net/i40e: extract non-x86 specific code from vector driver
authorJianbo Liu <jianbo.liu@linaro.org>
Fri, 14 Oct 2016 04:00:00 +0000 (09:30 +0530)
committerBruce Richardson <bruce.richardson@intel.com>
Wed, 26 Oct 2016 16:01:25 +0000 (18:01 +0200)
move scalar code which does not use x86 intrinsic functions to new file
"i40e_rxtx_vec_common.h", while keeping x86 code in i40e_rxtx_vec.c.
This allows the scalar code to to be shared among vector drivers for
different platforms.

Signed-off-by: Jianbo Liu <jianbo.liu@linaro.org>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
drivers/net/i40e/i40e_rxtx_vec.c
drivers/net/i40e/i40e_rxtx_vec_common.h [new file with mode: 0644]

index 0ee0241..3607312 100644 (file)
@@ -39,6 +39,7 @@
 #include "base/i40e_type.h"
 #include "i40e_ethdev.h"
 #include "i40e_rxtx.h"
+#include "i40e_rxtx_vec_common.h"
 
 #include <tmmintrin.h>
 
@@ -445,68 +446,6 @@ i40e_recv_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
        return _recv_raw_pkts_vec(rx_queue, rx_pkts, nb_pkts, NULL);
 }
 
-static inline uint16_t
-reassemble_packets(struct i40e_rx_queue *rxq, struct rte_mbuf **rx_bufs,
-                  uint16_t nb_bufs, uint8_t *split_flags)
-{
-       struct rte_mbuf *pkts[RTE_I40E_VPMD_RX_BURST]; /*finished pkts*/
-       struct rte_mbuf *start = rxq->pkt_first_seg;
-       struct rte_mbuf *end =  rxq->pkt_last_seg;
-       unsigned pkt_idx, buf_idx;
-
-       for (buf_idx = 0, pkt_idx = 0; buf_idx < nb_bufs; buf_idx++) {
-               if (end != NULL) {
-                       /* processing a split packet */
-                       end->next = rx_bufs[buf_idx];
-                       rx_bufs[buf_idx]->data_len += rxq->crc_len;
-
-                       start->nb_segs++;
-                       start->pkt_len += rx_bufs[buf_idx]->data_len;
-                       end = end->next;
-
-                       if (!split_flags[buf_idx]) {
-                               /* it's the last packet of the set */
-                               start->hash = end->hash;
-                               start->ol_flags = end->ol_flags;
-                               /* we need to strip crc for the whole packet */
-                               start->pkt_len -= rxq->crc_len;
-                               if (end->data_len > rxq->crc_len) {
-                                       end->data_len -= rxq->crc_len;
-                               } else {
-                                       /* free up last mbuf */
-                                       struct rte_mbuf *secondlast = start;
-
-                                       while (secondlast->next != end)
-                                               secondlast = secondlast->next;
-                                       secondlast->data_len -= (rxq->crc_len -
-                                                       end->data_len);
-                                       secondlast->next = NULL;
-                                       rte_pktmbuf_free_seg(end);
-                                       end = secondlast;
-                               }
-                               pkts[pkt_idx++] = start;
-                               start = end = NULL;
-                       }
-               } else {
-                       /* not processing a split packet */
-                       if (!split_flags[buf_idx]) {
-                               /* not a split packet, save and skip */
-                               pkts[pkt_idx++] = rx_bufs[buf_idx];
-                               continue;
-                       }
-                       end = start = rx_bufs[buf_idx];
-                       rx_bufs[buf_idx]->data_len += rxq->crc_len;
-                       rx_bufs[buf_idx]->pkt_len += rxq->crc_len;
-               }
-       }
-
-       /* save the partial packet for next time */
-       rxq->pkt_first_seg = start;
-       rxq->pkt_last_seg = end;
-       memcpy(rx_bufs, pkts, pkt_idx * (sizeof(*pkts)));
-       return pkt_idx;
-}
-
  /* vPMD receive routine that reassembles scattered packets
  * Notice:
  * - nb_pkts < RTE_I40E_DESCS_PER_LOOP, just return no packet
@@ -572,73 +511,6 @@ vtx(volatile struct i40e_tx_desc *txdp,
                vtx1(txdp, *pkt, flags);
 }
 
-static inline int __attribute__((always_inline))
-i40e_tx_free_bufs(struct i40e_tx_queue *txq)
-{
-       struct i40e_tx_entry *txep;
-       uint32_t n;
-       uint32_t i;
-       int nb_free = 0;
-       struct rte_mbuf *m, *free[RTE_I40E_TX_MAX_FREE_BUF_SZ];
-
-       /* check DD bits on threshold descriptor */
-       if ((txq->tx_ring[txq->tx_next_dd].cmd_type_offset_bsz &
-                       rte_cpu_to_le_64(I40E_TXD_QW1_DTYPE_MASK)) !=
-                       rte_cpu_to_le_64(I40E_TX_DESC_DTYPE_DESC_DONE))
-               return 0;
-
-       n = txq->tx_rs_thresh;
-
-        /* first buffer to free from S/W ring is at index
-         * tx_next_dd - (tx_rs_thresh-1)
-         */
-       txep = &txq->sw_ring[txq->tx_next_dd - (n - 1)];
-       m = __rte_pktmbuf_prefree_seg(txep[0].mbuf);
-       if (likely(m != NULL)) {
-               free[0] = m;
-               nb_free = 1;
-               for (i = 1; i < n; i++) {
-                       m = __rte_pktmbuf_prefree_seg(txep[i].mbuf);
-                       if (likely(m != NULL)) {
-                               if (likely(m->pool == free[0]->pool)) {
-                                       free[nb_free++] = m;
-                               } else {
-                                       rte_mempool_put_bulk(free[0]->pool,
-                                                            (void *)free,
-                                                            nb_free);
-                                       free[0] = m;
-                                       nb_free = 1;
-                               }
-                       }
-               }
-               rte_mempool_put_bulk(free[0]->pool, (void **)free, nb_free);
-       } else {
-               for (i = 1; i < n; i++) {
-                       m = __rte_pktmbuf_prefree_seg(txep[i].mbuf);
-                       if (m != NULL)
-                               rte_mempool_put(m->pool, m);
-               }
-       }
-
-       /* buffers were freed, update counters */
-       txq->nb_tx_free = (uint16_t)(txq->nb_tx_free + txq->tx_rs_thresh);
-       txq->tx_next_dd = (uint16_t)(txq->tx_next_dd + txq->tx_rs_thresh);
-       if (txq->tx_next_dd >= txq->nb_tx_desc)
-               txq->tx_next_dd = (uint16_t)(txq->tx_rs_thresh - 1);
-
-       return txq->tx_rs_thresh;
-}
-
-static inline void __attribute__((always_inline))
-tx_backlog_entry(struct i40e_tx_entry *txep,
-                struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
-{
-       int i;
-
-       for (i = 0; i < (int)nb_pkts; ++i)
-               txep[i].mbuf = tx_pkts[i];
-}
-
 uint16_t
 i40e_xmit_pkts_vec(void *tx_queue, struct rte_mbuf **tx_pkts,
                   uint16_t nb_pkts)
@@ -709,49 +581,13 @@ i40e_xmit_pkts_vec(void *tx_queue, struct rte_mbuf **tx_pkts,
 void __attribute__((cold))
 i40e_rx_queue_release_mbufs_vec(struct i40e_rx_queue *rxq)
 {
-       const unsigned mask = rxq->nb_rx_desc - 1;
-       unsigned i;
-
-       if (rxq->sw_ring == NULL || rxq->rxrearm_nb >= rxq->nb_rx_desc)
-               return;
-
-       /* free all mbufs that are valid in the ring */
-       if (rxq->rxrearm_nb == 0) {
-               for (i = 0; i < rxq->nb_rx_desc; i++) {
-                       if (rxq->sw_ring[i].mbuf != NULL)
-                               rte_pktmbuf_free_seg(rxq->sw_ring[i].mbuf);
-               }
-       } else {
-               for (i = rxq->rx_tail;
-                    i != rxq->rxrearm_start;
-                    i = (i + 1) & mask) {
-                       if (rxq->sw_ring[i].mbuf != NULL)
-                               rte_pktmbuf_free_seg(rxq->sw_ring[i].mbuf);
-               }
-       }
-
-       rxq->rxrearm_nb = rxq->nb_rx_desc;
-
-       /* set all entries to NULL */
-       memset(rxq->sw_ring, 0, sizeof(rxq->sw_ring[0]) * rxq->nb_rx_desc);
+       _i40e_rx_queue_release_mbufs_vec(rxq);
 }
 
 int __attribute__((cold))
 i40e_rxq_vec_setup(struct i40e_rx_queue *rxq)
 {
-       uintptr_t p;
-       struct rte_mbuf mb_def = { .buf_addr = 0 }; /* zeroed mbuf */
-
-       mb_def.nb_segs = 1;
-       mb_def.data_off = RTE_PKTMBUF_HEADROOM;
-       mb_def.port = rxq->port_id;
-       rte_mbuf_refcnt_set(&mb_def, 1);
-
-       /* prevent compiler reordering: rearm_data covers previous fields */
-       rte_compiler_barrier();
-       p = (uintptr_t)&mb_def.rearm_data;
-       rxq->mbuf_initializer = *(uint64_t *)p;
-       return 0;
+       return i40e_rxq_vec_setup_default(rxq);
 }
 
 int __attribute__((cold))
@@ -764,34 +600,10 @@ int __attribute__((cold))
 i40e_rx_vec_dev_conf_condition_check(struct rte_eth_dev *dev)
 {
 #ifndef RTE_LIBRTE_IEEE1588
-       struct rte_eth_rxmode *rxmode = &dev->data->dev_conf.rxmode;
-       struct rte_fdir_conf *fconf = &dev->data->dev_conf.fdir_conf;
-
        /* need SSE4.1 support */
        if (!rte_cpu_get_flag_enabled(RTE_CPUFLAG_SSE4_1))
                return -1;
-
-#ifndef RTE_LIBRTE_I40E_RX_OLFLAGS_ENABLE
-       /* whithout rx ol_flags, no VP flag report */
-       if (rxmode->hw_vlan_strip != 0 ||
-           rxmode->hw_vlan_extend != 0 ||
-           rxmode->hw_ip_checksum != 0)
-               return -1;
 #endif
 
-       /* no fdir support */
-       if (fconf->mode != RTE_FDIR_MODE_NONE)
-               return -1;
-
-        /* - no csum error report support
-        * - no header split support
-        */
-       if (rxmode->header_split == 1)
-               return -1;
-
-       return 0;
-#else
-       RTE_SET_USED(dev);
-       return -1;
-#endif
+       return i40e_rx_vec_dev_conf_condition_check_default(dev);
 }
diff --git a/drivers/net/i40e/i40e_rxtx_vec_common.h b/drivers/net/i40e/i40e_rxtx_vec_common.h
new file mode 100644 (file)
index 0000000..6cb5dce
--- /dev/null
@@ -0,0 +1,251 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright(c) 2010-2015 Intel Corporation. All rights reserved.
+ *   All rights reserved.
+ *
+ *   Redistribution and use in source and binary forms, with or without
+ *   modification, are permitted provided that the following conditions
+ *   are met:
+ *
+ *     * Redistributions of source code must retain the above copyright
+ *       notice, this list of conditions and the following disclaimer.
+ *     * Redistributions in binary form must reproduce the above copyright
+ *       notice, this list of conditions and the following disclaimer in
+ *       the documentation and/or other materials provided with the
+ *       distribution.
+ *     * Neither the name of Intel Corporation nor the names of its
+ *       contributors may be used to endorse or promote products derived
+ *       from this software without specific prior written permission.
+ *
+ *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ *   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ *   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ *   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ *   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ *   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ *   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ *   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ *   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ *   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ *   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef _I40E_RXTX_VEC_COMMON_H_
+#define _I40E_RXTX_VEC_COMMON_H_
+#include <stdint.h>
+#include <rte_ethdev.h>
+#include <rte_malloc.h>
+
+#include "i40e_ethdev.h"
+#include "i40e_rxtx.h"
+
+static inline uint16_t
+reassemble_packets(struct i40e_rx_queue *rxq, struct rte_mbuf **rx_bufs,
+                  uint16_t nb_bufs, uint8_t *split_flags)
+{
+       struct rte_mbuf *pkts[RTE_I40E_VPMD_RX_BURST]; /*finished pkts*/
+       struct rte_mbuf *start = rxq->pkt_first_seg;
+       struct rte_mbuf *end =  rxq->pkt_last_seg;
+       unsigned pkt_idx, buf_idx;
+
+       for (buf_idx = 0, pkt_idx = 0; buf_idx < nb_bufs; buf_idx++) {
+               if (end != NULL) {
+                       /* processing a split packet */
+                       end->next = rx_bufs[buf_idx];
+                       rx_bufs[buf_idx]->data_len += rxq->crc_len;
+
+                       start->nb_segs++;
+                       start->pkt_len += rx_bufs[buf_idx]->data_len;
+                       end = end->next;
+
+                       if (!split_flags[buf_idx]) {
+                               /* it's the last packet of the set */
+                               start->hash = end->hash;
+                               start->ol_flags = end->ol_flags;
+                               /* we need to strip crc for the whole packet */
+                               start->pkt_len -= rxq->crc_len;
+                               if (end->data_len > rxq->crc_len) {
+                                       end->data_len -= rxq->crc_len;
+                               } else {
+                                       /* free up last mbuf */
+                                       struct rte_mbuf *secondlast = start;
+
+                                       while (secondlast->next != end)
+                                               secondlast = secondlast->next;
+                                       secondlast->data_len -= (rxq->crc_len -
+                                                       end->data_len);
+                                       secondlast->next = NULL;
+                                       rte_pktmbuf_free_seg(end);
+                                       end = secondlast;
+                               }
+                               pkts[pkt_idx++] = start;
+                               start = end = NULL;
+                       }
+               } else {
+                       /* not processing a split packet */
+                       if (!split_flags[buf_idx]) {
+                               /* not a split packet, save and skip */
+                               pkts[pkt_idx++] = rx_bufs[buf_idx];
+                               continue;
+                       }
+                       end = start = rx_bufs[buf_idx];
+                       rx_bufs[buf_idx]->data_len += rxq->crc_len;
+                       rx_bufs[buf_idx]->pkt_len += rxq->crc_len;
+               }
+       }
+
+       /* save the partial packet for next time */
+       rxq->pkt_first_seg = start;
+       rxq->pkt_last_seg = end;
+       memcpy(rx_bufs, pkts, pkt_idx * (sizeof(*pkts)));
+       return pkt_idx;
+}
+
+static inline int __attribute__((always_inline))
+i40e_tx_free_bufs(struct i40e_tx_queue *txq)
+{
+       struct i40e_tx_entry *txep;
+       uint32_t n;
+       uint32_t i;
+       int nb_free = 0;
+       struct rte_mbuf *m, *free[RTE_I40E_TX_MAX_FREE_BUF_SZ];
+
+       /* check DD bits on threshold descriptor */
+       if ((txq->tx_ring[txq->tx_next_dd].cmd_type_offset_bsz &
+                       rte_cpu_to_le_64(I40E_TXD_QW1_DTYPE_MASK)) !=
+                       rte_cpu_to_le_64(I40E_TX_DESC_DTYPE_DESC_DONE))
+               return 0;
+
+       n = txq->tx_rs_thresh;
+
+        /* first buffer to free from S/W ring is at index
+         * tx_next_dd - (tx_rs_thresh-1)
+         */
+       txep = &txq->sw_ring[txq->tx_next_dd - (n - 1)];
+       m = __rte_pktmbuf_prefree_seg(txep[0].mbuf);
+       if (likely(m != NULL)) {
+               free[0] = m;
+               nb_free = 1;
+               for (i = 1; i < n; i++) {
+                       m = __rte_pktmbuf_prefree_seg(txep[i].mbuf);
+                       if (likely(m != NULL)) {
+                               if (likely(m->pool == free[0]->pool)) {
+                                       free[nb_free++] = m;
+                               } else {
+                                       rte_mempool_put_bulk(free[0]->pool,
+                                                            (void *)free,
+                                                            nb_free);
+                                       free[0] = m;
+                                       nb_free = 1;
+                               }
+                       }
+               }
+               rte_mempool_put_bulk(free[0]->pool, (void **)free, nb_free);
+       } else {
+               for (i = 1; i < n; i++) {
+                       m = __rte_pktmbuf_prefree_seg(txep[i].mbuf);
+                       if (m != NULL)
+                               rte_mempool_put(m->pool, m);
+               }
+       }
+
+       /* buffers were freed, update counters */
+       txq->nb_tx_free = (uint16_t)(txq->nb_tx_free + txq->tx_rs_thresh);
+       txq->tx_next_dd = (uint16_t)(txq->tx_next_dd + txq->tx_rs_thresh);
+       if (txq->tx_next_dd >= txq->nb_tx_desc)
+               txq->tx_next_dd = (uint16_t)(txq->tx_rs_thresh - 1);
+
+       return txq->tx_rs_thresh;
+}
+
+static inline void __attribute__((always_inline))
+tx_backlog_entry(struct i40e_tx_entry *txep,
+                struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
+{
+       int i;
+
+       for (i = 0; i < (int)nb_pkts; ++i)
+               txep[i].mbuf = tx_pkts[i];
+}
+
+static inline void
+_i40e_rx_queue_release_mbufs_vec(struct i40e_rx_queue *rxq)
+{
+       const unsigned mask = rxq->nb_rx_desc - 1;
+       unsigned i;
+
+       if (rxq->sw_ring == NULL || rxq->rxrearm_nb >= rxq->nb_rx_desc)
+               return;
+
+       /* free all mbufs that are valid in the ring */
+       if (rxq->rxrearm_nb == 0) {
+               for (i = 0; i < rxq->nb_rx_desc; i++) {
+                       if (rxq->sw_ring[i].mbuf != NULL)
+                               rte_pktmbuf_free_seg(rxq->sw_ring[i].mbuf);
+               }
+       } else {
+               for (i = rxq->rx_tail;
+                    i != rxq->rxrearm_start;
+                    i = (i + 1) & mask) {
+                       if (rxq->sw_ring[i].mbuf != NULL)
+                               rte_pktmbuf_free_seg(rxq->sw_ring[i].mbuf);
+               }
+       }
+
+       rxq->rxrearm_nb = rxq->nb_rx_desc;
+
+       /* set all entries to NULL */
+       memset(rxq->sw_ring, 0, sizeof(rxq->sw_ring[0]) * rxq->nb_rx_desc);
+}
+
+static inline int
+i40e_rxq_vec_setup_default(struct i40e_rx_queue *rxq)
+{
+       uintptr_t p;
+       struct rte_mbuf mb_def = { .buf_addr = 0 }; /* zeroed mbuf */
+
+       mb_def.nb_segs = 1;
+       mb_def.data_off = RTE_PKTMBUF_HEADROOM;
+       mb_def.port = rxq->port_id;
+       rte_mbuf_refcnt_set(&mb_def, 1);
+
+       /* prevent compiler reordering: rearm_data covers previous fields */
+       rte_compiler_barrier();
+       p = (uintptr_t)&mb_def.rearm_data;
+       rxq->mbuf_initializer = *(uint64_t *)p;
+       return 0;
+}
+
+static inline int
+i40e_rx_vec_dev_conf_condition_check_default(struct rte_eth_dev *dev)
+{
+#ifndef RTE_LIBRTE_IEEE1588
+       struct rte_eth_rxmode *rxmode = &dev->data->dev_conf.rxmode;
+       struct rte_fdir_conf *fconf = &dev->data->dev_conf.fdir_conf;
+
+#ifndef RTE_LIBRTE_I40E_RX_OLFLAGS_ENABLE
+       /* whithout rx ol_flags, no VP flag report */
+       if (rxmode->hw_vlan_strip != 0 ||
+           rxmode->hw_vlan_extend != 0 ||
+           rxmode->hw_ip_checksum != 0)
+               return -1;
+#endif
+
+       /* no fdir support */
+       if (fconf->mode != RTE_FDIR_MODE_NONE)
+               return -1;
+
+        /* - no csum error report support
+        * - no header split support
+        */
+       if (rxmode->header_split == 1)
+               return -1;
+
+       return 0;
+#else
+       RTE_SET_USED(dev);
+       return -1;
+#endif
+}
+#endif