Bug 138254 – Update scope related slow path code to use scope register added to opcodes

• Status: RESOLVED
• Resolution: FIXED
• Component: JavaScriptCore

[Michael Saboff]

Update the slow paths to use the scope operands added in <https://bugs.webkit.org/show_bug.cgi?id=138252> and <https://bugs.webkit.org/show_bug.cgi?id=138253>

[Michael Saboff]

Created attachment 240892 [details]
Patch

[Mark Lam]

Comment on attachment 240892 [details]
Patch

r=me

[Michael Saboff]

Committed r175509: <http://trac.webkit.org/changeset/175509>

[Chris Dumez]

Some bots seems to have trouble building after this change:
https://build.webkit.org/builders/EFL%20Linux%20ARMv7%20Thumb2%20Release%20%28Build%29/builds/8038/steps/compile-webkit/logs/stdio

[WebKit Commit Bot]

Re-opened since this is blocked by bug 138349

[Csaba Osztrogonác]

Created attachment 240917 [details]
buildfix for ARMv7

This patch fixed the ARMv7 Thumb2 Linux build for me, but got some regressions.
I'll upload the detailed results soon. (tests still running)

[Csaba Osztrogonác]

Created attachment 240918 [details]
JSC test log from ARMv7 Thumb2

[Michael Saboff]

(In reply to comment #7)
> Created attachment 240918 [details]
> JSC test log from ARMv7 Thumb2

Thanks for the build fix.  I added it to the original patch and ran tests on ARMv7 Thumb2.  I do not see the failures you are seeing.  I'm going to post the combined patch.

Could you clean build, rerun test and report back with the results?

[Michael Saboff]

Created attachment 240946 [details]
Patch updated with supplied build fixes.

Will run through the EWS bots before landing.

[Csaba Osztrogonác]

(In reply to comment #8)
> (In reply to comment #7)
> > Created attachment 240918 [details]
> > JSC test log from ARMv7 Thumb2
> 
> Thanks for the build fix.  I added it to the original patch and ran tests on
> ARMv7 Thumb2.  I do not see the failures you are seeing.  I'm going to post
> the combined patch.
> 
> Could you clean build, rerun test and report back with the results?

I tried it, the build works fine, but unfortunately the failures 
is still present on Linux with Thumb2 and ARM instruction set too.

[Csaba Osztrogonác]

Created attachment 241015 [details]
minimal example to debug

$ ./jsc --useLLInt=false --useDFGJIT=false -f foo.js
1.9097962116e-313

expected result: foo_string

[Csaba Osztrogonác]

Created attachment 241017 [details]
log on r175612 without patch

$ ./jsc --useLLInt=false --useDFGJIT=false --dumpGeneratedBytecodes=true --showDisassembly=true -f foo.js

[Csaba Osztrogonác]

Created attachment 241018 [details]
log on r175612 with patch applied

[Michael Saboff]

(In reply to comment #13)
> Created attachment 241018 [details]
> log on r175612 with patch applied

I'll look at the generated assembly today to see what is going on.

[Dániel Bátyai]

(In reply to comment #14)
> (In reply to comment #13)
> > Created attachment 241018 [details]
> > log on r175612 with patch applied
> 
> I'll look at the generated assembly today to see what is going on.

Hi, we have found the issue already, patch is on the way.

[Csaba Osztrogonác]

(In reply to comment #15)
> (In reply to comment #14)
> > (In reply to comment #13)
> > > Created attachment 241018 [details]
> > > log on r175612 with patch applied
> > 
> > I'll look at the generated assembly today to see what is going on.
> 
> Hi, we have found the issue already, patch is on the way.

Dániel already fixed the issue in bug138510. It's safe 
to land this one and then the fix from bug138510.

[Mark Lam]

Comment on attachment 240946 [details]
Patch updated with supplied build fixes.

r=me

[Michael Saboff]

Committed r175762: <http://trac.webkit.org/changeset/175762>