Re: Example miboot on new world rom problem

Re: Example miboot on new world rom problem

[jcarr]

> >I was working with a guy who was using startup disk on a new portable. It
> >was sending him into into miboot. I don't know how to fix this
> >problem. I think the only real option is to make a new startup disk
> >control panel for the macos.
> >
> >Basically this guy has to go into OF and do:
> >boot hd:9,\\:tbxi
> >
> >To get to yaboot. This problem is going to kick me in the butt. Heck it's
> >going to kick linux on the powerpc in general in the butt.
>
> Well, I have no problem triggering yaboot & the OF script via Startup
> Disk on the Pismo using a fake system folder.
> What is this user's setup exactly ?

Some new ibook I think. I get the same behavior on the iMac DV. Startup
disk doesn't change the boot-device value on there. Instead, it loads the
first TBXI file from the first HFS partition it finds. Then the MacOS ROM
finds out that "LinuxPPC Boot" is the name of the partition to boot and
then continues...

Thus miboot is loaded on a new world rom machine.

This is the reason I think either:

1) yaboot needs the code from pmac_support.c (aka SetPRAMBoot )

2) We'll have to write a new startup disk control panel( like system disk)

Jeff
jcarr@linuxppc.org


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Example miboot on new world rom problem

[Ethan Benson]

[-- Attachment #1: Type: text/plain, Size: 910 bytes --]

On Thu, Nov 09, 2000 at 08:57:21AM -0600, jcarr@mail.linuxppc.org wrote:
> 
> Some new ibook I think. I get the same behavior on the iMac DV. Startup
> disk doesn't change the boot-device value on there. Instead, it loads the

thats insane, probably a bug in apple's startup disk control panel...

> first TBXI file from the first HFS partition it finds. Then the MacOS ROM
> finds out that "LinuxPPC Boot" is the name of the partition to boot and
> then continues...
> 
> Thus miboot is loaded on a new world rom machine.
> 
> This is the reason I think either:
> 
> 1) yaboot needs the code from pmac_support.c (aka SetPRAMBoot )
> 
> 2) We'll have to write a new startup disk control panel( like system disk)

or 3)  setup the correct type of bootstrap for the current type of
machine.  then you don't have any of this nonsense.  

-- 
Ethan Benson
http://www.alaska.net/~erbenson/

[-- Attachment #2: Type: application/pgp-signature, Size: 0 bytes --]

Re: Example miboot on new world rom problem

[jcarr]

> >Some new ibook I think. I get the same behavior on the iMac DV. Startup
> >disk doesn't change the boot-device value on there. Instead, it loads the
> >first TBXI file from the first HFS partition it finds. Then the MacOS ROM
> >finds out that "LinuxPPC Boot" is the name of the partition to boot and
> >then continues...
> >
> >Thus miboot is loaded on a new world rom machine.
> >
> >This is the reason I think either:
> >
> >1) yaboot needs the code from pmac_support.c (aka SetPRAMBoot )
> >
> >2) We'll have to write a new startup disk control panel( like system disk)
>
> Whatever you do with SetPRAMBoot won't help since miBoot won't boot reliably
> the new machines. The problem is to have MacOS Startup Disk properly set boot
> to the yaboot image... I beleive you checked that you have the proper
> <COMPATIBLE> entries for those new machines in the OF boot script ?
> (tbxi should be a script that in turns load yaboot, look at ybin scripts).

Yes I need to look at the ybin script again. However, this doesn't fix the
problem of miBoot running on New World ROM machines and also does not
provide a solution to the problem either. The problem that the guy was
having with his ibook and I have with the iMac DV. The only way that this
guy is left to boot yaboot is to go into OF and do:

boot hd:9,\\:tbxi

The functionality of SetPRAMBoot(by that I mean the ability to do what
startup disk does) needs to be added to yaboot not for the reasons I
stated earlier. So I was wrong about the reasons I stated in the last
email. Sorry about that... this all can be rather convoluted! This
functionality needs to be there if yaboot is to ever allow users to select
thier MacOS boot partitions.

In the other case, miBoot needs to have the ability to change the
boot-device if it detects that it is running on a New World ROM
machine. Maybe Harry Eaton or Mike Tesch are still around and can help
grok bootvars code to add to miboot?

Jeff


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Example miboot on new world rom problem

[Joseph P. Garcia]

On  9 Nov, jcarr@mail.linuxppc.org wrote:
> The functionality of SetPRAMBoot(by that I mean the ability to do what
> startup disk does) needs to be added to yaboot not for the reasons I

'functionality'...  has anyone tested/written new code SPB for other
systems yet?

also must remember to add the bootvars/nvsetenv/ybin functionality for
the OF settings too.  Darwin seems to have its own System Disk program
that has XPRAM and OF covered.. except the pram part.  If that project
is open, might we want to try some code sharing?  Or does Apple own it?
Or might Apple be willing to share 9's Startup Disk info...

--
Joseph P. Garcia
http://www.execpc.com/~jpgarcia


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Example miboot on new world rom problem

[jcarr]

On Thu, 9 Nov 2000, Ethan Benson wrote:

> On Thu, Nov 09, 2000 at 08:57:21AM -0600, jcarr@mail.linuxppc.org wrote:
> >
> > Some new ibook I think. I get the same behavior on the iMac DV. Startup
> > disk doesn't change the boot-device value on there. Instead, it loads the
>
> thats insane, probably a bug in apple's startup disk control panel...

Tell me about it. :)

> > first TBXI file from the first HFS partition it finds. Then the MacOS ROM
> > finds out that "LinuxPPC Boot" is the name of the partition to boot and
> > then continues...
> >
> > Thus miboot is loaded on a new world rom machine.
> >
> > This is the reason I think either:
> >
> > 1) yaboot needs the code from pmac_support.c (aka SetPRAMBoot )
> >
> > 2) We'll have to write a new startup disk control panel( like system disk)
>
> or 3)  setup the correct type of bootstrap for the current type of
> machine.  then you don't have any of this nonsense.

OK - I'll take a look at the current incarnation
Jeff


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Example miboot on new world rom problem

[Ethan Benson]

[-- Attachment #1: Type: text/plain, Size: 1938 bytes --]

On Thu, Nov 09, 2000 at 11:12:09AM -0600, jcarr@mail.linuxppc.org wrote:

> Yes I need to look at the ybin script again. However, this doesn't fix the
> problem of miBoot running on New World ROM machines and also does not
> provide a solution to the problem either. The problem that the guy was
> having with his ibook and I have with the iMac DV. The only way that this
> guy is left to boot yaboot is to go into OF and do:

setup a 800K Apple_Bootstrap partition where the newworld bootstrap is
installed, that will solve your problem.  if you put it first you
don't need to do anything to OpenFirmwares configuration, just reset
it to defaults.  

> boot hd:9,\\:tbxi
> 
> The functionality of SetPRAMBoot(by that I mean the ability to do what
> startup disk does) needs to be added to yaboot not for the reasons I

huh?  i don't see why.  recent ybin creates a nice boot menu that
takes care of loading macos just fine without any of this StartupDisk garbage.

> stated earlier. So I was wrong about the reasons I stated in the last
> email. Sorry about that... this all can be rather convoluted! This
> functionality needs to be there if yaboot is to ever allow users to select
> thier MacOS boot partitions.

making yaboot load macos would be non-trivial.  however with current
ybin versions you can add `macos=/dev/hda10' to /etc/yaboot.conf and
ybin will build a menu that is displayed on bootup.  works very
nicely.  

> In the other case, miBoot needs to have the ability to change the
> boot-device if it detects that it is running on a New World ROM
> machine. Maybe Harry Eaton or Mike Tesch are still around and can help
> grok bootvars code to add to miboot?

miboot needs alot of things...  simplest way to deal with this is
don't install miboot on a newworld!  we don't install lilo on sparcs
or silo on x86 boxes afterall.  

-- 
Ethan Benson
http://www.alaska.net/~erbenson/

[-- Attachment #2: Type: application/pgp-signature, Size: 0 bytes --]

Re: Example miboot on new world rom problem

[jcarr]

> setup a 800K Apple_Bootstrap partition where the newworld bootstrap is
> installed, that will solve your problem.  if you put it first you
> don't need to do anything to OpenFirmwares configuration, just reset
> it to defaults.

Ok. That sounds workable. Is that what OSX does also? If so, will it
conflict?

> making yaboot load macos would be non-trivial.  however with current

True!


> > In the other case, miBoot needs to have the ability to change the
> > boot-device if it detects that it is running on a New World ROM
> > machine. Maybe Harry Eaton or Mike Tesch are still around and can help
> > grok bootvars code to add to miboot?
>
> miboot needs alot of things...  simplest way to deal with this is
> don't install miboot on a newworld!  we don't install lilo on sparcs
> or silo on x86 boxes afterall.

Well, I think that the macos will not like the system folder unless it has
a system file. But whatever we put in the system folder as the system
file, we will still have the problem we were having:

Anyone that has a partition visable to the macos and trys to boot off of
it will still get miboot. And, I think, It will be hard to get people not
to expect that to work.
Jeff


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Example miboot on new world rom problem

[Ethan Benson]

[-- Attachment #1: Type: text/plain, Size: 2014 bytes --]


On Fri, Nov 10, 2000 at 04:53:32PM -0600, jcarr@mail.linuxppc.org wrote:
> > setup a 800K Apple_Bootstrap partition where the newworld bootstrap is
> > installed, that will solve your problem.  if you put it first you
> > don't need to do anything to OpenFirmwares configuration, just reset
> > it to defaults.  
> 
> Ok. That sounds workable. Is that what OSX does also? If so, will it
> conflict?

when OSX is installed on an OldWorld or with UFS filesystems it
creates two bootstrap partitions, the first being Apple_Boot the
second being Apple_Secondary or something close to that.  since i have
been using Apple_Bootstrap OSX appears to ignore ours.  (the installer
does sometimes reorder all the partitions around to put its own first,
which is annoying since it hoses /etc/fstab among other things.  its
fixable with mac-fdisk though i added explanation on how to my
mac-fdisk doc)

> > miboot needs alot of things...  simplest way to deal with this is
> > don't install miboot on a newworld!  we don't install lilo on sparcs
> > or silo on x86 boxes afterall.  
> 
> Well, I think that the macos will not like the system folder unless it has
> a system file. But whatever we put in the system folder as the system

the MacOS does not like fake system folders of any kind, even these
fake ones you create are being deblessed and thus made unbootable by
MacOS very regularly.  i have a large mailbox of users asking me what
to do about this, i always tell them to change the partition type of
thier bootstrap partition to Apple_Bootstrap so MacOS can't see it
anymore.  that is the only solution that works 100%  

> file, we will still have the problem we were having:
> 
> Anyone that has a partition visable to the macos and trys to boot off of
> it will still get miboot. And, I think, It will be hard to get people not
> to expect that to work.
> Jeff

thats why you don't install miboot on a newworld ;-)

-- 
Ethan Benson
http://www.alaska.net/~erbenson/

[-- Attachment #2: Type: application/pgp-signature, Size: 0 bytes --]