[PATCH v7 10/23] mm/shmem: Handle uffd-wp during fork()

[Peter Xu <peterx@redhat.com>]

Normally we skip copy page when fork() for VM_SHARED shmem, but we can't skip
it anymore if uffd-wp is enabled on dst vma.  This should only happen when the
src uffd has UFFD_FEATURE_EVENT_FORK enabled on uffd-wp shmem vma, so that
VM_UFFD_WP will be propagated onto dst vma too, then we should copy the
pgtables with uffd-wp bit and pte markers, because these information will be
lost otherwise.

Since the condition checks will become even more complicated for deciding
"whether a vma needs to copy the pgtable during fork()", introduce a helper
vma_needs_copy() for it, so everything will be clearer.

Signed-off-by: Peter Xu <peterx@redhat.com>
---
 mm/memory.c | 49 +++++++++++++++++++++++++++++++++++++++++--------
 1 file changed, 41 insertions(+), 8 deletions(-)

diff --git a/mm/memory.c b/mm/memory.c
index e3e67e32eb8a..e9e335ecb5dc 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -857,6 +857,14 @@ copy_nonpresent_pte(struct mm_struct *dst_mm, struct mm_struct *src_mm,
 		if (try_restore_exclusive_pte(src_pte, src_vma, addr))
 			return -EBUSY;
 		return -ENOENT;
+	} else if (is_pte_marker_entry(entry)) {
+		/*
+		 * We're copying the pgtable should only because dst_vma has
+		 * uffd-wp enabled, do sanity check.
+		 */
+		WARN_ON_ONCE(!userfaultfd_wp(dst_vma));
+		set_pte_at(dst_mm, addr, dst_pte, pte);
+		return 0;
 	}
 	if (!userfaultfd_wp(dst_vma))
 		pte = pte_swp_clear_uffd_wp(pte);
@@ -1225,6 +1233,38 @@ copy_p4d_range(struct vm_area_struct *dst_vma, struct vm_area_struct *src_vma,
 	return 0;
 }
 
+/*
+ * Return true if the vma needs to copy the pgtable during this fork().  Return
+ * false when we can speed up fork() by allowing lazy page faults later until
+ * when the child accesses the memory range.
+ */
+bool
+vma_needs_copy(struct vm_area_struct *dst_vma, struct vm_area_struct *src_vma)
+{
+	/*
+	 * Always copy pgtables when dst_vma has uffd-wp enabled even if it's
+	 * file-backed (e.g. shmem). Because when uffd-wp is enabled, pgtable
+	 * contains uffd-wp protection information, that's something we can't
+	 * retrieve from page cache, and skip copying will lose those info.
+	 */
+	if (userfaultfd_wp(dst_vma))
+		return true;
+
+	if (src_vma->vm_flags & (VM_HUGETLB | VM_PFNMAP | VM_MIXEDMAP))
+		return true;
+
+	if (src_vma->anon_vma)
+		return true;
+
+	/*
+	 * Don't copy ptes where a page fault will fill them correctly.  Fork
+	 * becomes much lighter when there are big shared or private readonly
+	 * mappings. The tradeoff is that copy_page_range is more efficient
+	 * than faulting.
+	 */
+	return false;
+}
+
 int
 copy_page_range(struct vm_area_struct *dst_vma, struct vm_area_struct *src_vma)
 {
@@ -1238,14 +1278,7 @@ copy_page_range(struct vm_area_struct *dst_vma, struct vm_area_struct *src_vma)
 	bool is_cow;
 	int ret;
 
-	/*
-	 * Don't copy ptes where a page fault will fill them correctly.
-	 * Fork becomes much lighter when there are big shared or private
-	 * readonly mappings. The tradeoff is that copy_page_range is more
-	 * efficient than faulting.
-	 */
-	if (!(src_vma->vm_flags & (VM_HUGETLB | VM_PFNMAP | VM_MIXEDMAP)) &&
-	    !src_vma->anon_vma)
+	if (!vma_needs_copy(dst_vma, src_vma))
 		return 0;
 
 	if (is_vm_hugetlb_page(src_vma))
-- 
2.32.0