Repository navigation
Nullref in FindImplSlotForCurrentType #105304
Description
[FAIL] System.Collections.ObjectModel.Tests.CollectionTests.Insert_ZeroIndex
System.NullReferenceException : Object reference not set to an instance of an object.
at System.Runtime.DispatchResolve.FindImplSlotForCurrentType(MethodTable*, MethodTable*, UInt16, Boolean, UInt16*, MethodTable**) + 0x44
at System.Runtime.DispatchResolve.FindInterfaceMethodImplementationTarget(MethodTable*, MethodTable*, UInt16, MethodTable**) + 0x64
at System.Runtime.CachedInterfaceDispatch.RhResolveDispatchWorker(Object, Void*, DispatchCellInfo&) + 0x70
at System.Runtime.CachedInterfaceDispatch.RhpCidResolve_Worker(Object, IntPtr) + 0x30
at System.Collections.ObjectModel.Tests.CollectionTests.Insert_ZeroIndex() + 0x5c
Build Information
Build: https://dev.azure.com/dnceng-public/cbb18261-c48f-4abb-8651-8cdcb5474649/_build/results?buildId=748325
Build error leg or test failing: System.Reflection.Tests.WorkItemExecution
Pull request: #105160
Error Message
Fill the error message using step by step known issues guidance.
{
"ErrorMessage": "System.Runtime.DispatchResolve.FindImplSlotForCurrentType",
"ErrorPattern": "",
"BuildRetry": false,
"ExcludeConsoleLog": false
}Known issue validation
Build: 🔎 https://dev.azure.com/dnceng-public/public/_build/results?buildId=748325
Error message validated: [System.Runtime.DispatchResolve.FindImplSlotForCurrentType]
Result validation: ✅ Known issue matched with the provided build.
Validation performed at: 7/23/2024 9:51:38 AM UTC
Report
| Build | Definition | Test | Pull Request |
|---|---|---|---|
| 784563 | dotnet/runtime | System.Collections.ObjectModel.Tests.CollectionTests.Insert_ZeroIndex | #106752 |
| 782219 | dotnet/runtime | System.Collections.ObjectModel.Tests.CollectionTests.Insert_ZeroIndex | #106564 |
| 781179 | dotnet/runtime | System.Collections.ObjectModel.Tests.CollectionTests.Insert_ZeroIndex | #106673 |
| 780541 | dotnet/runtime | System.Collections.ObjectModel.Tests.CollectionTests.Insert_ZeroIndex | #106592 |
Summary
| 24-Hour Hit Count | 7-Day Hit Count | 1-Month Count |
|---|---|---|
| 0 | 0 | 4 |
Activity
dotnet-policy-service commentedon Jul 23, 2024
Tagging subscribers to this area: @agocke, @MichalStrehovsky, @jkotas
See info in area-owners.md if you want to be subscribed.
MichalStrehovsky commentedon Aug 2, 2024
We got a repro with a dump at https://dev.azure.com/dnceng-public/public/_build/results?buildId=763400&view=ms.vss-test-web.build-test-results-tab&runId=19408156&resultId=100001.
Can be pulled with
runfo get-helix-payload -j cbe1b7bb-c21c-4b8f-a98a-94a24a648442 -w System.Collections.Tests -o c:\helix_payload\System.Collections.TestsIn the dump, zooming into
RhResolveDispatchWorker, we have a validpObjectthat is of type 0x00007ff60260de60 (MethodTable forSortedSet<Int32>.Node). However,pInstanceTypeis 0x0043007400730065 and that's a bogus number. It should instead hold the MethodTable of the SortedSet in pObject. We then propagate this bugs number through a couple more frames and eventually AV.@VSadov could this be GCInfo related? I don't see many ways how a value held entirely on stack/registers (in the short time that it takes to call the next method where the value is passed by value) could be corrupted.
MichalStrehovsky commentedon Aug 2, 2024
Hmm, but the latest hit is from a RyuJIT PR that is messing with GC liveness so it might be related to that.
So that dump might not be related to the mac arm64 issue that we otherwise have no dumps for. But that would also mean we haven't seen that one for more than a week. One more week and we can close as no repro 🤞
MichalStrehovsky commentedon Aug 8, 2024
Oh joy, it reproed. And still arm64 macOS only.
MichalStrehovsky commentedon Sep 11, 2024
No repro in 3 weeks, closing.