Re: [Patch] LoongArch: Restore pointer_equality_needed for GOT data relocations
Project / Subsystem
binutils / loongarch
Date
2026-08-01
Proposer
Xi Ruoyao <xry111@xry111.site>
Source type
public_inbox
Consensus
Proposed
Sentiment
—/10
Technical tradeoffs
- • Balancing the fix for PR 29655 (function pointer equality in PIE) with the correct handling of data symbol access via GOT.
- • Determining the correct usage of `pointer_equality_needed` for different symbol types (function vs. data).
- • Ensuring ABI compatibility and runtime stability across various linking scenarios (static, dynamic, PIE).
All attributes
- project
- binutils
- subsystem
- loongarch
- patch_id
- —
- discussion_id
- 6a024beec3770ba7f563e1fc1632d41a44fa9264.camel@xry111.site
- source_type
- public_inbox
- title
- Re: [Patch] LoongArch: Restore pointer_equality_needed for GOT data relocations
- headline
- LoongArch: Restore pointer_equality_needed for GOT data relocations
- tldr
- Developer questions the rationale for restoring `pointer_equality_needed` for GOT data relocations, arguing it's irrelevant for data symbols.
- proposer
- Xi Ruoyao <xry111@xry111.site>
- consensus
- Proposed
- outcome
- proposed
- sentiment_score
- —
- technical_tradeoffs
-
- • Balancing the fix for PR 29655 (function pointer equality in PIE) with the correct handling of data symbol access via GOT.
- • Determining the correct usage of `pointer_equality_needed` for different symbol types (function vs. data).
- • Ensuring ABI compatibility and runtime stability across various linking scenarios (static, dynamic, PIE).
- series_id
- —
- series_role
- standalone
- series_parts
- []
- tags
-
- • binutils
- • loongarch
- • relocations
- • got
- • linker
- url
- https://inbox.sourceware.org/binutils/6a024beec3770ba7f563e1fc1632d41a44fa9264.camel@xry111.site
- bugzilla_url
- —
- date
- 2026-08-01T00:00:00.000Z
Re: [Patch] LoongArch: Restore pointer_equality_needed for GOT data relocations
Xi Ruoyao expresses skepticism regarding a proposed patch to restore pointer_equality_needed for LoongArch GOT data relocations. He argues that this flag is relevant only for function symbols to ensure pointer equality between PLT stubs and actual functions, and has no bearing on data symbols, suggesting the observed breakage for data symbols implies a more widespread issue than the patch addresses.