I have a USB PCI card, which shows up as this in `lspci`:
00:09.0 USB Controller: VIA Technologies, Inc. VT82C586B USB (rev 04)
... it appears that they tossed the whole southbridge chip onto a pci
board, and disabled everything but USB. Anyway, this device seems to be
semi-functional under 2.2.18 USB. I have a cheapie IBM usb camera that
works, but a USB scanner that does not -- always gives the following
errors:
usb_control/bulk_msg: timeout
scanner.c: write_scanner: NAK received.
The firmware upload seems to work ok, but nothing else does. The sane
tools actually segfault, and once the kernel oopsed (didn't manage to
get the output :( ).
I noticed a config setting for that chipset, but it was under block
devices, so I left it turned off. Judging form reading the LKML, VIA
chipsets are problematic. Has anyone has any positive experiences
getting this device to do the thing it's supposed to do under Linux? Any
tips?
-M
On 03 Mar 2001 12:54:36 +0100, Vojtech Pavlik wrote:
> No, they have a separate USB chip, but it has the same PCI ID as the
> builtin silicon in the southbridge.
Ah. I went and looked up that chip ID at via's website, and saw only
southbridge chips, no USB-only chips at all. But, my real question was,
is there a way to get it to work right under Linux? I have two machines;
one is an Athlon with built-in (Via) USB support, and it seems to work
ok. I also have a P-200 that's my print server, file server and
(hopefully) scanner server. It is an old HX-chipset pre-USB socket-7
intel machine, but I've put a Via-chipset 2-port USB card into it. I
keep getting timeouts on it. I've looked all over the net for clues as
to what I can do to get it to work, and nothing has. Both uhci and
usb-uhci have the same problems. Is there a generic cause for timeout
problems when doing bulk transfers?
On Sat, Mar 03, 2001 at 12:13:49AM -0500, Michael Rothwell wrote:
> I have a USB PCI card, which shows up as this in `lspci`:
>
> 00:09.0 USB Controller: VIA Technologies, Inc. VT82C586B USB (rev 04)
>
> ... it appears that they tossed the whole southbridge chip onto a pci
> board, and disabled everything but USB.
No, they have a separate USB chip, but it has the same PCI ID as the
builtin silicon in the southbridge.
--
Vojtech Pavlik
SuSE Labs