On 09.08.23 15:09, Matthew Wilcox wrote:
On Wed, Aug 09, 2023 at 09:35:01AM +0200, David Hildenbrand wrote:
On 08.08.23 19:09, Matthew Wilcox wrote:
Sid was incautious enough to say he'd like to take on fixing
walk_page_range() so that hugetlb isn't treated specially. This is
going to subject him to one of my rants, so I thought I'd share with
everyone before we meet to talk about it later today.
Are we only talking about walk_page_range() or also walk_page_range_vma() /
walk_page_vma() ?
I tend to like the VMA variants ...
We're talking about getting rid of mm_walk_ops. There aren't exactly a
lot of callers of either of those functions -- 4 of walk_page_vma()
and 1 of walk_page_range_vma().
Okay, I see. For some use cases, it's probably sufficient to walk folios.
But we do have some advanced users like fs/proc/task_mmu.c.
[I do have two more follow_page() -> walk_page_range_vma() conversions
lying around here; they want to also know if a page is mapped writable]
--
Cheers,
David / dhildenb
