On Thu, Jan 06, 2011 at 12:10:56PM +1100, Stephen Rothwell wrote:
> Hi Nick,
>
> Today's linux-next merge of the vfs-scale tree got a conflict in
> fs/xfs/xfs_iget.c between commits
> d95b7aaf9ab6738bef1ebcc52ab66563085e44ac ("xfs: rcu free inodes") and
> 1a3e8f3da09c7082d25b512a0ffe569391e4c09a ("xfs: convert inode cache
> lookups to use RCU locking") from the xfs tree and commit
> bb3e8c37a0af21d0a8fe54a0b0f17aca16335a82 ("fs: icache RCU free inodes")
> from the vfs-scale tree.
>
> OK, so looking at this, the first xfs tree patch above does the same as
> the vfs-scale tree patch (just using i_dentry instead of the (union
> eqivalent) i_rcu. I fixed it up (see below - the diff does not show that
> __xfs_inode_free has been removed) and can carry the fix as necessary.
> --
> Cheers,
> Stephen Rothwell sfr@xxxxxxxxxxxxxxxx
>
> diff --cc fs/xfs/xfs_iget.c
> index 3ecad00,d7de5a3..0000000
> --- a/fs/xfs/xfs_iget.c
> +++ b/fs/xfs/xfs_iget.c
> @@@ -157,17 -145,7 +156,17 @@@ xfs_inode_free
> ASSERT(!spin_is_locked(&ip->i_flags_lock));
> ASSERT(completion_done(&ip->i_flush));
>
> + /*
> + * Because we use RCU freeing we need to ensure the inode always
> + * appears to be reclaimed with an invalid inode number when in the
> + * free state. The ip->i_flags_lock provides the barrier against lookup
> + * races.
> + */
> + spin_lock(&ip->i_flags_lock);
> + ip->i_flags = XFS_IRECLAIM;
> + ip->i_ino = 0;
> + spin_unlock(&ip->i_flags_lock);
> - call_rcu((struct rcu_head *)&VFS_I(ip)->i_dentry, __xfs_inode_free);
> + call_rcu(&ip->i_vnode.i_rcu, xfs_inode_free_callback);
The fixed up call_rcu() shoul dbe:
+ call_rcu(&VFS_I(ip)->i_rcu, xfs_inode_free_callback);
Cheers,
Dave.
--
Dave Chinner
david@xxxxxxxxxxxxx
--
To unsubscribe from this list: send the line "unsubscribe linux-next" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
