speedtch usbatm.c,1.20,1.21 usbatm.h,1.13,1.14
rkagan at mail.ru
Fri Apr 29 01:49:28 EDT 2005
On Wed, Apr 27, 2005 at 04:37:53PM +0200, castet.matthieu at free.fr wrote:
> Selon Roman Kagan <rkagan at mail.ru>:
> > No, I thought about using static endpoint number from usbatm_driver.in
> > or .out and poking at the endpoint descriptor after .bind() (that can't
> > be done before .bind() because it may need to choose the proper
> > altsetting).
> You can do it in usb_probe so before the .bind(). That how it was done in my iso
> pipe patch.
Well, yes, but I thought that we considered driver's probe as workaround
for the usb core API which didn't allow for easy creation of extra
intermediate layers, which usbatm core was: if we could figure out the
subdriver in usbatm_usb_probe() without much hassle we wouldn't need
subdriver's usb_driver.probe() methods at all, only
usbatm_driver.bind(). So, generally speaking, the driver's probe should
only be a one-liner calling usbatm_usb_probe().
More information about the Usbatm