Bug 666687

+++ This bug was initially created as a clone of Bug #498228 +++

Description of problem:
Trying to scan images using xsane with a Brother scanner fails when running as non-root user. After accepting the program license an error is shown:

Failed to open device `brother3:bus1;dev1': Error during device I/O.

Version-Release number of selected component (if applicable):
xsane-0.991-5.el5
Brother MFC-290C scanner with brscan3-0.2.5-2 driver

How reproducible:
Always on that particular hardware

Steps to Reproduce:
1. Start xsane as normal user
2. Accept the user licence

  
Actual results:
Error dialog:
Failed to open device `brother3:bus1;dev1': Error during device I/O.

Expected results:
xsane starts and works

Additional info:
Running xsane as root work properly and I'm able to scan images and save the resulting files.

The device is connected via USB and the above error points me to:
$ ls -l /dev/bus/usb/001/001 
crw-r--r-- 1 root root 189, 0 29 апр 16,28 /dev/bus/usb/001/001

This appears to be access permissions issue.

SELinux is in enforcing mode but no denials occur /it should not be SELinux issue/.

brscan3 is a sane backend driver from the manufacturer and it works apparently when running as root.

--- Additional comment from atodorov on 2009-04-29 17:32:12 EEST ---

This is preventing regular users to use the scanner locally. They have to execute xsane with root parmissions which is not desirable. 

I can do the testing since I have the device required.

--- Additional comment from pknirsch on 2009-05-12 16:30:26 EEST ---

Isn't the device created by udev? If so then i think this is rather a udev problem than a xsane problem and we should change the component.

Thanks & regards, Phil

--- Additional comment from atodorov on 2009-05-12 16:40:28 EEST ---

(In reply to comment #2)
> Isn't the device created by udev? If so then i think this is rather a udev
> problem than a xsane problem and we should change the component.
> 

I wasn't quite sure which was the correct component. Feel free to change it if you think udev is the correct one.

--- Additional comment from pknirsch on 2009-05-12 18:28:36 EEST ---

Adding Harald Hoyer (udev maintainer) to the bug so he can comment on it before we possibly switch the component.

Thanks & regards, Phil

--- Additional comment from harald on 2009-05-12 18:48:03 EEST ---

I think the vendor/device ids are missing in /etc/udev/rules.d/60-libsane.rules 

$ rpm -qf /etc/udev/rules.d/60-libsane.rules 
sane-backends-1.0.18-5.el5

You could quickfix it by duplicating an entry with the correct ids.

--- Additional comment from pknirsch on 2009-05-18 15:33:10 EEST ---

For that we'd need the details for that scanner though, putting in NEEDINFO

Thanks & regards, Phil

--- Additional comment from atodorov on 2009-05-18 15:50:14 EEST ---

What info do you need and where/how to find it? What should be inside /etc/udev/rules.d/60-libsane.rules ?

--- Additional comment from pknirsch on 2009-05-18 18:18:29 EEST ---

The file /etc/udev/rules.d/60-libsane.rules can be found in the sane-backends package.

The format looks like this:

# Brother FAX-2900
SYSFS{idVendor}=="04f9", SYSFS{idProduct}=="0117", SYMLINK+="scanner-%k"

So we basically need the SYSFS vendor ID and the SYSFS product ID.

I just checked and the Brother MFC-290C doesn't seem to be in there yet.

So i've moved this bug to the sane-backends component as the change needs to go in there.

Thanks & regards, Phil

--- Additional comment from atodorov on 2009-05-19 11:06:30 EEST ---

Created attachment 344586 [details]
patch for /etc/udev/rules.d/60-libsane.rules

The above patch works for me. I'm able to start xsane as normal user.

--- Additional comment from pknirsch on 2009-05-19 15:17:46 EEST ---

Thanks Alexander! Patch looks good, granting Devel ACK.

Regards, Phil

--- Additional comment from pknirsch on 2009-05-21 19:23:09 EEST ---

Hm. As sane-backends is not an approved component and the deadline for FT erratas for RHEL-5.4 has run out i'll be reflagging it to RHEL-5.5 FT so that we won't forget about it for 5.5.

Thanks & regards, Phil

PS: And yes, imo the change is small and simple enough to make this a viable FT errata.

--- Additional comment from pknirsch on 2009-12-22 16:54:23 EET ---

Moving to RHEL-5.6 FT.

Thanks & regards, Phil

--- Additional comment from pm-rhel on 2010-10-05 11:04:34 EEST ---

This request is not going to be addressed in RHEL 5.6 FasTrack.

--- Additional comment from atodorov on 2010-10-05 11:28:02 EEST ---

Update summary and request for 5.7.