[turbofan] Remove another premature optimization from ControlReducer.

[turbofan] Remove another premature optimization from ControlReducer.

We never hit the phi case for DecideCondition in practice, since a more
general optimization is already performed by typing and constant
propagation.

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

Committed: https://crrev.com/a3106d228ecaf8f307a0c6fc070d163acc9d0b86
Cr-Commit-Position: refs/heads/master@{#29102}

[Benedikt Meurer, 2015-06-18 06:46:21]

Hey Michi,

As discussed offline, this code is completely dead in practice. Please take a
look.

Thanks,
Benedikt

[Michael Starzinger, 2015-06-18 06:48:10]

LGTM.

[Jarin, 2015-06-18 06:48:53]

lgtm

[Benedikt Meurer, 2015-06-18 06:59:50]

For the record: To verify that the optimization is completely useless, I added a
CHECK(result==kUnknown); right before the return result; and run the test suite
for TurboFan, plus all TurboFan relevant benchmarks and it never triggered.

[Benedikt Meurer, 2015-06-18 06:59:55]

The CQ bit was checked by bmeurer@chromium.org

[commit-bot: I haz the power, 2015-06-18 07:00:11]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1186973005/1

[commit-bot: I haz the power, 2015-06-18 07:13:41]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-06-18 07:13:54]

Patchset 1 (id:??) landed as
https://crrev.com/a3106d228ecaf8f307a0c6fc070d163acc9d0b86
Cr-Commit-Position: refs/heads/master@{#29102}