From patchwork Thu Dec 1 12:54:28 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Sunil Kovvuri X-Patchwork-Id: 9456123 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id A85636074E for ; Thu, 1 Dec 2016 12:57:32 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 97D7528249 for ; Thu, 1 Dec 2016 12:57:32 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 8A235284CB; Thu, 1 Dec 2016 12:57:32 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-4.1 required=2.0 tests=BAYES_00, DKIM_ADSP_CUSTOM_MED, DKIM_SIGNED,FREEMAIL_FROM,RCVD_IN_DNSWL_MED,T_DKIM_INVALID autolearn=unavailable version=3.3.1 Received: from bombadil.infradead.org (bombadil.infradead.org [198.137.202.9]) (using TLSv1.2 with cipher AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by mail.wl.linuxfoundation.org (Postfix) with ESMTPS id E35B928249 for ; Thu, 1 Dec 2016 12:57:31 +0000 (UTC) Received: from localhost ([127.0.0.1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.85_2 #1 (Red Hat Linux)) id 1cCQtC-00021Z-Fc; Thu, 01 Dec 2016 12:55:02 +0000 Received: from mail-pg0-x242.google.com ([2607:f8b0:400e:c05::242]) by bombadil.infradead.org with esmtps (Exim 4.85_2 #1 (Red Hat Linux)) id 1cCQt8-0001vy-Dx for linux-arm-kernel@lists.infradead.org; Thu, 01 Dec 2016 12:54:59 +0000 Received: by mail-pg0-x242.google.com with SMTP id p66so4597910pga.2 for ; Thu, 01 Dec 2016 04:54:37 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=from:to:cc:subject:date:message-id; bh=haR5/gizY20WhTsUHx+iRTmeVLHhaOrn0rRoO60tF7M=; b=WcmEHszm0PpGC+lj5ByvESMzCIouA0DzJD//AYzaozGCXhCD+TMuZVtO6grkS2fl91 b7OKYa02fJeF4zpmi5rmdSgxUcqJDYwyn3M1o71Udgfvie4GgJm+i63odtHHpK8xvc3c gEPcTnVJw/OtOz1ej67mEOUR6Jkz2/9Ks8mVE2FMj7K3eiNpr7W1xIpx7ZwXJEuQrMbr lWzS7nlbNZ2c3tAGQ4lmBdCalHtC6uc5FDEj38YtMBnaoaKyhCjN2H6BmK2oKWn4vK95 PZrzc+GGCRzz2RE0oF+fTpdYg/CD2zWP9bXAurA9mxzQo1ys1eeCWgMHmu74xmiQ2RNh q1Fw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20130820; h=x-gm-message-state:from:to:cc:subject:date:message-id; bh=haR5/gizY20WhTsUHx+iRTmeVLHhaOrn0rRoO60tF7M=; b=H6xEaFf3H+AyWjEYiRteKz5e6px3U4Y0D4e4FI4Wlh51UIGcGt24q6kjiAG0xnEWwb 2KffwGHAwspk7+Jt6Dl0mrvFVwRumfWm6qr29baMEEBXMKhrUugw3DN3hvPebVposF5l EwwuHSTWDQIAP9RrL/r5db1v0OjcAU9DeRHnWy9x/saoQ6HOVkv79X5/4YFFs2+pbX4n L8/83b6jfzuaPzP0f97LKMb3YB2XbWmpFFtNSTxEH5Gjr67EiAY/XKnERSBP7H5O0qRL 6CbJ9hd8vvwnfxZCYjIUmQbFmt8qwu1MpqFAwkE1e4Kyn7T3jso/S0PAz3Wq1xbUVat+ 2npA== X-Gm-Message-State: AKaTC02IAjQGxORTkeHA67v1q1GmzzbnX41oPWFMoxYUYWNIdUSJ5xHfgf6qPeKtUkuqrQ== X-Received: by 10.99.160.1 with SMTP id r1mr68304754pge.107.1480596876517; Thu, 01 Dec 2016 04:54:36 -0800 (PST) Received: from machine421.in.caveonetworks.com ([14.140.2.178]) by smtp.googlemail.com with ESMTPSA id y20sm248725pfj.26.2016.12.01.04.54.33 (version=TLS1_2 cipher=ECDHE-RSA-AES128-SHA bits=128/128); Thu, 01 Dec 2016 04:54:35 -0800 (PST) From: sunil.kovvuri@gmail.com To: netdev@vger.kernel.org Subject: [PATCH net-next v2] net: thunderx: Fix transmit queue timeout issue Date: Thu, 1 Dec 2016 18:24:28 +0530 Message-Id: <1480596868-17693-1-git-send-email-sunil.kovvuri@gmail.com> X-Mailer: git-send-email 2.7.4 X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20161201_045458_525113_7CE9F939 X-CRM114-Status: GOOD ( 16.33 ) X-BeenThere: linux-arm-kernel@lists.infradead.org X-Mailman-Version: 2.1.20 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Sunil Goutham , linux-kernel@vger.kernel.org, linux-arm-kernel@lists.infradead.org MIME-Version: 1.0 Sender: "linux-arm-kernel" Errors-To: linux-arm-kernel-bounces+patchwork-linux-arm=patchwork.kernel.org@lists.infradead.org X-Virus-Scanned: ClamAV using ClamSMTP From: Sunil Goutham Transmit queue timeout issue is seen in two cases - Due to a race condition btw setting stop_queue at xmit() and checking for stopped_queue in NAPI poll routine, at times transmission from a SQ comes to a halt. This is fixed by using barriers and also added a check for SQ free descriptors, incase SQ is stopped and there are only CQE_RX i.e no CQE_TX. - Contrary to an assumption, a HW errata where HW doesn't stop transmission even though there are not enough CQEs available for a CQE_TX is not fixed in T88 pass 2.x. This results in a Qset error with 'CQ_WR_FULL' stalling transmission. This is fixed by adjusting RXQ's RED levels for CQ level such that there is always enough space left for CQE_TXs. Signed-off-by: Sunil Goutham --- v2: As suggested by David, replaced netif_tx_start_queue with netif_tx_wake_queue. drivers/net/ethernet/cavium/thunder/nicvf_main.c | 52 ++++++++++++++++++---- drivers/net/ethernet/cavium/thunder/nicvf_queues.c | 24 ++-------- drivers/net/ethernet/cavium/thunder/nicvf_queues.h | 15 ++++--- 3 files changed, 54 insertions(+), 37 deletions(-) diff --git a/drivers/net/ethernet/cavium/thunder/nicvf_main.c b/drivers/net/ethernet/cavium/thunder/nicvf_main.c index 1eacec8..2006f58 100644 --- a/drivers/net/ethernet/cavium/thunder/nicvf_main.c +++ b/drivers/net/ethernet/cavium/thunder/nicvf_main.c @@ -644,6 +644,7 @@ static int nicvf_cq_intr_handler(struct net_device *netdev, u8 cq_idx, struct cmp_queue *cq = &qs->cq[cq_idx]; struct cqe_rx_t *cq_desc; struct netdev_queue *txq; + struct snd_queue *sq; unsigned int tx_pkts = 0, tx_bytes = 0; spin_lock_bh(&cq->lock); @@ -709,16 +710,20 @@ static int nicvf_cq_intr_handler(struct net_device *netdev, u8 cq_idx, done: /* Wakeup TXQ if its stopped earlier due to SQ full */ - if (tx_done) { + sq = &nic->qs->sq[cq_idx]; + if (tx_done || + (atomic_read(&sq->free_cnt) >= MIN_SQ_DESC_PER_PKT_XMIT)) { netdev = nic->pnicvf->netdev; txq = netdev_get_tx_queue(netdev, nicvf_netdev_qidx(nic, cq_idx)); if (tx_pkts) netdev_tx_completed_queue(txq, tx_pkts, tx_bytes); - nic = nic->pnicvf; + /* To read updated queue and carrier status */ + smp_mb(); if (netif_tx_queue_stopped(txq) && netif_carrier_ok(netdev)) { - netif_tx_start_queue(txq); + netif_tx_wake_queue(txq); + nic = nic->pnicvf; this_cpu_inc(nic->drv_stats->txq_wake); if (netif_msg_tx_err(nic)) netdev_warn(netdev, @@ -1054,6 +1059,9 @@ static netdev_tx_t nicvf_xmit(struct sk_buff *skb, struct net_device *netdev) struct nicvf *nic = netdev_priv(netdev); int qid = skb_get_queue_mapping(skb); struct netdev_queue *txq = netdev_get_tx_queue(netdev, qid); + struct nicvf *snic; + struct snd_queue *sq; + int tmp; /* Check for minimum packet length */ if (skb->len <= ETH_HLEN) { @@ -1061,13 +1069,39 @@ static netdev_tx_t nicvf_xmit(struct sk_buff *skb, struct net_device *netdev) return NETDEV_TX_OK; } - if (!netif_tx_queue_stopped(txq) && !nicvf_sq_append_skb(nic, skb)) { + snic = nic; + /* Get secondary Qset's SQ structure */ + if (qid >= MAX_SND_QUEUES_PER_QS) { + tmp = qid / MAX_SND_QUEUES_PER_QS; + snic = (struct nicvf *)nic->snicvf[tmp - 1]; + if (!snic) { + netdev_warn(nic->netdev, + "Secondary Qset#%d's ptr not initialized\n", + tmp - 1); + dev_kfree_skb(skb); + return NETDEV_TX_OK; + } + qid = qid % MAX_SND_QUEUES_PER_QS; + } + + sq = &snic->qs->sq[qid]; + if (!netif_tx_queue_stopped(txq) && + !nicvf_sq_append_skb(snic, sq, skb, qid)) { netif_tx_stop_queue(txq); - this_cpu_inc(nic->drv_stats->txq_stop); - if (netif_msg_tx_err(nic)) - netdev_warn(netdev, - "%s: Transmit ring full, stopping SQ%d\n", - netdev->name, qid); + + /* Barrier, so that stop_queue visible to other cpus */ + smp_mb(); + + /* Check again, incase another cpu freed descriptors */ + if (atomic_read(&sq->free_cnt) > MIN_SQ_DESC_PER_PKT_XMIT) { + netif_tx_wake_queue(txq); + } else { + this_cpu_inc(nic->drv_stats->txq_stop); + if (netif_msg_tx_err(nic)) + netdev_warn(netdev, + "%s: Transmit ring full, stopping SQ%d\n", + netdev->name, qid); + } return NETDEV_TX_BUSY; } diff --git a/drivers/net/ethernet/cavium/thunder/nicvf_queues.c b/drivers/net/ethernet/cavium/thunder/nicvf_queues.c index 7b336cd..d2ac133 100644 --- a/drivers/net/ethernet/cavium/thunder/nicvf_queues.c +++ b/drivers/net/ethernet/cavium/thunder/nicvf_queues.c @@ -1190,30 +1190,12 @@ static int nicvf_sq_append_tso(struct nicvf *nic, struct snd_queue *sq, } /* Append an skb to a SQ for packet transfer. */ -int nicvf_sq_append_skb(struct nicvf *nic, struct sk_buff *skb) +int nicvf_sq_append_skb(struct nicvf *nic, struct snd_queue *sq, + struct sk_buff *skb, u8 sq_num) { int i, size; int subdesc_cnt, tso_sqe = 0; - int sq_num, qentry; - struct queue_set *qs; - struct snd_queue *sq; - - sq_num = skb_get_queue_mapping(skb); - if (sq_num >= MAX_SND_QUEUES_PER_QS) { - /* Get secondary Qset's SQ structure */ - i = sq_num / MAX_SND_QUEUES_PER_QS; - if (!nic->snicvf[i - 1]) { - netdev_warn(nic->netdev, - "Secondary Qset#%d's ptr not initialized\n", - i - 1); - return 1; - } - nic = (struct nicvf *)nic->snicvf[i - 1]; - sq_num = sq_num % MAX_SND_QUEUES_PER_QS; - } - - qs = nic->qs; - sq = &qs->sq[sq_num]; + int qentry; subdesc_cnt = nicvf_sq_subdesc_required(nic, skb); if (subdesc_cnt > atomic_read(&sq->free_cnt)) diff --git a/drivers/net/ethernet/cavium/thunder/nicvf_queues.h b/drivers/net/ethernet/cavium/thunder/nicvf_queues.h index 20511f2..9e21046 100644 --- a/drivers/net/ethernet/cavium/thunder/nicvf_queues.h +++ b/drivers/net/ethernet/cavium/thunder/nicvf_queues.h @@ -88,13 +88,13 @@ /* RED and Backpressure levels of CQ for pkt reception * For CQ, level is a measure of emptiness i.e 0x0 means full - * eg: For CQ of size 4K, and for pass/drop levels of 128/96 - * HW accepts pkt if unused CQE >= 2048 - * RED accepts pkt if unused CQE < 2048 & >= 1536 - * DROPs pkts if unused CQE < 1536 + * eg: For CQ of size 4K, and for pass/drop levels of 160/144 + * HW accepts pkt if unused CQE >= 2560 + * RED accepts pkt if unused CQE < 2304 & >= 2560 + * DROPs pkts if unused CQE < 2304 */ -#define RQ_PASS_CQ_LVL 128ULL -#define RQ_DROP_CQ_LVL 96ULL +#define RQ_PASS_CQ_LVL 160ULL +#define RQ_DROP_CQ_LVL 144ULL /* RED and Backpressure levels of RBDR for pkt reception * For RBDR, level is a measure of fullness i.e 0x0 means empty @@ -306,7 +306,8 @@ void nicvf_sq_disable(struct nicvf *nic, int qidx); void nicvf_put_sq_desc(struct snd_queue *sq, int desc_cnt); void nicvf_sq_free_used_descs(struct net_device *netdev, struct snd_queue *sq, int qidx); -int nicvf_sq_append_skb(struct nicvf *nic, struct sk_buff *skb); +int nicvf_sq_append_skb(struct nicvf *nic, struct snd_queue *sq, + struct sk_buff *skb, u8 sq_num); struct sk_buff *nicvf_get_rcv_skb(struct nicvf *nic, struct cqe_rx_t *cqe_rx); void nicvf_rbdr_task(unsigned long data);