| View previous topic :: View next topic |
| Author |
Message |
shadowless
Tux's lil' helper
Joined: 27 Nov 2022
Posts: 101
|
 Posted: Fri Feb 23, 2024 5:55 pm Post subject: [SOLVED] Older Chromebook HDMI failure |
 |
|
I have an older Acer Chromebook (C710-2688) that I flashed with SeaBIOS (1.7.5-040914-johnlewis.ie) back in 2014. After recently upgrading the kernel (sys-kernel/gentoo-sources) to 6.6.13, the HDMI port no longer works. If plugged in on system boot, I get a black screen. If I plug the HDMI cable in after booting, dmesg shows an Oops. I'm not familiar with the inner workings of BIOS or hardware in general, so I don't know what I should do to try fixing this...
I've also installed sys-kernel/gentoo-kernel-bin-6.6.13, and it behaves exactly the same way. Previous working kernel was sys-kernel/gentoo-sources-6.1.67
dmesg
Last edited by shadowless on Mon Feb 26, 2024 2:54 am; edited 2 times in total |
|
| Back to top |
|
 |
Hu
Administrator
Joined: 06 Mar 2007
Posts: 21794
|
| Posted: Fri Feb 23, 2024 6:44 pm Post subject: |
|
|
An oops is good here. It means there is a kernel bug that can be diagnosed, reported, and fixed. Do I understand correctly that you never tried any of 6.2, 6.3, ... on this system? We know 6.1.67 is good and 6.6.13 is bad. There are 74732 non-merge commits in the suspect range. This makes guessing at the guilty commit difficult.
You could also report it as-is. Since it is reproducible on demand, debug kernels with more logging can be created. We can see that the fault occurs in the i915 driver. With some effort, this could be isolated to a single source statement, and someone familiar with the driver could identify why a null pointer is present here. |
|
| Back to top |
|
|
shadowless
Tux's lil' helper
Joined: 27 Nov 2022
Posts: 101
|
| Posted: Sat Feb 24, 2024 12:44 am Post subject: |
|
|
| Hu wrote: | | Do I understand correctly that you never tried any of 6.2, 6.3, ... on this system? |
Yes.
I don't believe any of those made it into the portage tree. Maybe I should try compiling some of these to test... before reporting the bug upstream. The ebuilds for the gentoo-sources or vanilla-sources don't look too complicated  |
|
| Back to top |
|
|
Hu
Administrator
Joined: 06 Mar 2007
Posts: 21794
|
| Posted: Sat Feb 24, 2024 2:04 am Post subject: |
|
|
| All of those were in the Gentoo Portage tree at various times. They may not have been marked as stable, though. |
|
| Back to top |
|
|
shadowless
Tux's lil' helper
Joined: 27 Nov 2022
Posts: 101
|
| Posted: Sat Feb 24, 2024 4:11 am Post subject: |
|
|
| Hu wrote: | | All of those were in the Gentoo Portage tree at various times. They may not have been marked as stable, though. |
Apologies. I was talking confidently about things I know of which I almost nothing... again. It's a bad habit I'm trying to kick  |
|
| Back to top |
|
|
shadowless
Tux's lil' helper
Joined: 27 Nov 2022
Posts: 101
|
| Posted: Mon Feb 26, 2024 2:53 am Post subject: |
|
|
So after reviewing the dmesg log again... I noticed this line here:
| Code: | | [ 0.033323] ACPI BIOS Warning (bug): 32/64X length mismatch in FADT/Gpe0Block: 128/64 (20230628/tbfadt-564) |
and decided to do some research on the BIOS. I suspected that a bug in that particular SeaBIOS build was causing this issue here. This isn't a daily driver machine, and I never really paid close attention to dmesg on it. It is possible that the bug warning there has been present for years and never caused any real issues until now.
I found a chromebook firmware updater script here https://mrchromebox.tech/#fwscript and gave it a try.
I flashed the device and replaced SeaBIOS with a coreboot UEFI. I carved some disk space out of my swap partition to create a new EFI partition, emerged grub with GRUB_PLATFORMS="efi-64", ran grub-install / grub-mkconfig ...
and everything seems to be working great now. The original BIOS bug warning in dmesg is gone and the HDMI now works on both sys-kernel/gentoo-sources-6.6.13 and sys-kernel/gentoo-kernel-bin-6.6.13 without issues.
I decided not to file a bug report upstream for i915, as I can no longer reproduce the issue (looks like it was likely my fault in the first place). |
|
| Back to top |
|
|
|
Kernel & Hardware
