fix( #1375 ): under-aligned buffers in large-arg range reduction#1377
Merged
serge-sans-paille merged 1 commit intoJul 9, 2026
Merged
Conversation
Contributor
|
<3 |
The generic trigonometric range-reduction kernel stored the batch into std::array buffers declared alignas(B). alignof(B) equals the alignment of the underlying SIMD register, which on some ABIs is smaller than the alignment store_aligned/load_aligned require (A::alignment()). On armhf NEON alignof(batch<float>) is 8 while neon::alignment() is 16, so the buffer was only 8-byte aligned and store_aligned tripped its alignment assertion whenever the stack happened to place it at an 8-mod-16 address. This is exactly what broke the Debian armhf build of 14.2.0 (test '[complex power] pow real complex' -> SIGABRT in store_aligned). Align the buffers to B::arch_type::alignment() instead, matching the convention used by every other internal store_aligned buffer.
DiamonDinoia
force-pushed
the
repro/armhf-complex-pow-align
branch
from
4722cf4 to
d7b6ba1
Compare
Contributor
|
Thanks for fixing this one. It has been hanging around for a long time! |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The generic trigonometric range-reduction kernel stored the batch into std::array buffers declared alignas(B). alignof(B) equals the alignment of the underlying SIMD register, which on some ABIs is smaller than the alignment store_aligned/load_aligned require (A::alignment()). On armhf NEON alignof(batch) is 8 while neon::alignment() is 16, so the buffer was only 8-byte aligned and store_aligned tripped its alignment assertion whenever the stack happened to place it at an 8-mod-16 address.
This is exactly what broke the Debian armhf build of 14.2.0 (test '[complex power] pow real complex' -> SIGABRT in store_aligned).