Linux-mm Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Jisheng Zhang" <jszhang@kernel.org>,
	"Catalin Marinas" <catalin.marinas@arm.com>,
	"Will Deacon" <will@kernel.org>,
	"Mark Rutland" <mark.rutland@arm.com>,
	"Andrey Ryabinin" <ryabinin.a.a@gmail.com>,
	"Alexander Potapenko" <glider@google.com>,
	"Andrey Konovalov" <andreyknvl@gmail.com>,
	"Dmitry Vyukov" <dvyukov@google.com>,
	"Vincenzo Frascino" <vincenzo.frascino@arm.com>,
	"Tejun Heo" <tj@kernel.org>, "Dennis Zhou" <dennis@kernel.org>,
	"Christoph Lameter" <cl@gentwo.org>
Cc: kasan-dev@googlegroups.com, linux-mm@kvack.org
Subject: Re: [PATCH 4/4] arm64: alternatives: Make branch and ADRP offset helpers noinst
Date: Sun, 08 Nov 2026 07:43:58 +0000	[thread overview]
Message-ID: <sashiko-outbox-203636@kernel.org> (raw)
In-Reply-To: <20261107071339.17467-5-jszhang@kernel.org>

Thank you for your contribution! Sashiko AI review found 1 potential
issue to consider:
- [Medium] arm64: insn: branch profiling instrumentation in noinstr
  offset helpers
--

commit b08429609214cdffaa257c9ef953b28e179973db
Author: Jisheng Zhang <jszhang@kernel.org>

arm64: alternatives: Make branch and ADRP offset helpers noinst

This commit marks the branch and ADRP offset helpers noinstr and introduces
__always_inline immediate encoding and decoding helpers so that
patch_alternative() does not call instrumentable code while modifying
instructions.

> diff --git a/arch/arm64/lib/insn.c b/arch/arm64/lib/insn.c
> index e70ac02385153..b98610fcab32e 100644
> --- a/arch/arm64/lib/insn.c
> +++ b/arch/arm64/lib/insn.c

[ ... ]

> @@ -1299,23 +1316,23 @@ u32 aarch64_insn_gen_adr(unsigned long pc, unsigned long addr,
>   * signed value (so it can be used when computing a new branch
>   * target).
>   */
> -s32 aarch64_get_branch_offset(u32 insn)
> +s32 noinstr aarch64_get_branch_offset(u32 insn)
>  {
>  	s32 imm;
>  
>  	if (aarch64_insn_is_b(insn) || aarch64_insn_is_bl(insn)) {
> -		imm = aarch64_insn_decode_immediate(AARCH64_INSN_IMM_26, insn);
> +		imm = __aarch64_insn_decode_immediate(AARCH64_INSN_IMM_26, insn);
>  		return (imm << 6) >> 4;
>  	}
>  
>  	if (aarch64_insn_is_cbz(insn) || aarch64_insn_is_cbnz(insn) ||
>  	    aarch64_insn_is_bcond(insn)) {
> -		imm = aarch64_insn_decode_immediate(AARCH64_INSN_IMM_19, insn);
> +		imm = __aarch64_insn_decode_immediate(AARCH64_INSN_IMM_19, insn);
>  		return (imm << 13) >> 11;
>  	}
>  
>  	if (aarch64_insn_is_tbz(insn) || aarch64_insn_is_tbnz(insn)) {
> -		imm = aarch64_insn_decode_immediate(AARCH64_INSN_IMM_14, insn);
> +		imm = __aarch64_insn_decode_immediate(AARCH64_INSN_IMM_14, insn);
>  		return (imm << 18) >> 16;
>  	}
>  
> @@ -1327,35 +1344,35 @@ s32 aarch64_get_branch_offset(u32 insn)
>   * Encode the displacement of a branch in the imm field and return the
>   * updated instruction.
>   */
> -u32 aarch64_set_branch_offset(u32 insn, s32 offset)
> +u32 noinstr aarch64_set_branch_offset(u32 insn, s32 offset)
>  {
>  	if (aarch64_insn_is_b(insn) || aarch64_insn_is_bl(insn))
> -		return aarch64_insn_encode_immediate(AARCH64_INSN_IMM_26, insn,
> +		return __aarch64_insn_encode_immediate(AARCH64_INSN_IMM_26, insn,
>  						     offset >> 2);
>  
>  	if (aarch64_insn_is_cbz(insn) || aarch64_insn_is_cbnz(insn) ||
>  	    aarch64_insn_is_bcond(insn))
> -		return aarch64_insn_encode_immediate(AARCH64_INSN_IMM_19, insn,
> +		return __aarch64_insn_encode_immediate(AARCH64_INSN_IMM_19, insn,
>  						     offset >> 2);
>  
>  	if (aarch64_insn_is_tbz(insn) || aarch64_insn_is_tbnz(insn))
> -		return aarch64_insn_encode_immediate(AARCH64_INSN_IMM_14, insn,
> +		return __aarch64_insn_encode_immediate(AARCH64_INSN_IMM_14, insn,
>  						     offset >> 2);
>  
>  	/* Unhandled instruction */
>  	BUG();
>  }
>  
> -s32 aarch64_insn_adrp_get_offset(u32 insn)
> +s32 noinstr aarch64_insn_adrp_get_offset(u32 insn)
>  {
>  	BUG_ON(!aarch64_insn_is_adrp(insn));
> -	return aarch64_insn_decode_immediate(AARCH64_INSN_IMM_ADR, insn) << 12;
> +	return __aarch64_insn_decode_immediate(AARCH64_INSN_IMM_ADR, insn) << 12;
>  }
>  
> -u32 aarch64_insn_adrp_set_offset(u32 insn, s32 offset)
> +u32 noinstr aarch64_insn_adrp_set_offset(u32 insn, s32 offset)
>  {
>  	BUG_ON(!aarch64_insn_is_adrp(insn));
> -	return aarch64_insn_encode_immediate(AARCH64_INSN_IMM_ADR, insn,
> +	return __aarch64_insn_encode_immediate(AARCH64_INSN_IMM_ADR, insn,
>  						offset >> 12);
>  }

[Severity: Medium]
Can these four noinstr helpers still call into ftrace_likely_update() when
CONFIG_TRACE_BRANCH_PROFILING is enabled?

Although f22c81bebf8b ("arm64: Disable branch profiling for all arm64 code")
added -DDISABLE_BRANCH_PROFILING to arch/arm64/Kbuild, arch/arm64/lib/ is
built via libs-y in arch/arm64/Makefile rather than arch/arm64/Kbuild, so
arch/arm64/lib/insn.o is compiled without DISABLE_BRANCH_PROFILING.

As a result, the generic BUG_ON() in aarch64_insn_adrp_get_offset() and
aarch64_insn_adrp_set_offset() expands to unlikely(), and the if statements
in aarch64_get_branch_offset() and aarch64_set_branch_offset() are
instrumented under CONFIG_PROFILE_ALL_BRANCHES.  When patch_alternative()
calls these helpers while patching live instructions before cache
maintenance completes, they call the instrumentable ftrace_likely_update().

Passing -DDISABLE_BRANCH_PROFILING in arch/arm64/lib/Makefile (or defining
DISABLE_BRANCH_PROFILING in arch/arm64/lib/insn.c) would prevent branch
profiling instrumentation here.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261107071339.17467-1-jszhang@kernel.org?part=4


      reply	other threads:[~2026-11-08  7:44 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-07  7:13 [PATCH 0/4] arm64: Fix instrumentation escapes from noinstr path Jisheng Zhang
2026-11-07  7:13 ` [PATCH 1/4] arm64: percpu: Always inline helpers used by noinstr code Jisheng Zhang
2026-11-08  7:43   ` sashiko-bot
2026-11-07  7:13 ` [PATCH 2/4] arm64: entry: Always inline helpers used by non-instrumentable paths Jisheng Zhang
2026-11-08  7:43   ` sashiko-bot
2026-11-07  7:13 ` [PATCH 3/4] arm64: stacktrace: Always inline internal stack boundary helpers Jisheng Zhang
2026-11-07  7:13 ` [PATCH 4/4] arm64: alternatives: Make branch and ADRP offset helpers noinst Jisheng Zhang
2026-11-08  7:43   ` sashiko-bot [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=sashiko-outbox-203636@kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=andreyknvl@gmail.com \
    --cc=catalin.marinas@arm.com \
    --cc=cl@gentwo.org \
    --cc=dennis@kernel.org \
    --cc=dvyukov@google.com \
    --cc=glider@google.com \
    --cc=jszhang@kernel.org \
    --cc=kasan-dev@googlegroups.com \
    --cc=linux-mm@kvack.org \
    --cc=mark.rutland@arm.com \
    --cc=ryabinin.a.a@gmail.com \
    --cc=sashiko-reviews@lists.linux.dev \
    --cc=tj@kernel.org \
    --cc=vincenzo.frascino@arm.com \
    --cc=will@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox