From: Karl Mehltretter <kmehltretter@gmail.com>
To: Helge Deller <deller@gmx.de>,
"James E . J . Bottomley" <James.Bottomley@HansenPartnership.com>
Cc: Karl Mehltretter <kmehltretter@gmail.com>,
Sven Schnelle <svens@stackframe.org>,
Steven Rostedt <rostedt@goodmis.org>,
Masami Hiramatsu <mhiramat@kernel.org>,
Mark Rutland <mark.rutland@arm.com>,
linux-parisc@vger.kernel.org, linux-trace-kernel@vger.kernel.org,
linux-kernel@vger.kernel.org
Subject: [PATCH] parisc: Fix register corruption in 32-bit dynamic ftrace
Date: Sat, 14 Nov 2026 19:02:50 +0100 [thread overview]
Message-ID: <e184f942e0933a1618f161abe672ad8a90bab529.1794672077.git.kmehltretter@gmail.com> (raw)
The 32-bit dynamic ftrace trampoline allocates a 64-byte frame.
ftrace_caller() saves %r26 at offset 20, which overlaps the third
argument's spill slot for ftrace_function_trampoline().
In the failing build, GCC spills org_sp_gr3 into that slot, overwriting
saved %r26 with the original stack pointer. Tracing preempt_count_add()
then passes that pointer as val and corrupts the preempt count.
Allocate two frames on 32-bit, as 64-bit already does, and update the
generated trampoline instruction to match.
Fixes: 6ca6366220ed ("parisc: add dynamic ftrace")
Cc: stable@vger.kernel.org
Signed-off-by: Karl Mehltretter <kmehltretter@gmail.com>
---
Tested with GCC 16.2.0 on a 32-bit SMP kernel under QEMU 10.0.13:
function tracing of preempt_count_add(), preempt_count_sub() and
update_rq_clock(), followed by a combined function/function-graph
workload. No warnings or faults.
I used an LLM to help isolate the failure, inspect the generated code,
write the test harness and prepare the commit message.
arch/parisc/kernel/entry.S | 4 ----
arch/parisc/kernel/ftrace.c | 2 +-
2 files changed, 1 insertion(+), 5 deletions(-)
diff --git a/arch/parisc/kernel/entry.S b/arch/parisc/kernel/entry.S
index e04c5d806c105..b0728dbee53d7 100644
--- a/arch/parisc/kernel/entry.S
+++ b/arch/parisc/kernel/entry.S
@@ -1949,11 +1949,7 @@ ENDPROC_CFI(mcount)
#ifdef CONFIG_DYNAMIC_FTRACE
-#ifdef CONFIG_64BIT
#define FTRACE_FRAME_SIZE (2*FRAME_SIZE)
-#else
-#define FTRACE_FRAME_SIZE FRAME_SIZE
-#endif
ENTRY_CFI(ftrace_caller, caller,frame=FTRACE_FRAME_SIZE,CALLS,SAVE_RP,SAVE_SP)
ftrace_caller:
.global ftrace_caller
diff --git a/arch/parisc/kernel/ftrace.c b/arch/parisc/kernel/ftrace.c
index 10fd5b3e63e75..963a741faa42d 100644
--- a/arch/parisc/kernel/ftrace.c
+++ b/arch/parisc/kernel/ftrace.c
@@ -146,7 +146,7 @@ int ftrace_make_call(struct dyn_ftrace *rec, unsigned long addr)
#else
u32 ftrace_trampoline[] = {
(u32)addr,
- 0x6fc10080, /* stw,ma r1,40(sp) */
+ 0x6fc10100, /* stw,ma r1,80(sp) */
0x48213fd1, /* ldw -18(r1),r1 */
0xe820c002, /* bv,n r0(r1) */
0xe83f1fdf, /* b,l,n .-c,r1 */
base-commit: 7475ea8b58616b33c04afb976cf0fbaad9dd63fb
--
2.39.5 (Apple Git-154)
next reply other threads:[~2026-11-14 18:03 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-14 18:02 Karl Mehltretter [this message]
2026-11-18 22:04 ` Helge Deller
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=e184f942e0933a1618f161abe672ad8a90bab529.1794672077.git.kmehltretter@gmail.com \
--to=kmehltretter@gmail.com \
--cc=James.Bottomley@HansenPartnership.com \
--cc=deller@gmx.de \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-parisc@vger.kernel.org \
--cc=linux-trace-kernel@vger.kernel.org \
--cc=mark.rutland@arm.com \
--cc=mhiramat@kernel.org \
--cc=rostedt@goodmis.org \
--cc=svens@stackframe.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®