On Thu, Feb 14, 2019 at 09:11:20AM +0100, Christoph Hellwig wrote: > On Wed, Feb 13, 2019 at 07:02:53PM +0100, Christoph Hellwig wrote: > > On Mon, Feb 11, 2019 at 04:03:37PM -0800, Darrick J. Wong wrote: > > > I pulled this series in and saw a bunch of new regressions in > > > generic/127 generic/269 generic/476 generic/521 generic/522 xfs/442 with > > > quota, rmap, and reflink enabled. The regressions were all the > > > XFS_WANT_CORRUPTED_GOTO near the top of xfs_rmap_unmap_shared. > > > > Quick status update: I managed to reproduce this ones on xfs/442 only > > yesterday. But as soon as I started bisecting it it disappeared again.. > > And another full overnight run couldn't report this even once. > > How reproducible is the issue for you? Are you sure it never happens > with the baseline, given that this series really doesn't touch anything > interacting with the rmap? Nearly every time, so I'll go digging and report back what I find. (It's odd, because I didn't see anything obvious in the patches that would cause this...) --D
