From: sashiko-bot@kernel.org
To: "Yunhui Cui" <cuiyunhui@bytedance.com>
Cc: bpf@vger.kernel.org
Subject: Re: [PATCH v8 3/3] riscv: store percpu offset into thread_info
Date: Fri, 03 Jul 2026 12:54:17 +0000 [thread overview]
Message-ID: <20260703125418.573231F000E9@smtp.kernel.org> (raw)
In-Reply-To: <20260703122832.15984-4-cuiyunhui@bytedance.com>
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Missing initialization of `pcpu_offset` in `arch_cpuhp_kick_ap_alive` causes secondary CPUs to use the boot CPU's per-cpu data during early boot, leading to memory corruption.
--
commit cfb659978a0a1c24330d1d94b1f3e36871d66657
Author: Yunhui Cui <cuiyunhui@bytedance.com>
riscv: store percpu offset into thread_info
RISC-V percpu addressing currently derives the base offset from the CPU
number and __per_cpu_offset[]. Cache the current CPU percpu offset in
thread_info so percpu accesses can load it directly. Keep the cached value
up to date for the boot CPU, context switches and secondary CPU bringup.
Link: https://lists.riscv.org/g/tech-privileged/topic/risc_v_tech_arch_review/113437553?page=2
> diff --git a/arch/riscv/kernel/smpboot.c b/arch/riscv/kernel/smpboot.c
> index f6ef57930b50a..7876854d16279 100644
> --- a/arch/riscv/kernel/smpboot.c
> +++ b/arch/riscv/kernel/smpboot.c
[ ... ]
> @@ -191,6 +191,7 @@ int __cpu_up(unsigned int cpu, struct task_struct *tidle)
> {
> int ret;
> tidle->thread_info.cpu = cpu;
> + tidle->thread_info.pcpu_offset = per_cpu_offset(cpu);
>
[Severity: High]
Does this code also need to update arch_cpuhp_kick_ap_alive() for the
CONFIG_HOTPLUG_PARALLEL case?
If CONFIG_HOTPLUG_PARALLEL is enabled, secondary CPUs are booted via
arch_cpuhp_kick_ap_alive() instead of __cpu_up(). Because the idle thread
is forked from init_task, it inherits CPU 0's pcpu_offset.
Without initializing tidle->thread_info.pcpu_offset in
arch_cpuhp_kick_ap_alive(), the secondary CPU will start executing with
CPU 0's per-cpu offset.
Any per-cpu access before __my_cpu_offset is explicitly re-initialized
in smp_callin() will concurrently modify CPU 0's per-cpu state without
locking, which could lead to memory corruption.
> ret = start_secondary_cpu(cpu, tidle);
> if (!ret) {
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260703122832.15984-1-cuiyunhui@bytedance.com?part=3
next prev parent reply other threads:[~2026-07-03 12:54 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-07-03 12:28 [PATCH v8 0/3] riscv: fix PIO helpers and add optimized percpu ops Yunhui Cui
2026-07-03 12:28 ` [PATCH v8 1/3] riscv: io: avoid null-pointer arithmetic in PIO helpers Yunhui Cui
2026-07-03 12:28 ` [PATCH v8 2/3] riscv: introduce percpu.h into include/asm Yunhui Cui
2026-07-03 12:28 ` [PATCH v8 3/3] riscv: store percpu offset into thread_info Yunhui Cui
2026-07-03 12:54 ` sashiko-bot [this message]
2026-07-03 21:02 ` Thomas Gleixner
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=20260703125418.573231F000E9@smtp.kernel.org \
--to=sashiko-bot@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=cuiyunhui@bytedance.com \
--cc=sashiko-reviews@lists.linux.dev \
/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