diff options
| author | Hugh Dickins <hugh@veritas.com> | 2005-10-29 18:16:40 -0700 |
|---|---|---|
| committer | Linus Torvalds <torvalds@g5.osdl.org> | 2005-10-29 21:40:42 -0700 |
| commit | 4c21e2f2441dc5fbb957b030333f5a3f2d02dea7 (patch) | |
| tree | 1f76d33bb1d76221c6424bc5fed080a4f91349a6 /mm/filemap.c | |
| parent | b38c6845b695141259019e2b7c0fe6c32a6e720d (diff) | |
[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 'mm/filemap.c')
| -rw-r--r-- | mm/filemap.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/mm/filemap.c b/mm/filemap.c index 8aa344e88489..f560b41c8f61 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -152,7 +152,7 @@ static int sync_page(void *word) * in the ->sync_page() methods make essential use of the * page_mapping(), merely passing the page down to the backing * device's unplug functions when it's non-NULL, which in turn - * ignore it for all cases but swap, where only page->private is + * ignore it for all cases but swap, where only page_private(page) is * of interest. When page_mapping() does go NULL, the entire * call stack gracefully ignores the page and returns. * -- wli |