MIPS64: Unify and improve Word32 compares to use same instructions as Word64 compares.

MIPS64: Unify and improve Word32 compares to use same instructions as Word64 compares.

The CL enables the same instructions are selected for Word32 and Word64 compare
operations which is possible due to a fact 32-bit inputs and produced values
are always sign-extended.

TEST=
BUG=

Committed: https://crrev.com/17ada20c17d548a2a98f55ba7b16a00978679d48
Cr-Commit-Position: refs/heads/master@{#27212}

[dusmil.imgtec, 2015-03-13 15:05:53]

dusan.milosavljevic@imgtec.com changed reviewers:
+ akos.palfi@imgtec.com, balazs.kilvady@imgtec.com, bmeurer@chromium.org,
danno@chromium.org, jkummerow@chromium.org, paul.lind@imgtec.com

[paul.l..., 2015-03-13 17:05:31]

LGTM.

https://codereview.chromium.org/1005123002/diff/1/src/compiler/mips64/code-ge...
File src/compiler/mips64/code-generator-mips64.cc (right):

https://codereview.chromium.org/1005123002/diff/1/src/compiler/mips64/code-ge...
src/compiler/mips64/code-generator-mips64.cc:825: // emit mips
psuedo-instructions, which are handled here by branch
nit: Here is another place I misspelled pseudo, please fix this one as well.

[paul.l..., 2015-03-13 17:08:06]

Also, this has a small change to cctest/compiler/call-tester.h that I cannot
approve. Can someone from v8 team PTAL?

[Benedikt Meurer, 2015-03-13 18:24:08]

lgtm

[dusmil.imgtec, 2015-03-16 10:32:32]

The CQ bit was checked by dusan.milosavljevic@imgtec.com

[commit-bot: I haz the power, 2015-03-16 10:32:45]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1005123002/1

[commit-bot: I haz the power, 2015-03-16 11:00:05]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-03-16 11:00:18]

Patchset 1 (id:??) landed as
https://crrev.com/17ada20c17d548a2a98f55ba7b16a00978679d48
Cr-Commit-Position: refs/heads/master@{#27212}

[titzer, 2015-03-24 14:37:21]

titzer@chromium.org changed reviewers:
+ titzer@chromium.org

[titzer, 2015-03-24 14:37:22]

https://codereview.chromium.org/1005123002/diff/1/test/cctest/compiler/call-t...
File test/cctest/compiler/call-tester.h (right):

https://codereview.chromium.org/1005123002/diff/1/test/cctest/compiler/call-t...
test/cctest/compiler/call-tester.h:131: #if V8_TARGET_ARCH_MIPS64
This should be good enough for all architectures, yes?

[dusmil.imgtec, 2015-03-24 18:41:12]

https://codereview.chromium.org/1005123002/diff/1/test/cctest/compiler/call-t...
File test/cctest/compiler/call-tester.h (right):

https://codereview.chromium.org/1005123002/diff/1/test/cctest/compiler/call-t...
test/cctest/compiler/call-tester.h:131: #if V8_TARGET_ARCH_MIPS64
On 2015/03/24 14:37:22, titzer wrote:
> This should be good enough for all architectures, yes?
Yes, after all this only concerns simulated architectures where host code and
generated code are different.  Particularly this solves problem for mips64
simulator builds where transition from x64 code to mips64 generated code is
done.

[titzer, 2015-03-24 18:43:33]

https://codereview.chromium.org/1005123002/diff/1/test/cctest/compiler/call-t...
File test/cctest/compiler/call-tester.h (right):

https://codereview.chromium.org/1005123002/diff/1/test/cctest/compiler/call-t...
test/cctest/compiler/call-tester.h:131: #if V8_TARGET_ARCH_MIPS64
On 2015/03/24 18:41:12, dusmil.imgtec wrote:
> On 2015/03/24 14:37:22, titzer wrote:
> > This should be good enough for all architectures, yes?
> Yes, after all this only concerns simulated architectures where host code and
> generated code are different.  Particularly this solves problem for mips64
> simulator builds where transition from x64 code to mips64 generated code is
> done.

If you remove the ifdef and it works on all platforms, I'll rubber-stamp it
tomorrow.

Sorry for the late comment, forgot to hit the button before it landed.