Ubuntu 22.04 LTS is currently available as an upgrade but not yet released.
It works out of the box once upgrade is complete.
*** Proceed at your own risk as this is pre-release *** I will update this post once the release is officially out.
I have noticed significant battery and HW support improvements, especially on Panel Self Refresh (PSR) support
Installation
You can follow the guide to get a base 20.04 install (or 21.x), then run:
$ do-release-upgrade -d
NB: I would recommend doing the do-release-upgrade from a console directly so you do not loose graphics - if you run it from gnome terminal or an xterm and your user interface disappears open a VT via ctrl+alt+F3, then run:
$ sudo -i # interactive sudo as root
$ screen -ls # dumps screen session list, you will see the release upgrade session
$ screen -d -r [SESSION_FROM_ABOVE] # will reconnect you to the upgrade process
Compatibility
Wifi works Bluetooth works Fingerprint Scanner works (see hw issue with old fingerprint scans fixed here ) PSR works Deep Sleep works Sensors works (lm-sensors for cpu and board temps) Headset works with the following command after setup:
$ echo "options snd-hda-intel model=dell-headset-multi" | sudo tee -a /etc/modprobe.d/alsa-base.conf
Fans
Still no great insight to fan speeds. i8kutils and similar chipset utils related to dell do NOT work with the fans on this board.
nbfc-linux appears to work with config LENOVO 20U9S19X00 but is lying - it will report the same fan speed consistently regardless of throttling and targets speeds are ignored
Outstanding Issues
Still seeing relatively high interrupt counts on idma64.2 and tick_sched_timer - the latter is not likely an issue and is probably tied to cpu frequency scaling. The former appears to be touchpad related, but I am not 100% on this yet, that may be a red herring
Misc
Running powertop --auto-tune on boot via rc.local or similar really helps with battery life - In my case (i7-1165G7 with 32GB Ram) I went from 5-6hrs of battery life on 20.04 to 8-10hrs on 22.04 and base powertop auto-tune applied. Normal usage for me is a reasonable amount of programming, and a lot of time in gmail / doc editing - not much media usage. Still seeing some drain during web conference calls, esp on teams but I believe that is browser related
some kernel flags to experiment with
nvme.noacpi=1 sets some better ssd behavior for sleep and power states
i915.enable_psr=1 will force PSR on
mem_sleep_default=deep explicitly states what sleep mode you want
intel_pstate=disable turns off processor level p-states and will let you more comfortably scale cpu freq to ~400mHz or so
acpi_enforce_resources=lax lightens up a few things to get better access under the hood - do NOT do this unless you know what you are doing
I experienced around 5-6 hours without any tuning on a mixed workload. Some writing, some compiling. Same as Mike.
I do still experience some major issues with the AX210 wifi card though, I had to disable the ax spec and now run it with just ac. That at least is working just fine.
Do you have the vPro Wi-Fi module? I am running the no vPro card and have had no observed issues with the Wi-Fi on 22.04. Any kind of insight or how you may have disabled the ax spec would be great for others if they are having the issues you ran into.
The non-Pro version. I observed frequent disconnects especially after a warm startup. I checked on my AP and it was using the full spectrum of AX features which is good aside from the disconnects.
There is a whole thread discussing the issues with the adapter on the forums and on the red hat but tracker but the specific workaround was this one [SOLVED] Using the AX210 with Linux on the Framework Laptop - #92 by Andrew_Marshall
I will try to occasionally enable the feature but for now I am fine with disabling ax.
Seeing some errors related to this on mine as well - nothing causing actual problems such as disconnects, just cluttering syslog for me. Also non-vPro module.
For fans, using the latest EC Tool works, for example
Set Fan to 40%:
$ ectool --interface=fwk fanduty 40
Set fan back to auto:
$ ectool --interface=fwk autofanctrl
Hopefully this makes it back upstream over time
Don’t necessarilly recommend non developers to tinker too much with ectool as you can do some damage unintentionally
Ok have installed the lasted ISO of Ubuntu 22.04, posted on the weekend! So far installation was very smooth (very fast!). Now I’ve been using Ubuntu for a long-time, but I wouldn’t call myself a power user at all. Still need help with lots of things. Since install on Saturday I haven’t done a thing and everything basic is simply working (incl WiFi). However I will do a few of the essential tweaks (headset port, deep sleep & powertop). If I need help will try to find existing answers, but if stuck I may come back here for a few pointers…!
OK first two above done fine, but with Powertop the output of “sudo powertop --auto-tune” leaves me unsure if it’s done anything. See below - does this look correct, or is something wrong?
modprobe cpufreq_stats failedLoaded 29 prior measurements
Cannot load from file /var/cache/powertop/saved_parameters.powertop
File will be loaded after taking minimum number of measurement(s) with battery only
RAPL device for cpu 0
RAPL Using PowerCap Sysfs : Domain Mask d
RAPL device for cpu 0
RAPL Using PowerCap Sysfs : Domain Mask d
Devfreq not enabled
glob returned GLOB_ABORTED
Cannot load from file /var/cache/powertop/saved_parameters.powertop
File will be loaded after taking minimum number of measurement(s) with battery only
To show power estimates do 346 measurement(s) connected to battery only
Leaving PowerTOP
Good to know! I don’t use a docking station at the moment, but my mouse which uses a USB dongle, is freezing a few secs after login. If I remove and re-insert, it’s fine for the rest of my session. I might try disabling the powertop command and see if that influences it.
Otherwise, a week in and so far so good! Really enjoying the upgrades/tweaks in 22.04 (from 20.04).
I’m having issues setting up hibernation for 22.04. I followed a few guides that I found that say to use a swapfile to suspend then hibernate. When I try to test hibernate using “sudo systemctl hibernate” it works but it immediately wakes back up and goes to the lock screen. Also the laptop does not seem to be suspending when I close the lid, the screen stays on. Tips?
Hi all,
I started with 20.04 (what I had on a USB stick from before), and managed to get the wifi (non vPro) working using the workarounds referred to here Ubuntu 20.04.4 LTS on the Framework Laptop
then I did an in-place upgrade to 22.04 using do-dist-upgrade, and ha ha joke’s on me wifi is now just not working at all.
ip link shows a wlp interface, but it never goes UP.
What did I mess up? The iwlwifi-ty-a0-gf-a0.pnvm file is in place in /lib/firmware, different from the patched one from the 20.04 update
Yay for getting it working even if it was the hard way. If you had WiFi working on Ubuntu 20.04 with latest updates, then probably it wasn’t the formula’s workaround fault. Upon upgrading to 5.13 via regular updates on 20.04 WiFi breaks due to the workaround. At that point you have to either disable it manually and restore the firmware file or rerun the formula which will do that for you. Since you had it working in order to do the upgrade, one or the other likely happened. If the firmware file was in place post-upgrade, then the workaround was disabled since all it does is remove it on every startup. So both pieces of evidence point to a problem elsewhere. Given that the interface was present also lends credibility to that hypothesis. My guess is something else dun fucked up. What exactly, we’ll never know since all the evidence was vaporized upon nuke detonation. With all that said, if I see other reports of the same problem, I’ll investigate. I have completed about 10 test upgrades and 2 for-real upgrades on machines that were salted with the formula and didn’t see any issues but as always, mileage varies and “works for me” doesn’t mean there are no defects.
Wifi works
Headset works with the following command after setup:
What exactly, we’ll never know since all the evidence was vaporized upon nuke detonation. With all that said, if I see other reports of the same problem, I’ll investigate. I have completed about 10 test upgrades and 2 for-real upgrades on machines that were salted with the formula and didn’t see any issues but as always, mileage varies and “works for me” doesn’t mean there are no defects.