[turbofan] Further fixing ES6 tail call elimination in Turbofan.

[turbofan] Further fixing ES6 tail call elimination in Turbofan.

In case when F tail calls G we should also remove the potential arguments adaptor frame for F.

This CL introduces two new machine instructions ArchTailCallCodeObjectFromJSFunction and ArchTailCallJSFunctionFromJSFunction which (unlike existing ArchTailCallCodeObject and ArchTailCallJSFunction) also drop arguments adaptor frame if it exists right before jumping to the target function.

BUG=v8:4698
LOG=N

Committed: https://crrev.com/2aae579cf04b24f605d1ae6b975d67d8dbbee672
Cr-Commit-Position: refs/heads/master@{#34566}

[Igor Sheludko, 2016-02-19 13:59:43]

Patchset #3 (id:40001) has been deleted

[Igor Sheludko, 2016-02-19 14:09:25]

Patchset #3 (id:60001) has been deleted

[Igor Sheludko, 2016-02-19 14:15:30]

Patchset #3 (id:80001) has been deleted

[Igor Sheludko, 2016-02-19 15:42:46]

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

[commit-bot: I haz the power, 2016-02-19 15:43:04]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1702423002/120001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1702423002/120001

[Igor Sheludko, 2016-02-19 15:58:20]

Description was changed from

==========
WIP: [turbofan] Further fixing ES6 tail calls in Turbofan.

In case when F tail calls G we should also remove the potential arguments
adaptor frame for F.

BUG=v8:4698
LOG=N
==========

to

==========
[turbofan] Further fixing ES6 tail calls in Turbofan.

In case when F tail calls G we should also remove the potential arguments
adaptor frame for F.

This CL introduces two new machine instructions
ArchTailCallCodeObjectFromJSFunction and ArchTailCallJSFunctionFromJSFunction
which (unlike existing ArchTailCallCodeObject and ArchTailCallJSFunction) also
drop arguments adaptor frame if it exists right before jumping to the target
function. 

BUG=v8:4698
LOG=N
==========

[commit-bot: I haz the power, 2016-02-19 16:08:44]

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

[commit-bot: I haz the power, 2016-02-19 16:08:45]

Dry run: This issue passed the CQ dry run.

[Igor Sheludko, 2016-02-19 16:23:23]

ishell@chromium.org changed reviewers:
+ bmeurer@chromium.org

[Igor Sheludko, 2016-02-19 16:23:23]

PTAL

[Benedikt Meurer, 2016-02-22 12:30:49]

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

[Benedikt Meurer, 2016-02-22 12:30:49]

bmeurer@chromium.org changed required reviewers:
+ mstarzinger@chromium.org

[Benedikt Meurer, 2016-02-22 12:30:50]

LGTM from my side. I'd like to have a second opinion from Michi.

[Igor Sheludko, 2016-03-07 09:36:07]

Patchset #5 (id:140001) has been deleted

[Igor Sheludko, 2016-03-07 11:34:07]

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

[commit-bot: I haz the power, 2016-03-07 11:34:09]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1702423002/160001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1702423002/160001

[Igor Sheludko, 2016-03-07 11:37:16]

Description was changed from

==========
[turbofan] Further fixing ES6 tail calls in Turbofan.

In case when F tail calls G we should also remove the potential arguments
adaptor frame for F.

This CL introduces two new machine instructions
ArchTailCallCodeObjectFromJSFunction and ArchTailCallJSFunctionFromJSFunction
which (unlike existing ArchTailCallCodeObject and ArchTailCallJSFunction) also
drop arguments adaptor frame if it exists right before jumping to the target
function. 

BUG=v8:4698
LOG=N
==========

to

==========
[turbofan] Further fixing ES6 tail call elimination in Turbofan.

In case when F tail calls G we should also remove the potential arguments
adaptor frame for F.

This CL introduces two new machine instructions
ArchTailCallCodeObjectFromJSFunction and ArchTailCallJSFunctionFromJSFunction
which (unlike existing ArchTailCallCodeObject and ArchTailCallJSFunction) also
drop arguments adaptor frame if it exists right before jumping to the target
function. 

BUG=v8:4698
LOG=N
==========

[commit-bot: I haz the power, 2016-03-07 11:47:03]

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

[commit-bot: I haz the power, 2016-03-07 11:47:04]

Dry run: Try jobs failed on following builders:
  v8_linux_rel_ng on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_rel_ng/builds/2443)

[Michael Starzinger, 2016-03-07 21:47:32]

LGTM with one optional suggestion.

https://codereview.chromium.org/1702423002/diff/160001/src/compiler/arm/code-...
File src/compiler/arm/code-generator-arm.cc (right):

https://codereview.chromium.org/1702423002/diff/160001/src/compiler/arm/code-...
src/compiler/arm/code-generator-arm.cc:455: case
kArchTailCallCodeObjectFromJSFunction: {
suggestion: I would slightly prefer the common code to be shared, to make sure
it stays in sync. I was thinking about the following pattern ...

case kArchTailCallCodeObjectFromJSFunction:
case kArchTailCallCodeObject: {
  [...]
  if (opcode == kArchTailCallCodeObjectFromJSFunction) {
    AssemblePopArgumentsAdaptorFrame(...);
  }
  [...]
  break;
}

But I am not an expert on the backends and would be fine either way. So if
Benedikt suggested to implement it like this, then that's fine with me and we
can leave it as it is, just wanted to have it mentioned.

[Igor Sheludko, 2016-03-07 23:18:28]

Patchset #6 (id:180001) has been deleted

[Igor Sheludko, 2016-03-07 23:19:23]

Thanks!

https://codereview.chromium.org/1702423002/diff/160001/src/compiler/arm/code-...
File src/compiler/arm/code-generator-arm.cc (right):

https://codereview.chromium.org/1702423002/diff/160001/src/compiler/arm/code-...
src/compiler/arm/code-generator-arm.cc:455: case
kArchTailCallCodeObjectFromJSFunction: {
On 2016/03/07 21:47:32, Michael Starzinger wrote:
> suggestion: I would slightly prefer the common code to be shared, to make sure
> it stays in sync. I was thinking about the following pattern ...
> 
> case kArchTailCallCodeObjectFromJSFunction:
> case kArchTailCallCodeObject: {
>   [...]
>   if (opcode == kArchTailCallCodeObjectFromJSFunction) {
>     AssemblePopArgumentsAdaptorFrame(...);
>   }
>   [...]
>   break;
> }
> 
> But I am not an expert on the backends and would be fine either way. So if
> Benedikt suggested to implement it like this, then that's fine with me and we
> can leave it as it is, just wanted to have it mentioned.

Good idea! Done.

[Igor Sheludko, 2016-03-07 23:19:32]

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

[commit-bot: I haz the power, 2016-03-07 23:19:39]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1702423002/200001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1702423002/200001

[commit-bot: I haz the power, 2016-03-07 23:23:53]

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

[commit-bot: I haz the power, 2016-03-07 23:23:54]

Dry run: Try jobs failed on following builders:
  v8_linux_arm64_rel on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_arm64_rel/builds/1...)

[Igor Sheludko, 2016-03-07 23:44:44]

Patchset #6 (id:200001) has been deleted

[Igor Sheludko, 2016-03-07 23:44:54]

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

[commit-bot: I haz the power, 2016-03-07 23:44:59]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1702423002/220001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1702423002/220001

[commit-bot: I haz the power, 2016-03-08 00:04:29]

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

[commit-bot: I haz the power, 2016-03-08 00:04:31]

Dry run: This issue passed the CQ dry run.

[Igor Sheludko, 2016-03-08 00:08:24]

The CQ bit was checked by ishell@chromium.org

[Igor Sheludko, 2016-03-08 00:08:24]

The patchset sent to the CQ was uploaded after l-g-t-m from
bmeurer@chromium.org, mstarzinger@chromium.org
Link to the patchset: https://codereview.chromium.org/1702423002/#ps220001
(title: "Addressing comments")

[commit-bot: I haz the power, 2016-03-08 00:08:29]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1702423002/220001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1702423002/220001

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

Description was changed from

==========
[turbofan] Further fixing ES6 tail call elimination in Turbofan.

In case when F tail calls G we should also remove the potential arguments
adaptor frame for F.

This CL introduces two new machine instructions
ArchTailCallCodeObjectFromJSFunction and ArchTailCallJSFunctionFromJSFunction
which (unlike existing ArchTailCallCodeObject and ArchTailCallJSFunction) also
drop arguments adaptor frame if it exists right before jumping to the target
function. 

BUG=v8:4698
LOG=N
==========

to

==========
[turbofan] Further fixing ES6 tail call elimination in Turbofan.

In case when F tail calls G we should also remove the potential arguments
adaptor frame for F.

This CL introduces two new machine instructions
ArchTailCallCodeObjectFromJSFunction and ArchTailCallJSFunctionFromJSFunction
which (unlike existing ArchTailCallCodeObject and ArchTailCallJSFunction) also
drop arguments adaptor frame if it exists right before jumping to the target
function. 

BUG=v8:4698
LOG=N
==========

[commit-bot: I haz the power, 2016-03-08 00:12:05]

Committed patchset #6 (id:220001)

[commit-bot: I haz the power, 2016-03-08 00:12:22]

Description was changed from

==========
[turbofan] Further fixing ES6 tail call elimination in Turbofan.

In case when F tail calls G we should also remove the potential arguments
adaptor frame for F.

This CL introduces two new machine instructions
ArchTailCallCodeObjectFromJSFunction and ArchTailCallJSFunctionFromJSFunction
which (unlike existing ArchTailCallCodeObject and ArchTailCallJSFunction) also
drop arguments adaptor frame if it exists right before jumping to the target
function. 

BUG=v8:4698
LOG=N
==========

to

==========
[turbofan] Further fixing ES6 tail call elimination in Turbofan.

In case when F tail calls G we should also remove the potential arguments
adaptor frame for F.

This CL introduces two new machine instructions
ArchTailCallCodeObjectFromJSFunction and ArchTailCallJSFunctionFromJSFunction
which (unlike existing ArchTailCallCodeObject and ArchTailCallJSFunction) also
drop arguments adaptor frame if it exists right before jumping to the target
function.

BUG=v8:4698
LOG=N

Committed: https://crrev.com/2aae579cf04b24f605d1ae6b975d67d8dbbee672
Cr-Commit-Position: refs/heads/master@{#34566}
==========

[commit-bot: I haz the power, 2016-03-08 00:12:24]

Patchset 6 (id:??) landed as
https://crrev.com/2aae579cf04b24f605d1ae6b975d67d8dbbee672
Cr-Commit-Position: refs/heads/master@{#34566}