KERNELBASE at 0xa0000000

Hi,
  anybody knows why in the MCPN765 port KERNELBASE is at 0xa0000000 instead
of 0xc0000000 ?
Thanks.

Stefano.


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

Stefano Coluccini wrote:

Hi,
  anybody knows why in the MCPN765 port KERNELBASE is at 0xa0000000 instead
of 0xc0000000 ?

Yep.  It was a quick hack to get a system with 1GByte of memory
working.  It works fine with all other memory configurations as well.

I don't see any reason anymore (once Matt Porter dumps all of the
PReP stuff :-) that we shouldn't be able to move all kernel base
addresses to 0x80000000 or to allow 3 G user spaces.  On PowerPC,
we are simply tossing away 1G of address space that should be
utilized to make our life easier.


	-- Dan

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

Stefano Coluccini wrote:

quoted

Hi,
  anybody knows why in the MCPN765 port KERNELBASE is at

0xa0000000 instead

quoted

of 0xc0000000 ?

Yep.  It was a quick hack to get a system with 1GByte of memory
working.  It works fine with all other memory configurations as well.

Thanks a lot for the reply.
I have hoped it was more platform-dependent because I have a MVME5100 (that
is, I think, a MCPN765 with a PCI-VME bridge) and with KERNELBASE at
0xa0000000 doesn't work (in head.S the kernel doesn't recognize that it is
running at 0 and it try to relocate itself).
Do you think that this is a warning of a more complicated problem ?


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