On Wed, 3 Jul 2002, Matthew Wilcox wrote:
> On Tue, Jul 02, 2002 at 06:25:47PM -0700, Paul Menage wrote:
> > This patch removes BKL protection from the invocation of the
> > super_operations ->statfs() method, and shifts it into the filesystems
> > where necessary. Any out-of-tree filesystems may need to take the BKL in
> > their statfs() methods if they were relying on it for synchronisation.
>
> Sure, makes sense to do. For real credit though, let's see how much we
> need the BKL. In ext2's statfs, we reference:
[snip]
> s_mount_opt doesn't actually need to be locked due to how it is
> modified & used. So it _looks_ like we only need to lock_super(sb); /
> unlock_super(sb); in ext2. Anyone more familiar with ext2 locking care
> to comment?
We actually don't need even that - the only places that might need
lock_super() (ext2_count_free_blocks()/ext2_count_free_inodes()) are
already taking it.
So BKL can be dropped in ext2_statfs() and no extra locks are needed.
> I bet most other filesystems can handle lock_super / unlock_super
> for themselves. See if some kerneljanitors are willing to help audit,
> perhaps?
Depending on filesystem, lock_super() might be a bad idea. Keep in
mind that these days ->s_lock is a private filesystem lock...
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sun Jul 07 2002 - 22:00:10 EST