[Webkit-unassigned] [Bug 261821] New: Resizable array buffer stress tests run out of memory on ppc64le

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Wed Sep 20 08:00:29 PDT 2023


            Bug ID: 261821
           Summary: Resizable array buffer stress tests run out of memory
                    on ppc64le
           Product: WebKit
           Version: WebKit Nightly Build
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: Normal
          Priority: P2
         Component: JavaScriptCore
          Assignee: webkit-unassigned at lists.webkit.org
          Reporter: mcatanzaro at redhat.com

Using a cloop build on ppc64le, the following JSC stress tests crash due to running out of memory:


All of the failures look like this:

Running stress/resizable-byteoffset.js.default
stress/resizable-byteoffset.js.default: Crashing because current footprint: 631242752 exceeds limit: 629145600
stress/resizable-byteoffset.js.default: test_script_452: line 2: 1830913 Aborted                 (core dumped) ( "$@" /home/jenkins/workspace/WebKit-JSC/label/ppc64le/WebKitBuild/Release/bin/jsc --useFTLJIT\=false --useFunctionDotArguments\=true --validateExceptionChecks\=true --useDollarVM\=true --maxPerThreadStackUsage\=1572864 --useFTLJIT\=true --useResizableArrayBuffer\=1 resizable-byteoffset.js )
stress/resizable-byteoffset.js.default: ERROR: Unexpected exit code: 134
FAIL: stress/resizable-byteoffset.js.default

My first instinct was to mark the tests as memoryLimited, since we commonly do this whenever a test requires a large amount of memory to run:

//@ skip if $memoryLimited

But then I realized that the tests do not look like they should require very much memory at all. E.g. the resizable-length test only creates one ArrayBuffer at a time, then resizes it up to 1024 bytes. It does this 10,000 times. I'm OK with skipping the tests, but I wonder if it's appropriate; perhaps something is wrong with garbage collection on this architecture, for example? Would any sort of debug info be helpful to decide?

You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20230920/526a475b/attachment.htm>

More information about the webkit-unassigned mailing list