Skip to contentSkip to conversation

binder: fix possible UAF when freeing buffer

PATCH In review Todd Kjos 2 participants 2 messages Jul 29, 2019 – Jul 31, 2019 Stable branch

lore
Share & tools
lore

Todd Kjos

author
#0 lore raw

PATCH 1/1 binder: fix possible UAF when freeing buffer

From: Todd Kjos <[email protected]>

commit a370003cc301d4361bae20c9ef615f89bf8d1e8a upstream

There is a race between the binder driver cleaning up a completed transaction via binder_free_transaction() and a user calling binder_ioctl(BC_FREE_BUFFER) to release a buffer. It doesn't matter which is first but they need to be protected against running concurrently which can result in a UAF.

---
 drivers/android/binder.c | 16 ++++++++++++++--
 1 file changed, 14 insertions(+), 2 deletions(-)
drivers/android/binder.c +14 -2
index 5d67f5fec6c1b..2decb1a5a8e2f 100644
@@ -1960,8 +1960,18 @@ static struct binder_thread *binder_get_txn_from_and_acq_inner(
 
 static void binder_free_transaction(struct binder_transaction *t)
 {
-	if (t->buffer)
-		t->buffer->transaction = NULL;
+	struct binder_proc *target_proc = t->to_proc;
+
+	if (target_proc) {
+		binder_inner_proc_lock(target_proc);
+		if (t->buffer)
+			t->buffer->transaction = NULL;
+		binder_inner_proc_unlock(target_proc);
+	}
+	/*
+	 * If the transaction has no target_proc, then
+	 * t->buffer->transaction has already been cleared.
+	 */
 	kfree(t);
 	binder_stats_deleted(BINDER_STAT_TRANSACTION);
 }
@@ -3484,10 +3494,12 @@ static int binder_thread_write(struct binder_proc *proc,
 				     buffer->debug_id,
 				     buffer->transaction ? "active" : "finished");
 
+			binder_inner_proc_lock(proc);
 			if (buffer->transaction) {
 				buffer->transaction->buffer = NULL;
 				buffer->transaction = NULL;
 			}
+			binder_inner_proc_unlock(proc);
 			if (buffer->async_transaction && buffer->target_node) {
 				struct binder_node *buf_node;
 				struct binder_work *w;
Signature · 1 line
2.22.0.709.g102302147b-goog
maintainer
#1 lore raw
Quoted 11 lines · Todd Kjos

From: Todd Kjos <[email protected]>

commit a370003cc301d4361bae20c9ef615f89bf8d1e8a upstream

There is a race between the binder driver cleaning up a completed transaction via binder_free_transaction() and a user calling binder_ioctl(BC_FREE_BUFFER) to release a buffer. It doesn't matter which is first but they need to be protected against running concurrently which can result in a UAF.

Quoted 6 lines
---
 drivers/android/binder.c | 16 ++++++++++++++--
 1 file changed, 14 insertions(+), 2 deletions(-)

Thanks for the backport, now queued up.

greg k-h