[SOLVED] Intel UHD used instead of Iris XE

Bitals · 2023-05-16 20:24:41

Hello,

I bought a Framework 13 12th gen laptop after my old one died, carried my old SSD with Arch already installed and am now unable to use Iris XE graphics driver.
I have dual channel memory.

lspci -nnkv | grep -i VGA  
00:02.0 VGA compatible controller [0300]: Intel Corporation Alder Lake-P Integrated Graphics Controller [8086:46a6] (rev 0c) (prog-if 00 [VGA controller])

According to the wiki (https://wiki.archlinux.org/title/Framew … op_13#i915) page on the laptop, Iris XE has a PCI ID of 8086:9a49, and it is not listed.
So understandably `vaainfo`, `glxinfo32`, `inxi` and everything else report i915, Iris XE is not even detected by kernel as a device.

I run TLP, tried disabling it and rebooting, no change.
Early KMS is set.

Linux MLa 6.3.2-zen1-1-zen #1 ZEN SMP PREEMPT_DYNAMIC Thu, 11 May 2023 16:40:19 +0000 x86_64 GNU/Linux

> sudo dmesg | grep i915
[    2.915619] i915 0000:00:02.0: [drm] VT-d active for gfx access
[    2.915788] i915 0000:00:02.0: vgaarb: deactivate vga console
[    2.915840] i915 0000:00:02.0: [drm] Using Transparent Hugepages
[    2.916325] i915 0000:00:02.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=io+mem:owns=io+mem
[    2.919672] i915 0000:00:02.0: [drm] Finished loading DMC firmware i915/adlp_dmc.bin (v2.19)
[    2.954322] i915 0000:00:02.0: [drm] GT0: GuC firmware i915/adlp_guc_70.bin version 70.5.1
[    2.954327] i915 0000:00:02.0: [drm] GT0: HuC firmware i915/tgl_huc.bin version 7.9.3
[    2.969319] i915 0000:00:02.0: [drm] HuC authenticated
[    2.970554] i915 0000:00:02.0: [drm] GT0: GuC submission enabled
[    2.970556] i915 0000:00:02.0: [drm] GT0: GuC SLPC enabled
[    2.971093] i915 0000:00:02.0: [drm] GuC RC: enabled
[    2.971912] i915 0000:00:02.0: [drm] Protected Xe Path (PXP) protected content support initialized
[    4.136796] [drm] Initialized i915 1.6.0 20201103 for 0000:00:02.0 on minor 0
[    4.143789] fbcon: i915drmfb (fb0) is primary device
[    4.143796] i915 0000:00:02.0: [drm] fb0: i915drmfb frame buffer device
[    4.888491] i915 0000:00:02.0: [drm] Selective fetch area calculation failed in pipe A
[    9.518589] snd_hda_intel 0000:00:1f.3: bound 0000:00:02.0 (ops i915_audio_component_bind_ops [i915])
[    9.522108] mei_pxp 0000:00:16.0-fbf6fcf1-96cf-4e2e-a6a6-1bab8cbe36b1: bound 0000:00:02.0 (ops i915_pxp_tee_component_ops [i915])
[    9.571236] mei_hdcp 0000:00:16.0-b638ab7e-94e2-4ea2-a552-d1c54b627f04: bound 0000:00:02.0 (ops i915_hdcp_component_ops [i915])

Works on W*ndows, so not a hardware/bios configuration issue.

https://share.bitals.xyz/s/EqDYPWTrCHT8 … 000856.png

Last edited by Bitals (2023-05-18 08:25:04)

V1del · 2023-05-16 22:19:51

Don't grep lspci, if it's just a 3D accelerator with no outputs it will show up as 3D controller not as a VGA port. FWIW that i915 is listed is completely normal that's the same kernel driver for all cards, and according to the windows screenshot you just have a single GPU. SO I don't see anything that's off or wrong here. If you hang yourself up on i915, then don't that's the same kernel driver but e.g. glxinfo -B should show the iris driver getting used. The takeaway in your dmesg that something Xe/Iris related is happening should be

 Protected Xe Path (PXP) protected content support initialized

Bitals · 2023-05-16 23:07:18

Thanks for the reply,

V1del wrote:

glxinfo -B should show the iris driver getting used.

It doesn't. I googled, Iris GPU is displayed as a separate PCI device with separate ID for people, indeed with i915 kernel driver, but that's not the point here. The issue is it isn't displayed at all, and every diagnostic utility reports Mesa (not kernel) UHD driver being used instead of Iris.

Sorry, I will quote the exact output tomorrow, it's 2 AM here.

V1del wrote:

The takeaway in your dmesg that something Xe/Iris related is happening should be
 Protected Xe Path (PXP) protected content support initialized

That's GuC and HuC firmware loading I enabled on my way debugging this. Didn't force Iris driver to load as I hoped it would.

Last edited by Bitals (2023-05-16 23:10:22)

V1del · 2023-05-16 23:16:13

Generally speaking, if you use the mesa package you are using the iris driver for any GPU that is newer than skylake most of the "marketing terms" are unlikely to be hugely relevant. But if you are indeed searching for a differing PID, and it's not showing up on lspci at all, then this is at a way earlier stage than mesa or the i915 kernel driver, so this is either BIOS/UEFI related or you'd have to check dmesg for some PCI loading failures.

Bitals · 2023-05-17 00:34:14

V1del wrote:

glxinfo -B should show the iris driver getting used.

> glxinfo -B  
name of display: :0
display: :0  screen: 0
direct rendering: Yes
Extended renderer info (GLX_MESA_query_renderer):
    Vendor: Intel (0x8086)
    Device: Mesa Intel(R) Graphics (ADL GT2) (0x46a6)
    Version: 23.0.3
    Accelerated: yes
    Video memory: 15703MB
    Unified memory: yes
    Preferred profile: core (0x1)
    Max core profile version: 4.6
    Max compat profile version: 4.6
    Max GLES1 profile version: 1.1
    Max GLES[23] profile version: 3.2
OpenGL vendor string: Intel
OpenGL renderer string: Mesa Intel(R) Graphics (ADL GT2)
OpenGL core profile version string: 4.6 (Core Profile) Mesa 23.0.3
OpenGL core profile shading language version string: 4.60
OpenGL core profile context flags: (none)
OpenGL core profile profile mask: core profile

OpenGL version string: 4.6 (Compatibility Profile) Mesa 23.0.3
OpenGL shading language version string: 4.60
OpenGL context flags: (none)
OpenGL profile mask: compatibility profile

OpenGL ES profile version string: OpenGL ES 3.2 Mesa 23.0.3
OpenGL ES profile shading language version string: OpenGL ES GLSL ES 3.20

> lspci -nnkv | grep 46a6
00:02.0 VGA compatible controller [0300]: Intel Corporation Alder Lake-P Integrated Graphics Controller [8086:46a6] (rev 0c) (prog-if 00 [VGA controller])
> lspci -nnkv | grep 9a49

V1del wrote:

so this is either BIOS/UEFI related

Shouldn't be available in Windows then

V1del wrote:

or you'd have to check dmesg for some PCI loading failures.

> sudo dmesg|grep -i PCI

https://pastebin.com/04FWe5K3

Last edited by Bitals (2023-05-17 02:00:49)

Lone_Wolf · 2023-05-17 11:16:16

Please stop grepping outputs unless specifically asked to use it.

Post the full output of lspci -k and (run with root rights) dmesg and journalctl -b . The contents of your /etc/mkinitcpio.conf would also be useful .

Bitals · 2023-05-17 11:30:35

Lone_Wolf wrote:

Post the full output of lspci -k and (run with root rights) dmesg and journalctl -b . The contents of your /etc/mkinitcpio.conf would also be useful .

https://share.bitals.xyz/s/QmSm9rw7KRWrJS2
Viewable in browser.

seth · 2023-05-17 13:16:17

Please use ix.io or 0x0.st

8086:9a49 is the TigerLake chip, you've an AlderLake which correctly shows up as 8086:46a6
https://en.wikipedia.org/wiki/Alder_Lake
https://en.wikipedia.org/wiki/Tiger_Lake

The windows screeshot is useless, look at the actual device and which PCI ID shows up.

Speaking of which:
3rd link below. Mandatory.
Disable it (it's NOT the BIOS setting!) and reboot windows and linux twice for voodo reasons.

Bitals · 2023-05-17 16:32:50

seth wrote:

Speaking of which:
3rd link below. Mandatory.
Disable it (it's NOT the BIOS setting!) and reboot windows and linux twice for voodo reasons.

I did this about 3 years ago on this installation. Are some logs indicating that I didn't? Just checked, FastStart didn't reenable itself or anything. Disabled hibernation additionally, but I am pretty sure haven't spend enough time in Windows to trigger it once for all this years.

seth wrote:

The windows screeshot is useless, look at the actual device and which PCI ID shows up.

Hm, from what I've seen iHD and Iris XE show as 2 different devices in Linux, well yes I got the PCI ID wrong, but it was weird for me only one was displayed.
PCI ID for Iris XE in Windows was indeed 46a6, but what weirded me out was no mention of the Iris driver in any diagnostic utuls (like glxinfo).

seth wrote:

0x0.st

Thanks, will switch to that on this forum.

seth · 2023-05-17 16:38:05

That's not weird.
nb. that the wiki for the device covers different HW that is available w/ different configurations of that notebook.

Check the xorg log for iris references.

Bitals · 2023-05-18 08:24:29

seth wrote:

That's not weird.
nb. that the wiki for the device covers different HW that is available w/ different configurations of that notebook.

I figured out that two PCI devices are the Iris Max jam. It shows XE and Max separately, and for some time (didn't dig about now because it's irrelevant for my laptop, but probably not anymore) it needed 2 different kernels to use them, so only one at a time.
So yeah, mobile Iris XE without the Max part is only one PCI device.

seth wrote:

Check the xorg log for iris references.

I do not use X on my laptop, but I stumbled upon Iris mention on the Firefox' about:support page while making sure my hw acceleration still worked.

GPU #1
Active	Yes
Description	Mesa Intel(R) Graphics (ADL GT2)
Vendor ID	0x8086
Device ID	0x46a6
Driver Vendor	mesa/iris
Driver Version	23.0.3.0
RAM	0

Arch Linux

#1 2023-05-16 20:24:41

[SOLVED] Intel UHD used instead of Iris XE

#2 2023-05-16 22:19:51

Re: [SOLVED] Intel UHD used instead of Iris XE

#3 2023-05-16 23:07:18

Re: [SOLVED] Intel UHD used instead of Iris XE

#4 2023-05-16 23:16:13

Re: [SOLVED] Intel UHD used instead of Iris XE

#5 2023-05-17 00:34:14

Re: [SOLVED] Intel UHD used instead of Iris XE

#6 2023-05-17 11:16:16

Re: [SOLVED] Intel UHD used instead of Iris XE

#7 2023-05-17 11:30:35

Re: [SOLVED] Intel UHD used instead of Iris XE

#8 2023-05-17 13:16:17

Re: [SOLVED] Intel UHD used instead of Iris XE

#9 2023-05-17 16:32:50

Re: [SOLVED] Intel UHD used instead of Iris XE

#10 2023-05-17 16:38:05

Re: [SOLVED] Intel UHD used instead of Iris XE

#11 2023-05-18 08:24:29

Re: [SOLVED] Intel UHD used instead of Iris XE

Board footer