From: Leizhen Zhang <lzsx618@gmail.com>
To: rostedt@goodmis.org
Cc: mhiramat@kernel.org, mathieu.desnoyers@efficios.com,
linux-trace-kernel@vger.kernel.org, linux-kernel@vger.kernel.org,
Leizhen Zhang <lzsx618@gmail.com>
Subject: [PATCH v3] tracepoint-update: check the section end before reading a string
Date: Mon, 9 Nov 2026 18:01:17 -0600 [thread overview]
Message-ID: <20261110000117.1720256-1-lzsx618@gmail.com> (raw)
for_each_shdr_str() evaluates strlen(str) before checking str < end.
A string section always ends in a NUL, so the last iteration always
steps str to end and then calls strlen() on the bytes just past the
section. This is not limited to malformed input: running the tool on
the vmlinux.o of a plain defconfig build, as "make UT=1" does, takes
that path in both loops that use the macro.
It does no harm today because what follows the section is another part
of the same mapped file, but it is still a read outside the section on
every such run. Check the bound first; the loop is otherwise unchanged.
The problem was found with ASan, and the analysis and fix were
prepared with the help of an AI coding assistant; I reviewed and
tested the result.
Fixes: e30f8e61e251 ("tracing: Add a tracepoint verification check at build time")
Signed-off-by: Leizhen Zhang <lzsx618@gmail.com>
---
Changes in v3:
- Dropped v2 1/2 ("do not reuse section pointers from a previous file"):
Kbuild only ever passes one file per invocation, so no real build can
reach it.
- Reworded the commit message to say where the read actually happens:
I checked with an instrumented copy of the tool, and both loops step
onto the section end when run on a defconfig vmlinux.o.
- Rebased onto v7.4-rc1.
- Replaced the Assisted-by trailer with a sentence in the changelog.
v2: https://lore.kernel.org/r/20261008164117.3670267-2-lzsx618@gmail.com
scripts/tracepoint-update.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/scripts/tracepoint-update.c b/scripts/tracepoint-update.c
index 5cf43c0aac89..c70c0b076554 100644
--- a/scripts/tracepoint-update.c
+++ b/scripts/tracepoint-update.c
@@ -74,7 +74,7 @@ static int add_string(const char *str, const char ***vals, int *count)
#define for_each_shdr_str(len, ehdr, sec) \
for (const char *str = (void *)(ehdr) + shdr_offset(sec), \
*end = str + shdr_size(sec); \
- len = strlen(str), str < end; \
+ str < end && (len = strlen(str), true); \
str += (len) + 1)
base-commit: 6628e7d4c6b25c8d18f1345e3a20d21829aae7e4
--
2.34.1
reply other threads:[~2026-11-10 0:01 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
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=20261110000117.1720256-1-lzsx618@gmail.com \
--to=lzsx618@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-trace-kernel@vger.kernel.org \
--cc=mathieu.desnoyers@efficios.com \
--cc=mhiramat@kernel.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
all inboxes | Powered by JetHome®