[PATCH] usb: gadget: aspeed_udc: cancel wake work on device removal

[Hyunjung Ko <hj351016@gmail.com>]

ast_udc_wakeup() schedules udc->wake_work to drive remote wakeup from
process context. The work handler ast_udc_wake_work() dereferences udc and
takes udc->lock.

udc is allocated with devm_kzalloc(), so devres frees it after
ast_udc_remove() returns. Removal never cancels wake_work, so a work queued
by a remote-wakeup request just before unbind can run after udc has been
freed and dereference freed memory -- a use-after-free.

Cancel wake_work in ast_udc_remove(), after the gadget is unregistered so
no new work can be queued, before the devm-allocated udc is freed.

Found by static analysis; not runtime-reproduced (needs Aspeed ast_udc
gadget hardware issuing a remote wakeup that races unbind).

Fixes: 055276c13205 ("usb: gadget: add Aspeed ast2600 udc driver")
Cc: stable@vger.kernel.org
Signed-off-by: Hyunjung Ko <hj351016@gmail.com>
---
 drivers/usb/gadget/udc/aspeed_udc.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/drivers/usb/gadget/udc/aspeed_udc.c b/drivers/usb/gadget/udc/aspeed_udc.c
index 81fa495..20dcbd9 100644
--- a/drivers/usb/gadget/udc/aspeed_udc.c
+++ b/drivers/usb/gadget/udc/aspeed_udc.c
@@ -1460,6 +1460,14 @@ static void ast_udc_remove(struct platform_device *pdev)
 	struct ast_udc_dev *udc = platform_get_drvdata(pdev);

 	usb_del_gadget_udc(&udc->gadget);
+
+	/*
+	 * ast_udc_wakeup() schedules wake_work, whose handler dereferences udc.
+	 * The gadget is unregistered above, so no new work is queued; drain a
+	 * pending one before the devm-allocated udc is freed.
+	 */
+	cancel_work_sync(&udc->wake_work);
+
 	if (udc->driver) {
 		/*
 		 * This is broken as only some cleanup is skipped, *udev is
--
2.43.0