[PATCH] mm: fix __vm_normal_page() to handle missing support for pmd_special()/pud_special()

["David Hildenbrand (Arm)" <david@kernel.org>]

On x86 32-bit with THP enabled, zap_huge_pmd() is seen to generate a
"WARNING: mm/memory.c:735 at __vm_normal_page+0x6a/0x7d", from the
VM_WARN_ON_ONCE(is_zero_pfn(pfn) || is_huge_zero_pfn(pfn)); followed
by "BUG: Bad rss-counter state"s, then later "BUG: Bad page state"s
when reclaim gets to call shrink_huge_zero_folio_scan().

It's as if the _PAGE_SPECIAL bit never got set in the huge_zero pmd:
and indeed, whereas pte_special() and pte_mkspecial() are subject to a
dedicated CONFIG_ARCH_HAS_PTE_SPECIAL, pmd_special() and pmd_mkspecial()
are subject to CONFIG_ARCH_SUPPORTS_PMD_PFNMAP, which is never enabled
on any 32-bit architecture.

While the problem was exposed through commit d80a9cb1a64a ("mm/huge_memory:
add and use normal_or_softleaf_folio_pmd()"), it was an oversight in commit
af38538801c6 ("mm/memory: factor out common code from vm_normal_page_*()")
and would result in other problems:
* huge zero folio accounted in smaps, pagemap (PAGE_IS_FILE) and
  numamaps as file-backed THP
* folio_walk_start() returning the folio even without FW_ZEROPAGE set.
  Callers seem to tolerate that, though.

... and triggering the VM_WARN_ON_ONE(), although never reported so far.

To fix it, teach vm_normal_page_pmd()/vm_normal_page_pud() to consider
whether pmd_special/pud_special is actually implemented.

Fixes: af38538801c6 ("mm/memory: factor out common code from vm_normal_page_*()")
Reported-by: Hugh Dickins <hughd@google.com>
Closes: https://lore.kernel.org/r/74a75b59-2e13-3985-ee99-d5521f39df2a@google.com
Reported-by: Bibo Mao <maobibo@loongson.cn>
Closes: https://lore.kernel.org/r/20260430041121.2839350-1-maobibo@loongson.cn
Debugged-by: Hugh Dickins <hughd@google.com>
Reviewed-by: Lance Yang <lance.yang@linux.dev>
Tested-by: Bibo Mao <maobibo@loongson.cn>
Cc: stable@vger.kernel.org
Signed-off-by: David Hildenbrand (Arm) <david@kernel.org>
---
This is an alternative to Hugh's patch, whereby we leave pmd_special()
be a NOP and instead teach __vm_normal_page() about lack of support for
pmd_special/pud_special.
---
 mm/memory.c | 22 +++++++++++++++++++---
 1 file changed, 19 insertions(+), 3 deletions(-)

diff --git a/mm/memory.c b/mm/memory.c
index 7322a40e73b9..4d84976fc7f4 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -612,6 +612,21 @@ static void print_bad_page_map(struct vm_area_struct *vma,
 	dump_stack();
 	add_taint(TAINT_BAD_PAGE, LOCKDEP_NOW_UNRELIABLE);
 }
+
+static inline bool pgtable_level_has_pxx_special(enum pgtable_level level)
+{
+	switch (level) {
+	case PGTABLE_LEVEL_PTE:
+		return IS_ENABLED(CONFIG_ARCH_HAS_PTE_SPECIAL);
+	case PGTABLE_LEVEL_PMD:
+		return IS_ENABLED(CONFIG_ARCH_SUPPORTS_PMD_PFNMAP);
+	case PGTABLE_LEVEL_PUD:
+		return IS_ENABLED(CONFIG_ARCH_SUPPORTS_PUD_PFNMAP);
+	default:
+		return false;
+	}
+}
+
 #define print_bad_pte(vma, addr, pte, page) \
 	print_bad_page_map(vma, addr, pte_val(pte), page, PGTABLE_LEVEL_PTE)
 
@@ -684,7 +699,7 @@ static inline struct page *__vm_normal_page(struct vm_area_struct *vma,
 		unsigned long addr, unsigned long pfn, bool special,
 		unsigned long long entry, enum pgtable_level level)
 {
-	if (IS_ENABLED(CONFIG_ARCH_HAS_PTE_SPECIAL)) {
+	if (pgtable_level_has_pxx_special(level)) {
 		if (unlikely(special)) {
 #ifdef CONFIG_FIND_NORMAL_PAGE
 			if (vma->vm_ops && vma->vm_ops->find_normal_page)
@@ -699,8 +714,9 @@ static inline struct page *__vm_normal_page(struct vm_area_struct *vma,
 			return NULL;
 		}
 		/*
-		 * With CONFIG_ARCH_HAS_PTE_SPECIAL, any special page table
-		 * mappings (incl. shared zero folios) are marked accordingly.
+		 * With working pte_special()/pmd_special()..., any special page
+		 * table mappings (incl. shared zero folios) are marked
+		 * accordingly.
 		 */
 	} else {
 		if (unlikely(vma->vm_flags & (VM_PFNMAP | VM_MIXEDMAP))) {

---

base-commit: d94322006a51b522dd361128a450bf9e75aad889

change-id: 20260430-pmd_special-610dbdd8ac3c

--

Cheers,

David