[turbofan] Call ArgumentsAccessStub to materialize arguments.

[turbofan] Call ArgumentsAccessStub to materialize arguments.

This lowers JSCreateArgument nodes to call the ArgumentsAccessStub for
help with materializing arguments objects when possible. Along the way
this changes the calling convention of said stub to take parameters in
registers instead of on the stack.

R=mvstanton@chromium.org

Committed: https://crrev.com/9b12ec9ac2aff2978d05df775ce621ba4d90cbe8
Cr-Commit-Position: refs/heads/master@{#30919}

[Michael Starzinger, 2015-09-16 15:19:20]

Still work in progress, but ready for high-level comments.

https://codereview.chromium.org/1348773002/diff/1/src/full-codegen/ia32/full-...
File src/full-codegen/ia32/full-codegen-ia32.cc (right):

https://codereview.chromium.org/1348773002/diff/1/src/full-codegen/ia32/full-...
src/full-codegen/ia32/full-codegen-ia32.cc:310: type =
ArgumentsAccessStub::NEW_SLOPPY_FAST;
I am still fighting with this third kind of stub, register pressure is high in
there. But a first round of comments on the rest on whether you agree with the
general direction of this would be nice.

[Michael Starzinger, 2015-09-17 11:16:45]

Finished for one arch, I'll wait with the other arch ports until first round of
comments.

https://codereview.chromium.org/1348773002/diff/1/src/full-codegen/ia32/full-...
File src/full-codegen/ia32/full-codegen-ia32.cc (right):

https://codereview.chromium.org/1348773002/diff/1/src/full-codegen/ia32/full-...
src/full-codegen/ia32/full-codegen-ia32.cc:310: type =
ArgumentsAccessStub::NEW_SLOPPY_FAST;
On 2015/09/16 15:19:20, Michael Starzinger wrote:
> I am still fighting with this third kind of stub, register pressure is high in
> there. But a first round of comments on the rest on whether you agree with the
> general direction of this would be nice.

Done.

[mvstanton, 2015-09-17 13:29:45]

One comment to address, but the platform stuff looks good.

https://codereview.chromium.org/1348773002/diff/20001/src/compiler/js-typed-l...
File src/compiler/js-typed-lowering.cc (right):

https://codereview.chromium.org/1348773002/diff/20001/src/compiler/js-typed-l...
src/compiler/js-typed-lowering.cc:1118: : ArgumentsAccessStub::NEW_SLOPPY_FAST;
Oh wow, I'm not a fan of this multiple ternary op thing - your nice formatting
helps but still. Also, it would be nice if this logic for chosing the right stub
is either encoded in the CodeFactory call (which would need as input
s->has_duplicate_parameters() and p.type() to select it's type), or in the
ArgumentsAccessStub as a helper function.

[Michael Starzinger, 2015-09-23 17:25:46]

mstarzinger@chromium.org changed reviewers:
+ v8-mips-ports@googlegroups.com

[Michael Starzinger, 2015-09-23 17:25:46]

Some architecture ports done. Comment addressed.
Michael: PTAL. This is ready for a final review.
MIPS-Folks: I might need your help with MIPS ports.

[paul.l..., 2015-09-24 05:10:55]

On 2015/09/23 17:25:46, Michael Starzinger wrote:
> MIPS-Folks: I might need your help with MIPS ports.

Hey Michael - There is a MIPS32 port here:
https://codereview.chromium.org/1366713003. I'm sorry but it took a while to
debug some issues, and we don't have a MIPS64 port yet. If you really want to
land this, we could fix MIPS64 after the fact, but if you can wait we should
have it. With some luck, Ákos will have that before your end of day.

[Michael Starzinger, 2015-09-24 07:52:14]

On 2015/09/24 05:10:55, paul.l... wrote:
> On 2015/09/23 17:25:46, Michael Starzinger wrote:
> > MIPS-Folks: I might need your help with MIPS ports.
> 
> Hey Michael - There is a MIPS32 port here:
> https://codereview.chromium.org/1366713003. I'm sorry but it took a while to
> debug some issues, and we don't have a MIPS64 port yet. If you really want to
> land this, we could fix MIPS64 after the fact, but if you can wait we should
> have it. With some luck, Ákos will have that before your end of day.

Hey Paul! Thanks for the MIPS32 port, I already integrated your patch into this
CL. I'll just wait for the MIPS64 port, don't stress yourself, I am not in a
hurry.

[akos.palfi.imgtec, 2015-09-24 13:40:26]

akos.palfi@imgtec.com changed reviewers:
+ akos.palfi@imgtec.com

[akos.palfi.imgtec, 2015-09-24 13:40:27]

Hi Michael,

I've finished the MIPS64 part, please find the port in:
https://codereview.chromium.org/1364223002

[mvstanton, 2015-09-24 13:43:17]

And the rest looks good to me. LGTM.

[Michael Starzinger, 2015-09-24 14:47:01]

On 2015/09/24 13:40:27, akos.palfi.imgtec wrote:
> Hi Michael,
> 
> I've finished the MIPS64 part, please find the port in:
> https://codereview.chromium.org/1364223002

Hey Akos! Thanks for the MIPS64 port, integrated the patch.

[Michael Starzinger, 2015-09-24 15:28:41]

The CQ bit was checked by mstarzinger@chromium.org

[Michael Starzinger, 2015-09-24 15:28:41]

The patchset sent to the CQ was uploaded after l-g-t-m from
mvstanton@chromium.org
Link to the patchset: https://codereview.chromium.org/1348773002/#ps140001
(title: "Rebased.")

[commit-bot: I haz the power, 2015-09-24 15:28:50]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1348773002/140001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1348773002/140001

[commit-bot: I haz the power, 2015-09-24 15:30:36]

Committed patchset #8 (id:140001)

[commit-bot: I haz the power, 2015-09-24 15:30:53]

Patchset 8 (id:??) landed as
https://crrev.com/9b12ec9ac2aff2978d05df775ce621ba4d90cbe8
Cr-Commit-Position: refs/heads/master@{#30919}