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 36FC7C19F2C for ; Thu, 28 Jul 2022 07:43:12 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S234648AbiG1HnK (ORCPT ); Thu, 28 Jul 2022 03:43:10 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:34772 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S234652AbiG1HnH (ORCPT ); Thu, 28 Jul 2022 03:43:07 -0400 Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.133.124]) by lindbergh.monkeyblade.net (Postfix) with ESMTP id EBF28606AD for ; Thu, 28 Jul 2022 00:43:05 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1658994184; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=MiRkDMIa3SaNcNAA4LhevtztHl/rFve9e8RNCF7+zFQ=; b=eFIIcTPhqtBqOXdkN8FJu7gTSWotVhGDlBI0DsmaEuQERG/U5ixemzBzKlf9wdltI5334P 5y0+S1nbso3PrlP27Q9ZaA3d6av2ixSrTH8KI9c11SXy4R9EH73U0lUplXJixexOfEH+is M0oj18pfomfXKE2+Wb5HkCpTE8K0zbU= Received: from mail-lf1-f70.google.com (mail-lf1-f70.google.com [209.85.167.70]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id us-mta-354-9X9IPlSqNVqFSIaizzyMiA-1; Thu, 28 Jul 2022 03:43:03 -0400 X-MC-Unique: 9X9IPlSqNVqFSIaizzyMiA-1 Received: by mail-lf1-f70.google.com with SMTP id w20-20020a194914000000b00489e66662d1so394346lfa.1 for ; Thu, 28 Jul 2022 00:43:03 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:mime-version:references:in-reply-to:from:date :message-id:subject:to:cc:content-transfer-encoding; bh=MiRkDMIa3SaNcNAA4LhevtztHl/rFve9e8RNCF7+zFQ=; b=T0+afmhZtttdDVLV4dTFkFMe+76qpQKFRoItlvW4YTqDif6c6wezjUcRnLGpbxMlyO h3yahHsqiNTCDjnqHbEY5fYVWu51thp0zm4xjoU54xhjZzixFw1I9zFc6MT2WAF4tNjC FgOImLbNIQ7FOGYn1lXyoie0nUK2Omfny9z7uek5lwuoNClU54ON0ykTM8OCbZNb+AL8 6QoHAHAh/9VLxa3dQ4N+Ptmq7CjWwcQgvqBlAg8fhMAHuR9qAmhCFvXDBJHfaeom1vBc alXU75GtvBA8n04DSTAA5vizW5znjHwtolOBSBRKT+KfgWh+K0iIbTLIOVoJ9V2XxBPp PwPw== X-Gm-Message-State: AJIora/diXAbxsNwdm+4/6+G3CPl9Il78T9HlkKZwwECHclnvOUk/pPe 4Whj+c7RxjZHuz02/6NPNETTjBPiVPzWKtd1r/B7KE9qxzYXTQOnx7YmiAty4iasrXuTt53EKfO soJGXay/EVTjnILgiN16XYqb2otASNvshHu7Oj1QGZX0+E8Np4A== X-Received: by 2002:a2e:a99e:0:b0:25e:a54:8328 with SMTP id x30-20020a2ea99e000000b0025e0a548328mr5494972ljq.141.1658994181686; Thu, 28 Jul 2022 00:43:01 -0700 (PDT) X-Google-Smtp-Source: AGRyM1u68nBKmGtn3ybA4uLX2AOBJSklXcUMjbmYLhRoC0rjTaSlVKsgAzpWU4s1tY/769/tc0LwTRbMsEcj8iXfSPw= X-Received: by 2002:a2e:a99e:0:b0:25e:a54:8328 with SMTP id x30-20020a2ea99e000000b0025e0a548328mr5494947ljq.141.1658994181424; Thu, 28 Jul 2022 00:43:01 -0700 (PDT) MIME-Version: 1.0 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: <1658992162.584327-1-xuanzhuo@linux.alibaba.com> From: Jason Wang Date: Thu, 28 Jul 2022 15:42:50 +0800 Message-ID: Subject: Re: [PATCH v13 16/42] virtio_ring: split: introduce virtqueue_resize_split() To: Xuan Zhuo 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 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, Jul 28, 2022 at 3:24 PM Xuan Zhuo wrot= e: > > On Thu, 28 Jul 2022 10:38:51 +0800, Jason Wang wrot= e: > > 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 n= ew num, > > > > > we will release the old vring. In any case, an error is returned, > > > > > 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/virtio= _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_virtqueu= e(unsigned int index, > > > > > void (*callback)(struc= t virtqueue *), > > > > > const char *name); > > > > > static struct vring_desc_extra *vring_alloc_desc_extra(unsigned= int num); > > > > > +static void vring_free(struct virtqueue *_vq); > > > > > > > > > > /* > > > > > * Helpers. > > > > > @@ -1117,6 +1118,39 @@ static struct virtqueue *vring_create_virt= queue_split( > > > > > return vq; > > > > > } > > > > > > > > > > +static int virtqueue_resize_split(struct virtqueue *_vq, u32 num= ) > > > > > +{ > > > > > + 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_b= ytes); > > > > size =3D sizeof(struct vring_desc_state_split) * vq->spli= t.vring.num; > > memset(vq->split.desc_state, 0, size); > > > > size =3D sizeof(struct vring_desc_extra) * vq->split.vrin= g.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 been = update > inside virtqueue_init(). We can simply keep free_head unchanged? Otherwise it's a bug somewhere I gu= ess. > > > > > virtqueue_init(vq, vq->split.vring.num); > > There are some operations in this, which can also be skipped, such as set= ting > use_dma_api. But I think calling this function directly will be more conv= enient > for maintenance. I don't see anything that is necessary here. > > > > 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 think= it=E2=80=99s worth > it, because this path will only be reached if the memory allocation fails= , 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 the = 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(). > > > > > > > > > >