linux-kernel - [PATCHv2 2/7] zsmalloc: re-shuffle zs

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20250130044455.2642465-3-senozhatsky@chromium.org>
Date: Thu, 30 Jan 2025 13:42:45 +0900
From: Sergey Senozhatsky <senozhatsky@...omium.org>
To: Andrew Morton <akpm@...ux-foundation.org>,
	Yosry Ahmed <yosry.ahmed@...ux.dev>
Cc: Minchan Kim <minchan@...nel.org>,
	Johannes Weiner <hannes@...xchg.org>,
	Nhat Pham <nphamcs@...il.com>,
	Uros Bizjak <ubizjak@...il.com>,
	linux-mm@...ck.org,
	linux-kernel@...r.kernel.org,
	Sergey Senozhatsky <senozhatsky@...omium.org>
Subject: [PATCHv2 2/7] zsmalloc: re-shuffle zs_pool members

There is no need in having ->name at offset 0, it's never user and is
being followed but a very large RO (after init) array of size classes,
so each time we need to lock pool->migrate_lock (which happens relatively
often) we need to jump over that RO region.  Swap ->migrate_lock and ->name.

Move ->pages_allocated is modified relatively often, move it up
right after RO size_classes arrays.

We modify ->compaction_in_progress and ->stats only from compaction
(defragmentation), which can run only on one CPU at a time and,
depending on the case, may or may not be common.  Move both to down.

Signed-off-by: Sergey Senozhatsky <senozhatsky@...omium.org>
---
 mm/zsmalloc.c | 20 +++++++++++---------
 1 file changed, 11 insertions(+), 9 deletions(-)

diff --git a/mm/zsmalloc.c b/mm/zsmalloc.c
index 8972b4b56cd3..2280ea17796b 100644
--- a/mm/zsmalloc.c
+++ b/mm/zsmalloc.c
@@ -204,18 +204,15 @@ struct link_free {
 };
 
 struct zs_pool {
-	const char *name;
+	/* protect page/zspage migration */
+	rwlock_t migrate_lock;
 
 	struct size_class *size_class[ZS_SIZE_CLASSES];
-	struct kmem_cache *handle_cachep;
-	struct kmem_cache *zspage_cachep;
 
 	atomic_long_t pages_allocated;
 
-	struct zs_pool_stats stats;
-
-	/* Compact classes */
-	struct shrinker *shrinker;
+	struct kmem_cache *handle_cachep;
+	struct kmem_cache *zspage_cachep;
 
 #ifdef CONFIG_ZSMALLOC_STAT
 	struct dentry *stat_dentry;
@@ -223,9 +220,14 @@ struct zs_pool {
 #ifdef CONFIG_COMPACTION
 	struct work_struct free_work;
 #endif
-	/* protect page/zspage migration */
-	rwlock_t migrate_lock;
+	const char *name;
+
+	/* Compact classes */
+	struct shrinker *shrinker;
+	/* Permit only once compaction at a time */
 	atomic_t compaction_in_progress;
+	/* compaction stats */
+	struct zs_pool_stats stats;
 };
 
 static void pool_write_unlock(struct zs_pool *pool)
-- 
2.48.1.262.g85cc9f2d1e-goog