linux-arch.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
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);



  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).