> > int auerchain_unlink_urb(struct auerchain *acp, struct urb *urb)
> > + spin_unlock_irqrestore(&acp->lock, flags);
> > + dbg("unlink waiting urb");
> > + urb->status = -ENOENT;
> > + urb->complete(urb);
> > Please no. Don't do this.
> >
> > I am afraid there are severe layering violations in this code.
>
> Can you explain? What do you mean? What should I do instead?
If you feel you need to emulate functionality the core should have, the
core should have that functionality. So go hence and expand the core :-)
For anything but usbcore to mess with the internals of an URB is
unacceptable.
But in your case I am not so sure about the need. It seems to me that you
need an approach with a little more abstraction. Something like:
struct chain_head {
struct device_descriptor *dc;
struct chain_element *first;
}
struct chain_element {
struct chain_element *next;
struct chain_head *chain;
struct urb *urb;
}
void complete_chain_element (struct urb *u)
{
struct chain_element *ce = (struct chain_element *)u->context;
if (u->status) {
error_in_chain(ce);
} else {
if (ce->next)
submit_chain_element(ce->next);
else
chain_done(ce->chain);
}
}
Just to give you an idea.
Regards
Oliver
-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
