ARM64: Slightly simplify LShiftI and LShiftS.

ARM64: Slightly simplify LShiftI and LShiftS.

BUG=
R=ulan@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=23277

[ulan, 2014-08-21 11:13:04]

lgtm

https://codereview.chromium.org/494053002/diff/1/src/arm64/lithium-codegen-ar...
File src/arm64/lithium-codegen-arm64.cc (right):

https://codereview.chromium.org/494053002/diff/1/src/arm64/lithium-codegen-ar...
src/arm64/lithium-codegen-arm64.cc:4941: UseScratchRegisterScope temps(masm());
Didn't we want to reduce usage of scratch registers in lithium? :) One the other
hand, it's nice to safe one register for SAR/SHL/SHR. Not sure if this an
improvement, but don't mind landing this.

[jbramley, 2014-08-21 13:02:12]

On 2014/08/21 11:13:04, ulan wrote:
> lgtm
> 
>
https://codereview.chromium.org/494053002/diff/1/src/arm64/lithium-codegen-ar...
> File src/arm64/lithium-codegen-arm64.cc (right):
> 
>
https://codereview.chromium.org/494053002/diff/1/src/arm64/lithium-codegen-ar...
> src/arm64/lithium-codegen-arm64.cc:4941: UseScratchRegisterScope
temps(masm());
> Didn't we want to reduce usage of scratch registers in lithium? :) One the
other
> hand, it's nice to safe one register for SAR/SHL/SHR. Not sure if this an
> improvement, but don't mind landing this.

Indeed! We've debated it quite a bit over the last year or so. Since the
introduction of UseScratchRegisterScope, it's a lot safer, since exhausting the
scratch register pool will trigger fatal assertions (even in release mode). We
aren't exactly short of registers in the arm64 back-end, but there are other
Lithium instructions where we use UseScratchRegisterScope for special cases,
rather than allocate a dedicated scratch register (or duplicate the selection
logic).

[jbramley, 2014-08-21 13:22:33]

Committed patchset #1 manually as 23277 (presubmit successful).