total computer freeze (Ryzen bug)

argiope · 2025-03-10 13:51:17

So I'm suffering from the well-known 'Ryzen freeze' bug, I've tried following everything I could from https://wiki.archlinux.org/title/Ryzen#Troubleshooting but to no avail so far (it's been about a month).

A bit of context: I bought a second-hand computer, transferred my old system from another machine into the current one without reinstalling archlinux, just by moving the hard drive and adding my system to the preexisting grub config (had to install a few drivers and firmware to make xorg work, as the previous machine had an intel CPU).

Current machine/system specs:
AMD Ryzen 3 2200G with Radeon Vega Graphics
ASRock B450 Gaming K4
Linux 6.12.18-1-lts

The problem: this new PC regularly freezes at completely random intervals, some days none at all, other days it'll freeze 2 or 3 times within the first few hours. The freeze is absolute and leaves me with no other option but to shut down my computer forcefully using the power supply off-switch. The keyboard is entirely unresponsive, I can't switch to any tty or even use the REISUB trick to shut down a bit more cleanly. The freeze happens mostly, if not only, when I'm browsing the web with firefox. My CPU is typically barely solicited before it happens (I did see a couple times my xmobar indicator show a red, 100% used core as the freeze happened), the swap is usually empty. Just once I managed to use REISUB before the freeze completely paralyzed my system, but most of the time I don't have time to react, the freeze takes over in a matter of seconds.

What I've tried:
- the following kernel parameters: rcu_nocbs=0-3 idle=nomwait processor.max_cstate=5 pci=nomsi (all together since I was really desperate to fix the issue and there's absolutely no way to check whether a solution works until the next freeze occurs)
- I've also installed disable-c6-systemd just in case. I tried using 'modprobe msr' as recommended but I'm not 100% sure it worked since `sudo lsmod | grep msr` lists `intel_rapl_msr` and not 'msr' alone, but perhaps it's normal (I know next to nothing about kernels if it's not already obvious).

Please help! I almost got a freeze this morning while updating my computer.

journalctl of boot before freeze: https://pastebin.com/i0an72Cj
journalctl of boot after freeze: https://pastebin.com/BHhqZDCB (I updated my computer then rebooted)

Last edited by argiope (2025-03-10 16:51:44)

Head_on_a_Stick · 2025-03-10 16:54:18

https://wiki.archlinux.org/title/Microcode?

Ketsui · 2025-03-10 16:58:11

Are you still getting frozen on version 6.12.18? The fix for the most recent freezing bug plaguing AMD APUs have finally made it to stable kernels:
https://web.git.kernel.org/pub/scm/linu … 30e0f3a0f3
If you're still getting them then I suggest you remove all those kernel commandline options, run memtest, make sure your RAM overclock is stable (mine wasn't) and that you're on BIOS version Combo-AM4 1.0.0.3 ABB (not sure about this one since you have different chip but my computer has been stable after downgrading), I have the 3200G and I've been getting upwards of 10 days of uptime on my system with the linked fix and this commandline:

root=UUID=96cd82d0-b8b5-4709-8c18-344aa801f00a rw resume=UUID=96cd82d0-b8b5-4709-8c18-344aa801f00a resume_offset=24

As you can see I have nothing fancy going on.

I learned that newer BIOS is buggy from here:
https://www.bleepingcomputer.com/forums … try5012252

seth · 2025-03-10 17:04:38

What I've tried:

doesn't include

the wiki wrote:

To solve this problem you need to supply higher voltage to your CPU so that it is stable when running at peak frequencies. The easiest way to achieve this is to use the AMD curve optimiser which is accessible via your motherboard's UEFI. Access it and put a positive offset…

argiope · 2025-03-10 18:11:58

seth wrote:

What I've tried:
doesn't include
the wiki wrote:
To solve this problem you need to supply higher voltage to your CPU so that it is stable when running at peak frequencies. The easiest way to achieve this is to use the AMD curve optimiser which is accessible via your motherboard's UEFI. Access it and put a positive offset…

I didn't try that because that's from a section about random reboots, which isn't the issue I'm encountering so far.

Ketsui wrote:

Are you still getting frozen on version 6.12.18?

So far no, but there's no guarantee it'll last... Here's hoping, of course! I'll try your ideas as soon as I encounter another freeze with the newer kernel. Thanks!

BTW, stupid question: how can I check my BIOS version? dmidecode doesn't show anything resembling 'Combo-AM4 1.0.0.3 ABB', it just says 'Version: P3.50 / BIOS Revision: 5.14'.

Head_on_a_Stick wrote:

https://wiki.archlinux.org/title/Microcode?

Thanks! I installed amd-ucode, gonna reboot, will keep you posted if I suffer another freeze.

seth · 2025-03-10 20:56:16

I didn't try that because that's from a section about random reboots, which isn't the issue I'm encountering so far.

The system doesn't softlock either - if the CPU enters erratic behavior, everything can happen (though spontanous reboots are the common symptom)

Wrt amd-ucode, what's the ouput of "lscpu"?
According to https://www.cpu-world.com/CPUs/Zen/AMD- … 2200G.html it's 0x17/0x11/0x0 which doesn't seem to be covered by those firmwares, https://gitlab.com/kernel-firmware/linu … type=heads

argiope · 2025-03-10 21:37:44

seth wrote:

Wrt amd-ucode, what's the ouput of "lscpu"?
According to https://www.cpu-world.com/CPUs/Zen/AMD- … 2200G.html it's 0x17/0x11/0x0 which doesn't seem to be covered by those firmwares, https://gitlab.com/kernel-firmware/linu … type=heads

You seem to be right, lscpu says 'family = 23 (ie 0x17), model = 17 (ie 0x11), stepping = 0x0'
So that means that the package I installed does nothing at all? If so, should I uninstall it or keep it in case in the future it gets updated? Can it do any harm?

seth wrote:

I didn't try that because that's from a section about random reboots, which isn't the issue I'm encountering so far.
The system doesn't softlock either - if the CPU enters erratic behavior, everything can happen (though spontanous reboots are the common symptom)

So... you're saying that even though I've never had a single random reboot, and only softlock-freezes (at least from a practical standpoint, maybe it's not technically the underlying issue but it's the symptom I've encountered), it could still happen and it's still the same problem even though it's a different troubleshooting section on that wiki page?

I'm a bit worried about changing the voltage of my CPU based on a fix for a reboot bug that I haven't experienced and that seems to concern a completely different Ryzen CPU family (5000 series)... Could you convince me it's a good idea or at least not dangerous? Because it's really not in the right troubleshooting section and I don't know what increasing the voltage of my CPU will do if it doesn't need it...

seth · 2025-03-10 21:51:01

You'd get out of a softlock freeze w/ sysrq+REISUB and it would show up as such in dmesg/the journal.
Yoiu're locking rather very hard here.

However, looking at the journals you posted

Mar 10 08:05:43 neptunia-hostname kernel: scsi 9:0:0:0: Direct-Access     ATA      TOSHIBA DT01ACA1 A750 PQ: 0 ANSI: 5
Mar 10 08:05:43 neptunia-hostname kernel: sd 9:0:0:0: [sdd] 1953525168 512-byte logical blocks: (1.00 TB/932 GiB)
Mar 10 08:05:43 neptunia-hostname kernel: sd 9:0:0:0: [sdd] 4096-byte physical blocks
Mar 10 08:05:43 neptunia-hostname kernel: sd 9:0:0:0: [sdd] Write Protect is off
Mar 10 08:05:43 neptunia-hostname kernel: sd 9:0:0:0: [sdd] Mode Sense: 00 3a 00 00
Mar 10 08:05:43 neptunia-hostname kernel: sd 9:0:0:0: [sdd] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
Mar 10 08:05:43 neptunia-hostname kernel: sd 9:0:0:0: [sdd] Preferred minimum I/O size 4096 bytes
Mar 10 08:05:43 neptunia-hostname kernel: ata10: log page 10h reported inactive tag 26
Mar 10 08:05:43 neptunia-hostname kernel: ata10.00: exception Emask 0x1 SAct 0x2 SErr 0x0 action 0x0
Mar 10 08:05:43 neptunia-hostname kernel: ata10.00: irq_stat 0x40000008
Mar 10 08:05:43 neptunia-hostname kernel: ata10.00: failed command: READ FPDMA QUEUED
Mar 10 08:05:43 neptunia-hostname kernel: ata10.00: cmd 60/08:08:88:6d:70/00:00:74:00:00/40 tag 1 ncq dma 4096 in
                                                  res 51/40:00:00:00:00/00:00:00:00:00/00 Emask 0x9 (media error)
Mar 10 08:05:43 neptunia-hostname kernel: ata10.00: status: { DRDY ERR }
Mar 10 08:05:43 neptunia-hostname kernel: ata10.00: error: { UNC }
Mar 10 08:05:43 neptunia-hostname kernel: ata10.00: configured for UDMA/133
Mar 10 08:05:43 neptunia-hostname kernel: sd 9:0:0:0: [sdd] tag#1 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_OK cmd_age=0s
Mar 10 08:05:43 neptunia-hostname kernel: sd 9:0:0:0: [sdd] tag#1 Sense Key : Medium Error [current] 
Mar 10 08:05:43 neptunia-hostname kernel: sd 9:0:0:0: [sdd] tag#1 Add. Sense: Unrecovered read error - auto reallocate failed
Mar 10 08:05:43 neptunia-hostname kernel: sd 9:0:0:0: [sdd] tag#1 CDB: Read(10) 28 00 74 70 6d 88 00 00 08 00
Mar 10 08:05:43 neptunia-hostname kernel: I/O error, dev sdd, sector 1953525128 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
Mar 10 08:05:43 neptunia-hostname kernel: Buffer I/O error on dev sdd, logical block 244190641, async page read
…
Mar 10 09:54:14 neptunia-hostname udisksd[114737]: Error probing device: Error sending ATA command IDENTIFY DEVICE to '/dev/sdf': Unexpected sense data returned:
                                                  0000: 00 00 00 00  00 00 00 00  00 00 00 00  00 00 00 00    ................
                                                  0010: 00 00 00 00  00 00 00 00  00 00 00 00  00 00 00 00    ................
                                                   (g-io-error-quark, 0)

That disk is likely toast (doesn't seem to be a relevant drive, though? And that would still not preclude the sysrq, even if syncing fails)

Just once I managed to use REISUB before the freeze completely paralyzed my system

Do you have the journal of that boot?

Gradually increasing the offset in the curve optimizer will not burn your CPU, holding a 220V cable at it will.
You're not supposed to overclock anything here, but adjust the settings within the specified range.

Ketsui · 2025-03-11 00:26:49

argiope wrote:

BTW, stupid question: how can I check my BIOS version? dmidecode doesn't show anything resembling 'Combo-AM4 1.0.0.3 ABB', it just says 'Version: P3.50 / BIOS Revision: 5.14'.

By looking at the support page for your motherboard model, if this is what you have then version 3.50 should be it:
https://www.asrock.com/mb/amd/fatal1ty% … x.asp#BIOS

agapito · 2025-03-11 09:05:58

seth wrote:

Gradually increasing the offset in the curve optimizer will not burn your CPU.

That's right. The problem here is Curve Optimizer is for Zen3/4/5 CPUs. The only thing that might work here, with that CPU, is a traditional positive voltage offset which does increase the voltage, but may solve the problem in case of CPU degradation.

Ketsui wrote:

argiope wrote:
BTW, stupid question: how can I check my BIOS version? dmidecode doesn't show anything resembling 'Combo-AM4 1.0.0.3 ABB', it just says 'Version: P3.50 / BIOS Revision: 5.14'.
By looking at the support page for your motherboard model, if this is what you have then version 3.50 should be it:
https://www.asrock.com/mb/amd/fatal1ty% … x.asp#BIOS

"ASRock do NOT recommend updating this BIOS if Pinnacle, Raven or Summit Ridge CPU is being used on your system."

2200G is a Raven Ridge CPU: https://www.techpowerup.com/cpu-specs/r … 200g.c1978

argiope wrote:

The freeze happens mostly, if not only, when I'm browsing the web with firefox.

Have you tried disabling graphics acceleration in Firefox?

Arch Linux

#1 2025-03-10 13:51:17

total computer freeze (Ryzen bug)

#2 2025-03-10 16:54:18

Re: total computer freeze (Ryzen bug)

#3 2025-03-10 16:58:11

Re: total computer freeze (Ryzen bug)

#4 2025-03-10 17:04:38

Re: total computer freeze (Ryzen bug)

#5 2025-03-10 18:11:58

Re: total computer freeze (Ryzen bug)

#6 2025-03-10 20:56:16

Re: total computer freeze (Ryzen bug)

#7 2025-03-10 21:37:44

Re: total computer freeze (Ryzen bug)

#8 2025-03-10 21:51:01

Re: total computer freeze (Ryzen bug)

#9 2025-03-11 00:26:49

Re: total computer freeze (Ryzen bug)

#10 2025-03-11 09:05:58

Re: total computer freeze (Ryzen bug)

Board footer