Thread (6 messages) 6 messages, 2 authors, 5d ago

Re: [PATCH v4 2/3] HID: wacom: Fix Use-After-Free in wacom_bamboo_pad

From: sashiko-bot@kernel.org
Date: 2026-06-16 09:43:12

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] Lockless access to `shared->pen` and `shared->touch` in `wacom_mode_change_work()` can lead to Use-After-Free.
--

commit e951961530e94aa8b795598b1d198c32fdbac428
Author: Lee Jones [off-list ref]

HID: wacom: Fix Use-After-Free in wacom_bamboo_pad

Resolves a use-after-free when events are processed while a sibling device
is removed by introducing RCU protection for the shared pen and touch
pointers.
quoted hunk ↗ jump to hunk
diff --git a/drivers/hid/wacom_sys.c b/drivers/hid/wacom_sys.c
index 86895f13dbaea..3990d8d0b40c0 100644
--- a/drivers/hid/wacom_sys.c
+++ b/drivers/hid/wacom_sys.c
[ ... ]
quoted hunk ↗ jump to hunk
@@ -2795,16 +2806,19 @@ static void wacom_mode_change_work(struct work_struct *work)
 	bool is_direct = wacom->wacom_wac.is_direct_mode;
 	int error = 0;
 
-	if (shared->pen) {
-		wacom1 = hid_get_drvdata(shared->pen);
+	struct hid_device *pen = rcu_access_pointer(shared->pen);
+	struct hid_device *touch = rcu_access_pointer(shared->touch);
+
+	if (pen) {
+		wacom1 = hid_get_drvdata(pen);
 		wacom_release_resources(wacom1);
 		hid_hw_stop(wacom1->hdev);
 		wacom1->wacom_wac.has_mode_change = true;
 		wacom1->wacom_wac.is_direct_mode = is_direct;
 	}
[Severity: High]
This is a pre-existing issue, but can this lockless access to the sibling
device pointers lead to a use-after-free if the sibling is concurrently
disconnected?

In wacom_mode_change_work(), rcu_access_pointer() reads the pointer outside
of an RCU read-side critical section. Since the pointer is then dereferenced
to fetch wacom1 via hid_get_drvdata(), what happens if wacom_remove_shared_data()
runs concurrently on the sibling?

The removal path would clear the shared pointer and call synchronize_rcu().
Because wacom_mode_change_work() is not holding the RCU read lock,
synchronize_rcu() will not wait for the work item to complete. This could
allow wacom1's driver data to be freed while wacom_mode_change_work() is
still using it to call wacom_release_resources(wacom1) and hid_hw_stop().

Since wacom_release_resources() and hid_hw_stop() can sleep, simply wrapping
this block in rcu_read_lock() wouldn't be sufficient. Is there a plan to
use reference counting or SRCU here to safely access the sibling device?
-	if (shared->touch) {
-		wacom2 = hid_get_drvdata(shared->touch);
+	if (touch) {
+		wacom2 = hid_get_drvdata(touch);
 		wacom_release_resources(wacom2);
 		hid_hw_stop(wacom2->hdev);
 		wacom2->wacom_wac.has_mode_change = true;
-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260616092658.1714548-1-lee@kernel.org?part=2
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help