[sane-devel] usblp trouble with all-in-one device

[Olaf Meeuwissen]

Johannes Meixner <jsmeix at suse.de> writes:

> Hello,

Hi!

> On Dec 15 16:27 Olaf Meeuwissen wrote (shortened):
>> Working on support for a new business oriented all-in-one for iscan, I
>> have hit a dead end and would appreciate some insights/suggestions.
>> 
>> When attaching the device, the kernel loads the usblp module and CUPS
>> has no trouble finding the printer part.  On the scanner side however,
>> the device is (usually) found but I can't get it to scan.
>
> I have no knowledge how the USB mess is supposed to work
> but perhaps issue "D)" in
> https://bugzilla.novell.com/show_bug.cgi?id=184798
> might be about the same problem?

That looks like it might be about the same or at least very similar
problem.  Thanks for the link.

>> Considering the target audience for this device, I can't very well say
>> "Oh, just rmmod usblp before you scan and insmod usblp again when you
>> are done".
>
> As far as I know this is exactly what the USB stuff in the
> current kernels requires.
>
>
>> Similarly, suggesting to compile a CUPS usb backend using
>> usb-libusb.c[1] is also not particularly user friendly.
>
> No ;-)
> As far as I know the only solution with the current kernels
> is that the CUPS usb backend uses libusb out of the box.
>
>
>> Is there a way to get SANE, CUPS and usblp to play nice _together_?
>
> As far as I know with the current kernels: No.
>
>
> But perhaps I am totally wrong and the issue is only because
> usblp and libusb try to communiucate with the same USB endpoint
> (i.e. with the whole all-in-one device) instead of usblp
> communicates only with the USB endpointwhitch matches to
> the printer interface of the device and libusb communicates
> only with the USB endpoint whitch matches to the scanner interface
> of the device (as far as I know all Epson all-in-one devices provide
> seperated USB interfaces for the printer unit and the scanner unit).
> But I don't know if such a kind of seperated communication is
> possible at all with the current kernels.

No, both usblp and libusb (try to) talk to different endpoints.  As
matter of fact, I tried with a little test program that explicitly
claims the interface with the scanner endpoints.  I even recompiled my
kernel with USB debugging support to see that usblp claims the
interface with the printer endpoints.

The odd thing is that for a number of other all-in-ones, usblp does
not seem to be (such) a problem.

Thanks for the info.
-- 
Olaf Meeuwissen                          EPSON AVASYS Corporation, SE1
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