mm: call arch_swap_restore() from unuse_pte()

We would like to move away from requiring architectures to restore
metadata from swap in the set_pte_at() implementation, as this is not only
error-prone but adds complexity to the arch-specific code.  This requires
us to call arch_swap_restore() before calling swap_free() whenever pages
are restored from swap.  We are currently doing so everywhere except in
unuse_pte(); do so there as well.

Link: https://lkml.kernel.org/r/[email protected]
Link: https://linux-review.googlesource.com/id/I68276653e612d64cde271ce1b5a99ae05d6bbc4f
Signed-off-by: Peter Collingbourne <[email protected]>
Suggested-by: David Hildenbrand <[email protected]>
Acked-by: David Hildenbrand <[email protected]>
Acked-by: "Huang, Ying" <[email protected]>
Reviewed-by: Steven Price <[email protected]>
Acked-by: Catalin Marinas <[email protected]>
Cc: Alexandru Elisei <[email protected]>
Cc: Chinwen Chang <[email protected]>
Cc: Evgenii Stepanov <[email protected]>
Cc: Greg Kroah-Hartman <[email protected]>
Cc: kasan-dev <[email protected]>
Cc: "Kuan-Ying Lee (李冠穎)" <[email protected]>
Cc: Qun-Wei Lin <[email protected]>
Cc: Suren Baghdasaryan <[email protected]>
Cc: Vincenzo Frascino <[email protected]>
Cc: Will Deacon <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>

mm/swapfile.c

@@ -1778,6 +1778,13 @@ static int unuse_pte(struct vm_area_struct *vma, pmd_t *pmd,
 		goto setpte;
 	}
 
+	/*
+	 * Some architectures may have to restore extra metadata to the page
+	 * when reading from swap. This metadata may be indexed by swap entry
+	 * so this must be called before swap_free().
+	 */
+	arch_swap_restore(entry, page_folio(page));
+
 	/* See do_swap_page() */
 	BUG_ON(!PageAnon(page) && PageMappedToDisk(page));
 	BUG_ON(PageAnon(page) && PageAnonExclusive(page));