Re: Loopback device hung [was Re: xfs deadlock on 3.9-rc5 running xfstests case #78]

[Jens Axboe <axboe@kernel.dk>]

On Tue, Apr 09 2013, Dave Chinner wrote:
> On Tue, Apr 09, 2013 at 09:01:39AM +0200, Jens Axboe wrote:
> > On Tue, Apr 09 2013, Dave Chinner wrote:
> > > On Thu, Apr 04, 2013 at 04:30:54PM -0400, Phillip Susi wrote:
> > > > > I have not tested it yet, but I am pretty sure it won't work.  It
> > > > > looks like the patch changes the BLKRRPART path to go ahead and remove
> > > > > existing partitions when GENHD_FL_NO_PARTSCAN is set.  loop doesn't
> > > > > issue the BLKRRPART ioctl when !LO_FLAGS_PARTSCAN so this won't help.
> > > > >  I think loop needs to set GENHD_FL_NO_PARTSCAN and then issue the
> > > > > ioctl regardless of the LO_FLAGS_PARTSCAN flag to get the partitions
> > > > > to be removed.  I will try to test tonight.
> > > > 
> > > > After testing, my initial thoughts appeared to have been correct.  I had
> > > > to modify the patch as follows.  To test, simply do:
> > > > 
> > > > truncate -s 10m img
> > > > losetup /dev/loop0 img
> > > > parted /dev/loop0
> > > > mklabel msdos
> > > > mkpart primary ext2 1m 2m
> > > > quit
> > > > ls /dev/loop0*
> > > > 
> > > > Note the /dev/loop0p1 node.  Run losetup -d /dev/loop0 and see if it is
> > > > still there.
> > > 
> > > Jens, can we get one of these fixes merged quickly? xfstests is
> > > unusable on any kernel more recent than 3.9-rc4 because of these
> > > problems....
> > 
> > Yep, did the revert yesterday and it's going out today.
> 
> Great, I didn't see an update in the latest -rc, so I wanted to make
> sure it hadn't fallen through he cracks....

Got pushed out yesterday, it is now upstream.

-- 
Jens Axboe