Summary: | Terminate WebProcesses if GPUProcess crashes more than twice in 30 seconds | ||||||||
---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Chris Dumez <cdumez> | ||||||
Component: | WebKit2 | Assignee: | Chris Dumez <cdumez> | ||||||
Status: | RESOLVED FIXED | ||||||||
Severity: | Normal | CC: | darin, ggaren, sam, simon.fraser, thorton, webkit-bug-importer | ||||||
Priority: | P2 | Keywords: | InRadar | ||||||
Version: | WebKit Nightly Build | ||||||||
Hardware: | Unspecified | ||||||||
OS: | Unspecified | ||||||||
See Also: |
https://bugs.webkit.org/show_bug.cgi?id=218856 https://bugs.webkit.org/show_bug.cgi?id=219719 |
||||||||
Bug Depends on: | 218806 | ||||||||
Bug Blocks: | |||||||||
Attachments: |
|
Description
Chris Dumez
2020-11-11 10:53:22 PST
Created attachment 413856 [details]
Patch
Comment on attachment 413856 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=413856&action=review > Source/WebKit/UIProcess/AuxiliaryProcessProxy.cpp:111 > + RELEASE_LOG_ERROR(Process, "AuxiliaryProcessProxy::terminate"); Can this log the pid of the thing being terminated? (In reply to Simon Fraser (smfr) from comment #3) > Comment on attachment 413856 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=413856&action=review > > > Source/WebKit/UIProcess/AuxiliaryProcessProxy.cpp:111 > > + RELEASE_LOG_ERROR(Process, "AuxiliaryProcessProxy::terminate"); > > Can this log the pid of the thing being terminated? good idea. Will fix. Created attachment 413858 [details]
Patch
Comment on attachment 413858 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=413858&action=review > Source/WebKit/UIProcess/WebProcessPool.cpp:270 > +#if ENABLE(GPU_PROCESS) > + , m_resetGPUProcessCrashCountTimer(RunLoop::main(), this, &WebProcessPool::resetGPUProcessCrashCount) > +#endif DeferrableTaskTimer is slightly nicer for this kind of thing because it doesn't require an initializer. > Source/WebKit/UIProcess/WebProcessPool.cpp:519 > +void WebProcessPool::resetGPUProcessCrashCount() > +{ > + m_recentGPUProcessCrashCount = 0; > } DeferrableTaskTimer is slightly nicer for this kind of thing because it doesn't require a separate function. Committed r269703: <https://trac.webkit.org/changeset/269703> All reviewed patches have been landed. Closing bug and clearing flags on attachment 413858 [details]. (In reply to Geoffrey Garen from comment #6) > Comment on attachment 413858 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=413858&action=review > > > Source/WebKit/UIProcess/WebProcessPool.cpp:270 > > +#if ENABLE(GPU_PROCESS) > > + , m_resetGPUProcessCrashCountTimer(RunLoop::main(), this, &WebProcessPool::resetGPUProcessCrashCount) > > +#endif > > DeferrableTaskTimer is slightly nicer for this kind of thing because it > doesn't require an initializer. > > > Source/WebKit/UIProcess/WebProcessPool.cpp:519 > > +void WebProcessPool::resetGPUProcessCrashCount() > > +{ > > + m_recentGPUProcessCrashCount = 0; > > } > > DeferrableTaskTimer is slightly nicer for this kind of thing because it > doesn't require a separate function. Following up in https://bugs.webkit.org/show_bug.cgi?id=218828. |