Fixes for various endianness issues #5302
Open
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Outdated suggestions cannot be applied.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.
The SPIR-V specifications is agnostic on the issue of endianness (at least at the 32-bit word level), and tooling has some supporting code for it. It sadly appears to suffer from quite a bit of rot, this PR tackles the following:
spirv-dis
fails to disassemble a SPIR-V file in the non-native endiannessMakeString
I ran and passed all tests on a real PowerPC G5, which is as fun a way to spend a day as any other. There's probably more I can do, but I'd like this to be reviewed before just so I know I'm on the right track.
Additionally, I'm not sure how relevant big-endian SPIR-V support really is considering the state of the tooling. Rather it would probably be beneficial to have big-endian architectures default to emitting little-endian files, as I doubt there are many, if any implementations that correctly implement endianness agnosticism. I think it needs to at least be an option!