Fixing test issue in test-cpu-profiler/TickLinesOptimized.

Fixing test issue in test-cpu-profiler/TickLinesOptimized.

The test was calling OptimizeFunctionOnNextCall on a function before
ever executing it - crankshaft therefore didn't have any type info and
was generating a soft deoptimization bailout. Make sure we execute the
function before calling OptimizeFunctionOnNextCall to avoid this issue.

BUG=

Committed: https://crrev.com/ca4707f775a7f67fa23808f0af1c5c4813911b14
Cr-Commit-Position: refs/heads/master@{#38171}

[georgia.kouveli, 2016-07-20 15:13:53]

georgia.kouveli@arm.com changed reviewers:
+ bmeurer@chromium.org, epertoso@chromium.org

[georgia.kouveli, 2016-07-20 15:13:54]

The test was calling OptimizeFunctionOnNextCall on a function before
ever executing it - crankshaft therefore didn't have any type info and
was generating a soft deoptimization bailout. The function would
eventually be marked to be optimized again, but the recompilation was not
guaranteed to happen before the test finishes execution and
func->abstract_code()
could still point to the CompileOptimized builtin.

This can happen on ARM64 (for certain revisions, in debug builds where the
size of the function ends up being larger than kMaxSizeEarlyOpt due to
--debug-code). Make sure we execute the function before calling
OptimizeFunctionOnNextCall to avoid this issue.

[Benedikt Meurer, 2016-07-20 17:39:28]

LGTM, thanks for fixing.

[georgia.kouveli, 2016-07-20 17:40:02]

The CQ bit was checked by georgia.kouveli@arm.com

[commit-bot: I haz the power, 2016-07-20 17:40:11]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-07-20 17:51:59]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-07-20 17:52:00]

Try jobs failed on following builders:
  v8_linux64_gyp_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_gyp_rel_ng/build...)
  v8_linux64_gyp_rel_ng_triggered on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_gyp_rel_ng_trigg...)

[georgia.kouveli, 2016-07-21 14:34:20]

I've been looking at the bot failures - I see three failing cases:
1) --nocrankshaft : 
 The assertion code->IsOptimized() that I added fails because the code obviously
doesn't get optimized. I can amend the assertion to check if crankshaft is
enabled (even though the test might not make sense if the code is not optimized,
it's essentially the same as TickLinesBaseline).
2) --ignition :
 Again the assertion code->IsOptimized() fails. func() gets executed with
ignition on the first call, then is optimized by crankshaft (which has no type
info so generates a soft deoptimization bailout), and the code falls back to
full-codegen.
3) --ignition --turbo --turbo-from-bytecode :
 Before my change, ignition wouldn't actually kick in for this test. The
function was directly compiled with turbofan (unless --ignition-eager was
specified, which would make the test fail before my changes too). Now func() is
first executed by ignition, then turbofan compiles an optimized version from
bytecode so doesn't have source info, and I get the same assertion failure as I
did on ARM64 when the code was not yet optimized (line_info->empty()).

What do you think? Would it be better to skip the test altogether when ignition
is enabled (in cctest.status) or change the test code to skip the test if
code->source_position_table()->length() is zero? (Looks like
https://codereview.chromium.org/2156753002 was also going to skip the test for
ignition, but got reverted).


On 2016/07/20 17:52:00, commit-bot: I haz the power wrote:
> Try jobs failed on following builders:
>   v8_linux64_gyp_rel_ng on master.tryserver.v8 (JOB_FAILED,
>
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_gyp_rel_ng/build...)
>   v8_linux64_gyp_rel_ng_triggered on master.tryserver.v8 (JOB_FAILED,
>
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_gyp_rel_ng_trigg...)

[Benedikt Meurer, 2016-07-22 04:01:22]

I think skipping the test with an appropriate TODO in the status file should be
OK for now.

[georgia.kouveli, 2016-07-22 12:54:07]

The CQ bit was checked by georgia.kouveli@arm.com to run a CQ dry run

[commit-bot: I haz the power, 2016-07-22 12:54:11]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-07-22 13:15:34]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-07-22 13:15:35]

Dry run: This issue passed the CQ dry run.

[georgia.kouveli, 2016-07-22 13:16:28]

PTAL!

[georgia.kouveli, 2016-07-29 10:00:40]

The CQ bit was checked by georgia.kouveli@arm.com

[georgia.kouveli, 2016-07-29 10:00:40]

The patchset sent to the CQ was uploaded after l-g-t-m from bmeurer@chromium.org
Link to the patchset: https://codereview.chromium.org/2168603003/#ps40001
(title: "Removed duplicate disabling of the test in cctest.status")

[commit-bot: I haz the power, 2016-07-29 10:00:48]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-07-29 10:21:22]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2016-07-29 10:22:11]

Description was changed from

==========
Fixing test issue in test-cpu-profiler/TickLinesOptimized.

The test was calling OptimizeFunctionOnNextCall on a function before
ever executing it - crankshaft therefore didn't have any type info and
was generating a soft deoptimization bailout. Make sure we execute the
function before calling OptimizeFunctionOnNextCall to avoid this issue.

BUG=
==========

to

==========
Fixing test issue in test-cpu-profiler/TickLinesOptimized.

The test was calling OptimizeFunctionOnNextCall on a function before
ever executing it - crankshaft therefore didn't have any type info and
was generating a soft deoptimization bailout. Make sure we execute the
function before calling OptimizeFunctionOnNextCall to avoid this issue.

BUG=

Committed: https://crrev.com/ca4707f775a7f67fa23808f0af1c5c4813911b14
Cr-Commit-Position: refs/heads/master@{#38171}
==========

[commit-bot: I haz the power, 2016-07-29 10:22:13]

Patchset 3 (id:??) landed as
https://crrev.com/ca4707f775a7f67fa23808f0af1c5c4813911b14
Cr-Commit-Position: refs/heads/master@{#38171}