Bug 286063
| Summary: | [GLIB] SSL API tests failing with ENABLE_ASSERTS=ON | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | Claudio Saavedra <csaavedra> |
| Component: | WebKitGTK | Assignee: | Nobody <webkit-unassigned> |
| Status: | RESOLVED FIXED | ||
| Severity: | Normal | CC: | bugs-noreply, clopez, fujii.hironori, pgriffis, zimmermann |
| Priority: | P2 | ||
| Version: | WebKit Nightly Build | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
Claudio Saavedra
We enabled assertions in the GTK bots and there are two SSL tests constantly failing due to an assertion reached:
/WebKitGTK/TestSSL:/webkit/WebKitWebView/web-socket-client-side-certificate
/WebKitGTK/TestSSL:/webkit/WebKitWebView/web-socket-tls-errors
The stacktrace is the same:
SHOULD NEVER BE REACHED
/app/webkit/Source/WebKit/NetworkProcess/NetworkProcess.cpp(464) : WebKit::NetworkProcess::AllowCookieAccess WebKit::NetworkProcess::allowsFirstPartyForCookies(WebCore::ProcessIdentifier, const RegistrableDomain&)
1 0x7f424a9f13aa WebKit::NetworkProcess::allowsFirstPartyForCookies(WTF::ObjectIdentifierGeneric<WebCore::ProcessIdentifierType, WTF::ObjectIdentifierMainThreadAccessTraits<unsigned long>, unsigned long>, WebCore::RegistrableDomain const&)
2 0x7f424a9f15c1 WebKit::NetworkProcess::allowsFirstPartyForCookies(WTF::ObjectIdentifierGeneric<WebCore::ProcessIdentifierType, WTF::ObjectIdentifierMainThreadAccessTraits<unsigned long>, unsigned long>, WTF::URL const&)
3 0x7f424a9cccb8 WebKit::NetworkConnectionToWebProcess::createSocketChannel(WebCore::ResourceRequest const&, WTF::String const&, WTF::ObjectIdentifierGeneric<WebCore::WebSocketIdentifierType, WTF::ObjectIdentifierThreadSafeAccessTraits<unsigned long>, unsigned long>, WTF::ObjectIdentifierGeneric<WebKit::WebPageProxyIdentifierType, WTF::ObjectIdentifierMainThreadAccessTraits<unsigned long>, unsigned long>, std::optional<WebCore::ProcessQualified<WTF::ObjectIdentifierGeneric<WebCore::FrameIdentifierType, WTF::ObjectIdentifierMainThreadAccessTraits<unsigned long>, unsigned long> > >, std::optional<WTF::ObjectIdentifierGeneric<WebCore::PageIdentifierType, WTF::ObjectIdentifierMainThreadAccessTraits<unsigned long>, unsigned long> >, WebCore::ClientOrigin const&, bool, bool, WTF::OptionSet<WebCore::AdvancedPrivacyProtections>, WebCore::ShouldRelaxThirdPartyCookieBlocking, WebCore::StoredCredentialsPolicy)
4 0x7f424a549f32 void IPC::handleMessage<Messages::NetworkConnectionToWebProcess::CreateSocketChannel, IPC::Connection, WebKit::NetworkConnectionToWebProcess, WebKit::NetworkConnectionToWebProcess, void (WebCore::ResourceRequest const&, WTF::String const&, WTF::ObjectIdentifierGeneric<WebCore::WebSocketIdentifierType, WTF::ObjectIdentifierThreadSafeAccessTraits<unsigned long>, unsigned long>, WTF::ObjectIdentifierGeneric<WebKit::WebPageProxyIdentifierType, WTF::ObjectIdentifierMainThreadAccessTraits<unsigned long>, unsigned long>, std::optional<WebCore::ProcessQualified<WTF::ObjectIdentifierGeneric<WebCore::FrameIdentifierType, WTF::ObjectIdentifierMainThreadAccessTraits<unsigned long>, unsigned long> > >, std::optional<WTF::ObjectIdentifierGeneric<WebCore::PageIdentifierType, WTF::ObjectIdentifierMainThreadAccessTraits<unsigned long>, unsigned long> >, WebCore::ClientOrigin const&, bool, bool, WTF::OptionSet<WebCore::AdvancedPrivacyProtections>, WebCore::ShouldRelaxThirdPartyCookieBlocking, WebCore::StoredCredentialsPolicy)>(IPC::Connection&, IPC::Decoder&, WebKit::NetworkConnectionToWebProcess*, void (WebKit::NetworkConnectionToWebProcess::*)(WebCore::ResourceRequest const&, WTF::String const&, WTF::ObjectIdentifierGeneric<WebCore::WebSocketIdentifierType, WTF::ObjectIdentifierThreadSafeAccessTraits<unsigned long>, unsigned long>, WTF::ObjectIdentifierGeneric<WebKit::WebPageProxyIdentifierType, WTF::ObjectIdentifierMainThreadAccessTraits<unsigned long>, unsigned long>, std::optional<WebCore::ProcessQualified<WTF::ObjectIdentifierGeneric<WebCore::FrameIdentifierType, WTF::ObjectIdentifierMainThreadAccessTraits<unsigned long>, unsigned long> > >, std::optional<WTF::ObjectIdentifierGeneric<WebCore::PageIdentifierType, WTF::ObjectIdentifierMainThreadAccessTraits<unsigned long>, unsigned long> >, WebCore::ClientOrigin const&, bool, bool, WTF::OptionSet<WebCore::AdvancedPrivacyProtections>, WebCore::ShouldRelaxThirdPartyCookieBlocking, WebCore::StoredCredentialsPolicy)) [clone .isra.0]
5 0x7f424a54b404 WebKit::NetworkConnectionToWebProcess::didReceiveMessage(IPC::Connection&, IPC::Decoder&)
6 0x7f424abf7c2d IPC::Connection::dispatchMessage(WTF::UniqueRef<IPC::Decoder>) [clone .part.0]
7 0x7f424abf8395 IPC::Connection::dispatchOneIncomingMessage()
8 0x7f4245176221 WTF::RunLoop::performWork()
9 0x7f4245252089 WTF::RunLoop::RunLoop()::{lambda(void*)#1}::_FUN(void*)
10 0x7f424525361f WTF::RunLoop::{lambda(_GSource*, int (*)(void*), void*)#1}::_FUN(_GSource*, int (*)(void*), void*)
11 0x7f423fd17d36 g_main_context_dispatch
12 0x7f423fd752b8 g_main_context_iterate.isra.0
13 0x7f423fd173ff g_main_loop_run
14 0x7f4245253750 WTF::RunLoop::run()
15 0x7f424ab4a63a int WebKit::AuxiliaryProcessMain<WebKit::NetworkProcessMainSoup>(int, char**)
16 0x7f423f75208a __libc_start_call_main
17 0x7f423f75214b __libc_start_main
18 0x5614fdc868a5 _start
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |
Claudio Saavedra
Gardened in https://github.com/WebKit/WebKit/pull/39131
Claudio Saavedra
They are also having the same problem in WPE.
Carlos Alberto Lopez Perez
This test is timing out on GTK (instead of failing/crashing/whatever) since bots were migrated to the new infra. See: https://bugs.webkit.org/show_bug.cgi?id=299598
Nikolas Zimmermann
Crash seen again in a different test: https://build.webkit.org/results/GTK-Linux-64-bit-Debug-Tests/304475@main%20(17735)/fast/forms/appearance-base/appearance-base-checkable-checkmark-dynamic-crash-log.txt
Nikolas Zimmermann
Same backtrace here: https://build.webkit.org/results/GTK-Linux-64-bit-Debug-Tests/304475@main%20(17735)/fast/mediastream/MediaStreamConstructor-crash-log.txt
Nikolas Zimmermann
(In reply to Nikolas Zimmermann from comment #5)
> Same backtrace here:
> https://build.webkit.org/results/GTK-Linux-64-bit-Debug-Tests/
> 304475@main%20(17735)/fast/mediastream/MediaStreamConstructor-crash-log.txt
Wrong paste, meant to paste this: https://build.webkit.org/results/GTK-Linux-64-bit-Debug-Tests/304475@main%20(17735)/fast/mediastream/RTCPeerConnection-page-cache-crash-log.txt
Nikolas Zimmermann
Same backtrace here: https://build.webkit.org/results/GTK-Linux-64-bit-Debug-Tests/304475@main%20(17735)/html5lib/generated/run-tests7-data-crash-log.txt
Nikolas Zimmermann
Same backtrace here: https://build.webkit.org/results/GTK-Linux-64-bit-Debug-Tests/304475@main%20(17735)/media/audio-only-video-intrinsic-size-crash-log.txt
Nikolas Zimmermann
Same backtrace here: https://build.webkit.org/results/GTK-Linux-64-bit-Debug-Tests/304475@main%20(17735)/storage/indexeddb/cursor-properties-crash-log.txt
Nikolas Zimmermann
Another one: https://build.webkit.org/results/GTK-Linux-64-bit-Debug-Tests/304475@main%20(17735)/storage/indexeddb/modern/idbobjectstore-delete-failures-private-crash-log.txt
Nikolas Zimmermann
Gardened the following tests in https://commits.webkit.org/304545@main:
fast/forms/appearance-base/appearance-base-checkable-checkmark-dynamic.html
fast/mediastream/RTCPeerConnection-page-cache.html
html5lib/generated/run-tests7-data.html
media/audio-only-video-intrinsic-size.html
storage/indexeddb/cursor-properties.html
storage/indexeddb/modern/idbobjectstore-delete-failures-private.html
Fujii Hironori
There are two bugs here. They have different backtraces.
NetworkProcess::allowsFirstPartyForCookie is called from
NetworkConnectionToWebProcess::createSocketChannel or
NetworkConnectionToWebProcess::scheduleResourceLoad.
Fujii Hironori
Pull request: https://github.com/WebKit/WebKit/pull/58586
EWS
Committed 307466@main (90216fdce034): <https://commits.webkit.org/307466@main>
Reviewed commits have been landed. Closing PR #58586 and removing active labels.
Claudio Saavedra
But the tests that Niko gardened are layout tests, are those unrelated to this bug then?
Fujii Hironori
yes, it's another bug. I'll file a bug.
Fujii Hironori
Created bug#308176 for the issue.