Revert of [turbofan] Fixes the code generation for branches on x64 when the condition is Word64Equal.

Revert of [turbofan] Fixes the code generation for branches on x64 when the condition is Word64Equal. (patchset #1 id:1 of https://codereview.chromium.org/1677503002/ )

Reason for revert:
Code like the example given in the CL description was produced, for example, by code-stub-assembler.cc.

Reverting this, and try to fix the root cause instead.

Original issue's description:
> [turbofan] Fixes the code generation for branches on x64 when the condition is Word64Equal.
>
> Before:
>
> REX.W cmpq r9,r8
> setzl r8l
> movzxbl r8,r8
> REX.W cmpq r8,0x0
> jz 185
>
> After:
>
> REX.W cmpq r9,r8
> jnz 149
>
> Committed: https://crrev.com/75cc8352d06aada2e9131fdae793299ef73fb639
> Cr-Commit-Position: refs/heads/master@{#33784}

TBR=bmeurer@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.

Committed: https://crrev.com/a979a5c0cfcc39de656c7f6e2de7127a833aa914
Cr-Commit-Position: refs/heads/master@{#33884}

[epertoso, 2016-02-11 09:22:42]

Created Revert of [turbofan] Fixes the code generation for branches on x64 when
the condition is Word64Equal.

[commit-bot: I haz the power, 2016-02-11 10:41:15]

Description was changed from

==========
Revert of [turbofan] Fixes the code generation for branches on x64 when the
condition is Word64Equal. (patchset #1 id:1 of
https://codereview.chromium.org/1677503002/ )

Reason for revert:
Code like the example given in the CL description was produced, for example, by
code-stub-assembler.cc.

Reverting this, and try to fix the root cause instead.

Original issue's description:
> [turbofan] Fixes the code generation for branches on x64 when the condition is
Word64Equal.
> 
> Before:
> 
> REX.W cmpq r9,r8
> setzl r8l
> movzxbl r8,r8
> REX.W cmpq r8,0x0
> jz 185
> 
> After:
> 
> REX.W cmpq r9,r8
> jnz 149
> 
> Committed: https://crrev.com/75cc8352d06aada2e9131fdae793299ef73fb639
> Cr-Commit-Position: refs/heads/master@{#33784}

TBR=bmeurer@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
==========

to

==========
Revert of [turbofan] Fixes the code generation for branches on x64 when the
condition is Word64Equal. (patchset #1 id:1 of
https://codereview.chromium.org/1677503002/ )

Reason for revert:
Code like the example given in the CL description was produced, for example, by
code-stub-assembler.cc.

Reverting this, and try to fix the root cause instead.

Original issue's description:
> [turbofan] Fixes the code generation for branches on x64 when the condition is
Word64Equal.
>
> Before:
>
> REX.W cmpq r9,r8
> setzl r8l
> movzxbl r8,r8
> REX.W cmpq r8,0x0
> jz 185
>
> After:
>
> REX.W cmpq r9,r8
> jnz 149
>
> Committed: https://crrev.com/75cc8352d06aada2e9131fdae793299ef73fb639
> Cr-Commit-Position: refs/heads/master@{#33784}

TBR=bmeurer@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.

Committed: https://crrev.com/a979a5c0cfcc39de656c7f6e2de7127a833aa914
Cr-Commit-Position: refs/heads/master@{#33884}
==========

[caitp, 2017-08-01 16:12:23]

On 2016/02/11 09:22:42, epertoso wrote:
> Created Revert of [turbofan] Fixes the code generation for branches on x64
when
> the condition is Word64Equal.

I'm wondering if there's any more information about this revert? The CL
description looks relevant to
https://chromium-review.googlesource.com/c/594500/6/src%252Fcode-stub-assembl...

---

I'm guessing this didn't actually produce the intended effect (at least for
CodeStubAssembler code?)

I was thinking about how this could be solved as an arch-specific
InstructionSelector::VisitPhi(), so that Intel arches could represent the Phi
using the status register, instead of a new/temporary reg. But I wasn't sure how
much work it would be, or if that approach would actually work at all. For the
case linked above, it doesn't really matter in practice, but it would be
interesting to get it generating better code.