Add native callbacks to FastAccessorAssembler.

Add native callbacks to FastAccessorAssembler.

(Trying to finish FastAccessorAssembler this week. This should make it easier to pick up the Blink side of this work later on.)

BUG=chromium:508898
SOUNDTRACK=http://youtu.be/i1EG-MKy4so
LOG=Y

Committed: https://crrev.com/d472cf448471dd6bfb7435ea8a69e4403da28233
Cr-Commit-Position: refs/heads/master@{#33671}

[vogelheim, 2016-01-22 14:11:24]

Description was changed from

==========
Add native callbacks to FastAccessorAssembler.

(Finishing off open work items. With this, we should be in a reasonable
 state to pick this up again on the Blink side.)

BUG=chromium:508898
==========

to

==========
Add native callbacks to FastAccessorAssembler.

(Finishing off open work items. With this, we should be in a reasonable
 state to pick this up again on the Blink side.)

BUG=chromium:508898
LOG=Y
==========

[vogelheim, 2016-01-22 17:17:14]

Description was changed from

==========
Add native callbacks to FastAccessorAssembler.

(Finishing off open work items. With this, we should be in a reasonable
 state to pick this up again on the Blink side.)

BUG=chromium:508898
LOG=Y
==========

to

==========
Add native callbacks to FastAccessorAssembler.

(Trying to finish FastAccessorAssembler this week. This should make it easier to
pick up the Blink side of this work later on.)

BUG=chromium:508898
SOUNDTRACK=http://youtu.be/i1EG-MKy4so
LOG=Y
==========

[vogelheim, 2016-01-22 17:17:14]

vogelheim@chromium.org changed reviewers:
+ epertoso@chromium.org, jochen@chromium.org

[epertoso, 2016-01-25 09:51:48]

lgtm

[jochen (gone - plz use gerrit), 2016-01-25 09:52:11]

nice

I was expecting that we'd need to write a new stub to get this? Or is that still
needed?

[jochen (gone - plz use gerrit), 2016-01-25 11:38:11]

lgtm after offline discussion

[vogelheim, 2016-01-25 17:10:32]

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

[vogelheim, 2016-01-25 17:10:33]

Benedikt... we've briefly talked about needing a stub for a C++ callback last
week. I've tried to re-use CallApiFunction stub instead. Does this make sense?

https://codereview.chromium.org/1620293002/diff/20001/src/compiler/fast-acces...
File src/compiler/fast-accessor-assembler.cc (right):

https://codereview.chromium.org/1620293002/diff/20001/src/compiler/fast-acces...
src/compiler/fast-accessor-assembler.cc:197: FromId(arg)
I'm not sure if I'm doing this right:

As I understand it, the stub itself has 5 params (described in the stub's
CallInterfaceDescriptor), the last being the # of additional params to pass to
the called function. So, logically, I have 6 parameters here. But as I read the
stub's code, it only expects five, and the rest on the stack (?). So I've added
a stack parameter (see below, 1 + stub.GetStackParameterCount() ), but now I
have to add this param twice to the args structure. That actually works, but it
looks wrong.

Honestly, I'm just guessing here... any comments?

[vogelheim, 2016-01-26 15:36:49]

>
https://codereview.chromium.org/1620293002/diff/20001/src/compiler/fast-acces...
> src/compiler/fast-accessor-assembler.cc:197: FromId(arg)
> I'm not sure if I'm doing this right:
> 
> As I understand it, the stub itself has 5 params (described in the stub's
> CallInterfaceDescriptor), the last being the # of additional params to pass to
> the called function. So, logically, I have 6 parameters here. But as I read
the
> stub's code, it only expects five, and the rest on the stack (?). So I've
added
> a stack parameter (see below, 1 + stub.GetStackParameterCount() ), but now I
> have to add this param twice to the args structure. That actually works, but
it
> looks wrong.
> 
> Honestly, I'm just guessing here... any comments?

So... the above doesn't make much sense; please ignore. What happened is that I
didn't know about the context parameter. I now changed the code & comments
accordingly.

Only.. I have no idea what I need to put in there. It looks like the stub always
expects it in a certain register and what I pass to CallN doesn't actually
matter, so I pass in UndefinedConstant().

[Benedikt Meurer, 2016-02-02 09:54:37]

LGTM according to the plan we discussed offline with danno.

[vogelheim, 2016-02-02 10:36:59]

The CQ bit was checked by vogelheim@chromium.org

[vogelheim, 2016-02-02 10:36:59]

The patchset sent to the CQ was uploaded after l-g-t-m from
epertoso@chromium.org, jochen@chromium.org
Link to the patchset: https://codereview.chromium.org/1620293002/#ps60001
(title: "clean-up parameter handling")

[commit-bot: I haz the power, 2016-02-02 10:37:30]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1620293002/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1620293002/60001

[commit-bot: I haz the power, 2016-02-02 10:59:30]

Description was changed from

==========
Add native callbacks to FastAccessorAssembler.

(Trying to finish FastAccessorAssembler this week. This should make it easier to
pick up the Blink side of this work later on.)

BUG=chromium:508898
SOUNDTRACK=http://youtu.be/i1EG-MKy4so
LOG=Y
==========

to

==========
Add native callbacks to FastAccessorAssembler.

(Trying to finish FastAccessorAssembler this week. This should make it easier to
pick up the Blink side of this work later on.)

BUG=chromium:508898
SOUNDTRACK=http://youtu.be/i1EG-MKy4so
LOG=Y
==========

[commit-bot: I haz the power, 2016-02-02 10:59:31]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2016-02-02 11:04:11]

Description was changed from

==========
Add native callbacks to FastAccessorAssembler.

(Trying to finish FastAccessorAssembler this week. This should make it easier to
pick up the Blink side of this work later on.)

BUG=chromium:508898
SOUNDTRACK=http://youtu.be/i1EG-MKy4so
LOG=Y
==========

to

==========
Add native callbacks to FastAccessorAssembler.

(Trying to finish FastAccessorAssembler this week. This should make it easier to
pick up the Blink side of this work later on.)

BUG=chromium:508898
SOUNDTRACK=http://youtu.be/i1EG-MKy4so
LOG=Y

Committed: https://crrev.com/d472cf448471dd6bfb7435ea8a69e4403da28233
Cr-Commit-Position: refs/heads/master@{#33671}
==========

[commit-bot: I haz the power, 2016-02-02 11:04:12]

Patchset 4 (id:??) landed as
https://crrev.com/d472cf448471dd6bfb7435ea8a69e4403da28233
Cr-Commit-Position: refs/heads/master@{#33671}