Re: [PATCH v3 0/3] arm64: Fixes for __READ_ONCE() with CONFIG_LTO=y

[PATCH v3 0/3] arm64: Fixes for __READ_ONCE() with CONFIG_LTO=y · Marco Elver <elver@google.com> · 2026-01-30
[PATCH v3 1/3] arm64: Fix non-atomic __READ_ONCE() with CONFIG_LTO=y · Marco Elver <elver@google.com> · 2026-01-30
Re: [PATCH v3 1/3] arm64: Fix non-atomic __READ_ONCE() with CONFIG_LTO=y · David Laight <hidden> · 2026-01-30
[PATCH v3 2/3] arm64: Optimize __READ_ONCE() with CONFIG_LTO=y · Marco Elver <elver@google.com> · 2026-01-30
Re: [PATCH v3 2/3] arm64: Optimize __READ_ONCE() with CONFIG_LTO=y · David Laight <hidden> · 2026-01-30
Re: [PATCH v3 2/3] arm64: Optimize __READ_ONCE() with CONFIG_LTO=y · Will Deacon <will@kernel.org> · 2026-02-02
Re: [PATCH v3 2/3] arm64: Optimize __READ_ONCE() with CONFIG_LTO=y · Peter Zijlstra <peterz@infradead.org> · 2026-02-02
Re: [PATCH v3 2/3] arm64: Optimize __READ_ONCE() with CONFIG_LTO=y · Will Deacon <will@kernel.org> · 2026-02-02
Re: [PATCH v3 2/3] arm64: Optimize __READ_ONCE() with CONFIG_LTO=y · Marco Elver <elver@google.com> · 2026-02-02
Re: [PATCH v3 2/3] arm64: Optimize __READ_ONCE() with CONFIG_LTO=y · David Laight <hidden> · 2026-02-02
[PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · Marco Elver <elver@google.com> · 2026-01-30
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · David Laight <hidden> · 2026-01-30
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · Will Deacon <will@kernel.org> · 2026-02-02
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · David Laight <hidden> · 2026-02-02
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · Will Deacon <will@kernel.org> · 2026-02-03
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · Marco Elver <elver@google.com> · 2026-02-04
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · Peter Zijlstra <peterz@infradead.org> · 2026-02-04
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · Will Deacon <will@kernel.org> · 2026-02-04
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · Marco Elver <elver@google.com> · 2026-02-06
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · David Laight <hidden> · 2026-02-06
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · Marco Elver <elver@google.com> · 2026-02-15
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · David Laight <hidden> · 2026-02-15
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · Marco Elver <elver@google.com> · 2026-02-15
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · David Laight <hidden> · 2026-02-15
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · Linus Torvalds <torvalds@linux-foundation.org> · 2026-02-15
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · David Laight <hidden> · 2026-02-16
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · Linus Torvalds <torvalds@linux-foundation.org> · 2026-02-16
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · David Laight <hidden> · 2026-02-16
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · Peter Zijlstra <peterz@infradead.org> · 2026-02-17
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · David Laight <hidden> · 2026-02-17
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · Linus Torvalds <torvalds@linux-foundation.org> · 2026-02-17
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · Linus Torvalds <torvalds@linux-foundation.org> · 2026-02-17
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · Boqun Feng <boqun@kernel.org> · 2026-02-18
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · Linus Torvalds <torvalds@linux-foundation.org> · 2026-02-18
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · "Gary Guo" <gary@garyguo.net> · 2026-02-19
Re: [PATCH v3 3/3] arm64, compiler-context-analysis: Permit alias analysis through __READ_ONCE() with CONFIG_LTO=y · Linus Torvalds <torvalds@linux-foundation.org> · 2026-02-19
Re: [PATCH v3 0/3] arm64: Fixes for __READ_ONCE() with CONFIG_LTO=y · Will Deacon <will@kernel.org> · 2026-02-02

From: Will Deacon <will@kernel.org>
Date: 2026-02-02 19:13:51
Also in: lkml, llvm

On Fri, 30 Jan 2026 14:28:23 +0100, Marco Elver wrote:

While investigating a Clang Context Analysis [1] false positive [2], I started
to dig deeper into arm64's __READ_ONCE() implementation with LTO. That rabbit
hole led me to find one critical bug with the current implementation (patch 1),
and subtle improvements that then enabled me to fix the original false positive.

Patch 1 fixes a bug where READ_ONCE() on types larger than 8 bytes (non-atomic
fallback) incorrectly qualified the pointer rather than the pointee as volatile.
This resulted in a lack of "once" semantics for large struct loads.

[...]

Applied first patch only to arm64 (for-next/core), thanks!

[1/3] arm64: Fix non-atomic __READ_ONCE() with CONFIG_LTO=y
      https://git.kernel.org/arm64/c/bb0c99e08ab9

Cheers,
-- 
Will

https://fixes.arm64.dev
https://next.arm64.dev
https://will.arm64.dev

`h`	back out one level
`j`	next message in thread
`k`	previous message in thread
`l`	drill in
`Esc`	close help / fold thread tree
`?`	toggle this help