linux_aura: Turn aura on on the builders.

linux_aura: Turn aura on on the builders.

This widens some suppressions from windows to linux, and suppresses some memory errors in third party libraries that we now use.

BUG=125106, 303342
R=cpu@chromium.org, mseaborn@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=227210

[Elliot Glaysher, 2013-10-02 00:58:50]

Carlos, I now have all green linux trybots with linux_aura by default,
suppressing 

Mark, we are making a run at shipping linux_aura in M32. These two tests are the
only nacl-y things that don't work. (The nacl_integration tests were never run
previously.) Both of the suppressions look identical, but I can file a different
bug for the second if you prefer.

[cpu_(ooo_6.6-7.5), 2013-10-02 02:49:27]

lgtm on my side adding Dirk to the CL, sorry :)

[Mark Seaborn, 2013-10-02 03:04:06]

On 1 October 2013 17:58, <erg@chromium.org> wrote:

> Reviewers: cpu, Mark Seaborn,
>
> Message:
> Carlos, I now have all green linux trybots with linux_aura by default,
> suppressing
>
> Mark, we are making a run at shipping linux_aura in M32. These two tests
> are the
> only nacl-y things that don't work. (The nacl_integration tests were never
> run
> previously.) Both of the suppressions look identical, but I can file a
> different
> bug for the second if you prefer.
>

When I look at the trybot log (
http://build.chromium.org/p/tryserver.chromium/builders/linux_rel/builds/1741...),
I see:

crash_dump_tester: Found 1 files
  /tmp/nacl_crash_dump_tester__Zb8Ct/.config/chromium/Crash
Reports/chromium-gpu-process-minidump-aa8c212dcb3701fa.dmp (size 132450)
crash_dump_tester: ERROR: Got 1 crash dumps but expected 0

crash_dump_tester: Found 1 files
  /tmp/nacl_crash_dump_tester_7mEzAM/.config/chromium/Crash
Reports/chromium-gpu-process-minidump-b69e6318da26c756.dmp (size 136546)
crash_dump_tester: ERROR: Got 1 crash dumps but expected 0

The fact that you're getting a crash report from the GPU process in a
GPU-related change is a pretty good sign that there is a real bug with the
GPU process that you should probably investigate, and that this isn't just
flakiness related to Breakpad, NaCl or the test infrastructure. :-)

Bear in mind that:

 * nacl_integration is one of the few test suites that runs the chrome
executable.  In contrast, browser_tests doesn't run the real chrome
executable, it just links in some libraries and calls those.

 * nacl_integration contains the only mostly-end-to-end test coverage of
Breakpad on the Chromium bots, AFAIK.  If the GPU process is crashing in
other test suites, that might very well not be reported as a test failure.

Cheers,
Mark

To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[Dirk Pranke, 2013-10-02 18:09:35]

I would expect this to break the blink bots. Can you confirm with a blink
(layout) tryjob? We probably need to manually force aura and ash off on linux on
those bots like we do on windows until I can get it working.

[Dirk Pranke, 2013-10-02 20:46:44]

On 2013/10/02 18:09:35, Dirk Pranke wrote:
> I would expect this to break the blink bots. Can you confirm with a blink
> (layout) tryjob? We probably need to manually force aura and ash off on linux
on
> those bots like we do on windows until I can get it working.

Okay, I posted a patch to force ash and aura off at
https://codereview.chromium.org/25792002/ . Once that lands and the bots are
restarted, this should be safe.

[Elliot Glaysher, 2013-10-02 21:35:56]

+piman

Mark, I talked to Antoine IRL and he volunteered to look at  the GPU post-land.
He thinks it may be a simple flag issue.

[Mark Seaborn, 2013-10-02 22:00:12]

On 2 October 2013 14:35, <erg@chromium.org> wrote:

> +piman
>
> Mark, I talked to Antoine IRL and he volunteered to look at  the GPU
> post-land.
> He thinks it may be a simple flag issue.


LGTM

Mark

To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[Elliot Glaysher, 2013-10-05 21:14:55]

Committed patchset #6 manually as r227210.