Re: [PATCH] VM routine fixes

[Andrew Morton <akpm@osdl.org>]

David Howells <dhowells@redhat.com> wrote:
>
> Compound pages seem to be in some way tied to the TLB entry coverage sizes
>  available (for hugetlb), so it's not obvious that it's permitted to have
>  compound pages not of these sizes, and as I need to allocate arbitrary
>  sizes...

We've considered enabling compound pages permanently.  We thought sparc64
might want that, and it simplifies coverage testing.  But the
conditionality has been left in for now as a microoptimisation.

>  If I am correct about this, then the !MMU problem would still exist - just
>  with adjacent sets of compound pages rather than adjacent sets of pages.

Why _does_ !CONFIG_MMU futz around with page counts in such weird ways
anyway?  Why does it have requirements for higher-order pages which differ
from !CONFIG_MMU?

If someone could explain the reasoning behind the current code, and the FRV
enhancements then perhaps we could work something out.

>  Compound pages might be nice, but they're overkill.

I don't expect they have significant performance overhead, because they'll
only add cycles for higher-order pages.  They're rare, and are already
rather inefficient.