Bug 204405 - Flaky JSC test: stress/stack-overflow-in-yarr-byteCompile.js.dfg-eager
Summary: Flaky JSC test: stress/stack-overflow-in-yarr-byteCompile.js.dfg-eager
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: New Bugs (show other bugs)
Version: Other
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Mark Lam
URL:
Keywords: InRadar
Depends on:
Blocks:
 
Reported: 2019-11-20 03:35 PST by Aakash Jain
Modified: 2019-11-22 10:36 PST (History)
10 users (show)

See Also:

Attachments
proposed patch. (2.56 KB, patch)
2019-11-20 09:15 PST, Mark Lam 		ap: 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 Aakash Jain 2019-11-20 03:35:37 PST 
stress/stack-overflow-in-yarr-byteCompile.js.dfg-eager seems flaky. In https://ews-build.webkit.org/#/builders/1/builds/328, the test failed in jscore-test step. However, in the immediately next retry step (jscore-test-rerun), it passed.
Comment 1 Aakash Jain 2019-11-20 03:49:46 PST 
Another instance of flakiness: https://ews-build.webkit.org/#/builders/1/builds/256
Comment 2 Aakash Jain 2019-11-20 03:50:12 PST 
This test was added recently in https://trac.webkit.org/changeset/252239/webkit
Comment 3 Mark Lam 2019-11-20 09:15:51 PST 
Created attachment 383966 [details]
proposed patch.
Comment 4 Alexey Proskuryakov 2019-11-20 10:31:06 PST 
Comment on attachment 383966 [details]
proposed patch.

View in context: https://bugs.webkit.org/attachment.cgi?id=383966&action=review

> Source/JavaScriptCore/ChangeLog:9
> +        $vm.allWithStackSize() manipulates the stack in ways that will freak out ASan.
> +        So, add SUPPRESS_ASAN to functionCallWithStackSize() to tell ASan to ignore it.

Was there a separate bug where this was reported?
Comment 5 Mark Lam 2019-11-20 10:47:45 PST 
(In reply to Alexey Proskuryakov from comment #4)
> Comment on attachment 383966 [details]
> proposed patch.
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=383966&action=review
> 
> > Source/JavaScriptCore/ChangeLog:9
> > +        $vm.allWithStackSize() manipulates the stack in ways that will freak out ASan.
> > +        So, add SUPPRESS_ASAN to functionCallWithStackSize() to tell ASan to ignore it.
> 
> Was there a separate bug where this was reported?

Nope.  I just happened to run an ASan build when I was trying to reproduce this issue, and noticed the ASan issue.  Thought I might as well fix it since it can be a source of failures for this test.
Comment 6 Mark Lam 2019-11-20 10:51:55 PST 
Thanks for the review.  Landed in r252694: <http://trac.webkit.org/r252694>.
Comment 7 Radar WebKit Bug Importer 2019-11-20 10:52:31 PST 
<rdar://problem/57365049>
Comment 8 Radar WebKit Bug Importer 2019-11-20 10:52:32 PST 
<rdar://problem/57365050>
Comment 9 Mark Lam 2019-11-20 10:54:00 PST 
<rdar://problem/57365049>
Comment 10 Michael Saboff 2019-11-20 13:21:16 PST 
Comment on attachment 383966 [details]
proposed patch.

lgtm
Comment 11 Guillaume Emont 2019-11-22 10:36:32 PST 
Comment on attachment 383966 [details]
proposed patch.

View in context: https://bugs.webkit.org/attachment.cgi?id=383966&action=review

> JSTests/stress/stack-overflow-in-yarr-byteCompile.js:2
> +//@ runDefault("--disableOptionsFreezingForTesting")
>  //@ skip if $hostOS != "darwin" or $architecture != "x86-64"

If runDefault() is before the skip, then it is always run. I corrected that in Bug 204522.