From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-23.3 required=3.0 tests=BAYES_00,DKIMWL_WL_MED, DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_CR_TRAILER,INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS, USER_IN_DEF_DKIM_WL autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 109A7C433B4 for ; Mon, 12 Apr 2021 23:11:12 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id DE3F761278 for ; Mon, 12 Apr 2021 23:11:11 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S243594AbhDLXL3 (ORCPT ); Mon, 12 Apr 2021 19:11:29 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:56656 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S237900AbhDLXLZ (ORCPT ); Mon, 12 Apr 2021 19:11:25 -0400 Received: from mail-lf1-x132.google.com (mail-lf1-x132.google.com [IPv6:2a00:1450:4864:20::132]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id D8D9CC061574 for ; Mon, 12 Apr 2021 16:11:06 -0700 (PDT) Received: by mail-lf1-x132.google.com with SMTP id f17so17054031lfu.7 for ; Mon, 12 Apr 2021 16:11:06 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20161025; h=mime-version:references:in-reply-to:from:date:message-id:subject:to :cc; bh=EB6M69ykRfOk8CNylVX2Xu1iWxE7slXxBQnBQQyoh0Q=; b=m/kjHFSDwzv3BbnwILdEAMEu34wkDzXHctkxqs41nUPCderXN+P0otyZ/WoqmpvyaW TYIBp8v6+r+6x5LJacRAqdsbIJCAgjXrYssLGqWZCmUb9eGhrTCTMWAcgzxg9lIU5qj1 F4ZfuoVmEuJ3bfPsaJ43rYq8BkCdkd+bIKPz/jUO37ijQy9qEAcEMDM01kcyIHWlFcDF UMTlwSGJGeuuYBZ/eUPPOybdNcFMZQ8AnNR6kErXat2ZPeYVRNMSjTlt8Kt+CBTOTqU+ jT2R/iUAnGSvAbYZxeVSbqND2Ldcavcbm/1ZTIUb7J2HkhjcfvL3t2C9zBRWfilXTSEm KYhQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:mime-version:references:in-reply-to:from:date :message-id:subject:to:cc; bh=EB6M69ykRfOk8CNylVX2Xu1iWxE7slXxBQnBQQyoh0Q=; b=OvpZXRW8NIUUMMELXkd34ntZtd000zExVN7wn0tYhMnPNXskKAHFDZNb/SN4hZvqQn oFXLQBQxmc9dmL6n/3icS4uNpeMt6Ze1SoiqiAhzMUBbIsYrp81sCkbOtKePOFdampc6 PX5wCrXglkThNEsPWRIuU5BsBg9AfBFNtokNfTetBC/4zkGXSvFWR6hrW7nv2KLNujE6 qWV/3waZTz/EV0HFWuHfakgnFMQdjzYkIE3L0UlVm1Cd6RyNup/NFbU09RxKAOlKfgRt eDqYG/FUy94qEFge/KCv+28tgHIndXnsvk5ItWFhy8IxTRxViortF7hBGS5RNTVj70PY XAZw== X-Gm-Message-State: AOAM53311W5mhF5fKO1hTX+9O9Jg0VNuMBIAQe6pEQkuhGynq7TVIbyd Fb1+FxWJkYRcIOVbnaa5nLaNaO9n0bMGOH8JXKr6kw== X-Google-Smtp-Source: ABdhPJzMX4OiuvyU6Y9R8hJ7f8iPiZlikv/N+32inGJEK1DkwPToxpRVbqW/4cVoydeOpfUpnT90qqYGopPMYu+cAXQ= X-Received: by 2002:a05:6512:2019:: with SMTP id a25mr11022613lfb.347.1618269065084; Mon, 12 Apr 2021 16:11:05 -0700 (PDT) MIME-Version: 1.0 References: <20210412225503.15119-1-longman@redhat.com> <20210412225503.15119-6-longman@redhat.com> In-Reply-To: <20210412225503.15119-6-longman@redhat.com> From: Shakeel Butt Date: Mon, 12 Apr 2021 16:10:53 -0700 Message-ID: Subject: Re: [PATCH v2 5/5] mm/memcg: Optimize user context object stock access To: Waiman Long Cc: Johannes Weiner , Michal Hocko , Vladimir Davydov , Andrew Morton , Tejun Heo , Christoph Lameter , Pekka Enberg , David Rientjes , Joonsoo Kim , Vlastimil Babka , Roman Gushchin , LKML , Cgroups , Linux MM , Muchun Song , Alex Shi , Chris Down , Yafang Shao , Wei Yang , Masayoshi Mizuma , Xing Zhengjun Content-Type: text/plain; charset="UTF-8" Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Mon, Apr 12, 2021 at 3:55 PM Waiman Long wrote: > > Most kmem_cache_alloc() calls are from user context. With instrumentation > enabled, the measured amount of kmem_cache_alloc() calls from non-task > context was about 0.01% of the total. > > The irq disable/enable sequence used in this case to access content > from object stock is slow. To optimize for user context access, there > are now two object stocks for task context and interrupt context access > respectively. > > The task context object stock can be accessed after disabling preemption > which is cheap in non-preempt kernel. The interrupt context object stock > can only be accessed after disabling interrupt. User context code can > access interrupt object stock, but not vice versa. > > The mod_objcg_state() function is also modified to make sure that memcg > and lruvec stat updates are done with interrupted disabled. > > The downside of this change is that there are more data stored in local > object stocks and not reflected in the charge counter and the vmstat > arrays. However, this is a small price to pay for better performance. > > Signed-off-by: Waiman Long > Acked-by: Roman Gushchin > --- > mm/memcontrol.c | 73 +++++++++++++++++++++++++++++++++++++++---------- > 1 file changed, 59 insertions(+), 14 deletions(-) > > diff --git a/mm/memcontrol.c b/mm/memcontrol.c > index 69f728383efe..29f2df76644a 100644 > --- a/mm/memcontrol.c > +++ b/mm/memcontrol.c > @@ -2229,7 +2229,8 @@ struct obj_stock { > struct memcg_stock_pcp { > struct mem_cgroup *cached; /* this never be root cgroup */ > unsigned int nr_pages; > - struct obj_stock obj; > + struct obj_stock task_obj; > + struct obj_stock irq_obj; > > struct work_struct work; > unsigned long flags; > @@ -2254,11 +2255,48 @@ static bool obj_stock_flush_required(struct memcg_stock_pcp *stock, > } > #endif > > +/* > + * Most kmem_cache_alloc() calls are from user context. The irq disable/enable > + * sequence used in this case to access content from object stock is slow. > + * To optimize for user context access, there are now two object stocks for > + * task context and interrupt context access respectively. > + * > + * The task context object stock can be accessed by disabling preemption only > + * which is cheap in non-preempt kernel. The interrupt context object stock > + * can only be accessed after disabling interrupt. User context code can > + * access interrupt object stock, but not vice versa. > + */ > static inline struct obj_stock *current_obj_stock(void) > { > struct memcg_stock_pcp *stock = this_cpu_ptr(&memcg_stock); > > - return &stock->obj; > + return in_task() ? &stock->task_obj : &stock->irq_obj; > +} > + > +#define get_obj_stock(flags) \ > +({ \ > + struct memcg_stock_pcp *stock; \ > + struct obj_stock *obj_stock; \ > + \ > + if (in_task()) { \ > + preempt_disable(); \ > + (flags) = -1L; \ > + stock = this_cpu_ptr(&memcg_stock); \ The above line was missing in the previous version. > + obj_stock = &stock->task_obj; \ > + } else { \ > + local_irq_save(flags); \ > + stock = this_cpu_ptr(&memcg_stock); \ > + obj_stock = &stock->irq_obj; \ > + } \ > + obj_stock; \ > +}) > + > +static inline void put_obj_stock(unsigned long flags) > +{ > + if (flags == -1L) > + preempt_enable(); > + else > + local_irq_restore(flags); > } > > /** > @@ -2327,7 +2365,9 @@ static void drain_local_stock(struct work_struct *dummy) > local_irq_save(flags); > > stock = this_cpu_ptr(&memcg_stock); > - drain_obj_stock(&stock->obj); > + drain_obj_stock(&stock->irq_obj); > + if (in_task()) > + drain_obj_stock(&stock->task_obj); > drain_stock(stock); > clear_bit(FLUSHING_CACHED_CHARGE, &stock->flags); > > @@ -3183,7 +3223,7 @@ static inline void mod_objcg_state(struct obj_cgroup *objcg, > memcg = obj_cgroup_memcg(objcg); > if (pgdat) > lruvec = mem_cgroup_lruvec(memcg, pgdat); > - __mod_memcg_lruvec_state(memcg, lruvec, idx, nr); > + mod_memcg_lruvec_state(memcg, lruvec, idx, nr); > rcu_read_unlock(); > } > > @@ -3193,7 +3233,7 @@ static bool consume_obj_stock(struct obj_cgroup *objcg, unsigned int nr_bytes) > unsigned long flags; > bool ret = false; > > - local_irq_save(flags); > + stock = get_obj_stock(flags); > > stock = current_obj_stock(); The above is redundant. > if (objcg == stock->cached_objcg && stock->nr_bytes >= nr_bytes) { > @@ -3201,7 +3241,7 @@ static bool consume_obj_stock(struct obj_cgroup *objcg, unsigned int nr_bytes) > ret = true; > } > > - local_irq_restore(flags); > + put_obj_stock(flags); > > return ret; > } > @@ -3254,8 +3294,13 @@ static bool obj_stock_flush_required(struct memcg_stock_pcp *stock, > { > struct mem_cgroup *memcg; > > - if (stock->obj.cached_objcg) { > - memcg = obj_cgroup_memcg(stock->obj.cached_objcg); > + if (in_task() && stock->task_obj.cached_objcg) { > + memcg = obj_cgroup_memcg(stock->task_obj.cached_objcg); > + if (memcg && mem_cgroup_is_descendant(memcg, root_memcg)) > + return true; > + } > + if (stock->irq_obj.cached_objcg) { > + memcg = obj_cgroup_memcg(stock->irq_obj.cached_objcg); > if (memcg && mem_cgroup_is_descendant(memcg, root_memcg)) > return true; > } > @@ -3283,9 +3328,9 @@ static void refill_obj_stock(struct obj_cgroup *objcg, unsigned int nr_bytes) > { > unsigned long flags; > > - local_irq_save(flags); > + get_obj_stock(flags); > __refill_obj_stock(objcg, nr_bytes); > - local_irq_restore(flags); > + put_obj_stock(flags); > } > > static void __mod_obj_stock_state(struct obj_cgroup *objcg, > @@ -3325,9 +3370,9 @@ void mod_obj_stock_state(struct obj_cgroup *objcg, struct pglist_data *pgdat, > { > unsigned long flags; > > - local_irq_save(flags); > + get_obj_stock(flags); > __mod_obj_stock_state(objcg, pgdat, idx, nr); > - local_irq_restore(flags); > + put_obj_stock(flags); > } > > int obj_cgroup_charge(struct obj_cgroup *objcg, gfp_t gfp, size_t size) > @@ -3380,10 +3425,10 @@ void obj_cgroup_uncharge_mod_state(struct obj_cgroup *objcg, size_t size, > { > unsigned long flags; > > - local_irq_save(flags); > + get_obj_stock(flags); > __refill_obj_stock(objcg, size); > __mod_obj_stock_state(objcg, pgdat, idx, -(int)size); > - local_irq_restore(flags); > + put_obj_stock(flags); > } > > #endif /* CONFIG_MEMCG_KMEM */ > -- > 2.18.1 > From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-23.3 required=3.0 tests=BAYES_00,DKIMWL_WL_MED, DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_CR_TRAILER,INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS, USER_IN_DEF_DKIM_WL autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id C62AEC433ED for ; Mon, 12 Apr 2021 23:11:08 +0000 (UTC) Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.kernel.org (Postfix) with ESMTP id 2B7C061287 for ; Mon, 12 Apr 2021 23:11:08 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 2B7C061287 Authentication-Results: mail.kernel.org; dmarc=fail (p=reject dis=none) header.from=google.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=owner-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix) id 9EBE06B0036; Mon, 12 Apr 2021 19:11:07 -0400 (EDT) Received: by kanga.kvack.org (Postfix, from userid 40) id 99B6B6B006E; Mon, 12 Apr 2021 19:11:07 -0400 (EDT) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 8188E6B0070; Mon, 12 Apr 2021 19:11:07 -0400 (EDT) X-Delivered-To: linux-mm@kvack.org Received: from forelay.hostedemail.com (smtprelay0023.hostedemail.com [216.40.44.23]) by kanga.kvack.org (Postfix) with ESMTP id 625ED6B0036 for ; Mon, 12 Apr 2021 19:11:07 -0400 (EDT) Received: from smtpin23.hostedemail.com (10.5.19.251.rfc1918.com [10.5.19.251]) by forelay04.hostedemail.com (Postfix) with ESMTP id 1C4E9246A for ; Mon, 12 Apr 2021 23:11:07 +0000 (UTC) X-FDA: 78025262574.23.83B6F60 Received: from mail-lf1-f46.google.com (mail-lf1-f46.google.com [209.85.167.46]) by imf25.hostedemail.com (Postfix) with ESMTP id 8ED1D6000108 for ; Mon, 12 Apr 2021 23:11:04 +0000 (UTC) Received: by mail-lf1-f46.google.com with SMTP id b4so24088721lfi.6 for ; Mon, 12 Apr 2021 16:11:06 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20161025; h=mime-version:references:in-reply-to:from:date:message-id:subject:to :cc; bh=EB6M69ykRfOk8CNylVX2Xu1iWxE7slXxBQnBQQyoh0Q=; b=m/kjHFSDwzv3BbnwILdEAMEu34wkDzXHctkxqs41nUPCderXN+P0otyZ/WoqmpvyaW TYIBp8v6+r+6x5LJacRAqdsbIJCAgjXrYssLGqWZCmUb9eGhrTCTMWAcgzxg9lIU5qj1 F4ZfuoVmEuJ3bfPsaJ43rYq8BkCdkd+bIKPz/jUO37ijQy9qEAcEMDM01kcyIHWlFcDF UMTlwSGJGeuuYBZ/eUPPOybdNcFMZQ8AnNR6kErXat2ZPeYVRNMSjTlt8Kt+CBTOTqU+ jT2R/iUAnGSvAbYZxeVSbqND2Ldcavcbm/1ZTIUb7J2HkhjcfvL3t2C9zBRWfilXTSEm KYhQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:mime-version:references:in-reply-to:from:date :message-id:subject:to:cc; bh=EB6M69ykRfOk8CNylVX2Xu1iWxE7slXxBQnBQQyoh0Q=; b=XLoPmhbcTbVucJ8VoyPQ0Z2zxU+72jWDHqWX3Vwm9RA4+vntru4IoUUm8lnTYMwNAs ASZeTgV7J+1KYrh2rpTY5pAw+D9MdviZJtQQhGmyYM13IfOGt208NHncjDC4EE+ZxQI4 L2zL3BXTl7JSTmgCncF1jcQ0lzn8j0Kn3z92m6OQQYcRVwFs/zy1CWlNBdwqGBhUjQ14 SMZs+kqKrCiVm4CPu2EEyvmzFEa3h+7fsg6ppc92ZK9RKwHgWH+6oy4R34QPecIhBTo1 5Fx9fBT844kWVdPkQYwLcltDzK7MR9dvs1OAP83rLy5Pu2Vd7jLd160LRAil4a4zCAmQ TiJA== X-Gm-Message-State: AOAM530BKgTVENjGI4MlJmuxjB7xtPgu2iNIqe7IOknfNxGhBRN3c65R RzQPaDwxP2q39yS0/6I2KZ+KXHFewc+4RtmieXGfXw== X-Google-Smtp-Source: ABdhPJzMX4OiuvyU6Y9R8hJ7f8iPiZlikv/N+32inGJEK1DkwPToxpRVbqW/4cVoydeOpfUpnT90qqYGopPMYu+cAXQ= X-Received: by 2002:a05:6512:2019:: with SMTP id a25mr11022613lfb.347.1618269065084; Mon, 12 Apr 2021 16:11:05 -0700 (PDT) MIME-Version: 1.0 References: <20210412225503.15119-1-longman@redhat.com> <20210412225503.15119-6-longman@redhat.com> In-Reply-To: <20210412225503.15119-6-longman@redhat.com> From: Shakeel Butt Date: Mon, 12 Apr 2021 16:10:53 -0700 Message-ID: Subject: Re: [PATCH v2 5/5] mm/memcg: Optimize user context object stock access To: Waiman Long Cc: Johannes Weiner , Michal Hocko , Vladimir Davydov , Andrew Morton , Tejun Heo , Christoph Lameter , Pekka Enberg , David Rientjes , Joonsoo Kim , Vlastimil Babka , Roman Gushchin , LKML , Cgroups , Linux MM , Muchun Song , Alex Shi , Chris Down , Yafang Shao , Wei Yang , Masayoshi Mizuma , Xing Zhengjun Content-Type: text/plain; charset="UTF-8" X-Stat-Signature: s8xnh4fdx7ze816m37didwijxw9huoa8 X-Rspamd-Server: rspam04 X-Rspamd-Queue-Id: 8ED1D6000108 Received-SPF: none (google.com>: No applicable sender policy available) receiver=imf25; identity=mailfrom; envelope-from=""; helo=mail-lf1-f46.google.com; client-ip=209.85.167.46 X-HE-DKIM-Result: pass/pass X-HE-Tag: 1618269064-22792 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: On Mon, Apr 12, 2021 at 3:55 PM Waiman Long wrote: > > Most kmem_cache_alloc() calls are from user context. With instrumentation > enabled, the measured amount of kmem_cache_alloc() calls from non-task > context was about 0.01% of the total. > > The irq disable/enable sequence used in this case to access content > from object stock is slow. To optimize for user context access, there > are now two object stocks for task context and interrupt context access > respectively. > > The task context object stock can be accessed after disabling preemption > which is cheap in non-preempt kernel. The interrupt context object stock > can only be accessed after disabling interrupt. User context code can > access interrupt object stock, but not vice versa. > > The mod_objcg_state() function is also modified to make sure that memcg > and lruvec stat updates are done with interrupted disabled. > > The downside of this change is that there are more data stored in local > object stocks and not reflected in the charge counter and the vmstat > arrays. However, this is a small price to pay for better performance. > > Signed-off-by: Waiman Long > Acked-by: Roman Gushchin > --- > mm/memcontrol.c | 73 +++++++++++++++++++++++++++++++++++++++---------- > 1 file changed, 59 insertions(+), 14 deletions(-) > > diff --git a/mm/memcontrol.c b/mm/memcontrol.c > index 69f728383efe..29f2df76644a 100644 > --- a/mm/memcontrol.c > +++ b/mm/memcontrol.c > @@ -2229,7 +2229,8 @@ struct obj_stock { > struct memcg_stock_pcp { > struct mem_cgroup *cached; /* this never be root cgroup */ > unsigned int nr_pages; > - struct obj_stock obj; > + struct obj_stock task_obj; > + struct obj_stock irq_obj; > > struct work_struct work; > unsigned long flags; > @@ -2254,11 +2255,48 @@ static bool obj_stock_flush_required(struct memcg_stock_pcp *stock, > } > #endif > > +/* > + * Most kmem_cache_alloc() calls are from user context. The irq disable/enable > + * sequence used in this case to access content from object stock is slow. > + * To optimize for user context access, there are now two object stocks for > + * task context and interrupt context access respectively. > + * > + * The task context object stock can be accessed by disabling preemption only > + * which is cheap in non-preempt kernel. The interrupt context object stock > + * can only be accessed after disabling interrupt. User context code can > + * access interrupt object stock, but not vice versa. > + */ > static inline struct obj_stock *current_obj_stock(void) > { > struct memcg_stock_pcp *stock = this_cpu_ptr(&memcg_stock); > > - return &stock->obj; > + return in_task() ? &stock->task_obj : &stock->irq_obj; > +} > + > +#define get_obj_stock(flags) \ > +({ \ > + struct memcg_stock_pcp *stock; \ > + struct obj_stock *obj_stock; \ > + \ > + if (in_task()) { \ > + preempt_disable(); \ > + (flags) = -1L; \ > + stock = this_cpu_ptr(&memcg_stock); \ The above line was missing in the previous version. > + obj_stock = &stock->task_obj; \ > + } else { \ > + local_irq_save(flags); \ > + stock = this_cpu_ptr(&memcg_stock); \ > + obj_stock = &stock->irq_obj; \ > + } \ > + obj_stock; \ > +}) > + > +static inline void put_obj_stock(unsigned long flags) > +{ > + if (flags == -1L) > + preempt_enable(); > + else > + local_irq_restore(flags); > } > > /** > @@ -2327,7 +2365,9 @@ static void drain_local_stock(struct work_struct *dummy) > local_irq_save(flags); > > stock = this_cpu_ptr(&memcg_stock); > - drain_obj_stock(&stock->obj); > + drain_obj_stock(&stock->irq_obj); > + if (in_task()) > + drain_obj_stock(&stock->task_obj); > drain_stock(stock); > clear_bit(FLUSHING_CACHED_CHARGE, &stock->flags); > > @@ -3183,7 +3223,7 @@ static inline void mod_objcg_state(struct obj_cgroup *objcg, > memcg = obj_cgroup_memcg(objcg); > if (pgdat) > lruvec = mem_cgroup_lruvec(memcg, pgdat); > - __mod_memcg_lruvec_state(memcg, lruvec, idx, nr); > + mod_memcg_lruvec_state(memcg, lruvec, idx, nr); > rcu_read_unlock(); > } > > @@ -3193,7 +3233,7 @@ static bool consume_obj_stock(struct obj_cgroup *objcg, unsigned int nr_bytes) > unsigned long flags; > bool ret = false; > > - local_irq_save(flags); > + stock = get_obj_stock(flags); > > stock = current_obj_stock(); The above is redundant. > if (objcg == stock->cached_objcg && stock->nr_bytes >= nr_bytes) { > @@ -3201,7 +3241,7 @@ static bool consume_obj_stock(struct obj_cgroup *objcg, unsigned int nr_bytes) > ret = true; > } > > - local_irq_restore(flags); > + put_obj_stock(flags); > > return ret; > } > @@ -3254,8 +3294,13 @@ static bool obj_stock_flush_required(struct memcg_stock_pcp *stock, > { > struct mem_cgroup *memcg; > > - if (stock->obj.cached_objcg) { > - memcg = obj_cgroup_memcg(stock->obj.cached_objcg); > + if (in_task() && stock->task_obj.cached_objcg) { > + memcg = obj_cgroup_memcg(stock->task_obj.cached_objcg); > + if (memcg && mem_cgroup_is_descendant(memcg, root_memcg)) > + return true; > + } > + if (stock->irq_obj.cached_objcg) { > + memcg = obj_cgroup_memcg(stock->irq_obj.cached_objcg); > if (memcg && mem_cgroup_is_descendant(memcg, root_memcg)) > return true; > } > @@ -3283,9 +3328,9 @@ static void refill_obj_stock(struct obj_cgroup *objcg, unsigned int nr_bytes) > { > unsigned long flags; > > - local_irq_save(flags); > + get_obj_stock(flags); > __refill_obj_stock(objcg, nr_bytes); > - local_irq_restore(flags); > + put_obj_stock(flags); > } > > static void __mod_obj_stock_state(struct obj_cgroup *objcg, > @@ -3325,9 +3370,9 @@ void mod_obj_stock_state(struct obj_cgroup *objcg, struct pglist_data *pgdat, > { > unsigned long flags; > > - local_irq_save(flags); > + get_obj_stock(flags); > __mod_obj_stock_state(objcg, pgdat, idx, nr); > - local_irq_restore(flags); > + put_obj_stock(flags); > } > > int obj_cgroup_charge(struct obj_cgroup *objcg, gfp_t gfp, size_t size) > @@ -3380,10 +3425,10 @@ void obj_cgroup_uncharge_mod_state(struct obj_cgroup *objcg, size_t size, > { > unsigned long flags; > > - local_irq_save(flags); > + get_obj_stock(flags); > __refill_obj_stock(objcg, size); > __mod_obj_stock_state(objcg, pgdat, idx, -(int)size); > - local_irq_restore(flags); > + put_obj_stock(flags); > } > > #endif /* CONFIG_MEMCG_KMEM */ > -- > 2.18.1 > From mboxrd@z Thu Jan 1 00:00:00 1970 From: Shakeel Butt Subject: Re: [PATCH v2 5/5] mm/memcg: Optimize user context object stock access Date: Mon, 12 Apr 2021 16:10:53 -0700 Message-ID: References: <20210412225503.15119-1-longman@redhat.com> <20210412225503.15119-6-longman@redhat.com> Mime-Version: 1.0 Return-path: DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20161025; h=mime-version:references:in-reply-to:from:date:message-id:subject:to :cc; bh=EB6M69ykRfOk8CNylVX2Xu1iWxE7slXxBQnBQQyoh0Q=; b=m/kjHFSDwzv3BbnwILdEAMEu34wkDzXHctkxqs41nUPCderXN+P0otyZ/WoqmpvyaW TYIBp8v6+r+6x5LJacRAqdsbIJCAgjXrYssLGqWZCmUb9eGhrTCTMWAcgzxg9lIU5qj1 F4ZfuoVmEuJ3bfPsaJ43rYq8BkCdkd+bIKPz/jUO37ijQy9qEAcEMDM01kcyIHWlFcDF UMTlwSGJGeuuYBZ/eUPPOybdNcFMZQ8AnNR6kErXat2ZPeYVRNMSjTlt8Kt+CBTOTqU+ jT2R/iUAnGSvAbYZxeVSbqND2Ldcavcbm/1ZTIUb7J2HkhjcfvL3t2C9zBRWfilXTSEm KYhQ== In-Reply-To: <20210412225503.15119-6-longman-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> List-ID: Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Waiman Long Cc: Johannes Weiner , Michal Hocko , Vladimir Davydov , Andrew Morton , Tejun Heo , Christoph Lameter , Pekka Enberg , David Rientjes , Joonsoo Kim , Vlastimil Babka , Roman Gushchin , LKML , Cgroups , Linux MM , Muchun Song , Alex Shi , Chris Down , Yafang Shao , Wei Yang , Masayoshi Mizuma , Xing Zhengjun On Mon, Apr 12, 2021 at 3:55 PM Waiman Long wrote: > > Most kmem_cache_alloc() calls are from user context. With instrumentation > enabled, the measured amount of kmem_cache_alloc() calls from non-task > context was about 0.01% of the total. > > The irq disable/enable sequence used in this case to access content > from object stock is slow. To optimize for user context access, there > are now two object stocks for task context and interrupt context access > respectively. > > The task context object stock can be accessed after disabling preemption > which is cheap in non-preempt kernel. The interrupt context object stock > can only be accessed after disabling interrupt. User context code can > access interrupt object stock, but not vice versa. > > The mod_objcg_state() function is also modified to make sure that memcg > and lruvec stat updates are done with interrupted disabled. > > The downside of this change is that there are more data stored in local > object stocks and not reflected in the charge counter and the vmstat > arrays. However, this is a small price to pay for better performance. > > Signed-off-by: Waiman Long > Acked-by: Roman Gushchin > --- > mm/memcontrol.c | 73 +++++++++++++++++++++++++++++++++++++++---------- > 1 file changed, 59 insertions(+), 14 deletions(-) > > diff --git a/mm/memcontrol.c b/mm/memcontrol.c > index 69f728383efe..29f2df76644a 100644 > --- a/mm/memcontrol.c > +++ b/mm/memcontrol.c > @@ -2229,7 +2229,8 @@ struct obj_stock { > struct memcg_stock_pcp { > struct mem_cgroup *cached; /* this never be root cgroup */ > unsigned int nr_pages; > - struct obj_stock obj; > + struct obj_stock task_obj; > + struct obj_stock irq_obj; > > struct work_struct work; > unsigned long flags; > @@ -2254,11 +2255,48 @@ static bool obj_stock_flush_required(struct memcg_stock_pcp *stock, > } > #endif > > +/* > + * Most kmem_cache_alloc() calls are from user context. The irq disable/enable > + * sequence used in this case to access content from object stock is slow. > + * To optimize for user context access, there are now two object stocks for > + * task context and interrupt context access respectively. > + * > + * The task context object stock can be accessed by disabling preemption only > + * which is cheap in non-preempt kernel. The interrupt context object stock > + * can only be accessed after disabling interrupt. User context code can > + * access interrupt object stock, but not vice versa. > + */ > static inline struct obj_stock *current_obj_stock(void) > { > struct memcg_stock_pcp *stock = this_cpu_ptr(&memcg_stock); > > - return &stock->obj; > + return in_task() ? &stock->task_obj : &stock->irq_obj; > +} > + > +#define get_obj_stock(flags) \ > +({ \ > + struct memcg_stock_pcp *stock; \ > + struct obj_stock *obj_stock; \ > + \ > + if (in_task()) { \ > + preempt_disable(); \ > + (flags) = -1L; \ > + stock = this_cpu_ptr(&memcg_stock); \ The above line was missing in the previous version. > + obj_stock = &stock->task_obj; \ > + } else { \ > + local_irq_save(flags); \ > + stock = this_cpu_ptr(&memcg_stock); \ > + obj_stock = &stock->irq_obj; \ > + } \ > + obj_stock; \ > +}) > + > +static inline void put_obj_stock(unsigned long flags) > +{ > + if (flags == -1L) > + preempt_enable(); > + else > + local_irq_restore(flags); > } > > /** > @@ -2327,7 +2365,9 @@ static void drain_local_stock(struct work_struct *dummy) > local_irq_save(flags); > > stock = this_cpu_ptr(&memcg_stock); > - drain_obj_stock(&stock->obj); > + drain_obj_stock(&stock->irq_obj); > + if (in_task()) > + drain_obj_stock(&stock->task_obj); > drain_stock(stock); > clear_bit(FLUSHING_CACHED_CHARGE, &stock->flags); > > @@ -3183,7 +3223,7 @@ static inline void mod_objcg_state(struct obj_cgroup *objcg, > memcg = obj_cgroup_memcg(objcg); > if (pgdat) > lruvec = mem_cgroup_lruvec(memcg, pgdat); > - __mod_memcg_lruvec_state(memcg, lruvec, idx, nr); > + mod_memcg_lruvec_state(memcg, lruvec, idx, nr); > rcu_read_unlock(); > } > > @@ -3193,7 +3233,7 @@ static bool consume_obj_stock(struct obj_cgroup *objcg, unsigned int nr_bytes) > unsigned long flags; > bool ret = false; > > - local_irq_save(flags); > + stock = get_obj_stock(flags); > > stock = current_obj_stock(); The above is redundant. > if (objcg == stock->cached_objcg && stock->nr_bytes >= nr_bytes) { > @@ -3201,7 +3241,7 @@ static bool consume_obj_stock(struct obj_cgroup *objcg, unsigned int nr_bytes) > ret = true; > } > > - local_irq_restore(flags); > + put_obj_stock(flags); > > return ret; > } > @@ -3254,8 +3294,13 @@ static bool obj_stock_flush_required(struct memcg_stock_pcp *stock, > { > struct mem_cgroup *memcg; > > - if (stock->obj.cached_objcg) { > - memcg = obj_cgroup_memcg(stock->obj.cached_objcg); > + if (in_task() && stock->task_obj.cached_objcg) { > + memcg = obj_cgroup_memcg(stock->task_obj.cached_objcg); > + if (memcg && mem_cgroup_is_descendant(memcg, root_memcg)) > + return true; > + } > + if (stock->irq_obj.cached_objcg) { > + memcg = obj_cgroup_memcg(stock->irq_obj.cached_objcg); > if (memcg && mem_cgroup_is_descendant(memcg, root_memcg)) > return true; > } > @@ -3283,9 +3328,9 @@ static void refill_obj_stock(struct obj_cgroup *objcg, unsigned int nr_bytes) > { > unsigned long flags; > > - local_irq_save(flags); > + get_obj_stock(flags); > __refill_obj_stock(objcg, nr_bytes); > - local_irq_restore(flags); > + put_obj_stock(flags); > } > > static void __mod_obj_stock_state(struct obj_cgroup *objcg, > @@ -3325,9 +3370,9 @@ void mod_obj_stock_state(struct obj_cgroup *objcg, struct pglist_data *pgdat, > { > unsigned long flags; > > - local_irq_save(flags); > + get_obj_stock(flags); > __mod_obj_stock_state(objcg, pgdat, idx, nr); > - local_irq_restore(flags); > + put_obj_stock(flags); > } > > int obj_cgroup_charge(struct obj_cgroup *objcg, gfp_t gfp, size_t size) > @@ -3380,10 +3425,10 @@ void obj_cgroup_uncharge_mod_state(struct obj_cgroup *objcg, size_t size, > { > unsigned long flags; > > - local_irq_save(flags); > + get_obj_stock(flags); > __refill_obj_stock(objcg, size); > __mod_obj_stock_state(objcg, pgdat, idx, -(int)size); > - local_irq_restore(flags); > + put_obj_stock(flags); > } > > #endif /* CONFIG_MEMCG_KMEM */ > -- > 2.18.1 >