conflict between USB and VGA

[Andrew Muhametshin]

Hello

RELENG_6_20070628
New USB stack (revision 528)
# dmesg -a | egrep "usb2|uscanner|nvidia|CPU:|avail memory"
CPU: AMD Duron(tm) Processor (700.03-MHz 686-class CPU)
avail memory = 511643648 (487 MB)
nvidia0: <GeForce4 MX 440 with AGP8X> mem 
0xdc000000-0xdcffffff,0xd0000000-0xd7ffffff irq 11 at device 0.0 on pci1
nvidia0: [GIANT-LOCKED]
usb2: EHCI version 0.95
usb2: <VIA VT6202 USB 2.0 controller> on ehci0
usb2: 480MBps High Speed USB v2.0
uscanner0: <EPSON EPSON Scanner, class 255/255, rev 2.00/1.00, addr 126>

After scanning, with use of the "Sane" program (through usanner), there 
are extraneous strips on the image, but much more often process of 
scanning hangs(interrupts).
But(!) I have found out, that this behaviour appears only at the working 
Xorg.  Having investigated more in detail this behaviour, I have found 
out dependence between activity during work in the Xwindow and problems 
of scanning.
I have tried to disconnect "Assign IRQ For VGA" in BIOS, but in this 
case the "Sane" program cannot define my scanner, in spite of the fact 
that system (at loading the uscanner-module) defines it.


And as there are two more problems and is possible they as that are 
interconnected:
1.
If I include line "device usb" in my kernel configuration file - my usb 
controller is not defined. But if I build kernel without this line, but 
I load "usb.ko" as the module - in this case usb controller is defined.
2.
Sane program does not define my scanner through the "ugen" device, but 
only through the "uscanner" device. In spite of the fact that system (at 
loading the ugen-module) defines it.


With best regards
Andrew M.

P.S.
I am sorry for my English. I am compelled to use the electronic translator.