On Wed, Dec 04, 2019 at 05:29:54PM +0100, David Sterba wrote:
> On Wed, Dec 04, 2019 at 05:17:23PM +0100, David Sterba wrote:
> > On Wed, Dec 04, 2019 at 02:08:14PM +0800, Qu Wenruo wrote:
> > > On 2019/12/4 下午2:06, Qu Wenruo wrote:
> > > > Found a strange failure at btrfs-progs selftest misc/016.
> > > >
> > > > It hangs at current master 63de37476ebd ("Merge tag
> > > > 'tag-chrome-platform-for-v5.5' of
> > > > git://git.kernel.org/pub/scm/linux/kernel/git/chrome-platform/linux").
> > > >
> > > > What makes it even more mysterious is, misc-5.5 passes.
> > > >
> > > > Since there is no other btrfs pull, it's caused by some upstream
> > > > non-btrfs commits.
> > >
> > > BTW, no blocked kernel thread, no CPU usage.
> > >
> > > And the hang can be canceled with a signal.
> > >
> > > >
> > > > Doing bisecting, but any clue will be appreciated.
> >
> > This looks like a problem caused by recent merge of updates to pipe
> > (commit 6a965666b7e7475c2f8c8e724703db58b8a8a445). I can reproduce that
> > and now building test kernel before that commit. As this also blocks all
> > fstests that use send, we have to wait until it's fixed in master
> > branch.
>
> That hangs too, so that's going to be a long bisect.
Well the test took slightly longer than expected, so
commit 6a965666b7e7 -- first bad merge
commit 32ef9553635a -- last good merge
There are 13 commits in the pipe branch so I'll have to bisect anyway
but at least it's going to be faster.
