Landing: acd18f524416
Project / Subsystem
binutils / x86
Date
2026-07-13
Author
"H.J. Lu" <hjl@sourceware.org>
Commit
acd18f5244167f458f5e23252c4a7b050157389a
Source
public_inbox
Perf win
No
Breaking
No
All attributes
- project
- binutils
- subsystem
- x86
- patch_id
- —
- commit_hash
- acd18f5244167f458f5e23252c4a7b050157389a
- source_type
- public_inbox
- headline
- Disable XCHG to MOV optimization in assembler
- tldr
- Assembler optimization for XCHG to MOV instruction is disabled on x86.
- author
- "H.J. Lu" <hjl@sourceware.org>
- outcome
- committed
- performance_win
- false
- breaking_change
- false
- series_id
- —
- series_parts
- []
- tags
-
- • x86
- • assembler
- • optimization
- discussion_id_link
- —
- bugzilla_pr
- —
- date
- 2026-07-13T00:00:00.000Z
This commit disables a specific optimization in the x86 assembler that rewrites XCHG instructions into MOV instructions when certain conditions are met. This optimization, enabled by -O, -O2, and -Os flags, was intended to shorten instruction encodings but has now been disabled to address potential issues. The change affects assembly code generation for x86 architectures.