Re: [PATCH 2/2] mm/hmm: make full use of walk_page_range()

From: Christoph Hellwig
Date: Wed Jul 24 2019 - 02:51:50 EST

On Tue, Jul 23, 2019 at 04:30:16PM -0700, Ralph Campbell wrote:
> hmm_range_snapshot() and hmm_range_fault() both call find_vma() and
> walk_page_range() in a loop. This is unnecessary duplication since
> walk_page_range() calls find_vma() in a loop already.
> Simplify hmm_range_snapshot() and hmm_range_fault() by defining a
> walk_test() callback function to filter unhandled vmas.

I like the approach a lot!

But we really need to sort out the duplication between hmm_range_fault
and hmm_range_snapshot first, as they are basically the same code. I
have patches here:

That being said we don't really have any users for the snapshot mode
or non-blocking faults, and I don't see any in the immediate pipeline
either. It might actually be a better idea to just kill that stuff off
for now until we have a user, as code without users is per definition
untested and will just bitrot and break.

> + const unsigned long device_vma = VM_IO | VM_PFNMAP | VM_MIXEDMAP;
> + struct hmm_vma_walk *hmm_vma_walk = walk->private;
> + struct hmm_range *range = hmm_vma_walk->range;
> + struct vm_area_struct *vma = walk->vma;
> +
> + /* If range is no longer valid, force retry. */
> + if (!range->valid)
> + return -EBUSY;
> +
> + if (vma->vm_flags & device_vma)

Can we just kill off this odd device_vma variable?

if (vma->vm_flags & (VM_IO | VM_PFNMAP | VM_MIXEDMAP))

and maybe add a comment on why we are skipping them (because they
don't have struct page backing I guess..)