[turbofan] Consume SignedSmall feedback for number operations.

[turbofan] Consume SignedSmall feedback for number operations.

So far we treated SignedSmall and Signed32 feedback the same for number
operations. However it would be beneficial to generate (a lot) less code
if we only do a Smi check on the inputs instead of doing the full Smi +
HeapNumber + conversion check that we need to do for Signed32 feedback.

R=epertoso@chromium.org
BUG=v8:4583
Committed: https://crrev.com/cf4b9307ad5fa3864244ec23883f0ac5094372cc
Cr-Commit-Position: refs/heads/master@{#38290}

[Benedikt Meurer, 2016-08-03 12:12:03]

The CQ bit was checked by bmeurer@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-08-03 12:12:14]

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

[epertoso, 2016-08-03 12:21:39]

lgtm

[Benedikt Meurer, 2016-08-03 12:23:22]

The CQ bit was unchecked by bmeurer@chromium.org

[Benedikt Meurer, 2016-08-03 12:23:25]

The CQ bit was checked by bmeurer@chromium.org

[commit-bot: I haz the power, 2016-08-03 12:23:35]

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, 2016-08-03 12:45:41]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-08-03 12:47:59]

Description was changed from

==========
[turbofan] Consume SignedSmall feedback for number operations.

So far we treated SignedSmall and Signed32 feedback the same for number
operations. However it would be beneficial to generate (a lot) less code
if we only do a Smi check on the inputs instead of doing the full Smi +
HeapNumber + conversion check that we need to do for Signed32 feedback.

R=epertoso@chromium.org
BUG=v8:4583
==========

to

==========
[turbofan] Consume SignedSmall feedback for number operations.

So far we treated SignedSmall and Signed32 feedback the same for number
operations. However it would be beneficial to generate (a lot) less code
if we only do a Smi check on the inputs instead of doing the full Smi +
HeapNumber + conversion check that we need to do for Signed32 feedback.

R=epertoso@chromium.org
BUG=v8:4583

Committed: https://crrev.com/cf4b9307ad5fa3864244ec23883f0ac5094372cc
Cr-Commit-Position: refs/heads/master@{#38290}
==========

[commit-bot: I haz the power, 2016-08-03 12:48:00]

Patchset 1 (id:??) landed as
https://crrev.com/cf4b9307ad5fa3864244ec23883f0ac5094372cc
Cr-Commit-Position: refs/heads/master@{#38290}

[Michael Achenbach, 2016-08-03 19:15:04]

machenbach@chromium.org changed reviewers:
+ machenbach@chromium.org

[Michael Achenbach, 2016-08-03 19:15:04]

Any chance this started here?
https://build.chromium.org/p/client.v8/builders/V8%20Linux%20-%20gc%20stress/...

[Benedikt Meurer, 2016-08-04 03:34:20]

On 2016/08/03 19:15:04, Michael Achenbach (slow) wrote:
> Any chance this started here?
>
https://build.chromium.org/p/client.v8/builders/V8%20Linux%20-%20gc%20stress/...

Highly unlikely, this test doesn't even compile a single function with TurboFan,
even when run with --turbo.

[Benedikt Meurer, 2016-08-04 03:40:42]

On 2016/08/04 03:34:20, Benedikt Meurer wrote:
> On 2016/08/03 19:15:04, Michael Achenbach (slow) wrote:
> > Any chance this started here?
> >
>
https://build.chromium.org/p/client.v8/builders/V8%20Linux%20-%20gc%20stress/...
> 
> Highly unlikely, this test doesn't even compile a single function with
TurboFan,
> even when run with --turbo.

Looks a lot like WASM calls to the runtime directly from a WASM function, then a
GC happens and relocates some of the parameters (which WASM doesn't mark as
tagged and so the GC doesn't update the pointer), and then the runtime function
accesses a stale pointer.