The patch titled
mm: optimize page_lock_anon_vma() fast-path
has been added to the -mm tree. Its filename is
mm-optimize-page_lock_anon_vma-fast-path.patch
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/SubmitChecklist when testing your code ***
See http://userweb.kernel.org/~akpm/stuff/added-to-mm.txt to find
out what to do about this
The current -mm tree may be found at http://userweb.kernel.org/~akpm/mmotm/
------------------------------------------------------
Subject: mm: optimize page_lock_anon_vma() fast-path
From: Peter Zijlstra <a.p.zijlstra@xxxxxxxxx>
Optimize the page_lock_anon_vma() fast path to be one atomic op, instead
of two.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@xxxxxxxxx>
Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
Cc: Benjamin Herrenschmidt <benh@xxxxxxxxxxxxxxxxxxx>
Cc: David Miller <davem@xxxxxxxxxxxxx>
Cc: Martin Schwidefsky <schwidefsky@xxxxxxxxxx>
Cc: Russell King <rmk@xxxxxxxxxxxxxxxx>
Cc: Paul Mundt <lethal@xxxxxxxxxxxx>
Cc: Jeff Dike <jdike@xxxxxxxxxxx>
Cc: Richard Weinberger <richard@xxxxxx>
Cc: Tony Luck <tony.luck@xxxxxxxxx>
Cc: Hugh Dickins <hughd@xxxxxxxxxx>
Cc: Mel Gorman <mel@xxxxxxxxx>
Cc: KOSAKI Motohiro <kosaki.motohiro@xxxxxxxxxxxxxx>
Cc: Nick Piggin <npiggin@xxxxxxxxx>
Cc: Namhyung Kim <namhyung@xxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
mm/rmap.c | 86 +++++++++++++++++++++++++++++++++++++++++++++++++---
1 file changed, 82 insertions(+), 4 deletions(-)
diff -puN mm/rmap.c~mm-optimize-page_lock_anon_vma-fast-path mm/rmap.c
--- a/mm/rmap.c~mm-optimize-page_lock_anon_vma-fast-path
+++ a/mm/rmap.c
@@ -86,6 +86,29 @@ static inline struct anon_vma *anon_vma_
static inline void anon_vma_free(struct anon_vma *anon_vma)
{
VM_BUG_ON(atomic_read(&anon_vma->refcount));
+
+ /*
+ * Synchronize against page_lock_anon_vma() such that
+ * we can safely hold the lock without the anon_vma getting
+ * freed.
+ *
+ * Relies on the full mb implied by the atomic_dec_and_test() from
+ * put_anon_vma() against the acquire barrier implied by
+ * mutex_trylock() from page_lock_anon_vma(). This orders:
+ *
+ * page_lock_anon_vma() VS put_anon_vma()
+ * mutex_trylock() atomic_dec_and_test()
+ * LOCK MB
+ * atomic_read() mutex_is_locked()
+ *
+ * LOCK should suffice since the actual taking of the lock must
+ * happen _before_ what follows.
+ */
+ if (mutex_is_locked(&anon_vma->root->mutex)) {
+ anon_vma_lock(anon_vma);
+ anon_vma_unlock(anon_vma);
+ }
+
kmem_cache_free(anon_vma_cachep, anon_vma);
}
@@ -372,20 +395,75 @@ out:
return anon_vma;
}
+/*
+ * Similar to page_get_anon_vma() except it locks the anon_vma.
+ *
+ * Its a little more complex as it tries to keep the fast path to a single
+ * atomic op -- the trylock. If we fail the trylock, we fall back to getting a
+ * reference like with page_get_anon_vma() and then block on the mutex.
+ */
struct anon_vma *page_lock_anon_vma(struct page *page)
{
- struct anon_vma *anon_vma = page_get_anon_vma(page);
+ struct anon_vma *anon_vma = NULL;
+ unsigned long anon_mapping;
- if (anon_vma)
- anon_vma_lock(anon_vma);
+ rcu_read_lock();
+ anon_mapping = (unsigned long) ACCESS_ONCE(page->mapping);
+ if ((anon_mapping & PAGE_MAPPING_FLAGS) != PAGE_MAPPING_ANON)
+ goto out;
+ if (!page_mapped(page))
+ goto out;
+
+ anon_vma = (struct anon_vma *) (anon_mapping - PAGE_MAPPING_ANON);
+ if (mutex_trylock(&anon_vma->root->mutex)) {
+ /*
+ * If we observe a !0 refcount, then holding the lock ensures
+ * the anon_vma will not go away, see __put_anon_vma().
+ */
+ if (!atomic_read(&anon_vma->refcount)) {
+ anon_vma_unlock(anon_vma);
+ anon_vma = NULL;
+ }
+ goto out;
+ }
+
+ /* trylock failed, we got to sleep */
+ if (!atomic_inc_not_zero(&anon_vma->refcount)) {
+ anon_vma = NULL;
+ goto out;
+ }
+ if (!page_mapped(page)) {
+ put_anon_vma(anon_vma);
+ anon_vma = NULL;
+ goto out;
+ }
+
+ /* we pinned the anon_vma, its safe to sleep */
+ rcu_read_unlock();
+ anon_vma_lock(anon_vma);
+
+ if (atomic_dec_and_test(&anon_vma->refcount)) {
+ /*
+ * Oops, we held the last refcount, release the lock
+ * and bail -- can't simply use put_anon_vma() because
+ * we'll deadlock on the anon_vma_lock() recursion.
+ */
+ anon_vma_unlock(anon_vma);
+ __put_anon_vma(anon_vma);
+ anon_vma = NULL;
+ }
+
+ return anon_vma;
+
+out:
+ rcu_read_unlock();
return anon_vma;
}
void page_unlock_anon_vma(struct anon_vma *anon_vma)
{
anon_vma_unlock(anon_vma);
- put_anon_vma(anon_vma);
}
/*
_
Patches currently in -mm which might be from a.p.zijlstra@xxxxxxxxx are
linux-next.patch
net-convert-%p-usage-to-%pk.patch
mm-mmu_gather-rework.patch
mm-mmu_gather-rework-fix.patch
powerpc-mmu_gather-rework.patch
sparc-mmu_gather-rework.patch
s390-mmu_gather-rework.patch
arm-mmu_gather-rework.patch
sh-mmu_gather-rework.patch
ia64-mmu_gather-rework.patch
um-mmu_gather-rework.patch
mm-now-that-all-old-mmu_gather-code-is-gone-remove-the-storage.patch
mm-powerpc-move-the-rcu-page-table-freeing-into-generic-code.patch
mm-extended-batches-for-generic-mmu_gather.patch
lockdep-mutex-provide-mutex_lock_nest_lock.patch
mm-remove-i_mmap_lock-lockbreak.patch
mm-convert-i_mmap_lock-to-a-mutex.patch
mm-revert-page_lock_anon_vma-lock-annotation.patch
mm-improve-page_lock_anon_vma-comment.patch
mm-use-refcounts-for-page_lock_anon_vma.patch
mm-convert-anon_vma-lock-to-a-mutex.patch
mm-optimize-page_lock_anon_vma-fast-path.patch
--
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
