kernel-ark/arch/arm/mm/cache-v4wb.S
Russell King 6ebbf2ce43 ARM: convert all "mov.* pc, reg" to "bx reg" for ARMv6+
ARMv6 and greater introduced a new instruction ("bx") which can be used
to return from function calls.  Recent CPUs perform better when the
"bx lr" instruction is used rather than the "mov pc, lr" instruction,
and this sequence is strongly recommended to be used by the ARM
architecture manual (section A.4.1.1).

We provide a new macro "ret" with all its variants for the condition
code which will resolve to the appropriate instruction.

Rather than doing this piecemeal, and miss some instances, change all
the "mov pc" instances to use the new macro, with the exception of
the "movs" instruction and the kprobes code.  This allows us to detect
the "mov pc, lr" case and fix it up - and also gives us the possibility
of deploying this for other registers depending on the CPU selection.

Reported-by: Will Deacon <will.deacon@arm.com>
Tested-by: Stephen Warren <swarren@nvidia.com> # Tegra Jetson TK1
Tested-by: Robert Jarzmik <robert.jarzmik@free.fr> # mioa701_bootresume.S
Tested-by: Andrew Lunn <andrew@lunn.ch> # Kirkwood
Tested-by: Shawn Guo <shawn.guo@freescale.com>
Tested-by: Tony Lindgren <tony@atomide.com> # OMAPs
Tested-by: Gregory CLEMENT <gregory.clement@free-electrons.com> # Armada XP, 375, 385
Acked-by: Sekhar Nori <nsekhar@ti.com> # DaVinci
Acked-by: Christoffer Dall <christoffer.dall@linaro.org> # kvm/hyp
Acked-by: Haojian Zhuang <haojian.zhuang@gmail.com> # PXA3xx
Acked-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com> # Xen
Tested-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> # ARMv7M
Tested-by: Simon Horman <horms+renesas@verge.net.au> # Shmobile
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2014-07-18 12:29:04 +01:00

262 lines
6.2 KiB
ArmAsm

/*
* linux/arch/arm/mm/cache-v4wb.S
*
* Copyright (C) 1997-2002 Russell king
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2 as
* published by the Free Software Foundation.
*/
#include <linux/linkage.h>
#include <linux/init.h>
#include <asm/assembler.h>
#include <asm/memory.h>
#include <asm/page.h>
#include "proc-macros.S"
/*
* The size of one data cache line.
*/
#define CACHE_DLINESIZE 32
/*
* The total size of the data cache.
*/
#if defined(CONFIG_CPU_SA110)
# define CACHE_DSIZE 16384
#elif defined(CONFIG_CPU_SA1100)
# define CACHE_DSIZE 8192
#else
# error Unknown cache size
#endif
/*
* This is the size at which it becomes more efficient to
* clean the whole cache, rather than using the individual
* cache line maintenance instructions.
*
* Size Clean (ticks) Dirty (ticks)
* 4096 21 20 21 53 55 54
* 8192 40 41 40 106 100 102
* 16384 77 77 76 140 140 138
* 32768 150 149 150 214 216 212 <---
* 65536 296 297 296 351 358 361
* 131072 591 591 591 656 657 651
* Whole 132 136 132 221 217 207 <---
*/
#define CACHE_DLIMIT (CACHE_DSIZE * 4)
.data
flush_base:
.long FLUSH_BASE
.text
/*
* flush_icache_all()
*
* Unconditionally clean and invalidate the entire icache.
*/
ENTRY(v4wb_flush_icache_all)
mov r0, #0
mcr p15, 0, r0, c7, c5, 0 @ invalidate I cache
ret lr
ENDPROC(v4wb_flush_icache_all)
/*
* flush_user_cache_all()
*
* Clean and invalidate all cache entries in a particular address
* space.
*/
ENTRY(v4wb_flush_user_cache_all)
/* FALLTHROUGH */
/*
* flush_kern_cache_all()
*
* Clean and invalidate the entire cache.
*/
ENTRY(v4wb_flush_kern_cache_all)
mov ip, #0
mcr p15, 0, ip, c7, c5, 0 @ invalidate I cache
__flush_whole_cache:
ldr r3, =flush_base
ldr r1, [r3, #0]
eor r1, r1, #CACHE_DSIZE
str r1, [r3, #0]
add r2, r1, #CACHE_DSIZE
1: ldr r3, [r1], #32
cmp r1, r2
blo 1b
#ifdef FLUSH_BASE_MINICACHE
add r2, r2, #FLUSH_BASE_MINICACHE - FLUSH_BASE
sub r1, r2, #512 @ only 512 bytes
1: ldr r3, [r1], #32
cmp r1, r2
blo 1b
#endif
mcr p15, 0, ip, c7, c10, 4 @ drain write buffer
ret lr
/*
* flush_user_cache_range(start, end, flags)
*
* Invalidate a range of cache entries in the specified
* address space.
*
* - start - start address (inclusive, page aligned)
* - end - end address (exclusive, page aligned)
* - flags - vma_area_struct flags describing address space
*/
ENTRY(v4wb_flush_user_cache_range)
mov ip, #0
sub r3, r1, r0 @ calculate total size
tst r2, #VM_EXEC @ executable region?
mcrne p15, 0, ip, c7, c5, 0 @ invalidate I cache
cmp r3, #CACHE_DLIMIT @ total size >= limit?
bhs __flush_whole_cache @ flush whole D cache
1: mcr p15, 0, r0, c7, c10, 1 @ clean D entry
mcr p15, 0, r0, c7, c6, 1 @ invalidate D entry
add r0, r0, #CACHE_DLINESIZE
cmp r0, r1
blo 1b
tst r2, #VM_EXEC
mcrne p15, 0, ip, c7, c10, 4 @ drain write buffer
ret lr
/*
* flush_kern_dcache_area(void *addr, size_t size)
*
* Ensure no D cache aliasing occurs, either with itself or
* the I cache
*
* - addr - kernel address
* - size - region size
*/
ENTRY(v4wb_flush_kern_dcache_area)
add r1, r0, r1
/* fall through */
/*
* coherent_kern_range(start, end)
*
* Ensure coherency between the Icache and the Dcache in the
* region described by start. If you have non-snooping
* Harvard caches, you need to implement this function.
*
* - start - virtual start address
* - end - virtual end address
*/
ENTRY(v4wb_coherent_kern_range)
/* fall through */
/*
* coherent_user_range(start, end)
*
* Ensure coherency between the Icache and the Dcache in the
* region described by start. If you have non-snooping
* Harvard caches, you need to implement this function.
*
* - start - virtual start address
* - end - virtual end address
*/
ENTRY(v4wb_coherent_user_range)
bic r0, r0, #CACHE_DLINESIZE - 1
1: mcr p15, 0, r0, c7, c10, 1 @ clean D entry
mcr p15, 0, r0, c7, c6, 1 @ invalidate D entry
add r0, r0, #CACHE_DLINESIZE
cmp r0, r1
blo 1b
mov r0, #0
mcr p15, 0, r0, c7, c5, 0 @ invalidate I cache
mcr p15, 0, r0, c7, c10, 4 @ drain WB
ret lr
/*
* 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
*/
v4wb_dma_inv_range:
tst r0, #CACHE_DLINESIZE - 1
bic r0, r0, #CACHE_DLINESIZE - 1
mcrne p15, 0, r0, c7, c10, 1 @ clean D entry
tst r1, #CACHE_DLINESIZE - 1
mcrne p15, 0, r1, c7, c10, 1 @ clean D entry
1: mcr p15, 0, r0, c7, c6, 1 @ invalidate D entry
add r0, r0, #CACHE_DLINESIZE
cmp r0, r1
blo 1b
mcr p15, 0, r0, c7, c10, 4 @ drain write buffer
ret lr
/*
* dma_clean_range(start, end)
*
* Clean (write back) the specified virtual address range.
*
* - start - virtual start address
* - end - virtual end address
*/
v4wb_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
cmp r0, r1
blo 1b
mcr p15, 0, r0, c7, c10, 4 @ drain write buffer
ret lr
/*
* dma_flush_range(start, end)
*
* Clean and invalidate the specified virtual address range.
*
* - start - virtual start address
* - end - virtual end address
*
* This is actually the same as v4wb_coherent_kern_range()
*/
.globl v4wb_dma_flush_range
.set v4wb_dma_flush_range, v4wb_coherent_kern_range
/*
* dma_map_area(start, size, dir)
* - start - kernel virtual start address
* - size - size of region
* - dir - DMA direction
*/
ENTRY(v4wb_dma_map_area)
add r1, r1, r0
cmp r2, #DMA_TO_DEVICE
beq v4wb_dma_clean_range
bcs v4wb_dma_inv_range
b v4wb_dma_flush_range
ENDPROC(v4wb_dma_map_area)
/*
* dma_unmap_area(start, size, dir)
* - start - kernel virtual start address
* - size - size of region
* - dir - DMA direction
*/
ENTRY(v4wb_dma_unmap_area)
ret lr
ENDPROC(v4wb_dma_unmap_area)
.globl v4wb_flush_kern_cache_louis
.equ v4wb_flush_kern_cache_louis, v4wb_flush_kern_cache_all
__INITDATA
@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
define_cache_functions v4wb