Commit 1bb3630e authored by Hugh Dickins's avatar Hugh Dickins Committed by Linus Torvalds

[PATCH] mm: ptd_alloc inline and out

It seems odd to me that, whereas pud_alloc and pmd_alloc test inline, only
calling out-of-line __pud_alloc __pmd_alloc if allocation needed,
pte_alloc_map and pte_alloc_kernel are entirely out-of-line.  Though it does
add a little to kernel size, change them to macros testing inline, calling
__pte_alloc or __pte_alloc_kernel to allocate out-of-line.  Mark none of them
as fastcalls, leave that to CONFIG_REGPARM or not.

It also seems more natural for the out-of-line functions to leave the offset
calculation and map to the inline, which has to do it anyway for the common
case.  At least mremap move wants __pte_alloc without _map.

Macros rather than inline functions, certainly to avoid the header file issues
which arise from CONFIG_HIGHPTE needing kmap_types.h, but also in case any
architectures I haven't built would have other such problems.
Signed-off-by: default avatarHugh Dickins <hugh@veritas.com>
Signed-off-by: default avatarAndrew Morton <akpm@osdl.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@osdl.org>
parent 872fec16
...@@ -11,13 +11,8 @@ ...@@ -11,13 +11,8 @@
#define pud_t pgd_t #define pud_t pgd_t
#define pmd_alloc(mm, pud, address) \ #define pmd_alloc(mm, pud, address) \
({ pmd_t *ret; \ ((unlikely(pgd_none(*(pud))) && __pmd_alloc(mm, pud, address))? \
if (pgd_none(*pud)) \ NULL: pmd_offset(pud, address))
ret = __pmd_alloc(mm, pud, address); \
else \
ret = pmd_offset(pud, address); \
ret; \
})
#define pud_alloc(mm, pgd, address) (pgd) #define pud_alloc(mm, pgd, address) (pgd)
#define pud_offset(pgd, start) (pgd) #define pud_offset(pgd, start) (pgd)
......
...@@ -704,10 +704,6 @@ static inline void unmap_shared_mapping_range(struct address_space *mapping, ...@@ -704,10 +704,6 @@ static inline void unmap_shared_mapping_range(struct address_space *mapping,
} }
extern int vmtruncate(struct inode * inode, loff_t offset); extern int vmtruncate(struct inode * inode, loff_t offset);
extern pud_t *FASTCALL(__pud_alloc(struct mm_struct *mm, pgd_t *pgd, unsigned long address));
extern pmd_t *FASTCALL(__pmd_alloc(struct mm_struct *mm, pud_t *pud, unsigned long address));
extern pte_t *FASTCALL(pte_alloc_kernel(pmd_t *pmd, unsigned long address));
extern pte_t *FASTCALL(pte_alloc_map(struct mm_struct *mm, pmd_t *pmd, unsigned long address));
extern int install_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long addr, struct page *page, pgprot_t prot); extern int install_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long addr, struct page *page, pgprot_t prot);
extern int install_file_pte(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long addr, unsigned long pgoff, pgprot_t prot); extern int install_file_pte(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long addr, unsigned long pgoff, pgprot_t prot);
extern int __handle_mm_fault(struct mm_struct *mm,struct vm_area_struct *vma, unsigned long address, int write_access); extern int __handle_mm_fault(struct mm_struct *mm,struct vm_area_struct *vma, unsigned long address, int write_access);
...@@ -760,32 +756,36 @@ struct shrinker; ...@@ -760,32 +756,36 @@ struct shrinker;
extern struct shrinker *set_shrinker(int, shrinker_t); extern struct shrinker *set_shrinker(int, shrinker_t);
extern void remove_shrinker(struct shrinker *shrinker); extern void remove_shrinker(struct shrinker *shrinker);
/* int __pud_alloc(struct mm_struct *mm, pgd_t *pgd, unsigned long address);
* On a two-level or three-level page table, this ends up being trivial. Thus int __pmd_alloc(struct mm_struct *mm, pud_t *pud, unsigned long address);
* the inlining and the symmetry break with pte_alloc_map() that does all int __pte_alloc(struct mm_struct *mm, pmd_t *pmd, unsigned long address);
* of this out-of-line. int __pte_alloc_kernel(pmd_t *pmd, unsigned long address);
*/
/* /*
* The following ifdef needed to get the 4level-fixup.h header to work. * The following ifdef needed to get the 4level-fixup.h header to work.
* Remove it when 4level-fixup.h has been removed. * Remove it when 4level-fixup.h has been removed.
*/ */
#ifdef CONFIG_MMU #if defined(CONFIG_MMU) && !defined(__ARCH_HAS_4LEVEL_HACK)
#ifndef __ARCH_HAS_4LEVEL_HACK
static inline pud_t *pud_alloc(struct mm_struct *mm, pgd_t *pgd, unsigned long address) static inline pud_t *pud_alloc(struct mm_struct *mm, pgd_t *pgd, unsigned long address)
{ {
if (pgd_none(*pgd)) return (unlikely(pgd_none(*pgd)) && __pud_alloc(mm, pgd, address))?
return __pud_alloc(mm, pgd, address); NULL: pud_offset(pgd, address);
return pud_offset(pgd, address);
} }
static inline pmd_t *pmd_alloc(struct mm_struct *mm, pud_t *pud, unsigned long address) static inline pmd_t *pmd_alloc(struct mm_struct *mm, pud_t *pud, unsigned long address)
{ {
if (pud_none(*pud)) return (unlikely(pud_none(*pud)) && __pmd_alloc(mm, pud, address))?
return __pmd_alloc(mm, pud, address); NULL: pmd_offset(pud, address);
return pmd_offset(pud, address);
} }
#endif #endif /* CONFIG_MMU && !__ARCH_HAS_4LEVEL_HACK */
#endif /* CONFIG_MMU */
#define pte_alloc_map(mm, pmd, address) \
((unlikely(!pmd_present(*(pmd))) && __pte_alloc(mm, pmd, address))? \
NULL: pte_offset_map(pmd, address))
#define pte_alloc_kernel(pmd, address) \
((unlikely(!pmd_present(*(pmd))) && __pte_alloc_kernel(pmd, address))? \
NULL: pte_offset_kernel(pmd, address))
extern void free_area_init(unsigned long * zones_size); extern void free_area_init(unsigned long * zones_size);
extern void free_area_init_node(int nid, pg_data_t *pgdat, extern void free_area_init_node(int nid, pg_data_t *pgdat,
......
...@@ -280,50 +280,39 @@ void free_pgtables(struct mmu_gather **tlb, struct vm_area_struct *vma, ...@@ -280,50 +280,39 @@ void free_pgtables(struct mmu_gather **tlb, struct vm_area_struct *vma,
} }
} }
pte_t fastcall *pte_alloc_map(struct mm_struct *mm, pmd_t *pmd, int __pte_alloc(struct mm_struct *mm, pmd_t *pmd, unsigned long address)
unsigned long address)
{ {
if (!pmd_present(*pmd)) {
struct page *new; struct page *new;
spin_unlock(&mm->page_table_lock); spin_unlock(&mm->page_table_lock);
new = pte_alloc_one(mm, address); new = pte_alloc_one(mm, address);
spin_lock(&mm->page_table_lock); spin_lock(&mm->page_table_lock);
if (!new) if (!new)
return NULL; return -ENOMEM;
/*
* Because we dropped the lock, we should re-check the if (pmd_present(*pmd)) /* Another has populated it */
* entry, as somebody else could have populated it..
*/
if (pmd_present(*pmd)) {
pte_free(new); pte_free(new);
goto out; else {
}
mm->nr_ptes++; mm->nr_ptes++;
inc_page_state(nr_page_table_pages); inc_page_state(nr_page_table_pages);
pmd_populate(mm, pmd, new); pmd_populate(mm, pmd, new);
} }
out: return 0;
return pte_offset_map(pmd, address);
} }
pte_t fastcall * pte_alloc_kernel(pmd_t *pmd, unsigned long address) int __pte_alloc_kernel(pmd_t *pmd, unsigned long address)
{ {
if (!pmd_present(*pmd)) { pte_t *new = pte_alloc_one_kernel(&init_mm, address);
pte_t *new;
new = pte_alloc_one_kernel(&init_mm, address);
if (!new) if (!new)
return NULL; return -ENOMEM;
spin_lock(&init_mm.page_table_lock); spin_lock(&init_mm.page_table_lock);
if (pmd_present(*pmd)) if (pmd_present(*pmd)) /* Another has populated it */
pte_free_kernel(new); pte_free_kernel(new);
else else
pmd_populate_kernel(&init_mm, pmd, new); pmd_populate_kernel(&init_mm, pmd, new);
spin_unlock(&init_mm.page_table_lock); spin_unlock(&init_mm.page_table_lock);
} return 0;
return pte_offset_kernel(pmd, address);
} }
static inline void add_mm_rss(struct mm_struct *mm, int file_rss, int anon_rss) static inline void add_mm_rss(struct mm_struct *mm, int file_rss, int anon_rss)
...@@ -2093,7 +2082,7 @@ int __handle_mm_fault(struct mm_struct *mm, struct vm_area_struct *vma, ...@@ -2093,7 +2082,7 @@ int __handle_mm_fault(struct mm_struct *mm, struct vm_area_struct *vma,
* Allocate page upper directory. * Allocate page upper directory.
* We've already handled the fast-path in-line. * We've already handled the fast-path in-line.
*/ */
pud_t fastcall *__pud_alloc(struct mm_struct *mm, pgd_t *pgd, unsigned long address) int __pud_alloc(struct mm_struct *mm, pgd_t *pgd, unsigned long address)
{ {
pud_t *new; pud_t *new;
...@@ -2103,19 +2092,17 @@ pud_t fastcall *__pud_alloc(struct mm_struct *mm, pgd_t *pgd, unsigned long addr ...@@ -2103,19 +2092,17 @@ pud_t fastcall *__pud_alloc(struct mm_struct *mm, pgd_t *pgd, unsigned long addr
if (!new) { if (!new) {
if (mm != &init_mm) /* Temporary bridging hack */ if (mm != &init_mm) /* Temporary bridging hack */
spin_lock(&mm->page_table_lock); spin_lock(&mm->page_table_lock);
return NULL; return -ENOMEM;
} }
spin_lock(&mm->page_table_lock); spin_lock(&mm->page_table_lock);
if (pgd_present(*pgd)) { if (pgd_present(*pgd)) /* Another has populated it */
pud_free(new); pud_free(new);
goto out; else
}
pgd_populate(mm, pgd, new); pgd_populate(mm, pgd, new);
out:
if (mm == &init_mm) /* Temporary bridging hack */ if (mm == &init_mm) /* Temporary bridging hack */
spin_unlock(&mm->page_table_lock); spin_unlock(&mm->page_table_lock);
return pud_offset(pgd, address); return 0;
} }
#endif /* __PAGETABLE_PUD_FOLDED */ #endif /* __PAGETABLE_PUD_FOLDED */
...@@ -2124,7 +2111,7 @@ pud_t fastcall *__pud_alloc(struct mm_struct *mm, pgd_t *pgd, unsigned long addr ...@@ -2124,7 +2111,7 @@ pud_t fastcall *__pud_alloc(struct mm_struct *mm, pgd_t *pgd, unsigned long addr
* Allocate page middle directory. * Allocate page middle directory.
* We've already handled the fast-path in-line. * We've already handled the fast-path in-line.
*/ */
pmd_t fastcall *__pmd_alloc(struct mm_struct *mm, pud_t *pud, unsigned long address) int __pmd_alloc(struct mm_struct *mm, pud_t *pud, unsigned long address)
{ {
pmd_t *new; pmd_t *new;
...@@ -2134,28 +2121,24 @@ pmd_t fastcall *__pmd_alloc(struct mm_struct *mm, pud_t *pud, unsigned long addr ...@@ -2134,28 +2121,24 @@ pmd_t fastcall *__pmd_alloc(struct mm_struct *mm, pud_t *pud, unsigned long addr
if (!new) { if (!new) {
if (mm != &init_mm) /* Temporary bridging hack */ if (mm != &init_mm) /* Temporary bridging hack */
spin_lock(&mm->page_table_lock); spin_lock(&mm->page_table_lock);
return NULL; return -ENOMEM;
} }
spin_lock(&mm->page_table_lock); spin_lock(&mm->page_table_lock);
#ifndef __ARCH_HAS_4LEVEL_HACK #ifndef __ARCH_HAS_4LEVEL_HACK
if (pud_present(*pud)) { if (pud_present(*pud)) /* Another has populated it */
pmd_free(new); pmd_free(new);
goto out; else
}
pud_populate(mm, pud, new); pud_populate(mm, pud, new);
#else #else
if (pgd_present(*pud)) { if (pgd_present(*pud)) /* Another has populated it */
pmd_free(new); pmd_free(new);
goto out; else
}
pgd_populate(mm, pud, new); pgd_populate(mm, pud, new);
#endif /* __ARCH_HAS_4LEVEL_HACK */ #endif /* __ARCH_HAS_4LEVEL_HACK */
out:
if (mm == &init_mm) /* Temporary bridging hack */ if (mm == &init_mm) /* Temporary bridging hack */
spin_unlock(&mm->page_table_lock); spin_unlock(&mm->page_table_lock);
return pmd_offset(pud, address); return 0;
} }
#endif /* __PAGETABLE_PMD_FOLDED */ #endif /* __PAGETABLE_PMD_FOLDED */
......
...@@ -51,7 +51,6 @@ static pmd_t *alloc_new_pmd(struct mm_struct *mm, unsigned long addr) ...@@ -51,7 +51,6 @@ static pmd_t *alloc_new_pmd(struct mm_struct *mm, unsigned long addr)
pgd_t *pgd; pgd_t *pgd;
pud_t *pud; pud_t *pud;
pmd_t *pmd = NULL; pmd_t *pmd = NULL;
pte_t *pte;
/* /*
* We do need page_table_lock: because allocators expect that. * We do need page_table_lock: because allocators expect that.
...@@ -66,12 +65,8 @@ static pmd_t *alloc_new_pmd(struct mm_struct *mm, unsigned long addr) ...@@ -66,12 +65,8 @@ static pmd_t *alloc_new_pmd(struct mm_struct *mm, unsigned long addr)
if (!pmd) if (!pmd)
goto out; goto out;
pte = pte_alloc_map(mm, pmd, addr); if (!pmd_present(*pmd) && __pte_alloc(mm, pmd, addr))
if (!pte) {
pmd = NULL; pmd = NULL;
goto out;
}
pte_unmap(pte);
out: out:
spin_unlock(&mm->page_table_lock); spin_unlock(&mm->page_table_lock);
return pmd; return pmd;
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment