So I'm testing my virtual hub drivers.
One problem I have is EP allocation.
As I mentioned earlier, I have a shared pool of endpoints shared
between multiple UDCs.
I've added code to "allocate" those EPs from my own udc->match_ep
callback. That works ok so far. When binding legacy mass storage,
I see it doing the EP autoconfig and grabbing those.
However, I've also added a hook to usb_ep_autoconfig_release() to
"recover" my EPs, and send them back to the pool. Finally, I've also
modified usb_ep_autoconfig_reset() to call the above rather than
manually clear the same fields that usb_ep_autoconfig_release()
clears.
Now, what I observe is that when the mass storage gets bound to the
UDC driver:
- First the endpoints are properly allocated by my match callback,
ie, epautoconf is called to allocate EPs to the function.
- But right away, composite.c calls usb_ep_autoconfig_release()
effectively releasing those 2 endpoints.
And they are never re-claimed again from there. IE. the driver keeps
a reference to those unclaimed endpoints and tries to enable & use
them.
The code that does that is in composite.c:usb_add_config
/* set_alt(), or next bind(), sets up ep->claimed as needed */
usb_ep_autoconfig_reset(cdev->gadget);
Except I observe that nothing calls epautoconf again, which afaik
is the only thing supposed to "claim" the EPs, or am I missing
something here ?
f->set_alt() in f_mass_storage() will enable/disable the EPs but
certainly won't re-claim them. Only bind() will do that but we aren't
getting a new bind() call.
So what am I missing ?
I also notice that nothing else seems to be calling either release
or reset on the EPs... from my understanding that should be done either
by the function ->unbind (since that's where they are requested in the
first place) or by the callers of that...
Cheers,
Ben.
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
