CONFIG_VMSPLIT_3G in RHEL5?

[Jarod Wilson]

On Wed, 2008-07-16 at 13:11 -0400, Michael ODonnell wrote:
> 
> >excerpt from arch/i386/Kconfig:
> >--
> >choice
> >        depends on EXPERIMENTAL && !X86_PAE
> >        prompt "Memory split" if EMBEDDED
> >        default VMSPLIT_3G
> >--
> >
> >Note the !X86_PAE part. At least above, you mentioned looking in a PAE
> >kernel config. :)
> 
> The default kernel appears to have PAE configured so that's the
> kernel whose config I was asked to look at.
> 
> I haven't thought this through but I can't think of anything
> inherent in the use of PAE (at least in principle, at the
> CPU/MMU level) that would make it impossible for Kernel/User to
> continue sharing page tables - would you know why the VMSPLIT
> option is only present when !X86_PAE ?


X86_PAE depends on HIGHMEM64G, which isn't compatible with the VMSPLIT
options, which all appear to require HIGHMEM4G.

Nb: per the above, I did read the i686 config slightly wrong before, the
VMSPLIT stuff should be compatible with HIGHMEM4G, but for some reason,
even in the RHEL5 i686 kernel config, I don't get any VMSPLIT options
exposed at the moment (searching shows they're known, but I dunno atm
how to get them actually twiddleable).




-- 
Jarod Wilson
jarod at wilsonet.com