| commit | 6bfb3d065f4c498c17a3a07f3dc08cedff53aff4 | |
| author | David Chinner <david@fromorbit.com> | |
| Thu, 30 Oct 2008 07:32:43 +0000 (30 18:32 +1100) | ||
| committer | Lachlan McIlroy <lachlan@redback.melbourne.sgi.com> | |
| Thu, 30 Oct 2008 07:32:43 +0000 (30 18:32 +1100) | ||
| tree | c5c528c77e44584616a175e0dcc89713e7b76d0a | treesnapshot (tar.gz zip) |
| parent | e0b8e8b65d578f5d5538465dff8392cf02e1cc5d | commitdiff |
[XFS] Fix race when looking up reclaimable inodes
If we get a race looking up a reclaimable inode, we can end up with the
winner proceeding to use the inode before it has been completely
re-initialised. This is a Bad Thing.
Fix the race by checking whether we are still initialising the inod eonce
we have a reference to it, and if so wait for the initialisation to
complete before continuing.
While there, fix a leaked reference count in the same code when
encountering an unlinked inode and we are not doing a lookup for a create
operation.
SGI-PV: 987246
SGI-Modid: xfs-linux-melb:xfs-kern:32429a
Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
If we get a race looking up a reclaimable inode, we can end up with the
winner proceeding to use the inode before it has been completely
re-initialised. This is a Bad Thing.
Fix the race by checking whether we are still initialising the inod eonce
we have a reference to it, and if so wait for the initialisation to
complete before continuing.
While there, fix a leaked reference count in the same code when
encountering an unlinked inode and we are not doing a lookup for a create
operation.
SGI-PV: 987246
SGI-Modid: xfs-linux-melb:xfs-kern:32429a
Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
| fs/xfs/linux-2.6/xfs_linux.h | diffblobblamehistory | |
| fs/xfs/xfs_iget.c | diffblobblamehistory |