[turbofan] Handle dead diamonds in scheduling and add a test.

[turbofan] Handle dead diamonds in scheduling and add a test.

The background here is that graphs generated from WASM are not trimmed.
That means there can be some floating control diamonds that are not
reachable from end. An assertion in the scheduler for phis from floating
diamonds checks that the use edge in this situation is the control edge,
but in general, any edge could cause this.

Scheduling still works without this assertion. The longer term fix
is to either trim the graphs (more compile time overhead for WASM)
or improve the scheduler's handling of dead code in the graph. Currently
it does not schedule dead code but the potential use positions of
dead code are used in the computation of the common dominator of uses. We could
recognize dead nodes in PrepareUses() and check in GetBlockForUse()
as per TODO.

R=bradnelson@chromium.org, mstarzinger@chromium.org
BUG=

Committed: https://crrev.com/45d75bca5c13f52850d9ac10ea58dae38fd0f1d7
Cr-Commit-Position: refs/heads/master@{#35245}

[titzer, 2016-03-31 12:03:12]

Description was changed from

==========
[turbofan] Handle dead diamonds in scheduling and add a test.

The background here is that graphs generated from WASM are not trimmed.
That means there can be some floating control diamonds that are not
reachable from end. An assertion in the scheduler for phis from floating
diamonds checks that the use edge in this situation is the control edge,
but in general, any edge could cause this.

Scheduling still works without this assertion. The longer term fix
is to either trim the graphs (more compile time overhead for WASM)
or improve the scheduler's handling of dead code in the graph. Currently
it does not schedule dead code but the potential use positions of
dead code are used in the common dominator calculation. We could
recognize dead nodes in PrepareUses() and check in GetBlockForUse()
as per TODO.

R=bradnelson@chromium.org, mstarzinger@chromium.org
BUG=
==========

to

==========
[turbofan] Handle dead diamonds in scheduling and add a test.

The background here is that graphs generated from WASM are not trimmed.
That means there can be some floating control diamonds that are not
reachable from end. An assertion in the scheduler for phis from floating
diamonds checks that the use edge in this situation is the control edge,
but in general, any edge could cause this.

Scheduling still works without this assertion. The longer term fix
is to either trim the graphs (more compile time overhead for WASM)
or improve the scheduler's handling of dead code in the graph. Currently
it does not schedule dead code but the potential use positions of
dead code are used in the computation of the common dominator of uses. We could
recognize dead nodes in PrepareUses() and check in GetBlockForUse()
as per TODO.

R=bradnelson@chromium.org, mstarzinger@chromium.org
BUG=
==========

[titzer, 2016-04-04 16:22:01]

On 2016/03/31 11:58:02, titzer wrote:

ping

[Michael Starzinger, 2016-04-04 17:07:43]

LGTM with one comment to address.

https://codereview.chromium.org/1846933002/diff/1/src/compiler/scheduler.cc
File src/compiler/scheduler.cc (left):

https://codereview.chromium.org/1846933002/diff/1/src/compiler/scheduler.cc#o...
src/compiler/scheduler.cc:1548: DCHECK_EQ(edge.to(),
NodeProperties::GetControlInput(use));
After the above TODO is being addressed, we really should bring back this
assertion again. Can you move the TODO here and just comment out the DCHECK for
now? Also the TODO should mention to re-enable this DCHECK.

[Michael Starzinger, 2016-04-04 17:09:22]

mstarzinger@chromium.org changed reviewers:
+ jarin@chromium.org

[Michael Starzinger, 2016-04-04 17:09:23]

+Jaro: FYI. As we discussed offline.

[titzer, 2016-04-04 17:17:11]

The CQ bit was checked by titzer@chromium.org

[titzer, 2016-04-04 17:17:12]

The patchset sent to the CQ was uploaded after l-g-t-m from
mstarzinger@chromium.org
Link to the patchset: https://codereview.chromium.org/1846933002/#ps40001
(title: " ")

[commit-bot: I haz the power, 2016-04-04 17:17:25]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1846933002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1846933002/40001

[commit-bot: I haz the power, 2016-04-04 17:44:37]

Description was changed from

==========
[turbofan] Handle dead diamonds in scheduling and add a test.

The background here is that graphs generated from WASM are not trimmed.
That means there can be some floating control diamonds that are not
reachable from end. An assertion in the scheduler for phis from floating
diamonds checks that the use edge in this situation is the control edge,
but in general, any edge could cause this.

Scheduling still works without this assertion. The longer term fix
is to either trim the graphs (more compile time overhead for WASM)
or improve the scheduler's handling of dead code in the graph. Currently
it does not schedule dead code but the potential use positions of
dead code are used in the computation of the common dominator of uses. We could
recognize dead nodes in PrepareUses() and check in GetBlockForUse()
as per TODO.

R=bradnelson@chromium.org, mstarzinger@chromium.org
BUG=
==========

to

==========
[turbofan] Handle dead diamonds in scheduling and add a test.

The background here is that graphs generated from WASM are not trimmed.
That means there can be some floating control diamonds that are not
reachable from end. An assertion in the scheduler for phis from floating
diamonds checks that the use edge in this situation is the control edge,
but in general, any edge could cause this.

Scheduling still works without this assertion. The longer term fix
is to either trim the graphs (more compile time overhead for WASM)
or improve the scheduler's handling of dead code in the graph. Currently
it does not schedule dead code but the potential use positions of
dead code are used in the computation of the common dominator of uses. We could
recognize dead nodes in PrepareUses() and check in GetBlockForUse()
as per TODO.

R=bradnelson@chromium.org, mstarzinger@chromium.org
BUG=
==========

[commit-bot: I haz the power, 2016-04-04 17:44:38]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2016-04-04 17:46:09]

Description was changed from

==========
[turbofan] Handle dead diamonds in scheduling and add a test.

The background here is that graphs generated from WASM are not trimmed.
That means there can be some floating control diamonds that are not
reachable from end. An assertion in the scheduler for phis from floating
diamonds checks that the use edge in this situation is the control edge,
but in general, any edge could cause this.

Scheduling still works without this assertion. The longer term fix
is to either trim the graphs (more compile time overhead for WASM)
or improve the scheduler's handling of dead code in the graph. Currently
it does not schedule dead code but the potential use positions of
dead code are used in the computation of the common dominator of uses. We could
recognize dead nodes in PrepareUses() and check in GetBlockForUse()
as per TODO.

R=bradnelson@chromium.org, mstarzinger@chromium.org
BUG=
==========

to

==========
[turbofan] Handle dead diamonds in scheduling and add a test.

The background here is that graphs generated from WASM are not trimmed.
That means there can be some floating control diamonds that are not
reachable from end. An assertion in the scheduler for phis from floating
diamonds checks that the use edge in this situation is the control edge,
but in general, any edge could cause this.

Scheduling still works without this assertion. The longer term fix
is to either trim the graphs (more compile time overhead for WASM)
or improve the scheduler's handling of dead code in the graph. Currently
it does not schedule dead code but the potential use positions of
dead code are used in the computation of the common dominator of uses. We could
recognize dead nodes in PrepareUses() and check in GetBlockForUse()
as per TODO.

R=bradnelson@chromium.org, mstarzinger@chromium.org
BUG=

Committed: https://crrev.com/45d75bca5c13f52850d9ac10ea58dae38fd0f1d7
Cr-Commit-Position: refs/heads/master@{#35245}
==========

[commit-bot: I haz the power, 2016-04-04 17:46:10]

Patchset 3 (id:??) landed as
https://crrev.com/45d75bca5c13f52850d9ac10ea58dae38fd0f1d7
Cr-Commit-Position: refs/heads/master@{#35245}