[dpdk-dev] [PATCH v2 1/2] vhost: support inflight share memory protocol feature

[Tiwei Bie]

On Wed, Jul 10, 2019 at 06:43:55PM +0800, JinYu wrote:
> This patch introduces two new messages VHOST_USER_GET_INFLIGHT_FD
> and VHOST_USER_SET_INFLIGHT_FD to support transferring a shared
> buffer between qemu and backend.
> 
> Firstly, qemu uses VHOST_USER_GET_INFLIGHT_FD to get the
> shared buffer from backend. Then qemu should send it back
> through VHOST_USER_SET_INFLIGHT_FD each time we start vhost-user.
> 
> This shared buffer is used to process inflight I/O when backend
> reconnect.

I saw you send several patch sets for v2. It's a bit
confusing. Please bump the version when you send a new
series. Besides, please keep the cover letter.

> 
> Signed-off-by: LinLi <lilin24 at baidu.com>
> Signed-off-by: XunNi <nixun at baidu.com>
> Signed-off-by: YuZhang <zhangyu31 at baidu.com>
> Signed-off-by: JinYu <jin.yu at intel.com>

There should be a space between FirstName and LastName.

[...]
>  
> +struct inflight_desc_split {
> +	uint8_t		inflight;
> +	uint8_t		padding[5];
> +	uint16_t	next;
> +	uint64_t	counter;
> +};
> +
> +struct inflight_info_split {
> +	uint64_t		features;
> +	uint16_t		version;
> +	uint16_t		desc_num;
> +	uint16_t		last_inflight_io;
> +	uint16_t		used_idx;
> +	struct inflight_desc_split desc[0];
> +};

I haven't looked into the details yet. Is it necessary to
expose above definitions to applications? It seems they are
not used in your example code.

> +
> +struct resubmit_desc {
> +	uint16_t index;
> +	uint64_t counter;
> +};
> +
> +struct resubmit_info {
> +	struct resubmit_desc	*resubmit_list;
> +	uint16_t		resubmit_num;
> +};

The rte_vhost_ prefix is needed for vhost APIs.

> +
> +struct rte_vhost_ring_inflight_split {
> +	union {
> +		struct inflight_info_split *inflight_split;
> +		/* TODO */
> +	};

Ditto. Seems it's not used in example code.

> +
> +	struct resubmit_info *resubmit_inflight_split;
> +};
> +
>  struct rte_vhost_vring {
>  	struct vring_desc	*desc;
>  	struct vring_avail	*avail;
> @@ -603,6 +642,22 @@ uint16_t rte_vhost_dequeue_burst(int vid, uint16_t queue_id,
>   */
>  int rte_vhost_get_mem_table(int vid, struct rte_vhost_memory **mem);
>  
> +/**
> + * Get guest inflight vring info, including inflight ring and resubmit list.
> + *
> + * @param vid
> + *  vhost device ID
> + * @param vring_idx
> + *  vring index
> + * @param vring
> + *  the structure to hold the requested inflight vring info
> + * @return
> + *  0 on success, -1 on failure
> + */
> +int __rte_experimental
> +rte_vhost_get_vhost_ring_inflight_split(int vid, uint16_t vring_idx,
> +			      struct rte_vhost_ring_inflight_split *vring);
> +
>  /**
>   * Get guest vring info, including the vring address, vring size, etc.
>   *
> @@ -631,6 +686,56 @@ int rte_vhost_get_vhost_vring(int vid, uint16_t vring_idx,
>   */
>  int rte_vhost_vring_call(int vid, uint16_t vring_idx);
>  
> +/**
> + * set inflight flag for a desc.
> + *
> + * @param vid
> + *  vhost device ID
> + * @param vring_idx
> + *  vring index
> + * @param idx
> + *  inflight entry index
> + * @return
> + *  0 on success, -1 on failure
> + */
> +int __rte_experimental
> +rte_vhost_set_inflight_desc_split(int vid, uint16_t vring_idx,
> +		uint16_t idx);
[...]
> +int __rte_experimental
> +rte_vhost_clr_inflight_desc_split(int vid, uint16_t vring_idx,
> +		uint16_t last_used_idx, uint16_t idx);
[...]
> +int __rte_experimental
> +rte_vhost_set_last_inflight_io_split(int vid, uint16_t vring_idx,
> +		uint16_t idx);

What will be different in the function prototype when we design
above three APIs for packed ring?