Re: [PATCH v2] hugetlbfs: skip VMAs without shareable locks in hugetlb_vmdelete_list

[Andrew Morton <akpm@linux-foundation.org>]

On Mon, 20 Oct 2025 18:52:11 +0100 Mark Brown <broonie@kernel.org> wrote:

> > For the past few days I've been seeing failures on Raspberry Pi 4 in
> > the hugetlbfs-madvise kselftest in -next which bisect to this patch.
> > The test reports:
> > 
> > # # -------------------------
> > # # running ./hugetlb-madvise
> > # # -------------------------
> > # # Unexpected number of free huge pages line 252
> > # # [FAIL]
> > # not ok 6 hugetlb-madvise # exit=1
> 
> This issue is now present in mainline:
> 
> Raspberry Pi 4: https://lava.sirena.org.uk/scheduler/job/1976561#L1798
> Orion O6:       https://lava.sirena.org.uk/scheduler/job/1977081#L1779
> 
> and still bisects to this patch.

Thanks.  Were you able to test the proposed fix?

From: Deepanshu Kartikey <kartikey406@gmail.com>
Subject: hugetlbfs: move lock assertions after early returns in huge_pmd_unshare()
Date: Tue, 14 Oct 2025 17:03:44 +0530

When hugetlb_vmdelete_list() processes VMAs during truncate operations, it
may encounter VMAs where huge_pmd_unshare() is called without the required
shareable lock.  This triggers an assertion failure in
hugetlb_vma_assert_locked().

The previous fix in commit dd83609b8898 ("hugetlbfs: skip VMAs without
shareable locks in hugetlb_vmdelete_list") skipped entire VMAs without
shareable locks to avoid the assertion.  However, this prevented pages
from being unmapped and freed, causing a regression in
fallocate(PUNCH_HOLE) operations where pages were not freed immediately,
as reported by Mark Brown.

Instead of checking locks in the caller or skipping VMAs, move the lock
assertions in huge_pmd_unshare() to after the early return checks.  The
assertions are only needed when actual PMD unsharing work will be
performed.  If the function returns early because sz != PMD_SIZE or the
PMD is not shared, no locks are required and assertions should not fire.

This approach reverts the VMA skipping logic from commit dd83609b8898
("hugetlbfs: skip VMAs without shareable locks in hugetlb_vmdelete_list")
while moving the assertions to avoid the assertion failure, keeping all
the logic within huge_pmd_unshare() itself and allowing page unmapping and
freeing to proceed for all VMAs.

Link: https://lkml.kernel.org/r/20251014113344.21194-1-kartikey406@gmail.com
Fixes: dd83609b8898 ("hugetlbfs: skip VMAs without shareable locks in hugetlb_vmdelete_list")
Signed-off-by: Deepanshu Kartikey <kartikey406@gmail.com>
Reported-by: <syzbot+f26d7c75c26ec19790e7@syzkaller.appspotmail.com>
Reported-by: Mark Brown <broonie@kernel.org>
Closes: https://syzkaller.appspot.com/bug?extid=f26d7c75c26ec19790e7
Suggested-by: David Hildenbrand <david@redhat.com>
Suggested-by: Oscar Salvador <osalvador@suse.de>
Tested-by: <syzbot+f26d7c75c26ec19790e7@syzkaller.appspotmail.com>
Acked-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---

 fs/hugetlbfs/inode.c |    9 ---------
 mm/hugetlb.c         |    5 ++---
 2 files changed, 2 insertions(+), 12 deletions(-)

--- a/fs/hugetlbfs/inode.c~hugetlbfs-move-lock-assertions-after-early-returns-in-huge_pmd_unshare
+++ a/fs/hugetlbfs/inode.c
@@ -478,14 +478,6 @@ hugetlb_vmdelete_list(struct rb_root_cac
 		if (!hugetlb_vma_trylock_write(vma))
 			continue;
 
-		/*
-		 * Skip VMAs without shareable locks. Per the design in commit
-		 * 40549ba8f8e0, these will be handled by remove_inode_hugepages()
-		 * called after this function with proper locking.
-		 */
-		if (!__vma_shareable_lock(vma))
-			goto skip;
-
 		v_start = vma_offset_start(vma, start);
 		v_end = vma_offset_end(vma, end);
 
@@ -496,7 +488,6 @@ hugetlb_vmdelete_list(struct rb_root_cac
 		 * vmas.  Therefore, lock is not held when calling
 		 * unmap_hugepage_range for private vmas.
 		 */
-skip:
 		hugetlb_vma_unlock_write(vma);
 	}
 }
--- a/mm/hugetlb.c~hugetlbfs-move-lock-assertions-after-early-returns-in-huge_pmd_unshare
+++ a/mm/hugetlb.c
@@ -7614,13 +7614,12 @@ int huge_pmd_unshare(struct mm_struct *m
 	p4d_t *p4d = p4d_offset(pgd, addr);
 	pud_t *pud = pud_offset(p4d, addr);
 
-	i_mmap_assert_write_locked(vma->vm_file->f_mapping);
-	hugetlb_vma_assert_locked(vma);
 	if (sz != PMD_SIZE)
 		return 0;
 	if (!ptdesc_pmd_is_shared(virt_to_ptdesc(ptep)))
 		return 0;
-
+	i_mmap_assert_write_locked(vma->vm_file->f_mapping);
+	hugetlb_vma_assert_locked(vma);
 	pud_clear(pud);
 	/*
 	 * Once our caller drops the rmap lock, some other process might be
_