Re: [dpdk-dev] [RFC] ethdev: configure SR-IOV VF from host

From: Thomas Monjalon <thomas@monjalon.net>
To: Jerin Jacob Kollanukkaran <jerinj@marvell.com>
Cc: Ferruh Yigit <ferruh.yigit@intel.com>,
	Andrew Rybchenko <arybchenko@solarflare.com>,
	"dev@dpdk.org" <dev@dpdk.org>,
	Bernard Iremonger <bernard.iremonger@intel.com>,
	Shahaf Shuler <shahafs@mellanox.com>,
	"E. Scott Daniels" <daniels@research.att.com>,
	Wenzhuo Lu <wenzhuo.lu@intel.com>, Alex Zelezniak <alexz@att.com>,
	Ajit Khaparde <ajit.khaparde@broadcom.com>,
	Declan Doherty <declan.doherty@intel.com>
Subject: Re: [dpdk-dev] [RFC] ethdev: configure SR-IOV VF from host
Date: Thu, 15 Aug 2019 19:59:18 +0200	[thread overview]
Message-ID: <3065298.OHD6ChtW5A@xps> (raw)
In-Reply-To: <BYAPR18MB24246F4379B7C00AC5B00D04C8AC0@BYAPR18MB2424.namprd18.prod.outlook.com>

15/08/2019 17:34, Jerin Jacob Kollanukkaran:
> From: Thomas Monjalon
> > 
> > In a virtual environment, the network controller may have to configure some
> > SR-IOV VF parameters for security reasons.
> > 
> > When the PF (host port) is drived by DPDK (OVS-DPDK case), we face two
> > different cases:
> > 	- driver is bifurcated (Mellanox case),
> > 	so the VF can be configured via the kernel.
> > 	- driver is on top of UIO or VFIO, so DPDK API is required.
> > 
> > This RFC proposes to use generic DPDK API for VF configuration.
> > The impacted functions are (can be extended):
> > 
> > 	- rte_eth_dev_is_valid_port
> > 	- rte_eth_promiscuous_enable
> > 	- rte_eth_promiscuous_disable
> > 	- rte_eth_promiscuous_get
> > 	- rte_eth_allmulticast_enable
> > 	- rte_eth_allmulticast_disable
> > 	- rte_eth_allmulticast_get
> > 	- rte_eth_dev_set_mc_addr_list
> > 	- rte_eth_dev_default_mac_addr_set
> > 	- rte_eth_macaddr_get
> > 	- rte_eth_dev_mac_addr_add
> > 	- rte_eth_dev_mac_addr_remove
> > 	- rte_eth_dev_vlan_filter
> > 	- rte_eth_dev_get_mtu
> > 	- rte_eth_dev_set_mtu
> > 
> > In order to target these functions to a VF (which has no port id in the host),
> > the higher bit of port id is reserved:
> > 
> > #define RTE_ETH_VF_PORT_FLAG (1 << 15)
> 
> Instead of changing the port number behavior, How about adding a bit field/
> I think, There is no ABI breakage as the parent type of bit field  is uint8_t and there
> is still more room.
> 
> --- a/lib/librte_ethdev/rte_ethdev_core.h
> +++ b/lib/librte_ethdev/rte_ethdev_core.h
> @@ -621,6 +621,7 @@ struct rte_eth_dev_data {
>                 all_multicast : 1, /**< RX all multicast mode ON(1) / OFF(0). */
>                 dev_started : 1,   /**< Device state: STARTED(1) / STOPPED(0). */
>                 lro         : 1;   /**< RX LRO is ON(1) / OFF(0) */
> +               vf         : 1;   /**< SR-IOV VF device */
>         uint8_t rx_queue_state[RTE_MAX_QUEUES_PER_PORT];
>                         /**< Queues state: STARTED(1) / STOPPED(0). */
>         uint8_t tx_queue_state[RTE_MAX_QUEUES_PER_PORT];

Sorry I don't understand how it can help.
We need to specify which VF we want to configure.

My proposal is to use the representor port,
which is connected to a VF.
We distinguish the representor and the VF with a flag in the port id
parameter passed to the functions.