Hello!

I would like to ask users running Debian on a real PowerMac machine (G3, G4, G5 etc)
to help me verify a theory on the boot mechanism. In particular, the question is
whether we can skip setting the boot-device in NVRAM in the grub-installer script
which causes incompatibilities with the PowerMac emulation in QEMU [1].

According to Apple's documentation [2], Open Firmware will automatically search for
available operating systems, so there is no need to explicitly set the path to the boot device.

To verify this, I set the boot device on my iBook G4 to NULL and checked whether the
machine would still boot which turns out to be true. However, before I change the
code in grub-installer, I want to make sure it won't break any other PowerMacs.

So, if you would like to help, please try the following.

As root, run:

root@ibook-g4:~# nvram --print-config=boot-device /pci@f4000000/ata-6@d/@0:2,\\:txbi
root@ibook-g4:~# nvram --update-config boot-device=""
root@ibook-g4:~# nvram --print-config=boot-device
root@ibook-g4:~# reboot

If your machine still comes up after changing "boot-device" to zero, we know that
it's safe to drop the NVRAM part from the grub-installer script for PowerMacs.

This will fix the remaining compatibility with QEMU.

Thanks,
Adrian

[1] https://lists.debian.org/debian-powerpc/2022/03/msg00029.html
[2] https://opensource.apple.com/source/bless/bless-37/README.BOOTING

--
.''`. John Paul Adrian Glaubitz
: :' : Debian Developer - [email protected]
`. `' Freie Universitaet Berlin - [email protected]
`- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

hi,

nvram --print-config='boot-device /pci@f4000000/ata-6@d/@0:2,\\:txbi'
nvram --update-config boot-device=""
nvram --print-config=boot-device
reboot

result is another login prompt

On Fri, Mar 25, 2022, 12:44 AM John Paul Adrian Glaubitz < [email protected]> wrote:

Hello!

I would like to ask users running Debian on a real PowerMac machine (G3,
G4, G5 etc)
to help me verify a theory on the boot mechanism. In particular, the
question is
whether we can skip setting the boot-device in NVRAM in the grub-installer script
which causes incompatibilities with the PowerMac emulation in QEMU [1].

According to Apple's documentation [2], Open Firmware will automatically search for
available operating systems, so there is no need to explicitly set the
path to
the boot device.

To verify this, I set the boot device on my iBook G4 to NULL and checked whether the
machine would still boot which turns out to be true. However, before I
change the
code in grub-installer, I want to make sure it won't break any other PowerMacs.

So, if you would like to help, please try the following.

As root, run:

root@ibook-g4:~# nvram --print-config=boot-device /pci@f4000000/ata-6@d/@0:2,\\:txbi
root@ibook-g4:~# nvram --update-config boot-device=""
root@ibook-g4:~# nvram --print-config=boot-device
root@ibook-g4:~# reboot

If your machine still comes up after changing "boot-device" to zero, we
know that
it's safe to drop the NVRAM part from the grub-installer script for PowerMacs.

This will fix the remaining compatibility with QEMU.

Thanks,
Adrian

[1] https://lists.debian.org/debian-powerpc/2022/03/msg00029.html
[2] https://opensource.apple.com/source/bless/bless-37/README.BOOTING

--
.''`. John Paul Adrian Glaubitz
: :' : Debian Developer - [email protected]
`. `' Freie Universitaet Berlin - [email protected]
`- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913

<div dir="auto"><div dir="auto"><div dir="auto"><div dir="auto">hi,</div><div dir="auto"><br></div><div dir="auto">nvram --print-config=&#39;boot-device /pci@f4000000/ata-6@d/@0:2,\\:txbi&#39;</div><div dir="auto">nvram --update-config boot-device=&quot;&
quot;</div><div dir="auto">nvram --print-config=boot-device</div></div></div><div dir="auto">reboot</div><div dir="auto"><br></div><div dir="auto">result is another login prompt</div><br><div class="gmail_quote" dir="auto"><div dir="ltr" class="gmail_
attr">On Fri, Mar 25, 2022, 12:44 AM John Paul Adrian Glaubitz &lt;<a href="mailto:[email protected]" target="_blank" rel="noreferrer">[email protected]</a>&gt; wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .
8ex;border-left:1px #ccc solid;padding-left:1ex">Hello!<br>

I would like to ask users running Debian on a real PowerMac machine (G3, G4, G5 etc)<br>
to help me verify a theory on the boot mechanism. In particular, the question is<br>
whether we can skip setting the boot-device in NVRAM in the grub-installer script<br>
which causes incompatibilities with the PowerMac emulation in QEMU [1].<br>

According to Apple&#39;s documentation [2], Open Firmware will automatically search for<br>
available operating systems, so there is no need to explicitly set the path to<br>
the boot device.<br>

To verify this, I set the boot device on my iBook G4 to NULL and checked whether the<br>
machine would still boot which turns out to be true. However, before I change the<br>
code in grub-installer, I want to make sure it won&#39;t break any other PowerMacs.<br>

So, if you would like to help, please try the following.<br>

As root, run:<br>

root@ibook-g4:~# nvram --print-config=boot-device<br> /pci@f4000000/ata-6@d/@0:2,\\:txbi<br>
root@ibook-g4:~# nvram --update-config boot-device=&quot;&quot;<br> root@ibook-g4:~# nvram --print-config=boot-device<br>
root@ibook-g4:~# reboot<br>

If your machine still comes up after changing &quot;boot-device&quot; to zero, we know that<br>
it&#39;s safe to drop the NVRAM part from the grub-installer script for PowerMacs.<br>

This will fix the remaining compatibility with QEMU.<br>

Thanks,<br>
Adrian<br>

&gt; [1] <a href="https://lists.debian.org/debian-powerpc/2022/03/msg00029.html" rel="noreferrer noreferrer noreferrer" target="_blank">https://lists.debian.org/debian-powerpc/2022/03/msg00029.html</a><br>
&gt; [2] <a href="https://opensource.apple.com/source/bless/bless-37/README.BOOTING" rel="noreferrer noreferrer noreferrer" target="_blank">https://opensource.apple.com/source/bless/bless-37/README.BOOTING</a><br>

-- <br>
 .&#39;&#39;`.  John Paul Adrian Glaubitz<br>
: :&#39; :  Debian Developer - <a href="mailto:[email protected]" rel="noreferrer noreferrer" target="_blank">[email protected]</a><br>
`. `&#39;   Freie Universitaet Berlin - <a href="mailto:[email protected]" rel="noreferrer noreferrer" target="_blank">[email protected]</a><br>
  `-    GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913<br>

</blockquote></div>
</div>

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hello Cameron!

On 3/25/22 09:40, Cameron MacPherson wrote:

nvram --print-config='boot-device /pci@f4000000/ata-6@d/@0:2,\\:txbi'
nvram --update-config boot-device=""
nvram --print-config=boot-device
reboot

result is another login prompt

On which machine exactly?

And by "another login prompt" you mean the machine booted normally?

Thanks,
Adrian

--
.''`. John Paul Adrian Glaubitz
: :' : Debian Developer - [email protected]
`. `' Freie Universitaet Berlin - [email protected]
`- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hello!

On 3/25/22 09:56, John Ogness wrote:

I suppose it would be better if someone did the test on a system where
*only* Linux is installed, to make sure that OpenFirmware is able to
find GRUB without boot-device being set.

I have already tested that on my iBook G4 and it still boots fine into
Linux with the boot-device set to NULL.

Adrian

--
.''`. John Paul Adrian Glaubitz
: :' : Debian Developer - [email protected]
`. `' Freie Universitaet Berlin - [email protected]
`- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 2022-03-25, John Paul Adrian Glaubitz <[email protected]> wrote:

I would like to ask users running Debian on a real PowerMac machine
(G3, G4, G5 etc) to help me verify a theory on the boot mechanism. In particular, the question is whether we can skip setting the
boot-device in NVRAM in the grub-installer script which causes incompatibilities with the PowerMac emulation in QEMU [1].

I have a Dual PowerMac G5 (PowerMac7,2), PPC970 revision 2.2 (pvr 0039
0202).

I have both MacOSX and Linux installed. After clearing the boot-device
variable in NVRAM, the machine booted normally (into MacOSX).

I suppose it would be better if someone did the test on a system where
*only* Linux is installed, to make sure that OpenFirmware is able to
find GRUB without boot-device being set.

John Ogness

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

El vie, 25-03-2022 a las 08:43 +0100, John Paul Adrian Glaubitz
escribió:

Hello!

I would like to ask users running Debian on a real PowerMac machine
(G3, G4, G5 etc)
to help me verify a theory on the boot mechanism. In particular, the
question is
whether we can skip setting the boot-device in NVRAM in the grub-
installer script
which causes incompatibilities with the PowerMac emulation in QEMU
[1].

According to Apple's documentation [2], Open Firmware will
automatically search for
available operating systems, so there is no need to explicitly set
the path to
the boot device.

To verify this, I set the boot device on my iBook G4 to NULL and
checked whether the
machine would still boot which turns out to be true. However, before
I change the
code in grub-installer, I want to make sure it won't break any other PowerMacs.

So, if you would like to help, please try the following.

As root, run:

root@ibook-g4:~# nvram --print-config=boot-device /pci@f4000000/ata-6@d/@0:2,\\:txbi
root@ibook-g4:~# nvram --update-config boot-device=""
root@ibook-g4:~# nvram --print-config=boot-device
root@ibook-g4:~# reboot

If your machine still comes up after changing "boot-device" to zero,
we know that
it's safe to drop the NVRAM part from the grub-installer script for PowerMacs.

This will fix the remaining compatibility with QEMU.

Thanks,
Adrian

Hi,

Powerbook G4

cpu : 7447/7457, altivec supported
clock : 1249.999000MHz
revision : 1.1 (pvr 8002 0101)
bogomips : 60.23

timebase : 18432000
platform : PowerMac
model : PowerBook5,2
machine : PowerBook5,2
motherboard : PowerBook5,2 MacRISC3 Power Macintosh
detected as : 287 (PowerBook G4 15")
pmac flags : 0000001a
L2 cache : 512K unified
pmac-generation : NewWorld
Memory : 512 MB

normal bootup after reboot.

[1] https://lists.debian.org/debian-powerpc/2022/03/msg00029.html
[2]
https://opensource.apple.com/source/bless/bless-37/README.BOOTING

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 3/25/22 03:43, John Paul Adrian Glaubitz wrote:

Hello!

I would like to ask users running Debian on a real PowerMac machine (G3, G4, G5 etc)
to help me verify a theory on the boot mechanism.

.
.
.

As root, run:

root@ibook-g4:~# nvram --print-config=boot-device /pci@f4000000/ata-6@d/@0:2,\\:txbi
root@ibook-g4:~# nvram --update-config boot-device=""
root@ibook-g4:~# nvram --print-config=boot-device
root@ibook-g4:~# reboot

If your machine still comes up after changing "boot-device" to zero, we know that
it's safe to drop the NVRAM part from the grub-installer script for PowerMacs.

From the "better late than never" department I have a PowerMac G5 "quad"
which runs great. Or at least it once did :

enceladus#
enceladus# nvram --print-config
"common" Partition
---------------------
little-endian?=false
real-mode?=false
auto-boot?=true
diag-switch?=false
fcode-debug?=false
oem-banner?=false
oem-logo?=false
use-nvramrc?=false
use-generic?=false
default-mac-address?=false
scroll-lock=true
skip-netboot?=false
real-base=-1
real-size=-1
load-base=0x800000
virt-base=-1
virt-size=-1
logger-base=-1
logger-size=-1
pci-probe-mask=-1
screen-#columns=100
screen-#rows=40
selftest-#megs=0
boot-volume=3
gmt-offset=0 boot-device=/ht@0,f2000000/pci@9/k2-sata-root@c/@0/@0:2,\powerpc-ieee1275\core.elf
boot-file=
boot-screen=screen
console-screen=screen
diag-device=enet
diag-file=,diags
input-device=keyboard
output-device=screen
input-device-1=/ipc
output-device-1=/ipc
mouse-device=mouse
oem-banner=
oem-logo=
nvramrc=
boot-command=mac-boot
default-client-ip=
default-server-ip=
default-gateway-ip=
default-subnet-mask=
default-router-ip=
boot-script=
aapl,pci=
boot-args=
aapl,tdm-units=
ram-size=0x28000000
boot-once=
boot-last-label=Linux

enceladus# cd
enceladus# nvi nvram.dat
enceladus#
enceladus#
enceladus#
enceladus#
enceladus# nvram --print-config=boot-device /ht@0,f2000000/pci@9/k2-sata-root@c/@0/@0:2,\powerpc-ieee1275\core.elf enceladus#
enceladus# nvram --update-config boot-device=""
enceladus#
enceladus# nvram --print-config=boot-device
enceladus#
enceladus# reboot
Connection to enceladus closed by remote host.
Connection to enceladus closed.


Then the machine DOES NOT boot.

I see :


Welcome to GRUB!

error: symbol `grub_disk_native_sectors' not found.
grub rescue>


Then, of course, the fans being their full speed noises.

So, any ideas on what to type at that prompt?


--
Dennis Clarke
RISC-V/SPARC/PPC/ARM/CISC
UNIX and Linux spoken
GreyBeard and suspenders optional


--
Dennis Clarke
RISC-V/SPARC/PPC/ARM/CISC
UNIX and Linux spoken
GreyBeard and suspenders optional

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Mar 27, 2022, at 10:51 PM, Dennis Clarke <[email protected]> wrote:

On 3/25/22 03:43, John Paul Adrian Glaubitz wrote:

Then the machine DOES NOT boot.

I see :

Welcome to GRUB!

error: symbol `grub_disk_native_sectors' not found.
grub rescue>

This change is unrelated to the change. It indicates a broken GRUB installation.

Then, of course, the fans being their full speed noises.

So, any ideas on what to type at that prompt?

Try invoking the boot menu by holding down <Alt> and see if you can find an alternative working boot loader.

If that doesn’t work, then my suspicion is that your installation is rather old and GRUB on this system was installed using the old hacky scripts.

I don’t think the G5 behaves differently here than my G4.

Adrian

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 3/27/22 17:07, John Paul Adrian Glaubitz wrote:

On Mar 27, 2022, at 10:51 PM, Dennis Clarke <[email protected]> wrote: >>
On 3/25/22 03:43, John Paul Adrian Glaubitz wrote:

Then the machine DOES NOT boot.

I see :


Welcome to GRUB!

error: symbol `grub_disk_native_sectors' not found.
grub rescue>

This change is unrelated to the change. It indicates a broken GRUB installation.

No surprise at all.

Then, of course, the fans being their full speed noises.

So, any ideas on what to type at that prompt?

Try invoking the boot menu by holding down <Alt> and see if you can find an alternative working boot loader.

If that doesn’t work, then my suspicion is that your installation is rather old and GRUB on this system was installed using the old hacky scripts.

I don’t think the G5 behaves differently here than my G4.

I just fetched the most recent netinst image ( 2022-03-24 ) where I am
guessing a full reinstall will be perfectly possible. I would rather
find a way to repair the old hacky GRUB but sometimes a full reinstall
is a "good thing"(tm).

--
Dennis Clarke
RISC-V/SPARC/PPC/ARM/CISC
UNIX and Linux spoken
GreyBeard and suspenders optional

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)