[sane-devel] Re: connecting to Epson PX-A550 (CX4600)

Olaf Meeuwissen olaf.meeuwissen@avasys.jp
Tue, 12 Apr 2005 09:40:50 +0900

davidsmith@acm.org (David D. Smith) writes:

>  "Olaf" == Olaf Meeuwissen <olaf.meeuwissen@avasys.jp> writes:
>   Olaf> Please note that the epkowa backend is NOT part of
>   Olaf> sane-backends.  It is part of Image Scan! for Linux (and
>   Olaf> Debian's libsane-extras package =  From your log it
>   Olaf> looks like you're using the latter.
> Correct, I compiled from the Debian libsane-extras= sources
> with debugging so I could trace scanimage in gdb. Should I try CVS?

That's probably not going to make much of a difference, seeing that
you already use the sources from unstable.

>   >> [epkowa] send buf, size = 2
>   >> [epkowa] buf[0] 1b .
>   >> [epkowa] buf[1] 40 @
>   >> [sanei_usb] sanei_usb_write_bulk: trying to write 2 bytes
>   >> [sanei_usb] 0000: 1B 40 .@..............  USB error: error
>   >> submitting URB: No such file or directory
>   Olaf> Looks like there is something fishy with your USB.  Does the
>   Olaf> device listed above (in this log libusb:002:014) exist?  That
>   Olaf> is, does /proc/bus/usb/002/014 exist?  What permissions are
>   Olaf> set on /proc/bus/usb/002?
> Yes, the device is present. The permissions are 644 root:root but if I
> run the test as root or if I change the perms to 664 root:users (and
> I'm in the users group), it fails either way with the same log,
> "submitting URB: No such file or directory."

Hmm, the device is there in /proc, yet libusb thinks it isn't.  Weird.
I know /proc is slated to be replaced by /sys, but with 2.6 being the
stable version I would be surprised if /proc is just plain ignored all
of a sudden.

Since you are looking at it with gdb, what file is this URB submitted
to anyway?

>   Olaf> FYI, on Debian testing, your scanner works fine with this
>   Olaf> version of the epkowa backend and kernel version 2.6.8.  #
>   Olaf> Compiled from the Image Scan! for Linux sources.
> I have a feeling it's something fishy with my USB host controller,
> although it has worked perfectly since 2.6.9 with anything I throw at
> it, but it could just be my laptop. Do you know how I can debug this
> further or get you more information? Everything that I can see with
> gdb and strace looks fine.

There is a remote possibility that something is wrong in libusb, but I
tend to agree that there is something fishy with your USB controller.
What happens if you go back to a 2.6.9 kernel?

I have no experience debug kernel drivers so I am afraid you have to
ask somewhere else.
Olaf Meeuwissen                          EPSON AVASYS Corporation, LAN
FSF Associate Member #1962           sign up at http://member.fsf.org/
GnuPG key: 6BE37D90/AB6B 0D1F 99E7 1BF5 EB97  976A 16C7 F27D 6BE3 7D90
Penguin's lib!       -- I hack, therefore I am --               LPIC-2