Hi Héctor,

Héctor Orón Martínez <hector.o...@gmail.com> (2021-01-26):
> Since 1.6.0 is available via backports and this issue only happens in
> stable, is this a common operation that must be fixed in stable?

Unsuspecting users like myself, seeing a segfault while running a simple
operation, might think one has to run a dtc version matching the kernel.
Especially so since there's a scripts/dtc/dtc there (as pointed out by
[1] for example).

 1. https://www.raspberrypi.org/documentation/configuration/device-tree.md

Upon asking on #debian-kernel, it turned out that the kernel version I
was looking at had a slightly earlier version (thus unaffected by the
bug), while dtc/stable was a tad more recent and affected by this bug.

So the whole issue was mostly surprising, and also time-consuming since
I had to cross-build parts of the kernel to get the in-tree dtc
executable.

Asking dtc to show what the current device tree looks like doesn't seem
like a crazy, corner-case-like operation, especially if one is
interesting in trying to build a DT overlay (and make it work). But of
course, you probably know what people do with dtc better than I do…

Having a more recent version in backports is nice but backports isn't
really there to fix bugs in stable…

I'm happy to look into preparing a fix for this bug in stable, and to
talk to the release team about it.


Cheers,
-- 
Cyril Brulebois -- Debian Consultant @ DEBAMAX -- https://debamax.com/

Attachment: signature.asc
Description: PGP signature

Reply via email to