Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755032AbdLOKJQ (ORCPT ); Fri, 15 Dec 2017 05:09:16 -0500 Received: from mail.linuxfoundation.org ([140.211.169.12]:36286 "EHLO mail.linuxfoundation.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932531AbdLOJyL (ORCPT ); Fri, 15 Dec 2017 04:54:11 -0500 From: Greg Kroah-Hartman To: linux-kernel@vger.kernel.org Cc: Greg Kroah-Hartman , stable@vger.kernel.org, Nils Hoppmann , Julian Wiedmann , "David S. Miller" Subject: [PATCH 4.14 20/52] s390/qeth: fix GSO throughput regression Date: Fri, 15 Dec 2017 10:51:57 +0100 Message-Id: <20171215092309.584072912@linuxfoundation.org> X-Mailer: git-send-email 2.15.1 In-Reply-To: <20171215092308.500651185@linuxfoundation.org> References: <20171215092308.500651185@linuxfoundation.org> User-Agent: quilt/0.65 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 5343 Lines: 140 4.14-stable review patch. If anyone has any objections, please let me know. ------------------ From: Julian Wiedmann [ Upstream commit 6d69b1f1eb7a2edf8a3547f361c61f2538e054bb ] Using GSO with small MTUs currently results in a substantial throughput regression - which is caused by how qeth needs to map non-linear skbs into its IO buffer elements: compared to a linear skb, each GSO-segmented skb effectively consumes twice as many buffer elements (ie two instead of one) due to the additional header-only part. This causes the Output Queue to be congested with low-utilized IO buffers. Fix this as follows: If the MSS is low enough so that a non-SG GSO segmentation produces order-0 skbs (currently ~3500 byte), opt out from NETIF_F_SG. This is where we anticipate the biggest savings, since an SG-enabled GSO segmentation produces skbs that always consume at least two buffer elements. Larger MSS values continue to get a SG-enabled GSO segmentation, since 1) the relative overhead of the additional header-only buffer element becomes less noticeable, and 2) the linearization overhead increases. With the throughput regression fixed, re-enable NETIF_F_SG by default to reap the significant CPU savings of GSO. Fixes: 5722963a8e83 ("qeth: do not turn on SG per default") Reported-by: Nils Hoppmann Signed-off-by: Julian Wiedmann Signed-off-by: David S. Miller Signed-off-by: Greg Kroah-Hartman --- drivers/s390/net/qeth_core.h | 3 +++ drivers/s390/net/qeth_core_main.c | 31 +++++++++++++++++++++++++++++++ drivers/s390/net/qeth_l2_main.c | 2 ++ drivers/s390/net/qeth_l3_main.c | 2 ++ 4 files changed, 38 insertions(+) --- a/drivers/s390/net/qeth_core.h +++ b/drivers/s390/net/qeth_core.h @@ -985,6 +985,9 @@ struct qeth_cmd_buffer *qeth_get_setassp int qeth_set_features(struct net_device *, netdev_features_t); int qeth_recover_features(struct net_device *); netdev_features_t qeth_fix_features(struct net_device *, netdev_features_t); +netdev_features_t qeth_features_check(struct sk_buff *skb, + struct net_device *dev, + netdev_features_t features); int qeth_vm_request_mac(struct qeth_card *card); int qeth_push_hdr(struct sk_buff *skb, struct qeth_hdr **hdr, unsigned int len); --- a/drivers/s390/net/qeth_core_main.c +++ b/drivers/s390/net/qeth_core_main.c @@ -19,6 +19,11 @@ #include #include #include +#include +#include +#include +#include + #include #include @@ -6505,6 +6510,32 @@ netdev_features_t qeth_fix_features(stru } EXPORT_SYMBOL_GPL(qeth_fix_features); +netdev_features_t qeth_features_check(struct sk_buff *skb, + struct net_device *dev, + netdev_features_t features) +{ + /* GSO segmentation builds skbs with + * a (small) linear part for the headers, and + * page frags for the data. + * Compared to a linear skb, the header-only part consumes an + * additional buffer element. This reduces buffer utilization, and + * hurts throughput. So compress small segments into one element. + */ + if (netif_needs_gso(skb, features)) { + /* match skb_segment(): */ + unsigned int doffset = skb->data - skb_mac_header(skb); + unsigned int hsize = skb_shinfo(skb)->gso_size; + unsigned int hroom = skb_headroom(skb); + + /* linearize only if resulting skb allocations are order-0: */ + if (SKB_DATA_ALIGN(hroom + doffset + hsize) <= SKB_MAX_HEAD(0)) + features &= ~NETIF_F_SG; + } + + return vlan_features_check(skb, features); +} +EXPORT_SYMBOL_GPL(qeth_features_check); + static int __init qeth_core_init(void) { int rc; --- a/drivers/s390/net/qeth_l2_main.c +++ b/drivers/s390/net/qeth_l2_main.c @@ -963,6 +963,7 @@ static const struct net_device_ops qeth_ .ndo_stop = qeth_l2_stop, .ndo_get_stats = qeth_get_stats, .ndo_start_xmit = qeth_l2_hard_start_xmit, + .ndo_features_check = qeth_features_check, .ndo_validate_addr = eth_validate_addr, .ndo_set_rx_mode = qeth_l2_set_rx_mode, .ndo_do_ioctl = qeth_do_ioctl, @@ -1009,6 +1010,7 @@ static int qeth_l2_setup_netdev(struct q if (card->info.type == QETH_CARD_TYPE_OSD && !card->info.guestlan) { card->dev->hw_features = NETIF_F_SG; card->dev->vlan_features = NETIF_F_SG; + card->dev->features |= NETIF_F_SG; /* OSA 3S and earlier has no RX/TX support */ if (qeth_is_supported(card, IPA_OUTBOUND_CHECKSUM)) { card->dev->hw_features |= NETIF_F_IP_CSUM; --- a/drivers/s390/net/qeth_l3_main.c +++ b/drivers/s390/net/qeth_l3_main.c @@ -2923,6 +2923,7 @@ static const struct net_device_ops qeth_ .ndo_stop = qeth_l3_stop, .ndo_get_stats = qeth_get_stats, .ndo_start_xmit = qeth_l3_hard_start_xmit, + .ndo_features_check = qeth_features_check, .ndo_validate_addr = eth_validate_addr, .ndo_set_rx_mode = qeth_l3_set_multicast_list, .ndo_do_ioctl = qeth_do_ioctl, @@ -2963,6 +2964,7 @@ static int qeth_l3_setup_netdev(struct q card->dev->vlan_features = NETIF_F_SG | NETIF_F_RXCSUM | NETIF_F_IP_CSUM | NETIF_F_TSO; + card->dev->features |= NETIF_F_SG; } } } else if (card->info.type == QETH_CARD_TYPE_IQD) {