[coLinux-devel] Re: [coLinux-users] coLinux on 2.6.15 ?

[Andreas <ge...@gm...>]

 
On 11:42 Thu 02 Feb     , Henry Nestler wrote:
> 
> 
> Andreas Schäfer wrote:
> 
> >Hi,
> >I've been toying around with the current revision from the Monotone
> >repository and tried to bump the kernel patch to version 2.6.15. It
> >seemed straightforward since only few hunks didn't succeed while
> >patching and even those seemed to be rather trivial.
> >Anyway, apparently it's not as easy. Although the kernel as well as
> >coLinux did compile smoothly, coLinux didn't boot. All I got was the
> >following error log:
> >>c:\harpy>colinux-daemon.exe mem=128 kernel=vmlinux initrd=initrd.gz hda1=smallroot root=/dev/hda1 hdb1=swap_device eth0=pcap-bridge,"Lan"
> >>Cooperative Linux Daemon, 0.7.1
> >>Compiled on Wed Feb  1 17:09:10 2006
> >>using 'vmlinux' as kernel image
> >>using 'initrd.gz' as initrd image
> >>configuring 128 MB of virtual RAM
> >>selected cobd0 for hda1, mapping to '\??\c:\coLinux\smallroot'
> >>selected cobd1 for hdb1, mapping to '\??\c:\coLinux\swap_device'
> >>configured PCAP bridge at 'Lan' device as eth0
> >>MAC address: auto generated
> >>kernel boot parameters: 'root=/dev/hda1'
> >>PID: 3984
> >>colinux: launching console
> >>colinux: booting
> >>colinux: Linux VM terminated
> >>colinux: BUG at arch/i386/mm/init.c:87
> 
> First boot without initrd and without swapfile.
> Some previous errors in MM was sometimes only on specific memory
> size. Try some more or less memory (mem=120).
>
> You can also boot a ramdisk only:
> colinux-daemon.exe mem=128 kernel=vmlinux initrd=initrd.gz root=/dev/ram0

I tried as well
  colinux-daemon.exe mem=500 kernel=vmlinux hda1=smallroot root=/dev/hda1
as well as 
  colinux-daemon.exe mem=50 kernel=vmlinux initrd=initrd.gz root=/dev/ram0
with different memory sizes but the error was always the same.

> Is failed function new in kernel 2.6.15?  If yes, check additional
> things in other functions, how they handle pseudo physical ram.
> Check the difference in file init.c between  
> kernel 2.6.12 and 2.6.15

No, the function is not new. In fact it has not changed during the two
versions. Unfortunately my kernel hacking skills are fairly weak and
so I'm not really sure where pseudo physical ram is being
handled. Anyway, I'll try and have a look at it.

Regards!
-Andreas