From: Peter Zijlstra <a.p.zijlstra@chello.nl>
To: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Tim Chen <tim.c.chen@linux.intel.com>,
Andi Kleen <ak@linux.intel.com>,
Shaohua Li <shaohua.li@intel.com>,
Andrew Morton <akpm@linux-foundation.org>,
Hugh Dickins <hughd@google.com>,
KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>,
Benjamin Herrenschmidt <benh@kernel.crashing.org>,
David Miller <davem@davemloft.net>,
Martin Schwidefsky <schwidefsky@de.ibm.com>,
Russell King <rmk@arm.linux.org.uk>,
Paul Mundt <lethal@linux-sh.org>, Jeff Dike <jdike@addtoit.com>,
Richard Weinberger <richard@nod.at>,
"Luck, Tony" <tony.luck@intel.com>,
KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>,
Mel Gorman <mel@csn.ul.ie>, Nick Piggin <npiggin@kernel.dk>,
Namhyung Kim <namhyung@gmail.com>,
"Shi, Alex" <alex.shi@intel.com>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
"linux-mm@kvack.org" <linux-mm@kvack.org>,
"Rafael J. Wysocki" <rjw@sisk.pl>
Subject: Re: REGRESSION: Performance regressions from switching anon_vma->lock to mutex
Date: Fri, 17 Jun 2011 13:28:00 +0200 [thread overview]
Message-ID: <1308310080.2355.19.camel@twins> (raw)
In-Reply-To: <BANLkTinUBTYWxrF5TCuDSQuFUAyivXJXjQ@mail.gmail.com>
On Thu, 2011-06-16 at 20:58 -0700, Linus Torvalds wrote:
> Ok, I'm still thinking. I have an approach that I think will handle it
> fairly cleanly, but that involves walking the same_vma list twice:
> once to actually unlink the anon_vma's under the lock, and then a
> second pass that does the rest. It should work.
Something like so? Compiles and runs the benchmark in question.
---
Index: linux-2.6/mm/rmap.c
===================================================================
--- linux-2.6.orig/mm/rmap.c
+++ linux-2.6/mm/rmap.c
@@ -324,36 +324,52 @@ int anon_vma_fork(struct vm_area_struct
return -ENOMEM;
}
-static void anon_vma_unlink(struct anon_vma_chain *anon_vma_chain)
+static int anon_vma_unlink(struct anon_vma_chain *anon_vma_chain, struct anon_vma *anon_vma)
{
- struct anon_vma *anon_vma = anon_vma_chain->anon_vma;
- int empty;
-
- /* If anon_vma_fork fails, we can get an empty anon_vma_chain. */
- if (!anon_vma)
- return;
-
- anon_vma_lock(anon_vma);
list_del(&anon_vma_chain->same_anon_vma);
/* We must garbage collect the anon_vma if it's empty */
- empty = list_empty(&anon_vma->head);
- anon_vma_unlock(anon_vma);
+ if (list_empty(&anon_vma->head))
+ return 1;
- if (empty)
- put_anon_vma(anon_vma);
+ return 0;
}
void unlink_anon_vmas(struct vm_area_struct *vma)
{
struct anon_vma_chain *avc, *next;
+ struct anon_vma *root = NULL;
/*
* Unlink each anon_vma chained to the VMA. This list is ordered
* from newest to oldest, ensuring the root anon_vma gets freed last.
*/
list_for_each_entry_safe(avc, next, &vma->anon_vma_chain, same_vma) {
- anon_vma_unlink(avc);
+ struct anon_vma *anon_vma = avc->anon_vma;
+
+ /* If anon_vma_fork fails, we can get an empty anon_vma_chain. */
+ if (anon_vma) {
+ root = lock_anon_vma_root(root, anon_vma);
+ /* Leave empty anon_vmas on the list. */
+ if (anon_vma_unlink(avc, anon_vma))
+ continue;
+ }
+ list_del(&avc->same_vma);
+ anon_vma_chain_free(avc);
+ }
+ unlock_anon_vma_root(root);
+
+ /*
+ * Iterate the list once more, it now only contains empty and unlinked
+ * anon_vmas, destroy them. Could not do before due to anon_vma_free()
+ * needing to acquire the anon_vma->root->mutex.
+ */
+ list_for_each_entry_safe(avc, next, &vma->anon_vma_chain, same_vma) {
+ struct anon_vma *anon_vma = avc->anon_vma;
+
+ if (anon_vma)
+ put_anon_vma(anon_vma);
+
list_del(&avc->same_vma);
anon_vma_chain_free(avc);
}
WARNING: multiple messages have this Message-ID (diff)
From: Peter Zijlstra <a.p.zijlstra@chello.nl>
To: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Tim Chen <tim.c.chen@linux.intel.com>,
Andi Kleen <ak@linux.intel.com>,
Shaohua Li <shaohua.li@intel.com>,
Andrew Morton <akpm@linux-foundation.org>,
Hugh Dickins <hughd@google.com>,
KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>,
Benjamin Herrenschmidt <benh@kernel.crashing.org>,
David Miller <davem@davemloft.net>,
Martin Schwidefsky <schwidefsky@de.ibm.com>,
Russell King <rmk@arm.linux.org.uk>,
Paul Mundt <lethal@linux-sh.org>, Jeff Dike <jdike@addtoit.com>,
Richard Weinberger <richard@nod.at>,
"Luck, Tony" <tony.luck@intel.com>,
KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>,
Mel Gorman <mel@csn.ul.ie>, Nick Piggin <npiggin@kernel.dk>,
Namhyung Kim <namhyung@gmail.com>,
"Shi, Alex" <alex.shi@intel.com>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
"linux-mm@kvack.org" <linux-mm@kvack.org>,
"Rafael J. Wysocki" <rjw@sisk.pl>
Subject: Re: REGRESSION: Performance regressions from switching anon_vma->lock to mutex
Date: Fri, 17 Jun 2011 13:28:00 +0200 [thread overview]
Message-ID: <1308310080.2355.19.camel@twins> (raw)
In-Reply-To: <BANLkTinUBTYWxrF5TCuDSQuFUAyivXJXjQ@mail.gmail.com>
On Thu, 2011-06-16 at 20:58 -0700, Linus Torvalds wrote:
> Ok, I'm still thinking. I have an approach that I think will handle it
> fairly cleanly, but that involves walking the same_vma list twice:
> once to actually unlink the anon_vma's under the lock, and then a
> second pass that does the rest. It should work.
Something like so? Compiles and runs the benchmark in question.
---
Index: linux-2.6/mm/rmap.c
===================================================================
--- linux-2.6.orig/mm/rmap.c
+++ linux-2.6/mm/rmap.c
@@ -324,36 +324,52 @@ int anon_vma_fork(struct vm_area_struct
return -ENOMEM;
}
-static void anon_vma_unlink(struct anon_vma_chain *anon_vma_chain)
+static int anon_vma_unlink(struct anon_vma_chain *anon_vma_chain, struct anon_vma *anon_vma)
{
- struct anon_vma *anon_vma = anon_vma_chain->anon_vma;
- int empty;
-
- /* If anon_vma_fork fails, we can get an empty anon_vma_chain. */
- if (!anon_vma)
- return;
-
- anon_vma_lock(anon_vma);
list_del(&anon_vma_chain->same_anon_vma);
/* We must garbage collect the anon_vma if it's empty */
- empty = list_empty(&anon_vma->head);
- anon_vma_unlock(anon_vma);
+ if (list_empty(&anon_vma->head))
+ return 1;
- if (empty)
- put_anon_vma(anon_vma);
+ return 0;
}
void unlink_anon_vmas(struct vm_area_struct *vma)
{
struct anon_vma_chain *avc, *next;
+ struct anon_vma *root = NULL;
/*
* Unlink each anon_vma chained to the VMA. This list is ordered
* from newest to oldest, ensuring the root anon_vma gets freed last.
*/
list_for_each_entry_safe(avc, next, &vma->anon_vma_chain, same_vma) {
- anon_vma_unlink(avc);
+ struct anon_vma *anon_vma = avc->anon_vma;
+
+ /* If anon_vma_fork fails, we can get an empty anon_vma_chain. */
+ if (anon_vma) {
+ root = lock_anon_vma_root(root, anon_vma);
+ /* Leave empty anon_vmas on the list. */
+ if (anon_vma_unlink(avc, anon_vma))
+ continue;
+ }
+ list_del(&avc->same_vma);
+ anon_vma_chain_free(avc);
+ }
+ unlock_anon_vma_root(root);
+
+ /*
+ * Iterate the list once more, it now only contains empty and unlinked
+ * anon_vmas, destroy them. Could not do before due to anon_vma_free()
+ * needing to acquire the anon_vma->root->mutex.
+ */
+ list_for_each_entry_safe(avc, next, &vma->anon_vma_chain, same_vma) {
+ struct anon_vma *anon_vma = avc->anon_vma;
+
+ if (anon_vma)
+ put_anon_vma(anon_vma);
+
list_del(&avc->same_vma);
anon_vma_chain_free(avc);
}
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2011-06-17 11:29 UTC|newest]
Thread overview: 166+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-06-15 0:29 REGRESSION: Performance regressions from switching anon_vma->lock to mutex Tim Chen
2011-06-15 0:29 ` Tim Chen
2011-06-15 0:36 ` Andi Kleen
2011-06-15 0:36 ` Andi Kleen
2011-06-17 19:07 ` Ingo Molnar
2011-06-17 19:07 ` Ingo Molnar
2011-06-15 1:21 ` Linus Torvalds
2011-06-15 1:21 ` Linus Torvalds
2011-06-15 3:42 ` Linus Torvalds
2011-06-15 1:26 ` Shaohua Li
2011-06-15 1:26 ` Shaohua Li
2011-06-15 11:52 ` Peter Zijlstra
2011-06-15 11:52 ` Peter Zijlstra
2011-06-15 12:49 ` Peter Zijlstra
2011-06-15 12:49 ` Peter Zijlstra
2011-06-15 16:18 ` Andi Kleen
2011-06-15 16:18 ` Andi Kleen
2011-06-15 16:45 ` Peter Zijlstra
2011-06-15 16:45 ` Peter Zijlstra
2011-06-15 16:47 ` Andi Kleen
2011-06-15 16:47 ` Andi Kleen
2011-06-15 18:43 ` Tim Chen
2011-06-15 18:43 ` Tim Chen
2011-06-15 20:32 ` Peter Zijlstra
2011-06-15 20:32 ` Peter Zijlstra
2011-06-15 20:57 ` Andi Kleen
2011-06-15 20:57 ` Andi Kleen
2011-06-15 21:12 ` Tim Chen
2011-06-15 21:12 ` Tim Chen
2011-06-15 21:37 ` Peter Zijlstra
2011-06-15 21:37 ` Peter Zijlstra
2011-06-15 21:51 ` Linus Torvalds
2011-06-15 21:51 ` Linus Torvalds
2011-06-15 22:19 ` Andi Kleen
2011-06-15 22:19 ` Andi Kleen
2011-06-16 0:16 ` Linus Torvalds
2011-06-16 0:16 ` Linus Torvalds
2011-06-16 20:14 ` Andi Kleen
2011-06-16 20:14 ` Andi Kleen
2011-06-16 20:37 ` Linus Torvalds
2011-06-16 20:37 ` Linus Torvalds
2011-06-17 0:24 ` Andi Kleen
2011-06-17 9:13 ` Ingo Molnar
2011-06-17 9:13 ` Ingo Molnar
2011-06-15 22:15 ` Andi Kleen
2011-06-15 22:15 ` Andi Kleen
2011-06-16 1:08 ` Tim Chen
2011-06-16 1:08 ` Tim Chen
2011-06-16 1:50 ` Linus Torvalds
2011-06-16 1:50 ` Linus Torvalds
2011-06-16 20:26 ` Tim Chen
2011-06-16 20:26 ` Tim Chen
2011-06-16 20:47 ` Linus Torvalds
2011-06-16 20:47 ` Linus Torvalds
2011-06-16 21:05 ` Linus Torvalds
2011-06-16 21:05 ` Linus Torvalds
2011-06-16 21:06 ` Linus Torvalds
2011-06-16 21:26 ` Linus Torvalds
2011-06-16 21:26 ` Linus Torvalds
2011-06-17 3:58 ` Linus Torvalds
2011-06-17 11:28 ` Peter Zijlstra [this message]
2011-06-17 11:28 ` Peter Zijlstra
2011-06-17 11:54 ` Peter Zijlstra
2011-06-17 11:54 ` Peter Zijlstra
2011-06-17 16:36 ` Linus Torvalds
2011-06-17 16:36 ` Linus Torvalds
2011-06-17 17:41 ` Hugh Dickins
2011-06-17 17:41 ` Hugh Dickins
2011-06-17 17:55 ` Peter Zijlstra
2011-06-17 17:55 ` Peter Zijlstra
2011-06-17 18:01 ` Linus Torvalds
2011-06-17 18:01 ` Linus Torvalds
2011-06-17 18:18 ` Peter Zijlstra
2011-06-17 18:18 ` Peter Zijlstra
2011-06-17 18:32 ` Peter Zijlstra
2011-06-17 18:32 ` Peter Zijlstra
2011-06-17 18:39 ` Linus Torvalds
2011-06-17 18:41 ` Linus Torvalds
2011-06-17 18:41 ` Linus Torvalds
2011-06-17 20:19 ` Tim Chen
2011-06-17 20:19 ` Tim Chen
2011-06-17 22:20 ` Hugh Dickins
2011-06-17 22:20 ` Hugh Dickins
2011-06-18 4:47 ` Linus Torvalds
2011-06-18 4:47 ` Linus Torvalds
2011-06-17 19:53 ` [PATCH] mm, memory-failure: Fix spinlock vs mutex order Peter Zijlstra
2011-06-17 19:53 ` Peter Zijlstra
2011-06-17 20:04 ` Andi Kleen
2011-06-17 20:04 ` Andi Kleen
2011-06-17 16:46 ` REGRESSION: Performance regressions from switching anon_vma->lock to mutex Linus Torvalds
2011-06-17 16:46 ` Linus Torvalds
2011-06-17 17:28 ` Linus Torvalds
2011-06-17 19:40 ` Andi Kleen
2011-06-17 19:40 ` Andi Kleen
2011-06-18 8:08 ` Ingo Molnar
2011-06-18 8:08 ` Ingo Molnar
2011-06-17 18:22 ` Tim Chen
2011-06-17 18:22 ` Tim Chen
2011-06-17 19:05 ` Ray Lee
2011-06-17 19:05 ` Ray Lee
2011-06-16 22:00 ` Andi Kleen
2011-06-16 22:00 ` Andi Kleen
2011-06-15 10:36 ` Peter Zijlstra
2011-06-15 10:36 ` Peter Zijlstra
2011-06-15 10:58 ` Peter Zijlstra
2011-06-15 10:58 ` Peter Zijlstra
2011-06-15 11:41 ` Peter Zijlstra
2011-06-15 11:41 ` Peter Zijlstra
2011-06-15 19:11 ` Linus Torvalds
2011-06-15 19:11 ` Linus Torvalds
2011-06-15 19:24 ` Andrew Morton
2011-06-15 19:24 ` Andrew Morton
2011-06-15 20:16 ` Ingo Molnar
2011-06-15 20:16 ` Ingo Molnar
2011-06-15 20:55 ` Linus Torvalds
2011-06-15 20:55 ` Linus Torvalds
2011-06-15 20:12 ` [GIT PULL] " Ingo Molnar
2011-06-15 20:12 ` Ingo Molnar
2011-06-15 20:29 ` Paul E. McKenney
2011-06-15 20:29 ` Paul E. McKenney
2011-06-15 20:47 ` Linus Torvalds
2011-06-15 20:47 ` Linus Torvalds
2011-06-15 20:54 ` Paul E. McKenney
2011-06-15 20:54 ` Paul E. McKenney
2011-06-15 21:05 ` Linus Torvalds
2011-06-15 21:05 ` Linus Torvalds
2011-06-15 21:15 ` Paul E. McKenney
2011-06-15 21:15 ` Paul E. McKenney
2011-06-15 21:27 ` Linus Torvalds
2011-06-15 21:27 ` Linus Torvalds
2011-06-16 7:03 ` Ingo Molnar
2011-06-16 7:03 ` Ingo Molnar
2011-06-16 17:16 ` Paul E. McKenney
2011-06-16 17:16 ` Paul E. McKenney
2011-06-16 20:25 ` Ingo Molnar
2011-06-16 20:25 ` Ingo Molnar
2011-06-16 21:01 ` Frederic Weisbecker
2011-06-16 21:01 ` Frederic Weisbecker
2011-06-16 23:02 ` Ingo Molnar
2011-06-16 23:02 ` Ingo Molnar
2011-06-17 15:19 ` Frederic Weisbecker
2011-06-17 15:19 ` Frederic Weisbecker
2011-06-16 21:02 ` Andi Kleen
2011-06-16 21:02 ` Andi Kleen
2011-06-16 22:21 ` Benjamin Herrenschmidt
2011-06-16 22:21 ` Benjamin Herrenschmidt
2011-06-16 22:38 ` Ingo Molnar
2011-06-16 22:38 ` Ingo Molnar
2011-06-16 22:47 ` Andi Kleen
2011-06-16 22:47 ` Andi Kleen
2011-06-16 22:58 ` Ingo Molnar
2011-06-16 22:58 ` Ingo Molnar
2011-06-17 0:45 ` Paul E. McKenney
2011-06-17 0:45 ` Paul E. McKenney
2011-06-17 9:43 ` Ingo Molnar
2011-06-17 9:43 ` Ingo Molnar
2011-06-17 16:48 ` Paul E. McKenney
2011-06-17 16:48 ` Paul E. McKenney
2011-06-16 23:37 ` Paul E. McKenney
2011-06-16 23:37 ` Paul E. McKenney
2011-06-15 20:13 ` Tim Chen
2011-06-15 20:13 ` Tim Chen
2011-06-15 20:17 ` Ingo Molnar
2011-06-15 20:17 ` Ingo Molnar
2011-06-15 20:21 ` Tim Chen
2011-06-15 20:21 ` Tim Chen
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=1308310080.2355.19.camel@twins \
--to=a.p.zijlstra@chello.nl \
--cc=ak@linux.intel.com \
--cc=akpm@linux-foundation.org \
--cc=alex.shi@intel.com \
--cc=benh@kernel.crashing.org \
--cc=davem@davemloft.net \
--cc=hughd@google.com \
--cc=jdike@addtoit.com \
--cc=kamezawa.hiroyu@jp.fujitsu.com \
--cc=kosaki.motohiro@jp.fujitsu.com \
--cc=lethal@linux-sh.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mel@csn.ul.ie \
--cc=namhyung@gmail.com \
--cc=npiggin@kernel.dk \
--cc=richard@nod.at \
--cc=rjw@sisk.pl \
--cc=rmk@arm.linux.org.uk \
--cc=schwidefsky@de.ibm.com \
--cc=shaohua.li@intel.com \
--cc=tim.c.chen@linux.intel.com \
--cc=tony.luck@intel.com \
--cc=torvalds@linux-foundation.org \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.