diff options
110 files changed, 13917 insertions, 9089 deletions
diff --git a/arch/powerpc/platforms/cell/spufs/sched.c b/arch/powerpc/platforms/cell/spufs/sched.c index 965d381abd75..25db92a8e1cf 100644 --- a/arch/powerpc/platforms/cell/spufs/sched.c +++ b/arch/powerpc/platforms/cell/spufs/sched.c @@ -1094,7 +1094,7 @@ static int show_spu_loadavg(struct seq_file *s, void *private) LOAD_INT(c), LOAD_FRAC(c), count_active_contexts(), atomic_read(&nr_spu_contexts), - current->nsproxy->pid_ns->last_pid); + task_active_pid_ns(current)->last_pid); return 0; } diff --git a/arch/um/drivers/mconsole_kern.c b/arch/um/drivers/mconsole_kern.c index 49e3b49e552f..4bd82ac0210f 100644 --- a/arch/um/drivers/mconsole_kern.c +++ b/arch/um/drivers/mconsole_kern.c @@ -123,7 +123,7 @@ void mconsole_log(struct mc_request *req) void mconsole_proc(struct mc_request *req) { - struct vfsmount *mnt = current->nsproxy->pid_ns->proc_mnt; + struct vfsmount *mnt = task_active_pid_ns(current)->proc_mnt; char *buf; int len; struct file *file; diff --git a/block/genhd.c b/block/genhd.c index 2a6fdf539a69..9a289d7c84bb 100644 --- a/block/genhd.c +++ b/block/genhd.c @@ -743,7 +743,6 @@ void __init printk_all_partitions(void) struct hd_struct *part; char name_buf[BDEVNAME_SIZE]; char devt_buf[BDEVT_SIZE]; - char uuid_buf[PARTITION_META_INFO_UUIDLTH * 2 + 5]; /* * Don't show empty devices or things that have been @@ -762,16 +761,11 @@ void __init printk_all_partitions(void) while ((part = disk_part_iter_next(&piter))) { bool is_part0 = part == &disk->part0; - uuid_buf[0] = '\0'; - if (part->info) - snprintf(uuid_buf, sizeof(uuid_buf), "%pU", - part->info->uuid); - printk("%s%s %10llu %s %s", is_part0 ? "" : " ", bdevt_str(part_devt(part), devt_buf), (unsigned long long)part_nr_sects_read(part) >> 1 , disk_name(disk, part->partno, name_buf), - uuid_buf); + part->info ? part->info->uuid : ""); if (is_part0) { if (disk->driverfs_dev != NULL && disk->driverfs_dev->driver != NULL) diff --git a/block/partitions/efi.c b/block/partitions/efi.c index 6296b403c67a..b62fb88b8711 100644 --- a/block/partitions/efi.c +++ b/block/partitions/efi.c @@ -620,7 +620,6 @@ int efi_partition(struct parsed_partitions *state) gpt_entry *ptes = NULL; u32 i; unsigned ssz = bdev_logical_block_size(state->bdev) / 512; - u8 unparsed_guid[37]; if (!find_valid_gpt(state, &gpt, &ptes) || !gpt || !ptes) { kfree(gpt); @@ -649,11 +648,7 @@ int efi_partition(struct parsed_partitions *state) state->parts[i + 1].flags = ADDPART_FLAG_RAID; info = &state->parts[i + 1].info; - /* Instead of doing a manual swap to big endian, reuse the - * common ASCII hex format as the interim. - */ - efi_guid_unparse(&ptes[i].unique_partition_guid, unparsed_guid); - part_pack_uuid(unparsed_guid, info->uuid); + efi_guid_unparse(&ptes[i].unique_partition_guid, info->uuid); /* Naively convert UTF16-LE to 7 bits. */ label_max = min(sizeof(info->volname) - 1, diff --git a/block/partitions/msdos.c b/block/partitions/msdos.c index 5f79a6677c69..8752a5d26565 100644 --- a/block/partitions/msdos.c +++ b/block/partitions/msdos.c @@ -94,6 +94,17 @@ static int aix_magic_present(struct parsed_partitions *state, unsigned char *p) return ret; } +static void set_info(struct parsed_partitions *state, int slot, + u32 disksig) +{ + struct partition_meta_info *info = &state->parts[slot].info; + + snprintf(info->uuid, sizeof(info->uuid), "%08x-%02x", disksig, + slot); + info->volname[0] = 0; + state->parts[slot].has_info = true; +} + /* * Create devices for each logical partition in an extended partition. * The logical partitions form a linked list, with each entry being @@ -106,7 +117,8 @@ static int aix_magic_present(struct parsed_partitions *state, unsigned char *p) */ static void parse_extended(struct parsed_partitions *state, - sector_t first_sector, sector_t first_size) + sector_t first_sector, sector_t first_size, + u32 disksig) { struct partition *p; Sector sect; @@ -166,6 +178,7 @@ static void parse_extended(struct parsed_partitions *state, } put_partition(state, state->next, next, size); + set_info(state, state->next, disksig); if (SYS_IND(p) == LINUX_RAID_PARTITION) state->parts[state->next].flags = ADDPART_FLAG_RAID; loopct = 0; @@ -437,6 +450,7 @@ int msdos_partition(struct parsed_partitions *state) struct partition *p; struct fat_boot_sector *fb; int slot; + u32 disksig; data = read_part_sector(state, 0, §); if (!data) @@ -491,6 +505,8 @@ int msdos_partition(struct parsed_partitions *state) #endif p = (struct partition *) (data + 0x1be); + disksig = le32_to_cpup((__le32 *)(data + 0x1b8)); + /* * Look for partitions in two passes: * First find the primary and DOS-type extended partitions. @@ -515,11 +531,12 @@ int msdos_partition(struct parsed_partitions *state) put_partition(state, slot, start, n); strlcat(state->pp_buf, " <", PAGE_SIZE); - parse_extended(state, start, size); + parse_extended(state, start, size, disksig); strlcat(state->pp_buf, " >", PAGE_SIZE); continue; } put_partition(state, slot, start, size); + set_info(state, slot, disksig); if (SYS_IND(p) == LINUX_RAID_PARTITION) state->parts[slot].flags = ADDPART_FLAG_RAID; if (SYS_IND(p) == DM6_PARTITION) diff --git a/drivers/block/cciss.c b/drivers/block/cciss.c index ca83f96756ad..6526157edafc 100644 --- a/drivers/block/cciss.c +++ b/drivers/block/cciss.c @@ -41,8 +41,9 @@ #include <linux/spinlock.h> #include <linux/compat.h> #include <linux/mutex.h> +#include <linux/bitmap.h> +#include <linux/io.h> #include <asm/uaccess.h> -#include <asm/io.h> #include <linux/dma-mapping.h> #include <linux/blkdev.h> @@ -978,8 +979,7 @@ static CommandList_struct *cmd_alloc(ctlr_info_t *h) i = find_first_zero_bit(h->cmd_pool_bits, h->nr_cmds); if (i == h->nr_cmds) return NULL; - } while (test_and_set_bit(i & (BITS_PER_LONG - 1), - h->cmd_pool_bits + (i / BITS_PER_LONG)) != 0); + } while (test_and_set_bit(i, h->cmd_pool_bits) != 0); c = h->cmd_pool + i; memset(c, 0, sizeof(CommandList_struct)); cmd_dma_handle = h->cmd_pool_dhandle + i * sizeof(CommandList_struct); @@ -1046,8 +1046,7 @@ static void cmd_free(ctlr_info_t *h, CommandList_struct *c) int i; i = c - h->cmd_pool; - clear_bit(i & (BITS_PER_LONG - 1), - h->cmd_pool_bits + (i / BITS_PER_LONG)); + clear_bit(i, h->cmd_pool_bits); h->nr_frees++; } @@ -4268,10 +4267,7 @@ static void __devinit cciss_find_board_params(ctlr_info_t *h) static inline bool CISS_signature_present(ctlr_info_t *h) { - if ((readb(&h->cfgtable->Signature[0]) != 'C') || - (readb(&h->cfgtable->Signature[1]) != 'I') || - (readb(&h->cfgtable->Signature[2]) != 'S') || - (readb(&h->cfgtable->Signature[3]) != 'S')) { + if (!check_signature(h->cfgtable->Signature, "CISS", 4)) { dev_warn(&h->pdev->dev, "not a valid CISS config table\n"); return false; } @@ -4812,8 +4808,7 @@ static __devinit int cciss_init_reset_devices(struct pci_dev *pdev) static __devinit int cciss_allocate_cmd_pool(ctlr_info_t *h) { - h->cmd_pool_bits = kmalloc( - DIV_ROUND_UP(h->nr_cmds, BITS_PER_LONG) * + h->cmd_pool_bits = kmalloc(BITS_TO_LONGS(h->nr_cmds) * sizeof(unsigned long), GFP_KERNEL); h->cmd_pool = pci_alloc_consistent(h->pdev, h->nr_cmds * sizeof(CommandList_struct), @@ -5068,9 +5063,7 @@ reinit_after_soft_reset: pci_set_drvdata(pdev, h); /* command and error info recs zeroed out before they are used */ - memset(h->cmd_pool_bits, 0, - DIV_ROUND_UP(h->nr_cmds, BITS_PER_LONG) - * sizeof(unsigned long)); + bitmap_zero(h->cmd_pool_bits, h->nr_cmds); h->num_luns = 0; h->highest_lun = -1; diff --git a/drivers/block/drbd/Kconfig b/drivers/block/drbd/Kconfig index df0983787390..7845bd6ee414 100644 --- a/drivers/block/drbd/Kconfig +++ b/drivers/block/drbd/Kconfig @@ -2,13 +2,14 @@ # DRBD device driver configuration # -comment "DRBD disabled because PROC_FS, INET or CONNECTOR not selected" - depends on PROC_FS='n' || INET='n' || CONNECTOR='n' +comment "DRBD disabled because PROC_FS or INET not selected" + depends on PROC_FS='n' || INET='n' config BLK_DEV_DRBD tristate "DRBD Distributed Replicated Block Device support" - depends on PROC_FS && INET && CONNECTOR + depends on PROC_FS && INET select LRU_CACHE + select LIBCRC32C default n help @@ -58,7 +59,8 @@ config DRBD_FAULT_INJECTION 32 data read 64 read ahead 128 kmalloc of bitmap - 256 allocation of EE (epoch_entries) + 256 allocation of peer_requests + 512 insert data corruption on receiving side fault_devs: bitmask of minor numbers fault_rate: frequency in percent diff --git a/drivers/block/drbd/Makefile b/drivers/block/drbd/Makefile index 0d3f337ff5ff..8b450338075e 100644 --- a/drivers/block/drbd/Makefile +++ b/drivers/block/drbd/Makefile @@ -1,5 +1,7 @@ drbd-y := drbd_bitmap.o drbd_proc.o drbd-y += drbd_worker.o drbd_receiver.o drbd_req.o drbd_actlog.o drbd-y += drbd_main.o drbd_strings.o drbd_nl.o +drbd-y += drbd_interval.o drbd_state.o +drbd-y += drbd_nla.o obj-$(CONFIG_BLK_DEV_DRBD) += drbd.o diff --git a/drivers/block/drbd/drbd_actlog.c b/drivers/block/drbd/drbd_actlog.c index 3fbef018ce55..92510f8ad013 100644 --- a/drivers/block/drbd/drbd_actlog.c +++ b/drivers/block/drbd/drbd_actlog.c @@ -24,21 +24,73 @@ */ #include <linux/slab.h> +#include <linux/crc32c.h> #include <linux/drbd.h> +#include <linux/drbd_limits.h> +#include <linux/dynamic_debug.h> #include "drbd_int.h" #include "drbd_wrappers.h" -/* We maintain a trivial checksum in our on disk activity log. - * With that we can ensure correct operation even when the storage - * device might do a partial (last) sector write while losing power. - */ -struct __packed al_transaction { - u32 magic; - u32 tr_number; - struct __packed { - u32 pos; - u32 extent; } updates[1 + AL_EXTENTS_PT]; - u32 xor_sum; + +enum al_transaction_types { + AL_TR_UPDATE = 0, + AL_TR_INITIALIZED = 0xffff +}; +/* all fields on disc in big endian */ +struct __packed al_transaction_on_disk { + /* don't we all like magic */ + __be32 magic; + + /* to identify the most recent transaction block + * in the on disk ring buffer */ + __be32 tr_number; + + /* checksum on the full 4k block, with this field set to 0. */ + __be32 crc32c; + + /* type of transaction, special transaction types like: + * purge-all, set-all-idle, set-all-active, ... to-be-defined + * see also enum al_transaction_types */ + __be16 transaction_type; + + /* we currently allow only a few thousand extents, + * so 16bit will be enough for the slot number. */ + + /* how many updates in this transaction */ + __be16 n_updates; + + /* maximum slot number, "al-extents" in drbd.conf speak. + * Having this in each transaction should make reconfiguration + * of that parameter easier. */ + __be16 context_size; + + /* slot number the context starts with */ + __be16 context_start_slot_nr; + + /* Some reserved bytes. Expected usage is a 64bit counter of + * sectors-written since device creation, and other data generation tag + * supporting usage */ + __be32 __reserved[4]; + + /* --- 36 byte used --- */ + + /* Reserve space for up to AL_UPDATES_PER_TRANSACTION changes + * in one transaction, then use the remaining byte in the 4k block for + * context information. "Flexible" number of updates per transaction + * does not help, as we have to account for the case when all update + * slots are used anyways, so it would only complicate code without + * additional benefit. + */ + __be16 update_slot_nr[AL_UPDATES_PER_TRANSACTION]; + + /* but the extent number is 32bit, which at an extent size of 4 MiB + * allows to cover device sizes of up to 2**54 Byte (16 PiB) */ + __be32 update_extent_nr[AL_UPDATES_PER_TRANSACTION]; + + /* --- 420 bytes used (36 + 64*6) --- */ + + /* 4096 - 420 = 3676 = 919 * 4 */ + __be32 context[AL_CONTEXT_PER_TRANSACTION]; }; struct update_odbm_work { @@ -48,22 +100,11 @@ struct update_odbm_work { struct update_al_work { struct drbd_work w; - struct lc_element *al_ext; struct completion event; - unsigned int enr; - /* if old_enr != LC_FREE, write corresponding bitmap sector, too */ - unsigned int old_enr; -}; - -struct drbd_atodb_wait { - atomic_t count; - struct completion io_done; - struct drbd_conf *mdev; - int error; + int err; }; - -int w_al_write_transaction(struct drbd_conf *, struct drbd_work *, int); +static int al_write_transaction(struct drbd_conf *mdev); void *drbd_md_get_buffer(struct drbd_conf *mdev) { @@ -82,22 +123,24 @@ void drbd_md_put_buffer(struct drbd_conf *mdev) wake_up(&mdev->misc_wait); } -static bool md_io_allowed(struct drbd_conf *mdev) -{ - enum drbd_disk_state ds = mdev->state.disk; - return ds >= D_NEGOTIATING || ds == D_ATTACHING; -} - -void wait_until_done_or_disk_failure(struct drbd_conf *mdev, struct drbd_backing_dev *bdev, +void wait_until_done_or_force_detached(struct drbd_conf *mdev, struct drbd_backing_dev *bdev, unsigned int *done) { - long dt = bdev->dc.disk_timeout * HZ / 10; + long dt; + + rcu_read_lock(); + dt = rcu_dereference(bdev->disk_conf)->disk_timeout; + rcu_read_unlock(); + dt = dt * HZ / 10; if (dt == 0) dt = MAX_SCHEDULE_TIMEOUT; - dt = wait_event_timeout(mdev->misc_wait, *done || !md_io_allowed(mdev), dt); - if (dt == 0) + dt = wait_event_timeout(mdev->misc_wait, + *done || test_bit(FORCE_DETACH, &mdev->flags), dt); + if (dt == 0) { dev_err(DEV, "meta-data IO operation timed out\n"); + drbd_chk_io_error(mdev, 1, DRBD_FORCE_DETACH); + } } static int _drbd_md_sync_page_io(struct drbd_conf *mdev, @@ -106,7 +149,7 @@ static int _drbd_md_sync_page_io(struct drbd_conf *mdev, int rw, int size) { struct bio *bio; - int ok; + int err; mdev->md_io.done = 0; mdev->md_io.error = -ENODEV; @@ -118,8 +161,8 @@ static int _drbd_md_sync_page_io(struct drbd_conf *mdev, bio = bio_alloc_drbd(GFP_NOIO); bio->bi_bdev = bdev->md_bdev; bio->bi_sector = sector; - ok = (bio_add_page(bio, page, size, 0) == size); - if (!ok) + err = -EIO; + if (bio_add_page(bio, page, size, 0) != size) goto out; bio->bi_private = &mdev->md_io; bio->bi_end_io = drbd_md_io_complete; @@ -127,7 +170,7 @@ static int _drbd_md_sync_page_io(struct drbd_conf *mdev, if (!get_ldev_if_state(mdev, D_ATTACHING)) { /* Corresponding put_ldev in drbd_md_io_complete() */ dev_err(DEV, "ASSERT FAILED: get_ldev_if_state() == 1 in _drbd_md_sync_page_io()\n"); - ok = 0; + err = -ENODEV; goto out; } @@ -137,86 +180,47 @@ static int _drbd_md_sync_page_io(struct drbd_conf *mdev, bio_endio(bio, -EIO); else submit_bio(rw, bio); - wait_until_done_or_disk_failure(mdev, bdev, &mdev->md_io.done); - ok = bio_flagged(bio, BIO_UPTODATE) && mdev->md_io.error == 0; + wait_until_done_or_force_detached(mdev, bdev, &mdev->md_io.done); + if (bio_flagged(bio, BIO_UPTODATE)) + err = mdev->md_io.error; out: bio_put(bio); - return ok; + return err; } int drbd_md_sync_page_io(struct drbd_conf *mdev, struct drbd_backing_dev *bdev, sector_t sector, int rw) { - int logical_block_size, mask, ok; - int offset = 0; + int err; struct page *iop = mdev->md_io_page; D_ASSERT(atomic_read(&mdev->md_io_in_use) == 1); BUG_ON(!bdev->md_bdev); - logical_block_size = bdev_logical_block_size(bdev->md_bdev); - if (logical_block_size == 0) - logical_block_size = MD_SECTOR_SIZE; - - /* in case logical_block_size != 512 [ s390 only? ] */ - if (logical_block_size != MD_SECTOR_SIZE) { - mask = (logical_block_size / MD_SECTOR_SIZE) - 1; - D_ASSERT(mask == 1 || mask == 3 || mask == 7); - D_ASSERT(logical_block_size == (mask+1) * MD_SECTOR_SIZE); - offset = sector & mask; - sector = sector & ~mask; - iop = mdev->md_io_tmpp; - - if (rw & WRITE) { - /* these are GFP_KERNEL pages, pre-allocated - * on device initialization */ - void *p = page_address(mdev->md_io_page); - void *hp = page_address(mdev->md_io_tmpp); - - ok = _drbd_md_sync_page_io(mdev, bdev, iop, sector, - READ, logical_block_size); - - if (unlikely(!ok)) { - dev_err(DEV, "drbd_md_sync_page_io(,%llus," - "READ [logical_block_size!=512]) failed!\n", - (unsigned long long)sector); - return 0; - } - - memcpy(hp + offset*MD_SECTOR_SIZE, p, MD_SECTOR_SIZE); - } - } + dev_dbg(DEV, "meta_data io: %s [%d]:%s(,%llus,%s)\n", + current->comm, current->pid, __func__, + (unsigned long long)sector, (rw & WRITE) ? "WRITE" : "READ"); if (sector < drbd_md_first_sector(bdev) || - sector > drbd_md_last_sector(bdev)) + sector + 7 > drbd_md_last_sector(bdev)) dev_alert(DEV, "%s [%d]:%s(,%llus,%s) out of range md access!\n", current->comm, current->pid, __func__, (unsigned long long)sector, (rw & WRITE) ? "WRITE" : "READ"); - ok = _drbd_md_sync_page_io(mdev, bdev, iop, sector, rw, logical_block_size); - if (unlikely(!ok)) { - dev_err(DEV, "drbd_md_sync_page_io(,%llus,%s) failed!\n", - (unsigned long long)sector, (rw & WRITE) ? "WRITE" : "READ"); - return 0; - } - - if (logical_block_size != MD_SECTOR_SIZE && !(rw & WRITE)) { - void *p = page_address(mdev->md_io_page); - void *hp = page_address(mdev->md_io_tmpp); - - memcpy(p, hp + offset*MD_SECTOR_SIZE, MD_SECTOR_SIZE); + err = _drbd_md_sync_page_io(mdev, bdev, iop, sector, rw, MD_BLOCK_SIZE); + if (err) { + dev_err(DEV, "drbd_md_sync_page_io(,%llus,%s) failed with error %d\n", + (unsigned long long)sector, (rw & WRITE) ? "WRITE" : "READ", err); } - - return ok; + return err; } static struct lc_element *_al_get(struct drbd_conf *mdev, unsigned int enr) { struct lc_element *al_ext; struct lc_element *tmp; - unsigned long al_flags = 0; int wake; spin_lock_irq(&mdev->al_lock); @@ -231,76 +235,92 @@ static struct lc_element *_al_get(struct drbd_conf *mdev, unsigned int enr) return NULL; } } - al_ext = lc_get(mdev->act_log, enr); - al_flags = mdev->act_log->flags; + al_ext = lc_get(mdev->act_log, enr); spin_unlock_irq(&mdev->al_lock); - - /* - if (!al_ext) { - if (al_flags & LC_STARVING) - dev_warn(DEV, "Have to wait for LRU element (AL too small?)\n"); - if (al_flags & LC_DIRTY) - dev_warn(DEV, "Ongoing AL update (AL device too slow?)\n"); - } - */ - return al_ext; } -void drbd_al_begin_io(struct drbd_conf *mdev, sector_t sector) +void drbd_al_begin_io(struct drbd_conf *mdev, struct drbd_interval *i) { - unsigned int enr = (sector >> (AL_EXTENT_SHIFT-9)); - struct lc_element *al_ext; - struct update_al_work al_work; + /* for bios crossing activity log extent boundaries, + * we may need to activate two extents in one go */ + unsigned first = i->sector >> (AL_EXTENT_SHIFT-9); + unsigned last = i->size == 0 ? first : (i->sector + (i->size >> 9) - 1) >> (AL_EXTENT_SHIFT-9); + unsigned enr; + bool locked = false; + + D_ASSERT(first <= last); D_ASSERT(atomic_read(&mdev->local_cnt) > 0); - wait_event(mdev->al_wait, (al_ext = _al_get(mdev, enr))); + for (enr = first; enr <= last; enr++) + wait_event(mdev->al_wait, _al_get(mdev, enr) != NULL); + + /* Serialize multiple transactions. + * This uses test_and_set_bit, memory barrier is implicit. + */ + wait_event(mdev->al_wait, + mdev->act_log->pending_changes == 0 || + (locked = lc_try_lock_for_transaction(mdev->act_log))); - if (al_ext->lc_number != enr) { + if (locked) { /* drbd_al_write_transaction(mdev,al_ext,enr); * recurses into generic_make_request(), which * disallows recursion, bios being serialized on the * current->bio_tail list now. * we have to delegate updates to the activity log * to the worker thread. */ - init_completion(&al_work.event); - al_work.al_ext = al_ext; - al_work.enr = enr; - al_work.old_enr = al_ext->lc_number; - al_work.w.cb = w_al_write_transaction; - drbd_queue_work_front(&mdev->data.work, &al_work.w); - wait_for_completion(&al_work.event); - - mdev->al_writ_cnt++; - - spin_lock_irq(&mdev->al_lock); - lc_changed(mdev->act_log, al_ext); - spin_unlock_irq(&mdev->al_lock); + + /* Double check: it may have been committed by someone else, + * while we have been waiting for the lock. */ + if (mdev->act_log->pending_changes) { + bool write_al_updates; + + rcu_read_lock(); + write_al_updates = rcu_dereference(mdev->ldev->disk_conf)->al_updates; + rcu_read_unlock(); + + if (write_al_updates) { + al_write_transaction(mdev); + mdev->al_writ_cnt++; + } + + spin_lock_irq(&mdev->al_lock); + /* FIXME + if (err) + we need an "lc_cancel" here; + */ + lc_committed(mdev->act_log); + spin_unlock_irq(&mdev->al_lock); + } + lc_unlock(mdev->act_log); wake_up(&mdev->al_wait); } } -void drbd_al_complete_io(struct drbd_conf *mdev, sector_t sector) +void drbd_al_complete_io(struct drbd_conf *mdev, struct drbd_interval *i) { - unsigned int enr = (sector >> (AL_EXTENT_SHIFT-9)); + /* for bios crossing activity log extent boundaries, + * we may need to activate two extents in one go */ + unsigned first = i->sector >> (AL_EXTENT_SHIFT-9); + unsigned last = i->size == 0 ? first : (i->sector + (i->size >> 9) - 1) >> (AL_EXTENT_SHIFT-9); + unsigned enr; struct lc_element *extent; unsigned long flags; + D_ASSERT(first <= last); spin_lock_irqsave(&mdev->al_lock, flags); - extent = lc_find(mdev->act_log, enr); - - if (!extent) { - spin_unlock_irqrestore(&mdev->al_lock, flags); - dev_err(DEV, "al_complete_io() called on inactive extent %u\n", enr); - return; + for (enr = first; enr <= last; enr++) { + extent = lc_find(mdev->act_log, enr); + if (!extent) { + dev_err(DEV, "al_complete_io() called on inactive extent %u\n", enr); + continue; + } + lc_put(mdev->act_log, extent); } - - if (lc_put(mdev->act_log, extent) == 0) - wake_up(&mdev->al_wait); - spin_unlock_irqrestore(&mdev->al_lock, flags); + wake_up(&mdev->al_wait); } #if (PAGE_SHIFT + 3) < (AL_EXTENT_SHIFT - BM_BLOCK_SHIFT) @@ -326,296 +346,148 @@ static unsigned int rs_extent_to_bm_page(unsigned int rs_enr) return rs_enr >> /* bit to page */ ((PAGE_SHIFT + 3) - - /* al extent number to bit */ + /* resync extent number to bit */ (BM_EXT_SHIFT - BM_BLOCK_SHIFT)); } -int -w_al_write_transaction(struct drbd_conf *mdev, struct drbd_work *w, int unused) +static int +_al_write_transaction(struct drbd_conf *mdev) { - struct update_al_work *aw = container_of(w, struct update_al_work, w); - struct lc_element *updated = aw->al_ext; - const unsigned int new_enr = aw->enr; - const unsigned int evicted = aw->old_enr; - struct al_transaction *buffer; + struct al_transaction_on_disk *buffer; + struct lc_element *e; sector_t sector; - int i, n, mx; - unsigned int extent_nr; - u32 xor_sum = 0; + int i, mx; + unsigned extent_nr; + unsigned crc = 0; + int err = 0; if (!get_ldev(mdev)) { - dev_err(DEV, - "disk is %s, cannot start al transaction (-%d +%d)\n", - drbd_disk_str(mdev->state.disk), evicted, new_enr); - complete(&((struct update_al_work *)w)->event); - return 1; + dev_err(DEV, "disk is %s, cannot start al transaction\n", + drbd_disk_str(mdev->state.disk)); + return -EIO; } - /* do we have to do a bitmap write, first? - * TODO reduce maximum latency: - * submit both bios, then wait for both, - * instead of doing two synchronous sector writes. - * For now, we must not write the transaction, - * if we cannot write out the bitmap of the evicted extent. */ - if (mdev->state.conn < C_CONNECTED && evicted != LC_FREE) - drbd_bm_write_page(mdev, al_extent_to_bm_page(evicted)); /* The bitmap write may have failed, causing a state change. */ if (mdev->state.disk < D_INCONSISTENT) { dev_err(DEV, - "disk is %s, cannot write al transaction (-%d +%d)\n", - drbd_disk_str(mdev->state.disk), evicted, new_enr); - complete(&((struct update_al_work *)w)->event); + "disk is %s, cannot write al transaction\n", + drbd_disk_str(mdev->state.disk)); put_ldev(mdev); - return 1; + return -EIO; } buffer = drbd_md_get_buffer(mdev); /* protects md_io_buffer, al_tr_cycle, ... */ if (!buffer) { dev_err(DEV, "disk failed while waiting for md_io buffer\n"); - complete(&((struct update_al_work *)w)->event); put_ldev(mdev); - return 1; + return -ENODEV; } - buffer->magic = __constant_cpu_to_be32(DRBD_MAGIC); + memset(buffer, 0, sizeof(*buffer)); + buffer->magic = cpu_to_be32(DRBD_AL_MAGIC); buffer->tr_number = cpu_to_be32(mdev->al_tr_number); - n = lc_index_of(mdev->act_log, updated); + i = 0; + + /* Even though no one can start to change this list + * once we set the LC_LOCKED -- from drbd_al_begin_io(), + * lc_try_lock_for_transaction() --, someone may still + * be in the process of changing it. */ + spin_lock_irq(&mdev->al_lock); + list_for_each_entry(e, &mdev->act_log->to_be_changed, list) { + if (i == AL_UPDATES_PER_TRANSACTION) { + i++; + break; + } + buffer->update_slot_nr[i] = cpu_to_be16(e->lc_index); + buffer->update_extent_nr[i] = cpu_to_be32(e->lc_new_number); + if (e->lc_number != LC_FREE) + drbd_bm_mark_for_writeout(mdev, + al_extent_to_bm_page(e->lc_number)); + i++; + } + spin_unlock_irq(&mdev->al_lock); + BUG_ON(i > AL_UPDATES_PER_TRANSACTION); - buffer->updates[0].pos = cpu_to_be32(n); - buffer->updates[0].extent = cpu_to_be32(new_enr); + buffer->n_updates = cpu_to_be16(i); + for ( ; i < AL_UPDATES_PER_TRANSACTION; i++) { + buffer->update_slot_nr[i] = cpu_to_be16(-1); + buffer->update_extent_nr[i] = cpu_to_be32(LC_FREE); + } - xor_sum ^= new_enr; + buffer->context_size = cpu_to_be16(mdev->act_log->nr_elements); + buffer->context_start_slot_nr = cpu_to_be16(mdev->al_tr_cycle); - mx = min_t(int, AL_EXTENTS_PT, + mx = min_t(int, AL_CONTEXT_PER_TRANSACTION, mdev->act_log->nr_elements - mdev->al_tr_cycle); for (i = 0; i < mx; i++) { unsigned idx = mdev->al_tr_cycle + i; extent_nr = lc_element_by_index(mdev->act_log, idx)->lc_number; - buffer->updates[i+1].pos = cpu_to_be32(idx); - buffer->updates[i+1].extent = cpu_to_be32(extent_nr); - xor_sum ^= extent_nr; - } - for (; i < AL_EXTENTS_PT; i++) { - buffer->updates[i+1].pos = __constant_cpu_to_be32(-1); - buffer->updates[i+1].extent = __constant_cpu_to_be32(LC_FREE); - xor_sum ^= LC_FREE; + buffer->context[i] = cpu_to_be32(extent_nr); } - mdev->al_tr_cycle += AL_EXTENTS_PT; + for (; i < AL_CONTEXT_PER_TRANSACTION; i++) + buffer->context[i] = cpu_to_be32(LC_FREE); + + mdev->al_tr_cycle += AL_CONTEXT_PER_TRANSACTION; if (mdev->al_tr_cycle >= mdev->act_log->nr_elements) mdev->al_tr_cycle = 0; - buffer->xor_sum = cpu_to_be32(xor_sum); - sector = mdev->ldev->md.md_offset - + mdev->ldev->md.al_offset + mdev->al_tr_pos; - - if (!drbd_md_sync_page_io(mdev, mdev->ldev, sector, WRITE)) - drbd_chk_io_error(mdev, 1, DRBD_META_IO_ERROR); + + mdev->ldev->md.al_offset + + mdev->al_tr_pos * (MD_BLOCK_SIZE>>9); - if (++mdev->al_tr_pos > - div_ceil(mdev->act_log->nr_elements, AL_EXTENTS_PT)) - mdev->al_tr_pos = 0; + crc = crc32c(0, buffer, 4096); + buffer->crc32c = cpu_to_be32(crc); - D_ASSERT(mdev->al_tr_pos < MD_AL_MAX_SIZE); - mdev->al_tr_number++; + if (drbd_bm_write_hinted(mdev)) + err = -EIO; + /* drbd_chk_io_error done already */ + else if (drbd_md_sync_page_io(mdev, mdev->ldev, sector, WRITE)) { + err = -EIO; + drbd_chk_io_error(mdev, 1, DRBD_META_IO_ERROR); + } else { + /* advance ringbuffer position and transaction counter */ + mdev->al_tr_pos = (mdev->al_tr_pos + 1) % (MD_AL_SECTORS*512/MD_BLOCK_SIZE); + mdev->al_tr_number++; + } drbd_md_put_buffer(mdev); - - complete(&((struct update_al_work *)w)->event); put_ldev(mdev); - return 1; + return err; } -/** - * drbd_al_read_tr() - Read a single transaction from the on disk activity log - * @mdev: DRBD device. - * @bdev: Block device to read form. - * @b: pointer to an al_transaction. - * @index: On disk slot of the transaction to read. - * - * Returns -1 on IO error, 0 on checksum error and 1 upon success. - */ -static int drbd_al_read_tr(struct drbd_conf *mdev, - struct drbd_backing_dev *bdev, - struct al_transaction *b, - int index) -{ - sector_t sector; - int rv, i; - u32 xor_sum = 0; - - sector = bdev->md.md_offset + bdev->md.al_offset + index; - - /* Dont process error normally, - * as this is done before disk is attached! */ - if (!drbd_md_sync_page_io(mdev, bdev, sector, READ)) - return -1; - - rv = (be32_to_cpu(b->magic) == DRBD_MAGIC); - - for (i = 0; i < AL_EXTENTS_PT + 1; i++) - xor_sum ^= be32_to_cpu(b->updates[i].extent); - rv &= (xor_sum == be32_to_cpu(b->xor_sum)); - return rv; -} - -/** - * drbd_al_read_log() - Restores the activity log from its on disk representation. - * @mdev: DRBD device. - * @bdev: Block device to read form. - * - * Returns 1 on success, returns 0 when reading the log failed due to IO errors. - */ -int drbd_al_read_log(struct drbd_conf *mdev, struct drbd_backing_dev *bdev) +static int w_al_write_transaction(struct drbd_work *w, int unused) { - struct al_transaction *buffer; - int i; - int rv; - int mx; - int active_extents = 0; - int transactions = 0; - int found_valid = 0; - int from = 0; - int to = 0; - u32 from_tnr = 0; - u32 to_tnr = 0; - u32 cnr; - - mx = div_ceil(mdev->act_log->nr_elements, AL_EXTENTS_PT); - - /* lock out all other meta data io for now, - * and make sure the page is mapped. - */ - buffer = drbd_md_get_buffer(mdev); - if (!buffer) - return 0; - - /* Find the valid transaction in the log */ - for (i = 0; i <= mx; i++) { - rv = drbd_al_read_tr(mdev, bdev, buffer, i); - if (rv == 0) - continue; - if (rv == -1) { - drbd_md_put_buffer(mdev); - return 0; - } - cnr = be32_to_cpu(buffer->tr_number); - - if (++found_valid == 1) { - from = i; - to = i; - from_tnr = cnr; - to_tnr = cnr; - continue; - } - if ((int)cnr - (int)from_tnr < 0) { - D_ASSERT(from_tnr - cnr + i - from == mx+1); - from = i; - from_tnr = cnr; - } - if ((int)cnr - (int)to_tnr > 0) { - D_ASSERT(cnr - to_tnr == i - to); - to = i; - to_tnr = cnr; - } - } - - if (!found_valid) { - dev_warn(DEV, "No usable activity log found.\n"); - drbd_md_put_buffer(mdev); - return 1; - } - - /* Read the valid transactions. - * dev_info(DEV, "Reading from %d to %d.\n",from,to); */ - i = from; - while (1) { - int j, pos; - unsigned int extent_nr; - unsigned int trn; - - rv = drbd_al_read_tr(mdev, bdev, buffer, i); - ERR_IF(rv == 0) goto cancel; - if (rv == -1) { - drbd_md_put_buffer(mdev); - return 0; - } - - trn = be32_to_cpu(buffer->tr_number); - - spin_lock_irq(&mdev->al_lock); - - /* This loop runs backwards because in the cyclic - elements there might be an old version of the - updated element (in slot 0). So the element in slot 0 - can overwrite old versions. */ - for (j = AL_EXTENTS_PT; j >= 0; j--) { - pos = be32_to_cpu(buffer->updates[j].pos); - extent_nr = be32_to_cpu(buffer->updates[j].extent); - - if (extent_nr == LC_FREE) - continue; - - lc_set(mdev->act_log, extent_nr, pos); - active_extents++; - } - spin_unlock_irq(&mdev->al_lock); - - transactions++; - -cancel: - if (i == to) - break; - i++; - if (i > mx) - i = 0; - } - - mdev->al_tr_number = to_tnr+1; - mdev->al_tr_pos = to; - if (++mdev->al_tr_pos > - div_ceil(mdev->act_log->nr_elements, AL_EXTENTS_PT)) - mdev->al_tr_pos = 0; - - /* ok, we are done with it */ - drbd_md_put_buffer(mdev); + struct update_al_work *aw = container_of(w, struct update_al_work, w); + struct drbd_conf *mdev = w->mdev; + int err; - dev_info(DEV, "Found %d transactions (%d active extents) in activity log.\n", - transactions, active_extents); + err = _al_write_transaction(mdev); + aw->err = err; + complete(&aw->event); - return 1; + return err != -EIO ? err : 0; } -/** - * drbd_al_apply_to_bm() - Sets the bitmap to diry(1) where covered ba active AL extents - * @mdev: DRBD device. - */ -void drbd_al_apply_to_bm(struct drbd_conf *mdev) +/* Calls from worker context (see w_restart_disk_io()) need to write the + transaction directly. Others came through generic_make_request(), + those need to delegate it to the worker. */ +static int al_write_transaction(struct drbd_conf *mdev) { - unsigned int enr; - unsigned long add = 0; - char ppb[10]; - int i, tmp; - - wait_event(mdev->al_wait, lc_try_lock(mdev->act_log)); + struct update_al_work al_work; - for (i = 0; i < mdev->act_log->nr_elements; i++) { - enr = lc_element_by_index(mdev->act_log, i)->lc_number; - if (enr == LC_FREE) - continue; - tmp = drbd_bm_ALe_set_all(mdev, enr); - dynamic_dev_dbg(DEV, "AL: set %d bits in extent %u\n", tmp, enr); - add += tmp; - } + if (current == mdev->tconn->worker.task) + return _al_write_transaction(mdev); - lc_unlock(mdev->act_log); - wake_up(&mdev->al_wait); + init_completion(&al_work.event); + al_work.w.cb = w_al_write_transaction; + al_work.w.mdev = mdev; + drbd_queue_work_front(&mdev->tconn->sender_work, &al_work.w); + wait_for_completion(&al_work.event); - dev_info(DEV, "Marked additional %s as out-of-sync based on AL.\n", - ppsize(ppb, Bit2KB(add))); + return al_work.err; } static int _try_lc_del(struct drbd_conf *mdev, struct lc_element *al_ext) @@ -645,7 +517,7 @@ void drbd_al_shrink(struct drbd_conf *mdev) struct lc_element *al_ext; int i; - D_ASSERT(test_bit(__LC_DIRTY, &mdev->act_log->flags)); + D_ASSERT(test_bit(__LC_LOCKED, &mdev->act_log->flags)); for (i = 0; i < mdev->act_log->nr_elements; i++) { al_ext = lc_element_by_index(mdev->act_log, i); @@ -657,15 +529,17 @@ void drbd_al_shrink(struct drbd_conf *mdev) wake_up(&mdev->al_wait); } -static int w_update_odbm(struct drbd_conf *mdev, struct drbd_work *w, int unused) +static int w_update_odbm(struct drbd_work *w, int unused) { struct update_odbm_work *udw = container_of(w, struct update_odbm_work, w); + struct drbd_conf *mdev = w->mdev; + struct sib_info sib = { .sib_reason = SIB_SYNC_PROGRESS, }; if (!get_ldev(mdev)) { if (__ratelimit(&drbd_ratelimit_state)) dev_warn(DEV, "Can not update on disk bitmap, local IO disabled.\n"); kfree(udw); - return 1; + return 0; } drbd_bm_write_page(mdev, rs_extent_to_bm_page(udw->enr)); @@ -683,9 +557,9 @@ static int w_update_odbm(struct drbd_conf *mdev, struct drbd_work *w, int unused break; } } - drbd_bcast_sync_progress(mdev); + drbd_bcast_event(mdev, &sib); - return 1; + return 0; } @@ -755,7 +629,9 @@ static void drbd_try_clear_on_disk_bm(struct drbd_conf *mdev, sector_t sector, } ext->rs_left = rs_left; ext->rs_failed = success ? 0 : count; - lc_changed(mdev->resync, &ext->lce); + /* we don't keep a persistent log of the resync lru, + * we can commit any change right away. */ + lc_committed(mdev->resync); } lc_put(mdev->resync, &ext->lce); /* no race, we are within the al_lock! */ @@ -767,7 +643,8 @@ static void drbd_try_clear_on_disk_bm(struct drbd_conf *mdev, sector_t sector, if (udw) { udw->enr = ext->lce.lc_number; udw->w.cb = w_update_odbm; - drbd_queue_work_front(&mdev->data.work, &udw->w); + udw->w.mdev = mdev; + drbd_queue_work_front(&mdev->tconn->sender_work, &udw->w); } else { dev_warn(DEV, "Could not kmalloc an udw\n"); } @@ -813,16 +690,22 @@ void __drbd_set_in_sync(struct drbd_conf *mdev, sector_t sector, int size, int wake_up = 0; unsigned long flags; - if (size <= 0 || (size & 0x1ff) != 0 || size > DRBD_MAX_BIO_SIZE) { + if (size <= 0 || !IS_ALIGNED(size, 512) || size > DRBD_MAX_BIO_SIZE) { dev_err(DEV, "drbd_set_in_sync: sector=%llus size=%d nonsense!\n", (unsigned long long)sector, size); return; } + + if (!get_ldev(mdev)) + return; /* no disk, no metadata, no bitmap to clear bits in */ + nr_sectors = drbd_get_capacity(mdev->this_bdev); esector = sector + (size >> 9) - 1; - ERR_IF(sector >= nr_sectors) return; - ERR_IF(esector >= nr_sectors) esector = (nr_sectors-1); + if (!expect(sector < nr_sectors)) + goto out; + if (!expect(esector < nr_sectors)) + esector = nr_sectors - 1; lbnr = BM_SECT_TO_BIT(nr_sectors-1); @@ -830,7 +713,7 @@ void __drbd_set_in_sync(struct drbd_conf *mdev, sector_t sector, int size, * round up start sector, round down end sector. we make sure we only * clear full, aligned, BM_BLOCK_SIZE (4K) blocks */ if (unlikely(esector < BM_SECT_PER_BIT-1)) - return; + goto out; if (unlikely(esector == (nr_sectors-1))) ebnr = lbnr; else @@ -838,14 +721,14 @@ void __drbd_set_in_sync(struct drbd_conf *mdev, sector_t sector, int size, sbnr = BM_SECT_TO_BIT(sector + BM_SECT_PER_BIT-1); if (sbnr > ebnr) - return; + goto out; /* * ok, (capacity & 7) != 0 sometimes, but who cares... * we count rs_{total,left} in bits, not sectors. */ count = drbd_bm_clear_bits(mdev, sbnr, ebnr); - if (count && get_ldev(mdev)) { + if (count) { drbd_advance_rs_marks(mdev, drbd_bm_total_weight(mdev)); spin_lock_irqsave(&mdev->al_lock, flags); drbd_try_clear_on_disk_bm(mdev, sector, count, true); @@ -854,8 +737,9 @@ void __drbd_set_in_sync(struct drbd_conf *mdev, sector_t sector, int size, /* just wake_up unconditional now, various lc_chaged(), * lc_put() in drbd_try_clear_on_disk_bm(). */ wake_up = 1; - put_ldev(mdev); } +out: + put_ldev(mdev); if (wake_up) wake_up(&mdev->al_wait); } @@ -871,7 +755,7 @@ void __drbd_set_in_sync(struct drbd_conf *mdev, sector_t sector, int size, int __drbd_set_out_of_sync(struct drbd_conf *mdev, sector_t sector, int size, const char *file, const unsigned int line) { - unsigned long sbnr, ebnr, lbnr, flags; + unsigned long sbnr, ebnr, flags; sector_t esector, nr_sectors; unsigned int enr, count = 0; struct lc_element *e; @@ -880,7 +764,7 @@ int __drbd_set_out_of_sync(struct drbd_conf *mdev, sector_t sector, int size, if (size == 0) return 0; - if (size < 0 || (size & 0x1ff) != 0 || size > DRBD_MAX_BIO_SIZE) { + if (size < 0 || !IS_ALIGNED(size, 512) || size > DRBD_MAX_BIO_SIZE) { dev_err(DEV, "sector: %llus, size: %d\n", (unsigned long long)sector, size); return 0; @@ -892,12 +776,10 @@ int __drbd_set_out_of_sync(struct drbd_conf *mdev, sector_t sector, int size, nr_sectors = drbd_get_capacity(mdev->this_bdev); esector = sector + (size >> 9) - 1; - ERR_IF(sector >= nr_sectors) + if (!expect(sector < nr_sectors)) goto out; - ERR_IF(esector >= nr_sectors) - esector = (nr_sectors-1); - - lbnr = BM_SECT_TO_BIT(nr_sectors-1); + if (!expect(esector < nr_sectors)) + esector = nr_sectors - 1; /* we set it out of sync, * we do not need to round anything here */ @@ -940,7 +822,7 @@ struct bm_extent *_bme_get(struct drbd_conf *mdev, unsigned int enr) if (bm_ext->lce.lc_number != enr) { bm_ext->rs_left = drbd_bm_e_weight(mdev, enr); bm_ext->rs_failed = 0; - lc_changed(mdev->resync, &bm_ext->lce); + lc_committed(mdev->resync); wakeup = 1; } if (bm_ext->lce.refcnt == 1) @@ -956,7 +838,7 @@ struct bm_extent *_bme_get(struct drbd_conf *mdev, unsigned int enr) if (rs_flags & LC_STARVING) dev_warn(DEV, "Have to wait for element" " (resync LRU too small?)\n"); - BUG_ON(rs_flags & LC_DIRTY); + BUG_ON(rs_flags & LC_LOCKED); } return bm_ext; @@ -964,26 +846,12 @@ struct bm_extent *_bme_get(struct drbd_conf *mdev, unsigned int enr) static int _is_in_al(struct drbd_conf *mdev, unsigned int enr) { - struct lc_element *al_ext; - int rv = 0; + int rv; spin_lock_irq(&mdev->al_lock); - if (unlikely(enr == mdev->act_log->new_number)) - rv = 1; - else { - al_ext = lc_find(mdev->act_log, enr); - if (al_ext) { - if (al_ext->refcnt) - rv = 1; - } - } + rv = lc_is_used(mdev->act_log, enr); spin_unlock_irq(&mdev->al_lock); - /* - if (unlikely(rv)) { - dev_info(DEV, "Delaying sync read until app's write is done\n"); - } - */ return rv; } @@ -1113,13 +981,13 @@ int drbd_try_rs_begin_io(struct drbd_conf *mdev, sector_t sector) if (rs_flags & LC_STARVING) dev_warn(DEV, "Have to wait for element" " (resync LRU too small?)\n"); - BUG_ON(rs_flags & LC_DIRTY); + BUG_ON(rs_flags & LC_LOCKED); goto try_again; } if (bm_ext->lce.lc_number != enr) { bm_ext->rs_left = drbd_bm_e_weight(mdev, enr); bm_ext->rs_failed = 0; - lc_changed(mdev->resync, &bm_ext->lce); + lc_committed(mdev->resync); wake_up(&mdev->al_wait); D_ASSERT(test_bit(BME_LOCKED, &bm_ext->flags) == 0); } @@ -1130,8 +998,6 @@ int drbd_try_rs_begin_io(struct drbd_conf *mdev, sector_t sector) } check_al: for (i = 0; i < AL_EXT_PER_BM_SECT; i++) { - if (unlikely(al_enr+i == mdev->act_log->new_number)) - goto try_again; if (lc_is_used(mdev->act_log, al_enr+i)) goto try_again; } @@ -1266,7 +1132,7 @@ void drbd_rs_failed_io(struct drbd_conf *mdev, sector_t sector, int size) sector_t esector, nr_sectors; int wake_up = 0; - if (size <= 0 || (size & 0x1ff) != 0 || size > DRBD_MAX_BIO_SIZE) { + if (size <= 0 || !IS_ALIGNED(size, 512) || size > DRBD_MAX_BIO_SIZE) { dev_err(DEV, "drbd_rs_failed_io: sector=%llus size=%d nonsense!\n", (unsigned long long)sector, size); return; @@ -1274,8 +1140,10 @@ void drbd_rs_failed_io(struct drbd_conf *mdev, sector_t sector, int size) nr_sectors = drbd_get_capacity(mdev->this_bdev); esector = sector + (size >> 9) - 1; - ERR_IF(sector >= nr_sectors) return; - ERR_IF(esector >= nr_sectors) esector = (nr_sectors-1); + if (!expect(sector < nr_sectors)) + return; + if (!expect(esector < nr_sectors)) + esector = nr_sectors - 1; lbnr = BM_SECT_TO_BIT(nr_sectors-1); diff --git a/drivers/block/drbd/drbd_bitmap.c b/drivers/block/drbd/drbd_bitmap.c index d84566496746..8dc29502dc08 100644 --- a/drivers/block/drbd/drbd_bitmap.c +++ b/drivers/block/drbd/drbd_bitmap.c @@ -119,13 +119,9 @@ static void __bm_print_lock_info(struct drbd_conf *mdev, const char *func) if (!__ratelimit(&drbd_ratelimit_state)) return; dev_err(DEV, "FIXME %s in %s, bitmap locked for '%s' by %s\n", - current == mdev->receiver.task ? "receiver" : - current == mdev->asender.task ? "asender" : - current == mdev->worker.task ? "worker" : current->comm, - func, b->bm_why ?: "?", - b->bm_task == mdev->receiver.task ? "receiver" : - b->bm_task == mdev->asender.task ? "asender" : - b->bm_task == mdev->worker.task ? "worker" : "?"); + drbd_task_to_thread_name(mdev->tconn, current), + func, b->bm_why ?: "?", + drbd_task_to_thread_name(mdev->tconn, b->bm_task)); } void drbd_bm_lock(struct drbd_conf *mdev, char *why, enum bm_flag flags) @@ -142,13 +138,9 @@ void drbd_bm_lock(struct drbd_conf *mdev, char *why, enum bm_flag flags) if (trylock_failed) { dev_warn(DEV, "%s going to '%s' but bitmap already locked for '%s' by %s\n", - current == mdev->receiver.task ? "receiver" : - current == mdev->asender.task ? "asender" : - current == mdev->worker.task ? "worker" : current->comm, - why, b->bm_why ?: "?", - b->bm_task == mdev->receiver.task ? "receiver" : - b->bm_task == mdev->asender.task ? "asender" : - b->bm_task == mdev->worker.task ? "worker" : "?"); + drbd_task_to_thread_name(mdev->tconn, current), + why, b->bm_why ?: "?", + drbd_task_to_thread_name(mdev->tconn, b->bm_task)); mutex_lock(&b->bm_change); } if (BM_LOCKED_MASK & b->bm_flags) @@ -196,6 +188,9 @@ void drbd_bm_unlock(struct drbd_conf *mdev) /* to mark for lazy writeout once syncer cleared all clearable bits, * we if bits have been cleared since last IO. */ #define BM_PAGE_LAZY_WRITEOUT 28 +/* pages marked with this "HINT" will be considered for writeout + * on activity log transactions */ +#define BM_PAGE_HINT_WRITEOUT 27 /* store_page_idx uses non-atomic assignment. It is only used directly after * allocating the page. All other bm_set_page_* and bm_clear_page_* need to @@ -227,8 +222,7 @@ static void bm_page_unlock_io(struct drbd_conf *mdev, int page_nr) { struct drbd_bitmap *b = mdev->bitmap; void *addr = &page_private(b->bm_pages[page_nr]); - clear_bit(BM_PAGE_IO_LOCK, addr); - smp_mb__after_clear_bit(); + clear_bit_unlock(BM_PAGE_IO_LOCK, addr); wake_up(&mdev->bitmap->bm_io_wait); } @@ -246,6 +240,27 @@ static void bm_set_page_need_writeout(struct page *page) set_bit(BM_PAGE_NEED_WRITEOUT, &page_private(page)); } +/** + * drbd_bm_mark_for_writeout() - mark a page with a "hint" to be considered for writeout + * @mdev: DRBD device. + * @page_nr: the bitmap page to mark with the "hint" flag + * + * From within an activity log transaction, we mark a few pages with these + * hints, then call drbd_bm_write_hinted(), which will only write out changed + * pages which are flagged with this mark. + */ +void drbd_bm_mark_for_writeout(struct drbd_conf *mdev, int page_nr) +{ + struct page *page; + if (page_nr >= mdev->bitmap->bm_number_of_pages) { + dev_warn(DEV, "BAD: page_nr: %u, number_of_pages: %u\n", + page_nr, (int)mdev->bitmap->bm_number_of_pages); + return; + } + page = mdev->bitmap->bm_pages[page_nr]; + set_bit(BM_PAGE_HINT_WRITEOUT, &page_private(page)); +} + static int bm_test_page_unchanged(struct page *page) { volatile const unsigned long *addr = &page_private(page); @@ -373,14 +388,16 @@ static struct page **bm_realloc_pages(struct drbd_bitmap *b, unsigned long want) return old_pages; /* Trying kmalloc first, falling back to vmalloc. - * GFP_KERNEL is ok, as this is done when a lower level disk is - * "attached" to the drbd. Context is receiver thread or cqueue - * thread. As we have no disk yet, we are not in the IO path, - * not even the IO path of the peer. */ + * GFP_NOIO, as this is called while drbd IO is "suspended", + * and during resize or attach on diskless Primary, + * we must not block on IO to ourselves. + * Context is receiver thread or dmsetup. */ bytes = sizeof(struct page *)*want; - new_pages = kzalloc(bytes, GFP_KERNEL); + new_pages = kzalloc(bytes, GFP_NOIO); if (!new_pages) { - new_pages = vzalloc(bytes); + new_pages = __vmalloc(bytes, + GFP_NOIO | __GFP_HIGHMEM | __GFP_ZERO, + PAGE_KERNEL); if (!new_pages) return NULL; vmalloced = 1; @@ -390,7 +407,7 @@ static struct page **bm_realloc_pages(struct drbd_bitmap *b, unsigned long want) for (i = 0; i < have; i++) new_pages[i] = old_pages[i]; for (; i < want; i++) { - page = alloc_page(GFP_HIGHUSER); + page = alloc_page(GFP_NOIO | __GFP_HIGHMEM); if (!page) { bm_free_pages(new_pages + have, i - have); bm_vk_free(new_pages, vmalloced); @@ -439,7 +456,8 @@ int drbd_bm_init(struct drbd_conf *mdev) sector_t drbd_bm_capacity(struct drbd_conf *mdev) { - ERR_IF(!mdev->bitmap) return 0; + if (!expect(mdev->bitmap)) + return 0; return mdev->bitmap->bm_dev_capacity; } @@ -447,7 +465,8 @@ sector_t drbd_bm_capacity(struct drbd_conf *mdev) */ void drbd_bm_cleanup(struct drbd_conf *mdev) { - ERR_IF (!mdev->bitmap) return; + if (!expect(mdev->bitmap)) + return; bm_free_pages(mdev->bitmap->bm_pages, mdev->bitmap->bm_number_of_pages); bm_vk_free(mdev->bitmap->bm_pages, (BM_P_VMALLOCED & mdev->bitmap->bm_flags)); kfree(mdev->bitmap); @@ -610,7 +629,8 @@ int drbd_bm_resize(struct drbd_conf *mdev, sector_t capacity, int set_new_bits) int err = 0, growing; int opages_vmalloced; - ERR_IF(!b) return -ENOMEM; + if (!expect(b)) + return -ENOMEM; drbd_bm_lock(mdev, "resize", BM_LOCKED_MASK); @@ -732,8 +752,10 @@ unsigned long _drbd_bm_total_weight(struct drbd_conf *mdev) unsigned long s; unsigned long flags; - ERR_IF(!b) return 0; - ERR_IF(!b->bm_pages) return 0; + if (!expect(b)) + return 0; + if (!expect(b->bm_pages)) + return 0; spin_lock_irqsave(&b->bm_lock, flags); s = b->bm_set; @@ -756,8 +778,10 @@ unsigned long drbd_bm_total_weight(struct drbd_conf *mdev) size_t drbd_bm_words(struct drbd_conf *mdev) { struct drbd_bitmap *b = mdev->bitmap; - ERR_IF(!b) return 0; - ERR_IF(!b->bm_pages) return 0; + if (!expect(b)) + return 0; + if (!expect(b->bm_pages)) + return 0; return b->bm_words; } @@ -765,7 +789,8 @@ size_t drbd_bm_words(struct drbd_conf *mdev) unsigned long drbd_bm_bits(struct drbd_conf *mdev) { struct drbd_bitmap *b = mdev->bitmap; - ERR_IF(!b) return 0; + if (!expect(b)) + return 0; return b->bm_bits; } @@ -786,8 +811,10 @@ void drbd_bm_merge_lel(struct drbd_conf *mdev, size_t offset, size_t number, end = offset + number; - ERR_IF(!b) return; - ERR_IF(!b->bm_pages) return; + if (!expect(b)) + return; + if (!expect(b->bm_pages)) + return; if (number == 0) return; WARN_ON(offset >= b->bm_words); @@ -831,8 +858,10 @@ void drbd_bm_get_lel(struct drbd_conf *mdev, size_t offset, size_t number, end = offset + number; - ERR_IF(!b) return; - ERR_IF(!b->bm_pages) return; + if (!expect(b)) + return; + if (!expect(b->bm_pages)) + return; spin_lock_irq(&b->bm_lock); if ((offset >= b->bm_words) || @@ -860,8 +889,10 @@ void drbd_bm_get_lel(struct drbd_conf *mdev, size_t offset, size_t number, void drbd_bm_set_all(struct drbd_conf *mdev) { struct drbd_bitmap *b = mdev->bitmap; - ERR_IF(!b) return; - ERR_IF(!b->bm_pages) return; + if (!expect(b)) + return; + if (!expect(b->bm_pages)) + return; spin_lock_irq(&b->bm_lock); bm_memset(b, 0, 0xff, b->bm_words); @@ -874,8 +905,10 @@ void drbd_bm_set_all(struct drbd_conf *mdev) void drbd_bm_clear_all(struct drbd_conf *mdev) { struct drbd_bitmap *b = mdev->bitmap; - ERR_IF(!b) return; - ERR_IF(!b->bm_pages) return; + if (!expect(b)) + return; + if (!expect(b->bm_pages)) + return; spin_lock_irq(&b->bm_lock); bm_memset(b, 0, 0, b->bm_words); @@ -889,7 +922,8 @@ struct bm_aio_ctx { unsigned int done; unsigned flags; #define BM_AIO_COPY_PAGES 1 -#define BM_WRITE_ALL_PAGES 2 +#define BM_AIO_WRITE_HINTED 2 +#define BM_WRITE_ALL_PAGES 4 int error; struct kref kref; }; @@ -977,17 +1011,11 @@ static void bm_page_io_async(struct bm_aio_ctx *ctx, int page_nr, int rw) __must bm_set_page_unchanged(b->bm_pages[page_nr]); if (ctx->flags & BM_AIO_COPY_PAGES) { - void *src, *dest; page = mempool_alloc(drbd_md_io_page_pool, __GFP_HIGHMEM|__GFP_WAIT); - dest = kmap_atomic(page); - src = kmap_atomic(b->bm_pages[page_nr]); - memcpy(dest, src, PAGE_SIZE); - kunmap_atomic(src); - kunmap_atomic(dest); + copy_highpage(page, b->bm_pages[page_nr]); bm_store_page_idx(page, page_nr); } else page = b->bm_pages[page_nr]; - bio->bi_bdev = mdev->ldev->md_bdev; bio->bi_sector = on_disk_sector; /* bio_add_page of a single page to an empty bio will always succeed, @@ -1060,6 +1088,11 @@ static int bm_rw(struct drbd_conf *mdev, int rw, unsigned flags, unsigned lazy_w if (lazy_writeout_upper_idx && i == lazy_writeout_upper_idx) break; if (rw & WRITE) { + if ((flags & BM_AIO_WRITE_HINTED) && + !test_and_clear_bit(BM_PAGE_HINT_WRITEOUT, + &page_private(b->bm_pages[i]))) + continue; + if (!(flags & BM_WRITE_ALL_PAGES) && bm_test_page_unchanged(b->bm_pages[i])) { dynamic_dev_dbg(DEV, "skipped bm write for idx %u\n", i); @@ -1088,13 +1121,15 @@ static int bm_rw(struct drbd_conf *mdev, int rw, unsigned flags, unsigned lazy_w * "in_flight reached zero, all done" event. */ if (!atomic_dec_and_test(&ctx->in_flight)) - wait_until_done_or_disk_failure(mdev, mdev->ldev, &ctx->done); + wait_until_done_or_force_detached(mdev, mdev->ldev, &ctx->done); else kref_put(&ctx->kref, &bm_aio_ctx_destroy); - dev_info(DEV, "bitmap %s of %u pages took %lu jiffies\n", - rw == WRITE ? "WRITE" : "READ", - count, jiffies - now); + /* summary for global bitmap IO */ + if (flags == 0) + dev_info(DEV, "bitmap %s of %u pages took %lu jiffies\n", + rw == WRITE ? "WRITE" : "READ", + count, jiffies - now); if (ctx->error) { dev_alert(DEV, "we had at least one MD IO ERROR during bitmap IO\n"); @@ -1103,7 +1138,7 @@ static int bm_rw(struct drbd_conf *mdev, int rw, unsigned flags, unsigned lazy_w } if (atomic_read(&ctx->in_flight)) - err = -EIO; /* Disk failed during IO... */ + err = -EIO; /* Disk timeout/force-detach during IO... */ now = jiffies; if (rw == WRITE) { @@ -1115,8 +1150,9 @@ static int bm_rw(struct drbd_conf *mdev, int rw, unsigned flags, unsigned lazy_w } now = b->bm_set; - dev_info(DEV, "%s (%lu bits) marked out-of-sync by on disk bit-map.\n", - ppsize(ppb, now << (BM_BLOCK_SHIFT-10)), now); + if (flags == 0) + dev_info(DEV, "%s (%lu bits) marked out-of-sync by on disk bit-map.\n", + ppsize(ppb, now << (BM_BLOCK_SHIFT-10)), now); kref_put(&ctx->kref, &bm_aio_ctx_destroy); return err; @@ -1179,9 +1215,17 @@ int drbd_bm_write_copy_pages(struct drbd_conf *mdev) __must_hold(local) return bm_rw(mdev, WRITE, BM_AIO_COPY_PAGES, 0); } +/** + * drbd_bm_write_hinted() - Write bitmap pages with "hint" marks, if they have changed. + * @mdev: DRBD device. + */ +int drbd_bm_write_hinted(struct drbd_conf *mdev) __must_hold(local) +{ + return bm_rw(mdev, WRITE, BM_AIO_WRITE_HINTED | BM_AIO_COPY_PAGES, 0); +} /** - * drbd_bm_write_page: Writes a PAGE_SIZE aligned piece of bitmap + * drbd_bm_write_page() - Writes a PAGE_SIZE aligned piece of bitmap * @mdev: DRBD device. * @idx: bitmap page index * @@ -1222,11 +1266,11 @@ int drbd_bm_write_page(struct drbd_conf *mdev, unsigned int idx) __must_hold(loc } bm_page_io_async(ctx, idx, WRITE_SYNC); - wait_until_done_or_disk_failure(mdev, mdev->ldev, &ctx->done); + wait_until_done_or_force_detached(mdev, mdev->ldev, &ctx->done); if (ctx->error) drbd_chk_io_error(mdev, 1, DRBD_META_IO_ERROR); - /* that should force detach, so the in memory bitmap will be + /* that causes us to detach, so the in memory bitmap will be * gone in a moment as well. */ mdev->bm_writ_cnt++; @@ -1289,8 +1333,10 @@ static unsigned long bm_find_next(struct drbd_conf *mdev, struct drbd_bitmap *b = mdev->bitmap; unsigned long i = DRBD_END_OF_BITMAP; - ERR_IF(!b) return i; - ERR_IF(!b->bm_pages) return i; + if (!expect(b)) + return i; + if (!expect(b->bm_pages)) + return i; spin_lock_irq(&b->bm_lock); if (BM_DONT_TEST & b->bm_flags) @@ -1391,8 +1437,10 @@ static int bm_change_bits_to(struct drbd_conf *mdev, const unsigned long s, struct drbd_bitmap *b = mdev->bitmap; int c = 0; - ERR_IF(!b) return 1; - ERR_IF(!b->bm_pages) return 0; + if (!expect(b)) + return 1; + if (!expect(b->bm_pages)) + return 0; spin_lock_irqsave(&b->bm_lock, flags); if ((val ? BM_DONT_SET : BM_DONT_CLEAR) & b->bm_flags) @@ -1423,13 +1471,21 @@ static inline void bm_set_full_words_within_one_page(struct drbd_bitmap *b, { int i; int bits; + int changed = 0; unsigned long *paddr = kmap_atomic(b->bm_pages[page_nr]); for (i = first_word; i < last_word; i++) { bits = hweight_long(paddr[i]); paddr[i] = ~0UL; - b->bm_set += BITS_PER_LONG - bits; + changed += BITS_PER_LONG - bits; } kunmap_atomic(paddr); + if (changed) { + /* We only need lazy writeout, the information is still in the + * remote bitmap as well, and is reconstructed during the next + * bitmap exchange, if lost locally due to a crash. */ + bm_set_page_lazy_writeout(b->bm_pages[page_nr]); + b->bm_set += changed; + } } /* Same thing as drbd_bm_set_bits, @@ -1524,8 +1580,10 @@ int drbd_bm_test_bit(struct drbd_conf *mdev, const unsigned long bitnr) unsigned long *p_addr; int i; - ERR_IF(!b) return 0; - ERR_IF(!b->bm_pages) return 0; + if (!expect(b)) + return 0; + if (!expect(b->bm_pages)) + return 0; spin_lock_irqsave(&b->bm_lock, flags); if (BM_DONT_TEST & b->bm_flags) @@ -1559,8 +1617,10 @@ int drbd_bm_count_bits(struct drbd_conf *mdev, const unsigned long s, const unsi * robust in case we screwed up elsewhere, in that case pretend there * was one dirty bit in the requested area, so we won't try to do a * local read there (no bitmap probably implies no disk) */ - ERR_IF(!b) return 1; - ERR_IF(!b->bm_pages) return 1; + if (!expect(b)) + return 1; + if (!expect(b->bm_pages)) + return 1; spin_lock_irqsave(&b->bm_lock, flags); if (BM_DONT_TEST & b->bm_flags) @@ -1573,11 +1633,10 @@ int drbd_bm_count_bits(struct drbd_conf *mdev, const unsigned long s, const unsi bm_unmap(p_addr); p_addr = bm_map_pidx(b, idx); } - ERR_IF (bitnr >= b->bm_bits) { - dev_err(DEV, "bitnr=%lu bm_bits=%lu\n", bitnr, b->bm_bits); - } else { + if (expect(bitnr < b->bm_bits)) c += (0 != test_bit_le(bitnr - (page_nr << (PAGE_SHIFT+3)), p_addr)); - } + else + dev_err(DEV, "bitnr=%lu bm_bits=%lu\n", bitnr, b->bm_bits); } if (p_addr) bm_unmap(p_addr); @@ -1607,8 +1666,10 @@ int drbd_bm_e_weight(struct drbd_conf *mdev, unsigned long enr) unsigned long flags; unsigned long *p_addr, *bm; - ERR_IF(!b) return 0; - ERR_IF(!b->bm_pages) return 0; + if (!expect(b)) + return 0; + if (!expect(b->bm_pages)) + return 0; spin_lock_irqsave(&b->bm_lock, flags); if (BM_DONT_TEST & b->bm_flags) @@ -1630,47 +1691,3 @@ int drbd_bm_e_weight(struct drbd_conf *mdev, unsigned long enr) spin_unlock_irqrestore(&b->bm_lock, flags); return count; } - -/* Set all bits covered by the AL-extent al_enr. - * Returns number of bits changed. */ -unsigned long drbd_bm_ALe_set_all(struct drbd_conf *mdev, unsigned long al_enr) -{ - struct drbd_bitmap *b = mdev->bitmap; - unsigned long *p_addr, *bm; - unsigned long weight; - unsigned long s, e; - int count, i, do_now; - ERR_IF(!b) return 0; - ERR_IF(!b->bm_pages) return 0; - - spin_lock_irq(&b->bm_lock); - if (BM_DONT_SET & b->bm_flags) - bm_print_lock_info(mdev); - weight = b->bm_set; - - s = al_enr * BM_WORDS_PER_AL_EXT; - e = min_t(size_t, s + BM_WORDS_PER_AL_EXT, b->bm_words); - /* assert that s and e are on the same page */ - D_ASSERT((e-1) >> (PAGE_SHIFT - LN2_BPL + 3) - == s >> (PAGE_SHIFT - LN2_BPL + 3)); - count = 0; - if (s < b->bm_words) { - i = do_now = e-s; - p_addr = bm_map_pidx(b, bm_word_to_page_idx(b, s)); - bm = p_addr + MLPP(s); - while (i--) { - count += hweight_long(*bm); - *bm = -1UL; - bm++; - } - bm_unmap(p_addr); - b->bm_set += do_now*BITS_PER_LONG - count; - if (e == b->bm_words) - b->bm_set -= bm_clear_surplus(b); - } else { - dev_err(DEV, "start offset (%lu) too large in drbd_bm_ALe_set_all\n", s); - } - weight = b->bm_set - weight; - spin_unlock_irq(&b->bm_lock); - return weight; -} diff --git a/drivers/block/drbd/drbd_int.h b/drivers/block/drbd/drbd_int.h index b953cc7c9c00..6b51afa1aae1 100644 --- a/drivers/block/drbd/drbd_int.h +++ b/drivers/block/drbd/drbd_int.h @@ -39,9 +39,13 @@ #include <linux/major.h> #include <linux/blkdev.h> #include <linux/genhd.h> +#include <linux/idr.h> #include <net/tcp.h> #include <linux/lru_cache.h> #include <linux/prefetch.h> +#include <linux/drbd_genl_api.h> +#include <linux/drbd.h> +#include "drbd_state.h" #ifdef __CHECKER__ # define __protected_by(x) __attribute__((require_context(x,1,999,"rdwr"))) @@ -61,7 +65,6 @@ extern unsigned int minor_count; extern bool disable_sendpage; extern bool allow_oos; -extern unsigned int cn_idx; #ifdef CONFIG_DRBD_FAULT_INJECTION extern int enable_faults; @@ -86,34 +89,44 @@ extern char usermode_helper[]; */ #define DRBD_SIGKILL SIGHUP -/* All EEs on the free list should have ID_VACANT (== 0) - * freshly allocated EEs get !ID_VACANT (== 1) - * so if it says "cannot dereference null pointer at address 0x00000001", - * it is most likely one of these :( */ - #define ID_IN_SYNC (4711ULL) #define ID_OUT_OF_SYNC (4712ULL) - #define ID_SYNCER (-1ULL) -#define ID_VACANT 0 -#define is_syncer_block_id(id) ((id) == ID_SYNCER) + #define UUID_NEW_BM_OFFSET ((u64)0x0001000000000000ULL) struct drbd_conf; +struct drbd_tconn; /* to shorten dev_warn(DEV, "msg"); and relatives statements */ #define DEV (disk_to_dev(mdev->vdisk)) +#define conn_printk(LEVEL, TCONN, FMT, ARGS...) \ + printk(LEVEL "d-con %s: " FMT, TCONN->name , ## ARGS) +#define conn_alert(TCONN, FMT, ARGS...) conn_printk(KERN_ALERT, TCONN, FMT, ## ARGS) +#define conn_crit(TCONN, FMT, ARGS...) conn_printk(KERN_CRIT, TCONN, FMT, ## ARGS) +#define conn_err(TCONN, FMT, ARGS...) conn_printk(KERN_ERR, TCONN, FMT, ## ARGS) +#define conn_warn(TCONN, FMT, ARGS...) conn_printk(KERN_WARNING, TCONN, FMT, ## ARGS) +#define conn_notice(TCONN, FMT, ARGS...) conn_printk(KERN_NOTICE, TCONN, FMT, ## ARGS) +#define conn_info(TCONN, FMT, ARGS...) conn_printk(KERN_INFO, TCONN, FMT, ## ARGS) +#define conn_dbg(TCONN, FMT, ARGS...) conn_printk(KERN_DEBUG, TCONN, FMT, ## ARGS) + #define D_ASSERT(exp) if (!(exp)) \ dev_err(DEV, "ASSERT( " #exp " ) in %s:%d\n", __FILE__, __LINE__) -#define ERR_IF(exp) if (({ \ - int _b = (exp) != 0; \ - if (_b) dev_err(DEV, "ASSERT FAILED: %s: (%s) in %s:%d\n", \ - __func__, #exp, __FILE__, __LINE__); \ - _b; \ - })) +/** + * expect - Make an assertion + * + * Unlike the assert macro, this macro returns a boolean result. + */ +#define expect(exp) ({ \ + bool _bool = (exp); \ + if (!_bool) \ + dev_err(DEV, "ASSERTION %s FAILED in %s\n", \ + #exp, __func__); \ + _bool; \ + }) /* Defines to control fault insertion */ enum { @@ -150,15 +163,12 @@ drbd_insert_fault(struct drbd_conf *mdev, unsigned int type) { /* usual integer division */ #define div_floor(A, B) ((A)/(B)) -/* drbd_meta-data.c (still in drbd_main.c) */ -/* 4th incarnation of the disk layout. */ -#define DRBD_MD_MAGIC (DRBD_MAGIC+4) - -extern struct drbd_conf **minor_table; extern struct ratelimit_state drbd_ratelimit_state; +extern struct idr minors; /* RCU, updates: genl_lock() */ +extern struct list_head drbd_tconns; /* RCU, updates: genl_lock() */ /* on the wire */ -enum drbd_packets { +enum drbd_packet { /* receiver (data socket) */ P_DATA = 0x00, P_DATA_REPLY = 0x01, /* Response to P_DATA_REQUEST */ @@ -186,7 +196,7 @@ enum drbd_packets { P_RECV_ACK = 0x15, /* Used in protocol B */ P_WRITE_ACK = 0x16, /* Used in protocol C */ P_RS_WRITE_ACK = 0x17, /* Is a P_WRITE_ACK, additionally call set_in_sync(). */ - P_DISCARD_ACK = 0x18, /* Used in proto C, two-primaries conflict detection */ + P_SUPERSEDED = 0x18, /* Used in proto C, two-primaries conflict detection */ P_NEG_ACK = 0x19, /* Sent if local disk is unusable */ P_NEG_DREPLY = 0x1a, /* Local disk is broken... */ P_NEG_RS_DREPLY = 0x1b, /* Local disk is broken... */ @@ -207,77 +217,23 @@ enum drbd_packets { P_DELAY_PROBE = 0x27, /* is used on BOTH sockets */ P_OUT_OF_SYNC = 0x28, /* Mark as out of sync (Outrunning), data socket */ P_RS_CANCEL = 0x29, /* meta: Used to cancel RS_DATA_REQUEST packet by SyncSource */ + P_CONN_ST_CHG_REQ = 0x2a, /* data sock: Connection wide state request */ + P_CONN_ST_CHG_REPLY = 0x2b, /* meta sock: Connection side state req reply */ + P_RETRY_WRITE = 0x2c, /* Protocol C: retry conflicting write request */ + P_PROTOCOL_UPDATE = 0x2d, /* data sock: is used in established connections */ - P_MAX_CMD = 0x2A, P_MAY_IGNORE = 0x100, /* Flag to test if (cmd > P_MAY_IGNORE) ... */ P_MAX_OPT_CMD = 0x101, /* special command ids for handshake */ - P_HAND_SHAKE_M = 0xfff1, /* First Packet on the MetaSock */ - P_HAND_SHAKE_S = 0xfff2, /* First Packet on the Socket */ + P_INITIAL_META = 0xfff1, /* First Packet on the MetaSock */ + P_INITIAL_DATA = 0xfff2, /* First Packet on the Socket */ - P_HAND_SHAKE = 0xfffe /* FIXED for the next century! */ + P_CONNECTION_FEATURES = 0xfffe /* FIXED for the next century! */ }; -static inline const char *cmdname(enum drbd_packets cmd) -{ - /* THINK may need to become several global tables - * when we want to support more than - * one PRO_VERSION */ - static const char *cmdnames[] = { - [P_DATA] = "Data", - [P_DATA_REPLY] = "DataReply", - [P_RS_DATA_REPLY] = "RSDataReply", - [P_BARRIER] = "Barrier", - [P_BITMAP] = "ReportBitMap", - [P_BECOME_SYNC_TARGET] = "BecomeSyncTarget", - [P_BECOME_SYNC_SOURCE] = "BecomeSyncSource", - [P_UNPLUG_REMOTE] = "UnplugRemote", - [P_DATA_REQUEST] = "DataRequest", - [P_RS_DATA_REQUEST] = "RSDataRequest", - [P_SYNC_PARAM] = "SyncParam", - [P_SYNC_PARAM89] = "SyncParam89", - [P_PROTOCOL] = "ReportProtocol", - [P_UUIDS] = "ReportUUIDs", - [P_SIZES] = "ReportSizes", - [P_STATE] = "ReportState", - [P_SYNC_UUID] = "ReportSyncUUID", - [P_AUTH_CHALLENGE] = "AuthChallenge", - [P_AUTH_RESPONSE] = "AuthResponse", - [P_PING] = "Ping", - [P_PING_ACK] = "PingAck", - [P_RECV_ACK] = "RecvAck", - [P_WRITE_ACK] = "WriteAck", - [P_RS_WRITE_ACK] = "RSWriteAck", - [P_DISCARD_ACK] = "DiscardAck", - [P_NEG_ACK] = "NegAck", - [P_NEG_DREPLY] = "NegDReply", - [P_NEG_RS_DREPLY] = "NegRSDReply", - [P_BARRIER_ACK] = "BarrierAck", - [P_STATE_CHG_REQ] = "StateChgRequest", - [P_STATE_CHG_REPLY] = "StateChgReply", - [P_OV_REQUEST] = "OVRequest", - [P_OV_REPLY] = "OVReply", - [P_OV_RESULT] = "OVResult", - [P_CSUM_RS_REQUEST] = "CsumRSRequest", - [P_RS_IS_IN_SYNC] = "CsumRSIsInSync", - [P_COMPRESSED_BITMAP] = "CBitmap", - [P_DELAY_PROBE] = "DelayProbe", - [P_OUT_OF_SYNC] = "OutOfSync", - [P_MAX_CMD] = NULL, - }; - - if (cmd == P_HAND_SHAKE_M) - return "HandShakeM"; - if (cmd == P_HAND_SHAKE_S) - return "HandShakeS"; - if (cmd == P_HAND_SHAKE) - return "HandShake"; - if (cmd >= P_MAX_CMD) - return "Unknown"; - return cmdnames[cmd]; -} +extern const char *cmdname(enum drbd_packet cmd); /* for sending/receiving the bitmap, * possibly in some encoding scheme */ @@ -337,37 +293,24 @@ struct p_header80 { u32 magic; u16 command; u16 length; /* bytes of data after this header */ - u8 payload[0]; } __packed; /* Header for big packets, Used for data packets exceeding 64kB */ struct p_header95 { u16 magic; /* use DRBD_MAGIC_BIG here */ u16 command; - u32 length; /* Use only 24 bits of that. Ignore the highest 8 bit. */ - u8 payload[0]; + u32 length; } __packed; -union p_header { - struct p_header80 h80; - struct p_header95 h95; -}; - -/* - * short commands, packets without payload, plain p_header: - * P_PING - * P_PING_ACK - * P_BECOME_SYNC_TARGET - * P_BECOME_SYNC_SOURCE - * P_UNPLUG_REMOTE - */ +struct p_header100 { + u32 magic; + u16 volume; + u16 command; + u32 length; + u32 pad; +} __packed; -/* - * commands with out-of-struct payload: - * P_BITMAP (no additional fields) - * P_DATA, P_DATA_REPLY (see p_data) - * P_COMPRESSED_BITMAP (see receive_compressed_bitmap) - */ +extern unsigned int drbd_header_size(struct drbd_tconn *tconn); /* these defines must not be changed without changing the protocol version */ #define DP_HARDBARRIER 1 /* depricated */ @@ -377,9 +320,10 @@ union p_header { #define DP_FUA 16 /* equals REQ_FUA */ #define DP_FLUSH 32 /* equals REQ_FLUSH */ #define DP_DISCARD 64 /* equals REQ_DISCARD */ +#define DP_SEND_RECEIVE_ACK 128 /* This is a proto B write request */ +#define DP_SEND_WRITE_ACK 256 /* This is a proto C write request */ struct p_data { - union p_header head; u64 sector; /* 64 bits sector number */ u64 block_id; /* to identify the request in protocol B&C */ u32 seq_num; @@ -390,21 +334,18 @@ struct p_data { * commands which share a struct: * p_block_ack: * P_RECV_ACK (proto B), P_WRITE_ACK (proto C), - * P_DISCARD_ACK (proto C, two-primaries conflict detection) + * P_SUPERSEDED (proto C, two-primaries conflict detection) * p_block_req: * P_DATA_REQUEST, P_RS_DATA_REQUEST */ struct p_block_ack { - struct p_header80 head; u64 sector; u64 block_id; u32 blksize; u32 seq_num; } __packed; - struct p_block_req { - struct p_header80 head; u64 sector; u64 block_id; u32 blksize; @@ -413,59 +354,52 @@ struct p_block_req { /* * commands with their own struct for additional fields: - * P_HAND_SHAKE + * P_CONNECTION_FEATURES * P_BARRIER * P_BARRIER_ACK * P_SYNC_PARAM * ReportParams */ -struct p_handshake { - struct p_header80 head; /* 8 bytes */ +struct p_connection_features { u32 protocol_min; u32 feature_flags; u32 protocol_max; /* should be more than enough for future enhancements - * for now, feature_flags and the reserverd array shall be zero. + * for now, feature_flags and the reserved array shall be zero. */ u32 _pad; - u64 reserverd[7]; + u64 reserved[7]; } __packed; -/* 80 bytes, FIXED for the next century */ struct p_barrier { - struct p_header80 head; u32 barrier; /* barrier number _handle_ only */ u32 pad; /* to multiple of 8 Byte */ } __packed; struct p_barrier_ack { - struct p_header80 head; u32 barrier; u32 set_size; } __packed; struct p_rs_param { - struct p_header80 head; - u32 rate; + u32 resync_rate; /* Since protocol version 88 and higher. */ char verify_alg[0]; } __packed; struct p_rs_param_89 { - struct p_header80 head; - u32 rate; + u32 resync_rate; /* protocol version 89: */ char verify_alg[SHARED_SECRET_MAX]; char csums_alg[SHARED_SECRET_MAX]; } __packed; struct p_rs_param_95 { - struct p_header80 head; - u32 rate; + u32 resync_rate; char verify_alg[SHARED_SECRET_MAX]; char csums_alg[SHARED_SECRET_MAX]; u32 c_plan_ahead; @@ -475,12 +409,11 @@ struct p_rs_param_95 { } __packed; enum drbd_conn_flags { - CF_WANT_LOSE = 1, + CF_DISCARD_MY_DATA = 1, CF_DRY_RUN = 2, }; struct p_protocol { - struct p_header80 head; u32 protocol; u32 after_sb_0p; u32 after_sb_1p; @@ -494,17 +427,14 @@ struct p_protocol { } __packed; struct p_uuids { - struct p_header80 head; u64 uuid[UI_EXTENDED_SIZE]; } __packed; struct p_rs_uuid { - struct p_header80 head; u64 uuid; } __packed; struct p_sizes { - struct p_header80 head; u64 d_size; /* size of disk */ u64 u_size; /* user requested size */ u64 c_size; /* current exported size */ @@ -514,18 +444,15 @@ struct p_sizes { } __packed; struct p_state { - struct p_header80 head; u32 state; } __packed; struct p_req_state { - struct p_header80 head; u32 mask; u32 val; } __packed; struct p_req_state_reply { - struct p_header80 head; u32 retcode; } __packed; @@ -539,15 +466,7 @@ struct p_drbd06_param { u32 bit_map_gen[5]; } __packed; -struct p_discard { - struct p_header80 head; - u64 block_id; - u32 seq_num; - u32 pad; -} __packed; - struct p_block_desc { - struct p_header80 head; u64 sector; u32 blksize; u32 pad; /* to multiple of 8 Byte */ @@ -563,7 +482,6 @@ enum drbd_bitmap_code { }; struct p_compressed_bm { - struct p_header80 head; /* (encoding & 0x0f): actual encoding, see enum drbd_bitmap_code * (encoding & 0x80): polarity (set/unset) of first runlength * ((encoding >> 4) & 0x07): pad_bits, number of trailing zero bits @@ -575,90 +493,22 @@ struct p_compressed_bm { } __packed; struct p_delay_probe93 { - struct p_header80 head; u32 seq_num; /* sequence number to match the two probe packets */ u32 offset; /* usecs the probe got sent after the reference time point */ } __packed; -/* DCBP: Drbd Compressed Bitmap Packet ... */ -static inline enum drbd_bitmap_code -DCBP_get_code(struct p_compressed_bm *p) -{ - return (enum drbd_bitmap_code)(p->encoding & 0x0f); -} - -static inline void -DCBP_set_code(struct p_compressed_bm *p, enum drbd_bitmap_code code) -{ - BUG_ON(code & ~0xf); - p->encoding = (p->encoding & ~0xf) | code; -} - -static inline int -DCBP_get_start(struct p_compressed_bm *p) -{ - return (p->encoding & 0x80) != 0; -} - -static inline void -DCBP_set_start(struct p_compressed_bm *p, int set) -{ - p->encoding = (p->encoding & ~0x80) | (set ? 0x80 : 0); -} - -static inline int -DCBP_get_pad_bits(struct p_compressed_bm *p) -{ - return (p->encoding >> 4) & 0x7; -} - -static inline void -DCBP_set_pad_bits(struct p_compressed_bm *p, int n) -{ - BUG_ON(n & ~0x7); - p->encoding = (p->encoding & (~0x7 << 4)) | (n << 4); -} - -/* one bitmap packet, including the p_header, - * should fit within one _architecture independend_ page. - * so we need to use the fixed size 4KiB page size - * most architectures have used for a long time. +/* + * Bitmap packets need to fit within a single page on the sender and receiver, + * so we are limited to 4 KiB (and not to PAGE_SIZE, which can be bigger). */ -#define BM_PACKET_PAYLOAD_BYTES (4096 - sizeof(struct p_header80)) -#define BM_PACKET_WORDS (BM_PACKET_PAYLOAD_BYTES/sizeof(long)) -#define BM_PACKET_VLI_BYTES_MAX (4096 - sizeof(struct p_compressed_bm)) -#if (PAGE_SIZE < 4096) -/* drbd_send_bitmap / receive_bitmap would break horribly */ -#error "PAGE_SIZE too small" -#endif - -union p_polymorph { - union p_header header; - struct p_handshake handshake; - struct p_data data; - struct p_block_ack block_ack; - struct p_barrier barrier; - struct p_barrier_ack barrier_ack; - struct p_rs_param_89 rs_param_89; - struct p_rs_param_95 rs_param_95; - struct p_protocol protocol; - struct p_sizes sizes; - struct p_uuids uuids; - struct p_state state; - struct p_req_state req_state; - struct p_req_state_reply req_state_reply; - struct p_block_req block_req; - struct p_delay_probe93 delay_probe93; - struct p_rs_uuid rs_uuid; - struct p_block_desc block_desc; -} __packed; +#define DRBD_SOCKET_BUFFER_SIZE 4096 /**********************************************************************/ enum drbd_thread_state { - None, - Running, - Exiting, - Restarting + NONE, + RUNNING, + EXITING, + RESTARTING }; struct drbd_thread { @@ -667,8 +517,9 @@ struct drbd_thread { struct completion stop; enum drbd_thread_state t_state; int (*function) (struct drbd_thread *); - struct drbd_conf *mdev; + struct drbd_tconn *tconn; int reset_cpu_mask; + char name[9]; }; static inline enum drbd_thread_state get_t_state(struct drbd_thread *thi) @@ -681,58 +532,54 @@ static inline enum drbd_thread_state get_t_state(struct drbd_thread *thi) return thi->t_state; } -struct drbd_work; -typedef int (*drbd_work_cb)(struct drbd_conf *, struct drbd_work *, int cancel); struct drbd_work { struct list_head list; - drbd_work_cb cb; + int (*cb)(struct drbd_work *, int cancel); + union { + struct drbd_conf *mdev; + struct drbd_tconn *tconn; + }; }; -struct drbd_tl_epoch; +#include "drbd_interval.h" + +extern int drbd_wait_misc(struct drbd_conf *, struct drbd_interval *); + struct drbd_request { struct drbd_work w; - struct drbd_conf *mdev; /* if local IO is not allowed, will be NULL. * if local IO _is_ allowed, holds the locally submitted bio clone, * or, after local IO completion, the ERR_PTR(error). - * see drbd_endio_pri(). */ + * see drbd_request_endio(). */ struct bio *private_bio; - struct hlist_node collision; - sector_t sector; - unsigned int size; - unsigned int epoch; /* barrier_nr */ + struct drbd_interval i; - /* barrier_nr: used to check on "completion" whether this req was in + /* epoch: used to check on "completion" whether this req was in * the current epoch, and we therefore have to close it, - * starting a new epoch... + * causing a p_barrier packet to be send, starting a new epoch. + * + * This corresponds to "barrier" in struct p_barrier[_ack], + * and to "barrier_nr" in struct drbd_epoch (and various + * comments/function parameters/local variable names). */ + unsigned int epoch; struct list_head tl_requests; /* ring list in the transfer log */ struct bio *master_bio; /* master bio pointer */ - unsigned long rq_state; /* see comments above _req_mod() */ unsigned long start_time; -}; - -struct drbd_tl_epoch { - struct drbd_work w; - struct list_head requests; /* requests before */ - struct drbd_tl_epoch *next; /* pointer to the next barrier */ - unsigned int br_number; /* the barriers identifier. */ - int n_writes; /* number of requests attached before this barrier */ -}; -struct drbd_request; + /* once it hits 0, we may complete the master_bio */ + atomic_t completion_ref; + /* once it hits 0, we may destroy this drbd_request object */ + struct kref kref; -/* These Tl_epoch_entries may be in one of 6 lists: - active_ee .. data packet being written - sync_ee .. syncer block being written - done_ee .. block written, need to send P_WRITE_ACK - read_ee .. [RS]P_DATA_REQUEST being read -*/ + unsigned rq_state; /* see comments above _req_mod() */ +}; struct drbd_epoch { + struct drbd_tconn *tconn; struct list_head list; unsigned int barrier_nr; atomic_t epoch_size; /* increased on every request added. */ @@ -762,17 +609,14 @@ struct digest_info { void *digest; }; -struct drbd_epoch_entry { +struct drbd_peer_request { struct drbd_work w; - struct hlist_node collision; struct drbd_epoch *epoch; /* for writes */ - struct drbd_conf *mdev; struct page *pages; atomic_t pending_bios; - unsigned int size; + struct drbd_interval i; /* see comments on ee flag bits below */ unsigned long flags; - sector_t sector; union { u64 block_id; struct digest_info *digest; @@ -793,31 +637,37 @@ enum { * we need to resubmit without the barrier flag. */ __EE_RESUBMITTED, - /* we may have several bios per epoch entry. + /* we may have several bios per peer request. * if any of those fail, we set this flag atomically * from the endio callback */ __EE_WAS_ERROR, /* This ee has a pointer to a digest instead of a block id */ __EE_HAS_DIGEST, + + /* Conflicting local requests need to be restarted after this request */ + __EE_RESTART_REQUESTS, + + /* The peer wants a write ACK for this (wire proto C) */ + __EE_SEND_WRITE_ACK, + + /* Is set when net_conf had two_primaries set while creating this peer_req */ + __EE_IN_INTERVAL_TREE, }; #define EE_CALL_AL_COMPLETE_IO (1<<__EE_CALL_AL_COMPLETE_IO) #define EE_MAY_SET_IN_SYNC (1<<__EE_MAY_SET_IN_SYNC) #define EE_RESUBMITTED (1<<__EE_RESUBMITTED) #define EE_WAS_ERROR (1<<__EE_WAS_ERROR) #define EE_HAS_DIGEST (1<<__EE_HAS_DIGEST) +#define EE_RESTART_REQUESTS (1<<__EE_RESTART_REQUESTS) +#define EE_SEND_WRITE_ACK (1<<__EE_SEND_WRITE_ACK) +#define EE_IN_INTERVAL_TREE (1<<__EE_IN_INTERVAL_TREE) -/* global flag bits */ +/* flag bits per mdev */ enum { - CREATE_BARRIER, /* next P_DATA is preceded by a P_BARRIER */ - SIGNAL_ASENDER, /* whether asender wants to be interrupted */ - SEND_PING, /* whether asender should send a ping asap */ - UNPLUG_REMOTE, /* sending a "UnplugRemote" could help */ MD_DIRTY, /* current uuids and flags not yet on disk */ - DISCARD_CONCURRENT, /* Set on one node, cleared on the peer! */ USE_DEGR_WFC_T, /* degr-wfc-timeout instead of wfc-timeout. */ - CLUSTER_ST_CHANGE, /* Cluster wide state change going on... */ CL_ST_CHG_SUCCESS, CL_ST_CHG_FAIL, CRASHED_PRIMARY, /* This node was a crashed primary. @@ -831,32 +681,18 @@ enum { once no more io in flight, start bitmap io */ BITMAP_IO_QUEUED, /* Started bitmap IO */ GO_DISKLESS, /* Disk is being detached, on io-error or admin request. */ - WAS_IO_ERROR, /* Local disk failed returned IO error */ + WAS_IO_ERROR, /* Local disk failed, returned IO error */ + WAS_READ_ERROR, /* Local disk READ failed (set additionally to the above) */ FORCE_DETACH, /* Force-detach from local disk, aborting any pending local IO */ RESYNC_AFTER_NEG, /* Resync after online grow after the attach&negotiate finished. */ - NET_CONGESTED, /* The data socket is congested */ - - CONFIG_PENDING, /* serialization of (re)configuration requests. - * if set, also prevents the device from dying */ - DEVICE_DYING, /* device became unconfigured, - * but worker thread is still handling the cleanup. - * reconfiguring (nl_disk_conf, nl_net_conf) is dissalowed, - * while this is set. */ RESIZE_PENDING, /* Size change detected locally, waiting for the response from * the peer, if it changed there as well. */ - CONN_DRY_RUN, /* Expect disconnect after resync handshake. */ - GOT_PING_ACK, /* set when we receive a ping_ack packet, misc wait gets woken */ NEW_CUR_UUID, /* Create new current UUID when thawing IO */ AL_SUSPENDED, /* Activity logging is currently suspended. */ AHEAD_TO_SYNC_SOURCE, /* Ahead -> SyncSource queued */ - STATE_SENT, /* Do not change state/UUIDs while this is set */ - - CALLBACK_PENDING, /* Whether we have a call_usermodehelper(, UMH_WAIT_PROC) - * pending, from drbd worker context. - * If set, bdi_write_congested() returns true, - * so shrink_page_list() would not recurse into, - * and potentially deadlock on, this drbd worker. - */ + B_RS_H_DONE, /* Before resync handler done (already executed) */ + DISCARD_MY_DATA, /* discard_my_data flag per volume */ + READ_BALANCE_RR, }; struct drbd_bitmap; /* opaque for drbd_conf */ @@ -894,24 +730,24 @@ enum bm_flag { struct drbd_work_queue { struct list_head q; - struct semaphore s; /* producers up it, worker down()s it */ spinlock_t q_lock; /* to protect the list. */ + wait_queue_head_t q_wait; }; struct drbd_socket { - struct drbd_work_queue work; struct mutex mutex; struct socket *socket; /* this way we get our * send/receive buffers off the stack */ - union p_polymorph sbuf; - union p_polymorph rbuf; + void *sbuf; + void *rbuf; }; struct drbd_md { u64 md_offset; /* sector offset to 'super' block */ u64 la_size_sect; /* last agreed size, unit sectors */ + spinlock_t uuid_lock; u64 uuid[UI_SIZE]; u64 device_uuid; u32 flags; @@ -921,24 +757,16 @@ struct drbd_md { s32 bm_offset; /* signed relative sector offset to bitmap */ /* u32 al_nr_extents; important for restoring the AL - * is stored into sync_conf.al_extents, which in turn + * is stored into ldev->dc.al_extents, which in turn * gets applied to act_log->nr_elements */ }; -/* for sync_conf and other types... */ -#define NL_PACKET(name, number, fields) struct name { fields }; -#define NL_INTEGER(pn,pr,member) int member; -#define NL_INT64(pn,pr,member) __u64 member; -#define NL_BIT(pn,pr,member) unsigned member:1; -#define NL_STRING(pn,pr,member,len) unsigned char member[len]; int member ## _len; -#include <linux/drbd_nl.h> - struct drbd_backing_dev { struct block_device *backing_bdev; struct block_device *md_bdev; struct drbd_md md; - struct disk_conf dc; /* The user provided config... */ + struct disk_conf *disk_conf; /* RCU, for updates: mdev->tconn->conf_update */ sector_t known_size; /* last known size of that backing device */ }; @@ -962,18 +790,116 @@ enum write_ordering_e { }; struct fifo_buffer { - int *values; unsigned int head_index; unsigned int size; + int total; /* sum of all values */ + int values[0]; +}; +extern struct fifo_buffer *fifo_alloc(int fifo_size); + +/* flag bits per tconn */ +enum { + NET_CONGESTED, /* The data socket is congested */ + RESOLVE_CONFLICTS, /* Set on one node, cleared on the peer! */ + SEND_PING, /* whether asender should send a ping asap */ + SIGNAL_ASENDER, /* whether asender wants to be interrupted */ + GOT_PING_ACK, /* set when we receive a ping_ack packet, ping_wait gets woken */ + CONN_WD_ST_CHG_REQ, /* A cluster wide state change on the connection is active */ + CONN_WD_ST_CHG_OKAY, + CONN_WD_ST_CHG_FAIL, + CONN_DRY_RUN, /* Expect disconnect after resync handshake. */ + CREATE_BARRIER, /* next P_DATA is preceded by a P_BARRIER */ + STATE_SENT, /* Do not change state/UUIDs while this is set */ + CALLBACK_PENDING, /* Whether we have a call_usermodehelper(, UMH_WAIT_PROC) + * pending, from drbd worker context. + * If set, bdi_write_congested() returns true, + * so shrink_page_list() would not recurse into, + * and potentially deadlock on, this drbd worker. + */ + DISCONNECT_SENT, +}; + +struct drbd_tconn { /* is a resource from the config file */ + char *name; /* Resource name */ + struct list_head all_tconn; /* linked on global drbd_tconns */ + struct kref kref; + struct idr volumes; /* <tconn, vnr> to mdev mapping */ + enum drbd_conns cstate; /* Only C_STANDALONE to C_WF_REPORT_PARAMS */ + unsigned susp:1; /* IO suspended by user */ + unsigned susp_nod:1; /* IO suspended because no data */ + unsigned susp_fen:1; /* IO suspended because fence peer handler runs */ + struct mutex cstate_mutex; /* Protects graceful disconnects */ + + unsigned long flags; + struct net_conf *net_conf; /* content protected by rcu */ + struct mutex conf_update; /* mutex for ready-copy-update of net_conf and disk_conf */ + wait_queue_head_t ping_wait; /* Woken upon reception of a ping, and a state change */ + struct res_opts res_opts; + + struct sockaddr_storage my_addr; + int my_addr_len; + struct sockaddr_storage peer_addr; + int peer_addr_len; + + struct drbd_socket data; /* data/barrier/cstate/parameter packets */ + struct drbd_socket meta; /* ping/ack (metadata) packets */ + int agreed_pro_version; /* actually used protocol version */ + unsigned long last_received; /* in jiffies, either socket */ + unsigned int ko_count; + + spinlock_t req_lock; + + struct list_head transfer_log; /* all requests not yet fully processed */ + + struct crypto_hash *cram_hmac_tfm; + struct crypto_hash *integrity_tfm; /* checksums we compute, updates protected by tconn->data->mutex */ + struct crypto_hash *peer_integrity_tfm; /* checksums we verify, only accessed from receiver thread */ + struct crypto_hash *csums_tfm; + struct crypto_hash *verify_tfm; + void *int_dig_in; + void *int_dig_vv; + + /* receiver side */ + struct drbd_epoch *current_epoch; + spinlock_t epoch_lock; + unsigned int epochs; + enum write_ordering_e write_ordering; + atomic_t current_tle_nr; /* transfer log epoch number */ + unsigned current_tle_writes; /* writes seen within this tl epoch */ + + unsigned long last_reconnect_jif; + struct drbd_thread receiver; + struct drbd_thread worker; + struct drbd_thread asender; + cpumask_var_t cpu_mask; + + /* sender side */ + struct drbd_work_queue sender_work; + + struct { + /* whether this sender thread + * has processed a single write yet. */ + bool seen_any_write_yet; + + /* Which barrier number to send with the next P_BARRIER */ + int current_epoch_nr; + + /* how many write requests have been sent + * with req->epoch == current_epoch_nr. + * If none, no P_BARRIER will be sent. */ + unsigned current_epoch_writes; + } send; }; struct drbd_conf { + struct drbd_tconn *tconn; + int vnr; /* volume number within the connection */ + struct kref kref; + /* things that are stored as / read from meta data on disk */ unsigned long flags; /* configured by drbdsetup */ - struct net_conf *net_conf; /* protected by get_net_conf() and put_net_conf() */ - struct syncer_conf sync_conf; struct drbd_backing_dev *ldev __protected_by(local); sector_t p_size; /* partner's disk size */ @@ -981,11 +907,7 @@ struct drbd_conf { struct block_device *this_bdev; struct gendisk *vdisk; - struct drbd_socket data; /* data/barrier/cstate/parameter packets */ - struct drbd_socket meta; /* ping/ack (metadata) packets */ - int agreed_pro_version; /* actually used protocol version */ - unsigned long last_received; /* in jiffies, either socket */ - unsigned int ko_count; + unsigned long last_reattach_jif; struct drbd_work resync_work, unplug_work, go_diskless, @@ -1005,10 +927,9 @@ struct drbd_conf { /* Used after attach while negotiating new disk state. */ union drbd_state new_state_tmp; - union drbd_state state; + union drbd_dev_state state; wait_queue_head_t misc_wait; wait_queue_head_t state_wait; /* upon each state change. */ - wait_queue_head_t net_cnt_wait; unsigned int send_cnt; unsigned int recv_cnt; unsigned int read_cnt; @@ -1018,17 +939,12 @@ struct drbd_conf { atomic_t ap_bio_cnt; /* Requests we need to complete */ atomic_t ap_pending_cnt; /* AP data packets on the wire, ack expected */ atomic_t rs_pending_cnt; /* RS request/data packets on the wire */ - atomic_t unacked_cnt; /* Need to send replys for */ + atomic_t unacked_cnt; /* Need to send replies for */ atomic_t local_cnt; /* Waiting for local completion */ - atomic_t net_cnt; /* Users of net_conf */ - spinlock_t req_lock; - struct drbd_tl_epoch *unused_spare_tle; /* for pre-allocation */ - struct drbd_tl_epoch *newest_tle; - struct drbd_tl_epoch *oldest_tle; - struct list_head out_of_sequence_requests; - struct list_head barrier_acked_requests; - struct hlist_head *tl_hash; - unsigned int tl_hash_s; + + /* Interval tree of pending local requests */ + struct rb_root read_requests; + struct rb_root write_requests; /* blocks to resync in this run [unit BM_BLOCK_SIZE] */ unsigned long rs_total; @@ -1048,9 +964,11 @@ struct drbd_conf { unsigned long rs_mark_time[DRBD_SYNC_MARKS]; /* current index into rs_mark_{left,time} */ int rs_last_mark; + unsigned long rs_last_bcast; /* [unit jiffies] */ /* where does the admin want us to start? (sector) */ sector_t ov_start_sector; + sector_t ov_stop_sector; /* where are we now? (sector) */ sector_t ov_position; /* Start sector of out of sync range (to merge printk reporting). */ @@ -1058,14 +976,7 @@ struct drbd_conf { /* size of out-of-sync range in sectors. */ sector_t ov_last_oos_size; unsigned long ov_left; /* in bits */ - struct crypto_hash *csums_tfm; - struct crypto_hash *verify_tfm; - unsigned long last_reattach_jif; - unsigned long last_reconnect_jif; - struct drbd_thread receiver; - struct drbd_thread worker; - struct drbd_thread asender; struct drbd_bitmap *bitmap; unsigned long bm_resync_fo; /* bit offset for drbd_bm_find_next */ @@ -1078,29 +989,19 @@ struct drbd_conf { int open_cnt; u64 *p_uuid; - struct drbd_epoch *current_epoch; - spinlock_t epoch_lock; - unsigned int epochs; - enum write_ordering_e write_ordering; + struct list_head active_ee; /* IO in progress (P_DATA gets written to disk) */ struct list_head sync_ee; /* IO in progress (P_RS_DATA_REPLY gets written to disk) */ - struct list_head done_ee; /* send ack */ - struct list_head read_ee; /* IO in progress (any read) */ + struct list_head done_ee; /* need to send P_WRITE_ACK */ + struct list_head read_ee; /* [RS]P_DATA_REQUEST being read */ struct list_head net_ee; /* zero-copy network send in progress */ - struct hlist_head *ee_hash; /* is proteced by req_lock! */ - unsigned int ee_hash_s; - - /* this one is protected by ee_lock, single thread */ - struct drbd_epoch_entry *last_write_w_barrier; int next_barrier_nr; - struct hlist_head *app_reads_hash; /* is proteced by req_lock */ struct list_head resync_reads; atomic_t pp_in_use; /* allocated from page pool */ atomic_t pp_in_use_by_net; /* sendpage()d, still referenced by tcp */ wait_queue_head_t ee_wait; struct page *md_io_page; /* one page buffer for md_io */ - struct page *md_io_tmpp; /* for logical_block_size != 512 */ struct drbd_md_io md_io; atomic_t md_io_in_use; /* protects the md_io, md_io_page and md_io_tmpp */ spinlock_t al_lock; @@ -1109,22 +1010,16 @@ struct drbd_conf { unsigned int al_tr_number; int al_tr_cycle; int al_tr_pos; /* position of the next transaction in the journal */ - struct crypto_hash *cram_hmac_tfm; - struct crypto_hash *integrity_w_tfm; /* to be used by the worker thread */ - struct crypto_hash *integrity_r_tfm; /* to be used by the receiver thread */ - void *int_dig_out; - void *int_dig_in; - void *int_dig_vv; wait_queue_head_t seq_wait; atomic_t packet_seq; unsigned int peer_seq; spinlock_t peer_seq_lock; unsigned int minor; unsigned long comm_bm_set; /* communicated number of set bits. */ - cpumask_var_t cpu_mask; struct bm_io_work bm_io_work; u64 ed_uuid; /* UUID of the exposed data */ - struct mutex state_mutex; + struct mutex own_state_mutex; + struct mutex *state_mutex; /* either own_state_mutex or mdev->tconn->cstate_mutex */ char congestion_reason; /* Why we where congested... */ atomic_t rs_sect_in; /* for incoming resync data rate, SyncTarget */ atomic_t rs_sect_ev; /* for submitted resync data rate, both */ @@ -1132,9 +1027,8 @@ struct drbd_conf { int rs_last_events; /* counter of read or write "events" (unit sectors) * on the lower level device when we last looked. */ int c_sync_rate; /* current resync rate after syncer throttle magic */ - struct fifo_buffer rs_plan_s; /* correction values of resync planer */ + struct fifo_buffer *rs_plan_s; /* correction values of resync planer (RCU, tconn->conn_update) */ int rs_in_flight; /* resync sectors in flight (to proxy, in proxy and from proxy) */ - int rs_planed; /* resync sectors already planned */ atomic_t ap_in_flight; /* App sectors in flight (waiting for ack) */ unsigned int peer_max_bio_size; unsigned int local_max_bio_size; @@ -1142,11 +1036,7 @@ struct drbd_conf { static inline struct drbd_conf *minor_to_mdev(unsigned int minor) { - struct drbd_conf *mdev; - - mdev = minor < minor_count ? minor_table[minor] : NULL; - - return mdev; + return (struct drbd_conf *)idr_find(&minors, minor); } static inline unsigned int mdev_to_minor(struct drbd_conf *mdev) @@ -1154,29 +1044,9 @@ static inline unsigned int mdev_to_minor(struct drbd_conf *mdev) return mdev->minor; } -/* returns 1 if it was successful, - * returns 0 if there was no data socket. - * so wherever you are going to use the data.socket, e.g. do - * if (!drbd_get_data_sock(mdev)) - * return 0; - * CODE(); - * drbd_put_data_sock(mdev); - */ -static inline int drbd_get_data_sock(struct drbd_conf *mdev) -{ - mutex_lock(&mdev->data.mutex); - /* drbd_disconnect() could have called drbd_free_sock() - * while we were waiting in down()... */ - if (unlikely(mdev->data.socket == NULL)) { - mutex_unlock(&mdev->data.mutex); - return 0; - } - return 1; -} - -static inline void drbd_put_data_sock(struct drbd_conf *mdev) +static inline struct drbd_conf *vnr_to_mdev(struct drbd_tconn *tconn, int vnr) { - mutex_unlock(&mdev->data.mutex); + return (struct drbd_conf *)idr_find(&tconn->volumes, vnr); } /* @@ -1185,106 +1055,77 @@ static inline void drbd_put_data_sock(struct drbd_conf *mdev) /* drbd_main.c */ -enum chg_state_flags { - CS_HARD = 1, - CS_VERBOSE = 2, - CS_WAIT_COMPLETE = 4, - CS_SERIALIZE = 8, - CS_ORDERED = CS_WAIT_COMPLETE + CS_SERIALIZE, -}; - enum dds_flags { DDSF_FORCED = 1, DDSF_NO_RESYNC = 2, /* Do not run a resync for the new space */ }; extern void drbd_init_set_defaults(struct drbd_conf *mdev); -extern enum drbd_state_rv drbd_change_state(struct drbd_conf *mdev, - enum chg_state_flags f, - union drbd_state mask, - union drbd_state val); -extern void drbd_force_state(struct drbd_conf *, union drbd_state, - union drbd_state); -extern enum drbd_state_rv _drbd_request_state(struct drbd_conf *, - union drbd_state, - union drbd_state, - enum chg_state_flags); -extern enum drbd_state_rv __drbd_set_state(struct drbd_conf *, union drbd_state, - enum chg_state_flags, - struct completion *done); -extern void print_st_err(struct drbd_conf *, union drbd_state, - union drbd_state, int); extern int drbd_thread_start(struct drbd_thread *thi); extern void _drbd_thread_stop(struct drbd_thread *thi, int restart, int wait); +extern char *drbd_task_to_thread_name(struct drbd_tconn *tconn, struct task_struct *task); #ifdef CONFIG_SMP -extern void drbd_thread_current_set_cpu(struct drbd_conf *mdev); -extern void drbd_calc_cpu_mask(struct drbd_conf *mdev); +extern void drbd_thread_current_set_cpu(struct drbd_thread *thi); +extern void drbd_calc_cpu_mask(struct drbd_tconn *tconn); #else #define drbd_thread_current_set_cpu(A) ({}) #define drbd_calc_cpu_mask(A) ({}) #endif -extern void drbd_free_resources(struct drbd_conf *mdev); -extern void tl_release(struct drbd_conf *mdev, unsigned int barrier_nr, +extern void tl_release(struct drbd_tconn *, unsigned int barrier_nr, unsigned int set_size); -extern void tl_clear(struct drbd_conf *mdev); -extern void _tl_add_barrier(struct drbd_conf *, struct drbd_tl_epoch *); -extern void drbd_free_sock(struct drbd_conf *mdev); -extern int drbd_send(struct drbd_conf *mdev, struct socket *sock, - void *buf, size_t size, unsigned msg_flags); -extern int drbd_send_protocol(struct drbd_conf *mdev); +extern void tl_clear(struct drbd_tconn *); +extern void drbd_free_sock(struct drbd_tconn *tconn); +extern int drbd_send(struct drbd_tconn *tconn, struct socket *sock, + void *buf, size_t size, unsigned msg_flags); +extern int drbd_send_all(struct drbd_tconn *, struct socket *, void *, size_t, + unsigned); + +extern int __drbd_send_protocol(struct drbd_tconn *tconn, enum drbd_packet cmd); +extern int drbd_send_protocol(struct drbd_tconn *tconn); extern int drbd_send_uuids(struct drbd_conf *mdev); extern int drbd_send_uuids_skip_initial_sync(struct drbd_conf *mdev); -extern int drbd_gen_and_send_sync_uuid(struct drbd_conf *mdev); +extern void drbd_gen_and_send_sync_uuid(struct drbd_conf *mdev); extern int drbd_send_sizes(struct drbd_conf *mdev, int trigger_reply, enum dds_flags flags); extern int drbd_send_state(struct drbd_conf *mdev, union drbd_state s); extern int drbd_send_current_state(struct drbd_conf *mdev); -extern int _drbd_send_cmd(struct drbd_conf *mdev, struct socket *sock, - enum drbd_packets cmd, struct p_header80 *h, - size_t size, unsigned msg_flags); -#define USE_DATA_SOCKET 1 -#define USE_META_SOCKET 0 -extern int drbd_send_cmd(struct drbd_conf *mdev, int use_data_socket, - enum drbd_packets cmd, struct p_header80 *h, - size_t size); -extern int drbd_send_cmd2(struct drbd_conf *mdev, enum drbd_packets cmd, - char *data, size_t size); -extern int drbd_send_sync_param(struct drbd_conf *mdev, struct syncer_conf *sc); -extern int drbd_send_b_ack(struct drbd_conf *mdev, u32 barrier_nr, - u32 set_size); -extern int drbd_send_ack(struct drbd_conf *mdev, enum drbd_packets cmd, - struct drbd_epoch_entry *e); -extern int drbd_send_ack_rp(struct drbd_conf *mdev, enum drbd_packets cmd, - struct p_block_req *rp); -extern int drbd_send_ack_dp(struct drbd_conf *mdev, enum drbd_packets cmd, - struct p_data *dp, int data_size); -extern int drbd_send_ack_ex(struct drbd_conf *mdev, enum drbd_packets cmd, +extern int drbd_send_sync_param(struct drbd_conf *mdev); +extern void drbd_send_b_ack(struct drbd_tconn *tconn, u32 barrier_nr, + u32 set_size); +extern int drbd_send_ack(struct drbd_conf *, enum drbd_packet, + struct drbd_peer_request *); +extern void drbd_send_ack_rp(struct drbd_conf *mdev, enum drbd_packet cmd, + struct p_block_req *rp); +extern void drbd_send_ack_dp(struct drbd_conf *mdev, enum drbd_packet cmd, + struct p_data *dp, int data_size); +extern int drbd_send_ack_ex(struct drbd_conf *mdev, enum drbd_packet cmd, sector_t sector, int blksize, u64 block_id); -extern int drbd_send_oos(struct drbd_conf *mdev, struct drbd_request *req); -extern int drbd_send_block(struct drbd_conf *mdev, enum drbd_packets cmd, - struct drbd_epoch_entry *e); +extern int drbd_send_out_of_sync(struct drbd_conf *, struct drbd_request *); +extern int drbd_send_block(struct drbd_conf *, enum drbd_packet, + struct drbd_peer_request *); extern int drbd_send_dblock(struct drbd_conf *mdev, struct drbd_request *req); extern int drbd_send_drequest(struct drbd_conf *mdev, int cmd, sector_t sector, int size, u64 block_id); -extern int drbd_send_drequest_csum(struct drbd_conf *mdev, - sector_t sector,int size, - void *digest, int digest_size, - enum drbd_packets cmd); +extern int drbd_send_drequest_csum(struct drbd_conf *mdev, sector_t sector, + int size, void *digest, int digest_size, + enum drbd_packet cmd); extern int drbd_send_ov_request(struct drbd_conf *mdev,sector_t sector,int size); extern int drbd_send_bitmap(struct drbd_conf *mdev); -extern int _drbd_send_bitmap(struct drbd_conf *mdev); -extern int drbd_send_sr_reply(struct drbd_conf *mdev, enum drbd_state_rv retcode); +extern void drbd_send_sr_reply(struct drbd_conf *mdev, enum drbd_state_rv retcode); +extern void conn_send_sr_reply(struct drbd_tconn *tconn, enum drbd_state_rv retcode); extern void drbd_free_bc(struct drbd_backing_dev *ldev); extern void drbd_mdev_cleanup(struct drbd_conf *mdev); void drbd_print_uuids(struct drbd_conf *mdev, const char *text); +extern void conn_md_sync(struct drbd_tconn *tconn); extern void drbd_md_sync(struct drbd_conf *mdev); extern int drbd_md_read(struct drbd_conf *mdev, struct drbd_backing_dev *bdev); extern void drbd_uuid_set(struct drbd_conf *mdev, int idx, u64 val) __must_hold(local); extern void _drbd_uuid_set(struct drbd_conf *mdev, int idx, u64 val) __must_hold(local); extern void drbd_uuid_new_current(struct drbd_conf *mdev) __must_hold(local); -extern void _drbd_uuid_new_current(struct drbd_conf *mdev) __must_hold(local); extern void drbd_uuid_set_bm(struct drbd_conf *mdev, u64 val) __must_hold(local); +extern void drbd_uuid_move_history(struct drbd_conf *mdev) __must_hold(local); +extern void __drbd_uuid_set(struct drbd_conf *mdev, int idx, u64 val) __must_hold(local); extern void drbd_md_set_flag(struct drbd_conf *mdev, int flags) __must_hold(local); extern void drbd_md_clear_flag(struct drbd_conf *mdev, int flags)__must_hold(local); extern int drbd_md_test_flag(struct drbd_backing_dev *, int); @@ -1302,33 +1143,52 @@ extern void drbd_queue_bitmap_io(struct drbd_conf *mdev, extern int drbd_bitmap_io(struct drbd_conf *mdev, int (*io_fn)(struct drbd_conf *), char *why, enum bm_flag flags); +extern int drbd_bitmap_io_from_worker(struct drbd_conf *mdev, + int (*io_fn)(struct drbd_conf *), + char *why, enum bm_flag flags); extern int drbd_bmio_set_n_write(struct drbd_conf *mdev); extern int drbd_bmio_clear_n_write(struct drbd_conf *mdev); extern void drbd_go_diskless(struct drbd_conf *mdev); extern void drbd_ldev_destroy(struct drbd_conf *mdev); - /* Meta data layout We reserve a 128MB Block (4k aligned) * either at the end of the backing device * or on a separate meta data device. */ -#define MD_RESERVED_SECT (128LU << 11) /* 128 MB, unit sectors */ /* The following numbers are sectors */ -#define MD_AL_OFFSET 8 /* 8 Sectors after start of meta area */ -#define MD_AL_MAX_SIZE 64 /* = 32 kb LOG ~ 3776 extents ~ 14 GB Storage */ -/* Allows up to about 3.8TB */ -#define MD_BM_OFFSET (MD_AL_OFFSET + MD_AL_MAX_SIZE) - -/* Since the smalles IO unit is usually 512 byte */ -#define MD_SECTOR_SHIFT 9 -#define MD_SECTOR_SIZE (1<<MD_SECTOR_SHIFT) - -/* activity log */ -#define AL_EXTENTS_PT ((MD_SECTOR_SIZE-12)/8-1) /* 61 ; Extents per 512B sector */ -#define AL_EXTENT_SHIFT 22 /* One extent represents 4M Storage */ +/* Allows up to about 3.8TB, so if you want more, + * you need to use the "flexible" meta data format. */ +#define MD_RESERVED_SECT (128LU << 11) /* 128 MB, unit sectors */ +#define MD_AL_OFFSET 8 /* 8 Sectors after start of meta area */ +#define MD_AL_SECTORS 64 /* = 32 kB on disk activity log ring buffer */ +#define MD_BM_OFFSET (MD_AL_OFFSET + MD_AL_SECTORS) + +/* we do all meta data IO in 4k blocks */ +#define MD_BLOCK_SHIFT 12 +#define MD_BLOCK_SIZE (1<<MD_BLOCK_SHIFT) + +/* One activity log extent represents 4M of storage */ +#define AL_EXTENT_SHIFT 22 #define AL_EXTENT_SIZE (1<<AL_EXTENT_SHIFT) +/* We could make these currently hardcoded constants configurable + * variables at create-md time (or even re-configurable at runtime?). + * Which will require some more changes to the DRBD "super block" + * and attach code. + * + * updates per transaction: + * This many changes to the active set can be logged with one transaction. + * This number is arbitrary. + * context per transaction: + * This many context extent numbers are logged with each transaction. + * This number is resulting from the transaction block size (4k), the layout + * of the transaction header, and the number of updates per transaction. + * See drbd_actlog.c:struct al_transaction_on_disk + * */ +#define AL_UPDATES_PER_TRANSACTION 64 // arbitrary +#define AL_CONTEXT_PER_TRANSACTION 919 // (4096 - 36 - 6*64)/4 + #if BITS_PER_LONG == 32 #define LN2_BPL 5 #define cpu_to_lel(A) cpu_to_le32(A) @@ -1364,11 +1224,14 @@ struct bm_extent { #define SLEEP_TIME (HZ/10) -#define BM_BLOCK_SHIFT 12 /* 4k per bit */ +/* We do bitmap IO in units of 4k blocks. + * We also still have a hardcoded 4k per bit relation. */ +#define BM_BLOCK_SHIFT 12 /* 4k per bit */ #define BM_BLOCK_SIZE (1<<BM_BLOCK_SHIFT) -/* (9+3) : 512 bytes @ 8 bits; representing 16M storage - * per sector of on disk bitmap */ -#define BM_EXT_SHIFT (BM_BLOCK_SHIFT + MD_SECTOR_SHIFT + 3) /* = 24 */ +/* mostly arbitrarily set the represented size of one bitmap extent, + * aka resync extent, to 16 MiB (which is also 512 Byte worth of bitmap + * at 4k per bit resolution) */ +#define BM_EXT_SHIFT 24 /* 16 MiB per resync extent */ #define BM_EXT_SIZE (1<<BM_EXT_SHIFT) #if (BM_EXT_SHIFT != 24) || (BM_BLOCK_SHIFT != 12) @@ -1436,17 +1299,20 @@ struct bm_extent { #endif #endif -/* Sector shift value for the "hash" functions of tl_hash and ee_hash tables. - * With a value of 8 all IO in one 128K block make it to the same slot of the - * hash table. */ -#define HT_SHIFT 8 -#define DRBD_MAX_BIO_SIZE (1U<<(9+HT_SHIFT)) +/* BIO_MAX_SIZE is 256 * PAGE_CACHE_SIZE, + * so for typical PAGE_CACHE_SIZE of 4k, that is (1<<20) Byte. + * Since we may live in a mixed-platform cluster, + * we limit us to a platform agnostic constant here for now. + * A followup commit may allow even bigger BIO sizes, + * once we thought that through. */ +#define DRBD_MAX_BIO_SIZE (1U << 20) +#if DRBD_MAX_BIO_SIZE > BIO_MAX_SIZE +#error Architecture not supported: DRBD_MAX_BIO_SIZE > BIO_MAX_SIZE +#endif #define DRBD_MAX_BIO_SIZE_SAFE (1U << 12) /* Works always = 4k */ -#define DRBD_MAX_SIZE_H80_PACKET (1U << 15) /* The old header only allows packets up to 32Kib data */ - -/* Number of elements in the app_reads_hash */ -#define APP_R_HSIZE 15 +#define DRBD_MAX_SIZE_H80_PACKET (1U << 15) /* Header 80 only allows packets up to 32KiB data */ +#define DRBD_MAX_BIO_SIZE_P95 (1U << 17) /* Protocol 95 to 99 allows bios up to 128KiB */ extern int drbd_bm_init(struct drbd_conf *mdev); extern int drbd_bm_resize(struct drbd_conf *mdev, sector_t sectors, int set_new_bits); @@ -1468,11 +1334,11 @@ extern int drbd_bm_test_bit(struct drbd_conf *mdev, unsigned long bitnr); extern int drbd_bm_e_weight(struct drbd_conf *mdev, unsigned long enr); extern int drbd_bm_write_page(struct drbd_conf *mdev, unsigned int idx) __must_hold(local); extern int drbd_bm_read(struct drbd_conf *mdev) __must_hold(local); +extern void drbd_bm_mark_for_writeout(struct drbd_conf *mdev, int page_nr); extern int drbd_bm_write(struct drbd_conf *mdev) __must_hold(local); +extern int drbd_bm_write_hinted(struct drbd_conf *mdev) __must_hold(local); extern int drbd_bm_write_all(struct drbd_conf *mdev) __must_hold(local); extern int drbd_bm_write_copy_pages(struct drbd_conf *mdev) __must_hold(local); -extern unsigned long drbd_bm_ALe_set_all(struct drbd_conf *mdev, - unsigned long al_enr); extern size_t drbd_bm_words(struct drbd_conf *mdev); extern unsigned long drbd_bm_bits(struct drbd_conf *mdev); extern sector_t drbd_bm_capacity(struct drbd_conf *mdev); @@ -1497,7 +1363,7 @@ extern void drbd_bm_unlock(struct drbd_conf *mdev); /* drbd_main.c */ extern struct kmem_cache *drbd_request_cache; -extern struct kmem_cache *drbd_ee_cache; /* epoch entries */ +extern struct kmem_cache *drbd_ee_cache; /* peer requests */ extern struct kmem_cache *drbd_bm_ext_cache; /* bitmap extents */ extern struct kmem_cache *drbd_al_ext_cache; /* activity log extents */ extern mempool_t *drbd_request_mempool; @@ -1537,12 +1403,22 @@ extern struct bio *bio_alloc_drbd(gfp_t gfp_mask); extern rwlock_t global_state_lock; -extern struct drbd_conf *drbd_new_device(unsigned int minor); -extern void drbd_free_mdev(struct drbd_conf *mdev); +extern int conn_lowest_minor(struct drbd_tconn *tconn); +enum drbd_ret_code conn_new_minor(struct drbd_tconn *tconn, unsigned int minor, int vnr); +extern void drbd_minor_destroy(struct kref *kref); + +extern int set_resource_options(struct drbd_tconn *tconn, struct res_opts *res_opts); +extern struct drbd_tconn *conn_create(const char *name, struct res_opts *res_opts); +extern void conn_destroy(struct kref *kref); +struct drbd_tconn *conn_get_by_name(const char *name); +extern struct drbd_tconn *conn_get_by_addrs(void *my_addr, int my_addr_len, + void *peer_addr, int peer_addr_len); +extern void conn_free_crypto(struct drbd_tconn *tconn); extern int proc_details; /* drbd_req */ +extern void __drbd_make_request(struct drbd_conf *, struct bio *, unsigned long); extern void drbd_make_request(struct request_queue *q, struct bio *bio); extern int drbd_read_remote(struct drbd_conf *mdev, struct drbd_request *req); extern int drbd_merge_bvec(struct request_queue *q, struct bvec_merge_data *bvm, struct bio_vec *bvec); @@ -1550,10 +1426,11 @@ extern int is_valid_ar_handle(struct drbd_request *, sector_t); /* drbd_nl.c */ +extern int drbd_msg_put_info(const char *info); extern void drbd_suspend_io(struct drbd_conf *mdev); extern void drbd_resume_io(struct drbd_conf *mdev); extern char *ppsize(char *buf, unsigned long long size); -extern sector_t drbd_new_dev_size(struct drbd_conf *, struct drbd_backing_dev *, int); +extern sector_t drbd_new_dev_size(struct drbd_conf *, struct drbd_backing_dev *, sector_t, int); enum determine_dev_size { dev_size_error = -1, unchanged = 0, shrunk = 1, grew = 2 }; extern enum determine_dev_size drbd_determine_dev_size(struct drbd_conf *, enum dds_flags) __must_hold(local); extern void resync_after_online_grow(struct drbd_conf *); @@ -1561,13 +1438,14 @@ extern void drbd_reconsider_max_bio_size(struct drbd_conf *mdev); extern enum drbd_state_rv drbd_set_role(struct drbd_conf *mdev, enum drbd_role new_role, int force); -extern enum drbd_disk_state drbd_try_outdate_peer(struct drbd_conf *mdev); -extern void drbd_try_outdate_peer_async(struct drbd_conf *mdev); +extern bool conn_try_outdate_peer(struct drbd_tconn *tconn); +extern void conn_try_outdate_peer_async(struct drbd_tconn *tconn); extern int drbd_khelper(struct drbd_conf *mdev, char *cmd); /* drbd_worker.c */ extern int drbd_worker(struct drbd_thread *thi); -extern int drbd_alter_sa(struct drbd_conf *mdev, int na); +enum drbd_ret_code drbd_resync_after_valid(struct drbd_conf *mdev, int o_minor); +void drbd_resync_after_changed(struct drbd_conf *mdev); extern void drbd_start_resync(struct drbd_conf *mdev, enum drbd_conns side); extern void resume_next_sg(struct drbd_conf *mdev); extern void suspend_other_sg(struct drbd_conf *mdev); @@ -1576,13 +1454,13 @@ extern int drbd_resync_finished(struct drbd_conf *mdev); extern void *drbd_md_get_buffer(struct drbd_conf *mdev); extern void drbd_md_put_buffer(struct drbd_conf *mdev); extern int drbd_md_sync_page_io(struct drbd_conf *mdev, - struct drbd_backing_dev *bdev, sector_t sector, int rw); -extern void wait_until_done_or_disk_failure(struct drbd_conf *mdev, struct drbd_backing_dev *bdev, - unsigned int *done); -extern void drbd_ov_oos_found(struct drbd_conf*, sector_t, int); + struct drbd_backing_dev *bdev, sector_t sector, int rw); +extern void drbd_ov_out_of_sync_found(struct drbd_conf *, sector_t, int); +extern void wait_until_done_or_force_detached(struct drbd_conf *mdev, + struct drbd_backing_dev *bdev, unsigned int *done); extern void drbd_rs_controller_reset(struct drbd_conf *mdev); -static inline void ov_oos_print(struct drbd_conf *mdev) +static inline void ov_out_of_sync_print(struct drbd_conf *mdev) { if (mdev->ov_last_oos_size) { dev_err(DEV, "Out of sync: start=%llu, size=%lu (sectors)\n", @@ -1594,97 +1472,102 @@ static inline void ov_oos_print(struct drbd_conf *mdev) extern void drbd_csum_bio(struct drbd_conf *, struct crypto_hash *, struct bio *, void *); -extern void drbd_csum_ee(struct drbd_conf *, struct crypto_hash *, struct drbd_epoch_entry *, void *); +extern void drbd_csum_ee(struct drbd_conf *, struct crypto_hash *, + struct drbd_peer_request *, void *); /* worker callbacks */ -extern int w_req_cancel_conflict(struct drbd_conf *, struct drbd_work *, int); -extern int w_read_retry_remote(struct drbd_conf *, struct drbd_work *, int); -extern int w_e_end_data_req(struct drbd_conf *, struct drbd_work *, int); -extern int w_e_end_rsdata_req(struct drbd_conf *, struct drbd_work *, int); -extern int w_e_end_csum_rs_req(struct drbd_conf *, struct drbd_work *, int); -extern int w_e_end_ov_reply(struct drbd_conf *, struct drbd_work *, int); -extern int w_e_end_ov_req(struct drbd_conf *, struct drbd_work *, int); -extern int w_ov_finished(struct drbd_conf *, struct drbd_work *, int); -extern int w_resync_timer(struct drbd_conf *, struct drbd_work *, int); -extern int w_resume_next_sg(struct drbd_conf *, struct drbd_work *, int); -extern int w_send_write_hint(struct drbd_conf *, struct drbd_work *, int); -extern int w_send_dblock(struct drbd_conf *, struct drbd_work *, int); -extern int w_send_barrier(struct drbd_conf *, struct drbd_work *, int); -extern int w_send_read_req(struct drbd_conf *, struct drbd_work *, int); -extern int w_prev_work_done(struct drbd_conf *, struct drbd_work *, int); -extern int w_e_reissue(struct drbd_conf *, struct drbd_work *, int); -extern int w_restart_disk_io(struct drbd_conf *, struct drbd_work *, int); -extern int w_send_oos(struct drbd_conf *, struct drbd_work *, int); -extern int w_start_resync(struct drbd_conf *, struct drbd_work *, int); +extern int w_e_end_data_req(struct drbd_work *, int); +extern int w_e_end_rsdata_req(struct drbd_work *, int); +extern int w_e_end_csum_rs_req(struct drbd_work *, int); +extern int w_e_end_ov_reply(struct drbd_work *, int); +extern int w_e_end_ov_req(struct drbd_work *, int); +extern int w_ov_finished(struct drbd_work *, int); +extern int w_resync_timer(struct drbd_work *, int); +extern int w_send_write_hint(struct drbd_work *, int); +extern int w_make_resync_request(struct drbd_work *, int); +extern int w_send_dblock(struct drbd_work *, int); +extern int w_send_read_req(struct drbd_work *, int); +extern int w_prev_work_done(struct drbd_work *, int); +extern int w_e_reissue(struct drbd_work *, int); +extern int w_restart_disk_io(struct drbd_work *, int); +extern int w_send_out_of_sync(struct drbd_work *, int); +extern int w_start_resync(struct drbd_work *, int); extern void resync_timer_fn(unsigned long data); extern void start_resync_timer_fn(unsigned long data); /* drbd_receiver.c */ extern int drbd_rs_should_slow_down(struct drbd_conf *mdev, sector_t sector); -extern int drbd_submit_ee(struct drbd_conf *mdev, struct drbd_epoch_entry *e, - const unsigned rw, const int fault_type); -extern int drbd_release_ee(struct drbd_conf *mdev, struct list_head *list); -extern struct drbd_epoch_entry *drbd_alloc_ee(struct drbd_conf *mdev, - u64 id, - sector_t sector, - unsigned int data_size, - gfp_t gfp_mask) __must_hold(local); -extern void drbd_free_some_ee(struct drbd_conf *mdev, struct drbd_epoch_entry *e, - int is_net); -#define drbd_free_ee(m,e) drbd_free_some_ee(m, e, 0) -#define drbd_free_net_ee(m,e) drbd_free_some_ee(m, e, 1) -extern void drbd_wait_ee_list_empty(struct drbd_conf *mdev, - struct list_head *head); -extern void _drbd_wait_ee_list_empty(struct drbd_conf *mdev, - struct list_head *head); +extern int drbd_submit_peer_request(struct drbd_conf *, + struct drbd_peer_request *, const unsigned, + const int); +extern int drbd_free_peer_reqs(struct drbd_conf *, struct list_head *); +extern struct drbd_peer_request *drbd_alloc_peer_req(struct drbd_conf *, u64, + sector_t, unsigned int, + gfp_t) __must_hold(local); +extern void __drbd_free_peer_req(struct drbd_conf *, struct drbd_peer_request *, + int); +#define drbd_free_peer_req(m,e) __drbd_free_peer_req(m, e, 0) +#define drbd_free_net_peer_req(m,e) __drbd_free_peer_req(m, e, 1) +extern struct page *drbd_alloc_pages(struct drbd_conf *, unsigned int, bool); extern void drbd_set_recv_tcq(struct drbd_conf *mdev, int tcq_enabled); extern void _drbd_clear_done_ee(struct drbd_conf *mdev, struct list_head *to_be_freed); -extern void drbd_flush_workqueue(struct drbd_conf *mdev); -extern void drbd_free_tl_hash(struct drbd_conf *mdev); +extern void conn_flush_workqueue(struct drbd_tconn *tconn); +extern int drbd_connected(struct drbd_conf *mdev); +static inline void drbd_flush_workqueue(struct drbd_conf *mdev) +{ + conn_flush_workqueue(mdev->tconn); +} -/* yes, there is kernel_setsockopt, but only since 2.6.18. we don't need to - * mess with get_fs/set_fs, we know we are KERNEL_DS always. */ +/* Yes, there is kernel_setsockopt, but only since 2.6.18. + * So we have our own copy of it here. */ static inline int drbd_setsockopt(struct socket *sock, int level, int optname, - char __user *optval, int optlen) + char *optval, int optlen) { + mm_segment_t oldfs = get_fs(); + char __user *uoptval; int err; + + uoptval = (char __user __force *)optval; + + set_fs(KERNEL_DS); if (level == SOL_SOCKET) - err = sock_setsockopt(sock, level, optname, optval, optlen); + err = sock_setsockopt(sock, level, optname, uoptval, optlen); else - err = sock->ops->setsockopt(sock, level, optname, optval, + err = sock->ops->setsockopt(sock, level, optname, uoptval, optlen); + set_fs(oldfs); return err; } static inline void drbd_tcp_cork(struct socket *sock) { - int __user val = 1; + int val = 1; (void) drbd_setsockopt(sock, SOL_TCP, TCP_CORK, - (char __user *)&val, sizeof(val)); + (char*)&val, sizeof(val)); } static inline void drbd_tcp_uncork(struct socket *sock) { - int __user val = 0; + int val = 0; (void) drbd_setsockopt(sock, SOL_TCP, TCP_CORK, - (char __user *)&val, sizeof(val)); + (char*)&val, sizeof(val)); } static inline void drbd_tcp_nodelay(struct socket *sock) { - int __user val = 1; + int val = 1; (void) drbd_setsockopt(sock, SOL_TCP, TCP_NODELAY, - (char __user *)&val, sizeof(val)); + (char*)&val, sizeof(val)); } static inline void drbd_tcp_quickack(struct socket *sock) { - int __user val = 2; + int val = 2; (void) drbd_setsockopt(sock, SOL_TCP, TCP_QUICKACK, - (char __user *)&val, sizeof(val)); + (char*)&val, sizeof(val)); } -void drbd_bump_write_ordering(struct drbd_conf *mdev, enum write_ordering_e wo); +void drbd_bump_write_ordering(struct drbd_tconn *tconn, enum write_ordering_e wo); /* drbd_proc.c */ extern struct proc_dir_entry *drbd_proc; @@ -1693,8 +1576,8 @@ extern const char *drbd_conn_str(enum drbd_conns s); extern const char *drbd_role_str(enum drbd_role s); /* drbd_actlog.c */ -extern void drbd_al_begin_io(struct drbd_conf *mdev, sector_t sector); -extern void drbd_al_complete_io(struct drbd_conf *mdev, sector_t sector); +extern void drbd_al_begin_io(struct drbd_conf *mdev, struct drbd_interval *i); +extern void drbd_al_complete_io(struct drbd_conf *mdev, struct drbd_interval *i); extern void drbd_rs_complete_io(struct drbd_conf *mdev, sector_t sector); extern int drbd_rs_begin_io(struct drbd_conf *mdev, sector_t sector); extern int drbd_try_rs_begin_io(struct drbd_conf *mdev, sector_t sector); @@ -1702,7 +1585,6 @@ extern void drbd_rs_cancel_all(struct drbd_conf *mdev); extern int drbd_rs_del_all(struct drbd_conf *mdev); extern void drbd_rs_failed_io(struct drbd_conf *mdev, sector_t sector, int size); -extern int drbd_al_read_log(struct drbd_conf *mdev, struct drbd_backing_dev *); extern void drbd_advance_rs_marks(struct drbd_conf *mdev, unsigned long still_to_go); extern void __drbd_set_in_sync(struct drbd_conf *mdev, sector_t sector, int size, const char *file, const unsigned int line); @@ -1712,73 +1594,24 @@ extern int __drbd_set_out_of_sync(struct drbd_conf *mdev, sector_t sector, int size, const char *file, const unsigned int line); #define drbd_set_out_of_sync(mdev, sector, size) \ __drbd_set_out_of_sync(mdev, sector, size, __FILE__, __LINE__) -extern void drbd_al_apply_to_bm(struct drbd_conf *mdev); extern void drbd_al_shrink(struct drbd_conf *mdev); - /* drbd_nl.c */ - -void drbd_nl_cleanup(void); -int __init drbd_nl_init(void); -void drbd_bcast_state(struct drbd_conf *mdev, union drbd_state); -void drbd_bcast_sync_progress(struct drbd_conf *mdev); -void drbd_bcast_ee(struct drbd_conf *mdev, - const char *reason, const int dgs, - const char* seen_hash, const char* calc_hash, - const struct drbd_epoch_entry* e); - - -/** - * DOC: DRBD State macros - * - * These macros are used to express state changes in easily readable form. - * - * The NS macros expand to a mask and a value, that can be bit ored onto the - * current state as soon as the spinlock (req_lock) was taken. - * - * The _NS macros are used for state functions that get called with the - * spinlock. These macros expand directly to the new state value. - * - * Besides the basic forms NS() and _NS() additional _?NS[23] are defined - * to express state changes that affect more than one aspect of the state. - * - * E.g. NS2(conn, C_CONNECTED, peer, R_SECONDARY) - * Means that the network connection was established and that the peer - * is in secondary role. - */ -#define role_MASK R_MASK -#define peer_MASK R_MASK -#define disk_MASK D_MASK -#define pdsk_MASK D_MASK -#define conn_MASK C_MASK -#define susp_MASK 1 -#define user_isp_MASK 1 -#define aftr_isp_MASK 1 -#define susp_nod_MASK 1 -#define susp_fen_MASK 1 - -#define NS(T, S) \ - ({ union drbd_state mask; mask.i = 0; mask.T = T##_MASK; mask; }), \ - ({ union drbd_state val; val.i = 0; val.T = (S); val; }) -#define NS2(T1, S1, T2, S2) \ - ({ union drbd_state mask; mask.i = 0; mask.T1 = T1##_MASK; \ - mask.T2 = T2##_MASK; mask; }), \ - ({ union drbd_state val; val.i = 0; val.T1 = (S1); \ - val.T2 = (S2); val; }) -#define NS3(T1, S1, T2, S2, T3, S3) \ - ({ union drbd_state mask; mask.i = 0; mask.T1 = T1##_MASK; \ - mask.T2 = T2##_MASK; mask.T3 = T3##_MASK; mask; }), \ - ({ union drbd_state val; val.i = 0; val.T1 = (S1); \ - val.T2 = (S2); val.T3 = (S3); val; }) - -#define _NS(D, T, S) \ - D, ({ union drbd_state __ns; __ns.i = D->state.i; __ns.T = (S); __ns; }) -#define _NS2(D, T1, S1, T2, S2) \ - D, ({ union drbd_state __ns; __ns.i = D->state.i; __ns.T1 = (S1); \ - __ns.T2 = (S2); __ns; }) -#define _NS3(D, T1, S1, T2, S2, T3, S3) \ - D, ({ union drbd_state __ns; __ns.i = D->state.i; __ns.T1 = (S1); \ - __ns.T2 = (S2); __ns.T3 = (S3); __ns; }) +/* state info broadcast */ +struct sib_info { + enum drbd_state_info_bcast_reason sib_reason; + union { + struct { + char *helper_name; + unsigned helper_exit_code; + }; + struct { + union drbd_state os; + union drbd_state ns; + }; + }; +}; +void drbd_bcast_event(struct drbd_conf *mdev, const struct sib_info *sib); /* * inline helper functions @@ -1795,9 +1628,10 @@ static inline struct page *page_chain_next(struct page *page) #define page_chain_for_each_safe(page, n) \ for (; page && ({ n = page_chain_next(page); 1; }); page = n) -static inline int drbd_ee_has_active_page(struct drbd_epoch_entry *e) + +static inline int drbd_peer_req_has_active_page(struct drbd_peer_request *peer_req) { - struct page *page = e->pages; + struct page *page = peer_req->pages; page_chain_for_each(page) { if (page_count(page) > 1) return 1; @@ -1805,18 +1639,6 @@ static inline int drbd_ee_has_active_page(struct drbd_epoch_entry *e) return 0; } -static inline void drbd_state_lock(struct drbd_conf *mdev) -{ - wait_event(mdev->misc_wait, - !test_and_set_bit(CLUSTER_ST_CHANGE, &mdev->flags)); -} - -static inline void drbd_state_unlock(struct drbd_conf *mdev) -{ - clear_bit(CLUSTER_ST_CHANGE, &mdev->flags); - wake_up(&mdev->misc_wait); -} - static inline enum drbd_state_rv _drbd_set_state(struct drbd_conf *mdev, union drbd_state ns, enum chg_state_flags flags, struct completion *done) @@ -1830,48 +1652,71 @@ _drbd_set_state(struct drbd_conf *mdev, union drbd_state ns, return rv; } -/** - * drbd_request_state() - Reqest a state change - * @mdev: DRBD device. - * @mask: mask of state bits to change. - * @val: value of new state bits. - * - * This is the most graceful way of requesting a state change. It is verbose - * quite verbose in case the state change is not possible, and all those - * state changes are globally serialized. - */ -static inline int drbd_request_state(struct drbd_conf *mdev, - union drbd_state mask, - union drbd_state val) +static inline union drbd_state drbd_read_state(struct drbd_conf *mdev) { - return _drbd_request_state(mdev, mask, val, CS_VERBOSE + CS_ORDERED); + union drbd_state rv; + + rv.i = mdev->state.i; + rv.susp = mdev->tconn->susp; + rv.susp_nod = mdev->tconn->susp_nod; + rv.susp_fen = mdev->tconn->susp_fen; + + return rv; } enum drbd_force_detach_flags { - DRBD_IO_ERROR, + DRBD_READ_ERROR, + DRBD_WRITE_ERROR, DRBD_META_IO_ERROR, DRBD_FORCE_DETACH, }; #define __drbd_chk_io_error(m,f) __drbd_chk_io_error_(m,f, __func__) static inline void __drbd_chk_io_error_(struct drbd_conf *mdev, - enum drbd_force_detach_flags forcedetach, + enum drbd_force_detach_flags df, const char *where) { - switch (mdev->ldev->dc.on_io_error) { - case EP_PASS_ON: - if (forcedetach == DRBD_IO_ERROR) { + enum drbd_io_error_p ep; + + rcu_read_lock(); + ep = rcu_dereference(mdev->ldev->disk_conf)->on_io_error; + rcu_read_unlock(); + switch (ep) { + case EP_PASS_ON: /* FIXME would this be better named "Ignore"? */ + if (df == DRBD_READ_ERROR || df == DRBD_WRITE_ERROR) { if (__ratelimit(&drbd_ratelimit_state)) dev_err(DEV, "Local IO failed in %s.\n", where); if (mdev->state.disk > D_INCONSISTENT) _drbd_set_state(_NS(mdev, disk, D_INCONSISTENT), CS_HARD, NULL); break; } - /* NOTE fall through to detach case if forcedetach set */ + /* NOTE fall through for DRBD_META_IO_ERROR or DRBD_FORCE_DETACH */ case EP_DETACH: case EP_CALL_HELPER: + /* Remember whether we saw a READ or WRITE error. + * + * Recovery of the affected area for WRITE failure is covered + * by the activity log. + * READ errors may fall outside that area though. Certain READ + * errors can be "healed" by writing good data to the affected + * blocks, which triggers block re-allocation in lower layers. + * + * If we can not write the bitmap after a READ error, + * we may need to trigger a full sync (see w_go_diskless()). + * + * Force-detach is not really an IO error, but rather a + * desperate measure to try to deal with a completely + * unresponsive lower level IO stack. + * Still it should be treated as a WRITE error. + * + * Meta IO error is always WRITE error: + * we read meta data only once during attach, + * which will fail in case of errors. + */ set_bit(WAS_IO_ERROR, &mdev->flags); - if (forcedetach == DRBD_FORCE_DETACH) + if (df == DRBD_READ_ERROR) + set_bit(WAS_READ_ERROR, &mdev->flags); + if (df == DRBD_FORCE_DETACH) set_bit(FORCE_DETACH, &mdev->flags); if (mdev->state.disk > D_FAILED) { _drbd_set_state(_NS(mdev, disk, D_FAILED), CS_HARD, NULL); @@ -1896,9 +1741,9 @@ static inline void drbd_chk_io_error_(struct drbd_conf *mdev, { if (error) { unsigned long flags; - spin_lock_irqsave(&mdev->req_lock, flags); + spin_lock_irqsave(&mdev->tconn->req_lock, flags); __drbd_chk_io_error_(mdev, forcedetach, where); - spin_unlock_irqrestore(&mdev->req_lock, flags); + spin_unlock_irqrestore(&mdev->tconn->req_lock, flags); } } @@ -1910,9 +1755,9 @@ static inline void drbd_chk_io_error_(struct drbd_conf *mdev, * BTW, for internal meta data, this happens to be the maximum capacity * we could agree upon with our peer node. */ -static inline sector_t drbd_md_first_sector(struct drbd_backing_dev *bdev) +static inline sector_t _drbd_md_first_sector(int meta_dev_idx, struct drbd_backing_dev *bdev) { - switch (bdev->dc.meta_dev_idx) { + switch (meta_dev_idx) { case DRBD_MD_INDEX_INTERNAL: case DRBD_MD_INDEX_FLEX_INT: return bdev->md.md_offset + bdev->md.bm_offset; @@ -1922,13 +1767,30 @@ static inline sector_t drbd_md_first_sector(struct drbd_backing_dev *bdev) } } +static inline sector_t drbd_md_first_sector(struct drbd_backing_dev *bdev) +{ + int meta_dev_idx; + + rcu_read_lock(); + meta_dev_idx = rcu_dereference(bdev->disk_conf)->meta_dev_idx; + rcu_read_unlock(); + + return _drbd_md_first_sector(meta_dev_idx, bdev); +} + /** * drbd_md_last_sector() - Return the last sector number of the meta data area * @bdev: Meta data block device. */ static inline sector_t drbd_md_last_sector(struct drbd_backing_dev *bdev) { - switch (bdev->dc.meta_dev_idx) { + int meta_dev_idx; + + rcu_read_lock(); + meta_dev_idx = rcu_dereference(bdev->disk_conf)->meta_dev_idx; + rcu_read_unlock(); + + switch (meta_dev_idx) { case DRBD_MD_INDEX_INTERNAL: case DRBD_MD_INDEX_FLEX_INT: return bdev->md.md_offset + MD_AL_OFFSET - 1; @@ -1956,12 +1818,18 @@ static inline sector_t drbd_get_capacity(struct block_device *bdev) static inline sector_t drbd_get_max_capacity(struct drbd_backing_dev *bdev) { sector_t s; - switch (bdev->dc.meta_dev_idx) { + int meta_dev_idx; + + rcu_read_lock(); + meta_dev_idx = rcu_dereference(bdev->disk_conf)->meta_dev_idx; + rcu_read_unlock(); + + switch (meta_dev_idx) { case DRBD_MD_INDEX_INTERNAL: case DRBD_MD_INDEX_FLEX_INT: s = drbd_get_capacity(bdev->backing_bdev) ? min_t(sector_t, DRBD_MAX_SECTORS_FLEX, - drbd_md_first_sector(bdev)) + _drbd_md_first_sector(meta_dev_idx, bdev)) : 0; break; case DRBD_MD_INDEX_FLEX_EXT: @@ -1987,9 +1855,15 @@ static inline sector_t drbd_get_max_capacity(struct drbd_backing_dev *bdev) static inline sector_t drbd_md_ss__(struct drbd_conf *mdev, struct drbd_backing_dev *bdev) { - switch (bdev->dc.meta_dev_idx) { + int meta_dev_idx; + + rcu_read_lock(); + meta_dev_idx = rcu_dereference(bdev->disk_conf)->meta_dev_idx; + rcu_read_unlock(); + + switch (meta_dev_idx) { default: /* external, some index */ - return MD_RESERVED_SECT * bdev->dc.meta_dev_idx; + return MD_RESERVED_SECT * meta_dev_idx; case DRBD_MD_INDEX_INTERNAL: /* with drbd08, internal meta data is always "flexible" */ case DRBD_MD_INDEX_FLEX_INT: @@ -2015,9 +1889,8 @@ drbd_queue_work_front(struct drbd_work_queue *q, struct drbd_work *w) unsigned long flags; spin_lock_irqsave(&q->q_lock, flags); list_add(&w->list, &q->q); - up(&q->s); /* within the spinlock, - see comment near end of drbd_worker() */ spin_unlock_irqrestore(&q->q_lock, flags); + wake_up(&q->q_wait); } static inline void @@ -2026,41 +1899,35 @@ drbd_queue_work(struct drbd_work_queue *q, struct drbd_work *w) unsigned long flags; spin_lock_irqsave(&q->q_lock, flags); list_add_tail(&w->list, &q->q); - up(&q->s); /* within the spinlock, - see comment near end of drbd_worker() */ spin_unlock_irqrestore(&q->q_lock, flags); + wake_up(&q->q_wait); } -static inline void wake_asender(struct drbd_conf *mdev) -{ - if (test_bit(SIGNAL_ASENDER, &mdev->flags)) - force_sig(DRBD_SIG, mdev->asender.task); -} - -static inline void request_ping(struct drbd_conf *mdev) +static inline void wake_asender(struct drbd_tconn *tconn) { - set_bit(SEND_PING, &mdev->flags); - wake_asender(mdev); + if (test_bit(SIGNAL_ASENDER, &tconn->flags)) + force_sig(DRBD_SIG, tconn->asender.task); } -static inline int drbd_send_short_cmd(struct drbd_conf *mdev, - enum drbd_packets cmd) +static inline void request_ping(struct drbd_tconn *tconn) { - struct p_header80 h; - return drbd_send_cmd(mdev, USE_DATA_SOCKET, cmd, &h, sizeof(h)); + set_bit(SEND_PING, &tconn->flags); + wake_asender(tconn); } -static inline int drbd_send_ping(struct drbd_conf *mdev) -{ - struct p_header80 h; - return drbd_send_cmd(mdev, USE_META_SOCKET, P_PING, &h, sizeof(h)); -} +extern void *conn_prepare_command(struct drbd_tconn *, struct drbd_socket *); +extern void *drbd_prepare_command(struct drbd_conf *, struct drbd_socket *); +extern int conn_send_command(struct drbd_tconn *, struct drbd_socket *, + enum drbd_packet, unsigned int, void *, + unsigned int); +extern int drbd_send_command(struct drbd_conf *, struct drbd_socket *, + enum drbd_packet, unsigned int, void *, + unsigned int); -static inline int drbd_send_ping_ack(struct drbd_conf *mdev) -{ - struct p_header80 h; - return drbd_send_cmd(mdev, USE_META_SOCKET, P_PING_ACK, &h, sizeof(h)); -} +extern int drbd_send_ping(struct drbd_tconn *tconn); +extern int drbd_send_ping_ack(struct drbd_tconn *tconn); +extern int drbd_send_state_req(struct drbd_conf *, union drbd_state, union drbd_state); +extern int conn_send_state_req(struct drbd_tconn *, union drbd_state, union drbd_state); static inline void drbd_thread_stop(struct drbd_thread *thi) { @@ -2082,21 +1949,21 @@ static inline void drbd_thread_restart_nowait(struct drbd_thread *thi) * or implicit barrier packets as necessary. * increased: * w_send_barrier - * _req_mod(req, queue_for_net_write or queue_for_net_read); + * _req_mod(req, QUEUE_FOR_NET_WRITE or QUEUE_FOR_NET_READ); * it is much easier and equally valid to count what we queue for the * worker, even before it actually was queued or send. * (drbd_make_request_common; recovery path on read io-error) * decreased: * got_BarrierAck (respective tl_clear, tl_clear_barrier) - * _req_mod(req, data_received) + * _req_mod(req, DATA_RECEIVED) * [from receive_DataReply] - * _req_mod(req, write_acked_by_peer or recv_acked_by_peer or neg_acked) + * _req_mod(req, WRITE_ACKED_BY_PEER or RECV_ACKED_BY_PEER or NEG_ACKED) * [from got_BlockAck (P_WRITE_ACK, P_RECV_ACK)] * for some reason it is NOT decreased in got_NegAck, * but in the resulting cleanup code from report_params. * we should try to remember the reason for that... - * _req_mod(req, send_failed or send_canceled) - * _req_mod(req, connection_lost_while_pending) + * _req_mod(req, SEND_FAILED or SEND_CANCELED) + * _req_mod(req, CONNECTION_LOST_WHILE_PENDING) * [from tl_clear_barrier] */ static inline void inc_ap_pending(struct drbd_conf *mdev) @@ -2104,17 +1971,19 @@ static inline void inc_ap_pending(struct drbd_conf *mdev) atomic_inc(&mdev->ap_pending_cnt); } -#define ERR_IF_CNT_IS_NEGATIVE(which) \ - if (atomic_read(&mdev->which) < 0) \ +#define ERR_IF_CNT_IS_NEGATIVE(which, func, line) \ + if (atomic_read(&mdev->which) < 0) \ dev_err(DEV, "in %s:%d: " #which " = %d < 0 !\n", \ - __func__ , __LINE__ , \ - atomic_read(&mdev->which)) + func, line, \ + atomic_read(&mdev->which)) -#define dec_ap_pending(mdev) do { \ - typecheck(struct drbd_conf *, mdev); \ - if (atomic_dec_and_test(&mdev->ap_pending_cnt)) \ - wake_up(&mdev->misc_wait); \ - ERR_IF_CNT_IS_NEGATIVE(ap_pending_cnt); } while (0) +#define dec_ap_pending(mdev) _dec_ap_pending(mdev, __FUNCTION__, __LINE__) +static inline void _dec_ap_pending(struct drbd_conf *mdev, const char *func, int line) +{ + if (atomic_dec_and_test(&mdev->ap_pending_cnt)) + wake_up(&mdev->misc_wait); + ERR_IF_CNT_IS_NEGATIVE(ap_pending_cnt, func, line); +} /* counts how many resync-related answers we still expect from the peer * increase decrease @@ -2127,10 +1996,12 @@ static inline void inc_rs_pending(struct drbd_conf *mdev) atomic_inc(&mdev->rs_pending_cnt); } -#define dec_rs_pending(mdev) do { \ - typecheck(struct drbd_conf *, mdev); \ - atomic_dec(&mdev->rs_pending_cnt); \ - ERR_IF_CNT_IS_NEGATIVE(rs_pending_cnt); } while (0) +#define dec_rs_pending(mdev) _dec_rs_pending(mdev, __FUNCTION__, __LINE__) +static inline void _dec_rs_pending(struct drbd_conf *mdev, const char *func, int line) +{ + atomic_dec(&mdev->rs_pending_cnt); + ERR_IF_CNT_IS_NEGATIVE(rs_pending_cnt, func, line); +} /* counts how many answers we still need to send to the peer. * increased on @@ -2146,38 +2017,18 @@ static inline void inc_unacked(struct drbd_conf *mdev) atomic_inc(&mdev->unacked_cnt); } -#define dec_unacked(mdev) do { \ - typecheck(struct drbd_conf *, mdev); \ - atomic_dec(&mdev->unacked_cnt); \ - ERR_IF_CNT_IS_NEGATIVE(unacked_cnt); } while (0) - -#define sub_unacked(mdev, n) do { \ - typecheck(struct drbd_conf *, mdev); \ - atomic_sub(n, &mdev->unacked_cnt); \ - ERR_IF_CNT_IS_NEGATIVE(unacked_cnt); } while (0) - - -static inline void put_net_conf(struct drbd_conf *mdev) +#define dec_unacked(mdev) _dec_unacked(mdev, __FUNCTION__, __LINE__) +static inline void _dec_unacked(struct drbd_conf *mdev, const char *func, int line) { - if (atomic_dec_and_test(&mdev->net_cnt)) - wake_up(&mdev->net_cnt_wait); + atomic_dec(&mdev->unacked_cnt); + ERR_IF_CNT_IS_NEGATIVE(unacked_cnt, func, line); } -/** - * get_net_conf() - Increase ref count on mdev->net_conf; Returns 0 if nothing there - * @mdev: DRBD device. - * - * You have to call put_net_conf() when finished working with mdev->net_conf. - */ -static inline int get_net_conf(struct drbd_conf *mdev) +#define sub_unacked(mdev, n) _sub_unacked(mdev, n, __FUNCTION__, __LINE__) +static inline void _sub_unacked(struct drbd_conf *mdev, int n, const char *func, int line) { - int have_net_conf; - - atomic_inc(&mdev->net_cnt); - have_net_conf = mdev->state.conn >= C_UNCONNECTED; - if (!have_net_conf) - put_net_conf(mdev); - return have_net_conf; + atomic_sub(n, &mdev->unacked_cnt); + ERR_IF_CNT_IS_NEGATIVE(unacked_cnt, func, line); } /** @@ -2281,17 +2132,20 @@ static inline void drbd_get_syncer_progress(struct drbd_conf *mdev, * maybe re-implement using semaphores? */ static inline int drbd_get_max_buffers(struct drbd_conf *mdev) { - int mxb = 1000000; /* arbitrary limit on open requests */ - if (get_net_conf(mdev)) { - mxb = mdev->net_conf->max_buffers; - put_net_conf(mdev); - } + struct net_conf *nc; + int mxb; + + rcu_read_lock(); + nc = rcu_dereference(mdev->tconn->net_conf); + mxb = nc ? nc->max_buffers : 1000000; /* arbitrary limit on open requests */ + rcu_read_unlock(); + return mxb; } static inline int drbd_state_is_stable(struct drbd_conf *mdev) { - union drbd_state s = mdev->state; + union drbd_dev_state s = mdev->state; /* DO NOT add a default clause, we want the compiler to warn us * for any newly introduced state we may have forgotten to add here */ @@ -2325,7 +2179,7 @@ static inline int drbd_state_is_stable(struct drbd_conf *mdev) /* Allow IO in BM exchange states with new protocols */ case C_WF_BITMAP_S: - if (mdev->agreed_pro_version < 96) + if (mdev->tconn->agreed_pro_version < 96) return 0; break; @@ -2347,7 +2201,7 @@ static inline int drbd_state_is_stable(struct drbd_conf *mdev) /* disk state is stable as well. */ break; - /* no new io accepted during tansitional states */ + /* no new io accepted during transitional states */ case D_ATTACHING: case D_NEGOTIATING: case D_UNKNOWN: @@ -2359,16 +2213,18 @@ static inline int drbd_state_is_stable(struct drbd_conf *mdev) return 1; } -static inline int is_susp(union drbd_state s) +static inline int drbd_suspended(struct drbd_conf *mdev) { - return s.susp || s.susp_nod || s.susp_fen; + struct drbd_tconn *tconn = mdev->tconn; + + return tconn->susp || tconn->susp_fen || tconn->susp_nod; } static inline bool may_inc_ap_bio(struct drbd_conf *mdev) { int mxb = drbd_get_max_buffers(mdev); - if (is_susp(mdev->state)) + if (drbd_suspended(mdev)) return false; if (test_bit(SUSPEND_IO, &mdev->flags)) return false; @@ -2390,30 +2246,30 @@ static inline bool may_inc_ap_bio(struct drbd_conf *mdev) return true; } -static inline bool inc_ap_bio_cond(struct drbd_conf *mdev, int count) +static inline bool inc_ap_bio_cond(struct drbd_conf *mdev) { bool rv = false; - spin_lock_irq(&mdev->req_lock); + spin_lock_irq(&mdev->tconn->req_lock); rv = may_inc_ap_bio(mdev); if (rv) - atomic_add(count, &mdev->ap_bio_cnt); - spin_unlock_irq(&mdev->req_lock); + atomic_inc(&mdev->ap_bio_cnt); + spin_unlock_irq(&mdev->tconn->req_lock); return rv; } -static inline void inc_ap_bio(struct drbd_conf *mdev, int count) +static inline void inc_ap_bio(struct drbd_conf *mdev) { /* we wait here * as long as the device is suspended * until the bitmap is no longer on the fly during connection - * handshake as long as we would exeed the max_buffer limit. + * handshake as long as we would exceed the max_buffer limit. * * to avoid races with the reconnect code, * we need to atomic_inc within the spinlock. */ - wait_event(mdev->misc_wait, inc_ap_bio_cond(mdev, count)); + wait_event(mdev->misc_wait, inc_ap_bio_cond(mdev)); } static inline void dec_ap_bio(struct drbd_conf *mdev) @@ -2425,7 +2281,7 @@ static inline void dec_ap_bio(struct drbd_conf *mdev) if (ap_bio == 0 && test_bit(BITMAP_IO, &mdev->flags)) { if (!test_and_set_bit(BITMAP_IO_QUEUED, &mdev->flags)) - drbd_queue_work(&mdev->data.work, &mdev->bm_io_work.w); + drbd_queue_work(&mdev->tconn->sender_work, &mdev->bm_io_work.w); } /* this currently does wake_up for every dec_ap_bio! @@ -2435,6 +2291,12 @@ static inline void dec_ap_bio(struct drbd_conf *mdev) wake_up(&mdev->misc_wait); } +static inline bool verify_can_do_stop_sector(struct drbd_conf *mdev) +{ + return mdev->tconn->agreed_pro_version >= 97 && + mdev->tconn->agreed_pro_version != 100; +} + static inline int drbd_set_ed_uuid(struct drbd_conf *mdev, u64 val) { int changed = mdev->ed_uuid != val; @@ -2442,40 +2304,6 @@ static inline int drbd_set_ed_uuid(struct drbd_conf *mdev, u64 val) return changed; } -static inline int seq_cmp(u32 a, u32 b) -{ - /* we assume wrap around at 32bit. - * for wrap around at 24bit (old atomic_t), - * we'd have to - * a <<= 8; b <<= 8; - */ - return (s32)(a) - (s32)(b); -} -#define seq_lt(a, b) (seq_cmp((a), (b)) < 0) -#define seq_gt(a, b) (seq_cmp((a), (b)) > 0) -#define seq_ge(a, b) (seq_cmp((a), (b)) >= 0) -#define seq_le(a, b) (seq_cmp((a), (b)) <= 0) -/* CAUTION: please no side effects in arguments! */ -#define seq_max(a, b) ((u32)(seq_gt((a), (b)) ? (a) : (b))) - -static inline void update_peer_seq(struct drbd_conf *mdev, unsigned int new_seq) -{ - unsigned int m; - spin_lock(&mdev->peer_seq_lock); - m = seq_max(mdev->peer_seq, new_seq); - mdev->peer_seq = m; - spin_unlock(&mdev->peer_seq_lock); - if (m == new_seq) - wake_up(&mdev->seq_wait); -} - -static inline void drbd_update_congested(struct drbd_conf *mdev) -{ - struct sock *sk = mdev->data.socket->sk; - if (sk->sk_wmem_queued > sk->sk_sndbuf * 4 / 5) - set_bit(NET_CONGESTED, &mdev->flags); -} - static inline int drbd_queue_order_type(struct drbd_conf *mdev) { /* sorry, we currently have no working implementation @@ -2490,10 +2318,15 @@ static inline void drbd_md_flush(struct drbd_conf *mdev) { int r; + if (mdev->ldev == NULL) { + dev_warn(DEV, "mdev->ldev == NULL in drbd_md_flush\n"); + return; + } + if (test_bit(MD_NO_FUA, &mdev->flags)) return; - r = blkdev_issue_flush(mdev->ldev->md_bdev, GFP_KERNEL, NULL); + r = blkdev_issue_flush(mdev->ldev->md_bdev, GFP_NOIO, NULL); if (r) { set_bit(MD_NO_FUA, &mdev->flags); dev_err(DEV, "meta data flush failed with status %d, disabling md-flushes\n", r); diff --git a/drivers/block/drbd/drbd_interval.c b/drivers/block/drbd/drbd_interval.c new file mode 100644 index 000000000000..89c497c630b4 --- /dev/null +++ b/drivers/block/drbd/drbd_interval.c @@ -0,0 +1,207 @@ +#include <asm/bug.h> +#include <linux/rbtree_augmented.h> +#include "drbd_interval.h" + +/** + * interval_end - return end of @node + */ +static inline +sector_t interval_end(struct rb_node *node) +{ + struct drbd_interval *this = rb_entry(node, struct drbd_interval, rb); + return this->end; +} + +/** + * compute_subtree_last - compute end of @node + * + * The end of an interval is the highest (start + (size >> 9)) value of this + * node and of its children. Called for @node and its parents whenever the end + * may have changed. + */ +static inline sector_t +compute_subtree_last(struct drbd_interval *node) +{ + sector_t max = node->sector + (node->size >> 9); + + if (node->rb.rb_left) { + sector_t left = interval_end(node->rb.rb_left); + if (left > max) + max = left; + } + if (node->rb.rb_right) { + sector_t right = interval_end(node->rb.rb_right); + if (right > max) + max = right; + } + return max; +} + +static void augment_propagate(struct rb_node *rb, struct rb_node *stop) +{ + while (rb != stop) { + struct drbd_interval *node = rb_entry(rb, struct drbd_interval, rb); + sector_t subtree_last = compute_subtree_last(node); + if (node->end == subtree_last) + break; + node->end = subtree_last; + rb = rb_parent(&node->rb); + } +} + +static void augment_copy(struct rb_node *rb_old, struct rb_node *rb_new) +{ + struct drbd_interval *old = rb_entry(rb_old, struct drbd_interval, rb); + struct drbd_interval *new = rb_entry(rb_new, struct drbd_interval, rb); + + new->end = old->end; +} + +static void augment_rotate(struct rb_node *rb_old, struct rb_node *rb_new) +{ + struct drbd_interval *old = rb_entry(rb_old, struct drbd_interval, rb); + struct drbd_interval *new = rb_entry(rb_new, struct drbd_interval, rb); + + new->end = old->end; + old->end = compute_subtree_last(old); +} + +static const struct rb_augment_callbacks augment_callbacks = { + augment_propagate, + augment_copy, + augment_rotate, +}; + +/** + * drbd_insert_interval - insert a new interval into a tree + */ +bool +drbd_insert_interval(struct rb_root *root, struct drbd_interval *this) +{ + struct rb_node **new = &root->rb_node, *parent = NULL; + + BUG_ON(!IS_ALIGNED(this->size, 512)); + + while (*new) { + struct drbd_interval *here = + rb_entry(*new, struct drbd_interval, rb); + + parent = *new; + if (this->sector < here->sector) + new = &(*new)->rb_left; + else if (this->sector > here->sector) + new = &(*new)->rb_right; + else if (this < here) + new = &(*new)->rb_left; + else if (this > here) + new = &(*new)->rb_right; + else + return false; + } + + rb_link_node(&this->rb, parent, new); + rb_insert_augmented(&this->rb, root, &augment_callbacks); + return true; +} + +/** + * drbd_contains_interval - check if a tree contains a given interval + * @sector: start sector of @interval + * @interval: may not be a valid pointer + * + * Returns if the tree contains the node @interval with start sector @start. + * Does not dereference @interval until @interval is known to be a valid object + * in @tree. Returns %false if @interval is in the tree but with a different + * sector number. + */ +bool +drbd_contains_interval(struct rb_root *root, sector_t sector, + struct drbd_interval *interval) +{ + struct rb_node *node = root->rb_node; + + while (node) { + struct drbd_interval *here = + rb_entry(node, struct drbd_interval, rb); + + if (sector < here->sector) + node = node->rb_left; + else if (sector > here->sector) + node = node->rb_right; + else if (interval < here) + node = node->rb_left; + else if (interval > here) + node = node->rb_right; + else + return true; + } + return false; +} + +/** + * drbd_remove_interval - remove an interval from a tree + */ +void +drbd_remove_interval(struct rb_root *root, struct drbd_interval *this) +{ + rb_erase_augmented(&this->rb, root, &augment_callbacks); +} + +/** + * drbd_find_overlap - search for an interval overlapping with [sector, sector + size) + * @sector: start sector + * @size: size, aligned to 512 bytes + * + * Returns an interval overlapping with [sector, sector + size), or NULL if + * there is none. When there is more than one overlapping interval in the + * tree, the interval with the lowest start sector is returned, and all other + * overlapping intervals will be on the right side of the tree, reachable with + * rb_next(). + */ +struct drbd_interval * +drbd_find_overlap(struct rb_root *root, sector_t sector, unsigned int size) +{ + struct rb_node *node = root->rb_node; + struct drbd_interval *overlap = NULL; + sector_t end = sector + (size >> 9); + + BUG_ON(!IS_ALIGNED(size, 512)); + + while (node) { + struct drbd_interval *here = + rb_entry(node, struct drbd_interval, rb); + + if (node->rb_left && + sector < interval_end(node->rb_left)) { + /* Overlap if any must be on left side */ + node = node->rb_left; + } else if (here->sector < end && + sector < here->sector + (here->size >> 9)) { + overlap = here; + break; + } else if (sector >= here->sector) { + /* Overlap if any must be on right side */ + node = node->rb_right; + } else + break; + } + return overlap; +} + +struct drbd_interval * +drbd_next_overlap(struct drbd_interval *i, sector_t sector, unsigned int size) +{ + sector_t end = sector + (size >> 9); + struct rb_node *node; + + for (;;) { + node = rb_next(&i->rb); + if (!node) + return NULL; + i = rb_entry(node, struct drbd_interval, rb); + if (i->sector >= end) + return NULL; + if (sector < i->sector + (i->size >> 9)) + return i; + } +} diff --git a/drivers/block/drbd/drbd_interval.h b/drivers/block/drbd/drbd_interval.h new file mode 100644 index 000000000000..f38fcb00c10d --- /dev/null +++ b/drivers/block/drbd/drbd_interval.h @@ -0,0 +1,40 @@ +#ifndef __DRBD_INTERVAL_H +#define __DRBD_INTERVAL_H + +#include <linux/types.h> +#include <linux/rbtree.h> + +struct drbd_interval { + struct rb_node rb; + sector_t sector; /* start sector of the interval */ + unsigned int size; /* size in bytes */ + sector_t end; /* highest interval end in subtree */ + int local:1 /* local or remote request? */; + int waiting:1; +}; + +static inline void drbd_clear_interval(struct drbd_interval *i) +{ + RB_CLEAR_NODE(&i->rb); +} + +static inline bool drbd_interval_empty(struct drbd_interval *i) +{ + return RB_EMPTY_NODE(&i->rb); +} + +extern bool drbd_insert_interval(struct rb_root *, struct drbd_interval *); +extern bool drbd_contains_interval(struct rb_root *, sector_t, + struct drbd_interval *); +extern void drbd_remove_interval(struct rb_root *, struct drbd_interval *); +extern struct drbd_interval *drbd_find_overlap(struct rb_root *, sector_t, + unsigned int); +extern struct drbd_interval *drbd_next_overlap(struct drbd_interval *, sector_t, + unsigned int); + +#define drbd_for_each_overlap(i, root, sector, size) \ + for (i = drbd_find_overlap(root, sector, size); \ + i; \ + i = drbd_next_overlap(i, sector, size)) + +#endif /* __DRBD_INTERVAL_H */ diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c index f55683ad4ffa..8c13eeb83c53 100644 --- a/drivers/block/drbd/drbd_main.c +++ b/drivers/block/drbd/drbd_main.c @@ -56,14 +56,6 @@ #include "drbd_vli.h" -struct after_state_chg_work { - struct drbd_work w; - union drbd_state os; - union drbd_state ns; - enum chg_state_flags flags; - struct completion *done; -}; - static DEFINE_MUTEX(drbd_main_mutex); int drbdd_init(struct drbd_thread *); int drbd_worker(struct drbd_thread *); @@ -72,21 +64,17 @@ int drbd_asender(struct drbd_thread *); int drbd_init(void); static int drbd_open(struct block_device *bdev, fmode_t mode); static int drbd_release(struct gendisk *gd, fmode_t mode); -static int w_after_state_ch(struct drbd_conf *mdev, struct drbd_work *w, int unused); -static void after_state_ch(struct drbd_conf *mdev, union drbd_state os, - union drbd_state ns, enum chg_state_flags flags); -static int w_md_sync(struct drbd_conf *mdev, struct drbd_work *w, int unused); +static int w_md_sync(struct drbd_work *w, int unused); static void md_sync_timer_fn(unsigned long data); -static int w_bitmap_io(struct drbd_conf *mdev, struct drbd_work *w, int unused); -static int w_go_diskless(struct drbd_conf *mdev, struct drbd_work *w, int unused); -static void _tl_clear(struct drbd_conf *mdev); +static int w_bitmap_io(struct drbd_work *w, int unused); +static int w_go_diskless(struct drbd_work *w, int unused); MODULE_AUTHOR("Philipp Reisner <phil@linbit.com>, " "Lars Ellenberg <lars@linbit.com>"); MODULE_DESCRIPTION("drbd - Distributed Replicated Block Device v" REL_VERSION); MODULE_VERSION(REL_VERSION); MODULE_LICENSE("GPL"); -MODULE_PARM_DESC(minor_count, "Maximum number of drbd devices (" +MODULE_PARM_DESC(minor_count, "Approximate number of drbd devices (" __stringify(DRBD_MINOR_COUNT_MIN) "-" __stringify(DRBD_MINOR_COUNT_MAX) ")"); MODULE_ALIAS_BLOCKDEV_MAJOR(DRBD_MAJOR); @@ -98,7 +86,6 @@ MODULE_PARM_DESC(allow_oos, "DONT USE!"); module_param(minor_count, uint, 0444); module_param(disable_sendpage, bool, 0644); module_param(allow_oos, bool, 0); -module_param(cn_idx, uint, 0444); module_param(proc_details, int, 0644); #ifdef CONFIG_DRBD_FAULT_INJECTION @@ -120,7 +107,6 @@ module_param(fault_devs, int, 0644); unsigned int minor_count = DRBD_MINOR_COUNT_DEF; bool disable_sendpage; bool allow_oos; -unsigned int cn_idx = CN_IDX_DRBD; int proc_details; /* Detail level in proc drbd*/ /* Module parameter for setting the user mode helper program @@ -132,10 +118,11 @@ module_param_string(usermode_helper, usermode_helper, sizeof(usermode_helper), 0 /* in 2.6.x, our device mapping and config info contains our virtual gendisks * as member "struct gendisk *vdisk;" */ -struct drbd_conf **minor_table; +struct idr minors; +struct list_head drbd_tconns; /* list of struct drbd_tconn */ struct kmem_cache *drbd_request_cache; -struct kmem_cache *drbd_ee_cache; /* epoch entries */ +struct kmem_cache *drbd_ee_cache; /* peer requests */ struct kmem_cache *drbd_bm_ext_cache; /* bitmap extents */ struct kmem_cache *drbd_al_ext_cache; /* activity log extents */ mempool_t *drbd_request_mempool; @@ -164,10 +151,15 @@ static const struct block_device_operations drbd_ops = { struct bio *bio_alloc_drbd(gfp_t gfp_mask) { + struct bio *bio; + if (!drbd_md_io_bio_set) return bio_alloc(gfp_mask, 1); - return bio_alloc_bioset(gfp_mask, 1, drbd_md_io_bio_set); + bio = bio_alloc_bioset(gfp_mask, 1, drbd_md_io_bio_set); + if (!bio) + return NULL; + return bio; } #ifdef __CHECKER__ @@ -190,158 +182,87 @@ int _get_ldev_if_state(struct drbd_conf *mdev, enum drbd_disk_state mins) #endif /** - * DOC: The transfer log - * - * The transfer log is a single linked list of &struct drbd_tl_epoch objects. - * mdev->newest_tle points to the head, mdev->oldest_tle points to the tail - * of the list. There is always at least one &struct drbd_tl_epoch object. - * - * Each &struct drbd_tl_epoch has a circular double linked list of requests - * attached. - */ -static int tl_init(struct drbd_conf *mdev) -{ - struct drbd_tl_epoch *b; - - /* during device minor initialization, we may well use GFP_KERNEL */ - b = kmalloc(sizeof(struct drbd_tl_epoch), GFP_KERNEL); - if (!b) - return 0; - INIT_LIST_HEAD(&b->requests); - INIT_LIST_HEAD(&b->w.list); - b->next = NULL; - b->br_number = 4711; - b->n_writes = 0; - b->w.cb = NULL; /* if this is != NULL, we need to dec_ap_pending in tl_clear */ - - mdev->oldest_tle = b; - mdev->newest_tle = b; - INIT_LIST_HEAD(&mdev->out_of_sequence_requests); - INIT_LIST_HEAD(&mdev->barrier_acked_requests); - - mdev->tl_hash = NULL; - mdev->tl_hash_s = 0; - - return 1; -} - -static void tl_cleanup(struct drbd_conf *mdev) -{ - D_ASSERT(mdev->oldest_tle == mdev->newest_tle); - D_ASSERT(list_empty(&mdev->out_of_sequence_requests)); - kfree(mdev->oldest_tle); - mdev->oldest_tle = NULL; - kfree(mdev->unused_spare_tle); - mdev->unused_spare_tle = NULL; - kfree(mdev->tl_hash); - mdev->tl_hash = NULL; - mdev->tl_hash_s = 0; -} - -/** - * _tl_add_barrier() - Adds a barrier to the transfer log - * @mdev: DRBD device. - * @new: Barrier to be added before the current head of the TL. - * - * The caller must hold the req_lock. - */ -void _tl_add_barrier(struct drbd_conf *mdev, struct drbd_tl_epoch *new) -{ - struct drbd_tl_epoch *newest_before; - - INIT_LIST_HEAD(&new->requests); - INIT_LIST_HEAD(&new->w.list); - new->w.cb = NULL; /* if this is != NULL, we need to dec_ap_pending in tl_clear */ - new->next = NULL; - new->n_writes = 0; - - newest_before = mdev->newest_tle; - new->br_number = newest_before->br_number+1; - if (mdev->newest_tle != new) { - mdev->newest_tle->next = new; - mdev->newest_tle = new; - } -} - -/** - * tl_release() - Free or recycle the oldest &struct drbd_tl_epoch object of the TL - * @mdev: DRBD device. + * tl_release() - mark as BARRIER_ACKED all requests in the corresponding transfer log epoch + * @tconn: DRBD connection. * @barrier_nr: Expected identifier of the DRBD write barrier packet. * @set_size: Expected number of requests before that barrier. * * In case the passed barrier_nr or set_size does not match the oldest - * &struct drbd_tl_epoch objects this function will cause a termination - * of the connection. + * epoch of not yet barrier-acked requests, this function will cause a + * termination of the connection. */ -void tl_release(struct drbd_conf *mdev, unsigned int barrier_nr, - unsigned int set_size) +void tl_release(struct drbd_tconn *tconn, unsigned int barrier_nr, + unsigned int set_size) { - struct drbd_tl_epoch *b, *nob; /* next old barrier */ - struct list_head *le, *tle; struct drbd_request *r; - - spin_lock_irq(&mdev->req_lock); - - b = mdev->oldest_tle; + struct drbd_request *req = NULL; + int expect_epoch = 0; + int expect_size = 0; + + spin_lock_irq(&tconn->req_lock); + + /* find oldest not yet barrier-acked write request, + * count writes in its epoch. */ + list_for_each_entry(r, &tconn->transfer_log, tl_requests) { + const unsigned s = r->rq_state; + if (!req) { + if (!(s & RQ_WRITE)) + continue; + if (!(s & RQ_NET_MASK)) + continue; + if (s & RQ_NET_DONE) + continue; + req = r; + expect_epoch = req->epoch; + expect_size ++; + } else { + if (r->epoch != expect_epoch) + break; + if (!(s & RQ_WRITE)) + continue; + /* if (s & RQ_DONE): not expected */ + /* if (!(s & RQ_NET_MASK)): not expected */ + expect_size++; + } + } /* first some paranoia code */ - if (b == NULL) { - dev_err(DEV, "BAD! BarrierAck #%u received, but no epoch in tl!?\n", - barrier_nr); + if (req == NULL) { + conn_err(tconn, "BAD! BarrierAck #%u received, but no epoch in tl!?\n", + barrier_nr); goto bail; } - if (b->br_number != barrier_nr) { - dev_err(DEV, "BAD! BarrierAck #%u received, expected #%u!\n", - barrier_nr, b->br_number); + if (expect_epoch != barrier_nr) { + conn_err(tconn, "BAD! BarrierAck #%u received, expected #%u!\n", + barrier_nr, expect_epoch); goto bail; } - if (b->n_writes != set_size) { - dev_err(DEV, "BAD! BarrierAck #%u received with n_writes=%u, expected n_writes=%u!\n", - barrier_nr, set_size, b->n_writes); + + if (expect_size != set_size) { + conn_err(tconn, "BAD! BarrierAck #%u received with n_writes=%u, expected n_writes=%u!\n", + barrier_nr, set_size, expect_size); goto bail; } - /* Clean up list of requests processed during current epoch */ - list_for_each_safe(le, tle, &b->requests) { - r = list_entry(le, struct drbd_request, tl_requests); - _req_mod(r, barrier_acked); - } - /* There could be requests on the list waiting for completion - of the write to the local disk. To avoid corruptions of - slab's data structures we have to remove the lists head. - - Also there could have been a barrier ack out of sequence, overtaking - the write acks - which would be a bug and violating write ordering. - To not deadlock in case we lose connection while such requests are - still pending, we need some way to find them for the - _req_mode(connection_lost_while_pending). - - These have been list_move'd to the out_of_sequence_requests list in - _req_mod(, barrier_acked) above. - */ - list_splice_init(&b->requests, &mdev->barrier_acked_requests); - - nob = b->next; - if (test_and_clear_bit(CREATE_BARRIER, &mdev->flags)) { - _tl_add_barrier(mdev, b); - if (nob) - mdev->oldest_tle = nob; - /* if nob == NULL b was the only barrier, and becomes the new - barrier. Therefore mdev->oldest_tle points already to b */ - } else { - D_ASSERT(nob != NULL); - mdev->oldest_tle = nob; - kfree(b); + /* Clean up list of requests processed during current epoch. */ + /* this extra list walk restart is paranoia, + * to catch requests being barrier-acked "unexpectedly". + * It usually should find the same req again, or some READ preceding it. */ + list_for_each_entry(req, &tconn->transfer_log, tl_requests) + if (req->epoch == expect_epoch) + break; + list_for_each_entry_safe_from(req, r, &tconn->transfer_log, tl_requests) { + if (req->epoch != expect_epoch) + break; + _req_mod(req, BARRIER_ACKED); } - - spin_unlock_irq(&mdev->req_lock); - dec_ap_pending(mdev); + spin_unlock_irq(&tconn->req_lock); return; bail: - spin_unlock_irq(&mdev->req_lock); - drbd_force_state(mdev, NS(conn, C_PROTOCOL_ERROR)); + spin_unlock_irq(&tconn->req_lock); + conn_request_state(tconn, NS(conn, C_PROTOCOL_ERROR), CS_HARD); } @@ -350,85 +271,24 @@ bail: * @mdev: DRBD device. * @what: The action/event to perform with all request objects * - * @what might be one of connection_lost_while_pending, resend, fail_frozen_disk_io, - * restart_frozen_disk_io. + * @what might be one of CONNECTION_LOST_WHILE_PENDING, RESEND, FAIL_FROZEN_DISK_IO, + * RESTART_FROZEN_DISK_IO. */ -static void _tl_restart(struct drbd_conf *mdev, enum drbd_req_event what) -{ - struct drbd_tl_epoch *b, *tmp, **pn; - struct list_head *le, *tle, carry_reads; - struct drbd_request *req; - int rv, n_writes, n_reads; - - b = mdev->oldest_tle; - pn = &mdev->oldest_tle; - while (b) { - n_writes = 0; - n_reads = 0; - INIT_LIST_HEAD(&carry_reads); - list_for_each_safe(le, tle, &b->requests) { - req = list_entry(le, struct drbd_request, tl_requests); - rv = _req_mod(req, what); - - n_writes += (rv & MR_WRITE) >> MR_WRITE_SHIFT; - n_reads += (rv & MR_READ) >> MR_READ_SHIFT; - } - tmp = b->next; - - if (n_writes) { - if (what == resend) { - b->n_writes = n_writes; - if (b->w.cb == NULL) { - b->w.cb = w_send_barrier; - inc_ap_pending(mdev); - set_bit(CREATE_BARRIER, &mdev->flags); - } - - drbd_queue_work(&mdev->data.work, &b->w); - } - pn = &b->next; - } else { - if (n_reads) - list_add(&carry_reads, &b->requests); - /* there could still be requests on that ring list, - * in case local io is still pending */ - list_del(&b->requests); - - /* dec_ap_pending corresponding to queue_barrier. - * the newest barrier may not have been queued yet, - * in which case w.cb is still NULL. */ - if (b->w.cb != NULL) - dec_ap_pending(mdev); - - if (b == mdev->newest_tle) { - /* recycle, but reinit! */ - D_ASSERT(tmp == NULL); - INIT_LIST_HEAD(&b->requests); - list_splice(&carry_reads, &b->requests); - INIT_LIST_HEAD(&b->w.list); - b->w.cb = NULL; - b->br_number = net_random(); - b->n_writes = 0; - - *pn = b; - break; - } - *pn = tmp; - kfree(b); - } - b = tmp; - list_splice(&carry_reads, &b->requests); - } - - /* Actions operating on the disk state, also want to work on - requests that got barrier acked. */ +/* must hold resource->req_lock */ +void _tl_restart(struct drbd_tconn *tconn, enum drbd_req_event what) +{ + struct drbd_request *req, *r; - list_for_each_safe(le, tle, &mdev->barrier_acked_requests) { - req = list_entry(le, struct drbd_request, tl_requests); + list_for_each_entry_safe(req, r, &tconn->transfer_log, tl_requests) _req_mod(req, what); - } } +void tl_restart(struct drbd_tconn *tconn, enum drbd_req_event what) +{ + spin_lock_irq(&tconn->req_lock); + _tl_restart(tconn, what); + spin_unlock_irq(&tconn->req_lock); +} /** * tl_clear() - Clears all requests and &struct drbd_tl_epoch objects out of the TL @@ -438,43 +298,9 @@ static void _tl_restart(struct drbd_conf *mdev, enum drbd_req_event what) * by the requests on the transfer gets marked as our of sync. Called from the * receiver thread and the worker thread. */ -void tl_clear(struct drbd_conf *mdev) +void tl_clear(struct drbd_tconn *tconn) { - spin_lock_irq(&mdev->req_lock); - _tl_clear(mdev); - spin_unlock_irq(&mdev->req_lock); -} - -static void _tl_clear(struct drbd_conf *mdev) -{ - struct list_head *le, *tle; - struct drbd_request *r; - - _tl_restart(mdev, connection_lost_while_pending); - - /* we expect this list to be empty. */ - D_ASSERT(list_empty(&mdev->out_of_sequence_requests)); - - /* but just in case, clean it up anyways! */ - list_for_each_safe(le, tle, &mdev->out_of_sequence_requests) { - r = list_entry(le, struct drbd_request, tl_requests); - /* It would be nice to complete outside of spinlock. - * But this is easier for now. */ - _req_mod(r, connection_lost_while_pending); - } - - /* ensure bit indicating barrier is required is clear */ - clear_bit(CREATE_BARRIER, &mdev->flags); - - memset(mdev->app_reads_hash, 0, APP_R_HSIZE*sizeof(void *)); - -} - -void tl_restart(struct drbd_conf *mdev, enum drbd_req_event what) -{ - spin_lock_irq(&mdev->req_lock); - _tl_restart(mdev, what); - spin_unlock_irq(&mdev->req_lock); + tl_restart(tconn, CONNECTION_LOST_WHILE_PENDING); } /** @@ -483,1377 +309,131 @@ void tl_restart(struct drbd_conf *mdev, enum drbd_req_event what) */ void tl_abort_disk_io(struct drbd_conf *mdev) { - struct drbd_tl_epoch *b; - struct list_head *le, *tle; - struct drbd_request *req; - - spin_lock_irq(&mdev->req_lock); - b = mdev->oldest_tle; - while (b) { - list_for_each_safe(le, tle, &b->requests) { - req = list_entry(le, struct drbd_request, tl_requests); - if (!(req->rq_state & RQ_LOCAL_PENDING)) - continue; - _req_mod(req, abort_disk_io); - } - b = b->next; - } + struct drbd_tconn *tconn = mdev->tconn; + struct drbd_request *req, *r; - list_for_each_safe(le, tle, &mdev->barrier_acked_requests) { - req = list_entry(le, struct drbd_request, tl_requests); + spin_lock_irq(&tconn->req_lock); + list_for_each_entry_safe(req, r, &tconn->transfer_log, tl_requests) { if (!(req->rq_state & RQ_LOCAL_PENDING)) continue; - _req_mod(req, abort_disk_io); - } - - spin_unlock_irq(&mdev->req_lock); -} - -/** - * cl_wide_st_chg() - true if the state change is a cluster wide one - * @mdev: DRBD device. - * @os: old (current) state. - * @ns: new (wanted) state. - */ -static int cl_wide_st_chg(struct drbd_conf *mdev, - union drbd_state os, union drbd_state ns) -{ - return (os.conn >= C_CONNECTED && ns.conn >= C_CONNECTED && - ((os.role != R_PRIMARY && ns.role == R_PRIMARY) || - (os.conn != C_STARTING_SYNC_T && ns.conn == C_STARTING_SYNC_T) || - (os.conn != C_STARTING_SYNC_S && ns.conn == C_STARTING_SYNC_S) || - (os.disk != D_FAILED && ns.disk == D_FAILED))) || - (os.conn >= C_CONNECTED && ns.conn == C_DISCONNECTING) || - (os.conn == C_CONNECTED && ns.conn == C_VERIFY_S); -} - -enum drbd_state_rv -drbd_change_state(struct drbd_conf *mdev, enum chg_state_flags f, - union drbd_state mask, union drbd_state val) -{ - unsigned long flags; - union drbd_state os, ns; - enum drbd_state_rv rv; - - spin_lock_irqsave(&mdev->req_lock, flags); - os = mdev->state; - ns.i = (os.i & ~mask.i) | val.i; - rv = _drbd_set_state(mdev, ns, f, NULL); - ns = mdev->state; - spin_unlock_irqrestore(&mdev->req_lock, flags); - - return rv; -} - -/** - * drbd_force_state() - Impose a change which happens outside our control on our state - * @mdev: DRBD device. - * @mask: mask of state bits to change. - * @val: value of new state bits. - */ -void drbd_force_state(struct drbd_conf *mdev, - union drbd_state mask, union drbd_state val) -{ - drbd_change_state(mdev, CS_HARD, mask, val); -} - -static enum drbd_state_rv is_valid_state(struct drbd_conf *, union drbd_state); -static enum drbd_state_rv is_valid_state_transition(struct drbd_conf *, - union drbd_state, - union drbd_state); -enum sanitize_state_warnings { - NO_WARNING, - ABORTED_ONLINE_VERIFY, - ABORTED_RESYNC, - CONNECTION_LOST_NEGOTIATING, - IMPLICITLY_UPGRADED_DISK, - IMPLICITLY_UPGRADED_PDSK, -}; -static union drbd_state sanitize_state(struct drbd_conf *mdev, union drbd_state os, - union drbd_state ns, enum sanitize_state_warnings *warn); -int drbd_send_state_req(struct drbd_conf *, - union drbd_state, union drbd_state); - -static enum drbd_state_rv -_req_st_cond(struct drbd_conf *mdev, union drbd_state mask, - union drbd_state val) -{ - union drbd_state os, ns; - unsigned long flags; - enum drbd_state_rv rv; - - if (test_and_clear_bit(CL_ST_CHG_SUCCESS, &mdev->flags)) - return SS_CW_SUCCESS; - - if (test_and_clear_bit(CL_ST_CHG_FAIL, &mdev->flags)) - return SS_CW_FAILED_BY_PEER; - - rv = 0; - spin_lock_irqsave(&mdev->req_lock, flags); - os = mdev->state; - ns.i = (os.i & ~mask.i) | val.i; - ns = sanitize_state(mdev, os, ns, NULL); - - if (!cl_wide_st_chg(mdev, os, ns)) - rv = SS_CW_NO_NEED; - if (!rv) { - rv = is_valid_state(mdev, ns); - if (rv == SS_SUCCESS) { - rv = is_valid_state_transition(mdev, ns, os); - if (rv == SS_SUCCESS) - rv = SS_UNKNOWN_ERROR; /* cont waiting, otherwise fail. */ - } - } - spin_unlock_irqrestore(&mdev->req_lock, flags); - - return rv; -} - -/** - * drbd_req_state() - Perform an eventually cluster wide state change - * @mdev: DRBD device. - * @mask: mask of state bits to change. - * @val: value of new state bits. - * @f: flags - * - * Should not be called directly, use drbd_request_state() or - * _drbd_request_state(). - */ -static enum drbd_state_rv -drbd_req_state(struct drbd_conf *mdev, union drbd_state mask, - union drbd_state val, enum chg_state_flags f) -{ - struct completion done; - unsigned long flags; - union drbd_state os, ns; - enum drbd_state_rv rv; - - init_completion(&done); - - if (f & CS_SERIALIZE) - mutex_lock(&mdev->state_mutex); - - spin_lock_irqsave(&mdev->req_lock, flags); - os = mdev->state; - ns.i = (os.i & ~mask.i) | val.i; - ns = sanitize_state(mdev, os, ns, NULL); - - if (cl_wide_st_chg(mdev, os, ns)) { - rv = is_valid_state(mdev, ns); - if (rv == SS_SUCCESS) - rv = is_valid_state_transition(mdev, ns, os); - spin_unlock_irqrestore(&mdev->req_lock, flags); - - if (rv < SS_SUCCESS) { - if (f & CS_VERBOSE) - print_st_err(mdev, os, ns, rv); - goto abort; - } - - drbd_state_lock(mdev); - if (!drbd_send_state_req(mdev, mask, val)) { - drbd_state_unlock(mdev); - rv = SS_CW_FAILED_BY_PEER; - if (f & CS_VERBOSE) - print_st_err(mdev, os, ns, rv); - goto abort; - } - - wait_event(mdev->state_wait, - (rv = _req_st_cond(mdev, mask, val))); - - if (rv < SS_SUCCESS) { - drbd_state_unlock(mdev); - if (f & CS_VERBOSE) - print_st_err(mdev, os, ns, rv); - goto abort; - } - spin_lock_irqsave(&mdev->req_lock, flags); - os = mdev->state; - ns.i = (os.i & ~mask.i) | val.i; - rv = _drbd_set_state(mdev, ns, f, &done); - drbd_state_unlock(mdev); - } else { - rv = _drbd_set_state(mdev, ns, f, &done); - } - - spin_unlock_irqrestore(&mdev->req_lock, flags); - - if (f & CS_WAIT_COMPLETE && rv == SS_SUCCESS) { - D_ASSERT(current != mdev->worker.task); - wait_for_completion(&done); - } - -abort: - if (f & CS_SERIALIZE) - mutex_unlock(&mdev->state_mutex); - - return rv; -} - -/** - * _drbd_request_state() - Request a state change (with flags) - * @mdev: DRBD device. - * @mask: mask of state bits to change. - * @val: value of new state bits. - * @f: flags - * - * Cousin of drbd_request_state(), useful with the CS_WAIT_COMPLETE - * flag, or when logging of failed state change requests is not desired. - */ -enum drbd_state_rv -_drbd_request_state(struct drbd_conf *mdev, union drbd_state mask, - union drbd_state val, enum chg_state_flags f) -{ - enum drbd_state_rv rv; - - wait_event(mdev->state_wait, - (rv = drbd_req_state(mdev, mask, val, f)) != SS_IN_TRANSIENT_STATE); - - return rv; -} - -static void print_st(struct drbd_conf *mdev, char *name, union drbd_state ns) -{ - dev_err(DEV, " %s = { cs:%s ro:%s/%s ds:%s/%s %c%c%c%c }\n", - name, - drbd_conn_str(ns.conn), - drbd_role_str(ns.role), - drbd_role_str(ns.peer), - drbd_disk_str(ns.disk), - drbd_disk_str(ns.pdsk), - is_susp(ns) ? 's' : 'r', - ns.aftr_isp ? 'a' : '-', - ns.peer_isp ? 'p' : '-', - ns.user_isp ? 'u' : '-' - ); -} - -void print_st_err(struct drbd_conf *mdev, union drbd_state os, - union drbd_state ns, enum drbd_state_rv err) -{ - if (err == SS_IN_TRANSIENT_STATE) - return; - dev_err(DEV, "State change failed: %s\n", drbd_set_st_err_str(err)); - print_st(mdev, " state", os); - print_st(mdev, "wanted", ns); -} - - -/** - * is_valid_state() - Returns an SS_ error code if ns is not valid - * @mdev: DRBD device. - * @ns: State to consider. - */ -static enum drbd_state_rv -is_valid_state(struct drbd_conf *mdev, union drbd_state ns) -{ - /* See drbd_state_sw_errors in drbd_strings.c */ - - enum drbd_fencing_p fp; - enum drbd_state_rv rv = SS_SUCCESS; - - fp = FP_DONT_CARE; - if (get_ldev(mdev)) { - fp = mdev->ldev->dc.fencing; - put_ldev(mdev); - } - - if (get_net_conf(mdev)) { - if (!mdev->net_conf->two_primaries && - ns.role == R_PRIMARY && ns.peer == R_PRIMARY) - rv = SS_TWO_PRIMARIES; - put_net_conf(mdev); - } - - if (rv <= 0) - /* already found a reason to abort */; - else if (ns.role == R_SECONDARY && mdev->open_cnt) - rv = SS_DEVICE_IN_USE; - - else if (ns.role == R_PRIMARY && ns.conn < C_CONNECTED && ns.disk < D_UP_TO_DATE) - rv = SS_NO_UP_TO_DATE_DISK; - - else if (fp >= FP_RESOURCE && - ns.role == R_PRIMARY && ns.conn < C_CONNECTED && ns.pdsk >= D_UNKNOWN) - rv = SS_PRIMARY_NOP; - - else if (ns.role == R_PRIMARY && ns.disk <= D_INCONSISTENT && ns.pdsk <= D_INCONSISTENT) - rv = SS_NO_UP_TO_DATE_DISK; - - else if (ns.conn > C_CONNECTED && ns.disk < D_INCONSISTENT) - rv = SS_NO_LOCAL_DISK; - - else if (ns.conn > C_CONNECTED && ns.pdsk < D_INCONSISTENT) - rv = SS_NO_REMOTE_DISK; - - else if (ns.conn > C_CONNECTED && ns.disk < D_UP_TO_DATE && ns.pdsk < D_UP_TO_DATE) - rv = SS_NO_UP_TO_DATE_DISK; - - else if ((ns.conn == C_CONNECTED || - ns.conn == C_WF_BITMAP_S || - ns.conn == C_SYNC_SOURCE || - ns.conn == C_PAUSED_SYNC_S) && - ns.disk == D_OUTDATED) - rv = SS_CONNECTED_OUTDATES; - - else if ((ns.conn == C_VERIFY_S || ns.conn == C_VERIFY_T) && - (mdev->sync_conf.verify_alg[0] == 0)) - rv = SS_NO_VERIFY_ALG; - - else if ((ns.conn == C_VERIFY_S || ns.conn == C_VERIFY_T) && - mdev->agreed_pro_version < 88) - rv = SS_NOT_SUPPORTED; - - else if (ns.conn >= C_CONNECTED && ns.pdsk == D_UNKNOWN) - rv = SS_CONNECTED_OUTDATES; - - return rv; -} - -/** - * is_valid_state_transition() - Returns an SS_ error code if the state transition is not possible - * @mdev: DRBD device. - * @ns: new state. - * @os: old state. - */ -static enum drbd_state_rv -is_valid_state_transition(struct drbd_conf *mdev, union drbd_state ns, - union drbd_state os) -{ - enum drbd_state_rv rv = SS_SUCCESS; - - if ((ns.conn == C_STARTING_SYNC_T || ns.conn == C_STARTING_SYNC_S) && - os.conn > C_CONNECTED) - rv = SS_RESYNC_RUNNING; - - if (ns.conn == C_DISCONNECTING && os.conn == C_STANDALONE) - rv = SS_ALREADY_STANDALONE; - - if (ns.disk > D_ATTACHING && os.disk == D_DISKLESS) - rv = SS_IS_DISKLESS; - - if (ns.conn == C_WF_CONNECTION && os.conn < C_UNCONNECTED) - rv = SS_NO_NET_CONFIG; - - if (ns.disk == D_OUTDATED && os.disk < D_OUTDATED && os.disk != D_ATTACHING) - rv = SS_LOWER_THAN_OUTDATED; - - if (ns.conn == C_DISCONNECTING && os.conn == C_UNCONNECTED) - rv = SS_IN_TRANSIENT_STATE; - - if (ns.conn == os.conn && ns.conn == C_WF_REPORT_PARAMS) - rv = SS_IN_TRANSIENT_STATE; - - /* While establishing a connection only allow cstate to change. - Delay/refuse role changes, detach attach etc... */ - if (test_bit(STATE_SENT, &mdev->flags) && - !(os.conn == C_WF_REPORT_PARAMS || - (ns.conn == C_WF_REPORT_PARAMS && os.conn == C_WF_CONNECTION))) - rv = SS_IN_TRANSIENT_STATE; - - if ((ns.conn == C_VERIFY_S || ns.conn == C_VERIFY_T) && os.conn < C_CONNECTED) - rv = SS_NEED_CONNECTION; - - if ((ns.conn == C_VERIFY_S || ns.conn == C_VERIFY_T) && - ns.conn != os.conn && os.conn > C_CONNECTED) - rv = SS_RESYNC_RUNNING; - - if ((ns.conn == C_STARTING_SYNC_S || ns.conn == C_STARTING_SYNC_T) && - os.conn < C_CONNECTED) - rv = SS_NEED_CONNECTION; - - if ((ns.conn == C_SYNC_TARGET || ns.conn == C_SYNC_SOURCE) - && os.conn < C_WF_REPORT_PARAMS) - rv = SS_NEED_CONNECTION; /* No NetworkFailure -> SyncTarget etc... */ - - return rv; -} - -static void print_sanitize_warnings(struct drbd_conf *mdev, enum sanitize_state_warnings warn) -{ - static const char *msg_table[] = { - [NO_WARNING] = "", - [ABORTED_ONLINE_VERIFY] = "Online-verify aborted.", - [ABORTED_RESYNC] = "Resync aborted.", - [CONNECTION_LOST_NEGOTIATING] = "Connection lost while negotiating, no data!", - [IMPLICITLY_UPGRADED_DISK] = "Implicitly upgraded disk", - [IMPLICITLY_UPGRADED_PDSK] = "Implicitly upgraded pdsk", - }; - - if (warn != NO_WARNING) - dev_warn(DEV, "%s\n", msg_table[warn]); -} - -/** - * sanitize_state() - Resolves implicitly necessary additional changes to a state transition - * @mdev: DRBD device. - * @os: old state. - * @ns: new state. - * @warn_sync_abort: - * - * When we loose connection, we have to set the state of the peers disk (pdsk) - * to D_UNKNOWN. This rule and many more along those lines are in this function. - */ -static union drbd_state sanitize_state(struct drbd_conf *mdev, union drbd_state os, - union drbd_state ns, enum sanitize_state_warnings *warn) -{ - enum drbd_fencing_p fp; - enum drbd_disk_state disk_min, disk_max, pdsk_min, pdsk_max; - - if (warn) - *warn = NO_WARNING; - - fp = FP_DONT_CARE; - if (get_ldev(mdev)) { - fp = mdev->ldev->dc.fencing; - put_ldev(mdev); - } - - /* Disallow Network errors to configure a device's network part */ - if ((ns.conn >= C_TIMEOUT && ns.conn <= C_TEAR_DOWN) && - os.conn <= C_DISCONNECTING) - ns.conn = os.conn; - - /* After a network error (+C_TEAR_DOWN) only C_UNCONNECTED or C_DISCONNECTING can follow. - * If you try to go into some Sync* state, that shall fail (elsewhere). */ - if (os.conn >= C_TIMEOUT && os.conn <= C_TEAR_DOWN && - ns.conn != C_UNCONNECTED && ns.conn != C_DISCONNECTING && ns.conn <= C_CONNECTED) - ns.conn = os.conn; - - /* we cannot fail (again) if we already detached */ - if (ns.disk == D_FAILED && os.disk == D_DISKLESS) - ns.disk = D_DISKLESS; - - /* After C_DISCONNECTING only C_STANDALONE may follow */ - if (os.conn == C_DISCONNECTING && ns.conn != C_STANDALONE) - ns.conn = os.conn; - - if (ns.conn < C_CONNECTED) { - ns.peer_isp = 0; - ns.peer = R_UNKNOWN; - if (ns.pdsk > D_UNKNOWN || ns.pdsk < D_INCONSISTENT) - ns.pdsk = D_UNKNOWN; - } - - /* Clear the aftr_isp when becoming unconfigured */ - if (ns.conn == C_STANDALONE && ns.disk == D_DISKLESS && ns.role == R_SECONDARY) - ns.aftr_isp = 0; - - /* Abort resync if a disk fails/detaches */ - if (os.conn > C_CONNECTED && ns.conn > C_CONNECTED && - (ns.disk <= D_FAILED || ns.pdsk <= D_FAILED)) { - if (warn) - *warn = os.conn == C_VERIFY_S || os.conn == C_VERIFY_T ? - ABORTED_ONLINE_VERIFY : ABORTED_RESYNC; - ns.conn = C_CONNECTED; - } - - /* Connection breaks down before we finished "Negotiating" */ - if (ns.conn < C_CONNECTED && ns.disk == D_NEGOTIATING && - get_ldev_if_state(mdev, D_NEGOTIATING)) { - if (mdev->ed_uuid == mdev->ldev->md.uuid[UI_CURRENT]) { - ns.disk = mdev->new_state_tmp.disk; - ns.pdsk = mdev->new_state_tmp.pdsk; - } else { - if (warn) - *warn = CONNECTION_LOST_NEGOTIATING; - ns.disk = D_DISKLESS; - ns.pdsk = D_UNKNOWN; - } - put_ldev(mdev); - } - - /* D_CONSISTENT and D_OUTDATED vanish when we get connected */ - if (ns.conn >= C_CONNECTED && ns.conn < C_AHEAD) { - if (ns.disk == D_CONSISTENT || ns.disk == D_OUTDATED) - ns.disk = D_UP_TO_DATE; - if (ns.pdsk == D_CONSISTENT || ns.pdsk == D_OUTDATED) - ns.pdsk = D_UP_TO_DATE; - } - - /* Implications of the connection stat on the disk states */ - disk_min = D_DISKLESS; - disk_max = D_UP_TO_DATE; - pdsk_min = D_INCONSISTENT; - pdsk_max = D_UNKNOWN; - switch ((enum drbd_conns)ns.conn) { - case C_WF_BITMAP_T: - case C_PAUSED_SYNC_T: - case C_STARTING_SYNC_T: - case C_WF_SYNC_UUID: - case C_BEHIND: - disk_min = D_INCONSISTENT; - disk_max = D_OUTDATED; - pdsk_min = D_UP_TO_DATE; - pdsk_max = D_UP_TO_DATE; - break; - case C_VERIFY_S: - case C_VERIFY_T: - disk_min = D_UP_TO_DATE; - disk_max = D_UP_TO_DATE; - pdsk_min = D_UP_TO_DATE; - pdsk_max = D_UP_TO_DATE; - break; - case C_CONNECTED: - disk_min = D_DISKLESS; - disk_max = D_UP_TO_DATE; - pdsk_min = D_DISKLESS; - pdsk_max = D_UP_TO_DATE; - break; - case C_WF_BITMAP_S: - case C_PAUSED_SYNC_S: - case C_STARTING_SYNC_S: - case C_AHEAD: - disk_min = D_UP_TO_DATE; - disk_max = D_UP_TO_DATE; - pdsk_min = D_INCONSISTENT; - pdsk_max = D_CONSISTENT; /* D_OUTDATED would be nice. But explicit outdate necessary*/ - break; - case C_SYNC_TARGET: - disk_min = D_INCONSISTENT; - disk_max = D_INCONSISTENT; - pdsk_min = D_UP_TO_DATE; - pdsk_max = D_UP_TO_DATE; - break; - case C_SYNC_SOURCE: - disk_min = D_UP_TO_DATE; - disk_max = D_UP_TO_DATE; - pdsk_min = D_INCONSISTENT; - pdsk_max = D_INCONSISTENT; - break; - case C_STANDALONE: - case C_DISCONNECTING: - case C_UNCONNECTED: - case C_TIMEOUT: - case C_BROKEN_PIPE: - case C_NETWORK_FAILURE: - case C_PROTOCOL_ERROR: - case C_TEAR_DOWN: - case C_WF_CONNECTION: - case C_WF_REPORT_PARAMS: - case C_MASK: - break; - } - if (ns.disk > disk_max) - ns.disk = disk_max; - - if (ns.disk < disk_min) { - if (warn) - *warn = IMPLICITLY_UPGRADED_DISK; - ns.disk = disk_min; - } - if (ns.pdsk > pdsk_max) - ns.pdsk = pdsk_max; - - if (ns.pdsk < pdsk_min) { - if (warn) - *warn = IMPLICITLY_UPGRADED_PDSK; - ns.pdsk = pdsk_min; - } - - if (fp == FP_STONITH && - (ns.role == R_PRIMARY && ns.conn < C_CONNECTED && ns.pdsk > D_OUTDATED) && - !(os.role == R_PRIMARY && os.conn < C_CONNECTED && os.pdsk > D_OUTDATED)) - ns.susp_fen = 1; /* Suspend IO while fence-peer handler runs (peer lost) */ - - if (mdev->sync_conf.on_no_data == OND_SUSPEND_IO && - (ns.role == R_PRIMARY && ns.disk < D_UP_TO_DATE && ns.pdsk < D_UP_TO_DATE) && - !(os.role == R_PRIMARY && os.disk < D_UP_TO_DATE && os.pdsk < D_UP_TO_DATE)) - ns.susp_nod = 1; /* Suspend IO while no data available (no accessible data available) */ - - if (ns.aftr_isp || ns.peer_isp || ns.user_isp) { - if (ns.conn == C_SYNC_SOURCE) - ns.conn = C_PAUSED_SYNC_S; - if (ns.conn == C_SYNC_TARGET) - ns.conn = C_PAUSED_SYNC_T; - } else { - if (ns.conn == C_PAUSED_SYNC_S) - ns.conn = C_SYNC_SOURCE; - if (ns.conn == C_PAUSED_SYNC_T) - ns.conn = C_SYNC_TARGET; - } - - return ns; -} - -/* helper for __drbd_set_state */ -static void set_ov_position(struct drbd_conf *mdev, enum drbd_conns cs) -{ - if (mdev->agreed_pro_version < 90) - mdev->ov_start_sector = 0; - mdev->rs_total = drbd_bm_bits(mdev); - mdev->ov_position = 0; - if (cs == C_VERIFY_T) { - /* starting online verify from an arbitrary position - * does not fit well into the existing protocol. - * on C_VERIFY_T, we initialize ov_left and friends - * implicitly in receive_DataRequest once the - * first P_OV_REQUEST is received */ - mdev->ov_start_sector = ~(sector_t)0; - } else { - unsigned long bit = BM_SECT_TO_BIT(mdev->ov_start_sector); - if (bit >= mdev->rs_total) { - mdev->ov_start_sector = - BM_BIT_TO_SECT(mdev->rs_total - 1); - mdev->rs_total = 1; - } else - mdev->rs_total -= bit; - mdev->ov_position = mdev->ov_start_sector; - } - mdev->ov_left = mdev->rs_total; -} - -static void drbd_resume_al(struct drbd_conf *mdev) -{ - if (test_and_clear_bit(AL_SUSPENDED, &mdev->flags)) - dev_info(DEV, "Resumed AL updates\n"); -} - -/** - * __drbd_set_state() - Set a new DRBD state - * @mdev: DRBD device. - * @ns: new state. - * @flags: Flags - * @done: Optional completion, that will get completed after the after_state_ch() finished - * - * Caller needs to hold req_lock, and global_state_lock. Do not call directly. - */ -enum drbd_state_rv -__drbd_set_state(struct drbd_conf *mdev, union drbd_state ns, - enum chg_state_flags flags, struct completion *done) -{ - union drbd_state os; - enum drbd_state_rv rv = SS_SUCCESS; - enum sanitize_state_warnings ssw; - struct after_state_chg_work *ascw; - - os = mdev->state; - - ns = sanitize_state(mdev, os, ns, &ssw); - - if (ns.i == os.i) - return SS_NOTHING_TO_DO; - - if (!(flags & CS_HARD)) { - /* pre-state-change checks ; only look at ns */ - /* See drbd_state_sw_errors in drbd_strings.c */ - - rv = is_valid_state(mdev, ns); - if (rv < SS_SUCCESS) { - /* If the old state was illegal as well, then let - this happen...*/ - - if (is_valid_state(mdev, os) == rv) - rv = is_valid_state_transition(mdev, ns, os); - } else - rv = is_valid_state_transition(mdev, ns, os); - } - - if (rv < SS_SUCCESS) { - if (flags & CS_VERBOSE) - print_st_err(mdev, os, ns, rv); - return rv; - } - - print_sanitize_warnings(mdev, ssw); - - { - char *pbp, pb[300]; - pbp = pb; - *pbp = 0; - if (ns.role != os.role) - pbp += sprintf(pbp, "role( %s -> %s ) ", - drbd_role_str(os.role), - drbd_role_str(ns.role)); - if (ns.peer != os.peer) - pbp += sprintf(pbp, "peer( %s -> %s ) ", - drbd_role_str(os.peer), - drbd_role_str(ns.peer)); - if (ns.conn != os.conn) - pbp += sprintf(pbp, "conn( %s -> %s ) ", - drbd_conn_str(os.conn), - drbd_conn_str(ns.conn)); - if (ns.disk != os.disk) - pbp += sprintf(pbp, "disk( %s -> %s ) ", - drbd_disk_str(os.disk), - drbd_disk_str(ns.disk)); - if (ns.pdsk != os.pdsk) - pbp += sprintf(pbp, "pdsk( %s -> %s ) ", - drbd_disk_str(os.pdsk), - drbd_disk_str(ns.pdsk)); - if (is_susp(ns) != is_susp(os)) - pbp += sprintf(pbp, "susp( %d -> %d ) ", - is_susp(os), - is_susp(ns)); - if (ns.aftr_isp != os.aftr_isp) - pbp += sprintf(pbp, "aftr_isp( %d -> %d ) ", - os.aftr_isp, - ns.aftr_isp); - if (ns.peer_isp != os.peer_isp) - pbp += sprintf(pbp, "peer_isp( %d -> %d ) ", - os.peer_isp, - ns.peer_isp); - if (ns.user_isp != os.user_isp) - pbp += sprintf(pbp, "user_isp( %d -> %d ) ", - os.user_isp, - ns.user_isp); - dev_info(DEV, "%s\n", pb); - } - - /* solve the race between becoming unconfigured, - * worker doing the cleanup, and - * admin reconfiguring us: - * on (re)configure, first set CONFIG_PENDING, - * then wait for a potentially exiting worker, - * start the worker, and schedule one no_op. - * then proceed with configuration. - */ - if (ns.disk == D_DISKLESS && - ns.conn == C_STANDALONE && - ns.role == R_SECONDARY && - !test_and_set_bit(CONFIG_PENDING, &mdev->flags)) - set_bit(DEVICE_DYING, &mdev->flags); - - /* if we are going -> D_FAILED or D_DISKLESS, grab one extra reference - * on the ldev here, to be sure the transition -> D_DISKLESS resp. - * drbd_ldev_destroy() won't happen before our corresponding - * after_state_ch works run, where we put_ldev again. */ - if ((os.disk != D_FAILED && ns.disk == D_FAILED) || - (os.disk != D_DISKLESS && ns.disk == D_DISKLESS)) - atomic_inc(&mdev->local_cnt); - - mdev->state = ns; - - if (os.disk == D_ATTACHING && ns.disk >= D_NEGOTIATING) - drbd_print_uuids(mdev, "attached to UUIDs"); - - wake_up(&mdev->misc_wait); - wake_up(&mdev->state_wait); - - /* aborted verify run. log the last position */ - if ((os.conn == C_VERIFY_S || os.conn == C_VERIFY_T) && - ns.conn < C_CONNECTED) { - mdev->ov_start_sector = - BM_BIT_TO_SECT(drbd_bm_bits(mdev) - mdev->ov_left); - dev_info(DEV, "Online Verify reached sector %llu\n", - (unsigned long long)mdev->ov_start_sector); - } - - if ((os.conn == C_PAUSED_SYNC_T || os.conn == C_PAUSED_SYNC_S) && - (ns.conn == C_SYNC_TARGET || ns.conn == C_SYNC_SOURCE)) { - dev_info(DEV, "Syncer continues.\n"); - mdev->rs_paused += (long)jiffies - -(long)mdev->rs_mark_time[mdev->rs_last_mark]; - if (ns.conn == C_SYNC_TARGET) - mod_timer(&mdev->resync_timer, jiffies); - } - - if ((os.conn == C_SYNC_TARGET || os.conn == C_SYNC_SOURCE) && - (ns.conn == C_PAUSED_SYNC_T || ns.conn == C_PAUSED_SYNC_S)) { - dev_info(DEV, "Resync suspended\n"); - mdev->rs_mark_time[mdev->rs_last_mark] = jiffies; - } - - if (os.conn == C_CONNECTED && - (ns.conn == C_VERIFY_S || ns.conn == C_VERIFY_T)) { - unsigned long now = jiffies; - int i; - - set_ov_position(mdev, ns.conn); - mdev->rs_start = now; - mdev->rs_last_events = 0; - mdev->rs_last_sect_ev = 0; - mdev->ov_last_oos_size = 0; - mdev->ov_last_oos_start = 0; - - for (i = 0; i < DRBD_SYNC_MARKS; i++) { - mdev->rs_mark_left[i] = mdev->ov_left; - mdev->rs_mark_time[i] = now; - } - - drbd_rs_controller_reset(mdev); - - if (ns.conn == C_VERIFY_S) { - dev_info(DEV, "Starting Online Verify from sector %llu\n", - (unsigned long long)mdev->ov_position); - mod_timer(&mdev->resync_timer, jiffies); - } - } - - if (get_ldev(mdev)) { - u32 mdf = mdev->ldev->md.flags & ~(MDF_CONSISTENT|MDF_PRIMARY_IND| - MDF_CONNECTED_IND|MDF_WAS_UP_TO_DATE| - MDF_PEER_OUT_DATED|MDF_CRASHED_PRIMARY); - - if (test_bit(CRASHED_PRIMARY, &mdev->flags)) - mdf |= MDF_CRASHED_PRIMARY; - if (mdev->state.role == R_PRIMARY || - (mdev->state.pdsk < D_INCONSISTENT && mdev->state.peer == R_PRIMARY)) - mdf |= MDF_PRIMARY_IND; - if (mdev->state.conn > C_WF_REPORT_PARAMS) - mdf |= MDF_CONNECTED_IND; - if (mdev->state.disk > D_INCONSISTENT) - mdf |= MDF_CONSISTENT; - if (mdev->state.disk > D_OUTDATED) - mdf |= MDF_WAS_UP_TO_DATE; - if (mdev->state.pdsk <= D_OUTDATED && mdev->state.pdsk >= D_INCONSISTENT) - mdf |= MDF_PEER_OUT_DATED; - if (mdf != mdev->ldev->md.flags) { - mdev->ldev->md.flags = mdf; - drbd_md_mark_dirty(mdev); - } - if (os.disk < D_CONSISTENT && ns.disk >= D_CONSISTENT) - drbd_set_ed_uuid(mdev, mdev->ldev->md.uuid[UI_CURRENT]); - put_ldev(mdev); - } - - /* Peer was forced D_UP_TO_DATE & R_PRIMARY, consider to resync */ - if (os.disk == D_INCONSISTENT && os.pdsk == D_INCONSISTENT && - os.peer == R_SECONDARY && ns.peer == R_PRIMARY) - set_bit(CONSIDER_RESYNC, &mdev->flags); - - /* Receiver should clean up itself */ - if (os.conn != C_DISCONNECTING && ns.conn == C_DISCONNECTING) - drbd_thread_stop_nowait(&mdev->receiver); - - /* Now the receiver finished cleaning up itself, it should die */ - if (os.conn != C_STANDALONE && ns.conn == C_STANDALONE) - drbd_thread_stop_nowait(&mdev->receiver); - - /* Upon network failure, we need to restart the receiver. */ - if (os.conn > C_WF_CONNECTION && - ns.conn <= C_TEAR_DOWN && ns.conn >= C_TIMEOUT) - drbd_thread_restart_nowait(&mdev->receiver); - - /* Resume AL writing if we get a connection */ - if (os.conn < C_CONNECTED && ns.conn >= C_CONNECTED) - drbd_resume_al(mdev); - - /* remember last connect and attach times so request_timer_fn() won't - * kill newly established sessions while we are still trying to thaw - * previously frozen IO */ - if (os.conn != C_WF_REPORT_PARAMS && ns.conn == C_WF_REPORT_PARAMS) - mdev->last_reconnect_jif = jiffies; - if ((os.disk == D_ATTACHING || os.disk == D_NEGOTIATING) && - ns.disk > D_NEGOTIATING) - mdev->last_reattach_jif = jiffies; - - ascw = kmalloc(sizeof(*ascw), GFP_ATOMIC); - if (ascw) { - ascw->os = os; - ascw->ns = ns; - ascw->flags = flags; - ascw->w.cb = w_after_state_ch; - ascw->done = done; - drbd_queue_work(&mdev->data.work, &ascw->w); - } else { - dev_warn(DEV, "Could not kmalloc an ascw\n"); - } - - return rv; -} - -static int w_after_state_ch(struct drbd_conf *mdev, struct drbd_work *w, int unused) -{ - struct after_state_chg_work *ascw = - container_of(w, struct after_state_chg_work, w); - after_state_ch(mdev, ascw->os, ascw->ns, ascw->flags); - if (ascw->flags & CS_WAIT_COMPLETE) { - D_ASSERT(ascw->done != NULL); - complete(ascw->done); - } - kfree(ascw); - - return 1; -} - -static void abw_start_sync(struct drbd_conf *mdev, int rv) -{ - if (rv) { - dev_err(DEV, "Writing the bitmap failed not starting resync.\n"); - _drbd_request_state(mdev, NS(conn, C_CONNECTED), CS_VERBOSE); - return; - } - - switch (mdev->state.conn) { - case C_STARTING_SYNC_T: - _drbd_request_state(mdev, NS(conn, C_WF_SYNC_UUID), CS_VERBOSE); - break; - case C_STARTING_SYNC_S: - drbd_start_resync(mdev, C_SYNC_SOURCE); - break; - } -} - -int drbd_bitmap_io_from_worker(struct drbd_conf *mdev, - int (*io_fn)(struct drbd_conf *), - char *why, enum bm_flag flags) -{ - int rv; - - D_ASSERT(current == mdev->worker.task); - - /* open coded non-blocking drbd_suspend_io(mdev); */ - set_bit(SUSPEND_IO, &mdev->flags); - - drbd_bm_lock(mdev, why, flags); - rv = io_fn(mdev); - drbd_bm_unlock(mdev); - - drbd_resume_io(mdev); - - return rv; -} - -/** - * after_state_ch() - Perform after state change actions that may sleep - * @mdev: DRBD device. - * @os: old state. - * @ns: new state. - * @flags: Flags - */ -static void after_state_ch(struct drbd_conf *mdev, union drbd_state os, - union drbd_state ns, enum chg_state_flags flags) -{ - enum drbd_fencing_p fp; - enum drbd_req_event what = nothing; - union drbd_state nsm = (union drbd_state){ .i = -1 }; - - if (os.conn != C_CONNECTED && ns.conn == C_CONNECTED) { - clear_bit(CRASHED_PRIMARY, &mdev->flags); - if (mdev->p_uuid) - mdev->p_uuid[UI_FLAGS] &= ~((u64)2); - } - - fp = FP_DONT_CARE; - if (get_ldev(mdev)) { - fp = mdev->ldev->dc.fencing; - put_ldev(mdev); - } - - /* Inform userspace about the change... */ - drbd_bcast_state(mdev, ns); - - if (!(os.role == R_PRIMARY && os.disk < D_UP_TO_DATE && os.pdsk < D_UP_TO_DATE) && - (ns.role == R_PRIMARY && ns.disk < D_UP_TO_DATE && ns.pdsk < D_UP_TO_DATE)) - drbd_khelper(mdev, "pri-on-incon-degr"); - - /* Here we have the actions that are performed after a - state change. This function might sleep */ - - if (os.disk <= D_NEGOTIATING && ns.disk > D_NEGOTIATING) - mod_timer(&mdev->request_timer, jiffies + HZ); - - nsm.i = -1; - if (ns.susp_nod) { - if (os.conn < C_CONNECTED && ns.conn >= C_CONNECTED) - what = resend; - - if ((os.disk == D_ATTACHING || os.disk == D_NEGOTIATING) && - ns.disk > D_NEGOTIATING) - what = restart_frozen_disk_io; - - if (what != nothing) - nsm.susp_nod = 0; - } - - if (ns.susp_fen) { - /* case1: The outdate peer handler is successful: */ - if (os.pdsk > D_OUTDATED && ns.pdsk <= D_OUTDATED) { - if (test_bit(NEW_CUR_UUID, &mdev->flags)) { - drbd_uuid_new_current(mdev); - clear_bit(NEW_CUR_UUID, &mdev->flags); - } - spin_lock_irq(&mdev->req_lock); - _tl_clear(mdev); - _drbd_set_state(_NS(mdev, susp_fen, 0), CS_VERBOSE, NULL); - spin_unlock_irq(&mdev->req_lock); - } - /* case2: The connection was established again: */ - if (os.conn < C_CONNECTED && ns.conn >= C_CONNECTED) { - clear_bit(NEW_CUR_UUID, &mdev->flags); - what = resend; - nsm.susp_fen = 0; - } - } - - if (what != nothing) { - spin_lock_irq(&mdev->req_lock); - _tl_restart(mdev, what); - nsm.i &= mdev->state.i; - _drbd_set_state(mdev, nsm, CS_VERBOSE, NULL); - spin_unlock_irq(&mdev->req_lock); - } - - /* Became sync source. With protocol >= 96, we still need to send out - * the sync uuid now. Need to do that before any drbd_send_state, or - * the other side may go "paused sync" before receiving the sync uuids, - * which is unexpected. */ - if ((os.conn != C_SYNC_SOURCE && os.conn != C_PAUSED_SYNC_S) && - (ns.conn == C_SYNC_SOURCE || ns.conn == C_PAUSED_SYNC_S) && - mdev->agreed_pro_version >= 96 && get_ldev(mdev)) { - drbd_gen_and_send_sync_uuid(mdev); - put_ldev(mdev); - } - - /* Do not change the order of the if above and the two below... */ - if (os.pdsk == D_DISKLESS && ns.pdsk > D_DISKLESS) { /* attach on the peer */ - /* we probably will start a resync soon. - * make sure those things are properly reset. */ - mdev->rs_total = 0; - mdev->rs_failed = 0; - atomic_set(&mdev->rs_pending_cnt, 0); - drbd_rs_cancel_all(mdev); - - drbd_send_uuids(mdev); - drbd_send_state(mdev, ns); - } - /* No point in queuing send_bitmap if we don't have a connection - * anymore, so check also the _current_ state, not only the new state - * at the time this work was queued. */ - if (os.conn != C_WF_BITMAP_S && ns.conn == C_WF_BITMAP_S && - mdev->state.conn == C_WF_BITMAP_S) - drbd_queue_bitmap_io(mdev, &drbd_send_bitmap, NULL, - "send_bitmap (WFBitMapS)", - BM_LOCKED_TEST_ALLOWED); - - /* Lost contact to peer's copy of the data */ - if ((os.pdsk >= D_INCONSISTENT && - os.pdsk != D_UNKNOWN && - os.pdsk != D_OUTDATED) - && (ns.pdsk < D_INCONSISTENT || - ns.pdsk == D_UNKNOWN || - ns.pdsk == D_OUTDATED)) { - if (get_ldev(mdev)) { - if ((ns.role == R_PRIMARY || ns.peer == R_PRIMARY) && - mdev->ldev->md.uuid[UI_BITMAP] == 0 && ns.disk >= D_UP_TO_DATE) { - if (is_susp(mdev->state)) { - set_bit(NEW_CUR_UUID, &mdev->flags); - } else { - drbd_uuid_new_current(mdev); - drbd_send_uuids(mdev); - } - } - put_ldev(mdev); - } - } - - if (ns.pdsk < D_INCONSISTENT && get_ldev(mdev)) { - if (os.peer == R_SECONDARY && ns.peer == R_PRIMARY && - mdev->ldev->md.uuid[UI_BITMAP] == 0 && ns.disk >= D_UP_TO_DATE) { - drbd_uuid_new_current(mdev); - drbd_send_uuids(mdev); - } - /* D_DISKLESS Peer becomes secondary */ - if (os.peer == R_PRIMARY && ns.peer == R_SECONDARY) - /* We may still be Primary ourselves. - * No harm done if the bitmap still changes, - * redirtied pages will follow later. */ - drbd_bitmap_io_from_worker(mdev, &drbd_bm_write, - "demote diskless peer", BM_LOCKED_SET_ALLOWED); - put_ldev(mdev); - } - - /* Write out all changed bits on demote. - * Though, no need to da that just yet - * if there is a resync going on still */ - if (os.role == R_PRIMARY && ns.role == R_SECONDARY && - mdev->state.conn <= C_CONNECTED && get_ldev(mdev)) { - /* No changes to the bitmap expected this time, so assert that, - * even though no harm was done if it did change. */ - drbd_bitmap_io_from_worker(mdev, &drbd_bm_write, - "demote", BM_LOCKED_TEST_ALLOWED); - put_ldev(mdev); - } - - /* Last part of the attaching process ... */ - if (ns.conn >= C_CONNECTED && - os.disk == D_ATTACHING && ns.disk == D_NEGOTIATING) { - drbd_send_sizes(mdev, 0, 0); /* to start sync... */ - drbd_send_uuids(mdev); - drbd_send_state(mdev, ns); - } - - /* We want to pause/continue resync, tell peer. */ - if (ns.conn >= C_CONNECTED && - ((os.aftr_isp != ns.aftr_isp) || - (os.user_isp != ns.user_isp))) - drbd_send_state(mdev, ns); - - /* In case one of the isp bits got set, suspend other devices. */ - if ((!os.aftr_isp && !os.peer_isp && !os.user_isp) && - (ns.aftr_isp || ns.peer_isp || ns.user_isp)) - suspend_other_sg(mdev); - - /* Make sure the peer gets informed about eventual state - changes (ISP bits) while we were in WFReportParams. */ - if (os.conn == C_WF_REPORT_PARAMS && ns.conn >= C_CONNECTED) - drbd_send_state(mdev, ns); - - if (os.conn != C_AHEAD && ns.conn == C_AHEAD) - drbd_send_state(mdev, ns); - - /* We are in the progress to start a full sync... */ - if ((os.conn != C_STARTING_SYNC_T && ns.conn == C_STARTING_SYNC_T) || - (os.conn != C_STARTING_SYNC_S && ns.conn == C_STARTING_SYNC_S)) - /* no other bitmap changes expected during this phase */ - drbd_queue_bitmap_io(mdev, - &drbd_bmio_set_n_write, &abw_start_sync, - "set_n_write from StartingSync", BM_LOCKED_TEST_ALLOWED); - - /* We are invalidating our self... */ - if (os.conn < C_CONNECTED && ns.conn < C_CONNECTED && - os.disk > D_INCONSISTENT && ns.disk == D_INCONSISTENT) - /* other bitmap operation expected during this phase */ - drbd_queue_bitmap_io(mdev, &drbd_bmio_set_n_write, NULL, - "set_n_write from invalidate", BM_LOCKED_MASK); - - /* first half of local IO error, failure to attach, - * or administrative detach */ - if (os.disk != D_FAILED && ns.disk == D_FAILED) { - enum drbd_io_error_p eh = EP_PASS_ON; - int was_io_error = 0; - /* corresponding get_ldev was in __drbd_set_state, to serialize - * our cleanup here with the transition to D_DISKLESS. - * But is is still not save to dreference ldev here, since - * we might come from an failed Attach before ldev was set. */ - if (mdev->ldev) { - eh = mdev->ldev->dc.on_io_error; - was_io_error = test_and_clear_bit(WAS_IO_ERROR, &mdev->flags); - - if (was_io_error && eh == EP_CALL_HELPER) - drbd_khelper(mdev, "local-io-error"); - - /* Immediately allow completion of all application IO, - * that waits for completion from the local disk, - * if this was a force-detach due to disk_timeout - * or administrator request (drbdsetup detach --force). - * Do NOT abort otherwise. - * Aborting local requests may cause serious problems, - * if requests are completed to upper layers already, - * and then later the already submitted local bio completes. - * This can cause DMA into former bio pages that meanwhile - * have been re-used for other things. - * So aborting local requests may cause crashes, - * or even worse, silent data corruption. - */ - if (test_and_clear_bit(FORCE_DETACH, &mdev->flags)) - tl_abort_disk_io(mdev); - - /* current state still has to be D_FAILED, - * there is only one way out: to D_DISKLESS, - * and that may only happen after our put_ldev below. */ - if (mdev->state.disk != D_FAILED) - dev_err(DEV, - "ASSERT FAILED: disk is %s during detach\n", - drbd_disk_str(mdev->state.disk)); - - if (ns.conn >= C_CONNECTED) - drbd_send_state(mdev, ns); - - drbd_rs_cancel_all(mdev); - - /* In case we want to get something to stable storage still, - * this may be the last chance. - * Following put_ldev may transition to D_DISKLESS. */ - drbd_md_sync(mdev); - } - put_ldev(mdev); - } - - /* second half of local IO error, failure to attach, - * or administrative detach, - * after local_cnt references have reached zero again */ - if (os.disk != D_DISKLESS && ns.disk == D_DISKLESS) { - /* We must still be diskless, - * re-attach has to be serialized with this! */ - if (mdev->state.disk != D_DISKLESS) - dev_err(DEV, - "ASSERT FAILED: disk is %s while going diskless\n", - drbd_disk_str(mdev->state.disk)); - - if (ns.conn >= C_CONNECTED) - drbd_send_state(mdev, ns); - - /* corresponding get_ldev in __drbd_set_state - * this may finally trigger drbd_ldev_destroy. */ - put_ldev(mdev); - } - - /* Notify peer that I had a local IO error, and did not detached.. */ - if (os.disk == D_UP_TO_DATE && ns.disk == D_INCONSISTENT && ns.conn >= C_CONNECTED) - drbd_send_state(mdev, ns); - - /* Disks got bigger while they were detached */ - if (ns.disk > D_NEGOTIATING && ns.pdsk > D_NEGOTIATING && - test_and_clear_bit(RESYNC_AFTER_NEG, &mdev->flags)) { - if (ns.conn == C_CONNECTED) - resync_after_online_grow(mdev); - } - - /* A resync finished or aborted, wake paused devices... */ - if ((os.conn > C_CONNECTED && ns.conn <= C_CONNECTED) || - (os.peer_isp && !ns.peer_isp) || - (os.user_isp && !ns.user_isp)) - resume_next_sg(mdev); - - /* sync target done with resync. Explicitly notify peer, even though - * it should (at least for non-empty resyncs) already know itself. */ - if (os.disk < D_UP_TO_DATE && os.conn >= C_SYNC_SOURCE && ns.conn == C_CONNECTED) - drbd_send_state(mdev, ns); - - /* Wake up role changes, that were delayed because of connection establishing */ - if (os.conn == C_WF_REPORT_PARAMS && ns.conn != C_WF_REPORT_PARAMS) { - clear_bit(STATE_SENT, &mdev->flags); - wake_up(&mdev->state_wait); - } - - /* This triggers bitmap writeout of potentially still unwritten pages - * if the resync finished cleanly, or aborted because of peer disk - * failure, or because of connection loss. - * For resync aborted because of local disk failure, we cannot do - * any bitmap writeout anymore. - * No harm done if some bits change during this phase. - */ - if (os.conn > C_CONNECTED && ns.conn <= C_CONNECTED && get_ldev(mdev)) { - drbd_queue_bitmap_io(mdev, &drbd_bm_write_copy_pages, NULL, - "write from resync_finished", BM_LOCKED_CHANGE_ALLOWED); - put_ldev(mdev); - } - - /* free tl_hash if we Got thawed and are C_STANDALONE */ - if (ns.conn == C_STANDALONE && !is_susp(ns) && mdev->tl_hash) - drbd_free_tl_hash(mdev); - - /* Upon network connection, we need to start the receiver */ - if (os.conn == C_STANDALONE && ns.conn == C_UNCONNECTED) - drbd_thread_start(&mdev->receiver); - - /* Terminate worker thread if we are unconfigured - it will be - restarted as needed... */ - if (ns.disk == D_DISKLESS && - ns.conn == C_STANDALONE && - ns.role == R_SECONDARY) { - if (os.aftr_isp != ns.aftr_isp) - resume_next_sg(mdev); - /* set in __drbd_set_state, unless CONFIG_PENDING was set */ - if (test_bit(DEVICE_DYING, &mdev->flags)) - drbd_thread_stop_nowait(&mdev->worker); + if (req->w.mdev != mdev) + continue; + _req_mod(req, ABORT_DISK_IO); } - - drbd_md_sync(mdev); + spin_unlock_irq(&tconn->req_lock); } - static int drbd_thread_setup(void *arg) { struct drbd_thread *thi = (struct drbd_thread *) arg; - struct drbd_conf *mdev = thi->mdev; + struct drbd_tconn *tconn = thi->tconn; unsigned long flags; int retval; + snprintf(current->comm, sizeof(current->comm), "drbd_%c_%s", + thi->name[0], thi->tconn->name); + restart: retval = thi->function(thi); spin_lock_irqsave(&thi->t_lock, flags); - /* if the receiver has been "Exiting", the last thing it did + /* if the receiver has been "EXITING", the last thing it did * was set the conn state to "StandAlone", * if now a re-connect request comes in, conn state goes C_UNCONNECTED, * and receiver thread will be "started". - * drbd_thread_start needs to set "Restarting" in that case. + * drbd_thread_start needs to set "RESTARTING" in that case. * t_state check and assignment needs to be within the same spinlock, - * so either thread_start sees Exiting, and can remap to Restarting, - * or thread_start see None, and can proceed as normal. + * so either thread_start sees EXITING, and can remap to RESTARTING, + * or thread_start see NONE, and can proceed as normal. */ - if (thi->t_state == Restarting) { - dev_info(DEV, "Restarting %s\n", current->comm); - thi->t_state = Running; + if (thi->t_state == RESTARTING) { + conn_info(tconn, "Restarting %s thread\n", thi->name); + thi->t_state = RUNNING; spin_unlock_irqrestore(&thi->t_lock, flags); goto restart; } thi->task = NULL; - thi->t_state = None; + thi->t_state = NONE; smp_mb(); - complete(&thi->stop); + complete_all(&thi->stop); spin_unlock_irqrestore(&thi->t_lock, flags); - dev_info(DEV, "Terminating %s\n", current->comm); + conn_info(tconn, "Terminating %s\n", current->comm); /* Release mod reference taken when thread was started */ + + kref_put(&tconn->kref, &conn_destroy); module_put(THIS_MODULE); return retval; } -static void drbd_thread_init(struct drbd_conf *mdev, struct drbd_thread *thi, - int (*func) (struct drbd_thread *)) +static void drbd_thread_init(struct drbd_tconn *tconn, struct drbd_thread *thi, + int (*func) (struct drbd_thread *), char *name) { spin_lock_init(&thi->t_lock); thi->task = NULL; - thi->t_state = None; + thi->t_state = NONE; thi->function = func; - thi->mdev = mdev; + thi->tconn = tconn; + strncpy(thi->name, name, ARRAY_SIZE(thi->name)); } int drbd_thread_start(struct drbd_thread *thi) { - struct drbd_conf *mdev = thi->mdev; + struct drbd_tconn *tconn = thi->tconn; struct task_struct *nt; unsigned long flags; - const char *me = - thi == &mdev->receiver ? "receiver" : - thi == &mdev->asender ? "asender" : - thi == &mdev->worker ? "worker" : "NONSENSE"; - /* is used from state engine doing drbd_thread_stop_nowait, * while holding the req lock irqsave */ spin_lock_irqsave(&thi->t_lock, flags); switch (thi->t_state) { - case None: - dev_info(DEV, "Starting %s thread (from %s [%d])\n", - me, current->comm, current->pid); + case NONE: + conn_info(tconn, "Starting %s thread (from %s [%d])\n", + thi->name, current->comm, current->pid); /* Get ref on module for thread - this is released when thread exits */ if (!try_module_get(THIS_MODULE)) { - dev_err(DEV, "Failed to get module reference in drbd_thread_start\n"); + conn_err(tconn, "Failed to get module reference in drbd_thread_start\n"); spin_unlock_irqrestore(&thi->t_lock, flags); return false; } + kref_get(&thi->tconn->kref); + init_completion(&thi->stop); - D_ASSERT(thi->task == NULL); thi->reset_cpu_mask = 1; - thi->t_state = Running; + thi->t_state = RUNNING; spin_unlock_irqrestore(&thi->t_lock, flags); flush_signals(current); /* otherw. may get -ERESTARTNOINTR */ nt = kthread_create(drbd_thread_setup, (void *) thi, - "drbd%d_%s", mdev_to_minor(mdev), me); + "drbd_%c_%s", thi->name[0], thi->tconn->name); if (IS_ERR(nt)) { - dev_err(DEV, "Couldn't start thread\n"); + conn_err(tconn, "Couldn't start thread\n"); + kref_put(&tconn->kref, &conn_destroy); module_put(THIS_MODULE); return false; } spin_lock_irqsave(&thi->t_lock, flags); thi->task = nt; - thi->t_state = Running; + thi->t_state = RUNNING; spin_unlock_irqrestore(&thi->t_lock, flags); wake_up_process(nt); break; - case Exiting: - thi->t_state = Restarting; - dev_info(DEV, "Restarting %s thread (from %s [%d])\n", - me, current->comm, current->pid); + case EXITING: + thi->t_state = RESTARTING; + conn_info(tconn, "Restarting %s thread (from %s [%d])\n", + thi->name, current->comm, current->pid); /* fall through */ - case Running: - case Restarting: + case RUNNING: + case RESTARTING: default: spin_unlock_irqrestore(&thi->t_lock, flags); break; @@ -1867,12 +447,12 @@ void _drbd_thread_stop(struct drbd_thread *thi, int restart, int wait) { unsigned long flags; - enum drbd_thread_state ns = restart ? Restarting : Exiting; + enum drbd_thread_state ns = restart ? RESTARTING : EXITING; /* may be called from state engine, holding the req lock irqsave */ spin_lock_irqsave(&thi->t_lock, flags); - if (thi->t_state == None) { + if (thi->t_state == NONE) { spin_unlock_irqrestore(&thi->t_lock, flags); if (restart) drbd_thread_start(thi); @@ -1890,7 +470,6 @@ void _drbd_thread_stop(struct drbd_thread *thi, int restart, int wait) init_completion(&thi->stop); if (thi->task != current) force_sig(DRBD_SIGKILL, thi->task); - } spin_unlock_irqrestore(&thi->t_lock, flags); @@ -1899,6 +478,35 @@ void _drbd_thread_stop(struct drbd_thread *thi, int restart, int wait) wait_for_completion(&thi->stop); } +static struct drbd_thread *drbd_task_to_thread(struct drbd_tconn *tconn, struct task_struct *task) +{ + struct drbd_thread *thi = + task == tconn->receiver.task ? &tconn->receiver : + task == tconn->asender.task ? &tconn->asender : + task == tconn->worker.task ? &tconn->worker : NULL; + + return thi; +} + +char *drbd_task_to_thread_name(struct drbd_tconn *tconn, struct task_struct *task) +{ + struct drbd_thread *thi = drbd_task_to_thread(tconn, task); + return thi ? thi->name : task->comm; +} + +int conn_lowest_minor(struct drbd_tconn *tconn) +{ + struct drbd_conf *mdev; + int vnr = 0, m; + + rcu_read_lock(); + mdev = idr_get_next(&tconn->volumes, &vnr); + m = mdev ? mdev_to_minor(mdev) : -1; + rcu_read_unlock(); + + return m; +} + #ifdef CONFIG_SMP /** * drbd_calc_cpu_mask() - Generate CPU masks, spread over all CPUs @@ -1907,238 +515,345 @@ void _drbd_thread_stop(struct drbd_thread *thi, int restart, int wait) * Forces all threads of a device onto the same CPU. This is beneficial for * DRBD's performance. May be overwritten by user's configuration. */ -void drbd_calc_cpu_mask(struct drbd_conf *mdev) +void drbd_calc_cpu_mask(struct drbd_tconn *tconn) { int ord, cpu; /* user override. */ - if (cpumask_weight(mdev->cpu_mask)) + if (cpumask_weight(tconn->cpu_mask)) return; - ord = mdev_to_minor(mdev) % cpumask_weight(cpu_online_mask); + ord = conn_lowest_minor(tconn) % cpumask_weight(cpu_online_mask); for_each_online_cpu(cpu) { if (ord-- == 0) { - cpumask_set_cpu(cpu, mdev->cpu_mask); + cpumask_set_cpu(cpu, tconn->cpu_mask); return; } } /* should not be reached */ - cpumask_setall(mdev->cpu_mask); + cpumask_setall(tconn->cpu_mask); } /** * drbd_thread_current_set_cpu() - modifies the cpu mask of the _current_ thread * @mdev: DRBD device. + * @thi: drbd_thread object * * call in the "main loop" of _all_ threads, no need for any mutex, current won't die * prematurely. */ -void drbd_thread_current_set_cpu(struct drbd_conf *mdev) +void drbd_thread_current_set_cpu(struct drbd_thread *thi) { struct task_struct *p = current; - struct drbd_thread *thi = - p == mdev->asender.task ? &mdev->asender : - p == mdev->receiver.task ? &mdev->receiver : - p == mdev->worker.task ? &mdev->worker : - NULL; - ERR_IF(thi == NULL) - return; + if (!thi->reset_cpu_mask) return; thi->reset_cpu_mask = 0; - set_cpus_allowed_ptr(p, mdev->cpu_mask); + set_cpus_allowed_ptr(p, thi->tconn->cpu_mask); } #endif -/* the appropriate socket mutex must be held already */ -int _drbd_send_cmd(struct drbd_conf *mdev, struct socket *sock, - enum drbd_packets cmd, struct p_header80 *h, - size_t size, unsigned msg_flags) +/** + * drbd_header_size - size of a packet header + * + * The header size is a multiple of 8, so any payload following the header is + * word aligned on 64-bit architectures. (The bitmap send and receive code + * relies on this.) + */ +unsigned int drbd_header_size(struct drbd_tconn *tconn) { - int sent, ok; + if (tconn->agreed_pro_version >= 100) { + BUILD_BUG_ON(!IS_ALIGNED(sizeof(struct p_header100), 8)); + return sizeof(struct p_header100); + } else { + BUILD_BUG_ON(sizeof(struct p_header80) != + sizeof(struct p_header95)); + BUILD_BUG_ON(!IS_ALIGNED(sizeof(struct p_header80), 8)); + return sizeof(struct p_header80); + } +} - ERR_IF(!h) return false; - ERR_IF(!size) return false; +static unsigned int prepare_header80(struct p_header80 *h, enum drbd_packet cmd, int size) +{ + h->magic = cpu_to_be32(DRBD_MAGIC); + h->command = cpu_to_be16(cmd); + h->length = cpu_to_be16(size); + return sizeof(struct p_header80); +} - h->magic = BE_DRBD_MAGIC; +static unsigned int prepare_header95(struct p_header95 *h, enum drbd_packet cmd, int size) +{ + h->magic = cpu_to_be16(DRBD_MAGIC_BIG); h->command = cpu_to_be16(cmd); - h->length = cpu_to_be16(size-sizeof(struct p_header80)); + h->length = cpu_to_be32(size); + return sizeof(struct p_header95); +} - sent = drbd_send(mdev, sock, h, size, msg_flags); +static unsigned int prepare_header100(struct p_header100 *h, enum drbd_packet cmd, + int size, int vnr) +{ + h->magic = cpu_to_be32(DRBD_MAGIC_100); + h->volume = cpu_to_be16(vnr); + h->command = cpu_to_be16(cmd); + h->length = cpu_to_be32(size); + h->pad = 0; + return sizeof(struct p_header100); +} - ok = (sent == size); - if (!ok && !signal_pending(current)) - dev_warn(DEV, "short sent %s size=%d sent=%d\n", - cmdname(cmd), (int)size, sent); - return ok; +static unsigned int prepare_header(struct drbd_tconn *tconn, int vnr, + void *buffer, enum drbd_packet cmd, int size) +{ + if (tconn->agreed_pro_version >= 100) + return prepare_header100(buffer, cmd, size, vnr); + else if (tconn->agreed_pro_version >= 95 && + size > DRBD_MAX_SIZE_H80_PACKET) + return prepare_header95(buffer, cmd, size); + else + return prepare_header80(buffer, cmd, size); } -/* don't pass the socket. we may only look at it - * when we hold the appropriate socket mutex. - */ -int drbd_send_cmd(struct drbd_conf *mdev, int use_data_socket, - enum drbd_packets cmd, struct p_header80 *h, size_t size) +static void *__conn_prepare_command(struct drbd_tconn *tconn, + struct drbd_socket *sock) { - int ok = 0; - struct socket *sock; + if (!sock->socket) + return NULL; + return sock->sbuf + drbd_header_size(tconn); +} - if (use_data_socket) { - mutex_lock(&mdev->data.mutex); - sock = mdev->data.socket; - } else { - mutex_lock(&mdev->meta.mutex); - sock = mdev->meta.socket; - } +void *conn_prepare_command(struct drbd_tconn *tconn, struct drbd_socket *sock) +{ + void *p; - /* drbd_disconnect() could have called drbd_free_sock() - * while we were waiting in down()... */ - if (likely(sock != NULL)) - ok = _drbd_send_cmd(mdev, sock, cmd, h, size, 0); + mutex_lock(&sock->mutex); + p = __conn_prepare_command(tconn, sock); + if (!p) + mutex_unlock(&sock->mutex); - if (use_data_socket) - mutex_unlock(&mdev->data.mutex); - else - mutex_unlock(&mdev->meta.mutex); - return ok; + return p; } -int drbd_send_cmd2(struct drbd_conf *mdev, enum drbd_packets cmd, char *data, - size_t size) +void *drbd_prepare_command(struct drbd_conf *mdev, struct drbd_socket *sock) { - struct p_header80 h; - int ok; + return conn_prepare_command(mdev->tconn, sock); +} - h.magic = BE_DRBD_MAGIC; - h.command = cpu_to_be16(cmd); - h.length = cpu_to_be16(size); +static int __send_command(struct drbd_tconn *tconn, int vnr, + struct drbd_socket *sock, enum drbd_packet cmd, + unsigned int header_size, void *data, + unsigned int size) +{ + int msg_flags; + int err; - if (!drbd_get_data_sock(mdev)) - return 0; + /* + * Called with @data == NULL and the size of the data blocks in @size + * for commands that send data blocks. For those commands, omit the + * MSG_MORE flag: this will increase the likelihood that data blocks + * which are page aligned on the sender will end up page aligned on the + * receiver. + */ + msg_flags = data ? MSG_MORE : 0; + + header_size += prepare_header(tconn, vnr, sock->sbuf, cmd, + header_size + size); + err = drbd_send_all(tconn, sock->socket, sock->sbuf, header_size, + msg_flags); + if (data && !err) + err = drbd_send_all(tconn, sock->socket, data, size, 0); + return err; +} - ok = (sizeof(h) == - drbd_send(mdev, mdev->data.socket, &h, sizeof(h), 0)); - ok = ok && (size == - drbd_send(mdev, mdev->data.socket, data, size, 0)); +static int __conn_send_command(struct drbd_tconn *tconn, struct drbd_socket *sock, + enum drbd_packet cmd, unsigned int header_size, + void *data, unsigned int size) +{ + return __send_command(tconn, 0, sock, cmd, header_size, data, size); +} - drbd_put_data_sock(mdev); +int conn_send_command(struct drbd_tconn *tconn, struct drbd_socket *sock, + enum drbd_packet cmd, unsigned int header_size, + void *data, unsigned int size) +{ + int err; - return ok; + err = __conn_send_command(tconn, sock, cmd, header_size, data, size); + mutex_unlock(&sock->mutex); + return err; } -int drbd_send_sync_param(struct drbd_conf *mdev, struct syncer_conf *sc) +int drbd_send_command(struct drbd_conf *mdev, struct drbd_socket *sock, + enum drbd_packet cmd, unsigned int header_size, + void *data, unsigned int size) { + int err; + + err = __send_command(mdev->tconn, mdev->vnr, sock, cmd, header_size, + data, size); + mutex_unlock(&sock->mutex); + return err; +} + +int drbd_send_ping(struct drbd_tconn *tconn) +{ + struct drbd_socket *sock; + + sock = &tconn->meta; + if (!conn_prepare_command(tconn, sock)) + return -EIO; + return conn_send_command(tconn, sock, P_PING, 0, NULL, 0); +} + +int drbd_send_ping_ack(struct drbd_tconn *tconn) +{ + struct drbd_socket *sock; + + sock = &tconn->meta; + if (!conn_prepare_command(tconn, sock)) + return -EIO; + return conn_send_command(tconn, sock, P_PING_ACK, 0, NULL, 0); +} + +int drbd_send_sync_param(struct drbd_conf *mdev) +{ + struct drbd_socket *sock; struct p_rs_param_95 *p; - struct socket *sock; - int size, rv; - const int apv = mdev->agreed_pro_version; + int size; + const int apv = mdev->tconn->agreed_pro_version; + enum drbd_packet cmd; + struct net_conf *nc; + struct disk_conf *dc; + + sock = &mdev->tconn->data; + p = drbd_prepare_command(mdev, sock); + if (!p) + return -EIO; + + rcu_read_lock(); + nc = rcu_dereference(mdev->tconn->net_conf); size = apv <= 87 ? sizeof(struct p_rs_param) : apv == 88 ? sizeof(struct p_rs_param) - + strlen(mdev->sync_conf.verify_alg) + 1 + + strlen(nc->verify_alg) + 1 : apv <= 94 ? sizeof(struct p_rs_param_89) : /* apv >= 95 */ sizeof(struct p_rs_param_95); - /* used from admin command context and receiver/worker context. - * to avoid kmalloc, grab the socket right here, - * then use the pre-allocated sbuf there */ - mutex_lock(&mdev->data.mutex); - sock = mdev->data.socket; - - if (likely(sock != NULL)) { - enum drbd_packets cmd = apv >= 89 ? P_SYNC_PARAM89 : P_SYNC_PARAM; - - p = &mdev->data.sbuf.rs_param_95; + cmd = apv >= 89 ? P_SYNC_PARAM89 : P_SYNC_PARAM; - /* initialize verify_alg and csums_alg */ - memset(p->verify_alg, 0, 2 * SHARED_SECRET_MAX); + /* initialize verify_alg and csums_alg */ + memset(p->verify_alg, 0, 2 * SHARED_SECRET_MAX); - p->rate = cpu_to_be32(sc->rate); - p->c_plan_ahead = cpu_to_be32(sc->c_plan_ahead); - p->c_delay_target = cpu_to_be32(sc->c_delay_target); - p->c_fill_target = cpu_to_be32(sc->c_fill_target); - p->c_max_rate = cpu_to_be32(sc->c_max_rate); - - if (apv >= 88) - strcpy(p->verify_alg, mdev->sync_conf.verify_alg); - if (apv >= 89) - strcpy(p->csums_alg, mdev->sync_conf.csums_alg); - - rv = _drbd_send_cmd(mdev, sock, cmd, &p->head, size, 0); - } else - rv = 0; /* not ok */ + if (get_ldev(mdev)) { + dc = rcu_dereference(mdev->ldev->disk_conf); + p->resync_rate = cpu_to_be32(dc->resync_rate); + p->c_plan_ahead = cpu_to_be32(dc->c_plan_ahead); + p->c_delay_target = cpu_to_be32(dc->c_delay_target); + p->c_fill_target = cpu_to_be32(dc->c_fill_target); + p->c_max_rate = cpu_to_be32(dc->c_max_rate); + put_ldev(mdev); + } else { + p->resync_rate = cpu_to_be32(DRBD_RESYNC_RATE_DEF); + p->c_plan_ahead = cpu_to_be32(DRBD_C_PLAN_AHEAD_DEF); + p->c_delay_target = cpu_to_be32(DRBD_C_DELAY_TARGET_DEF); + p->c_fill_target = cpu_to_be32(DRBD_C_FILL_TARGET_DEF); + p->c_max_rate = cpu_to_be32(DRBD_C_MAX_RATE_DEF); + } - mutex_unlock(&mdev->data.mutex); + if (apv >= 88) + strcpy(p->verify_alg, nc->verify_alg); + if (apv >= 89) + strcpy(p->csums_alg, nc->csums_alg); + rcu_read_unlock(); - return rv; + return drbd_send_command(mdev, sock, cmd, size, NULL, 0); } -int drbd_send_protocol(struct drbd_conf *mdev) +int __drbd_send_protocol(struct drbd_tconn *tconn, enum drbd_packet cmd) { + struct drbd_socket *sock; struct p_protocol *p; - int size, cf, rv; + struct net_conf *nc; + int size, cf; - size = sizeof(struct p_protocol); + sock = &tconn->data; + p = __conn_prepare_command(tconn, sock); + if (!p) + return -EIO; - if (mdev->agreed_pro_version >= 87) - size += strlen(mdev->net_conf->integrity_alg) + 1; + rcu_read_lock(); + nc = rcu_dereference(tconn->net_conf); - /* we must not recurse into our own queue, - * as that is blocked during handshake */ - p = kmalloc(size, GFP_NOIO); - if (p == NULL) - return 0; + if (nc->tentative && tconn->agreed_pro_version < 92) { + rcu_read_unlock(); + mutex_unlock(&sock->mutex); + conn_err(tconn, "--dry-run is not supported by peer"); + return -EOPNOTSUPP; + } - p->protocol = cpu_to_be32(mdev->net_conf->wire_protocol); - p->after_sb_0p = cpu_to_be32(mdev->net_conf->after_sb_0p); - p->after_sb_1p = cpu_to_be32(mdev->net_conf->after_sb_1p); - p->after_sb_2p = cpu_to_be32(mdev->net_conf->after_sb_2p); - p->two_primaries = cpu_to_be32(mdev->net_conf->two_primaries); + size = sizeof(*p); + if (tconn->agreed_pro_version >= 87) + size += strlen(nc->integrity_alg) + 1; + p->protocol = cpu_to_be32(nc->wire_protocol); + p->after_sb_0p = cpu_to_be32(nc->after_sb_0p); + p->after_sb_1p = cpu_to_be32(nc->after_sb_1p); + p->after_sb_2p = cpu_to_be32(nc->after_sb_2p); + p->two_primaries = cpu_to_be32(nc->two_primaries); cf = 0; - if (mdev->net_conf->want_lose) - cf |= CF_WANT_LOSE; - if (mdev->net_conf->dry_run) { - if (mdev->agreed_pro_version >= 92) - cf |= CF_DRY_RUN; - else { - dev_err(DEV, "--dry-run is not supported by peer"); - kfree(p); - return -1; - } - } + if (nc->discard_my_data) + cf |= CF_DISCARD_MY_DATA; + if (nc->tentative) + cf |= CF_DRY_RUN; p->conn_flags = cpu_to_be32(cf); - if (mdev->agreed_pro_version >= 87) - strcpy(p->integrity_alg, mdev->net_conf->integrity_alg); + if (tconn->agreed_pro_version >= 87) + strcpy(p->integrity_alg, nc->integrity_alg); + rcu_read_unlock(); - rv = drbd_send_cmd(mdev, USE_DATA_SOCKET, P_PROTOCOL, - (struct p_header80 *)p, size); - kfree(p); - return rv; + return __conn_send_command(tconn, sock, cmd, size, NULL, 0); +} + +int drbd_send_protocol(struct drbd_tconn *tconn) +{ + int err; + + mutex_lock(&tconn->data.mutex); + err = __drbd_send_protocol(tconn, P_PROTOCOL); + mutex_unlock(&tconn->data.mutex); + + return err; } int _drbd_send_uuids(struct drbd_conf *mdev, u64 uuid_flags) { - struct p_uuids p; + struct drbd_socket *sock; + struct p_uuids *p; int i; if (!get_ldev_if_state(mdev, D_NEGOTIATING)) - return 1; + return 0; + sock = &mdev->tconn->data; + p = drbd_prepare_command(mdev, sock); + if (!p) { + put_ldev(mdev); + return -EIO; + } + spin_lock_irq(&mdev->ldev->md.uuid_lock); for (i = UI_CURRENT; i < UI_SIZE; i++) - p.uuid[i] = mdev->ldev ? cpu_to_be64(mdev->ldev->md.uuid[i]) : 0; + p->uuid[i] = cpu_to_be64(mdev->ldev->md.uuid[i]); + spin_unlock_irq(&mdev->ldev->md.uuid_lock); mdev->comm_bm_set = drbd_bm_total_weight(mdev); - p.uuid[UI_SIZE] = cpu_to_be64(mdev->comm_bm_set); - uuid_flags |= mdev->net_conf->want_lose ? 1 : 0; + p->uuid[UI_SIZE] = cpu_to_be64(mdev->comm_bm_set); + rcu_read_lock(); + uuid_flags |= rcu_dereference(mdev->tconn->net_conf)->discard_my_data ? 1 : 0; + rcu_read_unlock(); uuid_flags |= test_bit(CRASHED_PRIMARY, &mdev->flags) ? 2 : 0; uuid_flags |= mdev->new_state_tmp.disk == D_INCONSISTENT ? 4 : 0; - p.uuid[UI_FLAGS] = cpu_to_be64(uuid_flags); + p->uuid[UI_FLAGS] = cpu_to_be64(uuid_flags); put_ldev(mdev); - - return drbd_send_cmd(mdev, USE_DATA_SOCKET, P_UUIDS, - (struct p_header80 *)&p, sizeof(p)); + return drbd_send_command(mdev, sock, P_UUIDS, sizeof(*p), NULL, 0); } int drbd_send_uuids(struct drbd_conf *mdev) @@ -2169,9 +884,10 @@ void drbd_print_uuids(struct drbd_conf *mdev, const char *text) } } -int drbd_gen_and_send_sync_uuid(struct drbd_conf *mdev) +void drbd_gen_and_send_sync_uuid(struct drbd_conf *mdev) { - struct p_rs_uuid p; + struct drbd_socket *sock; + struct p_rs_uuid *p; u64 uuid; D_ASSERT(mdev->state.disk == D_UP_TO_DATE); @@ -2184,24 +900,29 @@ int drbd_gen_and_send_sync_uuid(struct drbd_conf *mdev) drbd_uuid_set(mdev, UI_BITMAP, uuid); drbd_print_uuids(mdev, "updated sync UUID"); drbd_md_sync(mdev); - p.uuid = cpu_to_be64(uuid); - return drbd_send_cmd(mdev, USE_DATA_SOCKET, P_SYNC_UUID, - (struct p_header80 *)&p, sizeof(p)); + sock = &mdev->tconn->data; + p = drbd_prepare_command(mdev, sock); + if (p) { + p->uuid = cpu_to_be64(uuid); + drbd_send_command(mdev, sock, P_SYNC_UUID, sizeof(*p), NULL, 0); + } } int drbd_send_sizes(struct drbd_conf *mdev, int trigger_reply, enum dds_flags flags) { - struct p_sizes p; + struct drbd_socket *sock; + struct p_sizes *p; sector_t d_size, u_size; int q_order_type; unsigned int max_bio_size; - int ok; if (get_ldev_if_state(mdev, D_NEGOTIATING)) { D_ASSERT(mdev->ldev->backing_bdev); d_size = drbd_get_max_capacity(mdev->ldev); - u_size = mdev->ldev->dc.disk_size; + rcu_read_lock(); + u_size = rcu_dereference(mdev->ldev->disk_conf)->disk_size; + rcu_read_unlock(); q_order_type = drbd_queue_order_type(mdev); max_bio_size = queue_max_hw_sectors(mdev->ldev->backing_bdev->bd_disk->queue) << 9; max_bio_size = min(max_bio_size, DRBD_MAX_BIO_SIZE); @@ -2213,20 +934,23 @@ int drbd_send_sizes(struct drbd_conf *mdev, int trigger_reply, enum dds_flags fl max_bio_size = DRBD_MAX_BIO_SIZE; /* ... multiple BIOs per peer_request */ } - /* Never allow old drbd (up to 8.3.7) to see more than 32KiB */ - if (mdev->agreed_pro_version <= 94) - max_bio_size = min(max_bio_size, DRBD_MAX_SIZE_H80_PACKET); + sock = &mdev->tconn->data; + p = drbd_prepare_command(mdev, sock); + if (!p) + return -EIO; - p.d_size = cpu_to_be64(d_size); - p.u_size = cpu_to_be64(u_size); - p.c_size = cpu_to_be64(trigger_reply ? 0 : drbd_get_capacity(mdev->this_bdev)); - p.max_bio_size = cpu_to_be32(max_bio_size); - p.queue_order_type = cpu_to_be16(q_order_type); - p.dds_flags = cpu_to_be16(flags); + if (mdev->tconn->agreed_pro_version <= 94) + max_bio_size = min(max_bio_size, DRBD_MAX_SIZE_H80_PACKET); + else if (mdev->tconn->agreed_pro_version < 100) + max_bio_size = min(max_bio_size, DRBD_MAX_BIO_SIZE_P95); - ok = drbd_send_cmd(mdev, USE_DATA_SOCKET, P_SIZES, - (struct p_header80 *)&p, sizeof(p)); - return ok; + p->d_size = cpu_to_be64(d_size); + p->u_size = cpu_to_be64(u_size); + p->c_size = cpu_to_be64(trigger_reply ? 0 : drbd_get_capacity(mdev->this_bdev)); + p->max_bio_size = cpu_to_be32(max_bio_size); + p->queue_order_type = cpu_to_be16(q_order_type); + p->dds_flags = cpu_to_be16(flags); + return drbd_send_command(mdev, sock, P_SIZES, sizeof(*p), NULL, 0); } /** @@ -2235,34 +959,21 @@ int drbd_send_sizes(struct drbd_conf *mdev, int trigger_reply, enum dds_flags fl */ int drbd_send_current_state(struct drbd_conf *mdev) { - struct socket *sock; - struct p_state p; - int ok = 0; - - /* Grab state lock so we wont send state if we're in the middle - * of a cluster wide state change on another thread */ - drbd_state_lock(mdev); - - mutex_lock(&mdev->data.mutex); - - p.state = cpu_to_be32(mdev->state.i); /* Within the send mutex */ - sock = mdev->data.socket; + struct drbd_socket *sock; + struct p_state *p; - if (likely(sock != NULL)) { - ok = _drbd_send_cmd(mdev, sock, P_STATE, - (struct p_header80 *)&p, sizeof(p), 0); - } - - mutex_unlock(&mdev->data.mutex); - - drbd_state_unlock(mdev); - return ok; + sock = &mdev->tconn->data; + p = drbd_prepare_command(mdev, sock); + if (!p) + return -EIO; + p->state = cpu_to_be32(mdev->state.i); /* Within the send mutex */ + return drbd_send_command(mdev, sock, P_STATE, sizeof(*p), NULL, 0); } /** * drbd_send_state() - After a state change, sends the new state to the peer - * @mdev: DRBD device. - * @state: the state to send, not necessarily the current state. + * @mdev: DRBD device. + * @state: the state to send, not necessarily the current state. * * Each state change queues an "after_state_ch" work, which will eventually * send the resulting new state to the peer. If more state changes happen @@ -2271,50 +982,95 @@ int drbd_send_current_state(struct drbd_conf *mdev) */ int drbd_send_state(struct drbd_conf *mdev, union drbd_state state) { - struct socket *sock; - struct p_state p; - int ok = 0; + struct drbd_socket *sock; + struct p_state *p; - mutex_lock(&mdev->data.mutex); + sock = &mdev->tconn->data; + p = drbd_prepare_command(mdev, sock); + if (!p) + return -EIO; + p->state = cpu_to_be32(state.i); /* Within the send mutex */ + return drbd_send_command(mdev, sock, P_STATE, sizeof(*p), NULL, 0); +} - p.state = cpu_to_be32(state.i); - sock = mdev->data.socket; +int drbd_send_state_req(struct drbd_conf *mdev, union drbd_state mask, union drbd_state val) +{ + struct drbd_socket *sock; + struct p_req_state *p; - if (likely(sock != NULL)) { - ok = _drbd_send_cmd(mdev, sock, P_STATE, - (struct p_header80 *)&p, sizeof(p), 0); - } + sock = &mdev->tconn->data; + p = drbd_prepare_command(mdev, sock); + if (!p) + return -EIO; + p->mask = cpu_to_be32(mask.i); + p->val = cpu_to_be32(val.i); + return drbd_send_command(mdev, sock, P_STATE_CHG_REQ, sizeof(*p), NULL, 0); +} - mutex_unlock(&mdev->data.mutex); +int conn_send_state_req(struct drbd_tconn *tconn, union drbd_state mask, union drbd_state val) +{ + enum drbd_packet cmd; + struct drbd_socket *sock; + struct p_req_state *p; - return ok; + cmd = tconn->agreed_pro_version < 100 ? P_STATE_CHG_REQ : P_CONN_ST_CHG_REQ; + sock = &tconn->data; + p = conn_prepare_command(tconn, sock); + if (!p) + return -EIO; + p->mask = cpu_to_be32(mask.i); + p->val = cpu_to_be32(val.i); + return conn_send_command(tconn, sock, cmd, sizeof(*p), NULL, 0); } -int drbd_send_state_req(struct drbd_conf *mdev, - union drbd_state mask, union drbd_state val) +void drbd_send_sr_reply(struct drbd_conf *mdev, enum drbd_state_rv retcode) { - struct p_req_state p; + struct drbd_socket *sock; + struct p_req_state_reply *p; - p.mask = cpu_to_be32(mask.i); - p.val = cpu_to_be32(val.i); + sock = &mdev->tconn->meta; + p = drbd_prepare_command(mdev, sock); + if (p) { + p->retcode = cpu_to_be32(retcode); + drbd_send_command(mdev, sock, P_STATE_CHG_REPLY, sizeof(*p), NULL, 0); + } +} + +void conn_send_sr_reply(struct drbd_tconn *tconn, enum drbd_state_rv retcode) +{ + struct drbd_socket *sock; + struct p_req_state_reply *p; + enum drbd_packet cmd = tconn->agreed_pro_version < 100 ? P_STATE_CHG_REPLY : P_CONN_ST_CHG_REPLY; - return drbd_send_cmd(mdev, USE_DATA_SOCKET, P_STATE_CHG_REQ, - (struct p_header80 *)&p, sizeof(p)); + sock = &tconn->meta; + p = conn_prepare_command(tconn, sock); + if (p) { + p->retcode = cpu_to_be32(retcode); + conn_send_command(tconn, sock, cmd, sizeof(*p), NULL, 0); + } } -int drbd_send_sr_reply(struct drbd_conf *mdev, enum drbd_state_rv retcode) +static void dcbp_set_code(struct p_compressed_bm *p, enum drbd_bitmap_code code) { - struct p_req_state_reply p; + BUG_ON(code & ~0xf); + p->encoding = (p->encoding & ~0xf) | code; +} - p.retcode = cpu_to_be32(retcode); +static void dcbp_set_start(struct p_compressed_bm *p, int set) +{ + p->encoding = (p->encoding & ~0x80) | (set ? 0x80 : 0); +} - return drbd_send_cmd(mdev, USE_META_SOCKET, P_STATE_CHG_REPLY, - (struct p_header80 *)&p, sizeof(p)); +static void dcbp_set_pad_bits(struct p_compressed_bm *p, int n) +{ + BUG_ON(n & ~0x7); + p->encoding = (p->encoding & (~0x7 << 4)) | (n << 4); } int fill_bitmap_rle_bits(struct drbd_conf *mdev, - struct p_compressed_bm *p, - struct bm_xfer_ctx *c) + struct p_compressed_bm *p, + unsigned int size, + struct bm_xfer_ctx *c) { struct bitstream bs; unsigned long plain_bits; @@ -2322,19 +1078,21 @@ int fill_bitmap_rle_bits(struct drbd_conf *mdev, unsigned long rl; unsigned len; unsigned toggle; - int bits; + int bits, use_rle; /* may we use this feature? */ - if ((mdev->sync_conf.use_rle == 0) || - (mdev->agreed_pro_version < 90)) - return 0; + rcu_read_lock(); + use_rle = rcu_dereference(mdev->tconn->net_conf)->use_rle; + rcu_read_unlock(); + if (!use_rle || mdev->tconn->agreed_pro_version < 90) + return 0; if (c->bit_offset >= c->bm_bits) return 0; /* nothing to do. */ /* use at most thus many bytes */ - bitstream_init(&bs, p->code, BM_PACKET_VLI_BYTES_MAX, 0); - memset(p->code, 0, BM_PACKET_VLI_BYTES_MAX); + bitstream_init(&bs, p->code, size, 0); + memset(p->code, 0, size); /* plain bits covered in this code string */ plain_bits = 0; @@ -2356,12 +1114,12 @@ int fill_bitmap_rle_bits(struct drbd_conf *mdev, if (rl == 0) { /* the first checked bit was set, * store start value, */ - DCBP_set_start(p, 1); + dcbp_set_start(p, 1); /* but skip encoding of zero run length */ toggle = !toggle; continue; } - DCBP_set_start(p, 0); + dcbp_set_start(p, 0); } /* paranoia: catch zero runlength. @@ -2401,7 +1159,7 @@ int fill_bitmap_rle_bits(struct drbd_conf *mdev, bm_xfer_ctx_bit_to_word_offset(c); /* store pad_bits */ - DCBP_set_pad_bits(p, (8 - bs.cur.bit) & 0x7); + dcbp_set_pad_bits(p, (8 - bs.cur.bit) & 0x7); return len; } @@ -2413,48 +1171,52 @@ int fill_bitmap_rle_bits(struct drbd_conf *mdev, * code upon failure. */ static int -send_bitmap_rle_or_plain(struct drbd_conf *mdev, - struct p_header80 *h, struct bm_xfer_ctx *c) +send_bitmap_rle_or_plain(struct drbd_conf *mdev, struct bm_xfer_ctx *c) { - struct p_compressed_bm *p = (void*)h; - unsigned long num_words; - int len; - int ok; - - len = fill_bitmap_rle_bits(mdev, p, c); + struct drbd_socket *sock = &mdev->tconn->data; + unsigned int header_size = drbd_header_size(mdev->tconn); + struct p_compressed_bm *p = sock->sbuf + header_size; + int len, err; + len = fill_bitmap_rle_bits(mdev, p, + DRBD_SOCKET_BUFFER_SIZE - header_size - sizeof(*p), c); if (len < 0) return -EIO; if (len) { - DCBP_set_code(p, RLE_VLI_Bits); - ok = _drbd_send_cmd(mdev, mdev->data.socket, P_COMPRESSED_BITMAP, h, - sizeof(*p) + len, 0); - + dcbp_set_code(p, RLE_VLI_Bits); + err = __send_command(mdev->tconn, mdev->vnr, sock, + P_COMPRESSED_BITMAP, sizeof(*p) + len, + NULL, 0); c->packets[0]++; - c->bytes[0] += sizeof(*p) + len; + c->bytes[0] += header_size + sizeof(*p) + len; if (c->bit_offset >= c->bm_bits) len = 0; /* DONE */ } else { /* was not compressible. * send a buffer full of plain text bits instead. */ - num_words = min_t(size_t, BM_PACKET_WORDS, c->bm_words - c->word_offset); - len = num_words * sizeof(long); + unsigned int data_size; + unsigned long num_words; + unsigned long *p = sock->sbuf + header_size; + + data_size = DRBD_SOCKET_BUFFER_SIZE - header_size; + num_words = min_t(size_t, data_size / sizeof(*p), + c->bm_words - c->word_offset); + len = num_words * sizeof(*p); if (len) - drbd_bm_get_lel(mdev, c->word_offset, num_words, (unsigned long*)h->payload); - ok = _drbd_send_cmd(mdev, mdev->data.socket, P_BITMAP, - h, sizeof(struct p_header80) + len, 0); + drbd_bm_get_lel(mdev, c->word_offset, num_words, p); + err = __send_command(mdev->tconn, mdev->vnr, sock, P_BITMAP, len, NULL, 0); c->word_offset += num_words; c->bit_offset = c->word_offset * BITS_PER_LONG; c->packets[1]++; - c->bytes[1] += sizeof(struct p_header80) + len; + c->bytes[1] += header_size + len; if (c->bit_offset > c->bm_bits) c->bit_offset = c->bm_bits; } - if (ok) { + if (!err) { if (len == 0) { INFO_bm_xfer_stats(mdev, "send", c); return 0; @@ -2465,21 +1227,13 @@ send_bitmap_rle_or_plain(struct drbd_conf *mdev, } /* See the comment at receive_bitmap() */ -int _drbd_send_bitmap(struct drbd_conf *mdev) +static int _drbd_send_bitmap(struct drbd_conf *mdev) { struct bm_xfer_ctx c; - struct p_header80 *p; int err; - ERR_IF(!mdev->bitmap) return false; - - /* maybe we should use some per thread scratch page, - * and allocate that during initial device creation? */ - p = (struct p_header80 *) __get_free_page(GFP_NOIO); - if (!p) { - dev_err(DEV, "failed to allocate one page buffer in %s\n", __func__); + if (!expect(mdev->bitmap)) return false; - } if (get_ldev(mdev)) { if (drbd_md_test_flag(mdev->ldev, MDF_FULL_SYNC)) { @@ -2504,37 +1258,39 @@ int _drbd_send_bitmap(struct drbd_conf *mdev) }; do { - err = send_bitmap_rle_or_plain(mdev, p, &c); + err = send_bitmap_rle_or_plain(mdev, &c); } while (err > 0); - free_page((unsigned long) p); return err == 0; } int drbd_send_bitmap(struct drbd_conf *mdev) { - int err; + struct drbd_socket *sock = &mdev->tconn->data; + int err = -1; - if (!drbd_get_data_sock(mdev)) - return -1; - err = !_drbd_send_bitmap(mdev); - drbd_put_data_sock(mdev); + mutex_lock(&sock->mutex); + if (sock->socket) + err = !_drbd_send_bitmap(mdev); + mutex_unlock(&sock->mutex); return err; } -int drbd_send_b_ack(struct drbd_conf *mdev, u32 barrier_nr, u32 set_size) +void drbd_send_b_ack(struct drbd_tconn *tconn, u32 barrier_nr, u32 set_size) { - int ok; - struct p_barrier_ack p; + struct drbd_socket *sock; + struct p_barrier_ack *p; - p.barrier = barrier_nr; - p.set_size = cpu_to_be32(set_size); + if (tconn->cstate < C_WF_REPORT_PARAMS) + return; - if (mdev->state.conn < C_CONNECTED) - return false; - ok = drbd_send_cmd(mdev, USE_META_SOCKET, P_BARRIER_ACK, - (struct p_header80 *)&p, sizeof(p)); - return ok; + sock = &tconn->meta; + p = conn_prepare_command(tconn, sock); + if (!p) + return; + p->barrier = barrier_nr; + p->set_size = cpu_to_be32(set_size); + conn_send_command(tconn, sock, P_BARRIER_ACK, sizeof(*p), NULL, 0); } /** @@ -2545,62 +1301,62 @@ int drbd_send_b_ack(struct drbd_conf *mdev, u32 barrier_nr, u32 set_size) * @blksize: size in byte, needs to be in big endian byte order * @block_id: Id, big endian byte order */ -static int _drbd_send_ack(struct drbd_conf *mdev, enum drbd_packets cmd, - u64 sector, - u32 blksize, - u64 block_id) +static int _drbd_send_ack(struct drbd_conf *mdev, enum drbd_packet cmd, + u64 sector, u32 blksize, u64 block_id) { - int ok; - struct p_block_ack p; + struct drbd_socket *sock; + struct p_block_ack *p; - p.sector = sector; - p.block_id = block_id; - p.blksize = blksize; - p.seq_num = cpu_to_be32(atomic_add_return(1, &mdev->packet_seq)); + if (mdev->state.conn < C_CONNECTED) + return -EIO; - if (!mdev->meta.socket || mdev->state.conn < C_CONNECTED) - return false; - ok = drbd_send_cmd(mdev, USE_META_SOCKET, cmd, - (struct p_header80 *)&p, sizeof(p)); - return ok; + sock = &mdev->tconn->meta; + p = drbd_prepare_command(mdev, sock); + if (!p) + return -EIO; + p->sector = sector; + p->block_id = block_id; + p->blksize = blksize; + p->seq_num = cpu_to_be32(atomic_inc_return(&mdev->packet_seq)); + return drbd_send_command(mdev, sock, cmd, sizeof(*p), NULL, 0); } /* dp->sector and dp->block_id already/still in network byte order, * data_size is payload size according to dp->head, * and may need to be corrected for digest size. */ -int drbd_send_ack_dp(struct drbd_conf *mdev, enum drbd_packets cmd, - struct p_data *dp, int data_size) +void drbd_send_ack_dp(struct drbd_conf *mdev, enum drbd_packet cmd, + struct p_data *dp, int data_size) { - data_size -= (mdev->agreed_pro_version >= 87 && mdev->integrity_r_tfm) ? - crypto_hash_digestsize(mdev->integrity_r_tfm) : 0; - return _drbd_send_ack(mdev, cmd, dp->sector, cpu_to_be32(data_size), - dp->block_id); + if (mdev->tconn->peer_integrity_tfm) + data_size -= crypto_hash_digestsize(mdev->tconn->peer_integrity_tfm); + _drbd_send_ack(mdev, cmd, dp->sector, cpu_to_be32(data_size), + dp->block_id); } -int drbd_send_ack_rp(struct drbd_conf *mdev, enum drbd_packets cmd, - struct p_block_req *rp) +void drbd_send_ack_rp(struct drbd_conf *mdev, enum drbd_packet cmd, + struct p_block_req *rp) { - return _drbd_send_ack(mdev, cmd, rp->sector, rp->blksize, rp->block_id); + _drbd_send_ack(mdev, cmd, rp->sector, rp->blksize, rp->block_id); } /** * drbd_send_ack() - Sends an ack packet - * @mdev: DRBD device. - * @cmd: Packet command code. - * @e: Epoch entry. + * @mdev: DRBD device + * @cmd: packet command code + * @peer_req: peer request */ -int drbd_send_ack(struct drbd_conf *mdev, - enum drbd_packets cmd, struct drbd_epoch_entry *e) +int drbd_send_ack(struct drbd_conf *mdev, enum drbd_packet cmd, + struct drbd_peer_request *peer_req) { return _drbd_send_ack(mdev, cmd, - cpu_to_be64(e->sector), - cpu_to_be32(e->size), - e->block_id); + cpu_to_be64(peer_req->i.sector), + cpu_to_be32(peer_req->i.size), + peer_req->block_id); } /* This function misuses the block_id field to signal if the blocks * are is sync or not. */ -int drbd_send_ack_ex(struct drbd_conf *mdev, enum drbd_packets cmd, +int drbd_send_ack_ex(struct drbd_conf *mdev, enum drbd_packet cmd, sector_t sector, int blksize, u64 block_id) { return _drbd_send_ack(mdev, cmd, @@ -2612,85 +1368,87 @@ int drbd_send_ack_ex(struct drbd_conf *mdev, enum drbd_packets cmd, int drbd_send_drequest(struct drbd_conf *mdev, int cmd, sector_t sector, int size, u64 block_id) { - int ok; - struct p_block_req p; - - p.sector = cpu_to_be64(sector); - p.block_id = block_id; - p.blksize = cpu_to_be32(size); + struct drbd_socket *sock; + struct p_block_req *p; - ok = drbd_send_cmd(mdev, USE_DATA_SOCKET, cmd, - (struct p_header80 *)&p, sizeof(p)); - return ok; + sock = &mdev->tconn->data; + p = drbd_prepare_command(mdev, sock); + if (!p) + return -EIO; + p->sector = cpu_to_be64(sector); + p->block_id = block_id; + p->blksize = cpu_to_be32(size); + return drbd_send_command(mdev, sock, cmd, sizeof(*p), NULL, 0); } -int drbd_send_drequest_csum(struct drbd_conf *mdev, - sector_t sector, int size, - void *digest, int digest_size, - enum drbd_packets cmd) +int drbd_send_drequest_csum(struct drbd_conf *mdev, sector_t sector, int size, + void *digest, int digest_size, enum drbd_packet cmd) { - int ok; - struct p_block_req p; - - p.sector = cpu_to_be64(sector); - p.block_id = BE_DRBD_MAGIC + 0xbeef; - p.blksize = cpu_to_be32(size); - - p.head.magic = BE_DRBD_MAGIC; - p.head.command = cpu_to_be16(cmd); - p.head.length = cpu_to_be16(sizeof(p) - sizeof(struct p_header80) + digest_size); + struct drbd_socket *sock; + struct p_block_req *p; - mutex_lock(&mdev->data.mutex); + /* FIXME: Put the digest into the preallocated socket buffer. */ - ok = (sizeof(p) == drbd_send(mdev, mdev->data.socket, &p, sizeof(p), 0)); - ok = ok && (digest_size == drbd_send(mdev, mdev->data.socket, digest, digest_size, 0)); - - mutex_unlock(&mdev->data.mutex); - - return ok; + sock = &mdev->tconn->data; + p = drbd_prepare_command(mdev, sock); + if (!p) + return -EIO; + p->sector = cpu_to_be64(sector); + p->block_id = ID_SYNCER /* unused */; + p->blksize = cpu_to_be32(size); + return drbd_send_command(mdev, sock, cmd, sizeof(*p), + digest, digest_size); } int drbd_send_ov_request(struct drbd_conf *mdev, sector_t sector, int size) { - int ok; - struct p_block_req p; + struct drbd_socket *sock; + struct p_block_req *p; - p.sector = cpu_to_be64(sector); - p.block_id = BE_DRBD_MAGIC + 0xbabe; - p.blksize = cpu_to_be32(size); - - ok = drbd_send_cmd(mdev, USE_DATA_SOCKET, P_OV_REQUEST, - (struct p_header80 *)&p, sizeof(p)); - return ok; + sock = &mdev->tconn->data; + p = drbd_prepare_command(mdev, sock); + if (!p) + return -EIO; + p->sector = cpu_to_be64(sector); + p->block_id = ID_SYNCER /* unused */; + p->blksize = cpu_to_be32(size); + return drbd_send_command(mdev, sock, P_OV_REQUEST, sizeof(*p), NULL, 0); } /* called on sndtimeo * returns false if we should retry, * true if we think connection is dead */ -static int we_should_drop_the_connection(struct drbd_conf *mdev, struct socket *sock) +static int we_should_drop_the_connection(struct drbd_tconn *tconn, struct socket *sock) { int drop_it; /* long elapsed = (long)(jiffies - mdev->last_received); */ - drop_it = mdev->meta.socket == sock - || !mdev->asender.task - || get_t_state(&mdev->asender) != Running - || mdev->state.conn < C_CONNECTED; + drop_it = tconn->meta.socket == sock + || !tconn->asender.task + || get_t_state(&tconn->asender) != RUNNING + || tconn->cstate < C_WF_REPORT_PARAMS; if (drop_it) return true; - drop_it = !--mdev->ko_count; + drop_it = !--tconn->ko_count; if (!drop_it) { - dev_err(DEV, "[%s/%d] sock_sendmsg time expired, ko = %u\n", - current->comm, current->pid, mdev->ko_count); - request_ping(mdev); + conn_err(tconn, "[%s/%d] sock_sendmsg time expired, ko = %u\n", + current->comm, current->pid, tconn->ko_count); + request_ping(tconn); } return drop_it; /* && (mdev->state == R_PRIMARY) */; } +static void drbd_update_congested(struct drbd_tconn *tconn) +{ + struct sock *sk = tconn->data.socket->sk; + if (sk->sk_wmem_queued > sk->sk_sndbuf * 4 / 5) + set_bit(NET_CONGESTED, &tconn->flags); +} + /* The idea of sendpage seems to be to put some kind of reference * to the page into the skb, and to hand it over to the NIC. In * this process get_page() gets called. @@ -2713,21 +1471,28 @@ static int we_should_drop_the_connection(struct drbd_conf *mdev, struct socket * * with page_count == 0 or PageSlab. */ static int _drbd_no_send_page(struct drbd_conf *mdev, struct page *page, - int offset, size_t size, unsigned msg_flags) + int offset, size_t size, unsigned msg_flags) { - int sent = drbd_send(mdev, mdev->data.socket, kmap(page) + offset, size, msg_flags); + struct socket *socket; + void *addr; + int err; + + socket = mdev->tconn->data.socket; + addr = kmap(page) + offset; + err = drbd_send_all(mdev->tconn, socket, addr, size, msg_flags); kunmap(page); - if (sent == size) - mdev->send_cnt += size>>9; - return sent == size; + if (!err) + mdev->send_cnt += size >> 9; + return err; } static int _drbd_send_page(struct drbd_conf *mdev, struct page *page, int offset, size_t size, unsigned msg_flags) { + struct socket *socket = mdev->tconn->data.socket; mm_segment_t oldfs = get_fs(); - int sent, ok; int len = size; + int err = -EIO; /* e.g. XFS meta- & log-data is in slab pages, which have a * page_count of 0 and/or have PageSlab() set. @@ -2739,34 +1504,35 @@ static int _drbd_send_page(struct drbd_conf *mdev, struct page *page, return _drbd_no_send_page(mdev, page, offset, size, msg_flags); msg_flags |= MSG_NOSIGNAL; - drbd_update_congested(mdev); + drbd_update_congested(mdev->tconn); set_fs(KERNEL_DS); do { - sent = mdev->data.socket->ops->sendpage(mdev->data.socket, page, - offset, len, - msg_flags); - if (sent == -EAGAIN) { - if (we_should_drop_the_connection(mdev, - mdev->data.socket)) - break; - else - continue; - } + int sent; + + sent = socket->ops->sendpage(socket, page, offset, len, msg_flags); if (sent <= 0) { + if (sent == -EAGAIN) { + if (we_should_drop_the_connection(mdev->tconn, socket)) + break; + continue; + } dev_warn(DEV, "%s: size=%d len=%d sent=%d\n", __func__, (int)size, len, sent); + if (sent < 0) + err = sent; break; } len -= sent; offset += sent; } while (len > 0 /* THINK && mdev->cstate >= C_CONNECTED*/); set_fs(oldfs); - clear_bit(NET_CONGESTED, &mdev->flags); + clear_bit(NET_CONGESTED, &mdev->tconn->flags); - ok = (len == 0); - if (likely(ok)) - mdev->send_cnt += size>>9; - return ok; + if (len == 0) { + err = 0; + mdev->send_cnt += size >> 9; + } + return err; } static int _drbd_send_bio(struct drbd_conf *mdev, struct bio *bio) @@ -2775,12 +1541,15 @@ static int _drbd_send_bio(struct drbd_conf *mdev, struct bio *bio) int i; /* hint all but last page with MSG_MORE */ bio_for_each_segment(bvec, bio, i) { - if (!_drbd_no_send_page(mdev, bvec->bv_page, - bvec->bv_offset, bvec->bv_len, - i == bio->bi_vcnt -1 ? 0 : MSG_MORE)) - return 0; + int err; + + err = _drbd_no_send_page(mdev, bvec->bv_page, + bvec->bv_offset, bvec->bv_len, + i == bio->bi_vcnt - 1 ? 0 : MSG_MORE); + if (err) + return err; } - return 1; + return 0; } static int _drbd_send_zc_bio(struct drbd_conf *mdev, struct bio *bio) @@ -2789,32 +1558,40 @@ static int _drbd_send_zc_bio(struct drbd_conf *mdev, struct bio *bio) int i; /* hint all but last page with MSG_MORE */ bio_for_each_segment(bvec, bio, i) { - if (!_drbd_send_page(mdev, bvec->bv_page, - bvec->bv_offset, bvec->bv_len, - i == bio->bi_vcnt -1 ? 0 : MSG_MORE)) - return 0; + int err; + + err = _drbd_send_page(mdev, bvec->bv_page, + bvec->bv_offset, bvec->bv_len, + i == bio->bi_vcnt - 1 ? 0 : MSG_MORE); + if (err) + return err; } - return 1; + return 0; } -static int _drbd_send_zc_ee(struct drbd_conf *mdev, struct drbd_epoch_entry *e) +static int _drbd_send_zc_ee(struct drbd_conf *mdev, + struct drbd_peer_request *peer_req) { - struct page *page = e->pages; - unsigned len = e->size; + struct page *page = peer_req->pages; + unsigned len = peer_req->i.size; + int err; + /* hint all but last page with MSG_MORE */ page_chain_for_each(page) { unsigned l = min_t(unsigned, len, PAGE_SIZE); - if (!_drbd_send_page(mdev, page, 0, l, - page_chain_next(page) ? MSG_MORE : 0)) - return 0; + + err = _drbd_send_page(mdev, page, 0, l, + page_chain_next(page) ? MSG_MORE : 0); + if (err) + return err; len -= l; } - return 1; + return 0; } static u32 bio_flags_to_wire(struct drbd_conf *mdev, unsigned long bi_rw) { - if (mdev->agreed_pro_version >= 95) + if (mdev->tconn->agreed_pro_version >= 95) return (bi_rw & REQ_SYNC ? DP_RW_SYNC : 0) | (bi_rw & REQ_FUA ? DP_FUA : 0) | (bi_rw & REQ_FLUSH ? DP_FLUSH : 0) | @@ -2828,50 +1605,36 @@ static u32 bio_flags_to_wire(struct drbd_conf *mdev, unsigned long bi_rw) */ int drbd_send_dblock(struct drbd_conf *mdev, struct drbd_request *req) { - int ok = 1; - struct p_data p; + struct drbd_socket *sock; + struct p_data *p; unsigned int dp_flags = 0; - void *dgb; int dgs; + int err; - if (!drbd_get_data_sock(mdev)) - return 0; - - dgs = (mdev->agreed_pro_version >= 87 && mdev->integrity_w_tfm) ? - crypto_hash_digestsize(mdev->integrity_w_tfm) : 0; - - if (req->size <= DRBD_MAX_SIZE_H80_PACKET) { - p.head.h80.magic = BE_DRBD_MAGIC; - p.head.h80.command = cpu_to_be16(P_DATA); - p.head.h80.length = - cpu_to_be16(sizeof(p) - sizeof(union p_header) + dgs + req->size); - } else { - p.head.h95.magic = BE_DRBD_MAGIC_BIG; - p.head.h95.command = cpu_to_be16(P_DATA); - p.head.h95.length = - cpu_to_be32(sizeof(p) - sizeof(union p_header) + dgs + req->size); - } - - p.sector = cpu_to_be64(req->sector); - p.block_id = (unsigned long)req; - p.seq_num = cpu_to_be32(atomic_add_return(1, &mdev->packet_seq)); + sock = &mdev->tconn->data; + p = drbd_prepare_command(mdev, sock); + dgs = mdev->tconn->integrity_tfm ? crypto_hash_digestsize(mdev->tconn->integrity_tfm) : 0; + if (!p) + return -EIO; + p->sector = cpu_to_be64(req->i.sector); + p->block_id = (unsigned long)req; + p->seq_num = cpu_to_be32(atomic_inc_return(&mdev->packet_seq)); dp_flags = bio_flags_to_wire(mdev, req->master_bio->bi_rw); - if (mdev->state.conn >= C_SYNC_SOURCE && mdev->state.conn <= C_PAUSED_SYNC_T) dp_flags |= DP_MAY_SET_IN_SYNC; - - p.dp_flags = cpu_to_be32(dp_flags); - set_bit(UNPLUG_REMOTE, &mdev->flags); - ok = (sizeof(p) == - drbd_send(mdev, mdev->data.socket, &p, sizeof(p), dgs ? MSG_MORE : 0)); - if (ok && dgs) { - dgb = mdev->int_dig_out; - drbd_csum_bio(mdev, mdev->integrity_w_tfm, req->master_bio, dgb); - ok = dgs == drbd_send(mdev, mdev->data.socket, dgb, dgs, 0); - } - if (ok) { + if (mdev->tconn->agreed_pro_version >= 100) { + if (req->rq_state & RQ_EXP_RECEIVE_ACK) + dp_flags |= DP_SEND_RECEIVE_ACK; + if (req->rq_state & RQ_EXP_WRITE_ACK) + dp_flags |= DP_SEND_WRITE_ACK; + } + p->dp_flags = cpu_to_be32(dp_flags); + if (dgs) + drbd_csum_bio(mdev, mdev->tconn->integrity_tfm, req->master_bio, p + 1); + err = __send_command(mdev->tconn, mdev->vnr, sock, P_DATA, sizeof(*p) + dgs, NULL, req->i.size); + if (!err) { /* For protocol A, we have to memcpy the payload into * socket buffers, as we may complete right away * as soon as we handed it over to tcp, at which point the data @@ -2883,92 +1646,76 @@ int drbd_send_dblock(struct drbd_conf *mdev, struct drbd_request *req) * out ok after sending on this side, but does not fit on the * receiving side, we sure have detected corruption elsewhere. */ - if (mdev->net_conf->wire_protocol == DRBD_PROT_A || dgs) - ok = _drbd_send_bio(mdev, req->master_bio); + if (!(req->rq_state & (RQ_EXP_RECEIVE_ACK | RQ_EXP_WRITE_ACK)) || dgs) + err = _drbd_send_bio(mdev, req->master_bio); else - ok = _drbd_send_zc_bio(mdev, req->master_bio); + err = _drbd_send_zc_bio(mdev, req->master_bio); /* double check digest, sometimes buffers have been modified in flight. */ if (dgs > 0 && dgs <= 64) { /* 64 byte, 512 bit, is the largest digest size * currently supported in kernel crypto. */ unsigned char digest[64]; - drbd_csum_bio(mdev, mdev->integrity_w_tfm, req->master_bio, digest); - if (memcmp(mdev->int_dig_out, digest, dgs)) { + drbd_csum_bio(mdev, mdev->tconn->integrity_tfm, req->master_bio, digest); + if (memcmp(p + 1, digest, dgs)) { dev_warn(DEV, "Digest mismatch, buffer modified by upper layers during write: %llus +%u\n", - (unsigned long long)req->sector, req->size); + (unsigned long long)req->i.sector, req->i.size); } } /* else if (dgs > 64) { ... Be noisy about digest too large ... } */ } + mutex_unlock(&sock->mutex); /* locked by drbd_prepare_command() */ - drbd_put_data_sock(mdev); - - return ok; + return err; } /* answer packet, used to send data back for read requests: * Peer -> (diskless) R_PRIMARY (P_DATA_REPLY) * C_SYNC_SOURCE -> C_SYNC_TARGET (P_RS_DATA_REPLY) */ -int drbd_send_block(struct drbd_conf *mdev, enum drbd_packets cmd, - struct drbd_epoch_entry *e) +int drbd_send_block(struct drbd_conf *mdev, enum drbd_packet cmd, + struct drbd_peer_request *peer_req) { - int ok; - struct p_data p; - void *dgb; + struct drbd_socket *sock; + struct p_data *p; + int err; int dgs; - dgs = (mdev->agreed_pro_version >= 87 && mdev->integrity_w_tfm) ? - crypto_hash_digestsize(mdev->integrity_w_tfm) : 0; - - if (e->size <= DRBD_MAX_SIZE_H80_PACKET) { - p.head.h80.magic = BE_DRBD_MAGIC; - p.head.h80.command = cpu_to_be16(cmd); - p.head.h80.length = - cpu_to_be16(sizeof(p) - sizeof(struct p_header80) + dgs + e->size); - } else { - p.head.h95.magic = BE_DRBD_MAGIC_BIG; - p.head.h95.command = cpu_to_be16(cmd); - p.head.h95.length = - cpu_to_be32(sizeof(p) - sizeof(struct p_header80) + dgs + e->size); - } - - p.sector = cpu_to_be64(e->sector); - p.block_id = e->block_id; - /* p.seq_num = 0; No sequence numbers here.. */ - - /* Only called by our kernel thread. - * This one may be interrupted by DRBD_SIG and/or DRBD_SIGKILL - * in response to admin command or module unload. - */ - if (!drbd_get_data_sock(mdev)) - return 0; + sock = &mdev->tconn->data; + p = drbd_prepare_command(mdev, sock); - ok = sizeof(p) == drbd_send(mdev, mdev->data.socket, &p, sizeof(p), dgs ? MSG_MORE : 0); - if (ok && dgs) { - dgb = mdev->int_dig_out; - drbd_csum_ee(mdev, mdev->integrity_w_tfm, e, dgb); - ok = dgs == drbd_send(mdev, mdev->data.socket, dgb, dgs, 0); - } - if (ok) - ok = _drbd_send_zc_ee(mdev, e); + dgs = mdev->tconn->integrity_tfm ? crypto_hash_digestsize(mdev->tconn->integrity_tfm) : 0; - drbd_put_data_sock(mdev); + if (!p) + return -EIO; + p->sector = cpu_to_be64(peer_req->i.sector); + p->block_id = peer_req->block_id; + p->seq_num = 0; /* unused */ + p->dp_flags = 0; + if (dgs) + drbd_csum_ee(mdev, mdev->tconn->integrity_tfm, peer_req, p + 1); + err = __send_command(mdev->tconn, mdev->vnr, sock, cmd, sizeof(*p) + dgs, NULL, peer_req->i.size); + if (!err) + err = _drbd_send_zc_ee(mdev, peer_req); + mutex_unlock(&sock->mutex); /* locked by drbd_prepare_command() */ - return ok; + return err; } -int drbd_send_oos(struct drbd_conf *mdev, struct drbd_request *req) +int drbd_send_out_of_sync(struct drbd_conf *mdev, struct drbd_request *req) { - struct p_block_desc p; - - p.sector = cpu_to_be64(req->sector); - p.blksize = cpu_to_be32(req->size); + struct drbd_socket *sock; + struct p_block_desc *p; - return drbd_send_cmd(mdev, USE_DATA_SOCKET, P_OUT_OF_SYNC, &p.head, sizeof(p)); + sock = &mdev->tconn->data; + p = drbd_prepare_command(mdev, sock); + if (!p) + return -EIO; + p->sector = cpu_to_be64(req->i.sector); + p->blksize = cpu_to_be32(req->i.size); + return drbd_send_command(mdev, sock, P_OUT_OF_SYNC, sizeof(*p), NULL, 0); } /* @@ -2987,7 +1734,7 @@ int drbd_send_oos(struct drbd_conf *mdev, struct drbd_request *req) /* * you must have down()ed the appropriate [m]sock_mutex elsewhere! */ -int drbd_send(struct drbd_conf *mdev, struct socket *sock, +int drbd_send(struct drbd_tconn *tconn, struct socket *sock, void *buf, size_t size, unsigned msg_flags) { struct kvec iov; @@ -2995,7 +1742,7 @@ int drbd_send(struct drbd_conf *mdev, struct socket *sock, int rv, sent = 0; if (!sock) - return -1000; + return -EBADR; /* THINK if (signal_pending) return ... ? */ @@ -3008,9 +1755,11 @@ int drbd_send(struct drbd_conf *mdev, struct socket *sock, msg.msg_controllen = 0; msg.msg_flags = msg_flags | MSG_NOSIGNAL; - if (sock == mdev->data.socket) { - mdev->ko_count = mdev->net_conf->ko_count; - drbd_update_congested(mdev); + if (sock == tconn->data.socket) { + rcu_read_lock(); + tconn->ko_count = rcu_dereference(tconn->net_conf)->ko_count; + rcu_read_unlock(); + drbd_update_congested(tconn); } do { /* STRANGE @@ -3024,12 +1773,11 @@ int drbd_send(struct drbd_conf *mdev, struct socket *sock, */ rv = kernel_sendmsg(sock, &msg, &iov, 1, size); if (rv == -EAGAIN) { - if (we_should_drop_the_connection(mdev, sock)) + if (we_should_drop_the_connection(tconn, sock)) break; else continue; } - D_ASSERT(rv != 0); if (rv == -EINTR) { flush_signals(current); rv = 0; @@ -3041,22 +1789,40 @@ int drbd_send(struct drbd_conf *mdev, struct socket *sock, iov.iov_len -= rv; } while (sent < size); - if (sock == mdev->data.socket) - clear_bit(NET_CONGESTED, &mdev->flags); + if (sock == tconn->data.socket) + clear_bit(NET_CONGESTED, &tconn->flags); if (rv <= 0) { if (rv != -EAGAIN) { - dev_err(DEV, "%s_sendmsg returned %d\n", - sock == mdev->meta.socket ? "msock" : "sock", - rv); - drbd_force_state(mdev, NS(conn, C_BROKEN_PIPE)); + conn_err(tconn, "%s_sendmsg returned %d\n", + sock == tconn->meta.socket ? "msock" : "sock", + rv); + conn_request_state(tconn, NS(conn, C_BROKEN_PIPE), CS_HARD); } else - drbd_force_state(mdev, NS(conn, C_TIMEOUT)); + conn_request_state(tconn, NS(conn, C_TIMEOUT), CS_HARD); } return sent; } +/** + * drbd_send_all - Send an entire buffer + * + * Returns 0 upon success and a negative error value otherwise. + */ +int drbd_send_all(struct drbd_tconn *tconn, struct socket *sock, void *buffer, + size_t size, unsigned msg_flags) +{ + int err; + + err = drbd_send(tconn, sock, buffer, size, msg_flags); + if (err < 0) + return err; + if (err != size) + return -EIO; + return 0; +} + static int drbd_open(struct block_device *bdev, fmode_t mode) { struct drbd_conf *mdev = bdev->bd_disk->private_data; @@ -3064,7 +1830,7 @@ static int drbd_open(struct block_device *bdev, fmode_t mode) int rv = 0; mutex_lock(&drbd_main_mutex); - spin_lock_irqsave(&mdev->req_lock, flags); + spin_lock_irqsave(&mdev->tconn->req_lock, flags); /* to have a stable mdev->state.role * and no race with updating open_cnt */ @@ -3077,7 +1843,7 @@ static int drbd_open(struct block_device *bdev, fmode_t mode) if (!rv) mdev->open_cnt++; - spin_unlock_irqrestore(&mdev->req_lock, flags); + spin_unlock_irqrestore(&mdev->tconn->req_lock, flags); mutex_unlock(&drbd_main_mutex); return rv; @@ -3094,35 +1860,14 @@ static int drbd_release(struct gendisk *gd, fmode_t mode) static void drbd_set_defaults(struct drbd_conf *mdev) { - /* This way we get a compile error when sync_conf grows, - and we forgot to initialize it here */ - mdev->sync_conf = (struct syncer_conf) { - /* .rate = */ DRBD_RATE_DEF, - /* .after = */ DRBD_AFTER_DEF, - /* .al_extents = */ DRBD_AL_EXTENTS_DEF, - /* .verify_alg = */ {}, 0, - /* .cpu_mask = */ {}, 0, - /* .csums_alg = */ {}, 0, - /* .use_rle = */ 0, - /* .on_no_data = */ DRBD_ON_NO_DATA_DEF, - /* .c_plan_ahead = */ DRBD_C_PLAN_AHEAD_DEF, - /* .c_delay_target = */ DRBD_C_DELAY_TARGET_DEF, - /* .c_fill_target = */ DRBD_C_FILL_TARGET_DEF, - /* .c_max_rate = */ DRBD_C_MAX_RATE_DEF, - /* .c_min_rate = */ DRBD_C_MIN_RATE_DEF - }; - - /* Have to use that way, because the layout differs between - big endian and little endian */ - mdev->state = (union drbd_state) { + /* Beware! The actual layout differs + * between big endian and little endian */ + mdev->state = (union drbd_dev_state) { { .role = R_SECONDARY, .peer = R_UNKNOWN, .conn = C_STANDALONE, .disk = D_DISKLESS, .pdsk = D_UNKNOWN, - .susp = 0, - .susp_nod = 0, - .susp_fen = 0 } }; } @@ -3138,28 +1883,17 @@ void drbd_init_set_defaults(struct drbd_conf *mdev) atomic_set(&mdev->rs_pending_cnt, 0); atomic_set(&mdev->unacked_cnt, 0); atomic_set(&mdev->local_cnt, 0); - atomic_set(&mdev->net_cnt, 0); - atomic_set(&mdev->packet_seq, 0); - atomic_set(&mdev->pp_in_use, 0); atomic_set(&mdev->pp_in_use_by_net, 0); atomic_set(&mdev->rs_sect_in, 0); atomic_set(&mdev->rs_sect_ev, 0); atomic_set(&mdev->ap_in_flight, 0); atomic_set(&mdev->md_io_in_use, 0); - mutex_init(&mdev->data.mutex); - mutex_init(&mdev->meta.mutex); - sema_init(&mdev->data.work.s, 0); - sema_init(&mdev->meta.work.s, 0); - mutex_init(&mdev->state_mutex); - - spin_lock_init(&mdev->data.work.q_lock); - spin_lock_init(&mdev->meta.work.q_lock); + mutex_init(&mdev->own_state_mutex); + mdev->state_mutex = &mdev->own_state_mutex; spin_lock_init(&mdev->al_lock); - spin_lock_init(&mdev->req_lock); spin_lock_init(&mdev->peer_seq_lock); - spin_lock_init(&mdev->epoch_lock); INIT_LIST_HEAD(&mdev->active_ee); INIT_LIST_HEAD(&mdev->sync_ee); @@ -3167,8 +1901,6 @@ void drbd_init_set_defaults(struct drbd_conf *mdev) INIT_LIST_HEAD(&mdev->read_ee); INIT_LIST_HEAD(&mdev->net_ee); INIT_LIST_HEAD(&mdev->resync_reads); - INIT_LIST_HEAD(&mdev->data.work.q); - INIT_LIST_HEAD(&mdev->meta.work.q); INIT_LIST_HEAD(&mdev->resync_work.list); INIT_LIST_HEAD(&mdev->unplug_work.list); INIT_LIST_HEAD(&mdev->go_diskless.list); @@ -3182,6 +1914,14 @@ void drbd_init_set_defaults(struct drbd_conf *mdev) mdev->md_sync_work.cb = w_md_sync; mdev->bm_io_work.w.cb = w_bitmap_io; mdev->start_resync_work.cb = w_start_resync; + + mdev->resync_work.mdev = mdev; + mdev->unplug_work.mdev = mdev; + mdev->go_diskless.mdev = mdev; + mdev->md_sync_work.mdev = mdev; + mdev->bm_io_work.w.mdev = mdev; + mdev->start_resync_work.mdev = mdev; + init_timer(&mdev->resync_timer); init_timer(&mdev->md_sync_timer); init_timer(&mdev->start_resync_timer); @@ -3197,17 +1937,10 @@ void drbd_init_set_defaults(struct drbd_conf *mdev) init_waitqueue_head(&mdev->misc_wait); init_waitqueue_head(&mdev->state_wait); - init_waitqueue_head(&mdev->net_cnt_wait); init_waitqueue_head(&mdev->ee_wait); init_waitqueue_head(&mdev->al_wait); init_waitqueue_head(&mdev->seq_wait); - drbd_thread_init(mdev, &mdev->receiver, drbdd_init); - drbd_thread_init(mdev, &mdev->worker, drbd_worker); - drbd_thread_init(mdev, &mdev->asender, drbd_asender); - - mdev->agreed_pro_version = PRO_VERSION_MAX; - mdev->write_ordering = WO_bdev_flush; mdev->resync_wenr = LC_FREE; mdev->peer_max_bio_size = DRBD_MAX_BIO_SIZE_SAFE; mdev->local_max_bio_size = DRBD_MAX_BIO_SIZE_SAFE; @@ -3216,13 +1949,10 @@ void drbd_init_set_defaults(struct drbd_conf *mdev) void drbd_mdev_cleanup(struct drbd_conf *mdev) { int i; - if (mdev->receiver.t_state != None) + if (mdev->tconn->receiver.t_state != NONE) dev_err(DEV, "ASSERT FAILED: receiver t_state == %d expected 0.\n", - mdev->receiver.t_state); + mdev->tconn->receiver.t_state); - /* no need to lock it, I'm the only thread alive */ - if (atomic_read(&mdev->current_epoch->epoch_size) != 0) - dev_err(DEV, "epoch_size:%d\n", atomic_read(&mdev->current_epoch->epoch_size)); mdev->al_writ_cnt = mdev->bm_writ_cnt = mdev->read_cnt = @@ -3239,7 +1969,7 @@ void drbd_mdev_cleanup(struct drbd_conf *mdev) mdev->rs_mark_left[i] = 0; mdev->rs_mark_time[i] = 0; } - D_ASSERT(mdev->net_conf == NULL); + D_ASSERT(mdev->tconn->net_conf == NULL); drbd_set_my_capacity(mdev, 0); if (mdev->bitmap) { @@ -3248,21 +1978,18 @@ void drbd_mdev_cleanup(struct drbd_conf *mdev) drbd_bm_cleanup(mdev); } - drbd_free_resources(mdev); + drbd_free_bc(mdev->ldev); + mdev->ldev = NULL; + clear_bit(AL_SUSPENDED, &mdev->flags); - /* - * currently we drbd_init_ee only on module load, so - * we may do drbd_release_ee only on module unload! - */ D_ASSERT(list_empty(&mdev->active_ee)); D_ASSERT(list_empty(&mdev->sync_ee)); D_ASSERT(list_empty(&mdev->done_ee)); D_ASSERT(list_empty(&mdev->read_ee)); D_ASSERT(list_empty(&mdev->net_ee)); D_ASSERT(list_empty(&mdev->resync_reads)); - D_ASSERT(list_empty(&mdev->data.work.q)); - D_ASSERT(list_empty(&mdev->meta.work.q)); + D_ASSERT(list_empty(&mdev->tconn->sender_work.q)); D_ASSERT(list_empty(&mdev->resync_work.list)); D_ASSERT(list_empty(&mdev->unplug_work.list)); D_ASSERT(list_empty(&mdev->go_diskless.list)); @@ -3336,7 +2063,7 @@ static int drbd_create_mempools(void) goto Enomem; drbd_ee_cache = kmem_cache_create( - "drbd_ee", sizeof(struct drbd_epoch_entry), 0, 0, NULL); + "drbd_ee", sizeof(struct drbd_peer_request), 0, 0, NULL); if (drbd_ee_cache == NULL) goto Enomem; @@ -3351,11 +2078,9 @@ static int drbd_create_mempools(void) goto Enomem; /* mempools */ -#ifdef COMPAT_HAVE_BIOSET_CREATE drbd_md_io_bio_set = bioset_create(DRBD_MIN_POOL_PAGES, 0); if (drbd_md_io_bio_set == NULL) goto Enomem; -#endif drbd_md_io_page_pool = mempool_create_page_pool(DRBD_MIN_POOL_PAGES, 0); if (drbd_md_io_page_pool == NULL) @@ -3404,73 +2129,53 @@ static struct notifier_block drbd_notifier = { .notifier_call = drbd_notify_sys, }; -static void drbd_release_ee_lists(struct drbd_conf *mdev) +static void drbd_release_all_peer_reqs(struct drbd_conf *mdev) { int rr; - rr = drbd_release_ee(mdev, &mdev->active_ee); + rr = drbd_free_peer_reqs(mdev, &mdev->active_ee); if (rr) dev_err(DEV, "%d EEs in active list found!\n", rr); - rr = drbd_release_ee(mdev, &mdev->sync_ee); + rr = drbd_free_peer_reqs(mdev, &mdev->sync_ee); if (rr) dev_err(DEV, "%d EEs in sync list found!\n", rr); - rr = drbd_release_ee(mdev, &mdev->read_ee); + rr = drbd_free_peer_reqs(mdev, &mdev->read_ee); if (rr) dev_err(DEV, "%d EEs in read list found!\n", rr); - rr = drbd_release_ee(mdev, &mdev->done_ee); + rr = drbd_free_peer_reqs(mdev, &mdev->done_ee); if (rr) dev_err(DEV, "%d EEs in done list found!\n", rr); - rr = drbd_release_ee(mdev, &mdev->net_ee); + rr = drbd_free_peer_reqs(mdev, &mdev->net_ee); if (rr) dev_err(DEV, "%d EEs in net list found!\n", rr); } -/* caution. no locking. - * currently only used from module cleanup code. */ -static void drbd_delete_device(unsigned int minor) +/* caution. no locking. */ +void drbd_minor_destroy(struct kref *kref) { - struct drbd_conf *mdev = minor_to_mdev(minor); - - if (!mdev) - return; + struct drbd_conf *mdev = container_of(kref, struct drbd_conf, kref); + struct drbd_tconn *tconn = mdev->tconn; del_timer_sync(&mdev->request_timer); /* paranoia asserts */ - if (mdev->open_cnt != 0) - dev_err(DEV, "open_cnt = %d in %s:%u", mdev->open_cnt, - __FILE__ , __LINE__); - - ERR_IF (!list_empty(&mdev->data.work.q)) { - struct list_head *lp; - list_for_each(lp, &mdev->data.work.q) { - dev_err(DEV, "lp = %p\n", lp); - } - }; + D_ASSERT(mdev->open_cnt == 0); /* end paranoia asserts */ - del_gendisk(mdev->vdisk); - /* cleanup stuff that may have been allocated during * device (re-)configuration or state changes */ if (mdev->this_bdev) bdput(mdev->this_bdev); - drbd_free_resources(mdev); + drbd_free_bc(mdev->ldev); + mdev->ldev = NULL; - drbd_release_ee_lists(mdev); - - /* should be freed on disconnect? */ - kfree(mdev->ee_hash); - /* - mdev->ee_hash_s = 0; - mdev->ee_hash = NULL; - */ + drbd_release_all_peer_reqs(mdev); lc_destroy(mdev->act_log); lc_destroy(mdev->resync); @@ -3478,19 +2183,101 @@ static void drbd_delete_device(unsigned int minor) kfree(mdev->p_uuid); /* mdev->p_uuid = NULL; */ - kfree(mdev->int_dig_out); - kfree(mdev->int_dig_in); - kfree(mdev->int_dig_vv); + if (mdev->bitmap) /* should no longer be there. */ + drbd_bm_cleanup(mdev); + __free_page(mdev->md_io_page); + put_disk(mdev->vdisk); + blk_cleanup_queue(mdev->rq_queue); + kfree(mdev->rs_plan_s); + kfree(mdev); - /* cleanup the rest that has been - * allocated from drbd_new_device - * and actually free the mdev itself */ - drbd_free_mdev(mdev); + kref_put(&tconn->kref, &conn_destroy); } +/* One global retry thread, if we need to push back some bio and have it + * reinserted through our make request function. + */ +static struct retry_worker { + struct workqueue_struct *wq; + struct work_struct worker; + + spinlock_t lock; + struct list_head writes; +} retry; + +static void do_retry(struct work_struct *ws) +{ + struct retry_worker *retry = container_of(ws, struct retry_worker, worker); + LIST_HEAD(writes); + struct drbd_request *req, *tmp; + + spin_lock_irq(&retry->lock); + list_splice_init(&retry->writes, &writes); + spin_unlock_irq(&retry->lock); + + list_for_each_entry_safe(req, tmp, &writes, tl_requests) { + struct drbd_conf *mdev = req->w.mdev; + struct bio *bio = req->master_bio; + unsigned long start_time = req->start_time; + bool expected; + + expected = + expect(atomic_read(&req->completion_ref) == 0) && + expect(req->rq_state & RQ_POSTPONED) && + expect((req->rq_state & RQ_LOCAL_PENDING) == 0 || + (req->rq_state & RQ_LOCAL_ABORTED) != 0); + + if (!expected) + dev_err(DEV, "req=%p completion_ref=%d rq_state=%x\n", + req, atomic_read(&req->completion_ref), + req->rq_state); + + /* We still need to put one kref associated with the + * "completion_ref" going zero in the code path that queued it + * here. The request object may still be referenced by a + * frozen local req->private_bio, in case we force-detached. + */ + kref_put(&req->kref, drbd_req_destroy); + + /* A single suspended or otherwise blocking device may stall + * all others as well. Fortunately, this code path is to + * recover from a situation that "should not happen": + * concurrent writes in multi-primary setup. + * In a "normal" lifecycle, this workqueue is supposed to be + * destroyed without ever doing anything. + * If it turns out to be an issue anyways, we can do per + * resource (replication group) or per device (minor) retry + * workqueues instead. + */ + + /* We are not just doing generic_make_request(), + * as we want to keep the start_time information. */ + inc_ap_bio(mdev); + __drbd_make_request(mdev, bio, start_time); + } +} + +void drbd_restart_request(struct drbd_request *req) +{ + unsigned long flags; + spin_lock_irqsave(&retry.lock, flags); + list_move_tail(&req->tl_requests, &retry.writes); + spin_unlock_irqrestore(&retry.lock, flags); + + /* Drop the extra reference that would otherwise + * have been dropped by complete_master_bio. + * do_retry() needs to grab a new one. */ + dec_ap_bio(req->w.mdev); + + queue_work(retry.wq, &retry.worker); +} + + static void drbd_cleanup(void) { unsigned int i; + struct drbd_conf *mdev; + struct drbd_tconn *tconn, *tmp; unregister_reboot_notifier(&drbd_notifier); @@ -3505,19 +2292,31 @@ static void drbd_cleanup(void) if (drbd_proc) remove_proc_entry("drbd", NULL); - drbd_nl_cleanup(); + if (retry.wq) + destroy_workqueue(retry.wq); + + drbd_genl_unregister(); - if (minor_table) { - i = minor_count; - while (i--) - drbd_delete_device(i); - drbd_destroy_mempools(); + idr_for_each_entry(&minors, mdev, i) { + idr_remove(&minors, mdev_to_minor(mdev)); + idr_remove(&mdev->tconn->volumes, mdev->vnr); + del_gendisk(mdev->vdisk); + /* synchronize_rcu(); No other threads running at this point */ + kref_put(&mdev->kref, &drbd_minor_destroy); } - kfree(minor_table); + /* not _rcu since, no other updater anymore. Genl already unregistered */ + list_for_each_entry_safe(tconn, tmp, &drbd_tconns, all_tconn) { + list_del(&tconn->all_tconn); /* not _rcu no proc, not other threads */ + /* synchronize_rcu(); */ + kref_put(&tconn->kref, &conn_destroy); + } + drbd_destroy_mempools(); unregister_blkdev(DRBD_MAJOR, "drbd"); + idr_destroy(&minors); + printk(KERN_INFO "drbd: module cleanup done.\n"); } @@ -3542,7 +2341,7 @@ static int drbd_congested(void *congested_data, int bdi_bits) goto out; } - if (test_bit(CALLBACK_PENDING, &mdev->flags)) { + if (test_bit(CALLBACK_PENDING, &mdev->tconn->flags)) { r |= (1 << BDI_async_congested); /* Without good local data, we would need to read from remote, * and that would need the worker thread as well, which is @@ -3566,7 +2365,7 @@ static int drbd_congested(void *congested_data, int bdi_bits) reason = 'b'; } - if (bdi_bits & (1 << BDI_async_congested) && test_bit(NET_CONGESTED, &mdev->flags)) { + if (bdi_bits & (1 << BDI_async_congested) && test_bit(NET_CONGESTED, &mdev->tconn->flags)) { r |= (1 << BDI_async_congested); reason = reason == 'b' ? 'a' : 'n'; } @@ -3576,20 +2375,243 @@ out: return r; } -struct drbd_conf *drbd_new_device(unsigned int minor) +static void drbd_init_workqueue(struct drbd_work_queue* wq) +{ + spin_lock_init(&wq->q_lock); + INIT_LIST_HEAD(&wq->q); + init_waitqueue_head(&wq->q_wait); +} + +struct drbd_tconn *conn_get_by_name(const char *name) +{ + struct drbd_tconn *tconn; + + if (!name || !name[0]) + return NULL; + + rcu_read_lock(); + list_for_each_entry_rcu(tconn, &drbd_tconns, all_tconn) { + if (!strcmp(tconn->name, name)) { + kref_get(&tconn->kref); + goto found; + } + } + tconn = NULL; +found: + rcu_read_unlock(); + return tconn; +} + +struct drbd_tconn *conn_get_by_addrs(void *my_addr, int my_addr_len, + void *peer_addr, int peer_addr_len) +{ + struct drbd_tconn *tconn; + + rcu_read_lock(); + list_for_each_entry_rcu(tconn, &drbd_tconns, all_tconn) { + if (tconn->my_addr_len == my_addr_len && + tconn->peer_addr_len == peer_addr_len && + !memcmp(&tconn->my_addr, my_addr, my_addr_len) && + !memcmp(&tconn->peer_addr, peer_addr, peer_addr_len)) { + kref_get(&tconn->kref); + goto found; + } + } + tconn = NULL; +found: + rcu_read_unlock(); + return tconn; +} + +static int drbd_alloc_socket(struct drbd_socket *socket) +{ + socket->rbuf = (void *) __get_free_page(GFP_KERNEL); + if (!socket->rbuf) + return -ENOMEM; + socket->sbuf = (void *) __get_free_page(GFP_KERNEL); + if (!socket->sbuf) + return -ENOMEM; + return 0; +} + +static void drbd_free_socket(struct drbd_socket *socket) +{ + free_page((unsigned long) socket->sbuf); + free_page((unsigned long) socket->rbuf); +} + +void conn_free_crypto(struct drbd_tconn *tconn) +{ + drbd_free_sock(tconn); + + crypto_free_hash(tconn->csums_tfm); + crypto_free_hash(tconn->verify_tfm); + crypto_free_hash(tconn->cram_hmac_tfm); + crypto_free_hash(tconn->integrity_tfm); + crypto_free_hash(tconn->peer_integrity_tfm); + kfree(tconn->int_dig_in); + kfree(tconn->int_dig_vv); + + tconn->csums_tfm = NULL; + tconn->verify_tfm = NULL; + tconn->cram_hmac_tfm = NULL; + tconn->integrity_tfm = NULL; + tconn->peer_integrity_tfm = NULL; + tconn->int_dig_in = NULL; + tconn->int_dig_vv = NULL; +} + +int set_resource_options(struct drbd_tconn *tconn, struct res_opts *res_opts) +{ + cpumask_var_t new_cpu_mask; + int err; + + if (!zalloc_cpumask_var(&new_cpu_mask, GFP_KERNEL)) + return -ENOMEM; + /* + retcode = ERR_NOMEM; + drbd_msg_put_info("unable to allocate cpumask"); + */ + + /* silently ignore cpu mask on UP kernel */ + if (nr_cpu_ids > 1 && res_opts->cpu_mask[0] != 0) { + /* FIXME: Get rid of constant 32 here */ + err = bitmap_parse(res_opts->cpu_mask, 32, + cpumask_bits(new_cpu_mask), nr_cpu_ids); + if (err) { + conn_warn(tconn, "bitmap_parse() failed with %d\n", err); + /* retcode = ERR_CPU_MASK_PARSE; */ + goto fail; + } + } + tconn->res_opts = *res_opts; + if (!cpumask_equal(tconn->cpu_mask, new_cpu_mask)) { + cpumask_copy(tconn->cpu_mask, new_cpu_mask); + drbd_calc_cpu_mask(tconn); + tconn->receiver.reset_cpu_mask = 1; + tconn->asender.reset_cpu_mask = 1; + tconn->worker.reset_cpu_mask = 1; + } + err = 0; + +fail: + free_cpumask_var(new_cpu_mask); + return err; + +} + +/* caller must be under genl_lock() */ +struct drbd_tconn *conn_create(const char *name, struct res_opts *res_opts) +{ + struct drbd_tconn *tconn; + + tconn = kzalloc(sizeof(struct drbd_tconn), GFP_KERNEL); + if (!tconn) + return NULL; + + tconn->name = kstrdup(name, GFP_KERNEL); + if (!tconn->name) + goto fail; + + if (drbd_alloc_socket(&tconn->data)) + goto fail; + if (drbd_alloc_socket(&tconn->meta)) + goto fail; + + if (!zalloc_cpumask_var(&tconn->cpu_mask, GFP_KERNEL)) + goto fail; + + if (set_resource_options(tconn, res_opts)) + goto fail; + + tconn->current_epoch = kzalloc(sizeof(struct drbd_epoch), GFP_KERNEL); + if (!tconn->current_epoch) + goto fail; + + INIT_LIST_HEAD(&tconn->transfer_log); + + INIT_LIST_HEAD(&tconn->current_epoch->list); + tconn->epochs = 1; + spin_lock_init(&tconn->epoch_lock); + tconn->write_ordering = WO_bdev_flush; + + tconn->send.seen_any_write_yet = false; + tconn->send.current_epoch_nr = 0; + tconn->send.current_epoch_writes = 0; + + tconn->cstate = C_STANDALONE; + mutex_init(&tconn->cstate_mutex); + spin_lock_init(&tconn->req_lock); + mutex_init(&tconn->conf_update); + init_waitqueue_head(&tconn->ping_wait); + idr_init(&tconn->volumes); + + drbd_init_workqueue(&tconn->sender_work); + mutex_init(&tconn->data.mutex); + mutex_init(&tconn->meta.mutex); + + drbd_thread_init(tconn, &tconn->receiver, drbdd_init, "receiver"); + drbd_thread_init(tconn, &tconn->worker, drbd_worker, "worker"); + drbd_thread_init(tconn, &tconn->asender, drbd_asender, "asender"); + + kref_init(&tconn->kref); + list_add_tail_rcu(&tconn->all_tconn, &drbd_tconns); + + return tconn; + +fail: + kfree(tconn->current_epoch); + free_cpumask_var(tconn->cpu_mask); + drbd_free_socket(&tconn->meta); + drbd_free_socket(&tconn->data); + kfree(tconn->name); + kfree(tconn); + + return NULL; +} + +void conn_destroy(struct kref *kref) +{ + struct drbd_tconn *tconn = container_of(kref, struct drbd_tconn, kref); + + if (atomic_read(&tconn->current_epoch->epoch_size) != 0) + conn_err(tconn, "epoch_size:%d\n", atomic_read(&tconn->current_epoch->epoch_size)); + kfree(tconn->current_epoch); + + idr_destroy(&tconn->volumes); + + free_cpumask_var(tconn->cpu_mask); + drbd_free_socket(&tconn->meta); + drbd_free_socket(&tconn->data); + kfree(tconn->name); + kfree(tconn->int_dig_in); + kfree(tconn->int_dig_vv); + kfree(tconn); +} + +enum drbd_ret_code conn_new_minor(struct drbd_tconn *tconn, unsigned int minor, int vnr) { struct drbd_conf *mdev; struct gendisk *disk; struct request_queue *q; + int vnr_got = vnr; + int minor_got = minor; + enum drbd_ret_code err = ERR_NOMEM; + + mdev = minor_to_mdev(minor); + if (mdev) + return ERR_MINOR_EXISTS; /* GFP_KERNEL, we are outside of all write-out paths */ mdev = kzalloc(sizeof(struct drbd_conf), GFP_KERNEL); if (!mdev) - return NULL; - if (!zalloc_cpumask_var(&mdev->cpu_mask, GFP_KERNEL)) - goto out_no_cpumask; + return ERR_NOMEM; + + kref_get(&tconn->kref); + mdev->tconn = tconn; mdev->minor = minor; + mdev->vnr = vnr; drbd_init_set_defaults(mdev); @@ -3627,7 +2649,7 @@ struct drbd_conf *drbd_new_device(unsigned int minor) blk_queue_max_hw_sectors(q, DRBD_MAX_BIO_SIZE_SAFE >> 8); blk_queue_bounce_limit(q, BLK_BOUNCE_ANY); blk_queue_merge_bvec(q, drbd_merge_bvec); - q->queue_lock = &mdev->req_lock; + q->queue_lock = &mdev->tconn->req_lock; /* needed since we use */ mdev->md_io_page = alloc_page(GFP_KERNEL); if (!mdev->md_io_page) @@ -3635,30 +2657,44 @@ struct drbd_conf *drbd_new_device(unsigned int minor) if (drbd_bm_init(mdev)) goto out_no_bitmap; - /* no need to lock access, we are still initializing this minor device. */ - if (!tl_init(mdev)) - goto out_no_tl; - - mdev->app_reads_hash = kzalloc(APP_R_HSIZE*sizeof(void *), GFP_KERNEL); - if (!mdev->app_reads_hash) - goto out_no_app_reads; - - mdev->current_epoch = kzalloc(sizeof(struct drbd_epoch), GFP_KERNEL); - if (!mdev->current_epoch) - goto out_no_epoch; - - INIT_LIST_HEAD(&mdev->current_epoch->list); - mdev->epochs = 1; - - return mdev; - -/* out_whatever_else: - kfree(mdev->current_epoch); */ -out_no_epoch: - kfree(mdev->app_reads_hash); -out_no_app_reads: - tl_cleanup(mdev); -out_no_tl: + mdev->read_requests = RB_ROOT; + mdev->write_requests = RB_ROOT; + + if (!idr_pre_get(&minors, GFP_KERNEL)) + goto out_no_minor_idr; + if (idr_get_new_above(&minors, mdev, minor, &minor_got)) + goto out_no_minor_idr; + if (minor_got != minor) { + err = ERR_MINOR_EXISTS; + drbd_msg_put_info("requested minor exists already"); + goto out_idr_remove_minor; + } + + if (!idr_pre_get(&tconn->volumes, GFP_KERNEL)) + goto out_idr_remove_minor; + if (idr_get_new_above(&tconn->volumes, mdev, vnr, &vnr_got)) + goto out_idr_remove_minor; + if (vnr_got != vnr) { + err = ERR_INVALID_REQUEST; + drbd_msg_put_info("requested volume exists already"); + goto out_idr_remove_vol; + } + add_disk(disk); + kref_init(&mdev->kref); /* one ref for both idrs and the the add_disk */ + + /* inherit the connection state */ + mdev->state.conn = tconn->cstate; + if (mdev->state.conn == C_WF_REPORT_PARAMS) + drbd_connected(mdev); + + return NO_ERROR; + +out_idr_remove_vol: + idr_remove(&tconn->volumes, vnr_got); +out_idr_remove_minor: + idr_remove(&minors, minor_got); + synchronize_rcu(); +out_no_minor_idr: drbd_bm_cleanup(mdev); out_no_bitmap: __free_page(mdev->md_io_page); @@ -3667,55 +2703,25 @@ out_no_io_page: out_no_disk: blk_cleanup_queue(q); out_no_q: - free_cpumask_var(mdev->cpu_mask); -out_no_cpumask: - kfree(mdev); - return NULL; -} - -/* counterpart of drbd_new_device. - * last part of drbd_delete_device. */ -void drbd_free_mdev(struct drbd_conf *mdev) -{ - kfree(mdev->current_epoch); - kfree(mdev->app_reads_hash); - tl_cleanup(mdev); - if (mdev->bitmap) /* should no longer be there. */ - drbd_bm_cleanup(mdev); - __free_page(mdev->md_io_page); - put_disk(mdev->vdisk); - blk_cleanup_queue(mdev->rq_queue); - free_cpumask_var(mdev->cpu_mask); - drbd_free_tl_hash(mdev); kfree(mdev); + kref_put(&tconn->kref, &conn_destroy); + return err; } - int __init drbd_init(void) { int err; - if (sizeof(struct p_handshake) != 80) { - printk(KERN_ERR - "drbd: never change the size or layout " - "of the HandShake packet.\n"); - return -EINVAL; - } - if (minor_count < DRBD_MINOR_COUNT_MIN || minor_count > DRBD_MINOR_COUNT_MAX) { printk(KERN_ERR - "drbd: invalid minor_count (%d)\n", minor_count); + "drbd: invalid minor_count (%d)\n", minor_count); #ifdef MODULE return -EINVAL; #else - minor_count = 8; + minor_count = DRBD_MINOR_COUNT_DEF; #endif } - err = drbd_nl_init(); - if (err) - return err; - err = register_blkdev(DRBD_MAJOR, "drbd"); if (err) { printk(KERN_ERR @@ -3724,6 +2730,13 @@ int __init drbd_init(void) return err; } + err = drbd_genl_register(); + if (err) { + printk(KERN_ERR "drbd: unable to register generic netlink family\n"); + goto fail; + } + + register_reboot_notifier(&drbd_notifier); /* @@ -3734,22 +2747,29 @@ int __init drbd_init(void) init_waitqueue_head(&drbd_pp_wait); drbd_proc = NULL; /* play safe for drbd_cleanup */ - minor_table = kzalloc(sizeof(struct drbd_conf *)*minor_count, - GFP_KERNEL); - if (!minor_table) - goto Enomem; + idr_init(&minors); err = drbd_create_mempools(); if (err) - goto Enomem; + goto fail; drbd_proc = proc_create_data("drbd", S_IFREG | S_IRUGO , NULL, &drbd_proc_fops, NULL); if (!drbd_proc) { printk(KERN_ERR "drbd: unable to register proc file\n"); - goto Enomem; + goto fail; } rwlock_init(&global_state_lock); + INIT_LIST_HEAD(&drbd_tconns); + + retry.wq = create_singlethread_workqueue("drbd-reissue"); + if (!retry.wq) { + printk(KERN_ERR "drbd: unable to create retry workqueue\n"); + goto fail; + } + INIT_WORK(&retry.worker, do_retry); + spin_lock_init(&retry.lock); + INIT_LIST_HEAD(&retry.writes); printk(KERN_INFO "drbd: initialized. " "Version: " REL_VERSION " (api:%d/proto:%d-%d)\n", @@ -3757,11 +2777,10 @@ int __init drbd_init(void) printk(KERN_INFO "drbd: %s\n", drbd_buildtag()); printk(KERN_INFO "drbd: registered as block device major %d\n", DRBD_MAJOR); - printk(KERN_INFO "drbd: minor_table @ 0x%p\n", minor_table); return 0; /* Success! */ -Enomem: +fail: drbd_cleanup(); if (err == -ENOMEM) /* currently always the case */ @@ -3782,47 +2801,42 @@ void drbd_free_bc(struct drbd_backing_dev *ldev) kfree(ldev); } -void drbd_free_sock(struct drbd_conf *mdev) +void drbd_free_sock(struct drbd_tconn *tconn) { - if (mdev->data.socket) { - mutex_lock(&mdev->data.mutex); - kernel_sock_shutdown(mdev->data.socket, SHUT_RDWR); - sock_release(mdev->data.socket); - mdev->data.socket = NULL; - mutex_unlock(&mdev->data.mutex); + if (tconn->data.socket) { + mutex_lock(&tconn->data.mutex); + kernel_sock_shutdown(tconn->data.socket, SHUT_RDWR); + sock_release(tconn->data.socket); + tconn->data.socket = NULL; + mutex_unlock(&tconn->data.mutex); } - if (mdev->meta.socket) { - mutex_lock(&mdev->meta.mutex); - kernel_sock_shutdown(mdev->meta.socket, SHUT_RDWR); - sock_release(mdev->meta.socket); - mdev->meta.socket = NULL; - mutex_unlock(&mdev->meta.mutex); + if (tconn->meta.socket) { + mutex_lock(&tconn->meta.mutex); + kernel_sock_shutdown(tconn->meta.socket, SHUT_RDWR); + sock_release(tconn->meta.socket); + tconn->meta.socket = NULL; + mutex_unlock(&tconn->meta.mutex); } } +/* meta data management */ -void drbd_free_resources(struct drbd_conf *mdev) +void conn_md_sync(struct drbd_tconn *tconn) { - crypto_free_hash(mdev->csums_tfm); - mdev->csums_tfm = NULL; - crypto_free_hash(mdev->verify_tfm); - mdev->verify_tfm = NULL; - crypto_free_hash(mdev->cram_hmac_tfm); - mdev->cram_hmac_tfm = NULL; - crypto_free_hash(mdev->integrity_w_tfm); - mdev->integrity_w_tfm = NULL; - crypto_free_hash(mdev->integrity_r_tfm); - mdev->integrity_r_tfm = NULL; - - drbd_free_sock(mdev); + struct drbd_conf *mdev; + int vnr; - __no_warn(local, - drbd_free_bc(mdev->ldev); - mdev->ldev = NULL;); + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) { + kref_get(&mdev->kref); + rcu_read_unlock(); + drbd_md_sync(mdev); + kref_put(&mdev->kref, &drbd_minor_destroy); + rcu_read_lock(); + } + rcu_read_unlock(); } -/* meta data management */ - struct meta_data_on_disk { u64 la_size; /* last agreed size. */ u64 uuid[UI_SIZE]; /* UUIDs. */ @@ -3833,7 +2847,7 @@ struct meta_data_on_disk { u32 md_size_sect; u32 al_offset; /* offset to this block */ u32 al_nr_extents; /* important for restoring the AL */ - /* `-- act_log->nr_elements <-- sync_conf.al_extents */ + /* `-- act_log->nr_elements <-- ldev->dc.al_extents */ u32 bm_offset; /* offset to the bitmap, from here */ u32 bm_bytes_per_bit; /* BM_BLOCK_SIZE */ u32 la_peer_max_bio_size; /* last peer max_bio_size */ @@ -3871,7 +2885,7 @@ void drbd_md_sync(struct drbd_conf *mdev) for (i = UI_CURRENT; i < UI_SIZE; i++) buffer->uuid[i] = cpu_to_be64(mdev->ldev->md.uuid[i]); buffer->flags = cpu_to_be32(mdev->ldev->md.flags); - buffer->magic = cpu_to_be32(DRBD_MD_MAGIC); + buffer->magic = cpu_to_be32(DRBD_MD_MAGIC_84_UNCLEAN); buffer->md_size_sect = cpu_to_be32(mdev->ldev->md.md_size_sect); buffer->al_offset = cpu_to_be32(mdev->ldev->md.al_offset); @@ -3885,7 +2899,7 @@ void drbd_md_sync(struct drbd_conf *mdev) D_ASSERT(drbd_md_ss__(mdev, mdev->ldev) == mdev->ldev->md.md_offset); sector = mdev->ldev->md.md_offset; - if (!drbd_md_sync_page_io(mdev, mdev->ldev, sector, WRITE)) { + if (drbd_md_sync_page_io(mdev, mdev->ldev, sector, WRITE)) { /* this was a try anyways ... */ dev_err(DEV, "meta data update failed!\n"); drbd_chk_io_error(mdev, 1, DRBD_META_IO_ERROR); @@ -3906,11 +2920,12 @@ out: * @bdev: Device from which the meta data should be read in. * * Return 0 (NO_ERROR) on success, and an enum drbd_ret_code in case - * something goes wrong. Currently only: ERR_IO_MD_DISK, ERR_MD_INVALID. + * something goes wrong. */ int drbd_md_read(struct drbd_conf *mdev, struct drbd_backing_dev *bdev) { struct meta_data_on_disk *buffer; + u32 magic, flags; int i, rv = NO_ERROR; if (!get_ldev_if_state(mdev, D_ATTACHING)) @@ -3920,7 +2935,7 @@ int drbd_md_read(struct drbd_conf *mdev, struct drbd_backing_dev *bdev) if (!buffer) goto out; - if (!drbd_md_sync_page_io(mdev, bdev, bdev->md.md_offset, READ)) { + if (drbd_md_sync_page_io(mdev, bdev, bdev->md.md_offset, READ)) { /* NOTE: can't do normal error processing here as this is called BEFORE disk is attached */ dev_err(DEV, "Error while reading metadata.\n"); @@ -3928,8 +2943,20 @@ int drbd_md_read(struct drbd_conf *mdev, struct drbd_backing_dev *bdev) goto err; } - if (be32_to_cpu(buffer->magic) != DRBD_MD_MAGIC) { - dev_err(DEV, "Error while reading metadata, magic not found.\n"); + magic = be32_to_cpu(buffer->magic); + flags = be32_to_cpu(buffer->flags); + if (magic == DRBD_MD_MAGIC_84_UNCLEAN || + (magic == DRBD_MD_MAGIC_08 && !(flags & MDF_AL_CLEAN))) { + /* btw: that's Activity Log clean, not "all" clean. */ + dev_err(DEV, "Found unclean meta data. Did you \"drbdadm apply-al\"?\n"); + rv = ERR_MD_UNCLEAN; + goto err; + } + if (magic != DRBD_MD_MAGIC_08) { + if (magic == DRBD_MD_MAGIC_07) + dev_err(DEV, "Found old (0.7) meta data magic. Did you \"drbdadm create-md\"?\n"); + else + dev_err(DEV, "Meta data magic not found. Did you \"drbdadm create-md\"?\n"); rv = ERR_MD_INVALID; goto err; } @@ -3963,20 +2990,16 @@ int drbd_md_read(struct drbd_conf *mdev, struct drbd_backing_dev *bdev) for (i = UI_CURRENT; i < UI_SIZE; i++) bdev->md.uuid[i] = be64_to_cpu(buffer->uuid[i]); bdev->md.flags = be32_to_cpu(buffer->flags); - mdev->sync_conf.al_extents = be32_to_cpu(buffer->al_nr_extents); bdev->md.device_uuid = be64_to_cpu(buffer->device_uuid); - spin_lock_irq(&mdev->req_lock); + spin_lock_irq(&mdev->tconn->req_lock); if (mdev->state.conn < C_CONNECTED) { unsigned int peer; peer = be32_to_cpu(buffer->la_peer_max_bio_size); peer = max(peer, DRBD_MAX_BIO_SIZE_SAFE); mdev->peer_max_bio_size = peer; } - spin_unlock_irq(&mdev->req_lock); - - if (mdev->sync_conf.al_extents < 7) - mdev->sync_conf.al_extents = 127; + spin_unlock_irq(&mdev->tconn->req_lock); err: drbd_md_put_buffer(mdev); @@ -4011,7 +3034,7 @@ void drbd_md_mark_dirty(struct drbd_conf *mdev) } #endif -static void drbd_uuid_move_history(struct drbd_conf *mdev) __must_hold(local) +void drbd_uuid_move_history(struct drbd_conf *mdev) __must_hold(local) { int i; @@ -4019,7 +3042,7 @@ static void drbd_uuid_move_history(struct drbd_conf *mdev) __must_hold(local) mdev->ldev->md.uuid[i+1] = mdev->ldev->md.uuid[i]; } -void _drbd_uuid_set(struct drbd_conf *mdev, int idx, u64 val) __must_hold(local) +void __drbd_uuid_set(struct drbd_conf *mdev, int idx, u64 val) __must_hold(local) { if (idx == UI_CURRENT) { if (mdev->state.role == R_PRIMARY) @@ -4034,14 +3057,24 @@ void _drbd_uuid_set(struct drbd_conf *mdev, int idx, u64 val) __must_hold(local) drbd_md_mark_dirty(mdev); } +void _drbd_uuid_set(struct drbd_conf *mdev, int idx, u64 val) __must_hold(local) +{ + unsigned long flags; + spin_lock_irqsave(&mdev->ldev->md.uuid_lock, flags); + __drbd_uuid_set(mdev, idx, val); + spin_unlock_irqrestore(&mdev->ldev->md.uuid_lock, flags); +} void drbd_uuid_set(struct drbd_conf *mdev, int idx, u64 val) __must_hold(local) { + unsigned long flags; + spin_lock_irqsave(&mdev->ldev->md.uuid_lock, flags); if (mdev->ldev->md.uuid[idx]) { drbd_uuid_move_history(mdev); mdev->ldev->md.uuid[UI_HISTORY_START] = mdev->ldev->md.uuid[idx]; } - _drbd_uuid_set(mdev, idx, val); + __drbd_uuid_set(mdev, idx, val); + spin_unlock_irqrestore(&mdev->ldev->md.uuid_lock, flags); } /** @@ -4054,15 +3087,20 @@ void drbd_uuid_set(struct drbd_conf *mdev, int idx, u64 val) __must_hold(local) void drbd_uuid_new_current(struct drbd_conf *mdev) __must_hold(local) { u64 val; - unsigned long long bm_uuid = mdev->ldev->md.uuid[UI_BITMAP]; + unsigned long long bm_uuid; + + get_random_bytes(&val, sizeof(u64)); + + spin_lock_irq(&mdev->ldev->md.uuid_lock); + bm_uuid = mdev->ldev->md.uuid[UI_BITMAP]; if (bm_uuid) dev_warn(DEV, "bm UUID was already set: %llX\n", bm_uuid); mdev->ldev->md.uuid[UI_BITMAP] = mdev->ldev->md.uuid[UI_CURRENT]; + __drbd_uuid_set(mdev, UI_CURRENT, val); + spin_unlock_irq(&mdev->ldev->md.uuid_lock); - get_random_bytes(&val, sizeof(u64)); - _drbd_uuid_set(mdev, UI_CURRENT, val); drbd_print_uuids(mdev, "new current UUID"); /* get it to stable storage _now_ */ drbd_md_sync(mdev); @@ -4070,9 +3108,11 @@ void drbd_uuid_new_current(struct drbd_conf *mdev) __must_hold(local) void drbd_uuid_set_bm(struct drbd_conf *mdev, u64 val) __must_hold(local) { + unsigned long flags; if (mdev->ldev->md.uuid[UI_BITMAP] == 0 && val == 0) return; + spin_lock_irqsave(&mdev->ldev->md.uuid_lock, flags); if (val == 0) { drbd_uuid_move_history(mdev); mdev->ldev->md.uuid[UI_HISTORY_START] = mdev->ldev->md.uuid[UI_BITMAP]; @@ -4084,6 +3124,8 @@ void drbd_uuid_set_bm(struct drbd_conf *mdev, u64 val) __must_hold(local) mdev->ldev->md.uuid[UI_BITMAP] = val & ~((u64)1); } + spin_unlock_irqrestore(&mdev->ldev->md.uuid_lock, flags); + drbd_md_mark_dirty(mdev); } @@ -4135,9 +3177,10 @@ int drbd_bmio_clear_n_write(struct drbd_conf *mdev) return rv; } -static int w_bitmap_io(struct drbd_conf *mdev, struct drbd_work *w, int unused) +static int w_bitmap_io(struct drbd_work *w, int unused) { struct bm_io_work *work = container_of(w, struct bm_io_work, w); + struct drbd_conf *mdev = w->mdev; int rv = -EIO; D_ASSERT(atomic_read(&mdev->ap_bio_cnt) == 0); @@ -4149,8 +3192,7 @@ static int w_bitmap_io(struct drbd_conf *mdev, struct drbd_work *w, int unused) put_ldev(mdev); } - clear_bit(BITMAP_IO, &mdev->flags); - smp_mb__after_clear_bit(); + clear_bit_unlock(BITMAP_IO, &mdev->flags); wake_up(&mdev->misc_wait); if (work->done) @@ -4160,7 +3202,7 @@ static int w_bitmap_io(struct drbd_conf *mdev, struct drbd_work *w, int unused) work->why = NULL; work->flags = 0; - return 1; + return 0; } void drbd_ldev_destroy(struct drbd_conf *mdev) @@ -4173,29 +3215,51 @@ void drbd_ldev_destroy(struct drbd_conf *mdev) drbd_free_bc(mdev->ldev); mdev->ldev = NULL;); - if (mdev->md_io_tmpp) { - __free_page(mdev->md_io_tmpp); - mdev->md_io_tmpp = NULL; - } clear_bit(GO_DISKLESS, &mdev->flags); } -static int w_go_diskless(struct drbd_conf *mdev, struct drbd_work *w, int unused) +static int w_go_diskless(struct drbd_work *w, int unused) { + struct drbd_conf *mdev = w->mdev; + D_ASSERT(mdev->state.disk == D_FAILED); /* we cannot assert local_cnt == 0 here, as get_ldev_if_state will * inc/dec it frequently. Once we are D_DISKLESS, no one will touch * the protected members anymore, though, so once put_ldev reaches zero * again, it will be safe to free them. */ + + /* Try to write changed bitmap pages, read errors may have just + * set some bits outside the area covered by the activity log. + * + * If we have an IO error during the bitmap writeout, + * we will want a full sync next time, just in case. + * (Do we want a specific meta data flag for this?) + * + * If that does not make it to stable storage either, + * we cannot do anything about that anymore. + * + * We still need to check if both bitmap and ldev are present, we may + * end up here after a failed attach, before ldev was even assigned. + */ + if (mdev->bitmap && mdev->ldev) { + if (drbd_bitmap_io_from_worker(mdev, drbd_bm_write, + "detach", BM_LOCKED_MASK)) { + if (test_bit(WAS_READ_ERROR, &mdev->flags)) { + drbd_md_set_flag(mdev, MDF_FULL_SYNC); + drbd_md_sync(mdev); + } + } + } + drbd_force_state(mdev, NS(disk, D_DISKLESS)); - return 1; + return 0; } void drbd_go_diskless(struct drbd_conf *mdev) { D_ASSERT(mdev->state.disk == D_FAILED); if (!test_and_set_bit(GO_DISKLESS, &mdev->flags)) - drbd_queue_work(&mdev->data.work, &mdev->go_diskless); + drbd_queue_work(&mdev->tconn->sender_work, &mdev->go_diskless); } /** @@ -4215,7 +3279,7 @@ void drbd_queue_bitmap_io(struct drbd_conf *mdev, void (*done)(struct drbd_conf *, int), char *why, enum bm_flag flags) { - D_ASSERT(current == mdev->worker.task); + D_ASSERT(current == mdev->tconn->worker.task); D_ASSERT(!test_bit(BITMAP_IO_QUEUED, &mdev->flags)); D_ASSERT(!test_bit(BITMAP_IO, &mdev->flags)); @@ -4229,13 +3293,13 @@ void drbd_queue_bitmap_io(struct drbd_conf *mdev, mdev->bm_io_work.why = why; mdev->bm_io_work.flags = flags; - spin_lock_irq(&mdev->req_lock); + spin_lock_irq(&mdev->tconn->req_lock); set_bit(BITMAP_IO, &mdev->flags); if (atomic_read(&mdev->ap_bio_cnt) == 0) { if (!test_and_set_bit(BITMAP_IO_QUEUED, &mdev->flags)) - drbd_queue_work(&mdev->data.work, &mdev->bm_io_work.w); + drbd_queue_work(&mdev->tconn->sender_work, &mdev->bm_io_work.w); } - spin_unlock_irq(&mdev->req_lock); + spin_unlock_irq(&mdev->tconn->req_lock); } /** @@ -4252,7 +3316,7 @@ int drbd_bitmap_io(struct drbd_conf *mdev, int (*io_fn)(struct drbd_conf *), { int rv; - D_ASSERT(current != mdev->worker.task); + D_ASSERT(current != mdev->tconn->worker.task); if ((flags & BM_LOCKED_SET_ALLOWED) == 0) drbd_suspend_io(mdev); @@ -4291,18 +3355,127 @@ static void md_sync_timer_fn(unsigned long data) { struct drbd_conf *mdev = (struct drbd_conf *) data; - drbd_queue_work_front(&mdev->data.work, &mdev->md_sync_work); + /* must not double-queue! */ + if (list_empty(&mdev->md_sync_work.list)) + drbd_queue_work_front(&mdev->tconn->sender_work, &mdev->md_sync_work); } -static int w_md_sync(struct drbd_conf *mdev, struct drbd_work *w, int unused) +static int w_md_sync(struct drbd_work *w, int unused) { + struct drbd_conf *mdev = w->mdev; + dev_warn(DEV, "md_sync_timer expired! Worker calls drbd_md_sync().\n"); #ifdef DEBUG dev_warn(DEV, "last md_mark_dirty: %s:%u\n", mdev->last_md_mark_dirty.func, mdev->last_md_mark_dirty.line); #endif drbd_md_sync(mdev); - return 1; + return 0; +} + +const char *cmdname(enum drbd_packet cmd) +{ + /* THINK may need to become several global tables + * when we want to support more than + * one PRO_VERSION */ + static const char *cmdnames[] = { + [P_DATA] = "Data", + [P_DATA_REPLY] = "DataReply", + [P_RS_DATA_REPLY] = "RSDataReply", + [P_BARRIER] = "Barrier", + [P_BITMAP] = "ReportBitMap", + [P_BECOME_SYNC_TARGET] = "BecomeSyncTarget", + [P_BECOME_SYNC_SOURCE] = "BecomeSyncSource", + [P_UNPLUG_REMOTE] = "UnplugRemote", + [P_DATA_REQUEST] = "DataRequest", + [P_RS_DATA_REQUEST] = "RSDataRequest", + [P_SYNC_PARAM] = "SyncParam", + [P_SYNC_PARAM89] = "SyncParam89", + [P_PROTOCOL] = "ReportProtocol", + [P_UUIDS] = "ReportUUIDs", + [P_SIZES] = "ReportSizes", + [P_STATE] = "ReportState", + [P_SYNC_UUID] = "ReportSyncUUID", + [P_AUTH_CHALLENGE] = "AuthChallenge", + [P_AUTH_RESPONSE] = "AuthResponse", + [P_PING] = "Ping", + [P_PING_ACK] = "PingAck", + [P_RECV_ACK] = "RecvAck", + [P_WRITE_ACK] = "WriteAck", + [P_RS_WRITE_ACK] = "RSWriteAck", + [P_SUPERSEDED] = "Superseded", + [P_NEG_ACK] = "NegAck", + [P_NEG_DREPLY] = "NegDReply", + [P_NEG_RS_DREPLY] = "NegRSDReply", + [P_BARRIER_ACK] = "BarrierAck", + [P_STATE_CHG_REQ] = "StateChgRequest", + [P_STATE_CHG_REPLY] = "StateChgReply", + [P_OV_REQUEST] = "OVRequest", + [P_OV_REPLY] = "OVReply", + [P_OV_RESULT] = "OVResult", + [P_CSUM_RS_REQUEST] = "CsumRSRequest", + [P_RS_IS_IN_SYNC] = "CsumRSIsInSync", + [P_COMPRESSED_BITMAP] = "CBitmap", + [P_DELAY_PROBE] = "DelayProbe", + [P_OUT_OF_SYNC] = "OutOfSync", + [P_RETRY_WRITE] = "RetryWrite", + [P_RS_CANCEL] = "RSCancel", + [P_CONN_ST_CHG_REQ] = "conn_st_chg_req", + [P_CONN_ST_CHG_REPLY] = "conn_st_chg_reply", + [P_RETRY_WRITE] = "retry_write", + [P_PROTOCOL_UPDATE] = "protocol_update", + + /* enum drbd_packet, but not commands - obsoleted flags: + * P_MAY_IGNORE + * P_MAX_OPT_CMD + */ + }; + + /* too big for the array: 0xfffX */ + if (cmd == P_INITIAL_META) + return "InitialMeta"; + if (cmd == P_INITIAL_DATA) + return "InitialData"; + if (cmd == P_CONNECTION_FEATURES) + return "ConnectionFeatures"; + if (cmd >= ARRAY_SIZE(cmdnames)) + return "Unknown"; + return cmdnames[cmd]; +} + +/** + * drbd_wait_misc - wait for a request to make progress + * @mdev: device associated with the request + * @i: the struct drbd_interval embedded in struct drbd_request or + * struct drbd_peer_request + */ +int drbd_wait_misc(struct drbd_conf *mdev, struct drbd_interval *i) +{ + struct net_conf *nc; + DEFINE_WAIT(wait); + long timeout; + + rcu_read_lock(); + nc = rcu_dereference(mdev->tconn->net_conf); + if (!nc) { + rcu_read_unlock(); + return -ETIMEDOUT; + } + timeout = nc->ko_count ? nc->timeout * HZ / 10 * nc->ko_count : MAX_SCHEDULE_TIMEOUT; + rcu_read_unlock(); + + /* Indicate to wake up mdev->misc_wait on progress. */ + i->waiting = true; + prepare_to_wait(&mdev->misc_wait, &wait, TASK_INTERRUPTIBLE); + spin_unlock_irq(&mdev->tconn->req_lock); + timeout = schedule_timeout(timeout); + finish_wait(&mdev->misc_wait, &wait); + spin_lock_irq(&mdev->tconn->req_lock); + if (!timeout || mdev->state.conn < C_CONNECTED) + return -ETIMEDOUT; + if (signal_pending(current)) + return -ERESTARTSYS; + return 0; } #ifdef CONFIG_DRBD_FAULT_INJECTION diff --git a/drivers/block/drbd/drbd_nl.c b/drivers/block/drbd/drbd_nl.c index edb490aad8b4..2af26fc95280 100644 --- a/drivers/block/drbd/drbd_nl.c +++ b/drivers/block/drbd/drbd_nl.c @@ -29,159 +29,317 @@ #include <linux/fs.h> #include <linux/file.h> #include <linux/slab.h> -#include <linux/connector.h> #include <linux/blkpg.h> #include <linux/cpumask.h> #include "drbd_int.h" #include "drbd_req.h" #include "drbd_wrappers.h" #include <asm/unaligned.h> -#include <linux/drbd_tag_magic.h> #include <linux/drbd_limits.h> -#include <linux/compiler.h> #include <linux/kthread.h> -static unsigned short *tl_add_blob(unsigned short *, enum drbd_tags, const void *, int); -static unsigned short *tl_add_str(unsigned short *, enum drbd_tags, const char *); -static unsigned short *tl_add_int(unsigned short *, enum drbd_tags, const void *); - -/* see get_sb_bdev and bd_claim */ +#include <net/genetlink.h> + +/* .doit */ +// int drbd_adm_create_resource(struct sk_buff *skb, struct genl_info *info); +// int drbd_adm_delete_resource(struct sk_buff *skb, struct genl_info *info); + +int drbd_adm_add_minor(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_delete_minor(struct sk_buff *skb, struct genl_info *info); + +int drbd_adm_new_resource(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_del_resource(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_down(struct sk_buff *skb, struct genl_info *info); + +int drbd_adm_set_role(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_attach(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_disk_opts(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_detach(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_connect(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_net_opts(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_resize(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_start_ov(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_new_c_uuid(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_disconnect(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_invalidate(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_invalidate_peer(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_pause_sync(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_resume_sync(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_suspend_io(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_resume_io(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_outdate(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_resource_opts(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_get_status(struct sk_buff *skb, struct genl_info *info); +int drbd_adm_get_timeout_type(struct sk_buff *skb, struct genl_info *info); +/* .dumpit */ +int drbd_adm_get_status_all(struct sk_buff *skb, struct netlink_callback *cb); + +#include <linux/drbd_genl_api.h> +#include "drbd_nla.h" +#include <linux/genl_magic_func.h> + +/* used blkdev_get_by_path, to claim our meta data device(s) */ static char *drbd_m_holder = "Hands off! this is DRBD's meta data device."; -/* Generate the tag_list to struct functions */ -#define NL_PACKET(name, number, fields) \ -static int name ## _from_tags(struct drbd_conf *mdev, \ - unsigned short *tags, struct name *arg) __attribute__ ((unused)); \ -static int name ## _from_tags(struct drbd_conf *mdev, \ - unsigned short *tags, struct name *arg) \ -{ \ - int tag; \ - int dlen; \ - \ - while ((tag = get_unaligned(tags++)) != TT_END) { \ - dlen = get_unaligned(tags++); \ - switch (tag_number(tag)) { \ - fields \ - default: \ - if (tag & T_MANDATORY) { \ - dev_err(DEV, "Unknown tag: %d\n", tag_number(tag)); \ - return 0; \ - } \ - } \ - tags = (unsigned short *)((char *)tags + dlen); \ - } \ - return 1; \ -} -#define NL_INTEGER(pn, pr, member) \ - case pn: /* D_ASSERT( tag_type(tag) == TT_INTEGER ); */ \ - arg->member = get_unaligned((int *)(tags)); \ - break; -#define NL_INT64(pn, pr, member) \ - case pn: /* D_ASSERT( tag_type(tag) == TT_INT64 ); */ \ - arg->member = get_unaligned((u64 *)(tags)); \ +/* Configuration is strictly serialized, because generic netlink message + * processing is strictly serialized by the genl_lock(). + * Which means we can use one static global drbd_config_context struct. + */ +static struct drbd_config_context { + /* assigned from drbd_genlmsghdr */ + unsigned int minor; + /* assigned from request attributes, if present */ + unsigned int volume; +#define VOLUME_UNSPECIFIED (-1U) + /* pointer into the request skb, + * limited lifetime! */ + char *resource_name; + struct nlattr *my_addr; + struct nlattr *peer_addr; + + /* reply buffer */ + struct sk_buff *reply_skb; + /* pointer into reply buffer */ + struct drbd_genlmsghdr *reply_dh; + /* resolved from attributes, if possible */ + struct drbd_conf *mdev; + struct drbd_tconn *tconn; +} adm_ctx; + +static void drbd_adm_send_reply(struct sk_buff *skb, struct genl_info *info) +{ + genlmsg_end(skb, genlmsg_data(nlmsg_data(nlmsg_hdr(skb)))); + if (genlmsg_reply(skb, info)) + printk(KERN_ERR "drbd: error sending genl reply\n"); +} + +/* Used on a fresh "drbd_adm_prepare"d reply_skb, this cannot fail: The only + * reason it could fail was no space in skb, and there are 4k available. */ +int drbd_msg_put_info(const char *info) +{ + struct sk_buff *skb = adm_ctx.reply_skb; + struct nlattr *nla; + int err = -EMSGSIZE; + + if (!info || !info[0]) + return 0; + + nla = nla_nest_start(skb, DRBD_NLA_CFG_REPLY); + if (!nla) + return err; + + err = nla_put_string(skb, T_info_text, info); + if (err) { + nla_nest_cancel(skb, nla); + return err; + } else + nla_nest_end(skb, nla); + return 0; +} + +/* This would be a good candidate for a "pre_doit" hook, + * and per-family private info->pointers. + * But we need to stay compatible with older kernels. + * If it returns successfully, adm_ctx members are valid. + */ +#define DRBD_ADM_NEED_MINOR 1 +#define DRBD_ADM_NEED_RESOURCE 2 +#define DRBD_ADM_NEED_CONNECTION 4 +static int drbd_adm_prepare(struct sk_buff *skb, struct genl_info *info, + unsigned flags) +{ + struct drbd_genlmsghdr *d_in = info->userhdr; + const u8 cmd = info->genlhdr->cmd; + int err; + + memset(&adm_ctx, 0, sizeof(adm_ctx)); + + /* genl_rcv_msg only checks for CAP_NET_ADMIN on "GENL_ADMIN_PERM" :( */ + if (cmd != DRBD_ADM_GET_STATUS && !capable(CAP_NET_ADMIN)) + return -EPERM; + + adm_ctx.reply_skb = genlmsg_new(NLMSG_GOODSIZE, GFP_KERNEL); + if (!adm_ctx.reply_skb) { + err = -ENOMEM; + goto fail; + } + + adm_ctx.reply_dh = genlmsg_put_reply(adm_ctx.reply_skb, + info, &drbd_genl_family, 0, cmd); + /* put of a few bytes into a fresh skb of >= 4k will always succeed. + * but anyways */ + if (!adm_ctx.reply_dh) { + err = -ENOMEM; + goto fail; + } + + adm_ctx.reply_dh->minor = d_in->minor; + adm_ctx.reply_dh->ret_code = NO_ERROR; + + adm_ctx.volume = VOLUME_UNSPECIFIED; + if (info->attrs[DRBD_NLA_CFG_CONTEXT]) { + struct nlattr *nla; + /* parse and validate only */ + err = drbd_cfg_context_from_attrs(NULL, info); + if (err) + goto fail; + + /* It was present, and valid, + * copy it over to the reply skb. */ + err = nla_put_nohdr(adm_ctx.reply_skb, + info->attrs[DRBD_NLA_CFG_CONTEXT]->nla_len, + info->attrs[DRBD_NLA_CFG_CONTEXT]); + if (err) + goto fail; + + /* and assign stuff to the global adm_ctx */ + nla = nested_attr_tb[__nla_type(T_ctx_volume)]; + if (nla) + adm_ctx.volume = nla_get_u32(nla); + nla = nested_attr_tb[__nla_type(T_ctx_resource_name)]; + if (nla) + adm_ctx.resource_name = nla_data(nla); + adm_ctx.my_addr = nested_attr_tb[__nla_type(T_ctx_my_addr)]; + adm_ctx.peer_addr = nested_attr_tb[__nla_type(T_ctx_peer_addr)]; + if ((adm_ctx.my_addr && + nla_len(adm_ctx.my_addr) > sizeof(adm_ctx.tconn->my_addr)) || + (adm_ctx.peer_addr && + nla_len(adm_ctx.peer_addr) > sizeof(adm_ctx.tconn->peer_addr))) { + err = -EINVAL; + goto fail; + } + } + + adm_ctx.minor = d_in->minor; + adm_ctx.mdev = minor_to_mdev(d_in->minor); + adm_ctx.tconn = conn_get_by_name(adm_ctx.resource_name); + + if (!adm_ctx.mdev && (flags & DRBD_ADM_NEED_MINOR)) { + drbd_msg_put_info("unknown minor"); + return ERR_MINOR_INVALID; + } + if (!adm_ctx.tconn && (flags & DRBD_ADM_NEED_RESOURCE)) { + drbd_msg_put_info("unknown resource"); + return ERR_INVALID_REQUEST; + } + + if (flags & DRBD_ADM_NEED_CONNECTION) { + if (adm_ctx.tconn && !(flags & DRBD_ADM_NEED_RESOURCE)) { + drbd_msg_put_info("no resource name expected"); + return ERR_INVALID_REQUEST; + } + if (adm_ctx.mdev) { + drbd_msg_put_info("no minor number expected"); + return ERR_INVALID_REQUEST; + } + if (adm_ctx.my_addr && adm_ctx.peer_addr) + adm_ctx.tconn = conn_get_by_addrs(nla_data(adm_ctx.my_addr), + nla_len(adm_ctx.my_addr), + nla_data(adm_ctx.peer_addr), + nla_len(adm_ctx.peer_addr)); + if (!adm_ctx.tconn) { + drbd_msg_put_info("unknown connection"); + return ERR_INVALID_REQUEST; + } + } + + /* some more paranoia, if the request was over-determined */ + if (adm_ctx.mdev && adm_ctx.tconn && + adm_ctx.mdev->tconn != adm_ctx.tconn) { + pr_warning("request: minor=%u, resource=%s; but that minor belongs to connection %s\n", + adm_ctx.minor, adm_ctx.resource_name, + adm_ctx.mdev->tconn->name); + drbd_msg_put_info("minor exists in different resource"); + return ERR_INVALID_REQUEST; + } + if (adm_ctx.mdev && + adm_ctx.volume != VOLUME_UNSPECIFIED && + adm_ctx.volume != adm_ctx.mdev->vnr) { + pr_warning("request: minor=%u, volume=%u; but that minor is volume %u in %s\n", + adm_ctx.minor, adm_ctx.volume, + adm_ctx.mdev->vnr, adm_ctx.mdev->tconn->name); + drbd_msg_put_info("minor exists as different volume"); + return ERR_INVALID_REQUEST; + } + + return NO_ERROR; + +fail: + nlmsg_free(adm_ctx.reply_skb); + adm_ctx.reply_skb = NULL; + return err; +} + +static int drbd_adm_finish(struct genl_info *info, int retcode) +{ + if (adm_ctx.tconn) { + kref_put(&adm_ctx.tconn->kref, &conn_destroy); + adm_ctx.tconn = NULL; + } + + if (!adm_ctx.reply_skb) + return -ENOMEM; + + adm_ctx.reply_dh->ret_code = retcode; + drbd_adm_send_reply(adm_ctx.reply_skb, info); + return 0; +} + +static void setup_khelper_env(struct drbd_tconn *tconn, char **envp) +{ + char *afs; + + /* FIXME: A future version will not allow this case. */ + if (tconn->my_addr_len == 0 || tconn->peer_addr_len == 0) + return; + + switch (((struct sockaddr *)&tconn->peer_addr)->sa_family) { + case AF_INET6: + afs = "ipv6"; + snprintf(envp[4], 60, "DRBD_PEER_ADDRESS=%pI6", + &((struct sockaddr_in6 *)&tconn->peer_addr)->sin6_addr); break; -#define NL_BIT(pn, pr, member) \ - case pn: /* D_ASSERT( tag_type(tag) == TT_BIT ); */ \ - arg->member = *(char *)(tags) ? 1 : 0; \ + case AF_INET: + afs = "ipv4"; + snprintf(envp[4], 60, "DRBD_PEER_ADDRESS=%pI4", + &((struct sockaddr_in *)&tconn->peer_addr)->sin_addr); break; -#define NL_STRING(pn, pr, member, len) \ - case pn: /* D_ASSERT( tag_type(tag) == TT_STRING ); */ \ - if (dlen > len) { \ - dev_err(DEV, "arg too long: %s (%u wanted, max len: %u bytes)\n", \ - #member, dlen, (unsigned int)len); \ - return 0; \ - } \ - arg->member ## _len = dlen; \ - memcpy(arg->member, tags, min_t(size_t, dlen, len)); \ - break; -#include <linux/drbd_nl.h> - -/* Generate the struct to tag_list functions */ -#define NL_PACKET(name, number, fields) \ -static unsigned short* \ -name ## _to_tags(struct drbd_conf *mdev, \ - struct name *arg, unsigned short *tags) __attribute__ ((unused)); \ -static unsigned short* \ -name ## _to_tags(struct drbd_conf *mdev, \ - struct name *arg, unsigned short *tags) \ -{ \ - fields \ - return tags; \ -} - -#define NL_INTEGER(pn, pr, member) \ - put_unaligned(pn | pr | TT_INTEGER, tags++); \ - put_unaligned(sizeof(int), tags++); \ - put_unaligned(arg->member, (int *)tags); \ - tags = (unsigned short *)((char *)tags+sizeof(int)); -#define NL_INT64(pn, pr, member) \ - put_unaligned(pn | pr | TT_INT64, tags++); \ - put_unaligned(sizeof(u64), tags++); \ - put_unaligned(arg->member, (u64 *)tags); \ - tags = (unsigned short *)((char *)tags+sizeof(u64)); -#define NL_BIT(pn, pr, member) \ - put_unaligned(pn | pr | TT_BIT, tags++); \ - put_unaligned(sizeof(char), tags++); \ - *(char *)tags = arg->member; \ - tags = (unsigned short *)((char *)tags+sizeof(char)); -#define NL_STRING(pn, pr, member, len) \ - put_unaligned(pn | pr | TT_STRING, tags++); \ - put_unaligned(arg->member ## _len, tags++); \ - memcpy(tags, arg->member, arg->member ## _len); \ - tags = (unsigned short *)((char *)tags + arg->member ## _len); -#include <linux/drbd_nl.h> - -void drbd_bcast_ev_helper(struct drbd_conf *mdev, char *helper_name); -void drbd_nl_send_reply(struct cn_msg *, int); + default: + afs = "ssocks"; + snprintf(envp[4], 60, "DRBD_PEER_ADDRESS=%pI4", + &((struct sockaddr_in *)&tconn->peer_addr)->sin_addr); + } + snprintf(envp[3], 20, "DRBD_PEER_AF=%s", afs); +} int drbd_khelper(struct drbd_conf *mdev, char *cmd) { char *envp[] = { "HOME=/", "TERM=linux", "PATH=/sbin:/usr/sbin:/bin:/usr/bin", - NULL, /* Will be set to address family */ - NULL, /* Will be set to address */ + (char[20]) { }, /* address family */ + (char[60]) { }, /* address */ NULL }; - - char mb[12], af[20], ad[60], *afs; + char mb[12]; char *argv[] = {usermode_helper, cmd, mb, NULL }; + struct drbd_tconn *tconn = mdev->tconn; + struct sib_info sib; int ret; - if (current == mdev->worker.task) - set_bit(CALLBACK_PENDING, &mdev->flags); + if (current == tconn->worker.task) + set_bit(CALLBACK_PENDING, &tconn->flags); snprintf(mb, 12, "minor-%d", mdev_to_minor(mdev)); - - if (get_net_conf(mdev)) { - switch (((struct sockaddr *)mdev->net_conf->peer_addr)->sa_family) { - case AF_INET6: - afs = "ipv6"; - snprintf(ad, 60, "DRBD_PEER_ADDRESS=%pI6", - &((struct sockaddr_in6 *)mdev->net_conf->peer_addr)->sin6_addr); - break; - case AF_INET: - afs = "ipv4"; - snprintf(ad, 60, "DRBD_PEER_ADDRESS=%pI4", - &((struct sockaddr_in *)mdev->net_conf->peer_addr)->sin_addr); - break; - default: - afs = "ssocks"; - snprintf(ad, 60, "DRBD_PEER_ADDRESS=%pI4", - &((struct sockaddr_in *)mdev->net_conf->peer_addr)->sin_addr); - } - snprintf(af, 20, "DRBD_PEER_AF=%s", afs); - envp[3]=af; - envp[4]=ad; - put_net_conf(mdev); - } + setup_khelper_env(tconn, envp); /* The helper may take some time. * write out any unsynced meta data changes now */ drbd_md_sync(mdev); dev_info(DEV, "helper command: %s %s %s\n", usermode_helper, cmd, mb); - - drbd_bcast_ev_helper(mdev, cmd); + sib.sib_reason = SIB_HELPER_PRE; + sib.helper_name = cmd; + drbd_bcast_event(mdev, &sib); ret = call_usermodehelper(usermode_helper, argv, envp, UMH_WAIT_PROC); if (ret) dev_warn(DEV, "helper command: %s %s %s exit code %u (0x%x)\n", @@ -191,9 +349,46 @@ int drbd_khelper(struct drbd_conf *mdev, char *cmd) dev_info(DEV, "helper command: %s %s %s exit code %u (0x%x)\n", usermode_helper, cmd, mb, (ret >> 8) & 0xff, ret); + sib.sib_reason = SIB_HELPER_POST; + sib.helper_exit_code = ret; + drbd_bcast_event(mdev, &sib); + + if (current == tconn->worker.task) + clear_bit(CALLBACK_PENDING, &tconn->flags); + + if (ret < 0) /* Ignore any ERRNOs we got. */ + ret = 0; + + return ret; +} + +int conn_khelper(struct drbd_tconn *tconn, char *cmd) +{ + char *envp[] = { "HOME=/", + "TERM=linux", + "PATH=/sbin:/usr/sbin:/bin:/usr/bin", + (char[20]) { }, /* address family */ + (char[60]) { }, /* address */ + NULL }; + char *argv[] = {usermode_helper, cmd, tconn->name, NULL }; + int ret; + + setup_khelper_env(tconn, envp); + conn_md_sync(tconn); - if (current == mdev->worker.task) - clear_bit(CALLBACK_PENDING, &mdev->flags); + conn_info(tconn, "helper command: %s %s %s\n", usermode_helper, cmd, tconn->name); + /* TODO: conn_bcast_event() ?? */ + + ret = call_usermodehelper(usermode_helper, argv, envp, UMH_WAIT_PROC); + if (ret) + conn_warn(tconn, "helper command: %s %s %s exit code %u (0x%x)\n", + usermode_helper, cmd, tconn->name, + (ret >> 8) & 0xff, ret); + else + conn_info(tconn, "helper command: %s %s %s exit code %u (0x%x)\n", + usermode_helper, cmd, tconn->name, + (ret >> 8) & 0xff, ret); + /* TODO: conn_bcast_event() ?? */ if (ret < 0) /* Ignore any ERRNOs we got. */ ret = 0; @@ -201,116 +396,129 @@ int drbd_khelper(struct drbd_conf *mdev, char *cmd) return ret; } -enum drbd_disk_state drbd_try_outdate_peer(struct drbd_conf *mdev) +static enum drbd_fencing_p highest_fencing_policy(struct drbd_tconn *tconn) { + enum drbd_fencing_p fp = FP_NOT_AVAIL; + struct drbd_conf *mdev; + int vnr; + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) { + if (get_ldev_if_state(mdev, D_CONSISTENT)) { + fp = max_t(enum drbd_fencing_p, fp, + rcu_dereference(mdev->ldev->disk_conf)->fencing); + put_ldev(mdev); + } + } + rcu_read_unlock(); + + return fp; +} + +bool conn_try_outdate_peer(struct drbd_tconn *tconn) +{ + union drbd_state mask = { }; + union drbd_state val = { }; + enum drbd_fencing_p fp; char *ex_to_string; int r; - enum drbd_disk_state nps; - enum drbd_fencing_p fp; - D_ASSERT(mdev->state.pdsk == D_UNKNOWN); + if (tconn->cstate >= C_WF_REPORT_PARAMS) { + conn_err(tconn, "Expected cstate < C_WF_REPORT_PARAMS\n"); + return false; + } - if (get_ldev_if_state(mdev, D_CONSISTENT)) { - fp = mdev->ldev->dc.fencing; - put_ldev(mdev); - } else { - dev_warn(DEV, "Not fencing peer, I'm not even Consistent myself.\n"); - nps = mdev->state.pdsk; + fp = highest_fencing_policy(tconn); + switch (fp) { + case FP_NOT_AVAIL: + conn_warn(tconn, "Not fencing peer, I'm not even Consistent myself.\n"); goto out; + case FP_DONT_CARE: + return true; + default: ; } - r = drbd_khelper(mdev, "fence-peer"); + r = conn_khelper(tconn, "fence-peer"); switch ((r>>8) & 0xff) { case 3: /* peer is inconsistent */ ex_to_string = "peer is inconsistent or worse"; - nps = D_INCONSISTENT; + mask.pdsk = D_MASK; + val.pdsk = D_INCONSISTENT; break; case 4: /* peer got outdated, or was already outdated */ ex_to_string = "peer was fenced"; - nps = D_OUTDATED; + mask.pdsk = D_MASK; + val.pdsk = D_OUTDATED; break; case 5: /* peer was down */ - if (mdev->state.disk == D_UP_TO_DATE) { + if (conn_highest_disk(tconn) == D_UP_TO_DATE) { /* we will(have) create(d) a new UUID anyways... */ ex_to_string = "peer is unreachable, assumed to be dead"; - nps = D_OUTDATED; + mask.pdsk = D_MASK; + val.pdsk = D_OUTDATED; } else { ex_to_string = "peer unreachable, doing nothing since disk != UpToDate"; - nps = mdev->state.pdsk; } break; case 6: /* Peer is primary, voluntarily outdate myself. * This is useful when an unconnected R_SECONDARY is asked to * become R_PRIMARY, but finds the other peer being active. */ ex_to_string = "peer is active"; - dev_warn(DEV, "Peer is primary, outdating myself.\n"); - nps = D_UNKNOWN; - _drbd_request_state(mdev, NS(disk, D_OUTDATED), CS_WAIT_COMPLETE); + conn_warn(tconn, "Peer is primary, outdating myself.\n"); + mask.disk = D_MASK; + val.disk = D_OUTDATED; break; case 7: if (fp != FP_STONITH) - dev_err(DEV, "fence-peer() = 7 && fencing != Stonith !!!\n"); + conn_err(tconn, "fence-peer() = 7 && fencing != Stonith !!!\n"); ex_to_string = "peer was stonithed"; - nps = D_OUTDATED; + mask.pdsk = D_MASK; + val.pdsk = D_OUTDATED; break; default: /* The script is broken ... */ - nps = D_UNKNOWN; - dev_err(DEV, "fence-peer helper broken, returned %d\n", (r>>8)&0xff); - return nps; + conn_err(tconn, "fence-peer helper broken, returned %d\n", (r>>8)&0xff); + return false; /* Eventually leave IO frozen */ } - dev_info(DEV, "fence-peer helper returned %d (%s)\n", - (r>>8) & 0xff, ex_to_string); + conn_info(tconn, "fence-peer helper returned %d (%s)\n", + (r>>8) & 0xff, ex_to_string); -out: - if (mdev->state.susp_fen && nps >= D_UNKNOWN) { - /* The handler was not successful... unfreeze here, the - state engine can not unfreeze... */ - _drbd_request_state(mdev, NS(susp_fen, 0), CS_VERBOSE); - } + out: - return nps; + /* Not using + conn_request_state(tconn, mask, val, CS_VERBOSE); + here, because we might were able to re-establish the connection in the + meantime. */ + spin_lock_irq(&tconn->req_lock); + if (tconn->cstate < C_WF_REPORT_PARAMS && !test_bit(STATE_SENT, &tconn->flags)) + _conn_request_state(tconn, mask, val, CS_VERBOSE); + spin_unlock_irq(&tconn->req_lock); + + return conn_highest_pdsk(tconn) <= D_OUTDATED; } static int _try_outdate_peer_async(void *data) { - struct drbd_conf *mdev = (struct drbd_conf *)data; - enum drbd_disk_state nps; - union drbd_state ns; + struct drbd_tconn *tconn = (struct drbd_tconn *)data; - nps = drbd_try_outdate_peer(mdev); - - /* Not using - drbd_request_state(mdev, NS(pdsk, nps)); - here, because we might were able to re-establish the connection - in the meantime. This can only partially be solved in the state's - engine is_valid_state() and is_valid_state_transition() - functions. - - nps can be D_INCONSISTENT, D_OUTDATED or D_UNKNOWN. - pdsk == D_INCONSISTENT while conn >= C_CONNECTED is valid, - therefore we have to have the pre state change check here. - */ - spin_lock_irq(&mdev->req_lock); - ns = mdev->state; - if (ns.conn < C_WF_REPORT_PARAMS && !test_bit(STATE_SENT, &mdev->flags)) { - ns.pdsk = nps; - _drbd_set_state(mdev, ns, CS_VERBOSE, NULL); - } - spin_unlock_irq(&mdev->req_lock); + conn_try_outdate_peer(tconn); + kref_put(&tconn->kref, &conn_destroy); return 0; } -void drbd_try_outdate_peer_async(struct drbd_conf *mdev) +void conn_try_outdate_peer_async(struct drbd_tconn *tconn) { struct task_struct *opa; - opa = kthread_run(_try_outdate_peer_async, mdev, "drbd%d_a_helper", mdev_to_minor(mdev)); - if (IS_ERR(opa)) - dev_err(DEV, "out of mem, failed to invoke fence-peer helper\n"); + kref_get(&tconn->kref); + opa = kthread_run(_try_outdate_peer_async, tconn, "drbd_async_h"); + if (IS_ERR(opa)) { + conn_err(tconn, "out of mem, failed to invoke fence-peer helper\n"); + kref_put(&tconn->kref, &conn_destroy); + } } enum drbd_state_rv @@ -318,15 +526,15 @@ drbd_set_role(struct drbd_conf *mdev, enum drbd_role new_role, int force) { const int max_tries = 4; enum drbd_state_rv rv = SS_UNKNOWN_ERROR; + struct net_conf *nc; int try = 0; int forced = 0; union drbd_state mask, val; - enum drbd_disk_state nps; if (new_role == R_PRIMARY) - request_ping(mdev); /* Detect a dead peer ASAP */ + request_ping(mdev->tconn); /* Detect a dead peer ASAP */ - mutex_lock(&mdev->state_mutex); + mutex_lock(mdev->state_mutex); mask.i = 0; mask.role = R_MASK; val.i = 0; val.role = new_role; @@ -354,38 +562,34 @@ drbd_set_role(struct drbd_conf *mdev, enum drbd_role new_role, int force) if (rv == SS_NO_UP_TO_DATE_DISK && mdev->state.disk == D_CONSISTENT && mask.pdsk == 0) { D_ASSERT(mdev->state.pdsk == D_UNKNOWN); - nps = drbd_try_outdate_peer(mdev); - if (nps == D_OUTDATED || nps == D_INCONSISTENT) { + if (conn_try_outdate_peer(mdev->tconn)) { val.disk = D_UP_TO_DATE; mask.disk = D_MASK; } - - val.pdsk = nps; - mask.pdsk = D_MASK; - continue; } if (rv == SS_NOTHING_TO_DO) - goto fail; + goto out; if (rv == SS_PRIMARY_NOP && mask.pdsk == 0) { - nps = drbd_try_outdate_peer(mdev); - - if (force && nps > D_OUTDATED) { + if (!conn_try_outdate_peer(mdev->tconn) && force) { dev_warn(DEV, "Forced into split brain situation!\n"); - nps = D_OUTDATED; - } - - mask.pdsk = D_MASK; - val.pdsk = nps; + mask.pdsk = D_MASK; + val.pdsk = D_OUTDATED; + } continue; } if (rv == SS_TWO_PRIMARIES) { /* Maybe the peer is detected as dead very soon... retry at most once more in this case. */ - schedule_timeout_interruptible((mdev->net_conf->ping_timeo+1)*HZ/10); + int timeo; + rcu_read_lock(); + nc = rcu_dereference(mdev->tconn->net_conf); + timeo = nc ? (nc->ping_timeo + 1) * HZ / 10 : 1; + rcu_read_unlock(); + schedule_timeout_interruptible(timeo); if (try < max_tries) try = max_tries - 1; continue; @@ -394,13 +598,13 @@ drbd_set_role(struct drbd_conf *mdev, enum drbd_role new_role, int force) rv = _drbd_request_state(mdev, mask, val, CS_VERBOSE + CS_WAIT_COMPLETE); if (rv < SS_SUCCESS) - goto fail; + goto out; } break; } if (rv < SS_SUCCESS) - goto fail; + goto out; if (forced) dev_warn(DEV, "Forced to consider local data as UpToDate!\n"); @@ -408,6 +612,8 @@ drbd_set_role(struct drbd_conf *mdev, enum drbd_role new_role, int force) /* Wait until nothing is on the fly :) */ wait_event(mdev->misc_wait, atomic_read(&mdev->ap_pending_cnt) == 0); + /* FIXME also wait for all pending P_BARRIER_ACK? */ + if (new_role == R_SECONDARY) { set_disk_ro(mdev->vdisk, true); if (get_ldev(mdev)) { @@ -415,10 +621,12 @@ drbd_set_role(struct drbd_conf *mdev, enum drbd_role new_role, int force) put_ldev(mdev); } } else { - if (get_net_conf(mdev)) { - mdev->net_conf->want_lose = 0; - put_net_conf(mdev); - } + mutex_lock(&mdev->tconn->conf_update); + nc = mdev->tconn->net_conf; + if (nc) + nc->discard_my_data = 0; /* without copy; single bit op is atomic */ + mutex_unlock(&mdev->tconn->conf_update); + set_disk_ro(mdev->vdisk, false); if (get_ldev(mdev)) { if (((mdev->state.conn < C_CONNECTED || @@ -444,67 +652,47 @@ drbd_set_role(struct drbd_conf *mdev, enum drbd_role new_role, int force) drbd_md_sync(mdev); kobject_uevent(&disk_to_dev(mdev->vdisk)->kobj, KOBJ_CHANGE); - fail: - mutex_unlock(&mdev->state_mutex); +out: + mutex_unlock(mdev->state_mutex); return rv; } -static struct drbd_conf *ensure_mdev(int minor, int create) +static const char *from_attrs_err_to_txt(int err) { - struct drbd_conf *mdev; - - if (minor >= minor_count) - return NULL; - - mdev = minor_to_mdev(minor); - - if (!mdev && create) { - struct gendisk *disk = NULL; - mdev = drbd_new_device(minor); - - spin_lock_irq(&drbd_pp_lock); - if (minor_table[minor] == NULL) { - minor_table[minor] = mdev; - disk = mdev->vdisk; - mdev = NULL; - } /* else: we lost the race */ - spin_unlock_irq(&drbd_pp_lock); - - if (disk) /* we won the race above */ - /* in case we ever add a drbd_delete_device(), - * don't forget the del_gendisk! */ - add_disk(disk); - else /* we lost the race above */ - drbd_free_mdev(mdev); - - mdev = minor_to_mdev(minor); - } - - return mdev; + return err == -ENOMSG ? "required attribute missing" : + err == -EOPNOTSUPP ? "unknown mandatory attribute" : + err == -EEXIST ? "can not change invariant setting" : + "invalid attribute value"; } -static int drbd_nl_primary(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) +int drbd_adm_set_role(struct sk_buff *skb, struct genl_info *info) { - struct primary primary_args; - - memset(&primary_args, 0, sizeof(struct primary)); - if (!primary_from_tags(mdev, nlp->tag_list, &primary_args)) { - reply->ret_code = ERR_MANDATORY_TAG; - return 0; - } - - reply->ret_code = - drbd_set_role(mdev, R_PRIMARY, primary_args.primary_force); + struct set_role_parms parms; + int err; + enum drbd_ret_code retcode; - return 0; -} + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto out; -static int drbd_nl_secondary(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) -{ - reply->ret_code = drbd_set_role(mdev, R_SECONDARY, 0); + memset(&parms, 0, sizeof(parms)); + if (info->attrs[DRBD_NLA_SET_ROLE_PARMS]) { + err = set_role_parms_from_attrs(&parms, info); + if (err) { + retcode = ERR_MANDATORY_TAG; + drbd_msg_put_info(from_attrs_err_to_txt(err)); + goto out; + } + } + if (info->genlhdr->cmd == DRBD_ADM_PRIMARY) + retcode = drbd_set_role(adm_ctx.mdev, R_PRIMARY, parms.assume_uptodate); + else + retcode = drbd_set_role(adm_ctx.mdev, R_SECONDARY, 0); +out: + drbd_adm_finish(info, retcode); return 0; } @@ -514,7 +702,12 @@ static void drbd_md_set_sector_offsets(struct drbd_conf *mdev, struct drbd_backing_dev *bdev) { sector_t md_size_sect = 0; - switch (bdev->dc.meta_dev_idx) { + int meta_dev_idx; + + rcu_read_lock(); + meta_dev_idx = rcu_dereference(bdev->disk_conf)->meta_dev_idx; + + switch (meta_dev_idx) { default: /* v07 style fixed size indexed meta data */ bdev->md.md_size_sect = MD_RESERVED_SECT; @@ -533,7 +726,7 @@ static void drbd_md_set_sector_offsets(struct drbd_conf *mdev, case DRBD_MD_INDEX_FLEX_INT: bdev->md.md_offset = drbd_md_ss__(mdev, bdev); /* al size is still fixed */ - bdev->md.al_offset = -MD_AL_MAX_SIZE; + bdev->md.al_offset = -MD_AL_SECTORS; /* we need (slightly less than) ~ this much bitmap sectors: */ md_size_sect = drbd_get_capacity(bdev->backing_bdev); md_size_sect = ALIGN(md_size_sect, BM_SECT_PER_EXT); @@ -549,6 +742,7 @@ static void drbd_md_set_sector_offsets(struct drbd_conf *mdev, bdev->md.bm_offset = -md_size_sect + MD_AL_OFFSET; break; } + rcu_read_unlock(); } /* input size is expected to be in KB */ @@ -581,10 +775,16 @@ char *ppsize(char *buf, unsigned long long size) * R_PRIMARY D_INCONSISTENT, and C_SYNC_TARGET: * peer may not initiate a resize. */ +/* Note these are not to be confused with + * drbd_adm_suspend_io/drbd_adm_resume_io, + * which are (sub) state changes triggered by admin (drbdsetup), + * and can be long lived. + * This changes an mdev->flag, is triggered by drbd internals, + * and should be short-lived. */ void drbd_suspend_io(struct drbd_conf *mdev) { set_bit(SUSPEND_IO, &mdev->flags); - if (is_susp(mdev->state)) + if (drbd_suspended(mdev)) return; wait_event(mdev->misc_wait, !atomic_read(&mdev->ap_bio_cnt)); } @@ -605,7 +805,7 @@ void drbd_resume_io(struct drbd_conf *mdev) enum determine_dev_size drbd_determine_dev_size(struct drbd_conf *mdev, enum dds_flags flags) __must_hold(local) { sector_t prev_first_sect, prev_size; /* previous meta location */ - sector_t la_size; + sector_t la_size, u_size; sector_t size; char ppb[10]; @@ -633,7 +833,10 @@ enum determine_dev_size drbd_determine_dev_size(struct drbd_conf *mdev, enum dds /* TODO: should only be some assert here, not (re)init... */ drbd_md_set_sector_offsets(mdev, mdev->ldev); - size = drbd_new_dev_size(mdev, mdev->ldev, flags & DDSF_FORCED); + rcu_read_lock(); + u_size = rcu_dereference(mdev->ldev->disk_conf)->disk_size; + rcu_read_unlock(); + size = drbd_new_dev_size(mdev, mdev->ldev, u_size, flags & DDSF_FORCED); if (drbd_get_capacity(mdev->this_bdev) != size || drbd_bm_capacity(mdev) != size) { @@ -696,12 +899,12 @@ out: } sector_t -drbd_new_dev_size(struct drbd_conf *mdev, struct drbd_backing_dev *bdev, int assume_peer_has_space) +drbd_new_dev_size(struct drbd_conf *mdev, struct drbd_backing_dev *bdev, + sector_t u_size, int assume_peer_has_space) { sector_t p_size = mdev->p_size; /* partner's disk size. */ sector_t la_size = bdev->md.la_size_sect; /* last agreed size. */ sector_t m_size; /* my size */ - sector_t u_size = bdev->dc.disk_size; /* size requested by user. */ sector_t size = 0; m_size = drbd_get_max_capacity(bdev); @@ -750,24 +953,21 @@ drbd_new_dev_size(struct drbd_conf *mdev, struct drbd_backing_dev *bdev, int ass * failed, and 0 on success. You should call drbd_md_sync() after you called * this function. */ -static int drbd_check_al_size(struct drbd_conf *mdev) +static int drbd_check_al_size(struct drbd_conf *mdev, struct disk_conf *dc) { struct lru_cache *n, *t; struct lc_element *e; unsigned int in_use; int i; - ERR_IF(mdev->sync_conf.al_extents < 7) - mdev->sync_conf.al_extents = 127; - if (mdev->act_log && - mdev->act_log->nr_elements == mdev->sync_conf.al_extents) + mdev->act_log->nr_elements == dc->al_extents) return 0; in_use = 0; t = mdev->act_log; - n = lc_create("act_log", drbd_al_ext_cache, - mdev->sync_conf.al_extents, sizeof(struct lc_element), 0); + n = lc_create("act_log", drbd_al_ext_cache, AL_UPDATES_PER_TRANSACTION, + dc->al_extents, sizeof(struct lc_element), 0); if (n == NULL) { dev_err(DEV, "Cannot allocate act_log lru!\n"); @@ -808,7 +1008,9 @@ static void drbd_setup_queue_param(struct drbd_conf *mdev, unsigned int max_bio_ struct request_queue * const b = mdev->ldev->backing_bdev->bd_disk->queue; max_hw_sectors = min(queue_max_hw_sectors(b), max_bio_size >> 9); - max_segments = mdev->ldev->dc.max_bio_bvecs; + rcu_read_lock(); + max_segments = rcu_dereference(mdev->ldev->disk_conf)->max_bio_bvecs; + rcu_read_unlock(); put_ldev(mdev); } @@ -852,12 +1054,14 @@ void drbd_reconsider_max_bio_size(struct drbd_conf *mdev) Because new from 8.3.8 onwards the peer can use multiple BIOs for a single peer_request */ if (mdev->state.conn >= C_CONNECTED) { - if (mdev->agreed_pro_version < 94) { - peer = min(mdev->peer_max_bio_size, DRBD_MAX_SIZE_H80_PACKET); + if (mdev->tconn->agreed_pro_version < 94) + peer = min( mdev->peer_max_bio_size, DRBD_MAX_SIZE_H80_PACKET); /* Correct old drbd (up to 8.3.7) if it believes it can do more than 32KiB */ - } else if (mdev->agreed_pro_version == 94) + else if (mdev->tconn->agreed_pro_version == 94) peer = DRBD_MAX_SIZE_H80_PACKET; - else /* drbd 8.3.8 onwards */ + else if (mdev->tconn->agreed_pro_version < 100) + peer = DRBD_MAX_BIO_SIZE_P95; /* drbd 8.3.8 onwards, before 8.4.0 */ + else peer = DRBD_MAX_BIO_SIZE; } @@ -872,36 +1076,27 @@ void drbd_reconsider_max_bio_size(struct drbd_conf *mdev) drbd_setup_queue_param(mdev, new); } -/* serialize deconfig (worker exiting, doing cleanup) - * and reconfig (drbdsetup disk, drbdsetup net) - * - * Wait for a potentially exiting worker, then restart it, - * or start a new one. Flush any pending work, there may still be an - * after_state_change queued. - */ -static void drbd_reconfig_start(struct drbd_conf *mdev) +/* Starts the worker thread */ +static void conn_reconfig_start(struct drbd_tconn *tconn) { - wait_event(mdev->state_wait, !test_and_set_bit(CONFIG_PENDING, &mdev->flags)); - wait_event(mdev->state_wait, !test_bit(DEVICE_DYING, &mdev->flags)); - drbd_thread_start(&mdev->worker); - drbd_flush_workqueue(mdev); + drbd_thread_start(&tconn->worker); + conn_flush_workqueue(tconn); } -/* if still unconfigured, stops worker again. - * if configured now, clears CONFIG_PENDING. - * wakes potential waiters */ -static void drbd_reconfig_done(struct drbd_conf *mdev) +/* if still unconfigured, stops worker again. */ +static void conn_reconfig_done(struct drbd_tconn *tconn) { - spin_lock_irq(&mdev->req_lock); - if (mdev->state.disk == D_DISKLESS && - mdev->state.conn == C_STANDALONE && - mdev->state.role == R_SECONDARY) { - set_bit(DEVICE_DYING, &mdev->flags); - drbd_thread_stop_nowait(&mdev->worker); - } else - clear_bit(CONFIG_PENDING, &mdev->flags); - spin_unlock_irq(&mdev->req_lock); - wake_up(&mdev->state_wait); + bool stop_threads; + spin_lock_irq(&tconn->req_lock); + stop_threads = conn_all_vols_unconf(tconn) && + tconn->cstate == C_STANDALONE; + spin_unlock_irq(&tconn->req_lock); + if (stop_threads) { + /* asender is implicitly stopped by receiver + * in conn_disconnect() */ + drbd_thread_stop(&tconn->receiver); + drbd_thread_stop(&tconn->worker); + } } /* Make sure IO is suspended before calling this function(). */ @@ -909,42 +1104,187 @@ static void drbd_suspend_al(struct drbd_conf *mdev) { int s = 0; - if (lc_try_lock(mdev->act_log)) { - drbd_al_shrink(mdev); - lc_unlock(mdev->act_log); - } else { + if (!lc_try_lock(mdev->act_log)) { dev_warn(DEV, "Failed to lock al in drbd_suspend_al()\n"); return; } - spin_lock_irq(&mdev->req_lock); + drbd_al_shrink(mdev); + spin_lock_irq(&mdev->tconn->req_lock); if (mdev->state.conn < C_CONNECTED) s = !test_and_set_bit(AL_SUSPENDED, &mdev->flags); - - spin_unlock_irq(&mdev->req_lock); + spin_unlock_irq(&mdev->tconn->req_lock); + lc_unlock(mdev->act_log); if (s) dev_info(DEV, "Suspended AL updates\n"); } -/* does always return 0; - * interesting return code is in reply->ret_code */ -static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) + +static bool should_set_defaults(struct genl_info *info) +{ + unsigned flags = ((struct drbd_genlmsghdr*)info->userhdr)->flags; + return 0 != (flags & DRBD_GENL_F_SET_DEFAULTS); +} + +static void enforce_disk_conf_limits(struct disk_conf *dc) +{ + if (dc->al_extents < DRBD_AL_EXTENTS_MIN) + dc->al_extents = DRBD_AL_EXTENTS_MIN; + if (dc->al_extents > DRBD_AL_EXTENTS_MAX) + dc->al_extents = DRBD_AL_EXTENTS_MAX; + + if (dc->c_plan_ahead > DRBD_C_PLAN_AHEAD_MAX) + dc->c_plan_ahead = DRBD_C_PLAN_AHEAD_MAX; +} + +int drbd_adm_disk_opts(struct sk_buff *skb, struct genl_info *info) { enum drbd_ret_code retcode; + struct drbd_conf *mdev; + struct disk_conf *new_disk_conf, *old_disk_conf; + struct fifo_buffer *old_plan = NULL, *new_plan = NULL; + int err, fifo_size; + + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto out; + + mdev = adm_ctx.mdev; + + /* we also need a disk + * to change the options on */ + if (!get_ldev(mdev)) { + retcode = ERR_NO_DISK; + goto out; + } + + new_disk_conf = kmalloc(sizeof(struct disk_conf), GFP_KERNEL); + if (!new_disk_conf) { + retcode = ERR_NOMEM; + goto fail; + } + + mutex_lock(&mdev->tconn->conf_update); + old_disk_conf = mdev->ldev->disk_conf; + *new_disk_conf = *old_disk_conf; + if (should_set_defaults(info)) + set_disk_conf_defaults(new_disk_conf); + + err = disk_conf_from_attrs_for_change(new_disk_conf, info); + if (err && err != -ENOMSG) { + retcode = ERR_MANDATORY_TAG; + drbd_msg_put_info(from_attrs_err_to_txt(err)); + } + + if (!expect(new_disk_conf->resync_rate >= 1)) + new_disk_conf->resync_rate = 1; + + enforce_disk_conf_limits(new_disk_conf); + + fifo_size = (new_disk_conf->c_plan_ahead * 10 * SLEEP_TIME) / HZ; + if (fifo_size != mdev->rs_plan_s->size) { + new_plan = fifo_alloc(fifo_size); + if (!new_plan) { + dev_err(DEV, "kmalloc of fifo_buffer failed"); + retcode = ERR_NOMEM; + goto fail_unlock; + } + } + + drbd_suspend_io(mdev); + wait_event(mdev->al_wait, lc_try_lock(mdev->act_log)); + drbd_al_shrink(mdev); + err = drbd_check_al_size(mdev, new_disk_conf); + lc_unlock(mdev->act_log); + wake_up(&mdev->al_wait); + drbd_resume_io(mdev); + + if (err) { + retcode = ERR_NOMEM; + goto fail_unlock; + } + + write_lock_irq(&global_state_lock); + retcode = drbd_resync_after_valid(mdev, new_disk_conf->resync_after); + if (retcode == NO_ERROR) { + rcu_assign_pointer(mdev->ldev->disk_conf, new_disk_conf); + drbd_resync_after_changed(mdev); + } + write_unlock_irq(&global_state_lock); + + if (retcode != NO_ERROR) + goto fail_unlock; + + if (new_plan) { + old_plan = mdev->rs_plan_s; + rcu_assign_pointer(mdev->rs_plan_s, new_plan); + } + + mutex_unlock(&mdev->tconn->conf_update); + + if (new_disk_conf->al_updates) + mdev->ldev->md.flags &= ~MDF_AL_DISABLED; + else + mdev->ldev->md.flags |= MDF_AL_DISABLED; + + if (new_disk_conf->md_flushes) + clear_bit(MD_NO_FUA, &mdev->flags); + else + set_bit(MD_NO_FUA, &mdev->flags); + + drbd_bump_write_ordering(mdev->tconn, WO_bdev_flush); + + drbd_md_sync(mdev); + + if (mdev->state.conn >= C_CONNECTED) + drbd_send_sync_param(mdev); + + synchronize_rcu(); + kfree(old_disk_conf); + kfree(old_plan); + mod_timer(&mdev->request_timer, jiffies + HZ); + goto success; + +fail_unlock: + mutex_unlock(&mdev->tconn->conf_update); + fail: + kfree(new_disk_conf); + kfree(new_plan); +success: + put_ldev(mdev); + out: + drbd_adm_finish(info, retcode); + return 0; +} + +int drbd_adm_attach(struct sk_buff *skb, struct genl_info *info) +{ + struct drbd_conf *mdev; + int err; + enum drbd_ret_code retcode; enum determine_dev_size dd; sector_t max_possible_sectors; sector_t min_md_device_sectors; struct drbd_backing_dev *nbc = NULL; /* new_backing_conf */ + struct disk_conf *new_disk_conf = NULL; struct block_device *bdev; struct lru_cache *resync_lru = NULL; + struct fifo_buffer *new_plan = NULL; union drbd_state ns, os; enum drbd_state_rv rv; - int cp_discovered = 0; - int logical_block_size; + struct net_conf *nc; - drbd_reconfig_start(mdev); + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto finish; + + mdev = adm_ctx.mdev; + conn_reconfig_start(mdev->tconn); /* if you want to reconfigure, please tear down first */ if (mdev->state.disk > D_DISKLESS) { @@ -959,47 +1299,65 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp /* make sure there is no leftover from previous force-detach attempts */ clear_bit(FORCE_DETACH, &mdev->flags); + clear_bit(WAS_IO_ERROR, &mdev->flags); + clear_bit(WAS_READ_ERROR, &mdev->flags); /* and no leftover from previously aborted resync or verify, either */ mdev->rs_total = 0; mdev->rs_failed = 0; atomic_set(&mdev->rs_pending_cnt, 0); - /* allocation not in the IO path, cqueue thread context */ + /* allocation not in the IO path, drbdsetup context */ nbc = kzalloc(sizeof(struct drbd_backing_dev), GFP_KERNEL); if (!nbc) { retcode = ERR_NOMEM; goto fail; } + spin_lock_init(&nbc->md.uuid_lock); - nbc->dc.disk_size = DRBD_DISK_SIZE_SECT_DEF; - nbc->dc.on_io_error = DRBD_ON_IO_ERROR_DEF; - nbc->dc.fencing = DRBD_FENCING_DEF; - nbc->dc.max_bio_bvecs = DRBD_MAX_BIO_BVECS_DEF; + new_disk_conf = kzalloc(sizeof(struct disk_conf), GFP_KERNEL); + if (!new_disk_conf) { + retcode = ERR_NOMEM; + goto fail; + } + nbc->disk_conf = new_disk_conf; - if (!disk_conf_from_tags(mdev, nlp->tag_list, &nbc->dc)) { + set_disk_conf_defaults(new_disk_conf); + err = disk_conf_from_attrs(new_disk_conf, info); + if (err) { retcode = ERR_MANDATORY_TAG; + drbd_msg_put_info(from_attrs_err_to_txt(err)); goto fail; } - if (nbc->dc.meta_dev_idx < DRBD_MD_INDEX_FLEX_INT) { + enforce_disk_conf_limits(new_disk_conf); + + new_plan = fifo_alloc((new_disk_conf->c_plan_ahead * 10 * SLEEP_TIME) / HZ); + if (!new_plan) { + retcode = ERR_NOMEM; + goto fail; + } + + if (new_disk_conf->meta_dev_idx < DRBD_MD_INDEX_FLEX_INT) { retcode = ERR_MD_IDX_INVALID; goto fail; } - if (get_net_conf(mdev)) { - int prot = mdev->net_conf->wire_protocol; - put_net_conf(mdev); - if (nbc->dc.fencing == FP_STONITH && prot == DRBD_PROT_A) { + rcu_read_lock(); + nc = rcu_dereference(mdev->tconn->net_conf); + if (nc) { + if (new_disk_conf->fencing == FP_STONITH && nc->wire_protocol == DRBD_PROT_A) { + rcu_read_unlock(); retcode = ERR_STONITH_AND_PROT_A; goto fail; } } + rcu_read_unlock(); - bdev = blkdev_get_by_path(nbc->dc.backing_dev, + bdev = blkdev_get_by_path(new_disk_conf->backing_dev, FMODE_READ | FMODE_WRITE | FMODE_EXCL, mdev); if (IS_ERR(bdev)) { - dev_err(DEV, "open(\"%s\") failed with %ld\n", nbc->dc.backing_dev, + dev_err(DEV, "open(\"%s\") failed with %ld\n", new_disk_conf->backing_dev, PTR_ERR(bdev)); retcode = ERR_OPEN_DISK; goto fail; @@ -1014,12 +1372,12 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp * should check it for you already; but if you don't, or * someone fooled it, we need to double check here) */ - bdev = blkdev_get_by_path(nbc->dc.meta_dev, + bdev = blkdev_get_by_path(new_disk_conf->meta_dev, FMODE_READ | FMODE_WRITE | FMODE_EXCL, - (nbc->dc.meta_dev_idx < 0) ? + (new_disk_conf->meta_dev_idx < 0) ? (void *)mdev : (void *)drbd_m_holder); if (IS_ERR(bdev)) { - dev_err(DEV, "open(\"%s\") failed with %ld\n", nbc->dc.meta_dev, + dev_err(DEV, "open(\"%s\") failed with %ld\n", new_disk_conf->meta_dev, PTR_ERR(bdev)); retcode = ERR_OPEN_MD_DISK; goto fail; @@ -1027,14 +1385,14 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp nbc->md_bdev = bdev; if ((nbc->backing_bdev == nbc->md_bdev) != - (nbc->dc.meta_dev_idx == DRBD_MD_INDEX_INTERNAL || - nbc->dc.meta_dev_idx == DRBD_MD_INDEX_FLEX_INT)) { + (new_disk_conf->meta_dev_idx == DRBD_MD_INDEX_INTERNAL || + new_disk_conf->meta_dev_idx == DRBD_MD_INDEX_FLEX_INT)) { retcode = ERR_MD_IDX_INVALID; goto fail; } resync_lru = lc_create("resync", drbd_bm_ext_cache, - 61, sizeof(struct bm_extent), + 1, 61, sizeof(struct bm_extent), offsetof(struct bm_extent, lce)); if (!resync_lru) { retcode = ERR_NOMEM; @@ -1044,21 +1402,21 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp /* RT - for drbd_get_max_capacity() DRBD_MD_INDEX_FLEX_INT */ drbd_md_set_sector_offsets(mdev, nbc); - if (drbd_get_max_capacity(nbc) < nbc->dc.disk_size) { + if (drbd_get_max_capacity(nbc) < new_disk_conf->disk_size) { dev_err(DEV, "max capacity %llu smaller than disk size %llu\n", (unsigned long long) drbd_get_max_capacity(nbc), - (unsigned long long) nbc->dc.disk_size); + (unsigned long long) new_disk_conf->disk_size); retcode = ERR_DISK_TOO_SMALL; goto fail; } - if (nbc->dc.meta_dev_idx < 0) { + if (new_disk_conf->meta_dev_idx < 0) { max_possible_sectors = DRBD_MAX_SECTORS_FLEX; /* at least one MB, otherwise it does not make sense */ min_md_device_sectors = (2<<10); } else { max_possible_sectors = DRBD_MAX_SECTORS; - min_md_device_sectors = MD_RESERVED_SECT * (nbc->dc.meta_dev_idx + 1); + min_md_device_sectors = MD_RESERVED_SECT * (new_disk_conf->meta_dev_idx + 1); } if (drbd_get_capacity(nbc->md_bdev) < min_md_device_sectors) { @@ -1083,14 +1441,20 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp dev_warn(DEV, "==> truncating very big lower level device " "to currently maximum possible %llu sectors <==\n", (unsigned long long) max_possible_sectors); - if (nbc->dc.meta_dev_idx >= 0) + if (new_disk_conf->meta_dev_idx >= 0) dev_warn(DEV, "==>> using internal or flexible " "meta data may help <<==\n"); } drbd_suspend_io(mdev); /* also wait for the last barrier ack. */ - wait_event(mdev->misc_wait, !atomic_read(&mdev->ap_pending_cnt) || is_susp(mdev->state)); + /* FIXME see also https://daiquiri.linbit/cgi-bin/bugzilla/show_bug.cgi?id=171 + * We need a way to either ignore barrier acks for barriers sent before a device + * was attached, or a way to wait for all pending barrier acks to come in. + * As barriers are counted per resource, + * we'd need to suspend io on all devices of a resource. + */ + wait_event(mdev->misc_wait, !atomic_read(&mdev->ap_pending_cnt) || drbd_suspended(mdev)); /* and for any other previously queued work */ drbd_flush_workqueue(mdev); @@ -1105,25 +1469,6 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp drbd_md_set_sector_offsets(mdev, nbc); - /* allocate a second IO page if logical_block_size != 512 */ - logical_block_size = bdev_logical_block_size(nbc->md_bdev); - if (logical_block_size == 0) - logical_block_size = MD_SECTOR_SIZE; - - if (logical_block_size != MD_SECTOR_SIZE) { - if (!mdev->md_io_tmpp) { - struct page *page = alloc_page(GFP_NOIO); - if (!page) - goto force_diskless_dec; - - dev_warn(DEV, "Meta data's bdev logical_block_size = %d != %d\n", - logical_block_size, MD_SECTOR_SIZE); - dev_warn(DEV, "Workaround engaged (has performance impact).\n"); - - mdev->md_io_tmpp = page; - } - } - if (!mdev->bitmap) { if (drbd_bm_init(mdev)) { retcode = ERR_NOMEM; @@ -1145,30 +1490,25 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp } /* Since we are diskless, fix the activity log first... */ - if (drbd_check_al_size(mdev)) { + if (drbd_check_al_size(mdev, new_disk_conf)) { retcode = ERR_NOMEM; goto force_diskless_dec; } /* Prevent shrinking of consistent devices ! */ if (drbd_md_test_flag(nbc, MDF_CONSISTENT) && - drbd_new_dev_size(mdev, nbc, 0) < nbc->md.la_size_sect) { + drbd_new_dev_size(mdev, nbc, nbc->disk_conf->disk_size, 0) < nbc->md.la_size_sect) { dev_warn(DEV, "refusing to truncate a consistent device\n"); retcode = ERR_DISK_TOO_SMALL; goto force_diskless_dec; } - if (!drbd_al_read_log(mdev, nbc)) { - retcode = ERR_IO_MD_DISK; - goto force_diskless_dec; - } - /* Reset the "barriers don't work" bits here, then force meta data to * be written, to ensure we determine if barriers are supported. */ - if (nbc->dc.no_md_flush) - set_bit(MD_NO_FUA, &mdev->flags); - else + if (new_disk_conf->md_flushes) clear_bit(MD_NO_FUA, &mdev->flags); + else + set_bit(MD_NO_FUA, &mdev->flags); /* Point of no return reached. * Devices and memory are no longer released by error cleanup below. @@ -1177,11 +1517,13 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp D_ASSERT(mdev->ldev == NULL); mdev->ldev = nbc; mdev->resync = resync_lru; + mdev->rs_plan_s = new_plan; nbc = NULL; resync_lru = NULL; + new_disk_conf = NULL; + new_plan = NULL; - mdev->write_ordering = WO_bdev_flush; - drbd_bump_write_ordering(mdev, WO_bdev_flush); + drbd_bump_write_ordering(mdev->tconn, WO_bdev_flush); if (drbd_md_test_flag(mdev->ldev, MDF_CRASHED_PRIMARY)) set_bit(CRASHED_PRIMARY, &mdev->flags); @@ -1189,10 +1531,8 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp clear_bit(CRASHED_PRIMARY, &mdev->flags); if (drbd_md_test_flag(mdev->ldev, MDF_PRIMARY_IND) && - !(mdev->state.role == R_PRIMARY && mdev->state.susp_nod)) { + !(mdev->state.role == R_PRIMARY && mdev->tconn->susp_nod)) set_bit(CRASHED_PRIMARY, &mdev->flags); - cp_discovered = 1; - } mdev->send_cnt = 0; mdev->recv_cnt = 0; @@ -1228,7 +1568,9 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp } else if (dd == grew) set_bit(RESYNC_AFTER_NEG, &mdev->flags); - if (drbd_md_test_flag(mdev->ldev, MDF_FULL_SYNC)) { + if (drbd_md_test_flag(mdev->ldev, MDF_FULL_SYNC) || + (test_bit(CRASHED_PRIMARY, &mdev->flags) && + drbd_md_test_flag(mdev->ldev, MDF_AL_DISABLED))) { dev_info(DEV, "Assuming that all blocks are out of sync " "(aka FullSync)\n"); if (drbd_bitmap_io(mdev, &drbd_bmio_set_n_write, @@ -1238,16 +1580,7 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp } } else { if (drbd_bitmap_io(mdev, &drbd_bm_read, - "read from attaching", BM_LOCKED_MASK) < 0) { - retcode = ERR_IO_MD_DISK; - goto force_diskless_dec; - } - } - - if (cp_discovered) { - drbd_al_apply_to_bm(mdev); - if (drbd_bitmap_io(mdev, &drbd_bm_write, - "crashed primary apply AL", BM_LOCKED_MASK)) { + "read from attaching", BM_LOCKED_MASK)) { retcode = ERR_IO_MD_DISK; goto force_diskless_dec; } @@ -1256,9 +1589,9 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp if (_drbd_bm_total_weight(mdev) == drbd_bm_bits(mdev)) drbd_suspend_al(mdev); /* IO is still suspended here... */ - spin_lock_irq(&mdev->req_lock); - os = mdev->state; - ns.i = os.i; + spin_lock_irq(&mdev->tconn->req_lock); + os = drbd_read_state(mdev); + ns = os; /* If MDF_CONSISTENT is not set go into inconsistent state, otherwise investigate MDF_WasUpToDate... If MDF_WAS_UP_TO_DATE is not set go into D_OUTDATED disk state, @@ -1276,8 +1609,9 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp if (drbd_md_test_flag(mdev->ldev, MDF_PEER_OUT_DATED)) ns.pdsk = D_OUTDATED; - if ( ns.disk == D_CONSISTENT && - (ns.pdsk == D_OUTDATED || mdev->ldev->dc.fencing == FP_DONT_CARE)) + rcu_read_lock(); + if (ns.disk == D_CONSISTENT && + (ns.pdsk == D_OUTDATED || rcu_dereference(mdev->ldev->disk_conf)->fencing == FP_DONT_CARE)) ns.disk = D_UP_TO_DATE; /* All tests on MDF_PRIMARY_IND, MDF_CONNECTED_IND, @@ -1285,6 +1619,13 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp this point, because drbd_request_state() modifies these flags. */ + if (rcu_dereference(mdev->ldev->disk_conf)->al_updates) + mdev->ldev->md.flags &= ~MDF_AL_DISABLED; + else + mdev->ldev->md.flags |= MDF_AL_DISABLED; + + rcu_read_unlock(); + /* In case we are C_CONNECTED postpone any decision on the new disk state after the negotiation phase. */ if (mdev->state.conn == C_CONNECTED) { @@ -1300,12 +1641,13 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp } rv = _drbd_set_state(mdev, ns, CS_VERBOSE, NULL); - ns = mdev->state; - spin_unlock_irq(&mdev->req_lock); + spin_unlock_irq(&mdev->tconn->req_lock); if (rv < SS_SUCCESS) goto force_diskless_dec; + mod_timer(&mdev->request_timer, jiffies + HZ); + if (mdev->state.role == R_PRIMARY) mdev->ldev->md.uuid[UI_CURRENT] |= (u64)1; else @@ -1316,16 +1658,17 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp kobject_uevent(&disk_to_dev(mdev->vdisk)->kobj, KOBJ_CHANGE); put_ldev(mdev); - reply->ret_code = retcode; - drbd_reconfig_done(mdev); + conn_reconfig_done(mdev->tconn); + drbd_adm_finish(info, retcode); return 0; force_diskless_dec: put_ldev(mdev); force_diskless: - drbd_force_state(mdev, NS(disk, D_FAILED)); + drbd_force_state(mdev, NS(disk, D_DISKLESS)); drbd_md_sync(mdev); fail: + conn_reconfig_done(mdev->tconn); if (nbc) { if (nbc->backing_bdev) blkdev_put(nbc->backing_bdev, @@ -1335,34 +1678,24 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp FMODE_READ | FMODE_WRITE | FMODE_EXCL); kfree(nbc); } + kfree(new_disk_conf); lc_destroy(resync_lru); + kfree(new_plan); - reply->ret_code = retcode; - drbd_reconfig_done(mdev); + finish: + drbd_adm_finish(info, retcode); return 0; } -/* Detaching the disk is a process in multiple stages. First we need to lock - * out application IO, in-flight IO, IO stuck in drbd_al_begin_io. - * Then we transition to D_DISKLESS, and wait for put_ldev() to return all - * internal references as well. - * Only then we have finally detached. */ -static int drbd_nl_detach(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) +static int adm_detach(struct drbd_conf *mdev, int force) { - enum drbd_ret_code retcode; + enum drbd_state_rv retcode; int ret; - struct detach dt = {}; - if (!detach_from_tags(mdev, nlp->tag_list, &dt)) { - reply->ret_code = ERR_MANDATORY_TAG; - goto out; - } - - if (dt.detach_force) { + if (force) { set_bit(FORCE_DETACH, &mdev->flags); drbd_force_state(mdev, NS(disk, D_FAILED)); - reply->ret_code = SS_SUCCESS; + retcode = SS_SUCCESS; goto out; } @@ -1374,326 +1707,529 @@ static int drbd_nl_detach(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, ret = wait_event_interruptible(mdev->misc_wait, mdev->state.disk != D_FAILED); drbd_resume_io(mdev); - if ((int)retcode == (int)SS_IS_DISKLESS) retcode = SS_NOTHING_TO_DO; if (ret) retcode = ERR_INTR; - reply->ret_code = retcode; out: - return 0; + return retcode; } -static int drbd_nl_net_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) +/* Detaching the disk is a process in multiple stages. First we need to lock + * out application IO, in-flight IO, IO stuck in drbd_al_begin_io. + * Then we transition to D_DISKLESS, and wait for put_ldev() to return all + * internal references as well. + * Only then we have finally detached. */ +int drbd_adm_detach(struct sk_buff *skb, struct genl_info *info) { - int i, ns; enum drbd_ret_code retcode; - struct net_conf *new_conf = NULL; - struct crypto_hash *tfm = NULL; - struct crypto_hash *integrity_w_tfm = NULL; - struct crypto_hash *integrity_r_tfm = NULL; - struct hlist_head *new_tl_hash = NULL; - struct hlist_head *new_ee_hash = NULL; - struct drbd_conf *odev; - char hmac_name[CRYPTO_MAX_ALG_NAME]; - void *int_dig_out = NULL; - void *int_dig_in = NULL; - void *int_dig_vv = NULL; - struct sockaddr *new_my_addr, *new_peer_addr, *taken_addr; + struct detach_parms parms = { }; + int err; - drbd_reconfig_start(mdev); + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto out; - if (mdev->state.conn > C_STANDALONE) { - retcode = ERR_NET_CONFIGURED; - goto fail; + if (info->attrs[DRBD_NLA_DETACH_PARMS]) { + err = detach_parms_from_attrs(&parms, info); + if (err) { + retcode = ERR_MANDATORY_TAG; + drbd_msg_put_info(from_attrs_err_to_txt(err)); + goto out; + } } - /* allocation not in the IO path, cqueue thread context */ + retcode = adm_detach(adm_ctx.mdev, parms.force_detach); +out: + drbd_adm_finish(info, retcode); + return 0; +} + +static bool conn_resync_running(struct drbd_tconn *tconn) +{ + struct drbd_conf *mdev; + bool rv = false; + int vnr; + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) { + if (mdev->state.conn == C_SYNC_SOURCE || + mdev->state.conn == C_SYNC_TARGET || + mdev->state.conn == C_PAUSED_SYNC_S || + mdev->state.conn == C_PAUSED_SYNC_T) { + rv = true; + break; + } + } + rcu_read_unlock(); + + return rv; +} + +static bool conn_ov_running(struct drbd_tconn *tconn) +{ + struct drbd_conf *mdev; + bool rv = false; + int vnr; + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) { + if (mdev->state.conn == C_VERIFY_S || + mdev->state.conn == C_VERIFY_T) { + rv = true; + break; + } + } + rcu_read_unlock(); + + return rv; +} + +static enum drbd_ret_code +_check_net_options(struct drbd_tconn *tconn, struct net_conf *old_conf, struct net_conf *new_conf) +{ + struct drbd_conf *mdev; + int i; + + if (old_conf && tconn->cstate == C_WF_REPORT_PARAMS && tconn->agreed_pro_version < 100) { + if (new_conf->wire_protocol != old_conf->wire_protocol) + return ERR_NEED_APV_100; + + if (new_conf->two_primaries != old_conf->two_primaries) + return ERR_NEED_APV_100; + + if (strcmp(new_conf->integrity_alg, old_conf->integrity_alg)) + return ERR_NEED_APV_100; + } + + if (!new_conf->two_primaries && + conn_highest_role(tconn) == R_PRIMARY && + conn_highest_peer(tconn) == R_PRIMARY) + return ERR_NEED_ALLOW_TWO_PRI; + + if (new_conf->two_primaries && + (new_conf->wire_protocol != DRBD_PROT_C)) + return ERR_NOT_PROTO_C; + + idr_for_each_entry(&tconn->volumes, mdev, i) { + if (get_ldev(mdev)) { + enum drbd_fencing_p fp = rcu_dereference(mdev->ldev->disk_conf)->fencing; + put_ldev(mdev); + if (new_conf->wire_protocol == DRBD_PROT_A && fp == FP_STONITH) + return ERR_STONITH_AND_PROT_A; + } + if (mdev->state.role == R_PRIMARY && new_conf->discard_my_data) + return ERR_DISCARD_IMPOSSIBLE; + } + + if (new_conf->on_congestion != OC_BLOCK && new_conf->wire_protocol != DRBD_PROT_A) + return ERR_CONG_NOT_PROTO_A; + + return NO_ERROR; +} + +static enum drbd_ret_code +check_net_options(struct drbd_tconn *tconn, struct net_conf *new_conf) +{ + static enum drbd_ret_code rv; + struct drbd_conf *mdev; + int i; + + rcu_read_lock(); + rv = _check_net_options(tconn, rcu_dereference(tconn->net_conf), new_conf); + rcu_read_unlock(); + + /* tconn->volumes protected by genl_lock() here */ + idr_for_each_entry(&tconn->volumes, mdev, i) { + if (!mdev->bitmap) { + if(drbd_bm_init(mdev)) + return ERR_NOMEM; + } + } + + return rv; +} + +struct crypto { + struct crypto_hash *verify_tfm; + struct crypto_hash *csums_tfm; + struct crypto_hash *cram_hmac_tfm; + struct crypto_hash *integrity_tfm; +}; + +static int +alloc_hash(struct crypto_hash **tfm, char *tfm_name, int err_alg) +{ + if (!tfm_name[0]) + return NO_ERROR; + + *tfm = crypto_alloc_hash(tfm_name, 0, CRYPTO_ALG_ASYNC); + if (IS_ERR(*tfm)) { + *tfm = NULL; + return err_alg; + } + + return NO_ERROR; +} + +static enum drbd_ret_code +alloc_crypto(struct crypto *crypto, struct net_conf *new_conf) +{ + char hmac_name[CRYPTO_MAX_ALG_NAME]; + enum drbd_ret_code rv; + + rv = alloc_hash(&crypto->csums_tfm, new_conf->csums_alg, + ERR_CSUMS_ALG); + if (rv != NO_ERROR) + return rv; + rv = alloc_hash(&crypto->verify_tfm, new_conf->verify_alg, + ERR_VERIFY_ALG); + if (rv != NO_ERROR) + return rv; + rv = alloc_hash(&crypto->integrity_tfm, new_conf->integrity_alg, + ERR_INTEGRITY_ALG); + if (rv != NO_ERROR) + return rv; + if (new_conf->cram_hmac_alg[0] != 0) { + snprintf(hmac_name, CRYPTO_MAX_ALG_NAME, "hmac(%s)", + new_conf->cram_hmac_alg); + + rv = alloc_hash(&crypto->cram_hmac_tfm, hmac_name, + ERR_AUTH_ALG); + } + + return rv; +} + +static void free_crypto(struct crypto *crypto) +{ + crypto_free_hash(crypto->cram_hmac_tfm); + crypto_free_hash(crypto->integrity_tfm); + crypto_free_hash(crypto->csums_tfm); + crypto_free_hash(crypto->verify_tfm); +} + +int drbd_adm_net_opts(struct sk_buff *skb, struct genl_info *info) +{ + enum drbd_ret_code retcode; + struct drbd_tconn *tconn; + struct net_conf *old_conf, *new_conf = NULL; + int err; + int ovr; /* online verify running */ + int rsr; /* re-sync running */ + struct crypto crypto = { }; + + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_CONNECTION); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto out; + + tconn = adm_ctx.tconn; + new_conf = kzalloc(sizeof(struct net_conf), GFP_KERNEL); if (!new_conf) { retcode = ERR_NOMEM; + goto out; + } + + conn_reconfig_start(tconn); + + mutex_lock(&tconn->data.mutex); + mutex_lock(&tconn->conf_update); + old_conf = tconn->net_conf; + + if (!old_conf) { + drbd_msg_put_info("net conf missing, try connect"); + retcode = ERR_INVALID_REQUEST; goto fail; } - new_conf->timeout = DRBD_TIMEOUT_DEF; - new_conf->try_connect_int = DRBD_CONNECT_INT_DEF; - new_conf->ping_int = DRBD_PING_INT_DEF; - new_conf->max_epoch_size = DRBD_MAX_EPOCH_SIZE_DEF; - new_conf->max_buffers = DRBD_MAX_BUFFERS_DEF; - new_conf->unplug_watermark = DRBD_UNPLUG_WATERMARK_DEF; - new_conf->sndbuf_size = DRBD_SNDBUF_SIZE_DEF; - new_conf->rcvbuf_size = DRBD_RCVBUF_SIZE_DEF; - new_conf->ko_count = DRBD_KO_COUNT_DEF; - new_conf->after_sb_0p = DRBD_AFTER_SB_0P_DEF; - new_conf->after_sb_1p = DRBD_AFTER_SB_1P_DEF; - new_conf->after_sb_2p = DRBD_AFTER_SB_2P_DEF; - new_conf->want_lose = 0; - new_conf->two_primaries = 0; - new_conf->wire_protocol = DRBD_PROT_C; - new_conf->ping_timeo = DRBD_PING_TIMEO_DEF; - new_conf->rr_conflict = DRBD_RR_CONFLICT_DEF; - new_conf->on_congestion = DRBD_ON_CONGESTION_DEF; - new_conf->cong_extents = DRBD_CONG_EXTENTS_DEF; - - if (!net_conf_from_tags(mdev, nlp->tag_list, new_conf)) { + *new_conf = *old_conf; + if (should_set_defaults(info)) + set_net_conf_defaults(new_conf); + + err = net_conf_from_attrs_for_change(new_conf, info); + if (err && err != -ENOMSG) { retcode = ERR_MANDATORY_TAG; + drbd_msg_put_info(from_attrs_err_to_txt(err)); goto fail; } - if (new_conf->two_primaries - && (new_conf->wire_protocol != DRBD_PROT_C)) { - retcode = ERR_NOT_PROTO_C; + retcode = check_net_options(tconn, new_conf); + if (retcode != NO_ERROR) goto fail; - } - if (get_ldev(mdev)) { - enum drbd_fencing_p fp = mdev->ldev->dc.fencing; - put_ldev(mdev); - if (new_conf->wire_protocol == DRBD_PROT_A && fp == FP_STONITH) { - retcode = ERR_STONITH_AND_PROT_A; - goto fail; - } + /* re-sync running */ + rsr = conn_resync_running(tconn); + if (rsr && strcmp(new_conf->csums_alg, old_conf->csums_alg)) { + retcode = ERR_CSUMS_RESYNC_RUNNING; + goto fail; } - if (new_conf->on_congestion != OC_BLOCK && new_conf->wire_protocol != DRBD_PROT_A) { - retcode = ERR_CONG_NOT_PROTO_A; + /* online verify running */ + ovr = conn_ov_running(tconn); + if (ovr && strcmp(new_conf->verify_alg, old_conf->verify_alg)) { + retcode = ERR_VERIFY_RUNNING; goto fail; } - if (mdev->state.role == R_PRIMARY && new_conf->want_lose) { - retcode = ERR_DISCARD; + retcode = alloc_crypto(&crypto, new_conf); + if (retcode != NO_ERROR) goto fail; - } - retcode = NO_ERROR; + rcu_assign_pointer(tconn->net_conf, new_conf); - new_my_addr = (struct sockaddr *)&new_conf->my_addr; - new_peer_addr = (struct sockaddr *)&new_conf->peer_addr; - for (i = 0; i < minor_count; i++) { - odev = minor_to_mdev(i); - if (!odev || odev == mdev) - continue; - if (get_net_conf(odev)) { - taken_addr = (struct sockaddr *)&odev->net_conf->my_addr; - if (new_conf->my_addr_len == odev->net_conf->my_addr_len && - !memcmp(new_my_addr, taken_addr, new_conf->my_addr_len)) - retcode = ERR_LOCAL_ADDR; - - taken_addr = (struct sockaddr *)&odev->net_conf->peer_addr; - if (new_conf->peer_addr_len == odev->net_conf->peer_addr_len && - !memcmp(new_peer_addr, taken_addr, new_conf->peer_addr_len)) - retcode = ERR_PEER_ADDR; - - put_net_conf(odev); - if (retcode != NO_ERROR) - goto fail; - } + if (!rsr) { + crypto_free_hash(tconn->csums_tfm); + tconn->csums_tfm = crypto.csums_tfm; + crypto.csums_tfm = NULL; + } + if (!ovr) { + crypto_free_hash(tconn->verify_tfm); + tconn->verify_tfm = crypto.verify_tfm; + crypto.verify_tfm = NULL; } - if (new_conf->cram_hmac_alg[0] != 0) { - snprintf(hmac_name, CRYPTO_MAX_ALG_NAME, "hmac(%s)", - new_conf->cram_hmac_alg); - tfm = crypto_alloc_hash(hmac_name, 0, CRYPTO_ALG_ASYNC); - if (IS_ERR(tfm)) { - tfm = NULL; - retcode = ERR_AUTH_ALG; - goto fail; - } + crypto_free_hash(tconn->integrity_tfm); + tconn->integrity_tfm = crypto.integrity_tfm; + if (tconn->cstate >= C_WF_REPORT_PARAMS && tconn->agreed_pro_version >= 100) + /* Do this without trying to take tconn->data.mutex again. */ + __drbd_send_protocol(tconn, P_PROTOCOL_UPDATE); - if (!drbd_crypto_is_hash(crypto_hash_tfm(tfm))) { - retcode = ERR_AUTH_ALG_ND; - goto fail; - } - } + crypto_free_hash(tconn->cram_hmac_tfm); + tconn->cram_hmac_tfm = crypto.cram_hmac_tfm; - if (new_conf->integrity_alg[0]) { - integrity_w_tfm = crypto_alloc_hash(new_conf->integrity_alg, 0, CRYPTO_ALG_ASYNC); - if (IS_ERR(integrity_w_tfm)) { - integrity_w_tfm = NULL; - retcode=ERR_INTEGRITY_ALG; - goto fail; - } + mutex_unlock(&tconn->conf_update); + mutex_unlock(&tconn->data.mutex); + synchronize_rcu(); + kfree(old_conf); - if (!drbd_crypto_is_hash(crypto_hash_tfm(integrity_w_tfm))) { - retcode=ERR_INTEGRITY_ALG_ND; - goto fail; - } + if (tconn->cstate >= C_WF_REPORT_PARAMS) + drbd_send_sync_param(minor_to_mdev(conn_lowest_minor(tconn))); - integrity_r_tfm = crypto_alloc_hash(new_conf->integrity_alg, 0, CRYPTO_ALG_ASYNC); - if (IS_ERR(integrity_r_tfm)) { - integrity_r_tfm = NULL; - retcode=ERR_INTEGRITY_ALG; - goto fail; - } + goto done; + + fail: + mutex_unlock(&tconn->conf_update); + mutex_unlock(&tconn->data.mutex); + free_crypto(&crypto); + kfree(new_conf); + done: + conn_reconfig_done(tconn); + out: + drbd_adm_finish(info, retcode); + return 0; +} + +int drbd_adm_connect(struct sk_buff *skb, struct genl_info *info) +{ + struct drbd_conf *mdev; + struct net_conf *old_conf, *new_conf = NULL; + struct crypto crypto = { }; + struct drbd_tconn *tconn; + enum drbd_ret_code retcode; + int i; + int err; + + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_RESOURCE); + + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto out; + if (!(adm_ctx.my_addr && adm_ctx.peer_addr)) { + drbd_msg_put_info("connection endpoint(s) missing"); + retcode = ERR_INVALID_REQUEST; + goto out; } - ns = new_conf->max_epoch_size/8; - if (mdev->tl_hash_s != ns) { - new_tl_hash = kzalloc(ns*sizeof(void *), GFP_KERNEL); - if (!new_tl_hash) { - retcode = ERR_NOMEM; - goto fail; + /* No need for _rcu here. All reconfiguration is + * strictly serialized on genl_lock(). We are protected against + * concurrent reconfiguration/addition/deletion */ + list_for_each_entry(tconn, &drbd_tconns, all_tconn) { + if (nla_len(adm_ctx.my_addr) == tconn->my_addr_len && + !memcmp(nla_data(adm_ctx.my_addr), &tconn->my_addr, tconn->my_addr_len)) { + retcode = ERR_LOCAL_ADDR; + goto out; } - } - ns = new_conf->max_buffers/8; - if (new_conf->two_primaries && (mdev->ee_hash_s != ns)) { - new_ee_hash = kzalloc(ns*sizeof(void *), GFP_KERNEL); - if (!new_ee_hash) { - retcode = ERR_NOMEM; - goto fail; + if (nla_len(adm_ctx.peer_addr) == tconn->peer_addr_len && + !memcmp(nla_data(adm_ctx.peer_addr), &tconn->peer_addr, tconn->peer_addr_len)) { + retcode = ERR_PEER_ADDR; + goto out; } } - ((char *)new_conf->shared_secret)[SHARED_SECRET_MAX-1] = 0; + tconn = adm_ctx.tconn; + conn_reconfig_start(tconn); - if (integrity_w_tfm) { - i = crypto_hash_digestsize(integrity_w_tfm); - int_dig_out = kmalloc(i, GFP_KERNEL); - if (!int_dig_out) { - retcode = ERR_NOMEM; - goto fail; - } - int_dig_in = kmalloc(i, GFP_KERNEL); - if (!int_dig_in) { - retcode = ERR_NOMEM; - goto fail; - } - int_dig_vv = kmalloc(i, GFP_KERNEL); - if (!int_dig_vv) { - retcode = ERR_NOMEM; - goto fail; - } + if (tconn->cstate > C_STANDALONE) { + retcode = ERR_NET_CONFIGURED; + goto fail; } - if (!mdev->bitmap) { - if(drbd_bm_init(mdev)) { - retcode = ERR_NOMEM; - goto fail; - } + /* allocation not in the IO path, drbdsetup / netlink process context */ + new_conf = kzalloc(sizeof(*new_conf), GFP_KERNEL); + if (!new_conf) { + retcode = ERR_NOMEM; + goto fail; } - drbd_flush_workqueue(mdev); - spin_lock_irq(&mdev->req_lock); - if (mdev->net_conf != NULL) { - retcode = ERR_NET_CONFIGURED; - spin_unlock_irq(&mdev->req_lock); + set_net_conf_defaults(new_conf); + + err = net_conf_from_attrs(new_conf, info); + if (err && err != -ENOMSG) { + retcode = ERR_MANDATORY_TAG; + drbd_msg_put_info(from_attrs_err_to_txt(err)); goto fail; } - mdev->net_conf = new_conf; - mdev->send_cnt = 0; - mdev->recv_cnt = 0; + retcode = check_net_options(tconn, new_conf); + if (retcode != NO_ERROR) + goto fail; - if (new_tl_hash) { - kfree(mdev->tl_hash); - mdev->tl_hash_s = mdev->net_conf->max_epoch_size/8; - mdev->tl_hash = new_tl_hash; - } + retcode = alloc_crypto(&crypto, new_conf); + if (retcode != NO_ERROR) + goto fail; + + ((char *)new_conf->shared_secret)[SHARED_SECRET_MAX-1] = 0; + + conn_flush_workqueue(tconn); - if (new_ee_hash) { - kfree(mdev->ee_hash); - mdev->ee_hash_s = mdev->net_conf->max_buffers/8; - mdev->ee_hash = new_ee_hash; + mutex_lock(&tconn->conf_update); + old_conf = tconn->net_conf; + if (old_conf) { + retcode = ERR_NET_CONFIGURED; + mutex_unlock(&tconn->conf_update); + goto fail; } + rcu_assign_pointer(tconn->net_conf, new_conf); - crypto_free_hash(mdev->cram_hmac_tfm); - mdev->cram_hmac_tfm = tfm; + conn_free_crypto(tconn); + tconn->cram_hmac_tfm = crypto.cram_hmac_tfm; + tconn->integrity_tfm = crypto.integrity_tfm; + tconn->csums_tfm = crypto.csums_tfm; + tconn->verify_tfm = crypto.verify_tfm; - crypto_free_hash(mdev->integrity_w_tfm); - mdev->integrity_w_tfm = integrity_w_tfm; + tconn->my_addr_len = nla_len(adm_ctx.my_addr); + memcpy(&tconn->my_addr, nla_data(adm_ctx.my_addr), tconn->my_addr_len); + tconn->peer_addr_len = nla_len(adm_ctx.peer_addr); + memcpy(&tconn->peer_addr, nla_data(adm_ctx.peer_addr), tconn->peer_addr_len); - crypto_free_hash(mdev->integrity_r_tfm); - mdev->integrity_r_tfm = integrity_r_tfm; + mutex_unlock(&tconn->conf_update); - kfree(mdev->int_dig_out); - kfree(mdev->int_dig_in); - kfree(mdev->int_dig_vv); - mdev->int_dig_out=int_dig_out; - mdev->int_dig_in=int_dig_in; - mdev->int_dig_vv=int_dig_vv; - retcode = _drbd_set_state(_NS(mdev, conn, C_UNCONNECTED), CS_VERBOSE, NULL); - spin_unlock_irq(&mdev->req_lock); + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, i) { + mdev->send_cnt = 0; + mdev->recv_cnt = 0; + } + rcu_read_unlock(); - kobject_uevent(&disk_to_dev(mdev->vdisk)->kobj, KOBJ_CHANGE); - reply->ret_code = retcode; - drbd_reconfig_done(mdev); + retcode = conn_request_state(tconn, NS(conn, C_UNCONNECTED), CS_VERBOSE); + + conn_reconfig_done(tconn); + drbd_adm_finish(info, retcode); return 0; fail: - kfree(int_dig_out); - kfree(int_dig_in); - kfree(int_dig_vv); - crypto_free_hash(tfm); - crypto_free_hash(integrity_w_tfm); - crypto_free_hash(integrity_r_tfm); - kfree(new_tl_hash); - kfree(new_ee_hash); + free_crypto(&crypto); kfree(new_conf); - reply->ret_code = retcode; - drbd_reconfig_done(mdev); + conn_reconfig_done(tconn); +out: + drbd_adm_finish(info, retcode); return 0; } -static int drbd_nl_disconnect(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) +static enum drbd_state_rv conn_try_disconnect(struct drbd_tconn *tconn, bool force) { - int retcode; - struct disconnect dc; - - memset(&dc, 0, sizeof(struct disconnect)); - if (!disconnect_from_tags(mdev, nlp->tag_list, &dc)) { - retcode = ERR_MANDATORY_TAG; - goto fail; - } - - if (dc.force) { - spin_lock_irq(&mdev->req_lock); - if (mdev->state.conn >= C_WF_CONNECTION) - _drbd_set_state(_NS(mdev, conn, C_DISCONNECTING), CS_HARD, NULL); - spin_unlock_irq(&mdev->req_lock); - goto done; - } + enum drbd_state_rv rv; - retcode = _drbd_request_state(mdev, NS(conn, C_DISCONNECTING), CS_ORDERED); + rv = conn_request_state(tconn, NS(conn, C_DISCONNECTING), + force ? CS_HARD : 0); - if (retcode == SS_NOTHING_TO_DO) - goto done; - else if (retcode == SS_ALREADY_STANDALONE) - goto done; - else if (retcode == SS_PRIMARY_NOP) { - /* Our statche checking code wants to see the peer outdated. */ - retcode = drbd_request_state(mdev, NS2(conn, C_DISCONNECTING, - pdsk, D_OUTDATED)); - } else if (retcode == SS_CW_FAILED_BY_PEER) { + switch (rv) { + case SS_NOTHING_TO_DO: + break; + case SS_ALREADY_STANDALONE: + return SS_SUCCESS; + case SS_PRIMARY_NOP: + /* Our state checking code wants to see the peer outdated. */ + rv = conn_request_state(tconn, NS2(conn, C_DISCONNECTING, + pdsk, D_OUTDATED), CS_VERBOSE); + break; + case SS_CW_FAILED_BY_PEER: /* The peer probably wants to see us outdated. */ - retcode = _drbd_request_state(mdev, NS2(conn, C_DISCONNECTING, - disk, D_OUTDATED), - CS_ORDERED); - if (retcode == SS_IS_DISKLESS || retcode == SS_LOWER_THAN_OUTDATED) { - drbd_force_state(mdev, NS(conn, C_DISCONNECTING)); - retcode = SS_SUCCESS; + rv = conn_request_state(tconn, NS2(conn, C_DISCONNECTING, + disk, D_OUTDATED), 0); + if (rv == SS_IS_DISKLESS || rv == SS_LOWER_THAN_OUTDATED) { + rv = conn_request_state(tconn, NS(conn, C_DISCONNECTING), + CS_HARD); } + break; + default:; + /* no special handling necessary */ + } + + if (rv >= SS_SUCCESS) { + enum drbd_state_rv rv2; + /* No one else can reconfigure the network while I am here. + * The state handling only uses drbd_thread_stop_nowait(), + * we want to really wait here until the receiver is no more. + */ + drbd_thread_stop(&adm_ctx.tconn->receiver); + + /* Race breaker. This additional state change request may be + * necessary, if this was a forced disconnect during a receiver + * restart. We may have "killed" the receiver thread just + * after drbdd_init() returned. Typically, we should be + * C_STANDALONE already, now, and this becomes a no-op. + */ + rv2 = conn_request_state(tconn, NS(conn, C_STANDALONE), + CS_VERBOSE | CS_HARD); + if (rv2 < SS_SUCCESS) + conn_err(tconn, + "unexpected rv2=%d in conn_try_disconnect()\n", + rv2); } + return rv; +} - if (retcode < SS_SUCCESS) - goto fail; +int drbd_adm_disconnect(struct sk_buff *skb, struct genl_info *info) +{ + struct disconnect_parms parms; + struct drbd_tconn *tconn; + enum drbd_state_rv rv; + enum drbd_ret_code retcode; + int err; - if (wait_event_interruptible(mdev->state_wait, - mdev->state.conn != C_DISCONNECTING)) { - /* Do not test for mdev->state.conn == C_STANDALONE, since - someone else might connect us in the mean time! */ - retcode = ERR_INTR; + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_CONNECTION); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) goto fail; + + tconn = adm_ctx.tconn; + memset(&parms, 0, sizeof(parms)); + if (info->attrs[DRBD_NLA_DISCONNECT_PARMS]) { + err = disconnect_parms_from_attrs(&parms, info); + if (err) { + retcode = ERR_MANDATORY_TAG; + drbd_msg_put_info(from_attrs_err_to_txt(err)); + goto fail; + } } - done: - retcode = NO_ERROR; + rv = conn_try_disconnect(tconn, parms.force_disconnect); + if (rv < SS_SUCCESS) + retcode = rv; /* FIXME: Type mismatch. */ + else + retcode = NO_ERROR; fail: - drbd_md_sync(mdev); - reply->ret_code = retcode; + drbd_adm_finish(info, retcode); return 0; } @@ -1705,7 +2241,7 @@ void resync_after_online_grow(struct drbd_conf *mdev) if (mdev->state.role != mdev->state.peer) iass = (mdev->state.role == R_PRIMARY); else - iass = test_bit(DISCARD_CONCURRENT, &mdev->flags); + iass = test_bit(RESOLVE_CONFLICTS, &mdev->tconn->flags); if (iass) drbd_start_resync(mdev, C_SYNC_SOURCE); @@ -1713,20 +2249,34 @@ void resync_after_online_grow(struct drbd_conf *mdev) _drbd_request_state(mdev, NS(conn, C_WF_SYNC_UUID), CS_VERBOSE + CS_SERIALIZE); } -static int drbd_nl_resize(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) +int drbd_adm_resize(struct sk_buff *skb, struct genl_info *info) { - struct resize rs; - int retcode = NO_ERROR; + struct disk_conf *old_disk_conf, *new_disk_conf = NULL; + struct resize_parms rs; + struct drbd_conf *mdev; + enum drbd_ret_code retcode; enum determine_dev_size dd; enum dds_flags ddsf; + sector_t u_size; + int err; - memset(&rs, 0, sizeof(struct resize)); - if (!resize_from_tags(mdev, nlp->tag_list, &rs)) { - retcode = ERR_MANDATORY_TAG; + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) goto fail; + + memset(&rs, 0, sizeof(struct resize_parms)); + if (info->attrs[DRBD_NLA_RESIZE_PARMS]) { + err = resize_parms_from_attrs(&rs, info); + if (err) { + retcode = ERR_MANDATORY_TAG; + drbd_msg_put_info(from_attrs_err_to_txt(err)); + goto fail; + } } + mdev = adm_ctx.mdev; if (mdev->state.conn > C_CONNECTED) { retcode = ERR_RESIZE_RESYNC; goto fail; @@ -1743,15 +2293,36 @@ static int drbd_nl_resize(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, goto fail; } - if (rs.no_resync && mdev->agreed_pro_version < 93) { + if (rs.no_resync && mdev->tconn->agreed_pro_version < 93) { retcode = ERR_NEED_APV_93; goto fail_ldev; } + rcu_read_lock(); + u_size = rcu_dereference(mdev->ldev->disk_conf)->disk_size; + rcu_read_unlock(); + if (u_size != (sector_t)rs.resize_size) { + new_disk_conf = kmalloc(sizeof(struct disk_conf), GFP_KERNEL); + if (!new_disk_conf) { + retcode = ERR_NOMEM; + goto fail_ldev; + } + } + if (mdev->ldev->known_size != drbd_get_capacity(mdev->ldev->backing_bdev)) mdev->ldev->known_size = drbd_get_capacity(mdev->ldev->backing_bdev); - mdev->ldev->dc.disk_size = (sector_t)rs.resize_size; + if (new_disk_conf) { + mutex_lock(&mdev->tconn->conf_update); + old_disk_conf = mdev->ldev->disk_conf; + *new_disk_conf = *old_disk_conf; + new_disk_conf->disk_size = (sector_t)rs.resize_size; + rcu_assign_pointer(mdev->ldev->disk_conf, new_disk_conf); + mutex_unlock(&mdev->tconn->conf_update); + synchronize_rcu(); + kfree(old_disk_conf); + } + ddsf = (rs.resize_force ? DDSF_FORCED : 0) | (rs.no_resync ? DDSF_NO_RESYNC : 0); dd = drbd_determine_dev_size(mdev, ddsf); drbd_md_sync(mdev); @@ -1770,7 +2341,7 @@ static int drbd_nl_resize(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, } fail: - reply->ret_code = retcode; + drbd_adm_finish(info, retcode); return 0; fail_ldev: @@ -1778,204 +2349,55 @@ static int drbd_nl_resize(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, goto fail; } -static int drbd_nl_syncer_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) +int drbd_adm_resource_opts(struct sk_buff *skb, struct genl_info *info) { - int retcode = NO_ERROR; + enum drbd_ret_code retcode; + struct drbd_tconn *tconn; + struct res_opts res_opts; int err; - int ovr; /* online verify running */ - int rsr; /* re-sync running */ - struct crypto_hash *verify_tfm = NULL; - struct crypto_hash *csums_tfm = NULL; - struct syncer_conf sc; - cpumask_var_t new_cpu_mask; - int *rs_plan_s = NULL; - int fifo_size; - - if (!zalloc_cpumask_var(&new_cpu_mask, GFP_KERNEL)) { - retcode = ERR_NOMEM; - goto fail; - } - if (nlp->flags & DRBD_NL_SET_DEFAULTS) { - memset(&sc, 0, sizeof(struct syncer_conf)); - sc.rate = DRBD_RATE_DEF; - sc.after = DRBD_AFTER_DEF; - sc.al_extents = DRBD_AL_EXTENTS_DEF; - sc.on_no_data = DRBD_ON_NO_DATA_DEF; - sc.c_plan_ahead = DRBD_C_PLAN_AHEAD_DEF; - sc.c_delay_target = DRBD_C_DELAY_TARGET_DEF; - sc.c_fill_target = DRBD_C_FILL_TARGET_DEF; - sc.c_max_rate = DRBD_C_MAX_RATE_DEF; - sc.c_min_rate = DRBD_C_MIN_RATE_DEF; - } else - memcpy(&sc, &mdev->sync_conf, sizeof(struct syncer_conf)); - - if (!syncer_conf_from_tags(mdev, nlp->tag_list, &sc)) { - retcode = ERR_MANDATORY_TAG; - goto fail; - } - - /* re-sync running */ - rsr = ( mdev->state.conn == C_SYNC_SOURCE || - mdev->state.conn == C_SYNC_TARGET || - mdev->state.conn == C_PAUSED_SYNC_S || - mdev->state.conn == C_PAUSED_SYNC_T ); - - if (rsr && strcmp(sc.csums_alg, mdev->sync_conf.csums_alg)) { - retcode = ERR_CSUMS_RESYNC_RUNNING; - goto fail; - } - - if (!rsr && sc.csums_alg[0]) { - csums_tfm = crypto_alloc_hash(sc.csums_alg, 0, CRYPTO_ALG_ASYNC); - if (IS_ERR(csums_tfm)) { - csums_tfm = NULL; - retcode = ERR_CSUMS_ALG; - goto fail; - } - - if (!drbd_crypto_is_hash(crypto_hash_tfm(csums_tfm))) { - retcode = ERR_CSUMS_ALG_ND; - goto fail; - } - } - - /* online verify running */ - ovr = (mdev->state.conn == C_VERIFY_S || mdev->state.conn == C_VERIFY_T); - - if (ovr) { - if (strcmp(sc.verify_alg, mdev->sync_conf.verify_alg)) { - retcode = ERR_VERIFY_RUNNING; - goto fail; - } - } - - if (!ovr && sc.verify_alg[0]) { - verify_tfm = crypto_alloc_hash(sc.verify_alg, 0, CRYPTO_ALG_ASYNC); - if (IS_ERR(verify_tfm)) { - verify_tfm = NULL; - retcode = ERR_VERIFY_ALG; - goto fail; - } - - if (!drbd_crypto_is_hash(crypto_hash_tfm(verify_tfm))) { - retcode = ERR_VERIFY_ALG_ND; - goto fail; - } - } - - /* silently ignore cpu mask on UP kernel */ - if (nr_cpu_ids > 1 && sc.cpu_mask[0] != 0) { - err = bitmap_parse(sc.cpu_mask, 32, - cpumask_bits(new_cpu_mask), nr_cpu_ids); - if (err) { - dev_warn(DEV, "bitmap_parse() failed with %d\n", err); - retcode = ERR_CPU_MASK_PARSE; - goto fail; - } - } - - ERR_IF (sc.rate < 1) sc.rate = 1; - ERR_IF (sc.al_extents < 7) sc.al_extents = 127; /* arbitrary minimum */ -#define AL_MAX ((MD_AL_MAX_SIZE-1) * AL_EXTENTS_PT) - if (sc.al_extents > AL_MAX) { - dev_err(DEV, "sc.al_extents > %d\n", AL_MAX); - sc.al_extents = AL_MAX; - } -#undef AL_MAX - - /* to avoid spurious errors when configuring minors before configuring - * the minors they depend on: if necessary, first create the minor we - * depend on */ - if (sc.after >= 0) - ensure_mdev(sc.after, 1); - - /* most sanity checks done, try to assign the new sync-after - * dependency. need to hold the global lock in there, - * to avoid a race in the dependency loop check. */ - retcode = drbd_alter_sa(mdev, sc.after); + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_RESOURCE); + if (!adm_ctx.reply_skb) + return retcode; if (retcode != NO_ERROR) goto fail; + tconn = adm_ctx.tconn; - fifo_size = (sc.c_plan_ahead * 10 * SLEEP_TIME) / HZ; - if (fifo_size != mdev->rs_plan_s.size && fifo_size > 0) { - rs_plan_s = kzalloc(sizeof(int) * fifo_size, GFP_KERNEL); - if (!rs_plan_s) { - dev_err(DEV, "kmalloc of fifo_buffer failed"); - retcode = ERR_NOMEM; - goto fail; - } - } + res_opts = tconn->res_opts; + if (should_set_defaults(info)) + set_res_opts_defaults(&res_opts); - /* ok, assign the rest of it as well. - * lock against receive_SyncParam() */ - spin_lock(&mdev->peer_seq_lock); - mdev->sync_conf = sc; - - if (!rsr) { - crypto_free_hash(mdev->csums_tfm); - mdev->csums_tfm = csums_tfm; - csums_tfm = NULL; - } - - if (!ovr) { - crypto_free_hash(mdev->verify_tfm); - mdev->verify_tfm = verify_tfm; - verify_tfm = NULL; - } - - if (fifo_size != mdev->rs_plan_s.size) { - kfree(mdev->rs_plan_s.values); - mdev->rs_plan_s.values = rs_plan_s; - mdev->rs_plan_s.size = fifo_size; - mdev->rs_planed = 0; - rs_plan_s = NULL; + err = res_opts_from_attrs(&res_opts, info); + if (err && err != -ENOMSG) { + retcode = ERR_MANDATORY_TAG; + drbd_msg_put_info(from_attrs_err_to_txt(err)); + goto fail; } - spin_unlock(&mdev->peer_seq_lock); - - if (get_ldev(mdev)) { - wait_event(mdev->al_wait, lc_try_lock(mdev->act_log)); - drbd_al_shrink(mdev); - err = drbd_check_al_size(mdev); - lc_unlock(mdev->act_log); - wake_up(&mdev->al_wait); - - put_ldev(mdev); - drbd_md_sync(mdev); - - if (err) { + err = set_resource_options(tconn, &res_opts); + if (err) { + retcode = ERR_INVALID_REQUEST; + if (err == -ENOMEM) retcode = ERR_NOMEM; - goto fail; - } } - if (mdev->state.conn >= C_CONNECTED) - drbd_send_sync_param(mdev, &sc); - - if (!cpumask_equal(mdev->cpu_mask, new_cpu_mask)) { - cpumask_copy(mdev->cpu_mask, new_cpu_mask); - drbd_calc_cpu_mask(mdev); - mdev->receiver.reset_cpu_mask = 1; - mdev->asender.reset_cpu_mask = 1; - mdev->worker.reset_cpu_mask = 1; - } - - kobject_uevent(&disk_to_dev(mdev->vdisk)->kobj, KOBJ_CHANGE); fail: - kfree(rs_plan_s); - free_cpumask_var(new_cpu_mask); - crypto_free_hash(csums_tfm); - crypto_free_hash(verify_tfm); - reply->ret_code = retcode; + drbd_adm_finish(info, retcode); return 0; } -static int drbd_nl_invalidate(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) +int drbd_adm_invalidate(struct sk_buff *skb, struct genl_info *info) { - int retcode; + struct drbd_conf *mdev; + int retcode; /* enum drbd_ret_code rsp. enum drbd_state_rv */ + + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto out; + + mdev = adm_ctx.mdev; /* If there is still bitmap IO pending, probably because of a previous * resync just being finished, wait for it before requesting a new resync. @@ -1990,10 +2412,10 @@ static int drbd_nl_invalidate(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nl retcode = drbd_request_state(mdev, NS(conn, C_STARTING_SYNC_T)); while (retcode == SS_NEED_CONNECTION) { - spin_lock_irq(&mdev->req_lock); + spin_lock_irq(&mdev->tconn->req_lock); if (mdev->state.conn < C_CONNECTED) retcode = _drbd_set_state(_NS(mdev, disk, D_INCONSISTENT), CS_VERBOSE, NULL); - spin_unlock_irq(&mdev->req_lock); + spin_unlock_irq(&mdev->tconn->req_lock); if (retcode != SS_NEED_CONNECTION) break; @@ -2002,7 +2424,25 @@ static int drbd_nl_invalidate(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nl } drbd_resume_io(mdev); - reply->ret_code = retcode; +out: + drbd_adm_finish(info, retcode); + return 0; +} + +static int drbd_adm_simple_request_state(struct sk_buff *skb, struct genl_info *info, + union drbd_state mask, union drbd_state val) +{ + enum drbd_ret_code retcode; + + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto out; + + retcode = drbd_request_state(adm_ctx.mdev, mask, val); +out: + drbd_adm_finish(info, retcode); return 0; } @@ -2015,10 +2455,18 @@ static int drbd_bmio_set_susp_al(struct drbd_conf *mdev) return rv; } -static int drbd_nl_invalidate_peer(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) +int drbd_adm_invalidate_peer(struct sk_buff *skb, struct genl_info *info) { - int retcode; + int retcode; /* drbd_ret_code, drbd_state_rv */ + struct drbd_conf *mdev; + + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto out; + + mdev = adm_ctx.mdev; /* If there is still bitmap IO pending, probably because of a previous * resync just being finished, wait for it before requesting a new resync. @@ -2028,16 +2476,15 @@ static int drbd_nl_invalidate_peer(struct drbd_conf *mdev, struct drbd_nl_cfg_re drbd_flush_workqueue(mdev); retcode = _drbd_request_state(mdev, NS(conn, C_STARTING_SYNC_S), CS_ORDERED); - if (retcode < SS_SUCCESS) { if (retcode == SS_NEED_CONNECTION && mdev->state.role == R_PRIMARY) { - /* The peer will get a resync upon connect anyways. Just make that - into a full resync. */ + /* The peer will get a resync upon connect anyways. + * Just make that into a full resync. */ retcode = drbd_request_state(mdev, NS(pdsk, D_INCONSISTENT)); if (retcode >= SS_SUCCESS) { if (drbd_bitmap_io(mdev, &drbd_bmio_set_susp_al, - "set_n_write from invalidate_peer", - BM_LOCKED_SET_ALLOWED)) + "set_n_write from invalidate_peer", + BM_LOCKED_SET_ALLOWED)) retcode = ERR_IO_MD_DISK; } } else @@ -2045,30 +2492,41 @@ static int drbd_nl_invalidate_peer(struct drbd_conf *mdev, struct drbd_nl_cfg_re } drbd_resume_io(mdev); - reply->ret_code = retcode; +out: + drbd_adm_finish(info, retcode); return 0; } -static int drbd_nl_pause_sync(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) +int drbd_adm_pause_sync(struct sk_buff *skb, struct genl_info *info) { - int retcode = NO_ERROR; + enum drbd_ret_code retcode; - if (drbd_request_state(mdev, NS(user_isp, 1)) == SS_NOTHING_TO_DO) - retcode = ERR_PAUSE_IS_SET; + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto out; - reply->ret_code = retcode; + if (drbd_request_state(adm_ctx.mdev, NS(user_isp, 1)) == SS_NOTHING_TO_DO) + retcode = ERR_PAUSE_IS_SET; +out: + drbd_adm_finish(info, retcode); return 0; } -static int drbd_nl_resume_sync(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) +int drbd_adm_resume_sync(struct sk_buff *skb, struct genl_info *info) { - int retcode = NO_ERROR; - union drbd_state s; + union drbd_dev_state s; + enum drbd_ret_code retcode; - if (drbd_request_state(mdev, NS(user_isp, 0)) == SS_NOTHING_TO_DO) { - s = mdev->state; + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto out; + + if (drbd_request_state(adm_ctx.mdev, NS(user_isp, 0)) == SS_NOTHING_TO_DO) { + s = adm_ctx.mdev->state; if (s.conn == C_PAUSED_SYNC_S || s.conn == C_PAUSED_SYNC_T) { retcode = s.aftr_isp ? ERR_PIC_AFTER_DEP : s.peer_isp ? ERR_PIC_PEER_DEP : ERR_PAUSE_IS_CLEAR; @@ -2077,172 +2535,482 @@ static int drbd_nl_resume_sync(struct drbd_conf *mdev, struct drbd_nl_cfg_req *n } } - reply->ret_code = retcode; +out: + drbd_adm_finish(info, retcode); return 0; } -static int drbd_nl_suspend_io(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) +int drbd_adm_suspend_io(struct sk_buff *skb, struct genl_info *info) { - reply->ret_code = drbd_request_state(mdev, NS(susp, 1)); - - return 0; + return drbd_adm_simple_request_state(skb, info, NS(susp, 1)); } -static int drbd_nl_resume_io(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) +int drbd_adm_resume_io(struct sk_buff *skb, struct genl_info *info) { + struct drbd_conf *mdev; + int retcode; /* enum drbd_ret_code rsp. enum drbd_state_rv */ + + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto out; + + mdev = adm_ctx.mdev; if (test_bit(NEW_CUR_UUID, &mdev->flags)) { drbd_uuid_new_current(mdev); clear_bit(NEW_CUR_UUID, &mdev->flags); } drbd_suspend_io(mdev); - reply->ret_code = drbd_request_state(mdev, NS3(susp, 0, susp_nod, 0, susp_fen, 0)); - if (reply->ret_code == SS_SUCCESS) { + retcode = drbd_request_state(mdev, NS3(susp, 0, susp_nod, 0, susp_fen, 0)); + if (retcode == SS_SUCCESS) { if (mdev->state.conn < C_CONNECTED) - tl_clear(mdev); + tl_clear(mdev->tconn); if (mdev->state.disk == D_DISKLESS || mdev->state.disk == D_FAILED) - tl_restart(mdev, fail_frozen_disk_io); + tl_restart(mdev->tconn, FAIL_FROZEN_DISK_IO); } drbd_resume_io(mdev); +out: + drbd_adm_finish(info, retcode); return 0; } -static int drbd_nl_outdate(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) +int drbd_adm_outdate(struct sk_buff *skb, struct genl_info *info) { - reply->ret_code = drbd_request_state(mdev, NS(disk, D_OUTDATED)); - return 0; + return drbd_adm_simple_request_state(skb, info, NS(disk, D_OUTDATED)); } -static int drbd_nl_get_config(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) +int nla_put_drbd_cfg_context(struct sk_buff *skb, struct drbd_tconn *tconn, unsigned vnr) { - unsigned short *tl; + struct nlattr *nla; + nla = nla_nest_start(skb, DRBD_NLA_CFG_CONTEXT); + if (!nla) + goto nla_put_failure; + if (vnr != VOLUME_UNSPECIFIED && + nla_put_u32(skb, T_ctx_volume, vnr)) + goto nla_put_failure; + if (nla_put_string(skb, T_ctx_resource_name, tconn->name)) + goto nla_put_failure; + if (tconn->my_addr_len && + nla_put(skb, T_ctx_my_addr, tconn->my_addr_len, &tconn->my_addr)) + goto nla_put_failure; + if (tconn->peer_addr_len && + nla_put(skb, T_ctx_peer_addr, tconn->peer_addr_len, &tconn->peer_addr)) + goto nla_put_failure; + nla_nest_end(skb, nla); + return 0; - tl = reply->tag_list; +nla_put_failure: + if (nla) + nla_nest_cancel(skb, nla); + return -EMSGSIZE; +} - if (get_ldev(mdev)) { - tl = disk_conf_to_tags(mdev, &mdev->ldev->dc, tl); - put_ldev(mdev); - } +int nla_put_status_info(struct sk_buff *skb, struct drbd_conf *mdev, + const struct sib_info *sib) +{ + struct state_info *si = NULL; /* for sizeof(si->member); */ + struct net_conf *nc; + struct nlattr *nla; + int got_ldev; + int err = 0; + int exclude_sensitive; + + /* If sib != NULL, this is drbd_bcast_event, which anyone can listen + * to. So we better exclude_sensitive information. + * + * If sib == NULL, this is drbd_adm_get_status, executed synchronously + * in the context of the requesting user process. Exclude sensitive + * information, unless current has superuser. + * + * NOTE: for drbd_adm_get_status_all(), this is a netlink dump, and + * relies on the current implementation of netlink_dump(), which + * executes the dump callback successively from netlink_recvmsg(), + * always in the context of the receiving process */ + exclude_sensitive = sib || !capable(CAP_SYS_ADMIN); + + got_ldev = get_ldev(mdev); + + /* We need to add connection name and volume number information still. + * Minor number is in drbd_genlmsghdr. */ + if (nla_put_drbd_cfg_context(skb, mdev->tconn, mdev->vnr)) + goto nla_put_failure; + + if (res_opts_to_skb(skb, &mdev->tconn->res_opts, exclude_sensitive)) + goto nla_put_failure; + + rcu_read_lock(); + if (got_ldev) + if (disk_conf_to_skb(skb, rcu_dereference(mdev->ldev->disk_conf), exclude_sensitive)) + goto nla_put_failure; + + nc = rcu_dereference(mdev->tconn->net_conf); + if (nc) + err = net_conf_to_skb(skb, nc, exclude_sensitive); + rcu_read_unlock(); + if (err) + goto nla_put_failure; + + nla = nla_nest_start(skb, DRBD_NLA_STATE_INFO); + if (!nla) + goto nla_put_failure; + if (nla_put_u32(skb, T_sib_reason, sib ? sib->sib_reason : SIB_GET_STATUS_REPLY) || + nla_put_u32(skb, T_current_state, mdev->state.i) || + nla_put_u64(skb, T_ed_uuid, mdev->ed_uuid) || + nla_put_u64(skb, T_capacity, drbd_get_capacity(mdev->this_bdev)) || + nla_put_u64(skb, T_send_cnt, mdev->send_cnt) || + nla_put_u64(skb, T_recv_cnt, mdev->recv_cnt) || + nla_put_u64(skb, T_read_cnt, mdev->read_cnt) || + nla_put_u64(skb, T_writ_cnt, mdev->writ_cnt) || + nla_put_u64(skb, T_al_writ_cnt, mdev->al_writ_cnt) || + nla_put_u64(skb, T_bm_writ_cnt, mdev->bm_writ_cnt) || + nla_put_u32(skb, T_ap_bio_cnt, atomic_read(&mdev->ap_bio_cnt)) || + nla_put_u32(skb, T_ap_pending_cnt, atomic_read(&mdev->ap_pending_cnt)) || + nla_put_u32(skb, T_rs_pending_cnt, atomic_read(&mdev->rs_pending_cnt))) + goto nla_put_failure; + + if (got_ldev) { + int err; - if (get_net_conf(mdev)) { - tl = net_conf_to_tags(mdev, mdev->net_conf, tl); - put_net_conf(mdev); + spin_lock_irq(&mdev->ldev->md.uuid_lock); + err = nla_put(skb, T_uuids, sizeof(si->uuids), mdev->ldev->md.uuid); + spin_unlock_irq(&mdev->ldev->md.uuid_lock); + + if (err) + goto nla_put_failure; + + if (nla_put_u32(skb, T_disk_flags, mdev->ldev->md.flags) || + nla_put_u64(skb, T_bits_total, drbd_bm_bits(mdev)) || + nla_put_u64(skb, T_bits_oos, drbd_bm_total_weight(mdev))) + goto nla_put_failure; + if (C_SYNC_SOURCE <= mdev->state.conn && + C_PAUSED_SYNC_T >= mdev->state.conn) { + if (nla_put_u64(skb, T_bits_rs_total, mdev->rs_total) || + nla_put_u64(skb, T_bits_rs_failed, mdev->rs_failed)) + goto nla_put_failure; + } } - tl = syncer_conf_to_tags(mdev, &mdev->sync_conf, tl); - put_unaligned(TT_END, tl++); /* Close the tag list */ + if (sib) { + switch(sib->sib_reason) { + case SIB_SYNC_PROGRESS: + case SIB_GET_STATUS_REPLY: + break; + case SIB_STATE_CHANGE: + if (nla_put_u32(skb, T_prev_state, sib->os.i) || + nla_put_u32(skb, T_new_state, sib->ns.i)) + goto nla_put_failure; + break; + case SIB_HELPER_POST: + if (nla_put_u32(skb, T_helper_exit_code, + sib->helper_exit_code)) + goto nla_put_failure; + /* fall through */ + case SIB_HELPER_PRE: + if (nla_put_string(skb, T_helper, sib->helper_name)) + goto nla_put_failure; + break; + } + } + nla_nest_end(skb, nla); - return (int)((char *)tl - (char *)reply->tag_list); + if (0) +nla_put_failure: + err = -EMSGSIZE; + if (got_ldev) + put_ldev(mdev); + return err; } -static int drbd_nl_get_state(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) +int drbd_adm_get_status(struct sk_buff *skb, struct genl_info *info) { - unsigned short *tl = reply->tag_list; - union drbd_state s = mdev->state; - unsigned long rs_left; - unsigned int res; + enum drbd_ret_code retcode; + int err; - tl = get_state_to_tags(mdev, (struct get_state *)&s, tl); + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto out; - /* no local ref, no bitmap, no syncer progress. */ - if (s.conn >= C_SYNC_SOURCE && s.conn <= C_PAUSED_SYNC_T) { - if (get_ldev(mdev)) { - drbd_get_syncer_progress(mdev, &rs_left, &res); - tl = tl_add_int(tl, T_sync_progress, &res); - put_ldev(mdev); - } + err = nla_put_status_info(adm_ctx.reply_skb, adm_ctx.mdev, NULL); + if (err) { + nlmsg_free(adm_ctx.reply_skb); + return err; } - put_unaligned(TT_END, tl++); /* Close the tag list */ - - return (int)((char *)tl - (char *)reply->tag_list); +out: + drbd_adm_finish(info, retcode); + return 0; } -static int drbd_nl_get_uuids(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) +int get_one_status(struct sk_buff *skb, struct netlink_callback *cb) { - unsigned short *tl; - - tl = reply->tag_list; + struct drbd_conf *mdev; + struct drbd_genlmsghdr *dh; + struct drbd_tconn *pos = (struct drbd_tconn*)cb->args[0]; + struct drbd_tconn *tconn = NULL; + struct drbd_tconn *tmp; + unsigned volume = cb->args[1]; + + /* Open coded, deferred, iteration: + * list_for_each_entry_safe(tconn, tmp, &drbd_tconns, all_tconn) { + * idr_for_each_entry(&tconn->volumes, mdev, i) { + * ... + * } + * } + * where tconn is cb->args[0]; + * and i is cb->args[1]; + * + * cb->args[2] indicates if we shall loop over all resources, + * or just dump all volumes of a single resource. + * + * This may miss entries inserted after this dump started, + * or entries deleted before they are reached. + * + * We need to make sure the mdev won't disappear while + * we are looking at it, and revalidate our iterators + * on each iteration. + */ - if (get_ldev(mdev)) { - tl = tl_add_blob(tl, T_uuids, mdev->ldev->md.uuid, UI_SIZE*sizeof(u64)); - tl = tl_add_int(tl, T_uuids_flags, &mdev->ldev->md.flags); - put_ldev(mdev); + /* synchronize with conn_create()/conn_destroy() */ + rcu_read_lock(); + /* revalidate iterator position */ + list_for_each_entry_rcu(tmp, &drbd_tconns, all_tconn) { + if (pos == NULL) { + /* first iteration */ + pos = tmp; + tconn = pos; + break; + } + if (tmp == pos) { + tconn = pos; + break; + } } - put_unaligned(TT_END, tl++); /* Close the tag list */ + if (tconn) { +next_tconn: + mdev = idr_get_next(&tconn->volumes, &volume); + if (!mdev) { + /* No more volumes to dump on this tconn. + * Advance tconn iterator. */ + pos = list_entry_rcu(tconn->all_tconn.next, + struct drbd_tconn, all_tconn); + /* Did we dump any volume on this tconn yet? */ + if (volume != 0) { + /* If we reached the end of the list, + * or only a single resource dump was requested, + * we are done. */ + if (&pos->all_tconn == &drbd_tconns || cb->args[2]) + goto out; + volume = 0; + tconn = pos; + goto next_tconn; + } + } + + dh = genlmsg_put(skb, NETLINK_CB(cb->skb).portid, + cb->nlh->nlmsg_seq, &drbd_genl_family, + NLM_F_MULTI, DRBD_ADM_GET_STATUS); + if (!dh) + goto out; + + if (!mdev) { + /* This is a tconn without a single volume. + * Suprisingly enough, it may have a network + * configuration. */ + struct net_conf *nc; + dh->minor = -1U; + dh->ret_code = NO_ERROR; + if (nla_put_drbd_cfg_context(skb, tconn, VOLUME_UNSPECIFIED)) + goto cancel; + nc = rcu_dereference(tconn->net_conf); + if (nc && net_conf_to_skb(skb, nc, 1) != 0) + goto cancel; + goto done; + } + + D_ASSERT(mdev->vnr == volume); + D_ASSERT(mdev->tconn == tconn); + + dh->minor = mdev_to_minor(mdev); + dh->ret_code = NO_ERROR; - return (int)((char *)tl - (char *)reply->tag_list); + if (nla_put_status_info(skb, mdev, NULL)) { +cancel: + genlmsg_cancel(skb, dh); + goto out; + } +done: + genlmsg_end(skb, dh); + } + +out: + rcu_read_unlock(); + /* where to start the next iteration */ + cb->args[0] = (long)pos; + cb->args[1] = (pos == tconn) ? volume + 1 : 0; + + /* No more tconns/volumes/minors found results in an empty skb. + * Which will terminate the dump. */ + return skb->len; } -/** - * drbd_nl_get_timeout_flag() - Used by drbdsetup to find out which timeout value to use - * @mdev: DRBD device. - * @nlp: Netlink/connector packet from drbdsetup - * @reply: Reply packet for drbdsetup +/* + * Request status of all resources, or of all volumes within a single resource. + * + * This is a dump, as the answer may not fit in a single reply skb otherwise. + * Which means we cannot use the family->attrbuf or other such members, because + * dump is NOT protected by the genl_lock(). During dump, we only have access + * to the incoming skb, and need to opencode "parsing" of the nlattr payload. + * + * Once things are setup properly, we call into get_one_status(). */ -static int drbd_nl_get_timeout_flag(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) +int drbd_adm_get_status_all(struct sk_buff *skb, struct netlink_callback *cb) { - unsigned short *tl; - char rv; + const unsigned hdrlen = GENL_HDRLEN + GENL_MAGIC_FAMILY_HDRSZ; + struct nlattr *nla; + const char *resource_name; + struct drbd_tconn *tconn; + int maxtype; + + /* Is this a followup call? */ + if (cb->args[0]) { + /* ... of a single resource dump, + * and the resource iterator has been advanced already? */ + if (cb->args[2] && cb->args[2] != cb->args[0]) + return 0; /* DONE. */ + goto dump; + } + + /* First call (from netlink_dump_start). We need to figure out + * which resource(s) the user wants us to dump. */ + nla = nla_find(nlmsg_attrdata(cb->nlh, hdrlen), + nlmsg_attrlen(cb->nlh, hdrlen), + DRBD_NLA_CFG_CONTEXT); + + /* No explicit context given. Dump all. */ + if (!nla) + goto dump; + maxtype = ARRAY_SIZE(drbd_cfg_context_nl_policy) - 1; + nla = drbd_nla_find_nested(maxtype, nla, __nla_type(T_ctx_resource_name)); + if (IS_ERR(nla)) + return PTR_ERR(nla); + /* context given, but no name present? */ + if (!nla) + return -EINVAL; + resource_name = nla_data(nla); + tconn = conn_get_by_name(resource_name); + + if (!tconn) + return -ENODEV; + + kref_put(&tconn->kref, &conn_destroy); /* get_one_status() (re)validates tconn by itself */ + + /* prime iterators, and set "filter" mode mark: + * only dump this tconn. */ + cb->args[0] = (long)tconn; + /* cb->args[1] = 0; passed in this way. */ + cb->args[2] = (long)tconn; + +dump: + return get_one_status(skb, cb); +} - tl = reply->tag_list; +int drbd_adm_get_timeout_type(struct sk_buff *skb, struct genl_info *info) +{ + enum drbd_ret_code retcode; + struct timeout_parms tp; + int err; - rv = mdev->state.pdsk == D_OUTDATED ? UT_PEER_OUTDATED : - test_bit(USE_DEGR_WFC_T, &mdev->flags) ? UT_DEGRADED : UT_DEFAULT; + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto out; - tl = tl_add_blob(tl, T_use_degraded, &rv, sizeof(rv)); - put_unaligned(TT_END, tl++); /* Close the tag list */ + tp.timeout_type = + adm_ctx.mdev->state.pdsk == D_OUTDATED ? UT_PEER_OUTDATED : + test_bit(USE_DEGR_WFC_T, &adm_ctx.mdev->flags) ? UT_DEGRADED : + UT_DEFAULT; - return (int)((char *)tl - (char *)reply->tag_list); + err = timeout_parms_to_priv_skb(adm_ctx.reply_skb, &tp); + if (err) { + nlmsg_free(adm_ctx.reply_skb); + return err; + } +out: + drbd_adm_finish(info, retcode); + return 0; } -static int drbd_nl_start_ov(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) +int drbd_adm_start_ov(struct sk_buff *skb, struct genl_info *info) { - /* default to resume from last known position, if possible */ - struct start_ov args = - { .start_sector = mdev->ov_start_sector }; + struct drbd_conf *mdev; + enum drbd_ret_code retcode; + struct start_ov_parms parms; - if (!start_ov_from_tags(mdev, nlp->tag_list, &args)) { - reply->ret_code = ERR_MANDATORY_TAG; - return 0; + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto out; + + mdev = adm_ctx.mdev; + + /* resume from last known position, if possible */ + parms.ov_start_sector = mdev->ov_start_sector; + parms.ov_stop_sector = ULLONG_MAX; + if (info->attrs[DRBD_NLA_START_OV_PARMS]) { + int err = start_ov_parms_from_attrs(&parms, info); + if (err) { + retcode = ERR_MANDATORY_TAG; + drbd_msg_put_info(from_attrs_err_to_txt(err)); + goto out; + } } + /* w_make_ov_request expects position to be aligned */ + mdev->ov_start_sector = parms.ov_start_sector & ~(BM_SECT_PER_BIT-1); + mdev->ov_stop_sector = parms.ov_stop_sector; /* If there is still bitmap IO pending, e.g. previous resync or verify * just being finished, wait for it before requesting a new resync. */ drbd_suspend_io(mdev); wait_event(mdev->misc_wait, !test_bit(BITMAP_IO, &mdev->flags)); - - /* w_make_ov_request expects position to be aligned */ - mdev->ov_start_sector = args.start_sector & ~BM_SECT_PER_BIT; - reply->ret_code = drbd_request_state(mdev,NS(conn,C_VERIFY_S)); + retcode = drbd_request_state(mdev,NS(conn,C_VERIFY_S)); drbd_resume_io(mdev); +out: + drbd_adm_finish(info, retcode); return 0; } -static int drbd_nl_new_c_uuid(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp, - struct drbd_nl_cfg_reply *reply) +int drbd_adm_new_c_uuid(struct sk_buff *skb, struct genl_info *info) { - int retcode = NO_ERROR; + struct drbd_conf *mdev; + enum drbd_ret_code retcode; int skip_initial_sync = 0; int err; + struct new_c_uuid_parms args; - struct new_c_uuid args; + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto out_nolock; - memset(&args, 0, sizeof(struct new_c_uuid)); - if (!new_c_uuid_from_tags(mdev, nlp->tag_list, &args)) { - reply->ret_code = ERR_MANDATORY_TAG; - return 0; + mdev = adm_ctx.mdev; + memset(&args, 0, sizeof(args)); + if (info->attrs[DRBD_NLA_NEW_C_UUID_PARMS]) { + err = new_c_uuid_parms_from_attrs(&args, info); + if (err) { + retcode = ERR_MANDATORY_TAG; + drbd_msg_put_info(from_attrs_err_to_txt(err)); + goto out_nolock; + } } - mutex_lock(&mdev->state_mutex); /* Protects us against serialized state changes. */ + mutex_lock(mdev->state_mutex); /* Protects us against serialized state changes. */ if (!get_ldev(mdev)) { retcode = ERR_NO_DISK; @@ -2250,7 +3018,7 @@ static int drbd_nl_new_c_uuid(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nl } /* this is "skip initial sync", assume to be clean */ - if (mdev->state.conn == C_CONNECTED && mdev->agreed_pro_version >= 90 && + if (mdev->state.conn == C_CONNECTED && mdev->tconn->agreed_pro_version >= 90 && mdev->ldev->md.uuid[UI_CURRENT] == UUID_JUST_CREATED && args.clear_bm) { dev_info(DEV, "Preparing to skip initial sync\n"); skip_initial_sync = 1; @@ -2273,10 +3041,10 @@ static int drbd_nl_new_c_uuid(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nl drbd_send_uuids_skip_initial_sync(mdev); _drbd_uuid_set(mdev, UI_BITMAP, 0); drbd_print_uuids(mdev, "cleared bitmap UUID"); - spin_lock_irq(&mdev->req_lock); + spin_lock_irq(&mdev->tconn->req_lock); _drbd_set_state(_NS2(mdev, disk, D_UP_TO_DATE, pdsk, D_UP_TO_DATE), CS_VERBOSE, NULL); - spin_unlock_irq(&mdev->req_lock); + spin_unlock_irq(&mdev->tconn->req_lock); } } @@ -2284,416 +3052,284 @@ static int drbd_nl_new_c_uuid(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nl out_dec: put_ldev(mdev); out: - mutex_unlock(&mdev->state_mutex); - - reply->ret_code = retcode; + mutex_unlock(mdev->state_mutex); +out_nolock: + drbd_adm_finish(info, retcode); return 0; } -struct cn_handler_struct { - int (*function)(struct drbd_conf *, - struct drbd_nl_cfg_req *, - struct drbd_nl_cfg_reply *); - int reply_body_size; -}; - -static struct cn_handler_struct cnd_table[] = { - [ P_primary ] = { &drbd_nl_primary, 0 }, - [ P_secondary ] = { &drbd_nl_secondary, 0 }, - [ P_disk_conf ] = { &drbd_nl_disk_conf, 0 }, - [ P_detach ] = { &drbd_nl_detach, 0 }, - [ P_net_conf ] = { &drbd_nl_net_conf, 0 }, - [ P_disconnect ] = { &drbd_nl_disconnect, 0 }, - [ P_resize ] = { &drbd_nl_resize, 0 }, - [ P_syncer_conf ] = { &drbd_nl_syncer_conf, 0 }, - [ P_invalidate ] = { &drbd_nl_invalidate, 0 }, - [ P_invalidate_peer ] = { &drbd_nl_invalidate_peer, 0 }, - [ P_pause_sync ] = { &drbd_nl_pause_sync, 0 }, - [ P_resume_sync ] = { &drbd_nl_resume_sync, 0 }, - [ P_suspend_io ] = { &drbd_nl_suspend_io, 0 }, - [ P_resume_io ] = { &drbd_nl_resume_io, 0 }, - [ P_outdate ] = { &drbd_nl_outdate, 0 }, - [ P_get_config ] = { &drbd_nl_get_config, - sizeof(struct syncer_conf_tag_len_struct) + - sizeof(struct disk_conf_tag_len_struct) + - sizeof(struct net_conf_tag_len_struct) }, - [ P_get_state ] = { &drbd_nl_get_state, - sizeof(struct get_state_tag_len_struct) + - sizeof(struct sync_progress_tag_len_struct) }, - [ P_get_uuids ] = { &drbd_nl_get_uuids, - sizeof(struct get_uuids_tag_len_struct) }, - [ P_get_timeout_flag ] = { &drbd_nl_get_timeout_flag, - sizeof(struct get_timeout_flag_tag_len_struct)}, - [ P_start_ov ] = { &drbd_nl_start_ov, 0 }, - [ P_new_c_uuid ] = { &drbd_nl_new_c_uuid, 0 }, -}; - -static void drbd_connector_callback(struct cn_msg *req, struct netlink_skb_parms *nsp) +static enum drbd_ret_code +drbd_check_resource_name(const char *name) { - struct drbd_nl_cfg_req *nlp = (struct drbd_nl_cfg_req *)req->data; - struct cn_handler_struct *cm; - struct cn_msg *cn_reply; - struct drbd_nl_cfg_reply *reply; - struct drbd_conf *mdev; - int retcode, rr; - int reply_size = sizeof(struct cn_msg) - + sizeof(struct drbd_nl_cfg_reply) - + sizeof(short int); - - if (!try_module_get(THIS_MODULE)) { - printk(KERN_ERR "drbd: try_module_get() failed!\n"); - return; + if (!name || !name[0]) { + drbd_msg_put_info("resource name missing"); + return ERR_MANDATORY_TAG; } - - if (!capable(CAP_SYS_ADMIN)) { - retcode = ERR_PERM; - goto fail; - } - - mdev = ensure_mdev(nlp->drbd_minor, - (nlp->flags & DRBD_NL_CREATE_DEVICE)); - if (!mdev) { - retcode = ERR_MINOR_INVALID; - goto fail; + /* if we want to use these in sysfs/configfs/debugfs some day, + * we must not allow slashes */ + if (strchr(name, '/')) { + drbd_msg_put_info("invalid resource name"); + return ERR_INVALID_REQUEST; } + return NO_ERROR; +} - if (nlp->packet_type >= P_nl_after_last_packet || - nlp->packet_type == P_return_code_only) { - retcode = ERR_PACKET_NR; - goto fail; - } +int drbd_adm_new_resource(struct sk_buff *skb, struct genl_info *info) +{ + enum drbd_ret_code retcode; + struct res_opts res_opts; + int err; - cm = cnd_table + nlp->packet_type; + retcode = drbd_adm_prepare(skb, info, 0); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto out; - /* This may happen if packet number is 0: */ - if (cm->function == NULL) { - retcode = ERR_PACKET_NR; - goto fail; + set_res_opts_defaults(&res_opts); + err = res_opts_from_attrs(&res_opts, info); + if (err && err != -ENOMSG) { + retcode = ERR_MANDATORY_TAG; + drbd_msg_put_info(from_attrs_err_to_txt(err)); + goto out; } - reply_size += cm->reply_body_size; + retcode = drbd_check_resource_name(adm_ctx.resource_name); + if (retcode != NO_ERROR) + goto out; - /* allocation not in the IO path, cqueue thread context */ - cn_reply = kzalloc(reply_size, GFP_KERNEL); - if (!cn_reply) { - retcode = ERR_NOMEM; - goto fail; + if (adm_ctx.tconn) { + if (info->nlhdr->nlmsg_flags & NLM_F_EXCL) { + retcode = ERR_INVALID_REQUEST; + drbd_msg_put_info("resource exists"); + } + /* else: still NO_ERROR */ + goto out; } - reply = (struct drbd_nl_cfg_reply *) cn_reply->data; - - reply->packet_type = - cm->reply_body_size ? nlp->packet_type : P_return_code_only; - reply->minor = nlp->drbd_minor; - reply->ret_code = NO_ERROR; /* Might by modified by cm->function. */ - /* reply->tag_list; might be modified by cm->function. */ - - rr = cm->function(mdev, nlp, reply); - - cn_reply->id = req->id; - cn_reply->seq = req->seq; - cn_reply->ack = req->ack + 1; - cn_reply->len = sizeof(struct drbd_nl_cfg_reply) + rr; - cn_reply->flags = 0; - - rr = cn_netlink_send(cn_reply, CN_IDX_DRBD, GFP_KERNEL); - if (rr && rr != -ESRCH) - printk(KERN_INFO "drbd: cn_netlink_send()=%d\n", rr); - kfree(cn_reply); - module_put(THIS_MODULE); - return; - fail: - drbd_nl_send_reply(req, retcode); - module_put(THIS_MODULE); + if (!conn_create(adm_ctx.resource_name, &res_opts)) + retcode = ERR_NOMEM; +out: + drbd_adm_finish(info, retcode); + return 0; } -static atomic_t drbd_nl_seq = ATOMIC_INIT(2); /* two. */ - -static unsigned short * -__tl_add_blob(unsigned short *tl, enum drbd_tags tag, const void *data, - unsigned short len, int nul_terminated) +int drbd_adm_add_minor(struct sk_buff *skb, struct genl_info *info) { - unsigned short l = tag_descriptions[tag_number(tag)].max_len; - len = (len < l) ? len : l; - put_unaligned(tag, tl++); - put_unaligned(len, tl++); - memcpy(tl, data, len); - tl = (unsigned short*)((char*)tl + len); - if (nul_terminated) - *((char*)tl - 1) = 0; - return tl; -} + struct drbd_genlmsghdr *dh = info->userhdr; + enum drbd_ret_code retcode; -static unsigned short * -tl_add_blob(unsigned short *tl, enum drbd_tags tag, const void *data, int len) -{ - return __tl_add_blob(tl, tag, data, len, 0); -} + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_RESOURCE); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto out; -static unsigned short * -tl_add_str(unsigned short *tl, enum drbd_tags tag, const char *str) -{ - return __tl_add_blob(tl, tag, str, strlen(str)+1, 0); -} + if (dh->minor > MINORMASK) { + drbd_msg_put_info("requested minor out of range"); + retcode = ERR_INVALID_REQUEST; + goto out; + } + if (adm_ctx.volume > DRBD_VOLUME_MAX) { + drbd_msg_put_info("requested volume id out of range"); + retcode = ERR_INVALID_REQUEST; + goto out; + } -static unsigned short * -tl_add_int(unsigned short *tl, enum drbd_tags tag, const void *val) -{ - put_unaligned(tag, tl++); - switch(tag_type(tag)) { - case TT_INTEGER: - put_unaligned(sizeof(int), tl++); - put_unaligned(*(int *)val, (int *)tl); - tl = (unsigned short*)((char*)tl+sizeof(int)); - break; - case TT_INT64: - put_unaligned(sizeof(u64), tl++); - put_unaligned(*(u64 *)val, (u64 *)tl); - tl = (unsigned short*)((char*)tl+sizeof(u64)); - break; - default: - /* someone did something stupid. */ - ; + /* drbd_adm_prepare made sure already + * that mdev->tconn and mdev->vnr match the request. */ + if (adm_ctx.mdev) { + if (info->nlhdr->nlmsg_flags & NLM_F_EXCL) + retcode = ERR_MINOR_EXISTS; + /* else: still NO_ERROR */ + goto out; } - return tl; + + retcode = conn_new_minor(adm_ctx.tconn, dh->minor, adm_ctx.volume); +out: + drbd_adm_finish(info, retcode); + return 0; } -void drbd_bcast_state(struct drbd_conf *mdev, union drbd_state state) +static enum drbd_ret_code adm_delete_minor(struct drbd_conf *mdev) { - char buffer[sizeof(struct cn_msg)+ - sizeof(struct drbd_nl_cfg_reply)+ - sizeof(struct get_state_tag_len_struct)+ - sizeof(short int)]; - struct cn_msg *cn_reply = (struct cn_msg *) buffer; - struct drbd_nl_cfg_reply *reply = - (struct drbd_nl_cfg_reply *)cn_reply->data; - unsigned short *tl = reply->tag_list; - - /* dev_warn(DEV, "drbd_bcast_state() got called\n"); */ - - tl = get_state_to_tags(mdev, (struct get_state *)&state, tl); - - put_unaligned(TT_END, tl++); /* Close the tag list */ - - cn_reply->id.idx = CN_IDX_DRBD; - cn_reply->id.val = CN_VAL_DRBD; - - cn_reply->seq = atomic_add_return(1, &drbd_nl_seq); - cn_reply->ack = 0; /* not used here. */ - cn_reply->len = sizeof(struct drbd_nl_cfg_reply) + - (int)((char *)tl - (char *)reply->tag_list); - cn_reply->flags = 0; - - reply->packet_type = P_get_state; - reply->minor = mdev_to_minor(mdev); - reply->ret_code = NO_ERROR; - - cn_netlink_send(cn_reply, CN_IDX_DRBD, GFP_NOIO); + if (mdev->state.disk == D_DISKLESS && + /* no need to be mdev->state.conn == C_STANDALONE && + * we may want to delete a minor from a live replication group. + */ + mdev->state.role == R_SECONDARY) { + _drbd_request_state(mdev, NS(conn, C_WF_REPORT_PARAMS), + CS_VERBOSE + CS_WAIT_COMPLETE); + idr_remove(&mdev->tconn->volumes, mdev->vnr); + idr_remove(&minors, mdev_to_minor(mdev)); + del_gendisk(mdev->vdisk); + synchronize_rcu(); + kref_put(&mdev->kref, &drbd_minor_destroy); + return NO_ERROR; + } else + return ERR_MINOR_CONFIGURED; } -void drbd_bcast_ev_helper(struct drbd_conf *mdev, char *helper_name) +int drbd_adm_delete_minor(struct sk_buff *skb, struct genl_info *info) { - char buffer[sizeof(struct cn_msg)+ - sizeof(struct drbd_nl_cfg_reply)+ - sizeof(struct call_helper_tag_len_struct)+ - sizeof(short int)]; - struct cn_msg *cn_reply = (struct cn_msg *) buffer; - struct drbd_nl_cfg_reply *reply = - (struct drbd_nl_cfg_reply *)cn_reply->data; - unsigned short *tl = reply->tag_list; - - /* dev_warn(DEV, "drbd_bcast_state() got called\n"); */ - - tl = tl_add_str(tl, T_helper, helper_name); - put_unaligned(TT_END, tl++); /* Close the tag list */ - - cn_reply->id.idx = CN_IDX_DRBD; - cn_reply->id.val = CN_VAL_DRBD; - - cn_reply->seq = atomic_add_return(1, &drbd_nl_seq); - cn_reply->ack = 0; /* not used here. */ - cn_reply->len = sizeof(struct drbd_nl_cfg_reply) + - (int)((char *)tl - (char *)reply->tag_list); - cn_reply->flags = 0; + enum drbd_ret_code retcode; - reply->packet_type = P_call_helper; - reply->minor = mdev_to_minor(mdev); - reply->ret_code = NO_ERROR; + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto out; - cn_netlink_send(cn_reply, CN_IDX_DRBD, GFP_NOIO); + retcode = adm_delete_minor(adm_ctx.mdev); +out: + drbd_adm_finish(info, retcode); + return 0; } -void drbd_bcast_ee(struct drbd_conf *mdev, - const char *reason, const int dgs, - const char* seen_hash, const char* calc_hash, - const struct drbd_epoch_entry* e) +int drbd_adm_down(struct sk_buff *skb, struct genl_info *info) { - struct cn_msg *cn_reply; - struct drbd_nl_cfg_reply *reply; - unsigned short *tl; - struct page *page; - unsigned len; + int retcode; /* enum drbd_ret_code rsp. enum drbd_state_rv */ + struct drbd_conf *mdev; + unsigned i; - if (!e) - return; - if (!reason || !reason[0]) - return; + retcode = drbd_adm_prepare(skb, info, 0); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto out; - /* apparently we have to memcpy twice, first to prepare the data for the - * struct cn_msg, then within cn_netlink_send from the cn_msg to the - * netlink skb. */ - /* receiver thread context, which is not in the writeout path (of this node), - * but may be in the writeout path of the _other_ node. - * GFP_NOIO to avoid potential "distributed deadlock". */ - cn_reply = kzalloc( - sizeof(struct cn_msg)+ - sizeof(struct drbd_nl_cfg_reply)+ - sizeof(struct dump_ee_tag_len_struct)+ - sizeof(short int), - GFP_NOIO); - - if (!cn_reply) { - dev_err(DEV, "could not kmalloc buffer for drbd_bcast_ee, sector %llu, size %u\n", - (unsigned long long)e->sector, e->size); - return; + if (!adm_ctx.tconn) { + retcode = ERR_RES_NOT_KNOWN; + goto out; } - reply = (struct drbd_nl_cfg_reply*)cn_reply->data; - tl = reply->tag_list; - - tl = tl_add_str(tl, T_dump_ee_reason, reason); - tl = tl_add_blob(tl, T_seen_digest, seen_hash, dgs); - tl = tl_add_blob(tl, T_calc_digest, calc_hash, dgs); - tl = tl_add_int(tl, T_ee_sector, &e->sector); - tl = tl_add_int(tl, T_ee_block_id, &e->block_id); - - /* dump the first 32k */ - len = min_t(unsigned, e->size, 32 << 10); - put_unaligned(T_ee_data, tl++); - put_unaligned(len, tl++); - - page = e->pages; - page_chain_for_each(page) { - void *d = kmap_atomic(page); - unsigned l = min_t(unsigned, len, PAGE_SIZE); - memcpy(tl, d, l); - kunmap_atomic(d); - tl = (unsigned short*)((char*)tl + l); - len -= l; - if (len == 0) - break; + /* demote */ + idr_for_each_entry(&adm_ctx.tconn->volumes, mdev, i) { + retcode = drbd_set_role(mdev, R_SECONDARY, 0); + if (retcode < SS_SUCCESS) { + drbd_msg_put_info("failed to demote"); + goto out; + } } - put_unaligned(TT_END, tl++); /* Close the tag list */ - - cn_reply->id.idx = CN_IDX_DRBD; - cn_reply->id.val = CN_VAL_DRBD; - - cn_reply->seq = atomic_add_return(1,&drbd_nl_seq); - cn_reply->ack = 0; // not used here. - cn_reply->len = sizeof(struct drbd_nl_cfg_reply) + - (int)((char*)tl - (char*)reply->tag_list); - cn_reply->flags = 0; - - reply->packet_type = P_dump_ee; - reply->minor = mdev_to_minor(mdev); - reply->ret_code = NO_ERROR; - cn_netlink_send(cn_reply, CN_IDX_DRBD, GFP_NOIO); - kfree(cn_reply); -} - -void drbd_bcast_sync_progress(struct drbd_conf *mdev) -{ - char buffer[sizeof(struct cn_msg)+ - sizeof(struct drbd_nl_cfg_reply)+ - sizeof(struct sync_progress_tag_len_struct)+ - sizeof(short int)]; - struct cn_msg *cn_reply = (struct cn_msg *) buffer; - struct drbd_nl_cfg_reply *reply = - (struct drbd_nl_cfg_reply *)cn_reply->data; - unsigned short *tl = reply->tag_list; - unsigned long rs_left; - unsigned int res; + retcode = conn_try_disconnect(adm_ctx.tconn, 0); + if (retcode < SS_SUCCESS) { + drbd_msg_put_info("failed to disconnect"); + goto out; + } - /* no local ref, no bitmap, no syncer progress, no broadcast. */ - if (!get_ldev(mdev)) - return; - drbd_get_syncer_progress(mdev, &rs_left, &res); - put_ldev(mdev); + /* detach */ + idr_for_each_entry(&adm_ctx.tconn->volumes, mdev, i) { + retcode = adm_detach(mdev, 0); + if (retcode < SS_SUCCESS || retcode > NO_ERROR) { + drbd_msg_put_info("failed to detach"); + goto out; + } + } - tl = tl_add_int(tl, T_sync_progress, &res); - put_unaligned(TT_END, tl++); /* Close the tag list */ + /* If we reach this, all volumes (of this tconn) are Secondary, + * Disconnected, Diskless, aka Unconfigured. Make sure all threads have + * actually stopped, state handling only does drbd_thread_stop_nowait(). */ + drbd_thread_stop(&adm_ctx.tconn->worker); - cn_reply->id.idx = CN_IDX_DRBD; - cn_reply->id.val = CN_VAL_DRBD; + /* Now, nothing can fail anymore */ - cn_reply->seq = atomic_add_return(1, &drbd_nl_seq); - cn_reply->ack = 0; /* not used here. */ - cn_reply->len = sizeof(struct drbd_nl_cfg_reply) + - (int)((char *)tl - (char *)reply->tag_list); - cn_reply->flags = 0; + /* delete volumes */ + idr_for_each_entry(&adm_ctx.tconn->volumes, mdev, i) { + retcode = adm_delete_minor(mdev); + if (retcode != NO_ERROR) { + /* "can not happen" */ + drbd_msg_put_info("failed to delete volume"); + goto out; + } + } - reply->packet_type = P_sync_progress; - reply->minor = mdev_to_minor(mdev); - reply->ret_code = NO_ERROR; + /* delete connection */ + if (conn_lowest_minor(adm_ctx.tconn) < 0) { + list_del_rcu(&adm_ctx.tconn->all_tconn); + synchronize_rcu(); + kref_put(&adm_ctx.tconn->kref, &conn_destroy); - cn_netlink_send(cn_reply, CN_IDX_DRBD, GFP_NOIO); + retcode = NO_ERROR; + } else { + /* "can not happen" */ + retcode = ERR_RES_IN_USE; + drbd_msg_put_info("failed to delete connection"); + } + goto out; +out: + drbd_adm_finish(info, retcode); + return 0; } -int __init drbd_nl_init(void) +int drbd_adm_del_resource(struct sk_buff *skb, struct genl_info *info) { - static struct cb_id cn_id_drbd; - int err, try=10; + enum drbd_ret_code retcode; - cn_id_drbd.val = CN_VAL_DRBD; - do { - cn_id_drbd.idx = cn_idx; - err = cn_add_callback(&cn_id_drbd, "cn_drbd", &drbd_connector_callback); - if (!err) - break; - cn_idx = (cn_idx + CN_IDX_STEP); - } while (try--); + retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_RESOURCE); + if (!adm_ctx.reply_skb) + return retcode; + if (retcode != NO_ERROR) + goto out; - if (err) { - printk(KERN_ERR "drbd: cn_drbd failed to register\n"); - return err; + if (conn_lowest_minor(adm_ctx.tconn) < 0) { + list_del_rcu(&adm_ctx.tconn->all_tconn); + synchronize_rcu(); + kref_put(&adm_ctx.tconn->kref, &conn_destroy); + + retcode = NO_ERROR; + } else { + retcode = ERR_RES_IN_USE; } + if (retcode == NO_ERROR) + drbd_thread_stop(&adm_ctx.tconn->worker); +out: + drbd_adm_finish(info, retcode); return 0; } -void drbd_nl_cleanup(void) +void drbd_bcast_event(struct drbd_conf *mdev, const struct sib_info *sib) { - static struct cb_id cn_id_drbd; - - cn_id_drbd.idx = cn_idx; - cn_id_drbd.val = CN_VAL_DRBD; - - cn_del_callback(&cn_id_drbd); -} + static atomic_t drbd_genl_seq = ATOMIC_INIT(2); /* two. */ + struct sk_buff *msg; + struct drbd_genlmsghdr *d_out; + unsigned seq; + int err = -ENOMEM; + + if (sib->sib_reason == SIB_SYNC_PROGRESS) { + if (time_after(jiffies, mdev->rs_last_bcast + HZ)) + mdev->rs_last_bcast = jiffies; + else + return; + } -void drbd_nl_send_reply(struct cn_msg *req, int ret_code) -{ - char buffer[sizeof(struct cn_msg)+sizeof(struct drbd_nl_cfg_reply)]; - struct cn_msg *cn_reply = (struct cn_msg *) buffer; - struct drbd_nl_cfg_reply *reply = - (struct drbd_nl_cfg_reply *)cn_reply->data; - int rr; + seq = atomic_inc_return(&drbd_genl_seq); + msg = genlmsg_new(NLMSG_GOODSIZE, GFP_NOIO); + if (!msg) + goto failed; - memset(buffer, 0, sizeof(buffer)); - cn_reply->id = req->id; + err = -EMSGSIZE; + d_out = genlmsg_put(msg, 0, seq, &drbd_genl_family, 0, DRBD_EVENT); + if (!d_out) /* cannot happen, but anyways. */ + goto nla_put_failure; + d_out->minor = mdev_to_minor(mdev); + d_out->ret_code = NO_ERROR; - cn_reply->seq = req->seq; - cn_reply->ack = req->ack + 1; - cn_reply->len = sizeof(struct drbd_nl_cfg_reply); - cn_reply->flags = 0; + if (nla_put_status_info(msg, mdev, sib)) + goto nla_put_failure; + genlmsg_end(msg, d_out); + err = drbd_genl_multicast_events(msg, 0); + /* msg has been consumed or freed in netlink_broadcast() */ + if (err && err != -ESRCH) + goto failed; - reply->packet_type = P_return_code_only; - reply->minor = ((struct drbd_nl_cfg_req *)req->data)->drbd_minor; - reply->ret_code = ret_code; + return; - rr = cn_netlink_send(cn_reply, CN_IDX_DRBD, GFP_NOIO); - if (rr && rr != -ESRCH) - printk(KERN_INFO "drbd: cn_netlink_send()=%d\n", rr); +nla_put_failure: + nlmsg_free(msg); +failed: + dev_err(DEV, "Error %d while broadcasting event. " + "Event seq:%u sib_reason:%u\n", + err, seq, sib->sib_reason); } - diff --git a/drivers/block/drbd/drbd_nla.c b/drivers/block/drbd/drbd_nla.c new file mode 100644 index 000000000000..fa672b6df8d6 --- /dev/null +++ b/drivers/block/drbd/drbd_nla.c @@ -0,0 +1,55 @@ +#include "drbd_wrappers.h" +#include <linux/kernel.h> +#include <net/netlink.h> +#include <linux/drbd_genl_api.h> +#include "drbd_nla.h" + +static int drbd_nla_check_mandatory(int maxtype, struct nlattr *nla) +{ + struct nlattr *head = nla_data(nla); + int len = nla_len(nla); + int rem; + + /* + * validate_nla (called from nla_parse_nested) ignores attributes + * beyond maxtype, and does not understand the DRBD_GENLA_F_MANDATORY flag. + * In order to have it validate attributes with the DRBD_GENLA_F_MANDATORY + * flag set also, check and remove that flag before calling + * nla_parse_nested. + */ + + nla_for_each_attr(nla, head, len, rem) { + if (nla->nla_type & DRBD_GENLA_F_MANDATORY) { + nla->nla_type &= ~DRBD_GENLA_F_MANDATORY; + if (nla_type(nla) > maxtype) + return -EOPNOTSUPP; + } + } + return 0; +} + +int drbd_nla_parse_nested(struct nlattr *tb[], int maxtype, struct nlattr *nla, + const struct nla_policy *policy) +{ + int err; + + err = drbd_nla_check_mandatory(maxtype, nla); + if (!err) + err = nla_parse_nested(tb, maxtype, nla, policy); + + return err; +} + +struct nlattr *drbd_nla_find_nested(int maxtype, struct nlattr *nla, int attrtype) +{ + int err; + /* + * If any nested attribute has the DRBD_GENLA_F_MANDATORY flag set and + * we don't know about that attribute, reject all the nested + * attributes. + */ + err = drbd_nla_check_mandatory(maxtype, nla); + if (err) + return ERR_PTR(err); + return nla_find_nested(nla, attrtype); +} diff --git a/drivers/block/drbd/drbd_nla.h b/drivers/block/drbd/drbd_nla.h new file mode 100644 index 000000000000..679c2d5b4535 --- /dev/null +++ b/drivers/block/drbd/drbd_nla.h @@ -0,0 +1,8 @@ +#ifndef __DRBD_NLA_H +#define __DRBD_NLA_H + +extern int drbd_nla_parse_nested(struct nlattr *tb[], int maxtype, struct nlattr *nla, + const struct nla_policy *policy); +extern struct nlattr *drbd_nla_find_nested(int maxtype, struct nlattr *nla, int attrtype); + +#endif /* __DRBD_NLA_H */ diff --git a/drivers/block/drbd/drbd_proc.c b/drivers/block/drbd/drbd_proc.c index 5496104f90b9..56672a61eb94 100644 --- a/drivers/block/drbd/drbd_proc.c +++ b/drivers/block/drbd/drbd_proc.c @@ -167,18 +167,24 @@ static void drbd_syncer_progress(struct drbd_conf *mdev, struct seq_file *seq) * we convert to sectors in the display below. */ unsigned long bm_bits = drbd_bm_bits(mdev); unsigned long bit_pos; + unsigned long long stop_sector = 0; if (mdev->state.conn == C_VERIFY_S || - mdev->state.conn == C_VERIFY_T) + mdev->state.conn == C_VERIFY_T) { bit_pos = bm_bits - mdev->ov_left; - else + if (verify_can_do_stop_sector(mdev)) + stop_sector = mdev->ov_stop_sector; + } else bit_pos = mdev->bm_resync_fo; /* Total sectors may be slightly off for oddly * sized devices. So what. */ seq_printf(seq, - "\t%3d%% sector pos: %llu/%llu\n", + "\t%3d%% sector pos: %llu/%llu", (int)(bit_pos / (bm_bits/100+1)), (unsigned long long)bit_pos * BM_SECT_PER_BIT, (unsigned long long)bm_bits * BM_SECT_PER_BIT); + if (stop_sector != 0 && stop_sector != ULLONG_MAX) + seq_printf(seq, " stop sector: %llu", stop_sector); + seq_printf(seq, "\n"); } } @@ -194,9 +200,11 @@ static void resync_dump_detail(struct seq_file *seq, struct lc_element *e) static int drbd_seq_show(struct seq_file *seq, void *v) { - int i, hole = 0; + int i, prev_i = -1; const char *sn; struct drbd_conf *mdev; + struct net_conf *nc; + char wp; static char write_ordering_chars[] = { [WO_none] = 'n', @@ -227,16 +235,11 @@ static int drbd_seq_show(struct seq_file *seq, void *v) oos .. known out-of-sync kB */ - for (i = 0; i < minor_count; i++) { - mdev = minor_to_mdev(i); - if (!mdev) { - hole = 1; - continue; - } - if (hole) { - hole = 0; + rcu_read_lock(); + idr_for_each_entry(&minors, mdev, i) { + if (prev_i != i - 1) seq_printf(seq, "\n"); - } + prev_i = i; sn = drbd_conn_str(mdev->state.conn); @@ -248,6 +251,8 @@ static int drbd_seq_show(struct seq_file *seq, void *v) /* reset mdev->congestion_reason */ bdi_rw_congested(&mdev->rq_queue->backing_dev_info); + nc = rcu_dereference(mdev->tconn->net_conf); + wp = nc ? nc->wire_protocol - DRBD_PROT_A + 'A' : ' '; seq_printf(seq, "%2d: cs:%s ro:%s/%s ds:%s/%s %c %c%c%c%c%c%c\n" " ns:%u nr:%u dw:%u dr:%u al:%u bm:%u " @@ -257,9 +262,8 @@ static int drbd_seq_show(struct seq_file *seq, void *v) drbd_role_str(mdev->state.peer), drbd_disk_str(mdev->state.disk), drbd_disk_str(mdev->state.pdsk), - (mdev->net_conf == NULL ? ' ' : - (mdev->net_conf->wire_protocol - DRBD_PROT_A+'A')), - is_susp(mdev->state) ? 's' : 'r', + wp, + drbd_suspended(mdev) ? 's' : 'r', mdev->state.aftr_isp ? 'a' : '-', mdev->state.peer_isp ? 'p' : '-', mdev->state.user_isp ? 'u' : '-', @@ -276,8 +280,8 @@ static int drbd_seq_show(struct seq_file *seq, void *v) atomic_read(&mdev->rs_pending_cnt), atomic_read(&mdev->unacked_cnt), atomic_read(&mdev->ap_bio_cnt), - mdev->epochs, - write_ordering_chars[mdev->write_ordering] + mdev->tconn->epochs, + write_ordering_chars[mdev->tconn->write_ordering] ); seq_printf(seq, " oos:%llu\n", Bit2KB((unsigned long long) @@ -302,6 +306,7 @@ static int drbd_seq_show(struct seq_file *seq, void *v) } } } + rcu_read_unlock(); return 0; } diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c index c74ca2df7431..a9eccfc6079b 100644 --- a/drivers/block/drbd/drbd_receiver.c +++ b/drivers/block/drbd/drbd_receiver.c @@ -48,17 +48,25 @@ #include "drbd_vli.h" +struct packet_info { + enum drbd_packet cmd; + unsigned int size; + unsigned int vnr; + void *data; +}; + enum finish_epoch { FE_STILL_LIVE, FE_DESTROYED, FE_RECYCLED, }; -static int drbd_do_handshake(struct drbd_conf *mdev); -static int drbd_do_auth(struct drbd_conf *mdev); +static int drbd_do_features(struct drbd_tconn *tconn); +static int drbd_do_auth(struct drbd_tconn *tconn); +static int drbd_disconnected(struct drbd_conf *mdev); -static enum finish_epoch drbd_may_finish_epoch(struct drbd_conf *, struct drbd_epoch *, enum epoch_event); -static int e_end_block(struct drbd_conf *, struct drbd_work *, int); +static enum finish_epoch drbd_may_finish_epoch(struct drbd_tconn *, struct drbd_epoch *, enum epoch_event); +static int e_end_block(struct drbd_work *, int); #define GFP_TRY (__GFP_HIGHMEM | __GFP_NOWARN) @@ -142,11 +150,12 @@ static void page_chain_add(struct page **head, *head = chain_first; } -static struct page *drbd_pp_first_pages_or_try_alloc(struct drbd_conf *mdev, int number) +static struct page *__drbd_alloc_pages(struct drbd_conf *mdev, + unsigned int number) { struct page *page = NULL; struct page *tmp = NULL; - int i = 0; + unsigned int i = 0; /* Yes, testing drbd_pp_vacant outside the lock is racy. * So what. It saves a spin_lock. */ @@ -175,7 +184,7 @@ static struct page *drbd_pp_first_pages_or_try_alloc(struct drbd_conf *mdev, int return page; /* Not enough pages immediately available this time. - * No need to jump around here, drbd_pp_alloc will retry this + * No need to jump around here, drbd_alloc_pages will retry this * function "soon". */ if (page) { tmp = page_chain_tail(page, NULL); @@ -187,9 +196,10 @@ static struct page *drbd_pp_first_pages_or_try_alloc(struct drbd_conf *mdev, int return NULL; } -static void reclaim_net_ee(struct drbd_conf *mdev, struct list_head *to_be_freed) +static void reclaim_finished_net_peer_reqs(struct drbd_conf *mdev, + struct list_head *to_be_freed) { - struct drbd_epoch_entry *e; + struct drbd_peer_request *peer_req; struct list_head *le, *tle; /* The EEs are always appended to the end of the list. Since @@ -198,8 +208,8 @@ static void reclaim_net_ee(struct drbd_conf *mdev, struct list_head *to_be_freed stop to examine the list... */ list_for_each_safe(le, tle, &mdev->net_ee) { - e = list_entry(le, struct drbd_epoch_entry, w.list); - if (drbd_ee_has_active_page(e)) + peer_req = list_entry(le, struct drbd_peer_request, w.list); + if (drbd_peer_req_has_active_page(peer_req)) break; list_move(le, to_be_freed); } @@ -208,18 +218,18 @@ static void reclaim_net_ee(struct drbd_conf *mdev, struct list_head *to_be_freed static void drbd_kick_lo_and_reclaim_net(struct drbd_conf *mdev) { LIST_HEAD(reclaimed); - struct drbd_epoch_entry *e, *t; + struct drbd_peer_request *peer_req, *t; - spin_lock_irq(&mdev->req_lock); - reclaim_net_ee(mdev, &reclaimed); - spin_unlock_irq(&mdev->req_lock); + spin_lock_irq(&mdev->tconn->req_lock); + reclaim_finished_net_peer_reqs(mdev, &reclaimed); + spin_unlock_irq(&mdev->tconn->req_lock); - list_for_each_entry_safe(e, t, &reclaimed, w.list) - drbd_free_net_ee(mdev, e); + list_for_each_entry_safe(peer_req, t, &reclaimed, w.list) + drbd_free_net_peer_req(mdev, peer_req); } /** - * drbd_pp_alloc() - Returns @number pages, retries forever (or until signalled) + * drbd_alloc_pages() - Returns @number pages, retries forever (or until signalled) * @mdev: DRBD device. * @number: number of pages requested * @retry: whether to retry, if not enough pages are available right now @@ -230,23 +240,31 @@ static void drbd_kick_lo_and_reclaim_net(struct drbd_conf *mdev) * * Returns a page chain linked via page->private. */ -static struct page *drbd_pp_alloc(struct drbd_conf *mdev, unsigned number, bool retry) +struct page *drbd_alloc_pages(struct drbd_conf *mdev, unsigned int number, + bool retry) { struct page *page = NULL; + struct net_conf *nc; DEFINE_WAIT(wait); + int mxb; /* Yes, we may run up to @number over max_buffers. If we * follow it strictly, the admin will get it wrong anyways. */ - if (atomic_read(&mdev->pp_in_use) < mdev->net_conf->max_buffers) - page = drbd_pp_first_pages_or_try_alloc(mdev, number); + rcu_read_lock(); + nc = rcu_dereference(mdev->tconn->net_conf); + mxb = nc ? nc->max_buffers : 1000000; + rcu_read_unlock(); + + if (atomic_read(&mdev->pp_in_use) < mxb) + page = __drbd_alloc_pages(mdev, number); while (page == NULL) { prepare_to_wait(&drbd_pp_wait, &wait, TASK_INTERRUPTIBLE); drbd_kick_lo_and_reclaim_net(mdev); - if (atomic_read(&mdev->pp_in_use) < mdev->net_conf->max_buffers) { - page = drbd_pp_first_pages_or_try_alloc(mdev, number); + if (atomic_read(&mdev->pp_in_use) < mxb) { + page = __drbd_alloc_pages(mdev, number); if (page) break; } @@ -255,7 +273,7 @@ static struct page *drbd_pp_alloc(struct drbd_conf *mdev, unsigned number, bool break; if (signal_pending(current)) { - dev_warn(DEV, "drbd_pp_alloc interrupted!\n"); + dev_warn(DEV, "drbd_alloc_pages interrupted!\n"); break; } @@ -268,11 +286,11 @@ static struct page *drbd_pp_alloc(struct drbd_conf *mdev, unsigned number, bool return page; } -/* Must not be used from irq, as that may deadlock: see drbd_pp_alloc. - * Is also used from inside an other spin_lock_irq(&mdev->req_lock); +/* Must not be used from irq, as that may deadlock: see drbd_alloc_pages. + * Is also used from inside an other spin_lock_irq(&mdev->tconn->req_lock); * Either links the page chain back to the global pool, * or returns all pages to the system. */ -static void drbd_pp_free(struct drbd_conf *mdev, struct page *page, int is_net) +static void drbd_free_pages(struct drbd_conf *mdev, struct page *page, int is_net) { atomic_t *a = is_net ? &mdev->pp_in_use_by_net : &mdev->pp_in_use; int i; @@ -280,7 +298,7 @@ static void drbd_pp_free(struct drbd_conf *mdev, struct page *page, int is_net) if (page == NULL) return; - if (drbd_pp_vacant > (DRBD_MAX_BIO_SIZE/PAGE_SIZE)*minor_count) + if (drbd_pp_vacant > (DRBD_MAX_BIO_SIZE/PAGE_SIZE) * minor_count) i = page_chain_free(page); else { struct page *tmp; @@ -302,127 +320,130 @@ You need to hold the req_lock: _drbd_wait_ee_list_empty() You must not have the req_lock: - drbd_free_ee() - drbd_alloc_ee() - drbd_init_ee() - drbd_release_ee() + drbd_free_peer_req() + drbd_alloc_peer_req() + drbd_free_peer_reqs() drbd_ee_fix_bhs() - drbd_process_done_ee() + drbd_finish_peer_reqs() drbd_clear_done_ee() drbd_wait_ee_list_empty() */ -struct drbd_epoch_entry *drbd_alloc_ee(struct drbd_conf *mdev, - u64 id, - sector_t sector, - unsigned int data_size, - gfp_t gfp_mask) __must_hold(local) +struct drbd_peer_request * +drbd_alloc_peer_req(struct drbd_conf *mdev, u64 id, sector_t sector, + unsigned int data_size, gfp_t gfp_mask) __must_hold(local) { - struct drbd_epoch_entry *e; + struct drbd_peer_request *peer_req; struct page *page = NULL; unsigned nr_pages = (data_size + PAGE_SIZE -1) >> PAGE_SHIFT; if (drbd_insert_fault(mdev, DRBD_FAULT_AL_EE)) return NULL; - e = mempool_alloc(drbd_ee_mempool, gfp_mask & ~__GFP_HIGHMEM); - if (!e) { + peer_req = mempool_alloc(drbd_ee_mempool, gfp_mask & ~__GFP_HIGHMEM); + if (!peer_req) { if (!(gfp_mask & __GFP_NOWARN)) - dev_err(DEV, "alloc_ee: Allocation of an EE failed\n"); + dev_err(DEV, "%s: allocation failed\n", __func__); return NULL; } if (data_size) { - page = drbd_pp_alloc(mdev, nr_pages, (gfp_mask & __GFP_WAIT)); + page = drbd_alloc_pages(mdev, nr_pages, (gfp_mask & __GFP_WAIT)); if (!page) goto fail; } - INIT_HLIST_NODE(&e->collision); - e->epoch = NULL; - e->mdev = mdev; - e->pages = page; - atomic_set(&e->pending_bios, 0); - e->size = data_size; - e->flags = 0; - e->sector = sector; - e->block_id = id; + drbd_clear_interval(&peer_req->i); + peer_req->i.size = data_size; + peer_req->i.sector = sector; + peer_req->i.local = false; + peer_req->i.waiting = false; + + peer_req->epoch = NULL; + peer_req->w.mdev = mdev; + peer_req->pages = page; + atomic_set(&peer_req->pending_bios, 0); + peer_req->flags = 0; + /* + * The block_id is opaque to the receiver. It is not endianness + * converted, and sent back to the sender unchanged. + */ + peer_req->block_id = id; - return e; + return peer_req; fail: - mempool_free(e, drbd_ee_mempool); + mempool_free(peer_req, drbd_ee_mempool); return NULL; } -void drbd_free_some_ee(struct drbd_conf *mdev, struct drbd_epoch_entry *e, int is_net) +void __drbd_free_peer_req(struct drbd_conf *mdev, struct drbd_peer_request *peer_req, + int is_net) { - if (e->flags & EE_HAS_DIGEST) - kfree(e->digest); - drbd_pp_free(mdev, e->pages, is_net); - D_ASSERT(atomic_read(&e->pending_bios) == 0); - D_ASSERT(hlist_unhashed(&e->collision)); - mempool_free(e, drbd_ee_mempool); + if (peer_req->flags & EE_HAS_DIGEST) + kfree(peer_req->digest); + drbd_free_pages(mdev, peer_req->pages, is_net); + D_ASSERT(atomic_read(&peer_req->pending_bios) == 0); + D_ASSERT(drbd_interval_empty(&peer_req->i)); + mempool_free(peer_req, drbd_ee_mempool); } -int drbd_release_ee(struct drbd_conf *mdev, struct list_head *list) +int drbd_free_peer_reqs(struct drbd_conf *mdev, struct list_head *list) { LIST_HEAD(work_list); - struct drbd_epoch_entry *e, *t; + struct drbd_peer_request *peer_req, *t; int count = 0; int is_net = list == &mdev->net_ee; - spin_lock_irq(&mdev->req_lock); + spin_lock_irq(&mdev->tconn->req_lock); list_splice_init(list, &work_list); - spin_unlock_irq(&mdev->req_lock); + spin_unlock_irq(&mdev->tconn->req_lock); - list_for_each_entry_safe(e, t, &work_list, w.list) { - drbd_free_some_ee(mdev, e, is_net); + list_for_each_entry_safe(peer_req, t, &work_list, w.list) { + __drbd_free_peer_req(mdev, peer_req, is_net); count++; } return count; } - /* - * This function is called from _asender only_ - * but see also comments in _req_mod(,barrier_acked) - * and receive_Barrier. - * - * Move entries from net_ee to done_ee, if ready. - * Grab done_ee, call all callbacks, free the entries. - * The callbacks typically send out ACKs. + * See also comments in _req_mod(,BARRIER_ACKED) and receive_Barrier. */ -static int drbd_process_done_ee(struct drbd_conf *mdev) +static int drbd_finish_peer_reqs(struct drbd_conf *mdev) { LIST_HEAD(work_list); LIST_HEAD(reclaimed); - struct drbd_epoch_entry *e, *t; - int ok = (mdev->state.conn >= C_WF_REPORT_PARAMS); + struct drbd_peer_request *peer_req, *t; + int err = 0; - spin_lock_irq(&mdev->req_lock); - reclaim_net_ee(mdev, &reclaimed); + spin_lock_irq(&mdev->tconn->req_lock); + reclaim_finished_net_peer_reqs(mdev, &reclaimed); list_splice_init(&mdev->done_ee, &work_list); - spin_unlock_irq(&mdev->req_lock); + spin_unlock_irq(&mdev->tconn->req_lock); - list_for_each_entry_safe(e, t, &reclaimed, w.list) - drbd_free_net_ee(mdev, e); + list_for_each_entry_safe(peer_req, t, &reclaimed, w.list) + drbd_free_net_peer_req(mdev, peer_req); /* possible callbacks here: - * e_end_block, and e_end_resync_block, e_send_discard_ack. + * e_end_block, and e_end_resync_block, e_send_superseded. * all ignore the last argument. */ - list_for_each_entry_safe(e, t, &work_list, w.list) { + list_for_each_entry_safe(peer_req, t, &work_list, w.list) { + int err2; + /* list_del not necessary, next/prev members not touched */ - ok = e->w.cb(mdev, &e->w, !ok) && ok; - drbd_free_ee(mdev, e); + err2 = peer_req->w.cb(&peer_req->w, !!err); + if (!err) + err = err2; + drbd_free_peer_req(mdev, peer_req); } wake_up(&mdev->ee_wait); - return ok; + return err; } -void _drbd_wait_ee_list_empty(struct drbd_conf *mdev, struct list_head *head) +static void _drbd_wait_ee_list_empty(struct drbd_conf *mdev, + struct list_head *head) { DEFINE_WAIT(wait); @@ -430,55 +451,22 @@ void _drbd_wait_ee_list_empty(struct drbd_conf *mdev, struct list_head *head) * and calling prepare_to_wait in the fast path */ while (!list_empty(head)) { prepare_to_wait(&mdev->ee_wait, &wait, TASK_UNINTERRUPTIBLE); - spin_unlock_irq(&mdev->req_lock); + spin_unlock_irq(&mdev->tconn->req_lock); io_schedule(); finish_wait(&mdev->ee_wait, &wait); - spin_lock_irq(&mdev->req_lock); + spin_lock_irq(&mdev->tconn->req_lock); } } -void drbd_wait_ee_list_empty(struct drbd_conf *mdev, struct list_head *head) +static void drbd_wait_ee_list_empty(struct drbd_conf *mdev, + struct list_head *head) { - spin_lock_irq(&mdev->req_lock); + spin_lock_irq(&mdev->tconn->req_lock); _drbd_wait_ee_list_empty(mdev, head); - spin_unlock_irq(&mdev->req_lock); -} - -/* see also kernel_accept; which is only present since 2.6.18. - * also we want to log which part of it failed, exactly */ -static int drbd_accept(struct drbd_conf *mdev, const char **what, - struct socket *sock, struct socket **newsock) -{ - struct sock *sk = sock->sk; - int err = 0; - - *what = "listen"; - err = sock->ops->listen(sock, 5); - if (err < 0) - goto out; - - *what = "sock_create_lite"; - err = sock_create_lite(sk->sk_family, sk->sk_type, sk->sk_protocol, - newsock); - if (err < 0) - goto out; - - *what = "accept"; - err = sock->ops->accept(sock, *newsock, 0); - if (err < 0) { - sock_release(*newsock); - *newsock = NULL; - goto out; - } - (*newsock)->ops = sock->ops; - __module_get((*newsock)->ops->owner); - -out: - return err; + spin_unlock_irq(&mdev->tconn->req_lock); } -static int drbd_recv_short(struct drbd_conf *mdev, struct socket *sock, - void *buf, size_t size, int flags) +static int drbd_recv_short(struct socket *sock, void *buf, size_t size, int flags) { mm_segment_t oldfs; struct kvec iov = { @@ -500,59 +488,62 @@ static int drbd_recv_short(struct drbd_conf *mdev, struct socket *sock, return rv; } -static int drbd_recv(struct drbd_conf *mdev, void *buf, size_t size) +static int drbd_recv(struct drbd_tconn *tconn, void *buf, size_t size) { - mm_segment_t oldfs; - struct kvec iov = { - .iov_base = buf, - .iov_len = size, - }; - struct msghdr msg = { - .msg_iovlen = 1, - .msg_iov = (struct iovec *)&iov, - .msg_flags = MSG_WAITALL | MSG_NOSIGNAL - }; int rv; - oldfs = get_fs(); - set_fs(KERNEL_DS); + rv = drbd_recv_short(tconn->data.socket, buf, size, 0); - for (;;) { - rv = sock_recvmsg(mdev->data.socket, &msg, size, msg.msg_flags); - if (rv == size) - break; + if (rv < 0) { + if (rv == -ECONNRESET) + conn_info(tconn, "sock was reset by peer\n"); + else if (rv != -ERESTARTSYS) + conn_err(tconn, "sock_recvmsg returned %d\n", rv); + } else if (rv == 0) { + if (test_bit(DISCONNECT_SENT, &tconn->flags)) { + long t; + rcu_read_lock(); + t = rcu_dereference(tconn->net_conf)->ping_timeo * HZ/10; + rcu_read_unlock(); - /* Note: - * ECONNRESET other side closed the connection - * ERESTARTSYS (on sock) we got a signal - */ + t = wait_event_timeout(tconn->ping_wait, tconn->cstate < C_WF_REPORT_PARAMS, t); - if (rv < 0) { - if (rv == -ECONNRESET) - dev_info(DEV, "sock was reset by peer\n"); - else if (rv != -ERESTARTSYS) - dev_err(DEV, "sock_recvmsg returned %d\n", rv); - break; - } else if (rv == 0) { - dev_info(DEV, "sock was shut down by peer\n"); - break; - } else { - /* signal came in, or peer/link went down, - * after we read a partial message - */ - /* D_ASSERT(signal_pending(current)); */ - break; + if (t) + goto out; } - }; - - set_fs(oldfs); + conn_info(tconn, "sock was shut down by peer\n"); + } if (rv != size) - drbd_force_state(mdev, NS(conn, C_BROKEN_PIPE)); + conn_request_state(tconn, NS(conn, C_BROKEN_PIPE), CS_HARD); +out: return rv; } +static int drbd_recv_all(struct drbd_tconn *tconn, void *buf, size_t size) +{ + int err; + + err = drbd_recv(tconn, buf, size); + if (err != size) { + if (err >= 0) + err = -EIO; + } else + err = 0; + return err; +} + +static int drbd_recv_all_warn(struct drbd_tconn *tconn, void *buf, size_t size) +{ + int err; + + err = drbd_recv_all(tconn, buf, size); + if (err && !signal_pending(current)) + conn_warn(tconn, "short read (expected size %d)\n", (int)size); + return err; +} + /* quoting tcp(7): * On individual connections, the socket buffer size must be set prior to the * listen(2) or connect(2) calls in order to have it take effect. @@ -572,29 +563,50 @@ static void drbd_setbufsize(struct socket *sock, unsigned int snd, } } -static struct socket *drbd_try_connect(struct drbd_conf *mdev) +static struct socket *drbd_try_connect(struct drbd_tconn *tconn) { const char *what; struct socket *sock; struct sockaddr_in6 src_in6; - int err; + struct sockaddr_in6 peer_in6; + struct net_conf *nc; + int err, peer_addr_len, my_addr_len; + int sndbuf_size, rcvbuf_size, connect_int; int disconnect_on_error = 1; - if (!get_net_conf(mdev)) + rcu_read_lock(); + nc = rcu_dereference(tconn->net_conf); + if (!nc) { + rcu_read_unlock(); return NULL; + } + sndbuf_size = nc->sndbuf_size; + rcvbuf_size = nc->rcvbuf_size; + connect_int = nc->connect_int; + rcu_read_unlock(); + + my_addr_len = min_t(int, tconn->my_addr_len, sizeof(src_in6)); + memcpy(&src_in6, &tconn->my_addr, my_addr_len); + + if (((struct sockaddr *)&tconn->my_addr)->sa_family == AF_INET6) + src_in6.sin6_port = 0; + else + ((struct sockaddr_in *)&src_in6)->sin_port = 0; /* AF_INET & AF_SCI */ + + peer_addr_len = min_t(int, tconn->peer_addr_len, sizeof(src_in6)); + memcpy(&peer_in6, &tconn->peer_addr, peer_addr_len); what = "sock_create_kern"; - err = sock_create_kern(((struct sockaddr *)mdev->net_conf->my_addr)->sa_family, - SOCK_STREAM, IPPROTO_TCP, &sock); + err = sock_create_kern(((struct sockaddr *)&src_in6)->sa_family, + SOCK_STREAM, IPPROTO_TCP, &sock); if (err < 0) { sock = NULL; goto out; } sock->sk->sk_rcvtimeo = - sock->sk->sk_sndtimeo = mdev->net_conf->try_connect_int*HZ; - drbd_setbufsize(sock, mdev->net_conf->sndbuf_size, - mdev->net_conf->rcvbuf_size); + sock->sk->sk_sndtimeo = connect_int * HZ; + drbd_setbufsize(sock, sndbuf_size, rcvbuf_size); /* explicitly bind to the configured IP as source IP * for the outgoing connections. @@ -603,17 +615,8 @@ static struct socket *drbd_try_connect(struct drbd_conf *mdev) * Make sure to use 0 as port number, so linux selects * a free one dynamically. */ - memcpy(&src_in6, mdev->net_conf->my_addr, - min_t(int, mdev->net_conf->my_addr_len, sizeof(src_in6))); - if (((struct sockaddr *)mdev->net_conf->my_addr)->sa_family == AF_INET6) - src_in6.sin6_port = 0; - else - ((struct sockaddr_in *)&src_in6)->sin_port = 0; /* AF_INET & AF_SCI */ - what = "bind before connect"; - err = sock->ops->bind(sock, - (struct sockaddr *) &src_in6, - mdev->net_conf->my_addr_len); + err = sock->ops->bind(sock, (struct sockaddr *) &src_in6, my_addr_len); if (err < 0) goto out; @@ -621,9 +624,7 @@ static struct socket *drbd_try_connect(struct drbd_conf *mdev) * stay C_WF_CONNECTION, don't go Disconnecting! */ disconnect_on_error = 0; what = "connect"; - err = sock->ops->connect(sock, - (struct sockaddr *)mdev->net_conf->peer_addr, - mdev->net_conf->peer_addr_len, 0); + err = sock->ops->connect(sock, (struct sockaddr *) &peer_in6, peer_addr_len, 0); out: if (err < 0) { @@ -641,91 +642,174 @@ out: disconnect_on_error = 0; break; default: - dev_err(DEV, "%s failed, err = %d\n", what, err); + conn_err(tconn, "%s failed, err = %d\n", what, err); } if (disconnect_on_error) - drbd_force_state(mdev, NS(conn, C_DISCONNECTING)); + conn_request_state(tconn, NS(conn, C_DISCONNECTING), CS_HARD); } - put_net_conf(mdev); + return sock; } -static struct socket *drbd_wait_for_connect(struct drbd_conf *mdev) +struct accept_wait_data { + struct drbd_tconn *tconn; + struct socket *s_listen; + struct completion door_bell; + void (*original_sk_state_change)(struct sock *sk); + +}; + +static void drbd_incoming_connection(struct sock *sk) { - int timeo, err; - struct socket *s_estab = NULL, *s_listen; + struct accept_wait_data *ad = sk->sk_user_data; + void (*state_change)(struct sock *sk); + + state_change = ad->original_sk_state_change; + if (sk->sk_state == TCP_ESTABLISHED) + complete(&ad->door_bell); + state_change(sk); +} + +static int prepare_listen_socket(struct drbd_tconn *tconn, struct accept_wait_data *ad) +{ + int err, sndbuf_size, rcvbuf_size, my_addr_len; + struct sockaddr_in6 my_addr; + struct socket *s_listen; + struct net_conf *nc; const char *what; - if (!get_net_conf(mdev)) - return NULL; + rcu_read_lock(); + nc = rcu_dereference(tconn->net_conf); + if (!nc) { + rcu_read_unlock(); + return -EIO; + } + sndbuf_size = nc->sndbuf_size; + rcvbuf_size = nc->rcvbuf_size; + rcu_read_unlock(); + + my_addr_len = min_t(int, tconn->my_addr_len, sizeof(struct sockaddr_in6)); + memcpy(&my_addr, &tconn->my_addr, my_addr_len); what = "sock_create_kern"; - err = sock_create_kern(((struct sockaddr *)mdev->net_conf->my_addr)->sa_family, - SOCK_STREAM, IPPROTO_TCP, &s_listen); + err = sock_create_kern(((struct sockaddr *)&my_addr)->sa_family, + SOCK_STREAM, IPPROTO_TCP, &s_listen); if (err) { s_listen = NULL; goto out; } - timeo = mdev->net_conf->try_connect_int * HZ; - timeo += (random32() & 1) ? timeo / 7 : -timeo / 7; /* 28.5% random jitter */ - - s_listen->sk->sk_reuse = SK_CAN_REUSE; /* SO_REUSEADDR */ - s_listen->sk->sk_rcvtimeo = timeo; - s_listen->sk->sk_sndtimeo = timeo; - drbd_setbufsize(s_listen, mdev->net_conf->sndbuf_size, - mdev->net_conf->rcvbuf_size); + s_listen->sk->sk_reuse = SK_CAN_REUSE; /* SO_REUSEADDR */ + drbd_setbufsize(s_listen, sndbuf_size, rcvbuf_size); what = "bind before listen"; - err = s_listen->ops->bind(s_listen, - (struct sockaddr *) mdev->net_conf->my_addr, - mdev->net_conf->my_addr_len); + err = s_listen->ops->bind(s_listen, (struct sockaddr *)&my_addr, my_addr_len); if (err < 0) goto out; - err = drbd_accept(mdev, &what, s_listen, &s_estab); + ad->s_listen = s_listen; + write_lock_bh(&s_listen->sk->sk_callback_lock); + ad->original_sk_state_change = s_listen->sk->sk_state_change; + s_listen->sk->sk_state_change = drbd_incoming_connection; + s_listen->sk->sk_user_data = ad; + write_unlock_bh(&s_listen->sk->sk_callback_lock); + + what = "listen"; + err = s_listen->ops->listen(s_listen, 5); + if (err < 0) + goto out; + return 0; out: if (s_listen) sock_release(s_listen); if (err < 0) { if (err != -EAGAIN && err != -EINTR && err != -ERESTARTSYS) { - dev_err(DEV, "%s failed, err = %d\n", what, err); - drbd_force_state(mdev, NS(conn, C_DISCONNECTING)); + conn_err(tconn, "%s failed, err = %d\n", what, err); + conn_request_state(tconn, NS(conn, C_DISCONNECTING), CS_HARD); } } - put_net_conf(mdev); - return s_estab; + return -EIO; } -static int drbd_send_fp(struct drbd_conf *mdev, - struct socket *sock, enum drbd_packets cmd) +static void unregister_state_change(struct sock *sk, struct accept_wait_data *ad) { - struct p_header80 *h = &mdev->data.sbuf.header.h80; - - return _drbd_send_cmd(mdev, sock, cmd, h, sizeof(*h), 0); + write_lock_bh(&sk->sk_callback_lock); + sk->sk_state_change = ad->original_sk_state_change; + sk->sk_user_data = NULL; + write_unlock_bh(&sk->sk_callback_lock); } -static enum drbd_packets drbd_recv_fp(struct drbd_conf *mdev, struct socket *sock) +static struct socket *drbd_wait_for_connect(struct drbd_tconn *tconn, struct accept_wait_data *ad) { - struct p_header80 *h = &mdev->data.rbuf.header.h80; - int rr; + int timeo, connect_int, err = 0; + struct socket *s_estab = NULL; + struct net_conf *nc; + + rcu_read_lock(); + nc = rcu_dereference(tconn->net_conf); + if (!nc) { + rcu_read_unlock(); + return NULL; + } + connect_int = nc->connect_int; + rcu_read_unlock(); + + timeo = connect_int * HZ; + timeo += (random32() & 1) ? timeo / 7 : -timeo / 7; /* 28.5% random jitter */ + + err = wait_for_completion_interruptible_timeout(&ad->door_bell, timeo); + if (err <= 0) + return NULL; + + err = kernel_accept(ad->s_listen, &s_estab, 0); + if (err < 0) { + if (err != -EAGAIN && err != -EINTR && err != -ERESTARTSYS) { + conn_err(tconn, "accept failed, err = %d\n", err); + conn_request_state(tconn, NS(conn, C_DISCONNECTING), CS_HARD); + } + } + + if (s_estab) + unregister_state_change(s_estab->sk, ad); + + return s_estab; +} - rr = drbd_recv_short(mdev, sock, h, sizeof(*h), 0); +static int decode_header(struct drbd_tconn *, void *, struct packet_info *); - if (rr == sizeof(*h) && h->magic == BE_DRBD_MAGIC) - return be16_to_cpu(h->command); +static int send_first_packet(struct drbd_tconn *tconn, struct drbd_socket *sock, + enum drbd_packet cmd) +{ + if (!conn_prepare_command(tconn, sock)) + return -EIO; + return conn_send_command(tconn, sock, cmd, 0, NULL, 0); +} - return 0xffff; +static int receive_first_packet(struct drbd_tconn *tconn, struct socket *sock) +{ + unsigned int header_size = drbd_header_size(tconn); + struct packet_info pi; + int err; + + err = drbd_recv_short(sock, tconn->data.rbuf, header_size, 0); + if (err != header_size) { + if (err >= 0) + err = -EIO; + return err; + } + err = decode_header(tconn, tconn->data.rbuf, &pi); + if (err) + return err; + return pi.cmd; } /** * drbd_socket_okay() - Free the socket if its connection is not okay - * @mdev: DRBD device. * @sock: pointer to the pointer to the socket. */ -static int drbd_socket_okay(struct drbd_conf *mdev, struct socket **sock) +static int drbd_socket_okay(struct socket **sock) { int rr; char tb[4]; @@ -733,7 +817,7 @@ static int drbd_socket_okay(struct drbd_conf *mdev, struct socket **sock) if (!*sock) return false; - rr = drbd_recv_short(mdev, *sock, tb, 4, MSG_DONTWAIT | MSG_PEEK); + rr = drbd_recv_short(*sock, tb, 4, MSG_DONTWAIT | MSG_PEEK); if (rr > 0 || rr == -EAGAIN) { return true; @@ -743,6 +827,31 @@ static int drbd_socket_okay(struct drbd_conf *mdev, struct socket **sock) return false; } } +/* Gets called if a connection is established, or if a new minor gets created + in a connection */ +int drbd_connected(struct drbd_conf *mdev) +{ + int err; + + atomic_set(&mdev->packet_seq, 0); + mdev->peer_seq = 0; + + mdev->state_mutex = mdev->tconn->agreed_pro_version < 100 ? + &mdev->tconn->cstate_mutex : + &mdev->own_state_mutex; + + err = drbd_send_sync_param(mdev); + if (!err) + err = drbd_send_sizes(mdev, 0, 0); + if (!err) + err = drbd_send_uuids(mdev); + if (!err) + err = drbd_send_current_state(mdev); + clear_bit(USE_DEGR_WFC_T, &mdev->flags); + clear_bit(RESIZE_PENDING, &mdev->flags); + mod_timer(&mdev->request_timer, jiffies + HZ); /* just start it here. */ + return err; +} /* * return values: @@ -752,232 +861,315 @@ static int drbd_socket_okay(struct drbd_conf *mdev, struct socket **sock) * no point in trying again, please go standalone. * -2 We do not have a network config... */ -static int drbd_connect(struct drbd_conf *mdev) +static int conn_connect(struct drbd_tconn *tconn) { - struct socket *s, *sock, *msock; - int try, h, ok; + struct drbd_socket sock, msock; + struct drbd_conf *mdev; + struct net_conf *nc; + int vnr, timeout, h, ok; + bool discard_my_data; enum drbd_state_rv rv; + struct accept_wait_data ad = { + .tconn = tconn, + .door_bell = COMPLETION_INITIALIZER_ONSTACK(ad.door_bell), + }; - D_ASSERT(!mdev->data.socket); - - if (drbd_request_state(mdev, NS(conn, C_WF_CONNECTION)) < SS_SUCCESS) + clear_bit(DISCONNECT_SENT, &tconn->flags); + if (conn_request_state(tconn, NS(conn, C_WF_CONNECTION), CS_VERBOSE) < SS_SUCCESS) return -2; - clear_bit(DISCARD_CONCURRENT, &mdev->flags); + mutex_init(&sock.mutex); + sock.sbuf = tconn->data.sbuf; + sock.rbuf = tconn->data.rbuf; + sock.socket = NULL; + mutex_init(&msock.mutex); + msock.sbuf = tconn->meta.sbuf; + msock.rbuf = tconn->meta.rbuf; + msock.socket = NULL; + + /* Assume that the peer only understands protocol 80 until we know better. */ + tconn->agreed_pro_version = 80; - sock = NULL; - msock = NULL; + if (prepare_listen_socket(tconn, &ad)) + return 0; do { - for (try = 0;;) { - /* 3 tries, this should take less than a second! */ - s = drbd_try_connect(mdev); - if (s || ++try >= 3) - break; - /* give the other side time to call bind() & listen() */ - schedule_timeout_interruptible(HZ / 10); - } + struct socket *s; + s = drbd_try_connect(tconn); if (s) { - if (!sock) { - drbd_send_fp(mdev, s, P_HAND_SHAKE_S); - sock = s; - s = NULL; - } else if (!msock) { - drbd_send_fp(mdev, s, P_HAND_SHAKE_M); - msock = s; - s = NULL; + if (!sock.socket) { + sock.socket = s; + send_first_packet(tconn, &sock, P_INITIAL_DATA); + } else if (!msock.socket) { + clear_bit(RESOLVE_CONFLICTS, &tconn->flags); + msock.socket = s; + send_first_packet(tconn, &msock, P_INITIAL_META); } else { - dev_err(DEV, "Logic error in drbd_connect()\n"); + conn_err(tconn, "Logic error in conn_connect()\n"); goto out_release_sockets; } } - if (sock && msock) { - schedule_timeout_interruptible(mdev->net_conf->ping_timeo*HZ/10); - ok = drbd_socket_okay(mdev, &sock); - ok = drbd_socket_okay(mdev, &msock) && ok; + if (sock.socket && msock.socket) { + rcu_read_lock(); + nc = rcu_dereference(tconn->net_conf); + timeout = nc->ping_timeo * HZ / 10; + rcu_read_unlock(); + schedule_timeout_interruptible(timeout); + ok = drbd_socket_okay(&sock.socket); + ok = drbd_socket_okay(&msock.socket) && ok; if (ok) break; } retry: - s = drbd_wait_for_connect(mdev); + s = drbd_wait_for_connect(tconn, &ad); if (s) { - try = drbd_recv_fp(mdev, s); - drbd_socket_okay(mdev, &sock); - drbd_socket_okay(mdev, &msock); - switch (try) { - case P_HAND_SHAKE_S: - if (sock) { - dev_warn(DEV, "initial packet S crossed\n"); - sock_release(sock); + int fp = receive_first_packet(tconn, s); + drbd_socket_okay(&sock.socket); + drbd_socket_okay(&msock.socket); + switch (fp) { + case P_INITIAL_DATA: + if (sock.socket) { + conn_warn(tconn, "initial packet S crossed\n"); + sock_release(sock.socket); + sock.socket = s; + goto randomize; } - sock = s; + sock.socket = s; break; - case P_HAND_SHAKE_M: - if (msock) { - dev_warn(DEV, "initial packet M crossed\n"); - sock_release(msock); + case P_INITIAL_META: + set_bit(RESOLVE_CONFLICTS, &tconn->flags); + if (msock.socket) { + conn_warn(tconn, "initial packet M crossed\n"); + sock_release(msock.socket); + msock.socket = s; + goto randomize; } - msock = s; - set_bit(DISCARD_CONCURRENT, &mdev->flags); + msock.socket = s; break; default: - dev_warn(DEV, "Error receiving initial packet\n"); + conn_warn(tconn, "Error receiving initial packet\n"); sock_release(s); +randomize: if (random32() & 1) goto retry; } } - if (mdev->state.conn <= C_DISCONNECTING) + if (tconn->cstate <= C_DISCONNECTING) goto out_release_sockets; if (signal_pending(current)) { flush_signals(current); smp_rmb(); - if (get_t_state(&mdev->receiver) == Exiting) + if (get_t_state(&tconn->receiver) == EXITING) goto out_release_sockets; } - if (sock && msock) { - ok = drbd_socket_okay(mdev, &sock); - ok = drbd_socket_okay(mdev, &msock) && ok; - if (ok) - break; - } - } while (1); + ok = drbd_socket_okay(&sock.socket); + ok = drbd_socket_okay(&msock.socket) && ok; + } while (!ok); + + if (ad.s_listen) + sock_release(ad.s_listen); - msock->sk->sk_reuse = SK_CAN_REUSE; /* SO_REUSEADDR */ - sock->sk->sk_reuse = SK_CAN_REUSE; /* SO_REUSEADDR */ + sock.socket->sk->sk_reuse = SK_CAN_REUSE; /* SO_REUSEADDR */ + msock.socket->sk->sk_reuse = SK_CAN_REUSE; /* SO_REUSEADDR */ - sock->sk->sk_allocation = GFP_NOIO; - msock->sk->sk_allocation = GFP_NOIO; + sock.socket->sk->sk_allocation = GFP_NOIO; + msock.socket->sk->sk_allocation = GFP_NOIO; - sock->sk->sk_priority = TC_PRIO_INTERACTIVE_BULK; - msock->sk->sk_priority = TC_PRIO_INTERACTIVE; + sock.socket->sk->sk_priority = TC_PRIO_INTERACTIVE_BULK; + msock.socket->sk->sk_priority = TC_PRIO_INTERACTIVE; /* NOT YET ... - * sock->sk->sk_sndtimeo = mdev->net_conf->timeout*HZ/10; - * sock->sk->sk_rcvtimeo = MAX_SCHEDULE_TIMEOUT; - * first set it to the P_HAND_SHAKE timeout, + * sock.socket->sk->sk_sndtimeo = tconn->net_conf->timeout*HZ/10; + * sock.socket->sk->sk_rcvtimeo = MAX_SCHEDULE_TIMEOUT; + * first set it to the P_CONNECTION_FEATURES timeout, * which we set to 4x the configured ping_timeout. */ - sock->sk->sk_sndtimeo = - sock->sk->sk_rcvtimeo = mdev->net_conf->ping_timeo*4*HZ/10; + rcu_read_lock(); + nc = rcu_dereference(tconn->net_conf); - msock->sk->sk_sndtimeo = mdev->net_conf->timeout*HZ/10; - msock->sk->sk_rcvtimeo = mdev->net_conf->ping_int*HZ; + sock.socket->sk->sk_sndtimeo = + sock.socket->sk->sk_rcvtimeo = nc->ping_timeo*4*HZ/10; + + msock.socket->sk->sk_rcvtimeo = nc->ping_int*HZ; + timeout = nc->timeout * HZ / 10; + discard_my_data = nc->discard_my_data; + rcu_read_unlock(); + + msock.socket->sk->sk_sndtimeo = timeout; /* we don't want delays. * we use TCP_CORK where appropriate, though */ - drbd_tcp_nodelay(sock); - drbd_tcp_nodelay(msock); - - mdev->data.socket = sock; - mdev->meta.socket = msock; - mdev->last_received = jiffies; + drbd_tcp_nodelay(sock.socket); + drbd_tcp_nodelay(msock.socket); - D_ASSERT(mdev->asender.task == NULL); + tconn->data.socket = sock.socket; + tconn->meta.socket = msock.socket; + tconn->last_received = jiffies; - h = drbd_do_handshake(mdev); + h = drbd_do_features(tconn); if (h <= 0) return h; - if (mdev->cram_hmac_tfm) { + if (tconn->cram_hmac_tfm) { /* drbd_request_state(mdev, NS(conn, WFAuth)); */ - switch (drbd_do_auth(mdev)) { + switch (drbd_do_auth(tconn)) { case -1: - dev_err(DEV, "Authentication of peer failed\n"); + conn_err(tconn, "Authentication of peer failed\n"); return -1; case 0: - dev_err(DEV, "Authentication of peer failed, trying again.\n"); + conn_err(tconn, "Authentication of peer failed, trying again.\n"); return 0; } } - sock->sk->sk_sndtimeo = mdev->net_conf->timeout*HZ/10; - sock->sk->sk_rcvtimeo = MAX_SCHEDULE_TIMEOUT; + tconn->data.socket->sk->sk_sndtimeo = timeout; + tconn->data.socket->sk->sk_rcvtimeo = MAX_SCHEDULE_TIMEOUT; - atomic_set(&mdev->packet_seq, 0); - mdev->peer_seq = 0; - - if (drbd_send_protocol(mdev) == -1) + if (drbd_send_protocol(tconn) == -EOPNOTSUPP) return -1; - set_bit(STATE_SENT, &mdev->flags); - drbd_send_sync_param(mdev, &mdev->sync_conf); - drbd_send_sizes(mdev, 0, 0); - drbd_send_uuids(mdev); - drbd_send_current_state(mdev); - clear_bit(USE_DEGR_WFC_T, &mdev->flags); - clear_bit(RESIZE_PENDING, &mdev->flags); - spin_lock_irq(&mdev->req_lock); - rv = _drbd_set_state(_NS(mdev, conn, C_WF_REPORT_PARAMS), CS_VERBOSE, NULL); - if (mdev->state.conn != C_WF_REPORT_PARAMS) - clear_bit(STATE_SENT, &mdev->flags); - spin_unlock_irq(&mdev->req_lock); + set_bit(STATE_SENT, &tconn->flags); + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) { + kref_get(&mdev->kref); + /* Prevent a race between resync-handshake and + * being promoted to Primary. + * + * Grab and release the state mutex, so we know that any current + * drbd_set_role() is finished, and any incoming drbd_set_role + * will see the STATE_SENT flag, and wait for it to be cleared. + */ + mutex_lock(mdev->state_mutex); + mutex_unlock(mdev->state_mutex); + + rcu_read_unlock(); + + if (discard_my_data) + set_bit(DISCARD_MY_DATA, &mdev->flags); + else + clear_bit(DISCARD_MY_DATA, &mdev->flags); + + drbd_connected(mdev); + kref_put(&mdev->kref, &drbd_minor_destroy); + rcu_read_lock(); + } + rcu_read_unlock(); - if (rv < SS_SUCCESS) + rv = conn_request_state(tconn, NS(conn, C_WF_REPORT_PARAMS), CS_VERBOSE); + if (rv < SS_SUCCESS || tconn->cstate != C_WF_REPORT_PARAMS) { + clear_bit(STATE_SENT, &tconn->flags); return 0; + } - drbd_thread_start(&mdev->asender); - mod_timer(&mdev->request_timer, jiffies + HZ); /* just start it here. */ + drbd_thread_start(&tconn->asender); - return 1; + mutex_lock(&tconn->conf_update); + /* The discard_my_data flag is a single-shot modifier to the next + * connection attempt, the handshake of which is now well underway. + * No need for rcu style copying of the whole struct + * just to clear a single value. */ + tconn->net_conf->discard_my_data = 0; + mutex_unlock(&tconn->conf_update); + + return h; out_release_sockets: - if (sock) - sock_release(sock); - if (msock) - sock_release(msock); + if (ad.s_listen) + sock_release(ad.s_listen); + if (sock.socket) + sock_release(sock.socket); + if (msock.socket) + sock_release(msock.socket); return -1; } -static int drbd_recv_header(struct drbd_conf *mdev, enum drbd_packets *cmd, unsigned int *packet_size) +static int decode_header(struct drbd_tconn *tconn, void *header, struct packet_info *pi) { - union p_header *h = &mdev->data.rbuf.header; - int r; - - r = drbd_recv(mdev, h, sizeof(*h)); - if (unlikely(r != sizeof(*h))) { - if (!signal_pending(current)) - dev_warn(DEV, "short read expecting header on sock: r=%d\n", r); - return false; - } - - if (likely(h->h80.magic == BE_DRBD_MAGIC)) { - *cmd = be16_to_cpu(h->h80.command); - *packet_size = be16_to_cpu(h->h80.length); - } else if (h->h95.magic == BE_DRBD_MAGIC_BIG) { - *cmd = be16_to_cpu(h->h95.command); - *packet_size = be32_to_cpu(h->h95.length); + unsigned int header_size = drbd_header_size(tconn); + + if (header_size == sizeof(struct p_header100) && + *(__be32 *)header == cpu_to_be32(DRBD_MAGIC_100)) { + struct p_header100 *h = header; + if (h->pad != 0) { + conn_err(tconn, "Header padding is not zero\n"); + return -EINVAL; + } + pi->vnr = be16_to_cpu(h->volume); + pi->cmd = be16_to_cpu(h->command); + pi->size = be32_to_cpu(h->length); + } else if (header_size == sizeof(struct p_header95) && + *(__be16 *)header == cpu_to_be16(DRBD_MAGIC_BIG)) { + struct p_header95 *h = header; + pi->cmd = be16_to_cpu(h->command); + pi->size = be32_to_cpu(h->length); + pi->vnr = 0; + } else if (header_size == sizeof(struct p_header80) && + *(__be32 *)header == cpu_to_be32(DRBD_MAGIC)) { + struct p_header80 *h = header; + pi->cmd = be16_to_cpu(h->command); + pi->size = be16_to_cpu(h->length); + pi->vnr = 0; } else { - dev_err(DEV, "magic?? on data m: 0x%08x c: %d l: %d\n", - be32_to_cpu(h->h80.magic), - be16_to_cpu(h->h80.command), - be16_to_cpu(h->h80.length)); - return false; + conn_err(tconn, "Wrong magic value 0x%08x in protocol version %d\n", + be32_to_cpu(*(__be32 *)header), + tconn->agreed_pro_version); + return -EINVAL; } - mdev->last_received = jiffies; + pi->data = header + header_size; + return 0; +} - return true; +static int drbd_recv_header(struct drbd_tconn *tconn, struct packet_info *pi) +{ + void *buffer = tconn->data.rbuf; + int err; + + err = drbd_recv_all_warn(tconn, buffer, drbd_header_size(tconn)); + if (err) + return err; + + err = decode_header(tconn, buffer, pi); + tconn->last_received = jiffies; + + return err; } -static void drbd_flush(struct drbd_conf *mdev) +static void drbd_flush(struct drbd_tconn *tconn) { int rv; + struct drbd_conf *mdev; + int vnr; - if (mdev->write_ordering >= WO_bdev_flush && get_ldev(mdev)) { - rv = blkdev_issue_flush(mdev->ldev->backing_bdev, GFP_KERNEL, - NULL); - if (rv) { - dev_info(DEV, "local disk flush failed with status %d\n", rv); - /* would rather check on EOPNOTSUPP, but that is not reliable. - * don't try again for ANY return value != 0 - * if (rv == -EOPNOTSUPP) */ - drbd_bump_write_ordering(mdev, WO_drain_io); + if (tconn->write_ordering >= WO_bdev_flush) { + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) { + if (!get_ldev(mdev)) + continue; + kref_get(&mdev->kref); + rcu_read_unlock(); + + rv = blkdev_issue_flush(mdev->ldev->backing_bdev, + GFP_NOIO, NULL); + if (rv) { + dev_info(DEV, "local disk flush failed with status %d\n", rv); + /* would rather check on EOPNOTSUPP, but that is not reliable. + * don't try again for ANY return value != 0 + * if (rv == -EOPNOTSUPP) */ + drbd_bump_write_ordering(tconn, WO_drain_io); + } + put_ldev(mdev); + kref_put(&mdev->kref, &drbd_minor_destroy); + + rcu_read_lock(); + if (rv) + break; } - put_ldev(mdev); + rcu_read_unlock(); } } @@ -987,7 +1179,7 @@ static void drbd_flush(struct drbd_conf *mdev) * @epoch: Epoch object. * @ev: Epoch event. */ -static enum finish_epoch drbd_may_finish_epoch(struct drbd_conf *mdev, +static enum finish_epoch drbd_may_finish_epoch(struct drbd_tconn *tconn, struct drbd_epoch *epoch, enum epoch_event ev) { @@ -995,7 +1187,7 @@ static enum finish_epoch drbd_may_finish_epoch(struct drbd_conf *mdev, struct drbd_epoch *next_epoch; enum finish_epoch rv = FE_STILL_LIVE; - spin_lock(&mdev->epoch_lock); + spin_lock(&tconn->epoch_lock); do { next_epoch = NULL; @@ -1017,18 +1209,22 @@ static enum finish_epoch drbd_may_finish_epoch(struct drbd_conf *mdev, atomic_read(&epoch->active) == 0 && (test_bit(DE_HAVE_BARRIER_NUMBER, &epoch->flags) || ev & EV_CLEANUP)) { if (!(ev & EV_CLEANUP)) { - spin_unlock(&mdev->epoch_lock); - drbd_send_b_ack(mdev, epoch->barrier_nr, epoch_size); - spin_lock(&mdev->epoch_lock); + spin_unlock(&tconn->epoch_lock); + drbd_send_b_ack(epoch->tconn, epoch->barrier_nr, epoch_size); + spin_lock(&tconn->epoch_lock); } +#if 0 + /* FIXME: dec unacked on connection, once we have + * something to count pending connection packets in. */ if (test_bit(DE_HAVE_BARRIER_NUMBER, &epoch->flags)) - dec_unacked(mdev); + dec_unacked(epoch->tconn); +#endif - if (mdev->current_epoch != epoch) { + if (tconn->current_epoch != epoch) { next_epoch = list_entry(epoch->list.next, struct drbd_epoch, list); list_del(&epoch->list); ev = EV_BECAME_LAST | (ev & EV_CLEANUP); - mdev->epochs--; + tconn->epochs--; kfree(epoch); if (rv == FE_STILL_LIVE) @@ -1039,7 +1235,6 @@ static enum finish_epoch drbd_may_finish_epoch(struct drbd_conf *mdev, /* atomic_set(&epoch->active, 0); is already zero */ if (rv == FE_STILL_LIVE) rv = FE_RECYCLED; - wake_up(&mdev->ee_wait); } } @@ -1049,40 +1244,52 @@ static enum finish_epoch drbd_may_finish_epoch(struct drbd_conf *mdev, epoch = next_epoch; } while (1); - spin_unlock(&mdev->epoch_lock); + spin_unlock(&tconn->epoch_lock); return rv; } /** * drbd_bump_write_ordering() - Fall back to an other write ordering method - * @mdev: DRBD device. + * @tconn: DRBD connection. * @wo: Write ordering method to try. */ -void drbd_bump_write_ordering(struct drbd_conf *mdev, enum write_ordering_e wo) __must_hold(local) +void drbd_bump_write_ordering(struct drbd_tconn *tconn, enum write_ordering_e wo) { + struct disk_conf *dc; + struct drbd_conf *mdev; enum write_ordering_e pwo; + int vnr; static char *write_ordering_str[] = { [WO_none] = "none", [WO_drain_io] = "drain", [WO_bdev_flush] = "flush", }; - pwo = mdev->write_ordering; + pwo = tconn->write_ordering; wo = min(pwo, wo); - if (wo == WO_bdev_flush && mdev->ldev->dc.no_disk_flush) - wo = WO_drain_io; - if (wo == WO_drain_io && mdev->ldev->dc.no_disk_drain) - wo = WO_none; - mdev->write_ordering = wo; - if (pwo != mdev->write_ordering || wo == WO_bdev_flush) - dev_info(DEV, "Method to ensure write ordering: %s\n", write_ordering_str[mdev->write_ordering]); + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) { + if (!get_ldev_if_state(mdev, D_ATTACHING)) + continue; + dc = rcu_dereference(mdev->ldev->disk_conf); + + if (wo == WO_bdev_flush && !dc->disk_flushes) + wo = WO_drain_io; + if (wo == WO_drain_io && !dc->disk_drain) + wo = WO_none; + put_ldev(mdev); + } + rcu_read_unlock(); + tconn->write_ordering = wo; + if (pwo != tconn->write_ordering || wo == WO_bdev_flush) + conn_info(tconn, "Method to ensure write ordering: %s\n", write_ordering_str[tconn->write_ordering]); } /** - * drbd_submit_ee() + * drbd_submit_peer_request() * @mdev: DRBD device. - * @e: epoch entry + * @peer_req: peer request * @rw: flag field, see bio->bi_rw * * May spread the pages to multiple bios, @@ -1096,14 +1303,15 @@ void drbd_bump_write_ordering(struct drbd_conf *mdev, enum write_ordering_e wo) * on certain Xen deployments. */ /* TODO allocate from our own bio_set. */ -int drbd_submit_ee(struct drbd_conf *mdev, struct drbd_epoch_entry *e, - const unsigned rw, const int fault_type) +int drbd_submit_peer_request(struct drbd_conf *mdev, + struct drbd_peer_request *peer_req, + const unsigned rw, const int fault_type) { struct bio *bios = NULL; struct bio *bio; - struct page *page = e->pages; - sector_t sector = e->sector; - unsigned ds = e->size; + struct page *page = peer_req->pages; + sector_t sector = peer_req->i.sector; + unsigned ds = peer_req->i.size; unsigned n_bios = 0; unsigned nr_pages = (ds + PAGE_SIZE -1) >> PAGE_SHIFT; int err = -ENOMEM; @@ -1122,12 +1330,12 @@ next_bio: dev_err(DEV, "submit_ee: Allocation of a bio failed\n"); goto fail; } - /* > e->sector, unless this is the first bio */ + /* > peer_req->i.sector, unless this is the first bio */ bio->bi_sector = sector; bio->bi_bdev = mdev->ldev->backing_bdev; bio->bi_rw = rw; - bio->bi_private = e; - bio->bi_end_io = drbd_endio_sec; + bio->bi_private = peer_req; + bio->bi_end_io = drbd_peer_request_endio; bio->bi_next = bios; bios = bio; @@ -1156,7 +1364,7 @@ next_bio: D_ASSERT(page == NULL); D_ASSERT(ds == 0); - atomic_set(&e->pending_bios, n_bios); + atomic_set(&peer_req->pending_bios, n_bios); do { bio = bios; bios = bios->bi_next; @@ -1175,26 +1383,57 @@ fail: return err; } -static int receive_Barrier(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned int data_size) +static void drbd_remove_epoch_entry_interval(struct drbd_conf *mdev, + struct drbd_peer_request *peer_req) +{ + struct drbd_interval *i = &peer_req->i; + + drbd_remove_interval(&mdev->write_requests, i); + drbd_clear_interval(i); + + /* Wake up any processes waiting for this peer request to complete. */ + if (i->waiting) + wake_up(&mdev->misc_wait); +} + +void conn_wait_active_ee_empty(struct drbd_tconn *tconn) +{ + struct drbd_conf *mdev; + int vnr; + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) { + kref_get(&mdev->kref); + rcu_read_unlock(); + drbd_wait_ee_list_empty(mdev, &mdev->active_ee); + kref_put(&mdev->kref, &drbd_minor_destroy); + rcu_read_lock(); + } + rcu_read_unlock(); +} + +static int receive_Barrier(struct drbd_tconn *tconn, struct packet_info *pi) { int rv; - struct p_barrier *p = &mdev->data.rbuf.barrier; + struct p_barrier *p = pi->data; struct drbd_epoch *epoch; - inc_unacked(mdev); - - mdev->current_epoch->barrier_nr = p->barrier; - rv = drbd_may_finish_epoch(mdev, mdev->current_epoch, EV_GOT_BARRIER_NR); + /* FIXME these are unacked on connection, + * not a specific (peer)device. + */ + tconn->current_epoch->barrier_nr = p->barrier; + tconn->current_epoch->tconn = tconn; + rv = drbd_may_finish_epoch(tconn, tconn->current_epoch, EV_GOT_BARRIER_NR); /* P_BARRIER_ACK may imply that the corresponding extent is dropped from * the activity log, which means it would not be resynced in case the * R_PRIMARY crashes now. * Therefore we must send the barrier_ack after the barrier request was * completed. */ - switch (mdev->write_ordering) { + switch (tconn->write_ordering) { case WO_none: if (rv == FE_RECYCLED) - return true; + return 0; /* receiver context, in the writeout path of the other node. * avoid potential distributed deadlock */ @@ -1202,81 +1441,75 @@ static int receive_Barrier(struct drbd_conf *mdev, enum drbd_packets cmd, unsign if (epoch) break; else - dev_warn(DEV, "Allocation of an epoch failed, slowing down\n"); + conn_warn(tconn, "Allocation of an epoch failed, slowing down\n"); /* Fall through */ case WO_bdev_flush: case WO_drain_io: - drbd_wait_ee_list_empty(mdev, &mdev->active_ee); - drbd_flush(mdev); + conn_wait_active_ee_empty(tconn); + drbd_flush(tconn); - if (atomic_read(&mdev->current_epoch->epoch_size)) { + if (atomic_read(&tconn->current_epoch->epoch_size)) { epoch = kmalloc(sizeof(struct drbd_epoch), GFP_NOIO); if (epoch) break; } - epoch = mdev->current_epoch; - wait_event(mdev->ee_wait, atomic_read(&epoch->epoch_size) == 0); - - D_ASSERT(atomic_read(&epoch->active) == 0); - D_ASSERT(epoch->flags == 0); - - return true; + return 0; default: - dev_err(DEV, "Strangeness in mdev->write_ordering %d\n", mdev->write_ordering); - return false; + conn_err(tconn, "Strangeness in tconn->write_ordering %d\n", tconn->write_ordering); + return -EIO; } epoch->flags = 0; atomic_set(&epoch->epoch_size, 0); atomic_set(&epoch->active, 0); - spin_lock(&mdev->epoch_lock); - if (atomic_read(&mdev->current_epoch->epoch_size)) { - list_add(&epoch->list, &mdev->current_epoch->list); - mdev->current_epoch = epoch; - mdev->epochs++; + spin_lock(&tconn->epoch_lock); + if (atomic_read(&tconn->current_epoch->epoch_size)) { + list_add(&epoch->list, &tconn->current_epoch->list); + tconn->current_epoch = epoch; + tconn->epochs++; } else { /* The current_epoch got recycled while we allocated this one... */ kfree(epoch); } - spin_unlock(&mdev->epoch_lock); + spin_unlock(&tconn->epoch_lock); - return true; + return 0; } /* used from receive_RSDataReply (recv_resync_read) * and from receive_Data */ -static struct drbd_epoch_entry * -read_in_block(struct drbd_conf *mdev, u64 id, sector_t sector, int data_size) __must_hold(local) +static struct drbd_peer_request * +read_in_block(struct drbd_conf *mdev, u64 id, sector_t sector, + int data_size) __must_hold(local) { const sector_t capacity = drbd_get_capacity(mdev->this_bdev); - struct drbd_epoch_entry *e; + struct drbd_peer_request *peer_req; struct page *page; - int dgs, ds, rr; - void *dig_in = mdev->int_dig_in; - void *dig_vv = mdev->int_dig_vv; + int dgs, ds, err; + void *dig_in = mdev->tconn->int_dig_in; + void *dig_vv = mdev->tconn->int_dig_vv; unsigned long *data; - dgs = (mdev->agreed_pro_version >= 87 && mdev->integrity_r_tfm) ? - crypto_hash_digestsize(mdev->integrity_r_tfm) : 0; - - if (dgs) { - rr = drbd_recv(mdev, dig_in, dgs); - if (rr != dgs) { - if (!signal_pending(current)) - dev_warn(DEV, - "short read receiving data digest: read %d expected %d\n", - rr, dgs); + dgs = 0; + if (mdev->tconn->peer_integrity_tfm) { + dgs = crypto_hash_digestsize(mdev->tconn->peer_integrity_tfm); + /* + * FIXME: Receive the incoming digest into the receive buffer + * here, together with its struct p_data? + */ + err = drbd_recv_all_warn(mdev->tconn, dig_in, dgs); + if (err) return NULL; - } + data_size -= dgs; } - data_size -= dgs; - - ERR_IF(data_size & 0x1ff) return NULL; - ERR_IF(data_size > DRBD_MAX_BIO_SIZE) return NULL; + if (!expect(IS_ALIGNED(data_size, 512))) + return NULL; + if (!expect(data_size <= DRBD_MAX_BIO_SIZE)) + return NULL; /* even though we trust out peer, * we sometimes have to double check. */ @@ -1291,47 +1524,42 @@ read_in_block(struct drbd_conf *mdev, u64 id, sector_t sector, int data_size) __ /* GFP_NOIO, because we must not cause arbitrary write-out: in a DRBD * "criss-cross" setup, that might cause write-out on some other DRBD, * which in turn might block on the other node at this very place. */ - e = drbd_alloc_ee(mdev, id, sector, data_size, GFP_NOIO); - if (!e) + peer_req = drbd_alloc_peer_req(mdev, id, sector, data_size, GFP_NOIO); + if (!peer_req) return NULL; if (!data_size) - return e; + return peer_req; ds = data_size; - page = e->pages; + page = peer_req->pages; page_chain_for_each(page) { unsigned len = min_t(int, ds, PAGE_SIZE); data = kmap(page); - rr = drbd_recv(mdev, data, len); + err = drbd_recv_all_warn(mdev->tconn, data, len); if (drbd_insert_fault(mdev, DRBD_FAULT_RECEIVE)) { dev_err(DEV, "Fault injection: Corrupting data on receive\n"); data[0] = data[0] ^ (unsigned long)-1; } kunmap(page); - if (rr != len) { - drbd_free_ee(mdev, e); - if (!signal_pending(current)) - dev_warn(DEV, "short read receiving data: read %d expected %d\n", - rr, len); + if (err) { + drbd_free_peer_req(mdev, peer_req); return NULL; } - ds -= rr; + ds -= len; } if (dgs) { - drbd_csum_ee(mdev, mdev->integrity_r_tfm, e, dig_vv); + drbd_csum_ee(mdev, mdev->tconn->peer_integrity_tfm, peer_req, dig_vv); if (memcmp(dig_in, dig_vv, dgs)) { dev_err(DEV, "Digest integrity check FAILED: %llus +%u\n", (unsigned long long)sector, data_size); - drbd_bcast_ee(mdev, "digest failed", - dgs, dig_in, dig_vv, e); - drbd_free_ee(mdev, e); + drbd_free_peer_req(mdev, peer_req); return NULL; } } mdev->recv_cnt += data_size>>9; - return e; + return peer_req; } /* drbd_drain_block() just takes a data block @@ -1340,30 +1568,26 @@ read_in_block(struct drbd_conf *mdev, u64 id, sector_t sector, int data_size) __ static int drbd_drain_block(struct drbd_conf *mdev, int data_size) { struct page *page; - int rr, rv = 1; + int err = 0; void *data; if (!data_size) - return true; + return 0; - page = drbd_pp_alloc(mdev, 1, 1); + page = drbd_alloc_pages(mdev, 1, 1); data = kmap(page); while (data_size) { - rr = drbd_recv(mdev, data, min_t(int, data_size, PAGE_SIZE)); - if (rr != min_t(int, data_size, PAGE_SIZE)) { - rv = 0; - if (!signal_pending(current)) - dev_warn(DEV, - "short read receiving data: read %d expected %d\n", - rr, min_t(int, data_size, PAGE_SIZE)); + unsigned int len = min_t(int, data_size, PAGE_SIZE); + + err = drbd_recv_all_warn(mdev->tconn, data, len); + if (err) break; - } - data_size -= rr; + data_size -= len; } kunmap(page); - drbd_pp_free(mdev, page, 0); - return rv; + drbd_free_pages(mdev, page, 0); + return err; } static int recv_dless_read(struct drbd_conf *mdev, struct drbd_request *req, @@ -1371,26 +1595,19 @@ static int recv_dless_read(struct drbd_conf *mdev, struct drbd_request *req, { struct bio_vec *bvec; struct bio *bio; - int dgs, rr, i, expect; - void *dig_in = mdev->int_dig_in; - void *dig_vv = mdev->int_dig_vv; - - dgs = (mdev->agreed_pro_version >= 87 && mdev->integrity_r_tfm) ? - crypto_hash_digestsize(mdev->integrity_r_tfm) : 0; + int dgs, err, i, expect; + void *dig_in = mdev->tconn->int_dig_in; + void *dig_vv = mdev->tconn->int_dig_vv; - if (dgs) { - rr = drbd_recv(mdev, dig_in, dgs); - if (rr != dgs) { - if (!signal_pending(current)) - dev_warn(DEV, - "short read receiving data reply digest: read %d expected %d\n", - rr, dgs); - return 0; - } + dgs = 0; + if (mdev->tconn->peer_integrity_tfm) { + dgs = crypto_hash_digestsize(mdev->tconn->peer_integrity_tfm); + err = drbd_recv_all_warn(mdev->tconn, dig_in, dgs); + if (err) + return err; + data_size -= dgs; } - data_size -= dgs; - /* optimistically update recv_cnt. if receiving fails below, * we disconnect anyways, and counters will be reset. */ mdev->recv_cnt += data_size>>9; @@ -1399,63 +1616,61 @@ static int recv_dless_read(struct drbd_conf *mdev, struct drbd_request *req, D_ASSERT(sector == bio->bi_sector); bio_for_each_segment(bvec, bio, i) { + void *mapped = kmap(bvec->bv_page) + bvec->bv_offset; expect = min_t(int, data_size, bvec->bv_len); - rr = drbd_recv(mdev, - kmap(bvec->bv_page)+bvec->bv_offset, - expect); + err = drbd_recv_all_warn(mdev->tconn, mapped, expect); kunmap(bvec->bv_page); - if (rr != expect) { - if (!signal_pending(current)) - dev_warn(DEV, "short read receiving data reply: " - "read %d expected %d\n", - rr, expect); - return 0; - } - data_size -= rr; + if (err) + return err; + data_size -= expect; } if (dgs) { - drbd_csum_bio(mdev, mdev->integrity_r_tfm, bio, dig_vv); + drbd_csum_bio(mdev, mdev->tconn->peer_integrity_tfm, bio, dig_vv); if (memcmp(dig_in, dig_vv, dgs)) { dev_err(DEV, "Digest integrity check FAILED. Broken NICs?\n"); - return 0; + return -EINVAL; } } D_ASSERT(data_size == 0); - return 1; + return 0; } -/* e_end_resync_block() is called via - * drbd_process_done_ee() by asender only */ -static int e_end_resync_block(struct drbd_conf *mdev, struct drbd_work *w, int unused) +/* + * e_end_resync_block() is called in asender context via + * drbd_finish_peer_reqs(). + */ +static int e_end_resync_block(struct drbd_work *w, int unused) { - struct drbd_epoch_entry *e = (struct drbd_epoch_entry *)w; - sector_t sector = e->sector; - int ok; + struct drbd_peer_request *peer_req = + container_of(w, struct drbd_peer_request, w); + struct drbd_conf *mdev = w->mdev; + sector_t sector = peer_req->i.sector; + int err; - D_ASSERT(hlist_unhashed(&e->collision)); + D_ASSERT(drbd_interval_empty(&peer_req->i)); - if (likely((e->flags & EE_WAS_ERROR) == 0)) { - drbd_set_in_sync(mdev, sector, e->size); - ok = drbd_send_ack(mdev, P_RS_WRITE_ACK, e); + if (likely((peer_req->flags & EE_WAS_ERROR) == 0)) { + drbd_set_in_sync(mdev, sector, peer_req->i.size); + err = drbd_send_ack(mdev, P_RS_WRITE_ACK, peer_req); } else { /* Record failure to sync */ - drbd_rs_failed_io(mdev, sector, e->size); + drbd_rs_failed_io(mdev, sector, peer_req->i.size); - ok = drbd_send_ack(mdev, P_NEG_ACK, e); + err = drbd_send_ack(mdev, P_NEG_ACK, peer_req); } dec_unacked(mdev); - return ok; + return err; } static int recv_resync_read(struct drbd_conf *mdev, sector_t sector, int data_size) __releases(local) { - struct drbd_epoch_entry *e; + struct drbd_peer_request *peer_req; - e = read_in_block(mdev, ID_SYNCER, sector, data_size); - if (!e) + peer_req = read_in_block(mdev, ID_SYNCER, sector, data_size); + if (!peer_req) goto fail; dec_rs_pending(mdev); @@ -1464,64 +1679,88 @@ static int recv_resync_read(struct drbd_conf *mdev, sector_t sector, int data_si /* corresponding dec_unacked() in e_end_resync_block() * respective _drbd_clear_done_ee */ - e->w.cb = e_end_resync_block; + peer_req->w.cb = e_end_resync_block; - spin_lock_irq(&mdev->req_lock); - list_add(&e->w.list, &mdev->sync_ee); - spin_unlock_irq(&mdev->req_lock); + spin_lock_irq(&mdev->tconn->req_lock); + list_add(&peer_req->w.list, &mdev->sync_ee); + spin_unlock_irq(&mdev->tconn->req_lock); atomic_add(data_size >> 9, &mdev->rs_sect_ev); - if (drbd_submit_ee(mdev, e, WRITE, DRBD_FAULT_RS_WR) == 0) - return true; + if (drbd_submit_peer_request(mdev, peer_req, WRITE, DRBD_FAULT_RS_WR) == 0) + return 0; /* don't care for the reason here */ dev_err(DEV, "submit failed, triggering re-connect\n"); - spin_lock_irq(&mdev->req_lock); - list_del(&e->w.list); - spin_unlock_irq(&mdev->req_lock); + spin_lock_irq(&mdev->tconn->req_lock); + list_del(&peer_req->w.list); + spin_unlock_irq(&mdev->tconn->req_lock); - drbd_free_ee(mdev, e); + drbd_free_peer_req(mdev, peer_req); fail: put_ldev(mdev); - return false; + return -EIO; } -static int receive_DataReply(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned int data_size) +static struct drbd_request * +find_request(struct drbd_conf *mdev, struct rb_root *root, u64 id, + sector_t sector, bool missing_ok, const char *func) { struct drbd_request *req; + + /* Request object according to our peer */ + req = (struct drbd_request *)(unsigned long)id; + if (drbd_contains_interval(root, sector, &req->i) && req->i.local) + return req; + if (!missing_ok) { + dev_err(DEV, "%s: failed to find request 0x%lx, sector %llus\n", func, + (unsigned long)id, (unsigned long long)sector); + } + return NULL; +} + +static int receive_DataReply(struct drbd_tconn *tconn, struct packet_info *pi) +{ + struct drbd_conf *mdev; + struct drbd_request *req; sector_t sector; - int ok; - struct p_data *p = &mdev->data.rbuf.data; + int err; + struct p_data *p = pi->data; + + mdev = vnr_to_mdev(tconn, pi->vnr); + if (!mdev) + return -EIO; sector = be64_to_cpu(p->sector); - spin_lock_irq(&mdev->req_lock); - req = _ar_id_to_req(mdev, p->block_id, sector); - spin_unlock_irq(&mdev->req_lock); - if (unlikely(!req)) { - dev_err(DEV, "Got a corrupt block_id/sector pair(1).\n"); - return false; - } + spin_lock_irq(&mdev->tconn->req_lock); + req = find_request(mdev, &mdev->read_requests, p->block_id, sector, false, __func__); + spin_unlock_irq(&mdev->tconn->req_lock); + if (unlikely(!req)) + return -EIO; /* hlist_del(&req->collision) is done in _req_may_be_done, to avoid * special casing it there for the various failure cases. * still no race with drbd_fail_pending_reads */ - ok = recv_dless_read(mdev, req, sector, data_size); - - if (ok) - req_mod(req, data_received); + err = recv_dless_read(mdev, req, sector, pi->size); + if (!err) + req_mod(req, DATA_RECEIVED); /* else: nothing. handled from drbd_disconnect... * I don't think we may complete this just yet * in case we are "on-disconnect: freeze" */ - return ok; + return err; } -static int receive_RSDataReply(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned int data_size) +static int receive_RSDataReply(struct drbd_tconn *tconn, struct packet_info *pi) { + struct drbd_conf *mdev; sector_t sector; - int ok; - struct p_data *p = &mdev->data.rbuf.data; + int err; + struct p_data *p = pi->data; + + mdev = vnr_to_mdev(tconn, pi->vnr); + if (!mdev) + return -EIO; sector = be64_to_cpu(p->sector); D_ASSERT(p->block_id == ID_SYNCER); @@ -1529,42 +1768,63 @@ static int receive_RSDataReply(struct drbd_conf *mdev, enum drbd_packets cmd, un if (get_ldev(mdev)) { /* data is submitted to disk within recv_resync_read. * corresponding put_ldev done below on error, - * or in drbd_endio_write_sec. */ - ok = recv_resync_read(mdev, sector, data_size); + * or in drbd_peer_request_endio. */ + err = recv_resync_read(mdev, sector, pi->size); } else { if (__ratelimit(&drbd_ratelimit_state)) dev_err(DEV, "Can not write resync data to local disk.\n"); - ok = drbd_drain_block(mdev, data_size); + err = drbd_drain_block(mdev, pi->size); - drbd_send_ack_dp(mdev, P_NEG_ACK, p, data_size); + drbd_send_ack_dp(mdev, P_NEG_ACK, p, pi->size); } - atomic_add(data_size >> 9, &mdev->rs_sect_in); + atomic_add(pi->size >> 9, &mdev->rs_sect_in); - return ok; + return err; } -/* e_end_block() is called via drbd_process_done_ee(). - * this means this function only runs in the asender thread - */ -static int e_end_block(struct drbd_conf *mdev, struct drbd_work *w, int cancel) +static void restart_conflicting_writes(struct drbd_conf *mdev, + sector_t sector, int size) { - struct drbd_epoch_entry *e = (struct drbd_epoch_entry *)w; - sector_t sector = e->sector; - int ok = 1, pcmd; + struct drbd_interval *i; + struct drbd_request *req; - if (mdev->net_conf->wire_protocol == DRBD_PROT_C) { - if (likely((e->flags & EE_WAS_ERROR) == 0)) { + drbd_for_each_overlap(i, &mdev->write_requests, sector, size) { + if (!i->local) + continue; + req = container_of(i, struct drbd_request, i); + if (req->rq_state & RQ_LOCAL_PENDING || + !(req->rq_state & RQ_POSTPONED)) + continue; + /* as it is RQ_POSTPONED, this will cause it to + * be queued on the retry workqueue. */ + __req_mod(req, CONFLICT_RESOLVED, NULL); + } +} + +/* + * e_end_block() is called in asender context via drbd_finish_peer_reqs(). + */ +static int e_end_block(struct drbd_work *w, int cancel) +{ + struct drbd_peer_request *peer_req = + container_of(w, struct drbd_peer_request, w); + struct drbd_conf *mdev = w->mdev; + sector_t sector = peer_req->i.sector; + int err = 0, pcmd; + + if (peer_req->flags & EE_SEND_WRITE_ACK) { + if (likely((peer_req->flags & EE_WAS_ERROR) == 0)) { pcmd = (mdev->state.conn >= C_SYNC_SOURCE && mdev->state.conn <= C_PAUSED_SYNC_T && - e->flags & EE_MAY_SET_IN_SYNC) ? + peer_req->flags & EE_MAY_SET_IN_SYNC) ? P_RS_WRITE_ACK : P_WRITE_ACK; - ok &= drbd_send_ack(mdev, pcmd, e); + err = drbd_send_ack(mdev, pcmd, peer_req); if (pcmd == P_RS_WRITE_ACK) - drbd_set_in_sync(mdev, sector, e->size); + drbd_set_in_sync(mdev, sector, peer_req->i.size); } else { - ok = drbd_send_ack(mdev, P_NEG_ACK, e); + err = drbd_send_ack(mdev, P_NEG_ACK, peer_req); /* we expect it to be marked out of sync anyways... * maybe assert this? */ } @@ -1572,52 +1832,115 @@ static int e_end_block(struct drbd_conf *mdev, struct drbd_work *w, int cancel) } /* we delete from the conflict detection hash _after_ we sent out the * P_WRITE_ACK / P_NEG_ACK, to get the sequence number right. */ - if (mdev->net_conf->two_primaries) { - spin_lock_irq(&mdev->req_lock); - D_ASSERT(!hlist_unhashed(&e->collision)); - hlist_del_init(&e->collision); - spin_unlock_irq(&mdev->req_lock); - } else { - D_ASSERT(hlist_unhashed(&e->collision)); - } + if (peer_req->flags & EE_IN_INTERVAL_TREE) { + spin_lock_irq(&mdev->tconn->req_lock); + D_ASSERT(!drbd_interval_empty(&peer_req->i)); + drbd_remove_epoch_entry_interval(mdev, peer_req); + if (peer_req->flags & EE_RESTART_REQUESTS) + restart_conflicting_writes(mdev, sector, peer_req->i.size); + spin_unlock_irq(&mdev->tconn->req_lock); + } else + D_ASSERT(drbd_interval_empty(&peer_req->i)); - drbd_may_finish_epoch(mdev, e->epoch, EV_PUT + (cancel ? EV_CLEANUP : 0)); + drbd_may_finish_epoch(mdev->tconn, peer_req->epoch, EV_PUT + (cancel ? EV_CLEANUP : 0)); - return ok; + return err; } -static int e_send_discard_ack(struct drbd_conf *mdev, struct drbd_work *w, int unused) +static int e_send_ack(struct drbd_work *w, enum drbd_packet ack) { - struct drbd_epoch_entry *e = (struct drbd_epoch_entry *)w; - int ok = 1; + struct drbd_conf *mdev = w->mdev; + struct drbd_peer_request *peer_req = + container_of(w, struct drbd_peer_request, w); + int err; - D_ASSERT(mdev->net_conf->wire_protocol == DRBD_PROT_C); - ok = drbd_send_ack(mdev, P_DISCARD_ACK, e); + err = drbd_send_ack(mdev, ack, peer_req); + dec_unacked(mdev); - spin_lock_irq(&mdev->req_lock); - D_ASSERT(!hlist_unhashed(&e->collision)); - hlist_del_init(&e->collision); - spin_unlock_irq(&mdev->req_lock); + return err; +} - dec_unacked(mdev); +static int e_send_superseded(struct drbd_work *w, int unused) +{ + return e_send_ack(w, P_SUPERSEDED); +} + +static int e_send_retry_write(struct drbd_work *w, int unused) +{ + struct drbd_tconn *tconn = w->mdev->tconn; + + return e_send_ack(w, tconn->agreed_pro_version >= 100 ? + P_RETRY_WRITE : P_SUPERSEDED); +} + +static bool seq_greater(u32 a, u32 b) +{ + /* + * We assume 32-bit wrap-around here. + * For 24-bit wrap-around, we would have to shift: + * a <<= 8; b <<= 8; + */ + return (s32)a - (s32)b > 0; +} + +static u32 seq_max(u32 a, u32 b) +{ + return seq_greater(a, b) ? a : b; +} + +static bool need_peer_seq(struct drbd_conf *mdev) +{ + struct drbd_tconn *tconn = mdev->tconn; + int tp; - return ok; + /* + * We only need to keep track of the last packet_seq number of our peer + * if we are in dual-primary mode and we have the resolve-conflicts flag set; see + * handle_write_conflicts(). + */ + + rcu_read_lock(); + tp = rcu_dereference(mdev->tconn->net_conf)->two_primaries; + rcu_read_unlock(); + + return tp && test_bit(RESOLVE_CONFLICTS, &tconn->flags); } -static bool overlapping_resync_write(struct drbd_conf *mdev, struct drbd_epoch_entry *data_e) +static void update_peer_seq(struct drbd_conf *mdev, unsigned int peer_seq) { + unsigned int newest_peer_seq; - struct drbd_epoch_entry *rs_e; + if (need_peer_seq(mdev)) { + spin_lock(&mdev->peer_seq_lock); + newest_peer_seq = seq_max(mdev->peer_seq, peer_seq); + mdev->peer_seq = newest_peer_seq; + spin_unlock(&mdev->peer_seq_lock); + /* wake up only if we actually changed mdev->peer_seq */ + if (peer_seq == newest_peer_seq) + wake_up(&mdev->seq_wait); + } +} + +static inline int overlaps(sector_t s1, int l1, sector_t s2, int l2) +{ + return !((s1 + (l1>>9) <= s2) || (s1 >= s2 + (l2>>9))); +} + +/* maybe change sync_ee into interval trees as well? */ +static bool overlapping_resync_write(struct drbd_conf *mdev, struct drbd_peer_request *peer_req) +{ + struct drbd_peer_request *rs_req; bool rv = 0; - spin_lock_irq(&mdev->req_lock); - list_for_each_entry(rs_e, &mdev->sync_ee, w.list) { - if (overlaps(data_e->sector, data_e->size, rs_e->sector, rs_e->size)) { + spin_lock_irq(&mdev->tconn->req_lock); + list_for_each_entry(rs_req, &mdev->sync_ee, w.list) { + if (overlaps(peer_req->i.sector, peer_req->i.size, + rs_req->i.sector, rs_req->i.size)) { rv = 1; break; } } - spin_unlock_irq(&mdev->req_lock); + spin_unlock_irq(&mdev->tconn->req_lock); return rv; } @@ -1643,35 +1966,41 @@ static bool overlapping_resync_write(struct drbd_conf *mdev, struct drbd_epoch_e * * returns 0 if we may process the packet, * -ERESTARTSYS if we were interrupted (by disconnect signal). */ -static int drbd_wait_peer_seq(struct drbd_conf *mdev, const u32 packet_seq) +static int wait_for_and_update_peer_seq(struct drbd_conf *mdev, const u32 peer_seq) { DEFINE_WAIT(wait); - unsigned int p_seq; long timeout; - int ret = 0; + int ret; + + if (!need_peer_seq(mdev)) + return 0; + spin_lock(&mdev->peer_seq_lock); for (;;) { - prepare_to_wait(&mdev->seq_wait, &wait, TASK_INTERRUPTIBLE); - if (seq_le(packet_seq, mdev->peer_seq+1)) + if (!seq_greater(peer_seq - 1, mdev->peer_seq)) { + mdev->peer_seq = seq_max(mdev->peer_seq, peer_seq); + ret = 0; break; + } if (signal_pending(current)) { ret = -ERESTARTSYS; break; } - p_seq = mdev->peer_seq; + prepare_to_wait(&mdev->seq_wait, &wait, TASK_INTERRUPTIBLE); spin_unlock(&mdev->peer_seq_lock); - timeout = schedule_timeout(30*HZ); + rcu_read_lock(); + timeout = rcu_dereference(mdev->tconn->net_conf)->ping_timeo*HZ/10; + rcu_read_unlock(); + timeout = schedule_timeout(timeout); spin_lock(&mdev->peer_seq_lock); - if (timeout == 0 && p_seq == mdev->peer_seq) { + if (!timeout) { ret = -ETIMEDOUT; - dev_err(DEV, "ASSERT FAILED waited 30 seconds for sequence update, forcing reconnect\n"); + dev_err(DEV, "Timed out waiting for missing ack packets; disconnecting\n"); break; } } - finish_wait(&mdev->seq_wait, &wait); - if (mdev->peer_seq+1 == packet_seq) - mdev->peer_seq++; spin_unlock(&mdev->peer_seq_lock); + finish_wait(&mdev->seq_wait, &wait); return ret; } @@ -1686,233 +2015,277 @@ static unsigned long wire_flags_to_bio(struct drbd_conf *mdev, u32 dpf) (dpf & DP_DISCARD ? REQ_DISCARD : 0); } +static void fail_postponed_requests(struct drbd_conf *mdev, sector_t sector, + unsigned int size) +{ + struct drbd_interval *i; + + repeat: + drbd_for_each_overlap(i, &mdev->write_requests, sector, size) { + struct drbd_request *req; + struct bio_and_error m; + + if (!i->local) + continue; + req = container_of(i, struct drbd_request, i); + if (!(req->rq_state & RQ_POSTPONED)) + continue; + req->rq_state &= ~RQ_POSTPONED; + __req_mod(req, NEG_ACKED, &m); + spin_unlock_irq(&mdev->tconn->req_lock); + if (m.bio) + complete_master_bio(mdev, &m); + spin_lock_irq(&mdev->tconn->req_lock); + goto repeat; + } +} + +static int handle_write_conflicts(struct drbd_conf *mdev, + struct drbd_peer_request *peer_req) +{ + struct drbd_tconn *tconn = mdev->tconn; + bool resolve_conflicts = test_bit(RESOLVE_CONFLICTS, &tconn->flags); + sector_t sector = peer_req->i.sector; + const unsigned int size = peer_req->i.size; + struct drbd_interval *i; + bool equal; + int err; + + /* + * Inserting the peer request into the write_requests tree will prevent + * new conflicting local requests from being added. + */ + drbd_insert_interval(&mdev->write_requests, &peer_req->i); + + repeat: + drbd_for_each_overlap(i, &mdev->write_requests, sector, size) { + if (i == &peer_req->i) + continue; + + if (!i->local) { + /* + * Our peer has sent a conflicting remote request; this + * should not happen in a two-node setup. Wait for the + * earlier peer request to complete. + */ + err = drbd_wait_misc(mdev, i); + if (err) + goto out; + goto repeat; + } + + equal = i->sector == sector && i->size == size; + if (resolve_conflicts) { + /* + * If the peer request is fully contained within the + * overlapping request, it can be considered overwritten + * and thus superseded; otherwise, it will be retried + * once all overlapping requests have completed. + */ + bool superseded = i->sector <= sector && i->sector + + (i->size >> 9) >= sector + (size >> 9); + + if (!equal) + dev_alert(DEV, "Concurrent writes detected: " + "local=%llus +%u, remote=%llus +%u, " + "assuming %s came first\n", + (unsigned long long)i->sector, i->size, + (unsigned long long)sector, size, + superseded ? "local" : "remote"); + + inc_unacked(mdev); + peer_req->w.cb = superseded ? e_send_superseded : + e_send_retry_write; + list_add_tail(&peer_req->w.list, &mdev->done_ee); + wake_asender(mdev->tconn); + + err = -ENOENT; + goto out; + } else { + struct drbd_request *req = + container_of(i, struct drbd_request, i); + + if (!equal) + dev_alert(DEV, "Concurrent writes detected: " + "local=%llus +%u, remote=%llus +%u\n", + (unsigned long long)i->sector, i->size, + (unsigned long long)sector, size); + + if (req->rq_state & RQ_LOCAL_PENDING || + !(req->rq_state & RQ_POSTPONED)) { + /* + * Wait for the node with the discard flag to + * decide if this request has been superseded + * or needs to be retried. + * Requests that have been superseded will + * disappear from the write_requests tree. + * + * In addition, wait for the conflicting + * request to finish locally before submitting + * the conflicting peer request. + */ + err = drbd_wait_misc(mdev, &req->i); + if (err) { + _conn_request_state(mdev->tconn, + NS(conn, C_TIMEOUT), + CS_HARD); + fail_postponed_requests(mdev, sector, size); + goto out; + } + goto repeat; + } + /* + * Remember to restart the conflicting requests after + * the new peer request has completed. + */ + peer_req->flags |= EE_RESTART_REQUESTS; + } + } + err = 0; + + out: + if (err) + drbd_remove_epoch_entry_interval(mdev, peer_req); + return err; +} + /* mirrored write */ -static int receive_Data(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned int data_size) +static int receive_Data(struct drbd_tconn *tconn, struct packet_info *pi) { + struct drbd_conf *mdev; sector_t sector; - struct drbd_epoch_entry *e; - struct p_data *p = &mdev->data.rbuf.data; + struct drbd_peer_request *peer_req; + struct p_data *p = pi->data; + u32 peer_seq = be32_to_cpu(p->seq_num); int rw = WRITE; u32 dp_flags; + int err, tp; - if (!get_ldev(mdev)) { - spin_lock(&mdev->peer_seq_lock); - if (mdev->peer_seq+1 == be32_to_cpu(p->seq_num)) - mdev->peer_seq++; - spin_unlock(&mdev->peer_seq_lock); + mdev = vnr_to_mdev(tconn, pi->vnr); + if (!mdev) + return -EIO; - drbd_send_ack_dp(mdev, P_NEG_ACK, p, data_size); - atomic_inc(&mdev->current_epoch->epoch_size); - return drbd_drain_block(mdev, data_size); + if (!get_ldev(mdev)) { + int err2; + + err = wait_for_and_update_peer_seq(mdev, peer_seq); + drbd_send_ack_dp(mdev, P_NEG_ACK, p, pi->size); + atomic_inc(&tconn->current_epoch->epoch_size); + err2 = drbd_drain_block(mdev, pi->size); + if (!err) + err = err2; + return err; } - /* get_ldev(mdev) successful. - * Corresponding put_ldev done either below (on various errors), - * or in drbd_endio_write_sec, if we successfully submit the data at - * the end of this function. */ + /* + * Corresponding put_ldev done either below (on various errors), or in + * drbd_peer_request_endio, if we successfully submit the data at the + * end of this function. + */ sector = be64_to_cpu(p->sector); - e = read_in_block(mdev, p->block_id, sector, data_size); - if (!e) { + peer_req = read_in_block(mdev, p->block_id, sector, pi->size); + if (!peer_req) { put_ldev(mdev); - return false; + return -EIO; } - e->w.cb = e_end_block; + peer_req->w.cb = e_end_block; dp_flags = be32_to_cpu(p->dp_flags); rw |= wire_flags_to_bio(mdev, dp_flags); - if (e->pages == NULL) { - D_ASSERT(e->size == 0); + if (peer_req->pages == NULL) { + D_ASSERT(peer_req->i.size == 0); D_ASSERT(dp_flags & DP_FLUSH); } if (dp_flags & DP_MAY_SET_IN_SYNC) - e->flags |= EE_MAY_SET_IN_SYNC; - - spin_lock(&mdev->epoch_lock); - e->epoch = mdev->current_epoch; - atomic_inc(&e->epoch->epoch_size); - atomic_inc(&e->epoch->active); - spin_unlock(&mdev->epoch_lock); - - /* I'm the receiver, I do hold a net_cnt reference. */ - if (!mdev->net_conf->two_primaries) { - spin_lock_irq(&mdev->req_lock); - } else { - /* don't get the req_lock yet, - * we may sleep in drbd_wait_peer_seq */ - const int size = e->size; - const int discard = test_bit(DISCARD_CONCURRENT, &mdev->flags); - DEFINE_WAIT(wait); - struct drbd_request *i; - struct hlist_node *n; - struct hlist_head *slot; - int first; - - D_ASSERT(mdev->net_conf->wire_protocol == DRBD_PROT_C); - BUG_ON(mdev->ee_hash == NULL); - BUG_ON(mdev->tl_hash == NULL); - - /* conflict detection and handling: - * 1. wait on the sequence number, - * in case this data packet overtook ACK packets. - * 2. check our hash tables for conflicting requests. - * we only need to walk the tl_hash, since an ee can not - * have a conflict with an other ee: on the submitting - * node, the corresponding req had already been conflicting, - * and a conflicting req is never sent. - * - * Note: for two_primaries, we are protocol C, - * so there cannot be any request that is DONE - * but still on the transfer log. - * - * unconditionally add to the ee_hash. - * - * if no conflicting request is found: - * submit. - * - * if any conflicting request is found - * that has not yet been acked, - * AND I have the "discard concurrent writes" flag: - * queue (via done_ee) the P_DISCARD_ACK; OUT. - * - * if any conflicting request is found: - * block the receiver, waiting on misc_wait - * until no more conflicting requests are there, - * or we get interrupted (disconnect). - * - * we do not just write after local io completion of those - * requests, but only after req is done completely, i.e. - * we wait for the P_DISCARD_ACK to arrive! - * - * then proceed normally, i.e. submit. - */ - if (drbd_wait_peer_seq(mdev, be32_to_cpu(p->seq_num))) + peer_req->flags |= EE_MAY_SET_IN_SYNC; + + spin_lock(&tconn->epoch_lock); + peer_req->epoch = tconn->current_epoch; + atomic_inc(&peer_req->epoch->epoch_size); + atomic_inc(&peer_req->epoch->active); + spin_unlock(&tconn->epoch_lock); + + rcu_read_lock(); + tp = rcu_dereference(mdev->tconn->net_conf)->two_primaries; + rcu_read_unlock(); + if (tp) { + peer_req->flags |= EE_IN_INTERVAL_TREE; + err = wait_for_and_update_peer_seq(mdev, peer_seq); + if (err) goto out_interrupted; - - spin_lock_irq(&mdev->req_lock); - - hlist_add_head(&e->collision, ee_hash_slot(mdev, sector)); - -#define OVERLAPS overlaps(i->sector, i->size, sector, size) - slot = tl_hash_slot(mdev, sector); - first = 1; - for (;;) { - int have_unacked = 0; - int have_conflict = 0; - prepare_to_wait(&mdev->misc_wait, &wait, - TASK_INTERRUPTIBLE); - hlist_for_each_entry(i, n, slot, collision) { - if (OVERLAPS) { - /* only ALERT on first iteration, - * we may be woken up early... */ - if (first) - dev_alert(DEV, "%s[%u] Concurrent local write detected!" - " new: %llus +%u; pending: %llus +%u\n", - current->comm, current->pid, - (unsigned long long)sector, size, - (unsigned long long)i->sector, i->size); - if (i->rq_state & RQ_NET_PENDING) - ++have_unacked; - ++have_conflict; - } - } -#undef OVERLAPS - if (!have_conflict) - break; - - /* Discard Ack only for the _first_ iteration */ - if (first && discard && have_unacked) { - dev_alert(DEV, "Concurrent write! [DISCARD BY FLAG] sec=%llus\n", - (unsigned long long)sector); - inc_unacked(mdev); - e->w.cb = e_send_discard_ack; - list_add_tail(&e->w.list, &mdev->done_ee); - - spin_unlock_irq(&mdev->req_lock); - - /* we could probably send that P_DISCARD_ACK ourselves, - * but I don't like the receiver using the msock */ - + spin_lock_irq(&mdev->tconn->req_lock); + err = handle_write_conflicts(mdev, peer_req); + if (err) { + spin_unlock_irq(&mdev->tconn->req_lock); + if (err == -ENOENT) { put_ldev(mdev); - wake_asender(mdev); - finish_wait(&mdev->misc_wait, &wait); - return true; + return 0; } + goto out_interrupted; + } + } else + spin_lock_irq(&mdev->tconn->req_lock); + list_add(&peer_req->w.list, &mdev->active_ee); + spin_unlock_irq(&mdev->tconn->req_lock); - if (signal_pending(current)) { - hlist_del_init(&e->collision); - - spin_unlock_irq(&mdev->req_lock); - - finish_wait(&mdev->misc_wait, &wait); - goto out_interrupted; - } + if (mdev->state.conn == C_SYNC_TARGET) + wait_event(mdev->ee_wait, !overlapping_resync_write(mdev, peer_req)); - spin_unlock_irq(&mdev->req_lock); - if (first) { - first = 0; - dev_alert(DEV, "Concurrent write! [W AFTERWARDS] " - "sec=%llus\n", (unsigned long long)sector); - } else if (discard) { - /* we had none on the first iteration. - * there must be none now. */ - D_ASSERT(have_unacked == 0); - } - schedule(); - spin_lock_irq(&mdev->req_lock); + if (mdev->tconn->agreed_pro_version < 100) { + rcu_read_lock(); + switch (rcu_dereference(mdev->tconn->net_conf)->wire_protocol) { + case DRBD_PROT_C: + dp_flags |= DP_SEND_WRITE_ACK; + break; + case DRBD_PROT_B: + dp_flags |= DP_SEND_RECEIVE_ACK; + break; } - finish_wait(&mdev->misc_wait, &wait); + rcu_read_unlock(); } - list_add(&e->w.list, &mdev->active_ee); - spin_unlock_irq(&mdev->req_lock); - - if (mdev->state.conn == C_SYNC_TARGET) - wait_event(mdev->ee_wait, !overlapping_resync_write(mdev, e)); - - switch (mdev->net_conf->wire_protocol) { - case DRBD_PROT_C: + if (dp_flags & DP_SEND_WRITE_ACK) { + peer_req->flags |= EE_SEND_WRITE_ACK; inc_unacked(mdev); /* corresponding dec_unacked() in e_end_block() * respective _drbd_clear_done_ee */ - break; - case DRBD_PROT_B: + } + + if (dp_flags & DP_SEND_RECEIVE_ACK) { /* I really don't like it that the receiver thread * sends on the msock, but anyways */ - drbd_send_ack(mdev, P_RECV_ACK, e); - break; - case DRBD_PROT_A: - /* nothing to do */ - break; + drbd_send_ack(mdev, P_RECV_ACK, peer_req); } if (mdev->state.pdsk < D_INCONSISTENT) { /* In case we have the only disk of the cluster, */ - drbd_set_out_of_sync(mdev, e->sector, e->size); - e->flags |= EE_CALL_AL_COMPLETE_IO; - e->flags &= ~EE_MAY_SET_IN_SYNC; - drbd_al_begin_io(mdev, e->sector); + drbd_set_out_of_sync(mdev, peer_req->i.sector, peer_req->i.size); + peer_req->flags |= EE_CALL_AL_COMPLETE_IO; + peer_req->flags &= ~EE_MAY_SET_IN_SYNC; + drbd_al_begin_io(mdev, &peer_req->i); } - if (drbd_submit_ee(mdev, e, rw, DRBD_FAULT_DT_WR) == 0) - return true; + err = drbd_submit_peer_request(mdev, peer_req, rw, DRBD_FAULT_DT_WR); + if (!err) + return 0; /* don't care for the reason here */ dev_err(DEV, "submit failed, triggering re-connect\n"); - spin_lock_irq(&mdev->req_lock); - list_del(&e->w.list); - hlist_del_init(&e->collision); - spin_unlock_irq(&mdev->req_lock); - if (e->flags & EE_CALL_AL_COMPLETE_IO) - drbd_al_complete_io(mdev, e->sector); + spin_lock_irq(&mdev->tconn->req_lock); + list_del(&peer_req->w.list); + drbd_remove_epoch_entry_interval(mdev, peer_req); + spin_unlock_irq(&mdev->tconn->req_lock); + if (peer_req->flags & EE_CALL_AL_COMPLETE_IO) + drbd_al_complete_io(mdev, &peer_req->i); out_interrupted: - drbd_may_finish_epoch(mdev, e->epoch, EV_PUT + EV_CLEANUP); + drbd_may_finish_epoch(tconn, peer_req->epoch, EV_PUT + EV_CLEANUP); put_ldev(mdev); - drbd_free_ee(mdev, e); - return false; + drbd_free_peer_req(mdev, peer_req); + return err; } /* We may throttle resync, if the lower device seems to be busy, @@ -1933,9 +2306,14 @@ int drbd_rs_should_slow_down(struct drbd_conf *mdev, sector_t sector) struct lc_element *tmp; int curr_events; int throttle = 0; + unsigned int c_min_rate; + + rcu_read_lock(); + c_min_rate = rcu_dereference(mdev->ldev->disk_conf)->c_min_rate; + rcu_read_unlock(); /* feature disabled? */ - if (mdev->sync_conf.c_min_rate == 0) + if (c_min_rate == 0) return 0; spin_lock_irq(&mdev->al_lock); @@ -1975,40 +2353,46 @@ int drbd_rs_should_slow_down(struct drbd_conf *mdev, sector_t sector) db = mdev->rs_mark_left[i] - rs_left; dbdt = Bit2KB(db/dt); - if (dbdt > mdev->sync_conf.c_min_rate) + if (dbdt > c_min_rate) throttle = 1; } return throttle; } -static int receive_DataRequest(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned int digest_size) +static int receive_DataRequest(struct drbd_tconn *tconn, struct packet_info *pi) { + struct drbd_conf *mdev; sector_t sector; - const sector_t capacity = drbd_get_capacity(mdev->this_bdev); - struct drbd_epoch_entry *e; + sector_t capacity; + struct drbd_peer_request *peer_req; struct digest_info *di = NULL; int size, verb; unsigned int fault_type; - struct p_block_req *p = &mdev->data.rbuf.block_req; + struct p_block_req *p = pi->data; + + mdev = vnr_to_mdev(tconn, pi->vnr); + if (!mdev) + return -EIO; + capacity = drbd_get_capacity(mdev->this_bdev); sector = be64_to_cpu(p->sector); size = be32_to_cpu(p->blksize); - if (size <= 0 || (size & 0x1ff) != 0 || size > DRBD_MAX_BIO_SIZE) { + if (size <= 0 || !IS_ALIGNED(size, 512) || size > DRBD_MAX_BIO_SIZE) { dev_err(DEV, "%s:%d: sector: %llus, size: %u\n", __FILE__, __LINE__, (unsigned long long)sector, size); - return false; + return -EINVAL; } if (sector + (size>>9) > capacity) { dev_err(DEV, "%s:%d: sector: %llus, size: %u\n", __FILE__, __LINE__, (unsigned long long)sector, size); - return false; + return -EINVAL; } if (!get_ldev_if_state(mdev, D_UP_TO_DATE)) { verb = 1; - switch (cmd) { + switch (pi->cmd) { case P_DATA_REQUEST: drbd_send_ack_rp(mdev, P_NEG_DREPLY, p); break; @@ -2023,35 +2407,34 @@ static int receive_DataRequest(struct drbd_conf *mdev, enum drbd_packets cmd, un drbd_send_ack_ex(mdev, P_OV_RESULT, sector, size, ID_IN_SYNC); break; default: - dev_err(DEV, "unexpected command (%s) in receive_DataRequest\n", - cmdname(cmd)); + BUG(); } if (verb && __ratelimit(&drbd_ratelimit_state)) dev_err(DEV, "Can not satisfy peer's read request, " "no local data.\n"); /* drain possibly payload */ - return drbd_drain_block(mdev, digest_size); + return drbd_drain_block(mdev, pi->size); } /* GFP_NOIO, because we must not cause arbitrary write-out: in a DRBD * "criss-cross" setup, that might cause write-out on some other DRBD, * which in turn might block on the other node at this very place. */ - e = drbd_alloc_ee(mdev, p->block_id, sector, size, GFP_NOIO); - if (!e) { + peer_req = drbd_alloc_peer_req(mdev, p->block_id, sector, size, GFP_NOIO); + if (!peer_req) { put_ldev(mdev); - return false; + return -ENOMEM; } - switch (cmd) { + switch (pi->cmd) { case P_DATA_REQUEST: - e->w.cb = w_e_end_data_req; + peer_req->w.cb = w_e_end_data_req; fault_type = DRBD_FAULT_DT_RD; /* application IO, don't drbd_rs_begin_io */ goto submit; case P_RS_DATA_REQUEST: - e->w.cb = w_e_end_rsdata_req; + peer_req->w.cb = w_e_end_rsdata_req; fault_type = DRBD_FAULT_RS_RD; /* used in the sector offset progress display */ mdev->bm_resync_fo = BM_SECT_TO_BIT(sector); @@ -2060,28 +2443,28 @@ static int receive_DataRequest(struct drbd_conf *mdev, enum drbd_packets cmd, un case P_OV_REPLY: case P_CSUM_RS_REQUEST: fault_type = DRBD_FAULT_RS_RD; - di = kmalloc(sizeof(*di) + digest_size, GFP_NOIO); + di = kmalloc(sizeof(*di) + pi->size, GFP_NOIO); if (!di) goto out_free_e; - di->digest_size = digest_size; + di->digest_size = pi->size; di->digest = (((char *)di)+sizeof(struct digest_info)); - e->digest = di; - e->flags |= EE_HAS_DIGEST; + peer_req->digest = di; + peer_req->flags |= EE_HAS_DIGEST; - if (drbd_recv(mdev, di->digest, digest_size) != digest_size) + if (drbd_recv_all(mdev->tconn, di->digest, pi->size)) goto out_free_e; - if (cmd == P_CSUM_RS_REQUEST) { - D_ASSERT(mdev->agreed_pro_version >= 89); - e->w.cb = w_e_end_csum_rs_req; + if (pi->cmd == P_CSUM_RS_REQUEST) { + D_ASSERT(mdev->tconn->agreed_pro_version >= 89); + peer_req->w.cb = w_e_end_csum_rs_req; /* used in the sector offset progress display */ mdev->bm_resync_fo = BM_SECT_TO_BIT(sector); - } else if (cmd == P_OV_REPLY) { + } else if (pi->cmd == P_OV_REPLY) { /* track progress, we may need to throttle */ atomic_add(size >> 9, &mdev->rs_sect_in); - e->w.cb = w_e_end_ov_reply; + peer_req->w.cb = w_e_end_ov_reply; dec_rs_pending(mdev); /* drbd_rs_begin_io done when we sent this request, * but accounting still needs to be done. */ @@ -2091,7 +2474,7 @@ static int receive_DataRequest(struct drbd_conf *mdev, enum drbd_packets cmd, un case P_OV_REQUEST: if (mdev->ov_start_sector == ~(sector_t)0 && - mdev->agreed_pro_version >= 90) { + mdev->tconn->agreed_pro_version >= 90) { unsigned long now = jiffies; int i; mdev->ov_start_sector = sector; @@ -2105,15 +2488,12 @@ static int receive_DataRequest(struct drbd_conf *mdev, enum drbd_packets cmd, un dev_info(DEV, "Online Verify start sector: %llu\n", (unsigned long long)sector); } - e->w.cb = w_e_end_ov_req; + peer_req->w.cb = w_e_end_ov_req; fault_type = DRBD_FAULT_RS_RD; break; default: - dev_err(DEV, "unexpected command (%s) in receive_DataRequest\n", - cmdname(cmd)); - fault_type = DRBD_FAULT_MAX; - goto out_free_e; + BUG(); } /* Throttle, drbd_rs_begin_io and submit should become asynchronous @@ -2148,30 +2528,31 @@ submit_for_resync: submit: inc_unacked(mdev); - spin_lock_irq(&mdev->req_lock); - list_add_tail(&e->w.list, &mdev->read_ee); - spin_unlock_irq(&mdev->req_lock); + spin_lock_irq(&mdev->tconn->req_lock); + list_add_tail(&peer_req->w.list, &mdev->read_ee); + spin_unlock_irq(&mdev->tconn->req_lock); - if (drbd_submit_ee(mdev, e, READ, fault_type) == 0) - return true; + if (drbd_submit_peer_request(mdev, peer_req, READ, fault_type) == 0) + return 0; /* don't care for the reason here */ dev_err(DEV, "submit failed, triggering re-connect\n"); - spin_lock_irq(&mdev->req_lock); - list_del(&e->w.list); - spin_unlock_irq(&mdev->req_lock); + spin_lock_irq(&mdev->tconn->req_lock); + list_del(&peer_req->w.list); + spin_unlock_irq(&mdev->tconn->req_lock); /* no drbd_rs_complete_io(), we are dropping the connection anyways */ out_free_e: put_ldev(mdev); - drbd_free_ee(mdev, e); - return false; + drbd_free_peer_req(mdev, peer_req); + return -EIO; } static int drbd_asb_recover_0p(struct drbd_conf *mdev) __must_hold(local) { int self, peer, rv = -100; unsigned long ch_self, ch_peer; + enum drbd_after_sb_p after_sb_0p; self = mdev->ldev->md.uuid[UI_BITMAP] & 1; peer = mdev->p_uuid[UI_BITMAP] & 1; @@ -2179,10 +2560,14 @@ static int drbd_asb_recover_0p(struct drbd_conf *mdev) __must_hold(local) ch_peer = mdev->p_uuid[UI_SIZE]; ch_self = mdev->comm_bm_set; - switch (mdev->net_conf->after_sb_0p) { + rcu_read_lock(); + after_sb_0p = rcu_dereference(mdev->tconn->net_conf)->after_sb_0p; + rcu_read_unlock(); + switch (after_sb_0p) { case ASB_CONSENSUS: case ASB_DISCARD_SECONDARY: case ASB_CALL_HELPER: + case ASB_VIOLENTLY: dev_err(DEV, "Configuration error.\n"); break; case ASB_DISCONNECT: @@ -2211,14 +2596,14 @@ static int drbd_asb_recover_0p(struct drbd_conf *mdev) __must_hold(local) "Using discard-least-changes instead\n"); case ASB_DISCARD_ZERO_CHG: if (ch_peer == 0 && ch_self == 0) { - rv = test_bit(DISCARD_CONCURRENT, &mdev->flags) + rv = test_bit(RESOLVE_CONFLICTS, &mdev->tconn->flags) ? -1 : 1; break; } else { if (ch_peer == 0) { rv = 1; break; } if (ch_self == 0) { rv = -1; break; } } - if (mdev->net_conf->after_sb_0p == ASB_DISCARD_ZERO_CHG) + if (after_sb_0p == ASB_DISCARD_ZERO_CHG) break; case ASB_DISCARD_LEAST_CHG: if (ch_self < ch_peer) @@ -2227,7 +2612,7 @@ static int drbd_asb_recover_0p(struct drbd_conf *mdev) __must_hold(local) rv = 1; else /* ( ch_self == ch_peer ) */ /* Well, then use something else. */ - rv = test_bit(DISCARD_CONCURRENT, &mdev->flags) + rv = test_bit(RESOLVE_CONFLICTS, &mdev->tconn->flags) ? -1 : 1; break; case ASB_DISCARD_LOCAL: @@ -2243,13 +2628,18 @@ static int drbd_asb_recover_0p(struct drbd_conf *mdev) __must_hold(local) static int drbd_asb_recover_1p(struct drbd_conf *mdev) __must_hold(local) { int hg, rv = -100; + enum drbd_after_sb_p after_sb_1p; - switch (mdev->net_conf->after_sb_1p) { + rcu_read_lock(); + after_sb_1p = rcu_dereference(mdev->tconn->net_conf)->after_sb_1p; + rcu_read_unlock(); + switch (after_sb_1p) { case ASB_DISCARD_YOUNGER_PRI: case ASB_DISCARD_OLDER_PRI: case ASB_DISCARD_LEAST_CHG: case ASB_DISCARD_LOCAL: case ASB_DISCARD_REMOTE: + case ASB_DISCARD_ZERO_CHG: dev_err(DEV, "Configuration error.\n"); break; case ASB_DISCONNECT: @@ -2292,8 +2682,12 @@ static int drbd_asb_recover_1p(struct drbd_conf *mdev) __must_hold(local) static int drbd_asb_recover_2p(struct drbd_conf *mdev) __must_hold(local) { int hg, rv = -100; + enum drbd_after_sb_p after_sb_2p; - switch (mdev->net_conf->after_sb_2p) { + rcu_read_lock(); + after_sb_2p = rcu_dereference(mdev->tconn->net_conf)->after_sb_2p; + rcu_read_unlock(); + switch (after_sb_2p) { case ASB_DISCARD_YOUNGER_PRI: case ASB_DISCARD_OLDER_PRI: case ASB_DISCARD_LEAST_CHG: @@ -2301,6 +2695,7 @@ static int drbd_asb_recover_2p(struct drbd_conf *mdev) __must_hold(local) case ASB_DISCARD_REMOTE: case ASB_CONSENSUS: case ASB_DISCARD_SECONDARY: + case ASB_DISCARD_ZERO_CHG: dev_err(DEV, "Configuration error.\n"); break; case ASB_VIOLENTLY: @@ -2386,13 +2781,15 @@ static int drbd_uuid_compare(struct drbd_conf *mdev, int *rule_nr) __must_hold(l if (mdev->p_uuid[UI_BITMAP] == (u64)0 && mdev->ldev->md.uuid[UI_BITMAP] != (u64)0) { - if (mdev->agreed_pro_version < 91) + if (mdev->tconn->agreed_pro_version < 91) return -1091; if ((mdev->ldev->md.uuid[UI_BITMAP] & ~((u64)1)) == (mdev->p_uuid[UI_HISTORY_START] & ~((u64)1)) && (mdev->ldev->md.uuid[UI_HISTORY_START] & ~((u64)1)) == (mdev->p_uuid[UI_HISTORY_START + 1] & ~((u64)1))) { dev_info(DEV, "was SyncSource, missed the resync finished event, corrected myself:\n"); - drbd_uuid_set_bm(mdev, 0UL); + drbd_uuid_move_history(mdev); + mdev->ldev->md.uuid[UI_HISTORY_START] = mdev->ldev->md.uuid[UI_BITMAP]; + mdev->ldev->md.uuid[UI_BITMAP] = 0; drbd_uuid_dump(mdev, "self", mdev->ldev->md.uuid, mdev->state.disk >= D_NEGOTIATING ? drbd_bm_total_weight(mdev) : 0, 0); @@ -2407,7 +2804,7 @@ static int drbd_uuid_compare(struct drbd_conf *mdev, int *rule_nr) __must_hold(l if (mdev->ldev->md.uuid[UI_BITMAP] == (u64)0 && mdev->p_uuid[UI_BITMAP] != (u64)0) { - if (mdev->agreed_pro_version < 91) + if (mdev->tconn->agreed_pro_version < 91) return -1091; if ((mdev->ldev->md.uuid[UI_HISTORY_START] & ~((u64)1)) == (mdev->p_uuid[UI_BITMAP] & ~((u64)1)) && @@ -2440,7 +2837,7 @@ static int drbd_uuid_compare(struct drbd_conf *mdev, int *rule_nr) __must_hold(l case 1: /* self_pri && !peer_pri */ return 1; case 2: /* !self_pri && peer_pri */ return -1; case 3: /* self_pri && peer_pri */ - dc = test_bit(DISCARD_CONCURRENT, &mdev->flags); + dc = test_bit(RESOLVE_CONFLICTS, &mdev->tconn->flags); return dc ? -1 : 1; } } @@ -2453,14 +2850,14 @@ static int drbd_uuid_compare(struct drbd_conf *mdev, int *rule_nr) __must_hold(l *rule_nr = 51; peer = mdev->p_uuid[UI_HISTORY_START] & ~((u64)1); if (self == peer) { - if (mdev->agreed_pro_version < 96 ? + if (mdev->tconn->agreed_pro_version < 96 ? (mdev->ldev->md.uuid[UI_HISTORY_START] & ~((u64)1)) == (mdev->p_uuid[UI_HISTORY_START + 1] & ~((u64)1)) : peer + UUID_NEW_BM_OFFSET == (mdev->p_uuid[UI_BITMAP] & ~((u64)1))) { /* The last P_SYNC_UUID did not get though. Undo the last start of resync as sync source modifications of the peer's UUIDs. */ - if (mdev->agreed_pro_version < 91) + if (mdev->tconn->agreed_pro_version < 91) return -1091; mdev->p_uuid[UI_BITMAP] = mdev->p_uuid[UI_HISTORY_START]; @@ -2490,18 +2887,18 @@ static int drbd_uuid_compare(struct drbd_conf *mdev, int *rule_nr) __must_hold(l *rule_nr = 71; self = mdev->ldev->md.uuid[UI_HISTORY_START] & ~((u64)1); if (self == peer) { - if (mdev->agreed_pro_version < 96 ? + if (mdev->tconn->agreed_pro_version < 96 ? (mdev->ldev->md.uuid[UI_HISTORY_START + 1] & ~((u64)1)) == (mdev->p_uuid[UI_HISTORY_START] & ~((u64)1)) : self + UUID_NEW_BM_OFFSET == (mdev->ldev->md.uuid[UI_BITMAP] & ~((u64)1))) { /* The last P_SYNC_UUID did not get though. Undo the last start of resync as sync source modifications of our UUIDs. */ - if (mdev->agreed_pro_version < 91) + if (mdev->tconn->agreed_pro_version < 91) return -1091; - _drbd_uuid_set(mdev, UI_BITMAP, mdev->ldev->md.uuid[UI_HISTORY_START]); - _drbd_uuid_set(mdev, UI_HISTORY_START, mdev->ldev->md.uuid[UI_HISTORY_START + 1]); + __drbd_uuid_set(mdev, UI_BITMAP, mdev->ldev->md.uuid[UI_HISTORY_START]); + __drbd_uuid_set(mdev, UI_HISTORY_START, mdev->ldev->md.uuid[UI_HISTORY_START + 1]); dev_info(DEV, "Last syncUUID did not get through, corrected:\n"); drbd_uuid_dump(mdev, "self", mdev->ldev->md.uuid, @@ -2545,20 +2942,24 @@ static int drbd_uuid_compare(struct drbd_conf *mdev, int *rule_nr) __must_hold(l static enum drbd_conns drbd_sync_handshake(struct drbd_conf *mdev, enum drbd_role peer_role, enum drbd_disk_state peer_disk) __must_hold(local) { - int hg, rule_nr; enum drbd_conns rv = C_MASK; enum drbd_disk_state mydisk; + struct net_conf *nc; + int hg, rule_nr, rr_conflict, tentative; mydisk = mdev->state.disk; if (mydisk == D_NEGOTIATING) mydisk = mdev->new_state_tmp.disk; dev_info(DEV, "drbd_sync_handshake:\n"); + + spin_lock_irq(&mdev->ldev->md.uuid_lock); drbd_uuid_dump(mdev, "self", mdev->ldev->md.uuid, mdev->comm_bm_set, 0); drbd_uuid_dump(mdev, "peer", mdev->p_uuid, mdev->p_uuid[UI_SIZE], mdev->p_uuid[UI_FLAGS]); hg = drbd_uuid_compare(mdev, &rule_nr); + spin_unlock_irq(&mdev->ldev->md.uuid_lock); dev_info(DEV, "uuid_compare()=%d by rule %d\n", hg, rule_nr); @@ -2584,7 +2985,10 @@ static enum drbd_conns drbd_sync_handshake(struct drbd_conf *mdev, enum drbd_rol if (abs(hg) == 100) drbd_khelper(mdev, "initial-split-brain"); - if (hg == 100 || (hg == -100 && mdev->net_conf->always_asbp)) { + rcu_read_lock(); + nc = rcu_dereference(mdev->tconn->net_conf); + + if (hg == 100 || (hg == -100 && nc->always_asbp)) { int pcount = (mdev->state.role == R_PRIMARY) + (peer_role == R_PRIMARY); int forced = (hg == -100); @@ -2613,9 +3017,9 @@ static enum drbd_conns drbd_sync_handshake(struct drbd_conf *mdev, enum drbd_rol } if (hg == -100) { - if (mdev->net_conf->want_lose && !(mdev->p_uuid[UI_FLAGS]&1)) + if (test_bit(DISCARD_MY_DATA, &mdev->flags) && !(mdev->p_uuid[UI_FLAGS]&1)) hg = -1; - if (!mdev->net_conf->want_lose && (mdev->p_uuid[UI_FLAGS]&1)) + if (!test_bit(DISCARD_MY_DATA, &mdev->flags) && (mdev->p_uuid[UI_FLAGS]&1)) hg = 1; if (abs(hg) < 100) @@ -2623,6 +3027,9 @@ static enum drbd_conns drbd_sync_handshake(struct drbd_conf *mdev, enum drbd_rol "Sync from %s node\n", (hg < 0) ? "peer" : "this"); } + rr_conflict = nc->rr_conflict; + tentative = nc->tentative; + rcu_read_unlock(); if (hg == -100) { /* FIXME this log message is not correct if we end up here @@ -2641,7 +3048,7 @@ static enum drbd_conns drbd_sync_handshake(struct drbd_conf *mdev, enum drbd_rol if (hg < 0 && /* by intention we do not use mydisk here. */ mdev->state.role == R_PRIMARY && mdev->state.disk >= D_CONSISTENT) { - switch (mdev->net_conf->rr_conflict) { + switch (rr_conflict) { case ASB_CALL_HELPER: drbd_khelper(mdev, "pri-lost"); /* fall through */ @@ -2654,7 +3061,7 @@ static enum drbd_conns drbd_sync_handshake(struct drbd_conf *mdev, enum drbd_rol } } - if (mdev->net_conf->dry_run || test_bit(CONN_DRY_RUN, &mdev->flags)) { + if (tentative || test_bit(CONN_DRY_RUN, &mdev->tconn->flags)) { if (hg == 0) dev_info(DEV, "dry-run connect: No resync, would become Connected immediately.\n"); else @@ -2686,33 +3093,29 @@ static enum drbd_conns drbd_sync_handshake(struct drbd_conf *mdev, enum drbd_rol return rv; } -/* returns 1 if invalid */ -static int cmp_after_sb(enum drbd_after_sb_p peer, enum drbd_after_sb_p self) +static enum drbd_after_sb_p convert_after_sb(enum drbd_after_sb_p peer) { /* ASB_DISCARD_REMOTE - ASB_DISCARD_LOCAL is valid */ - if ((peer == ASB_DISCARD_REMOTE && self == ASB_DISCARD_LOCAL) || - (self == ASB_DISCARD_REMOTE && peer == ASB_DISCARD_LOCAL)) - return 0; + if (peer == ASB_DISCARD_REMOTE) + return ASB_DISCARD_LOCAL; /* any other things with ASB_DISCARD_REMOTE or ASB_DISCARD_LOCAL are invalid */ - if (peer == ASB_DISCARD_REMOTE || peer == ASB_DISCARD_LOCAL || - self == ASB_DISCARD_REMOTE || self == ASB_DISCARD_LOCAL) - return 1; + if (peer == ASB_DISCARD_LOCAL) + return ASB_DISCARD_REMOTE; /* everything else is valid if they are equal on both sides. */ - if (peer == self) - return 0; - - /* everything es is invalid. */ - return 1; + return peer; } -static int receive_protocol(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned int data_size) +static int receive_protocol(struct drbd_tconn *tconn, struct packet_info *pi) { - struct p_protocol *p = &mdev->data.rbuf.protocol; - int p_proto, p_after_sb_0p, p_after_sb_1p, p_after_sb_2p; - int p_want_lose, p_two_primaries, cf; - char p_integrity_alg[SHARED_SECRET_MAX] = ""; + struct p_protocol *p = pi->data; + enum drbd_after_sb_p p_after_sb_0p, p_after_sb_1p, p_after_sb_2p; + int p_proto, p_discard_my_data, p_two_primaries, cf; + struct net_conf *nc, *old_net_conf, *new_net_conf = NULL; + char integrity_alg[SHARED_SECRET_MAX] = ""; + struct crypto_hash *peer_integrity_tfm = NULL; + void *int_dig_in = NULL, *int_dig_vv = NULL; p_proto = be32_to_cpu(p->protocol); p_after_sb_0p = be32_to_cpu(p->after_sb_0p); @@ -2720,63 +3123,138 @@ static int receive_protocol(struct drbd_conf *mdev, enum drbd_packets cmd, unsig p_after_sb_2p = be32_to_cpu(p->after_sb_2p); p_two_primaries = be32_to_cpu(p->two_primaries); cf = be32_to_cpu(p->conn_flags); - p_want_lose = cf & CF_WANT_LOSE; - - clear_bit(CONN_DRY_RUN, &mdev->flags); + p_discard_my_data = cf & CF_DISCARD_MY_DATA; - if (cf & CF_DRY_RUN) - set_bit(CONN_DRY_RUN, &mdev->flags); + if (tconn->agreed_pro_version >= 87) { + int err; - if (p_proto != mdev->net_conf->wire_protocol) { - dev_err(DEV, "incompatible communication protocols\n"); - goto disconnect; + if (pi->size > sizeof(integrity_alg)) + return -EIO; + err = drbd_recv_all(tconn, integrity_alg, pi->size); + if (err) + return err; + integrity_alg[SHARED_SECRET_MAX - 1] = 0; } - if (cmp_after_sb(p_after_sb_0p, mdev->net_conf->after_sb_0p)) { - dev_err(DEV, "incompatible after-sb-0pri settings\n"); - goto disconnect; - } + if (pi->cmd != P_PROTOCOL_UPDATE) { + clear_bit(CONN_DRY_RUN, &tconn->flags); - if (cmp_after_sb(p_after_sb_1p, mdev->net_conf->after_sb_1p)) { - dev_err(DEV, "incompatible after-sb-1pri settings\n"); - goto disconnect; - } + if (cf & CF_DRY_RUN) + set_bit(CONN_DRY_RUN, &tconn->flags); - if (cmp_after_sb(p_after_sb_2p, mdev->net_conf->after_sb_2p)) { - dev_err(DEV, "incompatible after-sb-2pri settings\n"); - goto disconnect; - } + rcu_read_lock(); + nc = rcu_dereference(tconn->net_conf); - if (p_want_lose && mdev->net_conf->want_lose) { - dev_err(DEV, "both sides have the 'want_lose' flag set\n"); - goto disconnect; - } + if (p_proto != nc->wire_protocol) { + conn_err(tconn, "incompatible %s settings\n", "protocol"); + goto disconnect_rcu_unlock; + } - if (p_two_primaries != mdev->net_conf->two_primaries) { - dev_err(DEV, "incompatible setting of the two-primaries options\n"); - goto disconnect; + if (convert_after_sb(p_after_sb_0p) != nc->after_sb_0p) { + conn_err(tconn, "incompatible %s settings\n", "after-sb-0pri"); + goto disconnect_rcu_unlock; + } + + if (convert_after_sb(p_after_sb_1p) != nc->after_sb_1p) { + conn_err(tconn, "incompatible %s settings\n", "after-sb-1pri"); + goto disconnect_rcu_unlock; + } + + if (convert_after_sb(p_after_sb_2p) != nc->after_sb_2p) { + conn_err(tconn, "incompatible %s settings\n", "after-sb-2pri"); + goto disconnect_rcu_unlock; + } + + if (p_discard_my_data && nc->discard_my_data) { + conn_err(tconn, "incompatible %s settings\n", "discard-my-data"); + goto disconnect_rcu_unlock; + } + + if (p_two_primaries != nc->two_primaries) { + conn_err(tconn, "incompatible %s settings\n", "allow-two-primaries"); + goto disconnect_rcu_unlock; + } + + if (strcmp(integrity_alg, nc->integrity_alg)) { + conn_err(tconn, "incompatible %s settings\n", "data-integrity-alg"); + goto disconnect_rcu_unlock; + } + + rcu_read_unlock(); } - if (mdev->agreed_pro_version >= 87) { - unsigned char *my_alg = mdev->net_conf->integrity_alg; + if (integrity_alg[0]) { + int hash_size; - if (drbd_recv(mdev, p_integrity_alg, data_size) != data_size) - return false; + /* + * We can only change the peer data integrity algorithm + * here. Changing our own data integrity algorithm + * requires that we send a P_PROTOCOL_UPDATE packet at + * the same time; otherwise, the peer has no way to + * tell between which packets the algorithm should + * change. + */ - p_integrity_alg[SHARED_SECRET_MAX-1] = 0; - if (strcmp(p_integrity_alg, my_alg)) { - dev_err(DEV, "incompatible setting of the data-integrity-alg\n"); + peer_integrity_tfm = crypto_alloc_hash(integrity_alg, 0, CRYPTO_ALG_ASYNC); + if (!peer_integrity_tfm) { + conn_err(tconn, "peer data-integrity-alg %s not supported\n", + integrity_alg); goto disconnect; } - dev_info(DEV, "data-integrity-alg: %s\n", - my_alg[0] ? my_alg : (unsigned char *)"<not-used>"); + + hash_size = crypto_hash_digestsize(peer_integrity_tfm); + int_dig_in = kmalloc(hash_size, GFP_KERNEL); + int_dig_vv = kmalloc(hash_size, GFP_KERNEL); + if (!(int_dig_in && int_dig_vv)) { + conn_err(tconn, "Allocation of buffers for data integrity checking failed\n"); + goto disconnect; + } + } + + new_net_conf = kmalloc(sizeof(struct net_conf), GFP_KERNEL); + if (!new_net_conf) { + conn_err(tconn, "Allocation of new net_conf failed\n"); + goto disconnect; } - return true; + mutex_lock(&tconn->data.mutex); + mutex_lock(&tconn->conf_update); + old_net_conf = tconn->net_conf; + *new_net_conf = *old_net_conf; + + new_net_conf->wire_protocol = p_proto; + new_net_conf->after_sb_0p = convert_after_sb(p_after_sb_0p); + new_net_conf->after_sb_1p = convert_after_sb(p_after_sb_1p); + new_net_conf->after_sb_2p = convert_after_sb(p_after_sb_2p); + new_net_conf->two_primaries = p_two_primaries; + rcu_assign_pointer(tconn->net_conf, new_net_conf); + mutex_unlock(&tconn->conf_update); + mutex_unlock(&tconn->data.mutex); + + crypto_free_hash(tconn->peer_integrity_tfm); + kfree(tconn->int_dig_in); + kfree(tconn->int_dig_vv); + tconn->peer_integrity_tfm = peer_integrity_tfm; + tconn->int_dig_in = int_dig_in; + tconn->int_dig_vv = int_dig_vv; + + if (strcmp(old_net_conf->integrity_alg, integrity_alg)) + conn_info(tconn, "peer data-integrity-alg: %s\n", + integrity_alg[0] ? integrity_alg : "(none)"); + + synchronize_rcu(); + kfree(old_net_conf); + return 0; + +disconnect_rcu_unlock: + rcu_read_unlock(); disconnect: - drbd_force_state(mdev, NS(conn, C_DISCONNECTING)); - return false; + crypto_free_hash(peer_integrity_tfm); + kfree(int_dig_in); + kfree(int_dig_vv); + conn_request_state(tconn, NS(conn, C_DISCONNECTING), CS_HARD); + return -EIO; } /* helper function @@ -2798,24 +3276,64 @@ struct crypto_hash *drbd_crypto_alloc_digest_safe(const struct drbd_conf *mdev, alg, name, PTR_ERR(tfm)); return tfm; } - if (!drbd_crypto_is_hash(crypto_hash_tfm(tfm))) { - crypto_free_hash(tfm); - dev_err(DEV, "\"%s\" is not a digest (%s)\n", alg, name); - return ERR_PTR(-EINVAL); - } return tfm; } -static int receive_SyncParam(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned int packet_size) +static int ignore_remaining_packet(struct drbd_tconn *tconn, struct packet_info *pi) +{ + void *buffer = tconn->data.rbuf; + int size = pi->size; + + while (size) { + int s = min_t(int, size, DRBD_SOCKET_BUFFER_SIZE); + s = drbd_recv(tconn, buffer, s); + if (s <= 0) { + if (s < 0) + return s; + break; + } + size -= s; + } + if (size) + return -EIO; + return 0; +} + +/* + * config_unknown_volume - device configuration command for unknown volume + * + * When a device is added to an existing connection, the node on which the + * device is added first will send configuration commands to its peer but the + * peer will not know about the device yet. It will warn and ignore these + * commands. Once the device is added on the second node, the second node will + * send the same device configuration commands, but in the other direction. + * + * (We can also end up here if drbd is misconfigured.) + */ +static int config_unknown_volume(struct drbd_tconn *tconn, struct packet_info *pi) { - int ok = true; - struct p_rs_param_95 *p = &mdev->data.rbuf.rs_param_95; + conn_warn(tconn, "%s packet received for volume %u, which is not configured locally\n", + cmdname(pi->cmd), pi->vnr); + return ignore_remaining_packet(tconn, pi); +} + +static int receive_SyncParam(struct drbd_tconn *tconn, struct packet_info *pi) +{ + struct drbd_conf *mdev; + struct p_rs_param_95 *p; unsigned int header_size, data_size, exp_max_sz; struct crypto_hash *verify_tfm = NULL; struct crypto_hash *csums_tfm = NULL; - const int apv = mdev->agreed_pro_version; - int *rs_plan_s = NULL; + struct net_conf *old_net_conf, *new_net_conf = NULL; + struct disk_conf *old_disk_conf = NULL, *new_disk_conf = NULL; + const int apv = tconn->agreed_pro_version; + struct fifo_buffer *old_plan = NULL, *new_plan = NULL; int fifo_size = 0; + int err; + + mdev = vnr_to_mdev(tconn, pi->vnr); + if (!mdev) + return config_unknown_volume(tconn, pi); exp_max_sz = apv <= 87 ? sizeof(struct p_rs_param) : apv == 88 ? sizeof(struct p_rs_param) @@ -2823,32 +3341,49 @@ static int receive_SyncParam(struct drbd_conf *mdev, enum drbd_packets cmd, unsi : apv <= 94 ? sizeof(struct p_rs_param_89) : /* apv >= 95 */ sizeof(struct p_rs_param_95); - if (packet_size > exp_max_sz) { + if (pi->size > exp_max_sz) { dev_err(DEV, "SyncParam packet too long: received %u, expected <= %u bytes\n", - packet_size, exp_max_sz); - return false; + pi->size, exp_max_sz); + return -EIO; } if (apv <= 88) { - header_size = sizeof(struct p_rs_param) - sizeof(struct p_header80); - data_size = packet_size - header_size; + header_size = sizeof(struct p_rs_param); + data_size = pi->size - header_size; } else if (apv <= 94) { - header_size = sizeof(struct p_rs_param_89) - sizeof(struct p_header80); - data_size = packet_size - header_size; + header_size = sizeof(struct p_rs_param_89); + data_size = pi->size - header_size; D_ASSERT(data_size == 0); } else { - header_size = sizeof(struct p_rs_param_95) - sizeof(struct p_header80); - data_size = packet_size - header_size; + header_size = sizeof(struct p_rs_param_95); + data_size = pi->size - header_size; D_ASSERT(data_size == 0); } /* initialize verify_alg and csums_alg */ + p = pi->data; memset(p->verify_alg, 0, 2 * SHARED_SECRET_MAX); - if (drbd_recv(mdev, &p->head.payload, header_size) != header_size) - return false; + err = drbd_recv_all(mdev->tconn, p, header_size); + if (err) + return err; - mdev->sync_conf.rate = be32_to_cpu(p->rate); + mutex_lock(&mdev->tconn->conf_update); + old_net_conf = mdev->tconn->net_conf; + if (get_ldev(mdev)) { + new_disk_conf = kzalloc(sizeof(struct disk_conf), GFP_KERNEL); + if (!new_disk_conf) { + put_ldev(mdev); + mutex_unlock(&mdev->tconn->conf_update); + dev_err(DEV, "Allocation of new disk_conf failed\n"); + return -ENOMEM; + } + + old_disk_conf = mdev->ldev->disk_conf; + *new_disk_conf = *old_disk_conf; + + new_disk_conf->resync_rate = be32_to_cpu(p->resync_rate); + } if (apv >= 88) { if (apv == 88) { @@ -2856,12 +3391,13 @@ static int receive_SyncParam(struct drbd_conf *mdev, enum drbd_packets cmd, unsi dev_err(DEV, "verify-alg of wrong size, " "peer wants %u, accepting only up to %u byte\n", data_size, SHARED_SECRET_MAX); - return false; + err = -EIO; + goto reconnect; } - if (drbd_recv(mdev, p->verify_alg, data_size) != data_size) - return false; - + err = drbd_recv_all(mdev->tconn, p->verify_alg, data_size); + if (err) + goto reconnect; /* we expect NUL terminated string */ /* but just in case someone tries to be evil */ D_ASSERT(p->verify_alg[data_size-1] == 0); @@ -2876,10 +3412,10 @@ static int receive_SyncParam(struct drbd_conf *mdev, enum drbd_packets cmd, unsi p->csums_alg[SHARED_SECRET_MAX-1] = 0; } - if (strcmp(mdev->sync_conf.verify_alg, p->verify_alg)) { + if (strcmp(old_net_conf->verify_alg, p->verify_alg)) { if (mdev->state.conn == C_WF_REPORT_PARAMS) { dev_err(DEV, "Different verify-alg settings. me=\"%s\" peer=\"%s\"\n", - mdev->sync_conf.verify_alg, p->verify_alg); + old_net_conf->verify_alg, p->verify_alg); goto disconnect; } verify_tfm = drbd_crypto_alloc_digest_safe(mdev, @@ -2890,10 +3426,10 @@ static int receive_SyncParam(struct drbd_conf *mdev, enum drbd_packets cmd, unsi } } - if (apv >= 89 && strcmp(mdev->sync_conf.csums_alg, p->csums_alg)) { + if (apv >= 89 && strcmp(old_net_conf->csums_alg, p->csums_alg)) { if (mdev->state.conn == C_WF_REPORT_PARAMS) { dev_err(DEV, "Different csums-alg settings. me=\"%s\" peer=\"%s\"\n", - mdev->sync_conf.csums_alg, p->csums_alg); + old_net_conf->csums_alg, p->csums_alg); goto disconnect; } csums_tfm = drbd_crypto_alloc_digest_safe(mdev, @@ -2904,57 +3440,91 @@ static int receive_SyncParam(struct drbd_conf *mdev, enum drbd_packets cmd, unsi } } - if (apv > 94) { - mdev->sync_conf.rate = be32_to_cpu(p->rate); - mdev->sync_conf.c_plan_ahead = be32_to_cpu(p->c_plan_ahead); - mdev->sync_conf.c_delay_target = be32_to_cpu(p->c_delay_target); - mdev->sync_conf.c_fill_target = be32_to_cpu(p->c_fill_target); - mdev->sync_conf.c_max_rate = be32_to_cpu(p->c_max_rate); - - fifo_size = (mdev->sync_conf.c_plan_ahead * 10 * SLEEP_TIME) / HZ; - if (fifo_size != mdev->rs_plan_s.size && fifo_size > 0) { - rs_plan_s = kzalloc(sizeof(int) * fifo_size, GFP_KERNEL); - if (!rs_plan_s) { + if (apv > 94 && new_disk_conf) { + new_disk_conf->c_plan_ahead = be32_to_cpu(p->c_plan_ahead); + new_disk_conf->c_delay_target = be32_to_cpu(p->c_delay_target); + new_disk_conf->c_fill_target = be32_to_cpu(p->c_fill_target); + new_disk_conf->c_max_rate = be32_to_cpu(p->c_max_rate); + + fifo_size = (new_disk_conf->c_plan_ahead * 10 * SLEEP_TIME) / HZ; + if (fifo_size != mdev->rs_plan_s->size) { + new_plan = fifo_alloc(fifo_size); + if (!new_plan) { dev_err(DEV, "kmalloc of fifo_buffer failed"); + put_ldev(mdev); goto disconnect; } } } - spin_lock(&mdev->peer_seq_lock); - /* lock against drbd_nl_syncer_conf() */ - if (verify_tfm) { - strcpy(mdev->sync_conf.verify_alg, p->verify_alg); - mdev->sync_conf.verify_alg_len = strlen(p->verify_alg) + 1; - crypto_free_hash(mdev->verify_tfm); - mdev->verify_tfm = verify_tfm; - dev_info(DEV, "using verify-alg: \"%s\"\n", p->verify_alg); - } - if (csums_tfm) { - strcpy(mdev->sync_conf.csums_alg, p->csums_alg); - mdev->sync_conf.csums_alg_len = strlen(p->csums_alg) + 1; - crypto_free_hash(mdev->csums_tfm); - mdev->csums_tfm = csums_tfm; - dev_info(DEV, "using csums-alg: \"%s\"\n", p->csums_alg); - } - if (fifo_size != mdev->rs_plan_s.size) { - kfree(mdev->rs_plan_s.values); - mdev->rs_plan_s.values = rs_plan_s; - mdev->rs_plan_s.size = fifo_size; - mdev->rs_planed = 0; + if (verify_tfm || csums_tfm) { + new_net_conf = kzalloc(sizeof(struct net_conf), GFP_KERNEL); + if (!new_net_conf) { + dev_err(DEV, "Allocation of new net_conf failed\n"); + goto disconnect; + } + + *new_net_conf = *old_net_conf; + + if (verify_tfm) { + strcpy(new_net_conf->verify_alg, p->verify_alg); + new_net_conf->verify_alg_len = strlen(p->verify_alg) + 1; + crypto_free_hash(mdev->tconn->verify_tfm); + mdev->tconn->verify_tfm = verify_tfm; + dev_info(DEV, "using verify-alg: \"%s\"\n", p->verify_alg); + } + if (csums_tfm) { + strcpy(new_net_conf->csums_alg, p->csums_alg); + new_net_conf->csums_alg_len = strlen(p->csums_alg) + 1; + crypto_free_hash(mdev->tconn->csums_tfm); + mdev->tconn->csums_tfm = csums_tfm; + dev_info(DEV, "using csums-alg: \"%s\"\n", p->csums_alg); + } + rcu_assign_pointer(tconn->net_conf, new_net_conf); } - spin_unlock(&mdev->peer_seq_lock); } - return ok; + if (new_disk_conf) { + rcu_assign_pointer(mdev->ldev->disk_conf, new_disk_conf); + put_ldev(mdev); + } + + if (new_plan) { + old_plan = mdev->rs_plan_s; + rcu_assign_pointer(mdev->rs_plan_s, new_plan); + } + + mutex_unlock(&mdev->tconn->conf_update); + synchronize_rcu(); + if (new_net_conf) + kfree(old_net_conf); + kfree(old_disk_conf); + kfree(old_plan); + + return 0; + +reconnect: + if (new_disk_conf) { + put_ldev(mdev); + kfree(new_disk_conf); + } + mutex_unlock(&mdev->tconn->conf_update); + return -EIO; + disconnect: + kfree(new_plan); + if (new_disk_conf) { + put_ldev(mdev); + kfree(new_disk_conf); + } + mutex_unlock(&mdev->tconn->conf_update); /* just for completeness: actually not needed, * as this is not reached if csums_tfm was ok. */ crypto_free_hash(csums_tfm); /* but free the verify_tfm again, if csums_tfm did not work out */ crypto_free_hash(verify_tfm); - drbd_force_state(mdev, NS(conn, C_DISCONNECTING)); - return false; + conn_request_state(mdev->tconn, NS(conn, C_DISCONNECTING), CS_HARD); + return -EIO; } /* warn if the arguments differ by more than 12.5% */ @@ -2970,59 +3540,77 @@ static void warn_if_differ_considerably(struct drbd_conf *mdev, (unsigned long long)a, (unsigned long long)b); } -static int receive_sizes(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned int data_size) +static int receive_sizes(struct drbd_tconn *tconn, struct packet_info *pi) { - struct p_sizes *p = &mdev->data.rbuf.sizes; + struct drbd_conf *mdev; + struct p_sizes *p = pi->data; enum determine_dev_size dd = unchanged; sector_t p_size, p_usize, my_usize; int ldsc = 0; /* local disk size changed */ enum dds_flags ddsf; + mdev = vnr_to_mdev(tconn, pi->vnr); + if (!mdev) + return config_unknown_volume(tconn, pi); + p_size = be64_to_cpu(p->d_size); p_usize = be64_to_cpu(p->u_size); - if (p_size == 0 && mdev->state.disk == D_DISKLESS) { - dev_err(DEV, "some backing storage is needed\n"); - drbd_force_state(mdev, NS(conn, C_DISCONNECTING)); - return false; - } - /* just store the peer's disk size for now. * we still need to figure out whether we accept that. */ mdev->p_size = p_size; if (get_ldev(mdev)) { + rcu_read_lock(); + my_usize = rcu_dereference(mdev->ldev->disk_conf)->disk_size; + rcu_read_unlock(); + warn_if_differ_considerably(mdev, "lower level device sizes", p_size, drbd_get_max_capacity(mdev->ldev)); warn_if_differ_considerably(mdev, "user requested size", - p_usize, mdev->ldev->dc.disk_size); + p_usize, my_usize); /* if this is the first connect, or an otherwise expected * param exchange, choose the minimum */ if (mdev->state.conn == C_WF_REPORT_PARAMS) - p_usize = min_not_zero((sector_t)mdev->ldev->dc.disk_size, - p_usize); - - my_usize = mdev->ldev->dc.disk_size; - - if (mdev->ldev->dc.disk_size != p_usize) { - mdev->ldev->dc.disk_size = p_usize; - dev_info(DEV, "Peer sets u_size to %lu sectors\n", - (unsigned long)mdev->ldev->dc.disk_size); - } + p_usize = min_not_zero(my_usize, p_usize); /* Never shrink a device with usable data during connect. But allow online shrinking if we are connected. */ - if (drbd_new_dev_size(mdev, mdev->ldev, 0) < - drbd_get_capacity(mdev->this_bdev) && - mdev->state.disk >= D_OUTDATED && - mdev->state.conn < C_CONNECTED) { + if (drbd_new_dev_size(mdev, mdev->ldev, p_usize, 0) < + drbd_get_capacity(mdev->this_bdev) && + mdev->state.disk >= D_OUTDATED && + mdev->state.conn < C_CONNECTED) { dev_err(DEV, "The peer's disk size is too small!\n"); - drbd_force_state(mdev, NS(conn, C_DISCONNECTING)); - mdev->ldev->dc.disk_size = my_usize; + conn_request_state(mdev->tconn, NS(conn, C_DISCONNECTING), CS_HARD); put_ldev(mdev); - return false; + return -EIO; + } + + if (my_usize != p_usize) { + struct disk_conf *old_disk_conf, *new_disk_conf = NULL; + + new_disk_conf = kzalloc(sizeof(struct disk_conf), GFP_KERNEL); + if (!new_disk_conf) { + dev_err(DEV, "Allocation of new disk_conf failed\n"); + put_ldev(mdev); + return -ENOMEM; + } + + mutex_lock(&mdev->tconn->conf_update); + old_disk_conf = mdev->ldev->disk_conf; + *new_disk_conf = *old_disk_conf; + new_disk_conf->disk_size = p_usize; + + rcu_assign_pointer(mdev->ldev->disk_conf, new_disk_conf); + mutex_unlock(&mdev->tconn->conf_update); + synchronize_rcu(); + kfree(old_disk_conf); + + dev_info(DEV, "Peer sets u_size to %lu sectors\n", + (unsigned long)my_usize); } + put_ldev(mdev); } @@ -3031,7 +3619,7 @@ static int receive_sizes(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned dd = drbd_determine_dev_size(mdev, ddsf); put_ldev(mdev); if (dd == dev_size_error) - return false; + return -EIO; drbd_md_sync(mdev); } else { /* I am diskless, need to accept the peer's size. */ @@ -3070,16 +3658,25 @@ static int receive_sizes(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned } } - return true; + return 0; } -static int receive_uuids(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned int data_size) +static int receive_uuids(struct drbd_tconn *tconn, struct packet_info *pi) { - struct p_uuids *p = &mdev->data.rbuf.uuids; + struct drbd_conf *mdev; + struct p_uuids *p = pi->data; u64 *p_uuid; int i, updated_uuids = 0; + mdev = vnr_to_mdev(tconn, pi->vnr); + if (!mdev) + return config_unknown_volume(tconn, pi); + p_uuid = kmalloc(sizeof(u64)*UI_EXTENDED_SIZE, GFP_NOIO); + if (!p_uuid) { + dev_err(DEV, "kmalloc of p_uuid failed\n"); + return false; + } for (i = UI_CURRENT; i < UI_EXTENDED_SIZE; i++) p_uuid[i] = be64_to_cpu(p->uuid[i]); @@ -3093,14 +3690,14 @@ static int receive_uuids(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned (mdev->ed_uuid & ~((u64)1)) != (p_uuid[UI_CURRENT] & ~((u64)1))) { dev_err(DEV, "Can only connect to data with current UUID=%016llX\n", (unsigned long long)mdev->ed_uuid); - drbd_force_state(mdev, NS(conn, C_DISCONNECTING)); - return false; + conn_request_state(mdev->tconn, NS(conn, C_DISCONNECTING), CS_HARD); + return -EIO; } if (get_ldev(mdev)) { int skip_initial_sync = mdev->state.conn == C_CONNECTED && - mdev->agreed_pro_version >= 90 && + mdev->tconn->agreed_pro_version >= 90 && mdev->ldev->md.uuid[UI_CURRENT] == UUID_JUST_CREATED && (p_uuid[UI_FLAGS] & 8); if (skip_initial_sync) { @@ -3127,14 +3724,15 @@ static int receive_uuids(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned ongoing cluster wide state change is finished. That is important if we are primary and are detaching from our disk. We need to see the new disk state... */ - wait_event(mdev->misc_wait, !test_bit(CLUSTER_ST_CHANGE, &mdev->flags)); + mutex_lock(mdev->state_mutex); + mutex_unlock(mdev->state_mutex); if (mdev->state.conn >= C_CONNECTED && mdev->state.disk < D_INCONSISTENT) updated_uuids |= drbd_set_ed_uuid(mdev, p_uuid[UI_CURRENT]); if (updated_uuids) drbd_print_uuids(mdev, "receiver updated UUIDs to"); - return true; + return 0; } /** @@ -3146,6 +3744,7 @@ static union drbd_state convert_state(union drbd_state ps) union drbd_state ms; static enum drbd_conns c_tab[] = { + [C_WF_REPORT_PARAMS] = C_WF_REPORT_PARAMS, [C_CONNECTED] = C_CONNECTED, [C_STARTING_SYNC_S] = C_STARTING_SYNC_T, @@ -3167,40 +3766,74 @@ static union drbd_state convert_state(union drbd_state ps) return ms; } -static int receive_req_state(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned int data_size) +static int receive_req_state(struct drbd_tconn *tconn, struct packet_info *pi) { - struct p_req_state *p = &mdev->data.rbuf.req_state; + struct drbd_conf *mdev; + struct p_req_state *p = pi->data; union drbd_state mask, val; enum drbd_state_rv rv; + mdev = vnr_to_mdev(tconn, pi->vnr); + if (!mdev) + return -EIO; + mask.i = be32_to_cpu(p->mask); val.i = be32_to_cpu(p->val); - if (test_bit(DISCARD_CONCURRENT, &mdev->flags) && - test_bit(CLUSTER_ST_CHANGE, &mdev->flags)) { + if (test_bit(RESOLVE_CONFLICTS, &mdev->tconn->flags) && + mutex_is_locked(mdev->state_mutex)) { drbd_send_sr_reply(mdev, SS_CONCURRENT_ST_CHG); - return true; + return 0; } mask = convert_state(mask); val = convert_state(val); rv = drbd_change_state(mdev, CS_VERBOSE, mask, val); - drbd_send_sr_reply(mdev, rv); + drbd_md_sync(mdev); - return true; + return 0; } -static int receive_state(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned int data_size) +static int receive_req_conn_state(struct drbd_tconn *tconn, struct packet_info *pi) { - struct p_state *p = &mdev->data.rbuf.state; + struct p_req_state *p = pi->data; + union drbd_state mask, val; + enum drbd_state_rv rv; + + mask.i = be32_to_cpu(p->mask); + val.i = be32_to_cpu(p->val); + + if (test_bit(RESOLVE_CONFLICTS, &tconn->flags) && + mutex_is_locked(&tconn->cstate_mutex)) { + conn_send_sr_reply(tconn, SS_CONCURRENT_ST_CHG); + return 0; + } + + mask = convert_state(mask); + val = convert_state(val); + + rv = conn_request_state(tconn, mask, val, CS_VERBOSE | CS_LOCAL_ONLY | CS_IGN_OUTD_FAIL); + conn_send_sr_reply(tconn, rv); + + return 0; +} + +static int receive_state(struct drbd_tconn *tconn, struct packet_info *pi) +{ + struct drbd_conf *mdev; + struct p_state *p = pi->data; union drbd_state os, ns, peer_state; enum drbd_disk_state real_peer_disk; enum chg_state_flags cs_flags; int rv; + mdev = vnr_to_mdev(tconn, pi->vnr); + if (!mdev) + return config_unknown_volume(tconn, pi); + peer_state.i = be32_to_cpu(p->state); real_peer_disk = peer_state.disk; @@ -3209,16 +3842,16 @@ static int receive_state(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned dev_info(DEV, "real peer disk state = %s\n", drbd_disk_str(real_peer_disk)); } - spin_lock_irq(&mdev->req_lock); + spin_lock_irq(&mdev->tconn->req_lock); retry: - os = ns = mdev->state; - spin_unlock_irq(&mdev->req_lock); + os = ns = drbd_read_state(mdev); + spin_unlock_irq(&mdev->tconn->req_lock); /* If some other part of the code (asender thread, timeout) * already decided to close the connection again, * we must not "re-establish" it here. */ if (os.conn <= C_TEAR_DOWN) - return false; + return -ECONNRESET; /* If this is the "end of sync" confirmation, usually the peer disk * transitions from D_INCONSISTENT to D_UP_TO_DATE. For empty (0 bits @@ -3246,10 +3879,18 @@ static int receive_state(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned peer_state.conn == C_CONNECTED) { if (drbd_bm_total_weight(mdev) <= mdev->rs_failed) drbd_resync_finished(mdev); - return true; + return 0; } } + /* explicit verify finished notification, stop sector reached. */ + if (os.conn == C_VERIFY_T && os.disk == D_UP_TO_DATE && + peer_state.conn == C_CONNECTED && real_peer_disk == D_UP_TO_DATE) { + ov_out_of_sync_print(mdev); + drbd_resync_finished(mdev); + return 0; + } + /* peer says his disk is inconsistent, while we think it is uptodate, * and this happens while the peer still thinks we have a sync going on, * but we think we are already done with the sync. @@ -3298,17 +3939,17 @@ static int receive_state(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned peer_state.disk = D_DISKLESS; real_peer_disk = D_DISKLESS; } else { - if (test_and_clear_bit(CONN_DRY_RUN, &mdev->flags)) - return false; + if (test_and_clear_bit(CONN_DRY_RUN, &mdev->tconn->flags)) + return -EIO; D_ASSERT(os.conn == C_WF_REPORT_PARAMS); - drbd_force_state(mdev, NS(conn, C_DISCONNECTING)); - return false; + conn_request_state(mdev->tconn, NS(conn, C_DISCONNECTING), CS_HARD); + return -EIO; } } } - spin_lock_irq(&mdev->req_lock); - if (mdev->state.i != os.i) + spin_lock_irq(&mdev->tconn->req_lock); + if (os.i != drbd_read_state(mdev).i) goto retry; clear_bit(CONSIDER_RESYNC, &mdev->flags); ns.peer = peer_state.role; @@ -3317,25 +3958,25 @@ static int receive_state(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned if ((ns.conn == C_CONNECTED || ns.conn == C_WF_BITMAP_S) && ns.disk == D_NEGOTIATING) ns.disk = mdev->new_state_tmp.disk; cs_flags = CS_VERBOSE + (os.conn < C_CONNECTED && ns.conn >= C_CONNECTED ? 0 : CS_HARD); - if (ns.pdsk == D_CONSISTENT && is_susp(ns) && ns.conn == C_CONNECTED && os.conn < C_CONNECTED && + if (ns.pdsk == D_CONSISTENT && drbd_suspended(mdev) && ns.conn == C_CONNECTED && os.conn < C_CONNECTED && test_bit(NEW_CUR_UUID, &mdev->flags)) { - /* Do not allow tl_restart(resend) for a rebooted peer. We can only allow this + /* Do not allow tl_restart(RESEND) for a rebooted peer. We can only allow this for temporal network outages! */ - spin_unlock_irq(&mdev->req_lock); + spin_unlock_irq(&mdev->tconn->req_lock); dev_err(DEV, "Aborting Connect, can not thaw IO with an only Consistent peer\n"); - tl_clear(mdev); + tl_clear(mdev->tconn); drbd_uuid_new_current(mdev); clear_bit(NEW_CUR_UUID, &mdev->flags); - drbd_force_state(mdev, NS2(conn, C_PROTOCOL_ERROR, susp, 0)); - return false; + conn_request_state(mdev->tconn, NS2(conn, C_PROTOCOL_ERROR, susp, 0), CS_HARD); + return -EIO; } rv = _drbd_set_state(mdev, ns, cs_flags, NULL); - ns = mdev->state; - spin_unlock_irq(&mdev->req_lock); + ns = drbd_read_state(mdev); + spin_unlock_irq(&mdev->tconn->req_lock); if (rv < SS_SUCCESS) { - drbd_force_state(mdev, NS(conn, C_DISCONNECTING)); - return false; + conn_request_state(mdev->tconn, NS(conn, C_DISCONNECTING), CS_HARD); + return -EIO; } if (os.conn > C_WF_REPORT_PARAMS) { @@ -3349,16 +3990,21 @@ static int receive_state(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned } } - mdev->net_conf->want_lose = 0; + clear_bit(DISCARD_MY_DATA, &mdev->flags); drbd_md_sync(mdev); /* update connected indicator, la_size, ... */ - return true; + return 0; } -static int receive_sync_uuid(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned int data_size) +static int receive_sync_uuid(struct drbd_tconn *tconn, struct packet_info *pi) { - struct p_rs_uuid *p = &mdev->data.rbuf.rs_uuid; + struct drbd_conf *mdev; + struct p_rs_uuid *p = pi->data; + + mdev = vnr_to_mdev(tconn, pi->vnr); + if (!mdev) + return -EIO; wait_event(mdev->misc_wait, mdev->state.conn == C_WF_SYNC_UUID || @@ -3381,7 +4027,7 @@ static int receive_sync_uuid(struct drbd_conf *mdev, enum drbd_packets cmd, unsi } else dev_err(DEV, "Ignoring SyncUUID packet!\n"); - return true; + return 0; } /** @@ -3391,27 +4037,27 @@ static int receive_sync_uuid(struct drbd_conf *mdev, enum drbd_packets cmd, unsi * code upon failure. */ static int -receive_bitmap_plain(struct drbd_conf *mdev, unsigned int data_size, - unsigned long *buffer, struct bm_xfer_ctx *c) +receive_bitmap_plain(struct drbd_conf *mdev, unsigned int size, + unsigned long *p, struct bm_xfer_ctx *c) { - unsigned num_words = min_t(size_t, BM_PACKET_WORDS, c->bm_words - c->word_offset); - unsigned want = num_words * sizeof(long); + unsigned int data_size = DRBD_SOCKET_BUFFER_SIZE - + drbd_header_size(mdev->tconn); + unsigned int num_words = min_t(size_t, data_size / sizeof(*p), + c->bm_words - c->word_offset); + unsigned int want = num_words * sizeof(*p); int err; - if (want != data_size) { - dev_err(DEV, "%s:want (%u) != data_size (%u)\n", __func__, want, data_size); + if (want != size) { + dev_err(DEV, "%s:want (%u) != size (%u)\n", __func__, want, size); return -EIO; } if (want == 0) return 0; - err = drbd_recv(mdev, buffer, want); - if (err != want) { - if (err >= 0) - err = -EIO; + err = drbd_recv_all(mdev->tconn, p, want); + if (err) return err; - } - drbd_bm_merge_lel(mdev, c->word_offset, num_words, buffer); + drbd_bm_merge_lel(mdev, c->word_offset, num_words, p); c->word_offset += num_words; c->bit_offset = c->word_offset * BITS_PER_LONG; @@ -3421,6 +4067,21 @@ receive_bitmap_plain(struct drbd_conf *mdev, unsigned int data_size, return 1; } +static enum drbd_bitmap_code dcbp_get_code(struct p_compressed_bm *p) +{ + return (enum drbd_bitmap_code)(p->encoding & 0x0f); +} + +static int dcbp_get_start(struct p_compressed_bm *p) +{ + return (p->encoding & 0x80) != 0; +} + +static int dcbp_get_pad_bits(struct p_compressed_bm *p) +{ + return (p->encoding >> 4) & 0x7; +} + /** * recv_bm_rle_bits * @@ -3430,7 +4091,8 @@ receive_bitmap_plain(struct drbd_conf *mdev, unsigned int data_size, static int recv_bm_rle_bits(struct drbd_conf *mdev, struct p_compressed_bm *p, - struct bm_xfer_ctx *c) + struct bm_xfer_ctx *c, + unsigned int len) { struct bitstream bs; u64 look_ahead; @@ -3438,12 +4100,11 @@ recv_bm_rle_bits(struct drbd_conf *mdev, u64 tmp; unsigned long s = c->bit_offset; unsigned long e; - int len = be16_to_cpu(p->head.length) - (sizeof(*p) - sizeof(p->head)); - int toggle = DCBP_get_start(p); + int toggle = dcbp_get_start(p); int have; int bits; - bitstream_init(&bs, p->code, len, DCBP_get_pad_bits(p)); + bitstream_init(&bs, p->code, len, dcbp_get_pad_bits(p)); bits = bitstream_get_bits(&bs, &look_ahead, 64); if (bits < 0) @@ -3495,17 +4156,18 @@ recv_bm_rle_bits(struct drbd_conf *mdev, static int decode_bitmap_c(struct drbd_conf *mdev, struct p_compressed_bm *p, - struct bm_xfer_ctx *c) + struct bm_xfer_ctx *c, + unsigned int len) { - if (DCBP_get_code(p) == RLE_VLI_Bits) - return recv_bm_rle_bits(mdev, p, c); + if (dcbp_get_code(p) == RLE_VLI_Bits) + return recv_bm_rle_bits(mdev, p, c, len - sizeof(*p)); /* other variants had been implemented for evaluation, * but have been dropped as this one turned out to be "best" * during all our tests. */ dev_err(DEV, "receive_bitmap_c: unknown encoding %u\n", p->encoding); - drbd_force_state(mdev, NS(conn, C_PROTOCOL_ERROR)); + conn_request_state(mdev->tconn, NS(conn, C_PROTOCOL_ERROR), CS_HARD); return -EIO; } @@ -3513,11 +4175,13 @@ void INFO_bm_xfer_stats(struct drbd_conf *mdev, const char *direction, struct bm_xfer_ctx *c) { /* what would it take to transfer it "plaintext" */ - unsigned plain = sizeof(struct p_header80) * - ((c->bm_words+BM_PACKET_WORDS-1)/BM_PACKET_WORDS+1) - + c->bm_words * sizeof(long); - unsigned total = c->bytes[0] + c->bytes[1]; - unsigned r; + unsigned int header_size = drbd_header_size(mdev->tconn); + unsigned int data_size = DRBD_SOCKET_BUFFER_SIZE - header_size; + unsigned int plain = + header_size * (DIV_ROUND_UP(c->bm_words, data_size) + 1) + + c->bm_words * sizeof(unsigned long); + unsigned int total = c->bytes[0] + c->bytes[1]; + unsigned int r; /* total can not be zero. but just in case: */ if (total == 0) @@ -3551,67 +4215,63 @@ void INFO_bm_xfer_stats(struct drbd_conf *mdev, in order to be agnostic to the 32 vs 64 bits issue. returns 0 on failure, 1 if we successfully received it. */ -static int receive_bitmap(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned int data_size) +static int receive_bitmap(struct drbd_tconn *tconn, struct packet_info *pi) { + struct drbd_conf *mdev; struct bm_xfer_ctx c; - void *buffer; int err; - int ok = false; - struct p_header80 *h = &mdev->data.rbuf.header.h80; + + mdev = vnr_to_mdev(tconn, pi->vnr); + if (!mdev) + return -EIO; drbd_bm_lock(mdev, "receive bitmap", BM_LOCKED_SET_ALLOWED); /* you are supposed to send additional out-of-sync information * if you actually set bits during this phase */ - /* maybe we should use some per thread scratch page, - * and allocate that during initial device creation? */ - buffer = (unsigned long *) __get_free_page(GFP_NOIO); - if (!buffer) { - dev_err(DEV, "failed to allocate one page buffer in %s\n", __func__); - goto out; - } - c = (struct bm_xfer_ctx) { .bm_bits = drbd_bm_bits(mdev), .bm_words = drbd_bm_words(mdev), }; for(;;) { - if (cmd == P_BITMAP) { - err = receive_bitmap_plain(mdev, data_size, buffer, &c); - } else if (cmd == P_COMPRESSED_BITMAP) { + if (pi->cmd == P_BITMAP) + err = receive_bitmap_plain(mdev, pi->size, pi->data, &c); + else if (pi->cmd == P_COMPRESSED_BITMAP) { /* MAYBE: sanity check that we speak proto >= 90, * and the feature is enabled! */ - struct p_compressed_bm *p; + struct p_compressed_bm *p = pi->data; - if (data_size > BM_PACKET_PAYLOAD_BYTES) { + if (pi->size > DRBD_SOCKET_BUFFER_SIZE - drbd_header_size(tconn)) { dev_err(DEV, "ReportCBitmap packet too large\n"); + err = -EIO; goto out; } - /* use the page buff */ - p = buffer; - memcpy(p, h, sizeof(*h)); - if (drbd_recv(mdev, p->head.payload, data_size) != data_size) - goto out; - if (data_size <= (sizeof(*p) - sizeof(p->head))) { - dev_err(DEV, "ReportCBitmap packet too small (l:%u)\n", data_size); + if (pi->size <= sizeof(*p)) { + dev_err(DEV, "ReportCBitmap packet too small (l:%u)\n", pi->size); + err = -EIO; goto out; } - err = decode_bitmap_c(mdev, p, &c); + err = drbd_recv_all(mdev->tconn, p, pi->size); + if (err) + goto out; + err = decode_bitmap_c(mdev, p, &c, pi->size); } else { - dev_warn(DEV, "receive_bitmap: cmd neither ReportBitMap nor ReportCBitMap (is 0x%x)", cmd); + dev_warn(DEV, "receive_bitmap: cmd neither ReportBitMap nor ReportCBitMap (is 0x%x)", pi->cmd); + err = -EIO; goto out; } - c.packets[cmd == P_BITMAP]++; - c.bytes[cmd == P_BITMAP] += sizeof(struct p_header80) + data_size; + c.packets[pi->cmd == P_BITMAP]++; + c.bytes[pi->cmd == P_BITMAP] += drbd_header_size(tconn) + pi->size; if (err <= 0) { if (err < 0) goto out; break; } - if (!drbd_recv_header(mdev, &cmd, &data_size)) + err = drbd_recv_header(mdev->tconn, pi); + if (err) goto out; } @@ -3620,8 +4280,8 @@ static int receive_bitmap(struct drbd_conf *mdev, enum drbd_packets cmd, unsigne if (mdev->state.conn == C_WF_BITMAP_T) { enum drbd_state_rv rv; - ok = !drbd_send_bitmap(mdev); - if (!ok) + err = drbd_send_bitmap(mdev); + if (err) goto out; /* Omit CS_ORDERED with this state transition to avoid deadlocks. */ rv = _drbd_request_state(mdev, NS(conn, C_WF_SYNC_UUID), CS_VERBOSE); @@ -3632,47 +4292,40 @@ static int receive_bitmap(struct drbd_conf *mdev, enum drbd_packets cmd, unsigne dev_info(DEV, "unexpected cstate (%s) in receive_bitmap\n", drbd_conn_str(mdev->state.conn)); } + err = 0; - ok = true; out: drbd_bm_unlock(mdev); - if (ok && mdev->state.conn == C_WF_BITMAP_S) + if (!err && mdev->state.conn == C_WF_BITMAP_S) drbd_start_resync(mdev, C_SYNC_SOURCE); - free_page((unsigned long) buffer); - return ok; + return err; } -static int receive_skip(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned int data_size) +static int receive_skip(struct drbd_tconn *tconn, struct packet_info *pi) { - /* TODO zero copy sink :) */ - static char sink[128]; - int size, want, r; + conn_warn(tconn, "skipping unknown optional packet type %d, l: %d!\n", + pi->cmd, pi->size); - dev_warn(DEV, "skipping unknown optional packet type %d, l: %d!\n", - cmd, data_size); - - size = data_size; - while (size > 0) { - want = min_t(int, size, sizeof(sink)); - r = drbd_recv(mdev, sink, want); - ERR_IF(r <= 0) break; - size -= r; - } - return size == 0; + return ignore_remaining_packet(tconn, pi); } -static int receive_UnplugRemote(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned int data_size) +static int receive_UnplugRemote(struct drbd_tconn *tconn, struct packet_info *pi) { /* Make sure we've acked all the TCP data associated * with the data requests being unplugged */ - drbd_tcp_quickack(mdev->data.socket); + drbd_tcp_quickack(tconn->data.socket); - return true; + return 0; } -static int receive_out_of_sync(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned int data_size) +static int receive_out_of_sync(struct drbd_tconn *tconn, struct packet_info *pi) { - struct p_block_desc *p = &mdev->data.rbuf.block_desc; + struct drbd_conf *mdev; + struct p_block_desc *p = pi->data; + + mdev = vnr_to_mdev(tconn, pi->vnr); + if (!mdev) + return -EIO; switch (mdev->state.conn) { case C_WF_SYNC_UUID: @@ -3686,15 +4339,13 @@ static int receive_out_of_sync(struct drbd_conf *mdev, enum drbd_packets cmd, un drbd_set_out_of_sync(mdev, be64_to_cpu(p->sector), be32_to_cpu(p->blksize)); - return true; + return 0; } -typedef int (*drbd_cmd_handler_f)(struct drbd_conf *, enum drbd_packets cmd, unsigned int to_receive); - struct data_cmd { int expect_payload; size_t pkt_size; - drbd_cmd_handler_f function; + int (*fn)(struct drbd_tconn *, struct packet_info *); }; static struct data_cmd drbd_cmd_handler[] = { @@ -3702,13 +4353,13 @@ static struct data_cmd drbd_cmd_handler[] = { [P_DATA_REPLY] = { 1, sizeof(struct p_data), receive_DataReply }, [P_RS_DATA_REPLY] = { 1, sizeof(struct p_data), receive_RSDataReply } , [P_BARRIER] = { 0, sizeof(struct p_barrier), receive_Barrier } , - [P_BITMAP] = { 1, sizeof(struct p_header80), receive_bitmap } , - [P_COMPRESSED_BITMAP] = { 1, sizeof(struct p_header80), receive_bitmap } , - [P_UNPLUG_REMOTE] = { 0, sizeof(struct p_header80), receive_UnplugRemote }, + [P_BITMAP] = { 1, 0, receive_bitmap } , + [P_COMPRESSED_BITMAP] = { 1, 0, receive_bitmap } , + [P_UNPLUG_REMOTE] = { 0, 0, receive_UnplugRemote }, [P_DATA_REQUEST] = { 0, sizeof(struct p_block_req), receive_DataRequest }, [P_RS_DATA_REQUEST] = { 0, sizeof(struct p_block_req), receive_DataRequest }, - [P_SYNC_PARAM] = { 1, sizeof(struct p_header80), receive_SyncParam }, - [P_SYNC_PARAM89] = { 1, sizeof(struct p_header80), receive_SyncParam }, + [P_SYNC_PARAM] = { 1, 0, receive_SyncParam }, + [P_SYNC_PARAM89] = { 1, 0, receive_SyncParam }, [P_PROTOCOL] = { 1, sizeof(struct p_protocol), receive_protocol }, [P_UUIDS] = { 0, sizeof(struct p_uuids), receive_uuids }, [P_SIZES] = { 0, sizeof(struct p_sizes), receive_sizes }, @@ -3720,124 +4371,75 @@ static struct data_cmd drbd_cmd_handler[] = { [P_CSUM_RS_REQUEST] = { 1, sizeof(struct p_block_req), receive_DataRequest }, [P_DELAY_PROBE] = { 0, sizeof(struct p_delay_probe93), receive_skip }, [P_OUT_OF_SYNC] = { 0, sizeof(struct p_block_desc), receive_out_of_sync }, - /* anything missing from this table is in - * the asender_tbl, see get_asender_cmd */ - [P_MAX_CMD] = { 0, 0, NULL }, + [P_CONN_ST_CHG_REQ] = { 0, sizeof(struct p_req_state), receive_req_conn_state }, + [P_PROTOCOL_UPDATE] = { 1, sizeof(struct p_protocol), receive_protocol }, }; -/* All handler functions that expect a sub-header get that sub-heder in - mdev->data.rbuf.header.head.payload. - - Usually in mdev->data.rbuf.header.head the callback can find the usual - p_header, but they may not rely on that. Since there is also p_header95 ! - */ - -static void drbdd(struct drbd_conf *mdev) +static void drbdd(struct drbd_tconn *tconn) { - union p_header *header = &mdev->data.rbuf.header; - unsigned int packet_size; - enum drbd_packets cmd; + struct packet_info pi; size_t shs; /* sub header size */ - int rv; + int err; + + while (get_t_state(&tconn->receiver) == RUNNING) { + struct data_cmd *cmd; - while (get_t_state(&mdev->receiver) == Running) { - drbd_thread_current_set_cpu(mdev); - if (!drbd_recv_header(mdev, &cmd, &packet_size)) + drbd_thread_current_set_cpu(&tconn->receiver); + if (drbd_recv_header(tconn, &pi)) goto err_out; - if (unlikely(cmd >= P_MAX_CMD || !drbd_cmd_handler[cmd].function)) { - dev_err(DEV, "unknown packet type %d, l: %d!\n", cmd, packet_size); + cmd = &drbd_cmd_handler[pi.cmd]; + if (unlikely(pi.cmd >= ARRAY_SIZE(drbd_cmd_handler) || !cmd->fn)) { + conn_err(tconn, "Unexpected data packet %s (0x%04x)", + cmdname(pi.cmd), pi.cmd); goto err_out; } - shs = drbd_cmd_handler[cmd].pkt_size - sizeof(union p_header); - if (packet_size - shs > 0 && !drbd_cmd_handler[cmd].expect_payload) { - dev_err(DEV, "No payload expected %s l:%d\n", cmdname(cmd), packet_size); + shs = cmd->pkt_size; + if (pi.size > shs && !cmd->expect_payload) { + conn_err(tconn, "No payload expected %s l:%d\n", + cmdname(pi.cmd), pi.size); goto err_out; } if (shs) { - rv = drbd_recv(mdev, &header->h80.payload, shs); - if (unlikely(rv != shs)) { - if (!signal_pending(current)) - dev_warn(DEV, "short read while reading sub header: rv=%d\n", rv); + err = drbd_recv_all_warn(tconn, pi.data, shs); + if (err) goto err_out; - } + pi.size -= shs; } - rv = drbd_cmd_handler[cmd].function(mdev, cmd, packet_size - shs); - - if (unlikely(!rv)) { - dev_err(DEV, "error receiving %s, l: %d!\n", - cmdname(cmd), packet_size); + err = cmd->fn(tconn, &pi); + if (err) { + conn_err(tconn, "error receiving %s, e: %d l: %d!\n", + cmdname(pi.cmd), err, pi.size); goto err_out; } } + return; - if (0) { - err_out: - drbd_force_state(mdev, NS(conn, C_PROTOCOL_ERROR)); - } - /* If we leave here, we probably want to update at least the - * "Connected" indicator on stable storage. Do so explicitly here. */ - drbd_md_sync(mdev); + err_out: + conn_request_state(tconn, NS(conn, C_PROTOCOL_ERROR), CS_HARD); } -void drbd_flush_workqueue(struct drbd_conf *mdev) +void conn_flush_workqueue(struct drbd_tconn *tconn) { struct drbd_wq_barrier barr; barr.w.cb = w_prev_work_done; + barr.w.tconn = tconn; init_completion(&barr.done); - drbd_queue_work(&mdev->data.work, &barr.w); + drbd_queue_work(&tconn->sender_work, &barr.w); wait_for_completion(&barr.done); } -void drbd_free_tl_hash(struct drbd_conf *mdev) -{ - struct hlist_head *h; - - spin_lock_irq(&mdev->req_lock); - - if (!mdev->tl_hash || mdev->state.conn != C_STANDALONE) { - spin_unlock_irq(&mdev->req_lock); - return; - } - /* paranoia code */ - for (h = mdev->ee_hash; h < mdev->ee_hash + mdev->ee_hash_s; h++) - if (h->first) - dev_err(DEV, "ASSERT FAILED ee_hash[%u].first == %p, expected NULL\n", - (int)(h - mdev->ee_hash), h->first); - kfree(mdev->ee_hash); - mdev->ee_hash = NULL; - mdev->ee_hash_s = 0; - - /* We may not have had the chance to wait for all locally pending - * application requests. The hlist_add_fake() prevents access after - * free on master bio completion. */ - for (h = mdev->tl_hash; h < mdev->tl_hash + mdev->tl_hash_s; h++) { - struct drbd_request *req; - struct hlist_node *pos, *n; - hlist_for_each_entry_safe(req, pos, n, h, collision) { - hlist_del_init(&req->collision); - hlist_add_fake(&req->collision); - } - } - - kfree(mdev->tl_hash); - mdev->tl_hash = NULL; - mdev->tl_hash_s = 0; - spin_unlock_irq(&mdev->req_lock); -} - -static void drbd_disconnect(struct drbd_conf *mdev) +static void conn_disconnect(struct drbd_tconn *tconn) { - enum drbd_fencing_p fp; - union drbd_state os, ns; - int rv = SS_UNKNOWN_ERROR; - unsigned int i; + struct drbd_conf *mdev; + enum drbd_conns oc; + int vnr; - if (mdev->state.conn == C_STANDALONE) + if (tconn->cstate == C_STANDALONE) return; /* We are about to start the cleanup after connection loss. @@ -3845,18 +4447,54 @@ static void drbd_disconnect(struct drbd_conf *mdev) * Usually we should be in some network failure state already, * but just in case we are not, we fix it up here. */ - drbd_force_state(mdev, NS(conn, C_NETWORK_FAILURE)); + conn_request_state(tconn, NS(conn, C_NETWORK_FAILURE), CS_HARD); /* asender does not clean up anything. it must not interfere, either */ - drbd_thread_stop(&mdev->asender); - drbd_free_sock(mdev); + drbd_thread_stop(&tconn->asender); + drbd_free_sock(tconn); + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) { + kref_get(&mdev->kref); + rcu_read_unlock(); + drbd_disconnected(mdev); + kref_put(&mdev->kref, &drbd_minor_destroy); + rcu_read_lock(); + } + rcu_read_unlock(); + + if (!list_empty(&tconn->current_epoch->list)) + conn_err(tconn, "ASSERTION FAILED: tconn->current_epoch->list not empty\n"); + /* ok, no more ee's on the fly, it is safe to reset the epoch_size */ + atomic_set(&tconn->current_epoch->epoch_size, 0); + tconn->send.seen_any_write_yet = false; + + conn_info(tconn, "Connection closed\n"); + + if (conn_highest_role(tconn) == R_PRIMARY && conn_highest_pdsk(tconn) >= D_UNKNOWN) + conn_try_outdate_peer_async(tconn); + + spin_lock_irq(&tconn->req_lock); + oc = tconn->cstate; + if (oc >= C_UNCONNECTED) + _conn_request_state(tconn, NS(conn, C_UNCONNECTED), CS_VERBOSE); + + spin_unlock_irq(&tconn->req_lock); + + if (oc == C_DISCONNECTING) + conn_request_state(tconn, NS(conn, C_STANDALONE), CS_VERBOSE | CS_HARD); +} + +static int drbd_disconnected(struct drbd_conf *mdev) +{ + unsigned int i; /* wait for current activity to cease. */ - spin_lock_irq(&mdev->req_lock); + spin_lock_irq(&mdev->tconn->req_lock); _drbd_wait_ee_list_empty(mdev, &mdev->active_ee); _drbd_wait_ee_list_empty(mdev, &mdev->sync_ee); _drbd_wait_ee_list_empty(mdev, &mdev->read_ee); - spin_unlock_irq(&mdev->req_lock); + spin_unlock_irq(&mdev->tconn->req_lock); /* We do not have data structures that would allow us to * get the rs_pending_cnt down to 0 again. @@ -3874,7 +4512,6 @@ static void drbd_disconnect(struct drbd_conf *mdev) atomic_set(&mdev->rs_pending_cnt, 0); wake_up(&mdev->misc_wait); - /* make sure syncer is stopped and w_resume_next_sg queued */ del_timer_sync(&mdev->resync_timer); resync_timer_fn((unsigned long)mdev); @@ -3883,50 +4520,25 @@ static void drbd_disconnect(struct drbd_conf *mdev) * to be "canceled" */ drbd_flush_workqueue(mdev); - /* This also does reclaim_net_ee(). If we do this too early, we might - * miss some resync ee and pages.*/ - drbd_process_done_ee(mdev); + drbd_finish_peer_reqs(mdev); + + /* This second workqueue flush is necessary, since drbd_finish_peer_reqs() + might have issued a work again. The one before drbd_finish_peer_reqs() is + necessary to reclain net_ee in drbd_finish_peer_reqs(). */ + drbd_flush_workqueue(mdev); + + /* need to do it again, drbd_finish_peer_reqs() may have populated it + * again via drbd_try_clear_on_disk_bm(). */ + drbd_rs_cancel_all(mdev); kfree(mdev->p_uuid); mdev->p_uuid = NULL; - if (!is_susp(mdev->state)) - tl_clear(mdev); - - dev_info(DEV, "Connection closed\n"); + if (!drbd_suspended(mdev)) + tl_clear(mdev->tconn); drbd_md_sync(mdev); - fp = FP_DONT_CARE; - if (get_ldev(mdev)) { - fp = mdev->ldev->dc.fencing; - put_ldev(mdev); - } - - if (mdev->state.role == R_PRIMARY && fp >= FP_RESOURCE && mdev->state.pdsk >= D_UNKNOWN) - drbd_try_outdate_peer_async(mdev); - - spin_lock_irq(&mdev->req_lock); - os = mdev->state; - if (os.conn >= C_UNCONNECTED) { - /* Do not restart in case we are C_DISCONNECTING */ - ns = os; - ns.conn = C_UNCONNECTED; - rv = _drbd_set_state(mdev, ns, CS_VERBOSE, NULL); - } - spin_unlock_irq(&mdev->req_lock); - - if (os.conn == C_DISCONNECTING) { - wait_event(mdev->net_cnt_wait, atomic_read(&mdev->net_cnt) == 0); - - crypto_free_hash(mdev->cram_hmac_tfm); - mdev->cram_hmac_tfm = NULL; - - kfree(mdev->net_conf); - mdev->net_conf = NULL; - drbd_request_state(mdev, NS(conn, C_STANDALONE)); - } - /* serialize with bitmap writeout triggered by the state change, * if any. */ wait_event(mdev->misc_wait, !test_bit(BITMAP_IO, &mdev->flags)); @@ -3938,7 +4550,7 @@ static void drbd_disconnect(struct drbd_conf *mdev) * Actually we don't care for exactly when the network stack does its * put_page(), but release our reference on these pages right here. */ - i = drbd_release_ee(mdev, &mdev->net_ee); + i = drbd_free_peer_reqs(mdev, &mdev->net_ee); if (i) dev_info(DEV, "net_ee not empty, killed %u entries\n", i); i = atomic_read(&mdev->pp_in_use_by_net); @@ -3953,9 +4565,7 @@ static void drbd_disconnect(struct drbd_conf *mdev) D_ASSERT(list_empty(&mdev->sync_ee)); D_ASSERT(list_empty(&mdev->done_ee)); - /* ok, no more ee's on the fly, it is safe to reset the epoch_size */ - atomic_set(&mdev->current_epoch->epoch_size, 0); - D_ASSERT(list_empty(&mdev->current_epoch->list)); + return 0; } /* @@ -3967,29 +4577,19 @@ static void drbd_disconnect(struct drbd_conf *mdev) * * for now, they are expected to be zero, but ignored. */ -static int drbd_send_handshake(struct drbd_conf *mdev) +static int drbd_send_features(struct drbd_tconn *tconn) { - /* ASSERT current == mdev->receiver ... */ - struct p_handshake *p = &mdev->data.sbuf.handshake; - int ok; - - if (mutex_lock_interruptible(&mdev->data.mutex)) { - dev_err(DEV, "interrupted during initial handshake\n"); - return 0; /* interrupted. not ok. */ - } - - if (mdev->data.socket == NULL) { - mutex_unlock(&mdev->data.mutex); - return 0; - } + struct drbd_socket *sock; + struct p_connection_features *p; + sock = &tconn->data; + p = conn_prepare_command(tconn, sock); + if (!p) + return -EIO; memset(p, 0, sizeof(*p)); p->protocol_min = cpu_to_be32(PRO_VERSION_MIN); p->protocol_max = cpu_to_be32(PRO_VERSION_MAX); - ok = _drbd_send_cmd( mdev, mdev->data.socket, P_HAND_SHAKE, - (struct p_header80 *)p, sizeof(*p), 0 ); - mutex_unlock(&mdev->data.mutex); - return ok; + return conn_send_command(tconn, sock, P_CONNECTION_FEATURES, sizeof(*p), NULL, 0); } /* @@ -3999,42 +4599,38 @@ static int drbd_send_handshake(struct drbd_conf *mdev) * -1 peer talks different language, * no point in trying again, please go standalone. */ -static int drbd_do_handshake(struct drbd_conf *mdev) +static int drbd_do_features(struct drbd_tconn *tconn) { - /* ASSERT current == mdev->receiver ... */ - struct p_handshake *p = &mdev->data.rbuf.handshake; - const int expect = sizeof(struct p_handshake) - sizeof(struct p_header80); - unsigned int length; - enum drbd_packets cmd; - int rv; + /* ASSERT current == tconn->receiver ... */ + struct p_connection_features *p; + const int expect = sizeof(struct p_connection_features); + struct packet_info pi; + int err; - rv = drbd_send_handshake(mdev); - if (!rv) + err = drbd_send_features(tconn); + if (err) return 0; - rv = drbd_recv_header(mdev, &cmd, &length); - if (!rv) + err = drbd_recv_header(tconn, &pi); + if (err) return 0; - if (cmd != P_HAND_SHAKE) { - dev_err(DEV, "expected HandShake packet, received: %s (0x%04x)\n", - cmdname(cmd), cmd); + if (pi.cmd != P_CONNECTION_FEATURES) { + conn_err(tconn, "expected ConnectionFeatures packet, received: %s (0x%04x)\n", + cmdname(pi.cmd), pi.cmd); return -1; } - if (length != expect) { - dev_err(DEV, "expected HandShake length: %u, received: %u\n", - expect, length); + if (pi.size != expect) { + conn_err(tconn, "expected ConnectionFeatures length: %u, received: %u\n", + expect, pi.size); return -1; } - rv = drbd_recv(mdev, &p->head.payload, expect); - - if (rv != expect) { - if (!signal_pending(current)) - dev_warn(DEV, "short read receiving handshake packet: l=%u\n", rv); + p = pi.data; + err = drbd_recv_all_warn(tconn, p, expect); + if (err) return 0; - } p->protocol_min = be32_to_cpu(p->protocol_min); p->protocol_max = be32_to_cpu(p->protocol_max); @@ -4045,15 +4641,15 @@ static int drbd_do_handshake(struct drbd_conf *mdev) PRO_VERSION_MIN > p->protocol_max) goto incompat; - mdev->agreed_pro_version = min_t(int, PRO_VERSION_MAX, p->protocol_max); + tconn->agreed_pro_version = min_t(int, PRO_VERSION_MAX, p->protocol_max); - dev_info(DEV, "Handshake successful: " - "Agreed network protocol version %d\n", mdev->agreed_pro_version); + conn_info(tconn, "Handshake successful: " + "Agreed network protocol version %d\n", tconn->agreed_pro_version); return 1; incompat: - dev_err(DEV, "incompatible DRBD dialects: " + conn_err(tconn, "incompatible DRBD dialects: " "I support %d-%d, peer supports %d-%d\n", PRO_VERSION_MIN, PRO_VERSION_MAX, p->protocol_min, p->protocol_max); @@ -4061,7 +4657,7 @@ static int drbd_do_handshake(struct drbd_conf *mdev) } #if !defined(CONFIG_CRYPTO_HMAC) && !defined(CONFIG_CRYPTO_HMAC_MODULE) -static int drbd_do_auth(struct drbd_conf *mdev) +static int drbd_do_auth(struct drbd_tconn *tconn) { dev_err(DEV, "This kernel was build without CONFIG_CRYPTO_HMAC.\n"); dev_err(DEV, "You need to disable 'cram-hmac-alg' in drbd.conf.\n"); @@ -4076,121 +4672,139 @@ static int drbd_do_auth(struct drbd_conf *mdev) -1 - auth failed, don't try again. */ -static int drbd_do_auth(struct drbd_conf *mdev) +static int drbd_do_auth(struct drbd_tconn *tconn) { + struct drbd_socket *sock; char my_challenge[CHALLENGE_LEN]; /* 64 Bytes... */ struct scatterlist sg; char *response = NULL; char *right_response = NULL; char *peers_ch = NULL; - unsigned int key_len = strlen(mdev->net_conf->shared_secret); + unsigned int key_len; + char secret[SHARED_SECRET_MAX]; /* 64 byte */ unsigned int resp_size; struct hash_desc desc; - enum drbd_packets cmd; - unsigned int length; - int rv; + struct packet_info pi; + struct net_conf *nc; + int err, rv; + + /* FIXME: Put the challenge/response into the preallocated socket buffer. */ - desc.tfm = mdev->cram_hmac_tfm; + rcu_read_lock(); + nc = rcu_dereference(tconn->net_conf); + key_len = strlen(nc->shared_secret); + memcpy(secret, nc->shared_secret, key_len); + rcu_read_unlock(); + + desc.tfm = tconn->cram_hmac_tfm; desc.flags = 0; - rv = crypto_hash_setkey(mdev->cram_hmac_tfm, - (u8 *)mdev->net_conf->shared_secret, key_len); + rv = crypto_hash_setkey(tconn->cram_hmac_tfm, (u8 *)secret, key_len); if (rv) { - dev_err(DEV, "crypto_hash_setkey() failed with %d\n", rv); + conn_err(tconn, "crypto_hash_setkey() failed with %d\n", rv); rv = -1; goto fail; } get_random_bytes(my_challenge, CHALLENGE_LEN); - rv = drbd_send_cmd2(mdev, P_AUTH_CHALLENGE, my_challenge, CHALLENGE_LEN); + sock = &tconn->data; + if (!conn_prepare_command(tconn, sock)) { + rv = 0; + goto fail; + } + rv = !conn_send_command(tconn, sock, P_AUTH_CHALLENGE, 0, + my_challenge, CHALLENGE_LEN); if (!rv) goto fail; - rv = drbd_recv_header(mdev, &cmd, &length); - if (!rv) + err = drbd_recv_header(tconn, &pi); + if (err) { + rv = 0; goto fail; + } - if (cmd != P_AUTH_CHALLENGE) { - dev_err(DEV, "expected AuthChallenge packet, received: %s (0x%04x)\n", - cmdname(cmd), cmd); + if (pi.cmd != P_AUTH_CHALLENGE) { + conn_err(tconn, "expected AuthChallenge packet, received: %s (0x%04x)\n", + cmdname(pi.cmd), pi.cmd); rv = 0; goto fail; } - if (length > CHALLENGE_LEN * 2) { - dev_err(DEV, "expected AuthChallenge payload too big.\n"); + if (pi.size > CHALLENGE_LEN * 2) { + conn_err(tconn, "expected AuthChallenge payload too big.\n"); rv = -1; goto fail; } - peers_ch = kmalloc(length, GFP_NOIO); + peers_ch = kmalloc(pi.size, GFP_NOIO); if (peers_ch == NULL) { - dev_err(DEV, "kmalloc of peers_ch failed\n"); + conn_err(tconn, "kmalloc of peers_ch failed\n"); rv = -1; goto fail; } - rv = drbd_recv(mdev, peers_ch, length); - - if (rv != length) { - if (!signal_pending(current)) - dev_warn(DEV, "short read AuthChallenge: l=%u\n", rv); + err = drbd_recv_all_warn(tconn, peers_ch, pi.size); + if (err) { rv = 0; goto fail; } - resp_size = crypto_hash_digestsize(mdev->cram_hmac_tfm); + resp_size = crypto_hash_digestsize(tconn->cram_hmac_tfm); response = kmalloc(resp_size, GFP_NOIO); if (response == NULL) { - dev_err(DEV, "kmalloc of response failed\n"); + conn_err(tconn, "kmalloc of response failed\n"); rv = -1; goto fail; } sg_init_table(&sg, 1); - sg_set_buf(&sg, peers_ch, length); + sg_set_buf(&sg, peers_ch, pi.size); rv = crypto_hash_digest(&desc, &sg, sg.length, response); if (rv) { - dev_err(DEV, "crypto_hash_digest() failed with %d\n", rv); + conn_err(tconn, "crypto_hash_digest() failed with %d\n", rv); rv = -1; goto fail; } - rv = drbd_send_cmd2(mdev, P_AUTH_RESPONSE, response, resp_size); - if (!rv) + if (!conn_prepare_command(tconn, sock)) { + rv = 0; goto fail; - - rv = drbd_recv_header(mdev, &cmd, &length); + } + rv = !conn_send_command(tconn, sock, P_AUTH_RESPONSE, 0, + response, resp_size); if (!rv) goto fail; - if (cmd != P_AUTH_RESPONSE) { - dev_err(DEV, "expected AuthResponse packet, received: %s (0x%04x)\n", - cmdname(cmd), cmd); + err = drbd_recv_header(tconn, &pi); + if (err) { rv = 0; goto fail; } - if (length != resp_size) { - dev_err(DEV, "expected AuthResponse payload of wrong size\n"); + if (pi.cmd != P_AUTH_RESPONSE) { + conn_err(tconn, "expected AuthResponse packet, received: %s (0x%04x)\n", + cmdname(pi.cmd), pi.cmd); rv = 0; goto fail; } - rv = drbd_recv(mdev, response , resp_size); + if (pi.size != resp_size) { + conn_err(tconn, "expected AuthResponse payload of wrong size\n"); + rv = 0; + goto fail; + } - if (rv != resp_size) { - if (!signal_pending(current)) - dev_warn(DEV, "short read receiving AuthResponse: l=%u\n", rv); + err = drbd_recv_all_warn(tconn, response , resp_size); + if (err) { rv = 0; goto fail; } right_response = kmalloc(resp_size, GFP_NOIO); if (right_response == NULL) { - dev_err(DEV, "kmalloc of right_response failed\n"); + conn_err(tconn, "kmalloc of right_response failed\n"); rv = -1; goto fail; } @@ -4199,7 +4813,7 @@ static int drbd_do_auth(struct drbd_conf *mdev) rv = crypto_hash_digest(&desc, &sg, sg.length, right_response); if (rv) { - dev_err(DEV, "crypto_hash_digest() failed with %d\n", rv); + conn_err(tconn, "crypto_hash_digest() failed with %d\n", rv); rv = -1; goto fail; } @@ -4207,8 +4821,8 @@ static int drbd_do_auth(struct drbd_conf *mdev) rv = !memcmp(response, right_response, resp_size); if (rv) - dev_info(DEV, "Peer authenticated using %d bytes of '%s' HMAC\n", - resp_size, mdev->net_conf->cram_hmac_alg); + conn_info(tconn, "Peer authenticated using %d bytes HMAC\n", + resp_size); else rv = -1; @@ -4223,82 +4837,106 @@ static int drbd_do_auth(struct drbd_conf *mdev) int drbdd_init(struct drbd_thread *thi) { - struct drbd_conf *mdev = thi->mdev; - unsigned int minor = mdev_to_minor(mdev); + struct drbd_tconn *tconn = thi->tconn; int h; - sprintf(current->comm, "drbd%d_receiver", minor); - - dev_info(DEV, "receiver (re)started\n"); + conn_info(tconn, "receiver (re)started\n"); do { - h = drbd_connect(mdev); + h = conn_connect(tconn); if (h == 0) { - drbd_disconnect(mdev); + conn_disconnect(tconn); schedule_timeout_interruptible(HZ); } if (h == -1) { - dev_warn(DEV, "Discarding network configuration.\n"); - drbd_force_state(mdev, NS(conn, C_DISCONNECTING)); + conn_warn(tconn, "Discarding network configuration.\n"); + conn_request_state(tconn, NS(conn, C_DISCONNECTING), CS_HARD); } } while (h == 0); - if (h > 0) { - if (get_net_conf(mdev)) { - drbdd(mdev); - put_net_conf(mdev); - } - } + if (h > 0) + drbdd(tconn); - drbd_disconnect(mdev); + conn_disconnect(tconn); - dev_info(DEV, "receiver terminated\n"); + conn_info(tconn, "receiver terminated\n"); return 0; } /* ********* acknowledge sender ******** */ -static int got_RqSReply(struct drbd_conf *mdev, struct p_header80 *h) +static int got_conn_RqSReply(struct drbd_tconn *tconn, struct packet_info *pi) { - struct p_req_state_reply *p = (struct p_req_state_reply *)h; + struct p_req_state_reply *p = pi->data; + int retcode = be32_to_cpu(p->retcode); + + if (retcode >= SS_SUCCESS) { + set_bit(CONN_WD_ST_CHG_OKAY, &tconn->flags); + } else { + set_bit(CONN_WD_ST_CHG_FAIL, &tconn->flags); + conn_err(tconn, "Requested state change failed by peer: %s (%d)\n", + drbd_set_st_err_str(retcode), retcode); + } + wake_up(&tconn->ping_wait); + + return 0; +} +static int got_RqSReply(struct drbd_tconn *tconn, struct packet_info *pi) +{ + struct drbd_conf *mdev; + struct p_req_state_reply *p = pi->data; int retcode = be32_to_cpu(p->retcode); + mdev = vnr_to_mdev(tconn, pi->vnr); + if (!mdev) + return -EIO; + + if (test_bit(CONN_WD_ST_CHG_REQ, &tconn->flags)) { + D_ASSERT(tconn->agreed_pro_version < 100); + return got_conn_RqSReply(tconn, pi); + } + if (retcode >= SS_SUCCESS) { set_bit(CL_ST_CHG_SUCCESS, &mdev->flags); } else { set_bit(CL_ST_CHG_FAIL, &mdev->flags); dev_err(DEV, "Requested state change failed by peer: %s (%d)\n", - drbd_set_st_err_str(retcode), retcode); + drbd_set_st_err_str(retcode), retcode); } wake_up(&mdev->state_wait); - return true; + return 0; } -static int got_Ping(struct drbd_conf *mdev, struct p_header80 *h) +static int got_Ping(struct drbd_tconn *tconn, struct packet_info *pi) { - return drbd_send_ping_ack(mdev); + return drbd_send_ping_ack(tconn); } -static int got_PingAck(struct drbd_conf *mdev, struct p_header80 *h) +static int got_PingAck(struct drbd_tconn *tconn, struct packet_info *pi) { /* restore idle timeout */ - mdev->meta.socket->sk->sk_rcvtimeo = mdev->net_conf->ping_int*HZ; - if (!test_and_set_bit(GOT_PING_ACK, &mdev->flags)) - wake_up(&mdev->misc_wait); + tconn->meta.socket->sk->sk_rcvtimeo = tconn->net_conf->ping_int*HZ; + if (!test_and_set_bit(GOT_PING_ACK, &tconn->flags)) + wake_up(&tconn->ping_wait); - return true; + return 0; } -static int got_IsInSync(struct drbd_conf *mdev, struct p_header80 *h) +static int got_IsInSync(struct drbd_tconn *tconn, struct packet_info *pi) { - struct p_block_ack *p = (struct p_block_ack *)h; + struct drbd_conf *mdev; + struct p_block_ack *p = pi->data; sector_t sector = be64_to_cpu(p->sector); int blksize = be32_to_cpu(p->blksize); - D_ASSERT(mdev->agreed_pro_version >= 89); + mdev = vnr_to_mdev(tconn, pi->vnr); + if (!mdev) + return -EIO; + + D_ASSERT(mdev->tconn->agreed_pro_version >= 89); update_peer_seq(mdev, be32_to_cpu(p->seq_num)); @@ -4312,162 +4950,139 @@ static int got_IsInSync(struct drbd_conf *mdev, struct p_header80 *h) dec_rs_pending(mdev); atomic_add(blksize >> 9, &mdev->rs_sect_in); - return true; -} - -/* when we receive the ACK for a write request, - * verify that we actually know about it */ -static struct drbd_request *_ack_id_to_req(struct drbd_conf *mdev, - u64 id, sector_t sector) -{ - struct hlist_head *slot = tl_hash_slot(mdev, sector); - struct hlist_node *n; - struct drbd_request *req; - - hlist_for_each_entry(req, n, slot, collision) { - if ((unsigned long)req == (unsigned long)id) { - if (req->sector != sector) { - dev_err(DEV, "_ack_id_to_req: found req %p but it has " - "wrong sector (%llus versus %llus)\n", req, - (unsigned long long)req->sector, - (unsigned long long)sector); - break; - } - return req; - } - } - return NULL; + return 0; } -typedef struct drbd_request *(req_validator_fn) - (struct drbd_conf *mdev, u64 id, sector_t sector); - -static int validate_req_change_req_state(struct drbd_conf *mdev, - u64 id, sector_t sector, req_validator_fn validator, - const char *func, enum drbd_req_event what) +static int +validate_req_change_req_state(struct drbd_conf *mdev, u64 id, sector_t sector, + struct rb_root *root, const char *func, + enum drbd_req_event what, bool missing_ok) { struct drbd_request *req; struct bio_and_error m; - spin_lock_irq(&mdev->req_lock); - req = validator(mdev, id, sector); + spin_lock_irq(&mdev->tconn->req_lock); + req = find_request(mdev, root, id, sector, missing_ok, func); if (unlikely(!req)) { - spin_unlock_irq(&mdev->req_lock); - - dev_err(DEV, "%s: failed to find req %p, sector %llus\n", func, - (void *)(unsigned long)id, (unsigned long long)sector); - return false; + spin_unlock_irq(&mdev->tconn->req_lock); + return -EIO; } __req_mod(req, what, &m); - spin_unlock_irq(&mdev->req_lock); + spin_unlock_irq(&mdev->tconn->req_lock); if (m.bio) complete_master_bio(mdev, &m); - return true; + return 0; } -static int got_BlockAck(struct drbd_conf *mdev, struct p_header80 *h) +static int got_BlockAck(struct drbd_tconn *tconn, struct packet_info *pi) { - struct p_block_ack *p = (struct p_block_ack *)h; + struct drbd_conf *mdev; + struct p_block_ack *p = pi->data; sector_t sector = be64_to_cpu(p->sector); int blksize = be32_to_cpu(p->blksize); enum drbd_req_event what; + mdev = vnr_to_mdev(tconn, pi->vnr); + if (!mdev) + return -EIO; + update_peer_seq(mdev, be32_to_cpu(p->seq_num)); - if (is_syncer_block_id(p->block_id)) { + if (p->block_id == ID_SYNCER) { drbd_set_in_sync(mdev, sector, blksize); dec_rs_pending(mdev); - return true; + return 0; } - switch (be16_to_cpu(h->command)) { + switch (pi->cmd) { case P_RS_WRITE_ACK: - D_ASSERT(mdev->net_conf->wire_protocol == DRBD_PROT_C); - what = write_acked_by_peer_and_sis; + what = WRITE_ACKED_BY_PEER_AND_SIS; break; case P_WRITE_ACK: - D_ASSERT(mdev->net_conf->wire_protocol == DRBD_PROT_C); - what = write_acked_by_peer; + what = WRITE_ACKED_BY_PEER; break; case P_RECV_ACK: - D_ASSERT(mdev->net_conf->wire_protocol == DRBD_PROT_B); - what = recv_acked_by_peer; + what = RECV_ACKED_BY_PEER; break; - case P_DISCARD_ACK: - D_ASSERT(mdev->net_conf->wire_protocol == DRBD_PROT_C); - what = conflict_discarded_by_peer; + case P_SUPERSEDED: + what = CONFLICT_RESOLVED; + break; + case P_RETRY_WRITE: + what = POSTPONE_WRITE; break; default: - D_ASSERT(0); - return false; + BUG(); } return validate_req_change_req_state(mdev, p->block_id, sector, - _ack_id_to_req, __func__ , what); + &mdev->write_requests, __func__, + what, false); } -static int got_NegAck(struct drbd_conf *mdev, struct p_header80 *h) +static int got_NegAck(struct drbd_tconn *tconn, struct packet_info *pi) { - struct p_block_ack *p = (struct p_block_ack *)h; + struct drbd_conf *mdev; + struct p_block_ack *p = pi->data; sector_t sector = be64_to_cpu(p->sector); int size = be32_to_cpu(p->blksize); - struct drbd_request *req; - struct bio_and_error m; + int err; + + mdev = vnr_to_mdev(tconn, pi->vnr); + if (!mdev) + return -EIO; update_peer_seq(mdev, be32_to_cpu(p->seq_num)); - if (is_syncer_block_id(p->block_id)) { + if (p->block_id == ID_SYNCER) { dec_rs_pending(mdev); drbd_rs_failed_io(mdev, sector, size); - return true; + return 0; } - spin_lock_irq(&mdev->req_lock); - req = _ack_id_to_req(mdev, p->block_id, sector); - if (!req) { - spin_unlock_irq(&mdev->req_lock); - if (mdev->net_conf->wire_protocol == DRBD_PROT_A || - mdev->net_conf->wire_protocol == DRBD_PROT_B) { - /* Protocol A has no P_WRITE_ACKs, but has P_NEG_ACKs. - The master bio might already be completed, therefore the - request is no longer in the collision hash. - => Do not try to validate block_id as request. */ - /* In Protocol B we might already have got a P_RECV_ACK - but then get a P_NEG_ACK after wards. */ - drbd_set_out_of_sync(mdev, sector, size); - return true; - } else { - dev_err(DEV, "%s: failed to find req %p, sector %llus\n", __func__, - (void *)(unsigned long)p->block_id, (unsigned long long)sector); - return false; - } + err = validate_req_change_req_state(mdev, p->block_id, sector, + &mdev->write_requests, __func__, + NEG_ACKED, true); + if (err) { + /* Protocol A has no P_WRITE_ACKs, but has P_NEG_ACKs. + The master bio might already be completed, therefore the + request is no longer in the collision hash. */ + /* In Protocol B we might already have got a P_RECV_ACK + but then get a P_NEG_ACK afterwards. */ + drbd_set_out_of_sync(mdev, sector, size); } - __req_mod(req, neg_acked, &m); - spin_unlock_irq(&mdev->req_lock); - - if (m.bio) - complete_master_bio(mdev, &m); - return true; + return 0; } -static int got_NegDReply(struct drbd_conf *mdev, struct p_header80 *h) +static int got_NegDReply(struct drbd_tconn *tconn, struct packet_info *pi) { - struct p_block_ack *p = (struct p_block_ack *)h; + struct drbd_conf *mdev; + struct p_block_ack *p = pi->data; sector_t sector = be64_to_cpu(p->sector); + mdev = vnr_to_mdev(tconn, pi->vnr); + if (!mdev) + return -EIO; + update_peer_seq(mdev, be32_to_cpu(p->seq_num)); - dev_err(DEV, "Got NegDReply; Sector %llus, len %u; Fail original request.\n", + + dev_err(DEV, "Got NegDReply; Sector %llus, len %u.\n", (unsigned long long)sector, be32_to_cpu(p->blksize)); return validate_req_change_req_state(mdev, p->block_id, sector, - _ar_id_to_req, __func__ , neg_acked); + &mdev->read_requests, __func__, + NEG_ACKED, false); } -static int got_NegRSDReply(struct drbd_conf *mdev, struct p_header80 *h) +static int got_NegRSDReply(struct drbd_tconn *tconn, struct packet_info *pi) { + struct drbd_conf *mdev; sector_t sector; int size; - struct p_block_ack *p = (struct p_block_ack *)h; + struct p_block_ack *p = pi->data; + + mdev = vnr_to_mdev(tconn, pi->vnr); + if (!mdev) + return -EIO; sector = be64_to_cpu(p->sector); size = be32_to_cpu(p->blksize); @@ -4478,57 +5093,66 @@ static int got_NegRSDReply(struct drbd_conf *mdev, struct p_header80 *h) if (get_ldev_if_state(mdev, D_FAILED)) { drbd_rs_complete_io(mdev, sector); - switch (be16_to_cpu(h->command)) { + switch (pi->cmd) { case P_NEG_RS_DREPLY: drbd_rs_failed_io(mdev, sector, size); case P_RS_CANCEL: break; default: - D_ASSERT(0); - put_ldev(mdev); - return false; + BUG(); } put_ldev(mdev); } - return true; + return 0; } -static int got_BarrierAck(struct drbd_conf *mdev, struct p_header80 *h) +static int got_BarrierAck(struct drbd_tconn *tconn, struct packet_info *pi) { - struct p_barrier_ack *p = (struct p_barrier_ack *)h; - - tl_release(mdev, p->barrier, be32_to_cpu(p->set_size)); - - if (mdev->state.conn == C_AHEAD && - atomic_read(&mdev->ap_in_flight) == 0 && - !test_and_set_bit(AHEAD_TO_SYNC_SOURCE, &mdev->flags)) { - mdev->start_resync_timer.expires = jiffies + HZ; - add_timer(&mdev->start_resync_timer); + struct p_barrier_ack *p = pi->data; + struct drbd_conf *mdev; + int vnr; + + tl_release(tconn, p->barrier, be32_to_cpu(p->set_size)); + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) { + if (mdev->state.conn == C_AHEAD && + atomic_read(&mdev->ap_in_flight) == 0 && + !test_and_set_bit(AHEAD_TO_SYNC_SOURCE, &mdev->flags)) { + mdev->start_resync_timer.expires = jiffies + HZ; + add_timer(&mdev->start_resync_timer); + } } + rcu_read_unlock(); - return true; + return 0; } -static int got_OVResult(struct drbd_conf *mdev, struct p_header80 *h) +static int got_OVResult(struct drbd_tconn *tconn, struct packet_info *pi) { - struct p_block_ack *p = (struct p_block_ack *)h; + struct drbd_conf *mdev; + struct p_block_ack *p = pi->data; struct drbd_work *w; sector_t sector; int size; + mdev = vnr_to_mdev(tconn, pi->vnr); + if (!mdev) + return -EIO; + sector = be64_to_cpu(p->sector); size = be32_to_cpu(p->blksize); update_peer_seq(mdev, be32_to_cpu(p->seq_num)); if (be64_to_cpu(p->block_id) == ID_OUT_OF_SYNC) - drbd_ov_oos_found(mdev, sector, size); + drbd_ov_out_of_sync_found(mdev, sector, size); else - ov_oos_print(mdev); + ov_out_of_sync_print(mdev); if (!get_ldev(mdev)) - return true; + return 0; drbd_rs_complete_io(mdev, sector); dec_rs_pending(mdev); @@ -4543,114 +5167,137 @@ static int got_OVResult(struct drbd_conf *mdev, struct p_header80 *h) w = kmalloc(sizeof(*w), GFP_NOIO); if (w) { w->cb = w_ov_finished; - drbd_queue_work_front(&mdev->data.work, w); + w->mdev = mdev; + drbd_queue_work(&mdev->tconn->sender_work, w); } else { dev_err(DEV, "kmalloc(w) failed."); - ov_oos_print(mdev); + ov_out_of_sync_print(mdev); drbd_resync_finished(mdev); } } put_ldev(mdev); - return true; + return 0; +} + +static int got_skip(struct drbd_tconn *tconn, struct packet_info *pi) +{ + return 0; } -static int got_skip(struct drbd_conf *mdev, struct p_header80 *h) +static int tconn_finish_peer_reqs(struct drbd_tconn *tconn) { - return true; + struct drbd_conf *mdev; + int vnr, not_empty = 0; + + do { + clear_bit(SIGNAL_ASENDER, &tconn->flags); + flush_signals(current); + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) { + kref_get(&mdev->kref); + rcu_read_unlock(); + if (drbd_finish_peer_reqs(mdev)) { + kref_put(&mdev->kref, &drbd_minor_destroy); + return 1; + } + kref_put(&mdev->kref, &drbd_minor_destroy); + rcu_read_lock(); + } + set_bit(SIGNAL_ASENDER, &tconn->flags); + + spin_lock_irq(&tconn->req_lock); + idr_for_each_entry(&tconn->volumes, mdev, vnr) { + not_empty = !list_empty(&mdev->done_ee); + if (not_empty) + break; + } + spin_unlock_irq(&tconn->req_lock); + rcu_read_unlock(); + } while (not_empty); + + return 0; } struct asender_cmd { size_t pkt_size; - int (*process)(struct drbd_conf *mdev, struct p_header80 *h); + int (*fn)(struct drbd_tconn *tconn, struct packet_info *); }; -static struct asender_cmd *get_asender_cmd(int cmd) -{ - static struct asender_cmd asender_tbl[] = { - /* anything missing from this table is in - * the drbd_cmd_handler (drbd_default_handler) table, - * see the beginning of drbdd() */ - [P_PING] = { sizeof(struct p_header80), got_Ping }, - [P_PING_ACK] = { sizeof(struct p_header80), got_PingAck }, +static struct asender_cmd asender_tbl[] = { + [P_PING] = { 0, got_Ping }, + [P_PING_ACK] = { 0, got_PingAck }, [P_RECV_ACK] = { sizeof(struct p_block_ack), got_BlockAck }, [P_WRITE_ACK] = { sizeof(struct p_block_ack), got_BlockAck }, [P_RS_WRITE_ACK] = { sizeof(struct p_block_ack), got_BlockAck }, - [P_DISCARD_ACK] = { sizeof(struct p_block_ack), got_BlockAck }, + [P_SUPERSEDED] = { sizeof(struct p_block_ack), got_BlockAck }, [P_NEG_ACK] = { sizeof(struct p_block_ack), got_NegAck }, [P_NEG_DREPLY] = { sizeof(struct p_block_ack), got_NegDReply }, - [P_NEG_RS_DREPLY] = { sizeof(struct p_block_ack), got_NegRSDReply}, + [P_NEG_RS_DREPLY] = { sizeof(struct p_block_ack), got_NegRSDReply }, [P_OV_RESULT] = { sizeof(struct p_block_ack), got_OVResult }, [P_BARRIER_ACK] = { sizeof(struct p_barrier_ack), got_BarrierAck }, [P_STATE_CHG_REPLY] = { sizeof(struct p_req_state_reply), got_RqSReply }, [P_RS_IS_IN_SYNC] = { sizeof(struct p_block_ack), got_IsInSync }, [P_DELAY_PROBE] = { sizeof(struct p_delay_probe93), got_skip }, - [P_RS_CANCEL] = { sizeof(struct p_block_ack), got_NegRSDReply}, - [P_MAX_CMD] = { 0, NULL }, - }; - if (cmd > P_MAX_CMD || asender_tbl[cmd].process == NULL) - return NULL; - return &asender_tbl[cmd]; -} + [P_RS_CANCEL] = { sizeof(struct p_block_ack), got_NegRSDReply }, + [P_CONN_ST_CHG_REPLY]={ sizeof(struct p_req_state_reply), got_conn_RqSReply }, + [P_RETRY_WRITE] = { sizeof(struct p_block_ack), got_BlockAck }, +}; int drbd_asender(struct drbd_thread *thi) { - struct drbd_conf *mdev = thi->mdev; - struct p_header80 *h = &mdev->meta.rbuf.header.h80; + struct drbd_tconn *tconn = thi->tconn; struct asender_cmd *cmd = NULL; - - int rv, len; - void *buf = h; + struct packet_info pi; + int rv; + void *buf = tconn->meta.rbuf; int received = 0; - int expect = sizeof(struct p_header80); - int empty; - int ping_timeout_active = 0; - - sprintf(current->comm, "drbd%d_asender", mdev_to_minor(mdev)); + unsigned int header_size = drbd_header_size(tconn); + int expect = header_size; + bool ping_timeout_active = false; + struct net_conf *nc; + int ping_timeo, tcp_cork, ping_int; current->policy = SCHED_RR; /* Make this a realtime task! */ current->rt_priority = 2; /* more important than all other tasks */ - while (get_t_state(thi) == Running) { - drbd_thread_current_set_cpu(mdev); - if (test_and_clear_bit(SEND_PING, &mdev->flags)) { - ERR_IF(!drbd_send_ping(mdev)) goto reconnect; - mdev->meta.socket->sk->sk_rcvtimeo = - mdev->net_conf->ping_timeo*HZ/10; - ping_timeout_active = 1; - } + while (get_t_state(thi) == RUNNING) { + drbd_thread_current_set_cpu(thi); - /* conditionally cork; - * it may hurt latency if we cork without much to send */ - if (!mdev->net_conf->no_cork && - 3 < atomic_read(&mdev->unacked_cnt)) - drbd_tcp_cork(mdev->meta.socket); - while (1) { - clear_bit(SIGNAL_ASENDER, &mdev->flags); - flush_signals(current); - if (!drbd_process_done_ee(mdev)) + rcu_read_lock(); + nc = rcu_dereference(tconn->net_conf); + ping_timeo = nc->ping_timeo; + tcp_cork = nc->tcp_cork; + ping_int = nc->ping_int; + rcu_read_unlock(); + + if (test_and_clear_bit(SEND_PING, &tconn->flags)) { + if (drbd_send_ping(tconn)) { + conn_err(tconn, "drbd_send_ping has failed\n"); goto reconnect; - /* to avoid race with newly queued ACKs */ - set_bit(SIGNAL_ASENDER, &mdev->flags); - spin_lock_irq(&mdev->req_lock); - empty = list_empty(&mdev->done_ee); - spin_unlock_irq(&mdev->req_lock); - /* new ack may have been queued right here, - * but then there is also a signal pending, - * and we start over... */ - if (empty) - break; + } + tconn->meta.socket->sk->sk_rcvtimeo = ping_timeo * HZ / 10; + ping_timeout_active = true; + } + + /* TODO: conditionally cork; it may hurt latency if we cork without + much to send */ + if (tcp_cork) + drbd_tcp_cork(tconn->meta.socket); + if (tconn_finish_peer_reqs(tconn)) { + conn_err(tconn, "tconn_finish_peer_reqs() failed\n"); + goto reconnect; } /* but unconditionally uncork unless disabled */ - if (!mdev->net_conf->no_cork) - drbd_tcp_uncork(mdev->meta.socket); + if (tcp_cork) + drbd_tcp_uncork(tconn->meta.socket); /* short circuit, recv_msg would return EINTR anyways. */ if (signal_pending(current)) continue; - rv = drbd_recv_short(mdev, mdev->meta.socket, - buf, expect-received, 0); - clear_bit(SIGNAL_ASENDER, &mdev->flags); + rv = drbd_recv_short(tconn->meta.socket, buf, expect-received, 0); + clear_bit(SIGNAL_ASENDER, &tconn->flags); flush_signals(current); @@ -4668,80 +5315,91 @@ int drbd_asender(struct drbd_thread *thi) received += rv; buf += rv; } else if (rv == 0) { - dev_err(DEV, "meta connection shut down by peer.\n"); + if (test_bit(DISCONNECT_SENT, &tconn->flags)) { + long t; + rcu_read_lock(); + t = rcu_dereference(tconn->net_conf)->ping_timeo * HZ/10; + rcu_read_unlock(); + + t = wait_event_timeout(tconn->ping_wait, + tconn->cstate < C_WF_REPORT_PARAMS, + t); + if (t) + break; + } + conn_err(tconn, "meta connection shut down by peer.\n"); goto reconnect; } else if (rv == -EAGAIN) { /* If the data socket received something meanwhile, * that is good enough: peer is still alive. */ - if (time_after(mdev->last_received, - jiffies - mdev->meta.socket->sk->sk_rcvtimeo)) + if (time_after(tconn->last_received, + jiffies - tconn->meta.socket->sk->sk_rcvtimeo)) continue; if (ping_timeout_active) { - dev_err(DEV, "PingAck did not arrive in time.\n"); + conn_err(tconn, "PingAck did not arrive in time.\n"); goto reconnect; } - set_bit(SEND_PING, &mdev->flags); + set_bit(SEND_PING, &tconn->flags); continue; } else if (rv == -EINTR) { continue; } else { - dev_err(DEV, "sock_recvmsg returned %d\n", rv); + conn_err(tconn, "sock_recvmsg returned %d\n", rv); goto reconnect; } if (received == expect && cmd == NULL) { - if (unlikely(h->magic != BE_DRBD_MAGIC)) { - dev_err(DEV, "magic?? on meta m: 0x%08x c: %d l: %d\n", - be32_to_cpu(h->magic), - be16_to_cpu(h->command), - be16_to_cpu(h->length)); + if (decode_header(tconn, tconn->meta.rbuf, &pi)) goto reconnect; - } - cmd = get_asender_cmd(be16_to_cpu(h->command)); - len = be16_to_cpu(h->length); - if (unlikely(cmd == NULL)) { - dev_err(DEV, "unknown command?? on meta m: 0x%08x c: %d l: %d\n", - be32_to_cpu(h->magic), - be16_to_cpu(h->command), - be16_to_cpu(h->length)); + cmd = &asender_tbl[pi.cmd]; + if (pi.cmd >= ARRAY_SIZE(asender_tbl) || !cmd->fn) { + conn_err(tconn, "Unexpected meta packet %s (0x%04x)\n", + cmdname(pi.cmd), pi.cmd); goto disconnect; } - expect = cmd->pkt_size; - ERR_IF(len != expect-sizeof(struct p_header80)) + expect = header_size + cmd->pkt_size; + if (pi.size != expect - header_size) { + conn_err(tconn, "Wrong packet size on meta (c: %d, l: %d)\n", + pi.cmd, pi.size); goto reconnect; + } } if (received == expect) { - mdev->last_received = jiffies; - D_ASSERT(cmd != NULL); - if (!cmd->process(mdev, h)) + bool err; + + err = cmd->fn(tconn, &pi); + if (err) { + conn_err(tconn, "%pf failed\n", cmd->fn); goto reconnect; + } + + tconn->last_received = jiffies; - /* the idle_timeout (ping-int) - * has been restored in got_PingAck() */ - if (cmd == get_asender_cmd(P_PING_ACK)) - ping_timeout_active = 0; + if (cmd == &asender_tbl[P_PING_ACK]) { + /* restore idle timeout */ + tconn->meta.socket->sk->sk_rcvtimeo = ping_int * HZ; + ping_timeout_active = false; + } - buf = h; + buf = tconn->meta.rbuf; received = 0; - expect = sizeof(struct p_header80); + expect = header_size; cmd = NULL; } } if (0) { reconnect: - drbd_force_state(mdev, NS(conn, C_NETWORK_FAILURE)); - drbd_md_sync(mdev); + conn_request_state(tconn, NS(conn, C_NETWORK_FAILURE), CS_HARD); + conn_md_sync(tconn); } if (0) { disconnect: - drbd_force_state(mdev, NS(conn, C_DISCONNECTING)); - drbd_md_sync(mdev); + conn_request_state(tconn, NS(conn, C_DISCONNECTING), CS_HARD); } - clear_bit(SIGNAL_ASENDER, &mdev->flags); + clear_bit(SIGNAL_ASENDER, &tconn->flags); - D_ASSERT(mdev->state.conn < C_CONNECTED); - dev_info(DEV, "asender terminated\n"); + conn_info(tconn, "asender terminated\n"); return 0; } diff --git a/drivers/block/drbd/drbd_req.c b/drivers/block/drbd/drbd_req.c index 01b2ac641c7b..f58a4a4b4dfb 100644 --- a/drivers/block/drbd/drbd_req.c +++ b/drivers/block/drbd/drbd_req.c @@ -31,6 +31,8 @@ #include "drbd_req.h" +static bool drbd_may_do_local_read(struct drbd_conf *mdev, sector_t sector, int size); + /* Update disk stats at start of I/O request */ static void _drbd_start_io_acct(struct drbd_conf *mdev, struct drbd_request *req, struct bio *bio) { @@ -40,6 +42,8 @@ static void _drbd_start_io_acct(struct drbd_conf *mdev, struct drbd_request *req part_round_stats(cpu, &mdev->vdisk->part0); part_stat_inc(cpu, &mdev->vdisk->part0, ios[rw]); part_stat_add(cpu, &mdev->vdisk->part0, sectors[rw], bio_sectors(bio)); + (void) cpu; /* The macro invocations above want the cpu argument, I do not like + the compiler warning about cpu only assigned but never used... */ part_inc_in_flight(&mdev->vdisk->part0, rw); part_stat_unlock(); } @@ -57,9 +61,51 @@ static void _drbd_end_io_acct(struct drbd_conf *mdev, struct drbd_request *req) part_stat_unlock(); } -static void _req_is_done(struct drbd_conf *mdev, struct drbd_request *req, const int rw) +static struct drbd_request *drbd_req_new(struct drbd_conf *mdev, + struct bio *bio_src) +{ + struct drbd_request *req; + + req = mempool_alloc(drbd_request_mempool, GFP_NOIO); + if (!req) + return NULL; + + drbd_req_make_private_bio(req, bio_src); + req->rq_state = bio_data_dir(bio_src) == WRITE ? RQ_WRITE : 0; + req->w.mdev = mdev; + req->master_bio = bio_src; + req->epoch = 0; + + drbd_clear_interval(&req->i); + req->i.sector = bio_src->bi_sector; + req->i.size = bio_src->bi_size; + req->i.local = true; + req->i.waiting = false; + + INIT_LIST_HEAD(&req->tl_requests); + INIT_LIST_HEAD(&req->w.list); + + /* one reference to be put by __drbd_make_request */ + atomic_set(&req->completion_ref, 1); + /* one kref as long as completion_ref > 0 */ + kref_init(&req->kref); + return req; +} + +void drbd_req_destroy(struct kref *kref) { - const unsigned long s = req->rq_state; + struct drbd_request *req = container_of(kref, struct drbd_request, kref); + struct drbd_conf *mdev = req->w.mdev; + const unsigned s = req->rq_state; + + if ((req->master_bio && !(s & RQ_POSTPONED)) || + atomic_read(&req->completion_ref) || + (s & RQ_LOCAL_PENDING) || + ((s & RQ_NET_MASK) && !(s & RQ_NET_DONE))) { + dev_err(DEV, "drbd_req_destroy: Logic BUG rq_state = 0x%x, completion_ref = %d\n", + s, atomic_read(&req->completion_ref)); + return; + } /* remove it from the transfer log. * well, only if it had been there in the first @@ -67,24 +113,33 @@ static void _req_is_done(struct drbd_conf *mdev, struct drbd_request *req, const * and never sent), it should still be "empty" as * initialized in drbd_req_new(), so we can list_del() it * here unconditionally */ - list_del(&req->tl_requests); + list_del_init(&req->tl_requests); /* if it was a write, we may have to set the corresponding * bit(s) out-of-sync first. If it had a local part, we need to * release the reference to the activity log. */ - if (rw == WRITE) { + if (s & RQ_WRITE) { /* Set out-of-sync unless both OK flags are set * (local only or remote failed). * Other places where we set out-of-sync: * READ with local io-error */ - if (!(s & RQ_NET_OK) || !(s & RQ_LOCAL_OK)) - drbd_set_out_of_sync(mdev, req->sector, req->size); - if ((s & RQ_NET_OK) && (s & RQ_LOCAL_OK) && (s & RQ_NET_SIS)) - drbd_set_in_sync(mdev, req->sector, req->size); + /* There is a special case: + * we may notice late that IO was suspended, + * and postpone, or schedule for retry, a write, + * before it even was submitted or sent. + * In that case we do not want to touch the bitmap at all. + */ + if ((s & (RQ_POSTPONED|RQ_LOCAL_MASK|RQ_NET_MASK)) != RQ_POSTPONED) { + if (!(s & RQ_NET_OK) || !(s & RQ_LOCAL_OK)) + drbd_set_out_of_sync(mdev, req->i.sector, req->i.size); + + if ((s & RQ_NET_OK) && (s & RQ_LOCAL_OK) && (s & RQ_NET_SIS)) + drbd_set_in_sync(mdev, req->i.sector, req->i.size); + } /* one might be tempted to move the drbd_al_complete_io - * to the local io completion callback drbd_endio_pri. + * to the local io completion callback drbd_request_endio. * but, if this was a mirror write, we may only * drbd_al_complete_io after this is RQ_NET_DONE, * otherwise the extent could be dropped from the al @@ -93,109 +148,35 @@ static void _req_is_done(struct drbd_conf *mdev, struct drbd_request *req, const * but after the extent has been dropped from the al, * we would forget to resync the corresponding extent. */ - if (s & RQ_LOCAL_MASK) { + if (s & RQ_IN_ACT_LOG) { if (get_ldev_if_state(mdev, D_FAILED)) { - if (s & RQ_IN_ACT_LOG) - drbd_al_complete_io(mdev, req->sector); + drbd_al_complete_io(mdev, &req->i); put_ldev(mdev); } else if (__ratelimit(&drbd_ratelimit_state)) { - dev_warn(DEV, "Should have called drbd_al_complete_io(, %llu), " - "but my Disk seems to have failed :(\n", - (unsigned long long) req->sector); + dev_warn(DEV, "Should have called drbd_al_complete_io(, %llu, %u), " + "but my Disk seems to have failed :(\n", + (unsigned long long) req->i.sector, req->i.size); } } } - drbd_req_free(req); + mempool_free(req, drbd_request_mempool); } -static void queue_barrier(struct drbd_conf *mdev) -{ - struct drbd_tl_epoch *b; - - /* We are within the req_lock. Once we queued the barrier for sending, - * we set the CREATE_BARRIER bit. It is cleared as soon as a new - * barrier/epoch object is added. This is the only place this bit is - * set. It indicates that the barrier for this epoch is already queued, - * and no new epoch has been created yet. */ - if (test_bit(CREATE_BARRIER, &mdev->flags)) - return; - - b = mdev->newest_tle; - b->w.cb = w_send_barrier; - /* inc_ap_pending done here, so we won't - * get imbalanced on connection loss. - * dec_ap_pending will be done in got_BarrierAck - * or (on connection loss) in tl_clear. */ - inc_ap_pending(mdev); - drbd_queue_work(&mdev->data.work, &b->w); - set_bit(CREATE_BARRIER, &mdev->flags); +static void wake_all_senders(struct drbd_tconn *tconn) { + wake_up(&tconn->sender_work.q_wait); } -static void _about_to_complete_local_write(struct drbd_conf *mdev, - struct drbd_request *req) +/* must hold resource->req_lock */ +static void start_new_tl_epoch(struct drbd_tconn *tconn) { - const unsigned long s = req->rq_state; - struct drbd_request *i; - struct drbd_epoch_entry *e; - struct hlist_node *n; - struct hlist_head *slot; - - /* Before we can signal completion to the upper layers, - * we may need to close the current epoch. - * We can skip this, if this request has not even been sent, because we - * did not have a fully established connection yet/anymore, during - * bitmap exchange, or while we are C_AHEAD due to congestion policy. - */ - if (mdev->state.conn >= C_CONNECTED && - (s & RQ_NET_SENT) != 0 && - req->epoch == mdev->newest_tle->br_number) - queue_barrier(mdev); - - /* we need to do the conflict detection stuff, - * if we have the ee_hash (two_primaries) and - * this has been on the network */ - if ((s & RQ_NET_DONE) && mdev->ee_hash != NULL) { - const sector_t sector = req->sector; - const int size = req->size; - - /* ASSERT: - * there must be no conflicting requests, since - * they must have been failed on the spot */ -#define OVERLAPS overlaps(sector, size, i->sector, i->size) - slot = tl_hash_slot(mdev, sector); - hlist_for_each_entry(i, n, slot, collision) { - if (OVERLAPS) { - dev_alert(DEV, "LOGIC BUG: completed: %p %llus +%u; " - "other: %p %llus +%u\n", - req, (unsigned long long)sector, size, - i, (unsigned long long)i->sector, i->size); - } - } + /* no point closing an epoch, if it is empty, anyways. */ + if (tconn->current_tle_writes == 0) + return; - /* maybe "wake" those conflicting epoch entries - * that wait for this request to finish. - * - * currently, there can be only _one_ such ee - * (well, or some more, which would be pending - * P_DISCARD_ACK not yet sent by the asender...), - * since we block the receiver thread upon the - * first conflict detection, which will wait on - * misc_wait. maybe we want to assert that? - * - * anyways, if we found one, - * we just have to do a wake_up. */ -#undef OVERLAPS -#define OVERLAPS overlaps(sector, size, e->sector, e->size) - slot = ee_hash_slot(mdev, req->sector); - hlist_for_each_entry(e, n, slot, collision) { - if (OVERLAPS) { - wake_up(&mdev->misc_wait); - break; - } - } - } -#undef OVERLAPS + tconn->current_tle_writes = 0; + atomic_inc(&tconn->current_tle_nr); + wake_all_senders(tconn); } void complete_master_bio(struct drbd_conf *mdev, @@ -205,17 +186,33 @@ void complete_master_bio(struct drbd_conf *mdev, dec_ap_bio(mdev); } + +static void drbd_remove_request_interval(struct rb_root *root, + struct drbd_request *req) +{ + struct drbd_conf *mdev = req->w.mdev; + struct drbd_interval *i = &req->i; + + drbd_remove_interval(root, i); + + /* Wake up any processes waiting for this request to complete. */ + if (i->waiting) + wake_up(&mdev->misc_wait); +} + /* Helper for __req_mod(). * Set m->bio to the master bio, if it is fit to be completed, * or leave it alone (it is initialized to NULL in __req_mod), * if it has already been completed, or cannot be completed yet. * If m->bio is set, the error status to be returned is placed in m->error. */ -void _req_may_be_done(struct drbd_request *req, struct bio_and_error *m) +static +void drbd_req_complete(struct drbd_request *req, struct bio_and_error *m) { - const unsigned long s = req->rq_state; - struct drbd_conf *mdev = req->mdev; - int rw = req->rq_state & RQ_WRITE ? WRITE : READ; + const unsigned s = req->rq_state; + struct drbd_conf *mdev = req->w.mdev; + int rw; + int error, ok; /* we must not complete the master bio, while it is * still being processed by _drbd_send_zc_bio (drbd_send_dblock) @@ -226,165 +223,220 @@ void _req_may_be_done(struct drbd_request *req, struct bio_and_error *m) * the receiver, * the bio_endio completion callbacks. */ - if (s & RQ_NET_QUEUED) - return; - if (s & RQ_NET_PENDING) + if ((s & RQ_LOCAL_PENDING && !(s & RQ_LOCAL_ABORTED)) || + (s & RQ_NET_QUEUED) || (s & RQ_NET_PENDING) || + (s & RQ_COMPLETION_SUSP)) { + dev_err(DEV, "drbd_req_complete: Logic BUG rq_state = 0x%x\n", s); return; - if (s & RQ_LOCAL_PENDING && !(s & RQ_LOCAL_ABORTED)) + } + + if (!req->master_bio) { + dev_err(DEV, "drbd_req_complete: Logic BUG, master_bio == NULL!\n"); return; + } - if (req->master_bio) { - /* this is data_received (remote read) - * or protocol C P_WRITE_ACK - * or protocol B P_RECV_ACK - * or protocol A "handed_over_to_network" (SendAck) - * or canceled or failed, - * or killed from the transfer log due to connection loss. - */ + rw = bio_rw(req->master_bio); - /* - * figure out whether to report success or failure. - * - * report success when at least one of the operations succeeded. - * or, to put the other way, - * only report failure, when both operations failed. - * - * what to do about the failures is handled elsewhere. - * what we need to do here is just: complete the master_bio. - * - * local completion error, if any, has been stored as ERR_PTR - * in private_bio within drbd_endio_pri. - */ - int ok = (s & RQ_LOCAL_OK) || (s & RQ_NET_OK); - int error = PTR_ERR(req->private_bio); + /* + * figure out whether to report success or failure. + * + * report success when at least one of the operations succeeded. + * or, to put the other way, + * only report failure, when both operations failed. + * + * what to do about the failures is handled elsewhere. + * what we need to do here is just: complete the master_bio. + * + * local completion error, if any, has been stored as ERR_PTR + * in private_bio within drbd_request_endio. + */ + ok = (s & RQ_LOCAL_OK) || (s & RQ_NET_OK); + error = PTR_ERR(req->private_bio); - /* remove the request from the conflict detection - * respective block_id verification hash */ - if (!hlist_unhashed(&req->collision)) - hlist_del(&req->collision); - else - D_ASSERT((s & (RQ_NET_MASK & ~RQ_NET_DONE)) == 0); + /* remove the request from the conflict detection + * respective block_id verification hash */ + if (!drbd_interval_empty(&req->i)) { + struct rb_root *root; - /* for writes we need to do some extra housekeeping */ if (rw == WRITE) - _about_to_complete_local_write(mdev, req); + root = &mdev->write_requests; + else + root = &mdev->read_requests; + drbd_remove_request_interval(root, req); + } else if (!(s & RQ_POSTPONED)) + D_ASSERT((s & (RQ_NET_MASK & ~RQ_NET_DONE)) == 0); - /* Update disk stats */ - _drbd_end_io_acct(mdev, req); + /* Before we can signal completion to the upper layers, + * we may need to close the current transfer log epoch. + * We are within the request lock, so we can simply compare + * the request epoch number with the current transfer log + * epoch number. If they match, increase the current_tle_nr, + * and reset the transfer log epoch write_cnt. + */ + if (rw == WRITE && + req->epoch == atomic_read(&mdev->tconn->current_tle_nr)) + start_new_tl_epoch(mdev->tconn); + + /* Update disk stats */ + _drbd_end_io_acct(mdev, req); + + /* If READ failed, + * have it be pushed back to the retry work queue, + * so it will re-enter __drbd_make_request(), + * and be re-assigned to a suitable local or remote path, + * or failed if we do not have access to good data anymore. + * + * Unless it was failed early by __drbd_make_request(), + * because no path was available, in which case + * it was not even added to the transfer_log. + * + * READA may fail, and will not be retried. + * + * WRITE should have used all available paths already. + */ + if (!ok && rw == READ && !list_empty(&req->tl_requests)) + req->rq_state |= RQ_POSTPONED; + if (!(req->rq_state & RQ_POSTPONED)) { m->error = ok ? 0 : (error ?: -EIO); m->bio = req->master_bio; req->master_bio = NULL; } +} - if (s & RQ_LOCAL_PENDING) - return; +static int drbd_req_put_completion_ref(struct drbd_request *req, struct bio_and_error *m, int put) +{ + struct drbd_conf *mdev = req->w.mdev; + D_ASSERT(m || (req->rq_state & RQ_POSTPONED)); + + if (!atomic_sub_and_test(put, &req->completion_ref)) + return 0; - if ((s & RQ_NET_MASK) == 0 || (s & RQ_NET_DONE)) { - /* this is disconnected (local only) operation, - * or protocol C P_WRITE_ACK, - * or protocol A or B P_BARRIER_ACK, - * or killed from the transfer log due to connection loss. */ - _req_is_done(mdev, req, rw); + drbd_req_complete(req, m); + + if (req->rq_state & RQ_POSTPONED) { + /* don't destroy the req object just yet, + * but queue it for retry */ + drbd_restart_request(req); + return 0; } - /* else: network part and not DONE yet. that is - * protocol A or B, barrier ack still pending... */ + + return 1; } -static void _req_may_be_done_not_susp(struct drbd_request *req, struct bio_and_error *m) +/* I'd like this to be the only place that manipulates + * req->completion_ref and req->kref. */ +static void mod_rq_state(struct drbd_request *req, struct bio_and_error *m, + int clear, int set) { - struct drbd_conf *mdev = req->mdev; + struct drbd_conf *mdev = req->w.mdev; + unsigned s = req->rq_state; + int c_put = 0; + int k_put = 0; - if (!is_susp(mdev->state)) - _req_may_be_done(req, m); -} + if (drbd_suspended(mdev) && !((s | clear) & RQ_COMPLETION_SUSP)) + set |= RQ_COMPLETION_SUSP; -/* - * checks whether there was an overlapping request - * or ee already registered. - * - * if so, return 1, in which case this request is completed on the spot, - * without ever being submitted or send. - * - * return 0 if it is ok to submit this request. - * - * NOTE: - * paranoia: assume something above us is broken, and issues different write - * requests for the same block simultaneously... - * - * To ensure these won't be reordered differently on both nodes, resulting in - * diverging data sets, we discard the later one(s). Not that this is supposed - * to happen, but this is the rationale why we also have to check for - * conflicting requests with local origin, and why we have to do so regardless - * of whether we allowed multiple primaries. - * - * BTW, in case we only have one primary, the ee_hash is empty anyways, and the - * second hlist_for_each_entry becomes a noop. This is even simpler than to - * grab a reference on the net_conf, and check for the two_primaries flag... - */ -static int _req_conflicts(struct drbd_request *req) -{ - struct drbd_conf *mdev = req->mdev; - const sector_t sector = req->sector; - const int size = req->size; - struct drbd_request *i; - struct drbd_epoch_entry *e; - struct hlist_node *n; - struct hlist_head *slot; + /* apply */ - D_ASSERT(hlist_unhashed(&req->collision)); + req->rq_state &= ~clear; + req->rq_state |= set; - if (!get_net_conf(mdev)) - return 0; + /* no change? */ + if (req->rq_state == s) + return; - /* BUG_ON */ - ERR_IF (mdev->tl_hash_s == 0) - goto out_no_conflict; - BUG_ON(mdev->tl_hash == NULL); - -#define OVERLAPS overlaps(i->sector, i->size, sector, size) - slot = tl_hash_slot(mdev, sector); - hlist_for_each_entry(i, n, slot, collision) { - if (OVERLAPS) { - dev_alert(DEV, "%s[%u] Concurrent local write detected! " - "[DISCARD L] new: %llus +%u; " - "pending: %llus +%u\n", - current->comm, current->pid, - (unsigned long long)sector, size, - (unsigned long long)i->sector, i->size); - goto out_conflict; - } + /* intent: get references */ + + if (!(s & RQ_LOCAL_PENDING) && (set & RQ_LOCAL_PENDING)) + atomic_inc(&req->completion_ref); + + if (!(s & RQ_NET_PENDING) && (set & RQ_NET_PENDING)) { + inc_ap_pending(mdev); + atomic_inc(&req->completion_ref); } - if (mdev->ee_hash_s) { - /* now, check for overlapping requests with remote origin */ - BUG_ON(mdev->ee_hash == NULL); -#undef OVERLAPS -#define OVERLAPS overlaps(e->sector, e->size, sector, size) - slot = ee_hash_slot(mdev, sector); - hlist_for_each_entry(e, n, slot, collision) { - if (OVERLAPS) { - dev_alert(DEV, "%s[%u] Concurrent remote write detected!" - " [DISCARD L] new: %llus +%u; " - "pending: %llus +%u\n", - current->comm, current->pid, - (unsigned long long)sector, size, - (unsigned long long)e->sector, e->size); - goto out_conflict; - } - } + if (!(s & RQ_NET_QUEUED) && (set & RQ_NET_QUEUED)) + atomic_inc(&req->completion_ref); + + if (!(s & RQ_EXP_BARR_ACK) && (set & RQ_EXP_BARR_ACK)) + kref_get(&req->kref); /* wait for the DONE */ + + if (!(s & RQ_NET_SENT) && (set & RQ_NET_SENT)) + atomic_add(req->i.size >> 9, &mdev->ap_in_flight); + + if (!(s & RQ_COMPLETION_SUSP) && (set & RQ_COMPLETION_SUSP)) + atomic_inc(&req->completion_ref); + + /* progress: put references */ + + if ((s & RQ_COMPLETION_SUSP) && (clear & RQ_COMPLETION_SUSP)) + ++c_put; + + if (!(s & RQ_LOCAL_ABORTED) && (set & RQ_LOCAL_ABORTED)) { + D_ASSERT(req->rq_state & RQ_LOCAL_PENDING); + /* local completion may still come in later, + * we need to keep the req object around. */ + kref_get(&req->kref); + ++c_put; + } + + if ((s & RQ_LOCAL_PENDING) && (clear & RQ_LOCAL_PENDING)) { + if (req->rq_state & RQ_LOCAL_ABORTED) + ++k_put; + else + ++c_put; } -#undef OVERLAPS -out_no_conflict: - /* this is like it should be, and what we expected. - * our users do behave after all... */ - put_net_conf(mdev); - return 0; + if ((s & RQ_NET_PENDING) && (clear & RQ_NET_PENDING)) { + dec_ap_pending(mdev); + ++c_put; + } -out_conflict: - put_net_conf(mdev); - return 1; + if ((s & RQ_NET_QUEUED) && (clear & RQ_NET_QUEUED)) + ++c_put; + + if ((s & RQ_EXP_BARR_ACK) && !(s & RQ_NET_DONE) && (set & RQ_NET_DONE)) { + if (req->rq_state & RQ_NET_SENT) + atomic_sub(req->i.size >> 9, &mdev->ap_in_flight); + ++k_put; + } + + /* potentially complete and destroy */ + + if (k_put || c_put) { + /* Completion does it's own kref_put. If we are going to + * kref_sub below, we need req to be still around then. */ + int at_least = k_put + !!c_put; + int refcount = atomic_read(&req->kref.refcount); + if (refcount < at_least) + dev_err(DEV, + "mod_rq_state: Logic BUG: %x -> %x: refcount = %d, should be >= %d\n", + s, req->rq_state, refcount, at_least); + } + + /* If we made progress, retry conflicting peer requests, if any. */ + if (req->i.waiting) + wake_up(&mdev->misc_wait); + + if (c_put) + k_put += drbd_req_put_completion_ref(req, m, c_put); + if (k_put) + kref_sub(&req->kref, k_put, drbd_req_destroy); +} + +static void drbd_report_io_error(struct drbd_conf *mdev, struct drbd_request *req) +{ + char b[BDEVNAME_SIZE]; + + if (!__ratelimit(&drbd_ratelimit_state)) + return; + + dev_warn(DEV, "local %s IO error sector %llu+%u on %s\n", + (req->rq_state & RQ_WRITE) ? "WRITE" : "READ", + (unsigned long long)req->i.sector, + req->i.size >> 9, + bdevname(mdev->ldev->backing_bdev, b)); } /* obviously this could be coded as many single functions @@ -402,9 +454,12 @@ out_conflict: int __req_mod(struct drbd_request *req, enum drbd_req_event what, struct bio_and_error *m) { - struct drbd_conf *mdev = req->mdev; - int rv = 0; - m->bio = NULL; + struct drbd_conf *mdev = req->w.mdev; + struct net_conf *nc; + int p, rv = 0; + + if (m) + m->bio = NULL; switch (what) { default: @@ -413,116 +468,91 @@ int __req_mod(struct drbd_request *req, enum drbd_req_event what, /* does not happen... * initialization done in drbd_req_new - case created: + case CREATED: break; */ - case to_be_send: /* via network */ - /* reached via drbd_make_request_common + case TO_BE_SENT: /* via network */ + /* reached via __drbd_make_request * and from w_read_retry_remote */ D_ASSERT(!(req->rq_state & RQ_NET_MASK)); - req->rq_state |= RQ_NET_PENDING; - inc_ap_pending(mdev); + rcu_read_lock(); + nc = rcu_dereference(mdev->tconn->net_conf); + p = nc->wire_protocol; + rcu_read_unlock(); + req->rq_state |= + p == DRBD_PROT_C ? RQ_EXP_WRITE_ACK : + p == DRBD_PROT_B ? RQ_EXP_RECEIVE_ACK : 0; + mod_rq_state(req, m, 0, RQ_NET_PENDING); break; - case to_be_submitted: /* locally */ - /* reached via drbd_make_request_common */ + case TO_BE_SUBMITTED: /* locally */ + /* reached via __drbd_make_request */ D_ASSERT(!(req->rq_state & RQ_LOCAL_MASK)); - req->rq_state |= RQ_LOCAL_PENDING; + mod_rq_state(req, m, 0, RQ_LOCAL_PENDING); break; - case completed_ok: + case COMPLETED_OK: if (req->rq_state & RQ_WRITE) - mdev->writ_cnt += req->size>>9; + mdev->writ_cnt += req->i.size >> 9; else - mdev->read_cnt += req->size>>9; + mdev->read_cnt += req->i.size >> 9; - req->rq_state |= (RQ_LOCAL_COMPLETED|RQ_LOCAL_OK); - req->rq_state &= ~RQ_LOCAL_PENDING; - - _req_may_be_done_not_susp(req, m); + mod_rq_state(req, m, RQ_LOCAL_PENDING, + RQ_LOCAL_COMPLETED|RQ_LOCAL_OK); break; - case abort_disk_io: - req->rq_state |= RQ_LOCAL_ABORTED; - if (req->rq_state & RQ_WRITE) - _req_may_be_done_not_susp(req, m); - else - goto goto_queue_for_net_read; + case ABORT_DISK_IO: + mod_rq_state(req, m, 0, RQ_LOCAL_ABORTED); break; - case write_completed_with_error: - req->rq_state |= RQ_LOCAL_COMPLETED; - req->rq_state &= ~RQ_LOCAL_PENDING; - - __drbd_chk_io_error(mdev, DRBD_IO_ERROR); - _req_may_be_done_not_susp(req, m); + case WRITE_COMPLETED_WITH_ERROR: + drbd_report_io_error(mdev, req); + __drbd_chk_io_error(mdev, DRBD_WRITE_ERROR); + mod_rq_state(req, m, RQ_LOCAL_PENDING, RQ_LOCAL_COMPLETED); break; - case read_ahead_completed_with_error: - /* it is legal to fail READA */ - req->rq_state |= RQ_LOCAL_COMPLETED; - req->rq_state &= ~RQ_LOCAL_PENDING; - _req_may_be_done_not_susp(req, m); + case READ_COMPLETED_WITH_ERROR: + drbd_set_out_of_sync(mdev, req->i.sector, req->i.size); + drbd_report_io_error(mdev, req); + __drbd_chk_io_error(mdev, DRBD_READ_ERROR); + /* fall through. */ + case READ_AHEAD_COMPLETED_WITH_ERROR: + /* it is legal to fail READA, no __drbd_chk_io_error in that case. */ + mod_rq_state(req, m, RQ_LOCAL_PENDING, RQ_LOCAL_COMPLETED); break; - case read_completed_with_error: - drbd_set_out_of_sync(mdev, req->sector, req->size); - - req->rq_state |= RQ_LOCAL_COMPLETED; - req->rq_state &= ~RQ_LOCAL_PENDING; - - if (req->rq_state & RQ_LOCAL_ABORTED) { - _req_may_be_done(req, m); - break; - } - - __drbd_chk_io_error(mdev, DRBD_IO_ERROR); - - goto_queue_for_net_read: - - D_ASSERT(!(req->rq_state & RQ_NET_MASK)); - - /* no point in retrying if there is no good remote data, - * or we have no connection. */ - if (mdev->state.pdsk != D_UP_TO_DATE) { - _req_may_be_done_not_susp(req, m); - break; - } - - /* _req_mod(req,to_be_send); oops, recursion... */ - req->rq_state |= RQ_NET_PENDING; - inc_ap_pending(mdev); - /* fall through: _req_mod(req,queue_for_net_read); */ - - case queue_for_net_read: + case QUEUE_FOR_NET_READ: /* READ or READA, and * no local disk, * or target area marked as invalid, * or just got an io-error. */ - /* from drbd_make_request_common + /* from __drbd_make_request * or from bio_endio during read io-error recovery */ - /* so we can verify the handle in the answer packet - * corresponding hlist_del is in _req_may_be_done() */ - hlist_add_head(&req->collision, ar_hash_slot(mdev, req->sector)); + /* So we can verify the handle in the answer packet. + * Corresponding drbd_remove_request_interval is in + * drbd_req_complete() */ + D_ASSERT(drbd_interval_empty(&req->i)); + drbd_insert_interval(&mdev->read_requests, &req->i); set_bit(UNPLUG_REMOTE, &mdev->flags); D_ASSERT(req->rq_state & RQ_NET_PENDING); - req->rq_state |= RQ_NET_QUEUED; - req->w.cb = (req->rq_state & RQ_LOCAL_MASK) - ? w_read_retry_remote - : w_send_read_req; - drbd_queue_work(&mdev->data.work, &req->w); + D_ASSERT((req->rq_state & RQ_LOCAL_MASK) == 0); + mod_rq_state(req, m, 0, RQ_NET_QUEUED); + req->w.cb = w_send_read_req; + drbd_queue_work(&mdev->tconn->sender_work, &req->w); break; - case queue_for_net_write: + case QUEUE_FOR_NET_WRITE: /* assert something? */ - /* from drbd_make_request_common only */ + /* from __drbd_make_request only */ - hlist_add_head(&req->collision, tl_hash_slot(mdev, req->sector)); - /* corresponding hlist_del is in _req_may_be_done() */ + /* Corresponding drbd_remove_request_interval is in + * drbd_req_complete() */ + D_ASSERT(drbd_interval_empty(&req->i)); + drbd_insert_interval(&mdev->write_requests, &req->i); /* NOTE * In case the req ended up on the transfer log before being @@ -533,7 +563,7 @@ int __req_mod(struct drbd_request *req, enum drbd_req_event what, * * _req_add_to_epoch(req); this has to be after the * _maybe_start_new_epoch(req); which happened in - * drbd_make_request_common, because we now may set the bit + * __drbd_make_request, because we now may set the bit * again ourselves to close the current epoch. * * Add req to the (now) current epoch (barrier). */ @@ -543,202 +573,187 @@ int __req_mod(struct drbd_request *req, enum drbd_req_event what, * hurting performance. */ set_bit(UNPLUG_REMOTE, &mdev->flags); - /* see drbd_make_request_common, - * just after it grabs the req_lock */ - D_ASSERT(test_bit(CREATE_BARRIER, &mdev->flags) == 0); - - req->epoch = mdev->newest_tle->br_number; - - /* increment size of current epoch */ - mdev->newest_tle->n_writes++; - /* queue work item to send data */ D_ASSERT(req->rq_state & RQ_NET_PENDING); - req->rq_state |= RQ_NET_QUEUED; + mod_rq_state(req, m, 0, RQ_NET_QUEUED|RQ_EXP_BARR_ACK); req->w.cb = w_send_dblock; - drbd_queue_work(&mdev->data.work, &req->w); + drbd_queue_work(&mdev->tconn->sender_work, &req->w); /* close the epoch, in case it outgrew the limit */ - if (mdev->newest_tle->n_writes >= mdev->net_conf->max_epoch_size) - queue_barrier(mdev); + rcu_read_lock(); + nc = rcu_dereference(mdev->tconn->net_conf); + p = nc->max_epoch_size; + rcu_read_unlock(); + if (mdev->tconn->current_tle_writes >= p) + start_new_tl_epoch(mdev->tconn); break; - case queue_for_send_oos: - req->rq_state |= RQ_NET_QUEUED; - req->w.cb = w_send_oos; - drbd_queue_work(&mdev->data.work, &req->w); + case QUEUE_FOR_SEND_OOS: + mod_rq_state(req, m, 0, RQ_NET_QUEUED); + req->w.cb = w_send_out_of_sync; + drbd_queue_work(&mdev->tconn->sender_work, &req->w); break; - case read_retry_remote_canceled: - case send_canceled: - case send_failed: + case READ_RETRY_REMOTE_CANCELED: + case SEND_CANCELED: + case SEND_FAILED: /* real cleanup will be done from tl_clear. just update flags * so it is no longer marked as on the worker queue */ - req->rq_state &= ~RQ_NET_QUEUED; - /* if we did it right, tl_clear should be scheduled only after - * this, so this should not be necessary! */ - _req_may_be_done_not_susp(req, m); + mod_rq_state(req, m, RQ_NET_QUEUED, 0); break; - case handed_over_to_network: + case HANDED_OVER_TO_NETWORK: /* assert something? */ - if (bio_data_dir(req->master_bio) == WRITE) - atomic_add(req->size>>9, &mdev->ap_in_flight); - if (bio_data_dir(req->master_bio) == WRITE && - mdev->net_conf->wire_protocol == DRBD_PROT_A) { + !(req->rq_state & (RQ_EXP_RECEIVE_ACK | RQ_EXP_WRITE_ACK))) { /* this is what is dangerous about protocol A: * pretend it was successfully written on the peer. */ - if (req->rq_state & RQ_NET_PENDING) { - dec_ap_pending(mdev); - req->rq_state &= ~RQ_NET_PENDING; - req->rq_state |= RQ_NET_OK; - } /* else: neg-ack was faster... */ + if (req->rq_state & RQ_NET_PENDING) + mod_rq_state(req, m, RQ_NET_PENDING, RQ_NET_OK); + /* else: neg-ack was faster... */ /* it is still not yet RQ_NET_DONE until the * corresponding epoch barrier got acked as well, * so we know what to dirty on connection loss */ } - req->rq_state &= ~RQ_NET_QUEUED; - req->rq_state |= RQ_NET_SENT; - _req_may_be_done_not_susp(req, m); + mod_rq_state(req, m, RQ_NET_QUEUED, RQ_NET_SENT); break; - case oos_handed_to_network: + case OOS_HANDED_TO_NETWORK: /* Was not set PENDING, no longer QUEUED, so is now DONE * as far as this connection is concerned. */ - req->rq_state &= ~RQ_NET_QUEUED; - req->rq_state |= RQ_NET_DONE; - _req_may_be_done_not_susp(req, m); + mod_rq_state(req, m, RQ_NET_QUEUED, RQ_NET_DONE); break; - case connection_lost_while_pending: + case CONNECTION_LOST_WHILE_PENDING: /* transfer log cleanup after connection loss */ - /* assert something? */ - if (req->rq_state & RQ_NET_PENDING) - dec_ap_pending(mdev); - req->rq_state &= ~(RQ_NET_OK|RQ_NET_PENDING); - req->rq_state |= RQ_NET_DONE; - if (req->rq_state & RQ_NET_SENT && req->rq_state & RQ_WRITE) - atomic_sub(req->size>>9, &mdev->ap_in_flight); - - /* if it is still queued, we may not complete it here. - * it will be canceled soon. */ - if (!(req->rq_state & RQ_NET_QUEUED)) - _req_may_be_done(req, m); /* Allowed while state.susp */ + mod_rq_state(req, m, + RQ_NET_OK|RQ_NET_PENDING|RQ_COMPLETION_SUSP, + RQ_NET_DONE); break; - case conflict_discarded_by_peer: - /* for discarded conflicting writes of multiple primaries, + case CONFLICT_RESOLVED: + /* for superseded conflicting writes of multiple primaries, * there is no need to keep anything in the tl, potential - * node crashes are covered by the activity log. */ - if (what == conflict_discarded_by_peer) - dev_alert(DEV, "Got DiscardAck packet %llus +%u!" - " DRBD is not a random data generator!\n", - (unsigned long long)req->sector, req->size); - req->rq_state |= RQ_NET_DONE; - /* fall through */ - case write_acked_by_peer_and_sis: - case write_acked_by_peer: - if (what == write_acked_by_peer_and_sis) - req->rq_state |= RQ_NET_SIS; + * node crashes are covered by the activity log. + * + * If this request had been marked as RQ_POSTPONED before, + * it will actually not be completed, but "restarted", + * resubmitted from the retry worker context. */ + D_ASSERT(req->rq_state & RQ_NET_PENDING); + D_ASSERT(req->rq_state & RQ_EXP_WRITE_ACK); + mod_rq_state(req, m, RQ_NET_PENDING, RQ_NET_DONE|RQ_NET_OK); + break; + + case WRITE_ACKED_BY_PEER_AND_SIS: + req->rq_state |= RQ_NET_SIS; + case WRITE_ACKED_BY_PEER: + D_ASSERT(req->rq_state & RQ_EXP_WRITE_ACK); /* protocol C; successfully written on peer. * Nothing more to do here. * We want to keep the tl in place for all protocols, to cater * for volatile write-back caches on lower level devices. */ - case recv_acked_by_peer: + goto ack_common; + case RECV_ACKED_BY_PEER: + D_ASSERT(req->rq_state & RQ_EXP_RECEIVE_ACK); /* protocol B; pretends to be successfully written on peer. - * see also notes above in handed_over_to_network about + * see also notes above in HANDED_OVER_TO_NETWORK about * protocol != C */ - req->rq_state |= RQ_NET_OK; + ack_common: D_ASSERT(req->rq_state & RQ_NET_PENDING); - dec_ap_pending(mdev); - atomic_sub(req->size>>9, &mdev->ap_in_flight); - req->rq_state &= ~RQ_NET_PENDING; - _req_may_be_done_not_susp(req, m); + mod_rq_state(req, m, RQ_NET_PENDING, RQ_NET_OK); break; - case neg_acked: - /* assert something? */ - if (req->rq_state & RQ_NET_PENDING) { - dec_ap_pending(mdev); - atomic_sub(req->size>>9, &mdev->ap_in_flight); - } - req->rq_state &= ~(RQ_NET_OK|RQ_NET_PENDING); + case POSTPONE_WRITE: + D_ASSERT(req->rq_state & RQ_EXP_WRITE_ACK); + /* If this node has already detected the write conflict, the + * worker will be waiting on misc_wait. Wake it up once this + * request has completed locally. + */ + D_ASSERT(req->rq_state & RQ_NET_PENDING); + req->rq_state |= RQ_POSTPONED; + if (req->i.waiting) + wake_up(&mdev->misc_wait); + /* Do not clear RQ_NET_PENDING. This request will make further + * progress via restart_conflicting_writes() or + * fail_postponed_requests(). Hopefully. */ + break; - req->rq_state |= RQ_NET_DONE; - _req_may_be_done_not_susp(req, m); - /* else: done by handed_over_to_network */ + case NEG_ACKED: + mod_rq_state(req, m, RQ_NET_OK|RQ_NET_PENDING, 0); break; - case fail_frozen_disk_io: + case FAIL_FROZEN_DISK_IO: if (!(req->rq_state & RQ_LOCAL_COMPLETED)) break; - - _req_may_be_done(req, m); /* Allowed while state.susp */ + mod_rq_state(req, m, RQ_COMPLETION_SUSP, 0); break; - case restart_frozen_disk_io: + case RESTART_FROZEN_DISK_IO: if (!(req->rq_state & RQ_LOCAL_COMPLETED)) break; - req->rq_state &= ~RQ_LOCAL_COMPLETED; + mod_rq_state(req, m, + RQ_COMPLETION_SUSP|RQ_LOCAL_COMPLETED, + RQ_LOCAL_PENDING); rv = MR_READ; if (bio_data_dir(req->master_bio) == WRITE) rv = MR_WRITE; - get_ldev(mdev); + get_ldev(mdev); /* always succeeds in this call path */ req->w.cb = w_restart_disk_io; - drbd_queue_work(&mdev->data.work, &req->w); + drbd_queue_work(&mdev->tconn->sender_work, &req->w); break; - case resend: + case RESEND: /* Simply complete (local only) READs. */ if (!(req->rq_state & RQ_WRITE) && !req->w.cb) { - _req_may_be_done(req, m); + mod_rq_state(req, m, RQ_COMPLETION_SUSP, 0); break; } /* If RQ_NET_OK is already set, we got a P_WRITE_ACK or P_RECV_ACK - before the connection loss (B&C only); only P_BARRIER_ACK was missing. - Trowing them out of the TL here by pretending we got a BARRIER_ACK - We ensure that the peer was not rebooted */ + before the connection loss (B&C only); only P_BARRIER_ACK + (or the local completion?) was missing when we suspended. + Throwing them out of the TL here by pretending we got a BARRIER_ACK. + During connection handshake, we ensure that the peer was not rebooted. */ if (!(req->rq_state & RQ_NET_OK)) { + /* FIXME could this possibly be a req->w.cb == w_send_out_of_sync? + * in that case we must not set RQ_NET_PENDING. */ + + mod_rq_state(req, m, RQ_COMPLETION_SUSP, RQ_NET_QUEUED|RQ_NET_PENDING); if (req->w.cb) { - drbd_queue_work(&mdev->data.work, &req->w); + drbd_queue_work(&mdev->tconn->sender_work, &req->w); rv = req->rq_state & RQ_WRITE ? MR_WRITE : MR_READ; - } + } /* else: FIXME can this happen? */ break; } - /* else, fall through to barrier_acked */ + /* else, fall through to BARRIER_ACKED */ - case barrier_acked: + case BARRIER_ACKED: + /* barrier ack for READ requests does not make sense */ if (!(req->rq_state & RQ_WRITE)) break; if (req->rq_state & RQ_NET_PENDING) { - /* barrier came in before all requests have been acked. + /* barrier came in before all requests were acked. * this is bad, because if the connection is lost now, * we won't be able to clean them up... */ - dev_err(DEV, "FIXME (barrier_acked but pending)\n"); - list_move(&req->tl_requests, &mdev->out_of_sequence_requests); + dev_err(DEV, "FIXME (BARRIER_ACKED but pending)\n"); } - if ((req->rq_state & RQ_NET_MASK) != 0) { - req->rq_state |= RQ_NET_DONE; - if (mdev->net_conf->wire_protocol == DRBD_PROT_A) - atomic_sub(req->size>>9, &mdev->ap_in_flight); - } - _req_may_be_done(req, m); /* Allowed while state.susp */ + /* Allowed to complete requests, even while suspended. + * As this is called for all requests within a matching epoch, + * we need to filter, and only set RQ_NET_DONE for those that + * have actually been on the wire. */ + mod_rq_state(req, m, RQ_COMPLETION_SUSP, + (req->rq_state & RQ_NET_MASK) ? RQ_NET_DONE : 0); break; - case data_received: + case DATA_RECEIVED: D_ASSERT(req->rq_state & RQ_NET_PENDING); - dec_ap_pending(mdev); - req->rq_state &= ~RQ_NET_PENDING; - req->rq_state |= (RQ_NET_OK|RQ_NET_DONE); - _req_may_be_done_not_susp(req, m); + mod_rq_state(req, m, RQ_NET_PENDING, RQ_NET_OK|RQ_NET_DONE); break; }; @@ -752,75 +767,265 @@ int __req_mod(struct drbd_request *req, enum drbd_req_event what, * since size may be bigger than BM_BLOCK_SIZE, * we may need to check several bits. */ -static int drbd_may_do_local_read(struct drbd_conf *mdev, sector_t sector, int size) +static bool drbd_may_do_local_read(struct drbd_conf *mdev, sector_t sector, int size) { unsigned long sbnr, ebnr; sector_t esector, nr_sectors; if (mdev->state.disk == D_UP_TO_DATE) - return 1; - if (mdev->state.disk >= D_OUTDATED) - return 0; - if (mdev->state.disk < D_INCONSISTENT) - return 0; - /* state.disk == D_INCONSISTENT We will have a look at the BitMap */ - nr_sectors = drbd_get_capacity(mdev->this_bdev); + return true; + if (mdev->state.disk != D_INCONSISTENT) + return false; esector = sector + (size >> 9) - 1; - + nr_sectors = drbd_get_capacity(mdev->this_bdev); D_ASSERT(sector < nr_sectors); D_ASSERT(esector < nr_sectors); sbnr = BM_SECT_TO_BIT(sector); ebnr = BM_SECT_TO_BIT(esector); - return 0 == drbd_bm_count_bits(mdev, sbnr, ebnr); + return drbd_bm_count_bits(mdev, sbnr, ebnr) == 0; +} + +static bool remote_due_to_read_balancing(struct drbd_conf *mdev, sector_t sector, + enum drbd_read_balancing rbm) +{ + struct backing_dev_info *bdi; + int stripe_shift; + + switch (rbm) { + case RB_CONGESTED_REMOTE: + bdi = &mdev->ldev->backing_bdev->bd_disk->queue->backing_dev_info; + return bdi_read_congested(bdi); + case RB_LEAST_PENDING: + return atomic_read(&mdev->local_cnt) > + atomic_read(&mdev->ap_pending_cnt) + atomic_read(&mdev->rs_pending_cnt); + case RB_32K_STRIPING: /* stripe_shift = 15 */ + case RB_64K_STRIPING: + case RB_128K_STRIPING: + case RB_256K_STRIPING: + case RB_512K_STRIPING: + case RB_1M_STRIPING: /* stripe_shift = 20 */ + stripe_shift = (rbm - RB_32K_STRIPING + 15); + return (sector >> (stripe_shift - 9)) & 1; + case RB_ROUND_ROBIN: + return test_and_change_bit(READ_BALANCE_RR, &mdev->flags); + case RB_PREFER_REMOTE: + return true; + case RB_PREFER_LOCAL: + default: + return false; + } +} + +/* + * complete_conflicting_writes - wait for any conflicting write requests + * + * The write_requests tree contains all active write requests which we + * currently know about. Wait for any requests to complete which conflict with + * the new one. + * + * Only way out: remove the conflicting intervals from the tree. + */ +static void complete_conflicting_writes(struct drbd_request *req) +{ + DEFINE_WAIT(wait); + struct drbd_conf *mdev = req->w.mdev; + struct drbd_interval *i; + sector_t sector = req->i.sector; + int size = req->i.size; + + i = drbd_find_overlap(&mdev->write_requests, sector, size); + if (!i) + return; + + for (;;) { + prepare_to_wait(&mdev->misc_wait, &wait, TASK_UNINTERRUPTIBLE); + i = drbd_find_overlap(&mdev->write_requests, sector, size); + if (!i) + break; + /* Indicate to wake up device->misc_wait on progress. */ + i->waiting = true; + spin_unlock_irq(&mdev->tconn->req_lock); + schedule(); + spin_lock_irq(&mdev->tconn->req_lock); + } + finish_wait(&mdev->misc_wait, &wait); } +/* called within req_lock and rcu_read_lock() */ static void maybe_pull_ahead(struct drbd_conf *mdev) { - int congested = 0; + struct drbd_tconn *tconn = mdev->tconn; + struct net_conf *nc; + bool congested = false; + enum drbd_on_congestion on_congestion; + + nc = rcu_dereference(tconn->net_conf); + on_congestion = nc ? nc->on_congestion : OC_BLOCK; + if (on_congestion == OC_BLOCK || + tconn->agreed_pro_version < 96) + return; /* If I don't even have good local storage, we can not reasonably try * to pull ahead of the peer. We also need the local reference to make * sure mdev->act_log is there. - * Note: caller has to make sure that net_conf is there. */ if (!get_ldev_if_state(mdev, D_UP_TO_DATE)) return; - if (mdev->net_conf->cong_fill && - atomic_read(&mdev->ap_in_flight) >= mdev->net_conf->cong_fill) { + if (nc->cong_fill && + atomic_read(&mdev->ap_in_flight) >= nc->cong_fill) { dev_info(DEV, "Congestion-fill threshold reached\n"); - congested = 1; + congested = true; } - if (mdev->act_log->used >= mdev->net_conf->cong_extents) { + if (mdev->act_log->used >= nc->cong_extents) { dev_info(DEV, "Congestion-extents threshold reached\n"); - congested = 1; + congested = true; } if (congested) { - queue_barrier(mdev); /* last barrier, after mirrored writes */ + /* start a new epoch for non-mirrored writes */ + start_new_tl_epoch(mdev->tconn); - if (mdev->net_conf->on_congestion == OC_PULL_AHEAD) + if (on_congestion == OC_PULL_AHEAD) _drbd_set_state(_NS(mdev, conn, C_AHEAD), 0, NULL); - else /*mdev->net_conf->on_congestion == OC_DISCONNECT */ + else /*nc->on_congestion == OC_DISCONNECT */ _drbd_set_state(_NS(mdev, conn, C_DISCONNECTING), 0, NULL); } put_ldev(mdev); } -static int drbd_make_request_common(struct drbd_conf *mdev, struct bio *bio, unsigned long start_time) +/* If this returns false, and req->private_bio is still set, + * this should be submitted locally. + * + * If it returns false, but req->private_bio is not set, + * we do not have access to good data :( + * + * Otherwise, this destroys req->private_bio, if any, + * and returns true. + */ +static bool do_remote_read(struct drbd_request *req) +{ + struct drbd_conf *mdev = req->w.mdev; + enum drbd_read_balancing rbm; + + if (req->private_bio) { + if (!drbd_may_do_local_read(mdev, + req->i.sector, req->i.size)) { + bio_put(req->private_bio); + req->private_bio = NULL; + put_ldev(mdev); + } + } + + if (mdev->state.pdsk != D_UP_TO_DATE) + return false; + + if (req->private_bio == NULL) + return true; + + /* TODO: improve read balancing decisions, take into account drbd + * protocol, pending requests etc. */ + + rcu_read_lock(); + rbm = rcu_dereference(mdev->ldev->disk_conf)->read_balancing; + rcu_read_unlock(); + + if (rbm == RB_PREFER_LOCAL && req->private_bio) + return false; /* submit locally */ + + if (remote_due_to_read_balancing(mdev, req->i.sector, rbm)) { + if (req->private_bio) { + bio_put(req->private_bio); + req->private_bio = NULL; + put_ldev(mdev); + } + return true; + } + + return false; +} + +/* returns number of connections (== 1, for drbd 8.4) + * expected to actually write this data, + * which does NOT include those that we are L_AHEAD for. */ +static int drbd_process_write_request(struct drbd_request *req) +{ + struct drbd_conf *mdev = req->w.mdev; + int remote, send_oos; + + rcu_read_lock(); + remote = drbd_should_do_remote(mdev->state); + if (remote) { + maybe_pull_ahead(mdev); + remote = drbd_should_do_remote(mdev->state); + } + send_oos = drbd_should_send_out_of_sync(mdev->state); + rcu_read_unlock(); + + /* Need to replicate writes. Unless it is an empty flush, + * which is better mapped to a DRBD P_BARRIER packet, + * also for drbd wire protocol compatibility reasons. + * If this was a flush, just start a new epoch. + * Unless the current epoch was empty anyways, or we are not currently + * replicating, in which case there is no point. */ + if (unlikely(req->i.size == 0)) { + /* The only size==0 bios we expect are empty flushes. */ + D_ASSERT(req->master_bio->bi_rw & REQ_FLUSH); + if (remote) + start_new_tl_epoch(mdev->tconn); + return 0; + } + + if (!remote && !send_oos) + return 0; + + D_ASSERT(!(remote && send_oos)); + + if (remote) { + _req_mod(req, TO_BE_SENT); + _req_mod(req, QUEUE_FOR_NET_WRITE); + } else if (drbd_set_out_of_sync(mdev, req->i.sector, req->i.size)) + _req_mod(req, QUEUE_FOR_SEND_OOS); + + return remote; +} + +static void +drbd_submit_req_private_bio(struct drbd_request *req) +{ + struct drbd_conf *mdev = req->w.mdev; + struct bio *bio = req->private_bio; + const int rw = bio_rw(bio); + + bio->bi_bdev = mdev->ldev->backing_bdev; + + /* State may have changed since we grabbed our reference on the + * ->ldev member. Double check, and short-circuit to endio. + * In case the last activity log transaction failed to get on + * stable storage, and this is a WRITE, we may not even submit + * this bio. */ + if (get_ldev(mdev)) { + if (drbd_insert_fault(mdev, + rw == WRITE ? DRBD_FAULT_DT_WR + : rw == READ ? DRBD_FAULT_DT_RD + : DRBD_FAULT_DT_RA)) + bio_endio(bio, -EIO); + else + generic_make_request(bio); + put_ldev(mdev); + } else + bio_endio(bio, -EIO); +} + +void __drbd_make_request(struct drbd_conf *mdev, struct bio *bio, unsigned long start_time) { const int rw = bio_rw(bio); - const int size = bio->bi_size; - const sector_t sector = bio->bi_sector; - struct drbd_tl_epoch *b = NULL; + struct bio_and_error m = { NULL, }; struct drbd_request *req; - int local, remote, send_oos = 0; - int err = -EIO; - int ret = 0; - union drbd_state s; + bool no_remote = false; /* allocate outside of all locks; */ req = drbd_req_new(mdev, bio); @@ -830,55 +1035,14 @@ static int drbd_make_request_common(struct drbd_conf *mdev, struct bio *bio, uns * if user cannot handle io errors, that's not our business. */ dev_err(DEV, "could not kmalloc() req\n"); bio_endio(bio, -ENOMEM); - return 0; + return; } req->start_time = start_time; - local = get_ldev(mdev); - if (!local) { - bio_put(req->private_bio); /* or we get a bio leak */ + if (!get_ldev(mdev)) { + bio_put(req->private_bio); req->private_bio = NULL; } - if (rw == WRITE) { - /* Need to replicate writes. Unless it is an empty flush, - * which is better mapped to a DRBD P_BARRIER packet, - * also for drbd wire protocol compatibility reasons. */ - if (unlikely(size == 0)) { - /* The only size==0 bios we expect are empty flushes. */ - D_ASSERT(bio->bi_rw & REQ_FLUSH); - remote = 0; - } else - remote = 1; - } else { - /* READ || READA */ - if (local) { - if (!drbd_may_do_local_read(mdev, sector, size)) { - /* we could kick the syncer to - * sync this extent asap, wait for - * it, then continue locally. - * Or just issue the request remotely. - */ - local = 0; - bio_put(req->private_bio); - req->private_bio = NULL; - put_ldev(mdev); - } - } - remote = !local && mdev->state.pdsk >= D_UP_TO_DATE; - } - - /* If we have a disk, but a READA request is mapped to remote, - * we are R_PRIMARY, D_INCONSISTENT, SyncTarget. - * Just fail that READA request right here. - * - * THINK: maybe fail all READA when not local? - * or make this configurable... - * if network is slow, READA won't do any good. - */ - if (rw == READA && mdev->state.disk >= D_INCONSISTENT && !local) { - err = -EWOULDBLOCK; - goto fail_and_free_req; - } /* For WRITES going to the local disk, grab a reference on the target * extent. This waits for any resync activity in the corresponding @@ -887,348 +1051,131 @@ static int drbd_make_request_common(struct drbd_conf *mdev, struct bio *bio, uns * of transactional on-disk meta data updates. * Empty flushes don't need to go into the activity log, they can only * flush data for pending writes which are already in there. */ - if (rw == WRITE && local && size + if (rw == WRITE && req->private_bio && req->i.size && !test_bit(AL_SUSPENDED, &mdev->flags)) { req->rq_state |= RQ_IN_ACT_LOG; - drbd_al_begin_io(mdev, sector); - } - - s = mdev->state; - remote = remote && drbd_should_do_remote(s); - send_oos = rw == WRITE && drbd_should_send_oos(s); - D_ASSERT(!(remote && send_oos)); - - if (!(local || remote) && !is_susp(mdev->state)) { - if (__ratelimit(&drbd_ratelimit_state)) - dev_err(DEV, "IO ERROR: neither local nor remote disk\n"); - goto fail_free_complete; + drbd_al_begin_io(mdev, &req->i); } - /* For WRITE request, we have to make sure that we have an - * unused_spare_tle, in case we need to start a new epoch. - * I try to be smart and avoid to pre-allocate always "just in case", - * but there is a race between testing the bit and pointer outside the - * spinlock, and grabbing the spinlock. - * if we lost that race, we retry. */ - if (rw == WRITE && (remote || send_oos) && - mdev->unused_spare_tle == NULL && - test_bit(CREATE_BARRIER, &mdev->flags)) { -allocate_barrier: - b = kmalloc(sizeof(struct drbd_tl_epoch), GFP_NOIO); - if (!b) { - dev_err(DEV, "Failed to alloc barrier.\n"); - err = -ENOMEM; - goto fail_free_complete; - } + spin_lock_irq(&mdev->tconn->req_lock); + if (rw == WRITE) { + /* This may temporarily give up the req_lock, + * but will re-aquire it before it returns here. + * Needs to be before the check on drbd_suspended() */ + complete_conflicting_writes(req); } - /* GOOD, everything prepared, grab the spin_lock */ - spin_lock_irq(&mdev->req_lock); - - if (is_susp(mdev->state)) { - /* If we got suspended, use the retry mechanism of - drbd_make_request() to restart processing of this - bio. In the next call to drbd_make_request - we sleep in inc_ap_bio() */ - ret = 1; - spin_unlock_irq(&mdev->req_lock); - goto fail_free_complete; - } + /* no more giving up req_lock from now on! */ - if (remote || send_oos) { - remote = drbd_should_do_remote(mdev->state); - send_oos = rw == WRITE && drbd_should_send_oos(mdev->state); - D_ASSERT(!(remote && send_oos)); - - if (!(remote || send_oos)) - dev_warn(DEV, "lost connection while grabbing the req_lock!\n"); - if (!(local || remote)) { - dev_err(DEV, "IO ERROR: neither local nor remote disk\n"); - spin_unlock_irq(&mdev->req_lock); - goto fail_free_complete; + if (drbd_suspended(mdev)) { + /* push back and retry: */ + req->rq_state |= RQ_POSTPONED; + if (req->private_bio) { + bio_put(req->private_bio); + req->private_bio = NULL; + put_ldev(mdev); } + goto out; } - if (b && mdev->unused_spare_tle == NULL) { - mdev->unused_spare_tle = b; - b = NULL; - } - if (rw == WRITE && (remote || send_oos) && - mdev->unused_spare_tle == NULL && - test_bit(CREATE_BARRIER, &mdev->flags)) { - /* someone closed the current epoch - * while we were grabbing the spinlock */ - spin_unlock_irq(&mdev->req_lock); - goto allocate_barrier; - } - - /* Update disk stats */ _drbd_start_io_acct(mdev, req, bio); - /* _maybe_start_new_epoch(mdev); - * If we need to generate a write barrier packet, we have to add the - * new epoch (barrier) object, and queue the barrier packet for sending, - * and queue the req's data after it _within the same lock_, otherwise - * we have race conditions were the reorder domains could be mixed up. - * - * Even read requests may start a new epoch and queue the corresponding - * barrier packet. To get the write ordering right, we only have to - * make sure that, if this is a write request and it triggered a - * barrier packet, this request is queued within the same spinlock. */ - if ((remote || send_oos) && mdev->unused_spare_tle && - test_and_clear_bit(CREATE_BARRIER, &mdev->flags)) { - _tl_add_barrier(mdev, mdev->unused_spare_tle); - mdev->unused_spare_tle = NULL; - } else { - D_ASSERT(!(remote && rw == WRITE && - test_bit(CREATE_BARRIER, &mdev->flags))); + /* We fail READ/READA early, if we can not serve it. + * We must do this before req is registered on any lists. + * Otherwise, drbd_req_complete() will queue failed READ for retry. */ + if (rw != WRITE) { + if (!do_remote_read(req) && !req->private_bio) + goto nodata; } - /* NOTE - * Actually, 'local' may be wrong here already, since we may have failed - * to write to the meta data, and may become wrong anytime because of - * local io-error for some other request, which would lead to us - * "detaching" the local disk. - * - * 'remote' may become wrong any time because the network could fail. - * - * This is a harmless race condition, though, since it is handled - * correctly at the appropriate places; so it just defers the failure - * of the respective operation. - */ - - /* mark them early for readability. - * this just sets some state flags. */ - if (remote) - _req_mod(req, to_be_send); - if (local) - _req_mod(req, to_be_submitted); - - /* check this request on the collision detection hash tables. - * if we have a conflict, just complete it here. - * THINK do we want to check reads, too? (I don't think so...) */ - if (rw == WRITE && _req_conflicts(req)) - goto fail_conflicting; + /* which transfer log epoch does this belong to? */ + req->epoch = atomic_read(&mdev->tconn->current_tle_nr); /* no point in adding empty flushes to the transfer log, * they are mapped to drbd barriers already. */ - if (likely(size!=0)) - list_add_tail(&req->tl_requests, &mdev->newest_tle->requests); + if (likely(req->i.size!=0)) { + if (rw == WRITE) + mdev->tconn->current_tle_writes++; - /* NOTE remote first: to get the concurrent write detection right, - * we must register the request before start of local IO. */ - if (remote) { - /* either WRITE and C_CONNECTED, - * or READ, and no local disk, - * or READ, but not in sync. - */ - _req_mod(req, (rw == WRITE) - ? queue_for_net_write - : queue_for_net_read); + list_add_tail(&req->tl_requests, &mdev->tconn->transfer_log); } - if (send_oos && drbd_set_out_of_sync(mdev, sector, size)) - _req_mod(req, queue_for_send_oos); - if (remote && - mdev->net_conf->on_congestion != OC_BLOCK && mdev->agreed_pro_version >= 96) - maybe_pull_ahead(mdev); - - /* If this was a flush, queue a drbd barrier/start a new epoch. - * Unless the current epoch was empty anyways, or we are not currently - * replicating, in which case there is no point. */ - if (unlikely(bio->bi_rw & REQ_FLUSH) - && mdev->newest_tle->n_writes - && drbd_should_do_remote(mdev->state)) - queue_barrier(mdev); - - spin_unlock_irq(&mdev->req_lock); - kfree(b); /* if someone else has beaten us to it... */ - - if (local) { - req->private_bio->bi_bdev = mdev->ldev->backing_bdev; - - /* State may have changed since we grabbed our reference on the - * mdev->ldev member. Double check, and short-circuit to endio. - * In case the last activity log transaction failed to get on - * stable storage, and this is a WRITE, we may not even submit - * this bio. */ - if (get_ldev(mdev)) { - if (drbd_insert_fault(mdev, rw == WRITE ? DRBD_FAULT_DT_WR - : rw == READ ? DRBD_FAULT_DT_RD - : DRBD_FAULT_DT_RA)) - bio_endio(req->private_bio, -EIO); - else - generic_make_request(req->private_bio); - put_ldev(mdev); + if (rw == WRITE) { + if (!drbd_process_write_request(req)) + no_remote = true; + } else { + /* We either have a private_bio, or we can read from remote. + * Otherwise we had done the goto nodata above. */ + if (req->private_bio == NULL) { + _req_mod(req, TO_BE_SENT); + _req_mod(req, QUEUE_FOR_NET_READ); } else - bio_endio(req->private_bio, -EIO); + no_remote = true; } - return 0; - -fail_conflicting: - /* this is a conflicting request. - * even though it may have been only _partially_ - * overlapping with one of the currently pending requests, - * without even submitting or sending it, we will - * pretend that it was successfully served right now. - */ - _drbd_end_io_acct(mdev, req); - spin_unlock_irq(&mdev->req_lock); - if (remote) - dec_ap_pending(mdev); - /* THINK: do we want to fail it (-EIO), or pretend success? - * this pretends success. */ - err = 0; - -fail_free_complete: - if (req->rq_state & RQ_IN_ACT_LOG) - drbd_al_complete_io(mdev, sector); -fail_and_free_req: - if (local) { - bio_put(req->private_bio); - req->private_bio = NULL; - put_ldev(mdev); + if (req->private_bio) { + /* needs to be marked within the same spinlock */ + _req_mod(req, TO_BE_SUBMITTED); + /* but we need to give up the spinlock to submit */ + spin_unlock_irq(&mdev->tconn->req_lock); + drbd_submit_req_private_bio(req); + spin_lock_irq(&mdev->tconn->req_lock); + } else if (no_remote) { +nodata: + if (__ratelimit(&drbd_ratelimit_state)) + dev_err(DEV, "IO ERROR: neither local nor remote data, sector %llu+%u\n", + (unsigned long long)req->i.sector, req->i.size >> 9); + /* A write may have been queued for send_oos, however. + * So we can not simply free it, we must go through drbd_req_put_completion_ref() */ } - if (!ret) - bio_endio(bio, err); - - drbd_req_free(req); - dec_ap_bio(mdev); - kfree(b); - - return ret; -} -/* helper function for drbd_make_request - * if we can determine just by the mdev (state) that this request will fail, - * return 1 - * otherwise return 0 - */ -static int drbd_fail_request_early(struct drbd_conf *mdev, int is_write) -{ - if (mdev->state.role != R_PRIMARY && - (!allow_oos || is_write)) { - if (__ratelimit(&drbd_ratelimit_state)) { - dev_err(DEV, "Process %s[%u] tried to %s; " - "since we are not in Primary state, " - "we cannot allow this\n", - current->comm, current->pid, - is_write ? "WRITE" : "READ"); - } - return 1; - } +out: + if (drbd_req_put_completion_ref(req, &m, 1)) + kref_put(&req->kref, drbd_req_destroy); + spin_unlock_irq(&mdev->tconn->req_lock); - return 0; + if (m.bio) + complete_master_bio(mdev, &m); + return; } void drbd_make_request(struct request_queue *q, struct bio *bio) { - unsigned int s_enr, e_enr; struct drbd_conf *mdev = (struct drbd_conf *) q->queuedata; unsigned long start_time; - if (drbd_fail_request_early(mdev, bio_data_dir(bio) & WRITE)) { - bio_endio(bio, -EPERM); - return; - } - start_time = jiffies; /* * what we "blindly" assume: */ - D_ASSERT((bio->bi_size & 0x1ff) == 0); - - /* to make some things easier, force alignment of requests within the - * granularity of our hash tables */ - s_enr = bio->bi_sector >> HT_SHIFT; - e_enr = bio->bi_size ? (bio->bi_sector+(bio->bi_size>>9)-1) >> HT_SHIFT : s_enr; - - if (likely(s_enr == e_enr)) { - do { - inc_ap_bio(mdev, 1); - } while (drbd_make_request_common(mdev, bio, start_time)); - return; - } - - /* can this bio be split generically? - * Maybe add our own split-arbitrary-bios function. */ - if (bio->bi_vcnt != 1 || bio->bi_idx != 0 || bio->bi_size > DRBD_MAX_BIO_SIZE) { - /* rather error out here than BUG in bio_split */ - dev_err(DEV, "bio would need to, but cannot, be split: " - "(vcnt=%u,idx=%u,size=%u,sector=%llu)\n", - bio->bi_vcnt, bio->bi_idx, bio->bi_size, - (unsigned long long)bio->bi_sector); - bio_endio(bio, -EINVAL); - } else { - /* This bio crosses some boundary, so we have to split it. */ - struct bio_pair *bp; - /* works for the "do not cross hash slot boundaries" case - * e.g. sector 262269, size 4096 - * s_enr = 262269 >> 6 = 4097 - * e_enr = (262269+8-1) >> 6 = 4098 - * HT_SHIFT = 6 - * sps = 64, mask = 63 - * first_sectors = 64 - (262269 & 63) = 3 - */ - const sector_t sect = bio->bi_sector; - const int sps = 1 << HT_SHIFT; /* sectors per slot */ - const int mask = sps - 1; - const sector_t first_sectors = sps - (sect & mask); - bp = bio_split(bio, first_sectors); + D_ASSERT(IS_ALIGNED(bio->bi_size, 512)); - /* we need to get a "reference count" (ap_bio_cnt) - * to avoid races with the disconnect/reconnect/suspend code. - * In case we need to split the bio here, we need to get three references - * atomically, otherwise we might deadlock when trying to submit the - * second one! */ - inc_ap_bio(mdev, 3); - - D_ASSERT(e_enr == s_enr + 1); - - while (drbd_make_request_common(mdev, &bp->bio1, start_time)) - inc_ap_bio(mdev, 1); - - while (drbd_make_request_common(mdev, &bp->bio2, start_time)) - inc_ap_bio(mdev, 1); - - dec_ap_bio(mdev); - - bio_pair_release(bp); - } + inc_ap_bio(mdev); + __drbd_make_request(mdev, bio, start_time); } -/* This is called by bio_add_page(). With this function we reduce - * the number of BIOs that span over multiple DRBD_MAX_BIO_SIZEs - * units (was AL_EXTENTs). +/* This is called by bio_add_page(). + * + * q->max_hw_sectors and other global limits are already enforced there. * - * we do the calculation within the lower 32bit of the byte offsets, - * since we don't care for actual offset, but only check whether it - * would cross "activity log extent" boundaries. + * We need to call down to our lower level device, + * in case it has special restrictions. + * + * We also may need to enforce configured max-bio-bvecs limits. * * As long as the BIO is empty we have to allow at least one bvec, - * regardless of size and offset. so the resulting bio may still - * cross extent boundaries. those are dealt with (bio_split) in - * drbd_make_request. + * regardless of size and offset, so no need to ask lower levels. */ int drbd_merge_bvec(struct request_queue *q, struct bvec_merge_data *bvm, struct bio_vec *bvec) { struct drbd_conf *mdev = (struct drbd_conf *) q->queuedata; - unsigned int bio_offset = - (unsigned int)bvm->bi_sector << 9; /* 32 bit */ unsigned int bio_size = bvm->bi_size; - int limit, backing_limit; - - limit = DRBD_MAX_BIO_SIZE - - ((bio_offset & (DRBD_MAX_BIO_SIZE-1)) + bio_size); - if (limit < 0) - limit = 0; - if (bio_size == 0) { - if (limit <= bvec->bv_len) - limit = bvec->bv_len; - } else if (limit && get_ldev(mdev)) { + int limit = DRBD_MAX_BIO_SIZE; + int backing_limit; + + if (bio_size && get_ldev(mdev)) { struct request_queue * const b = mdev->ldev->backing_bdev->bd_disk->queue; if (b->merge_bvec_fn) { @@ -1240,24 +1187,38 @@ int drbd_merge_bvec(struct request_queue *q, struct bvec_merge_data *bvm, struct return limit; } +struct drbd_request *find_oldest_request(struct drbd_tconn *tconn) +{ + /* Walk the transfer log, + * and find the oldest not yet completed request */ + struct drbd_request *r; + list_for_each_entry(r, &tconn->transfer_log, tl_requests) { + if (atomic_read(&r->completion_ref)) + return r; + } + return NULL; +} + void request_timer_fn(unsigned long data) { struct drbd_conf *mdev = (struct drbd_conf *) data; + struct drbd_tconn *tconn = mdev->tconn; struct drbd_request *req; /* oldest request */ - struct list_head *le; + struct net_conf *nc; unsigned long ent = 0, dt = 0, et, nt; /* effective timeout = ko_count * timeout */ unsigned long now; - if (get_net_conf(mdev)) { - if (mdev->state.conn >= C_WF_REPORT_PARAMS) - ent = mdev->net_conf->timeout*HZ/10 - * mdev->net_conf->ko_count; - put_net_conf(mdev); - } + rcu_read_lock(); + nc = rcu_dereference(tconn->net_conf); + if (nc && mdev->state.conn >= C_WF_REPORT_PARAMS) + ent = nc->timeout * HZ/10 * nc->ko_count; + if (get_ldev(mdev)) { /* implicit state.disk >= D_INCONSISTENT */ - dt = mdev->ldev->dc.disk_timeout * HZ / 10; + dt = rcu_dereference(mdev->ldev->disk_conf)->disk_timeout * HZ / 10; put_ldev(mdev); } + rcu_read_unlock(); + et = min_not_zero(dt, ent); if (!et) @@ -1265,17 +1226,14 @@ void request_timer_fn(unsigned long data) now = jiffies; - spin_lock_irq(&mdev->req_lock); - le = &mdev->oldest_tle->requests; - if (list_empty(le)) { - spin_unlock_irq(&mdev->req_lock); + spin_lock_irq(&tconn->req_lock); + req = find_oldest_request(tconn); + if (!req) { + spin_unlock_irq(&tconn->req_lock); mod_timer(&mdev->request_timer, now + et); return; } - le = le->prev; - req = list_entry(le, struct drbd_request, tl_requests); - /* The request is considered timed out, if * - we have some effective timeout from the configuration, * with above state restrictions applied, @@ -1294,17 +1252,17 @@ void request_timer_fn(unsigned long data) */ if (ent && req->rq_state & RQ_NET_PENDING && time_after(now, req->start_time + ent) && - !time_in_range(now, mdev->last_reconnect_jif, mdev->last_reconnect_jif + ent)) { + !time_in_range(now, tconn->last_reconnect_jif, tconn->last_reconnect_jif + ent)) { dev_warn(DEV, "Remote failed to finish a request within ko-count * timeout\n"); _drbd_set_state(_NS(mdev, conn, C_TIMEOUT), CS_VERBOSE | CS_HARD, NULL); } - if (dt && req->rq_state & RQ_LOCAL_PENDING && + if (dt && req->rq_state & RQ_LOCAL_PENDING && req->w.mdev == mdev && time_after(now, req->start_time + dt) && !time_in_range(now, mdev->last_reattach_jif, mdev->last_reattach_jif + dt)) { dev_warn(DEV, "Local backing device failed to meet the disk-timeout\n"); __drbd_chk_io_error(mdev, DRBD_FORCE_DETACH); } nt = (time_after(now, req->start_time + et) ? now : req->start_time) + et; - spin_unlock_irq(&mdev->req_lock); + spin_unlock_irq(&tconn->req_lock); mod_timer(&mdev->request_timer, nt); } diff --git a/drivers/block/drbd/drbd_req.h b/drivers/block/drbd/drbd_req.h index 3d2111919486..016de6b8bb57 100644 --- a/drivers/block/drbd/drbd_req.h +++ b/drivers/block/drbd/drbd_req.h @@ -77,40 +77,41 @@ */ enum drbd_req_event { - created, - to_be_send, - to_be_submitted, + CREATED, + TO_BE_SENT, + TO_BE_SUBMITTED, /* XXX yes, now I am inconsistent... * these are not "events" but "actions" * oh, well... */ - queue_for_net_write, - queue_for_net_read, - queue_for_send_oos, - - send_canceled, - send_failed, - handed_over_to_network, - oos_handed_to_network, - connection_lost_while_pending, - read_retry_remote_canceled, - recv_acked_by_peer, - write_acked_by_peer, - write_acked_by_peer_and_sis, /* and set_in_sync */ - conflict_discarded_by_peer, - neg_acked, - barrier_acked, /* in protocol A and B */ - data_received, /* (remote read) */ - - read_completed_with_error, - read_ahead_completed_with_error, - write_completed_with_error, - abort_disk_io, - completed_ok, - resend, - fail_frozen_disk_io, - restart_frozen_disk_io, - nothing, /* for tracing only */ + QUEUE_FOR_NET_WRITE, + QUEUE_FOR_NET_READ, + QUEUE_FOR_SEND_OOS, + + SEND_CANCELED, + SEND_FAILED, + HANDED_OVER_TO_NETWORK, + OOS_HANDED_TO_NETWORK, + CONNECTION_LOST_WHILE_PENDING, + READ_RETRY_REMOTE_CANCELED, + RECV_ACKED_BY_PEER, + WRITE_ACKED_BY_PEER, + WRITE_ACKED_BY_PEER_AND_SIS, /* and set_in_sync */ + CONFLICT_RESOLVED, + POSTPONE_WRITE, + NEG_ACKED, + BARRIER_ACKED, /* in protocol A and B */ + DATA_RECEIVED, /* (remote read) */ + + READ_COMPLETED_WITH_ERROR, + READ_AHEAD_COMPLETED_WITH_ERROR, + WRITE_COMPLETED_WITH_ERROR, + ABORT_DISK_IO, + COMPLETED_OK, + RESEND, + FAIL_FROZEN_DISK_IO, + RESTART_FROZEN_DISK_IO, + NOTHING, }; /* encoding of request states for now. we don't actually need that many bits. @@ -142,8 +143,8 @@ enum drbd_req_state_bits { * recv_ack (B) or implicit "ack" (A), * still waiting for the barrier ack. * master_bio may already be completed and invalidated. - * 11100: write_acked (C), - * data_received (for remote read, any protocol) + * 11100: write acked (C), + * data received (for remote read, any protocol) * or finally the barrier ack has arrived (B,A)... * request can be freed * 01100: neg-acked (write, protocol C) @@ -198,6 +199,22 @@ enum drbd_req_state_bits { /* Should call drbd_al_complete_io() for this request... */ __RQ_IN_ACT_LOG, + + /* The peer has sent a retry ACK */ + __RQ_POSTPONED, + + /* would have been completed, + * but was not, because of drbd_suspended() */ + __RQ_COMPLETION_SUSP, + + /* We expect a receive ACK (wire proto B) */ + __RQ_EXP_RECEIVE_ACK, + + /* We expect a write ACK (wite proto C) */ + __RQ_EXP_WRITE_ACK, + + /* waiting for a barrier ack, did an extra kref_get */ + __RQ_EXP_BARR_ACK, }; #define RQ_LOCAL_PENDING (1UL << __RQ_LOCAL_PENDING) @@ -219,56 +236,16 @@ enum drbd_req_state_bits { #define RQ_WRITE (1UL << __RQ_WRITE) #define RQ_IN_ACT_LOG (1UL << __RQ_IN_ACT_LOG) +#define RQ_POSTPONED (1UL << __RQ_POSTPONED) +#define RQ_COMPLETION_SUSP (1UL << __RQ_COMPLETION_SUSP) +#define RQ_EXP_RECEIVE_ACK (1UL << __RQ_EXP_RECEIVE_ACK) +#define RQ_EXP_WRITE_ACK (1UL << __RQ_EXP_WRITE_ACK) +#define RQ_EXP_BARR_ACK (1UL << __RQ_EXP_BARR_ACK) /* For waking up the frozen transfer log mod_req() has to return if the request should be counted in the epoch object*/ -#define MR_WRITE_SHIFT 0 -#define MR_WRITE (1 << MR_WRITE_SHIFT) -#define MR_READ_SHIFT 1 -#define MR_READ (1 << MR_READ_SHIFT) - -/* epoch entries */ -static inline -struct hlist_head *ee_hash_slot(struct drbd_conf *mdev, sector_t sector) -{ - BUG_ON(mdev->ee_hash_s == 0); - return mdev->ee_hash + - ((unsigned int)(sector>>HT_SHIFT) % mdev->ee_hash_s); -} - -/* transfer log (drbd_request objects) */ -static inline -struct hlist_head *tl_hash_slot(struct drbd_conf *mdev, sector_t sector) -{ - BUG_ON(mdev->tl_hash_s == 0); - return mdev->tl_hash + - ((unsigned int)(sector>>HT_SHIFT) % mdev->tl_hash_s); -} - -/* application reads (drbd_request objects) */ -static struct hlist_head *ar_hash_slot(struct drbd_conf *mdev, sector_t sector) -{ - return mdev->app_reads_hash - + ((unsigned int)(sector) % APP_R_HSIZE); -} - -/* when we receive the answer for a read request, - * verify that we actually know about it */ -static inline struct drbd_request *_ar_id_to_req(struct drbd_conf *mdev, - u64 id, sector_t sector) -{ - struct hlist_head *slot = ar_hash_slot(mdev, sector); - struct hlist_node *n; - struct drbd_request *req; - - hlist_for_each_entry(req, n, slot, collision) { - if ((unsigned long)req == (unsigned long)id) { - D_ASSERT(req->sector == sector); - return req; - } - } - return NULL; -} +#define MR_WRITE 1 +#define MR_READ 2 static inline void drbd_req_make_private_bio(struct drbd_request *req, struct bio *bio_src) { @@ -278,41 +255,10 @@ static inline void drbd_req_make_private_bio(struct drbd_request *req, struct bi req->private_bio = bio; bio->bi_private = req; - bio->bi_end_io = drbd_endio_pri; + bio->bi_end_io = drbd_request_endio; bio->bi_next = NULL; } -static inline struct drbd_request *drbd_req_new(struct drbd_conf *mdev, - struct bio *bio_src) -{ - struct drbd_request *req = - mempool_alloc(drbd_request_mempool, GFP_NOIO); - if (likely(req)) { - drbd_req_make_private_bio(req, bio_src); - - req->rq_state = bio_data_dir(bio_src) == WRITE ? RQ_WRITE : 0; - req->mdev = mdev; - req->master_bio = bio_src; - req->epoch = 0; - req->sector = bio_src->bi_sector; - req->size = bio_src->bi_size; - INIT_HLIST_NODE(&req->collision); - INIT_LIST_HEAD(&req->tl_requests); - INIT_LIST_HEAD(&req->w.list); - } - return req; -} - -static inline void drbd_req_free(struct drbd_request *req) -{ - mempool_free(req, drbd_request_mempool); -} - -static inline int overlaps(sector_t s1, int l1, sector_t s2, int l2) -{ - return !((s1 + (l1>>9) <= s2) || (s1 >= s2 + (l2>>9))); -} - /* Short lived temporary struct on the stack. * We could squirrel the error to be returned into * bio->bi_size, or similar. But that would be too ugly. */ @@ -321,6 +267,7 @@ struct bio_and_error { int error; }; +extern void drbd_req_destroy(struct kref *kref); extern void _req_may_be_done(struct drbd_request *req, struct bio_and_error *m); extern int __req_mod(struct drbd_request *req, enum drbd_req_event what, @@ -328,13 +275,17 @@ extern int __req_mod(struct drbd_request *req, enum drbd_req_event what, extern void complete_master_bio(struct drbd_conf *mdev, struct bio_and_error *m); extern void request_timer_fn(unsigned long data); -extern void tl_restart(struct drbd_conf *mdev, enum drbd_req_event what); +extern void tl_restart(struct drbd_tconn *tconn, enum drbd_req_event what); +extern void _tl_restart(struct drbd_tconn *tconn, enum drbd_req_event what); + +/* this is in drbd_main.c */ +extern void drbd_restart_request(struct drbd_request *req); /* use this if you don't want to deal with calling complete_master_bio() * outside the spinlock, e.g. when walking some list on cleanup. */ static inline int _req_mod(struct drbd_request *req, enum drbd_req_event what) { - struct drbd_conf *mdev = req->mdev; + struct drbd_conf *mdev = req->w.mdev; struct bio_and_error m; int rv; @@ -354,13 +305,13 @@ static inline int req_mod(struct drbd_request *req, enum drbd_req_event what) { unsigned long flags; - struct drbd_conf *mdev = req->mdev; + struct drbd_conf *mdev = req->w.mdev; struct bio_and_error m; int rv; - spin_lock_irqsave(&mdev->req_lock, flags); + spin_lock_irqsave(&mdev->tconn->req_lock, flags); rv = __req_mod(req, what, &m); - spin_unlock_irqrestore(&mdev->req_lock, flags); + spin_unlock_irqrestore(&mdev->tconn->req_lock, flags); if (m.bio) complete_master_bio(mdev, &m); @@ -368,7 +319,7 @@ static inline int req_mod(struct drbd_request *req, return rv; } -static inline bool drbd_should_do_remote(union drbd_state s) +static inline bool drbd_should_do_remote(union drbd_dev_state s) { return s.pdsk == D_UP_TO_DATE || (s.pdsk >= D_INCONSISTENT && @@ -378,7 +329,7 @@ static inline bool drbd_should_do_remote(union drbd_state s) That is equivalent since before 96 IO was frozen in the C_WF_BITMAP* states. */ } -static inline bool drbd_should_send_oos(union drbd_state s) +static inline bool drbd_should_send_out_of_sync(union drbd_dev_state s) { return s.conn == C_AHEAD || s.conn == C_WF_BITMAP_S; /* pdsk = D_INCONSISTENT as a consequence. Protocol 96 check not necessary diff --git a/drivers/block/drbd/drbd_state.c b/drivers/block/drbd/drbd_state.c new file mode 100644 index 000000000000..53bf6182bac4 --- /dev/null +++ b/drivers/block/drbd/drbd_state.c @@ -0,0 +1,1856 @@ +/* + drbd_state.c + + This file is part of DRBD by Philipp Reisner and Lars Ellenberg. + + Copyright (C) 2001-2008, LINBIT Information Technologies GmbH. + Copyright (C) 1999-2008, Philipp Reisner <philipp.reisner@linbit.com>. + Copyright (C) 2002-2008, Lars Ellenberg <lars.ellenberg@linbit.com>. + + Thanks to Carter Burden, Bart Grantham and Gennadiy Nerubayev + from Logicworks, Inc. for making SDP replication support possible. + + drbd is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2, or (at your option) + any later version. + + drbd is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with drbd; see the file COPYING. If not, write to + the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. + */ + +#include <linux/drbd_limits.h> +#include "drbd_int.h" +#include "drbd_req.h" + +/* in drbd_main.c */ +extern void tl_abort_disk_io(struct drbd_conf *mdev); + +struct after_state_chg_work { + struct drbd_work w; + union drbd_state os; + union drbd_state ns; + enum chg_state_flags flags; + struct completion *done; +}; + +enum sanitize_state_warnings { + NO_WARNING, + ABORTED_ONLINE_VERIFY, + ABORTED_RESYNC, + CONNECTION_LOST_NEGOTIATING, + IMPLICITLY_UPGRADED_DISK, + IMPLICITLY_UPGRADED_PDSK, +}; + +static int w_after_state_ch(struct drbd_work *w, int unused); +static void after_state_ch(struct drbd_conf *mdev, union drbd_state os, + union drbd_state ns, enum chg_state_flags flags); +static enum drbd_state_rv is_valid_state(struct drbd_conf *, union drbd_state); +static enum drbd_state_rv is_valid_soft_transition(union drbd_state, union drbd_state, struct drbd_tconn *); +static enum drbd_state_rv is_valid_transition(union drbd_state os, union drbd_state ns); +static union drbd_state sanitize_state(struct drbd_conf *mdev, union drbd_state ns, + enum sanitize_state_warnings *warn); + +static inline bool is_susp(union drbd_state s) +{ + return s.susp || s.susp_nod || s.susp_fen; +} + +bool conn_all_vols_unconf(struct drbd_tconn *tconn) +{ + struct drbd_conf *mdev; + bool rv = true; + int vnr; + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) { + if (mdev->state.disk != D_DISKLESS || + mdev->state.conn != C_STANDALONE || + mdev->state.role != R_SECONDARY) { + rv = false; + break; + } + } + rcu_read_unlock(); + + return rv; +} + +/* Unfortunately the states where not correctly ordered, when + they where defined. therefore can not use max_t() here. */ +static enum drbd_role max_role(enum drbd_role role1, enum drbd_role role2) +{ + if (role1 == R_PRIMARY || role2 == R_PRIMARY) + return R_PRIMARY; + if (role1 == R_SECONDARY || role2 == R_SECONDARY) + return R_SECONDARY; + return R_UNKNOWN; +} +static enum drbd_role min_role(enum drbd_role role1, enum drbd_role role2) +{ + if (role1 == R_UNKNOWN || role2 == R_UNKNOWN) + return R_UNKNOWN; + if (role1 == R_SECONDARY || role2 == R_SECONDARY) + return R_SECONDARY; + return R_PRIMARY; +} + +enum drbd_role conn_highest_role(struct drbd_tconn *tconn) +{ + enum drbd_role role = R_UNKNOWN; + struct drbd_conf *mdev; + int vnr; + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) + role = max_role(role, mdev->state.role); + rcu_read_unlock(); + + return role; +} + +enum drbd_role conn_highest_peer(struct drbd_tconn *tconn) +{ + enum drbd_role peer = R_UNKNOWN; + struct drbd_conf *mdev; + int vnr; + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) + peer = max_role(peer, mdev->state.peer); + rcu_read_unlock(); + + return peer; +} + +enum drbd_disk_state conn_highest_disk(struct drbd_tconn *tconn) +{ + enum drbd_disk_state ds = D_DISKLESS; + struct drbd_conf *mdev; + int vnr; + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) + ds = max_t(enum drbd_disk_state, ds, mdev->state.disk); + rcu_read_unlock(); + + return ds; +} + +enum drbd_disk_state conn_lowest_disk(struct drbd_tconn *tconn) +{ + enum drbd_disk_state ds = D_MASK; + struct drbd_conf *mdev; + int vnr; + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) + ds = min_t(enum drbd_disk_state, ds, mdev->state.disk); + rcu_read_unlock(); + + return ds; +} + +enum drbd_disk_state conn_highest_pdsk(struct drbd_tconn *tconn) +{ + enum drbd_disk_state ds = D_DISKLESS; + struct drbd_conf *mdev; + int vnr; + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) + ds = max_t(enum drbd_disk_state, ds, mdev->state.pdsk); + rcu_read_unlock(); + + return ds; +} + +enum drbd_conns conn_lowest_conn(struct drbd_tconn *tconn) +{ + enum drbd_conns conn = C_MASK; + struct drbd_conf *mdev; + int vnr; + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) + conn = min_t(enum drbd_conns, conn, mdev->state.conn); + rcu_read_unlock(); + + return conn; +} + +static bool no_peer_wf_report_params(struct drbd_tconn *tconn) +{ + struct drbd_conf *mdev; + int vnr; + bool rv = true; + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) + if (mdev->state.conn == C_WF_REPORT_PARAMS) { + rv = false; + break; + } + rcu_read_unlock(); + + return rv; +} + + +/** + * cl_wide_st_chg() - true if the state change is a cluster wide one + * @mdev: DRBD device. + * @os: old (current) state. + * @ns: new (wanted) state. + */ +static int cl_wide_st_chg(struct drbd_conf *mdev, + union drbd_state os, union drbd_state ns) +{ + return (os.conn >= C_CONNECTED && ns.conn >= C_CONNECTED && + ((os.role != R_PRIMARY && ns.role == R_PRIMARY) || + (os.conn != C_STARTING_SYNC_T && ns.conn == C_STARTING_SYNC_T) || + (os.conn != C_STARTING_SYNC_S && ns.conn == C_STARTING_SYNC_S) || + (os.disk != D_FAILED && ns.disk == D_FAILED))) || + (os.conn >= C_CONNECTED && ns.conn == C_DISCONNECTING) || + (os.conn == C_CONNECTED && ns.conn == C_VERIFY_S) || + (os.conn == C_CONNECTED && ns.conn == C_WF_REPORT_PARAMS); +} + +static union drbd_state +apply_mask_val(union drbd_state os, union drbd_state mask, union drbd_state val) +{ + union drbd_state ns; + ns.i = (os.i & ~mask.i) | val.i; + return ns; +} + +enum drbd_state_rv +drbd_change_state(struct drbd_conf *mdev, enum chg_state_flags f, + union drbd_state mask, union drbd_state val) +{ + unsigned long flags; + union drbd_state ns; + enum drbd_state_rv rv; + + spin_lock_irqsave(&mdev->tconn->req_lock, flags); + ns = apply_mask_val(drbd_read_state(mdev), mask, val); + rv = _drbd_set_state(mdev, ns, f, NULL); + spin_unlock_irqrestore(&mdev->tconn->req_lock, flags); + + return rv; +} + +/** + * drbd_force_state() - Impose a change which happens outside our control on our state + * @mdev: DRBD device. + * @mask: mask of state bits to change. + * @val: value of new state bits. + */ +void drbd_force_state(struct drbd_conf *mdev, + union drbd_state mask, union drbd_state val) +{ + drbd_change_state(mdev, CS_HARD, mask, val); +} + +static enum drbd_state_rv +_req_st_cond(struct drbd_conf *mdev, union drbd_state mask, + union drbd_state val) +{ + union drbd_state os, ns; + unsigned long flags; + enum drbd_state_rv rv; + + if (test_and_clear_bit(CL_ST_CHG_SUCCESS, &mdev->flags)) + return SS_CW_SUCCESS; + + if (test_and_clear_bit(CL_ST_CHG_FAIL, &mdev->flags)) + return SS_CW_FAILED_BY_PEER; + + spin_lock_irqsave(&mdev->tconn->req_lock, flags); + os = drbd_read_state(mdev); + ns = sanitize_state(mdev, apply_mask_val(os, mask, val), NULL); + rv = is_valid_transition(os, ns); + if (rv >= SS_SUCCESS) + rv = SS_UNKNOWN_ERROR; /* cont waiting, otherwise fail. */ + + if (!cl_wide_st_chg(mdev, os, ns)) + rv = SS_CW_NO_NEED; + if (rv == SS_UNKNOWN_ERROR) { + rv = is_valid_state(mdev, ns); + if (rv >= SS_SUCCESS) { + rv = is_valid_soft_transition(os, ns, mdev->tconn); + if (rv >= SS_SUCCESS) + rv = SS_UNKNOWN_ERROR; /* cont waiting, otherwise fail. */ + } + } + spin_unlock_irqrestore(&mdev->tconn->req_lock, flags); + + return rv; +} + +/** + * drbd_req_state() - Perform an eventually cluster wide state change + * @mdev: DRBD device. + * @mask: mask of state bits to change. + * @val: value of new state bits. + * @f: flags + * + * Should not be called directly, use drbd_request_state() or + * _drbd_request_state(). + */ +static enum drbd_state_rv +drbd_req_state(struct drbd_conf *mdev, union drbd_state mask, + union drbd_state val, enum chg_state_flags f) +{ + struct completion done; + unsigned long flags; + union drbd_state os, ns; + enum drbd_state_rv rv; + + init_completion(&done); + + if (f & CS_SERIALIZE) + mutex_lock(mdev->state_mutex); + + spin_lock_irqsave(&mdev->tconn->req_lock, flags); + os = drbd_read_state(mdev); + ns = sanitize_state(mdev, apply_mask_val(os, mask, val), NULL); + rv = is_valid_transition(os, ns); + if (rv < SS_SUCCESS) { + spin_unlock_irqrestore(&mdev->tconn->req_lock, flags); + goto abort; + } + + if (cl_wide_st_chg(mdev, os, ns)) { + rv = is_valid_state(mdev, ns); + if (rv == SS_SUCCESS) + rv = is_valid_soft_transition(os, ns, mdev->tconn); + spin_unlock_irqrestore(&mdev->tconn->req_lock, flags); + + if (rv < SS_SUCCESS) { + if (f & CS_VERBOSE) + print_st_err(mdev, os, ns, rv); + goto abort; + } + + if (drbd_send_state_req(mdev, mask, val)) { + rv = SS_CW_FAILED_BY_PEER; + if (f & CS_VERBOSE) + print_st_err(mdev, os, ns, rv); + goto abort; + } + + wait_event(mdev->state_wait, + (rv = _req_st_cond(mdev, mask, val))); + + if (rv < SS_SUCCESS) { + if (f & CS_VERBOSE) + print_st_err(mdev, os, ns, rv); + goto abort; + } + spin_lock_irqsave(&mdev->tconn->req_lock, flags); + ns = apply_mask_val(drbd_read_state(mdev), mask, val); + rv = _drbd_set_state(mdev, ns, f, &done); + } else { + rv = _drbd_set_state(mdev, ns, f, &done); + } + + spin_unlock_irqrestore(&mdev->tconn->req_lock, flags); + + if (f & CS_WAIT_COMPLETE && rv == SS_SUCCESS) { + D_ASSERT(current != mdev->tconn->worker.task); + wait_for_completion(&done); + } + +abort: + if (f & CS_SERIALIZE) + mutex_unlock(mdev->state_mutex); + + return rv; +} + +/** + * _drbd_request_state() - Request a state change (with flags) + * @mdev: DRBD device. + * @mask: mask of state bits to change. + * @val: value of new state bits. + * @f: flags + * + * Cousin of drbd_request_state(), useful with the CS_WAIT_COMPLETE + * flag, or when logging of failed state change requests is not desired. + */ +enum drbd_state_rv +_drbd_request_state(struct drbd_conf *mdev, union drbd_state mask, + union drbd_state val, enum chg_state_flags f) +{ + enum drbd_state_rv rv; + + wait_event(mdev->state_wait, + (rv = drbd_req_state(mdev, mask, val, f)) != SS_IN_TRANSIENT_STATE); + + return rv; +} + +static void print_st(struct drbd_conf *mdev, char *name, union drbd_state ns) +{ + dev_err(DEV, " %s = { cs:%s ro:%s/%s ds:%s/%s %c%c%c%c%c%c }\n", + name, + drbd_conn_str(ns.conn), + drbd_role_str(ns.role), + drbd_role_str(ns.peer), + drbd_disk_str(ns.disk), + drbd_disk_str(ns.pdsk), + is_susp(ns) ? 's' : 'r', + ns.aftr_isp ? 'a' : '-', + ns.peer_isp ? 'p' : '-', + ns.user_isp ? 'u' : '-', + ns.susp_fen ? 'F' : '-', + ns.susp_nod ? 'N' : '-' + ); +} + +void print_st_err(struct drbd_conf *mdev, union drbd_state os, + union drbd_state ns, enum drbd_state_rv err) +{ + if (err == SS_IN_TRANSIENT_STATE) + return; + dev_err(DEV, "State change failed: %s\n", drbd_set_st_err_str(err)); + print_st(mdev, " state", os); + print_st(mdev, "wanted", ns); +} + +static long print_state_change(char *pb, union drbd_state os, union drbd_state ns, + enum chg_state_flags flags) +{ + char *pbp; + pbp = pb; + *pbp = 0; + + if (ns.role != os.role && flags & CS_DC_ROLE) + pbp += sprintf(pbp, "role( %s -> %s ) ", + drbd_role_str(os.role), + drbd_role_str(ns.role)); + if (ns.peer != os.peer && flags & CS_DC_PEER) + pbp += sprintf(pbp, "peer( %s -> %s ) ", + drbd_role_str(os.peer), + drbd_role_str(ns.peer)); + if (ns.conn != os.conn && flags & CS_DC_CONN) + pbp += sprintf(pbp, "conn( %s -> %s ) ", + drbd_conn_str(os.conn), + drbd_conn_str(ns.conn)); + if (ns.disk != os.disk && flags & CS_DC_DISK) + pbp += sprintf(pbp, "disk( %s -> %s ) ", + drbd_disk_str(os.disk), + drbd_disk_str(ns.disk)); + if (ns.pdsk != os.pdsk && flags & CS_DC_PDSK) + pbp += sprintf(pbp, "pdsk( %s -> %s ) ", + drbd_disk_str(os.pdsk), + drbd_disk_str(ns.pdsk)); + + return pbp - pb; +} + +static void drbd_pr_state_change(struct drbd_conf *mdev, union drbd_state os, union drbd_state ns, + enum chg_state_flags flags) +{ + char pb[300]; + char *pbp = pb; + + pbp += print_state_change(pbp, os, ns, flags ^ CS_DC_MASK); + + if (ns.aftr_isp != os.aftr_isp) + pbp += sprintf(pbp, "aftr_isp( %d -> %d ) ", + os.aftr_isp, + ns.aftr_isp); + if (ns.peer_isp != os.peer_isp) + pbp += sprintf(pbp, "peer_isp( %d -> %d ) ", + os.peer_isp, + ns.peer_isp); + if (ns.user_isp != os.user_isp) + pbp += sprintf(pbp, "user_isp( %d -> %d ) ", + os.user_isp, + ns.user_isp); + + if (pbp != pb) + dev_info(DEV, "%s\n", pb); +} + +static void conn_pr_state_change(struct drbd_tconn *tconn, union drbd_state os, union drbd_state ns, + enum chg_state_flags flags) +{ + char pb[300]; + char *pbp = pb; + + pbp += print_state_change(pbp, os, ns, flags); + + if (is_susp(ns) != is_susp(os) && flags & CS_DC_SUSP) + pbp += sprintf(pbp, "susp( %d -> %d ) ", + is_susp(os), + is_susp(ns)); + + if (pbp != pb) + conn_info(tconn, "%s\n", pb); +} + + +/** + * is_valid_state() - Returns an SS_ error code if ns is not valid + * @mdev: DRBD device. + * @ns: State to consider. + */ +static enum drbd_state_rv +is_valid_state(struct drbd_conf *mdev, union drbd_state ns) +{ + /* See drbd_state_sw_errors in drbd_strings.c */ + + enum drbd_fencing_p fp; + enum drbd_state_rv rv = SS_SUCCESS; + struct net_conf *nc; + + rcu_read_lock(); + fp = FP_DONT_CARE; + if (get_ldev(mdev)) { + fp = rcu_dereference(mdev->ldev->disk_conf)->fencing; + put_ldev(mdev); + } + + nc = rcu_dereference(mdev->tconn->net_conf); + if (nc) { + if (!nc->two_primaries && ns.role == R_PRIMARY) { + if (ns.peer == R_PRIMARY) + rv = SS_TWO_PRIMARIES; + else if (conn_highest_peer(mdev->tconn) == R_PRIMARY) + rv = SS_O_VOL_PEER_PRI; + } + } + + if (rv <= 0) + /* already found a reason to abort */; + else if (ns.role == R_SECONDARY && mdev->open_cnt) + rv = SS_DEVICE_IN_USE; + + else if (ns.role == R_PRIMARY && ns.conn < C_CONNECTED && ns.disk < D_UP_TO_DATE) + rv = SS_NO_UP_TO_DATE_DISK; + + else if (fp >= FP_RESOURCE && + ns.role == R_PRIMARY && ns.conn < C_CONNECTED && ns.pdsk >= D_UNKNOWN) + rv = SS_PRIMARY_NOP; + + else if (ns.role == R_PRIMARY && ns.disk <= D_INCONSISTENT && ns.pdsk <= D_INCONSISTENT) + rv = SS_NO_UP_TO_DATE_DISK; + + else if (ns.conn > C_CONNECTED && ns.disk < D_INCONSISTENT) + rv = SS_NO_LOCAL_DISK; + + else if (ns.conn > C_CONNECTED && ns.pdsk < D_INCONSISTENT) + rv = SS_NO_REMOTE_DISK; + + else if (ns.conn > C_CONNECTED && ns.disk < D_UP_TO_DATE && ns.pdsk < D_UP_TO_DATE) + rv = SS_NO_UP_TO_DATE_DISK; + + else if ((ns.conn == C_CONNECTED || + ns.conn == C_WF_BITMAP_S || + ns.conn == C_SYNC_SOURCE || + ns.conn == C_PAUSED_SYNC_S) && + ns.disk == D_OUTDATED) + rv = SS_CONNECTED_OUTDATES; + + else if ((ns.conn == C_VERIFY_S || ns.conn == C_VERIFY_T) && + (nc->verify_alg[0] == 0)) + rv = SS_NO_VERIFY_ALG; + + else if ((ns.conn == C_VERIFY_S || ns.conn == C_VERIFY_T) && + mdev->tconn->agreed_pro_version < 88) + rv = SS_NOT_SUPPORTED; + + else if (ns.conn >= C_CONNECTED && ns.pdsk == D_UNKNOWN) + rv = SS_CONNECTED_OUTDATES; + + rcu_read_unlock(); + + return rv; +} + +/** + * is_valid_soft_transition() - Returns an SS_ error code if the state transition is not possible + * This function limits state transitions that may be declined by DRBD. I.e. + * user requests (aka soft transitions). + * @mdev: DRBD device. + * @ns: new state. + * @os: old state. + */ +static enum drbd_state_rv +is_valid_soft_transition(union drbd_state os, union drbd_state ns, struct drbd_tconn *tconn) +{ + enum drbd_state_rv rv = SS_SUCCESS; + + if ((ns.conn == C_STARTING_SYNC_T || ns.conn == C_STARTING_SYNC_S) && + os.conn > C_CONNECTED) + rv = SS_RESYNC_RUNNING; + + if (ns.conn == C_DISCONNECTING && os.conn == C_STANDALONE) + rv = SS_ALREADY_STANDALONE; + + if (ns.disk > D_ATTACHING && os.disk == D_DISKLESS) + rv = SS_IS_DISKLESS; + + if (ns.conn == C_WF_CONNECTION && os.conn < C_UNCONNECTED) + rv = SS_NO_NET_CONFIG; + + if (ns.disk == D_OUTDATED && os.disk < D_OUTDATED && os.disk != D_ATTACHING) + rv = SS_LOWER_THAN_OUTDATED; + + if (ns.conn == C_DISCONNECTING && os.conn == C_UNCONNECTED) + rv = SS_IN_TRANSIENT_STATE; + + /* if (ns.conn == os.conn && ns.conn == C_WF_REPORT_PARAMS) + rv = SS_IN_TRANSIENT_STATE; */ + + /* While establishing a connection only allow cstate to change. + Delay/refuse role changes, detach attach etc... */ + if (test_bit(STATE_SENT, &tconn->flags) && + !(os.conn == C_WF_REPORT_PARAMS || + (ns.conn == C_WF_REPORT_PARAMS && os.conn == C_WF_CONNECTION))) + rv = SS_IN_TRANSIENT_STATE; + + if ((ns.conn == C_VERIFY_S || ns.conn == C_VERIFY_T) && os.conn < C_CONNECTED) + rv = SS_NEED_CONNECTION; + + if ((ns.conn == C_VERIFY_S || ns.conn == C_VERIFY_T) && + ns.conn != os.conn && os.conn > C_CONNECTED) + rv = SS_RESYNC_RUNNING; + + if ((ns.conn == C_STARTING_SYNC_S || ns.conn == C_STARTING_SYNC_T) && + os.conn < C_CONNECTED) + rv = SS_NEED_CONNECTION; + + if ((ns.conn == C_SYNC_TARGET || ns.conn == C_SYNC_SOURCE) + && os.conn < C_WF_REPORT_PARAMS) + rv = SS_NEED_CONNECTION; /* No NetworkFailure -> SyncTarget etc... */ + + return rv; +} + +static enum drbd_state_rv +is_valid_conn_transition(enum drbd_conns oc, enum drbd_conns nc) +{ + /* no change -> nothing to do, at least for the connection part */ + if (oc == nc) + return SS_NOTHING_TO_DO; + + /* disconnect of an unconfigured connection does not make sense */ + if (oc == C_STANDALONE && nc == C_DISCONNECTING) + return SS_ALREADY_STANDALONE; + + /* from C_STANDALONE, we start with C_UNCONNECTED */ + if (oc == C_STANDALONE && nc != C_UNCONNECTED) + return SS_NEED_CONNECTION; + + /* When establishing a connection we need to go through WF_REPORT_PARAMS! + Necessary to do the right thing upon invalidate-remote on a disconnected resource */ + if (oc < C_WF_REPORT_PARAMS && nc >= C_CONNECTED) + return SS_NEED_CONNECTION; + + /* After a network error only C_UNCONNECTED or C_DISCONNECTING may follow. */ + if (oc >= C_TIMEOUT && oc <= C_TEAR_DOWN && nc != C_UNCONNECTED && nc != C_DISCONNECTING) + return SS_IN_TRANSIENT_STATE; + + /* After C_DISCONNECTING only C_STANDALONE may follow */ + if (oc == C_DISCONNECTING && nc != C_STANDALONE) + return SS_IN_TRANSIENT_STATE; + + return SS_SUCCESS; +} + + +/** + * is_valid_transition() - Returns an SS_ error code if the state transition is not possible + * This limits hard state transitions. Hard state transitions are facts there are + * imposed on DRBD by the environment. E.g. disk broke or network broke down. + * But those hard state transitions are still not allowed to do everything. + * @ns: new state. + * @os: old state. + */ +static enum drbd_state_rv +is_valid_transition(union drbd_state os, union drbd_state ns) +{ + enum drbd_state_rv rv; + + rv = is_valid_conn_transition(os.conn, ns.conn); + + /* we cannot fail (again) if we already detached */ + if (ns.disk == D_FAILED && os.disk == D_DISKLESS) + rv = SS_IS_DISKLESS; + + return rv; +} + +static void print_sanitize_warnings(struct drbd_conf *mdev, enum sanitize_state_warnings warn) +{ + static const char *msg_table[] = { + [NO_WARNING] = "", + [ABORTED_ONLINE_VERIFY] = "Online-verify aborted.", + [ABORTED_RESYNC] = "Resync aborted.", + [CONNECTION_LOST_NEGOTIATING] = "Connection lost while negotiating, no data!", + [IMPLICITLY_UPGRADED_DISK] = "Implicitly upgraded disk", + [IMPLICITLY_UPGRADED_PDSK] = "Implicitly upgraded pdsk", + }; + + if (warn != NO_WARNING) + dev_warn(DEV, "%s\n", msg_table[warn]); +} + +/** + * sanitize_state() - Resolves implicitly necessary additional changes to a state transition + * @mdev: DRBD device. + * @os: old state. + * @ns: new state. + * @warn_sync_abort: + * + * When we loose connection, we have to set the state of the peers disk (pdsk) + * to D_UNKNOWN. This rule and many more along those lines are in this function. + */ +static union drbd_state sanitize_state(struct drbd_conf *mdev, union drbd_state ns, + enum sanitize_state_warnings *warn) +{ + enum drbd_fencing_p fp; + enum drbd_disk_state disk_min, disk_max, pdsk_min, pdsk_max; + + if (warn) + *warn = NO_WARNING; + + fp = FP_DONT_CARE; + if (get_ldev(mdev)) { + rcu_read_lock(); + fp = rcu_dereference(mdev->ldev->disk_conf)->fencing; + rcu_read_unlock(); + put_ldev(mdev); + } + + /* Implications from connection to peer and peer_isp */ + if (ns.conn < C_CONNECTED) { + ns.peer_isp = 0; + ns.peer = R_UNKNOWN; + if (ns.pdsk > D_UNKNOWN || ns.pdsk < D_INCONSISTENT) + ns.pdsk = D_UNKNOWN; + } + + /* Clear the aftr_isp when becoming unconfigured */ + if (ns.conn == C_STANDALONE && ns.disk == D_DISKLESS && ns.role == R_SECONDARY) + ns.aftr_isp = 0; + + /* An implication of the disk states onto the connection state */ + /* Abort resync if a disk fails/detaches */ + if (ns.conn > C_CONNECTED && (ns.disk <= D_FAILED || ns.pdsk <= D_FAILED)) { + if (warn) + *warn = ns.conn == C_VERIFY_S || ns.conn == C_VERIFY_T ? + ABORTED_ONLINE_VERIFY : ABORTED_RESYNC; + ns.conn = C_CONNECTED; + } + + /* Connection breaks down before we finished "Negotiating" */ + if (ns.conn < C_CONNECTED && ns.disk == D_NEGOTIATING && + get_ldev_if_state(mdev, D_NEGOTIATING)) { + if (mdev->ed_uuid == mdev->ldev->md.uuid[UI_CURRENT]) { + ns.disk = mdev->new_state_tmp.disk; + ns.pdsk = mdev->new_state_tmp.pdsk; + } else { + if (warn) + *warn = CONNECTION_LOST_NEGOTIATING; + ns.disk = D_DISKLESS; + ns.pdsk = D_UNKNOWN; + } + put_ldev(mdev); + } + + /* D_CONSISTENT and D_OUTDATED vanish when we get connected */ + if (ns.conn >= C_CONNECTED && ns.conn < C_AHEAD) { + if (ns.disk == D_CONSISTENT || ns.disk == D_OUTDATED) + ns.disk = D_UP_TO_DATE; + if (ns.pdsk == D_CONSISTENT || ns.pdsk == D_OUTDATED) + ns.pdsk = D_UP_TO_DATE; + } + + /* Implications of the connection stat on the disk states */ + disk_min = D_DISKLESS; + disk_max = D_UP_TO_DATE; + pdsk_min = D_INCONSISTENT; + pdsk_max = D_UNKNOWN; + switch ((enum drbd_conns)ns.conn) { + case C_WF_BITMAP_T: + case C_PAUSED_SYNC_T: + case C_STARTING_SYNC_T: + case C_WF_SYNC_UUID: + case C_BEHIND: + disk_min = D_INCONSISTENT; + disk_max = D_OUTDATED; + pdsk_min = D_UP_TO_DATE; + pdsk_max = D_UP_TO_DATE; + break; + case C_VERIFY_S: + case C_VERIFY_T: + disk_min = D_UP_TO_DATE; + disk_max = D_UP_TO_DATE; + pdsk_min = D_UP_TO_DATE; + pdsk_max = D_UP_TO_DATE; + break; + case C_CONNECTED: + disk_min = D_DISKLESS; + disk_max = D_UP_TO_DATE; + pdsk_min = D_DISKLESS; + pdsk_max = D_UP_TO_DATE; + break; + case C_WF_BITMAP_S: + case C_PAUSED_SYNC_S: + case C_STARTING_SYNC_S: + case C_AHEAD: + disk_min = D_UP_TO_DATE; + disk_max = D_UP_TO_DATE; + pdsk_min = D_INCONSISTENT; + pdsk_max = D_CONSISTENT; /* D_OUTDATED would be nice. But explicit outdate necessary*/ + break; + case C_SYNC_TARGET: + disk_min = D_INCONSISTENT; + disk_max = D_INCONSISTENT; + pdsk_min = D_UP_TO_DATE; + pdsk_max = D_UP_TO_DATE; + break; + case C_SYNC_SOURCE: + disk_min = D_UP_TO_DATE; + disk_max = D_UP_TO_DATE; + pdsk_min = D_INCONSISTENT; + pdsk_max = D_INCONSISTENT; + break; + case C_STANDALONE: + case C_DISCONNECTING: + case C_UNCONNECTED: + case C_TIMEOUT: + case C_BROKEN_PIPE: + case C_NETWORK_FAILURE: + case C_PROTOCOL_ERROR: + case C_TEAR_DOWN: + case C_WF_CONNECTION: + case C_WF_REPORT_PARAMS: + case C_MASK: + break; + } + if (ns.disk > disk_max) + ns.disk = disk_max; + + if (ns.disk < disk_min) { + if (warn) + *warn = IMPLICITLY_UPGRADED_DISK; + ns.disk = disk_min; + } + if (ns.pdsk > pdsk_max) + ns.pdsk = pdsk_max; + + if (ns.pdsk < pdsk_min) { + if (warn) + *warn = IMPLICITLY_UPGRADED_PDSK; + ns.pdsk = pdsk_min; + } + + if (fp == FP_STONITH && + (ns.role == R_PRIMARY && ns.conn < C_CONNECTED && ns.pdsk > D_OUTDATED)) + ns.susp_fen = 1; /* Suspend IO while fence-peer handler runs (peer lost) */ + + if (mdev->tconn->res_opts.on_no_data == OND_SUSPEND_IO && + (ns.role == R_PRIMARY && ns.disk < D_UP_TO_DATE && ns.pdsk < D_UP_TO_DATE)) + ns.susp_nod = 1; /* Suspend IO while no data available (no accessible data available) */ + + if (ns.aftr_isp || ns.peer_isp || ns.user_isp) { + if (ns.conn == C_SYNC_SOURCE) + ns.conn = C_PAUSED_SYNC_S; + if (ns.conn == C_SYNC_TARGET) + ns.conn = C_PAUSED_SYNC_T; + } else { + if (ns.conn == C_PAUSED_SYNC_S) + ns.conn = C_SYNC_SOURCE; + if (ns.conn == C_PAUSED_SYNC_T) + ns.conn = C_SYNC_TARGET; + } + + return ns; +} + +void drbd_resume_al(struct drbd_conf *mdev) +{ + if (test_and_clear_bit(AL_SUSPENDED, &mdev->flags)) + dev_info(DEV, "Resumed AL updates\n"); +} + +/* helper for __drbd_set_state */ +static void set_ov_position(struct drbd_conf *mdev, enum drbd_conns cs) +{ + if (mdev->tconn->agreed_pro_version < 90) + mdev->ov_start_sector = 0; + mdev->rs_total = drbd_bm_bits(mdev); + mdev->ov_position = 0; + if (cs == C_VERIFY_T) { + /* starting online verify from an arbitrary position + * does not fit well into the existing protocol. + * on C_VERIFY_T, we initialize ov_left and friends + * implicitly in receive_DataRequest once the + * first P_OV_REQUEST is received */ + mdev->ov_start_sector = ~(sector_t)0; + } else { + unsigned long bit = BM_SECT_TO_BIT(mdev->ov_start_sector); + if (bit >= mdev->rs_total) { + mdev->ov_start_sector = + BM_BIT_TO_SECT(mdev->rs_total - 1); + mdev->rs_total = 1; + } else + mdev->rs_total -= bit; + mdev->ov_position = mdev->ov_start_sector; + } + mdev->ov_left = mdev->rs_total; +} + +/** + * __drbd_set_state() - Set a new DRBD state + * @mdev: DRBD device. + * @ns: new state. + * @flags: Flags + * @done: Optional completion, that will get completed after the after_state_ch() finished + * + * Caller needs to hold req_lock, and global_state_lock. Do not call directly. + */ +enum drbd_state_rv +__drbd_set_state(struct drbd_conf *mdev, union drbd_state ns, + enum chg_state_flags flags, struct completion *done) +{ + union drbd_state os; + enum drbd_state_rv rv = SS_SUCCESS; + enum sanitize_state_warnings ssw; + struct after_state_chg_work *ascw; + + os = drbd_read_state(mdev); + + ns = sanitize_state(mdev, ns, &ssw); + if (ns.i == os.i) + return SS_NOTHING_TO_DO; + + rv = is_valid_transition(os, ns); + if (rv < SS_SUCCESS) + return rv; + + if (!(flags & CS_HARD)) { + /* pre-state-change checks ; only look at ns */ + /* See drbd_state_sw_errors in drbd_strings.c */ + + rv = is_valid_state(mdev, ns); + if (rv < SS_SUCCESS) { + /* If the old state was illegal as well, then let + this happen...*/ + + if (is_valid_state(mdev, os) == rv) + rv = is_valid_soft_transition(os, ns, mdev->tconn); + } else + rv = is_valid_soft_transition(os, ns, mdev->tconn); + } + + if (rv < SS_SUCCESS) { + if (flags & CS_VERBOSE) + print_st_err(mdev, os, ns, rv); + return rv; + } + + print_sanitize_warnings(mdev, ssw); + + drbd_pr_state_change(mdev, os, ns, flags); + + /* Display changes to the susp* flags that where caused by the call to + sanitize_state(). Only display it here if we where not called from + _conn_request_state() */ + if (!(flags & CS_DC_SUSP)) + conn_pr_state_change(mdev->tconn, os, ns, (flags & ~CS_DC_MASK) | CS_DC_SUSP); + + /* if we are going -> D_FAILED or D_DISKLESS, grab one extra reference + * on the ldev here, to be sure the transition -> D_DISKLESS resp. + * drbd_ldev_destroy() won't happen before our corresponding + * after_state_ch works run, where we put_ldev again. */ + if ((os.disk != D_FAILED && ns.disk == D_FAILED) || + (os.disk != D_DISKLESS && ns.disk == D_DISKLESS)) + atomic_inc(&mdev->local_cnt); + + mdev->state.i = ns.i; + mdev->tconn->susp = ns.susp; + mdev->tconn->susp_nod = ns.susp_nod; + mdev->tconn->susp_fen = ns.susp_fen; + + if (os.disk == D_ATTACHING && ns.disk >= D_NEGOTIATING) + drbd_print_uuids(mdev, "attached to UUIDs"); + + /* Wake up role changes, that were delayed because of connection establishing */ + if (os.conn == C_WF_REPORT_PARAMS && ns.conn != C_WF_REPORT_PARAMS && + no_peer_wf_report_params(mdev->tconn)) + clear_bit(STATE_SENT, &mdev->tconn->flags); + + wake_up(&mdev->misc_wait); + wake_up(&mdev->state_wait); + wake_up(&mdev->tconn->ping_wait); + + /* Aborted verify run, or we reached the stop sector. + * Log the last position, unless end-of-device. */ + if ((os.conn == C_VERIFY_S || os.conn == C_VERIFY_T) && + ns.conn <= C_CONNECTED) { + mdev->ov_start_sector = + BM_BIT_TO_SECT(drbd_bm_bits(mdev) - mdev->ov_left); + if (mdev->ov_left) + dev_info(DEV, "Online Verify reached sector %llu\n", + (unsigned long long)mdev->ov_start_sector); + } + + if ((os.conn == C_PAUSED_SYNC_T || os.conn == C_PAUSED_SYNC_S) && + (ns.conn == C_SYNC_TARGET || ns.conn == C_SYNC_SOURCE)) { + dev_info(DEV, "Syncer continues.\n"); + mdev->rs_paused += (long)jiffies + -(long)mdev->rs_mark_time[mdev->rs_last_mark]; + if (ns.conn == C_SYNC_TARGET) + mod_timer(&mdev->resync_timer, jiffies); + } + + if ((os.conn == C_SYNC_TARGET || os.conn == C_SYNC_SOURCE) && + (ns.conn == C_PAUSED_SYNC_T || ns.conn == C_PAUSED_SYNC_S)) { + dev_info(DEV, "Resync suspended\n"); + mdev->rs_mark_time[mdev->rs_last_mark] = jiffies; + } + + if (os.conn == C_CONNECTED && + (ns.conn == C_VERIFY_S || ns.conn == C_VERIFY_T)) { + unsigned long now = jiffies; + int i; + + set_ov_position(mdev, ns.conn); + mdev->rs_start = now; + mdev->rs_last_events = 0; + mdev->rs_last_sect_ev = 0; + mdev->ov_last_oos_size = 0; + mdev->ov_last_oos_start = 0; + + for (i = 0; i < DRBD_SYNC_MARKS; i++) { + mdev->rs_mark_left[i] = mdev->ov_left; + mdev->rs_mark_time[i] = now; + } + + drbd_rs_controller_reset(mdev); + + if (ns.conn == C_VERIFY_S) { + dev_info(DEV, "Starting Online Verify from sector %llu\n", + (unsigned long long)mdev->ov_position); + mod_timer(&mdev->resync_timer, jiffies); + } + } + + if (get_ldev(mdev)) { + u32 mdf = mdev->ldev->md.flags & ~(MDF_CONSISTENT|MDF_PRIMARY_IND| + MDF_CONNECTED_IND|MDF_WAS_UP_TO_DATE| + MDF_PEER_OUT_DATED|MDF_CRASHED_PRIMARY); + + mdf &= ~MDF_AL_CLEAN; + if (test_bit(CRASHED_PRIMARY, &mdev->flags)) + mdf |= MDF_CRASHED_PRIMARY; + if (mdev->state.role == R_PRIMARY || + (mdev->state.pdsk < D_INCONSISTENT && mdev->state.peer == R_PRIMARY)) + mdf |= MDF_PRIMARY_IND; + if (mdev->state.conn > C_WF_REPORT_PARAMS) + mdf |= MDF_CONNECTED_IND; + if (mdev->state.disk > D_INCONSISTENT) + mdf |= MDF_CONSISTENT; + if (mdev->state.disk > D_OUTDATED) + mdf |= MDF_WAS_UP_TO_DATE; + if (mdev->state.pdsk <= D_OUTDATED && mdev->state.pdsk >= D_INCONSISTENT) + mdf |= MDF_PEER_OUT_DATED; + if (mdf != mdev->ldev->md.flags) { + mdev->ldev->md.flags = mdf; + drbd_md_mark_dirty(mdev); + } + if (os.disk < D_CONSISTENT && ns.disk >= D_CONSISTENT) + drbd_set_ed_uuid(mdev, mdev->ldev->md.uuid[UI_CURRENT]); + put_ldev(mdev); + } + + /* Peer was forced D_UP_TO_DATE & R_PRIMARY, consider to resync */ + if (os.disk == D_INCONSISTENT && os.pdsk == D_INCONSISTENT && + os.peer == R_SECONDARY && ns.peer == R_PRIMARY) + set_bit(CONSIDER_RESYNC, &mdev->flags); + + /* Receiver should clean up itself */ + if (os.conn != C_DISCONNECTING && ns.conn == C_DISCONNECTING) + drbd_thread_stop_nowait(&mdev->tconn->receiver); + + /* Now the receiver finished cleaning up itself, it should die */ + if (os.conn != C_STANDALONE && ns.conn == C_STANDALONE) + drbd_thread_stop_nowait(&mdev->tconn->receiver); + + /* Upon network failure, we need to restart the receiver. */ + if (os.conn > C_WF_CONNECTION && + ns.conn <= C_TEAR_DOWN && ns.conn >= C_TIMEOUT) + drbd_thread_restart_nowait(&mdev->tconn->receiver); + + /* Resume AL writing if we get a connection */ + if (os.conn < C_CONNECTED && ns.conn >= C_CONNECTED) + drbd_resume_al(mdev); + + /* remember last attach time so request_timer_fn() won't + * kill newly established sessions while we are still trying to thaw + * previously frozen IO */ + if ((os.disk == D_ATTACHING || os.disk == D_NEGOTIATING) && + ns.disk > D_NEGOTIATING) + mdev->last_reattach_jif = jiffies; + + ascw = kmalloc(sizeof(*ascw), GFP_ATOMIC); + if (ascw) { + ascw->os = os; + ascw->ns = ns; + ascw->flags = flags; + ascw->w.cb = w_after_state_ch; + ascw->w.mdev = mdev; + ascw->done = done; + drbd_queue_work(&mdev->tconn->sender_work, &ascw->w); + } else { + dev_err(DEV, "Could not kmalloc an ascw\n"); + } + + return rv; +} + +static int w_after_state_ch(struct drbd_work *w, int unused) +{ + struct after_state_chg_work *ascw = + container_of(w, struct after_state_chg_work, w); + struct drbd_conf *mdev = w->mdev; + + after_state_ch(mdev, ascw->os, ascw->ns, ascw->flags); + if (ascw->flags & CS_WAIT_COMPLETE) { + D_ASSERT(ascw->done != NULL); + complete(ascw->done); + } + kfree(ascw); + + return 0; +} + +static void abw_start_sync(struct drbd_conf *mdev, int rv) +{ + if (rv) { + dev_err(DEV, "Writing the bitmap failed not starting resync.\n"); + _drbd_request_state(mdev, NS(conn, C_CONNECTED), CS_VERBOSE); + return; + } + + switch (mdev->state.conn) { + case C_STARTING_SYNC_T: + _drbd_request_state(mdev, NS(conn, C_WF_SYNC_UUID), CS_VERBOSE); + break; + case C_STARTING_SYNC_S: + drbd_start_resync(mdev, C_SYNC_SOURCE); + break; + } +} + +int drbd_bitmap_io_from_worker(struct drbd_conf *mdev, + int (*io_fn)(struct drbd_conf *), + char *why, enum bm_flag flags) +{ + int rv; + + D_ASSERT(current == mdev->tconn->worker.task); + + /* open coded non-blocking drbd_suspend_io(mdev); */ + set_bit(SUSPEND_IO, &mdev->flags); + + drbd_bm_lock(mdev, why, flags); + rv = io_fn(mdev); + drbd_bm_unlock(mdev); + + drbd_resume_io(mdev); + + return rv; +} + +/** + * after_state_ch() - Perform after state change actions that may sleep + * @mdev: DRBD device. + * @os: old state. + * @ns: new state. + * @flags: Flags + */ +static void after_state_ch(struct drbd_conf *mdev, union drbd_state os, + union drbd_state ns, enum chg_state_flags flags) +{ + struct sib_info sib; + + sib.sib_reason = SIB_STATE_CHANGE; + sib.os = os; + sib.ns = ns; + + if (os.conn != C_CONNECTED && ns.conn == C_CONNECTED) { + clear_bit(CRASHED_PRIMARY, &mdev->flags); + if (mdev->p_uuid) + mdev->p_uuid[UI_FLAGS] &= ~((u64)2); + } + + /* Inform userspace about the change... */ + drbd_bcast_event(mdev, &sib); + + if (!(os.role == R_PRIMARY && os.disk < D_UP_TO_DATE && os.pdsk < D_UP_TO_DATE) && + (ns.role == R_PRIMARY && ns.disk < D_UP_TO_DATE && ns.pdsk < D_UP_TO_DATE)) + drbd_khelper(mdev, "pri-on-incon-degr"); + + /* Here we have the actions that are performed after a + state change. This function might sleep */ + + if (ns.susp_nod) { + struct drbd_tconn *tconn = mdev->tconn; + enum drbd_req_event what = NOTHING; + + spin_lock_irq(&tconn->req_lock); + if (os.conn < C_CONNECTED && conn_lowest_conn(tconn) >= C_CONNECTED) + what = RESEND; + + if ((os.disk == D_ATTACHING || os.disk == D_NEGOTIATING) && + conn_lowest_disk(tconn) > D_NEGOTIATING) + what = RESTART_FROZEN_DISK_IO; + + if (tconn->susp_nod && what != NOTHING) { + _tl_restart(tconn, what); + _conn_request_state(tconn, + (union drbd_state) { { .susp_nod = 1 } }, + (union drbd_state) { { .susp_nod = 0 } }, + CS_VERBOSE); + } + spin_unlock_irq(&tconn->req_lock); + } + + if (ns.susp_fen) { + struct drbd_tconn *tconn = mdev->tconn; + + spin_lock_irq(&tconn->req_lock); + if (tconn->susp_fen && conn_lowest_conn(tconn) >= C_CONNECTED) { + /* case2: The connection was established again: */ + struct drbd_conf *odev; + int vnr; + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, odev, vnr) + clear_bit(NEW_CUR_UUID, &odev->flags); + rcu_read_unlock(); + _tl_restart(tconn, RESEND); + _conn_request_state(tconn, + (union drbd_state) { { .susp_fen = 1 } }, + (union drbd_state) { { .susp_fen = 0 } }, + CS_VERBOSE); + } + spin_unlock_irq(&tconn->req_lock); + } + + /* Became sync source. With protocol >= 96, we still need to send out + * the sync uuid now. Need to do that before any drbd_send_state, or + * the other side may go "paused sync" before receiving the sync uuids, + * which is unexpected. */ + if ((os.conn != C_SYNC_SOURCE && os.conn != C_PAUSED_SYNC_S) && + (ns.conn == C_SYNC_SOURCE || ns.conn == C_PAUSED_SYNC_S) && + mdev->tconn->agreed_pro_version >= 96 && get_ldev(mdev)) { + drbd_gen_and_send_sync_uuid(mdev); + put_ldev(mdev); + } + + /* Do not change the order of the if above and the two below... */ + if (os.pdsk == D_DISKLESS && + ns.pdsk > D_DISKLESS && ns.pdsk != D_UNKNOWN) { /* attach on the peer */ + /* we probably will start a resync soon. + * make sure those things are properly reset. */ + mdev->rs_total = 0; + mdev->rs_failed = 0; + atomic_set(&mdev->rs_pending_cnt, 0); + drbd_rs_cancel_all(mdev); + + drbd_send_uuids(mdev); + drbd_send_state(mdev, ns); + } + /* No point in queuing send_bitmap if we don't have a connection + * anymore, so check also the _current_ state, not only the new state + * at the time this work was queued. */ + if (os.conn != C_WF_BITMAP_S && ns.conn == C_WF_BITMAP_S && + mdev->state.conn == C_WF_BITMAP_S) + drbd_queue_bitmap_io(mdev, &drbd_send_bitmap, NULL, + "send_bitmap (WFBitMapS)", + BM_LOCKED_TEST_ALLOWED); + + /* Lost contact to peer's copy of the data */ + if ((os.pdsk >= D_INCONSISTENT && + os.pdsk != D_UNKNOWN && + os.pdsk != D_OUTDATED) + && (ns.pdsk < D_INCONSISTENT || + ns.pdsk == D_UNKNOWN || + ns.pdsk == D_OUTDATED)) { + if (get_ldev(mdev)) { + if ((ns.role == R_PRIMARY || ns.peer == R_PRIMARY) && + mdev->ldev->md.uuid[UI_BITMAP] == 0 && ns.disk >= D_UP_TO_DATE) { + if (drbd_suspended(mdev)) { + set_bit(NEW_CUR_UUID, &mdev->flags); + } else { + drbd_uuid_new_current(mdev); + drbd_send_uuids(mdev); + } + } + put_ldev(mdev); + } + } + + if (ns.pdsk < D_INCONSISTENT && get_ldev(mdev)) { + if (os.peer == R_SECONDARY && ns.peer == R_PRIMARY && + mdev->ldev->md.uuid[UI_BITMAP] == 0 && ns.disk >= D_UP_TO_DATE) { + drbd_uuid_new_current(mdev); + drbd_send_uuids(mdev); + } + /* D_DISKLESS Peer becomes secondary */ + if (os.peer == R_PRIMARY && ns.peer == R_SECONDARY) + /* We may still be Primary ourselves. + * No harm done if the bitmap still changes, + * redirtied pages will follow later. */ + drbd_bitmap_io_from_worker(mdev, &drbd_bm_write, + "demote diskless peer", BM_LOCKED_SET_ALLOWED); + put_ldev(mdev); + } + + /* Write out all changed bits on demote. + * Though, no need to da that just yet + * if there is a resync going on still */ + if (os.role == R_PRIMARY && ns.role == R_SECONDARY && + mdev->state.conn <= C_CONNECTED && get_ldev(mdev)) { + /* No changes to the bitmap expected this time, so assert that, + * even though no harm was done if it did change. */ + drbd_bitmap_io_from_worker(mdev, &drbd_bm_write, + "demote", BM_LOCKED_TEST_ALLOWED); + put_ldev(mdev); + } + + /* Last part of the attaching process ... */ + if (ns.conn >= C_CONNECTED && + os.disk == D_ATTACHING && ns.disk == D_NEGOTIATING) { + drbd_send_sizes(mdev, 0, 0); /* to start sync... */ + drbd_send_uuids(mdev); + drbd_send_state(mdev, ns); + } + + /* We want to pause/continue resync, tell peer. */ + if (ns.conn >= C_CONNECTED && + ((os.aftr_isp != ns.aftr_isp) || + (os.user_isp != ns.user_isp))) + drbd_send_state(mdev, ns); + + /* In case one of the isp bits got set, suspend other devices. */ + if ((!os.aftr_isp && !os.peer_isp && !os.user_isp) && + (ns.aftr_isp || ns.peer_isp || ns.user_isp)) + suspend_other_sg(mdev); + + /* Make sure the peer gets informed about eventual state + changes (ISP bits) while we were in WFReportParams. */ + if (os.conn == C_WF_REPORT_PARAMS && ns.conn >= C_CONNECTED) + drbd_send_state(mdev, ns); + + if (os.conn != C_AHEAD && ns.conn == C_AHEAD) + drbd_send_state(mdev, ns); + + /* We are in the progress to start a full sync... */ + if ((os.conn != C_STARTING_SYNC_T && ns.conn == C_STARTING_SYNC_T) || + (os.conn != C_STARTING_SYNC_S && ns.conn == C_STARTING_SYNC_S)) + /* no other bitmap changes expected during this phase */ + drbd_queue_bitmap_io(mdev, + &drbd_bmio_set_n_write, &abw_start_sync, + "set_n_write from StartingSync", BM_LOCKED_TEST_ALLOWED); + + /* We are invalidating our self... */ + if (os.conn < C_CONNECTED && ns.conn < C_CONNECTED && + os.disk > D_INCONSISTENT && ns.disk == D_INCONSISTENT) + /* other bitmap operation expected during this phase */ + drbd_queue_bitmap_io(mdev, &drbd_bmio_set_n_write, NULL, + "set_n_write from invalidate", BM_LOCKED_MASK); + + /* first half of local IO error, failure to attach, + * or administrative detach */ + if (os.disk != D_FAILED && ns.disk == D_FAILED) { + enum drbd_io_error_p eh = EP_PASS_ON; + int was_io_error = 0; + /* corresponding get_ldev was in __drbd_set_state, to serialize + * our cleanup here with the transition to D_DISKLESS. + * But is is still not save to dreference ldev here, since + * we might come from an failed Attach before ldev was set. */ + if (mdev->ldev) { + rcu_read_lock(); + eh = rcu_dereference(mdev->ldev->disk_conf)->on_io_error; + rcu_read_unlock(); + + was_io_error = test_and_clear_bit(WAS_IO_ERROR, &mdev->flags); + + if (was_io_error && eh == EP_CALL_HELPER) + drbd_khelper(mdev, "local-io-error"); + + /* Immediately allow completion of all application IO, + * that waits for completion from the local disk, + * if this was a force-detach due to disk_timeout + * or administrator request (drbdsetup detach --force). + * Do NOT abort otherwise. + * Aborting local requests may cause serious problems, + * if requests are completed to upper layers already, + * and then later the already submitted local bio completes. + * This can cause DMA into former bio pages that meanwhile + * have been re-used for other things. + * So aborting local requests may cause crashes, + * or even worse, silent data corruption. + */ + if (test_and_clear_bit(FORCE_DETACH, &mdev->flags)) + tl_abort_disk_io(mdev); + + /* current state still has to be D_FAILED, + * there is only one way out: to D_DISKLESS, + * and that may only happen after our put_ldev below. */ + if (mdev->state.disk != D_FAILED) + dev_err(DEV, + "ASSERT FAILED: disk is %s during detach\n", + drbd_disk_str(mdev->state.disk)); + + if (ns.conn >= C_CONNECTED) + drbd_send_state(mdev, ns); + + drbd_rs_cancel_all(mdev); + + /* In case we want to get something to stable storage still, + * this may be the last chance. + * Following put_ldev may transition to D_DISKLESS. */ + drbd_md_sync(mdev); + } + put_ldev(mdev); + } + + /* second half of local IO error, failure to attach, + * or administrative detach, + * after local_cnt references have reached zero again */ + if (os.disk != D_DISKLESS && ns.disk == D_DISKLESS) { + /* We must still be diskless, + * re-attach has to be serialized with this! */ + if (mdev->state.disk != D_DISKLESS) + dev_err(DEV, + "ASSERT FAILED: disk is %s while going diskless\n", + drbd_disk_str(mdev->state.disk)); + + if (ns.conn >= C_CONNECTED) + drbd_send_state(mdev, ns); + /* corresponding get_ldev in __drbd_set_state + * this may finally trigger drbd_ldev_destroy. */ + put_ldev(mdev); + } + + /* Notify peer that I had a local IO error, and did not detached.. */ + if (os.disk == D_UP_TO_DATE && ns.disk == D_INCONSISTENT && ns.conn >= C_CONNECTED) + drbd_send_state(mdev, ns); + + /* Disks got bigger while they were detached */ + if (ns.disk > D_NEGOTIATING && ns.pdsk > D_NEGOTIATING && + test_and_clear_bit(RESYNC_AFTER_NEG, &mdev->flags)) { + if (ns.conn == C_CONNECTED) + resync_after_online_grow(mdev); + } + + /* A resync finished or aborted, wake paused devices... */ + if ((os.conn > C_CONNECTED && ns.conn <= C_CONNECTED) || + (os.peer_isp && !ns.peer_isp) || + (os.user_isp && !ns.user_isp)) + resume_next_sg(mdev); + + /* sync target done with resync. Explicitly notify peer, even though + * it should (at least for non-empty resyncs) already know itself. */ + if (os.disk < D_UP_TO_DATE && os.conn >= C_SYNC_SOURCE && ns.conn == C_CONNECTED) + drbd_send_state(mdev, ns); + + /* Verify finished, or reached stop sector. Peer did not know about + * the stop sector, and we may even have changed the stop sector during + * verify to interrupt/stop early. Send the new state. */ + if (os.conn == C_VERIFY_S && ns.conn == C_CONNECTED + && verify_can_do_stop_sector(mdev)) + drbd_send_state(mdev, ns); + + /* This triggers bitmap writeout of potentially still unwritten pages + * if the resync finished cleanly, or aborted because of peer disk + * failure, or because of connection loss. + * For resync aborted because of local disk failure, we cannot do + * any bitmap writeout anymore. + * No harm done if some bits change during this phase. + */ + if (os.conn > C_CONNECTED && ns.conn <= C_CONNECTED && get_ldev(mdev)) { + drbd_queue_bitmap_io(mdev, &drbd_bm_write_copy_pages, NULL, + "write from resync_finished", BM_LOCKED_CHANGE_ALLOWED); + put_ldev(mdev); + } + + if (ns.disk == D_DISKLESS && + ns.conn == C_STANDALONE && + ns.role == R_SECONDARY) { + if (os.aftr_isp != ns.aftr_isp) + resume_next_sg(mdev); + } + + drbd_md_sync(mdev); +} + +struct after_conn_state_chg_work { + struct drbd_work w; + enum drbd_conns oc; + union drbd_state ns_min; + union drbd_state ns_max; /* new, max state, over all mdevs */ + enum chg_state_flags flags; +}; + +static int w_after_conn_state_ch(struct drbd_work *w, int unused) +{ + struct after_conn_state_chg_work *acscw = + container_of(w, struct after_conn_state_chg_work, w); + struct drbd_tconn *tconn = w->tconn; + enum drbd_conns oc = acscw->oc; + union drbd_state ns_max = acscw->ns_max; + struct drbd_conf *mdev; + int vnr; + + kfree(acscw); + + /* Upon network configuration, we need to start the receiver */ + if (oc == C_STANDALONE && ns_max.conn == C_UNCONNECTED) + drbd_thread_start(&tconn->receiver); + + if (oc == C_DISCONNECTING && ns_max.conn == C_STANDALONE) { + struct net_conf *old_conf; + + mutex_lock(&tconn->conf_update); + old_conf = tconn->net_conf; + tconn->my_addr_len = 0; + tconn->peer_addr_len = 0; + rcu_assign_pointer(tconn->net_conf, NULL); + conn_free_crypto(tconn); + mutex_unlock(&tconn->conf_update); + + synchronize_rcu(); + kfree(old_conf); + } + + if (ns_max.susp_fen) { + /* case1: The outdate peer handler is successful: */ + if (ns_max.pdsk <= D_OUTDATED) { + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) { + if (test_bit(NEW_CUR_UUID, &mdev->flags)) { + drbd_uuid_new_current(mdev); + clear_bit(NEW_CUR_UUID, &mdev->flags); + } + } + rcu_read_unlock(); + spin_lock_irq(&tconn->req_lock); + _tl_restart(tconn, CONNECTION_LOST_WHILE_PENDING); + _conn_request_state(tconn, + (union drbd_state) { { .susp_fen = 1 } }, + (union drbd_state) { { .susp_fen = 0 } }, + CS_VERBOSE); + spin_unlock_irq(&tconn->req_lock); + } + } + kref_put(&tconn->kref, &conn_destroy); + + conn_md_sync(tconn); + + return 0; +} + +void conn_old_common_state(struct drbd_tconn *tconn, union drbd_state *pcs, enum chg_state_flags *pf) +{ + enum chg_state_flags flags = ~0; + struct drbd_conf *mdev; + int vnr, first_vol = 1; + union drbd_dev_state os, cs = { + { .role = R_SECONDARY, + .peer = R_UNKNOWN, + .conn = tconn->cstate, + .disk = D_DISKLESS, + .pdsk = D_UNKNOWN, + } }; + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) { + os = mdev->state; + + if (first_vol) { + cs = os; + first_vol = 0; + continue; + } + + if (cs.role != os.role) + flags &= ~CS_DC_ROLE; + + if (cs.peer != os.peer) + flags &= ~CS_DC_PEER; + + if (cs.conn != os.conn) + flags &= ~CS_DC_CONN; + + if (cs.disk != os.disk) + flags &= ~CS_DC_DISK; + + if (cs.pdsk != os.pdsk) + flags &= ~CS_DC_PDSK; + } + rcu_read_unlock(); + + *pf |= CS_DC_MASK; + *pf &= flags; + (*pcs).i = cs.i; +} + +static enum drbd_state_rv +conn_is_valid_transition(struct drbd_tconn *tconn, union drbd_state mask, union drbd_state val, + enum chg_state_flags flags) +{ + enum drbd_state_rv rv = SS_SUCCESS; + union drbd_state ns, os; + struct drbd_conf *mdev; + int vnr; + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) { + os = drbd_read_state(mdev); + ns = sanitize_state(mdev, apply_mask_val(os, mask, val), NULL); + + if (flags & CS_IGN_OUTD_FAIL && ns.disk == D_OUTDATED && os.disk < D_OUTDATED) + ns.disk = os.disk; + + if (ns.i == os.i) + continue; + + rv = is_valid_transition(os, ns); + if (rv < SS_SUCCESS) + break; + + if (!(flags & CS_HARD)) { + rv = is_valid_state(mdev, ns); + if (rv < SS_SUCCESS) { + if (is_valid_state(mdev, os) == rv) + rv = is_valid_soft_transition(os, ns, tconn); + } else + rv = is_valid_soft_transition(os, ns, tconn); + } + if (rv < SS_SUCCESS) + break; + } + rcu_read_unlock(); + + if (rv < SS_SUCCESS && flags & CS_VERBOSE) + print_st_err(mdev, os, ns, rv); + + return rv; +} + +void +conn_set_state(struct drbd_tconn *tconn, union drbd_state mask, union drbd_state val, + union drbd_state *pns_min, union drbd_state *pns_max, enum chg_state_flags flags) +{ + union drbd_state ns, os, ns_max = { }; + union drbd_state ns_min = { + { .role = R_MASK, + .peer = R_MASK, + .conn = val.conn, + .disk = D_MASK, + .pdsk = D_MASK + } }; + struct drbd_conf *mdev; + enum drbd_state_rv rv; + int vnr, number_of_volumes = 0; + + if (mask.conn == C_MASK) { + /* remember last connect time so request_timer_fn() won't + * kill newly established sessions while we are still trying to thaw + * previously frozen IO */ + if (tconn->cstate != C_WF_REPORT_PARAMS && val.conn == C_WF_REPORT_PARAMS) + tconn->last_reconnect_jif = jiffies; + + tconn->cstate = val.conn; + } + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) { + number_of_volumes++; + os = drbd_read_state(mdev); + ns = apply_mask_val(os, mask, val); + ns = sanitize_state(mdev, ns, NULL); + + if (flags & CS_IGN_OUTD_FAIL && ns.disk == D_OUTDATED && os.disk < D_OUTDATED) + ns.disk = os.disk; + + rv = __drbd_set_state(mdev, ns, flags, NULL); + if (rv < SS_SUCCESS) + BUG(); + + ns.i = mdev->state.i; + ns_max.role = max_role(ns.role, ns_max.role); + ns_max.peer = max_role(ns.peer, ns_max.peer); + ns_max.conn = max_t(enum drbd_conns, ns.conn, ns_max.conn); + ns_max.disk = max_t(enum drbd_disk_state, ns.disk, ns_max.disk); + ns_max.pdsk = max_t(enum drbd_disk_state, ns.pdsk, ns_max.pdsk); + + ns_min.role = min_role(ns.role, ns_min.role); + ns_min.peer = min_role(ns.peer, ns_min.peer); + ns_min.conn = min_t(enum drbd_conns, ns.conn, ns_min.conn); + ns_min.disk = min_t(enum drbd_disk_state, ns.disk, ns_min.disk); + ns_min.pdsk = min_t(enum drbd_disk_state, ns.pdsk, ns_min.pdsk); + } + rcu_read_unlock(); + + if (number_of_volumes == 0) { + ns_min = ns_max = (union drbd_state) { { + .role = R_SECONDARY, + .peer = R_UNKNOWN, + .conn = val.conn, + .disk = D_DISKLESS, + .pdsk = D_UNKNOWN + } }; + } + + ns_min.susp = ns_max.susp = tconn->susp; + ns_min.susp_nod = ns_max.susp_nod = tconn->susp_nod; + ns_min.susp_fen = ns_max.susp_fen = tconn->susp_fen; + + *pns_min = ns_min; + *pns_max = ns_max; +} + +static enum drbd_state_rv +_conn_rq_cond(struct drbd_tconn *tconn, union drbd_state mask, union drbd_state val) +{ + enum drbd_state_rv rv; + + if (test_and_clear_bit(CONN_WD_ST_CHG_OKAY, &tconn->flags)) + return SS_CW_SUCCESS; + + if (test_and_clear_bit(CONN_WD_ST_CHG_FAIL, &tconn->flags)) + return SS_CW_FAILED_BY_PEER; + + rv = tconn->cstate != C_WF_REPORT_PARAMS ? SS_CW_NO_NEED : SS_UNKNOWN_ERROR; + + if (rv == SS_UNKNOWN_ERROR) + rv = conn_is_valid_transition(tconn, mask, val, 0); + + if (rv == SS_SUCCESS) + rv = SS_UNKNOWN_ERROR; /* cont waiting, otherwise fail. */ + + return rv; +} + +enum drbd_state_rv +_conn_request_state(struct drbd_tconn *tconn, union drbd_state mask, union drbd_state val, + enum chg_state_flags flags) +{ + enum drbd_state_rv rv = SS_SUCCESS; + struct after_conn_state_chg_work *acscw; + enum drbd_conns oc = tconn->cstate; + union drbd_state ns_max, ns_min, os; + bool have_mutex = false; + + if (mask.conn) { + rv = is_valid_conn_transition(oc, val.conn); + if (rv < SS_SUCCESS) + goto abort; + } + + rv = conn_is_valid_transition(tconn, mask, val, flags); + if (rv < SS_SUCCESS) + goto abort; + + if (oc == C_WF_REPORT_PARAMS && val.conn == C_DISCONNECTING && + !(flags & (CS_LOCAL_ONLY | CS_HARD))) { + + /* This will be a cluster-wide state change. + * Need to give up the spinlock, grab the mutex, + * then send the state change request, ... */ + spin_unlock_irq(&tconn->req_lock); + mutex_lock(&tconn->cstate_mutex); + have_mutex = true; + + set_bit(CONN_WD_ST_CHG_REQ, &tconn->flags); + if (conn_send_state_req(tconn, mask, val)) { + /* sending failed. */ + clear_bit(CONN_WD_ST_CHG_REQ, &tconn->flags); + rv = SS_CW_FAILED_BY_PEER; + /* need to re-aquire the spin lock, though */ + goto abort_unlocked; + } + + if (val.conn == C_DISCONNECTING) + set_bit(DISCONNECT_SENT, &tconn->flags); + + /* ... and re-aquire the spinlock. + * If _conn_rq_cond() returned >= SS_SUCCESS, we must call + * conn_set_state() within the same spinlock. */ + spin_lock_irq(&tconn->req_lock); + wait_event_lock_irq(tconn->ping_wait, + (rv = _conn_rq_cond(tconn, mask, val)), + tconn->req_lock); + clear_bit(CONN_WD_ST_CHG_REQ, &tconn->flags); + if (rv < SS_SUCCESS) + goto abort; + } + + conn_old_common_state(tconn, &os, &flags); + flags |= CS_DC_SUSP; + conn_set_state(tconn, mask, val, &ns_min, &ns_max, flags); + conn_pr_state_change(tconn, os, ns_max, flags); + + acscw = kmalloc(sizeof(*acscw), GFP_ATOMIC); + if (acscw) { + acscw->oc = os.conn; + acscw->ns_min = ns_min; + acscw->ns_max = ns_max; + acscw->flags = flags; + acscw->w.cb = w_after_conn_state_ch; + kref_get(&tconn->kref); + acscw->w.tconn = tconn; + drbd_queue_work(&tconn->sender_work, &acscw->w); + } else { + conn_err(tconn, "Could not kmalloc an acscw\n"); + } + + abort: + if (have_mutex) { + /* mutex_unlock() "... must not be used in interrupt context.", + * so give up the spinlock, then re-aquire it */ + spin_unlock_irq(&tconn->req_lock); + abort_unlocked: + mutex_unlock(&tconn->cstate_mutex); + spin_lock_irq(&tconn->req_lock); + } + if (rv < SS_SUCCESS && flags & CS_VERBOSE) { + conn_err(tconn, "State change failed: %s\n", drbd_set_st_err_str(rv)); + conn_err(tconn, " mask = 0x%x val = 0x%x\n", mask.i, val.i); + conn_err(tconn, " old_conn:%s wanted_conn:%s\n", drbd_conn_str(oc), drbd_conn_str(val.conn)); + } + return rv; +} + +enum drbd_state_rv +conn_request_state(struct drbd_tconn *tconn, union drbd_state mask, union drbd_state val, + enum chg_state_flags flags) +{ + enum drbd_state_rv rv; + + spin_lock_irq(&tconn->req_lock); + rv = _conn_request_state(tconn, mask, val, flags); + spin_unlock_irq(&tconn->req_lock); + + return rv; +} diff --git a/drivers/block/drbd/drbd_state.h b/drivers/block/drbd/drbd_state.h new file mode 100644 index 000000000000..a3c361bbc4b6 --- /dev/null +++ b/drivers/block/drbd/drbd_state.h @@ -0,0 +1,161 @@ +#ifndef DRBD_STATE_H +#define DRBD_STATE_H + +struct drbd_conf; +struct drbd_tconn; + +/** + * DOC: DRBD State macros + * + * These macros are used to express state changes in easily readable form. + * + * The NS macros expand to a mask and a value, that can be bit ored onto the + * current state as soon as the spinlock (req_lock) was taken. + * + * The _NS macros are used for state functions that get called with the + * spinlock. These macros expand directly to the new state value. + * + * Besides the basic forms NS() and _NS() additional _?NS[23] are defined + * to express state changes that affect more than one aspect of the state. + * + * E.g. NS2(conn, C_CONNECTED, peer, R_SECONDARY) + * Means that the network connection was established and that the peer + * is in secondary role. + */ +#define role_MASK R_MASK +#define peer_MASK R_MASK +#define disk_MASK D_MASK +#define pdsk_MASK D_MASK +#define conn_MASK C_MASK +#define susp_MASK 1 +#define user_isp_MASK 1 +#define aftr_isp_MASK 1 +#define susp_nod_MASK 1 +#define susp_fen_MASK 1 + +#define NS(T, S) \ + ({ union drbd_state mask; mask.i = 0; mask.T = T##_MASK; mask; }), \ + ({ union drbd_state val; val.i = 0; val.T = (S); val; }) +#define NS2(T1, S1, T2, S2) \ + ({ union drbd_state mask; mask.i = 0; mask.T1 = T1##_MASK; \ + mask.T2 = T2##_MASK; mask; }), \ + ({ union drbd_state val; val.i = 0; val.T1 = (S1); \ + val.T2 = (S2); val; }) +#define NS3(T1, S1, T2, S2, T3, S3) \ + ({ union drbd_state mask; mask.i = 0; mask.T1 = T1##_MASK; \ + mask.T2 = T2##_MASK; mask.T3 = T3##_MASK; mask; }), \ + ({ union drbd_state val; val.i = 0; val.T1 = (S1); \ + val.T2 = (S2); val.T3 = (S3); val; }) + +#define _NS(D, T, S) \ + D, ({ union drbd_state __ns; __ns = drbd_read_state(D); __ns.T = (S); __ns; }) +#define _NS2(D, T1, S1, T2, S2) \ + D, ({ union drbd_state __ns; __ns = drbd_read_state(D); __ns.T1 = (S1); \ + __ns.T2 = (S2); __ns; }) +#define _NS3(D, T1, S1, T2, S2, T3, S3) \ + D, ({ union drbd_state __ns; __ns = drbd_read_state(D); __ns.T1 = (S1); \ + __ns.T2 = (S2); __ns.T3 = (S3); __ns; }) + +enum chg_state_flags { + CS_HARD = 1 << 0, + CS_VERBOSE = 1 << 1, + CS_WAIT_COMPLETE = 1 << 2, + CS_SERIALIZE = 1 << 3, + CS_ORDERED = CS_WAIT_COMPLETE + CS_SERIALIZE, + CS_LOCAL_ONLY = 1 << 4, /* Do not consider a device pair wide state change */ + CS_DC_ROLE = 1 << 5, /* DC = display as connection state change */ + CS_DC_PEER = 1 << 6, + CS_DC_CONN = 1 << 7, + CS_DC_DISK = 1 << 8, + CS_DC_PDSK = 1 << 9, + CS_DC_SUSP = 1 << 10, + CS_DC_MASK = CS_DC_ROLE + CS_DC_PEER + CS_DC_CONN + CS_DC_DISK + CS_DC_PDSK, + CS_IGN_OUTD_FAIL = 1 << 11, +}; + +/* drbd_dev_state and drbd_state are different types. This is to stress the + small difference. There is no suspended flag (.susp), and no suspended + while fence handler runs flas (susp_fen). */ +union drbd_dev_state { + struct { +#if defined(__LITTLE_ENDIAN_BITFIELD) + unsigned role:2 ; /* 3/4 primary/secondary/unknown */ + unsigned peer:2 ; /* 3/4 primary/secondary/unknown */ + unsigned conn:5 ; /* 17/32 cstates */ + unsigned disk:4 ; /* 8/16 from D_DISKLESS to D_UP_TO_DATE */ + unsigned pdsk:4 ; /* 8/16 from D_DISKLESS to D_UP_TO_DATE */ + unsigned _unused:1 ; + unsigned aftr_isp:1 ; /* isp .. imposed sync pause */ + unsigned peer_isp:1 ; + unsigned user_isp:1 ; + unsigned _pad:11; /* 0 unused */ +#elif defined(__BIG_ENDIAN_BITFIELD) + unsigned _pad:11; + unsigned user_isp:1 ; + unsigned peer_isp:1 ; + unsigned aftr_isp:1 ; /* isp .. imposed sync pause */ + unsigned _unused:1 ; + unsigned pdsk:4 ; /* 8/16 from D_DISKLESS to D_UP_TO_DATE */ + unsigned disk:4 ; /* 8/16 from D_DISKLESS to D_UP_TO_DATE */ + unsigned conn:5 ; /* 17/32 cstates */ + unsigned peer:2 ; /* 3/4 primary/secondary/unknown */ + unsigned role:2 ; /* 3/4 primary/secondary/unknown */ +#else +# error "this endianess is not supported" +#endif + }; + unsigned int i; +}; + +extern enum drbd_state_rv drbd_change_state(struct drbd_conf *mdev, + enum chg_state_flags f, + union drbd_state mask, + union drbd_state val); +extern void drbd_force_state(struct drbd_conf *, union drbd_state, + union drbd_state); +extern enum drbd_state_rv _drbd_request_state(struct drbd_conf *, + union drbd_state, + union drbd_state, + enum chg_state_flags); +extern enum drbd_state_rv __drbd_set_state(struct drbd_conf *, union drbd_state, + enum chg_state_flags, + struct completion *done); +extern void print_st_err(struct drbd_conf *, union drbd_state, + union drbd_state, int); + +enum drbd_state_rv +_conn_request_state(struct drbd_tconn *tconn, union drbd_state mask, union drbd_state val, + enum chg_state_flags flags); + +enum drbd_state_rv +conn_request_state(struct drbd_tconn *tconn, union drbd_state mask, union drbd_state val, + enum chg_state_flags flags); + +extern void drbd_resume_al(struct drbd_conf *mdev); +extern bool conn_all_vols_unconf(struct drbd_tconn *tconn); + +/** + * drbd_request_state() - Reqest a state change + * @mdev: DRBD device. + * @mask: mask of state bits to change. + * @val: value of new state bits. + * + * This is the most graceful way of requesting a state change. It is verbose + * quite verbose in case the state change is not possible, and all those + * state changes are globally serialized. + */ +static inline int drbd_request_state(struct drbd_conf *mdev, + union drbd_state mask, + union drbd_state val) +{ + return _drbd_request_state(mdev, mask, val, CS_VERBOSE + CS_ORDERED); +} + +enum drbd_role conn_highest_role(struct drbd_tconn *tconn); +enum drbd_role conn_highest_peer(struct drbd_tconn *tconn); +enum drbd_disk_state conn_highest_disk(struct drbd_tconn *tconn); +enum drbd_disk_state conn_lowest_disk(struct drbd_tconn *tconn); +enum drbd_disk_state conn_highest_pdsk(struct drbd_tconn *tconn); +enum drbd_conns conn_lowest_conn(struct drbd_tconn *tconn); + +#endif diff --git a/drivers/block/drbd/drbd_strings.c b/drivers/block/drbd/drbd_strings.c index c44a2a602772..9a664bd27404 100644 --- a/drivers/block/drbd/drbd_strings.c +++ b/drivers/block/drbd/drbd_strings.c @@ -89,6 +89,7 @@ static const char *drbd_state_sw_errors[] = { [-SS_LOWER_THAN_OUTDATED] = "Disk state is lower than outdated", [-SS_IN_TRANSIENT_STATE] = "In transient state, retry after next state change", [-SS_CONCURRENT_ST_CHG] = "Concurrent state changes detected and aborted", + [-SS_O_VOL_PEER_PRI] = "Other vol primary on peer not allowed by config", }; const char *drbd_conn_str(enum drbd_conns s) diff --git a/drivers/block/drbd/drbd_worker.c b/drivers/block/drbd/drbd_worker.c index 6bce2cc179d4..424dc7bdf9b7 100644 --- a/drivers/block/drbd/drbd_worker.c +++ b/drivers/block/drbd/drbd_worker.c @@ -38,16 +38,13 @@ #include "drbd_int.h" #include "drbd_req.h" -static int w_make_ov_request(struct drbd_conf *mdev, struct drbd_work *w, int cancel); -static int w_make_resync_request(struct drbd_conf *mdev, - struct drbd_work *w, int cancel); - +static int w_make_ov_request(struct drbd_work *w, int cancel); /* endio handlers: * drbd_md_io_complete (defined here) - * drbd_endio_pri (defined here) - * drbd_endio_sec (defined here) + * drbd_request_endio (defined here) + * drbd_peer_request_endio (defined here) * bm_async_io_complete (defined in drbd_bitmap.c) * * For all these callbacks, note the following: @@ -60,7 +57,7 @@ static int w_make_resync_request(struct drbd_conf *mdev, /* About the global_state_lock Each state transition on an device holds a read lock. In case we have - to evaluate the sync after dependencies, we grab a write lock, because + to evaluate the resync after dependencies, we grab a write lock, because we need stable states on all devices for that. */ rwlock_t global_state_lock; @@ -98,97 +95,93 @@ void drbd_md_io_complete(struct bio *bio, int error) /* reads on behalf of the partner, * "submitted" by the receiver */ -void drbd_endio_read_sec_final(struct drbd_epoch_entry *e) __releases(local) +void drbd_endio_read_sec_final(struct drbd_peer_request *peer_req) __releases(local) { unsigned long flags = 0; - struct drbd_conf *mdev = e->mdev; - - D_ASSERT(e->block_id != ID_VACANT); + struct drbd_conf *mdev = peer_req->w.mdev; - spin_lock_irqsave(&mdev->req_lock, flags); - mdev->read_cnt += e->size >> 9; - list_del(&e->w.list); + spin_lock_irqsave(&mdev->tconn->req_lock, flags); + mdev->read_cnt += peer_req->i.size >> 9; + list_del(&peer_req->w.list); if (list_empty(&mdev->read_ee)) wake_up(&mdev->ee_wait); - if (test_bit(__EE_WAS_ERROR, &e->flags)) - __drbd_chk_io_error(mdev, DRBD_IO_ERROR); - spin_unlock_irqrestore(&mdev->req_lock, flags); + if (test_bit(__EE_WAS_ERROR, &peer_req->flags)) + __drbd_chk_io_error(mdev, DRBD_READ_ERROR); + spin_unlock_irqrestore(&mdev->tconn->req_lock, flags); - drbd_queue_work(&mdev->data.work, &e->w); + drbd_queue_work(&mdev->tconn->sender_work, &peer_req->w); put_ldev(mdev); } /* writes on behalf of the partner, or resync writes, * "submitted" by the receiver, final stage. */ -static void drbd_endio_write_sec_final(struct drbd_epoch_entry *e) __releases(local) +static void drbd_endio_write_sec_final(struct drbd_peer_request *peer_req) __releases(local) { unsigned long flags = 0; - struct drbd_conf *mdev = e->mdev; - sector_t e_sector; + struct drbd_conf *mdev = peer_req->w.mdev; + struct drbd_interval i; int do_wake; - int is_syncer_req; + u64 block_id; int do_al_complete_io; - D_ASSERT(e->block_id != ID_VACANT); - - /* after we moved e to done_ee, + /* after we moved peer_req to done_ee, * we may no longer access it, * it may be freed/reused already! * (as soon as we release the req_lock) */ - e_sector = e->sector; - do_al_complete_io = e->flags & EE_CALL_AL_COMPLETE_IO; - is_syncer_req = is_syncer_block_id(e->block_id); + i = peer_req->i; + do_al_complete_io = peer_req->flags & EE_CALL_AL_COMPLETE_IO; + block_id = peer_req->block_id; - spin_lock_irqsave(&mdev->req_lock, flags); - mdev->writ_cnt += e->size >> 9; - list_del(&e->w.list); /* has been on active_ee or sync_ee */ - list_add_tail(&e->w.list, &mdev->done_ee); + spin_lock_irqsave(&mdev->tconn->req_lock, flags); + mdev->writ_cnt += peer_req->i.size >> 9; + list_move_tail(&peer_req->w.list, &mdev->done_ee); - /* No hlist_del_init(&e->collision) here, we did not send the Ack yet, - * neither did we wake possibly waiting conflicting requests. - * done from "drbd_process_done_ee" within the appropriate w.cb - * (e_end_block/e_end_resync_block) or from _drbd_clear_done_ee */ + /* + * Do not remove from the write_requests tree here: we did not send the + * Ack yet and did not wake possibly waiting conflicting requests. + * Removed from the tree from "drbd_process_done_ee" within the + * appropriate w.cb (e_end_block/e_end_resync_block) or from + * _drbd_clear_done_ee. + */ - do_wake = is_syncer_req - ? list_empty(&mdev->sync_ee) - : list_empty(&mdev->active_ee); + do_wake = list_empty(block_id == ID_SYNCER ? &mdev->sync_ee : &mdev->active_ee); - if (test_bit(__EE_WAS_ERROR, &e->flags)) - __drbd_chk_io_error(mdev, DRBD_IO_ERROR); - spin_unlock_irqrestore(&mdev->req_lock, flags); + if (test_bit(__EE_WAS_ERROR, &peer_req->flags)) + __drbd_chk_io_error(mdev, DRBD_WRITE_ERROR); + spin_unlock_irqrestore(&mdev->tconn->req_lock, flags); - if (is_syncer_req) - drbd_rs_complete_io(mdev, e_sector); + if (block_id == ID_SYNCER) + drbd_rs_complete_io(mdev, i.sector); if (do_wake) wake_up(&mdev->ee_wait); if (do_al_complete_io) - drbd_al_complete_io(mdev, e_sector); + drbd_al_complete_io(mdev, &i); - wake_asender(mdev); + wake_asender(mdev->tconn); put_ldev(mdev); } /* writes on behalf of the partner, or resync writes, * "submitted" by the receiver. */ -void drbd_endio_sec(struct bio *bio, int error) +void drbd_peer_request_endio(struct bio *bio, int error) { - struct drbd_epoch_entry *e = bio->bi_private; - struct drbd_conf *mdev = e->mdev; + struct drbd_peer_request *peer_req = bio->bi_private; + struct drbd_conf *mdev = peer_req->w.mdev; int uptodate = bio_flagged(bio, BIO_UPTODATE); int is_write = bio_data_dir(bio) == WRITE; if (error && __ratelimit(&drbd_ratelimit_state)) dev_warn(DEV, "%s: error=%d s=%llus\n", is_write ? "write" : "read", error, - (unsigned long long)e->sector); + (unsigned long long)peer_req->i.sector); if (!error && !uptodate) { if (__ratelimit(&drbd_ratelimit_state)) dev_warn(DEV, "%s: setting error to -EIO s=%llus\n", is_write ? "write" : "read", - (unsigned long long)e->sector); + (unsigned long long)peer_req->i.sector); /* strange behavior of some lower level drivers... * fail the request by clearing the uptodate flag, * but do not return any error?! */ @@ -196,24 +189,24 @@ void drbd_endio_sec(struct bio *bio, int error) } if (error) - set_bit(__EE_WAS_ERROR, &e->flags); + set_bit(__EE_WAS_ERROR, &peer_req->flags); bio_put(bio); /* no need for the bio anymore */ - if (atomic_dec_and_test(&e->pending_bios)) { + if (atomic_dec_and_test(&peer_req->pending_bios)) { if (is_write) - drbd_endio_write_sec_final(e); + drbd_endio_write_sec_final(peer_req); else - drbd_endio_read_sec_final(e); + drbd_endio_read_sec_final(peer_req); } } /* read, readA or write requests on R_PRIMARY coming from drbd_make_request */ -void drbd_endio_pri(struct bio *bio, int error) +void drbd_request_endio(struct bio *bio, int error) { unsigned long flags; struct drbd_request *req = bio->bi_private; - struct drbd_conf *mdev = req->mdev; + struct drbd_conf *mdev = req->w.mdev; struct bio_and_error m; enum drbd_req_event what; int uptodate = bio_flagged(bio, BIO_UPTODATE); @@ -227,53 +220,72 @@ void drbd_endio_pri(struct bio *bio, int error) error = -EIO; } + + /* If this request was aborted locally before, + * but now was completed "successfully", + * chances are that this caused arbitrary data corruption. + * + * "aborting" requests, or force-detaching the disk, is intended for + * completely blocked/hung local backing devices which do no longer + * complete requests at all, not even do error completions. In this + * situation, usually a hard-reset and failover is the only way out. + * + * By "aborting", basically faking a local error-completion, + * we allow for a more graceful swichover by cleanly migrating services. + * Still the affected node has to be rebooted "soon". + * + * By completing these requests, we allow the upper layers to re-use + * the associated data pages. + * + * If later the local backing device "recovers", and now DMAs some data + * from disk into the original request pages, in the best case it will + * just put random data into unused pages; but typically it will corrupt + * meanwhile completely unrelated data, causing all sorts of damage. + * + * Which means delayed successful completion, + * especially for READ requests, + * is a reason to panic(). + * + * We assume that a delayed *error* completion is OK, + * though we still will complain noisily about it. + */ + if (unlikely(req->rq_state & RQ_LOCAL_ABORTED)) { + if (__ratelimit(&drbd_ratelimit_state)) + dev_emerg(DEV, "delayed completion of aborted local request; disk-timeout may be too aggressive\n"); + + if (!error) + panic("possible random memory corruption caused by delayed completion of aborted local request\n"); + } + /* to avoid recursion in __req_mod */ if (unlikely(error)) { what = (bio_data_dir(bio) == WRITE) - ? write_completed_with_error + ? WRITE_COMPLETED_WITH_ERROR : (bio_rw(bio) == READ) - ? read_completed_with_error - : read_ahead_completed_with_error; + ? READ_COMPLETED_WITH_ERROR + : READ_AHEAD_COMPLETED_WITH_ERROR; } else - what = completed_ok; + what = COMPLETED_OK; bio_put(req->private_bio); req->private_bio = ERR_PTR(error); /* not req_mod(), we need irqsave here! */ - spin_lock_irqsave(&mdev->req_lock, flags); + spin_lock_irqsave(&mdev->tconn->req_lock, flags); __req_mod(req, what, &m); - spin_unlock_irqrestore(&mdev->req_lock, flags); + spin_unlock_irqrestore(&mdev->tconn->req_lock, flags); put_ldev(mdev); if (m.bio) complete_master_bio(mdev, &m); } -int w_read_retry_remote(struct drbd_conf *mdev, struct drbd_work *w, int cancel) -{ - struct drbd_request *req = container_of(w, struct drbd_request, w); - - /* We should not detach for read io-error, - * but try to WRITE the P_DATA_REPLY to the failed location, - * to give the disk the chance to relocate that block */ - - spin_lock_irq(&mdev->req_lock); - if (cancel || mdev->state.pdsk != D_UP_TO_DATE) { - _req_mod(req, read_retry_remote_canceled); - spin_unlock_irq(&mdev->req_lock); - return 1; - } - spin_unlock_irq(&mdev->req_lock); - - return w_send_read_req(mdev, w, 0); -} - -void drbd_csum_ee(struct drbd_conf *mdev, struct crypto_hash *tfm, struct drbd_epoch_entry *e, void *digest) +void drbd_csum_ee(struct drbd_conf *mdev, struct crypto_hash *tfm, + struct drbd_peer_request *peer_req, void *digest) { struct hash_desc desc; struct scatterlist sg; - struct page *page = e->pages; + struct page *page = peer_req->pages; struct page *tmp; unsigned len; @@ -290,7 +302,7 @@ void drbd_csum_ee(struct drbd_conf *mdev, struct crypto_hash *tfm, struct drbd_e page = tmp; } /* and now the last, possibly only partially used page */ - len = e->size & (PAGE_SIZE - 1); + len = peer_req->i.size & (PAGE_SIZE - 1); sg_set_page(&sg, page, len ?: PAGE_SIZE, 0); crypto_hash_update(&desc, &sg, sg.length); crypto_hash_final(&desc, digest); @@ -316,59 +328,58 @@ void drbd_csum_bio(struct drbd_conf *mdev, struct crypto_hash *tfm, struct bio * crypto_hash_final(&desc, digest); } -/* TODO merge common code with w_e_end_ov_req */ -int w_e_send_csum(struct drbd_conf *mdev, struct drbd_work *w, int cancel) +/* MAYBE merge common code with w_e_end_ov_req */ +static int w_e_send_csum(struct drbd_work *w, int cancel) { - struct drbd_epoch_entry *e = container_of(w, struct drbd_epoch_entry, w); + struct drbd_peer_request *peer_req = container_of(w, struct drbd_peer_request, w); + struct drbd_conf *mdev = w->mdev; int digest_size; void *digest; - int ok = 1; - - D_ASSERT(e->block_id == DRBD_MAGIC + 0xbeef); + int err = 0; if (unlikely(cancel)) goto out; - if (likely((e->flags & EE_WAS_ERROR) != 0)) + if (unlikely((peer_req->flags & EE_WAS_ERROR) != 0)) goto out; - digest_size = crypto_hash_digestsize(mdev->csums_tfm); + digest_size = crypto_hash_digestsize(mdev->tconn->csums_tfm); digest = kmalloc(digest_size, GFP_NOIO); if (digest) { - sector_t sector = e->sector; - unsigned int size = e->size; - drbd_csum_ee(mdev, mdev->csums_tfm, e, digest); - /* Free e and pages before send. + sector_t sector = peer_req->i.sector; + unsigned int size = peer_req->i.size; + drbd_csum_ee(mdev, mdev->tconn->csums_tfm, peer_req, digest); + /* Free peer_req and pages before send. * In case we block on congestion, we could otherwise run into * some distributed deadlock, if the other side blocks on * congestion as well, because our receiver blocks in - * drbd_pp_alloc due to pp_in_use > max_buffers. */ - drbd_free_ee(mdev, e); - e = NULL; + * drbd_alloc_pages due to pp_in_use > max_buffers. */ + drbd_free_peer_req(mdev, peer_req); + peer_req = NULL; inc_rs_pending(mdev); - ok = drbd_send_drequest_csum(mdev, sector, size, - digest, digest_size, - P_CSUM_RS_REQUEST); + err = drbd_send_drequest_csum(mdev, sector, size, + digest, digest_size, + P_CSUM_RS_REQUEST); kfree(digest); } else { dev_err(DEV, "kmalloc() of digest failed.\n"); - ok = 0; + err = -ENOMEM; } out: - if (e) - drbd_free_ee(mdev, e); + if (peer_req) + drbd_free_peer_req(mdev, peer_req); - if (unlikely(!ok)) + if (unlikely(err)) dev_err(DEV, "drbd_send_drequest(..., csum) failed\n"); - return ok; + return err; } #define GFP_TRY (__GFP_HIGHMEM | __GFP_NOWARN) static int read_for_csum(struct drbd_conf *mdev, sector_t sector, int size) { - struct drbd_epoch_entry *e; + struct drbd_peer_request *peer_req; if (!get_ldev(mdev)) return -EIO; @@ -378,45 +389,47 @@ static int read_for_csum(struct drbd_conf *mdev, sector_t sector, int size) /* GFP_TRY, because if there is no memory available right now, this may * be rescheduled for later. It is "only" background resync, after all. */ - e = drbd_alloc_ee(mdev, DRBD_MAGIC+0xbeef, sector, size, GFP_TRY); - if (!e) + peer_req = drbd_alloc_peer_req(mdev, ID_SYNCER /* unused */, sector, + size, GFP_TRY); + if (!peer_req) goto defer; - e->w.cb = w_e_send_csum; - spin_lock_irq(&mdev->req_lock); - list_add(&e->w.list, &mdev->read_ee); - spin_unlock_irq(&mdev->req_lock); + peer_req->w.cb = w_e_send_csum; + spin_lock_irq(&mdev->tconn->req_lock); + list_add(&peer_req->w.list, &mdev->read_ee); + spin_unlock_irq(&mdev->tconn->req_lock); atomic_add(size >> 9, &mdev->rs_sect_ev); - if (drbd_submit_ee(mdev, e, READ, DRBD_FAULT_RS_RD) == 0) + if (drbd_submit_peer_request(mdev, peer_req, READ, DRBD_FAULT_RS_RD) == 0) return 0; /* If it failed because of ENOMEM, retry should help. If it failed * because bio_add_page failed (probably broken lower level driver), * retry may or may not help. * If it does not, you may need to force disconnect. */ - spin_lock_irq(&mdev->req_lock); - list_del(&e->w.list); - spin_unlock_irq(&mdev->req_lock); + spin_lock_irq(&mdev->tconn->req_lock); + list_del(&peer_req->w.list); + spin_unlock_irq(&mdev->tconn->req_lock); - drbd_free_ee(mdev, e); + drbd_free_peer_req(mdev, peer_req); defer: put_ldev(mdev); return -EAGAIN; } -int w_resync_timer(struct drbd_conf *mdev, struct drbd_work *w, int cancel) +int w_resync_timer(struct drbd_work *w, int cancel) { + struct drbd_conf *mdev = w->mdev; switch (mdev->state.conn) { case C_VERIFY_S: - w_make_ov_request(mdev, w, cancel); + w_make_ov_request(w, cancel); break; case C_SYNC_TARGET: - w_make_resync_request(mdev, w, cancel); + w_make_resync_request(w, cancel); break; } - return 1; + return 0; } void resync_timer_fn(unsigned long data) @@ -424,7 +437,7 @@ void resync_timer_fn(unsigned long data) struct drbd_conf *mdev = (struct drbd_conf *) data; if (list_empty(&mdev->resync_work.list)) - drbd_queue_work(&mdev->data.work, &mdev->resync_work); + drbd_queue_work(&mdev->tconn->sender_work, &mdev->resync_work); } static void fifo_set(struct fifo_buffer *fb, int value) @@ -456,8 +469,24 @@ static void fifo_add_val(struct fifo_buffer *fb, int value) fb->values[i] += value; } +struct fifo_buffer *fifo_alloc(int fifo_size) +{ + struct fifo_buffer *fb; + + fb = kzalloc(sizeof(struct fifo_buffer) + sizeof(int) * fifo_size, GFP_NOIO); + if (!fb) + return NULL; + + fb->head_index = 0; + fb->size = fifo_size; + fb->total = 0; + + return fb; +} + static int drbd_rs_controller(struct drbd_conf *mdev) { + struct disk_conf *dc; unsigned int sect_in; /* Number of sectors that came in since the last turn */ unsigned int want; /* The number of sectors we want in the proxy */ int req_sect; /* Number of sectors to request in this turn */ @@ -466,38 +495,39 @@ static int drbd_rs_controller(struct drbd_conf *mdev) int steps; /* Number of time steps to plan ahead */ int curr_corr; int max_sect; + struct fifo_buffer *plan; sect_in = atomic_xchg(&mdev->rs_sect_in, 0); /* Number of sectors that came in */ mdev->rs_in_flight -= sect_in; - spin_lock(&mdev->peer_seq_lock); /* get an atomic view on mdev->rs_plan_s */ + dc = rcu_dereference(mdev->ldev->disk_conf); + plan = rcu_dereference(mdev->rs_plan_s); - steps = mdev->rs_plan_s.size; /* (mdev->sync_conf.c_plan_ahead * 10 * SLEEP_TIME) / HZ; */ + steps = plan->size; /* (dc->c_plan_ahead * 10 * SLEEP_TIME) / HZ; */ if (mdev->rs_in_flight + sect_in == 0) { /* At start of resync */ - want = ((mdev->sync_conf.rate * 2 * SLEEP_TIME) / HZ) * steps; + want = ((dc->resync_rate * 2 * SLEEP_TIME) / HZ) * steps; } else { /* normal path */ - want = mdev->sync_conf.c_fill_target ? mdev->sync_conf.c_fill_target : - sect_in * mdev->sync_conf.c_delay_target * HZ / (SLEEP_TIME * 10); + want = dc->c_fill_target ? dc->c_fill_target : + sect_in * dc->c_delay_target * HZ / (SLEEP_TIME * 10); } - correction = want - mdev->rs_in_flight - mdev->rs_planed; + correction = want - mdev->rs_in_flight - plan->total; /* Plan ahead */ cps = correction / steps; - fifo_add_val(&mdev->rs_plan_s, cps); - mdev->rs_planed += cps * steps; + fifo_add_val(plan, cps); + plan->total += cps * steps; /* What we do in this step */ - curr_corr = fifo_push(&mdev->rs_plan_s, 0); - spin_unlock(&mdev->peer_seq_lock); - mdev->rs_planed -= curr_corr; + curr_corr = fifo_push(plan, 0); + plan->total -= curr_corr; req_sect = sect_in + curr_corr; if (req_sect < 0) req_sect = 0; - max_sect = (mdev->sync_conf.c_max_rate * 2 * SLEEP_TIME) / HZ; + max_sect = (dc->c_max_rate * 2 * SLEEP_TIME) / HZ; if (req_sect > max_sect) req_sect = max_sect; @@ -513,22 +543,25 @@ static int drbd_rs_controller(struct drbd_conf *mdev) static int drbd_rs_number_requests(struct drbd_conf *mdev) { int number; - if (mdev->rs_plan_s.size) { /* mdev->sync_conf.c_plan_ahead */ + + rcu_read_lock(); + if (rcu_dereference(mdev->rs_plan_s)->size) { number = drbd_rs_controller(mdev) >> (BM_BLOCK_SHIFT - 9); mdev->c_sync_rate = number * HZ * (BM_BLOCK_SIZE / 1024) / SLEEP_TIME; } else { - mdev->c_sync_rate = mdev->sync_conf.rate; + mdev->c_sync_rate = rcu_dereference(mdev->ldev->disk_conf)->resync_rate; number = SLEEP_TIME * mdev->c_sync_rate / ((BM_BLOCK_SIZE / 1024) * HZ); } + rcu_read_unlock(); /* ignore the amount of pending requests, the resync controller should * throttle down to incoming reply rate soon enough anyways. */ return number; } -static int w_make_resync_request(struct drbd_conf *mdev, - struct drbd_work *w, int cancel) +int w_make_resync_request(struct drbd_work *w, int cancel) { + struct drbd_conf *mdev = w->mdev; unsigned long bit; sector_t sector; const sector_t capacity = drbd_get_capacity(mdev->this_bdev); @@ -538,12 +571,12 @@ static int w_make_resync_request(struct drbd_conf *mdev, int i = 0; if (unlikely(cancel)) - return 1; + return 0; if (mdev->rs_total == 0) { /* empty resync? */ drbd_resync_finished(mdev); - return 1; + return 0; } if (!get_ldev(mdev)) { @@ -552,7 +585,7 @@ static int w_make_resync_request(struct drbd_conf *mdev, to continue resync with a broken disk makes no sense at all */ dev_err(DEV, "Disk broke down during resync!\n"); - return 1; + return 0; } max_bio_size = queue_max_hw_sectors(mdev->rq_queue) << 9; @@ -562,15 +595,15 @@ static int w_make_resync_request(struct drbd_conf *mdev, for (i = 0; i < number; i++) { /* Stop generating RS requests, when half of the send buffer is filled */ - mutex_lock(&mdev->data.mutex); - if (mdev->data.socket) { - queued = mdev->data.socket->sk->sk_wmem_queued; - sndbuf = mdev->data.socket->sk->sk_sndbuf; + mutex_lock(&mdev->tconn->data.mutex); + if (mdev->tconn->data.socket) { + queued = mdev->tconn->data.socket->sk->sk_wmem_queued; + sndbuf = mdev->tconn->data.socket->sk->sk_sndbuf; } else { queued = 1; sndbuf = 0; } - mutex_unlock(&mdev->data.mutex); + mutex_unlock(&mdev->tconn->data.mutex); if (queued > sndbuf / 2) goto requeue; @@ -581,7 +614,7 @@ next_sector: if (bit == DRBD_END_OF_BITMAP) { mdev->bm_resync_fo = drbd_bm_bits(mdev); put_ldev(mdev); - return 1; + return 0; } sector = BM_BIT_TO_SECT(bit); @@ -640,11 +673,11 @@ next_sector: /* adjust very last sectors, in case we are oddly sized */ if (sector + (size>>9) > capacity) size = (capacity-sector)<<9; - if (mdev->agreed_pro_version >= 89 && mdev->csums_tfm) { + if (mdev->tconn->agreed_pro_version >= 89 && mdev->tconn->csums_tfm) { switch (read_for_csum(mdev, sector, size)) { case -EIO: /* Disk failure */ put_ldev(mdev); - return 0; + return -EIO; case -EAGAIN: /* allocation failed, or ldev busy */ drbd_rs_complete_io(mdev, sector); mdev->bm_resync_fo = BM_SECT_TO_BIT(sector); @@ -657,13 +690,16 @@ next_sector: BUG(); } } else { + int err; + inc_rs_pending(mdev); - if (!drbd_send_drequest(mdev, P_RS_DATA_REQUEST, - sector, size, ID_SYNCER)) { + err = drbd_send_drequest(mdev, P_RS_DATA_REQUEST, + sector, size, ID_SYNCER); + if (err) { dev_err(DEV, "drbd_send_drequest() failed, aborting...\n"); dec_rs_pending(mdev); put_ldev(mdev); - return 0; + return err; } } } @@ -676,21 +712,23 @@ next_sector: * until then resync "work" is "inactive" ... */ put_ldev(mdev); - return 1; + return 0; } requeue: mdev->rs_in_flight += (i << (BM_BLOCK_SHIFT - 9)); mod_timer(&mdev->resync_timer, jiffies + SLEEP_TIME); put_ldev(mdev); - return 1; + return 0; } -static int w_make_ov_request(struct drbd_conf *mdev, struct drbd_work *w, int cancel) +static int w_make_ov_request(struct drbd_work *w, int cancel) { + struct drbd_conf *mdev = w->mdev; int number, i, size; sector_t sector; const sector_t capacity = drbd_get_capacity(mdev->this_bdev); + bool stop_sector_reached = false; if (unlikely(cancel)) return 1; @@ -699,9 +737,17 @@ static int w_make_ov_request(struct drbd_conf *mdev, struct drbd_work *w, int ca sector = mdev->ov_position; for (i = 0; i < number; i++) { - if (sector >= capacity) { + if (sector >= capacity) return 1; - } + + /* We check for "finished" only in the reply path: + * w_e_end_ov_reply(). + * We need to send at least one request out. */ + stop_sector_reached = i > 0 + && verify_can_do_stop_sector(mdev) + && sector >= mdev->ov_stop_sector; + if (stop_sector_reached) + break; size = BM_BLOCK_SIZE; @@ -715,7 +761,7 @@ static int w_make_ov_request(struct drbd_conf *mdev, struct drbd_work *w, int ca size = (capacity-sector)<<9; inc_rs_pending(mdev); - if (!drbd_send_ov_request(mdev, sector, size)) { + if (drbd_send_ov_request(mdev, sector, size)) { dec_rs_pending(mdev); return 0; } @@ -725,56 +771,39 @@ static int w_make_ov_request(struct drbd_conf *mdev, struct drbd_work *w, int ca requeue: mdev->rs_in_flight += (i << (BM_BLOCK_SHIFT - 9)); - mod_timer(&mdev->resync_timer, jiffies + SLEEP_TIME); - return 1; -} - - -void start_resync_timer_fn(unsigned long data) -{ - struct drbd_conf *mdev = (struct drbd_conf *) data; - - drbd_queue_work(&mdev->data.work, &mdev->start_resync_work); -} - -int w_start_resync(struct drbd_conf *mdev, struct drbd_work *w, int cancel) -{ - if (atomic_read(&mdev->unacked_cnt) || atomic_read(&mdev->rs_pending_cnt)) { - dev_warn(DEV, "w_start_resync later...\n"); - mdev->start_resync_timer.expires = jiffies + HZ/10; - add_timer(&mdev->start_resync_timer); - return 1; - } - - drbd_start_resync(mdev, C_SYNC_SOURCE); - clear_bit(AHEAD_TO_SYNC_SOURCE, &mdev->flags); + if (i == 0 || !stop_sector_reached) + mod_timer(&mdev->resync_timer, jiffies + SLEEP_TIME); return 1; } -int w_ov_finished(struct drbd_conf *mdev, struct drbd_work *w, int cancel) +int w_ov_finished(struct drbd_work *w, int cancel) { + struct drbd_conf *mdev = w->mdev; kfree(w); - ov_oos_print(mdev); + ov_out_of_sync_print(mdev); drbd_resync_finished(mdev); - return 1; + return 0; } -static int w_resync_finished(struct drbd_conf *mdev, struct drbd_work *w, int cancel) +static int w_resync_finished(struct drbd_work *w, int cancel) { + struct drbd_conf *mdev = w->mdev; kfree(w); drbd_resync_finished(mdev); - return 1; + return 0; } static void ping_peer(struct drbd_conf *mdev) { - clear_bit(GOT_PING_ACK, &mdev->flags); - request_ping(mdev); - wait_event(mdev->misc_wait, - test_bit(GOT_PING_ACK, &mdev->flags) || mdev->state.conn < C_CONNECTED); + struct drbd_tconn *tconn = mdev->tconn; + + clear_bit(GOT_PING_ACK, &tconn->flags); + request_ping(tconn); + wait_event(tconn->ping_wait, + test_bit(GOT_PING_ACK, &tconn->flags) || mdev->state.conn < C_CONNECTED); } int drbd_resync_finished(struct drbd_conf *mdev) @@ -799,7 +828,8 @@ int drbd_resync_finished(struct drbd_conf *mdev) w = kmalloc(sizeof(struct drbd_work), GFP_ATOMIC); if (w) { w->cb = w_resync_finished; - drbd_queue_work(&mdev->data.work, w); + w->mdev = mdev; + drbd_queue_work(&mdev->tconn->sender_work, w); return 1; } dev_err(DEV, "Warn failed to drbd_rs_del_all() and to kmalloc(w).\n"); @@ -808,7 +838,12 @@ int drbd_resync_finished(struct drbd_conf *mdev) dt = (jiffies - mdev->rs_start - mdev->rs_paused) / HZ; if (dt <= 0) dt = 1; + db = mdev->rs_total; + /* adjust for verify start and stop sectors, respective reached position */ + if (mdev->state.conn == C_VERIFY_S || mdev->state.conn == C_VERIFY_T) + db -= mdev->ov_left; + dbdt = Bit2KB(db/dt); mdev->rs_paused /= HZ; @@ -817,8 +852,8 @@ int drbd_resync_finished(struct drbd_conf *mdev) ping_peer(mdev); - spin_lock_irq(&mdev->req_lock); - os = mdev->state; + spin_lock_irq(&mdev->tconn->req_lock); + os = drbd_read_state(mdev); verify_done = (os.conn == C_VERIFY_S || os.conn == C_VERIFY_T); @@ -831,7 +866,7 @@ int drbd_resync_finished(struct drbd_conf *mdev) ns.conn = C_CONNECTED; dev_info(DEV, "%s done (total %lu sec; paused %lu sec; %lu K/sec)\n", - verify_done ? "Online verify " : "Resync", + verify_done ? "Online verify" : "Resync", dt + mdev->rs_paused, mdev->rs_paused, dbdt); n_oos = drbd_bm_total_weight(mdev); @@ -848,7 +883,7 @@ int drbd_resync_finished(struct drbd_conf *mdev) if (os.conn == C_SYNC_TARGET || os.conn == C_PAUSED_SYNC_T) khelper_cmd = "after-resync-target"; - if (mdev->csums_tfm && mdev->rs_total) { + if (mdev->tconn->csums_tfm && mdev->rs_total) { const unsigned long s = mdev->rs_same_csum; const unsigned long t = mdev->rs_total; const int ratio = @@ -906,13 +941,15 @@ int drbd_resync_finished(struct drbd_conf *mdev) _drbd_set_state(mdev, ns, CS_VERBOSE, NULL); out_unlock: - spin_unlock_irq(&mdev->req_lock); + spin_unlock_irq(&mdev->tconn->req_lock); put_ldev(mdev); out: mdev->rs_total = 0; mdev->rs_failed = 0; mdev->rs_paused = 0; - if (verify_done) + + /* reset start sector, if we reached end of device */ + if (verify_done && mdev->ov_left == 0) mdev->ov_start_sector = 0; drbd_md_sync(mdev); @@ -924,19 +961,19 @@ out: } /* helper */ -static void move_to_net_ee_or_free(struct drbd_conf *mdev, struct drbd_epoch_entry *e) +static void move_to_net_ee_or_free(struct drbd_conf *mdev, struct drbd_peer_request *peer_req) { - if (drbd_ee_has_active_page(e)) { + if (drbd_peer_req_has_active_page(peer_req)) { /* This might happen if sendpage() has not finished */ - int i = (e->size + PAGE_SIZE -1) >> PAGE_SHIFT; + int i = (peer_req->i.size + PAGE_SIZE -1) >> PAGE_SHIFT; atomic_add(i, &mdev->pp_in_use_by_net); atomic_sub(i, &mdev->pp_in_use); - spin_lock_irq(&mdev->req_lock); - list_add_tail(&e->w.list, &mdev->net_ee); - spin_unlock_irq(&mdev->req_lock); + spin_lock_irq(&mdev->tconn->req_lock); + list_add_tail(&peer_req->w.list, &mdev->net_ee); + spin_unlock_irq(&mdev->tconn->req_lock); wake_up(&drbd_pp_wait); } else - drbd_free_ee(mdev, e); + drbd_free_peer_req(mdev, peer_req); } /** @@ -945,174 +982,177 @@ static void move_to_net_ee_or_free(struct drbd_conf *mdev, struct drbd_epoch_ent * @w: work object. * @cancel: The connection will be closed anyways */ -int w_e_end_data_req(struct drbd_conf *mdev, struct drbd_work *w, int cancel) +int w_e_end_data_req(struct drbd_work *w, int cancel) { - struct drbd_epoch_entry *e = container_of(w, struct drbd_epoch_entry, w); - int ok; + struct drbd_peer_request *peer_req = container_of(w, struct drbd_peer_request, w); + struct drbd_conf *mdev = w->mdev; + int err; if (unlikely(cancel)) { - drbd_free_ee(mdev, e); + drbd_free_peer_req(mdev, peer_req); dec_unacked(mdev); - return 1; + return 0; } - if (likely((e->flags & EE_WAS_ERROR) == 0)) { - ok = drbd_send_block(mdev, P_DATA_REPLY, e); + if (likely((peer_req->flags & EE_WAS_ERROR) == 0)) { + err = drbd_send_block(mdev, P_DATA_REPLY, peer_req); } else { if (__ratelimit(&drbd_ratelimit_state)) dev_err(DEV, "Sending NegDReply. sector=%llus.\n", - (unsigned long long)e->sector); + (unsigned long long)peer_req->i.sector); - ok = drbd_send_ack(mdev, P_NEG_DREPLY, e); + err = drbd_send_ack(mdev, P_NEG_DREPLY, peer_req); } dec_unacked(mdev); - move_to_net_ee_or_free(mdev, e); + move_to_net_ee_or_free(mdev, peer_req); - if (unlikely(!ok)) + if (unlikely(err)) dev_err(DEV, "drbd_send_block() failed\n"); - return ok; + return err; } /** - * w_e_end_rsdata_req() - Worker callback to send a P_RS_DATA_REPLY packet in response to a P_RS_DATA_REQUESTRS + * w_e_end_rsdata_req() - Worker callback to send a P_RS_DATA_REPLY packet in response to a P_RS_DATA_REQUEST * @mdev: DRBD device. * @w: work object. * @cancel: The connection will be closed anyways */ -int w_e_end_rsdata_req(struct drbd_conf *mdev, struct drbd_work *w, int cancel) +int w_e_end_rsdata_req(struct drbd_work *w, int cancel) { - struct drbd_epoch_entry *e = container_of(w, struct drbd_epoch_entry, w); - int ok; + struct drbd_peer_request *peer_req = container_of(w, struct drbd_peer_request, w); + struct drbd_conf *mdev = w->mdev; + int err; if (unlikely(cancel)) { - drbd_free_ee(mdev, e); + drbd_free_peer_req(mdev, peer_req); dec_unacked(mdev); - return 1; + return 0; } if (get_ldev_if_state(mdev, D_FAILED)) { - drbd_rs_complete_io(mdev, e->sector); + drbd_rs_complete_io(mdev, peer_req->i.sector); put_ldev(mdev); } if (mdev->state.conn == C_AHEAD) { - ok = drbd_send_ack(mdev, P_RS_CANCEL, e); - } else if (likely((e->flags & EE_WAS_ERROR) == 0)) { + err = drbd_send_ack(mdev, P_RS_CANCEL, peer_req); + } else if (likely((peer_req->flags & EE_WAS_ERROR) == 0)) { if (likely(mdev->state.pdsk >= D_INCONSISTENT)) { inc_rs_pending(mdev); - ok = drbd_send_block(mdev, P_RS_DATA_REPLY, e); + err = drbd_send_block(mdev, P_RS_DATA_REPLY, peer_req); } else { if (__ratelimit(&drbd_ratelimit_state)) dev_err(DEV, "Not sending RSDataReply, " "partner DISKLESS!\n"); - ok = 1; + err = 0; } } else { if (__ratelimit(&drbd_ratelimit_state)) dev_err(DEV, "Sending NegRSDReply. sector %llus.\n", - (unsigned long long)e->sector); + (unsigned long long)peer_req->i.sector); - ok = drbd_send_ack(mdev, P_NEG_RS_DREPLY, e); + err = drbd_send_ack(mdev, P_NEG_RS_DREPLY, peer_req); /* update resync data with failure */ - drbd_rs_failed_io(mdev, e->sector, e->size); + drbd_rs_failed_io(mdev, peer_req->i.sector, peer_req->i.size); } dec_unacked(mdev); - move_to_net_ee_or_free(mdev, e); + move_to_net_ee_or_free(mdev, peer_req); - if (unlikely(!ok)) + if (unlikely(err)) dev_err(DEV, "drbd_send_block() failed\n"); - return ok; + return err; } -int w_e_end_csum_rs_req(struct drbd_conf *mdev, struct drbd_work *w, int cancel) +int w_e_end_csum_rs_req(struct drbd_work *w, int cancel) { - struct drbd_epoch_entry *e = container_of(w, struct drbd_epoch_entry, w); + struct drbd_peer_request *peer_req = container_of(w, struct drbd_peer_request, w); + struct drbd_conf *mdev = w->mdev; struct digest_info *di; int digest_size; void *digest = NULL; - int ok, eq = 0; + int err, eq = 0; if (unlikely(cancel)) { - drbd_free_ee(mdev, e); + drbd_free_peer_req(mdev, peer_req); dec_unacked(mdev); - return 1; + return 0; } if (get_ldev(mdev)) { - drbd_rs_complete_io(mdev, e->sector); + drbd_rs_complete_io(mdev, peer_req->i.sector); put_ldev(mdev); } - di = e->digest; + di = peer_req->digest; - if (likely((e->flags & EE_WAS_ERROR) == 0)) { + if (likely((peer_req->flags & EE_WAS_ERROR) == 0)) { /* quick hack to try to avoid a race against reconfiguration. * a real fix would be much more involved, * introducing more locking mechanisms */ - if (mdev->csums_tfm) { - digest_size = crypto_hash_digestsize(mdev->csums_tfm); + if (mdev->tconn->csums_tfm) { + digest_size = crypto_hash_digestsize(mdev->tconn->csums_tfm); D_ASSERT(digest_size == di->digest_size); digest = kmalloc(digest_size, GFP_NOIO); } if (digest) { - drbd_csum_ee(mdev, mdev->csums_tfm, e, digest); + drbd_csum_ee(mdev, mdev->tconn->csums_tfm, peer_req, digest); eq = !memcmp(digest, di->digest, digest_size); kfree(digest); } if (eq) { - drbd_set_in_sync(mdev, e->sector, e->size); + drbd_set_in_sync(mdev, peer_req->i.sector, peer_req->i.size); /* rs_same_csums unit is BM_BLOCK_SIZE */ - mdev->rs_same_csum += e->size >> BM_BLOCK_SHIFT; - ok = drbd_send_ack(mdev, P_RS_IS_IN_SYNC, e); + mdev->rs_same_csum += peer_req->i.size >> BM_BLOCK_SHIFT; + err = drbd_send_ack(mdev, P_RS_IS_IN_SYNC, peer_req); } else { inc_rs_pending(mdev); - e->block_id = ID_SYNCER; /* By setting block_id, digest pointer becomes invalid! */ - e->flags &= ~EE_HAS_DIGEST; /* This e no longer has a digest pointer */ + peer_req->block_id = ID_SYNCER; /* By setting block_id, digest pointer becomes invalid! */ + peer_req->flags &= ~EE_HAS_DIGEST; /* This peer request no longer has a digest pointer */ kfree(di); - ok = drbd_send_block(mdev, P_RS_DATA_REPLY, e); + err = drbd_send_block(mdev, P_RS_DATA_REPLY, peer_req); } } else { - ok = drbd_send_ack(mdev, P_NEG_RS_DREPLY, e); + err = drbd_send_ack(mdev, P_NEG_RS_DREPLY, peer_req); if (__ratelimit(&drbd_ratelimit_state)) dev_err(DEV, "Sending NegDReply. I guess it gets messy.\n"); } dec_unacked(mdev); - move_to_net_ee_or_free(mdev, e); + move_to_net_ee_or_free(mdev, peer_req); - if (unlikely(!ok)) + if (unlikely(err)) dev_err(DEV, "drbd_send_block/ack() failed\n"); - return ok; + return err; } -/* TODO merge common code with w_e_send_csum */ -int w_e_end_ov_req(struct drbd_conf *mdev, struct drbd_work *w, int cancel) +int w_e_end_ov_req(struct drbd_work *w, int cancel) { - struct drbd_epoch_entry *e = container_of(w, struct drbd_epoch_entry, w); - sector_t sector = e->sector; - unsigned int size = e->size; + struct drbd_peer_request *peer_req = container_of(w, struct drbd_peer_request, w); + struct drbd_conf *mdev = w->mdev; + sector_t sector = peer_req->i.sector; + unsigned int size = peer_req->i.size; int digest_size; void *digest; - int ok = 1; + int err = 0; if (unlikely(cancel)) goto out; - digest_size = crypto_hash_digestsize(mdev->verify_tfm); + digest_size = crypto_hash_digestsize(mdev->tconn->verify_tfm); digest = kmalloc(digest_size, GFP_NOIO); if (!digest) { - ok = 0; /* terminate the connection in case the allocation failed */ + err = 1; /* terminate the connection in case the allocation failed */ goto out; } - if (likely(!(e->flags & EE_WAS_ERROR))) - drbd_csum_ee(mdev, mdev->verify_tfm, e, digest); + if (likely(!(peer_req->flags & EE_WAS_ERROR))) + drbd_csum_ee(mdev, mdev->tconn->verify_tfm, peer_req, digest); else memset(digest, 0, digest_size); @@ -1120,25 +1160,23 @@ int w_e_end_ov_req(struct drbd_conf *mdev, struct drbd_work *w, int cancel) * In case we block on congestion, we could otherwise run into * some distributed deadlock, if the other side blocks on * congestion as well, because our receiver blocks in - * drbd_pp_alloc due to pp_in_use > max_buffers. */ - drbd_free_ee(mdev, e); - e = NULL; + * drbd_alloc_pages due to pp_in_use > max_buffers. */ + drbd_free_peer_req(mdev, peer_req); + peer_req = NULL; inc_rs_pending(mdev); - ok = drbd_send_drequest_csum(mdev, sector, size, - digest, digest_size, - P_OV_REPLY); - if (!ok) + err = drbd_send_drequest_csum(mdev, sector, size, digest, digest_size, P_OV_REPLY); + if (err) dec_rs_pending(mdev); kfree(digest); out: - if (e) - drbd_free_ee(mdev, e); + if (peer_req) + drbd_free_peer_req(mdev, peer_req); dec_unacked(mdev); - return ok; + return err; } -void drbd_ov_oos_found(struct drbd_conf *mdev, sector_t sector, int size) +void drbd_ov_out_of_sync_found(struct drbd_conf *mdev, sector_t sector, int size) { if (mdev->ov_last_oos_start + mdev->ov_last_oos_size == sector) { mdev->ov_last_oos_size += size>>9; @@ -1149,36 +1187,38 @@ void drbd_ov_oos_found(struct drbd_conf *mdev, sector_t sector, int size) drbd_set_out_of_sync(mdev, sector, size); } -int w_e_end_ov_reply(struct drbd_conf *mdev, struct drbd_work *w, int cancel) +int w_e_end_ov_reply(struct drbd_work *w, int cancel) { - struct drbd_epoch_entry *e = container_of(w, struct drbd_epoch_entry, w); + struct drbd_peer_request *peer_req = container_of(w, struct drbd_peer_request, w); + struct drbd_conf *mdev = w->mdev; struct digest_info *di; void *digest; - sector_t sector = e->sector; - unsigned int size = e->size; + sector_t sector = peer_req->i.sector; + unsigned int size = peer_req->i.size; int digest_size; - int ok, eq = 0; + int err, eq = 0; + bool stop_sector_reached = false; if (unlikely(cancel)) { - drbd_free_ee(mdev, e); + drbd_free_peer_req(mdev, peer_req); dec_unacked(mdev); - return 1; + return 0; } /* after "cancel", because after drbd_disconnect/drbd_rs_cancel_all * the resync lru has been cleaned up already */ if (get_ldev(mdev)) { - drbd_rs_complete_io(mdev, e->sector); + drbd_rs_complete_io(mdev, peer_req->i.sector); put_ldev(mdev); } - di = e->digest; + di = peer_req->digest; - if (likely((e->flags & EE_WAS_ERROR) == 0)) { - digest_size = crypto_hash_digestsize(mdev->verify_tfm); + if (likely((peer_req->flags & EE_WAS_ERROR) == 0)) { + digest_size = crypto_hash_digestsize(mdev->tconn->verify_tfm); digest = kmalloc(digest_size, GFP_NOIO); if (digest) { - drbd_csum_ee(mdev, mdev->verify_tfm, e, digest); + drbd_csum_ee(mdev, mdev->tconn->verify_tfm, peer_req, digest); D_ASSERT(digest_size == di->digest_size); eq = !memcmp(digest, di->digest, digest_size); @@ -1186,19 +1226,19 @@ int w_e_end_ov_reply(struct drbd_conf *mdev, struct drbd_work *w, int cancel) } } - /* Free e and pages before send. - * In case we block on congestion, we could otherwise run into - * some distributed deadlock, if the other side blocks on - * congestion as well, because our receiver blocks in - * drbd_pp_alloc due to pp_in_use > max_buffers. */ - drbd_free_ee(mdev, e); + /* Free peer_req and pages before send. + * In case we block on congestion, we could otherwise run into + * some distributed deadlock, if the other side blocks on + * congestion as well, because our receiver blocks in + * drbd_alloc_pages due to pp_in_use > max_buffers. */ + drbd_free_peer_req(mdev, peer_req); if (!eq) - drbd_ov_oos_found(mdev, sector, size); + drbd_ov_out_of_sync_found(mdev, sector, size); else - ov_oos_print(mdev); + ov_out_of_sync_print(mdev); - ok = drbd_send_ack_ex(mdev, P_OV_RESULT, sector, size, - eq ? ID_IN_SYNC : ID_OUT_OF_SYNC); + err = drbd_send_ack_ex(mdev, P_OV_RESULT, sector, size, + eq ? ID_IN_SYNC : ID_OUT_OF_SYNC); dec_unacked(mdev); @@ -1208,73 +1248,102 @@ int w_e_end_ov_reply(struct drbd_conf *mdev, struct drbd_work *w, int cancel) if ((mdev->ov_left & 0x200) == 0x200) drbd_advance_rs_marks(mdev, mdev->ov_left); - if (mdev->ov_left == 0) { - ov_oos_print(mdev); + stop_sector_reached = verify_can_do_stop_sector(mdev) && + (sector + (size>>9)) >= mdev->ov_stop_sector; + + if (mdev->ov_left == 0 || stop_sector_reached) { + ov_out_of_sync_print(mdev); drbd_resync_finished(mdev); } - return ok; + return err; } -int w_prev_work_done(struct drbd_conf *mdev, struct drbd_work *w, int cancel) +int w_prev_work_done(struct drbd_work *w, int cancel) { struct drbd_wq_barrier *b = container_of(w, struct drbd_wq_barrier, w); + complete(&b->done); - return 1; + return 0; } -int w_send_barrier(struct drbd_conf *mdev, struct drbd_work *w, int cancel) +/* FIXME + * We need to track the number of pending barrier acks, + * and to be able to wait for them. + * See also comment in drbd_adm_attach before drbd_suspend_io. + */ +int drbd_send_barrier(struct drbd_tconn *tconn) { - struct drbd_tl_epoch *b = container_of(w, struct drbd_tl_epoch, w); - struct p_barrier *p = &mdev->data.sbuf.barrier; - int ok = 1; - - /* really avoid racing with tl_clear. w.cb may have been referenced - * just before it was reassigned and re-queued, so double check that. - * actually, this race was harmless, since we only try to send the - * barrier packet here, and otherwise do nothing with the object. - * but compare with the head of w_clear_epoch */ - spin_lock_irq(&mdev->req_lock); - if (w->cb != w_send_barrier || mdev->state.conn < C_CONNECTED) - cancel = 1; - spin_unlock_irq(&mdev->req_lock); - if (cancel) - return 1; + struct p_barrier *p; + struct drbd_socket *sock; - if (!drbd_get_data_sock(mdev)) - return 0; - p->barrier = b->br_number; - /* inc_ap_pending was done where this was queued. - * dec_ap_pending will be done in got_BarrierAck - * or (on connection loss) in w_clear_epoch. */ - ok = _drbd_send_cmd(mdev, mdev->data.socket, P_BARRIER, - (struct p_header80 *)p, sizeof(*p), 0); - drbd_put_data_sock(mdev); - - return ok; + sock = &tconn->data; + p = conn_prepare_command(tconn, sock); + if (!p) + return -EIO; + p->barrier = tconn->send.current_epoch_nr; + p->pad = 0; + tconn->send.current_epoch_writes = 0; + + return conn_send_command(tconn, sock, P_BARRIER, sizeof(*p), NULL, 0); } -int w_send_write_hint(struct drbd_conf *mdev, struct drbd_work *w, int cancel) +int w_send_write_hint(struct drbd_work *w, int cancel) { + struct drbd_conf *mdev = w->mdev; + struct drbd_socket *sock; + if (cancel) - return 1; - return drbd_send_short_cmd(mdev, P_UNPLUG_REMOTE); + return 0; + sock = &mdev->tconn->data; + if (!drbd_prepare_command(mdev, sock)) + return -EIO; + return drbd_send_command(mdev, sock, P_UNPLUG_REMOTE, 0, NULL, 0); } -int w_send_oos(struct drbd_conf *mdev, struct drbd_work *w, int cancel) +static void re_init_if_first_write(struct drbd_tconn *tconn, unsigned int epoch) +{ + if (!tconn->send.seen_any_write_yet) { + tconn->send.seen_any_write_yet = true; + tconn->send.current_epoch_nr = epoch; + tconn->send.current_epoch_writes = 0; + } +} + +static void maybe_send_barrier(struct drbd_tconn *tconn, unsigned int epoch) +{ + /* re-init if first write on this connection */ + if (!tconn->send.seen_any_write_yet) + return; + if (tconn->send.current_epoch_nr != epoch) { + if (tconn->send.current_epoch_writes) + drbd_send_barrier(tconn); + tconn->send.current_epoch_nr = epoch; + } +} + +int w_send_out_of_sync(struct drbd_work *w, int cancel) { struct drbd_request *req = container_of(w, struct drbd_request, w); - int ok; + struct drbd_conf *mdev = w->mdev; + struct drbd_tconn *tconn = mdev->tconn; + int err; if (unlikely(cancel)) { - req_mod(req, send_canceled); - return 1; + req_mod(req, SEND_CANCELED); + return 0; } - ok = drbd_send_oos(mdev, req); - req_mod(req, oos_handed_to_network); + /* this time, no tconn->send.current_epoch_writes++; + * If it was sent, it was the closing barrier for the last + * replicated epoch, before we went into AHEAD mode. + * No more barriers will be sent, until we leave AHEAD mode again. */ + maybe_send_barrier(tconn, req->epoch); + + err = drbd_send_out_of_sync(mdev, req); + req_mod(req, OOS_HANDED_TO_NETWORK); - return ok; + return err; } /** @@ -1283,20 +1352,26 @@ int w_send_oos(struct drbd_conf *mdev, struct drbd_work *w, int cancel) * @w: work object. * @cancel: The connection will be closed anyways */ -int w_send_dblock(struct drbd_conf *mdev, struct drbd_work *w, int cancel) +int w_send_dblock(struct drbd_work *w, int cancel) { struct drbd_request *req = container_of(w, struct drbd_request, w); - int ok; + struct drbd_conf *mdev = w->mdev; + struct drbd_tconn *tconn = mdev->tconn; + int err; if (unlikely(cancel)) { - req_mod(req, send_canceled); - return 1; + req_mod(req, SEND_CANCELED); + return 0; } - ok = drbd_send_dblock(mdev, req); - req_mod(req, ok ? handed_over_to_network : send_failed); + re_init_if_first_write(tconn, req->epoch); + maybe_send_barrier(tconn, req->epoch); + tconn->send.current_epoch_writes++; + + err = drbd_send_dblock(mdev, req); + req_mod(req, err ? SEND_FAILED : HANDED_OVER_TO_NETWORK); - return ok; + return err; } /** @@ -1305,57 +1380,61 @@ int w_send_dblock(struct drbd_conf *mdev, struct drbd_work *w, int cancel) * @w: work object. * @cancel: The connection will be closed anyways */ -int w_send_read_req(struct drbd_conf *mdev, struct drbd_work *w, int cancel) +int w_send_read_req(struct drbd_work *w, int cancel) { struct drbd_request *req = container_of(w, struct drbd_request, w); - int ok; + struct drbd_conf *mdev = w->mdev; + struct drbd_tconn *tconn = mdev->tconn; + int err; if (unlikely(cancel)) { - req_mod(req, send_canceled); - return 1; + req_mod(req, SEND_CANCELED); + return 0; } - ok = drbd_send_drequest(mdev, P_DATA_REQUEST, req->sector, req->size, - (unsigned long)req); + /* Even read requests may close a write epoch, + * if there was any yet. */ + maybe_send_barrier(tconn, req->epoch); - if (!ok) { - /* ?? we set C_TIMEOUT or C_BROKEN_PIPE in drbd_send(); - * so this is probably redundant */ - if (mdev->state.conn >= C_CONNECTED) - drbd_force_state(mdev, NS(conn, C_NETWORK_FAILURE)); - } - req_mod(req, ok ? handed_over_to_network : send_failed); + err = drbd_send_drequest(mdev, P_DATA_REQUEST, req->i.sector, req->i.size, + (unsigned long)req); + + req_mod(req, err ? SEND_FAILED : HANDED_OVER_TO_NETWORK); - return ok; + return err; } -int w_restart_disk_io(struct drbd_conf *mdev, struct drbd_work *w, int cancel) +int w_restart_disk_io(struct drbd_work *w, int cancel) { struct drbd_request *req = container_of(w, struct drbd_request, w); + struct drbd_conf *mdev = w->mdev; if (bio_data_dir(req->master_bio) == WRITE && req->rq_state & RQ_IN_ACT_LOG) - drbd_al_begin_io(mdev, req->sector); - /* Calling drbd_al_begin_io() out of the worker might deadlocks - theoretically. Practically it can not deadlock, since this is - only used when unfreezing IOs. All the extents of the requests - that made it into the TL are already active */ + drbd_al_begin_io(mdev, &req->i); drbd_req_make_private_bio(req, req->master_bio); req->private_bio->bi_bdev = mdev->ldev->backing_bdev; generic_make_request(req->private_bio); - return 1; + return 0; } static int _drbd_may_sync_now(struct drbd_conf *mdev) { struct drbd_conf *odev = mdev; + int resync_after; while (1) { - if (odev->sync_conf.after == -1) + if (!odev->ldev) + return 1; + rcu_read_lock(); + resync_after = rcu_dereference(odev->ldev->disk_conf)->resync_after; + rcu_read_unlock(); + if (resync_after == -1) + return 1; + odev = minor_to_mdev(resync_after); + if (!expect(odev)) return 1; - odev = minor_to_mdev(odev->sync_conf.after); - ERR_IF(!odev) return 1; if ((odev->state.conn >= C_SYNC_SOURCE && odev->state.conn <= C_PAUSED_SYNC_T) || odev->state.aftr_isp || odev->state.peer_isp || @@ -1375,16 +1454,15 @@ static int _drbd_pause_after(struct drbd_conf *mdev) struct drbd_conf *odev; int i, rv = 0; - for (i = 0; i < minor_count; i++) { - odev = minor_to_mdev(i); - if (!odev) - continue; + rcu_read_lock(); + idr_for_each_entry(&minors, odev, i) { if (odev->state.conn == C_STANDALONE && odev->state.disk == D_DISKLESS) continue; if (!_drbd_may_sync_now(odev)) rv |= (__drbd_set_state(_NS(odev, aftr_isp, 1), CS_HARD, NULL) != SS_NOTHING_TO_DO); } + rcu_read_unlock(); return rv; } @@ -1400,10 +1478,8 @@ static int _drbd_resume_next(struct drbd_conf *mdev) struct drbd_conf *odev; int i, rv = 0; - for (i = 0; i < minor_count; i++) { - odev = minor_to_mdev(i); - if (!odev) - continue; + rcu_read_lock(); + idr_for_each_entry(&minors, odev, i) { if (odev->state.conn == C_STANDALONE && odev->state.disk == D_DISKLESS) continue; if (odev->state.aftr_isp) { @@ -1413,6 +1489,7 @@ static int _drbd_resume_next(struct drbd_conf *mdev) != SS_NOTHING_TO_DO) ; } } + rcu_read_unlock(); return rv; } @@ -1430,57 +1507,86 @@ void suspend_other_sg(struct drbd_conf *mdev) write_unlock_irq(&global_state_lock); } -static int sync_after_error(struct drbd_conf *mdev, int o_minor) +/* caller must hold global_state_lock */ +enum drbd_ret_code drbd_resync_after_valid(struct drbd_conf *mdev, int o_minor) { struct drbd_conf *odev; + int resync_after; if (o_minor == -1) return NO_ERROR; if (o_minor < -1 || minor_to_mdev(o_minor) == NULL) - return ERR_SYNC_AFTER; + return ERR_RESYNC_AFTER; /* check for loops */ odev = minor_to_mdev(o_minor); while (1) { if (odev == mdev) - return ERR_SYNC_AFTER_CYCLE; + return ERR_RESYNC_AFTER_CYCLE; + rcu_read_lock(); + resync_after = rcu_dereference(odev->ldev->disk_conf)->resync_after; + rcu_read_unlock(); /* dependency chain ends here, no cycles. */ - if (odev->sync_conf.after == -1) + if (resync_after == -1) return NO_ERROR; /* follow the dependency chain */ - odev = minor_to_mdev(odev->sync_conf.after); + odev = minor_to_mdev(resync_after); } } -int drbd_alter_sa(struct drbd_conf *mdev, int na) +/* caller must hold global_state_lock */ +void drbd_resync_after_changed(struct drbd_conf *mdev) { int changes; - int retcode; - write_lock_irq(&global_state_lock); - retcode = sync_after_error(mdev, na); - if (retcode == NO_ERROR) { - mdev->sync_conf.after = na; - do { - changes = _drbd_pause_after(mdev); - changes |= _drbd_resume_next(mdev); - } while (changes); - } - write_unlock_irq(&global_state_lock); - return retcode; + do { + changes = _drbd_pause_after(mdev); + changes |= _drbd_resume_next(mdev); + } while (changes); } void drbd_rs_controller_reset(struct drbd_conf *mdev) { + struct fifo_buffer *plan; + atomic_set(&mdev->rs_sect_in, 0); atomic_set(&mdev->rs_sect_ev, 0); mdev->rs_in_flight = 0; - mdev->rs_planed = 0; - spin_lock(&mdev->peer_seq_lock); - fifo_set(&mdev->rs_plan_s, 0); - spin_unlock(&mdev->peer_seq_lock); + + /* Updating the RCU protected object in place is necessary since + this function gets called from atomic context. + It is valid since all other updates also lead to an completely + empty fifo */ + rcu_read_lock(); + plan = rcu_dereference(mdev->rs_plan_s); + plan->total = 0; + fifo_set(plan, 0); + rcu_read_unlock(); +} + +void start_resync_timer_fn(unsigned long data) +{ + struct drbd_conf *mdev = (struct drbd_conf *) data; + + drbd_queue_work(&mdev->tconn->sender_work, &mdev->start_resync_work); +} + +int w_start_resync(struct drbd_work *w, int cancel) +{ + struct drbd_conf *mdev = w->mdev; + + if (atomic_read(&mdev->unacked_cnt) || atomic_read(&mdev->rs_pending_cnt)) { + dev_warn(DEV, "w_start_resync later...\n"); + mdev->start_resync_timer.expires = jiffies + HZ/10; + add_timer(&mdev->start_resync_timer); + return 0; + } + + drbd_start_resync(mdev, C_SYNC_SOURCE); + clear_bit(AHEAD_TO_SYNC_SOURCE, &mdev->flags); + return 0; } /** @@ -1501,43 +1607,58 @@ void drbd_start_resync(struct drbd_conf *mdev, enum drbd_conns side) return; } - if (side == C_SYNC_TARGET) { - /* Since application IO was locked out during C_WF_BITMAP_T and - C_WF_SYNC_UUID we are still unmodified. Before going to C_SYNC_TARGET - we check that we might make the data inconsistent. */ - r = drbd_khelper(mdev, "before-resync-target"); - r = (r >> 8) & 0xff; - if (r > 0) { - dev_info(DEV, "before-resync-target handler returned %d, " - "dropping connection.\n", r); - drbd_force_state(mdev, NS(conn, C_DISCONNECTING)); - return; - } - } else /* C_SYNC_SOURCE */ { - r = drbd_khelper(mdev, "before-resync-source"); - r = (r >> 8) & 0xff; - if (r > 0) { - if (r == 3) { - dev_info(DEV, "before-resync-source handler returned %d, " - "ignoring. Old userland tools?", r); - } else { - dev_info(DEV, "before-resync-source handler returned %d, " + if (!test_bit(B_RS_H_DONE, &mdev->flags)) { + if (side == C_SYNC_TARGET) { + /* Since application IO was locked out during C_WF_BITMAP_T and + C_WF_SYNC_UUID we are still unmodified. Before going to C_SYNC_TARGET + we check that we might make the data inconsistent. */ + r = drbd_khelper(mdev, "before-resync-target"); + r = (r >> 8) & 0xff; + if (r > 0) { + dev_info(DEV, "before-resync-target handler returned %d, " "dropping connection.\n", r); - drbd_force_state(mdev, NS(conn, C_DISCONNECTING)); + conn_request_state(mdev->tconn, NS(conn, C_DISCONNECTING), CS_HARD); return; } + } else /* C_SYNC_SOURCE */ { + r = drbd_khelper(mdev, "before-resync-source"); + r = (r >> 8) & 0xff; + if (r > 0) { + if (r == 3) { + dev_info(DEV, "before-resync-source handler returned %d, " + "ignoring. Old userland tools?", r); + } else { + dev_info(DEV, "before-resync-source handler returned %d, " + "dropping connection.\n", r); + conn_request_state(mdev->tconn, NS(conn, C_DISCONNECTING), CS_HARD); + return; + } + } } } - drbd_state_lock(mdev); + if (current == mdev->tconn->worker.task) { + /* The worker should not sleep waiting for state_mutex, + that can take long */ + if (!mutex_trylock(mdev->state_mutex)) { + set_bit(B_RS_H_DONE, &mdev->flags); + mdev->start_resync_timer.expires = jiffies + HZ/5; + add_timer(&mdev->start_resync_timer); + return; + } + } else { + mutex_lock(mdev->state_mutex); + } + clear_bit(B_RS_H_DONE, &mdev->flags); + write_lock_irq(&global_state_lock); if (!get_ldev_if_state(mdev, D_NEGOTIATING)) { write_unlock_irq(&global_state_lock); - drbd_state_unlock(mdev); + mutex_unlock(mdev->state_mutex); return; } - ns.i = mdev->state.i; + ns = drbd_read_state(mdev); ns.aftr_isp = !_drbd_may_sync_now(mdev); @@ -1549,7 +1670,7 @@ void drbd_start_resync(struct drbd_conf *mdev, enum drbd_conns side) ns.pdsk = D_INCONSISTENT; r = __drbd_set_state(mdev, ns, CS_VERBOSE, NULL); - ns = mdev->state; + ns = drbd_read_state(mdev); if (ns.conn < C_CONNECTED) r = SS_UNKNOWN_ERROR; @@ -1575,6 +1696,10 @@ void drbd_start_resync(struct drbd_conf *mdev, enum drbd_conns side) write_unlock_irq(&global_state_lock); if (r == SS_SUCCESS) { + /* reset rs_last_bcast when a resync or verify is started, + * to deal with potential jiffies wrap. */ + mdev->rs_last_bcast = jiffies - HZ; + dev_info(DEV, "Began resync as %s (will sync %lu KB [%lu bits set]).\n", drbd_conn_str(ns.conn), (unsigned long) mdev->rs_total << (BM_BLOCK_SHIFT-10), @@ -1589,10 +1714,10 @@ void drbd_start_resync(struct drbd_conf *mdev, enum drbd_conns side) * drbd_resync_finished from here in that case. * We drbd_gen_and_send_sync_uuid here for protocol < 96, * and from after_state_ch otherwise. */ - if (side == C_SYNC_SOURCE && mdev->agreed_pro_version < 96) + if (side == C_SYNC_SOURCE && mdev->tconn->agreed_pro_version < 96) drbd_gen_and_send_sync_uuid(mdev); - if (mdev->agreed_pro_version < 95 && mdev->rs_total == 0) { + if (mdev->tconn->agreed_pro_version < 95 && mdev->rs_total == 0) { /* This still has a race (about when exactly the peers * detect connection loss) that can lead to a full sync * on next handshake. In 8.3.9 we fixed this with explicit @@ -1603,10 +1728,16 @@ void drbd_start_resync(struct drbd_conf *mdev, enum drbd_conns side) * detect connection loss, then waiting for a ping * response (implicit in drbd_resync_finished) reduces * the race considerably, but does not solve it. */ - if (side == C_SYNC_SOURCE) - schedule_timeout_interruptible( - mdev->net_conf->ping_int * HZ + - mdev->net_conf->ping_timeo*HZ/9); + if (side == C_SYNC_SOURCE) { + struct net_conf *nc; + int timeo; + + rcu_read_lock(); + nc = rcu_dereference(mdev->tconn->net_conf); + timeo = nc->ping_int * HZ + nc->ping_timeo * HZ / 9; + rcu_read_unlock(); + schedule_timeout_interruptible(timeo); + } drbd_resync_finished(mdev); } @@ -1621,114 +1752,180 @@ void drbd_start_resync(struct drbd_conf *mdev, enum drbd_conns side) drbd_md_sync(mdev); } put_ldev(mdev); - drbd_state_unlock(mdev); + mutex_unlock(mdev->state_mutex); } -int drbd_worker(struct drbd_thread *thi) +/* If the resource already closed the current epoch, but we did not + * (because we have not yet seen new requests), we should send the + * corresponding barrier now. Must be checked within the same spinlock + * that is used to check for new requests. */ +bool need_to_send_barrier(struct drbd_tconn *connection) { - struct drbd_conf *mdev = thi->mdev; - struct drbd_work *w = NULL; - LIST_HEAD(work_list); - int intr = 0, i; + if (!connection->send.seen_any_write_yet) + return false; + + /* Skip barriers that do not contain any writes. + * This may happen during AHEAD mode. */ + if (!connection->send.current_epoch_writes) + return false; + + /* ->req_lock is held when requests are queued on + * connection->sender_work, and put into ->transfer_log. + * It is also held when ->current_tle_nr is increased. + * So either there are already new requests queued, + * and corresponding barriers will be send there. + * Or nothing new is queued yet, so the difference will be 1. + */ + if (atomic_read(&connection->current_tle_nr) != + connection->send.current_epoch_nr + 1) + return false; + + return true; +} + +bool dequeue_work_batch(struct drbd_work_queue *queue, struct list_head *work_list) +{ + spin_lock_irq(&queue->q_lock); + list_splice_init(&queue->q, work_list); + spin_unlock_irq(&queue->q_lock); + return !list_empty(work_list); +} - sprintf(current->comm, "drbd%d_worker", mdev_to_minor(mdev)); +bool dequeue_work_item(struct drbd_work_queue *queue, struct list_head *work_list) +{ + spin_lock_irq(&queue->q_lock); + if (!list_empty(&queue->q)) + list_move(queue->q.next, work_list); + spin_unlock_irq(&queue->q_lock); + return !list_empty(work_list); +} - while (get_t_state(thi) == Running) { - drbd_thread_current_set_cpu(mdev); +void wait_for_work(struct drbd_tconn *connection, struct list_head *work_list) +{ + DEFINE_WAIT(wait); + struct net_conf *nc; + int uncork, cork; - if (down_trylock(&mdev->data.work.s)) { - mutex_lock(&mdev->data.mutex); - if (mdev->data.socket && !mdev->net_conf->no_cork) - drbd_tcp_uncork(mdev->data.socket); - mutex_unlock(&mdev->data.mutex); + dequeue_work_item(&connection->sender_work, work_list); + if (!list_empty(work_list)) + return; - intr = down_interruptible(&mdev->data.work.s); + /* Still nothing to do? + * Maybe we still need to close the current epoch, + * even if no new requests are queued yet. + * + * Also, poke TCP, just in case. + * Then wait for new work (or signal). */ + rcu_read_lock(); + nc = rcu_dereference(connection->net_conf); + uncork = nc ? nc->tcp_cork : 0; + rcu_read_unlock(); + if (uncork) { + mutex_lock(&connection->data.mutex); + if (connection->data.socket) + drbd_tcp_uncork(connection->data.socket); + mutex_unlock(&connection->data.mutex); + } - mutex_lock(&mdev->data.mutex); - if (mdev->data.socket && !mdev->net_conf->no_cork) - drbd_tcp_cork(mdev->data.socket); - mutex_unlock(&mdev->data.mutex); + for (;;) { + int send_barrier; + prepare_to_wait(&connection->sender_work.q_wait, &wait, TASK_INTERRUPTIBLE); + spin_lock_irq(&connection->req_lock); + spin_lock(&connection->sender_work.q_lock); /* FIXME get rid of this one? */ + /* dequeue single item only, + * we still use drbd_queue_work_front() in some places */ + if (!list_empty(&connection->sender_work.q)) + list_move(connection->sender_work.q.next, work_list); + spin_unlock(&connection->sender_work.q_lock); /* FIXME get rid of this one? */ + if (!list_empty(work_list) || signal_pending(current)) { + spin_unlock_irq(&connection->req_lock); + break; } + send_barrier = need_to_send_barrier(connection); + spin_unlock_irq(&connection->req_lock); + if (send_barrier) { + drbd_send_barrier(connection); + connection->send.current_epoch_nr++; + } + schedule(); + /* may be woken up for other things but new work, too, + * e.g. if the current epoch got closed. + * In which case we send the barrier above. */ + } + finish_wait(&connection->sender_work.q_wait, &wait); + + /* someone may have changed the config while we have been waiting above. */ + rcu_read_lock(); + nc = rcu_dereference(connection->net_conf); + cork = nc ? nc->tcp_cork : 0; + rcu_read_unlock(); + mutex_lock(&connection->data.mutex); + if (connection->data.socket) { + if (cork) + drbd_tcp_cork(connection->data.socket); + else if (!uncork) + drbd_tcp_uncork(connection->data.socket); + } + mutex_unlock(&connection->data.mutex); +} - if (intr) { - D_ASSERT(intr == -EINTR); +int drbd_worker(struct drbd_thread *thi) +{ + struct drbd_tconn *tconn = thi->tconn; + struct drbd_work *w = NULL; + struct drbd_conf *mdev; + LIST_HEAD(work_list); + int vnr; + + while (get_t_state(thi) == RUNNING) { + drbd_thread_current_set_cpu(thi); + + /* as long as we use drbd_queue_work_front(), + * we may only dequeue single work items here, not batches. */ + if (list_empty(&work_list)) + wait_for_work(tconn, &work_list); + + if (signal_pending(current)) { flush_signals(current); - ERR_IF (get_t_state(thi) == Running) + if (get_t_state(thi) == RUNNING) { + conn_warn(tconn, "Worker got an unexpected signal\n"); continue; + } break; } - if (get_t_state(thi) != Running) + if (get_t_state(thi) != RUNNING) break; - /* With this break, we have done a down() but not consumed - the entry from the list. The cleanup code takes care of - this... */ - - w = NULL; - spin_lock_irq(&mdev->data.work.q_lock); - ERR_IF(list_empty(&mdev->data.work.q)) { - /* something terribly wrong in our logic. - * we were able to down() the semaphore, - * but the list is empty... doh. - * - * what is the best thing to do now? - * try again from scratch, restarting the receiver, - * asender, whatnot? could break even more ugly, - * e.g. when we are primary, but no good local data. - * - * I'll try to get away just starting over this loop. - */ - spin_unlock_irq(&mdev->data.work.q_lock); - continue; - } - w = list_entry(mdev->data.work.q.next, struct drbd_work, list); - list_del_init(&w->list); - spin_unlock_irq(&mdev->data.work.q_lock); - - if (!w->cb(mdev, w, mdev->state.conn < C_CONNECTED)) { - /* dev_warn(DEV, "worker: a callback failed! \n"); */ - if (mdev->state.conn >= C_CONNECTED) - drbd_force_state(mdev, - NS(conn, C_NETWORK_FAILURE)); + + while (!list_empty(&work_list)) { + w = list_first_entry(&work_list, struct drbd_work, list); + list_del_init(&w->list); + if (w->cb(w, tconn->cstate < C_WF_REPORT_PARAMS) == 0) + continue; + if (tconn->cstate >= C_WF_REPORT_PARAMS) + conn_request_state(tconn, NS(conn, C_NETWORK_FAILURE), CS_HARD); } } - D_ASSERT(test_bit(DEVICE_DYING, &mdev->flags)); - D_ASSERT(test_bit(CONFIG_PENDING, &mdev->flags)); - - spin_lock_irq(&mdev->data.work.q_lock); - i = 0; - while (!list_empty(&mdev->data.work.q)) { - list_splice_init(&mdev->data.work.q, &work_list); - spin_unlock_irq(&mdev->data.work.q_lock); + do { while (!list_empty(&work_list)) { - w = list_entry(work_list.next, struct drbd_work, list); + w = list_first_entry(&work_list, struct drbd_work, list); list_del_init(&w->list); - w->cb(mdev, w, 1); - i++; /* dead debugging code */ + w->cb(w, 1); } - - spin_lock_irq(&mdev->data.work.q_lock); + dequeue_work_batch(&tconn->sender_work, &work_list); + } while (!list_empty(&work_list)); + + rcu_read_lock(); + idr_for_each_entry(&tconn->volumes, mdev, vnr) { + D_ASSERT(mdev->state.disk == D_DISKLESS && mdev->state.conn == C_STANDALONE); + kref_get(&mdev->kref); + rcu_read_unlock(); + drbd_mdev_cleanup(mdev); + kref_put(&mdev->kref, &drbd_minor_destroy); + rcu_read_lock(); } - sema_init(&mdev->data.work.s, 0); - /* DANGEROUS race: if someone did queue his work within the spinlock, - * but up() ed outside the spinlock, we could get an up() on the - * semaphore without corresponding list entry. - * So don't do that. - */ - spin_unlock_irq(&mdev->data.work.q_lock); - - D_ASSERT(mdev->state.disk == D_DISKLESS && mdev->state.conn == C_STANDALONE); - /* _drbd_set_state only uses stop_nowait. - * wait here for the Exiting receiver. */ - drbd_thread_stop(&mdev->receiver); - drbd_mdev_cleanup(mdev); - - dev_info(DEV, "worker terminated\n"); - - clear_bit(DEVICE_DYING, &mdev->flags); - clear_bit(CONFIG_PENDING, &mdev->flags); - wake_up(&mdev->state_wait); + rcu_read_unlock(); return 0; } diff --git a/drivers/block/drbd/drbd_wrappers.h b/drivers/block/drbd/drbd_wrappers.h index 151f1a37478f..328f18e4b4ee 100644 --- a/drivers/block/drbd/drbd_wrappers.h +++ b/drivers/block/drbd/drbd_wrappers.h @@ -3,6 +3,7 @@ #include <linux/ctype.h> #include <linux/mm.h> +#include "drbd_int.h" /* see get_sb_bdev and bd_claim */ extern char *drbd_sec_holder; @@ -20,8 +21,8 @@ static inline void drbd_set_my_capacity(struct drbd_conf *mdev, /* bi_end_io handlers */ extern void drbd_md_io_complete(struct bio *bio, int error); -extern void drbd_endio_sec(struct bio *bio, int error); -extern void drbd_endio_pri(struct bio *bio, int error); +extern void drbd_peer_request_endio(struct bio *bio, int error); +extern void drbd_request_endio(struct bio *bio, int error); /* * used to submit our private bio @@ -45,12 +46,6 @@ static inline void drbd_generic_make_request(struct drbd_conf *mdev, generic_make_request(bio); } -static inline int drbd_crypto_is_hash(struct crypto_tfm *tfm) -{ - return (crypto_tfm_alg_type(tfm) & CRYPTO_ALG_TYPE_HASH_MASK) - == CRYPTO_ALG_TYPE_HASH; -} - #ifndef __CHECKER__ # undef __cond_lock # define __cond_lock(x,c) (c) diff --git a/drivers/block/loop.c b/drivers/block/loop.c index 54046e51160a..ae1251270624 100644 --- a/drivers/block/loop.c +++ b/drivers/block/loop.c @@ -463,6 +463,7 @@ out: */ static void loop_add_bio(struct loop_device *lo, struct bio *bio) { + lo->lo_bio_count++; bio_list_add(&lo->lo_bio_list, bio); } @@ -471,6 +472,7 @@ static void loop_add_bio(struct loop_device *lo, struct bio *bio) */ static struct bio *loop_get_bio(struct loop_device *lo) { + lo->lo_bio_count--; return bio_list_pop(&lo->lo_bio_list); } @@ -489,6 +491,10 @@ static void loop_make_request(struct request_queue *q, struct bio *old_bio) goto out; if (unlikely(rw == WRITE && (lo->lo_flags & LO_FLAGS_READ_ONLY))) goto out; + if (lo->lo_bio_count >= q->nr_congestion_on) + wait_event_lock_irq(lo->lo_req_wait, + lo->lo_bio_count < q->nr_congestion_off, + lo->lo_lock); loop_add_bio(lo, old_bio); wake_up(&lo->lo_event); spin_unlock_irq(&lo->lo_lock); @@ -546,6 +552,8 @@ static int loop_thread(void *data) continue; spin_lock_irq(&lo->lo_lock); bio = loop_get_bio(lo); + if (lo->lo_bio_count < lo->lo_queue->nr_congestion_off) + wake_up(&lo->lo_req_wait); spin_unlock_irq(&lo->lo_lock); BUG_ON(!bio); @@ -873,6 +881,7 @@ static int loop_set_fd(struct loop_device *lo, fmode_t mode, lo->transfer = transfer_none; lo->ioctl = NULL; lo->lo_sizelimit = 0; + lo->lo_bio_count = 0; lo->old_gfp_mask = mapping_gfp_mask(mapping); mapping_set_gfp_mask(mapping, lo->old_gfp_mask & ~(__GFP_IO|__GFP_FS)); @@ -1673,6 +1682,7 @@ static int loop_add(struct loop_device **l, int i) lo->lo_number = i; lo->lo_thread = NULL; init_waitqueue_head(&lo->lo_event); + init_waitqueue_head(&lo->lo_req_wait); spin_lock_init(&lo->lo_lock); disk->major = LOOP_MAJOR; disk->first_minor = i << part_shift; diff --git a/drivers/block/xen-blkback/blkback.c b/drivers/block/xen-blkback/blkback.c index 280a13846e6c..74374fb762aa 100644 --- a/drivers/block/xen-blkback/blkback.c +++ b/drivers/block/xen-blkback/blkback.c @@ -39,6 +39,7 @@ #include <linux/list.h> #include <linux/delay.h> #include <linux/freezer.h> +#include <linux/bitmap.h> #include <xen/events.h> #include <xen/page.h> @@ -79,6 +80,7 @@ struct pending_req { unsigned short operation; int status; struct list_head free_list; + DECLARE_BITMAP(unmap_seg, BLKIF_MAX_SEGMENTS_PER_REQUEST); }; #define BLKBACK_INVALID_HANDLE (~0) @@ -99,6 +101,36 @@ struct xen_blkbk { static struct xen_blkbk *blkbk; /* + * Maximum number of grant pages that can be mapped in blkback. + * BLKIF_MAX_SEGMENTS_PER_REQUEST * RING_SIZE is the maximum number of + * pages that blkback will persistently map. + * Currently, this is: + * RING_SIZE = 32 (for all known ring types) + * BLKIF_MAX_SEGMENTS_PER_REQUEST = 11 + * sizeof(struct persistent_gnt) = 48 + * So the maximum memory used to store the grants is: + * 32 * 11 * 48 = 16896 bytes + */ +static inline unsigned int max_mapped_grant_pages(enum blkif_protocol protocol) +{ + switch (protocol) { + case BLKIF_PROTOCOL_NATIVE: + return __CONST_RING_SIZE(blkif, PAGE_SIZE) * + BLKIF_MAX_SEGMENTS_PER_REQUEST; + case BLKIF_PROTOCOL_X86_32: + return __CONST_RING_SIZE(blkif_x86_32, PAGE_SIZE) * + BLKIF_MAX_SEGMENTS_PER_REQUEST; + case BLKIF_PROTOCOL_X86_64: + return __CONST_RING_SIZE(blkif_x86_64, PAGE_SIZE) * + BLKIF_MAX_SEGMENTS_PER_REQUEST; + default: + BUG(); + } + return 0; +} + + +/* * Little helpful macro to figure out the index and virtual address of the * pending_pages[..]. For each 'pending_req' we have have up to * BLKIF_MAX_SEGMENTS_PER_REQUEST (11) pages. The seg would be from 0 through @@ -129,6 +161,90 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif, static void make_response(struct xen_blkif *blkif, u64 id, unsigned short op, int st); +#define foreach_grant(pos, rbtree, node) \ + for ((pos) = container_of(rb_first((rbtree)), typeof(*(pos)), node); \ + &(pos)->node != NULL; \ + (pos) = container_of(rb_next(&(pos)->node), typeof(*(pos)), node)) + + +static void add_persistent_gnt(struct rb_root *root, + struct persistent_gnt *persistent_gnt) +{ + struct rb_node **new = &(root->rb_node), *parent = NULL; + struct persistent_gnt *this; + + /* Figure out where to put new node */ + while (*new) { + this = container_of(*new, struct persistent_gnt, node); + + parent = *new; + if (persistent_gnt->gnt < this->gnt) + new = &((*new)->rb_left); + else if (persistent_gnt->gnt > this->gnt) + new = &((*new)->rb_right); + else { + pr_alert(DRV_PFX " trying to add a gref that's already in the tree\n"); + BUG(); + } + } + + /* Add new node and rebalance tree. */ + rb_link_node(&(persistent_gnt->node), parent, new); + rb_insert_color(&(persistent_gnt->node), root); +} + +static struct persistent_gnt *get_persistent_gnt(struct rb_root *root, + grant_ref_t gref) +{ + struct persistent_gnt *data; + struct rb_node *node = root->rb_node; + + while (node) { + data = container_of(node, struct persistent_gnt, node); + + if (gref < data->gnt) + node = node->rb_left; + else if (gref > data->gnt) + node = node->rb_right; + else + return data; + } + return NULL; +} + +static void free_persistent_gnts(struct rb_root *root, unsigned int num) +{ + struct gnttab_unmap_grant_ref unmap[BLKIF_MAX_SEGMENTS_PER_REQUEST]; + struct page *pages[BLKIF_MAX_SEGMENTS_PER_REQUEST]; + struct persistent_gnt *persistent_gnt; + int ret = 0; + int segs_to_unmap = 0; + + foreach_grant(persistent_gnt, root, node) { + BUG_ON(persistent_gnt->handle == + BLKBACK_INVALID_HANDLE); + gnttab_set_unmap_op(&unmap[segs_to_unmap], + (unsigned long) pfn_to_kaddr(page_to_pfn( + persistent_gnt->page)), + GNTMAP_host_map, + persistent_gnt->handle); + + pages[segs_to_unmap] = persistent_gnt->page; + rb_erase(&persistent_gnt->node, root); + kfree(persistent_gnt); + num--; + + if (++segs_to_unmap == BLKIF_MAX_SEGMENTS_PER_REQUEST || + !rb_next(&persistent_gnt->node)) { + ret = gnttab_unmap_refs(unmap, NULL, pages, + segs_to_unmap); + BUG_ON(ret); + segs_to_unmap = 0; + } + } + BUG_ON(num != 0); +} + /* * Retrieve from the 'pending_reqs' a free pending_req structure to be used. */ @@ -302,6 +418,14 @@ int xen_blkif_schedule(void *arg) print_stats(blkif); } + /* Free all persistent grant pages */ + if (!RB_EMPTY_ROOT(&blkif->persistent_gnts)) + free_persistent_gnts(&blkif->persistent_gnts, + blkif->persistent_gnt_c); + + BUG_ON(!RB_EMPTY_ROOT(&blkif->persistent_gnts)); + blkif->persistent_gnt_c = 0; + if (log_stats) print_stats(blkif); @@ -328,6 +452,8 @@ static void xen_blkbk_unmap(struct pending_req *req) int ret; for (i = 0; i < req->nr_pages; i++) { + if (!test_bit(i, req->unmap_seg)) + continue; handle = pending_handle(req, i); if (handle == BLKBACK_INVALID_HANDLE) continue; @@ -344,12 +470,26 @@ static void xen_blkbk_unmap(struct pending_req *req) static int xen_blkbk_map(struct blkif_request *req, struct pending_req *pending_req, - struct seg_buf seg[]) + struct seg_buf seg[], + struct page *pages[]) { struct gnttab_map_grant_ref map[BLKIF_MAX_SEGMENTS_PER_REQUEST]; - int i; + struct persistent_gnt *persistent_gnts[BLKIF_MAX_SEGMENTS_PER_REQUEST]; + struct page *pages_to_gnt[BLKIF_MAX_SEGMENTS_PER_REQUEST]; + struct persistent_gnt *persistent_gnt = NULL; + struct xen_blkif *blkif = pending_req->blkif; + phys_addr_t addr = 0; + int i, j; + bool new_map; int nseg = req->u.rw.nr_segments; + int segs_to_map = 0; int ret = 0; + int use_persistent_gnts; + + use_persistent_gnts = (blkif->vbd.feature_gnt_persistent); + + BUG_ON(blkif->persistent_gnt_c > + max_mapped_grant_pages(pending_req->blkif->blk_protocol)); /* * Fill out preq.nr_sects with proper amount of sectors, and setup @@ -359,36 +499,146 @@ static int xen_blkbk_map(struct blkif_request *req, for (i = 0; i < nseg; i++) { uint32_t flags; - flags = GNTMAP_host_map; - if (pending_req->operation != BLKIF_OP_READ) - flags |= GNTMAP_readonly; - gnttab_set_map_op(&map[i], vaddr(pending_req, i), flags, - req->u.rw.seg[i].gref, - pending_req->blkif->domid); + if (use_persistent_gnts) + persistent_gnt = get_persistent_gnt( + &blkif->persistent_gnts, + req->u.rw.seg[i].gref); + + if (persistent_gnt) { + /* + * We are using persistent grants and + * the grant is already mapped + */ + new_map = false; + } else if (use_persistent_gnts && + blkif->persistent_gnt_c < + max_mapped_grant_pages(blkif->blk_protocol)) { + /* + * We are using persistent grants, the grant is + * not mapped but we have room for it + */ + new_map = true; + persistent_gnt = kmalloc( + sizeof(struct persistent_gnt), + GFP_KERNEL); + if (!persistent_gnt) + return -ENOMEM; + persistent_gnt->page = alloc_page(GFP_KERNEL); + if (!persistent_gnt->page) { + kfree(persistent_gnt); + return -ENOMEM; + } + persistent_gnt->gnt = req->u.rw.seg[i].gref; + persistent_gnt->handle = BLKBACK_INVALID_HANDLE; + + pages_to_gnt[segs_to_map] = + persistent_gnt->page; + addr = (unsigned long) pfn_to_kaddr( + page_to_pfn(persistent_gnt->page)); + + add_persistent_gnt(&blkif->persistent_gnts, + persistent_gnt); + blkif->persistent_gnt_c++; + pr_debug(DRV_PFX " grant %u added to the tree of persistent grants, using %u/%u\n", + persistent_gnt->gnt, blkif->persistent_gnt_c, + max_mapped_grant_pages(blkif->blk_protocol)); + } else { + /* + * We are either using persistent grants and + * hit the maximum limit of grants mapped, + * or we are not using persistent grants. + */ + if (use_persistent_gnts && + !blkif->vbd.overflow_max_grants) { + blkif->vbd.overflow_max_grants = 1; + pr_alert(DRV_PFX " domain %u, device %#x is using maximum number of persistent grants\n", + blkif->domid, blkif->vbd.handle); + } + new_map = true; + pages[i] = blkbk->pending_page(pending_req, i); + addr = vaddr(pending_req, i); + pages_to_gnt[segs_to_map] = + blkbk->pending_page(pending_req, i); + } + + if (persistent_gnt) { + pages[i] = persistent_gnt->page; + persistent_gnts[i] = persistent_gnt; + } else { + persistent_gnts[i] = NULL; + } + + if (new_map) { + flags = GNTMAP_host_map; + if (!persistent_gnt && + (pending_req->operation != BLKIF_OP_READ)) + flags |= GNTMAP_readonly; + gnttab_set_map_op(&map[segs_to_map++], addr, + flags, req->u.rw.seg[i].gref, + blkif->domid); + } } - ret = gnttab_map_refs(map, NULL, &blkbk->pending_page(pending_req, 0), nseg); - BUG_ON(ret); + if (segs_to_map) { + ret = gnttab_map_refs(map, NULL, pages_to_gnt, segs_to_map); + BUG_ON(ret); + } /* * Now swizzle the MFN in our domain with the MFN from the other domain * so that when we access vaddr(pending_req,i) it has the contents of * the page from the other domain. */ - for (i = 0; i < nseg; i++) { - if (unlikely(map[i].status != 0)) { - pr_debug(DRV_PFX "invalid buffer -- could not remap it\n"); - map[i].handle = BLKBACK_INVALID_HANDLE; - ret |= 1; + bitmap_zero(pending_req->unmap_seg, BLKIF_MAX_SEGMENTS_PER_REQUEST); + for (i = 0, j = 0; i < nseg; i++) { + if (!persistent_gnts[i] || + persistent_gnts[i]->handle == BLKBACK_INVALID_HANDLE) { + /* This is a newly mapped grant */ + BUG_ON(j >= segs_to_map); + if (unlikely(map[j].status != 0)) { + pr_debug(DRV_PFX "invalid buffer -- could not remap it\n"); + map[j].handle = BLKBACK_INVALID_HANDLE; + ret |= 1; + if (persistent_gnts[i]) { + rb_erase(&persistent_gnts[i]->node, + &blkif->persistent_gnts); + blkif->persistent_gnt_c--; + kfree(persistent_gnts[i]); + persistent_gnts[i] = NULL; + } + } + } + if (persistent_gnts[i]) { + if (persistent_gnts[i]->handle == + BLKBACK_INVALID_HANDLE) { + /* + * If this is a new persistent grant + * save the handler + */ + persistent_gnts[i]->handle = map[j].handle; + persistent_gnts[i]->dev_bus_addr = + map[j++].dev_bus_addr; + } + pending_handle(pending_req, i) = + persistent_gnts[i]->handle; + + if (ret) + continue; + + seg[i].buf = persistent_gnts[i]->dev_bus_addr | + (req->u.rw.seg[i].first_sect << 9); + } else { + pending_handle(pending_req, i) = map[j].handle; + bitmap_set(pending_req->unmap_seg, i, 1); + + if (ret) { + j++; + continue; + } + + seg[i].buf = map[j++].dev_bus_addr | + (req->u.rw.seg[i].first_sect << 9); } - - pending_handle(pending_req, i) = map[i].handle; - - if (ret) - continue; - - seg[i].buf = map[i].dev_bus_addr | - (req->u.rw.seg[i].first_sect << 9); } return ret; } @@ -591,6 +841,7 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif, int operation; struct blk_plug plug; bool drain = false; + struct page *pages[BLKIF_MAX_SEGMENTS_PER_REQUEST]; switch (req->operation) { case BLKIF_OP_READ: @@ -677,7 +928,7 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif, * the hypercall to unmap the grants - that is all done in * xen_blkbk_unmap. */ - if (xen_blkbk_map(req, pending_req, seg)) + if (xen_blkbk_map(req, pending_req, seg, pages)) goto fail_flush; /* @@ -689,7 +940,7 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif, for (i = 0; i < nseg; i++) { while ((bio == NULL) || (bio_add_page(bio, - blkbk->pending_page(pending_req, i), + pages[i], seg[i].nsec << 9, seg[i].buf & ~PAGE_MASK) == 0)) { diff --git a/drivers/block/xen-blkback/common.h b/drivers/block/xen-blkback/common.h index 9a54623e52d7..6072390c7f57 100644 --- a/drivers/block/xen-blkback/common.h +++ b/drivers/block/xen-blkback/common.h @@ -34,6 +34,7 @@ #include <linux/vmalloc.h> #include <linux/wait.h> #include <linux/io.h> +#include <linux/rbtree.h> #include <asm/setup.h> #include <asm/pgalloc.h> #include <asm/hypervisor.h> @@ -160,10 +161,21 @@ struct xen_vbd { sector_t size; unsigned int flush_support:1; unsigned int discard_secure:1; + unsigned int feature_gnt_persistent:1; + unsigned int overflow_max_grants:1; }; struct backend_info; + +struct persistent_gnt { + struct page *page; + grant_ref_t gnt; + grant_handle_t handle; + uint64_t dev_bus_addr; + struct rb_node node; +}; + struct xen_blkif { /* Unique identifier for this interface. */ domid_t domid; @@ -190,6 +202,10 @@ struct xen_blkif { struct task_struct *xenblkd; unsigned int waiting_reqs; + /* tree to store persistent grants */ + struct rb_root persistent_gnts; + unsigned int persistent_gnt_c; + /* statistics */ unsigned long st_print; int st_rd_req; diff --git a/drivers/block/xen-blkback/xenbus.c b/drivers/block/xen-blkback/xenbus.c index f58434c2617c..63980722db41 100644 --- a/drivers/block/xen-blkback/xenbus.c +++ b/drivers/block/xen-blkback/xenbus.c @@ -117,6 +117,7 @@ static struct xen_blkif *xen_blkif_alloc(domid_t domid) atomic_set(&blkif->drain, 0); blkif->st_print = jiffies; init_waitqueue_head(&blkif->waiting_to_free); + blkif->persistent_gnts.rb_node = NULL; return blkif; } @@ -672,6 +673,13 @@ again: xen_blkbk_barrier(xbt, be, be->blkif->vbd.flush_support); + err = xenbus_printf(xbt, dev->nodename, "feature-persistent", "%u", 1); + if (err) { + xenbus_dev_fatal(dev, err, "writing %s/feature-persistent", + dev->nodename); + goto abort; + } + err = xenbus_printf(xbt, dev->nodename, "sectors", "%llu", (unsigned long long)vbd_sz(&be->blkif->vbd)); if (err) { @@ -720,6 +728,7 @@ static int connect_ring(struct backend_info *be) struct xenbus_device *dev = be->dev; unsigned long ring_ref; unsigned int evtchn; + unsigned int pers_grants; char protocol[64] = ""; int err; @@ -749,8 +758,18 @@ static int connect_ring(struct backend_info *be) xenbus_dev_fatal(dev, err, "unknown fe protocol %s", protocol); return -1; } - pr_info(DRV_PFX "ring-ref %ld, event-channel %d, protocol %d (%s)\n", - ring_ref, evtchn, be->blkif->blk_protocol, protocol); + err = xenbus_gather(XBT_NIL, dev->otherend, + "feature-persistent", "%u", + &pers_grants, NULL); + if (err) + pers_grants = 0; + + be->blkif->vbd.feature_gnt_persistent = pers_grants; + be->blkif->vbd.overflow_max_grants = 0; + + pr_info(DRV_PFX "ring-ref %ld, event-channel %d, protocol %d (%s) %s\n", + ring_ref, evtchn, be->blkif->blk_protocol, protocol, + pers_grants ? "persistent grants" : ""); /* Map the shared frame, irq etc. */ err = xen_blkif_map(be->blkif, ring_ref, evtchn); diff --git a/drivers/block/xen-blkfront.c b/drivers/block/xen-blkfront.c index 007db8986e84..96e9b00db081 100644 --- a/drivers/block/xen-blkfront.c +++ b/drivers/block/xen-blkfront.c @@ -44,6 +44,7 @@ #include <linux/mutex.h> #include <linux/scatterlist.h> #include <linux/bitmap.h> +#include <linux/llist.h> #include <xen/xen.h> #include <xen/xenbus.h> @@ -64,10 +65,17 @@ enum blkif_state { BLKIF_STATE_SUSPENDED, }; +struct grant { + grant_ref_t gref; + unsigned long pfn; + struct llist_node node; +}; + struct blk_shadow { struct blkif_request req; struct request *request; unsigned long frame[BLKIF_MAX_SEGMENTS_PER_REQUEST]; + struct grant *grants_used[BLKIF_MAX_SEGMENTS_PER_REQUEST]; }; static DEFINE_MUTEX(blkfront_mutex); @@ -97,6 +105,8 @@ struct blkfront_info struct work_struct work; struct gnttab_free_callback callback; struct blk_shadow shadow[BLK_RING_SIZE]; + struct llist_head persistent_gnts; + unsigned int persistent_gnts_c; unsigned long shadow_free; unsigned int feature_flush; unsigned int flush_op; @@ -104,6 +114,7 @@ struct blkfront_info unsigned int feature_secdiscard:1; unsigned int discard_granularity; unsigned int discard_alignment; + unsigned int feature_persistent:1; int is_ready; }; @@ -287,21 +298,36 @@ static int blkif_queue_request(struct request *req) unsigned long id; unsigned int fsect, lsect; int i, ref; + + /* + * Used to store if we are able to queue the request by just using + * existing persistent grants, or if we have to get new grants, + * as there are not sufficiently many free. + */ + bool new_persistent_gnts; grant_ref_t gref_head; + struct page *granted_page; + struct grant *gnt_list_entry = NULL; struct scatterlist *sg; if (unlikely(info->connected != BLKIF_STATE_CONNECTED)) return 1; - if (gnttab_alloc_grant_references( - BLKIF_MAX_SEGMENTS_PER_REQUEST, &gref_head) < 0) { - gnttab_request_free_callback( - &info->callback, - blkif_restart_queue_callback, - info, - BLKIF_MAX_SEGMENTS_PER_REQUEST); - return 1; - } + /* Check if we have enought grants to allocate a requests */ + if (info->persistent_gnts_c < BLKIF_MAX_SEGMENTS_PER_REQUEST) { + new_persistent_gnts = 1; + if (gnttab_alloc_grant_references( + BLKIF_MAX_SEGMENTS_PER_REQUEST - info->persistent_gnts_c, + &gref_head) < 0) { + gnttab_request_free_callback( + &info->callback, + blkif_restart_queue_callback, + info, + BLKIF_MAX_SEGMENTS_PER_REQUEST); + return 1; + } + } else + new_persistent_gnts = 0; /* Fill out a communications ring structure. */ ring_req = RING_GET_REQUEST(&info->ring, info->ring.req_prod_pvt); @@ -341,18 +367,73 @@ static int blkif_queue_request(struct request *req) BLKIF_MAX_SEGMENTS_PER_REQUEST); for_each_sg(info->sg, sg, ring_req->u.rw.nr_segments, i) { - buffer_mfn = pfn_to_mfn(page_to_pfn(sg_page(sg))); fsect = sg->offset >> 9; lsect = fsect + (sg->length >> 9) - 1; - /* install a grant reference. */ - ref = gnttab_claim_grant_reference(&gref_head); - BUG_ON(ref == -ENOSPC); - gnttab_grant_foreign_access_ref( - ref, + if (info->persistent_gnts_c) { + BUG_ON(llist_empty(&info->persistent_gnts)); + gnt_list_entry = llist_entry( + llist_del_first(&info->persistent_gnts), + struct grant, node); + + ref = gnt_list_entry->gref; + buffer_mfn = pfn_to_mfn(gnt_list_entry->pfn); + info->persistent_gnts_c--; + } else { + ref = gnttab_claim_grant_reference(&gref_head); + BUG_ON(ref == -ENOSPC); + + gnt_list_entry = + kmalloc(sizeof(struct grant), + GFP_ATOMIC); + if (!gnt_list_entry) + return -ENOMEM; + + granted_page = alloc_page(GFP_ATOMIC); + if (!granted_page) { + kfree(gnt_list_entry); + return -ENOMEM; + } + + gnt_list_entry->pfn = + page_to_pfn(granted_page); + gnt_list_entry->gref = ref; + + buffer_mfn = pfn_to_mfn(page_to_pfn( + granted_page)); + gnttab_grant_foreign_access_ref(ref, info->xbdev->otherend_id, - buffer_mfn, - rq_data_dir(req)); + buffer_mfn, 0); + } + + info->shadow[id].grants_used[i] = gnt_list_entry; + + if (rq_data_dir(req)) { + char *bvec_data; + void *shared_data; + + BUG_ON(sg->offset + sg->length > PAGE_SIZE); + + shared_data = kmap_atomic( + pfn_to_page(gnt_list_entry->pfn)); + bvec_data = kmap_atomic(sg_page(sg)); + + /* + * this does not wipe data stored outside the + * range sg->offset..sg->offset+sg->length. + * Therefore, blkback *could* see data from + * previous requests. This is OK as long as + * persistent grants are shared with just one + * domain. It may need refactoring if this + * changes + */ + memcpy(shared_data + sg->offset, + bvec_data + sg->offset, + sg->length); + + kunmap_atomic(bvec_data); + kunmap_atomic(shared_data); + } info->shadow[id].frame[i] = mfn_to_pfn(buffer_mfn); ring_req->u.rw.seg[i] = @@ -368,7 +449,8 @@ static int blkif_queue_request(struct request *req) /* Keep a private copy so we can reissue requests when recovering. */ info->shadow[id].req = *ring_req; - gnttab_free_grant_references(gref_head); + if (new_persistent_gnts) + gnttab_free_grant_references(gref_head); return 0; } @@ -480,12 +562,13 @@ static int xlvbd_init_blk_queue(struct gendisk *gd, u16 sector_size) static void xlvbd_flush(struct blkfront_info *info) { blk_queue_flush(info->rq, info->feature_flush); - printk(KERN_INFO "blkfront: %s: %s: %s\n", + printk(KERN_INFO "blkfront: %s: %s: %s %s\n", info->gd->disk_name, info->flush_op == BLKIF_OP_WRITE_BARRIER ? "barrier" : (info->flush_op == BLKIF_OP_FLUSH_DISKCACHE ? "flush diskcache" : "barrier or flush"), - info->feature_flush ? "enabled" : "disabled"); + info->feature_flush ? "enabled" : "disabled", + info->feature_persistent ? "using persistent grants" : ""); } static int xen_translate_vdev(int vdevice, int *minor, unsigned int *offset) @@ -707,6 +790,9 @@ static void blkif_restart_queue(struct work_struct *work) static void blkif_free(struct blkfront_info *info, int suspend) { + struct llist_node *all_gnts; + struct grant *persistent_gnt; + /* Prevent new requests being issued until we fix things up. */ spin_lock_irq(&info->io_lock); info->connected = suspend ? @@ -714,6 +800,18 @@ static void blkif_free(struct blkfront_info *info, int suspend) /* No more blkif_request(). */ if (info->rq) blk_stop_queue(info->rq); + + /* Remove all persistent grants */ + if (info->persistent_gnts_c) { + all_gnts = llist_del_all(&info->persistent_gnts); + llist_for_each_entry(persistent_gnt, all_gnts, node) { + gnttab_end_foreign_access(persistent_gnt->gref, 0, 0UL); + __free_page(pfn_to_page(persistent_gnt->pfn)); + kfree(persistent_gnt); + } + info->persistent_gnts_c = 0; + } + /* No more gnttab callback work. */ gnttab_cancel_free_callback(&info->callback); spin_unlock_irq(&info->io_lock); @@ -734,13 +832,43 @@ static void blkif_free(struct blkfront_info *info, int suspend) } -static void blkif_completion(struct blk_shadow *s) +static void blkif_completion(struct blk_shadow *s, struct blkfront_info *info, + struct blkif_response *bret) { int i; - /* Do not let BLKIF_OP_DISCARD as nr_segment is in the same place - * flag. */ - for (i = 0; i < s->req.u.rw.nr_segments; i++) - gnttab_end_foreign_access(s->req.u.rw.seg[i].gref, 0, 0UL); + struct bio_vec *bvec; + struct req_iterator iter; + unsigned long flags; + char *bvec_data; + void *shared_data; + unsigned int offset = 0; + + if (bret->operation == BLKIF_OP_READ) { + /* + * Copy the data received from the backend into the bvec. + * Since bv_offset can be different than 0, and bv_len different + * than PAGE_SIZE, we have to keep track of the current offset, + * to be sure we are copying the data from the right shared page. + */ + rq_for_each_segment(bvec, s->request, iter) { + BUG_ON((bvec->bv_offset + bvec->bv_len) > PAGE_SIZE); + i = offset >> PAGE_SHIFT; + BUG_ON(i >= s->req.u.rw.nr_segments); + shared_data = kmap_atomic( + pfn_to_page(s->grants_used[i]->pfn)); + bvec_data = bvec_kmap_irq(bvec, &flags); + memcpy(bvec_data, shared_data + bvec->bv_offset, + bvec->bv_len); + bvec_kunmap_irq(bvec_data, &flags); + kunmap_atomic(shared_data); + offset += bvec->bv_len; + } + } + /* Add the persistent grant into the list of free grants */ + for (i = 0; i < s->req.u.rw.nr_segments; i++) { + llist_add(&s->grants_used[i]->node, &info->persistent_gnts); + info->persistent_gnts_c++; + } } static irqreturn_t blkif_interrupt(int irq, void *dev_id) @@ -783,7 +911,7 @@ static irqreturn_t blkif_interrupt(int irq, void *dev_id) req = info->shadow[id].request; if (bret->operation != BLKIF_OP_DISCARD) - blkif_completion(&info->shadow[id]); + blkif_completion(&info->shadow[id], info, bret); if (add_id_to_freelist(info, id)) { WARN(1, "%s: response to %s (id %ld) couldn't be recycled!\n", @@ -942,6 +1070,11 @@ again: message = "writing protocol"; goto abort_transaction; } + err = xenbus_printf(xbt, dev->nodename, + "feature-persistent", "%u", 1); + if (err) + dev_warn(&dev->dev, + "writing persistent grants feature to xenbus"); err = xenbus_transaction_end(xbt, 0); if (err) { @@ -1029,6 +1162,8 @@ static int blkfront_probe(struct xenbus_device *dev, spin_lock_init(&info->io_lock); info->xbdev = dev; info->vdevice = vdevice; + init_llist_head(&info->persistent_gnts); + info->persistent_gnts_c = 0; info->connected = BLKIF_STATE_DISCONNECTED; INIT_WORK(&info->work, blkif_restart_queue); @@ -1093,7 +1228,7 @@ static int blkif_recover(struct blkfront_info *info) req->u.rw.seg[j].gref, info->xbdev->otherend_id, pfn_to_mfn(info->shadow[req->u.rw.id].frame[j]), - rq_data_dir(info->shadow[req->u.rw.id].request)); + 0); } info->shadow[req->u.rw.id].req = *req; @@ -1225,7 +1360,7 @@ static void blkfront_connect(struct blkfront_info *info) unsigned long sector_size; unsigned int binfo; int err; - int barrier, flush, discard; + int barrier, flush, discard, persistent; switch (info->connected) { case BLKIF_STATE_CONNECTED: @@ -1303,6 +1438,14 @@ static void blkfront_connect(struct blkfront_info *info) if (!err && discard) blkfront_setup_discard(info); + err = xenbus_gather(XBT_NIL, info->xbdev->otherend, + "feature-persistent", "%u", &persistent, + NULL); + if (err) + info->feature_persistent = 0; + else + info->feature_persistent = persistent; + err = xlvbd_alloc_gendisk(sectors, info, binfo, sector_size); if (err) { xenbus_dev_fatal(info->xbdev, err, "xlvbd_add at %s", diff --git a/drivers/firmware/efivars.c b/drivers/firmware/efivars.c index 52c5d8956d7d..7b1c37497c9a 100644 --- a/drivers/firmware/efivars.c +++ b/drivers/firmware/efivars.c @@ -883,7 +883,6 @@ static struct inode *efivarfs_get_inode(struct super_block *sb, if (inode) { inode->i_ino = get_next_ino(); - inode->i_uid = inode->i_gid = 0; inode->i_mode = mode; inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME; switch (mode & S_IFMT) { diff --git a/drivers/md/md.c b/drivers/md/md.c index bd8bf0953fe3..4843b004c558 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -452,7 +452,7 @@ void md_flush_request(struct mddev *mddev, struct bio *bio) spin_lock_irq(&mddev->write_lock); wait_event_lock_irq(mddev->sb_wait, !mddev->flush_bio, - mddev->write_lock, /*nothing*/); + mddev->write_lock); mddev->flush_bio = bio; spin_unlock_irq(&mddev->write_lock); diff --git a/drivers/md/md.h b/drivers/md/md.h index af443ab868db..1e2fc3d9c74c 100644 --- a/drivers/md/md.h +++ b/drivers/md/md.h @@ -551,32 +551,6 @@ struct md_thread { #define THREAD_WAKEUP 0 -#define __wait_event_lock_irq(wq, condition, lock, cmd) \ -do { \ - wait_queue_t __wait; \ - init_waitqueue_entry(&__wait, current); \ - \ - add_wait_queue(&wq, &__wait); \ - for (;;) { \ - set_current_state(TASK_UNINTERRUPTIBLE); \ - if (condition) \ - break; \ - spin_unlock_irq(&lock); \ - cmd; \ - schedule(); \ - spin_lock_irq(&lock); \ - } \ - current->state = TASK_RUNNING; \ - remove_wait_queue(&wq, &__wait); \ -} while (0) - -#define wait_event_lock_irq(wq, condition, lock, cmd) \ -do { \ - if (condition) \ - break; \ - __wait_event_lock_irq(wq, condition, lock, cmd); \ -} while (0) - static inline void safe_put_page(struct page *p) { if (p) put_page(p); diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c index a0f73092176e..d5bddfc4010e 100644 --- a/drivers/md/raid1.c +++ b/drivers/md/raid1.c @@ -822,7 +822,7 @@ static void raise_barrier(struct r1conf *conf) /* Wait until no block IO is waiting */ wait_event_lock_irq(conf->wait_barrier, !conf->nr_waiting, - conf->resync_lock, ); + conf->resync_lock); /* block any new IO from starting */ conf->barrier++; @@ -830,7 +830,7 @@ static void raise_barrier(struct r1conf *conf) /* Now wait for all pending IO to complete */ wait_event_lock_irq(conf->wait_barrier, !conf->nr_pending && conf->barrier < RESYNC_DEPTH, - conf->resync_lock, ); + conf->resync_lock); spin_unlock_irq(&conf->resync_lock); } @@ -864,8 +864,7 @@ static void wait_barrier(struct r1conf *conf) (conf->nr_pending && current->bio_list && !bio_list_empty(current->bio_list)), - conf->resync_lock, - ); + conf->resync_lock); conf->nr_waiting--; } conf->nr_pending++; @@ -898,10 +897,10 @@ static void freeze_array(struct r1conf *conf) spin_lock_irq(&conf->resync_lock); conf->barrier++; conf->nr_waiting++; - wait_event_lock_irq(conf->wait_barrier, - conf->nr_pending == conf->nr_queued+1, - conf->resync_lock, - flush_pending_writes(conf)); + wait_event_lock_irq_cmd(conf->wait_barrier, + conf->nr_pending == conf->nr_queued+1, + conf->resync_lock, + flush_pending_writes(conf)); spin_unlock_irq(&conf->resync_lock); } static void unfreeze_array(struct r1conf *conf) diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c index c9acbd717131..64d48249c03b 100644 --- a/drivers/md/raid10.c +++ b/drivers/md/raid10.c @@ -952,7 +952,7 @@ static void raise_barrier(struct r10conf *conf, int force) /* Wait until no block IO is waiting (unless 'force') */ wait_event_lock_irq(conf->wait_barrier, force || !conf->nr_waiting, - conf->resync_lock, ); + conf->resync_lock); /* block any new IO from starting */ conf->barrier++; @@ -960,7 +960,7 @@ static void raise_barrier(struct r10conf *conf, int force) /* Now wait for all pending IO to complete */ wait_event_lock_irq(conf->wait_barrier, !conf->nr_pending && conf->barrier < RESYNC_DEPTH, - conf->resync_lock, ); + conf->resync_lock); spin_unlock_irq(&conf->resync_lock); } @@ -993,8 +993,7 @@ static void wait_barrier(struct r10conf *conf) (conf->nr_pending && current->bio_list && !bio_list_empty(current->bio_list)), - conf->resync_lock, - ); + conf->resync_lock); conf->nr_waiting--; } conf->nr_pending++; @@ -1027,10 +1026,10 @@ static void freeze_array(struct r10conf *conf) spin_lock_irq(&conf->resync_lock); conf->barrier++; conf->nr_waiting++; - wait_event_lock_irq(conf->wait_barrier, - conf->nr_pending == conf->nr_queued+1, - conf->resync_lock, - flush_pending_writes(conf)); + wait_event_lock_irq_cmd(conf->wait_barrier, + conf->nr_pending == conf->nr_queued+1, + conf->resync_lock, + flush_pending_writes(conf)); spin_unlock_irq(&conf->resync_lock); } diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index 3380372c0393..8d8555bf3e1d 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -466,7 +466,7 @@ get_active_stripe(struct r5conf *conf, sector_t sector, do { wait_event_lock_irq(conf->wait_for_stripe, conf->quiesce == 0 || noquiesce, - conf->device_lock, /* nothing */); + conf->device_lock); sh = __find_stripe(conf, sector, conf->generation - previous); if (!sh) { if (!conf->inactive_blocked) @@ -480,8 +480,7 @@ get_active_stripe(struct r5conf *conf, sector_t sector, (atomic_read(&conf->active_stripes) < (conf->max_nr_stripes *3/4) || !conf->inactive_blocked), - conf->device_lock, - ); + conf->device_lock); conf->inactive_blocked = 0; } else init_stripe(sh, sector, previous); @@ -1646,8 +1645,7 @@ static int resize_stripes(struct r5conf *conf, int newsize) spin_lock_irq(&conf->device_lock); wait_event_lock_irq(conf->wait_for_stripe, !list_empty(&conf->inactive_list), - conf->device_lock, - ); + conf->device_lock); osh = get_free_stripe(conf); spin_unlock_irq(&conf->device_lock); atomic_set(&nsh->count, 1); @@ -4003,7 +4001,7 @@ static int chunk_aligned_read(struct mddev *mddev, struct bio * raid_bio) spin_lock_irq(&conf->device_lock); wait_event_lock_irq(conf->wait_for_stripe, conf->quiesce == 0, - conf->device_lock, /* nothing */); + conf->device_lock); atomic_inc(&conf->active_aligned_reads); spin_unlock_irq(&conf->device_lock); @@ -6095,7 +6093,7 @@ static void raid5_quiesce(struct mddev *mddev, int state) wait_event_lock_irq(conf->wait_for_stripe, atomic_read(&conf->active_stripes) == 0 && atomic_read(&conf->active_aligned_reads) == 0, - conf->device_lock, /* nothing */); + conf->device_lock); conf->quiesce = 1; spin_unlock_irq(&conf->device_lock); /* allow reshape to continue */ diff --git a/drivers/staging/android/binder.c b/drivers/staging/android/binder.c index 4a36e9ab8cf7..2d12e8a1f82e 100644 --- a/drivers/staging/android/binder.c +++ b/drivers/staging/android/binder.c @@ -35,6 +35,7 @@ #include <linux/uaccess.h> #include <linux/vmalloc.h> #include <linux/slab.h> +#include <linux/pid_namespace.h> #include "binder.h" #include "binder_trace.h" @@ -2320,7 +2321,7 @@ retry: if (t->from) { struct task_struct *sender = t->from->proc->tsk; tr.sender_pid = task_tgid_nr_ns(sender, - current->nsproxy->pid_ns); + task_active_pid_ns(current)); } else { tr.sender_pid = 0; } diff --git a/fs/attr.c b/fs/attr.c index cce7df53b694..1449adb14ef6 100644 --- a/fs/attr.c +++ b/fs/attr.c @@ -49,14 +49,15 @@ int inode_change_ok(const struct inode *inode, struct iattr *attr) /* Make sure a caller can chown. */ if ((ia_valid & ATTR_UID) && (!uid_eq(current_fsuid(), inode->i_uid) || - !uid_eq(attr->ia_uid, inode->i_uid)) && !capable(CAP_CHOWN)) + !uid_eq(attr->ia_uid, inode->i_uid)) && + !inode_capable(inode, CAP_CHOWN)) return -EPERM; /* Make sure caller can chgrp. */ if ((ia_valid & ATTR_GID) && (!uid_eq(current_fsuid(), inode->i_uid) || (!in_group_p(attr->ia_gid) && !gid_eq(attr->ia_gid, inode->i_gid))) && - !capable(CAP_CHOWN)) + !inode_capable(inode, CAP_CHOWN)) return -EPERM; /* Make sure a caller can chmod. */ @@ -65,7 +66,8 @@ int inode_change_ok(const struct inode *inode, struct iattr *attr) return -EPERM; /* Also check the setgid bit! */ if (!in_group_p((ia_valid & ATTR_GID) ? attr->ia_gid : - inode->i_gid) && !capable(CAP_FSETID)) + inode->i_gid) && + !inode_capable(inode, CAP_FSETID)) attr->ia_mode &= ~S_ISGID; } @@ -157,7 +159,8 @@ void setattr_copy(struct inode *inode, const struct iattr *attr) if (ia_valid & ATTR_MODE) { umode_t mode = attr->ia_mode; - if (!in_group_p(inode->i_gid) && !capable(CAP_FSETID)) + if (!in_group_p(inode->i_gid) && + !inode_capable(inode, CAP_FSETID)) mode &= ~S_ISGID; inode->i_mode = mode; } diff --git a/fs/autofs4/autofs_i.h b/fs/autofs4/autofs_i.h index 908e18455413..b785e7707959 100644 --- a/fs/autofs4/autofs_i.h +++ b/fs/autofs4/autofs_i.h @@ -74,8 +74,8 @@ struct autofs_info { unsigned long last_used; atomic_t count; - uid_t uid; - gid_t gid; + kuid_t uid; + kgid_t gid; }; #define AUTOFS_INF_EXPIRING (1<<0) /* dentry is in the process of expiring */ @@ -89,8 +89,8 @@ struct autofs_wait_queue { struct qstr name; u32 dev; u64 ino; - uid_t uid; - gid_t gid; + kuid_t uid; + kgid_t gid; pid_t pid; pid_t tgid; /* This is for status reporting upon return */ diff --git a/fs/autofs4/dev-ioctl.c b/fs/autofs4/dev-ioctl.c index a16214109d31..9f68a37bb2b2 100644 --- a/fs/autofs4/dev-ioctl.c +++ b/fs/autofs4/dev-ioctl.c @@ -437,8 +437,8 @@ static int autofs_dev_ioctl_requester(struct file *fp, err = 0; autofs4_expire_wait(path.dentry); spin_lock(&sbi->fs_lock); - param->requester.uid = ino->uid; - param->requester.gid = ino->gid; + param->requester.uid = from_kuid_munged(current_user_ns(), ino->uid); + param->requester.gid = from_kgid_munged(current_user_ns(), ino->gid); spin_unlock(&sbi->fs_lock); } path_put(&path); diff --git a/fs/autofs4/inode.c b/fs/autofs4/inode.c index 8a4fed8ead30..b104726e2d0a 100644 --- a/fs/autofs4/inode.c +++ b/fs/autofs4/inode.c @@ -36,8 +36,8 @@ struct autofs_info *autofs4_new_ino(struct autofs_sb_info *sbi) void autofs4_clean_ino(struct autofs_info *ino) { - ino->uid = 0; - ino->gid = 0; + ino->uid = GLOBAL_ROOT_UID; + ino->gid = GLOBAL_ROOT_GID; ino->last_used = jiffies; } @@ -79,10 +79,12 @@ static int autofs4_show_options(struct seq_file *m, struct dentry *root) return 0; seq_printf(m, ",fd=%d", sbi->pipefd); - if (root_inode->i_uid != 0) - seq_printf(m, ",uid=%u", root_inode->i_uid); - if (root_inode->i_gid != 0) - seq_printf(m, ",gid=%u", root_inode->i_gid); + if (!uid_eq(root_inode->i_uid, GLOBAL_ROOT_UID)) + seq_printf(m, ",uid=%u", + from_kuid_munged(&init_user_ns, root_inode->i_uid)); + if (!gid_eq(root_inode->i_gid, GLOBAL_ROOT_GID)) + seq_printf(m, ",gid=%u", + from_kgid_munged(&init_user_ns, root_inode->i_gid)); seq_printf(m, ",pgrp=%d", sbi->oz_pgrp); seq_printf(m, ",timeout=%lu", sbi->exp_timeout/HZ); seq_printf(m, ",minproto=%d", sbi->min_proto); @@ -126,7 +128,7 @@ static const match_table_t tokens = { {Opt_err, NULL} }; -static int parse_options(char *options, int *pipefd, uid_t *uid, gid_t *gid, +static int parse_options(char *options, int *pipefd, kuid_t *uid, kgid_t *gid, pid_t *pgrp, unsigned int *type, int *minproto, int *maxproto) { char *p; @@ -159,12 +161,16 @@ static int parse_options(char *options, int *pipefd, uid_t *uid, gid_t *gid, case Opt_uid: if (match_int(args, &option)) return 1; - *uid = option; + *uid = make_kuid(current_user_ns(), option); + if (!uid_valid(*uid)) + return 1; break; case Opt_gid: if (match_int(args, &option)) return 1; - *gid = option; + *gid = make_kgid(current_user_ns(), option); + if (!gid_valid(*gid)) + return 1; break; case Opt_pgrp: if (match_int(args, &option)) diff --git a/fs/autofs4/waitq.c b/fs/autofs4/waitq.c index dce436e595c1..03bc1d347d8e 100644 --- a/fs/autofs4/waitq.c +++ b/fs/autofs4/waitq.c @@ -154,6 +154,7 @@ static void autofs4_notify_daemon(struct autofs_sb_info *sbi, case autofs_ptype_expire_direct: { struct autofs_v5_packet *packet = &pkt.v5_pkt.v5_packet; + struct user_namespace *user_ns = sbi->pipe->f_cred->user_ns; pktsz = sizeof(*packet); @@ -163,8 +164,8 @@ static void autofs4_notify_daemon(struct autofs_sb_info *sbi, packet->name[wq->name.len] = '\0'; packet->dev = wq->dev; packet->ino = wq->ino; - packet->uid = wq->uid; - packet->gid = wq->gid; + packet->uid = from_kuid_munged(user_ns, wq->uid); + packet->gid = from_kgid_munged(user_ns, wq->gid); packet->pid = wq->pid; packet->tgid = wq->tgid; break; diff --git a/fs/exec.c b/fs/exec.c index d5eb9e605ffd..d8e1191cb112 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -1266,14 +1266,13 @@ int prepare_binprm(struct linux_binprm *bprm) bprm->cred->egid = current_egid(); if (!(bprm->file->f_path.mnt->mnt_flags & MNT_NOSUID) && - !current->no_new_privs) { + !current->no_new_privs && + kuid_has_mapping(bprm->cred->user_ns, inode->i_uid) && + kgid_has_mapping(bprm->cred->user_ns, inode->i_gid)) { /* Set-uid? */ if (mode & S_ISUID) { - if (!kuid_has_mapping(bprm->cred->user_ns, inode->i_uid)) - return -EPERM; bprm->per_clear |= PER_CLEAR_ON_SETID; bprm->cred->euid = inode->i_uid; - } /* Set-gid? */ @@ -1283,8 +1282,6 @@ int prepare_binprm(struct linux_binprm *bprm) * executable. */ if ((mode & (S_ISGID | S_IXGRP)) == (S_ISGID | S_IXGRP)) { - if (!kgid_has_mapping(bprm->cred->user_ns, inode->i_gid)) - return -EPERM; bprm->per_clear |= PER_CLEAR_ON_SETID; bprm->cred->egid = inode->i_gid; } diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c index 8c23fa7a91e6..c16335315e5d 100644 --- a/fs/fuse/dev.c +++ b/fs/fuse/dev.c @@ -92,8 +92,8 @@ static void __fuse_put_request(struct fuse_req *req) static void fuse_req_init_context(struct fuse_req *req) { - req->in.h.uid = current_fsuid(); - req->in.h.gid = current_fsgid(); + req->in.h.uid = from_kuid_munged(&init_user_ns, current_fsuid()); + req->in.h.gid = from_kgid_munged(&init_user_ns, current_fsgid()); req->in.h.pid = current->pid; } diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 324bc0850534..b7c09f9eb40c 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -818,8 +818,8 @@ static void fuse_fillattr(struct inode *inode, struct fuse_attr *attr, stat->ino = attr->ino; stat->mode = (inode->i_mode & S_IFMT) | (attr->mode & 07777); stat->nlink = attr->nlink; - stat->uid = attr->uid; - stat->gid = attr->gid; + stat->uid = make_kuid(&init_user_ns, attr->uid); + stat->gid = make_kgid(&init_user_ns, attr->gid); stat->rdev = inode->i_rdev; stat->atime.tv_sec = attr->atime; stat->atime.tv_nsec = attr->atimensec; @@ -1007,12 +1007,12 @@ int fuse_allow_task(struct fuse_conn *fc, struct task_struct *task) rcu_read_lock(); ret = 0; cred = __task_cred(task); - if (cred->euid == fc->user_id && - cred->suid == fc->user_id && - cred->uid == fc->user_id && - cred->egid == fc->group_id && - cred->sgid == fc->group_id && - cred->gid == fc->group_id) + if (uid_eq(cred->euid, fc->user_id) && + uid_eq(cred->suid, fc->user_id) && + uid_eq(cred->uid, fc->user_id) && + gid_eq(cred->egid, fc->group_id) && + gid_eq(cred->sgid, fc->group_id) && + gid_eq(cred->gid, fc->group_id)) ret = 1; rcu_read_unlock(); @@ -1306,9 +1306,9 @@ static void iattr_to_fattr(struct iattr *iattr, struct fuse_setattr_in *arg) if (ivalid & ATTR_MODE) arg->valid |= FATTR_MODE, arg->mode = iattr->ia_mode; if (ivalid & ATTR_UID) - arg->valid |= FATTR_UID, arg->uid = iattr->ia_uid; + arg->valid |= FATTR_UID, arg->uid = from_kuid(&init_user_ns, iattr->ia_uid); if (ivalid & ATTR_GID) - arg->valid |= FATTR_GID, arg->gid = iattr->ia_gid; + arg->valid |= FATTR_GID, arg->gid = from_kgid(&init_user_ns, iattr->ia_gid); if (ivalid & ATTR_SIZE) arg->valid |= FATTR_SIZE, arg->size = iattr->ia_size; if (ivalid & ATTR_ATIME) { diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h index e24dd74e3068..e105a53fc72d 100644 --- a/fs/fuse/fuse_i.h +++ b/fs/fuse/fuse_i.h @@ -333,10 +333,10 @@ struct fuse_conn { atomic_t count; /** The user id for this mount */ - uid_t user_id; + kuid_t user_id; /** The group id for this mount */ - gid_t group_id; + kgid_t group_id; /** The fuse mount flags for this mount */ unsigned flags; diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c index f0eda124cffb..73ca6b72beaf 100644 --- a/fs/fuse/inode.c +++ b/fs/fuse/inode.c @@ -60,8 +60,8 @@ MODULE_PARM_DESC(max_user_congthresh, struct fuse_mount_data { int fd; unsigned rootmode; - unsigned user_id; - unsigned group_id; + kuid_t user_id; + kgid_t group_id; unsigned fd_present:1; unsigned rootmode_present:1; unsigned user_id_present:1; @@ -164,8 +164,8 @@ void fuse_change_attributes_common(struct inode *inode, struct fuse_attr *attr, inode->i_ino = fuse_squash_ino(attr->ino); inode->i_mode = (inode->i_mode & S_IFMT) | (attr->mode & 07777); set_nlink(inode, attr->nlink); - inode->i_uid = attr->uid; - inode->i_gid = attr->gid; + inode->i_uid = make_kuid(&init_user_ns, attr->uid); + inode->i_gid = make_kgid(&init_user_ns, attr->gid); inode->i_blocks = attr->blocks; inode->i_atime.tv_sec = attr->atime; inode->i_atime.tv_nsec = attr->atimensec; @@ -492,14 +492,18 @@ static int parse_fuse_opt(char *opt, struct fuse_mount_data *d, int is_bdev) case OPT_USER_ID: if (match_int(&args[0], &value)) return 0; - d->user_id = value; + d->user_id = make_kuid(current_user_ns(), value); + if (!uid_valid(d->user_id)) + return 0; d->user_id_present = 1; break; case OPT_GROUP_ID: if (match_int(&args[0], &value)) return 0; - d->group_id = value; + d->group_id = make_kgid(current_user_ns(), value); + if (!gid_valid(d->group_id)) + return 0; d->group_id_present = 1; break; @@ -540,8 +544,8 @@ static int fuse_show_options(struct seq_file *m, struct dentry *root) struct super_block *sb = root->d_sb; struct fuse_conn *fc = get_fuse_conn_super(sb); - seq_printf(m, ",user_id=%u", fc->user_id); - seq_printf(m, ",group_id=%u", fc->group_id); + seq_printf(m, ",user_id=%u", from_kuid_munged(&init_user_ns, fc->user_id)); + seq_printf(m, ",group_id=%u", from_kgid_munged(&init_user_ns, fc->group_id)); if (fc->flags & FUSE_DEFAULT_PERMISSIONS) seq_puts(m, ",default_permissions"); if (fc->flags & FUSE_ALLOW_OTHER) @@ -989,7 +993,8 @@ static int fuse_fill_super(struct super_block *sb, void *data, int silent) if (!file) goto err; - if (file->f_op != &fuse_dev_operations) + if ((file->f_op != &fuse_dev_operations) || + (file->f_cred->user_ns != &init_user_ns)) goto err_fput; fc = kmalloc(sizeof(*fc), GFP_KERNEL); diff --git a/fs/hppfs/hppfs.c b/fs/hppfs/hppfs.c index 78f21f8dc2ec..43b315f2002b 100644 --- a/fs/hppfs/hppfs.c +++ b/fs/hppfs/hppfs.c @@ -710,7 +710,7 @@ static int hppfs_fill_super(struct super_block *sb, void *d, int silent) struct vfsmount *proc_mnt; int err = -ENOENT; - proc_mnt = mntget(current->nsproxy->pid_ns->proc_mnt); + proc_mnt = mntget(task_active_pid_ns(current)->proc_mnt); if (IS_ERR(proc_mnt)) goto out; diff --git a/fs/mount.h b/fs/mount.h index 4f291f9de641..cd5007980400 100644 --- a/fs/mount.h +++ b/fs/mount.h @@ -4,8 +4,11 @@ struct mnt_namespace { atomic_t count; + unsigned int proc_inum; struct mount * root; struct list_head list; + struct user_namespace *user_ns; + u64 seq; /* Sequence number to prevent loops */ wait_queue_head_t poll; int event; }; diff --git a/fs/namespace.c b/fs/namespace.c index 24960626bb6b..c1bbe86f4920 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -12,6 +12,7 @@ #include <linux/export.h> #include <linux/capability.h> #include <linux/mnt_namespace.h> +#include <linux/user_namespace.h> #include <linux/namei.h> #include <linux/security.h> #include <linux/idr.h> @@ -20,6 +21,7 @@ #include <linux/fs_struct.h> /* get_fs_root et.al. */ #include <linux/fsnotify.h> /* fsnotify_vfsmount_delete */ #include <linux/uaccess.h> +#include <linux/proc_fs.h> #include "pnode.h" #include "internal.h" @@ -784,7 +786,7 @@ static struct mount *clone_mnt(struct mount *old, struct dentry *root, if (!mnt) return ERR_PTR(-ENOMEM); - if (flag & (CL_SLAVE | CL_PRIVATE)) + if (flag & (CL_SLAVE | CL_PRIVATE | CL_SHARED_TO_SLAVE)) mnt->mnt_group_id = 0; /* not a peer of original */ else mnt->mnt_group_id = old->mnt_group_id; @@ -805,7 +807,8 @@ static struct mount *clone_mnt(struct mount *old, struct dentry *root, list_add_tail(&mnt->mnt_instance, &sb->s_mounts); br_write_unlock(&vfsmount_lock); - if (flag & CL_SLAVE) { + if ((flag & CL_SLAVE) || + ((flag & CL_SHARED_TO_SLAVE) && IS_MNT_SHARED(old))) { list_add(&mnt->mnt_slave, &old->mnt_slave_list); mnt->mnt_master = old; CLEAR_MNT_SHARED(mnt); @@ -1266,7 +1269,7 @@ SYSCALL_DEFINE2(umount, char __user *, name, int, flags) goto dput_and_out; retval = -EPERM; - if (!capable(CAP_SYS_ADMIN)) + if (!ns_capable(mnt->mnt_ns->user_ns, CAP_SYS_ADMIN)) goto dput_and_out; retval = do_umount(mnt, flags); @@ -1292,7 +1295,7 @@ SYSCALL_DEFINE1(oldumount, char __user *, name) static int mount_is_safe(struct path *path) { - if (capable(CAP_SYS_ADMIN)) + if (ns_capable(real_mount(path->mnt)->mnt_ns->user_ns, CAP_SYS_ADMIN)) return 0; return -EPERM; #ifdef notyet @@ -1308,6 +1311,26 @@ static int mount_is_safe(struct path *path) #endif } +static bool mnt_ns_loop(struct path *path) +{ + /* Could bind mounting the mount namespace inode cause a + * mount namespace loop? + */ + struct inode *inode = path->dentry->d_inode; + struct proc_inode *ei; + struct mnt_namespace *mnt_ns; + + if (!proc_ns_inode(inode)) + return false; + + ei = PROC_I(inode); + if (ei->ns_ops != &mntns_operations) + return false; + + mnt_ns = ei->ns; + return current->nsproxy->mnt_ns->seq >= mnt_ns->seq; +} + struct mount *copy_tree(struct mount *mnt, struct dentry *dentry, int flag) { @@ -1610,7 +1633,7 @@ static int do_change_type(struct path *path, int flag) int type; int err = 0; - if (!capable(CAP_SYS_ADMIN)) + if (!ns_capable(mnt->mnt_ns->user_ns, CAP_SYS_ADMIN)) return -EPERM; if (path->dentry != path->mnt->mnt_root) @@ -1655,6 +1678,10 @@ static int do_loopback(struct path *path, const char *old_name, if (err) return err; + err = -EINVAL; + if (mnt_ns_loop(&old_path)) + goto out; + err = lock_mount(path); if (err) goto out; @@ -1770,7 +1797,7 @@ static int do_move_mount(struct path *path, const char *old_name) struct mount *p; struct mount *old; int err = 0; - if (!capable(CAP_SYS_ADMIN)) + if (!ns_capable(real_mount(path->mnt)->mnt_ns->user_ns, CAP_SYS_ADMIN)) return -EPERM; if (!old_name || !*old_name) return -EINVAL; @@ -1857,21 +1884,6 @@ static struct vfsmount *fs_set_subtype(struct vfsmount *mnt, const char *fstype) return ERR_PTR(err); } -static struct vfsmount * -do_kern_mount(const char *fstype, int flags, const char *name, void *data) -{ - struct file_system_type *type = get_fs_type(fstype); - struct vfsmount *mnt; - if (!type) - return ERR_PTR(-ENODEV); - mnt = vfs_kern_mount(type, flags, name, data); - if (!IS_ERR(mnt) && (type->fs_flags & FS_HAS_SUBTYPE) && - !mnt->mnt_sb->s_subtype) - mnt = fs_set_subtype(mnt, fstype); - put_filesystem(type); - return mnt; -} - /* * add a mount into a namespace's mount tree */ @@ -1917,20 +1929,46 @@ unlock: * create a new mount for userspace and request it to be added into the * namespace's tree */ -static int do_new_mount(struct path *path, const char *type, int flags, +static int do_new_mount(struct path *path, const char *fstype, int flags, int mnt_flags, const char *name, void *data) { + struct file_system_type *type; + struct user_namespace *user_ns; struct vfsmount *mnt; int err; - if (!type) + if (!fstype) return -EINVAL; /* we need capabilities... */ - if (!capable(CAP_SYS_ADMIN)) + user_ns = real_mount(path->mnt)->mnt_ns->user_ns; + if (!ns_capable(user_ns, CAP_SYS_ADMIN)) return -EPERM; - mnt = do_kern_mount(type, flags, name, data); + type = get_fs_type(fstype); + if (!type) + return -ENODEV; + + if (user_ns != &init_user_ns) { + if (!(type->fs_flags & FS_USERNS_MOUNT)) { + put_filesystem(type); + return -EPERM; + } + /* Only in special cases allow devices from mounts + * created outside the initial user namespace. + */ + if (!(type->fs_flags & FS_USERNS_DEV_MOUNT)) { + flags |= MS_NODEV; + mnt_flags |= MNT_NODEV; + } + } + + mnt = vfs_kern_mount(type, flags, name, data); + if (!IS_ERR(mnt) && (type->fs_flags & FS_HAS_SUBTYPE) && + !mnt->mnt_sb->s_subtype) + mnt = fs_set_subtype(mnt, fstype); + + put_filesystem(type); if (IS_ERR(mnt)) return PTR_ERR(mnt); @@ -2261,18 +2299,42 @@ dput_out: return retval; } -static struct mnt_namespace *alloc_mnt_ns(void) +static void free_mnt_ns(struct mnt_namespace *ns) +{ + proc_free_inum(ns->proc_inum); + put_user_ns(ns->user_ns); + kfree(ns); +} + +/* + * Assign a sequence number so we can detect when we attempt to bind + * mount a reference to an older mount namespace into the current + * mount namespace, preventing reference counting loops. A 64bit + * number incrementing at 10Ghz will take 12,427 years to wrap which + * is effectively never, so we can ignore the possibility. + */ +static atomic64_t mnt_ns_seq = ATOMIC64_INIT(1); + +static struct mnt_namespace *alloc_mnt_ns(struct user_namespace *user_ns) { struct mnt_namespace *new_ns; + int ret; new_ns = kmalloc(sizeof(struct mnt_namespace), GFP_KERNEL); if (!new_ns) return ERR_PTR(-ENOMEM); + ret = proc_alloc_inum(&new_ns->proc_inum); + if (ret) { + kfree(new_ns); + return ERR_PTR(ret); + } + new_ns->seq = atomic64_add_return(1, &mnt_ns_seq); atomic_set(&new_ns->count, 1); new_ns->root = NULL; INIT_LIST_HEAD(&new_ns->list); init_waitqueue_head(&new_ns->poll); new_ns->event = 0; + new_ns->user_ns = get_user_ns(user_ns); return new_ns; } @@ -2281,24 +2343,28 @@ static struct mnt_namespace *alloc_mnt_ns(void) * copied from the namespace of the passed in task structure. */ static struct mnt_namespace *dup_mnt_ns(struct mnt_namespace *mnt_ns, - struct fs_struct *fs) + struct user_namespace *user_ns, struct fs_struct *fs) { struct mnt_namespace *new_ns; struct vfsmount *rootmnt = NULL, *pwdmnt = NULL; struct mount *p, *q; struct mount *old = mnt_ns->root; struct mount *new; + int copy_flags; - new_ns = alloc_mnt_ns(); + new_ns = alloc_mnt_ns(user_ns); if (IS_ERR(new_ns)) return new_ns; down_write(&namespace_sem); /* First pass: copy the tree topology */ - new = copy_tree(old, old->mnt.mnt_root, CL_COPY_ALL | CL_EXPIRE); + copy_flags = CL_COPY_ALL | CL_EXPIRE; + if (user_ns != mnt_ns->user_ns) + copy_flags |= CL_SHARED_TO_SLAVE; + new = copy_tree(old, old->mnt.mnt_root, copy_flags); if (IS_ERR(new)) { up_write(&namespace_sem); - kfree(new_ns); + free_mnt_ns(new_ns); return ERR_CAST(new); } new_ns->root = new; @@ -2339,7 +2405,7 @@ static struct mnt_namespace *dup_mnt_ns(struct mnt_namespace *mnt_ns, } struct mnt_namespace *copy_mnt_ns(unsigned long flags, struct mnt_namespace *ns, - struct fs_struct *new_fs) + struct user_namespace *user_ns, struct fs_struct *new_fs) { struct mnt_namespace *new_ns; @@ -2349,7 +2415,7 @@ struct mnt_namespace *copy_mnt_ns(unsigned long flags, struct mnt_namespace *ns, if (!(flags & CLONE_NEWNS)) return ns; - new_ns = dup_mnt_ns(ns, new_fs); + new_ns = dup_mnt_ns(ns, user_ns, new_fs); put_mnt_ns(ns); return new_ns; @@ -2361,7 +2427,7 @@ struct mnt_namespace *copy_mnt_ns(unsigned long flags, struct mnt_namespace *ns, */ static struct mnt_namespace *create_mnt_ns(struct vfsmount *m) { - struct mnt_namespace *new_ns = alloc_mnt_ns(); + struct mnt_namespace *new_ns = alloc_mnt_ns(&init_user_ns); if (!IS_ERR(new_ns)) { struct mount *mnt = real_mount(m); mnt->mnt_ns = new_ns; @@ -2501,7 +2567,7 @@ SYSCALL_DEFINE2(pivot_root, const char __user *, new_root, struct mount *new_mnt, *root_mnt; int error; - if (!capable(CAP_SYS_ADMIN)) + if (!ns_capable(current->nsproxy->mnt_ns->user_ns, CAP_SYS_ADMIN)) return -EPERM; error = user_path_dir(new_root, &new); @@ -2583,8 +2649,13 @@ static void __init init_mount_tree(void) struct vfsmount *mnt; struct mnt_namespace *ns; struct path root; + struct file_system_type *type; - mnt = do_kern_mount("rootfs", 0, "rootfs", NULL); + type = get_fs_type("rootfs"); + if (!type) + panic("Can't find rootfs type"); + mnt = vfs_kern_mount(type, 0, "rootfs", NULL); + put_filesystem(type); if (IS_ERR(mnt)) panic("Can't create rootfs"); @@ -2647,7 +2718,7 @@ void put_mnt_ns(struct mnt_namespace *ns) br_write_unlock(&vfsmount_lock); up_write(&namespace_sem); release_mounts(&umount_list); - kfree(ns); + free_mnt_ns(ns); } struct vfsmount *kern_mount_data(struct file_system_type *type, void *data) @@ -2681,3 +2752,71 @@ bool our_mnt(struct vfsmount *mnt) { return check_mnt(real_mount(mnt)); } + +static void *mntns_get(struct task_struct *task) +{ + struct mnt_namespace *ns = NULL; + struct nsproxy *nsproxy; + + rcu_read_lock(); + nsproxy = task_nsproxy(task); + if (nsproxy) { + ns = nsproxy->mnt_ns; + get_mnt_ns(ns); + } + rcu_read_unlock(); + + return ns; +} + +static void mntns_put(void *ns) +{ + put_mnt_ns(ns); +} + +static int mntns_install(struct nsproxy *nsproxy, void *ns) +{ + struct fs_struct *fs = current->fs; + struct mnt_namespace *mnt_ns = ns; + struct path root; + + if (!ns_capable(mnt_ns->user_ns, CAP_SYS_ADMIN) || + !nsown_capable(CAP_SYS_CHROOT)) + return -EPERM; + + if (fs->users != 1) + return -EINVAL; + + get_mnt_ns(mnt_ns); + put_mnt_ns(nsproxy->mnt_ns); + nsproxy->mnt_ns = mnt_ns; + + /* Find the root */ + root.mnt = &mnt_ns->root->mnt; + root.dentry = mnt_ns->root->mnt.mnt_root; + path_get(&root); + while(d_mountpoint(root.dentry) && follow_down_one(&root)) + ; + + /* Update the pwd and root */ + set_fs_pwd(fs, &root); + set_fs_root(fs, &root); + + path_put(&root); + return 0; +} + +static unsigned int mntns_inum(void *ns) +{ + struct mnt_namespace *mnt_ns = ns; + return mnt_ns->proc_inum; +} + +const struct proc_ns_operations mntns_operations = { + .name = "mnt", + .type = CLONE_NEWNS, + .get = mntns_get, + .put = mntns_put, + .install = mntns_install, + .inum = mntns_inum, +}; diff --git a/fs/open.c b/fs/open.c index 59071f55bf7f..182d8667b7bd 100644 --- a/fs/open.c +++ b/fs/open.c @@ -435,7 +435,7 @@ SYSCALL_DEFINE1(chroot, const char __user *, filename) goto dput_and_out; error = -EPERM; - if (!capable(CAP_SYS_CHROOT)) + if (!nsown_capable(CAP_SYS_CHROOT)) goto dput_and_out; error = security_path_chroot(&path); if (error) diff --git a/fs/pnode.h b/fs/pnode.h index 65c60979d541..19b853a3445c 100644 --- a/fs/pnode.h +++ b/fs/pnode.h @@ -22,6 +22,7 @@ #define CL_COPY_ALL 0x04 #define CL_MAKE_SHARED 0x08 #define CL_PRIVATE 0x10 +#define CL_SHARED_TO_SLAVE 0x20 static inline void set_mnt_shared(struct mount *mnt) { diff --git a/fs/proc/Makefile b/fs/proc/Makefile index 99349efbbc2b..981b05601931 100644 --- a/fs/proc/Makefile +++ b/fs/proc/Makefile @@ -21,6 +21,7 @@ proc-y += uptime.o proc-y += version.o proc-y += softirqs.o proc-y += namespaces.o +proc-y += self.o proc-$(CONFIG_PROC_SYSCTL) += proc_sysctl.o proc-$(CONFIG_NET) += proc_net.o proc-$(CONFIG_PROC_KCORE) += kcore.o diff --git a/fs/proc/array.c b/fs/proc/array.c index 060a56a91278..6a91e6ffbcbd 100644 --- a/fs/proc/array.c +++ b/fs/proc/array.c @@ -162,7 +162,7 @@ static inline const char *get_task_state(struct task_struct *tsk) static inline void task_state(struct seq_file *m, struct pid_namespace *ns, struct pid *pid, struct task_struct *p) { - struct user_namespace *user_ns = current_user_ns(); + struct user_namespace *user_ns = seq_user_ns(m); struct group_info *group_info; int g; struct fdtable *fdt = NULL; diff --git a/fs/proc/base.c b/fs/proc/base.c index aa63d25157b8..5a5a0be40e40 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -2345,146 +2345,6 @@ static const struct file_operations proc_coredump_filter_operations = { }; #endif -/* - * /proc/self: - */ -static int proc_self_readlink(struct dentry *dentry, char __user *buffer, - int buflen) -{ - struct pid_namespace *ns = dentry->d_sb->s_fs_info; - pid_t tgid = task_tgid_nr_ns(current, ns); - char tmp[PROC_NUMBUF]; - if (!tgid) - return -ENOENT; - sprintf(tmp, "%d", tgid); - return vfs_readlink(dentry,buffer,buflen,tmp); -} - -static void *proc_self_follow_link(struct dentry *dentry, struct nameidata *nd) -{ - struct pid_namespace *ns = dentry->d_sb->s_fs_info; - pid_t tgid = task_tgid_nr_ns(current, ns); - char *name = ERR_PTR(-ENOENT); - if (tgid) { - /* 11 for max length of signed int in decimal + NULL term */ - name = kmalloc(12, GFP_KERNEL); - if (!name) - name = ERR_PTR(-ENOMEM); - else - sprintf(name, "%d", tgid); - } - nd_set_link(nd, name); - return NULL; -} - -static void proc_self_put_link(struct dentry *dentry, struct nameidata *nd, - void *cookie) -{ - char *s = nd_get_link(nd); - if (!IS_ERR(s)) - kfree(s); -} - -static const struct inode_operations proc_self_inode_operations = { - .readlink = proc_self_readlink, - .follow_link = proc_self_follow_link, - .put_link = proc_self_put_link, -}; - -/* - * proc base - * - * These are the directory entries in the root directory of /proc - * that properly belong to the /proc filesystem, as they describe - * describe something that is process related. - */ -static const struct pid_entry proc_base_stuff[] = { - NOD("self", S_IFLNK|S_IRWXUGO, - &proc_self_inode_operations, NULL, {}), -}; - -static struct dentry *proc_base_instantiate(struct inode *dir, - struct dentry *dentry, struct task_struct *task, const void *ptr) -{ - const struct pid_entry *p = ptr; - struct inode *inode; - struct proc_inode *ei; - struct dentry *error; - - /* Allocate the inode */ - error = ERR_PTR(-ENOMEM); - inode = new_inode(dir->i_sb); - if (!inode) - goto out; - - /* Initialize the inode */ - ei = PROC_I(inode); - inode->i_ino = get_next_ino(); - inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME; - - /* - * grab the reference to the task. - */ - ei->pid = get_task_pid(task, PIDTYPE_PID); - if (!ei->pid) - goto out_iput; - - inode->i_mode = p->mode; - if (S_ISDIR(inode->i_mode)) - set_nlink(inode, 2); - if (S_ISLNK(inode->i_mode)) - inode->i_size = 64; - if (p->iop) - inode->i_op = p->iop; - if (p->fop) - inode->i_fop = p->fop; - ei->op = p->op; - d_add(dentry, inode); - error = NULL; -out: - return error; -out_iput: - iput(inode); - goto out; -} - -static struct dentry *proc_base_lookup(struct inode *dir, struct dentry *dentry) -{ - struct dentry *error; - struct task_struct *task = get_proc_task(dir); - const struct pid_entry *p, *last; - - error = ERR_PTR(-ENOENT); - - if (!task) - goto out_no_task; - - /* Lookup the directory entry */ - last = &proc_base_stuff[ARRAY_SIZE(proc_base_stuff) - 1]; - for (p = proc_base_stuff; p <= last; p++) { - if (p->len != dentry->d_name.len) - continue; - if (!memcmp(dentry->d_name.name, p->name, p->len)) - break; - } - if (p > last) - goto out; - - error = proc_base_instantiate(dir, dentry, task, p); - -out: - put_task_struct(task); -out_no_task: - return error; -} - -static int proc_base_fill_cache(struct file *filp, void *dirent, - filldir_t filldir, struct task_struct *task, const struct pid_entry *p) -{ - return proc_fill_cache(filp, dirent, filldir, p->name, p->len, - proc_base_instantiate, task, p); -} - #ifdef CONFIG_TASK_IO_ACCOUNTING static int do_io_accounting(struct task_struct *task, char *buffer, int whole) { @@ -2839,10 +2699,6 @@ void proc_flush_task(struct task_struct *task) proc_flush_task_mnt(upid->ns->proc_mnt, upid->nr, tgid->numbers[i].nr); } - - upid = &pid->numbers[pid->level]; - if (upid->nr == 1) - pid_ns_release_proc(upid->ns); } static struct dentry *proc_pid_instantiate(struct inode *dir, @@ -2876,15 +2732,11 @@ out: struct dentry *proc_pid_lookup(struct inode *dir, struct dentry * dentry, unsigned int flags) { - struct dentry *result; + struct dentry *result = NULL; struct task_struct *task; unsigned tgid; struct pid_namespace *ns; - result = proc_base_lookup(dir, dentry); - if (!IS_ERR(result) || PTR_ERR(result) != -ENOENT) - goto out; - tgid = name_to_int(dentry); if (tgid == ~0U) goto out; @@ -2947,7 +2799,7 @@ retry: return iter; } -#define TGID_OFFSET (FIRST_PROCESS_ENTRY + ARRAY_SIZE(proc_base_stuff)) +#define TGID_OFFSET (FIRST_PROCESS_ENTRY) static int proc_pid_fill_cache(struct file *filp, void *dirent, filldir_t filldir, struct tgid_iter iter) @@ -2967,25 +2819,12 @@ static int fake_filldir(void *buf, const char *name, int namelen, /* for the /proc/ directory itself, after non-process stuff has been done */ int proc_pid_readdir(struct file * filp, void * dirent, filldir_t filldir) { - unsigned int nr; - struct task_struct *reaper; struct tgid_iter iter; struct pid_namespace *ns; filldir_t __filldir; if (filp->f_pos >= PID_MAX_LIMIT + TGID_OFFSET) - goto out_no_task; - nr = filp->f_pos - FIRST_PROCESS_ENTRY; - - reaper = get_proc_task(filp->f_path.dentry->d_inode); - if (!reaper) - goto out_no_task; - - for (; nr < ARRAY_SIZE(proc_base_stuff); filp->f_pos++, nr++) { - const struct pid_entry *p = &proc_base_stuff[nr]; - if (proc_base_fill_cache(filp, dirent, filldir, reaper, p) < 0) - goto out; - } + goto out; ns = filp->f_dentry->d_sb->s_fs_info; iter.task = NULL; @@ -3006,8 +2845,6 @@ int proc_pid_readdir(struct file * filp, void * dirent, filldir_t filldir) } filp->f_pos = PID_MAX_LIMIT + TGID_OFFSET; out: - put_task_struct(reaper); -out_no_task: return 0; } diff --git a/fs/proc/generic.c b/fs/proc/generic.c index 0d80cef4cfb9..7b3ae3cc0ef9 100644 --- a/fs/proc/generic.c +++ b/fs/proc/generic.c @@ -350,14 +350,14 @@ static DEFINE_SPINLOCK(proc_inum_lock); /* protects the above */ * Return an inode number between PROC_DYNAMIC_FIRST and * 0xffffffff, or zero on failure. */ -static unsigned int get_inode_number(void) +int proc_alloc_inum(unsigned int *inum) { unsigned int i; int error; retry: - if (ida_pre_get(&proc_inum_ida, GFP_KERNEL) == 0) - return 0; + if (!ida_pre_get(&proc_inum_ida, GFP_KERNEL)) + return -ENOMEM; spin_lock(&proc_inum_lock); error = ida_get_new(&proc_inum_ida, &i); @@ -365,18 +365,19 @@ retry: if (error == -EAGAIN) goto retry; else if (error) - return 0; + return error; if (i > UINT_MAX - PROC_DYNAMIC_FIRST) { spin_lock(&proc_inum_lock); ida_remove(&proc_inum_ida, i); spin_unlock(&proc_inum_lock); - return 0; + return -ENOSPC; } - return PROC_DYNAMIC_FIRST + i; + *inum = PROC_DYNAMIC_FIRST + i; + return 0; } -static void release_inode_number(unsigned int inum) +void proc_free_inum(unsigned int inum) { spin_lock(&proc_inum_lock); ida_remove(&proc_inum_ida, inum - PROC_DYNAMIC_FIRST); @@ -554,13 +555,12 @@ static const struct inode_operations proc_dir_inode_operations = { static int proc_register(struct proc_dir_entry * dir, struct proc_dir_entry * dp) { - unsigned int i; struct proc_dir_entry *tmp; + int ret; - i = get_inode_number(); - if (i == 0) - return -EAGAIN; - dp->low_ino = i; + ret = proc_alloc_inum(&dp->low_ino); + if (ret) + return ret; if (S_ISDIR(dp->mode)) { if (dp->proc_iops == NULL) { @@ -764,7 +764,7 @@ EXPORT_SYMBOL(proc_create_data); static void free_proc_entry(struct proc_dir_entry *de) { - release_inode_number(de->low_ino); + proc_free_inum(de->low_ino); if (S_ISLNK(de->mode)) kfree(de->data); diff --git a/fs/proc/inode.c b/fs/proc/inode.c index 3b22bbdee9ec..439ae6886507 100644 --- a/fs/proc/inode.c +++ b/fs/proc/inode.c @@ -31,6 +31,7 @@ static void proc_evict_inode(struct inode *inode) struct proc_dir_entry *de; struct ctl_table_header *head; const struct proc_ns_operations *ns_ops; + void *ns; truncate_inode_pages(&inode->i_data, 0); clear_inode(inode); @@ -49,8 +50,9 @@ static void proc_evict_inode(struct inode *inode) } /* Release any associated namespace */ ns_ops = PROC_I(inode)->ns_ops; - if (ns_ops && ns_ops->put) - ns_ops->put(PROC_I(inode)->ns); + ns = PROC_I(inode)->ns; + if (ns_ops && ns) + ns_ops->put(ns); } static struct kmem_cache * proc_inode_cachep; diff --git a/fs/proc/internal.h b/fs/proc/internal.h index 43973b084abf..252544c05207 100644 --- a/fs/proc/internal.h +++ b/fs/proc/internal.h @@ -15,6 +15,7 @@ struct ctl_table_header; struct mempolicy; extern struct proc_dir_entry proc_root; +extern void proc_self_init(void); #ifdef CONFIG_PROC_SYSCTL extern int proc_sys_init(void); extern void sysctl_head_put(struct ctl_table_header *head); diff --git a/fs/proc/namespaces.c b/fs/proc/namespaces.c index b178ed733c36..b7a47196c8c3 100644 --- a/fs/proc/namespaces.c +++ b/fs/proc/namespaces.c @@ -11,6 +11,7 @@ #include <net/net_namespace.h> #include <linux/ipc_namespace.h> #include <linux/pid_namespace.h> +#include <linux/user_namespace.h> #include "internal.h" @@ -24,12 +25,168 @@ static const struct proc_ns_operations *ns_entries[] = { #ifdef CONFIG_IPC_NS &ipcns_operations, #endif +#ifdef CONFIG_PID_NS + &pidns_operations, +#endif +#ifdef CONFIG_USER_NS + &userns_operations, +#endif + &mntns_operations, }; static const struct file_operations ns_file_operations = { .llseek = no_llseek, }; +static const struct inode_operations ns_inode_operations = { + .setattr = proc_setattr, +}; + +static int ns_delete_dentry(const struct dentry *dentry) +{ + /* Don't cache namespace inodes when not in use */ + return 1; +} + +static char *ns_dname(struct dentry *dentry, char *buffer, int buflen) +{ + struct inode *inode = dentry->d_inode; + const struct proc_ns_operations *ns_ops = PROC_I(inode)->ns_ops; + + return dynamic_dname(dentry, buffer, buflen, "%s:[%lu]", + ns_ops->name, inode->i_ino); +} + +const struct dentry_operations ns_dentry_operations = +{ + .d_delete = ns_delete_dentry, + .d_dname = ns_dname, +}; + +static struct dentry *proc_ns_get_dentry(struct super_block *sb, + struct task_struct *task, const struct proc_ns_operations *ns_ops) +{ + struct dentry *dentry, *result; + struct inode *inode; + struct proc_inode *ei; + struct qstr qname = { .name = "", }; + void *ns; + + ns = ns_ops->get(task); + if (!ns) + return ERR_PTR(-ENOENT); + + dentry = d_alloc_pseudo(sb, &qname); + if (!dentry) { + ns_ops->put(ns); + return ERR_PTR(-ENOMEM); + } + + inode = iget_locked(sb, ns_ops->inum(ns)); + if (!inode) { + dput(dentry); + ns_ops->put(ns); + return ERR_PTR(-ENOMEM); + } + + ei = PROC_I(inode); + if (inode->i_state & I_NEW) { + inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME; + inode->i_op = &ns_inode_operations; + inode->i_mode = S_IFREG | S_IRUGO; + inode->i_fop = &ns_file_operations; + ei->ns_ops = ns_ops; + ei->ns = ns; + unlock_new_inode(inode); + } else { + ns_ops->put(ns); + } + + d_set_d_op(dentry, &ns_dentry_operations); + result = d_instantiate_unique(dentry, inode); + if (result) { + dput(dentry); + dentry = result; + } + + return dentry; +} + +static void *proc_ns_follow_link(struct dentry *dentry, struct nameidata *nd) +{ + struct inode *inode = dentry->d_inode; + struct super_block *sb = inode->i_sb; + struct proc_inode *ei = PROC_I(inode); + struct task_struct *task; + struct dentry *ns_dentry; + void *error = ERR_PTR(-EACCES); + + task = get_proc_task(inode); + if (!task) + goto out; + + if (!ptrace_may_access(task, PTRACE_MODE_READ)) + goto out_put_task; + + ns_dentry = proc_ns_get_dentry(sb, task, ei->ns_ops); + if (IS_ERR(ns_dentry)) { + error = ERR_CAST(ns_dentry); + goto out_put_task; + } + + dput(nd->path.dentry); + nd->path.dentry = ns_dentry; + error = NULL; + +out_put_task: + put_task_struct(task); +out: + return error; +} + +static int proc_ns_readlink(struct dentry *dentry, char __user *buffer, int buflen) +{ + struct inode *inode = dentry->d_inode; + struct proc_inode *ei = PROC_I(inode); + const struct proc_ns_operations *ns_ops = ei->ns_ops; + struct task_struct *task; + void *ns; + char name[50]; + int len = -EACCES; + + task = get_proc_task(inode); + if (!task) + goto out; + + if (!ptrace_may_access(task, PTRACE_MODE_READ)) + goto out_put_task; + + len = -ENOENT; + ns = ns_ops->get(task); + if (!ns) + goto out_put_task; + + snprintf(name, sizeof(name), "%s:[%u]", ns_ops->name, ns_ops->inum(ns)); + len = strlen(name); + + if (len > buflen) + len = buflen; + if (copy_to_user(buffer, name, len)) + len = -EFAULT; + + ns_ops->put(ns); +out_put_task: + put_task_struct(task); +out: + return len; +} + +static const struct inode_operations proc_ns_link_inode_operations = { + .readlink = proc_ns_readlink, + .follow_link = proc_ns_follow_link, + .setattr = proc_setattr, +}; + static struct dentry *proc_ns_instantiate(struct inode *dir, struct dentry *dentry, struct task_struct *task, const void *ptr) { @@ -37,21 +194,15 @@ static struct dentry *proc_ns_instantiate(struct inode *dir, struct inode *inode; struct proc_inode *ei; struct dentry *error = ERR_PTR(-ENOENT); - void *ns; inode = proc_pid_make_inode(dir->i_sb, task); if (!inode) goto out; - ns = ns_ops->get(task); - if (!ns) - goto out_iput; - ei = PROC_I(inode); - inode->i_mode = S_IFREG|S_IRUSR; - inode->i_fop = &ns_file_operations; - ei->ns_ops = ns_ops; - ei->ns = ns; + inode->i_mode = S_IFLNK|S_IRWXUGO; + inode->i_op = &proc_ns_link_inode_operations; + ei->ns_ops = ns_ops; d_set_d_op(dentry, &pid_dentry_operations); d_add(dentry, inode); @@ -60,9 +211,6 @@ static struct dentry *proc_ns_instantiate(struct inode *dir, error = NULL; out: return error; -out_iput: - iput(inode); - goto out; } static int proc_ns_fill_cache(struct file *filp, void *dirent, @@ -89,10 +237,6 @@ static int proc_ns_dir_readdir(struct file *filp, void *dirent, if (!task) goto out_no_task; - ret = -EPERM; - if (!ptrace_may_access(task, PTRACE_MODE_READ)) - goto out; - ret = 0; i = filp->f_pos; switch (i) { @@ -152,10 +296,6 @@ static struct dentry *proc_ns_dir_lookup(struct inode *dir, if (!task) goto out_no_task; - error = ERR_PTR(-EPERM); - if (!ptrace_may_access(task, PTRACE_MODE_READ)) - goto out; - last = &ns_entries[ARRAY_SIZE(ns_entries)]; for (entry = ns_entries; entry < last; entry++) { if (strlen((*entry)->name) != len) @@ -163,7 +303,6 @@ static struct dentry *proc_ns_dir_lookup(struct inode *dir, if (!memcmp(dentry->d_name.name, (*entry)->name, len)) break; } - error = ERR_PTR(-ENOENT); if (entry == last) goto out; @@ -198,3 +337,7 @@ out_invalid: return ERR_PTR(-EINVAL); } +bool proc_ns_inode(struct inode *inode) +{ + return inode->i_fop == &ns_file_operations; +} diff --git a/fs/proc/root.c b/fs/proc/root.c index 9889a92d2e01..c6e9fac26bac 100644 --- a/fs/proc/root.c +++ b/fs/proc/root.c @@ -100,14 +100,13 @@ static struct dentry *proc_mount(struct file_system_type *fs_type, int err; struct super_block *sb; struct pid_namespace *ns; - struct proc_inode *ei; char *options; if (flags & MS_KERNMOUNT) { ns = (struct pid_namespace *)data; options = NULL; } else { - ns = current->nsproxy->pid_ns; + ns = task_active_pid_ns(current); options = data; } @@ -130,13 +129,6 @@ static struct dentry *proc_mount(struct file_system_type *fs_type, sb->s_flags |= MS_ACTIVE; } - ei = PROC_I(sb->s_root->d_inode); - if (!ei->pid) { - rcu_read_lock(); - ei->pid = get_pid(find_pid_ns(1, ns)); - rcu_read_unlock(); - } - return dget(sb->s_root); } @@ -153,6 +145,7 @@ static struct file_system_type proc_fs_type = { .name = "proc", .mount = proc_mount, .kill_sb = proc_kill_sb, + .fs_flags = FS_USERNS_MOUNT, }; void __init proc_root_init(void) @@ -163,12 +156,8 @@ void __init proc_root_init(void) err = register_filesystem(&proc_fs_type); if (err) return; - err = pid_ns_prepare_proc(&init_pid_ns); - if (err) { - unregister_filesystem(&proc_fs_type); - return; - } + proc_self_init(); proc_symlink("mounts", NULL, "self/mounts"); proc_net_init(); diff --git a/fs/proc/self.c b/fs/proc/self.c new file mode 100644 index 000000000000..aa5cc3bff140 --- /dev/null +++ b/fs/proc/self.c @@ -0,0 +1,59 @@ +#include <linux/proc_fs.h> +#include <linux/sched.h> +#include <linux/namei.h> + +/* + * /proc/self: + */ +static int proc_self_readlink(struct dentry *dentry, char __user *buffer, + int buflen) +{ + struct pid_namespace *ns = dentry->d_sb->s_fs_info; + pid_t tgid = task_tgid_nr_ns(current, ns); + char tmp[PROC_NUMBUF]; + if (!tgid) + return -ENOENT; + sprintf(tmp, "%d", tgid); + return vfs_readlink(dentry,buffer,buflen,tmp); +} + +static void *proc_self_follow_link(struct dentry *dentry, struct nameidata *nd) +{ + struct pid_namespace *ns = dentry->d_sb->s_fs_info; + pid_t tgid = task_tgid_nr_ns(current, ns); + char *name = ERR_PTR(-ENOENT); + if (tgid) { + /* 11 for max length of signed int in decimal + NULL term */ + name = kmalloc(12, GFP_KERNEL); + if (!name) + name = ERR_PTR(-ENOMEM); + else + sprintf(name, "%d", tgid); + } + nd_set_link(nd, name); + return NULL; +} + +static void proc_self_put_link(struct dentry *dentry, struct nameidata *nd, + void *cookie) +{ + char *s = nd_get_link(nd); + if (!IS_ERR(s)) + kfree(s); +} + +static const struct inode_operations proc_self_inode_operations = { + .readlink = proc_self_readlink, + .follow_link = proc_self_follow_link, + .put_link = proc_self_put_link, +}; + +void __init proc_self_init(void) +{ + struct proc_dir_entry *proc_self_symlink; + mode_t mode; + + mode = S_IFLNK | S_IRWXUGO; + proc_self_symlink = proc_create("self", mode, NULL, NULL ); + proc_self_symlink->proc_iops = &proc_self_inode_operations; +} diff --git a/fs/sysfs/mount.c b/fs/sysfs/mount.c index 71eb7e253927..db940a9be045 100644 --- a/fs/sysfs/mount.c +++ b/fs/sysfs/mount.c @@ -149,6 +149,7 @@ static struct file_system_type sysfs_fs_type = { .name = "sysfs", .mount = sysfs_mount, .kill_sb = sysfs_kill_sb, + .fs_flags = FS_USERNS_MOUNT, }; int __init sysfs_init(void) diff --git a/include/linux/cred.h b/include/linux/cred.h index 0142aacb70b7..abb2cd50f6b2 100644 --- a/include/linux/cred.h +++ b/include/linux/cred.h @@ -344,10 +344,8 @@ static inline void put_cred(const struct cred *_cred) extern struct user_namespace init_user_ns; #ifdef CONFIG_USER_NS #define current_user_ns() (current_cred_xxx(user_ns)) -#define task_user_ns(task) (task_cred_xxx((task), user_ns)) #else #define current_user_ns() (&init_user_ns) -#define task_user_ns(task) (&init_user_ns) #endif diff --git a/include/linux/drbd.h b/include/linux/drbd.h index 47e3d4850584..0c5a18ec322c 100644 --- a/include/linux/drbd.h +++ b/include/linux/drbd.h @@ -51,12 +51,11 @@ #endif - extern const char *drbd_buildtag(void); -#define REL_VERSION "8.3.13" -#define API_VERSION 88 +#define REL_VERSION "8.4.2" +#define API_VERSION 1 #define PRO_VERSION_MIN 86 -#define PRO_VERSION_MAX 96 +#define PRO_VERSION_MAX 101 enum drbd_io_error_p { @@ -66,7 +65,8 @@ enum drbd_io_error_p { }; enum drbd_fencing_p { - FP_DONT_CARE, + FP_NOT_AVAIL = -1, /* Not a policy */ + FP_DONT_CARE = 0, FP_RESOURCE, FP_STONITH }; @@ -102,6 +102,20 @@ enum drbd_on_congestion { OC_DISCONNECT, }; +enum drbd_read_balancing { + RB_PREFER_LOCAL, + RB_PREFER_REMOTE, + RB_ROUND_ROBIN, + RB_LEAST_PENDING, + RB_CONGESTED_REMOTE, + RB_32K_STRIPING, + RB_64K_STRIPING, + RB_128K_STRIPING, + RB_256K_STRIPING, + RB_512K_STRIPING, + RB_1M_STRIPING, +}; + /* KEEP the order, do not delete or insert. Only append. */ enum drbd_ret_code { ERR_CODE_BASE = 100, @@ -122,7 +136,7 @@ enum drbd_ret_code { ERR_AUTH_ALG = 120, ERR_AUTH_ALG_ND = 121, ERR_NOMEM = 122, - ERR_DISCARD = 123, + ERR_DISCARD_IMPOSSIBLE = 123, ERR_DISK_CONFIGURED = 124, ERR_NET_CONFIGURED = 125, ERR_MANDATORY_TAG = 126, @@ -130,8 +144,8 @@ enum drbd_ret_code { ERR_INTR = 129, /* EINTR */ ERR_RESIZE_RESYNC = 130, ERR_NO_PRIMARY = 131, - ERR_SYNC_AFTER = 132, - ERR_SYNC_AFTER_CYCLE = 133, + ERR_RESYNC_AFTER = 132, + ERR_RESYNC_AFTER_CYCLE = 133, ERR_PAUSE_IS_SET = 134, ERR_PAUSE_IS_CLEAR = 135, ERR_PACKET_NR = 137, @@ -155,6 +169,14 @@ enum drbd_ret_code { ERR_CONG_NOT_PROTO_A = 155, ERR_PIC_AFTER_DEP = 156, ERR_PIC_PEER_DEP = 157, + ERR_RES_NOT_KNOWN = 158, + ERR_RES_IN_USE = 159, + ERR_MINOR_CONFIGURED = 160, + ERR_MINOR_EXISTS = 161, + ERR_INVALID_REQUEST = 162, + ERR_NEED_APV_100 = 163, + ERR_NEED_ALLOW_TWO_PRI = 164, + ERR_MD_UNCLEAN = 165, /* insert new ones above this line */ AFTER_LAST_ERR_CODE @@ -296,7 +318,8 @@ enum drbd_state_rv { SS_NOT_SUPPORTED = -17, /* drbd-8.2 only */ SS_IN_TRANSIENT_STATE = -18, /* Retry after the next state change */ SS_CONCURRENT_ST_CHG = -19, /* Concurrent cluster side state change! */ - SS_AFTER_LAST_ERROR = -20, /* Keep this at bottom */ + SS_O_VOL_PEER_PRI = -20, + SS_AFTER_LAST_ERROR = -21, /* Keep this at bottom */ }; /* from drbd_strings.c */ @@ -313,7 +336,9 @@ extern const char *drbd_set_st_err_str(enum drbd_state_rv); #define MDF_FULL_SYNC (1 << 3) #define MDF_WAS_UP_TO_DATE (1 << 4) #define MDF_PEER_OUT_DATED (1 << 5) -#define MDF_CRASHED_PRIMARY (1 << 6) +#define MDF_CRASHED_PRIMARY (1 << 6) +#define MDF_AL_CLEAN (1 << 7) +#define MDF_AL_DISABLED (1 << 8) enum drbd_uuid_index { UI_CURRENT, @@ -333,37 +358,23 @@ enum drbd_timeout_flag { #define UUID_JUST_CREATED ((__u64)4) +/* magic numbers used in meta data and network packets */ #define DRBD_MAGIC 0x83740267 -#define BE_DRBD_MAGIC __constant_cpu_to_be32(DRBD_MAGIC) #define DRBD_MAGIC_BIG 0x835a -#define BE_DRBD_MAGIC_BIG __constant_cpu_to_be16(DRBD_MAGIC_BIG) +#define DRBD_MAGIC_100 0x8620ec20 + +#define DRBD_MD_MAGIC_07 (DRBD_MAGIC+3) +#define DRBD_MD_MAGIC_08 (DRBD_MAGIC+4) +#define DRBD_MD_MAGIC_84_UNCLEAN (DRBD_MAGIC+5) + + +/* how I came up with this magic? + * base64 decode "actlog==" ;) */ +#define DRBD_AL_MAGIC 0x69cb65a2 /* these are of type "int" */ #define DRBD_MD_INDEX_INTERNAL -1 #define DRBD_MD_INDEX_FLEX_EXT -2 #define DRBD_MD_INDEX_FLEX_INT -3 -/* Start of the new netlink/connector stuff */ - -#define DRBD_NL_CREATE_DEVICE 0x01 -#define DRBD_NL_SET_DEFAULTS 0x02 - - -/* For searching a vacant cn_idx value */ -#define CN_IDX_STEP 6977 - -struct drbd_nl_cfg_req { - int packet_type; - unsigned int drbd_minor; - int flags; - unsigned short tag_list[]; -}; - -struct drbd_nl_cfg_reply { - int packet_type; - unsigned int minor; - int ret_code; /* enum ret_code or set_st_err_t */ - unsigned short tag_list[]; /* only used with get_* calls */ -}; - #endif diff --git a/include/linux/drbd_genl.h b/include/linux/drbd_genl.h new file mode 100644 index 000000000000..d0d8fac8a6e4 --- /dev/null +++ b/include/linux/drbd_genl.h @@ -0,0 +1,378 @@ +/* + * General overview: + * full generic netlink message: + * |nlmsghdr|genlmsghdr|<payload> + * + * payload: + * |optional fixed size family header|<sequence of netlink attributes> + * + * sequence of netlink attributes: + * I chose to have all "top level" attributes NLA_NESTED, + * corresponding to some real struct. + * So we have a sequence of |tla, len|<nested nla sequence> + * + * nested nla sequence: + * may be empty, or contain a sequence of netlink attributes + * representing the struct fields. + * + * The tag number of any field (regardless of containing struct) + * will be available as T_ ## field_name, + * so you cannot have the same field name in two differnt structs. + * + * The tag numbers themselves are per struct, though, + * so should always begin at 1 (not 0, that is the special "NLA_UNSPEC" type, + * which we won't use here). + * The tag numbers are used as index in the respective nla_policy array. + * + * GENL_struct(tag_name, tag_number, struct name, struct fields) - struct and policy + * genl_magic_struct.h + * generates the struct declaration, + * generates an entry in the tla enum, + * genl_magic_func.h + * generates an entry in the static tla policy + * with .type = NLA_NESTED + * generates the static <struct_name>_nl_policy definition, + * and static conversion functions + * + * genl_magic_func.h + * + * GENL_mc_group(group) + * genl_magic_struct.h + * does nothing + * genl_magic_func.h + * defines and registers the mcast group, + * and provides a send helper + * + * GENL_notification(op_name, op_num, mcast_group, tla list) + * These are notifications to userspace. + * + * genl_magic_struct.h + * generates an entry in the genl_ops enum, + * genl_magic_func.h + * does nothing + * + * mcast group: the name of the mcast group this notification should be + * expected on + * tla list: the list of expected top level attributes, + * for documentation and sanity checking. + * + * GENL_op(op_name, op_num, flags and handler, tla list) - "genl operations" + * These are requests from userspace. + * + * _op and _notification share the same "number space", + * op_nr will be assigned to "genlmsghdr->cmd" + * + * genl_magic_struct.h + * generates an entry in the genl_ops enum, + * genl_magic_func.h + * generates an entry in the static genl_ops array, + * and static register/unregister functions to + * genl_register_family_with_ops(). + * + * flags and handler: + * GENL_op_init( .doit = x, .dumpit = y, .flags = something) + * GENL_doit(x) => .dumpit = NULL, .flags = GENL_ADMIN_PERM + * tla list: the list of expected top level attributes, + * for documentation and sanity checking. + */ + +/* + * STRUCTS + */ + +/* this is sent kernel -> userland on various error conditions, and contains + * informational textual info, which is supposedly human readable. + * The computer relevant return code is in the drbd_genlmsghdr. + */ +GENL_struct(DRBD_NLA_CFG_REPLY, 1, drbd_cfg_reply, + /* "arbitrary" size strings, nla_policy.len = 0 */ + __str_field(1, DRBD_GENLA_F_MANDATORY, info_text, 0) +) + +/* Configuration requests typically need a context to operate on. + * Possible keys are device minor (fits in the drbd_genlmsghdr), + * the replication link (aka connection) name, + * and/or the replication group (aka resource) name, + * and the volume id within the resource. */ +GENL_struct(DRBD_NLA_CFG_CONTEXT, 2, drbd_cfg_context, + __u32_field(1, DRBD_GENLA_F_MANDATORY, ctx_volume) + __str_field(2, DRBD_GENLA_F_MANDATORY, ctx_resource_name, 128) + __bin_field(3, DRBD_GENLA_F_MANDATORY, ctx_my_addr, 128) + __bin_field(4, DRBD_GENLA_F_MANDATORY, ctx_peer_addr, 128) +) + +GENL_struct(DRBD_NLA_DISK_CONF, 3, disk_conf, + __str_field(1, DRBD_F_REQUIRED | DRBD_F_INVARIANT, backing_dev, 128) + __str_field(2, DRBD_F_REQUIRED | DRBD_F_INVARIANT, meta_dev, 128) + __s32_field(3, DRBD_F_REQUIRED | DRBD_F_INVARIANT, meta_dev_idx) + + /* use the resize command to try and change the disk_size */ + __u64_field(4, DRBD_GENLA_F_MANDATORY | DRBD_F_INVARIANT, disk_size) + /* we could change the max_bio_bvecs, + * but it won't propagate through the stack */ + __u32_field(5, DRBD_GENLA_F_MANDATORY | DRBD_F_INVARIANT, max_bio_bvecs) + + __u32_field_def(6, DRBD_GENLA_F_MANDATORY, on_io_error, DRBD_ON_IO_ERROR_DEF) + __u32_field_def(7, DRBD_GENLA_F_MANDATORY, fencing, DRBD_FENCING_DEF) + + __u32_field_def(8, DRBD_GENLA_F_MANDATORY, resync_rate, DRBD_RESYNC_RATE_DEF) + __s32_field_def(9, DRBD_GENLA_F_MANDATORY, resync_after, DRBD_MINOR_NUMBER_DEF) + __u32_field_def(10, DRBD_GENLA_F_MANDATORY, al_extents, DRBD_AL_EXTENTS_DEF) + __u32_field_def(11, DRBD_GENLA_F_MANDATORY, c_plan_ahead, DRBD_C_PLAN_AHEAD_DEF) + __u32_field_def(12, DRBD_GENLA_F_MANDATORY, c_delay_target, DRBD_C_DELAY_TARGET_DEF) + __u32_field_def(13, DRBD_GENLA_F_MANDATORY, c_fill_target, DRBD_C_FILL_TARGET_DEF) + __u32_field_def(14, DRBD_GENLA_F_MANDATORY, c_max_rate, DRBD_C_MAX_RATE_DEF) + __u32_field_def(15, DRBD_GENLA_F_MANDATORY, c_min_rate, DRBD_C_MIN_RATE_DEF) + + __flg_field_def(16, DRBD_GENLA_F_MANDATORY, disk_barrier, DRBD_DISK_BARRIER_DEF) + __flg_field_def(17, DRBD_GENLA_F_MANDATORY, disk_flushes, DRBD_DISK_FLUSHES_DEF) + __flg_field_def(18, DRBD_GENLA_F_MANDATORY, disk_drain, DRBD_DISK_DRAIN_DEF) + __flg_field_def(19, DRBD_GENLA_F_MANDATORY, md_flushes, DRBD_MD_FLUSHES_DEF) + __u32_field_def(20, DRBD_GENLA_F_MANDATORY, disk_timeout, DRBD_DISK_TIMEOUT_DEF) + __u32_field_def(21, 0 /* OPTIONAL */, read_balancing, DRBD_READ_BALANCING_DEF) + /* 9: __u32_field_def(22, DRBD_GENLA_F_MANDATORY, unplug_watermark, DRBD_UNPLUG_WATERMARK_DEF) */ + __flg_field_def(23, 0 /* OPTIONAL */, al_updates, DRBD_AL_UPDATES_DEF) +) + +GENL_struct(DRBD_NLA_RESOURCE_OPTS, 4, res_opts, + __str_field_def(1, DRBD_GENLA_F_MANDATORY, cpu_mask, 32) + __u32_field_def(2, DRBD_GENLA_F_MANDATORY, on_no_data, DRBD_ON_NO_DATA_DEF) +) + +GENL_struct(DRBD_NLA_NET_CONF, 5, net_conf, + __str_field_def(1, DRBD_GENLA_F_MANDATORY | DRBD_F_SENSITIVE, + shared_secret, SHARED_SECRET_MAX) + __str_field_def(2, DRBD_GENLA_F_MANDATORY, cram_hmac_alg, SHARED_SECRET_MAX) + __str_field_def(3, DRBD_GENLA_F_MANDATORY, integrity_alg, SHARED_SECRET_MAX) + __str_field_def(4, DRBD_GENLA_F_MANDATORY, verify_alg, SHARED_SECRET_MAX) + __str_field_def(5, DRBD_GENLA_F_MANDATORY, csums_alg, SHARED_SECRET_MAX) + __u32_field_def(6, DRBD_GENLA_F_MANDATORY, wire_protocol, DRBD_PROTOCOL_DEF) + __u32_field_def(7, DRBD_GENLA_F_MANDATORY, connect_int, DRBD_CONNECT_INT_DEF) + __u32_field_def(8, DRBD_GENLA_F_MANDATORY, timeout, DRBD_TIMEOUT_DEF) + __u32_field_def(9, DRBD_GENLA_F_MANDATORY, ping_int, DRBD_PING_INT_DEF) + __u32_field_def(10, DRBD_GENLA_F_MANDATORY, ping_timeo, DRBD_PING_TIMEO_DEF) + __u32_field_def(11, DRBD_GENLA_F_MANDATORY, sndbuf_size, DRBD_SNDBUF_SIZE_DEF) + __u32_field_def(12, DRBD_GENLA_F_MANDATORY, rcvbuf_size, DRBD_RCVBUF_SIZE_DEF) + __u32_field_def(13, DRBD_GENLA_F_MANDATORY, ko_count, DRBD_KO_COUNT_DEF) + __u32_field_def(14, DRBD_GENLA_F_MANDATORY, max_buffers, DRBD_MAX_BUFFERS_DEF) + __u32_field_def(15, DRBD_GENLA_F_MANDATORY, max_epoch_size, DRBD_MAX_EPOCH_SIZE_DEF) + __u32_field_def(16, DRBD_GENLA_F_MANDATORY, unplug_watermark, DRBD_UNPLUG_WATERMARK_DEF) + __u32_field_def(17, DRBD_GENLA_F_MANDATORY, after_sb_0p, DRBD_AFTER_SB_0P_DEF) + __u32_field_def(18, DRBD_GENLA_F_MANDATORY, after_sb_1p, DRBD_AFTER_SB_1P_DEF) + __u32_field_def(19, DRBD_GENLA_F_MANDATORY, after_sb_2p, DRBD_AFTER_SB_2P_DEF) + __u32_field_def(20, DRBD_GENLA_F_MANDATORY, rr_conflict, DRBD_RR_CONFLICT_DEF) + __u32_field_def(21, DRBD_GENLA_F_MANDATORY, on_congestion, DRBD_ON_CONGESTION_DEF) + __u32_field_def(22, DRBD_GENLA_F_MANDATORY, cong_fill, DRBD_CONG_FILL_DEF) + __u32_field_def(23, DRBD_GENLA_F_MANDATORY, cong_extents, DRBD_CONG_EXTENTS_DEF) + __flg_field_def(24, DRBD_GENLA_F_MANDATORY, two_primaries, DRBD_ALLOW_TWO_PRIMARIES_DEF) + __flg_field(25, DRBD_GENLA_F_MANDATORY | DRBD_F_INVARIANT, discard_my_data) + __flg_field_def(26, DRBD_GENLA_F_MANDATORY, tcp_cork, DRBD_TCP_CORK_DEF) + __flg_field_def(27, DRBD_GENLA_F_MANDATORY, always_asbp, DRBD_ALWAYS_ASBP_DEF) + __flg_field(28, DRBD_GENLA_F_MANDATORY | DRBD_F_INVARIANT, tentative) + __flg_field_def(29, DRBD_GENLA_F_MANDATORY, use_rle, DRBD_USE_RLE_DEF) + /* 9: __u32_field_def(30, DRBD_GENLA_F_MANDATORY, fencing_policy, DRBD_FENCING_DEF) */ +) + +GENL_struct(DRBD_NLA_SET_ROLE_PARMS, 6, set_role_parms, + __flg_field(1, DRBD_GENLA_F_MANDATORY, assume_uptodate) +) + +GENL_struct(DRBD_NLA_RESIZE_PARMS, 7, resize_parms, + __u64_field(1, DRBD_GENLA_F_MANDATORY, resize_size) + __flg_field(2, DRBD_GENLA_F_MANDATORY, resize_force) + __flg_field(3, DRBD_GENLA_F_MANDATORY, no_resync) +) + +GENL_struct(DRBD_NLA_STATE_INFO, 8, state_info, + /* the reason of the broadcast, + * if this is an event triggered broadcast. */ + __u32_field(1, DRBD_GENLA_F_MANDATORY, sib_reason) + __u32_field(2, DRBD_F_REQUIRED, current_state) + __u64_field(3, DRBD_GENLA_F_MANDATORY, capacity) + __u64_field(4, DRBD_GENLA_F_MANDATORY, ed_uuid) + + /* These are for broadcast from after state change work. + * prev_state and new_state are from the moment the state change took + * place, new_state is not neccessarily the same as current_state, + * there may have been more state changes since. Which will be + * broadcasted soon, in their respective after state change work. */ + __u32_field(5, DRBD_GENLA_F_MANDATORY, prev_state) + __u32_field(6, DRBD_GENLA_F_MANDATORY, new_state) + + /* if we have a local disk: */ + __bin_field(7, DRBD_GENLA_F_MANDATORY, uuids, (UI_SIZE*sizeof(__u64))) + __u32_field(8, DRBD_GENLA_F_MANDATORY, disk_flags) + __u64_field(9, DRBD_GENLA_F_MANDATORY, bits_total) + __u64_field(10, DRBD_GENLA_F_MANDATORY, bits_oos) + /* and in case resync or online verify is active */ + __u64_field(11, DRBD_GENLA_F_MANDATORY, bits_rs_total) + __u64_field(12, DRBD_GENLA_F_MANDATORY, bits_rs_failed) + + /* for pre and post notifications of helper execution */ + __str_field(13, DRBD_GENLA_F_MANDATORY, helper, 32) + __u32_field(14, DRBD_GENLA_F_MANDATORY, helper_exit_code) + + __u64_field(15, 0, send_cnt) + __u64_field(16, 0, recv_cnt) + __u64_field(17, 0, read_cnt) + __u64_field(18, 0, writ_cnt) + __u64_field(19, 0, al_writ_cnt) + __u64_field(20, 0, bm_writ_cnt) + __u32_field(21, 0, ap_bio_cnt) + __u32_field(22, 0, ap_pending_cnt) + __u32_field(23, 0, rs_pending_cnt) +) + +GENL_struct(DRBD_NLA_START_OV_PARMS, 9, start_ov_parms, + __u64_field(1, DRBD_GENLA_F_MANDATORY, ov_start_sector) + __u64_field(2, DRBD_GENLA_F_MANDATORY, ov_stop_sector) +) + +GENL_struct(DRBD_NLA_NEW_C_UUID_PARMS, 10, new_c_uuid_parms, + __flg_field(1, DRBD_GENLA_F_MANDATORY, clear_bm) +) + +GENL_struct(DRBD_NLA_TIMEOUT_PARMS, 11, timeout_parms, + __u32_field(1, DRBD_F_REQUIRED, timeout_type) +) + +GENL_struct(DRBD_NLA_DISCONNECT_PARMS, 12, disconnect_parms, + __flg_field(1, DRBD_GENLA_F_MANDATORY, force_disconnect) +) + +GENL_struct(DRBD_NLA_DETACH_PARMS, 13, detach_parms, + __flg_field(1, DRBD_GENLA_F_MANDATORY, force_detach) +) + +/* + * Notifications and commands (genlmsghdr->cmd) + */ +GENL_mc_group(events) + + /* kernel -> userspace announcement of changes */ +GENL_notification( + DRBD_EVENT, 1, events, + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED) + GENL_tla_expected(DRBD_NLA_STATE_INFO, DRBD_F_REQUIRED) + GENL_tla_expected(DRBD_NLA_NET_CONF, DRBD_GENLA_F_MANDATORY) + GENL_tla_expected(DRBD_NLA_DISK_CONF, DRBD_GENLA_F_MANDATORY) + GENL_tla_expected(DRBD_NLA_SYNCER_CONF, DRBD_GENLA_F_MANDATORY) +) + + /* query kernel for specific or all info */ +GENL_op( + DRBD_ADM_GET_STATUS, 2, + GENL_op_init( + .doit = drbd_adm_get_status, + .dumpit = drbd_adm_get_status_all, + /* anyone may ask for the status, + * it is broadcasted anyways */ + ), + /* To select the object .doit. + * Or a subset of objects in .dumpit. */ + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_GENLA_F_MANDATORY) +) + + /* add DRBD minor devices as volumes to resources */ +GENL_op(DRBD_ADM_NEW_MINOR, 5, GENL_doit(drbd_adm_add_minor), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)) +GENL_op(DRBD_ADM_DEL_MINOR, 6, GENL_doit(drbd_adm_delete_minor), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)) + + /* add or delete resources */ +GENL_op(DRBD_ADM_NEW_RESOURCE, 7, GENL_doit(drbd_adm_new_resource), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)) +GENL_op(DRBD_ADM_DEL_RESOURCE, 8, GENL_doit(drbd_adm_del_resource), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)) + +GENL_op(DRBD_ADM_RESOURCE_OPTS, 9, + GENL_doit(drbd_adm_resource_opts), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED) + GENL_tla_expected(DRBD_NLA_RESOURCE_OPTS, DRBD_GENLA_F_MANDATORY) +) + +GENL_op( + DRBD_ADM_CONNECT, 10, + GENL_doit(drbd_adm_connect), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED) + GENL_tla_expected(DRBD_NLA_NET_CONF, DRBD_F_REQUIRED) +) + +GENL_op( + DRBD_ADM_CHG_NET_OPTS, 29, + GENL_doit(drbd_adm_net_opts), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED) + GENL_tla_expected(DRBD_NLA_NET_CONF, DRBD_F_REQUIRED) +) + +GENL_op(DRBD_ADM_DISCONNECT, 11, GENL_doit(drbd_adm_disconnect), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)) + +GENL_op(DRBD_ADM_ATTACH, 12, + GENL_doit(drbd_adm_attach), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED) + GENL_tla_expected(DRBD_NLA_DISK_CONF, DRBD_F_REQUIRED) +) + +GENL_op(DRBD_ADM_CHG_DISK_OPTS, 28, + GENL_doit(drbd_adm_disk_opts), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED) + GENL_tla_expected(DRBD_NLA_DISK_OPTS, DRBD_F_REQUIRED) +) + +GENL_op( + DRBD_ADM_RESIZE, 13, + GENL_doit(drbd_adm_resize), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED) + GENL_tla_expected(DRBD_NLA_RESIZE_PARMS, DRBD_GENLA_F_MANDATORY) +) + +GENL_op( + DRBD_ADM_PRIMARY, 14, + GENL_doit(drbd_adm_set_role), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED) + GENL_tla_expected(DRBD_NLA_SET_ROLE_PARMS, DRBD_F_REQUIRED) +) + +GENL_op( + DRBD_ADM_SECONDARY, 15, + GENL_doit(drbd_adm_set_role), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED) + GENL_tla_expected(DRBD_NLA_SET_ROLE_PARMS, DRBD_F_REQUIRED) +) + +GENL_op( + DRBD_ADM_NEW_C_UUID, 16, + GENL_doit(drbd_adm_new_c_uuid), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED) + GENL_tla_expected(DRBD_NLA_NEW_C_UUID_PARMS, DRBD_GENLA_F_MANDATORY) +) + +GENL_op( + DRBD_ADM_START_OV, 17, + GENL_doit(drbd_adm_start_ov), + GENL_tla_expected(DRBD_NLA_START_OV_PARMS, DRBD_GENLA_F_MANDATORY) +) + +GENL_op(DRBD_ADM_DETACH, 18, GENL_doit(drbd_adm_detach), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED) + GENL_tla_expected(DRBD_NLA_DETACH_PARMS, DRBD_GENLA_F_MANDATORY)) + +GENL_op(DRBD_ADM_INVALIDATE, 19, GENL_doit(drbd_adm_invalidate), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)) +GENL_op(DRBD_ADM_INVAL_PEER, 20, GENL_doit(drbd_adm_invalidate_peer), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)) +GENL_op(DRBD_ADM_PAUSE_SYNC, 21, GENL_doit(drbd_adm_pause_sync), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)) +GENL_op(DRBD_ADM_RESUME_SYNC, 22, GENL_doit(drbd_adm_resume_sync), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)) +GENL_op(DRBD_ADM_SUSPEND_IO, 23, GENL_doit(drbd_adm_suspend_io), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)) +GENL_op(DRBD_ADM_RESUME_IO, 24, GENL_doit(drbd_adm_resume_io), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)) +GENL_op(DRBD_ADM_OUTDATE, 25, GENL_doit(drbd_adm_outdate), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)) +GENL_op(DRBD_ADM_GET_TIMEOUT_TYPE, 26, GENL_doit(drbd_adm_get_timeout_type), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)) +GENL_op(DRBD_ADM_DOWN, 27, GENL_doit(drbd_adm_down), + GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, DRBD_F_REQUIRED)) diff --git a/include/linux/drbd_genl_api.h b/include/linux/drbd_genl_api.h new file mode 100644 index 000000000000..9ef50d51e34e --- /dev/null +++ b/include/linux/drbd_genl_api.h @@ -0,0 +1,55 @@ +#ifndef DRBD_GENL_STRUCT_H +#define DRBD_GENL_STRUCT_H + +/** + * struct drbd_genlmsghdr - DRBD specific header used in NETLINK_GENERIC requests + * @minor: + * For admin requests (user -> kernel): which minor device to operate on. + * For (unicast) replies or informational (broadcast) messages + * (kernel -> user): which minor device the information is about. + * If we do not operate on minors, but on connections or resources, + * the minor value shall be (~0), and the attribute DRBD_NLA_CFG_CONTEXT + * is used instead. + * @flags: possible operation modifiers (relevant only for user->kernel): + * DRBD_GENL_F_SET_DEFAULTS + * @volume: + * When creating a new minor (adding it to a resource), the resource needs + * to know which volume number within the resource this is supposed to be. + * The volume number corresponds to the same volume number on the remote side, + * whereas the minor number on the remote side may be different + * (union with flags). + * @ret_code: kernel->userland unicast cfg reply return code (union with flags); + */ +struct drbd_genlmsghdr { + __u32 minor; + union { + __u32 flags; + __s32 ret_code; + }; +}; + +/* To be used in drbd_genlmsghdr.flags */ +enum { + DRBD_GENL_F_SET_DEFAULTS = 1, +}; + +enum drbd_state_info_bcast_reason { + SIB_GET_STATUS_REPLY = 1, + SIB_STATE_CHANGE = 2, + SIB_HELPER_PRE = 3, + SIB_HELPER_POST = 4, + SIB_SYNC_PROGRESS = 5, +}; + +/* hack around predefined gcc/cpp "linux=1", + * we cannot possibly include <1/drbd_genl.h> */ +#undef linux + +#include <linux/drbd.h> +#define GENL_MAGIC_VERSION API_VERSION +#define GENL_MAGIC_FAMILY drbd +#define GENL_MAGIC_FAMILY_HDRSZ sizeof(struct drbd_genlmsghdr) +#define GENL_MAGIC_INCLUDE_FILE <linux/drbd_genl.h> +#include <linux/genl_magic_struct.h> + +#endif diff --git a/include/linux/drbd_limits.h b/include/linux/drbd_limits.h index fb670bf603f7..1fa19c5f5e64 100644 --- a/include/linux/drbd_limits.h +++ b/include/linux/drbd_limits.h @@ -16,29 +16,37 @@ #define DEBUG_RANGE_CHECK 0 #define DRBD_MINOR_COUNT_MIN 1 -#define DRBD_MINOR_COUNT_MAX 256 +#define DRBD_MINOR_COUNT_MAX 255 #define DRBD_MINOR_COUNT_DEF 32 +#define DRBD_MINOR_COUNT_SCALE '1' + +#define DRBD_VOLUME_MAX 65535 #define DRBD_DIALOG_REFRESH_MIN 0 #define DRBD_DIALOG_REFRESH_MAX 600 +#define DRBD_DIALOG_REFRESH_SCALE '1' /* valid port number */ #define DRBD_PORT_MIN 1 #define DRBD_PORT_MAX 0xffff +#define DRBD_PORT_SCALE '1' /* startup { */ /* if you want more than 3.4 days, disable */ #define DRBD_WFC_TIMEOUT_MIN 0 #define DRBD_WFC_TIMEOUT_MAX 300000 #define DRBD_WFC_TIMEOUT_DEF 0 +#define DRBD_WFC_TIMEOUT_SCALE '1' #define DRBD_DEGR_WFC_TIMEOUT_MIN 0 #define DRBD_DEGR_WFC_TIMEOUT_MAX 300000 #define DRBD_DEGR_WFC_TIMEOUT_DEF 0 +#define DRBD_DEGR_WFC_TIMEOUT_SCALE '1' #define DRBD_OUTDATED_WFC_TIMEOUT_MIN 0 #define DRBD_OUTDATED_WFC_TIMEOUT_MAX 300000 #define DRBD_OUTDATED_WFC_TIMEOUT_DEF 0 +#define DRBD_OUTDATED_WFC_TIMEOUT_SCALE '1' /* }*/ /* net { */ @@ -47,75 +55,91 @@ #define DRBD_TIMEOUT_MIN 1 #define DRBD_TIMEOUT_MAX 600 #define DRBD_TIMEOUT_DEF 60 /* 6 seconds */ +#define DRBD_TIMEOUT_SCALE '1' /* If backing disk takes longer than disk_timeout, mark the disk as failed */ #define DRBD_DISK_TIMEOUT_MIN 0 /* 0 = disabled */ #define DRBD_DISK_TIMEOUT_MAX 6000 /* 10 Minutes */ #define DRBD_DISK_TIMEOUT_DEF 0 /* disabled */ +#define DRBD_DISK_TIMEOUT_SCALE '1' /* active connection retries when C_WF_CONNECTION */ #define DRBD_CONNECT_INT_MIN 1 #define DRBD_CONNECT_INT_MAX 120 #define DRBD_CONNECT_INT_DEF 10 /* seconds */ +#define DRBD_CONNECT_INT_SCALE '1' /* keep-alive probes when idle */ #define DRBD_PING_INT_MIN 1 #define DRBD_PING_INT_MAX 120 #define DRBD_PING_INT_DEF 10 +#define DRBD_PING_INT_SCALE '1' /* timeout for the ping packets.*/ #define DRBD_PING_TIMEO_MIN 1 #define DRBD_PING_TIMEO_MAX 300 #define DRBD_PING_TIMEO_DEF 5 +#define DRBD_PING_TIMEO_SCALE '1' /* max number of write requests between write barriers */ #define DRBD_MAX_EPOCH_SIZE_MIN 1 #define DRBD_MAX_EPOCH_SIZE_MAX 20000 #define DRBD_MAX_EPOCH_SIZE_DEF 2048 +#define DRBD_MAX_EPOCH_SIZE_SCALE '1' /* I don't think that a tcp send buffer of more than 10M is useful */ #define DRBD_SNDBUF_SIZE_MIN 0 #define DRBD_SNDBUF_SIZE_MAX (10<<20) #define DRBD_SNDBUF_SIZE_DEF 0 +#define DRBD_SNDBUF_SIZE_SCALE '1' #define DRBD_RCVBUF_SIZE_MIN 0 #define DRBD_RCVBUF_SIZE_MAX (10<<20) #define DRBD_RCVBUF_SIZE_DEF 0 +#define DRBD_RCVBUF_SIZE_SCALE '1' /* @4k PageSize -> 128kB - 512MB */ #define DRBD_MAX_BUFFERS_MIN 32 #define DRBD_MAX_BUFFERS_MAX 131072 #define DRBD_MAX_BUFFERS_DEF 2048 +#define DRBD_MAX_BUFFERS_SCALE '1' /* @4k PageSize -> 4kB - 512MB */ #define DRBD_UNPLUG_WATERMARK_MIN 1 #define DRBD_UNPLUG_WATERMARK_MAX 131072 #define DRBD_UNPLUG_WATERMARK_DEF (DRBD_MAX_BUFFERS_DEF/16) +#define DRBD_UNPLUG_WATERMARK_SCALE '1' /* 0 is disabled. * 200 should be more than enough even for very short timeouts */ #define DRBD_KO_COUNT_MIN 0 #define DRBD_KO_COUNT_MAX 200 -#define DRBD_KO_COUNT_DEF 0 +#define DRBD_KO_COUNT_DEF 7 +#define DRBD_KO_COUNT_SCALE '1' /* } */ /* syncer { */ /* FIXME allow rate to be zero? */ -#define DRBD_RATE_MIN 1 +#define DRBD_RESYNC_RATE_MIN 1 /* channel bonding 10 GbE, or other hardware */ -#define DRBD_RATE_MAX (4 << 20) -#define DRBD_RATE_DEF 250 /* kb/second */ +#define DRBD_RESYNC_RATE_MAX (4 << 20) +#define DRBD_RESYNC_RATE_DEF 250 +#define DRBD_RESYNC_RATE_SCALE 'k' /* kilobytes */ /* less than 7 would hit performance unnecessarily. - * 3833 is the largest prime that still does fit - * into 64 sectors of activity log */ + * 919 slots context information per transaction, + * 32k activity log, 4k transaction size, + * one transaction in flight: + * 919 * 7 = 6433 */ #define DRBD_AL_EXTENTS_MIN 7 -#define DRBD_AL_EXTENTS_MAX 3833 -#define DRBD_AL_EXTENTS_DEF 127 +#define DRBD_AL_EXTENTS_MAX 6433 +#define DRBD_AL_EXTENTS_DEF 1237 +#define DRBD_AL_EXTENTS_SCALE '1' -#define DRBD_AFTER_MIN -1 -#define DRBD_AFTER_MAX 255 -#define DRBD_AFTER_DEF -1 +#define DRBD_MINOR_NUMBER_MIN -1 +#define DRBD_MINOR_NUMBER_MAX ((1 << 20) - 1) +#define DRBD_MINOR_NUMBER_DEF -1 +#define DRBD_MINOR_NUMBER_SCALE '1' /* } */ @@ -124,11 +148,12 @@ * the upper limit with 64bit kernel, enough ram and flexible meta data * is 1 PiB, currently. */ /* DRBD_MAX_SECTORS */ -#define DRBD_DISK_SIZE_SECT_MIN 0 -#define DRBD_DISK_SIZE_SECT_MAX (1 * (2LLU << 40)) -#define DRBD_DISK_SIZE_SECT_DEF 0 /* = disabled = no user size... */ +#define DRBD_DISK_SIZE_MIN 0 +#define DRBD_DISK_SIZE_MAX (1 * (2LLU << 40)) +#define DRBD_DISK_SIZE_DEF 0 /* = disabled = no user size... */ +#define DRBD_DISK_SIZE_SCALE 's' /* sectors */ -#define DRBD_ON_IO_ERROR_DEF EP_PASS_ON +#define DRBD_ON_IO_ERROR_DEF EP_DETACH #define DRBD_FENCING_DEF FP_DONT_CARE #define DRBD_AFTER_SB_0P_DEF ASB_DISCONNECT #define DRBD_AFTER_SB_1P_DEF ASB_DISCONNECT @@ -136,38 +161,59 @@ #define DRBD_RR_CONFLICT_DEF ASB_DISCONNECT #define DRBD_ON_NO_DATA_DEF OND_IO_ERROR #define DRBD_ON_CONGESTION_DEF OC_BLOCK +#define DRBD_READ_BALANCING_DEF RB_PREFER_LOCAL #define DRBD_MAX_BIO_BVECS_MIN 0 #define DRBD_MAX_BIO_BVECS_MAX 128 #define DRBD_MAX_BIO_BVECS_DEF 0 +#define DRBD_MAX_BIO_BVECS_SCALE '1' #define DRBD_C_PLAN_AHEAD_MIN 0 #define DRBD_C_PLAN_AHEAD_MAX 300 -#define DRBD_C_PLAN_AHEAD_DEF 0 /* RS rate controller disabled by default */ +#define DRBD_C_PLAN_AHEAD_DEF 20 +#define DRBD_C_PLAN_AHEAD_SCALE '1' #define DRBD_C_DELAY_TARGET_MIN 1 #define DRBD_C_DELAY_TARGET_MAX 100 #define DRBD_C_DELAY_TARGET_DEF 10 +#define DRBD_C_DELAY_TARGET_SCALE '1' #define DRBD_C_FILL_TARGET_MIN 0 #define DRBD_C_FILL_TARGET_MAX (1<<20) /* 500MByte in sec */ -#define DRBD_C_FILL_TARGET_DEF 0 /* By default disabled -> controlled by delay_target */ +#define DRBD_C_FILL_TARGET_DEF 100 /* Try to place 50KiB in socket send buffer during resync */ +#define DRBD_C_FILL_TARGET_SCALE 's' /* sectors */ -#define DRBD_C_MAX_RATE_MIN 250 /* kByte/sec */ +#define DRBD_C_MAX_RATE_MIN 250 #define DRBD_C_MAX_RATE_MAX (4 << 20) #define DRBD_C_MAX_RATE_DEF 102400 +#define DRBD_C_MAX_RATE_SCALE 'k' /* kilobytes */ -#define DRBD_C_MIN_RATE_MIN 0 /* kByte/sec */ +#define DRBD_C_MIN_RATE_MIN 0 #define DRBD_C_MIN_RATE_MAX (4 << 20) -#define DRBD_C_MIN_RATE_DEF 4096 +#define DRBD_C_MIN_RATE_DEF 250 +#define DRBD_C_MIN_RATE_SCALE 'k' /* kilobytes */ #define DRBD_CONG_FILL_MIN 0 #define DRBD_CONG_FILL_MAX (10<<21) /* 10GByte in sectors */ #define DRBD_CONG_FILL_DEF 0 +#define DRBD_CONG_FILL_SCALE 's' /* sectors */ #define DRBD_CONG_EXTENTS_MIN DRBD_AL_EXTENTS_MIN #define DRBD_CONG_EXTENTS_MAX DRBD_AL_EXTENTS_MAX #define DRBD_CONG_EXTENTS_DEF DRBD_AL_EXTENTS_DEF +#define DRBD_CONG_EXTENTS_SCALE DRBD_AL_EXTENTS_SCALE + +#define DRBD_PROTOCOL_DEF DRBD_PROT_C + +#define DRBD_DISK_BARRIER_DEF 0 +#define DRBD_DISK_FLUSHES_DEF 1 +#define DRBD_DISK_DRAIN_DEF 1 +#define DRBD_MD_FLUSHES_DEF 1 +#define DRBD_TCP_CORK_DEF 1 +#define DRBD_AL_UPDATES_DEF 1 + +#define DRBD_ALLOW_TWO_PRIMARIES_DEF 0 +#define DRBD_ALWAYS_ASBP_DEF 0 +#define DRBD_USE_RLE_DEF 1 -#undef RANGE #endif diff --git a/include/linux/drbd_nl.h b/include/linux/drbd_nl.h deleted file mode 100644 index a8706f08ab36..000000000000 --- a/include/linux/drbd_nl.h +++ /dev/null @@ -1,163 +0,0 @@ -/* - PAKET( name, - TYPE ( pn, pr, member ) - ... - ) - - You may never reissue one of the pn arguments -*/ - -#if !defined(NL_PACKET) || !defined(NL_STRING) || !defined(NL_INTEGER) || !defined(NL_BIT) || !defined(NL_INT64) -#error "The macros NL_PACKET, NL_STRING, NL_INTEGER, NL_INT64 and NL_BIT needs to be defined" -#endif - -NL_PACKET(primary, 1, - NL_BIT( 1, T_MAY_IGNORE, primary_force) -) - -NL_PACKET(secondary, 2, ) - -NL_PACKET(disk_conf, 3, - NL_INT64( 2, T_MAY_IGNORE, disk_size) - NL_STRING( 3, T_MANDATORY, backing_dev, 128) - NL_STRING( 4, T_MANDATORY, meta_dev, 128) - NL_INTEGER( 5, T_MANDATORY, meta_dev_idx) - NL_INTEGER( 6, T_MAY_IGNORE, on_io_error) - NL_INTEGER( 7, T_MAY_IGNORE, fencing) - NL_BIT( 37, T_MAY_IGNORE, use_bmbv) - NL_BIT( 53, T_MAY_IGNORE, no_disk_flush) - NL_BIT( 54, T_MAY_IGNORE, no_md_flush) - /* 55 max_bio_size was available in 8.2.6rc2 */ - NL_INTEGER( 56, T_MAY_IGNORE, max_bio_bvecs) - NL_BIT( 57, T_MAY_IGNORE, no_disk_barrier) - NL_BIT( 58, T_MAY_IGNORE, no_disk_drain) - NL_INTEGER( 89, T_MAY_IGNORE, disk_timeout) -) - -NL_PACKET(detach, 4, - NL_BIT( 88, T_MANDATORY, detach_force) -) - -NL_PACKET(net_conf, 5, - NL_STRING( 8, T_MANDATORY, my_addr, 128) - NL_STRING( 9, T_MANDATORY, peer_addr, 128) - NL_STRING( 10, T_MAY_IGNORE, shared_secret, SHARED_SECRET_MAX) - NL_STRING( 11, T_MAY_IGNORE, cram_hmac_alg, SHARED_SECRET_MAX) - NL_STRING( 44, T_MAY_IGNORE, integrity_alg, SHARED_SECRET_MAX) - NL_INTEGER( 14, T_MAY_IGNORE, timeout) - NL_INTEGER( 15, T_MANDATORY, wire_protocol) - NL_INTEGER( 16, T_MAY_IGNORE, try_connect_int) - NL_INTEGER( 17, T_MAY_IGNORE, ping_int) - NL_INTEGER( 18, T_MAY_IGNORE, max_epoch_size) - NL_INTEGER( 19, T_MAY_IGNORE, max_buffers) - NL_INTEGER( 20, T_MAY_IGNORE, unplug_watermark) - NL_INTEGER( 21, T_MAY_IGNORE, sndbuf_size) - NL_INTEGER( 22, T_MAY_IGNORE, ko_count) - NL_INTEGER( 24, T_MAY_IGNORE, after_sb_0p) - NL_INTEGER( 25, T_MAY_IGNORE, after_sb_1p) - NL_INTEGER( 26, T_MAY_IGNORE, after_sb_2p) - NL_INTEGER( 39, T_MAY_IGNORE, rr_conflict) - NL_INTEGER( 40, T_MAY_IGNORE, ping_timeo) - NL_INTEGER( 67, T_MAY_IGNORE, rcvbuf_size) - NL_INTEGER( 81, T_MAY_IGNORE, on_congestion) - NL_INTEGER( 82, T_MAY_IGNORE, cong_fill) - NL_INTEGER( 83, T_MAY_IGNORE, cong_extents) - /* 59 addr_family was available in GIT, never released */ - NL_BIT( 60, T_MANDATORY, mind_af) - NL_BIT( 27, T_MAY_IGNORE, want_lose) - NL_BIT( 28, T_MAY_IGNORE, two_primaries) - NL_BIT( 41, T_MAY_IGNORE, always_asbp) - NL_BIT( 61, T_MAY_IGNORE, no_cork) - NL_BIT( 62, T_MANDATORY, auto_sndbuf_size) - NL_BIT( 70, T_MANDATORY, dry_run) -) - -NL_PACKET(disconnect, 6, - NL_BIT( 84, T_MAY_IGNORE, force) -) - -NL_PACKET(resize, 7, - NL_INT64( 29, T_MAY_IGNORE, resize_size) - NL_BIT( 68, T_MAY_IGNORE, resize_force) - NL_BIT( 69, T_MANDATORY, no_resync) -) - -NL_PACKET(syncer_conf, 8, - NL_INTEGER( 30, T_MAY_IGNORE, rate) - NL_INTEGER( 31, T_MAY_IGNORE, after) - NL_INTEGER( 32, T_MAY_IGNORE, al_extents) -/* NL_INTEGER( 71, T_MAY_IGNORE, dp_volume) - * NL_INTEGER( 72, T_MAY_IGNORE, dp_interval) - * NL_INTEGER( 73, T_MAY_IGNORE, throttle_th) - * NL_INTEGER( 74, T_MAY_IGNORE, hold_off_th) - * feature will be reimplemented differently with 8.3.9 */ - NL_STRING( 52, T_MAY_IGNORE, verify_alg, SHARED_SECRET_MAX) - NL_STRING( 51, T_MAY_IGNORE, cpu_mask, 32) - NL_STRING( 64, T_MAY_IGNORE, csums_alg, SHARED_SECRET_MAX) - NL_BIT( 65, T_MAY_IGNORE, use_rle) - NL_INTEGER( 75, T_MAY_IGNORE, on_no_data) - NL_INTEGER( 76, T_MAY_IGNORE, c_plan_ahead) - NL_INTEGER( 77, T_MAY_IGNORE, c_delay_target) - NL_INTEGER( 78, T_MAY_IGNORE, c_fill_target) - NL_INTEGER( 79, T_MAY_IGNORE, c_max_rate) - NL_INTEGER( 80, T_MAY_IGNORE, c_min_rate) -) - -NL_PACKET(invalidate, 9, ) -NL_PACKET(invalidate_peer, 10, ) -NL_PACKET(pause_sync, 11, ) -NL_PACKET(resume_sync, 12, ) -NL_PACKET(suspend_io, 13, ) -NL_PACKET(resume_io, 14, ) -NL_PACKET(outdate, 15, ) -NL_PACKET(get_config, 16, ) -NL_PACKET(get_state, 17, - NL_INTEGER( 33, T_MAY_IGNORE, state_i) -) - -NL_PACKET(get_uuids, 18, - NL_STRING( 34, T_MAY_IGNORE, uuids, (UI_SIZE*sizeof(__u64))) - NL_INTEGER( 35, T_MAY_IGNORE, uuids_flags) -) - -NL_PACKET(get_timeout_flag, 19, - NL_BIT( 36, T_MAY_IGNORE, use_degraded) -) - -NL_PACKET(call_helper, 20, - NL_STRING( 38, T_MAY_IGNORE, helper, 32) -) - -/* Tag nr 42 already allocated in drbd-8.1 development. */ - -NL_PACKET(sync_progress, 23, - NL_INTEGER( 43, T_MAY_IGNORE, sync_progress) -) - -NL_PACKET(dump_ee, 24, - NL_STRING( 45, T_MAY_IGNORE, dump_ee_reason, 32) - NL_STRING( 46, T_MAY_IGNORE, seen_digest, SHARED_SECRET_MAX) - NL_STRING( 47, T_MAY_IGNORE, calc_digest, SHARED_SECRET_MAX) - NL_INT64( 48, T_MAY_IGNORE, ee_sector) - NL_INT64( 49, T_MAY_IGNORE, ee_block_id) - NL_STRING( 50, T_MAY_IGNORE, ee_data, 32 << 10) -) - -NL_PACKET(start_ov, 25, - NL_INT64( 66, T_MAY_IGNORE, start_sector) -) - -NL_PACKET(new_c_uuid, 26, - NL_BIT( 63, T_MANDATORY, clear_bm) -) - -#ifdef NL_RESPONSE -NL_RESPONSE(return_code_only, 27) -#endif - -#undef NL_PACKET -#undef NL_INTEGER -#undef NL_INT64 -#undef NL_BIT -#undef NL_STRING -#undef NL_RESPONSE diff --git a/include/linux/drbd_tag_magic.h b/include/linux/drbd_tag_magic.h deleted file mode 100644 index 82de1f9e48b1..000000000000 --- a/include/linux/drbd_tag_magic.h +++ /dev/null @@ -1,84 +0,0 @@ -#ifndef DRBD_TAG_MAGIC_H -#define DRBD_TAG_MAGIC_H - -#define TT_END 0 -#define TT_REMOVED 0xE000 - -/* declare packet_type enums */ -enum packet_types { -#define NL_PACKET(name, number, fields) P_ ## name = number, -#define NL_RESPONSE(name, number) P_ ## name = number, -#define NL_INTEGER(pn, pr, member) -#define NL_INT64(pn, pr, member) -#define NL_BIT(pn, pr, member) -#define NL_STRING(pn, pr, member, len) -#include <linux/drbd_nl.h> - P_nl_after_last_packet, -}; - -/* These struct are used to deduce the size of the tag lists: */ -#define NL_PACKET(name, number, fields) \ - struct name ## _tag_len_struct { fields }; -#define NL_INTEGER(pn, pr, member) \ - int member; int tag_and_len ## member; -#define NL_INT64(pn, pr, member) \ - __u64 member; int tag_and_len ## member; -#define NL_BIT(pn, pr, member) \ - unsigned char member:1; int tag_and_len ## member; -#define NL_STRING(pn, pr, member, len) \ - unsigned char member[len]; int member ## _len; \ - int tag_and_len ## member; -#include <linux/drbd_nl.h> - -/* declare tag-list-sizes */ -static const int tag_list_sizes[] = { -#define NL_PACKET(name, number, fields) 2 fields , -#define NL_INTEGER(pn, pr, member) + 4 + 4 -#define NL_INT64(pn, pr, member) + 4 + 8 -#define NL_BIT(pn, pr, member) + 4 + 1 -#define NL_STRING(pn, pr, member, len) + 4 + (len) -#include <linux/drbd_nl.h> -}; - -/* The two highest bits are used for the tag type */ -#define TT_MASK 0xC000 -#define TT_INTEGER 0x0000 -#define TT_INT64 0x4000 -#define TT_BIT 0x8000 -#define TT_STRING 0xC000 -/* The next bit indicates if processing of the tag is mandatory */ -#define T_MANDATORY 0x2000 -#define T_MAY_IGNORE 0x0000 -#define TN_MASK 0x1fff -/* The remaining 13 bits are used to enumerate the tags */ - -#define tag_type(T) ((T) & TT_MASK) -#define tag_number(T) ((T) & TN_MASK) - -/* declare tag enums */ -#define NL_PACKET(name, number, fields) fields -enum drbd_tags { -#define NL_INTEGER(pn, pr, member) T_ ## member = pn | TT_INTEGER | pr , -#define NL_INT64(pn, pr, member) T_ ## member = pn | TT_INT64 | pr , -#define NL_BIT(pn, pr, member) T_ ## member = pn | TT_BIT | pr , -#define NL_STRING(pn, pr, member, len) T_ ## member = pn | TT_STRING | pr , -#include <linux/drbd_nl.h> -}; - -struct tag { - const char *name; - int type_n_flags; - int max_len; -}; - -/* declare tag names */ -#define NL_PACKET(name, number, fields) fields -static const struct tag tag_descriptions[] = { -#define NL_INTEGER(pn, pr, member) [ pn ] = { #member, TT_INTEGER | pr, sizeof(int) }, -#define NL_INT64(pn, pr, member) [ pn ] = { #member, TT_INT64 | pr, sizeof(__u64) }, -#define NL_BIT(pn, pr, member) [ pn ] = { #member, TT_BIT | pr, sizeof(int) }, -#define NL_STRING(pn, pr, member, len) [ pn ] = { #member, TT_STRING | pr, (len) }, -#include <linux/drbd_nl.h> -}; - -#endif diff --git a/include/linux/fs.h b/include/linux/fs.h index 5abf703d06ba..a823d4be38e7 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1810,6 +1810,8 @@ struct file_system_type { #define FS_REQUIRES_DEV 1 #define FS_BINARY_MOUNTDATA 2 #define FS_HAS_SUBTYPE 4 +#define FS_USERNS_MOUNT 8 /* Can be mounted by userns root */ +#define FS_USERNS_DEV_MOUNT 16 /* A userns mount does not imply MNT_NODEV */ #define FS_REVAL_DOT 16384 /* Check the paths ".", ".." for staleness */ #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move() during rename() internally. */ struct dentry *(*mount) (struct file_system_type *, int, diff --git a/include/linux/genhd.h b/include/linux/genhd.h index 4f440b3e89fe..79b8bba19363 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -88,10 +88,14 @@ struct disk_stats { }; #define PARTITION_META_INFO_VOLNAMELTH 64 -#define PARTITION_META_INFO_UUIDLTH 16 +/* + * Enough for the string representation of any kind of UUID plus NULL. + * EFI UUID is 36 characters. MSDOS UUID is 11 characters. + */ +#define PARTITION_META_INFO_UUIDLTH 37 struct partition_meta_info { - u8 uuid[PARTITION_META_INFO_UUIDLTH]; /* always big endian */ + char uuid[PARTITION_META_INFO_UUIDLTH]; u8 volname[PARTITION_META_INFO_VOLNAMELTH]; }; diff --git a/include/linux/genl_magic_func.h b/include/linux/genl_magic_func.h new file mode 100644 index 000000000000..023bc346b877 --- /dev/null +++ b/include/linux/genl_magic_func.h @@ -0,0 +1,422 @@ +#ifndef GENL_MAGIC_FUNC_H +#define GENL_MAGIC_FUNC_H + +#include <linux/genl_magic_struct.h> + +/* + * Magic: declare tla policy {{{1 + * Magic: declare nested policies + * {{{2 + */ +#undef GENL_mc_group +#define GENL_mc_group(group) + +#undef GENL_notification +#define GENL_notification(op_name, op_num, mcast_group, tla_list) + +#undef GENL_op +#define GENL_op(op_name, op_num, handler, tla_list) + +#undef GENL_struct +#define GENL_struct(tag_name, tag_number, s_name, s_fields) \ + [tag_name] = { .type = NLA_NESTED }, + +static struct nla_policy CONCAT_(GENL_MAGIC_FAMILY, _tla_nl_policy)[] = { +#include GENL_MAGIC_INCLUDE_FILE +}; + +#undef GENL_struct +#define GENL_struct(tag_name, tag_number, s_name, s_fields) \ +static struct nla_policy s_name ## _nl_policy[] __read_mostly = \ +{ s_fields }; + +#undef __field +#define __field(attr_nr, attr_flag, name, nla_type, _type, __get, \ + __put, __is_signed) \ + [attr_nr] = { .type = nla_type }, + +#undef __array +#define __array(attr_nr, attr_flag, name, nla_type, _type, maxlen, \ + __get, __put, __is_signed) \ + [attr_nr] = { .type = nla_type, \ + .len = maxlen - (nla_type == NLA_NUL_STRING) }, + +#include GENL_MAGIC_INCLUDE_FILE + +#ifndef __KERNEL__ +#ifndef pr_info +#define pr_info(args...) fprintf(stderr, args); +#endif +#endif + +#ifdef GENL_MAGIC_DEBUG +static void dprint_field(const char *dir, int nla_type, + const char *name, void *valp) +{ + __u64 val = valp ? *(__u32 *)valp : 1; + switch (nla_type) { + case NLA_U8: val = (__u8)val; + case NLA_U16: val = (__u16)val; + case NLA_U32: val = (__u32)val; + pr_info("%s attr %s: %d 0x%08x\n", dir, + name, (int)val, (unsigned)val); + break; + case NLA_U64: + val = *(__u64*)valp; + pr_info("%s attr %s: %lld 0x%08llx\n", dir, + name, (long long)val, (unsigned long long)val); + break; + case NLA_FLAG: + if (val) + pr_info("%s attr %s: set\n", dir, name); + break; + } +} + +static void dprint_array(const char *dir, int nla_type, + const char *name, const char *val, unsigned len) +{ + switch (nla_type) { + case NLA_NUL_STRING: + if (len && val[len-1] == '\0') + len--; + pr_info("%s attr %s: [len:%u] '%s'\n", dir, name, len, val); + break; + default: + /* we can always show 4 byte, + * thats what nlattr are aligned to. */ + pr_info("%s attr %s: [len:%u] %02x%02x%02x%02x ...\n", + dir, name, len, val[0], val[1], val[2], val[3]); + } +} + +#define DPRINT_TLA(a, op, b) pr_info("%s %s %s\n", a, op, b); + +/* Name is a member field name of the struct s. + * If s is NULL (only parsing, no copy requested in *_from_attrs()), + * nla is supposed to point to the attribute containing the information + * corresponding to that struct member. */ +#define DPRINT_FIELD(dir, nla_type, name, s, nla) \ + do { \ + if (s) \ + dprint_field(dir, nla_type, #name, &s->name); \ + else if (nla) \ + dprint_field(dir, nla_type, #name, \ + (nla_type == NLA_FLAG) ? NULL \ + : nla_data(nla)); \ + } while (0) + +#define DPRINT_ARRAY(dir, nla_type, name, s, nla) \ + do { \ + if (s) \ + dprint_array(dir, nla_type, #name, \ + s->name, s->name ## _len); \ + else if (nla) \ + dprint_array(dir, nla_type, #name, \ + nla_data(nla), nla_len(nla)); \ + } while (0) +#else +#define DPRINT_TLA(a, op, b) do {} while (0) +#define DPRINT_FIELD(dir, nla_type, name, s, nla) do {} while (0) +#define DPRINT_ARRAY(dir, nla_type, name, s, nla) do {} while (0) +#endif + +/* + * Magic: provide conversion functions {{{1 + * populate struct from attribute table: + * {{{2 + */ + +/* processing of generic netlink messages is serialized. + * use one static buffer for parsing of nested attributes */ +static struct nlattr *nested_attr_tb[128]; + +#ifndef BUILD_BUG_ON +/* Force a compilation error if condition is true */ +#define BUILD_BUG_ON(condition) ((void)BUILD_BUG_ON_ZERO(condition)) +/* Force a compilation error if condition is true, but also produce a + result (of value 0 and type size_t), so the expression can be used + e.g. in a structure initializer (or where-ever else comma expressions + aren't permitted). */ +#define BUILD_BUG_ON_ZERO(e) (sizeof(struct { int:-!!(e); })) +#define BUILD_BUG_ON_NULL(e) ((void *)sizeof(struct { int:-!!(e); })) +#endif + +#undef GENL_struct +#define GENL_struct(tag_name, tag_number, s_name, s_fields) \ +/* *_from_attrs functions are static, but potentially unused */ \ +static int __ ## s_name ## _from_attrs(struct s_name *s, \ + struct genl_info *info, bool exclude_invariants) \ +{ \ + const int maxtype = ARRAY_SIZE(s_name ## _nl_policy)-1; \ + struct nlattr *tla = info->attrs[tag_number]; \ + struct nlattr **ntb = nested_attr_tb; \ + struct nlattr *nla; \ + int err; \ + BUILD_BUG_ON(ARRAY_SIZE(s_name ## _nl_policy) > ARRAY_SIZE(nested_attr_tb)); \ + if (!tla) \ + return -ENOMSG; \ + DPRINT_TLA(#s_name, "<=-", #tag_name); \ + err = drbd_nla_parse_nested(ntb, maxtype, tla, s_name ## _nl_policy); \ + if (err) \ + return err; \ + \ + s_fields \ + return 0; \ +} __attribute__((unused)) \ +static int s_name ## _from_attrs(struct s_name *s, \ + struct genl_info *info) \ +{ \ + return __ ## s_name ## _from_attrs(s, info, false); \ +} __attribute__((unused)) \ +static int s_name ## _from_attrs_for_change(struct s_name *s, \ + struct genl_info *info) \ +{ \ + return __ ## s_name ## _from_attrs(s, info, true); \ +} __attribute__((unused)) \ + +#define __assign(attr_nr, attr_flag, name, nla_type, type, assignment...) \ + nla = ntb[attr_nr]; \ + if (nla) { \ + if (exclude_invariants && ((attr_flag) & DRBD_F_INVARIANT)) { \ + pr_info("<< must not change invariant attr: %s\n", #name); \ + return -EEXIST; \ + } \ + assignment; \ + } else if (exclude_invariants && ((attr_flag) & DRBD_F_INVARIANT)) { \ + /* attribute missing from payload, */ \ + /* which was expected */ \ + } else if ((attr_flag) & DRBD_F_REQUIRED) { \ + pr_info("<< missing attr: %s\n", #name); \ + return -ENOMSG; \ + } + +#undef __field +#define __field(attr_nr, attr_flag, name, nla_type, type, __get, __put, \ + __is_signed) \ + __assign(attr_nr, attr_flag, name, nla_type, type, \ + if (s) \ + s->name = __get(nla); \ + DPRINT_FIELD("<<", nla_type, name, s, nla)) + +/* validate_nla() already checked nla_len <= maxlen appropriately. */ +#undef __array +#define __array(attr_nr, attr_flag, name, nla_type, type, maxlen, \ + __get, __put, __is_signed) \ + __assign(attr_nr, attr_flag, name, nla_type, type, \ + if (s) \ + s->name ## _len = \ + __get(s->name, nla, maxlen); \ + DPRINT_ARRAY("<<", nla_type, name, s, nla)) + +#include GENL_MAGIC_INCLUDE_FILE + +#undef GENL_struct +#define GENL_struct(tag_name, tag_number, s_name, s_fields) + +/* + * Magic: define op number to op name mapping {{{1 + * {{{2 + */ +const char *CONCAT_(GENL_MAGIC_FAMILY, _genl_cmd_to_str)(__u8 cmd) +{ + switch (cmd) { +#undef GENL_op +#define GENL_op(op_name, op_num, handler, tla_list) \ + case op_num: return #op_name; +#include GENL_MAGIC_INCLUDE_FILE + default: + return "unknown"; + } +} + +#ifdef __KERNEL__ +#include <linux/stringify.h> +/* + * Magic: define genl_ops {{{1 + * {{{2 + */ + +#undef GENL_op +#define GENL_op(op_name, op_num, handler, tla_list) \ +{ \ + handler \ + .cmd = op_name, \ + .policy = CONCAT_(GENL_MAGIC_FAMILY, _tla_nl_policy), \ +}, + +#define ZZZ_genl_ops CONCAT_(GENL_MAGIC_FAMILY, _genl_ops) +static struct genl_ops ZZZ_genl_ops[] __read_mostly = { +#include GENL_MAGIC_INCLUDE_FILE +}; + +#undef GENL_op +#define GENL_op(op_name, op_num, handler, tla_list) + +/* + * Define the genl_family, multicast groups, {{{1 + * and provide register/unregister functions. + * {{{2 + */ +#define ZZZ_genl_family CONCAT_(GENL_MAGIC_FAMILY, _genl_family) +static struct genl_family ZZZ_genl_family __read_mostly = { + .id = GENL_ID_GENERATE, + .name = __stringify(GENL_MAGIC_FAMILY), + .version = GENL_MAGIC_VERSION, +#ifdef GENL_MAGIC_FAMILY_HDRSZ + .hdrsize = NLA_ALIGN(GENL_MAGIC_FAMILY_HDRSZ), +#endif + .maxattr = ARRAY_SIZE(drbd_tla_nl_policy)-1, +}; + +/* + * Magic: define multicast groups + * Magic: define multicast group registration helper + */ +#undef GENL_mc_group +#define GENL_mc_group(group) \ +static struct genl_multicast_group \ +CONCAT_(GENL_MAGIC_FAMILY, _mcg_ ## group) __read_mostly = { \ + .name = #group, \ +}; \ +static int CONCAT_(GENL_MAGIC_FAMILY, _genl_multicast_ ## group)( \ + struct sk_buff *skb, gfp_t flags) \ +{ \ + unsigned int group_id = \ + CONCAT_(GENL_MAGIC_FAMILY, _mcg_ ## group).id; \ + if (!group_id) \ + return -EINVAL; \ + return genlmsg_multicast(skb, 0, group_id, flags); \ +} + +#include GENL_MAGIC_INCLUDE_FILE + +int CONCAT_(GENL_MAGIC_FAMILY, _genl_register)(void) +{ + int err = genl_register_family_with_ops(&ZZZ_genl_family, + ZZZ_genl_ops, ARRAY_SIZE(ZZZ_genl_ops)); + if (err) + return err; +#undef GENL_mc_group +#define GENL_mc_group(group) \ + err = genl_register_mc_group(&ZZZ_genl_family, \ + &CONCAT_(GENL_MAGIC_FAMILY, _mcg_ ## group)); \ + if (err) \ + goto fail; \ + else \ + pr_info("%s: mcg %s: %u\n", #group, \ + __stringify(GENL_MAGIC_FAMILY), \ + CONCAT_(GENL_MAGIC_FAMILY, _mcg_ ## group).id); + +#include GENL_MAGIC_INCLUDE_FILE + +#undef GENL_mc_group +#define GENL_mc_group(group) + return 0; +fail: + genl_unregister_family(&ZZZ_genl_family); + return err; +} + +void CONCAT_(GENL_MAGIC_FAMILY, _genl_unregister)(void) +{ + genl_unregister_family(&ZZZ_genl_family); +} + +/* + * Magic: provide conversion functions {{{1 + * populate skb from struct. + * {{{2 + */ + +#undef GENL_op +#define GENL_op(op_name, op_num, handler, tla_list) + +#undef GENL_struct +#define GENL_struct(tag_name, tag_number, s_name, s_fields) \ +static int s_name ## _to_skb(struct sk_buff *skb, struct s_name *s, \ + const bool exclude_sensitive) \ +{ \ + struct nlattr *tla = nla_nest_start(skb, tag_number); \ + if (!tla) \ + goto nla_put_failure; \ + DPRINT_TLA(#s_name, "-=>", #tag_name); \ + s_fields \ + nla_nest_end(skb, tla); \ + return 0; \ + \ +nla_put_failure: \ + if (tla) \ + nla_nest_cancel(skb, tla); \ + return -EMSGSIZE; \ +} \ +static inline int s_name ## _to_priv_skb(struct sk_buff *skb, \ + struct s_name *s) \ +{ \ + return s_name ## _to_skb(skb, s, 0); \ +} \ +static inline int s_name ## _to_unpriv_skb(struct sk_buff *skb, \ + struct s_name *s) \ +{ \ + return s_name ## _to_skb(skb, s, 1); \ +} + + +#undef __field +#define __field(attr_nr, attr_flag, name, nla_type, type, __get, __put, \ + __is_signed) \ + if (!exclude_sensitive || !((attr_flag) & DRBD_F_SENSITIVE)) { \ + DPRINT_FIELD(">>", nla_type, name, s, NULL); \ + if (__put(skb, attr_nr, s->name)) \ + goto nla_put_failure; \ + } + +#undef __array +#define __array(attr_nr, attr_flag, name, nla_type, type, maxlen, \ + __get, __put, __is_signed) \ + if (!exclude_sensitive || !((attr_flag) & DRBD_F_SENSITIVE)) { \ + DPRINT_ARRAY(">>",nla_type, name, s, NULL); \ + if (__put(skb, attr_nr, min_t(int, maxlen, \ + s->name ## _len + (nla_type == NLA_NUL_STRING)),\ + s->name)) \ + goto nla_put_failure; \ + } + +#include GENL_MAGIC_INCLUDE_FILE + + +/* Functions for initializing structs to default values. */ + +#undef __field +#define __field(attr_nr, attr_flag, name, nla_type, type, __get, __put, \ + __is_signed) +#undef __array +#define __array(attr_nr, attr_flag, name, nla_type, type, maxlen, \ + __get, __put, __is_signed) +#undef __u32_field_def +#define __u32_field_def(attr_nr, attr_flag, name, default) \ + x->name = default; +#undef __s32_field_def +#define __s32_field_def(attr_nr, attr_flag, name, default) \ + x->name = default; +#undef __flg_field_def +#define __flg_field_def(attr_nr, attr_flag, name, default) \ + x->name = default; +#undef __str_field_def +#define __str_field_def(attr_nr, attr_flag, name, maxlen) \ + memset(x->name, 0, sizeof(x->name)); \ + x->name ## _len = 0; +#undef GENL_struct +#define GENL_struct(tag_name, tag_number, s_name, s_fields) \ +static void set_ ## s_name ## _defaults(struct s_name *x) __attribute__((unused)); \ +static void set_ ## s_name ## _defaults(struct s_name *x) { \ +s_fields \ +} + +#include GENL_MAGIC_INCLUDE_FILE + +#endif /* __KERNEL__ */ + +/* }}}1 */ +#endif /* GENL_MAGIC_FUNC_H */ +/* vim: set foldmethod=marker foldlevel=1 nofoldenable : */ diff --git a/include/linux/genl_magic_struct.h b/include/linux/genl_magic_struct.h new file mode 100644 index 000000000000..eecd19b37001 --- /dev/null +++ b/include/linux/genl_magic_struct.h @@ -0,0 +1,277 @@ +#ifndef GENL_MAGIC_STRUCT_H +#define GENL_MAGIC_STRUCT_H + +#ifndef GENL_MAGIC_FAMILY +# error "you need to define GENL_MAGIC_FAMILY before inclusion" +#endif + +#ifndef GENL_MAGIC_VERSION +# error "you need to define GENL_MAGIC_VERSION before inclusion" +#endif + +#ifndef GENL_MAGIC_INCLUDE_FILE +# error "you need to define GENL_MAGIC_INCLUDE_FILE before inclusion" +#endif + +#include <linux/genetlink.h> +#include <linux/types.h> + +#define CONCAT__(a,b) a ## b +#define CONCAT_(a,b) CONCAT__(a,b) + +extern int CONCAT_(GENL_MAGIC_FAMILY, _genl_register)(void); +extern void CONCAT_(GENL_MAGIC_FAMILY, _genl_unregister)(void); + +/* + * Extension of genl attribute validation policies {{{2 + */ + +/* + * @DRBD_GENLA_F_MANDATORY: By default, netlink ignores attributes it does not + * know about. This flag can be set in nlattr->nla_type to indicate that this + * attribute must not be ignored. + * + * We check and remove this flag in drbd_nla_check_mandatory() before + * validating the attribute types and lengths via nla_parse_nested(). + */ +#define DRBD_GENLA_F_MANDATORY (1 << 14) + +/* + * Flags specific to drbd and not visible at the netlink layer, used in + * <struct>_from_attrs and <struct>_to_skb: + * + * @DRBD_F_REQUIRED: Attribute is required; a request without this attribute is + * invalid. + * + * @DRBD_F_SENSITIVE: Attribute includes sensitive information and must not be + * included in unpriviledged get requests or broadcasts. + * + * @DRBD_F_INVARIANT: Attribute is set when an object is initially created, but + * cannot subsequently be changed. + */ +#define DRBD_F_REQUIRED (1 << 0) +#define DRBD_F_SENSITIVE (1 << 1) +#define DRBD_F_INVARIANT (1 << 2) + +#define __nla_type(x) ((__u16)((x) & NLA_TYPE_MASK & ~DRBD_GENLA_F_MANDATORY)) + +/* }}}1 + * MAGIC + * multi-include macro expansion magic starts here + */ + +/* MAGIC helpers {{{2 */ + +/* possible field types */ +#define __flg_field(attr_nr, attr_flag, name) \ + __field(attr_nr, attr_flag, name, NLA_U8, char, \ + nla_get_u8, nla_put_u8, false) +#define __u8_field(attr_nr, attr_flag, name) \ + __field(attr_nr, attr_flag, name, NLA_U8, unsigned char, \ + nla_get_u8, nla_put_u8, false) +#define __u16_field(attr_nr, attr_flag, name) \ + __field(attr_nr, attr_flag, name, NLA_U16, __u16, \ + nla_get_u16, nla_put_u16, false) +#define __u32_field(attr_nr, attr_flag, name) \ + __field(attr_nr, attr_flag, name, NLA_U32, __u32, \ + nla_get_u32, nla_put_u32, false) +#define __s32_field(attr_nr, attr_flag, name) \ + __field(attr_nr, attr_flag, name, NLA_U32, __s32, \ + nla_get_u32, nla_put_u32, true) +#define __u64_field(attr_nr, attr_flag, name) \ + __field(attr_nr, attr_flag, name, NLA_U64, __u64, \ + nla_get_u64, nla_put_u64, false) +#define __str_field(attr_nr, attr_flag, name, maxlen) \ + __array(attr_nr, attr_flag, name, NLA_NUL_STRING, char, maxlen, \ + nla_strlcpy, nla_put, false) +#define __bin_field(attr_nr, attr_flag, name, maxlen) \ + __array(attr_nr, attr_flag, name, NLA_BINARY, char, maxlen, \ + nla_memcpy, nla_put, false) + +/* fields with default values */ +#define __flg_field_def(attr_nr, attr_flag, name, default) \ + __flg_field(attr_nr, attr_flag, name) +#define __u32_field_def(attr_nr, attr_flag, name, default) \ + __u32_field(attr_nr, attr_flag, name) +#define __s32_field_def(attr_nr, attr_flag, name, default) \ + __s32_field(attr_nr, attr_flag, name) +#define __str_field_def(attr_nr, attr_flag, name, maxlen) \ + __str_field(attr_nr, attr_flag, name, maxlen) + +#define GENL_op_init(args...) args +#define GENL_doit(handler) \ + .doit = handler, \ + .flags = GENL_ADMIN_PERM, +#define GENL_dumpit(handler) \ + .dumpit = handler, \ + .flags = GENL_ADMIN_PERM, + +/* }}}1 + * Magic: define the enum symbols for genl_ops + * Magic: define the enum symbols for top level attributes + * Magic: define the enum symbols for nested attributes + * {{{2 + */ + +#undef GENL_struct +#define GENL_struct(tag_name, tag_number, s_name, s_fields) + +#undef GENL_mc_group +#define GENL_mc_group(group) + +#undef GENL_notification +#define GENL_notification(op_name, op_num, mcast_group, tla_list) \ + op_name = op_num, + +#undef GENL_op +#define GENL_op(op_name, op_num, handler, tla_list) \ + op_name = op_num, + +enum { +#include GENL_MAGIC_INCLUDE_FILE +}; + +#undef GENL_notification +#define GENL_notification(op_name, op_num, mcast_group, tla_list) + +#undef GENL_op +#define GENL_op(op_name, op_num, handler, attr_list) + +#undef GENL_struct +#define GENL_struct(tag_name, tag_number, s_name, s_fields) \ + tag_name = tag_number, + +enum { +#include GENL_MAGIC_INCLUDE_FILE +}; + +#undef GENL_struct +#define GENL_struct(tag_name, tag_number, s_name, s_fields) \ +enum { \ + s_fields \ +}; + +#undef __field +#define __field(attr_nr, attr_flag, name, nla_type, type, \ + __get, __put, __is_signed) \ + T_ ## name = (__u16)(attr_nr | ((attr_flag) & DRBD_GENLA_F_MANDATORY)), + +#undef __array +#define __array(attr_nr, attr_flag, name, nla_type, type, \ + maxlen, __get, __put, __is_signed) \ + T_ ## name = (__u16)(attr_nr | ((attr_flag) & DRBD_GENLA_F_MANDATORY)), + +#include GENL_MAGIC_INCLUDE_FILE + +/* }}}1 + * Magic: compile time assert unique numbers for operations + * Magic: -"- unique numbers for top level attributes + * Magic: -"- unique numbers for nested attributes + * {{{2 + */ + +#undef GENL_struct +#define GENL_struct(tag_name, tag_number, s_name, s_fields) + +#undef GENL_op +#define GENL_op(op_name, op_num, handler, attr_list) \ + case op_name: + +#undef GENL_notification +#define GENL_notification(op_name, op_num, mcast_group, tla_list) \ + case op_name: + +static inline void ct_assert_unique_operations(void) +{ + switch (0) { +#include GENL_MAGIC_INCLUDE_FILE + ; + } +} + +#undef GENL_op +#define GENL_op(op_name, op_num, handler, attr_list) + +#undef GENL_notification +#define GENL_notification(op_name, op_num, mcast_group, tla_list) + +#undef GENL_struct +#define GENL_struct(tag_name, tag_number, s_name, s_fields) \ + case tag_number: + +static inline void ct_assert_unique_top_level_attributes(void) +{ + switch (0) { +#include GENL_MAGIC_INCLUDE_FILE + ; + } +} + +#undef GENL_struct +#define GENL_struct(tag_name, tag_number, s_name, s_fields) \ +static inline void ct_assert_unique_ ## s_name ## _attributes(void) \ +{ \ + switch (0) { \ + s_fields \ + ; \ + } \ +} + +#undef __field +#define __field(attr_nr, attr_flag, name, nla_type, type, __get, __put, \ + __is_signed) \ + case attr_nr: + +#undef __array +#define __array(attr_nr, attr_flag, name, nla_type, type, maxlen, \ + __get, __put, __is_signed) \ + case attr_nr: + +#include GENL_MAGIC_INCLUDE_FILE + +/* }}}1 + * Magic: declare structs + * struct <name> { + * fields + * }; + * {{{2 + */ + +#undef GENL_struct +#define GENL_struct(tag_name, tag_number, s_name, s_fields) \ +struct s_name { s_fields }; + +#undef __field +#define __field(attr_nr, attr_flag, name, nla_type, type, __get, __put, \ + __is_signed) \ + type name; + +#undef __array +#define __array(attr_nr, attr_flag, name, nla_type, type, maxlen, \ + __get, __put, __is_signed) \ + type name[maxlen]; \ + __u32 name ## _len; + +#include GENL_MAGIC_INCLUDE_FILE + +#undef GENL_struct +#define GENL_struct(tag_name, tag_number, s_name, s_fields) \ +enum { \ + s_fields \ +}; + +#undef __field +#define __field(attr_nr, attr_flag, name, nla_type, type, __get, __put, \ + is_signed) \ + F_ ## name ## _IS_SIGNED = is_signed, + +#undef __array +#define __array(attr_nr, attr_flag, name, nla_type, type, maxlen, \ + __get, __put, is_signed) \ + F_ ## name ## _IS_SIGNED = is_signed, + +#include GENL_MAGIC_INCLUDE_FILE + +/* }}}1 */ +#endif /* GENL_MAGIC_STRUCT_H */ +/* vim: set foldmethod=marker nofoldenable : */ diff --git a/include/linux/idr.h b/include/linux/idr.h index 87259a44c251..de7e190f1af4 100644 --- a/include/linux/idr.h +++ b/include/linux/idr.h @@ -152,4 +152,15 @@ void ida_simple_remove(struct ida *ida, unsigned int id); void __init idr_init_cache(void); +/** + * idr_for_each_entry - iterate over an idr's elements of a given type + * @idp: idr handle + * @entry: the type * to use as cursor + * @id: id entry's key + */ +#define idr_for_each_entry(idp, entry, id) \ + for (id = 0, entry = (typeof(entry))idr_get_next((idp), &(id)); \ + entry != NULL; \ + ++id, entry = (typeof(entry))idr_get_next((idp), &(id))) + #endif /* __IDR_H__ */ diff --git a/include/linux/ipc_namespace.h b/include/linux/ipc_namespace.h index 5499c92a9153..fe771978e877 100644 --- a/include/linux/ipc_namespace.h +++ b/include/linux/ipc_namespace.h @@ -67,6 +67,8 @@ struct ipc_namespace { /* user_ns which owns the ipc ns */ struct user_namespace *user_ns; + + unsigned int proc_inum; }; extern struct ipc_namespace init_ipc_ns; @@ -133,7 +135,8 @@ static inline int mq_init_ns(struct ipc_namespace *ns) { return 0; } #if defined(CONFIG_IPC_NS) extern struct ipc_namespace *copy_ipcs(unsigned long flags, - struct task_struct *tsk); + struct user_namespace *user_ns, struct ipc_namespace *ns); + static inline struct ipc_namespace *get_ipc_ns(struct ipc_namespace *ns) { if (ns) @@ -144,12 +147,12 @@ static inline struct ipc_namespace *get_ipc_ns(struct ipc_namespace *ns) extern void put_ipc_ns(struct ipc_namespace *ns); #else static inline struct ipc_namespace *copy_ipcs(unsigned long flags, - struct task_struct *tsk) + struct user_namespace *user_ns, struct ipc_namespace *ns) { if (flags & CLONE_NEWIPC) return ERR_PTR(-EINVAL); - return tsk->nsproxy->ipc_ns; + return ns; } static inline struct ipc_namespace *get_ipc_ns(struct ipc_namespace *ns) diff --git a/include/linux/loop.h b/include/linux/loop.h index 6492181bcb1d..460b60fa7adf 100644 --- a/include/linux/loop.h +++ b/include/linux/loop.h @@ -53,10 +53,13 @@ struct loop_device { spinlock_t lo_lock; struct bio_list lo_bio_list; + unsigned int lo_bio_count; int lo_state; struct mutex lo_ctl_mutex; struct task_struct *lo_thread; wait_queue_head_t lo_event; + /* wait queue for incoming requests */ + wait_queue_head_t lo_req_wait; struct request_queue *lo_queue; struct gendisk *lo_disk; diff --git a/include/linux/lru_cache.h b/include/linux/lru_cache.h index cafc7f99e124..4019013c6593 100644 --- a/include/linux/lru_cache.h +++ b/include/linux/lru_cache.h @@ -166,9 +166,11 @@ struct lc_element { /* if we want to track a larger set of objects, * it needs to become arch independend u64 */ unsigned lc_number; - /* special label when on free list */ #define LC_FREE (~0U) + + /* for pending changes */ + unsigned lc_new_number; }; struct lru_cache { @@ -176,6 +178,7 @@ struct lru_cache { struct list_head lru; struct list_head free; struct list_head in_use; + struct list_head to_be_changed; /* the pre-created kmem cache to allocate the objects from */ struct kmem_cache *lc_cache; @@ -186,7 +189,7 @@ struct lru_cache { size_t element_off; /* number of elements (indices) */ - unsigned int nr_elements; + unsigned int nr_elements; /* Arbitrary limit on maximum tracked objects. Practical limit is much * lower due to allocation failures, probably. For typical use cases, * nr_elements should be a few thousand at most. @@ -194,18 +197,19 @@ struct lru_cache { * 8 high bits of .lc_index to be overloaded with flags in the future. */ #define LC_MAX_ACTIVE (1<<24) + /* allow to accumulate a few (index:label) changes, + * but no more than max_pending_changes */ + unsigned int max_pending_changes; + /* number of elements currently on to_be_changed list */ + unsigned int pending_changes; + /* statistics */ - unsigned used; /* number of lelements currently on in_use list */ - unsigned long hits, misses, starving, dirty, changed; + unsigned used; /* number of elements currently on in_use list */ + unsigned long hits, misses, starving, locked, changed; /* see below: flag-bits for lru_cache */ unsigned long flags; - /* when changing the label of an index element */ - unsigned int new_number; - - /* for paranoia when changing the label of an index element */ - struct lc_element *changing_element; void *lc_private; const char *name; @@ -221,10 +225,15 @@ enum { /* debugging aid, to catch concurrent access early. * user needs to guarantee exclusive access by proper locking! */ __LC_PARANOIA, - /* if we need to change the set, but currently there is a changing - * transaction pending, we are "dirty", and must deferr further - * changing requests */ + + /* annotate that the set is "dirty", possibly accumulating further + * changes, until a transaction is finally triggered */ __LC_DIRTY, + + /* Locked, no further changes allowed. + * Also used to serialize changing transactions. */ + __LC_LOCKED, + /* if we need to change the set, but currently there is no free nor * unused element available, we are "starving", and must not give out * further references, to guarantee that eventually some refcnt will @@ -236,9 +245,11 @@ enum { }; #define LC_PARANOIA (1<<__LC_PARANOIA) #define LC_DIRTY (1<<__LC_DIRTY) +#define LC_LOCKED (1<<__LC_LOCKED) #define LC_STARVING (1<<__LC_STARVING) extern struct lru_cache *lc_create(const char *name, struct kmem_cache *cache, + unsigned max_pending_changes, unsigned e_count, size_t e_size, size_t e_off); extern void lc_reset(struct lru_cache *lc); extern void lc_destroy(struct lru_cache *lc); @@ -249,7 +260,7 @@ extern struct lc_element *lc_try_get(struct lru_cache *lc, unsigned int enr); extern struct lc_element *lc_find(struct lru_cache *lc, unsigned int enr); extern struct lc_element *lc_get(struct lru_cache *lc, unsigned int enr); extern unsigned int lc_put(struct lru_cache *lc, struct lc_element *e); -extern void lc_changed(struct lru_cache *lc, struct lc_element *e); +extern void lc_committed(struct lru_cache *lc); struct seq_file; extern size_t lc_seq_printf_stats(struct seq_file *seq, struct lru_cache *lc); @@ -258,32 +269,40 @@ extern void lc_seq_dump_details(struct seq_file *seq, struct lru_cache *lc, char void (*detail) (struct seq_file *, struct lc_element *)); /** - * lc_try_lock - can be used to stop lc_get() from changing the tracked set + * lc_try_lock_for_transaction - can be used to stop lc_get() from changing the tracked set * @lc: the lru cache to operate on * - * Note that the reference counts and order on the active and lru lists may - * still change. Returns true if we acquired the lock. + * Allows (expects) the set to be "dirty". Note that the reference counts and + * order on the active and lru lists may still change. Used to serialize + * changing transactions. Returns true if we aquired the lock. */ -static inline int lc_try_lock(struct lru_cache *lc) +static inline int lc_try_lock_for_transaction(struct lru_cache *lc) { - return !test_and_set_bit(__LC_DIRTY, &lc->flags); + return !test_and_set_bit(__LC_LOCKED, &lc->flags); } /** + * lc_try_lock - variant to stop lc_get() from changing the tracked set + * @lc: the lru cache to operate on + * + * Note that the reference counts and order on the active and lru lists may + * still change. Only works on a "clean" set. Returns true if we aquired the + * lock, which means there are no pending changes, and any further attempt to + * change the set will not succeed until the next lc_unlock(). + */ +extern int lc_try_lock(struct lru_cache *lc); + +/** * lc_unlock - unlock @lc, allow lc_get() to change the set again * @lc: the lru cache to operate on */ static inline void lc_unlock(struct lru_cache *lc) { clear_bit(__LC_DIRTY, &lc->flags); - smp_mb__after_clear_bit(); + clear_bit_unlock(__LC_LOCKED, &lc->flags); } -static inline int lc_is_used(struct lru_cache *lc, unsigned int enr) -{ - struct lc_element *e = lc_find(lc, enr); - return e && e->refcnt; -} +extern bool lc_is_used(struct lru_cache *lc, unsigned int enr); #define lc_entry(ptr, type, member) \ container_of(ptr, type, member) diff --git a/include/linux/mnt_namespace.h b/include/linux/mnt_namespace.h index 5a8e3903d770..12b2ab510323 100644 --- a/include/linux/mnt_namespace.h +++ b/include/linux/mnt_namespace.h @@ -4,9 +4,10 @@ struct mnt_namespace; struct fs_struct; +struct user_namespace; extern struct mnt_namespace *copy_mnt_ns(unsigned long, struct mnt_namespace *, - struct fs_struct *); + struct user_namespace *, struct fs_struct *); extern void put_mnt_ns(struct mnt_namespace *ns); extern const struct file_operations proc_mounts_operations; diff --git a/include/linux/nsproxy.h b/include/linux/nsproxy.h index cc37a55ad004..10e5947491c7 100644 --- a/include/linux/nsproxy.h +++ b/include/linux/nsproxy.h @@ -67,7 +67,7 @@ void exit_task_namespaces(struct task_struct *tsk); void switch_task_namespaces(struct task_struct *tsk, struct nsproxy *new); void free_nsproxy(struct nsproxy *ns); int unshare_nsproxy_namespaces(unsigned long, struct nsproxy **, - struct fs_struct *); + struct cred *, struct fs_struct *); int __init nsproxy_cache_init(void); static inline void put_nsproxy(struct nsproxy *ns) diff --git a/include/linux/pid_namespace.h b/include/linux/pid_namespace.h index 65e3e87eacc5..bf285999273a 100644 --- a/include/linux/pid_namespace.h +++ b/include/linux/pid_namespace.h @@ -21,6 +21,7 @@ struct pid_namespace { struct kref kref; struct pidmap pidmap[PIDMAP_ENTRIES]; int last_pid; + int nr_hashed; struct task_struct *child_reaper; struct kmem_cache *pid_cachep; unsigned int level; @@ -31,9 +32,12 @@ struct pid_namespace { #ifdef CONFIG_BSD_PROCESS_ACCT struct bsd_acct_struct *bacct; #endif + struct user_namespace *user_ns; + struct work_struct proc_work; kgid_t pid_gid; int hide_pid; int reboot; /* group exit code if this pidns was rebooted */ + unsigned int proc_inum; }; extern struct pid_namespace init_pid_ns; @@ -46,7 +50,8 @@ static inline struct pid_namespace *get_pid_ns(struct pid_namespace *ns) return ns; } -extern struct pid_namespace *copy_pid_ns(unsigned long flags, struct pid_namespace *ns); +extern struct pid_namespace *copy_pid_ns(unsigned long flags, + struct user_namespace *user_ns, struct pid_namespace *ns); extern void zap_pid_ns_processes(struct pid_namespace *pid_ns); extern int reboot_pid_ns(struct pid_namespace *pid_ns, int cmd); extern void put_pid_ns(struct pid_namespace *ns); @@ -59,8 +64,8 @@ static inline struct pid_namespace *get_pid_ns(struct pid_namespace *ns) return ns; } -static inline struct pid_namespace * -copy_pid_ns(unsigned long flags, struct pid_namespace *ns) +static inline struct pid_namespace *copy_pid_ns(unsigned long flags, + struct user_namespace *user_ns, struct pid_namespace *ns) { if (flags & CLONE_NEWPID) ns = ERR_PTR(-EINVAL); diff --git a/include/linux/proc_fs.h b/include/linux/proc_fs.h index b4f70f0a9a48..32676b35d2f5 100644 --- a/include/linux/proc_fs.h +++ b/include/linux/proc_fs.h @@ -28,7 +28,11 @@ struct mm_struct; */ enum { - PROC_ROOT_INO = 1, + PROC_ROOT_INO = 1, + PROC_IPC_INIT_INO = 0xEFFFFFFFU, + PROC_UTS_INIT_INO = 0xEFFFFFFEU, + PROC_USER_INIT_INO = 0xEFFFFFFDU, + PROC_PID_INIT_INO = 0xEFFFFFFCU, }; /* @@ -174,7 +178,10 @@ extern struct proc_dir_entry *proc_net_mkdir(struct net *net, const char *name, struct proc_dir_entry *parent); extern struct file *proc_ns_fget(int fd); +extern bool proc_ns_inode(struct inode *inode); +extern int proc_alloc_inum(unsigned int *pino); +extern void proc_free_inum(unsigned int inum); #else #define proc_net_fops_create(net, name, mode, fops) ({ (void)(mode), NULL; }) @@ -229,6 +236,19 @@ static inline struct file *proc_ns_fget(int fd) return ERR_PTR(-EINVAL); } +static inline bool proc_ns_inode(struct inode *inode) +{ + return false; +} + +static inline int proc_alloc_inum(unsigned int *inum) +{ + *inum = 1; + return 0; +} +static inline void proc_free_inum(unsigned int inum) +{ +} #endif /* CONFIG_PROC_FS */ #if !defined(CONFIG_PROC_KCORE) @@ -247,10 +267,14 @@ struct proc_ns_operations { void *(*get)(struct task_struct *task); void (*put)(void *ns); int (*install)(struct nsproxy *nsproxy, void *ns); + unsigned int (*inum)(void *ns); }; extern const struct proc_ns_operations netns_operations; extern const struct proc_ns_operations utsns_operations; extern const struct proc_ns_operations ipcns_operations; +extern const struct proc_ns_operations pidns_operations; +extern const struct proc_ns_operations userns_operations; +extern const struct proc_ns_operations mntns_operations; union proc_op { int (*proc_get_link)(struct dentry *, struct path *); diff --git a/include/linux/user_namespace.h b/include/linux/user_namespace.h index 95142cae446a..b9bd2e6c73cc 100644 --- a/include/linux/user_namespace.h +++ b/include/linux/user_namespace.h @@ -25,6 +25,7 @@ struct user_namespace { struct user_namespace *parent; kuid_t owner; kgid_t group; + unsigned int proc_inum; }; extern struct user_namespace init_user_ns; @@ -39,6 +40,7 @@ static inline struct user_namespace *get_user_ns(struct user_namespace *ns) } extern int create_user_ns(struct cred *new); +extern int unshare_userns(unsigned long unshare_flags, struct cred **new_cred); extern void free_user_ns(struct kref *kref); static inline void put_user_ns(struct user_namespace *ns) @@ -66,6 +68,14 @@ static inline int create_user_ns(struct cred *new) return -EINVAL; } +static inline int unshare_userns(unsigned long unshare_flags, + struct cred **new_cred) +{ + if (unshare_flags & CLONE_NEWUSER) + return -EINVAL; + return 0; +} + static inline void put_user_ns(struct user_namespace *ns) { } diff --git a/include/linux/utsname.h b/include/linux/utsname.h index 2b345206722a..239e27733d6c 100644 --- a/include/linux/utsname.h +++ b/include/linux/utsname.h @@ -23,6 +23,7 @@ struct uts_namespace { struct kref kref; struct new_utsname name; struct user_namespace *user_ns; + unsigned int proc_inum; }; extern struct uts_namespace init_uts_ns; @@ -33,7 +34,7 @@ static inline void get_uts_ns(struct uts_namespace *ns) } extern struct uts_namespace *copy_utsname(unsigned long flags, - struct task_struct *tsk); + struct user_namespace *user_ns, struct uts_namespace *old_ns); extern void free_uts_ns(struct kref *kref); static inline void put_uts_ns(struct uts_namespace *ns) @@ -50,12 +51,12 @@ static inline void put_uts_ns(struct uts_namespace *ns) } static inline struct uts_namespace *copy_utsname(unsigned long flags, - struct task_struct *tsk) + struct user_namespace *user_ns, struct uts_namespace *old_ns) { if (flags & CLONE_NEWUTS) return ERR_PTR(-EINVAL); - return tsk->nsproxy->uts_ns; + return old_ns; } #endif diff --git a/include/linux/wait.h b/include/linux/wait.h index 168dfe122dd3..7cb64d4b499d 100644 --- a/include/linux/wait.h +++ b/include/linux/wait.h @@ -550,6 +550,170 @@ do { \ __ret; \ }) + +#define __wait_event_lock_irq(wq, condition, lock, cmd) \ +do { \ + DEFINE_WAIT(__wait); \ + \ + for (;;) { \ + prepare_to_wait(&wq, &__wait, TASK_UNINTERRUPTIBLE); \ + if (condition) \ + break; \ + spin_unlock_irq(&lock); \ + cmd; \ + schedule(); \ + spin_lock_irq(&lock); \ + } \ + finish_wait(&wq, &__wait); \ +} while (0) + +/** + * wait_event_lock_irq_cmd - sleep until a condition gets true. The + * condition is checked under the lock. This + * is expected to be called with the lock + * taken. + * @wq: the waitqueue to wait on + * @condition: a C expression for the event to wait for + * @lock: a locked spinlock_t, which will be released before cmd + * and schedule() and reacquired afterwards. + * @cmd: a command which is invoked outside the critical section before + * sleep + * + * The process is put to sleep (TASK_UNINTERRUPTIBLE) until the + * @condition evaluates to true. The @condition is checked each time + * the waitqueue @wq is woken up. + * + * wake_up() has to be called after changing any variable that could + * change the result of the wait condition. + * + * This is supposed to be called while holding the lock. The lock is + * dropped before invoking the cmd and going to sleep and is reacquired + * afterwards. + */ +#define wait_event_lock_irq_cmd(wq, condition, lock, cmd) \ +do { \ + if (condition) \ + break; \ + __wait_event_lock_irq(wq, condition, lock, cmd); \ +} while (0) + +/** + * wait_event_lock_irq - sleep until a condition gets true. The + * condition is checked under the lock. This + * is expected to be called with the lock + * taken. + * @wq: the waitqueue to wait on + * @condition: a C expression for the event to wait for + * @lock: a locked spinlock_t, which will be released before schedule() + * and reacquired afterwards. + * + * The process is put to sleep (TASK_UNINTERRUPTIBLE) until the + * @condition evaluates to true. The @condition is checked each time + * the waitqueue @wq is woken up. + * + * wake_up() has to be called after changing any variable that could + * change the result of the wait condition. + * + * This is supposed to be called while holding the lock. The lock is + * dropped before going to sleep and is reacquired afterwards. + */ +#define wait_event_lock_irq(wq, condition, lock) \ +do { \ + if (condition) \ + break; \ + __wait_event_lock_irq(wq, condition, lock, ); \ +} while (0) + + +#define __wait_event_interruptible_lock_irq(wq, condition, \ + lock, ret, cmd) \ +do { \ + DEFINE_WAIT(__wait); \ + \ + for (;;) { \ + prepare_to_wait(&wq, &__wait, TASK_INTERRUPTIBLE); \ + if (condition) \ + break; \ + if (signal_pending(current)) { \ + ret = -ERESTARTSYS; \ + break; \ + } \ + spin_unlock_irq(&lock); \ + cmd; \ + schedule(); \ + spin_lock_irq(&lock); \ + } \ + finish_wait(&wq, &__wait); \ +} while (0) + +/** + * wait_event_interruptible_lock_irq_cmd - sleep until a condition gets true. + * The condition is checked under the lock. This is expected to + * be called with the lock taken. + * @wq: the waitqueue to wait on + * @condition: a C expression for the event to wait for + * @lock: a locked spinlock_t, which will be released before cmd and + * schedule() and reacquired afterwards. + * @cmd: a command which is invoked outside the critical section before + * sleep + * + * The process is put to sleep (TASK_INTERRUPTIBLE) until the + * @condition evaluates to true or a signal is received. The @condition is + * checked each time the waitqueue @wq is woken up. + * + * wake_up() has to be called after changing any variable that could + * change the result of the wait condition. + * + * This is supposed to be called while holding the lock. The lock is + * dropped before invoking the cmd and going to sleep and is reacquired + * afterwards. + * + * The macro will return -ERESTARTSYS if it was interrupted by a signal + * and 0 if @condition evaluated to true. + */ +#define wait_event_interruptible_lock_irq_cmd(wq, condition, lock, cmd) \ +({ \ + int __ret = 0; \ + \ + if (!(condition)) \ + __wait_event_interruptible_lock_irq(wq, condition, \ + lock, __ret, cmd); \ + __ret; \ +}) + +/** + * wait_event_interruptible_lock_irq - sleep until a condition gets true. + * The condition is checked under the lock. This is expected + * to be called with the lock taken. + * @wq: the waitqueue to wait on + * @condition: a C expression for the event to wait for + * @lock: a locked spinlock_t, which will be released before schedule() + * and reacquired afterwards. + * + * The process is put to sleep (TASK_INTERRUPTIBLE) until the + * @condition evaluates to true or signal is received. The @condition is + * checked each time the waitqueue @wq is woken up. + * + * wake_up() has to be called after changing any variable that could + * change the result of the wait condition. + * + * This is supposed to be called while holding the lock. The lock is + * dropped before going to sleep and is reacquired afterwards. + * + * The macro will return -ERESTARTSYS if it was interrupted by a signal + * and 0 if @condition evaluated to true. + */ +#define wait_event_interruptible_lock_irq(wq, condition, lock) \ +({ \ + int __ret = 0; \ + \ + if (!(condition)) \ + __wait_event_interruptible_lock_irq(wq, condition, \ + lock, __ret, ); \ + __ret; \ +}) + + /* * These are the old interfaces to sleep waiting for an event. * They are racy. DO NOT use them, use the wait_event* interfaces above. diff --git a/include/net/net_namespace.h b/include/net/net_namespace.h index c5a43f56b796..de644bcd8613 100644 --- a/include/net/net_namespace.h +++ b/include/net/net_namespace.h @@ -56,6 +56,8 @@ struct net { struct user_namespace *user_ns; /* Owning user namespace */ + unsigned int proc_inum; + struct proc_dir_entry *proc_net; struct proc_dir_entry *proc_net_stat; diff --git a/init/Kconfig b/init/Kconfig index 1a207efca591..675d8a2326cf 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -1069,11 +1069,9 @@ config UIDGID_CONVERTED # Filesystems depends on 9P_FS = n depends on AFS_FS = n - depends on AUTOFS4_FS = n depends on CEPH_FS = n depends on CIFS = n depends on CODA_FS = n - depends on FUSE_FS = n depends on GFS2_FS = n depends on NCP_FS = n depends on NFSD = n diff --git a/init/do_mounts.c b/init/do_mounts.c index f8a66424360d..1d1b6348f903 100644 --- a/init/do_mounts.c +++ b/init/do_mounts.c @@ -69,23 +69,28 @@ __setup("ro", readonly); __setup("rw", readwrite); #ifdef CONFIG_BLOCK +struct uuidcmp { + const char *uuid; + int len; +}; + /** * match_dev_by_uuid - callback for finding a partition using its uuid * @dev: device passed in by the caller - * @data: opaque pointer to a 36 byte char array with a UUID + * @data: opaque pointer to the desired struct uuidcmp to match * * Returns 1 if the device matches, and 0 otherwise. */ static int match_dev_by_uuid(struct device *dev, void *data) { - u8 *uuid = data; + struct uuidcmp *cmp = data; struct hd_struct *part = dev_to_part(dev); if (!part->info) goto no_match; - if (memcmp(uuid, part->info->uuid, sizeof(part->info->uuid))) - goto no_match; + if (strncasecmp(cmp->uuid, part->info->uuid, cmp->len)) + goto no_match; return 1; no_match: @@ -95,7 +100,7 @@ no_match: /** * devt_from_partuuid - looks up the dev_t of a partition by its UUID - * @uuid: min 36 byte char array containing a hex ascii UUID + * @uuid: char array containing ascii UUID * * The function will return the first partition which contains a matching * UUID value in its partition_meta_info struct. This does not search @@ -106,38 +111,41 @@ no_match: * * Returns the matching dev_t on success or 0 on failure. */ -static dev_t devt_from_partuuid(char *uuid_str) +static dev_t devt_from_partuuid(const char *uuid_str) { dev_t res = 0; + struct uuidcmp cmp; struct device *dev = NULL; - u8 uuid[16]; struct gendisk *disk; struct hd_struct *part; int offset = 0; + bool clear_root_wait = false; + char *slash; - if (strlen(uuid_str) < 36) - goto done; + cmp.uuid = uuid_str; + slash = strchr(uuid_str, '/'); /* Check for optional partition number offset attributes. */ - if (uuid_str[36]) { + if (slash) { char c = 0; /* Explicitly fail on poor PARTUUID syntax. */ - if (sscanf(&uuid_str[36], - "/PARTNROFF=%d%c", &offset, &c) != 1) { - printk(KERN_ERR "VFS: PARTUUID= is invalid.\n" - "Expected PARTUUID=<valid-uuid-id>[/PARTNROFF=%%d]\n"); - if (root_wait) - printk(KERN_ERR - "Disabling rootwait; root= is invalid.\n"); - root_wait = 0; + if (sscanf(slash + 1, + "PARTNROFF=%d%c", &offset, &c) != 1) { + clear_root_wait = true; goto done; } + cmp.len = slash - uuid_str; + } else { + cmp.len = strlen(uuid_str); } - /* Pack the requested UUID in the expected format. */ - part_pack_uuid(uuid_str, uuid); + if (!cmp.len) { + clear_root_wait = true; + goto done; + } - dev = class_find_device(&block_class, NULL, uuid, &match_dev_by_uuid); + dev = class_find_device(&block_class, NULL, &cmp, + &match_dev_by_uuid); if (!dev) goto done; @@ -158,6 +166,13 @@ static dev_t devt_from_partuuid(char *uuid_str) no_offset: put_device(dev); done: + if (clear_root_wait) { + pr_err("VFS: PARTUUID= is invalid.\n" + "Expected PARTUUID=<valid-uuid-id>[/PARTNROFF=%%d]\n"); + if (root_wait) + pr_err("Disabling rootwait; root= is invalid.\n"); + root_wait = 0; + } return res; } #endif @@ -174,6 +189,10 @@ done: * used when disk name of partitioned disk ends on a digit. * 6) PARTUUID=00112233-4455-6677-8899-AABBCCDDEEFF representing the * unique id of a partition if the partition table provides it. + * The UUID may be either an EFI/GPT UUID, or refer to an MSDOS + * partition using the format SSSSSSSS-PP, where SSSSSSSS is a zero- + * filled hex representation of the 32-bit "NT disk signature", and PP + * is a zero-filled hex representation of the 1-based partition number. * 7) PARTUUID=<UUID>/PARTNROFF=<int> to select a partition in relation to * a partition with a known unique id. * diff --git a/init/main.c b/init/main.c index 63ae904a99a8..baf1f0f5c461 100644 --- a/init/main.c +++ b/init/main.c @@ -812,7 +812,6 @@ static int __ref kernel_init(void *unused) system_state = SYSTEM_RUNNING; numa_default_policy(); - current->signal->flags |= SIGNAL_UNKILLABLE; flush_delayed_fput(); if (ramdisk_execute_command) { diff --git a/init/version.c b/init/version.c index 86fe0ccb997a..58170f18912d 100644 --- a/init/version.c +++ b/init/version.c @@ -12,6 +12,7 @@ #include <linux/utsname.h> #include <generated/utsrelease.h> #include <linux/version.h> +#include <linux/proc_fs.h> #ifndef CONFIG_KALLSYMS #define version(a) Version_ ## a @@ -34,6 +35,7 @@ struct uts_namespace init_uts_ns = { .domainname = UTS_DOMAINNAME, }, .user_ns = &init_user_ns, + .proc_inum = PROC_UTS_INIT_INO, }; EXPORT_SYMBOL_GPL(init_uts_ns); diff --git a/ipc/msgutil.c b/ipc/msgutil.c index 26143d377c95..6471f1bdae96 100644 --- a/ipc/msgutil.c +++ b/ipc/msgutil.c @@ -16,6 +16,7 @@ #include <linux/msg.h> #include <linux/ipc_namespace.h> #include <linux/utsname.h> +#include <linux/proc_fs.h> #include <asm/uaccess.h> #include "util.h" @@ -30,6 +31,7 @@ DEFINE_SPINLOCK(mq_lock); struct ipc_namespace init_ipc_ns = { .count = ATOMIC_INIT(1), .user_ns = &init_user_ns, + .proc_inum = PROC_IPC_INIT_INO, }; atomic_t nr_ipc_ns = ATOMIC_INIT(1); diff --git a/ipc/namespace.c b/ipc/namespace.c index f362298c5ce4..cf3386a51de2 100644 --- a/ipc/namespace.c +++ b/ipc/namespace.c @@ -16,7 +16,7 @@ #include "util.h" -static struct ipc_namespace *create_ipc_ns(struct task_struct *tsk, +static struct ipc_namespace *create_ipc_ns(struct user_namespace *user_ns, struct ipc_namespace *old_ns) { struct ipc_namespace *ns; @@ -26,9 +26,16 @@ static struct ipc_namespace *create_ipc_ns(struct task_struct *tsk, if (ns == NULL) return ERR_PTR(-ENOMEM); + err = proc_alloc_inum(&ns->proc_inum); + if (err) { + kfree(ns); + return ERR_PTR(err); + } + atomic_set(&ns->count, 1); err = mq_init_ns(ns); if (err) { + proc_free_inum(ns->proc_inum); kfree(ns); return ERR_PTR(err); } @@ -46,19 +53,17 @@ static struct ipc_namespace *create_ipc_ns(struct task_struct *tsk, ipcns_notify(IPCNS_CREATED); register_ipcns_notifier(ns); - ns->user_ns = get_user_ns(task_cred_xxx(tsk, user_ns)); + ns->user_ns = get_user_ns(user_ns); return ns; } struct ipc_namespace *copy_ipcs(unsigned long flags, - struct task_struct *tsk) + struct user_namespace *user_ns, struct ipc_namespace *ns) { - struct ipc_namespace *ns = tsk->nsproxy->ipc_ns; - if (!(flags & CLONE_NEWIPC)) return get_ipc_ns(ns); - return create_ipc_ns(tsk, ns); + return create_ipc_ns(user_ns, ns); } /* @@ -113,6 +118,7 @@ static void free_ipc_ns(struct ipc_namespace *ns) */ ipcns_notify(IPCNS_REMOVED); put_user_ns(ns->user_ns); + proc_free_inum(ns->proc_inum); kfree(ns); } @@ -161,8 +167,12 @@ static void ipcns_put(void *ns) return put_ipc_ns(ns); } -static int ipcns_install(struct nsproxy *nsproxy, void *ns) +static int ipcns_install(struct nsproxy *nsproxy, void *new) { + struct ipc_namespace *ns = new; + if (!ns_capable(ns->user_ns, CAP_SYS_ADMIN)) + return -EPERM; + /* Ditch state from the old ipc namespace */ exit_sem(current); put_ipc_ns(nsproxy->ipc_ns); @@ -170,10 +180,18 @@ static int ipcns_install(struct nsproxy *nsproxy, void *ns) return 0; } +static unsigned int ipcns_inum(void *vp) +{ + struct ipc_namespace *ns = vp; + + return ns->proc_inum; +} + const struct proc_ns_operations ipcns_operations = { .name = "ipc", .type = CLONE_NEWIPC, .get = ipcns_get, .put = ipcns_put, .install = ipcns_install, + .inum = ipcns_inum, }; diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 571264830a29..4855892798fd 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -3408,7 +3408,7 @@ static struct cgroup_pidlist *cgroup_pidlist_find(struct cgroup *cgrp, { struct cgroup_pidlist *l; /* don't need task_nsproxy() if we're looking at ourself */ - struct pid_namespace *ns = current->nsproxy->pid_ns; + struct pid_namespace *ns = task_active_pid_ns(current); /* * We can't drop the pidlist_mutex before taking the l->mutex in case diff --git a/kernel/events/core.c b/kernel/events/core.c index f9ff5493171d..301079d06f24 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -6155,7 +6155,7 @@ perf_event_alloc(struct perf_event_attr *attr, int cpu, event->parent = parent_event; - event->ns = get_pid_ns(current->nsproxy->pid_ns); + event->ns = get_pid_ns(task_active_pid_ns(current)); event->id = atomic64_inc_return(&perf_event_id); event->state = PERF_EVENT_STATE_INACTIVE; diff --git a/kernel/exit.c b/kernel/exit.c index 50d2e93c36ea..b4df21937216 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -72,18 +72,6 @@ static void __unhash_process(struct task_struct *p, bool group_dead) list_del_rcu(&p->tasks); list_del_init(&p->sibling); __this_cpu_dec(process_counts); - /* - * If we are the last child process in a pid namespace to be - * reaped, notify the reaper sleeping zap_pid_ns_processes(). - */ - if (IS_ENABLED(CONFIG_PID_NS)) { - struct task_struct *parent = p->real_parent; - - if ((task_active_pid_ns(parent)->child_reaper == parent) && - list_empty(&parent->children) && - (parent->flags & PF_EXITING)) - wake_up_process(parent); - } } list_del_rcu(&p->thread_group); } diff --git a/kernel/fork.c b/kernel/fork.c index 115d6c2e4cca..c36c4e301efe 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1044,8 +1044,6 @@ static int copy_signal(unsigned long clone_flags, struct task_struct *tsk) atomic_set(&sig->live, 1); atomic_set(&sig->sigcnt, 1); init_waitqueue_head(&sig->wait_chldexit); - if (clone_flags & CLONE_NEWPID) - sig->flags |= SIGNAL_UNKILLABLE; sig->curr_target = tsk; init_sigpending(&sig->shared_pending); INIT_LIST_HEAD(&sig->posix_timers); @@ -1438,8 +1436,10 @@ static struct task_struct *copy_process(unsigned long clone_flags, ptrace_init_task(p, (clone_flags & CLONE_PTRACE) || trace); if (thread_group_leader(p)) { - if (is_child_reaper(pid)) - p->nsproxy->pid_ns->child_reaper = p; + if (is_child_reaper(pid)) { + ns_of_pid(pid)->child_reaper = p; + p->signal->flags |= SIGNAL_UNKILLABLE; + } p->signal->leader_pid = pid; p->signal->tty = tty_kref_get(current->signal->tty); @@ -1473,8 +1473,6 @@ bad_fork_cleanup_io: if (p->io_context) exit_io_context(p); bad_fork_cleanup_namespaces: - if (unlikely(clone_flags & CLONE_NEWPID)) - pid_ns_release_proc(p->nsproxy->pid_ns); exit_task_namespaces(p); bad_fork_cleanup_mm: if (p->mm) @@ -1554,15 +1552,9 @@ long do_fork(unsigned long clone_flags, * Do some preliminary argument and permissions checking before we * actually start allocating stuff */ - if (clone_flags & CLONE_NEWUSER) { - if (clone_flags & CLONE_THREAD) + if (clone_flags & (CLONE_NEWUSER | CLONE_NEWPID)) { + if (clone_flags & (CLONE_THREAD|CLONE_PARENT)) return -EINVAL; - /* hopefully this check will go away when userns support is - * complete - */ - if (!capable(CAP_SYS_ADMIN) || !capable(CAP_SETUID) || - !capable(CAP_SETGID)) - return -EPERM; } /* @@ -1724,7 +1716,8 @@ static int check_unshare_flags(unsigned long unshare_flags) { if (unshare_flags & ~(CLONE_THREAD|CLONE_FS|CLONE_NEWNS|CLONE_SIGHAND| CLONE_VM|CLONE_FILES|CLONE_SYSVSEM| - CLONE_NEWUTS|CLONE_NEWIPC|CLONE_NEWNET)) + CLONE_NEWUTS|CLONE_NEWIPC|CLONE_NEWNET| + CLONE_NEWUSER|CLONE_NEWPID)) return -EINVAL; /* * Not implemented, but pretend it works if there is nothing to @@ -1791,19 +1784,40 @@ SYSCALL_DEFINE1(unshare, unsigned long, unshare_flags) { struct fs_struct *fs, *new_fs = NULL; struct files_struct *fd, *new_fd = NULL; + struct cred *new_cred = NULL; struct nsproxy *new_nsproxy = NULL; int do_sysvsem = 0; int err; - err = check_unshare_flags(unshare_flags); - if (err) - goto bad_unshare_out; - + /* + * If unsharing a user namespace must also unshare the thread. + */ + if (unshare_flags & CLONE_NEWUSER) + unshare_flags |= CLONE_THREAD; + /* + * If unsharing a pid namespace must also unshare the thread. + */ + if (unshare_flags & CLONE_NEWPID) + unshare_flags |= CLONE_THREAD; + /* + * If unsharing a thread from a thread group, must also unshare vm. + */ + if (unshare_flags & CLONE_THREAD) + unshare_flags |= CLONE_VM; + /* + * If unsharing vm, must also unshare signal handlers. + */ + if (unshare_flags & CLONE_VM) + unshare_flags |= CLONE_SIGHAND; /* * If unsharing namespace, must also unshare filesystem information. */ if (unshare_flags & CLONE_NEWNS) unshare_flags |= CLONE_FS; + + err = check_unshare_flags(unshare_flags); + if (err) + goto bad_unshare_out; /* * CLONE_NEWIPC must also detach from the undolist: after switching * to a new ipc namespace, the semaphore arrays from the old @@ -1817,11 +1831,15 @@ SYSCALL_DEFINE1(unshare, unsigned long, unshare_flags) err = unshare_fd(unshare_flags, &new_fd); if (err) goto bad_unshare_cleanup_fs; - err = unshare_nsproxy_namespaces(unshare_flags, &new_nsproxy, new_fs); + err = unshare_userns(unshare_flags, &new_cred); if (err) goto bad_unshare_cleanup_fd; + err = unshare_nsproxy_namespaces(unshare_flags, &new_nsproxy, + new_cred, new_fs); + if (err) + goto bad_unshare_cleanup_cred; - if (new_fs || new_fd || do_sysvsem || new_nsproxy) { + if (new_fs || new_fd || do_sysvsem || new_cred || new_nsproxy) { if (do_sysvsem) { /* * CLONE_SYSVSEM is equivalent to sys_exit(). @@ -1854,11 +1872,20 @@ SYSCALL_DEFINE1(unshare, unsigned long, unshare_flags) } task_unlock(current); + + if (new_cred) { + /* Install the new user namespace */ + commit_creds(new_cred); + new_cred = NULL; + } } if (new_nsproxy) put_nsproxy(new_nsproxy); +bad_unshare_cleanup_cred: + if (new_cred) + put_cred(new_cred); bad_unshare_cleanup_fd: if (new_fd) put_files_struct(new_fd); diff --git a/kernel/nsproxy.c b/kernel/nsproxy.c index 7e1c3de1ce45..78e2ecb20165 100644 --- a/kernel/nsproxy.c +++ b/kernel/nsproxy.c @@ -57,7 +57,8 @@ static inline struct nsproxy *create_nsproxy(void) * leave it to the caller to do proper locking and attach it to task. */ static struct nsproxy *create_new_namespaces(unsigned long flags, - struct task_struct *tsk, struct fs_struct *new_fs) + struct task_struct *tsk, struct user_namespace *user_ns, + struct fs_struct *new_fs) { struct nsproxy *new_nsp; int err; @@ -66,31 +67,31 @@ static struct nsproxy *create_new_namespaces(unsigned long flags, if (!new_nsp) return ERR_PTR(-ENOMEM); - new_nsp->mnt_ns = copy_mnt_ns(flags, tsk->nsproxy->mnt_ns, new_fs); + new_nsp->mnt_ns = copy_mnt_ns(flags, tsk->nsproxy->mnt_ns, user_ns, new_fs); if (IS_ERR(new_nsp->mnt_ns)) { err = PTR_ERR(new_nsp->mnt_ns); goto out_ns; } - new_nsp->uts_ns = copy_utsname(flags, tsk); + new_nsp->uts_ns = copy_utsname(flags, user_ns, tsk->nsproxy->uts_ns); if (IS_ERR(new_nsp->uts_ns)) { err = PTR_ERR(new_nsp->uts_ns); goto out_uts; } - new_nsp->ipc_ns = copy_ipcs(flags, tsk); + new_nsp->ipc_ns = copy_ipcs(flags, user_ns, tsk->nsproxy->ipc_ns); if (IS_ERR(new_nsp->ipc_ns)) { err = PTR_ERR(new_nsp->ipc_ns); goto out_ipc; } - new_nsp->pid_ns = copy_pid_ns(flags, task_active_pid_ns(tsk)); + new_nsp->pid_ns = copy_pid_ns(flags, user_ns, tsk->nsproxy->pid_ns); if (IS_ERR(new_nsp->pid_ns)) { err = PTR_ERR(new_nsp->pid_ns); goto out_pid; } - new_nsp->net_ns = copy_net_ns(flags, task_cred_xxx(tsk, user_ns), tsk->nsproxy->net_ns); + new_nsp->net_ns = copy_net_ns(flags, user_ns, tsk->nsproxy->net_ns); if (IS_ERR(new_nsp->net_ns)) { err = PTR_ERR(new_nsp->net_ns); goto out_net; @@ -122,6 +123,7 @@ out_ns: int copy_namespaces(unsigned long flags, struct task_struct *tsk) { struct nsproxy *old_ns = tsk->nsproxy; + struct user_namespace *user_ns = task_cred_xxx(tsk, user_ns); struct nsproxy *new_ns; int err = 0; @@ -134,7 +136,7 @@ int copy_namespaces(unsigned long flags, struct task_struct *tsk) CLONE_NEWPID | CLONE_NEWNET))) return 0; - if (!capable(CAP_SYS_ADMIN)) { + if (!ns_capable(user_ns, CAP_SYS_ADMIN)) { err = -EPERM; goto out; } @@ -151,7 +153,8 @@ int copy_namespaces(unsigned long flags, struct task_struct *tsk) goto out; } - new_ns = create_new_namespaces(flags, tsk, tsk->fs); + new_ns = create_new_namespaces(flags, tsk, + task_cred_xxx(tsk, user_ns), tsk->fs); if (IS_ERR(new_ns)) { err = PTR_ERR(new_ns); goto out; @@ -183,19 +186,21 @@ void free_nsproxy(struct nsproxy *ns) * On success, returns the new nsproxy. */ int unshare_nsproxy_namespaces(unsigned long unshare_flags, - struct nsproxy **new_nsp, struct fs_struct *new_fs) + struct nsproxy **new_nsp, struct cred *new_cred, struct fs_struct *new_fs) { + struct user_namespace *user_ns; int err = 0; if (!(unshare_flags & (CLONE_NEWNS | CLONE_NEWUTS | CLONE_NEWIPC | - CLONE_NEWNET))) + CLONE_NEWNET | CLONE_NEWPID))) return 0; - if (!capable(CAP_SYS_ADMIN)) + user_ns = new_cred ? new_cred->user_ns : current_user_ns(); + if (!ns_capable(user_ns, CAP_SYS_ADMIN)) return -EPERM; - *new_nsp = create_new_namespaces(unshare_flags, current, - new_fs ? new_fs : current->fs); + *new_nsp = create_new_namespaces(unshare_flags, current, user_ns, + new_fs ? new_fs : current->fs); if (IS_ERR(*new_nsp)) { err = PTR_ERR(*new_nsp); goto out; @@ -241,9 +246,6 @@ SYSCALL_DEFINE2(setns, int, fd, int, nstype) struct file *file; int err; - if (!capable(CAP_SYS_ADMIN)) - return -EPERM; - file = proc_ns_fget(fd); if (IS_ERR(file)) return PTR_ERR(file); @@ -254,7 +256,7 @@ SYSCALL_DEFINE2(setns, int, fd, int, nstype) if (nstype && (ops->type != nstype)) goto out; - new_nsproxy = create_new_namespaces(0, tsk, tsk->fs); + new_nsproxy = create_new_namespaces(0, tsk, current_user_ns(), tsk->fs); if (IS_ERR(new_nsproxy)) { err = PTR_ERR(new_nsproxy); goto out; diff --git a/kernel/pid.c b/kernel/pid.c index a54a1123c7cf..36aa02ff17d6 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -36,6 +36,7 @@ #include <linux/pid_namespace.h> #include <linux/init_task.h> #include <linux/syscalls.h> +#include <linux/proc_fs.h> #define pid_hashfn(nr, ns) \ hash_long((unsigned long)nr + (unsigned long)ns, pidhash_shift) @@ -78,6 +79,8 @@ struct pid_namespace init_pid_ns = { .last_pid = 0, .level = 0, .child_reaper = &init_task, + .user_ns = &init_user_ns, + .proc_inum = PROC_PID_INIT_INO, }; EXPORT_SYMBOL_GPL(init_pid_ns); @@ -254,8 +257,24 @@ void free_pid(struct pid *pid) unsigned long flags; spin_lock_irqsave(&pidmap_lock, flags); - for (i = 0; i <= pid->level; i++) - hlist_del_rcu(&pid->numbers[i].pid_chain); + for (i = 0; i <= pid->level; i++) { + struct upid *upid = pid->numbers + i; + struct pid_namespace *ns = upid->ns; + hlist_del_rcu(&upid->pid_chain); + switch(--ns->nr_hashed) { + case 1: + /* When all that is left in the pid namespace + * is the reaper wake up the reaper. The reaper + * may be sleeping in zap_pid_ns_processes(). + */ + wake_up_process(ns->child_reaper); + break; + case 0: + ns->nr_hashed = -1; + schedule_work(&ns->proc_work); + break; + } + } spin_unlock_irqrestore(&pidmap_lock, flags); for (i = 0; i <= pid->level; i++) @@ -277,6 +296,7 @@ struct pid *alloc_pid(struct pid_namespace *ns) goto out; tmp = ns; + pid->level = ns->level; for (i = ns->level; i >= 0; i--) { nr = alloc_pidmap(tmp); if (nr < 0) @@ -287,22 +307,32 @@ struct pid *alloc_pid(struct pid_namespace *ns) tmp = tmp->parent; } + if (unlikely(is_child_reaper(pid))) { + if (pid_ns_prepare_proc(ns)) + goto out_free; + } + get_pid_ns(ns); - pid->level = ns->level; atomic_set(&pid->count, 1); for (type = 0; type < PIDTYPE_MAX; ++type) INIT_HLIST_HEAD(&pid->tasks[type]); upid = pid->numbers + ns->level; spin_lock_irq(&pidmap_lock); - for ( ; upid >= pid->numbers; --upid) + if (ns->nr_hashed < 0) + goto out_unlock; + for ( ; upid >= pid->numbers; --upid) { hlist_add_head_rcu(&upid->pid_chain, &pid_hash[pid_hashfn(upid->nr, upid->ns)]); + upid->ns->nr_hashed++; + } spin_unlock_irq(&pidmap_lock); out: return pid; +out_unlock: + spin_unlock(&pidmap_lock); out_free: while (++i <= ns->level) free_pidmap(pid->numbers + i); @@ -329,7 +359,7 @@ EXPORT_SYMBOL_GPL(find_pid_ns); struct pid *find_vpid(int nr) { - return find_pid_ns(nr, current->nsproxy->pid_ns); + return find_pid_ns(nr, task_active_pid_ns(current)); } EXPORT_SYMBOL_GPL(find_vpid); @@ -413,7 +443,7 @@ struct task_struct *find_task_by_pid_ns(pid_t nr, struct pid_namespace *ns) struct task_struct *find_task_by_vpid(pid_t vnr) { - return find_task_by_pid_ns(vnr, current->nsproxy->pid_ns); + return find_task_by_pid_ns(vnr, task_active_pid_ns(current)); } struct pid *get_task_pid(struct task_struct *task, enum pid_type type) @@ -468,7 +498,7 @@ EXPORT_SYMBOL_GPL(pid_nr_ns); pid_t pid_vnr(struct pid *pid) { - return pid_nr_ns(pid, current->nsproxy->pid_ns); + return pid_nr_ns(pid, task_active_pid_ns(current)); } EXPORT_SYMBOL_GPL(pid_vnr); @@ -479,7 +509,7 @@ pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type, rcu_read_lock(); if (!ns) - ns = current->nsproxy->pid_ns; + ns = task_active_pid_ns(current); if (likely(pid_alive(task))) { if (type != PIDTYPE_PID) task = task->group_leader; @@ -554,6 +584,7 @@ void __init pidmap_init(void) /* Reserve PID 0. We never call free_pidmap(0) */ set_bit(0, init_pid_ns.pidmap[0].page); atomic_dec(&init_pid_ns.pidmap[0].nr_free); + init_pid_ns.nr_hashed = 1; init_pid_ns.pid_cachep = KMEM_CACHE(pid, SLAB_HWCACHE_ALIGN | SLAB_PANIC); diff --git a/kernel/pid_namespace.c b/kernel/pid_namespace.c index 7b07cc0dfb75..560da0dab230 100644 --- a/kernel/pid_namespace.c +++ b/kernel/pid_namespace.c @@ -10,6 +10,7 @@ #include <linux/pid.h> #include <linux/pid_namespace.h> +#include <linux/user_namespace.h> #include <linux/syscalls.h> #include <linux/err.h> #include <linux/acct.h> @@ -71,10 +72,17 @@ err_alloc: return NULL; } +static void proc_cleanup_work(struct work_struct *work) +{ + struct pid_namespace *ns = container_of(work, struct pid_namespace, proc_work); + pid_ns_release_proc(ns); +} + /* MAX_PID_NS_LEVEL is needed for limiting size of 'struct pid' */ #define MAX_PID_NS_LEVEL 32 -static struct pid_namespace *create_pid_namespace(struct pid_namespace *parent_pid_ns) +static struct pid_namespace *create_pid_namespace(struct user_namespace *user_ns, + struct pid_namespace *parent_pid_ns) { struct pid_namespace *ns; unsigned int level = parent_pid_ns->level + 1; @@ -99,9 +107,15 @@ static struct pid_namespace *create_pid_namespace(struct pid_namespace *parent_p if (ns->pid_cachep == NULL) goto out_free_map; + err = proc_alloc_inum(&ns->proc_inum); + if (err) + goto out_free_map; + kref_init(&ns->kref); ns->level = level; ns->parent = get_pid_ns(parent_pid_ns); + ns->user_ns = get_user_ns(user_ns); + INIT_WORK(&ns->proc_work, proc_cleanup_work); set_bit(0, ns->pidmap[0].page); atomic_set(&ns->pidmap[0].nr_free, BITS_PER_PAGE - 1); @@ -109,14 +123,8 @@ static struct pid_namespace *create_pid_namespace(struct pid_namespace *parent_p for (i = 1; i < PIDMAP_ENTRIES; i++) atomic_set(&ns->pidmap[i].nr_free, BITS_PER_PAGE); - err = pid_ns_prepare_proc(ns); - if (err) - goto out_put_parent_pid_ns; - return ns; -out_put_parent_pid_ns: - put_pid_ns(parent_pid_ns); out_free_map: kfree(ns->pidmap[0].page); out_free: @@ -129,18 +137,21 @@ static void destroy_pid_namespace(struct pid_namespace *ns) { int i; + proc_free_inum(ns->proc_inum); for (i = 0; i < PIDMAP_ENTRIES; i++) kfree(ns->pidmap[i].page); + put_user_ns(ns->user_ns); kmem_cache_free(pid_ns_cachep, ns); } -struct pid_namespace *copy_pid_ns(unsigned long flags, struct pid_namespace *old_ns) +struct pid_namespace *copy_pid_ns(unsigned long flags, + struct user_namespace *user_ns, struct pid_namespace *old_ns) { if (!(flags & CLONE_NEWPID)) return get_pid_ns(old_ns); - if (flags & (CLONE_THREAD|CLONE_PARENT)) + if (task_active_pid_ns(current) != old_ns) return ERR_PTR(-EINVAL); - return create_pid_namespace(old_ns); + return create_pid_namespace(user_ns, old_ns); } static void free_pid_ns(struct kref *kref) @@ -211,22 +222,15 @@ void zap_pid_ns_processes(struct pid_namespace *pid_ns) /* * sys_wait4() above can't reap the TASK_DEAD children. - * Make sure they all go away, see __unhash_process(). + * Make sure they all go away, see free_pid(). */ for (;;) { - bool need_wait = false; - - read_lock(&tasklist_lock); - if (!list_empty(¤t->children)) { - __set_current_state(TASK_UNINTERRUPTIBLE); - need_wait = true; - } - read_unlock(&tasklist_lock); - - if (!need_wait) + set_current_state(TASK_UNINTERRUPTIBLE); + if (pid_ns->nr_hashed == 1) break; schedule(); } + __set_current_state(TASK_RUNNING); if (pid_ns->reboot) current->signal->group_exit_code = pid_ns->reboot; @@ -239,9 +243,10 @@ void zap_pid_ns_processes(struct pid_namespace *pid_ns) static int pid_ns_ctl_handler(struct ctl_table *table, int write, void __user *buffer, size_t *lenp, loff_t *ppos) { + struct pid_namespace *pid_ns = task_active_pid_ns(current); struct ctl_table tmp = *table; - if (write && !capable(CAP_SYS_ADMIN)) + if (write && !ns_capable(pid_ns->user_ns, CAP_SYS_ADMIN)) return -EPERM; /* @@ -250,7 +255,7 @@ static int pid_ns_ctl_handler(struct ctl_table *table, int write, * it should synchronize its usage with external means. */ - tmp.data = ¤t->nsproxy->pid_ns->last_pid; + tmp.data = &pid_ns->last_pid; return proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos); } @@ -299,6 +304,67 @@ int reboot_pid_ns(struct pid_namespace *pid_ns, int cmd) return 0; } +static void *pidns_get(struct task_struct *task) +{ + struct pid_namespace *ns; + + rcu_read_lock(); + ns = get_pid_ns(task_active_pid_ns(task)); + rcu_read_unlock(); + + return ns; +} + +static void pidns_put(void *ns) +{ + put_pid_ns(ns); +} + +static int pidns_install(struct nsproxy *nsproxy, void *ns) +{ + struct pid_namespace *active = task_active_pid_ns(current); + struct pid_namespace *ancestor, *new = ns; + + if (!ns_capable(new->user_ns, CAP_SYS_ADMIN)) + return -EPERM; + + /* + * Only allow entering the current active pid namespace + * or a child of the current active pid namespace. + * + * This is required for fork to return a usable pid value and + * this maintains the property that processes and their + * children can not escape their current pid namespace. + */ + if (new->level < active->level) + return -EINVAL; + + ancestor = new; + while (ancestor->level > active->level) + ancestor = ancestor->parent; + if (ancestor != active) + return -EINVAL; + + put_pid_ns(nsproxy->pid_ns); + nsproxy->pid_ns = get_pid_ns(new); + return 0; +} + +static unsigned int pidns_inum(void *ns) +{ + struct pid_namespace *pid_ns = ns; + return pid_ns->proc_inum; +} + +const struct proc_ns_operations pidns_operations = { + .name = "pid", + .type = CLONE_NEWPID, + .get = pidns_get, + .put = pidns_put, + .install = pidns_install, + .inum = pidns_inum, +}; + static __init int pid_namespaces_init(void) { pid_ns_cachep = KMEM_CACHE(pid_namespace, SLAB_PANIC); diff --git a/kernel/ptrace.c b/kernel/ptrace.c index ec8118ab2a47..1599157336a6 100644 --- a/kernel/ptrace.c +++ b/kernel/ptrace.c @@ -215,8 +215,12 @@ ok: smp_rmb(); if (task->mm) dumpable = get_dumpable(task->mm); - if (!dumpable && !ptrace_has_cap(task_user_ns(task), mode)) + rcu_read_lock(); + if (!dumpable && !ptrace_has_cap(__task_cred(task)->user_ns, mode)) { + rcu_read_unlock(); return -EPERM; + } + rcu_read_unlock(); return security_ptrace_access_check(task, mode); } @@ -280,8 +284,10 @@ static int ptrace_attach(struct task_struct *task, long request, if (seize) flags |= PT_SEIZED; - if (ns_capable(task_user_ns(task), CAP_SYS_PTRACE)) + rcu_read_lock(); + if (ns_capable(__task_cred(task)->user_ns, CAP_SYS_PTRACE)) flags |= PT_PTRACE_CAP; + rcu_read_unlock(); task->ptrace = flags; __ptrace_link(task, current); diff --git a/kernel/sched/core.c b/kernel/sched/core.c index c1fb82104bfb..257002c13bb0 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -4097,8 +4097,14 @@ long sched_setaffinity(pid_t pid, const struct cpumask *in_mask) goto out_free_cpus_allowed; } retval = -EPERM; - if (!check_same_owner(p) && !ns_capable(task_user_ns(p), CAP_SYS_NICE)) - goto out_unlock; + if (!check_same_owner(p)) { + rcu_read_lock(); + if (!ns_capable(__task_cred(p)->user_ns, CAP_SYS_NICE)) { + rcu_read_unlock(); + goto out_unlock; + } + rcu_read_unlock(); + } retval = security_task_setscheduler(p); if (retval) diff --git a/kernel/signal.c b/kernel/signal.c index a49c7f36ceb3..580a91e63471 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -1753,7 +1753,7 @@ static void do_notify_parent_cldstop(struct task_struct *tsk, * see comment in do_notify_parent() about the following 4 lines */ rcu_read_lock(); - info.si_pid = task_pid_nr_ns(tsk, parent->nsproxy->pid_ns); + info.si_pid = task_pid_nr_ns(tsk, task_active_pid_ns(parent)); info.si_uid = from_kuid_munged(task_cred_xxx(parent, user_ns), task_uid(tsk)); rcu_read_unlock(); diff --git a/kernel/sysctl_binary.c b/kernel/sysctl_binary.c index 65bdcf198d4e..5a6384450501 100644 --- a/kernel/sysctl_binary.c +++ b/kernel/sysctl_binary.c @@ -1344,7 +1344,7 @@ static ssize_t binary_sysctl(const int *name, int nlen, goto out_putname; } - mnt = current->nsproxy->pid_ns->proc_mnt; + mnt = task_active_pid_ns(current)->proc_mnt; file = file_open_root(mnt->mnt_root, mnt, pathname, flags); result = PTR_ERR(file); if (IS_ERR(file)) diff --git a/kernel/user.c b/kernel/user.c index 750acffbe9ec..33acb5e53a5f 100644 --- a/kernel/user.c +++ b/kernel/user.c @@ -16,6 +16,7 @@ #include <linux/interrupt.h> #include <linux/export.h> #include <linux/user_namespace.h> +#include <linux/proc_fs.h> /* * userns count is 1 for root user, 1 for init_uts_ns, @@ -51,6 +52,7 @@ struct user_namespace init_user_ns = { }, .owner = GLOBAL_ROOT_UID, .group = GLOBAL_ROOT_GID, + .proc_inum = PROC_USER_INIT_INO, }; EXPORT_SYMBOL_GPL(init_user_ns); diff --git a/kernel/user_namespace.c b/kernel/user_namespace.c index 456a6b9fba34..f5975ccf9348 100644 --- a/kernel/user_namespace.c +++ b/kernel/user_namespace.c @@ -9,6 +9,7 @@ #include <linux/nsproxy.h> #include <linux/slab.h> #include <linux/user_namespace.h> +#include <linux/proc_fs.h> #include <linux/highuid.h> #include <linux/cred.h> #include <linux/securebits.h> @@ -26,6 +27,24 @@ static struct kmem_cache *user_ns_cachep __read_mostly; static bool new_idmap_permitted(struct user_namespace *ns, int cap_setid, struct uid_gid_map *map); +static void set_cred_user_ns(struct cred *cred, struct user_namespace *user_ns) +{ + /* Start with the same capabilities as init but useless for doing + * anything as the capabilities are bound to the new user namespace. + */ + cred->securebits = SECUREBITS_DEFAULT; + cred->cap_inheritable = CAP_EMPTY_SET; + cred->cap_permitted = CAP_FULL_SET; + cred->cap_effective = CAP_FULL_SET; + cred->cap_bset = CAP_FULL_SET; +#ifdef CONFIG_KEYS + key_put(cred->request_key_auth); + cred->request_key_auth = NULL; +#endif + /* tgcred will be cleared in our caller bc CLONE_THREAD won't be set */ + cred->user_ns = user_ns; +} + /* * Create a new user namespace, deriving the creator from the user in the * passed credentials, and replacing that user with the new root user for the @@ -39,6 +58,7 @@ int create_user_ns(struct cred *new) struct user_namespace *ns, *parent_ns = new->user_ns; kuid_t owner = new->euid; kgid_t group = new->egid; + int ret; /* The creator needs a mapping in the parent user namespace * or else we won't be able to reasonably tell userspace who @@ -52,38 +72,45 @@ int create_user_ns(struct cred *new) if (!ns) return -ENOMEM; + ret = proc_alloc_inum(&ns->proc_inum); + if (ret) { + kmem_cache_free(user_ns_cachep, ns); + return ret; + } + kref_init(&ns->kref); + /* Leave the new->user_ns reference with the new user namespace. */ ns->parent = parent_ns; ns->owner = owner; ns->group = group; - /* Start with the same capabilities as init but useless for doing - * anything as the capabilities are bound to the new user namespace. - */ - new->securebits = SECUREBITS_DEFAULT; - new->cap_inheritable = CAP_EMPTY_SET; - new->cap_permitted = CAP_FULL_SET; - new->cap_effective = CAP_FULL_SET; - new->cap_bset = CAP_FULL_SET; -#ifdef CONFIG_KEYS - key_put(new->request_key_auth); - new->request_key_auth = NULL; -#endif - /* tgcred will be cleared in our caller bc CLONE_THREAD won't be set */ - - /* Leave the new->user_ns reference with the new user namespace. */ - /* Leave the reference to our user_ns with the new cred. */ - new->user_ns = ns; + set_cred_user_ns(new, ns); return 0; } +int unshare_userns(unsigned long unshare_flags, struct cred **new_cred) +{ + struct cred *cred; + + if (!(unshare_flags & CLONE_NEWUSER)) + return 0; + + cred = prepare_creds(); + if (!cred) + return -ENOMEM; + + *new_cred = cred; + return create_user_ns(cred); +} + void free_user_ns(struct kref *kref) { struct user_namespace *parent, *ns = container_of(kref, struct user_namespace, kref); parent = ns->parent; + proc_free_inum(ns->proc_inum); kmem_cache_free(user_ns_cachep, ns); put_user_ns(parent); } @@ -372,7 +399,7 @@ static int uid_m_show(struct seq_file *seq, void *v) struct user_namespace *lower_ns; uid_t lower; - lower_ns = current_user_ns(); + lower_ns = seq_user_ns(seq); if ((lower_ns == ns) && lower_ns->parent) lower_ns = lower_ns->parent; @@ -393,7 +420,7 @@ static int gid_m_show(struct seq_file *seq, void *v) struct user_namespace *lower_ns; gid_t lower; - lower_ns = current_user_ns(); + lower_ns = seq_user_ns(seq); if ((lower_ns == ns) && lower_ns->parent) lower_ns = lower_ns->parent; @@ -669,10 +696,14 @@ ssize_t proc_uid_map_write(struct file *file, const char __user *buf, size_t siz { struct seq_file *seq = file->private_data; struct user_namespace *ns = seq->private; + struct user_namespace *seq_ns = seq_user_ns(seq); if (!ns->parent) return -EPERM; + if ((seq_ns != ns) && (seq_ns != ns->parent)) + return -EPERM; + return map_write(file, buf, size, ppos, CAP_SETUID, &ns->uid_map, &ns->parent->uid_map); } @@ -681,10 +712,14 @@ ssize_t proc_gid_map_write(struct file *file, const char __user *buf, size_t siz { struct seq_file *seq = file->private_data; struct user_namespace *ns = seq->private; + struct user_namespace *seq_ns = seq_user_ns(seq); if (!ns->parent) return -EPERM; + if ((seq_ns != ns) && (seq_ns != ns->parent)) + return -EPERM; + return map_write(file, buf, size, ppos, CAP_SETGID, &ns->gid_map, &ns->parent->gid_map); } @@ -709,6 +744,21 @@ ssize_t proc_projid_map_write(struct file *file, const char __user *buf, size_t static bool new_idmap_permitted(struct user_namespace *ns, int cap_setid, struct uid_gid_map *new_map) { + /* Allow mapping to your own filesystem ids */ + if ((new_map->nr_extents == 1) && (new_map->extent[0].count == 1)) { + u32 id = new_map->extent[0].lower_first; + if (cap_setid == CAP_SETUID) { + kuid_t uid = make_kuid(ns->parent, id); + if (uid_eq(uid, current_fsuid())) + return true; + } + else if (cap_setid == CAP_SETGID) { + kgid_t gid = make_kgid(ns->parent, id); + if (gid_eq(gid, current_fsgid())) + return true; + } + } + /* Allow anyone to set a mapping that doesn't require privilege */ if (!cap_valid(cap_setid)) return true; @@ -722,6 +772,65 @@ static bool new_idmap_permitted(struct user_namespace *ns, int cap_setid, return false; } +static void *userns_get(struct task_struct *task) +{ + struct user_namespace *user_ns; + + rcu_read_lock(); + user_ns = get_user_ns(__task_cred(task)->user_ns); + rcu_read_unlock(); + + return user_ns; +} + +static void userns_put(void *ns) +{ + put_user_ns(ns); +} + +static int userns_install(struct nsproxy *nsproxy, void *ns) +{ + struct user_namespace *user_ns = ns; + struct cred *cred; + + /* Don't allow gaining capabilities by reentering + * the same user namespace. + */ + if (user_ns == current_user_ns()) + return -EINVAL; + + /* Threaded many not enter a different user namespace */ + if (atomic_read(¤t->mm->mm_users) > 1) + return -EINVAL; + + if (!ns_capable(user_ns, CAP_SYS_ADMIN)) + return -EPERM; + + cred = prepare_creds(); + if (!cred) + return -ENOMEM; + + put_user_ns(cred->user_ns); + set_cred_user_ns(cred, get_user_ns(user_ns)); + + return commit_creds(cred); +} + +static unsigned int userns_inum(void *ns) +{ + struct user_namespace *user_ns = ns; + return user_ns->proc_inum; +} + +const struct proc_ns_operations userns_operations = { + .name = "user", + .type = CLONE_NEWUSER, + .get = userns_get, + .put = userns_put, + .install = userns_install, + .inum = userns_inum, +}; + static __init int user_namespaces_init(void) { user_ns_cachep = KMEM_CACHE(user_namespace, SLAB_PANIC); diff --git a/kernel/utsname.c b/kernel/utsname.c index 679d97a5d3fd..f6336d51d64c 100644 --- a/kernel/utsname.c +++ b/kernel/utsname.c @@ -32,18 +32,25 @@ static struct uts_namespace *create_uts_ns(void) * @old_ns: namespace to clone * Return NULL on error (failure to kmalloc), new ns otherwise */ -static struct uts_namespace *clone_uts_ns(struct task_struct *tsk, +static struct uts_namespace *clone_uts_ns(struct user_namespace *user_ns, struct uts_namespace *old_ns) { struct uts_namespace *ns; + int err; ns = create_uts_ns(); if (!ns) return ERR_PTR(-ENOMEM); + err = proc_alloc_inum(&ns->proc_inum); + if (err) { + kfree(ns); + return ERR_PTR(err); + } + down_read(&uts_sem); memcpy(&ns->name, &old_ns->name, sizeof(ns->name)); - ns->user_ns = get_user_ns(task_cred_xxx(tsk, user_ns)); + ns->user_ns = get_user_ns(user_ns); up_read(&uts_sem); return ns; } @@ -55,9 +62,8 @@ static struct uts_namespace *clone_uts_ns(struct task_struct *tsk, * versa. */ struct uts_namespace *copy_utsname(unsigned long flags, - struct task_struct *tsk) + struct user_namespace *user_ns, struct uts_namespace *old_ns) { - struct uts_namespace *old_ns = tsk->nsproxy->uts_ns; struct uts_namespace *new_ns; BUG_ON(!old_ns); @@ -66,7 +72,7 @@ struct uts_namespace *copy_utsname(unsigned long flags, if (!(flags & CLONE_NEWUTS)) return old_ns; - new_ns = clone_uts_ns(tsk, old_ns); + new_ns = clone_uts_ns(user_ns, old_ns); put_uts_ns(old_ns); return new_ns; @@ -78,6 +84,7 @@ void free_uts_ns(struct kref *kref) ns = container_of(kref, struct uts_namespace, kref); put_user_ns(ns->user_ns); + proc_free_inum(ns->proc_inum); kfree(ns); } @@ -102,19 +109,31 @@ static void utsns_put(void *ns) put_uts_ns(ns); } -static int utsns_install(struct nsproxy *nsproxy, void *ns) +static int utsns_install(struct nsproxy *nsproxy, void *new) { + struct uts_namespace *ns = new; + + if (!ns_capable(ns->user_ns, CAP_SYS_ADMIN)) + return -EPERM; + get_uts_ns(ns); put_uts_ns(nsproxy->uts_ns); nsproxy->uts_ns = ns; return 0; } +static unsigned int utsns_inum(void *vp) +{ + struct uts_namespace *ns = vp; + + return ns->proc_inum; +} + const struct proc_ns_operations utsns_operations = { .name = "uts", .type = CLONE_NEWUTS, .get = utsns_get, .put = utsns_put, .install = utsns_install, + .inum = utsns_inum, }; - diff --git a/lib/lru_cache.c b/lib/lru_cache.c index a07e7268d7ed..d71d89498943 100644 --- a/lib/lru_cache.c +++ b/lib/lru_cache.c @@ -44,8 +44,8 @@ MODULE_LICENSE("GPL"); } while (0) #define RETURN(x...) do { \ - clear_bit(__LC_PARANOIA, &lc->flags); \ - smp_mb__after_clear_bit(); return x ; } while (0) + clear_bit_unlock(__LC_PARANOIA, &lc->flags); \ + return x ; } while (0) /* BUG() if e is not one of the elements tracked by lc */ #define PARANOIA_LC_ELEMENT(lc, e) do { \ @@ -55,9 +55,40 @@ MODULE_LICENSE("GPL"); BUG_ON(i >= lc_->nr_elements); \ BUG_ON(lc_->lc_element[i] != e_); } while (0) + +/* We need to atomically + * - try to grab the lock (set LC_LOCKED) + * - only if there is no pending transaction + * (neither LC_DIRTY nor LC_STARVING is set) + * Because of PARANOIA_ENTRY() above abusing lc->flags as well, + * it is not sufficient to just say + * return 0 == cmpxchg(&lc->flags, 0, LC_LOCKED); + */ +int lc_try_lock(struct lru_cache *lc) +{ + unsigned long val; + do { + val = cmpxchg(&lc->flags, 0, LC_LOCKED); + } while (unlikely (val == LC_PARANOIA)); + /* Spin until no-one is inside a PARANOIA_ENTRY()/RETURN() section. */ + return 0 == val; +#if 0 + /* Alternative approach, spin in case someone enters or leaves a + * PARANOIA_ENTRY()/RETURN() section. */ + unsigned long old, new, val; + do { + old = lc->flags & LC_PARANOIA; + new = old | LC_LOCKED; + val = cmpxchg(&lc->flags, old, new); + } while (unlikely (val == (old ^ LC_PARANOIA))); + return old == val; +#endif +} + /** * lc_create - prepares to track objects in an active set * @name: descriptive name only used in lc_seq_printf_stats and lc_seq_dump_details + * @max_pending_changes: maximum changes to accumulate until a transaction is required * @e_count: number of elements allowed to be active simultaneously * @e_size: size of the tracked objects * @e_off: offset to the &struct lc_element member in a tracked object @@ -66,6 +97,7 @@ MODULE_LICENSE("GPL"); * or NULL on (allocation) failure. */ struct lru_cache *lc_create(const char *name, struct kmem_cache *cache, + unsigned max_pending_changes, unsigned e_count, size_t e_size, size_t e_off) { struct hlist_head *slot = NULL; @@ -98,12 +130,13 @@ struct lru_cache *lc_create(const char *name, struct kmem_cache *cache, INIT_LIST_HEAD(&lc->in_use); INIT_LIST_HEAD(&lc->lru); INIT_LIST_HEAD(&lc->free); + INIT_LIST_HEAD(&lc->to_be_changed); lc->name = name; lc->element_size = e_size; lc->element_off = e_off; lc->nr_elements = e_count; - lc->new_number = LC_FREE; + lc->max_pending_changes = max_pending_changes; lc->lc_cache = cache; lc->lc_element = element; lc->lc_slot = slot; @@ -117,6 +150,7 @@ struct lru_cache *lc_create(const char *name, struct kmem_cache *cache, e = p + e_off; e->lc_index = i; e->lc_number = LC_FREE; + e->lc_new_number = LC_FREE; list_add(&e->list, &lc->free); element[i] = e; } @@ -175,15 +209,15 @@ void lc_reset(struct lru_cache *lc) INIT_LIST_HEAD(&lc->in_use); INIT_LIST_HEAD(&lc->lru); INIT_LIST_HEAD(&lc->free); + INIT_LIST_HEAD(&lc->to_be_changed); lc->used = 0; lc->hits = 0; lc->misses = 0; lc->starving = 0; - lc->dirty = 0; + lc->locked = 0; lc->changed = 0; + lc->pending_changes = 0; lc->flags = 0; - lc->changing_element = NULL; - lc->new_number = LC_FREE; memset(lc->lc_slot, 0, sizeof(struct hlist_head) * lc->nr_elements); for (i = 0; i < lc->nr_elements; i++) { @@ -194,6 +228,7 @@ void lc_reset(struct lru_cache *lc) /* re-init it */ e->lc_index = i; e->lc_number = LC_FREE; + e->lc_new_number = LC_FREE; list_add(&e->list, &lc->free); } } @@ -208,14 +243,14 @@ size_t lc_seq_printf_stats(struct seq_file *seq, struct lru_cache *lc) /* NOTE: * total calls to lc_get are * (starving + hits + misses) - * misses include "dirty" count (update from an other thread in + * misses include "locked" count (update from an other thread in * progress) and "changed", when this in fact lead to an successful * update of the cache. */ return seq_printf(seq, "\t%s: used:%u/%u " - "hits:%lu misses:%lu starving:%lu dirty:%lu changed:%lu\n", + "hits:%lu misses:%lu starving:%lu locked:%lu changed:%lu\n", lc->name, lc->used, lc->nr_elements, - lc->hits, lc->misses, lc->starving, lc->dirty, lc->changed); + lc->hits, lc->misses, lc->starving, lc->locked, lc->changed); } static struct hlist_head *lc_hash_slot(struct lru_cache *lc, unsigned int enr) @@ -224,16 +259,8 @@ static struct hlist_head *lc_hash_slot(struct lru_cache *lc, unsigned int enr) } -/** - * lc_find - find element by label, if present in the hash table - * @lc: The lru_cache object - * @enr: element number - * - * Returns the pointer to an element, if the element with the requested - * "label" or element number is present in the hash table, - * or NULL if not found. Does not change the refcnt. - */ -struct lc_element *lc_find(struct lru_cache *lc, unsigned int enr) +static struct lc_element *__lc_find(struct lru_cache *lc, unsigned int enr, + bool include_changing) { struct hlist_node *n; struct lc_element *e; @@ -241,29 +268,48 @@ struct lc_element *lc_find(struct lru_cache *lc, unsigned int enr) BUG_ON(!lc); BUG_ON(!lc->nr_elements); hlist_for_each_entry(e, n, lc_hash_slot(lc, enr), colision) { - if (e->lc_number == enr) + /* "about to be changed" elements, pending transaction commit, + * are hashed by their "new number". "Normal" elements have + * lc_number == lc_new_number. */ + if (e->lc_new_number != enr) + continue; + if (e->lc_new_number == e->lc_number || include_changing) return e; + break; } return NULL; } -/* returned element will be "recycled" immediately */ -static struct lc_element *lc_evict(struct lru_cache *lc) +/** + * lc_find - find element by label, if present in the hash table + * @lc: The lru_cache object + * @enr: element number + * + * Returns the pointer to an element, if the element with the requested + * "label" or element number is present in the hash table, + * or NULL if not found. Does not change the refcnt. + * Ignores elements that are "about to be used", i.e. not yet in the active + * set, but still pending transaction commit. + */ +struct lc_element *lc_find(struct lru_cache *lc, unsigned int enr) { - struct list_head *n; - struct lc_element *e; - - if (list_empty(&lc->lru)) - return NULL; - - n = lc->lru.prev; - e = list_entry(n, struct lc_element, list); - - PARANOIA_LC_ELEMENT(lc, e); + return __lc_find(lc, enr, 0); +} - list_del(&e->list); - hlist_del(&e->colision); - return e; +/** + * lc_is_used - find element by label + * @lc: The lru_cache object + * @enr: element number + * + * Returns true, if the element with the requested "label" or element number is + * present in the hash table, and is used (refcnt > 0). + * Also finds elements that are not _currently_ used but only "about to be + * used", i.e. on the "to_be_changed" list, pending transaction commit. + */ +bool lc_is_used(struct lru_cache *lc, unsigned int enr) +{ + struct lc_element *e = __lc_find(lc, enr, 1); + return e && e->refcnt; } /** @@ -280,22 +326,34 @@ void lc_del(struct lru_cache *lc, struct lc_element *e) PARANOIA_LC_ELEMENT(lc, e); BUG_ON(e->refcnt); - e->lc_number = LC_FREE; + e->lc_number = e->lc_new_number = LC_FREE; hlist_del_init(&e->colision); list_move(&e->list, &lc->free); RETURN(); } -static struct lc_element *lc_get_unused_element(struct lru_cache *lc) +static struct lc_element *lc_prepare_for_change(struct lru_cache *lc, unsigned new_number) { struct list_head *n; + struct lc_element *e; + + if (!list_empty(&lc->free)) + n = lc->free.next; + else if (!list_empty(&lc->lru)) + n = lc->lru.prev; + else + return NULL; + + e = list_entry(n, struct lc_element, list); + PARANOIA_LC_ELEMENT(lc, e); - if (list_empty(&lc->free)) - return lc_evict(lc); + e->lc_new_number = new_number; + if (!hlist_unhashed(&e->colision)) + __hlist_del(&e->colision); + hlist_add_head(&e->colision, lc_hash_slot(lc, new_number)); + list_move(&e->list, &lc->to_be_changed); - n = lc->free.next; - list_del(n); - return list_entry(n, struct lc_element, list); + return e; } static int lc_unused_element_available(struct lru_cache *lc) @@ -308,45 +366,7 @@ static int lc_unused_element_available(struct lru_cache *lc) return 0; } - -/** - * lc_get - get element by label, maybe change the active set - * @lc: the lru cache to operate on - * @enr: the label to look up - * - * Finds an element in the cache, increases its usage count, - * "touches" and returns it. - * - * In case the requested number is not present, it needs to be added to the - * cache. Therefore it is possible that an other element becomes evicted from - * the cache. In either case, the user is notified so he is able to e.g. keep - * a persistent log of the cache changes, and therefore the objects in use. - * - * Return values: - * NULL - * The cache was marked %LC_STARVING, - * or the requested label was not in the active set - * and a changing transaction is still pending (@lc was marked %LC_DIRTY). - * Or no unused or free element could be recycled (@lc will be marked as - * %LC_STARVING, blocking further lc_get() operations). - * - * pointer to the element with the REQUESTED element number. - * In this case, it can be used right away - * - * pointer to an UNUSED element with some different element number, - * where that different number may also be %LC_FREE. - * - * In this case, the cache is marked %LC_DIRTY (blocking further changes), - * and the returned element pointer is removed from the lru list and - * hash collision chains. The user now should do whatever housekeeping - * is necessary. - * Then he must call lc_changed(lc,element_pointer), to finish - * the change. - * - * NOTE: The user needs to check the lc_number on EACH use, so he recognizes - * any cache set change. - */ -struct lc_element *lc_get(struct lru_cache *lc, unsigned int enr) +static struct lc_element *__lc_get(struct lru_cache *lc, unsigned int enr, bool may_change) { struct lc_element *e; @@ -356,8 +376,12 @@ struct lc_element *lc_get(struct lru_cache *lc, unsigned int enr) RETURN(NULL); } - e = lc_find(lc, enr); - if (e) { + e = __lc_find(lc, enr, 1); + /* if lc_new_number != lc_number, + * this enr is currently being pulled in already, + * and will be available once the pending transaction + * has been committed. */ + if (e && e->lc_new_number == e->lc_number) { ++lc->hits; if (e->refcnt++ == 0) lc->used++; @@ -366,6 +390,26 @@ struct lc_element *lc_get(struct lru_cache *lc, unsigned int enr) } ++lc->misses; + if (!may_change) + RETURN(NULL); + + /* It has been found above, but on the "to_be_changed" list, not yet + * committed. Don't pull it in twice, wait for the transaction, then + * try again */ + if (e) + RETURN(NULL); + + /* To avoid races with lc_try_lock(), first, mark us dirty + * (using test_and_set_bit, as it implies memory barriers), ... */ + test_and_set_bit(__LC_DIRTY, &lc->flags); + + /* ... only then check if it is locked anyways. If lc_unlock clears + * the dirty bit again, that's not a problem, we will come here again. + */ + if (test_bit(__LC_LOCKED, &lc->flags)) { + ++lc->locked; + RETURN(NULL); + } /* In case there is nothing available and we can not kick out * the LRU element, we have to wait ... @@ -375,71 +419,109 @@ struct lc_element *lc_get(struct lru_cache *lc, unsigned int enr) RETURN(NULL); } - /* it was not present in the active set. - * we are going to recycle an unused (or even "free") element. - * user may need to commit a transaction to record that change. - * we serialize on flags & TF_DIRTY */ - if (test_and_set_bit(__LC_DIRTY, &lc->flags)) { - ++lc->dirty; + /* It was not present in the active set. We are going to recycle an + * unused (or even "free") element, but we won't accumulate more than + * max_pending_changes changes. */ + if (lc->pending_changes >= lc->max_pending_changes) RETURN(NULL); - } - e = lc_get_unused_element(lc); + e = lc_prepare_for_change(lc, enr); BUG_ON(!e); clear_bit(__LC_STARVING, &lc->flags); BUG_ON(++e->refcnt != 1); lc->used++; - - lc->changing_element = e; - lc->new_number = enr; + lc->pending_changes++; RETURN(e); } -/* similar to lc_get, - * but only gets a new reference on an existing element. - * you either get the requested element, or NULL. - * will be consolidated into one function. +/** + * lc_get - get element by label, maybe change the active set + * @lc: the lru cache to operate on + * @enr: the label to look up + * + * Finds an element in the cache, increases its usage count, + * "touches" and returns it. + * + * In case the requested number is not present, it needs to be added to the + * cache. Therefore it is possible that an other element becomes evicted from + * the cache. In either case, the user is notified so he is able to e.g. keep + * a persistent log of the cache changes, and therefore the objects in use. + * + * Return values: + * NULL + * The cache was marked %LC_STARVING, + * or the requested label was not in the active set + * and a changing transaction is still pending (@lc was marked %LC_DIRTY). + * Or no unused or free element could be recycled (@lc will be marked as + * %LC_STARVING, blocking further lc_get() operations). + * + * pointer to the element with the REQUESTED element number. + * In this case, it can be used right away + * + * pointer to an UNUSED element with some different element number, + * where that different number may also be %LC_FREE. + * + * In this case, the cache is marked %LC_DIRTY, + * so lc_try_lock() will no longer succeed. + * The returned element pointer is moved to the "to_be_changed" list, + * and registered with the new element number on the hash collision chains, + * so it is possible to pick it up from lc_is_used(). + * Up to "max_pending_changes" (see lc_create()) can be accumulated. + * The user now should do whatever housekeeping is necessary, + * typically serialize on lc_try_lock_for_transaction(), then call + * lc_committed(lc) and lc_unlock(), to finish the change. + * + * NOTE: The user needs to check the lc_number on EACH use, so he recognizes + * any cache set change. */ -struct lc_element *lc_try_get(struct lru_cache *lc, unsigned int enr) +struct lc_element *lc_get(struct lru_cache *lc, unsigned int enr) { - struct lc_element *e; - - PARANOIA_ENTRY(); - if (lc->flags & LC_STARVING) { - ++lc->starving; - RETURN(NULL); - } + return __lc_get(lc, enr, 1); +} - e = lc_find(lc, enr); - if (e) { - ++lc->hits; - if (e->refcnt++ == 0) - lc->used++; - list_move(&e->list, &lc->in_use); /* Not evictable... */ - } - RETURN(e); +/** + * lc_try_get - get element by label, if present; do not change the active set + * @lc: the lru cache to operate on + * @enr: the label to look up + * + * Finds an element in the cache, increases its usage count, + * "touches" and returns it. + * + * Return values: + * NULL + * The cache was marked %LC_STARVING, + * or the requested label was not in the active set + * + * pointer to the element with the REQUESTED element number. + * In this case, it can be used right away + */ +struct lc_element *lc_try_get(struct lru_cache *lc, unsigned int enr) +{ + return __lc_get(lc, enr, 0); } /** - * lc_changed - tell @lc that the change has been recorded + * lc_committed - tell @lc that pending changes have been recorded * @lc: the lru cache to operate on - * @e: the element pending label change + * + * User is expected to serialize on explicit lc_try_lock_for_transaction() + * before the transaction is started, and later needs to lc_unlock() explicitly + * as well. */ -void lc_changed(struct lru_cache *lc, struct lc_element *e) +void lc_committed(struct lru_cache *lc) { + struct lc_element *e, *tmp; + PARANOIA_ENTRY(); - BUG_ON(e != lc->changing_element); - PARANOIA_LC_ELEMENT(lc, e); - ++lc->changed; - e->lc_number = lc->new_number; - list_add(&e->list, &lc->in_use); - hlist_add_head(&e->colision, lc_hash_slot(lc, lc->new_number)); - lc->changing_element = NULL; - lc->new_number = LC_FREE; - clear_bit(__LC_DIRTY, &lc->flags); - smp_mb__after_clear_bit(); + list_for_each_entry_safe(e, tmp, &lc->to_be_changed, list) { + /* count number of changes, not number of transactions */ + ++lc->changed; + e->lc_number = e->lc_new_number; + list_move(&e->list, &lc->in_use); + } + lc->pending_changes = 0; RETURN(); } @@ -458,13 +540,12 @@ unsigned int lc_put(struct lru_cache *lc, struct lc_element *e) PARANOIA_ENTRY(); PARANOIA_LC_ELEMENT(lc, e); BUG_ON(e->refcnt == 0); - BUG_ON(e == lc->changing_element); + BUG_ON(e->lc_number != e->lc_new_number); if (--e->refcnt == 0) { /* move it to the front of LRU. */ list_move(&e->list, &lc->lru); lc->used--; - clear_bit(__LC_STARVING, &lc->flags); - smp_mb__after_clear_bit(); + clear_bit_unlock(__LC_STARVING, &lc->flags); } RETURN(e->refcnt); } @@ -504,16 +585,24 @@ unsigned int lc_index_of(struct lru_cache *lc, struct lc_element *e) void lc_set(struct lru_cache *lc, unsigned int enr, int index) { struct lc_element *e; + struct list_head *lh; if (index < 0 || index >= lc->nr_elements) return; e = lc_element_by_index(lc, index); - e->lc_number = enr; + BUG_ON(e->lc_number != e->lc_new_number); + BUG_ON(e->refcnt != 0); + e->lc_number = e->lc_new_number = enr; hlist_del_init(&e->colision); - hlist_add_head(&e->colision, lc_hash_slot(lc, enr)); - list_move(&e->list, e->refcnt ? &lc->in_use : &lc->lru); + if (enr == LC_FREE) + lh = &lc->free; + else { + hlist_add_head(&e->colision, lc_hash_slot(lc, enr)); + lh = &lc->lru; + } + list_move(&e->list, lh); } /** @@ -553,8 +642,10 @@ EXPORT_SYMBOL(lc_try_get); EXPORT_SYMBOL(lc_find); EXPORT_SYMBOL(lc_get); EXPORT_SYMBOL(lc_put); -EXPORT_SYMBOL(lc_changed); +EXPORT_SYMBOL(lc_committed); EXPORT_SYMBOL(lc_element_by_index); EXPORT_SYMBOL(lc_index_of); EXPORT_SYMBOL(lc_seq_printf_stats); EXPORT_SYMBOL(lc_seq_dump_details); +EXPORT_SYMBOL(lc_try_lock); +EXPORT_SYMBOL(lc_is_used); diff --git a/mm/migrate.c b/mm/migrate.c index 32efd8028bc9..3b676b0c5c3e 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -1734,7 +1734,7 @@ int migrate_misplaced_transhuge_page(struct mm_struct *mm, page_add_new_anon_rmap(new_page, vma, haddr); set_pmd_at(mm, haddr, pmd, entry); - update_mmu_cache_pmd(vma, address, entry); + update_mmu_cache_pmd(vma, address, &entry); page_remove_rmap(page); /* * Finish the charge transaction under the page table lock to diff --git a/net/core/net_namespace.c b/net/core/net_namespace.c index 6456439cbbd9..2e9a3132b8dd 100644 --- a/net/core/net_namespace.c +++ b/net/core/net_namespace.c @@ -381,6 +381,21 @@ struct net *get_net_ns_by_pid(pid_t pid) } EXPORT_SYMBOL_GPL(get_net_ns_by_pid); +static __net_init int net_ns_net_init(struct net *net) +{ + return proc_alloc_inum(&net->proc_inum); +} + +static __net_exit void net_ns_net_exit(struct net *net) +{ + proc_free_inum(net->proc_inum); +} + +static struct pernet_operations __net_initdata net_ns_ops = { + .init = net_ns_net_init, + .exit = net_ns_net_exit, +}; + static int __init net_ns_init(void) { struct net_generic *ng; @@ -412,6 +427,8 @@ static int __init net_ns_init(void) mutex_unlock(&net_mutex); + register_pernet_subsys(&net_ns_ops); + return 0; } @@ -630,16 +647,28 @@ static void netns_put(void *ns) static int netns_install(struct nsproxy *nsproxy, void *ns) { + struct net *net = ns; + + if (!ns_capable(net->user_ns, CAP_SYS_ADMIN)) + return -EPERM; + put_net(nsproxy->net_ns); - nsproxy->net_ns = get_net(ns); + nsproxy->net_ns = get_net(net); return 0; } +static unsigned int netns_inum(void *ns) +{ + struct net *net = ns; + return net->proc_inum; +} + const struct proc_ns_operations netns_operations = { .name = "net", .type = CLONE_NEWNET, .get = netns_get, .put = netns_put, .install = netns_install, + .inum = netns_inum, }; #endif diff --git a/security/yama/yama_lsm.c b/security/yama/yama_lsm.c index 2663145d1197..23414b93771f 100644 --- a/security/yama/yama_lsm.c +++ b/security/yama/yama_lsm.c @@ -298,14 +298,18 @@ int yama_ptrace_access_check(struct task_struct *child, /* No additional restrictions. */ break; case YAMA_SCOPE_RELATIONAL: + rcu_read_lock(); if (!task_is_descendant(current, child) && !ptracer_exception_found(current, child) && - !ns_capable(task_user_ns(child), CAP_SYS_PTRACE)) + !ns_capable(__task_cred(child)->user_ns, CAP_SYS_PTRACE)) rc = -EPERM; + rcu_read_unlock(); break; case YAMA_SCOPE_CAPABILITY: - if (!ns_capable(task_user_ns(child), CAP_SYS_PTRACE)) + rcu_read_lock(); + if (!ns_capable(__task_cred(child)->user_ns, CAP_SYS_PTRACE)) rc = -EPERM; + rcu_read_unlock(); break; case YAMA_SCOPE_NO_ATTACH: default: @@ -343,8 +347,10 @@ int yama_ptrace_traceme(struct task_struct *parent) /* Only disallow PTRACE_TRACEME on more aggressive settings. */ switch (ptrace_scope) { case YAMA_SCOPE_CAPABILITY: - if (!ns_capable(task_user_ns(parent), CAP_SYS_PTRACE)) + rcu_read_lock(); + if (!ns_capable(__task_cred(parent)->user_ns, CAP_SYS_PTRACE)) rc = -EPERM; + rcu_read_unlock(); break; case YAMA_SCOPE_NO_ATTACH: rc = -EPERM; |