[turbofan] Fix a bug in the RawMachineAssembler

[turbofan] Fix a bug in the RawMachineAssembler

This was causing code like:

REX.W cmpq r9,r8
setzl r8l
movzxbl r8,r8
REX.W cmpq r8,0x0
jz 185

(note the cmpq instead of cmpl above) on x64 instead of:

REX.W cmpq r9,r8
jnz 149

http://crrev.com/1677503002 is now obsolete and has been reverted.

Committed: https://crrev.com/61a4c528b705f36899b72b02fb9b959b70494c00
Cr-Commit-Position: refs/heads/master@{#33934}

[epertoso, 2016-02-10 18:16:55]

Patchset #1 (id:1) has been deleted

[epertoso, 2016-02-10 18:19:07]

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

[epertoso, 2016-02-10 18:19:07]

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

[Benedikt Meurer, 2016-02-11 05:10:51]

Description was changed from

==========
Fix a bug in the x64 instruction selector introduced in
https://codereview.chromium.org/1677503002.

While visiting a branch, a 64-bit word comparison between a Phi node and 0
yielded a kX64Cmp32 instruction.

BUG=
==========

to

==========
[turbofan] Fix a bug in the x64 instruction selector introduced in
https://crrev.com/1677503002.

While visiting a branch, a 64-bit word comparison between a Phi node and 0
yielded a kX64Cmp32 instruction.
==========

[Benedikt Meurer, 2016-02-11 05:13:53]

https://codereview.chromium.org/1685183003/diff/20001/src/compiler/x64/instru...
File src/compiler/x64/instruction-selector-x64.cc (right):

https://codereview.chromium.org/1685183003/diff/20001/src/compiler/x64/instru...
src/compiler/x64/instruction-selector-x64.cc:1457: case IrOpcode::kLoad:
This looks rather hacky to me, esp. hunting for Load/Phi and special casing
those. Only Word32Equal with 0 is ever used to negate a condition (actually
maybe we should really consider introducing BitEqual to avoid the confusion), so
performing this optimization for Word64Equal will probably not help anyway.

[epertoso, 2016-02-11 13:46:43]

Patchset #2 (id:40001) has been deleted

[epertoso, 2016-02-11 13:47:47]

PTAL

https://codereview.chromium.org/1685183003/diff/20001/src/compiler/x64/instru...
File src/compiler/x64/instruction-selector-x64.cc (right):

https://codereview.chromium.org/1685183003/diff/20001/src/compiler/x64/instru...
src/compiler/x64/instruction-selector-x64.cc:1457: case IrOpcode::kLoad:
On 2016/02/11 at 05:13:53, Benedikt Meurer wrote:
> This looks rather hacky to me, esp. hunting for Load/Phi and special casing
those. Only Word32Equal with 0 is ever used to negate a condition (actually
maybe we should really consider introducing BitEqual to avoid the confusion), so
performing this optimization for Word64Equal will probably not help anyway.

Yeah, I traced back the issue to the use of Word64BinaryNot. I've removed it
from the RawMachineAssembler.

[epertoso, 2016-02-11 13:59:33]

Description was changed from

==========
[turbofan] Fix a bug in the x64 instruction selector introduced in
https://crrev.com/1677503002.

While visiting a branch, a 64-bit word comparison between a Phi node and 0
yielded a kX64Cmp32 instruction.
==========

to

==========
[turbofan] Fix a bug in the RawMachineAssembler

This was causing code like:

REX.W cmpq r9,r8
setzl r8l
movzxbl r8,r8
REX.W cmpq r8,0x0
jz 185

(note the cmpq instead of cmpl above) on x64 instead of:

REX.W cmpq r9,r8
jnz 149


http://crrev.com/1677503002 is now obsolete and has been reverted.
==========

[Benedikt Meurer, 2016-02-11 17:34:58]

Nice. LGTM

[epertoso, 2016-02-12 08:52:55]

The CQ bit was checked by epertoso@chromium.org

[commit-bot: I haz the power, 2016-02-12 08:53:05]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1685183003/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1685183003/60001

[commit-bot: I haz the power, 2016-02-12 09:03:23]

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

[commit-bot: I haz the power, 2016-02-12 09:03:23]

Try jobs failed on following builders:
  v8_linux_arm_rel on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_arm_rel/builds/13647)

[epertoso, 2016-02-12 09:33:13]

The CQ bit was checked by epertoso@chromium.org

[epertoso, 2016-02-12 09:33:14]

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/1685183003/#ps80001
(title: "Fixes one ARM unit test.")

[commit-bot: I haz the power, 2016-02-12 09:33:22]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1685183003/80001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1685183003/80001

[commit-bot: I haz the power, 2016-02-12 10:03:02]

Committed patchset #3 (id:80001)

[commit-bot: I haz the power, 2016-02-12 10:03:19]

Description was changed from

==========
[turbofan] Fix a bug in the RawMachineAssembler

This was causing code like:

REX.W cmpq r9,r8
setzl r8l
movzxbl r8,r8
REX.W cmpq r8,0x0
jz 185

(note the cmpq instead of cmpl above) on x64 instead of:

REX.W cmpq r9,r8
jnz 149


http://crrev.com/1677503002 is now obsolete and has been reverted.
==========

to

==========
[turbofan] Fix a bug in the RawMachineAssembler

This was causing code like:

REX.W cmpq r9,r8
setzl r8l
movzxbl r8,r8
REX.W cmpq r8,0x0
jz 185

(note the cmpq instead of cmpl above) on x64 instead of:

REX.W cmpq r9,r8
jnz 149

http://crrev.com/1677503002 is now obsolete and has been reverted.

Committed: https://crrev.com/61a4c528b705f36899b72b02fb9b959b70494c00
Cr-Commit-Position: refs/heads/master@{#33934}
==========

[commit-bot: I haz the power, 2016-02-12 10:03:20]

Patchset 3 (id:??) landed as
https://crrev.com/61a4c528b705f36899b72b02fb9b959b70494c00
Cr-Commit-Position: refs/heads/master@{#33934}