Subject: Re: mol rpm and startmol problem
From: Alex Samaras (tadpole_005@mac.com)
Date: Sun Mar 26 2000 - 22:53:56 MST
I was having the same problem with the kernel patching with MOL, which you
posted with this error message:
>[root@beautybox root]# startmol
>Using MOL library directory /usr/lib/mol
>This kernel is NOT runtime patched
>Trying to apply MOL runtime patches.
>**** Examining '/boot/System.map' ****
>**** Success ****
>Patching the kernel...
>Rule 7: *** FAILED ***
>No patching was performed
What I gathered from this is that since MOL is patching the kernel in ram on
the fly, it needs a model to know what to look for in RAM. Since (if you are
using bootX) the kernal is loaded into RAM from your mac disk, MOL doesn't
know what to look for in ram, since the kernel is usually in the /boot
directory. So, my solution (which worked, btw) was to copy the kernel I was
using from my mac:system folder:linux kernals in to my /boot directory on my
linux partition. Now that MOL has a model in linux to look at and compair to
RAM, it no longer has a problem patching the kernel.
MOL now starts up cleanly, however, I am having a problem. For some reason
linux will not recognize that my Mac startup disk is even in my computer.
The disk is a 1.2GB quantum fireball(on a PPC 7500) at SCSI ID 0 formatted
with an HFS+ partition(w/ OS9 installed on it). I understand that LinuxPPC
can't recognize HFS+ partitions, but for some reason it won't even recognize
that the device is there. When I run dmesg, it tells me something like:
MESH: SCSI 0 try aborted
[some 4 digit number]: SCSI 1 mesh
and then proceeds to list the partitions on that drive(which are "swap" "/"
and a normal mac HFS partition). Randomly for no apparent reason today, I
ran MOL and it recognized the disk and tried to start up OS9 but froze on
startup and MOL then proceeded to give me the same errors that it gives when
there is no Mac system disk available(almost like MOL started MacOS, and
then "realized" that it shouldn't be able to access that partition due to
the problem above). Does anyone have any ideas on this? A friend of mine
said that it might be some sort of Master/Slave problem(like on IDE), but I
was not aware that SCSI had Master/Slave relationships. I didn't have time
to record all of the fine details of these problems today because I had to
go to a study group, but I will get them next time I try to run MOL. Any
ideas?
One last problem(I know...I'm practically writing a book here). This isn't a
MOL problem, but perhaps some of you can help anyway. I use a cable modem,
and I have some trouble connecting to the net w/ LinuxPPC. On start up, my
computer does communicate with the DHCP server and gets me and IP(which is
apparent when I run ifconfig), but for some reason, netcfg will not get the
IP, so I have to manually enter it into netcfg and reactivate it. This works
about 75% of the time. Is there some way I can script this to happen
automatically or configure netcfg to do this automatically? It is just kind
of annoying to have to do this manually every time I reboot my machine(which
is kind of often since I enjoy switching between MacOS and Linux).
Thanks for your help.
shh listen,
Alex
This archive was generated by hypermail 2a24 : Sun Apr 02 2000 - 21:06:30 MDT