From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jason Wang Subject: Re: [PATCH net-next] virtio_net: ethtool tx napi configuration Date: Fri, 28 Sep 2018 07:39:57 +0800 Message-ID: <33e52452-3c11-2162-b403-cc1f81502e52@redhat.com> References: <20180909224449.203593-1-willemdebruijn.kernel@gmail.com> <2fd8d46f-7466-e507-af31-c587693beb6e@gmail.com> <4a3d69be-8651-e36d-bc14-5a3f1f23d155@redhat.com> <6db3c755-a1dc-dcc9-e110-bfc38143e83d@redhat.com> <0af0043d-c13c-e68e-795d-ef62901f57fc@redhat.com> <841d0fc6-d9a2-84d0-f794-7a7b2019fa17@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: 8bit Cc: Florian Fainelli , Network Development , David Miller , caleb.raitto@gmail.com, "Michael S. Tsirkin" , "Jon Olson (Google Drive)" , Willem de Bruijn To: Willem de Bruijn Return-path: Received: from mx1.redhat.com ([209.132.183.28]:33950 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1725917AbeI1GAz (ORCPT ); Fri, 28 Sep 2018 02:00:55 -0400 In-Reply-To: Content-Language: en-US Sender: netdev-owner@vger.kernel.org List-ID: On 2018年09月27日 21:53, Willem de Bruijn wrote: > On Thu, Sep 27, 2018 at 4:51 AM Jason Wang wrote: >> >> >> On 2018年09月14日 12:46, Willem de Bruijn wrote: >>>> I'm not sure I get this. If we don't enable tx napi, we tend to delay TX >>>> interrupt if we found the ring is about to full to avoid interrupt >>>> storm, so we're probably ok in this case. >>> I'm only concerned about the transition state when converting from >>> napi to no-napi when the queue is stopped and tx interrupt disabled. >>> >>> With napi mode the interrupt is only disabled if napi is scheduled, >>> in which case it will eventually reenable the interrupt. But when >>> switching to no-napi mode in this state no progress will be made. >>> >>> But it seems this cannot happen. When converting to no-napi >>> mode, set_coalesce waits for napi to complete in napi_disable. >>> So the interrupt should always start enabled when transitioning >>> into no-napi mode. >> An update, I meet a hang in napi_disalbe(). But it's hard to be >> reproduced. I tend to choose a easy way like V1 that only allow the >> switching when device is down. > I agree. > >> I will post the patch after a vacation. (or you can post if it was >> urgent for you). > If you have time to review and add your signed-off-by, I can post it. > It's a pretty small diff at this point. > > But no rush, we can also wait until after your vacation. Then let me post it after the vacation. > > I also need to look at a patch to toggle LRO using ethtool, btw. Interesting, we've already did something similar during XDP. The GUEST_TSO_XXX part may need some private flags I believe. Thanks