On Sat, Nov 17, 2001 at 03:19:45PM -0500, Justin S. Peavey wrote: > I've copied the stats included in my original message again here for > convenience. > > Windows does not have the same speed problem when running natively > (SMP or UP), but it does when running under SMP & VMWare 3.0, which > uses usbcore.o, usb-uhci.o ro uhci.o to access the USB devices but not > SANE.
Which version of Windows did you use to test with? vmware 3.0 uses usbdevfs for Windows USB accesses when vmware is running Windows on top of Linux, and Linux does not grab the device itself. usbdevfs is not exactly a speed deamon :) So the vmware benchmark is interesting, but doesn't really pertain. (Side note, vmware 3.0 using usbdevfs can let developers sniff the USB protocol of Windows drivers to help in reverse engineering some USB devices) > I'll be doing some packet transfer tests shortly, recompiling kernel > with CONFIG_USB_DEBUG=y now. I'd recommend rebuiding scanner.c after enabling the #define DEBUG line in scanner.h That should give you more info on what is going on. Have you contacted the author of the scanner driver about this? thanks, greg k-h _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-users
