From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from dpdk.org (dpdk.org [92.243.14.124]) by inbox.dpdk.org (Postfix) with ESMTP id 4A058A00BE; Tue, 7 Jul 2020 10:22:28 +0200 (CEST) Received: from [92.243.14.124] (localhost [127.0.0.1]) by dpdk.org (Postfix) with ESMTP id C0FEC1DAC4; Tue, 7 Jul 2020 10:22:26 +0200 (CEST) Received: from mga11.intel.com (mga11.intel.com [192.55.52.93]) by dpdk.org (Postfix) with ESMTP id B50991DAC4 for ; Tue, 7 Jul 2020 10:22:24 +0200 (CEST) IronPort-SDR: KTTGAbKDUIwkOKXDMa//aRDUlBcoTm6TszqeGiz59qzzLkRryyAzIGvTV2G3fDQIZdILOAUXU+ WgjGJl13V7AQ== X-IronPort-AV: E=McAfee;i="6000,8403,9674"; a="145656633" X-IronPort-AV: E=Sophos;i="5.75,323,1589266800"; d="scan'208";a="145656633" X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from orsmga007.jf.intel.com ([10.7.209.58]) by fmsmga102.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 07 Jul 2020 01:22:23 -0700 IronPort-SDR: dFuvU/EYTXYFOZ49KzNLPkPDBFTtIRGfaDruB3YZZEuwu/eQWjpr3z34CGiifIQq/FXxPcHuic 94tyJKokqiWA== X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.75,323,1589266800"; d="scan'208";a="323468696" Received: from fmsmsx103.amr.corp.intel.com ([10.18.124.201]) by orsmga007.jf.intel.com with ESMTP; 07 Jul 2020 01:22:23 -0700 Received: from fmsmsx607.amr.corp.intel.com (10.18.126.87) by FMSMSX103.amr.corp.intel.com (10.18.124.201) with Microsoft SMTP Server (TLS) id 14.3.439.0; Tue, 7 Jul 2020 01:22:22 -0700 Received: from fmsmsx611.amr.corp.intel.com (10.18.126.91) by fmsmsx607.amr.corp.intel.com (10.18.126.87) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.1.1713.5; Tue, 7 Jul 2020 01:22:22 -0700 Received: from FMSEDG001.ED.cps.intel.com (10.1.192.133) by fmsmsx611.amr.corp.intel.com (10.18.126.91) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256) id 15.1.1713.5 via Frontend Transport; Tue, 7 Jul 2020 01:22:22 -0700 Received: from NAM10-DM6-obe.outbound.protection.outlook.com (104.47.58.105) by edgegateway.intel.com (192.55.55.68) with Microsoft SMTP Server (TLS) id 14.3.439.0; Tue, 7 Jul 2020 01:22:22 -0700 ARC-Seal: i=1; a=rsa-sha256; s=arcselector9901; d=microsoft.com; cv=none; b=fClPnyFpYwlrFN06eR2/QjYPWnlR/C3Bmr4q69TowcvaDnXqmzJhjh76biRMCVtAA6ykZm+mNmIAi6BCxdqipPhfAwNXAiGPlT/J0DI3+zgu89WLWKnjlTyLzZE3Q1WDyJLomtl7gJ3hmR+gOsEZyvQWtwcTU5B8TVaZwXS1A+sx+WqgYIqBScXswM1Upa0j+W9SC54OczzuzfTlAEB3KNB0YJArGCQOOGNPLE3Gw+CAR0BtyvjGLVryM3vqHW7knsPem5zh6Smnql0AQx31LpTm2oOKR3i+8qw2rOSH8k04SvAD8NcqcxoANX8cM2k8IamZIkLmfHoZ4Z7dZBwwkw== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=microsoft.com; s=arcselector9901; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-SenderADCheck; bh=j1RvUohZBFpQatVMSYMgmNwvFcJ+3rNBQMqVe8w8KxY=; b=Nwn93bhznwdy8gENXDK5Kxm7WKHP4jV6QNgzfSfdNhBYYwJjwMcuzxkZRqHDXYU0hEb0GdANYvB6uHrnHBlhuAatTppCZhcGRL1D+UMXDXTQBF+mM+QkcAlGHhIchDsNBWdPdbGZ4/P3rOYxGc/DIIkUUobyxrl5d3LMZO9CMr9W3mJUkj9xb9v0+agLhqjIW4VoaPv6+X9paWKfcSBXgdJ44ANbUYHQ3UFvJdxhxQ8LE1yeo6S2YrDi0ktycGbfNRhpi1IAdtB3MbSUWCxc45moMyWjGuh/LGBMTh2jX/VDyUue+95y/XNnQw2++itsrcFuAna14Sc02CL2ZuBqJg== ARC-Authentication-Results: i=1; mx.microsoft.com 1; spf=pass smtp.mailfrom=intel.com; dmarc=pass action=none header.from=intel.com; dkim=pass header.d=intel.com; arc=none DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=intel.onmicrosoft.com; s=selector2-intel-onmicrosoft-com; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-SenderADCheck; bh=j1RvUohZBFpQatVMSYMgmNwvFcJ+3rNBQMqVe8w8KxY=; b=CKxrHSAQAlSJsyzFb1aXyEtlPf+7sk9JQQs9cegapzgKrElhutkwL9OycrI21x+wC9F1ZaB620y95tgsu3EU8W/MAtH1+e/sA2aVn4NHn78YxxbXS31zcEn1Pv494Oyfv3/qDXWU4EMUJ0UNv2eaZFuzdkBaE3jrKV1ch3IblMI= Received: from MN2PR11MB4063.namprd11.prod.outlook.com (2603:10b6:208:13f::22) by MN2PR11MB3712.namprd11.prod.outlook.com (2603:10b6:208:f6::29) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.3153.20; Tue, 7 Jul 2020 08:22:20 +0000 Received: from MN2PR11MB4063.namprd11.prod.outlook.com ([fe80::7cde:8326:5010:c47e]) by MN2PR11MB4063.namprd11.prod.outlook.com ([fe80::7cde:8326:5010:c47e%7]) with mapi id 15.20.3153.029; Tue, 7 Jul 2020 08:22:20 +0000 From: "Xia, Chenbo" To: "Fu, Patrick" , "dev@dpdk.org" , "maxime.coquelin@redhat.com" , "Wang, Zhihong" CC: "Wang, Yinan" , "Jiang, Cheng1" , "Liang, Cunming" Thread-Topic: [PATCH v6 2/2] vhost: introduce async enqueue for split ring Thread-Index: AQHWVByCXb5WPe7jLE603+/BKReqyKj7xzCw Date: Tue, 7 Jul 2020 08:22:19 +0000 Message-ID: References: <1591869725-13331-1-git-send-email-patrick.fu@intel.com> <20200707050709.205480-1-patrick.fu@intel.com> <20200707050709.205480-3-patrick.fu@intel.com> In-Reply-To: <20200707050709.205480-3-patrick.fu@intel.com> Accept-Language: zh-CN, en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: authentication-results: intel.com; dkim=none (message not signed) header.d=none;intel.com; dmarc=none action=none header.from=intel.com; x-originating-ip: [192.198.147.218] x-ms-publictraffictype: Email x-ms-office365-filtering-correlation-id: e64fef8c-19d9-42c1-8158-08d8224ede00 x-ms-traffictypediagnostic: MN2PR11MB3712: x-ms-exchange-transport-forked: True x-microsoft-antispam-prvs: x-ms-oob-tlc-oobclassifiers: OLM:4303; x-forefront-prvs: 0457F11EAF x-ms-exchange-senderadcheck: 1 x-microsoft-antispam: BCL:0; x-microsoft-antispam-message-info: SIv/hZT3RuikK9D6ji4bERHdzLyFVQkw/7pkn1VMd8//sH3sxgg/AJfr5RkmXWywVmyCO9fXSbosDSbXPaEnmBgqZUvXmVjhFOBG8TCluZJ+xqyL+PAR3TiIAn3Fy9cU4IbndrK++NgX2UcEYfcjEREkdh1JWSWuogBiGFl+/1amu+4x+s2n7zvnZE2u0Hc1wMzsFw0uE85nXgh0mRpb/R5YrGq/YrBMp/7RPbVaf7VUZFBYKcSLNGNjzgKp0kn3jl3vB9TA08uwszpTIzqXGvWN7MlywFuo0Uy8tm5UFM0AHCmASK1iy/vgpbdjn1K6 x-forefront-antispam-report: CIP:255.255.255.255; CTRY:; LANG:en; SCL:1; SRV:; IPV:NLI; SFV:NSPM; H:MN2PR11MB4063.namprd11.prod.outlook.com; PTR:; CAT:NONE; SFTY:; SFS:(4636009)(39860400002)(136003)(366004)(346002)(396003)(376002)(26005)(2906002)(110136005)(316002)(6636002)(478600001)(66946007)(64756008)(66446008)(66556008)(66476007)(7696005)(4326008)(6506007)(54906003)(76116006)(107886003)(30864003)(53546011)(33656002)(52536014)(8936002)(186003)(8676002)(86362001)(55016002)(5660300002)(9686003)(83380400001)(71200400001); DIR:OUT; SFP:1102; x-ms-exchange-antispam-messagedata: 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 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 X-MS-Exchange-CrossTenant-AuthAs: Internal X-MS-Exchange-CrossTenant-AuthSource: MN2PR11MB4063.namprd11.prod.outlook.com X-MS-Exchange-CrossTenant-Network-Message-Id: e64fef8c-19d9-42c1-8158-08d8224ede00 X-MS-Exchange-CrossTenant-originalarrivaltime: 07 Jul 2020 08:22:19.9244 (UTC) X-MS-Exchange-CrossTenant-fromentityheader: Hosted X-MS-Exchange-CrossTenant-id: 46c98d88-e344-4ed4-8496-4ed7712e255d X-MS-Exchange-CrossTenant-mailboxtype: HOSTED X-MS-Exchange-CrossTenant-userprincipalname: qkN/EFBH3IEAwE/iKlnxcLZfuNtGo33xf6GHsHKxte6IjxBj9XNyY/GP7BGdV2oP7p/vdoz+NMPMKhrqEj4PLA== X-MS-Exchange-Transport-CrossTenantHeadersStamped: MN2PR11MB3712 X-OriginatorOrg: intel.com Subject: Re: [dpdk-dev] [PATCH v6 2/2] vhost: introduce async enqueue for split ring X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" > -----Original Message----- > From: Fu, Patrick > Sent: Tuesday, July 7, 2020 1:07 PM > To: dev@dpdk.org; maxime.coquelin@redhat.com; Xia, Chenbo > ; Wang, Zhihong > Cc: Fu, Patrick ; Wang, Yinan ; > Jiang, Cheng1 ; Liang, Cunming > > Subject: [PATCH v6 2/2] vhost: introduce async enqueue for split ring >=20 > From: Patrick Fu >=20 > This patch implements async enqueue data path for split ring. 2 new async= data > path APIs are defined, by which applications can submit and poll packets = to/from > async engines. The async engine is either a physical DMA device or it cou= ld also > be a software emulated backend. > The async enqueue data path leverages callback functions registered by > applications to work with the async engine. >=20 > Signed-off-by: Patrick Fu > --- > lib/librte_vhost/rte_vhost_async.h | 40 +++ > lib/librte_vhost/virtio_net.c | 551 ++++++++++++++++++++++++++++- > 2 files changed, 589 insertions(+), 2 deletions(-) >=20 > diff --git a/lib/librte_vhost/rte_vhost_async.h > b/lib/librte_vhost/rte_vhost_async.h > index d5a59279a..c8ad8dbc7 100644 > --- a/lib/librte_vhost/rte_vhost_async.h > +++ b/lib/librte_vhost/rte_vhost_async.h > @@ -133,4 +133,44 @@ int rte_vhost_async_channel_register(int vid, uint16= _t > queue_id, __rte_experimental int rte_vhost_async_channel_unregister(int= vid, > uint16_t queue_id); >=20 > +/** > + * This function submit enqueue data to async engine. This function has > + * no guranttee to the transfer completion upon return. Applications > + * should poll transfer status by rte_vhost_poll_enqueue_completed() > + * > + * @param vid > + * id of vhost device to enqueue data > + * @param queue_id > + * queue id to enqueue data > + * @param pkts > + * array of packets to be enqueued > + * @param count > + * packets num to be enqueued > + * @return > + * num of packets enqueued > + */ > +__rte_experimental > +uint16_t rte_vhost_submit_enqueue_burst(int vid, uint16_t queue_id, > + struct rte_mbuf **pkts, uint16_t count); > + > +/** > + * This function check async completion status for a specific vhost > + * device queue. Packets which finish copying (enqueue) operation > + * will be returned in an array. > + * > + * @param vid > + * id of vhost device to enqueue data > + * @param queue_id > + * queue id to enqueue data > + * @param pkts > + * blank array to get return packet pointer > + * @param count > + * size of the packet array > + * @return > + * num of packets returned > + */ > +__rte_experimental > +uint16_t rte_vhost_poll_enqueue_completed(int vid, uint16_t queue_id, > + struct rte_mbuf **pkts, uint16_t count); > + > #endif /* _RTE_VHOST_ASYNC_H_ */ > diff --git a/lib/librte_vhost/virtio_net.c b/lib/librte_vhost/virtio_net.= c index > 751c1f373..236498f71 100644 > --- a/lib/librte_vhost/virtio_net.c > +++ b/lib/librte_vhost/virtio_net.c > @@ -17,14 +17,15 @@ > #include > #include > #include > +#include >=20 > #include "iotlb.h" > #include "vhost.h" >=20 > -#define MAX_PKT_BURST 32 > - > #define MAX_BATCH_LEN 256 >=20 > +#define VHOST_ASYNC_BATCH_THRESHOLD 32 > + > static __rte_always_inline bool > rxvq_is_mergeable(struct virtio_net *dev) { @@ -116,6 +117,31 @@ > flush_shadow_used_ring_split(struct virtio_net *dev, struct vhost_virtque= ue *vq) > sizeof(vq->used->idx)); > } >=20 > +static __rte_always_inline void > +async_flush_shadow_used_ring_split(struct virtio_net *dev, > + struct vhost_virtqueue *vq) > +{ > + uint16_t used_idx =3D vq->last_used_idx & (vq->size - 1); > + > + if (used_idx + vq->shadow_used_idx <=3D vq->size) { > + do_flush_shadow_used_ring_split(dev, vq, used_idx, 0, > + vq->shadow_used_idx); > + } else { > + uint16_t size; > + > + /* update used ring interval [used_idx, vq->size] */ > + size =3D vq->size - used_idx; > + do_flush_shadow_used_ring_split(dev, vq, used_idx, 0, size); > + > + /* update the left half used ring interval [0, left_size] */ > + do_flush_shadow_used_ring_split(dev, vq, 0, size, > + vq->shadow_used_idx - size); > + } > + > + vq->last_used_idx +=3D vq->shadow_used_idx; > + vq->shadow_used_idx =3D 0; > +} > + > static __rte_always_inline void > update_shadow_used_ring_split(struct vhost_virtqueue *vq, > uint16_t desc_idx, uint32_t len) > @@ -905,6 +931,200 @@ copy_mbuf_to_desc(struct virtio_net *dev, struct > vhost_virtqueue *vq, > return error; > } >=20 > +static __rte_always_inline void > +async_fill_vec(struct iovec *v, void *base, size_t len) { > + v->iov_base =3D base; > + v->iov_len =3D len; > +} > + > +static __rte_always_inline void > +async_fill_iter(struct rte_vhost_iov_iter *it, size_t count, > + struct iovec *vec, unsigned long nr_seg) { > + it->offset =3D 0; > + it->count =3D count; > + > + if (count) { > + it->iov =3D vec; > + it->nr_segs =3D nr_seg; > + } else { > + it->iov =3D 0; > + it->nr_segs =3D 0; > + } > +} > + > +static __rte_always_inline void > +async_fill_desc(struct rte_vhost_async_desc *desc, > + struct rte_vhost_iov_iter *src, struct rte_vhost_iov_iter *dst) { > + desc->src =3D src; > + desc->dst =3D dst; > +} > + > +static __rte_always_inline int > +async_mbuf_to_desc(struct virtio_net *dev, struct vhost_virtqueue *vq, > + struct rte_mbuf *m, struct buf_vector *buf_vec, > + uint16_t nr_vec, uint16_t num_buffers, > + struct iovec *src_iovec, struct iovec *dst_iovec, > + struct rte_vhost_iov_iter *src_it, > + struct rte_vhost_iov_iter *dst_it) > +{ > + uint32_t vec_idx =3D 0; > + uint32_t mbuf_offset, mbuf_avail; > + uint32_t buf_offset, buf_avail; > + uint64_t buf_addr, buf_iova, buf_len; > + uint32_t cpy_len, cpy_threshold; > + uint64_t hdr_addr; > + struct rte_mbuf *hdr_mbuf; > + struct batch_copy_elem *batch_copy =3D vq->batch_copy_elems; > + struct virtio_net_hdr_mrg_rxbuf tmp_hdr, *hdr =3D NULL; > + int error =3D 0; > + > + uint32_t tlen =3D 0; > + int tvec_idx =3D 0; > + void *hpa; > + > + if (unlikely(m =3D=3D NULL)) { > + error =3D -1; > + goto out; > + } > + > + cpy_threshold =3D vq->async_threshold; > + > + buf_addr =3D buf_vec[vec_idx].buf_addr; > + buf_iova =3D buf_vec[vec_idx].buf_iova; > + buf_len =3D buf_vec[vec_idx].buf_len; > + > + if (unlikely(buf_len < dev->vhost_hlen && nr_vec <=3D 1)) { > + error =3D -1; > + goto out; > + } > + > + hdr_mbuf =3D m; > + hdr_addr =3D buf_addr; > + if (unlikely(buf_len < dev->vhost_hlen)) > + hdr =3D &tmp_hdr; > + else > + hdr =3D (struct virtio_net_hdr_mrg_rxbuf *)(uintptr_t)hdr_addr; > + > + VHOST_LOG_DATA(DEBUG, "(%d) RX: num merge buffers %d\n", > + dev->vid, num_buffers); > + > + if (unlikely(buf_len < dev->vhost_hlen)) { > + buf_offset =3D dev->vhost_hlen - buf_len; > + vec_idx++; > + buf_addr =3D buf_vec[vec_idx].buf_addr; > + buf_iova =3D buf_vec[vec_idx].buf_iova; > + buf_len =3D buf_vec[vec_idx].buf_len; > + buf_avail =3D buf_len - buf_offset; > + } else { > + buf_offset =3D dev->vhost_hlen; > + buf_avail =3D buf_len - dev->vhost_hlen; > + } > + > + mbuf_avail =3D rte_pktmbuf_data_len(m); > + mbuf_offset =3D 0; > + > + while (mbuf_avail !=3D 0 || m->next !=3D NULL) { > + /* done with current buf, get the next one */ > + if (buf_avail =3D=3D 0) { > + vec_idx++; > + if (unlikely(vec_idx >=3D nr_vec)) { > + error =3D -1; > + goto out; > + } > + > + buf_addr =3D buf_vec[vec_idx].buf_addr; > + buf_iova =3D buf_vec[vec_idx].buf_iova; > + buf_len =3D buf_vec[vec_idx].buf_len; > + > + buf_offset =3D 0; > + buf_avail =3D buf_len; > + } > + > + /* done with current mbuf, get the next one */ > + if (mbuf_avail =3D=3D 0) { > + m =3D m->next; > + > + mbuf_offset =3D 0; > + mbuf_avail =3D rte_pktmbuf_data_len(m); > + } > + > + if (hdr_addr) { > + virtio_enqueue_offload(hdr_mbuf, &hdr->hdr); > + if (rxvq_is_mergeable(dev)) > + ASSIGN_UNLESS_EQUAL(hdr->num_buffers, > + num_buffers); > + > + if (unlikely(hdr =3D=3D &tmp_hdr)) { > + copy_vnet_hdr_to_desc(dev, vq, buf_vec, hdr); > + } else { > + PRINT_PACKET(dev, (uintptr_t)hdr_addr, > + dev->vhost_hlen, 0); > + vhost_log_cache_write_iova(dev, vq, > + buf_vec[0].buf_iova, > + dev->vhost_hlen); > + } > + > + hdr_addr =3D 0; > + } > + > + cpy_len =3D RTE_MIN(buf_avail, mbuf_avail); > + > + if (unlikely(cpy_len >=3D cpy_threshold)) { > + hpa =3D (void *)(uintptr_t)gpa_to_hpa(dev, > + buf_iova + buf_offset, cpy_len); > + > + if (unlikely(!hpa)) { > + error =3D -1; > + goto out; > + } > + > + async_fill_vec(src_iovec + tvec_idx, > + (void *)(uintptr_t)rte_pktmbuf_iova_offset(m, > + mbuf_offset), cpy_len); > + > + async_fill_vec(dst_iovec + tvec_idx, hpa, cpy_len); > + > + tlen +=3D cpy_len; > + tvec_idx++; > + } else { > + if (unlikely(vq->batch_copy_nb_elems >=3D vq->size)) { > + rte_memcpy( > + (void *)((uintptr_t)(buf_addr + buf_offset)), > + rte_pktmbuf_mtod_offset(m, void *, > mbuf_offset), > + cpy_len); > + > + PRINT_PACKET(dev, > + (uintptr_t)(buf_addr + buf_offset), > + cpy_len, 0); > + } else { > + batch_copy[vq->batch_copy_nb_elems].dst =3D > + (void *)((uintptr_t)(buf_addr + buf_offset)); > + batch_copy[vq->batch_copy_nb_elems].src =3D > + rte_pktmbuf_mtod_offset(m, void *, > mbuf_offset); > + batch_copy[vq- > >batch_copy_nb_elems].log_addr =3D > + buf_iova + buf_offset; > + batch_copy[vq->batch_copy_nb_elems].len =3D > + cpy_len; > + vq->batch_copy_nb_elems++; > + } > + } > + > + mbuf_avail -=3D cpy_len; > + mbuf_offset +=3D cpy_len; > + buf_avail -=3D cpy_len; > + buf_offset +=3D cpy_len; > + } > + > +out: > + async_fill_iter(src_it, tlen, src_iovec, tvec_idx); > + async_fill_iter(dst_it, tlen, dst_iovec, tvec_idx); > + > + return error; > +} > + > static __rte_always_inline int > vhost_enqueue_single_packed(struct virtio_net *dev, > struct vhost_virtqueue *vq, > @@ -1236,6 +1456,333 @@ rte_vhost_enqueue_burst(int vid, uint16_t > queue_id, > return virtio_dev_rx(dev, queue_id, pkts, count); } >=20 > +static __rte_always_inline uint16_t > +virtio_dev_rx_async_get_info_idx(uint16_t pkts_idx, > + uint16_t vq_size, uint16_t n_inflight) { > + return pkts_idx > n_inflight ? (pkts_idx - n_inflight) : > + (vq_size - n_inflight + pkts_idx) & (vq_size - 1); } > + > +static __rte_always_inline void > +virtio_dev_rx_async_submit_split_err(struct virtio_net *dev, > + struct vhost_virtqueue *vq, uint16_t queue_id, > + uint16_t last_idx, uint16_t shadow_idx) { > + uint16_t start_idx, pkts_idx, vq_size; > + uint64_t *async_pending_info; > + > + pkts_idx =3D vq->async_pkts_idx; > + async_pending_info =3D vq->async_pending_info; > + vq_size =3D vq->size; > + start_idx =3D virtio_dev_rx_async_get_info_idx(pkts_idx, > + vq_size, vq->async_pkts_inflight_n); > + > + while (likely((start_idx & (vq_size - 1)) !=3D pkts_idx)) { > + uint64_t n_seg =3D > + async_pending_info[(start_idx) & (vq_size - 1)] >> > + ASYNC_PENDING_INFO_N_SFT; > + > + while (n_seg) > + n_seg -=3D vq->async_ops.check_completed_copies(dev- > >vid, > + queue_id, 0, 1); > + } > + > + vq->async_pkts_inflight_n =3D 0; > + vq->batch_copy_nb_elems =3D 0; > + > + vq->shadow_used_idx =3D shadow_idx; > + vq->last_avail_idx =3D last_idx; > +} > + > +static __rte_noinline uint32_t > +virtio_dev_rx_async_submit_split(struct virtio_net *dev, > + struct vhost_virtqueue *vq, uint16_t queue_id, > + struct rte_mbuf **pkts, uint32_t count) { > + uint32_t pkt_idx =3D 0, pkt_burst_idx =3D 0; > + uint16_t num_buffers; > + struct buf_vector buf_vec[BUF_VECTOR_MAX]; > + uint16_t avail_head, last_idx, shadow_idx; > + > + struct rte_vhost_iov_iter *it_pool =3D vq->it_pool; > + struct iovec *vec_pool =3D vq->vec_pool; > + struct rte_vhost_async_desc tdes[MAX_PKT_BURST]; > + struct iovec *src_iovec =3D vec_pool; > + struct iovec *dst_iovec =3D vec_pool + (VHOST_MAX_ASYNC_VEC >> 1); > + struct rte_vhost_iov_iter *src_it =3D it_pool; > + struct rte_vhost_iov_iter *dst_it =3D it_pool + 1; > + uint16_t n_free_slot, slot_idx; > + int n_pkts =3D 0; > + > + avail_head =3D __atomic_load_n(&vq->avail->idx, __ATOMIC_ACQUIRE); > + last_idx =3D vq->last_avail_idx; > + shadow_idx =3D vq->shadow_used_idx; > + > + /* > + * The ordering between avail index and > + * desc reads needs to be enforced. > + */ > + rte_smp_rmb(); > + > + rte_prefetch0(&vq->avail->ring[vq->last_avail_idx & (vq->size - 1)]); > + > + for (pkt_idx =3D 0; pkt_idx < count; pkt_idx++) { > + uint32_t pkt_len =3D pkts[pkt_idx]->pkt_len + dev->vhost_hlen; > + uint16_t nr_vec =3D 0; > + > + if (unlikely(reserve_avail_buf_split(dev, vq, > + pkt_len, buf_vec, > &num_buffers, > + avail_head, &nr_vec) < 0)) { > + VHOST_LOG_DATA(DEBUG, > + "(%d) failed to get enough desc from vring\n", > + dev->vid); > + vq->shadow_used_idx -=3D num_buffers; > + break; > + } > + > + VHOST_LOG_DATA(DEBUG, "(%d) current index %d | end > index %d\n", > + dev->vid, vq->last_avail_idx, > + vq->last_avail_idx + num_buffers); > + > + if (async_mbuf_to_desc(dev, vq, pkts[pkt_idx], > + buf_vec, nr_vec, num_buffers, > + src_iovec, dst_iovec, src_it, dst_it) < 0) { > + vq->shadow_used_idx -=3D num_buffers; > + break; > + } > + > + slot_idx =3D (vq->async_pkts_idx + pkt_idx) & (vq->size - 1); > + if (src_it->count) { > + async_fill_desc(&tdes[pkt_burst_idx], src_it, dst_it); > + pkt_burst_idx++; > + vq->async_pending_info[slot_idx] =3D > + num_buffers | (src_it->nr_segs << 16); > + src_iovec +=3D src_it->nr_segs; > + dst_iovec +=3D dst_it->nr_segs; > + src_it +=3D 2; > + dst_it +=3D 2; > + } else { > + vq->async_pending_info[slot_idx] =3D num_buffers; > + vq->async_pkts_inflight_n++; > + } > + > + vq->last_avail_idx +=3D num_buffers; > + > + if (pkt_burst_idx >=3D VHOST_ASYNC_BATCH_THRESHOLD || > + (pkt_idx =3D=3D count - 1 && pkt_burst_idx)) { > + n_pkts =3D vq->async_ops.transfer_data(dev->vid, > + queue_id, tdes, 0, pkt_burst_idx); > + src_iovec =3D vec_pool; > + dst_iovec =3D vec_pool + (VHOST_MAX_ASYNC_VEC >> 1); > + src_it =3D it_pool; > + dst_it =3D it_pool + 1; > + > + if (unlikely(n_pkts < (int)pkt_burst_idx)) { > + vq->async_pkts_inflight_n +=3D > + n_pkts > 0 ? n_pkts : 0; > + virtio_dev_rx_async_submit_split_err(dev, > + vq, queue_id, last_idx, shadow_idx); > + return 0; > + } > + > + pkt_burst_idx =3D 0; > + vq->async_pkts_inflight_n +=3D n_pkts; > + } > + } > + > + if (pkt_burst_idx) { > + n_pkts =3D vq->async_ops.transfer_data(dev->vid, > + queue_id, tdes, 0, pkt_burst_idx); > + if (unlikely(n_pkts < (int)pkt_burst_idx)) { > + vq->async_pkts_inflight_n +=3D n_pkts > 0 ? n_pkts : 0; > + virtio_dev_rx_async_submit_split_err(dev, vq, queue_id, > + last_idx, shadow_idx); > + return 0; > + } > + > + vq->async_pkts_inflight_n +=3D n_pkts; > + } > + > + do_data_copy_enqueue(dev, vq); > + > + n_free_slot =3D vq->size - vq->async_pkts_idx; > + if (n_free_slot > pkt_idx) { > + rte_memcpy(&vq->async_pkts_pending[vq->async_pkts_idx], > + pkts, pkt_idx * sizeof(uintptr_t)); > + vq->async_pkts_idx +=3D pkt_idx; > + } else { > + rte_memcpy(&vq->async_pkts_pending[vq->async_pkts_idx], > + pkts, n_free_slot * sizeof(uintptr_t)); > + rte_memcpy(&vq->async_pkts_pending[0], > + &pkts[n_free_slot], > + (pkt_idx - n_free_slot) * sizeof(uintptr_t)); > + vq->async_pkts_idx =3D pkt_idx - n_free_slot; > + } > + > + if (likely(vq->shadow_used_idx)) > + async_flush_shadow_used_ring_split(dev, vq); > + > + return pkt_idx; > +} > + > +uint16_t rte_vhost_poll_enqueue_completed(int vid, uint16_t queue_id, > + struct rte_mbuf **pkts, uint16_t count) { > + struct virtio_net *dev =3D get_device(vid); > + struct vhost_virtqueue *vq; > + uint16_t n_pkts_cpl, n_pkts_put =3D 0, n_descs =3D 0; > + uint16_t start_idx, pkts_idx, vq_size; > + uint64_t *async_pending_info; > + > + VHOST_LOG_DATA(DEBUG, "(%d) %s\n", dev->vid, __func__); > + if (unlikely(!is_valid_virt_queue_idx(queue_id, 0, dev->nr_vring))) { > + VHOST_LOG_DATA(ERR, "(%d) %s: invalid virtqueue idx %d.\n", > + dev->vid, __func__, queue_id); > + return 0; > + } > + > + vq =3D dev->virtqueue[queue_id]; > + > + rte_spinlock_lock(&vq->access_lock); > + > + pkts_idx =3D vq->async_pkts_idx; > + async_pending_info =3D vq->async_pending_info; > + vq_size =3D vq->size; > + start_idx =3D virtio_dev_rx_async_get_info_idx(pkts_idx, > + vq_size, vq->async_pkts_inflight_n); > + > + n_pkts_cpl =3D > + vq->async_ops.check_completed_copies(vid, queue_id, 0, > count); > + > + rte_smp_wmb(); > + > + while (likely(((start_idx + n_pkts_put) & (vq_size - 1)) !=3D pkts_idx)= ) { > + uint64_t info =3D async_pending_info[ > + (start_idx + n_pkts_put) & (vq_size - 1)]; > + uint64_t n_segs; > + n_pkts_put++; > + n_descs +=3D info & ASYNC_PENDING_INFO_N_MSK; > + n_segs =3D info >> ASYNC_PENDING_INFO_N_SFT; > + > + if (n_segs) { > + if (!n_pkts_cpl || n_pkts_cpl < n_segs) { > + n_pkts_put--; > + n_descs -=3D info & > ASYNC_PENDING_INFO_N_MSK; > + if (n_pkts_cpl) { > + async_pending_info[ > + (start_idx + n_pkts_put) & > + (vq_size - 1)] =3D > + ((n_segs - n_pkts_cpl) << > + ASYNC_PENDING_INFO_N_SFT) | > + (info & > ASYNC_PENDING_INFO_N_MSK); > + n_pkts_cpl =3D 0; > + } > + break; > + } > + n_pkts_cpl -=3D n_segs; > + } > + } > + > + if (n_pkts_put) { > + vq->async_pkts_inflight_n -=3D n_pkts_put; > + __atomic_add_fetch(&vq->used->idx, n_descs, > __ATOMIC_RELEASE); > + > + vhost_vring_call_split(dev, vq); > + } > + > + if (start_idx + n_pkts_put <=3D vq_size) { > + rte_memcpy(pkts, &vq->async_pkts_pending[start_idx], > + n_pkts_put * sizeof(uintptr_t)); > + } else { > + rte_memcpy(pkts, &vq->async_pkts_pending[start_idx], > + (vq_size - start_idx) * sizeof(uintptr_t)); > + rte_memcpy(&pkts[vq_size - start_idx], vq- > >async_pkts_pending, > + (n_pkts_put - vq_size + start_idx) * sizeof(uintptr_t)); > + } > + > + rte_spinlock_unlock(&vq->access_lock); > + > + return n_pkts_put; > +} > + > +static __rte_always_inline uint32_t > +virtio_dev_rx_async_submit(struct virtio_net *dev, uint16_t queue_id, > + struct rte_mbuf **pkts, uint32_t count) { > + struct vhost_virtqueue *vq; > + uint32_t nb_tx =3D 0; > + bool drawback =3D false; > + > + VHOST_LOG_DATA(DEBUG, "(%d) %s\n", dev->vid, __func__); > + if (unlikely(!is_valid_virt_queue_idx(queue_id, 0, dev->nr_vring))) { > + VHOST_LOG_DATA(ERR, "(%d) %s: invalid virtqueue idx %d.\n", > + dev->vid, __func__, queue_id); > + return 0; > + } > + > + vq =3D dev->virtqueue[queue_id]; > + > + rte_spinlock_lock(&vq->access_lock); > + > + if (unlikely(vq->enabled =3D=3D 0)) > + goto out_access_unlock; > + > + if (unlikely(!vq->async_registered)) { > + drawback =3D true; > + goto out_access_unlock; > + } > + > + if (dev->features & (1ULL << VIRTIO_F_IOMMU_PLATFORM)) > + vhost_user_iotlb_rd_lock(vq); > + > + if (unlikely(vq->access_ok =3D=3D 0)) > + if (unlikely(vring_translate(dev, vq) < 0)) > + goto out; > + > + count =3D RTE_MIN((uint32_t)MAX_PKT_BURST, count); > + if (count =3D=3D 0) > + goto out; > + > + /* TODO: packed queue not implemented */ > + if (vq_is_packed(dev)) > + nb_tx =3D 0; > + else > + nb_tx =3D virtio_dev_rx_async_submit_split(dev, > + vq, queue_id, pkts, count); > + > +out: > + if (dev->features & (1ULL << VIRTIO_F_IOMMU_PLATFORM)) > + vhost_user_iotlb_rd_unlock(vq); > + > +out_access_unlock: > + rte_spinlock_unlock(&vq->access_lock); > + > + if (drawback) > + return rte_vhost_enqueue_burst(dev->vid, queue_id, pkts, > count); > + > + return nb_tx; > +} > + > +uint16_t > +rte_vhost_submit_enqueue_burst(int vid, uint16_t queue_id, > + struct rte_mbuf **pkts, uint16_t count) { > + struct virtio_net *dev =3D get_device(vid); > + > + if (!dev) > + return 0; > + > + if (unlikely(!(dev->flags & VIRTIO_DEV_BUILTIN_VIRTIO_NET))) { > + VHOST_LOG_DATA(ERR, > + "(%d) %s: built-in vhost net backend is disabled.\n", > + dev->vid, __func__); > + return 0; > + } > + > + return virtio_dev_rx_async_submit(dev, queue_id, pkts, count); } > + > static inline bool > virtio_net_with_host_offload(struct virtio_net *dev) { > -- > 2.18.4 Reviewed-by: Chenbo Xia