Bib-lost's questions

My laptop (Dell XPS 13, 9350) has a single USB-C port, which I am used to connecting to a docking station which is in turn connected to peripherals (USB devices, audio, display, power supply). One day when booting up the computer, the display had stopped working: the laptop still 'sees' that an external monitor is connected, but the monitor reports that it receives no signal.

Since monitor and docking station were working fine, I feared the worst: a broken USB-C port. But when I booted my laptop from a live USB with Ubuntu, suddenly the external monitor was working. So the issue must be driver related. The problem started a few weeks ago and I honestly did not expect this issue to be driver related, so I am afraid I cannot reconstuct which changes were made to the software shortly before the issue started occurring.

In summary: Dell XPS 13 9350, Ubuntu 20.04.3, 64bit, kernel 5.11.0-27-generic, integrated Intel HD Graphics 520. USB-C port still works for power supply and to connect USB and audio devices, but no longer properly sends the Displayport signal.

Update: Based on the following log entry

i915 0000:00:02.0: [drm] *ERROR* Link Training Unsuccessful

I found the following related bug report on Launchpad. Supposedly, the problem should be solved in kernel version 5.11.

Curiously, I found that booting in the old kernel 5.4.0-81-generic instead resolved the issue. I would rather not downgrade so many kernel versions, and my technical knowledge is too limited to read from the discussion on the Launchpad bug report what to do to patch the problem manually.

After upgrading my system from Ubuntu 18.04 to 20.04, I had some graphical glitches and had to reinstall xorg (following this answer). Everything seemed to work fine, except that I could not launch Steam anymore. I tried reinstalling Steam, including removing all the Steam related folders in /home, but to no avail. First time booting Steam after reinstall, a window with a progress bar pops up stating that "Pins potentially out of date ... rebuilding" and then nothing. All other times, there is no graphical feedback whatsoever.

The terminal output at first glance does not seem to point towards any missing packages or anything either, so I am out of ideas. Most of the related reports I find online turn out to be related to proprietary graphics drivers of dedicated GPU's (typically NVidia), but I only have an integrated Intel chipset and want to play some indie games

Running Steam on ubuntu 20.04 64-bit
STEAM_RUNTIME is enabled automatically
Pins up-to-date!
Can't find 'steam-runtime-check-requirements', continuing anyway
/home/nicolas/.local/share/Steam/ubuntu12_32/steam
[2020-11-07 13:26:07] Startup - updater built Jul 28 2020 22:37:33
ILocalize::AddFile() failed to load file "public/steambootstrapper_english.txt".
Installing breakpad exception handler for appid(steam)/version(0)
free(): invalid pointer
No minidump written, nothing to upload.
/home/nicolas/.local/share/Steam/steam.sh: regel 750: 19858 Afgebroken              (geheugendump gemaakt) $STEAM_DEBUGGER "$STEAMROOT/$STEAMEXEPATH" "$@"

In 2016 I bought a Dell XPS 13 9350, which came with Ubuntu 14.04 out of the box. By now it has been upgraded to 20.04. Firmware updates are available from the Software Centre by default.

Lately, however, firmware updates are no longer installing. I found out out that the problem might be the size of the EFI boot partition, which is only 48 MB, supposedly too small for the firmware updates, according to the error message of the Software Centre.

What would be the safest way to continue? Resizing partitions which are at the beginning of the disk is notoriously risky. Attached is a picture of the partition (from GParted)

I have set up an OpenVPN server to which I can succesfully connect from Android and iOS devices.

I am having troubling setting up my Ubuntu 18.04 machine as a client. I installed the gnome OpenVPN client so I could connect via the Network Manager. Having done so with the openvpn file and password provided by the server, I get the message that I succesfully connected to the VPN server.

However, I cannot actually reach the VPN network; in fact, I do not even get a response when I ping the VPN server on my client machine.

What could be causing this? There is no firewall enabled on the client side.

Edit: just tried connecting with

sudo openvpn --config [ovpn file name] 

instead. Worked flawlessly. So the problem is with the network manager. Does it somehow not import the right settings from the *.ovpn file?

Edit 2 on request: When connected to VPN through CLI command: I can ping 8.8.8.8 and any machine on the local network. When supposedly "connected" to the VPN through network manager GUI I can ping neither 8.8.8.8, nor any machine on the local network, nor the server itself.

I currently switch between the following two set-ups on a daily basis:

1) Being connected to the internet via a WiFi network.

2) Having my (Ubuntu 16.04) laptop connected to the internet via a wired connection, then sharing this connection with my other devices through the built-in hot-spot functionality of Ubuntu.

At the moment, switching from 1 to 2 requires me to manually go through the "connect to hidden network" dialogue in the Network Manager; similarly switching from 2 to 1 requires me to manually disconnect from the hotspot and connect to the WiFi network. It's not a lot of work, but I do it often enough that I would like to have it be automated.

It is not as simple as enabling autoconnect for the hotspot, since then it overrides the autoconnect functionality of other networks instead of connecting me to a WiFi network when I am not connected via a wired connection, and I end up with a hot-spot but no internet. So I need a solution which:

• connects me to the hot-spot whenever I am connected to the internet via a wired connection.
• disconnects me from the hot-spot when there is no ethernet connection.
• autoconnects me to one of the networks with autoconnect=true whenever they are available and there is no ethernet connection.

A full explanation on how to do this would be great, but a reference to a good (non-expert) manual for setting up these kinds of rules would also be very much appreciated.

My (Ubuntu 16.04) laptop has the following quirk: usually shutdown will take under 5 seconds, occasionally it will take over a minute. Never anything in between.

Having observed this phenomenon for almost half a year now, the only thing I am sure of is this: the long shutdowns happen exclusively if - since booting up - the WiFi has at some point been enabled. They do not always happen if WiFi has been enabled, but it seems to be a necessary condition. On the other hand, disabling the WiFi in the network manager GUI right before shutdown is not sufficient to prevent the minute-long delay.

Yet, looking at the logging during shutdown, no mention of anything WiFi-related is made. During such long shutdowns, I immediately get

[ OK ] Started Show Plymouth Power Off Screen
[ OK ] Stopped Thermal Deamon Service

Then it stalls for one or two minutes, then the rest of the shutdown procedure happens in less than a second, starting with

[ OK ] Stopped session c2 of user *username*
[ OK ] Removed slice User slice of *username*

and ending with

[ OK ] Stopped Remount Root and Kernel Systems
[ OK ] Reached Target Shutdown

Nowhere is there any mention of the WiFi hardware or any error message. CUPS is somewhere in the middle of these logs; I know that it is often the culprit in slow-shutdown situations, but I do not have any printers on the network.

Let me know if you need more details on the shutdown logs, I recorded it on camera and did not feel like retyping all of it.

I always tell people that 'nothing can go wrong' and that 'their computer won't even remember' if I boot their systems from my Ubuntu-installed pendrive, but I recently found out that this is not true. Booting from a pendrive has the possibility of changing the hardware clock of the computer it is used on.

As explained e.g. in these help pages, Linux and Windows interpret the hardware clock as UTC and local time, respectively. This means that in a dual-boot system, one wants to make Linux read the system clock as local time too, to prevent a mismatch in Windows. The problem is that I want to use my pendrive both in computers with Linux as their main OS and computers with Windows as their main OS, so whatever decision I make for my pendrive (local time vs UTC), it's not going to fit all computers.

Is there a way to simply prevent Ubuntu from changing the system clock, so at least I don't mess up other people's setups? If that means the pendrive system sometimes does not display the correct time, then so be it.

NOTE TO EAGER FLAGGERS: I know that this question has been asked a thousand times on this site by dual-booters (see the two Q&A's below), but for as far as I could find the proposed solution was always to adjust Linux settings to interpret the hardware clock as local time (or make Windows interpret it as UTC). This is not an option for me since I want a portable system that does not change any clocks.

• Time keeps changing between BIOS/WIndows 7/Ubuntu 14.04

• Clock time is off on dual boot

Steam installs games to the home folder. I sort of get why this decision was made, although I would prefer they installed the games to /opt and maybe only the personal files to /home. The current situation means that having an encrypted home folder is not really an option.

If my laptop gets stolen, I want my documents and passwords to be safe. I don't care if my saved games are stolen, let alone that people might know what games I play. Is there a way to unencrypt a part of the home folder? Or should I leave the entire home folder unencrypted by default and only encrypt subfolders? I am not an expert on this to be honest, I just always use the option to encrypt the home folder provided during install.

I'd like to make my (Windows 10) desktop computer into a dual-boot system Windows-Ubuntu. However, since Windows 10 will be the most used operating system by far, I wonder if it's possible to do this in such a way that the computer always automatically boots into Windows, preferably without showing any bootmanager and without a significant decrease in speed compared to the current single-boot situation. Ubuntu would then have to be accessed by a one-time-boot menu, triggered by pressing a button during startup (probably F12 on my system).

Is this possible at all and how should I go about this?

I have noticed that the way my laptop (Dell XPS 13, Ubuntu 16.04, Unity) suspends when I manually order it to (i.e. via the graphical interface or terminal command sudo pm-suspend) is different from the suspension which happens when I close the lid. The most notable difference is that the power button continuously flashes in the second case, but goes completely dark in the first case. Besides that, I think wake-up time is also ever so slightly higher in the lid-close case, but I'm not 100% sure. I'd like to know why these two distinct suspension modes exist at all.

I am trying to revamp an old HP Mini netbook and make it into a useable mini laptop with Lubuntu 16.04. The laptop was unuseably slow when it still ran Windows 7, now with Lubuntu it can be taken seriously once again. Thank you, Lubuntu team.

There is one thing so far which I have found works worse than it did on Windows 7, this is playing Youtube videos. Even when using the HTML5 player, I cannot get past 360p without getting immense framerate drops. The laptop used to be able to handle 480p without issues, and this is what I'd like to get to. The screen is too small and crappy to enjoy benefits of even higher resolutions anyway.

I have read through this high-quality answer. I haven't fully followed the guide yet of watching youtube through an external player, but it is true that downloaded video can be played significantly more smoothly.

Before resorting to 'work-arounds' like embedding youtube videos in external players, is there more I could do to get the web player to work? Going from Firefox to Chromium already helped (on Firefox I could only watch videos on 144p, too little to be able to read subtitles). The laptop's 1 GB of RAM won't be the problem I think, but the 1.66 Ghz Intel Atom N455 processor is a bottleneck. It's a 32 bit system. Internet speed is not a problem. Let me know if you need more specs, I am pretty sure this is the full list of specs.

I am able to connect to a Wireless network using the graphical interface of Ubuntu 16.04 (Unity) by manually configuring a network whose name I know.

I can also scan for networks and their information using the terminal command sudo iwlist scanning | less and this works just fine. But there used to be a nice graphical way to look at available wireless networks, just by opening the drop-down menu in the upper right corner of the screen. Even when I am connected to a wireless network, I no longer see any networks (or signal strengths) in this drop-down menu. What could have happened? I am working on a Dell XPS 13 laptop, the Wifi is Intel Wireless 8260, the kernel version is 4.4.0.

EDIT: It would seem that this problem sometimes appears and disappears after rebooting or connecting/disconnecting with certain networks. JPS' solution provides a temporary workaround, but problem comes back once in a while. I will try to figure out what exactly triggers the problem, but sometimes it is just there when the computer boots up.

I know this question has been asked before, but most questions with answers seem outdated (like this one) and I was wondering if anything has changed.

In Ubuntu 16.04 (Unity), is it possible to have pinch-to-zoom gestures on a touchpad, provided that the hardware allows it? I have a Dell XPS 13 (which shipped with 14.04, but I upgraded it) and the touchpad allows two-finger scrolling just fine, but I do miss the more fancy multi-touch functionalities. This post explains how to check the number of fingers the touchpad supports using the geis-tools package; I found my laptop actually detects five, so it should be possible in principle. Off-topic: Who even uses five-finger gestures?

Background (can be skipped): Because of some (probably hardware) defect, my laptop sometimes disconnects from the battery without warning, causing an immediate and insecure shutdown if the AC adapter is not attached to computer and power supply. Several so-called computer repairers have proven completely unable to locate the problem (one tried to make me change my harddrive, claiming that 80% of it was corrupted even though I could still access all my files) so after having wasted a lot of money and getting nothing, I decided to just live with it and only use the laptop when it's connected to an external power supply.

Actual question: How can I make Ubuntu (16.04) immediately notify me (by text message and/or sound) whenever the power supply through the AC adapter is interrupted?

Own attempt: I wrote the following script

notify-send -i face-worried "No power!" "Connect the adapter."

which can be run from /opt/custom_scripts/warning_AC.sh both as root and as user (I tried both in bash). I have then tried to mimic the method explained in this answer on unix SA to let this script be run whenever the AC chord is plugged in. That is, I have monitored

sudo udevadm info --path=/sys/class/power_supply/AC0

to identify the changes made when plugging adapter in or out. I saw (among others)

E: POWER_SUPPLY_NAME=AC0
E: POWER_SUPPLY_ONLINE=0
E: SUBSYSTEM=power_supply

where in the second line the 0 changed to 1 when the power chord was plugged in and vice versa when it was plugged out. I thus decided the script I needed to place in /etc/udev/rules.d/80-power.rules was

SUBSYSTEM=="power_supply", ENV{POWER_SUPPLY_NAME}=="AC0", ENV{POWER_SUPPLY_ONLINE}=="0", RUN+="/opt/custom_scripts/warning_AC.sh"

I rebooted the computer, but nothing happened when plugging the power in or out.

I am very new to Linux and Ubuntu and willing to learn as much as I can about commands and scripts and stuff but this is as far as I got. Thanks for your help.