From Newsgroup: comp.os.linux.misc


I did some research. There are known issues that prevent this working
out of the box..

They seem to be due to Trixie having dropped some libraries that the
current vbox needs and a need to remove the kvm nodule,

Whose 'fault' this is is a worm can I will refrain from opening...
--
Canada is all right really, though not for the whole weekend.

"Saki"
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

The Natural Philosopher <tnp@invalid.invalid> writes:

I did some research. There are known issues that prevent this working
out of the box..

They seem to be due to Trixie having dropped some libraries that the
current vbox needs and a need to remove the kvm nodule,

Whose 'fault' this is is a worm can I will refrain from opening...

I followed the instructions for Debian-based distributions at https://www.virtualbox.org/wiki/Linux_Downloads.

* There were no dependency issues at least for VirtualBox 7.1, using the
virtualbox.org archive. If you installed the .deb manually you might
have to install some dependencies manually too, but thatrCOs just
question of whether you want to make your life easy or hard.

* Because IrCOm using Secure Boot, I had to create and install a driver
signing key before it could install the VirtualBox kernel modules. The
diagnostic from the package install included full instructions and it
worked first time. An end user who is not using Secure Boot would
presumably not need this step.

* I had to manually remove the kvm and kvm_internal kernel modules.
ItrCOs a one-liner to remove them from the running kernel, I didnrCOt
persist the change because IrCOll want to re-enable kvm at the end of
this exercise.

ItrCOs installing Ubuntu 24.04 into a test VM as I write.


The KVM incompatibility is evidently real, although it doesnrCOt seem particularly surprising that running two hypervisors in one kernel isnrCOt
a working configuration. Apart from that minor issue, the idea that
VirtualBox has insurmountable problems with Debian trixie, or the other
way around, doesnrCOt seem to hold water.
--
https://www.greenend.org.uk/rjk/
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

On Wed, 25 Feb 2026 12:13:31 +0000, The Natural Philosopher wrote:

I did some research. There are known issues that prevent this working
out of the box..

They seem to be due to Trixie having dropped some libraries that the
current vbox needs and a need to remove the kvm nodule,

Whose 'fault' this is is a worm can I will refrain from opening...

https://forums.virtualbox.org/viewtopic.php?t=113902

Yeah, it does seems to be a deeper problem than the unloading the kvm
module that worked for people on other distros.
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

On 2/25/26 07:13, The Natural Philosopher wrote:

I did some research. There are known issues that prevent this working
out of the box..

They seem to be due to Trixie having dropped some libraries that the
current vbox needs and a need to remove the kvm nodule,

Whose 'fault' this is is a worm can I will refrain from opening...

Deb used to be fairly 'stable' - what would
work 20 years ago would still work fine, you
didn't have to go chasing this months suite
of tools to get things done.

Now ... starting with BullsEye ... that solid
rock rep seems to be crumbling. This is made
even more critical because so many other distros
use Deb as their base.

I will admit that while BullsEye gave me fits when
it first came out it DID get its act together pretty
well later on. Trixie, like Bullseye, may have just
been born a bit premature.

We shall see.

There's clearly some pressure for orgs to pop the
newest version out the door. However this leads
to issues that can run down the rep of your distro,
counter-productive.

Mint is based on a LTS version of Ubuntu, which is
apparently based on a mature Bullseye. VBox runs
just fine on it. Now I just have to figure out why
the Manjaro VM won't mount my NFS shares even though
it can see the machine. Maybe I didn't load some
library ... gotta fool around with it. Next up
is the GhostBSD VM. In the end there can be only
one but I'm gonna give both a fair shake.

--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

On Wed, 25 Feb 2026 20:03:48 -0500, c186282 wrote:

Now I just have to figure out why
the Manjaro VM won't mount my NFS shares even though it can see the
machine.

Is the NFS server exporting to *?
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

Richard Kettlewell <invalid@invalid.invalid> wrote:

The Natural Philosopher <tnp@invalid.invalid> writes:

I did some research. There are known issues that prevent this working
out of the box..

They seem to be due to Trixie having dropped some libraries that the
current vbox needs

did we _drop_ those libraries or are we just not installing them any
more in the default system?

and a need to remove the kvm nodule,

Is known why the kvm module gets loaded? Sadly, I don't have a
reference system handy and am not willing to spend my private time to
solve "c18"'s problems.

* There were no dependency issues at least for VirtualBox 7.1, using the
virtualbox.org archive. If you installed the .deb manually you might
have to install some dependencies manually too, but thatrCOs just
question of whether you want to make your life easy or hard.

If the Virtualbox 7.1 deb needs libraries, they should declare those dependencies and apt would take care of that automatically. Things are different if Virtualbox needs libs that we are not shipping, then I'd
like to know which libraries those are to be able to say why we are
not shipping them.

* Because IrCOm using Secure Boot, I had to create and install a driver
signing key before it could install the VirtualBox kernel modules. The
diagnostic from the package install included full instructions and it
worked first time. An end user who is not using Secure Boot would
presumably not need this step.

Out-of-tree Kernel Modules do not play well with secure boot for very
obvious reasons. That's one of the reasons why I am happy to having
switched away from Virtualbox a decade ago.

Greetings
Marc
-- ---------------------------------------------------------------------------- Marc Haber | " Questions are the | Mailadresse im Header Rhein-Neckar, DE | Beginning of Wisdom " |
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 6224 1600402
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

On 2/26/26 01:04, rbowman wrote:

On Wed, 25 Feb 2026 20:03:48 -0500, c186282 wrote:

Now I just have to figure out why
the Manjaro VM won't mount my NFS shares even though it can see the
machine.

Is the NFS server exporting to *?

Nevermind ... installed a couple of nfs-related
pkgs - which brought in some more - then did the
full system update. It now works - I can export
my ffmpeg results to where I want.

The NFS server exports to anyone. Just need to
add a line in fstab.

No, this is a 'home'/hobby system ... not so
worried about 'security' issues. NFS doesn't
have much 'security' anyhow ....... a relic
of a kinder and gentler past :-)

Tomorrow, the actual ffmpeg app ......

DO keep having probs with VBox adding the
'extension pack'. Not the first time. It
kind of starts, gets about 60%, then says
it can't do it. Forget how I solved that
the last time ... will have to study the
issue again.

In any case, VBox makes some aspects of VMs
and their maint much easier than KVM. So, if
possible, I'll stick with VBox for now.
Somewhere I was able to make a CP/M-86 install,
would like to do that again. Turbo Pascal had
a CP/M version early on ... should be fun.

Yea yea ... some here don't feel fulfilled unless
they have to micro-edit a dozen config scripts and
kernel drivers to make something work :-) Getting
too old for that, just want it To Work ...

Remember spending DAYS tweaking 'X' so it'd run
on ancient RedHat .....

OTHER prob, non-Linux, my new ATT 'Air' router
keeps going into flashing yellow mode - esp in
mid-evening when consumer usage is probably
the highest. Hardly ever get more than one bar
on the thing (metal roof). Did a hard reset,
had to re-do all my passwords and the subnet
stuff. Didn't help here (though the SPEED when
it is connected improved). Competitors all
want bank routing numbers for billing. You get
many protections in USA with credit-cards, but
none if you give routing numbers. Might have to
create a new small billing acct with Just Enough
money in it ........

Did look into StarLink ... it doesn't offer much
in the way of performance and the BOX is very
expensive. Maybe I'd be paying for free boxes
for Ukraine and Iran ......

Do want to try yer KVM-kernel blacklist fixes ...
may set up Trixie somewhere else and give that
a try, OR just try the full KVM thing. I've got
a fair number of those BMax/BeeLink mini boxes
plus some PIs (and a decent desktop I installed
and haven't used since :-).

Have one BeeLink N95 still in the box with
Winders on it.

--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

On 2/26/26 01:48, Marc Haber wrote:

Richard Kettlewell <invalid@invalid.invalid> wrote:

The Natural Philosopher <tnp@invalid.invalid> writes:

I did some research. There are known issues that prevent this working
out of the box..

They seem to be due to Trixie having dropped some libraries that the
current vbox needs

did we _drop_ those libraries or are we just not installing them any
more in the default system?

Almost amounts to the same thing ... then we
have to GUESS what libs/apps are needed.

and a need to remove the kvm nodule,

Is known why the kvm module gets loaded? Sadly, I don't have a
reference system handy and am not willing to spend my private time to
solve "c18"'s problems.

The KVM stuff really does seem 'baked into' the kernel
at this point. Disabling it ... tried some modprobe
solutions that didn't work

Likely, by the next point version, Trixie will be much
better. But for now ...

* There were no dependency issues at least for VirtualBox 7.1, using the
virtualbox.org archive. If you installed the .deb manually you might
have to install some dependencies manually too, but thatrCOs just
question of whether you want to make your life easy or hard.

If the Virtualbox 7.1 deb needs libraries, they should declare those dependencies and apt would take care of that automatically. Things are different if Virtualbox needs libs that we are not shipping, then I'd
like to know which libraries those are to be able to say why we are
not shipping them.

I think rbowman knows ... or has made a fair guess.

* Because IrCOm using Secure Boot, I had to create and install a driver
signing key before it could install the VirtualBox kernel modules. The
diagnostic from the package install included full instructions and it
worked first time. An end user who is not using Secure Boot would
presumably not need this step.

Out-of-tree Kernel Modules do not play well with secure boot for very
obvious reasons. That's one of the reasons why I am happy to having
switched away from Virtualbox a decade ago.

Well, I don't need secure boot anymore ... so that's
not the issue.

Anyway, got VBox to run on Mint, and another Linux to
run as a VM and see my NFS shares within the VM. I'm set
for now.

--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

Marc Haber <mh+usenetspam1118@zugschl.us> writes:

Richard Kettlewell <invalid@invalid.invalid> wrote:

The Natural Philosopher <tnp@invalid.invalid> writes:

I did some research. There are known issues that prevent this working
out of the box..

They seem to be due to Trixie having dropped some libraries that the
current vbox needs

did we _drop_ those libraries or are we just not installing them any
more in the default system?

There are no missing libraries, or at least there werenrCOt yesterday. I expect there was an interval between trixie being released abnd
VirtualBox being recompiled for trixie, but whatever happened in the
past, itrCOs not relevant any more.

and a need to remove the kvm nodule,

Is known why the kvm module gets loaded? Sadly, I don't have a
reference system handy and am not willing to spend my private time to
solve "c18"'s problems.

DonrCOt know, donrCOt care.

* There were no dependency issues at least for VirtualBox 7.1, using the
virtualbox.org archive. If you installed the .deb manually you might
have to install some dependencies manually too, but thatrCOs just
question of whether you want to make your life easy or hard.

If the Virtualbox 7.1 deb needs libraries, they should declare those dependencies and apt would take care of that automatically. Things are different if Virtualbox needs libs that we are not shipping, then I'd
like to know which libraries those are to be able to say why we are
not shipping them.

I literally just wrote that there were no dependency issues. It does
declare its dependencies and apt satisfies them in the normal way.

* Because IrCOm using Secure Boot, I had to create and install a driver
signing key before it could install the VirtualBox kernel modules. The
diagnostic from the package install included full instructions and it
worked first time. An end user who is not using Secure Boot would
presumably not need this step.

Out-of-tree Kernel Modules do not play well with secure boot for very
obvious reasons. That's one of the reasons why I am happy to having
switched away from Virtualbox a decade ago.

They play fine with secure boot, provided you properly create and
install a signing key. ItrCOs not difficult, but itrCOs an unfamiliar step
and in general you might have to chase down some documentation (although VirtualBoxrCOs setup bypasses even that).
--
https://www.greenend.org.uk/rjk/
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

On 26/02/2026 06:48, Marc Haber wrote:

Richard Kettlewell <invalid@invalid.invalid> wrote:

The Natural Philosopher <tnp@invalid.invalid> writes:

I did some research. There are known issues that prevent this working
out of the box..

They seem to be due to Trixie having dropped some libraries that the
current vbox needs

did we _drop_ those libraries or are we just not installing them any
more in the default system?

Marc. I simply do not know. What I *think* happened is that the version numbers moved on and VirtualBox hasn't kept up.

And that is the problem with Debian and non-free software. Debian does
not include a 'certified working' copy in its distro.

But Mint has no such scruples.

and a need to remove the kvm nodule,

Is known why the kvm module gets loaded? Sadly, I don't have a
reference system handy and am not willing to spend my private time to
solve "c18"'s problems.

Again, I do not know. Richard is more on top of this than I am.

* There were no dependency issues at least for VirtualBox 7.1, using the
virtualbox.org archive. If you installed the .deb manually you might
have to install some dependencies manually too, but thatrCOs just
question of whether you want to make your life easy or hard.

If the Virtualbox 7.1 deb needs libraries, they should declare those dependencies and apt would take care of that automatically. Things are different if Virtualbox needs libs that we are not shipping, then I'd
like to know which libraries those are to be able to say why we are
not shipping them.

As i said it is because you ship versions "x.3" and current vbox wants "version "x.25" AIUI.

I think one poster referred to linking the two names together.

There is info on the virtualbox forum somewhere.

* Because IrCOm using Secure Boot, I had to create and install a driver
signing key before it could install the VirtualBox kernel modules. The
diagnostic from the package install included full instructions and it
worked first time. An end user who is not using Secure Boot would
presumably not need this step.

Out-of-tree Kernel Modules do not play well with secure boot for very
obvious reasons. That's one of the reasons why I am happy to having
switched away from Virtualbox a decade ago.

Greetings
Marc

--
Future generations will wonder in bemused amazement that the early twenty-first centuryrCOs developed world went into hysterical panic over a globally average temperature increase of a few tenths of a degree, and,
on the basis of gross exaggerations of highly uncertain computer
projections combined into implausible chains of inference, proceeded to contemplate a rollback of the industrial age.

Richard Lindzen

--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

The Natural Philosopher <tnp@invalid.invalid> writes:

And that is the problem with Debian and non-free software. Debian does
not include a 'certified working' copy in its distro.

Except for unstable (Sid). And looks like there's a fast track repo now
by Debian too where a VirtualBox package can also be installed from.

BTW, the Virtualbox situation in Debian isn't about freedom, it's about Oracle's unwillingness to provide security updates for Virtualbox and
Debian's insistence on same. As Debian Stable doesn't accept new
versions, Virtualbox can't be in it and so we end up with these
workarounds.
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

On 26/02/2026 12:32, Anssi Saari wrote:

The Natural Philosopher <tnp@invalid.invalid> writes:

And that is the problem with Debian and non-free software. Debian does
not include a 'certified working' copy in its distro.

Except for unstable (Sid). And looks like there's a fast track repo now
by Debian too where a VirtualBox package can also be installed from.

Ok, that is worth knowing

BTW, the Virtualbox situation in Debian isn't about freedom, it's about Oracle's unwillingness to provide security updates for Virtualbox and Debian's insistence on same. As Debian Stable doesn't accept new
versions, Virtualbox can't be in it and so we end up with these
workarounds.

Right, thanks for clarification.

But in essence the problem remains incompatibility between Oracle and
Debian's policy. Both of which are understandable.
--
WOKE is an acronym... Without Originality, Knowledge or Education.

--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

Richard Kettlewell <invalid@invalid.invalid> wrote:

Marc Haber <mh+usenetspam1118@zugschl.us> writes:

Richard Kettlewell <invalid@invalid.invalid> wrote:

The Natural Philosopher <tnp@invalid.invalid> writes:

I did some research. There are known issues that prevent this working
out of the box..

They seem to be due to Trixie having dropped some libraries that the
current vbox needs

did we _drop_ those libraries or are we just not installing them any
more in the default system?

There are no missing libraries, or at least there werenrCOt yesterday. I >expect there was an interval between trixie being released abnd
VirtualBox being recompiled for trixie,

There are official contrib packages at least for the Virtualbox
usersapce, so I'd expect them to be available right away.

but whatever happened in the
past, itrCOs not relevant any more.

trixie is our current stable version, so if there is something to fix,
I'll happily passt that on. Needs more report quality than what that
c18 anonymous is willing to give, though.

If the Virtualbox 7.1 deb needs libraries, they should declare those
dependencies and apt would take care of that automatically. Things are
different if Virtualbox needs libs that we are not shipping, then I'd
like to know which libraries those are to be able to say why we are
not shipping them.

I literally just wrote that there were no dependency issues. It does
declare its dependencies and apt satisfies them in the normal way.

Good, then I misunderstood.

* Because IrCOm using Secure Boot, I had to create and install a driver
signing key before it could install the VirtualBox kernel modules. The
diagnostic from the package install included full instructions and it
worked first time. An end user who is not using Secure Boot would
presumably not need this step.

Out-of-tree Kernel Modules do not play well with secure boot for very
obvious reasons. That's one of the reasons why I am happy to having
switched away from Virtualbox a decade ago.

They play fine with secure boot, provided you properly create and
install a signing key. ItrCOs not difficult, but itrCOs an unfamiliar step >and in general you might have to chase down some documentation (although >VirtualBoxrCOs setup bypasses even that).

I'd love that user experience to be a bit smoother.

Greetings
Marc
-- ---------------------------------------------------------------------------- Marc Haber | " Questions are the | Mailadresse im Header Rhein-Neckar, DE | Beginning of Wisdom " |
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 6224 1600402
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

On Thu, 26 Feb 2026 02:20:26 -0500, c186282 wrote:

Do want to try yer KVM-kernel blacklist fixes ...
may set up Trixie somewhere else and give that a try, OR just try the
full KVM thing. I've got a fair number of those BMax/BeeLink mini
boxes plus some PIs (and a decent desktop I installed and haven't
used since .

https://linuxhint.com/kvm_virtualization_raspberry_pi4/

I'm running the Raspberry Pi OS based on Trixie on the RPi 5.

'sudo apt -install virt-manager' pulls down all the necessary packages.
'sudo usermod -a -G libvirt $(whoami)'

Virtual Machine Manager shows up under Systems Tools.

It looks the same as Mint. I may try installing the Ubuntu ARM iso.

The references are old but I think VMWare had something for the Pi but I
don't know if it survived Broadcom.


--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

On Thu, 26 Feb 2026 07:48:06 +0100, Marc Haber wrote:

Is known why the kvm module gets loaded? Sadly, I don't have a reference system handy and am not willing to spend my private time to solve
"c18"'s problems.

Looking at my various boxes the kvm module is loaded by default. Try
'lsmod | grep kvm'
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

On Thu, 26 Feb 2026 11:40:15 +0000, The Natural Philosopher wrote:

And that is the problem with Debian and non-free software. Debian does
not include a 'certified working' copy in its distro.

Debian doesn't have a problem with non-free software. In fact that's why
they have had issues with the FSF.

https://itsfoss.com/news/fsf-does-not-consider-debian-a-free-distribution/

Then there is Stallman.

https://www.debian.org/vote/2021/vote_002

Lot of history there with Perens' and Raymond's rift with Stallman and Persens' and Raymond's subsequent feud. Raymond likes guns; Perens
doesn't.

--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

On 26 Feb 2026 21:41:17 GMT, rbowman wrote:

On Thu, 26 Feb 2026 02:20:26 -0500, c186282 wrote:

Do want to try yer KVM-kernel blacklist fixes ...
may set up Trixie somewhere else and give that a try, OR just try
the full KVM thing. I've got a fair number of those BMax/BeeLink
mini boxes plus some PIs (and a decent desktop I installed and
haven't used since .

https://linuxhint.com/kvm_virtualization_raspberry_pi4/

I'm running the Raspberry Pi OS based on Trixie on the RPi 5.

'sudo apt -install virt-manager' pulls down all the necessary packages. 'sudo usermod -a -G libvirt $(whoami)'

Virtual Machine Manager shows up under Systems Tools.

It looks the same as Mint. I may try installing the Ubuntu ARM iso.

The references are old but I think VMWare had something for the Pi but I don't know if it survived Broadcom.

The Ubuntu image seemed to install in a VM but fails to start with some
UEFI error.
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

On 2/26/26 08:03, The Natural Philosopher wrote:

On 26/02/2026 12:32, Anssi Saari wrote:

The Natural Philosopher <tnp@invalid.invalid> writes:

And that is the problem with Debian and non-free software. Debian does
not include a 'certified working' copy in its distro.

Except for unstable (Sid). And looks like there's a fast track repo now
by Debian too where a VirtualBox package can also be installed from.

Ok, that is worth knowing

Note : on the Oracle site there's a clearly
labeled Trixie version.

But it didn't work ... at least not 10 days ago.

As said, Trixie is very young, so it may be a little
while before everything is brought over and tuned.
I'll use VBox with Manjaro for now, check back
with Trixie in a few months.

BTW, the Virtualbox situation in Debian isn't about freedom, it's about
Oracle's unwillingness to provide security updates for Virtualbox and
Debian's insistence on same. As Debian Stable doesn't accept new
versions, Virtualbox can't be in it and so we end up with these
workarounds.

Right, thanks for clarification.

But in essence the problem remains incompatibility between Oracle and Debian's policy. Both of which are understandable.

It is 'policy' ... or just not getting everything tweaked-in
yet for a new distro ?

Lots of people use VBox, lots of people use KVM. The 'policy'
is probably to try and make EVERYBODY happy - they're just
not there yet.

--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

c186282 <c186282@nnada.net> writes:

On 2/26/26 08:03, The Natural Philosopher wrote:

On 26/02/2026 12:32, Anssi Saari wrote:

The Natural Philosopher <tnp@invalid.invalid> writes:

And that is the problem with Debian and non-free software. Debian
does not include a 'certified working' copy in its distro.

Except for unstable (Sid). And looks like there's a fast track repo
now by Debian too where a VirtualBox package can also be installed
from.

Ok, that is worth knowing

Note : on the Oracle site there's a clearly
labeled Trixie version.

But it didn't work ... at least not 10 days ago.

It works fine, see my other post. Probably a PEBKAC situation.
--
https://www.greenend.org.uk/rjk/
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

rbowman <bowman@montana.com> wrote:

On Thu, 26 Feb 2026 07:48:06 +0100, Marc Haber wrote:

Is known why the kvm module gets loaded? Sadly, I don't have a reference
system handy and am not willing to spend my private time to solve
"c18"'s problems.

Looking at my various boxes the kvm module is loaded by default. Try
'lsmod | grep kvm'

Thank you for giving me the opportunity to read up on this. It
obviously the case since many years that kernel modules declare what hardware/CPU feature etc the support. Here:
alias: cpu:type:x86,ven*fam*mod*:feature:*0085*

systemd-udevd in later version then goes ahead and loads all modules
that are available and document to make a CPU feature that is present
in the current CPU available.

Learned: It is in fact the case that the kvm module gets autoloaded on reasonably modern Linux systems without having software that actually
makes use of KVM installed. That looks to me as a function of rather
low system-level software, so it is probably the best solution to
blacklist kvm, kvm_intel and kvm_amd if you want to use Virtualbox.

For Debian, I can say that this behavior is present at least from
Debian 12 ("bookworm") upwards and such also is there in Debian 13
("trixie") and most probably also in our future releases. That is
unlikely to change. There is no bug in the Debian BTS complainig about
that behavior (I looked in the source package systemd, which also
delivers udev, for the string "kvm" and didn't find anything).

Debian 10 ("buster") doesn't autoload the kvm modules, and I don't
have any Debian 11 ("bullseye") systems left in my fleet.

I was too lazy to see whether other distributions are doing it
differently and how they have configured their way. Maybe others can
chime in here. If Debian is indeed the exception in regarding allowing systemd-udev to autoload the kvm modules, I wold be willing to file a
bug. But, Captain, I need facts for that.

Greetings
Marc
-- ---------------------------------------------------------------------------- Marc Haber | " Questions are the | Mailadresse im Header Rhein-Neckar, DE | Beginning of Wisdom " |
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 6224 1600402
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

rbowman <bowman@montana.com> wrote:

'sudo apt -install virt-manager' pulls down all the necessary packages. >'sudo usermod -a -G libvirt $(whoami)'

Virtual Machine Manager shows up under Systems Tools.

I THINK that you also need to install the correct qemu-system package
(I guess it would be qemu-system-arm64) to actually be able to start
VMs on the Raspi. Probably more (see recommend and suggests lists of
the packages in question). Virt-Manager can also be used to control a
remote hypervisor.

Greetings
Marc
-- ---------------------------------------------------------------------------- Marc Haber | " Questions are the | Mailadresse im Header Rhein-Neckar, DE | Beginning of Wisdom " |
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 6224 1600402
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

rbowman <bowman@montana.com> wrote:

The Ubuntu image seemed to install in a VM but fails to start with some
UEFI error.

If you're talking about the Ubuntu Image for the Raspberry Pi, I am
not surprise that that doesn't start on a virtualized arm64 system
with UEFI. The Raspberry Pi's boot mechanics are more bizarre than how
Windows 95 booted.

nThere is a UEFI firmware for the Raspi available, but a lot of
hardware including graphics doesn't work when the Raspi is started via
UEFI. That's a real pity, but that's how things are.

Greetings
Marc
-- ---------------------------------------------------------------------------- Marc Haber | " Questions are the | Mailadresse im Header Rhein-Neckar, DE | Beginning of Wisdom " |
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 6224 1600402
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

Marc Haber <mh+usenetspam1118@zugschl.us> writes:

For Debian, I can say that this behavior is present at least from
Debian 12 ("bookworm") upwards and such also is there in Debian 13
("trixie") and most probably also in our future releases. That is
unlikely to change. There is no bug in the Debian BTS complainig about
that behavior (I looked in the source package systemd, which also
delivers udev, for the string "kvm" and didn't find anything).

I also found a Reddit thread from 2020 (https://www.reddit.com/r/linuxadmin/comments/e6ox9t/on_most_linux_systems_why_is_the_kvm_kernel/)
so it goes back to that far, the OP had Ubuntu 19.10. There's a link to
a kernel commit too in there which explains things a
little. https://git.kernel.org/linus/644e9cbbe3fc032cc92d0936057e166a994dc246

OTOH, there was some web page in nz where someone explained the same
issue: VBox worked in Bookworm but not in Trixie and he had to
blacklist. So, it seems there might be something there. I dunno, all my Bookworm and Trixie systems load kvm and kvm_amd or kvm_intel by default.

Could it be there was some postinstall script in Bookworm times which
handled blacklisting and unloaded the kernel modules so Virtualbox was
ready to go?
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

On Fri, 27 Feb 2026 11:03:17 +0100, Marc Haber wrote:

rbowman <bowman@montana.com> wrote:

'sudo apt -install virt-manager' pulls down all the necessary packages. >>'sudo usermod -a -G libvirt $(whoami)'

Virtual Machine Manager shows up under Systems Tools.

I THINK that you also need to install the correct qemu-system package (I guess it would be qemu-system-arm64) to actually be able to start VMs on
the Raspi. Probably more (see recommend and suggests lists of the
packages in question). Virt-Manager can also be used to control a remote hypervisor.

The virt-manager install does install many qemu packages. What I've been trying to run are images, not isos. I'm going to try the Ubuntu Server
iso.
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

On Fri, 27 Feb 2026 11:01:20 +0100, Marc Haber wrote:

I was too lazy to see whether other distributions are doing it
differently and how they have configured their way. Maybe others can
chime in here. If Debian is indeed the exception in regarding allowing systemd-udev to autoload the kvm modules, I wold be willing to file a
bug. But, Captain, I need facts for that.

I've been playing around so it's not a definitive answer but my Ubuntu,
Arch, Fedora, and Mint boxes all show kvm loaded. I do have a openSUSE
Leap 16 instance in a VM on the Fedora box that I haven't messed with and
it also shows kvm loaded.

lsmod on the Pi does not show it although 'modinfo kvm' shows it as a
builtin. That might be a Pi kernel tweak.
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

Anssi Saari <anssi.saari@usenet.mail.kapsi.fi> wrote:

OTOH, there was some web page in nz where someone explained the same
issue: VBox worked in Bookworm but not in Trixie and he had to
blacklist.

I consider having to blacklist a kernel module a fully valid, and
trivial, and well-known workaround. I am not following the "Virtualbox
does not work".

Could it be there was some postinstall script in Bookworm times which
handled blacklisting and unloaded the kernel modules so Virtualbox was
ready to go?

If I would make Virtualbox packages for Debian, I'd probably ship the
blacklist file with my package, yes, as a dpkg-conffile, and advise my
users to reboot before starting Virtualbox for the first time.

Greetings
Marc
-- ---------------------------------------------------------------------------- Marc Haber | " Questions are the | Mailadresse im Header Rhein-Neckar, DE | Beginning of Wisdom " |
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 6224 1600402
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

rbowman <bowman@montana.com> wrote:

lsmod on the Pi does not show it although 'modinfo kvm' shows it as a >builtin. That might be a Pi kernel tweak.

grep KVM /boot/config-$KERNELVERSION or somewhere in /proc, I think
it's /proc/sys/kernel/config or something similar.

Greetings
Marc
-- ---------------------------------------------------------------------------- Marc Haber | " Questions are the | Mailadresse im Header Rhein-Neckar, DE | Beginning of Wisdom " |
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 6224 1600402
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

rbowman <bowman@montana.com> wrote:

On Fri, 27 Feb 2026 11:03:17 +0100, Marc Haber wrote:

rbowman <bowman@montana.com> wrote:

'sudo apt -install virt-manager' pulls down all the necessary packages. >>>'sudo usermod -a -G libvirt $(whoami)'

Virtual Machine Manager shows up under Systems Tools.

I THINK that you also need to install the correct qemu-system package (I
guess it would be qemu-system-arm64) to actually be able to start VMs on
the Raspi. Probably more (see recommend and suggests lists of the
packages in question). Virt-Manager can also be used to control a remote
hypervisor.

The virt-manager install does install many qemu packages. What I've been >trying to run are images, not isos. I'm going to try the Ubuntu Server
iso.

Whether you're booting your system from a hard disk image or an ISO is irrelevant to the fact that you need the right KVM/qemu userspace to
even start your VM. You can run a live system from an ISO without
having a virtual hard disk configure and you can of course boot an
installed system from a virtual hard disk.

Greetings
Marc
-- ---------------------------------------------------------------------------- Marc Haber | " Questions are the | Mailadresse im Header Rhein-Neckar, DE | Beginning of Wisdom " |
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 6224 1600402
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

On Fri, 27 Feb 2026 22:19:44 +0100, Marc Haber wrote:

grep KVM /boot/config-$KERNELVERSION or somewhere in /proc, I think
it's /proc/sys/kernel/config or something similar.

As I recall, the ability to interrogate the kernel itself to discover
its build configuration settings is not enabled on DebianrCOs kernel
builds.
--- Synchronet 3.21b-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

On Fri, 27 Feb 2026 22:21:23 +0100, Marc Haber wrote:

rbowman <bowman@montana.com> wrote:

On Fri, 27 Feb 2026 11:03:17 +0100, Marc Haber wrote:

rbowman <bowman@montana.com> wrote:

'sudo apt -install virt-manager' pulls down all the necessary >>>>packages.
'sudo usermod -a -G libvirt $(whoami)'

Virtual Machine Manager shows up under Systems Tools.

I THINK that you also need to install the correct qemu-system package
(I guess it would be qemu-system-arm64) to actually be able to start
VMs on the Raspi. Probably more (see recommend and suggests lists of
the packages in question). Virt-Manager can also be used to control a
remote hypervisor.

The virt-manager install does install many qemu packages. What I've been >>trying to run are images, not isos. I'm going to try the Ubuntu Server
iso.

Whether you're booting your system from a hard disk image or an ISO is irrelevant to the fact that you need the right KVM/qemu userspace to
even start your VM. You can run a live system from an ISO without having
a virtual hard disk configure and you can of course boot an installed
system from a virtual hard disk.

fwiw, I was able to create a VM on the Pi using the Ubuntu amr64 iso. It's
not very useful other than a proof of concept but I didn't want a full fat iso. It came up, went through the usual installation and rebooted successfully. I can stop the VM from virt-manager, start it again, and it reboots. I can then log in normally as the user I created during
installation. Looking at /boot it does have an efi subdirectory.

It's been fun but it's time to get back to my regularly scheduled
programming. There are a few things I may research at a later data. With Raspberry Pi OS (Trixie) lsmod doesn't show kvm although the virt-manager install and functionality works the same as on x64.

The other probably gets into NAT and isn't something I want to deal with.
The virtual NAT works as far as accessing the outside world. However all
my boxes use wireless connections and I don't think the bridging schemes
work when the host's connection is wireless.

It was fine playing with Leap but when I brought up Ubuntu Server the
question crossed my mind 'exactly how am I going to serve anything but the host?'
--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

On Fri, 27 Feb 2026 22:19:44 +0100, Marc Haber wrote:

rbowman <bowman@montana.com> wrote:

lsmod on the Pi does not show it although 'modinfo kvm' shows it as a >>builtin. That might be a Pi kernel tweak.

grep KVM /boot/config-$KERNELVERSION or somewhere in /proc, I think it's /proc/sys/kernel/config or something similar.

Greetings Marc

There are 19 lines in the config, CONFIG_KVM_* and CONFIG_HAVE_KVM_* and
all are 'y'

The Ubuntu box has 49 but the Pi seems to have enough to get the job
done.
--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

Lawrence D-|Oliveiro <ldo@nz.invalid> wrote:

On Fri, 27 Feb 2026 22:19:44 +0100, Marc Haber wrote:

grep KVM /boot/config-$KERNELVERSION or somewhere in /proc, I think
it's /proc/sys/kernel/config or something similar.

As I recall, the ability to interrogate the kernel itself to discover
its build configuration settings is not enabled on DebianrCOs kernel
builds.

Yes, right. Because we ship that as a file, the first grep option I
listed.

Greetings
Marc
-- ---------------------------------------------------------------------------- Marc Haber | " Questions are the | Mailadresse im Header Rhein-Neckar, DE | Beginning of Wisdom " |
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 6224 1600402
--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

rbowman <bowman@montana.com> wrote:

On Fri, 27 Feb 2026 22:19:44 +0100, Marc Haber wrote:

rbowman <bowman@montana.com> wrote:

lsmod on the Pi does not show it although 'modinfo kvm' shows it as a >>>builtin. That might be a Pi kernel tweak.

grep KVM /boot/config-$KERNELVERSION or somewhere in /proc, I think it's
/proc/sys/kernel/config or something similar.

There are 19 lines in the config, CONFIG_KVM_* and CONFIG_HAVE_KVM_* and
all are 'y'

Interesting is CONFIG_KVM.

Greetings
Marc
-- ---------------------------------------------------------------------------- Marc Haber | " Questions are the | Mailadresse im Header Rhein-Neckar, DE | Beginning of Wisdom " |
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 6224 1600402
--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

rbowman <bowman@montana.com> wrote:

fwiw, I was able to create a VM on the Pi using the Ubuntu amr64 iso.

Yes. The VM has UEFI and gets the hardware emulated.

The actual hardware boots very differently.

It's been fun but it's time to get back to my regularly scheduled >programming. There are a few things I may research at a later data. With >Raspberry Pi OS (Trixie) lsmod doesn't show kvm although the virt-manager >install and functionality works the same as on x64.

grep KVM= /boot/config-$KERNELVERSION will show =y, as we have learned
in another part of the thread.

The other probably gets into NAT and isn't something I want to deal with. >The virtual NAT works as far as accessing the outside world. However all
my boxes use wireless connections and I don't think the bridging schemes >work when the host's connection is wireless.

It may work. At least OpenWRT happily bridges ethernets and Wifis
together. Beyond that (it's the VPN router that connects my network to
one of my customers), I don't have any machine with Wifi that has the
Wifi in a bridge, sorry.

It was fine playing with Leap but when I brought up Ubuntu Server the >question crossed my mind 'exactly how am I going to serve anything but the >host?'

If nothing else works, with destination NAT.

Greetings
Marc
-- ---------------------------------------------------------------------------- Marc Haber | " Questions are the | Mailadresse im Header Rhein-Neckar, DE | Beginning of Wisdom " |
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 6224 1600402
--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

At Wed, 25 Feb 2026 20:03:48 -0500, c186282 <c186282@nnada.net> wrote:

On 2/25/26 07:13, The Natural Philosopher wrote:

I did some research. There are known issues that prevent this working
out of the box..

They seem to be due to Trixie having dropped some libraries that the current vbox needs and a need to remove the kvm nodule,

Whose 'fault' this is is a worm can I will refrain from opening...

Deb used to be fairly 'stable' - what would
work 20 years ago would still work fine, you
didn't have to go chasing this months suite
of tools to get things done.

Now ... starting with BullsEye ... that solid
rock rep seems to be crumbling. This is made
even more critical because so many other distros
use Deb as their base.

I will admit that while BullsEye gave me fits when
it first came out it DID get its act together pretty
well later on. Trixie, like Bullseye, may have just
been born a bit premature.

We shall see.

Well, I don't agree with your opinion, but I respect it.

Personally, I wonder why you'd not avail yourself of the
opportunity to, you know, talk to a Debian dev. Maybe even
offer to help?

There's clearly some pressure for orgs to pop the
newest version out the door. However this leads
to issues that can run down the rep of your distro,
counter-productive.

We certainly wouldn't want to be counter-productive.

Mint is based on a LTS version of Ubuntu, which is
apparently based on a mature Bullseye. VBox runs
just fine on it. Now I just have to figure out why
the Manjaro VM won't mount my NFS shares even though
it can see the machine. Maybe I didn't load some
library ... gotta fool around with it.

It's probably the way NAT works with your vhost's virtual
network -- it could be that your source port is helpfully
translated to a source port outside the "secure" ports,
so called.

"man exports", and take a look at the discussion of the
"secure/nosecure" exports option.

Next up
is the GhostBSD VM. In the end there can be only
one but I'm gonna give both a fair shake.

--
-v System76 Thelio Mega v1.1 x86_64 Mem: 258G
OS: Linux 6.19.5 D: Mint 22.3 DE: Xfce 4.18 (X11)
NVIDIA GeForce RTX 3090Ti (24G) (580.126.18)
"Move your vowels every day or you'll get consonated."
--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

On 3/2/26 22:41, vallor wrote:

At Wed, 25 Feb 2026 20:03:48 -0500, c186282 <c186282@nnada.net> wrote:

On 2/25/26 07:13, The Natural Philosopher wrote:

I did some research. There are known issues that prevent this working
out of the box..

They seem to be due to Trixie having dropped some libraries that the
current vbox needs and a need to remove the kvm nodule,

Whose 'fault' this is is a worm can I will refrain from opening...

Deb used to be fairly 'stable' - what would
work 20 years ago would still work fine, you
didn't have to go chasing this months suite
of tools to get things done.

Now ... starting with BullsEye ... that solid
rock rep seems to be crumbling. This is made
even more critical because so many other distros
use Deb as their base.

I will admit that while BullsEye gave me fits when
it first came out it DID get its act together pretty
well later on. Trixie, like Bullseye, may have just
been born a bit premature.

We shall see.

Well, I don't agree with your opinion, but I respect it.

Personally, I wonder why you'd not avail yourself of the
opportunity to, you know, talk to a Debian dev. Maybe even
offer to help?

I was implementing Linux boxes at my old workplace
long before most anybody had even HEARD of it -
because I saw the superiority. Red Hat and old
pre-Open SUSE.

Now I can complain to Deb ... but I'm just one guy.
I think someone in management is pressuring the
too-early release of new versions. As said, Bullseye
gave me lots of shit (issues posted here a couple
years ago) and now it's Trixie. Had a very unflattering
alt name for Bullseye :-)

Deb USED to be famous for getting things dead stable
before a new release. Now, not so much. That's BAD,
esp because SO many distros are based on Deb.

There's clearly some pressure for orgs to pop the
newest version out the door. However this leads
to issues that can run down the rep of your distro,
counter-productive.

We certainly wouldn't want to be counter-productive.

Well ... Deb in particular SEEMS to be sabotaging
the rep of its own product. However I've seen a
Fedora 'released too soon' as well. There seems
to be a 'version fever' at work.

Hey - get it RIGHT first, THEN release !

Mint is based on a LTS version of Ubuntu, which is
apparently based on a mature Bullseye. VBox runs
just fine on it. Now I just have to figure out why
the Manjaro VM won't mount my NFS shares even though
it can see the machine. Maybe I didn't load some
library ... gotta fool around with it.

It's probably the way NAT works with your vhost's virtual
network -- it could be that your source port is helpfully
translated to a source port outside the "secure" ports,
so called.

VirtualBox is good at setting up a bridge connection
so the VM can be on your home subnet. KVM does NOT
seem good at bridges. Alas, with Trixie, VBox does
not work well ... even the newest versions straight
from the Oracle site that CLAIM they're for Trixie.

Maybe in six months Deb will include an 'official'
VBox in its repo that WILL make the needed adjustments
to work properly. But that's "the future" ...

"man exports", and take a look at the discussion of the
"secure/nosecure" exports option.

Next up
is the GhostBSD VM. In the end there can be only
one but I'm gonna give both a fair shake.

Haven't installed GBSD on Trixie KVM yet ... I'll
get around to it maybe this week. I've looked at
GBSD before and it's a pretty nice/friendly take
on BSD, even more than its native FreeBSD.

However there's still the networking issue. That
is KVM, not the particular distro.

BSDs ... Ghost and Dragonfly are probably the most
'accessible' to Linux users. Free/Open are more
'industrial' and often more work than it's worth
unless you're doing For-Money servers.

As mentioned to rbowman ... it came to me that I have
at least one wifi->wired 'bridge' thingie around the
house. I *could* plug that into the wired port on
my box and set it to link to my main local wifi ...
convert the hard port KVM likes to de-facto wifi.
It's on my decent, but totally unused, desktop unit
in the back room. All I do now is laptops and those
little BMax/BeeLink boxes (love those for Linux).

--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

c186282 <c186282@nnada.net> writes:

VirtualBox is good at setting up a bridge connection so the VM can
be on your home subnet. KVM does NOT seem good at bridges.

Bridging works fine with KVM, IrCOve been using it that way for years.

Alas, with Trixie, VBox does not work well ... even the newest
versions straight from the Oracle site that CLAIM they're for
Trixie.

It worked fine when I tried it.
--
https://www.greenend.org.uk/rjk/
--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

On 3/3/26 03:42, Richard Kettlewell wrote:

c186282<c186282@nnada.net> writes:

VirtualBox is good at setting up a bridge connection so the VM can
be on your home subnet. KVM does NOT seem good at bridges.

Bridging works fine with KVM, IrCOve been using it that way for years.

Alas, with Trixie, VBox does not work well ... even the newest
versions straight from the Oracle site that CLAIM they're for
Trixie.

It worked fine when I tried it.

-- https://www.greenend.org.uk/rjk/

Try it NOW.

My host is generic Deb 'Trixie'.

KVM does NOT see/offer any 'bridge'
connection that WORK ... and I've
spent HOURS on it. Also no wifi
connections. Wired connections - they
can exist, but aren't on MY subnet.

VBox makes it easy ... but, so far, does
not work right on Trixie.

Maybe my need has just fallen through the
disto cracks - Trixie NOT REALLY READY.
Had similar disappointments with Bullseye
right out of the box.

I'd LIKE to stay current ... but Deb in
particular now seems to release before the
distro is ACTUALLY good/stable. Not great.

Didn't used to be THIS hard ....

--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

vallor <vallor@vallor.earth> wrote:

Personally, I wonder why you'd not avail yourself of the
opportunity to, you know, talk to a Debian dev. Maybe even
offer to help?

That person CAN talk to a Debian Developer, actually they are doing
this in this very second. But since they cannot put any technical
content into their complaints AND do have substantial misunderstanding
about Debian works, the Debian Developer in this group is rapidly
losing motivation. Actually already fully has.

Greetings
Marc
-- ---------------------------------------------------------------------------- Marc Haber | " Questions are the | Mailadresse im Header Rhein-Neckar, DE | Beginning of Wisdom " |
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 6224 1600402
--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.misc

On 03/03/2026 09:52, Marc Haber wrote:

vallor <vallor@vallor.earth> wrote:

Personally, I wonder why you'd not avail yourself of the
opportunity to, you know, talk to a Debian dev. Maybe even
offer to help?

That person CAN talk to a Debian Developer, actually they are doing
this in this very second. But since they cannot put any technical
content into their complaints AND do have substantial misunderstanding
about Debian works, the Debian Developer in this group is rapidly
losing motivation. Actually already fully has.

Greetings
Marc

And anyway, it is a question of Debian policy wrt to Oracle more than a 'technical ' issue

Its like every time I plug my android phone to linux [Mint] USB it pops
up an notification box saying No such interface "org.gtk.vfs.Mount"

Linux developers say it is all the fault of Android.

It all still works fine despite the error.
--
Of what good are dead warriors? rCa Warriors are those who desire battle
more than peace. Those who seek battle despite peace. Those who thump
their spears on the ground and talk of honor. Those who leap high the
battle dance and dream of glory rCa The good of dead warriors, Mother, is
that they are dead.
Sheri S Tepper: The Awakeners.

--- Synchronet 3.21d-Linux NewsLink 1.2