Thread (9 messages) read the whole thread 9 messages, 3 authors, 2025-05-28

Re: [PATCH 1/2] ftrace: fix UAF when lookup kallsym after ftrace disabled

From: Steven Rostedt <rostedt@goodmis.org>
Date: 2025-05-23 17:54:04
Also in: lkml

On Fri, 23 May 2025 16:39:44 +0800
Ye Bin [off-list ref] wrote:
Above issue may happens as follow:
(1) Add kprobe trace point;
(2) insmod test.ko;
(3) Trigger ftrace disabled;
This is the bug. How was ftrace_disabled triggered? That should never
happen. Was test.ko buggy?
(4) rmmod test.ko;
(5) cat /proc/kallsyms; --> Will trigger UAF as test.ko already removed;
ftrace_mod_get_kallsym()
...
strscpy(module_name, mod_map->mod->name, MODULE_NAME_LEN);
...

As ftrace_release_mod() judge 'ftrace_disabled' is true will return, and
'mod_map' will remaining in ftrace_mod_maps. 'mod_map' has no chance to
release. Therefore, this also causes residual resources to accumulate.
To solve above issue, unconditionally clean up'mod_map'.

Fixes: aba4b5c22cba ("ftrace: Save module init functions kallsyms symbols for tracing")
This is *not* a fix. ftrace_disabled gets set when a bug is triggered. If
this prevents ftrace_disabled from getting set, then it would be a fix. But
if something else happens when ftrace_disabled is set, it just fixes a
symptom and not the bug itself.

quoted hunk ↗ jump to hunk
Signed-off-by: Ye Bin <redacted>
---
 kernel/trace/ftrace.c | 3 ---
 1 file changed, 3 deletions(-)
diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
index a3d4dfad0cbc..ff5d9d73a4a7 100644
--- a/kernel/trace/ftrace.c
+++ b/kernel/trace/ftrace.c
@@ -7438,9 +7438,6 @@ void ftrace_release_mod(struct module *mod)
 
 	mutex_lock(&ftrace_lock);
 
-	if (ftrace_disabled)
-		goto out_unlock;
-
Here you delete the check, and the next patch you have:

+	if (ftrace_disabled || (mod && !mod->num_ftrace_callsites)) {
+		mutex_unlock(&ftrace_lock);
+		return;
+	}
+

Why the two patches where the second patch just adds back the check and
then adds some more stuff around it. This should be a single patch.

Also, why not just keep the goto unlock, that has:

 out_unlock:
	mutex_unlock(&ftrace_lock);

	/* Need to synchronize with ftrace_location_range() */
	if (tmp_page)
		synchronize_rcu();
	for (pg = tmp_page; pg; pg = tmp_page) {

		/* Needs to be called outside of ftrace_lock */
		clear_mod_from_hashes(pg);

		if (pg->records) {
			free_pages((unsigned long)pg->records, pg->order);
			ftrace_number_of_pages -= 1 << pg->order;
		}
		tmp_page = pg->next;
		kfree(pg);
		ftrace_number_of_groups--;
	}
}

And tmp_page is set to NULL before that jump, so the if and for loop will
both be nops.

Why all this extra churn?

-- Steve

 	list_for_each_entry_safe(mod_map, n, &ftrace_mod_maps, list) {
 		if (mod_map->mod == mod) {
 			list_del_rcu(&mod_map->list);
  
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help