From: Mel Gorman <mgorman@suse.de>
To: Richard Davies <richard@arachsys.com>
Cc: kvm@vger.kernel.org, qemu-devel@nongnu.org, linux-mm@kvack.org,
Avi Kivity <avi@redhat.com>, Shaohua Li <shli@kernel.org>
Subject: Re: [Qemu-devel] Windows VM slow boot
Date: Wed, 12 Sep 2012 13:25:41 +0100 [thread overview]
Message-ID: <20120912122541.GO11266@suse.de> (raw)
In-Reply-To: <20120912105659.GA23818@alpha.arachsys.com>
On Wed, Sep 12, 2012 at 11:56:59AM +0100, Richard Davies wrote:
> [ adding linux-mm - previously at http://marc.info/?t=134511509400003 ]
>
> Hi Rik,
>
I'm not Rik but hi anyway.
> Since qemu-kvm 1.2.0 and Linux 3.6.0-rc5 came out, I thought that I would
> retest with these.
>
Ok. 3.6.0-rc5 contains [c67fe375: mm: compaction: Abort async compaction
if locks are contended or taking too long] that should have helped mitigate
some of the lock contention problem but not all of it as we'll see later.
> The typical symptom now appears to be that the Windows VMs boot reasonably
> fast,
I see that this is an old-ish bug but I did not read the full history.
Is it now booting faster than 3.5.0 was? I'm asking because I'm
interested to see if commit c67fe375 helped your particular case.
> but then there is high CPU use and load for many minutes afterwards -
> the high CPU use is both for the qemu-kvm processes themselves and also for
> % sys.
>
Ok, I cannot comment on the userspace portion of things but the kernel
portion still indicates that there is a high percentage of time on what
appears to be lock contention.
> I attach a perf report which seems to show that the high CPU use is in the
> memory manager.
>
A follow-on from commit c67fe375 was the following patch (author cc'd)
which addresses lock contention in isolate_migratepages_range where your
perf report indicates that we're spending 95% of the time. Would you be
willing to test it please?
---8<---
From: Shaohua Li <shli@kernel.org>
Subject: mm: compaction: check lock contention first before taking lock
isolate_migratepages_range will take zone->lru_lock first and check if the
lock is contented, if yes, it will release the lock. This isn't
efficient. If the lock is truly contented, a lock/unlock pair will
increase the lock contention. We'd better check if the lock is contended
first. compact_trylock_irqsave perfectly meets the requirement.
Signed-off-by: Shaohua Li <shli@fusionio.com>
Acked-by: Mel Gorman <mgorman@suse.de>
Acked-by: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
mm/compaction.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff -puN mm/compaction.c~mm-compaction-check-lock-contention-first-before-taking-lock mm/compaction.c
--- a/mm/compaction.c~mm-compaction-check-lock-contention-first-before-taking-lock
+++ a/mm/compaction.c
@@ -349,8 +349,9 @@ isolate_migratepages_range(struct zone *
/* Time to isolate some pages for migration */
cond_resched();
- spin_lock_irqsave(&zone->lru_lock, flags);
- locked = true;
+ locked = compact_trylock_irqsave(&zone->lru_lock, &flags, cc);
+ if (!locked)
+ return 0;
for (; low_pfn < end_pfn; low_pfn++) {
struct page *page;
next prev parent reply other threads:[~2012-09-12 12:25 UTC|newest]
Thread overview: 46+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-08-16 10:47 [Qemu-devel] Windows slow boot: contractor wanted Richard Davies
2012-08-16 11:39 ` Avi Kivity
2012-08-17 12:36 ` Richard Davies
2012-08-17 13:02 ` Robert Vineyard
2012-08-18 14:44 ` Richard Davies
2012-08-19 5:02 ` Brian Jackson
2012-08-20 8:16 ` Richard Davies
2012-08-19 8:40 ` Avi Kivity
2012-08-19 8:51 ` Richard Davies
2012-08-19 14:04 ` Avi Kivity
2012-08-20 13:56 ` Richard Davies
2012-08-21 9:00 ` Avi Kivity
2012-08-21 15:21 ` Richard Davies
2012-08-21 15:39 ` Troy Benjegerdes
2012-08-22 9:08 ` Avi Kivity
2012-08-22 12:40 ` Richard Davies
2012-08-22 12:44 ` Avi Kivity
2012-08-22 14:41 ` Richard Davies
2012-08-22 14:53 ` Avi Kivity
2012-08-22 15:26 ` Richard Davies
2012-08-22 17:22 ` Troy Benjegerdes
2012-08-25 17:51 ` Richard Davies
2012-08-22 15:21 ` Rik van Riel
2012-08-22 15:34 ` Richard Davies
2012-08-25 17:45 ` Richard Davies
2012-08-25 18:11 ` Rik van Riel
2012-08-26 10:58 ` Richard Davies
2012-09-06 9:20 ` Richard Davies
2012-09-12 10:56 ` [Qemu-devel] Windows VM slow boot Richard Davies
2012-09-12 12:25 ` Mel Gorman [this message]
2012-09-12 16:46 ` Richard Davies
2012-09-13 9:50 ` Mel Gorman
2012-09-13 19:47 ` [Qemu-devel] [PATCH 1/2] Revert "mm: have order > 0 compaction start near a pageblock with free pages" Rik van Riel
2012-09-13 19:48 ` [Qemu-devel] [PATCH 2/2] make the compaction "skip ahead" logic robust Rik van Riel
2012-09-13 19:54 ` [Qemu-devel] [PATCH -v2 " Rik van Riel
2012-09-15 15:55 ` Richard Davies
2012-09-16 19:12 ` Richard Davies
2012-09-17 12:26 ` Mel Gorman
2012-09-18 8:14 ` Richard Davies
2012-09-18 11:21 ` Mel Gorman
2012-09-18 17:58 ` Richard Davies
2012-09-17 13:50 ` Rik van Riel
2012-09-17 14:07 ` Mel Gorman
2012-08-16 14:10 ` [Qemu-devel] Windows slow boot: contractor wanted Benoît Canet
2012-08-16 15:53 ` Troy Benjegerdes
2012-09-18 15:12 ` [Qemu-devel] Windows slow boot Michael Tokarev
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=20120912122541.GO11266@suse.de \
--to=mgorman@suse.de \
--cc=avi@redhat.com \
--cc=kvm@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=qemu-devel@nongnu.org \
--cc=richard@arachsys.com \
--cc=shli@kernel.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 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).