[turbofan] Add support for SOFT deopts and use that for property access.

[turbofan] Add support for SOFT deopts and use that for property access.

Up until now we were unable to (re)optimize code when we hit
uninitialized (Keyed)Load/StoreICs in the code. We always put an IC
there (sharing the feedback vector with fullcodegen at least) and called
it a day. But we never deoptimized the code object when we gathered more
feedback. This doesn't work very well in practice, esp. with hot code
relying on this. So until we have a proper mechanism to express the need
to reoptimize after we gathered additional feedback from optimized code,
we follow the Crankshaft approach instead and install a SOFT deopt, so
we can not only learn but also utilize the new feedback.

R=mstarzinger@chromium.org
BUG=v8:4470
LOG=n

Committed: https://crrev.com/bda527b5ffd58b05ad6e6202c3d9b1632a20a4c8
Cr-Commit-Position: refs/heads/master@{#34178}

[mvstanton, 2015-12-11 14:26:02]

mvstanton@chromium.org changed reviewers:
+ mvstanton@chromium.org

[mvstanton, 2015-12-11 14:26:03]

LGTM holds, as long as test fix isn't more than a couple lines :p

[Michael Starzinger, 2015-12-11 14:44:23]

LGTM.

[Benedikt Meurer, 2016-02-20 19:04:06]

The CQ bit was checked by bmeurer@chromium.org

[Benedikt Meurer, 2016-02-20 19:04:06]

The patchset sent to the CQ was uploaded after l-g-t-m from
mvstanton@chromium.org, mstarzinger@chromium.org
Link to the patchset: https://codereview.chromium.org/1518013002/#ps40001
(title: "Simplification")

[commit-bot: I haz the power, 2016-02-20 19:04:13]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1518013002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1518013002/40001

[commit-bot: I haz the power, 2016-02-20 19:05:53]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2016-02-20 19:06:23]

Description was changed from

==========
[turbofan] Add support for SOFT deopts and use that for property access.

Up until now we were unable to (re)optimize code when we hit
uninitialized (Keyed)Load/StoreICs in the code. We always put an IC
there (sharing the feedback vector with fullcodegen at least) and called
it a day. But we never deoptimized the code object when we gathered more
feedback. This doesn't work very well in practice, esp. with hot code
relying on this. So until we have a proper mechanism to express the need
to reoptimize after we gathered additional feedback from optimized code,
we follow the Crankshaft approach instead and install a SOFT deopt, so
we can not only learn but also utilize the new feedback.

R=mstarzinger@chromium.org
BUG=v8:4470
LOG=n
==========

to

==========
[turbofan] Add support for SOFT deopts and use that for property access.

Up until now we were unable to (re)optimize code when we hit
uninitialized (Keyed)Load/StoreICs in the code. We always put an IC
there (sharing the feedback vector with fullcodegen at least) and called
it a day. But we never deoptimized the code object when we gathered more
feedback. This doesn't work very well in practice, esp. with hot code
relying on this. So until we have a proper mechanism to express the need
to reoptimize after we gathered additional feedback from optimized code,
we follow the Crankshaft approach instead and install a SOFT deopt, so
we can not only learn but also utilize the new feedback.

R=mstarzinger@chromium.org
BUG=v8:4470
LOG=n

Committed: https://crrev.com/bda527b5ffd58b05ad6e6202c3d9b1632a20a4c8
Cr-Commit-Position: refs/heads/master@{#34178}
==========

[commit-bot: I haz the power, 2016-02-20 19:06:24]

Patchset 3 (id:??) landed as
https://crrev.com/bda527b5ffd58b05ad6e6202c3d9b1632a20a4c8
Cr-Commit-Position: refs/heads/master@{#34178}