Fix a couple .NET 9 testing issues in the SOS tests in the diagnostics repo #99255
Conversation
|
Tagging subscribers to this area: @tommcdon Issue DetailsFix SOS DAC GetAssemblyData failure from ClassLoader not being in core dump Bump the MacOS stack overflow stack size
|
| @@ -849,7 +849,7 @@ HijackFaultingThread( | |||
| if (fIsStackOverflow) | |||
| { | |||
| // Allocate the minimal stack necessary for handling stack overflow | |||
| int stackOverflowStackSize = 7 * 4096; | |||
| int stackOverflowStackSize = 15 * 4096; | |||
There was a problem hiding this comment.
Yes, @mikem8361 has hit the issue, so I've told him he could make this part of the change if he wants to. I'll fix the stackoverflow detection in the presence of the macos issue.
There was a problem hiding this comment.
I wanted to get the couple of changes in so I can get my diagnostics repo .NET testing in soon.
| @@ -2577,7 +2577,7 @@ ClrDataAccess::GetAssemblyData(CLRDATA_ADDRESS cdBaseDomainPtr, CLRDATA_ADDRESS | |||
| } | |||
|
|
|||
| assemblyData->AssemblyPtr = HOST_CDADDR(pAssembly); | |||
| assemblyData->ClassLoader = HOST_CDADDR(pAssembly->GetLoader()); | |||
| assemblyData->ClassLoader = 0; | |||
There was a problem hiding this comment.
This change LGTM. Are you going to make a matching change in the SOS to delete ClassLoader from the DumpAssembly output?
There was a problem hiding this comment.
I thought I would leave it in SOS for now. It will still print something valid for older runtimes and 0 for .NET 9. But if you insist, I'll remove it.
There was a problem hiding this comment.
I think it would be an improvement to delete it from SOS output. The information that it prints for older runtimes is not very useful.
Fix SOS DAC GetAssemblyData failure from ClassLoader not being in core dump
Bump the MacOS stack overflow stack size