Re: [PATCH v10] mm: fix ASSERT_EXCLUSIVE_BITS by passing memdesc_flags_t by pointer

[David Hildenbrand (Arm)]

On 7/8/26 04:15, Hui Zhu wrote:
> From: Hui Zhu <zhuhui@xxxxxxxxxx>
> 
> KCSAN reports a data race between page_to_nid()/folio_pgdat() reading
> page->flags and folio_trylock()/folio_lock() concurrently doing
> test_and_set_bit_lock(PG_locked, ...) on the same word, e.g.:
> 
>   BUG: KCSAN: data-race in __lruvec_stat_mod_folio / shmem_get_folio_gfp
> 
> The race is benign: nid/zone bits are set once at page init and never
> overlap with PG_locked.  However, ASSERT_EXCLUSIVE_BITS() inside
> memdesc_nid/zonenum() was checking a by-value copy of the flags word,
> not the live page->flags, so it failed to annotate the real access.
> 
> Change memdesc_nid(), memdesc_zonenum(), memdesc_section(), and
> memdesc_is_zone_device() to take a const memdesc_flags_t * and update
> all callers to pass &page->flags / &folio->flags, so
> ASSERT_EXCLUSIVE_BITS() operates on the actual shared word.
> 
> Guard the ASSERT_EXCLUSIVE_BITS() call in memdesc_zonenum() under
> ZONES_WIDTH != 0 to avoid a zero-mask check on configs where the zone
> field is absent.  memdesc_section() needs no such guard, since
> SECTIONS_WIDTH is never 0 wherever SECTION_IN_PAGE_FLAGS is defined.
> Under CONFIG_NUMA=n, memdesc_nid() itself is stubbed to "return 0"
> instead of reading page->flags, since NODES_MASK is 0 and the check
> can never fire; page_to_nid()/folio_nid() now just call memdesc_nid()
> unconditionally and rely on that stub, instead of duplicating the
> CONFIG_NUMA split at each call site.
> 
> Co-developed-by: David Hildenbrand (Arm) <david@xxxxxxxxxx>
> Signed-off-by: David Hildenbrand (Arm) <david@xxxxxxxxxx>
> Signed-off-by: Hui Zhu <zhuhui@xxxxxxxxxx>
> ---
> Changelog:
> v10:
> According to the comments of David, drop the redundant CONFIG_NUMA split
> in page_to_nid()/folio_nid() and remove the SECTIONS_WIDTH != 0 guard
> around ASSERT_EXCLUSIVE_BITS() in memdesc_section().
> v9:
> Add the SECTIONS_WIDTH check to memdesc_section.
> v8:
> According to the comments of Andrew, include kcsan-checks.h in mm.h.
> Incorporate David's patch that switch memdesc_nid(), memdesc_zonenum(),
> memdesc_section() and memdesc_is_zone_device() to take a const
> memdesc_flags_t * instead of using a per-accessor macro/call-site hack.
> Update all callers accordingly and extend the same exclusive-bits check
> to memdesc_section() and memdesc_is_zone_device(), guarded by
> SECTIONS_WIDTH != 0 / reusing ZONES_WIDTH != 0 to avoid zero-mask checks
> on configs without the corresponding field.
> v7:
> According to the comments of Sashiko, restrict the memdesc_nid() macro
> to CONFIG_NUMA, keeping a plain "return 0" static inline stub otherwise,
> and re-add a local page pointer in page_to_nid() to avoid evaluating
> PF_POISONED_CHECK(page) twice.
> v6:
> According to the comments of David, turn memdesc_nid() from a static
> inline function into a macro so ASSERT_EXCLUSIVE_BITS() can check the
> caller's page->flags/folio->flags directly.
> v5:
> According to the comments of Sashiko, guard the ASSERT_EXCLUSIVE_BITS()
> calls with #ifndef NODE_NOT_IN_PAGE_FLAGS (for nid) and #if
> ZONES_WIDTH != 0 (for zonenum).
> According to the comments of David, avoid calling
> PF_POISONED_CHECK(page) twice in page_to_nid().
> According to the warning of lkp, switch the CONFIG_NUMA=n
> page_to_nid()/folio_nid() stubs from macros to static inline functions.
> v4:
> According to the comments of Andrew and Sashiko, set
> page_to_nid()/folio_nid() as static inline stubs returning 0
> under CONFIG_NUMA=n.
> v3:
> According to the comments of Andrew and Sashiko, move
> ASSERT_EXCLUSIVE_BITS out of memdesc_nid()/memdesc_zonenum()
> into the page/folio call sites.
> v2:
> According to the comments of David, remove useless comments and use
> ASSERT_EXCLUSIVE_BITS() in memdesc_nid() instead of data_race() in
> page_to_nid().
> 
>  include/asm-generic/memory_model.h |  2 +-
>  include/linux/mm.h                 | 28 ++++++++++++++++++++--------
>  include/linux/mm_inline.h          |  4 ++--
>  include/linux/mmzone.h             | 26 ++++++++++++++------------
>  mm/page_alloc.c                    |  6 +++---
>  mm/slab.h                          |  2 +-
>  mm/sparse.c                        |  2 +-
>  7 files changed, 42 insertions(+), 28 deletions(-)
> 
> diff --git a/include/asm-generic/memory_model.h b/include/asm-generic/memory_model.h
> index fd74de50b054..c6b4eafaf4cb 100644
> --- a/include/asm-generic/memory_model.h
> +++ b/include/asm-generic/memory_model.h
> @@ -53,7 +53,7 @@ static inline int pfn_valid(unsigned long pfn)
>   */
>  #define __page_to_pfn(pg)					\
>  ({	const struct page *__pg = (pg);				\
> -	int __sec = memdesc_section(__pg->flags);		\
> +	int __sec = memdesc_section(&__pg->flags);		\
>  	(unsigned long)(__pg - __section_mem_map_addr(__nr_to_section(__sec)));	\
>  })
>  
> diff --git a/include/linux/mm.h b/include/linux/mm.h
> index 2101e5205fc0..01683d50b3cd 100644
> --- a/include/linux/mm.h
> +++ b/include/linux/mm.h
> @@ -37,6 +37,7 @@
>  #include <linux/bitmap.h>
>  #include <linux/bitops.h>
>  #include <linux/iommu-debug-pagealloc.h>
> +#include <linux/kcsan-checks.h>
>  
>  struct mempolicy;
>  struct anon_vma;
> @@ -2286,22 +2287,32 @@ static inline int page_zone_id(struct page *page)
>  }
>  
>  #ifdef NODE_NOT_IN_PAGE_FLAGS
> -int memdesc_nid(memdesc_flags_t mdf);
> +int memdesc_nid(const memdesc_flags_t *mdf);
>  #else
> -static inline int memdesc_nid(memdesc_flags_t mdf)
> +#ifdef CONFIG_NUMA
> +static inline int memdesc_nid(const memdesc_flags_t *mdf)
>  {
> -	return (mdf.f >> NODES_PGSHIFT) & NODES_MASK;
> +	ASSERT_EXCLUSIVE_BITS(mdf->f, NODES_MASK << NODES_PGSHIFT);
> +	return (mdf->f >> NODES_PGSHIFT) & NODES_MASK;
>  }
> +#else
> +static inline int memdesc_nid(const memdesc_flags_t *mdf)
> +{
> +	return 0;
> +}
> +#endif
>  #endif
>  
>  static inline int page_to_nid(const struct page *page)
>  {
> -	return memdesc_nid(PF_POISONED_CHECK(page)->flags);
> +	const struct page *p = PF_POISONED_CHECK(page);
> +
> +	return memdesc_nid(&p->flags);
>  }

Nit: Why not keep it as

return memdesc_nid(&(PF_POISONED_CHECK(page)->flags));

Apart from that

Acked-by: David Hildenbrand (Arm) <david@xxxxxxxxxx>

-- 
Cheers,

David