Bug 152695 - Turn off Internal Function inlining in the DFG for super calls.
Summary: Turn off Internal Function inlining in the DFG for super calls.
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: New Bugs (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Keith Miller
URL:
Keywords:
Depends on:
Blocks: 152706
  Show dependency tree / graph
 
Reported: 2016-01-04 10:00 PST by Keith Miller
Modified: 2016-01-04 16:35 PST (History)
4 users (show)

See Also:

Attachments
Patch (4.76 KB, patch)
2016-01-04 10:11 PST, Keith Miller 		ggaren: review+
Details | Formatted Diff | Diff
Benchmark results (64.06 KB, text/plain)
2016-01-04 10:20 PST, Keith Miller 		no flags Details
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Keith Miller 2016-01-04 10:00:12 PST 
Turn off Internal Function inlining in the DFG for super calls.
Comment 1 Keith Miller 2016-01-04 10:11:59 PST 
Created attachment 268203 [details]
Patch
Comment 2 Keith Miller 2016-01-04 10:20:41 PST 
Created attachment 268204 [details]
Benchmark results
Comment 3 Geoffrey Garen 2016-01-04 11:00:28 PST 
Comment on attachment 268203 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=268203&action=review

r- because this needs a test case.

> Source/JavaScriptCore/dfg/DFGByteCodeParser.cpp:2433
> +        // FIXME: If we have inlined super calls up to the call site, however, we should be able to figure out the structure.

You should file a bug for this fixme.
Comment 4 Keith Miller 2016-01-04 11:13:34 PST 
> r- because this needs a test case.
> 

I'm not sure what behavior you want me to test. This patch just ensures that we take a slow path when subclassing an InternalFunction. There is no non-performance difference here.

> > Source/JavaScriptCore/dfg/DFGByteCodeParser.cpp:2433
> > +        // FIXME: If we have inlined super calls up to the call site, however, we should be able to figure out the structure.
> 
> You should file a bug for this fix me.

Fixed. https://bugs.webkit.org/show_bug.cgi?id=152700
Comment 5 Geoffrey Garen 2016-01-04 15:15:11 PST 
Comment on attachment 268203 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=268203&action=review

r=me

I guess the test case will come later, once the slow path starts to do the right thing.

>> Source/JavaScriptCore/dfg/DFGByteCodeParser.cpp:2433
>> +        // FIXME: If we have inlined super calls up to the call site, however, we should be able to figure out the structure.
> 
> You should file a bug for this fixme.

Please paste the bug link in the source code.

> Source/JavaScriptCore/dfg/DFGByteCodeParser.cpp:2441
> -        
> +

Revert.
Comment 6 Keith Miller 2016-01-04 16:21:01 PST 
(In reply to comment #5)
> Comment on attachment 268203 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=268203&action=review
> 
> r=me
> 
> I guess the test case will come later, once the slow path starts to do the
> right thing.
> 
> >> Source/JavaScriptCore/dfg/DFGByteCodeParser.cpp:2433
> >> +        // FIXME: If we have inlined super calls up to the call site, however, we should be able to figure out the structure.
> > 
> > You should file a bug for this fixme.
> 
> Please paste the bug link in the source code.

It's already there.

> 
> > Source/JavaScriptCore/dfg/DFGByteCodeParser.cpp:2441
> > -        
> > +
> 
> Revert.

Fixed.
Comment 7 Keith Miller 2016-01-04 16:35:30 PST 
Committed r194565: <http://trac.webkit.org/changeset/194565>