From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jason Wang Subject: Re: [PATCH net-next 2/2] virtio-net: simplify XDP handling in small buffer Date: Thu, 1 Mar 2018 17:24:35 +0800 Message-ID: <5c20a958-55c6-5a5f-c0e9-dc6f7c0e4db5@redhat.com> References: <1519874345-10235-1-git-send-email-jasowang@redhat.com> <1519874345-10235-3-git-send-email-jasowang@redhat.com> <20180301090206.04e13a71@redhat.com> <20180301101553.55c4c6bd@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: 8bit Cc: mst@redhat.com, virtualization@lists.linux-foundation.org, netdev@vger.kernel.org, linux-kernel@vger.kernel.org, john.fastabend@gmail.com To: Jesper Dangaard Brouer Return-path: In-Reply-To: <20180301101553.55c4c6bd@redhat.com> Content-Language: en-US Sender: linux-kernel-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On 2018年03月01日 17:15, Jesper Dangaard Brouer wrote: > On Thu, 1 Mar 2018 16:49:24 +0800 > Jason Wang wrote: > >>> 2. This can easily cause out-of-order packets. >> I may miss something, but it looks to me packets were still delivered >> in order? Or you mean the packets that was dropped by cpumap? > No. Packets can now travel two code paths to the egress device. (1) XDP > native via ndp_xdp_xmit via direct delivery into a lockfree/dedicated > TX queue, (2) via normal network stack which can involve being queue in > a qdisc. Do you see the possibility of the reorder now? > I see, thanks. But consider this could only happen for first few packets, not sure it was worth to worry about it. Thanks