On Tue, Jul 23, 2019 at 11:57:52PM +0000, Jason Gunthorpe wrote:
> diff --git a/mm/hmm.c b/mm/hmm.c
> index 16b6731a34db79..3d8cdfb67a6ab8 100644
> --- a/mm/hmm.c
> +++ b/mm/hmm.c
> @@ -285,8 +285,9 @@ struct hmm_vma_walk {
> struct hmm_range *range;
> struct dev_pagemap *pgmap;
> unsigned long last;
> - bool fault;
> - bool block;
> + bool fault : 1;
> + bool block : 1;
> + bool hugetlb : 1;
I don't think we should even keep these bools around. I have something
like this hiding in a branche, which properly cleans much of this up:
http://git.infradead.org/users/hch/misc.git/shortlog/refs/heads/hmm-dma-cleanup
Notably:
http://git.infradead.org/users/hch/misc.git/commitdiff/2abdc0ac8f9f32149246957121ebccbe5c0a729d
http://git.infradead.org/users/hch/misc.git/commitdiff/a34ccd30ee8a8a3111d9e91711c12901ed7dea74
http://git.infradead.org/users/hch/misc.git/commitdiff/81f442ebac7170815af7770a1efa9c4ab662137e
This doesn't go all the way yet - the page_walk infrastructure is
built around the idea of doing its own vma lookups, and we should
eventually kill the lookup inside hmm entirely.
