i915 0000:00:02.0: Device initialization failed (-16) with 82852/855GM Integrated Graphics Device

This is on a (pretty ancient, sure) Panasonic Thinkpad CF-51, running openSUSE Tumbleweed LegacyX86 (it's an 32bit Pentium M Dothan processor machine) The last known working kernel was openSUSE's kernel-pae-6.7.7 It broke for me with the next tested kernel-pae-6.9.5 (not tests in between, the machine is not used regularly). It is still broken in kernel-pae-6.10.rc4 Due to the limits of this machine, self-building complete kernels for a bisection would be a major undertaking for me, but if you think it is necessary, I can try that.

The i915 messages in dmesg are

toughbook:~ # grep i915 dmesg-915-broken.txt 
[   23.136721] [  T594] i915 0000:00:02.0: vgaarb: deactivate vga console
[   23.325049] [  T594] genirq: Flags mismatch irq 9. 00000080 (i915) vs. 00002080 (acpi)
[   23.466113] [  T594] i915 0000:00:02.0: Device initialization failed (-16)
[   23.466126] [  T594] i915 0000:00:02.0: Please file a bug on drm/i915; see https://drm.pages.freedesktop.org/intel-docs/how-to-file-i915-bugs.html for details.
[   23.466135] [  T594] i915 0000:00:02.0: probe with driver i915 failed with error -16

I can boot the machine with "nomodeset" and it will come up with xorg running, not sure if a different xorg driver is used then. Console stays at "legacy VGA" (80x25) in this case.

requested info:

toughbook:~ # uname -a
Linux toughbook 6.10.0-rc4-1.g3306b36-pae #1 SMP PREEMPT_DYNAMIC Sun Jun 16 20:59:28 UTC 2024 (3306b36) i686 i686 i386 GNU/Linux
toughbook:~ # lspci -vnn -d :*:0300
00:02.0 VGA compatible controller [0300]: Intel Corporation 82852/855GM Integrated Graphics Device [8086:3582] (rev 02) (prog-if 00 [VGA controller])
        Subsystem: Matsushita Electric Industrial Co., Ltd. Device [10f7:8338]
        Flags: fast devsel, IRQ 9
        Memory at e8000000 (32-bit, prefetchable) [size=128M]
        Memory at e0000000 (32-bit, non-prefetchable) [size=512K]
        I/O ports at 1880 [size=8]
        Expansion ROM at 000c0000 [virtual] [disabled] [size=128K]
        Capabilities: [d0] Power Management version 1
        Kernel modules: i915

dmesg with drm.debug=0xe log_buf_len=4M ignore_loglevel will be attached.

dmesg-i915-drmdebug.txt

Note that I just noticed other pci devices also not working on 6.9 or later like the built in ethernet (r8169) and the cardbus slots, mabe related to the genirq messages?

[   10.982961] [  T303] genirq: Flags mismatch irq 9. 00000080 (firewire_ohci) vs. 00002080 (acpi)
[   11.002834] [  T301] genirq: Flags mismatch irq 9. 00000080 (yenta) vs. 00002080 (acpi)
[   11.181629] [  T312] genirq: Flags mismatch irq 9. 00000080 (uhci_hcd:usb1) vs. 00002080 (acpi)
[   11.308156] [   T47] genirq: Flags mismatch irq 9. 00000080 (ehci_hcd:usb1) vs. 00002080 (acpi)
[   11.324937] [  T312] genirq: Flags mismatch irq 9. 00000080 (uhci_hcd:usb1) vs. 00002080 (acpi)
[   11.325214] [  T312] genirq: Flags mismatch irq 9. 00000080 (uhci_hcd:usb1) vs. 00002080 (acpi)
[   11.455385] [  T301] genirq: Flags mismatch irq 9. 00000080 (yenta) vs. 00002080 (acpi)
[   23.168202] [  T623] genirq: Flags mismatch irq 9. 00000080 (snd_intel8x0m) vs. 00002080 (acpi)
[   24.261543] [  T629] genirq: Flags mismatch irq 9. 00000080 (snd_intel8x0) vs. 00002080 (acpi)
[   27.306585] [  T631] genirq: Flags mismatch irq 9. 00000080 (i915) vs. 00002080 (acpi)
[   30.309313] [  T872] genirq: Flags mismatch irq 9. 00000080 (enp2s1) vs. 00002080 (acpi)

added Community label

Looks like a regression from

commit 655a6e7c0d83d47c36218525708c9fcfdd7f4b43
Author: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Date:   Fri Dec 15 12:26:33 2023 +0100

    ACPI: EC: Use a threaded handler for dedicated IRQ

You should report this to the ACPI folks.

@seife Please reach out to Rafael J. Wysocki`as suggested by Ville.

Thanks, I'll do. I already suspected it's not i915's fault after finding other broken pci drivers. Sorry for the noise and thanks for pointing me into the right direction.

closed