summaryrefslogtreecommitdiffstats
path: root/include
diff options
context:
space:
mode:
authorHugh Dickins <hugh@veritas.com>2005-10-29 18:16:40 -0700
committerLinus Torvalds <torvalds@g5.osdl.org>2005-10-29 21:40:42 -0700
commit4c21e2f2441dc5fbb957b030333f5a3f2d02dea7 (patch)
tree1f76d33bb1d76221c6424bc5fed080a4f91349a6 /include
parentb38c6845b695141259019e2b7c0fe6c32a6e720d (diff)
downloadop-kernel-dev-4c21e2f2441dc5fbb957b030333f5a3f2d02dea7.zip
op-kernel-dev-4c21e2f2441dc5fbb957b030333f5a3f2d02dea7.tar.gz
[PATCH] mm: split page table lock
Christoph Lameter demonstrated very poor scalability on the SGI 512-way, with a many-threaded application which concurrently initializes different parts of a large anonymous area. This patch corrects that, by using a separate spinlock per page table page, to guard the page table entries in that page, instead of using the mm's single page_table_lock. (But even then, page_table_lock is still used to guard page table allocation, and anon_vma allocation.) In this implementation, the spinlock is tucked inside the struct page of the page table page: with a BUILD_BUG_ON in case it overflows - which it would in the case of 32-bit PA-RISC with spinlock debugging enabled. Splitting the lock is not quite for free: another cacheline access. Ideally, I suppose we would use split ptlock only for multi-threaded processes on multi-cpu machines; but deciding that dynamically would have its own costs. So for now enable it by config, at some number of cpus - since the Kconfig language doesn't support inequalities, let preprocessor compare that with NR_CPUS. But I don't think it's worth being user-configurable: for good testing of both split and unsplit configs, split now at 4 cpus, and perhaps change that to 8 later. There is a benefit even for singly threaded processes: kswapd can be attacking one part of the mm while another part is busy faulting. Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Diffstat (limited to 'include')
-rw-r--r--include/linux/buffer_head.h6
-rw-r--r--include/linux/mm.h46
2 files changed, 41 insertions, 11 deletions
diff --git a/include/linux/buffer_head.h b/include/linux/buffer_head.h
index 88af42f..c937d6e 100644
--- a/include/linux/buffer_head.h
+++ b/include/linux/buffer_head.h
@@ -126,8 +126,8 @@ BUFFER_FNS(Eopnotsupp, eopnotsupp)
/* If we *know* page->private refers to buffer_heads */
#define page_buffers(page) \
({ \
- BUG_ON(!PagePrivate(page)); \
- ((struct buffer_head *)(page)->private); \
+ BUG_ON(!PagePrivate(page)); \
+ ((struct buffer_head *)page_private(page)); \
})
#define page_has_buffers(page) PagePrivate(page)
@@ -219,7 +219,7 @@ static inline void attach_page_buffers(struct page *page,
{
page_cache_get(page);
SetPagePrivate(page);
- page->private = (unsigned long)head;
+ set_page_private(page, (unsigned long)head);
}
static inline void get_bh(struct buffer_head *bh)
diff --git a/include/linux/mm.h b/include/linux/mm.h
index e8d1424..8a514ec 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -226,13 +226,18 @@ struct page {
* to show when page is mapped
* & limit reverse map searches.
*/
- unsigned long private; /* Mapping-private opaque data:
+ union {
+ unsigned long private; /* Mapping-private opaque data:
* usually used for buffer_heads
* if PagePrivate set; used for
* swp_entry_t if PageSwapCache
* When page is free, this indicates
* order in the buddy system.
*/
+#if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
+ spinlock_t ptl;
+#endif
+ } u;
struct address_space *mapping; /* If low bit clear, points to
* inode address_space, or NULL.
* If page mapped as anonymous
@@ -260,6 +265,9 @@ struct page {
#endif /* WANT_PAGE_VIRTUAL */
};
+#define page_private(page) ((page)->u.private)
+#define set_page_private(page, v) ((page)->u.private = (v))
+
/*
* FIXME: take this include out, include page-flags.h in
* files which need it (119 of them)
@@ -311,17 +319,17 @@ extern void FASTCALL(__page_cache_release(struct page *));
#ifdef CONFIG_HUGETLB_PAGE
-static inline int page_count(struct page *p)
+static inline int page_count(struct page *page)
{
- if (PageCompound(p))
- p = (struct page *)p->private;
- return atomic_read(&(p)->_count) + 1;
+ if (PageCompound(page))
+ page = (struct page *)page_private(page);
+ return atomic_read(&page->_count) + 1;
}
static inline void get_page(struct page *page)
{
if (unlikely(PageCompound(page)))
- page = (struct page *)page->private;
+ page = (struct page *)page_private(page);
atomic_inc(&page->_count);
}
@@ -587,7 +595,7 @@ static inline int PageAnon(struct page *page)
static inline pgoff_t page_index(struct page *page)
{
if (unlikely(PageSwapCache(page)))
- return page->private;
+ return page_private(page);
return page->index;
}
@@ -779,9 +787,31 @@ static inline pmd_t *pmd_alloc(struct mm_struct *mm, pud_t *pud, unsigned long a
}
#endif /* CONFIG_MMU && !__ARCH_HAS_4LEVEL_HACK */
+#if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
+/*
+ * We tuck a spinlock to guard each pagetable page into its struct page,
+ * at page->private, with BUILD_BUG_ON to make sure that this will not
+ * overflow into the next struct page (as it might with DEBUG_SPINLOCK).
+ * When freeing, reset page->mapping so free_pages_check won't complain.
+ */
+#define __pte_lockptr(page) &((page)->u.ptl)
+#define pte_lock_init(_page) do { \
+ spin_lock_init(__pte_lockptr(_page)); \
+} while (0)
+#define pte_lock_deinit(page) ((page)->mapping = NULL)
+#define pte_lockptr(mm, pmd) ({(void)(mm); __pte_lockptr(pmd_page(*(pmd)));})
+#else
+/*
+ * We use mm->page_table_lock to guard all pagetable pages of the mm.
+ */
+#define pte_lock_init(page) do {} while (0)
+#define pte_lock_deinit(page) do {} while (0)
+#define pte_lockptr(mm, pmd) ({(void)(pmd); &(mm)->page_table_lock;})
+#endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
+
#define pte_offset_map_lock(mm, pmd, address, ptlp) \
({ \
- spinlock_t *__ptl = &(mm)->page_table_lock; \
+ spinlock_t *__ptl = pte_lockptr(mm, pmd); \
pte_t *__pte = pte_offset_map(pmd, address); \
*(ptlp) = __ptl; \
spin_lock(__ptl); \
OpenPOWER on IntegriCloud