Epson NX515, scanner not found by sane-find-scanner, drakex etc.

[Almost-retired]

New post, meaningful subject ;-)

About 6 months back, the scanner disappeared from my Epson NX515.

Posted about at the time, as a continuance of an older thread, neither of which has been resolved.

When I turn on this MFD, I get this in the logs:

Mar  1 10:50:39 coyote klogd: usb 1-5.4.3: new high speed USB device using ehci_hcd and address 29
Mar  1 10:50:39 coyote klogd: usb 1-5.4.3: New USB device found, idVendor=04b8, idProduct=0856
Mar  1 10:50:39 coyote klogd: usb 1-5.4.3: New USB device strings: Mfr=1, Product=2, SerialNumber=3
Mar  1 10:50:39 coyote klogd: usb 1-5.4.3: Product: USB2.0 MFP(Hi-Speed)
Mar  1 10:50:39 coyote klogd: usb 1-5.4.3: Manufacturer: EPSON
Mar  1 10:50:39 coyote klogd: usb 1-5.4.3: SerialNumber: 4C454C593037313524
Mar  1 10:50:39 coyote klogd: usblp2: USB Bidirectional printer dev 29 if 1 alt 0 proto 2 vid 0x04B8 pid 0x0856
Mar  1 10:50:39 coyote klogd: scsi10 : usb-storage 1-5.4.3:1.2
Mar  1 10:50:40 coyote udev-configure-printer: invalid or missing IEEE 1284 Device ID
Is this the real problem? ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Mar  1 10:50:40 coyote klogd: scsi 10:0:0:0: Direct-Access     EPSON    Storage          1.00 PQ: 0 ANSI: 2
Mar  1 10:50:40 coyote klogd: sd 10:0:0:0: Attached scsi generic sg5 type 0
Mar  1 10:50:40 coyote klogd: sd 10:0:0:0: [sde] Attached SCSI removable disk
Mar  1 10:50:41 coyote klogd: usb 1-5.4.3: usbfs: interface 1 claimed by usblp while 'usb' sets config #1
Mar  1 10:50:43 coyote udev-configure-printer: Re-enabled printer ipp://localhost:631/printers/Epson_Stylus_NX515

So the vid, pid above hasn't changed.
sane-find-scanner does see it:
[root@coyote udev]# sane-find-scanner

  # sane-find-scanner will now attempt to detect your scanner. If the
  # result is different from what you expected, first make sure your
  # scanner is powered up and properly connected to your computer.

  # No SCSI scanners found. If you expected something different, make sure that
  # you have loaded a kernel SCSI driver for your SCSI adapter.

found USB scanner (vendor=0x0403 [FTDI], product=0x6001 [USB FAST SERIAL ADAPTER]) at libusb:001:028
found USB scanner (vendor=0x04b8 [EPSON], product=0x010f [EPSON Scanner 010F], chip=LM9832/3) at libusb:001:026
=======this is the NX515, I have 2 scanners========
found USB scanner (vendor=0x04b8, product=0x0856) at libusb:001:029
==================================
found USB scanner (vendor=0x0403 [FTDI], product=0x6001 [USB HS SERIAL CONVERTER], chip=GT-6816?) at libusb:001:013
  # Your USB scanner was (probably) detected. It may or may not be supported by
  # SANE. Try scanimage -L and read the backend's manpage.

But scanimage -l:
device `v4l:/dev/video0' is a Noname pcHDTV HD3000 HDTV virtual device
device `plustek:libusb:001:026' is a Epson Perfection 1250/Photo flatbed scanner

I called up DrakeX but in the add manual list, the NX515 is conspiciously missing.

Someone suggested I boot the liveCD, but I need a link to DL the latest one.

Thanks, Gene

[pags]

The current (as of now) full KDE iso is here.

Do you have the proper PPD for this model?

*EDIT*

Unverified source for drivers...(only posting this because I do not see anything in Synaptic, nor already installed that is obviously for the NX515):
http://www.avasys.jp/lx-bin2/linux_e/spc/DL1.do

 

...you could test these drivers in the LiveCD environment, without installing.  That way, if anything craps, just shutdown and all is back to normal...

[Almost-retired]

I half expected it to work from the liveCD, and it did.  Perfectly.

The discovery lines in dmesg are identical.

So, before I go off into the sunset poking 50 cal holes in the sky and checking every installed version of this stuff against the CD's version,  is there anything else this tells us?

The kernel I am booted to ATM is one version older,  2.6.38.8-pclos2.pae.bfs, not running 2.6.38.8-pclos3.pae.bfs because something broke the nvidia install for it, and my requests for assistance with that have been ignored.
I have ripped both out, rebooted to the older version, reinstalled the 3 version, and tried several times to get dkms to install the latest driver but it always bails out saying its already installed.  But its borked somehow.

Hopefully, if and when we ever get a much newer kernel, that will resolve itself.  Right now if I try to boot to the pclos3 version, it thinks nvidia is installed, but fails to bring it up, then proceeds to totally trash my /etc/X11/xorg.conf trying to get one of the OS drivers to work, failing and falling back the a tty screen.

This is why I tend to play mix-n-match with my questions.

I don't recall whether the 2.6.38.8-pclos3 on the cd is 'pae.bfs' or not.  bfs is sweet, I have my machine back doing what I want it to do, not 10 seconds later like the other kernels.

The stuff installed that is not from your repo is a short list, none of which walks & quacks like it fits this duck:
Amanda-4.0.0.alpha, built from the tarball, I play canary in the coal mine for amanda for the last 12 years.
Cadsoft's eagle, I occasionally design & mill small printed circuits.

A general sampling of other cad programs, heekscad/heekscnc, freecad, qcad etc etc.
I tried to install xbmc from your repos, but it refuses to start.

None of the above should have anything to do with this scanner.

Thanks, Gene

[AS]

Hi Gene,

try to diff-check the files /etc/sane.d/dll.conf ... between your installation and the livecd ...

About nvidia ... exactly which nvidia driver ? 290.10 ? 295.20 ? 173.x ? 96.x ?

AS

[Almost-retired]

Try to diff-check the files /etc/sane.d/dll.conf ... between your installation and the livecd ...

No need to, there was both an rpmnew, and an rpmsave versions in there.  Manually moving the rpmnew into effect was sufficient to resuscitate it.  So that is fixed I believe.  Thank you very much.

Now, the nvidia.  dkms-nvidia-current-290.10-1pclos2011.i586.rpm is what is currently installed.  That hasn't been updated for several months now, no newer seems to be available.  This one is dated Nov 23, 2011.  No clue if the lack of updates is the pass shutdown or what...  You asked about 295, when & where is that one?

Speaking of pass, I haven't seen an announcement that its back, did I miss that?

Thanks, Gene

[AS]

Try to diff-check the files /etc/sane.d/dll.conf ... between your installation and the livecd ...

No need to, there was both an rpmnew, and an rpmsave versions in there.  Manually moving the rpmnew into effect was sufficient to resuscitate it.  So that is fixed I believe.  Thank you very much.

Now, the nvidia.  dkms-nvidia-current-290.10-1pclos2011.i586.rpm is what is currently installed.  That hasn't been updated for several months now, no newer seems to be available.  This one is dated Nov 23, 2011.  No clue if the lack of updates is the pass shutdown or what...  You asked about 295, when & where is that one?

Speaking of pass, I haven't seen an announcement that its back, did I miss that?

Thanks, Gene

Nvdia 295.20 is actually in testing section, so far has been reported working from a few members, here the thread:
http://www.pclinuxos.com/forum/index.php/topic,102865.0.html

Havent'd experienced myself problems with nvidia drivers, except when self compiling kernels ... in which case I need to unset the following, otherwise the nvidia-module doesn't load:

# CONFIG_MODVERSIONS is not set
# CONFIG_MODULE_SRCVERSION_ALL is not set

AS

[Almost-retired]

Its been out so long I had forgotten that the pclos3 version was in fact a homebrew, built and installed by a script I've been fine tuning on since about Red Hat 6.0, so another automatic build is underway.  NDI why it takes so long on this quad core phenom, but its well over half an hour, and that is when using ccache!  Without ccache, its more like an hour & a half to build a recent kernel.

# CONFIG_MODVERSIONS is not set
# CONFIG_MODULE_SRCVERSION_ALL is not set

I had both set, couldn't recall the second line above, but this build is running, on the naveau(SP?) driver ATM, so at least I have X while the next one is building.

I'll post if this works, otherwise I'll start a fresh thread as I still have another unrelated problem with missing libraries.

Thanks for the clues AS, most appreciated.

Cheers, Gene

[pags]

I half expected it to work from the liveCD, and it did.  Perfectly.
...
The kernel I am booted to ATM is one version older,  2.6.38.8-pclos2.pae.bfs, not running 2.6.38.8-pclos3.pae.bfs because something broke the nvidia install for it, and my requests for assistance with that have been ignored.
I have ripped both out, rebooted to the older version, reinstalled the 3 version, and tried several times to get dkms to install the latest driver but it always bails out saying its already installed.  But its borked somehow.
...

Is there a particluar need for the .pae kernel (do you have more than 4G RAM you need access to)?

...
I don't recall whether the 2.6.38.8-pclos3 on the cd is 'pae.bfs' or not.  bfs is sweet, I have my machine back doing what I want it to do, not 10 seconds later like the other kernels.
...

The LiveCD kernel is .bfs, but not .pae.

...
A general sampling of other cad programs, heekscad/heekscnc, freecad, qcad etc etc.
I tried to install xbmc from your repos, but it refuses to start.
...

IIRC, there were some issues with nVidia and XBMC.  I thought there was a post from Texstar about moving some libs, but I haven't found it, yet.  Did come across this, however (about the 295 drivers).

Its been out so long I had forgotten that the pclos3 version was in fact a homebrew, built and installed by a script I've been fine tuning on since about Red Hat 6.0, so another automatic build is underway.
...

Again, do you have a particluar need for a custom kernel (some esoteric piece of HW, etc. that isn't supported)?  You reported it working from the LiveCD, so I would presume that kernel, at least, is functional for you.

...
NDI why it takes so long on this quad core phenom, but its well over half an hour, and that is when using ccache!  Without ccache, its more like an hour & a half to build a recent kernel.
...

It has been a long time since I've compiled a kernel (not since running Gentoo).  I seem to recall a setting or variable or flag to set the number of threads ("j"?), and the recommendattion (then) was to set it to one more than the number of cores+threads your CPU(s) supported in total.