[sane-devel] VIA chipset and HP4100C USB scanner

[Henning Meier-Geinitz]

Hi,

On Tue, Jun 29, 2004 at 02:45:51PM -0700, Bill Moseley wrote:
> I've got a MSI KT266-Pro2 motherboard with a VIA chipset.  I'm wondering if
> my hardware is the problem or if there's other things to check.  I'm
> able to sync a USB Palm device with Jpilot without any problems, but
> my scanner generates errors.

Maybe it's hardware but I rather think it's a kernel problem. There
have been many reports about similar trouble with 2.6 kernels and VIA
hardware.

> Then I run sane-find-scanner and it reports:
> 
>     found USB scanner (vendor=0x03f0, product=0x0101) at libusb:001:002
>       # Your USB scanner was (probably) detected. It may or may not be supported by
>       # SANE. Try scanimage -L and read the backend's manpage.
> 
> 
> and in syslog I see:
> 
>     Jun 28 21:30:41 bumby kernel: usb 1-2: control timeout on ep0in
>     Jun 28 21:30:41 bumby kernel: usbfs: USBDEVFS_CONTROL failed cmd sane-find-scann dev 2 rqt 128 rq 6 len 100 ret -110
>     Jun 28 21:30:42 bumby kernel: usb 1-2: control timeout on ep0in
>     Jun 28 21:30:42 bumby kernel: usbfs: USBDEVFS_CONTROL failed cmd sane-find-scann dev 2 rqt 128 rq 6 len 100 ret -110

sane-find-scanner just looks for USB devices. It doesn't send any
device-specific commands (at least in this case). So it's most
probably not a SANE problem.

> Sometimes this works:
> 
>     moseley@bumby:~$ scanimage -L
>     device `hp:libusb:001:002' is a Hewlett-Packard ScanJet 4100C flatbed scanner
> 
> and entries in syslog from scanimage:
> 
>     Jun 28 21:34:11 bumby kernel: Linux video capture interface: v1.00
>     Jun 28 21:34:18 bumby kernel: usb 1-2: control timeout on ep0out
> 
> Not sure why Linux video capture was loaded.  I don't have any video
> capture cards.  lsmod shows "videodev".  Is hotplug confused about what
> module should be loaded?

SANE can also "scan" from video cards that are supported by video4linux.

>     Jun 28 21:37:34 bumby kernel: usb 1-2: control timeout on ep0out
> 
> But sometimes scanimage -L reports that a scanner can't be found, and
> sometimes scanimage -L hangs with the process in "D" status which can't
> be kill -9'ed (requires a reboot, so it seems).  I've also had scanimage
> -L do this:
> 
>     moseley@bumby:~$ scanimage -L
>     Segmentation fault

Bad. I haven't seen reports about such behaviour until now.

> And that's where I'm stuck.  Any suggestions?

Ask the Linux kernel guys. As a workaround, use kernel 2.4.*

> I'm able to scan on my laptop, also runnning Debian Sid and same software.
> But, once in a while I do see similar problems -- so I wonder if it's a
> scanner issue.
> 
> Jun 29 09:29:22 localhost kernel: usbfs: USBDEVFS_CONTROL failed cmd sane-find-scann dev 2 rqt 128 r
> Jun 29 09:29:23 localhost kernel: usb 1-1: control timeout on ep0in

Control timeouts mean that something went wrong with the communication
of the host and the scanner. The reason can be a wrong command, but as
these problems only occur on 2.6 it looks like a kernel issue for me.
Unfortunately nobody has yet found out the real cause.

Bye,
  Henning