Fix ARM cast codegen#18063
Conversation
ARM cast codegen is rather convoluted and sometimes does the wrong thing by applying GTF_UNSIGNED to the destination type even though this flag is only about the source type.
These tests are intended to support casts with contained operands. They're also useful to catch issues caused by load nodes having small types and generally improve test coverage for casts.
|
@dotnet/jit-contrib |
CarolEidt
left a comment
CarolEidt
left a comment
There was a problem hiding this comment.
LGTM, but I'll hold off on merging as @BruceForstall has self-requested a review.
|
@dotnet-bot test Ubuntu arm Cross Checked normal Build and Test @dotnet-bot test Windows_NT arm64 Cross Checked normal Build and Test |
|
Presumably the failure of FixedAddressValueType is a glitch? I tried to run it locally with AltJit and I don't see any differences in the generated code. |
|
Perhaps, although I don't see any other example of it failing. Try again: @dotnet-bot test Ubuntu arm Cross Checked jitstress1 Build and Test |
|
ping @BruceForstall - are you still wanting to review this? |
|
Hopefully you aren't trying to review those 11000 lines of test code :) |
|
All the tests I triggered passed, so I'll sign off. (Was mostly out today) |
* Fix ARM cast codegen ARM cast codegen is rather convoluted and sometimes does the wrong thing by applying GTF_UNSIGNED to the destination type even though this flag is only about the source type. * Add more conversion tests These tests are intended to support casts with contained operands. They're also useful to catch issues caused by load nodes having small types and generally improve test coverage for casts. Commit migrated from dotnet/coreclr@4cdb810
3 participants
ARM cast codegen is rather convoluted and sometimes does the wrong thing by applying
GTF_UNSIGNEDto the destination type even though this flag is only about the source type.Fixes #18040