[Turbofan] Refactor GapResolver tests in preparation for FP aliasing.

[Turbofan] Refactor GapResolver tests in preparation for FP aliasing.
- Changes tests to canonicalize FP slot/register moves, to simplify
testing any implementations that may fragment FP register moves.
- Adds code to generate correct ParallelMoves (e.g. no slot sources
of different reps overlapping.)
- Refactors test functions, so we can add manually generated tests
to current fuzzed tests.
- Adds SIMD operands, since these can be tested here now.

LOG=N
BUG=v8:4124
Committed: https://crrev.com/00ac990cc35850f36f67211faf6d8b722d7a7057
Cr-Commit-Position: refs/heads/master@{#39943}

[bbudge, 2016-09-23 19:06:33]

The CQ bit was checked by bbudge@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-09-23 19:06:41]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-23 19:56:57]

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

[commit-bot: I haz the power, 2016-09-23 19:56:59]

Dry run: Try jobs failed on following builders:
  v8_linux_arm_rel_ng on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_arm_rel_ng/builds/...)
  v8_linux_arm_rel_ng_triggered on master.tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_arm_rel_ng_trigger...)

[bbudge, 2016-09-23 21:50:11]

The CQ bit was checked by bbudge@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-09-23 21:50:21]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-23 22:24:23]

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

[commit-bot: I haz the power, 2016-09-23 22:24:24]

Dry run: This issue passed the CQ dry run.

[bbudge, 2016-09-23 22:42:38]

Description was changed from

==========
[Turbofan] Modify GapResolver tests to accept fragmented moves.
- Changes test to fragment FP slot/register moves into sub-moves, to
accept implementations that do that.
- Adds manually generated Aliasing test for specific combinations that
may not be generated by FuzzResolver.

LOG=N
BUG=v8:4124
==========

to

==========
[Turbofan] Refactor GapResolver tests in preparation for FP aliasing.
- Changes test to canonicalize FP slot/register moves, to test any
implementations that do that.
- Refactors test running, so we can add tests beyond FuzzResolver.

LOG=N
BUG=v8:4124
==========

[bbudge, 2016-09-23 22:47:09]

Description was changed from

==========
[Turbofan] Refactor GapResolver tests in preparation for FP aliasing.
- Changes test to canonicalize FP slot/register moves, to test any
implementations that do that.
- Refactors test running, so we can add tests beyond FuzzResolver.

LOG=N
BUG=v8:4124
==========

to

==========
[Turbofan] Refactor GapResolver tests in preparation for FP aliasing.
- Changes tests to canonicalize FP slot/register moves, to simplify
testing any implementations that may fragment FP register moves.
- Adds code to generate correct ParallelMoves (e.g. no slot sources
of different reps overlapping.)
- Refactors test functions, so we can add manually generated tests
to current fuzzed tests.
- Adds SIMD operands, since these can be tested here now.

LOG=N
BUG=v8:4124
==========

[bbudge, 2016-09-29 18:35:51]

The CQ bit was checked by bbudge@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-09-29 18:35:55]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-29 19:32:44]

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

[commit-bot: I haz the power, 2016-09-29 19:32:44]

Dry run: This issue passed the CQ dry run.

[bbudge, 2016-09-30 00:32:09]

The CQ bit was checked by bbudge@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-09-30 00:32:17]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[bbudge, 2016-09-30 00:45:07]

bbudge@chromium.org changed reviewers:
+ danno@chromium.org, mtrofin@chromium.org

[commit-bot: I haz the power, 2016-09-30 00:56:07]

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

[commit-bot: I haz the power, 2016-09-30 00:56:08]

Dry run: This issue passed the CQ dry run.

[Mircea Trofin, 2016-09-30 04:56:13]

lgtm

https://codereview.chromium.org/2365983002/diff/60001/test/cctest/compiler/te...
File test/cctest/compiler/test-gap-resolver.cc (right):

https://codereview.chromium.org/2365983002/diff/60001/test/cctest/compiler/te...
test/cctest/compiler/test-gap-resolver.cc:16: bool IsFPLocation(const
InstructionOperand& op) {
don't we have this defined on InstructionOperand? Or could we have it there,
instead?

[bbudge, 2016-09-30 18:16:31]

https://codereview.chromium.org/2365983002/diff/60001/test/cctest/compiler/te...
File test/cctest/compiler/test-gap-resolver.cc (right):

https://codereview.chromium.org/2365983002/diff/60001/test/cctest/compiler/te...
test/cctest/compiler/test-gap-resolver.cc:16: bool IsFPLocation(const
InstructionOperand& op) {
On 2016/09/30 04:56:13, Mircea Trofin wrote:
> don't we have this defined on InstructionOperand? Or could we have it there,
> instead?

Since the only usage is in this test, I just implemented it here. However, it is
going to be more useful when aliasing is added, so I have another CL that adds
IsFPLocation (and also IsLocation) and rearranges the types and enums to reduce
the number of comparisons that we have to do in these methods. I'll send you
that CL next.

[bbudge, 2016-10-03 22:36:53]

The CQ bit was checked by bbudge@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-03 22:37:01]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-10-03 23:04:51]

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

[commit-bot: I haz the power, 2016-10-03 23:04:53]

Dry run: This issue passed the CQ dry run.

[bbudge, 2016-10-03 23:43:35]

The CQ bit was checked by bbudge@chromium.org

[bbudge, 2016-10-03 23:43:35]

The patchset sent to the CQ was uploaded after l-g-t-m from mtrofin@chromium.org
Link to the patchset: https://codereview.chromium.org/2365983002/#ps80001
(title: "Rebase.")

[commit-bot: I haz the power, 2016-10-03 23:43:40]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[bbudge, 2016-10-03 23:45:50]

https://codereview.chromium.org/2365983002/diff/60001/test/cctest/compiler/te...
File test/cctest/compiler/test-gap-resolver.cc (right):

https://codereview.chromium.org/2365983002/diff/60001/test/cctest/compiler/te...
test/cctest/compiler/test-gap-resolver.cc:16: bool IsFPLocation(const
InstructionOperand& op) {
On 2016/09/30 18:16:31, bbudge wrote:
> On 2016/09/30 04:56:13, Mircea Trofin wrote:
> > don't we have this defined on InstructionOperand? Or could we have it there,
> > instead?
> 
> Since the only usage is in this test, I just implemented it here. However, it
is
> going to be more useful when aliasing is added, so I have another CL that adds
> IsFPLocation (and also IsLocation) and rearranges the types and enums to
reduce
> the number of comparisons that we have to do in these methods. I'll send you
> that CL next.

I merged in the IsFPLocationOperand method from my other CL.

Done.

[commit-bot: I haz the power, 2016-10-03 23:45:54]

Description was changed from

==========
[Turbofan] Refactor GapResolver tests in preparation for FP aliasing.
- Changes tests to canonicalize FP slot/register moves, to simplify
testing any implementations that may fragment FP register moves.
- Adds code to generate correct ParallelMoves (e.g. no slot sources
of different reps overlapping.)
- Refactors test functions, so we can add manually generated tests
to current fuzzed tests.
- Adds SIMD operands, since these can be tested here now.

LOG=N
BUG=v8:4124
==========

to

==========
[Turbofan] Refactor GapResolver tests in preparation for FP aliasing.
- Changes tests to canonicalize FP slot/register moves, to simplify
testing any implementations that may fragment FP register moves.
- Adds code to generate correct ParallelMoves (e.g. no slot sources
of different reps overlapping.)
- Refactors test functions, so we can add manually generated tests
to current fuzzed tests.
- Adds SIMD operands, since these can be tested here now.

LOG=N
BUG=v8:4124
==========

[commit-bot: I haz the power, 2016-10-03 23:45:55]

Committed patchset #5 (id:80001)

[commit-bot: I haz the power, 2016-10-03 23:46:12]

Description was changed from

==========
[Turbofan] Refactor GapResolver tests in preparation for FP aliasing.
- Changes tests to canonicalize FP slot/register moves, to simplify
testing any implementations that may fragment FP register moves.
- Adds code to generate correct ParallelMoves (e.g. no slot sources
of different reps overlapping.)
- Refactors test functions, so we can add manually generated tests
to current fuzzed tests.
- Adds SIMD operands, since these can be tested here now.

LOG=N
BUG=v8:4124
==========

to

==========
[Turbofan] Refactor GapResolver tests in preparation for FP aliasing.
- Changes tests to canonicalize FP slot/register moves, to simplify
testing any implementations that may fragment FP register moves.
- Adds code to generate correct ParallelMoves (e.g. no slot sources
of different reps overlapping.)
- Refactors test functions, so we can add manually generated tests
to current fuzzed tests.
- Adds SIMD operands, since these can be tested here now.

LOG=N
BUG=v8:4124

Committed: https://crrev.com/00ac990cc35850f36f67211faf6d8b722d7a7057
Cr-Commit-Position: refs/heads/master@{#39943}
==========

[commit-bot: I haz the power, 2016-10-03 23:46:13]

Patchset 5 (id:??) landed as
https://crrev.com/00ac990cc35850f36f67211faf6d8b722d7a7057
Cr-Commit-Position: refs/heads/master@{#39943}