[ignition] Add lookup fast path to generated turbofan graph

[ignition] Add lookup fast path to generated turbofan graph

Adds a fast-path test and branch for the turbofan graph generated by
BytecodeGraphBuilder for dynamic local lookups.

BUG=v8:5263
Committed: https://crrev.com/d8d964baa226881480e58c13d519b018697c4511
Cr-Commit-Position: refs/heads/master@{#39832}

[Leszek Swirski, 2016-09-28 12:30:54]

The CQ bit was checked by leszeks@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-09-28 12:30:58]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-28 12:58:07]

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

[commit-bot: I haz the power, 2016-09-28 12:58:07]

Dry run: This issue passed the CQ dry run.

[Leszek Swirski, 2016-09-28 13:19:21]

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

[Michael Starzinger, 2016-09-28 13:55:21]

https://codereview.chromium.org/2378653003/diff/1/src/compiler/bytecode-graph...
File src/compiler/bytecode-graph-builder.cc (right):

https://codereview.chromium.org/2378653003/diff/1/src/compiler/bytecode-graph...
src/compiler/bytecode-graph-builder.cc:897: for (uint32_t d = 0; d < depth; d++)
{
Shouldn't this be "<=" here? IIUC we also want to do an extension check on the
actual target context itself as well. Unless I am missing something.

Also, is InterpreterAssembler::GotoIfHasContextExtensionUpToDepth suffering from
the same problem?

[Leszek Swirski, 2016-09-28 13:58:34]

On 2016/09/28 13:55:21, Michael Starzinger wrote:
>
https://codereview.chromium.org/2378653003/diff/1/src/compiler/bytecode-graph...
> File src/compiler/bytecode-graph-builder.cc (right):
> 
>
https://codereview.chromium.org/2378653003/diff/1/src/compiler/bytecode-graph...
> src/compiler/bytecode-graph-builder.cc:897: for (uint32_t d = 0; d < depth;
d++)
> {
> Shouldn't this be "<=" here? IIUC we also want to do an extension check on the
> actual target context itself as well. Unless I am missing something.
> 
> Also, is InterpreterAssembler::GotoIfHasContextExtensionUpToDepth suffering
from
> the same problem?

I'm glad you asked! In fact, I had the very same thought yesterday, but after
talking to Adam Klein we decided (well, he decided) that it's impossible for an
eval to shadow a local in the same scope, as logically any 'var' in the eval is
hoisted up to the top of the function -- so, you never need to check the
variable's own scope for context extensions which may shadow it.

[Michael Starzinger, 2016-09-28 14:22:13]

On 2016/09/28 13:58:34, Leszek Swirski wrote:
> On 2016/09/28 13:55:21, Michael Starzinger wrote:
> >
>
https://codereview.chromium.org/2378653003/diff/1/src/compiler/bytecode-graph...
> > File src/compiler/bytecode-graph-builder.cc (right):
> > 
> >
>
https://codereview.chromium.org/2378653003/diff/1/src/compiler/bytecode-graph...
> > src/compiler/bytecode-graph-builder.cc:897: for (uint32_t d = 0; d < depth;
> d++)
> > {
> > Shouldn't this be "<=" here? IIUC we also want to do an extension check on
the
> > actual target context itself as well. Unless I am missing something.
> > 
> > Also, is InterpreterAssembler::GotoIfHasContextExtensionUpToDepth suffering
> from
> > the same problem?
> 
> I'm glad you asked! In fact, I had the very same thought yesterday, but after
> talking to Adam Klein we decided (well, he decided) that it's impossible for
an
> eval to shadow a local in the same scope, as logically any 'var' in the eval
is
> hoisted up to the top of the function -- so, you never need to check the
> variable's own scope for context extensions which may shadow it.

Ack, got it, thanks for the explanation. As discussed offline, lets leave a
comment here in that regard to record our train of thought.

[Michael Starzinger, 2016-09-28 14:22:25]

LGTM otherwise.

[Leszek Swirski, 2016-09-28 14:29:11]

The CQ bit was checked by leszeks@chromium.org

[Leszek Swirski, 2016-09-28 14:29:12]

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/2378653003/#ps20001
(title: "Add explanatory comment for context search")

[commit-bot: I haz the power, 2016-09-28 14:29:14]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-28 14:55:22]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2016-09-28 14:55:42]

Description was changed from

==========
[ignition] Add lookup fast path to generated turbofan graph

Adds a fast-path test and branch for the turbofan graph generated by
BytecodeGraphBuilder for dynamic local lookups.

BUG=v8:5263
==========

to

==========
[ignition] Add lookup fast path to generated turbofan graph

Adds a fast-path test and branch for the turbofan graph generated by
BytecodeGraphBuilder for dynamic local lookups.

BUG=v8:5263

Committed: https://crrev.com/d8d964baa226881480e58c13d519b018697c4511
Cr-Commit-Position: refs/heads/master@{#39832}
==========

[commit-bot: I haz the power, 2016-09-28 14:55:43]

Patchset 2 (id:??) landed as
https://crrev.com/d8d964baa226881480e58c13d519b018697c4511
Cr-Commit-Position: refs/heads/master@{#39832}