FS#80207 - [linux] 6.6.x i915 driver crash
Attached to Project:
Arch Linux
Opened by D (Nebulosa) - Thursday, 09 November 2023, 11:21 GMT
Last edited by Buggy McBugFace (bugbot) - Saturday, 25 November 2023, 20:13 GMT
Opened by D (Nebulosa) - Thursday, 09 November 2023, 11:21 GMT
Last edited by Buggy McBugFace (bugbot) - Saturday, 25 November 2023, 20:13 GMT
|
Details
Description:
After upgrading to v6.6.1 from v6.5.9 (also tried downgrading to v6.6) - KDE starts, but audio not detects (after a couple minutes it detects finally), switching to other tty (CTRL+ALT+F1, ..F3, F4 etc) not working. Also, I can't reboot or even poweroff my PC - it just continue working with a blank screen. In GRUB shell or on linux kernel v6.5.9 rebooting and poweroff is working as expected. Steps to reproduce: Upgrade to latest linux kernel and reboot. |
This task depends upon
Closed by Buggy McBugFace (bugbot)
Saturday, 25 November 2023, 20:13 GMT
Reason for closing: Moved
Additional comments about closing: https://gitlab.archlinux.org/archlinux/p ackaging/packages/linux/issues/5
Saturday, 25 November 2023, 20:13 GMT
Reason for closing: Moved
Additional comments about closing: https://gitlab.archlinux.org/archlinux/p ackaging/packages/linux/issues/5
[1]: https://wiki.archlinux.org/title/Kernel#Debugging_regressions
[1] https://gitlab.freedesktop.org/drm/intel
ACPI: video: [Firmware Bug]: Duplicate ACPI video bus devices for the same VGA controller, please try module parameter "video.allow_duplicates=1" if the current driver doesn't work.
Hardware name: Sony Corporation VPCSB2X9R/VAIO, BIOS R2087H4 06/15/2012
RIP: 0010:video_get_max_state+0x17/0x30 [video]
Code: Unable to access opcode bytes at 0xffffffffc04a5ffd.
Call Trace:
<TASK>
? __die+0x23/0x70
? page_fault_oops+0x171/0x4e0
? exc_page_fault+0x7f/0x180
? asm_exc_page_fault+0x26/0x30
? video_get_max_state+0x17/0x30 [video 7b201efe94c3bedceddd3a2b0b46fc87b7fd7ce0]
__thermal_cooling_device_register.part.0+0xf2/0x2f0
acpi_video_bus_register_backlight.part.0.isra.0+0x414/0x570 [video 7b201efe94c3bedceddd3a2b0b46fc87b7fd7ce0]
acpi_video_register_backlight+0x57/0x80 [video 7b201efe94c3bedceddd3a2b0b46fc87b7fd7ce0]
intel_acpi_video_register+0x68/0x90 [i915 91011d57f4e61d512e7a501bb95f35e3223bce0f]
intel_display_driver_register+0x28/0x50 [i915 91011d57f4e61d512e7a501bb95f35e3223bce0f]
i915_driver_probe+0x790/0xb90 [i915 91011d57f4e61d512e7a501bb95f35e3223bce0f]
[1]: https://gitlab.freedesktop.org/drm/intel/uploads/fcccbfe2833d6f4679dff3608c735ee2/0001-drm-i915-Also-check-for-VGA-converter-in-eDP-probe.patch
[2]: https://gitlab.freedesktop.org/drm/intel/-/issues/9636
Downgraded kernel again on 6.5.9.
https://drive.google.com/file/d/1cSs8kTxkHY1qEzqM99ChvFADurTUnU3d/view?usp=sharing linux-6.6.1.arch1-1.2-x86_64.pkg.tar.zst
https://drive.google.com/file/d/1mf8lQvhIYLNpzP0zOVT3uH4U4gpO_AQW/view?usp=sharing linux-headers-6.6.1.arch1-1.2-x86_64.pkg.tar.zst
[1] https://gitlab.freedesktop.org/drm/intel/uploads/fcccbfe2833d6f4679dff3608c735ee2/0001-drm-i915-Also-check-for-VGA-converter-in-eDP-probe.patch
So, I took PKGBUILD from: https://archlinux.org/packages/core/x86_64/linux/, added patch 0001-drm-i915-Also-check-for-VGA-converter-in-eDP-probe.patch, made sure that the changes were applied and installed the linux, linux-headers packages (version 6.6.1-arch1-2 in my case). Regression is still here. Dmesg: https://0x0.st/Hvjp.txt
Also, I tried to install your packages '6.6.1.arch1-1.2' but nothing changed. Dmesg: https://0x0.st/Hvjx.txt
6.5.9 is still the latest kernel that worked.
[1] https://gitlab.freedesktop.org/drm/intel/-/issues/9636
Meanwhile, I read wiki, forum, etc., installed ccache and modprobed-db, removed docs building from PKGBUILD and started procedure of bisecting.
If you are checking out the commits and using makepkg -e:
git checkout 49d4648b65d03752904ac945aefa60044329a9a3 # test if the commit is bad
git show 49d4648b65d03752904ac945aefa60044329a9a3 # confirm the commit is not a merge
git rev-parse 49d4648b65d03752904ac945aefa60044329a9a3~ # get the commits first parent
git checkout 9856308c94ca821fdc6f3440e4d4de069b09677c # test if parent of the commit above is good
After git checkout, I should build kernel again, right?
$ cd ../..
$ makepkg -efsi
Continue bisection proccess...
[1]: https://cgit.freedesktop.org/drm-intel/commit/?h=for-linux-next&id=fcd479a79120bf0cd507d85f898297a3b868dda6
https://github.com/torvalds/linux/commit/0d16710146a10cf62b3efddee8ffd006432d5d7e - is first bad commit
What should I do next?
Assuming you can no longer produce the issue with the revert you need to decide if you want to report the issue against the i915 driver [1] or acpi [2] possibly on the https://bugzilla.kernel.org product ACPI component ACPICA-Core and add to the CC list Michal Wilczynski <michal.wilczynski@intel.com> plus Rafael J. Wysocki <rafael.j.wysocki@intel.com> or the linux-acpi mailing list possibly by replying to https://lore.kernel.org/linux-acpi/20230703080252.2899090-3-michal.wilczynski%40intel.com/
$ perl scripts/get_maintainer.pl drivers/acpi/bus.c
"Rafael J. Wysocki" <rafael@kernel.org> (supporter:ACPI)
Len Brown <lenb@kernel.org> (reviewer:ACPI)
linux-acpi@vger.kernel.org (open list:ACPI)
linux-kernel@vger.kernel.org (open list)
[1]: https://drm.pages.freedesktop.org/intel-docs/how-to-file-i915-bugs.html
[2]: https://docs.kernel.org/admin-guide/reporting-issues.html
Trivial patch in attach.
I think it's i915-specific bug, on my other instances (with Nvidia card or with other several vpses with virtual cards) there is no regression at all, so right decision will be reporting issue in gitlab.freedesktop org at first.