From patchwork Mon Jul 5 07:19:09 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Jason Wang X-Patchwork-Id: 12358367 X-Patchwork-Delegate: kuba@kernel.org Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-14.4 required=3.0 tests=BAYES_00,DKIMWL_WL_HIGH, DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_CR_TRAILER,INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS, UNWANTED_LANGUAGE_BODY autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id AC510C07E98 for ; Mon, 5 Jul 2021 07:19:28 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 851186108D for ; Mon, 5 Jul 2021 07:19:28 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S229934AbhGEHWD (ORCPT ); Mon, 5 Jul 2021 03:22:03 -0400 Received: from us-smtp-delivery-124.mimecast.com ([216.205.24.124]:41162 "EHLO us-smtp-delivery-124.mimecast.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229925AbhGEHWC (ORCPT ); Mon, 5 Jul 2021 03:22:02 -0400 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1625469566; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version: content-transfer-encoding:content-transfer-encoding; bh=vAxaFW9a7xpmG8B2uRDAj3X/DHOBu21mlVjLtanDQ7M=; b=UbG5KCYuHSXiXBGSwB3RLgfJ5+v/uupXdPlfLXdh+U3S+vagpZNzT10/0rUa1clYTiazH1 qESzOP584r+SLgQaWqu0YbPKeHnrlUw1IBZKoLiLIVtCnC3qOfcArDugmlztzhAOkCjPnl ofJoARXbIgeHBAetk2Jqk3Q4NAAqK08= Received: from mimecast-mx01.redhat.com (mimecast-mx01.redhat.com [209.132.183.4]) (Using TLS) by relay.mimecast.com with ESMTP id us-mta-237-9nB3Jf5cMJyCjT3G0w6Ckg-1; Mon, 05 Jul 2021 03:19:24 -0400 X-MC-Unique: 9nB3Jf5cMJyCjT3G0w6Ckg-1 Received: from smtp.corp.redhat.com (int-mx04.intmail.prod.int.phx2.redhat.com [10.5.11.14]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mimecast-mx01.redhat.com (Postfix) with ESMTPS id 68C0D801B0F; Mon, 5 Jul 2021 07:19:23 +0000 (UTC) Received: from localhost.localdomain (ovpn-13-21.pek2.redhat.com [10.72.13.21]) by smtp.corp.redhat.com (Postfix) with ESMTP id B0B785D9DC; Mon, 5 Jul 2021 07:19:13 +0000 (UTC) From: Jason Wang To: mst@redhat.com, jasowang@redhat.com, virtualization@lists.linux-foundation.org, linux-kernel@vger.kernel.org, kvm@vger.kernel.org Cc: netdev@vger.kernel.org, xieyongji@bytedance.com, stefanha@redhat.com Subject: [PATCH 1/2] vdpa: support per virtqueue max queue size Date: Mon, 5 Jul 2021 15:19:09 +0800 Message-Id: <20210705071910.31965-1-jasowang@redhat.com> MIME-Version: 1.0 X-Scanned-By: MIMEDefang 2.79 on 10.5.11.14 Precedence: bulk List-ID: X-Mailing-List: netdev@vger.kernel.org Virtio spec allows the device to specify the per virtqueue max queue size. vDPA needs to adapt to this flexibility. E.g Qemu advertise a small control virtqueue for virtio-net. So this patch adds a index parameter to get_vq_num_max bus operations for the device to report its per virtqueue max queue size. Both VHOST_VDPA_GET_VRING_NUM and VDPA_ATTR_DEV_MAX_VQ_SIZE assume a global maximum size. So we iterate all the virtqueues to return the minimal size in this case. Actually, the VHOST_VDPA_GET_VRING_NUM is not a must for the userspace. Userspace may choose to check the VHOST_SET_VRING_NUM for proving or validating the maximum virtqueue size. Anyway, we can invent a per vq version of VHOST_VDPA_GET_VRING_NUM in the future if it's necessary. Signed-off-by: Jason Wang Reviewed-by: Stefan Hajnoczi --- drivers/vdpa/ifcvf/ifcvf_main.c | 2 +- drivers/vdpa/mlx5/net/mlx5_vnet.c | 2 +- drivers/vdpa/vdpa.c | 22 +++++++++++++++++++++- drivers/vdpa/vdpa_sim/vdpa_sim.c | 2 +- drivers/vdpa/virtio_pci/vp_vdpa.c | 2 +- drivers/vhost/vdpa.c | 9 ++++++--- drivers/virtio/virtio_vdpa.c | 2 +- include/linux/vdpa.h | 5 ++++- 8 files changed, 36 insertions(+), 10 deletions(-) diff --git a/drivers/vdpa/ifcvf/ifcvf_main.c b/drivers/vdpa/ifcvf/ifcvf_main.c index ab0ab5cf0f6e..646b340db2af 100644 --- a/drivers/vdpa/ifcvf/ifcvf_main.c +++ b/drivers/vdpa/ifcvf/ifcvf_main.c @@ -254,7 +254,7 @@ static void ifcvf_vdpa_set_status(struct vdpa_device *vdpa_dev, u8 status) ifcvf_set_status(vf, status); } -static u16 ifcvf_vdpa_get_vq_num_max(struct vdpa_device *vdpa_dev) +static u16 ifcvf_vdpa_get_vq_num_max(struct vdpa_device *vdpa_dev, u16 qid) { return IFCVF_QUEUE_MAX; } diff --git a/drivers/vdpa/mlx5/net/mlx5_vnet.c b/drivers/vdpa/mlx5/net/mlx5_vnet.c index dda5dc6f7737..afd6114d07b0 100644 --- a/drivers/vdpa/mlx5/net/mlx5_vnet.c +++ b/drivers/vdpa/mlx5/net/mlx5_vnet.c @@ -1584,7 +1584,7 @@ static void mlx5_vdpa_set_config_cb(struct vdpa_device *vdev, struct vdpa_callba } #define MLX5_VDPA_MAX_VQ_ENTRIES 256 -static u16 mlx5_vdpa_get_vq_num_max(struct vdpa_device *vdev) +static u16 mlx5_vdpa_get_vq_num_max(struct vdpa_device *vdev, u16 idx) { return MLX5_VDPA_MAX_VQ_ENTRIES; } diff --git a/drivers/vdpa/vdpa.c b/drivers/vdpa/vdpa.c index bb3f1d1f0422..d77d59811389 100644 --- a/drivers/vdpa/vdpa.c +++ b/drivers/vdpa/vdpa.c @@ -239,6 +239,26 @@ void vdpa_unregister_driver(struct vdpa_driver *drv) } EXPORT_SYMBOL_GPL(vdpa_unregister_driver); +/** + * vdpa_get_vq_num_max - get the maximum virtqueue size + * @vdev: vdpa device + */ +u16 vdpa_get_vq_num_max(struct vdpa_device *vdev) +{ + const struct vdpa_config_ops *ops = vdev->config; + u16 s, size = ops->get_vq_num_max(vdev, 0); + int i; + + for (i = 1; i < vdev->nvqs; i++) { + s = ops->get_vq_num_max(vdev, i); + if (s && s < size) + size = s; + } + + return size; +} +EXPORT_SYMBOL_GPL(vdpa_get_vq_num_max); + /** * vdpa_mgmtdev_register - register a vdpa management device * @@ -502,7 +522,7 @@ vdpa_dev_fill(struct vdpa_device *vdev, struct sk_buff *msg, u32 portid, u32 seq device_id = vdev->config->get_device_id(vdev); vendor_id = vdev->config->get_vendor_id(vdev); - max_vq_size = vdev->config->get_vq_num_max(vdev); + max_vq_size = vdpa_get_vq_num_max(vdev); err = -EMSGSIZE; if (nla_put_string(msg, VDPA_ATTR_DEV_NAME, dev_name(&vdev->dev))) diff --git a/drivers/vdpa/vdpa_sim/vdpa_sim.c b/drivers/vdpa/vdpa_sim/vdpa_sim.c index 98f793bc9376..49e29056f164 100644 --- a/drivers/vdpa/vdpa_sim/vdpa_sim.c +++ b/drivers/vdpa/vdpa_sim/vdpa_sim.c @@ -422,7 +422,7 @@ static void vdpasim_set_config_cb(struct vdpa_device *vdpa, /* We don't support config interrupt */ } -static u16 vdpasim_get_vq_num_max(struct vdpa_device *vdpa) +static u16 vdpasim_get_vq_num_max(struct vdpa_device *vdpa, u16 idx) { return VDPASIM_QUEUE_MAX; } diff --git a/drivers/vdpa/virtio_pci/vp_vdpa.c b/drivers/vdpa/virtio_pci/vp_vdpa.c index c76ebb531212..2926641fb586 100644 --- a/drivers/vdpa/virtio_pci/vp_vdpa.c +++ b/drivers/vdpa/virtio_pci/vp_vdpa.c @@ -195,7 +195,7 @@ static void vp_vdpa_set_status(struct vdpa_device *vdpa, u8 status) vp_vdpa_free_irq(vp_vdpa); } -static u16 vp_vdpa_get_vq_num_max(struct vdpa_device *vdpa) +static u16 vp_vdpa_get_vq_num_max(struct vdpa_device *vdpa, u16 qid) { return VP_VDPA_QUEUE_MAX; } diff --git a/drivers/vhost/vdpa.c b/drivers/vhost/vdpa.c index fb41db3da611..c9ec395b8c42 100644 --- a/drivers/vhost/vdpa.c +++ b/drivers/vhost/vdpa.c @@ -289,11 +289,14 @@ static long vhost_vdpa_set_features(struct vhost_vdpa *v, u64 __user *featurep) static long vhost_vdpa_get_vring_num(struct vhost_vdpa *v, u16 __user *argp) { - struct vdpa_device *vdpa = v->vdpa; - const struct vdpa_config_ops *ops = vdpa->config; u16 num; - num = ops->get_vq_num_max(vdpa); + /* + * VHOST_VDPA_GET_VRING_NUM asssumes a global max virtqueue + * size. So we need to iterate all the virtqueue and return + * the minimal one here. + */ + num = vdpa_get_vq_num_max(v->vdpa); if (copy_to_user(argp, &num, sizeof(num))) return -EFAULT; diff --git a/drivers/virtio/virtio_vdpa.c b/drivers/virtio/virtio_vdpa.c index e28acf482e0c..a43bd5ec5ac6 100644 --- a/drivers/virtio/virtio_vdpa.c +++ b/drivers/virtio/virtio_vdpa.c @@ -158,7 +158,7 @@ virtio_vdpa_setup_vq(struct virtio_device *vdev, unsigned int index, if (!info) return ERR_PTR(-ENOMEM); - num = ops->get_vq_num_max(vdpa); + num = ops->get_vq_num_max(vdpa, index); if (num == 0) { err = -ENOENT; goto error_new_virtqueue; diff --git a/include/linux/vdpa.h b/include/linux/vdpa.h index f311d227aa1b..0f8977a54783 100644 --- a/include/linux/vdpa.h +++ b/include/linux/vdpa.h @@ -137,6 +137,7 @@ struct vdpa_iova_range { * @cb: virtio-vdev interrupt callback structure * @get_vq_num_max: Get the max size of virtqueue * @vdev: vdpa device + * @idx: virtqueue index * Returns u16: max size of virtqueue * @get_device_id: Get virtio device id * @vdev: vdpa device @@ -229,7 +230,7 @@ struct vdpa_config_ops { int (*set_features)(struct vdpa_device *vdev, u64 features); void (*set_config_cb)(struct vdpa_device *vdev, struct vdpa_callback *cb); - u16 (*get_vq_num_max)(struct vdpa_device *vdev); + u16 (*get_vq_num_max)(struct vdpa_device *vdev, u16 idx); u32 (*get_device_id)(struct vdpa_device *vdev); u32 (*get_vendor_id)(struct vdpa_device *vdev); u8 (*get_status)(struct vdpa_device *vdev); @@ -270,6 +271,8 @@ void vdpa_unregister_device(struct vdpa_device *vdev); int _vdpa_register_device(struct vdpa_device *vdev, int nvqs); void _vdpa_unregister_device(struct vdpa_device *vdev); +u16 vdpa_get_vq_num_max(struct vdpa_device *vdev); + /** * struct vdpa_driver - operations for a vDPA driver * @driver: underlying device driver From patchwork Mon Jul 5 07:19:10 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Jason Wang X-Patchwork-Id: 12358369 X-Patchwork-Delegate: kuba@kernel.org Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-17.2 required=3.0 tests=BAYES_00,DKIMWL_WL_HIGH, DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_CR_TRAILER,INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id E3FACC07E9A for ; Mon, 5 Jul 2021 07:19:31 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id CE3E0613FC for ; Mon, 5 Jul 2021 07:19:31 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S230003AbhGEHWH (ORCPT ); Mon, 5 Jul 2021 03:22:07 -0400 Received: from us-smtp-delivery-124.mimecast.com ([170.10.133.124]:55144 "EHLO us-smtp-delivery-124.mimecast.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229986AbhGEHWG (ORCPT ); Mon, 5 Jul 2021 03:22:06 -0400 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1625469570; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=AsZSFMJJkoJGWjbavb6OT4THQJwUV8GZhfv+a3hztGw=; b=BuG0yfRk291+TgiriLkxeTmUyY+11wK33rjA0oA/mooeOKlR0qDaMlLXvXcsSDm3//cuKK hm0ablxOxmjtZYlqv9mDAxO4k6FAWHjiTqMEn+c1IM6C5VUHdfucmEsf42dVHLg20/Mi0i OL6DYagAcuXJ373kP6VLED/tIguj8PI= Received: from mimecast-mx01.redhat.com (mimecast-mx01.redhat.com [209.132.183.4]) (Using TLS) by relay.mimecast.com with ESMTP id us-mta-92-88fkwEcTN6ymt989fu6u9Q-1; Mon, 05 Jul 2021 03:19:28 -0400 X-MC-Unique: 88fkwEcTN6ymt989fu6u9Q-1 Received: from smtp.corp.redhat.com (int-mx04.intmail.prod.int.phx2.redhat.com [10.5.11.14]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mimecast-mx01.redhat.com (Postfix) with ESMTPS id B913E79EDC; Mon, 5 Jul 2021 07:19:27 +0000 (UTC) Received: from localhost.localdomain (ovpn-13-21.pek2.redhat.com [10.72.13.21]) by smtp.corp.redhat.com (Postfix) with ESMTP id E95D85D9FC; Mon, 5 Jul 2021 07:19:23 +0000 (UTC) From: Jason Wang To: mst@redhat.com, jasowang@redhat.com, virtualization@lists.linux-foundation.org, linux-kernel@vger.kernel.org, kvm@vger.kernel.org Cc: netdev@vger.kernel.org, xieyongji@bytedance.com, stefanha@redhat.com Subject: [PATCH 2/2] vdpa: vp_vdpa: don't use hard-coded maximum virtqueue size Date: Mon, 5 Jul 2021 15:19:10 +0800 Message-Id: <20210705071910.31965-2-jasowang@redhat.com> In-Reply-To: <20210705071910.31965-1-jasowang@redhat.com> References: <20210705071910.31965-1-jasowang@redhat.com> MIME-Version: 1.0 X-Scanned-By: MIMEDefang 2.79 on 10.5.11.14 Precedence: bulk List-ID: X-Mailing-List: netdev@vger.kernel.org This patch switch to read virtqueue size from the capability instead of depending on the hardcoded value. This allows the per virtqueue size could be advertised. Signed-off-by: Jason Wang --- drivers/vdpa/virtio_pci/vp_vdpa.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/drivers/vdpa/virtio_pci/vp_vdpa.c b/drivers/vdpa/virtio_pci/vp_vdpa.c index 2926641fb586..198f7076e4d9 100644 --- a/drivers/vdpa/virtio_pci/vp_vdpa.c +++ b/drivers/vdpa/virtio_pci/vp_vdpa.c @@ -18,7 +18,6 @@ #include #include -#define VP_VDPA_QUEUE_MAX 256 #define VP_VDPA_DRIVER_NAME "vp_vdpa" #define VP_VDPA_NAME_SIZE 256 @@ -197,7 +196,10 @@ static void vp_vdpa_set_status(struct vdpa_device *vdpa, u8 status) static u16 vp_vdpa_get_vq_num_max(struct vdpa_device *vdpa, u16 qid) { - return VP_VDPA_QUEUE_MAX; + struct vp_vdpa *vp_vdpa = vdpa_to_vp(vdpa); + struct virtio_pci_modern_device *mdev = &vp_vdpa->mdev; + + return vp_modern_get_queue_size(mdev, qid); } static int vp_vdpa_get_vq_state(struct vdpa_device *vdpa, u16 qid,