android_kernel_xiaomi_sm7250/arch/sh/mm/cache-sh7705.c
Paul Mundt 26b7a78c55 sh: Lazy dcache writeback optimizations.
This converts the lazy dcache handling to the model described in
Documentation/cachetlb.txt and drops the ptep_get_and_clear() hacks
used for the aliasing dcaches on SH-4 and SH7705 in 32kB mode. As a
bonus, this slightly cuts down on the cache flushing frequency.

With that and the PTEA handling out of the way, the update_mmu_cache()
implementations can be consolidated, and we no longer have to worry
about which configuration the cache is in for the SH7705 case.

And finally, explicitly disable the lazy writeback on SMP (SH-4A).

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-02-13 10:54:44 +09:00

203 lines
4.9 KiB
C

/*
* arch/sh/mm/cache-sh7705.c
*
* Copyright (C) 1999, 2000 Niibe Yutaka
* Copyright (C) 2004 Alex Song
* Copyright (C) 2006 Paul Mundt
*
* This file is subject to the terms and conditions of the GNU General Public
* License. See the file "COPYING" in the main directory of this archive
* for more details.
*/
#include <linux/init.h>
#include <linux/mman.h>
#include <linux/mm.h>
#include <linux/threads.h>
#include <asm/addrspace.h>
#include <asm/page.h>
#include <asm/pgtable.h>
#include <asm/processor.h>
#include <asm/cache.h>
#include <asm/io.h>
#include <asm/uaccess.h>
#include <asm/pgalloc.h>
#include <asm/mmu_context.h>
#include <asm/cacheflush.h>
/*
* The 32KB cache on the SH7705 suffers from the same synonym problem
* as SH4 CPUs
*/
static inline void cache_wback_all(void)
{
unsigned long ways, waysize, addrstart;
ways = cpu_data->dcache.ways;
waysize = cpu_data->dcache.sets;
waysize <<= cpu_data->dcache.entry_shift;
addrstart = CACHE_OC_ADDRESS_ARRAY;
do {
unsigned long addr;
for (addr = addrstart;
addr < addrstart + waysize;
addr += cpu_data->dcache.linesz) {
unsigned long data;
int v = SH_CACHE_UPDATED | SH_CACHE_VALID;
data = ctrl_inl(addr);
if ((data & v) == v)
ctrl_outl(data & ~v, addr);
}
addrstart += cpu_data->dcache.way_incr;
} while (--ways);
}
/*
* Write back the range of D-cache, and purge the I-cache.
*
* Called from kernel/module.c:sys_init_module and routine for a.out format.
*/
void flush_icache_range(unsigned long start, unsigned long end)
{
__flush_wback_region((void *)start, end - start);
}
/*
* Writeback&Invalidate the D-cache of the page
*/
static void __flush_dcache_page(unsigned long phys)
{
unsigned long ways, waysize, addrstart;
unsigned long flags;
phys |= SH_CACHE_VALID;
/*
* Here, phys is the physical address of the page. We check all the
* tags in the cache for those with the same page number as this page
* (by masking off the lowest 2 bits of the 19-bit tag; these bits are
* derived from the offset within in the 4k page). Matching valid
* entries are invalidated.
*
* Since 2 bits of the cache index are derived from the virtual page
* number, knowing this would reduce the number of cache entries to be
* searched by a factor of 4. However this function exists to deal with
* potential cache aliasing, therefore the optimisation is probably not
* possible.
*/
local_irq_save(flags);
jump_to_P2();
ways = cpu_data->dcache.ways;
waysize = cpu_data->dcache.sets;
waysize <<= cpu_data->dcache.entry_shift;
addrstart = CACHE_OC_ADDRESS_ARRAY;
do {
unsigned long addr;
for (addr = addrstart;
addr < addrstart + waysize;
addr += cpu_data->dcache.linesz) {
unsigned long data;
data = ctrl_inl(addr) & (0x1ffffC00 | SH_CACHE_VALID);
if (data == phys) {
data &= ~(SH_CACHE_VALID | SH_CACHE_UPDATED);
ctrl_outl(data, addr);
}
}
addrstart += cpu_data->dcache.way_incr;
} while (--ways);
back_to_P1();
local_irq_restore(flags);
}
/*
* Write back & invalidate the D-cache of the page.
* (To avoid "alias" issues)
*/
void flush_dcache_page(struct page *page)
{
struct address_space *mapping = page_mapping(page);
if (mapping && !mapping_mapped(mapping))
set_bit(PG_dcache_dirty, &page->flags);
else
__flush_dcache_page(PHYSADDR(page_address(page)));
}
void flush_cache_all(void)
{
unsigned long flags;
local_irq_save(flags);
jump_to_P2();
cache_wback_all();
back_to_P1();
local_irq_restore(flags);
}
void flush_cache_mm(struct mm_struct *mm)
{
/* Is there any good way? */
/* XXX: possibly call flush_cache_range for each vm area */
flush_cache_all();
}
/*
* Write back and invalidate D-caches.
*
* START, END: Virtual Address (U0 address)
*
* NOTE: We need to flush the _physical_ page entry.
* Flushing the cache lines for U0 only isn't enough.
* We need to flush for P1 too, which may contain aliases.
*/
void flush_cache_range(struct vm_area_struct *vma, unsigned long start,
unsigned long end)
{
/*
* We could call flush_cache_page for the pages of these range,
* but it's not efficient (scan the caches all the time...).
*
* We can't use A-bit magic, as there's the case we don't have
* valid entry on TLB.
*/
flush_cache_all();
}
/*
* Write back and invalidate I/D-caches for the page.
*
* ADDRESS: Virtual Address (U0 address)
*/
void flush_cache_page(struct vm_area_struct *vma, unsigned long address,
unsigned long pfn)
{
__flush_dcache_page(pfn << PAGE_SHIFT);
}
/*
* This is called when a page-cache page is about to be mapped into a
* user process' address space. It offers an opportunity for a
* port to ensure d-cache/i-cache coherency if necessary.
*
* Not entirely sure why this is necessary on SH3 with 32K cache but
* without it we get occasional "Memory fault" when loading a program.
*/
void flush_icache_page(struct vm_area_struct *vma, struct page *page)
{
__flush_purge_region(page_address(page), PAGE_SIZE);
}