Landing: ea0f94f1737e

Project / Subsystem

gcc / aarch64

Date

2026-09-05

Author

Artemiy Volkov

Commit

ea0f94f1737e44f58a30ddedb3a76bca4d30ce2c

Source

github

Perf win

No

Breaking

No

All attributes

project
gcc
subsystem
aarch64
patch_id
—
commit_hash
ea0f94f1737e44f58a30ddedb3a76bca4d30ce2c
source_type
github
headline
Fix FP16 combine patterns for aarch64, preventing assembly failures.
tldr
Corrects aarch64 assembly generation for FP16 types, avoiding failures when the fp16 extension is not enabled.
author
Artemiy Volkov
outcome
committed
performance_win
false
breaking_change
false
series_id
—
series_parts
[]
tags
  • • aarch64
  • • compiler
  • • assembly
  • • fp16
discussion_id_link
—
bugzilla_pr
—
date
2026-09-05T00:00:00.000Z

This commit addresses an issue on aarch64 where the combinez patterns for scalar and partial vector FP16 types incorrectly used the FMOV instruction even when the fp16 extension was not available. This led to assembly failures. The fix specializes these patterns for HF and BF modes and uses the INS instruction as a fallback, resolving the bug reported in PR126983.