From: "Darrick J. Wong" <djwong@kernel.org>
To: Dave Chinner <david@fromorbit.com>
Cc: linux-xfs@vger.kernel.org
Subject: Re: [PATCH 5/6] xfs: xfs_do_force_shutdown needs to block racing shutdowns
Date: Mon, 28 Mar 2022 17:19:17 -0700 [thread overview]
Message-ID: <20220329001917.GD27713@magnolia> (raw)
In-Reply-To: <20220324002103.710477-6-david@fromorbit.com>
On Thu, Mar 24, 2022 at 11:21:02AM +1100, Dave Chinner wrote:
> From: Dave Chinner <dchinner@redhat.com>
>
> When we call xfs_forced_shutdown(), the caller often expects the
> filesystem to be completely shut down when it returns. However,
> if we have racing xfs_forced_shutdown() calls, the first caller sets
> the mount shutdown flag then goes to shutdown the log. The second
> caller sees the mount shutdown flag and returns immediately - it
> does not wait for the log to be shut down.
>
> Unfortunately, xfs_forced_shutdown() is used in some places that
> expect it to completely shut down the filesystem before it returns
> (e.g. xfs_trans_log_inode()). As such, returning before the log has
> been shut down leaves us in a place where the transaction failed to
> complete correctly but we still call xfs_trans_commit(). This
> situation arises because xfs_trans_log_inode() does not return an
> error and instead calls xfs_force_shutdown() to ensure that the
> transaction being committed is aborted.
>
> Unfortunately, we have a race condition where xfs_trans_commit()
> needs to check xlog_is_shutdown() because it can't abort log items
> before the log is shut down, but it needs to use xfs_is_shutdown()
> because xfs_forced_shutdown() does not block waiting for the log to
> shut down.
>
> To fix this conundrum, first we make all calls to
> xfs_forced_shutdown() block until the log is also shut down. This
> means we can then safely use xfs_forced_shutdown() as a mechanism
> that ensures the currently running transaction will be aborted by
> xfs_trans_commit() regardless of the shutdown check it uses.
>
> Signed-off-by: Dave Chinner <dchinner@redhat.com>
I always feel dirty looking at what __var_waitqueue does, but this does
make sense that everything must be shut down by the time
xfs_force_shutdown returns.
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
--D
> ---
> fs/xfs/xfs_fsops.c | 6 +++++-
> fs/xfs/xfs_log.c | 1 +
> fs/xfs/xfs_log_priv.h | 11 +++++++++++
> 3 files changed, 17 insertions(+), 1 deletion(-)
>
> diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c
> index 33e26690a8c4..093a44aecec0 100644
> --- a/fs/xfs/xfs_fsops.c
> +++ b/fs/xfs/xfs_fsops.c
> @@ -17,6 +17,7 @@
> #include "xfs_fsops.h"
> #include "xfs_trans_space.h"
> #include "xfs_log.h"
> +#include "xfs_log_priv.h"
> #include "xfs_ag.h"
> #include "xfs_ag_resv.h"
> #include "xfs_trace.h"
> @@ -528,8 +529,11 @@ xfs_do_force_shutdown(
> int tag;
> const char *why;
>
> - if (test_and_set_bit(XFS_OPSTATE_SHUTDOWN, &mp->m_opstate))
> +
> + if (test_and_set_bit(XFS_OPSTATE_SHUTDOWN, &mp->m_opstate)) {
> + xlog_shutdown_wait(mp->m_log);
> return;
> + }
> if (mp->m_sb_bp)
> mp->m_sb_bp->b_flags |= XBF_DONE;
>
> diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c
> index 6879d6d19d68..4336eb804a4a 100644
> --- a/fs/xfs/xfs_log.c
> +++ b/fs/xfs/xfs_log.c
> @@ -3909,6 +3909,7 @@ xlog_force_shutdown(
> xlog_state_shutdown_callbacks(log);
> spin_unlock(&log->l_icloglock);
>
> + wake_up_var(&log->l_opstate);
> return log_error;
> }
>
> diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h
> index 80d77aac6fe4..401cdc400980 100644
> --- a/fs/xfs/xfs_log_priv.h
> +++ b/fs/xfs/xfs_log_priv.h
> @@ -484,6 +484,17 @@ xlog_is_shutdown(struct xlog *log)
> return test_bit(XLOG_IO_ERROR, &log->l_opstate);
> }
>
> +/*
> + * Wait until the xlog_force_shutdown() has marked the log as shut down
> + * so xlog_is_shutdown() will always return true.
> + */
> +static inline void
> +xlog_shutdown_wait(
> + struct xlog *log)
> +{
> + wait_var_event(&log->l_opstate, xlog_is_shutdown(log));
> +}
> +
> /* common routines */
> extern int
> xlog_recover(
> --
> 2.35.1
>
next prev parent reply other threads:[~2022-03-29 0:19 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-03-24 0:20 [PATCH 0/6 v2] xfs: more shutdown/recovery fixes Dave Chinner
2022-03-24 0:20 ` [PATCH 1/6] xfs: aborting inodes on shutdown may need buffer lock Dave Chinner
2022-03-28 22:44 ` Darrick J. Wong
2022-03-28 23:11 ` Dave Chinner
2022-03-30 1:20 ` Darrick J. Wong
2022-03-24 0:20 ` [PATCH 2/6] xfs: shutdown in intent recovery has non-intent items in the AIL Dave Chinner
2022-03-28 22:46 ` Darrick J. Wong
2022-03-24 0:21 ` [PATCH 3/6] xfs: run callbacks before waking waiters in xlog_state_shutdown_callbacks Dave Chinner
2022-03-28 23:05 ` Darrick J. Wong
2022-03-28 23:13 ` Dave Chinner
2022-03-28 23:36 ` Darrick J. Wong
2022-03-24 0:21 ` [PATCH 4/6] xfs: log shutdown triggers should only shut down the log Dave Chinner
2022-03-29 0:14 ` Darrick J. Wong
2022-03-24 0:21 ` [PATCH 5/6] xfs: xfs_do_force_shutdown needs to block racing shutdowns Dave Chinner
2022-03-29 0:19 ` Darrick J. Wong [this message]
2022-03-24 0:21 ` [PATCH 6/6] xfs: xfs_trans_commit() path must check for log shutdown Dave Chinner
2022-03-29 0:36 ` Darrick J. Wong
2022-03-29 3:08 ` Dave Chinner
2022-03-27 22:55 ` [PATCH 7/6] xfs: xfs: shutdown during log recovery needs to mark the " Dave Chinner
2022-03-29 0:37 ` Darrick J. Wong
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=20220329001917.GD27713@magnolia \
--to=djwong@kernel.org \
--cc=david@fromorbit.com \
--cc=linux-xfs@vger.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).