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=-2.2 required=3.0 tests=HEADER_FROM_DIFFERENT_DOMAINS, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,USER_AGENT_SANE_1 autolearn=no 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 3D9A8C43331 for ; Thu, 26 Mar 2020 06:59:39 +0000 (UTC) Received: from lists.xenproject.org (lists.xenproject.org [192.237.175.120]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 124ED20719 for ; Thu, 26 Mar 2020 06:59:38 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 124ED20719 Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=suse.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=xen-devel-bounces@lists.xenproject.org Received: from localhost ([127.0.0.1] helo=lists.xenproject.org) by lists.xenproject.org with esmtp (Exim 4.89) (envelope-from ) id 1jHMTY-00019o-RU; Thu, 26 Mar 2020 06:58:48 +0000 Received: from us1-rack-iad1.inumbo.com ([172.99.69.81]) by lists.xenproject.org with esmtp (Exim 4.89) (envelope-from ) id 1jHMTY-00019j-2s for xen-devel@lists.xenproject.org; Thu, 26 Mar 2020 06:58:48 +0000 X-Inumbo-ID: 3c413964-6f2f-11ea-b34e-bc764e2007e4 Received: from mx2.suse.de (unknown [195.135.220.15]) by us1-rack-iad1.inumbo.com (Halon) with ESMTPS id 3c413964-6f2f-11ea-b34e-bc764e2007e4; Thu, 26 Mar 2020 06:58:46 +0000 (UTC) X-Virus-Scanned: by amavisd-new at test-mx.suse.de Received: from relay2.suse.de (unknown [195.135.220.254]) by mx2.suse.de (Postfix) with ESMTP id D01EDAD82; Thu, 26 Mar 2020 06:58:44 +0000 (UTC) To: Julien Grall References: <20200325105511.20882-1-jgross@suse.com> <20200325105511.20882-3-jgross@suse.com> <1c688233-2f64-9dd0-7d98-4a0840489293@xen.org> From: Jan Beulich Message-ID: <5950cabe-0065-904c-6c61-0e6eab8f8060@suse.com> Date: Thu, 26 Mar 2020 07:58:39 +0100 User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:68.0) Gecko/20100101 Thunderbird/68.6.0 MIME-Version: 1.0 In-Reply-To: <1c688233-2f64-9dd0-7d98-4a0840489293@xen.org> Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 8bit Subject: Re: [Xen-devel] [PATCH v7 2/5] xen/rcu: don't use stop_machine_run() for rcu_barrier() X-BeenThere: xen-devel@lists.xenproject.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: Xen developer discussion List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Cc: Juergen Gross , Stefano Stabellini , Wei Liu , Andrew Cooper , Ian Jackson , George Dunlap , xen-devel@lists.xenproject.org Errors-To: xen-devel-bounces@lists.xenproject.org Sender: "Xen-devel" On 25.03.2020 17:13, Julien Grall wrote: > On 25/03/2020 10:55, Juergen Gross wrote: >> @@ -143,51 +143,90 @@ static int qhimark = 10000; >>   static int qlowmark = 100; >>   static int rsinterval = 1000; >>   -struct rcu_barrier_data { >> -    struct rcu_head head; >> -    atomic_t *cpu_count; >> -}; >> +/* >> + * rcu_barrier() handling: >> + * Two counters are used to synchronize rcu_barrier() work: >> + * - cpu_count holds the number of cpus required to finish barrier handling. >> + *   It is decremented by each cpu when it has performed all pending rcu calls. >> + * - pending_count shows whether any rcu_barrier() activity is running and >> + *   it is used to synchronize leaving rcu_barrier() only after all cpus >> + *   have finished their processing. pending_count is initialized to nr_cpus + 1 >> + *   and it is decremented by each cpu when it has seen that cpu_count has >> + *   reached 0. The cpu where rcu_barrier() has been called will wait until >> + *   pending_count has been decremented to 1 (so all cpus have seen cpu_count >> + *   reaching 0) and will then set pending_count to 0 indicating there is no >> + *   rcu_barrier() running. >> + * Cpus are synchronized via softirq mechanism. rcu_barrier() is regarded to >> + * be active if pending_count is not zero. In case rcu_barrier() is called on >> + * multiple cpus it is enough to check for pending_count being not zero on entry >> + * and to call process_pending_softirqs() in a loop until pending_count drops to >> + * zero, before starting the new rcu_barrier() processing. >> + */ >> +static atomic_t cpu_count = ATOMIC_INIT(0); >> +static atomic_t pending_count = ATOMIC_INIT(0); >>     static void rcu_barrier_callback(struct rcu_head *head) >>   { >> -    struct rcu_barrier_data *data = container_of( >> -        head, struct rcu_barrier_data, head); >> -    atomic_inc(data->cpu_count); >> +    smp_mb__before_atomic();     /* Make all writes visible to other cpus. */ > > smp_mb__before_atomic() will order both read and write. However, the > comment suggest only the write are required to be ordered. > > So either the barrier is too strong or the comment is incorrect. Can > you clarify it? Neither is the case, I guess: There simply is no smp_wmb__before_atomic() in Linux, and if we want to follow their model we shouldn't have one either. I'd rather take the comment to indicate that if one appeared, it could be used here. >> +    atomic_dec(&cpu_count); >>   } >>   -static int rcu_barrier_action(void *_cpu_count) >> +static void rcu_barrier_action(void) >>   { >> -    struct rcu_barrier_data data = { .cpu_count = _cpu_count }; >> - >> -    ASSERT(!local_irq_is_enabled()); >> -    local_irq_enable(); >> +    struct rcu_head head; >>         /* >>        * When callback is executed, all previously-queued RCU work on this CPU >> -     * is completed. When all CPUs have executed their callback, data.cpu_count >> -     * will have been incremented to include every online CPU. >> +     * is completed. When all CPUs have executed their callback, cpu_count >> +     * will have been decremented to 0. >>        */ >> -    call_rcu(&data.head, rcu_barrier_callback); >> +    call_rcu(&head, rcu_barrier_callback); >>   -    while ( atomic_read(data.cpu_count) != num_online_cpus() ) >> +    while ( atomic_read(&cpu_count) ) >>       { >>           process_pending_softirqs(); >>           cpu_relax(); >>       } >>   -    local_irq_disable(); >> - >> -    return 0; >> +    smp_mb__before_atomic(); >> +    atomic_dec(&pending_count); >>   } >>   -/* >> - * As rcu_barrier() is using stop_machine_run() it is allowed to be used in >> - * idle context only (see comment for stop_machine_run()). >> - */ >> -int rcu_barrier(void) >> +void rcu_barrier(void) >>   { >> -    atomic_t cpu_count = ATOMIC_INIT(0); >> -    return stop_machine_run(rcu_barrier_action, &cpu_count, NR_CPUS); >> +    unsigned int n_cpus; >> + >> +    ASSERT(!in_irq() && local_irq_is_enabled()); >> + >> +    for ( ; ; ) >> +    { >> +        if ( !atomic_read(&pending_count) && get_cpu_maps() ) >> +        { >> +            n_cpus = num_online_cpus(); >> + >> +            if ( atomic_cmpxchg(&pending_count, 0, n_cpus + 1) == 0 ) >> +                break; >> + >> +            put_cpu_maps(); >> +        } >> + >> +        process_pending_softirqs(); >> +        cpu_relax(); >> +    } >> + >> +    smp_mb__before_atomic(); > > Our semantic of atomic_cmpxchg() is exactly the same as Linux. I.e > it will contain a full barrier when the cmpxchg succeed. So why do you need this barrier? I was me I think to have (wrongly) suggested a barrier was missing here, sorry. Jan