Re: [PATCH] vfs: don't hold s_umount over close_bdev_exclusive() call

[Al Viro <viro@ZenIV.linux.org.uk>]

On Tue, May 25, 2010 at 10:30:03AM +0200, Jens Axboe wrote:
> On Sat, May 22 2010, Tejun Heo wrote:
> > This patch fixes an obscure AB-BA deadlock in get_sb_bdev().
> > 
> > When a superblock is mounted more than once get_sb_bdev() calls
> > close_bdev_exclusive() to drop the extra bdev reference while holding
> > s_umount.  However, sb->s_umount nests inside bd_mutex during
> > __invalidate_device() and close_bdev_exclusive() acquires bd_mutex
> > during blkdev_put(); thus creating an AB-BA deadlock.
> > 
> > This condition doesn't trigger frequently.  For this condition to be
> > visible to lockdep, the filesystem must occupy the whole device (as
> > __invalidate_device() only grabs bd_mutex for the whole device), the
> > FS must be mounted more than once and partition rescan should be
> > issued while the FS is still mounted.
> > 
> > Fix it by dropping s_umount over close_bdev_exclusive().
> 
> Looks safe to me, since it has (as you note) an elevated ref count.

Ehh...  It's probably OK, but I'm worried about the interplay with
->bd_fsfreeze_mutex logics there ;-/