d/numam-dpdk
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
bc91e8db12
numam-dpdk/drivers/net/mlx4/mlx4_rxtx.h
Yongseok Koh 31912d9924 net/mlx4: support externally allocated static memory 
When MLX PMD registers memory for DMA, it accesses the global memseg list
of DPDK to maximize the range of registration so that LKey search can be
more efficient. Granularity of MR registration is per page.

Externally allocated memory shouldn't be used for DMA because it can't be
searched in the memseg list and free event can't be tracked by DPDK. If it
is used, the following error will occur:

	net_mlx5: port 0 unable to find virtually contiguous chunk for
	address (0x5600017587c0). rte_memseg_contig_walk() failed.

There's a pending patchset [1] which enables externally allocated memory.
Once it is merged, users can register their own memory out of EAL then that
will resolve this issue.

Meanwhile, if the external memory is static (allocated on startup and never
freed), such memory can also be registered by little tweak in the code.

[1] http://patches.dpdk.org/project/dpdk/list/?series=1415

This patch is not a bug fix but needs to be included in stable versions.

Fixes: 9797bfcce1 ("net/mlx4: add new memory region support")
Cc: stable@dpdk.org

Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
2018-10-11 18:53:49 +02:00

261 lines
8.7 KiB
C
Raw Blame History

/* SPDX-License-Identifier: BSD-3-Clause
* Copyright 2017 6WIND S.A.
* Copyright 2017 Mellanox Technologies, Ltd
*/
#ifndef MLX4_RXTX_H_
#define MLX4_RXTX_H_
#include <stdint.h>
#include <sys/queue.h>
/* Verbs headers do not support -pedantic. */
#ifdef PEDANTIC
#pragma GCC diagnostic ignored "-Wpedantic"
#endif
#include <infiniband/mlx4dv.h>
#include <infiniband/verbs.h>
#ifdef PEDANTIC
#pragma GCC diagnostic error "-Wpedantic"
#endif
#include <rte_ethdev_driver.h>
#include <rte_mbuf.h>
#include <rte_mempool.h>
#include "mlx4.h"
#include "mlx4_prm.h"
#include "mlx4_mr.h"
/** Rx queue counters. */
struct mlx4_rxq_stats {
unsigned int idx; /**< Mapping index. */
uint64_t ipackets; /**< Total of successfully received packets. */
uint64_t ibytes; /**< Total of successfully received bytes. */
uint64_t idropped; /**< Total of packets dropped when Rx ring full. */
uint64_t rx_nombuf; /**< Total of Rx mbuf allocation failures. */
};
/** Rx queue descriptor. */
struct rxq {
struct priv *priv; /**< Back pointer to private data. */
struct rte_mempool *mp; /**< Memory pool for allocations. */
struct ibv_cq *cq; /**< Completion queue. */
struct ibv_wq *wq; /**< Work queue. */
struct ibv_comp_channel *channel; /**< Rx completion channel. */
uint16_t rq_ci; /**< Saved RQ consumer index. */
uint16_t port_id; /**< Port ID for incoming packets. */
uint16_t sges_n; /**< Number of segments per packet (log2 value). */
uint16_t elts_n; /**< Mbuf queue size (log2 value). */
struct mlx4_mr_ctrl mr_ctrl; /* MR control descriptor. */
struct rte_mbuf *(*elts)[]; /**< Rx elements. */
volatile struct mlx4_wqe_data_seg (*wqes)[]; /**< HW queue entries. */
volatile uint32_t *rq_db; /**< RQ doorbell record. */
uint32_t csum:1; /**< Enable checksum offloading. */
uint32_t csum_l2tun:1; /**< Same for L2 tunnels. */
uint32_t crc_present:1; /**< CRC must be subtracted. */
uint32_t l2tun_offload:1; /**< L2 tunnel offload is enabled. */
struct mlx4_cq mcq; /**< Info for directly manipulating the CQ. */
struct mlx4_rxq_stats stats; /**< Rx queue counters. */
unsigned int socket; /**< CPU socket ID for allocations. */
uint32_t usecnt; /**< Number of users relying on queue resources. */
uint8_t data[]; /**< Remaining queue resources. */
};
/** Shared flow target for Rx queues. */
struct mlx4_rss {
LIST_ENTRY(mlx4_rss) next; /**< Next entry in list. */
struct priv *priv; /**< Back pointer to private data. */
uint32_t refcnt; /**< Reference count for this object. */
uint32_t usecnt; /**< Number of users relying on @p qp and @p ind. */
struct ibv_qp *qp; /**< Queue pair. */
struct ibv_rwq_ind_table *ind; /**< Indirection table. */
uint64_t fields; /**< Fields for RSS processing (Verbs format). */
uint8_t key[MLX4_RSS_HASH_KEY_SIZE]; /**< Hash key to use. */
uint16_t queues; /**< Number of target queues. */
uint16_t queue_id[]; /**< Target queues. */
};
/** Tx element. */
struct txq_elt {
struct rte_mbuf *buf; /**< Buffer. */
union {
volatile struct mlx4_wqe_ctrl_seg *wqe; /**< SQ WQE. */
volatile uint32_t *eocb; /**< End of completion burst. */
};
};
/** Tx queue counters. */
struct mlx4_txq_stats {
unsigned int idx; /**< Mapping index. */
uint64_t opackets; /**< Total of successfully sent packets. */
uint64_t obytes; /**< Total of successfully sent bytes. */
uint64_t odropped; /**< Total number of packets failed to transmit. */
};
/** Tx queue descriptor. */
struct txq {
struct mlx4_sq msq; /**< Info for directly manipulating the SQ. */
struct mlx4_cq mcq; /**< Info for directly manipulating the CQ. */
unsigned int elts_head; /**< Current index in (*elts)[]. */
unsigned int elts_tail; /**< First element awaiting completion. */
int elts_comp_cd; /**< Countdown for next completion. */
unsigned int elts_comp_cd_init; /**< Initial value for countdown. */
unsigned int elts_n; /**< (*elts)[] length. */
struct mlx4_mr_ctrl mr_ctrl; /* MR control descriptor. */
struct txq_elt (*elts)[]; /**< Tx elements. */
struct mlx4_txq_stats stats; /**< Tx queue counters. */
uint32_t max_inline; /**< Max inline send size. */
uint32_t csum:1; /**< Enable checksum offloading. */
uint32_t csum_l2tun:1; /**< Same for L2 tunnels. */
uint32_t lb:1; /**< Whether packets should be looped back by eSwitch. */
uint8_t *bounce_buf;
/**< Memory used for storing the first DWORD of data TXBBs. */
struct priv *priv; /**< Back pointer to private data. */
unsigned int socket; /**< CPU socket ID for allocations. */
struct ibv_cq *cq; /**< Completion queue. */
struct ibv_qp *qp; /**< Queue pair. */
uint8_t data[]; /**< Remaining queue resources. */
};
/* mlx4_rxq.c */
uint8_t mlx4_rss_hash_key_default[MLX4_RSS_HASH_KEY_SIZE];
int mlx4_rss_init(struct priv *priv);
void mlx4_rss_deinit(struct priv *priv);
struct mlx4_rss *mlx4_rss_get(struct priv *priv, uint64_t fields,
const uint8_t key[MLX4_RSS_HASH_KEY_SIZE],
uint16_t queues, const uint16_t queue_id[]);
void mlx4_rss_put(struct mlx4_rss *rss);
int mlx4_rss_attach(struct mlx4_rss *rss);
void mlx4_rss_detach(struct mlx4_rss *rss);
int mlx4_rxq_attach(struct rxq *rxq);
void mlx4_rxq_detach(struct rxq *rxq);
uint64_t mlx4_get_rx_port_offloads(struct priv *priv);
uint64_t mlx4_get_rx_queue_offloads(struct priv *priv);
int mlx4_rx_queue_setup(struct rte_eth_dev *dev, uint16_t idx,
uint16_t desc, unsigned int socket,
const struct rte_eth_rxconf *conf,
struct rte_mempool *mp);
void mlx4_rx_queue_release(void *dpdk_rxq);
/* mlx4_rxtx.c */
uint16_t mlx4_tx_burst(void *dpdk_txq, struct rte_mbuf **pkts,
uint16_t pkts_n);
uint16_t mlx4_rx_burst(void *dpdk_rxq, struct rte_mbuf **pkts,
uint16_t pkts_n);
uint16_t mlx4_tx_burst_removed(void *dpdk_txq, struct rte_mbuf **pkts,
uint16_t pkts_n);
uint16_t mlx4_rx_burst_removed(void *dpdk_rxq, struct rte_mbuf **pkts,
uint16_t pkts_n);
/* mlx4_txq.c */
uint64_t mlx4_get_tx_port_offloads(struct priv *priv);
int mlx4_tx_queue_setup(struct rte_eth_dev *dev, uint16_t idx,
uint16_t desc, unsigned int socket,
const struct rte_eth_txconf *conf);
void mlx4_tx_queue_release(void *dpdk_txq);
/* mlx4_mr.c */
void mlx4_mr_flush_local_cache(struct mlx4_mr_ctrl *mr_ctrl);
uint32_t mlx4_rx_addr2mr_bh(struct rxq *rxq, uintptr_t addr);
uint32_t mlx4_tx_addr2mr_bh(struct txq *txq, uintptr_t addr);
uint32_t mlx4_tx_update_ext_mp(struct txq *txq, uintptr_t addr,
struct rte_mempool *mp);
/**
* Get Memory Pool (MP) from mbuf. If mbuf is indirect, the pool from which the
* cloned mbuf is allocated is returned instead.
*
* @param buf
* Pointer to mbuf.
*
* @return
* Memory pool where data is located for given mbuf.
*/
static struct rte_mempool *
mlx4_mb2mp(struct rte_mbuf *buf)
{
if (unlikely(RTE_MBUF_INDIRECT(buf)))
return rte_mbuf_from_indirect(buf)->pool;
return buf->pool;
}
/**
* Query LKey from a packet buffer for Rx. No need to flush local caches for Rx
* as mempool is pre-configured and static.
*
* @param rxq
* Pointer to Rx queue structure.
* @param addr
* Address to search.
*
* @return
* Searched LKey on success, UINT32_MAX on no match.
*/
static __rte_always_inline uint32_t
mlx4_rx_addr2mr(struct rxq *rxq, uintptr_t addr)
{
struct mlx4_mr_ctrl *mr_ctrl = &rxq->mr_ctrl;
uint32_t lkey;
/* Linear search on MR cache array. */
lkey = mlx4_mr_lookup_cache(mr_ctrl->cache, &mr_ctrl->mru,
MLX4_MR_CACHE_N, addr);
if (likely(lkey != UINT32_MAX))
return lkey;
/* Take slower bottom-half (Binary Search) on miss. */
return mlx4_rx_addr2mr_bh(rxq, addr);
}
#define mlx4_rx_mb2mr(rxq, mb) mlx4_rx_addr2mr(rxq, (uintptr_t)((mb)->buf_addr))
/**
* Query LKey from a packet buffer for Tx. If not found, add the mempool.
*
* @param txq
* Pointer to Tx queue structure.
* @param addr
* Address to search.
*
* @return
* Searched LKey on success, UINT32_MAX on no match.
*/
static __rte_always_inline uint32_t
mlx4_tx_addr2mr(struct txq *txq, uintptr_t addr)
{
struct mlx4_mr_ctrl *mr_ctrl = &txq->mr_ctrl;
uint32_t lkey;
/* Check generation bit to see if there's any change on existing MRs. */
if (unlikely(*mr_ctrl->dev_gen_ptr != mr_ctrl->cur_gen))
mlx4_mr_flush_local_cache(mr_ctrl);
/* Linear search on MR cache array. */
lkey = mlx4_mr_lookup_cache(mr_ctrl->cache, &mr_ctrl->mru,
MLX4_MR_CACHE_N, addr);
if (likely(lkey != UINT32_MAX))
return lkey;
/* Take slower bottom-half (binary search) on miss. */
return mlx4_tx_addr2mr_bh(txq, addr);
}
static __rte_always_inline uint32_t
mlx4_tx_mb2mr(struct txq *txq, struct rte_mbuf *mb)
{
uintptr_t addr = (uintptr_t)mb->buf_addr;
uint32_t lkey = mlx4_tx_addr2mr(txq, addr);
if (likely(lkey != UINT32_MAX))
return lkey;
if (rte_errno == ENXIO) {
/* Mempool may have externally allocated memory. */
lkey = mlx4_tx_update_ext_mp(txq, addr, mlx4_mb2mp(mb));
}
return lkey;
}
#endif /* MLX4_RXTX_H_ */
Reference in New Issue View Git Blame Copy Permalink