From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 9861EC19F2C for ; Thu, 28 Jul 2022 08:18:25 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S235102AbiG1ISW (ORCPT ); Thu, 28 Jul 2022 04:18:22 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:34578 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S234999AbiG1IST (ORCPT ); Thu, 28 Jul 2022 04:18:19 -0400 Received: from out30-42.freemail.mail.aliyun.com (out30-42.freemail.mail.aliyun.com [115.124.30.42]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 0C9D55A2DE; Thu, 28 Jul 2022 01:18:15 -0700 (PDT) X-Alimail-AntiSpam: AC=PASS;BC=-1|-1;BR=01201311R721e4;CH=green;DM=||false|;DS=||;FP=0|-1|-1|-1|0|-1|-1|-1;HT=ay29a033018045168;MF=xuanzhuo@linux.alibaba.com;NM=1;PH=DS;RN=36;SR=0;TI=SMTPD_---0VKefuGf_1658996289; Received: from localhost(mailfrom:xuanzhuo@linux.alibaba.com fp:SMTPD_---0VKefuGf_1658996289) by smtp.aliyun-inc.com; Thu, 28 Jul 2022 16:18:10 +0800 Message-ID: <1658995783.1026692-1-xuanzhuo@linux.alibaba.com> Subject: Re: [PATCH v13 16/42] virtio_ring: split: introduce virtqueue_resize_split() Date: Thu, 28 Jul 2022 16:09:43 +0800 From: Xuan Zhuo To: Jason Wang Cc: Richard Weinberger , Anton Ivanov , Johannes Berg , "Michael S. Tsirkin" , "David S. Miller" , Eric Dumazet , Jakub Kicinski , Paolo Abeni , Hans de Goede , Mark Gross , Vadim Pasternak , Bjorn Andersson , Mathieu Poirier , Cornelia Huck , Halil Pasic , Eric Farman , Heiko Carstens , Vasily Gorbik , Alexander Gordeev , Christian Borntraeger , Sven Schnelle , Alexei Starovoitov , Daniel Borkmann , Jesper Dangaard Brouer , John Fastabend , Vincent Whitchurch , linux-um@lists.infradead.org, netdev , platform-driver-x86@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, kvm , "open list:XDP (eXpress Data Path)" , Kangjie Xu , virtualization References: <20220726072225.19884-1-xuanzhuo@linux.alibaba.com> <20220726072225.19884-17-xuanzhuo@linux.alibaba.com> <15aa26f2-f8af-5dbd-f2b2-9270ad873412@redhat.com> <1658907413.1860468-2-xuanzhuo@linux.alibaba.com> <1658992162.584327-1-xuanzhuo@linux.alibaba.com> In-Reply-To: Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Precedence: bulk List-ID: X-Mailing-List: platform-driver-x86@vger.kernel.org On Thu, 28 Jul 2022 15:42:50 +0800, Jason Wang wrote: > On Thu, Jul 28, 2022 at 3:24 PM Xuan Zhuo wr= ote: > > > > On Thu, 28 Jul 2022 10:38:51 +0800, Jason Wang wr= ote: > > > On Wed, Jul 27, 2022 at 3:44 PM Xuan Zhuo wrote: > > > > > > > > On Wed, 27 Jul 2022 11:12:19 +0800, Jason Wang wrote: > > > > > > > > > > =E5=9C=A8 2022/7/26 15:21, Xuan Zhuo =E5=86=99=E9=81=93: > > > > > > virtio ring split supports resize. > > > > > > > > > > > > Only after the new vring is successfully allocated based on the= new num, > > > > > > we will release the old vring. In any case, an error is returne= d, > > > > > > indicating that the vring still points to the old vring. > > > > > > > > > > > > In the case of an error, re-initialize(virtqueue_reinit_split()= ) the > > > > > > virtqueue to ensure that the vring can be used. > > > > > > > > > > > > Signed-off-by: Xuan Zhuo > > > > > > Acked-by: Jason Wang > > > > > > --- > > > > > > drivers/virtio/virtio_ring.c | 34 +++++++++++++++++++++++++++= +++++++ > > > > > > 1 file changed, 34 insertions(+) > > > > > > > > > > > > diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virt= io_ring.c > > > > > > index b6fda91c8059..58355e1ac7d7 100644 > > > > > > --- a/drivers/virtio/virtio_ring.c > > > > > > +++ b/drivers/virtio/virtio_ring.c > > > > > > @@ -220,6 +220,7 @@ static struct virtqueue *__vring_new_virtqu= eue(unsigned int index, > > > > > > void (*callback)(str= uct virtqueue *), > > > > > > const char *name); > > > > > > static struct vring_desc_extra *vring_alloc_desc_extra(unsign= ed int num); > > > > > > +static void vring_free(struct virtqueue *_vq); > > > > > > > > > > > > /* > > > > > > * Helpers. > > > > > > @@ -1117,6 +1118,39 @@ static struct virtqueue *vring_create_vi= rtqueue_split( > > > > > > return vq; > > > > > > } > > > > > > > > > > > > +static int virtqueue_resize_split(struct virtqueue *_vq, u32 n= um) > > > > > > +{ > > > > > > + struct vring_virtqueue_split vring_split =3D {}; > > > > > > + struct vring_virtqueue *vq =3D to_vvq(_vq); > > > > > > + struct virtio_device *vdev =3D _vq->vdev; > > > > > > + int err; > > > > > > + > > > > > > + err =3D vring_alloc_queue_split(&vring_split, vdev, num, > > > > > > + vq->split.vring_align, > > > > > > + vq->split.may_reduce_num); > > > > > > + if (err) > > > > > > + goto err; > > > > > > > > > > > > > > > I think we don't need to do anything here? > > > > > > > > Am I missing something? > > > > > > I meant it looks to me most of the virtqueue_reinit() is unnecessary. > > > We probably only need to reinit avail/used idx there. > > > > > > In this function, we can indeed remove some code. > > > > > static void virtqueue_reinit_split(struct vring_virtqueue *vq) > > > { > > > int size, i; > > > > > > memset(vq->split.vring.desc, 0, vq->split.queue_size_in= _bytes); > > > > > > size =3D sizeof(struct vring_desc_state_split) * vq->sp= lit.vring.num; > > > memset(vq->split.desc_state, 0, size); > > > > > > size =3D sizeof(struct vring_desc_extra) * vq->split.vr= ing.num; > > > memset(vq->split.desc_extra, 0, size); > > > > These memsets can be removed, and theoretically it will not cause any > > exceptions. > > Yes, otherwise we have bugs in detach_buf(). > > > > > > > > > > > > > > > for (i =3D 0; i < vq->split.vring.num - 1; i++) > > > vq->split.desc_extra[i].next =3D i + 1; > > > > This can also be removed, but we need to record free_head that will bee= n update > > inside virtqueue_init(). > > We can simply keep free_head unchanged? Otherwise it's a bug somewhere I = guess. > > > > > > > > > > virtqueue_init(vq, vq->split.vring.num); > > > > There are some operations in this, which can also be skipped, such as s= etting > > use_dma_api. But I think calling this function directly will be more co= nvenient > > for maintenance. > > I don't see anything that is necessary here. These three are currently inside virtqueue_init() vq->last_used_idx =3D 0; vq->event_triggered =3D false; vq->num_added =3D 0; Thanks. > > > > > > > > virtqueue_vring_init_split(&vq->split, vq); > > > > virtqueue_vring_init_split() is necessary. > > Right. > > > > > > } > > > > Another method, we can take out all the variables to be reinitialized > > separately, and repackage them into a new function. I don=E2=80=99t thi= nk it=E2=80=99s worth > > it, because this path will only be reached if the memory allocation fai= ls, which > > is a rare occurrence. In this case, doing so will increase the cost of > > maintenance. If you think so also, I will remove the above memset in th= e next > > version. > > I agree. > > Thanks > > > > > Thanks. > > > > > > > > > > Thanks > > > > > > > > > > > > > > > > > > > > > > > + > > > > > > + err =3D vring_alloc_state_extra_split(&vring_split); > > > > > > + if (err) { > > > > > > + vring_free_split(&vring_split, vdev); > > > > > > + goto err; > > > > > > > > > > > > > > > I suggest to move vring_free_split() into a dedicated error label. > > > > > > > > Will change. > > > > > > > > Thanks. > > > > > > > > > > > > > > > > > > Thanks > > > > > > > > > > > > > > > > + } > > > > > > + > > > > > > + vring_free(&vq->vq); > > > > > > + > > > > > > + virtqueue_vring_init_split(&vring_split, vq); > > > > > > + > > > > > > + virtqueue_init(vq, vring_split.vring.num); > > > > > > + virtqueue_vring_attach_split(vq, &vring_split); > > > > > > + > > > > > > + return 0; > > > > > > + > > > > > > +err: > > > > > > + virtqueue_reinit_split(vq); > > > > > > + return -ENOMEM; > > > > > > +} > > > > > > + > > > > > > > > > > > > /* > > > > > > * Packed ring specific functions - *_packed(). > > > > > > > > > > > > > > >