Repository navigation
Linux arm GCStress=C ZapDisable JitStress=2 failures #38230
Copy link
Copy link
Closed
Labels
GCStressJitStressCLR JIT issues involving JIT internal stress modesCLR JIT issues involving JIT internal stress modesarch-arm32area-CodeGen-coreclrCLR JIT compiler in src/coreclr/src/jit and related components such as SuperPMICLR JIT compiler in src/coreclr/src/jit and related components such as SuperPMIos-linuxLinux OS (any supported distro)Linux OS (any supported distro)
Milestone
Description
Over 4000 tests fail on Linux arm with:
export COMPlus_TieredCompilation=0
export COMPlus_GCStress=0xC
export COMPlus_JitStress=2
export COMPlus_ReadyToRun=0
export COMPlus_ZapDisable=1
E.g.:
+ dotnet /root/helix/work/correlation/xunit.console.dll JIT/Directed/JIT.Directed.XUnitWrapper.dll -parallel collections -nocolor -noshadow -xml testResults.xml -trait TestGroup=JIT.Directed
Microsoft.DotNet.XUnitConsoleRunner v2.5.0 (32-bit .NET 5.0.0-preview.6.20305.6)
Discovering: JIT.Directed.XUnitWrapper
Discovered: JIT.Directed.XUnitWrapper
Starting: JIT.Directed.XUnitWrapper
JIT/Directed/refbyref/ref2iu_il_d/ref2iu_il_d.sh [FAIL]
Assert failure(PID 102 [0x00000066], Thread: 102 [0x0066]): SanityCheck()
File: /__w/1/s/src/coreclr/src/vm/methodtable.cpp Line: 9357
Image: /root/helix/work/correlation/corerun
/root/helix/work/workitem/JIT/Directed/refbyref/ref2iu_il_d/ref2iu_il_d.sh: line 317: 102 Aborted $LAUNCHER $ExePath "${CLRTestExecutionArguments[@]}"
Return code: 1
Raw output file: /root/helix/work/workitem/JIT/Directed/Reports/JIT.Directed/refbyref/ref2iu_il_d/ref2iu_il_d.output.txt
Raw output:
BEGIN EXECUTION
/root/helix/work/correlation/corerun ref2iu_il_d.dll ''
Writing minidump to file /home/helixbot/dotnetbuild/dumps/coredump.102.dmp
Written 52105216 bytes (12721 pages) to core file
Dump successfully written
Expected: 100
Actual: 134
END EXECUTION - FAILED
Test Harness Exitcode is : 1
To run the test:
> set CORE_ROOT=/root/helix/work/correlation
> /root/helix/work/workitem/JIT/Directed/refbyref/ref2iu_il_d/ref2iu_il_d.sh
Expected: True
Actual: False
Stack Trace:
/__w/1/s/artifacts/tests/coreclr/Linux.arm.Checked/TestWrappers/JIT.Directed/JIT.Directed.XUnitWrapper.cs(45536,0): at JIT_Directed._refbyref_ref2iu_il_d_ref2iu_il_d_._refbyref_ref2iu_il_d_ref2iu_il_d_sh()
Output:
Assert failure(PID 102 [0x00000066], Thread: 102 [0x0066]): SanityCheck()
File: /__w/1/s/src/coreclr/src/vm/methodtable.cpp Line: 9357
Image: /root/helix/work/correlation/corerun
/root/helix/work/workitem/JIT/Directed/refbyref/ref2iu_il_d/ref2iu_il_d.sh: line 317: 102 Aborted $LAUNCHER $ExePath "${CLRTestExecutionArguments[@]}"
Return code: 1
Raw output file: /root/helix/work/workitem/JIT/Directed/Reports/JIT.Directed/refbyref/ref2iu_il_d/ref2iu_il_d.output.txt
Raw output:
BEGIN EXECUTION
/root/helix/work/correlation/corerun ref2iu_il_d.dll ''
Writing minidump to file /home/helixbot/dotnetbuild/dumps/coredump.102.dmp
Written 52105216 bytes (12721 pages) to core file
Dump successfully written
Expected: 100
Actual: 134
END EXECUTION - FAILED
Test Harness Exitcode is : 1
To run the test:
> set CORE_ROOT=/root/helix/work/correlation
> /root/helix/work/workitem/JIT/Directed/refbyref/ref2iu_il_d/ref2iu_il_d.sh
Not all failure modes are identical, but the ones I looked at look like classic GCStress failure modes.
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
GCStressJitStressCLR JIT issues involving JIT internal stress modesCLR JIT issues involving JIT internal stress modesarch-arm32area-CodeGen-coreclrCLR JIT compiler in src/coreclr/src/jit and related components such as SuperPMICLR JIT compiler in src/coreclr/src/jit and related components such as SuperPMIos-linuxLinux OS (any supported distro)Linux OS (any supported distro)
Activity
AndyAyersMS commentedon Jun 22, 2020
@BruceForstall Seems like this has to be a recent regression...?
BruceForstall commentedon Jun 22, 2020
Looks like the failure started between 20200607.1 and 20200613.1. Unfortunately, we haven't paid enough attention to these results because there have been so many failures here, and more to the point, so many failures in our non-gcstress runs.
AndyAyersMS commentedon Jun 30, 2020
@BruceForstall let me know if you want help with investigating this one.
AndyAyersMS commentedon Jul 1, 2020
As noted over in #36971, I suspect this is a (stress induced) exit profiler hook issue, perhaps introduced by #37679.
I'm going to provisionally undo that and see what happens.
AndyAyersMS commentedon Jul 20, 2020
Closing this as we've reverted #37679.