Something I want (eventually) to see between steps 2 & 4 is notifying the hcd to clear each endpoint's configuration data. That state is known to hardware (OHCI and EHCI) so changing it needs hardware synchronization.
Interesting. Could you point me at the code in question?
The 2.4 OHCI code has it in sohci_free_dev(), which is where it's always been but let's not restart that flamewar.
The "hcd" style drivers have that in the free_config() code. But again, usbcore doesn't call that at the right places, so there are problems.
The same problem comes up with (re)setting an interface to some altsetting. There's no probe/disconnect relevant there since the interface doesn't go away, but the hardware does need to clear out its state.
How about usb_reset_device() ?
That needs to clear out the same kind of state as changing a configuration, yes.
To do so it seems to me, we need another entry in the bus operations. What do you think?
Yes, I agree. I think a blocking "clear_state(dev,endpoint)" would be the simplest solution all around. Call it for each endpoint in the interface, or configuration, as needed.
- Dave
Regards Oliver
------------------------------------------------------- This SF.net email is sponsored by: Scholarships for Techies! Can't afford IT training? All 2003 ictp students receive scholarships. Get hands-on training in Microsoft, Cisco, Sun, Linux/UNIX, and more. www.ictp.com/training/sourceforge.asp _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
