Skip to contentSkip to conversation

ftrace: fix use-after-free of mod->name in function_stat_show()

PATCH In review Xiang Gao 2 participants 3 messages Apr 16 – Apr 22

Subsystems (2)
lore
Share & tools
lore

Xiang Gao

author
#0 lore raw

PATCH 1/1 ftrace: fix use-after-free of mod->name in function_stat_show()

From: Xiang Gao <[email protected]>

function_stat_show() uses guard(rcu)() inside the else block to hold the RCU read lock while calling __module_text_address() and accessing mod->name. However, guard(rcu)() ties the RCU read lock lifetime to the scope of the else block. The original code stores mod->name into refsymbol and uses it in snprintf() after the else block exits, at which point the RCU read lock has already been released. If the module is concurrently unloaded, mod->name is freed, causing a use-after-free.

Fix by moving the snprintf() call into each branch of the if/else, so that mod->name is only accessed while the RCU read lock is held. refsymbol now points to the local str buffer (which already contains the formatted string) rather than to mod->name, and is only used afterwards as a non-NULL indicator to skip the kallsyms_lookup() fallback.

---
 kernel/trace/ftrace.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)
kernel/trace/ftrace.c +6 -4
index 413310912609..6217b363203c 100644
@@ -559,21 +559,23 @@ static int function_stat_show(struct seq_file *m, void *v)
 		unsigned long offset;
 
 		if (core_kernel_text(rec->ip)) {
-			refsymbol = "_text";
 			offset = rec->ip - (unsigned long)_text;
+			snprintf(str, sizeof(str), "  %s+%#lx",
+				 "_text", offset);
+			refsymbol = str;
 		} else {
 			struct module *mod;
 
 			guard(rcu)();
 			mod = __module_text_address(rec->ip);
 			if (mod) {
-				refsymbol = mod->name;
 				/* Calculate offset from module's text entry address. */
 				offset = rec->ip - (unsigned long)mod->mem[MOD_TEXT].base;
+				snprintf(str, sizeof(str), "  %s+%#lx",
+					 mod->name, offset);
+				refsymbol = str;
 			}
 		}
-		if (refsymbol)
-			snprintf(str, sizeof(str), "  %s+%#lx", refsymbol, offset);
 	}
 	if (!refsymbol)
 		kallsyms_lookup(rec->ip, NULL, NULL, NULL, str);
Signature · 1 line
2.34.1
maintainer
#1 lore raw

The tracing subsystem expects subjects to start with a capital letter:

ftrace: Fix use-after-free of mod-name in function_stat_show()

Quoted 11 lines · On Thu, 16 Apr 2026 16:33:35 +0800 Xiang Gao <[email protected]> wrote:

From: Xiang Gao <[email protected]>

function_stat_show() uses guard(rcu)() inside the else block to hold the RCU read lock while calling __module_text_address() and accessing mod->name. However, guard(rcu)() ties the RCU read lock lifetime to the scope of the else block. The original code stores mod->name into refsymbol and uses it in snprintf() after the else block exits, at which point the RCU read lock has already been released. If the module is concurrently unloaded, mod->name is freed, causing a use-after-free.

Quoted 6 lines

Fix by moving the snprintf() call into each branch of the if/else, so that mod->name is only accessed while the RCU read lock is held. refsymbol now points to the local str buffer (which already contains the formatted string) rather than to mod->name, and is only used afterwards as a non-NULL indicator to skip the kallsyms_lookup() fallback.

Was AI used for any part of this patch? Including finding the bug? If so, it must be disclosed.

Quoted 13 lines
---
 kernel/trace/ftrace.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)
kernel/trace/ftrace.c +0 -0
index 413310912609..6217b363203c 100644
@@ -559,21 +559,23 @@ static int function_stat_show(struct seq_file *m, void *v)
 		unsigned long offset;
 

kernel/trace/ftrace.c static int function_stat_show(struct seq_file *m, void *v) · line 569

Quoted 10 lines
kernel/trace/ftrace.c +3 -1
 		if (core_kernel_text(rec->ip)) {
-			refsymbol = "_text";
 			offset = rec->ip - (unsigned long)_text;
+			snprintf(str, sizeof(str), "  %s+%#lx",
+				 "_text", offset);
+			refsymbol = str;
 		} else {
 			struct module *mod;
 
 			guard(rcu)();

Just move guard(rcu) out of this if statement to include the below reference. No need to make the code worse. This really looks like AI slop :-(

-- Steve

Quoted 15 lines
kernel/trace/ftrace.c +3 -3
 			mod = __module_text_address(rec->ip);
 			if (mod) {
-				refsymbol = mod->name;
 				/* Calculate offset from module's text entry address. */
 				offset = rec->ip - (unsigned long)mod->mem[MOD_TEXT].base;
+				snprintf(str, sizeof(str), "  %s+%#lx",
+					 mod->name, offset);
+				refsymbol = str;
 			}
 		}
-		if (refsymbol)
-			snprintf(str, sizeof(str), "  %s+%#lx", refsymbol, offset);
 	}
 	if (!refsymbol)
 		kallsyms_lookup(rec->ip, NULL, NULL, NULL, str);
author
#2 lore raw
Quoted 2 lines · Steven Rostedt

Was AI used for any part of this patch? Including finding the bug? If so, it must be disclosed.

Yes, AI was used. Claude (claude-opus-4-7) assisted in both finding the bug and drafting the fix. I reviewed the analysis and took responsibility for the submission, but I should have disclosed this up front per Documentation/process/coding-assistants.rst. I apologize for the oversight, and I will add an

Assisted-by Claude:claude-opus-4-7 tag in the follow-up.
Quoted 3 lines

Just move guard(rcu) out of this if statement to include the below reference. No need to make the code worse. This really looks like AI slop :-(

You are right. Hoisting guard(rcu)() to the top of the

  if (tr->trace_flags & TRACE_ITER(PROF_TEXT_OFFSET)) {

block so its scope covers the single snprintf() after the if/else is the correct fix -- +1/-1, net zero, instead of duplicating snprintf() into both branches as I did. I should have recognized this instead of submitting the first plausible-looking approach.

I will send a follow-up patch that restores the single snprintf() after the if/else and hoists guard(rcu)() to cover it, with the Subject capitalized ("ftrace: Fix ...") and

Assisted-by Claude:claude-opus-4-7 added.

Thanks for the review and for pushing back on the approach.

Xiang