From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751681AbbH1CmJ (ORCPT ); Thu, 27 Aug 2015 22:42:09 -0400 Received: from mx1.redhat.com ([209.132.183.28]:33151 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751365AbbH1CmG (ORCPT ); Thu, 27 Aug 2015 22:42:06 -0400 Message-ID: <55DFCA7B.8050908@redhat.com> Date: Fri, 28 Aug 2015 10:42:03 +0800 From: Jason Wang User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.8.0 MIME-Version: 1.0 To: "Michael S. Tsirkin" CC: vyasevic@redhat.com, netdev@vger.kernel.org, linux-kernel@vger.kernel.org, Patrick McHardy Subject: Re: [PATCH net-next] macvtap/macvlan: use IFF_NO_QUEUE References: <1440405192-25926-1-git-send-email-jasowang@redhat.com> <20150825131608-mutt-send-email-mst@redhat.com> <55DC51C1.5090201@redhat.com> <55DC98A3.7010601@redhat.com> <55DD527A.8040806@redhat.com> <20150827134102-mutt-send-email-mst@redhat.com> In-Reply-To: <20150827134102-mutt-send-email-mst@redhat.com> Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 08/27/2015 06:43 PM, Michael S. Tsirkin wrote: > On Wed, Aug 26, 2015 at 01:45:30PM +0800, Jason Wang wrote: >> >> On 08/26/2015 12:32 AM, Vlad Yasevich wrote: >>> On 08/25/2015 07:30 AM, Jason Wang wrote: >>>> On 08/25/2015 06:17 PM, Michael S. Tsirkin wrote: >>>>> On Mon, Aug 24, 2015 at 04:33:12PM +0800, Jason Wang wrote: >>>>>>> For macvlan, switch to use IFF_NO_QUEUE instead of tx_queue_len = 0. >>>>>>> >>>>>>> For macvtap, after commit 6acf54f1cf0a6747bac9fea26f34cfc5a9029523 >>>>>>> ("macvtap: Add support of packet capture on macvtap >>>>>>> device."). Multiqueue macvtap suffers from single qdisc lock >>>>>>> contention. This is because macvtap claims a non zero tx_queue_len and >>>>>>> it reuses this value as it socket receive queue size.Thanks to >>>>>>> IFF_NO_QUEUE, we can remove the lock contention without breaking >>>>>>> existing socket receive queue length logic. >>>>>>> >>>>>>> Cc: Patrick McHardy >>>>>>> Cc: Vladislav Yasevich >>>>>>> Cc: Michael S. Tsirkin >>>>>>> Signed-off-by: Jason Wang >>>>> Seems to make sense. Give me a day or two to get over the jet lag >>>>> (and get out from under the pile of mail accumulated while I was traveling), >>>>> I'll review properly and ack. >>>>> >>>> A note on this patch: only default qdisc were removed but we don't lose >>>> the ability to attach a qdisc to macvtap (though it may cause lock >>>> contention on multiqueue case). >>>> >>> Wouldn't that lock contention be solved if we really had multiple queues >>> for multi-queue macvtaps? >>> >>> -vlad >> Yes, but this introduce another layer of txq locks contention? > I don't follow - why does it? Could you clarify please? I believe Vlad wants to remove NETIF_F_LLTX. If yes, core will do an extra tx lock at macvlan layer. > >> And it >> also needs macvlan multiqueue support. We used to do something like this >> but switch to NETIF_F_LLTX finally. You may refer: >> >> 2c11455321f37da6fe6cc36353149f9ac9183334 macvlan: add multiqueue capability >> 8ffab51b3dfc54876f145f15b351c41f3f703195 macvlan: lockless tx path > My concern is that the moment someone configures a non-standard qdisc > scalability suddenly disappears. That would also be tricky to debug in the > field as not a lot of developers use non-standard qdiscs. > What do you think? Probably not an issue. Non-standard qdisc may need be attached manually after device creation, and we don't lose this ability with this patch. (Unless somebody changes default_qdisc). Actually, user before 6acf54f1cf0a6747bac9fea26f34cfc5a9029523 does not expect any qdisc work for macvtap like other stacked devices. This patch also restore this.