Bug 217730

Summary: REGRESSION(r268423) SHOULD NEVER BE REACHED: GPUConnectionToWebProcess::didReceiveMessage(IPC::Connection &, IPC::Decoder &)
Product: WebKit Reporter: Hector Lopez <hector_i_lopez>
Component: New BugsAssignee: Chris Dumez <cdumez>
Status: RESOLVED FIXED    
Severity: Normal CC: cdumez, ggaren, peng.liu6, sam, webkit-bot-watchers-bugzilla, webkit-bug-importer, youennf
Priority: P2 Keywords: InRadar
Version: WebKit Nightly Build   
Hardware: Unspecified   
OS: Unspecified   
Attachments:
Description Flags
Patch
none
Patch none

Hector Lopez
Reported 2020-10-14 14:22:23 PDT
fast/mediastream/captureInGPUProcess.html Test is a constant crash according to history on macOS and iOS Debug. First occurrence of a crash is at r268423 for iOS and has also been a flaky timeout. First occurrence of a crash on macOS is at r268431 but may be due to a lack of results from r268421. History: https://results.webkit.org/?suite=layout-tests&test=fast%2Fmediastream%2FcaptureInGPUProcess.html&platform=ios&platform=mac&style=debug stderr log: SHOULD NEVER BE REACHED /Volumes/Data/slave/ios-simulator-14-debug/build/WebKitBuild/Debug-iphonesimulator/DerivedSources/WebKit2/GPUConnectionToWebProcessMessageReceiver.cpp(96) : virtual void WebKit::GPUConnectionToWebProcess::didReceiveMessage(IPC::Connection &, IPC::Decoder &) 1 0x11f1bbe89 WTFCrash 2 0x10d1cc10b WTFCrashWithInfo(int, char const*, char const*, int) 3 0x10d5e9725 WebKit::GPUConnectionToWebProcess::didReceiveMessage(IPC::Connection&, IPC::Decoder&) 4 0x10d24127f IPC::Connection::dispatchMessage(IPC::Decoder&) 5 0x10d241bd0 IPC::Connection::dispatchMessage(std::__1::unique_ptr<IPC::Decoder, std::__1::default_delete<IPC::Decoder> >) 6 0x10d242260 IPC::Connection::dispatchOneIncomingMessage() 7 0x10d260cf8 IPC::Connection::enqueueIncomingMessage(std::__1::unique_ptr<IPC::Decoder, std::__1::default_delete<IPC::Decoder> >)::$_7::operator()() 8 0x10d260c0e WTF::Detail::CallableWrapper<IPC::Connection::enqueueIncomingMessage(std::__1::unique_ptr<IPC::Decoder, std::__1::default_delete<IPC::Decoder> >)::$_7, void>::call() 9 0x11f1e75a2 WTF::Function<void ()>::operator()() const 10 0x11f264f05 WTF::RunLoop::performWork() 11 0x11f269641 WTF::RunLoop::performWork(void*) 12 0x108ca3845 __CFRUNLOOP_IS_CALLING_OUT_TO_A_SOURCE0_PERFORM_FUNCTION__ 13 0x108ca373d __CFRunLoopDoSource0 14 0x108ca2c1f __CFRunLoopDoSources0 15 0x108c9d3f7 __CFRunLoopRun 16 0x108c9cb9e CFRunLoopRunSpecific 17 0x10868ce61 -[NSRunLoop(NSRunLoop) runMode:beforeDate:] 18 0x10868d075 -[NSRunLoop(NSRunLoop) run] 19 0x10a5d0506 _xpc_objc_main 20 0x10a5d24aa xpc_main 21 0x10da66ad5 WebKit::XPCServiceMain(int, char const**) 22 0x10ef4fbdb WKXPCServiceMain 23 0x1084bdd62 main 24 0x10a264415 start LEAK: 2 WebPageProxy
Attachments
Patch (7.11 KB, patch)
2020-10-14 15:40 PDT, Chris Dumez
no flags
Patch (7.12 KB, patch)
2020-10-14 16:19 PDT, Chris Dumez
no flags
Radar WebKit Bug Importer
Comment 1 2020-10-14 14:23:01 PDT
Chris Dumez
Comment 2 2020-10-14 14:51:11 PDT
It appears the IPC message that is not handled is RemoteAudioMediaStreamTrackRenderer::SetVolume. At first sight, I don't see a direct link to my change.
Chris Dumez
Comment 3 2020-10-14 14:55:29 PDT
Ok. I think I know what's going on.
Chris Dumez
Comment 4 2020-10-14 15:40:42 PDT
Peng Liu
Comment 5 2020-10-14 16:08:20 PDT
Comment on attachment 411380 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=411380&action=review > Source/WebKit/ChangeLog:25 > + good to have consistency between all kinds of receivers. Good fix! > Source/WebKit/Platform/IPC/Connection.cpp:1013 > + // First check if there is a global message receiver and return it there is one. Nit. s/it there is one/it if there is one/
Chris Dumez
Comment 6 2020-10-14 16:19:09 PDT
Geoffrey Garen
Comment 7 2020-10-14 16:33:26 PDT
Comment on attachment 411384 [details] Patch r=me
EWS
Comment 8 2020-10-14 17:09:23 PDT
Committed r268500: <https://trac.webkit.org/changeset/268500> All reviewed patches have been landed. Closing bug and clearing flags on attachment 411384 [details].
Note You need to log in before you can comment on or make changes to this bug.