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 1F686A04C0; Tue, 29 Sep 2020 08:39:25 +0200 (CEST) Received: from [92.243.14.124] (localhost [127.0.0.1]) by dpdk.org (Postfix) with ESMTP id B23EE1D71D; Tue, 29 Sep 2020 08:38:47 +0200 (CEST) Received: from mga14.intel.com (mga14.intel.com [192.55.52.115]) by dpdk.org (Postfix) with ESMTP id 787201D704 for ; Tue, 29 Sep 2020 08:38:35 +0200 (CEST) IronPort-SDR: y5LN6Psr8UxrG1GbIZQpBRH7NRA2ho5xnKdkN5l804XCULzKbjVd/lUlHL6IlBqQ229qzVcBa8 G052khLbno2Q== X-IronPort-AV: E=McAfee;i="6000,8403,9758"; a="161356245" X-IronPort-AV: E=Sophos;i="5.77,317,1596524400"; d="scan'208";a="161356245" X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from orsmga008.jf.intel.com ([10.7.209.65]) by fmsmga103.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 28 Sep 2020 23:38:35 -0700 IronPort-SDR: NSi4iUq3wQFcVUrlKkaFI0jJPAYQB7juO/UyEzSBrWsmi0DDcsNMrVDBdyLphbebomB6evTs1+ aDklzhiikD6A== X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.77,317,1596524400"; d="scan'208";a="340745847" Received: from npg-dpdk-patrickfu-casc2.sh.intel.com ([10.67.119.92]) by orsmga008.jf.intel.com with ESMTP; 28 Sep 2020 23:38:32 -0700 From: Patrick Fu To: dev@dpdk.org, maxime.coquelin@redhat.com, chenbo.xia@intel.com Cc: zhihong.wang@intel.com, cheng1.jiang@intel.com, patrick.fu@intel.com Date: Tue, 29 Sep 2020 14:29:57 +0800 Message-Id: <20200929062958.2832351-4-patrick.fu@intel.com> X-Mailer: git-send-email 2.18.4 In-Reply-To: <20200929062958.2832351-1-patrick.fu@intel.com> References: <20200911015316.1903181-1-patrick.fu@intel.com> <20200929062958.2832351-1-patrick.fu@intel.com> Subject: [dpdk-dev] [PATCH v2 3/4] vhost: fix async vector buffer overrun 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" Add check on the async vector buffer usage to prevent the buf overrun. If the unused vector buffer is not sufficient to prepare for next packet's iov creation, an async transfer will be triggered immediately to free the vector buffer. Fixes: 78639d54563a ("vhost: introduce async enqueue registration API") Signed-off-by: Patrick Fu --- lib/librte_vhost/vhost.h | 2 +- lib/librte_vhost/virtio_net.c | 13 ++++++++++++- 2 files changed, 13 insertions(+), 2 deletions(-) diff --git a/lib/librte_vhost/vhost.h b/lib/librte_vhost/vhost.h index f0ee00c73..34197ce99 100644 --- a/lib/librte_vhost/vhost.h +++ b/lib/librte_vhost/vhost.h @@ -47,7 +47,7 @@ #define MAX_PKT_BURST 32 #define VHOST_MAX_ASYNC_IT (MAX_PKT_BURST * 2) -#define VHOST_MAX_ASYNC_VEC (BUF_VECTOR_MAX * 2) +#define VHOST_MAX_ASYNC_VEC (BUF_VECTOR_MAX * 4) #define PACKED_DESC_ENQUEUE_USED_FLAG(w) \ ((w) ? (VRING_DESC_F_AVAIL | VRING_DESC_F_USED | VRING_DESC_F_WRITE) : \ diff --git a/lib/librte_vhost/virtio_net.c b/lib/librte_vhost/virtio_net.c index 68ead9a71..6ba64c12e 100644 --- a/lib/librte_vhost/virtio_net.c +++ b/lib/librte_vhost/virtio_net.c @@ -1492,6 +1492,7 @@ virtio_dev_rx_async_submit_split(struct virtio_net *dev, struct rte_vhost_iov_iter *dst_it = it_pool + 1; uint16_t n_free_slot, slot_idx; uint16_t pkt_err = 0; + uint16_t segs_await = 0; struct async_inflight_info *pkts_info = vq->async_pkts_info; int n_pkts = 0; @@ -1540,6 +1541,7 @@ virtio_dev_rx_async_submit_split(struct virtio_net *dev, dst_iovec += dst_it->nr_segs; src_it += 2; dst_it += 2; + segs_await += src_it->nr_segs; } else { pkts_info[slot_idx].info = num_buffers; vq->async_pkts_inflight_n++; @@ -1547,14 +1549,23 @@ virtio_dev_rx_async_submit_split(struct virtio_net *dev, vq->last_avail_idx += num_buffers; + /* + * conditions to trigger async device transfer: + * - bufferred packet number reaches transfer threshold + * - this is the last packet in the burst enqueue + * - unused async iov number is less than max vhost vector + */ if (pkt_burst_idx >= VHOST_ASYNC_BATCH_THRESHOLD || - (pkt_idx == count - 1 && pkt_burst_idx)) { + (pkt_idx == count - 1 && pkt_burst_idx) || + (VHOST_MAX_ASYNC_VEC / 2 - segs_await < + BUF_VECTOR_MAX)) { n_pkts = vq->async_ops.transfer_data(dev->vid, queue_id, tdes, 0, pkt_burst_idx); src_iovec = vec_pool; dst_iovec = vec_pool + (VHOST_MAX_ASYNC_VEC >> 1); src_it = it_pool; dst_it = it_pool + 1; + segs_await = 0; vq->async_pkts_inflight_n += n_pkts; if (unlikely(n_pkts < (int)pkt_burst_idx)) { -- 2.18.4