Bug 168587 – ASSERTION failure: !masqueradesAsUndefinedWatchpointIsStillValid() || !isKnownCell(operand.node())

• Component: JavaScriptCore
• Status: RESOLVED DUPLICATE

[Mark Lam]

When running the JSC tests with a debug build (WebKit trunk r212625 in my case), we sometimes get this failure:

stress/test-finally.js.ftl-eager: ASSERTION FAILED: The Compare should have been eliminated, it is known to be always false.
stress/test-finally.js.ftl-eager: !masqueradesAsUndefinedWatchpointIsStillValid() || !isKnownCell(operand.node())
stress/test-finally.js.ftl-eager: /Volumes/Data/ws3/OpenSource/Source/JavaScriptCore/dfg/DFGSpeculativeJIT64.cpp(222) : void JSC::DFG::SpeculativeJIT::nonSpeculativeNonPeepholeCompareNullOrUndefined(JSC::DFG::Edge)
stress/test-finally.js.ftl-eager: 1   0x108aec19d WTFCrash
stress/test-finally.js.ftl-eager: 2   0x108161af9 JSC::DFG::SpeculativeJIT::nonSpeculativeNonPeepholeCompareNullOrUndefined(JSC::DFG::Edge)
stress/test-finally.js.ftl-eager: 3   0x1080fdbdb JSC::DFG::SpeculativeJIT::compare(JSC::DFG::Node*, JSC::MacroAssemblerX86Common::RelationalCondition, JSC::MacroAssemblerX86Common::DoubleCondition, unsigned long (*)(JSC::ExecState*, long long, long long))
stress/test-finally.js.ftl-eager: 4   0x108171617 JSC::DFG::SpeculativeJIT::compile(JSC::DFG::Node*)
stress/test-finally.js.ftl-eager: 5   0x1080e30af JSC::DFG::SpeculativeJIT::compileCurrentBlock()
stress/test-finally.js.ftl-eager: 6   0x1080e3c43 JSC::DFG::SpeculativeJIT::compile()
stress/test-finally.js.ftl-eager: 7   0x107fa27e7 JSC::DFG::JITCompiler::compileBody()
stress/test-finally.js.ftl-eager: 8   0x107fa78a3 JSC::DFG::JITCompiler::compileFunction()
stress/test-finally.js.ftl-eager: 9   0x108094b4e JSC::DFG::Plan::compileInThreadImpl(JSC::DFG::LongLivedState&)
stress/test-finally.js.ftl-eager: 10  0x108091e49 JSC::DFG::Plan::compileInThread(JSC::DFG::LongLivedState&, JSC::DFG::ThreadData*)
stress/test-finally.js.ftl-eager: 11  0x108207325 JSC::DFG::Worklist::ThreadBody::work()
stress/test-finally.js.ftl-eager: 12  0x108afa25d WTF::AutomaticThread::start(WTF::Locker<WTF::LockBase> const&)::$_0::operator()() const
stress/test-finally.js.ftl-eager: 13  0x108afa01d void std::__1::__invoke_void_return_wrapper<void>::__call<WTF::AutomaticThread::start(WTF::Locker<WTF::LockBase> const&)::$_0&>(WTF::AutomaticThread::start(WTF::Locker<WTF::LockBase> const&)::$_0&&&)
stress/test-finally.js.ftl-eager: 14  0x108af9db9 std::__1::__function::__func<WTF::AutomaticThread::start(WTF::Locker<WTF::LockBase> const&)::$_0, std::__1::allocator<WTF::AutomaticThread::start(WTF::Locker<WTF::LockBase> const&)::$_0>, void ()>::operator()()
stress/test-finally.js.ftl-eager: 15  0x1080d9e5a std::__1::function<void ()>::operator()() const
stress/test-finally.js.ftl-eager: 16  0x108b5c4a7 WTF::threadEntryPoint(void*)
stress/test-finally.js.ftl-eager: 17  0x108b5e081 WTF::wtfThreadEntryPoint(void*)
stress/test-finally.js.ftl-eager: 18  0x7fffd076eaab _pthread_body
stress/test-finally.js.ftl-eager: 19  0x7fffd076e9f7 _pthread_body
stress/test-finally.js.ftl-eager: 20  0x7fffd076e1fd thread_start
stress/test-finally.js.ftl-eager: test_script_16641: line 2: 42646 Segmentation fault: 11  ( "$@" ../../.vm/JavaScriptCore.framework/Resources/jsc --useFTLJIT\=false --useFunctionDotArguments\=true --maxPerThreadStackUsage\=1572864 --useFTLJIT\=true --thresholdForJITAfterWarmUp\=10 --thresholdForJITSoon\=10 --thresholdForOptimizeAfterWarmUp\=20 --thresholdForOptimizeAfterLongWarmUp\=20 --thresholdForOptimizeSoon\=20 --thresholdForFTLOptimizeAfterWarmUp\=20 --thresholdForFTLOptimizeSoon\=20 --maximumEvalCacheableSourceLength\=150000 --useEagerCodeBlockJettisonTiming\=true test-finally.js )
stress/test-finally.js.ftl-eager: ERROR: Unexpected exit code: 139

[Ryan Haddad]

This flaky assertion still appears on the Debug JSC bot:
https://build.webkit.org/builders/Apple%20El%20Capitan%20Debug%20JSC%20%28Tests%29/builds/333

[Radar WebKit Bug Importer]

<rdar://problem/32201671>

[Ryan Haddad]

Skipped test in http://trac.webkit.org/projects/webkit/changeset/217621

[Filip Pizlo]

(In reply to Ryan Haddad from comment #3)
> Skipped test in http://trac.webkit.org/projects/webkit/changeset/217621

Thanks!

[Mark Lam]

Looks like this issue is solved by https://bugs.webkit.org/show_bug.cgi?id=172673.  Will un-skip the test once the fix lands.

[Mark Lam]

Since the fix has landed in r217896: <http://trac.webkit.org/changeset/217896>, I've unskipped the test in r217900: <http://trac.webkit.org/r217900>.

[Mark Lam]

*** This bug has been marked as a duplicate of bug 172673 ***