Am Mittwoch, 10. Juli 2002 09:30 schrieb Thomas Sailer:
> On Tue, 2002-07-09 at 23:48, Oliver Neukum wrote:
> > Here we have an open with no protection at all against reentrancy.
> > The comment is absolutely bogus as open can sleep. Very bad.
>
> What problem is there if usbdev_open is reentered? You failed to explain
> this. I agree that the GFP_KERNEL is a potential problem.
Here:
spin_lock_init(&ps->lock);
INIT_LIST_HEAD(&ps->async_pending);
INIT_LIST_HEAD(&ps->async_completed);
init_waitqueue_head(&ps->wait);
init_rwsem(&ps->devsem);
you unconditionally reinit locks taht might already be in use.
Tasks blocking on those are doomed.
Regards
Oliver
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Two, two, TWO treats in one.
http://thinkgeek.com/sf
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
