From patchwork Thu Feb 15 21:14:57 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Jianfeng Wang X-Patchwork-Id: 13559153 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by smtp.lore.kernel.org (Postfix) with ESMTP id 4DA30C4829E for ; Thu, 15 Feb 2024 21:15:09 +0000 (UTC) Received: by kanga.kvack.org (Postfix) id D30496B0081; Thu, 15 Feb 2024 16:15:08 -0500 (EST) Received: by kanga.kvack.org (Postfix, from userid 40) id CE0066B0082; Thu, 15 Feb 2024 16:15:08 -0500 (EST) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id B80956B0087; Thu, 15 Feb 2024 16:15:08 -0500 (EST) X-Delivered-To: linux-mm@kvack.org Received: from relay.hostedemail.com (smtprelay0013.hostedemail.com [216.40.44.13]) by kanga.kvack.org (Postfix) with ESMTP id A4C4C6B0081 for ; Thu, 15 Feb 2024 16:15:08 -0500 (EST) Received: from smtpin16.hostedemail.com (a10.router.float.18 [10.200.18.1]) by unirelay09.hostedemail.com (Postfix) with ESMTP id 75C12801C5 for ; Thu, 15 Feb 2024 21:15:08 +0000 (UTC) X-FDA: 81795293496.16.1C94259 Received: from mx0b-00069f02.pphosted.com (mx0b-00069f02.pphosted.com [205.220.177.32]) by imf05.hostedemail.com (Postfix) with ESMTP id 8777B10000C for ; Thu, 15 Feb 2024 21:15:05 +0000 (UTC) Authentication-Results: imf05.hostedemail.com; dkim=pass header.d=oracle.com header.s=corp-2023-11-20 header.b=mQ5zcsdz; dmarc=pass (policy=none) header.from=oracle.com; spf=pass (imf05.hostedemail.com: domain of jianfeng.w.wang@oracle.com designates 205.220.177.32 as permitted sender) smtp.mailfrom=jianfeng.w.wang@oracle.com ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=hostedemail.com; s=arc-20220608; t=1708031705; h=from:from:sender:reply-to:subject:subject:date:date: message-id:message-id:to:to:cc:cc:mime-version:mime-version: content-type:content-transfer-encoding:content-transfer-encoding: in-reply-to:references:dkim-signature; bh=QwOz67E/OHYRy6Ud5RwkMTPCcemq8UDvvWptjvYQl9I=; b=AEnybKM9kCquH3kB7rpu3jrWXC/gd4wYRZKP0KeROPJWEWYqtXWL8OtIn7dxntH3Yq/Pi6 +Bpmx0RshOjjcUJxTHD9XWAJkjjIrW8nrdP1LbXLeGZjXURYlkxfenIyC82ch8LuIm3ijY VnTHx5rflsuf1DfEhNWrYiNcq9VI558= ARC-Authentication-Results: i=1; imf05.hostedemail.com; dkim=pass header.d=oracle.com header.s=corp-2023-11-20 header.b=mQ5zcsdz; dmarc=pass (policy=none) header.from=oracle.com; spf=pass (imf05.hostedemail.com: domain of jianfeng.w.wang@oracle.com designates 205.220.177.32 as permitted sender) smtp.mailfrom=jianfeng.w.wang@oracle.com ARC-Seal: i=1; s=arc-20220608; d=hostedemail.com; t=1708031705; a=rsa-sha256; cv=none; b=Tq6L1sWgteneLJaM6oT88WwBlZIRAGFk7nzcbtoF5pN65PSB9La+vlbGCqadoWrFpBukeT YUqpmOfYVpNJEA819xH53+C2JGEb37t6mooJQ4KJZUOIL+W2V+uOoH2rekIYJSThyHhsDs pUkHdJgMvkT7kBuZxi4IU5kO7SebeAE= Received: from pps.filterd (m0246631.ppops.net [127.0.0.1]) by mx0b-00069f02.pphosted.com (8.17.1.19/8.17.1.19) with ESMTP id 41FFStrN006126; Thu, 15 Feb 2024 21:15:00 GMT DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=oracle.com; h=from : to : cc : subject : date : message-id : mime-version : content-transfer-encoding; s=corp-2023-11-20; bh=QwOz67E/OHYRy6Ud5RwkMTPCcemq8UDvvWptjvYQl9I=; b=mQ5zcsdzKlohCg/+z3teCpPZbiHZPLsZ/yWSRnNfyg2BsZxc1bD/yBdnM6KrFmbGCdUQ xRyWIjW8BrzAhg6ij7wwIyRfFu4Pm268HVzAH6F2xAHrRCrFtWHx/SeIf5fL+6gRC3zw huD2xhcr8qSGsBY01xula1KOm93HjFbINlf1rXrgUXS0LJthA0Zrp1wINUSsdaodUwX0 1zy7L8lK9L3YEyDoTFDjJoUVfaAbOHvgiBkPRzAQ2LXDT0FU6CQQe7jeOtpotxgyY+0p 4AKl99TroLiFIu67olqW+k/Pt9FkzZae/AescBiyzCHgblXuB/vgV+HU+wEC+R1ipzu8 SA== Received: from phxpaimrmta02.imrmtpd1.prodappphxaev1.oraclevcn.com (phxpaimrmta02.appoci.oracle.com [147.154.114.232]) by mx0b-00069f02.pphosted.com (PPS) with ESMTPS id 3w930139ft-1 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=OK); Thu, 15 Feb 2024 21:15:00 +0000 Received: from pps.filterd (phxpaimrmta02.imrmtpd1.prodappphxaev1.oraclevcn.com [127.0.0.1]) by phxpaimrmta02.imrmtpd1.prodappphxaev1.oraclevcn.com (8.17.1.19/8.17.1.19) with ESMTP id 41FL9JMI014984; Thu, 15 Feb 2024 21:14:59 GMT Received: from pps.reinject (localhost [127.0.0.1]) by phxpaimrmta02.imrmtpd1.prodappphxaev1.oraclevcn.com (PPS) with ESMTPS id 3w5ykav2pa-1 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=OK); Thu, 15 Feb 2024 21:14:59 +0000 Received: from phxpaimrmta02.imrmtpd1.prodappphxaev1.oraclevcn.com (phxpaimrmta02.imrmtpd1.prodappphxaev1.oraclevcn.com [127.0.0.1]) by pps.reinject (8.17.1.5/8.17.1.5) with ESMTP id 41FLExx9014286; Thu, 15 Feb 2024 21:14:59 GMT Received: from jfwang-mac.us.oracle.com (dhcp-10-159-156-163.vpn.oracle.com [10.159.156.163]) by phxpaimrmta02.imrmtpd1.prodappphxaev1.oraclevcn.com (PPS) with ESMTP id 3w5ykav2nh-1; Thu, 15 Feb 2024 21:14:59 +0000 From: Jianfeng Wang To: cl@linux.com, penberg@kernel.org, rientjes@google.com, iamjoonsoo.kim@lge.com, akpm@linux-foundation.org, vbabka@suse.cz, roman.gushchin@linux.dev, 42.hyeyoo@gmail.com Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org, Jianfeng Wang Subject: [PATCH] slub: avoid scanning all partial slabs in get_slabinfo() Date: Thu, 15 Feb 2024 13:14:57 -0800 Message-ID: <20240215211457.32172-1-jianfeng.w.wang@oracle.com> X-Mailer: git-send-email 2.42.1 MIME-Version: 1.0 X-Proofpoint-Virus-Version: vendor=baseguard engine=ICAP:2.0.272,Aquarius:18.0.1011,Hydra:6.0.619,FMLib:17.11.176.26 definitions=2024-02-15_20,2024-02-14_01,2023-05-22_02 X-Proofpoint-Spam-Details: rule=notspam policy=default score=0 adultscore=0 spamscore=0 mlxlogscore=999 bulkscore=0 phishscore=0 mlxscore=0 suspectscore=0 malwarescore=0 classifier=spam adjust=0 reason=mlx scancount=1 engine=8.12.0-2311290000 definitions=main-2402150168 X-Proofpoint-GUID: wUxNxrDlcZA1LNy_BFJnVnTGBApNpkWc X-Proofpoint-ORIG-GUID: wUxNxrDlcZA1LNy_BFJnVnTGBApNpkWc X-Rspam-User: X-Stat-Signature: pnas3mm3otf9iobbh745r83kcpemxt9x X-Rspamd-Server: rspam07 X-Rspamd-Queue-Id: 8777B10000C X-HE-Tag: 1708031705-226524 X-HE-Meta: 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 ImDxe+ou SbYIDYYPIaoNoF5ZjY9JH5jwzGzKvKATqj/+72H11YCoynRpJvOmF3LaqLyMeZUchhTS2jxR/lQLLLvM+wq93Nl96+0XLefA0KPdiE9LGrF2l215Oh7d+K1fNEjZ8RUqRhA8L85Eo+i8JAlM2paKa9iszA/ixzHQDLy30PYNu3aabOkgWYsAVPsXhNj20YOhrsn3iKT3OUThV0QyF+n/W7iwEX0s9yGH/MCR9rWHImTTb/sl+mY3mN9ACLslTANNmDT7OpaeXE59EpAZ4B3DDRCq3C2B7eGTy9BQO5BQWu4BcIPuctz7dFFLDNWx+CZIjKE93cQFe87kKDmltdjH7yO2une8rYWIjp3w66Mrj8U414S30gFI5tyQfoZf0bUtteRZHJt735rPkFZAQYACP85U5U1jY6LfwXc2T4LzT4yogNc5js2fRgFSYPm5zIjn92q8U X-Bogosity: Ham, tests=bogofilter, spamicity=0.000000, version=1.2.4 Sender: owner-linux-mm@kvack.org Precedence: bulk X-Loop: owner-majordomo@kvack.org List-ID: List-Subscribe: List-Unsubscribe: When reading "/proc/slabinfo", the kernel needs to report the number of free objects for each kmem_cache. The current implementation relies on count_partial() that counts the number of free objects by scanning each kmem_cache_node's partial slab list and summing free objects from all partial slabs in the list. This process must hold per kmem_cache_node spinlock and disable IRQ. Consequently, it can block slab allocation requests on other CPU cores and cause timeouts for network devices etc., if the partial slab list is long. In production, even NMI watchdog can be triggered because some slab caches have a long partial list: e.g., for "buffer_head", the number of partial slabs was observed to be ~1M in one kmem_cache_node. This problem was also observed by several others [1-2] in the past. The fix is to maintain a counter of free objects for each kmem_cache. Then, in get_slabinfo(), use the counter rather than count_partial() when reporting the number of free objects for a slab cache. per-cpu counter is used to minimize atomic or lock operation. Benchmark: run hackbench on a dual-socket 72-CPU bare metal machine with 256 GB memory and Intel(R) Xeon(R) CPU E5-2699 v3 @ 2.3 GHz. The command is "hackbench 18 thread 20000". Each group gets 10 runs. Results: - Mainline: 21.0381 +- 0.0325 seconds time elapsed ( +- 0.15% ) - Mainline w/ this patch: 21.1878 +- 0.0239 seconds time elapsed ( +- 0.11% ) [1] https://lore.kernel.org/linux-mm/ alpine.DEB.2.21.2003031602460.1537@www.lameter.com/T/ [2] https://lore.kernel.org/lkml/ alpine.DEB.2.22.394.2008071258020.55871@www.lameter.com/T/ Signed-off-by: Jianfeng Wang --- mm/slab.h | 4 ++++ mm/slub.c | 31 +++++++++++++++++++++++++++++-- 2 files changed, 33 insertions(+), 2 deletions(-) diff --git a/mm/slab.h b/mm/slab.h index 54deeb0428c6..a0e7672ba648 100644 --- a/mm/slab.h +++ b/mm/slab.h @@ -11,6 +11,7 @@ #include #include #include +#include /* * Internal slab definitions @@ -277,6 +278,9 @@ struct kmem_cache { unsigned int red_left_pad; /* Left redzone padding size */ const char *name; /* Name (only for display!) */ struct list_head list; /* List of slab caches */ +#ifdef CONFIG_SLUB_DEBUG + struct percpu_counter free_objects; +#endif #ifdef CONFIG_SYSFS struct kobject kobj; /* For sysfs */ #endif diff --git a/mm/slub.c b/mm/slub.c index 2ef88bbf56a3..44f8ded96574 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -736,6 +736,12 @@ static inline bool slab_update_freelist(struct kmem_cache *s, struct slab *slab, static unsigned long object_map[BITS_TO_LONGS(MAX_OBJS_PER_PAGE)]; static DEFINE_SPINLOCK(object_map_lock); +static inline void +__update_kmem_cache_free_objs(struct kmem_cache *s, s64 delta) +{ + percpu_counter_add_batch(&s->free_objects, delta, INT_MAX); +} + static void __fill_map(unsigned long *obj_map, struct kmem_cache *s, struct slab *slab) { @@ -1829,6 +1835,9 @@ slab_flags_t kmem_cache_flags(unsigned int object_size, return flags | slub_debug_local; } #else /* !CONFIG_SLUB_DEBUG */ +static inline void +__update_kmem_cache_free_objs(struct kmem_cache *s, s64 delta) {} + static inline void setup_object_debug(struct kmem_cache *s, void *object) {} static inline void setup_slab_debug(struct kmem_cache *s, struct slab *slab, void *addr) {} @@ -2369,6 +2378,7 @@ static struct slab *allocate_slab(struct kmem_cache *s, gfp_t flags, int node) slab->inuse = 0; slab->frozen = 0; + __update_kmem_cache_free_objs(s, slab->objects); account_slab(slab, oo_order(oo), s, flags); slab->slab_cache = s; @@ -2445,6 +2455,7 @@ static void free_slab(struct kmem_cache *s, struct slab *slab) call_rcu(&slab->rcu_head, rcu_free_slab); else __free_slab(s, slab); + __update_kmem_cache_free_objs(s, -slab->objects); } static void discard_slab(struct kmem_cache *s, struct slab *slab) @@ -3859,6 +3870,8 @@ static __fastpath_inline void *slab_alloc_node(struct kmem_cache *s, struct list */ slab_post_alloc_hook(s, objcg, gfpflags, 1, &object, init, orig_size); + if (object) + __update_kmem_cache_free_objs(s, -1); return object; } @@ -4235,6 +4248,7 @@ static __always_inline void do_slab_free(struct kmem_cache *s, unsigned long tid; void **freelist; + __update_kmem_cache_free_objs(s, cnt); redo: /* * Determine the currently cpus per cpu slab. @@ -4286,6 +4300,7 @@ static void do_slab_free(struct kmem_cache *s, struct slab *slab, void *head, void *tail, int cnt, unsigned long addr) { + __update_kmem_cache_free_objs(s, cnt); __slab_free(s, slab, head, tail, cnt, addr); } #endif /* CONFIG_SLUB_TINY */ @@ -4658,6 +4673,7 @@ int kmem_cache_alloc_bulk(struct kmem_cache *s, gfp_t flags, size_t size, memcg_slab_alloc_error_hook(s, size, objcg); } + __update_kmem_cache_free_objs(s, -i); return i; } EXPORT_SYMBOL(kmem_cache_alloc_bulk); @@ -4899,6 +4915,9 @@ void __kmem_cache_release(struct kmem_cache *s) cache_random_seq_destroy(s); #ifndef CONFIG_SLUB_TINY free_percpu(s->cpu_slab); +#endif +#ifdef CONFIG_SLUB_DEBUG + percpu_counter_destroy(&s->free_objects); #endif free_kmem_cache_nodes(s); } @@ -5109,6 +5128,14 @@ static int kmem_cache_open(struct kmem_cache *s, slab_flags_t flags) s->random = get_random_long(); #endif +#ifdef CONFIG_SLUB_DEBUG + int ret; + + ret = percpu_counter_init(&s->free_objects, 0, GFP_KERNEL); + if (ret) + return ret; +#endif + if (!calculate_sizes(s)) goto error; if (disable_higher_order_debug) { @@ -7100,15 +7127,15 @@ void get_slabinfo(struct kmem_cache *s, struct slabinfo *sinfo) { unsigned long nr_slabs = 0; unsigned long nr_objs = 0; - unsigned long nr_free = 0; + unsigned long nr_free; int node; struct kmem_cache_node *n; for_each_kmem_cache_node(s, node, n) { nr_slabs += node_nr_slabs(n); nr_objs += node_nr_objs(n); - nr_free += count_partial(n, count_free); } + nr_free = percpu_counter_sum_positive(&s->free_objects); sinfo->active_objs = nr_objs - nr_free; sinfo->num_objs = nr_objs;