_kms(&i915->drm, "\tSU dirty FIFO underrun\n"); + drm_dbg_kms(display->drm, "\tSU dirty FIFO underrun\n"); if (val & PSR_EVENT_SU_CRC_FIFO_UNDERRUN) - drm_dbg_kms(&i915->drm, "\tSU CRC FIFO underrun\n"); + drm_dbg_kms(display->drm, "\tSU CRC FIFO underrun\n"); if (val & PSR_EVENT_GRAPHICS_RESET) - drm_dbg_kms(&i915->drm, "\tGraphics reset\n"); + drm_dbg_kms(display->drm, "\tGraphics reset\n"); if (val & PSR_EVENT_PCH_INTERRUPT) - drm_dbg_kms(&i915->drm, "\tPCH interrupt\n"); + drm_dbg_kms(display->drm, "\tPCH interrupt\n"); if (val & PSR_EVENT_MEMORY_UP) - drm_dbg_kms(&i915->drm, "\tMemory up\n"); + drm_dbg_kms(display->drm, "\tMemory up\n"); if (val & PSR_EVENT_FRONT_BUFFER_MODIFY) - drm_dbg_kms(&i915->drm, "\tFront buffer modification\n"); + drm_dbg_kms(display->drm, "\tFront buffer modification\n"); if (val & PSR_EVENT_WD_TIMER_EXPIRE) - drm_dbg_kms(&i915->drm, "\tPSR watchdog timer expired\n"); + drm_dbg_kms(display->drm, "\tPSR watchdog timer expired\n"); if (val & PSR_EVENT_PIPE_REGISTERS_UPDATE) - drm_dbg_kms(&i915->drm, "\tPIPE registers updated\n"); + drm_dbg_kms(display->drm, "\tPIPE registers updated\n"); if (val & PSR_EVENT_REGISTER_UPDATE) - drm_dbg_kms(&i915->drm, "\tRegister updated\n"); + drm_dbg_kms(display->drm, "\tRegister updated\n"); if (val & PSR_EVENT_HDCP_ENABLE) - drm_dbg_kms(&i915->drm, "\tHDCP enabled\n"); + drm_dbg_kms(display->drm, "\tHDCP enabled\n"); if (val & PSR_EVENT_KVMR_SESSION_ENABLE) - drm_dbg_kms(&i915->drm, "\tKVMR session enabled\n"); + drm_dbg_kms(display->drm, "\tKVMR session enabled\n"); if (val & PSR_EVENT_VBI_ENABLE) - drm_dbg_kms(&i915->drm, "\tVBI enabled\n"); + drm_dbg_kms(display->drm, "\tVBI enabled\n"); if (val & PSR_EVENT_LPSP_MODE_EXIT) - drm_dbg_kms(&i915->drm, "\tLPSP mode exited\n"); + drm_dbg_kms(display->drm, "\tLPSP mode exited\n"); if ((val & PSR_EVENT_PSR_DISABLE) && !sel_update_enabled) - drm_dbg_kms(&i915->drm, "\tPSR disabled\n"); + drm_dbg_kms(display->drm, "\tPSR disabled\n"); } void intel_psr_irq_handler(struct intel_dp *intel_dp, u32 psr_iir) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); enum transcoder cpu_transcoder = intel_dp->psr.transcoder; ktime_t time_ns = ktime_get(); if (psr_iir & psr_irq_pre_entry_bit_get(intel_dp)) { intel_dp->psr.last_entry_attempt = time_ns; - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "[transcoder %s] PSR entry attempt in 2 vblanks\n", transcoder_name(cpu_transcoder)); } if (psr_iir & psr_irq_post_exit_bit_get(intel_dp)) { intel_dp->psr.last_exit = time_ns; - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "[transcoder %s] PSR exit completed\n", transcoder_name(cpu_transcoder)); - if (DISPLAY_VER(dev_priv) >= 9) { + if (DISPLAY_VER(display) >= 9) { u32 val; val = intel_de_rmw(dev_priv, PSR_EVENT(dev_priv, cpu_transcoder), 0, 0); - psr_event_print(dev_priv, val, intel_dp->psr.sel_update_enabled); + psr_event_print(display, val, intel_dp->psr.sel_update_enabled); } } if (psr_iir & psr_irq_psr_error_bit_get(intel_dp)) { - drm_warn(&dev_priv->drm, "[transcoder %s] PSR aux error\n", + drm_warn(display->drm, "[transcoder %s] PSR aux error\n", transcoder_name(cpu_transcoder)); intel_dp->psr.irq_aux_error = true; @@ -459,7 +460,7 @@ void intel_psr_irq_handler(struct intel_dp *intel_dp, u32 psr_iir) * again so we don't care about unmask the interruption * or unset irq_aux_error. */ - intel_de_rmw(dev_priv, psr_imr_reg(dev_priv, cpu_transcoder), + intel_de_rmw(display, psr_imr_reg(display, cpu_transcoder), 0, psr_irq_psr_error_bit_get(intel_dp)); queue_work(dev_priv->unordered_wq, &intel_dp->psr.work); @@ -468,14 +469,14 @@ void intel_psr_irq_handler(struct intel_dp *intel_dp, u32 psr_iir) static u8 intel_dp_get_sink_sync_latency(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); u8 val = 8; /* assume the worst if we can't read the value */ if (drm_dp_dpcd_readb(&intel_dp->aux, DP_SYNCHRONIZATION_LATENCY_IN_SINK, &val) == 1) val &= DP_MAX_RESYNC_FRAME_COUNT_MASK; else - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "Unable to get sink synchronization latency, assuming 8 frames\n"); return val; } @@ -516,7 +517,7 @@ intel_dp_get_su_y_granularity_offset(struct intel_dp *intel_dp) */ static void intel_dp_get_su_granularity(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); ssize_t r; u16 w; u8 y; @@ -542,7 +543,7 @@ static void intel_dp_get_su_granularity(struct intel_dp *intel_dp) intel_dp_get_su_x_granularity_offset(intel_dp), &w, 2); if (r != 2) - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "Unable to read selective update x granularity\n"); /* * Spec says that if the value read is 0 the default granularity should @@ -555,7 +556,7 @@ static void intel_dp_get_su_granularity(struct intel_dp *intel_dp) intel_dp_get_su_y_granularity_offset(intel_dp), &y, 1); if (r != 1) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "Unable to read selective update y granularity\n"); y = 4; } @@ -569,17 +570,17 @@ exit: static void _panel_replay_init_dpcd(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); if (intel_dp_is_edp(intel_dp)) { if (!intel_alpm_aux_less_wake_supported(intel_dp)) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "Panel doesn't support AUX-less ALPM, eDP Panel Replay not possible\n"); return; } if (!(intel_dp->pr_dpcd & DP_PANEL_REPLAY_EARLY_TRANSPORT_SUPPORT)) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "Panel doesn't support early transport, eDP Panel Replay not possible\n"); return; } @@ -590,7 +591,7 @@ static void _panel_replay_init_dpcd(struct intel_dp *intel_dp) if (intel_dp->pr_dpcd & DP_PANEL_REPLAY_SU_SUPPORT) intel_dp->psr.sink_panel_replay_su_support = true; - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "Panel replay %sis supported by panel\n", intel_dp->psr.sink_panel_replay_su_support ? "selective_update " : ""); @@ -598,20 +599,19 @@ static void _panel_replay_init_dpcd(struct intel_dp *intel_dp) static void _psr_init_dpcd(struct intel_dp *intel_dp) { - struct drm_i915_private *i915 = - to_i915(dp_to_dig_port(intel_dp)->base.base.dev); + struct intel_display *display = to_intel_display(intel_dp); - drm_dbg_kms(&i915->drm, "eDP panel supports PSR version %x\n", + drm_dbg_kms(display->drm, "eDP panel supports PSR version %x\n", intel_dp->psr_dpcd[0]); if (drm_dp_has_quirk(&intel_dp->desc, DP_DPCD_QUIRK_NO_PSR)) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "PSR support not currently available for this panel\n"); return; } if (!(intel_dp->edp_dpcd[1] & DP_EDP_SET_POWER_CAP)) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "Panel lacks power state control, PSR cannot be enabled\n"); return; } @@ -620,7 +620,7 @@ static void _psr_init_dpcd(struct intel_dp *intel_dp) intel_dp->psr.sink_sync_latency = intel_dp_get_sink_sync_latency(intel_dp); - if (DISPLAY_VER(i915) >= 9 && + if (DISPLAY_VER(display) >= 9 && intel_dp->psr_dpcd[0] >= DP_PSR2_WITH_Y_COORD_IS_SUPPORTED) { bool y_req = intel_dp->psr_dpcd[1] & DP_PSR2_SU_Y_COORDINATE_REQUIRED; @@ -638,7 +638,7 @@ static void _psr_init_dpcd(struct intel_dp *intel_dp) */ intel_dp->psr.sink_psr2_support = y_req && intel_alpm_aux_wake_supported(intel_dp); - drm_dbg_kms(&i915->drm, "PSR2 %ssupported\n", + drm_dbg_kms(display->drm, "PSR2 %ssupported\n", intel_dp->psr.sink_psr2_support ? "" : "not "); } } @@ -663,7 +663,8 @@ void intel_psr_init_dpcd(struct intel_dp *intel_dp) static void hsw_psr_setup_aux(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); enum transcoder cpu_transcoder = intel_dp->psr.transcoder; u32 aux_clock_divider, aux_ctl; /* write DP_SET_POWER=D0 */ @@ -679,7 +680,7 @@ static void hsw_psr_setup_aux(struct intel_dp *intel_dp) BUILD_BUG_ON(sizeof(aux_msg) > 20); for (i = 0; i < sizeof(aux_msg); i += 4) intel_de_write(dev_priv, - psr_aux_data_reg(dev_priv, cpu_transcoder, i >> 2), + psr_aux_data_reg(display, cpu_transcoder, i >> 2), intel_dp_aux_pack(&aux_msg[i], sizeof(aux_msg) - i)); aux_clock_divider = intel_dp->get_aux_clock_divider(intel_dp, 0); @@ -694,15 +695,15 @@ static void hsw_psr_setup_aux(struct intel_dp *intel_dp) EDP_PSR_AUX_CTL_PRECHARGE_2US_MASK | EDP_PSR_AUX_CTL_BIT_CLOCK_2X_MASK; - intel_de_write(dev_priv, psr_aux_ctl_reg(dev_priv, cpu_transcoder), + intel_de_write(display, psr_aux_ctl_reg(display, cpu_transcoder), aux_ctl); } static bool psr2_su_region_et_valid(struct intel_dp *intel_dp, bool panel_replay) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); - if (DISPLAY_VER(i915) < 20 || !intel_dp_is_edp(intel_dp) || + if (DISPLAY_VER(display) < 20 || !intel_dp_is_edp(intel_dp) || intel_dp->psr.debug & I915_PSR_DEBUG_SU_REGION_ET_DISABLE) return false; @@ -741,7 +742,7 @@ static void _panel_replay_enable_sink(struct intel_dp *intel_dp, static void _psr_enable_sink(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); u8 val = DP_PSR_ENABLE; if (crtc_state->has_sel_update) { @@ -750,7 +751,7 @@ static void _psr_enable_sink(struct intel_dp *intel_dp, if (intel_dp->psr.link_standby) val |= DP_PSR_MAIN_LINK_ACTIVE; - if (DISPLAY_VER(i915) >= 8) + if (DISPLAY_VER(display) >= 8) val |= DP_PSR_CRC_VERIFICATION; } @@ -802,14 +803,15 @@ void intel_psr_enable_sink(struct intel_dp *intel_dp, static u32 intel_psr1_get_tp_time(struct intel_dp *intel_dp) { + struct intel_display *display = to_intel_display(intel_dp); struct intel_connector *connector = intel_dp->attached_connector; - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); u32 val = 0; - if (DISPLAY_VER(dev_priv) >= 11) + if (DISPLAY_VER(display) >= 11) val |= EDP_PSR_TP4_TIME_0us; - if (dev_priv->display.params.psr_safest_params) { + if (display->params.psr_safest_params) { val |= EDP_PSR_TP1_TIME_2500us; val |= EDP_PSR_TP2_TP3_TIME_2500us; goto check_tp3_sel; @@ -854,8 +856,8 @@ check_tp3_sel: static u8 psr_compute_idle_frames(struct intel_dp *intel_dp) { + struct intel_display *display = to_intel_display(intel_dp); struct intel_connector *connector = intel_dp->attached_connector; - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); int idle_frames; /* Let's use 6 as the minimum to cover all known cases including the @@ -864,7 +866,7 @@ static u8 psr_compute_idle_frames(struct intel_dp *intel_dp) idle_frames = max(6, connector->panel.vbt.psr.idle_frames); idle_frames = max(idle_frames, intel_dp->psr.sink_sync_latency + 1); - if (drm_WARN_ON(&dev_priv->drm, idle_frames > 0xf)) + if (drm_WARN_ON(display->drm, idle_frames > 0xf)) idle_frames = 0xf; return idle_frames; @@ -872,14 +874,15 @@ static u8 psr_compute_idle_frames(struct intel_dp *intel_dp) static void hsw_activate_psr1(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); enum transcoder cpu_transcoder = intel_dp->psr.transcoder; u32 max_sleep_time = 0x1f; u32 val = EDP_PSR_ENABLE; val |= EDP_PSR_IDLE_FRAMES(psr_compute_idle_frames(intel_dp)); - if (DISPLAY_VER(dev_priv) < 20) + if (DISPLAY_VER(display) < 20) val |= EDP_PSR_MAX_SLEEP_TIME(max_sleep_time); if (IS_HASWELL(dev_priv)) @@ -890,23 +893,23 @@ static void hsw_activate_psr1(struct intel_dp *intel_dp) val |= intel_psr1_get_tp_time(intel_dp); - if (DISPLAY_VER(dev_priv) >= 8) + if (DISPLAY_VER(display) >= 8) val |= EDP_PSR_CRC_ENABLE; - if (DISPLAY_VER(dev_priv) >= 20) + if (DISPLAY_VER(display) >= 20) val |= LNL_EDP_PSR_ENTRY_SETUP_FRAMES(intel_dp->psr.entry_setup_frames); - intel_de_rmw(dev_priv, psr_ctl_reg(dev_priv, cpu_transcoder), + intel_de_rmw(display, psr_ctl_reg(display, cpu_transcoder), ~EDP_PSR_RESTORE_PSR_ACTIVE_CTX_MASK, val); } static u32 intel_psr2_get_tp_time(struct intel_dp *intel_dp) { + struct intel_display *display = to_intel_display(intel_dp); struct intel_connector *connector = intel_dp->attached_connector; - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); u32 val = 0; - if (dev_priv->display.params.psr_safest_params) + if (display->params.psr_safest_params) return EDP_PSR2_TP2_TIME_2500us; if (connector->panel.vbt.psr.psr2_tp2_tp3_wakeup_time_us >= 0 && @@ -950,7 +953,7 @@ static u8 frames_before_su_entry(struct intel_dp *intel_dp) static void dg2_activate_panel_replay(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct intel_psr *psr = &intel_dp->psr; enum transcoder cpu_transcoder = intel_dp->psr.transcoder; @@ -961,38 +964,39 @@ static void dg2_activate_panel_replay(struct intel_dp *intel_dp) if (intel_dp->psr.req_psr2_sdp_prior_scanline) val |= EDP_PSR2_SU_SDP_SCANLINE; - intel_de_write(dev_priv, EDP_PSR2_CTL(dev_priv, cpu_transcoder), + intel_de_write(display, EDP_PSR2_CTL(display, cpu_transcoder), val); } - intel_de_rmw(dev_priv, - PSR2_MAN_TRK_CTL(dev_priv, intel_dp->psr.transcoder), + intel_de_rmw(display, + PSR2_MAN_TRK_CTL(display, intel_dp->psr.transcoder), 0, ADLP_PSR2_MAN_TRK_CTL_SF_CONTINUOS_FULL_FRAME); - intel_de_rmw(dev_priv, TRANS_DP2_CTL(intel_dp->psr.transcoder), 0, + intel_de_rmw(display, TRANS_DP2_CTL(intel_dp->psr.transcoder), 0, TRANS_DP2_PANEL_REPLAY_ENABLE); } static void hsw_activate_psr2(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); enum transcoder cpu_transcoder = intel_dp->psr.transcoder; u32 val = EDP_PSR2_ENABLE; u32 psr_val = 0; val |= EDP_PSR2_IDLE_FRAMES(psr_compute_idle_frames(intel_dp)); - if (DISPLAY_VER(dev_priv) < 14 && !IS_ALDERLAKE_P(dev_priv)) + if (DISPLAY_VER(display) < 14 && !IS_ALDERLAKE_P(dev_priv)) val |= EDP_SU_TRACK_ENABLE; - if (DISPLAY_VER(dev_priv) >= 10 && DISPLAY_VER(dev_priv) < 13) + if (DISPLAY_VER(display) >= 10 && DISPLAY_VER(display) < 13) val |= EDP_Y_COORDINATE_ENABLE; val |= EDP_PSR2_FRAME_BEFORE_SU(frames_before_su_entry(intel_dp)); val |= intel_psr2_get_tp_time(intel_dp); - if (DISPLAY_VER(dev_priv) >= 12 && DISPLAY_VER(dev_priv) < 20) { + if (DISPLAY_VER(display) >= 12 && DISPLAY_VER(display) < 20) { if (psr2_block_count(intel_dp) > 2) val |= TGL_EDP_PSR2_BLOCK_COUNT_NUM_3; else @@ -1000,7 +1004,7 @@ static void hsw_activate_psr2(struct intel_dp *intel_dp) } /* Wa_22012278275:adl-p */ - if (IS_ALDERLAKE_P(dev_priv) && IS_DISPLAY_STEP(dev_priv, STEP_A0, STEP_E0)) { + if (IS_ALDERLAKE_P(dev_priv) && IS_DISPLAY_STEP(display, STEP_A0, STEP_E0)) { static const u8 map[] = { 2, /* 5 lines */ 1, /* 6 lines */ @@ -1023,12 +1027,12 @@ static void hsw_activate_psr2(struct intel_dp *intel_dp) tmp = map[intel_dp->alpm_parameters.fast_wake_lines - TGL_EDP_PSR2_FAST_WAKE_MIN_LINES]; val |= TGL_EDP_PSR2_FAST_WAKE(tmp + TGL_EDP_PSR2_FAST_WAKE_MIN_LINES); - } else if (DISPLAY_VER(dev_priv) >= 20) { + } else if (DISPLAY_VER(display) >= 20) { val |= LNL_EDP_PSR2_IO_BUFFER_WAKE(intel_dp->alpm_parameters.io_wake_lines); - } else if (DISPLAY_VER(dev_priv) >= 12) { + } else if (DISPLAY_VER(display) >= 12) { val |= TGL_EDP_PSR2_IO_BUFFER_WAKE(intel_dp->alpm_parameters.io_wake_lines); val |= TGL_EDP_PSR2_FAST_WAKE(intel_dp->alpm_parameters.fast_wake_lines); - } else if (DISPLAY_VER(dev_priv) >= 9) { + } else if (DISPLAY_VER(display) >= 9) { val |= EDP_PSR2_IO_BUFFER_WAKE(intel_dp->alpm_parameters.io_wake_lines); val |= EDP_PSR2_FAST_WAKE(intel_dp->alpm_parameters.fast_wake_lines); } @@ -1036,18 +1040,18 @@ static void hsw_activate_psr2(struct intel_dp *intel_dp) if (intel_dp->psr.req_psr2_sdp_prior_scanline) val |= EDP_PSR2_SU_SDP_SCANLINE; - if (DISPLAY_VER(dev_priv) >= 20) + if (DISPLAY_VER(display) >= 20) psr_val |= LNL_EDP_PSR_ENTRY_SETUP_FRAMES(intel_dp->psr.entry_setup_frames); if (intel_dp->psr.psr2_sel_fetch_enabled) { u32 tmp; - tmp = intel_de_read(dev_priv, - PSR2_MAN_TRK_CTL(dev_priv, cpu_transcoder)); - drm_WARN_ON(&dev_priv->drm, !(tmp & PSR2_MAN_TRK_CTL_ENABLE)); - } else if (HAS_PSR2_SEL_FETCH(dev_priv)) { - intel_de_write(dev_priv, - PSR2_MAN_TRK_CTL(dev_priv, cpu_transcoder), 0); + tmp = intel_de_read(display, + PSR2_MAN_TRK_CTL(display, cpu_transcoder)); + drm_WARN_ON(display->drm, !(tmp & PSR2_MAN_TRK_CTL_ENABLE)); + } else if (HAS_PSR2_SEL_FETCH(display)) { + intel_de_write(display, + PSR2_MAN_TRK_CTL(display, cpu_transcoder), 0); } if (intel_dp->psr.su_region_et_enabled) @@ -1057,19 +1061,21 @@ static void hsw_activate_psr2(struct intel_dp *intel_dp) * PSR2 HW is incorrectly using EDP_PSR_TP1_TP3_SEL and BSpec is * recommending keep this bit unset while PSR2 is enabled. */ - intel_de_write(dev_priv, psr_ctl_reg(dev_priv, cpu_transcoder), psr_val); + intel_de_write(display, psr_ctl_reg(display, cpu_transcoder), psr_val); - intel_de_write(dev_priv, EDP_PSR2_CTL(dev_priv, cpu_transcoder), val); + intel_de_write(display, EDP_PSR2_CTL(display, cpu_transcoder), val); } static bool -transcoder_has_psr2(struct drm_i915_private *dev_priv, enum transcoder cpu_transcoder) +transcoder_has_psr2(struct intel_display *display, enum transcoder cpu_transcoder) { - if (IS_ALDERLAKE_P(dev_priv) || DISPLAY_VER(dev_priv) >= 14) + struct drm_i915_private *dev_priv = to_i915(display->drm); + + if (IS_ALDERLAKE_P(dev_priv) || DISPLAY_VER(display) >= 14) return cpu_transcoder == TRANSCODER_A || cpu_transcoder == TRANSCODER_B; - else if (DISPLAY_VER(dev_priv) >= 12) + else if (DISPLAY_VER(display) >= 12) return cpu_transcoder == TRANSCODER_A; - else if (DISPLAY_VER(dev_priv) >= 9) + else if (DISPLAY_VER(display) >= 9) return cpu_transcoder == TRANSCODER_EDP; else return false; @@ -1087,17 +1093,18 @@ static u32 intel_get_frame_time_us(const struct intel_crtc_state *crtc_state) static void psr2_program_idle_frames(struct intel_dp *intel_dp, u32 idle_frames) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); enum transcoder cpu_transcoder = intel_dp->psr.transcoder; - intel_de_rmw(dev_priv, EDP_PSR2_CTL(dev_priv, cpu_transcoder), + intel_de_rmw(display, EDP_PSR2_CTL(display, cpu_transcoder), EDP_PSR2_IDLE_FRAMES_MASK, EDP_PSR2_IDLE_FRAMES(idle_frames)); } static void tgl_psr2_enable_dc3co(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); psr2_program_idle_frames(intel_dp, 0); intel_display_power_set_target_dc_state(dev_priv, DC_STATE_EN_DC3CO); @@ -1105,7 +1112,8 @@ static void tgl_psr2_enable_dc3co(struct intel_dp *intel_dp) static void tgl_psr2_disable_dc3co(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); intel_display_power_set_target_dc_state(dev_priv, DC_STATE_EN_UPTO_DC6); psr2_program_idle_frames(intel_dp, psr_compute_idle_frames(intel_dp)); @@ -1140,12 +1148,13 @@ static bool dc3co_is_pipe_port_compatible(struct intel_dp *intel_dp, struct intel_crtc_state *crtc_state) { + struct intel_display *display = to_intel_display(intel_dp); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); enum pipe pipe = to_intel_crtc(crtc_state->uapi.crtc)->pipe; - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); enum port port = dig_port->base.port; - if (IS_ALDERLAKE_P(dev_priv) || DISPLAY_VER(dev_priv) >= 14) + if (IS_ALDERLAKE_P(dev_priv) || DISPLAY_VER(display) >= 14) return pipe <= PIPE_B && port <= PORT_B; else return pipe == PIPE_A && port == PORT_A; @@ -1155,9 +1164,10 @@ static void tgl_dc3co_exitline_compute_config(struct intel_dp *intel_dp, struct intel_crtc_state *crtc_state) { + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); const u32 crtc_vdisplay = crtc_state->uapi.adjusted_mode.crtc_vdisplay; - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); - struct i915_power_domains *power_domains = &dev_priv->display.power.domains; + struct i915_power_domains *power_domains = &display->power.domains; u32 exit_scanlines; /* @@ -1181,7 +1191,7 @@ tgl_dc3co_exitline_compute_config(struct intel_dp *intel_dp, return; /* Wa_16011303918:adl-p */ - if (IS_ALDERLAKE_P(dev_priv) && IS_DISPLAY_STEP(dev_priv, STEP_A0, STEP_B0)) + if (IS_ALDERLAKE_P(dev_priv) && IS_DISPLAY_STEP(display, STEP_A0, STEP_B0)) return; /* @@ -1191,7 +1201,7 @@ tgl_dc3co_exitline_compute_config(struct intel_dp *intel_dp, exit_scanlines = intel_usecs_to_scanlines(&crtc_state->uapi.adjusted_mode, 200) + 1; - if (drm_WARN_ON(&dev_priv->drm, exit_scanlines > crtc_vdisplay)) + if (drm_WARN_ON(display->drm, exit_scanlines > crtc_vdisplay)) return; crtc_state->dc3co_exitline = crtc_vdisplay - exit_scanlines; @@ -1200,17 +1210,17 @@ tgl_dc3co_exitline_compute_config(struct intel_dp *intel_dp, static bool intel_psr2_sel_fetch_config_valid(struct intel_dp *intel_dp, struct intel_crtc_state *crtc_state) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); - if (!dev_priv->display.params.enable_psr2_sel_fetch && + if (!display->params.enable_psr2_sel_fetch && intel_dp->psr.debug != I915_PSR_DEBUG_ENABLE_SEL_FETCH) { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "PSR2 sel fetch not enabled, disabled by parameter\n"); return false; } if (crtc_state->uapi.async_flip) { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "PSR2 sel fetch not enabled, async flip enabled\n"); return false; } @@ -1221,7 +1231,8 @@ static bool intel_psr2_sel_fetch_config_valid(struct intel_dp *intel_dp, static bool psr2_granularity_check(struct intel_dp *intel_dp, struct intel_crtc_state *crtc_state) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); const struct drm_dsc_config *vdsc_cfg = &crtc_state->dsc.config; const int crtc_hdisplay = crtc_state->hw.adjusted_mode.crtc_hdisplay; const int crtc_vdisplay = crtc_state->hw.adjusted_mode.crtc_vdisplay; @@ -1243,7 +1254,7 @@ static bool psr2_granularity_check(struct intel_dp *intel_dp, * For other platforms with SW tracking we can adjust the y coordinates * to match sink requirement if multiple of 4. */ - if (IS_ALDERLAKE_P(dev_priv) || DISPLAY_VER(dev_priv) >= 14) + if (IS_ALDERLAKE_P(dev_priv) || DISPLAY_VER(display) >= 14) y_granularity = intel_dp->psr.su_y_granularity; else if (intel_dp->psr.su_y_granularity <= 2) y_granularity = 4; @@ -1264,8 +1275,8 @@ static bool psr2_granularity_check(struct intel_dp *intel_dp, static bool _compute_psr2_sdp_prior_scanline_indication(struct intel_dp *intel_dp, struct intel_crtc_state *crtc_state) { + struct intel_display *display = to_intel_display(intel_dp); const struct drm_display_mode *adjusted_mode = &crtc_state->uapi.adjusted_mode; - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); u32 hblank_total, hblank_ns, req_ns; hblank_total = adjusted_mode->crtc_hblank_end - adjusted_mode->crtc_hblank_start; @@ -1278,7 +1289,7 @@ static bool _compute_psr2_sdp_prior_scanline_indication(struct intel_dp *intel_d return true; /* Not supported <13 / Wa_22012279113:adl-p */ - if (DISPLAY_VER(dev_priv) < 14 || intel_dp->edp_dpcd[0] < DP_EDP_14b) + if (DISPLAY_VER(display) < 14 || intel_dp->edp_dpcd[0] < DP_EDP_14b) return false; crtc_state->req_psr2_sdp_prior_scanline = true; @@ -1288,12 +1299,12 @@ static bool _compute_psr2_sdp_prior_scanline_indication(struct intel_dp *intel_d static int intel_psr_entry_setup_frames(struct intel_dp *intel_dp, const struct drm_display_mode *adjusted_mode) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); int psr_setup_time = drm_dp_psr_setup_time(intel_dp->psr_dpcd); int entry_setup_frames = 0; if (psr_setup_time < 0) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "PSR condition failed: Invalid PSR setup time (0x%02x)\n", intel_dp->psr_dpcd[1]); return -ETIME; @@ -1301,14 +1312,14 @@ static int intel_psr_entry_setup_frames(struct intel_dp *intel_dp, if (intel_usecs_to_scanlines(adjusted_mode, psr_setup_time) > adjusted_mode->crtc_vtotal - adjusted_mode->crtc_vdisplay - 1) { - if (DISPLAY_VER(i915) >= 20) { + if (DISPLAY_VER(display) >= 20) { /* setup entry frames can be up to 3 frames */ entry_setup_frames = 1; - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "PSR setup entry frames %d\n", entry_setup_frames); } else { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "PSR condition failed: PSR setup time (%d us) too long\n", psr_setup_time); return -ETIME; @@ -1322,7 +1333,7 @@ static bool wake_lines_fit_into_vblank(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state, bool aux_less) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); int vblank = crtc_state->hw.adjusted_mode.crtc_vblank_end - crtc_state->hw.adjusted_mode.crtc_vblank_start; int wake_lines; @@ -1330,7 +1341,7 @@ static bool wake_lines_fit_into_vblank(struct intel_dp *intel_dp, if (aux_less) wake_lines = intel_dp->alpm_parameters.aux_less_wake_lines; else - wake_lines = DISPLAY_VER(i915) < 20 ? + wake_lines = DISPLAY_VER(display) < 20 ? psr2_block_count_lines(intel_dp) : intel_dp->alpm_parameters.io_wake_lines; @@ -1348,16 +1359,16 @@ static bool alpm_config_valid(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state, bool aux_less) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); if (!intel_alpm_compute_params(intel_dp, crtc_state)) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "PSR2/Panel Replay not enabled, Unable to use long enough wake times\n"); return false; } if (!wake_lines_fit_into_vblank(intel_dp, crtc_state, aux_less)) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "PSR2/Panel Replay not enabled, too short vblank time\n"); return false; } @@ -1368,7 +1379,8 @@ static bool alpm_config_valid(struct intel_dp *intel_dp, static bool intel_psr2_config_valid(struct intel_dp *intel_dp, struct intel_crtc_state *crtc_state) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); int crtc_hdisplay = crtc_state->hw.adjusted_mode.crtc_hdisplay; int crtc_vdisplay = crtc_state->hw.adjusted_mode.crtc_vdisplay; int psr_max_h = 0, psr_max_v = 0, max_bpp = 0; @@ -1378,24 +1390,26 @@ static bool intel_psr2_config_valid(struct intel_dp *intel_dp, /* JSL and EHL only supports eDP 1.3 */ if (IS_JASPERLAKE(dev_priv) || IS_ELKHARTLAKE(dev_priv)) { - drm_dbg_kms(&dev_priv->drm, "PSR2 not supported by phy\n"); + drm_dbg_kms(display->drm, "PSR2 not supported by phy\n"); return false; } /* Wa_16011181250 */ if (IS_ROCKETLAKE(dev_priv) || IS_ALDERLAKE_S(dev_priv) || IS_DG2(dev_priv)) { - drm_dbg_kms(&dev_priv->drm, "PSR2 is defeatured for this platform\n"); + drm_dbg_kms(display->drm, + "PSR2 is defeatured for this platform\n"); return false; } - if (IS_ALDERLAKE_P(dev_priv) && IS_DISPLAY_STEP(dev_priv, STEP_A0, STEP_B0)) { - drm_dbg_kms(&dev_priv->drm, "PSR2 not completely functional in this stepping\n"); + if (IS_ALDERLAKE_P(dev_priv) && IS_DISPLAY_STEP(display, STEP_A0, STEP_B0)) { + drm_dbg_kms(display->drm, + "PSR2 not completely functional in this stepping\n"); return false; } - if (!transcoder_has_psr2(dev_priv, crtc_state->cpu_transcoder)) { - drm_dbg_kms(&dev_priv->drm, + if (!transcoder_has_psr2(display, crtc_state->cpu_transcoder)) { + drm_dbg_kms(display->drm, "PSR2 not supported in transcoder %s\n", transcoder_name(crtc_state->cpu_transcoder)); return false; @@ -1407,28 +1421,28 @@ static bool intel_psr2_config_valid(struct intel_dp *intel_dp, * over PSR2. */ if (crtc_state->dsc.compression_enable && - (DISPLAY_VER(dev_priv) < 14 && !IS_ALDERLAKE_P(dev_priv))) { - drm_dbg_kms(&dev_priv->drm, + (DISPLAY_VER(display) < 14 && !IS_ALDERLAKE_P(dev_priv))) { + drm_dbg_kms(display->drm, "PSR2 cannot be enabled since DSC is enabled\n"); return false; } - if (DISPLAY_VER(dev_priv) >= 12) { + if (DISPLAY_VER(display) >= 12) { psr_max_h = 5120; psr_max_v = 3200; max_bpp = 30; - } else if (DISPLAY_VER(dev_priv) >= 10) { + } else if (DISPLAY_VER(display) >= 10) { psr_max_h = 4096; psr_max_v = 2304; max_bpp = 24; - } else if (DISPLAY_VER(dev_priv) == 9) { + } else if (DISPLAY_VER(display) == 9) { psr_max_h = 3640; psr_max_v = 2304; max_bpp = 24; } if (crtc_state->pipe_bpp > max_bpp) { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "PSR2 not enabled, pipe bpp %d > max supported %d\n", crtc_state->pipe_bpp, max_bpp); return false; @@ -1436,8 +1450,8 @@ static bool intel_psr2_config_valid(struct intel_dp *intel_dp, /* Wa_16011303918:adl-p */ if (crtc_state->vrr.enable && - IS_ALDERLAKE_P(dev_priv) && IS_DISPLAY_STEP(dev_priv, STEP_A0, STEP_B0)) { - drm_dbg_kms(&dev_priv->drm, + IS_ALDERLAKE_P(dev_priv) && IS_DISPLAY_STEP(display, STEP_A0, STEP_B0)) { + drm_dbg_kms(display->drm, "PSR2 not enabled, not compatible with HW stepping + VRR\n"); return false; } @@ -1447,7 +1461,7 @@ static bool intel_psr2_config_valid(struct intel_dp *intel_dp, if (!crtc_state->enable_psr2_sel_fetch && (crtc_hdisplay > psr_max_h || crtc_vdisplay > psr_max_v)) { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "PSR2 not enabled, resolution %dx%d > max supported %dx%d\n", crtc_hdisplay, crtc_vdisplay, psr_max_h, psr_max_v); @@ -1462,18 +1476,19 @@ static bool intel_psr2_config_valid(struct intel_dp *intel_dp, static bool intel_sel_update_config_valid(struct intel_dp *intel_dp, struct intel_crtc_state *crtc_state) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); - if (HAS_PSR2_SEL_FETCH(dev_priv) && + if (HAS_PSR2_SEL_FETCH(display) && !intel_psr2_sel_fetch_config_valid(intel_dp, crtc_state) && - !HAS_PSR_HW_TRACKING(dev_priv)) { - drm_dbg_kms(&dev_priv->drm, + !HAS_PSR_HW_TRACKING(display)) { + drm_dbg_kms(display->drm, "Selective update not enabled, selective fetch not valid and no HW tracking available\n"); goto unsupported; } if (!psr2_global_enabled(intel_dp)) { - drm_dbg_kms(&dev_priv->drm, "Selective update disabled by flag\n"); + drm_dbg_kms(display->drm, + "Selective update disabled by flag\n"); goto unsupported; } @@ -1481,23 +1496,23 @@ static bool intel_sel_update_config_valid(struct intel_dp *intel_dp, goto unsupported; if (!_compute_psr2_sdp_prior_scanline_indication(intel_dp, crtc_state)) { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "Selective update not enabled, SDP indication do not fit in hblank\n"); goto unsupported; } - if (crtc_state->has_panel_replay && (DISPLAY_VER(dev_priv) < 14 || + if (crtc_state->has_panel_replay && (DISPLAY_VER(display) < 14 || !intel_dp->psr.sink_panel_replay_su_support)) goto unsupported; if (crtc_state->crc_enabled) { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "Selective update not enabled because it would inhibit pipe CRC calculation\n"); goto unsupported; } if (!psr2_granularity_check(intel_dp, crtc_state)) { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "Selective update not enabled, SU granularity not compatible\n"); goto unsupported; } @@ -1515,7 +1530,7 @@ unsupported: static bool _psr_compute_config(struct intel_dp *intel_dp, struct intel_crtc_state *crtc_state) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); const struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode; int entry_setup_frames; @@ -1534,7 +1549,7 @@ static bool _psr_compute_config(struct intel_dp *intel_dp, if (entry_setup_frames >= 0) { intel_dp->psr.entry_setup_frames = entry_setup_frames; } else { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "PSR condition failed: PSR setup timing not met\n"); return false; } @@ -1547,7 +1562,7 @@ _panel_replay_compute_config(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state, const struct drm_connector_state *conn_state) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct intel_connector *connector = to_intel_connector(conn_state->connector); struct intel_hdcp *hdcp = &connector->hdcp; @@ -1556,7 +1571,7 @@ _panel_replay_compute_config(struct intel_dp *intel_dp, return false; if (!panel_replay_global_enabled(intel_dp)) { - drm_dbg_kms(&i915->drm, "Panel Replay disabled by flag\n"); + drm_dbg_kms(display->drm, "Panel Replay disabled by flag\n"); return false; } @@ -1567,7 +1582,7 @@ _panel_replay_compute_config(struct intel_dp *intel_dp, /* 128b/132b Panel Replay is not supported on eDP */ if (intel_dp_is_uhbr(crtc_state)) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "Panel Replay is not supported with 128b/132b\n"); return false; } @@ -1578,7 +1593,7 @@ _panel_replay_compute_config(struct intel_dp *intel_dp, (conn_state->content_protection == DRM_MODE_CONTENT_PROTECTION_ENABLED && hdcp->value == DRM_MODE_CONTENT_PROTECTION_UNDESIRED)) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "Panel Replay is not supported with HDCP\n"); return false; } @@ -1587,7 +1602,7 @@ _panel_replay_compute_config(struct intel_dp *intel_dp, return false; if (crtc_state->crc_enabled) { - drm_dbg_kms(&i915->drm, + drm_dbg_kms(display->drm, "Panel Replay not enabled because it would inhibit pipe CRC calculation\n"); return false; } @@ -1599,22 +1614,22 @@ void intel_psr_compute_config(struct intel_dp *intel_dp, struct intel_crtc_state *crtc_state, struct drm_connector_state *conn_state) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); const struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode; if (!psr_global_enabled(intel_dp)) { - drm_dbg_kms(&dev_priv->drm, "PSR disabled by flag\n"); + drm_dbg_kms(display->drm, "PSR disabled by flag\n"); return; } if (intel_dp->psr.sink_not_reliable) { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "PSR sink implementation is not reliable\n"); return; } if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "PSR condition failed: Interlaced mode enabled\n"); return; } @@ -1625,7 +1640,7 @@ void intel_psr_compute_config(struct intel_dp *intel_dp, * PSR is a transcoder level feature. */ if (crtc_state->joiner_pipes) { - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "PSR disabled due to joiner\n"); return; } @@ -1646,7 +1661,7 @@ void intel_psr_compute_config(struct intel_dp *intel_dp, void intel_psr_get_config(struct intel_encoder *encoder, struct intel_crtc_state *pipe_config) { - struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_display *display = to_intel_display(encoder); struct intel_digital_port *dig_port = enc_to_dig_port(encoder); enum transcoder cpu_transcoder = pipe_config->cpu_transcoder; struct intel_dp *intel_dp; @@ -1679,18 +1694,18 @@ void intel_psr_get_config(struct intel_encoder *encoder, if (!intel_dp->psr.sel_update_enabled) goto unlock; - if (HAS_PSR2_SEL_FETCH(dev_priv)) { - val = intel_de_read(dev_priv, - PSR2_MAN_TRK_CTL(dev_priv, cpu_transcoder)); + if (HAS_PSR2_SEL_FETCH(display)) { + val = intel_de_read(display, + PSR2_MAN_TRK_CTL(display, cpu_transcoder)); if (val & PSR2_MAN_TRK_CTL_ENABLE) pipe_config->enable_psr2_sel_fetch = true; } pipe_config->enable_psr2_su_region_et = intel_dp->psr.su_region_et_enabled; - if (DISPLAY_VER(dev_priv) >= 12) { - val = intel_de_read(dev_priv, - TRANS_EXITLINE(dev_priv, cpu_transcoder)); + if (DISPLAY_VER(display) >= 12) { + val = intel_de_read(display, + TRANS_EXITLINE(display, cpu_transcoder)); pipe_config->dc3co_exitline = REG_FIELD_GET(EXITLINE_MASK, val); } unlock: @@ -1699,17 +1714,17 @@ unlock: static void intel_psr_activate(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); enum transcoder cpu_transcoder = intel_dp->psr.transcoder; - drm_WARN_ON(&dev_priv->drm, - transcoder_has_psr2(dev_priv, cpu_transcoder) && - intel_de_read(dev_priv, EDP_PSR2_CTL(dev_priv, cpu_transcoder)) & EDP_PSR2_ENABLE); + drm_WARN_ON(display->drm, + transcoder_has_psr2(display, cpu_transcoder) && + intel_de_read(display, EDP_PSR2_CTL(display, cpu_transcoder)) & EDP_PSR2_ENABLE); - drm_WARN_ON(&dev_priv->drm, - intel_de_read(dev_priv, psr_ctl_reg(dev_priv, cpu_transcoder)) & EDP_PSR_ENABLE); + drm_WARN_ON(display->drm, + intel_de_read(display, psr_ctl_reg(display, cpu_transcoder)) & EDP_PSR_ENABLE); - drm_WARN_ON(&dev_priv->drm, intel_dp->psr.active); + drm_WARN_ON(display->drm, intel_dp->psr.active); lockdep_assert_held(&intel_dp->psr.lock); @@ -1748,30 +1763,31 @@ static u32 wa_16013835468_bit_get(struct intel_dp *intel_dp) static void wm_optimization_wa(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); bool set_wa_bit = false; /* Wa_14015648006 */ - if (IS_DISPLAY_VER(dev_priv, 11, 14)) + if (IS_DISPLAY_VER(display, 11, 14)) set_wa_bit |= crtc_state->wm_level_disabled; /* Wa_16013835468 */ - if (DISPLAY_VER(dev_priv) == 12) + if (DISPLAY_VER(display) == 12) set_wa_bit |= crtc_state->hw.adjusted_mode.crtc_vblank_start != crtc_state->hw.adjusted_mode.crtc_vdisplay; if (set_wa_bit) - intel_de_rmw(dev_priv, GEN8_CHICKEN_DCPR_1, + intel_de_rmw(display, GEN8_CHICKEN_DCPR_1, 0, wa_16013835468_bit_get(intel_dp)); else - intel_de_rmw(dev_priv, GEN8_CHICKEN_DCPR_1, + intel_de_rmw(display, GEN8_CHICKEN_DCPR_1, wa_16013835468_bit_get(intel_dp), 0); } static void intel_psr_enable_source(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); enum transcoder cpu_transcoder = intel_dp->psr.transcoder; u32 mask = 0; @@ -1779,7 +1795,7 @@ static void intel_psr_enable_source(struct intel_dp *intel_dp, * Only HSW and BDW have PSR AUX registers that need to be setup. * SKL+ use hardcoded values PSR AUX transactions */ - if (DISPLAY_VER(dev_priv) < 9) + if (DISPLAY_VER(display) < 9) hsw_psr_setup_aux(intel_dp); /* @@ -1796,7 +1812,7 @@ static void intel_psr_enable_source(struct intel_dp *intel_dp, * Panel Replay on DP: No bits are applicable * Panel Replay on eDP: All bits are applicable */ - if (DISPLAY_VER(dev_priv) < 20 || intel_dp_is_edp(intel_dp)) + if (DISPLAY_VER(display) < 20 || intel_dp_is_edp(intel_dp)) mask = EDP_PSR_DEBUG_MASK_HPD; if (intel_dp_is_edp(intel_dp)) { @@ -1810,17 +1826,17 @@ static void intel_psr_enable_source(struct intel_dp *intel_dp, * As a workaround leave LPSP unmasked to prevent PSR entry * when external displays are active. */ - if (DISPLAY_VER(dev_priv) >= 8 || IS_HASWELL_ULT(dev_priv)) + if (DISPLAY_VER(display) >= 8 || IS_HASWELL_ULT(dev_priv)) mask |= EDP_PSR_DEBUG_MASK_LPSP; - if (DISPLAY_VER(dev_priv) < 20) + if (DISPLAY_VER(display) < 20) mask |= EDP_PSR_DEBUG_MASK_MAX_SLEEP; /* * No separate pipe reg write mask on hsw/bdw, so have to unmask all * registers in order to keep the CURSURFLIVE tricks working :( */ - if (IS_DISPLAY_VER(dev_priv, 9, 10)) + if (IS_DISPLAY_VER(display, 9, 10)) mask |= EDP_PSR_DEBUG_MASK_DISP_REG_WRITE; /* allow PSR with sprite enabled */ @@ -1828,7 +1844,7 @@ static void intel_psr_enable_source(struct intel_dp *intel_dp, mask |= EDP_PSR_DEBUG_MASK_SPRITE_ENABLE; } - intel_de_write(dev_priv, psr_debug_reg(dev_priv, cpu_transcoder), mask); + intel_de_write(display, psr_debug_reg(display, cpu_transcoder), mask); psr_irq_control(intel_dp); @@ -1837,13 +1853,13 @@ static void intel_psr_enable_source(struct intel_dp *intel_dp, * transcoder, EXITLINE will need to be unset when disabling PSR */ if (intel_dp->psr.dc3co_exitline) - intel_de_rmw(dev_priv, - TRANS_EXITLINE(dev_priv, cpu_transcoder), + intel_de_rmw(display, + TRANS_EXITLINE(display, cpu_transcoder), EXITLINE_MASK, intel_dp->psr.dc3co_exitline << EXITLINE_SHIFT | EXITLINE_ENABLE); - if (HAS_PSR_HW_TRACKING(dev_priv) && HAS_PSR2_SEL_FETCH(dev_priv)) - intel_de_rmw(dev_priv, CHICKEN_PAR1_1, IGNORE_PSR2_HW_TRACKING, + if (HAS_PSR_HW_TRACKING(display) && HAS_PSR2_SEL_FETCH(display)) + intel_de_rmw(display, CHICKEN_PAR1_1, IGNORE_PSR2_HW_TRACKING, intel_dp->psr.psr2_sel_fetch_enabled ? IGNORE_PSR2_HW_TRACKING : 0); @@ -1857,8 +1873,8 @@ static void intel_psr_enable_source(struct intel_dp *intel_dp, wm_optimization_wa(intel_dp, crtc_state); if (intel_dp->psr.sel_update_enabled) { - if (DISPLAY_VER(dev_priv) == 9) - intel_de_rmw(dev_priv, CHICKEN_TRANS(cpu_transcoder), 0, + if (DISPLAY_VER(display) == 9) + intel_de_rmw(display, CHICKEN_TRANS(cpu_transcoder), 0, PSR2_VSC_ENABLE_PROG_HEADER | PSR2_ADD_VERTICAL_LINE_COUNT); @@ -1868,27 +1884,27 @@ static void intel_psr_enable_source(struct intel_dp *intel_dp, * cause issues if non-supported panels are used. */ if (!intel_dp->psr.panel_replay_enabled && - (IS_DISPLAY_VER_STEP(dev_priv, IP_VER(14, 0), STEP_A0, STEP_B0) || + (IS_DISPLAY_VER_STEP(display, IP_VER(14, 0), STEP_A0, STEP_B0) || IS_ALDERLAKE_P(dev_priv))) - intel_de_rmw(dev_priv, hsw_chicken_trans_reg(dev_priv, cpu_transcoder), + intel_de_rmw(display, hsw_chicken_trans_reg(dev_priv, cpu_transcoder), 0, ADLP_1_BASED_X_GRANULARITY); /* Wa_16012604467:adlp,mtl[a0,b0] */ if (!intel_dp->psr.panel_replay_enabled && - IS_DISPLAY_VER_STEP(dev_priv, IP_VER(14, 0), STEP_A0, STEP_B0)) - intel_de_rmw(dev_priv, - MTL_CLKGATE_DIS_TRANS(dev_priv, cpu_transcoder), + IS_DISPLAY_VER_STEP(display, IP_VER(14, 0), STEP_A0, STEP_B0)) + intel_de_rmw(display, + MTL_CLKGATE_DIS_TRANS(display, cpu_transcoder), 0, MTL_CLKGATE_DIS_TRANS_DMASC_GATING_DIS); else if (IS_ALDERLAKE_P(dev_priv)) - intel_de_rmw(dev_priv, CLKGATE_DIS_MISC, 0, + intel_de_rmw(display, CLKGATE_DIS_MISC, 0, CLKGATE_DIS_MISC_DMASC_GATING_DIS); } } static bool psr_interrupt_error_check(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); enum transcoder cpu_transcoder = intel_dp->psr.transcoder; u32 val; @@ -1903,11 +1919,11 @@ static bool psr_interrupt_error_check(struct intel_dp *intel_dp) * first time that PSR HW tries to activate so lets keep PSR disabled * to avoid any rendering problems. */ - val = intel_de_read(dev_priv, psr_iir_reg(dev_priv, cpu_transcoder)); + val = intel_de_read(display, psr_iir_reg(display, cpu_transcoder)); val &= psr_irq_psr_error_bit_get(intel_dp); if (val) { intel_dp->psr.sink_not_reliable = true; - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "PSR interruption error set, not enabling PSR\n"); return false; } @@ -1919,11 +1935,11 @@ no_err: static void intel_psr_enable_locked(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state) { + struct intel_display *display = to_intel_display(intel_dp); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); u32 val; - drm_WARN_ON(&dev_priv->drm, intel_dp->psr.enabled); + drm_WARN_ON(display->drm, intel_dp->psr.enabled); intel_dp->psr.sel_update_enabled = crtc_state->has_sel_update; intel_dp->psr.panel_replay_enabled = crtc_state->has_panel_replay; @@ -1944,9 +1960,9 @@ static void intel_psr_enable_locked(struct intel_dp *intel_dp, return; if (intel_dp->psr.panel_replay_enabled) { - drm_dbg_kms(&dev_priv->drm, "Enabling Panel Replay\n"); + drm_dbg_kms(display->drm, "Enabling Panel Replay\n"); } else { - drm_dbg_kms(&dev_priv->drm, "Enabling PSR%s\n", + drm_dbg_kms(display->drm, "Enabling PSR%s\n", intel_dp->psr.sel_update_enabled ? "2" : "1"); /* @@ -1968,68 +1984,71 @@ static void intel_psr_enable_locked(struct intel_dp *intel_dp, static void intel_psr_exit(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); enum transcoder cpu_transcoder = intel_dp->psr.transcoder; u32 val; if (!intel_dp->psr.active) { - if (transcoder_has_psr2(dev_priv, cpu_transcoder)) { - val = intel_de_read(dev_priv, - EDP_PSR2_CTL(dev_priv, cpu_transcoder)); - drm_WARN_ON(&dev_priv->drm, val & EDP_PSR2_ENABLE); + if (transcoder_has_psr2(display, cpu_transcoder)) { + val = intel_de_read(display, + EDP_PSR2_CTL(display, cpu_transcoder)); + drm_WARN_ON(display->drm, val & EDP_PSR2_ENABLE); } - val = intel_de_read(dev_priv, psr_ctl_reg(dev_priv, cpu_transcoder)); - drm_WARN_ON(&dev_priv->drm, val & EDP_PSR_ENABLE); + val = intel_de_read(display, + psr_ctl_reg(display, cpu_transcoder)); + drm_WARN_ON(display->drm, val & EDP_PSR_ENABLE); return; } if (intel_dp->psr.panel_replay_enabled) { - intel_de_rmw(dev_priv, TRANS_DP2_CTL(intel_dp->psr.transcoder), + intel_de_rmw(display, TRANS_DP2_CTL(intel_dp->psr.transcoder), TRANS_DP2_PANEL_REPLAY_ENABLE, 0); } else if (intel_dp->psr.sel_update_enabled) { tgl_disallow_dc3co_on_psr2_exit(intel_dp); - val = intel_de_rmw(dev_priv, - EDP_PSR2_CTL(dev_priv, cpu_transcoder), + val = intel_de_rmw(display, + EDP_PSR2_CTL(display, cpu_transcoder), EDP_PSR2_ENABLE, 0); - drm_WARN_ON(&dev_priv->drm, !(val & EDP_PSR2_ENABLE)); + drm_WARN_ON(display->drm, !(val & EDP_PSR2_ENABLE)); } else { - val = intel_de_rmw(dev_priv, psr_ctl_reg(dev_priv, cpu_transcoder), + val = intel_de_rmw(display, + psr_ctl_reg(display, cpu_transcoder), EDP_PSR_ENABLE, 0); - drm_WARN_ON(&dev_priv->drm, !(val & EDP_PSR_ENABLE)); + drm_WARN_ON(display->drm, !(val & EDP_PSR_ENABLE)); } intel_dp->psr.active = false; } static void intel_psr_wait_exit_locked(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); enum transcoder cpu_transcoder = intel_dp->psr.transcoder; i915_reg_t psr_status; u32 psr_status_mask; if (intel_dp_is_edp(intel_dp) && (intel_dp->psr.sel_update_enabled || intel_dp->psr.panel_replay_enabled)) { - psr_status = EDP_PSR2_STATUS(dev_priv, cpu_transcoder); + psr_status = EDP_PSR2_STATUS(display, cpu_transcoder); psr_status_mask = EDP_PSR2_STATUS_STATE_MASK; } else { - psr_status = psr_status_reg(dev_priv, cpu_transcoder); + psr_status = psr_status_reg(display, cpu_transcoder); psr_status_mask = EDP_PSR_STATUS_STATE_MASK; } /* Wait till PSR is idle */ - if (intel_de_wait_for_clear(dev_priv, psr_status, + if (intel_de_wait_for_clear(display, psr_status, psr_status_mask, 2000)) - drm_err(&dev_priv->drm, "Timed out waiting PSR idle state\n"); + drm_err(display->drm, "Timed out waiting PSR idle state\n"); } static void intel_psr_disable_locked(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); enum transcoder cpu_transcoder = intel_dp->psr.transcoder; lockdep_assert_held(&intel_dp->psr.lock); @@ -2038,9 +2057,9 @@ static void intel_psr_disable_locked(struct intel_dp *intel_dp) return; if (intel_dp->psr.panel_replay_enabled) - drm_dbg_kms(&dev_priv->drm, "Disabling Panel Replay\n"); + drm_dbg_kms(display->drm, "Disabling Panel Replay\n"); else - drm_dbg_kms(&dev_priv->drm, "Disabling PSR%s\n", + drm_dbg_kms(display->drm, "Disabling PSR%s\n", intel_dp->psr.sel_update_enabled ? "2" : "1"); intel_psr_exit(intel_dp); @@ -2050,19 +2069,19 @@ static void intel_psr_disable_locked(struct intel_dp *intel_dp) * Wa_16013835468 * Wa_14015648006 */ - if (DISPLAY_VER(dev_priv) >= 11) - intel_de_rmw(dev_priv, GEN8_CHICKEN_DCPR_1, + if (DISPLAY_VER(display) >= 11) + intel_de_rmw(display, GEN8_CHICKEN_DCPR_1, wa_16013835468_bit_get(intel_dp), 0); if (intel_dp->psr.sel_update_enabled) { /* Wa_16012604467:adlp,mtl[a0,b0] */ if (!intel_dp->psr.panel_replay_enabled && - IS_DISPLAY_VER_STEP(dev_priv, IP_VER(14, 0), STEP_A0, STEP_B0)) - intel_de_rmw(dev_priv, - MTL_CLKGATE_DIS_TRANS(dev_priv, cpu_transcoder), + IS_DISPLAY_VER_STEP(display, IP_VER(14, 0), STEP_A0, STEP_B0)) + intel_de_rmw(display, + MTL_CLKGATE_DIS_TRANS(display, cpu_transcoder), MTL_CLKGATE_DIS_TRANS_DMASC_GATING_DIS, 0); else if (IS_ALDERLAKE_P(dev_priv)) - intel_de_rmw(dev_priv, CLKGATE_DIS_MISC, + intel_de_rmw(display, CLKGATE_DIS_MISC, CLKGATE_DIS_MISC_DMASC_GATING_DIS, 0); } @@ -2071,12 +2090,12 @@ static void intel_psr_disable_locked(struct intel_dp *intel_dp) /* Panel Replay on eDP is always using ALPM aux less. */ if (intel_dp->psr.panel_replay_enabled && intel_dp_is_edp(intel_dp)) { - intel_de_rmw(dev_priv, ALPM_CTL(dev_priv, cpu_transcoder), + intel_de_rmw(display, ALPM_CTL(display, cpu_transcoder), ALPM_CTL_ALPM_ENABLE | ALPM_CTL_ALPM_AUX_LESS_ENABLE, 0); - intel_de_rmw(dev_priv, - PORT_ALPM_CTL(dev_priv, cpu_transcoder), + intel_de_rmw(display, + PORT_ALPM_CTL(display, cpu_transcoder), PORT_ALPM_CTL_ALPM_AUX_LESS_ENABLE, 0); } @@ -2107,12 +2126,12 @@ static void intel_psr_disable_locked(struct intel_dp *intel_dp) void intel_psr_disable(struct intel_dp *intel_dp, const struct intel_crtc_state *old_crtc_state) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); if (!old_crtc_state->has_psr) return; - if (drm_WARN_ON(&dev_priv->drm, !CAN_PSR(intel_dp))) + if (drm_WARN_ON(display->drm, !CAN_PSR(intel_dp))) return; mutex_lock(&intel_dp->psr.lock); @@ -2132,7 +2151,7 @@ void intel_psr_disable(struct intel_dp *intel_dp, */ void intel_psr_pause(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct intel_psr *psr = &intel_dp->psr; if (!CAN_PSR(intel_dp) && !CAN_PANEL_REPLAY(intel_dp)) @@ -2146,7 +2165,7 @@ void intel_psr_pause(struct intel_dp *intel_dp) } /* If we ever hit this, we will need to add refcount to pause/resume */ - drm_WARN_ON(&dev_priv->drm, psr->paused); + drm_WARN_ON(display->drm, psr->paused); intel_psr_exit(intel_dp); intel_psr_wait_exit_locked(intel_dp); @@ -2183,45 +2202,53 @@ unlock: mutex_unlock(&psr->lock); } -static u32 man_trk_ctl_enable_bit_get(struct drm_i915_private *dev_priv) +static u32 man_trk_ctl_enable_bit_get(struct intel_display *display) { - return IS_ALDERLAKE_P(dev_priv) || DISPLAY_VER(dev_priv) >= 14 ? 0 : + struct drm_i915_private *dev_priv = to_i915(display->drm); + + return IS_ALDERLAKE_P(dev_priv) || DISPLAY_VER(display) >= 14 ? 0 : PSR2_MAN_TRK_CTL_ENABLE; } -static u32 man_trk_ctl_single_full_frame_bit_get(struct drm_i915_private *dev_priv) +static u32 man_trk_ctl_single_full_frame_bit_get(struct intel_display *display) { - return IS_ALDERLAKE_P(dev_priv) || DISPLAY_VER(dev_priv) >= 14 ? + struct drm_i915_private *dev_priv = to_i915(display->drm); + + return IS_ALDERLAKE_P(dev_priv) || DISPLAY_VER(display) >= 14 ? ADLP_PSR2_MAN_TRK_CTL_SF_SINGLE_FULL_FRAME : PSR2_MAN_TRK_CTL_SF_SINGLE_FULL_FRAME; } -static u32 man_trk_ctl_partial_frame_bit_get(struct drm_i915_private *dev_priv) +static u32 man_trk_ctl_partial_frame_bit_get(struct intel_display *display) { - return IS_ALDERLAKE_P(dev_priv) || DISPLAY_VER(dev_priv) >= 14 ? + struct drm_i915_private *dev_priv = to_i915(display->drm); + + return IS_ALDERLAKE_P(dev_priv) || DISPLAY_VER(display) >= 14 ? ADLP_PSR2_MAN_TRK_CTL_SF_PARTIAL_FRAME_UPDATE : PSR2_MAN_TRK_CTL_SF_PARTIAL_FRAME_UPDATE; } -static u32 man_trk_ctl_continuos_full_frame(struct drm_i915_private *dev_priv) +static u32 man_trk_ctl_continuos_full_frame(struct intel_display *display) { - return IS_ALDERLAKE_P(dev_priv) || DISPLAY_VER(dev_priv) >= 14 ? + struct drm_i915_private *dev_priv = to_i915(display->drm); + + return IS_ALDERLAKE_P(dev_priv) || DISPLAY_VER(display) >= 14 ? ADLP_PSR2_MAN_TRK_CTL_SF_CONTINUOS_FULL_FRAME : PSR2_MAN_TRK_CTL_SF_CONTINUOS_FULL_FRAME; } static void psr_force_hw_tracking_exit(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); enum transcoder cpu_transcoder = intel_dp->psr.transcoder; if (intel_dp->psr.psr2_sel_fetch_enabled) - intel_de_write(dev_priv, - PSR2_MAN_TRK_CTL(dev_priv, cpu_transcoder), - man_trk_ctl_enable_bit_get(dev_priv) | - man_trk_ctl_partial_frame_bit_get(dev_priv) | - man_trk_ctl_single_full_frame_bit_get(dev_priv) | - man_trk_ctl_continuos_full_frame(dev_priv)); + intel_de_write(display, + PSR2_MAN_TRK_CTL(display, cpu_transcoder), + man_trk_ctl_enable_bit_get(display) | + man_trk_ctl_partial_frame_bit_get(display) | + man_trk_ctl_single_full_frame_bit_get(display) | + man_trk_ctl_continuos_full_frame(display)); /* * Display WA #0884: skl+ @@ -2236,20 +2263,20 @@ static void psr_force_hw_tracking_exit(struct intel_dp *intel_dp) * but testing proved that it works for up display 13, for newer * than that testing will be needed. */ - intel_de_write(dev_priv, CURSURFLIVE(dev_priv, intel_dp->psr.pipe), 0); + intel_de_write(display, CURSURFLIVE(display, intel_dp->psr.pipe), 0); } void intel_psr2_program_trans_man_trk_ctl(const struct intel_crtc_state *crtc_state) { + struct intel_display *display = to_intel_display(crtc_state); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); - struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev); enum transcoder cpu_transcoder = crtc_state->cpu_transcoder; struct intel_encoder *encoder; if (!crtc_state->enable_psr2_sel_fetch) return; - for_each_intel_encoder_mask_with_psr(&dev_priv->drm, encoder, + for_each_intel_encoder_mask_with_psr(display->drm, encoder, crtc_state->uapi.encoder_mask) { struct intel_dp *intel_dp = enc_to_intel_dp(encoder); @@ -2259,36 +2286,37 @@ void intel_psr2_program_trans_man_trk_ctl(const struct intel_crtc_state *crtc_st break; } - intel_de_write(dev_priv, PSR2_MAN_TRK_CTL(dev_priv, cpu_transcoder), + intel_de_write(display, PSR2_MAN_TRK_CTL(display, cpu_transcoder), crtc_state->psr2_man_track_ctl); if (!crtc_state->enable_psr2_su_region_et) return; - intel_de_write(dev_priv, PIPE_SRCSZ_ERLY_TPT(crtc->pipe), + intel_de_write(display, PIPE_SRCSZ_ERLY_TPT(crtc->pipe), crtc_state->pipe_srcsz_early_tpt); } static void psr2_man_trk_ctl_calc(struct intel_crtc_state *crtc_state, bool full_update) { + struct intel_display *display = to_intel_display(crtc_state); struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); - u32 val = man_trk_ctl_enable_bit_get(dev_priv); + u32 val = man_trk_ctl_enable_bit_get(display); /* SF partial frame enable has to be set even on full update */ - val |= man_trk_ctl_partial_frame_bit_get(dev_priv); + val |= man_trk_ctl_partial_frame_bit_get(display); if (full_update) { - val |= man_trk_ctl_single_full_frame_bit_get(dev_priv); - val |= man_trk_ctl_continuos_full_frame(dev_priv); + val |= man_trk_ctl_single_full_frame_bit_get(display); + val |= man_trk_ctl_continuos_full_frame(display); goto exit; } if (crtc_state->psr2_su_area.y1 == -1) goto exit; - if (IS_ALDERLAKE_P(dev_priv) || DISPLAY_VER(dev_priv) >= 14) { + if (IS_ALDERLAKE_P(dev_priv) || DISPLAY_VER(display) >= 14) { val |= ADLP_PSR2_MAN_TRK_CTL_SU_REGION_START_ADDR(crtc_state->psr2_su_area.y1); val |= ADLP_PSR2_MAN_TRK_CTL_SU_REGION_END_ADDR(crtc_state->psr2_su_area.y2 - 1); } else { @@ -2341,13 +2369,14 @@ static void clip_area_update(struct drm_rect *overlap_damage_area, static void intel_psr2_sel_fetch_pipe_alignment(struct intel_crtc_state *crtc_state) { + struct intel_display *display = to_intel_display(crtc_state); struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev); const struct drm_dsc_config *vdsc_cfg = &crtc_state->dsc.config; u16 y_alignment; /* ADLP aligns the SU region to vdsc slice height in case dsc is enabled */ if (crtc_state->dsc.compression_enable && - (IS_ALDERLAKE_P(dev_priv) || DISPLAY_VER(dev_priv) >= 14)) + (IS_ALDERLAKE_P(dev_priv) || DISPLAY_VER(display) >= 14)) y_alignment = vdsc_cfg->slice_height; else y_alignment = crtc_state->su_y_granularity; @@ -2435,6 +2464,7 @@ static bool psr2_sel_fetch_pipe_state_supported(const struct intel_crtc_state *c int intel_psr2_sel_fetch_update(struct intel_atomic_state *state, struct intel_crtc *crtc) { + struct intel_display *display = to_intel_display(state); struct drm_i915_private *dev_priv = to_i915(state->base.dev); struct intel_crtc_state *crtc_state = intel_atomic_get_new_crtc_state(state, crtc); struct intel_plane_state *new_plane_state, *old_plane_state; @@ -2531,7 +2561,7 @@ int intel_psr2_sel_fetch_update(struct intel_atomic_state *state, * calculation for those. */ if (crtc_state->psr2_su_area.y1 == -1) { - drm_info_once(&dev_priv->drm, + drm_info_once(display->drm, "Selective fetch area calculation failed in pipe %c\n", pipe_name(crtc->pipe)); full_update = true; @@ -2542,7 +2572,7 @@ int intel_psr2_sel_fetch_update(struct intel_atomic_state *state, /* Wa_14014971492 */ if (!crtc_state->has_panel_replay && - ((IS_DISPLAY_VER_STEP(dev_priv, IP_VER(14, 0), STEP_A0, STEP_B0) || + ((IS_DISPLAY_VER_STEP(display, IP_VER(14, 0), STEP_A0, STEP_B0) || IS_ALDERLAKE_P(dev_priv) || IS_TIGERLAKE(dev_priv))) && crtc_state->splitter.enable) crtc_state->psr2_su_area.y1 = 0; @@ -2628,6 +2658,7 @@ skip_sel_fetch_set_loop: void intel_psr_pre_plane_update(struct intel_atomic_state *state, struct intel_crtc *crtc) { + struct intel_display *display = to_intel_display(state); struct drm_i915_private *i915 = to_i915(state->base.dev); const struct intel_crtc_state *old_crtc_state = intel_atomic_get_old_crtc_state(state, crtc); @@ -2635,7 +2666,7 @@ void intel_psr_pre_plane_update(struct intel_atomic_state *state, intel_atomic_get_new_crtc_state(state, crtc); struct intel_encoder *encoder; - if (!HAS_PSR(i915)) + if (!HAS_PSR(display)) return; for_each_intel_encoder_mask_with_psr(state->base.dev, encoder, @@ -2676,7 +2707,7 @@ void intel_psr_pre_plane_update(struct intel_atomic_state *state, void intel_psr_post_plane_update(struct intel_atomic_state *state, struct intel_crtc *crtc) { - struct drm_i915_private *dev_priv = to_i915(state->base.dev); + struct intel_display *display = to_intel_display(state); const struct intel_crtc_state *crtc_state = intel_atomic_get_new_crtc_state(state, crtc); struct intel_encoder *encoder; @@ -2692,13 +2723,14 @@ void intel_psr_post_plane_update(struct intel_atomic_state *state, mutex_lock(&psr->lock); - drm_WARN_ON(&dev_priv->drm, psr->enabled && !crtc_state->active_planes); + drm_WARN_ON(display->drm, + psr->enabled && !crtc_state->active_planes); keep_disabled |= psr->sink_not_reliable; keep_disabled |= !crtc_state->active_planes; /* Display WA #1136: skl, bxt */ - keep_disabled |= DISPLAY_VER(dev_priv) < 11 && + keep_disabled |= DISPLAY_VER(display) < 11 && crtc_state->wm_level_disabled; if (!psr->enabled && !keep_disabled) @@ -2723,7 +2755,7 @@ void intel_psr_post_plane_update(struct intel_atomic_state *state, static int _psr2_ready_for_pipe_update_locked(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); enum transcoder cpu_transcoder = intel_dp->psr.transcoder; /* @@ -2731,14 +2763,14 @@ static int _psr2_ready_for_pipe_update_locked(struct intel_dp *intel_dp) * As all higher states has bit 4 of PSR2 state set we can just wait for * EDP_PSR2_STATUS_STATE_DEEP_SLEEP to be cleared. */ - return intel_de_wait_for_clear(dev_priv, - EDP_PSR2_STATUS(dev_priv, cpu_transcoder), + return intel_de_wait_for_clear(display, + EDP_PSR2_STATUS(display, cpu_transcoder), EDP_PSR2_STATUS_STATE_DEEP_SLEEP, 50); } static int _psr1_ready_for_pipe_update_locked(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); enum transcoder cpu_transcoder = intel_dp->psr.transcoder; /* @@ -2747,8 +2779,8 @@ static int _psr1_ready_for_pipe_update_locked(struct intel_dp *intel_dp) * exit training time + 1.5 ms of aux channel handshake. 50 ms is * defensive enough to cover everything. */ - return intel_de_wait_for_clear(dev_priv, - psr_status_reg(dev_priv, cpu_transcoder), + return intel_de_wait_for_clear(display, + psr_status_reg(display, cpu_transcoder), EDP_PSR_STATUS_STATE_MASK, 50); } @@ -2768,13 +2800,13 @@ static int _panel_replay_ready_for_pipe_update_locked(struct intel_dp *intel_dp) */ void intel_psr_wait_for_idle_locked(const struct intel_crtc_state *new_crtc_state) { - struct drm_i915_private *dev_priv = to_i915(new_crtc_state->uapi.crtc->dev); + struct intel_display *display = to_intel_display(new_crtc_state); struct intel_encoder *encoder; if (!new_crtc_state->has_psr) return; - for_each_intel_encoder_mask_with_psr(&dev_priv->drm, encoder, + for_each_intel_encoder_mask_with_psr(display->drm, encoder, new_crtc_state->uapi.encoder_mask) { struct intel_dp *intel_dp = enc_to_intel_dp(encoder); int ret; @@ -2792,13 +2824,14 @@ void intel_psr_wait_for_idle_locked(const struct intel_crtc_state *new_crtc_stat ret = _psr1_ready_for_pipe_update_locked(intel_dp); if (ret) - drm_err(&dev_priv->drm, "PSR wait timed out, atomic update may fail\n"); + drm_err(display->drm, + "PSR wait timed out, atomic update may fail\n"); } } static bool __psr_wait_for_idle_locked(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); enum transcoder cpu_transcoder = intel_dp->psr.transcoder; i915_reg_t reg; u32 mask; @@ -2809,18 +2842,18 @@ static bool __psr_wait_for_idle_locked(struct intel_dp *intel_dp) if (intel_dp_is_edp(intel_dp) && (intel_dp->psr.sel_update_enabled || intel_dp->psr.panel_replay_enabled)) { - reg = EDP_PSR2_STATUS(dev_priv, cpu_transcoder); + reg = EDP_PSR2_STATUS(display, cpu_transcoder); mask = EDP_PSR2_STATUS_STATE_MASK; } else { - reg = psr_status_reg(dev_priv, cpu_transcoder); + reg = psr_status_reg(display, cpu_transcoder); mask = EDP_PSR_STATUS_STATE_MASK; } mutex_unlock(&intel_dp->psr.lock); - err = intel_de_wait_for_clear(dev_priv, reg, mask, 50); + err = intel_de_wait_for_clear(display, reg, mask, 50); if (err) - drm_err(&dev_priv->drm, + drm_err(display->drm, "Timed out waiting for PSR Idle for re-enable\n"); /* After the unlocked wait, verify that PSR is still wanted! */ @@ -2828,7 +2861,7 @@ static bool __psr_wait_for_idle_locked(struct intel_dp *intel_dp) return err == 0 && intel_dp->psr.enabled; } -static int intel_psr_fastset_force(struct drm_i915_private *dev_priv) +static int intel_psr_fastset_force(struct intel_display *display) { struct drm_connector_list_iter conn_iter; struct drm_modeset_acquire_ctx ctx; @@ -2836,7 +2869,7 @@ static int intel_psr_fastset_force(struct drm_i915_private *dev_priv) struct drm_connector *conn; int err = 0; - state = drm_atomic_state_alloc(&dev_priv->drm); + state = drm_atomic_state_alloc(display->drm); if (!state) return -ENOMEM; @@ -2846,7 +2879,7 @@ static int intel_psr_fastset_force(struct drm_i915_private *dev_priv) to_intel_atomic_state(state)->internal = true; retry: - drm_connector_list_iter_begin(&dev_priv->drm, &conn_iter); + drm_connector_list_iter_begin(display->drm, &conn_iter); drm_for_each_connector_iter(conn, &conn_iter) { struct drm_connector_state *conn_state; struct drm_crtc_state *crtc_state; @@ -2893,7 +2926,7 @@ retry: int intel_psr_debug_set(struct intel_dp *intel_dp, u64 val) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); const u32 mode = val & I915_PSR_DEBUG_MODE_MASK; const u32 disable_bits = val & (I915_PSR_DEBUG_SU_REGION_ET_DISABLE | I915_PSR_DEBUG_PANEL_REPLAY_DISABLE); @@ -2904,7 +2937,7 @@ int intel_psr_debug_set(struct intel_dp *intel_dp, u64 val) I915_PSR_DEBUG_PANEL_REPLAY_DISABLE | I915_PSR_DEBUG_MODE_MASK) || mode > I915_PSR_DEBUG_ENABLE_SEL_FETCH) { - drm_dbg_kms(&dev_priv->drm, "Invalid debug mask %llx\n", val); + drm_dbg_kms(display->drm, "Invalid debug mask %llx\n", val); return -EINVAL; } @@ -2929,7 +2962,7 @@ int intel_psr_debug_set(struct intel_dp *intel_dp, u64 val) mutex_unlock(&intel_dp->psr.lock); if (old_mode != mode || old_disable_bits != disable_bits) - ret = intel_psr_fastset_force(dev_priv); + ret = intel_psr_fastset_force(display); return ret; } @@ -2981,7 +3014,7 @@ unlock: static void _psr_invalidate_handle(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); enum transcoder cpu_transcoder = intel_dp->psr.transcoder; if (intel_dp->psr.psr2_sel_fetch_enabled) { @@ -2989,20 +3022,20 @@ static void _psr_invalidate_handle(struct intel_dp *intel_dp) if (intel_dp->psr.psr2_sel_fetch_cff_enabled) { /* Send one update otherwise lag is observed in screen */ - intel_de_write(dev_priv, - CURSURFLIVE(dev_priv, intel_dp->psr.pipe), + intel_de_write(display, + CURSURFLIVE(display, intel_dp->psr.pipe), 0); return; } - val = man_trk_ctl_enable_bit_get(dev_priv) | - man_trk_ctl_partial_frame_bit_get(dev_priv) | - man_trk_ctl_continuos_full_frame(dev_priv); - intel_de_write(dev_priv, - PSR2_MAN_TRK_CTL(dev_priv, cpu_transcoder), + val = man_trk_ctl_enable_bit_get(display) | + man_trk_ctl_partial_frame_bit_get(display) | + man_trk_ctl_continuos_full_frame(display); + intel_de_write(display, + PSR2_MAN_TRK_CTL(display, cpu_transcoder), val); - intel_de_write(dev_priv, - CURSURFLIVE(dev_priv, intel_dp->psr.pipe), 0); + intel_de_write(display, + CURSURFLIVE(display, intel_dp->psr.pipe), 0); intel_dp->psr.psr2_sel_fetch_cff_enabled = true; } else { intel_psr_exit(intel_dp); @@ -3011,7 +3044,7 @@ static void _psr_invalidate_handle(struct intel_dp *intel_dp) /** * intel_psr_invalidate - Invalidate PSR - * @dev_priv: i915 device + * @display: display device * @frontbuffer_bits: frontbuffer plane tracking bits * @origin: which operation caused the invalidate * @@ -3022,7 +3055,7 @@ static void _psr_invalidate_handle(struct intel_dp *intel_dp) * * Dirty frontbuffers relevant to PSR are tracked in busy_frontbuffer_bits." */ -void intel_psr_invalidate(struct drm_i915_private *dev_priv, +void intel_psr_invalidate(struct intel_display *display, unsigned frontbuffer_bits, enum fb_op_origin origin) { struct intel_encoder *encoder; @@ -3030,7 +3063,7 @@ void intel_psr_invalidate(struct drm_i915_private *dev_priv, if (origin == ORIGIN_FLIP) return; - for_each_intel_encoder_with_psr(&dev_priv->drm, encoder) { + for_each_intel_encoder_with_psr(display->drm, encoder) { unsigned int pipe_frontbuffer_bits = frontbuffer_bits; struct intel_dp *intel_dp = enc_to_intel_dp(encoder); @@ -3060,7 +3093,8 @@ static void tgl_dc3co_flush_locked(struct intel_dp *intel_dp, unsigned int frontbuffer_bits, enum fb_op_origin origin) { - struct drm_i915_private *i915 = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *i915 = to_i915(display->drm); if (!intel_dp->psr.dc3co_exitline || !intel_dp->psr.sel_update_enabled || !intel_dp->psr.active) @@ -3081,17 +3115,18 @@ tgl_dc3co_flush_locked(struct intel_dp *intel_dp, unsigned int frontbuffer_bits, static void _psr_flush_handle(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); enum transcoder cpu_transcoder = intel_dp->psr.transcoder; if (intel_dp->psr.psr2_sel_fetch_enabled) { if (intel_dp->psr.psr2_sel_fetch_cff_enabled) { /* can we turn CFF off? */ if (intel_dp->psr.busy_frontbuffer_bits == 0) { - u32 val = man_trk_ctl_enable_bit_get(dev_priv) | - man_trk_ctl_partial_frame_bit_get(dev_priv) | - man_trk_ctl_single_full_frame_bit_get(dev_priv) | - man_trk_ctl_continuos_full_frame(dev_priv); + u32 val = man_trk_ctl_enable_bit_get(display) | + man_trk_ctl_partial_frame_bit_get(display) | + man_trk_ctl_single_full_frame_bit_get(display) | + man_trk_ctl_continuos_full_frame(display); /* * Set psr2_sel_fetch_cff_enabled as false to allow selective @@ -3099,11 +3134,11 @@ static void _psr_flush_handle(struct intel_dp *intel_dp) * SU configuration in case update is sent for any reason after * sff bit gets cleared by the HW on next vblank. */ - intel_de_write(dev_priv, - PSR2_MAN_TRK_CTL(dev_priv, cpu_transcoder), + intel_de_write(display, + PSR2_MAN_TRK_CTL(display, cpu_transcoder), val); - intel_de_write(dev_priv, - CURSURFLIVE(dev_priv, intel_dp->psr.pipe), + intel_de_write(display, + CURSURFLIVE(display, intel_dp->psr.pipe), 0); intel_dp->psr.psr2_sel_fetch_cff_enabled = false; } @@ -3124,7 +3159,7 @@ static void _psr_flush_handle(struct intel_dp *intel_dp) /** * intel_psr_flush - Flush PSR - * @dev_priv: i915 device + * @display: display device * @frontbuffer_bits: frontbuffer plane tracking bits * @origin: which operation caused the flush * @@ -3135,12 +3170,12 @@ static void _psr_flush_handle(struct intel_dp *intel_dp) * * Dirty frontbuffers relevant to PSR are tracked in busy_frontbuffer_bits. */ -void intel_psr_flush(struct drm_i915_private *dev_priv, +void intel_psr_flush(struct intel_display *display, unsigned frontbuffer_bits, enum fb_op_origin origin) { struct intel_encoder *encoder; - for_each_intel_encoder_with_psr(&dev_priv->drm, encoder) { + for_each_intel_encoder_with_psr(display->drm, encoder) { unsigned int pipe_frontbuffer_bits = frontbuffer_bits; struct intel_dp *intel_dp = enc_to_intel_dp(encoder); @@ -3189,11 +3224,12 @@ unlock: */ void intel_psr_init(struct intel_dp *intel_dp) { + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_connector *connector = intel_dp->attached_connector; struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); - if (!(HAS_PSR(dev_priv) || HAS_DP20(dev_priv))) + if (!(HAS_PSR(display) || HAS_DP20(dev_priv))) return; /* @@ -3205,21 +3241,21 @@ void intel_psr_init(struct intel_dp *intel_dp) * So lets keep it hardcoded to PORT_A for BDW, GEN9 and GEN11. * But GEN12 supports a instance of PSR registers per transcoder. */ - if (DISPLAY_VER(dev_priv) < 12 && dig_port->base.port != PORT_A) { - drm_dbg_kms(&dev_priv->drm, + if (DISPLAY_VER(display) < 12 && dig_port->base.port != PORT_A) { + drm_dbg_kms(display->drm, "PSR condition failed: Port not supported\n"); return; } if ((HAS_DP20(dev_priv) && !intel_dp_is_edp(intel_dp)) || - DISPLAY_VER(dev_priv) >= 20) + DISPLAY_VER(display) >= 20) intel_dp->psr.source_panel_replay_support = true; - if (HAS_PSR(dev_priv) && intel_dp_is_edp(intel_dp)) + if (HAS_PSR(display) && intel_dp_is_edp(intel_dp)) intel_dp->psr.source_support = true; /* Set link_standby x link_off defaults */ - if (DISPLAY_VER(dev_priv) < 12) + if (DISPLAY_VER(display) < 12) /* For new platforms up to TGL let's respect VBT back again */ intel_dp->psr.link_standby = connector->panel.vbt.psr.full_link; @@ -3256,7 +3292,7 @@ static int psr_get_status_and_error_status(struct intel_dp *intel_dp, static void psr_alpm_check(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct drm_dp_aux *aux = &intel_dp->aux; struct intel_psr *psr = &intel_dp->psr; u8 val; @@ -3267,14 +3303,14 @@ static void psr_alpm_check(struct intel_dp *intel_dp) r = drm_dp_dpcd_readb(aux, DP_RECEIVER_ALPM_STATUS, &val); if (r != 1) { - drm_err(&dev_priv->drm, "Error reading ALPM status\n"); + drm_err(display->drm, "Error reading ALPM status\n"); return; } if (val & DP_ALPM_LOCK_TIMEOUT_ERROR) { intel_psr_disable_locked(intel_dp); psr->sink_not_reliable = true; - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "ALPM lock timeout error, disabling PSR\n"); /* Clearing error */ @@ -3284,21 +3320,21 @@ static void psr_alpm_check(struct intel_dp *intel_dp) static void psr_capability_changed_check(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct intel_psr *psr = &intel_dp->psr; u8 val; int r; r = drm_dp_dpcd_readb(&intel_dp->aux, DP_PSR_ESI, &val); if (r != 1) { - drm_err(&dev_priv->drm, "Error reading DP_PSR_ESI\n"); + drm_err(display->drm, "Error reading DP_PSR_ESI\n"); return; } if (val & DP_PSR_CAPS_CHANGE) { intel_psr_disable_locked(intel_dp); psr->sink_not_reliable = true; - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "Sink PSR capability changed, disabling PSR\n"); /* Clearing it */ @@ -3315,7 +3351,7 @@ static void psr_capability_changed_check(struct intel_dp *intel_dp) */ void intel_psr_short_pulse(struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); struct intel_psr *psr = &intel_dp->psr; u8 status, error_status; const u8 errors = DP_PSR_RFB_STORAGE_ERROR | @@ -3331,7 +3367,7 @@ void intel_psr_short_pulse(struct intel_dp *intel_dp) goto exit; if (psr_get_status_and_error_status(intel_dp, &status, &error_status)) { - drm_err(&dev_priv->drm, + drm_err(display->drm, "Error reading PSR status or error status\n"); goto exit; } @@ -3344,20 +3380,20 @@ void intel_psr_short_pulse(struct intel_dp *intel_dp) if (!psr->panel_replay_enabled && status == DP_PSR_SINK_INTERNAL_ERROR && !error_status) - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "PSR sink internal error, disabling PSR\n"); if (error_status & DP_PSR_RFB_STORAGE_ERROR) - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "PSR RFB storage error, disabling PSR\n"); if (error_status & DP_PSR_VSC_SDP_UNCORRECTABLE_ERROR) - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "PSR VSC SDP uncorrectable error, disabling PSR\n"); if (error_status & DP_PSR_LINK_CRC_ERROR) - drm_dbg_kms(&dev_priv->drm, + drm_dbg_kms(display->drm, "PSR Link CRC error, disabling PSR\n"); if (error_status & ~errors) - drm_err(&dev_priv->drm, + drm_err(display->drm, "PSR_ERROR_STATUS unhandled errors %x\n", error_status & ~errors); /* clear status register */ @@ -3396,13 +3432,13 @@ bool intel_psr_enabled(struct intel_dp *intel_dp) */ void intel_psr_lock(const struct intel_crtc_state *crtc_state) { - struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev); + struct intel_display *display = to_intel_display(crtc_state); struct intel_encoder *encoder; if (!crtc_state->has_psr) return; - for_each_intel_encoder_mask_with_psr(&i915->drm, encoder, + for_each_intel_encoder_mask_with_psr(display->drm, encoder, crtc_state->uapi.encoder_mask) { struct intel_dp *intel_dp = enc_to_intel_dp(encoder); @@ -3419,13 +3455,13 @@ void intel_psr_lock(const struct intel_crtc_state *crtc_state) */ void intel_psr_unlock(const struct intel_crtc_state *crtc_state) { - struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev); + struct intel_display *display = to_intel_display(crtc_state); struct intel_encoder *encoder; if (!crtc_state->has_psr) return; - for_each_intel_encoder_mask_with_psr(&i915->drm, encoder, + for_each_intel_encoder_mask_with_psr(display->drm, encoder, crtc_state->uapi.encoder_mask) { struct intel_dp *intel_dp = enc_to_intel_dp(encoder); @@ -3437,7 +3473,7 @@ void intel_psr_unlock(const struct intel_crtc_state *crtc_state) static void psr_source_status(struct intel_dp *intel_dp, struct seq_file *m) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); enum transcoder cpu_transcoder = intel_dp->psr.transcoder; const char *status = "unknown"; u32 val, status_val; @@ -3457,8 +3493,8 @@ psr_source_status(struct intel_dp *intel_dp, struct seq_file *m) "BUF_ON", "TG_ON" }; - val = intel_de_read(dev_priv, - EDP_PSR2_STATUS(dev_priv, cpu_transcoder)); + val = intel_de_read(display, + EDP_PSR2_STATUS(display, cpu_transcoder)); status_val = REG_FIELD_GET(EDP_PSR2_STATUS_STATE_MASK, val); if (status_val < ARRAY_SIZE(live_status)) status = live_status[status_val]; @@ -3473,7 +3509,8 @@ psr_source_status(struct intel_dp *intel_dp, struct seq_file *m) "SRDOFFACK", "SRDENT_ON", }; - val = intel_de_read(dev_priv, psr_status_reg(dev_priv, cpu_transcoder)); + val = intel_de_read(display, + psr_status_reg(display, cpu_transcoder)); status_val = REG_FIELD_GET(EDP_PSR_STATUS_STATE_MASK, val); if (status_val < ARRAY_SIZE(live_status)) status = live_status[status_val]; @@ -3534,7 +3571,8 @@ static void intel_psr_print_mode(struct intel_dp *intel_dp, static int intel_psr_status(struct seq_file *m, struct intel_dp *intel_dp) { - struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + struct intel_display *display = to_intel_display(intel_dp); + struct drm_i915_private *dev_priv = to_i915(display->drm); enum transcoder cpu_transcoder = intel_dp->psr.transcoder; struct intel_psr *psr = &intel_dp->psr; intel_wakeref_t wakeref; @@ -3559,20 +3597,20 @@ static int intel_psr_status(struct seq_file *m, struct intel_dp *intel_dp) } if (psr->panel_replay_enabled) { - val = intel_de_read(dev_priv, TRANS_DP2_CTL(cpu_transcoder)); + val = intel_de_read(display, TRANS_DP2_CTL(cpu_transcoder)); if (intel_dp_is_edp(intel_dp)) - psr2_ctl = intel_de_read(dev_priv, - EDP_PSR2_CTL(dev_priv, + psr2_ctl = intel_de_read(display, + EDP_PSR2_CTL(display, cpu_transcoder)); enabled = val & TRANS_DP2_PANEL_REPLAY_ENABLE; } else if (psr->sel_update_enabled) { - val = intel_de_read(dev_priv, - EDP_PSR2_CTL(dev_priv, cpu_transcoder)); + val = intel_de_read(display, + EDP_PSR2_CTL(display, cpu_transcoder)); enabled = val & EDP_PSR2_ENABLE; } else { - val = intel_de_read(dev_priv, psr_ctl_reg(dev_priv, cpu_transcoder)); + val = intel_de_read(display, psr_ctl_reg(display, cpu_transcoder)); enabled = val & EDP_PSR_ENABLE; } seq_printf(m, "Source PSR/PanelReplay ctl: %s [0x%08x]\n", @@ -3587,7 +3625,7 @@ static int intel_psr_status(struct seq_file *m, struct intel_dp *intel_dp) /* * SKL+ Perf counter is reset to 0 everytime DC state is entered */ - val = intel_de_read(dev_priv, psr_perf_cnt_reg(dev_priv, cpu_transcoder)); + val = intel_de_read(display, psr_perf_cnt_reg(display, cpu_transcoder)); seq_printf(m, "Performance counter: %u\n", REG_FIELD_GET(EDP_PSR_PERF_CNT_MASK, val)); @@ -3606,8 +3644,8 @@ static int intel_psr_status(struct seq_file *m, struct intel_dp *intel_dp) * frame boundary between register reads */ for (frame = 0; frame < PSR2_SU_STATUS_FRAMES; frame += 3) { - val = intel_de_read(dev_priv, - PSR2_SU_STATUS(dev_priv, cpu_transcoder, frame)); + val = intel_de_read(display, + PSR2_SU_STATUS(display, cpu_transcoder, frame)); su_frames_val[frame / 3] = val; } @@ -3635,15 +3673,15 @@ unlock: static int i915_edp_psr_status_show(struct seq_file *m, void *data) { - struct drm_i915_private *dev_priv = m->private; + struct intel_display *display = m->private; struct intel_dp *intel_dp = NULL; struct intel_encoder *encoder; - if (!HAS_PSR(dev_priv)) + if (!HAS_PSR(display)) return -ENODEV; /* Find the first EDP which supports PSR */ - for_each_intel_encoder_with_psr(&dev_priv->drm, encoder) { + for_each_intel_encoder_with_psr(display->drm, encoder) { intel_dp = enc_to_intel_dp(encoder); break; } @@ -3658,18 +3696,19 @@ DEFINE_SHOW_ATTRIBUTE(i915_edp_psr_status); static int i915_edp_psr_debug_set(void *data, u64 val) { - struct drm_i915_private *dev_priv = data; + struct intel_display *display = data; + struct drm_i915_private *dev_priv = to_i915(display->drm); struct intel_encoder *encoder; intel_wakeref_t wakeref; int ret = -ENODEV; - if (!HAS_PSR(dev_priv)) + if (!HAS_PSR(display)) return ret; - for_each_intel_encoder_with_psr(&dev_priv->drm, encoder) { + for_each_intel_encoder_with_psr(display->drm, encoder) { struct intel_dp *intel_dp = enc_to_intel_dp(encoder); - drm_dbg_kms(&dev_priv->drm, "Setting PSR debug to %llx\n", val); + drm_dbg_kms(display->drm, "Setting PSR debug to %llx\n", val); wakeref = intel_runtime_pm_get(&dev_priv->runtime_pm); @@ -3685,13 +3724,13 @@ i915_edp_psr_debug_set(void *data, u64 val) static int i915_edp_psr_debug_get(void *data, u64 *val) { - struct drm_i915_private *dev_priv = data; + struct intel_display *display = data; struct intel_encoder *encoder; - if (!HAS_PSR(dev_priv)) + if (!HAS_PSR(display)) return -ENODEV; - for_each_intel_encoder_with_psr(&dev_priv->drm, encoder) { + for_each_intel_encoder_with_psr(display->drm, encoder) { struct intel_dp *intel_dp = enc_to_intel_dp(encoder); // TODO: split to each transcoder's PSR debug state @@ -3706,15 +3745,15 @@ DEFINE_SIMPLE_ATTRIBUTE(i915_edp_psr_debug_fops, i915_edp_psr_debug_get, i915_edp_psr_debug_set, "%llu\n"); -void intel_psr_debugfs_register(struct drm_i915_private *i915) +void intel_psr_debugfs_register(struct intel_display *display) { - struct drm_minor *minor = i915->drm.primary; + struct drm_minor *minor = display->drm->primary; debugfs_create_file("i915_edp_psr_debug", 0644, minor->debugfs_root, - i915, &i915_edp_psr_debug_fops); + display, &i915_edp_psr_debug_fops); debugfs_create_file("i915_edp_psr_status", 0444, minor->debugfs_root, - i915, &i915_edp_psr_status_fops); + display, &i915_edp_psr_status_fops); } static const char *psr_mode_str(struct intel_dp *intel_dp) @@ -3795,6 +3834,7 @@ DEFINE_SHOW_ATTRIBUTE(i915_psr_status); void intel_psr_connector_debugfs_add(struct intel_connector *connector) { + struct intel_display *display = to_intel_display(connector); struct drm_i915_private *i915 = to_i915(connector->base.dev); struct dentry *root = connector->base.debugfs_entry; @@ -3807,7 +3847,7 @@ void intel_psr_connector_debugfs_add(struct intel_connector *connector) debugfs_create_file("i915_psr_sink_status", 0444, root, connector, &i915_psr_sink_status_fops); - if (HAS_PSR(i915) || HAS_DP20(i915)) + if (HAS_PSR(display) || HAS_DP20(i915)) debugfs_create_file("i915_psr_status", 0444, root, connector, &i915_psr_status_fops); } diff --git a/drivers/gpu/drm/i915/display/intel_psr.h b/drivers/gpu/drm/i915/display/intel_psr.h index d483c85870e1..4e09c10908e4 100644 --- a/drivers/gpu/drm/i915/display/intel_psr.h +++ b/drivers/gpu/drm/i915/display/intel_psr.h @@ -11,11 +11,11 @@ enum fb_op_origin; struct drm_connector; struct drm_connector_state; -struct drm_i915_private; struct intel_atomic_state; struct intel_connector; struct intel_crtc; struct intel_crtc_state; +struct intel_display; struct intel_dp; struct intel_encoder; struct intel_plane; @@ -35,10 +35,10 @@ void intel_psr_post_plane_update(struct intel_atomic_state *state, void intel_psr_disable(struct intel_dp *intel_dp, const struct intel_crtc_state *old_crtc_state); int intel_psr_debug_set(struct intel_dp *intel_dp, u64 value); -void intel_psr_invalidate(struct drm_i915_private *dev_priv, +void intel_psr_invalidate(struct intel_display *display, unsigned frontbuffer_bits, enum fb_op_origin origin); -void intel_psr_flush(struct drm_i915_private *dev_priv, +void intel_psr_flush(struct intel_display *display, unsigned frontbuffer_bits, enum fb_op_origin origin); void intel_psr_init(struct intel_dp *intel_dp); @@ -60,6 +60,6 @@ void intel_psr_resume(struct intel_dp *intel_dp); void intel_psr_lock(const struct intel_crtc_state *crtc_state); void intel_psr_unlock(const struct intel_crtc_state *crtc_state); void intel_psr_connector_debugfs_add(struct intel_connector *connector); -void intel_psr_debugfs_register(struct drm_i915_private *i915); +void intel_psr_debugfs_register(struct intel_display *display); #endif /* __INTEL_PSR_H__ */ diff --git a/drivers/gpu/drm/i915/display/intel_quirks.c b/drivers/gpu/drm/i915/display/intel_quirks.c index 14d5fefc9c5b..29b56d53a340 100644 --- a/drivers/gpu/drm/i915/display/intel_quirks.c +++ b/drivers/gpu/drm/i915/display/intel_quirks.c @@ -14,6 +14,11 @@ static void intel_set_quirk(struct intel_display *display, enum intel_quirk_id q display->quirks.mask |= BIT(quirk); } +static void intel_set_dpcd_quirk(struct intel_dp *intel_dp, enum intel_quirk_id quirk) +{ + intel_dp->quirks.mask |= BIT(quirk); +} + /* * Some machines (Lenovo U160) do not work with SSC on LVDS for some reason */ @@ -65,6 +70,14 @@ static void quirk_no_pps_backlight_power_hook(struct intel_display *display) drm_info(display->drm, "Applying no pps backlight power quirk\n"); } +static void quirk_fw_sync_len(struct intel_dp *intel_dp) +{ + struct intel_display *display = to_intel_display(intel_dp); + + intel_set_dpcd_quirk(intel_dp, QUIRK_FW_SYNC_LEN); + drm_info(display->drm, "Applying Fast Wake sync pulse count quirk\n"); +} + struct intel_quirk { int device; int subsystem_vendor; @@ -72,6 +85,21 @@ struct intel_quirk { void (*hook)(struct intel_display *display); }; +struct intel_dpcd_quirk { + int device; + int subsystem_vendor; + int subsystem_device; + u8 sink_oui[3]; + u8 sink_device_id[6]; + void (*hook)(struct intel_dp *intel_dp); +}; + +#define SINK_OUI(first, second, third) { (first), (second), (third) } +#define SINK_DEVICE_ID(first, second, third, fourth, fifth, sixth) \ + { (first), (second), (third), (fourth), (fifth), (sixth) } + +#define SINK_DEVICE_ID_ANY SINK_DEVICE_ID(0, 0, 0, 0, 0, 0) + /* For systems that don't have a meaningful PCI subdevice/subvendor ID */ struct intel_dmi_quirk { void (*hook)(struct intel_display *display); @@ -203,6 +231,18 @@ static struct intel_quirk intel_quirks[] = { { 0x0f31, 0x103c, 0x220f, quirk_invert_brightness }, }; +static struct intel_dpcd_quirk intel_dpcd_quirks[] = { + /* Dell Precision 5490 */ + { + .device = 0x7d55, + .subsystem_vendor = 0x1028, + .subsystem_device = 0x0cc7, + .sink_oui = SINK_OUI(0x38, 0xec, 0x11), + .hook = quirk_fw_sync_len, + }, + +}; + void intel_init_quirks(struct intel_display *display) { struct pci_dev *d = to_pci_dev(display->drm->dev); @@ -224,7 +264,35 @@ void intel_init_quirks(struct intel_display *display) } } +void intel_init_dpcd_quirks(struct intel_dp *intel_dp, + const struct drm_dp_dpcd_ident *ident) +{ + struct intel_display *display = to_intel_display(intel_dp); + struct pci_dev *d = to_pci_dev(display->drm->dev); + int i; + + for (i = 0; i < ARRAY_SIZE(intel_dpcd_quirks); i++) { + struct intel_dpcd_quirk *q = &intel_dpcd_quirks[i]; + + if (d->device == q->device && + (d->subsystem_vendor == q->subsystem_vendor || + q->subsystem_vendor == PCI_ANY_ID) && + (d->subsystem_device == q->subsystem_device || + q->subsystem_device == PCI_ANY_ID) && + !memcmp(q->sink_oui, ident->oui, sizeof(ident->oui)) && + (!memcmp(q->sink_device_id, ident->device_id, + sizeof(ident->device_id)) || + mem_is_zero(q->sink_device_id, sizeof(q->sink_device_id)))) + q->hook(intel_dp); + } +} + bool intel_has_quirk(struct intel_display *display, enum intel_quirk_id quirk) { return display->quirks.mask & BIT(quirk); } + +bool intel_has_dpcd_quirk(struct intel_dp *intel_dp, enum intel_quirk_id quirk) +{ + return intel_dp->quirks.mask & BIT(quirk); +} diff --git a/drivers/gpu/drm/i915/display/intel_quirks.h b/drivers/gpu/drm/i915/display/intel_quirks.h index 151c8f4ae576..cafdebda7535 100644 --- a/drivers/gpu/drm/i915/display/intel_quirks.h +++ b/drivers/gpu/drm/i915/display/intel_quirks.h @@ -9,6 +9,8 @@ #include <linux/types.h> struct intel_display; +struct intel_dp; +struct drm_dp_dpcd_ident; enum intel_quirk_id { QUIRK_BACKLIGHT_PRESENT, @@ -17,9 +19,13 @@ enum intel_quirk_id { QUIRK_INVERT_BRIGHTNESS, QUIRK_LVDS_SSC_DISABLE, QUIRK_NO_PPS_BACKLIGHT_POWER_HOOK, + QUIRK_FW_SYNC_LEN, }; void intel_init_quirks(struct intel_display *display); +void intel_init_dpcd_quirks(struct intel_dp *intel_dp, + const struct drm_dp_dpcd_ident *ident); bool intel_has_quirk(struct intel_display *display, enum intel_quirk_id quirk); +bool intel_has_dpcd_quirk(struct intel_dp *intel_dp, enum intel_quirk_id quirk); #endif /* __INTEL_QUIRKS_H__ */ diff --git a/drivers/gpu/drm/i915/display/intel_vbt_defs.h b/drivers/gpu/drm/i915/display/intel_vbt_defs.h index e613288937e4..42022756bbd5 100644 --- a/drivers/gpu/drm/i915/display/intel_vbt_defs.h +++ b/drivers/gpu/drm/i915/display/intel_vbt_defs.h @@ -1080,6 +1080,8 @@ struct bdb_edp { u16 edp_fast_link_training_rate[16]; /* 224+ */ u16 edp_max_port_link_rate[16]; /* 244+ */ u16 edp_dsc_disable; /* 251+ */ + u16 t6_delay_support; /* 260+ */ + u16 link_idle_time[16]; /* 260+ */ } __packed; /* @@ -1321,7 +1323,7 @@ struct als_data_entry { } __packed; struct aggressiveness_profile_entry { - u8 dpst_aggressiveness : 4; + u8 dpst_aggressiveness : 4; /* (228/252)-256 */ u8 lace_aggressiveness : 4; } __packed; @@ -1330,12 +1332,27 @@ struct aggressiveness_profile2_entry { u8 elp_aggressiveness : 4; } __packed; +struct aggressiveness_profile3_entry { + u8 apd_aggressiveness:4; + u8 pixoptix_aggressiveness:4; +} __packed; + +struct aggressiveness_profile4_entry { + u8 xpst_aggressiveness:4; + u8 tcon_aggressiveness:4; +} __packed; + +struct panel_identification { + u8 panel_technology:4; + u8 reserved:4; +} __packed; + struct bdb_lfp_power { struct lfp_power_features features; /* ???-227 */ struct als_data_entry als[5]; u8 lace_aggressiveness_profile:3; /* 210-227 */ u8 reserved1:5; - u16 dpst; /* 228+ */ + u16 dpst; /* 228-256 */ u16 psr; /* 228+ */ u16 drrs; /* 228+ */ u16 lace_support; /* 228+ */ @@ -1343,12 +1360,20 @@ struct bdb_lfp_power { u16 dmrrs; /* 228+ */ u16 adb; /* 228+ */ u16 lace_enabled_status; /* 228+ */ - struct aggressiveness_profile_entry aggressiveness[16]; /* 228+ */ + struct aggressiveness_profile_entry aggressiveness[16]; u16 hobl; /* 232+ */ u16 vrr_feature_enabled; /* 233+ */ - u16 elp; /* 247+ */ - u16 opst; /* 247+ */ - struct aggressiveness_profile2_entry aggressiveness2[16]; /* 247+ */ + u16 elp; /* 247-256 */ + u16 opst; /* 247-256 */ + struct aggressiveness_profile2_entry aggressiveness2[16]; /* 247-256 */ + u16 apd; /* 253-256 */ + u16 pixoptix; /* 253-256 */ + struct aggressiveness_profile3_entry aggressiveness3[16]; /* 253-256 */ + struct panel_identification panel_identification[16]; /* 257+ */ + u16 xpst_support; /* 257+ */ + u16 tcon_based_backlight_optimization; /* 257+ */ + struct aggressiveness_profile4_entry aggressiveness4[16]; /* 257+ */ + u16 tcon_backlight_xpst_coexistence; /* 257+ */ } __packed; /* diff --git a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c index 01b7587dd1f8..a3b83cfe1726 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c @@ -825,7 +825,7 @@ static int eb_select_context(struct i915_execbuffer *eb) struct i915_gem_context *ctx; ctx = i915_gem_context_lookup(eb->file->driver_priv, eb->args->rsvd1); - if (unlikely(IS_ERR(ctx))) + if (IS_ERR(ctx)) return PTR_ERR(ctx); eb->gem_context = ctx; diff --git a/drivers/gpu/drm/i915/gt/intel_gt_regs.h b/drivers/gpu/drm/i915/gt/intel_gt_regs.h index e42b3a5d4e63..57a3c83d3655 100644 --- a/drivers/gpu/drm/i915/gt/intel_gt_regs.h +++ b/drivers/gpu/drm/i915/gt/intel_gt_regs.h @@ -1553,6 +1553,8 @@ #define VLV_RENDER_C0_COUNT _MMIO(0x138118) #define VLV_MEDIA_C0_COUNT _MMIO(0x13811c) +#define PCU_PWM_FAN_SPEED _MMIO(0x138140) + #define GEN12_RPSTAT1 _MMIO(0x1381b4) #define GEN12_VOLTAGE_MASK REG_GENMASK(10, 0) #define GEN12_CAGF_MASK REG_GENMASK(19, 11) diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.c b/drivers/gpu/drm/i915/gt/intel_workarounds.c index bfe6d8fc820f..e539a656cfc3 100644 --- a/drivers/gpu/drm/i915/gt/intel_workarounds.c +++ b/drivers/gpu/drm/i915/gt/intel_workarounds.c @@ -111,9 +111,8 @@ static void wa_init_finish(struct i915_wa_list *wal) { /* Trim unused entries. */ if (!IS_ALIGNED(wal->count, WA_LIST_CHUNK)) { - struct i915_wa *list = kmemdup(wal->list, - wal->count * sizeof(*list), - GFP_KERNEL); + struct i915_wa *list = kmemdup_array(wal->list, wal->count, + sizeof(*list), GFP_KERNEL); if (list) { kfree(wal->list); @@ -2072,7 +2071,7 @@ static void dg2_whitelist_build(struct intel_engine_cs *engine) case RENDER_CLASS: /* Required by recommended tuning setting (not a workaround) */ whitelist_mcr_reg(w, XEHP_COMMON_SLICE_CHICKEN3); - + whitelist_reg(w, GEN7_COMMON_SLICE_CHICKEN1); break; default: break; @@ -2087,7 +2086,7 @@ static void xelpg_whitelist_build(struct intel_engine_cs *engine) case RENDER_CLASS: /* Required by recommended tuning setting (not a workaround) */ whitelist_mcr_reg(w, XEHP_COMMON_SLICE_CHICKEN3); - + whitelist_reg(w, GEN7_COMMON_SLICE_CHICKEN1); break; default: break; diff --git a/drivers/gpu/drm/i915/gt/sysfs_engines.c b/drivers/gpu/drm/i915/gt/sysfs_engines.c index 021f51d9b456..aab2759067d2 100644 --- a/drivers/gpu/drm/i915/gt/sysfs_engines.c +++ b/drivers/gpu/drm/i915/gt/sysfs_engines.c @@ -530,9 +530,8 @@ void intel_engines_add_sysfs(struct drm_i915_private *i915) err_object: kobject_put(kobj); err_engine: - dev_err(kdev, "Failed to add sysfs engine '%s'\n", - engine->name); - break; + dev_warn(kdev, "Failed to add sysfs engine '%s'\n", + engine->name); } } } diff --git a/drivers/gpu/drm/i915/gt/uc/intel_gsc_uc.c b/drivers/gpu/drm/i915/gt/uc/intel_gsc_uc.c index 453d855dd1de..3d3191deb0ab 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_gsc_uc.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_gsc_uc.c @@ -302,7 +302,7 @@ void intel_gsc_uc_load_start(struct intel_gsc_uc *gsc) { struct intel_gt *gt = gsc_uc_to_gt(gsc); - if (!intel_uc_fw_is_loadable(&gsc->fw)) + if (!intel_uc_fw_is_loadable(&gsc->fw) || intel_uc_fw_is_in_error(&gsc->fw)) return; if (intel_gsc_uc_fw_init_done(gsc)) diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c index c3a5d9e1288e..8322c913bc3c 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c @@ -4507,7 +4507,7 @@ static void guc_default_vfuncs(struct intel_engine_cs *engine) /* Wa_16019325821 */ /* Wa_14019159160 */ if ((engine->class == COMPUTE_CLASS || engine->class == RENDER_CLASS) && - IS_GFX_GT_IP_RANGE(engine->gt, IP_VER(12, 70), IP_VER(12, 71))) + IS_GFX_GT_IP_RANGE(engine->gt, IP_VER(12, 70), IP_VER(12, 74))) engine->flags |= I915_ENGINE_USES_WA_HOLD_SWITCHOUT; /* diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.h b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.h index 9a431726c8d5..ac7b3aad2222 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.h +++ b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.h @@ -258,6 +258,11 @@ static inline bool intel_uc_fw_is_running(struct intel_uc_fw *uc_fw) return __intel_uc_fw_status(uc_fw) == INTEL_UC_FIRMWARE_RUNNING; } +static inline bool intel_uc_fw_is_in_error(struct intel_uc_fw *uc_fw) +{ + return intel_uc_fw_status_to_error(__intel_uc_fw_status(uc_fw)) != 0; +} + static inline bool intel_uc_fw_is_overridden(const struct intel_uc_fw *uc_fw) { return uc_fw->user_overridden; diff --git a/drivers/gpu/drm/i915/i915_driver.c b/drivers/gpu/drm/i915/i915_driver.c index ccdd2983cfb5..fe905d65ddf7 100644 --- a/drivers/gpu/drm/i915/i915_driver.c +++ b/drivers/gpu/drm/i915/i915_driver.c @@ -723,7 +723,7 @@ i915_driver_create(struct pci_dev *pdev, const struct pci_device_id *ent) if (IS_ERR(i915)) return i915; - pci_set_drvdata(pdev, i915); + pci_set_drvdata(pdev, &i915->drm); /* Device parameters start as a copy of module parameters. */ i915_params_copy(&i915->params, &i915_modparams); @@ -1167,7 +1167,7 @@ static int i915_drm_resume(struct drm_device *dev) intel_dmc_resume(dev_priv); i915_restore_display(dev_priv); - intel_pps_unlock_regs_wa(dev_priv); + intel_pps_unlock_regs_wa(display); intel_init_pch_refclk(dev_priv); diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index d772cbe15fec..39f6614a0a99 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -365,12 +365,16 @@ static inline struct drm_i915_private *to_i915(const struct drm_device *dev) static inline struct drm_i915_private *kdev_to_i915(struct device *kdev) { - return dev_get_drvdata(kdev); + struct drm_device *drm = dev_get_drvdata(kdev); + + return drm ? to_i915(drm) : NULL; } static inline struct drm_i915_private *pdev_to_i915(struct pci_dev *pdev) { - return pci_get_drvdata(pdev); + struct drm_device *drm = pci_get_drvdata(pdev); + + return drm ? to_i915(drm) : NULL; } static inline struct intel_gt *to_gt(const struct drm_i915_private *i915) diff --git a/drivers/gpu/drm/i915/i915_hwmon.c b/drivers/gpu/drm/i915/i915_hwmon.c index 49db3e09826c..17d30f6b84b0 100644 --- a/drivers/gpu/drm/i915/i915_hwmon.c +++ b/drivers/gpu/drm/i915/i915_hwmon.c @@ -5,6 +5,7 @@ #include <linux/hwmon.h> #include <linux/hwmon-sysfs.h> +#include <linux/jiffies.h> #include <linux/types.h> #include "i915_drv.h" @@ -36,6 +37,7 @@ struct hwm_reg { i915_reg_t pkg_rapl_limit; i915_reg_t energy_status_all; i915_reg_t energy_status_tile; + i915_reg_t fan_speed; }; struct hwm_energy_info { @@ -43,11 +45,17 @@ struct hwm_energy_info { long accum_energy; /* Accumulated energy for energy1_input */ }; +struct hwm_fan_info { + u32 reg_val_prev; + u64 time_prev; +}; + struct hwm_drvdata { struct i915_hwmon *hwmon; struct intel_uncore *uncore; struct device *hwmon_dev; struct hwm_energy_info ei; /* Energy info for energy1_input */ + struct hwm_fan_info fi; /* Fan info for fan1_input */ char name[12]; int gt_n; bool reset_in_progress; @@ -276,6 +284,7 @@ static const struct hwmon_channel_info * const hwm_info[] = { HWMON_CHANNEL_INFO(power, HWMON_P_MAX | HWMON_P_RATED_MAX | HWMON_P_CRIT), HWMON_CHANNEL_INFO(energy, HWMON_E_INPUT), HWMON_CHANNEL_INFO(curr, HWMON_C_CRIT), + HWMON_CHANNEL_INFO(fan, HWMON_F_INPUT), NULL }; @@ -614,6 +623,69 @@ hwm_curr_write(struct hwm_drvdata *ddat, u32 attr, long val) } static umode_t +hwm_fan_is_visible(const struct hwm_drvdata *ddat, u32 attr) +{ + struct i915_hwmon *hwmon = ddat->hwmon; + + if (attr == hwmon_fan_input && i915_mmio_reg_valid(hwmon->rg.fan_speed)) + return 0444; + + return 0; +} + +static int +hwm_fan_input_read(struct hwm_drvdata *ddat, long *val) +{ + struct i915_hwmon *hwmon = ddat->hwmon; + struct hwm_fan_info *fi = &ddat->fi; + u64 rotations, time_now, time; + intel_wakeref_t wakeref; + u32 reg_val; + int ret = 0; + + wakeref = intel_runtime_pm_get(ddat->uncore->rpm); + mutex_lock(&hwmon->hwmon_lock); + + reg_val = intel_uncore_read(ddat->uncore, hwmon->rg.fan_speed); + time_now = get_jiffies_64(); + + /* + * HW register value is accumulated count of pulses from + * PWM fan with the scale of 2 pulses per rotation. + */ + rotations = (reg_val - fi->reg_val_prev) / 2; + + time = jiffies_delta_to_msecs(time_now - fi->time_prev); + if (unlikely(!time)) { + ret = -EAGAIN; + goto exit; + } + + /* + * Calculate fan speed in RPM by time averaging two subsequent + * readings in minutes. + * RPM = number of rotations * msecs per minute / time in msecs + */ + *val = DIV_ROUND_UP_ULL(rotations * (MSEC_PER_SEC * 60), time); + + fi->reg_val_prev = reg_val; + fi->time_prev = time_now; +exit: + mutex_unlock(&hwmon->hwmon_lock); + intel_runtime_pm_put(ddat->uncore->rpm, wakeref); + return ret; +} + +static int +hwm_fan_read(struct hwm_drvdata *ddat, u32 attr, long *val) +{ + if (attr == hwmon_fan_input) + return hwm_fan_input_read(ddat, val); + + return -EOPNOTSUPP; +} + +static umode_t hwm_is_visible(const void *drvdata, enum hwmon_sensor_types type, u32 attr, int channel) { @@ -628,6 +700,8 @@ hwm_is_visible(const void *drvdata, enum hwmon_sensor_types type, return hwm_energy_is_visible(ddat, attr); case hwmon_curr: return hwm_curr_is_visible(ddat, attr); + case hwmon_fan: + return hwm_fan_is_visible(ddat, attr); default: return 0; } @@ -648,6 +722,8 @@ hwm_read(struct device *dev, enum hwmon_sensor_types type, u32 attr, return hwm_energy_read(ddat, attr, val); case hwmon_curr: return hwm_curr_read(ddat, attr, val); + case hwmon_fan: + return hwm_fan_read(ddat, attr, val); default: return -EOPNOTSUPP; } @@ -739,12 +815,14 @@ hwm_get_preregistration_info(struct drm_i915_private *i915) hwmon->rg.pkg_rapl_limit = PCU_PACKAGE_RAPL_LIMIT; hwmon->rg.energy_status_all = PCU_PACKAGE_ENERGY_STATUS; hwmon->rg.energy_status_tile = INVALID_MMIO_REG; + hwmon->rg.fan_speed = PCU_PWM_FAN_SPEED; } else { hwmon->rg.pkg_power_sku_unit = INVALID_MMIO_REG; hwmon->rg.pkg_power_sku = INVALID_MMIO_REG; hwmon->rg.pkg_rapl_limit = INVALID_MMIO_REG; hwmon->rg.energy_status_all = INVALID_MMIO_REG; hwmon->rg.energy_status_tile = INVALID_MMIO_REG; + hwmon->rg.fan_speed = INVALID_MMIO_REG; } with_intel_runtime_pm(uncore->rpm, wakeref) { @@ -755,6 +833,16 @@ hwm_get_preregistration_info(struct drm_i915_private *i915) if (i915_mmio_reg_valid(hwmon->rg.pkg_power_sku_unit)) val_sku_unit = intel_uncore_read(uncore, hwmon->rg.pkg_power_sku_unit); + + /* + * Store the initial fan register value, so that we can use it for + * initial fan speed calculation. + */ + if (i915_mmio_reg_valid(hwmon->rg.fan_speed)) { + ddat->fi.reg_val_prev = intel_uncore_read(uncore, + hwmon->rg.fan_speed); + ddat->fi.time_prev = get_jiffies_64(); + } } hwmon->scl_shift_power = REG_FIELD_GET(PKG_PWR_UNIT, val_sku_unit); diff --git a/drivers/gpu/drm/i915/i915_module.c b/drivers/gpu/drm/i915/i915_module.c index 65acd7bf75d0..7ed6d70389af 100644 --- a/drivers/gpu/drm/i915/i915_module.c +++ b/drivers/gpu/drm/i915/i915_module.c @@ -30,15 +30,20 @@ static int i915_check_nomodeset(void) */ if (i915_modparams.modeset == 0) + pr_warn("i915.modeset=0 is deprecated. Please use the 'nomodeset' kernel parameter instead.\n"); + else if (i915_modparams.modeset != -1) + pr_warn("i915.modeset=%d is deprecated. Please remove it and the 'nomodeset' kernel parameter instead.\n", + i915_modparams.modeset); + + if (i915_modparams.modeset == 0) use_kms = false; if (drm_firmware_drivers_only() && i915_modparams.modeset == -1) use_kms = false; if (!use_kms) { - /* Silently fail loading to not upset userspace. */ DRM_DEBUG_DRIVER("KMS disabled.\n"); - return 1; + return -ENODEV; } return 0; diff --git a/drivers/gpu/drm/i915/i915_params.c b/drivers/gpu/drm/i915/i915_params.c index 316e55f3e87b..37746dd619fd 100644 --- a/drivers/gpu/drm/i915/i915_params.c +++ b/drivers/gpu/drm/i915/i915_params.c @@ -64,8 +64,7 @@ struct i915_params i915_modparams __read_mostly = { */ i915_param_named(modeset, int, 0400, - "Use kernel modesetting [KMS] (0=disable, " - "1=on, -1=force vga console preference [default])"); + "Deprecated. Use the 'nomodeset' kernel parameter instead."); i915_param_named_unsafe(reset, uint, 0400, "Attempt GPU resets (0=disabled, 1=full gpu reset, 2=engine reset [default])"); diff --git a/drivers/gpu/drm/i915/i915_sw_fence.c b/drivers/gpu/drm/i915/i915_sw_fence.c index 8a9aad523eec..1d4cc91c0e40 100644 --- a/drivers/gpu/drm/i915/i915_sw_fence.c +++ b/drivers/gpu/drm/i915/i915_sw_fence.c @@ -51,7 +51,7 @@ static inline void debug_fence_init(struct i915_sw_fence *fence) debug_object_init(fence, &i915_sw_fence_debug_descr); } -static inline void debug_fence_init_onstack(struct i915_sw_fence *fence) +static inline __maybe_unused void debug_fence_init_onstack(struct i915_sw_fence *fence) { debug_object_init_on_stack(fence, &i915_sw_fence_debug_descr); } @@ -77,7 +77,7 @@ static inline void debug_fence_destroy(struct i915_sw_fence *fence) debug_object_destroy(fence, &i915_sw_fence_debug_descr); } -static inline void debug_fence_free(struct i915_sw_fence *fence) +static inline __maybe_unused void debug_fence_free(struct i915_sw_fence *fence) { debug_object_free(fence, &i915_sw_fence_debug_descr); smp_wmb(); /* flush the change in state before reallocation */ @@ -94,7 +94,7 @@ static inline void debug_fence_init(struct i915_sw_fence *fence) { } -static inline void debug_fence_init_onstack(struct i915_sw_fence *fence) +static inline __maybe_unused void debug_fence_init_onstack(struct i915_sw_fence *fence) { } @@ -115,7 +115,7 @@ static inline void debug_fence_destroy(struct i915_sw_fence *fence) { } -static inline void debug_fence_free(struct i915_sw_fence *fence) +static inline __maybe_unused void debug_fence_free(struct i915_sw_fence *fence) { } diff --git a/drivers/gpu/drm/i915/selftests/mock_gem_device.c b/drivers/gpu/drm/i915/selftests/mock_gem_device.c index 0bd29846873b..91794ca17a58 100644 --- a/drivers/gpu/drm/i915/selftests/mock_gem_device.c +++ b/drivers/gpu/drm/i915/selftests/mock_gem_device.c @@ -172,7 +172,7 @@ struct drm_i915_private *mock_gem_device(void) return NULL; } - pci_set_drvdata(pdev, i915); + pci_set_drvdata(pdev, &i915->drm); /* Device parameters start as a copy of module parameters. */ i915_params_copy(&i915->params, &i915_modparams); diff --git a/drivers/gpu/drm/imagination/pvr_vm.c b/drivers/gpu/drm/imagination/pvr_vm.c index ffd466509d0b..39773991710a 100644 --- a/drivers/gpu/drm/imagination/pvr_vm.c +++ b/drivers/gpu/drm/imagination/pvr_vm.c @@ -114,6 +114,8 @@ struct pvr_vm_gpuva { struct drm_gpuva base; }; +#define to_pvr_vm_gpuva(va) container_of_const(va, struct pvr_vm_gpuva, base) + enum pvr_vm_bind_type { PVR_VM_BIND_TYPE_MAP, PVR_VM_BIND_TYPE_UNMAP, @@ -386,6 +388,7 @@ pvr_vm_gpuva_unmap(struct drm_gpuva_op *op, void *op_ctx) drm_gpuva_unmap(&op->unmap); drm_gpuva_unlink(op->unmap.va); + kfree(to_pvr_vm_gpuva(op->unmap.va)); return 0; } @@ -433,6 +436,7 @@ pvr_vm_gpuva_remap(struct drm_gpuva_op *op, void *op_ctx) } drm_gpuva_unlink(op->remap.unmap->va); + kfree(to_pvr_vm_gpuva(op->remap.unmap->va)); return 0; } diff --git a/drivers/gpu/drm/imx/dcss/Kconfig b/drivers/gpu/drm/imx/dcss/Kconfig index 3ffc061d392b..59e3b6a1dff0 100644 --- a/drivers/gpu/drm/imx/dcss/Kconfig +++ b/drivers/gpu/drm/imx/dcss/Kconfig @@ -2,6 +2,8 @@ config DRM_IMX_DCSS tristate "i.MX8MQ DCSS" select IMX_IRQSTEER select DRM_KMS_HELPER + select DRM_DISPLAY_HELPER + select DRM_BRIDGE_CONNECTOR select DRM_GEM_DMA_HELPER select VIDEOMODE_HELPERS depends on DRM && ARCH_MXC && ARM64 diff --git a/drivers/gpu/drm/imx/lcdc/Kconfig b/drivers/gpu/drm/imx/lcdc/Kconfig index 7e57922bbd9d..9c28bb0f4662 100644 --- a/drivers/gpu/drm/imx/lcdc/Kconfig +++ b/drivers/gpu/drm/imx/lcdc/Kconfig @@ -3,5 +3,7 @@ config DRM_IMX_LCDC depends on DRM && (ARCH_MXC || COMPILE_TEST) select DRM_GEM_DMA_HELPER select DRM_KMS_HELPER + select DRM_DISPLAY_HELPER + select DRM_BRIDGE_CONNECTOR help Found on i.MX1, i.MX21, i.MX25 and i.MX27. diff --git a/drivers/gpu/drm/ingenic/Kconfig b/drivers/gpu/drm/ingenic/Kconfig index 3db117c5edd9..8cd7b750dffe 100644 --- a/drivers/gpu/drm/ingenic/Kconfig +++ b/drivers/gpu/drm/ingenic/Kconfig @@ -8,6 +8,8 @@ config DRM_INGENIC select DRM_BRIDGE select DRM_PANEL_BRIDGE select DRM_KMS_HELPER + select DRM_DISPLAY_HELPER + select DRM_BRIDGE_CONNECTOR select DRM_GEM_DMA_HELPER select REGMAP select REGMAP_MMIO diff --git a/drivers/gpu/drm/kmb/Kconfig b/drivers/gpu/drm/kmb/Kconfig index fd011367db1d..e5ae3ec52392 100644 --- a/drivers/gpu/drm/kmb/Kconfig +++ b/drivers/gpu/drm/kmb/Kconfig @@ -3,6 +3,8 @@ config DRM_KMB_DISPLAY depends on DRM depends on ARCH_KEEMBAY || COMPILE_TEST select DRM_KMS_HELPER + select DRM_DISPLAY_HELPER + select DRM_BRIDGE_CONNECTOR select DRM_GEM_DMA_HELPER select DRM_MIPI_DSI help diff --git a/drivers/gpu/drm/mediatek/Kconfig b/drivers/gpu/drm/mediatek/Kconfig index d6449ebae838..417ac8c9af41 100644 --- a/drivers/gpu/drm/mediatek/Kconfig +++ b/drivers/gpu/drm/mediatek/Kconfig @@ -9,6 +9,8 @@ config DRM_MEDIATEK depends on MTK_MMSYS select DRM_GEM_DMA_HELPER if DRM_FBDEV_EMULATION select DRM_KMS_HELPER + select DRM_DISPLAY_HELPER + select DRM_BRIDGE_CONNECTOR select DRM_MIPI_DSI select DRM_PANEL select MEMORY diff --git a/drivers/gpu/drm/mediatek/mtk_crtc.c b/drivers/gpu/drm/mediatek/mtk_crtc.c index 6f34f573e127..175b00e5a253 100644 --- a/drivers/gpu/drm/mediatek/mtk_crtc.c +++ b/drivers/gpu/drm/mediatek/mtk_crtc.c @@ -69,6 +69,8 @@ struct mtk_crtc { /* lock for display hardware access */ struct mutex hw_lock; bool config_updating; + /* lock for config_updating to cmd buffer */ + spinlock_t config_lock; }; struct mtk_crtc_state { @@ -106,51 +108,18 @@ static void mtk_crtc_finish_page_flip(struct mtk_crtc *mtk_crtc) static void mtk_drm_finish_page_flip(struct mtk_crtc *mtk_crtc) { + unsigned long flags; + drm_crtc_handle_vblank(&mtk_crtc->base); + + spin_lock_irqsave(&mtk_crtc->config_lock, flags); if (!mtk_crtc->config_updating && mtk_crtc->pending_needs_vblank) { mtk_crtc_finish_page_flip(mtk_crtc); mtk_crtc->pending_needs_vblank = false; } + spin_unlock_irqrestore(&mtk_crtc->config_lock, flags); } -#if IS_REACHABLE(CONFIG_MTK_CMDQ) -static int mtk_drm_cmdq_pkt_create(struct cmdq_client *client, struct cmdq_pkt *pkt, - size_t size) -{ - struct device *dev; - dma_addr_t dma_addr; - - pkt->va_base = kzalloc(size, GFP_KERNEL); - if (!pkt->va_base) - return -ENOMEM; - - pkt->buf_size = size; - pkt->cl = (void *)client; - - dev = client->chan->mbox->dev; - dma_addr = dma_map_single(dev, pkt->va_base, pkt->buf_size, - DMA_TO_DEVICE); - if (dma_mapping_error(dev, dma_addr)) { - dev_err(dev, "dma map failed, size=%u\n", (u32)(u64)size); - kfree(pkt->va_base); - return -ENOMEM; - } - - pkt->pa_base = dma_addr; - - return 0; -} - -static void mtk_drm_cmdq_pkt_destroy(struct cmdq_pkt *pkt) -{ - struct cmdq_client *client = (struct cmdq_client *)pkt->cl; - - dma_unmap_single(client->chan->mbox->dev, pkt->pa_base, pkt->buf_size, - DMA_TO_DEVICE); - kfree(pkt->va_base); -} -#endif - static void mtk_crtc_destroy(struct drm_crtc *crtc) { struct mtk_crtc *mtk_crtc = to_mtk_crtc(crtc); @@ -158,7 +127,7 @@ static void mtk_crtc_destroy(struct drm_crtc *crtc) mtk_mutex_put(mtk_crtc->mutex); #if IS_REACHABLE(CONFIG_MTK_CMDQ) - mtk_drm_cmdq_pkt_destroy(&mtk_crtc->cmdq_handle); + cmdq_pkt_destroy(&mtk_crtc->cmdq_client, &mtk_crtc->cmdq_handle); if (mtk_crtc->cmdq_client.chan) { mbox_free_channel(mtk_crtc->cmdq_client.chan); @@ -308,12 +277,19 @@ static void ddp_cmdq_cb(struct mbox_client *cl, void *mssg) struct mtk_crtc *mtk_crtc = container_of(cmdq_cl, struct mtk_crtc, cmdq_client); struct mtk_crtc_state *state; unsigned int i; + unsigned long flags; if (data->sta < 0) return; state = to_mtk_crtc_state(mtk_crtc->base.state); + spin_lock_irqsave(&mtk_crtc->config_lock, flags); + if (mtk_crtc->config_updating) { + spin_unlock_irqrestore(&mtk_crtc->config_lock, flags); + goto ddp_cmdq_cb_out; + } + state->pending_config = false; if (mtk_crtc->pending_planes) { @@ -340,6 +316,10 @@ static void ddp_cmdq_cb(struct mbox_client *cl, void *mssg) mtk_crtc->pending_async_planes = false; } + spin_unlock_irqrestore(&mtk_crtc->config_lock, flags); + +ddp_cmdq_cb_out: + mtk_crtc->cmdq_vblank_cnt = 0; wake_up(&mtk_crtc->cb_blocking_queue); } @@ -449,6 +429,7 @@ static void mtk_crtc_ddp_hw_fini(struct mtk_crtc *mtk_crtc) { struct drm_device *drm = mtk_crtc->base.dev; struct drm_crtc *crtc = &mtk_crtc->base; + unsigned long flags; int i; for (i = 0; i < mtk_crtc->ddp_comp_nr; i++) { @@ -480,10 +461,10 @@ static void mtk_crtc_ddp_hw_fini(struct mtk_crtc *mtk_crtc) pm_runtime_put(drm->dev); if (crtc->state->event && !crtc->state->active) { - spin_lock_irq(&crtc->dev->event_lock); + spin_lock_irqsave(&crtc->dev->event_lock, flags); drm_crtc_send_vblank_event(crtc, crtc->state->event); crtc->state->event = NULL; - spin_unlock_irq(&crtc->dev->event_lock); + spin_unlock_irqrestore(&crtc->dev->event_lock, flags); } } @@ -569,9 +550,14 @@ static void mtk_crtc_update_config(struct mtk_crtc *mtk_crtc, bool needs_vblank) struct mtk_drm_private *priv = crtc->dev->dev_private; unsigned int pending_planes = 0, pending_async_planes = 0; int i; + unsigned long flags; mutex_lock(&mtk_crtc->hw_lock); + + spin_lock_irqsave(&mtk_crtc->config_lock, flags); mtk_crtc->config_updating = true; + spin_unlock_irqrestore(&mtk_crtc->config_lock, flags); + if (needs_vblank) mtk_crtc->pending_needs_vblank = true; @@ -607,7 +593,7 @@ static void mtk_crtc_update_config(struct mtk_crtc *mtk_crtc, bool needs_vblank) cmdq_pkt_clear_event(cmdq_handle, mtk_crtc->cmdq_event); cmdq_pkt_wfe(cmdq_handle, mtk_crtc->cmdq_event, false); mtk_crtc_ddp_config(crtc, cmdq_handle); - cmdq_pkt_finalize(cmdq_handle); + cmdq_pkt_eoc(cmdq_handle); dma_sync_single_for_device(mtk_crtc->cmdq_client.chan->mbox->dev, cmdq_handle->pa_base, cmdq_handle->cmd_buf_size, @@ -625,7 +611,10 @@ static void mtk_crtc_update_config(struct mtk_crtc *mtk_crtc, bool needs_vblank) mbox_client_txdone(mtk_crtc->cmdq_client.chan, 0); } #endif + spin_lock_irqsave(&mtk_crtc->config_lock, flags); mtk_crtc->config_updating = false; + spin_unlock_irqrestore(&mtk_crtc->config_lock, flags); + mutex_unlock(&mtk_crtc->hw_lock); } @@ -925,7 +914,7 @@ static int mtk_crtc_init_comp_planes(struct drm_device *drm_dev, mtk_crtc_plane_type(mtk_crtc->layer_nr, num_planes), mtk_ddp_comp_supported_rotations(comp), mtk_ddp_comp_get_formats(comp), - mtk_ddp_comp_get_num_formats(comp)); + mtk_ddp_comp_get_num_formats(comp), i); if (ret) return ret; @@ -1068,6 +1057,7 @@ int mtk_crtc_create(struct drm_device *drm_dev, const unsigned int *path, drm_mode_crtc_set_gamma_size(&mtk_crtc->base, gamma_lut_size); drm_crtc_enable_color_mgmt(&mtk_crtc->base, 0, has_ctm, gamma_lut_size); mutex_init(&mtk_crtc->hw_lock); + spin_lock_init(&mtk_crtc->config_lock); #if IS_REACHABLE(CONFIG_MTK_CMDQ) i = priv->mbox_index++; @@ -1094,9 +1084,9 @@ int mtk_crtc_create(struct drm_device *drm_dev, const unsigned int *path, mbox_free_channel(mtk_crtc->cmdq_client.chan); mtk_crtc->cmdq_client.chan = NULL; } else { - ret = mtk_drm_cmdq_pkt_create(&mtk_crtc->cmdq_client, - &mtk_crtc->cmdq_handle, - PAGE_SIZE); + ret = cmdq_pkt_create(&mtk_crtc->cmdq_client, + &mtk_crtc->cmdq_handle, + PAGE_SIZE); if (ret) { dev_dbg(dev, "mtk_crtc %d failed to create cmdq packet\n", drm_crtc_index(&mtk_crtc->base)); diff --git a/drivers/gpu/drm/mediatek/mtk_disp_ovl.c b/drivers/gpu/drm/mediatek/mtk_disp_ovl.c index 9d6d9fd8342e..89b439dcf3a6 100644 --- a/drivers/gpu/drm/mediatek/mtk_disp_ovl.c +++ b/drivers/gpu/drm/mediatek/mtk_disp_ovl.c @@ -56,8 +56,12 @@ #define GMC_THRESHOLD_HIGH ((1 << GMC_THRESHOLD_BITS) / 4) #define GMC_THRESHOLD_LOW ((1 << GMC_THRESHOLD_BITS) / 8) +#define OVL_CON_CLRFMT_MAN BIT(23) #define OVL_CON_BYTE_SWAP BIT(24) -#define OVL_CON_MTX_YUV_TO_RGB (6 << 16) + +/* OVL_CON_RGB_SWAP works only if OVL_CON_CLRFMT_MAN is enabled */ +#define OVL_CON_RGB_SWAP BIT(25) + #define OVL_CON_CLRFMT_RGB (1 << 12) #define OVL_CON_CLRFMT_ARGB8888 (2 << 12) #define OVL_CON_CLRFMT_RGBA8888 (3 << 12) @@ -65,6 +69,11 @@ #define OVL_CON_CLRFMT_BGRA8888 (OVL_CON_CLRFMT_ARGB8888 | OVL_CON_BYTE_SWAP) #define OVL_CON_CLRFMT_UYVY (4 << 12) #define OVL_CON_CLRFMT_YUYV (5 << 12) +#define OVL_CON_MTX_YUV_TO_RGB (6 << 16) +#define OVL_CON_CLRFMT_PARGB8888 ((3 << 12) | OVL_CON_CLRFMT_MAN) +#define OVL_CON_CLRFMT_PABGR8888 (OVL_CON_CLRFMT_PARGB8888 | OVL_CON_RGB_SWAP) +#define OVL_CON_CLRFMT_PBGRA8888 (OVL_CON_CLRFMT_PARGB8888 | OVL_CON_BYTE_SWAP) +#define OVL_CON_CLRFMT_PRGBA8888 (OVL_CON_CLRFMT_PABGR8888 | OVL_CON_BYTE_SWAP) #define OVL_CON_CLRFMT_RGB565(ovl) ((ovl)->data->fmt_rgb565_is_0 ? \ 0 : OVL_CON_CLRFMT_RGB) #define OVL_CON_CLRFMT_RGB888(ovl) ((ovl)->data->fmt_rgb565_is_0 ? \ @@ -377,7 +386,8 @@ void mtk_ovl_layer_off(struct device *dev, unsigned int idx, DISP_REG_OVL_RDMA_CTRL(idx)); } -static unsigned int ovl_fmt_convert(struct mtk_disp_ovl *ovl, unsigned int fmt) +static unsigned int ovl_fmt_convert(struct mtk_disp_ovl *ovl, unsigned int fmt, + unsigned int blend_mode) { /* The return value in switch "MEM_MODE_INPUT_FORMAT_XXX" * is defined in mediatek HW data sheet. @@ -398,22 +408,30 @@ static unsigned int ovl_fmt_convert(struct mtk_disp_ovl *ovl, unsigned int fmt) case DRM_FORMAT_RGBA8888: case DRM_FORMAT_RGBX1010102: case DRM_FORMAT_RGBA1010102: - return OVL_CON_CLRFMT_RGBA8888; + return blend_mode == DRM_MODE_BLEND_COVERAGE ? + OVL_CON_CLRFMT_RGBA8888 : + OVL_CON_CLRFMT_PRGBA8888; case DRM_FORMAT_BGRX8888: case DRM_FORMAT_BGRA8888: case DRM_FORMAT_BGRX1010102: case DRM_FORMAT_BGRA1010102: - return OVL_CON_CLRFMT_BGRA8888; + return blend_mode == DRM_MODE_BLEND_COVERAGE ? + OVL_CON_CLRFMT_BGRA8888 : + OVL_CON_CLRFMT_PBGRA8888; case DRM_FORMAT_XRGB8888: case DRM_FORMAT_ARGB8888: case DRM_FORMAT_XRGB2101010: case DRM_FORMAT_ARGB2101010: - return OVL_CON_CLRFMT_ARGB8888; + return blend_mode == DRM_MODE_BLEND_COVERAGE ? + OVL_CON_CLRFMT_ARGB8888 : + OVL_CON_CLRFMT_PARGB8888; case DRM_FORMAT_XBGR8888: case DRM_FORMAT_ABGR8888: case DRM_FORMAT_XBGR2101010: case DRM_FORMAT_ABGR2101010: - return OVL_CON_CLRFMT_ABGR8888; + return blend_mode == DRM_MODE_BLEND_COVERAGE ? + OVL_CON_CLRFMT_ABGR8888 : + OVL_CON_CLRFMT_PABGR8888; case DRM_FORMAT_UYVY: return OVL_CON_CLRFMT_UYVY | OVL_CON_MTX_YUV_TO_RGB; case DRM_FORMAT_YUYV: @@ -434,6 +452,7 @@ void mtk_ovl_layer_config(struct device *dev, unsigned int idx, unsigned int fmt = pending->format; unsigned int offset = (pending->y << 16) | pending->x; unsigned int src_size = (pending->height << 16) | pending->width; + unsigned int blend_mode = state->base.pixel_blend_mode; unsigned int ignore_pixel_alpha = 0; unsigned int con; bool is_afbc = pending->modifier != DRM_FORMAT_MOD_LINEAR; @@ -452,7 +471,7 @@ void mtk_ovl_layer_config(struct device *dev, unsigned int idx, return; } - con = ovl_fmt_convert(ovl, fmt); + con = ovl_fmt_convert(ovl, fmt, blend_mode); if (state->base.fb) { con |= OVL_CON_AEN; con |= state->base.alpha & OVL_CON_ALPHA; @@ -463,7 +482,8 @@ void mtk_ovl_layer_config(struct device *dev, unsigned int idx, * For RGB888 related formats, whether CONST_BLD is enabled or not won't * affect the result. Therefore we use !has_alpha as the condition. */ - if (state->base.fb && !state->base.fb->format->has_alpha) + if ((state->base.fb && !state->base.fb->format->has_alpha) || + blend_mode == DRM_MODE_BLEND_PIXEL_NONE) ignore_pixel_alpha = OVL_CONST_BLEND; if (pending->rotation & DRM_MODE_REFLECT_Y) { diff --git a/drivers/gpu/drm/mediatek/mtk_disp_ovl_adaptor.c b/drivers/gpu/drm/mediatek/mtk_disp_ovl_adaptor.c index 1a2a73757370..c6768210b08b 100644 --- a/drivers/gpu/drm/mediatek/mtk_disp_ovl_adaptor.c +++ b/drivers/gpu/drm/mediatek/mtk_disp_ovl_adaptor.c @@ -17,7 +17,6 @@ #include <linux/soc/mediatek/mtk-mmsys.h> #include <linux/soc/mediatek/mtk-mutex.h> -#include "mtk_crtc.h" #include "mtk_ddp_comp.h" #include "mtk_disp_drv.h" #include "mtk_drm_drv.h" @@ -494,12 +493,12 @@ static int compare_of(struct device *dev, void *data) static int ovl_adaptor_comp_init(struct device *dev, struct component_match **match) { struct mtk_disp_ovl_adaptor *priv = dev_get_drvdata(dev); - struct device_node *node, *parent; + struct device_node *parent; struct platform_device *comp_pdev; parent = dev->parent->parent->of_node->parent; - for_each_child_of_node(parent, node) { + for_each_child_of_node_scoped(parent, node) { const struct of_device_id *of_id; enum mtk_ovl_adaptor_comp_type type; int id; diff --git a/drivers/gpu/drm/mediatek/mtk_disp_rdma.c b/drivers/gpu/drm/mediatek/mtk_disp_rdma.c index 634bbba5d43f..07243f372260 100644 --- a/drivers/gpu/drm/mediatek/mtk_disp_rdma.c +++ b/drivers/gpu/drm/mediatek/mtk_disp_rdma.c @@ -341,14 +341,11 @@ static int mtk_disp_rdma_probe(struct platform_device *pdev) dev_dbg(dev, "get mediatek,gce-client-reg fail!\n"); #endif - if (of_find_property(dev->of_node, "mediatek,rdma-fifo-size", &ret)) { - ret = of_property_read_u32(dev->of_node, - "mediatek,rdma-fifo-size", - &priv->fifo_size); - if (ret) - return dev_err_probe(dev, ret, - "Failed to get rdma fifo size\n"); - } + ret = of_property_read_u32(dev->of_node, + "mediatek,rdma-fifo-size", + &priv->fifo_size); + if (ret && (ret != -EINVAL)) + return dev_err_probe(dev, ret, "Failed to get rdma fifo size\n"); /* Disable and clear pending interrupts */ writel(0x0, priv->regs + DISP_REG_RDMA_INT_ENABLE); diff --git a/drivers/gpu/drm/mediatek/mtk_dsi.c b/drivers/gpu/drm/mediatek/mtk_dsi.c index b6e3c011a12d..eeec641cab60 100644 --- a/drivers/gpu/drm/mediatek/mtk_dsi.c +++ b/drivers/gpu/drm/mediatek/mtk_dsi.c @@ -88,12 +88,15 @@ #define DSI_HSA_WC 0x50 #define DSI_HBP_WC 0x54 #define DSI_HFP_WC 0x58 +#define HFP_HS_VB_PS_WC GENMASK(30, 16) +#define HFP_HS_EN BIT(31) #define DSI_CMDQ_SIZE 0x60 #define CMDQ_SIZE 0x3f #define CMDQ_SIZE_SEL BIT(15) #define DSI_HSTX_CKL_WC 0x64 +#define HSTX_CKL_WC GENMASK(15, 2) #define DSI_RX_DATA0 0x74 #define DSI_RX_DATA1 0x78 @@ -187,6 +190,7 @@ struct mtk_dsi_driver_data { bool has_shadow_ctl; bool has_size_ctl; bool cmdq_long_packet_ctl; + bool support_per_frame_lp; }; struct mtk_dsi { @@ -426,7 +430,75 @@ static void mtk_dsi_ps_control(struct mtk_dsi *dsi, bool config_vact) writel(ps_val, dsi->regs + DSI_PSCTRL); } -static void mtk_dsi_config_vdo_timing(struct mtk_dsi *dsi) +static void mtk_dsi_config_vdo_timing_per_frame_lp(struct mtk_dsi *dsi) +{ + u32 horizontal_sync_active_byte; + u32 horizontal_backporch_byte; + u32 horizontal_frontporch_byte; + u32 hfp_byte_adjust, v_active_adjust; + u32 cklp_wc_min_adjust, cklp_wc_max_adjust; + u32 dsi_tmp_buf_bpp; + unsigned int da_hs_trail; + unsigned int ps_wc, hs_vb_ps_wc; + u32 v_active_roundup, hstx_cklp_wc; + u32 hstx_cklp_wc_max, hstx_cklp_wc_min; + struct videomode *vm = &dsi->vm; + + if (dsi->format == MIPI_DSI_FMT_RGB565) + dsi_tmp_buf_bpp = 2; + else + dsi_tmp_buf_bpp = 3; + + da_hs_trail = dsi->phy_timing.da_hs_trail; + ps_wc = vm->hactive * dsi_tmp_buf_bpp; + + if (dsi->mode_flags & MIPI_DSI_MODE_VIDEO_SYNC_PULSE) { + horizontal_sync_active_byte = + vm->hsync_len * dsi_tmp_buf_bpp - 10; + horizontal_backporch_byte = + vm->hback_porch * dsi_tmp_buf_bpp - 10; + hfp_byte_adjust = 12; + v_active_adjust = 32 + horizontal_sync_active_byte; + cklp_wc_min_adjust = 12 + 2 + 4 + horizontal_sync_active_byte; + cklp_wc_max_adjust = 20 + 6 + 4 + horizontal_sync_active_byte; + } else { + horizontal_sync_active_byte = vm->hsync_len * dsi_tmp_buf_bpp - 4; + horizontal_backporch_byte = (vm->hback_porch + vm->hsync_len) * + dsi_tmp_buf_bpp - 10; + cklp_wc_min_adjust = 4; + cklp_wc_max_adjust = 12 + 4 + 4; + if (dsi->mode_flags & MIPI_DSI_MODE_VIDEO_BURST) { + hfp_byte_adjust = 18; + v_active_adjust = 28; + } else { + hfp_byte_adjust = 12; + v_active_adjust = 22; + } + } + horizontal_frontporch_byte = vm->hfront_porch * dsi_tmp_buf_bpp - hfp_byte_adjust; + v_active_roundup = (v_active_adjust + horizontal_backporch_byte + ps_wc + + horizontal_frontporch_byte) % dsi->lanes; + if (v_active_roundup) + horizontal_backporch_byte += dsi->lanes - v_active_roundup; + hstx_cklp_wc_min = (DIV_ROUND_UP(cklp_wc_min_adjust, dsi->lanes) + da_hs_trail + 1) + * dsi->lanes / 6 - 1; + hstx_cklp_wc_max = (DIV_ROUND_UP((cklp_wc_max_adjust + horizontal_backporch_byte + + ps_wc), dsi->lanes) + da_hs_trail + 1) * dsi->lanes / 6 - 1; + + hstx_cklp_wc = FIELD_PREP(HSTX_CKL_WC, (hstx_cklp_wc_min + hstx_cklp_wc_max) / 2); + writel(hstx_cklp_wc, dsi->regs + DSI_HSTX_CKL_WC); + + hs_vb_ps_wc = ps_wc - (dsi->phy_timing.lpx + dsi->phy_timing.da_hs_exit + + dsi->phy_timing.da_hs_prepare + dsi->phy_timing.da_hs_zero + 2) * dsi->lanes; + horizontal_frontporch_byte |= FIELD_PREP(HFP_HS_EN, 1) | + FIELD_PREP(HFP_HS_VB_PS_WC, hs_vb_ps_wc); + + writel(horizontal_sync_active_byte, dsi->regs + DSI_HSA_WC); + writel(horizontal_backporch_byte, dsi->regs + DSI_HBP_WC); + writel(horizontal_frontporch_byte, dsi->regs + DSI_HFP_WC); +} + +static void mtk_dsi_config_vdo_timing_per_line_lp(struct mtk_dsi *dsi) { u32 horizontal_sync_active_byte; u32 horizontal_backporch_byte; @@ -436,7 +508,6 @@ static void mtk_dsi_config_vdo_timing(struct mtk_dsi *dsi) u32 dsi_tmp_buf_bpp, data_phy_cycles; u32 delta; struct mtk_phy_timing *timing = &dsi->phy_timing; - struct videomode *vm = &dsi->vm; if (dsi->format == MIPI_DSI_FMT_RGB565) @@ -444,16 +515,6 @@ static void mtk_dsi_config_vdo_timing(struct mtk_dsi *dsi) else dsi_tmp_buf_bpp = 3; - writel(vm->vsync_len, dsi->regs + DSI_VSA_NL); - writel(vm->vback_porch, dsi->regs + DSI_VBP_NL); - writel(vm->vfront_porch, dsi->regs + DSI_VFP_NL); - writel(vm->vactive, dsi->regs + DSI_VACT_NL); - - if (dsi->driver_data->has_size_ctl) - writel(FIELD_PREP(DSI_HEIGHT, vm->vactive) | - FIELD_PREP(DSI_WIDTH, vm->hactive), - dsi->regs + DSI_SIZE_CON); - horizontal_sync_active_byte = (vm->hsync_len * dsi_tmp_buf_bpp - 10); if (dsi->mode_flags & MIPI_DSI_MODE_VIDEO_SYNC_PULSE) @@ -499,6 +560,26 @@ static void mtk_dsi_config_vdo_timing(struct mtk_dsi *dsi) writel(horizontal_sync_active_byte, dsi->regs + DSI_HSA_WC); writel(horizontal_backporch_byte, dsi->regs + DSI_HBP_WC); writel(horizontal_frontporch_byte, dsi->regs + DSI_HFP_WC); +} + +static void mtk_dsi_config_vdo_timing(struct mtk_dsi *dsi) +{ + struct videomode *vm = &dsi->vm; + + writel(vm->vsync_len, dsi->regs + DSI_VSA_NL); + writel(vm->vback_porch, dsi->regs + DSI_VBP_NL); + writel(vm->vfront_porch, dsi->regs + DSI_VFP_NL); + writel(vm->vactive, dsi->regs + DSI_VACT_NL); + + if (dsi->driver_data->has_size_ctl) + writel(FIELD_PREP(DSI_HEIGHT, vm->vactive) | + FIELD_PREP(DSI_WIDTH, vm->hactive), + dsi->regs + DSI_SIZE_CON); + + if (dsi->driver_data->support_per_frame_lp) + mtk_dsi_config_vdo_timing_per_frame_lp(dsi); + else + mtk_dsi_config_vdo_timing_per_line_lp(dsi); mtk_dsi_ps_control(dsi, false); } @@ -1197,6 +1278,7 @@ static const struct mtk_dsi_driver_data mt8188_dsi_driver_data = { .has_shadow_ctl = true, .has_size_ctl = true, .cmdq_long_packet_ctl = true, + .support_per_frame_lp = true, }; static const struct of_device_id mtk_dsi_of_match[] = { diff --git a/drivers/gpu/drm/mediatek/mtk_ethdr.c b/drivers/gpu/drm/mediatek/mtk_ethdr.c index 9dfd13d32dfa..d1d9cf8b10e1 100644 --- a/drivers/gpu/drm/mediatek/mtk_ethdr.c +++ b/drivers/gpu/drm/mediatek/mtk_ethdr.c @@ -3,6 +3,7 @@ * Copyright (c) 2021 MediaTek Inc. */ +#include <drm/drm_blend.h> #include <drm/drm_fourcc.h> #include <drm/drm_framebuffer.h> #include <linux/clk.h> @@ -35,6 +36,7 @@ #define MIX_SRC_L0_EN BIT(0) #define MIX_L_SRC_CON(n) (0x28 + 0x18 * (n)) #define NON_PREMULTI_SOURCE (2 << 12) +#define PREMULTI_SOURCE (3 << 12) #define MIX_L_SRC_SIZE(n) (0x30 + 0x18 * (n)) #define MIX_L_SRC_OFFSET(n) (0x34 + 0x18 * (n)) #define MIX_FUNC_DCM0 0x120 @@ -175,7 +177,13 @@ void mtk_ethdr_layer_config(struct device *dev, unsigned int idx, alpha_con |= state->base.alpha & MIXER_ALPHA; } - if (state->base.fb && !state->base.fb->format->has_alpha) { + if (state->base.pixel_blend_mode == DRM_MODE_BLEND_PREMULTI) + alpha_con |= PREMULTI_SOURCE; + else + alpha_con |= NON_PREMULTI_SOURCE; + + if ((state->base.fb && !state->base.fb->format->has_alpha) || + state->base.pixel_blend_mode == DRM_MODE_BLEND_PIXEL_NONE) { /* * Mixer doesn't support CONST_BLD mode, * use a trick to make the output equivalent @@ -191,8 +199,7 @@ void mtk_ethdr_layer_config(struct device *dev, unsigned int idx, mtk_ddp_write(cmdq_pkt, pending->height << 16 | align_width, &mixer->cmdq_base, mixer->regs, MIX_L_SRC_SIZE(idx)); mtk_ddp_write(cmdq_pkt, offset, &mixer->cmdq_base, mixer->regs, MIX_L_SRC_OFFSET(idx)); - mtk_ddp_write_mask(cmdq_pkt, alpha_con, &mixer->cmdq_base, mixer->regs, MIX_L_SRC_CON(idx), - 0x1ff); + mtk_ddp_write(cmdq_pkt, alpha_con, &mixer->cmdq_base, mixer->regs, MIX_L_SRC_CON(idx)); mtk_ddp_write_mask(cmdq_pkt, BIT(idx), &mixer->cmdq_base, mixer->regs, MIX_SRC_CON, BIT(idx)); } diff --git a/drivers/gpu/drm/mediatek/mtk_plane.c b/drivers/gpu/drm/mediatek/mtk_plane.c index 1723d4333f37..7d2cb4e0fafa 100644 --- a/drivers/gpu/drm/mediatek/mtk_plane.c +++ b/drivers/gpu/drm/mediatek/mtk_plane.c @@ -321,7 +321,7 @@ static const struct drm_plane_helper_funcs mtk_plane_helper_funcs = { int mtk_plane_init(struct drm_device *dev, struct drm_plane *plane, unsigned long possible_crtcs, enum drm_plane_type type, unsigned int supported_rotations, const u32 *formats, - size_t num_formats) + size_t num_formats, unsigned int plane_idx) { int err; @@ -338,6 +338,22 @@ int mtk_plane_init(struct drm_device *dev, struct drm_plane *plane, return err; } + /* + * The hardware does not support repositioning planes by muxing: their + * Z-position is infact fixed and the only way to change the actual + * order is to swap the contents of the entire register set of one + * overlay with another, which may be more expensive than desired. + * + * With no repositioning, the caller of this function guarantees that + * the plane_idx is correct. This means that, for example, the PRIMARY + * plane fed to this function will always have plane_idx zero. + */ + err = drm_plane_create_zpos_immutable_property(plane, plane_idx); + if (err) { + DRM_ERROR("Failed to create zpos property for plane %u\n", plane_idx); + return err; + } + if (supported_rotations) { err = drm_plane_create_rotation_property(plane, DRM_MODE_ROTATE_0, @@ -346,6 +362,17 @@ int mtk_plane_init(struct drm_device *dev, struct drm_plane *plane, DRM_INFO("Create rotation property failed\n"); } + err = drm_plane_create_alpha_property(plane); + if (err) + DRM_ERROR("failed to create property: alpha\n"); + + err = drm_plane_create_blend_mode_property(plane, + BIT(DRM_MODE_BLEND_PREMULTI) | + BIT(DRM_MODE_BLEND_COVERAGE) | + BIT(DRM_MODE_BLEND_PIXEL_NONE)); + if (err) + DRM_ERROR("failed to create property: blend_mode\n"); + drm_plane_helper_add(plane, &mtk_plane_helper_funcs); return 0; diff --git a/drivers/gpu/drm/mediatek/mtk_plane.h b/drivers/gpu/drm/mediatek/mtk_plane.h index 231bb7aac947..5b177eac67b7 100644 --- a/drivers/gpu/drm/mediatek/mtk_plane.h +++ b/drivers/gpu/drm/mediatek/mtk_plane.h @@ -49,6 +49,5 @@ to_mtk_plane_state(struct drm_plane_state *state) int mtk_plane_init(struct drm_device *dev, struct drm_plane *plane, unsigned long possible_crtcs, enum drm_plane_type type, unsigned int supported_rotations, const u32 *formats, - size_t num_formats); - + size_t num_formats, unsigned int plane_idx); #endif diff --git a/drivers/gpu/drm/meson/Kconfig b/drivers/gpu/drm/meson/Kconfig index 615fdd0ce41b..2544756538cc 100644 --- a/drivers/gpu/drm/meson/Kconfig +++ b/drivers/gpu/drm/meson/Kconfig @@ -4,6 +4,8 @@ config DRM_MESON depends on DRM && OF && (ARM || ARM64) depends on ARCH_MESON || COMPILE_TEST select DRM_KMS_HELPER + select DRM_DISPLAY_HELPER + select DRM_BRIDGE_CONNECTOR select DRM_GEM_DMA_HELPER select DRM_DISPLAY_CONNECTOR select VIDEOMODE_HELPERS diff --git a/drivers/gpu/drm/msm/Kconfig b/drivers/gpu/drm/msm/Kconfig index 8bfebd579a74..c8dda0ebd043 100644 --- a/drivers/gpu/drm/msm/Kconfig +++ b/drivers/gpu/drm/msm/Kconfig @@ -17,6 +17,7 @@ config DRM_MSM select DRM_DISPLAY_DP_AUX_BUS select DRM_DISPLAY_DP_HELPER select DRM_DISPLAY_HELPER + select DRM_BRIDGE_CONNECTOR select DRM_EXEC select DRM_KMS_HELPER select DRM_PANEL diff --git a/drivers/gpu/drm/msm/Makefile b/drivers/gpu/drm/msm/Makefile index f5e2838c6a76..13110fcc46a8 100644 --- a/drivers/gpu/drm/msm/Makefile +++ b/drivers/gpu/drm/msm/Makefile @@ -37,6 +37,7 @@ msm-display-$(CONFIG_DRM_MSM_HDMI) += \ hdmi/hdmi_phy.o \ hdmi/hdmi_phy_8960.o \ hdmi/hdmi_phy_8996.o \ + hdmi/hdmi_phy_8998.o \ hdmi/hdmi_phy_8x60.o \ hdmi/hdmi_phy_8x74.o \ hdmi/hdmi_pll_8960.o \ diff --git a/drivers/gpu/drm/msm/adreno/a3xx_catalog.c b/drivers/gpu/drm/msm/adreno/a3xx_catalog.c index 0de8465b6cf0..2eb6c3e93748 100644 --- a/drivers/gpu/drm/msm/adreno/a3xx_catalog.c +++ b/drivers/gpu/drm/msm/adreno/a3xx_catalog.c @@ -42,6 +42,17 @@ static const struct adreno_info a3xx_gpus[] = { .inactive_period = DRM_MSM_INACTIVE_PERIOD, .init = a3xx_gpu_init, }, { + .chip_ids = ADRENO_CHIP_IDS(0x03000620), + .family = ADRENO_3XX, + .revn = 308, + .fw = { + [ADRENO_FW_PM4] = "a300_pm4.fw", + [ADRENO_FW_PFP] = "a300_pfp.fw", + }, + .gmem = SZ_128K, + .inactive_period = DRM_MSM_INACTIVE_PERIOD, + .init = a3xx_gpu_init, + }, { .chip_ids = ADRENO_CHIP_IDS( 0x03020000, 0x03020001, diff --git a/drivers/gpu/drm/msm/adreno/a3xx_gpu.c b/drivers/gpu/drm/msm/adreno/a3xx_gpu.c index 5273dc849838..b46ff49f47cf 100644 --- a/drivers/gpu/drm/msm/adreno/a3xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a3xx_gpu.c @@ -145,6 +145,10 @@ static int a3xx_hw_init(struct msm_gpu *gpu) gpu_write(gpu, REG_A3XX_VBIF_ROUND_ROBIN_QOS_ARB, 0x0003); gpu_write(gpu, REG_A3XX_VBIF_OUT_RD_LIM_CONF0, 0x0000000a); gpu_write(gpu, REG_A3XX_VBIF_OUT_WR_LIM_CONF0, 0x0000000a); + } else if (adreno_is_a306a(adreno_gpu)) { + gpu_write(gpu, REG_A3XX_VBIF_ROUND_ROBIN_QOS_ARB, 0x0003); + gpu_write(gpu, REG_A3XX_VBIF_OUT_RD_LIM_CONF0, 0x00000010); + gpu_write(gpu, REG_A3XX_VBIF_OUT_WR_LIM_CONF0, 0x00000010); } else if (adreno_is_a320(adreno_gpu)) { /* Set up 16 deep read/write request queues: */ gpu_write(gpu, REG_A3XX_VBIF_IN_RD_LIM_CONF0, 0x10101010); @@ -237,7 +241,9 @@ static int a3xx_hw_init(struct msm_gpu *gpu) gpu_write(gpu, REG_A3XX_UCHE_CACHE_MODE_CONTROL_REG, 0x00000001); /* Enable Clock gating: */ - if (adreno_is_a305b(adreno_gpu) || adreno_is_a306(adreno_gpu)) + if (adreno_is_a305b(adreno_gpu) || + adreno_is_a306(adreno_gpu) || + adreno_is_a306a(adreno_gpu)) gpu_write(gpu, REG_A3XX_RBBM_CLOCK_CTL, 0xaaaaaaaa); else if (adreno_is_a320(adreno_gpu)) gpu_write(gpu, REG_A3XX_RBBM_CLOCK_CTL, 0xbfffffff); @@ -334,8 +340,10 @@ static int a3xx_hw_init(struct msm_gpu *gpu) gpu_write(gpu, REG_A3XX_CP_PFP_UCODE_DATA, ptr[i]); /* CP ROQ queue sizes (bytes) - RB:16, ST:16, IB1:32, IB2:64 */ - if (adreno_is_a305(adreno_gpu) || adreno_is_a306(adreno_gpu) || - adreno_is_a320(adreno_gpu)) { + if (adreno_is_a305(adreno_gpu) || + adreno_is_a306(adreno_gpu) || + adreno_is_a306a(adreno_gpu) || + adreno_is_a320(adreno_gpu)) { gpu_write(gpu, REG_AXXX_CP_QUEUE_THRESHOLDS, AXXX_CP_QUEUE_THRESHOLDS_CSQ_IB1_START(2) | AXXX_CP_QUEUE_THRESHOLDS_CSQ_IB2_START(6) | diff --git a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c index c0b5373e90d7..e09044930547 100644 --- a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c @@ -65,6 +65,8 @@ void a5xx_flush(struct msm_gpu *gpu, struct msm_ringbuffer *ring, static void a5xx_submit_in_rb(struct msm_gpu *gpu, struct msm_gem_submit *submit) { + struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu); + struct a5xx_gpu *a5xx_gpu = to_a5xx_gpu(adreno_gpu); struct msm_ringbuffer *ring = submit->ring; struct drm_gem_object *obj; uint32_t *ptr, dwords; @@ -109,6 +111,7 @@ static void a5xx_submit_in_rb(struct msm_gpu *gpu, struct msm_gem_submit *submit } } + a5xx_gpu->last_seqno[ring->id] = submit->seqno; a5xx_flush(gpu, ring, true); a5xx_preempt_trigger(gpu); @@ -150,9 +153,13 @@ static void a5xx_submit(struct msm_gpu *gpu, struct msm_gem_submit *submit) OUT_PKT7(ring, CP_SET_PROTECTED_MODE, 1); OUT_RING(ring, 1); - /* Enable local preemption for finegrain preemption */ + /* + * Disable local preemption by default because it requires + * user-space to be aware of it and provide additional handling + * to restore rendering state or do various flushes on switch. + */ OUT_PKT7(ring, CP_PREEMPT_ENABLE_LOCAL, 1); - OUT_RING(ring, 0x1); + OUT_RING(ring, 0x0); /* Allow CP_CONTEXT_SWITCH_YIELD packets in the IB2 */ OUT_PKT7(ring, CP_YIELD_ENABLE, 1); @@ -206,6 +213,7 @@ static void a5xx_submit(struct msm_gpu *gpu, struct msm_gem_submit *submit) /* Write the fence to the scratch register */ OUT_PKT4(ring, REG_A5XX_CP_SCRATCH_REG(2), 1); OUT_RING(ring, submit->seqno); + a5xx_gpu->last_seqno[ring->id] = submit->seqno; /* * Execute a CACHE_FLUSH_TS event. This will ensure that the @@ -1793,5 +1801,9 @@ struct msm_gpu *a5xx_gpu_init(struct drm_device *dev) else adreno_gpu->ubwc_config.highest_bank_bit = 14; + /* a5xx only supports UBWC 1.0, these are not configurable */ + adreno_gpu->ubwc_config.macrotile_mode = 0; + adreno_gpu->ubwc_config.ubwc_swizzle = 0x7; + return gpu; } diff --git a/drivers/gpu/drm/msm/adreno/a5xx_gpu.h b/drivers/gpu/drm/msm/adreno/a5xx_gpu.h index c7187bcc5e90..9c0d701fe4b8 100644 --- a/drivers/gpu/drm/msm/adreno/a5xx_gpu.h +++ b/drivers/gpu/drm/msm/adreno/a5xx_gpu.h @@ -34,8 +34,10 @@ struct a5xx_gpu { struct drm_gem_object *preempt_counters_bo[MSM_GPU_MAX_RINGS]; struct a5xx_preempt_record *preempt[MSM_GPU_MAX_RINGS]; uint64_t preempt_iova[MSM_GPU_MAX_RINGS]; + uint32_t last_seqno[MSM_GPU_MAX_RINGS]; atomic_t preempt_state; + spinlock_t preempt_start_lock; struct timer_list preempt_timer; struct drm_gem_object *shadow_bo; diff --git a/drivers/gpu/drm/msm/adreno/a5xx_preempt.c b/drivers/gpu/drm/msm/adreno/a5xx_preempt.c index f58dd564d122..0469fea55010 100644 --- a/drivers/gpu/drm/msm/adreno/a5xx_preempt.c +++ b/drivers/gpu/drm/msm/adreno/a5xx_preempt.c @@ -55,6 +55,8 @@ static inline void update_wptr(struct msm_gpu *gpu, struct msm_ringbuffer *ring) /* Return the highest priority ringbuffer with something in it */ static struct msm_ringbuffer *get_next_ring(struct msm_gpu *gpu) { + struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu); + struct a5xx_gpu *a5xx_gpu = to_a5xx_gpu(adreno_gpu); unsigned long flags; int i; @@ -64,6 +66,8 @@ static struct msm_ringbuffer *get_next_ring(struct msm_gpu *gpu) spin_lock_irqsave(&ring->preempt_lock, flags); empty = (get_wptr(ring) == gpu->funcs->get_rptr(gpu, ring)); + if (!empty && ring == a5xx_gpu->cur_ring) + empty = ring->memptrs->fence == a5xx_gpu->last_seqno[i]; spin_unlock_irqrestore(&ring->preempt_lock, flags); if (!empty) @@ -98,11 +102,18 @@ void a5xx_preempt_trigger(struct msm_gpu *gpu) return; /* + * Serialize preemption start to ensure that we always make + * decision on latest state. Otherwise we can get stuck in + * lower priority or empty ring. + */ + spin_lock_irqsave(&a5xx_gpu->preempt_start_lock, flags); + + /* * Try to start preemption by moving from NONE to START. If * unsuccessful, a preemption is already in flight */ if (!try_preempt_state(a5xx_gpu, PREEMPT_NONE, PREEMPT_START)) - return; + goto out; /* Get the next ring to preempt to */ ring = get_next_ring(gpu); @@ -127,9 +138,11 @@ void a5xx_preempt_trigger(struct msm_gpu *gpu) set_preempt_state(a5xx_gpu, PREEMPT_ABORT); update_wptr(gpu, a5xx_gpu->cur_ring); set_preempt_state(a5xx_gpu, PREEMPT_NONE); - return; + goto out; } + spin_unlock_irqrestore(&a5xx_gpu->preempt_start_lock, flags); + /* Make sure the wptr doesn't update while we're in motion */ spin_lock_irqsave(&ring->preempt_lock, flags); a5xx_gpu->preempt[ring->id]->wptr = get_wptr(ring); @@ -152,6 +165,10 @@ void a5xx_preempt_trigger(struct msm_gpu *gpu) /* And actually start the preemption */ gpu_write(gpu, REG_A5XX_CP_CONTEXT_SWITCH_CNTL, 1); + return; + +out: + spin_unlock_irqrestore(&a5xx_gpu->preempt_start_lock, flags); } void a5xx_preempt_irq(struct msm_gpu *gpu) @@ -188,6 +205,12 @@ void a5xx_preempt_irq(struct msm_gpu *gpu) update_wptr(gpu, a5xx_gpu->cur_ring); set_preempt_state(a5xx_gpu, PREEMPT_NONE); + + /* + * Try to trigger preemption again in case there was a submit or + * retire during ring switch + */ + a5xx_preempt_trigger(gpu); } void a5xx_preempt_hw_init(struct msm_gpu *gpu) @@ -204,6 +227,8 @@ void a5xx_preempt_hw_init(struct msm_gpu *gpu) return; for (i = 0; i < gpu->nr_rings; i++) { + a5xx_gpu->preempt[i]->data = 0; + a5xx_gpu->preempt[i]->info = 0; a5xx_gpu->preempt[i]->wptr = 0; a5xx_gpu->preempt[i]->rptr = 0; a5xx_gpu->preempt[i]->rbase = gpu->rb[i]->iova; @@ -298,5 +323,6 @@ void a5xx_preempt_init(struct msm_gpu *gpu) } } + spin_lock_init(&a5xx_gpu->preempt_start_lock); timer_setup(&a5xx_gpu->preempt_timer, a5xx_preempt_timer, 0); } diff --git a/drivers/gpu/drm/msm/adreno/a6xx_catalog.c b/drivers/gpu/drm/msm/adreno/a6xx_catalog.c index 68ba9aed5506..0312b6ee0356 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_catalog.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_catalog.c @@ -129,6 +129,59 @@ static const struct adreno_reglist a615_hwcg[] = { {}, }; +static const struct adreno_reglist a620_hwcg[] = { + {REG_A6XX_RBBM_CLOCK_CNTL_SP0, 0x02222222}, + {REG_A6XX_RBBM_CLOCK_CNTL2_SP0, 0x02222220}, + {REG_A6XX_RBBM_CLOCK_DELAY_SP0, 0x00000080}, + {REG_A6XX_RBBM_CLOCK_HYST_SP0, 0x0000f3cf}, + {REG_A6XX_RBBM_CLOCK_CNTL_TP0, 0x02222222}, + {REG_A6XX_RBBM_CLOCK_CNTL2_TP0, 0x22222222}, + {REG_A6XX_RBBM_CLOCK_CNTL3_TP0, 0x22222222}, + {REG_A6XX_RBBM_CLOCK_CNTL4_TP0, 0x00022222}, + {REG_A6XX_RBBM_CLOCK_DELAY_TP0, 0x11111111}, + {REG_A6XX_RBBM_CLOCK_DELAY2_TP0, 0x11111111}, + {REG_A6XX_RBBM_CLOCK_DELAY3_TP0, 0x11111111}, + {REG_A6XX_RBBM_CLOCK_DELAY4_TP0, 0x00011111}, + {REG_A6XX_RBBM_CLOCK_HYST_TP0, 0x77777777}, + {REG_A6XX_RBBM_CLOCK_HYST2_TP0, 0x77777777}, + {REG_A6XX_RBBM_CLOCK_HYST3_TP0, 0x77777777}, + {REG_A6XX_RBBM_CLOCK_HYST4_TP0, 0x00077777}, + {REG_A6XX_RBBM_CLOCK_CNTL_RB0, 0x22222222}, + {REG_A6XX_RBBM_CLOCK_CNTL2_RB0, 0x01002222}, + {REG_A6XX_RBBM_CLOCK_CNTL_CCU0, 0x00002220}, + {REG_A6XX_RBBM_CLOCK_HYST_RB_CCU0, 0x00040f00}, + {REG_A6XX_RBBM_CLOCK_CNTL_RAC, 0x25222022}, + {REG_A6XX_RBBM_CLOCK_CNTL2_RAC, 0x00005555}, + {REG_A6XX_RBBM_CLOCK_DELAY_RAC, 0x00000011}, + {REG_A6XX_RBBM_CLOCK_HYST_RAC, 0x00445044}, + {REG_A6XX_RBBM_CLOCK_CNTL_TSE_RAS_RBBM, 0x04222222}, + {REG_A6XX_RBBM_CLOCK_MODE_VFD, 0x00002222}, + {REG_A6XX_RBBM_CLOCK_MODE_GPC, 0x00222222}, + {REG_A6XX_RBBM_CLOCK_DELAY_HLSQ_2, 0x00000002}, + {REG_A6XX_RBBM_CLOCK_MODE_HLSQ, 0x00002222}, + {REG_A6XX_RBBM_CLOCK_DELAY_TSE_RAS_RBBM, 0x00004000}, + {REG_A6XX_RBBM_CLOCK_DELAY_VFD, 0x00002222}, + {REG_A6XX_RBBM_CLOCK_DELAY_GPC, 0x00000200}, + {REG_A6XX_RBBM_CLOCK_DELAY_HLSQ, 0x00000000}, + {REG_A6XX_RBBM_CLOCK_HYST_TSE_RAS_RBBM, 0x00000000}, + {REG_A6XX_RBBM_CLOCK_HYST_VFD, 0x00000000}, + {REG_A6XX_RBBM_CLOCK_HYST_GPC, 0x04104004}, + {REG_A6XX_RBBM_CLOCK_HYST_HLSQ, 0x00000000}, + {REG_A6XX_RBBM_CLOCK_CNTL_TEX_FCHE, 0x00000222}, + {REG_A6XX_RBBM_CLOCK_DELAY_TEX_FCHE, 0x00000111}, + {REG_A6XX_RBBM_CLOCK_HYST_TEX_FCHE, 0x00000777}, + {REG_A6XX_RBBM_CLOCK_CNTL_UCHE, 0x22222222}, + {REG_A6XX_RBBM_CLOCK_HYST_UCHE, 0x00000004}, + {REG_A6XX_RBBM_CLOCK_DELAY_UCHE, 0x00000002}, + {REG_A6XX_RBBM_ISDB_CNT, 0x00000182}, + {REG_A6XX_RBBM_RAC_THRESHOLD_CNT, 0x00000000}, + {REG_A6XX_RBBM_SP_HYST_CNT, 0x00000000}, + {REG_A6XX_RBBM_CLOCK_CNTL_GMU_GX, 0x00000222}, + {REG_A6XX_RBBM_CLOCK_DELAY_GMU_GX, 0x00000111}, + {REG_A6XX_RBBM_CLOCK_HYST_GMU_GX, 0x00000555}, + {}, +}; + static const struct adreno_reglist a630_hwcg[] = { {REG_A6XX_RBBM_CLOCK_CNTL_SP0, 0x22222222}, {REG_A6XX_RBBM_CLOCK_CNTL_SP1, 0x22222222}, @@ -448,7 +501,6 @@ static const struct adreno_reglist a690_hwcg[] = { {REG_A6XX_RBBM_CLOCK_CNTL_GMU_GX, 0x00000222}, {REG_A6XX_RBBM_CLOCK_DELAY_GMU_GX, 0x00000111}, {REG_A6XX_RBBM_CLOCK_HYST_GMU_GX, 0x00000555}, - {REG_A6XX_GPU_GMU_AO_GMU_CGC_MODE_CNTL, 0x20200}, {REG_A6XX_GPU_GMU_AO_GMU_CGC_DELAY_CNTL, 0x10111}, {REG_A6XX_GPU_GMU_AO_GMU_CGC_HYST_CNTL, 0x5555}, {} @@ -491,7 +543,6 @@ static const u32 a630_protect_regs[] = { }; DECLARE_ADRENO_PROTECT(a630_protect, 32); -/* These are for a620 and a650 */ static const u32 a650_protect_regs[] = { A6XX_PROTECT_RDONLY(0x00000, 0x04ff), A6XX_PROTECT_RDONLY(0x00501, 0x0005), @@ -636,6 +687,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a612_hwcg, .protect = &a630_protect, + .gmu_cgc_mode = 0x00020202, + .prim_fifo_threshold = 0x00080000, }, /* * There are (at least) three SoCs implementing A610: SM6125 @@ -652,6 +705,35 @@ static const struct adreno_info a6xx_gpus[] = { { 127, 4 }, ), }, { + .chip_ids = ADRENO_CHIP_IDS(0x06010500), + .family = ADRENO_6XX_GEN1, + .revn = 615, + .fw = { + [ADRENO_FW_SQE] = "a630_sqe.fw", + [ADRENO_FW_GMU] = "a630_gmu.bin", + }, + .gmem = SZ_512K, + .inactive_period = DRM_MSM_INACTIVE_PERIOD, + .init = a6xx_gpu_init, + .zapfw = "a615_zap.mdt", + .a6xx = &(const struct a6xx_info) { + .hwcg = a615_hwcg, + .protect = &a630_protect, + .gmu_cgc_mode = 0x00000222, + .prim_fifo_threshold = 0x0018000, + }, + .speedbins = ADRENO_SPEEDBINS( + /* + * The default speed bin (0) has the same values as + * speed bin 90 which goes up to 432 MHz. + */ + { 0, 0 }, + { 90, 0 }, + { 105, 1 }, + { 146, 2 }, + { 163, 3 }, + ), + }, { .machine = "qcom,sm7150", .chip_ids = ADRENO_CHIP_IDS(0x06010800), .family = ADRENO_6XX_GEN1, @@ -667,6 +749,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a615_hwcg, .protect = &a630_protect, + .gmu_cgc_mode = 0x00000222, + .prim_fifo_threshold = 0x00180000, }, .speedbins = ADRENO_SPEEDBINS( { 0, 0 }, @@ -689,6 +773,8 @@ static const struct adreno_info a6xx_gpus[] = { .init = a6xx_gpu_init, .a6xx = &(const struct a6xx_info) { .protect = &a630_protect, + .gmu_cgc_mode = 0x00000222, + .prim_fifo_threshold = 0x00180000, }, .speedbins = ADRENO_SPEEDBINS( { 0, 0 }, @@ -711,6 +797,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a615_hwcg, .protect = &a630_protect, + .gmu_cgc_mode = 0x00000222, + .prim_fifo_threshold = 0x00018000, }, .speedbins = ADRENO_SPEEDBINS( { 0, 0 }, @@ -733,6 +821,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a615_hwcg, .protect = &a630_protect, + .gmu_cgc_mode = 0x00000222, + .prim_fifo_threshold = 0x00018000, }, .speedbins = ADRENO_SPEEDBINS( { 0, 0 }, @@ -755,6 +845,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a615_hwcg, .protect = &a630_protect, + .gmu_cgc_mode = 0x00000222, + .prim_fifo_threshold = 0x00018000, }, .speedbins = ADRENO_SPEEDBINS( { 0, 0 }, @@ -764,6 +856,30 @@ static const struct adreno_info a6xx_gpus[] = { { 180, 1 }, ), }, { + .chip_ids = ADRENO_CHIP_IDS(0x06020100), + .family = ADRENO_6XX_GEN3, + .fw = { + [ADRENO_FW_SQE] = "a650_sqe.fw", + [ADRENO_FW_GMU] = "a621_gmu.bin", + }, + .gmem = SZ_512K, + .inactive_period = DRM_MSM_INACTIVE_PERIOD, + .quirks = ADRENO_QUIRK_HAS_CACHED_COHERENT | + ADRENO_QUIRK_HAS_HW_APRIV, + .init = a6xx_gpu_init, + .zapfw = "a620_zap.mbn", + .a6xx = &(const struct a6xx_info) { + .hwcg = a620_hwcg, + .protect = &a650_protect, + .gmu_cgc_mode = 0x00020200, + .prim_fifo_threshold = 0x00010000, + }, + .address_space_size = SZ_16G, + .speedbins = ADRENO_SPEEDBINS( + { 0, 0 }, + { 137, 1 }, + ), + }, { .chip_ids = ADRENO_CHIP_IDS( 0x06030001, 0x06030002 @@ -782,6 +898,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a630_hwcg, .protect = &a630_protect, + .gmu_cgc_mode = 0x00020202, + .prim_fifo_threshold = 0x00180000, }, }, { .chip_ids = ADRENO_CHIP_IDS(0x06040001), @@ -799,6 +917,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a640_hwcg, .protect = &a630_protect, + .gmu_cgc_mode = 0x00020202, + .prim_fifo_threshold = 0x00180000, }, .speedbins = ADRENO_SPEEDBINS( { 0, 0 }, @@ -821,6 +941,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a650_hwcg, .protect = &a650_protect, + .gmu_cgc_mode = 0x00020202, + .prim_fifo_threshold = 0x00300200, }, .address_space_size = SZ_16G, .speedbins = ADRENO_SPEEDBINS( @@ -846,6 +968,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a660_hwcg, .protect = &a660_protect, + .gmu_cgc_mode = 0x00020000, + .prim_fifo_threshold = 0x00300200, }, .address_space_size = SZ_16G, }, { @@ -864,11 +988,14 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a660_hwcg, .protect = &a660_protect, + .gmu_cgc_mode = 0x00020202, + .prim_fifo_threshold = 0x00200200, }, .address_space_size = SZ_16G, .speedbins = ADRENO_SPEEDBINS( { 0, 0 }, { 117, 0 }, + { 129, 4 }, { 172, 2 }, /* Called speedbin 1 downstream, but let's not break things! */ { 190, 1 }, ), @@ -888,6 +1015,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a640_hwcg, .protect = &a630_protect, + .gmu_cgc_mode = 0x00020202, + .prim_fifo_threshold = 0x00200200, }, }, { .chip_ids = ADRENO_CHIP_IDS(0x06090000), @@ -905,6 +1034,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a690_hwcg, .protect = &a690_protect, + .gmu_cgc_mode = 0x00020200, + .prim_fifo_threshold = 0x00800200, }, .address_space_size = SZ_16G, } @@ -1165,6 +1296,8 @@ static const struct adreno_info a7xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a702_hwcg, .protect = &a650_protect, + .gmu_cgc_mode = 0x00020202, + .prim_fifo_threshold = 0x0000c000, }, .speedbins = ADRENO_SPEEDBINS( { 0, 0 }, @@ -1188,6 +1321,7 @@ static const struct adreno_info a7xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a730_hwcg, .protect = &a730_protect, + .gmu_cgc_mode = 0x00020000, }, .address_space_size = SZ_16G, }, { @@ -1207,6 +1341,7 @@ static const struct adreno_info a7xx_gpus[] = { .hwcg = a740_hwcg, .protect = &a730_protect, .gmu_chipid = 0x7020100, + .gmu_cgc_mode = 0x00020202, }, .address_space_size = SZ_16G, }, { @@ -1225,6 +1360,7 @@ static const struct adreno_info a7xx_gpus[] = { .hwcg = a740_hwcg, .protect = &a730_protect, .gmu_chipid = 0x7050001, + .gmu_cgc_mode = 0x00020202, }, .address_space_size = SZ_256G, }, { @@ -1243,6 +1379,7 @@ static const struct adreno_info a7xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .protect = &a730_protect, .gmu_chipid = 0x7090100, + .gmu_cgc_mode = 0x00020202, }, .address_space_size = SZ_16G, } diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c index cb538a262d1c..37927bdd6fbe 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c @@ -423,6 +423,20 @@ static int a6xx_gmu_gfx_rail_on(struct a6xx_gmu *gmu) return a6xx_gmu_set_oob(gmu, GMU_OOB_BOOT_SLUMBER); } +static void a6xx_gemnoc_workaround(struct a6xx_gmu *gmu) +{ + struct a6xx_gpu *a6xx_gpu = container_of(gmu, struct a6xx_gpu, gmu); + struct adreno_gpu *adreno_gpu = &a6xx_gpu->base; + + /* + * GEMNoC can power collapse whilst the GPU is being powered down, resulting + * in the power down sequence not being fully executed. That in turn can + * prevent CX_GDSC from collapsing. Assert Qactive to avoid this. + */ + if (adreno_is_a621(adreno_gpu) || adreno_is_7c3(adreno_gpu)) + gmu_write(gmu, REG_A6XX_GMU_AO_AHB_FENCE_CTRL, BIT(0)); +} + /* Let the GMU know that we are about to go into slumber */ static int a6xx_gmu_notify_slumber(struct a6xx_gmu *gmu) { @@ -456,6 +470,8 @@ static int a6xx_gmu_notify_slumber(struct a6xx_gmu *gmu) } out: + a6xx_gemnoc_workaround(gmu); + /* Put fence into allow mode */ gmu_write(gmu, REG_A6XX_GMU_AO_AHB_FENCE_CTRL, 0); return ret; @@ -525,8 +541,7 @@ static void a6xx_gmu_rpmh_init(struct a6xx_gmu *gmu) if (IS_ERR(pdcptr)) goto err; - if (adreno_is_a650(adreno_gpu) || - adreno_is_a660_family(adreno_gpu) || + if (adreno_is_a650_family(adreno_gpu) || adreno_is_a7xx(adreno_gpu)) pdc_in_aop = true; else if (adreno_is_a618(adreno_gpu) || adreno_is_a640_family(adreno_gpu)) @@ -946,6 +961,8 @@ static void a6xx_gmu_force_off(struct a6xx_gmu *gmu) /* Force off SPTP in case the GMU is managing it */ a6xx_sptprac_disable(gmu); + a6xx_gemnoc_workaround(gmu); + /* Make sure there are no outstanding RPMh votes */ a6xx_gmu_rpmh_off(gmu); diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c index bcaec86ac67a..06cab2c6fd66 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c @@ -402,7 +402,8 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state) struct a6xx_gmu *gmu = &a6xx_gpu->gmu; const struct adreno_reglist *reg; unsigned int i; - u32 val, clock_cntl_on, cgc_mode; + u32 cgc_delay, cgc_hyst; + u32 val, clock_cntl_on; if (!(adreno_gpu->info->a6xx->hwcg || adreno_is_a7xx(adreno_gpu))) return; @@ -416,16 +417,15 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state) else clock_cntl_on = 0x8aa8aa82; - if (adreno_is_a7xx(adreno_gpu)) { - cgc_mode = adreno_is_a740_family(adreno_gpu) ? 0x20222 : 0x20000; - - gmu_write(&a6xx_gpu->gmu, REG_A6XX_GPU_GMU_AO_GMU_CGC_MODE_CNTL, - state ? cgc_mode : 0); - gmu_write(&a6xx_gpu->gmu, REG_A6XX_GPU_GMU_AO_GMU_CGC_DELAY_CNTL, - state ? 0x10111 : 0); - gmu_write(&a6xx_gpu->gmu, REG_A6XX_GPU_GMU_AO_GMU_CGC_HYST_CNTL, - state ? 0x5555 : 0); - } + cgc_delay = adreno_is_a615_family(adreno_gpu) ? 0x111 : 0x10111; + cgc_hyst = adreno_is_a615_family(adreno_gpu) ? 0x555 : 0x5555; + + gmu_write(&a6xx_gpu->gmu, REG_A6XX_GPU_GMU_AO_GMU_CGC_MODE_CNTL, + state ? adreno_gpu->info->a6xx->gmu_cgc_mode : 0); + gmu_write(&a6xx_gpu->gmu, REG_A6XX_GPU_GMU_AO_GMU_CGC_DELAY_CNTL, + state ? cgc_delay : 0); + gmu_write(&a6xx_gpu->gmu, REG_A6XX_GPU_GMU_AO_GMU_CGC_HYST_CNTL, + state ? cgc_hyst : 0); if (!adreno_gpu->info->a6xx->hwcg) { gpu_write(gpu, REG_A7XX_RBBM_CLOCK_CNTL_GLOBAL, 1); @@ -493,24 +493,17 @@ static void a6xx_set_cp_protect(struct msm_gpu *gpu) static void a6xx_calc_ubwc_config(struct adreno_gpu *gpu) { - /* Unknown, introduced with A650 family, related to UBWC mode/ver 4 */ gpu->ubwc_config.rgb565_predicator = 0; - /* Unknown, introduced with A650 family */ gpu->ubwc_config.uavflagprd_inv = 0; - /* Whether the minimum access length is 64 bits */ gpu->ubwc_config.min_acc_len = 0; - /* Entirely magic, per-GPU-gen value */ - gpu->ubwc_config.ubwc_mode = 0; - /* - * The Highest Bank Bit value represents the bit of the highest DDR bank. - * This should ideally use DRAM type detection. - */ + gpu->ubwc_config.ubwc_swizzle = 0x6; + gpu->ubwc_config.macrotile_mode = 0; gpu->ubwc_config.highest_bank_bit = 15; if (adreno_is_a610(gpu)) { gpu->ubwc_config.highest_bank_bit = 13; gpu->ubwc_config.min_acc_len = 1; - gpu->ubwc_config.ubwc_mode = 1; + gpu->ubwc_config.ubwc_swizzle = 0x7; } if (adreno_is_a618(gpu)) @@ -523,9 +516,18 @@ static void a6xx_calc_ubwc_config(struct adreno_gpu *gpu) if (adreno_is_a619_holi(gpu)) gpu->ubwc_config.highest_bank_bit = 13; + if (adreno_is_a621(gpu)) { + gpu->ubwc_config.highest_bank_bit = 13; + gpu->ubwc_config.amsbc = 1; + gpu->ubwc_config.uavflagprd_inv = 2; + } + if (adreno_is_a640_family(gpu)) gpu->ubwc_config.amsbc = 1; + if (adreno_is_a680(gpu)) + gpu->ubwc_config.macrotile_mode = 1; + if (adreno_is_a650(gpu) || adreno_is_a660(gpu) || adreno_is_a690(gpu) || @@ -536,6 +538,7 @@ static void a6xx_calc_ubwc_config(struct adreno_gpu *gpu) gpu->ubwc_config.amsbc = 1; gpu->ubwc_config.rgb565_predicator = 1; gpu->ubwc_config.uavflagprd_inv = 2; + gpu->ubwc_config.macrotile_mode = 1; } if (adreno_is_7c3(gpu)) { @@ -543,12 +546,12 @@ static void a6xx_calc_ubwc_config(struct adreno_gpu *gpu) gpu->ubwc_config.amsbc = 1; gpu->ubwc_config.rgb565_predicator = 1; gpu->ubwc_config.uavflagprd_inv = 2; + gpu->ubwc_config.macrotile_mode = 1; } if (adreno_is_a702(gpu)) { gpu->ubwc_config.highest_bank_bit = 14; gpu->ubwc_config.min_acc_len = 1; - gpu->ubwc_config.ubwc_mode = 0; } } @@ -564,21 +567,26 @@ static void a6xx_set_ubwc_config(struct msm_gpu *gpu) u32 hbb = adreno_gpu->ubwc_config.highest_bank_bit - 13; u32 hbb_hi = hbb >> 2; u32 hbb_lo = hbb & 3; + u32 ubwc_mode = adreno_gpu->ubwc_config.ubwc_swizzle & 1; + u32 level2_swizzling_dis = !(adreno_gpu->ubwc_config.ubwc_swizzle & 2); gpu_write(gpu, REG_A6XX_RB_NC_MODE_CNTL, + level2_swizzling_dis << 12 | adreno_gpu->ubwc_config.rgb565_predicator << 11 | hbb_hi << 10 | adreno_gpu->ubwc_config.amsbc << 4 | adreno_gpu->ubwc_config.min_acc_len << 3 | - hbb_lo << 1 | adreno_gpu->ubwc_config.ubwc_mode); + hbb_lo << 1 | ubwc_mode); - gpu_write(gpu, REG_A6XX_TPL1_NC_MODE_CNTL, hbb_hi << 4 | + gpu_write(gpu, REG_A6XX_TPL1_NC_MODE_CNTL, + level2_swizzling_dis << 6 | hbb_hi << 4 | adreno_gpu->ubwc_config.min_acc_len << 3 | - hbb_lo << 1 | adreno_gpu->ubwc_config.ubwc_mode); + hbb_lo << 1 | ubwc_mode); - gpu_write(gpu, REG_A6XX_SP_NC_MODE_CNTL, hbb_hi << 10 | + gpu_write(gpu, REG_A6XX_SP_NC_MODE_CNTL, + level2_swizzling_dis << 12 | hbb_hi << 10 | adreno_gpu->ubwc_config.uavflagprd_inv << 4 | adreno_gpu->ubwc_config.min_acc_len << 3 | - hbb_lo << 1 | adreno_gpu->ubwc_config.ubwc_mode); + hbb_lo << 1 | ubwc_mode); if (adreno_is_a7xx(adreno_gpu)) gpu_write(gpu, REG_A7XX_GRAS_NC_MODE_CNTL, @@ -586,6 +594,9 @@ static void a6xx_set_ubwc_config(struct msm_gpu *gpu) gpu_write(gpu, REG_A6XX_UCHE_MODE_CNTL, adreno_gpu->ubwc_config.min_acc_len << 23 | hbb_lo << 21); + + gpu_write(gpu, REG_A6XX_RBBM_NC_MODE_CNTL, + adreno_gpu->ubwc_config.macrotile_mode); } static int a6xx_cp_init(struct msm_gpu *gpu) @@ -976,25 +987,11 @@ static int hw_init(struct msm_gpu *gpu) } else if (!adreno_is_a7xx(adreno_gpu)) gpu_write(gpu, REG_A6XX_CP_MEM_POOL_SIZE, 128); - /* Setting the primFifo thresholds default values, - * and vccCacheSkipDis=1 bit (0x200) for A640 and newer - */ - if (adreno_is_a702(adreno_gpu)) - gpu_write(gpu, REG_A6XX_PC_DBG_ECO_CNTL, 0x0000c000); - else if (adreno_is_a690(adreno_gpu)) - gpu_write(gpu, REG_A6XX_PC_DBG_ECO_CNTL, 0x00800200); - else if (adreno_is_a650(adreno_gpu) || adreno_is_a660(adreno_gpu)) - gpu_write(gpu, REG_A6XX_PC_DBG_ECO_CNTL, 0x00300200); - else if (adreno_is_a640_family(adreno_gpu) || adreno_is_7c3(adreno_gpu)) - gpu_write(gpu, REG_A6XX_PC_DBG_ECO_CNTL, 0x00200200); - else if (adreno_is_a650(adreno_gpu) || adreno_is_a660(adreno_gpu)) - gpu_write(gpu, REG_A6XX_PC_DBG_ECO_CNTL, 0x00300200); - else if (adreno_is_a619(adreno_gpu)) - gpu_write(gpu, REG_A6XX_PC_DBG_ECO_CNTL, 0x00018000); - else if (adreno_is_a610(adreno_gpu)) - gpu_write(gpu, REG_A6XX_PC_DBG_ECO_CNTL, 0x00080000); - else if (!adreno_is_a7xx(adreno_gpu)) - gpu_write(gpu, REG_A6XX_PC_DBG_ECO_CNTL, 0x00180000); + + /* Set the default primFifo threshold values */ + if (adreno_gpu->info->a6xx->prim_fifo_threshold) + gpu_write(gpu, REG_A6XX_PC_DBG_ECO_CNTL, + adreno_gpu->info->a6xx->prim_fifo_threshold); /* Set the AHB default slave response to "ERROR" */ gpu_write(gpu, REG_A6XX_CP_AHB_CNTL, 0x1); diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu.h b/drivers/gpu/drm/msm/adreno/a6xx_gpu.h index e3e5c53ae8af..0fb7febf70e7 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.h +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.h @@ -22,6 +22,8 @@ struct a6xx_info { const struct adreno_reglist *hwcg; const struct adreno_protect *protect; u32 gmu_chipid; + u32 gmu_cgc_mode; + u32 prim_fifo_threshold; }; struct a6xx_gpu { diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu_state.c b/drivers/gpu/drm/msm/adreno/a6xx_gpu_state.c index 789a11416f7a..0fcae53c0b14 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu_state.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu_state.c @@ -388,18 +388,18 @@ static void a7xx_get_debugbus_blocks(struct msm_gpu *gpu, const u32 *debugbus_blocks, *gbif_debugbus_blocks; int i; - if (adreno_is_a730(adreno_gpu)) { + if (adreno_gpu->info->family == ADRENO_7XX_GEN1) { debugbus_blocks = gen7_0_0_debugbus_blocks; debugbus_blocks_count = ARRAY_SIZE(gen7_0_0_debugbus_blocks); gbif_debugbus_blocks = a7xx_gbif_debugbus_blocks; gbif_debugbus_blocks_count = ARRAY_SIZE(a7xx_gbif_debugbus_blocks); - } else if (adreno_is_a740_family(adreno_gpu)) { + } else if (adreno_gpu->info->family == ADRENO_7XX_GEN2) { debugbus_blocks = gen7_2_0_debugbus_blocks; debugbus_blocks_count = ARRAY_SIZE(gen7_2_0_debugbus_blocks); gbif_debugbus_blocks = a7xx_gbif_debugbus_blocks; gbif_debugbus_blocks_count = ARRAY_SIZE(a7xx_gbif_debugbus_blocks); } else { - BUG_ON(!adreno_is_a750(adreno_gpu)); + BUG_ON(adreno_gpu->info->family != ADRENO_7XX_GEN3); debugbus_blocks = gen7_9_0_debugbus_blocks; debugbus_blocks_count = ARRAY_SIZE(gen7_9_0_debugbus_blocks); gbif_debugbus_blocks = gen7_9_0_gbif_debugbus_blocks; @@ -509,7 +509,7 @@ static void a6xx_get_debugbus(struct msm_gpu *gpu, const struct a6xx_debugbus_block *cx_debugbus_blocks; if (adreno_is_a7xx(adreno_gpu)) { - BUG_ON(!(adreno_is_a730(adreno_gpu) || adreno_is_a740_family(adreno_gpu))); + BUG_ON(adreno_gpu->info->family > ADRENO_7XX_GEN3); cx_debugbus_blocks = a7xx_cx_debugbus_blocks; nr_cx_debugbus_blocks = ARRAY_SIZE(a7xx_cx_debugbus_blocks); } else { @@ -660,13 +660,16 @@ static void a7xx_get_dbgahb_clusters(struct msm_gpu *gpu, const struct gen7_sptp_cluster_registers *dbgahb_clusters; unsigned dbgahb_clusters_size; - if (adreno_is_a730(adreno_gpu)) { + if (adreno_gpu->info->family == ADRENO_7XX_GEN1) { dbgahb_clusters = gen7_0_0_sptp_clusters; dbgahb_clusters_size = ARRAY_SIZE(gen7_0_0_sptp_clusters); - } else { - BUG_ON(!adreno_is_a740_family(adreno_gpu)); + } else if (adreno_gpu->info->family == ADRENO_7XX_GEN2) { dbgahb_clusters = gen7_2_0_sptp_clusters; dbgahb_clusters_size = ARRAY_SIZE(gen7_2_0_sptp_clusters); + } else { + BUG_ON(adreno_gpu->info->family != ADRENO_7XX_GEN3); + dbgahb_clusters = gen7_9_0_sptp_clusters; + dbgahb_clusters_size = ARRAY_SIZE(gen7_9_0_sptp_clusters); } a6xx_state->dbgahb_clusters = state_kcalloc(a6xx_state, @@ -818,14 +821,14 @@ static void a7xx_get_clusters(struct msm_gpu *gpu, const struct gen7_cluster_registers *clusters; unsigned clusters_size; - if (adreno_is_a730(adreno_gpu)) { + if (adreno_gpu->info->family == ADRENO_7XX_GEN1) { clusters = gen7_0_0_clusters; clusters_size = ARRAY_SIZE(gen7_0_0_clusters); - } else if (adreno_is_a740_family(adreno_gpu)) { + } else if (adreno_gpu->info->family == ADRENO_7XX_GEN2) { clusters = gen7_2_0_clusters; clusters_size = ARRAY_SIZE(gen7_2_0_clusters); } else { - BUG_ON(!adreno_is_a750(adreno_gpu)); + BUG_ON(adreno_gpu->info->family != ADRENO_7XX_GEN3); clusters = gen7_9_0_clusters; clusters_size = ARRAY_SIZE(gen7_9_0_clusters); } @@ -893,7 +896,7 @@ static void a7xx_get_shader_block(struct msm_gpu *gpu, if (WARN_ON(datasize > A6XX_CD_DATA_SIZE)) return; - if (adreno_is_a730(adreno_gpu)) { + if (adreno_gpu->info->family == ADRENO_7XX_GEN1) { gpu_rmw(gpu, REG_A7XX_SP_DBG_CNTL, GENMASK(1, 0), 3); } @@ -923,7 +926,7 @@ static void a7xx_get_shader_block(struct msm_gpu *gpu, datasize); out: - if (adreno_is_a730(adreno_gpu)) { + if (adreno_gpu->info->family == ADRENO_7XX_GEN1) { gpu_rmw(gpu, REG_A7XX_SP_DBG_CNTL, GENMASK(1, 0), 0); } } @@ -956,14 +959,14 @@ static void a7xx_get_shaders(struct msm_gpu *gpu, unsigned num_shader_blocks; int i; - if (adreno_is_a730(adreno_gpu)) { + if (adreno_gpu->info->family == ADRENO_7XX_GEN1) { shader_blocks = gen7_0_0_shader_blocks; num_shader_blocks = ARRAY_SIZE(gen7_0_0_shader_blocks); - } else if (adreno_is_a740_family(adreno_gpu)) { + } else if (adreno_gpu->info->family == ADRENO_7XX_GEN2) { shader_blocks = gen7_2_0_shader_blocks; num_shader_blocks = ARRAY_SIZE(gen7_2_0_shader_blocks); } else { - BUG_ON(!adreno_is_a750(adreno_gpu)); + BUG_ON(adreno_gpu->info->family != ADRENO_7XX_GEN3); shader_blocks = gen7_9_0_shader_blocks; num_shader_blocks = ARRAY_SIZE(gen7_9_0_shader_blocks); } @@ -1348,14 +1351,14 @@ static void a7xx_get_registers(struct msm_gpu *gpu, const u32 *pre_crashdumper_regs; const struct gen7_reg_list *reglist; - if (adreno_is_a730(adreno_gpu)) { + if (adreno_gpu->info->family == ADRENO_7XX_GEN1) { reglist = gen7_0_0_reg_list; pre_crashdumper_regs = gen7_0_0_pre_crashdumper_gpu_registers; - } else if (adreno_is_a740_family(adreno_gpu)) { + } else if (adreno_gpu->info->family == ADRENO_7XX_GEN2) { reglist = gen7_2_0_reg_list; pre_crashdumper_regs = gen7_0_0_pre_crashdumper_gpu_registers; } else { - BUG_ON(!adreno_is_a750(adreno_gpu)); + BUG_ON(adreno_gpu->info->family != ADRENO_7XX_GEN3); reglist = gen7_9_0_reg_list; pre_crashdumper_regs = gen7_9_0_pre_crashdumper_gpu_registers; } @@ -1405,8 +1408,7 @@ static void a7xx_get_post_crashdumper_registers(struct msm_gpu *gpu, struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu); const u32 *regs; - BUG_ON(!(adreno_is_a730(adreno_gpu) || adreno_is_a740_family(adreno_gpu) || - adreno_is_a750(adreno_gpu))); + BUG_ON(adreno_gpu->info->family > ADRENO_7XX_GEN3); regs = gen7_0_0_post_crashdumper_registers; a7xx_get_ahb_gpu_registers(gpu, @@ -1514,11 +1516,11 @@ static void a7xx_get_indexed_registers(struct msm_gpu *gpu, const struct a6xx_indexed_registers *indexed_regs; int i, indexed_count, mempool_count; - if (adreno_is_a730(adreno_gpu) || adreno_is_a740_family(adreno_gpu)) { + if (adreno_gpu->info->family <= ADRENO_7XX_GEN2) { indexed_regs = a7xx_indexed_reglist; indexed_count = ARRAY_SIZE(a7xx_indexed_reglist); } else { - BUG_ON(!adreno_is_a750(adreno_gpu)); + BUG_ON(adreno_gpu->info->family != ADRENO_7XX_GEN3); indexed_regs = gen7_9_0_cp_indexed_reg_list; indexed_count = ARRAY_SIZE(gen7_9_0_cp_indexed_reg_list); } diff --git a/drivers/gpu/drm/msm/adreno/adreno_gen7_9_0_snapshot.h b/drivers/gpu/drm/msm/adreno/adreno_gen7_9_0_snapshot.h index 260d66eccfec..9a327d543f27 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_gen7_9_0_snapshot.h +++ b/drivers/gpu/drm/msm/adreno/adreno_gen7_9_0_snapshot.h @@ -1303,7 +1303,7 @@ static struct a6xx_indexed_registers gen7_9_0_cp_indexed_reg_list[] = { REG_A6XX_CP_ROQ_DBG_DATA, 0x00800}, { "CP_UCODE_DBG_DATA", REG_A6XX_CP_SQE_UCODE_DBG_ADDR, REG_A6XX_CP_SQE_UCODE_DBG_DATA, 0x08000}, - { "CP_BV_SQE_STAT_ADDR", REG_A7XX_CP_BV_DRAW_STATE_ADDR, + { "CP_BV_DRAW_STATE_ADDR", REG_A7XX_CP_BV_DRAW_STATE_ADDR, REG_A7XX_CP_BV_DRAW_STATE_DATA, 0x00200}, { "CP_BV_ROQ_DBG_ADDR", REG_A7XX_CP_BV_ROQ_DBG_ADDR, REG_A7XX_CP_BV_ROQ_DBG_DATA, 0x00800}, diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.c b/drivers/gpu/drm/msm/adreno/adreno_gpu.c index ecc3fc5cec22..465a4cd14a43 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_gpu.c +++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.c @@ -379,6 +379,12 @@ int adreno_get_param(struct msm_gpu *gpu, struct msm_file_private *ctx, case MSM_PARAM_RAYTRACING: *value = adreno_gpu->has_ray_tracing; return 0; + case MSM_PARAM_UBWC_SWIZZLE: + *value = adreno_gpu->ubwc_config.ubwc_swizzle; + return 0; + case MSM_PARAM_MACROTILE_MODE: + *value = adreno_gpu->ubwc_config.macrotile_mode; + return 0; default: DBG("%s: invalid param: %u", gpu->name, param); return -EINVAL; @@ -478,7 +484,7 @@ adreno_request_fw(struct adreno_gpu *adreno_gpu, const char *fwname) ret = request_firmware_direct(&fw, fwname, drm->dev); if (!ret) { DRM_DEV_INFO(drm->dev, "loaded %s from legacy location\n", - newname); + fwname); adreno_gpu->fwloc = FW_LOCATION_LEGACY; goto out; } else if (adreno_gpu->fwloc != FW_LOCATION_UNKNOWN) { @@ -688,11 +694,9 @@ int adreno_gpu_state_get(struct msm_gpu *gpu, struct msm_gpu_state *state) size = j + 1; if (size) { - state->ring[i].data = kvmalloc(size << 2, GFP_KERNEL); - if (state->ring[i].data) { - memcpy(state->ring[i].data, gpu->rb[i]->start, size << 2); + state->ring[i].data = kvmemdup(gpu->rb[i]->start, size << 2, GFP_KERNEL); + if (state->ring[i].data) state->ring[i].data_size = size << 2; - } } } @@ -1083,6 +1087,7 @@ int adreno_gpu_init(struct drm_device *drm, struct platform_device *pdev, adreno_gpu->chip_id = config->chip_id; gpu->allow_relocs = config->info->family < ADRENO_6XX_GEN1; + gpu->pdev = pdev; /* Only handle the core clock when GMU is not in use (or is absent). */ if (adreno_has_gmu_wrapper(adreno_gpu) || diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.h b/drivers/gpu/drm/msm/adreno/adreno_gpu.h index 1ab523a163a0..58d7e7915c57 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_gpu.h +++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.h @@ -191,12 +191,42 @@ struct adreno_gpu { const struct firmware *fw[ADRENO_FW_MAX]; struct { + /** + * @rgb565_predicator: Unknown, introduced with A650 family, + * related to UBWC mode/ver 4 + */ u32 rgb565_predicator; + /** @uavflagprd_inv: Unknown, introduced with A650 family */ u32 uavflagprd_inv; + /** @min_acc_len: Whether the minimum access length is 64 bits */ u32 min_acc_len; - u32 ubwc_mode; + /** + * @ubwc_swizzle: Whether to enable level 1, 2 & 3 bank swizzling. + * + * UBWC 1.0 always enables all three levels. + * UBWC 2.0 removes level 1 bank swizzling, leaving levels 2 & 3. + * UBWC 4.0 adds the optional ability to disable levels 2 & 3. + * + * This is a bitmask where BIT(0) enables level 1, BIT(1) + * controls level 2, and BIT(2) enables level 3. + */ + u32 ubwc_swizzle; + /** + * @highest_bank_bit: Highest Bank Bit + * + * The Highest Bank Bit value represents the bit of the highest + * DDR bank. This should ideally use DRAM type detection. + */ u32 highest_bank_bit; u32 amsbc; + /** + * @macrotile_mode: Macrotile Mode + * + * Whether to use 4-channel macrotiling mode or the newer + * 8-channel macrotiling mode introduced in UBWC 3.1. 0 is + * 4-channel and 1 is 8-channel. + */ + u32 macrotile_mode; } ubwc_config; /* @@ -294,6 +324,12 @@ static inline bool adreno_is_a306(const struct adreno_gpu *gpu) return adreno_is_revn(gpu, 307); } +static inline bool adreno_is_a306a(const struct adreno_gpu *gpu) +{ + /* a306a (marketing name is a308) */ + return adreno_is_revn(gpu, 308); +} + static inline bool adreno_is_a320(const struct adreno_gpu *gpu) { return adreno_is_revn(gpu, 320); @@ -384,6 +420,11 @@ static inline int adreno_is_a619_holi(const struct adreno_gpu *gpu) return adreno_is_a619(gpu) && adreno_has_gmu_wrapper(gpu); } +static inline int adreno_is_a621(const struct adreno_gpu *gpu) +{ + return gpu->info->chip_ids[0] == 0x06020100; +} + static inline int adreno_is_a630(const struct adreno_gpu *gpu) { return adreno_is_revn(gpu, 630); @@ -433,7 +474,13 @@ static inline int adreno_is_a610_family(const struct adreno_gpu *gpu) return adreno_is_a610(gpu) || adreno_is_a702(gpu); } -/* check for a615, a616, a618, a619 or any a630 derivatives */ +/* TODO: 615/616 */ +static inline int adreno_is_a615_family(const struct adreno_gpu *gpu) +{ + return adreno_is_a618(gpu) || + adreno_is_a619(gpu); +} + static inline int adreno_is_a630_family(const struct adreno_gpu *gpu) { if (WARN_ON_ONCE(!gpu->info)) diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_0_sm8150.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_0_sm8150.h index 145f3d5953a3..6ccfde82fecd 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_0_sm8150.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_0_sm8150.h @@ -33,6 +33,7 @@ static const struct dpu_mdp_cfg sm8150_mdp = { [DPU_CLK_CTRL_DMA1] = { .reg_off = 0x2b4, .bit_off = 8 }, [DPU_CLK_CTRL_DMA2] = { .reg_off = 0x2bc, .bit_off = 8 }, [DPU_CLK_CTRL_DMA3] = { .reg_off = 0x2c4, .bit_off = 8 }, + [DPU_CLK_CTRL_WB2] = { .reg_off = 0x2bc, .bit_off = 16 }, }, }; @@ -290,6 +291,21 @@ static const struct dpu_dsc_cfg sm8150_dsc[] = { }, }; +static const struct dpu_wb_cfg sm8150_wb[] = { + { + .name = "wb_2", .id = WB_2, + .base = 0x65000, .len = 0x2c8, + .features = WB_SDM845_MASK, + .format_list = wb2_formats_rgb, + .num_formats = ARRAY_SIZE(wb2_formats_rgb), + .clk_ctrl = DPU_CLK_CTRL_WB2, + .xin_id = 6, + .vbif_idx = VBIF_RT, + .maxlinewidth = 4096, + .intr_wb_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 4), + }, +}; + static const struct dpu_intf_cfg sm8150_intf[] = { { .name = "intf_0", .id = INTF_0, @@ -384,6 +400,8 @@ const struct dpu_mdss_cfg dpu_sm8150_cfg = { .pingpong = sm8150_pp, .merge_3d_count = ARRAY_SIZE(sm8150_merge_3d), .merge_3d = sm8150_merge_3d, + .wb_count = ARRAY_SIZE(sm8150_wb), + .wb = sm8150_wb, .intf_count = ARRAY_SIZE(sm8150_intf), .intf = sm8150_intf, .vbif_count = ARRAY_SIZE(sdm845_vbif), diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_1_sc8180x.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_1_sc8180x.h index 9e3bec8bc121..bab19ddd1d4f 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_1_sc8180x.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_1_sc8180x.h @@ -33,6 +33,7 @@ static const struct dpu_mdp_cfg sc8180x_mdp = { [DPU_CLK_CTRL_DMA1] = { .reg_off = 0x2b4, .bit_off = 8 }, [DPU_CLK_CTRL_DMA2] = { .reg_off = 0x2bc, .bit_off = 8 }, [DPU_CLK_CTRL_DMA3] = { .reg_off = 0x2c4, .bit_off = 8 }, + [DPU_CLK_CTRL_WB2] = { .reg_off = 0x2bc, .bit_off = 16 }, }, }; @@ -297,6 +298,21 @@ static const struct dpu_dsc_cfg sc8180x_dsc[] = { }, }; +static const struct dpu_wb_cfg sc8180x_wb[] = { + { + .name = "wb_2", .id = WB_2, + .base = 0x65000, .len = 0x2c8, + .features = WB_SDM845_MASK, + .format_list = wb2_formats_rgb, + .num_formats = ARRAY_SIZE(wb2_formats_rgb), + .clk_ctrl = DPU_CLK_CTRL_WB2, + .xin_id = 6, + .vbif_idx = VBIF_RT, + .maxlinewidth = 4096, + .intr_wb_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 4), + }, +}; + static const struct dpu_intf_cfg sc8180x_intf[] = { { .name = "intf_0", .id = INTF_0, @@ -410,6 +426,8 @@ const struct dpu_mdss_cfg dpu_sc8180x_cfg = { .pingpong = sc8180x_pp, .merge_3d_count = ARRAY_SIZE(sc8180x_merge_3d), .merge_3d = sc8180x_merge_3d, + .wb_count = ARRAY_SIZE(sc8180x_wb), + .wb = sc8180x_wb, .intf_count = ARRAY_SIZE(sc8180x_intf), .intf = sc8180x_intf, .vbif_count = ARRAY_SIZE(sdm845_vbif), diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_4_sm6125.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_4_sm6125.h index 76b2ec0d2489..d039b96beb97 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_4_sm6125.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_4_sm6125.h @@ -27,6 +27,7 @@ static const struct dpu_mdp_cfg sm6125_mdp = { [DPU_CLK_CTRL_VIG0] = { .reg_off = 0x2ac, .bit_off = 0 }, [DPU_CLK_CTRL_DMA0] = { .reg_off = 0x2ac, .bit_off = 8 }, [DPU_CLK_CTRL_DMA1] = { .reg_off = 0x2b4, .bit_off = 8 }, + [DPU_CLK_CTRL_WB2] = { .reg_off = 0x2bc, .bit_off = 16 }, }, }; @@ -139,6 +140,21 @@ static const struct dpu_pingpong_cfg sm6125_pp[] = { }, }; +static const struct dpu_wb_cfg sm6125_wb[] = { + { + .name = "wb_2", .id = WB_2, + .base = 0x65000, .len = 0x2c8, + .features = WB_SDM845_MASK, + .format_list = wb2_formats_rgb, + .num_formats = ARRAY_SIZE(wb2_formats_rgb), + .clk_ctrl = DPU_CLK_CTRL_WB2, + .xin_id = 6, + .vbif_idx = VBIF_RT, + .maxlinewidth = 2160, + .intr_wb_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 4), + }, +}; + static const struct dpu_intf_cfg sm6125_intf[] = { { .name = "intf_0", .id = INTF_0, @@ -210,6 +226,8 @@ const struct dpu_mdss_cfg dpu_sm6125_cfg = { .dspp = sm6125_dspp, .pingpong_count = ARRAY_SIZE(sm6125_pp), .pingpong = sm6125_pp, + .wb_count = ARRAY_SIZE(sm6125_wb), + .wb = sm6125_wb, .intf_count = ARRAY_SIZE(sm6125_intf), .intf = sm6125_intf, .vbif_count = ARRAY_SIZE(sdm845_vbif), diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_4_sm6350.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_4_sm6350.h index e17a30be7525..0502cee2f116 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_4_sm6350.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_4_sm6350.h @@ -26,6 +26,7 @@ static const struct dpu_mdp_cfg sm6350_mdp = { [DPU_CLK_CTRL_DMA0] = { .reg_off = 0x2ac, .bit_off = 8 }, [DPU_CLK_CTRL_DMA1] = { .reg_off = 0x2b4, .bit_off = 8 }, [DPU_CLK_CTRL_DMA2] = { .reg_off = 0x2c4, .bit_off = 8 }, + [DPU_CLK_CTRL_WB2] = { .reg_off = 0x2bc, .bit_off = 16 }, [DPU_CLK_CTRL_REG_DMA] = { .reg_off = 0x2bc, .bit_off = 20 }, }, }; @@ -145,6 +146,21 @@ static const struct dpu_dsc_cfg sm6350_dsc[] = { }, }; +static const struct dpu_wb_cfg sm6350_wb[] = { + { + .name = "wb_2", .id = WB_2, + .base = 0x65000, .len = 0x2c8, + .features = WB_SM8250_MASK, + .format_list = wb2_formats_rgb, + .num_formats = ARRAY_SIZE(wb2_formats_rgb), + .clk_ctrl = DPU_CLK_CTRL_WB2, + .xin_id = 6, + .vbif_idx = VBIF_RT, + .maxlinewidth = 1920, + .intr_wb_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 4), + }, +}; + static const struct dpu_intf_cfg sm6350_intf[] = { { .name = "intf_0", .id = INTF_0, @@ -218,6 +234,8 @@ const struct dpu_mdss_cfg dpu_sm6350_cfg = { .dsc = sm6350_dsc, .pingpong_count = ARRAY_SIZE(sm6350_pp), .pingpong = sm6350_pp, + .wb_count = ARRAY_SIZE(sm6350_wb), + .wb = sm6350_wb, .intf_count = ARRAY_SIZE(sm6350_intf), .intf = sm6350_intf, .vbif_count = ARRAY_SIZE(sdm845_vbif), diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c index 648c8d0a4c36..dcb4fd85e73b 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c @@ -96,14 +96,16 @@ #define INTF_SC7280_MASK (INTF_SC7180_MASK) -#define WB_SM8250_MASK (BIT(DPU_WB_LINE_MODE) | \ +#define WB_SDM845_MASK (BIT(DPU_WB_LINE_MODE) | \ BIT(DPU_WB_UBWC) | \ BIT(DPU_WB_YUV_CONFIG) | \ BIT(DPU_WB_PIPE_ALPHA) | \ BIT(DPU_WB_XY_ROI_OFFSET) | \ BIT(DPU_WB_QOS) | \ BIT(DPU_WB_QOS_8LVL) | \ - BIT(DPU_WB_CDP) | \ + BIT(DPU_WB_CDP)) + +#define WB_SM8250_MASK (WB_SDM845_MASK | \ BIT(DPU_WB_INPUT_CTRL)) #define DEFAULT_PIXEL_RAM_SIZE (50 * 1024) diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.c index 6e2ac50b94a4..0f40eea7f5e2 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.c @@ -2,6 +2,8 @@ /* Copyright (c) 2015-2018, The Linux Foundation. All rights reserved. */ +#include <linux/bitfield.h> + #include <drm/drm_managed.h> #include "dpu_hwio.h" @@ -231,8 +233,38 @@ static void dpu_hw_intf_audio_select(struct dpu_hw_mdp *mdp) DPU_REG_WRITE(c, HDMI_DP_CORE_SELECT, 0x1); } +static void dpu_hw_dp_phy_intf_sel(struct dpu_hw_mdp *mdp, + enum dpu_dp_phy_sel phys[2]) +{ + struct dpu_hw_blk_reg_map *c = &mdp->hw; + unsigned int intf; + u32 sel = 0; + + sel |= FIELD_PREP(MDP_DP_PHY_INTF_SEL_INTF0, phys[0]); + sel |= FIELD_PREP(MDP_DP_PHY_INTF_SEL_INTF1, phys[1]); + + for (intf = 0; intf < 2; intf++) { + switch (phys[intf]) { + case DPU_DP_PHY_0: + sel |= FIELD_PREP(MDP_DP_PHY_INTF_SEL_PHY0, intf + 1); + break; + case DPU_DP_PHY_1: + sel |= FIELD_PREP(MDP_DP_PHY_INTF_SEL_PHY1, intf + 1); + break; + case DPU_DP_PHY_2: + sel |= FIELD_PREP(MDP_DP_PHY_INTF_SEL_PHY2, intf + 1); + break; + default: + /* ignore */ + break; + } + } + + DPU_REG_WRITE(c, MDP_DP_PHY_INTF_SEL, sel); +} + static void _setup_mdp_ops(struct dpu_hw_mdp_ops *ops, - unsigned long cap) + unsigned long cap, const struct dpu_mdss_version *mdss_rev) { ops->setup_split_pipe = dpu_hw_setup_split_pipe; ops->setup_clk_force_ctrl = dpu_hw_setup_clk_force_ctrl; @@ -245,6 +277,9 @@ static void _setup_mdp_ops(struct dpu_hw_mdp_ops *ops, ops->get_safe_status = dpu_hw_get_safe_status; + if (mdss_rev->core_major_ver >= 5) + ops->dp_phy_intf_sel = dpu_hw_dp_phy_intf_sel; + if (cap & BIT(DPU_MDP_AUDIO_SELECT)) ops->intf_audio_select = dpu_hw_intf_audio_select; } @@ -252,7 +287,7 @@ static void _setup_mdp_ops(struct dpu_hw_mdp_ops *ops, struct dpu_hw_mdp *dpu_hw_mdptop_init(struct drm_device *dev, const struct dpu_mdp_cfg *cfg, void __iomem *addr, - const struct dpu_mdss_cfg *m) + const struct dpu_mdss_version *mdss_rev) { struct dpu_hw_mdp *mdp; @@ -270,7 +305,7 @@ struct dpu_hw_mdp *dpu_hw_mdptop_init(struct drm_device *dev, * Assign ops */ mdp->caps = cfg; - _setup_mdp_ops(&mdp->ops, mdp->caps->features); + _setup_mdp_ops(&mdp->ops, mdp->caps->features, mdss_rev); return mdp; } diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.h index 5c9a7ede991e..f1ab9fd106e5 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.h @@ -67,6 +67,13 @@ struct dpu_vsync_source_cfg { enum dpu_vsync_source vsync_source; }; +enum dpu_dp_phy_sel { + DPU_DP_PHY_NONE, + DPU_DP_PHY_0, + DPU_DP_PHY_1, + DPU_DP_PHY_2, +}; + /** * struct dpu_hw_mdp_ops - interface to the MDP TOP Hw driver functions * Assumption is these functions will be called after clocks are enabled. @@ -126,6 +133,13 @@ struct dpu_hw_mdp_ops { struct dpu_danger_safe_status *status); /** + * dp_phy_intf_sel - configure intf to phy mapping + * @mdp: mdp top context driver + * @phys: list of phys the DP interfaces should be connected to. 0 disables the INTF. + */ + void (*dp_phy_intf_sel)(struct dpu_hw_mdp *mdp, enum dpu_dp_phy_sel phys[2]); + + /** * intf_audio_select - select the external interface for audio * @mdp: mdp top context driver */ @@ -148,12 +162,12 @@ struct dpu_hw_mdp { * @dev: Corresponding device for devres management * @cfg: MDP TOP configuration from catalog * @addr: Mapped register io address of MDP - * @m: Pointer to mdss catalog data + * @mdss_rev: dpu core's major and minor versions */ struct dpu_hw_mdp *dpu_hw_mdptop_init(struct drm_device *dev, const struct dpu_mdp_cfg *cfg, void __iomem *addr, - const struct dpu_mdss_cfg *m); + const struct dpu_mdss_version *mdss_rev); void dpu_hw_mdp_destroy(struct dpu_hw_mdp *mdp); diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hwio.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hwio.h index 5acd5683d25a..054fe097ebf8 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hwio.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hwio.h @@ -60,6 +60,13 @@ #define MDP_WD_TIMER_4_LOAD_VALUE 0x448 #define DCE_SEL 0x450 +#define MDP_DP_PHY_INTF_SEL 0x460 +#define MDP_DP_PHY_INTF_SEL_INTF0 GENMASK(2, 0) +#define MDP_DP_PHY_INTF_SEL_INTF1 GENMASK(5, 3) +#define MDP_DP_PHY_INTF_SEL_PHY0 GENMASK(8, 6) +#define MDP_DP_PHY_INTF_SEL_PHY1 GENMASK(11, 9) +#define MDP_DP_PHY_INTF_SEL_PHY2 GENMASK(14, 12) + #define MDP_PERIPH_TOP0 MDP_WD_TIMER_0_CTL #define MDP_PERIPH_TOP0_END CLK_CTRL3 diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c index d1e2143110f2..9bcae53c4f45 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c @@ -1146,7 +1146,7 @@ static int dpu_kms_hw_init(struct msm_kms *kms) dpu_kms->hw_mdp = dpu_hw_mdptop_init(dev, dpu_kms->catalog->mdp, dpu_kms->mmio, - dpu_kms->catalog); + dpu_kms->catalog->mdss_ver); if (IS_ERR(dpu_kms->hw_mdp)) { rc = PTR_ERR(dpu_kms->hw_mdp); DPU_ERROR("failed to get hw_mdp: %d\n", rc); @@ -1181,6 +1181,16 @@ static int dpu_kms_hw_init(struct msm_kms *kms) goto err_pm_put; } + /* + * We need to program DP <-> PHY relationship only for SC8180X since it + * has fewer DP controllers than DP PHYs. + * If any other platform requires the same kind of programming, or if + * the INTF <->DP relationship isn't static anymore, this needs to be + * configured through the DT. + */ + if (of_device_is_compatible(dpu_kms->pdev->dev.of_node, "qcom,sc8180x-dpu")) + dpu_kms->hw_mdp->ops.dp_phy_intf_sel(dpu_kms->hw_mdp, (unsigned int[]){ 1, 2, }); + dpu_kms->hw_intr = dpu_hw_intr_init(dev, dpu_kms->mmio, dpu_kms->catalog); if (IS_ERR(dpu_kms->hw_intr)) { rc = PTR_ERR(dpu_kms->hw_intr); diff --git a/drivers/gpu/drm/msm/disp/mdp5/mdp5_smp.c b/drivers/gpu/drm/msm/disp/mdp5/mdp5_smp.c index 3a7f7edda96b..500b7dc895d0 100644 --- a/drivers/gpu/drm/msm/disp/mdp5/mdp5_smp.c +++ b/drivers/gpu/drm/msm/disp/mdp5/mdp5_smp.c @@ -351,7 +351,7 @@ void mdp5_smp_dump(struct mdp5_smp *smp, struct drm_printer *p, drm_printf(p, "%s:%d\t%d\t%s\n", pipe2name(pipe), j, inuse, - plane ? plane->name : NULL); + plane ? plane->name : "(null)"); total += inuse; } diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c index 9622e58dce3e..e1228fb093ee 100644 --- a/drivers/gpu/drm/msm/dp/dp_display.c +++ b/drivers/gpu/drm/msm/dp/dp_display.c @@ -119,7 +119,7 @@ struct msm_dp_desc { }; static const struct msm_dp_desc sc7180_dp_descs[] = { - { .io_start = 0x0ae90000, .id = MSM_DP_CONTROLLER_0 }, + { .io_start = 0x0ae90000, .id = MSM_DP_CONTROLLER_0, .wide_bus_supported = true }, {} }; @@ -130,9 +130,9 @@ static const struct msm_dp_desc sc7280_dp_descs[] = { }; static const struct msm_dp_desc sc8180x_dp_descs[] = { - { .io_start = 0x0ae90000, .id = MSM_DP_CONTROLLER_0 }, - { .io_start = 0x0ae98000, .id = MSM_DP_CONTROLLER_1 }, - { .io_start = 0x0ae9a000, .id = MSM_DP_CONTROLLER_2 }, + { .io_start = 0x0ae90000, .id = MSM_DP_CONTROLLER_0, .wide_bus_supported = true }, + { .io_start = 0x0ae98000, .id = MSM_DP_CONTROLLER_1, .wide_bus_supported = true }, + { .io_start = 0x0ae9a000, .id = MSM_DP_CONTROLLER_2, .wide_bus_supported = true }, {} }; @@ -149,7 +149,7 @@ static const struct msm_dp_desc sc8280xp_dp_descs[] = { }; static const struct msm_dp_desc sm8650_dp_descs[] = { - { .io_start = 0x0af54000, .id = MSM_DP_CONTROLLER_0 }, + { .io_start = 0x0af54000, .id = MSM_DP_CONTROLLER_0, .wide_bus_supported = true }, {} }; diff --git a/drivers/gpu/drm/msm/dsi/phy/dsi_phy_7nm.c b/drivers/gpu/drm/msm/dsi/phy/dsi_phy_7nm.c index 3b59137ca674..031446c87dae 100644 --- a/drivers/gpu/drm/msm/dsi/phy/dsi_phy_7nm.c +++ b/drivers/gpu/drm/msm/dsi/phy/dsi_phy_7nm.c @@ -135,7 +135,7 @@ static void dsi_pll_calc_dec_frac(struct dsi_pll_7nm *pll, struct dsi_pll_config config->pll_clock_inverters = 0x00; else config->pll_clock_inverters = 0x40; - } else { + } else if (pll->phy->cfg->quirks & DSI_PHY_7NM_QUIRK_V4_1) { if (pll_freq <= 1000000000ULL) config->pll_clock_inverters = 0xa0; else if (pll_freq <= 2500000000ULL) @@ -144,6 +144,16 @@ static void dsi_pll_calc_dec_frac(struct dsi_pll_7nm *pll, struct dsi_pll_config config->pll_clock_inverters = 0x00; else config->pll_clock_inverters = 0x40; + } else { + /* 4.2, 4.3 */ + if (pll_freq <= 1000000000ULL) + config->pll_clock_inverters = 0xa0; + else if (pll_freq <= 2500000000ULL) + config->pll_clock_inverters = 0x20; + else if (pll_freq <= 3500000000ULL) + config->pll_clock_inverters = 0x00; + else + config->pll_clock_inverters = 0x40; } config->decimal_div_start = dec; diff --git a/drivers/gpu/drm/msm/hdmi/hdmi.c b/drivers/gpu/drm/msm/hdmi/hdmi.c index 24abcb7254cc..0bfee41c2e71 100644 --- a/drivers/gpu/drm/msm/hdmi/hdmi.c +++ b/drivers/gpu/drm/msm/hdmi/hdmi.c @@ -549,6 +549,7 @@ static void msm_hdmi_dev_remove(struct platform_device *pdev) } static const struct of_device_id msm_hdmi_dt_match[] = { + { .compatible = "qcom,hdmi-tx-8998", .data = &hdmi_tx_8974_config }, { .compatible = "qcom,hdmi-tx-8996", .data = &hdmi_tx_8974_config }, { .compatible = "qcom,hdmi-tx-8994", .data = &hdmi_tx_8974_config }, { .compatible = "qcom,hdmi-tx-8084", .data = &hdmi_tx_8974_config }, diff --git a/drivers/gpu/drm/msm/hdmi/hdmi.h b/drivers/gpu/drm/msm/hdmi/hdmi.h index 4586baf36415..a62d2aedfbb7 100644 --- a/drivers/gpu/drm/msm/hdmi/hdmi.h +++ b/drivers/gpu/drm/msm/hdmi/hdmi.h @@ -137,6 +137,7 @@ enum hdmi_phy_type { MSM_HDMI_PHY_8960, MSM_HDMI_PHY_8x74, MSM_HDMI_PHY_8996, + MSM_HDMI_PHY_8998, MSM_HDMI_PHY_MAX, }; @@ -154,6 +155,7 @@ extern const struct hdmi_phy_cfg msm_hdmi_phy_8x60_cfg; extern const struct hdmi_phy_cfg msm_hdmi_phy_8960_cfg; extern const struct hdmi_phy_cfg msm_hdmi_phy_8x74_cfg; extern const struct hdmi_phy_cfg msm_hdmi_phy_8996_cfg; +extern const struct hdmi_phy_cfg msm_hdmi_phy_8998_cfg; struct hdmi_phy { struct platform_device *pdev; @@ -184,6 +186,7 @@ void __exit msm_hdmi_phy_driver_unregister(void); #ifdef CONFIG_COMMON_CLK int msm_hdmi_pll_8960_init(struct platform_device *pdev); int msm_hdmi_pll_8996_init(struct platform_device *pdev); +int msm_hdmi_pll_8998_init(struct platform_device *pdev); #else static inline int msm_hdmi_pll_8960_init(struct platform_device *pdev) { @@ -194,6 +197,11 @@ static inline int msm_hdmi_pll_8996_init(struct platform_device *pdev) { return -ENODEV; } + +static inline int msm_hdmi_pll_8998_init(struct platform_device *pdev) +{ + return -ENODEV; +} #endif /* diff --git a/drivers/gpu/drm/msm/hdmi/hdmi_phy.c b/drivers/gpu/drm/msm/hdmi/hdmi_phy.c index 88a3423b7f24..95b3f7535d84 100644 --- a/drivers/gpu/drm/msm/hdmi/hdmi_phy.c +++ b/drivers/gpu/drm/msm/hdmi/hdmi_phy.c @@ -118,6 +118,9 @@ static int msm_hdmi_phy_pll_init(struct platform_device *pdev, case MSM_HDMI_PHY_8996: ret = msm_hdmi_pll_8996_init(pdev); break; + case MSM_HDMI_PHY_8998: + ret = msm_hdmi_pll_8998_init(pdev); + break; /* * we don't have PLL support for these, don't report an error for now */ @@ -193,6 +196,8 @@ static const struct of_device_id msm_hdmi_phy_dt_match[] = { .data = &msm_hdmi_phy_8x74_cfg }, { .compatible = "qcom,hdmi-phy-8996", .data = &msm_hdmi_phy_8996_cfg }, + { .compatible = "qcom,hdmi-phy-8998", + .data = &msm_hdmi_phy_8998_cfg }, {} }; diff --git a/drivers/gpu/drm/msm/hdmi/hdmi_phy_8998.c b/drivers/gpu/drm/msm/hdmi/hdmi_phy_8998.c new file mode 100644 index 000000000000..0e3a2b16a2ce --- /dev/null +++ b/drivers/gpu/drm/msm/hdmi/hdmi_phy_8998.c @@ -0,0 +1,779 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2016, The Linux Foundation. All rights reserved. + * Copyright (c) 2024 Freebox SAS + */ + +#include <linux/clk-provider.h> +#include <linux/delay.h> + +#include "hdmi.h" + +#define HDMI_VCO_MAX_FREQ 12000000000UL +#define HDMI_VCO_MIN_FREQ 8000000000UL + +#define HDMI_PCLK_MAX_FREQ 600000000 +#define HDMI_PCLK_MIN_FREQ 25000000 + +#define HDMI_HIGH_FREQ_BIT_CLK_THRESHOLD 3400000000UL +#define HDMI_DIG_FREQ_BIT_CLK_THRESHOLD 1500000000UL +#define HDMI_MID_FREQ_BIT_CLK_THRESHOLD 750000000UL +#define HDMI_CORECLK_DIV 5 +#define HDMI_DEFAULT_REF_CLOCK 19200000 +#define HDMI_PLL_CMP_CNT 1024 + +#define HDMI_PLL_POLL_MAX_READS 100 +#define HDMI_PLL_POLL_TIMEOUT_US 150 + +#define HDMI_NUM_TX_CHANNEL 4 + +struct hdmi_pll_8998 { + struct platform_device *pdev; + struct clk_hw clk_hw; + unsigned long rate; + + /* pll mmio base */ + void __iomem *mmio_qserdes_com; + /* tx channel base */ + void __iomem *mmio_qserdes_tx[HDMI_NUM_TX_CHANNEL]; +}; + +#define hw_clk_to_pll(x) container_of(x, struct hdmi_pll_8998, clk_hw) + +struct hdmi_8998_phy_pll_reg_cfg { + u32 com_svs_mode_clk_sel; + u32 com_hsclk_sel; + u32 com_pll_cctrl_mode0; + u32 com_pll_rctrl_mode0; + u32 com_cp_ctrl_mode0; + u32 com_dec_start_mode0; + u32 com_div_frac_start1_mode0; + u32 com_div_frac_start2_mode0; + u32 com_div_frac_start3_mode0; + u32 com_integloop_gain0_mode0; + u32 com_integloop_gain1_mode0; + u32 com_lock_cmp_en; + u32 com_lock_cmp1_mode0; + u32 com_lock_cmp2_mode0; + u32 com_lock_cmp3_mode0; + u32 com_core_clk_en; + u32 com_coreclk_div_mode0; + + u32 tx_lx_tx_band[HDMI_NUM_TX_CHANNEL]; + u32 tx_lx_tx_drv_lvl[HDMI_NUM_TX_CHANNEL]; + u32 tx_lx_tx_emp_post1_lvl[HDMI_NUM_TX_CHANNEL]; + u32 tx_lx_pre_driver_1[HDMI_NUM_TX_CHANNEL]; + u32 tx_lx_pre_driver_2[HDMI_NUM_TX_CHANNEL]; + u32 tx_lx_res_code_offset[HDMI_NUM_TX_CHANNEL]; + + u32 phy_mode; +}; + +struct hdmi_8998_post_divider { + u64 vco_freq; + int hsclk_divsel; + int vco_ratio; + int tx_band_sel; + int half_rate_mode; +}; + +static inline struct hdmi_phy *pll_get_phy(struct hdmi_pll_8998 *pll) +{ + return platform_get_drvdata(pll->pdev); +} + +static inline void hdmi_pll_write(struct hdmi_pll_8998 *pll, int offset, + u32 data) +{ + writel(data, pll->mmio_qserdes_com + offset); +} + +static inline u32 hdmi_pll_read(struct hdmi_pll_8998 *pll, int offset) +{ + return readl(pll->mmio_qserdes_com + offset); +} + +static inline void hdmi_tx_chan_write(struct hdmi_pll_8998 *pll, int channel, + int offset, int data) +{ + writel(data, pll->mmio_qserdes_tx[channel] + offset); +} + +static inline u32 pll_get_cpctrl(u64 frac_start, unsigned long ref_clk, + bool gen_ssc) +{ + if ((frac_start != 0) || gen_ssc) + return 0x8; + + return 0x30; +} + +static inline u32 pll_get_rctrl(u64 frac_start, bool gen_ssc) +{ + if ((frac_start != 0) || gen_ssc) + return 0x16; + + return 0x18; +} + +static inline u32 pll_get_cctrl(u64 frac_start, bool gen_ssc) +{ + if ((frac_start != 0) || gen_ssc) + return 0x34; + + return 0x2; +} + +static inline u32 pll_get_integloop_gain(u64 frac_start, u64 bclk, u32 ref_clk, + bool gen_ssc) +{ + int digclk_divsel = bclk > HDMI_DIG_FREQ_BIT_CLK_THRESHOLD ? 1 : 2; + u64 base; + + if ((frac_start != 0) || gen_ssc) + base = 0x3F; + else + base = 0xC4; + + base <<= (digclk_divsel == 2 ? 1 : 0); + + return (base <= 2046 ? base : 2046); +} + +static inline u32 pll_get_pll_cmp(u64 fdata, unsigned long ref_clk) +{ + u64 dividend = HDMI_PLL_CMP_CNT * fdata; + u32 divisor = ref_clk * 10; + u32 rem; + + rem = do_div(dividend, divisor); + if (rem > (divisor >> 1)) + dividend++; + + return dividend - 1; +} + +static inline u64 pll_cmp_to_fdata(u32 pll_cmp, unsigned long ref_clk) +{ + u64 fdata = ((u64)pll_cmp) * ref_clk * 10; + + do_div(fdata, HDMI_PLL_CMP_CNT); + + return fdata; +} + +#define HDMI_REF_CLOCK_HZ ((u64)19200000) +#define HDMI_MHZ_TO_HZ ((u64)1000000) +static int pll_get_post_div(struct hdmi_8998_post_divider *pd, u64 bclk) +{ + u32 const ratio_list[] = {1, 2, 3, 4, 5, 6, + 9, 10, 12, 15, 25}; + u32 const band_list[] = {0, 1, 2, 3}; + u32 const sz_ratio = ARRAY_SIZE(ratio_list); + u32 const sz_band = ARRAY_SIZE(band_list); + u32 const cmp_cnt = 1024; + u32 const th_min = 500, th_max = 1000; + u32 half_rate_mode = 0; + u32 list_elements; + int optimal_index; + u32 i, j, k; + u32 found_hsclk_divsel = 0, found_vco_ratio; + u32 found_tx_band_sel; + u64 const min_freq = HDMI_VCO_MIN_FREQ, max_freq = HDMI_VCO_MAX_FREQ; + u64 freq_list[ARRAY_SIZE(ratio_list) * ARRAY_SIZE(band_list)]; + u64 found_vco_freq; + u64 freq_optimal; + +find_optimal_index: + freq_optimal = max_freq; + optimal_index = -1; + list_elements = 0; + + for (i = 0; i < sz_ratio; i++) { + for (j = 0; j < sz_band; j++) { + u64 freq = div_u64(bclk, (1 << half_rate_mode)); + + freq *= (ratio_list[i] * (1 << band_list[j])); + freq_list[list_elements++] = freq; + } + } + + for (k = 0; k < ARRAY_SIZE(freq_list); k++) { + u32 const clks_pll_div = 2, core_clk_div = 5; + u32 const rng1 = 16, rng2 = 8; + u32 th1, th2; + u64 core_clk, rvar1, rem; + + core_clk = div_u64(freq_list[k], + ratio_list[k / sz_band] * clks_pll_div * + core_clk_div); + + rvar1 = HDMI_REF_CLOCK_HZ * rng1 * HDMI_MHZ_TO_HZ; + rvar1 = div64_u64_rem(rvar1, (cmp_cnt * core_clk), &rem); + if (rem > ((cmp_cnt * core_clk) >> 1)) + rvar1++; + th1 = rvar1; + + rvar1 = HDMI_REF_CLOCK_HZ * rng2 * HDMI_MHZ_TO_HZ; + rvar1 = div64_u64_rem(rvar1, (cmp_cnt * core_clk), &rem); + if (rem > ((cmp_cnt * core_clk) >> 1)) + rvar1++; + th2 = rvar1; + + if (freq_list[k] >= min_freq && + freq_list[k] <= max_freq) { + if ((th1 >= th_min && th1 <= th_max) || + (th2 >= th_min && th2 <= th_max)) { + if (freq_list[k] <= freq_optimal) { + freq_optimal = freq_list[k]; + optimal_index = k; + } + } + } + } + + if (optimal_index == -1) { + if (!half_rate_mode) { + half_rate_mode = 1; + goto find_optimal_index; + } else { + return -EINVAL; + } + } else { + found_vco_ratio = ratio_list[optimal_index / sz_band]; + found_tx_band_sel = band_list[optimal_index % sz_band]; + found_vco_freq = freq_optimal; + } + + switch (found_vco_ratio) { + case 1: + found_hsclk_divsel = 15; + break; + case 2: + found_hsclk_divsel = 0; + break; + case 3: + found_hsclk_divsel = 4; + break; + case 4: + found_hsclk_divsel = 8; + break; + case 5: + found_hsclk_divsel = 12; + break; + case 6: + found_hsclk_divsel = 1; + break; + case 9: + found_hsclk_divsel = 5; + break; + case 10: + found_hsclk_divsel = 2; + break; + case 12: + found_hsclk_divsel = 9; + break; + case 15: + found_hsclk_divsel = 13; + break; + case 25: + found_hsclk_divsel = 14; + break; + }; + + pd->vco_freq = found_vco_freq; + pd->tx_band_sel = found_tx_band_sel; + pd->vco_ratio = found_vco_ratio; + pd->hsclk_divsel = found_hsclk_divsel; + + return 0; +} + +static int pll_calculate(unsigned long pix_clk, unsigned long ref_clk, + struct hdmi_8998_phy_pll_reg_cfg *cfg) +{ + struct hdmi_8998_post_divider pd; + u64 bclk; + u64 dec_start; + u64 frac_start; + u64 fdata; + u32 pll_divisor; + u32 rem; + u32 cpctrl; + u32 rctrl; + u32 cctrl; + u32 integloop_gain; + u32 pll_cmp; + int i, ret; + + /* bit clk = 10 * pix_clk */ + bclk = ((u64)pix_clk) * 10; + + ret = pll_get_post_div(&pd, bclk); + if (ret) + return ret; + + dec_start = pd.vco_freq; + pll_divisor = 4 * ref_clk; + do_div(dec_start, pll_divisor); + + frac_start = pd.vco_freq * (1 << 20); + + rem = do_div(frac_start, pll_divisor); + frac_start -= dec_start * (1 << 20); + if (rem > (pll_divisor >> 1)) + frac_start++; + + cpctrl = pll_get_cpctrl(frac_start, ref_clk, false); + rctrl = pll_get_rctrl(frac_start, false); + cctrl = pll_get_cctrl(frac_start, false); + integloop_gain = pll_get_integloop_gain(frac_start, bclk, + ref_clk, false); + + fdata = pd.vco_freq; + do_div(fdata, pd.vco_ratio); + + pll_cmp = pll_get_pll_cmp(fdata, ref_clk); + + /* Convert these values to register specific values */ + if (bclk > HDMI_DIG_FREQ_BIT_CLK_THRESHOLD) + cfg->com_svs_mode_clk_sel = 1; + else + cfg->com_svs_mode_clk_sel = 2; + + cfg->com_hsclk_sel = (0x20 | pd.hsclk_divsel); + cfg->com_pll_cctrl_mode0 = cctrl; + cfg->com_pll_rctrl_mode0 = rctrl; + cfg->com_cp_ctrl_mode0 = cpctrl; + cfg->com_dec_start_mode0 = dec_start; + cfg->com_div_frac_start1_mode0 = (frac_start & 0xff); + cfg->com_div_frac_start2_mode0 = ((frac_start & 0xff00) >> 8); + cfg->com_div_frac_start3_mode0 = ((frac_start & 0xf0000) >> 16); + cfg->com_integloop_gain0_mode0 = (integloop_gain & 0xff); + cfg->com_integloop_gain1_mode0 = ((integloop_gain & 0xf00) >> 8); + cfg->com_lock_cmp1_mode0 = (pll_cmp & 0xff); + cfg->com_lock_cmp2_mode0 = ((pll_cmp & 0xff00) >> 8); + cfg->com_lock_cmp3_mode0 = ((pll_cmp & 0x30000) >> 16); + cfg->com_lock_cmp_en = 0x0; + cfg->com_core_clk_en = 0x2c; + cfg->com_coreclk_div_mode0 = HDMI_CORECLK_DIV; + cfg->phy_mode = (bclk > HDMI_HIGH_FREQ_BIT_CLK_THRESHOLD) ? 0x5 : 0x4; + + for (i = 0; i < HDMI_NUM_TX_CHANNEL; i++) + cfg->tx_lx_tx_band[i] = pd.tx_band_sel; + + if (bclk > HDMI_HIGH_FREQ_BIT_CLK_THRESHOLD) { + cfg->tx_lx_tx_drv_lvl[0] = 0x0f; + cfg->tx_lx_tx_drv_lvl[1] = 0x0f; + cfg->tx_lx_tx_drv_lvl[2] = 0x0f; + cfg->tx_lx_tx_drv_lvl[3] = 0x0f; + cfg->tx_lx_tx_emp_post1_lvl[0] = 0x03; + cfg->tx_lx_tx_emp_post1_lvl[1] = 0x02; + cfg->tx_lx_tx_emp_post1_lvl[2] = 0x03; + cfg->tx_lx_tx_emp_post1_lvl[3] = 0x00; + cfg->tx_lx_pre_driver_1[0] = 0x00; + cfg->tx_lx_pre_driver_1[1] = 0x00; + cfg->tx_lx_pre_driver_1[2] = 0x00; + cfg->tx_lx_pre_driver_1[3] = 0x00; + cfg->tx_lx_pre_driver_2[0] = 0x1C; + cfg->tx_lx_pre_driver_2[1] = 0x1C; + cfg->tx_lx_pre_driver_2[2] = 0x1C; + cfg->tx_lx_pre_driver_2[3] = 0x00; + cfg->tx_lx_res_code_offset[0] = 0x03; + cfg->tx_lx_res_code_offset[1] = 0x00; + cfg->tx_lx_res_code_offset[2] = 0x00; + cfg->tx_lx_res_code_offset[3] = 0x03; + } else if (bclk > HDMI_DIG_FREQ_BIT_CLK_THRESHOLD) { + cfg->tx_lx_tx_drv_lvl[0] = 0x0f; + cfg->tx_lx_tx_drv_lvl[1] = 0x0f; + cfg->tx_lx_tx_drv_lvl[2] = 0x0f; + cfg->tx_lx_tx_drv_lvl[3] = 0x0f; + cfg->tx_lx_tx_emp_post1_lvl[0] = 0x03; + cfg->tx_lx_tx_emp_post1_lvl[1] = 0x03; + cfg->tx_lx_tx_emp_post1_lvl[2] = 0x03; + cfg->tx_lx_tx_emp_post1_lvl[3] = 0x00; + cfg->tx_lx_pre_driver_1[0] = 0x00; + cfg->tx_lx_pre_driver_1[1] = 0x00; + cfg->tx_lx_pre_driver_1[2] = 0x00; + cfg->tx_lx_pre_driver_1[3] = 0x00; + cfg->tx_lx_pre_driver_2[0] = 0x16; + cfg->tx_lx_pre_driver_2[1] = 0x16; + cfg->tx_lx_pre_driver_2[2] = 0x16; + cfg->tx_lx_pre_driver_2[3] = 0x18; + cfg->tx_lx_res_code_offset[0] = 0x03; + cfg->tx_lx_res_code_offset[1] = 0x00; + cfg->tx_lx_res_code_offset[2] = 0x00; + cfg->tx_lx_res_code_offset[3] = 0x00; + } else if (bclk > HDMI_MID_FREQ_BIT_CLK_THRESHOLD) { + cfg->tx_lx_tx_drv_lvl[0] = 0x0f; + cfg->tx_lx_tx_drv_lvl[1] = 0x0f; + cfg->tx_lx_tx_drv_lvl[2] = 0x0f; + cfg->tx_lx_tx_drv_lvl[3] = 0x0f; + cfg->tx_lx_tx_emp_post1_lvl[0] = 0x05; + cfg->tx_lx_tx_emp_post1_lvl[1] = 0x05; + cfg->tx_lx_tx_emp_post1_lvl[2] = 0x05; + cfg->tx_lx_tx_emp_post1_lvl[3] = 0x00; + cfg->tx_lx_pre_driver_1[0] = 0x00; + cfg->tx_lx_pre_driver_1[1] = 0x00; + cfg->tx_lx_pre_driver_1[2] = 0x00; + cfg->tx_lx_pre_driver_1[3] = 0x00; + cfg->tx_lx_pre_driver_2[0] = 0x0E; + cfg->tx_lx_pre_driver_2[1] = 0x0E; + cfg->tx_lx_pre_driver_2[2] = 0x0E; + cfg->tx_lx_pre_driver_2[3] = 0x0E; + cfg->tx_lx_res_code_offset[0] = 0x00; + cfg->tx_lx_res_code_offset[1] = 0x00; + cfg->tx_lx_res_code_offset[2] = 0x00; + cfg->tx_lx_res_code_offset[3] = 0x00; + } else { + cfg->tx_lx_tx_drv_lvl[0] = 0x01; + cfg->tx_lx_tx_drv_lvl[1] = 0x01; + cfg->tx_lx_tx_drv_lvl[2] = 0x01; + cfg->tx_lx_tx_drv_lvl[3] = 0x00; + cfg->tx_lx_tx_emp_post1_lvl[0] = 0x00; + cfg->tx_lx_tx_emp_post1_lvl[1] = 0x00; + cfg->tx_lx_tx_emp_post1_lvl[2] = 0x00; + cfg->tx_lx_tx_emp_post1_lvl[3] = 0x00; + cfg->tx_lx_pre_driver_1[0] = 0x00; + cfg->tx_lx_pre_driver_1[1] = 0x00; + cfg->tx_lx_pre_driver_1[2] = 0x00; + cfg->tx_lx_pre_driver_1[3] = 0x00; + cfg->tx_lx_pre_driver_2[0] = 0x16; + cfg->tx_lx_pre_driver_2[1] = 0x16; + cfg->tx_lx_pre_driver_2[2] = 0x16; + cfg->tx_lx_pre_driver_2[3] = 0x18; + cfg->tx_lx_res_code_offset[0] = 0x00; + cfg->tx_lx_res_code_offset[1] = 0x00; + cfg->tx_lx_res_code_offset[2] = 0x00; + cfg->tx_lx_res_code_offset[3] = 0x00; + } + + return 0; +} + +static int hdmi_8998_pll_set_clk_rate(struct clk_hw *hw, unsigned long rate, + unsigned long parent_rate) +{ + struct hdmi_pll_8998 *pll = hw_clk_to_pll(hw); + struct hdmi_phy *phy = pll_get_phy(pll); + struct hdmi_8998_phy_pll_reg_cfg cfg = {}; + int i, ret; + + ret = pll_calculate(rate, parent_rate, &cfg); + if (ret) { + DRM_ERROR("PLL calculation failed\n"); + return ret; + } + + /* Initially shut down PHY */ + hdmi_phy_write(phy, REG_HDMI_8998_PHY_PD_CTL, 0x0); + udelay(500); + + /* Power up sequence */ + hdmi_phy_write(phy, REG_HDMI_8998_PHY_PD_CTL, 0x1); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_RESETSM_CNTRL, 0x20); + hdmi_phy_write(phy, REG_HDMI_8998_PHY_CMN_CTRL, 0x6); + + for (i = 0; i < HDMI_NUM_TX_CHANNEL; i++) { + hdmi_tx_chan_write(pll, i, + REG_HDMI_8998_PHY_TXn_INTERFACE_SELECT_TX_BAND, + cfg.tx_lx_tx_band[i]); + hdmi_tx_chan_write(pll, i, + REG_HDMI_8998_PHY_TXn_CLKBUF_TERM_ENABLE, + 0x1); + hdmi_tx_chan_write(pll, i, + REG_HDMI_8998_PHY_TXn_LANE_MODE, + 0x20); + } + + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_SYSCLK_BUF_ENABLE, 0x02); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_BIAS_EN_CLKBUFLR_EN, 0x0B); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_SYSCLK_EN_SEL, 0x37); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_SYS_CLK_CTRL, 0x02); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_CLK_ENABLE1, 0x0E); + + /* Bypass VCO calibration */ + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_SVS_MODE_CLK_SEL, + cfg.com_svs_mode_clk_sel); + + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_PLL_IVCO, 0x07); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_VCO_TUNE_CTRL, 0x00); + + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_CLK_SEL, 0x30); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_HSCLK_SEL, + cfg.com_hsclk_sel); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_LOCK_CMP_EN, + cfg.com_lock_cmp_en); + + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_PLL_CCTRL_MODE0, + cfg.com_pll_cctrl_mode0); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_PLL_RCTRL_MODE0, + cfg.com_pll_rctrl_mode0); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_CP_CTRL_MODE0, + cfg.com_cp_ctrl_mode0); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_DEC_START_MODE0, + cfg.com_dec_start_mode0); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_DIV_FRAC_START1_MODE0, + cfg.com_div_frac_start1_mode0); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_DIV_FRAC_START2_MODE0, + cfg.com_div_frac_start2_mode0); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_DIV_FRAC_START3_MODE0, + cfg.com_div_frac_start3_mode0); + + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_INTEGLOOP_GAIN0_MODE0, + cfg.com_integloop_gain0_mode0); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_INTEGLOOP_GAIN1_MODE0, + cfg.com_integloop_gain1_mode0); + + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_LOCK_CMP1_MODE0, + cfg.com_lock_cmp1_mode0); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_LOCK_CMP2_MODE0, + cfg.com_lock_cmp2_mode0); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_LOCK_CMP3_MODE0, + cfg.com_lock_cmp3_mode0); + + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_VCO_TUNE_MAP, 0x00); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_CORE_CLK_EN, + cfg.com_core_clk_en); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_CORECLK_DIV_MODE0, + cfg.com_coreclk_div_mode0); + + /* TX lanes setup (TX 0/1/2/3) */ + for (i = 0; i < HDMI_NUM_TX_CHANNEL; i++) { + hdmi_tx_chan_write(pll, i, + REG_HDMI_8998_PHY_TXn_DRV_LVL, + cfg.tx_lx_tx_drv_lvl[i]); + hdmi_tx_chan_write(pll, i, + REG_HDMI_8998_PHY_TXn_EMP_POST1_LVL, + cfg.tx_lx_tx_emp_post1_lvl[i]); + hdmi_tx_chan_write(pll, i, + REG_HDMI_8998_PHY_TXn_PRE_DRIVER_1, + cfg.tx_lx_pre_driver_1[i]); + hdmi_tx_chan_write(pll, i, + REG_HDMI_8998_PHY_TXn_PRE_DRIVER_2, + cfg.tx_lx_pre_driver_2[i]); + hdmi_tx_chan_write(pll, i, + REG_HDMI_8998_PHY_TXn_DRV_LVL_RES_CODE_OFFSET, + cfg.tx_lx_res_code_offset[i]); + } + + hdmi_phy_write(phy, REG_HDMI_8998_PHY_MODE, cfg.phy_mode); + + for (i = 0; i < HDMI_NUM_TX_CHANNEL; i++) { + hdmi_tx_chan_write(pll, i, + REG_HDMI_8998_PHY_TXn_LANE_CONFIG, + 0x10); + } + + /* + * Ensure that vco configuration gets flushed to hardware before + * enabling the PLL + */ + wmb(); + + pll->rate = rate; + + return 0; +} + +static int hdmi_8998_phy_ready_status(struct hdmi_phy *phy) +{ + u32 nb_tries = HDMI_PLL_POLL_MAX_READS; + unsigned long timeout = HDMI_PLL_POLL_TIMEOUT_US; + u32 status; + int phy_ready = 0; + + while (nb_tries--) { + status = hdmi_phy_read(phy, REG_HDMI_8998_PHY_STATUS); + phy_ready = status & BIT(0); + + if (phy_ready) + break; + + udelay(timeout); + } + + return phy_ready; +} + +static int hdmi_8998_pll_lock_status(struct hdmi_pll_8998 *pll) +{ + u32 status; + int nb_tries = HDMI_PLL_POLL_MAX_READS; + unsigned long timeout = HDMI_PLL_POLL_TIMEOUT_US; + int pll_locked = 0; + + while (nb_tries--) { + status = hdmi_pll_read(pll, + REG_HDMI_8998_PHY_QSERDES_COM_C_READY_STATUS); + pll_locked = status & BIT(0); + + if (pll_locked) + break; + + udelay(timeout); + } + + return pll_locked; +} + +static int hdmi_8998_pll_prepare(struct clk_hw *hw) +{ + struct hdmi_pll_8998 *pll = hw_clk_to_pll(hw); + struct hdmi_phy *phy = pll_get_phy(pll); + int i, ret = 0; + + hdmi_phy_write(phy, REG_HDMI_8998_PHY_CFG, 0x1); + udelay(100); + + hdmi_phy_write(phy, REG_HDMI_8998_PHY_CFG, 0x59); + udelay(100); + + ret = hdmi_8998_pll_lock_status(pll); + if (!ret) + return ret; + + for (i = 0; i < HDMI_NUM_TX_CHANNEL; i++) { + hdmi_tx_chan_write(pll, i, + REG_HDMI_8998_PHY_TXn_LANE_CONFIG, 0x1F); + } + + /* Ensure all registers are flushed to hardware */ + wmb(); + + ret = hdmi_8998_phy_ready_status(phy); + if (!ret) + return ret; + + /* Restart the retiming buffer */ + hdmi_phy_write(phy, REG_HDMI_8998_PHY_CFG, 0x58); + udelay(1); + hdmi_phy_write(phy, REG_HDMI_8998_PHY_CFG, 0x59); + + /* Ensure all registers are flushed to hardware */ + wmb(); + + return 0; +} + +static long hdmi_8998_pll_round_rate(struct clk_hw *hw, + unsigned long rate, + unsigned long *parent_rate) +{ + if (rate < HDMI_PCLK_MIN_FREQ) + return HDMI_PCLK_MIN_FREQ; + else if (rate > HDMI_PCLK_MAX_FREQ) + return HDMI_PCLK_MAX_FREQ; + else + return rate; +} + +static unsigned long hdmi_8998_pll_recalc_rate(struct clk_hw *hw, + unsigned long parent_rate) +{ + struct hdmi_pll_8998 *pll = hw_clk_to_pll(hw); + return pll->rate; +} + +static void hdmi_8998_pll_unprepare(struct clk_hw *hw) +{ + struct hdmi_pll_8998 *pll = hw_clk_to_pll(hw); + struct hdmi_phy *phy = pll_get_phy(pll); + + hdmi_phy_write(phy, REG_HDMI_8998_PHY_PD_CTL, 0); + usleep_range(100, 150); +} + +static int hdmi_8998_pll_is_enabled(struct clk_hw *hw) +{ + struct hdmi_pll_8998 *pll = hw_clk_to_pll(hw); + u32 status; + int pll_locked; + + status = hdmi_pll_read(pll, REG_HDMI_8998_PHY_QSERDES_COM_C_READY_STATUS); + pll_locked = status & BIT(0); + + return pll_locked; +} + +static const struct clk_ops hdmi_8998_pll_ops = { + .set_rate = hdmi_8998_pll_set_clk_rate, + .round_rate = hdmi_8998_pll_round_rate, + .recalc_rate = hdmi_8998_pll_recalc_rate, + .prepare = hdmi_8998_pll_prepare, + .unprepare = hdmi_8998_pll_unprepare, + .is_enabled = hdmi_8998_pll_is_enabled, +}; + +static const struct clk_init_data pll_init = { + .name = "hdmipll", + .ops = &hdmi_8998_pll_ops, + .parent_data = (const struct clk_parent_data[]){ + { .fw_name = "xo", .name = "xo_board" }, + }, + .num_parents = 1, + .flags = CLK_IGNORE_UNUSED, +}; + +int msm_hdmi_pll_8998_init(struct platform_device *pdev) +{ + struct device *dev = &pdev->dev; + struct hdmi_pll_8998 *pll; + int ret, i; + + pll = devm_kzalloc(dev, sizeof(*pll), GFP_KERNEL); + if (!pll) + return -ENOMEM; + + pll->pdev = pdev; + + pll->mmio_qserdes_com = msm_ioremap(pdev, "hdmi_pll"); + if (IS_ERR(pll->mmio_qserdes_com)) { + DRM_DEV_ERROR(dev, "failed to map pll base\n"); + return -ENOMEM; + } + + for (i = 0; i < HDMI_NUM_TX_CHANNEL; i++) { + char name[32]; + + snprintf(name, sizeof(name), "hdmi_tx_l%d", i); + + pll->mmio_qserdes_tx[i] = msm_ioremap(pdev, name); + if (IS_ERR(pll->mmio_qserdes_tx[i])) { + DRM_DEV_ERROR(dev, "failed to map pll base\n"); + return -ENOMEM; + } + } + pll->clk_hw.init = &pll_init; + + ret = devm_clk_hw_register(dev, &pll->clk_hw); + if (ret) { + DRM_DEV_ERROR(dev, "failed to register pll clock\n"); + return ret; + } + + ret = devm_of_clk_add_hw_provider(dev, of_clk_hw_simple_get, &pll->clk_hw); + if (ret) { + DRM_DEV_ERROR(dev, "failed to register clk provider: %d\n", ret); + return ret; + } + + return 0; +} + +static const char * const hdmi_phy_8998_reg_names[] = { + "vddio", + "vcca", +}; + +static const char * const hdmi_phy_8998_clk_names[] = { + "iface", "ref", "xo", +}; + +const struct hdmi_phy_cfg msm_hdmi_phy_8998_cfg = { + .type = MSM_HDMI_PHY_8998, + .reg_names = hdmi_phy_8998_reg_names, + .num_regs = ARRAY_SIZE(hdmi_phy_8998_reg_names), + .clk_names = hdmi_phy_8998_clk_names, + .num_clks = ARRAY_SIZE(hdmi_phy_8998_clk_names), +}; diff --git a/drivers/gpu/drm/msm/msm_drv.h b/drivers/gpu/drm/msm/msm_drv.h index be016d7b4ef1..0d3adf398bc1 100644 --- a/drivers/gpu/drm/msm/msm_drv.h +++ b/drivers/gpu/drm/msm/msm_drv.h @@ -215,8 +215,6 @@ struct msm_drm_private { struct notifier_block vmap_notifier; struct shrinker *shrinker; - struct drm_atomic_state *pm_state; - /** * hangcheck_period: For hang detection, in ms * @@ -254,8 +252,6 @@ void msm_atomic_destroy_pending_timer(struct msm_pending_timer *timer); void msm_atomic_commit_tail(struct drm_atomic_state *state); int msm_atomic_check(struct drm_device *dev, struct drm_atomic_state *state); struct drm_atomic_state *msm_atomic_state_alloc(struct drm_device *dev); -void msm_atomic_state_clear(struct drm_atomic_state *state); -void msm_atomic_state_free(struct drm_atomic_state *state); int msm_crtc_enable_vblank(struct drm_crtc *crtc); void msm_crtc_disable_vblank(struct drm_crtc *crtc); diff --git a/drivers/gpu/drm/msm/msm_gpu.c b/drivers/gpu/drm/msm/msm_gpu.c index 3666b42b4ecd..a274b8466423 100644 --- a/drivers/gpu/drm/msm/msm_gpu.c +++ b/drivers/gpu/drm/msm/msm_gpu.c @@ -931,7 +931,6 @@ int msm_gpu_init(struct drm_device *drm, struct platform_device *pdev, if (IS_ERR(gpu->gpu_cx)) gpu->gpu_cx = NULL; - gpu->pdev = pdev; platform_set_drvdata(pdev, &gpu->adreno_smmu); msm_devfreq_init(gpu); diff --git a/drivers/gpu/drm/msm/registers/adreno/a6xx.xml b/drivers/gpu/drm/msm/registers/adreno/a6xx.xml index 2dfe6913ab4f..97608603ea62 100644 --- a/drivers/gpu/drm/msm/registers/adreno/a6xx.xml +++ b/drivers/gpu/drm/msm/registers/adreno/a6xx.xml @@ -1198,6 +1198,1027 @@ to upconvert to 32b float internally? <value value="0x3" name="TESS_CCW_TRIS"/> </enum> +<enum name="a7xx_cp_perfcounter_select"> + <value value="0" name="A7XX_PERF_CP_ALWAYS_COUNT"/> + <value value="1" name="A7XX_PERF_CP_BUSY_GFX_CORE_IDLE"/> + <value value="2" name="A7XX_PERF_CP_BUSY_CYCLES"/> + <value value="3" name="A7XX_PERF_CP_NUM_PREEMPTIONS"/> + <value value="4" name="A7XX_PERF_CP_PREEMPTION_REACTION_DELAY"/> + <value value="5" name="A7XX_PERF_CP_PREEMPTION_SWITCH_OUT_TIME"/> + <value value="6" name="A7XX_PERF_CP_PREEMPTION_SWITCH_IN_TIME"/> + <value value="7" name="A7XX_PERF_CP_DEAD_DRAWS_IN_BIN_RENDER"/> + <value value="8" name="A7XX_PERF_CP_PREDICATED_DRAWS_KILLED"/> + <value value="9" name="A7XX_PERF_CP_MODE_SWITCH"/> + <value value="10" name="A7XX_PERF_CP_ZPASS_DONE"/> + <value value="11" name="A7XX_PERF_CP_CONTEXT_DONE"/> + <value value="12" name="A7XX_PERF_CP_CACHE_FLUSH"/> + <value value="13" name="A7XX_PERF_CP_LONG_PREEMPTIONS"/> + <value value="14" name="A7XX_PERF_CP_SQE_I_CACHE_STARVE"/> + <value value="15" name="A7XX_PERF_CP_SQE_IDLE"/> + <value value="16" name="A7XX_PERF_CP_SQE_PM4_STARVE_RB_IB"/> + <value value="17" name="A7XX_PERF_CP_SQE_PM4_STARVE_SDS"/> + <value value="18" name="A7XX_PERF_CP_SQE_MRB_STARVE"/> + <value value="19" name="A7XX_PERF_CP_SQE_RRB_STARVE"/> + <value value="20" name="A7XX_PERF_CP_SQE_VSD_STARVE"/> + <value value="21" name="A7XX_PERF_CP_VSD_DECODE_STARVE"/> + <value value="22" name="A7XX_PERF_CP_SQE_PIPE_OUT_STALL"/> + <value value="23" name="A7XX_PERF_CP_SQE_SYNC_STALL"/> + <value value="24" name="A7XX_PERF_CP_SQE_PM4_WFI_STALL"/> + <value value="25" name="A7XX_PERF_CP_SQE_SYS_WFI_STALL"/> + <value value="26" name="A7XX_PERF_CP_SQE_T4_EXEC"/> + <value value="27" name="A7XX_PERF_CP_SQE_LOAD_STATE_EXEC"/> + <value value="28" name="A7XX_PERF_CP_SQE_SAVE_SDS_STATE"/> + <value value="29" name="A7XX_PERF_CP_SQE_DRAW_EXEC"/> + <value value="30" name="A7XX_PERF_CP_SQE_CTXT_REG_BUNCH_EXEC"/> + <value value="31" name="A7XX_PERF_CP_SQE_EXEC_PROFILED"/> + <value value="32" name="A7XX_PERF_CP_MEMORY_POOL_EMPTY"/> + <value value="33" name="A7XX_PERF_CP_MEMORY_POOL_SYNC_STALL"/> + <value value="34" name="A7XX_PERF_CP_MEMORY_POOL_ABOVE_THRESH"/> + <value value="35" name="A7XX_PERF_CP_AHB_WR_STALL_PRE_DRAWS"/> + <value value="36" name="A7XX_PERF_CP_AHB_STALL_SQE_GMU"/> + <value value="37" name="A7XX_PERF_CP_AHB_STALL_SQE_WR_OTHER"/> + <value value="38" name="A7XX_PERF_CP_AHB_STALL_SQE_RD_OTHER"/> + <value value="39" name="A7XX_PERF_CP_CLUSTER0_EMPTY"/> + <value value="40" name="A7XX_PERF_CP_CLUSTER1_EMPTY"/> + <value value="41" name="A7XX_PERF_CP_CLUSTER2_EMPTY"/> + <value value="42" name="A7XX_PERF_CP_CLUSTER3_EMPTY"/> + <value value="43" name="A7XX_PERF_CP_CLUSTER4_EMPTY"/> + <value value="44" name="A7XX_PERF_CP_CLUSTER5_EMPTY"/> + <value value="45" name="A7XX_PERF_CP_PM4_DATA"/> + <value value="46" name="A7XX_PERF_CP_PM4_HEADERS"/> + <value value="47" name="A7XX_PERF_CP_VBIF_READ_BEATS"/> + <value value="48" name="A7XX_PERF_CP_VBIF_WRITE_BEATS"/> + <value value="49" name="A7XX_PERF_CP_SQE_INSTR_COUNTER"/> + <value value="50" name="A7XX_PERF_CP_RESERVED_50"/> + <value value="51" name="A7XX_PERF_CP_RESERVED_51"/> + <value value="52" name="A7XX_PERF_CP_RESERVED_52"/> + <value value="53" name="A7XX_PERF_CP_RESERVED_53"/> + <value value="54" name="A7XX_PERF_CP_RESERVED_54"/> + <value value="55" name="A7XX_PERF_CP_RESERVED_55"/> + <value value="56" name="A7XX_PERF_CP_RESERVED_56"/> + <value value="57" name="A7XX_PERF_CP_RESERVED_57"/> + <value value="58" name="A7XX_PERF_CP_RESERVED_58"/> + <value value="59" name="A7XX_PERF_CP_RESERVED_59"/> + <value value="60" name="A7XX_PERF_CP_CLUSTER0_FULL"/> + <value value="61" name="A7XX_PERF_CP_CLUSTER1_FULL"/> + <value value="62" name="A7XX_PERF_CP_CLUSTER2_FULL"/> + <value value="63" name="A7XX_PERF_CP_CLUSTER3_FULL"/> + <value value="64" name="A7XX_PERF_CP_CLUSTER4_FULL"/> + <value value="65" name="A7XX_PERF_CP_CLUSTER5_FULL"/> + <value value="66" name="A7XX_PERF_CP_CLUSTER6_FULL"/> + <value value="67" name="A7XX_PERF_CP_CLUSTER6_EMPTY"/> + <value value="68" name="A7XX_PERF_CP_ICACHE_MISSES"/> + <value value="69" name="A7XX_PERF_CP_ICACHE_HITS"/> + <value value="70" name="A7XX_PERF_CP_ICACHE_STALL"/> + <value value="71" name="A7XX_PERF_CP_DCACHE_MISSES"/> + <value value="72" name="A7XX_PERF_CP_DCACHE_HITS"/> + <value value="73" name="A7XX_PERF_CP_DCACHE_STALLS"/> + <value value="74" name="A7XX_PERF_CP_AQE_SQE_STALL"/> + <value value="75" name="A7XX_PERF_CP_SQE_AQE_STARVE"/> + <value value="76" name="A7XX_PERF_CP_PREEMPT_LATENCY"/> + <value value="77" name="A7XX_PERF_CP_SQE_MD8_STALL_CYCLES"/> + <value value="78" name="A7XX_PERF_CP_SQE_MESH_EXEC_CYCLES"/> + <value value="79" name="A7XX_PERF_CP_AQE_NUM_AS_CHUNKS"/> + <value value="80" name="A7XX_PERF_CP_AQE_NUM_MS_CHUNKS"/> +</enum> + +<enum name="a7xx_rbbm_perfcounter_select"> + <value value="0" name="A7XX_PERF_RBBM_ALWAYS_COUNT"/> + <value value="1" name="A7XX_PERF_RBBM_ALWAYS_ON"/> + <value value="2" name="A7XX_PERF_RBBM_TSE_BUSY"/> + <value value="3" name="A7XX_PERF_RBBM_RAS_BUSY"/> + <value value="4" name="A7XX_PERF_RBBM_PC_DCALL_BUSY"/> + <value value="5" name="A7XX_PERF_RBBM_PC_VSD_BUSY"/> + <value value="6" name="A7XX_PERF_RBBM_STATUS_MASKED"/> + <value value="7" name="A7XX_PERF_RBBM_COM_BUSY"/> + <value value="8" name="A7XX_PERF_RBBM_DCOM_BUSY"/> + <value value="9" name="A7XX_PERF_RBBM_VBIF_BUSY"/> + <value value="10" name="A7XX_PERF_RBBM_VSC_BUSY"/> + <value value="11" name="A7XX_PERF_RBBM_TESS_BUSY"/> + <value value="12" name="A7XX_PERF_RBBM_UCHE_BUSY"/> + <value value="13" name="A7XX_PERF_RBBM_HLSQ_BUSY"/> +</enum> + +<enum name="a7xx_pc_perfcounter_select"> + <value value="0" name="A7XX_PERF_PC_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_PC_WORKING_CYCLES"/> + <value value="2" name="A7XX_PERF_PC_STALL_CYCLES_VFD"/> + <value value="3" name="A7XX_PERF_PC_RESERVED"/> + <value value="4" name="A7XX_PERF_PC_STALL_CYCLES_VPC"/> + <value value="5" name="A7XX_PERF_PC_STALL_CYCLES_UCHE"/> + <value value="6" name="A7XX_PERF_PC_STALL_CYCLES_TESS"/> + <value value="7" name="A7XX_PERF_PC_STALL_CYCLES_VFD_ONLY"/> + <value value="8" name="A7XX_PERF_PC_STALL_CYCLES_VPC_ONLY"/> + <value value="9" name="A7XX_PERF_PC_PASS1_TF_STALL_CYCLES"/> + <value value="10" name="A7XX_PERF_PC_STARVE_CYCLES_FOR_INDEX"/> + <value value="11" name="A7XX_PERF_PC_STARVE_CYCLES_FOR_TESS_FACTOR"/> + <value value="12" name="A7XX_PERF_PC_STARVE_CYCLES_FOR_VIZ_STREAM"/> + <value value="13" name="A7XX_PERF_PC_STARVE_CYCLES_DI"/> + <value value="14" name="A7XX_PERF_PC_VIS_STREAMS_LOADED"/> + <value value="15" name="A7XX_PERF_PC_INSTANCES"/> + <value value="16" name="A7XX_PERF_PC_VPC_PRIMITIVES"/> + <value value="17" name="A7XX_PERF_PC_DEAD_PRIM"/> + <value value="18" name="A7XX_PERF_PC_LIVE_PRIM"/> + <value value="19" name="A7XX_PERF_PC_VERTEX_HITS"/> + <value value="20" name="A7XX_PERF_PC_IA_VERTICES"/> + <value value="21" name="A7XX_PERF_PC_IA_PRIMITIVES"/> + <value value="22" name="A7XX_PERF_PC_RESERVED_22"/> + <value value="23" name="A7XX_PERF_PC_HS_INVOCATIONS"/> + <value value="24" name="A7XX_PERF_PC_DS_INVOCATIONS"/> + <value value="25" name="A7XX_PERF_PC_VS_INVOCATIONS"/> + <value value="26" name="A7XX_PERF_PC_GS_INVOCATIONS"/> + <value value="27" name="A7XX_PERF_PC_DS_PRIMITIVES"/> + <value value="28" name="A7XX_PERF_PC_3D_DRAWCALLS"/> + <value value="29" name="A7XX_PERF_PC_2D_DRAWCALLS"/> + <value value="30" name="A7XX_PERF_PC_NON_DRAWCALL_GLOBAL_EVENTS"/> + <value value="31" name="A7XX_PERF_PC_TESS_BUSY_CYCLES"/> + <value value="32" name="A7XX_PERF_PC_TESS_WORKING_CYCLES"/> + <value value="33" name="A7XX_PERF_PC_TESS_STALL_CYCLES_PC"/> + <value value="34" name="A7XX_PERF_PC_TESS_STARVE_CYCLES_PC"/> + <value value="35" name="A7XX_PERF_PC_TESS_SINGLE_PRIM_CYCLES"/> + <value value="36" name="A7XX_PERF_PC_TESS_PC_UV_TRANS"/> + <value value="37" name="A7XX_PERF_PC_TESS_PC_UV_PATCHES"/> + <value value="38" name="A7XX_PERF_PC_TESS_FACTOR_TRANS"/> + <value value="39" name="A7XX_PERF_PC_TAG_CHECKED_VERTICES"/> + <value value="40" name="A7XX_PERF_PC_MESH_VS_WAVES"/> + <value value="41" name="A7XX_PERF_PC_MESH_DRAWS"/> + <value value="42" name="A7XX_PERF_PC_MESH_DEAD_DRAWS"/> + <value value="43" name="A7XX_PERF_PC_MESH_MVIS_EN_DRAWS"/> + <value value="44" name="A7XX_PERF_PC_MESH_DEAD_PRIM"/> + <value value="45" name="A7XX_PERF_PC_MESH_LIVE_PRIM"/> + <value value="46" name="A7XX_PERF_PC_MESH_PA_EN_PRIM"/> + <value value="47" name="A7XX_PERF_PC_STARVE_CYCLES_FOR_MVIS_STREAM"/> + <value value="48" name="A7XX_PERF_PC_STARVE_CYCLES_PREDRAW"/> + <value value="49" name="A7XX_PERF_PC_STALL_CYCLES_COMPUTE_GFX"/> + <value value="50" name="A7XX_PERF_PC_STALL_CYCLES_GFX_COMPUTE"/> + <value value="51" name="A7XX_PERF_PC_TESS_PC_MULTI_PATCH_TRANS"/> +</enum> + +<enum name="a7xx_vfd_perfcounter_select"> + <value value="0" name="A7XX_PERF_VFD_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_VFD_STALL_CYCLES_UCHE"/> + <value value="2" name="A7XX_PERF_VFD_STALL_CYCLES_VPC_ALLOC"/> + <value value="3" name="A7XX_PERF_VFD_STALL_CYCLES_SP_INFO"/> + <value value="4" name="A7XX_PERF_VFD_STALL_CYCLES_SP_ATTR"/> + <value value="5" name="A7XX_PERF_VFD_STARVE_CYCLES_UCHE"/> + <value value="6" name="A7XX_PERF_VFD_RBUFFER_FULL"/> + <value value="7" name="A7XX_PERF_VFD_ATTR_INFO_FIFO_FULL"/> + <value value="8" name="A7XX_PERF_VFD_DECODED_ATTRIBUTE_BYTES"/> + <value value="9" name="A7XX_PERF_VFD_NUM_ATTRIBUTES"/> + <value value="10" name="A7XX_PERF_VFD_UPPER_SHADER_FIBERS"/> + <value value="11" name="A7XX_PERF_VFD_LOWER_SHADER_FIBERS"/> + <value value="12" name="A7XX_PERF_VFD_MODE_0_FIBERS"/> + <value value="13" name="A7XX_PERF_VFD_MODE_1_FIBERS"/> + <value value="14" name="A7XX_PERF_VFD_MODE_2_FIBERS"/> + <value value="15" name="A7XX_PERF_VFD_MODE_3_FIBERS"/> + <value value="16" name="A7XX_PERF_VFD_MODE_4_FIBERS"/> + <value value="17" name="A7XX_PERF_VFD_TOTAL_VERTICES"/> + <value value="18" name="A7XX_PERF_VFDP_STALL_CYCLES_VFD"/> + <value value="19" name="A7XX_PERF_VFDP_STALL_CYCLES_VFD_INDEX"/> + <value value="20" name="A7XX_PERF_VFDP_STALL_CYCLES_VFD_PROG"/> + <value value="21" name="A7XX_PERF_VFDP_STARVE_CYCLES_PC"/> + <value value="22" name="A7XX_PERF_VFDP_VS_STAGE_WAVES"/> + <value value="23" name="A7XX_PERF_VFD_STALL_CYCLES_PRG_END_FE"/> + <value value="24" name="A7XX_PERF_VFD_STALL_CYCLES_CBSYNC"/> +</enum> + +<enum name="a7xx_hlsq_perfcounter_select"> + <value value="0" name="A7XX_PERF_HLSQ_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_HLSQ_STALL_CYCLES_UCHE"/> + <value value="2" name="A7XX_PERF_HLSQ_STALL_CYCLES_SP_STATE"/> + <value value="3" name="A7XX_PERF_HLSQ_STALL_CYCLES_SP_FS_STAGE"/> + <value value="4" name="A7XX_PERF_HLSQ_UCHE_LATENCY_CYCLES"/> + <value value="5" name="A7XX_PERF_HLSQ_UCHE_LATENCY_COUNT"/> + <value value="6" name="A7XX_PERF_HLSQ_RESERVED_6"/> + <value value="7" name="A7XX_PERF_HLSQ_RESERVED_7"/> + <value value="8" name="A7XX_PERF_HLSQ_RESERVED_8"/> + <value value="9" name="A7XX_PERF_HLSQ_RESERVED_9"/> + <value value="10" name="A7XX_PERF_HLSQ_COMPUTE_DRAWCALLS"/> + <value value="11" name="A7XX_PERF_HLSQ_FS_DATA_WAIT_PROGRAMMING"/> + <value value="12" name="A7XX_PERF_HLSQ_DUAL_FS_PROG_ACTIVE"/> + <value value="13" name="A7XX_PERF_HLSQ_DUAL_VS_PROG_ACTIVE"/> + <value value="14" name="A7XX_PERF_HLSQ_FS_BATCH_COUNT_ZERO"/> + <value value="15" name="A7XX_PERF_HLSQ_VS_BATCH_COUNT_ZERO"/> + <value value="16" name="A7XX_PERF_HLSQ_WAVE_PENDING_NO_QUAD"/> + <value value="17" name="A7XX_PERF_HLSQ_WAVE_PENDING_NO_PRIM_BASE"/> + <value value="18" name="A7XX_PERF_HLSQ_STALL_CYCLES_VPC"/> + <value value="19" name="A7XX_PERF_HLSQ_RESERVED_19"/> + <value value="20" name="A7XX_PERF_HLSQ_DRAW_MODE_SWITCH_VSFS_SYNC"/> + <value value="21" name="A7XX_PERF_HLSQ_VSBR_STALL_CYCLES"/> + <value value="22" name="A7XX_PERF_HLSQ_FS_STALL_CYCLES"/> + <value value="23" name="A7XX_PERF_HLSQ_LPAC_STALL_CYCLES"/> + <value value="24" name="A7XX_PERF_HLSQ_BV_STALL_CYCLES"/> + <value value="25" name="A7XX_PERF_HLSQ_VSBR_DEREF_CYCLES"/> + <value value="26" name="A7XX_PERF_HLSQ_FS_DEREF_CYCLES"/> + <value value="27" name="A7XX_PERF_HLSQ_LPAC_DEREF_CYCLES"/> + <value value="28" name="A7XX_PERF_HLSQ_BV_DEREF_CYCLES"/> + <value value="29" name="A7XX_PERF_HLSQ_VSBR_S2W_CYCLES"/> + <value value="30" name="A7XX_PERF_HLSQ_FS_S2W_CYCLES"/> + <value value="31" name="A7XX_PERF_HLSQ_LPAC_S2W_CYCLES"/> + <value value="32" name="A7XX_PERF_HLSQ_BV_S2W_CYCLES"/> + <value value="33" name="A7XX_PERF_HLSQ_VSBR_WAIT_FS_S2W"/> + <value value="34" name="A7XX_PERF_HLSQ_FS_WAIT_VS_S2W"/> + <value value="35" name="A7XX_PERF_HLSQ_LPAC_WAIT_VS_S2W"/> + <value value="36" name="A7XX_PERF_HLSQ_BV_WAIT_FS_S2W"/> + <value value="37" name="A7XX_PERF_HLSQ_VS_WAIT_CONST_RESOURCE"/> + <value value="38" name="A7XX_PERF_HLSQ_FS_WAIT_SAME_VS_S2W"/> + <value value="39" name="A7XX_PERF_HLSQ_FS_STARVING_SP"/> + <value value="40" name="A7XX_PERF_HLSQ_VS_DATA_WAIT_PROGRAMMING"/> + <value value="41" name="A7XX_PERF_HLSQ_BV_DATA_WAIT_PROGRAMMING"/> + <value value="42" name="A7XX_PERF_HLSQ_STPROC_WAVE_CONTEXTS_VS"/> + <value value="43" name="A7XX_PERF_HLSQ_STPROC_WAVE_CONTEXT_CYCLES_VS"/> + <value value="44" name="A7XX_PERF_HLSQ_STPROC_WAVE_CONTEXTS_FS"/> + <value value="45" name="A7XX_PERF_HLSQ_STPROC_WAVE_CONTEXT_CYCLES_FS"/> + <value value="46" name="A7XX_PERF_HLSQ_STPROC_WAVE_CONTEXTS_BV"/> + <value value="47" name="A7XX_PERF_HLSQ_STPROC_WAVE_CONTEXT_CYCLES_BV"/> + <value value="48" name="A7XX_PERF_HLSQ_STPROC_WAVE_CONTEXTS_LPAC"/> + <value value="49" name="A7XX_PERF_HLSQ_STPROC_WAVE_CONTEXT_CYCLES_LPAC"/> + <value value="50" name="A7XX_PERF_HLSQ_SPTROC_STCHE_WARMUP_INC_VS"/> + <value value="51" name="A7XX_PERF_HLSQ_SPTROC_STCHE_WARMUP_INC_FS"/> + <value value="52" name="A7XX_PERF_HLSQ_SPTROC_STCHE_WARMUP_INC_BV"/> + <value value="53" name="A7XX_PERF_HLSQ_SPTROC_STCHE_WARMUP_INC_LPAC"/> + <value value="54" name="A7XX_PERF_HLSQ_SPTROC_STCHE_MISS_INC_VS"/> + <value value="55" name="A7XX_PERF_HLSQ_SPTROC_STCHE_MISS_INC_FS"/> + <value value="56" name="A7XX_PERF_HLSQ_SPTROC_STCHE_MISS_INC_BV"/> + <value value="57" name="A7XX_PERF_HLSQ_SPTROC_STCHE_MISS_INC_LPAC"/> +</enum> + +<enum name="a7xx_vpc_perfcounter_select"> + <value value="0" name="A7XX_PERF_VPC_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_VPC_WORKING_CYCLES"/> + <value value="2" name="A7XX_PERF_VPC_STALL_CYCLES_UCHE"/> + <value value="3" name="A7XX_PERF_VPC_STALL_CYCLES_VFD_WACK"/> + <value value="4" name="A7XX_PERF_VPC_STALL_CYCLES_HLSQ_PRIM_ALLOC"/> + <value value="5" name="A7XX_PERF_VPC_RESERVED_5"/> + <value value="6" name="A7XX_PERF_VPC_STALL_CYCLES_SP_LM"/> + <value value="7" name="A7XX_PERF_VPC_STARVE_CYCLES_SP"/> + <value value="8" name="A7XX_PERF_VPC_STARVE_CYCLES_LRZ"/> + <value value="9" name="A7XX_PERF_VPC_PC_PRIMITIVES"/> + <value value="10" name="A7XX_PERF_VPC_SP_COMPONENTS"/> + <value value="11" name="A7XX_PERF_VPC_STALL_CYCLES_VPCRAM_POS"/> + <value value="12" name="A7XX_PERF_VPC_LRZ_ASSIGN_PRIMITIVES"/> + <value value="13" name="A7XX_PERF_VPC_RB_VISIBLE_PRIMITIVES"/> + <value value="14" name="A7XX_PERF_VPC_LM_TRANSACTION"/> + <value value="15" name="A7XX_PERF_VPC_STREAMOUT_TRANSACTION"/> + <value value="16" name="A7XX_PERF_VPC_VS_BUSY_CYCLES"/> + <value value="17" name="A7XX_PERF_VPC_PS_BUSY_CYCLES"/> + <value value="18" name="A7XX_PERF_VPC_VS_WORKING_CYCLES"/> + <value value="19" name="A7XX_PERF_VPC_PS_WORKING_CYCLES"/> + <value value="20" name="A7XX_PERF_VPC_STARVE_CYCLES_RB"/> + <value value="21" name="A7XX_PERF_VPC_NUM_VPCRAM_READ_POS"/> + <value value="22" name="A7XX_PERF_VPC_WIT_FULL_CYCLES"/> + <value value="23" name="A7XX_PERF_VPC_VPCRAM_FULL_CYCLES"/> + <value value="24" name="A7XX_PERF_VPC_LM_FULL_WAIT_FOR_INTP_END"/> + <value value="25" name="A7XX_PERF_VPC_NUM_VPCRAM_WRITE"/> + <value value="26" name="A7XX_PERF_VPC_NUM_VPCRAM_READ_SO"/> + <value value="27" name="A7XX_PERF_VPC_NUM_ATTR_REQ_LM"/> + <value value="28" name="A7XX_PERF_VPC_STALL_CYCLE_TSE"/> + <value value="29" name="A7XX_PERF_VPC_TSE_PRIMITIVES"/> + <value value="30" name="A7XX_PERF_VPC_GS_PRIMITIVES"/> + <value value="31" name="A7XX_PERF_VPC_TSE_TRANSACTIONS"/> + <value value="32" name="A7XX_PERF_VPC_STALL_CYCLES_CCU"/> + <value value="33" name="A7XX_PERF_VPC_NUM_WM_HIT"/> + <value value="34" name="A7XX_PERF_VPC_STALL_DQ_WACK"/> + <value value="35" name="A7XX_PERF_VPC_STALL_CYCLES_CCHE"/> + <value value="36" name="A7XX_PERF_VPC_STARVE_CYCLES_CCHE"/> + <value value="37" name="A7XX_PERF_VPC_NUM_PA_REQ"/> + <value value="38" name="A7XX_PERF_VPC_NUM_LM_REQ_HIT"/> + <value value="39" name="A7XX_PERF_VPC_CCHE_REQBUF_FULL"/> + <value value="40" name="A7XX_PERF_VPC_STALL_CYCLES_LM_ACK"/> + <value value="41" name="A7XX_PERF_VPC_STALL_CYCLES_PRG_END_FE"/> + <value value="42" name="A7XX_PERF_VPC_STALL_CYCLES_PRG_END_PCVS"/> + <value value="43" name="A7XX_PERF_VPC_STALL_CYCLES_PRG_END_VPCPS"/> +</enum> + +<enum name="a7xx_tse_perfcounter_select"> + <value value="0" name="A7XX_PERF_TSE_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_TSE_CLIPPING_CYCLES"/> + <value value="2" name="A7XX_PERF_TSE_STALL_CYCLES_RAS"/> + <value value="3" name="A7XX_PERF_TSE_STALL_CYCLES_LRZ_BARYPLANE"/> + <value value="4" name="A7XX_PERF_TSE_STALL_CYCLES_LRZ_ZPLANE"/> + <value value="5" name="A7XX_PERF_TSE_STARVE_CYCLES_PC"/> + <value value="6" name="A7XX_PERF_TSE_INPUT_PRIM"/> + <value value="7" name="A7XX_PERF_TSE_INPUT_NULL_PRIM"/> + <value value="8" name="A7XX_PERF_TSE_TRIVAL_REJ_PRIM"/> + <value value="9" name="A7XX_PERF_TSE_CLIPPED_PRIM"/> + <value value="10" name="A7XX_PERF_TSE_ZERO_AREA_PRIM"/> + <value value="11" name="A7XX_PERF_TSE_FACENESS_CULLED_PRIM"/> + <value value="12" name="A7XX_PERF_TSE_ZERO_PIXEL_PRIM"/> + <value value="13" name="A7XX_PERF_TSE_OUTPUT_NULL_PRIM"/> + <value value="14" name="A7XX_PERF_TSE_OUTPUT_VISIBLE_PRIM"/> + <value value="15" name="A7XX_PERF_TSE_CINVOCATION"/> + <value value="16" name="A7XX_PERF_TSE_CPRIMITIVES"/> + <value value="17" name="A7XX_PERF_TSE_2D_INPUT_PRIM"/> + <value value="18" name="A7XX_PERF_TSE_2D_ALIVE_CYCLES"/> + <value value="19" name="A7XX_PERF_TSE_CLIP_PLANES"/> +</enum> + +<enum name="a7xx_ras_perfcounter_select"> + <value value="0" name="A7XX_PERF_RAS_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_RAS_SUPERTILE_ACTIVE_CYCLES"/> + <value value="2" name="A7XX_PERF_RAS_STALL_CYCLES_LRZ"/> + <value value="3" name="A7XX_PERF_RAS_STARVE_CYCLES_TSE"/> + <value value="4" name="A7XX_PERF_RAS_SUPER_TILES"/> + <value value="5" name="A7XX_PERF_RAS_8X4_TILES"/> + <value value="6" name="A7XX_PERF_RAS_MASKGEN_ACTIVE"/> + <value value="7" name="A7XX_PERF_RAS_FULLY_COVERED_SUPER_TILES"/> + <value value="8" name="A7XX_PERF_RAS_FULLY_COVERED_8X4_TILES"/> + <value value="9" name="A7XX_PERF_RAS_PRIM_KILLED_INVISILBE"/> + <value value="10" name="A7XX_PERF_RAS_SUPERTILE_GEN_ACTIVE_CYCLES"/> + <value value="11" name="A7XX_PERF_RAS_LRZ_INTF_WORKING_CYCLES"/> + <value value="12" name="A7XX_PERF_RAS_BLOCKS"/> + <value value="13" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_0_WORKING_CC_l2"/> + <value value="14" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_1_WORKING_CC_l2"/> + <value value="15" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_2_WORKING_CC_l2"/> + <value value="16" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_3_WORKING_CC_l2"/> + <value value="17" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_4_WORKING_CC_l2"/> + <value value="18" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_5_WORKING_CC_l2"/> + <value value="19" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_6_WORKING_CC_l2"/> + <value value="20" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_7_WORKING_CC_l2"/> + <value value="21" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_8_WORKING_CC_l2"/> + <value value="22" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_9_WORKING_CC_l2"/> + <value value="23" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_10_WORKING_CC_l2"/> + <value value="24" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_11_WORKING_CC_l2"/> + <value value="25" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_12_WORKING_CC_l2"/> + <value value="26" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_13_WORKING_CC_l2"/> + <value value="27" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_14_WORKING_CC_l2"/> + <value value="28" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_15_WORKING_CC_l2"/> + <value value="29" name="A7XX_PERF_RAS_FALSE_PARTIAL_STILE"/> + +</enum> + +<enum name="a7xx_uche_perfcounter_select"> + <value value="0" name="A7XX_PERF_UCHE_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_UCHE_STALL_CYCLES_ARBITER"/> + <value value="2" name="A7XX_PERF_UCHE_VBIF_LATENCY_CYCLES"/> + <value value="3" name="A7XX_PERF_UCHE_VBIF_LATENCY_SAMPLES"/> + <value value="4" name="A7XX_PERF_UCHE_VBIF_READ_BEATS_TP"/> + <value value="5" name="A7XX_PERF_UCHE_VBIF_READ_BEATS_VFD"/> + <value value="6" name="A7XX_PERF_UCHE_VBIF_READ_BEATS_HLSQ"/> + <value value="7" name="A7XX_PERF_UCHE_VBIF_READ_BEATS_LRZ"/> + <value value="8" name="A7XX_PERF_UCHE_VBIF_READ_BEATS_SP"/> + <value value="9" name="A7XX_PERF_UCHE_READ_REQUESTS_TP"/> + <value value="10" name="A7XX_PERF_UCHE_READ_REQUESTS_VFD"/> + <value value="11" name="A7XX_PERF_UCHE_READ_REQUESTS_HLSQ"/> + <value value="12" name="A7XX_PERF_UCHE_READ_REQUESTS_LRZ"/> + <value value="13" name="A7XX_PERF_UCHE_READ_REQUESTS_SP"/> + <value value="14" name="A7XX_PERF_UCHE_WRITE_REQUESTS_LRZ"/> + <value value="15" name="A7XX_PERF_UCHE_WRITE_REQUESTS_SP"/> + <value value="16" name="A7XX_PERF_UCHE_WRITE_REQUESTS_VPC"/> + <value value="17" name="A7XX_PERF_UCHE_WRITE_REQUESTS_VSC"/> + <value value="18" name="A7XX_PERF_UCHE_EVICTS"/> + <value value="19" name="A7XX_PERF_UCHE_BANK_REQ0"/> + <value value="20" name="A7XX_PERF_UCHE_BANK_REQ1"/> + <value value="21" name="A7XX_PERF_UCHE_BANK_REQ2"/> + <value value="22" name="A7XX_PERF_UCHE_BANK_REQ3"/> + <value value="23" name="A7XX_PERF_UCHE_BANK_REQ4"/> + <value value="24" name="A7XX_PERF_UCHE_BANK_REQ5"/> + <value value="25" name="A7XX_PERF_UCHE_BANK_REQ6"/> + <value value="26" name="A7XX_PERF_UCHE_BANK_REQ7"/> + <value value="27" name="A7XX_PERF_UCHE_VBIF_READ_BEATS_CH0"/> + <value value="28" name="A7XX_PERF_UCHE_VBIF_READ_BEATS_CH1"/> + <value value="29" name="A7XX_PERF_UCHE_GMEM_READ_BEATS"/> + <value value="30" name="A7XX_PERF_UCHE_TPH_REF_FULL"/> + <value value="31" name="A7XX_PERF_UCHE_TPH_VICTIM_FULL"/> + <value value="32" name="A7XX_PERF_UCHE_TPH_EXT_FULL"/> + <value value="33" name="A7XX_PERF_UCHE_VBIF_STALL_WRITE_DATA"/> + <value value="34" name="A7XX_PERF_UCHE_DCMP_LATENCY_SAMPLES"/> + <value value="35" name="A7XX_PERF_UCHE_DCMP_LATENCY_CYCLES"/> + <value value="36" name="A7XX_PERF_UCHE_VBIF_READ_BEATS_PC"/> + <value value="37" name="A7XX_PERF_UCHE_READ_REQUESTS_PC"/> + <value value="38" name="A7XX_PERF_UCHE_RAM_READ_REQ"/> + <value value="39" name="A7XX_PERF_UCHE_RAM_WRITE_REQ"/> + <value value="40" name="A7XX_PERF_UCHE_STARVED_CYCLES_VBIF_DECMP"/> + <value value="41" name="A7XX_PERF_UCHE_STALL_CYCLES_DECMP"/> + <value value="42" name="A7XX_PERF_UCHE_ARBITER_STALL_CYCLES_VBIF"/> + <value value="43" name="A7XX_PERF_UCHE_READ_REQUESTS_TP_UBWC"/> + <value value="44" name="A7XX_PERF_UCHE_READ_REQUESTS_TP_NONUBWC"/> + <value value="45" name="A7XX_PERF_UCHE_READ_REQUESTS_TP_GMEM"/> + <value value="46" name="A7XX_PERF_UCHE_LONG_LINE_ALL_EVICTS_KAILUA"/> + <value value="47" name="A7XX_PERF_UCHE_LONG_LINE_PARTIAL_EVICTS_KAILUA"/> + <value value="48" name="A7XX_PERF_UCHE_TPH_CONFLICT_CL_CCHE"/> + <value value="49" name="A7XX_PERF_UCHE_TPH_CONFLICT_CL_OTHER_KAILUA"/> + <value value="50" name="A7XX_PERF_UCHE_DBANK_CONFLICT_CL_CCHE"/> + <value value="51" name="A7XX_PERF_UCHE_DBANK_CONFLICT_CL_OTHER_CLIENTS"/> + <value value="52" name="A7XX_PERF_UCHE_VBIF_WRITE_BEATS_CH0"/> + <value value="53" name="A7XX_PERF_UCHE_VBIF_WRITE_BEATS_CH1"/> + <value value="54" name="A7XX_PERF_UCHE_CCHE_TPH_QUEUE_FULL"/> + <value value="55" name="A7XX_PERF_UCHE_CCHE_DPH_QUEUE_FULL"/> + <value value="56" name="A7XX_PERF_UCHE_GMEM_WRITE_BEATS"/> + <value value="57" name="A7XX_PERF_UCHE_UBWC_READ_BEATS"/> + <value value="58" name="A7XX_PERF_UCHE_UBWC_WRITE_BEATS"/> +</enum> + +<enum name="a7xx_tp_perfcounter_select"> + <value value="0" name="A7XX_PERF_TP_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_TP_STALL_CYCLES_UCHE"/> + <value value="2" name="A7XX_PERF_TP_LATENCY_CYCLES"/> + <value value="3" name="A7XX_PERF_TP_LATENCY_TRANS"/> + <value value="4" name="A7XX_PERF_TP_FLAG_FIFO_DELAY_SAMPLES"/> + <value value="5" name="A7XX_PERF_TP_FLAG_FIFO_DELAY_CYCLES"/> + <value value="6" name="A7XX_PERF_TP_L1_CACHELINE_REQUESTS"/> + <value value="7" name="A7XX_PERF_TP_L1_CACHELINE_MISSES"/> + <value value="8" name="A7XX_PERF_TP_SP_TP_TRANS"/> + <value value="9" name="A7XX_PERF_TP_TP_SP_TRANS"/> + <value value="10" name="A7XX_PERF_TP_OUTPUT_PIXELS"/> + <value value="11" name="A7XX_PERF_TP_FILTER_WORKLOAD_16BIT"/> + <value value="12" name="A7XX_PERF_TP_FILTER_WORKLOAD_32BIT"/> + <value value="13" name="A7XX_PERF_TP_QUADS_RECEIVED"/> + <value value="14" name="A7XX_PERF_TP_QUADS_OFFSET"/> + <value value="15" name="A7XX_PERF_TP_QUADS_SHADOW"/> + <value value="16" name="A7XX_PERF_TP_QUADS_ARRAY"/> + <value value="17" name="A7XX_PERF_TP_QUADS_GRADIENT"/> + <value value="18" name="A7XX_PERF_TP_QUADS_1D"/> + <value value="19" name="A7XX_PERF_TP_QUADS_2D"/> + <value value="20" name="A7XX_PERF_TP_QUADS_BUFFER"/> + <value value="21" name="A7XX_PERF_TP_QUADS_3D"/> + <value value="22" name="A7XX_PERF_TP_QUADS_CUBE"/> + <value value="23" name="A7XX_PERF_TP_DIVERGENT_QUADS_RECEIVED"/> + <value value="24" name="A7XX_PERF_TP_PRT_NON_RESIDENT_EVENTS"/> + <value value="25" name="A7XX_PERF_TP_OUTPUT_PIXELS_POINT"/> + <value value="26" name="A7XX_PERF_TP_OUTPUT_PIXELS_BILINEAR"/> + <value value="27" name="A7XX_PERF_TP_OUTPUT_PIXELS_MIP"/> + <value value="28" name="A7XX_PERF_TP_OUTPUT_PIXELS_ANISO"/> + <value value="29" name="A7XX_PERF_TP_OUTPUT_PIXELS_ZERO_LOD"/> + <value value="30" name="A7XX_PERF_TP_FLAG_CACHE_REQUESTS"/> + <value value="31" name="A7XX_PERF_TP_FLAG_CACHE_MISSES"/> + <value value="32" name="A7XX_PERF_TP_L1_5_L2_REQUESTS"/> + <value value="33" name="A7XX_PERF_TP_2D_OUTPUT_PIXELS"/> + <value value="34" name="A7XX_PERF_TP_2D_OUTPUT_PIXELS_POINT"/> + <value value="35" name="A7XX_PERF_TP_2D_OUTPUT_PIXELS_BILINEAR"/> + <value value="36" name="A7XX_PERF_TP_2D_FILTER_WORKLOAD_16BIT"/> + <value value="37" name="A7XX_PERF_TP_2D_FILTER_WORKLOAD_32BIT"/> + <value value="38" name="A7XX_PERF_TP_TPA2TPC_TRANS"/> + <value value="39" name="A7XX_PERF_TP_L1_MISSES_ASTC_1TILE"/> + <value value="40" name="A7XX_PERF_TP_L1_MISSES_ASTC_2TILE"/> + <value value="41" name="A7XX_PERF_TP_L1_MISSES_ASTC_4TILE"/> + <value value="42" name="A7XX_PERF_TP_L1_5_COMPRESS_REQS"/> + <value value="43" name="A7XX_PERF_TP_L1_5_L2_COMPRESS_MISS"/> + <value value="44" name="A7XX_PERF_TP_L1_BANK_CONFLICT"/> + <value value="45" name="A7XX_PERF_TP_L1_5_MISS_LATENCY_CYCLES"/> + <value value="46" name="A7XX_PERF_TP_L1_5_MISS_LATENCY_TRANS"/> + <value value="47" name="A7XX_PERF_TP_QUADS_CONSTANT_MULTIPLIED"/> + <value value="48" name="A7XX_PERF_TP_FRONTEND_WORKING_CYCLES"/> + <value value="49" name="A7XX_PERF_TP_L1_TAG_WORKING_CYCLES"/> + <value value="50" name="A7XX_PERF_TP_L1_DATA_WRITE_WORKING_CYCLES"/> + <value value="51" name="A7XX_PERF_TP_PRE_L1_DECOM_WORKING_CYCLES"/> + <value value="52" name="A7XX_PERF_TP_BACKEND_WORKING_CYCLES"/> + <value value="53" name="A7XX_PERF_TP_L1_5_CACHE_WORKING_CYCLES"/> + <value value="54" name="A7XX_PERF_TP_STARVE_CYCLES_SP"/> + <value value="55" name="A7XX_PERF_TP_STARVE_CYCLES_UCHE"/> + <value value="56" name="A7XX_PERF_TP_STALL_CYCLES_UFC"/> + <value value="57" name="A7XX_PERF_TP_FORMAT_DECOMP"/> + <value value="58" name="A7XX_PERF_TP_FILTER_POINT_FP16"/> + <value value="59" name="A7XX_PERF_TP_FILTER_POINT_FP32"/> + <value value="60" name="A7XX_PERF_TP_LATENCY_FIFO_FULL"/> + <value value="61" name="A7XX_PERF_TP_RESERVED_61"/> + <value value="62" name="A7XX_PERF_TP_RESERVED_62"/> + <value value="63" name="A7XX_PERF_TP_RESERVED_63"/> + <value value="64" name="A7XX_PERF_TP_RESERVED_64"/> + <value value="65" name="A7XX_PERF_TP_RESERVED_65"/> + <value value="66" name="A7XX_PERF_TP_RESERVED_66"/> + <value value="67" name="A7XX_PERF_TP_RESERVED_67"/> + <value value="68" name="A7XX_PERF_TP_RESERVED_68"/> + <value value="69" name="A7XX_PERF_TP_RESERVED_69"/> + <value value="70" name="A7XX_PERF_TP_RESERVED_70"/> + <value value="71" name="A7XX_PERF_TP_RESERVED_71"/> + <value value="72" name="A7XX_PERF_TP_RESERVED_72"/> + <value value="73" name="A7XX_PERF_TP_RESERVED_73"/> + <value value="74" name="A7XX_PERF_TP_RESERVED_74"/> + <value value="75" name="A7XX_PERF_TP_RESERVED_75"/> + <value value="76" name="A7XX_PERF_TP_RESERVED_76"/> + <value value="77" name="A7XX_PERF_TP_RESERVED_77"/> + <value value="78" name="A7XX_PERF_TP_RESERVED_78"/> + <value value="79" name="A7XX_PERF_TP_RESERVED_79"/> + <value value="80" name="A7XX_PERF_TP_RESERVED_80"/> + <value value="81" name="A7XX_PERF_TP_RESERVED_81"/> + <value value="82" name="A7XX_PERF_TP_RESERVED_82"/> + <value value="83" name="A7XX_PERF_TP_RESERVED_83"/> + <value value="84" name="A7XX_PERF_TP_RESERVED_84"/> + <value value="85" name="A7XX_PERF_TP_RESERVED_85"/> + <value value="86" name="A7XX_PERF_TP_RESERVED_86"/> + <value value="87" name="A7XX_PERF_TP_RESERVED_87"/> + <value value="88" name="A7XX_PERF_TP_RESERVED_88"/> + <value value="89" name="A7XX_PERF_TP_RESERVED_89"/> + <value value="90" name="A7XX_PERF_TP_RESERVED_90"/> + <value value="91" name="A7XX_PERF_TP_RESERVED_91"/> + <value value="92" name="A7XX_PERF_TP_RESERVED_92"/> + <value value="93" name="A7XX_PERF_TP_RESERVED_93"/> + <value value="94" name="A7XX_PERF_TP_RESERVED_94"/> + <value value="95" name="A7XX_PERF_TP_RESERVED_95"/> + <value value="96" name="A7XX_PERF_TP_RESERVED_96"/> + <value value="97" name="A7XX_PERF_TP_RESERVED_97"/> + <value value="98" name="A7XX_PERF_TP_RESERVED_98"/> + <value value="99" name="A7XX_PERF_TP_RESERVED_99"/> + <value value="100" name="A7XX_PERF_TP_RESERVED_100"/> + <value value="101" name="A7XX_PERF_TP_RESERVED_101"/> + <value value="102" name="A7XX_PERF_TP_RESERVED_102"/> + <value value="103" name="A7XX_PERF_TP_RESERVED_103"/> + <value value="104" name="A7XX_PERF_TP_RESERVED_104"/> + <value value="105" name="A7XX_PERF_TP_RESERVED_105"/> + <value value="106" name="A7XX_PERF_TP_RESERVED_106"/> + <value value="107" name="A7XX_PERF_TP_RESERVED_107"/> + <value value="108" name="A7XX_PERF_TP_RESERVED_108"/> + <value value="109" name="A7XX_PERF_TP_RESERVED_109"/> + <value value="110" name="A7XX_PERF_TP_RESERVED_110"/> + <value value="111" name="A7XX_PERF_TP_RESERVED_111"/> + <value value="112" name="A7XX_PERF_TP_RESERVED_112"/> + <value value="113" name="A7XX_PERF_TP_RESERVED_113"/> + <value value="114" name="A7XX_PERF_TP_RESERVED_114"/> + <value value="115" name="A7XX_PERF_TP_RESERVED_115"/> + <value value="116" name="A7XX_PERF_TP_RESERVED_116"/> + <value value="117" name="A7XX_PERF_TP_RESERVED_117"/> + <value value="118" name="A7XX_PERF_TP_RESERVED_118"/> + <value value="119" name="A7XX_PERF_TP_RESERVED_119"/> + <value value="120" name="A7XX_PERF_TP_RESERVED_120"/> + <value value="121" name="A7XX_PERF_TP_RESERVED_121"/> + <value value="122" name="A7XX_PERF_TP_RESERVED_122"/> + <value value="123" name="A7XX_PERF_TP_RESERVED_123"/> + <value value="124" name="A7XX_PERF_TP_RESERVED_124"/> + <value value="125" name="A7XX_PERF_TP_RESERVED_125"/> + <value value="126" name="A7XX_PERF_TP_RESERVED_126"/> + <value value="127" name="A7XX_PERF_TP_RESERVED_127"/> + <value value="128" name="A7XX_PERF_TP_FORMAT_DECOMP_BILINEAR"/> + <value value="129" name="A7XX_PERF_TP_PACKED_POINT_BOTH_VALID_FP16"/> + <value value="130" name="A7XX_PERF_TP_PACKED_POINT_SINGLE_VALID_FP16"/> + <value value="131" name="A7XX_PERF_TP_PACKED_POINT_BOTH_VALID_FP32"/> + <value value="132" name="A7XX_PERF_TP_PACKED_POINT_SINGLE_VALID_FP32"/> +</enum> + +<enum name="a7xx_sp_perfcounter_select"> + <value value="0" name="A7XX_PERF_SP_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_SP_ALU_WORKING_CYCLES"/> + <value value="2" name="A7XX_PERF_SP_EFU_WORKING_CYCLES"/> + <value value="3" name="A7XX_PERF_SP_STALL_CYCLES_VPC"/> + <value value="4" name="A7XX_PERF_SP_STALL_CYCLES_TP"/> + <value value="5" name="A7XX_PERF_SP_STALL_CYCLES_UCHE"/> + <value value="6" name="A7XX_PERF_SP_STALL_CYCLES_RB"/> + <value value="7" name="A7XX_PERF_SP_NON_EXECUTION_CYCLES"/> + <value value="8" name="A7XX_PERF_SP_WAVE_CONTEXTS"/> + <value value="9" name="A7XX_PERF_SP_WAVE_CONTEXT_CYCLES"/> + <value value="10" name="A7XX_PERF_SP_STAGE_WAVE_CYCLES"/> + <value value="11" name="A7XX_PERF_SP_STAGE_WAVE_SAMPLES"/> + <value value="12" name="A7XX_PERF_SP_VS_STAGE_WAVE_CYCLES"/> + <value value="13" name="A7XX_PERF_SP_VS_STAGE_WAVE_SAMPLES"/> + <value value="14" name="A7XX_PERF_SP_FS_STAGE_DURATION_CYCLES"/> + <value value="15" name="A7XX_PERF_SP_VS_STAGE_DURATION_CYCLES"/> + <value value="16" name="A7XX_PERF_SP_WAVE_CTRL_CYCLES"/> + <value value="17" name="A7XX_PERF_SP_WAVE_LOAD_CYCLES"/> + <value value="18" name="A7XX_PERF_SP_WAVE_EMIT_CYCLES"/> + <value value="19" name="A7XX_PERF_SP_WAVE_NOP_CYCLES"/> + <value value="20" name="A7XX_PERF_SP_WAVE_WAIT_CYCLES"/> + <value value="21" name="A7XX_PERF_SP_WAVE_FETCH_CYCLES"/> + <value value="22" name="A7XX_PERF_SP_WAVE_IDLE_CYCLES"/> + <value value="23" name="A7XX_PERF_SP_WAVE_END_CYCLES"/> + <value value="24" name="A7XX_PERF_SP_WAVE_LONG_SYNC_CYCLES"/> + <value value="25" name="A7XX_PERF_SP_WAVE_SHORT_SYNC_CYCLES"/> + <value value="26" name="A7XX_PERF_SP_WAVE_JOIN_CYCLES"/> + <value value="27" name="A7XX_PERF_SP_LM_LOAD_INSTRUCTIONS"/> + <value value="28" name="A7XX_PERF_SP_LM_STORE_INSTRUCTIONS"/> + <value value="29" name="A7XX_PERF_SP_LM_ATOMICS"/> + <value value="30" name="A7XX_PERF_SP_GM_LOAD_INSTRUCTIONS"/> + <value value="31" name="A7XX_PERF_SP_GM_STORE_INSTRUCTIONS"/> + <value value="32" name="A7XX_PERF_SP_GM_ATOMICS"/> + <value value="33" name="A7XX_PERF_SP_VS_STAGE_TEX_INSTRUCTIONS"/> + <value value="34" name="A7XX_PERF_SP_VS_STAGE_EFU_INSTRUCTIONS"/> + <value value="35" name="A7XX_PERF_SP_VS_STAGE_FULL_ALU_INSTRUCTIONS"/> + <value value="36" name="A7XX_PERF_SP_VS_STAGE_HALF_ALU_INSTRUCTIONS"/> + <value value="37" name="A7XX_PERF_SP_FS_STAGE_TEX_INSTRUCTIONS"/> + <value value="38" name="A7XX_PERF_SP_FS_STAGE_CFLOW_INSTRUCTIONS"/> + <value value="39" name="A7XX_PERF_SP_FS_STAGE_EFU_INSTRUCTIONS"/> + <value value="40" name="A7XX_PERF_SP_FS_STAGE_FULL_ALU_INSTRUCTIONS"/> + <value value="41" name="A7XX_PERF_SP_FS_STAGE_HALF_ALU_INSTRUCTIONS"/> + <value value="42" name="A7XX_PERF_SP_FS_STAGE_BARY_INSTRUCTIONS"/> + <value value="43" name="A7XX_PERF_SP_VS_INSTRUCTIONS"/> + <value value="44" name="A7XX_PERF_SP_FS_INSTRUCTIONS"/> + <value value="45" name="A7XX_PERF_SP_ADDR_LOCK_COUNT"/> + <value value="46" name="A7XX_PERF_SP_UCHE_READ_TRANS"/> + <value value="47" name="A7XX_PERF_SP_UCHE_WRITE_TRANS"/> + <value value="48" name="A7XX_PERF_SP_EXPORT_VPC_TRANS"/> + <value value="49" name="A7XX_PERF_SP_EXPORT_RB_TRANS"/> + <value value="50" name="A7XX_PERF_SP_PIXELS_KILLED"/> + <value value="51" name="A7XX_PERF_SP_ICL1_REQUESTS"/> + <value value="52" name="A7XX_PERF_SP_ICL1_MISSES"/> + <value value="53" name="A7XX_PERF_SP_HS_INSTRUCTIONS"/> + <value value="54" name="A7XX_PERF_SP_DS_INSTRUCTIONS"/> + <value value="55" name="A7XX_PERF_SP_GS_INSTRUCTIONS"/> + <value value="56" name="A7XX_PERF_SP_CS_INSTRUCTIONS"/> + <value value="57" name="A7XX_PERF_SP_GPR_READ"/> + <value value="58" name="A7XX_PERF_SP_GPR_WRITE"/> + <value value="59" name="A7XX_PERF_SP_FS_STAGE_HALF_EFU_INSTRUCTIONS"/> + <value value="60" name="A7XX_PERF_SP_VS_STAGE_HALF_EFU_INSTRUCTIONS"/> + <value value="61" name="A7XX_PERF_SP_LM_BANK_CONFLICTS"/> + <value value="62" name="A7XX_PERF_SP_TEX_CONTROL_WORKING_CYCLES"/> + <value value="63" name="A7XX_PERF_SP_LOAD_CONTROL_WORKING_CYCLES"/> + <value value="64" name="A7XX_PERF_SP_FLOW_CONTROL_WORKING_CYCLES"/> + <value value="65" name="A7XX_PERF_SP_LM_WORKING_CYCLES"/> + <value value="66" name="A7XX_PERF_SP_DISPATCHER_WORKING_CYCLES"/> + <value value="67" name="A7XX_PERF_SP_SEQUENCER_WORKING_CYCLES"/> + <value value="68" name="A7XX_PERF_SP_LOW_EFFICIENCY_STARVED_BY_TP"/> + <value value="69" name="A7XX_PERF_SP_STARVE_CYCLES_HLSQ"/> + <value value="70" name="A7XX_PERF_SP_NON_EXECUTION_LS_CYCLES"/> + <value value="71" name="A7XX_PERF_SP_WORKING_EU"/> + <value value="72" name="A7XX_PERF_SP_ANY_EU_WORKING"/> + <value value="73" name="A7XX_PERF_SP_WORKING_EU_FS_STAGE"/> + <value value="74" name="A7XX_PERF_SP_ANY_EU_WORKING_FS_STAGE"/> + <value value="75" name="A7XX_PERF_SP_WORKING_EU_VS_STAGE"/> + <value value="76" name="A7XX_PERF_SP_ANY_EU_WORKING_VS_STAGE"/> + <value value="77" name="A7XX_PERF_SP_WORKING_EU_CS_STAGE"/> + <value value="78" name="A7XX_PERF_SP_ANY_EU_WORKING_CS_STAGE"/> + <value value="79" name="A7XX_PERF_SP_GPR_READ_PREFETCH"/> + <value value="80" name="A7XX_PERF_SP_GPR_READ_CONFLICT"/> + <value value="81" name="A7XX_PERF_SP_GPR_WRITE_CONFLICT"/> + <value value="82" name="A7XX_PERF_SP_GM_LOAD_LATENCY_CYCLES"/> + <value value="83" name="A7XX_PERF_SP_GM_LOAD_LATENCY_SAMPLES"/> + <value value="84" name="A7XX_PERF_SP_EXECUTABLE_WAVES"/> + <value value="85" name="A7XX_PERF_SP_ICL1_MISS_FETCH_CYCLES"/> + <value value="86" name="A7XX_PERF_SP_WORKING_EU_LPAC"/> + <value value="87" name="A7XX_PERF_SP_BYPASS_BUSY_CYCLES"/> + <value value="88" name="A7XX_PERF_SP_ANY_EU_WORKING_LPAC"/> + <value value="89" name="A7XX_PERF_SP_WAVE_ALU_CYCLES"/> + <value value="90" name="A7XX_PERF_SP_WAVE_EFU_CYCLES"/> + <value value="91" name="A7XX_PERF_SP_WAVE_INT_CYCLES"/> + <value value="92" name="A7XX_PERF_SP_WAVE_CSP_CYCLES"/> + <value value="93" name="A7XX_PERF_SP_EWAVE_CONTEXTS"/> + <value value="94" name="A7XX_PERF_SP_EWAVE_CONTEXT_CYCLES"/> + <value value="95" name="A7XX_PERF_SP_LPAC_BUSY_CYCLES"/> + <value value="96" name="A7XX_PERF_SP_LPAC_INSTRUCTIONS"/> + <value value="97" name="A7XX_PERF_SP_FS_STAGE_1X_WAVES"/> + <value value="98" name="A7XX_PERF_SP_FS_STAGE_2X_WAVES"/> + <value value="99" name="A7XX_PERF_SP_QUADS"/> + <value value="100" name="A7XX_PERF_SP_CS_INVOCATIONS"/> + <value value="101" name="A7XX_PERF_SP_PIXELS"/> + <value value="102" name="A7XX_PERF_SP_LPAC_DRAWCALLS"/> + <value value="103" name="A7XX_PERF_SP_PI_WORKING_CYCLES"/> + <value value="104" name="A7XX_PERF_SP_WAVE_INPUT_CYCLES"/> + <value value="105" name="A7XX_PERF_SP_WAVE_OUTPUT_CYCLES"/> + <value value="106" name="A7XX_PERF_SP_WAVE_HWAVE_WAIT_CYCLES"/> + <value value="107" name="A7XX_PERF_SP_WAVE_HWAVE_SYNC"/> + <value value="108" name="A7XX_PERF_SP_OUTPUT_3D_PIXELS"/> + <value value="109" name="A7XX_PERF_SP_FULL_ALU_MAD_INSTRUCTIONS"/> + <value value="110" name="A7XX_PERF_SP_HALF_ALU_MAD_INSTRUCTIONS"/> + <value value="111" name="A7XX_PERF_SP_FULL_ALU_MUL_INSTRUCTIONS"/> + <value value="112" name="A7XX_PERF_SP_HALF_ALU_MUL_INSTRUCTIONS"/> + <value value="113" name="A7XX_PERF_SP_FULL_ALU_ADD_INSTRUCTIONS"/> + <value value="114" name="A7XX_PERF_SP_HALF_ALU_ADD_INSTRUCTIONS"/> + <value value="115" name="A7XX_PERF_SP_BARY_FP32_INSTRUCTIONS"/> + <value value="116" name="A7XX_PERF_SP_ALU_GPR_READ_CYCLES"/> + <value value="117" name="A7XX_PERF_SP_ALU_DATA_FORWARDING_CYCLES"/> + <value value="118" name="A7XX_PERF_SP_LM_FULL_CYCLES"/> + <value value="119" name="A7XX_PERF_SP_TEXTURE_FETCH_LATENCY_CYCLES"/> + <value value="120" name="A7XX_PERF_SP_TEXTURE_FETCH_LATENCY_SAMPLES"/> + <value value="121" name="A7XX_PERF_SP_FS_STAGE_PI_TEX_INSTRUCTION"/> + <value value="122" name="A7XX_PERF_SP_RAY_QUERY_INSTRUCTIONS"/> + <value value="123" name="A7XX_PERF_SP_RBRT_KICKOFF_FIBERS"/> + <value value="124" name="A7XX_PERF_SP_RBRT_KICKOFF_DQUADS"/> + <value value="125" name="A7XX_PERF_SP_RTU_BUSY_CYCLES"/> + <value value="126" name="A7XX_PERF_SP_RTU_L0_HITS"/> + <value value="127" name="A7XX_PERF_SP_RTU_L0_MISSES"/> + <value value="128" name="A7XX_PERF_SP_RTU_L0_HIT_ON_MISS"/> + <value value="129" name="A7XX_PERF_SP_RTU_STALL_CYCLES_WAVE_QUEUE"/> + <value value="130" name="A7XX_PERF_SP_RTU_STALL_CYCLES_L0_HIT_QUEUE"/> + <value value="131" name="A7XX_PERF_SP_RTU_STALL_CYCLES_L0_MISS_QUEUE"/> + <value value="132" name="A7XX_PERF_SP_RTU_STALL_CYCLES_L0D_IDX_QUEUE"/> + <value value="133" name="A7XX_PERF_SP_RTU_STALL_CYCLES_L0DATA"/> + <value value="134" name="A7XX_PERF_SP_RTU_STALL_CYCLES_REPLACE_CNT"/> + <value value="135" name="A7XX_PERF_SP_RTU_STALL_CYCLES_MRG_CNT"/> + <value value="136" name="A7XX_PERF_SP_RTU_STALL_CYCLES_UCHE"/> + <value value="137" name="A7XX_PERF_SP_RTU_OPERAND_FETCH_STALL_CYCLES_L0"/> + <value value="138" name="A7XX_PERF_SP_RTU_OPERAND_FETCH_STALL_CYCLES_INS_FIFO"/> + <value value="139" name="A7XX_PERF_SP_RTU_BVH_FETCH_LATENCY_CYCLES"/> + <value value="140" name="A7XX_PERF_SP_RTU_BVH_FETCH_LATENCY_SAMPLES"/> + <value value="141" name="A7XX_PERF_SP_STCHE_MISS_INC_VS"/> + <value value="142" name="A7XX_PERF_SP_STCHE_MISS_INC_FS"/> + <value value="143" name="A7XX_PERF_SP_STCHE_MISS_INC_BV"/> + <value value="144" name="A7XX_PERF_SP_STCHE_MISS_INC_LPAC"/> + <value value="145" name="A7XX_PERF_SP_VGPR_ACTIVE_CONTEXTS"/> + <value value="146" name="A7XX_PERF_SP_PGPR_ALLOC_CONTEXTS"/> + <value value="147" name="A7XX_PERF_SP_VGPR_ALLOC_CONTEXTS"/> + <value value="148" name="A7XX_PERF_SP_RTU_RAY_BOX_INTERSECTIONS"/> + <value value="149" name="A7XX_PERF_SP_RTU_RAY_TRIANGLE_INTERSECTIONS"/> + <value value="150" name="A7XX_PERF_SP_SCH_STALL_CYCLES_RTU"/> +</enum> + +<enum name="a7xx_rb_perfcounter_select"> + <value value="0" name="A7XX_PERF_RB_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_RB_STALL_CYCLES_HLSQ"/> + <value value="2" name="A7XX_PERF_RB_STALL_CYCLES_FIFO0_FULL"/> + <value value="3" name="A7XX_PERF_RB_STALL_CYCLES_FIFO1_FULL"/> + <value value="4" name="A7XX_PERF_RB_STALL_CYCLES_FIFO2_FULL"/> + <value value="5" name="A7XX_PERF_RB_STARVE_CYCLES_SP"/> + <value value="6" name="A7XX_PERF_RB_STARVE_CYCLES_LRZ_TILE"/> + <value value="7" name="A7XX_PERF_RB_STARVE_CYCLES_CCU"/> + <value value="8" name="A7XX_PERF_RB_STARVE_CYCLES_Z_PLANE"/> + <value value="9" name="A7XX_PERF_RB_STARVE_CYCLES_BARY_PLANE"/> + <value value="10" name="A7XX_PERF_RB_Z_WORKLOAD"/> + <value value="11" name="A7XX_PERF_RB_HLSQ_ACTIVE"/> + <value value="12" name="A7XX_PERF_RB_Z_READ"/> + <value value="13" name="A7XX_PERF_RB_Z_WRITE"/> + <value value="14" name="A7XX_PERF_RB_C_READ"/> + <value value="15" name="A7XX_PERF_RB_C_WRITE"/> + <value value="16" name="A7XX_PERF_RB_TOTAL_PASS"/> + <value value="17" name="A7XX_PERF_RB_Z_PASS"/> + <value value="18" name="A7XX_PERF_RB_Z_FAIL"/> + <value value="19" name="A7XX_PERF_RB_S_FAIL"/> + <value value="20" name="A7XX_PERF_RB_BLENDED_FXP_COMPONENTS"/> + <value value="21" name="A7XX_PERF_RB_BLENDED_FP16_COMPONENTS"/> + <value value="22" name="A7XX_PERF_RB_PS_INVOCATIONS"/> + <value value="23" name="A7XX_PERF_RB_2D_ALIVE_CYCLES"/> + <value value="24" name="A7XX_PERF_RB_2D_STALL_CYCLES_A2D"/> + <value value="25" name="A7XX_PERF_RB_2D_STARVE_CYCLES_SRC"/> + <value value="26" name="A7XX_PERF_RB_2D_STARVE_CYCLES_SP"/> + <value value="27" name="A7XX_PERF_RB_2D_STARVE_CYCLES_DST"/> + <value value="28" name="A7XX_PERF_RB_2D_VALID_PIXELS"/> + <value value="29" name="A7XX_PERF_RB_3D_PIXELS"/> + <value value="30" name="A7XX_PERF_RB_BLENDER_WORKING_CYCLES"/> + <value value="31" name="A7XX_PERF_RB_ZPROC_WORKING_CYCLES"/> + <value value="32" name="A7XX_PERF_RB_CPROC_WORKING_CYCLES"/> + <value value="33" name="A7XX_PERF_RB_SAMPLER_WORKING_CYCLES"/> + <value value="34" name="A7XX_PERF_RB_STALL_CYCLES_CCU_COLOR_READ"/> + <value value="35" name="A7XX_PERF_RB_STALL_CYCLES_CCU_COLOR_WRITE"/> + <value value="36" name="A7XX_PERF_RB_STALL_CYCLES_CCU_DEPTH_READ"/> + <value value="37" name="A7XX_PERF_RB_STALL_CYCLES_CCU_DEPTH_WRITE"/> + <value value="38" name="A7XX_PERF_RB_STALL_CYCLES_VPC"/> + <value value="39" name="A7XX_PERF_RB_2D_INPUT_TRANS"/> + <value value="40" name="A7XX_PERF_RB_2D_OUTPUT_RB_DST_TRANS"/> + <value value="41" name="A7XX_PERF_RB_2D_OUTPUT_RB_SRC_TRANS"/> + <value value="42" name="A7XX_PERF_RB_BLENDED_FP32_COMPONENTS"/> + <value value="43" name="A7XX_PERF_RB_COLOR_PIX_TILES"/> + <value value="44" name="A7XX_PERF_RB_STALL_CYCLES_CCU"/> + <value value="45" name="A7XX_PERF_RB_EARLY_Z_ARB3_GRANT"/> + <value value="46" name="A7XX_PERF_RB_LATE_Z_ARB3_GRANT"/> + <value value="47" name="A7XX_PERF_RB_EARLY_Z_SKIP_GRANT"/> + <value value="48" name="A7XX_PERF_RB_VRS_1x1_QUADS"/> + <value value="49" name="A7XX_PERF_RB_VRS_2x1_QUADS"/> + <value value="50" name="A7XX_PERF_RB_VRS_1x2_QUADS"/> + <value value="51" name="A7XX_PERF_RB_VRS_2x2_QUADS"/> + <value value="52" name="A7XX_PERF_RB_VRS_4x2_QUADS"/> + <value value="53" name="A7XX_PERF_RB_VRS_4x4_QUADS"/> +</enum> + +<enum name="a7xx_vsc_perfcounter_select"> + <value value="0" name="A7XX_PERF_VSC_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_VSC_WORKING_CYCLES"/> + <value value="2" name="A7XX_PERF_VSC_STALL_CYCLES_UCHE"/> + <value value="3" name="A7XX_PERF_VSC_EOT_NUM"/> + <value value="4" name="A7XX_PERF_VSC_INPUT_TILES"/> +</enum> + +<enum name="a7xx_ccu_perfcounter_select"> + <value value="0" name="A7XX_PERF_CCU_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_CCU_STALL_CYCLES_RB_DEPTH_RETURN"/> + <value value="2" name="A7XX_PERF_CCU_STALL_CYCLES_RB_COLOR_RETURN"/> + <value value="3" name="A7XX_PERF_CCU_DEPTH_BLOCKS"/> + <value value="4" name="A7XX_PERF_CCU_COLOR_BLOCKS"/> + <value value="5" name="A7XX_PERF_CCU_DEPTH_BLOCK_HIT"/> + <value value="6" name="A7XX_PERF_CCU_COLOR_BLOCK_HIT"/> + <value value="7" name="A7XX_PERF_CCU_PARTIAL_BLOCK_READ"/> + <value value="8" name="A7XX_PERF_CCU_GMEM_READ"/> + <value value="9" name="A7XX_PERF_CCU_GMEM_WRITE"/> + <value value="10" name="A7XX_PERF_CCU_2D_RD_REQ"/> + <value value="11" name="A7XX_PERF_CCU_2D_WR_REQ"/> + <value value="12" name="A7XX_PERF_CCU_UBWC_COLOR_BLOCKS_CONCURRENT"/> + <value value="13" name="A7XX_PERF_CCU_UBWC_DEPTH_BLOCKS_CONCURRENT"/> + <value value="14" name="A7XX_PERF_CCU_COLOR_RESOLVE_DROPPED"/> + <value value="15" name="A7XX_PERF_CCU_DEPTH_RESOLVE_DROPPED"/> + <value value="16" name="A7XX_PERF_CCU_COLOR_RENDER_CONCURRENT"/> + <value value="17" name="A7XX_PERF_CCU_DEPTH_RENDER_CONCURRENT"/> + <value value="18" name="A7XX_PERF_CCU_COLOR_RESOLVE_AFTER_RENDER"/> + <value value="19" name="A7XX_PERF_CCU_DEPTH_RESOLVE_AFTER_RENDER"/> + <value value="20" name="A7XX_PERF_CCU_GMEM_EXTRA_DEPTH_READ"/> + <value value="21" name="A7XX_PERF_CCU_GMEM_COLOR_READ_4AA"/> + <value value="22" name="A7XX_PERF_CCU_GMEM_COLOR_READ_4AA_FULL"/> +</enum> + +<enum name="a7xx_lrz_perfcounter_select"> + <value value="0" name="A7XX_PERF_LRZ_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_LRZ_STARVE_CYCLES_RAS"/> + <value value="2" name="A7XX_PERF_LRZ_STALL_CYCLES_RB"/> + <value value="3" name="A7XX_PERF_LRZ_STALL_CYCLES_VSC"/> + <value value="4" name="A7XX_PERF_LRZ_STALL_CYCLES_VPC"/> + <value value="5" name="A7XX_PERF_LRZ_STALL_CYCLES_FLAG_PREFETCH"/> + <value value="6" name="A7XX_PERF_LRZ_STALL_CYCLES_UCHE"/> + <value value="7" name="A7XX_PERF_LRZ_LRZ_READ"/> + <value value="8" name="A7XX_PERF_LRZ_LRZ_WRITE"/> + <value value="9" name="A7XX_PERF_LRZ_READ_LATENCY"/> + <value value="10" name="A7XX_PERF_LRZ_MERGE_CACHE_UPDATING"/> + <value value="11" name="A7XX_PERF_LRZ_PRIM_KILLED_BY_MASKGEN"/> + <value value="12" name="A7XX_PERF_LRZ_PRIM_KILLED_BY_LRZ"/> + <value value="13" name="A7XX_PERF_LRZ_VISIBLE_PRIM_AFTER_LRZ"/> + <value value="14" name="A7XX_PERF_LRZ_FULL_8X8_TILES"/> + <value value="15" name="A7XX_PERF_LRZ_PARTIAL_8X8_TILES"/> + <value value="16" name="A7XX_PERF_LRZ_TILE_KILLED"/> + <value value="17" name="A7XX_PERF_LRZ_TOTAL_PIXEL"/> + <value value="18" name="A7XX_PERF_LRZ_VISIBLE_PIXEL_AFTER_LRZ"/> + <value value="19" name="A7XX_PERF_LRZ_FEEDBACK_ACCEPT"/> + <value value="20" name="A7XX_PERF_LRZ_FEEDBACK_DISCARD"/> + <value value="21" name="A7XX_PERF_LRZ_FEEDBACK_STALL"/> + <value value="22" name="A7XX_PERF_LRZ_STALL_CYCLES_RB_ZPLANE"/> + <value value="23" name="A7XX_PERF_LRZ_STALL_CYCLES_RB_BPLANE"/> + <value value="24" name="A7XX_PERF_LRZ_RAS_MASK_TRANS"/> + <value value="25" name="A7XX_PERF_LRZ_STALL_CYCLES_MVC"/> + <value value="26" name="A7XX_PERF_LRZ_TILE_KILLED_BY_IMAGE_VRS"/> + <value value="27" name="A7XX_PERF_LRZ_TILE_KILLED_BY_Z"/> +</enum> + +<enum name="a7xx_cmp_perfcounter_select"> + <value value="0" name="A7XX_PERF_CMPDECMP_STALL_CYCLES_ARB"/> + <value value="1" name="A7XX_PERF_CMPDECMP_VBIF_LATENCY_CYCLES"/> + <value value="2" name="A7XX_PERF_CMPDECMP_VBIF_LATENCY_SAMPLES"/> + <value value="3" name="A7XX_PERF_CMPDECMP_VBIF_READ_DATA_CCU"/> + <value value="4" name="A7XX_PERF_CMPDECMP_VBIF_WRITE_DATA_CCU"/> + <value value="5" name="A7XX_PERF_CMPDECMP_VBIF_READ_REQUEST"/> + <value value="6" name="A7XX_PERF_CMPDECMP_VBIF_WRITE_REQUEST"/> + <value value="7" name="A7XX_PERF_CMPDECMP_VBIF_READ_DATA"/> + <value value="8" name="A7XX_PERF_CMPDECMP_VBIF_WRITE_DATA"/> + <value value="9" name="A7XX_PERF_CMPDECMP_DEPTH_WRITE_FLAG1_COUNT"/> + <value value="10" name="A7XX_PERF_CMPDECMP_DEPTH_WRITE_FLAG2_COUNT"/> + <value value="11" name="A7XX_PERF_CMPDECMP_DEPTH_WRITE_FLAG3_COUNT"/> + <value value="12" name="A7XX_PERF_CMPDECMP_DEPTH_WRITE_FLAG4_COUNT"/> + <value value="13" name="A7XX_PERF_CMPDECMP_DEPTH_WRITE_FLAG5_COUNT"/> + <value value="14" name="A7XX_PERF_CMPDECMP_DEPTH_WRITE_FLAG6_COUNT"/> + <value value="15" name="A7XX_PERF_CMPDECMP_DEPTH_WRITE_FLAG8_COUNT"/> + <value value="16" name="A7XX_PERF_CMPDECMP_COLOR_WRITE_FLAG1_COUNT"/> + <value value="17" name="A7XX_PERF_CMPDECMP_COLOR_WRITE_FLAG2_COUNT"/> + <value value="18" name="A7XX_PERF_CMPDECMP_COLOR_WRITE_FLAG3_COUNT"/> + <value value="19" name="A7XX_PERF_CMPDECMP_COLOR_WRITE_FLAG4_COUNT"/> + <value value="20" name="A7XX_PERF_CMPDECMP_COLOR_WRITE_FLAG5_COUNT"/> + <value value="21" name="A7XX_PERF_CMPDECMP_COLOR_WRITE_FLAG6_COUNT"/> + <value value="22" name="A7XX_PERF_CMPDECMP_COLOR_WRITE_FLAG8_COUNT"/> + <value value="23" name="A7XX_PERF_CMPDECMP_VBIF_READ_DATA_UCHE_CH0"/> + <value value="24" name="A7XX_PERF_CMPDECMP_VBIF_READ_DATA_UCHE_CH1"/> + <value value="25" name="A7XX_PERF_CMPDECMP_VBIF_WRITE_DATA_UCHE"/> + <value value="26" name="A7XX_PERF_CMPDECMP_DEPTH_WRITE_FLAG0_COUNT"/> + <value value="27" name="A7XX_PERF_CMPDECMP_COLOR_WRITE_FLAG0_COUNT"/> + <value value="28" name="A7XX_PERF_CMPDECMP_COLOR_WRITE_FLAGALPHA_COUNT"/> + <value value="29" name="A7XX_PERF_CMPDECMP_RESOLVE_EVENTS"/> + <value value="30" name="A7XX_PERF_CMPDECMP_CONCURRENT_RESOLVE_EVENTS"/> + <value value="31" name="A7XX_PERF_CMPDECMP_DROPPED_CLEAR_EVENTS"/> + <value value="32" name="A7XX_PERF_CMPDECMP_ST_BLOCKS_CONCURRENT"/> + <value value="33" name="A7XX_PERF_CMPDECMP_LRZ_ST_BLOCKS_CONCURRENT"/> + <value value="34" name="A7XX_PERF_CMPDECMP_DEPTH_READ_FLAG0_COUNT"/> + <value value="35" name="A7XX_PERF_CMPDECMP_DEPTH_READ_FLAG1_COUNT"/> + <value value="36" name="A7XX_PERF_CMPDECMP_DEPTH_READ_FLAG2_COUNT"/> + <value value="37" name="A7XX_PERF_CMPDECMP_DEPTH_READ_FLAG3_COUNT"/> + <value value="38" name="A7XX_PERF_CMPDECMP_DEPTH_READ_FLAG4_COUNT"/> + <value value="39" name="A7XX_PERF_CMPDECMP_DEPTH_READ_FLAG5_COUNT"/> + <value value="40" name="A7XX_PERF_CMPDECMP_DEPTH_READ_FLAG6_COUNT"/> + <value value="41" name="A7XX_PERF_CMPDECMP_DEPTH_READ_FLAG8_COUNT"/> + <value value="42" name="A7XX_PERF_CMPDECMP_COLOR_READ_FLAG0_COUNT"/> + <value value="43" name="A7XX_PERF_CMPDECMP_COLOR_READ_FLAG1_COUNT"/> + <value value="44" name="A7XX_PERF_CMPDECMP_COLOR_READ_FLAG2_COUNT"/> + <value value="45" name="A7XX_PERF_CMPDECMP_COLOR_READ_FLAG3_COUNT"/> + <value value="46" name="A7XX_PERF_CMPDECMP_COLOR_READ_FLAG4_COUNT"/> + <value value="47" name="A7XX_PERF_CMPDECMP_COLOR_READ_FLAG5_COUNT"/> + <value value="48" name="A7XX_PERF_CMPDECMP_COLOR_READ_FLAG6_COUNT"/> + <value value="49" name="A7XX_PERF_CMPDECMP_COLOR_READ_FLAG8_COUNT"/> +</enum> + +<enum name="a7xx_gbif_perfcounter_select"> + <value value="0" name="A7XX_PERF_GBIF_RESERVED_0"/> + <value value="1" name="A7XX_PERF_GBIF_RESERVED_1"/> + <value value="2" name="A7XX_PERF_GBIF_RESERVED_2"/> + <value value="3" name="A7XX_PERF_GBIF_RESERVED_3"/> + <value value="4" name="A7XX_PERF_GBIF_RESERVED_4"/> + <value value="5" name="A7XX_PERF_GBIF_RESERVED_5"/> + <value value="6" name="A7XX_PERF_GBIF_RESERVED_6"/> + <value value="7" name="A7XX_PERF_GBIF_RESERVED_7"/> + <value value="8" name="A7XX_PERF_GBIF_RESERVED_8"/> + <value value="9" name="A7XX_PERF_GBIF_RESERVED_9"/> + <value value="10" name="A7XX_PERF_GBIF_AXI0_READ_REQUESTS_TOTAL"/> + <value value="11" name="A7XX_PERF_GBIF_AXI1_READ_REQUESTS_TOTAL"/> + <value value="12" name="A7XX_PERF_GBIF_RESERVED_12"/> + <value value="13" name="A7XX_PERF_GBIF_RESERVED_13"/> + <value value="14" name="A7XX_PERF_GBIF_RESERVED_14"/> + <value value="15" name="A7XX_PERF_GBIF_RESERVED_15"/> + <value value="16" name="A7XX_PERF_GBIF_RESERVED_16"/> + <value value="17" name="A7XX_PERF_GBIF_RESERVED_17"/> + <value value="18" name="A7XX_PERF_GBIF_RESERVED_18"/> + <value value="19" name="A7XX_PERF_GBIF_RESERVED_19"/> + <value value="20" name="A7XX_PERF_GBIF_RESERVED_20"/> + <value value="21" name="A7XX_PERF_GBIF_RESERVED_21"/> + <value value="22" name="A7XX_PERF_GBIF_AXI0_WRITE_REQUESTS_TOTAL"/> + <value value="23" name="A7XX_PERF_GBIF_AXI1_WRITE_REQUESTS_TOTAL"/> + <value value="24" name="A7XX_PERF_GBIF_RESERVED_24"/> + <value value="25" name="A7XX_PERF_GBIF_RESERVED_25"/> + <value value="26" name="A7XX_PERF_GBIF_RESERVED_26"/> + <value value="27" name="A7XX_PERF_GBIF_RESERVED_27"/> + <value value="28" name="A7XX_PERF_GBIF_RESERVED_28"/> + <value value="29" name="A7XX_PERF_GBIF_RESERVED_29"/> + <value value="30" name="A7XX_PERF_GBIF_RESERVED_30"/> + <value value="31" name="A7XX_PERF_GBIF_RESERVED_31"/> + <value value="32" name="A7XX_PERF_GBIF_RESERVED_32"/> + <value value="33" name="A7XX_PERF_GBIF_RESERVED_33"/> + <value value="34" name="A7XX_PERF_GBIF_AXI0_READ_DATA_BEATS_TOTAL"/> + <value value="35" name="A7XX_PERF_GBIF_AXI1_READ_DATA_BEATS_TOTAL"/> + <value value="36" name="A7XX_PERF_GBIF_RESERVED_36"/> + <value value="37" name="A7XX_PERF_GBIF_RESERVED_37"/> + <value value="38" name="A7XX_PERF_GBIF_RESERVED_38"/> + <value value="39" name="A7XX_PERF_GBIF_RESERVED_39"/> + <value value="40" name="A7XX_PERF_GBIF_RESERVED_40"/> + <value value="41" name="A7XX_PERF_GBIF_RESERVED_41"/> + <value value="42" name="A7XX_PERF_GBIF_RESERVED_42"/> + <value value="43" name="A7XX_PERF_GBIF_RESERVED_43"/> + <value value="44" name="A7XX_PERF_GBIF_RESERVED_44"/> + <value value="45" name="A7XX_PERF_GBIF_RESERVED_45"/> + <value value="46" name="A7XX_PERF_GBIF_AXI0_WRITE_DATA_BEATS_TOTAL"/> + <value value="47" name="A7XX_PERF_GBIF_AXI1_WRITE_DATA_BEATS_TOTAL"/> + <value value="48" name="A7XX_PERF_GBIF_RESERVED_48"/> + <value value="49" name="A7XX_PERF_GBIF_RESERVED_49"/> + <value value="50" name="A7XX_PERF_GBIF_RESERVED_50"/> + <value value="51" name="A7XX_PERF_GBIF_RESERVED_51"/> + <value value="52" name="A7XX_PERF_GBIF_RESERVED_52"/> + <value value="53" name="A7XX_PERF_GBIF_RESERVED_53"/> + <value value="54" name="A7XX_PERF_GBIF_RESERVED_54"/> + <value value="55" name="A7XX_PERF_GBIF_RESERVED_55"/> + <value value="56" name="A7XX_PERF_GBIF_RESERVED_56"/> + <value value="57" name="A7XX_PERF_GBIF_RESERVED_57"/> + <value value="58" name="A7XX_PERF_GBIF_RESERVED_58"/> + <value value="59" name="A7XX_PERF_GBIF_RESERVED_59"/> + <value value="60" name="A7XX_PERF_GBIF_RESERVED_60"/> + <value value="61" name="A7XX_PERF_GBIF_RESERVED_61"/> + <value value="62" name="A7XX_PERF_GBIF_RESERVED_62"/> + <value value="63" name="A7XX_PERF_GBIF_RESERVED_63"/> + <value value="64" name="A7XX_PERF_GBIF_RESERVED_64"/> + <value value="65" name="A7XX_PERF_GBIF_RESERVED_65"/> + <value value="66" name="A7XX_PERF_GBIF_RESERVED_66"/> + <value value="67" name="A7XX_PERF_GBIF_RESERVED_67"/> + <value value="68" name="A7XX_PERF_GBIF_CYCLES_CH0_HELD_OFF_RD_ALL"/> + <value value="69" name="A7XX_PERF_GBIF_CYCLES_CH1_HELD_OFF_RD_ALL"/> + <value value="70" name="A7XX_PERF_GBIF_CYCLES_CH0_HELD_OFF_WR_ALL"/> + <value value="71" name="A7XX_PERF_GBIF_CYCLES_CH1_HELD_OFF_WR_ALL"/> + <value value="72" name="A7XX_PERF_GBIF_AXI_CH0_REQUEST_HELD_OFF"/> + <value value="73" name="A7XX_PERF_GBIF_AXI_CH1_REQUEST_HELD_OFF"/> + <value value="74" name="A7XX_PERF_GBIF_AXI_REQUEST_HELD_OFF"/> + <value value="75" name="A7XX_PERF_GBIF_AXI_CH0_WRITE_DATA_HELD_OFF"/> + <value value="76" name="A7XX_PERF_GBIF_AXI_CH1_WRITE_DATA_HELD_OFF"/> + <value value="77" name="A7XX_PERF_GBIF_AXI_ALL_WRITE_DATA_HELD_OFF"/> + <value value="78" name="A7XX_PERF_GBIF_AXI_ALL_READ_BEATS"/> + <value value="79" name="A7XX_PERF_GBIF_AXI_ALL_WRITE_BEATS"/> + <value value="80" name="A7XX_PERF_GBIF_AXI_ALL_BEATS"/> +</enum> + +<enum name="a7xx_ufc_perfcounter_select"> + <value value="0" name="A7XX_PERF_UFC_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_UFC_READ_DATA_VBIF"/> + <value value="2" name="A7XX_PERF_UFC_WRITE_DATA_VBIF"/> + <value value="3" name="A7XX_PERF_UFC_READ_REQUEST_VBIF"/> + <value value="4" name="A7XX_PERF_UFC_WRITE_REQUEST_VBIF"/> + <value value="5" name="A7XX_PERF_UFC_LRZ_FILTER_HIT"/> + <value value="6" name="A7XX_PERF_UFC_LRZ_FILTER_MISS"/> + <value value="7" name="A7XX_PERF_UFC_CRE_FILTER_HIT"/> + <value value="8" name="A7XX_PERF_UFC_CRE_FILTER_MISS"/> + <value value="9" name="A7XX_PERF_UFC_SP_FILTER_HIT"/> + <value value="10" name="A7XX_PERF_UFC_SP_FILTER_MISS"/> + <value value="11" name="A7XX_PERF_UFC_SP_REQUESTS"/> + <value value="12" name="A7XX_PERF_UFC_TP_FILTER_HIT"/> + <value value="13" name="A7XX_PERF_UFC_TP_FILTER_MISS"/> + <value value="14" name="A7XX_PERF_UFC_TP_REQUESTS"/> + <value value="15" name="A7XX_PERF_UFC_MAIN_HIT_LRZ_PREFETCH"/> + <value value="16" name="A7XX_PERF_UFC_MAIN_HIT_CRE_PREFETCH"/> + <value value="17" name="A7XX_PERF_UFC_MAIN_HIT_SP_PREFETCH"/> + <value value="18" name="A7XX_PERF_UFC_MAIN_HIT_TP_PREFETCH"/> + <value value="19" name="A7XX_PERF_UFC_MAIN_HIT_UBWC_READ"/> + <value value="20" name="A7XX_PERF_UFC_MAIN_HIT_UBWC_WRITE"/> + <value value="21" name="A7XX_PERF_UFC_MAIN_MISS_LRZ_PREFETCH"/> + <value value="22" name="A7XX_PERF_UFC_MAIN_MISS_CRE_PREFETCH"/> + <value value="23" name="A7XX_PERF_UFC_MAIN_MISS_SP_PREFETCH"/> + <value value="24" name="A7XX_PERF_UFC_MAIN_MISS_TP_PREFETCH"/> + <value value="25" name="A7XX_PERF_UFC_MAIN_MISS_UBWC_READ"/> + <value value="26" name="A7XX_PERF_UFC_MAIN_MISS_UBWC_WRITE"/> + <value value="27" name="A7XX_PERF_UFC_UBWC_READ_UFC_TRANS"/> + <value value="28" name="A7XX_PERF_UFC_UBWC_WRITE_UFC_TRANS"/> + <value value="29" name="A7XX_PERF_UFC_STALL_CYCLES_GBIF_CMD"/> + <value value="30" name="A7XX_PERF_UFC_STALL_CYCLES_GBIF_RDATA"/> + <value value="31" name="A7XX_PERF_UFC_STALL_CYCLES_GBIF_WDATA"/> + <value value="32" name="A7XX_PERF_UFC_STALL_CYCLES_UBWC_WR_FLAG"/> + <value value="33" name="A7XX_PERF_UFC_STALL_CYCLES_UBWC_FLAG_RTN"/> + <value value="34" name="A7XX_PERF_UFC_STALL_CYCLES_UBWC_EVENT"/> + <value value="35" name="A7XX_PERF_UFC_LRZ_PREFETCH_STALLED_CYCLES"/> + <value value="36" name="A7XX_PERF_UFC_CRE_PREFETCH_STALLED_CYCLES"/> + <value value="37" name="A7XX_PERF_UFC_SPTP_PREFETCH_STALLED_CYCLES"/> + <value value="38" name="A7XX_PERF_UFC_UBWC_RD_STALLED_CYCLES"/> + <value value="39" name="A7XX_PERF_UFC_UBWC_WR_STALLED_CYCLES"/> + <value value="40" name="A7XX_PERF_UFC_PREFETCH_STALLED_CYCLES"/> + <value value="41" name="A7XX_PERF_UFC_EVICTION_STALLED_CYCLES"/> + <value value="42" name="A7XX_PERF_UFC_LOCK_STALLED_CYCLES"/> + <value value="43" name="A7XX_PERF_UFC_MISS_LATENCY_CYCLES"/> + <value value="44" name="A7XX_PERF_UFC_MISS_LATENCY_SAMPLES"/> + <value value="45" name="A7XX_PERF_UFC_UBWC_REQ_STALLED_CYCLES"/> + <value value="46" name="A7XX_PERF_UFC_TP_HINT_TAG_MISS"/> + <value value="47" name="A7XX_PERF_UFC_TP_HINT_TAG_HIT_RDY"/> + <value value="48" name="A7XX_PERF_UFC_TP_HINT_TAG_HIT_NRDY"/> + <value value="49" name="A7XX_PERF_UFC_TP_HINT_IS_FCLEAR"/> + <value value="50" name="A7XX_PERF_UFC_TP_HINT_IS_ALPHA0"/> + <value value="51" name="A7XX_PERF_UFC_SP_L1_FILTER_HIT"/> + <value value="52" name="A7XX_PERF_UFC_SP_L1_FILTER_MISS"/> + <value value="53" name="A7XX_PERF_UFC_SP_L1_FILTER_REQUESTS"/> + <value value="54" name="A7XX_PERF_UFC_TP_L1_TAG_HIT_RDY"/> + <value value="55" name="A7XX_PERF_UFC_TP_L1_TAG_HIT_NRDY"/> + <value value="56" name="A7XX_PERF_UFC_TP_L1_TAG_MISS"/> + <value value="57" name="A7XX_PERF_UFC_TP_L1_FILTER_REQUESTS"/> +</enum> + <domain name="A6XX" width="32" prefix="variant" varset="chip"> <bitset name="A6XX_RBBM_INT_0_MASK" inline="no" varset="chip"> <bitfield name="RBBM_GPU_IDLE" pos="0" type="boolean"/> @@ -1584,7 +2605,7 @@ to upconvert to 32b float internally? <reg32 offset="0x050e" name="RBBM_PERFCTR_SRAM_INIT_CMD"/> <reg32 offset="0x050f" name="RBBM_PERFCTR_SRAM_INIT_STATUS"/> <reg32 offset="0x0533" name="RBBM_ISDB_CNT"/> - <reg32 offset="0x0534" name="RBBM_NC_MODE_CNTL" variants="A7XX-"/> + <reg32 offset="0x0534" name="RBBM_NC_MODE_CNTL"/> <reg32 offset="0x0535" name="RBBM_SNAPSHOT_STATUS" variants="A7XX-"/> <!--- @@ -2184,13 +3205,28 @@ to upconvert to 32b float internally? <value value="3" name="BUFFERS_IN_SYSMEM"/> </enum> + <enum name="a6xx_lrz_feedback_mask"> + <value value="0x0" name="LRZ_FEEDBACK_NONE"/> + <value value="0x1" name="LRZ_FEEDBACK_EARLY_Z"/> + <value value="0x2" name="LRZ_FEEDBACK_EARLY_LRZ_LATE_Z"/> + <!-- We don't have a flag type and this flags combination is often used --> + <value value="0x3" name="LRZ_FEEDBACK_EARLY_Z_OR_EARLY_LRZ_LATE_Z"/> + <value value="0x4" name="LRZ_FEEDBACK_LATE_Z"/> + </enum> + <reg32 offset="0x80a1" name="GRAS_BIN_CONTROL" usage="rp_blit"> <bitfield name="BINW" low="0" high="5" shr="5" type="uint"/> <bitfield name="BINH" low="8" high="14" shr="4" type="uint"/> <bitfield name="RENDER_MODE" low="18" high="20" type="a6xx_render_mode"/> + <doc>Disable LRZ feedback writes</doc> <bitfield name="FORCE_LRZ_WRITE_DIS" pos="21" type="boolean"/> <bitfield name="BUFFERS_LOCATION" low="22" high="23" type="a6xx_buffers_location" variants="A6XX"/> - <bitfield name="LRZ_FEEDBACK_ZMODE_MASK" low="24" high="26"/> + <doc> + Allows draws that don't have GRAS_LRZ_CNTL.LRZ_WRITE but have + GRAS_LRZ_CNTL.ENABLE to contribute to LRZ during RENDERING pass. + In sysmem mode GRAS_LRZ_CNTL.LRZ_WRITE is not considered. + </doc> + <bitfield name="LRZ_FEEDBACK_ZMODE_MASK" low="24" high="26" type="a6xx_lrz_feedback_mask"/> <bitfield name="UNK27" pos="27"/> </reg32> @@ -2270,7 +3306,7 @@ to upconvert to 32b float internally? - 0.0 if GREATER - 1.0 if LESS </doc> - <bitfield name="FC_ENABLE" pos="3" type="boolean"/> + <bitfield name="FC_ENABLE" pos="3" type="boolean" variants="A6XX"/> <!-- set when depth-test + depth-write enabled --> <bitfield name="Z_TEST_ENABLE" pos="4" type="boolean"/> <bitfield name="Z_BOUNDS_ENABLE" pos="5" type="boolean"/> @@ -2284,7 +3320,7 @@ to upconvert to 32b float internally? Disable LRZ based on previous direction and the current one. If DIR_WRITE is not enabled - there is no write to direction buffer. </doc> - <bitfield name="DISABLE_ON_WRONG_DIR" pos="9" type="boolean"/> + <bitfield name="DISABLE_ON_WRONG_DIR" pos="9" type="boolean" variants="A6XX"/> <bitfield name="Z_FUNC" low="11" high="13" type="adreno_compare_func" variants="A7XX-"/> </reg32> @@ -2357,7 +3393,10 @@ to upconvert to 32b float internally? <bitfield name="BASE_MIP_LEVEL" low="28" high="31" type="uint"/> </reg32> - <reg32 offset="0x810b" name="GRAS_UNKNOWN_810B" variants="A7XX-" usage="cmd"/> + <reg32 offset="0x810b" name="GRAS_LRZ_CNTL2" variants="A7XX-" usage="rp_blit"> + <bitfield name="DISABLE_ON_WRONG_DIR" pos="0" type="boolean"/> + <bitfield name="FC_ENABLE" pos="1" type="boolean"/> + </reg32> <!-- 0x810c-0x810f invalid --> @@ -2366,7 +3405,10 @@ to upconvert to 32b float internally? <!-- A bit tentative but it's a color and it is followed by LRZ_CLEAR --> <reg32 offset="0x8111" name="GRAS_LRZ_CLEAR_DEPTH_F32" type="float" variants="A7XX-"/> - <reg32 offset="0x8113" name="GRAS_LRZ_DEPTH_BUFFER_INFO" variants="A7XX-" usage="rp_blit"/> + <reg32 offset="0x8113" name="GRAS_LRZ_DEPTH_BUFFER_INFO" variants="A7XX-" usage="rp_blit"> + <bitfield name="DEPTH_FORMAT" low="0" high="2" type="a6xx_depth_format"/> + <bitfield name="UNK3" pos="3"/> + </reg32> <!-- Always written together and always equal 09510840 00000a62 --> <reg32 offset="0x8120" name="GRAS_UNKNOWN_8120" variants="A7XX-" usage="cmd"/> @@ -2440,7 +3482,7 @@ to upconvert to 32b float internally? <bitfield name="RENDER_MODE" low="18" high="20" type="a6xx_render_mode"/> <bitfield name="FORCE_LRZ_WRITE_DIS" pos="21" type="boolean"/> <bitfield name="BUFFERS_LOCATION" low="22" high="23" type="a6xx_buffers_location"/> - <bitfield name="LRZ_FEEDBACK_ZMODE_MASK" low="24" high="26"/> + <bitfield name="LRZ_FEEDBACK_ZMODE_MASK" low="24" high="26" type="a6xx_lrz_feedback_mask"/> </reg32> <reg32 offset="0x8800" name="RB_BIN_CONTROL" variants="A7XX-" usage="rp_blit"> @@ -2448,7 +3490,7 @@ to upconvert to 32b float internally? <bitfield name="BINH" low="8" high="14" shr="4" type="uint"/> <bitfield name="RENDER_MODE" low="18" high="20" type="a6xx_render_mode"/> <bitfield name="FORCE_LRZ_WRITE_DIS" pos="21" type="boolean"/> - <bitfield name="LRZ_FEEDBACK_ZMODE_MASK" low="24" high="26"/> + <bitfield name="LRZ_FEEDBACK_ZMODE_MASK" low="24" high="26" type="a6xx_lrz_feedback_mask"/> </reg32> <reg32 offset="0x8801" name="RB_RENDER_CNTL" variants="A6XX" usage="rp_blit"> @@ -2605,6 +3647,7 @@ to upconvert to 32b float internally? <bitfield name="UNK10" pos="10"/> <bitfield name="LOSSLESSCOMPEN" pos="11" type="boolean"/> <bitfield name="COLOR_SWAP" low="13" high="14" type="a3xx_color_swap"/> + <bitfield name="MUTABLEEN" pos="16" type="boolean" variants="A7XX-"/> </reg32> <!-- at least in gmem, things seem to be aligned to pitch of 64.. @@ -2770,6 +3813,7 @@ to upconvert to 32b float internally? <bitfield name="COLOR_SWAP" low="5" high="6" type="a3xx_color_swap"/> <bitfield name="COLOR_FORMAT" low="7" high="14" type="a6xx_format"/> <bitfield name="UNK15" pos="15" type="boolean"/> + <bitfield name="MUTABLEEN" pos="16" type="boolean" variants="A7XX-"/> </reg32> <reg64 offset="0x88d8" name="RB_BLIT_DST" type="waddress" align="64" usage="rp_blit"/> <reg32 offset="0x88da" name="RB_BLIT_DST_PITCH" low="0" high="15" shr="6" type="uint" usage="rp_blit"/> @@ -2886,13 +3930,12 @@ to upconvert to 32b float internally? <reg32 offset="0x8c00" name="RB_2D_BLIT_CNTL" type="a6xx_2d_blit_cntl" usage="rp_blit"/> <reg32 offset="0x8c01" name="RB_2D_UNKNOWN_8C01" low="0" high="31" usage="rp_blit"/> - <bitset name="a6xx_2d_surf_info" inline="yes"> + <bitset name="a6xx_2d_src_surf_info" inline="yes"> <bitfield name="COLOR_FORMAT" low="0" high="7" type="a6xx_format"/> <bitfield name="TILE_MODE" low="8" high="9" type="a6xx_tile_mode"/> <bitfield name="COLOR_SWAP" low="10" high="11" type="a3xx_color_swap"/> <bitfield name="FLAGS" pos="12" type="boolean"/> <bitfield name="SRGB" pos="13" type="boolean"/> - <!-- the rest is only for src --> <bitfield name="SAMPLES" low="14" high="15" type="a3xx_msaa_samples"/> <bitfield name="FILTER" pos="16" type="boolean"/> <bitfield name="UNK17" pos="17" type="boolean"/> @@ -2903,11 +3946,21 @@ to upconvert to 32b float internally? <bitfield name="UNK22" pos="22" type="boolean"/> <bitfield name="UNK23" low="23" high="26"/> <bitfield name="UNK28" pos="28" type="boolean"/> + <bitfield name="MUTABLEEN" pos="29" type="boolean" variants="A7XX-"/> + </bitset> + + <bitset name="a6xx_2d_dst_surf_info" inline="yes"> + <bitfield name="COLOR_FORMAT" low="0" high="7" type="a6xx_format"/> + <bitfield name="TILE_MODE" low="8" high="9" type="a6xx_tile_mode"/> + <bitfield name="COLOR_SWAP" low="10" high="11" type="a3xx_color_swap"/> + <bitfield name="FLAGS" pos="12" type="boolean"/> + <bitfield name="SRGB" pos="13" type="boolean"/> + <bitfield name="SAMPLES" low="14" high="15" type="a3xx_msaa_samples"/> + <bitfield name="MUTABLEEN" pos="17" type="boolean" variants="A7XX-"/> </bitset> <!-- 0x8c02-0x8c16 invalid --> - <!-- TODO: RB_2D_DST_INFO has 17 valid bits (doesn't match a6xx_2d_surf_info) --> - <reg32 offset="0x8c17" name="RB_2D_DST_INFO" type="a6xx_2d_surf_info" usage="rp_blit"/> + <reg32 offset="0x8c17" name="RB_2D_DST_INFO" type="a6xx_2d_dst_surf_info" usage="rp_blit"/> <reg64 offset="0x8c18" name="RB_2D_DST" type="waddress" align="64" usage="rp_blit"/> <reg32 offset="0x8c1a" name="RB_2D_DST_PITCH" low="0" high="15" shr="6" type="uint" usage="rp_blit"/> <!-- this is a guess but seems likely (for NV12/IYUV): --> @@ -2927,7 +3980,10 @@ to upconvert to 32b float internally? <reg32 offset="0x8c2d" name="RB_2D_SRC_SOLID_C1" usage="rp_blit"/> <reg32 offset="0x8c2e" name="RB_2D_SRC_SOLID_C2" usage="rp_blit"/> <reg32 offset="0x8c2f" name="RB_2D_SRC_SOLID_C3" usage="rp_blit"/> - <!-- 0x8c34-0x8dff invalid --> + + <reg32 offset="0x8c34" name="RB_UNKNOWN_8C34" variants="A7XX-" usage="cmd"/> + + <!-- 0x8c35-0x8dff invalid --> <!-- always 0x1 ? either doesn't exist for a650 or write-only: --> <reg32 offset="0x8e01" name="RB_UNKNOWN_8E01" usage="cmd"/> @@ -4275,7 +5331,7 @@ to upconvert to 32b float internally? badly named or the functionality moved in a6xx. But downstream kernel calls this "a6xx_sp_ps_tp_2d_cluster" --> - <reg32 offset="0xb4c0" name="SP_PS_2D_SRC_INFO" type="a6xx_2d_surf_info" variants="A6XX" usage="rp_blit"/> + <reg32 offset="0xb4c0" name="SP_PS_2D_SRC_INFO" type="a6xx_2d_src_surf_info" variants="A6XX" usage="rp_blit"/> <reg32 offset="0xb4c1" name="SP_PS_2D_SRC_SIZE" variants="A6XX" usage="rp_blit"> <bitfield name="WIDTH" low="0" high="14" type="uint"/> <bitfield name="HEIGHT" low="15" high="29" type="uint"/> @@ -4286,7 +5342,7 @@ to upconvert to 32b float internally? <bitfield name="PITCH" low="9" high="23" shr="6" type="uint"/> </reg32> - <reg32 offset="0xb2c0" name="SP_PS_2D_SRC_INFO" type="a6xx_2d_surf_info" variants="A7XX-" usage="rp_blit"/> + <reg32 offset="0xb2c0" name="SP_PS_2D_SRC_INFO" type="a6xx_2d_src_surf_info" variants="A7XX-" usage="rp_blit"/> <reg32 offset="0xb2c1" name="SP_PS_2D_SRC_SIZE" variants="A7XX"> <bitfield name="WIDTH" low="0" high="14" type="uint"/> <bitfield name="HEIGHT" low="15" high="29" type="uint"/> @@ -4329,7 +5385,12 @@ to upconvert to 32b float internally? <!-- always 0x100000 or 0x1000000? --> <reg32 offset="0xb600" name="TPL1_DBG_ECO_CNTL" low="0" high="25" usage="cmd"/> <reg32 offset="0xb601" name="TPL1_ADDR_MODE_CNTL" type="a5xx_address_mode"/> - <reg32 offset="0xb602" name="TPL1_DBG_ECO_CNTL1" usage="cmd"/> + <reg32 offset="0xb602" name="TPL1_DBG_ECO_CNTL1" usage="cmd"> + <!-- Affects UBWC in some way, if BLIT_OP_SCALE is done with this bit set + and if other blit is done without it - UBWC image may be copied incorrectly. + --> + <bitfield name="TP_UBWC_FLAG_HINT" pos="18" type="boolean"/> + </reg32> <reg32 offset="0xb604" name="TPL1_NC_MODE_CNTL"> <bitfield name="MODE" pos="0" type="boolean"/> <bitfield name="LOWER_BIT" low="1" high="2" type="uint"/> @@ -4351,7 +5412,8 @@ to upconvert to 32b float internally? <reg32 offset="0xb60b" name="TPL1_BICUBIC_WEIGHTS_TABLE_3" low="0" high="29" variants="A7XX" usage="cmd"/> <reg32 offset="0xb60c" name="TPL1_BICUBIC_WEIGHTS_TABLE_4" low="0" high="29" variants="A7XX" usage="cmd"/> - <array offset="0xb610" name="TPL1_PERFCTR_TP_SEL" stride="1" length="12"/> + <array offset="0xb610" name="TPL1_PERFCTR_TP_SEL" stride="1" length="12" variants="A6XX"/> + <array offset="0xb610" name="TPL1_PERFCTR_TP_SEL" stride="1" length="18" variants="A7XX"/> <!-- TODO: 4 more perfcntr sel at 0xb620 ? --> @@ -4582,15 +5644,15 @@ to upconvert to 32b float internally? <bitfield name="UNK6" pos="6" type="boolean"/> </reg32> - <reg32 offset="0xbb00" name="HLSQ_DRAW_CMD"> + <reg32 offset="0xbb00" name="HLSQ_DRAW_CMD" variants="A6XX"> <bitfield name="STATE_ID" low="0" high="7"/> </reg32> - <reg32 offset="0xbb01" name="HLSQ_DISPATCH_CMD"> + <reg32 offset="0xbb01" name="HLSQ_DISPATCH_CMD" variants="A6XX"> <bitfield name="STATE_ID" low="0" high="7"/> </reg32> - <reg32 offset="0xbb02" name="HLSQ_EVENT_CMD"> + <reg32 offset="0xbb02" name="HLSQ_EVENT_CMD" variants="A6XX"> <!-- I think only the low bit is actually used? --> <bitfield name="STATE_ID" low="16" high="23"/> <bitfield name="EVENT" low="0" high="6" type="vgt_event_type"/> @@ -4623,6 +5685,19 @@ to upconvert to 32b float internally? <bitfield name="GFX_BINDLESS" low="14" high="18" type="hex"/> </reg32> + <reg32 offset="0xab1c" name="HLSQ_DRAW_CMD" variants="A7XX-"> + <bitfield name="STATE_ID" low="0" high="7"/> + </reg32> + + <reg32 offset="0xab1d" name="HLSQ_DISPATCH_CMD" variants="A7XX-"> + <bitfield name="STATE_ID" low="0" high="7"/> + </reg32> + + <reg32 offset="0xab1e" name="HLSQ_EVENT_CMD" variants="A7XX-"> + <bitfield name="STATE_ID" low="16" high="23"/> + <bitfield name="EVENT" low="0" high="6" type="vgt_event_type"/> + </reg32> + <reg32 offset="0xab1f" name="HLSQ_INVALIDATE_CMD" variants="A7XX-" usage="cmd"> <doc> This register clears pending loads queued up by @@ -4791,7 +5866,7 @@ to upconvert to 32b float internally? <reg32 offset="3" name="3"/> </domain> -<domain name="A6XX_TEX_CONST" width="32"> +<domain name="A6XX_TEX_CONST" width="32" varset="chip"> <doc>Texture constant dwords</doc> <enum name="a6xx_tex_swiz"> <!-- same as a4xx? --> <value name="A6XX_TEX_X" value="0"/> @@ -4831,6 +5906,7 @@ to upconvert to 32b float internally? <reg32 offset="1" name="1"> <bitfield name="WIDTH" low="0" high="14" type="uint"/> <bitfield name="HEIGHT" low="15" high="29" type="uint"/> + <bitfield name="MUTABLEEN" pos="31" type="boolean" variants="A7XX-"/> </reg32> <reg32 offset="2" name="2"> <!-- diff --git a/drivers/gpu/drm/msm/registers/display/hdmi.xml b/drivers/gpu/drm/msm/registers/display/hdmi.xml index 6c81581016c7..1cf1b14fbd91 100644 --- a/drivers/gpu/drm/msm/registers/display/hdmi.xml +++ b/drivers/gpu/drm/msm/registers/display/hdmi.xml @@ -1012,4 +1012,93 @@ xsi:schemaLocation="https://gitlab.freedesktop.org/freedreno/ rules-fd.xsd"> <reg32 offset="0x00110" name="TX_ALOG_INTF_OBSV"/> </domain> +<domain name="HDMI_8998_PHY" width="32"> + <reg32 offset="0x00000" name="CFG"/> + <reg32 offset="0x00004" name="PD_CTL"/> + <reg32 offset="0x00010" name="MODE"/> + <reg32 offset="0x0005C" name="CLOCK"/> + <reg32 offset="0x00068" name="CMN_CTRL"/> + <reg32 offset="0x000B4" name="STATUS"/> +</domain> + +<domain name="HDMI_8998_PHY_QSERDES_COM" width="32"> + <reg32 offset="0x0000" name="ATB_SEL1"/> + <reg32 offset="0x0004" name="ATB_SEL2"/> + <reg32 offset="0x0008" name="FREQ_UPDATE"/> + <reg32 offset="0x000C" name="BG_TIMER"/> + <reg32 offset="0x0010" name="SSC_EN_CENTER"/> + <reg32 offset="0x0014" name="SSC_ADJ_PER1"/> + <reg32 offset="0x0018" name="SSC_ADJ_PER2"/> + <reg32 offset="0x001C" name="SSC_PER1"/> + <reg32 offset="0x0020" name="SSC_PER2"/> + <reg32 offset="0x0024" name="SSC_STEP_SIZE1"/> + <reg32 offset="0x0028" name="SSC_STEP_SIZE2"/> + <reg32 offset="0x002C" name="POST_DIV"/> + <reg32 offset="0x0030" name="POST_DIV_MUX"/> + <reg32 offset="0x0034" name="BIAS_EN_CLKBUFLR_EN"/> + <reg32 offset="0x0038" name="CLK_ENABLE1"/> + <reg32 offset="0x003C" name="SYS_CLK_CTRL"/> + <reg32 offset="0x0040" name="SYSCLK_BUF_ENABLE"/> + <reg32 offset="0x0044" name="PLL_EN"/> + <reg32 offset="0x0048" name="PLL_IVCO"/> + <reg32 offset="0x004C" name="CMN_IETRIM"/> + <reg32 offset="0x0050" name="CMN_IPTRIM"/> + <reg32 offset="0x0060" name="CP_CTRL_MODE0"/> + <reg32 offset="0x0064" name="CP_CTRL_MODE1"/> + <reg32 offset="0x0068" name="PLL_RCTRL_MODE0"/> + <reg32 offset="0x006C" name="PLL_RCTRL_MODE1"/> + <reg32 offset="0x0070" name="PLL_CCTRL_MODE0"/> + <reg32 offset="0x0074" name="PLL_CCTRL_MODE1"/> + <reg32 offset="0x0078" name="PLL_CNTRL"/> + <reg32 offset="0x007C" name="BIAS_EN_CTRL_BY_PSM"/> + <reg32 offset="0x0080" name="SYSCLK_EN_SEL"/> + <reg32 offset="0x0084" name="CML_SYSCLK_SEL"/> + <reg32 offset="0x0088" name="RESETSM_CNTRL"/> + <reg32 offset="0x008C" name="RESETSM_CNTRL2"/> + <reg32 offset="0x0090" name="LOCK_CMP_EN"/> + <reg32 offset="0x0094" name="LOCK_CMP_CFG"/> + <reg32 offset="0x0098" name="LOCK_CMP1_MODE0"/> + <reg32 offset="0x009C" name="LOCK_CMP2_MODE0"/> + <reg32 offset="0x00A0" name="LOCK_CMP3_MODE0"/> + <reg32 offset="0x00B0" name="DEC_START_MODE0"/> + <reg32 offset="0x00B4" name="DEC_START_MODE1"/> + <reg32 offset="0x00B8" name="DIV_FRAC_START1_MODE0"/> + <reg32 offset="0x00BC" name="DIV_FRAC_START2_MODE0"/> + <reg32 offset="0x00C0" name="DIV_FRAC_START3_MODE0"/> + <reg32 offset="0x00C4" name="DIV_FRAC_START1_MODE1"/> + <reg32 offset="0x00C8" name="DIV_FRAC_START2_MODE1"/> + <reg32 offset="0x00CC" name="DIV_FRAC_START3_MODE1"/> + <reg32 offset="0x00D0" name="INTEGLOOP_INITVAL"/> + <reg32 offset="0x00D4" name="INTEGLOOP_EN"/> + <reg32 offset="0x00D8" name="INTEGLOOP_GAIN0_MODE0"/> + <reg32 offset="0x00DC" name="INTEGLOOP_GAIN1_MODE0"/> + <reg32 offset="0x00E0" name="INTEGLOOP_GAIN0_MODE1"/> + <reg32 offset="0x00E4" name="INTEGLOOP_GAIN1_MODE1"/> + <reg32 offset="0x00E8" name="VCOCAL_DEADMAN_CTRL"/> + <reg32 offset="0x00EC" name="VCO_TUNE_CTRL"/> + <reg32 offset="0x00F0" name="VCO_TUNE_MAP"/> + <reg32 offset="0x0124" name="CMN_STATUS"/> + <reg32 offset="0x0128" name="RESET_SM_STATUS"/> + <reg32 offset="0x0138" name="CLK_SEL"/> + <reg32 offset="0x013C" name="HSCLK_SEL"/> + <reg32 offset="0x0148" name="CORECLK_DIV_MODE0"/> + <reg32 offset="0x0150" name="SW_RESET"/> + <reg32 offset="0x0154" name="CORE_CLK_EN"/> + <reg32 offset="0x0158" name="C_READY_STATUS"/> + <reg32 offset="0x015C" name="CMN_CONFIG"/> + <reg32 offset="0x0164" name="SVS_MODE_CLK_SEL"/> +</domain> + +<domain name="HDMI_8998_PHY_TXn" width="32"> + <reg32 offset="0x0000" name="EMP_POST1_LVL"/> + <reg32 offset="0x0008" name="INTERFACE_SELECT_TX_BAND"/> + <reg32 offset="0x000C" name="CLKBUF_TERM_ENABLE"/> + <reg32 offset="0x0014" name="DRV_LVL_RES_CODE_OFFSET"/> + <reg32 offset="0x0018" name="DRV_LVL"/> + <reg32 offset="0x001C" name="LANE_CONFIG"/> + <reg32 offset="0x0024" name="PRE_DRIVER_1"/> + <reg32 offset="0x0028" name="PRE_DRIVER_2"/> + <reg32 offset="0x002C" name="LANE_MODE"/> +</domain> + </database> diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/fwsec.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/fwsec.c index 330d72b1a4af..52412965fac1 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/fwsec.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/fwsec.c @@ -324,7 +324,7 @@ nvkm_gsp_fwsec_sb(struct nvkm_gsp *gsp) return ret; /* Verify. */ - err = nvkm_rd32(device, 0x001400 + (0xf * 4)) & 0x0000ffff; + err = nvkm_rd32(device, 0x001400 + (0x15 * 4)) & 0x0000ffff; if (err) { nvkm_error(subdev, "fwsec-sb: 0x%04x\n", err); return -EIO; diff --git a/drivers/gpu/drm/omapdrm/Kconfig b/drivers/gpu/drm/omapdrm/Kconfig index 64e440a2649b..fbd9af758581 100644 --- a/drivers/gpu/drm/omapdrm/Kconfig +++ b/drivers/gpu/drm/omapdrm/Kconfig @@ -5,6 +5,8 @@ config DRM_OMAP depends on DRM && OF depends on ARCH_OMAP2PLUS || (COMPILE_TEST && PAGE_SIZE_LESS_THAN_64KB) select DRM_KMS_HELPER + select DRM_DISPLAY_HELPER + select DRM_BRIDGE_CONNECTOR select FB_DMAMEM_HELPERS_DEFERRED if DRM_FBDEV_EMULATION select VIDEOMODE_HELPERS select HDMI diff --git a/drivers/gpu/drm/panel/panel-newvision-nv3052c.c b/drivers/gpu/drm/panel/panel-newvision-nv3052c.c index c5d3ead38555..d3baccfe6286 100644 --- a/drivers/gpu/drm/panel/panel-newvision-nv3052c.c +++ b/drivers/gpu/drm/panel/panel-newvision-nv3052c.c @@ -925,7 +925,7 @@ MODULE_DEVICE_TABLE(spi, nv3052c_ids); static const struct of_device_id nv3052c_of_match[] = { { .compatible = "leadtek,ltk035c5444t", .data = <k035c5444t_panel_info }, { .compatible = "fascontek,fs035vg158", .data = &fs035vg158_panel_info }, - { .compatible = "wl-355608-a8", .data = &wl_355608_a8_panel_info }, + { .compatible = "anbernic,rg35xx-plus-panel", .data = &wl_355608_a8_panel_info }, { /* sentinel */ } }; MODULE_DEVICE_TABLE(of, nv3052c_of_match); diff --git a/drivers/gpu/drm/panthor/panthor_drv.c b/drivers/gpu/drm/panthor/panthor_drv.c index 444e3bb1cfb5..0caf9e9a8c45 100644 --- a/drivers/gpu/drm/panthor/panthor_drv.c +++ b/drivers/gpu/drm/panthor/panthor_drv.c @@ -14,6 +14,7 @@ #include <linux/platform_device.h> #include <linux/pm_runtime.h> +#include <drm/drm_auth.h> #include <drm/drm_debugfs.h> #include <drm/drm_drv.h> #include <drm/drm_exec.h> @@ -1036,6 +1037,24 @@ static int panthor_ioctl_group_destroy(struct drm_device *ddev, void *data, return panthor_group_destroy(pfile, args->group_handle); } +static int group_priority_permit(struct drm_file *file, + u8 priority) +{ + /* Ensure that priority is valid */ + if (priority > PANTHOR_GROUP_PRIORITY_HIGH) + return -EINVAL; + + /* Medium priority and below are always allowed */ + if (priority <= PANTHOR_GROUP_PRIORITY_MEDIUM) + return 0; + + /* Higher priorities require CAP_SYS_NICE or DRM_MASTER */ + if (capable(CAP_SYS_NICE) || drm_is_current_master(file)) + return 0; + + return -EACCES; +} + static int panthor_ioctl_group_create(struct drm_device *ddev, void *data, struct drm_file *file) { @@ -1051,6 +1070,10 @@ static int panthor_ioctl_group_create(struct drm_device *ddev, void *data, if (ret) return ret; + ret = group_priority_permit(file, args->priority); + if (ret) + return ret; + ret = panthor_group_create(pfile, args, queue_args); if (ret >= 0) { args->group_handle = ret; diff --git a/drivers/gpu/drm/panthor/panthor_fw.c b/drivers/gpu/drm/panthor/panthor_fw.c index 857f3f11258a..ef232c0c2049 100644 --- a/drivers/gpu/drm/panthor/panthor_fw.c +++ b/drivers/gpu/drm/panthor/panthor_fw.c @@ -1089,6 +1089,12 @@ int panthor_fw_post_reset(struct panthor_device *ptdev) panthor_fw_stop(ptdev); ptdev->fw->fast_reset = false; drm_err(&ptdev->base, "FW fast reset failed, trying a slow reset"); + + ret = panthor_vm_flush_all(ptdev->fw->vm); + if (ret) { + drm_err(&ptdev->base, "FW slow reset failed (couldn't flush FW's AS l2cache)"); + return ret; + } } /* Reload all sections, including RO ones. We're not supposed @@ -1099,7 +1105,7 @@ int panthor_fw_post_reset(struct panthor_device *ptdev) ret = panthor_fw_start(ptdev); if (ret) { - drm_err(&ptdev->base, "FW slow reset failed"); + drm_err(&ptdev->base, "FW slow reset failed (couldn't start the FW )"); return ret; } diff --git a/drivers/gpu/drm/panthor/panthor_mmu.c b/drivers/gpu/drm/panthor/panthor_mmu.c index 7d0a90559182..37f1885c54c7 100644 --- a/drivers/gpu/drm/panthor/panthor_mmu.c +++ b/drivers/gpu/drm/panthor/panthor_mmu.c @@ -576,6 +576,12 @@ static int mmu_hw_do_operation_locked(struct panthor_device *ptdev, int as_nr, if (as_nr < 0) return 0; + /* + * If the AS number is greater than zero, then we can be sure + * the device is up and running, so we don't need to explicitly + * power it up + */ + if (op != AS_COMMAND_UNLOCK) lock_region(ptdev, as_nr, iova, size); @@ -874,14 +880,23 @@ static int panthor_vm_flush_range(struct panthor_vm *vm, u64 iova, u64 size) if (!drm_dev_enter(&ptdev->base, &cookie)) return 0; - /* Flush the PTs only if we're already awake */ - if (pm_runtime_active(ptdev->base.dev)) - ret = mmu_hw_do_operation(vm, iova, size, AS_COMMAND_FLUSH_PT); + ret = mmu_hw_do_operation(vm, iova, size, AS_COMMAND_FLUSH_PT); drm_dev_exit(cookie); return ret; } +/** + * panthor_vm_flush_all() - Flush L2 caches for the entirety of a VM's AS + * @vm: VM whose cache to flush + * + * Return: 0 on success, a negative error code if flush failed. + */ +int panthor_vm_flush_all(struct panthor_vm *vm) +{ + return panthor_vm_flush_range(vm, vm->base.mm_start, vm->base.mm_range); +} + static int panthor_vm_unmap_pages(struct panthor_vm *vm, u64 iova, u64 size) { struct panthor_device *ptdev = vm->ptdev; diff --git a/drivers/gpu/drm/panthor/panthor_mmu.h b/drivers/gpu/drm/panthor/panthor_mmu.h index f3c1ed19f973..6788771071e3 100644 --- a/drivers/gpu/drm/panthor/panthor_mmu.h +++ b/drivers/gpu/drm/panthor/panthor_mmu.h @@ -31,6 +31,7 @@ panthor_vm_get_bo_for_va(struct panthor_vm *vm, u64 va, u64 *bo_offset); int panthor_vm_active(struct panthor_vm *vm); void panthor_vm_idle(struct panthor_vm *vm); int panthor_vm_as(struct panthor_vm *vm); +int panthor_vm_flush_all(struct panthor_vm *vm); struct panthor_heap_pool * panthor_vm_get_heap_pool(struct panthor_vm *vm, bool create); diff --git a/drivers/gpu/drm/panthor/panthor_sched.c b/drivers/gpu/drm/panthor/panthor_sched.c index d3246f7d9591..42afdf0ddb7e 100644 --- a/drivers/gpu/drm/panthor/panthor_sched.c +++ b/drivers/gpu/drm/panthor/panthor_sched.c @@ -3092,7 +3092,7 @@ int panthor_group_create(struct panthor_file *pfile, if (group_args->pad) return -EINVAL; - if (group_args->priority > PANTHOR_CSG_PRIORITY_HIGH) + if (group_args->priority >= PANTHOR_CSG_PRIORITY_COUNT) return -EINVAL; if ((group_args->compute_core_mask & ~ptdev->gpu_info.shader_present) || diff --git a/drivers/gpu/drm/renesas/rcar-du/Kconfig b/drivers/gpu/drm/renesas/rcar-du/Kconfig index be86ecb9f559..e1f41468a9a6 100644 --- a/drivers/gpu/drm/renesas/rcar-du/Kconfig +++ b/drivers/gpu/drm/renesas/rcar-du/Kconfig @@ -5,6 +5,8 @@ config DRM_RCAR_DU depends on ARM || ARM64 || COMPILE_TEST depends on ARCH_RENESAS || COMPILE_TEST select DRM_KMS_HELPER + select DRM_DISPLAY_HELPER + select DRM_BRIDGE_CONNECTOR select DRM_GEM_DMA_HELPER select VIDEOMODE_HELPERS help diff --git a/drivers/gpu/drm/renesas/rz-du/Kconfig b/drivers/gpu/drm/renesas/rz-du/Kconfig index 8ec14271ebba..89bdb598e0ae 100644 --- a/drivers/gpu/drm/renesas/rz-du/Kconfig +++ b/drivers/gpu/drm/renesas/rz-du/Kconfig @@ -6,6 +6,8 @@ config DRM_RZG2L_DU depends on VIDEO_RENESAS_VSP1 select DRM_GEM_DMA_HELPER select DRM_KMS_HELPER + select DRM_DISPLAY_HELPER + select DRM_BRIDGE_CONNECTOR select VIDEOMODE_HELPERS help Choose this option if you have an RZ/G2L alike chipset. diff --git a/drivers/gpu/drm/renesas/shmobile/Kconfig b/drivers/gpu/drm/renesas/shmobile/Kconfig index 027220b8fe1c..c329ab8a7a8b 100644 --- a/drivers/gpu/drm/renesas/shmobile/Kconfig +++ b/drivers/gpu/drm/renesas/shmobile/Kconfig @@ -5,6 +5,8 @@ config DRM_SHMOBILE depends on ARCH_RENESAS || ARCH_SHMOBILE || COMPILE_TEST select BACKLIGHT_CLASS_DEVICE select DRM_KMS_HELPER + select DRM_DISPLAY_HELPER + select DRM_BRIDGE_CONNECTOR select DRM_GEM_DMA_HELPER select VIDEOMODE_HELPERS help diff --git a/drivers/gpu/drm/rockchip/Kconfig b/drivers/gpu/drm/rockchip/Kconfig index 7df875e38517..23c49e91f1cc 100644 --- a/drivers/gpu/drm/rockchip/Kconfig +++ b/drivers/gpu/drm/rockchip/Kconfig @@ -86,6 +86,8 @@ config ROCKCHIP_LVDS bool "Rockchip LVDS support" depends on DRM_ROCKCHIP depends on PINCTRL && OF + select DRM_DISPLAY_HELPER + select DRM_BRIDGE_CONNECTOR help Choose this option to enable support for Rockchip LVDS controllers. Rockchip rk3288 SoC has LVDS TX Controller can be used, and it @@ -96,6 +98,8 @@ config ROCKCHIP_RGB bool "Rockchip RGB support" depends on DRM_ROCKCHIP depends on PINCTRL + select DRM_DISPLAY_HELPER + select DRM_BRIDGE_CONNECTOR help Choose this option to enable support for Rockchip RGB output. Some Rockchip CRTCs, like rv1108, can directly output parallel diff --git a/drivers/gpu/drm/tegra/Kconfig b/drivers/gpu/drm/tegra/Kconfig index 782f51d3044a..e688d8104652 100644 --- a/drivers/gpu/drm/tegra/Kconfig +++ b/drivers/gpu/drm/tegra/Kconfig @@ -8,6 +8,7 @@ config DRM_TEGRA select DRM_DISPLAY_DP_HELPER select DRM_DISPLAY_HDMI_HELPER select DRM_DISPLAY_HELPER + select DRM_BRIDGE_CONNECTOR select DRM_DISPLAY_DP_AUX_BUS select DRM_KMS_HELPER select DRM_MIPI_DSI diff --git a/drivers/gpu/drm/tidss/Kconfig b/drivers/gpu/drm/tidss/Kconfig index 378600806167..2385c56493b9 100644 --- a/drivers/gpu/drm/tidss/Kconfig +++ b/drivers/gpu/drm/tidss/Kconfig @@ -3,6 +3,8 @@ config DRM_TIDSS depends on DRM && OF depends on ARM || ARM64 || COMPILE_TEST select DRM_KMS_HELPER + select DRM_DISPLAY_HELPER + select DRM_BRIDGE_CONNECTOR select DRM_GEM_DMA_HELPER help The TI Keystone family SoCs introduced a new generation of diff --git a/drivers/gpu/drm/v3d/v3d_sched.c b/drivers/gpu/drm/v3d/v3d_sched.c index 090bbaebb496..f58c38d7de29 100644 --- a/drivers/gpu/drm/v3d/v3d_sched.c +++ b/drivers/gpu/drm/v3d/v3d_sched.c @@ -136,6 +136,8 @@ v3d_job_start_stats(struct v3d_job *job, enum v3d_queue queue) struct v3d_stats *local_stats = &file->stats[queue]; u64 now = local_clock(); + preempt_disable(); + write_seqcount_begin(&local_stats->lock); local_stats->start_ns = now; write_seqcount_end(&local_stats->lock); @@ -143,6 +145,8 @@ v3d_job_start_stats(struct v3d_job *job, enum v3d_queue queue) write_seqcount_begin(&global_stats->lock); global_stats->start_ns = now; write_seqcount_end(&global_stats->lock); + + preempt_enable(); } static void @@ -164,8 +168,10 @@ v3d_job_update_stats(struct v3d_job *job, enum v3d_queue queue) struct v3d_stats *local_stats = &file->stats[queue]; u64 now = local_clock(); + preempt_disable(); v3d_stats_update(local_stats, now); v3d_stats_update(global_stats, now); + preempt_enable(); } static struct dma_fence *v3d_bin_job_run(struct drm_sched_job *sched_job) diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_blit.c b/drivers/gpu/drm/vmwgfx/vmwgfx_blit.c index 717d624e9a05..890a66a2361f 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_blit.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_blit.c @@ -27,6 +27,8 @@ **************************************************************************/ #include "vmwgfx_drv.h" + +#include "vmwgfx_bo.h" #include <linux/highmem.h> /* @@ -420,13 +422,105 @@ static int vmw_bo_cpu_blit_line(struct vmw_bo_blit_line_data *d, return 0; } +static void *map_external(struct vmw_bo *bo, struct iosys_map *map) +{ + struct vmw_private *vmw = + container_of(bo->tbo.bdev, struct vmw_private, bdev); + void *ptr = NULL; + int ret; + + if (bo->tbo.base.import_attach) { + ret = dma_buf_vmap(bo->tbo.base.dma_buf, map); + if (ret) { + drm_dbg_driver(&vmw->drm, + "Wasn't able to map external bo!\n"); + goto out; + } + ptr = map->vaddr; + } else { + ptr = vmw_bo_map_and_cache(bo); + } + +out: + return ptr; +} + +static void unmap_external(struct vmw_bo *bo, struct iosys_map *map) +{ + if (bo->tbo.base.import_attach) + dma_buf_vunmap(bo->tbo.base.dma_buf, map); + else + vmw_bo_unmap(bo); +} + +static int vmw_external_bo_copy(struct vmw_bo *dst, u32 dst_offset, + u32 dst_stride, struct vmw_bo *src, + u32 src_offset, u32 src_stride, + u32 width_in_bytes, u32 height, + struct vmw_diff_cpy *diff) +{ + struct vmw_private *vmw = + container_of(dst->tbo.bdev, struct vmw_private, bdev); + size_t dst_size = dst->tbo.resource->size; + size_t src_size = src->tbo.resource->size; + struct iosys_map dst_map = {0}; + struct iosys_map src_map = {0}; + int ret, i; + int x_in_bytes; + u8 *vsrc; + u8 *vdst; + + vsrc = map_external(src, &src_map); + if (!vsrc) { + drm_dbg_driver(&vmw->drm, "Wasn't able to map src\n"); + ret = -ENOMEM; + goto out; + } + + vdst = map_external(dst, &dst_map); + if (!vdst) { + drm_dbg_driver(&vmw->drm, "Wasn't able to map dst\n"); + ret = -ENOMEM; + goto out; + } + + vsrc += src_offset; + vdst += dst_offset; + if (src_stride == dst_stride) { + dst_size -= dst_offset; + src_size -= src_offset; + memcpy(vdst, vsrc, + min(dst_stride * height, min(dst_size, src_size))); + } else { + WARN_ON(dst_stride < width_in_bytes); + for (i = 0; i < height; ++i) { + memcpy(vdst, vsrc, width_in_bytes); + vsrc += src_stride; + vdst += dst_stride; + } + } + + x_in_bytes = (dst_offset % dst_stride); + diff->rect.x1 = x_in_bytes / diff->cpp; + diff->rect.y1 = ((dst_offset - x_in_bytes) / dst_stride); + diff->rect.x2 = diff->rect.x1 + width_in_bytes / diff->cpp; + diff->rect.y2 = diff->rect.y1 + height; + + ret = 0; +out: + unmap_external(src, &src_map); + unmap_external(dst, &dst_map); + + return ret; +} + /** * vmw_bo_cpu_blit - in-kernel cpu blit. * - * @dst: Destination buffer object. + * @vmw_dst: Destination buffer object. * @dst_offset: Destination offset of blit start in bytes. * @dst_stride: Destination stride in bytes. - * @src: Source buffer object. + * @vmw_src: Source buffer object. * @src_offset: Source offset of blit start in bytes. * @src_stride: Source stride in bytes. * @w: Width of blit. @@ -444,13 +538,15 @@ static int vmw_bo_cpu_blit_line(struct vmw_bo_blit_line_data *d, * Neither of the buffer objects may be placed in PCI memory * (Fixed memory in TTM terminology) when using this function. */ -int vmw_bo_cpu_blit(struct ttm_buffer_object *dst, +int vmw_bo_cpu_blit(struct vmw_bo *vmw_dst, u32 dst_offset, u32 dst_stride, - struct ttm_buffer_object *src, + struct vmw_bo *vmw_src, u32 src_offset, u32 src_stride, u32 w, u32 h, struct vmw_diff_cpy *diff) { + struct ttm_buffer_object *src = &vmw_src->tbo; + struct ttm_buffer_object *dst = &vmw_dst->tbo; struct ttm_operation_ctx ctx = { .interruptible = false, .no_wait_gpu = false @@ -460,6 +556,11 @@ int vmw_bo_cpu_blit(struct ttm_buffer_object *dst, int ret = 0; struct page **dst_pages = NULL; struct page **src_pages = NULL; + bool src_external = (src->ttm->page_flags & TTM_TT_FLAG_EXTERNAL) != 0; + bool dst_external = (dst->ttm->page_flags & TTM_TT_FLAG_EXTERNAL) != 0; + + if (WARN_ON(dst == src)) + return -EINVAL; /* Buffer objects need to be either pinned or reserved: */ if (!(dst->pin_count)) @@ -479,6 +580,11 @@ int vmw_bo_cpu_blit(struct ttm_buffer_object *dst, return ret; } + if (src_external || dst_external) + return vmw_external_bo_copy(vmw_dst, dst_offset, dst_stride, + vmw_src, src_offset, src_stride, + w, h, diff); + if (!src->ttm->pages && src->ttm->sg) { src_pages = kvmalloc_array(src->ttm->num_pages, sizeof(struct page *), GFP_KERNEL); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c index f42ebc4a7c22..a0e433fbcba6 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c @@ -360,6 +360,8 @@ void *vmw_bo_map_and_cache_size(struct vmw_bo *vbo, size_t size) void *virtual; int ret; + atomic_inc(&vbo->map_count); + virtual = ttm_kmap_obj_virtual(&vbo->map, ¬_used); if (virtual) return virtual; @@ -383,11 +385,17 @@ void *vmw_bo_map_and_cache_size(struct vmw_bo *vbo, size_t size) */ void vmw_bo_unmap(struct vmw_bo *vbo) { + int map_count; + if (vbo->map.bo == NULL) return; - ttm_bo_kunmap(&vbo->map); - vbo->map.bo = NULL; + map_count = atomic_dec_return(&vbo->map_count); + + if (!map_count) { + ttm_bo_kunmap(&vbo->map); + vbo->map.bo = NULL; + } } @@ -421,6 +429,7 @@ static int vmw_bo_init(struct vmw_private *dev_priv, vmw_bo->tbo.priority = 3; vmw_bo->res_tree = RB_ROOT; xa_init(&vmw_bo->detached_resources); + atomic_set(&vmw_bo->map_count, 0); params->size = ALIGN(params->size, PAGE_SIZE); drm_gem_private_object_init(vdev, &vmw_bo->tbo.base, params->size); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.h b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.h index 62b4342d5f7c..43b5439ec9f7 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.h @@ -71,6 +71,8 @@ struct vmw_bo_params { * @map: Kmap object for semi-persistent mappings * @res_tree: RB tree of resources using this buffer object as a backing MOB * @res_prios: Eviction priority counts for attached resources + * @map_count: The number of currently active maps. Will differ from the + * cpu_writers because it includes kernel maps. * @cpu_writers: Number of synccpu write grabs. Protected by reservation when * increased. May be decreased without reservation. * @dx_query_ctx: DX context if this buffer object is used as a DX query MOB @@ -90,6 +92,7 @@ struct vmw_bo { u32 res_prios[TTM_MAX_BO_PRIORITY]; struct xarray detached_resources; + atomic_t map_count; atomic_t cpu_writers; /* Not ref-counted. Protected by binding_mutex */ struct vmw_resource *dx_query_ctx; diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h index 32f50e595809..3f4719b3c268 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h @@ -1353,9 +1353,9 @@ void vmw_diff_memcpy(struct vmw_diff_cpy *diff, u8 *dest, const u8 *src, void vmw_memcpy(struct vmw_diff_cpy *diff, u8 *dest, const u8 *src, size_t n); -int vmw_bo_cpu_blit(struct ttm_buffer_object *dst, +int vmw_bo_cpu_blit(struct vmw_bo *dst, u32 dst_offset, u32 dst_stride, - struct ttm_buffer_object *src, + struct vmw_bo *src, u32 src_offset, u32 src_stride, u32 w, u32 h, struct vmw_diff_cpy *diff); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c b/drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c index 5453f7cf0e2d..fab155a68054 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c @@ -502,7 +502,7 @@ static void vmw_stdu_bo_cpu_commit(struct vmw_kms_dirty *dirty) container_of(dirty->unit, typeof(*stdu), base); s32 width, height; s32 src_pitch, dst_pitch; - struct ttm_buffer_object *src_bo, *dst_bo; + struct vmw_bo *src_bo, *dst_bo; u32 src_offset, dst_offset; struct vmw_diff_cpy diff = VMW_CPU_BLIT_DIFF_INITIALIZER(stdu->cpp); @@ -517,11 +517,11 @@ static void vmw_stdu_bo_cpu_commit(struct vmw_kms_dirty *dirty) /* Assume we are blitting from Guest (bo) to Host (display_srf) */ src_pitch = stdu->display_srf->metadata.base_size.width * stdu->cpp; - src_bo = &stdu->display_srf->res.guest_memory_bo->tbo; + src_bo = stdu->display_srf->res.guest_memory_bo; src_offset = ddirty->top * src_pitch + ddirty->left * stdu->cpp; dst_pitch = ddirty->pitch; - dst_bo = &ddirty->buf->tbo; + dst_bo = ddirty->buf; dst_offset = ddirty->fb_top * dst_pitch + ddirty->fb_left * stdu->cpp; (void) vmw_bo_cpu_blit(dst_bo, dst_offset, dst_pitch, @@ -1170,7 +1170,7 @@ vmw_stdu_bo_populate_update_cpu(struct vmw_du_update_plane *update, void *cmd, struct vmw_diff_cpy diff = VMW_CPU_BLIT_DIFF_INITIALIZER(0); struct vmw_stdu_update_gb_image *cmd_img = cmd; struct vmw_stdu_update *cmd_update; - struct ttm_buffer_object *src_bo, *dst_bo; + struct vmw_bo *src_bo, *dst_bo; u32 src_offset, dst_offset; s32 src_pitch, dst_pitch; s32 width, height; @@ -1184,11 +1184,11 @@ vmw_stdu_bo_populate_update_cpu(struct vmw_du_update_plane *update, void *cmd, diff.cpp = stdu->cpp; - dst_bo = &stdu->display_srf->res.guest_memory_bo->tbo; + dst_bo = stdu->display_srf->res.guest_memory_bo; dst_pitch = stdu->display_srf->metadata.base_size.width * stdu->cpp; dst_offset = bb->y1 * dst_pitch + bb->x1 * stdu->cpp; - src_bo = &vfbbo->buffer->tbo; + src_bo = vfbbo->buffer; src_pitch = update->vfb->base.pitches[0]; src_offset = bo_update->fb_top * src_pitch + bo_update->fb_left * stdu->cpp; diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_surface.c b/drivers/gpu/drm/vmwgfx/vmwgfx_surface.c index 8ae6a761c900..1625b30d9970 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_surface.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_surface.c @@ -2283,9 +2283,11 @@ int vmw_dumb_create(struct drm_file *file_priv, /* * Without mob support we're just going to use raw memory buffer * because we wouldn't be able to support full surface coherency - * without mobs + * without mobs. There also no reason to support surface coherency + * without 3d (i.e. gpu usage on the host) because then all the + * contents is going to be rendered guest side. */ - if (!dev_priv->has_mob) { + if (!dev_priv->has_mob || !vmw_supports_3d(dev_priv)) { int cpp = DIV_ROUND_UP(args->bpp, 8); switch (cpp) { diff --git a/drivers/gpu/drm/xe/Makefile b/drivers/gpu/drm/xe/Makefile index b9670ae09a9e..edfd812e0f41 100644 --- a/drivers/gpu/drm/xe/Makefile +++ b/drivers/gpu/drm/xe/Makefile @@ -40,6 +40,7 @@ xe-y += xe_bb.o \ xe_ggtt.o \ xe_gpu_scheduler.o \ xe_gsc.o \ + xe_gsc_debugfs.o \ xe_gsc_proxy.o \ xe_gsc_submit.o \ xe_gt.o \ diff --git a/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h b/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h index b7b12b20e390..f27a2c75b56d 100644 --- a/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h +++ b/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h @@ -21,11 +21,6 @@ static inline struct drm_i915_private *to_i915(const struct drm_device *dev) return container_of(dev, struct drm_i915_private, drm); } -static inline struct drm_i915_private *kdev_to_i915(struct device *kdev) -{ - return dev_get_drvdata(kdev); -} - #define IS_PLATFORM(xe, x) ((xe)->info.platform == x) #define INTEL_INFO(dev_priv) (&((dev_priv)->info)) #define IS_I830(dev_priv) (dev_priv && 0) @@ -80,9 +75,9 @@ static inline struct drm_i915_private *kdev_to_i915(struct device *kdev) #define IS_MOBILE(xe) (xe && 0) -#define IS_LP(xe) (0) -#define IS_GEN9_LP(xe) (0) -#define IS_GEN9_BC(xe) (0) +#define IS_LP(xe) ((xe) && 0) +#define IS_GEN9_LP(xe) ((xe) && 0) +#define IS_GEN9_BC(xe) ((xe) && 0) #define IS_TIGERLAKE_UY(xe) (xe && 0) #define IS_COMETLAKE_ULX(xe) (xe && 0) @@ -110,8 +105,6 @@ struct i915_sched_attr { }; #define i915_gem_fence_wait_priority(fence, attr) do { (void) attr; } while (0) -#define pdev_to_i915 pdev_to_xe_device - #define FORCEWAKE_ALL XE_FORCEWAKE_ALL #ifdef CONFIG_ARM64 diff --git a/drivers/gpu/drm/xe/compat-i915-headers/intel_pcode.h b/drivers/gpu/drm/xe/compat-i915-headers/intel_pcode.h index 0c47661bdc6a..a473aa6697d0 100644 --- a/drivers/gpu/drm/xe/compat-i915-headers/intel_pcode.h +++ b/drivers/gpu/drm/xe/compat-i915-headers/intel_pcode.h @@ -13,7 +13,7 @@ static inline int snb_pcode_write_timeout(struct intel_uncore *uncore, u32 mbox, u32 val, int fast_timeout_us, int slow_timeout_ms) { - return xe_pcode_write_timeout(__compat_uncore_to_gt(uncore), mbox, val, + return xe_pcode_write_timeout(__compat_uncore_to_tile(uncore), mbox, val, slow_timeout_ms ?: 1); } @@ -21,13 +21,13 @@ static inline int snb_pcode_write(struct intel_uncore *uncore, u32 mbox, u32 val) { - return xe_pcode_write(__compat_uncore_to_gt(uncore), mbox, val); + return xe_pcode_write(__compat_uncore_to_tile(uncore), mbox, val); } static inline int snb_pcode_read(struct intel_uncore *uncore, u32 mbox, u32 *val, u32 *val1) { - return xe_pcode_read(__compat_uncore_to_gt(uncore), mbox, val, val1); + return xe_pcode_read(__compat_uncore_to_tile(uncore), mbox, val, val1); } static inline int @@ -35,7 +35,7 @@ skl_pcode_request(struct intel_uncore *uncore, u32 mbox, u32 request, u32 reply_mask, u32 reply, int timeout_base_ms) { - return xe_pcode_request(__compat_uncore_to_gt(uncore), mbox, request, reply_mask, reply, + return xe_pcode_request(__compat_uncore_to_tile(uncore), mbox, request, reply_mask, reply, timeout_base_ms); } diff --git a/drivers/gpu/drm/xe/compat-i915-headers/intel_uncore.h b/drivers/gpu/drm/xe/compat-i915-headers/intel_uncore.h index 083c4da2ea41..eb5b5f0e4bd9 100644 --- a/drivers/gpu/drm/xe/compat-i915-headers/intel_uncore.h +++ b/drivers/gpu/drm/xe/compat-i915-headers/intel_uncore.h @@ -17,6 +17,13 @@ static inline struct xe_gt *__compat_uncore_to_gt(struct intel_uncore *uncore) return xe_root_mmio_gt(xe); } +static inline struct xe_tile *__compat_uncore_to_tile(struct intel_uncore *uncore) +{ + struct xe_device *xe = container_of(uncore, struct xe_device, uncore); + + return xe_device_get_root_tile(xe); +} + static inline u32 intel_uncore_read(struct intel_uncore *uncore, i915_reg_t i915_reg) { diff --git a/drivers/gpu/drm/xe/display/intel_fbdev_fb.c b/drivers/gpu/drm/xe/display/intel_fbdev_fb.c index cd8948c08661..99499d6c0256 100644 --- a/drivers/gpu/drm/xe/display/intel_fbdev_fb.c +++ b/drivers/gpu/drm/xe/display/intel_fbdev_fb.c @@ -8,7 +8,6 @@ #include "intel_display_types.h" #include "intel_fbdev_fb.h" #include "xe_bo.h" -#include "xe_gt.h" #include "xe_ttm_stolen_mgr.h" #include "xe_wa.h" diff --git a/drivers/gpu/drm/xe/display/xe_display.c b/drivers/gpu/drm/xe/display/xe_display.c index 78a884ddd499..75736faf2a80 100644 --- a/drivers/gpu/drm/xe/display/xe_display.c +++ b/drivers/gpu/drm/xe/display/xe_display.c @@ -10,7 +10,7 @@ #include <drm/drm_drv.h> #include <drm/drm_managed.h> -#include <drm/xe_drm.h> +#include <uapi/drm/xe_drm.h> #include "soc/intel_dram.h" #include "i915_drv.h" /* FIXME: HAS_DISPLAY() depends on this */ @@ -341,16 +341,14 @@ void xe_display_pm_suspend(struct xe_device *xe, bool runtime) xe_display_flush_cleanup_work(xe); - xe_display_flush_cleanup_work(xe); - intel_dp_mst_suspend(xe); intel_hpd_cancel_work(xe); - if (!runtime && has_display(xe)) + if (!runtime && has_display(xe)) { intel_display_driver_suspend_access(xe); - - intel_encoder_suspend_all(&xe->display); + intel_encoder_suspend_all(&xe->display); + } intel_opregion_suspend(display, s2idle ? PCI_D1 : PCI_D3cold); diff --git a/drivers/gpu/drm/xe/display/xe_dsb_buffer.c b/drivers/gpu/drm/xe/display/xe_dsb_buffer.c index ccd0d87d438a..f99d901a3214 100644 --- a/drivers/gpu/drm/xe/display/xe_dsb_buffer.c +++ b/drivers/gpu/drm/xe/display/xe_dsb_buffer.c @@ -9,7 +9,6 @@ #include "xe_bo.h" #include "xe_device.h" #include "xe_device_types.h" -#include "xe_gt.h" u32 intel_dsb_buffer_ggtt_offset(struct intel_dsb_buffer *dsb_buf) { diff --git a/drivers/gpu/drm/xe/display/xe_fb_pin.c b/drivers/gpu/drm/xe/display/xe_fb_pin.c index d650c5ac41a4..b58fc4ba2aac 100644 --- a/drivers/gpu/drm/xe/display/xe_fb_pin.c +++ b/drivers/gpu/drm/xe/display/xe_fb_pin.c @@ -12,7 +12,6 @@ #include "xe_bo.h" #include "xe_device.h" #include "xe_ggtt.h" -#include "xe_gt.h" #include "xe_pm.h" static void diff --git a/drivers/gpu/drm/xe/display/xe_hdcp_gsc.c b/drivers/gpu/drm/xe/display/xe_hdcp_gsc.c index 0af667ebebf9..6619a40aed15 100644 --- a/drivers/gpu/drm/xe/display/xe_hdcp_gsc.c +++ b/drivers/gpu/drm/xe/display/xe_hdcp_gsc.c @@ -16,7 +16,6 @@ #include "xe_force_wake.h" #include "xe_gsc_proxy.h" #include "xe_gsc_submit.h" -#include "xe_gt.h" #include "xe_map.h" #include "xe_pm.h" #include "xe_uc_fw.h" diff --git a/drivers/gpu/drm/xe/regs/xe_gsc_regs.h b/drivers/gpu/drm/xe/regs/xe_gsc_regs.h index e2a925be137c..7702364b65f1 100644 --- a/drivers/gpu/drm/xe/regs/xe_gsc_regs.h +++ b/drivers/gpu/drm/xe/regs/xe_gsc_regs.h @@ -32,8 +32,12 @@ #define HECI1_FWSTS1_CURRENT_STATE_RESET 0 #define HECI1_FWSTS1_PROXY_STATE_NORMAL 5 #define HECI1_FWSTS1_INIT_COMPLETE REG_BIT(9) +#define HECI_FWSTS2(base) XE_REG((base) + 0xc48) +#define HECI_FWSTS3(base) XE_REG((base) + 0xc60) +#define HECI_FWSTS4(base) XE_REG((base) + 0xc64) #define HECI_FWSTS5(base) XE_REG((base) + 0xc68) #define HECI1_FWSTS5_HUC_AUTH_DONE REG_BIT(19) +#define HECI_FWSTS6(base) XE_REG((base) + 0xc6c) #define HECI_H_GS1(base) XE_REG((base) + 0xc4c) #define HECI_H_GS1_ER_PREP REG_BIT(0) diff --git a/drivers/gpu/drm/xe/tests/xe_dma_buf.c b/drivers/gpu/drm/xe/tests/xe_dma_buf.c index 13db6c0530b3..cedd3e88a6fb 100644 --- a/drivers/gpu/drm/xe/tests/xe_dma_buf.c +++ b/drivers/gpu/drm/xe/tests/xe_dma_buf.c @@ -3,7 +3,7 @@ * Copyright © 2022 Intel Corporation */ -#include <drm/xe_drm.h> +#include <uapi/drm/xe_drm.h> #include <kunit/test.h> #include <kunit/visibility.h> diff --git a/drivers/gpu/drm/xe/xe_assert.h b/drivers/gpu/drm/xe/xe_assert.h index 8b0cc1bc9327..e22bbf57fca7 100644 --- a/drivers/gpu/drm/xe/xe_assert.h +++ b/drivers/gpu/drm/xe/xe_assert.h @@ -81,7 +81,7 @@ #if IS_ENABLED(CONFIG_DRM_XE_DEBUG) #define __xe_assert_msg(xe, condition, msg, arg...) ({ \ - (void)drm_WARN(&(xe)->drm, !(condition), "[" DRM_NAME "] Assertion `%s` failed!\n" msg, \ + (void)drm_WARN(&(xe)->drm, !(condition), "Assertion `%s` failed!\n" msg, \ __stringify(condition), ## arg); \ }) #else diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c index 25d0c939ba31..06911e9a3bf5 100644 --- a/drivers/gpu/drm/xe/xe_bo.c +++ b/drivers/gpu/drm/xe/xe_bo.c @@ -13,7 +13,7 @@ #include <drm/ttm/ttm_device.h> #include <drm/ttm/ttm_placement.h> #include <drm/ttm/ttm_tt.h> -#include <drm/xe_drm.h> +#include <uapi/drm/xe_drm.h> #include "xe_device.h" #include "xe_dma_buf.h" @@ -758,7 +758,16 @@ static int xe_bo_move(struct ttm_buffer_object *ttm_bo, bool evict, xe_assert(xe, migrate); trace_xe_bo_move(bo, new_mem->mem_type, old_mem_type, move_lacks_source); - xe_pm_runtime_get_noresume(xe); + if (xe_rpm_reclaim_safe(xe)) { + /* + * We might be called through swapout in the validation path of + * another TTM device, so unconditionally acquire rpm here. + */ + xe_pm_runtime_get(xe); + } else { + drm_WARN_ON(&xe->drm, handle_system_ccs); + xe_pm_runtime_get_noresume(xe); + } if (xe_bo_is_pinned(bo) && !xe_bo_is_user(bo)) { /* diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c index b6db7e082d88..1a0d7fdd094b 100644 --- a/drivers/gpu/drm/xe/xe_device.c +++ b/drivers/gpu/drm/xe/xe_device.c @@ -15,7 +15,7 @@ #include <drm/drm_ioctl.h> #include <drm/drm_managed.h> #include <drm/drm_print.h> -#include <drm/xe_drm.h> +#include <uapi/drm/xe_drm.h> #include "display/xe_display.h" #include "instructions/xe_gpu_commands.h" diff --git a/drivers/gpu/drm/xe/xe_device.h b/drivers/gpu/drm/xe/xe_device.h index f052c06a2d2f..894f04770454 100644 --- a/drivers/gpu/drm/xe/xe_device.h +++ b/drivers/gpu/drm/xe/xe_device.h @@ -17,12 +17,16 @@ static inline struct xe_device *to_xe_device(const struct drm_device *dev) static inline struct xe_device *kdev_to_xe_device(struct device *kdev) { - return dev_get_drvdata(kdev); + struct drm_device *drm = dev_get_drvdata(kdev); + + return drm ? to_xe_device(drm) : NULL; } static inline struct xe_device *pdev_to_xe_device(struct pci_dev *pdev) { - return pci_get_drvdata(pdev); + struct drm_device *drm = pci_get_drvdata(pdev); + + return drm ? to_xe_device(drm) : NULL; } static inline struct xe_device *xe_device_const_cast(const struct xe_device *xe) diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h index e73fb0c23932..ec7eb7811126 100644 --- a/drivers/gpu/drm/xe/xe_device_types.h +++ b/drivers/gpu/drm/xe/xe_device_types.h @@ -208,6 +208,12 @@ struct xe_tile { } vf; } sriov; + /** @pcode: tile's PCODE */ + struct { + /** @pcode.lock: protecting tile's PCODE mailbox data */ + struct mutex lock; + } pcode; + /** @migrate: Migration helper for vram blits and clearing */ struct xe_migrate *migrate; diff --git a/drivers/gpu/drm/xe/xe_drm_client.c b/drivers/gpu/drm/xe/xe_drm_client.c index 7ddd59908334..e64f4b645e2e 100644 --- a/drivers/gpu/drm/xe/xe_drm_client.c +++ b/drivers/gpu/drm/xe/xe_drm_client.c @@ -5,7 +5,7 @@ #include "xe_drm_client.h" #include <drm/drm_print.h> -#include <drm/xe_drm.h> +#include <uapi/drm/xe_drm.h> #include <linux/kernel.h> #include <linux/slab.h> #include <linux/types.h> diff --git a/drivers/gpu/drm/xe/xe_exec.c b/drivers/gpu/drm/xe/xe_exec.c index 484acfbe0e61..7b38485817dc 100644 --- a/drivers/gpu/drm/xe/xe_exec.c +++ b/drivers/gpu/drm/xe/xe_exec.c @@ -8,7 +8,7 @@ #include <drm/drm_device.h> #include <drm/drm_exec.h> #include <drm/drm_file.h> -#include <drm/xe_drm.h> +#include <uapi/drm/xe_drm.h> #include <linux/delay.h> #include "xe_bo.h" diff --git a/drivers/gpu/drm/xe/xe_exec_queue.c b/drivers/gpu/drm/xe/xe_exec_queue.c index e53937fafd14..5a9cbc97f0be 100644 --- a/drivers/gpu/drm/xe/xe_exec_queue.c +++ b/drivers/gpu/drm/xe/xe_exec_queue.c @@ -9,7 +9,7 @@ #include <drm/drm_device.h> #include <drm/drm_file.h> -#include <drm/xe_drm.h> +#include <uapi/drm/xe_drm.h> #include "xe_device.h" #include "xe_gt.h" diff --git a/drivers/gpu/drm/xe/xe_execlist.c b/drivers/gpu/drm/xe/xe_execlist.c index 7502e3486eaf..6a59165b9569 100644 --- a/drivers/gpu/drm/xe/xe_execlist.c +++ b/drivers/gpu/drm/xe/xe_execlist.c @@ -123,8 +123,8 @@ static void __xe_execlist_port_idle(struct xe_execlist_port *port) if (!port->running_exl) return; - xe_lrc_write_ring(port->hwe->kernel_lrc, noop, sizeof(noop)); - __start_lrc(port->hwe, port->hwe->kernel_lrc, 0); + xe_lrc_write_ring(port->lrc, noop, sizeof(noop)); + __start_lrc(port->hwe, port->lrc, 0); port->running_exl = NULL; } @@ -254,14 +254,22 @@ struct xe_execlist_port *xe_execlist_port_create(struct xe_device *xe, { struct drm_device *drm = &xe->drm; struct xe_execlist_port *port; - int i; + int i, err; port = drmm_kzalloc(drm, sizeof(*port), GFP_KERNEL); - if (!port) - return ERR_PTR(-ENOMEM); + if (!port) { + err = -ENOMEM; + goto err; + } port->hwe = hwe; + port->lrc = xe_lrc_create(hwe, NULL, SZ_16K); + if (IS_ERR(port->lrc)) { + err = PTR_ERR(port->lrc); + goto err; + } + spin_lock_init(&port->lock); for (i = 0; i < ARRAY_SIZE(port->active); i++) INIT_LIST_HEAD(&port->active[i]); @@ -277,6 +285,9 @@ struct xe_execlist_port *xe_execlist_port_create(struct xe_device *xe, add_timer(&port->irq_fail); return port; + +err: + return ERR_PTR(err); } void xe_execlist_port_destroy(struct xe_execlist_port *port) @@ -287,6 +298,8 @@ void xe_execlist_port_destroy(struct xe_execlist_port *port) spin_lock_irq(>_to_xe(port->hwe->gt)->irq.lock); port->hwe->irq_handler = NULL; spin_unlock_irq(>_to_xe(port->hwe->gt)->irq.lock); + + xe_lrc_put(port->lrc); } static struct dma_fence * diff --git a/drivers/gpu/drm/xe/xe_execlist_types.h b/drivers/gpu/drm/xe/xe_execlist_types.h index f94bbf4c53e4..415140936f11 100644 --- a/drivers/gpu/drm/xe/xe_execlist_types.h +++ b/drivers/gpu/drm/xe/xe_execlist_types.h @@ -27,6 +27,8 @@ struct xe_execlist_port { struct xe_execlist_exec_queue *running_exl; struct timer_list irq_fail; + + struct xe_lrc *lrc; }; struct xe_execlist_exec_queue { diff --git a/drivers/gpu/drm/xe/xe_gsc.c b/drivers/gpu/drm/xe/xe_gsc.c index 8a137cb83318..6fbea70d3d36 100644 --- a/drivers/gpu/drm/xe/xe_gsc.c +++ b/drivers/gpu/drm/xe/xe_gsc.c @@ -8,6 +8,7 @@ #include <linux/delay.h> #include <drm/drm_managed.h> +#include <drm/drm_print.h> #include <generated/xe_wa_oob.h> @@ -165,10 +166,11 @@ static int query_compatibility_version(struct xe_gsc *gsc) return err; } - compat->major = version_query_rd(xe, &bo->vmap, rd_offset, compat_major); - compat->minor = version_query_rd(xe, &bo->vmap, rd_offset, compat_minor); + compat->major = version_query_rd(xe, &bo->vmap, rd_offset, proj_major); + compat->minor = version_query_rd(xe, &bo->vmap, rd_offset, compat_major); + compat->patch = version_query_rd(xe, &bo->vmap, rd_offset, compat_minor); - xe_gt_info(gt, "found GSC cv%u.%u\n", compat->major, compat->minor); + xe_gt_info(gt, "found GSC cv%u.%u.%u\n", compat->major, compat->minor, compat->patch); out_bo: xe_bo_unpin_map_no_vm(bo); @@ -333,9 +335,11 @@ static int gsc_er_complete(struct xe_gt *gt) if (er_status == GSCI_TIMER_STATUS_TIMER_EXPIRED) { /* * XXX: we should trigger an FLR here, but we don't have support - * for that yet. + * for that yet. Since we can't recover from the error, we + * declare the device as wedged. */ xe_gt_err(gt, "GSC ER timed out!\n"); + xe_device_declare_wedged(gt_to_xe(gt)); return -EIO; } @@ -513,13 +517,28 @@ out_bo: void xe_gsc_load_start(struct xe_gsc *gsc) { struct xe_gt *gt = gsc_to_gt(gsc); + struct xe_device *xe = gt_to_xe(gt); if (!xe_uc_fw_is_loadable(&gsc->fw) || !gsc->q) return; + /* + * The GSC HW is only reset by driver FLR or D3cold entry. We don't + * support the former at runtime, while the latter is only supported on + * DGFX, for which we don't support GSC. Therefore, if GSC failed to + * load previously there is no need to try again because the HW is + * stuck in the error state. + */ + xe_assert(xe, !IS_DGFX(xe)); + if (xe_uc_fw_is_in_error_state(&gsc->fw)) + return; + /* GSC FW survives GT reset and D3Hot */ if (gsc_fw_is_loaded(gt)) { - xe_uc_fw_change_status(&gsc->fw, XE_UC_FIRMWARE_TRANSFERRED); + if (xe_gsc_proxy_init_done(gsc)) + xe_uc_fw_change_status(&gsc->fw, XE_UC_FIRMWARE_RUNNING); + else + xe_uc_fw_change_status(&gsc->fw, XE_UC_FIRMWARE_TRANSFERRED); return; } @@ -571,3 +590,35 @@ void xe_gsc_wa_14015076503(struct xe_gt *gt, bool prep) msleep(200); } } + +/** + * xe_gsc_print_info - print info about GSC FW status + * @gsc: the GSC structure + * @p: the printer to be used to print the info + */ +void xe_gsc_print_info(struct xe_gsc *gsc, struct drm_printer *p) +{ + struct xe_gt *gt = gsc_to_gt(gsc); + int err; + + xe_uc_fw_print(&gsc->fw, p); + + drm_printf(p, "\tfound security version %u\n", gsc->security_version); + + if (!xe_uc_fw_is_enabled(&gsc->fw)) + return; + + err = xe_force_wake_get(gt_to_fw(gt), XE_FW_GSC); + if (err) + return; + + drm_printf(p, "\nHECI1 FWSTS: 0x%08x 0x%08x 0x%08x 0x%08x 0x%08x 0x%08x\n", + xe_mmio_read32(gt, HECI_FWSTS1(MTL_GSC_HECI1_BASE)), + xe_mmio_read32(gt, HECI_FWSTS2(MTL_GSC_HECI1_BASE)), + xe_mmio_read32(gt, HECI_FWSTS3(MTL_GSC_HECI1_BASE)), + xe_mmio_read32(gt, HECI_FWSTS4(MTL_GSC_HECI1_BASE)), + xe_mmio_read32(gt, HECI_FWSTS5(MTL_GSC_HECI1_BASE)), + xe_mmio_read32(gt, HECI_FWSTS6(MTL_GSC_HECI1_BASE))); + + xe_force_wake_put(gt_to_fw(gt), XE_FW_GSC); +} diff --git a/drivers/gpu/drm/xe/xe_gsc.h b/drivers/gpu/drm/xe/xe_gsc.h index 1c7a623faf11..e282b9ef6ec4 100644 --- a/drivers/gpu/drm/xe/xe_gsc.h +++ b/drivers/gpu/drm/xe/xe_gsc.h @@ -8,6 +8,7 @@ #include <linux/types.h> +struct drm_printer; struct xe_gsc; struct xe_gt; struct xe_hw_engine; @@ -21,4 +22,6 @@ void xe_gsc_hwe_irq_handler(struct xe_hw_engine *hwe, u16 intr_vec); void xe_gsc_wa_14015076503(struct xe_gt *gt, bool prep); +void xe_gsc_print_info(struct xe_gsc *gsc, struct drm_printer *p); + #endif diff --git a/drivers/gpu/drm/xe/xe_gsc_debugfs.c b/drivers/gpu/drm/xe/xe_gsc_debugfs.c new file mode 100644 index 000000000000..461d7e99c2b3 --- /dev/null +++ b/drivers/gpu/drm/xe/xe_gsc_debugfs.c @@ -0,0 +1,71 @@ +// SPDX-License-Identifier: MIT +/* + * Copyright © 2022 Intel Corporation + */ + +#include "xe_gsc_debugfs.h" + +#include <drm/drm_debugfs.h> +#include <drm/drm_managed.h> + +#include "xe_device.h" +#include "xe_gt.h" +#include "xe_gsc.h" +#include "xe_macros.h" +#include "xe_pm.h" + +static struct xe_gt * +gsc_to_gt(struct xe_gsc *gsc) +{ + return container_of(gsc, struct xe_gt, uc.gsc); +} + +static struct xe_device * +gsc_to_xe(struct xe_gsc *gsc) +{ + return gt_to_xe(gsc_to_gt(gsc)); +} + +static struct xe_gsc *node_to_gsc(struct drm_info_node *node) +{ + return node->info_ent->data; +} + +static int gsc_info(struct seq_file *m, void *data) +{ + struct xe_gsc *gsc = node_to_gsc(m->private); + struct xe_device *xe = gsc_to_xe(gsc); + struct drm_printer p = drm_seq_file_printer(m); + + xe_pm_runtime_get(xe); + xe_gsc_print_info(gsc, &p); + xe_pm_runtime_put(xe); + + return 0; +} + +static const struct drm_info_list debugfs_list[] = { + {"gsc_info", gsc_info, 0}, +}; + +void xe_gsc_debugfs_register(struct xe_gsc *gsc, struct dentry *parent) +{ + struct drm_minor *minor = gsc_to_xe(gsc)->drm.primary; + struct drm_info_list *local; + int i; + +#define DEBUGFS_SIZE (ARRAY_SIZE(debugfs_list) * sizeof(struct drm_info_list)) + local = drmm_kmalloc(&gsc_to_xe(gsc)->drm, DEBUGFS_SIZE, GFP_KERNEL); + if (!local) + return; + + memcpy(local, debugfs_list, DEBUGFS_SIZE); +#undef DEBUGFS_SIZE + + for (i = 0; i < ARRAY_SIZE(debugfs_list); ++i) + local[i].data = gsc; + + drm_debugfs_create_files(local, + ARRAY_SIZE(debugfs_list), + parent, minor); +} diff --git a/drivers/gpu/drm/xe/xe_gsc_debugfs.h b/drivers/gpu/drm/xe/xe_gsc_debugfs.h new file mode 100644 index 000000000000..c2e2645dc705 --- /dev/null +++ b/drivers/gpu/drm/xe/xe_gsc_debugfs.h @@ -0,0 +1,14 @@ +/* SPDX-License-Identifier: MIT */ +/* + * Copyright © 2024 Intel Corporation + */ + +#ifndef _XE_GSC_DEBUGFS_H_ +#define _XE_GSC_DEBUGFS_H_ + +struct dentry; +struct xe_gsc; + +void xe_gsc_debugfs_register(struct xe_gsc *gsc, struct dentry *parent); + +#endif diff --git a/drivers/gpu/drm/xe/xe_gt.c b/drivers/gpu/drm/xe/xe_gt.c index 08a004d698d4..dd96dec95b19 100644 --- a/drivers/gpu/drm/xe/xe_gt.c +++ b/drivers/gpu/drm/xe/xe_gt.c @@ -8,7 +8,7 @@ #include <linux/minmax.h> #include <drm/drm_managed.h> -#include <drm/xe_drm.h> +#include <uapi/drm/xe_drm.h> #include <generated/xe_wa_oob.h> @@ -48,7 +48,6 @@ #include "xe_migrate.h" #include "xe_mmio.h" #include "xe_pat.h" -#include "xe_pcode.h" #include "xe_pm.h" #include "xe_mocs.h" #include "xe_reg_sr.h" @@ -388,7 +387,6 @@ int xe_gt_init_early(struct xe_gt *gt) xe_tuning_process_gt(gt); xe_force_wake_init_gt(gt, gt_to_fw(gt)); - xe_pcode_init(gt); spin_lock_init(>->global_invl_lock); return 0; @@ -756,12 +754,13 @@ static int gt_reset(struct xe_gt *gt) xe_gt_info(gt, "reset started\n"); + xe_pm_runtime_get(gt_to_xe(gt)); + if (xe_fault_inject_gt_reset()) { err = -ECANCELED; goto err_fail; } - xe_pm_runtime_get(gt_to_xe(gt)); xe_gt_sanitize(gt); err = xe_force_wake_get(gt_to_fw(gt), XE_FORCEWAKE_ALL); @@ -796,11 +795,11 @@ err_out: XE_WARN_ON(xe_force_wake_put(gt_to_fw(gt), XE_FORCEWAKE_ALL)); err_msg: XE_WARN_ON(xe_uc_start(>->uc)); - xe_pm_runtime_put(gt_to_xe(gt)); err_fail: xe_gt_err(gt, "reset failed (%pe)\n", ERR_PTR(err)); xe_device_declare_wedged(gt_to_xe(gt)); + xe_pm_runtime_put(gt_to_xe(gt)); return err; } diff --git a/drivers/gpu/drm/xe/xe_gt_pagefault.c b/drivers/gpu/drm/xe/xe_gt_pagefault.c index 0be4687bfc20..730eec07795e 100644 --- a/drivers/gpu/drm/xe/xe_gt_pagefault.c +++ b/drivers/gpu/drm/xe/xe_gt_pagefault.c @@ -388,20 +388,17 @@ static void pagefault_fini(void *arg) { struct xe_gt *gt = arg; struct xe_device *xe = gt_to_xe(gt); - int i; if (!xe->info.has_usm) return; destroy_workqueue(gt->usm.acc_wq); destroy_workqueue(gt->usm.pf_wq); - - for (i = 0; i < NUM_PF_QUEUE; ++i) - kfree(gt->usm.pf_queue[i].data); } static int xe_alloc_pf_queue(struct xe_gt *gt, struct pf_queue *pf_queue) { + struct xe_device *xe = gt_to_xe(gt); xe_dss_mask_t all_dss; int num_dss, num_eus; @@ -417,7 +414,8 @@ static int xe_alloc_pf_queue(struct xe_gt *gt, struct pf_queue *pf_queue) (num_eus + XE_NUM_HW_ENGINES) * PF_MSG_LEN_DW; pf_queue->gt = gt; - pf_queue->data = kcalloc(pf_queue->num_dw, sizeof(u32), GFP_KERNEL); + pf_queue->data = devm_kcalloc(xe->drm.dev, pf_queue->num_dw, + sizeof(u32), GFP_KERNEL); if (!pf_queue->data) return -ENOMEM; diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_pf.c b/drivers/gpu/drm/xe/xe_gt_sriov_pf.c index ef239440963c..905f409db74b 100644 --- a/drivers/gpu/drm/xe/xe_gt_sriov_pf.c +++ b/drivers/gpu/drm/xe/xe_gt_sriov_pf.c @@ -9,6 +9,7 @@ #include "xe_gt_sriov_pf.h" #include "xe_gt_sriov_pf_config.h" +#include "xe_gt_sriov_pf_control.h" #include "xe_gt_sriov_pf_helpers.h" #include "xe_gt_sriov_pf_service.h" #include "xe_mmio.h" @@ -57,6 +58,10 @@ int xe_gt_sriov_pf_init_early(struct xe_gt *gt) if (err) return err; + err = xe_gt_sriov_pf_control_init(gt); + if (err) + return err; + return 0; } @@ -93,4 +98,5 @@ void xe_gt_sriov_pf_init_hw(struct xe_gt *gt) void xe_gt_sriov_pf_restart(struct xe_gt *gt) { xe_gt_sriov_pf_config_restart(gt); + xe_gt_sriov_pf_control_restart(gt); } diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c b/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c index 41ed07b153b5..a95e546b7744 100644 --- a/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c +++ b/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c @@ -29,6 +29,7 @@ #include "xe_guc_submit.h" #include "xe_lmtt.h" #include "xe_map.h" +#include "xe_migrate.h" #include "xe_sriov.h" #include "xe_ttm_vram_mgr.h" #include "xe_wopcm.h" @@ -276,6 +277,14 @@ static u32 encode_config(u32 *cfg, const struct xe_gt_sriov_config *config) cfg[n++] = PREP_GUC_KLV_TAG(VF_CFG_PREEMPT_TIMEOUT); cfg[n++] = config->preempt_timeout; +#define encode_threshold_config(TAG, ...) ({ \ + cfg[n++] = PREP_GUC_KLV_TAG(VF_CFG_THRESHOLD_##TAG); \ + cfg[n++] = config->thresholds[MAKE_XE_GUC_KLV_THRESHOLD_INDEX(TAG)]; \ +}); + + MAKE_XE_GUC_KLV_THRESHOLDS_SET(encode_threshold_config); +#undef encode_threshold_config + return n; } @@ -1833,6 +1842,18 @@ u32 xe_gt_sriov_pf_config_get_threshold(struct xe_gt *gt, unsigned int vfid, return value; } +static void pf_reset_config_thresholds(struct xe_gt *gt, struct xe_gt_sriov_config *config) +{ + lockdep_assert_held(xe_gt_sriov_pf_master_mutex(gt)); + +#define reset_threshold_config(TAG, ...) ({ \ + config->thresholds[MAKE_XE_GUC_KLV_THRESHOLD_INDEX(TAG)] = 0; \ +}); + + MAKE_XE_GUC_KLV_THRESHOLDS_SET(reset_threshold_config); +#undef reset_threshold_config +} + static void pf_release_vf_config(struct xe_gt *gt, unsigned int vfid) { struct xe_gt_sriov_config *config = pf_pick_vf_config(gt, vfid); @@ -1848,6 +1869,7 @@ static void pf_release_vf_config(struct xe_gt *gt, unsigned int vfid) pf_release_config_ctxs(gt, config); pf_release_config_dbs(gt, config); pf_reset_config_sched(gt, config); + pf_reset_config_thresholds(gt, config); } /** @@ -1881,6 +1903,87 @@ int xe_gt_sriov_pf_config_release(struct xe_gt *gt, unsigned int vfid, bool forc return force ? 0 : err; } +static void pf_sanitize_ggtt(struct xe_ggtt_node *ggtt_region, unsigned int vfid) +{ + if (xe_ggtt_node_allocated(ggtt_region)) + xe_ggtt_assign(ggtt_region, vfid); +} + +static int pf_sanitize_lmem(struct xe_tile *tile, struct xe_bo *bo, long timeout) +{ + struct xe_migrate *m = tile->migrate; + struct dma_fence *fence; + int err; + + if (!bo) + return 0; + + xe_bo_lock(bo, false); + fence = xe_migrate_clear(m, bo, bo->ttm.resource, XE_MIGRATE_CLEAR_FLAG_FULL); + if (IS_ERR(fence)) { + err = PTR_ERR(fence); + } else if (!fence) { + err = -ENOMEM; + } else { + long ret = dma_fence_wait_timeout(fence, false, timeout); + + err = ret > 0 ? 0 : ret < 0 ? ret : -ETIMEDOUT; + dma_fence_put(fence); + if (!err) + xe_gt_sriov_dbg_verbose(tile->primary_gt, "LMEM cleared in %dms\n", + jiffies_to_msecs(timeout - ret)); + } + xe_bo_unlock(bo); + + return err; +} + +static int pf_sanitize_vf_resources(struct xe_gt *gt, u32 vfid, long timeout) +{ + struct xe_gt_sriov_config *config = pf_pick_vf_config(gt, vfid); + struct xe_tile *tile = gt_to_tile(gt); + struct xe_device *xe = gt_to_xe(gt); + int err = 0; + + /* + * Only GGTT and LMEM requires to be cleared by the PF. + * GuC doorbell IDs and context IDs do not need any clearing. + */ + if (!xe_gt_is_media_type(gt)) { + pf_sanitize_ggtt(config->ggtt_region, vfid); + if (IS_DGFX(xe)) + err = pf_sanitize_lmem(tile, config->lmem_obj, timeout); + } + + return err; +} + +/** + * xe_gt_sriov_pf_config_sanitize() - Sanitize VF's resources. + * @gt: the &xe_gt + * @vfid: the VF identifier (can't be PF) + * @timeout: maximum timeout to wait for completion in jiffies + * + * This function can only be called on PF. + * + * Return: 0 on success or a negative error code on failure. + */ +int xe_gt_sriov_pf_config_sanitize(struct xe_gt *gt, unsigned int vfid, long timeout) +{ + int err; + + xe_gt_assert(gt, vfid != PFID); + + mutex_lock(xe_gt_sriov_pf_master_mutex(gt)); + err = pf_sanitize_vf_resources(gt, vfid, timeout); + mutex_unlock(xe_gt_sriov_pf_master_mutex(gt)); + + if (unlikely(err)) + xe_gt_sriov_notice(gt, "VF%u resource sanitizing failed (%pe)\n", + vfid, ERR_PTR(err)); + return err; +} + /** * xe_gt_sriov_pf_config_push - Reprovision VF's configuration. * @gt: the &xe_gt diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.h b/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.h index c0e6e4743dc2..42e64769f666 100644 --- a/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.h +++ b/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.h @@ -50,6 +50,7 @@ int xe_gt_sriov_pf_config_set_threshold(struct xe_gt *gt, unsigned int vfid, enum xe_guc_klv_threshold_index index, u32 value); int xe_gt_sriov_pf_config_set_fair(struct xe_gt *gt, unsigned int vfid, unsigned int num_vfs); +int xe_gt_sriov_pf_config_sanitize(struct xe_gt *gt, unsigned int vfid, long timeout); int xe_gt_sriov_pf_config_release(struct xe_gt *gt, unsigned int vfid, bool force); int xe_gt_sriov_pf_config_push(struct xe_gt *gt, unsigned int vfid, bool refresh); diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_pf_control.c b/drivers/gpu/drm/xe/xe_gt_sriov_pf_control.c index ebf06e037750..02f7328bd6ce 100644 --- a/drivers/gpu/drm/xe/xe_gt_sriov_pf_control.c +++ b/drivers/gpu/drm/xe/xe_gt_sriov_pf_control.c @@ -3,11 +3,17 @@ * Copyright © 2023-2024 Intel Corporation */ +#include <drm/drm_managed.h> + #include "abi/guc_actions_sriov_abi.h" #include "xe_device.h" #include "xe_gt.h" +#include "xe_gt_sriov_pf_config.h" #include "xe_gt_sriov_pf_control.h" +#include "xe_gt_sriov_pf_helpers.h" +#include "xe_gt_sriov_pf_monitor.h" +#include "xe_gt_sriov_pf_service.h" #include "xe_gt_sriov_printk.h" #include "xe_guc_ct.h" #include "xe_sriov.h" @@ -41,10 +47,6 @@ static int guc_action_vf_control_cmd(struct xe_guc *guc, u32 vfid, u32 cmd) }; int ret; - /* XXX those two commands are now sent from the G2H handler */ - if (cmd == GUC_PF_TRIGGER_VF_FLR_START || cmd == GUC_PF_TRIGGER_VF_FLR_FINISH) - return xe_guc_ct_send_g2h_handler(&guc->ct, request, ARRAY_SIZE(request)); - ret = xe_guc_ct_send_block(&guc->ct, request, ARRAY_SIZE(request)); return ret > 0 ? -EPROTO : ret; } @@ -54,6 +56,8 @@ static int pf_send_vf_control_cmd(struct xe_gt *gt, unsigned int vfid, u32 cmd) int err; xe_gt_assert(gt, vfid != PFID); + xe_gt_sriov_dbg_verbose(gt, "sending VF%u control command %s\n", + vfid, control_cmd_to_string(cmd)); err = guc_action_vf_control_cmd(>->uc.guc, vfid, cmd); if (unlikely(err)) @@ -88,6 +92,456 @@ static int pf_send_vf_flr_finish(struct xe_gt *gt, unsigned int vfid) } /** + * DOC: The VF state machine + * + * The simplified VF state machine could be presented as:: + * + * pause--------------------------o + * / | + * / v + * (READY)<------------------resume-----(PAUSED) + * ^ \ / / + * | \ / / + * | stop---->(STOPPED)<----stop / + * | / / + * | / / + * o--------<-----flr / + * \ / + * o------<--------------------flr + * + * Where: + * + * * READY - represents a state in which VF is fully operable + * * PAUSED - represents a state in which VF activity is temporarily suspended + * * STOPPED - represents a state in which VF activity is definitely halted + * * pause - represents a request to temporarily suspend VF activity + * * resume - represents a request to resume VF activity + * * stop - represents a request to definitely halt VF activity + * * flr - represents a request to perform VF FLR to restore VF activity + * + * However, each state transition requires additional steps that involves + * communication with GuC that might fail or be interrupted by other requests:: + * + * .................................WIP.... + * : : + * pause--------------------->PAUSE_WIP----------------------------o + * / : / \ : | + * / : o----<---stop flr--o : | + * / : | \ / | : V + * (READY,RESUMED)<--------+------------RESUME_WIP<----+--<-----resume--(PAUSED) + * ^ \ \ : | | : / / + * | \ \ : | | : / / + * | \ \ : | | : / / + * | \ \ : o----<----------------------+--<-------stop / + * | \ \ : | | : / + * | \ \ : V | : / + * | \ stop----->STOP_WIP---------flr--->-----o : / + * | \ : | | : / + * | \ : | V : / + * | flr--------+----->----------------->FLR_WIP<-----flr + * | : | / ^ : + * | : | / | : + * o--------<-------:----+-----<----------------o | : + * : | | : + * :....|...........................|.....: + * | | + * V | + * (STOPPED)--------------------flr + * + * For details about each internal WIP state machine see: + * + * * `The VF PAUSE state machine`_ + * * `The VF RESUME state machine`_ + * * `The VF STOP state machine`_ + * * `The VF FLR state machine`_ + */ + +#ifdef CONFIG_DRM_XE_DEBUG_SRIOV +static const char *control_bit_to_string(enum xe_gt_sriov_control_bits bit) +{ + switch (bit) { +#define CASE2STR(_X) \ + case XE_GT_SRIOV_STATE_##_X: return #_X + CASE2STR(WIP); + CASE2STR(FLR_WIP); + CASE2STR(FLR_SEND_START); + CASE2STR(FLR_WAIT_GUC); + CASE2STR(FLR_GUC_DONE); + CASE2STR(FLR_RESET_CONFIG); + CASE2STR(FLR_RESET_DATA); + CASE2STR(FLR_RESET_MMIO); + CASE2STR(FLR_SEND_FINISH); + CASE2STR(FLR_FAILED); + CASE2STR(PAUSE_WIP); + CASE2STR(PAUSE_SEND_PAUSE); + CASE2STR(PAUSE_WAIT_GUC); + CASE2STR(PAUSE_GUC_DONE); + CASE2STR(PAUSE_FAILED); + CASE2STR(PAUSED); + CASE2STR(RESUME_WIP); + CASE2STR(RESUME_SEND_RESUME); + CASE2STR(RESUME_FAILED); + CASE2STR(RESUMED); + CASE2STR(STOP_WIP); + CASE2STR(STOP_SEND_STOP); + CASE2STR(STOP_FAILED); + CASE2STR(STOPPED); + CASE2STR(MISMATCH); +#undef CASE2STR + default: return "?"; + } +} +#endif + +static unsigned long pf_get_default_timeout(enum xe_gt_sriov_control_bits bit) +{ + switch (bit) { + case XE_GT_SRIOV_STATE_FLR_WAIT_GUC: + case XE_GT_SRIOV_STATE_PAUSE_WAIT_GUC: + return HZ / 2; + case XE_GT_SRIOV_STATE_FLR_WIP: + case XE_GT_SRIOV_STATE_FLR_RESET_CONFIG: + return 5 * HZ; + default: + return HZ; + } +} + +static struct xe_gt_sriov_control_state *pf_pick_vf_control(struct xe_gt *gt, unsigned int vfid) +{ + xe_gt_assert(gt, IS_SRIOV_PF(gt_to_xe(gt))); + xe_gt_assert(gt, vfid <= xe_gt_sriov_pf_get_totalvfs(gt)); + + return >->sriov.pf.vfs[vfid].control; +} + +static unsigned long *pf_peek_vf_state(struct xe_gt *gt, unsigned int vfid) +{ + struct xe_gt_sriov_control_state *cs = pf_pick_vf_control(gt, vfid); + + return &cs->state; +} + +static bool pf_check_vf_state(struct xe_gt *gt, unsigned int vfid, + enum xe_gt_sriov_control_bits bit) +{ + return test_bit(bit, pf_peek_vf_state(gt, vfid)); +} + +static void pf_dump_vf_state(struct xe_gt *gt, unsigned int vfid) +{ + unsigned long state = *pf_peek_vf_state(gt, vfid); + enum xe_gt_sriov_control_bits bit; + + if (state) { + xe_gt_sriov_dbg_verbose(gt, "VF%u state %#lx%s%*pbl\n", + vfid, state, state ? " bits " : "", + (int)BITS_PER_LONG, &state); + for_each_set_bit(bit, &state, BITS_PER_LONG) + xe_gt_sriov_dbg_verbose(gt, "VF%u state %s(%d)\n", + vfid, control_bit_to_string(bit), bit); + } else { + xe_gt_sriov_dbg_verbose(gt, "VF%u state READY\n", vfid); + } +} + +static bool pf_expect_vf_state(struct xe_gt *gt, unsigned int vfid, + enum xe_gt_sriov_control_bits bit) +{ + bool result = pf_check_vf_state(gt, vfid, bit); + + if (unlikely(!result)) + pf_dump_vf_state(gt, vfid); + + return result; +} + +static bool pf_expect_vf_not_state(struct xe_gt *gt, unsigned int vfid, + enum xe_gt_sriov_control_bits bit) +{ + bool result = !pf_check_vf_state(gt, vfid, bit); + + if (unlikely(!result)) + pf_dump_vf_state(gt, vfid); + + return result; +} + +static bool pf_enter_vf_state(struct xe_gt *gt, unsigned int vfid, + enum xe_gt_sriov_control_bits bit) +{ + if (!test_and_set_bit(bit, pf_peek_vf_state(gt, vfid))) { + xe_gt_sriov_dbg_verbose(gt, "VF%u state %s(%d) enter\n", + vfid, control_bit_to_string(bit), bit); + return true; + } + return false; +} + +static bool pf_exit_vf_state(struct xe_gt *gt, unsigned int vfid, + enum xe_gt_sriov_control_bits bit) +{ + if (test_and_clear_bit(bit, pf_peek_vf_state(gt, vfid))) { + xe_gt_sriov_dbg_verbose(gt, "VF%u state %s(%d) exit\n", + vfid, control_bit_to_string(bit), bit); + return true; + } + return false; +} + +static void pf_escape_vf_state(struct xe_gt *gt, unsigned int vfid, + enum xe_gt_sriov_control_bits bit) +{ + if (pf_exit_vf_state(gt, vfid, bit)) + xe_gt_sriov_dbg_verbose(gt, "VF%u state %s(%d) escaped by %ps\n", + vfid, control_bit_to_string(bit), bit, + __builtin_return_address(0)); +} + +static void pf_enter_vf_mismatch(struct xe_gt *gt, unsigned int vfid) +{ + if (pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_MISMATCH)) { + xe_gt_sriov_dbg(gt, "VF%u state mismatch detected by %ps\n", + vfid, __builtin_return_address(0)); + pf_dump_vf_state(gt, vfid); + } +} + +static void pf_exit_vf_mismatch(struct xe_gt *gt, unsigned int vfid) +{ + if (pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_MISMATCH)) + xe_gt_sriov_dbg(gt, "VF%u state mismatch cleared by %ps\n", + vfid, __builtin_return_address(0)); + + pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_STOP_FAILED); + pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSE_FAILED); + pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_RESUME_FAILED); + pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_FAILED); +} + +#define pf_enter_vf_state_machine_bug(gt, vfid) ({ \ + pf_enter_vf_mismatch((gt), (vfid)); \ +}) + +static void pf_queue_control_worker(struct xe_gt *gt) +{ + struct xe_device *xe = gt_to_xe(gt); + + xe_gt_assert(gt, IS_SRIOV_PF(xe)); + + queue_work(xe->sriov.wq, >->sriov.pf.control.worker); +} + +static void pf_queue_vf(struct xe_gt *gt, unsigned int vfid) +{ + struct xe_gt_sriov_pf_control *pfc = >->sriov.pf.control; + + xe_gt_assert(gt, IS_SRIOV_PF(gt_to_xe(gt))); + + spin_lock(&pfc->lock); + list_move_tail(>->sriov.pf.vfs[vfid].control.link, &pfc->list); + spin_unlock(&pfc->lock); + + pf_queue_control_worker(gt); +} + +static void pf_exit_vf_flr_wip(struct xe_gt *gt, unsigned int vfid); +static void pf_exit_vf_stop_wip(struct xe_gt *gt, unsigned int vfid); +static void pf_exit_vf_pause_wip(struct xe_gt *gt, unsigned int vfid); +static void pf_exit_vf_resume_wip(struct xe_gt *gt, unsigned int vfid); + +static bool pf_enter_vf_wip(struct xe_gt *gt, unsigned int vfid) +{ + if (pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_WIP)) { + struct xe_gt_sriov_control_state *cs = pf_pick_vf_control(gt, vfid); + + reinit_completion(&cs->done); + return true; + } + return false; +} + +static void pf_exit_vf_wip(struct xe_gt *gt, unsigned int vfid) +{ + if (pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_WIP)) { + struct xe_gt_sriov_control_state *cs = pf_pick_vf_control(gt, vfid); + + pf_exit_vf_flr_wip(gt, vfid); + pf_exit_vf_stop_wip(gt, vfid); + pf_exit_vf_pause_wip(gt, vfid); + pf_exit_vf_resume_wip(gt, vfid); + + complete_all(&cs->done); + } +} + +static int pf_wait_vf_wip_done(struct xe_gt *gt, unsigned int vfid, unsigned long timeout) +{ + struct xe_gt_sriov_control_state *cs = pf_pick_vf_control(gt, vfid); + + return wait_for_completion_timeout(&cs->done, timeout) ? 0 : -ETIMEDOUT; +} + +static void pf_enter_vf_ready(struct xe_gt *gt, unsigned int vfid) +{ + pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSED); + pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_STOPPED); + pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_RESUMED); + pf_exit_vf_mismatch(gt, vfid); + pf_exit_vf_wip(gt, vfid); +} + +/** + * DOC: The VF PAUSE state machine + * + * The VF PAUSE state machine looks like:: + * + * (READY,RESUMED)<-------------<---------------------o---------o + * | \ \ + * pause \ \ + * | \ \ + * ....V...........................PAUSE_WIP........ \ \ + * : \ : o \ + * : \ o------<-----busy : | \ + * : \ / / : | | + * : PAUSE_SEND_PAUSE ---failed--->----------o--->(PAUSE_FAILED) | + * : | \ : | | + * : acked rejected---->----------o--->(MISMATCH) / + * : | : / + * : v : / + * : PAUSE_WAIT_GUC : / + * : | : / + * : done : / + * : | : / + * : v : / + * : PAUSE_GUC_DONE o-----restart + * : / : + * : / : + * :....o..............o...............o...........: + * | | | + * completed flr stop + * | | | + * V .....V..... ......V..... + * (PAUSED) : FLR_WIP : : STOP_WIP : + * :.........: :..........: + * + * For the full state machine view, see `The VF state machine`_. + */ + +static void pf_exit_vf_pause_wip(struct xe_gt *gt, unsigned int vfid) +{ + if (pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSE_WIP)) { + pf_escape_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSE_SEND_PAUSE); + pf_escape_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSE_WAIT_GUC); + pf_escape_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSE_GUC_DONE); + } +} + +static void pf_enter_vf_paused(struct xe_gt *gt, unsigned int vfid) +{ + if (!pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSED)) + pf_enter_vf_state_machine_bug(gt, vfid); + + pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_RESUMED); + pf_exit_vf_mismatch(gt, vfid); + pf_exit_vf_wip(gt, vfid); +} + +static void pf_enter_vf_pause_completed(struct xe_gt *gt, unsigned int vfid) +{ + pf_enter_vf_paused(gt, vfid); +} + +static void pf_enter_vf_pause_failed(struct xe_gt *gt, unsigned int vfid) +{ + pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSE_FAILED); + pf_exit_vf_wip(gt, vfid); +} + +static void pf_enter_vf_pause_rejected(struct xe_gt *gt, unsigned int vfid) +{ + pf_enter_vf_mismatch(gt, vfid); + pf_enter_vf_pause_failed(gt, vfid); +} + +static bool pf_exit_vf_pause_guc_done(struct xe_gt *gt, unsigned int vfid) +{ + if (!pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSE_GUC_DONE)) + return false; + + pf_enter_vf_pause_completed(gt, vfid); + return true; +} + +static void pf_enter_vf_pause_guc_done(struct xe_gt *gt, unsigned int vfid) +{ + if (pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSE_GUC_DONE)) + pf_queue_vf(gt, vfid); +} + +static void pf_enter_pause_wait_guc(struct xe_gt *gt, unsigned int vfid) +{ + if (!pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSE_WAIT_GUC)) + pf_enter_vf_state_machine_bug(gt, vfid); +} + +static bool pf_exit_pause_wait_guc(struct xe_gt *gt, unsigned int vfid) +{ + return pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSE_WAIT_GUC); +} + +static void pf_enter_vf_pause_send_pause(struct xe_gt *gt, unsigned int vfid) +{ + if (!pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSE_SEND_PAUSE)) + pf_enter_vf_state_machine_bug(gt, vfid); + + pf_queue_vf(gt, vfid); +} + +static bool pf_exit_vf_pause_send_pause(struct xe_gt *gt, unsigned int vfid) +{ + int err; + + if (!pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSE_SEND_PAUSE)) + return false; + + /* GuC may actually send a PAUSE_DONE before we get a RESPONSE */ + pf_enter_pause_wait_guc(gt, vfid); + + err = pf_send_vf_pause(gt, vfid); + if (err) { + /* send failed, so we shouldn't expect PAUSE_DONE from GuC */ + pf_exit_pause_wait_guc(gt, vfid); + + if (err == -EBUSY) + pf_enter_vf_pause_send_pause(gt, vfid); + else if (err == -EIO) + pf_enter_vf_pause_rejected(gt, vfid); + else + pf_enter_vf_pause_failed(gt, vfid); + } else { + /* + * we have already moved to WAIT_GUC, maybe even to GUC_DONE + * but since GuC didn't complain, we may clear MISMATCH + */ + pf_exit_vf_mismatch(gt, vfid); + } + + return true; +} + +static bool pf_enter_vf_pause_wip(struct xe_gt *gt, unsigned int vfid) +{ + if (pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSE_WIP)) { + pf_enter_vf_wip(gt, vfid); + pf_enter_vf_pause_send_pause(gt, vfid); + return true; + } + + return false; +} + +/** * xe_gt_sriov_pf_control_pause_vf - Pause a VF. * @gt: the &xe_gt * @vfid: the VF identifier @@ -98,7 +552,140 @@ static int pf_send_vf_flr_finish(struct xe_gt *gt, unsigned int vfid) */ int xe_gt_sriov_pf_control_pause_vf(struct xe_gt *gt, unsigned int vfid) { - return pf_send_vf_pause(gt, vfid); + unsigned long timeout = pf_get_default_timeout(XE_GT_SRIOV_STATE_PAUSE_WIP); + int err; + + if (pf_check_vf_state(gt, vfid, XE_GT_SRIOV_STATE_STOPPED)) { + xe_gt_sriov_dbg(gt, "VF%u is stopped!\n", vfid); + return -EPERM; + } + + if (pf_check_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSED)) { + xe_gt_sriov_dbg(gt, "VF%u was already paused!\n", vfid); + return -ESTALE; + } + + if (!pf_enter_vf_pause_wip(gt, vfid)) { + xe_gt_sriov_dbg(gt, "VF%u pause already in progress!\n", vfid); + return -EALREADY; + } + + err = pf_wait_vf_wip_done(gt, vfid, timeout); + if (err) { + xe_gt_sriov_dbg(gt, "VF%u pause didn't finish in %u ms (%pe)\n", + vfid, jiffies_to_msecs(timeout), ERR_PTR(err)); + return err; + } + + if (pf_expect_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSED)) { + xe_gt_sriov_info(gt, "VF%u paused!\n", vfid); + return 0; + } + + if (pf_check_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSE_FAILED)) { + xe_gt_sriov_dbg(gt, "VF%u pause failed!\n", vfid); + return -EIO; + } + + xe_gt_sriov_dbg(gt, "VF%u pause was canceled!\n", vfid); + return -ECANCELED; +} + +/** + * DOC: The VF RESUME state machine + * + * The VF RESUME state machine looks like:: + * + * (PAUSED)<-----------------<------------------------o + * | \ + * resume \ + * | \ + * ....V............................RESUME_WIP...... \ + * : \ : o + * : \ o-------<-----busy : | + * : \ / / : | + * : RESUME_SEND_RESUME ---failed--->--------o--->(RESUME_FAILED) + * : / \ : | + * : acked rejected---->---------o--->(MISMATCH) + * : / : + * :....o..............o...............o.....o.....: + * | | | \ + * completed flr stop restart-->(READY) + * | | | + * V .....V..... ......V..... + * (RESUMED) : FLR_WIP : : STOP_WIP : + * :.........: :..........: + * + * For the full state machine view, see `The VF state machine`_. + */ + +static void pf_exit_vf_resume_wip(struct xe_gt *gt, unsigned int vfid) +{ + if (pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_RESUME_WIP)) + pf_escape_vf_state(gt, vfid, XE_GT_SRIOV_STATE_RESUME_SEND_RESUME); +} + +static void pf_enter_vf_resumed(struct xe_gt *gt, unsigned int vfid) +{ + pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_RESUMED); + pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSED); + pf_exit_vf_mismatch(gt, vfid); + pf_exit_vf_wip(gt, vfid); +} + +static void pf_enter_vf_resume_completed(struct xe_gt *gt, unsigned int vfid) +{ + pf_enter_vf_resumed(gt, vfid); +} + +static void pf_enter_vf_resume_failed(struct xe_gt *gt, unsigned int vfid) +{ + pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_RESUME_FAILED); + pf_exit_vf_wip(gt, vfid); +} + +static void pf_enter_vf_resume_rejected(struct xe_gt *gt, unsigned int vfid) +{ + pf_enter_vf_mismatch(gt, vfid); + pf_enter_vf_resume_failed(gt, vfid); +} + +static void pf_enter_vf_resume_send_resume(struct xe_gt *gt, unsigned int vfid) +{ + if (!pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_RESUME_SEND_RESUME)) + pf_enter_vf_state_machine_bug(gt, vfid); + + pf_queue_vf(gt, vfid); +} + +static bool pf_exit_vf_resume_send_resume(struct xe_gt *gt, unsigned int vfid) +{ + int err; + + if (!pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_RESUME_SEND_RESUME)) + return false; + + err = pf_send_vf_resume(gt, vfid); + if (err == -EBUSY) + pf_enter_vf_resume_send_resume(gt, vfid); + else if (err == -EIO) + pf_enter_vf_resume_rejected(gt, vfid); + else if (err) + pf_enter_vf_resume_failed(gt, vfid); + else + pf_enter_vf_resume_completed(gt, vfid); + return true; +} + +static bool pf_enter_vf_resume_wip(struct xe_gt *gt, unsigned int vfid) +{ + if (pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_RESUME_WIP)) { + pf_enter_vf_wip(gt, vfid); + pf_enter_vf_resume_send_resume(gt, vfid); + return true; + } + + return false; } /** @@ -112,7 +699,134 @@ int xe_gt_sriov_pf_control_pause_vf(struct xe_gt *gt, unsigned int vfid) */ int xe_gt_sriov_pf_control_resume_vf(struct xe_gt *gt, unsigned int vfid) { - return pf_send_vf_resume(gt, vfid); + unsigned long timeout = pf_get_default_timeout(XE_GT_SRIOV_STATE_RESUME_WIP); + int err; + + if (!pf_check_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSED)) { + xe_gt_sriov_dbg(gt, "VF%u is not paused!\n", vfid); + return -EPERM; + } + + if (!pf_enter_vf_resume_wip(gt, vfid)) { + xe_gt_sriov_dbg(gt, "VF%u resume already in progress!\n", vfid); + return -EALREADY; + } + + err = pf_wait_vf_wip_done(gt, vfid, timeout); + if (err) + return err; + + if (pf_expect_vf_state(gt, vfid, XE_GT_SRIOV_STATE_RESUMED)) { + xe_gt_sriov_info(gt, "VF%u resumed!\n", vfid); + return 0; + } + + if (pf_check_vf_state(gt, vfid, XE_GT_SRIOV_STATE_RESUME_FAILED)) { + xe_gt_sriov_dbg(gt, "VF%u resume failed!\n", vfid); + return -EIO; + } + + xe_gt_sriov_dbg(gt, "VF%u resume was canceled!\n", vfid); + return -ECANCELED; +} + +/** + * DOC: The VF STOP state machine + * + * The VF STOP state machine looks like:: + * + * (READY,PAUSED,RESUMED)<-------<--------------------o + * | \ + * stop \ + * | \ + * ....V..............................STOP_WIP...... \ + * : \ : o + * : \ o----<----busy : | + * : \ / / : | + * : STOP_SEND_STOP--------failed--->--------o--->(STOP_FAILED) + * : / \ : | + * : acked rejected-------->--------o--->(MISMATCH) + * : / : + * :....o..............o...............o...........: + * | | | + * completed flr restart + * | | | + * V .....V..... V + * (STOPPED) : FLR_WIP : (READY) + * :.........: + * + * For the full state machine view, see `The VF state machine`_. + */ + +static void pf_exit_vf_stop_wip(struct xe_gt *gt, unsigned int vfid) +{ + if (pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_STOP_WIP)) + pf_escape_vf_state(gt, vfid, XE_GT_SRIOV_STATE_STOP_SEND_STOP); +} + +static void pf_enter_vf_stopped(struct xe_gt *gt, unsigned int vfid) +{ + if (!pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_STOPPED)) + pf_enter_vf_state_machine_bug(gt, vfid); + + pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_RESUMED); + pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSED); + pf_exit_vf_mismatch(gt, vfid); + pf_exit_vf_wip(gt, vfid); +} + +static void pf_enter_vf_stop_completed(struct xe_gt *gt, unsigned int vfid) +{ + pf_enter_vf_stopped(gt, vfid); +} + +static void pf_enter_vf_stop_failed(struct xe_gt *gt, unsigned int vfid) +{ + pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_STOP_FAILED); + pf_exit_vf_wip(gt, vfid); +} + +static void pf_enter_vf_stop_rejected(struct xe_gt *gt, unsigned int vfid) +{ + pf_enter_vf_mismatch(gt, vfid); + pf_enter_vf_stop_failed(gt, vfid); +} + +static void pf_enter_vf_stop_send_stop(struct xe_gt *gt, unsigned int vfid) +{ + if (!pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_STOP_SEND_STOP)) + pf_enter_vf_state_machine_bug(gt, vfid); + + pf_queue_vf(gt, vfid); +} + +static bool pf_exit_vf_stop_send_stop(struct xe_gt *gt, unsigned int vfid) +{ + int err; + + if (!pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_STOP_SEND_STOP)) + return false; + + err = pf_send_vf_stop(gt, vfid); + if (err == -EBUSY) + pf_enter_vf_stop_send_stop(gt, vfid); + else if (err == -EIO) + pf_enter_vf_stop_rejected(gt, vfid); + else if (err) + pf_enter_vf_stop_failed(gt, vfid); + else + pf_enter_vf_stop_completed(gt, vfid); + return true; +} + +static bool pf_enter_vf_stop_wip(struct xe_gt *gt, unsigned int vfid) +{ + if (pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_STOP_WIP)) { + pf_enter_vf_wip(gt, vfid); + pf_enter_vf_stop_send_stop(gt, vfid); + return true; + } + return false; } /** @@ -126,7 +840,280 @@ int xe_gt_sriov_pf_control_resume_vf(struct xe_gt *gt, unsigned int vfid) */ int xe_gt_sriov_pf_control_stop_vf(struct xe_gt *gt, unsigned int vfid) { - return pf_send_vf_stop(gt, vfid); + unsigned long timeout = pf_get_default_timeout(XE_GT_SRIOV_STATE_STOP_WIP); + int err; + + if (pf_check_vf_state(gt, vfid, XE_GT_SRIOV_STATE_STOPPED)) { + xe_gt_sriov_dbg(gt, "VF%u was already stopped!\n", vfid); + return -ESTALE; + } + + if (!pf_enter_vf_stop_wip(gt, vfid)) { + xe_gt_sriov_dbg(gt, "VF%u stop already in progress!\n", vfid); + return -EALREADY; + } + + err = pf_wait_vf_wip_done(gt, vfid, timeout); + if (err) + return err; + + if (pf_expect_vf_state(gt, vfid, XE_GT_SRIOV_STATE_STOPPED)) { + xe_gt_sriov_info(gt, "VF%u stopped!\n", vfid); + return 0; + } + + if (pf_check_vf_state(gt, vfid, XE_GT_SRIOV_STATE_STOP_FAILED)) { + xe_gt_sriov_dbg(gt, "VF%u stop failed!\n", vfid); + return -EIO; + } + + xe_gt_sriov_dbg(gt, "VF%u stop was canceled!\n", vfid); + return -ECANCELED; +} + +/** + * DOC: The VF FLR state machine + * + * The VF FLR state machine looks like:: + * + * (READY,PAUSED,STOPPED)<------------<--------------o + * | \ + * flr \ + * | \ + * ....V..........................FLR_WIP........... \ + * : \ : \ + * : \ o----<----busy : | + * : \ / / : | + * : FLR_SEND_START---failed----->-----------o--->(FLR_FAILED)<---o + * : | \ : | | + * : acked rejected----->-----------o--->(MISMATCH) | + * : | : ^ | + * : v : | | + * : FLR_WAIT_GUC : | | + * : | : | | + * : done : | | + * : | : | | + * : v : | | + * : FLR_GUC_DONE : | | + * : | : | | + * : FLR_RESET_CONFIG---failed--->-----------o--------+-----------o + * : | : | | + * : FLR_RESET_DATA : | | + * : | : | | + * : FLR_RESET_MMIO : | | + * : | : | | + * : | o----<----busy : | | + * : |/ / : | | + * : FLR_SEND_FINISH----failed--->-----------o--------+-----------o + * : / \ : | + * : acked rejected----->-----------o--------o + * : / : + * :....o..............................o...........: + * | | + * completed restart + * | / + * V / + * (READY)<----------<------------o + * + * For the full state machine view, see `The VF state machine`_. + */ + +static void pf_enter_vf_flr_send_start(struct xe_gt *gt, unsigned int vfid) +{ + if (!pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_SEND_START)) + pf_enter_vf_state_machine_bug(gt, vfid); + + pf_queue_vf(gt, vfid); +} + +static void pf_enter_vf_flr_wip(struct xe_gt *gt, unsigned int vfid) +{ + if (!pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_WIP)) { + xe_gt_sriov_dbg(gt, "VF%u FLR is already in progress\n", vfid); + return; + } + + pf_enter_vf_wip(gt, vfid); + pf_enter_vf_flr_send_start(gt, vfid); +} + +static void pf_exit_vf_flr_wip(struct xe_gt *gt, unsigned int vfid) +{ + if (pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_WIP)) { + pf_escape_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_SEND_FINISH); + pf_escape_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_RESET_MMIO); + pf_escape_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_RESET_DATA); + pf_escape_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_RESET_CONFIG); + pf_escape_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_GUC_DONE); + pf_escape_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_WAIT_GUC); + pf_escape_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_SEND_START); + } +} + +static void pf_enter_vf_flr_completed(struct xe_gt *gt, unsigned int vfid) +{ + pf_enter_vf_ready(gt, vfid); +} + +static void pf_enter_vf_flr_failed(struct xe_gt *gt, unsigned int vfid) +{ + if (pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_FAILED)) + xe_gt_sriov_notice(gt, "VF%u FLR failed!\n", vfid); + pf_exit_vf_wip(gt, vfid); +} + +static void pf_enter_vf_flr_rejected(struct xe_gt *gt, unsigned int vfid) +{ + pf_enter_vf_mismatch(gt, vfid); + pf_enter_vf_flr_failed(gt, vfid); +} + +static void pf_enter_vf_flr_send_finish(struct xe_gt *gt, unsigned int vfid) +{ + if (!pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_SEND_FINISH)) + pf_enter_vf_state_machine_bug(gt, vfid); + + pf_queue_vf(gt, vfid); +} + +static bool pf_exit_vf_flr_send_finish(struct xe_gt *gt, unsigned int vfid) +{ + int err; + + if (!pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_SEND_FINISH)) + return false; + + err = pf_send_vf_flr_finish(gt, vfid); + if (err == -EBUSY) + pf_enter_vf_flr_send_finish(gt, vfid); + else if (err == -EIO) + pf_enter_vf_flr_rejected(gt, vfid); + else if (err) + pf_enter_vf_flr_failed(gt, vfid); + else + pf_enter_vf_flr_completed(gt, vfid); + return true; +} + +static void pf_enter_vf_flr_reset_mmio(struct xe_gt *gt, unsigned int vfid) +{ + if (!pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_RESET_MMIO)) + pf_enter_vf_state_machine_bug(gt, vfid); + + pf_queue_vf(gt, vfid); +} + +static bool pf_exit_vf_flr_reset_mmio(struct xe_gt *gt, unsigned int vfid) +{ + if (!pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_RESET_MMIO)) + return false; + + /* XXX: placeholder */ + + pf_enter_vf_flr_send_finish(gt, vfid); + return true; +} + +static void pf_enter_vf_flr_reset_data(struct xe_gt *gt, unsigned int vfid) +{ + if (!pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_RESET_DATA)) + pf_enter_vf_state_machine_bug(gt, vfid); + + pf_queue_vf(gt, vfid); +} + +static bool pf_exit_vf_flr_reset_data(struct xe_gt *gt, unsigned int vfid) +{ + if (!pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_RESET_DATA)) + return false; + + xe_gt_sriov_pf_service_reset(gt, vfid); + xe_gt_sriov_pf_monitor_flr(gt, vfid); + + pf_enter_vf_flr_reset_mmio(gt, vfid); + return true; +} + +static void pf_enter_vf_flr_reset_config(struct xe_gt *gt, unsigned int vfid) +{ + if (!pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_RESET_CONFIG)) + pf_enter_vf_state_machine_bug(gt, vfid); + + pf_queue_vf(gt, vfid); +} + +static bool pf_exit_vf_flr_reset_config(struct xe_gt *gt, unsigned int vfid) +{ + unsigned long timeout = pf_get_default_timeout(XE_GT_SRIOV_STATE_FLR_RESET_CONFIG); + int err; + + if (!pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_RESET_CONFIG)) + return false; + + err = xe_gt_sriov_pf_config_sanitize(gt, vfid, timeout); + if (err) + pf_enter_vf_flr_failed(gt, vfid); + else + pf_enter_vf_flr_reset_data(gt, vfid); + return true; +} + +static void pf_enter_vf_flr_wait_guc(struct xe_gt *gt, unsigned int vfid) +{ + if (!pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_WAIT_GUC)) + pf_enter_vf_state_machine_bug(gt, vfid); +} + +static bool pf_exit_vf_flr_wait_guc(struct xe_gt *gt, unsigned int vfid) +{ + return pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_WAIT_GUC); +} + +static bool pf_exit_vf_flr_send_start(struct xe_gt *gt, unsigned int vfid) +{ + int err; + + if (!pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_SEND_START)) + return false; + + /* GuC may actually send a FLR_DONE before we get a RESPONSE */ + pf_enter_vf_flr_wait_guc(gt, vfid); + + err = pf_send_vf_flr_start(gt, vfid); + if (err) { + /* send failed, so we shouldn't expect FLR_DONE from GuC */ + pf_exit_vf_flr_wait_guc(gt, vfid); + + if (err == -EBUSY) + pf_enter_vf_flr_send_start(gt, vfid); + else if (err == -EIO) + pf_enter_vf_flr_rejected(gt, vfid); + else + pf_enter_vf_flr_failed(gt, vfid); + } else { + /* + * we have already moved to WAIT_GUC, maybe even to GUC_DONE + * but since GuC didn't complain, we may clear MISMATCH + */ + pf_exit_vf_mismatch(gt, vfid); + } + + return true; +} + +static bool pf_exit_vf_flr_guc_done(struct xe_gt *gt, unsigned int vfid) +{ + if (!pf_exit_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_GUC_DONE)) + return false; + + pf_enter_vf_flr_reset_config(gt, vfid); + return true; +} + +static void pf_enter_vf_flr_guc_done(struct xe_gt *gt, unsigned int vfid) +{ + if (pf_enter_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_GUC_DONE)) + pf_queue_vf(gt, vfid); } /** @@ -140,46 +1127,56 @@ int xe_gt_sriov_pf_control_stop_vf(struct xe_gt *gt, unsigned int vfid) */ int xe_gt_sriov_pf_control_trigger_flr(struct xe_gt *gt, unsigned int vfid) { + unsigned long timeout = pf_get_default_timeout(XE_GT_SRIOV_STATE_FLR_WIP); int err; - /* XXX pf_send_vf_flr_start() expects ct->lock */ - mutex_lock(>->uc.guc.ct.lock); - err = pf_send_vf_flr_start(gt, vfid); - mutex_unlock(>->uc.guc.ct.lock); + pf_enter_vf_flr_wip(gt, vfid); - return err; + err = pf_wait_vf_wip_done(gt, vfid, timeout); + if (err) { + xe_gt_sriov_notice(gt, "VF%u FLR didn't finish in %u ms (%pe)\n", + vfid, jiffies_to_msecs(timeout), ERR_PTR(err)); + return err; + } + + if (!pf_expect_vf_not_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_FAILED)) + return -EIO; + + return 0; } /** * DOC: The VF FLR Flow with GuC * - * PF GUC PCI - * ======================================================== - * | | | - * (1) | [ ] <----- FLR --| - * | [ ] : - * (2) [ ] <-------- NOTIFY FLR --[ ] - * [ ] | - * (3) [ ] | - * [ ] | - * [ ]-- START FLR ---------> [ ] - * | [ ] - * (4) | [ ] - * | [ ] - * [ ] <--------- FLR DONE -- [ ] - * [ ] | - * (5) [ ] | - * [ ] | - * [ ]-- FINISH FLR --------> [ ] - * | | - * - * Step 1: PCI HW generates interrupt to the GuC about VF FLR - * Step 2: GuC FW sends G2H notification to the PF about VF FLR - * Step 2a: on some platforms G2H is only received from root GuC - * Step 3: PF sends H2G request to the GuC to start VF FLR sequence - * Step 3a: on some platforms PF must send H2G to all other GuCs - * Step 4: GuC FW performs VF FLR cleanups and notifies the PF when done - * Step 5: PF performs VF FLR cleanups and notifies the GuC FW when finished + * The VF FLR flow includes several steps:: + * + * PF GUC PCI + * ======================================================== + * | | | + * (1) | [ ] <----- FLR --| + * | [ ] : + * (2) [ ] <-------- NOTIFY FLR --[ ] + * [ ] | + * (3) [ ] | + * [ ] | + * [ ]-- START FLR ---------> [ ] + * | [ ] + * (4) | [ ] + * | [ ] + * [ ] <--------- FLR DONE -- [ ] + * [ ] | + * (5) [ ] | + * [ ] | + * [ ]-- FINISH FLR --------> [ ] + * | | + * + * * Step 1: PCI HW generates interrupt to the GuC about VF FLR + * * Step 2: GuC FW sends G2H notification to the PF about VF FLR + * * Step 2a: on some platforms G2H is only received from root GuC + * * Step 3: PF sends H2G request to the GuC to start VF FLR sequence + * * Step 3a: on some platforms PF must send H2G to all other GuCs + * * Step 4: GuC FW performs VF FLR cleanups and notifies the PF when done + * * Step 5: PF performs VF FLR cleanups and notifies the GuC FW when finished */ static bool needs_dispatch_flr(struct xe_device *xe) @@ -197,19 +1194,41 @@ static void pf_handle_vf_flr(struct xe_gt *gt, u32 vfid) if (needs_dispatch_flr(xe)) { for_each_gt(gtit, xe, gtid) - pf_send_vf_flr_start(gtit, vfid); + pf_enter_vf_flr_wip(gtit, vfid); } else { - pf_send_vf_flr_start(gt, vfid); + pf_enter_vf_flr_wip(gt, vfid); } } static void pf_handle_vf_flr_done(struct xe_gt *gt, u32 vfid) { - pf_send_vf_flr_finish(gt, vfid); + if (!pf_exit_vf_flr_wait_guc(gt, vfid)) { + xe_gt_sriov_dbg(gt, "Received out of order 'VF%u FLR done'\n", vfid); + pf_enter_vf_mismatch(gt, vfid); + return; + } + + pf_enter_vf_flr_guc_done(gt, vfid); +} + +static void pf_handle_vf_pause_done(struct xe_gt *gt, u32 vfid) +{ + if (!pf_exit_pause_wait_guc(gt, vfid)) { + xe_gt_sriov_dbg(gt, "Received out of order 'VF%u PAUSE done'\n", vfid); + pf_enter_vf_mismatch(gt, vfid); + return; + } + + pf_enter_vf_pause_guc_done(gt, vfid); } static int pf_handle_vf_event(struct xe_gt *gt, u32 vfid, u32 eventid) { + xe_gt_sriov_dbg_verbose(gt, "received VF%u event %#x\n", vfid, eventid); + + if (vfid > xe_gt_sriov_pf_get_totalvfs(gt)) + return -EPROTO; + switch (eventid) { case GUC_PF_NOTIFY_VF_FLR: pf_handle_vf_flr(gt, vfid); @@ -218,6 +1237,7 @@ static int pf_handle_vf_event(struct xe_gt *gt, u32 vfid, u32 eventid) pf_handle_vf_flr_done(gt, vfid); break; case GUC_PF_NOTIFY_VF_PAUSE_DONE: + pf_handle_vf_pause_done(gt, vfid); break; case GUC_PF_NOTIFY_VF_FIXUP_DONE: break; @@ -276,3 +1296,159 @@ int xe_gt_sriov_pf_control_process_guc2pf(struct xe_gt *gt, const u32 *msg, u32 return vfid ? pf_handle_vf_event(gt, vfid, eventid) : pf_handle_pf_event(gt, eventid); } + +static bool pf_process_vf_state_machine(struct xe_gt *gt, unsigned int vfid) +{ + if (pf_exit_vf_flr_send_start(gt, vfid)) + return true; + + if (pf_check_vf_state(gt, vfid, XE_GT_SRIOV_STATE_FLR_WAIT_GUC)) { + xe_gt_sriov_dbg_verbose(gt, "VF%u in %s\n", vfid, + control_bit_to_string(XE_GT_SRIOV_STATE_FLR_WAIT_GUC)); + return false; + } + + if (pf_exit_vf_flr_guc_done(gt, vfid)) + return true; + + if (pf_exit_vf_flr_reset_config(gt, vfid)) + return true; + + if (pf_exit_vf_flr_reset_data(gt, vfid)) + return true; + + if (pf_exit_vf_flr_reset_mmio(gt, vfid)) + return true; + + if (pf_exit_vf_flr_send_finish(gt, vfid)) + return true; + + if (pf_exit_vf_stop_send_stop(gt, vfid)) + return true; + + if (pf_exit_vf_pause_send_pause(gt, vfid)) + return true; + + if (pf_check_vf_state(gt, vfid, XE_GT_SRIOV_STATE_PAUSE_WAIT_GUC)) { + xe_gt_sriov_dbg_verbose(gt, "VF%u in %s\n", vfid, + control_bit_to_string(XE_GT_SRIOV_STATE_PAUSE_WAIT_GUC)); + return true; + } + + if (pf_exit_vf_pause_guc_done(gt, vfid)) + return true; + + if (pf_exit_vf_resume_send_resume(gt, vfid)) + return true; + + return false; +} + +static unsigned int pf_control_state_index(struct xe_gt *gt, + struct xe_gt_sriov_control_state *cs) +{ + return container_of(cs, struct xe_gt_sriov_metadata, control) - gt->sriov.pf.vfs; +} + +static void pf_worker_find_work(struct xe_gt *gt) +{ + struct xe_gt_sriov_pf_control *pfc = >->sriov.pf.control; + struct xe_gt_sriov_control_state *cs; + unsigned int vfid; + bool empty; + bool more; + + spin_lock(&pfc->lock); + cs = list_first_entry_or_null(&pfc->list, struct xe_gt_sriov_control_state, link); + if (cs) + list_del_init(&cs->link); + empty = list_empty(&pfc->list); + spin_unlock(&pfc->lock); + + if (!cs) + return; + + /* VF metadata structures are indexed by the VFID */ + vfid = pf_control_state_index(gt, cs); + xe_gt_assert(gt, vfid <= xe_gt_sriov_pf_get_totalvfs(gt)); + + more = pf_process_vf_state_machine(gt, vfid); + if (more) + pf_queue_vf(gt, vfid); + else if (!empty) + pf_queue_control_worker(gt); +} + +static void control_worker_func(struct work_struct *w) +{ + struct xe_gt *gt = container_of(w, struct xe_gt, sriov.pf.control.worker); + + xe_gt_assert(gt, IS_SRIOV_PF(gt_to_xe(gt))); + pf_worker_find_work(gt); +} + +static void pf_stop_worker(struct xe_gt *gt) +{ + xe_gt_assert(gt, IS_SRIOV_PF(gt_to_xe(gt))); + cancel_work_sync(>->sriov.pf.control.worker); +} + +static void control_fini_action(struct drm_device *dev, void *data) +{ + struct xe_gt *gt = data; + + pf_stop_worker(gt); +} + +/** + * xe_gt_sriov_pf_control_init() - Initialize PF's control data. + * @gt: the &xe_gt + * + * This function is for PF only. + * + * Return: 0 on success or a negative error code on failure. + */ +int xe_gt_sriov_pf_control_init(struct xe_gt *gt) +{ + struct xe_device *xe = gt_to_xe(gt); + unsigned int n, totalvfs; + + xe_gt_assert(gt, IS_SRIOV_PF(xe)); + + totalvfs = xe_sriov_pf_get_totalvfs(xe); + for (n = 0; n <= totalvfs; n++) { + struct xe_gt_sriov_control_state *cs = pf_pick_vf_control(gt, n); + + init_completion(&cs->done); + INIT_LIST_HEAD(&cs->link); + } + + spin_lock_init(>->sriov.pf.control.lock); + INIT_LIST_HEAD(>->sriov.pf.control.list); + INIT_WORK(>->sriov.pf.control.worker, control_worker_func); + + return drmm_add_action_or_reset(&xe->drm, control_fini_action, gt); +} + +/** + * xe_gt_sriov_pf_control_restart() - Restart SR-IOV control data after a GT reset. + * @gt: the &xe_gt + * + * Any per-VF status maintained by the PF or any ongoing VF control activity + * performed by the PF must be reset or cancelled when the GT is reset. + * + * This function is for PF only. + */ +void xe_gt_sriov_pf_control_restart(struct xe_gt *gt) +{ + struct xe_device *xe = gt_to_xe(gt); + unsigned int n, totalvfs; + + xe_gt_assert(gt, IS_SRIOV_PF(xe)); + + pf_stop_worker(gt); + + totalvfs = xe_sriov_pf_get_totalvfs(xe); + for (n = 1; n <= totalvfs; n++) + pf_enter_vf_ready(gt, n); +} diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_pf_control.h b/drivers/gpu/drm/xe/xe_gt_sriov_pf_control.h index 405d1586f991..c85e64f099cc 100644 --- a/drivers/gpu/drm/xe/xe_gt_sriov_pf_control.h +++ b/drivers/gpu/drm/xe/xe_gt_sriov_pf_control.h @@ -11,6 +11,9 @@ struct xe_gt; +int xe_gt_sriov_pf_control_init(struct xe_gt *gt); +void xe_gt_sriov_pf_control_restart(struct xe_gt *gt); + int xe_gt_sriov_pf_control_pause_vf(struct xe_gt *gt, unsigned int vfid); int xe_gt_sriov_pf_control_resume_vf(struct xe_gt *gt, unsigned int vfid); int xe_gt_sriov_pf_control_stop_vf(struct xe_gt *gt, unsigned int vfid); diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_pf_control_types.h b/drivers/gpu/drm/xe/xe_gt_sriov_pf_control_types.h new file mode 100644 index 000000000000..11830aafea45 --- /dev/null +++ b/drivers/gpu/drm/xe/xe_gt_sriov_pf_control_types.h @@ -0,0 +1,107 @@ +/* SPDX-License-Identifier: MIT */ +/* + * Copyright © 2024 Intel Corporation + */ + +#ifndef _XE_GT_SRIOV_PF_CONTROL_TYPES_H_ +#define _XE_GT_SRIOV_PF_CONTROL_TYPES_H_ + +#include <linux/completion.h> +#include <linux/spinlock.h> +#include <linux/workqueue_types.h> + +/** + * enum xe_gt_sriov_control_bits - Various bits used by the PF to represent a VF state + * + * @XE_GT_SRIOV_STATE_WIP: indicates that some operations are in progress. + * @XE_GT_SRIOV_STATE_FLR_WIP: indicates that a VF FLR is in progress. + * @XE_GT_SRIOV_STATE_FLR_SEND_START: indicates that the PF wants to send a FLR START command. + * @XE_GT_SRIOV_STATE_FLR_WAIT_GUC: indicates that the PF awaits for a response from the GuC. + * @XE_GT_SRIOV_STATE_FLR_GUC_DONE: indicates that the PF has received a response from the GuC. + * @XE_GT_SRIOV_STATE_FLR_RESET_CONFIG: indicates that the PF needs to clear VF's resources. + * @XE_GT_SRIOV_STATE_FLR_RESET_DATA: indicates that the PF needs to clear VF's data. + * @XE_GT_SRIOV_STATE_FLR_RESET_MMIO: indicates that the PF needs to reset VF's registers. + * @XE_GT_SRIOV_STATE_FLR_SEND_FINISH: indicates that the PF wants to send a FLR FINISH message. + * @XE_GT_SRIOV_STATE_FLR_FAILED: indicates that VF FLR sequence failed. + * @XE_GT_SRIOV_STATE_PAUSE_WIP: indicates that a VF pause operation is in progress. + * @XE_GT_SRIOV_STATE_PAUSE_SEND_PAUSE: indicates that the PF is about to send a PAUSE command. + * @XE_GT_SRIOV_STATE_PAUSE_WAIT_GUC: indicates that the PF awaits for a response from the GuC. + * @XE_GT_SRIOV_STATE_PAUSE_GUC_DONE: indicates that the PF has received a response from the GuC. + * @XE_GT_SRIOV_STATE_PAUSE_FAILED: indicates that a VF pause operation has failed. + * @XE_GT_SRIOV_STATE_PAUSED: indicates that the VF is paused. + * @XE_GT_SRIOV_STATE_RESUME_WIP: indicates the a VF resume operation is in progress. + * @XE_GT_SRIOV_STATE_RESUME_SEND_RESUME: indicates that the PF is about to send RESUME command. + * @XE_GT_SRIOV_STATE_RESUME_FAILED: indicates that a VF resume operation has failed. + * @XE_GT_SRIOV_STATE_RESUMED: indicates that the VF was resumed. + * @XE_GT_SRIOV_STATE_STOP_WIP: indicates that a VF stop operation is in progress. + * @XE_GT_SRIOV_STATE_STOP_SEND_STOP: indicates that the PF wants to send a STOP command. + * @XE_GT_SRIOV_STATE_STOP_FAILED: indicates that the VF stop operation has failed + * @XE_GT_SRIOV_STATE_STOPPED: indicates that the VF was stopped. + * @XE_GT_SRIOV_STATE_MISMATCH: indicates that the PF has detected a VF state mismatch. + */ +enum xe_gt_sriov_control_bits { + XE_GT_SRIOV_STATE_WIP = 1, + + XE_GT_SRIOV_STATE_FLR_WIP, + XE_GT_SRIOV_STATE_FLR_SEND_START, + XE_GT_SRIOV_STATE_FLR_WAIT_GUC, + XE_GT_SRIOV_STATE_FLR_GUC_DONE, + XE_GT_SRIOV_STATE_FLR_RESET_CONFIG, + XE_GT_SRIOV_STATE_FLR_RESET_DATA, + XE_GT_SRIOV_STATE_FLR_RESET_MMIO, + XE_GT_SRIOV_STATE_FLR_SEND_FINISH, + XE_GT_SRIOV_STATE_FLR_FAILED, + + XE_GT_SRIOV_STATE_PAUSE_WIP, + XE_GT_SRIOV_STATE_PAUSE_SEND_PAUSE, + XE_GT_SRIOV_STATE_PAUSE_WAIT_GUC, + XE_GT_SRIOV_STATE_PAUSE_GUC_DONE, + XE_GT_SRIOV_STATE_PAUSE_FAILED, + XE_GT_SRIOV_STATE_PAUSED, + + XE_GT_SRIOV_STATE_RESUME_WIP, + XE_GT_SRIOV_STATE_RESUME_SEND_RESUME, + XE_GT_SRIOV_STATE_RESUME_FAILED, + XE_GT_SRIOV_STATE_RESUMED, + + XE_GT_SRIOV_STATE_STOP_WIP, + XE_GT_SRIOV_STATE_STOP_SEND_STOP, + XE_GT_SRIOV_STATE_STOP_FAILED, + XE_GT_SRIOV_STATE_STOPPED, + + XE_GT_SRIOV_STATE_MISMATCH = BITS_PER_LONG - 1, +}; + +/** + * struct xe_gt_sriov_control_state - GT-level per-VF control state. + * + * Used by the PF driver to maintain per-VF control data. + */ +struct xe_gt_sriov_control_state { + /** @state: VF state bits */ + unsigned long state; + + /** @done: completion of async operations */ + struct completion done; + + /** @link: link into worker list */ + struct list_head link; +}; + +/** + * struct xe_gt_sriov_pf_control - GT-level control data. + * + * Used by the PF driver to maintain its data. + */ +struct xe_gt_sriov_pf_control { + /** @worker: worker that executes a VF operations */ + struct work_struct worker; + + /** @list: list of VF entries that have a pending work */ + struct list_head list; + + /** @lock: protects VF pending list */ + spinlock_t lock; +}; + +#endif diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_pf_types.h b/drivers/gpu/drm/xe/xe_gt_sriov_pf_types.h index 40cbaea3ef44..28e1b130bf87 100644 --- a/drivers/gpu/drm/xe/xe_gt_sriov_pf_types.h +++ b/drivers/gpu/drm/xe/xe_gt_sriov_pf_types.h @@ -9,6 +9,7 @@ #include <linux/types.h> #include "xe_gt_sriov_pf_config_types.h" +#include "xe_gt_sriov_pf_control_types.h" #include "xe_gt_sriov_pf_monitor_types.h" #include "xe_gt_sriov_pf_policy_types.h" #include "xe_gt_sriov_pf_service_types.h" @@ -23,6 +24,9 @@ struct xe_gt_sriov_metadata { /** @monitor: per-VF monitoring data. */ struct xe_gt_sriov_monitor monitor; + /** @control: per-VF control data. */ + struct xe_gt_sriov_control_state control; + /** @version: negotiated VF/PF ABI version */ struct xe_gt_sriov_pf_service_version version; }; @@ -30,12 +34,14 @@ struct xe_gt_sriov_metadata { /** * struct xe_gt_sriov_pf - GT level PF virtualization data. * @service: service data. + * @control: control data. * @policy: policy data. * @spare: PF-only provisioning configuration. * @vfs: metadata for all VFs. */ struct xe_gt_sriov_pf { struct xe_gt_sriov_pf_service service; + struct xe_gt_sriov_pf_control control; struct xe_gt_sriov_pf_policy policy; struct xe_gt_sriov_spare_config spare; struct xe_gt_sriov_metadata *vfs; diff --git a/drivers/gpu/drm/xe/xe_gt_types.h b/drivers/gpu/drm/xe/xe_gt_types.h index 31946d7fe701..3d1c51de0268 100644 --- a/drivers/gpu/drm/xe/xe_gt_types.h +++ b/drivers/gpu/drm/xe/xe_gt_types.h @@ -329,12 +329,6 @@ struct xe_gt { /** @eclass: per hardware engine class interface on the GT */ struct xe_hw_engine_class_intf eclass[XE_ENGINE_CLASS_MAX]; - /** @pcode: GT's PCODE */ - struct { - /** @pcode.lock: protecting GT's PCODE mailbox data */ - struct mutex lock; - } pcode; - /** @sysfs: sysfs' kobj used by xe_gt_sysfs */ struct kobject *sysfs; diff --git a/drivers/gpu/drm/xe/xe_guc_pc.c b/drivers/gpu/drm/xe/xe_guc_pc.c index def503abeed5..034b29984d5e 100644 --- a/drivers/gpu/drm/xe/xe_guc_pc.c +++ b/drivers/gpu/drm/xe/xe_guc_pc.c @@ -915,7 +915,7 @@ static void pc_init_pcode_freq(struct xe_guc_pc *pc) u32 min = DIV_ROUND_CLOSEST(pc->rpn_freq, GT_FREQUENCY_MULTIPLIER); u32 max = DIV_ROUND_CLOSEST(pc->rp0_freq, GT_FREQUENCY_MULTIPLIER); - XE_WARN_ON(xe_pcode_init_min_freq_table(pc_to_gt(pc), min, max)); + XE_WARN_ON(xe_pcode_init_min_freq_table(gt_to_tile(pc_to_gt(pc)), min, max)); } static int pc_init_freqs(struct xe_guc_pc *pc) diff --git a/drivers/gpu/drm/xe/xe_hw_engine.c b/drivers/gpu/drm/xe/xe_hw_engine.c index 18980238a2ea..c9c3beb3ce8d 100644 --- a/drivers/gpu/drm/xe/xe_hw_engine.c +++ b/drivers/gpu/drm/xe/xe_hw_engine.c @@ -8,7 +8,7 @@ #include <linux/nospec.h> #include <drm/drm_managed.h> -#include <drm/xe_drm.h> +#include <uapi/drm/xe_drm.h> #include "regs/xe_engine_regs.h" #include "regs/xe_gt_regs.h" @@ -273,7 +273,6 @@ static void hw_engine_fini(void *arg) if (hwe->exl_port) xe_execlist_port_destroy(hwe->exl_port); - xe_lrc_put(hwe->kernel_lrc); hwe->gt = NULL; } @@ -558,21 +557,13 @@ static int hw_engine_init(struct xe_gt *gt, struct xe_hw_engine *hwe, goto err_name; } - hwe->kernel_lrc = xe_lrc_create(hwe, NULL, SZ_16K); - if (IS_ERR(hwe->kernel_lrc)) { - err = PTR_ERR(hwe->kernel_lrc); - goto err_hwsp; - } - if (!xe_device_uc_enabled(xe)) { hwe->exl_port = xe_execlist_port_create(xe, hwe); if (IS_ERR(hwe->exl_port)) { err = PTR_ERR(hwe->exl_port); - goto err_kernel_lrc; + goto err_hwsp; } - } - - if (xe_device_uc_enabled(xe)) { + } else { /* GSCCS has a special interrupt for reset */ if (hwe->class == XE_ENGINE_CLASS_OTHER) hwe->irq_handler = xe_gsc_hwe_irq_handler; @@ -587,8 +578,6 @@ static int hw_engine_init(struct xe_gt *gt, struct xe_hw_engine *hwe, return devm_add_action_or_reset(xe->drm.dev, hw_engine_fini, hwe); -err_kernel_lrc: - xe_lrc_put(hwe->kernel_lrc); err_hwsp: xe_bo_unpin_map_no_vm(hwe->hwsp); err_name: diff --git a/drivers/gpu/drm/xe/xe_hw_engine_types.h b/drivers/gpu/drm/xe/xe_hw_engine_types.h index 39f24012d0f4..8be6d420ece4 100644 --- a/drivers/gpu/drm/xe/xe_hw_engine_types.h +++ b/drivers/gpu/drm/xe/xe_hw_engine_types.h @@ -136,8 +136,6 @@ struct xe_hw_engine { enum xe_force_wake_domains domain; /** @hwsp: hardware status page buffer object */ struct xe_bo *hwsp; - /** @kernel_lrc: Kernel LRC (should be replaced /w an xe_engine) */ - struct xe_lrc *kernel_lrc; /** @exl_port: execlists port */ struct xe_execlist_port *exl_port; /** @fence_irq: fence IRQ to run when a hw engine IRQ is received */ diff --git a/drivers/gpu/drm/xe/xe_hwmon.c b/drivers/gpu/drm/xe/xe_hwmon.c index 832ea81faeee..aa11728e7e79 100644 --- a/drivers/gpu/drm/xe/xe_hwmon.c +++ b/drivers/gpu/drm/xe/xe_hwmon.c @@ -12,7 +12,6 @@ #include "regs/xe_mchbar_regs.h" #include "regs/xe_pcode_regs.h" #include "xe_device.h" -#include "xe_gt.h" #include "xe_hwmon.h" #include "xe_mmio.h" #include "xe_pcode.h" @@ -65,8 +64,8 @@ struct xe_hwmon_energy_info { struct xe_hwmon { /** @hwmon_dev: hwmon device for xe */ struct device *hwmon_dev; - /** @gt: primary gt */ - struct xe_gt *gt; + /** @xe: Xe device */ + struct xe_device *xe; /** @hwmon_lock: lock for rw attributes*/ struct mutex hwmon_lock; /** @scl_shift_power: pkg power unit */ @@ -82,7 +81,7 @@ struct xe_hwmon { static struct xe_reg xe_hwmon_get_reg(struct xe_hwmon *hwmon, enum xe_hwmon_reg hwmon_reg, int channel) { - struct xe_device *xe = gt_to_xe(hwmon->gt); + struct xe_device *xe = hwmon->xe; switch (hwmon_reg) { case REG_PKG_RAPL_LIMIT: @@ -148,8 +147,9 @@ static struct xe_reg xe_hwmon_get_reg(struct xe_hwmon *hwmon, enum xe_hwmon_reg static void xe_hwmon_power_max_read(struct xe_hwmon *hwmon, int channel, long *value) { u64 reg_val, min, max; - struct xe_device *xe = gt_to_xe(hwmon->gt); + struct xe_device *xe = hwmon->xe; struct xe_reg rapl_limit, pkg_power_sku; + struct xe_gt *mmio = xe_root_mmio_gt(xe); rapl_limit = xe_hwmon_get_reg(hwmon, REG_PKG_RAPL_LIMIT, channel); pkg_power_sku = xe_hwmon_get_reg(hwmon, REG_PKG_POWER_SKU, channel); @@ -166,7 +166,7 @@ static void xe_hwmon_power_max_read(struct xe_hwmon *hwmon, int channel, long *v mutex_lock(&hwmon->hwmon_lock); - reg_val = xe_mmio_read32(hwmon->gt, rapl_limit); + reg_val = xe_mmio_read32(mmio, rapl_limit); /* Check if PL1 limit is disabled */ if (!(reg_val & PKG_PWR_LIM_1_EN)) { *value = PL1_DISABLE; @@ -176,7 +176,7 @@ static void xe_hwmon_power_max_read(struct xe_hwmon *hwmon, int channel, long *v reg_val = REG_FIELD_GET(PKG_PWR_LIM_1, reg_val); *value = mul_u64_u32_shr(reg_val, SF_POWER, hwmon->scl_shift_power); - reg_val = xe_mmio_read64_2x32(hwmon->gt, pkg_power_sku); + reg_val = xe_mmio_read64_2x32(mmio, pkg_power_sku); min = REG_FIELD_GET(PKG_MIN_PWR, reg_val); min = mul_u64_u32_shr(min, SF_POWER, hwmon->scl_shift_power); max = REG_FIELD_GET(PKG_MAX_PWR, reg_val); @@ -190,6 +190,7 @@ unlock: static int xe_hwmon_power_max_write(struct xe_hwmon *hwmon, int channel, long value) { + struct xe_gt *mmio = xe_root_mmio_gt(hwmon->xe); int ret = 0; u64 reg_val; struct xe_reg rapl_limit; @@ -200,10 +201,10 @@ static int xe_hwmon_power_max_write(struct xe_hwmon *hwmon, int channel, long va /* Disable PL1 limit and verify, as limit cannot be disabled on all platforms */ if (value == PL1_DISABLE) { - reg_val = xe_mmio_rmw32(hwmon->gt, rapl_limit, PKG_PWR_LIM_1_EN, 0); - reg_val = xe_mmio_read32(hwmon->gt, rapl_limit); + reg_val = xe_mmio_rmw32(mmio, rapl_limit, PKG_PWR_LIM_1_EN, 0); + reg_val = xe_mmio_read32(mmio, rapl_limit); if (reg_val & PKG_PWR_LIM_1_EN) { - drm_warn(>_to_xe(hwmon->gt)->drm, "PL1 disable is not supported!\n"); + drm_warn(&hwmon->xe->drm, "PL1 disable is not supported!\n"); ret = -EOPNOTSUPP; } goto unlock; @@ -212,7 +213,7 @@ static int xe_hwmon_power_max_write(struct xe_hwmon *hwmon, int channel, long va /* Computation in 64-bits to avoid overflow. Round to nearest. */ reg_val = DIV_ROUND_CLOSEST_ULL((u64)value << hwmon->scl_shift_power, SF_POWER); reg_val = PKG_PWR_LIM_1_EN | REG_FIELD_PREP(PKG_PWR_LIM_1, reg_val); - reg_val = xe_mmio_rmw32(hwmon->gt, rapl_limit, PKG_PWR_LIM_1_EN | PKG_PWR_LIM_1, reg_val); + reg_val = xe_mmio_rmw32(mmio, rapl_limit, PKG_PWR_LIM_1_EN | PKG_PWR_LIM_1, reg_val); unlock: mutex_unlock(&hwmon->hwmon_lock); @@ -221,6 +222,7 @@ unlock: static void xe_hwmon_power_rated_max_read(struct xe_hwmon *hwmon, int channel, long *value) { + struct xe_gt *mmio = xe_root_mmio_gt(hwmon->xe); struct xe_reg reg = xe_hwmon_get_reg(hwmon, REG_PKG_POWER_SKU, channel); u64 reg_val; @@ -229,7 +231,7 @@ static void xe_hwmon_power_rated_max_read(struct xe_hwmon *hwmon, int channel, l * for this register can be skipped. * See xe_hwmon_power_is_visible. */ - reg_val = xe_mmio_read32(hwmon->gt, reg); + reg_val = xe_mmio_read32(mmio, reg); reg_val = REG_FIELD_GET(PKG_TDP, reg_val); *value = mul_u64_u32_shr(reg_val, SF_POWER, hwmon->scl_shift_power); } @@ -257,11 +259,12 @@ static void xe_hwmon_power_rated_max_read(struct xe_hwmon *hwmon, int channel, l static void xe_hwmon_energy_get(struct xe_hwmon *hwmon, int channel, long *energy) { + struct xe_gt *mmio = xe_root_mmio_gt(hwmon->xe); struct xe_hwmon_energy_info *ei = &hwmon->ei[channel]; u64 reg_val; - reg_val = xe_mmio_read32(hwmon->gt, xe_hwmon_get_reg(hwmon, REG_PKG_ENERGY_STATUS, - channel)); + reg_val = xe_mmio_read32(mmio, xe_hwmon_get_reg(hwmon, REG_PKG_ENERGY_STATUS, + channel)); if (reg_val >= ei->reg_val_prev) ei->accum_energy += reg_val - ei->reg_val_prev; @@ -279,19 +282,20 @@ xe_hwmon_power_max_interval_show(struct device *dev, struct device_attribute *at char *buf) { struct xe_hwmon *hwmon = dev_get_drvdata(dev); + struct xe_gt *mmio = xe_root_mmio_gt(hwmon->xe); u32 x, y, x_w = 2; /* 2 bits */ u64 r, tau4, out; int sensor_index = to_sensor_dev_attr(attr)->index; - xe_pm_runtime_get(gt_to_xe(hwmon->gt)); + xe_pm_runtime_get(hwmon->xe); mutex_lock(&hwmon->hwmon_lock); - r = xe_mmio_read32(hwmon->gt, xe_hwmon_get_reg(hwmon, REG_PKG_RAPL_LIMIT, sensor_index)); + r = xe_mmio_read32(mmio, xe_hwmon_get_reg(hwmon, REG_PKG_RAPL_LIMIT, sensor_index)); mutex_unlock(&hwmon->hwmon_lock); - xe_pm_runtime_put(gt_to_xe(hwmon->gt)); + xe_pm_runtime_put(hwmon->xe); x = REG_FIELD_GET(PKG_PWR_LIM_1_TIME_X, r); y = REG_FIELD_GET(PKG_PWR_LIM_1_TIME_Y, r); @@ -319,6 +323,7 @@ xe_hwmon_power_max_interval_store(struct device *dev, struct device_attribute *a const char *buf, size_t count) { struct xe_hwmon *hwmon = dev_get_drvdata(dev); + struct xe_gt *mmio = xe_root_mmio_gt(hwmon->xe); u32 x, y, rxy, x_w = 2; /* 2 bits */ u64 tau4, r, max_win; unsigned long val; @@ -371,16 +376,16 @@ xe_hwmon_power_max_interval_store(struct device *dev, struct device_attribute *a rxy = REG_FIELD_PREP(PKG_PWR_LIM_1_TIME_X, x) | REG_FIELD_PREP(PKG_PWR_LIM_1_TIME_Y, y); - xe_pm_runtime_get(gt_to_xe(hwmon->gt)); + xe_pm_runtime_get(hwmon->xe); mutex_lock(&hwmon->hwmon_lock); - r = xe_mmio_rmw32(hwmon->gt, xe_hwmon_get_reg(hwmon, REG_PKG_RAPL_LIMIT, sensor_index), + r = xe_mmio_rmw32(mmio, xe_hwmon_get_reg(hwmon, REG_PKG_RAPL_LIMIT, sensor_index), PKG_PWR_LIM_1_TIME, rxy); mutex_unlock(&hwmon->hwmon_lock); - xe_pm_runtime_put(gt_to_xe(hwmon->gt)); + xe_pm_runtime_put(hwmon->xe); return count; } @@ -406,11 +411,11 @@ static umode_t xe_hwmon_attributes_visible(struct kobject *kobj, struct xe_hwmon *hwmon = dev_get_drvdata(dev); int ret = 0; - xe_pm_runtime_get(gt_to_xe(hwmon->gt)); + xe_pm_runtime_get(hwmon->xe); ret = xe_reg_is_valid(xe_hwmon_get_reg(hwmon, REG_PKG_RAPL_LIMIT, index)) ? attr->mode : 0; - xe_pm_runtime_put(gt_to_xe(hwmon->gt)); + xe_pm_runtime_put(hwmon->xe); return ret; } @@ -435,22 +440,26 @@ static const struct hwmon_channel_info * const hwmon_info[] = { }; /* I1 is exposed as power_crit or as curr_crit depending on bit 31 */ -static int xe_hwmon_pcode_read_i1(struct xe_gt *gt, u32 *uval) +static int xe_hwmon_pcode_read_i1(const struct xe_hwmon *hwmon, u32 *uval) { + struct xe_tile *root_tile = xe_device_get_root_tile(hwmon->xe); + /* Avoid Illegal Subcommand error */ - if (gt_to_xe(gt)->info.platform == XE_DG2) + if (hwmon->xe->info.platform == XE_DG2) return -ENXIO; - return xe_pcode_read(gt, PCODE_MBOX(PCODE_POWER_SETUP, + return xe_pcode_read(root_tile, PCODE_MBOX(PCODE_POWER_SETUP, POWER_SETUP_SUBCOMMAND_READ_I1, 0), uval, NULL); } -static int xe_hwmon_pcode_write_i1(struct xe_gt *gt, u32 uval) +static int xe_hwmon_pcode_write_i1(const struct xe_hwmon *hwmon, u32 uval) { - return xe_pcode_write(gt, PCODE_MBOX(PCODE_POWER_SETUP, + struct xe_tile *root_tile = xe_device_get_root_tile(hwmon->xe); + + return xe_pcode_write(root_tile, PCODE_MBOX(PCODE_POWER_SETUP, POWER_SETUP_SUBCOMMAND_WRITE_I1, 0), - uval); + (uval & POWER_SETUP_I1_DATA_MASK)); } static int xe_hwmon_power_curr_crit_read(struct xe_hwmon *hwmon, int channel, @@ -461,7 +470,7 @@ static int xe_hwmon_power_curr_crit_read(struct xe_hwmon *hwmon, int channel, mutex_lock(&hwmon->hwmon_lock); - ret = xe_hwmon_pcode_read_i1(hwmon->gt, &uval); + ret = xe_hwmon_pcode_read_i1(hwmon, &uval); if (ret) goto unlock; @@ -481,7 +490,7 @@ static int xe_hwmon_power_curr_crit_write(struct xe_hwmon *hwmon, int channel, mutex_lock(&hwmon->hwmon_lock); uval = DIV_ROUND_CLOSEST_ULL(value << POWER_SETUP_I1_SHIFT, scale_factor); - ret = xe_hwmon_pcode_write_i1(hwmon->gt, uval); + ret = xe_hwmon_pcode_write_i1(hwmon, uval); mutex_unlock(&hwmon->hwmon_lock); return ret; @@ -489,9 +498,10 @@ static int xe_hwmon_power_curr_crit_write(struct xe_hwmon *hwmon, int channel, static void xe_hwmon_get_voltage(struct xe_hwmon *hwmon, int channel, long *value) { + struct xe_gt *mmio = xe_root_mmio_gt(hwmon->xe); u64 reg_val; - reg_val = xe_mmio_read32(hwmon->gt, xe_hwmon_get_reg(hwmon, REG_GT_PERF_STATUS, channel)); + reg_val = xe_mmio_read32(mmio, xe_hwmon_get_reg(hwmon, REG_GT_PERF_STATUS, channel)); /* HW register value in units of 2.5 millivolt */ *value = DIV_ROUND_CLOSEST(REG_FIELD_GET(VOLTAGE_MASK, reg_val) * 2500, SF_VOLTAGE); } @@ -510,7 +520,7 @@ xe_hwmon_power_is_visible(struct xe_hwmon *hwmon, u32 attr, int channel) channel)) ? 0444 : 0; case hwmon_power_crit: if (channel == CHANNEL_PKG) - return (xe_hwmon_pcode_read_i1(hwmon->gt, &uval) || + return (xe_hwmon_pcode_read_i1(hwmon, &uval) || !(uval & POWER_SETUP_I1_WATTS)) ? 0 : 0644; break; case hwmon_power_label: @@ -563,10 +573,10 @@ xe_hwmon_curr_is_visible(const struct xe_hwmon *hwmon, u32 attr, int channel) switch (attr) { case hwmon_curr_crit: - return (xe_hwmon_pcode_read_i1(hwmon->gt, &uval) || + return (xe_hwmon_pcode_read_i1(hwmon, &uval) || (uval & POWER_SETUP_I1_WATTS)) ? 0 : 0644; case hwmon_curr_label: - return (xe_hwmon_pcode_read_i1(hwmon->gt, &uval) || + return (xe_hwmon_pcode_read_i1(hwmon, &uval) || (uval & POWER_SETUP_I1_WATTS)) ? 0 : 0444; break; default: @@ -654,7 +664,7 @@ xe_hwmon_is_visible(const void *drvdata, enum hwmon_sensor_types type, struct xe_hwmon *hwmon = (struct xe_hwmon *)drvdata; int ret; - xe_pm_runtime_get(gt_to_xe(hwmon->gt)); + xe_pm_runtime_get(hwmon->xe); switch (type) { case hwmon_power: @@ -674,7 +684,7 @@ xe_hwmon_is_visible(const void *drvdata, enum hwmon_sensor_types type, break; } - xe_pm_runtime_put(gt_to_xe(hwmon->gt)); + xe_pm_runtime_put(hwmon->xe); return ret; } @@ -686,7 +696,7 @@ xe_hwmon_read(struct device *dev, enum hwmon_sensor_types type, u32 attr, struct xe_hwmon *hwmon = dev_get_drvdata(dev); int ret; - xe_pm_runtime_get(gt_to_xe(hwmon->gt)); + xe_pm_runtime_get(hwmon->xe); switch (type) { case hwmon_power: @@ -706,7 +716,7 @@ xe_hwmon_read(struct device *dev, enum hwmon_sensor_types type, u32 attr, break; } - xe_pm_runtime_put(gt_to_xe(hwmon->gt)); + xe_pm_runtime_put(hwmon->xe); return ret; } @@ -718,7 +728,7 @@ xe_hwmon_write(struct device *dev, enum hwmon_sensor_types type, u32 attr, struct xe_hwmon *hwmon = dev_get_drvdata(dev); int ret; - xe_pm_runtime_get(gt_to_xe(hwmon->gt)); + xe_pm_runtime_get(hwmon->xe); switch (type) { case hwmon_power: @@ -732,7 +742,7 @@ xe_hwmon_write(struct device *dev, enum hwmon_sensor_types type, u32 attr, break; } - xe_pm_runtime_put(gt_to_xe(hwmon->gt)); + xe_pm_runtime_put(hwmon->xe); return ret; } @@ -771,6 +781,7 @@ static const struct hwmon_chip_info hwmon_chip_info = { static void xe_hwmon_get_preregistration_info(struct xe_device *xe) { + struct xe_gt *mmio = xe_root_mmio_gt(xe); struct xe_hwmon *hwmon = xe->hwmon; long energy; u64 val_sku_unit = 0; @@ -783,7 +794,7 @@ xe_hwmon_get_preregistration_info(struct xe_device *xe) */ pkg_power_sku_unit = xe_hwmon_get_reg(hwmon, REG_PKG_POWER_SKU_UNIT, 0); if (xe_reg_is_valid(pkg_power_sku_unit)) { - val_sku_unit = xe_mmio_read32(hwmon->gt, pkg_power_sku_unit); + val_sku_unit = xe_mmio_read32(mmio, pkg_power_sku_unit); hwmon->scl_shift_power = REG_FIELD_GET(PKG_PWR_UNIT, val_sku_unit); hwmon->scl_shift_energy = REG_FIELD_GET(PKG_ENERGY_UNIT, val_sku_unit); hwmon->scl_shift_time = REG_FIELD_GET(PKG_TIME_UNIT, val_sku_unit); @@ -828,8 +839,8 @@ void xe_hwmon_register(struct xe_device *xe) if (devm_add_action_or_reset(dev, xe_hwmon_mutex_destroy, hwmon)) return; - /* primary GT to access device level properties */ - hwmon->gt = xe->tiles[0].primary_gt; + /* There's only one instance of hwmon per device */ + hwmon->xe = xe; xe_hwmon_get_preregistration_info(xe); diff --git a/drivers/gpu/drm/xe/xe_migrate.c b/drivers/gpu/drm/xe/xe_migrate.c index cbf54be224c9..cfd31ae49cc1 100644 --- a/drivers/gpu/drm/xe/xe_migrate.c +++ b/drivers/gpu/drm/xe/xe_migrate.c @@ -10,7 +10,7 @@ #include <drm/drm_managed.h> #include <drm/ttm/ttm_tt.h> -#include <drm/xe_drm.h> +#include <uapi/drm/xe_drm.h> #include <generated/xe_wa_oob.h> diff --git a/drivers/gpu/drm/xe/xe_oa.c b/drivers/gpu/drm/xe/xe_oa.c index 4d4541e0b24c..63286ed8457f 100644 --- a/drivers/gpu/drm/xe/xe_oa.c +++ b/drivers/gpu/drm/xe/xe_oa.c @@ -10,7 +10,7 @@ #include <drm/drm_drv.h> #include <drm/drm_managed.h> -#include <drm/xe_drm.h> +#include <uapi/drm/xe_drm.h> #include "abi/guc_actions_slpc_abi.h" #include "instructions/xe_mi_commands.h" diff --git a/drivers/gpu/drm/xe/xe_oa_types.h b/drivers/gpu/drm/xe/xe_oa_types.h index 540c3ec53a6d..8862eca73fbe 100644 --- a/drivers/gpu/drm/xe/xe_oa_types.h +++ b/drivers/gpu/drm/xe/xe_oa_types.h @@ -11,7 +11,7 @@ #include <linux/mutex.h> #include <linux/types.h> -#include <drm/xe_drm.h> +#include <uapi/drm/xe_drm.h> #include "regs/xe_reg_defs.h" #include "xe_hw_engine_types.h" diff --git a/drivers/gpu/drm/xe/xe_observation.c b/drivers/gpu/drm/xe/xe_observation.c index a78c92a44ec2..8ec1b84cbb9e 100644 --- a/drivers/gpu/drm/xe/xe_observation.c +++ b/drivers/gpu/drm/xe/xe_observation.c @@ -6,7 +6,7 @@ #include <linux/errno.h> #include <linux/sysctl.h> -#include <drm/xe_drm.h> +#include <uapi/drm/xe_drm.h> #include "xe_oa.h" #include "xe_observation.h" diff --git a/drivers/gpu/drm/xe/xe_pat.c b/drivers/gpu/drm/xe/xe_pat.c index 722278cc23fc..f291a1730024 100644 --- a/drivers/gpu/drm/xe/xe_pat.c +++ b/drivers/gpu/drm/xe/xe_pat.c @@ -5,7 +5,7 @@ #include "xe_pat.h" -#include <drm/xe_drm.h> +#include <uapi/drm/xe_drm.h> #include <generated/xe_wa_oob.h> diff --git a/drivers/gpu/drm/xe/xe_pci.c b/drivers/gpu/drm/xe/xe_pci.c index f276194d9c4e..937c3e064f0d 100644 --- a/drivers/gpu/drm/xe/xe_pci.c +++ b/drivers/gpu/drm/xe/xe_pci.c @@ -793,7 +793,7 @@ static int xe_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent) if (IS_ERR(xe)) return PTR_ERR(xe); - pci_set_drvdata(pdev, xe); + pci_set_drvdata(pdev, &xe->drm); xe_pm_assert_unbounded_bridge(xe); subplatform_desc = find_subplatform(xe, desc); diff --git a/drivers/gpu/drm/xe/xe_pcode.c b/drivers/gpu/drm/xe/xe_pcode.c index 9c4eefdf6642..7397d556996a 100644 --- a/drivers/gpu/drm/xe/xe_pcode.c +++ b/drivers/gpu/drm/xe/xe_pcode.c @@ -12,7 +12,6 @@ #include "xe_assert.h" #include "xe_device.h" -#include "xe_gt.h" #include "xe_mmio.h" #include "xe_pcode_api.h" @@ -30,7 +29,7 @@ * - PCODE for display operations */ -static int pcode_mailbox_status(struct xe_gt *gt) +static int pcode_mailbox_status(struct xe_tile *tile) { u32 err; static const struct pcode_err_decode err_decode[] = { @@ -45,9 +44,9 @@ static int pcode_mailbox_status(struct xe_gt *gt) [PCODE_ERROR_MASK] = {-EPROTO, "Unknown"}, }; - err = xe_mmio_read32(gt, PCODE_MAILBOX) & PCODE_ERROR_MASK; + err = xe_mmio_read32(tile->primary_gt, PCODE_MAILBOX) & PCODE_ERROR_MASK; if (err) { - drm_err(>_to_xe(gt)->drm, "PCODE Mailbox failed: %d %s", err, + drm_err(&tile_to_xe(tile)->drm, "PCODE Mailbox failed: %d %s", err, err_decode[err].str ?: "Unknown"); return err_decode[err].errno ?: -EPROTO; } @@ -55,84 +54,85 @@ static int pcode_mailbox_status(struct xe_gt *gt) return 0; } -static int __pcode_mailbox_rw(struct xe_gt *gt, u32 mbox, u32 *data0, u32 *data1, +static int __pcode_mailbox_rw(struct xe_tile *tile, u32 mbox, u32 *data0, u32 *data1, unsigned int timeout_ms, bool return_data, bool atomic) { + struct xe_gt *mmio = tile->primary_gt; int err; - if (gt_to_xe(gt)->info.skip_pcode) + if (tile_to_xe(tile)->info.skip_pcode) return 0; - if ((xe_mmio_read32(gt, PCODE_MAILBOX) & PCODE_READY) != 0) + if ((xe_mmio_read32(mmio, PCODE_MAILBOX) & PCODE_READY) != 0) return -EAGAIN; - xe_mmio_write32(gt, PCODE_DATA0, *data0); - xe_mmio_write32(gt, PCODE_DATA1, data1 ? *data1 : 0); - xe_mmio_write32(gt, PCODE_MAILBOX, PCODE_READY | mbox); + xe_mmio_write32(mmio, PCODE_DATA0, *data0); + xe_mmio_write32(mmio, PCODE_DATA1, data1 ? *data1 : 0); + xe_mmio_write32(mmio, PCODE_MAILBOX, PCODE_READY | mbox); - err = xe_mmio_wait32(gt, PCODE_MAILBOX, PCODE_READY, 0, + err = xe_mmio_wait32(mmio, PCODE_MAILBOX, PCODE_READY, 0, timeout_ms * USEC_PER_MSEC, NULL, atomic); if (err) return err; if (return_data) { - *data0 = xe_mmio_read32(gt, PCODE_DATA0); + *data0 = xe_mmio_read32(mmio, PCODE_DATA0); if (data1) - *data1 = xe_mmio_read32(gt, PCODE_DATA1); + *data1 = xe_mmio_read32(mmio, PCODE_DATA1); } - return pcode_mailbox_status(gt); + return pcode_mailbox_status(tile); } -static int pcode_mailbox_rw(struct xe_gt *gt, u32 mbox, u32 *data0, u32 *data1, +static int pcode_mailbox_rw(struct xe_tile *tile, u32 mbox, u32 *data0, u32 *data1, unsigned int timeout_ms, bool return_data, bool atomic) { - if (gt_to_xe(gt)->info.skip_pcode) + if (tile_to_xe(tile)->info.skip_pcode) return 0; - lockdep_assert_held(>->pcode.lock); + lockdep_assert_held(&tile->pcode.lock); - return __pcode_mailbox_rw(gt, mbox, data0, data1, timeout_ms, return_data, atomic); + return __pcode_mailbox_rw(tile, mbox, data0, data1, timeout_ms, return_data, atomic); } -int xe_pcode_write_timeout(struct xe_gt *gt, u32 mbox, u32 data, int timeout) +int xe_pcode_write_timeout(struct xe_tile *tile, u32 mbox, u32 data, int timeout) { int err; - mutex_lock(>->pcode.lock); - err = pcode_mailbox_rw(gt, mbox, &data, NULL, timeout, false, false); - mutex_unlock(>->pcode.lock); + mutex_lock(&tile->pcode.lock); + err = pcode_mailbox_rw(tile, mbox, &data, NULL, timeout, false, false); + mutex_unlock(&tile->pcode.lock); return err; } -int xe_pcode_read(struct xe_gt *gt, u32 mbox, u32 *val, u32 *val1) +int xe_pcode_read(struct xe_tile *tile, u32 mbox, u32 *val, u32 *val1) { int err; - mutex_lock(>->pcode.lock); - err = pcode_mailbox_rw(gt, mbox, val, val1, 1, true, false); - mutex_unlock(>->pcode.lock); + mutex_lock(&tile->pcode.lock); + err = pcode_mailbox_rw(tile, mbox, val, val1, 1, true, false); + mutex_unlock(&tile->pcode.lock); return err; } -static int pcode_try_request(struct xe_gt *gt, u32 mbox, +static int pcode_try_request(struct xe_tile *tile, u32 mbox, u32 request, u32 reply_mask, u32 reply, u32 *status, bool atomic, int timeout_us, bool locked) { int slept, wait = 10; - xe_gt_assert(gt, timeout_us > 0); + xe_tile_assert(tile, timeout_us > 0); for (slept = 0; slept < timeout_us; slept += wait) { if (locked) - *status = pcode_mailbox_rw(gt, mbox, &request, NULL, 1, true, + *status = pcode_mailbox_rw(tile, mbox, &request, NULL, 1, true, atomic); else - *status = __pcode_mailbox_rw(gt, mbox, &request, NULL, 1, true, + *status = __pcode_mailbox_rw(tile, mbox, &request, NULL, 1, true, atomic); if ((*status == 0) && ((request & reply_mask) == reply)) return 0; @@ -149,7 +149,7 @@ static int pcode_try_request(struct xe_gt *gt, u32 mbox, /** * xe_pcode_request - send PCODE request until acknowledgment - * @gt: gt + * @tile: tile * @mbox: PCODE mailbox ID the request is targeted for * @request: request ID * @reply_mask: mask used to check for request acknowledgment @@ -166,17 +166,17 @@ static int pcode_try_request(struct xe_gt *gt, u32 mbox, * Returns 0 on success, %-ETIMEDOUT in case of a timeout, <0 in case of some * other error as reported by PCODE. */ -int xe_pcode_request(struct xe_gt *gt, u32 mbox, u32 request, - u32 reply_mask, u32 reply, int timeout_base_ms) +int xe_pcode_request(struct xe_tile *tile, u32 mbox, u32 request, + u32 reply_mask, u32 reply, int timeout_base_ms) { u32 status; int ret; - xe_gt_assert(gt, timeout_base_ms <= 3); + xe_tile_assert(tile, timeout_base_ms <= 3); - mutex_lock(>->pcode.lock); + mutex_lock(&tile->pcode.lock); - ret = pcode_try_request(gt, mbox, request, reply_mask, reply, &status, + ret = pcode_try_request(tile, mbox, request, reply_mask, reply, &status, false, timeout_base_ms * 1000, true); if (!ret) goto out; @@ -191,20 +191,20 @@ int xe_pcode_request(struct xe_gt *gt, u32 mbox, u32 request, * requests, and for any quirks of the PCODE firmware that delays * the request completion. */ - drm_err(>_to_xe(gt)->drm, + drm_err(&tile_to_xe(tile)->drm, "PCODE timeout, retrying with preemption disabled\n"); preempt_disable(); - ret = pcode_try_request(gt, mbox, request, reply_mask, reply, &status, + ret = pcode_try_request(tile, mbox, request, reply_mask, reply, &status, true, 50 * 1000, true); preempt_enable(); out: - mutex_unlock(>->pcode.lock); + mutex_unlock(&tile->pcode.lock); return status ? status : ret; } /** * xe_pcode_init_min_freq_table - Initialize PCODE's QOS frequency table - * @gt: gt instance + * @tile: tile instance * @min_gt_freq: Minimal (RPn) GT frequency in units of 50MHz. * @max_gt_freq: Maximal (RP0) GT frequency in units of 50MHz. * @@ -227,30 +227,30 @@ out: * - -EACCES, "PCODE Rejected" * - -EPROTO, "Unknown" */ -int xe_pcode_init_min_freq_table(struct xe_gt *gt, u32 min_gt_freq, +int xe_pcode_init_min_freq_table(struct xe_tile *tile, u32 min_gt_freq, u32 max_gt_freq) { int ret; u32 freq; - if (!gt_to_xe(gt)->info.has_llc) + if (!tile_to_xe(tile)->info.has_llc) return 0; if (max_gt_freq <= min_gt_freq) return -EINVAL; - mutex_lock(>->pcode.lock); + mutex_lock(&tile->pcode.lock); for (freq = min_gt_freq; freq <= max_gt_freq; freq++) { u32 data = freq << PCODE_FREQ_RING_RATIO_SHIFT | freq; - ret = pcode_mailbox_rw(gt, PCODE_WRITE_MIN_FREQ_TABLE, + ret = pcode_mailbox_rw(tile, PCODE_WRITE_MIN_FREQ_TABLE, &data, NULL, 1, false, false); if (ret) goto unlock; } unlock: - mutex_unlock(>->pcode.lock); + mutex_unlock(&tile->pcode.lock); return ret; } @@ -270,7 +270,7 @@ unlock: int xe_pcode_ready(struct xe_device *xe, bool locked) { u32 status, request = DGFX_GET_INIT_STATUS; - struct xe_gt *gt = xe_root_mmio_gt(xe); + struct xe_tile *tile = xe_device_get_root_tile(xe); int timeout_us = 180000000; /* 3 min */ int ret; @@ -281,15 +281,15 @@ int xe_pcode_ready(struct xe_device *xe, bool locked) return 0; if (locked) - mutex_lock(>->pcode.lock); + mutex_lock(&tile->pcode.lock); - ret = pcode_try_request(gt, DGFX_PCODE_STATUS, request, + ret = pcode_try_request(tile, DGFX_PCODE_STATUS, request, DGFX_INIT_STATUS_COMPLETE, DGFX_INIT_STATUS_COMPLETE, &status, false, timeout_us, locked); if (locked) - mutex_unlock(>->pcode.lock); + mutex_unlock(&tile->pcode.lock); if (ret) drm_err(&xe->drm, @@ -300,14 +300,14 @@ int xe_pcode_ready(struct xe_device *xe, bool locked) /** * xe_pcode_init: initialize components of PCODE - * @gt: gt instance + * @tile: tile instance * * This function initializes the xe_pcode component. * To be called once only during probe. */ -void xe_pcode_init(struct xe_gt *gt) +void xe_pcode_init(struct xe_tile *tile) { - drmm_mutex_init(>_to_xe(gt)->drm, >->pcode.lock); + drmm_mutex_init(&tile_to_xe(tile)->drm, &tile->pcode.lock); } /** diff --git a/drivers/gpu/drm/xe/xe_pcode.h b/drivers/gpu/drm/xe/xe_pcode.h index 3f54c6d2a57d..ba33991d72a7 100644 --- a/drivers/gpu/drm/xe/xe_pcode.h +++ b/drivers/gpu/drm/xe/xe_pcode.h @@ -7,21 +7,21 @@ #define _XE_PCODE_H_ #include <linux/types.h> -struct xe_gt; +struct xe_tile; struct xe_device; -void xe_pcode_init(struct xe_gt *gt); +void xe_pcode_init(struct xe_tile *tile); int xe_pcode_probe_early(struct xe_device *xe); int xe_pcode_ready(struct xe_device *xe, bool locked); -int xe_pcode_init_min_freq_table(struct xe_gt *gt, u32 min_gt_freq, +int xe_pcode_init_min_freq_table(struct xe_tile *tile, u32 min_gt_freq, u32 max_gt_freq); -int xe_pcode_read(struct xe_gt *gt, u32 mbox, u32 *val, u32 *val1); -int xe_pcode_write_timeout(struct xe_gt *gt, u32 mbox, u32 val, +int xe_pcode_read(struct xe_tile *tile, u32 mbox, u32 *val, u32 *val1); +int xe_pcode_write_timeout(struct xe_tile *tile, u32 mbox, u32 val, int timeout_ms); -#define xe_pcode_write(gt, mbox, val) \ - xe_pcode_write_timeout(gt, mbox, val, 1) +#define xe_pcode_write(tile, mbox, val) \ + xe_pcode_write_timeout(tile, mbox, val, 1) -int xe_pcode_request(struct xe_gt *gt, u32 mbox, u32 request, +int xe_pcode_request(struct xe_tile *tile, u32 mbox, u32 request, u32 reply_mask, u32 reply, int timeout_ms); #define PCODE_MBOX(mbcmd, param1, param2)\ diff --git a/drivers/gpu/drm/xe/xe_pm.c b/drivers/gpu/drm/xe/xe_pm.c index 2e2accd76fb2..e518557e0eec 100644 --- a/drivers/gpu/drm/xe/xe_pm.c +++ b/drivers/gpu/drm/xe/xe_pm.c @@ -79,7 +79,14 @@ static struct lockdep_map xe_pm_runtime_nod3cold_map = { }; #endif -static bool __maybe_unused xe_rpm_reclaim_safe(const struct xe_device *xe) +/** + * xe_rpm_reclaim_safe() - Whether runtime resume can be done from reclaim context + * @xe: The xe device. + * + * Return: true if it is safe to runtime resume from reclaim context. + * false otherwise. + */ +bool xe_rpm_reclaim_safe(const struct xe_device *xe) { return !xe->d3cold.capable && !xe->info.has_sriov; } @@ -392,8 +399,6 @@ int xe_pm_runtime_suspend(struct xe_device *xe) xe_display_pm_runtime_suspend(xe); if (xe->d3cold.allowed) { - xe_display_pm_suspend(xe, true); - err = xe_bo_evict_all(xe); if (err) goto out; diff --git a/drivers/gpu/drm/xe/xe_pm.h b/drivers/gpu/drm/xe/xe_pm.h index 9aef673b1c8a..998d1ed64556 100644 --- a/drivers/gpu/drm/xe/xe_pm.h +++ b/drivers/gpu/drm/xe/xe_pm.h @@ -31,6 +31,7 @@ bool xe_pm_runtime_resume_and_get(struct xe_device *xe); void xe_pm_assert_unbounded_bridge(struct xe_device *xe); int xe_pm_set_vram_threshold(struct xe_device *xe, u32 threshold); void xe_pm_d3cold_allowed_toggle(struct xe_device *xe); +bool xe_rpm_reclaim_safe(const struct xe_device *xe); struct task_struct *xe_pm_read_callback_task(struct xe_device *xe); int xe_pm_module_init(void); diff --git a/drivers/gpu/drm/xe/xe_pt.c b/drivers/gpu/drm/xe/xe_pt.c index 579ed31b46db..d6353e8969f0 100644 --- a/drivers/gpu/drm/xe/xe_pt.c +++ b/drivers/gpu/drm/xe/xe_pt.c @@ -3,6 +3,8 @@ * Copyright © 2022 Intel Corporation */ +#include <linux/dma-fence-array.h> + #include "xe_pt.h" #include "regs/xe_gtt_defs.h" @@ -1627,9 +1629,11 @@ xe_pt_update_ops_rfence_interval(struct xe_vm_pgtable_update_ops *pt_update_ops, static int vma_reserve_fences(struct xe_device *xe, struct xe_vma *vma) { + int shift = xe_device_get_root_tile(xe)->media_gt ? 1 : 0; + if (!xe_vma_has_no_bo(vma) && !xe_vma_bo(vma)->vm) return dma_resv_reserve_fences(xe_vma_bo(vma)->ttm.base.resv, - xe->info.tile_count); + xe->info.tile_count << shift); return 0; } @@ -1816,6 +1820,7 @@ int xe_pt_update_ops_prepare(struct xe_tile *tile, struct xe_vma_ops *vops) struct xe_vm_pgtable_update_ops *pt_update_ops = &vops->pt_update_ops[tile->id]; struct xe_vma_op *op; + int shift = tile->media_gt ? 1 : 0; int err; lockdep_assert_held(&vops->vm->lock); @@ -1824,7 +1829,7 @@ int xe_pt_update_ops_prepare(struct xe_tile *tile, struct xe_vma_ops *vops) xe_pt_update_ops_init(pt_update_ops); err = dma_resv_reserve_fences(xe_vm_resv(vops->vm), - tile_to_xe(tile)->info.tile_count); + tile_to_xe(tile)->info.tile_count << shift); if (err) return err; @@ -1849,13 +1854,20 @@ int xe_pt_update_ops_prepare(struct xe_tile *tile, struct xe_vma_ops *vops) static void bind_op_commit(struct xe_vm *vm, struct xe_tile *tile, struct xe_vm_pgtable_update_ops *pt_update_ops, - struct xe_vma *vma, struct dma_fence *fence) + struct xe_vma *vma, struct dma_fence *fence, + struct dma_fence *fence2) { - if (!xe_vma_has_no_bo(vma) && !xe_vma_bo(vma)->vm) + if (!xe_vma_has_no_bo(vma) && !xe_vma_bo(vma)->vm) { dma_resv_add_fence(xe_vma_bo(vma)->ttm.base.resv, fence, pt_update_ops->wait_vm_bookkeep ? DMA_RESV_USAGE_KERNEL : DMA_RESV_USAGE_BOOKKEEP); + if (fence2) + dma_resv_add_fence(xe_vma_bo(vma)->ttm.base.resv, fence2, + pt_update_ops->wait_vm_bookkeep ? + DMA_RESV_USAGE_KERNEL : + DMA_RESV_USAGE_BOOKKEEP); + } vma->tile_present |= BIT(tile->id); vma->tile_staged &= ~BIT(tile->id); if (xe_vma_is_userptr(vma)) { @@ -1875,13 +1887,20 @@ static void bind_op_commit(struct xe_vm *vm, struct xe_tile *tile, static void unbind_op_commit(struct xe_vm *vm, struct xe_tile *tile, struct xe_vm_pgtable_update_ops *pt_update_ops, - struct xe_vma *vma, struct dma_fence *fence) + struct xe_vma *vma, struct dma_fence *fence, + struct dma_fence *fence2) { - if (!xe_vma_has_no_bo(vma) && !xe_vma_bo(vma)->vm) + if (!xe_vma_has_no_bo(vma) && !xe_vma_bo(vma)->vm) { dma_resv_add_fence(xe_vma_bo(vma)->ttm.base.resv, fence, pt_update_ops->wait_vm_bookkeep ? DMA_RESV_USAGE_KERNEL : DMA_RESV_USAGE_BOOKKEEP); + if (fence2) + dma_resv_add_fence(xe_vma_bo(vma)->ttm.base.resv, fence2, + pt_update_ops->wait_vm_bookkeep ? + DMA_RESV_USAGE_KERNEL : + DMA_RESV_USAGE_BOOKKEEP); + } vma->tile_present &= ~BIT(tile->id); if (!vma->tile_present) { list_del_init(&vma->combined_links.rebind); @@ -1898,7 +1917,8 @@ static void unbind_op_commit(struct xe_vm *vm, struct xe_tile *tile, static void op_commit(struct xe_vm *vm, struct xe_tile *tile, struct xe_vm_pgtable_update_ops *pt_update_ops, - struct xe_vma_op *op, struct dma_fence *fence) + struct xe_vma_op *op, struct dma_fence *fence, + struct dma_fence *fence2) { xe_vm_assert_held(vm); @@ -1907,26 +1927,28 @@ static void op_commit(struct xe_vm *vm, if (!op->map.immediate && xe_vm_in_fault_mode(vm)) break; - bind_op_commit(vm, tile, pt_update_ops, op->map.vma, fence); + bind_op_commit(vm, tile, pt_update_ops, op->map.vma, fence, + fence2); break; case DRM_GPUVA_OP_REMAP: unbind_op_commit(vm, tile, pt_update_ops, - gpuva_to_vma(op->base.remap.unmap->va), fence); + gpuva_to_vma(op->base.remap.unmap->va), fence, + fence2); if (op->remap.prev) bind_op_commit(vm, tile, pt_update_ops, op->remap.prev, - fence); + fence, fence2); if (op->remap.next) bind_op_commit(vm, tile, pt_update_ops, op->remap.next, - fence); + fence, fence2); break; case DRM_GPUVA_OP_UNMAP: unbind_op_commit(vm, tile, pt_update_ops, - gpuva_to_vma(op->base.unmap.va), fence); + gpuva_to_vma(op->base.unmap.va), fence, fence2); break; case DRM_GPUVA_OP_PREFETCH: bind_op_commit(vm, tile, pt_update_ops, - gpuva_to_vma(op->base.prefetch.va), fence); + gpuva_to_vma(op->base.prefetch.va), fence, fence2); break; default: drm_warn(&vm->xe->drm, "NOT POSSIBLE"); @@ -1963,7 +1985,9 @@ xe_pt_update_ops_run(struct xe_tile *tile, struct xe_vma_ops *vops) struct xe_vm_pgtable_update_ops *pt_update_ops = &vops->pt_update_ops[tile->id]; struct dma_fence *fence; - struct invalidation_fence *ifence = NULL; + struct invalidation_fence *ifence = NULL, *mfence = NULL; + struct dma_fence **fences = NULL; + struct dma_fence_array *cf = NULL; struct xe_range_fence *rfence; struct xe_vma_op *op; int err = 0, i; @@ -1996,6 +2020,23 @@ xe_pt_update_ops_run(struct xe_tile *tile, struct xe_vma_ops *vops) err = -ENOMEM; goto kill_vm_tile1; } + if (tile->media_gt) { + mfence = kzalloc(sizeof(*ifence), GFP_KERNEL); + if (!mfence) { + err = -ENOMEM; + goto free_ifence; + } + fences = kmalloc_array(2, sizeof(*fences), GFP_KERNEL); + if (!fences) { + err = -ENOMEM; + goto free_ifence; + } + cf = dma_fence_array_alloc(2); + if (!cf) { + err = -ENOMEM; + goto free_ifence; + } + } } rfence = kzalloc(sizeof(*rfence), GFP_KERNEL); @@ -2027,19 +2068,50 @@ xe_pt_update_ops_run(struct xe_tile *tile, struct xe_vma_ops *vops) /* tlb invalidation must be done before signaling rebind */ if (ifence) { + if (mfence) + dma_fence_get(fence); invalidation_fence_init(tile->primary_gt, ifence, fence, pt_update_ops->start, pt_update_ops->last, vm->usm.asid); - fence = &ifence->base.base; + if (mfence) { + invalidation_fence_init(tile->media_gt, mfence, fence, + pt_update_ops->start, + pt_update_ops->last, vm->usm.asid); + fences[0] = &ifence->base.base; + fences[1] = &mfence->base.base; + dma_fence_array_init(cf, 2, fences, + vm->composite_fence_ctx, + vm->composite_fence_seqno++, + false); + fence = &cf->base; + } else { + fence = &ifence->base.base; + } } - dma_resv_add_fence(xe_vm_resv(vm), fence, - pt_update_ops->wait_vm_bookkeep ? - DMA_RESV_USAGE_KERNEL : - DMA_RESV_USAGE_BOOKKEEP); + if (!mfence) { + dma_resv_add_fence(xe_vm_resv(vm), fence, + pt_update_ops->wait_vm_bookkeep ? + DMA_RESV_USAGE_KERNEL : + DMA_RESV_USAGE_BOOKKEEP); - list_for_each_entry(op, &vops->list, link) - op_commit(vops->vm, tile, pt_update_ops, op, fence); + list_for_each_entry(op, &vops->list, link) + op_commit(vops->vm, tile, pt_update_ops, op, fence, NULL); + } else { + dma_resv_add_fence(xe_vm_resv(vm), &ifence->base.base, + pt_update_ops->wait_vm_bookkeep ? + DMA_RESV_USAGE_KERNEL : + DMA_RESV_USAGE_BOOKKEEP); + + dma_resv_add_fence(xe_vm_resv(vm), &mfence->base.base, + pt_update_ops->wait_vm_bookkeep ? + DMA_RESV_USAGE_KERNEL : + DMA_RESV_USAGE_BOOKKEEP); + + list_for_each_entry(op, &vops->list, link) + op_commit(vops->vm, tile, pt_update_ops, op, + &ifence->base.base, &mfence->base.base); + } if (pt_update_ops->needs_userptr_lock) up_read(&vm->userptr.notifier_lock); @@ -2049,6 +2121,9 @@ xe_pt_update_ops_run(struct xe_tile *tile, struct xe_vma_ops *vops) free_rfence: kfree(rfence); free_ifence: + kfree(cf); + kfree(fences); + kfree(mfence); kfree(ifence); kill_vm_tile1: if (err != -EAGAIN && tile->id) diff --git a/drivers/gpu/drm/xe/xe_query.c b/drivers/gpu/drm/xe/xe_query.c index 73ef6e4c2dc9..28d9bb3b825d 100644 --- a/drivers/gpu/drm/xe/xe_query.c +++ b/drivers/gpu/drm/xe/xe_query.c @@ -9,7 +9,7 @@ #include <linux/sched/clock.h> #include <drm/ttm/ttm_placement.h> -#include <drm/xe_drm.h> +#include <uapi/drm/xe_drm.h> #include "regs/xe_engine_regs.h" #include "regs/xe_gt_regs.h" diff --git a/drivers/gpu/drm/xe/xe_rtp.c b/drivers/gpu/drm/xe/xe_rtp.c index e78ba324dd18..86c705d18c0d 100644 --- a/drivers/gpu/drm/xe/xe_rtp.c +++ b/drivers/gpu/drm/xe/xe_rtp.c @@ -7,7 +7,7 @@ #include <kunit/visibility.h> -#include <drm/xe_drm.h> +#include <uapi/drm/xe_drm.h> #include "xe_gt.h" #include "xe_gt_topology.h" diff --git a/drivers/gpu/drm/xe/xe_sched_job.c b/drivers/gpu/drm/xe/xe_sched_job.c index 55d47450b2c6..eeccc1c318ae 100644 --- a/drivers/gpu/drm/xe/xe_sched_job.c +++ b/drivers/gpu/drm/xe/xe_sched_job.c @@ -5,7 +5,7 @@ #include "xe_sched_job.h" -#include <drm/xe_drm.h> +#include <uapi/drm/xe_drm.h> #include <linux/dma-fence-chain.h> #include <linux/slab.h> diff --git a/drivers/gpu/drm/xe/xe_sync.c b/drivers/gpu/drm/xe/xe_sync.c index 436faff09bac..bb3c2a830362 100644 --- a/drivers/gpu/drm/xe/xe_sync.c +++ b/drivers/gpu/drm/xe/xe_sync.c @@ -12,7 +12,7 @@ #include <drm/drm_print.h> #include <drm/drm_syncobj.h> -#include <drm/xe_drm.h> +#include <uapi/drm/xe_drm.h> #include "xe_device_types.h" #include "xe_exec_queue.h" diff --git a/drivers/gpu/drm/xe/xe_tile.c b/drivers/gpu/drm/xe/xe_tile.c index 15ea0a942f67..dda5268507d8 100644 --- a/drivers/gpu/drm/xe/xe_tile.c +++ b/drivers/gpu/drm/xe/xe_tile.c @@ -9,6 +9,7 @@ #include "xe_ggtt.h" #include "xe_gt.h" #include "xe_migrate.h" +#include "xe_pcode.h" #include "xe_sa.h" #include "xe_tile.h" #include "xe_tile_sysfs.h" @@ -124,6 +125,8 @@ int xe_tile_init_early(struct xe_tile *tile, struct xe_device *xe, u8 id) if (IS_ERR(tile->primary_gt)) return PTR_ERR(tile->primary_gt); + xe_pcode_init(tile); + return 0; } diff --git a/drivers/gpu/drm/xe/xe_uc_debugfs.c b/drivers/gpu/drm/xe/xe_uc_debugfs.c index 78eb8db73791..24a4209051ee 100644 --- a/drivers/gpu/drm/xe/xe_uc_debugfs.c +++ b/drivers/gpu/drm/xe/xe_uc_debugfs.c @@ -8,6 +8,7 @@ #include <drm/drm_debugfs.h> #include "xe_gt.h" +#include "xe_gsc_debugfs.h" #include "xe_guc_debugfs.h" #include "xe_huc_debugfs.h" #include "xe_macros.h" @@ -23,6 +24,7 @@ void xe_uc_debugfs_register(struct xe_uc *uc, struct dentry *parent) return; } + xe_gsc_debugfs_register(&uc->gsc, root); xe_guc_debugfs_register(&uc->guc, root); xe_huc_debugfs_register(&uc->huc, root); } diff --git a/drivers/gpu/drm/xe/xe_uc_fw.c b/drivers/gpu/drm/xe/xe_uc_fw.c index 4bb2a4a80ddc..d431d0031185 100644 --- a/drivers/gpu/drm/xe/xe_uc_fw.c +++ b/drivers/gpu/drm/xe/xe_uc_fw.c @@ -129,8 +129,8 @@ struct fw_blobs_by_type { /* for the GSC FW we match the compatibility version and not the release one */ #define XE_GSC_FIRMWARE_DEFS(fw_def, major_ver) \ - fw_def(LUNARLAKE, major_ver(xe, gsc, lnl, 1, 0, 0)) \ - fw_def(METEORLAKE, major_ver(i915, gsc, mtl, 1, 0, 0)) + fw_def(LUNARLAKE, major_ver(xe, gsc, lnl, 104, 1, 0)) \ + fw_def(METEORLAKE, major_ver(i915, gsc, mtl, 102, 1, 0)) #define MAKE_FW_PATH(dir__, uc__, shortname__, version__) \ __stringify(dir__) "/" __stringify(shortname__) "_" __stringify(uc__) version__ ".bin" @@ -141,6 +141,8 @@ struct fw_blobs_by_type { MAKE_FW_PATH(dir_, uc_, shortname_, "_" __stringify(a)) #define fw_filename_no_ver(dir_, uc_, shortname_) \ MAKE_FW_PATH(dir_, uc_, shortname_, "") +#define fw_filename_gsc(dir_, uc_, shortname_, a, b, c) \ + MAKE_FW_PATH(dir_, uc_, shortname_, "_" __stringify(b)) #define uc_fw_entry_mmp_ver(dir_, uc_, shortname_, a, b, c) \ { fw_filename_mmp_ver(dir_, uc_, shortname_, a, b, c), \ @@ -151,6 +153,9 @@ struct fw_blobs_by_type { #define uc_fw_entry_no_ver(dir_, uc_, shortname_) \ { fw_filename_no_ver(dir_, uc_, shortname_), \ 0, 0 } +#define uc_fw_entry_gsc(dir_, uc_, shortname_, a, b, c) \ + { fw_filename_gsc(dir_, uc_, shortname_, a, b, c), \ + a, b, c } /* All blobs need to be declared via MODULE_FIRMWARE() */ #define XE_UC_MODULE_FIRMWARE(platform__, fw_filename) \ @@ -166,7 +171,7 @@ XE_GUC_FIRMWARE_DEFS(XE_UC_MODULE_FIRMWARE, fw_filename_mmp_ver, fw_filename_major_ver) XE_HUC_FIRMWARE_DEFS(XE_UC_MODULE_FIRMWARE, fw_filename_mmp_ver, fw_filename_no_ver) -XE_GSC_FIRMWARE_DEFS(XE_UC_MODULE_FIRMWARE, fw_filename_major_ver) +XE_GSC_FIRMWARE_DEFS(XE_UC_MODULE_FIRMWARE, fw_filename_gsc) static struct xe_gt * __uc_fw_to_gt(struct xe_uc_fw *uc_fw, enum xe_uc_fw_type type) @@ -209,7 +214,7 @@ uc_fw_auto_select(struct xe_device *xe, struct xe_uc_fw *uc_fw) uc_fw_entry_no_ver) }; static const struct uc_fw_entry entries_gsc[] = { - XE_GSC_FIRMWARE_DEFS(XE_UC_FW_ENTRY, uc_fw_entry_major_ver) + XE_GSC_FIRMWARE_DEFS(XE_UC_FW_ENTRY, uc_fw_entry_gsc) }; static const struct fw_blobs_by_type blobs_all[XE_UC_FW_NUM_TYPES] = { [XE_UC_FW_TYPE_GUC] = { entries_guc, ARRAY_SIZE(entries_guc) }, diff --git a/drivers/gpu/drm/xe/xe_uc_fw.h b/drivers/gpu/drm/xe/xe_uc_fw.h index c108e9d08e70..6195e353f269 100644 --- a/drivers/gpu/drm/xe/xe_uc_fw.h +++ b/drivers/gpu/drm/xe/xe_uc_fw.h @@ -65,7 +65,7 @@ const char *xe_uc_fw_status_repr(enum xe_uc_fw_status status) return "<invalid>"; } -static inline int xe_uc_fw_status_to_error(enum xe_uc_fw_status status) +static inline int xe_uc_fw_status_to_error(const enum xe_uc_fw_status status) { switch (status) { case XE_UC_FIRMWARE_NOT_SUPPORTED: @@ -108,7 +108,7 @@ static inline const char *xe_uc_fw_type_repr(enum xe_uc_fw_type type) } static inline enum xe_uc_fw_status -__xe_uc_fw_status(struct xe_uc_fw *uc_fw) +__xe_uc_fw_status(const struct xe_uc_fw *uc_fw) { /* shouldn't call this before checking hw/blob availability */ XE_WARN_ON(uc_fw->status == XE_UC_FIRMWARE_UNINITIALIZED); @@ -156,6 +156,11 @@ static inline bool xe_uc_fw_is_overridden(const struct xe_uc_fw *uc_fw) return uc_fw->user_overridden; } +static inline bool xe_uc_fw_is_in_error_state(const struct xe_uc_fw *uc_fw) +{ + return xe_uc_fw_status_to_error(__xe_uc_fw_status(uc_fw)) < 0; +} + static inline void xe_uc_fw_sanitize(struct xe_uc_fw *uc_fw) { if (xe_uc_fw_is_loadable(uc_fw)) diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c index 3eb76d874eb2..7acd5fc9d032 100644 --- a/drivers/gpu/drm/xe/xe_vm.c +++ b/drivers/gpu/drm/xe/xe_vm.c @@ -12,7 +12,7 @@ #include <drm/drm_print.h> #include <drm/ttm/ttm_execbuf_util.h> #include <drm/ttm/ttm_tt.h> -#include <drm/xe_drm.h> +#include <uapi/drm/xe_drm.h> #include <linux/ascii85.h> #include <linux/delay.h> #include <linux/kthread.h> diff --git a/drivers/gpu/drm/xe/xe_vram_freq.c b/drivers/gpu/drm/xe/xe_vram_freq.c index 99ff95e408e0..b26e26d73dae 100644 --- a/drivers/gpu/drm/xe/xe_vram_freq.c +++ b/drivers/gpu/drm/xe/xe_vram_freq.c @@ -34,7 +34,6 @@ static ssize_t max_freq_show(struct device *dev, struct device_attribute *attr, char *buf) { struct xe_tile *tile = dev_to_tile(dev); - struct xe_gt *gt = tile->primary_gt; u32 val, mbox; int err; @@ -42,7 +41,7 @@ static ssize_t max_freq_show(struct device *dev, struct device_attribute *attr, | REG_FIELD_PREP(PCODE_MB_PARAM1, PCODE_MBOX_FC_SC_READ_FUSED_P0) | REG_FIELD_PREP(PCODE_MB_PARAM2, PCODE_MBOX_DOMAIN_HBM); - err = xe_pcode_read(gt, mbox, &val, NULL); + err = xe_pcode_read(tile, mbox, &val, NULL); if (err) return err; @@ -57,7 +56,6 @@ static ssize_t min_freq_show(struct device *dev, struct device_attribute *attr, char *buf) { struct xe_tile *tile = dev_to_tile(dev); - struct xe_gt *gt = tile->primary_gt; u32 val, mbox; int err; @@ -65,7 +63,7 @@ static ssize_t min_freq_show(struct device *dev, struct device_attribute *attr, | REG_FIELD_PREP(PCODE_MB_PARAM1, PCODE_MBOX_FC_SC_READ_FUSED_PN) | REG_FIELD_PREP(PCODE_MB_PARAM2, PCODE_MBOX_DOMAIN_HBM); - err = xe_pcode_read(gt, mbox, &val, NULL); + err = xe_pcode_read(tile, mbox, &val, NULL); if (err) return err; diff --git a/drivers/gpu/drm/xe/xe_wait_user_fence.c b/drivers/gpu/drm/xe/xe_wait_user_fence.c index f69721339201..d46fa8374980 100644 --- a/drivers/gpu/drm/xe/xe_wait_user_fence.c +++ b/drivers/gpu/drm/xe/xe_wait_user_fence.c @@ -8,7 +8,7 @@ #include <drm/drm_device.h> #include <drm/drm_file.h> #include <drm/drm_utils.h> -#include <drm/xe_drm.h> +#include <uapi/drm/xe_drm.h> #include "xe_device.h" #include "xe_gt.h" diff --git a/drivers/gpu/drm/xlnx/Kconfig b/drivers/gpu/drm/xlnx/Kconfig index 68ee897de9d7..626e5ac4c33d 100644 --- a/drivers/gpu/drm/xlnx/Kconfig +++ b/drivers/gpu/drm/xlnx/Kconfig @@ -8,6 +8,7 @@ config DRM_ZYNQMP_DPSUB select DMA_ENGINE select DRM_DISPLAY_DP_HELPER select DRM_DISPLAY_HELPER + select DRM_BRIDGE_CONNECTOR select DRM_GEM_DMA_HELPER select DRM_KMS_HELPER select GENERIC_PHY |