Re: Hang in XFS reclaim on 3.7.0-rc3

[Torsten Kaiser]

On Mon, Oct 29, 2012 at 11:26 PM, Dave Chinner <david@xxxxxxxxxxxxx> wrote:
> On Mon, Oct 29, 2012 at 09:03:15PM +0100, Torsten Kaiser wrote:
>> After experiencing a hang of all IO yesterday (
>> http://marc.info/?l=linux-kernel&m=135142236520624&w=2 ), I turned on
>> LOCKDEP after upgrading to -rc3.
>>
>> I then tried to replicate the load that hung yesterday and got the
>> following lockdep report, implicating XFS instead of by stacking swap
>> onto dm-crypt and md.
>>
>> [ 2844.971913]
>> [ 2844.971920] =================================
>> [ 2844.971921] [ INFO: inconsistent lock state ]
>> [ 2844.971924] 3.7.0-rc3 #1 Not tainted
>> [ 2844.971925] ---------------------------------
>> [ 2844.971927] inconsistent {RECLAIM_FS-ON-W} -> {IN-RECLAIM_FS-W} usage.
>> [ 2844.971929] kswapd0/725 [HC0[0]:SC0[0]:HE1:SE1] takes:
>> [ 2844.971931] (&(&ip->i_lock)->mr_lock){++++?.}, at: [<ffffffff811e7ef4>] xfs_ilock+0x84/0xb0
>> [ 2844.971941] {RECLAIM_FS-ON-W} state was registered at:
>> [ 2844.971942]   [<ffffffff8108137e>] mark_held_locks+0x7e/0x130
>> [ 2844.971947]   [<ffffffff81081a63>] lockdep_trace_alloc+0x63/0xc0
>> [ 2844.971949]   [<ffffffff810e9dd5>] kmem_cache_alloc+0x35/0xe0
>> [ 2844.971952]   [<ffffffff810dba31>] vm_map_ram+0x271/0x770
>> [ 2844.971955]   [<ffffffff811e10a6>] _xfs_buf_map_pages+0x46/0xe0
>> [ 2844.971959]   [<ffffffff811e1fba>] xfs_buf_get_map+0x8a/0x130
>> [ 2844.971961]   [<ffffffff81233849>] xfs_trans_get_buf_map+0xa9/0xd0
>> [ 2844.971964]   [<ffffffff8121e339>] xfs_ifree_cluster+0x129/0x670
>> [ 2844.971967]   [<ffffffff8121f959>] xfs_ifree+0xe9/0xf0
>> [ 2844.971969]   [<ffffffff811f4abf>] xfs_inactive+0x2af/0x480
>> [ 2844.971972]   [<ffffffff811efb90>] xfs_fs_evict_inode+0x70/0x80
>> [ 2844.971974]   [<ffffffff8110cb8f>] evict+0xaf/0x1b0
>> [ 2844.971977]   [<ffffffff8110cd95>] iput+0x105/0x210
>> [ 2844.971979]   [<ffffffff811070d0>] dentry_iput+0xa0/0xe0
>> [ 2844.971981]   [<ffffffff81108310>] dput+0x150/0x280
>> [ 2844.971983]   [<ffffffff811020fb>] sys_renameat+0x21b/0x290
>> [ 2844.971986]   [<ffffffff81102186>] sys_rename+0x16/0x20
>> [ 2844.971988]   [<ffffffff816b2292>] system_call_fastpath+0x16/0x1b
>
> We shouldn't be mapping pages there. See if the patch below fixes
> it.

Applying your fix and rerunning my test workload did not trigger this
or any other LOCKDEP reports.
While I'm not 100% sure about my test case always hitting this, your
description makes me quite confident, that it really fixed this issue.

I will keep LOCKDEP enabled on that system, and if there really is
another splat, I will report back here. But I rather doubt that this
will be needed.

Thanks for the very quick fix!

Torsten

> Fundamentally, though, the lockdep warning has come about because
> vm_map_ram is doing a GFP_KERNEL allocation when we need it to be
> doing GFP_NOFS - we are within a transaction here, so memory reclaim
> is not allowed to recurse back into the filesystem.
>
> mm-folk: can we please get this vmalloc/gfp_flags passing API
> fixed once and for all? This is the fourth time in the last month or
> so that I've seen XFS bug reports with silent hangs and associated
> lockdep output that implicate GFP_KERNEL allocations from vm_map_ram
> in GFP_NOFS conditions as the potential cause....
>
> Cheers,
>
> Dave.
> --
> Dave Chinner
> david@xxxxxxxxxxxxx
>
> xfs: don't vmap inode cluster buffers during free
>
> From: Dave Chinner <dchinner@xxxxxxxxxx>
>
> Signed-off-by: Dave Chinner <dchinner@xxxxxxxxxx>
> ---
>  fs/xfs/xfs_inode.c |    3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
> index c4add46..82f6e5d 100644
> --- a/fs/xfs/xfs_inode.c
> +++ b/fs/xfs/xfs_inode.c
> @@ -1781,7 +1781,8 @@ xfs_ifree_cluster(
>                  * to mark all the active inodes on the buffer stale.
>                  */
>                 bp = xfs_trans_get_buf(tp, mp->m_ddev_targp, blkno,
> -                                       mp->m_bsize * blks_per_cluster, 0);
> +                                       mp->m_bsize * blks_per_cluster,
> +                                       XBF_UNMAPPED);
>
>                 if (!bp)
>                         return ENOMEM;
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/