On Mon, Dec 03, 2001, Mark S. Mathews <[EMAIL PROTECTED]> wrote: > I'm working on a driver for Intersil Prism2 wireless devices. I'm > having trouble with the device reset in hub.c. > > The Prism device doesn't seem to work properly unless the reset is held > for 500-700 msec. > > Up until now, I've gotten around it by doing my own reset inside my > probe() function. Unfortunately, various versions of usb-uhci and > usbcore seem to really dislike having the driver perform that reset > while other versions don't mind. > > I began looking at the device detection and initialization code in hub.c > and tried modifying the device reset delay to 500 msec and that worked > quite nicely (my probe() reset was disabled). > > Here's the quandry: It's clear that many devices work well with the > device reset code (and delays) that are currently in hub.c. The Prism2 > device is probably an oddity, so I don't think modifying hub.c for this > one device is the right thing to do. Does anyone have a suggestion for > a way to tune the reset delay in a sane manner that would allow the > Prism devices to work correctly _and_ allow things to continue the way > they are? > > The most obvious suggestion is to make the reset delay parameterized via > a module argument (for module compile) and/or Kernel config item (for > static compile). > > If any of the core USB gurus/maintainers think that this is an OK idea, > I'll be happy to submit a patch.
So the device sort of works? Enough to enumerate? But it doesn't work correctly unless you make the delay longer? JE _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
