Merge branch 'cache' (early part)

29 files changed, 224 insertions, 155 deletions

diff --git a/arch/arm/common/dmabounce.c b/arch/arm/common/dmabounce.c
index 5a375e5fef21..bc90364a96c7 100644
--- a/arch/arm/common/dmabounce.c
+++ b/arch/arm/common/dmabounce.c
@@ -308,15 +308,11 @@ static inline void unmap_single(struct device *dev, dma_addr_t dma_addr,
 			memcpy(ptr, buf->safe, size);
 
 			/*
-			 * DMA buffers must have the same cache properties
-			 * as if they were really used for DMA - which means
-			 * data must be written back to RAM.  Note that
-			 * we don't use dmac_flush_range() here for the
-			 * bidirectional case because we know the cache
-			 * lines will be coherent with the data written.
+			 * Since we may have written to a page cache page,
+			 * we need to ensure that the data will be coherent
+			 * with user mappings.
 			 */
-			dmac_clean_range(ptr, ptr + size);
-			outer_clean_range(__pa(ptr), __pa(ptr) + size);
+			__cpuc_flush_kernel_dcache_area(ptr, size);
 		}
 		free_safe_buffer(dev->archdata.dmabounce, buf);
 	}
diff --git a/arch/arm/include/asm/cacheflush.h b/arch/arm/include/asm/cacheflush.h
index 73eceb87e588..730aefcfbee3 100644
--- a/arch/arm/include/asm/cacheflush.h
+++ b/arch/arm/include/asm/cacheflush.h
@@ -211,7 +211,7 @@ struct cpu_cache_fns {
 
 	void (*coherent_kern_range)(unsigned long, unsigned long);
 	void (*coherent_user_range)(unsigned long, unsigned long);
-	void (*flush_kern_dcache_page)(void *);
+	void (*flush_kern_dcache_area)(void *, size_t);
 
 	void (*dma_inv_range)(const void *, const void *);
 	void (*dma_clean_range)(const void *, const void *);
@@ -236,7 +236,7 @@ extern struct cpu_cache_fns cpu_cache;
 #define __cpuc_flush_user_range		cpu_cache.flush_user_range
 #define __cpuc_coherent_kern_range	cpu_cache.coherent_kern_range
 #define __cpuc_coherent_user_range	cpu_cache.coherent_user_range
-#define __cpuc_flush_dcache_page	cpu_cache.flush_kern_dcache_page
+#define __cpuc_flush_dcache_area	cpu_cache.flush_kern_dcache_area
 
 /*
  * These are private to the dma-mapping API.  Do not use directly.
@@ -255,14 +255,14 @@ extern struct cpu_cache_fns cpu_cache;
 #define __cpuc_flush_user_range		__glue(_CACHE,_flush_user_cache_range)
 #define __cpuc_coherent_kern_range	__glue(_CACHE,_coherent_kern_range)
 #define __cpuc_coherent_user_range	__glue(_CACHE,_coherent_user_range)
-#define __cpuc_flush_dcache_page	__glue(_CACHE,_flush_kern_dcache_page)
+#define __cpuc_flush_dcache_area	__glue(_CACHE,_flush_kern_dcache_area)
 
 extern void __cpuc_flush_kern_all(void);
 extern void __cpuc_flush_user_all(void);
 extern void __cpuc_flush_user_range(unsigned long, unsigned long, unsigned int);
 extern void __cpuc_coherent_kern_range(unsigned long, unsigned long);
 extern void __cpuc_coherent_user_range(unsigned long, unsigned long);
-extern void __cpuc_flush_dcache_page(void *);
+extern void __cpuc_flush_dcache_area(void *, size_t);
 
 /*
  * These are private to the dma-mapping API.  Do not use directly.
@@ -448,7 +448,7 @@ static inline void flush_kernel_dcache_page(struct page *page)
 {
 	/* highmem pages are always flushed upon kunmap already */
 	if ((cache_is_vivt() || cache_is_vipt_aliasing()) && !PageHighMem(page))
-		__cpuc_flush_dcache_page(page_address(page));
+		__cpuc_flush_dcache_area(page_address(page), PAGE_SIZE);
 }
 
 #define flush_dcache_mmap_lock(mapping) \
@@ -465,13 +465,6 @@ static inline void flush_kernel_dcache_page(struct page *page)
  */
 #define flush_icache_page(vma,page)	do { } while (0)
 
-static inline void flush_ioremap_region(unsigned long phys, void __iomem *virt,
-	unsigned offset, size_t size)
-{
-	const void *start = (void __force *)virt + offset;
-	dmac_inv_range(start, start + size);
-}
-
 /*
  * flush_cache_vmap() is used when creating mappings (eg, via vmap,
  * vmalloc, ioremap etc) in kernel space for pages.  On non-VIPT
diff --git a/arch/arm/mm/cache-fa.S b/arch/arm/mm/cache-fa.S
index b63a8f7b95cf..a89444a3c016 100644
--- a/arch/arm/mm/cache-fa.S
+++ b/arch/arm/mm/cache-fa.S
@@ -127,15 +127,16 @@ ENTRY(fa_coherent_user_range)
 	mov	pc, lr
 
 /*
- *	flush_kern_dcache_page(kaddr)
+ *	flush_kern_dcache_area(void *addr, size_t size)
  *
  *	Ensure that the data held in the page kaddr is written back
  *	to the page in question.
  *
- *	- kaddr   - kernel address (guaranteed to be page aligned)
+ *	- addr	- kernel address
+ *	- size	- size of region
  */
-ENTRY(fa_flush_kern_dcache_page)
-	add	r1, r0, #PAGE_SZ
+ENTRY(fa_flush_kern_dcache_area)
+	add	r1, r0, r1
 1:	mcr	p15, 0, r0, c7, c14, 1		@ clean & invalidate D line
 	add	r0, r0, #CACHE_DLINESIZE
 	cmp	r0, r1
@@ -213,7 +214,7 @@ ENTRY(fa_cache_fns)
 	.long	fa_flush_user_cache_range
 	.long	fa_coherent_kern_range
 	.long	fa_coherent_user_range
-	.long	fa_flush_kern_dcache_page
+	.long	fa_flush_kern_dcache_area
 	.long	fa_dma_inv_range
 	.long	fa_dma_clean_range
 	.long	fa_dma_flush_range
diff --git a/arch/arm/mm/cache-l2x0.c b/arch/arm/mm/cache-l2x0.c
index 747f9a9021bb..cb8fc6573b1b 100644
--- a/arch/arm/mm/cache-l2x0.c
+++ b/arch/arm/mm/cache-l2x0.c
@@ -28,69 +28,120 @@
 static void __iomem *l2x0_base;
 static DEFINE_SPINLOCK(l2x0_lock);
 
-static inline void sync_writel(unsigned long val, unsigned long reg,
-			       unsigned long complete_mask)
+static inline void cache_wait(void __iomem *reg, unsigned long mask)
 {
-	unsigned long flags;
-
-	spin_lock_irqsave(&l2x0_lock, flags);
-	writel(val, l2x0_base + reg);
 	/* wait for the operation to complete */
-	while (readl(l2x0_base + reg) & complete_mask)
+	while (readl(reg) & mask)
 		;
-	spin_unlock_irqrestore(&l2x0_lock, flags);
 }
 
 static inline void cache_sync(void)
 {
-	sync_writel(0, L2X0_CACHE_SYNC, 1);
+	void __iomem *base = l2x0_base;
+	writel(0, base + L2X0_CACHE_SYNC);
+	cache_wait(base + L2X0_CACHE_SYNC, 1);
 }
 
 static inline void l2x0_inv_all(void)
 {
+	unsigned long flags;
+
 	/* invalidate all ways */
-	sync_writel(0xff, L2X0_INV_WAY, 0xff);
+	spin_lock_irqsave(&l2x0_lock, flags);
+	writel(0xff, l2x0_base + L2X0_INV_WAY);
+	cache_wait(l2x0_base + L2X0_INV_WAY, 0xff);
 	cache_sync();
+	spin_unlock_irqrestore(&l2x0_lock, flags);
 }
 
 static void l2x0_inv_range(unsigned long start, unsigned long end)
 {
-	unsigned long addr;
+	void __iomem *base = l2x0_base;
+	unsigned long flags;
 
+	spin_lock_irqsave(&l2x0_lock, flags);
 	if (start & (CACHE_LINE_SIZE - 1)) {
 		start &= ~(CACHE_LINE_SIZE - 1);
-		sync_writel(start, L2X0_CLEAN_INV_LINE_PA, 1);
+		cache_wait(base + L2X0_CLEAN_INV_LINE_PA, 1);
+		writel(start, base + L2X0_CLEAN_INV_LINE_PA);
 		start += CACHE_LINE_SIZE;
 	}
 
 	if (end & (CACHE_LINE_SIZE - 1)) {
 		end &= ~(CACHE_LINE_SIZE - 1);
-		sync_writel(end, L2X0_CLEAN_INV_LINE_PA, 1);
+		cache_wait(base + L2X0_CLEAN_INV_LINE_PA, 1);
+		writel(end, base + L2X0_CLEAN_INV_LINE_PA);
 	}
 
-	for (addr = start; addr < end; addr += CACHE_LINE_SIZE)
-		sync_writel(addr, L2X0_INV_LINE_PA, 1);
+	while (start < end) {
+		unsigned long blk_end = start + min(end - start, 4096UL);
+
+		while (start < blk_end) {
+			cache_wait(base + L2X0_INV_LINE_PA, 1);
+			writel(start, base + L2X0_INV_LINE_PA);
+			start += CACHE_LINE_SIZE;
+		}
+
+		if (blk_end < end) {
+			spin_unlock_irqrestore(&l2x0_lock, flags);
+			spin_lock_irqsave(&l2x0_lock, flags);
+		}
+	}
+	cache_wait(base + L2X0_INV_LINE_PA, 1);
 	cache_sync();
+	spin_unlock_irqrestore(&l2x0_lock, flags);
 }
 
 static void l2x0_clean_range(unsigned long start, unsigned long end)
 {
-	unsigned long addr;
+	void __iomem *base = l2x0_base;
+	unsigned long flags;
 
+	spin_lock_irqsave(&l2x0_lock, flags);
 	start &= ~(CACHE_LINE_SIZE - 1);
-	for (addr = start; addr < end; addr += CACHE_LINE_SIZE)
-		sync_writel(addr, L2X0_CLEAN_LINE_PA, 1);
+	while (start < end) {
+		unsigned long blk_end = start + min(end - start, 4096UL);
+
+		while (start < blk_end) {
+			cache_wait(base + L2X0_CLEAN_LINE_PA, 1);
+			writel(start, base + L2X0_CLEAN_LINE_PA);
+			start += CACHE_LINE_SIZE;
+		}
+
+		if (blk_end < end) {
+			spin_unlock_irqrestore(&l2x0_lock, flags);
+			spin_lock_irqsave(&l2x0_lock, flags);
+		}
+	}
+	cache_wait(base + L2X0_CLEAN_LINE_PA, 1);
 	cache_sync();
+	spin_unlock_irqrestore(&l2x0_lock, flags);
 }
 
 static void l2x0_flush_range(unsigned long start, unsigned long end)
 {
-	unsigned long addr;
+	void __iomem *base = l2x0_base;
+	unsigned long flags;
 
+	spin_lock_irqsave(&l2x0_lock, flags);
 	start &= ~(CACHE_LINE_SIZE - 1);
-	for (addr = start; addr < end; addr += CACHE_LINE_SIZE)
-		sync_writel(addr, L2X0_CLEAN_INV_LINE_PA, 1);
+	while (start < end) {
+		unsigned long blk_end = start + min(end - start, 4096UL);
+
+		while (start < blk_end) {
+			cache_wait(base + L2X0_CLEAN_INV_LINE_PA, 1);
+			writel(start, base + L2X0_CLEAN_INV_LINE_PA);
+			start += CACHE_LINE_SIZE;
+		}
+
+		if (blk_end < end) {
+			spin_unlock_irqrestore(&l2x0_lock, flags);
+			spin_lock_irqsave(&l2x0_lock, flags);
+		}
+	}
+	cache_wait(base + L2X0_CLEAN_INV_LINE_PA, 1);
 	cache_sync();
+	spin_unlock_irqrestore(&l2x0_lock, flags);
 }
 
 void __init l2x0_init(void __iomem *base, __u32 aux_val, __u32 aux_mask)
diff --git a/arch/arm/mm/cache-v3.S b/arch/arm/mm/cache-v3.S
index 8a4abebc478a..2a482731ea36 100644
--- a/arch/arm/mm/cache-v3.S
+++ b/arch/arm/mm/cache-v3.S
@@ -72,14 +72,15 @@ ENTRY(v3_coherent_user_range)
 	mov	pc, lr
 
 /*
- *	flush_kern_dcache_page(void *page)
+ *	flush_kern_dcache_area(void *page, size_t size)
  *
  *	Ensure no D cache aliasing occurs, either with itself or
  *	the I cache
  *
- *	- addr	- page aligned address
+ *	- addr	- kernel address
+ *	- size	- region size
  */
-ENTRY(v3_flush_kern_dcache_page)
+ENTRY(v3_flush_kern_dcache_area)
 	/* FALLTHROUGH */
 
 /*
@@ -129,7 +130,7 @@ ENTRY(v3_cache_fns)
 	.long	v3_flush_user_cache_range
 	.long	v3_coherent_kern_range
 	.long	v3_coherent_user_range
-	.long	v3_flush_kern_dcache_page
+	.long	v3_flush_kern_dcache_area
 	.long	v3_dma_inv_range
 	.long	v3_dma_clean_range
 	.long	v3_dma_flush_range
diff --git a/arch/arm/mm/cache-v4.S b/arch/arm/mm/cache-v4.S
index 3668611cb400..5c7da3e372e9 100644
--- a/arch/arm/mm/cache-v4.S
+++ b/arch/arm/mm/cache-v4.S
@@ -82,14 +82,15 @@ ENTRY(v4_coherent_user_range)
 	mov	pc, lr
 
 /*
- *	flush_kern_dcache_page(void *page)
+ *	flush_kern_dcache_area(void *addr, size_t size)
  *
  *	Ensure no D cache aliasing occurs, either with itself or
  *	the I cache
  *
- *	- addr	- page aligned address
+ *	- addr	- kernel address
+ *	- size	- region size
  */
-ENTRY(v4_flush_kern_dcache_page)
+ENTRY(v4_flush_kern_dcache_area)
 	/* FALLTHROUGH */
 
 /*
@@ -141,7 +142,7 @@ ENTRY(v4_cache_fns)
 	.long	v4_flush_user_cache_range
 	.long	v4_coherent_kern_range
 	.long	v4_coherent_user_range
-	.long	v4_flush_kern_dcache_page
+	.long	v4_flush_kern_dcache_area
 	.long	v4_dma_inv_range
 	.long	v4_dma_clean_range
 	.long	v4_dma_flush_range
diff --git a/arch/arm/mm/cache-v4wb.S b/arch/arm/mm/cache-v4wb.S
index 2ebc1b3bf856..3dbedf1ec0e7 100644
--- a/arch/arm/mm/cache-v4wb.S
+++ b/arch/arm/mm/cache-v4wb.S
@@ -114,15 +114,16 @@ ENTRY(v4wb_flush_user_cache_range)
 	mov	pc, lr
 
 /*
- *	flush_kern_dcache_page(void *page)
+ *	flush_kern_dcache_area(void *addr, size_t size)
  *
  *	Ensure no D cache aliasing occurs, either with itself or
  *	the I cache
  *
- *	- addr	- page aligned address
+ *	- addr	- kernel address
+ *	- size	- region size
  */
-ENTRY(v4wb_flush_kern_dcache_page)
-	add	r1, r0, #PAGE_SZ
+ENTRY(v4wb_flush_kern_dcache_area)
+	add	r1, r0, r1
 	/* fall through */
 
 /*
@@ -224,7 +225,7 @@ ENTRY(v4wb_cache_fns)
 	.long	v4wb_flush_user_cache_range
 	.long	v4wb_coherent_kern_range
 	.long	v4wb_coherent_user_range
-	.long	v4wb_flush_kern_dcache_page
+	.long	v4wb_flush_kern_dcache_area
 	.long	v4wb_dma_inv_range
 	.long	v4wb_dma_clean_range
 	.long	v4wb_dma_flush_range
diff --git a/arch/arm/mm/cache-v4wt.S b/arch/arm/mm/cache-v4wt.S
index c54fa2cc40e6..b3b7410270b4 100644
--- a/arch/arm/mm/cache-v4wt.S
+++ b/arch/arm/mm/cache-v4wt.S
@@ -117,17 +117,18 @@ ENTRY(v4wt_coherent_user_range)
 	mov	pc, lr
 
 /*
- *	flush_kern_dcache_page(void *page)
+ *	flush_kern_dcache_area(void *addr, size_t size)
  *
  *	Ensure no D cache aliasing occurs, either with itself or
  *	the I cache
  *
- *	- addr	- page aligned address
+ *	- addr	- kernel address
+ *	- size	- region size
  */
-ENTRY(v4wt_flush_kern_dcache_page)
+ENTRY(v4wt_flush_kern_dcache_area)
 	mov	r2, #0
 	mcr	p15, 0, r2, c7, c5, 0		@ invalidate I cache
-	add	r1, r0, #PAGE_SZ
+	add	r1, r0, r1
 	/* fallthrough */
 
 /*
@@ -180,7 +181,7 @@ ENTRY(v4wt_cache_fns)
 	.long	v4wt_flush_user_cache_range
 	.long	v4wt_coherent_kern_range
 	.long	v4wt_coherent_user_range
-	.long	v4wt_flush_kern_dcache_page
+	.long	v4wt_flush_kern_dcache_area
 	.long	v4wt_dma_inv_range
 	.long	v4wt_dma_clean_range
 	.long	v4wt_dma_flush_range
diff --git a/arch/arm/mm/cache-v6.S b/arch/arm/mm/cache-v6.S
index 295e25dd6381..4ba0a24ce6f5 100644
--- a/arch/arm/mm/cache-v6.S
+++ b/arch/arm/mm/cache-v6.S
@@ -159,15 +159,16 @@ ENDPROC(v6_coherent_user_range)
 ENDPROC(v6_coherent_kern_range)
 
 /*
- *	v6_flush_kern_dcache_page(kaddr)
+ *	v6_flush_kern_dcache_area(void *addr, size_t size)
  *
  *	Ensure that the data held in the page kaddr is written back
  *	to the page in question.
  *
- *	- kaddr   - kernel address (guaranteed to be page aligned)
+ *	- addr	- kernel address
+ *	- size	- region size
  */
-ENTRY(v6_flush_kern_dcache_page)
-	add	r1, r0, #PAGE_SZ
+ENTRY(v6_flush_kern_dcache_area)
+	add	r1, r0, r1
 1:
 #ifdef HARVARD_CACHE
 	mcr	p15, 0, r0, c7, c14, 1		@ clean & invalidate D line
@@ -271,7 +272,7 @@ ENTRY(v6_cache_fns)
 	.long	v6_flush_user_cache_range
 	.long	v6_coherent_kern_range
 	.long	v6_coherent_user_range
-	.long	v6_flush_kern_dcache_page
+	.long	v6_flush_kern_dcache_area
 	.long	v6_dma_inv_range
 	.long	v6_dma_clean_range
 	.long	v6_dma_flush_range
diff --git a/arch/arm/mm/cache-v7.S b/arch/arm/mm/cache-v7.S
index e1bd9759617f..9073db849fb4 100644
--- a/arch/arm/mm/cache-v7.S
+++ b/arch/arm/mm/cache-v7.S
@@ -186,16 +186,17 @@ ENDPROC(v7_coherent_kern_range)
 ENDPROC(v7_coherent_user_range)
 
 /*
- *	v7_flush_kern_dcache_page(kaddr)
+ *	v7_flush_kern_dcache_area(void *addr, size_t size)
  *
  *	Ensure that the data held in the page kaddr is written back
  *	to the page in question.
  *
- *	- kaddr   - kernel address (guaranteed to be page aligned)
+ *	- addr	- kernel address
+ *	- size	- region size
  */
-ENTRY(v7_flush_kern_dcache_page)
+ENTRY(v7_flush_kern_dcache_area)
 	dcache_line_size r2, r3
-	add	r1, r0, #PAGE_SZ
+	add	r1, r0, r1
 1:
 	mcr	p15, 0, r0, c7, c14, 1		@ clean & invalidate D line / unified line
 	add	r0, r0, r2
@@ -203,7 +204,7 @@ ENTRY(v7_flush_kern_dcache_page)
 	blo	1b
 	dsb
 	mov	pc, lr
-ENDPROC(v7_flush_kern_dcache_page)
+ENDPROC(v7_flush_kern_dcache_area)
 
 /*
  *	v7_dma_inv_range(start,end)
@@ -279,7 +280,7 @@ ENTRY(v7_cache_fns)
 	.long	v7_flush_user_cache_range
 	.long	v7_coherent_kern_range
 	.long	v7_coherent_user_range
-	.long	v7_flush_kern_dcache_page
+	.long	v7_flush_kern_dcache_area
 	.long	v7_dma_inv_range
 	.long	v7_dma_clean_range
 	.long	v7_dma_flush_range
diff --git a/arch/arm/mm/flush.c b/arch/arm/mm/flush.c
index 329594e760cd..6f3a4b7a3b82 100644
--- a/arch/arm/mm/flush.c
+++ b/arch/arm/mm/flush.c
@@ -131,7 +131,7 @@ void __flush_dcache_page(struct address_space *mapping, struct page *page)
 	 */
 	if (addr)
 #endif
-		__cpuc_flush_dcache_page(addr);
+		__cpuc_flush_dcache_area(addr, PAGE_SIZE);
 
 	/*
 	 * If this is a page cache page, and we have an aliasing VIPT cache,
@@ -258,5 +258,5 @@ void __flush_anon_page(struct vm_area_struct *vma, struct page *page, unsigned l
 	 * in this mapping of the page.  FIXME: this is overkill
 	 * since we actually ask for a write-back and invalidate.
 	 */
-	__cpuc_flush_dcache_page(page_address(page));
+	__cpuc_flush_dcache_area(page_address(page), PAGE_SIZE);
 }
diff --git a/arch/arm/mm/highmem.c b/arch/arm/mm/highmem.c
index 30f82fb5918c..2be1ec7c1b41 100644
--- a/arch/arm/mm/highmem.c
+++ b/arch/arm/mm/highmem.c
@@ -79,7 +79,7 @@ void kunmap_atomic(void *kvaddr, enum km_type type)
 	unsigned int idx = type + KM_TYPE_NR * smp_processor_id();
 
 	if (kvaddr >= (void *)FIXADDR_START) {
-		__cpuc_flush_dcache_page((void *)vaddr);
+		__cpuc_flush_dcache_area((void *)vaddr, PAGE_SIZE);
 #ifdef CONFIG_DEBUG_HIGHMEM
 		BUG_ON(vaddr != __fix_to_virt(FIX_KMAP_BEGIN + idx));
 		set_pte_ext(TOP_PTE(vaddr), __pte(0), 0);
diff --git a/arch/arm/mm/nommu.c b/arch/arm/mm/nommu.c
index 900811cc9130..374a8311bc84 100644
--- a/arch/arm/mm/nommu.c
+++ b/arch/arm/mm/nommu.c
@@ -61,7 +61,7 @@ void setup_mm_for_reboot(char mode)
 
 void flush_dcache_page(struct page *page)
 {
-	__cpuc_flush_dcache_page(page_address(page));
+	__cpuc_flush_dcache_area(page_address(page), PAGE_SIZE);
 }
 EXPORT_SYMBOL(flush_dcache_page);
 
diff --git a/arch/arm/mm/proc-arm1020.S b/arch/arm/mm/proc-arm1020.S
index d9fb4b98c49f..8012e24282b2 100644
--- a/arch/arm/mm/proc-arm1020.S
+++ b/arch/arm/mm/proc-arm1020.S
@@ -231,17 +231,18 @@ ENTRY(arm1020_coherent_user_range)
 	mov	pc, lr
 
 /*
- *	flush_kern_dcache_page(void *page)
+ *	flush_kern_dcache_area(void *addr, size_t size)
  *
  *	Ensure no D cache aliasing occurs, either with itself or
  *	the I cache
  *
- *	- page	- page aligned address
+ *	- addr	- kernel address
+ *	- size	- region size
  */
-ENTRY(arm1020_flush_kern_dcache_page)
+ENTRY(arm1020_flush_kern_dcache_area)
 	mov	ip, #0
 #ifndef CONFIG_CPU_DCACHE_DISABLE
-	add	r1, r0, #PAGE_SZ
+	add	r1, r0, r1
 1:	mcr	p15, 0, r0, c7, c14, 1		@ clean+invalidate D entry
 	mcr	p15, 0, ip, c7, c10, 4		@ drain WB
 	add	r0, r0, #CACHE_DLINESIZE
@@ -335,7 +336,7 @@ ENTRY(arm1020_cache_fns)
 	.long	arm1020_flush_user_cache_range
 	.long	arm1020_coherent_kern_range
 	.long	arm1020_coherent_user_range
-	.long	arm1020_flush_kern_dcache_page
+	.long	arm1020_flush_kern_dcache_area
 	.long	arm1020_dma_inv_range
 	.long	arm1020_dma_clean_range
 	.long	arm1020_dma_flush_range
diff --git a/arch/arm/mm/proc-arm1020e.S b/arch/arm/mm/proc-arm1020e.S
index 7453b75dcea5..41fe25d234f5 100644
--- a/arch/arm/mm/proc-arm1020e.S
+++ b/arch/arm/mm/proc-arm1020e.S
@@ -225,17 +225,18 @@ ENTRY(arm1020e_coherent_user_range)
 	mov	pc, lr
 
 /*
- *	flush_kern_dcache_page(void *page)
+ *	flush_kern_dcache_area(void *addr, size_t size)
  *
  *	Ensure no D cache aliasing occurs, either with itself or
  *	the I cache
  *
- *	- page	- page aligned address
+ *	- addr	- kernel address
+ *	- size	- region size
  */
-ENTRY(arm1020e_flush_kern_dcache_page)
+ENTRY(arm1020e_flush_kern_dcache_area)
 	mov	ip, #0
 #ifndef CONFIG_CPU_DCACHE_DISABLE
-	add	r1, r0, #PAGE_SZ
+	add	r1, r0, r1
 1:	mcr	p15, 0, r0, c7, c14, 1		@ clean+invalidate D entry
 	add	r0, r0, #CACHE_DLINESIZE
 	cmp	r0, r1
@@ -321,7 +322,7 @@ ENTRY(arm1020e_cache_fns)
 	.long	arm1020e_flush_user_cache_range
 	.long	arm1020e_coherent_kern_range
 	.long	arm1020e_coherent_user_range
-	.long	arm1020e_flush_kern_dcache_page
+	.long	arm1020e_flush_kern_dcache_area
 	.long	arm1020e_dma_inv_range
 	.long	arm1020e_dma_clean_range
 	.long	arm1020e_dma_flush_range
diff --git a/arch/arm/mm/proc-arm1022.S b/arch/arm/mm/proc-arm1022.S
index 8eb72d75a8b6..20a5b1b31a70 100644
--- a/arch/arm/mm/proc-arm1022.S
+++ b/arch/arm/mm/proc-arm1022.S
@@ -214,17 +214,18 @@ ENTRY(arm1022_coherent_user_range)
 	mov	pc, lr
 
 /*
- *	flush_kern_dcache_page(void *page)
+ *	flush_kern_dcache_area(void *addr, size_t size)
  *
  *	Ensure no D cache aliasing occurs, either with itself or
  *	the I cache
  *
- *	- page	- page aligned address
+ *	- addr	- kernel address
+ *	- size	- region size
  */
-ENTRY(arm1022_flush_kern_dcache_page)
+ENTRY(arm1022_flush_kern_dcache_area)
 	mov	ip, #0
 #ifndef CONFIG_CPU_DCACHE_DISABLE
-	add	r1, r0, #PAGE_SZ
+	add	r1, r0, r1
 1:	mcr	p15, 0, r0, c7, c14, 1		@ clean+invalidate D entry
 	add	r0, r0, #CACHE_DLINESIZE
 	cmp	r0, r1
@@ -310,7 +311,7 @@ ENTRY(arm1022_cache_fns)
 	.long	arm1022_flush_user_cache_range
 	.long	arm1022_coherent_kern_range
 	.long	arm1022_coherent_user_range
-	.long	arm1022_flush_kern_dcache_page
+	.long	arm1022_flush_kern_dcache_area
 	.long	arm1022_dma_inv_range
 	.long	arm1022_dma_clean_range
 	.long	arm1022_dma_flush_range
diff --git a/arch/arm/mm/proc-arm1026.S b/arch/arm/mm/proc-arm1026.S
index 3b59f0d67139..96aedb10fcc4 100644
--- a/arch/arm/mm/proc-arm1026.S
+++ b/arch/arm/mm/proc-arm1026.S
@@ -208,17 +208,18 @@ ENTRY(arm1026_coherent_user_range)
 	mov	pc, lr
 
 /*
- *	flush_kern_dcache_page(void *page)
+ *	flush_kern_dcache_area(void *addr, size_t size)
  *
  *	Ensure no D cache aliasing occurs, either with itself or
  *	the I cache
  *
- *	- page	- page aligned address
+ *	- addr	- kernel address
+ *	- size	- region size
  */
-ENTRY(arm1026_flush_kern_dcache_page)
+ENTRY(arm1026_flush_kern_dcache_area)
 	mov	ip, #0
 #ifndef CONFIG_CPU_DCACHE_DISABLE
-	add	r1, r0, #PAGE_SZ
+	add	r1, r0, r1
 1:	mcr	p15, 0, r0, c7, c14, 1		@ clean+invalidate D entry
 	add	r0, r0, #CACHE_DLINESIZE
 	cmp	r0, r1
@@ -304,7 +305,7 @@ ENTRY(arm1026_cache_fns)
 	.long	arm1026_flush_user_cache_range
 	.long	arm1026_coherent_kern_range
 	.long	arm1026_coherent_user_range
-	.long	arm1026_flush_kern_dcache_page
+	.long	arm1026_flush_kern_dcache_area
 	.long	arm1026_dma_inv_range
 	.long	arm1026_dma_clean_range
 	.long	arm1026_dma_flush_range
diff --git a/arch/arm/mm/proc-arm920.S b/arch/arm/mm/proc-arm920.S
index 2b7c197cc58d..471669e2d7cb 100644
--- a/arch/arm/mm/proc-arm920.S
+++ b/arch/arm/mm/proc-arm920.S
@@ -207,15 +207,16 @@ ENTRY(arm920_coherent_user_range)
 	mov	pc, lr
 
 /*
- *	flush_kern_dcache_page(void *page)
+ *	flush_kern_dcache_area(void *addr, size_t size)
  *
  *	Ensure no D cache aliasing occurs, either with itself or
  *	the I cache
  *
- *	- addr	- page aligned address
+ *	- addr	- kernel address
+ *	- size	- region size
  */
-ENTRY(arm920_flush_kern_dcache_page)
-	add	r1, r0, #PAGE_SZ
+ENTRY(arm920_flush_kern_dcache_area)
+	add	r1, r0, r1
 1:	mcr	p15, 0, r0, c7, c14, 1		@ clean+invalidate D entry
 	add	r0, r0, #CACHE_DLINESIZE
 	cmp	r0, r1
@@ -293,7 +294,7 @@ ENTRY(arm920_cache_fns)
 	.long	arm920_flush_user_cache_range
 	.long	arm920_coherent_kern_range
 	.long	arm920_coherent_user_range
-	.long	arm920_flush_kern_dcache_page
+	.long	arm920_flush_kern_dcache_area
 	.long	arm920_dma_inv_range
 	.long	arm920_dma_clean_range
 	.long	arm920_dma_flush_range
diff --git a/arch/arm/mm/proc-arm922.S b/arch/arm/mm/proc-arm922.S
index 06a1aa4e3398..ee111b00fa41 100644
--- a/arch/arm/mm/proc-arm922.S
+++ b/arch/arm/mm/proc-arm922.S
@@ -209,15 +209,16 @@ ENTRY(arm922_coherent_user_range)
 	mov	pc, lr
 
 /*
- *	flush_kern_dcache_page(void *page)
+ *	flush_kern_dcache_area(void *addr, size_t size)
  *
  *	Ensure no D cache aliasing occurs, either with itself or
  *	the I cache
  *
- *	- addr	- page aligned address
+ *	- addr	- kernel address
+ *	- size	- region size
  */
-ENTRY(arm922_flush_kern_dcache_page)
-	add	r1, r0, #PAGE_SZ
+ENTRY(arm922_flush_kern_dcache_area)
+	add	r1, r0, r1
 1:	mcr	p15, 0, r0, c7, c14, 1		@ clean+invalidate D entry
 	add	r0, r0, #CACHE_DLINESIZE
 	cmp	r0, r1
@@ -295,7 +296,7 @@ ENTRY(arm922_cache_fns)
 	.long	arm922_flush_user_cache_range
 	.long	arm922_coherent_kern_range
 	.long	arm922_coherent_user_range
-	.long	arm922_flush_kern_dcache_page
+	.long	arm922_flush_kern_dcache_area
 	.long	arm922_dma_inv_range
 	.long	arm922_dma_clean_range
 	.long	arm922_dma_flush_range
diff --git a/arch/arm/mm/proc-arm925.S b/arch/arm/mm/proc-arm925.S
index cb53435a85ae..8deb5bde58e4 100644
--- a/arch/arm/mm/proc-arm925.S
+++ b/arch/arm/mm/proc-arm925.S
@@ -251,15 +251,16 @@ ENTRY(arm925_coherent_user_range)
 	mov	pc, lr
 
 /*
- *	flush_kern_dcache_page(void *page)
+ *	flush_kern_dcache_area(void *addr, size_t size)
  *
  *	Ensure no D cache aliasing occurs, either with itself or
  *	the I cache
  *
- *	- addr	- page aligned address
+ *	- addr	- kernel address
+ *	- size	- region size
  */
-ENTRY(arm925_flush_kern_dcache_page)
-	add	r1, r0, #PAGE_SZ
+ENTRY(arm925_flush_kern_dcache_area)
+	add	r1, r0, r1
 1:	mcr	p15, 0, r0, c7, c14, 1		@ clean+invalidate D entry
 	add	r0, r0, #CACHE_DLINESIZE
 	cmp	r0, r1
@@ -346,7 +347,7 @@ ENTRY(arm925_cache_fns)
 	.long	arm925_flush_user_cache_range
 	.long	arm925_coherent_kern_range
 	.long	arm925_coherent_user_range
-	.long	arm925_flush_kern_dcache_page
+	.long	arm925_flush_kern_dcache_area
 	.long	arm925_dma_inv_range
 	.long	arm925_dma_clean_range
 	.long	arm925_dma_flush_range
diff --git a/arch/arm/mm/proc-arm926.S b/arch/arm/mm/proc-arm926.S
index 1c4848704bb3..64db6e275a44 100644
--- a/arch/arm/mm/proc-arm926.S
+++ b/arch/arm/mm/proc-arm926.S
@@ -214,15 +214,16 @@ ENTRY(arm926_coherent_user_range)
 	mov	pc, lr
 
 /*
- *	flush_kern_dcache_page(void *page)
+ *	flush_kern_dcache_area(void *addr, size_t size)
  *
  *	Ensure no D cache aliasing occurs, either with itself or
  *	the I cache
  *
- *	- addr	- page aligned address
+ *	- addr	- kernel address
+ *	- size	- region size
  */
-ENTRY(arm926_flush_kern_dcache_page)
-	add	r1, r0, #PAGE_SZ
+ENTRY(arm926_flush_kern_dcache_area)
+	add	r1, r0, r1
 1:	mcr	p15, 0, r0, c7, c14, 1		@ clean+invalidate D entry
 	add	r0, r0, #CACHE_DLINESIZE
 	cmp	r0, r1
@@ -309,7 +310,7 @@ ENTRY(arm926_cache_fns)
 	.long	arm926_flush_user_cache_range
 	.long	arm926_coherent_kern_range
 	.long	arm926_coherent_user_range
-	.long	arm926_flush_kern_dcache_page
+	.long	arm926_flush_kern_dcache_area
 	.long	arm926_dma_inv_range
 	.long	arm926_dma_clean_range
 	.long	arm926_dma_flush_range
diff --git a/arch/arm/mm/proc-arm940.S b/arch/arm/mm/proc-arm940.S
index 5b0f8464c8f2..8196b9f401fb 100644
--- a/arch/arm/mm/proc-arm940.S
+++ b/arch/arm/mm/proc-arm940.S
@@ -141,14 +141,15 @@ ENTRY(arm940_coherent_user_range)
 	/* FALLTHROUGH */
 
 /*
- *	flush_kern_dcache_page(void *page)
+ *	flush_kern_dcache_area(void *addr, size_t size)
  *
  *	Ensure no D cache aliasing occurs, either with itself or
  *	the I cache
  *
- *	- addr	- page aligned address
+ *	- addr	- kernel address
+ *	- size	- region size
  */
-ENTRY(arm940_flush_kern_dcache_page)
+ENTRY(arm940_flush_kern_dcache_area)
 	mov	ip, #0
 	mov	r1, #(CACHE_DSEGMENTS - 1) << 4	@ 4 segments
 1:	orr	r3, r1, #(CACHE_DENTRIES - 1) << 26 @ 64 entries
@@ -238,7 +239,7 @@ ENTRY(arm940_cache_fns)
 	.long	arm940_flush_user_cache_range
 	.long	arm940_coherent_kern_range
 	.long	arm940_coherent_user_range
-	.long	arm940_flush_kern_dcache_page
+	.long	arm940_flush_kern_dcache_area
 	.long	arm940_dma_inv_range
 	.long	arm940_dma_clean_range
 	.long	arm940_dma_flush_range
diff --git a/arch/arm/mm/proc-arm946.S b/arch/arm/mm/proc-arm946.S
index 40c0449a139b..9a951239c86c 100644
--- a/arch/arm/mm/proc-arm946.S
+++ b/arch/arm/mm/proc-arm946.S
@@ -183,16 +183,17 @@ ENTRY(arm946_coherent_user_range)
 	mov	pc, lr
 
 /*
- *	flush_kern_dcache_page(void *page)
+ *	flush_kern_dcache_area(void *addr, size_t size)
  *
  *	Ensure no D cache aliasing occurs, either with itself or
  *	the I cache
  *
- *	- addr	- page aligned address
+ *	- addr	- kernel address
+ *	- size	- region size
  * (same as arm926)
  */
-ENTRY(arm946_flush_kern_dcache_page)
-	add	r1, r0, #PAGE_SZ
+ENTRY(arm946_flush_kern_dcache_area)
+	add	r1, r0, r1
 1:	mcr	p15, 0, r0, c7, c14, 1		@ clean+invalidate D entry
 	add	r0, r0, #CACHE_DLINESIZE
 	cmp	r0, r1
@@ -280,7 +281,7 @@ ENTRY(arm946_cache_fns)
 	.long	arm946_flush_user_cache_range
 	.long	arm946_coherent_kern_range
 	.long	arm946_coherent_user_range
-	.long	arm946_flush_kern_dcache_page
+	.long	arm946_flush_kern_dcache_area
 	.long	arm946_dma_inv_range
 	.long	arm946_dma_clean_range
 	.long	arm946_dma_flush_range
diff --git a/arch/arm/mm/proc-feroceon.S b/arch/arm/mm/proc-feroceon.S
index d0d7795200fc..dbc39383e66a 100644
--- a/arch/arm/mm/proc-feroceon.S
+++ b/arch/arm/mm/proc-feroceon.S
@@ -226,16 +226,17 @@ ENTRY(feroceon_coherent_user_range)
 	mov	pc, lr
 
 /*
- *	flush_kern_dcache_page(void *page)
+ *	flush_kern_dcache_area(void *addr, size_t size)
  *
  *	Ensure no D cache aliasing occurs, either with itself or
  *	the I cache
  *
- *	- addr	- page aligned address
+ *	- addr	- kernel address
+ *	- size	- region size
  */
 	.align	5
-ENTRY(feroceon_flush_kern_dcache_page)
-	add	r1, r0, #PAGE_SZ
+ENTRY(feroceon_flush_kern_dcache_area)
+	add	r1, r0, r1
 1:	mcr	p15, 0, r0, c7, c14, 1		@ clean+invalidate D entry
 	add	r0, r0, #CACHE_DLINESIZE
 	cmp	r0, r1
@@ -246,7 +247,7 @@ ENTRY(feroceon_flush_kern_dcache_page)
 	mov	pc, lr
 
 	.align	5
-ENTRY(feroceon_range_flush_kern_dcache_page)
+ENTRY(feroceon_range_flush_kern_dcache_area)
 	mrs	r2, cpsr
 	add	r1, r0, #PAGE_SZ - CACHE_DLINESIZE	@ top addr is inclusive
 	orr	r3, r2, #PSR_I_BIT
@@ -372,7 +373,7 @@ ENTRY(feroceon_cache_fns)
 	.long	feroceon_flush_user_cache_range
 	.long	feroceon_coherent_kern_range
 	.long	feroceon_coherent_user_range
-	.long	feroceon_flush_kern_dcache_page
+	.long	feroceon_flush_kern_dcache_area
 	.long	feroceon_dma_inv_range
 	.long	feroceon_dma_clean_range
 	.long	feroceon_dma_flush_range
@@ -383,7 +384,7 @@ ENTRY(feroceon_range_cache_fns)
 	.long	feroceon_flush_user_cache_range
 	.long	feroceon_coherent_kern_range
 	.long	feroceon_coherent_user_range
-	.long	feroceon_range_flush_kern_dcache_page
+	.long	feroceon_range_flush_kern_dcache_area
 	.long	feroceon_range_dma_inv_range
 	.long	feroceon_range_dma_clean_range
 	.long	feroceon_range_dma_flush_range
diff --git a/arch/arm/mm/proc-mohawk.S b/arch/arm/mm/proc-mohawk.S
index 52b5fd74fbb3..9674d36cc97d 100644
--- a/arch/arm/mm/proc-mohawk.S
+++ b/arch/arm/mm/proc-mohawk.S
@@ -186,15 +186,16 @@ ENTRY(mohawk_coherent_user_range)
 	mov	pc, lr
 
 /*
- *	flush_kern_dcache_page(void *page)
+ *	flush_kern_dcache_area(void *addr, size_t size)
  *
  *	Ensure no D cache aliasing occurs, either with itself or
  *	the I cache
  *
- *	- addr	- page aligned address
+ *	- addr	- kernel address
+ *	- size	- region size
  */
-ENTRY(mohawk_flush_kern_dcache_page)
-	add	r1, r0, #PAGE_SZ
+ENTRY(mohawk_flush_kern_dcache_area)
+	add	r1, r0, r1
 1:	mcr	p15, 0, r0, c7, c14, 1		@ clean+invalidate D entry
 	add	r0, r0, #CACHE_DLINESIZE
 	cmp	r0, r1
@@ -273,7 +274,7 @@ ENTRY(mohawk_cache_fns)
 	.long	mohawk_flush_user_cache_range
 	.long	mohawk_coherent_kern_range
 	.long	mohawk_coherent_user_range
-	.long	mohawk_flush_kern_dcache_page
+	.long	mohawk_flush_kern_dcache_area
 	.long	mohawk_dma_inv_range
 	.long	mohawk_dma_clean_range
 	.long	mohawk_dma_flush_range
diff --git a/arch/arm/mm/proc-syms.c b/arch/arm/mm/proc-syms.c
index ac5c80062b70..3e6210b4d6d4 100644
--- a/arch/arm/mm/proc-syms.c
+++ b/arch/arm/mm/proc-syms.c
@@ -27,8 +27,7 @@ EXPORT_SYMBOL(__cpuc_flush_kern_all);
 EXPORT_SYMBOL(__cpuc_flush_user_all);
 EXPORT_SYMBOL(__cpuc_flush_user_range);
 EXPORT_SYMBOL(__cpuc_coherent_kern_range);
-EXPORT_SYMBOL(__cpuc_flush_dcache_page);
-EXPORT_SYMBOL(dmac_inv_range);  /* because of flush_ioremap_region() */
+EXPORT_SYMBOL(__cpuc_flush_dcache_area);
 #else
 EXPORT_SYMBOL(cpu_cache);
 #endif
diff --git a/arch/arm/mm/proc-xsc3.S b/arch/arm/mm/proc-xsc3.S
index fab134e29826..96456f548798 100644
--- a/arch/arm/mm/proc-xsc3.S
+++ b/arch/arm/mm/proc-xsc3.S
@@ -226,15 +226,16 @@ ENTRY(xsc3_coherent_user_range)
 	mov	pc, lr
 
 /*
- *	flush_kern_dcache_page(void *page)
+ *	flush_kern_dcache_area(void *addr, size_t size)
  *
  *	Ensure no D cache aliasing occurs, either with itself or
  *	the I cache.
  *
- *	- addr	- page aligned address
+ *	- addr	- kernel address
+ *	- size	- region size
  */
-ENTRY(xsc3_flush_kern_dcache_page)
-	add	r1, r0, #PAGE_SZ
+ENTRY(xsc3_flush_kern_dcache_area)
+	add	r1, r0, r1
 1:	mcr	p15, 0, r0, c7, c14, 1		@ clean/invalidate L1 D line
 	add	r0, r0, #CACHELINESIZE
 	cmp	r0, r1
@@ -309,7 +310,7 @@ ENTRY(xsc3_cache_fns)
 	.long	xsc3_flush_user_cache_range
 	.long	xsc3_coherent_kern_range
 	.long	xsc3_coherent_user_range
-	.long	xsc3_flush_kern_dcache_page
+	.long	xsc3_flush_kern_dcache_area
 	.long	xsc3_dma_inv_range
 	.long	xsc3_dma_clean_range
 	.long	xsc3_dma_flush_range
diff --git a/arch/arm/mm/proc-xscale.S b/arch/arm/mm/proc-xscale.S
index f056c283682d..93df47265f2d 100644
--- a/arch/arm/mm/proc-xscale.S
+++ b/arch/arm/mm/proc-xscale.S
@@ -284,15 +284,16 @@ ENTRY(xscale_coherent_user_range)
 	mov	pc, lr
 
 /*
- *	flush_kern_dcache_page(void *page)
+ *	flush_kern_dcache_area(void *addr, size_t size)
  *
  *	Ensure no D cache aliasing occurs, either with itself or
  *	the I cache
  *
- *	- addr	- page aligned address
+ *	- addr	- kernel address
+ *	- size	- region size
  */
-ENTRY(xscale_flush_kern_dcache_page)
-	add	r1, r0, #PAGE_SZ
+ENTRY(xscale_flush_kern_dcache_area)
+	add	r1, r0, r1
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
 	mcr	p15, 0, r0, c7, c6, 1		@ invalidate D entry
 	add	r0, r0, #CACHELINESIZE
@@ -368,7 +369,7 @@ ENTRY(xscale_cache_fns)
 	.long	xscale_flush_user_cache_range
 	.long	xscale_coherent_kern_range
 	.long	xscale_coherent_user_range
-	.long	xscale_flush_kern_dcache_page
+	.long	xscale_flush_kern_dcache_area
 	.long	xscale_dma_inv_range
 	.long	xscale_dma_clean_range
 	.long	xscale_dma_flush_range
@@ -392,7 +393,7 @@ ENTRY(xscale_80200_A0_A1_cache_fns)
 	.long	xscale_flush_user_cache_range
 	.long	xscale_coherent_kern_range
 	.long	xscale_coherent_user_range
-	.long	xscale_flush_kern_dcache_page
+	.long	xscale_flush_kern_dcache_area
 	.long	xscale_dma_flush_range
 	.long	xscale_dma_clean_range
 	.long	xscale_dma_flush_range
diff --git a/drivers/mtd/maps/pxa2xx-flash.c b/drivers/mtd/maps/pxa2xx-flash.c
index 74fa075c838a..b13f6417b5b2 100644
--- a/drivers/mtd/maps/pxa2xx-flash.c
+++ b/drivers/mtd/maps/pxa2xx-flash.c
@@ -20,14 +20,23 @@
 
 #include <asm/io.h>
 #include <mach/hardware.h>
-#include <asm/cacheflush.h>
 
 #include <asm/mach/flash.h>
 
+#define CACHELINESIZE	32
+
 static void pxa2xx_map_inval_cache(struct map_info *map, unsigned long from,
 				      ssize_t len)
 {
-	flush_ioremap_region(map->phys, map->cached, from, len);
+	unsigned long start = (unsigned long)map->cached + from;
+	unsigned long end = start + len;
+
+	start &= ~(CACHELINESIZE - 1);
+	while (start < end) {
+		/* invalidate D cache line */
+		asm volatile ("mcr p15, 0, %0, c7, c6, 1" : : "r" (start));
+		start += CACHELINESIZE;
+	}
 }
 
 struct pxa2xx_flash_info {