Re: [dpdk-dev] [PATCH 01/13] ethdev: support setup function for hairpin queue

[Ori Kam <orika@mellanox.com>]

Hi Andrew,
Thanks for your comments PSB.
 
> -----Original Message-----
> From: Andrew Rybchenko <arybchenko@solarflare.com>
> On 9/26/19 9:28 AM, Ori Kam wrote:
> > This commit introduce the RX/TX hairpin setup function.
>
> > RX/TX should be Rx/Tx here and everywhere below.
> >
> > Hairpin is RX/TX queue that is used by the nic in order to offload
> > wire to wire traffic.
> >
> > Each hairpin queue is binded to one or more queues from other type.
> > For example TX hairpin queue should be binded to at least 1 RX hairpin
> > queue and vice versa.
>
> How should application find out that hairpin queues are supported?

It should be stated in the release note of the DPDK, when manufacture adds support for this.
In addition if the application try to set hairpin queue and it fails it can mean depending on the
error that the hairpin is not supported.

> How many?

There is no limit to the number of hairpin queues from application all queues can be hairpin queues.

> How should application find out which ports/queues could be used for
> pining?

All ports and queues can be supported, if the application request invalid combination, for example
in current Mellanox implementation binding between two ports then the setup function will  fail.

If you would like I can add capability for this, but there are too many options. For example number
of queues, binding limitations all of those will be very hard to declare. 


> Is hair-pinning domain on device level sufficient to expose limitations?
>
I'm sorry but I don’t understand your question.

> > Signed-off-by: Ori Kam <orika@mellanox.com>
> > ---
> >   lib/librte_ethdev/rte_ethdev.c           | 213
>> +++++++++++++++++++++++++++++++
> >   lib/librte_ethdev/rte_ethdev.h           | 145 +++++++++++++++++++++
> >   lib/librte_ethdev/rte_ethdev_core.h      |  18 +++
> >   lib/librte_ethdev/rte_ethdev_version.map |   4 +
> >   4 files changed, 380 insertions(+)
> >
> > diff --git a/lib/librte_ethdev/rte_ethdev.c
> > b/lib/librte_ethdev/rte_ethdev.c index 30b0c78..4021f38 100644
> > --- a/lib/librte_ethdev/rte_ethdev.c
> > +++ b/lib/librte_ethdev/rte_ethdev.c
> > @@ -1701,6 +1701,115 @@ struct rte_eth_dev *
> >   }
> >
> >   int
> > +rte_eth_rx_hairpin_queue_setup(uint16_t port_id, uint16_t
> > rx_queue_id,
> > +			       uint16_t nb_rx_desc, unsigned int socket_id,
> > +			       const struct rte_eth_rxconf *rx_conf,
> > +			       const struct rte_eth_hairpin_conf *hairpin_conf)
>
> > Below code duplicates rte_eth_rx_queue_setup() a lot and it is very bad
> > from maintenance point of view. Similar problem with Tx hairpin queue
> > setup.
> >

I'm aware of that. The reasons I choose it are: (same goes to Tx)
1. use the same function approach, meaning to use the current  setup function
    the issues with this are:
     * API break.
     * It will have extra parameters, for example mempool will not be used
        for hairpin and hairpin configuration will not be used for normal queue.
        It is possible to use a struct but again API break and some fields are not used.
     * we are just starting with the hairpin, most likely there will be modification so
         it is better to have a different function.
     * From application he undertand that this is a different kind of queue, which shouldn't be 
         used by the application.

> > +{
> > +	int ret;
> > +	struct rte_eth_dev *dev;
> > +	struct rte_eth_dev_info dev_info;
> > +	struct rte_eth_rxconf local_conf;
> > +	void **rxq;
> > +
> > +	RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -EINVAL);
> > +
> > +	dev = &rte_eth_devices[port_id];
> > +	if (rx_queue_id >= dev->data->nb_rx_queues) {
> > +		RTE_ETHDEV_LOG(ERR, "Invalid RX queue_id=%u\n",
> rx_queue_id);
> > +		return -EINVAL;
> > +	}
> > +
> > +	RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->dev_infos_get, -
> ENOTSUP);
> > +	RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops-
> >rx_hairpin_queue_setup,
> > +				-ENOTSUP);
> > +
> > +	rte_eth_dev_info_get(port_id, &dev_info);
> > +
> > +	/* Use default specified by driver, if nb_rx_desc is zero */
> > +	if (nb_rx_desc == 0) {
> > +		nb_rx_desc = dev_info.default_rxportconf.ring_size;
> > +		/* If driver default is also zero, fall back on EAL default */
> > +		if (nb_rx_desc == 0)
> > +			nb_rx_desc =
> RTE_ETH_DEV_FALLBACK_RX_RINGSIZE;
> > +	}
> > +
> > +	if (nb_rx_desc > dev_info.rx_desc_lim.nb_max ||
> > +			nb_rx_desc < dev_info.rx_desc_lim.nb_min ||
> > +			nb_rx_desc % dev_info.rx_desc_lim.nb_align != 0) {
> > +
> > +		RTE_ETHDEV_LOG(ERR,
> > +			       "Invalid value for nb_rx_desc(=%hu), should be: "
> > +			       "<= %hu, >= %hu, and a product of %hu\n",
> > +			nb_rx_desc, dev_info.rx_desc_lim.nb_max,
> > +			dev_info.rx_desc_lim.nb_min,
> > +			dev_info.rx_desc_lim.nb_align);
> > +		return -EINVAL;
> > +	}
> > +
> > +	if (dev->data->dev_started &&
> > +		!(dev_info.dev_capa &
> > +
> 	RTE_ETH_DEV_CAPA_RUNTIME_RX_QUEUE_SETUP))
> > +		return -EBUSY;
> > +
> > +	if (dev->data->dev_started &&
> > +		(dev->data->rx_queue_state[rx_queue_id] !=
> > +			RTE_ETH_QUEUE_STATE_STOPPED))
> > +		return -EBUSY;
> > +
> > +	rxq = dev->data->rx_queues;
> > +	if (rxq[rx_queue_id]) {
> > +		RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops-
> >rx_queue_release,
> > +					-ENOTSUP);
> > +		(*dev->dev_ops->rx_queue_release)(rxq[rx_queue_id]);
> > +		rxq[rx_queue_id] = NULL;
> > +	}
> > +
> > +	if (rx_conf == NULL)
> > +		rx_conf = &dev_info.default_rxconf;
> > +
> > +	local_conf = *rx_conf;
> > +
> > +	/*
> > +	 * If an offloading has already been enabled in
> > +	 * rte_eth_dev_configure(), it has been enabled on all queues,
> > +	 * so there is no need to enable it in this queue again.
> > +	 * The local_conf.offloads input to underlying PMD only carries
> > +	 * those offloadings which are only enabled on this queue and
> > +	 * not enabled on all queues.
> > +	 */
> > +	local_conf.offloads &= ~dev->data->dev_conf.rxmode.offloads;
> > +
> > +	/*
> > +	 * New added offloadings for this queue are those not enabled in
> > +	 * rte_eth_dev_configure() and they must be per-queue type.
> > +	 * A pure per-port offloading can't be enabled on a queue while
> > +	 * disabled on another queue. A pure per-port offloading can't
> > +	 * be enabled for any queue as new added one if it hasn't been
> > +	 * enabled in rte_eth_dev_configure().
> > +	 */
> > +	if ((local_conf.offloads & dev_info.rx_queue_offload_capa) !=
> > +	     local_conf.offloads) {
> > +		RTE_ETHDEV_LOG(ERR,
> > +			"Ethdev port_id=%d rx_queue_id=%d, "
> > +			"new added offloads 0x%"PRIx64" must be "
> > +			"within per-queue offload capabilities "
> > +			"0x%"PRIx64" in %s()\n",
> > +			port_id, rx_queue_id, local_conf.offloads,
> > +			dev_info.rx_queue_offload_capa,
> > +			__func__);
> > +		return -EINVAL;
> > +	}
> > +
> > +	ret = (*dev->dev_ops->rx_hairpin_queue_setup)(dev,
> rx_queue_id,
> > +						      nb_rx_desc, socket_id,
> > +						      &local_conf,
> > +						      hairpin_conf);
> > +
> > +	return eth_err(port_id, ret);
> > +}
> > +
> > +int
> >   rte_eth_tx_queue_setup(uint16_t port_id, uint16_t tx_queue_id,
> >   		       uint16_t nb_tx_desc, unsigned int socket_id,
> >   		       const struct rte_eth_txconf *tx_conf) @@ -1799,6
> +1908,110
> > @@ struct rte_eth_dev *
> >   		       tx_queue_id, nb_tx_desc, socket_id, &local_conf));
> >   }
> >
> > +int
> > +rte_eth_tx_hairpin_queue_setup(uint16_t port_id, uint16_t
> tx_queue_id,
> > +			       uint16_t nb_tx_desc, unsigned int socket_id,
> > +			       const struct rte_eth_txconf *tx_conf,
> > +			       const struct rte_eth_hairpin_conf *hairpin_conf)
> {
> > +	struct rte_eth_dev *dev;
> > +	struct rte_eth_dev_info dev_info;
> > +	struct rte_eth_txconf local_conf;
> > +	void **txq;
> > +
> > +	RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -EINVAL);
> > +
> > +	dev = &rte_eth_devices[port_id];
> > +	if (tx_queue_id >= dev->data->nb_tx_queues) {
> > +		RTE_ETHDEV_LOG(ERR, "Invalid TX queue_id=%u\n",
> tx_queue_id);
> > +		return -EINVAL;
> > +	}
> > +
> > +	RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->dev_infos_get, -
> ENOTSUP);
> > +	RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops-
> >tx_hairpin_queue_setup,
> > +				-ENOTSUP);
> > +
> > +	rte_eth_dev_info_get(port_id, &dev_info);
> > +
> > +	/* Use default specified by driver, if nb_tx_desc is zero */
> > +	if (nb_tx_desc == 0) {
> > +		nb_tx_desc = dev_info.default_txportconf.ring_size;
> > +		/* If driver default is zero, fall back on EAL default */
> > +		if (nb_tx_desc == 0)
> > +			nb_tx_desc =
> RTE_ETH_DEV_FALLBACK_TX_RINGSIZE;
> > +	}
> > +	if (nb_tx_desc > dev_info.tx_desc_lim.nb_max ||
> > +	    nb_tx_desc < dev_info.tx_desc_lim.nb_min ||
> > +	    nb_tx_desc % dev_info.tx_desc_lim.nb_align != 0) {
> > +		RTE_ETHDEV_LOG(ERR,
> > +			       "Invalid value for nb_tx_desc(=%hu), "
> > +			       "should be: <= %hu, >= %hu, and a product of "
> > +			       " %hu\n",
> > +			       nb_tx_desc, dev_info.tx_desc_lim.nb_max,
> > +			       dev_info.tx_desc_lim.nb_min,
> > +			       dev_info.tx_desc_lim.nb_align);
> > +		return -EINVAL;
> > +	}
> > +
> > +	if (dev->data->dev_started &&
> > +		!(dev_info.dev_capa &
> > +		  RTE_ETH_DEV_CAPA_RUNTIME_TX_QUEUE_SETUP))
> > +		return -EBUSY;
> > +
> > +	if (dev->data->dev_started &&
> > +		(dev->data->tx_queue_state[tx_queue_id] !=
> > +		 RTE_ETH_QUEUE_STATE_STOPPED))
> > +		return -EBUSY;
> > +
> > +	txq = dev->data->tx_queues;
> > +	if (txq[tx_queue_id]) {
> > +		RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops-
> >tx_queue_release,
> > +					-ENOTSUP);
> > +		(*dev->dev_ops->tx_queue_release)(txq[tx_queue_id]);
> > +		txq[tx_queue_id] = NULL;
> > +	}
> > +
> > +	if (tx_conf == NULL)
> > +		tx_conf = &dev_info.default_txconf;
> > +
> > +	local_conf = *tx_conf;
> > +
> > +	/*
> > +	 * If an offloading has already been enabled in
> > +	 * rte_eth_dev_configure(), it has been enabled on all queues,
> > +	 * so there is no need to enable it in this queue again.
> > +	 * The local_conf.offloads input to underlying PMD only carries
> > +	 * those offloadings which are only enabled on this queue and
> > +	 * not enabled on all queues.
> > +	 */
> > +	local_conf.offloads &= ~dev->data->dev_conf.txmode.offloads;
> > +
> > +	/*
> > +	 * New added offloadings for this queue are those not enabled in
> > +	 * rte_eth_dev_configure() and they must be per-queue type.
> > +	 * A pure per-port offloading can't be enabled on a queue while
> > +	 * disabled on another queue. A pure per-port offloading can't
> > +	 * be enabled for any queue as new added one if it hasn't been
> > +	 * enabled in rte_eth_dev_configure().
> > +	 */
> > +	if ((local_conf.offloads & dev_info.tx_queue_offload_capa) !=
> > +	     local_conf.offloads) {
> > +		RTE_ETHDEV_LOG(ERR,
> > +			       "Ethdev port_id=%d tx_queue_id=%d, new
> added "
> > +			       "offloads 0x%"PRIx64" must be within "
> > +			       "per-queue offload capabilities 0x%"PRIx64" "
> > +			       "in %s()\n",
> > +			       port_id, tx_queue_id, local_conf.offloads,
> > +			       dev_info.tx_queue_offload_capa,
> > +			       __func__);
> > +		return -EINVAL;
> > +	}
> > +
> > +	return eth_err(port_id, (*dev->dev_ops->tx_hairpin_queue_setup)
> > +		       (dev, tx_queue_id, nb_tx_desc, socket_id, &local_conf,
> > +			hairpin_conf));
> > +}
> > +
> >   void
> >   rte_eth_tx_buffer_drop_callback(struct rte_mbuf **pkts, uint16_t
> unsent,
> >   		void *userdata __rte_unused)
> > diff --git a/lib/librte_ethdev/rte_ethdev.h
> > b/lib/librte_ethdev/rte_ethdev.h index 475dbda..b3b1597 100644
> > --- a/lib/librte_ethdev/rte_ethdev.h
> > +++ b/lib/librte_ethdev/rte_ethdev.h
> > @@ -803,6 +803,30 @@ struct rte_eth_txconf {
> >   	uint64_t offloads;
> >   };
> >
> > +#define RTE_ETH_MAX_HAIRPIN_PEERS 32
> > +
> > +/**
> > + * @warning
> > + * @b EXPERIMENTAL: this API may change, or be removed, without prior
> > +notice
> > + *
> > + * A structure used to hold hairpin peer data.
> > + */
> > +struct rte_eth_hairpin_peer {
> > +	uint16_t port; /**< Peer port. */
> > +	uint16_t queue; /**< Peer queue. */
> > +};
> > +
> > +/**
> > + * @warning
> > + * @b EXPERIMENTAL: this API may change, or be removed, without prior
> > +notice
> > + *
> > + * A structure used to configure hairpin binding.
> > + */
> > +struct rte_eth_hairpin_conf {
> > +	uint16_t peer_n; /**< The number of peers. */
> > +	struct rte_eth_hairpin_peer peers[RTE_ETH_MAX_HAIRPIN_PEERS];
> };
> > +
> >   /**
> >    * A structure contains information about HW descriptor ring limitations.
> >    */
> > @@ -1769,6 +1793,60 @@ int rte_eth_rx_queue_setup(uint16_t port_id,
> uint16_t rx_queue_id,
> >   		struct rte_mempool *mb_pool);
> >
> >   /**
> > + * @warning
> > + * @b EXPERIMENTAL: this API may change, or be removed, without prior
> > + notice
> > + *
> > + * Allocate and set up a hairpin receive queue for an Ethernet device.
> > + *
> > + * The function set up the selected queue to be used in hairpin.
> > + *
> > + * @param port_id
> > + *   The port identifier of the Ethernet device.
> > + * @param rx_queue_id
> > + *   The index of the receive queue to set up.
> > + *   The value must be in the range [0, nb_rx_queue - 1] previously
> supplied
> > + *   to rte_eth_dev_configure().
>
> Is any Rx queue may be setup as hairpin queue?
> Can it be still used for regular traffic?
>

No if a queue is used as hairpin it can't be used for normal traffic.
This is also why I like the idea of two different functions, in order to create
This distinction.

> > + * @param nb_rx_desc
> > + *   The number of receive descriptors to allocate for the receive ring.
>
> Does it still make sense for hairpin queue?
>

Yes, since it can affect memory size used by the device, and can affect performance.

> > + * @param socket_id
> > + *   The *socket_id* argument is the socket identifier in case of NUMA.
> > + *   The value can be *SOCKET_ID_ANY* if there is no NUMA constraint
> for
> > + *   the DMA memory allocated for the receive descriptors of the ring.
>
> Is it still required to be provided for hairpin Rx queue?
>

Yes, for internal PMD structures to be allocated, but we can if pressed remove it.

> > + * @param rx_conf
> > + *   The pointer to the configuration data to be used for the receive
> queue.
> > + *   NULL value is allowed, in which case default RX configuration
> > + *   will be used.
> > + *   The *rx_conf* structure contains an *rx_thresh* structure with the
> values
> > + *   of the Prefetch, Host, and Write-Back threshold registers of the
> receive
> > + *   ring.
> > + *   In addition it contains the hardware offloads features to activate using
> > + *   the DEV_RX_OFFLOAD_* flags.
> > + *   If an offloading set in rx_conf->offloads
> > + *   hasn't been set in the input argument eth_conf->rxmode.offloads
> > + *   to rte_eth_dev_configure(), it is a new added offloading, it must be
> > + *   per-queue type and it is enabled for the queue.
> > + *   No need to repeat any bit in rx_conf->offloads which has already been
> > + *   enabled in rte_eth_dev_configure() at port level. An offloading
> enabled
> > + *   at port level can't be disabled at queue level.
>
> Which offloads still make sense in the case of hairpin Rx queue?
> What about threshhods, drop enable?
>

Drop and thresholds make sense, for example the application can state that,
in case of back pressure to start dropping packets in order not to affect the
entire nic.
regarding offloads mainly vlan strip or vlan insert but those can also 
be used in rte_flow.
But future offloads like QoS or other maybe shared.

> > + * @param hairpin_conf
> > + *   The pointer to the hairpin binding configuration.
> > + * @return
> > + *   - 0: Success, receive queue correctly set up.
> > + *   - -EINVAL: The size of network buffers which can be allocated from the
> > + *      memory pool does not fit the various buffer sizes allowed by the
> > + *      device controller.
> > + *   - -ENOMEM: Unable to allocate the receive ring descriptors or to
> > + *      allocate network memory buffers from the memory pool when
> > + *      initializing receive descriptors.
> > + */
> > +__rte_experimental
> > +int rte_eth_rx_hairpin_queue_setup
> > +	(uint16_t port_id, uint16_t rx_queue_id,
> > +	 uint16_t nb_rx_desc, unsigned int socket_id,
> > +	 const struct rte_eth_rxconf *rx_conf,
> > +	 const struct rte_eth_hairpin_conf *hairpin_conf);
> > +
> > +/**
> >    * Allocate and set up a transmit queue for an Ethernet device.
> >    *
> >    * @param port_id
> > @@ -1821,6 +1899,73 @@ int rte_eth_tx_queue_setup(uint16_t port_id,
> uint16_t tx_queue_id,
> >   		const struct rte_eth_txconf *tx_conf);
> >
> >   /**
> > + * @warning
> > + * @b EXPERIMENTAL: this API may change, or be removed, without prior
> > + notice
> > + *
> > + * Allocate and set up a transmit hairpin queue for an Ethernet device.
> > + *
> > + * @param port_id
> > + *   The port identifier of the Ethernet device.
> > + * @param tx_queue_id
> > + *   The index of the transmit queue to set up.
> > + *   The value must be in the range [0, nb_tx_queue - 1] previously
> supplied
> > + *   to rte_eth_dev_configure().
>
> Is any Tx queue may be setup as hairpin queue?
>

Yes just like any Rx queue.

> > + * @param nb_tx_desc
> > + *   The number of transmit descriptors to allocate for the transmit ring.
>
> Is it really required for hairpin queue? Are min/max/align limits still the
> same?
>
The number of descriptors can effect memory and performance.
Regarding min/max/align I guess this depends on the implementation in the nic.

> > + * @param socket_id
> > + *   The *socket_id* argument is the socket identifier in case of NUMA.
> > + *   Its value can be *SOCKET_ID_ANY* if there is no NUMA constraint for
> > + *   the DMA memory allocated for the transmit descriptors of the ring.
>
> Does it still make sense for Tx hairpin queue?
>

Same as for the RX, it is used for internal PMD structures, but maybe on 
other nics they can use this.

> > + * @param tx_conf
> > + *   The pointer to the configuration data to be used for the transmit
> queue.
> > + *   NULL value is allowed, in which case default RX configuration
> > + *   will be used.
> > + *   The *tx_conf* structure contains the following data:
> > + *   - The *tx_thresh* structure with the values of the Prefetch, Host, and
> > + *     Write-Back threshold registers of the transmit ring.
> > + *     When setting Write-Back threshold to the value greater then zero,
> > + *     *tx_rs_thresh* value should be explicitly set to one.
> > + *   - The *tx_free_thresh* value indicates the [minimum] number of
> network
> > + *     buffers that must be pending in the transmit ring to trigger their
> > + *     [implicit] freeing by the driver transmit function.
> > + *   - The *tx_rs_thresh* value indicates the [minimum] number of
> transmit
> > + *     descriptors that must be pending in the transmit ring before setting
> the
> > + *     RS bit on a descriptor by the driver transmit function.
> > + *     The *tx_rs_thresh* value should be less or equal then
> > + *     *tx_free_thresh* value, and both of them should be less then
> > + *     *nb_tx_desc* - 3.
>
> I'm not sure that everything above makes sense for hairpin Tx queue.
>

You are right not all of them make sense,
But since I don't know other nics I prefer to give them those values, if they need them.
If you wish I can change the documentation.

> > + *   - The *txq_flags* member contains flags to pass to the TX queue
> setup
> > + *     function to configure the behavior of the TX queue. This should be
> set
> > + *     to 0 if no special configuration is required.
> > + *     This API is obsolete and will be deprecated. Applications
> > + *     should set it to ETH_TXQ_FLAGS_IGNORE and use
> > + *     the offloads field below.
>
> There is no txq_flags for a long time already. So, I'm wondering when it was
> copies from rte_eth_tx_queue_setup().
>
My bad from 17.11. will fix.

> > + *   - The *offloads* member contains Tx offloads to be enabled.
> > + *     If an offloading set in tx_conf->offloads
> > + *     hasn't been set in the input argument eth_conf->txmode.offloads
> > + *     to rte_eth_dev_configure(), it is a new added offloading, it must be
> > + *     per-queue type and it is enabled for the queue.
> > + *     No need to repeat any bit in tx_conf->offloads which has already
> been
> > + *     enabled in rte_eth_dev_configure() at port level. An offloading
> enabled
> > + *     at port level can't be disabled at queue level.
>
> Which offloads do really make sense and valid to use for hairpin Tx queues?
> Do we need separate caps for hairpin offloads?
>
I'm sure that we will need caps for example QoS but I don't know which yet.


> > + *
> > + *     Note that setting *tx_free_thresh* or *tx_rs_thresh* value to 0
> forces
> > + *     the transmit function to use default values.
> > + * @param hairpin_conf
> > + *   The hairpin binding configuration.
> > + *
> > + * @return
> > + *   - 0: Success, the transmit queue is correctly set up.
> > + *   - -ENOMEM: Unable to allocate the transmit ring descriptors.
> > + */
> > +__rte_experimental
> > +int rte_eth_tx_hairpin_queue_setup
> > +	(uint16_t port_id, uint16_t tx_queue_id,
> > +	 uint16_t nb_tx_desc, unsigned int socket_id,
> > +	 const struct rte_eth_txconf *tx_conf,
> > +	 const struct rte_eth_hairpin_conf *hairpin_conf);
> > +
> > +/**
> >    * Return the NUMA socket to which an Ethernet device is connected
> >    *
> >    * @param port_id
> >
>
> [snip]