[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Updated installation images for Debian Ports (2025-03-16)



On Fri, 2025-03-28 at 16:16 +1100, Damien Stewart wrote:
> > Not just Apple OpenFirmware, but also PowerKVM. When you try to install
> > Debian on a ppc64el KVM instance you won't get a console output either.
> > 
> > An engineer from IBM wanted to look into it, but it took him forever to
> > reproduce the problem and it also seemed that he wasn't really interested
> > in fixing this problem.
> > 
> Okay, I can understand for Apple, but I expected a PowerKVM instance to be
> more prevalent. So it seems as if no one needs to install Debian into a new
> KVM right now?

This bug was introduced in Debian Trixie which is not released yet. I assume
that most users won't install their POWER servers with Debian Testing.

> > I assume that we will just have to wait until Trixie has been released
> > and people start reporting this issue when they are switching to Trixie.
> > I have no clue why I'm not being able to convince anyone to fix this problem
> > even though it's clearly a bug.
> > 
> I've noticed changes to standards causing issues, such as the change requiring
> child nodes in device trees to have size properties, which I got caught up in.

OK, but that's a different topic.

> > For what is worth, the patch that Debian's kernel package is currently
> > carrying was even rejected upstream because of the very problem it had
> > with hard-wiring the name in such an incompatible way.
> 
> Would be good if this solved itself by being rejected then. But I wonder if
> it the prom init code needs changing to account for it if the problem persists?
> I was reading some particular code that avoids closing OF handles as it causes
> trouble on some Apple models when Quiescing. Looks like prom_init.c is popular
> for all these fixup patches.

The patch is being reverted in Debian's kernel package now.

Adrian

-- 
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer
`. `'   Physicist
  `-    GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913


Reply to: