Bug 127201 - CStack Branch: Fix Baseline JIT for X86-32
Summary: CStack Branch: Fix Baseline JIT for X86-32
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: JavaScriptCore (show other bugs)
Version: 528+ (Nightly build)
Hardware: All All
: P2 Normal
Assignee: Michael Saboff
URL:
Keywords:
Depends on: 127071
Blocks: 127205
  Show dependency tree / graph
 
Reported: 2014-01-17 15:46 PST by Michael Saboff
Modified: 2014-01-17 21:06 PST (History)
0 users

See Also:

Attachments
Patch (21.52 KB, patch)
2014-01-17 16:20 PST, Michael Saboff 		ggaren: 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 Michael Saboff 2014-01-17 15:46:13 PST 
This is to get the baseline JIT working again.
Comment 1 Michael Saboff 2014-01-17 16:20:22 PST 
Created attachment 221508 [details]
Patch
Comment 2 Geoffrey Garen 2014-01-17 16:39:42 PST 
Comment on attachment 221508 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=221508&action=review

r=me

> Source/JavaScriptCore/ChangeLog:9
> +        epilogues) and making room on the stack for outgoiing arguments.

Should be "outgoing".

> Source/JavaScriptCore/jit/ThunkGenerators.cpp:523
> +    // Install the new return PC.
> +    // FIXME: I don't think currentReturnThunkPC is used and should be deleted.
> +#  if 0
> +    jit.loadPtr(&vm->currentReturnThunkPC, GPRInfo::regT2);
> +    jit.storePtr(GPRInfo::regT2, JSInterfaceJIT::Address(JSInterfaceJIT::callFrameRegister, CallFrame::returnPCOffset()));
> +#   else
> +    jit.storePtr(GPRInfo::regT5, JSInterfaceJIT::Address(JSInterfaceJIT::callFrameRegister, CallFrame::returnPCOffset()));
> +#   endif

Need a bug about this.
Comment 3 Michael Saboff 2014-01-17 16:44:55 PST 
(In reply to comment #2)
> (From update of attachment 221508 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=221508&action=review
> 
> r=me
> 
> > Source/JavaScriptCore/ChangeLog:9
> > +        epilogues) and making room on the stack for outgoiing arguments.
> 
> Should be "outgoing".

Fixed.
 
> > Source/JavaScriptCore/jit/ThunkGenerators.cpp:523
> > +    // Install the new return PC.
> > +    // FIXME: I don't think currentReturnThunkPC is used and should be deleted.
> > +#  if 0
> > +    jit.loadPtr(&vm->currentReturnThunkPC, GPRInfo::regT2);
> > +    jit.storePtr(GPRInfo::regT2, JSInterfaceJIT::Address(JSInterfaceJIT::callFrameRegister, CallFrame::returnPCOffset()));
> > +#   else
> > +    jit.storePtr(GPRInfo::regT5, JSInterfaceJIT::Address(JSInterfaceJIT::callFrameRegister, CallFrame::returnPCOffset()));
> > +#   endif
> 
> Need a bug about this.

Added <https://bugs.webkit.org/show_bug.cgi?id=127205> - CStack Branch: VM::currentReturnThunkPC appears to be unused and should be removed
Comment 4 Michael Saboff 2014-01-17 21:06:05 PST 
Committed r162242: <http://trac.webkit.org/changeset/162242>