HOWTO Install Samsung Unified Printer Driver

[Kempe]

Printing with the non-legacy drivers now work for me. Thanks for a great job Tweedledee.

[urusha]

It' would be nice to add strings like theese in the head of "Release" file in the repo:

Origin: suldr
Codename: stable

Or some another values. The idea is to make unattended-upgrades work by adding next line to Allowed-Origins in /etc/apt/apt.conf.d/50unattended-upgrades:

"suldr stable";

Values should be without spaces (to properly use it for releases earlier then 10.10 or debian 6). We use your repo on many machines and so, automatic updates should work, I think...
Thanks for attantion and for the great repo.

[cyrillo]

A few days I've bought a Samsung CLX-3185 printer. As it has been suggested here, I have installed the unified driver and had some problems. Because I have nowhere found anything about this problem, I'd like to share my experience.

When you get the following error message:
Stopping job because the scheduler could not execute a filter.

Then have a look in the error-log and you will find something like this:

[Job 80] Started filter /usr/lib/cups/filter/pstops (PID 20165)
Unable to execute /usr/lib/cups/filter/pstosecps: no execute permissions (0100644)
Unable to start filter "pstosecps" - Operation not permitted.
Discarding unused job-state-changed event...
[Job 80] Stopping job because the scheduler could not execute a filter.

It looks like that you just have to change the permission of /usr/lib/cups/filter/pstosecps but this filter-file is not there.

Copy the filter-file from the directory where you have unpacked the Unified-Printer-Driver to the cups filters.

 sudo cp ./cdroot/Linux/i386/at_root/usr/lib/cups/filter/pstosecps /usr/lib/cups/filter/

Then change the permission:

 sudo chmod 755 /usr/lib/cups/filter/pstosecps

With this I brought the printer to work and I'm very happy with my CLX-3185.

I hope my contribution can help someone with the same problem.

[tweedledee]

A few days I've bought a Samsung CLX-3185 printer. As it has been suggested here, I have installed the unified driver and had some problems. Because I have nowhere found anything about this problem, I'd like to share my experience.

Just so I'm clear - how did you install the driver and which version? I believe that file is only associated with the newest drivers, which suggests you installed the package directly from Samsung - and if so, the latest drivers from them appear to be very buggy. (The file you mention is one I had to "fix" during the brief period I tried to package and distribute the latest drivers, for the reason you describe.)

[tweedledee]

It' would be nice to add strings like theese in the head of "Release" file in the repo:

Done. Origin is "suldr" and codename remains "debian" as it has been.

[miltondp]

Hi guys. First of all, thank you very much for your work. It really was very useful for me. Now, I am facing with a problem I would like to share with you, maybe you can help me.

I have two Samsung ML-1665 printers working, connected to the same PC. First, my problem was that Ubuntu couldn't distinguish them. So if I added one printer it works. Although the second was detected too, it had the same device URI, so all printing were send to the first one only. This are some post of people with the same problem:

http://ubuntuforums.org/showthread.php?t=1232097
http://wwww.ubuntuforums.org/showthread.php?p=6715273
http://ubuntuforums.org/showthread.php?t=1390046

Then I could fix it, changing the device URI in /etc/cups/printers.conf. Here there is a post explaining the fix:

http://www.henningweiler.de/2009/01/...ical-printers/

The device URIs look like this: file:/dev/usblp0, etc.

Now, my problem is the following: although both printers are working, sometimes the system change the usb ports, and the usblp0 device points to the printer B, and the usblp1 one to the printer A. This happens randomly. It may be working two months without problems, and then it gets crazy again for a few days.

I saw that Ubuntu puts a "serial" parameter in the device URI for some EPSON and HP printers. Running lpusb -v, I see that my printers have a serial number, but it's not used by CUPS when they are installed. I will try to put it manually, and see what happens, but it does not seems to work, as someone said in some post.

So, any help? Did one of you guys have the same problem?

As I don't have direct access to the machine with these printers installed, what I will be trying as soon as I have the possibility is:

• Try to unload the usblp module, and installing the printers again.
• Update to the latest driver you have posted in the repositories.
• Try to use a different usb port for one of the printers. Right now, both are connected in the front of the PC.
• Try with serial parameter in the device URI.

Thank you.

[tweedledee]

• Try to unload the usblp module, and installing the printers again.
• Update to the latest driver you have posted in the repositories.
• Try to use a different usb port for one of the printers. Right now, both are connected in the front of the PC.
• Try with serial parameter in the device URI.

Thank you.

The first won't be a permanent fix, the second won't help, the third might help depending on how the USB hubs are recognized internally, and the fourth I have no idea.

It sounds to me like what you are really looking for is a set of rules to add to udev so that the USB ports are always numbered exactly the same way, which isn't really something that has to do with printing or CUPS. At one point in time I had to do some of that for a different problem (occasional random renaming of the wireless device upon boot), but the computer that required those modifications died a while back and I no longer seem to have the notes. I would suggest looking for solutions along those lines rather than something specific to CUPS or printing.

[hokiejp]

Ok, I now have scanning working in 10.10 amd64. Here's what I did:

1. Download the source for eglibc 2.10.1 from packages.ubuntu.com
2. Build a 32bit version of eglibc (roughly following the package rules for libc6-i386). This is the hardest part, see my next post for detailed instructions.
3. Installed the new libc in a folder I made called /opt/Samsung/tools/
4. copied the 32 bit version of netdiscovery from the Samsung drivers package to /opt/Samsung/mfp/bin/netdiscovery32
5. Backed up my old netdiscovery to netdiscovery64
6. Wrote a script called /opt/Samsung/mfp/bin/netdiscovery, contents are:
   

   LD_LIBRARY_PATH=/opt/Samsung/tools/lib /opt/Samsung/mfp/bin/netdiscovery32 $*

Then I ran xsane and started scanning random things around my house, just because I finally could. This solution will work as long as one can still build eglibc 2.10.1, has no implications for system stability, and merely wastes about 100MB of disk space. That could probably be reduced by keeping only the files that were strictly required.

[hokiejp]

Build steps for eglibc 2.10.1 (from previous post)

[EDIT: I have not done a careful careful check of the packages required to do this build, but tweedledee reports that build-essential, gcc-multilib, and gawk should be all that are required. If you discover any other dependencies or if the build fails for you for any reason, please contact me and I will update these instructions.]

1. Make a working folder and cd into it
2. Download the sources:

   wget http://archive.ubuntu.com/ubuntu/pool/main/e/eglibc/eglibc_2.10.1.orig.tar.gz

3. Build a 32bit libc with the following steps. This takes a fairly long time to compile.
   

   tar -zxvf eglibc_2.10.1.orig.tar.gz
   mkdir build
   cd build
   CC="gcc-4.4 -fno-stack-protector -U_FORTIFY_SOURCE -m32";export CC
   CFLAGS="-O2 -pipe -fstrict-aliasing -march=i686 -mtune=generic"; export CFLAGS
   ../eglibc-2.10.1/configure --prefix=/opt/Samsung/tools/ --disable-profile --enable-multiarch --host=i686-linux-gnu
   make

4. Make the install folder and install with the following commands. It's probably fine to `make install` as root since the prefix was set in the configure step, but it's definitely safer not to.
   

   sudo mkdir /opt/Samsung/tools
   sudo chown $USERNAME:$USERNAME /opt/Samsung/tools
   make install

[tweedledee]

Ok, I now have scanning working in 10.10 amd64.

This seems to be a solution I can safely package and distribute. Over the next week or so I will look into doing so.