Re: [v3 02/11] mm/thp: zone_device awareness in THP handling code

[Matthew Brost <matthew.brost@intel.com>]

On Thu, Aug 28, 2025 at 10:12:53PM +0200, David Hildenbrand wrote:
> On 28.08.25 22:05, Matthew Brost wrote:
> > On Tue, Aug 12, 2025 at 12:40:27PM +1000, Balbir Singh wrote:
> > > Make THP handling code in the mm subsystem for THP pages aware of zone
> > > device pages. Although the code is designed to be generic when it comes
> > > to handling splitting of pages, the code is designed to work for THP
> > > page sizes corresponding to HPAGE_PMD_NR.
> > > 
> > > Modify page_vma_mapped_walk() to return true when a zone device huge
> > > entry is present, enabling try_to_migrate() and other code migration
> > > paths to appropriately process the entry. page_vma_mapped_walk() will
> > > return true for zone device private large folios only when
> > > PVMW_THP_DEVICE_PRIVATE is passed. This is to prevent locations that are
> > > not zone device private pages from having to add awareness. The key
> > > callback that needs this flag is try_to_migrate_one(). The other
> > > callbacks page idle, damon use it for setting young/dirty bits, which is
> > > not significant when it comes to pmd level bit harvesting.
> > > 
> > > pmd_pfn() does not work well with zone device entries, use
> > > pfn_pmd_entry_to_swap() for checking and comparison as for zone device
> > > entries.
> > > 
> > > Support partial unmapping of zone device private entries, which happens
> > > via munmap(). munmap() causes the device private entry pmd to be split,
> > > but the corresponding folio is not split. Deferred split does not work for
> > > zone device private folios due to the need to split during fault
> > > handling. Get migrate_vma_collect_pmd() to handle this case by splitting
> > > partially unmapped device private folios.
> > > 
> > > Cc: Andrew Morton <akpm@linux-foundation.org>
> > > Cc: David Hildenbrand <david@redhat.com>
> > > Cc: Zi Yan <ziy@nvidia.com>
> > > Cc: Joshua Hahn <joshua.hahnjy@gmail.com>
> > > Cc: Rakie Kim <rakie.kim@sk.com>
> > > Cc: Byungchul Park <byungchul@sk.com>
> > > Cc: Gregory Price <gourry@gourry.net>
> > > Cc: Ying Huang <ying.huang@linux.alibaba.com>
> > > Cc: Alistair Popple <apopple@nvidia.com>
> > > Cc: Oscar Salvador <osalvador@suse.de>
> > > Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
> > > Cc: Baolin Wang <baolin.wang@linux.alibaba.com>
> > > Cc: "Liam R. Howlett" <Liam.Howlett@oracle.com>
> > > Cc: Nico Pache <npache@redhat.com>
> > > Cc: Ryan Roberts <ryan.roberts@arm.com>
> > > Cc: Dev Jain <dev.jain@arm.com>
> > > Cc: Barry Song <baohua@kernel.org>
> > > Cc: Lyude Paul <lyude@redhat.com>
> > > Cc: Danilo Krummrich <dakr@kernel.org>
> > > Cc: David Airlie <airlied@gmail.com>
> > > Cc: Simona Vetter <simona@ffwll.ch>
> > > Cc: Ralph Campbell <rcampbell@nvidia.com>
> > > Cc: Mika Penttilä <mpenttil@redhat.com>
> > > Cc: Matthew Brost <matthew.brost@intel.com>
> > > Cc: Francois Dugast <francois.dugast@intel.com>
> > > 
> > > Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> > > Signed-off-by: Balbir Singh <balbirs@nvidia.com>
> > > ---
> > >   include/linux/rmap.h    |   2 +
> > >   include/linux/swapops.h |  17 ++++
> > >   lib/test_hmm.c          |   2 +-
> > >   mm/huge_memory.c        | 214 +++++++++++++++++++++++++++++++---------
> > >   mm/migrate_device.c     |  47 +++++++++
> > >   mm/page_vma_mapped.c    |  13 ++-
> > >   mm/pgtable-generic.c    |   6 ++
> > >   mm/rmap.c               |  24 ++++-
> > >   8 files changed, 272 insertions(+), 53 deletions(-)
> > > 
> > > diff --git a/include/linux/rmap.h b/include/linux/rmap.h
> > > index 6cd020eea37a..dfb7aae3d77b 100644
> > > --- a/include/linux/rmap.h
> > > +++ b/include/linux/rmap.h
> > > @@ -927,6 +927,8 @@ struct page *make_device_exclusive(struct mm_struct *mm, unsigned long addr,
> > >   #define PVMW_SYNC		(1 << 0)
> > >   /* Look for migration entries rather than present PTEs */
> > >   #define PVMW_MIGRATION		(1 << 1)
> > > +/* Look for device private THP entries */
> > > +#define PVMW_THP_DEVICE_PRIVATE	(1 << 2)
> > >   struct page_vma_mapped_walk {
> > >   	unsigned long pfn;
> > > diff --git a/include/linux/swapops.h b/include/linux/swapops.h
> > > index 64ea151a7ae3..2641c01bd5d2 100644
> > > --- a/include/linux/swapops.h
> > > +++ b/include/linux/swapops.h
> > > @@ -563,6 +563,7 @@ static inline int is_pmd_migration_entry(pmd_t pmd)
> > >   {
> > >   	return is_swap_pmd(pmd) && is_migration_entry(pmd_to_swp_entry(pmd));
> > >   }
> > > +
> > >   #else  /* CONFIG_ARCH_ENABLE_THP_MIGRATION */
> > >   static inline int set_pmd_migration_entry(struct page_vma_mapped_walk *pvmw,
> > >   		struct page *page)
> > > @@ -594,6 +595,22 @@ static inline int is_pmd_migration_entry(pmd_t pmd)
> > >   }
> > >   #endif  /* CONFIG_ARCH_ENABLE_THP_MIGRATION */
> > > +#if defined(CONFIG_ZONE_DEVICE) && defined(CONFIG_ARCH_ENABLE_THP_MIGRATION)
> > > +
> > > +static inline int is_pmd_device_private_entry(pmd_t pmd)
> > > +{
> > > +	return is_swap_pmd(pmd) && is_device_private_entry(pmd_to_swp_entry(pmd));
> > > +}
> > > +
> > > +#else /* CONFIG_ZONE_DEVICE && CONFIG_ARCH_ENABLE_THP_MIGRATION */
> > > +
> > > +static inline int is_pmd_device_private_entry(pmd_t pmd)
> > > +{
> > > +	return 0;
> > > +}
> > > +
> > > +#endif /* CONFIG_ZONE_DEVICE && CONFIG_ARCH_ENABLE_THP_MIGRATION */
> > > +
> > >   static inline int non_swap_entry(swp_entry_t entry)
> > >   {
> > >   	return swp_type(entry) >= MAX_SWAPFILES;
> > > diff --git a/lib/test_hmm.c b/lib/test_hmm.c
> > > index 761725bc713c..297f1e034045 100644
> > > --- a/lib/test_hmm.c
> > > +++ b/lib/test_hmm.c
> > > @@ -1408,7 +1408,7 @@ static vm_fault_t dmirror_devmem_fault(struct vm_fault *vmf)
> > >   	 * the mirror but here we use it to hold the page for the simulated
> > >   	 * device memory and that page holds the pointer to the mirror.
> > >   	 */
> > > -	rpage = vmf->page->zone_device_data;
> > > +	rpage = folio_page(page_folio(vmf->page), 0)->zone_device_data;
> > >   	dmirror = rpage->zone_device_data;
> > >   	/* FIXME demonstrate how we can adjust migrate range */
> > > diff --git a/mm/huge_memory.c b/mm/huge_memory.c
> > > index 9c38a95e9f09..2495e3fdbfae 100644
> > > --- a/mm/huge_memory.c
> > > +++ b/mm/huge_memory.c
> > > @@ -1711,8 +1711,11 @@ int copy_huge_pmd(struct mm_struct *dst_mm, struct mm_struct *src_mm,
> > >   	if (unlikely(is_swap_pmd(pmd))) {
> > >   		swp_entry_t entry = pmd_to_swp_entry(pmd);
> > > -		VM_BUG_ON(!is_pmd_migration_entry(pmd));
> > > -		if (!is_readable_migration_entry(entry)) {
> > > +		VM_WARN_ON(!is_pmd_migration_entry(pmd) &&
> > > +				!is_pmd_device_private_entry(pmd));
> > > +
> > > +		if (is_migration_entry(entry) &&
> > > +			is_writable_migration_entry(entry)) {
> > >   			entry = make_readable_migration_entry(
> > >   							swp_offset(entry));
> > >   			pmd = swp_entry_to_pmd(entry);
> > > @@ -1722,6 +1725,32 @@ int copy_huge_pmd(struct mm_struct *dst_mm, struct mm_struct *src_mm,
> > >   				pmd = pmd_swp_mkuffd_wp(pmd);
> > >   			set_pmd_at(src_mm, addr, src_pmd, pmd);
> > >   		}
> > > +
> > > +		if (is_device_private_entry(entry)) {
> > > +			if (is_writable_device_private_entry(entry)) {
> > > +				entry = make_readable_device_private_entry(
> > > +					swp_offset(entry));
> > > +				pmd = swp_entry_to_pmd(entry);
> > > +
> > > +				if (pmd_swp_soft_dirty(*src_pmd))
> > > +					pmd = pmd_swp_mksoft_dirty(pmd);
> > > +				if (pmd_swp_uffd_wp(*src_pmd))
> > > +					pmd = pmd_swp_mkuffd_wp(pmd);
> > > +				set_pmd_at(src_mm, addr, src_pmd, pmd);
> > > +			}
> > > +
> > > +			src_folio = pfn_swap_entry_folio(entry);
> > > +			VM_WARN_ON(!folio_test_large(src_folio));
> > > +
> > > +			folio_get(src_folio);
> > > +			/*
> > > +			 * folio_try_dup_anon_rmap_pmd does not fail for
> > > +			 * device private entries.
> > > +			 */
> > > +			VM_WARN_ON(folio_try_dup_anon_rmap_pmd(src_folio,
> > > +					  &src_folio->page, dst_vma, src_vma));
> > 
> > VM_WARN_ON compiles out in non-debug builds. I hit this running the
> > fork self I shared with a non-debug build.
> 
> 
> folio_try_dup_anon_rmap_pmd() will never fail for
> folio_is_device_private(folio) -- unless something is deeply messed up that
> we wouldn't identify this folio as being device-private.
> 
> Can you elaborate, what were you able to trigger, and in what kind of
> environment?
> 

Maybe this was bad phrasing. I compilied the kernel with a non-debug
build and fork() broke for THP device pages because the above call to
folio_try_dup_anon_rmap_pmd compiled out (i.e., it wasn't called).

Matt

> -- 
> Cheers
> 
> David / dhildenb
>