Re: LoongArch: Enforce 4-byte align for machine instructions

Project / Subsystem

binutils / loongarch

Date

2026-05-01

Proposer

Alan Modra <amodra@gmail.com>

Source type

public_inbox

Consensus

Proposed

Sentiment

/10

Technical tradeoffs

  • Fixes a broken test case, ensuring proper instruction alignment is enforced.
  • The looser regex may be less precise, but still confirms the essential alignment property.

All attributes

project
binutils
subsystem
loongarch
patch_id
discussion_id
afSgc4Rp5QmZdQBD@squeak.grove.modra.org
source_type
public_inbox
title
Re: LoongArch: Enforce 4-byte align for machine instructions
headline
LoongArch: Fix test for instruction alignment
tldr
Fixes a test case for LoongArch instruction alignment that was over-specified, causing failures on 32-bit objects.
proposer
Alan Modra <amodra@gmail.com>
consensus
Proposed
outcome
proposed
sentiment_score
technical_tradeoffs
  • Fixes a broken test case, ensuring proper instruction alignment is enforced.
  • The looser regex may be less precise, but still confirms the essential alignment property.
series_id
series_role
standalone
series_parts
[]
tags
  • binutils
  • assembler
  • gas
  • loongarch
  • test case
bugzilla_url
date
2026-05-01T00:00:00.000Z

Re: LoongArch: Enforce 4-byte align for machine instructions

Alan Modra corrects a test case (insn_align_4.d) for LoongArch instruction alignment. The original test was too specific and failed on 32-bit objects due to differences in readelf output. The patch relaxes the test to match the .text section alignment of 4 bytes without specifying other details.