[sane-devel] How o make genesys backend make use of symlink via udev-rule?

[Ralph Little]

Hi,
Let's see if I can summarise what you want to do:

1. The scanner works locally via USB using the genesys backend and you can
verify this with scanimage and the device is "genesys:libusb:002:005".

2. The scanner works over the network using saned/net backend which you
verified also using scanimage and the device is "net:192.168.168.245:
genesys:libusb:002:005".

This is where I start to lose track of things because you are changing udev
settings, although you already state that you can access the scanner
locally and via the NET protocol.

What is it that you are trying to do?

Cheers,
Ralph

On Fri, Aug 6, 2021 at 9:51 AM Heinz Rohde <5420108 at mailbox.org> wrote:

> Dear Ralph,
>
>
>
> Hello,
>
> I am very sorry for beeing so short with my detailed information.
>
>
> Here we go:
>
> - The Scanner is a Canon CanoScan LiDE 60
> - The Scanner is
> connected to an ols ThinkPad r60e via USB (All USB ports are USB 2.0)
> -
> The Scanner is handeled by genesys backend.
> - The OS is Debian Buster 64
> bit, all updates installed.
> - The IP number of the Thinkpad is
> 192.168.168.245
>
> - The scanner works locally
> hr at r60e:~$ sudo scanimage -L
> d
> evice `genesys:libusb:002:005' is a Canon LiDE 60 flatbed scanner
>
>
> - The
> scanner works over the network, my 2nd laptop is a Dell E6440 running
> Debian Buster
> hr at e6440:~$ scanimage -L
> device
> `net:192.168.168.245:genesys:libusb:002:005' is a Canon LiDE 60 flatbed
> scanner
>
>
> - This is my udev-rule in /usr/lib/udev/rules.d
>
> less 90-
> scanner.rules
>
> # change device file permission to group scanner for sane
> access
> ATTRS{idVendor}=="04a9", ATTRS{idProduct}=="221c",
> SUBSYSTEMS=="usb", ACTION=="add", MODE="0664", GROUP="scanner",
> SYMLINK+="MyLide60"
>
> but this does not work - it creates no symlink in
> /dev
>
>
>
> - This does work and is my udev-rule in /etc/udev/rules.d
>
> less 83-
> scanner.rules
>
> SUBSYSTEM=="usb", ATTR{product}=="CanoScan",
> ATTR{idProduct}=="221c", ATTR{idVendor}=="04a9", SYMLINK+="myLide60"
> GROUP+="scanner"
>
> It creates a symlink in /dev/
> lrwxrwxrwx  1 root root
> 15 Aug  6 18:18 myLide60 -> bus/usb/002/005
>
>
> It's hardly possilbe to
> document here all the links I visited during the last 5 days, but tere
> are Links that suggested to use
> hr at r60e:~$ sudo scanimage -d
> /dev/myLide60
>
> which obv iously does not work, because a UNIX Device is
> not to be used within the backend systax.
>
> A invocation like
> hr at r60e:~$
> sudo scanimage -d genesys:libusb:/dev/myLide60 results also in an
> error, though it was recommended in a forum blogpost by someone more or
> less savvy.
>
>
> The reason for this is, that I have some windows 10 clients
> in my network.
> They can be used with sane-Scanning-Solution via NAPS2
> Scanning Software and the windows WIA-Sane driver
> https://github.com/mback2k/wiasane/wiki/Downloads
>
> But one need to
> configure the scanning device by hand. Unfortunatly the scanner will be
> assigned deviating USB Bus and Device numbers if the scanner is
> unplugged and replugged or if the scanserver is rebootet. You cannot
> rely on the persistence of USB bus/device number assignment.
>
> That leads
> to the situation, that the users of the windows clients won't be able
> to scan at a certain day or hour noone can predict.
> To assign the then
> valid usb bus / device numbers, you'll have to reinstall the WIA-Sane
> driver each time it is necessary.
>
> Therefore I'd like to have the option
> to configure the windows WIA Sane Driver to the symbolic link.
>
> But until
> today this endeavour was not successful.
>
>
> Important to know:
> SCSI-
> Scanners seem to behave other than USB-Scanners.
> SCSI-Scanners for
> example may have a /dev/sg1, which is a UNIX-Device.
> For SCSI-Scanners
> the symlink may work. I unfortunatly cannot try this out, but several
> postings do claim that.
>
>
> I am now looking for a method to get a
> persistent USB device enumeration for my scanner.
>
> A already tried the
> ALIASES option via dll.aliases.
>
> I wrote this script, which updates the
> dll.aliases in /dev/sane.d
>
> -----
> #!/bin/bash
>
> ALIASFILE=/etc/sane.d/dll.ali
> ases
>
> lsusb | grep LiDE | awk '{ print "alias MyLiDE
> genesys:libusb:"$2":"$4}' | sed 's/.$//' > $ALIASFILE
> -----
>
> The scanner
> then is recognised by its alias, but the windows clients cannot make
> use of it.
> All that the windows WIA Sane Driver sees by autotetection is
> the genesys:libusb:002:005 which I can select as only Option during
> WIA-Sane Setup.
>
>
>
> This is the long version, perhaps you have ideas how I
> should proceed.
>
> THX,
> Heinz
>
> Am Freitag, den 06.08.2021, 07:43 -0700 schrieb Ralph Little:
> > Hi,
> >
> > On 2021-08-04 1:44 p.m., Heinz Rohde via sane-devel wrote:
> > > Hello,
> > >
> > > I created a symlink via udev rule calles MyScanner and points to
> > > /dev/bus/001/002 (for example)
> > >
> > > I cannot get genesys to use this link, or in other words, scanimage
> > > -d
> > > /dev/MyScanner resukts in an error.
> > >
> > The argument to -d is not a device file in the Unix sense, it is a
> > backend-specific string. The format of that string is determined by
> > the
> > backend itself and can vary widely.
> > This is because backends can provide services that are not from a
> > local
> > device, such as the many devices that can be accessed through
> > ethernet
> > or WiFi. Another example would be the test backend which is not
> > associated with any backend. For the test backend, the device would
> > be
> > "test:0" or "test:1".
> >
> > The common way to determine the device name is to run scanimage -L
> > which
> > will probe for available devices and list them out.
> >
> > Can you tell us which model of device you are trying to scan from?
> >
> > Cheers,
> > Ralph
> >
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://alioth-lists.debian.net/pipermail/sane-devel/attachments/20210806/c1a6563b/attachment-0001.htm>