systemd failure na Ryzen 3000

AMD Ryzen 7 3700X + Ryzen 9 3900X Offer Incredible Linux Performance But With A Big Caveat Review - Phoronix

So what’s the problem? On newer Linux distributions, there’s a hard regression either within the kernel but more likely some cross-kernel/user-space interaction issue leaving newer Linux distributions unbootable.

While Ubuntu 18.04 LTS and older Linux distributions boot Zen 2, to date I have not been able to successfully boot the likes of Ubuntu 19.04, Manjaro Linux, and Fedora Workstation 31. On all newer Linux distributions I’ve tried on two different systems built around the Ryzen 7 3700X and Ryzen 9 3900X, each time early in the boot process as soon as trying to start systemd services, all systemd services fail to start.

I’ve confirmed with AMD they do have an open issue surrounding “5.0.9” (the stock kernel of Ubuntu 19.04) but as of writing hadn’t shed any light into the issue. AMD has said their testing has been mostly focused on Ubuntu 18.04 given its LTS status. I’ve also confirmed the same behavior with some other Windows reviewers who occasionally dabble with Linux.

So unfortunately not being able to boot newer Linux distributions is a huge pain. I’ve spent days trying different BIOS versions/options, different kernel command line parameters, and other options to no avail. On some Linux distributions after roughly 20~30 minutes of waiting after all systemd services fail to start, sometimes there will be a kernel panic but that hadn’t occurred on all systems at least not within that time-frame.

Complicating this issue further is that when trying Linux 5.0/5.1/5.2 kernels on Ubuntu 18.04 LTS, they all booted up fine! Hence it doesn’t appear to simply be a kernel regression but either some kernel+systemd interaction issue happening just on AMD Ryzen 3000 series processors or some awkward thing happening with the X570 chipset. When asking AMD again on the day prior to launch, they were still investigating the issue and apparently unsure of the root cause at this time.

I’ve spent much time trying to figure out this issue and hoping some BIOS upgrade or settings adjustment or kernel command-line parameter exists to workaround this problem but so far have been unsuccessful in the time available. Otherwise basically all recent/new Linux distributions appear affected, at least those using systemd, appear impacted,


Eh, systemd, sta faking init sistem ima sa hardware-om?

Eh, systemd, sta faking init sistem ima sa hardware-om?

Khm…čiji je ovo komentar?

Moj, nisam izvadio iz quote. Ispalo je da systemd koristi rdrand umesto /dev/urandom, a ovi novi cipovi vracaju -1 kad nema entropije
sto ga zbunjuje i onda fejluje.

Кад налетим на овакве преписке… осећам се као ћопави слепац који прелази ибарску магистралу суботом у 2 сата ујутру… Ја се грчим ли грчим са неким глупостима

Kod Linuxa je sve lepo odvojeno i složeno, no izgleda da je systemd napravljen tako da preuzme sve bitne funkcije sistema i da nije samo init

SystemD je teška prevara koja je napravljena da uništi slobodan softver

systemd nije u duhu Unix filozofije gde se sistem sastoji iz delova koji rade pojedinacan posao dobro. systemd je hobotnica koja
hoce da preuzme ceo sistem i zbog toga pavi mnogo problema. Pazi systemd preuzima i dns?! Koliko god da sam indiferentan
systemd pocinje da smeta…

Знам да је овде већина (ових искуснијих) против системД-а, ал реко само да поделим кратак исечак. Дакле почев од 20мин и 30 секунди до неких 21мин и 20 секунди. За то се мора одати признање ;). Мислим да сам наштеловао линк да пусти од правог места:;t=1230