From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <dev-bounces@dpdk.org>
Received: from dpdk.org (dpdk.org [92.243.14.124])
	by inbox.dpdk.org (Postfix) with ESMTP id 5C40FA2E1B
	for <public@inbox.dpdk.org>; Wed,  4 Sep 2019 08:56:42 +0200 (CEST)
Received: from [92.243.14.124] (localhost [127.0.0.1])
	by dpdk.org (Postfix) with ESMTP id 5CF341E8BF;
	Wed,  4 Sep 2019 08:56:41 +0200 (CEST)
Received: from mx1.redhat.com (mx1.redhat.com [209.132.183.28])
 by dpdk.org (Postfix) with ESMTP id C5F0E1E8B1;
 Wed,  4 Sep 2019 08:56:39 +0200 (CEST)
Received: from smtp.corp.redhat.com (int-mx08.intmail.prod.int.phx2.redhat.com
 [10.5.11.23])
 (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits))
 (No client certificate requested)
 by mx1.redhat.com (Postfix) with ESMTPS id 1BAF4307D88C;
 Wed,  4 Sep 2019 06:56:39 +0000 (UTC)
Received: from [10.36.112.25] (ovpn-112-25.ams2.redhat.com [10.36.112.25])
 by smtp.corp.redhat.com (Postfix) with ESMTPS id B4B5A19C4F;
 Wed,  4 Sep 2019 06:56:34 +0000 (UTC)
To: Jason Wang <jasowang@redhat.com>, Tiwei Bie <tiwei.bie@intel.com>
Cc: zhihong.wang@intel.com, amorenoz@redhat.com, xiao.w.wang@intel.com,
 dev@dpdk.org, jfreimann@redhat.com, stable@dpdk.org
References: <20190829080000.20806-1-maxime.coquelin@redhat.com>
 <20190829080000.20806-11-maxime.coquelin@redhat.com>
 <20190903053006.GA32701@___>
 <b7b58dd1-7ac5-9ba3-32cd-d749d695617b@redhat.com> <20190903084909.GA9622@___>
 <718b4bc8-1a04-689f-5a3a-b6d80c517720@redhat.com>
From: Maxime Coquelin <maxime.coquelin@redhat.com>
Openpgp: preference=signencrypt
Autocrypt: addr=maxime.coquelin@redhat.com; keydata=
 mQINBFOEQQIBEADjNLYZZqghYuWv1nlLisptPJp+TSxE/KuP7x47e1Gr5/oMDJ1OKNG8rlNg
 kLgBQUki3voWhUbMb69ybqdMUHOl21DGCj0BTU3lXwapYXOAnsh8q6RRM+deUpasyT+Jvf3a
 gU35dgZcomRh5HPmKMU4KfeA38cVUebsFec1HuJAWzOb/UdtQkYyZR4rbzw8SbsOemtMtwOx
 YdXodneQD7KuRU9IhJKiEfipwqk2pufm2VSGl570l5ANyWMA/XADNhcEXhpkZ1Iwj3TWO7XR
 uH4xfvPl8nBsLo/EbEI7fbuUULcAnHfowQslPUm6/yaGv6cT5160SPXT1t8U9QDO6aTSo59N
 jH519JS8oeKZB1n1eLDslCfBpIpWkW8ZElGkOGWAN0vmpLfdyiqBNNyS3eGAfMkJ6b1A24un
 /TKc6j2QxM0QK4yZGfAxDxtvDv9LFXec8ENJYsbiR6WHRHq7wXl/n8guyh5AuBNQ3LIK44x0
 KjGXP1FJkUhUuruGyZsMrDLBRHYi+hhDAgRjqHgoXi5XGETA1PAiNBNnQwMf5aubt+mE2Q5r
 qLNTgwSo2dpTU3+mJ3y3KlsIfoaxYI7XNsPRXGnZi4hbxmeb2NSXgdCXhX3nELUNYm4ArKBP
 LugOIT/zRwk0H0+RVwL2zHdMO1Tht1UOFGfOZpvuBF60jhMzbQARAQABtCxNYXhpbWUgQ29x
 dWVsaW4gPG1heGltZS5jb3F1ZWxpbkByZWRoYXQuY29tPokCOAQTAQIAIgUCV3u/5QIbAwYL
 CQgHAwIGFQgCCQoLBBYCAwECHgECF4AACgkQyjiNKEaHD4ma2g/+P+Hg9WkONPaY1J4AR7Uf
 kBneosS4NO3CRy0x4WYmUSLYMLx1I3VH6SVjqZ6uBoYy6Fs6TbF6SHNc7QbB6Qjo3neqnQR1
 71Ua1MFvIob8vUEl3jAR/+oaE1UJKrxjWztpppQTukIk4oJOmXbL0nj3d8dA2QgHdTyttZ1H
 xzZJWWz6vqxCrUqHU7RSH9iWg9R2iuTzii4/vk1oi4Qz7y/q8ONOq6ffOy/t5xSZOMtZCspu
 Mll2Szzpc/trFO0pLH4LZZfz/nXh2uuUbk8qRIJBIjZH3ZQfACffgfNefLe2PxMqJZ8mFJXc
 RQO0ONZvwoOoHL6CcnFZp2i0P5ddduzwPdGsPq1bnIXnZqJSl3dUfh3xG5ArkliZ/++zGF1O
 wvpGvpIuOgLqjyCNNRoR7cP7y8F24gWE/HqJBXs1qzdj/5Hr68NVPV1Tu/l2D1KMOcL5sOrz
 2jLXauqDWn1Okk9hkXAP7+0Cmi6QwAPuBT3i6t2e8UdtMtCE4sLesWS/XohnSFFscZR6Vaf3
 gKdWiJ/fW64L6b9gjkWtHd4jAJBAIAx1JM6xcA1xMbAFsD8gA2oDBWogHGYcScY/4riDNKXi
 lw92d6IEHnSf6y7KJCKq8F+Jrj2BwRJiFKTJ6ChbOpyyR6nGTckzsLgday2KxBIyuh4w+hMq
 TGDSp2rmWGJjASq5Ag0EVPSbkwEQAMkaNc084Qvql+XW+wcUIY+Dn9A2D1gMr2BVwdSfVDN7
 0ZYxo9PvSkzh6eQmnZNQtl8WSHl3VG3IEDQzsMQ2ftZn2sxjcCadexrQQv3Lu60Tgj7YVYRM
 H+fLYt9W5YuWduJ+FPLbjIKynBf6JCRMWr75QAOhhhaI0tsie3eDsKQBA0w7WCuPiZiheJaL
 4MDe9hcH4rM3ybnRW7K2dLszWNhHVoYSFlZGYh+MGpuODeQKDS035+4H2rEWgg+iaOwqD7bg
 CQXwTZ1kSrm8NxIRVD3MBtzp9SZdUHLfmBl/tLVwDSZvHZhhvJHC6Lj6VL4jPXF5K2+Nn/Su
 CQmEBisOmwnXZhhu8ulAZ7S2tcl94DCo60ReheDoPBU8PR2TLg8rS5f9w6mLYarvQWL7cDtT
 d2eX3Z6TggfNINr/RTFrrAd7NHl5h3OnlXj7PQ1f0kfufduOeCQddJN4gsQfxo/qvWVB7PaE
 1WTIggPmWS+Xxijk7xG6x9McTdmGhYaPZBpAxewK8ypl5+yubVsE9yOOhKMVo9DoVCjh5To5
 aph7CQWfQsV7cd9PfSJjI2lXI0dhEXhQ7lRCFpf3V3mD6CyrhpcJpV6XVGjxJvGUale7+IOp
 sQIbPKUHpB2F+ZUPWds9yyVxGwDxD8WLqKKy0WLIjkkSsOb9UBNzgRyzrEC9lgQ/ABEBAAGJ
 Ah8EGAECAAkFAlT0m5MCGwwACgkQyjiNKEaHD4nU8hAAtt0xFJAy0sOWqSmyxTc7FUcX+pbD
 KVyPlpl6urKKMk1XtVMUPuae/+UwvIt0urk1mXi6DnrAN50TmQqvdjcPTQ6uoZ8zjgGeASZg
 jj0/bJGhgUr9U7oG7Hh2F8vzpOqZrdd65MRkxmc7bWj1k81tOU2woR/Gy8xLzi0k0KUa8ueB
 iYOcZcIGTcs9CssVwQjYaXRoeT65LJnTxYZif2pfNxfINFzCGw42s3EtZFteczClKcVSJ1+L
 +QUY/J24x0/ocQX/M1PwtZbB4c/2Pg/t5FS+s6UB1Ce08xsJDcwyOPIH6O3tccZuriHgvqKP
 yKz/Ble76+NFlTK1mpUlfM7PVhD5XzrDUEHWRTeTJSvJ8TIPL4uyfzhjHhlkCU0mw7Pscyxn
 DE8G0UYMEaNgaZap8dcGMYH/96EfE5s/nTX0M6MXV0yots7U2BDb4soLCxLOJz4tAFDtNFtA
 wLBhXRSvWhdBJZiig/9CG3dXmKfi2H+wdUCSvEFHRpgo7GK8/Kh3vGhgKmnnxhl8ACBaGy9n
 fxjSxjSO6rj4/MeenmlJw1yebzkX8ZmaSi8BHe+n6jTGEFNrbiOdWpJgc5yHIZZnwXaW54QT
 UhhSjDL1rV2B4F28w30jYmlRmm2RdN7iCZfbyP3dvFQTzQ4ySquuPkIGcOOHrvZzxbRjzMx1
 Mwqu3GQ=
Message-ID: <4d9888a3-ec38-ceb2-93f8-e85c68017422@redhat.com>
Date: Wed, 4 Sep 2019 08:56:32 +0200
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101
 Thunderbird/60.8.0
MIME-Version: 1.0
In-Reply-To: <718b4bc8-1a04-689f-5a3a-b6d80c517720@redhat.com>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 8bit
X-Scanned-By: MIMEDefang 2.84 on 10.5.11.23
X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16
 (mx1.redhat.com [10.5.110.48]); Wed, 04 Sep 2019 06:56:39 +0000 (UTC)
Subject: Re: [dpdk-dev] [PATCH 10/15] net/virtio: add vDPA op to configure
 and start the device
X-BeenThere: dev@dpdk.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: DPDK patches and discussions <dev.dpdk.org>
List-Unsubscribe: <https://mails.dpdk.org/options/dev>,
 <mailto:dev-request@dpdk.org?subject=unsubscribe>
List-Archive: <http://mails.dpdk.org/archives/dev/>
List-Post: <mailto:dev@dpdk.org>
List-Help: <mailto:dev-request@dpdk.org?subject=help>
List-Subscribe: <https://mails.dpdk.org/listinfo/dev>,
 <mailto:dev-request@dpdk.org?subject=subscribe>
Errors-To: dev-bounces@dpdk.org
Sender: "dev" <dev-bounces@dpdk.org>



On 9/4/19 6:06 AM, Jason Wang wrote:
> 
> On 2019/9/3 下午4:49, Tiwei Bie wrote:
>> On Tue, Sep 03, 2019 at 09:40:25AM +0200, Maxime Coquelin wrote:
>>> On 9/3/19 7:30 AM, Tiwei Bie wrote:
>>>> On Thu, Aug 29, 2019 at 09:59:55AM +0200, Maxime Coquelin wrote:
>>>>> In order to support multi-queue, we need to implement the control
>>>>> path. The problem is that both the Vhost-user master and slave use
>>>>> VAs in their processes address spaces as IOVAs, which creates
>>>>> collusions between the data rings IOVAs managed the master, and
>>>>> the Control ring IOVAs. The trick here is to remmap the Control
>>>>> ring memory to another range, after the slave is aware of master's
>>>>> ranges.
>>>>>
>>>>> Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
>>>>> ---
>>>>>   drivers/net/virtio/virtio_vdpa.c | 255
>>>>> +++++++++++++++++++++++++++++++
>>>>>   1 file changed, 255 insertions(+)
>>>>>
>>>>> diff --git a/drivers/net/virtio/virtio_vdpa.c
>>>>> b/drivers/net/virtio/virtio_vdpa.c
>>>>> index fc52a8e92..13b4dd07d 100644
>>>>> --- a/drivers/net/virtio/virtio_vdpa.c
>>>>> +++ b/drivers/net/virtio/virtio_vdpa.c
>>>>> @@ -106,6 +106,127 @@ find_internal_resource_by_dev(struct
>>>>> rte_pci_device *pdev)
>>>>>       return list;
>>>>>   }
>>>>>   +static int
>>>>> +virtio_vdpa_dma_map_ctrl_queue(struct virtio_vdpa_device *dev, int
>>>>> do_map,
>>>>> +        uint64_t iova)
>>>>> +{
>>>>> +    const struct rte_memzone *mz;
>>>>> +    int ret;
>>>>> +
>>>>> +    /*
>>>>> +     * IOVAs are processes VAs. We cannot use them as the Data and
>>>>> Control
>>>>> +     * paths are run in different processes, which may (does) lead to
>>>>> +     * collusions. The trick here is to fixup Ctrl path IOVAs so
>>>>> that they
>>>>> +     * start after the Data path ranges.
>>>>> +     */
>>>>> +    if (do_map) {
>>>>> +        mz = dev->cvq->cq.mz;
>>>>> +        ret = rte_vfio_container_dma_map(dev->vfio_container_fd,
>>>>> +                (uint64_t)(uintptr_t)mz->addr,
>>>>> +                iova, mz->len);
>>>>> +        if (ret < 0) {
>>>>> +            DRV_LOG(ERR, "Failed to map ctrl ring (%d)", ret);
>>>>> +            return ret;
>>>>> +        }
>>>>> +
>>>>> +        dev->cvq->vq_ring_mem = iova;
>>>>> +        iova += mz->len;
>>>>> +
>>>>> +        mz = dev->cvq->cq.virtio_net_hdr_mz;
>>>>> +        ret = rte_vfio_container_dma_map(dev->vfio_container_fd,
>>>>> +                (uint64_t)(uintptr_t)mz->addr,
>>>>> +                iova, mz->len);
>>>>> +        if (ret < 0) {
>>>>> +            DRV_LOG(ERR, "Failed to map ctrl headers (%d)", ret);
>>>>> +            return ret;
>>>>> +        }
>>>> This will allow guest to access the cq.mz and cq.virtio_net_hdr_mz
>>>> via the device which may have potential risks.
>>> I get what you mean, but I'm not sure to see how we could avoid that.
>>> AFAIU, we need to map the control queue in the device IOMMU, otherwise
>>> how could the host (in case of virtual device) or the NIC (in case of
>>> Virtio offload), could access the ring?
>>> Any thoughts?
>> I also don't see a way to avoid that. That's why I said in below
>> thread that I think the control queue based interface seems not a
>> quite good interface for a backend device:
>>
>> https://lkml.org/lkml/2019/9/2/934
>>
>> In IFCVF NIC, we added a MMIO based interface to replace control
>> queue for the multiqueue setup in vDPA mode.
>>
>> Jason is proposing some changes to make virtio device suitable
>> for backend device. I'm not sure whether it's possible to cover
>> this case as well..
> 
> 
> A silly question, can we do dynamic mapping like what kernel driver did
> here?

Not silly at all, it is of course possible.
I will implement that in my v2.

Thanks!
Maxime

> Thanks
> 
> 
>>
>> Regards,
>> Tiwei
>>
>>> Thanks,
>>> Maxime
>>>> Regards,
>>>> Tiwei
>>>>
>>>>> +
>>>>> +        dev->cvq->cq.virtio_net_hdr_mem = iova;
>>>>> +    } else {
>>>>> +        mz = dev->cvq->cq.mz;
>>>>> +        ret = rte_vfio_container_dma_unmap(dev->vfio_container_fd,
>>>>> +                (uint64_t)(uintptr_t)mz->addr,
>>>>> +                iova, mz->len);
>>>>> +        if (ret < 0) {
>>>>> +            DRV_LOG(ERR, "Failed to unmap ctrl ring (%d)", ret);
>>>>> +            return ret;
>>>>> +        }
>>>>> +
>>>>> +        dev->cvq->vq_ring_mem = 0;
>>>>> +        iova += mz->len;
>>>>> +
>>>>> +        mz = dev->cvq->cq.virtio_net_hdr_mz;
>>>>> +        ret = rte_vfio_container_dma_unmap(dev->vfio_container_fd,
>>>>> +                (uint64_t)(uintptr_t)mz->addr,
>>>>> +                iova, mz->len);
>>>>> +        if (ret < 0) {
>>>>> +            DRV_LOG(ERR, "Failed to unmap ctrl headers (%d)", ret);
>>>>> +            return ret;
>>>>> +        }
>>>>> +
>>>>> +        dev->cvq->cq.virtio_net_hdr_mem = 0;
>>>>> +    }
>>>>> +
>>>>> +    return 0;
>>>>> +}
>>>>> +
>>>>> +static int
>>>>> +virtio_vdpa_dma_map(struct virtio_vdpa_device *dev, int do_map)
>>>>> +{
>>>>> +    uint32_t i;
>>>>> +    int ret;
>>>>> +    struct rte_vhost_memory *mem = NULL;
>>>>> +    int vfio_container_fd;
>>>>> +    uint64_t avail_iova = 0;
>>>>> +
>>>>> +    ret = rte_vhost_get_mem_table(dev->vid, &mem);
>>>>> +    if (ret < 0 || !mem) {
>>>>> +        DRV_LOG(ERR, "failed to get VM memory layout.");
>>>>> +        return ret;
>>>>> +    }
>>>>> +
>>>>> +    vfio_container_fd = dev->vfio_container_fd;
>>>>> +
>>>>> +    for (i = 0; i < mem->nregions; i++) {
>>>>> +        struct rte_vhost_mem_region *reg;
>>>>> +
>>>>> +        reg = &mem->regions[i];
>>>>> +        DRV_LOG(INFO, "%s, region %u: HVA 0x%" PRIx64 ", "
>>>>> +            "GPA 0x%" PRIx64 ", size 0x%" PRIx64 ".",
>>>>> +            do_map ? "DMA map" : "DMA unmap", i,
>>>>> +            reg->host_user_addr, reg->guest_phys_addr, reg->size);
>>>>> +
>>>>> +        if (reg->guest_phys_addr + reg->size > avail_iova)
>>>>> +            avail_iova = reg->guest_phys_addr + reg->size;
>>>>> +
>>>>> +        if (do_map) {
>>>>> +            ret = rte_vfio_container_dma_map(vfio_container_fd,
>>>>> +                reg->host_user_addr, reg->guest_phys_addr,
>>>>> +                reg->size);
>>>>> +            if (ret < 0) {
>>>>> +                DRV_LOG(ERR, "DMA map failed.");
>>>>> +                goto exit;
>>>>> +            }
>>>>> +        } else {
>>>>> +            ret = rte_vfio_container_dma_unmap(vfio_container_fd,
>>>>> +                reg->host_user_addr, reg->guest_phys_addr,
>>>>> +                reg->size);
>>>>> +            if (ret < 0) {
>>>>> +                DRV_LOG(ERR, "DMA unmap failed.");
>>>>> +                goto exit;
>>>>> +            }
>>>>> +        }
>>>>> +    }
>>>>> +
>>>>> +    if (dev->cvq)
>>>>> +        ret = virtio_vdpa_dma_map_ctrl_queue(dev, do_map,
>>>>> avail_iova);
>>>>> +
>>>>> +exit:
>>>>> +    free(mem);
>>>>> +
>>>>> +    return ret;
>>>>> +}
>>>>> +
>>>>>   static int
>>>>>   virtio_vdpa_vfio_setup(struct virtio_vdpa_device *dev)
>>>>>   {
>>>>> @@ -216,10 +337,144 @@ virtio_vdpa_get_protocol_features(int did
>>>>> __rte_unused, uint64_t *features)
>>>>>       return 0;
>>>>>   }
>>>>>   +static uint64_t
>>>>> +hva_to_gpa(int vid, uint64_t hva)
>>>>> +{
>>>>> +    struct rte_vhost_memory *mem = NULL;
>>>>> +    struct rte_vhost_mem_region *reg;
>>>>> +    uint32_t i;
>>>>> +    uint64_t gpa = 0;
>>>>> +
>>>>> +    if (rte_vhost_get_mem_table(vid, &mem) < 0)
>>>>> +        goto exit;
>>>>> +
>>>>> +    for (i = 0; i < mem->nregions; i++) {
>>>>> +        reg = &mem->regions[i];
>>>>> +
>>>>> +        if (hva >= reg->host_user_addr &&
>>>>> +                hva < reg->host_user_addr + reg->size) {
>>>>> +            gpa = hva - reg->host_user_addr + reg->guest_phys_addr;
>>>>> +            break;
>>>>> +        }
>>>>> +    }
>>>>> +
>>>>> +exit:
>>>>> +    if (mem)
>>>>> +        free(mem);
>>>>> +    return gpa;
>>>>> +}
>>>>> +
>>>>> +static int
>>>>> +virtio_vdpa_start(struct virtio_vdpa_device *dev)
>>>>> +{
>>>>> +    struct virtio_hw *hw = &dev->hw;
>>>>> +    int i, vid, nr_vring, ret;
>>>>> +    struct rte_vhost_vring vr;
>>>>> +    struct virtio_pmd_ctrl ctrl;
>>>>> +    int dlen[1];
>>>>> +
>>>>> +    vid = dev->vid;
>>>>> +    nr_vring = rte_vhost_get_vring_num(vid);
>>>>> +
>>>>> +    if (dev->vqs)
>>>>> +        rte_free(dev->vqs);
>>>>> +
>>>>> +    dev->vqs = rte_zmalloc("virtio_vdpa", sizeof(*dev->vqs) *
>>>>> nr_vring, 0);
>>>>> +
>>>>> +    for (i = 0; i < nr_vring; i++) {
>>>>> +        struct virtqueue *vq = &dev->vqs[i];
>>>>> +
>>>>> +        rte_vhost_get_vhost_vring(vid, i, &vr);
>>>>> +
>>>>> +        vq->vq_queue_index = i;
>>>>> +        vq->vq_nentries = vr.size;
>>>>> +        vq->vq_ring_mem = hva_to_gpa(vid,
>>>>> (uint64_t)(uintptr_t)vr.desc);
>>>>> +        if (vq->vq_ring_mem  == 0) {
>>>>> +            DRV_LOG(ERR, "Fail to get GPA for descriptor ring.");
>>>>> +            ret = -1;
>>>>> +            goto out_free_vqs;
>>>>> +        }
>>>>> +
>>>>> +        ret = VTPCI_OPS(hw)->setup_queue(hw, vq);
>>>>> +        if (ret) {
>>>>> +            DRV_LOG(ERR, "Fail to setup queue.");
>>>>> +            goto out_free_vqs;
>>>>> +        }
>>>>> +    }
>>>>> +
>>>>> +    if (dev->cvq) {
>>>>> +        ret = VTPCI_OPS(hw)->setup_queue(hw, dev->cvq);
>>>>> +        if (ret) {
>>>>> +            DRV_LOG(ERR, "Fail to setup ctrl queue.");
>>>>> +            goto out_free_vqs;
>>>>> +        }
>>>>> +    }
>>>>> +
>>>>> +    vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_DRIVER_OK);
>>>>> +
>>>>> +    if (!dev->cvq)
>>>>> +        return 0;
>>>>> +
>>>>> +    ctrl.hdr.class = VIRTIO_NET_CTRL_MQ;
>>>>> +    ctrl.hdr.cmd = VIRTIO_NET_CTRL_MQ_VQ_PAIRS_SET;
>>>>> +    memcpy(ctrl.data, &dev->max_queue_pairs, sizeof(uint16_t));
>>>>> +
>>>>> +    dlen[0] = sizeof(uint16_t);
>>>>> +
>>>>> +    ret = virtio_send_command(hw->cvq, &ctrl, dlen, 1);
>>>>> +    if (ret) {
>>>>> +        DRV_LOG(ERR, "Multiqueue configured but send command "
>>>>> +              "failed, this is too late now...");
>>>>> +        ret = -EINVAL;
>>>>> +        goto out_free_vqs;
>>>>> +    }
>>>>> +
>>>>> +    return 0;
>>>>> +out_free_vqs:
>>>>> +    rte_free(dev->vqs);
>>>>> +
>>>>> +    return ret;
>>>>> +}
>>>>> +
>>>>> +static int
>>>>> +virtio_vdpa_dev_config(int vid)
>>>>> +{
>>>>> +    int did, ret;
>>>>> +    struct internal_list *list;
>>>>> +    struct virtio_vdpa_device *dev;
>>>>> +
>>>>> +    did = rte_vhost_get_vdpa_device_id(vid);
>>>>> +    list = find_internal_resource_by_did(did);
>>>>> +    if (list == NULL) {
>>>>> +        DRV_LOG(ERR, "Invalid device id: %d", did);
>>>>> +        return -1;
>>>>> +    }
>>>>> +
>>>>> +    dev = list->dev;
>>>>> +    dev->vid = vid;
>>>>> +
>>>>> +    rte_spinlock_lock(&dev->lock);
>>>>> +
>>>>> +    ret = virtio_vdpa_dma_map(dev, 1);
>>>>> +    if (ret)
>>>>> +        goto out_unlock;
>>>>> +
>>>>> +    ret = virtio_vdpa_start(dev);
>>>>> +
>>>>> +    if (rte_vhost_host_notifier_ctrl(vid, true) != 0)
>>>>> +        DRV_LOG(NOTICE, "vDPA (%d): software relay is used.", did);
>>>>> +
>>>>> +out_unlock:
>>>>> +    rte_spinlock_unlock(&dev->lock);
>>>>> +
>>>>> +    return ret;
>>>>> +}
>>>>> +
>>>>>   static struct rte_vdpa_dev_ops virtio_vdpa_ops = {
>>>>>       .get_queue_num = virtio_vdpa_get_queue_num,
>>>>>       .get_features = virtio_vdpa_get_features,
>>>>>       .get_protocol_features = virtio_vdpa_get_protocol_features,
>>>>> +    .dev_conf = virtio_vdpa_dev_config,
>>>>>   };
>>>>>     static inline int
>>>>> -- 
>>>>> 2.21.0
>>>>>