2003-07-09 09:52:33

by gigag

Subject: Known problems for 3.5/0.5 virtual space split???

I wonder if there are any known problems for applying
2.4.21rc8aa1 patch from Andrea?

I'm trying to boot a Xeon 2.4 box with 4GB of RAM, IDE CD and
Adaptec AIC7XXX SCSI devices.

I compile the kernel with 1/3 split and it boots OK. When I simply
change the split to 3.5/0.5, it gets stuck during the boot not being
able to mount root partition.

Any ideas? Any other working kernel/patch configurations for h/w
like I have?

Thanks
Giga

2003-07-09 14:05:51

by Martin J. Bligh

Subject: Re: Known problems for 3.5/0.5 virtual space split???

> I wonder if there are any known problems for applying
> 2.4.21rc8aa1 patch from Andrea?
>
> I'm trying to boot a Xeon 2.4 box with 4GB of RAM, IDE CD and
> Adaptec AIC7XXX SCSI devices.
>
> I compile the kernel with 1/3 split and it boots OK. When I simply
> change the split to 3.5/0.5, it gets stuck during the boot not being
> able to mount root partition.
>
> Any ideas? Any other working kernel/patch configurations for h/w
> like I have?

Odd - should work. Post a boot log (use a serial console).

Also, if you can try 2.5-mjb tree, that should have a forward port of
the same concept ... does it work there?

M.

2003-07-10 13:58:28

by gigag

Subject: Re: Known problems for 3.5/0.5 virtual space split???

>Also, if you can try 2.5-mjb tree, that should have a forward port of
>the same concept ... does it work there?
Yes, it worked with 2.5.70 kernel with the proper mjb patch. I had
to change root=LABEL=/ token to root=/dev/sda2 one in order to
make the system boot from SCSI disk. Also, I compiled AIC7XXX
drivers statically into the kernel.

I tried to do the same trick for 2.4.21 kernel, but it seemed not to
help. I didn't play with it for too long though. I indend to experiment
with 2.4.21 kernel in about a week after I finish with 2.5.70, which
is currently working for me.

Thanks a lot for the help
Giga