From: "Rui Qi" <qirui.001@bytedance.com>
To: <palmer@dabbelt.com>, <pjw@kernel.org>, <aou@eecs.berkeley.edu>,
<alex@ghiti.fr>, "Conor Dooley" <conor@kernel.org>
Cc: <linux-riscv@lists.infradead.org>, <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH] riscv: fix frame pointer in call_on_irq_stack for RV32
Date: Mon, 22 Jun 2026 10:54:25 +0800 [thread overview]
Message-ID: <d3aeae54-9b87-4a9c-b8a6-9769a858ff63@bytedance.com> (raw)
In-Reply-To: <20260603035305.564823-1-qirui.001@bytedance.com>
Hi Palmer, Paul, Albert, Alexandre,
Gentle ping on this patch. It has been posted for over two weeks
now and I would appreciate any feedback.
This fixes a frame pointer bug in call_on_irq_stack on RV32 where
the unwinder reads from the alignment padding instead of the
actual saved fp/ra, causing broken stack traces on 32-bit RISC-V.
Please let me know if there are any concerns or if further changes
are needed.
+Cc: Conor Dooley -- adding RISC-V subsystem reviewer.
On 6/3/26 11:53 AM, Rui Qi wrote:
> The frame pointer (s0/fp) in call_on_irq_stack is set using
> STACKFRAME_SIZE_ON_STACK, which equals ALIGN(sizeof(struct stackframe),
> STACK_ALIGN). On RV64, sizeof(struct stackframe) is 16 and the aligned
> size is also 16, so there is no difference. However, on RV32,
> sizeof(struct stackframe) is 8 while STACKFRAME_SIZE_ON_STACK is 16 due
> to 8 bytes of alignment padding.
>
> The stack unwinder does 'frame = (struct stackframe *)fp - 1', which
> reads from 'fp - sizeof(struct stackframe)'. On RV32, with fp set to
> sp + STACKFRAME_SIZE_ON_STACK (sp + 16), the unwinder reads from
> sp + 8, which falls in the alignment padding rather than where the
> saved fp/ra are actually stored (sp + 0 and sp + 4).
>
> Fix this by introducing STACKFRAME_SIZE (the unaligned sizeof) and using
> it for frame pointer setup and restoration, while keeping
> STACKFRAME_SIZE_ON_STACK for the stack pointer allocation/deallocation
> which must remain 16-byte aligned.
>
> Signed-off-by: Rui Qi <qirui.001@bytedance.com>
> ---
> arch/riscv/kernel/asm-offsets.c | 1 +
> arch/riscv/kernel/entry.S | 4 ++--
> 2 files changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/arch/riscv/kernel/asm-offsets.c b/arch/riscv/kernel/asm-offsets.c
> index af827448a609..c1b5f7eb03fd 100644
> --- a/arch/riscv/kernel/asm-offsets.c
> +++ b/arch/riscv/kernel/asm-offsets.c
> @@ -500,6 +500,7 @@ void asm_offsets(void)
> OFFSET(SBI_HART_BOOT_TASK_PTR_OFFSET, sbi_hart_boot_data, task_ptr);
> OFFSET(SBI_HART_BOOT_STACK_PTR_OFFSET, sbi_hart_boot_data, stack_ptr);
>
> + DEFINE(STACKFRAME_SIZE, sizeof(struct stackframe));
> DEFINE(STACKFRAME_SIZE_ON_STACK, ALIGN(sizeof(struct stackframe), STACK_ALIGN));
> OFFSET(STACKFRAME_FP, stackframe, fp);
> OFFSET(STACKFRAME_RA, stackframe, ra);
> diff --git a/arch/riscv/kernel/entry.S b/arch/riscv/kernel/entry.S
> index d011fb51c59a..e8b654e2b7b5 100644
> --- a/arch/riscv/kernel/entry.S
> +++ b/arch/riscv/kernel/entry.S
> @@ -383,7 +383,7 @@ SYM_FUNC_START(call_on_irq_stack)
> addi sp, sp, -STACKFRAME_SIZE_ON_STACK
> REG_S ra, STACKFRAME_RA(sp)
> REG_S s0, STACKFRAME_FP(sp)
> - addi s0, sp, STACKFRAME_SIZE_ON_STACK
> + addi s0, sp, STACKFRAME_SIZE
>
> /* Switch to the per-CPU shadow call stack */
> scs_save_current
> @@ -399,7 +399,7 @@ SYM_FUNC_START(call_on_irq_stack)
> scs_load_current
>
> /* Switch back to the thread stack and restore ra and s0 */
> - addi sp, s0, -STACKFRAME_SIZE_ON_STACK
> + addi sp, s0, -STACKFRAME_SIZE
> REG_L ra, STACKFRAME_RA(sp)
> REG_L s0, STACKFRAME_FP(sp)
> addi sp, sp, STACKFRAME_SIZE_ON_STACK
next prev parent reply other threads:[~2026-06-22 2:54 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-06-03 3:53 Rui Qi
2026-06-22 2:54 ` Rui Qi [this message]
2026-06-22 18:08 ` Nam Cao
2026-06-24 11:19 ` Rui Qi
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=d3aeae54-9b87-4a9c-b8a6-9769a858ff63@bytedance.com \
--to=qirui.001@bytedance.com \
--cc=alex@ghiti.fr \
--cc=aou@eecs.berkeley.edu \
--cc=conor@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-riscv@lists.infradead.org \
--cc=palmer@dabbelt.com \
--cc=pjw@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
all inboxes | Powered by JetHome®