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=-17.1 required=3.0 tests=BAYES_00,DKIMWL_WL_HIGH, DKIM_SIGNED,DKIM_VALID,HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER, INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED, USER_AGENT_GIT 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 F148EC49EA5 for ; Sat, 26 Jun 2021 11:05:44 +0000 (UTC) Received: from bombadil.infradead.org (bombadil.infradead.org [198.137.202.133]) (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 B89E361474 for ; Sat, 26 Jun 2021 11:05:44 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org B89E361474 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=huawei.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-arm-kernel-bounces+linux-arm-kernel=archiver.kernel.org@lists.infradead.org DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=bombadil.20210309; h=Sender: Content-Transfer-Encoding:Content-Type:List-Subscribe:List-Help:List-Post: List-Archive:List-Unsubscribe:List-Id:MIME-Version:References:In-Reply-To: Message-ID:Date:Subject:CC:To:From:Reply-To:Content-ID:Content-Description: Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID: List-Owner; bh=ZfjsJhopyn/C30QOi/R9rs7keNZ5OAyDQl6D5JVfNZU=; b=SBYWAfbrNfcc86 q8RcGfn4q9JGl5O+ELst6MxDZx5dJwm3lSEH+HBVU1Ch4ifIR4oOYRv0/1hPE1hlmZMg8oMtq0lMf gv0wvnvTekGn+/tQmIiYcKd7MTaQFPV56yGik2sHK5HFxDHttQ1pTxQqC6HvePFivXsjFvS+lzg70 tnoBez2JXmnonbxAi1KmU1GeIwrjejXnfjDtzNy3/5qICztV7BESJ/AGU+knebeZolHnGP8fP+7x9 ftu9TkwEXp3rzf5XuppvFlmL+oRDMWtNiCfl7D6TR+WYu/KBJqju0cXlWZWkpHadRDdWMV33uKT7M rzlldP/ksja08OdQE6AA==; Received: from localhost ([::1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.94.2 #2 (Red Hat Linux)) id 1lx66h-003zhW-1h; Sat, 26 Jun 2021 11:04:15 +0000 Received: from szxga01-in.huawei.com ([45.249.212.187]) by bombadil.infradead.org with esmtps (Exim 4.94.2 #2 (Red Hat Linux)) id 1lx64q-003yiz-V0 for linux-arm-kernel@lists.infradead.org; Sat, 26 Jun 2021 11:02:26 +0000 Received: from dggemv711-chm.china.huawei.com (unknown [172.30.72.54]) by szxga01-in.huawei.com (SkyGuard) with ESMTP id 4GBrRG2D4hzXlHg; Sat, 26 Jun 2021 18:56:58 +0800 (CST) Received: from dggpemm500006.china.huawei.com (7.185.36.236) by dggemv711-chm.china.huawei.com (10.1.198.66) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.1.2176.2; Sat, 26 Jun 2021 19:02:12 +0800 Received: from thunder-town.china.huawei.com (10.174.179.0) by dggpemm500006.china.huawei.com (7.185.36.236) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.1.2176.2; Sat, 26 Jun 2021 19:02:12 +0800 From: Zhen Lei To: Will Deacon , Robin Murphy , "Joerg Roedel" , linux-arm-kernel , iommu , linux-kernel CC: Zhen Lei Subject: [PATCH RFC 8/8] iommu/arm-smmu-v3: Add support for less than one ECMDQ per core Date: Sat, 26 Jun 2021 19:01:30 +0800 Message-ID: <20210626110130.2416-9-thunder.leizhen@huawei.com> X-Mailer: git-send-email 2.26.0.windows.1 In-Reply-To: <20210626110130.2416-1-thunder.leizhen@huawei.com> References: <20210626110130.2416-1-thunder.leizhen@huawei.com> MIME-Version: 1.0 X-Originating-IP: [10.174.179.0] X-ClientProxiedBy: dggems701-chm.china.huawei.com (10.3.19.178) To dggpemm500006.china.huawei.com (7.185.36.236) X-CFilter-Loop: Reflected X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20210626_040221_495830_C6934DE0 X-CRM114-Status: GOOD ( 17.85 ) X-BeenThere: linux-arm-kernel@lists.infradead.org X-Mailman-Version: 2.1.34 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Sender: "linux-arm-kernel" Errors-To: linux-arm-kernel-bounces+linux-arm-kernel=archiver.kernel.org@lists.infradead.org Due to limited hardware resources, the number of ECMDQs may be less than the number of cores. If the number of ECMDQs is greater than the number of numa nodes, ensure that each node has at least one ECMDQ. This is because ECMDQ queue memory is requested from the NUMA node where it resides, which may result in better command filling and insertion performance. The current ECMDQ implementation reuses the command insertion function arm_smmu_cmdq_issue_cmdlist() of the normal CMDQ. This function already supports multiple cores concurrent insertion commands. Signed-off-by: Zhen Lei --- drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c | 101 ++++++++++++++++++-- 1 file changed, 92 insertions(+), 9 deletions(-) diff --git a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c index a088f2479fc6223..55f651ce42e7a51 100644 --- a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c +++ b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c @@ -3636,14 +3636,15 @@ static int arm_smmu_device_reset(struct arm_smmu_device *smmu, bool bypass) static int arm_smmu_ecmdq_layout(struct arm_smmu_device *smmu) { - int cpu; - struct arm_smmu_ecmdq *ecmdq; + int cpu, node, nr_remain, nr_nodes = 0; + int *nr_ecmdqs; + struct arm_smmu_ecmdq *ecmdq, **ecmdqs; - if (num_possible_cpus() <= smmu->nr_ecmdq) { - ecmdq = devm_alloc_percpu(smmu->dev, *ecmdq); - if (!ecmdq) - return -ENOMEM; + ecmdq = devm_alloc_percpu(smmu->dev, *ecmdq); + if (!ecmdq) + return -ENOMEM; + if (num_possible_cpus() <= smmu->nr_ecmdq) { for_each_possible_cpu(cpu) *per_cpu_ptr(smmu->ecmdq, cpu) = per_cpu_ptr(ecmdq, cpu); @@ -3653,7 +3654,79 @@ static int arm_smmu_ecmdq_layout(struct arm_smmu_device *smmu) return 0; } - return -ENOSPC; + for_each_node(node) + if (nr_cpus_node(node)) + nr_nodes++; + + if (nr_nodes >= smmu->nr_ecmdq) { + dev_err(smmu->dev, "%d ECMDQs is less than %d nodes\n", smmu->nr_ecmdq, nr_nodes); + return -ENOSPC; + } + + nr_ecmdqs = kcalloc(MAX_NUMNODES, sizeof(int), GFP_KERNEL); + if (!nr_ecmdqs) + return -ENOMEM; + + ecmdqs = kcalloc(smmu->nr_ecmdq, sizeof(*ecmdqs), GFP_KERNEL); + if (!ecmdqs) { + kfree(nr_ecmdqs); + return -ENOMEM; + } + + /* [1] Ensure that each node has at least one ECMDQ */ + nr_remain = smmu->nr_ecmdq - nr_nodes; + for_each_node(node) { + /* + * Calculate the number of ECMDQs to be allocated to this node. + * NR_ECMDQS_PER_CPU = nr_remain / num_possible_cpus(); + * When nr_cpus_node(node) is not zero, less than one ECMDQ + * may be left due to truncation rounding. + */ + nr_ecmdqs[node] = nr_cpus_node(node) * nr_remain / num_possible_cpus(); + nr_remain -= nr_ecmdqs[node]; + } + + /* Divide the remaining ECMDQs */ + while (nr_remain) { + for_each_node(node) { + if (!nr_remain) + break; + + if (nr_ecmdqs[node] >= nr_cpus_node(node)) + continue; + + nr_ecmdqs[node]++; + nr_remain--; + } + } + + for_each_node(node) { + int i, round, shared = 0; + + if (!nr_cpus_node(node)) + continue; + + /* An ECMDQ has been reserved for each node at above [1] */ + nr_ecmdqs[node]++; + + if (nr_ecmdqs[node] < nr_cpus_node(node)) + shared = 1; + + i = 0; + for_each_cpu(cpu, cpumask_of_node(node)) { + round = i % nr_ecmdqs[node]; + if (i++ < nr_ecmdqs[node]) { + ecmdqs[round] = per_cpu_ptr(ecmdq, cpu); + ecmdqs[round]->cmdq.shared = shared; + } + *per_cpu_ptr(smmu->ecmdq, cpu) = ecmdqs[round]; + } + } + + kfree(nr_ecmdqs); + kfree(ecmdqs); + + return 0; } static int arm_smmu_ecmdq_probe(struct arm_smmu_device *smmu) @@ -3718,10 +3791,20 @@ static int arm_smmu_ecmdq_probe(struct arm_smmu_device *smmu) struct arm_smmu_queue *q; ecmdq = *per_cpu_ptr(smmu->ecmdq, cpu); - ecmdq->base = cp_base + addr; - q = &ecmdq->cmdq.q; + /* + * The boot option "maxcpus=" can limit the number of online + * CPUs. The CPUs that are not selected are not showed in + * cpumask_of_node(node), their 'ecmdq' may be NULL. + * + * (q->ecmdq_prod & ECMDQ_PROD_EN) indicates that the ECMDQ is + * shared by multiple cores and has been initialized. + */ + if (!ecmdq || (q->ecmdq_prod & ECMDQ_PROD_EN)) + continue; + ecmdq->base = cp_base + addr; + q->llq.max_n_shift = ECMDQ_MAX_SZ_SHIFT + shift_increment; ret = arm_smmu_init_one_queue(smmu, q, ecmdq->base, ARM_SMMU_ECMDQ_PROD, ARM_SMMU_ECMDQ_CONS, CMDQ_ENT_DWORDS, "ecmdq"); -- 2.26.0.106.g9fadedd _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel