From patchwork Mon Mar 16 10:09:32 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Yuri Benditovich X-Patchwork-Id: 11440093 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id C2F2D1668 for ; Mon, 16 Mar 2020 10:26:29 +0000 (UTC) Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) (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 86980205ED for ; Mon, 16 Mar 2020 10:26:29 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (2048-bit key) header.d=daynix-com.20150623.gappssmtp.com header.i=@daynix-com.20150623.gappssmtp.com header.b="MUUD3KIP" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 86980205ED Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=daynix.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org Received: from localhost ([::1]:36680 helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1jDmx2-0003zW-66 for patchwork-qemu-devel@patchwork.kernel.org; Mon, 16 Mar 2020 06:26:28 -0400 Received: from eggs.gnu.org ([2001:470:142:3::10]:38347) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1jDmgw-00063b-Gb for qemu-devel@nongnu.org; Mon, 16 Mar 2020 06:09:53 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1jDmgu-0004Nh-GN for qemu-devel@nongnu.org; Mon, 16 Mar 2020 06:09:50 -0400 Received: from mail-wm1-x341.google.com ([2a00:1450:4864:20::341]:39280) by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_128_CBC_SHA1:16) (Exim 4.71) (envelope-from ) id 1jDmgu-0004FG-6z for qemu-devel@nongnu.org; Mon, 16 Mar 2020 06:09:48 -0400 Received: by mail-wm1-x341.google.com with SMTP id f7so17288097wml.4 for ; Mon, 16 Mar 2020 03:09:48 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=daynix-com.20150623.gappssmtp.com; s=20150623; h=from:to:cc:subject:date:message-id:in-reply-to:references; bh=c1wTfXu/2hbExbmLg7C4XuLU62RQFo9tK0K81KxmXHM=; b=MUUD3KIPu5J28Ne7jIrSrZKZ6YDPuL1UyCJhzp/vPG3gJU0DVRkDlzIWP/RX3BpvpH g28cscDZ3VtyfB7hFfzmMTikoWKL5GxD1w9chtOS6BAPFtdA0JFNv8/dz+8e3W51mvI7 QUEMDNjrujbLU7l9glqsBvQSJVYyQd5qJB92ggVAm6NNPtk8btVXmw4jTWp5InzL5F1x 8ntjvmmwuz9ZnfSfeaPkGatLAeeCBmALC2C4C7yDoJTsGQe+/nOekUHRkPMbvp/zUxRc UeRTcq8cJjjO/w1mF6+nct8dHaZViqVUodZ+p55l8SxMz3h4dBnNAzOp4+LnF+aOaNLE RAcg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references; bh=c1wTfXu/2hbExbmLg7C4XuLU62RQFo9tK0K81KxmXHM=; b=iLm9fpcFHK+6Gx13QovVhKstOnGO6Ek4WB5s8khCCAGyOlgt8Bm4AGquBXpUTneOMK HFIjYyujTRKeJw6My+sLYx0ZpoqagVJ5FPBQ7IYJADUI6pEvmFDNxkMugqiyy2v+ex1H uTpu/Ehtvpbc1vT7xzZ+/RV0Mht58UCg0chnPO28jONKUm3HKBmFDlrZQ4Szi+Nd6KCq ZqgedookX1LJhQL/spFVFOxPr3Ph7GNfyCP7KfrRClw2yOesje6NYtHGpyYpHcuUh9JY qHUi3S1g//6wJRLY+IqfBM3fhq+7S49gk25+oYrYwQJLFqqOt1jC1NsLqbhoRNDNO7bs hV0g== X-Gm-Message-State: ANhLgQ2pynbXiBB7OYGNA0rY1PlrV0fqEg3WwMZF+sKgPMEG0X8hSwU5 rzSFkeWB1fo4y8hsBy6I16p5/rREq1qmzA== X-Google-Smtp-Source: ADFU+vs2dErK4xDpNUS1FuEgUI02mz0eThomvm/jO/PHRFcOH1LfezyUFdlWeQsxLoMPJNQN8IqUgw== X-Received: by 2002:a7b:ce94:: with SMTP id q20mr26050931wmj.157.1584353386666; Mon, 16 Mar 2020 03:09:46 -0700 (PDT) Received: from f2.redhat.com (bzq-79-183-43-120.red.bezeqint.net. [79.183.43.120]) by smtp.gmail.com with ESMTPSA id l7sm22083171wrw.33.2020.03.16.03.09.45 (version=TLS1_2 cipher=ECDHE-RSA-CHACHA20-POLY1305 bits=256/256); Mon, 16 Mar 2020 03:09:45 -0700 (PDT) From: Yuri Benditovich To: qemu-devel@nongnu.org, mst@redhat.com, jasowang@redhat.com Subject: [PATCH v4 5/6] virtio-net: reference implementation of hash report Date: Mon, 16 Mar 2020 12:09:32 +0200 Message-Id: <20200316100933.11499-6-yuri.benditovich@daynix.com> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20200316100933.11499-1-yuri.benditovich@daynix.com> References: <20200316100933.11499-1-yuri.benditovich@daynix.com> X-detected-operating-system: by eggs.gnu.org: Genre and OS details not recognized. X-Received-From: 2a00:1450:4864:20::341 X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: yan@daynix.com Errors-To: qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org Sender: "Qemu-devel" Suggest VIRTIO_NET_F_HASH_REPORT if specified in device parameters. If the VIRTIO_NET_F_HASH_REPORT is set, the device extends configuration space. If the feature is negotiated, the packet layout is extended to accomodate the hash information. In this case deliver packet's hash value and report type in virtio header extension. Use for configuration the same procedure as already used for RSS. We add two fields in rss_data that controls what the device does with the calculated hash if rss_data.enabled is set. If field 'populate' is set the hash is set in the packet, if field 'redirect' is set the hash is used to decide the queue to place the packet to. Signed-off-by: Yuri Benditovich --- hw/net/virtio-net.c | 99 +++++++++++++++++++++++++++------- include/hw/virtio/virtio-net.h | 2 + 2 files changed, 81 insertions(+), 20 deletions(-) diff --git a/hw/net/virtio-net.c b/hw/net/virtio-net.c index de2d68d4ca..a0614ad4e6 100644 --- a/hw/net/virtio-net.c +++ b/hw/net/virtio-net.c @@ -184,7 +184,7 @@ static VirtIOFeature feature_sizes[] = { .end = endof(struct virtio_net_config, mtu)}, {.flags = 1ULL << VIRTIO_NET_F_SPEED_DUPLEX, .end = endof(struct virtio_net_config, duplex)}, - {.flags = 1ULL << VIRTIO_NET_F_RSS, + {.flags = (1ULL << VIRTIO_NET_F_RSS) | (1ULL << VIRTIO_NET_F_HASH_REPORT), .end = endof(struct virtio_net_config_with_rss, supported_hash_types)}, {} }; @@ -218,7 +218,8 @@ static void virtio_net_get_config(VirtIODevice *vdev, uint8_t *config) netcfg.cfg.duplex = n->net_conf.duplex; netcfg.rss_max_key_size = VIRTIO_NET_RSS_MAX_KEY_SIZE; virtio_stw_p(vdev, &netcfg.rss_max_indirection_table_length, - VIRTIO_NET_RSS_MAX_TABLE_LEN); + virtio_host_has_feature(vdev, VIRTIO_NET_F_RSS) ? + VIRTIO_NET_RSS_MAX_TABLE_LEN : 1); virtio_stl_p(vdev, &netcfg.supported_hash_types, VIRTIO_NET_RSS_SUPPORTED_HASHES); memcpy(config, &netcfg, n->config_size); @@ -644,7 +645,7 @@ static int peer_has_ufo(VirtIONet *n) } static void virtio_net_set_mrg_rx_bufs(VirtIONet *n, int mergeable_rx_bufs, - int version_1) + int version_1, int hash_report) { int i; NetClientState *nc; @@ -652,7 +653,10 @@ static void virtio_net_set_mrg_rx_bufs(VirtIONet *n, int mergeable_rx_bufs, n->mergeable_rx_bufs = mergeable_rx_bufs; if (version_1) { - n->guest_hdr_len = sizeof(struct virtio_net_hdr_mrg_rxbuf); + n->guest_hdr_len = hash_report ? + sizeof(struct virtio_net_hdr_v1_hash) : + sizeof(struct virtio_net_hdr_mrg_rxbuf); + n->rss_data.populate_hash = true; } else { n->guest_hdr_len = n->mergeable_rx_bufs ? sizeof(struct virtio_net_hdr_mrg_rxbuf) : @@ -773,6 +777,8 @@ static uint64_t virtio_net_get_features(VirtIODevice *vdev, uint64_t features, virtio_clear_feature(&features, VIRTIO_NET_F_GUEST_TSO4); virtio_clear_feature(&features, VIRTIO_NET_F_GUEST_TSO6); virtio_clear_feature(&features, VIRTIO_NET_F_GUEST_ECN); + + virtio_clear_feature(&features, VIRTIO_NET_F_HASH_REPORT); } if (!peer_has_vnet_hdr(n) || !peer_has_ufo(n)) { @@ -785,6 +791,7 @@ static uint64_t virtio_net_get_features(VirtIODevice *vdev, uint64_t features, } virtio_clear_feature(&features, VIRTIO_NET_F_RSS); + virtio_clear_feature(&features, VIRTIO_NET_F_HASH_REPORT); features = vhost_net_get_features(get_vhost_net(nc->peer), features); vdev->backend_features = features; @@ -951,12 +958,15 @@ static void virtio_net_set_features(VirtIODevice *vdev, uint64_t features) virtio_has_feature(features, VIRTIO_NET_F_MRG_RXBUF), virtio_has_feature(features, - VIRTIO_F_VERSION_1)); + VIRTIO_F_VERSION_1), + virtio_has_feature(features, + VIRTIO_NET_F_HASH_REPORT)); n->rsc4_enabled = virtio_has_feature(features, VIRTIO_NET_F_RSC_EXT) && virtio_has_feature(features, VIRTIO_NET_F_GUEST_TSO4); n->rsc6_enabled = virtio_has_feature(features, VIRTIO_NET_F_RSC_EXT) && virtio_has_feature(features, VIRTIO_NET_F_GUEST_TSO6); + n->rss_data.redirect = virtio_has_feature(features, VIRTIO_NET_F_RSS); if (n->has_vnet_hdr) { n->curr_guest_offloads = @@ -1230,7 +1240,9 @@ static void virtio_net_disable_rss(VirtIONet *n) } static uint16_t virtio_net_handle_rss(VirtIONet *n, - struct iovec *iov, unsigned int iov_cnt) + struct iovec *iov, + unsigned int iov_cnt, + bool do_rss) { VirtIODevice *vdev = VIRTIO_DEVICE(n); struct virtio_net_rss_config cfg; @@ -1243,10 +1255,14 @@ static uint16_t virtio_net_handle_rss(VirtIONet *n, const char *err_msg = ""; uint32_t err_value = 0; - if (!virtio_vdev_has_feature(vdev, VIRTIO_NET_F_RSS)) { + if (do_rss && !virtio_vdev_has_feature(vdev, VIRTIO_NET_F_RSS)) { err_msg = "RSS is not negotiated"; goto error; } + if (!do_rss && !virtio_vdev_has_feature(vdev, VIRTIO_NET_F_HASH_REPORT)) { + err_msg = "Hash report is not negotiated"; + goto error; + } size_get = offsetof(struct virtio_net_rss_config, indirection_table); s = iov_to_buf(iov, iov_cnt, offset, &cfg, size_get); if (s != size_get) { @@ -1258,6 +1274,9 @@ static uint16_t virtio_net_handle_rss(VirtIONet *n, n->rss_data.indirections_len = virtio_lduw_p(vdev, &cfg.indirection_table_mask); n->rss_data.indirections_len++; + if (!do_rss) { + n->rss_data.indirections_len = 1; + } if (!is_power_of_2(n->rss_data.indirections_len)) { err_msg = "Invalid size of indirection table"; err_value = n->rss_data.indirections_len; @@ -1268,8 +1287,8 @@ static uint16_t virtio_net_handle_rss(VirtIONet *n, err_value = n->rss_data.indirections_len; goto error; } - n->rss_data.default_queue = - virtio_lduw_p(vdev, &cfg.unclassified_queue); + n->rss_data.default_queue = do_rss ? + virtio_lduw_p(vdev, &cfg.unclassified_queue) : 0; if (n->rss_data.default_queue >= n->max_queues) { err_msg = "Invalid default queue"; err_value = n->rss_data.default_queue; @@ -1303,7 +1322,7 @@ static uint16_t virtio_net_handle_rss(VirtIONet *n, err_value = (uint32_t)s; goto error; } - queues = virtio_lduw_p(vdev, &temp.us); + queues = do_rss ? virtio_lduw_p(vdev, &temp.us) : n->curr_queues; if (queues == 0 || queues > n->max_queues) { err_msg = "Invalid number of queues"; err_value = queues; @@ -1349,8 +1368,12 @@ static int virtio_net_handle_mq(VirtIONet *n, uint8_t cmd, uint16_t queues; virtio_net_disable_rss(n); + if (cmd == VIRTIO_NET_CTRL_MQ_HASH_CONFIG) { + queues = virtio_net_handle_rss(n, iov, iov_cnt, false); + return queues ? VIRTIO_NET_OK : VIRTIO_NET_ERR; + } if (cmd == VIRTIO_NET_CTRL_MQ_RSS_CONFIG) { - queues = virtio_net_handle_rss(n, iov, iov_cnt); + queues = virtio_net_handle_rss(n, iov, iov_cnt, true); } else if (cmd == VIRTIO_NET_CTRL_MQ_VQ_PAIRS_SET) { struct virtio_net_ctrl_mq mq; size_t s; @@ -1637,15 +1660,34 @@ static uint8_t virtio_net_get_hash_type(bool isip4, return 0xff; } +static void virtio_set_packet_hash(const uint8_t *buf, uint8_t report, + uint32_t hash) +{ + struct virtio_net_hdr_v1_hash *hdr = (void *)buf; + hdr->hash_value = hash; + hdr->hash_report = report; +} + static int virtio_net_process_rss(NetClientState *nc, const uint8_t *buf, size_t size) { VirtIONet *n = qemu_get_nic_opaque(nc); - unsigned int index = nc->queue_index, new_index; + unsigned int index = nc->queue_index, new_index = index; struct NetRxPkt *pkt = n->rx_pkt; uint8_t net_hash_type; uint32_t hash; bool isip4, isip6, isudp, istcp; + static const uint8_t reports[NetPktRssIpV6UdpEx + 1] = { + VIRTIO_NET_HASH_REPORT_IPv4, + VIRTIO_NET_HASH_REPORT_TCPv4, + VIRTIO_NET_HASH_REPORT_TCPv6, + VIRTIO_NET_HASH_REPORT_IPv6, + VIRTIO_NET_HASH_REPORT_IPv6_EX, + VIRTIO_NET_HASH_REPORT_TCPv6_EX, + VIRTIO_NET_HASH_REPORT_UDPv4, + VIRTIO_NET_HASH_REPORT_UDPv6, + VIRTIO_NET_HASH_REPORT_UDPv6_EX + }; net_rx_pkt_set_protocols(pkt, buf + n->host_hdr_len, size - n->host_hdr_len); @@ -1659,16 +1701,24 @@ static int virtio_net_process_rss(NetClientState *nc, const uint8_t *buf, net_hash_type = virtio_net_get_hash_type(isip4, isip6, isudp, istcp, n->rss_data.hash_types); if (net_hash_type > NetPktRssIpV6UdpEx) { - return n->rss_data.default_queue; + if (n->rss_data.populate_hash) { + virtio_set_packet_hash(buf, VIRTIO_NET_HASH_REPORT_NONE, 0); + } + return n->rss_data.redirect ? n->rss_data.default_queue : -1; } hash = net_rx_pkt_calc_rss_hash(pkt, net_hash_type, n->rss_data.key); - new_index = hash & (n->rss_data.indirections_len - 1); - new_index = n->rss_data.indirections_table[new_index]; - if (index == new_index) { - return -1; + + if (n->rss_data.populate_hash) { + virtio_set_packet_hash(buf, reports[net_hash_type], hash); } - return new_index; + + if (n->rss_data.redirect) { + new_index = hash & (n->rss_data.indirections_len - 1); + new_index = n->rss_data.indirections_table[new_index]; + } + + return (index == new_index) ? -1 : new_index; } static ssize_t virtio_net_receive_rcu(NetClientState *nc, const uint8_t *buf, @@ -1744,6 +1794,11 @@ static ssize_t virtio_net_receive_rcu(NetClientState *nc, const uint8_t *buf, } receive_header(n, sg, elem->in_num, buf, size); + if (n->rss_data.populate_hash) { + offset = sizeof(mhdr); + iov_from_buf(sg, elem->in_num, offset, + buf + offset, n->host_hdr_len - sizeof(mhdr)); + } offset = n->host_hdr_len; total += n->guest_hdr_len; guest_offset = n->guest_hdr_len; @@ -2736,7 +2791,9 @@ static int virtio_net_post_load_device(void *opaque, int version_id) trace_virtio_net_post_load_device(); virtio_net_set_mrg_rx_bufs(n, n->mergeable_rx_bufs, virtio_vdev_has_feature(vdev, - VIRTIO_F_VERSION_1)); + VIRTIO_F_VERSION_1), + virtio_vdev_has_feature(vdev, + VIRTIO_NET_F_HASH_REPORT)); /* MAC_TABLE_ENTRIES may be different from the saved image */ if (n->mac_table.in_use > MAC_TABLE_ENTRIES) { @@ -3352,7 +3409,7 @@ static void virtio_net_device_realize(DeviceState *dev, Error **errp) n->vqs[0].tx_waiting = 0; n->tx_burst = n->net_conf.txburst; - virtio_net_set_mrg_rx_bufs(n, 0, 0); + virtio_net_set_mrg_rx_bufs(n, 0, 0, 0); n->promisc = 1; /* for compatibility */ n->mac_table.macs = g_malloc0(MAC_TABLE_ENTRIES * ETH_ALEN); @@ -3507,6 +3564,8 @@ static Property virtio_net_properties[] = { DEFINE_PROP_BIT64("mq", VirtIONet, host_features, VIRTIO_NET_F_MQ, false), DEFINE_PROP_BIT64("rss", VirtIONet, host_features, VIRTIO_NET_F_RSS, false), + DEFINE_PROP_BIT64("hash", VirtIONet, host_features, + VIRTIO_NET_F_HASH_REPORT, false), DEFINE_PROP_BIT64("guest_rsc_ext", VirtIONet, host_features, VIRTIO_NET_F_RSC_EXT, false), DEFINE_PROP_UINT32("rsc_interval", VirtIONet, rsc_timeout, diff --git a/include/hw/virtio/virtio-net.h b/include/hw/virtio/virtio-net.h index c92697bfd5..a5b99167b1 100644 --- a/include/hw/virtio/virtio-net.h +++ b/include/hw/virtio/virtio-net.h @@ -131,6 +131,8 @@ typedef struct VirtioNetRscChain { typedef struct VirtioNetRssData { bool enabled; + bool redirect; + bool populate_hash; uint32_t hash_types; uint8_t key[VIRTIO_NET_RSS_MAX_KEY_SIZE]; uint32_t indirections_len;