sh: Migrate from PG_mapped to PG_dcache_dirty.
This inverts the delayed dcache flush a bit to be more in line with other
platforms. At the same time this also gives us the ability to do some
more optimizations and cleanup. Now that the update_mmu_cache() callsite
only tests for the bit, the implementation can gradually be split out and
made generic, rather than relying on special implementations for each of
the peculiar CPU types.
SH7705 in 32kB mode and SH-4 still need slightly different handling, but
this is something that can remain isolated in the varying page copy/clear
routines. On top of that, SH-X3 is dcache coherent, so there is no need
to bother with any of these tests in the PTEAEX version of
update_mmu_cache(), so we kill that off too.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
diff --git a/arch/sh/mm/pg-sh4.c b/arch/sh/mm/pg-sh4.c
index 2fe14da..f3c4b2a 100644
--- a/arch/sh/mm/pg-sh4.c
+++ b/arch/sh/mm/pg-sh4.c
@@ -15,8 +15,6 @@
#include <asm/mmu_context.h>
#include <asm/cacheflush.h>
-#define CACHE_ALIAS (current_cpu_data.dcache.alias_mask)
-
#define kmap_get_fixmap_pte(vaddr) \
pte_offset_kernel(pmd_offset(pud_offset(pgd_offset_k(vaddr), (vaddr)), (vaddr)), (vaddr))
@@ -68,10 +66,9 @@
*/
void clear_user_page(void *to, unsigned long address, struct page *page)
{
- __set_bit(PG_mapped, &page->flags);
-
clear_page(to);
- if ((((address & PAGE_MASK) ^ (unsigned long)to) & CACHE_ALIAS))
+
+ if (pages_do_alias((unsigned long)to, address & PAGE_MASK))
__flush_wback_region(to, PAGE_SIZE);
}
@@ -79,13 +76,14 @@
unsigned long vaddr, void *dst, const void *src,
unsigned long len)
{
- void *vto;
-
- __set_bit(PG_mapped, &page->flags);
-
- vto = kmap_coherent(page, vaddr) + (vaddr & ~PAGE_MASK);
- memcpy(vto, src, len);
- kunmap_coherent(vto);
+ if (page_mapped(page) && !test_bit(PG_dcache_dirty, &page->flags)) {
+ void *vto = kmap_coherent(page, vaddr) + (vaddr & ~PAGE_MASK);
+ memcpy(vto, src, len);
+ kunmap_coherent(vto);
+ } else {
+ memcpy(dst, src, len);
+ set_bit(PG_dcache_dirty, &page->flags);
+ }
if (vma->vm_flags & VM_EXEC)
flush_cache_page(vma, vaddr, page_to_pfn(page));
@@ -95,13 +93,14 @@
unsigned long vaddr, void *dst, const void *src,
unsigned long len)
{
- void *vfrom;
-
- __set_bit(PG_mapped, &page->flags);
-
- vfrom = kmap_coherent(page, vaddr) + (vaddr & ~PAGE_MASK);
- memcpy(dst, vfrom, len);
- kunmap_coherent(vfrom);
+ if (page_mapped(page) && !test_bit(PG_dcache_dirty, &page->flags)) {
+ void *vfrom = kmap_coherent(page, vaddr) + (vaddr & ~PAGE_MASK);
+ memcpy(dst, vfrom, len);
+ kunmap_coherent(vfrom);
+ } else {
+ memcpy(dst, src, len);
+ set_bit(PG_dcache_dirty, &page->flags);
+ }
}
void copy_user_highpage(struct page *to, struct page *from,
@@ -109,14 +108,19 @@
{
void *vfrom, *vto;
- __set_bit(PG_mapped, &to->flags);
-
vto = kmap_atomic(to, KM_USER1);
- vfrom = kmap_coherent(from, vaddr);
- copy_page(vto, vfrom);
- kunmap_coherent(vfrom);
- if (((vaddr ^ (unsigned long)vto) & CACHE_ALIAS))
+ if (page_mapped(from) && !test_bit(PG_dcache_dirty, &from->flags)) {
+ vfrom = kmap_coherent(from, vaddr);
+ copy_page(vto, vfrom);
+ kunmap_coherent(vfrom);
+ } else {
+ vfrom = kmap_atomic(from, KM_USER0);
+ copy_page(vto, vfrom);
+ kunmap_atomic(vfrom, KM_USER0);
+ }
+
+ if (pages_do_alias((unsigned long)vto, vaddr & PAGE_MASK))
__flush_wback_region(vto, PAGE_SIZE);
kunmap_atomic(vto, KM_USER1);
@@ -124,23 +128,3 @@
smp_wmb();
}
EXPORT_SYMBOL(copy_user_highpage);
-
-/*
- * For SH-4, we have our own implementation for ptep_get_and_clear
- */
-pte_t ptep_get_and_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep)
-{
- pte_t pte = *ptep;
-
- pte_clear(mm, addr, ptep);
- if (!pte_not_present(pte)) {
- unsigned long pfn = pte_pfn(pte);
- if (pfn_valid(pfn)) {
- struct page *page = pfn_to_page(pfn);
- struct address_space *mapping = page_mapping(page);
- if (!mapping || !mapping_writably_mapped(mapping))
- __clear_bit(PG_mapped, &page->flags);
- }
- }
- return pte;
-}