Am Dienstag, 22. September 2009 12:31:01 schrieb Alan Cox:
> > Any other suggestions?
>
> For those cases the last set of patches merged added
>
> extern inline int tty_port_users(struct tty_port *port)
> {
> return port->count + port->blocked_open;
> }
>
> however using that without care (eg without holding the port lock) will
This makes it hard to use for USB drivers. The need to hold a mutex
was the precise reason the sierra driver got the opened flag protected
with a spinlock.
> leave races. There is of course no reason the USB layer shouldn't either
> keep another flag that indicates what USB needs to know or just adjust
> where it sets ASYNCB_INITIALIZED.
>
> However - you need to think hard about what the rules are for the URB
> retransmission versus a parallel hangup/close/open sequence. Trying to do
> it in the interrupt handler seems rather bogus when the URB could be
> cancelled in the close down call ?
As long as an URB resubmits only itself that is safe, because in the close
down call usb_kill_urb() is used which prevents a resubmission. If a completion
handler submits another URB, as at least one driver does, things get hairy.
It seems to me that you best set a flag for an ongoing closure for these
drivers.
In fact for the common case it would be simplest but brutal to unconditionally
resubmit and be prepared to deal with -EPERM returned from usb_submit_urb()
Regards
Oliver
--
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
