Freeze

seth · 2024-12-24 20:21:45

Dec 24 11:34:21 aPC kernel: pci 0000:03:00.0: vgaarb: VGA device added: decodes=io+mem,owns=none,locks=none
Dec 24 11:34:21 aPC kernel: pci 0000:0e:00.0: vgaarb: bridge control possible
Dec 24 11:34:21 aPC kernel: pci 0000:0e:00.0: vgaarb: VGA device added: decodes=io+mem,owns=none,locks=none
---
Dec 24 11:34:21 aPC kernel: amdgpu 0000:03:00.0: amdgpu: Fetched VBIOS from VFCT
Dec 24 11:34:21 aPC kernel: amdgpu: ATOM BIOS: 113-EXT90455-100
Dec 24 11:34:21 aPC kernel: amdgpu 0000:03:00.0: amdgpu: CP RS64 enable
Dec 24 11:34:21 aPC kernel: amdgpu 0000:0e:00.0: enabling device (0006 -> 0007)
Dec 24 11:34:21 aPC kernel: amdgpu 0000:0e:00.0: amdgpu: Fetched VBIOS from VFCT
Dec 24 11:34:21 aPC kernel: amdgpu: ATOM BIOS: 102-RAPHAEL-008
---
Dec 24 11:34:21 aPC kernel: [drm] Initialized amdgpu 3.59.0 for 0000:03:00.0 on minor 1
Dec 24 11:34:21 aPC kernel: fbcon: amdgpudrmfb (fb0) is primary device
Dec 24 11:34:21 aPC kernel: [drm] Initialized amdgpu 3.59.0 for 0000:0e:00.0 on minor 0
Dec 24 11:34:21 aPC kernel: amdgpu 0000:0e:00.0: [drm] fb1: amdgpudrmfb frame buffer device
---
Dec 24 11:34:26 aPC kernel: amdgpu 0000:0e:00.0: [drm] REG_WAIT timeout 1us * 100000 tries - optc31_disable_crtc line:145
Dec 24 12:08:48 aPC kernel: amdgpu 0000:03:00.0: [drm] *ERROR* [CRTC:79:crtc-0] flip_done timed out

would be the CPU's

Yes.

which I thought was off

No

irarch · 2024-12-24 20:25:30

OK thank you for catching that. I did have integrated Graphics on AUTO and it is now disabled.
Time will tell if it effects the freeze but I'm glad to have it disabled in any case.

irarch · 2024-12-25 00:51:14

It still freezes with only 1 GPU enabled. So I've re-enabled the integrated graphics. I remembered the reason I had it on was so I could use the onboard HDMI port as an audio out.

OJaksch · 2024-12-25 06:14:47

Alas, after editing my post in #46 yesterday I turned over to Discord and Elements and just a sec later my system froze again, so amdgpu.mcbp=0 isn't the definitive solution.

seth · 2024-12-25 09:01:48

Unfortunately fits https://bbs.archlinux.org/viewtopic.php … 7#p2216437

Is it only electron clients that cause this?
And do you also have "amdgpu_device_ip_suspend_phase1" in the most recent kernel backtrace then?

OJaksch · 2024-12-25 10:56:47

seth wrote:

Unfortunately fits https://bbs.archlinux.org/viewtopic.php … 7#p2216437
Is it only electron clients that cause this?

It looks like and this is also my thought...

And do you also have "amdgpu_device_ip_suspend_phase1" in the most recent kernel backtrace then?

Unsure, nor that nor similar messages(s) in the journal. It is almost a "amdgpu: Dumping IP State" to the beginning of a freeze.
Used Kernel parameters:

acpi_os_name="Linux" acpi_osi="Linux" clocksource=hpet mitigations=off quiet amdgpu.noretry=0 amdgpu.mcbp=0

Dez 24 21:10:34 ... akonadiserver[4010]: org.kde.pim.akonadiserver.search: Executing search "searchUpdate-1735071034"
Dez 24 21:10:34 ... akonadiserver[4010]: org.kde.pim.akonadiserver.search: Search  "searchUpdate-1735071034" done (without remote search)
Dez 24 21:10:34 ... akonadiserver[4010]: org.kde.pim.akonadiserver.search: Search update for collection "Letzte Suche" ( 546 ) finished: all results:  11516 , removed results: 0
Dez 24 21:12:13 ... element-desktop[2882]: (node:2882) MaxListenersExceededWarning: Possible EventEmitter memory leak detected. 11 serverSupportedVersions listeners added to [IpcMainI>
Dez 24 21:12:14 ... element-desktop[2882]: (node:2882) MaxListenersExceededWarning: Possible EventEmitter memory leak detected. 11 userAccessToken listeners added to [IpcMainImpl]. Ma>
Dez 24 21:12:14 ... element-desktop[2882]: (node:2882) MaxListenersExceededWarning: Possible EventEmitter memory leak detected. 11 homeserverUrl listeners added to [IpcMainImpl]. MaxL>
Dez 24 21:12:14 ... element-desktop[2882]: (node:2882) MaxListenersExceededWarning: Possible EventEmitter memory leak detected. 11 userAccessToken listeners added to [IpcMainImpl]. Ma>
Dez 24 21:13:52 ... kernel: amdgpu 0000:04:00.0: amdgpu: Dumping IP State
Dez 24 21:13:57 ... kernel: [drm:amdgpu_dm_atomic_check [amdgpu]] *ERROR* [CRTC:73:crtc-0] hw_done or flip_done timed out
Dez 24 21:13:58 ... pipewire[2637]: spa.alsa: front:2p: (0 suppressed) snd_pcm_avail after recover: Datenübergabe unterbrochen (broken pipe)
Dez 24 21:14:04 ... kernel: INFO: NMI handler (perf_event_nmi_handler) took too long to run: 71.426 msecs
Dez 24 21:14:04 ... kernel: perf: interrupt took too long (562878 > 6132), lowering kernel.perf_event_max_sample_rate to 300
Dez 24 21:14:04 ... kernel: hrtimer: interrupt took 142880345 ns
Dez 24 21:14:08 ... kernel: [drm:amdgpu_dm_atomic_check [amdgpu]] *ERROR* [CRTC:77:crtc-1] hw_done or flip_done timed out
Dez 24 21:14:10 ... kernel: perf: interrupt took too long (750403 > 703597), lowering kernel.perf_event_max_sample_rate to 300
Dez 24 21:14:17 ... pipewire[2637]: spa.alsa: front:2p: (1 suppressed) snd_pcm_avail after recover: Datenübergabe unterbrochen (broken pipe)
Dez 24 21:14:17 ... pipewire[2637]: spa.alsa: front:2p: snd_pcm_mmap_commit error: Datenübergabe unterbrochen (broken pipe)
Dez 24 21:14:20 ... kernel: perf: interrupt took too long (1201097 > 938003), lowering kernel.perf_event_max_sample_rate to 300
Dez 24 21:14:24 ... pipewire[2637]: pw.node: (alsa_output.pci-0000_04_00.6.analog-stereo-63) graph xrun not-triggered (0 suppressed)
Dez 24 21:14:24 ... pipewire[2637]: pw.node: (alsa_output.pci-0000_04_00.6.analog-stereo-63) xrun state:0x78eb63d44008 pending:1/2 s:57902787364731 a:57902787502320 f:57902787529838 w>
Dez 24 21:14:24 ... pipewire[2637]: pw.node: (FluidSynth-82) xrun state:0x78eb635a4008 pending:0/1 s:57907216069178 a:57907216087617 f:57902787484860 waiting:18439 process:18446744069>
Dez 24 21:14:26 ... pipewire[2637]: spa.alsa: front:2p: (0 suppressed) snd_pcm_avail after recover: Datenübergabe unterbrochen (broken pipe)
Dez 24 21:14:36 ... pipewire[2637]: pw.node: (alsa_output.pci-0000_04_00.6.analog-stereo-63) graph xrun not-triggered (16 suppressed)
Dez 24 21:14:36 ... pipewire[2637]: pw.node: (alsa_output.pci-0000_04_00.6.analog-stereo-63) xrun state:0x78eb63d44008 pending:1/2 s:57919430567723 a:57919430736671 f:57919430772639 w>
Dez 24 21:14:36 ... pipewire[2637]: pw.node: (FluidSynth-82) xrun state:0x78eb635a4008 pending:0/1 s:57919501861295 a:57919501877359 f:57919430726753 waiting:16064 process:18446744073>larr

seth · 2024-12-25 14:22:58

That looks more like the page flip timeouts irarch and others experience like in https://bbs.archlinux.org/viewtopic.php … 9#p2214959
Would you say amdgpu.mcbp=0 had some impact on frequency or trigger conditions at all? Or was the quiet phase an isolate fluke?

Also, is this X11 or wayland (and are those electron clients you mentioned maybe running and causing this via xwayland)?

OJaksch · 2024-12-25 15:23:50

seth wrote:

Would you say amdgpu.mcbp=0 had some impact on frequency or trigger conditions at all? Or was the quiet phase an isolate fluke?

Hadn't done any measurement but system felt normal, so "if it's fast enough then it's okay". I'm no gamer, building and compiling is of more interest

Also, is this X11 or wayland (and are those electron clients you mentioned maybe running and causing this via xwayland)?

This is X11/Plasma. The programs I'm using on a normal day, means most of the time, are LibreWolf, Kontact, Yakuake, Element, Vesktop and a self made Nativefier version of HomeAssistant. Dunno if one of them is using xwayland per default, but Plasma depends on it.
Sometimes I'm playing around with Wayland/Plasme, but I'm not happy with that (with the behaviour of windows management) for now.

seth · 2024-12-25 15:30:09

No, I menat "frequency of the crashes"

Dunno if one of them is using xwayland per default

Nothing uses xwayland in an X11 session (unless you're nesting a wayland server and there run an X11 client w/o access to the surrounding X11 server)

OJaksch · 2024-12-25 15:53:53

seth wrote:

No, I menat "frequency of the crashes"

I should read the whole sentence, not only the buzzwords Hmm, I'd say it depends on how often and how fast I'm switching to the programs and using them. Yesterday was a good day while using amdgpu.mcbp=0... 16 hrs of undisturbed work, sometimes with playing music. But that was xmas and not so much tasks and less switching between applications. No hibernation or similar done. My displays are turning off themself after 5 mins w/o presence in front of them.
But at the end I turned over to Vesktop (OSS Discord) while moving the mouse across the screens and everything froze.

The days before using amdgpu.mcbp=0 were bad days. Only some minutes of work could be done before the system hangs.

Nothing uses xwayland in an X11 session (unless you're nesting a wayland server and there run an X11 client w/o access to the surrounding X11 server)

Thanks, that has now been clarified.

Last edited by OJaksch (2024-12-25 15:56:57)

Arch Linux

#51 2024-12-24 20:21:45

Re: Freeze

#52 2024-12-24 20:25:30

Re: Freeze

#53 2024-12-25 00:51:14

Re: Freeze

#54 2024-12-25 06:14:47

Re: Freeze

#55 2024-12-25 09:01:48

Re: Freeze

#56 2024-12-25 10:56:47

Re: Freeze

#57 2024-12-25 14:22:58

Re: Freeze

#58 2024-12-25 15:23:50

Re: Freeze

#59 2024-12-25 15:30:09

Re: Freeze

#60 2024-12-25 15:53:53

Re: Freeze

Board footer