[runtime] Deprecate RUNTIME_ASSERT from atomics.

[runtime] Deprecate RUNTIME_ASSERT from atomics.

This removes explicit uses of the RUNTIME_ASSERT macro from some runtime
function. The implicit ones in CONVERT_FOO_ARG_CHECKED will be addressed
in a separate CL for all runtime modules at once.

R=binji@chromium.org
BUG=v8:5066
Committed: https://crrev.com/21d7ec05692d705480b70dae563ff778579f6515
Cr-Commit-Position: refs/heads/master@{#36810}

[Michael Starzinger, 2016-06-03 13:21:21]

Description was changed from

==========
[runtime] Deprecate RUNTIME_ASSERT from atomics.

This removes explicit uses of the RUNTIME_ASSERT macro from some runtime
function. The implicit ones in CONVERT_FOO_ARG_CHECKED will be addressed
in a separate CL for all runtime modules at once.

R=
BUG=v8:5066
==========

to

==========
[runtime] Deprecate RUNTIME_ASSERT from atomics.

This removes explicit uses of the RUNTIME_ASSERT macro from some runtime
function. The implicit ones in CONVERT_FOO_ARG_CHECKED will be addressed
in a separate CL for all runtime modules at once.

R=binji@chromium.org
BUG=v8:5066
==========

[Michael Starzinger, 2016-06-03 13:21:21]

mstarzinger@chromium.org changed reviewers:
+ binji@chromium.org

[binji, 2016-06-03 18:01:42]

lgtm

Though I remember that it is possible for the fuzzer to generate code that will
trigger these (via the JS native functions), so it was necessary that it didn't
crash. Will CHECK crash?

[Michael Starzinger, 2016-06-06 09:16:41]

On 2016/06/03 18:01:42, binji wrote:
> lgtm
> 
> Though I remember that it is possible for the fuzzer to generate code that
will
> trigger these (via the JS native functions), so it was necessary that it
didn't
> crash. Will CHECK crash?

Just to clarify, by "JS native functions" do you mean the "%FooBar" syntax being
emitted by the fuzzer directly? AFAICT the fuzzer should remove all such direct
natives calls from the fuzzed snippets by now. Do you happen to have a link to
one of these ClusterFuzz reports when this was happening? Then I could
double-check.

To answer your question: Yes, a CHECK will crash and ClusterFuzz will report it.
So if there is any path how this could trigger (without fuzzed direct "%FooBar"
calls) then ClusterFuzz would find it and report it as a bug.

[binji, 2016-06-07 20:20:10]

On 2016/06/06 09:16:41, Michael Starzinger wrote:
> On 2016/06/03 18:01:42, binji wrote:
> > lgtm
> > 
> > Though I remember that it is possible for the fuzzer to generate code that
> will
> > trigger these (via the JS native functions), so it was necessary that it
> didn't
> > crash. Will CHECK crash?
> 
> Just to clarify, by "JS native functions" do you mean the "%FooBar" syntax
being
> emitted by the fuzzer directly? AFAICT the fuzzer should remove all such
direct
> natives calls from the fuzzed snippets by now. Do you happen to have a link to
> one of these ClusterFuzz reports when this was happening? Then I could
> double-check.

No, I just remember getting bugs about this when I was originally writing this a
year ago. It sounds like it has been fixed (yay!)

> 
> To answer your question: Yes, a CHECK will crash and ClusterFuzz will report
it.
> So if there is any path how this could trigger (without fuzzed direct
"%FooBar"
> calls) then ClusterFuzz would find it and report it as a bug.

Cool, sgtm

[Michael Starzinger, 2016-06-08 07:09:39]

The CQ bit was checked by mstarzinger@chromium.org

[commit-bot: I haz the power, 2016-06-08 07:09:44]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/2034063002/1

[commit-bot: I haz the power, 2016-06-08 07:38:44]

Description was changed from

==========
[runtime] Deprecate RUNTIME_ASSERT from atomics.

This removes explicit uses of the RUNTIME_ASSERT macro from some runtime
function. The implicit ones in CONVERT_FOO_ARG_CHECKED will be addressed
in a separate CL for all runtime modules at once.

R=binji@chromium.org
BUG=v8:5066
==========

to

==========
[runtime] Deprecate RUNTIME_ASSERT from atomics.

This removes explicit uses of the RUNTIME_ASSERT macro from some runtime
function. The implicit ones in CONVERT_FOO_ARG_CHECKED will be addressed
in a separate CL for all runtime modules at once.

R=binji@chromium.org
BUG=v8:5066
==========

[commit-bot: I haz the power, 2016-06-08 07:38:45]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-06-08 07:40:40]

Description was changed from

==========
[runtime] Deprecate RUNTIME_ASSERT from atomics.

This removes explicit uses of the RUNTIME_ASSERT macro from some runtime
function. The implicit ones in CONVERT_FOO_ARG_CHECKED will be addressed
in a separate CL for all runtime modules at once.

R=binji@chromium.org
BUG=v8:5066
==========

to

==========
[runtime] Deprecate RUNTIME_ASSERT from atomics.

This removes explicit uses of the RUNTIME_ASSERT macro from some runtime
function. The implicit ones in CONVERT_FOO_ARG_CHECKED will be addressed
in a separate CL for all runtime modules at once.

R=binji@chromium.org
BUG=v8:5066

Committed: https://crrev.com/21d7ec05692d705480b70dae563ff778579f6515
Cr-Commit-Position: refs/heads/master@{#36810}
==========

[commit-bot: I haz the power, 2016-06-08 07:40:40]

Patchset 1 (id:??) landed as
https://crrev.com/21d7ec05692d705480b70dae563ff778579f6515
Cr-Commit-Position: refs/heads/master@{#36810}