XFS corruption in 3.14.72

[jorgus]

Hi

Is anyone experiencing the same problem? Since I upgraded to 3.14.72 + grsecurity-3.1-3.14.72-201606081603.patch I started observing XFS corruption problems on my servers:

XFS (dm-4): xfs_do_force_shutdown(0x8) called from line 3136 of file fs/xfs/xfs_inode.c. Return address = 0xffffffffa0364fef
XFS (dm-4): Corruption of in-memory data detected. Shutting down filesystem
XFS (dm-4): Please umount the filesystem and rectify the problem(s)
XFS (dm-4): xfs_log_force: error 5 returned.
XFS (dm-4): xfs_log_force: error 5 returned.
XFS (dm-4): xfs_log_force: error 5 returned.
....

I'm not sure whether this is a bug in the vanilla kernel (I can't find anything about it in the internet but there have been XFS changes there) or maybe grsecurity patch contributed to that.

[PaX Team]

our changes to xfs are minimal and have been there for a long time so it's more likely that this has something to do with the stable changes that went into 3.14.72. can you try a vanilla kernel (assuming it's easily reproducible)?

[jorgus]

It is not that easily reproducible as you just need to wait till it happens. I think I will give the upstream kernel maintainers a few days and maybe they will realise they broke something. I'm not that keen on running a vanilla kernel on my production machines for extensive period of time for obvious reasons.

[spender]

I see the problem, it's an upstream bug, so congrats: you are apparently the only person in the world using XFS on 3.14.72

-Brad