[turbofan] ARM64: Use zr to store immediate zero

[turbofan] ARM64: Use zr to store immediate zero

When storing an immediate integer or floating point zero, use the zero register
as the source value. This avoids the need to sometimes allocate a new register.

BUG=

Committed: https://crrev.com/0322c20d178db0b3360af0353972bc338b2d4b18
Cr-Commit-Position: refs/heads/master@{#36013}

[martyn.capewell, 2016-05-03 17:18:57]

martyn.capewell@arm.com changed reviewers:
+ bmeurer@chromium.org, danno@chromium.org, jarin@chromium.org

[Benedikt Meurer, 2016-05-03 17:33:20]

LGTM with comments.

https://codereview.chromium.org/1945783002/diff/1/src/compiler/arm64/code-gen...
File src/compiler/arm64/code-generator-arm64.cc (right):

https://codereview.chromium.org/1945783002/diff/1/src/compiler/arm64/code-gen...
src/compiler/arm64/code-generator-arm64.cc:40: (copysignf(1.0f,
InputFloat32(index)) > 0.0f));
Use std::signbit or bit_cast<int32_t>(InputFloat32(index)) == 0

https://codereview.chromium.org/1945783002/diff/1/src/compiler/arm64/code-gen...
src/compiler/arm64/code-generator-arm64.cc:51: (copysign(1.0,
InputDouble(index)) > 0.0));
Use std::signbit or bit_cast<int64_t>(InputFloat64(index)) == 0

https://codereview.chromium.org/1945783002/diff/1/src/compiler/arm64/instruct...
File src/compiler/arm64/instruction-selector-arm64.cc (right):

https://codereview.chromium.org/1945783002/diff/1/src/compiler/arm64/instruct...
src/compiler/arm64/instruction-selector-arm64.cc:45: (copysign(1.0,
GetFloatConstantValue(node)) > 0.0))) {
Use std::signbit here, or even easier use
bit_cast<int64_t>(GetFloatConstantValue(node)) == V8_INT64_C(0)

https://codereview.chromium.org/1945783002/diff/1/src/compiler/arm64/instruct...
src/compiler/arm64/instruction-selector-arm64.cc:82: DCHECK(node->opcode() ==
IrOpcode::kFloat64Constant);
Nit: Use DCHECK_EQ(IrOpcode::kFloat64Constant, node->opcode())

[martyn.capewell, 2016-05-04 09:47:47]

https://codereview.chromium.org/1945783002/diff/1/src/compiler/arm64/code-gen...
File src/compiler/arm64/code-generator-arm64.cc (right):

https://codereview.chromium.org/1945783002/diff/1/src/compiler/arm64/code-gen...
src/compiler/arm64/code-generator-arm64.cc:40: (copysignf(1.0f,
InputFloat32(index)) > 0.0f));
On 2016/05/03 17:33:20, Benedikt Meurer wrote:
> Use std::signbit or bit_cast<int32_t>(InputFloat32(index)) == 0

Done.

https://codereview.chromium.org/1945783002/diff/1/src/compiler/arm64/code-gen...
src/compiler/arm64/code-generator-arm64.cc:51: (copysign(1.0,
InputDouble(index)) > 0.0));
On 2016/05/03 17:33:20, Benedikt Meurer wrote:
> Use std::signbit or bit_cast<int64_t>(InputFloat64(index)) == 0

Done.

https://codereview.chromium.org/1945783002/diff/1/src/compiler/arm64/instruct...
File src/compiler/arm64/instruction-selector-arm64.cc (right):

https://codereview.chromium.org/1945783002/diff/1/src/compiler/arm64/instruct...
src/compiler/arm64/instruction-selector-arm64.cc:45: (copysign(1.0,
GetFloatConstantValue(node)) > 0.0))) {
On 2016/05/03 17:33:20, Benedikt Meurer wrote:
> Use std::signbit here, or even easier use
> bit_cast<int64_t>(GetFloatConstantValue(node)) == V8_INT64_C(0)

Done.

https://codereview.chromium.org/1945783002/diff/1/src/compiler/arm64/instruct...
src/compiler/arm64/instruction-selector-arm64.cc:82: DCHECK(node->opcode() ==
IrOpcode::kFloat64Constant);
On 2016/05/03 17:33:20, Benedikt Meurer wrote:
> Nit: Use DCHECK_EQ(IrOpcode::kFloat64Constant, node->opcode())

Done.

[martyn.capewell, 2016-05-04 09:54:17]

The CQ bit was checked by martyn.capewell@arm.com

[martyn.capewell, 2016-05-04 09:54:18]

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/1945783002/#ps40001
(title: "Remove superfluous DCHECK clause")

[commit-bot: I haz the power, 2016-05-04 09:54:19]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1945783002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1945783002/40001

[commit-bot: I haz the power, 2016-05-04 10:17:38]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2016-05-04 10:19:59]

Description was changed from

==========
[turbofan] ARM64: Use zr to store immediate zero

When storing an immediate integer or floating point zero, use the zero register
as the source value. This avoids the need to sometimes allocate a new register.

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

to

==========
[turbofan] ARM64: Use zr to store immediate zero

When storing an immediate integer or floating point zero, use the zero register
as the source value. This avoids the need to sometimes allocate a new register.

BUG=

Committed: https://crrev.com/0322c20d178db0b3360af0353972bc338b2d4b18
Cr-Commit-Position: refs/heads/master@{#36013}
==========

[commit-bot: I haz the power, 2016-05-04 10:20:00]

Patchset 3 (id:??) landed as
https://crrev.com/0322c20d178db0b3360af0353972bc338b2d4b18
Cr-Commit-Position: refs/heads/master@{#36013}