[PATCH] xhci: Cancel cmd_timer before cleanup the command queue in xhci_hc_died()

[Jeffy Chen <jeffy.chen@rock-chips.com>]

When the cmd_timer fired, it would try to access the command struct.
So cancel it before cleanup the command queue in xhci_hc_died(), to
avoid use-after-free reported by KASAN:
[  176.952537] BUG: KASAN: use-after-free in xhci_handle_command_timeout+0x68/0x224
[  176.960846] Write of size 4 at addr ffffffc0cbb01608 by task kworker/3:3/1680
...
[  177.180644] Freed by task 0:
[  177.183882]  kasan_slab_free+0x90/0x15c
[  177.188194]  kfree+0x114/0x28c
[  177.191630]  xhci_cleanup_command_queue+0xc8/0xf8
[  177.196916]  xhci_hc_died+0x84/0x358

Fixes: d9f11ba9f107 ("xhci: Rework how we handle unresponsive or hoptlug removed hosts")
Signed-off-by: Jeffy Chen <jeffy.chen@rock-chips.com>
---

 drivers/usb/host/xhci-ring.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/usb/host/xhci-ring.c b/drivers/usb/host/xhci-ring.c
index a9443651ce0f..69ac3deffb7b 100644
--- a/drivers/usb/host/xhci-ring.c
+++ b/drivers/usb/host/xhci-ring.c
@@ -914,6 +914,7 @@ void xhci_hc_died(struct xhci_hcd *xhci)
 	xhci_err(xhci, "xHCI host controller not responding, assume dead\n");
 	xhci->xhc_state |= XHCI_STATE_DYING;
 
+	cancel_delayed_work_sync(&xhci->cmd_timer);
 	xhci_cleanup_command_queue(xhci);
 
 	/* return any pending urbs, remove may be waiting for them */
-- 
2.11.0