Revert "memcg: enable accounting for file lock caches"
authorLinus Torvalds <torvalds@linux-foundation.org>
Tue, 7 Sep 2021 18:21:48 +0000 (11:21 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Tue, 7 Sep 2021 18:21:48 +0000 (11:21 -0700)
This reverts commit 0f12156dff2862ac54235fc72703f18770769042.

The kernel test robot reports a sizeable performance regression for this
commit, and while it clearly does the rigth thing in theory, we'll need
to look at just how to avoid or minimize the performance overhead of the
memcg accounting.

People already have suggestions on how to do that, but it's "future
work".

So revert it for now.

Link: https://lore.kernel.org/lkml/20210907150757.GE17617@xsang-OptiPlex-9020/
Acked-by: Jens Axboe <axboe@kernel.dk>
Acked-by: Shakeel Butt <shakeelb@google.com>
Acked-by: Roman Gushchin <guro@fb.com>
Cc: Tejun Heo <tj@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
fs/locks.c

index 51a5b72ef3027d54ed85fcf4592801e0ac9abacc..3d6fb4ae847b4c6084c4f9745f2effce71ea6d78 100644 (file)
@@ -2941,12 +2941,10 @@ static int __init filelock_init(void)
        int i;
 
        flctx_cache = kmem_cache_create("file_lock_ctx",
-                       sizeof(struct file_lock_context), 0,
-                       SLAB_PANIC | SLAB_ACCOUNT, NULL);
+                       sizeof(struct file_lock_context), 0, SLAB_PANIC, NULL);
 
        filelock_cache = kmem_cache_create("file_lock_cache",
-                       sizeof(struct file_lock), 0,
-                       SLAB_PANIC | SLAB_ACCOUNT, NULL);
+                       sizeof(struct file_lock), 0, SLAB_PANIC, NULL);
 
        for_each_possible_cpu(i) {
                struct file_lock_list_struct *fll = per_cpu_ptr(&file_lock_list, i);