mm: prevent userfaults to be handled under per-vma lock

Due to the possibility of handle_userfault dropping mmap_lock, avoid fault
handling under VMA lock and retry holding mmap_lock.  This can be handled
more gracefully in the future.

Link: https://lkml.kernel.org/r/20230227173632.3292573-28-surenb@google.com
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Suggested-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

mm/memory.c

@@ -5266,6 +5266,15 @@ struct vm_area_struct *lock_vma_under_rcu(struct mm_struct *mm,
 	if (!vma_start_read(vma))
 		goto inval;
 
+	/*
+	 * Due to the possibility of userfault handler dropping mmap_lock, avoid
+	 * it for now and fall back to page fault handling under mmap_lock.
+	 */
+	if (userfaultfd_armed(vma)) {
+		vma_end_read(vma);
+		goto inval;
+	}
+
 	/* Check since vm_start/vm_end might change before we lock the VMA */
 	if (unlikely(address < vma->vm_start || address >= vma->vm_end)) {
 		vma_end_read(vma);