Hi, I'm using the linux-2.4.18-pre6 source tree from the bitkeeper repository.
Problem #1: I'm running some tests on a custom device using 1 control transfer followed by 6 bulk-out followed by 6 bulk-in transfers. The bulk transfers are 64 bytes each (max packet size). With uhci.o about 1 in every 4,000 urbs submitted for bulk-in never get their callbacks invoked. I have a CATC and have traces which show the transfer completing with an ACK, so it seems like I should get that callback. With usb-uhci.o I have yet to see this problem. Problem #2: I have another test where I intentionally attempt an unsupported control read transfer. The device STALLs the data and status stage, as expected, but usb-uhci.o returns the urb status as zero, indicating that no problems occurred. uchi.o behaves as I expect, setting the urb status to -EPIPE. I'm not prepared to debug these problems myself, but will help anyone who's interested in fixing them with more info as required. Regards, --gmcnutt _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
