Prevent devirtualizing into abstract classes never derived from #1700
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…from We defer looking at virtual methods on abstract classes and instead only look at them once we have a non-abstract class deriving from them. This lets us optimize away method bodies on abstract classes that are overriden by all descendants and therefore unreachable. If there's no non-abstract descendant, we didn't look at any virtuals. But RyuJIT might still attempt to devirtualize into such classes. We need to prevent that to avoid compilation failures (and bloat).
MichalStrehovsky
added
the
area-NativeAOT-coreclr
MichalStrehovsky
added a commit
to MichalStrehovsky/runtime
that referenced
this pull request
Aug 12, 2022
If we were in a situation like in the regression test, we would devirtualize the `GrabObject` call to `SomeUnallocatedClass.GrabObject`. But because `SomeUnallocatedClass` was never allocated, the scanner didn't scan it, and bad things would happen. Prevent devirtualizing into types that were not seen as allocated. This is not a real issue for non-generic (non-shareable) types because the tentative instance method optimization generates throwing bodies for these. But tentative method optimization doesn't run for shared generics: https://github.com/dotnet/runtime/blob/4cbe6f99d23e04c56a89251d49de1b0f14000427/src/coreclr/tools/Common/Compiler/MethodExtensions.cs#L115 This was rare enough that we haven't seen it until I did dotnet#73683 and there was one useless constructor that we stopped generating and triggered this. This also includes what is essentially a rollback of dotnet/runtimelab#1700. This should have been rolled back with dotnet#66145 but I forgot we had this. It was not needed.
MichalStrehovsky
added a commit
to dotnet/runtime
that referenced
this pull request
Aug 12, 2022
If we were in a situation like in the regression test, we would devirtualize the `GrabObject` call to `SomeUnallocatedClass.GrabObject`. But because `SomeUnallocatedClass` was never allocated, the scanner didn't scan it, and bad things would happen. Prevent devirtualizing into types that were not seen as allocated. This is not a real issue for non-generic (non-shareable) types because the tentative instance method optimization generates throwing bodies for these. But tentative method optimization doesn't run for shared generics: https://github.com/dotnet/runtime/blob/4cbe6f99d23e04c56a89251d49de1b0f14000427/src/coreclr/tools/Common/Compiler/MethodExtensions.cs#L115 This was rare enough that we haven't seen it until I did #73683 and there was one useless constructor that we stopped generating and triggered this. This also includes what is essentially a rollback of dotnet/runtimelab#1700. This should have been rolled back with #66145 but I forgot we had this. It was not needed. * Update tests.proj
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
We defer looking at virtual methods on abstract classes and instead only look at them once we have a non-abstract class deriving from them. This lets us optimize away method bodies on abstract classes that are overriden by all descendants and therefore unreachable.
If there's no non-abstract descendant, we didn't look at any virtuals. But RyuJIT might still attempt to devirtualize into such classes. We need to prevent that to avoid compilation failures (and bloat).