kernel_optimize_test/arch/arm/mm/cache-v6.S
Nicolas Pitre 18afea04f1 [ARM] 3294/1: don't invalidate individual BTB entries on ARMv6
Patch from Nicolas Pitre

Doing so adds a much larger cost to the loop than the cost implied by
simply invalidating the whole BTB at once.

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-02-01 19:26:01 +00:00

229 lines
5.4 KiB
ArmAsm

/*
* linux/arch/arm/mm/cache-v6.S
*
* Copyright (C) 2001 Deep Blue Solutions Ltd.
*
* 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.
*
* This is the "shell" of the ARMv6 processor support.
*/
#include <linux/linkage.h>
#include <linux/init.h>
#include <asm/assembler.h>
#include "proc-macros.S"
#define HARVARD_CACHE
#define CACHE_LINE_SIZE 32
#define D_CACHE_LINE_SIZE 32
#define BTB_FLUSH_SIZE 8
/*
* v6_flush_cache_all()
*
* Flush the entire cache.
*
* It is assumed that:
*/
ENTRY(v6_flush_kern_cache_all)
mov r0, #0
#ifdef HARVARD_CACHE
mcr p15, 0, r0, c7, c14, 0 @ D cache clean+invalidate
mcr p15, 0, r0, c7, c5, 0 @ I+BTB cache invalidate
#else
mcr p15, 0, r0, c7, c15, 0 @ Cache clean+invalidate
#endif
mov pc, lr
/*
* v6_flush_cache_all()
*
* Flush all TLB entries in a particular address space
*
* - mm - mm_struct describing address space
*/
ENTRY(v6_flush_user_cache_all)
/*FALLTHROUGH*/
/*
* v6_flush_cache_range(start, end, flags)
*
* Flush a range of TLB entries in the specified address space.
*
* - start - start address (may not be aligned)
* - end - end address (exclusive, may not be aligned)
* - flags - vm_area_struct flags describing address space
*
* It is assumed that:
* - we have a VIPT cache.
*/
ENTRY(v6_flush_user_cache_range)
mov pc, lr
/*
* v6_coherent_kern_range(start,end)
*
* Ensure that the I and D caches are coherent within specified
* region. This is typically used when code has been written to
* a memory region, and will be executed.
*
* - start - virtual start address of region
* - end - virtual end address of region
*
* It is assumed that:
* - the Icache does not read data from the write buffer
*/
ENTRY(v6_coherent_kern_range)
/* FALLTHROUGH */
/*
* v6_coherent_user_range(start,end)
*
* Ensure that the I and D caches are coherent within specified
* region. This is typically used when code has been written to
* a memory region, and will be executed.
*
* - start - virtual start address of region
* - end - virtual end address of region
*
* It is assumed that:
* - the Icache does not read data from the write buffer
*/
ENTRY(v6_coherent_user_range)
#ifdef HARVARD_CACHE
bic r0, r0, #CACHE_LINE_SIZE - 1
1: mcr p15, 0, r0, c7, c10, 1 @ clean D line
mcr p15, 0, r0, c7, c5, 1 @ invalidate I line
add r0, r0, #CACHE_LINE_SIZE
cmp r0, r1
blo 1b
#endif
mcr p15, 0, r0, c7, c5, 6 @ invalidate BTB
#ifdef HARVARD_CACHE
mov r0, #0
mcr p15, 0, r0, c7, c10, 4 @ drain write buffer
#endif
mov pc, lr
/*
* v6_flush_kern_dcache_page(kaddr)
*
* 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)
*/
ENTRY(v6_flush_kern_dcache_page)
add r1, r0, #PAGE_SZ
1:
#ifdef HARVARD_CACHE
mcr p15, 0, r0, c7, c14, 1 @ clean & invalidate D line
#else
mcr p15, 0, r0, c7, c15, 1 @ clean & invalidate unified line
#endif
add r0, r0, #D_CACHE_LINE_SIZE
cmp r0, r1
blo 1b
#ifdef HARVARD_CACHE
mov r0, #0
mcr p15, 0, r0, c7, c10, 4
#endif
mov pc, lr
/*
* v6_dma_inv_range(start,end)
*
* Invalidate the data cache within the specified region; we will
* be performing a DMA operation in this region and we want to
* purge old data in the cache.
*
* - start - virtual start address of region
* - end - virtual end address of region
*/
ENTRY(v6_dma_inv_range)
tst r0, #D_CACHE_LINE_SIZE - 1
bic r0, r0, #D_CACHE_LINE_SIZE - 1
#ifdef HARVARD_CACHE
mcrne p15, 0, r0, c7, c10, 1 @ clean D line
#else
mcrne p15, 0, r0, c7, c11, 1 @ clean unified line
#endif
tst r1, #D_CACHE_LINE_SIZE - 1
bic r1, r1, #D_CACHE_LINE_SIZE - 1
#ifdef HARVARD_CACHE
mcrne p15, 0, r1, c7, c14, 1 @ clean & invalidate D line
#else
mcrne p15, 0, r1, c7, c15, 1 @ clean & invalidate unified line
#endif
1:
#ifdef HARVARD_CACHE
mcr p15, 0, r0, c7, c6, 1 @ invalidate D line
#else
mcr p15, 0, r0, c7, c7, 1 @ invalidate unified line
#endif
add r0, r0, #D_CACHE_LINE_SIZE
cmp r0, r1
blo 1b
mov r0, #0
mcr p15, 0, r0, c7, c10, 4 @ drain write buffer
mov pc, lr
/*
* v6_dma_clean_range(start,end)
* - start - virtual start address of region
* - end - virtual end address of region
*/
ENTRY(v6_dma_clean_range)
bic r0, r0, #D_CACHE_LINE_SIZE - 1
1:
#ifdef HARVARD_CACHE
mcr p15, 0, r0, c7, c10, 1 @ clean D line
#else
mcr p15, 0, r0, c7, c11, 1 @ clean unified line
#endif
add r0, r0, #D_CACHE_LINE_SIZE
cmp r0, r1
blo 1b
mov r0, #0
mcr p15, 0, r0, c7, c10, 4 @ drain write buffer
mov pc, lr
/*
* v6_dma_flush_range(start,end)
* - start - virtual start address of region
* - end - virtual end address of region
*/
ENTRY(v6_dma_flush_range)
bic r0, r0, #D_CACHE_LINE_SIZE - 1
1:
#ifdef HARVARD_CACHE
mcr p15, 0, r0, c7, c14, 1 @ clean & invalidate D line
#else
mcr p15, 0, r0, c7, c15, 1 @ clean & invalidate line
#endif
add r0, r0, #D_CACHE_LINE_SIZE
cmp r0, r1
blo 1b
mov r0, #0
mcr p15, 0, r0, c7, c10, 4 @ drain write buffer
mov pc, lr
__INITDATA
.type v6_cache_fns, #object
ENTRY(v6_cache_fns)
.long v6_flush_kern_cache_all
.long v6_flush_user_cache_all
.long v6_flush_user_cache_range
.long v6_coherent_kern_range
.long v6_coherent_user_range
.long v6_flush_kern_dcache_page
.long v6_dma_inv_range
.long v6_dma_clean_range
.long v6_dma_flush_range
.size v6_cache_fns, . - v6_cache_fns