[PATCH v2 0/2] mm: stop calling pmd_folio() on special PMDs

[PATCH v2 0/2] mm: stop calling pmd_folio() on special PMDs

[Gregory Price]

Two page table walkers resolve the folio behind a PMD with pmd_folio(),
which is only valid for a PMD mapping a refcounted struct page:

  madvise_cold_or_pageout_pte_range()   mm/madvise.c
  queue_folios_pmd()                    mm/mempolicy.c

vmf_insert_pfn_pmd() installs special PMDs holding a raw pfn that need not
have a memmap entry at all.  Both walkers can reach one and fault on the
first folio field read.  The PTE halves of both already use
vm_normal_folio(); these two patches make the PMD halves match.

The four callers of vmf_insert_pfn_pmd(), and which walker each reaches:

  drivers/vfio/pci/vfio_pci_core.c        VM_PFNMAP     mempolicy
  drivers/gpu/drm/drm_gem_shmem_helper.c  VM_PFNMAP     mempolicy
  drivers/gpu/drm/panthor/panthor_gem.c   VM_PFNMAP     mempolicy
  drivers/hv/mshv_vtl_main.c              VM_MIXEDMAP   both

can_madv_lru_vma() rejects VM_PFNMAP, so only mshv_vtl_low reaches the
madvise walker, and that needs CAP_SYS_ADMIN.  queue_pages_walk_ops
supplies its own ->test_walk, so walk_page_test()'s generic VM_PFNMAP skip
never runs and vfio-pci is reachable by any process holding the device fd.
Hence the different stable tags.

One behaviour change: mbind(MPOL_MF_STRICT) over a PMD mapped VM_PFNMAP
region now returns 0 rather than -EIO.  The PTE loop already returned 0
there.  drm_gem_shmem and panthor are where this is observable, since they
PMD map pages that do have a memmap entry and so never faulted.

Reproducer
==========

No hardware needed.  An out of tree module stands in for the drivers above:
three misc devices, each with a ->huge_fault calling vmf_insert_pfn_pmd(),
plus VM_HUGEPAGE so the fault path takes the PMD branch.

  /dev/pmdspec_mixed    VM_MIXEDMAP, pfn at the 1 TiB mark, no memmap
  /dev/pmdspec_pfnmap   VM_PFNMAP,   pfn at the 1 TiB mark, no memmap
  /dev/pmdspec_real     VM_PFNMAP,   real alloc_pages(PMD_ORDER) on node 0

Userspace maps the device into a PMD aligned window, reads one byte to
fault the PMD in, checks a module parameter to confirm it went in, then
issues the operation.

  vng --run <bzImage> --user root --memory 4G --verbose \
      --append "numa=fake=2" \
      --exec "insmod pmdspec.ko && ./pmdspec_test <subtest>"

numa=fake=2 gives a node 1 to bind to; the module allocates its real page
on node 0, which is what makes queue_folio_required() true.

  subtest        operation                              parent   series
  --------------------------------------------------------------------
  madv_cold      madvise(MADV_COLD)                     oops     ret=0
  madv_pageout   madvise(MADV_PAGEOUT)                  oops     ret=0
  mbind_mixed    mbind(MPOL_BIND, n1, MPOL_MF_MOVE)     oops     ret=0
  mbind_pfnmap   mbind(MPOL_BIND, n1, MPOL_MF_STRICT)   oops     ret=0
  mbind_real     mbind(MPOL_BIND, n1, MPOL_MF_STRICT)   -EIO     ret=0

Two things the table shows that are easy to miss in the code:

  - mbind_mixed passes only MPOL_MF_MOVE.  MPOL_MF_STRICT is not needed for
    a VM_MIXEDMAP vma: walk_page_test() only skips VM_PFNMAP, and
    vma_migratable() is true for VM_MIXEDMAP.

  - mbind_real demonstrates the user visible change (-EIO -> 0)

Reported-by: sashiko-bot <sashiko-bot@kernel.org>
Closes: https://sashiko.dev/#/patchset/20260817220810.1175596-1-gourry%40gourry.net
Assisted-by: LLM

---
v2: improvements from David

Gregory Price (2):
  mm/mempolicy: use vm_normal_folio_pmd() in queue_folios_pmd()
  mm/madvise: use vm_normal_folio_pmd() in cold/pageout PMD range

 mm/madvise.c   |  7 +++----
 mm/mempolicy.c | 16 +++++++++-------
 2 files changed, 12 insertions(+), 11 deletions(-)

-- 
2.55.0

[PATCH v2 1/2] mm/mempolicy: use vm_normal_folio_pmd() in queue_folios_pmd()

[Gregory Price]

mmap a VM_PFNMAP region whose ->huge_fault installs a PMD through
vmf_insert_pfn_pmd() - a vfio-pci MMIO BAR does this - then

	mbind(p, len, MPOL_BIND, &mask, maxnode, MPOL_MF_STRICT);

With a stand-in module for the driver:

  BUG: unable to handle page fault for address: fffff96dc0000008
  RIP: 0010:queue_folios_pte_range+0xaf/0x440
   walk_pgd_range+0x52b/0xaf0
   __walk_page_range+0x6a/0x1d0
   walk_page_range_mm_unsafe+0x193/0x230
   queue_pages_range+0x64/0xa0
   do_mbind+0x25e/0x640

queue_folios_pmd(), inlined above, calls pmd_folio() on that PMD.  The pfn
is raw MMIO with no memmap entry, so the folio lands in unpopulated
vmemmap.  Neither guard stops the walk:

  walk_page_test()         skips VM_PFNMAP, but queue_pages_walk_ops
                           supplies ->test_walk, so it never runs
  queue_pages_test_walk()  honours vma_migratable(), but only while
                           MPOL_MF_STRICT is clear

A VM_MIXEDMAP vma needs neither flag, being vma_migratable(), so plain
mbind(MPOL_MF_MOVE) reaches this too - and there the bad folio carries on
into migrate_folio_add() and folio_isolate_lru().  mshv_vtl_low is such a
mapping.

Use vm_normal_folio_pmd() and skip on NULL, as the PTE loop in
queue_folios_pte_range() already does with vm_normal_folio().  On the NULL
path, retain ACTION_CONTINUE handling for the huge zero PMD.

mbind(MPOL_MF_STRICT) over a PMD mapped VM_PFNMAP region now returns 0
rather than -EIO.  The PTE loop already returned 0 there.

Fixes: 3c8e44c9b369 ("mm: mark special bits for huge pfn mappings when inject")
Reported-by: sashiko-bot <sashiko-bot@kernel.org>
Closes: https://sashiko.dev/#/patchset/20260817220810.1175596-1-gourry%40gourry.net
Cc: stable@vger.kernel.org
Assisted-by: LLM
Signed-off-by: Gregory Price (Meta) <gourry@gourry.net>
Acked-by: David Hildenbrand (Arm) <david@kernel.org>
---
 mm/mempolicy.c | 16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)

diff --git a/mm/mempolicy.c b/mm/mempolicy.c
index 2ad0a5f18280..8fc8a975657e 100644
--- a/mm/mempolicy.c
+++ b/mm/mempolicy.c
@@ -667,7 +667,8 @@ static inline bool queue_folio_required(struct folio *folio,
 	return node_isset(nid, *qp->nmask) == !(flags & MPOL_MF_INVERT);
 }
 
-static void queue_folios_pmd(pmd_t *pmd, struct mm_walk *walk)
+static void queue_folios_pmd(pmd_t *pmd, unsigned long addr,
+			     struct mm_walk *walk)
 {
 	struct folio *folio;
 	struct queue_pages *qp = walk->private;
@@ -678,13 +679,14 @@ static void queue_folios_pmd(pmd_t *pmd, struct mm_walk *walk)
 			qp->nr_failed++;
 		return;
 	}
-	folio = pmd_folio(pmdval);
-	if (folio_is_zone_device(folio))
-		return;
-	if (is_huge_zero_folio(folio)) {
-		walk->action = ACTION_CONTINUE;
+	folio = vm_normal_folio_pmd(walk->vma, addr, pmdval);
+	if (!folio) {
+		if (is_huge_zero_pmd(pmdval))
+			walk->action = ACTION_CONTINUE;
 		return;
 	}
+	if (folio_is_zone_device(folio))
+		return;
 	if (!queue_folio_required(folio, qp))
 		return;
 	if (!(qp->flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL)) ||
@@ -717,7 +719,7 @@ static int queue_folios_pte_range(pmd_t *pmd, unsigned long addr,
 
 	ptl = pmd_trans_huge_lock(pmd, vma);
 	if (ptl) {
-		queue_folios_pmd(pmd, walk);
+		queue_folios_pmd(pmd, addr, walk);
 		spin_unlock(ptl);
 		goto out;
 	}
-- 
2.55.0

[PATCH v2 2/2] mm/madvise: use vm_normal_folio_pmd() in cold/pageout PMD range

[Gregory Price]

mmap a VM_MIXEDMAP region whose ->huge_fault installs a PMD through
vmf_insert_pfn_pmd() - mshv_vtl_low does this, and needs CAP_SYS_ADMIN
to open - then:

  madvise(p, PMD_SIZE, MADV_PAGEOUT);

With a stand-in module for the driver:

  BUG: unable to handle page fault for address: fffff587c0000008
  RIP: 0010:madvise_cold_or_pageout_pte_range+0x410/0x9b0
   walk_pgd_range+0x52b/0xaf0
   __walk_page_range+0x6a/0x1d0
   walk_page_range_vma_unsafe+0x8e/0x120
   madvise_pageout+0xb2/0x180
   madvise_vma_behavior+0x46b/0xa90
   do_madvise+0x108/0x190
   __x64_sys_madvise+0x26/0x30

Nothing validates the pfn on the way in:

  can_madv_lru_vma()     rejects VM_PFNMAP, but not VM_MIXEDMAP
  can_fault()            *pfn = vmf->pgoff & ~(mask >> PAGE_SHIFT);
  vmf_insert_pfn_pmd()   no pfn_valid() check
  pmd_folio()            pfn_to_page() -> unpopulated vmemmap

Even with a valid pfn the path is wrong.  The mapping carries no rmap, so
folio_maybe_mapped_shared() sees mapcount 0, and the walker goes on to
folio_deactivate(), or folio_isolate_lru() plus reclaim_pages(), against a
folio this mapping does not own.

Use vm_normal_folio_pmd() and skip on NULL, as the PTE half of this same
walker already does with vm_normal_folio().  This also filters the huge
zero PMD, so its separate check is no longer needed.

Fixes: 3c8e44c9b369 ("mm: mark special bits for huge pfn mappings when inject")
Reported-by: sashiko-bot <sashiko-bot@kernel.org>
Closes: https://sashiko.dev/#/patchset/20260817220810.1175596-1-gourry%40gourry.net
Cc: stable@vger.kernel.org # v6.19+
Assisted-by: LLM
Signed-off-by: Gregory Price (Meta) <gourry@gourry.net>
---
 mm/madvise.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/mm/madvise.c b/mm/madvise.c
index f75a9d139980..fbb72ab49aa6 100644
--- a/mm/madvise.c
+++ b/mm/madvise.c
@@ -395,16 +395,15 @@ static int madvise_cold_or_pageout_pte_range(pmd_t *pmd,
 			return 0;
 
 		orig_pmd = *pmd;
-		if (is_huge_zero_pmd(orig_pmd))
-			goto huge_unlock;
-
 		if (unlikely(!pmd_present(orig_pmd))) {
 			VM_WARN_ON_ONCE(!pmd_is_migration_entry(orig_pmd) &&
 					!pmd_is_device_private_entry(orig_pmd));
 			goto huge_unlock;
 		}
 
-		folio = pmd_folio(orig_pmd);
+		folio = vm_normal_folio_pmd(vma, addr, orig_pmd);
+		if (!folio)
+			goto huge_unlock;
 
 		if (folio_is_zone_device(folio))
 			goto huge_unlock;
-- 
2.55.0

Re: [PATCH v2 0/2] mm: stop calling pmd_folio() on special PMDs

[Andrew Morton]

On Fri, 11 Sep 2026 23:48:31 -0400 Gregory Price <gourry@gourry.net> wrote:

> Reproducer
> ==========
> 
> No hardware needed.  An out of tree module stands in for the drivers above:
> three misc devices, each with a ->huge_fault calling vmf_insert_pfn_pmd(),
> plus VM_HUGEPAGE so the fault path takes the PMD branch.

I take it there's no known way of triggering this in current mainline?

Re: [PATCH v2 0/2] mm: stop calling pmd_folio() on special PMDs

[Gregory Price]

On Fri, Sep 11, 2026 at 09:25:03PM -0700, Andrew Morton wrote:
> On Fri, 11 Sep 2026 23:48:31 -0400 Gregory Price <gourry@gourry.net> wrote:
> 
> > Reproducer
> > ==========
> > 
> > No hardware needed.  An out of tree module stands in for the drivers above:
> > three misc devices, each with a ->huge_fault calling vmf_insert_pfn_pmd(),
> > plus VM_HUGEPAGE so the fault path takes the PMD branch.
> 
> I take it there's no known way of triggering this in current mainline?

As far as I can tell there are no in-tree modules that reach this state,
at least not obviously.  

~Gregory

Re: [PATCH v2 0/2] mm: stop calling pmd_folio() on special PMDs

[Gregory Price]

On Sat, Sep 12, 2026 at 06:35:08AM -0400, Gregory Price wrote:
> On Fri, Sep 11, 2026 at 09:25:03PM -0700, Andrew Morton wrote:
> > On Fri, 11 Sep 2026 23:48:31 -0400 Gregory Price <gourry@gourry.net> wrote:
> > 
> > > Reproducer
> > > ==========
> > > 
> > > No hardware needed.  An out of tree module stands in for the drivers above:
> > > three misc devices, each with a ->huge_fault calling vmf_insert_pfn_pmd(),
> > > plus VM_HUGEPAGE so the fault path takes the PMD branch.
> > 
> > I take it there's no known way of triggering this in current mainline?
> 
> As far as I can tell there are no in-tree modules that reach this state,
> at least not obviously.  
> 

Ah - sashiko did report a similar issue in fs/proc/task_mmu.c though, so
I'm going to take a run at doing a larger detection run for this
problem.  These patches are good to go though.

~Gregory

Re: [PATCH v2 1/2] mm/mempolicy: use vm_normal_folio_pmd() in queue_folios_pmd()

[Zi Yan]

On Fri Sep 11, 2026 at 11:48 PM EDT, Gregory Price wrote:
> mmap a VM_PFNMAP region whose ->huge_fault installs a PMD through
> vmf_insert_pfn_pmd() - a vfio-pci MMIO BAR does this - then
>
> 	mbind(p, len, MPOL_BIND, &mask, maxnode, MPOL_MF_STRICT);
>
> With a stand-in module for the driver:
>
>   BUG: unable to handle page fault for address: fffff96dc0000008
>   RIP: 0010:queue_folios_pte_range+0xaf/0x440
>    walk_pgd_range+0x52b/0xaf0
>    __walk_page_range+0x6a/0x1d0
>    walk_page_range_mm_unsafe+0x193/0x230
>    queue_pages_range+0x64/0xa0
>    do_mbind+0x25e/0x640
>
> queue_folios_pmd(), inlined above, calls pmd_folio() on that PMD.  The pfn
> is raw MMIO with no memmap entry, so the folio lands in unpopulated
> vmemmap.  Neither guard stops the walk:
>
>   walk_page_test()         skips VM_PFNMAP, but queue_pages_walk_ops
>                            supplies ->test_walk, so it never runs
>   queue_pages_test_walk()  honours vma_migratable(), but only while
>                            MPOL_MF_STRICT is clear
>
> A VM_MIXEDMAP vma needs neither flag, being vma_migratable(), so plain
> mbind(MPOL_MF_MOVE) reaches this too - and there the bad folio carries on
> into migrate_folio_add() and folio_isolate_lru().  mshv_vtl_low is such a
> mapping.
>
> Use vm_normal_folio_pmd() and skip on NULL, as the PTE loop in
> queue_folios_pte_range() already does with vm_normal_folio().  On the NULL
> path, retain ACTION_CONTINUE handling for the huge zero PMD.
>
> mbind(MPOL_MF_STRICT) over a PMD mapped VM_PFNMAP region now returns 0
> rather than -EIO.  The PTE loop already returned 0 there.
>
> Fixes: 3c8e44c9b369 ("mm: mark special bits for huge pfn mappings when inject")
> Reported-by: sashiko-bot <sashiko-bot@kernel.org>
> Closes: https://sashiko.dev/#/patchset/20260817220810.1175596-1-gourry%40gourry.net
> Cc: stable@vger.kernel.org
> Assisted-by: LLM
> Signed-off-by: Gregory Price (Meta) <gourry@gourry.net>
> Acked-by: David Hildenbrand (Arm) <david@kernel.org>
> ---
>  mm/mempolicy.c | 16 +++++++++-------
>  1 file changed, 9 insertions(+), 7 deletions(-)
>
LGTM.

Reviewed-by: Zi Yan <ziy@nvidia.com>

-- 
Best Regards,
Yan, Zi

Re: [PATCH v2 2/2] mm/madvise: use vm_normal_folio_pmd() in cold/pageout PMD range

[Zi Yan]

On Fri Sep 11, 2026 at 11:48 PM EDT, Gregory Price wrote:
> mmap a VM_MIXEDMAP region whose ->huge_fault installs a PMD through
> vmf_insert_pfn_pmd() - mshv_vtl_low does this, and needs CAP_SYS_ADMIN
> to open - then:
>
>   madvise(p, PMD_SIZE, MADV_PAGEOUT);
>
> With a stand-in module for the driver:
>
>   BUG: unable to handle page fault for address: fffff587c0000008
>   RIP: 0010:madvise_cold_or_pageout_pte_range+0x410/0x9b0
>    walk_pgd_range+0x52b/0xaf0
>    __walk_page_range+0x6a/0x1d0
>    walk_page_range_vma_unsafe+0x8e/0x120
>    madvise_pageout+0xb2/0x180
>    madvise_vma_behavior+0x46b/0xa90
>    do_madvise+0x108/0x190
>    __x64_sys_madvise+0x26/0x30
>
> Nothing validates the pfn on the way in:
>
>   can_madv_lru_vma()     rejects VM_PFNMAP, but not VM_MIXEDMAP
>   can_fault()            *pfn = vmf->pgoff & ~(mask >> PAGE_SHIFT);
>   vmf_insert_pfn_pmd()   no pfn_valid() check
>   pmd_folio()            pfn_to_page() -> unpopulated vmemmap
>
> Even with a valid pfn the path is wrong.  The mapping carries no rmap, so
> folio_maybe_mapped_shared() sees mapcount 0, and the walker goes on to
> folio_deactivate(), or folio_isolate_lru() plus reclaim_pages(), against a
> folio this mapping does not own.
>
> Use vm_normal_folio_pmd() and skip on NULL, as the PTE half of this same
> walker already does with vm_normal_folio().  This also filters the huge
> zero PMD, so its separate check is no longer needed.
>
> Fixes: 3c8e44c9b369 ("mm: mark special bits for huge pfn mappings when inject")
> Reported-by: sashiko-bot <sashiko-bot@kernel.org>
> Closes: https://sashiko.dev/#/patchset/20260817220810.1175596-1-gourry%40gourry.net
> Cc: stable@vger.kernel.org # v6.19+
> Assisted-by: LLM
> Signed-off-by: Gregory Price (Meta) <gourry@gourry.net>
> ---
>  mm/madvise.c | 7 +++----
>  1 file changed, 3 insertions(+), 4 deletions(-)
>
LGTM.

Reviewed-by: Zi Yan <ziy@nvidia.com>

-- 
Best Regards,
Yan, Zi

Re: [PATCH v2 2/2] mm/madvise: use vm_normal_folio_pmd() in cold/pageout PMD range

[David Hildenbrand (Arm)]

On 9/12/26 05:48, Gregory Price wrote:
> mmap a VM_MIXEDMAP region whose ->huge_fault installs a PMD through
> vmf_insert_pfn_pmd() - mshv_vtl_low does this, and needs CAP_SYS_ADMIN
> to open - then:
> 
>   madvise(p, PMD_SIZE, MADV_PAGEOUT);
> 
> With a stand-in module for the driver:
> 
>   BUG: unable to handle page fault for address: fffff587c0000008
>   RIP: 0010:madvise_cold_or_pageout_pte_range+0x410/0x9b0
>    walk_pgd_range+0x52b/0xaf0
>    __walk_page_range+0x6a/0x1d0
>    walk_page_range_vma_unsafe+0x8e/0x120
>    madvise_pageout+0xb2/0x180
>    madvise_vma_behavior+0x46b/0xa90
>    do_madvise+0x108/0x190
>    __x64_sys_madvise+0x26/0x30
> 
> Nothing validates the pfn on the way in:
> 
>   can_madv_lru_vma()     rejects VM_PFNMAP, but not VM_MIXEDMAP
>   can_fault()            *pfn = vmf->pgoff & ~(mask >> PAGE_SHIFT);
>   vmf_insert_pfn_pmd()   no pfn_valid() check
>   pmd_folio()            pfn_to_page() -> unpopulated vmemmap
> 
> Even with a valid pfn the path is wrong.  The mapping carries no rmap, so
> folio_maybe_mapped_shared() sees mapcount 0, and the walker goes on to
> folio_deactivate(), or folio_isolate_lru() plus reclaim_pages(), against a
> folio this mapping does not own.
> 
> Use vm_normal_folio_pmd() and skip on NULL, as the PTE half of this same
> walker already does with vm_normal_folio().  This also filters the huge
> zero PMD, so its separate check is no longer needed.
> 
> Fixes: 3c8e44c9b369 ("mm: mark special bits for huge pfn mappings when inject")
> Reported-by: sashiko-bot <sashiko-bot@kernel.org>
> Closes: https://sashiko.dev/#/patchset/20260817220810.1175596-1-gourry%40gourry.net
> Cc: stable@vger.kernel.org # v6.19+
> Assisted-by: LLM
> Signed-off-by: Gregory Price (Meta) <gourry@gourry.net>
> ---

Acked-by: David Hildenbrand (Arm) <david@kernel.org>

-- 
Cheers,

David

Re: [PATCH v2 1/2] mm/mempolicy: use vm_normal_folio_pmd() in queue_folios_pmd()

[Lorenzo Stoakes (ARM)]

On Fri, Sep 11, 2026 at 11:48:32PM -0400, Gregory Price wrote:
> mmap a VM_PFNMAP region whose ->huge_fault installs a PMD through
> vmf_insert_pfn_pmd() - a vfio-pci MMIO BAR does this - then
>
> 	mbind(p, len, MPOL_BIND, &mask, maxnode, MPOL_MF_STRICT);
>
> With a stand-in module for the driver:
>
>   BUG: unable to handle page fault for address: fffff96dc0000008
>   RIP: 0010:queue_folios_pte_range+0xaf/0x440
>    walk_pgd_range+0x52b/0xaf0
>    __walk_page_range+0x6a/0x1d0
>    walk_page_range_mm_unsafe+0x193/0x230
>    queue_pages_range+0x64/0xa0
>    do_mbind+0x25e/0x640
>
> queue_folios_pmd(), inlined above, calls pmd_folio() on that PMD.  The pfn

This makes me wonder if pmd_folio() is just generally defunct in favour of
vm_normal_folio_pmd()?

> is raw MMIO with no memmap entry, so the folio lands in unpopulated
> vmemmap.  Neither guard stops the walk:
>
>   walk_page_test()         skips VM_PFNMAP, but queue_pages_walk_ops
>                            supplies ->test_walk, so it never runs
>   queue_pages_test_walk()  honours vma_migratable(), but only while
>                            MPOL_MF_STRICT is clear
>
> A VM_MIXEDMAP vma needs neither flag, being vma_migratable(), so plain
> mbind(MPOL_MF_MOVE) reaches this too - and there the bad folio carries on
> into migrate_folio_add() and folio_isolate_lru().  mshv_vtl_low is such a
> mapping.
>
> Use vm_normal_folio_pmd() and skip on NULL, as the PTE loop in
> queue_folios_pte_range() already does with vm_normal_folio().  On the NULL
> path, retain ACTION_CONTINUE handling for the huge zero PMD.
>
> mbind(MPOL_MF_STRICT) over a PMD mapped VM_PFNMAP region now returns 0
> rather than -EIO.  The PTE loop already returned 0 there.
>
> Fixes: 3c8e44c9b369 ("mm: mark special bits for huge pfn mappings when inject")
> Reported-by: sashiko-bot <sashiko-bot@kernel.org>
> Closes: https://sashiko.dev/#/patchset/20260817220810.1175596-1-gourry%40gourry.net
> Cc: stable@vger.kernel.org
> Assisted-by: LLM
> Signed-off-by: Gregory Price (Meta) <gourry@gourry.net>
> Acked-by: David Hildenbrand (Arm) <david@kernel.org>
> ---
>  mm/mempolicy.c | 16 +++++++++-------
>  1 file changed, 9 insertions(+), 7 deletions(-)
>
> diff --git a/mm/mempolicy.c b/mm/mempolicy.c
> index 2ad0a5f18280..8fc8a975657e 100644
> --- a/mm/mempolicy.c
> +++ b/mm/mempolicy.c
> @@ -667,7 +667,8 @@ static inline bool queue_folio_required(struct folio *folio,
>  	return node_isset(nid, *qp->nmask) == !(flags & MPOL_MF_INVERT);
>  }
>
> -static void queue_folios_pmd(pmd_t *pmd, struct mm_walk *walk)
> +static void queue_folios_pmd(pmd_t *pmd, unsigned long addr,
> +			     struct mm_walk *walk)
>  {
>  	struct folio *folio;
>  	struct queue_pages *qp = walk->private;
> @@ -678,13 +679,14 @@ static void queue_folios_pmd(pmd_t *pmd, struct mm_walk *walk)
>  			qp->nr_failed++;
>  		return;
>  	}
> -	folio = pmd_folio(pmdval);
> -	if (folio_is_zone_device(folio))
> -		return;
> -	if (is_huge_zero_folio(folio)) {
> -		walk->action = ACTION_CONTINUE;
> +	folio = vm_normal_folio_pmd(walk->vma, addr, pmdval);
> +	if (!folio) {
> +		if (is_huge_zero_pmd(pmdval))
> +			walk->action = ACTION_CONTINUE;

Since you've already identified that this is a leaf PMD, is there really any
difference between setting walk->action to ACTION_CONTINUE or not?

There's nothing to descend to in the subtree here so either way you continue to
the next PMD entry don't you?

So there's a pmd_entry specified for both set of walk ops:

static const struct mm_walk_ops queue_pages_walk_ops = {
	.hugetlb_entry		= queue_folios_hugetlb,
	.pmd_entry		= queue_folios_pte_range,
	.test_walk		= queue_pages_test_walk,
	.walk_lock		= PGWALK_RDLOCK,
};

static const struct mm_walk_ops queue_pages_lock_vma_walk_ops = {
	.hugetlb_entry		= queue_folios_hugetlb,
	.pmd_entry		= queue_folios_pte_range,
	.test_walk		= queue_pages_test_walk,
	.walk_lock		= PGWALK_WRLOCK,
};

But no pte_entry or install_pte handlers.

Reading walk_pmd_range() in mm/pagewalk.c:

static int walk_pmd_range(pud_t *pud, unsigned long addr, unsigned long end,
			  struct mm_walk *walk)
{
	bool has_handler = ops->pte_entry;
	...
	bool has_install = ops->install_pte;
	...

	do {
again:
		walk->action = ACTION_SUBTREE;
		...
		if (ops->pmd_entry)
			err = ops->pmd_entry(pmd, addr, next, walk);
		if (err)
			break;

		if (walk->action == ACTION_AGAIN)
			goto again;
		if (walk->action == ACTION_CONTINUE)
			continue;

		if (!has_handler) { /* No handlers for lower page tables. */
			if (!has_install)
				continue; /* Nothing to do. */
			...
		}
	...
	} while (pmd++, addr = next, addr != end);
	...
}

It seems that setting ACTION_CONTINUE or not makes no difference.

This is really just a long way of saying maybe you could just do:

	folio = vm_normal_folio_pmd(walk->vma, addr, pmdval);
	if (!folio)
		return;

? :)

>  		return;
>  	}
> +	if (folio_is_zone_device(folio))
> +		return;
>  	if (!queue_folio_required(folio, qp))
>  		return;
>  	if (!(qp->flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL)) ||
> @@ -717,7 +719,7 @@ static int queue_folios_pte_range(pmd_t *pmd, unsigned long addr,
>
>  	ptl = pmd_trans_huge_lock(pmd, vma);
>  	if (ptl) {
> -		queue_folios_pmd(pmd, walk);
> +		queue_folios_pmd(pmd, addr, walk);
>  		spin_unlock(ptl);
>  		goto out;
>  	}
> --
> 2.55.0
>

--
Cheers, Lorenzo

Re: [PATCH v2 2/2] mm/madvise: use vm_normal_folio_pmd() in cold/pageout PMD range

[Lorenzo Stoakes (ARM)]

On Fri, Sep 11, 2026 at 11:48:33PM -0400, Gregory Price wrote:
> mmap a VM_MIXEDMAP region whose ->huge_fault installs a PMD through

Gosh, shock, horror! VMA_MIXEDMAP_BIT sir! :P

(It's fine it's fine)

> vmf_insert_pfn_pmd() - mshv_vtl_low does this, and needs CAP_SYS_ADMIN
> to open - then:
>
>   madvise(p, PMD_SIZE, MADV_PAGEOUT);
>
> With a stand-in module for the driver:
>
>   BUG: unable to handle page fault for address: fffff587c0000008
>   RIP: 0010:madvise_cold_or_pageout_pte_range+0x410/0x9b0
>    walk_pgd_range+0x52b/0xaf0
>    __walk_page_range+0x6a/0x1d0
>    walk_page_range_vma_unsafe+0x8e/0x120
>    madvise_pageout+0xb2/0x180
>    madvise_vma_behavior+0x46b/0xa90
>    do_madvise+0x108/0x190
>    __x64_sys_madvise+0x26/0x30
>
> Nothing validates the pfn on the way in:
>
>   can_madv_lru_vma()     rejects VM_PFNMAP, but not VM_MIXEDMAP
>   can_fault()            *pfn = vmf->pgoff & ~(mask >> PAGE_SHIFT);
>   vmf_insert_pfn_pmd()   no pfn_valid() check
>   pmd_folio()            pfn_to_page() -> unpopulated vmemmap

I definitely suggest checking out the small series [0] I sent which changes
how these kinds of semantics are expressed where I... ugh what I missed
can_madv_lru_vma()! Damn it.

Noted as a follow up :)

[0]:https://lore.kernel.org/linux-mm/20260914-b4-mmap-prepare-vma-flag-sanify-v2-0-7d9781ed5361@kernel.org/

>
> Even with a valid pfn the path is wrong.  The mapping carries no rmap, so

Isn't a non-rmappable page not a folio?

I mean the fact that vm_normal_folio_pmd() returns NULL is kinda saying
that :)

> folio_maybe_mapped_shared() sees mapcount 0, and the walker goes on to
> folio_deactivate(), or folio_isolate_lru() plus reclaim_pages(), against a
> folio this mapping does not own.
>
> Use vm_normal_folio_pmd() and skip on NULL, as the PTE half of this same
> walker already does with vm_normal_folio().  This also filters the huge
> zero PMD, so its separate check is no longer needed.
>
> Fixes: 3c8e44c9b369 ("mm: mark special bits for huge pfn mappings when inject")
> Reported-by: sashiko-bot <sashiko-bot@kernel.org>
> Closes: https://sashiko.dev/#/patchset/20260817220810.1175596-1-gourry%40gourry.net
> Cc: stable@vger.kernel.org # v6.19+
> Assisted-by: LLM
> Signed-off-by: Gregory Price (Meta) <gourry@gourry.net>

LGTM in general so:

Reviewed-by: Lorenzo Stoakes (ARM) <ljs@kernel.org>

> ---
>  mm/madvise.c | 7 +++----
>  1 file changed, 3 insertions(+), 4 deletions(-)
>
> diff --git a/mm/madvise.c b/mm/madvise.c
> index f75a9d139980..fbb72ab49aa6 100644
> --- a/mm/madvise.c
> +++ b/mm/madvise.c
> @@ -395,16 +395,15 @@ static int madvise_cold_or_pageout_pte_range(pmd_t *pmd,
>  			return 0;
>

Oh God! This function again!

>  		orig_pmd = *pmd;
> -		if (is_huge_zero_pmd(orig_pmd))
> -			goto huge_unlock;
> -
>  		if (unlikely(!pmd_present(orig_pmd))) {
>  			VM_WARN_ON_ONCE(!pmd_is_migration_entry(orig_pmd) &&
>  					!pmd_is_device_private_entry(orig_pmd));
>  			goto huge_unlock;
>  		}
>
> -		folio = pmd_folio(orig_pmd);

Again I'm wondering if pmd_folio() is just a code smell in general?

> +		folio = vm_normal_folio_pmd(vma, addr, orig_pmd);
> +		if (!folio)
> +			goto huge_unlock;
>
>  		if (folio_is_zone_device(folio))
>  			goto huge_unlock;
> --
> 2.55.0
>

--
Cheers, Lorenzo