On Mon, Dec 8, 2025 at 10:25 AM Stanley J. Johnson
[off-list ref] wrote:

Hello,

Please let me know if you have any information regarding this issue.

The fix[1] which I CC-ed you on is waiting on the PPC maintainers to pick up.

Rob

[1] https://lore.kernel.org/all/20251029174047.1620073-1-robh@kernel.org/ (local)

thanks

-Stan Johnson

On 10/30/25 4:14 PM, Stan Johnson wrote:

quoted

Attached are the dtc output files for a PB Lombard and a PB 3400c. If
you need any other information, please let me know.

Thanks for looking into this.

-Stan Johnson

-----

On 10/29/25 11:00 AM, Stan Johnson wrote:

quoted

On 10/29/25 1:29 AM, Segher Boessenkool wrote:

quoted

On Tue, Oct 28, 2025 at 08:17:27PM -0500, Rob Herring wrote:

quoted

On Tue, Oct 28, 2025 at 7:05 PM Stan Johnson [off-list ref] wrote:

quoted

-------- Forwarded Message --------
Subject: Excluded List for "#size-cells" warning
Date: Tue, 28 Oct 2025 10:00:25 -0600
From: Stan Johnson <redacted>
To: debian-powerpc@lists.debian.org
CC: Finn Thain <fthain@linux-m68k.org>, Christophe Leroy
[off-list ref]

Hello,

On a PowerBook G3 Pismo running the latest Debian SID, dmesg
reports the
warning shown below. I've also seen the warning on PowerBook
Lombard and
Wallstreet systems. I haven't checked PowerBook 3400c or Kanga.

Can you send me a dump of the device tree on these systems:

dtc -O dts /proc/device-tree

Please see the attached compressed files containing dtc output for a
Wallstreet (dtc_wallstreet.txt) and a Pismo (dtc_pismo.txt).

quoted

quoted

We've been fixing up these cases such as in commit 7e67ef889c9a
("powerpc/prom_init: Fixup missing #size-cells on PowerBook6,7")

And of course it is perfectly fine for an actual Open Firmware to *not*
repeat the defaults.  As the documentation (the main IEEE 1275 thing)
says: "A missing “#size-cells” property signifies the default value of
one."  There are many other places in OF geared towards this default
btw, take for example the "reg" word, that silently assumes your node's
#size-cells is 1, and does completely the wrong thing if not.

Flattened device trees are a fine thing, but the gratuitous ways it
differs from OF, are not.


Segher