aboutsummaryrefslogtreecommitdiff
path: root/arch/arm/mm
diff options
context:
space:
mode:
Diffstat (limited to 'arch/arm/mm')
-rw-r--r--arch/arm/mm/Kconfig22
-rw-r--r--arch/arm/mm/Makefile16
-rw-r--r--arch/arm/mm/cache-fa.S43
-rw-r--r--arch/arm/mm/cache-l2x0.c118
-rw-r--r--arch/arm/mm/cache-tauros2.c263
-rw-r--r--arch/arm/mm/cache-v3.S52
-rw-r--r--arch/arm/mm/cache-v4.S52
-rw-r--r--arch/arm/mm/cache-v4wb.S43
-rw-r--r--arch/arm/mm/cache-v4wt.S51
-rw-r--r--arch/arm/mm/cache-v6.S65
-rw-r--r--arch/arm/mm/cache-v7.S66
-rw-r--r--arch/arm/mm/cache-xsc3l2.c11
-rw-r--r--arch/arm/mm/context.c5
-rw-r--r--arch/arm/mm/copypage-v6.c8
-rw-r--r--arch/arm/mm/dma-mapping.c656
-rw-r--r--arch/arm/mm/fault-armv.c108
-rw-r--r--arch/arm/mm/fault.c5
-rw-r--r--arch/arm/mm/flush.c117
-rw-r--r--arch/arm/mm/highmem.c4
-rw-r--r--arch/arm/mm/init.c24
-rw-r--r--arch/arm/mm/mm.h2
-rw-r--r--arch/arm/mm/mmap.c3
-rw-r--r--arch/arm/mm/mmu.c14
-rw-r--r--arch/arm/mm/nommu.c2
-rw-r--r--arch/arm/mm/proc-arm1020.S43
-rw-r--r--arch/arm/mm/proc-arm1020e.S43
-rw-r--r--arch/arm/mm/proc-arm1022.S43
-rw-r--r--arch/arm/mm/proc-arm1026.S43
-rw-r--r--arch/arm/mm/proc-arm920.S43
-rw-r--r--arch/arm/mm/proc-arm922.S43
-rw-r--r--arch/arm/mm/proc-arm925.S43
-rw-r--r--arch/arm/mm/proc-arm926.S43
-rw-r--r--arch/arm/mm/proc-arm940.S41
-rw-r--r--arch/arm/mm/proc-arm946.S43
-rw-r--r--arch/arm/mm/proc-feroceon.S69
-rw-r--r--arch/arm/mm/proc-mohawk.S43
-rw-r--r--arch/arm/mm/proc-syms.c3
-rw-r--r--arch/arm/mm/proc-v6.S39
-rw-r--r--arch/arm/mm/proc-v7.S14
-rw-r--r--arch/arm/mm/proc-xsc3.S52
-rw-r--r--arch/arm/mm/proc-xscale.S62
-rw-r--r--arch/arm/mm/tlb-v7.S1
-rw-r--r--arch/arm/mm/vmregion.c131
-rw-r--r--arch/arm/mm/vmregion.h29
44 files changed, 1813 insertions, 808 deletions
diff --git a/arch/arm/mm/Kconfig b/arch/arm/mm/Kconfig
index e993140edd8..baf638487a2 100644
--- a/arch/arm/mm/Kconfig
+++ b/arch/arm/mm/Kconfig
@@ -1,9 +1,5 @@
comment "Processor Type"
-config CPU_32
- bool
- default y
-
# Select CPU types depending on the architecture selected. This selects
# which CPUs we support in the kernel image, and the compiler instruction
# optimiser behaviour.
@@ -122,10 +118,7 @@ config CPU_ARM920T
select CPU_TLB_V4WBI if MMU
help
The ARM920T is licensed to be produced by numerous vendors,
- and is used in the Maverick EP9312 and the Samsung S3C2410.
-
- More information on the Maverick EP9312 at
- <http://linuxdevices.com/products/PD2382866068.html>.
+ and is used in the Cirrus EP93xx and the Samsung S3C2410.
Say Y if you want support for the ARM920T processor.
Otherwise, say N.
@@ -391,7 +384,7 @@ config CPU_FEROCEON_OLD_ID
# ARMv6
config CPU_V6
- bool "Support ARM V6 processor" if ARCH_INTEGRATOR || MACH_REALVIEW_EB || MACH_REALVIEW_PBX
+ bool "Support ARM V6 processor" if ARCH_INTEGRATOR || MACH_REALVIEW_EB || MACH_REALVIEW_PBX || ARCH_DOVE
select CPU_32v6
select CPU_ABRT_EV6
select CPU_PABRT_V6
@@ -767,6 +760,15 @@ config CACHE_L2X0
help
This option enables the L2x0 PrimeCell.
+config CACHE_TAUROS2
+ bool "Enable the Tauros2 L2 cache controller"
+ depends on ARCH_DOVE
+ default y
+ select OUTER_CACHE
+ help
+ This option enables the Tauros2 L2 cache controller (as
+ found on PJ1/PJ4).
+
config CACHE_XSC3L2
bool "Enable the L2 cache on XScale3"
depends on CPU_XSC3
@@ -777,5 +779,5 @@ config CACHE_XSC3L2
config ARM_L1_CACHE_SHIFT
int
- default 6 if ARCH_OMAP3
+ default 6 if ARCH_OMAP3 || ARCH_S5PC1XX
default 5
diff --git a/arch/arm/mm/Makefile b/arch/arm/mm/Makefile
index 055cb2aa813..e8d34a80851 100644
--- a/arch/arm/mm/Makefile
+++ b/arch/arm/mm/Makefile
@@ -6,7 +6,7 @@ obj-y := dma-mapping.o extable.o fault.o init.o \
iomap.o
obj-$(CONFIG_MMU) += fault-armv.o flush.o ioremap.o mmap.o \
- pgd.o mmu.o
+ pgd.o mmu.o vmregion.o
ifneq ($(CONFIG_MMU),y)
obj-y += nommu.o
@@ -27,6 +27,9 @@ obj-$(CONFIG_CPU_ABRT_EV5TJ) += abort-ev5tj.o
obj-$(CONFIG_CPU_ABRT_EV6) += abort-ev6.o
obj-$(CONFIG_CPU_ABRT_EV7) += abort-ev7.o
+AFLAGS_abort-ev6.o :=-Wa,-march=armv6k
+AFLAGS_abort-ev7.o :=-Wa,-march=armv7-a
+
obj-$(CONFIG_CPU_PABRT_LEGACY) += pabort-legacy.o
obj-$(CONFIG_CPU_PABRT_V6) += pabort-v6.o
obj-$(CONFIG_CPU_PABRT_V7) += pabort-v7.o
@@ -39,6 +42,9 @@ obj-$(CONFIG_CPU_CACHE_V6) += cache-v6.o
obj-$(CONFIG_CPU_CACHE_V7) += cache-v7.o
obj-$(CONFIG_CPU_CACHE_FA) += cache-fa.o
+AFLAGS_cache-v6.o :=-Wa,-march=armv6
+AFLAGS_cache-v7.o :=-Wa,-march=armv7-a
+
obj-$(CONFIG_CPU_COPY_V3) += copypage-v3.o
obj-$(CONFIG_CPU_COPY_V4WT) += copypage-v4wt.o
obj-$(CONFIG_CPU_COPY_V4WB) += copypage-v4wb.o
@@ -58,6 +64,9 @@ obj-$(CONFIG_CPU_TLB_V6) += tlb-v6.o
obj-$(CONFIG_CPU_TLB_V7) += tlb-v7.o
obj-$(CONFIG_CPU_TLB_FA) += tlb-fa.o
+AFLAGS_tlb-v6.o :=-Wa,-march=armv6
+AFLAGS_tlb-v7.o :=-Wa,-march=armv7-a
+
obj-$(CONFIG_CPU_ARM610) += proc-arm6_7.o
obj-$(CONFIG_CPU_ARM710) += proc-arm6_7.o
obj-$(CONFIG_CPU_ARM7TDMI) += proc-arm7tdmi.o
@@ -84,7 +93,10 @@ obj-$(CONFIG_CPU_FEROCEON) += proc-feroceon.o
obj-$(CONFIG_CPU_V6) += proc-v6.o
obj-$(CONFIG_CPU_V7) += proc-v7.o
+AFLAGS_proc-v6.o :=-Wa,-march=armv6
+AFLAGS_proc-v7.o :=-Wa,-march=armv7-a
+
obj-$(CONFIG_CACHE_FEROCEON_L2) += cache-feroceon-l2.o
obj-$(CONFIG_CACHE_L2X0) += cache-l2x0.o
obj-$(CONFIG_CACHE_XSC3L2) += cache-xsc3l2.o
-
+obj-$(CONFIG_CACHE_TAUROS2) += cache-tauros2.o
diff --git a/arch/arm/mm/cache-fa.S b/arch/arm/mm/cache-fa.S
index b63a8f7b95c..7148e53e607 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
@@ -156,7 +157,7 @@ ENTRY(fa_flush_kern_dcache_page)
* - start - virtual start address
* - end - virtual end address
*/
-ENTRY(fa_dma_inv_range)
+fa_dma_inv_range:
tst r0, #CACHE_DLINESIZE - 1
bic r0, r0, #CACHE_DLINESIZE - 1
mcrne p15, 0, r0, c7, c14, 1 @ clean & invalidate D entry
@@ -179,7 +180,7 @@ ENTRY(fa_dma_inv_range)
* - start - virtual start address
* - end - virtual end address
*/
-ENTRY(fa_dma_clean_range)
+fa_dma_clean_range:
bic r0, r0, #CACHE_DLINESIZE - 1
1: mcr p15, 0, r0, c7, c10, 1 @ clean D entry
add r0, r0, #CACHE_DLINESIZE
@@ -204,6 +205,30 @@ ENTRY(fa_dma_flush_range)
mcr p15, 0, r0, c7, c10, 4 @ drain write buffer
mov pc, lr
+/*
+ * dma_map_area(start, size, dir)
+ * - start - kernel virtual start address
+ * - size - size of region
+ * - dir - DMA direction
+ */
+ENTRY(fa_dma_map_area)
+ add r1, r1, r0
+ cmp r2, #DMA_TO_DEVICE
+ beq fa_dma_clean_range
+ bcs fa_dma_inv_range
+ b fa_dma_flush_range
+ENDPROC(fa_dma_map_area)
+
+/*
+ * dma_unmap_area(start, size, dir)
+ * - start - kernel virtual start address
+ * - size - size of region
+ * - dir - DMA direction
+ */
+ENTRY(fa_dma_unmap_area)
+ mov pc, lr
+ENDPROC(fa_dma_unmap_area)
+
__INITDATA
.type fa_cache_fns, #object
@@ -213,8 +238,8 @@ 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_dma_inv_range
- .long fa_dma_clean_range
+ .long fa_flush_kern_dcache_area
+ .long fa_dma_map_area
+ .long fa_dma_unmap_area
.long fa_dma_flush_range
.size fa_cache_fns, . - fa_cache_fns
diff --git a/arch/arm/mm/cache-l2x0.c b/arch/arm/mm/cache-l2x0.c
index b480f1d3591..cb8fc6573b1 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)
@@ -99,18 +150,25 @@ void __init l2x0_init(void __iomem *base, __u32 aux_val, __u32 aux_mask)
l2x0_base = base;
- /* disable L2X0 */
- writel(0, l2x0_base + L2X0_CTRL);
+ /*
+ * Check if l2x0 controller is already enabled.
+ * If you are booting from non-secure mode
+ * accessing the below registers will fault.
+ */
+ if (!(readl(l2x0_base + L2X0_CTRL) & 1)) {
+
+ /* l2x0 controller is disabled */
- aux = readl(l2x0_base + L2X0_AUX_CTRL);
- aux &= aux_mask;
- aux |= aux_val;
- writel(aux, l2x0_base + L2X0_AUX_CTRL);
+ aux = readl(l2x0_base + L2X0_AUX_CTRL);
+ aux &= aux_mask;
+ aux |= aux_val;
+ writel(aux, l2x0_base + L2X0_AUX_CTRL);
- l2x0_inv_all();
+ l2x0_inv_all();
- /* enable L2X0 */
- writel(1, l2x0_base + L2X0_CTRL);
+ /* enable L2X0 */
+ writel(1, l2x0_base + L2X0_CTRL);
+ }
outer_cache.inv_range = l2x0_inv_range;
outer_cache.clean_range = l2x0_clean_range;
diff --git a/arch/arm/mm/cache-tauros2.c b/arch/arm/mm/cache-tauros2.c
new file mode 100644
index 00000000000..50868651890
--- /dev/null
+++ b/arch/arm/mm/cache-tauros2.c
@@ -0,0 +1,263 @@
+/*
+ * arch/arm/mm/cache-tauros2.c - Tauros2 L2 cache controller support
+ *
+ * Copyright (C) 2008 Marvell Semiconductor
+ *
+ * This file is licensed under the terms of the GNU General Public
+ * License version 2. This program is licensed "as is" without any
+ * warranty of any kind, whether express or implied.
+ *
+ * References:
+ * - PJ1 CPU Core Datasheet,
+ * Document ID MV-S104837-01, Rev 0.7, January 24 2008.
+ * - PJ4 CPU Core Datasheet,
+ * Document ID MV-S105190-00, Rev 0.7, March 14 2008.
+ */
+
+#include <linux/init.h>
+#include <asm/cacheflush.h>
+#include <asm/hardware/cache-tauros2.h>
+
+
+/*
+ * When Tauros2 is used on a CPU that supports the v7 hierarchical
+ * cache operations, the cache handling code in proc-v7.S takes care
+ * of everything, including handling DMA coherency.
+ *
+ * So, we only need to register outer cache operations here if we're
+ * being used on a pre-v7 CPU, and we only need to build support for
+ * outer cache operations into the kernel image if the kernel has been
+ * configured to support a pre-v7 CPU.
+ */
+#if __LINUX_ARM_ARCH__ < 7
+/*
+ * Low-level cache maintenance operations.
+ */
+static inline void tauros2_clean_pa(unsigned long addr)
+{
+ __asm__("mcr p15, 1, %0, c7, c11, 3" : : "r" (addr));
+}
+
+static inline void tauros2_clean_inv_pa(unsigned long addr)
+{
+ __asm__("mcr p15, 1, %0, c7, c15, 3" : : "r" (addr));
+}
+
+static inline void tauros2_inv_pa(unsigned long addr)
+{
+ __asm__("mcr p15, 1, %0, c7, c7, 3" : : "r" (addr));
+}
+
+
+/*
+ * Linux primitives.
+ *
+ * Note that the end addresses passed to Linux primitives are
+ * noninclusive.
+ */
+#define CACHE_LINE_SIZE 32
+
+static void tauros2_inv_range(unsigned long start, unsigned long end)
+{
+ /*
+ * Clean and invalidate partial first cache line.
+ */
+ if (start & (CACHE_LINE_SIZE - 1)) {
+ tauros2_clean_inv_pa(start & ~(CACHE_LINE_SIZE - 1));
+ start = (start | (CACHE_LINE_SIZE - 1)) + 1;
+ }
+
+ /*
+ * Clean and invalidate partial last cache line.
+ */
+ if (end & (CACHE_LINE_SIZE - 1)) {
+ tauros2_clean_inv_pa(end & ~(CACHE_LINE_SIZE - 1));
+ end &= ~(CACHE_LINE_SIZE - 1);
+ }
+
+ /*
+ * Invalidate all full cache lines between 'start' and 'end'.
+ */
+ while (start < end) {
+ tauros2_inv_pa(start);
+ start += CACHE_LINE_SIZE;
+ }
+
+ dsb();
+}
+
+static void tauros2_clean_range(unsigned long start, unsigned long end)
+{
+ start &= ~(CACHE_LINE_SIZE - 1);
+ while (start < end) {
+ tauros2_clean_pa(start);
+ start += CACHE_LINE_SIZE;
+ }
+
+ dsb();
+}
+
+static void tauros2_flush_range(unsigned long start, unsigned long end)
+{
+ start &= ~(CACHE_LINE_SIZE - 1);
+ while (start < end) {
+ tauros2_clean_inv_pa(start);
+ start += CACHE_LINE_SIZE;
+ }
+
+ dsb();
+}
+#endif
+
+static inline u32 __init read_extra_features(void)
+{
+ u32 u;
+
+ __asm__("mrc p15, 1, %0, c15, c1, 0" : "=r" (u));
+
+ return u;
+}
+
+static inline void __init write_extra_features(u32 u)
+{
+ __asm__("mcr p15, 1, %0, c15, c1, 0" : : "r" (u));
+}
+
+static void __init disable_l2_prefetch(void)
+{
+ u32 u;
+
+ /*
+ * Read the CPU Extra Features register and verify that the
+ * Disable L2 Prefetch bit is set.
+ */
+ u = read_extra_features();
+ if (!(u & 0x01000000)) {
+ printk(KERN_INFO "Tauros2: Disabling L2 prefetch.\n");
+ write_extra_features(u | 0x01000000);
+ }
+}
+
+static inline int __init cpuid_scheme(void)
+{
+ extern int processor_id;
+
+ return !!((processor_id & 0x000f0000) == 0x000f0000);
+}
+
+static inline u32 __init read_mmfr3(void)
+{
+ u32 mmfr3;
+
+ __asm__("mrc p15, 0, %0, c0, c1, 7\n" : "=r" (mmfr3));
+
+ return mmfr3;
+}
+
+static inline u32 __init read_actlr(void)
+{
+ u32 actlr;
+
+ __asm__("mrc p15, 0, %0, c1, c0, 1\n" : "=r" (actlr));
+
+ return actlr;
+}
+
+static inline void __init write_actlr(u32 actlr)
+{
+ __asm__("mcr p15, 0, %0, c1, c0, 1\n" : : "r" (actlr));
+}
+
+void __init tauros2_init(void)
+{
+ extern int processor_id;
+ char *mode;
+
+ disable_l2_prefetch();
+
+#ifdef CONFIG_CPU_32v5
+ if ((processor_id & 0xff0f0000) == 0x56050000) {
+ u32 feat;
+
+ /*
+ * v5 CPUs with Tauros2 have the L2 cache enable bit
+ * located in the CPU Extra Features register.
+ */
+ feat = read_extra_features();
+ if (!(feat & 0x00400000)) {
+ printk(KERN_INFO "Tauros2: Enabling L2 cache.\n");
+ write_extra_features(feat | 0x00400000);
+ }
+
+ mode = "ARMv5";
+ outer_cache.inv_range = tauros2_inv_range;
+ outer_cache.clean_range = tauros2_clean_range;
+ outer_cache.flush_range = tauros2_flush_range;
+ }
+#endif
+
+#ifdef CONFIG_CPU_32v6
+ /*
+ * Check whether this CPU lacks support for the v7 hierarchical
+ * cache ops. (PJ4 is in its v6 personality mode if the MMFR3
+ * register indicates no support for the v7 hierarchical cache
+ * ops.)
+ */
+ if (cpuid_scheme() && (read_mmfr3() & 0xf) == 0) {
+ /*
+ * When Tauros2 is used in an ARMv6 system, the L2
+ * enable bit is in the ARMv6 ARM-mandated position
+ * (bit [26] of the System Control Register).
+ */
+ if (!(get_cr() & 0x04000000)) {
+ printk(KERN_INFO "Tauros2: Enabling L2 cache.\n");
+ adjust_cr(0x04000000, 0x04000000);
+ }
+
+ mode = "ARMv6";
+ outer_cache.inv_range = tauros2_inv_range;
+ outer_cache.clean_range = tauros2_clean_range;
+ outer_cache.flush_range = tauros2_flush_range;
+ }
+#endif
+
+#ifdef CONFIG_CPU_32v7
+ /*
+ * Check whether this CPU has support for the v7 hierarchical
+ * cache ops. (PJ4 is in its v7 personality mode if the MMFR3
+ * register indicates support for the v7 hierarchical cache
+ * ops.)
+ *
+ * (Although strictly speaking there may exist CPUs that
+ * implement the v7 cache ops but are only ARMv6 CPUs (due to
+ * not complying with all of the other ARMv7 requirements),
+ * there are no real-life examples of Tauros2 being used on
+ * such CPUs as of yet.)
+ */
+ if (cpuid_scheme() && (read_mmfr3() & 0xf) == 1) {
+ u32 actlr;
+
+ /*
+ * When Tauros2 is used in an ARMv7 system, the L2
+ * enable bit is located in the Auxiliary System Control
+ * Register (which is the only register allowed by the
+ * ARMv7 spec to contain fine-grained cache control bits).
+ */
+ actlr = read_actlr();
+ if (!(actlr & 0x00000002)) {
+ printk(KERN_INFO "Tauros2: Enabling L2 cache.\n");
+ write_actlr(actlr | 0x00000002);
+ }
+
+ mode = "ARMv7";
+ }
+#endif
+
+ if (mode == NULL) {
+ printk(KERN_CRIT "Tauros2: Unable to detect CPU mode.\n");
+ return;
+ }
+
+ printk(KERN_INFO "Tauros2: L2 cache support initialised "
+ "in %s mode.\n", mode);
+}
diff --git a/arch/arm/mm/cache-v3.S b/arch/arm/mm/cache-v3.S
index 8a4abebc478..c2ff3c599fe 100644
--- a/arch/arm/mm/cache-v3.S
+++ b/arch/arm/mm/cache-v3.S
@@ -72,28 +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)
- /* FALLTHROUGH */
-
-/*
- * dma_inv_range(start, end)
- *
- * Invalidate (discard) the specified virtual address range.
- * May not write back any entries. If 'start' or 'end'
- * are not cache line aligned, those lines must be written
- * back.
- *
- * - start - virtual start address
- * - end - virtual end address
- */
-ENTRY(v3_dma_inv_range)
+ENTRY(v3_flush_kern_dcache_area)
/* FALLTHROUGH */
/*
@@ -107,18 +94,29 @@ ENTRY(v3_dma_inv_range)
ENTRY(v3_dma_flush_range)
mov r0, #0
mcr p15, 0, r0, c7, c0, 0 @ flush ID cache
+ mov pc, lr
+
+/*
+ * dma_unmap_area(start, size, dir)
+ * - start - kernel virtual start address
+ * - size - size of region
+ * - dir - DMA direction
+ */
+ENTRY(v3_dma_unmap_area)
+ teq r2, #DMA_TO_DEVICE
+ bne v3_dma_flush_range
/* FALLTHROUGH */
/*
- * dma_clean_range(start, end)
- *
- * Clean (write back) the specified virtual address range.
- *
- * - start - virtual start address
- * - end - virtual end address
+ * dma_map_area(start, size, dir)
+ * - start - kernel virtual start address
+ * - size - size of region
+ * - dir - DMA direction
*/
-ENTRY(v3_dma_clean_range)
+ENTRY(v3_dma_map_area)
mov pc, lr
+ENDPROC(v3_dma_unmap_area)
+ENDPROC(v3_dma_map_area)
__INITDATA
@@ -129,8 +127,8 @@ 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_dma_inv_range
- .long v3_dma_clean_range
+ .long v3_flush_kern_dcache_area
+ .long v3_dma_map_area
+ .long v3_dma_unmap_area
.long v3_dma_flush_range
.size v3_cache_fns, . - v3_cache_fns
diff --git a/arch/arm/mm/cache-v4.S b/arch/arm/mm/cache-v4.S
index 3668611cb40..4810f7e3e81 100644
--- a/arch/arm/mm/cache-v4.S
+++ b/arch/arm/mm/cache-v4.S
@@ -82,28 +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)
- /* FALLTHROUGH */
-
-/*
- * dma_inv_range(start, end)
- *
- * Invalidate (discard) the specified virtual address range.
- * May not write back any entries. If 'start' or 'end'
- * are not cache line aligned, those lines must be written
- * back.
- *
- * - start - virtual start address
- * - end - virtual end address
- */
-ENTRY(v4_dma_inv_range)
+ENTRY(v4_flush_kern_dcache_area)
/* FALLTHROUGH */
/*
@@ -119,18 +106,29 @@ ENTRY(v4_dma_flush_range)
mov r0, #0
mcr p15, 0, r0, c7, c7, 0 @ flush ID cache
#endif
+ mov pc, lr
+
+/*
+ * dma_unmap_area(start, size, dir)
+ * - start - kernel virtual start address
+ * - size - size of region
+ * - dir - DMA direction
+ */