:::: Its less likely this is a permission thing for  $USER (you) under X  ::::
(Though maybe its one of those console-kit pieces of helpful rolling breakage)
The  include snippet _may_ help ..

However  there may be something here and you can test (see: next rambling post)

I was checking that input was working and devtmpfs was _ensured_ .

Which should mean "input" is working under /dev/ etc.
(not relying on say X to start something or  some other helpful complicator.)

For clarity 
I should have said:
What does mount say about option mode=nnn  and devtmpfs or maybe
(little old school )

cat /proc/mounts | grep devtmpfs
ls -lah /dev/input/mice

mode 755 for the first
notice group access for the ls -lah
But given you have  a /dev tree the above is a wander off to nowhere as :

I'm not in the group input ... mouse works with X
Or plugdev ... ditto

So perhaps another mechanism is at work.
You can check "id"  and add yourself to plugdev and netdev

FWIW

## /etc/X11/xorg.conf.d/30-mouse.conf

Section "InputDevice"
        identifier      "Mouse0"
        driver          "mouse"
        option          "AutoServerLayout"      "on"

        # /dev/input/mice will provide support for hotpluging mouse.
        # >>>Without needs to restart X server.<<<< 
        option  "device"        "/dev/input/mice"
        #option "device"       "/dev/input/mouse0" 
        #option "device"        "/dev/input/psaux"
        # To config fancy stuff like disable acceleration,
        # put into your .xinitrc something like 'xset m 1 1 &'
EndSection

(And the "AutoServerLayout on" probably  makes that reliable )
(also the psaux  is not likely

There is genius

It has a command line only mode.
I think it can also work in RPN
Its way more than I need.

#### wandering off topic #####
units is hours of fun

Also I dont think I'm alone in firing up python 
for some simple stuff. (which is all I need)

hmmm or most  interpreters actually.
Though watch out for divide by zero on some.

or even plain old bash:

ANS=$(( 2 * 2 ))
echo $ANS

oh and emacs of course it has a -few-  calcs
(someone had to say it)

Bug fixed

Hopefully its useful.

It may be that you do not have the "contrib" and "non-free"  components  enabled for your install.
synaptic can do that for you.
It is  "non-free" accept no other spelling. .

aside:

/aside

Meantime here is a walk through of one method of finding and installing software.

1./ lspci (list hardware on your PCI buses.)
     you visually grepped out the atheros bit...

2./ use YOUR system tools first
     apt-cache  is a good place to start or synaptic => search
     apt-cache search  foo
     apt-cache show fooz
There is also :
     dpkg -S packagefooz
     dpkg -L packagefooz
There are man pages or even plain old --help -h

synaptic actually uses these in the background ... well OK sometimes they are built in versions ...

3./ look to your local  devuan
    /usr/share/doc/fooz/ or fooz-doc
    apropos fooz
    man fooz  and or info fooz
   
4./ look upstream debian and then :
     There are links that may be helpful
     In this case as its firmware less useful
     But you may find information that helps with the last step.

5./ your favourite search engine
(rinse repeat)
Follow barely understood instructions down cut and pasted roads to confusion.
Drink too much cool aid and eventually realise NO ONE really understands all of this.
Not even Linus

Always backup any config files and you will hardly ever break stuff.
I break stuff all of the time.
Its how I  learn things best :  by making mistakes and f i x i n g them

minimalish example

1 [Wed Jun 28 05:45 PM] pete@mingus:~$ lspci

2 [Wed Jun 28 05:45 PM] pete@mingus:~$ apt-cache search atheros
firmware-atheros - Binary firmware for Atheros wireless cards

3 [Wed Jun 28 05:46 PM] pete@mingus:~$ apt-cache show firmware-atheros
Package: firmware-atheros
Source: firmware-nonfree
Version: 20161130-3
Installed-Size: 8047
Maintainer: Debian Kernel Team <debian-kernel@lists.debian.org>
Architecture: all
Suggests: initramfs-tools
Description: Binary firmware for Atheros wireless cards
Description-md5: cc64d612094e73015861cf2915345869
Multi-Arch: foreign
Homepage: https://git.kernel.org/cgit/linux/kernel/git/firmware/linux-firmware.git
Section: non-free/kernel
Priority: optional
Filename: pool/DEBIAN/non-free/f/firmware-nonfree/firmware-atheros_20161130-3_all.deb
Size: 3131536
MD5sum: 1b4ff70d8293e5e1b2d3480ee9cc6a3b
SHA256: 98ecdada8228a647007a816a30daec33436bb58e53057b5e8f8ad45b9a85b881

There is a second version available to my system that I have snipped.
However this blob of terse info tells you much that is useful and wise.

Including:
     where the source is maintained (kernel.org )
     That it is non-free software (multiple times)
    There are also mentions of "initramfs-tools" "firmware-nonfree"
     and "linux-firmware" 
So Its worth also have an apt-cache search for those just for some context.

Probably all you need to do is:
NOTE I do NOT need this this package and it will be removed
((my atheros is a nice girl and never reads nasty dark code 
She was also cheap.))

The following is edited for sanities sake

4 pete@mingus:~$ sudo apt-get install firmware-atheros
<snips>
The following NEW packages will be installed:
  firmware-atheros
0 upgraded, 1 newly installed, 0 to remove and LOTS not upgraded.
Need to get 3131 kB of archives.
After this operation, 8191 kB of additional disk space will be used.
Get:1 http://au.mirror.devuan.org/merged jessie-backports/non-free i386 firmware-atheros all 20161130-3~bpo8+1 [3131 kB]
Fetched 3131 kB in 12s (260 kB/s)
Selecting previously unselected package firmware-atheros.
(Reading database ... 149922 files and directories currently installed.)
Preparing to unpack .../firmware-atheros_20161130-3~bpo8+1_all.deb ...
Unpacking firmware-atheros (20161130-3~bpo8+1) ...
Setting up firmware-atheros (20161130-3~bpo8+1) ...
 
5  pete@mingus:~$ apropos atheros
atheros: nothing appropriate.

6 pete@mingus:~$ apropos firmware
firmware: nothing appropriate.

7 barry@white:~$ ls /usr/share/doc/fir[TAB][TAB]
firefox-esr/            firmware-intel-sound/   firmware-linux-nonfree/
firmware-amd-graphics/  firmware-linux/         firmware-misc-nonfree/
firmware-atheros/       firmware-linux-free/

7 pete@mingus:~$ ls /usr/share/doc/firmware-atheros/
changelog.Debian.gz  copyright

8 pete@mingus:~$ zless /usr/share/doc/firmware-atheros/changelog.Debian.gz

Which has nothing useful for me at this time.
Proving once again that non-free stuff totally sucks.
However sometimes you have to use it.
The installer may even taken the time to add this module to your initramfs
But if you can boot without it (and unless you booting over your wireless network ....)

try :  ip l     and or   ip a
or finally wih sudo or su :

[2] [root@mingus]: /home/pete$ iwconfig
wlan0     IEEE 802.11  Mode:Master  Tx-Power=20 dBm
          Retry short limit:7   RTS thr:off   Fragment thr:off
          Power Management:off

eth1      no wireless extensions.

lo        no wireless extensions.

br0       no wireless extensions.

eth2      no wireless extensions.

dummy0    no wireless extensions.

eth0      no wireless extensions.

As fsmithred and greenjeans  have said you do not need an antivirus for linux.
(or *BSDs)

_the rest of this is possibly not relevant_ but as I wrote it over my first coffee ...
Here it is anyway. Random observations  thats all.

Its not impossible to write a virus that actually does something,  just very difficult to do without _social__tricks_.
Like convincing you to run your browser or office suite  as root or something startling...
(yeah I guess this could happen, we are only humans. )

The easier attack vector for "Nixen" is where they do their work on networks.
(so packet filtering and simply not running services you don't need).
Or via exploits in existing software or via the package management system or even firmware blobs.
(say in a closed driver for some zappy new IoT thing)

rkhunter
is an _excellent_ well tested tool to scan for , rootkits.

It is particularly useful where you have many users with mixed skill sets.
e.g. you are  running webservers or have shell access accounts (!)  etc.
As _one_ layer of defence along with others.

clamav
is  for scanning email and _attachments_  for windows targeted payloads.
It would also be useful if your file sharing with windows networks etc.
Or collaborating with MSoffice users a lot.

i.e. Clamav looks for _known_ bad payloads in files you probably share..
You can help protect _them_ with clamav.
This works very well.
Now all you need to do is get them to keep _their_ anti virus suites up to date.
(this is not possible)

As a general user running with no special system privileges your
web browser is the weakest link.
( other than _yourself_ and,  say sudo

There are other approaches like tripwire or entire suites of stuff that audit and help validate your system.
Some of them are almost attacks in their own right as they grind your cpu's down to dust... checksumming  away.

These absolutely have a real place and are critical in some use cases...
Most of those assume your software doesn't change too often.
(thats a blurry line)

So real security starts with a clean kernel and a good package manager.

Like Devuan  

Sorry I do rabbit on . .
You only need a virus scanner for other peoples stuff.
(or because your boss/client/user is an box checking idiot)
(which is just a special sub set of the world of stupid really)

The type of use case as discussed above would most probably use a Display Manager in any case.
(and dbus and udev etc)

I am not suggesting that a chopped down xinit type login would have all  conveniences.
I am certainly not suggesting this is a default setting for anybody but those who are curious
or working toward smaller footprints and perhaps developing  debugging.

Its also a dying skill set
(and I'm not excluding myself)
In any case the next post should go some way to covering your absolutely correct note.
(untested as I have no dbus to test )
And probably even buggier now .

I need to trace startx a little more carefully as well.

Its just a hack. Not for user land.

It would not hurt to do the caching machine first if only so that _its_ caching software is up to date.
However it shouldn't really matter.

I _did_ have issues a while ago with apt-cacher-ng and switched to a squid plugin type system called  squid-deb-proxy.
However that was appropriate as I was running squid3 in any case.
Just in case (though unless you want squid3 ... )

You should not have issues with keeping debian devuan ubuntii on <em>either</em>. 
Well I never did with apt-cacher-* and still don't with the squid proxy.
(beyond drive space !)
apt-cacher keeps nice easy to understand mirrors which you can also use as repros
The squid thing is way harder to grok ! Though fully automatic, and near zero thought required after its all working.

For what its worth my home network is a similar size even has the RPies
The proxies are also handy for upgrading _other_ peoples systems (which are usually debian or ubuntu ).
I guess one day I will explore the avahi auto thing for it but there is no need  and I'm always breaking "helpful" guff like avahi

Replying to myself

Xbindkeys looks promising it will even write a config file for you in either plain text or guile
however I think my window manger is grabbing (back) the keys or I have to  log out in again ... no idea

This _should_ let you build you own user specific .xbindkeys
I'm not using TWM (openbox)

apt-get install xbindkeys   
man xbindkeys

has examples