Re: [PATCH v3] xfs: load uncached unlinked inodes into memory on demand

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Fri, Sep 01, 2023 at 08:03:11AM -0700, Darrick J. Wong wrote:
> From: Darrick J. Wong <djwong@xxxxxxxxxx>
> 
> shrikanth hegde reports that filesystems fail shortly after mount with
> the following failure:
> 
> 	WARNING: CPU: 56 PID: 12450 at fs/xfs/xfs_inode.c:1839 xfs_iunlink_lookup+0x58/0x80 [xfs]
> 
> This of course is the WARN_ON_ONCE in xfs_iunlink_lookup:
> 
> 	ip = radix_tree_lookup(&pag->pag_ici_root, agino);
> 	if (WARN_ON_ONCE(!ip || !ip->i_ino)) { ... }
> 
> From diagnostic data collected by the bug reporters, it would appear
> that we cleanly mounted a filesystem that contained unlinked inodes.
> Unlinked inodes are only processed as a final step of log recovery,
> which means that clean mounts do not process the unlinked list at all.
> 
> Prior to the introduction of the incore unlinked lists, this wasn't a
> problem because the unlink code would (very expensively) traverse the
> entire ondisk metadata iunlink chain to keep things up to date.
> However, the incore unlinked list code complains when it realizes that
> it is out of sync with the ondisk metadata and shuts down the fs, which
> is bad.
> 
> Ritesh proposed to solve this problem by unconditionally parsing the
> unlinked lists at mount time, but this imposes a mount time cost for
> every filesystem to catch something that should be very infrequent.
> Instead, let's target the places where we can encounter a next_unlinked
> pointer that refers to an inode that is not in cache, and load it into
> cache.
> 
> Note: This patch does not address the problem of iget loading an inode
> from the middle of the iunlink list and needing to set i_prev_unlinked
> correctly.
> 
> Eric Sandeen adds:
> 
> "One way to end up in this situation is to have at one point run a very
> old kernel which did not contain this commit, merged in kernel v4.14:
> 
> commit 6f4a1eefdd0ad4561543270a7fceadabcca075dd
> Author: Eric Sandeen <sandeen@xxxxxxxxxxx>
> Date:   Tue Aug 8 18:21:49 2017 -0700
> 
>     xfs: toggle readonly state around xfs_log_mount_finish
> 
>     When we do log recovery on a readonly mount, unlinked inode
>     processing does not happen due to the readonly checks in
>     xfs_inactive(), which are trying to prevent any I/O on a
>     readonly mount.
> 
>     This is misguided - we do I/O on readonly mounts all the time,
>     for consistency; for example, log recovery.  So do the same
>     RDONLY flag twiddling around xfs_log_mount_finish() as we
>     do around xfs_log_mount(), for the same reason.
> 
>     This all cries out for a big rework but for now this is a
>     simple fix to an obvious problem.
> 
>     Signed-off-by: Eric Sandeen <sandeen@xxxxxxxxxx>
>     Reviewed-by: Brian Foster <bfoster@xxxxxxxxxx>
>     Reviewed-by: Christoph Hellwig <hch@xxxxxx>
>     Reviewed-by: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
>     Signed-off-by: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
> 
> "so if you:
> 
> 1) Crash with unlinked inodes
> 2) mount -o ro <recovers log but skips unlinked inode recovery>
> 3) mount -o remount,rw
> 4) umount <writes clean log record>
> 
> "You now have a filesystem with on-disk unlinked inodes and a clean log,
> and those inodes won't get cleaned up until log recovery runs again or
> xfs_repair is run.
> 
> "And in testing an old OS (RHEL7) it does seem that the root filesystem
> goes through a mount -o ro, mount -o remount,rw transition at boot time.
> So this situation may be somewhat common."
> 
> Reported-by: shrikanth hegde <sshegde@xxxxxxxxxxxxxxxxxx>
> Triaged-by: Ritesh Harjani <ritesh.list@xxxxxxxxx>
> Signed-off-by: Darrick J. Wong <djwong@xxxxxxxxxx>
> Reviewed-by: Dave Chinner <dchinner@xxxxxxxxxx>
> Reviewed-by: Bill O'Donnell <bodonnel@xxxxxxxxxx>
> ---
> v3: add RVB tags and historical context from sandeen
> v2: log that we're doing runtime recovery, dont mess with DONTCACHE,
>     and actually return ENOLINK
> ---
>  fs/xfs/xfs_inode.c |   75 +++++++++++++++++++++++++++++++++++++++++++++++++---
>  fs/xfs/xfs_trace.h |   25 +++++++++++++++++
>  2 files changed, 96 insertions(+), 4 deletions(-)
> 
> diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
> index 6ee266be45d4..2942002560b5 100644
> --- a/fs/xfs/xfs_inode.c
> +++ b/fs/xfs/xfs_inode.c
> @@ -1829,12 +1829,17 @@ xfs_iunlink_lookup(

<sigh> Somehow between the four copies of this patch that I have flying
around (djwong-dev, patchmail, linus TOT, old 6.6 merge branch) I once
again lost the comment change for this function, and apparently never
actually sent that to the list.

"A poor workman blames his tools", etc.

I bet that workman doesn't have to do this much manual paperwork
either...

--D

>  
>  	rcu_read_lock();
>  	ip = radix_tree_lookup(&pag->pag_ici_root, agino);
> +	if (!ip) {
> +		/* Caller can handle inode not being in memory. */
> +		rcu_read_unlock();
> +		return NULL;
> +	}
>  
>  	/*
> -	 * Inode not in memory or in RCU freeing limbo should not happen.
> -	 * Warn about this and let the caller handle the failure.
> +	 * Inode in RCU freeing limbo should not happen.  Warn about this and
> +	 * let the caller handle the failure.
>  	 */
> -	if (WARN_ON_ONCE(!ip || !ip->i_ino)) {
> +	if (WARN_ON_ONCE(!ip->i_ino)) {
>  		rcu_read_unlock();
>  		return NULL;
>  	}
> @@ -1858,7 +1863,8 @@ xfs_iunlink_update_backref(
>  
>  	ip = xfs_iunlink_lookup(pag, next_agino);
>  	if (!ip)
> -		return -EFSCORRUPTED;
> +		return -ENOLINK;
> +
>  	ip->i_prev_unlinked = prev_agino;
>  	return 0;
>  }
> @@ -1902,6 +1908,62 @@ xfs_iunlink_update_bucket(
>  	return 0;
>  }
>  
> +/*
> + * Load the inode @next_agino into the cache and set its prev_unlinked pointer
> + * to @prev_agino.  Caller must hold the AGI to synchronize with other changes
> + * to the unlinked list.
> + */
> +STATIC int
> +xfs_iunlink_reload_next(
> +	struct xfs_trans	*tp,
> +	struct xfs_buf		*agibp,
> +	xfs_agino_t		prev_agino,
> +	xfs_agino_t		next_agino)
> +{
> +	struct xfs_perag	*pag = agibp->b_pag;
> +	struct xfs_mount	*mp = pag->pag_mount;
> +	struct xfs_inode	*next_ip = NULL;
> +	xfs_ino_t		ino;
> +	int			error;
> +
> +	ASSERT(next_agino != NULLAGINO);
> +
> +#ifdef DEBUG
> +	rcu_read_lock();
> +	next_ip = radix_tree_lookup(&pag->pag_ici_root, next_agino);
> +	ASSERT(next_ip == NULL);
> +	rcu_read_unlock();
> +#endif
> +
> +	xfs_info_ratelimited(mp,
> + "Found unrecovered unlinked inode 0x%x in AG 0x%x.  Initiating recovery.",
> +			next_agino, pag->pag_agno);
> +
> +	/*
> +	 * Use an untrusted lookup just to be cautious in case the AGI has been
> +	 * corrupted and now points at a free inode.  That shouldn't happen,
> +	 * but we'd rather shut down now since we're already running in a weird
> +	 * situation.
> +	 */
> +	ino = XFS_AGINO_TO_INO(mp, pag->pag_agno, next_agino);
> +	error = xfs_iget(mp, tp, ino, XFS_IGET_UNTRUSTED, 0, &next_ip);
> +	if (error)
> +		return error;
> +
> +	/* If this is not an unlinked inode, something is very wrong. */
> +	if (VFS_I(next_ip)->i_nlink != 0) {
> +		error = -EFSCORRUPTED;
> +		goto rele;
> +	}
> +
> +	next_ip->i_prev_unlinked = prev_agino;
> +	trace_xfs_iunlink_reload_next(next_ip);
> +rele:
> +	ASSERT(!(VFS_I(next_ip)->i_state & I_DONTCACHE));
> +	xfs_irele(next_ip);
> +	return error;
> +}
> +
>  static int
>  xfs_iunlink_insert_inode(
>  	struct xfs_trans	*tp,
> @@ -1933,6 +1995,8 @@ xfs_iunlink_insert_inode(
>  	 * inode.
>  	 */
>  	error = xfs_iunlink_update_backref(pag, agino, next_agino);
> +	if (error == -ENOLINK)
> +		error = xfs_iunlink_reload_next(tp, agibp, agino, next_agino);
>  	if (error)
>  		return error;
>  
> @@ -2027,6 +2091,9 @@ xfs_iunlink_remove_inode(
>  	 */
>  	error = xfs_iunlink_update_backref(pag, ip->i_prev_unlinked,
>  			ip->i_next_unlinked);
> +	if (error == -ENOLINK)
> +		error = xfs_iunlink_reload_next(tp, agibp, ip->i_prev_unlinked,
> +				ip->i_next_unlinked);
>  	if (error)
>  		return error;
>  
> diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
> index 36bd42ed9ec8..f4e46bac9b91 100644
> --- a/fs/xfs/xfs_trace.h
> +++ b/fs/xfs/xfs_trace.h
> @@ -3832,6 +3832,31 @@ TRACE_EVENT(xfs_iunlink_update_dinode,
>  		  __entry->new_ptr)
>  );
>  
> +TRACE_EVENT(xfs_iunlink_reload_next,
> +	TP_PROTO(struct xfs_inode *ip),
> +	TP_ARGS(ip),
> +	TP_STRUCT__entry(
> +		__field(dev_t, dev)
> +		__field(xfs_agnumber_t, agno)
> +		__field(xfs_agino_t, agino)
> +		__field(xfs_agino_t, prev_agino)
> +		__field(xfs_agino_t, next_agino)
> +	),
> +	TP_fast_assign(
> +		__entry->dev = ip->i_mount->m_super->s_dev;
> +		__entry->agno = XFS_INO_TO_AGNO(ip->i_mount, ip->i_ino);
> +		__entry->agino = XFS_INO_TO_AGINO(ip->i_mount, ip->i_ino);
> +		__entry->prev_agino = ip->i_prev_unlinked;
> +		__entry->next_agino = ip->i_next_unlinked;
> +	),
> +	TP_printk("dev %d:%d agno 0x%x agino 0x%x prev_unlinked 0x%x next_unlinked 0x%x",
> +		  MAJOR(__entry->dev), MINOR(__entry->dev),
> +		  __entry->agno,
> +		  __entry->agino,
> +		  __entry->prev_agino,
> +		  __entry->next_agino)
> +);
> +
>  DECLARE_EVENT_CLASS(xfs_ag_inode_class,
>  	TP_PROTO(struct xfs_inode *ip),
>  	TP_ARGS(ip),



[Index of Archives]     [XFS Filesystem Development (older mail)]     [Linux Filesystem Development]     [Linux Audio Users]     [Yosemite Trails]     [Linux Kernel]     [Linux RAID]     [Linux SCSI]


  Powered by Linux