net: add macro for MAC address print
[dpdk.git] / drivers / net / enic / enic_res.c
index 84c5d33..0493e09 100644 (file)
@@ -1,39 +1,10 @@
-/*
- * Copyright 2008-2010 Cisco Systems, Inc.  All rights reserved.
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright 2008-2017 Cisco Systems, Inc.  All rights reserved.
  * Copyright 2007 Nuova Systems, Inc.  All rights reserved.
- *
- * Copyright (c) 2014, Cisco Systems, Inc.
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * 1. Redistributions of source code must retain the above copyright
- * notice, this list of conditions and the following disclaimer.
- *
- * 2. 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.
- *
- * 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 HOLDER 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.
- *
  */
 
 #include "enic_compat.h"
-#include "rte_ethdev.h"
+#include "ethdev_driver.h"
 #include "wq_enet_desc.h"
 #include "rq_enet_desc.h"
 #include "cq_enet_desc.h"
@@ -54,6 +25,7 @@ int enic_get_vnic_config(struct enic *enic)
 {
        struct vnic_enet_config *c = &enic->config;
        int err;
+       uint64_t sizes;
 
        err = vnic_dev_get_mac_addr(enic->vdev, enic->mac_addr);
        if (err) {
@@ -62,6 +34,7 @@ int enic_get_vnic_config(struct enic *enic)
                return err;
        }
 
+
 #define GET_CONFIG(m) \
        do { \
                err = vnic_dev_spec(enic->vdev, \
@@ -88,33 +61,59 @@ int enic_get_vnic_config(struct enic *enic)
        /* max packet size is only defined in newer VIC firmware
         * and will be 0 for legacy firmware and VICs
         */
-       if (c->max_pkt_size > ENIC_DEFAULT_MAX_PKT_SIZE)
-               enic->max_mtu = c->max_pkt_size - (ETHER_HDR_LEN + 4);
+       if (c->max_pkt_size > ENIC_DEFAULT_RX_MAX_PKT_SIZE)
+               enic->max_mtu = c->max_pkt_size - RTE_ETHER_HDR_LEN;
        else
-               enic->max_mtu = ENIC_DEFAULT_MAX_PKT_SIZE - (ETHER_HDR_LEN + 4);
+               enic->max_mtu = ENIC_DEFAULT_RX_MAX_PKT_SIZE -
+                       RTE_ETHER_HDR_LEN;
        if (c->mtu == 0)
                c->mtu = 1500;
 
-       enic->rte_dev->data->mtu = min_t(u16, enic->max_mtu,
-                                        max_t(u16, ENIC_MIN_MTU, c->mtu));
+       enic->rte_dev->data->mtu = RTE_MIN(enic->max_mtu,
+                               RTE_MAX((uint16_t)ENIC_MIN_MTU, c->mtu));
+
+       enic->adv_filters = vnic_dev_capable_adv_filters(enic->vdev);
+       dev_info(enic, "Advanced Filters %savailable\n", ((enic->adv_filters)
+                ? "" : "not "));
 
-       c->wq_desc_count =
-               min_t(u32, ENIC_MAX_WQ_DESCS,
-               max_t(u32, ENIC_MIN_WQ_DESCS,
-               c->wq_desc_count));
+       err = vnic_dev_capable_filter_mode(enic->vdev, &enic->flow_filter_mode,
+                                          &enic->filter_actions);
+       if (err) {
+               dev_err(enic_get_dev(enic),
+                       "Error getting filter modes, %d\n", err);
+               return err;
+       }
+       vnic_dev_capable_udp_rss_weak(enic->vdev, &enic->nic_cfg_chk,
+                                     &enic->udp_rss_weak);
+
+       dev_info(enic, "Flow api filter mode: %s Actions: %s%s%s%s\n",
+               ((enic->flow_filter_mode == FILTER_FLOWMAN) ? "FLOWMAN" :
+               ((enic->flow_filter_mode == FILTER_DPDK_1) ? "DPDK" :
+               ((enic->flow_filter_mode == FILTER_USNIC_IP) ? "USNIC" :
+               ((enic->flow_filter_mode == FILTER_IPV4_5TUPLE) ? "5TUPLE" :
+               "NONE")))),
+               ((enic->filter_actions & FILTER_ACTION_RQ_STEERING_FLAG) ?
+                "steer " : ""),
+               ((enic->filter_actions & FILTER_ACTION_FILTER_ID_FLAG) ?
+                "tag " : ""),
+               ((enic->filter_actions & FILTER_ACTION_DROP_FLAG) ?
+                "drop " : ""),
+               ((enic->filter_actions & FILTER_ACTION_COUNTER_FLAG) ?
+                "count " : ""));
+
+       c->wq_desc_count = RTE_MIN((uint32_t)ENIC_MAX_WQ_DESCS,
+                       RTE_MAX((uint32_t)ENIC_MIN_WQ_DESCS, c->wq_desc_count));
        c->wq_desc_count &= 0xffffffe0; /* must be aligned to groups of 32 */
 
-       c->rq_desc_count =
-               min_t(u32, ENIC_MAX_RQ_DESCS,
-               max_t(u32, ENIC_MIN_RQ_DESCS,
-               c->rq_desc_count));
+       c->rq_desc_count = RTE_MIN((uint32_t)ENIC_MAX_RQ_DESCS,
+                       RTE_MAX((uint32_t)ENIC_MIN_RQ_DESCS, c->rq_desc_count));
        c->rq_desc_count &= 0xffffffe0; /* must be aligned to groups of 32 */
 
-       c->intr_timer_usec = min_t(u32, c->intr_timer_usec,
-               vnic_dev_get_intr_coal_timer_max(enic->vdev));
+       c->intr_timer_usec = RTE_MIN(c->intr_timer_usec,
+                                 vnic_dev_get_intr_coal_timer_max(enic->vdev));
 
        dev_info(enic_get_dev(enic),
-               "vNIC MAC addr %02x:%02x:%02x:%02x:%02x:%02x "
+               "vNIC MAC addr " RTE_ETHER_ADDR_PRT_FMT
                "wq/rq %d/%d mtu %d, max mtu:%d\n",
                enic->mac_addr[0], enic->mac_addr[1], enic->mac_addr[2],
                enic->mac_addr[3], enic->mac_addr[4], enic->mac_addr[5],
@@ -125,7 +124,10 @@ int enic_get_vnic_config(struct enic *enic)
                "loopback tag 0x%04x\n",
                ENIC_SETTING(enic, TXCSUM) ? "yes" : "no",
                ENIC_SETTING(enic, RXCSUM) ? "yes" : "no",
-               ENIC_SETTING(enic, RSS) ? "yes" : "no",
+               ENIC_SETTING(enic, RSS) ?
+                       (ENIC_SETTING(enic, RSSHASH_UDPIPV4) ? "+UDP" :
+                       ((enic->udp_rss_weak ? "+udp" :
+                       "yes"))) : "no",
                c->intr_mode == VENET_INTR_MODE_INTX ? "INTx" :
                c->intr_mode == VENET_INTR_MODE_MSI ? "MSI" :
                c->intr_mode == VENET_INTR_MODE_ANY ? "any" :
@@ -136,41 +138,102 @@ int enic_get_vnic_config(struct enic *enic)
                c->intr_timer_usec,
                c->loop_tag);
 
-       return 0;
-}
+       /* RSS settings from vNIC */
+       enic->reta_size = ENIC_RSS_RETA_SIZE;
+       enic->hash_key_size = ENIC_RSS_HASH_KEY_SIZE;
+       enic->flow_type_rss_offloads = 0;
+       if (ENIC_SETTING(enic, RSSHASH_IPV4))
+               /*
+                * IPV4 hash type handles both non-frag and frag packet types.
+                * TCP/UDP is controlled via a separate flag below.
+                */
+               enic->flow_type_rss_offloads |= ETH_RSS_IPV4 |
+                       ETH_RSS_FRAG_IPV4 | ETH_RSS_NONFRAG_IPV4_OTHER;
+       if (ENIC_SETTING(enic, RSSHASH_TCPIPV4))
+               enic->flow_type_rss_offloads |= ETH_RSS_NONFRAG_IPV4_TCP;
+       if (ENIC_SETTING(enic, RSSHASH_IPV6))
+               /*
+                * The VIC adapter can perform RSS on IPv6 packets with and
+                * without extension headers. An IPv6 "fragment" is an IPv6
+                * packet with the fragment extension header.
+                */
+               enic->flow_type_rss_offloads |= ETH_RSS_IPV6 |
+                       ETH_RSS_IPV6_EX | ETH_RSS_FRAG_IPV6 |
+                       ETH_RSS_NONFRAG_IPV6_OTHER;
+       if (ENIC_SETTING(enic, RSSHASH_TCPIPV6))
+               enic->flow_type_rss_offloads |= ETH_RSS_NONFRAG_IPV6_TCP |
+                       ETH_RSS_IPV6_TCP_EX;
+       if (enic->udp_rss_weak)
+               enic->flow_type_rss_offloads |=
+                       ETH_RSS_NONFRAG_IPV4_UDP | ETH_RSS_NONFRAG_IPV6_UDP |
+                       ETH_RSS_IPV6_UDP_EX;
+       if (ENIC_SETTING(enic, RSSHASH_UDPIPV4))
+               enic->flow_type_rss_offloads |= ETH_RSS_NONFRAG_IPV4_UDP;
+       if (ENIC_SETTING(enic, RSSHASH_UDPIPV6))
+               enic->flow_type_rss_offloads |= ETH_RSS_NONFRAG_IPV6_UDP |
+                       ETH_RSS_IPV6_UDP_EX;
 
-int enic_add_vlan(struct enic *enic, u16 vlanid)
-{
-       u64 a0 = vlanid, a1 = 0;
-       int wait = 1000;
-       int err;
+       /* Zero offloads if RSS is not enabled */
+       if (!ENIC_SETTING(enic, RSS))
+               enic->flow_type_rss_offloads = 0;
 
-       err = vnic_dev_cmd(enic->vdev, CMD_VLAN_ADD, &a0, &a1, wait);
-       if (err)
-               dev_err(enic_get_dev(enic), "Can't add vlan id, %d\n", err);
+       enic->vxlan = ENIC_SETTING(enic, VXLAN) &&
+               vnic_dev_capable_vxlan(enic->vdev);
+       enic->geneve = ENIC_SETTING(enic, GENEVE) &&
+               vnic_dev_capable_geneve(enic->vdev);
 
-       return err;
-}
-
-int enic_del_vlan(struct enic *enic, u16 vlanid)
-{
-       u64 a0 = vlanid, a1 = 0;
-       int wait = 1000;
-       int err;
+       /* Supported CQ entry sizes */
+       enic->cq_entry_sizes = vnic_dev_capable_cq_entry_size(enic->vdev);
+       sizes = enic->cq_entry_sizes;
+       dev_debug(NULL, "Supported CQ entry sizes:%s%s%s\n",
+                 (sizes & VNIC_RQ_CQ_ENTRY_SIZE_16_CAPABLE) ? " 16" : "",
+                 (sizes & VNIC_RQ_CQ_ENTRY_SIZE_32_CAPABLE) ? " 32" : "",
+                 (sizes & VNIC_RQ_CQ_ENTRY_SIZE_64_CAPABLE) ? " 64" : "");
+       /* Use 64B entry if requested and available */
+       enic->cq64 = enic->cq64_request &&
+               (sizes & VNIC_RQ_CQ_ENTRY_SIZE_64_CAPABLE);
+       dev_debug(NULL, "Using %sB CQ entry size\n", enic->cq64 ? "64" : "16");
 
-       err = vnic_dev_cmd(enic->vdev, CMD_VLAN_DEL, &a0, &a1, wait);
-       if (err)
-               dev_err(enic_get_dev(enic), "Can't delete vlan id, %d\n", err);
+       /*
+        * Default hardware capabilities. enic_dev_init() may add additional
+        * flags if it enables overlay offloads.
+        */
+       enic->tx_queue_offload_capa = 0;
+       enic->tx_offload_capa =
+               enic->tx_queue_offload_capa |
+               DEV_TX_OFFLOAD_MULTI_SEGS |
+               DEV_TX_OFFLOAD_VLAN_INSERT |
+               DEV_TX_OFFLOAD_IPV4_CKSUM |
+               DEV_TX_OFFLOAD_UDP_CKSUM |
+               DEV_TX_OFFLOAD_TCP_CKSUM |
+               DEV_TX_OFFLOAD_TCP_TSO;
+       enic->rx_offload_capa =
+               DEV_RX_OFFLOAD_SCATTER |
+               DEV_RX_OFFLOAD_JUMBO_FRAME |
+               DEV_RX_OFFLOAD_VLAN_STRIP |
+               DEV_RX_OFFLOAD_IPV4_CKSUM |
+               DEV_RX_OFFLOAD_UDP_CKSUM |
+               DEV_RX_OFFLOAD_TCP_CKSUM |
+               DEV_RX_OFFLOAD_RSS_HASH;
+       enic->tx_offload_mask =
+               PKT_TX_IPV6 |
+               PKT_TX_IPV4 |
+               PKT_TX_VLAN |
+               PKT_TX_IP_CKSUM |
+               PKT_TX_L4_MASK |
+               PKT_TX_TCP_SEG;
 
-       return err;
+       return 0;
 }
 
-int enic_set_nic_cfg(struct enic *enic, u8 rss_default_cpu, u8 rss_hash_type,
-       u8 rss_hash_bits, u8 rss_base_cpu, u8 rss_enable, u8 tso_ipid_split_en,
-       u8 ig_vlan_strip_en)
+int enic_set_nic_cfg(struct enic *enic, uint8_t rss_default_cpu,
+                    uint8_t rss_hash_type, uint8_t rss_hash_bits,
+                    uint8_t rss_base_cpu, uint8_t rss_enable,
+                    uint8_t tso_ipid_split_en, uint8_t ig_vlan_strip_en)
 {
-       u64 a0, a1;
-       u32 nic_cfg;
+       enum vnic_devcmd_cmd cmd;
+       uint64_t a0, a1;
+       uint32_t nic_cfg;
        int wait = 1000;
 
        vnic_set_nic_cfg(&nic_cfg, rss_default_cpu,
@@ -179,21 +242,21 @@ int enic_set_nic_cfg(struct enic *enic, u8 rss_default_cpu, u8 rss_hash_type,
 
        a0 = nic_cfg;
        a1 = 0;
-
-       return vnic_dev_cmd(enic->vdev, CMD_NIC_CFG, &a0, &a1, wait);
+       cmd = enic->nic_cfg_chk ? CMD_NIC_CFG_CHK : CMD_NIC_CFG;
+       return vnic_dev_cmd(enic->vdev, cmd, &a0, &a1, wait);
 }
 
-int enic_set_rss_key(struct enic *enic, dma_addr_t key_pa, u64 len)
+int enic_set_rss_key(struct enic *enic, dma_addr_t key_pa, uint64_t len)
 {
-       u64 a0 = (u64)key_pa, a1 = len;
+       uint64_t a0 = (uint64_t)key_pa, a1 = len;
        int wait = 1000;
 
        return vnic_dev_cmd(enic->vdev, CMD_RSS_KEY, &a0, &a1, wait);
 }
 
-int enic_set_rss_cpu(struct enic *enic, dma_addr_t cpu_pa, u64 len)
+int enic_set_rss_cpu(struct enic *enic, dma_addr_t cpu_pa, uint64_t len)
 {
-       u64 a0 = (u64)cpu_pa, a1 = len;
+       uint64_t a0 = (uint64_t)cpu_pa, a1 = len;
        int wait = 1000;
 
        return vnic_dev_cmd(enic->vdev, CMD_RSS_CPU, &a0, &a1, wait);
@@ -210,7 +273,8 @@ void enic_free_vnic_resources(struct enic *enic)
                        vnic_rq_free(&enic->rq[i]);
        for (i = 0; i < enic->cq_count; i++)
                vnic_cq_free(&enic->cq[i]);
-       vnic_intr_free(&enic->intr);
+       for (i = 0; i < enic->intr_count; i++)
+               vnic_intr_free(&enic->intr[i]);
 }
 
 void enic_get_res_counts(struct enic *enic)