Re: bfin, frv, m68k, vax: got creation in check_relocs
Project / Subsystem
binutils / binutils
Date
2026-08-11
Proposer
"Maciej W. Rozycki" <macro@orcam.me.uk>
Source type
public_inbox
Consensus
Proposed
Sentiment
—/10
All attributes
- project
- binutils
- subsystem
- binutils
- patch_id
- —
- discussion_id
- alpine.DEB.2.21.2608112135560.48788@angie.orcam.me.uk
- source_type
- public_inbox
- title
- Re: bfin, frv, m68k, vax: got creation in check_relocs
- headline
- bfin, frv, m68k, vax: got creation in check_relocs
- tldr
- Call to create GOT section within `check_relocs` is acceptable due to idempotency.
- proposer
- "Maciej W. Rozycki" <macro@orcam.me.uk>
- consensus
- Proposed
- outcome
- proposed
- sentiment_score
- —
- technical_tradeoffs
- []
- series_id
- —
- series_role
- standalone
- series_parts
- []
- tags
-
- • binutils
- • bfd
- • relocation
- • got
- • code
- bugzilla_url
- —
- date
- 2026-08-11T00:00:00.000Z
Re: bfin, frv, m68k, vax: got creation in check_relocs
Alan Modra questioned the correctness of calling GOT section creation functions within a !dynobj check in check_relocs, as it was not obvious if the GOT section would always be created. Maciej W. Rozycki confirms that while the original intent might have been to call it only once, the functions are idempotent, making the current calls acceptable.