| Message ID | 1408057206-8854-1-git-send-email-khoroshilov@ispras.ru (mailing list archive) |
|---|---|
| State | Not Applicable, archived |
| Headers | show |
diff --git a/drivers/net/wireless/at76c50x-usb.c b/drivers/net/wireless/at76c50x-usb.c index 334c2ece855a..da92bfa76b7c 100644 --- a/drivers/net/wireless/at76c50x-usb.c +++ b/drivers/net/wireless/at76c50x-usb.c @@ -2423,8 +2423,6 @@ static void at76_delete_device(struct at76_priv *priv) kfree_skb(priv->rx_skb); - usb_put_dev(priv->udev); - at76_dbg(DBG_PROC_ENTRY, "%s: before freeing priv/ieee80211_hw", __func__); ieee80211_free_hw(priv->hw); @@ -2558,6 +2556,7 @@ static void at76_disconnect(struct usb_interface *interface) wiphy_info(priv->hw->wiphy, "disconnecting\n"); at76_delete_device(priv); + usb_put_dev(priv->udev); dev_info(&interface->dev, "disconnected\n"); }
After commit 174beab7d445 ("at76c50x-usb: Don't perform DMA from stack memory") at76_delete_device() and usb_put_dev() are called both if at76_init_new_device() fails in at76_probe(). But at76_delete_device() does usb_put_dev(priv->dev) itself that means double usb_put_dev(). The patch avoids the problem by moving usb_put_dev() from at76_delete_device() to at76_disconnect(). Found by Linux Driver Verification project (linuxtesting.org). Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru> --- drivers/net/wireless/at76c50x-usb.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)