From: Andrew Donnellan <ajd@linux.ibm.com>
To: linuxppc-dev@lists.ozlabs.org, linux-kernel@vger.kernel.org,
maddy@linux.ibm.com, mpe@ellerman.id.au,
christophe.leroy@csgroup.eu, peterz@infradead.org,
jpoimboe@kernel.org, jbaron@akamai.com
Cc: npiggin@gmail.com, rostedt@goodmis.org, ardb@kernel.org,
Erhard Furtner <erhard_f@mailbox.org>
Subject: [PATCH RFC] powerpc: Panic on jump label code patching failure
Date: Fri, 5 Sep 2025 16:11:35 +1000 [thread overview]
Message-ID: <20250905061135.1451362-1-ajd@linux.ibm.com> (raw)
If patch_branch() or patch_instruction() fails while updating a jump
label, we presently fail silently, leading to unpredictable behaviour
later on.
Change arch_jump_label_transform() to panic on a code patching failure,
matching the existing behaviour of arch_static_call_transform().
Reported-by: Erhard Furtner <erhard_f@mailbox.org>
Signed-off-by: Andrew Donnellan <ajd@linux.ibm.com>
---
Ran into this while debugging an issue that Erhard reported to me about my
PAGE_TABLE_CHECK series on a G4, where updating a static key failed
silently, but only for one call site, leading to an incorrect reference
count later on. This looks to be due to the issue fixed in [0]. A loud
failure would have saved us all considerable debugging time.
Should I change the return type of arch_jump_label_transform() and handle
this in an arch-independent way? Are there other users of code patching
in powerpc that ought to be hardened?
Or is this excessive?
[0] https://patchwork.ozlabs.org/project/linuxppc-dev/patch/4b5e6eb281d7b1ea77619bee17095f905a125168.1757003584.git.christophe.leroy@csgroup.eu/
---
arch/powerpc/kernel/jump_label.c | 9 +++++++--
1 file changed, 7 insertions(+), 2 deletions(-)
diff --git a/arch/powerpc/kernel/jump_label.c b/arch/powerpc/kernel/jump_label.c
index 2659e1ac8604..80d41ed7ac50 100644
--- a/arch/powerpc/kernel/jump_label.c
+++ b/arch/powerpc/kernel/jump_label.c
@@ -12,9 +12,14 @@ void arch_jump_label_transform(struct jump_entry *entry,
enum jump_label_type type)
{
u32 *addr = (u32 *)jump_entry_code(entry);
+ int err;
if (type == JUMP_LABEL_JMP)
- patch_branch(addr, jump_entry_target(entry), 0);
+ err = patch_branch(addr, jump_entry_target(entry), 0);
else
- patch_instruction(addr, ppc_inst(PPC_RAW_NOP()));
+ err = patch_instruction(addr, ppc_inst(PPC_RAW_NOP()));
+
+ if (err)
+ panic("%s: patching failed, err %d, type %d, addr %pS, target %pS\n",
+ __func__, err, type, addr, (void *)jump_entry_target(entry));
}
--
2.51.0
next reply other threads:[~2025-09-05 6:12 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-09-05 6:11 Andrew Donnellan [this message]
2025-09-05 6:34 ` [PATCH RFC] powerpc: Panic on jump label code patching failure Christophe Leroy
2025-09-06 3:52 ` Ritesh Harjani
2025-09-06 6:42 ` Christophe Leroy
2025-09-08 1:45 ` Andrew Donnellan
2025-09-09 3:19 ` Ritesh Harjani
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=20250905061135.1451362-1-ajd@linux.ibm.com \
--to=ajd@linux.ibm.com \
--cc=ardb@kernel.org \
--cc=christophe.leroy@csgroup.eu \
--cc=erhard_f@mailbox.org \
--cc=jbaron@akamai.com \
--cc=jpoimboe@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=maddy@linux.ibm.com \
--cc=mpe@ellerman.id.au \
--cc=npiggin@gmail.com \
--cc=peterz@infradead.org \
--cc=rostedt@goodmis.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