[turbofan] Proper dead code elimination as regular reducer.

[turbofan] Proper dead code elimination as regular reducer.

The three different concerns that the ControlReducer used to deal with
are now properly separated into

  a.) DeadCodeElimination, which is a regular AdvancedReducer, that
      propagates Dead via control edges,
  b.) CommonOperatorReducer, which does strength reduction on common
      operators (i.e. Branch, Phi, and friends), and
  c.) GraphTrimming, which removes dead->live edges from the graph.

This will make it possible to run the DeadCodeElimination together with
other passes that actually introduce Dead nodes, i.e. typed lowering;
and it opens the door for general inlining without two stage fix point
iteration.

To make the DeadCodeElimination easier and more uniform, we basically
reverted the introduction of DeadValue and DeadEffect, and changed the
Dead operator to produce control, value and effect. Note however that
this is not a requirement, but merely a way to make dead propagation
easier and more uniform. We could always go back and decide to have
different Dead operators if some other change requires that.

Note that there are several additional opportunities for cleanup now,
i.e. OSR deconstruction could be a regular reducer now, and we don't
need to use TheHole as dead value marker in the GraphReducer. And we can
actually run the dead code elimination together with the other passes
instead of using separate passes over the graph.  We will do this in
follow up CLs.

R=jarin@chromium.org, mstarzinger@chromium.org

Committed: https://crrev.com/733a2463865c8db539d8c71dabdb7e8a3cc8de0a
Cr-Commit-Position: refs/heads/master@{#29146}

[Benedikt Meurer, 2015-06-19 08:13:07]

Hey Michi, Jaro,

This is the key change to make dead code elimination play well with the other
reductions. It should be functionally equivalent; we can do the actual cleanup
later.
Please take a look.

Thanks,
Benedikt

[Jarin, 2015-06-19 09:50:51]

lgtm (although I do not pretend I really understand the OSR magic)

https://codereview.chromium.org/1193833002/diff/1/src/compiler/dead-code-elim...
File src/compiler/dead-code-elimination.cc (right):

https://codereview.chromium.org/1193833002/diff/1/src/compiler/dead-code-elim...
src/compiler/dead-code-elimination.cc:52: if (live_input_count == 0) {
Out of curiosity - when does this happen?

https://codereview.chromium.org/1193833002/diff/1/src/compiler/js-graph.cc
File src/compiler/js-graph.cc (right):

https://codereview.chromium.org/1193833002/diff/1/src/compiler/js-graph.cc#ne...
src/compiler/js-graph.cc:206: Node* JSGraph::DeadControl() {
Does it still make sense to call this DeadControl?

[Benedikt Meurer, 2015-06-19 09:55:31]

https://codereview.chromium.org/1193833002/diff/1/src/compiler/dead-code-elim...
File src/compiler/dead-code-elimination.cc (right):

https://codereview.chromium.org/1193833002/diff/1/src/compiler/dead-code-elim...
src/compiler/dead-code-elimination.cc:52: if (live_input_count == 0) {
I guess this can never happen. But I wanted to keep the existing functionality
of the ControlReducer. Maybe we want to replace this with a CHECK at some point?

https://codereview.chromium.org/1193833002/diff/1/src/compiler/js-graph.cc
File src/compiler/js-graph.cc (right):

https://codereview.chromium.org/1193833002/diff/1/src/compiler/js-graph.cc#ne...
src/compiler/js-graph.cc:206: Node* JSGraph::DeadControl() {
Renamed to Dead.

[Benedikt Meurer, 2015-06-19 12:05:03]

The CQ bit was checked by bmeurer@chromium.org

[Benedikt Meurer, 2015-06-19 12:05:03]

The patchset sent to the CQ was uploaded after l-g-t-m from jarin@chromium.org
Link to the patchset: https://codereview.chromium.org/1193833002/#ps20001
(title: "Renamed DeadControl to Dead.")

[commit-bot: I haz the power, 2015-06-19 12:05:14]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1193833002/20001

[commit-bot: I haz the power, 2015-06-19 12:07:24]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-06-19 12:07:36]

Patchset 2 (id:??) landed as
https://crrev.com/733a2463865c8db539d8c71dabdb7e8a3cc8de0a
Cr-Commit-Position: refs/heads/master@{#29146}

[Jarin, 2015-06-19 12:12:11]

https://codereview.chromium.org/1193833002/diff/1/src/compiler/dead-code-elim...
File src/compiler/dead-code-elimination.cc (right):

https://codereview.chromium.org/1193833002/diff/1/src/compiler/dead-code-elim...
src/compiler/dead-code-elimination.cc:52: if (live_input_count == 0) {
On 2015/06/19 09:55:31, Benedikt Meurer wrote:
> I guess this can never happen. But I wanted to keep the existing functionality
> of the ControlReducer. Maybe we want to replace this with a CHECK at some
point?

Indeed, that was my thinking...