2 * Copyright(c) 2017 Intel Corporation.
4 * This file is provided under a dual BSD/GPLv2 license. When using or
5 * redistributing this file, you may do so under either license.
9 * This program is free software; you can redistribute it and/or modify
10 * it under the terms of version 2 of the GNU General Public License as
11 * published by the Free Software Foundation.
13 * This program is distributed in the hope that it will be useful, but
14 * WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 * General Public License for more details.
20 * Redistribution and use in source and binary forms, with or without
21 * modification, are permitted provided that the following conditions
24 * - Redistributions of source code must retain the above copyright
25 * notice, this list of conditions and the following disclaimer.
26 * - Redistributions in binary form must reproduce the above copyright
27 * notice, this list of conditions and the following disclaimer in
28 * the documentation and/or other materials provided with the
30 * - Neither the name of Intel Corporation nor the names of its
31 * contributors may be used to endorse or promote products derived
32 * from this software without specific prior written permission.
34 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
35 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
36 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
37 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
38 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
39 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
40 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
41 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
42 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
43 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
44 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
49 * This file contains OPA Virtual Network Interface Controller (VNIC) driver
50 * netdev functionality.
53 #include <linux/module.h>
54 #include <linux/if_vlan.h>
55 #include <linux/crc32.h>
57 #include "opa_vnic_internal.h"
59 #define OPA_TX_TIMEOUT_MS 1000
61 #define OPA_VNIC_SKB_HEADROOM \
62 ALIGN((OPA_VNIC_HDR_LEN + OPA_VNIC_SKB_MDATA_LEN), 8)
64 /* This function is overloaded for opa_vnic specific implementation */
65 static void opa_vnic_get_stats64(struct net_device
*netdev
,
66 struct rtnl_link_stats64
*stats
)
68 struct opa_vnic_adapter
*adapter
= opa_vnic_priv(netdev
);
69 struct opa_vnic_stats vstats
;
71 memset(&vstats
, 0, sizeof(vstats
));
72 mutex_lock(&adapter
->stats_lock
);
73 adapter
->rn_ops
->ndo_get_stats64(netdev
, &vstats
.netstats
);
74 mutex_unlock(&adapter
->stats_lock
);
75 memcpy(stats
, &vstats
.netstats
, sizeof(*stats
));
78 /* opa_netdev_start_xmit - transmit function */
79 static netdev_tx_t
opa_netdev_start_xmit(struct sk_buff
*skb
,
80 struct net_device
*netdev
)
82 struct opa_vnic_adapter
*adapter
= opa_vnic_priv(netdev
);
84 v_dbg("xmit: queue %d skb len %d\n", skb
->queue_mapping
, skb
->len
);
85 /* pad to ensure mininum ethernet packet length */
86 if (unlikely(skb
->len
< ETH_ZLEN
)) {
87 if (skb_padto(skb
, ETH_ZLEN
))
90 skb_put(skb
, ETH_ZLEN
- skb
->len
);
93 opa_vnic_encap_skb(adapter
, skb
);
94 return adapter
->rn_ops
->ndo_start_xmit(skb
, netdev
);
97 static u16
opa_vnic_select_queue(struct net_device
*netdev
, struct sk_buff
*skb
,
99 select_queue_fallback_t fallback
)
101 struct opa_vnic_adapter
*adapter
= opa_vnic_priv(netdev
);
102 struct opa_vnic_skb_mdata
*mdata
;
105 /* pass entropy and vl as metadata in skb */
106 mdata
= (struct opa_vnic_skb_mdata
*)skb_push(skb
, sizeof(*mdata
));
107 mdata
->entropy
= opa_vnic_calc_entropy(adapter
, skb
);
108 mdata
->vl
= opa_vnic_get_vl(adapter
, skb
);
109 rc
= adapter
->rn_ops
->ndo_select_queue(netdev
, skb
,
110 accel_priv
, fallback
);
111 skb_pull(skb
, sizeof(*mdata
));
115 /* opa_vnic_process_vema_config - process vema configuration updates */
116 void opa_vnic_process_vema_config(struct opa_vnic_adapter
*adapter
)
118 struct __opa_veswport_info
*info
= &adapter
->info
;
119 struct rdma_netdev
*rn
= netdev_priv(adapter
->netdev
);
120 u8 port_num
[OPA_VESW_MAX_NUM_DEF_PORT
] = { 0 };
121 struct net_device
*netdev
= adapter
->netdev
;
122 u8 i
, port_count
= 0;
125 /* If the base_mac_addr is changed, update the interface mac address */
126 if (memcmp(info
->vport
.base_mac_addr
, adapter
->vema_mac_addr
,
127 ARRAY_SIZE(info
->vport
.base_mac_addr
))) {
128 struct sockaddr saddr
;
130 memcpy(saddr
.sa_data
, info
->vport
.base_mac_addr
,
131 ARRAY_SIZE(info
->vport
.base_mac_addr
));
132 mutex_lock(&adapter
->lock
);
133 eth_mac_addr(netdev
, &saddr
);
134 memcpy(adapter
->vema_mac_addr
,
135 info
->vport
.base_mac_addr
, ETH_ALEN
);
136 mutex_unlock(&adapter
->lock
);
139 rn
->set_id(netdev
, info
->vesw
.vesw_id
);
141 /* Handle MTU limit change */
143 netdev
->max_mtu
= max_t(unsigned int, info
->vesw
.eth_mtu_non_vlan
,
145 if (netdev
->mtu
> netdev
->max_mtu
)
146 dev_set_mtu(netdev
, netdev
->max_mtu
);
149 /* Update flow to default port redirection table */
150 port_mask
= info
->vesw
.def_port_mask
;
151 for (i
= 0; i
< OPA_VESW_MAX_NUM_DEF_PORT
; i
++) {
153 port_num
[port_count
++] = i
;
158 * Build the flow table. Flow table is required when destination LID
159 * is not available. Up to OPA_VNIC_FLOW_TBL_SIZE flows supported.
160 * Each flow need a default port number to get its dlid from the
161 * u_ucast_dlid array.
163 for (i
= 0; i
< OPA_VNIC_FLOW_TBL_SIZE
; i
++)
164 adapter
->flow_tbl
[i
] = port_count
? port_num
[i
% port_count
] :
165 OPA_VNIC_INVALID_PORT
;
167 /* Operational state can only be DROP_ALL or FORWARDING */
168 if (info
->vport
.config_state
== OPA_VNIC_STATE_FORWARDING
) {
169 info
->vport
.oper_state
= OPA_VNIC_STATE_FORWARDING
;
170 netif_dormant_off(netdev
);
172 info
->vport
.oper_state
= OPA_VNIC_STATE_DROP_ALL
;
173 netif_dormant_on(netdev
);
178 * Set the power on default values in adapter's vema interface structure.
180 static inline void opa_vnic_set_pod_values(struct opa_vnic_adapter
*adapter
)
182 adapter
->info
.vport
.max_mac_tbl_ent
= OPA_VNIC_MAC_TBL_MAX_ENTRIES
;
183 adapter
->info
.vport
.max_smac_ent
= OPA_VNIC_MAX_SMAC_LIMIT
;
184 adapter
->info
.vport
.config_state
= OPA_VNIC_STATE_DROP_ALL
;
185 adapter
->info
.vport
.eth_link_status
= OPA_VNIC_ETH_LINK_DOWN
;
188 /* opa_vnic_set_mac_addr - change mac address */
189 static int opa_vnic_set_mac_addr(struct net_device
*netdev
, void *addr
)
191 struct opa_vnic_adapter
*adapter
= opa_vnic_priv(netdev
);
192 struct sockaddr
*sa
= addr
;
195 if (!memcmp(netdev
->dev_addr
, sa
->sa_data
, ETH_ALEN
))
198 mutex_lock(&adapter
->lock
);
199 rc
= eth_mac_addr(netdev
, addr
);
200 mutex_unlock(&adapter
->lock
);
204 adapter
->info
.vport
.uc_macs_gen_count
++;
205 opa_vnic_vema_report_event(adapter
,
206 OPA_VESWPORT_TRAP_IFACE_UCAST_MAC_CHANGE
);
211 * opa_vnic_mac_send_event - post event on possible mac list exchange
212 * Send trap when digest from uc/mc mac list differs from previous run.
213 * Digest is evaluated similar to how cksum does.
215 static void opa_vnic_mac_send_event(struct net_device
*netdev
, u8 event
)
217 struct opa_vnic_adapter
*adapter
= opa_vnic_priv(netdev
);
218 struct netdev_hw_addr
*ha
;
219 struct netdev_hw_addr_list
*hw_list
;
224 case OPA_VESWPORT_TRAP_IFACE_UCAST_MAC_CHANGE
:
225 hw_list
= &netdev
->uc
;
226 adapter
->info
.vport
.uc_macs_gen_count
++;
227 ref_crc
= &adapter
->umac_hash
;
229 case OPA_VESWPORT_TRAP_IFACE_MCAST_MAC_CHANGE
:
230 hw_list
= &netdev
->mc
;
231 adapter
->info
.vport
.mc_macs_gen_count
++;
232 ref_crc
= &adapter
->mmac_hash
;
237 netdev_hw_addr_list_for_each(ha
, hw_list
) {
238 crc
= crc32_le(crc
, ha
->addr
, ETH_ALEN
);
240 l
= netdev_hw_addr_list_count(hw_list
) * ETH_ALEN
;
241 crc
= ~crc32_le(crc
, (void *)&l
, sizeof(l
));
243 if (crc
!= *ref_crc
) {
245 opa_vnic_vema_report_event(adapter
, event
);
249 /* opa_vnic_set_rx_mode - handle uc/mc mac list change */
250 static void opa_vnic_set_rx_mode(struct net_device
*netdev
)
252 opa_vnic_mac_send_event(netdev
,
253 OPA_VESWPORT_TRAP_IFACE_UCAST_MAC_CHANGE
);
255 opa_vnic_mac_send_event(netdev
,
256 OPA_VESWPORT_TRAP_IFACE_MCAST_MAC_CHANGE
);
259 /* opa_netdev_open - activate network interface */
260 static int opa_netdev_open(struct net_device
*netdev
)
262 struct opa_vnic_adapter
*adapter
= opa_vnic_priv(netdev
);
265 rc
= adapter
->rn_ops
->ndo_open(adapter
->netdev
);
267 v_dbg("open failed %d\n", rc
);
271 /* Update eth link status and send trap */
272 adapter
->info
.vport
.eth_link_status
= OPA_VNIC_ETH_LINK_UP
;
273 opa_vnic_vema_report_event(adapter
,
274 OPA_VESWPORT_TRAP_ETH_LINK_STATUS_CHANGE
);
278 /* opa_netdev_close - disable network interface */
279 static int opa_netdev_close(struct net_device
*netdev
)
281 struct opa_vnic_adapter
*adapter
= opa_vnic_priv(netdev
);
284 rc
= adapter
->rn_ops
->ndo_stop(adapter
->netdev
);
286 v_dbg("close failed %d\n", rc
);
290 /* Update eth link status and send trap */
291 adapter
->info
.vport
.eth_link_status
= OPA_VNIC_ETH_LINK_DOWN
;
292 opa_vnic_vema_report_event(adapter
,
293 OPA_VESWPORT_TRAP_ETH_LINK_STATUS_CHANGE
);
298 static const struct net_device_ops opa_netdev_ops
= {
299 .ndo_open
= opa_netdev_open
,
300 .ndo_stop
= opa_netdev_close
,
301 .ndo_start_xmit
= opa_netdev_start_xmit
,
302 .ndo_get_stats64
= opa_vnic_get_stats64
,
303 .ndo_set_rx_mode
= opa_vnic_set_rx_mode
,
304 .ndo_select_queue
= opa_vnic_select_queue
,
305 .ndo_set_mac_address
= opa_vnic_set_mac_addr
,
308 /* opa_vnic_add_netdev - create vnic netdev interface */
309 struct opa_vnic_adapter
*opa_vnic_add_netdev(struct ib_device
*ibdev
,
310 u8 port_num
, u8 vport_num
)
312 struct opa_vnic_adapter
*adapter
;
313 struct net_device
*netdev
;
314 struct rdma_netdev
*rn
;
317 netdev
= ibdev
->alloc_rdma_netdev(ibdev
, port_num
,
318 RDMA_NETDEV_OPA_VNIC
,
319 "veth%d", NET_NAME_UNKNOWN
,
322 return ERR_PTR(-ENOMEM
);
323 else if (IS_ERR(netdev
))
324 return ERR_CAST(netdev
);
326 adapter
= kzalloc(sizeof(*adapter
), GFP_KERNEL
);
332 rn
= netdev_priv(netdev
);
333 rn
->clnt_priv
= adapter
;
335 rn
->port_num
= port_num
;
336 adapter
->netdev
= netdev
;
337 adapter
->ibdev
= ibdev
;
338 adapter
->port_num
= port_num
;
339 adapter
->vport_num
= vport_num
;
340 adapter
->rn_ops
= netdev
->netdev_ops
;
342 netdev
->netdev_ops
= &opa_netdev_ops
;
343 netdev
->priv_flags
|= IFF_LIVE_ADDR_CHANGE
;
344 netdev
->hard_header_len
+= OPA_VNIC_SKB_HEADROOM
;
345 mutex_init(&adapter
->lock
);
346 mutex_init(&adapter
->mactbl_lock
);
347 mutex_init(&adapter
->stats_lock
);
349 SET_NETDEV_DEV(netdev
, ibdev
->dev
.parent
);
351 opa_vnic_set_ethtool_ops(netdev
);
353 opa_vnic_set_pod_values(adapter
);
355 rc
= register_netdev(netdev
);
359 netif_carrier_off(netdev
);
360 netif_dormant_on(netdev
);
361 v_info("initialized\n");
365 mutex_destroy(&adapter
->lock
);
366 mutex_destroy(&adapter
->mactbl_lock
);
367 mutex_destroy(&adapter
->stats_lock
);
370 ibdev
->free_rdma_netdev(netdev
);
375 /* opa_vnic_rem_netdev - remove vnic netdev interface */
376 void opa_vnic_rem_netdev(struct opa_vnic_adapter
*adapter
)
378 struct net_device
*netdev
= adapter
->netdev
;
379 struct ib_device
*ibdev
= adapter
->ibdev
;
381 v_info("removing\n");
382 unregister_netdev(netdev
);
383 opa_vnic_release_mac_tbl(adapter
);
384 mutex_destroy(&adapter
->lock
);
385 mutex_destroy(&adapter
->mactbl_lock
);
386 mutex_destroy(&adapter
->stats_lock
);
388 ibdev
->free_rdma_netdev(netdev
);