Separate CodeAssembler and CodeStubAssembler

Separate CodeAssembler and CodeStubAssembler

This separation is needed to make two goals possible simultaneously:
* is should be possible to offer V8 components a simple, clean
  interface to TurboFan's low-level code generation that doesn't
  expose details about the TF.
* it should be possible to easily create new CodeAssembler "macros"
  that don't require a review from an OWNER of the compiler directory.

Committed: https://crrev.com/973fc5b9b0a15c1a1f1fd3e3748258400c1b8349
Cr-Commit-Position: refs/heads/master@{#35576}

[danno, 2016-04-11 13:10:07]

Description was changed from

==========
Separate CodeAssembler and CodeStubAssembler
==========

to

==========
Separate CodeAssembler and CodeStubAssembler

This separation is needed to make two goals possible simultaneously:
* is should be possible to offer V8 components a simple, clean
  interface to TurboFan's low-level code generation that doesn't
  expose details about the TF.
* it should be possible to easily create new CodeAssembler "macros"
  that don't require a review from an OWNER of the compiler directory.
==========

[danno, 2016-04-11 13:10:07]

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

[danno, 2016-04-11 13:16:50]

As discussed last week. PTAL.

[Benedikt Meurer, 2016-04-11 18:12:09]

LGTM from my side modulo the Smi stuff and a nit.

https://codereview.chromium.org/1875583003/diff/80001/src/compiler/code-assem...
File src/compiler/code-assembler.h (left):

https://codereview.chromium.org/1875583003/diff/80001/src/compiler/code-assem...
src/compiler/code-assembler.h:280: // Tag a Word as a Smi value.
I think those Smi helpers should be in the CodeAssembler too, esp. since we have
SmiTag/SmiUntag there as well.

https://codereview.chromium.org/1875583003/diff/80001/src/compiler/code-assem...
File src/compiler/code-assembler.h (right):

https://codereview.chromium.org/1875583003/diff/80001/src/compiler/code-assem...
src/compiler/code-assembler.h:149: bool Is64();
Nit: const

[Michael Starzinger, 2016-04-14 09:04:23]

LGTM. Just a coupla nits.

https://codereview.chromium.org/1875583003/diff/80001/BUILD.gn
File BUILD.gn (right):

https://codereview.chromium.org/1875583003/diff/80001/BUILD.gn#newcode787
BUILD.gn:787: "src/compiler/code-stub-assembler.cc",
nit: These two files are gone now, let's drop them.

https://codereview.chromium.org/1875583003/diff/80001/src/fast-accessor-assem...
File src/fast-accessor-assembler.cc (right):

https://codereview.chromium.org/1875583003/diff/80001/src/fast-accessor-assem...
src/fast-accessor-assembler.cc:13: using v8::internal::CodeStubAssembler;
nit: The implementation inside this file now lives in the same namespace, this
using directive should no longer be needed here AFAICT.

https://codereview.chromium.org/1875583003/diff/80001/tools/gyp/v8.gyp
File tools/gyp/v8.gyp (right):

https://codereview.chromium.org/1875583003/diff/80001/tools/gyp/v8.gyp#newcod...
tools/gyp/v8.gyp:498: '../../src/compiler/code-assembler.cc',
nit: Please alpha-sort.

[danno, 2016-04-18 11:00:04]

Feedback addressed, I'll land this shortly.

https://codereview.chromium.org/1875583003/diff/80001/BUILD.gn
File BUILD.gn (right):

https://codereview.chromium.org/1875583003/diff/80001/BUILD.gn#newcode787
BUILD.gn:787: "src/compiler/code-stub-assembler.cc",
On 2016/04/14 09:04:23, Michael Starzinger wrote:
> nit: These two files are gone now, let's drop them.

Done.

https://codereview.chromium.org/1875583003/diff/80001/src/compiler/code-assem...
File src/compiler/code-assembler.h (left):

https://codereview.chromium.org/1875583003/diff/80001/src/compiler/code-assem...
src/compiler/code-assembler.h:280: // Tag a Word as a Smi value.
On 2016/04/11 18:12:09, Benedikt Meurer wrote:
> I think those Smi helpers should be in the CodeAssembler too, esp. since we
have
> SmiTag/SmiUntag there as well.
As discussed, I see them there too, but I can't move them until we move the
allocation wrappers out, too, and that's more work because they use the
RawMachineAssembler directly. I'll address that separately.

https://codereview.chromium.org/1875583003/diff/80001/src/compiler/code-assem...
File src/compiler/code-assembler.h (right):

https://codereview.chromium.org/1875583003/diff/80001/src/compiler/code-assem...
src/compiler/code-assembler.h:149: bool Is64();
On 2016/04/11 18:12:09, Benedikt Meurer wrote:
> Nit: const

Done.

https://codereview.chromium.org/1875583003/diff/80001/src/fast-accessor-assem...
File src/fast-accessor-assembler.cc (right):

https://codereview.chromium.org/1875583003/diff/80001/src/fast-accessor-assem...
src/fast-accessor-assembler.cc:13: using v8::internal::CodeStubAssembler;
On 2016/04/14 09:04:23, Michael Starzinger wrote:
> nit: The implementation inside this file now lives in the same namespace, this
> using directive should no longer be needed here AFAICT.

Done.

https://codereview.chromium.org/1875583003/diff/80001/tools/gyp/v8.gyp
File tools/gyp/v8.gyp (right):

https://codereview.chromium.org/1875583003/diff/80001/tools/gyp/v8.gyp#newcod...
tools/gyp/v8.gyp:498: '../../src/compiler/code-assembler.cc',
On 2016/04/14 09:04:23, Michael Starzinger wrote:
> nit: Please alpha-sort.

Done.

[danno, 2016-04-18 11:55:03]

The CQ bit was checked by danno@chromium.org

[danno, 2016-04-18 11:55:04]

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/1875583003/#ps100002
(title: "Fix gn build. Again.")

[commit-bot: I haz the power, 2016-04-18 11:55:06]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1875583003/100002
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1875583003/100002

[commit-bot: I haz the power, 2016-04-18 11:57:10]

Description was changed from

==========
Separate CodeAssembler and CodeStubAssembler

This separation is needed to make two goals possible simultaneously:
* is should be possible to offer V8 components a simple, clean
  interface to TurboFan's low-level code generation that doesn't
  expose details about the TF.
* it should be possible to easily create new CodeAssembler "macros"
  that don't require a review from an OWNER of the compiler directory.
==========

to

==========
Separate CodeAssembler and CodeStubAssembler

This separation is needed to make two goals possible simultaneously:
* is should be possible to offer V8 components a simple, clean
  interface to TurboFan's low-level code generation that doesn't
  expose details about the TF.
* it should be possible to easily create new CodeAssembler "macros"
  that don't require a review from an OWNER of the compiler directory.
==========

[commit-bot: I haz the power, 2016-04-18 11:57:11]

Committed patchset #8 (id:100002)

[commit-bot: I haz the power, 2016-04-18 11:58:11]

Description was changed from

==========
Separate CodeAssembler and CodeStubAssembler

This separation is needed to make two goals possible simultaneously:
* is should be possible to offer V8 components a simple, clean
  interface to TurboFan's low-level code generation that doesn't
  expose details about the TF.
* it should be possible to easily create new CodeAssembler "macros"
  that don't require a review from an OWNER of the compiler directory.
==========

to

==========
Separate CodeAssembler and CodeStubAssembler

This separation is needed to make two goals possible simultaneously:
* is should be possible to offer V8 components a simple, clean
  interface to TurboFan's low-level code generation that doesn't
  expose details about the TF.
* it should be possible to easily create new CodeAssembler "macros"
  that don't require a review from an OWNER of the compiler directory.

Committed: https://crrev.com/973fc5b9b0a15c1a1f1fd3e3748258400c1b8349
Cr-Commit-Position: refs/heads/master@{#35576}
==========

[commit-bot: I haz the power, 2016-04-18 11:58:12]

Patchset 8 (id:??) landed as
https://crrev.com/973fc5b9b0a15c1a1f1fd3e3748258400c1b8349
Cr-Commit-Position: refs/heads/master@{#35576}