When I start chromium, the display (Xorg with no config file) hangs for
10 or so seconds, the screen goes black for a few seconds, and it
returns.  This hang--blank--return cycle repeats twice more.  Then all
is fine.

Here are relevant lines from the dmessages.  (I've also attached the GPU
error state.)

It's an old laptop (Thinkpad T61), but this problem appeared only after
a dist-upgrade a few days ago (while running kernel 6.12).  Maybe it's
triggered by the upgrade of chromium (went from v140.0 to v141.0)?

$ sudo dmesg | egrep -i 'gpu|i915|drm'
[    3.122844] ata3.00: supports DRM functions and may not be fully accessible
[    3.124414] ata3.00: supports DRM functions and may not be fully accessible
[   11.713176] systemd[1]: Starting modprobe@drm.service - Load Kernel Module drm...
[   12.047508] ACPI: bus type drm_connector registered
[   12.056673] systemd[1]: modprobe@drm.service: Deactivated successfully.
[   12.071499] systemd[1]: Finished modprobe@drm.service - Load Kernel Module drm.
[   18.306504] i915 0000:00:02.0: [drm] Found i965gm (device ID 2a02) integrated display version 4.00 stepping N/A
[   18.389390] i915 0000:00:02.0: vgaarb: deactivate vga console
[   18.799500] i915 0000:00:02.0: vgaarb: VGA decodes changed: olddecodes=io+mem,decodes=io+mem:owns=io+mem
[   18.835046] i915 0000:00:02.0: [drm] Initialized overlay support.
[   18.847486] i915 0000:00:02.0: [drm] Skipping intel_backlight registration
[   18.847597] [drm] Initialized i915 1.6.0 for 0000:00:02.0 on minor 0
[   18.931838] fbcon: i915drmfb (fb0) is primary device
[   19.455586] i915 0000:00:02.0: [drm] fb0: i915drmfb frame buffer device
[  187.069814] i915 0000:00:02.0: [drm] GPU HANG: ecode 4:1:9f47f9fd, in chromium [4847]
[  187.069827] i915 0000:00:02.0: [drm] GPU error state saved to /sys/class/drm/card0/error
[  197.607423] i915 0000:00:02.0: [drm] GT0: Resetting chip for stopped heartbeat on rcs0
[  197.709968] i915 0000:00:02.0: [drm] chromium[4847] context reset due to GPU hang
[  205.245883] i915 0000:00:02.0: [drm] GPU HANG: ecode 4:1:ab95f0fb, in chromium [5012]
[  215.767378] i915 0000:00:02.0: [drm] GT0: Resetting chip for stopped heartbeat on rcs0
[  215.869954] i915 0000:00:02.0: [drm] chromium[5012] context reset due to GPU hang
[  225.213762] i915 0000:00:02.0: [drm] GPU HANG: ecode 4:1:9f47f9fd, in chromium [5059]
[  235.743344] i915 0000:00:02.0: [drm] GT0: Resetting chip for stopped heartbeat on rcs0
[  235.845956] i915 0000:00:02.0: [drm] chromium[5059] context reset due to GPU hang

Hi

So do I understand correctly that the kernel was no changed while
noticing this behaviour change? In this case please do roll back as
well first the chromium version.

If not please do downgrade to the preivous kernel you were running and
see if you can trigger the problem. If you only by chaning the kernel
then are able to trigger the problem in one version and not in the
other this might be a good start for a bisect of upstream versions.

Let's clarify first the above.

Regards,
Salvatore

Right, the kernel stayed at 6.12, but chromium was upgraded.  That
upgrade was enough to trigger the problem.

When I rebooted, it rebooted into the upgraded kernel (6.16), and it had
the same GPU hangs.  So, I will try earlier chromiums (with kernel 6.16
at first).  I think that it was chromium 140, but I will try a few
versions.

I've just tried chromium v140 (with a current kernel -- see below), and
it worked fine (no GPU hang).

$ dpkg -l 'chromium*' | grep '^ii'
ii  chromium                 140.0.7339.185-1 amd64        web browser
ii  chromium-common          140.0.7339.185-1 amd64        web browser - common resources used by the chromium packages
ii  chromium-sandbox         140.0.7339.185-1 amd64        web browser - setuid security sandbox for chromium

$ uname -a
Linux insight 6.16.9+deb14-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.16.9-1 (2025-09-27) x86_64 GNU/Linux

Control: tags -1 - moreinfo

Thanks for reporting back, so likely this is actually a an older i915
related bug (there are similar older hanging GPU reports), but just
now triggered on at least your system.

I'm removing the moreinfo tag, and we now have to thinker more on how
to proceed.

Regards,
Salvatore

At least it’s reproducible here, so it might be solvable. I can keep those kernel and chromium packages around for further tests.

The same GPU hang just happened with the older chromium
(140.0.7339.185-1 amd64 debian package).  It happened just after waking
up from suspend (S3). Chromium was already running before the suspend.

Hi

Not sure we will be sucessful, but it might give us additional hints
if you can provide the full dmesg from boot, to suspend and wakeup
again where this has happened.

Can you provide the log?

Regards,
Salvatore

Sure, here it is (attached).  These GPU hangs start at 23544.253696
seconds (after boot).  For these hangs, no GPU error state was saved.
Maybe it is because an error state was saved earlier in the boot (from
the GPU hangs at 187.069827 seconds after boot).