DPDK patches and discussions
 help / color / mirror / Atom feed
From: Maxime Coquelin <maxime.coquelin@redhat.com>
To: Fan Zhang <roy.fan.zhang@intel.com>, dev@dpdk.org
Cc: chenbo.xia@intel.com, changpeng.liu@intel.com,
	ferruh.yigit@intel.com, stable@dpdk.org
Subject: Re: [dpdk-dev] [dpdk-dev v2 1/2] vhost: add backend type in driver start
Date: Tue, 6 Oct 2020 09:53:57 +0200	[thread overview]
Message-ID: <7c1121c2-8474-0878-b56b-152fd4854392@redhat.com> (raw)
In-Reply-To: <20201002153601.84097-2-roy.fan.zhang@intel.com>

Hi Fan,

On 10/2/20 5:36 PM, Fan Zhang wrote:
> This patch adds an internal driver start function with a newly
> added backend type identifier as parameter. With this way
> different built-in driver types (net, crypto) can be identified.
> 
> Signed-off-by: Fan Zhang <roy.fan.zhang@intel.com>
> ---
>  lib/librte_vhost/socket.c     | 39 ++++++++++++++++++++++++++++-------
>  lib/librte_vhost/vhost_user.h |  8 +++++++
>  2 files changed, 39 insertions(+), 8 deletions(-)
> 
> diff --git a/lib/librte_vhost/socket.c b/lib/librte_vhost/socket.c
> index 0169d3648..a53e536ac 100644
> --- a/lib/librte_vhost/socket.c
> +++ b/lib/librte_vhost/socket.c
> @@ -38,7 +38,7 @@ struct vhost_user_socket {
>  	bool is_server;
>  	bool reconnect;
>  	bool iommu_support;
> -	bool use_builtin_virtio_net;
> +	enum virtio_backend_type backend_type;
>  	bool extbuf;
>  	bool linearbuf;
>  	bool async_copy;
> @@ -224,7 +224,9 @@ vhost_user_add_connection(int fd, struct vhost_user_socket *vsocket)
>  	size = strnlen(vsocket->path, PATH_MAX);
>  	vhost_set_ifname(vid, vsocket->path, size);
>  
> -	vhost_set_builtin_virtio_net(vid, vsocket->use_builtin_virtio_net);
> +	vhost_set_builtin_virtio_net(vid,
> +			vsocket->backend_type == VIRTIO_DEV_BUILTIN_NET ?
> +					true : false);
>  
>  	vhost_attach_vdpa_device(vid, vsocket->vdpa_dev);
>  
> @@ -632,7 +634,7 @@ rte_vhost_driver_disable_features(const char *path, uint64_t features)
>  	pthread_mutex_lock(&vhost_user.mutex);
>  	vsocket = find_vhost_user_socket(path);
>  
> -	/* Note that use_builtin_virtio_net is not affected by this function
> +	/* Note that backend type is not affected by this function
>  	 * since callers may want to selectively disable features of the
>  	 * built-in vhost net device backend.
>  	 */
> @@ -681,7 +683,8 @@ rte_vhost_driver_set_features(const char *path, uint64_t features)
>  		/* Anyone setting feature bits is implementing their own vhost
>  		 * device backend.
>  		 */
> -		vsocket->use_builtin_virtio_net = false;
> +		if (vsocket->backend_type == VIRTIO_DEV_BUILTIN_NET)
> +			vsocket->backend_type = VIRTIO_DEV_UNKNOWN;
>  	}
>  	pthread_mutex_unlock(&vhost_user.mutex);
>  
> @@ -899,7 +902,7 @@ rte_vhost_driver_register(const char *path, uint64_t flags)
>  	 * rte_vhost_driver_set_features(), which will overwrite following
>  	 * two values.
>  	 */
> -	vsocket->use_builtin_virtio_net = true;
> +	vsocket->backend_type = VIRTIO_DEV_BUILTIN_NET;
>  	vsocket->supported_features = VIRTIO_NET_SUPPORTED_FEATURES;
>  	vsocket->features           = VIRTIO_NET_SUPPORTED_FEATURES;
>  	vsocket->protocol_features  = VHOST_USER_PROTOCOL_FEATURES;
> @@ -1117,10 +1120,17 @@ vhost_driver_callback_get(const char *path)
>  }
>  
>  int
> -rte_vhost_driver_start(const char *path)
> +vhost_driver_start(const char *path, enum virtio_backend_type backend_type)
>  {
>  	struct vhost_user_socket *vsocket;
>  	static pthread_t fdset_tid;
> +	int ret;
> +
> +	if (backend_type <= VIRTIO_DEV_UNKNOWN ||
> +			backend_type > VIRTIO_DEV_BUILTIN_CRYPTO) {
> +		VHOST_LOG_CONFIG(ERR, "Wrong backend type\n");
> +		return -1;
> +	}
>  
>  	pthread_mutex_lock(&vhost_user.mutex);
>  	vsocket = find_vhost_user_socket(path);
> @@ -1153,7 +1163,20 @@ rte_vhost_driver_start(const char *path)
>  	}
>  
>  	if (vsocket->is_server)
> -		return vhost_user_start_server(vsocket);
> +		ret = vhost_user_start_server(vsocket);
>  	else
> -		return vhost_user_start_client(vsocket);
> +		ret = vhost_user_start_client(vsocket);
> +
> +	if (ret < 0)
> +		return ret;
> +
> +	vsocket->backend_type = backend_type;
> +
> +	return 0;

Looking at it again, I think patch 1 is not necessary.
Patch 2 should be enough and would be easier to backport.

What do you think?

Regards,
Maxime

> +}
> +
> +int
> +rte_vhost_driver_start(const char *path)
> +{
> +	return vhost_driver_start(path, VIRTIO_DEV_BUILTIN_NET);
>  }
> diff --git a/lib/librte_vhost/vhost_user.h b/lib/librte_vhost/vhost_user.h
> index 16fe03f88..9f526e676 100644
> --- a/lib/librte_vhost/vhost_user.h
> +++ b/lib/librte_vhost/vhost_user.h
> @@ -158,6 +158,12 @@ typedef struct VhostUserMsg {
>  /* The version of the protocol we support */
>  #define VHOST_USER_VERSION    0x1
>  
> +/* virtio backend types */
> +enum virtio_backend_type {
> +	VIRTIO_DEV_UNKNOWN = 0, /* Likely external */
> +	VIRTIO_DEV_BUILTIN_NET, /* Virtio-net device */
> +	VIRTIO_DEV_BUILTIN_CRYPTO, /* Virtio-crypto device */
> +};
>  
>  /* vhost_user.c */
>  int vhost_user_msg_handler(int vid, int fd);
> @@ -167,5 +173,7 @@ int vhost_user_iotlb_miss(struct virtio_net *dev, uint64_t iova, uint8_t perm);
>  int read_fd_message(int sockfd, char *buf, int buflen, int *fds, int max_fds,
>  		int *fd_num);
>  int send_fd_message(int sockfd, char *buf, int buflen, int *fds, int fd_num);
> +int vhost_driver_start(const char *path,
> +		enum virtio_backend_type backend_type);
>  
>  #endif
> 


  reply	other threads:[~2020-10-06  7:54 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-10-02  8:36 [dpdk-dev] vhost/crypto: fix initialization Fan Zhang
2020-10-02 12:17 ` Maxime Coquelin
2020-10-02 12:38   ` Zhang, Roy Fan
2020-10-02 15:35 ` [dpdk-dev] [dpdk-dev v2 0/2] " Fan Zhang
2020-10-02 15:36   ` [dpdk-dev] [dpdk-dev v2 1/2] vhost: add backend type in driver start Fan Zhang
2020-10-06  7:53     ` Maxime Coquelin [this message]
2020-10-06  8:56       ` Zhang, Roy Fan
2020-10-02 15:36   ` [dpdk-dev] [dpdk-dev v2 2/2] vhost/crypto: fix feature negotiation Fan Zhang
2020-10-06  8:09     ` Maxime Coquelin
2020-10-06  8:37       ` Zhang, Roy Fan
2020-10-09  7:36         ` Maxime Coquelin
2020-10-10  9:28           ` Jiang, YuX
2020-10-09  6:39     ` Maxime Coquelin
2020-10-09  7:24     ` Maxime Coquelin

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=7c1121c2-8474-0878-b56b-152fd4854392@redhat.com \
    --to=maxime.coquelin@redhat.com \
    --cc=changpeng.liu@intel.com \
    --cc=chenbo.xia@intel.com \
    --cc=dev@dpdk.org \
    --cc=ferruh.yigit@intel.com \
    --cc=roy.fan.zhang@intel.com \
    --cc=stable@dpdk.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).