> > Is this only for usb-uhci? A system with a definitely used usb-ohci also
> > shows a zero-usage:
> >
> > usb-ohci 18880 0 (autoclean) (unused)
> > dabusb 53584 2 (autoclean) <--- uses OHCI
> > usbcore 52208 1 [usb-ohci dabusb]
>
> No, it happens with my uhci module as well. So it�s probably a design
> issue....
That behavior is intentional -- it's possible to "rmmod" the
host controller drivers, thereby resetting them. It's handy
for developers of those HCs.
I think the "rmmod -a" style module removal approach is
problematic in any case. Assume for a moment that some
host controller driver is loaded, but there's no USB device
hooked up. That wouldn't make it reasonable to unload
the HCD; users can connect USB devices at any moment.
Similarly, if a driver is loaded, but no usermode software
has bound to it (through a /dev file) and caused its module
use count to increase, it's bad to unload it; a user
might start such an application at any moment.
Now, if some "rmmod" mode knew not to remove drivers
that were bound to hardware, even if the refcount were zero,
such a mode would be reasonable to use in a cron job. Other
than that, I don't see a good fix to such problems.
- Dave
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
