zsmalloc: rework zspage chain size selection

Patch series "zsmalloc: make zspage chain size configurable".

Computers are bad at division.  We currently decide the best zspage chain
size (max number of physical pages per-zspage) by looking at a `used
percentage` value.  This is not enough as we lose precision during usage
percentage calculations For example, let's look at size class 208:

pages per zspage       wasted bytes         used%
       1                   144               96
       2                    80               99
       3                    16               99
       4                   160               99

Current algorithm will select 2 page per zspage configuration, as it's the
first one to reach 99%.  However, 3 pages per zspage waste less memory.

Change algorithm and select zspage configuration that has lowest wasted
value.

Link: https://lkml.kernel.org/r/[email protected]
Link: https://lkml.kernel.org/r/[email protected]
Signed-off-by: Sergey Senozhatsky <[email protected]>
Acked-by: Minchan Kim <[email protected]>
Cc: Mike Kravetz <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>

mm/zsmalloc.c

@@ -820,42 +820,6 @@ static enum fullness_group fix_fullness_group(struct size_class *class,
 	return newfg;
 }
 
-/*
- * We have to decide on how many pages to link together
- * to form a zspage for each size class. This is important
- * to reduce wastage due to unusable space left at end of
- * each zspage which is given as:
- *     wastage = Zp % class_size
- *     usage = Zp - wastage
- * where Zp = zspage size = k * PAGE_SIZE where k = 1, 2, ...
- *
- * For example, for size class of 3/8 * PAGE_SIZE, we should
- * link together 3 PAGE_SIZE sized pages to form a zspage
- * since then we can perfectly fit in 8 such objects.
- */
-static int get_pages_per_zspage(int class_size)
-{
-	int i, max_usedpc = 0;
-	/* zspage order which gives maximum used size per KB */
-	int max_usedpc_order = 1;
-
-	for (i = 1; i <= ZS_MAX_PAGES_PER_ZSPAGE; i++) {
-		int zspage_size;
-		int waste, usedpc;
-
-		zspage_size = i * PAGE_SIZE;
-		waste = zspage_size % class_size;
-		usedpc = (zspage_size - waste) * 100 / zspage_size;
-
-		if (usedpc > max_usedpc) {
-			max_usedpc = usedpc;
-			max_usedpc_order = i;
-		}
-	}
-
-	return max_usedpc_order;
-}
-
 static struct zspage *get_zspage(struct page *page)
 {
 	struct zspage *zspage = (struct zspage *)page->private;
@@ -2312,6 +2276,24 @@ static int zs_register_shrinker(struct zs_pool *pool)
 	return register_shrinker(&pool->shrinker);
 }
 
+static int calculate_zspage_chain_size(int class_size)
+{
+	int i, min_waste = INT_MAX;
+	int chain_size = 1;
+
+	for (i = 1; i <= ZS_MAX_PAGES_PER_ZSPAGE; i++) {
+		int waste;
+
+		waste = (i * PAGE_SIZE) % class_size;
+		if (waste < min_waste) {
+			min_waste = waste;
+			chain_size = i;
+		}
+	}
+
+	return chain_size;
+}
+
 /**
  * zs_create_pool - Creates an allocation pool to work from.
  * @name: pool name to be created
@@ -2356,7 +2338,7 @@ struct zs_pool *zs_create_pool(const char *name)
 		size = ZS_MIN_ALLOC_SIZE + i * ZS_SIZE_CLASS_DELTA;
 		if (size > ZS_MAX_ALLOC_SIZE)
 			size = ZS_MAX_ALLOC_SIZE;
-		pages_per_zspage = get_pages_per_zspage(size);
+		pages_per_zspage = calculate_zspage_chain_size(size);
 		objs_per_zspage = pages_per_zspage * PAGE_SIZE / size;
 
 		/*