diff options
Diffstat (limited to 'arch/arm/mm')
-rw-r--r-- | arch/arm/mm/cache-feroceon-l2.c | 37 | ||||
-rw-r--r-- | arch/arm/mm/cache-v6.S | 28 | ||||
-rw-r--r-- | arch/arm/mm/cache-v7.S | 27 | ||||
-rw-r--r-- | arch/arm/mm/cache-xsc3l2.c | 57 | ||||
-rw-r--r-- | arch/arm/mm/dma-mapping.c | 7 | ||||
-rw-r--r-- | arch/arm/mm/flush.c | 7 | ||||
-rw-r--r-- | arch/arm/mm/highmem.c | 87 | ||||
-rw-r--r-- | arch/arm/mm/proc-macros.S | 22 | ||||
-rw-r--r-- | arch/arm/mm/proc-v7.S | 4 |
9 files changed, 103 insertions, 173 deletions
diff --git a/arch/arm/mm/cache-feroceon-l2.c b/arch/arm/mm/cache-feroceon-l2.c index 6e77c042d8e9..e0b0e7a4ec68 100644 --- a/arch/arm/mm/cache-feroceon-l2.c +++ b/arch/arm/mm/cache-feroceon-l2.c @@ -13,13 +13,9 @@ */ #include <linux/init.h> +#include <linux/highmem.h> #include <asm/cacheflush.h> -#include <asm/kmap_types.h> -#include <asm/fixmap.h> -#include <asm/pgtable.h> -#include <asm/tlbflush.h> #include <plat/cache-feroceon-l2.h> -#include "mm.h" /* * Low-level cache maintenance operations. @@ -39,27 +35,30 @@ * between which we don't want to be preempted. */ -static inline unsigned long l2_start_va(unsigned long paddr) +static inline unsigned long l2_get_va(unsigned long paddr) { #ifdef CONFIG_HIGHMEM /* - * Let's do our own fixmap stuff in a minimal way here. * Because range ops can't be done on physical addresses, * we simply install a virtual mapping for it only for the * TLB lookup to occur, hence no need to flush the untouched - * memory mapping. This is protected with the disabling of - * interrupts by the caller. + * memory mapping afterwards (note: a cache flush may happen + * in some circumstances depending on the path taken in kunmap_atomic). */ - unsigned long idx = KM_L2_CACHE + KM_TYPE_NR * smp_processor_id(); - unsigned long vaddr = __fix_to_virt(FIX_KMAP_BEGIN + idx); - set_pte_ext(TOP_PTE(vaddr), pfn_pte(paddr >> PAGE_SHIFT, PAGE_KERNEL), 0); - local_flush_tlb_kernel_page(vaddr); - return vaddr + (paddr & ~PAGE_MASK); + void *vaddr = kmap_atomic_pfn(paddr >> PAGE_SHIFT); + return (unsigned long)vaddr + (paddr & ~PAGE_MASK); #else return __phys_to_virt(paddr); #endif } +static inline void l2_put_va(unsigned long vaddr) +{ +#ifdef CONFIG_HIGHMEM + kunmap_atomic((void *)vaddr); +#endif +} + static inline void l2_clean_pa(unsigned long addr) { __asm__("mcr p15, 1, %0, c15, c9, 3" : : "r" (addr)); @@ -76,13 +75,14 @@ static inline void l2_clean_pa_range(unsigned long start, unsigned long end) */ BUG_ON((start ^ end) >> PAGE_SHIFT); - raw_local_irq_save(flags); - va_start = l2_start_va(start); + va_start = l2_get_va(start); va_end = va_start + (end - start); + raw_local_irq_save(flags); __asm__("mcr p15, 1, %0, c15, c9, 4\n\t" "mcr p15, 1, %1, c15, c9, 5" : : "r" (va_start), "r" (va_end)); raw_local_irq_restore(flags); + l2_put_va(va_start); } static inline void l2_clean_inv_pa(unsigned long addr) @@ -106,13 +106,14 @@ static inline void l2_inv_pa_range(unsigned long start, unsigned long end) */ BUG_ON((start ^ end) >> PAGE_SHIFT); - raw_local_irq_save(flags); - va_start = l2_start_va(start); + va_start = l2_get_va(start); va_end = va_start + (end - start); + raw_local_irq_save(flags); __asm__("mcr p15, 1, %0, c15, c11, 4\n\t" "mcr p15, 1, %1, c15, c11, 5" : : "r" (va_start), "r" (va_end)); raw_local_irq_restore(flags); + l2_put_va(va_start); } static inline void l2_inv_all(void) diff --git a/arch/arm/mm/cache-v6.S b/arch/arm/mm/cache-v6.S index 99fa688dfadd..c96fa1b3f49f 100644 --- a/arch/arm/mm/cache-v6.S +++ b/arch/arm/mm/cache-v6.S @@ -203,6 +203,10 @@ ENTRY(v6_flush_kern_dcache_area) * - end - virtual end address of region */ v6_dma_inv_range: +#ifdef CONFIG_DMA_CACHE_RWFO + ldrb r2, [r0] @ read for ownership + strb r2, [r0] @ write for ownership +#endif tst r0, #D_CACHE_LINE_SIZE - 1 bic r0, r0, #D_CACHE_LINE_SIZE - 1 #ifdef HARVARD_CACHE @@ -211,6 +215,10 @@ v6_dma_inv_range: mcrne p15, 0, r0, c7, c11, 1 @ clean unified line #endif tst r1, #D_CACHE_LINE_SIZE - 1 +#ifdef CONFIG_DMA_CACHE_RWFO + ldrneb r2, [r1, #-1] @ read for ownership + strneb r2, [r1, #-1] @ write for ownership +#endif bic r1, r1, #D_CACHE_LINE_SIZE - 1 #ifdef HARVARD_CACHE mcrne p15, 0, r1, c7, c14, 1 @ clean & invalidate D line @@ -218,10 +226,6 @@ v6_dma_inv_range: mcrne p15, 0, r1, c7, c15, 1 @ clean & invalidate unified line #endif 1: -#ifdef CONFIG_DMA_CACHE_RWFO - ldr r2, [r0] @ read for ownership - str r2, [r0] @ write for ownership -#endif #ifdef HARVARD_CACHE mcr p15, 0, r0, c7, c6, 1 @ invalidate D line #else @@ -229,6 +233,10 @@ v6_dma_inv_range: #endif add r0, r0, #D_CACHE_LINE_SIZE cmp r0, r1 +#ifdef CONFIG_DMA_CACHE_RWFO + ldrlo r2, [r0] @ read for ownership + strlo r2, [r0] @ write for ownership +#endif blo 1b mov r0, #0 mcr p15, 0, r0, c7, c10, 4 @ drain write buffer @@ -263,12 +271,12 @@ v6_dma_clean_range: * - end - virtual end address of region */ ENTRY(v6_dma_flush_range) - bic r0, r0, #D_CACHE_LINE_SIZE - 1 -1: #ifdef CONFIG_DMA_CACHE_RWFO - ldr r2, [r0] @ read for ownership - str r2, [r0] @ write for ownership + ldrb r2, [r0] @ read for ownership + strb r2, [r0] @ write for ownership #endif + bic r0, r0, #D_CACHE_LINE_SIZE - 1 +1: #ifdef HARVARD_CACHE mcr p15, 0, r0, c7, c14, 1 @ clean & invalidate D line #else @@ -276,6 +284,10 @@ ENTRY(v6_dma_flush_range) #endif add r0, r0, #D_CACHE_LINE_SIZE cmp r0, r1 +#ifdef CONFIG_DMA_CACHE_RWFO + ldrlob r2, [r0] @ read for ownership + strlob r2, [r0] @ write for ownership +#endif blo 1b mov r0, #0 mcr p15, 0, r0, c7, c10, 4 @ drain write buffer diff --git a/arch/arm/mm/cache-v7.S b/arch/arm/mm/cache-v7.S index a3ebf7a4f49b..6136e68ce953 100644 --- a/arch/arm/mm/cache-v7.S +++ b/arch/arm/mm/cache-v7.S @@ -173,15 +173,22 @@ ENTRY(v7_coherent_user_range) UNWIND(.fnstart ) dcache_line_size r2, r3 sub r3, r2, #1 - bic r0, r0, r3 + bic r12, r0, r3 1: - USER( mcr p15, 0, r0, c7, c11, 1 ) @ clean D line to the point of unification + USER( mcr p15, 0, r12, c7, c11, 1 ) @ clean D line to the point of unification + add r12, r12, r2 + cmp r12, r1 + blo 1b dsb - USER( mcr p15, 0, r0, c7, c5, 1 ) @ invalidate I line - add r0, r0, r2 + icache_line_size r2, r3 + sub r3, r2, #1 + bic r12, r0, r3 2: - cmp r0, r1 - blo 1b + USER( mcr p15, 0, r12, c7, c5, 1 ) @ invalidate I line + add r12, r12, r2 + cmp r12, r1 + blo 2b +3: mov r0, #0 ALT_SMP(mcr p15, 0, r0, c7, c1, 6) @ invalidate BTB Inner Shareable ALT_UP(mcr p15, 0, r0, c7, c5, 6) @ invalidate BTB @@ -194,10 +201,10 @@ ENTRY(v7_coherent_user_range) * isn't mapped, just try the next page. */ 9001: - mov r0, r0, lsr #12 - mov r0, r0, lsl #12 - add r0, r0, #4096 - b 2b + mov r12, r12, lsr #12 + mov r12, r12, lsl #12 + add r12, r12, #4096 + b 3b UNWIND(.fnend ) ENDPROC(v7_coherent_kern_range) ENDPROC(v7_coherent_user_range) diff --git a/arch/arm/mm/cache-xsc3l2.c b/arch/arm/mm/cache-xsc3l2.c index c3154928bccd..5a32020471e3 100644 --- a/arch/arm/mm/cache-xsc3l2.c +++ b/arch/arm/mm/cache-xsc3l2.c @@ -17,14 +17,10 @@ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */ #include <linux/init.h> +#include <linux/highmem.h> #include <asm/system.h> #include <asm/cputype.h> #include <asm/cacheflush.h> -#include <asm/kmap_types.h> -#include <asm/fixmap.h> -#include <asm/pgtable.h> -#include <asm/tlbflush.h> -#include "mm.h" #define CR_L2 (1 << 26) @@ -71,16 +67,15 @@ static inline void xsc3_l2_inv_all(void) dsb(); } +static inline void l2_unmap_va(unsigned long va) +{ #ifdef CONFIG_HIGHMEM -#define l2_map_save_flags(x) raw_local_save_flags(x) -#define l2_map_restore_flags(x) raw_local_irq_restore(x) -#else -#define l2_map_save_flags(x) ((x) = 0) -#define l2_map_restore_flags(x) ((void)(x)) + if (va != -1) + kunmap_atomic((void *)va); #endif +} -static inline unsigned long l2_map_va(unsigned long pa, unsigned long prev_va, - unsigned long flags) +static inline unsigned long l2_map_va(unsigned long pa, unsigned long prev_va) { #ifdef CONFIG_HIGHMEM unsigned long va = prev_va & PAGE_MASK; @@ -89,17 +84,10 @@ static inline unsigned long l2_map_va(unsigned long pa, unsigned long prev_va, /* * Switching to a new page. Because cache ops are * using virtual addresses only, we must put a mapping - * in place for it. We also enable interrupts for a - * short while and disable them again to protect this - * mapping. + * in place for it. */ - unsigned long idx; - raw_local_irq_restore(flags); - idx = KM_L2_CACHE + KM_TYPE_NR * smp_processor_id(); - va = __fix_to_virt(FIX_KMAP_BEGIN + idx); - raw_local_irq_restore(flags | PSR_I_BIT); - set_pte_ext(TOP_PTE(va), pfn_pte(pa >> PAGE_SHIFT, PAGE_KERNEL), 0); - local_flush_tlb_kernel_page(va); + l2_unmap_va(prev_va); + va = (unsigned long)kmap_atomic_pfn(pa >> PAGE_SHIFT); } return va + (pa_offset >> (32 - PAGE_SHIFT)); #else @@ -109,7 +97,7 @@ static inline unsigned long l2_map_va(unsigned long pa, unsigned long prev_va, static void xsc3_l2_inv_range(unsigned long start, unsigned long end) { - unsigned long vaddr, flags; + unsigned long vaddr; if (start == 0 && end == -1ul) { xsc3_l2_inv_all(); @@ -117,13 +105,12 @@ static void xsc3_l2_inv_range(unsigned long start, unsigned long end) } vaddr = -1; /* to force the first mapping */ - l2_map_save_flags(flags); /* * Clean and invalidate partial first cache line. */ if (start & (CACHE_LINE_SIZE - 1)) { - vaddr = l2_map_va(start & ~(CACHE_LINE_SIZE - 1), vaddr, flags); + vaddr = l2_map_va(start & ~(CACHE_LINE_SIZE - 1), vaddr); xsc3_l2_clean_mva(vaddr); xsc3_l2_inv_mva(vaddr); start = (start | (CACHE_LINE_SIZE - 1)) + 1; @@ -133,7 +120,7 @@ static void xsc3_l2_inv_range(unsigned long start, unsigned long end) * Invalidate all full cache lines between 'start' and 'end'. */ while (start < (end & ~(CACHE_LINE_SIZE - 1))) { - vaddr = l2_map_va(start, vaddr, flags); + vaddr = l2_map_va(start, vaddr); xsc3_l2_inv_mva(vaddr); start += CACHE_LINE_SIZE; } @@ -142,31 +129,30 @@ static void xsc3_l2_inv_range(unsigned long start, unsigned long end) * Clean and invalidate partial last cache line. */ if (start < end) { - vaddr = l2_map_va(start, vaddr, flags); + vaddr = l2_map_va(start, vaddr); xsc3_l2_clean_mva(vaddr); xsc3_l2_inv_mva(vaddr); } - l2_map_restore_flags(flags); + l2_unmap_va(vaddr); dsb(); } static void xsc3_l2_clean_range(unsigned long start, unsigned long end) { - unsigned long vaddr, flags; + unsigned long vaddr; vaddr = -1; /* to force the first mapping */ - l2_map_save_flags(flags); start &= ~(CACHE_LINE_SIZE - 1); while (start < end) { - vaddr = l2_map_va(start, vaddr, flags); + vaddr = l2_map_va(start, vaddr); xsc3_l2_clean_mva(vaddr); start += CACHE_LINE_SIZE; } - l2_map_restore_flags(flags); + l2_unmap_va(vaddr); dsb(); } @@ -193,7 +179,7 @@ static inline void xsc3_l2_flush_all(void) static void xsc3_l2_flush_range(unsigned long start, unsigned long end) { - unsigned long vaddr, flags; + unsigned long vaddr; if (start == 0 && end == -1ul) { xsc3_l2_flush_all(); @@ -201,17 +187,16 @@ static void xsc3_l2_flush_range(unsigned long start, unsigned long end) } vaddr = -1; /* to force the first mapping */ - l2_map_save_flags(flags); start &= ~(CACHE_LINE_SIZE - 1); while (start < end) { - vaddr = l2_map_va(start, vaddr, flags); + vaddr = l2_map_va(start, vaddr); xsc3_l2_clean_mva(vaddr); xsc3_l2_inv_mva(vaddr); start += CACHE_LINE_SIZE; } - l2_map_restore_flags(flags); + l2_unmap_va(vaddr); dsb(); } diff --git a/arch/arm/mm/dma-mapping.c b/arch/arm/mm/dma-mapping.c index e24fe6d9c712..6b48e0a3d7aa 100644 --- a/arch/arm/mm/dma-mapping.c +++ b/arch/arm/mm/dma-mapping.c @@ -17,6 +17,7 @@ #include <linux/init.h> #include <linux/device.h> #include <linux/dma-mapping.h> +#include <linux/highmem.h> #include <asm/memory.h> #include <asm/highmem.h> @@ -480,10 +481,10 @@ static void dma_cache_maint_page(struct page *page, unsigned long offset, op(vaddr, len, dir); kunmap_high(page); } else if (cache_is_vipt()) { - pte_t saved_pte; - vaddr = kmap_high_l1_vipt(page, &saved_pte); + /* unmapped pages might still be cached */ + vaddr = kmap_atomic(page); op(vaddr + offset, len, dir); - kunmap_high_l1_vipt(page, saved_pte); + kunmap_atomic(vaddr); } } else { vaddr = page_address(page) + offset; diff --git a/arch/arm/mm/flush.c b/arch/arm/mm/flush.c index 391ffae75098..c29f2839f1d2 100644 --- a/arch/arm/mm/flush.c +++ b/arch/arm/mm/flush.c @@ -10,6 +10,7 @@ #include <linux/module.h> #include <linux/mm.h> #include <linux/pagemap.h> +#include <linux/highmem.h> #include <asm/cacheflush.h> #include <asm/cachetype.h> @@ -180,10 +181,10 @@ void __flush_dcache_page(struct address_space *mapping, struct page *page) __cpuc_flush_dcache_area(addr, PAGE_SIZE); kunmap_high(page); } else if (cache_is_vipt()) { - pte_t saved_pte; - addr = kmap_high_l1_vipt(page, &saved_pte); + /* unmapped pages might still be cached */ + addr = kmap_atomic(page); __cpuc_flush_dcache_area(addr, PAGE_SIZE); - kunmap_high_l1_vipt(page, saved_pte); + kunmap_atomic(addr); } } diff --git a/arch/arm/mm/highmem.c b/arch/arm/mm/highmem.c index c435fd9e1da9..807c0573abbe 100644 --- a/arch/arm/mm/highmem.c +++ b/arch/arm/mm/highmem.c @@ -140,90 +140,3 @@ struct page *kmap_atomic_to_page(const void *ptr) pte = TOP_PTE(vaddr); return pte_page(*pte); } - -#ifdef CONFIG_CPU_CACHE_VIPT - -#include <linux/percpu.h> - -/* - * The VIVT cache of a highmem page is always flushed before the page - * is unmapped. Hence unmapped highmem pages need no cache maintenance - * in that case. - * - * However unmapped pages may still be cached with a VIPT cache, and - * it is not possible to perform cache maintenance on them using physical - * addresses unfortunately. So we have no choice but to set up a temporary - * virtual mapping for that purpose. - * - * Yet this VIPT cache maintenance may be triggered from DMA support - * functions which are possibly called from interrupt context. As we don't - * want to keep interrupt disabled all the time when such maintenance is - * taking place, we therefore allow for some reentrancy by preserving and - * restoring the previous fixmap entry before the interrupted context is - * resumed. If the reentrancy depth is 0 then there is no need to restore - * the previous fixmap, and leaving the current one in place allow it to - * be reused the next time without a TLB flush (common with DMA). - */ - -static DEFINE_PER_CPU(int, kmap_high_l1_vipt_depth); - -void *kmap_high_l1_vipt(struct page *page, pte_t *saved_pte) -{ - unsigned int idx, cpu; - int *depth; - unsigned long vaddr, flags; - pte_t pte, *ptep; - - if (!in_interrupt()) - preempt_disable(); - - cpu = smp_processor_id(); - depth = &per_cpu(kmap_high_l1_vipt_depth, cpu); - - idx = KM_L1_CACHE + KM_TYPE_NR * cpu; - vaddr = __fix_to_virt(FIX_KMAP_BEGIN + idx); - ptep = TOP_PTE(vaddr); - pte = mk_pte(page, kmap_prot); - - raw_local_irq_save(flags); - (*depth)++; - if (pte_val(*ptep) == pte_val(pte)) { - *saved_pte = pte; - } else { - *saved_pte = *ptep; - set_pte_ext(ptep, pte, 0); - local_flush_tlb_kernel_page(vaddr); - } - raw_local_irq_restore(flags); - - return (void *)vaddr; -} - -void kunmap_high_l1_vipt(struct page *page, pte_t saved_pte) -{ - unsigned int idx, cpu = smp_processor_id(); - int *depth = &per_cpu(kmap_high_l1_vipt_depth, cpu); - unsigned long vaddr, flags; - pte_t pte, *ptep; - - idx = KM_L1_CACHE + KM_TYPE_NR * cpu; - vaddr = __fix_to_virt(FIX_KMAP_BEGIN + idx); - ptep = TOP_PTE(vaddr); - pte = mk_pte(page, kmap_prot); - - BUG_ON(pte_val(*ptep) != pte_val(pte)); - BUG_ON(*depth <= 0); - - raw_local_irq_save(flags); - (*depth)--; - if (*depth != 0 && pte_val(pte) != pte_val(saved_pte)) { - set_pte_ext(ptep, saved_pte, 0); - local_flush_tlb_kernel_page(vaddr); - } - raw_local_irq_restore(flags); - - if (!in_interrupt()) - preempt_enable(); -} - -#endif /* CONFIG_CPU_CACHE_VIPT */ diff --git a/arch/arm/mm/proc-macros.S b/arch/arm/mm/proc-macros.S index 337f10256cd6..f8f777df8d72 100644 --- a/arch/arm/mm/proc-macros.S +++ b/arch/arm/mm/proc-macros.S @@ -61,17 +61,27 @@ .endm /* - * cache_line_size - get the cache line size from the CSIDR register - * (available on ARMv7+). It assumes that the CSSR register was configured - * to access the L1 data cache CSIDR. + * dcache_line_size - get the minimum D-cache line size from the CTR register + * on ARMv7. */ .macro dcache_line_size, reg, tmp - mrc p15, 1, \tmp, c0, c0, 0 @ read CSIDR - and \tmp, \tmp, #7 @ cache line size encoding - mov \reg, #16 @ size offset + mrc p15, 0, \tmp, c0, c0, 1 @ read ctr + lsr \tmp, \tmp, #16 + and \tmp, \tmp, #0xf @ cache line size encoding + mov \reg, #4 @ bytes per word mov \reg, \reg, lsl \tmp @ actual cache line size .endm +/* + * icache_line_size - get the minimum I-cache line size from the CTR register + * on ARMv7. + */ + .macro icache_line_size, reg, tmp + mrc p15, 0, \tmp, c0, c0, 1 @ read ctr + and \tmp, \tmp, #0xf @ cache line size encoding + mov \reg, #4 @ bytes per word + mov \reg, \reg, lsl \tmp @ actual cache line size + .endm /* * Sanity check the PTE configuration for the code below - which makes diff --git a/arch/arm/mm/proc-v7.S b/arch/arm/mm/proc-v7.S index 2b5b20baf80d..7401f4d7e676 100644 --- a/arch/arm/mm/proc-v7.S +++ b/arch/arm/mm/proc-v7.S @@ -386,7 +386,7 @@ __v7_ca9mp_proc_info: PMD_SECT_XN | \ PMD_SECT_AP_WRITE | \ PMD_SECT_AP_READ - b __v7_ca9mp_setup + W(b) __v7_ca9mp_setup .long cpu_arch_name .long cpu_elf_name .long HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP|HWCAP_TLS @@ -418,7 +418,7 @@ __v7_proc_info: PMD_SECT_XN | \ PMD_SECT_AP_WRITE | \ PMD_SECT_AP_READ - b __v7_setup + W(b) __v7_setup .long cpu_arch_name .long cpu_elf_name .long HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP|HWCAP_TLS |