Re: Porting LinuxPPC

[Ralph Blach <rcblach@raleigh.ibm.com>]

Daris,

I disagree.  I beleive that the Kernel load simply must be above
0x80000000.
And I you want to put a lot of memory in one of these systems, it had
better be movable down from 0xc0000000.  On the IBM405gp and I assume,
the 8xx systems, all of real memory is pinned
into a to TLB's.  Well, if one can't move the memory base down the the
real memory available is from 0xc0000000 to whereve the IO starts.  In
the IBM405gp, this is at 0xe80000000.

Chip




Daris A Nevil wrote:
>
> Graham,
>
> I fell vicitm to the same mistake when I was writing boot code
> for MPC850.  I think adding that to the HOWTO is a great
> idea.  It would have save me 2 or 3 weeks of effort :-).
>
> Daris Nevil
> SiSIC Inc/SNMC
>
> Graham Stoney wrote:
>
> > Hi Daniel,
> >
> > Daniel Wu writes:
> > > I compiled the mpc8xx-2.2.13 sources using the attached configuration for a
> > > MPC860T based board with 32MB of RAM. The output vmlinux file was downloaded
> > > to the target using a BDM debugger. I changed the KERNELLOAD and KERNELBASE
> > > to 1000000 instead of c0000000 so that the code is located at the start of
> > > the 16MB region.
> >
> > To quote Grant Erickson in
> > http://lists.linuxppc.org/listarcs/linuxppc-embedded/200001/msg00123.html:
> >
> > "You don't ever want to change KERNELLOAD or KERNELBASE, otherwise the
> > virtual memory and MMU code will all break."
> >
> > This seems to be a common error, so I might add mention of it in the HOWTO.
> >
> > Regards,
> > Graham
> >
>
> --
>
> Daris A Nevil
> SiSIC, Inc., f.k.a. Simple Network Magic Corporation
> 401 Kentucky Lane
> McKinney, TX 75069
> 214-793-7757
> dnevil@snmc.com
>

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