From: Andrew Morton <akpm@osdl.org>
To: Badari Pulavarty <pbadari@us.ibm.com>
Cc: linux-mm <linux-mm@kvack.org>, ext4 <linux-ext4@vger.kernel.org>,
lkml <linux-kernel@vger.kernel.org>
Subject: Re: pagefault in generic_file_buffered_write() causing deadlock
Date: Wed, 15 Nov 2006 11:29:57 -0800 [thread overview]
Message-ID: <20061115112957.e38539e9.akpm@osdl.org> (raw)
In-Reply-To: <455B5A7B.6010807@us.ibm.com>
On Wed, 15 Nov 2006 10:20:43 -0800
Badari Pulavarty <pbadari@us.ibm.com> wrote:
> Andrew Morton wrote:
> > On Wed, 15 Nov 2006 07:57:45 -0800
> > Badari Pulavarty <pbadari@us.ibm.com> wrote:
> >
> >
> >> We are looking at a customer situation (on 2.6.16-based distro) - where
> >> system becomes almost useless while running some java & stress tests.
> >>
> >> Root cause seems to be taking a pagefault in generic_file_buffered_write
> >> () after calling prepare_write. I am wondering
> >>
> >> 1) Why & How this can happen - since we made sure to fault the user
> >> buffer before prepare write.
> >>
> >
> > When using writev() we only fault in the first segment of the iovec. If
> > the second or succesive segment isn't mapped into pagetables we're
> > vulnerable to the deadlock.
> >
> >
> Hmm.. Not it :(
> Its coming from write() not writev().
>
> [C00000002ABBF290] [C00000000039D58C] .do_page_fault+0x2e4/0x75c
> [C00000002ABBF460] [C000000000004860] .handle_page_fault+0x20/0x54
> --- Exception: 301 at .__copy_tofrom_user+0x11c/0x580
> LR = .generic_file_buffered_write+0x39c/0x7c8
> [C00000002ABBF750] [C000000000095A94]
> .generic_file_buffered_write+0x2c0/0x7c8 (
> unreliable)
> [C00000002ABBF8F0] [C0000000000962EC]
> .__generic_file_aio_write_nolock+0x350/0x3
> e0
> [C00000002ABBFA20] [C000000000096908] .generic_file_aio_write+0x78/0x104
> [C00000002ABBFAE0] [C0000000001649F0] .ext3_file_write+0x2c/0xd4
> [C00000002ABBFB70] [C0000000000C5168] .do_sync_write+0xd4/0x130
> [C00000002ABBFCF0] [C0000000000C5ED4] .vfs_write+0x128/0x20c
> [C00000002ABBFD90] [C0000000000C664C] .sys_write+0x4c/0x8c
> [C00000002ABBFE30] [C00000000000871C] syscall_exit+0x0/0x40
>
Oh well. If it's a deadlock (this is not clear from your description) then
please gather backtraces of all affected tasks.
There is an ab/ba deadlock with journal_start() and lock_page(), iirc.
Chris and I had a look at that a while back and collapsed in exhaustion -
it isn't pretty.
next prev parent reply other threads:[~2006-11-15 19:30 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-11-15 15:57 pagefault in generic_file_buffered_write() causing deadlock Badari Pulavarty
2006-11-15 17:00 ` Andrew Morton
2006-11-15 18:16 ` Badari Pulavarty
2006-11-15 18:20 ` Badari Pulavarty
2006-11-15 19:29 ` Andrew Morton [this message]
2006-11-15 20:39 ` Chris Mason
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=20061115112957.e38539e9.akpm@osdl.org \
--to=akpm@osdl.org \
--cc=linux-ext4@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=pbadari@us.ibm.com \
/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).