| Summary: | [JSC] Return DisposableCallSiteIndex when destroying GCAwareJITStubRoutineWithExceptionHandler | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Product: | WebKit | Reporter: | Yusuke Suzuki <ysuzuki> | ||||||
| Component: | New Bugs | Assignee: | Yusuke Suzuki <ysuzuki> | ||||||
| Status: | RESOLVED FIXED | ||||||||
| Severity: | Normal | CC: | annulen, ews-watchlist, gyuyoung.kim, keith_miller, mark.lam, msaboff, ryuan.choi, saam, sergio, tzagallo, webkit-bug-importer | ||||||
| Priority: | P2 | Keywords: | InRadar | ||||||
| Version: | WebKit Nightly Build | ||||||||
| Hardware: | Unspecified | ||||||||
| OS: | Unspecified | ||||||||
| Attachments: |
|
||||||||
|
Description
Yusuke Suzuki
2020-06-10 23:15:10 PDT
Created attachment 401622 [details]
Patch
Comment on attachment 401622 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=401622&action=review > Source/JavaScriptCore/dfg/DFGCodeOriginPool.h:51 > + HashSet<unsigned, WTF::IntHash<unsigned>, WTF::UnsignedWithZeroKeyHashTraits<unsigned>> m_callSiteIndexFreeList; I don't know why it is HashSet. Maybe, we can just make it Vector. Created attachment 401624 [details]
Patch
mac-debug-wk1 is recently crashing. https://results.webkit.org/?suite=layout-tests&test=js%2Fdom%2Fpromise-stack-overflow.html And it seems that crash is unrelated to this. https://build.webkit.org/results/Apple-Catalina-Release-WK1-Tests/r262781%20(6547)/js/dom/promise-stack-overflow-crash-log.txt Comment on attachment 401624 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=401624&action=review > Source/JavaScriptCore/jit/GCAwareJITStubRoutine.cpp:132 > + m_codeOriginPool = nullptr; When is this function called? Comment on attachment 401624 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=401624&action=review >> Source/JavaScriptCore/jit/GCAwareJITStubRoutine.cpp:132 >> + m_codeOriginPool = nullptr; > > When is this function called? This function is called when CodeBlock is destroying whole JIT related data. 1. CodeBlock itself is getting destroyed. 2. Destroying entire JITData via resetJITData So, we can just make it nullptr, since everything will be destroyed. Comment on attachment 401624 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=401624&action=review r=me >>> Source/JavaScriptCore/jit/GCAwareJITStubRoutine.cpp:132 >>> + m_codeOriginPool = nullptr; >> >> When is this function called? > > This function is called when CodeBlock is destroying whole JIT related data. > > 1. CodeBlock itself is getting destroyed. > 2. Destroying entire JITData via resetJITData > > So, we can just make it nullptr, since everything will be destroyed. Right. That's what I remembered, just wanted to make sure. Committed r262920: <https://trac.webkit.org/changeset/262920> All reviewed patches have been landed. Closing bug and clearing flags on attachment 401624 [details]. |