Intel iGPU Using Both i915 and xe Kernel Modules, Is This an Issue? How to Resolve?

[arazorbacks]

Because the i915 driver looked old, I went back to the same linux-firmware site as before and grabbed the latest i915 driver. Here’s what it shows now (I went from 70.44.1 to 70.49.4). We’ll see if this helps or not. If you’ve had a chance to review my kern.log, it would be great to hear any ideas.

user@server:~$ sudo dmesg | grep -i guc
[    6.238218] xe 0000:00:02.0: [drm] GT0: Using GuC firmware from i915/tgl_guc_70.bin version 70.49.4
[    6.483538] xe 0000:03:00.0: [drm] GT0: Using GuC firmware from xe/bmg_guc_70.bin version 70.58.0
[    6.589417] xe 0000:03:00.0: [drm] GT1: Using GuC firmware from xe/bmg_guc_70.bin version 70.58.0

Edit: Still crashing.

Edit 2: I ran apt list --installed and it doesn’t show any of the Intel packages as installed anymore. I also removed the repo using sudo add-apt-repository --remove ppa:kobuk-team/intel-graphics. Anything else I should check?

[arazorbacks]

Hi, your .deb is causing my machine to crash multiple times a day. I really need you to pop back in here and tell me how to remove whatever it is your .deb installed. My machine was more stable before installing it.

[arazorbacks]

Ok, so no responses from anyone, so here’s what I’ve done. Since I wasn’t sure what all was in juergh’s .deb, I ran the below. Turns out it’s a linux-firmware package with nothing else.

dpkg -f linux-firmware_20240318.git3b128b60-0ubuntu2.25+lp2085434_amd64.deb Package

Knowing that it seemed the best thing to do was to update linux-firmware to the latest release to try to write over it.

sudo apt upgrade linux-firmware

Then I confirmed it’s still using the latest firmware versions which I manually loaded.

user@server:~$ sudo dmesg | grep -i guc
[    6.238218] xe 0000:00:02.0: [drm] GT0: Using GuC firmware from i915/tgl_guc_70.bin version 70.49.4
[    6.483538] xe 0000:03:00.0: [drm] GT0: Using GuC firmware from xe/bmg_guc_70.bin version 70.58.0
[    6.589417] xe 0000:03:00.0: [drm] GT1: Using GuC firmware from xe/bmg_guc_70.bin version 70.58.0

So, who knows, maybe it’ll be more stable. I’ll report back either with another crash or with a note that it’s been stable.

[arazorbacks]

It crashed.

I assume there’s some conflicting driver/firmware from a previous build that’s getting loaded and causing issues. At this point I don’t really have any hope of figuring out specifically what it is, so I decided to just reformat the box. I reinstalled Server LTS 24.04 with the HWE kernel (6.17). The GPU was immediately identified and is working out of the box. It’s been up for a couple days now with no crashes. I’m assuming it’s “fixed”.