From: "Jun'ichi Nomura" <j-nomura@ce.jp.nec.com>
To: device-mapper development <dm-devel@redhat.com>,
Mike Snitzer <snitzer@redhat.com>
Subject: Re: [PATCH] dm: use revalidate_disk to update device size after set_capacity
Date: Wed, 20 Oct 2010 14:42:19 +0900 [thread overview]
Message-ID: <4CBE813B.6050105@ce.jp.nec.com> (raw)
In-Reply-To: <20101019220711.GA25169@redhat.com>
Hi Mike,
(10/20/10 07:07), Mike Snitzer wrote:
> Avoid taking md->bdev->bd_inode->i_mutex to update the DM block device's
> size. Doing so eliminates the potential for deadlock if an fsync is
> racing with a device resize.
>
> Signed-off-by: Mike Snitzer <snitzer@redhat.com>
> ---
> drivers/md/dm.c | 5 +----
> 1 files changed, 1 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/md/dm.c b/drivers/md/dm.c
> index f934e98..fd315a7 100644
> --- a/drivers/md/dm.c
> +++ b/drivers/md/dm.c
> @@ -1995,10 +1995,7 @@ static void event_callback(void *context)
> static void __set_size(struct mapped_device *md, sector_t size)
> {
> set_capacity(md->disk, size);
> -
> - mutex_lock(&md->bdev->bd_inode->i_mutex);
> - i_size_write(md->bdev->bd_inode, (loff_t)size << SECTOR_SHIFT);
> - mutex_unlock(&md->bdev->bd_inode->i_mutex);
> + revalidate_disk(md->disk);
> }
Some concerns/questions:
- revalidate_disk() calls bdget() inside it.
Does bdget() no longer block on I/O?
Sometime ago, bdget() has been blocked by I_LOCK,
where process holding I_LOCK blocked by resize.
Since I_LOCK has disappeared, I suspect this might not
be a valid concern anymore.
FYI, past discussion on this topic:
http://www.redhat.com/archives/dm-devel/2007-October/msg00134.html
(it's not my intention to push the patch in the above URL)
- revalidate_disk() ends up with get_super():
revalidate_disk
check_disk_size_change
flush_disk
__invalidate_device
get_super
and get_super() takes sb->s_umount.
OTOH, there are codes which wait on I/O holding s_umount
exclusively. E.g. freeze_super() calls sync_filesystem().
So it seems there is a possible deadlock if you use
revalidate_disk from dm_swap_table.
I've been away from that part of the code for a while.
So it's nice if the above is just a false alarm...
Thanks,
--
Jun'ichi Nomura, NEC Corporation
next prev parent reply other threads:[~2010-10-20 5:42 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-10-19 22:07 [PATCH] dm: use revalidate_disk to update device size after set_capacity Mike Snitzer
2010-10-20 5:42 ` Jun'ichi Nomura [this message]
2010-10-28 1:16 ` Mike Snitzer
2010-10-28 12:15 ` Jun'ichi Nomura
2010-10-28 19:54 ` Mike Snitzer
2010-10-28 22:18 ` Mike Snitzer
2010-10-29 3:00 ` Jun'ichi Nomura
2010-10-29 21:50 ` dm: lock bd_mutex when setting device size Mike Snitzer
2010-11-01 7:19 ` Jun'ichi Nomura
2010-11-01 13:14 ` Mike Snitzer
2010-11-03 18:06 ` [PATCH] dm: remove extra locking when changing " Mike Snitzer
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=4CBE813B.6050105@ce.jp.nec.com \
--to=j-nomura@ce.jp.nec.com \
--cc=dm-devel@redhat.com \
--cc=snitzer@redhat.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 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.