From patchwork Tue Dec 11 13:48:01 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Maxime Coquelin X-Patchwork-Id: 48636 Return-Path: X-Original-To: patchwork@dpdk.org Delivered-To: patchwork@dpdk.org Received: from [92.243.14.124] (localhost [127.0.0.1]) by dpdk.org (Postfix) with ESMTP id EF4334F9C; Tue, 11 Dec 2018 14:48:26 +0100 (CET) Received: from mx1.redhat.com (mx1.redhat.com [209.132.183.28]) by dpdk.org (Postfix) with ESMTP id 4F3794CB5 for ; Tue, 11 Dec 2018 14:48:25 +0100 (CET) Received: from smtp.corp.redhat.com (int-mx02.intmail.prod.int.phx2.redhat.com [10.5.11.12]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id 96A4158E29; Tue, 11 Dec 2018 13:48:24 +0000 (UTC) Received: from localhost.localdomain (ovpn-112-28.ams2.redhat.com [10.36.112.28]) by smtp.corp.redhat.com (Postfix) with ESMTP id B8FCC60C64; Tue, 11 Dec 2018 13:48:07 +0000 (UTC) From: Maxime Coquelin To: dev@dpdk.org, jfreimann@redhat.com, tiwei.bie@intel.com, zhihong.wang@intel.com Cc: Maxime Coquelin Date: Tue, 11 Dec 2018 14:48:01 +0100 Message-Id: <20181211134804.10318-1-maxime.coquelin@redhat.com> X-Scanned-By: MIMEDefang 2.79 on 10.5.11.12 X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.39]); Tue, 11 Dec 2018 13:48:24 +0000 (UTC) Subject: [dpdk-dev] [PATCH v2 0/3] net/virtio: Rx paths improvements 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" Initial version of this series did merge out-of-order mergeable and non-mergeable receive paths, but Intel STV team highlighted some performance regression when using multiqueue with two cores enqueueing descs on host, while a single core dequeues the two queues. I didn't manage to close the performance gap, so I decided to give-up this refactoring. But while trying to optimize, I reworked the meargeable function so that it looks like the in-order one. I.e. descriptors are now dequeued in batches, so are descriptors refilled. Doing that, I measure a perfromance gain of 6% when doing rxonly microbenchmark with two cores on host, one in guest. The other two patches of the series haven't been modified. Maxime Coquelin (3): net/virtio: inline refill and offload helpers net/virtio: add non-mergeable support to in-order path net/virtio: improve batching in mergeable path drivers/net/virtio/virtio_ethdev.c | 11 +- drivers/net/virtio/virtio_ethdev.h | 2 +- drivers/net/virtio/virtio_rxtx.c | 255 ++++++++++++++++------------- 3 files changed, 141 insertions(+), 127 deletions(-)