Skip to content

Some more changes in stdarch-gen-arm#2105

Merged
folkertdev merged 3 commits intorust-lang:mainfrom
sayantn:gen-arm-improvements
May 7, 2026
Merged

Some more changes in stdarch-gen-arm#2105
folkertdev merged 3 commits intorust-lang:mainfrom
sayantn:gen-arm-improvements

Conversation

@sayantn
Copy link
Copy Markdown
Contributor

@sayantn sayantn commented May 7, 2026

attempting to make push some more changes to stdarch-gen-arm standalone before #2063. Again, all commits are by @davidtwco.

r? @folkertdev

@sayantn sayantn force-pushed the gen-arm-improvements branch 2 times, most recently from 8bc2159 to 49c2820 Compare May 7, 2026 19:38
davidtwco added 3 commits May 8, 2026 01:16
Changes the definition of `vabdl_high_u{8,16,32}` to match that of
`vabdl_high_s{8,16,32}` so that the `big_endian_inverse` transformation
can apply.
Writing these intrinsics in terms of the `vqtbl_$ty` functions rather
than the `vqtbl` helper function preserves the behaviour while making
them work in big endian.
Writing these intrinsics in terms of the `vqtbx_$ty` functions rather
than the `vqtbx` helper function preserves the behaviour while making
them work in big endian.
@sayantn sayantn force-pushed the gen-arm-improvements branch from 49c2820 to 8f84a30 Compare May 7, 2026 19:46
@sayantn sayantn marked this pull request as ready for review May 7, 2026 19:55
Copy link
Copy Markdown
Contributor

@folkertdev folkertdev left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks fine, and I'm happy to unblock the main PR. Some things we could follow up on

  • crate::mem::zeroed() can probably be replaced by vec_ty::ZERO
  • functions like vreinterpret_u8_s8 probably don't need a separate big-endian implementation (in most cases?)

View changes since this review

@folkertdev folkertdev added this pull request to the merge queue May 7, 2026
@sayantn
Copy link
Copy Markdown
Contributor Author

sayantn commented May 7, 2026

I tried to put the vreinterpret changes here, CI was not too happy with it, I will try again in a different PR

@folkertdev
Copy link
Copy Markdown
Contributor

Perhaps it should use simd_as instead of a transmute?

Merged via the queue into rust-lang:main with commit 5569035 May 7, 2026
76 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants