On 05/16/12 12:13pm, Alan Stern wrote:
>On Thu, 17 May 2012, Huajun Li wrote:
>
>> On 05/15/12 07:55pm, Oncaphillis wrote:
>> >On 05/16/2012 02:56 AM, Huajun Li wrote:
>> >>Thanks for your reminder, and here is the draft patch for review:
>> >
>> >Didn't do the trck. I patched 3.3.4 .. Should i use 3.4-rc7 ?
>> >
>>
>> From 3.3.4 to 3.4-rc7, devio.c has no much changes, maybe this is
>> not the root cause of your problem.
>>
>> However, I think there still exists another race in destroy_async_on_interface(),
>
>What race are you thinking of?
>
I ever thought the 'as' may be moved from async_pending to hitlist, then
in usbdev_release(), calling destroy_all_async() would not destroy all the 'as'.
However, it is not true. destroy_async_on_interface() can destroy these
'as'.
>> @@ -491,21 +497,22 @@ static void destroy_async(struct dev_state *ps, struct list_head *list)
>> static void destroy_async_on_interface(struct dev_state *ps,
>> unsigned int ifnum)
>> {
>> - struct list_head *p, *q, hitlist;
>> + struct list_head *p, *q;
>> unsigned long flags;
>>
>> - INIT_LIST_HEAD(&hitlist);
>> + INIT_LIST_HEAD(&ps->async_intf);
>> spin_lock_irqsave(&ps->lock, flags);
>> list_for_each_safe(p, q, &ps->async_pending)
>> if (ifnum == list_entry(p, struct async, asynclist)->ifnum)
>> - list_move_tail(p, &hitlist);
>> + list_move_tail(p, &ps->async_intf);
>> spin_unlock_irqrestore(&ps->lock, flags);
>> - destroy_async(ps, &hitlist);
>> + destroy_async(ps, &ps->async_intf);
>> }
>
>What difference does this make?
>
Just want to call destroy_async(ps, &ps->async_intf) again in
destroy_all_async(). but this is not needed as the reason mentioned above.
--
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
