mm/rmap: minimize folio->_nr_pages_mapped updates when batching PTE (un)mapping

It is not immediately obvious, but we can move the folio->_nr_pages_mapped
update out of the loop and reduce the number of atomic ops without
affecting the stats.

The important point to realize is that only removing the last PMD mapping
will result in _nr_pages_mapped going below ENTIRELY_MAPPED, not the
individual atomic_inc_return_relaxed() calls.  Concurrent races with
removal of PMD mappings should be handled as expected, just like when we
would have such races right now on a single mapcount update.

In a simple munmap() microbenchmark [1] on 1 GiB of memory backed by the
same PTE-mapped folio size (only mapped by a single process such that they
will get completely unmapped), this change results in a speedup (positive
is good) per folio size on a x86-64 Intel machine of roughly (a bit of
noise expected):

* 16 KiB: +10%
* 32 KiB: +15%
* 64 KiB: +17%
* 128 KiB: +21%
* 256 KiB: +22%
* 512 KiB: +22%
* 1024 KiB: +23%
* 2048 KiB: +27%

[1] https://gitlab.com/davidhildenbrand/scratchspace/-/blob/main/pte-mapped-folio-benchmarks.c

Link: https://lkml.kernel.org/r/20240807115515.1640951-1-david@redhat.comSigned-off-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

mm/rmap.c

@@ -1158,7 +1158,7 @@ static __always_inline unsigned int __folio_add_rmap(struct folio *folio,
 {
 	atomic_t *mapped = &folio->_nr_pages_mapped;
 	const int orig_nr_pages = nr_pages;
-	int first, nr = 0;
+	int first = 0, nr = 0;
 
 	__folio_rmap_sanity_checks(folio, page, nr_pages, level);
 
@@ -1170,13 +1170,13 @@ static __always_inline unsigned int __folio_add_rmap(struct folio *folio,
 		}
 
 		do {
-			first = atomic_inc_and_test(&page->_mapcount);
-			if (first) {
-				first = atomic_inc_return_relaxed(mapped);
-				if (first < ENTIRELY_MAPPED)
-					nr++;
-			}
+			first += atomic_inc_and_test(&page->_mapcount);
 		} while (page++, --nr_pages > 0);
+
+		if (first &&
+		    atomic_add_return_relaxed(first, mapped) < ENTIRELY_MAPPED)
+			nr = first;
+
 		atomic_add(orig_nr_pages, &folio->_large_mapcount);
 		break;
 	case RMAP_LEVEL_PMD:
@@ -1527,7 +1527,7 @@ static __always_inline void __folio_remove_rmap(struct folio *folio,
 		enum rmap_level level)
 {
 	atomic_t *mapped = &folio->_nr_pages_mapped;
-	int last, nr = 0, nr_pmdmapped = 0;
+	int last = 0, nr = 0, nr_pmdmapped = 0;
 	bool partially_mapped = false;
 
 	__folio_rmap_sanity_checks(folio, page, nr_pages, level);
@@ -1541,14 +1541,13 @@ static __always_inline void __folio_remove_rmap(struct folio *folio,
 
 		atomic_sub(nr_pages, &folio->_large_mapcount);
 		do {
-			last = atomic_add_negative(-1, &page->_mapcount);
-			if (last) {
-				last = atomic_dec_return_relaxed(mapped);
-				if (last < ENTIRELY_MAPPED)
-					nr++;
-			}
+			last += atomic_add_negative(-1, &page->_mapcount);
 		} while (page++, --nr_pages > 0);
 
+		if (last &&
+		    atomic_sub_return_relaxed(last, mapped) < ENTIRELY_MAPPED)
+			nr = last;
+
 		partially_mapped = nr && atomic_read(mapped);
 		break;
 	case RMAP_LEVEL_PMD: