Upgrade to linux kernel 6.1 to fix AMDGPU issue

Support
1

Hi guys,
I have been hit by the amdgpu bug which is described in detail in this thread:

Journalctl -S -1h | grep amdgpu yielded this info in line with the aforementioned thread.

mar 16 20:17:12 parrot kernel: amdgpu 0000:07:00.0: amdgpu: [mmhub0] no-retry page fault (src_id:0 ring:40 vmid:5 pasid:0, for process  pid 0 thread  pid 0)
mar 16 20:17:12 parrot kernel: amdgpu 0000:07:00.0: amdgpu:   in page starting at address 0x0000800109820000 from IH client 0x12 (VMC)
mar 16 20:17:12 parrot kernel: amdgpu 0000:07:00.0: amdgpu: VM_L2_PROTECTION_FAULT_STATUS:0x00540050
mar 16 20:17:12 parrot kernel: amdgpu 0000:07:00.0: amdgpu:          Faulty UTCL2 client ID: MP1 (0x0)
mar 16 20:17:12 parrot kernel: amdgpu 0000:07:00.0: amdgpu:          MORE_FAULTS: 0x0
mar 16 20:17:12 parrot kernel: amdgpu 0000:07:00.0: amdgpu:          WALKER_ERROR: 0x0
mar 16 20:17:12 parrot kernel: amdgpu 0000:07:00.0: amdgpu:          PERMISSION_FAULTS: 0x5
mar 16 20:17:12 parrot kernel: amdgpu 0000:07:00.0: amdgpu:          MAPPING_ERROR: 0x0
mar 16 20:17:12 parrot kernel: amdgpu 0000:07:00.0: amdgpu:          RW: 0x1

It seems that the Linux Kernel 6.1 has the fix, and since I’m running the 6.0 included in Parrot 5.2 I got the bug a few times the last few weeks. Interestingly both times it happened it was when I open the lid on my laptop and browse the slack app (from snap store).

How do I install the 6.1 linux Kernel they mention?

uname -a yields:
Linux parrot 6.0.0-12parrot1-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.0.12-1parrot1 (2023-01-12) x86_64 GNU/Linux

2

It seems there’s a debian package available in apt by the name linux-image-6.1.0-0.deb11.5-amd64-unsigned

If I install that with the package manager do I get linux 6.1 (and of course select that option in GRUB at boot)? Sorry for the newbie question :slight_smile:

3

I went ahead and installed the package. I got the following message during installation:

Error! Your kernel headers for kernel 6.1.0-0.deb11.5-amd64 cannot be found at /lib/modules/6.1.0-0.deb11.5-amd64/build or /lib/modules/6.1.0-0.deb11.5-amd64/source.
Please install the linux-headers-6.1.0-0.deb11.5-amd64 package or use the --kernelsourcedir option to tell DKMS where it’s located.

I went ahead and installed the mentioned package as well. I’m now running:

└──╼ $uname -a
Linux parrot 6.1.0-0.deb11.5-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.1.12-1~bpo11+1 (2023-03-05) x86_64 GNU/Linux

I will report back what happens. I’m using Slack on a daily basis, so if I have a problem still I expect it within 2 weeks.

4

You can install another kernel, yes, if you do it manually, then you ‘may’ have to alter your boot config to suit.
This will help you to configurte a different boot kernel;

1 Like
5

Thank you. That was helpful :slight_smile:

It made me realize I had chosen a wrong boot entry because the uppermost one did not mention the kernel version. But it is the 6.1 kernel in the first option as grub-customizer shows me. So the package installation must have gone alright.

I booted the Noveau/Nvidia Off boot option (second from top), which I guess is pretty much without effect since I have a Ryzen CPU with integrated GPU :smiley:

6

I thought I’d give an update.

The system has an uptime for 19 days currently. No crashes yet. It seems better than before at least.

I will report back again later.

└──╼ $uptime
 00:20:21 up 19 days, 57 min,  1 user,  load average: 1,34, 1,20, 1,08
7

Second update. No news. Stable still.

──╼ $uptime
11:20:43 up 40 days, 11:57, 1 user, load average: 1,01, 0,57, 0,39

8

Seems like you didn’t install/upgrade linux-headers when you upgraded the kernel image?

9

Yes I think that’s what happened. I manually installed the header immediately after (before a reboot). It worked out in the end.

10

So i guess everything is fine now :smiley:
p/s: just in case you didn’t know: linux-kernel-image is the binary file of the kernel. The header is short of “table of contents” so the drivers can use the data in kernel’s libraries.

11

This is getting slightly boring. Stable still. The problem is probably fixed.

──╼ $uptime
 22:47:34 up 61 days, 23:24,  1 user,  load average: 1,07, 1,00, 1,01
1 Like
12

Same story … it seems solid. No crash for 88 days. Not sure if anyone is still interested in more updates :smiley:

──╼ $uptime
 14:52:07 up 88 days, 15:28,  1 user,  load average: 0,20, 0,30, 0,42
13

I think its safe to say, “It’s stable.” :laughing:

14

This topic was automatically closed 120 days after the last reply. New replies are no longer allowed.