Dear Maintainer,

after upgrading my setup from trixie to forky, the system stopped
coming up. Just switching boot.scr to load the old kernel
(6.12.43+deb13) was enough to get the systems up again. (This report is
created with 6.12.43+deb13 running.)

After attaching a serial console output was captured - see below.
Note: I did the investigation with 7.1.8+deb14.1. I then upgraded the
kernel to 7.1.12+deb14 and can confirm that 7.1.12+deb14 is affected
too. (However, I didn't get the serial output of 7.1.12+deb14, because
the setup with the serial console is not at the same location.)

A few noteworthy facts about serial output:
* The output just stops without any error message.
* The output stopped at slightly different times of the boot process
every time I tried.
* The hardware itself mostly seems up. LEDs are on, USB is powered etc.
* If I type something into the serial console, I actually get the echo
back.

I don't know, what to make of this combination of symptoms or how to
investigate further.

Attached you find:
boot_fail*.log - the serial output of two attempts with a broken
kernel.
boot_okay.log - the output when booting the last known to work version.

Thanks for your work maintaining the kernel package!
Harald


As noted above this information pertains to the working version for
obvious reasons:

Hallo Harald,

I didn't compare your two logs, but I suspect that the difference is
just the order of the messages because drivers are loaded asynchronously
and thus the timing is slightly different each time.

A few ideas for debugging:

 - How long did you wait for further output once it stopped?

 - please add

 	sysrq_always_enabled

   to the kernel commandline and when the hung state is reached send a
   break on the serial console and then a `t`. Ideally that provides a
   bunch of output.

 - Does it help if you add

 	root=/dev/mmcblk0p2

   to the kernel command line? (Is mmcblk0p2 the right device?)
   Comparing the boot logs I see that 7.1 has considerably less messages
   for ata (i.e. hard disks).

 - Can you provide the boot log of a boot with

 	debug

   on the commandline?

 - Can you please provide the 7.1 initrd?

If anything here is unclear, don't hesitate to ask. Either here, or if
you prefer in #debian-arm (highlight me (= ukleinek)).

Best regards
Uwe

Hallo Uwe,

danke für die ausführliche Hilfe.

Am Sonntag, dem 13.09.2026 um 20:10 +0200 schrieb Uwe Kleine-König:
With serial console: Once at least 15 minutes.

It did, but I can't interpret this. See attached log.

No significant changes.

(And yes: mmcblk0p2 is the right device. I used to have a SATA disk at
/data, back when the board was new and useful for heavy stuff. The disk
is long gone now, but I guess some of the modules stayed in the old
initrd.)

See again the attached log.


Thanks for your detailed instructions and best regards,
Harald