Mailinglist Archive: opensuse-factory (745 mails)

< Previous Next >
[opensuse-factory] amd ryzen low load hard freeze kernel
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

since nearly 6 month now i have trouble with
my unstable ryzen systems.

randmom crashes after a couple of hours (<8), (nearly) no load, most times only
hard reset
is possible to start system new.
most times nothing special in system-log
(as i know, maybe somebody who know more could see suspected things.)

after searching the net:

because i am not really familiar with kernel-boot parameter settings,
and i have read its necessary to build own kernel (on lists of other
distributions)

i like to ask:

how to fix (work around) or if maybe included in
the tumbleweed-kernel the "low load freeze of ryzen"
here: ryzen (1700) 8 core processor systems with tumbleweed.
i read several links found by google:

https://bugzilla.kernel.org/show_bug.cgi?id=196683
https://bugs.launchpad.net/linux/+bug/1690085/comments/69
https://forums.fedoraforum.org/showthread.php?315887-Random-Crashing

they suggest to include rcu_nocbs=0-15
is /etc/defaults/grub file
correct?
GRUB_CMDLINE_LINUX_DEFAULT="video=1920x1200 splash=silent quiet showopts"
change to:
GRUB_CMDLINE_LINUX_DEFAULT="video=1920x1200 splash=silent quiet showopts
rcu_nocbs=0-15 "
??
and run:
grub2-mkconfig?

is this correct will this work? or is there a better/other solution for
tumbleweed?

(at the moment here not the newest version of tumbleweed):
how to check the tumbleweed version?
uname -a
Linux becherer1 4.14.2-1-default #1 SMP PREEMPT Fri Nov 24 08:20:07 UTC 2017
(b0610fc) x86_64 x86_64 x86_64 GNU/Linux

============
by the way, as suggested by this mailing-list i have replaced the amd ryzen
prozessors because of the
segfault problem (i found this problem for my systemsy after the hint on this
list).
tumbleweed is now running fine with HIGH load (kill-ryzen.sh) script (more than
48 hours)

and for others who read inside internet "amd ryzen production date "after week
25 (sometimes is written week30) is fine"
THIS is NOT TRUE. at least this type:
ryzen 1700 UA 1733PGS
has this bug, (2017 week 33) (i have had after a first replacement 4 of this
processors all have
had this bug) see also here:
https://forum.level1techs.com/t/ryzen-pre-week-25-fabrication-rma-issue/118658/98
==============================

would be happy to get an info how to fix the low load problem to get stable
tumbleweed-ryzen-system's.

thanks in advance,

simoN




-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)
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=aB7T
-----END PGP SIGNATURE-----
--
To unsubscribe, e-mail: opensuse-factory+unsubscribe@xxxxxxxxxxxx
To contact the owner, e-mail: opensuse-factory+owner@xxxxxxxxxxxx

< Previous Next >
Follow Ups