mm: Remove pointless barrier() after pmdp_get_lockless()

pmdp_get_lockless() should itself imply any ordering required.
Signed-off-by: NPeter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20221022114425.298833095%40infradead.org

mm/hmm.c

@@ -362,7 +362,6 @@ static int hmm_vma_walk_pmd(pmd_t *pmdp,
 		 * values.
 		 */
 		pmd = pmdp_get_lockless(pmdp);
-		barrier();
 		if (!pmd_devmap(pmd) && !pmd_trans_huge(pmd))
 			goto again;
 

mm/vmscan.c

@@ -4041,9 +4041,6 @@ static void walk_pmd_range(pud_t *pud, unsigned long start, unsigned long end,
 	for (i = pmd_index(start), addr = start; addr != end; i++, addr = next) {
 		pmd_t val = pmdp_get_lockless(pmd + i);
 
-		/* for pmdp_get_lockless() */
-		barrier();
-
 		next = pmd_addr_end(addr, end);
 
 		if (!pmd_present(val) || is_huge_zero_pmd(val)) {