On Tue, 24 May 2005, Henk wrote: > On Tue, May 24, 2005 at 07:52:28AM +0200, Martin Diehl wrote: > > On Mon, 23 May 2005, Henk wrote: > > > > > > The variant in the kernel module will crash with something like: > > > > ... host controller process error. something bad > > > > ... host controller halted. very bad > > > > > > > > /* KERNEL variant */ > > > > err = usb_control_msg(chip->udev, usb_sndctrlpipe(chip->udev, 0), > > [...] > > > > CMD[URB_INIT],URB_LENGTH, /* init string 16 bytes long*/ > > > > Just a wild guess: are you sure the CMD-buffer is properly kmalloc'ed, in > > contrast to the stack or other location which is invalid for busmastering? > > I've seen this mistake leading to HC-errors on UHCI in the past. > > > > Martin > > > No the cmd strings are staticly defined data strings .... > I never thought of this before, but I'll bet you're absolutely right. > > I will test this tonight.
On x86 it won't make any difference, but for other architectures it's an important distinction. The host controller might try to access locations not present or not mapped to physical RAM, which would cause the sort of errors you saw. Alan Stern ------------------------------------------------------- This SF.Net email is sponsored by Yahoo. Introducing Yahoo! Search Developer Network - Create apps using Yahoo! Search APIs Find out how you can build Yahoo! directly into your own Applications - visit http://developer.yahoo.net/?fr=offad-ysdn-ostg-q22005 _______________________________________________ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
