[fullcodegen] Factor out EmitNamedPropertyLoad.

[fullcodegen] Factor out EmitNamedPropertyLoad.

This makes the FullCodeGenerator::EmitNamedPropertyLoad be architecture
independent by adding MacroAssembler::Move helpers.

R=bmeurer@chromium.org

Committed: https://crrev.com/9c53fcad7867640d051e2a3ede1afacb836ff19d
Cr-Commit-Position: refs/heads/master@{#34259}

[Michael Starzinger, 2016-02-24 13:28:51]

https://codereview.chromium.org/1734643002/diff/1/src/ia32/macro-assembler-ia...
File src/ia32/macro-assembler-ia32.h (right):

https://codereview.chromium.org/1734643002/diff/1/src/ia32/macro-assembler-ia...
src/ia32/macro-assembler-ia32.h:799: void Move(Register dst, Handle<Object> x) {
mov(dst, Immediate(x)); }
I am using the "mov" here that can only deal with handles referencing a
HeapObject in old-space. On some architectures the existing "Move" is more
generic. We could delegate to "LoadObject" to have this be the generic version.
Either is fine with me. Please advise which is preferred here.

[Benedikt Meurer, 2016-02-24 13:38:30]

Nice. LGTM.

https://codereview.chromium.org/1734643002/diff/1/src/ia32/macro-assembler-ia...
File src/ia32/macro-assembler-ia32.h (right):

https://codereview.chromium.org/1734643002/diff/1/src/ia32/macro-assembler-ia...
src/ia32/macro-assembler-ia32.h:799: void Move(Register dst, Handle<Object> x) {
mov(dst, Immediate(x)); }
I don't care. Either is fine with me.

[Igor Sheludko, 2016-02-24 13:56:46]

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

[Igor Sheludko, 2016-02-24 13:56:46]

https://codereview.chromium.org/1734643002/diff/1/src/ia32/macro-assembler-ia...
File src/ia32/macro-assembler-ia32.h (right):

https://codereview.chromium.org/1734643002/diff/1/src/ia32/macro-assembler-ia...
src/ia32/macro-assembler-ia32.h:799: void Move(Register dst, Handle<Object> x) {
mov(dst, Immediate(x)); }
On 2016/02/24 13:28:51, Michael Starzinger wrote:
> I am using the "mov" here that can only deal with handles referencing a
> HeapObject in old-space. On some architectures the existing "Move" is more
> generic. We could delegate to "LoadObject" to have this be the generic
version.
> Either is fine with me. Please advise which is preferred here.

I think we should make it safe for any kind of handles, like on x64 for example,
otherwise it would be so easy to misuse it during poring x64 change to ia32.

[Michael Starzinger, 2016-02-24 17:13:58]

Thanks. Addressed comments. Landing.

https://codereview.chromium.org/1734643002/diff/1/src/ia32/macro-assembler-ia...
File src/ia32/macro-assembler-ia32.h (right):

https://codereview.chromium.org/1734643002/diff/1/src/ia32/macro-assembler-ia...
src/ia32/macro-assembler-ia32.h:799: void Move(Register dst, Handle<Object> x) {
mov(dst, Immediate(x)); }
On 2016/02/24 13:56:46, Igor Sheludko wrote:
> On 2016/02/24 13:28:51, Michael Starzinger wrote:
> > I am using the "mov" here that can only deal with handles referencing a
> > HeapObject in old-space. On some architectures the existing "Move" is more
> > generic. We could delegate to "LoadObject" to have this be the generic
> version.
> > Either is fine with me. Please advise which is preferred here.
> 
> I think we should make it safe for any kind of handles, like on x64 for
example,
> otherwise it would be so easy to misuse it during poring x64 change to ia32.

Done. Thanks for the data point. Done for all architecture where it applies
(i.e. ARM64, ia32, MIPS, MIPS64 and x87).

[Michael Starzinger, 2016-02-24 18:10:47]

The CQ bit was checked by mstarzinger@chromium.org

[Michael Starzinger, 2016-02-24 18:10:48]

The patchset sent to the CQ was uploaded after l-g-t-m from bmeurer@chromium.org
Link to the patchset: https://codereview.chromium.org/1734643002/#ps60001
(title: "Rebased.")

[commit-bot: I haz the power, 2016-02-24 18:11:05]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1734643002/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1734643002/60001

[commit-bot: I haz the power, 2016-02-24 18:18:19]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2016-02-24 18:18:50]

Description was changed from

==========
[fullcodegen] Factor out EmitNamedPropertyLoad.

This makes the FullCodeGenerator::EmitNamedPropertyLoad be architecture
independent by adding MacroAssembler::Move helpers.

R=bmeurer@chromium.org
==========

to

==========
[fullcodegen] Factor out EmitNamedPropertyLoad.

This makes the FullCodeGenerator::EmitNamedPropertyLoad be architecture
independent by adding MacroAssembler::Move helpers.

R=bmeurer@chromium.org

Committed: https://crrev.com/9c53fcad7867640d051e2a3ede1afacb836ff19d
Cr-Commit-Position: refs/heads/master@{#34259}
==========

[commit-bot: I haz the power, 2016-02-24 18:18:52]

Patchset 4 (id:??) landed as
https://crrev.com/9c53fcad7867640d051e2a3ede1afacb836ff19d
Cr-Commit-Position: refs/heads/master@{#34259}