Network scanner not detected by xsane after kernel upgrade

[Anthony Buckley]

Hopefully I'm reporting this in the correct area/group. Sorry if not.
After upgrading to 16.04, my scanner is no longer detected by xsane (and other scan apps.) The model is 'Epson WF-3520' multi-function and is connected wirelessly to the network.
The problem occurs on all my computers: A Dell 64 bit desktop, An Asus 64 bit laptop, A Medion 32 bit laptop.
I've upgraded to the latest software released by the manufacturer on 2 computers, but it makes no difference.
Printing works OK and also the scanner is detected if I connect using a USB cable (a second printer is installed though).
The problem only occurs using the 'linux-image-4.4.0...' series of kernels. If I boot with any of the previous kernels available, 'linux-image-4.2.0...' or 'linux-image-3.19.0...' everything works OK.
I am able to ping the scanner.
I've tried a number of suggested fixes from searching around various forums, although many of these relate to 'back-end' scanners connected to a host for network access which is not applicable as mine is connected directly to the network. In any case none have worked.
I have tried the following:
. Removing all manufacturer software and /usr/sane.d/epkowa.conf
. Comment out all entries in the /usr/sane.d/dll (Back-end related)
. Comment out all entries in the /usr/sane.d/dll.conf.conf except 'net' and adding 'epkowa' (Back-end related)
. Disable the firewall (ufw)
. Specifically adding a port to use (1865) (ie. net 192.1... 1865
. Linking the libraries in /usr/lib/sane to /usr/lib/x86_64-linux-gnu/sane

Other than that I'm at a loss and sticking with older kernels.

Regards.

ProblemType: Bug
DistroRelease: Ubuntu 16.04
Package: linux-image-4.4.0-34-generic 4.4.0-34.53
ProcVersionSignature: Ubuntu 4.4.0-34.53-generic 4.4.15
Uname: Linux 4.4.0-34-generic x86_64
ApportVersion: 2.20.1-0ubuntu2.1
Architecture: amd64
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC0: tony 4520 F.... pulseaudio
 /dev/snd/controlC1: tony 4520 F.... pulseaudio
CurrentDesktop: Unity
Date: Sun Aug 14 16:37:16 2016
HibernationDevice: RESUME=UUID=762f657a-d174-4173-b09d-52d580c09a91
InstallationDate: Installed on 2013-10-01 (1047 days ago)
InstallationMedia: Ubuntu 13.04 "Raring Ringtail" - Release amd64 (20130424)
MachineType: ASUSTeK Computer Inc. N71Jq
ProcFB: 0 radeondrmfb
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-4.4.0-34-generic root=UUID=dea23950-d6b3-498c-abba-a94513b58579 ro quiet splash vt.handoff=7
RelatedPackageVersions:
 linux-restricted-modules-4.4.0-34-generic N/A
 linux-backports-modules-4.4.0-34-generic N/A
 linux-firmware 1.157.3
SourcePackage: linux
UpgradeStatus: Upgraded to xenial on 2016-07-30 (14 days ago)
dmi.bios.date: 12/24/2009
dmi.bios.vendor: American Megatrends Inc.
dmi.bios.version: N71Jq.204
dmi.board.asset.tag: ATN12345678901234567
dmi.board.name: N71Jq
dmi.board.vendor: ASUSTeK Computer Inc.
dmi.board.version: 1.0
dmi.chassis.asset.tag: No Asset Tag
dmi.chassis.type: 10
dmi.chassis.vendor: ASUSTeK Computer Inc.
dmi.chassis.version: 1.0
dmi.modalias: dmi:bvnAmericanMegatrendsInc.:bvrN71Jq.204:bd12/24/2009:svnASUSTeKComputerInc.:pnN71Jq:pvr1.0:rvnASUSTeKComputerInc.:rnN71Jq:rvr1.0:cvnASUSTeKComputerInc.:ct10:cvr1.0:
dmi.product.name: N71Jq
dmi.product.version: 1.0
dmi.sys.vendor: ASUSTeK Computer Inc.

[Brad Figg (brad-figg)]

This change was made by a bot.

[Joseph Salisbury (jsalisbury)]

Would it be possible for you to test the latest upstream kernel? Refer to https://wiki.ubuntu.com/KernelMainlineBuilds . Please test the latest v4.8 kernel[0].

If this bug is fixed in the mainline kernel, please add the following tag 'kernel-fixed-upstream'.

If the mainline kernel does not fix this bug, please add the tag: 'kernel-bug-exists-upstream'.

Once testing of the upstream kernel is complete, please mark this bug as "Confirmed".

Thanks in advance.

[0] http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.8-rc2/

[Anthony Buckley (tony-buckley)]

I've tried the above mainline (http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.8-rc2/) on both 64 bit and 32 bit PC's, but the problem has not been fixed. If I get some time and the kernels are available for download I'll try to identify at exactly what kernel it started.

Thanks.

[Anthony Buckley (tony-buckley)]

OK, I've tested a few of the mainlines (binary chop) and as far as I can tell the last kernel that worked OK was 4.3.6-040306-generic (http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.3.6-wily/). So it looks to me as if its since the start of the 4.4... series.

[Anthony Buckley (tony-buckley)]

May be related to bug #1577137

[Anthony Buckley (tony-buckley)]

Hopefully this will be of assistance, in case anyone looks at this bug. I've run the following commands under the current kernel (4.4.0-34-generic) and also a previous kernel (4.2.0-41-generic)

strace scanimage -L log_curr_kernel_not_work 2>&1 Scanner not detected
strace scanimage -L log2_prev_kernel_works 2>&1 Scanner works fine

Files attached.

Diffing the logs shows the interesting stuff at line 3072 (log_curr_kernel_not_work) and line 3232 (log2_prev_kernel_works). It appears that there is an issue connecting under the current kernel that doesn't happen in previous kernels. Why? Beyond me I'm afraid, but its disappointing not to have scanning available now.
Any help appreciated. Thanks.

[Anthony Buckley (tony-buckley)]

Second log for above post

[penalvch (penalvch)]

Anthony Buckley, to keep this relevant to upstream, one would want to test the latest mainline kernel (now 4.8-rc4) as it is released.

Could you please advise?

[Anthony Buckley (tony-buckley)]

4.8-rc5 was the latest when I checked, so I tested this instead. Problem still exists.

[penalvch (penalvch)]

Anthony Buckley, the next step is to fully commit bisect from kernel 4.2 to 4.4 in order to identify the last good kernel commit, followed immediately by the first bad one. This will allow for a more expedited analysis of the root cause of your issue. Could you please do this following https://wiki.ubuntu.com/Kernel/KernelBisection ?

Please note, finding adjacent kernel versions is not fully commit bisecting.

Also, the kernel release names are irrelevant for the purposes of bisecting.

It is most helpful that after the fix commit (not kernel version) has been identified, you then mark this report Status Confirmed.

Thank you for your help.

[Anthony Buckley (tony-buckley)]

Hello Christopher,
Just to clear what you require here. I'm a little confused. I've been thru the Mainline builds and established where I'm having a problem (v4.4...) and not having it (4.3.6-040306-generic).

So the next step next is to run 'git clone...', 'git bisect...' and start compiling kernels? Um, I'll give it a go, but I've not done this before.

I started looking at 'https://wiki.ubuntu.com/Kernel/KernelBisection', followed (with changes) some of the links and ended up at 'https://launchpad.net/ubuntu/xenial/+source/linux' and followed on into some of the Releases (eg. https://launchpad.net/ubuntu/+source/linux/4.4.0-1.15/+build/8880827). I was (initially) under the impression, you wanted me to test these release kernels to establish where my problem started. However installing packages always failed with dependency errors (for example again with 4.4.0-1.15 'linux-headers-4.4.0-1-generic depends on linux-headers-4.4.0-1'). So I guess this isn't what you need and its the git / compile actions.

Could you please confirm?
Regards

[Anthony Buckley (tony-buckley)]

I need a little assistance - anyone. I've managed to get started (I think) and test the first bisect and tried to mark as 'GOOD', but got the following:

git bisect good
Bisecting: 147 revisions left to test after this (roughly 7 steps)
error: Your local changes to the following files would be overwritten by checkout:
        debian.master/changelog
Please, commit your changes or stash them before you can switch branches.
Aborting

What is the correct action, there's nothing in the guide about this?

[Anthony Buckley (tony-buckley)]

Well after many days and heaven alone knows how many hours, I'm beginning to go completely loopy. After getting a number of odd messages (as in above) I finally got the following:

debian/rules.d/4-checks.mk:12: recipe for target 'module-check-generic' failed
make: *** [module-check-generic] Error 1

I tried a number of things like adjusting the HEAD, but nothing changed. So I figured I must have made a mistake somewhere and started again. More odd messages. Not wanting to go down that path again I tried playing a little with the bisect range, but all this showed was what I've always suspected - the problem lies between 4.3.0-7.18 and 4.4.0-1.15. When I try to bisect these to confirm, the 'debian.master' directory gets deleted, which appears to be because they are non-linear (whatever that means). Looking at the instructions, they get pretty vague here. Seems to me they are fine for straightforward things, but quite esoteric when things go awry. Is it telling me that I now need to commit bisect the mainline kernels? It mentions something about mapping one to the other, but this proved confusing and useless.

I'd like some confirmation about this as all I really know where mainline kernels are concerned is that I have no problem at v4.3.6 (http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.3.6-wily) and have a problem at v4.4-rc1+cod1 (http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.4-rc1+cod1-wily).

To be honest, I'm not even sure that posting here is the proper way to seek assistance (I'm starting to suspect it isn't in which case I'm really lost), but will attempt the mainline bisect if I here nothing.

[Anthony Buckley (tony-buckley)]

Apologies in advance for marking this as confirmed, but I'm pretty much out of options. I tried to bisect using mainline kernels (above), but got nowhere. I followed the instructions as given:-

git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
cd linux
git checkout 49f9ae92af54217fd9821a7613342178893a72af (NB also tried v4.3.6 and more)

This commit was from the last known build where I had no problem - v4.3.6 (http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.3.6-wily)

and got this error:-
fatal: reference is not a tree: 49f9ae92af54217fd9821a7613342178893a72af

I've tried lots of searching to try resolve my bisect issues, but its been of limited help. So I'm really stumped and can't proceed with anything, which is disappointing because I'd really like scanner back.
Regards.

[Anthony Buckley (tony-buckley)]

Just set status back - thought of something else to try.

[Anthony Buckley (tony-buckley)]

OK, after all my problems trying bisect (see previous comments) I've finally managed to successfully complete one. See the attached bisect log for details of the last good commit (10cbc8f179177c1a6d5f56a46ebddc8f602ce5ac) and the first bad one (ef547f2ac16bd9d77a780a0e7c70857e69e8f23f).

Quite a job to muddle through. Hopefully its been worth it.

Regards.
Tony

[Anthony Buckley (tony-buckley)]

Sorry just to elaborate on what I did for the above in case its unclear.

git checkout master
git bisect start
git bisect bad v4.4-rc1
git bisect good v4.3

NB. - I built kernels for each of these to confirm the bad / good status of each beforehand.

Then ran built kernels, then ran git bisect good / bad as required until completion.

[penalvch (penalvch)]

Anthony Buckley, could you please advise to the results of 4.9-rc2? If it is still reproducible in this, then this would be an upstream bug also.

[Anthony Buckley (tony-buckley)]

4.9-rc4 is now the latest. Problem not fixed.
Regards.

[penalvch (penalvch)]

Anthony Buckley, the issue you are reporting is an upstream one. Could you please report this problem following the instructions verbatim at https://wiki.ubuntu.com/Bugs/Upstream/kernel to the appropriate mailing list (TO: Eric Dumazet, and David S. Miller, CC netdev)?

Please provide a direct URL to your post to the mailing list when it becomes available so that it may be tracked.

Thank you for your help.

[Anthony Buckley (tony-buckley)]

There have been some developments with this problem so I thought I should
post the latest news.

Firstly, I have reported the problem as directed above to the people concerned
and the netdev team twice, but there has been no response.

Secondly, I managed to narrow down the cause of the problem in the codeline
only to then discover that its a known issue. A suggested patch to fix
the problem was rejected with the expectation that applications would
have to be changed to handle the new kernel network code.

Further details may be viewed here:- https://patchwork.ozlabs.org/patch/618164/

<begin rant> No doubt the sentiments expressed in the post are valid, but they
seem rather inflexible and harsh to me given that the failing applications were
working. Ultimately if an application is not changed, users will be stuck. In
my case its no network scanner. Seems crazy to me when a simple change will
resolve the problem. <end rant>

The application in point here (bug 1613027) is the Epson network plugin.

Unfortunately Epson or the authoring 3rd party do not publish the source code
and my efforts through Epson support to try to raise the issue or trace the
origin of the network plugin have been fruitless.

So for anyone experiencing the problem here, and it looks like there is at least
one, what to do...

1. Do nothing and only have a usb scanner.
2. Download the Linux kernel source code every time there is a new release
    and apply the suggested patch yourself.
3. Override the system 'listen' function with 'dlsym' whenever the
    network plugin is invoked. This is what I have done.
    To this end I have created a repository on my
    git home page here:- https://github.com/mr-headwind/ListenOverride
    This should only be viewed as a temporary workaround, but given the
    circumstances I suspect it may be permanent.

Regards.
Tony

[Zippo (peter-henninger)]

I experienced the same problem, with ubuntu 14.04 last year in November.
My scanner was no longer detected by xsane (and other scan apps.)
The model is 'Canon MX870' multi-function and is connected via LAN with a fix ip-adress to the network. After some weeks the problem disappeared by system updates!

Since I installed ubuntu 16.04 I have the same problem again.
My scanner is not detected by xsane and other scan applications!
It looks like an still unresolved issue.

[penalvch (penalvch)]

Anthony Buckley, to clarify:

1) Does this still happen when using the latest iscan 1.0.4 from http://download.ebz.epson.net/dsc/search/01/search/searchModule following http://download.ebz.epson.net/man/linux/iscan_e.html#sec5 ?
2) What did Epson say specifically when you contacted them?

[Anthony Buckley (tony-buckley)]

Hello Christopher,
Sorry for the delay in responding. I've been away.
1. I tested 1.0.4 shortly after release and I can confirm that the problem still persists. The network plug-in in the package (iscan-network-nt) that I am fairly sure is the problem has not been changed since 2013. This plug-in is proprietary and we don't have source code (that I can find anyway).
2. My interactions with Epson support bordered on being comical.
I first contacted them shortly after I noticed the problem on the off chance they may have had some input. The response was that they do not support Linux and simply provided a link to the driver download page.
I next contacted them when I was pretty sure what the problem was, explaining that I understood there was no Linux support, but that I didn't want support as such. I explained in general terms what was happening, what needed to be done and that I was trying to trace the origins / ownership of the network plug-in or could they follow it up. The response I received was that they are very supportive of Linux and then proceeded to tell me all the options for "printing" under Linux and that my best option was to use GIMP! When I responded (politely) that nowhere in my mail did I mention printing, but that I had mentioned a scanner problem several times, they responded that they thought I should contact my computer support - Linux!

So, in effect, I got nowhere. At least I have a workaround.