From: Peter Zijlstra <a.p.zijlstra@chello.nl>
To: Andrea Arcangeli <aarcange@redhat.com>,
Avi Kivity <avi@redhat.com>, Thomas Gleixner <tglx@linutronix.de>,
Rik van Riel <riel@redhat.com>, Ingo Molnar <mingo@elte.hu>,
akpm@linux-fou
Cc: linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org,
Benjamin Herrenschmidt <benh@kernel.crashing.org>,
David Miller <davem@davemloft.net>,
Hugh Dickins <hugh.dickins@tiscali.co.uk>,
Mel Gorman <mel@csn.ul.ie>, Nick Piggin <npiggin@suse.de>,
Peter Zijlstra <a.p.zijlstra@chello.nl>,
"Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Subject: [PATCH 28/28] mm: Optimize page_lock_anon_vma() fast-path
Date: Mon, 07 Jun 2010 13:07:22 +0200 [thread overview]
Message-ID: <20100607111408.955268806@chello.nl> (raw)
In-Reply-To: 20100607110654.606530953@chello.nl
[-- Attachment #1: mm-opt-page_lock_anon_vma.patch --]
[-- Type: text/plain, Size: 2722 bytes --]
Optimize the page_lock_anon_vma() fast path to be one LOCKed op,
instead of two.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
---
mm/rmap.c | 65 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++----
1 file changed, 61 insertions(+), 4 deletions(-)
Index: linux-2.6/mm/rmap.c
===================================================================
--- linux-2.6.orig/mm/rmap.c
+++ linux-2.6/mm/rmap.c
@@ -351,20 +351,69 @@ 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;
+
+ 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->lock)) {
+ /*
+ * 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;
+ }
+
+ /* 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;
+ }
- if (anon_vma)
- anon_vma_lock(anon_vma);
+ 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);
}
/*
@@ -1422,6 +1471,14 @@ int try_to_munlock(struct page *page)
void __put_anon_vma(struct anon_vma *anon_vma)
{
+ /*
+ * Synchronize against page_lock_anon_vma() such that
+ * we can safely hold the lock without the anon_vma getting
+ * freed.
+ */
+ anon_vma_lock(anon_vma);
+ anon_vma_unlock(anon_vma);
+
if (anon_vma->root != anon_vma)
put_anon_vma(anon_vma->root);
anon_vma_free(anon_vma);
WARNING: multiple messages have this Message-ID (diff)
From: Peter Zijlstra <a.p.zijlstra@chello.nl>
To: Andrea Arcangeli <aarcange@redhat.com>,
Avi Kivity <avi@redhat.com>, Thomas Gleixner <tglx@linutronix.de>,
Rik van Riel <riel@redhat.com>, Ingo Molnar <mingo@elte.hu>,
akpm@linux-foundation.org,
Linus Torvalds <torvalds@linux-foundation.org>
Cc: linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org,
Benjamin Herrenschmidt <benh@kernel.crashing.org>,
David Miller <davem@davemloft.net>,
Hugh Dickins <hugh.dickins@tiscali.co.uk>,
Mel Gorman <mel@csn.ul.ie>, Nick Piggin <npiggin@suse.de>,
Peter Zijlstra <a.p.zijlstra@chello.nl>,
"Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Subject: [PATCH 28/28] mm: Optimize page_lock_anon_vma() fast-path
Date: Mon, 07 Jun 2010 13:07:22 +0200 [thread overview]
Message-ID: <20100607111408.955268806@chello.nl> (raw)
Message-ID: <20100607110722.heOsozbjCUeMydJ0eHdw1OuKu7EnVRFETRapb1WDfnc@z> (raw)
In-Reply-To: 20100607110654.606530953@chello.nl
[-- Attachment #1: mm-opt-page_lock_anon_vma.patch --]
[-- Type: text/plain, Size: 2724 bytes --]
Optimize the page_lock_anon_vma() fast path to be one LOCKed op,
instead of two.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
---
mm/rmap.c | 65 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++----
1 file changed, 61 insertions(+), 4 deletions(-)
Index: linux-2.6/mm/rmap.c
===================================================================
--- linux-2.6.orig/mm/rmap.c
+++ linux-2.6/mm/rmap.c
@@ -351,20 +351,69 @@ 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;
+
+ 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->lock)) {
+ /*
+ * 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;
+ }
+
+ /* 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;
+ }
- if (anon_vma)
- anon_vma_lock(anon_vma);
+ 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);
}
/*
@@ -1422,6 +1471,14 @@ int try_to_munlock(struct page *page)
void __put_anon_vma(struct anon_vma *anon_vma)
{
+ /*
+ * Synchronize against page_lock_anon_vma() such that
+ * we can safely hold the lock without the anon_vma getting
+ * freed.
+ */
+ anon_vma_lock(anon_vma);
+ anon_vma_unlock(anon_vma);
+
if (anon_vma->root != anon_vma)
put_anon_vma(anon_vma->root);
anon_vma_free(anon_vma);
next prev parent reply other threads:[~2010-06-07 11:07 UTC|newest]
Thread overview: 81+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-06-07 11:06 [PATCH 00/28] mm: preemptibility -v3 Peter Zijlstra
2010-06-07 11:06 ` Peter Zijlstra
2010-06-07 11:06 ` [PATCH 01/28] powerpc: Use call_rcu_sched() for pagetables Peter Zijlstra
2010-06-07 11:06 ` Peter Zijlstra
2010-06-07 11:06 ` [PATCH 02/28] mm: Improve page_lock_anon_vma() comment Peter Zijlstra
2010-06-07 11:06 ` Peter Zijlstra
2010-06-09 10:50 ` Mel Gorman
2010-06-07 11:06 ` [PATCH 03/28] rename anon_vma_lock to vma_lock_anon_vma Peter Zijlstra
2010-06-07 11:06 ` Peter Zijlstra
2010-06-07 15:01 ` Peter Zijlstra
2010-06-09 10:59 ` Mel Gorman
2010-06-07 11:06 ` [PATCH 04/28] change direct call of spin_lock(anon_vma->lock) to inline function Peter Zijlstra
2010-06-07 11:06 ` Peter Zijlstra
2010-06-07 11:06 ` [PATCH 05/28] track the root (oldest) anon_vma Peter Zijlstra
2010-06-07 11:06 ` Peter Zijlstra
2010-06-07 11:07 ` [PATCH 06/28] always lock " Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-07 11:07 ` [PATCH 07/28] extend KSM refcounts to the anon_vma root Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-09 10:55 ` Mel Gorman
2010-06-07 11:07 ` [PATCH 08/28] mm: Rename drop_anon_vma to put_anon_vma Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-07 11:07 ` [PATCH 09/28] mm: Move anon_vma ref out from under CONFIG_KSM Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-09 11:05 ` Mel Gorman
2010-06-09 11:05 ` Mel Gorman
2010-06-07 11:07 ` [PATCH 10/28] mm: Make use of the anon_vma ref count Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-09 13:22 ` Mel Gorman
2010-06-09 13:32 ` Peter Zijlstra
2010-06-07 11:07 ` [PATCH 11/28] mm: Preemptible mmu_gather Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-07 11:07 ` [PATCH 12/28] powerpc: " Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-07 11:07 ` [PATCH 13/28] sparc: " Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-07 11:07 ` [PATCH 14/28] s390: preemptible mmu_gather Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-07 11:07 ` [PATCH 15/28] arm: Preemptible mmu_gather Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-07 11:07 ` [PATCH 16/28] sh: " Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-07 11:38 ` Peter Zijlstra
2010-06-07 11:38 ` Peter Zijlstra
2010-06-07 11:07 ` [PATCH 17/28] um: " Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-07 11:07 ` [PATCH 18/28] ia64: " Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-07 11:07 ` [PATCH 19/28] mm, powerpc: Move the RCU page-table freeing into generic code Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-07 11:07 ` [PATCH 20/28] sparc64: Kill page table quicklists Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-07 11:07 ` [PATCH 21/28] sparc64: Use RCU page table freeing Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-07 11:07 ` [PATCH 22/28] sparc64: Add support for _PAGE_SPECIAL Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-07 11:07 ` [PATCH 23/28] sparc64: Implement get_user_pages_fast() Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-07 11:07 ` [PATCH 24/28] lockdep, mutex: Provide mutex_lock_nest_lock Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-07 11:07 ` [PATCH 25/28] mutex: Provide mutex_is_contended Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-07 11:07 ` [PATCH 26/28] mm: Convert i_mmap_lock and anon_vma->lock to mutexes Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-07 11:07 ` [PATCH 27/28] mm: Extended batches for generic mmu_gather Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra
2010-06-07 11:07 ` Peter Zijlstra [this message]
2010-06-07 11:07 ` [PATCH 28/28] mm: Optimize page_lock_anon_vma() fast-path Peter Zijlstra
2010-06-07 13:57 ` [PATCH 00/28] mm: preemptibility -v3 Sam Ravnborg
2010-06-07 15:02 ` Peter Zijlstra
2010-06-07 16:36 ` Andi Kleen
2010-06-07 16:39 ` Peter Zijlstra
2010-06-07 16:39 ` Peter Zijlstra
2010-06-10 1:45 ` Zhang, Yanmin
2010-06-10 6:52 ` Peter Zijlstra
2010-06-10 6:59 ` Zhang, Yanmin
2010-06-21 10:21 ` Peter Zijlstra
2010-06-24 9:55 ` Peter Zijlstra
2010-06-29 7:40 ` Zhang, Yanmin
2010-06-29 7:40 ` Zhang, Yanmin
2010-06-29 7:48 ` Peter Zijlstra
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20100607111408.955268806@chello.nl \
--to=a.p.zijlstra@chello.nl \
--cc=aarcange@redhat.com \
--cc=akpm@linux-fou \
--cc=avi@redhat.com \
--cc=benh@kernel.crashing.org \
--cc=davem@davemloft.net \
--cc=hugh.dickins@tiscali.co.uk \
--cc=linux-arch@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mel@csn.ul.ie \
--cc=mingo@elte.hu \
--cc=npiggin@suse.de \
--cc=paulmck@linux.vnet.ibm.com \
--cc=riel@redhat.com \
--cc=tglx@linutronix.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).