[ Solved] Fail to restart after suspend

[waynus]

An update tried the 5.13.0-22 kernel.

Does not seem to have made a difference.

https://termbin.com/4x1c


To clarify what I observe the computer goes into suspend, then on touching the keyboard it and the mouse light up the hard drive light shows activity the monitors activate so must be getting a signal, but that is it, nothing else happens and I have to hard reboot to get any response from the computer.

[SMG]

To clarify what I observe the computer goes into suspend,

It looks like you let it stay suspended for 65 seconds. Does that sound correct?

then on touching the keyboard it and the mouse light up the hard drive light shows activity the monitors activate so must be getting a signal, but that is it, nothing else happens and I have to hard reboot to get any response from the computer.

Unfortunately, I see pretty much the same errors as in the prior logs. It seems that proposed kernel change did not apply for your system.

Code: Select all

Nov 30 05:19:16 wayne-System-Product-Name kernel: amdgpu 0000:09:00.0: amdgpu: RAP: optional rap ta ucode is not available
Nov 30 05:19:16 wayne-System-Product-Name kernel: amdgpu 0000:09:00.0: amdgpu: RAS: optional ras ta ucode is not available
Nov 30 05:19:16 wayne-System-Product-Name kernel: amdgpu 0000:09:00.0: amdgpu: SECUREDISPLAY: securedisplay ta ucode is not available
Nov 30 05:19:16 wayne-System-Product-Name kernel: amdgpu 0000:09:00.0: amdgpu: RunBtc failed!
Nov 30 05:19:16 wayne-System-Product-Name kernel: amdgpu 0000:09:00.0: amdgpu: Failed to setup smc hw!
Nov 30 05:19:16 wayne-System-Product-Name kernel: [drm:amdgpu_device_ip_resume_phase2 [amdgpu]] *ERROR* resume of IP block <smu> failed -62
Nov 30 05:19:16 wayne-System-Product-Name kernel: amdgpu 0000:09:00.0: amdgpu: amdgpu_device_ip_resume failed (-62).
Nov 30 05:19:16 wayne-System-Product-Name kernel: PM: dpm_run_callback(): pci_pm_resume+0x0/0x90 returns -62
Nov 30 05:19:16 wayne-System-Product-Name kernel: amdgpu 0000:09:00.0: PM: failed to resume async: error -62

And then the next set. There was more after that, but it was all kinds of repeats. You are correct that the system itself returns from suspend, but the GPU is not.

Code: Select all

Nov 30 05:19:16 wayne-System-Product-Name kernel: amdgpu: Move buffer fallback to memcpy unavailable
Nov 30 05:19:16 wayne-System-Product-Name kernel: [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to process the buffer list -19!
Nov 30 05:19:16 wayne-System-Product-Name kernel: amdgpu: Move buffer fallback to memcpy unavailable
Nov 30 05:19:16 wayne-System-Product-Name kernel: [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to process the buffer list -19!
Nov 30 05:19:17 wayne-System-Product-Name kernel: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_hubp_pg_control line:539
Nov 30 05:19:17 wayne-System-Product-Name kernel: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_dpp_pg_control line:465
Nov 30 05:19:17 wayne-System-Product-Name kernel: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_hubp_pg_control line:531
Nov 30 05:19:17 wayne-System-Product-Name kernel: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_dpp_pg_control line:457
Nov 30 05:19:17 wayne-System-Product-Name kernel: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_hubp_pg_control line:523
Nov 30 05:19:17 wayne-System-Product-Name kernel: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_dpp_pg_control line:449
Nov 30 05:19:17 wayne-System-Product-Name kernel: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_hubp_pg_control line:515
Nov 30 05:19:17 wayne-System-Product-Name kernel: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_dpp_pg_control line:441
Nov 30 05:19:17 wayne-System-Product-Name kernel: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_hubp_pg_control line:507
Nov 30 05:19:17 wayne-System-Product-Name kernel: [drm] REG_WAIT timeout 1us * 1000 tries - dcn20_dpp_pg_control line:433
Nov 30 05:19:17 wayne-System-Product-Name kernel: [drm] perform_link_training_with_retries: Link training attempt 1 of 4 failed
Nov 30 05:19:18 wayne-System-Product-Name kernel: [drm] perform_link_training_with_retries: Link training attempt 2 of 4 failed
Nov 30 05:19:18 wayne-System-Product-Name kernel: [drm] perform_link_training_with_retries: Link training attempt 3 of 4 failed
Nov 30 05:19:19 wayne-System-Product-Name kernel: [drm] enabling link 0 failed: 15

[waynus]

Yes about two minutes so 65 seconds seems correct.

I am thinking it will be a waiting game, found the same error type of problem in many places with amdgpu.

Much more of an issue though in laptops where it is doing this on closing the lid.

Found a couple of ideas
1) https://archived.forum.manjaro.org/t/so ... eep/152183

Which suggest disabling IOMMU in the bios. I don't think it will hurt but not sure how it would help.

2) https://github.com/jrandiny/yoga-slim7-ubuntu

A more involved approach with s3 but again I am doubtful it will do the trick either.

[SMG]

I am thinking it will be a waiting game, found the same error type of problem in many places with amdgpu.

Unfortunately, that is all I can suggest.

The links you listed seem to have slightly different messages, but the first one is also a desktop, so it might be worth a try. There are specific messages in a log when disabling IOMMU might be needed and I didn't see those in yours, but I do not have any other ideas.

[waynus]

Thanks.

Such is the nature of computers. At least the system is useable and I can do all that I need to on it.

[waynus]

I tried the new Manjaro.

Nice distribution and an easy entry to Arch.

Sadly same problem.

[SMG]

I tried the new Manjaro...Sadly same problem.

What kernel version is it using?

[waynus]

I didn't check. Will do later.

[deepakdeshp]

Code: Select all

 uname -r

will give the version you are running.

I didn't check. Will do later.

[waynus]

The kernel is 15.15.7-1 and 5.16 rc4

Both do exactly the same as on mint.

It is disappointing as I bought this as my main machine to run Linux and I am back to Windows 11 because it works properly and I leave the machine on 24/7.

I run Mint on a number of other machines without a problem, but on this machine it looks like a bust. That it did work at one time and now doesnt is frustrating.

[SMG]

The kernel is 15.15.7-1 and 5.16 rc4

Both do exactly the same as on mint.

The amdgpu driver is in the kernel so it is not encouraging new the problem is still there in later mainline kernels. I do not know if Canonical is working with AMD to try and get a fix in the Ubuntu/Mint kernels.

[waynus]

I tried disabling IOMMU.

It had 3 settings enabled, disabled and auto. It was set to auto.

When I set it to disabled suspend works.

Problem solved. Of course the question is why?

[SMG]

I tried disabling IOMMU.

It had 3 settings enabled, disabled and auto. It was set to auto.

When I set it to disabled suspend works.

Does it work for any kernel or just specific ones?

[waynus]

Works with 5.4 and 5.13.

[SMG]

Thanks for the feedback.
I do not know how IOMMU relates to your specific motherboard to know why that setting is making a difference.

[Da Londo]

Just leaving a note here about the issue, got the same with 5.13 but with 5.11 all is well, checked IOMMU and was disabled by default.
Asrock B450M Pro with AMD Ryzen 3 3200G.

[SMG]

Moderator note: linuxovios issue can now be found here Fail to restart after suspend.