Bug 142229 – DFG IR should refer to FunctionExecutables directly and not via the CodeBlock

• Component: JavaScriptCore
• Status: RESOLVED FIXED

[Filip Pizlo]

This will make function-creating operations more inlineable and much simpler to reason about.

[Filip Pizlo]

Created attachment 247783 [details]
the patch

[Mark Lam]

Comment on attachment 247783 [details]
the patch

View in context: https://bugs.webkit.org/attachment.cgi?id=247783&action=review

r=me

> Source/JavaScriptCore/ChangeLog:20
> +        For this reason, those operations were not inlineable. But the functin tables in CodeBlock

s/functin/function/

[Build Bot]

Comment on attachment 247783 [details]
the patch

Attachment 247783 [details] did not pass mac-ews (mac):
Output: http://webkit-queues.appspot.com/results/6094596501143552

New failing tests:
webgl/1.0.2/conformance/more/conformance/quickCheckAPI-L_S.html

[Build Bot]

Created attachment 247792 [details]
Archive of layout-test-results from ews101 for mac-mavericks

The attached test failures were seen while running run-webkit-tests on the mac-ews.
Bot: ews101  Port: mac-mavericks  Platform: Mac OS X 10.9.5

[Build Bot]

Comment on attachment 247783 [details]
the patch

Attachment 247783 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://webkit-queues.appspot.com/results/5470569965289472

New failing tests:
webgl/1.0.2/conformance/more/conformance/quickCheckAPI-L_S.html

[Build Bot]

Created attachment 247793 [details]
Archive of layout-test-results from ews105 for mac-mavericks-wk2

The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews.
Bot: ews105  Port: mac-mavericks-wk2  Platform: Mac OS X 10.9.5

[Filip Pizlo]

(In reply to comment #6)
> Created attachment 247793 [details]
> Archive of layout-test-results from ews105 for mac-mavericks-wk2
> 
> The attached test failures were seen while running run-webkit-tests on the
> mac-wk2-ews.
> Bot: ews105  Port: mac-mavericks-wk2  Platform: Mac OS X 10.9.5

I would have had these fixed by now if it wasn't for https://bugs.webkit.org/show_bug.cgi?id=142234.

[Filip Pizlo]

It turns out that enabling inlining of NewFunctionExpr causes our handling of scopeRegister in inlined functions to go bananas.  We currently constant fold *accesses* to the scopeRegister, which is badly wrong - it makes us lose the OSR exit liveness information.

To fix this, I'm doing: https://bugs.webkit.org/show_bug.cgi?id=106202

[Filip Pizlo]

I fixed the bugs.  I will land.

[Filip Pizlo]

Landed in http://trac.webkit.org/changeset/180993