[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>

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);				\