On Sat, 13 Sep 2003, Dmitri Katchalov wrote: > I'm having some problems with usb_reset_device: > > 1) I have this particular device - Atmel at76c503a 802.11b adapter which > requires firmware download followed by a reset. Upon reset device descriptor > does not change however 2 more endpoints appear in config descriptor. > usb_reset_device() only checks for device descriptor and does not reload > configuration and the whole thing goes out of sync with reality. > > 2) If device descriptor is the same then usb_set_configuration() is not > called > and as a result the device is left in USB_STATE_ADDRESS state as set by > a previous call to usb_set_address(). This causes subsequent > usb_submit_urb() > to fail. > > 3) If configuration does change but usb_get_configuration() fails then the > function > returns 1 whereas it looks like it should return negative. > > I wonder if this attempt to keep the current configuration really worth the > efforts.
Dmitri: I'm planning to revise the usb_reset_device() code Real Soon Now. The revised version will compare all the descriptors, not just the device descriptor. But it will take a while to make the changes, and in any case I'm waiting for some of David Brownell's submitted patches (that touch parts of the same code) to be accepted before I begin. Alan Stern ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
