[Turbofan] Fix perf regression

[Turbofan] Fix perf regression introduced by per-range change.

When the range ends just at the gap of a non-deferred block, the last
instruction the range covers is in the predecessor. If that predecessor is
a deferred block, before this CL, we would splinter the remainder of the
range all the way to the end. That leads to inefficient codegen, because
we still want a split inside the deferred block.

Also, opportunistically added a trace before we splinter, for better
diagnostics.

BUG=chromium:546416
LOG=N

Committed: https://crrev.com/32b6e085e74a8fcf94a01d20740fe4fdede07a86
Cr-Commit-Position: refs/heads/master@{#31529}

[Mircea Trofin, 2015-10-23 05:39:36]

Description was changed from

==========
[Turbofan] Fix perf regression


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

to

==========
[Turbofan] Fix perf regression introduced by per-range change.

When the range ends just at the gap of a non-deferred block, the last 
instruction the range covers is in the predecessor. If that predecessor is 
a deferred block, before this CL, we would splinter the remainder of the 
range all the way to the end. That leads to inefficient codegen, because
we still want a split inside the deferred block.

BUG= chromium:546416
==========

[Mircea Trofin, 2015-10-23 05:39:37]

mtrofin@chromium.org changed reviewers:
+ bmeurer@chromium.org, jarin@chromium.org

[Mircea Trofin, 2015-10-23 05:41:33]

Description was changed from

==========
[Turbofan] Fix perf regression introduced by per-range change.

When the range ends just at the gap of a non-deferred block, the last 
instruction the range covers is in the predecessor. If that predecessor is 
a deferred block, before this CL, we would splinter the remainder of the 
range all the way to the end. That leads to inefficient codegen, because
we still want a split inside the deferred block.

BUG= chromium:546416
==========

to

==========
[Turbofan] Fix perf regression introduced by per-range change.

When the range ends just at the gap of a non-deferred block, the last 
instruction the range covers is in the predecessor. If that predecessor is 
a deferred block, before this CL, we would splinter the remainder of the 
range all the way to the end. That leads to inefficient codegen, because
we still want a split inside the deferred block.

Also, opportunistically added a trace before we splinter, for better
diagnostics.

BUG= chromium:546416
==========

[Benedikt Meurer, 2015-10-23 09:14:05]

LGTM

[Mircea Trofin, 2015-10-23 15:11:15]

The CQ bit was checked by mtrofin@chromium.org

[commit-bot: I haz the power, 2015-10-23 15:11:23]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1412123009/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1412123009/1

[commit-bot: I haz the power, 2015-10-23 15:13:28]

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

[commit-bot: I haz the power, 2015-10-23 15:13:29]

Try jobs failed on following builders:
  v8_presubmit on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_presubmit/builds/7028)

[Mircea Trofin, 2015-10-23 15:45:31]

Description was changed from

==========
[Turbofan] Fix perf regression introduced by per-range change.

When the range ends just at the gap of a non-deferred block, the last 
instruction the range covers is in the predecessor. If that predecessor is 
a deferred block, before this CL, we would splinter the remainder of the 
range all the way to the end. That leads to inefficient codegen, because
we still want a split inside the deferred block.

Also, opportunistically added a trace before we splinter, for better
diagnostics.

BUG= chromium:546416
==========

to

==========
[Turbofan] Fix perf regression introduced by per-range change.

When the range ends just at the gap of a non-deferred block, the last 
instruction the range covers is in the predecessor. If that predecessor is 
a deferred block, before this CL, we would splinter the remainder of the 
range all the way to the end. That leads to inefficient codegen, because
we still want a split inside the deferred block.

Also, opportunistically added a trace before we splinter, for better
diagnostics.

BUG= chromium:546416
LOG=N
==========

[Mircea Trofin, 2015-10-23 15:45:36]

The CQ bit was checked by mtrofin@chromium.org

[commit-bot: I haz the power, 2015-10-23 15:45:42]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1412123009/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1412123009/1

[commit-bot: I haz the power, 2015-10-23 15:46:48]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-10-23 15:47:26]

Patchset 1 (id:??) landed as
https://crrev.com/32b6e085e74a8fcf94a01d20740fe4fdede07a86
Cr-Commit-Position: refs/heads/master@{#31529}

[Michael Achenbach, 2015-10-23 17:39:55]

A revert of this CL (patchset #1 id:1) has been created in
https://codereview.chromium.org/1412893007/ by machenbach@chromium.org.

The reason for reverting is: [sheriff] breaks benchmarks:
http://build.chromium.org/p/client.v8/builders/V8%20Linux%20-%20debug/builds/....