On Thu, Mar 07, 2013 at 07:47:31AM +0100, Lukáš Czerner wrote:
> On Wed, 6 Mar 2013, Theodore Ts'o wrote:
>
> > Date: Wed, 6 Mar 2013 17:58:18 -0500
> > From: Theodore Ts'o <tytso@xxxxxxx>
> > To: Zheng Liu <gnehzuil.liu@xxxxxxxxx>
> > Cc: linux-ext4@xxxxxxxxxxxxxxx, Zheng Liu <wenqing.lz@xxxxxxxxxx>,
> > Dmitry Monakhov <dmonakhov@xxxxxxxxxx>
> > Subject: Dev branch regressions
> >
> > On Wed, Mar 06, 2013 at 10:17:10PM +0800, Zheng Liu wrote:
> > >
> > > *Big Note*
> > > When I am testing this patch series, I found some regressions in dev branch.
> > > Here is a note. These regressions could be hitted by running test case
> > > serveral times. So If we just run xfstests one time, they could be missed.
> > >
> > > - xfstests #74 with data=journal
> > >
> > > - xfstests #247 with data=journal
> > > Some warning messages are printed by ext4_releasepage. We hit
> > > WARN_ON(PageChecked(page)) in this function. But the test case itself can
> > > pass.
> > >
> > > - xfstests #269 with dioread_nolock
> > > The system will hang
> >
> > I'm going to guess that you were running this using your SSD test
> > setup? I just ran:
> >
> > kvm-xfstests -c data_journal 74,74,74,74,74,247,247,247,247,247
> >
> > using my standard hdd setup, and didn't see any failures or warnings.
> >
> > How frequently are you seeing these failures? When I have a chance
> > I'll try running these tests with a tmpfs image and see if I have any
> > better luck reproducing the problem there.
> >
> > I did manage to get a hang (preceded with a soft lockup for the
> > dioread_nolock with test 269).
> >
> > > - xfstests #83 with bigalloc
> > > Some threads could be blocked for 120s.
> >
> > I've seen this test blocked for hours (but without managing to trigger
> > the 120s soft lockup warning), but I'm not entirely sure this was a
> > regression. I believe I've seen a similar hang with 3.8.0-rc3 if I
> > recall correctly. I had been hoping the changes with the extent
> > status tree would fix it, but apparently no such luck. :-(
>
> You're right this is not a regression the problem has always been
> there, however now with some bigalloc fixes it becomes more obvious.
> I have some patches to address this issue, though it's not ready
> yet.
Hi Ted, Lukas,
I take a close look at these problems, and I can confirm that they are
not regression because in 3.8 kernel they have been there. Honestly
they are hard to be hitted because I need to run several times to
trigger them. IMHO most users don't use this features (data=journal,
bigalloc, dioread_nolock). So let us fix them one by one. I will file
these bugs in mailing list to let folks know.
Regards,
- Zheng
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
