Re: [PATCH 0/7] [RFC] hugetlb: pagetable_operations API (V2)

[Hugh Dickins <hugh@veritas.com>]

On Wed, 21 Mar 2007, William Lee Irwin III wrote:
> 
> What sort of uglier stuff are you concerned about this enabling? My
> wild guess is precisely the prospective users in my queue of features
> to implement that I've neglected on account of the lack of such an
> interface. It might be a good idea for me to take whatever distaste for
> them exists into account before belting out the code for them.

I'll probably hate it whatever it is, don't worry about me ;)  I'm
rather weary of all the grand mm schemes people are toting these days.

I guess my distaste is for letting a door open, for all kinds of
drivers (or odd corners of architectures etc) to take control of
the mm in ways we've never anticipated, and which we'll forever
after be stumbling over.

It would be a good idea for you to belt out the code for a few of them,
to give everyone else an idea of what's to be let in through this door:
I haven't the faintest idea what's envisaged.

Seeing what you have, maybe everyone will react that of course
Adam's page table ops are the way to go, or maybe the reverse.

I would be rather surprised if the adhoc collection of divergences
you found necessary for hugetlb turn out to have a fundamental
applicability.  Perhaps we'll need subvert_the_core_ops.

Hugh

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>