Filtering bits in set_pte_at()

[David Gibson]

On Tue, Nov 03, 2009 at 09:19:27AM +1100, Benjamin Herrenschmidt wrote:
> On Mon, 2009-11-02 at 13:27 +0000, Hugh Dickins wrote:
> > On Sat, 31 Oct 2009, Benjamin Herrenschmidt wrote:
> > 
> > > Hi folks !
> > > 
> > > So I have a little problem on powerpc ... :-)
> > 
> > Thanks a lot for running this by us.
> 
> Heh, I though you may have been bored :-)
> 
> > I've not looked to see if there are more such issues in arch/powerpc
> > itself, but those instances you mention are the only ones I managed
> > to find: uses of update_mmu_cache() and that hugetlb_cow() one.
> 
> Right, that's all I spotted so far
> 
> > The hugetlb_cow() one involves not set_pte_at() but set_huge_pte_at(),
> > so you'd want to change that too?  And presumably set_pte_at_notify()?
> > It all seems a lot of tedium, when so very few places are interested
> > in the pte after they've set it.
> 
> We need to change set_huge_pte_at() too. Currently, David fixed the
> problem in a local tree by making hugetlb_cow() re-read the PTE . 

Well, actually I have another cleanup patch in the queue which makes
set_huge_pte_at() equal to set_pte_at() on powerpc, and I was using
that on the tree where this problem became apparent.

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson