于 2012年12月17日 09:37, Chen Gang 写道:
> 于 2012年12月16日 01:04, Alan Stern 写道:
>> Also, don't forget that the first think usb_hcd_giveback_urb does is
>> set urb->hcpriv to NULL.
>>
>
> in finish_urb:
> when call usb_hcd_giveback_urb, need unlock firstly (lock again, after finish).
> kfree urb->hcpriv in lock protected, but set it NULL out of lock protected.
> (at least, it seems not a good idea)
> (since we already agree to sending a patch, it is not necessary to reply).
>
oh, sorry, I forgot 2 things:
I originally said "not find set urb->hcpriv to NULL within finish_urb".
it is incorrect (as your reply: usb_hcd_giveback_urb does in finish_urb)
it is my fault.
:-)
are all kfree for urb->hcpriv in lock protected ?
(set urb->hcpriv = NULL in usb_hcd_giveback_urb, not in lock protected).
can you be sure that it is no synchronization issue ?
if you can be sure
would you please explain to me (I am learning).
else
I will check it in details to make sure whether it can cause issue, or not.
thanks.
--
Chen Gang
Asianux Corporation
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
