Bug 181999 - DFG should always flush `this`
Summary: DFG should always flush `this`
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: JavaScriptCore (show other bugs)
Version: WebKit Nightly Build
Hardware: All All
: P2 Normal
Assignee: Filip Pizlo
URL:
Keywords: InRadar
Depends on:
Blocks:
 
Reported: 2018-01-23 11:54 PST by Filip Pizlo
Modified: 2018-01-23 12:17 PST (History)
6 users (show)

See Also:

Attachments
the patch (10.69 KB, patch)
2018-01-23 11:58 PST, Filip Pizlo 		saam: review+
Details | Formatted Diff | Diff
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 Filip Pizlo 2018-01-23 11:54:32 PST 
Patch forthcoming.
Comment 1 Filip Pizlo 2018-01-23 11:58:38 PST 
Created attachment 332053 [details]
the patch
Comment 2 Saam Barati 2018-01-23 12:06:44 PST 
Comment on attachment 332053 [details]
the patch

View in context: https://bugs.webkit.org/attachment.cgi?id=332053&action=review

> Source/JavaScriptCore/dfg/DFGByteCodeParser.cpp:539
> +        addFlushDirect(remapOperand(inlineCallFrame, virtualRegisterForArgument(0)));

This should just go in the above loop and make it loop over the zero index
Comment 3 Mark Lam 2018-01-23 12:09:46 PST 
Comment on attachment 332053 [details]
the patch

View in context: https://bugs.webkit.org/attachment.cgi?id=332053&action=review

r=me with suggested simplification.

> Source/JavaScriptCore/dfg/DFGByteCodeParser.cpp:539
> +        for (unsigned argument = numArguments; argument-- > 1;)
> +            addFlushDirect(remapOperand(inlineCallFrame, virtualRegisterForArgument(argument)));
> +
> +        addFlushDirect(remapOperand(inlineCallFrame, virtualRegisterForArgument(0)));

Why not just do all the flushing in the for loop?

    for (unsigned argument = numArguments; argument--;)
        addFlushDirect(remapOperand(inlineCallFrame, virtualRegisterForArgument(argument)));
Comment 4 Filip Pizlo 2018-01-23 12:14:51 PST 
(In reply to Saam Barati from comment #2)
> Comment on attachment 332053 [details]
> the patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=332053&action=review
> 
> > Source/JavaScriptCore/dfg/DFGByteCodeParser.cpp:539
> > +        addFlushDirect(remapOperand(inlineCallFrame, virtualRegisterForArgument(0)));
> 
> This should just go in the above loop and make it loop over the zero index

Oh. Right.
Comment 5 Radar WebKit Bug Importer 2018-01-23 12:15:10 PST 
<rdar://problem/36787959>
Comment 6 Filip Pizlo 2018-01-23 12:17:18 PST 
Landed in https://trac.webkit.org/changeset/227431/webkit