Separate GPU Trace Test into 2 tests.

Separate GPU Trace Test into 2 tests.

Not all devices support GPU Tracing, but it is still useful to
be able run a trace test which tests whether or not traces are
properly plumbed all the way through the stack.

This CL separates out trace tests into 2 different tests, the
original "trace_test" only tests the CPU side and does not test
anything on the GPU. The "device_trace_test" will be used to test
whether or not GPU traces are working properly on devices which
support it.

Currently the device trace test is set to skip unconditionally.
Once we have the recipes set up to run the "device_trace_test" we
can enable this for devices that support it.

BUG=455324
TEST=trybots

Committed: https://crrev.com/f14b5b44f87fd607a05289a98df05860d862159a
Cr-Commit-Position: refs/heads/master@{#314614}

[David Yen, 2015-01-27 19:30:59]

dyen@chromium.org changed reviewers:
+ kbr@chromium.org, vmiura@chromium.org

[David Yen, 2015-01-27 19:33:03]

I've decided to move out the GPU related testing to a separate
"device_trace_test", that way we can have the basic "trace_test" run on all
machines regardless of whether the GPU supports tracing or not.

Once this CL lands I will create another CL which adds the device_trace_test to
the recipes that are run on the bots.

PTAL.

[Ken Russell (switch to Gerrit), 2015-01-30 20:37:05]

kbr@chromium.org changed reviewers:
+ bajones@chromium.org, zmo@chromium.org

[Ken Russell (switch to Gerrit), 2015-01-30 20:37:05]

+bajones and zmo as alternate reviewers

Looking for vmiura@ to do the first review of this new test.

[David Yen, 2015-02-02 23:37:00]

On 2015/01/30 20:37:05, Ken Russell wrote:
> +bajones and zmo as alternate reviewers
> 
> Looking for vmiura@ to do the first review of this new test.

ping, looking forward to getting these tests in so we don't regress on traces
crashing.

[vmiura, 2015-02-02 23:42:15]

LGTM + nit.

https://codereview.chromium.org/874023003/diff/60001/content/test/gpu/gpu_tes...
File content/test/gpu/gpu_tests/trace_test.py (right):

https://codereview.chromium.org/874023003/diff/60001/content/test/gpu/gpu_tes...
content/test/gpu/gpu_tests/trace_test.py:65: class
_TraceValidator(_TraceValidatorBase):
nit: "Service"TraceValidator, and similar in the other places?

[David Yen, 2015-02-02 23:52:46]

https://codereview.chromium.org/874023003/diff/60001/content/test/gpu/gpu_tes...
File content/test/gpu/gpu_tests/trace_test.py (right):

https://codereview.chromium.org/874023003/diff/60001/content/test/gpu/gpu_tes...
content/test/gpu/gpu_tests/trace_test.py:65: class
_TraceValidator(_TraceValidatorBase):
On 2015/02/02 23:42:15, vmiura wrote:
> nit: "Service"TraceValidator, and similar in the other places?

I actually had this name originally, but I think a plain "trace test" is more
applicable than "service trace test" here. The reason being these tests aren't
necessarily testing "service" vs "device" traces (gpu_tracer_unittest is in
charge of testing that those 2 varieties are properly written).

These tests are trying to test if traceBeginCHROMIUM() is plumbed all the way
from the client to GPUTracer in the GPU process. A simple "Trace Test" name
makes more sense here IMHO.

[Ken Russell (switch to Gerrit), 2015-02-04 01:49:38]

LGTM.

I suggest filing tracking bugs for work that spans multiple CLs.

[David Yen, 2015-02-04 19:26:10]

On 2015/02/04 01:49:38, Ken Russell wrote:
> LGTM.
> 
> I suggest filing tracking bugs for work that spans multiple CLs.

Done. I filled out a tracking bug.

[David Yen, 2015-02-04 19:26:33]

The CQ bit was checked by dyen@chromium.org

[commit-bot: I haz the power, 2015-02-04 19:26:50]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/874023003/60001

[commit-bot: I haz the power, 2015-02-04 20:07:00]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2015-02-04 20:08:36]

Patchset 4 (id:??) landed as
https://crrev.com/f14b5b44f87fd607a05289a98df05860d862159a
Cr-Commit-Position: refs/heads/master@{#314614}