On Fri, May 30, 2025 at 11:37:21AM +0200, David Hildenbrand wrote:

On 29.05.25 08:32, Alistair Popple wrote:

quoted

Currently dax is the only user of pmd and pud mapped ZONE_DEVICE
pages. Therefore page walkers that want to exclude DAX pages can check
pmd_devmap or pud_devmap. However soon dax will no longer set PFN_DEV,
meaning dax pages are mapped as normal pages.

Ensure page walkers that currently use pXd_devmap to skip DAX pages
continue to do so by adding explicit checks of the VMA instead.

Signed-off-by: Alistair Popple <apopple@nvidia.com>
---
  fs/userfaultfd.c | 2 +-
  mm/hmm.c         | 2 +-
  mm/userfaultfd.c | 2 +-
  3 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/fs/userfaultfd.c b/fs/userfaultfd.c
index 22f4bf9..de671d3 100644
--- a/fs/userfaultfd.c
+++ b/fs/userfaultfd.c

@@ -304,7 +304,7 @@ static inline bool userfaultfd_must_wait(struct userfaultfd_ctx *ctx,
  		goto out;
  	ret = false;
-	if (!pmd_present(_pmd) || pmd_devmap(_pmd))
+	if (!pmd_present(_pmd) || vma_is_dax(vmf->vma))
  		goto out;
  	if (pmd_trans_huge(_pmd)) {

diff --git a/mm/hmm.c b/mm/hmm.c
index 082f7b7..db12c0a 100644
--- a/mm/hmm.c
+++ b/mm/hmm.c

@@ -429,7 +429,7 @@ static int hmm_vma_walk_pud(pud_t *pudp, unsigned long start, unsigned long end,
  		return hmm_vma_walk_hole(start, end, -1, walk);
  	}
-	if (pud_leaf(pud) && pud_devmap(pud)) {
+	if (pud_leaf(pud) && vma_is_dax(walk->vma)) {
  		unsigned long i, npages, pfn;
  		unsigned int required_fault;
  		unsigned long *hmm_pfns;

diff --git a/mm/userfaultfd.c b/mm/userfaultfd.c
index e0db855..133f750 100644
--- a/mm/userfaultfd.c
+++ b/mm/userfaultfd.c

@@ -1791,7 +1791,7 @@ ssize_t move_pages(struct userfaultfd_ctx *ctx, unsigned long dst_start,
  		ptl = pmd_trans_huge_lock(src_pmd, src_vma);
  		if (ptl) {
-			if (pmd_devmap(*src_pmd)) {
+			if (vma_is_dax(src_vma)) {
  				spin_unlock(ptl);
  				err = -ENOENT;
  				break;

I assume we could also just refuse dax folios, right?

Yep.

If we decide to check VMAs, we should probably check earlier.

Ok, that makes sense.
 

But I wonder, what about anonymous non-dax pages in COW mappings? Is it
possible? Not supported?

You mean other non-dax ZONE_DEVICE pages? Currently not possible, because
non-dax ZONE_DEVICE pages can't be pmd mapped (although it is a future
enhancement I'd like to make).

If supported, checking the actual folio would be the right thing to do.

-- 
Cheers,

David / dhildenb