XFCE / Thunar (times FIFTEEN CHARS!)

OK - so - I built Kali 2023.3 on a Raspberry Pi 4B (8 GB model) - runs well (also got it running, also arm64, in UTM / QEMU on my MacBook - probably running a tad better there after installing spice tools in the guest).

Default user on the Pi image is / was “kali” - which I didn’t like - but sometimes changing a user name creates all sorts of issues (installing the arm64 Kali image from the install iso asks you what user name you want).

So - I created a new user “x” alongside “kali”, but useradd didn’t read the skeleton stuff out of /etc/ (not only that - it didn’t even create home!) - I also can NEVER REMEMBER which one to use, useradd, or adduser, and depending on the vagaries of what distro, or release you’re on, sometimes, one of those two will run you through a series of questions - seriously - WHICH ONE? useradd? adduser? I think on some distros, they’re the same thing…

So - “me being me” I created /home/x manually, copied /home/kali contents into it. Then (as root) cd /home/x ; chown -Rf x:x . - but when I try to fire up the file manager (I would assume “thunar”) - it barfs 'cause somewhere in its configuration it’s hard coded /home/kali .

And I can’t find the file that has that hardcoded in it. e.g. ~/.config/Thunar there’s only one file “accels.scm” whatever that is :

cat Thunar/accels.scm
; Thunar GtkAccelMap rc-file         -*- scheme -*-
; this file is an automated accelerator map dump
;
(gtk_accel_path "<Actions>/ThunarActions/uca-action-2-2" "F4")
(gtk_accel_path "<Actions>/ThunarActions/uca-action-1-1" "F4"

The files user-dirs* located in /home/x/.config have $HOME in them e.g.

# This file is written by xdg-user-dirs-update
# If you want to change or add directories, just edit the line you're
# interested in. All local changes will be retained on the next run.
# Format is XDG_xxx_DIR="$HOME/yyy", where yyy is a shell-escaped
# homedir-relative path, or XDG_xxx_DIR="/yyy", where /yyy is an
# absolute path. No other format is supported.
# 
XDG_DESKTOP_DIR="$HOME/Desktop"
XDG_DOWNLOAD_DIR="$HOME/Downloads"
XDG_TEMPLATES_DIR="$HOME/Templates"
XDG_PUBLICSHARE_DIR="$HOME/Public"
XDG_DOCUMENTS_DIR="$HOME/Documents"
XDG_MUSIC_DIR="$HOME/Music"
XDG_PICTURES_DIR="$HOME/Pictures"
XDG_VIDEOS_DIR="$HOME/Videos"

So WHERE the HELL is Thunar getting /home/kali from ? I don’t even know if it’s that - it might be ~kali?

I looked in ~kali, and can’t find any file with that username hardcoded into it :

┌──(root㉿ilak)-[/home/kali]
└─# pwd
/home/kali
┌──(root㉿ilak)-[/home/kali]
└─# find . -iname \.\* -type f -exec grep kali {} \;
Xsession: X session started for kali at Sun Sep 24 01:20:38 AM UTC 2023
dbus-update-activation-environment: setting XAUTHORITY=/home/kali/.Xauthority
localuser:kali being added to access control list
grep: ./.Xauthority: binary file matches

Surely it can’t be in .Xauthority? It’s a binary file… How does one edit it? Can it be deleted and will it be created on startup / desktop restart ???

I renamed .Xauthority and then ran an X app over SSH (xeyes) and that looks to have rebuilt the MIT Magic Cookie file $HOME/.Xauthority…

But I’m still getting the error :
“Failed to execute the … blah blah” because it’s got /home/kali hardcoded into it - and I can’t find which file has that in it…

Just tried to purge thunar - and guess what? It also purged xfce4-desktop* (yes and a whole bunch of other stuff!)… I reckon what I should have done is maybe installed nemo, set that as the preferred file manager…

Trying to re-install it now… It’s actully still running - surprising that it let me pull the rug out from under it - while it was still running… Let’s see how I go…

I might have to try this :
See if the Kali arm64 installer will boot off a USB thumb drive on a Pi4 (the Pi does have USB boot enabled) - and then allow me to install to SD-Card…

I’d rather not though - just remembered I spent a couple of hours yesterday trying to get a “hat” (I guess its kinda like an ISA [definitely not like a PCI) card that sits atop the GPIO pins on a Pi) that works as a UPS - a “PiJuice” hat - it’s basically so that when I close the lid on the NextDock the Pi his using as display keyboard and mouse, it doesn’t shutoff power to the Pi.

** “(times FIFTEEN CHARS!)” only refers to the Discourse prompt telling me my subject must be at at least 15 chars long…


I’ve given up trying to fix Thunar - I just removed the icon from the top panel, and added Nemo to my plank dock…


Note also : There’s an option in XFCE Settings Manager for “File Manager Settings” but nothing there about which folder to default to… Anyway - I’m glad, I kinda hate the icon for Thunar anyway…

When you login as ‘x’ what is the value of $HOME ?

┌──(x㉿ilak)-[~/bin]
└─$ echo $HOME   
/home/x

I can live without Thunar…

Now I hit another issue that’s miffing me - I’d like a way to unify the clipboard, and the X select buffer…

I can sync them with : xclip…

But - scenario :

Kali running as VM in QEMU (UTM - aarch64)

When I select some text in the Host OS in e.g. terminal (iTerm2) which is MacOS 13 - it goes into the clipboard automatically (MacOS only has one clipboard) - but when I go to the Kali VM - I have shared clipboard enabled in QEMU and spice-agentd installed and running… It knows about the clipboard - but the Host OS (MacOS) clipboard contents, are NOT in the Guest OS X Select buffer for middle button pasting in e.g. a Terminal or a Text Editor.

Note : Synergy does a fine job sorting this - e.g. the clipboard contents in MacOS become BOTH X Select Buffer - AND - Clipboard in Linux with Synergy clipboard sharing enabled… And I have Synergy client running on the Pi4 that’s also running aarch64 version of Kali, and it works fine there, I can select something on MacOS and Synergy shares that clipboard to everything else running Synergy (so I can middle click to paste the clipboard of my Mac, into a terminal window in Kali running on Pi 4).

I don’t wanna have to run a shell script that does : xclip -o -selection clipboard | xclip -i every time I need it…

With a linux host and linux guest, that works every time.
Therefore your problem is in the MacOS host… does it have all the required spice packages installed? I needs to know something about spice to be able to talk to the spice-vdagent in the guest.

Another thought… do you use qemu-ga in the guest?

Clipboard works 100% via spice-agentd…

That’s no the issue… the issue is that the guest O/S doesn’t integrate the clipboard and the X select buffer…

Honestly - I’d rather have X select buffer in the guest VM shared with the host machine - in this case MacOS.

I reckon I’d have the same issue with Windows as the hosting machine too…

I don’t know what “qemu-ga” is - if its for sharing directories and files, I don’t need it - I already have my preferred cloud sync solution working in the guest (Resilio Sync)…

I get it now, sorry
qemu-ga may help
When you install qeme-ga in guest
you need to setup a Channel for it in virt-manager
so you use the add hardware thing
choose channel
then choose qemu-ga channel
and it offers a box to tick for clipboard.

So that may do it. No guarentees.

My workaround is going to be : then I want to paste into the terminal, I’ll just SSH from the host MacOS…

Bigger fish to fry right now - my Pop!_OS (22.04 Ryzen 7, discrete AMD GPU: Radeon RX 6800/6800 XT / 6900 XT) desktop seems to have gotten its knickers twisted…

Was trying to do USB passthrough to a VM in VirtualBox (for more Kali testing), added my user to the group vboxusers, and instead of just logging out of X, I rebooted…

Then X wouldn’t start… At all… i.e. after unlocking my LUKS “/” - just a black screen…

ssh’d in from my MacBook, checked logs, nothing obvious - decided to update / upgrade - that took ages, the longest bit was it generating a massive list of Locales I would NEVER use - and - I saw it complain about missing firmware for AMD GPU…

Rebooted again… Nada… Powered off… Powered on…

Then it eventually displayed a login prompt on both monitors (a kinda weird looking prompt too) - tried to login - kept going back to the same login screen (which is unusual, as I have it set to autologin)… Tried 4 times… So there’s obviously some issue with the DM cannot load the DE.

Fired up a virtual TTY (ctrl+alt+f3) - logged in as “me” then issued startx and my desktop loaded… everything looks like it’s where it should be - still seems to be using “accelerated” graphics from my GPU (using open source AMD drivers) :

I think I’ll just leave it like this for the time being and try and dig to the bottom of the issue some other time (busy busy busy with work right now - and this machine is my KVM (keyboard, mouse, [video]) server to drive my other computers I use for work.

If it looks like too much work - I’ll probably just backup my Steam and Lutris game libraries somewhere (e.g. my NAS) and rebuild… And maybe not Pop!_OS… Maybe Ubuntu 23.04 (or just wait for 23.10)…

The reason I wanted to test Kali on VirtualBox - is that the monitor-mode compatible USB WiFi dongle I have - when passed through to QEMU (it works) seems to only do 2.4 Ghz WiFi - and I know for a fact the Raspberry Pi (all models that have WiFi) can also only do 2.4 Ghz - but I will need BOTH for what I’m planning… Just verified - the dongle only does 2.4 Ghz WiFi… Doh!

1 Like

Damn - it’s contagious - my Ryzen Thinkpad with Ubu 23.04 won’t boot up now… and it’s worse than my desktop machine - unable to backdoor in via virtual TTY console…

First time around it was showing stuff about IRQ 7 - and - TPM…

Funny, I don’t remember enabling TPM… Went into the BIOS and it’s enabled, so I disabled it (but how did it get enabled?) - TPM message is now gone, but still getting something like :
# irq 7: nobody cared (try booting with the "irqpoll" option) - copied from elsewhere - as I can’t login to it. Other people have complained of this symptom but they can still login. I can’t even ctrl+alt+f3, f4 or f5 - just goes back to tty1 with the message about IRQ 7 - and maybe something about AMD (the main noticeable issue on my desktop machine, is also about AMD or something - but I can still start x and use acceleration - it’s working - tried playing Quake [but with a much more recent game engine])…

I might have to try a recent live image on there - like a beta of Ubuntu 23.10 perhaps? Something with kernel 6 anyway - as I reckon both issues are maybe something to do with kernel 6 and possibly AMD - I can always swap in a decent NVidia GPU in the desktop machine - but don’t have that option on the Thinkpad, it’s AMD only, and the AMD is vega8 APU (i.e. GPU cores are built into the Ryzen 5 CPU).

Note also - that Thinkpad has been complaining for a fortnight or more that I need to reboot it 'cause of an update - it was working perfectly before…

I have kernel 6 in Void and Gentoo and MX. No problems there… and I have an amd gpu, but not an amd cpu.

1 Like

Uncanny - I let the Ubu 23.04 Thinkpad sit there “overnight” with that TTY message on the screen - got up this morning, and it’s now presenting a graphical login - but - exactly the same symptom as my Pop! 22.04 - loops back to the login prompt…

CTRL+ALT+F4, login, type startx and I’ve got my DE…

Surely others (outside of this forum) must be experiencing similar symptoms…

The similarities are too close to be co-incidence - IMHO some update has screwed up both my systems…

Ubuntu 23.04 ThinkPad Kernel 6.2x Ryzen 5 AMD GPU, NVMe SSD 32 GB DDR4 :
Linux fenriz 6.2.0-33-generic #33-Ubuntu SMP PREEMPT_DYNAMIC Tue Sep 5 14:49:19 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux

GPU info “synopsis” from inxi -Fxz :

Graphics:
  Device-1: AMD Picasso/Raven 2 [Radeon Vega Series / Radeon Mobile Series]
    vendor: Lenovo ThinkPad E595 driver: amdgpu v: kernel arch: GCN-5 bus-ID: 05:00.0 temp: 50.0 C
  Device-2: Acer Integrated Camera type: USB driver: uvcvideo bus-ID: 3-2:3
  Display: server: X.Org v: 1.21.1.4 with: Xwayland v: 22.1.8 driver: X: loaded: amdgpu
    unloaded: fbdev,modesetting,radeon,vesa dri: swrast gpu: amdgpu resolution: 1: 2560x1440~60Hz
    2: 2560x1440~60Hz
  API: OpenGL v: 4.5 Mesa 23.0.4-0ubuntu1~23.04.1 renderer: llvmpipe (LLVM 15.0.7 256 bits)
    direct-render: Yes

Pop!_OS 22.04 Desktop Kernel 6.4 Ryzen 7 AMD GPU, NVM3 SSD 16 GB DDR4 (SO-DIMM) :
Linux titan 6.4.6-76060406-generic #202307241739~1695312586~22.04~4d8f3e9 SMP PREEMPT_DYNAMIC Thu S x86_64 x86_64 x86_64 GNU/Linux

GPU info “synopsis” from inxi -Fxz :

Graphics:
  Device-1: AMD Navi 23 [Radeon RX 6600/6600 XT/6600M] vendor: Sapphire
    driver: amdgpu v: kernel bus-ID: 2f:00.0
  Display: server: X.Org v: 1.21.1.4 driver: X: loaded: amdgpu,ati
    unloaded: fbdev,modesetting,radeon,vesa gpu: amdgpu resolution:
    1: 2560x1440~60Hz 2: 2560x1440~60Hz
  OpenGL: renderer: AMD Radeon RX 6600 (navi23 LLVM 15.0.7 DRM 3.52
  6.4.6-76060406-generic)
    v: 4.6 Mesa 23.1.3-1pop0~1689084530~22.04~0618746 direct render: Yes

I just noticed - Ubu 23.04 is trying to do “Xwayland” support…

Anyway - both are “sort of” working for now, I can live like this for a while… Given I hardly ever powerdown or even logout - I may not even notice the difference :smiley:


Nearly all the “fixes” seem to be to remove ~/.config - but I don’t want to do that! Might as well rebuild / reinstall…

In fact - I might leave things as they are - when 23.10 comes out - I’ll see if the Thinkpad will do a “do-release-upgrade” - if not - maybe I’ll just install 23.10 from scratch…

I don’t think Pop!_OS supports inplace upgrades, so if I want Pop!_OS 24.04 next year on this desktop machine, I’ll probably have to do a fresh install - can I wait that long? I reckon I can…

1 Like

Disable the login manager…it loops like that when

  • elogind is playing up
  • starting X fails
  • filesystem out of space

Not sure what “elogind” is - I use whatever the default DM is that Ubuntu / Gnome uses - I assume “gdm”.
Starting X doesn’t fail (i.e. “startx” works)
Defo not out of space… :smiley:

As I said I can live with it like this for the time being…

It is the systemd version of logind… runs as a daemon
it has a config file somewhere… try fiddling with that
I had trouble with elogind in a laptop once… something to do with suspend or hibernate

1 Like

My desktop machine with Pop!_OS :

╭─x@titan ~
╰─➤  ps -ef |grep logind|grep -v grep root        1301       1  0 Sep27 ?        00:00:03 /lib/systemd/systemd-logind
╭─x@titan ~/.config
╰─➤  systemd-analyze cat-config systemd/logind.conf
# /etc/systemd/logind.conf
#  This file is part of systemd.
#
#  systemd is free software; you can redistribute it and/or modify it under the
#  terms of the GNU Lesser General Public License as published by the Free
#  Software Foundation; either version 2.1 of the License, or (at your option)
#  any later version.
#
# Entries in this file show the compile time defaults. Local configuration
# should be created by either modifying this file, or by creating "drop-ins" in
# the logind.conf.d/ subdirectory. The latter is generally recommended.
# Defaults can be restored by simply deleting this file and all drop-ins.
#
# Use 'systemd-analyze cat-config systemd/logind.conf' to display the full config.
#
# See logind.conf(5) for details.

[Login]
#NAutoVTs=6
#ReserveVT=6
#KillUserProcesses=no
#KillOnlyUsers=
#KillExcludeUsers=root
#InhibitDelayMaxSec=5
#UserStopDelaySec=10
#HandlePowerKey=poweroff
#HandleSuspendKey=suspend
#HandleHibernateKey=hibernate
#HandleLidSwitch=suspend
#HandleLidSwitchExternalPower=suspend
#HandleLidSwitchDocked=ignore
#HandleRebootKey=reboot
#PowerKeyIgnoreInhibited=no
#SuspendKeyIgnoreInhibited=no
#HibernateKeyIgnoreInhibited=no
#LidSwitchIgnoreInhibited=yes
#RebootKeyIgnoreInhibited=no
#HoldoffTimeoutSec=30s
#IdleAction=ignore
#IdleActionSec=30min
#RuntimeDirectorySize=10%
#RuntimeDirectoryInodesMax=400k
#RemoveIPC=yes

Just see a lot of handlers commented out…

My Thinkpad’s powered off (tried Kali on there via USB 3 connected SSD and it was pretty much unusable - i.e. slower and 10x more laggy than a Raspberry Pi 4 running Kali) - and I’m working in the office for the first time in 2-3 weeks (I can SSH to my desktop machine from work using my 24x7 “headless” Pi4)…

They are all commentd out
Get rid of the # and set them all to ‘ignore’

I might give that a go next time the piece of crap allows me to login… Can’t believe a simple software update can demote a working computer to a boot-loop pile of crap… Just powered it up about 3 hours ago - decrypted my LUKS root partition - and - it’s been doing the boot loop thing ad nauseam since - I can’t “ctrl+alt+f2,3,4” cause when I do it just loops back to the shonky AMD GPU and “IRQ7” message on TTY1…

The solution on my desktop system might be “switch out the AMD PCIe card for an NVidia one” - it will probably run with the Nouveau driver for a bit… But whatever it is - it hasn’t bricked my desktop machine.

Note : nearly identical issues on Pop!_OS 22.04 and Ubuntu 23.04… Pop! running on a PCIe “Radeon RX 6600/6600 XT/6600M”, Ubuntu on Vega8 GPU from the AMD Ryzen 5 “APU”…

I’m tempted to try out one of the forks of Valve’s “Arch based” distributions on the Thinkpad - nearly identical specs as the SteamDeck (that was one of the main reasons I bought it off e-bay)…

That may be a good argument for not using a login manager.

Do you follow the footy. Th nrl GF was a nail biter

I only follow footy if it’s Aussie Rules (despite being born in NSW - I’ve lived most of my life in WA or Victoria), and then only if it’s a team I like, and I LOATHE Collingwood :smiley: - their last foray in the GF was against The Eagles, who I also happen to despise somewhat (even though I live in “their town”) - was the only time I’ve supported The Eagles, that’s how much I loathe Collingwood.

Note : boot loop - still going some 12-16 hours later… I suspect this thing can’t be resurrected… But how do I avoid this ever happening in the future?

I may end up just not trying to fix it and see if Ubuntu 23.10 live boot is smart enough to upgrade…

While I was replying here - I decided to see if I could ping it using avahi (ping fenriz.local) and it worked - so I tried SSH and that worked…

So - it’s available as a headless thing on the network … But there’s no updates from Canonical… I’m convinced an update broke it - was hoping a subsequent update might fix it…

I’ve just set all those dick handler values to uncommented and “ignore” and rebooted…

Note : changing those values in logind.conf made zero difference… I can SSH to it - but I can NEVER get it to stay on virtual TTY - e.g. soon as I hit CTRL+ALT+F2, F3, F4, F5 etc - within 2-3 seconds, it’s flicked back to the boot console with the ugly message about IRQ7 :

2023-10-02T07:35:55.151425+08:00 fenriz kernel: [    1.252825] irq 7: nobody cared (try booting with the "irqpoll" option)
2023-10-02T07:43:12.626827+08:00 fenriz kernel: [    1.225689] irq 7: nobody cared (try booting with the "irqpoll" option)

Note : since I started typing this reply - I kinda made a workaround…

(from remote ssh session) :
sudo systemctl stop gdm.service
sudo systemctl disable gdm.service

On the laptop :
ctrl+alt+F3
(after stopping GDM it stops looping back to virtual TTY 1)
Login and type “startx”…

But it’s too much of a co-incidence that TWO of my AMD GPU devices have very similar symptoms… There are no AMD GPU driver updates…

That is how most feel. It is strange what attracts about a team. Some of it is just the local team effect, but there are other things about the style of play.

You machine is running, its just the login manager that is looping. Why not remove it?

IT’s the DM not the login manager… Disabled GDM and I can now login via console and manually start X… It’s a workaround - not ideal…