Is ARCH_HAS_MMU going away?
Robin Getz
rgetz at blackfin.uclinux.org
Wed Aug 5 16:32:57 UTC 2009
On Sun 2 Aug 2009 22:54, Rob Landley pondered:
> On Sunday 02 August 2009 19:58:41 Robin Getz wrote:
> > On Mon 27 Jul 2009 21:36, Rob Landley pondered:
> > > I don't understand the difference between ARCH_HAS_MMU and ARCH_USE_MMU.
> > > (For a kernel, sure. For a C library, not so much.)
> >
> > I'm not sure if you are saying they are redundant (which I understand),
> > and we only need one, or if you think they both are unnecessary?
>
> I think they're redundant.
>
> Having ARCH_USE_MMU, and nommu support in general, makes sense for embedded
> systems. But I'm unaware of any instance where userspace cares about the
> difference between having and using an MMU, and I don't see it expressed
> anywhere in the code either.
I think that is correct - the C library only needs to know if it should use
fork/sbrk/brk/etc - it shouldn't care if the hardware has/doesn't have the
hardware that actually supports it....
More information about the uClibc
mailing list