xhci: fix possible memleak on setup address fails.
authorIkjoon Jang <ikjn@chromium.org>
Fri, 30 Aug 2019 13:39:16 +0000 (16:39 +0300)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Tue, 3 Sep 2019 13:49:12 +0000 (15:49 +0200)
Xhci re-enables a slot on transaction error in set_address using
xhci_disable_slot() + xhci_alloc_dev().

But in this case, xhci_alloc_dev() creates debugfs entries upon an
existing device without cleaning up old entries, thus memory leaks.

So this patch simply moves calling xhci_debugfs_free_dev() from
xhci_free_dev() to xhci_disable_slot().

[added "possible" to header as this is about failure codepath -Mathias]
Signed-off-by: Ikjoon Jang <ikjn@chromium.org>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Link: https://lore.kernel.org/r/1567172356-12915-5-git-send-email-mathias.nyman@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/usb/host/xhci.c

index e315c0158e90a50e264179c1a5b89853b0c99b21..5008659756877f28b5bdd2dba0f33ae6cfad532e 100644 (file)
@@ -3814,7 +3814,6 @@ static void xhci_free_dev(struct usb_hcd *hcd, struct usb_device *udev)
                virt_dev->eps[i].ep_state &= ~EP_STOP_CMD_PENDING;
                del_timer_sync(&virt_dev->eps[i].stop_cmd_timer);
        }
-       xhci_debugfs_remove_slot(xhci, udev->slot_id);
        virt_dev->udev = NULL;
        ret = xhci_disable_slot(xhci, udev->slot_id);
        if (ret)
@@ -3832,6 +3831,8 @@ int xhci_disable_slot(struct xhci_hcd *xhci, u32 slot_id)
        if (!command)
                return -ENOMEM;
 
+       xhci_debugfs_remove_slot(xhci, slot_id);
+
        spin_lock_irqsave(&xhci->lock, flags);
        /* Don't disable the slot if the host controller is dead. */
        state = readl(&xhci->op_regs->status);