Description
CoreCLR crashes when running our tests, I believe it's because the test is throwing an exception, although I haven't confirmed this.
Reproduction Steps
$ git clone https://github.com/rolfbjarne/macios
$ cd macios
$ git checkout -b x64-crash-clr-interp
$ ./repro.sh
Expected behavior
No crash.
Actual behavior
Crash report: https://gist.github.com/rolfbjarne/e086cc123dbe19dc134494d04d290726
Looks related to exception handling:
Thread 0 Crashed:: Dispatch queue: com.apple.main-thread
0 <translation info unavailable> 0x103195688 ???
1 libcoreclr.dylib 0x10be1f4d0 memcpy_s + 128
2 libcoreclr.dylib 0x10be0ea90 PAL_VirtualUnwind + 672
3 libcoreclr.dylib 0x10bba96f5 InterpreterCodeManager::ResumeAfterCatch(_CONTEXT*, unsigned long, bool) + 69
4 libcoreclr.dylib 0x10bd3ee41 CallCatchFunclet(Object*, unsigned char*, REGDISPLAY*, ExInfo*) + 625
5 libcoreclr.dylib 0x10bd3dc35 DispatchExSecondPass(ExInfo*) + 245
6 libcoreclr.dylib 0x10bd3d586 DispatchManagedException(Object*, _CONTEXT*, _EXCEPTION_RECORD*) + 406
7 libcoreclr.dylib 0x10bd3dd3b DispatchManagedException(Object*) + 59
8 libcoreclr.dylib 0x10bcc0d30 InterpExecMethod(InterpreterFrame*, InterpMethodContextFrame*, InterpThreadContext*, ExceptionClauseArgs*) + 25216
9 libcoreclr.dylib 0x10bc21a87 ExecuteInterpretedMethod + 343
10 libcoreclr.dylib 0x10bdee61f InterpreterStubRetI8 + 15
might also be some kind of stack overflow, because the remaining 500 frames are identical:
[...]
509 libcoreclr.dylib 0x10bdee61f InterpreterStubRetI8 + 15
510 libcoreclr.dylib 0x10bdee61f InterpreterStubRetI8 + 15
Regression?
No response
Known Workarounds
No response
Configuration
$ dotnet --version
11.0.100-alpha.1.25614.102
Other information
No response
Description
CoreCLR crashes when running our tests, I believe it's because the test is throwing an exception, although I haven't confirmed this.
Reproduction Steps
Expected behavior
No crash.
Actual behavior
Crash report: https://gist.github.com/rolfbjarne/e086cc123dbe19dc134494d04d290726
Looks related to exception handling:
might also be some kind of stack overflow, because the remaining 500 frames are identical:
Regression?
No response
Known Workarounds
No response
Configuration
Other information
No response