systemd-udevd 261.3-1 severe boot hang/slowdown, USB+DisplayPort broke

Elmario · Today 06:05:38

Hello!

I experienced a severe regression immediately after a recent system update and rebooting involving systemd 261.3-1 / systemd-udevd 261.3-1.

The system became almost unusable and showed the following symptoms:

Booting became extremely slow at the point where systemd-udevd starts:
"Starting systemd-udevd version 261.3-1". There it was stuck for like two minutes.

Then the boot log progressed extremely slowly, only one or a few lines at a time, with long pauses, like a C64 would do.

After reaching the desktop, the entire system was sluggish and became increasingly unresponsive.

My second monitor, connected via USB-C using DisplayPort Alt Mode, was no longer detected, Only the internal laptop display remained active. USB-devices connected to the same external USB-hub where still working - slowly, like the keyboard typing was very slow.

Disconnecting USB devices it became even slower I normally have custom udev rules that trigger actions when USB devices are connected or disconnected, and these actions were also heavily delayed.

Reconnecting the USB-C connection to the external monitor caused the entire system to freeze completely, requiring a forced power-off.

Subsequent boots showed the same severe delay around systemd-udevd, even trying CachyOS and Zen kernel in addition toe the Arch kernel that I have set as default.
Also I tried pulling all the USB devices while booting.

My notebook is now working again, after successfully restoring it using timeshift, systemd is restored to version 261.2-1.

Has anyone else experienced problems with systemd/udev 261.3-1, particularly involving USB device events, device initialization, or blocking during udev startup?

Thank you for help with encircling finding and fixing the issue!

Last edited by Elmario (Today 06:10:45)

seth · Today 07:41:56

involving systemd 261.3-1 / systemd-udevd 261.3-1

For clarification, systemd-udevd is not a thing, udev is part of the systemd package. Or are you using different builds/repos?
Have you confirmed them being the offenders by downgrading only systemd*?

There it was stuck for like two minutes.

90s - waiting from some device or service?

My notebook is now working again, after successfully restoring it using timeshift, systemd is restored to version 261.2-1.

Do you happend to have preserved the system journals of one of the affected boots?
(Otherwise the implication is you'll have to restore the broken state, dump the journal and either timeshift back or isolate the regression by downgrading

* systemd is split into 'systemd' 'systemd-libs' 'systemd-resolvconf' 'systemd-sysvcompat' 'systemd-tests' and 'systemd-ukify'
You'll have to up/downgrade the ones you've installed together

Elmario · Today 08:29:19

seth wrote:

For clarification, systemd-udevd is not a thing, udev is part of the systemd package. Or are you using different builds/repos?

No. I just don't know better, sorry

seth wrote:

Have you confirmed them being the offenders by downgrading only systemd*?

Not, yet, as my notebook not booting usually make me panic and my first action always is getting it do it's job again as fast as possible. So from the current position on, I could try the opposite, if you think this makes sense. This would be just upgrading just systemd.

seth wrote:

90s - waiting from some device or service?

seth wrote:

Do you happend to have preserved the system journals of one of the affected boots?

I fear timeshift probably deleted these, doesn't it?

seth wrote:

(Otherwise the implication is you'll have to restore the broken state, dump the journal and either timeshift back or isolate the regression by downgrading
* systemd is split into 'systemd' 'systemd-libs' 'systemd-resolvconf' 'systemd-sysvcompat' 'systemd-tests' and 'systemd-ukify'
You'll have to up/downgrade the ones you've installed together

OK, in case this repeats with my next trial updating the system, I will report back with these diagnostics.

Elmario · Today 10:58:24

OK: Following your advice I upgraded all again, and then tried subsequently downgrading suspicious packages, and with the third I found the culprit:

It's linux-firmware.
Added it to my worryingly growing IgnorePkg list:
IgnorePkg = linux-firmware linux-firmware-amdgpu lib32-libpcap lib32-libnl libpcap libcupsfilters

It's working again, thank you very much!

seth · Today 12:49:00

https://bbs.archlinux.org/viewtopic.php?id=314887
https://bbs.archlinux.org/viewtopic.php … 3#p2309463

likely only linux-firmware-amdgpu ?

Elmario · Today 14:05:40

Seems very probable, but their dependency forced uninstalling and ignoring both.

Last edited by Elmario (Today 14:06:49)

seth · Today 15:06:09

https://archlinux.org/packages/core/any/linux-firmware/ is a meta-package that drags all others in - you can probably remote that and the firmware packages you don't need and save like 200MB disk space

Edit: https://gitlab.archlinux.org/archlinux/ … k_items/52

Last edited by seth (Today 15:07:05)

Elmario · Today 16:21:46

Thank you. Good to it's already spreading.

Arch Linux

#1 Today 06:05:38

systemd-udevd 261.3-1 severe boot hang/slowdown, USB+DisplayPort broke

#2 Today 07:41:56

Re: systemd-udevd 261.3-1 severe boot hang/slowdown, USB+DisplayPort broke

#3 Today 08:29:19

Re: systemd-udevd 261.3-1 severe boot hang/slowdown, USB+DisplayPort broke

#4 Today 10:58:24

Re: systemd-udevd 261.3-1 severe boot hang/slowdown, USB+DisplayPort broke

#5 Today 12:49:00

Re: systemd-udevd 261.3-1 severe boot hang/slowdown, USB+DisplayPort broke

#6 Today 14:05:40

Re: systemd-udevd 261.3-1 severe boot hang/slowdown, USB+DisplayPort broke

#7 Today 15:06:09

Re: systemd-udevd 261.3-1 severe boot hang/slowdown, USB+DisplayPort broke

#8 Today 16:21:46

Re: systemd-udevd 261.3-1 severe boot hang/slowdown, USB+DisplayPort broke

Board footer