Skip to content

Commit

Permalink
Disable NEON intrinsics on big-endian ARM
Browse files Browse the repository at this point in the history
These are currently broken because the order of elements inside
vectors is reversed on big-endian systems: the ARM ABI requires that
element 0 is located at the highest address of the vector type. However
LLVM intrinsics expect element 0 to be located at the lowest address.

See https://llvm.org/docs/BigEndianNEON.html and `arm_neon.h` in
Clang for more details.

Although this is a breaking change, this is acceptable for 2 reasons:
- big endian ARM targets are only tier 3.
- it is preferable to stop existing code from compiling than to let it
run and produce incorrect results.
  • Loading branch information
Amanieu committed Oct 19, 2023
1 parent f4528dd commit f91abb5
Show file tree
Hide file tree
Showing 6 changed files with 14 additions and 98 deletions.
3 changes: 3 additions & 0 deletions crates/core_arch/src/aarch64/mod.rs
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,10 @@
//! [arm_ref]: http://infocenter.arm.com/help/topic/com.arm.doc.ihi0073a/IHI0073A_arm_neon_intrinsics_ref.pdf
//! [arm_dat]: https://developer.arm.com/technologies/neon/intrinsics
// NEON intrinsics are currently broken on big-endian, so don't expose them. (#1484)
#[cfg(target_endian = "little")]
mod neon;
#[cfg(target_endian = "little")]
pub use self::neon::*;

mod tme;
Expand Down
Loading

0 comments on commit f91abb5

Please sign in to comment.