spectrum_router.c 218.6 KB
Newer Older
1 2
// SPDX-License-Identifier: BSD-3-Clause OR GPL-2.0
/* Copyright (c) 2016-2018 Mellanox Technologies. All rights reserved */
3 4 5

#include <linux/kernel.h>
#include <linux/types.h>
6 7 8
#include <linux/rhashtable.h>
#include <linux/bitops.h>
#include <linux/in6.h>
9
#include <linux/notifier.h>
10
#include <linux/inetdevice.h>
11
#include <linux/netdevice.h>
12
#include <linux/if_bridge.h>
13
#include <linux/socket.h>
14
#include <linux/route.h>
15
#include <linux/gcd.h>
16
#include <linux/if_macvlan.h>
17
#include <linux/refcount.h>
18
#include <linux/jhash.h>
19
#include <linux/net_namespace.h>
20
#include <linux/mutex.h>
21
#include <net/netevent.h>
22 23
#include <net/neighbour.h>
#include <net/arp.h>
24
#include <net/ip_fib.h>
25
#include <net/ip6_fib.h>
26
#include <net/nexthop.h>
27
#include <net/fib_rules.h>
28
#include <net/ip_tunnels.h>
29
#include <net/l3mdev.h>
30
#include <net/addrconf.h>
31 32
#include <net/ndisc.h>
#include <net/ipv6.h>
33
#include <net/fib_notifier.h>
34
#include <net/switchdev.h>
35 36 37 38

#include "spectrum.h"
#include "core.h"
#include "reg.h"
39 40
#include "spectrum_cnt.h"
#include "spectrum_dpipe.h"
41
#include "spectrum_ipip.h"
42 43
#include "spectrum_mr.h"
#include "spectrum_mr_tcam.h"
44
#include "spectrum_router.h"
45
#include "spectrum_span.h"
46

47
struct mlxsw_sp_fib;
48 49
struct mlxsw_sp_vr;
struct mlxsw_sp_lpm_tree;
50
struct mlxsw_sp_rif_ops;
51

52 53 54
struct mlxsw_sp_rif {
	struct list_head nexthop_list;
	struct list_head neigh_list;
55
	struct net_device *dev; /* NULL for underlay RIF */
56
	struct mlxsw_sp_fid *fid;
57 58
	unsigned char addr[ETH_ALEN];
	int mtu;
59
	u16 rif_index;
60
	u16 vr_id;
61 62 63
	const struct mlxsw_sp_rif_ops *ops;
	struct mlxsw_sp *mlxsw_sp;

64 65 66 67
	unsigned int counter_ingress;
	bool counter_ingress_valid;
	unsigned int counter_egress;
	bool counter_egress_valid;
68 69
};

70 71 72 73 74 75 76 77 78 79
struct mlxsw_sp_rif_params {
	struct net_device *dev;
	union {
		u16 system_port;
		u16 lag_id;
	};
	u16 vid;
	bool lag;
};

80 81
struct mlxsw_sp_rif_subport {
	struct mlxsw_sp_rif common;
82
	refcount_t ref_count;
83 84 85 86 87 88 89 90
	union {
		u16 system_port;
		u16 lag_id;
	};
	u16 vid;
	bool lag;
};

91 92 93 94
struct mlxsw_sp_rif_ipip_lb {
	struct mlxsw_sp_rif common;
	struct mlxsw_sp_rif_ipip_lb_config lb_config;
	u16 ul_vr_id; /* Reserved for Spectrum-2. */
95
	u16 ul_rif_id; /* Reserved for Spectrum. */
96 97 98 99 100 101 102
};

struct mlxsw_sp_rif_params_ipip_lb {
	struct mlxsw_sp_rif_params common;
	struct mlxsw_sp_rif_ipip_lb_config lb_config;
};

103 104 105 106 107 108 109 110
struct mlxsw_sp_rif_ops {
	enum mlxsw_sp_rif_type type;
	size_t rif_size;

	void (*setup)(struct mlxsw_sp_rif *rif,
		      const struct mlxsw_sp_rif_params *params);
	int (*configure)(struct mlxsw_sp_rif *rif);
	void (*deconfigure)(struct mlxsw_sp_rif *rif);
111 112
	struct mlxsw_sp_fid * (*fid_get)(struct mlxsw_sp_rif *rif,
					 struct netlink_ext_ack *extack);
113
	void (*fdb_del)(struct mlxsw_sp_rif *rif, const char *mac);
114 115
};

116 117 118
static struct mlxsw_sp_rif *
mlxsw_sp_rif_find_by_dev(const struct mlxsw_sp *mlxsw_sp,
			 const struct net_device *dev);
119
static void mlxsw_sp_rif_destroy(struct mlxsw_sp_rif *rif);
120 121 122 123 124 125 126 127 128
static void mlxsw_sp_lpm_tree_hold(struct mlxsw_sp_lpm_tree *lpm_tree);
static void mlxsw_sp_lpm_tree_put(struct mlxsw_sp *mlxsw_sp,
				  struct mlxsw_sp_lpm_tree *lpm_tree);
static int mlxsw_sp_vr_lpm_tree_bind(struct mlxsw_sp *mlxsw_sp,
				     const struct mlxsw_sp_fib *fib,
				     u8 tree_id);
static int mlxsw_sp_vr_lpm_tree_unbind(struct mlxsw_sp *mlxsw_sp,
				       const struct mlxsw_sp_fib *fib);

129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263
static unsigned int *
mlxsw_sp_rif_p_counter_get(struct mlxsw_sp_rif *rif,
			   enum mlxsw_sp_rif_counter_dir dir)
{
	switch (dir) {
	case MLXSW_SP_RIF_COUNTER_EGRESS:
		return &rif->counter_egress;
	case MLXSW_SP_RIF_COUNTER_INGRESS:
		return &rif->counter_ingress;
	}
	return NULL;
}

static bool
mlxsw_sp_rif_counter_valid_get(struct mlxsw_sp_rif *rif,
			       enum mlxsw_sp_rif_counter_dir dir)
{
	switch (dir) {
	case MLXSW_SP_RIF_COUNTER_EGRESS:
		return rif->counter_egress_valid;
	case MLXSW_SP_RIF_COUNTER_INGRESS:
		return rif->counter_ingress_valid;
	}
	return false;
}

static void
mlxsw_sp_rif_counter_valid_set(struct mlxsw_sp_rif *rif,
			       enum mlxsw_sp_rif_counter_dir dir,
			       bool valid)
{
	switch (dir) {
	case MLXSW_SP_RIF_COUNTER_EGRESS:
		rif->counter_egress_valid = valid;
		break;
	case MLXSW_SP_RIF_COUNTER_INGRESS:
		rif->counter_ingress_valid = valid;
		break;
	}
}

static int mlxsw_sp_rif_counter_edit(struct mlxsw_sp *mlxsw_sp, u16 rif_index,
				     unsigned int counter_index, bool enable,
				     enum mlxsw_sp_rif_counter_dir dir)
{
	char ritr_pl[MLXSW_REG_RITR_LEN];
	bool is_egress = false;
	int err;

	if (dir == MLXSW_SP_RIF_COUNTER_EGRESS)
		is_egress = true;
	mlxsw_reg_ritr_rif_pack(ritr_pl, rif_index);
	err = mlxsw_reg_query(mlxsw_sp->core, MLXSW_REG(ritr), ritr_pl);
	if (err)
		return err;

	mlxsw_reg_ritr_counter_pack(ritr_pl, counter_index, enable,
				    is_egress);
	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ritr), ritr_pl);
}

int mlxsw_sp_rif_counter_value_get(struct mlxsw_sp *mlxsw_sp,
				   struct mlxsw_sp_rif *rif,
				   enum mlxsw_sp_rif_counter_dir dir, u64 *cnt)
{
	char ricnt_pl[MLXSW_REG_RICNT_LEN];
	unsigned int *p_counter_index;
	bool valid;
	int err;

	valid = mlxsw_sp_rif_counter_valid_get(rif, dir);
	if (!valid)
		return -EINVAL;

	p_counter_index = mlxsw_sp_rif_p_counter_get(rif, dir);
	if (!p_counter_index)
		return -EINVAL;
	mlxsw_reg_ricnt_pack(ricnt_pl, *p_counter_index,
			     MLXSW_REG_RICNT_OPCODE_NOP);
	err = mlxsw_reg_query(mlxsw_sp->core, MLXSW_REG(ricnt), ricnt_pl);
	if (err)
		return err;
	*cnt = mlxsw_reg_ricnt_good_unicast_packets_get(ricnt_pl);
	return 0;
}

static int mlxsw_sp_rif_counter_clear(struct mlxsw_sp *mlxsw_sp,
				      unsigned int counter_index)
{
	char ricnt_pl[MLXSW_REG_RICNT_LEN];

	mlxsw_reg_ricnt_pack(ricnt_pl, counter_index,
			     MLXSW_REG_RICNT_OPCODE_CLEAR);
	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ricnt), ricnt_pl);
}

int mlxsw_sp_rif_counter_alloc(struct mlxsw_sp *mlxsw_sp,
			       struct mlxsw_sp_rif *rif,
			       enum mlxsw_sp_rif_counter_dir dir)
{
	unsigned int *p_counter_index;
	int err;

	p_counter_index = mlxsw_sp_rif_p_counter_get(rif, dir);
	if (!p_counter_index)
		return -EINVAL;
	err = mlxsw_sp_counter_alloc(mlxsw_sp, MLXSW_SP_COUNTER_SUB_POOL_RIF,
				     p_counter_index);
	if (err)
		return err;

	err = mlxsw_sp_rif_counter_clear(mlxsw_sp, *p_counter_index);
	if (err)
		goto err_counter_clear;

	err = mlxsw_sp_rif_counter_edit(mlxsw_sp, rif->rif_index,
					*p_counter_index, true, dir);
	if (err)
		goto err_counter_edit;
	mlxsw_sp_rif_counter_valid_set(rif, dir, true);
	return 0;

err_counter_edit:
err_counter_clear:
	mlxsw_sp_counter_free(mlxsw_sp, MLXSW_SP_COUNTER_SUB_POOL_RIF,
			      *p_counter_index);
	return err;
}

void mlxsw_sp_rif_counter_free(struct mlxsw_sp *mlxsw_sp,
			       struct mlxsw_sp_rif *rif,
			       enum mlxsw_sp_rif_counter_dir dir)
{
	unsigned int *p_counter_index;

264 265 266
	if (!mlxsw_sp_rif_counter_valid_get(rif, dir))
		return;

267 268 269 270 271 272 273 274 275 276
	p_counter_index = mlxsw_sp_rif_p_counter_get(rif, dir);
	if (WARN_ON(!p_counter_index))
		return;
	mlxsw_sp_rif_counter_edit(mlxsw_sp, rif->rif_index,
				  *p_counter_index, false, dir);
	mlxsw_sp_counter_free(mlxsw_sp, MLXSW_SP_COUNTER_SUB_POOL_RIF,
			      *p_counter_index);
	mlxsw_sp_rif_counter_valid_set(rif, dir, false);
}

277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295
static void mlxsw_sp_rif_counters_alloc(struct mlxsw_sp_rif *rif)
{
	struct mlxsw_sp *mlxsw_sp = rif->mlxsw_sp;
	struct devlink *devlink;

	devlink = priv_to_devlink(mlxsw_sp->core);
	if (!devlink_dpipe_table_counter_enabled(devlink,
						 MLXSW_SP_DPIPE_TABLE_NAME_ERIF))
		return;
	mlxsw_sp_rif_counter_alloc(mlxsw_sp, rif, MLXSW_SP_RIF_COUNTER_EGRESS);
}

static void mlxsw_sp_rif_counters_free(struct mlxsw_sp_rif *rif)
{
	struct mlxsw_sp *mlxsw_sp = rif->mlxsw_sp;

	mlxsw_sp_rif_counter_free(mlxsw_sp, rif, MLXSW_SP_RIF_COUNTER_EGRESS);
}

296
#define MLXSW_SP_PREFIX_COUNT (sizeof(struct in6_addr) * BITS_PER_BYTE + 1)
297 298 299 300 301

struct mlxsw_sp_prefix_usage {
	DECLARE_BITMAP(b, MLXSW_SP_PREFIX_COUNT);
};

302 303 304 305 306 307 308 309 310 311
#define mlxsw_sp_prefix_usage_for_each(prefix, prefix_usage) \
	for_each_set_bit(prefix, (prefix_usage)->b, MLXSW_SP_PREFIX_COUNT)

static bool
mlxsw_sp_prefix_usage_eq(struct mlxsw_sp_prefix_usage *prefix_usage1,
			 struct mlxsw_sp_prefix_usage *prefix_usage2)
{
	return !memcmp(prefix_usage1, prefix_usage2, sizeof(*prefix_usage1));
}

312 313 314 315 316 317 318
static void
mlxsw_sp_prefix_usage_cpy(struct mlxsw_sp_prefix_usage *prefix_usage1,
			  struct mlxsw_sp_prefix_usage *prefix_usage2)
{
	memcpy(prefix_usage1, prefix_usage2, sizeof(*prefix_usage1));
}

319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337
static void
mlxsw_sp_prefix_usage_set(struct mlxsw_sp_prefix_usage *prefix_usage,
			  unsigned char prefix_len)
{
	set_bit(prefix_len, prefix_usage->b);
}

static void
mlxsw_sp_prefix_usage_clear(struct mlxsw_sp_prefix_usage *prefix_usage,
			    unsigned char prefix_len)
{
	clear_bit(prefix_len, prefix_usage->b);
}

struct mlxsw_sp_fib_key {
	unsigned char addr[sizeof(struct in6_addr)];
	unsigned char prefix_len;
};

338 339 340 341
enum mlxsw_sp_fib_entry_type {
	MLXSW_SP_FIB_ENTRY_TYPE_REMOTE,
	MLXSW_SP_FIB_ENTRY_TYPE_LOCAL,
	MLXSW_SP_FIB_ENTRY_TYPE_TRAP,
342
	MLXSW_SP_FIB_ENTRY_TYPE_BLACKHOLE,
343
	MLXSW_SP_FIB_ENTRY_TYPE_UNREACHABLE,
344 345 346 347 348 349 350 351

	/* This is a special case of local delivery, where a packet should be
	 * decapsulated on reception. Note that there is no corresponding ENCAP,
	 * because that's a type of next hop, not of FIB entry. (There can be
	 * several next hops in a REMOTE entry, and some of them may be
	 * encapsulating entries.)
	 */
	MLXSW_SP_FIB_ENTRY_TYPE_IPIP_DECAP,
352
	MLXSW_SP_FIB_ENTRY_TYPE_NVE_DECAP,
353 354
};

355
struct mlxsw_sp_nexthop_group;
356
struct mlxsw_sp_fib_entry;
357

358
struct mlxsw_sp_fib_node {
359
	struct mlxsw_sp_fib_entry *fib_entry;
360
	struct list_head list;
361
	struct rhash_head ht_node;
362
	struct mlxsw_sp_fib *fib;
363
	struct mlxsw_sp_fib_key key;
364 365
};

366 367 368 369 370
struct mlxsw_sp_fib_entry_decap {
	struct mlxsw_sp_ipip_entry *ipip_entry;
	u32 tunnel_index;
};

371 372
struct mlxsw_sp_fib_entry {
	struct mlxsw_sp_fib_node *fib_node;
373
	enum mlxsw_sp_fib_entry_type type;
374 375
	struct list_head nexthop_group_node;
	struct mlxsw_sp_nexthop_group *nh_group;
376
	struct mlxsw_sp_fib_entry_decap decap; /* Valid for decap entries. */
377 378
};

379 380 381 382 383 384 385 386
struct mlxsw_sp_fib4_entry {
	struct mlxsw_sp_fib_entry common;
	u32 tb_id;
	u32 prio;
	u8 tos;
	u8 type;
};

387 388 389 390 391 392 393 394
struct mlxsw_sp_fib6_entry {
	struct mlxsw_sp_fib_entry common;
	struct list_head rt6_list;
	unsigned int nrt6;
};

struct mlxsw_sp_rt6 {
	struct list_head list;
395
	struct fib6_info *rt;
396 397
};

398 399 400 401
struct mlxsw_sp_lpm_tree {
	u8 id; /* tree ID */
	unsigned int ref_count;
	enum mlxsw_sp_l3proto proto;
402
	unsigned long prefix_ref_count[MLXSW_SP_PREFIX_COUNT];
403 404 405
	struct mlxsw_sp_prefix_usage prefix_usage;
};

406 407
struct mlxsw_sp_fib {
	struct rhashtable ht;
408
	struct list_head node_list;
409 410 411
	struct mlxsw_sp_vr *vr;
	struct mlxsw_sp_lpm_tree *lpm_tree;
	enum mlxsw_sp_l3proto proto;
412 413
};

414 415 416 417 418
struct mlxsw_sp_vr {
	u16 id; /* virtual router ID */
	u32 tb_id; /* kernel fib table id */
	unsigned int rif_count;
	struct mlxsw_sp_fib *fib4;
419
	struct mlxsw_sp_fib *fib6;
420
	struct mlxsw_sp_mr_table *mr_table[MLXSW_SP_L3_PROTO_MAX];
421 422
	struct mlxsw_sp_rif *ul_rif;
	refcount_t ul_rif_refcnt;
423 424
};

425
static const struct rhashtable_params mlxsw_sp_fib_ht_params;
426

427 428
static struct mlxsw_sp_fib *mlxsw_sp_fib_create(struct mlxsw_sp *mlxsw_sp,
						struct mlxsw_sp_vr *vr,
429
						enum mlxsw_sp_l3proto proto)
430
{
431
	struct mlxsw_sp_lpm_tree *lpm_tree;
432 433 434
	struct mlxsw_sp_fib *fib;
	int err;

435
	lpm_tree = mlxsw_sp->router->lpm.proto_trees[proto];
436 437 438 439 440 441
	fib = kzalloc(sizeof(*fib), GFP_KERNEL);
	if (!fib)
		return ERR_PTR(-ENOMEM);
	err = rhashtable_init(&fib->ht, &mlxsw_sp_fib_ht_params);
	if (err)
		goto err_rhashtable_init;
442
	INIT_LIST_HEAD(&fib->node_list);
443 444
	fib->proto = proto;
	fib->vr = vr;
445 446 447 448 449
	fib->lpm_tree = lpm_tree;
	mlxsw_sp_lpm_tree_hold(lpm_tree);
	err = mlxsw_sp_vr_lpm_tree_bind(mlxsw_sp, fib, lpm_tree->id);
	if (err)
		goto err_lpm_tree_bind;
450 451
	return fib;

452 453
err_lpm_tree_bind:
	mlxsw_sp_lpm_tree_put(mlxsw_sp, lpm_tree);
454 455 456 457 458
err_rhashtable_init:
	kfree(fib);
	return ERR_PTR(err);
}

459 460
static void mlxsw_sp_fib_destroy(struct mlxsw_sp *mlxsw_sp,
				 struct mlxsw_sp_fib *fib)
461
{
462 463
	mlxsw_sp_vr_lpm_tree_unbind(mlxsw_sp, fib);
	mlxsw_sp_lpm_tree_put(mlxsw_sp, fib->lpm_tree);
464
	WARN_ON(!list_empty(&fib->node_list));
465 466 467 468
	rhashtable_destroy(&fib->ht);
	kfree(fib);
}

469
static struct mlxsw_sp_lpm_tree *
470
mlxsw_sp_lpm_tree_find_unused(struct mlxsw_sp *mlxsw_sp)
471 472 473 474
{
	static struct mlxsw_sp_lpm_tree *lpm_tree;
	int i;

475 476
	for (i = 0; i < mlxsw_sp->router->lpm.tree_count; i++) {
		lpm_tree = &mlxsw_sp->router->lpm.trees[i];
477 478
		if (lpm_tree->ref_count == 0)
			return lpm_tree;
479 480 481 482 483 484 485 486 487
	}
	return NULL;
}

static int mlxsw_sp_lpm_tree_alloc(struct mlxsw_sp *mlxsw_sp,
				   struct mlxsw_sp_lpm_tree *lpm_tree)
{
	char ralta_pl[MLXSW_REG_RALTA_LEN];

488 489 490
	mlxsw_reg_ralta_pack(ralta_pl, true,
			     (enum mlxsw_reg_ralxx_protocol) lpm_tree->proto,
			     lpm_tree->id);
491 492 493
	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ralta), ralta_pl);
}

494 495
static void mlxsw_sp_lpm_tree_free(struct mlxsw_sp *mlxsw_sp,
				   struct mlxsw_sp_lpm_tree *lpm_tree)
496 497 498
{
	char ralta_pl[MLXSW_REG_RALTA_LEN];

499 500 501
	mlxsw_reg_ralta_pack(ralta_pl, false,
			     (enum mlxsw_reg_ralxx_protocol) lpm_tree->proto,
			     lpm_tree->id);
502
	mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ralta), ralta_pl);
503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531
}

static int
mlxsw_sp_lpm_tree_left_struct_set(struct mlxsw_sp *mlxsw_sp,
				  struct mlxsw_sp_prefix_usage *prefix_usage,
				  struct mlxsw_sp_lpm_tree *lpm_tree)
{
	char ralst_pl[MLXSW_REG_RALST_LEN];
	u8 root_bin = 0;
	u8 prefix;
	u8 last_prefix = MLXSW_REG_RALST_BIN_NO_CHILD;

	mlxsw_sp_prefix_usage_for_each(prefix, prefix_usage)
		root_bin = prefix;

	mlxsw_reg_ralst_pack(ralst_pl, root_bin, lpm_tree->id);
	mlxsw_sp_prefix_usage_for_each(prefix, prefix_usage) {
		if (prefix == 0)
			continue;
		mlxsw_reg_ralst_bin_pack(ralst_pl, prefix, last_prefix,
					 MLXSW_REG_RALST_BIN_NO_CHILD);
		last_prefix = prefix;
	}
	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ralst), ralst_pl);
}

static struct mlxsw_sp_lpm_tree *
mlxsw_sp_lpm_tree_create(struct mlxsw_sp *mlxsw_sp,
			 struct mlxsw_sp_prefix_usage *prefix_usage,
532
			 enum mlxsw_sp_l3proto proto)
533 534 535 536
{
	struct mlxsw_sp_lpm_tree *lpm_tree;
	int err;

537
	lpm_tree = mlxsw_sp_lpm_tree_find_unused(mlxsw_sp);
538 539 540 541 542 543 544 545 546 547 548
	if (!lpm_tree)
		return ERR_PTR(-EBUSY);
	lpm_tree->proto = proto;
	err = mlxsw_sp_lpm_tree_alloc(mlxsw_sp, lpm_tree);
	if (err)
		return ERR_PTR(err);

	err = mlxsw_sp_lpm_tree_left_struct_set(mlxsw_sp, prefix_usage,
						lpm_tree);
	if (err)
		goto err_left_struct_set;
549 550
	memcpy(&lpm_tree->prefix_usage, prefix_usage,
	       sizeof(lpm_tree->prefix_usage));
551 552 553
	memset(&lpm_tree->prefix_ref_count, 0,
	       sizeof(lpm_tree->prefix_ref_count));
	lpm_tree->ref_count = 1;
554 555 556 557 558 559 560
	return lpm_tree;

err_left_struct_set:
	mlxsw_sp_lpm_tree_free(mlxsw_sp, lpm_tree);
	return ERR_PTR(err);
}

561 562
static void mlxsw_sp_lpm_tree_destroy(struct mlxsw_sp *mlxsw_sp,
				      struct mlxsw_sp_lpm_tree *lpm_tree)
563
{
564
	mlxsw_sp_lpm_tree_free(mlxsw_sp, lpm_tree);
565 566 567 568 569
}

static struct mlxsw_sp_lpm_tree *
mlxsw_sp_lpm_tree_get(struct mlxsw_sp *mlxsw_sp,
		      struct mlxsw_sp_prefix_usage *prefix_usage,
570
		      enum mlxsw_sp_l3proto proto)
571 572 573 574
{
	struct mlxsw_sp_lpm_tree *lpm_tree;
	int i;

575 576
	for (i = 0; i < mlxsw_sp->router->lpm.tree_count; i++) {
		lpm_tree = &mlxsw_sp->router->lpm.trees[i];
577 578
		if (lpm_tree->ref_count != 0 &&
		    lpm_tree->proto == proto &&
579
		    mlxsw_sp_prefix_usage_eq(&lpm_tree->prefix_usage,
580 581
					     prefix_usage)) {
			mlxsw_sp_lpm_tree_hold(lpm_tree);
582
			return lpm_tree;
583
		}
584
	}
585 586
	return mlxsw_sp_lpm_tree_create(mlxsw_sp, prefix_usage, proto);
}
587

588 589
static void mlxsw_sp_lpm_tree_hold(struct mlxsw_sp_lpm_tree *lpm_tree)
{
590 591 592
	lpm_tree->ref_count++;
}

593 594
static void mlxsw_sp_lpm_tree_put(struct mlxsw_sp *mlxsw_sp,
				  struct mlxsw_sp_lpm_tree *lpm_tree)
595 596
{
	if (--lpm_tree->ref_count == 0)
597
		mlxsw_sp_lpm_tree_destroy(mlxsw_sp, lpm_tree);
598 599
}

600
#define MLXSW_SP_LPM_TREE_MIN 1 /* tree 0 is reserved */
601 602

static int mlxsw_sp_lpm_init(struct mlxsw_sp *mlxsw_sp)
603
{
604
	struct mlxsw_sp_prefix_usage req_prefix_usage = {{ 0 } };
605
	struct mlxsw_sp_lpm_tree *lpm_tree;
606
	u64 max_trees;
607
	int err, i;
608

609 610 611 612
	if (!MLXSW_CORE_RES_VALID(mlxsw_sp->core, MAX_LPM_TREES))
		return -EIO;

	max_trees = MLXSW_CORE_RES_GET(mlxsw_sp->core, MAX_LPM_TREES);
613 614
	mlxsw_sp->router->lpm.tree_count = max_trees - MLXSW_SP_LPM_TREE_MIN;
	mlxsw_sp->router->lpm.trees = kcalloc(mlxsw_sp->router->lpm.tree_count,
615 616
					     sizeof(struct mlxsw_sp_lpm_tree),
					     GFP_KERNEL);
617
	if (!mlxsw_sp->router->lpm.trees)
618 619
		return -ENOMEM;

620 621
	for (i = 0; i < mlxsw_sp->router->lpm.tree_count; i++) {
		lpm_tree = &mlxsw_sp->router->lpm.trees[i];
622 623
		lpm_tree->id = i + MLXSW_SP_LPM_TREE_MIN;
	}
624

625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640
	lpm_tree = mlxsw_sp_lpm_tree_get(mlxsw_sp, &req_prefix_usage,
					 MLXSW_SP_L3_PROTO_IPV4);
	if (IS_ERR(lpm_tree)) {
		err = PTR_ERR(lpm_tree);
		goto err_ipv4_tree_get;
	}
	mlxsw_sp->router->lpm.proto_trees[MLXSW_SP_L3_PROTO_IPV4] = lpm_tree;

	lpm_tree = mlxsw_sp_lpm_tree_get(mlxsw_sp, &req_prefix_usage,
					 MLXSW_SP_L3_PROTO_IPV6);
	if (IS_ERR(lpm_tree)) {
		err = PTR_ERR(lpm_tree);
		goto err_ipv6_tree_get;
	}
	mlxsw_sp->router->lpm.proto_trees[MLXSW_SP_L3_PROTO_IPV6] = lpm_tree;

641
	return 0;
642 643 644 645 646 647 648

err_ipv6_tree_get:
	lpm_tree = mlxsw_sp->router->lpm.proto_trees[MLXSW_SP_L3_PROTO_IPV4];
	mlxsw_sp_lpm_tree_put(mlxsw_sp, lpm_tree);
err_ipv4_tree_get:
	kfree(mlxsw_sp->router->lpm.trees);
	return err;
649 650 651 652
}

static void mlxsw_sp_lpm_fini(struct mlxsw_sp *mlxsw_sp)
{
653 654 655 656 657 658 659 660
	struct mlxsw_sp_lpm_tree *lpm_tree;

	lpm_tree = mlxsw_sp->router->lpm.proto_trees[MLXSW_SP_L3_PROTO_IPV6];
	mlxsw_sp_lpm_tree_put(mlxsw_sp, lpm_tree);

	lpm_tree = mlxsw_sp->router->lpm.proto_trees[MLXSW_SP_L3_PROTO_IPV4];
	mlxsw_sp_lpm_tree_put(mlxsw_sp, lpm_tree);

661
	kfree(mlxsw_sp->router->lpm.trees);
662 663
}

664 665
static bool mlxsw_sp_vr_is_used(const struct mlxsw_sp_vr *vr)
{
666 667 668
	return !!vr->fib4 || !!vr->fib6 ||
	       !!vr->mr_table[MLXSW_SP_L3_PROTO_IPV4] ||
	       !!vr->mr_table[MLXSW_SP_L3_PROTO_IPV6];
669 670
}

671 672 673 674 675
static struct mlxsw_sp_vr *mlxsw_sp_vr_find_unused(struct mlxsw_sp *mlxsw_sp)
{
	struct mlxsw_sp_vr *vr;
	int i;

J
Jiri Pirko 已提交
676
	for (i = 0; i < MLXSW_CORE_RES_GET(mlxsw_sp->core, MAX_VRS); i++) {
677
		vr = &mlxsw_sp->router->vrs[i];
678
		if (!mlxsw_sp_vr_is_used(vr))
679 680 681 682 683 684
			return vr;
	}
	return NULL;
}

static int mlxsw_sp_vr_lpm_tree_bind(struct mlxsw_sp *mlxsw_sp,
685
				     const struct mlxsw_sp_fib *fib, u8 tree_id)
686 687 688
{
	char raltb_pl[MLXSW_REG_RALTB_LEN];

689 690
	mlxsw_reg_raltb_pack(raltb_pl, fib->vr->id,
			     (enum mlxsw_reg_ralxx_protocol) fib->proto,
691
			     tree_id);
692 693 694 695
	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(raltb), raltb_pl);
}

static int mlxsw_sp_vr_lpm_tree_unbind(struct mlxsw_sp *mlxsw_sp,
696
				       const struct mlxsw_sp_fib *fib)
697 698 699 700
{
	char raltb_pl[MLXSW_REG_RALTB_LEN];

	/* Bind to tree 0 which is default */
701 702
	mlxsw_reg_raltb_pack(raltb_pl, fib->vr->id,
			     (enum mlxsw_reg_ralxx_protocol) fib->proto, 0);
703 704 705 706 707
	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(raltb), raltb_pl);
}

static u32 mlxsw_sp_fix_tb_id(u32 tb_id)
{
708 709
	/* For our purpose, squash main, default and local tables into one */
	if (tb_id == RT_TABLE_LOCAL || tb_id == RT_TABLE_DEFAULT)
710 711 712 713 714
		tb_id = RT_TABLE_MAIN;
	return tb_id;
}

static struct mlxsw_sp_vr *mlxsw_sp_vr_find(struct mlxsw_sp *mlxsw_sp,
715
					    u32 tb_id)
716 717 718 719 720
{
	struct mlxsw_sp_vr *vr;
	int i;

	tb_id = mlxsw_sp_fix_tb_id(tb_id);
721

J
Jiri Pirko 已提交
722
	for (i = 0; i < MLXSW_CORE_RES_GET(mlxsw_sp->core, MAX_VRS); i++) {
723
		vr = &mlxsw_sp->router->vrs[i];
724
		if (mlxsw_sp_vr_is_used(vr) && vr->tb_id == tb_id)
725 726 727 728 729
			return vr;
	}
	return NULL;
}

730 731 732 733 734 735 736 737 738 739 740 741 742
int mlxsw_sp_router_tb_id_vr_id(struct mlxsw_sp *mlxsw_sp, u32 tb_id,
				u16 *vr_id)
{
	struct mlxsw_sp_vr *vr;

	vr = mlxsw_sp_vr_find(mlxsw_sp, tb_id);
	if (!vr)
		return -ESRCH;
	*vr_id = vr->id;

	return 0;
}

743 744 745 746 747 748 749
static struct mlxsw_sp_fib *mlxsw_sp_vr_fib(const struct mlxsw_sp_vr *vr,
					    enum mlxsw_sp_l3proto proto)
{
	switch (proto) {
	case MLXSW_SP_L3_PROTO_IPV4:
		return vr->fib4;
	case MLXSW_SP_L3_PROTO_IPV6:
750
		return vr->fib6;
751 752 753 754
	}
	return NULL;
}

755
static struct mlxsw_sp_vr *mlxsw_sp_vr_create(struct mlxsw_sp *mlxsw_sp,
756 757
					      u32 tb_id,
					      struct netlink_ext_ack *extack)
758
{
759
	struct mlxsw_sp_mr_table *mr4_table, *mr6_table;
760 761
	struct mlxsw_sp_fib *fib4;
	struct mlxsw_sp_fib *fib6;
762
	struct mlxsw_sp_vr *vr;
763
	int err;
764 765

	vr = mlxsw_sp_vr_find_unused(mlxsw_sp);
766
	if (!vr) {
767
		NL_SET_ERR_MSG_MOD(extack, "Exceeded number of supported virtual routers");
768
		return ERR_PTR(-EBUSY);
769
	}
770 771 772 773 774 775
	fib4 = mlxsw_sp_fib_create(mlxsw_sp, vr, MLXSW_SP_L3_PROTO_IPV4);
	if (IS_ERR(fib4))
		return ERR_CAST(fib4);
	fib6 = mlxsw_sp_fib_create(mlxsw_sp, vr, MLXSW_SP_L3_PROTO_IPV6);
	if (IS_ERR(fib6)) {
		err = PTR_ERR(fib6);
776 777
		goto err_fib6_create;
	}
778 779 780 781
	mr4_table = mlxsw_sp_mr_table_create(mlxsw_sp, vr->id,
					     MLXSW_SP_L3_PROTO_IPV4);
	if (IS_ERR(mr4_table)) {
		err = PTR_ERR(mr4_table);
782
		goto err_mr4_table_create;
783
	}
784 785 786 787 788 789 790
	mr6_table = mlxsw_sp_mr_table_create(mlxsw_sp, vr->id,
					     MLXSW_SP_L3_PROTO_IPV6);
	if (IS_ERR(mr6_table)) {
		err = PTR_ERR(mr6_table);
		goto err_mr6_table_create;
	}

791 792
	vr->fib4 = fib4;
	vr->fib6 = fib6;
793 794
	vr->mr_table[MLXSW_SP_L3_PROTO_IPV4] = mr4_table;
	vr->mr_table[MLXSW_SP_L3_PROTO_IPV6] = mr6_table;
795 796
	vr->tb_id = tb_id;
	return vr;
797

798 799 800
err_mr6_table_create:
	mlxsw_sp_mr_table_destroy(mr4_table);
err_mr4_table_create:
801
	mlxsw_sp_fib_destroy(mlxsw_sp, fib6);
802
err_fib6_create:
803
	mlxsw_sp_fib_destroy(mlxsw_sp, fib4);
804
	return ERR_PTR(err);
805 806
}

807 808
static void mlxsw_sp_vr_destroy(struct mlxsw_sp *mlxsw_sp,
				struct mlxsw_sp_vr *vr)
809
{
810 811 812 813
	mlxsw_sp_mr_table_destroy(vr->mr_table[MLXSW_SP_L3_PROTO_IPV6]);
	vr->mr_table[MLXSW_SP_L3_PROTO_IPV6] = NULL;
	mlxsw_sp_mr_table_destroy(vr->mr_table[MLXSW_SP_L3_PROTO_IPV4]);
	vr->mr_table[MLXSW_SP_L3_PROTO_IPV4] = NULL;
814
	mlxsw_sp_fib_destroy(mlxsw_sp, vr->fib6);
815
	vr->fib6 = NULL;
816
	mlxsw_sp_fib_destroy(mlxsw_sp, vr->fib4);
817
	vr->fib4 = NULL;
818 819
}

820 821
static struct mlxsw_sp_vr *mlxsw_sp_vr_get(struct mlxsw_sp *mlxsw_sp, u32 tb_id,
					   struct netlink_ext_ack *extack)
822 823 824 825
{
	struct mlxsw_sp_vr *vr;

	tb_id = mlxsw_sp_fix_tb_id(tb_id);
826 827
	vr = mlxsw_sp_vr_find(mlxsw_sp, tb_id);
	if (!vr)
828
		vr = mlxsw_sp_vr_create(mlxsw_sp, tb_id, extack);
829 830 831
	return vr;
}

832
static void mlxsw_sp_vr_put(struct mlxsw_sp *mlxsw_sp, struct mlxsw_sp_vr *vr)
833
{
834
	if (!vr->rif_count && list_empty(&vr->fib4->node_list) &&
835
	    list_empty(&vr->fib6->node_list) &&
836 837
	    mlxsw_sp_mr_table_empty(vr->mr_table[MLXSW_SP_L3_PROTO_IPV4]) &&
	    mlxsw_sp_mr_table_empty(vr->mr_table[MLXSW_SP_L3_PROTO_IPV6]))
838
		mlxsw_sp_vr_destroy(mlxsw_sp, vr);
839 840
}

841 842 843 844 845 846 847 848
static bool
mlxsw_sp_vr_lpm_tree_should_replace(struct mlxsw_sp_vr *vr,
				    enum mlxsw_sp_l3proto proto, u8 tree_id)
{
	struct mlxsw_sp_fib *fib = mlxsw_sp_vr_fib(vr, proto);

	if (!mlxsw_sp_vr_is_used(vr))
		return false;
849
	if (fib->lpm_tree->id == tree_id)
850 851 852 853 854 855 856 857 858 859 860 861 862
		return true;
	return false;
}

static int mlxsw_sp_vr_lpm_tree_replace(struct mlxsw_sp *mlxsw_sp,
					struct mlxsw_sp_fib *fib,
					struct mlxsw_sp_lpm_tree *new_tree)
{
	struct mlxsw_sp_lpm_tree *old_tree = fib->lpm_tree;
	int err;

	fib->lpm_tree = new_tree;
	mlxsw_sp_lpm_tree_hold(new_tree);
863 864 865
	err = mlxsw_sp_vr_lpm_tree_bind(mlxsw_sp, fib, new_tree->id);
	if (err)
		goto err_tree_bind;
866 867
	mlxsw_sp_lpm_tree_put(mlxsw_sp, old_tree);
	return 0;
868 869 870 871 872

err_tree_bind:
	mlxsw_sp_lpm_tree_put(mlxsw_sp, new_tree);
	fib->lpm_tree = old_tree;
	return err;
873 874 875 876 877 878 879
}

static int mlxsw_sp_vrs_lpm_tree_replace(struct mlxsw_sp *mlxsw_sp,
					 struct mlxsw_sp_fib *fib,
					 struct mlxsw_sp_lpm_tree *new_tree)
{
	enum mlxsw_sp_l3proto proto = fib->proto;
880
	struct mlxsw_sp_lpm_tree *old_tree;
881 882 883 884
	u8 old_id, new_id = new_tree->id;
	struct mlxsw_sp_vr *vr;
	int i, err;

885
	old_tree = mlxsw_sp->router->lpm.proto_trees[proto];
886 887 888 889 890 891 892 893 894 895 896 897 898
	old_id = old_tree->id;

	for (i = 0; i < MLXSW_CORE_RES_GET(mlxsw_sp->core, MAX_VRS); i++) {
		vr = &mlxsw_sp->router->vrs[i];
		if (!mlxsw_sp_vr_lpm_tree_should_replace(vr, proto, old_id))
			continue;
		err = mlxsw_sp_vr_lpm_tree_replace(mlxsw_sp,
						   mlxsw_sp_vr_fib(vr, proto),
						   new_tree);
		if (err)
			goto err_tree_replace;
	}

899 900 901 902 903
	memcpy(new_tree->prefix_ref_count, old_tree->prefix_ref_count,
	       sizeof(new_tree->prefix_ref_count));
	mlxsw_sp->router->lpm.proto_trees[proto] = new_tree;
	mlxsw_sp_lpm_tree_put(mlxsw_sp, old_tree);

904 905 906 907 908 909 910 911 912 913 914 915 916
	return 0;

err_tree_replace:
	for (i--; i >= 0; i--) {
		if (!mlxsw_sp_vr_lpm_tree_should_replace(vr, proto, new_id))
			continue;
		mlxsw_sp_vr_lpm_tree_replace(mlxsw_sp,
					     mlxsw_sp_vr_fib(vr, proto),
					     old_tree);
	}
	return err;
}

917
static int mlxsw_sp_vrs_init(struct mlxsw_sp *mlxsw_sp)
918 919
{
	struct mlxsw_sp_vr *vr;
J
Jiri Pirko 已提交
920
	u64 max_vrs;
921 922
	int i;

J
Jiri Pirko 已提交
923
	if (!MLXSW_CORE_RES_VALID(mlxsw_sp->core, MAX_VRS))
924 925
		return -EIO;

J
Jiri Pirko 已提交
926
	max_vrs = MLXSW_CORE_RES_GET(mlxsw_sp->core, MAX_VRS);
927 928 929
	mlxsw_sp->router->vrs = kcalloc(max_vrs, sizeof(struct mlxsw_sp_vr),
					GFP_KERNEL);
	if (!mlxsw_sp->router->vrs)
930 931
		return -ENOMEM;

J
Jiri Pirko 已提交
932
	for (i = 0; i < max_vrs; i++) {
933
		vr = &mlxsw_sp->router->vrs[i];
934 935
		vr->id = i;
	}
936 937 938 939

	return 0;
}

940 941
static void mlxsw_sp_router_fib_flush(struct mlxsw_sp *mlxsw_sp);

942 943
static void mlxsw_sp_vrs_fini(struct mlxsw_sp *mlxsw_sp)
{
944 945 946 947 948 949 950 951
	/* At this stage we're guaranteed not to have new incoming
	 * FIB notifications and the work queue is free from FIBs
	 * sitting on top of mlxsw netdevs. However, we can still
	 * have other FIBs queued. Flush the queue before flushing
	 * the device's tables. No need for locks, as we're the only
	 * writer.
	 */
	mlxsw_core_flush_owq();
952
	mlxsw_sp_router_fib_flush(mlxsw_sp);
953
	kfree(mlxsw_sp->router->vrs);
954 955
}

956 957 958 959 960 961
static struct net_device *
__mlxsw_sp_ipip_netdev_ul_dev_get(const struct net_device *ol_dev)
{
	struct ip_tunnel *tun = netdev_priv(ol_dev);
	struct net *net = dev_net(ol_dev);

962
	return dev_get_by_index_rcu(net, tun->parms.link);
963 964
}

965
u32 mlxsw_sp_ipip_dev_ul_tb_id(const struct net_device *ol_dev)
966
{
967 968
	struct net_device *d;
	u32 tb_id;
969

970 971
	rcu_read_lock();
	d = __mlxsw_sp_ipip_netdev_ul_dev_get(ol_dev);
972
	if (d)
973
		tb_id = l3mdev_fib_table(d) ? : RT_TABLE_MAIN;
974
	else
975 976 977 978
		tb_id = RT_TABLE_MAIN;
	rcu_read_unlock();

	return tb_id;
979 980
}

981 982
static struct mlxsw_sp_rif *
mlxsw_sp_rif_create(struct mlxsw_sp *mlxsw_sp,
983 984
		    const struct mlxsw_sp_rif_params *params,
		    struct netlink_ext_ack *extack);
985 986 987 988

static struct mlxsw_sp_rif_ipip_lb *
mlxsw_sp_ipip_ol_ipip_lb_create(struct mlxsw_sp *mlxsw_sp,
				enum mlxsw_sp_ipip_type ipipt,
989 990
				struct net_device *ol_dev,
				struct netlink_ext_ack *extack)
991 992 993 994 995 996 997 998 999 1000 1001 1002
{
	struct mlxsw_sp_rif_params_ipip_lb lb_params;
	const struct mlxsw_sp_ipip_ops *ipip_ops;
	struct mlxsw_sp_rif *rif;

	ipip_ops = mlxsw_sp->router->ipip_ops_arr[ipipt];
	lb_params = (struct mlxsw_sp_rif_params_ipip_lb) {
		.common.dev = ol_dev,
		.common.lag = false,
		.lb_config = ipip_ops->ol_loopback_config(mlxsw_sp, ol_dev),
	};

1003
	rif = mlxsw_sp_rif_create(mlxsw_sp, &lb_params.common, extack);
1004 1005 1006 1007 1008 1009 1010 1011 1012 1013
	if (IS_ERR(rif))
		return ERR_CAST(rif);
	return container_of(rif, struct mlxsw_sp_rif_ipip_lb, common);
}

static struct mlxsw_sp_ipip_entry *
mlxsw_sp_ipip_entry_alloc(struct mlxsw_sp *mlxsw_sp,
			  enum mlxsw_sp_ipip_type ipipt,
			  struct net_device *ol_dev)
{
1014
	const struct mlxsw_sp_ipip_ops *ipip_ops;
1015 1016 1017
	struct mlxsw_sp_ipip_entry *ipip_entry;
	struct mlxsw_sp_ipip_entry *ret = NULL;

1018
	ipip_ops = mlxsw_sp->router->ipip_ops_arr[ipipt];
1019 1020 1021 1022 1023
	ipip_entry = kzalloc(sizeof(*ipip_entry), GFP_KERNEL);
	if (!ipip_entry)
		return ERR_PTR(-ENOMEM);

	ipip_entry->ol_lb = mlxsw_sp_ipip_ol_ipip_lb_create(mlxsw_sp, ipipt,
1024
							    ol_dev, NULL);
1025 1026 1027 1028 1029 1030 1031
	if (IS_ERR(ipip_entry->ol_lb)) {
		ret = ERR_CAST(ipip_entry->ol_lb);
		goto err_ol_ipip_lb_create;
	}

	ipip_entry->ipipt = ipipt;
	ipip_entry->ol_dev = ol_dev;
1032 1033 1034 1035 1036 1037 1038 1039 1040

	switch (ipip_ops->ul_proto) {
	case MLXSW_SP_L3_PROTO_IPV4:
		ipip_entry->parms4 = mlxsw_sp_ipip_netdev_parms4(ol_dev);
		break;
	case MLXSW_SP_L3_PROTO_IPV6:
		WARN_ON(1);
		break;
	}
1041 1042 1043 1044 1045 1046 1047 1048 1049

	return ipip_entry;

err_ol_ipip_lb_create:
	kfree(ipip_entry);
	return ret;
}

static void
1050
mlxsw_sp_ipip_entry_dealloc(struct mlxsw_sp_ipip_entry *ipip_entry)
1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074
{
	mlxsw_sp_rif_destroy(&ipip_entry->ol_lb->common);
	kfree(ipip_entry);
}

static bool
mlxsw_sp_ipip_entry_saddr_matches(struct mlxsw_sp *mlxsw_sp,
				  const enum mlxsw_sp_l3proto ul_proto,
				  union mlxsw_sp_l3addr saddr,
				  u32 ul_tb_id,
				  struct mlxsw_sp_ipip_entry *ipip_entry)
{
	u32 tun_ul_tb_id = mlxsw_sp_ipip_dev_ul_tb_id(ipip_entry->ol_dev);
	enum mlxsw_sp_ipip_type ipipt = ipip_entry->ipipt;
	union mlxsw_sp_l3addr tun_saddr;

	if (mlxsw_sp->router->ipip_ops_arr[ipipt]->ul_proto != ul_proto)
		return false;

	tun_saddr = mlxsw_sp_ipip_netdev_saddr(ul_proto, ipip_entry->ol_dev);
	return tun_ul_tb_id == ul_tb_id &&
	       mlxsw_sp_l3addr_eq(&tun_saddr, &saddr);
}

1075 1076 1077 1078 1079 1080 1081 1082
static int
mlxsw_sp_fib_entry_decap_init(struct mlxsw_sp *mlxsw_sp,
			      struct mlxsw_sp_fib_entry *fib_entry,
			      struct mlxsw_sp_ipip_entry *ipip_entry)
{
	u32 tunnel_index;
	int err;

1083 1084
	err = mlxsw_sp_kvdl_alloc(mlxsw_sp, MLXSW_SP_KVDL_ENTRY_TYPE_ADJ,
				  1, &tunnel_index);
1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099
	if (err)
		return err;

	ipip_entry->decap_fib_entry = fib_entry;
	fib_entry->decap.ipip_entry = ipip_entry;
	fib_entry->decap.tunnel_index = tunnel_index;
	return 0;
}

static void mlxsw_sp_fib_entry_decap_fini(struct mlxsw_sp *mlxsw_sp,
					  struct mlxsw_sp_fib_entry *fib_entry)
{
	/* Unlink this node from the IPIP entry that it's the decap entry of. */
	fib_entry->decap.ipip_entry->decap_fib_entry = NULL;
	fib_entry->decap.ipip_entry = NULL;
1100
	mlxsw_sp_kvdl_free(mlxsw_sp, MLXSW_SP_KVDL_ENTRY_TYPE_ADJ,
1101
			   1, fib_entry->decap.tunnel_index);
1102 1103
}

1104 1105 1106
static struct mlxsw_sp_fib_node *
mlxsw_sp_fib_node_lookup(struct mlxsw_sp_fib *fib, const void *addr,
			 size_t addr_len, unsigned char prefix_len);
1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121
static int mlxsw_sp_fib_entry_update(struct mlxsw_sp *mlxsw_sp,
				     struct mlxsw_sp_fib_entry *fib_entry);

static void
mlxsw_sp_ipip_entry_demote_decap(struct mlxsw_sp *mlxsw_sp,
				 struct mlxsw_sp_ipip_entry *ipip_entry)
{
	struct mlxsw_sp_fib_entry *fib_entry = ipip_entry->decap_fib_entry;

	mlxsw_sp_fib_entry_decap_fini(mlxsw_sp, fib_entry);
	fib_entry->type = MLXSW_SP_FIB_ENTRY_TYPE_TRAP;

	mlxsw_sp_fib_entry_update(mlxsw_sp, fib_entry);
}

1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135
static void
mlxsw_sp_ipip_entry_promote_decap(struct mlxsw_sp *mlxsw_sp,
				  struct mlxsw_sp_ipip_entry *ipip_entry,
				  struct mlxsw_sp_fib_entry *decap_fib_entry)
{
	if (mlxsw_sp_fib_entry_decap_init(mlxsw_sp, decap_fib_entry,
					  ipip_entry))
		return;
	decap_fib_entry->type = MLXSW_SP_FIB_ENTRY_TYPE_IPIP_DECAP;

	if (mlxsw_sp_fib_entry_update(mlxsw_sp, decap_fib_entry))
		mlxsw_sp_ipip_entry_demote_decap(mlxsw_sp, ipip_entry);
}

1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169
static struct mlxsw_sp_fib_entry *
mlxsw_sp_router_ip2me_fib_entry_find(struct mlxsw_sp *mlxsw_sp, u32 tb_id,
				     enum mlxsw_sp_l3proto proto,
				     const union mlxsw_sp_l3addr *addr,
				     enum mlxsw_sp_fib_entry_type type)
{
	struct mlxsw_sp_fib_node *fib_node;
	unsigned char addr_prefix_len;
	struct mlxsw_sp_fib *fib;
	struct mlxsw_sp_vr *vr;
	const void *addrp;
	size_t addr_len;
	u32 addr4;

	vr = mlxsw_sp_vr_find(mlxsw_sp, tb_id);
	if (!vr)
		return NULL;
	fib = mlxsw_sp_vr_fib(vr, proto);

	switch (proto) {
	case MLXSW_SP_L3_PROTO_IPV4:
		addr4 = be32_to_cpu(addr->addr4);
		addrp = &addr4;
		addr_len = 4;
		addr_prefix_len = 32;
		break;
	case MLXSW_SP_L3_PROTO_IPV6: /* fall through */
	default:
		WARN_ON(1);
		return NULL;
	}

	fib_node = mlxsw_sp_fib_node_lookup(fib, addrp, addr_len,
					    addr_prefix_len);
1170
	if (!fib_node || fib_node->fib_entry->type != type)
1171 1172
		return NULL;

1173
	return fib_node->fib_entry;
1174 1175
}

1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216
/* Given an IPIP entry, find the corresponding decap route. */
static struct mlxsw_sp_fib_entry *
mlxsw_sp_ipip_entry_find_decap(struct mlxsw_sp *mlxsw_sp,
			       struct mlxsw_sp_ipip_entry *ipip_entry)
{
	static struct mlxsw_sp_fib_node *fib_node;
	const struct mlxsw_sp_ipip_ops *ipip_ops;
	unsigned char saddr_prefix_len;
	union mlxsw_sp_l3addr saddr;
	struct mlxsw_sp_fib *ul_fib;
	struct mlxsw_sp_vr *ul_vr;
	const void *saddrp;
	size_t saddr_len;
	u32 ul_tb_id;
	u32 saddr4;

	ipip_ops = mlxsw_sp->router->ipip_ops_arr[ipip_entry->ipipt];

	ul_tb_id = mlxsw_sp_ipip_dev_ul_tb_id(ipip_entry->ol_dev);
	ul_vr = mlxsw_sp_vr_find(mlxsw_sp, ul_tb_id);
	if (!ul_vr)
		return NULL;

	ul_fib = mlxsw_sp_vr_fib(ul_vr, ipip_ops->ul_proto);
	saddr = mlxsw_sp_ipip_netdev_saddr(ipip_ops->ul_proto,
					   ipip_entry->ol_dev);

	switch (ipip_ops->ul_proto) {
	case MLXSW_SP_L3_PROTO_IPV4:
		saddr4 = be32_to_cpu(saddr.addr4);
		saddrp = &saddr4;
		saddr_len = 4;
		saddr_prefix_len = 32;
		break;
	case MLXSW_SP_L3_PROTO_IPV6:
		WARN_ON(1);
		return NULL;
	}

	fib_node = mlxsw_sp_fib_node_lookup(ul_fib, saddrp, saddr_len,
					    saddr_prefix_len);
1217 1218
	if (!fib_node ||
	    fib_node->fib_entry->type != MLXSW_SP_FIB_ENTRY_TYPE_TRAP)
1219 1220
		return NULL;

1221
	return fib_node->fib_entry;
1222 1223
}

1224
static struct mlxsw_sp_ipip_entry *
1225 1226 1227
mlxsw_sp_ipip_entry_create(struct mlxsw_sp *mlxsw_sp,
			   enum mlxsw_sp_ipip_type ipipt,
			   struct net_device *ol_dev)
1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241
{
	struct mlxsw_sp_ipip_entry *ipip_entry;

	ipip_entry = mlxsw_sp_ipip_entry_alloc(mlxsw_sp, ipipt, ol_dev);
	if (IS_ERR(ipip_entry))
		return ipip_entry;

	list_add_tail(&ipip_entry->ipip_list_node,
		      &mlxsw_sp->router->ipip_list);

	return ipip_entry;
}

static void
1242 1243
mlxsw_sp_ipip_entry_destroy(struct mlxsw_sp *mlxsw_sp,
			    struct mlxsw_sp_ipip_entry *ipip_entry)
1244
{
1245 1246
	list_del(&ipip_entry->ipip_list_node);
	mlxsw_sp_ipip_entry_dealloc(ipip_entry);
1247 1248
}

1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262
static bool
mlxsw_sp_ipip_entry_matches_decap(struct mlxsw_sp *mlxsw_sp,
				  const struct net_device *ul_dev,
				  enum mlxsw_sp_l3proto ul_proto,
				  union mlxsw_sp_l3addr ul_dip,
				  struct mlxsw_sp_ipip_entry *ipip_entry)
{
	u32 ul_tb_id = l3mdev_fib_table(ul_dev) ? : RT_TABLE_MAIN;
	enum mlxsw_sp_ipip_type ipipt = ipip_entry->ipipt;

	if (mlxsw_sp->router->ipip_ops_arr[ipipt]->ul_proto != ul_proto)
		return false;

	return mlxsw_sp_ipip_entry_saddr_matches(mlxsw_sp, ul_proto, ul_dip,
1263
						 ul_tb_id, ipip_entry);
1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284
}

/* Given decap parameters, find the corresponding IPIP entry. */
static struct mlxsw_sp_ipip_entry *
mlxsw_sp_ipip_entry_find_by_decap(struct mlxsw_sp *mlxsw_sp,
				  const struct net_device *ul_dev,
				  enum mlxsw_sp_l3proto ul_proto,
				  union mlxsw_sp_l3addr ul_dip)
{
	struct mlxsw_sp_ipip_entry *ipip_entry;

	list_for_each_entry(ipip_entry, &mlxsw_sp->router->ipip_list,
			    ipip_list_node)
		if (mlxsw_sp_ipip_entry_matches_decap(mlxsw_sp, ul_dev,
						      ul_proto, ul_dip,
						      ipip_entry))
			return ipip_entry;

	return NULL;
}

1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303
static bool mlxsw_sp_netdev_ipip_type(const struct mlxsw_sp *mlxsw_sp,
				      const struct net_device *dev,
				      enum mlxsw_sp_ipip_type *p_type)
{
	struct mlxsw_sp_router *router = mlxsw_sp->router;
	const struct mlxsw_sp_ipip_ops *ipip_ops;
	enum mlxsw_sp_ipip_type ipipt;

	for (ipipt = 0; ipipt < MLXSW_SP_IPIP_TYPE_MAX; ++ipipt) {
		ipip_ops = router->ipip_ops_arr[ipipt];
		if (dev->type == ipip_ops->dev_type) {
			if (p_type)
				*p_type = ipipt;
			return true;
		}
	}
	return false;
}

1304 1305
bool mlxsw_sp_netdev_is_ipip_ol(const struct mlxsw_sp *mlxsw_sp,
				const struct net_device *dev)
1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323
{
	return mlxsw_sp_netdev_ipip_type(mlxsw_sp, dev, NULL);
}

static struct mlxsw_sp_ipip_entry *
mlxsw_sp_ipip_entry_find_by_ol_dev(struct mlxsw_sp *mlxsw_sp,
				   const struct net_device *ol_dev)
{
	struct mlxsw_sp_ipip_entry *ipip_entry;

	list_for_each_entry(ipip_entry, &mlxsw_sp->router->ipip_list,
			    ipip_list_node)
		if (ipip_entry->ol_dev == ol_dev)
			return ipip_entry;

	return NULL;
}

1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334
static struct mlxsw_sp_ipip_entry *
mlxsw_sp_ipip_entry_find_by_ul_dev(const struct mlxsw_sp *mlxsw_sp,
				   const struct net_device *ul_dev,
				   struct mlxsw_sp_ipip_entry *start)
{
	struct mlxsw_sp_ipip_entry *ipip_entry;

	ipip_entry = list_prepare_entry(start, &mlxsw_sp->router->ipip_list,
					ipip_list_node);
	list_for_each_entry_continue(ipip_entry, &mlxsw_sp->router->ipip_list,
				     ipip_list_node) {
1335 1336 1337 1338 1339 1340
		struct net_device *ol_dev = ipip_entry->ol_dev;
		struct net_device *ipip_ul_dev;

		rcu_read_lock();
		ipip_ul_dev = __mlxsw_sp_ipip_netdev_ul_dev_get(ol_dev);
		rcu_read_unlock();
1341 1342 1343 1344 1345 1346 1347 1348

		if (ipip_ul_dev == ul_dev)
			return ipip_entry;
	}

	return NULL;
}

1349
bool mlxsw_sp_netdev_is_ipip_ul(struct mlxsw_sp *mlxsw_sp,
1350 1351
				const struct net_device *dev)
{
1352 1353 1354 1355 1356 1357 1358
	bool is_ipip_ul;

	mutex_lock(&mlxsw_sp->router->lock);
	is_ipip_ul = mlxsw_sp_ipip_entry_find_by_ul_dev(mlxsw_sp, dev, NULL);
	mutex_unlock(&mlxsw_sp->router->lock);

	return is_ipip_ul;
1359 1360
}

1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374
static bool mlxsw_sp_netdevice_ipip_can_offload(struct mlxsw_sp *mlxsw_sp,
						const struct net_device *ol_dev,
						enum mlxsw_sp_ipip_type ipipt)
{
	const struct mlxsw_sp_ipip_ops *ops
		= mlxsw_sp->router->ipip_ops_arr[ipipt];

	/* For deciding whether decap should be offloaded, we don't care about
	 * overlay protocol, so ask whether either one is supported.
	 */
	return ops->can_offload(mlxsw_sp, ol_dev, MLXSW_SP_L3_PROTO_IPV4) ||
	       ops->can_offload(mlxsw_sp, ol_dev, MLXSW_SP_L3_PROTO_IPV6);
}

1375 1376
static int mlxsw_sp_netdevice_ipip_ol_reg_event(struct mlxsw_sp *mlxsw_sp,
						struct net_device *ol_dev)
1377 1378
{
	struct mlxsw_sp_ipip_entry *ipip_entry;
1379
	enum mlxsw_sp_l3proto ul_proto;
1380
	enum mlxsw_sp_ipip_type ipipt;
1381 1382
	union mlxsw_sp_l3addr saddr;
	u32 ul_tb_id;
1383 1384

	mlxsw_sp_netdev_ipip_type(mlxsw_sp, ol_dev, &ipipt);
1385
	if (mlxsw_sp_netdevice_ipip_can_offload(mlxsw_sp, ol_dev, ipipt)) {
1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396
		ul_tb_id = mlxsw_sp_ipip_dev_ul_tb_id(ol_dev);
		ul_proto = mlxsw_sp->router->ipip_ops_arr[ipipt]->ul_proto;
		saddr = mlxsw_sp_ipip_netdev_saddr(ul_proto, ol_dev);
		if (!mlxsw_sp_ipip_demote_tunnel_by_saddr(mlxsw_sp, ul_proto,
							  saddr, ul_tb_id,
							  NULL)) {
			ipip_entry = mlxsw_sp_ipip_entry_create(mlxsw_sp, ipipt,
								ol_dev);
			if (IS_ERR(ipip_entry))
				return PTR_ERR(ipip_entry);
		}
1397 1398 1399 1400 1401
	}

	return 0;
}

1402 1403
static void mlxsw_sp_netdevice_ipip_ol_unreg_event(struct mlxsw_sp *mlxsw_sp,
						   struct net_device *ol_dev)
1404 1405 1406 1407 1408
{
	struct mlxsw_sp_ipip_entry *ipip_entry;

	ipip_entry = mlxsw_sp_ipip_entry_find_by_ol_dev(mlxsw_sp, ol_dev);
	if (ipip_entry)
1409
		mlxsw_sp_ipip_entry_destroy(mlxsw_sp, ipip_entry);
1410 1411
}

1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423
static void
mlxsw_sp_ipip_entry_ol_up_event(struct mlxsw_sp *mlxsw_sp,
				struct mlxsw_sp_ipip_entry *ipip_entry)
{
	struct mlxsw_sp_fib_entry *decap_fib_entry;

	decap_fib_entry = mlxsw_sp_ipip_entry_find_decap(mlxsw_sp, ipip_entry);
	if (decap_fib_entry)
		mlxsw_sp_ipip_entry_promote_decap(mlxsw_sp, ipip_entry,
						  decap_fib_entry);
}

1424
static int
1425 1426
mlxsw_sp_rif_ipip_lb_op(struct mlxsw_sp_rif_ipip_lb *lb_rif, u16 ul_vr_id,
			u16 ul_rif_id, bool enable)
1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440
{
	struct mlxsw_sp_rif_ipip_lb_config lb_cf = lb_rif->lb_config;
	struct mlxsw_sp_rif *rif = &lb_rif->common;
	struct mlxsw_sp *mlxsw_sp = rif->mlxsw_sp;
	char ritr_pl[MLXSW_REG_RITR_LEN];
	u32 saddr4;

	switch (lb_cf.ul_protocol) {
	case MLXSW_SP_L3_PROTO_IPV4:
		saddr4 = be32_to_cpu(lb_cf.saddr.addr4);
		mlxsw_reg_ritr_pack(ritr_pl, enable, MLXSW_REG_RITR_LOOPBACK_IF,
				    rif->rif_index, rif->vr_id, rif->dev->mtu);
		mlxsw_reg_ritr_loopback_ipip4_pack(ritr_pl, lb_cf.lb_ipipt,
			    MLXSW_REG_RITR_LOOPBACK_IPIP_OPTIONS_GRE_KEY_PRESET,
1441
			    ul_vr_id, ul_rif_id, saddr4, lb_cf.okey);
1442 1443 1444 1445 1446 1447 1448 1449 1450
		break;

	case MLXSW_SP_L3_PROTO_IPV6:
		return -EAFNOSUPPORT;
	}

	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ritr), ritr_pl);
}

1451 1452 1453 1454 1455 1456 1457 1458 1459 1460
static int mlxsw_sp_netdevice_ipip_ol_update_mtu(struct mlxsw_sp *mlxsw_sp,
						 struct net_device *ol_dev)
{
	struct mlxsw_sp_ipip_entry *ipip_entry;
	struct mlxsw_sp_rif_ipip_lb *lb_rif;
	int err = 0;

	ipip_entry = mlxsw_sp_ipip_entry_find_by_ol_dev(mlxsw_sp, ol_dev);
	if (ipip_entry) {
		lb_rif = ipip_entry->ol_lb;
1461 1462
		err = mlxsw_sp_rif_ipip_lb_op(lb_rif, lb_rif->ul_vr_id,
					      lb_rif->ul_rif_id, true);
1463 1464 1465 1466 1467 1468 1469 1470 1471
		if (err)
			goto out;
		lb_rif->common.mtu = ol_dev->mtu;
	}

out:
	return err;
}

1472 1473
static void mlxsw_sp_netdevice_ipip_ol_up_event(struct mlxsw_sp *mlxsw_sp,
						struct net_device *ol_dev)
1474 1475 1476 1477
{
	struct mlxsw_sp_ipip_entry *ipip_entry;

	ipip_entry = mlxsw_sp_ipip_entry_find_by_ol_dev(mlxsw_sp, ol_dev);
1478 1479
	if (ipip_entry)
		mlxsw_sp_ipip_entry_ol_up_event(mlxsw_sp, ipip_entry);
1480 1481
}

1482 1483 1484 1485 1486 1487 1488 1489
static void
mlxsw_sp_ipip_entry_ol_down_event(struct mlxsw_sp *mlxsw_sp,
				  struct mlxsw_sp_ipip_entry *ipip_entry)
{
	if (ipip_entry->decap_fib_entry)
		mlxsw_sp_ipip_entry_demote_decap(mlxsw_sp, ipip_entry);
}

1490 1491
static void mlxsw_sp_netdevice_ipip_ol_down_event(struct mlxsw_sp *mlxsw_sp,
						  struct net_device *ol_dev)
1492 1493 1494 1495
{
	struct mlxsw_sp_ipip_entry *ipip_entry;

	ipip_entry = mlxsw_sp_ipip_entry_find_by_ol_dev(mlxsw_sp, ol_dev);
1496 1497
	if (ipip_entry)
		mlxsw_sp_ipip_entry_ol_down_event(mlxsw_sp, ipip_entry);
1498 1499
}

1500 1501 1502
static void mlxsw_sp_nexthop_rif_migrate(struct mlxsw_sp *mlxsw_sp,
					 struct mlxsw_sp_rif *old_rif,
					 struct mlxsw_sp_rif *new_rif);
1503 1504 1505
static int
mlxsw_sp_ipip_entry_ol_lb_update(struct mlxsw_sp *mlxsw_sp,
				 struct mlxsw_sp_ipip_entry *ipip_entry,
1506
				 bool keep_encap,
1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518
				 struct netlink_ext_ack *extack)
{
	struct mlxsw_sp_rif_ipip_lb *old_lb_rif = ipip_entry->ol_lb;
	struct mlxsw_sp_rif_ipip_lb *new_lb_rif;

	new_lb_rif = mlxsw_sp_ipip_ol_ipip_lb_create(mlxsw_sp,
						     ipip_entry->ipipt,
						     ipip_entry->ol_dev,
						     extack);
	if (IS_ERR(new_lb_rif))
		return PTR_ERR(new_lb_rif);
	ipip_entry->ol_lb = new_lb_rif;
1519

1520 1521 1522
	if (keep_encap)
		mlxsw_sp_nexthop_rif_migrate(mlxsw_sp, &old_lb_rif->common,
					     &new_lb_rif->common);
1523

1524
	mlxsw_sp_rif_destroy(&old_lb_rif->common);
1525

1526 1527 1528
	return 0;
}

1529 1530 1531
static void mlxsw_sp_nexthop_rif_update(struct mlxsw_sp *mlxsw_sp,
					struct mlxsw_sp_rif *rif);

1532 1533 1534 1535 1536 1537 1538 1539 1540
/**
 * Update the offload related to an IPIP entry. This always updates decap, and
 * in addition to that it also:
 * @recreate_loopback: recreates the associated loopback RIF
 * @keep_encap: updates next hops that use the tunnel netdevice. This is only
 *              relevant when recreate_loopback is true.
 * @update_nexthops: updates next hops, keeping the current loopback RIF. This
 *                   is only relevant when recreate_loopback is false.
 */
1541 1542
int __mlxsw_sp_ipip_entry_update_tunnel(struct mlxsw_sp *mlxsw_sp,
					struct mlxsw_sp_ipip_entry *ipip_entry,
1543 1544 1545
					bool recreate_loopback,
					bool keep_encap,
					bool update_nexthops,
1546 1547 1548
					struct netlink_ext_ack *extack)
{
	int err;
1549

1550 1551 1552 1553
	/* RIFs can't be edited, so to update loopback, we need to destroy and
	 * recreate it. That creates a window of opportunity where RALUE and
	 * RATR registers end up referencing a RIF that's already gone. RATRs
	 * are handled in mlxsw_sp_ipip_entry_ol_lb_update(), and to take care
1554 1555 1556 1557 1558
	 * of RALUE, demote the decap route back.
	 */
	if (ipip_entry->decap_fib_entry)
		mlxsw_sp_ipip_entry_demote_decap(mlxsw_sp, ipip_entry);

1559 1560 1561 1562 1563 1564 1565 1566 1567
	if (recreate_loopback) {
		err = mlxsw_sp_ipip_entry_ol_lb_update(mlxsw_sp, ipip_entry,
						       keep_encap, extack);
		if (err)
			return err;
	} else if (update_nexthops) {
		mlxsw_sp_nexthop_rif_update(mlxsw_sp,
					    &ipip_entry->ol_lb->common);
	}
1568 1569 1570

	if (ipip_entry->ol_dev->flags & IFF_UP)
		mlxsw_sp_ipip_entry_ol_up_event(mlxsw_sp, ipip_entry);
1571 1572 1573 1574

	return 0;
}

1575 1576 1577 1578 1579 1580 1581 1582 1583
static int mlxsw_sp_netdevice_ipip_ol_vrf_event(struct mlxsw_sp *mlxsw_sp,
						struct net_device *ol_dev,
						struct netlink_ext_ack *extack)
{
	struct mlxsw_sp_ipip_entry *ipip_entry =
		mlxsw_sp_ipip_entry_find_by_ol_dev(mlxsw_sp, ol_dev);

	if (!ipip_entry)
		return 0;
1584

1585
	return __mlxsw_sp_ipip_entry_update_tunnel(mlxsw_sp, ipip_entry,
1586
						   true, false, false, extack);
1587 1588
}

1589 1590 1591 1592
static int
mlxsw_sp_netdevice_ipip_ul_vrf_event(struct mlxsw_sp *mlxsw_sp,
				     struct mlxsw_sp_ipip_entry *ipip_entry,
				     struct net_device *ul_dev,
1593
				     bool *demote_this,
1594 1595
				     struct netlink_ext_ack *extack)
{
1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611
	u32 ul_tb_id = l3mdev_fib_table(ul_dev) ? : RT_TABLE_MAIN;
	enum mlxsw_sp_l3proto ul_proto;
	union mlxsw_sp_l3addr saddr;

	/* Moving underlay to a different VRF might cause local address
	 * conflict, and the conflicting tunnels need to be demoted.
	 */
	ul_proto = mlxsw_sp->router->ipip_ops_arr[ipip_entry->ipipt]->ul_proto;
	saddr = mlxsw_sp_ipip_netdev_saddr(ul_proto, ipip_entry->ol_dev);
	if (mlxsw_sp_ipip_demote_tunnel_by_saddr(mlxsw_sp, ul_proto,
						 saddr, ul_tb_id,
						 ipip_entry)) {
		*demote_this = true;
		return 0;
	}

1612 1613 1614 1615
	return __mlxsw_sp_ipip_entry_update_tunnel(mlxsw_sp, ipip_entry,
						   true, true, false, extack);
}

1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637
static int
mlxsw_sp_netdevice_ipip_ul_up_event(struct mlxsw_sp *mlxsw_sp,
				    struct mlxsw_sp_ipip_entry *ipip_entry,
				    struct net_device *ul_dev)
{
	return __mlxsw_sp_ipip_entry_update_tunnel(mlxsw_sp, ipip_entry,
						   false, false, true, NULL);
}

static int
mlxsw_sp_netdevice_ipip_ul_down_event(struct mlxsw_sp *mlxsw_sp,
				      struct mlxsw_sp_ipip_entry *ipip_entry,
				      struct net_device *ul_dev)
{
	/* A down underlay device causes encapsulated packets to not be
	 * forwarded, but decap still works. So refresh next hops without
	 * touching anything else.
	 */
	return __mlxsw_sp_ipip_entry_update_tunnel(mlxsw_sp, ipip_entry,
						   false, false, true, NULL);
}

1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666
static int
mlxsw_sp_netdevice_ipip_ol_change_event(struct mlxsw_sp *mlxsw_sp,
					struct net_device *ol_dev,
					struct netlink_ext_ack *extack)
{
	const struct mlxsw_sp_ipip_ops *ipip_ops;
	struct mlxsw_sp_ipip_entry *ipip_entry;
	int err;

	ipip_entry = mlxsw_sp_ipip_entry_find_by_ol_dev(mlxsw_sp, ol_dev);
	if (!ipip_entry)
		/* A change might make a tunnel eligible for offloading, but
		 * that is currently not implemented. What falls to slow path
		 * stays there.
		 */
		return 0;

	/* A change might make a tunnel not eligible for offloading. */
	if (!mlxsw_sp_netdevice_ipip_can_offload(mlxsw_sp, ol_dev,
						 ipip_entry->ipipt)) {
		mlxsw_sp_ipip_entry_demote_tunnel(mlxsw_sp, ipip_entry);
		return 0;
	}

	ipip_ops = mlxsw_sp->router->ipip_ops_arr[ipip_entry->ipipt];
	err = ipip_ops->ol_netdev_change(mlxsw_sp, ipip_entry, extack);
	return err;
}

1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704
void mlxsw_sp_ipip_entry_demote_tunnel(struct mlxsw_sp *mlxsw_sp,
				       struct mlxsw_sp_ipip_entry *ipip_entry)
{
	struct net_device *ol_dev = ipip_entry->ol_dev;

	if (ol_dev->flags & IFF_UP)
		mlxsw_sp_ipip_entry_ol_down_event(mlxsw_sp, ipip_entry);
	mlxsw_sp_ipip_entry_destroy(mlxsw_sp, ipip_entry);
}

/* The configuration where several tunnels have the same local address in the
 * same underlay table needs special treatment in the HW. That is currently not
 * implemented in the driver. This function finds and demotes the first tunnel
 * with a given source address, except the one passed in in the argument
 * `except'.
 */
bool
mlxsw_sp_ipip_demote_tunnel_by_saddr(struct mlxsw_sp *mlxsw_sp,
				     enum mlxsw_sp_l3proto ul_proto,
				     union mlxsw_sp_l3addr saddr,
				     u32 ul_tb_id,
				     const struct mlxsw_sp_ipip_entry *except)
{
	struct mlxsw_sp_ipip_entry *ipip_entry, *tmp;

	list_for_each_entry_safe(ipip_entry, tmp, &mlxsw_sp->router->ipip_list,
				 ipip_list_node) {
		if (ipip_entry != except &&
		    mlxsw_sp_ipip_entry_saddr_matches(mlxsw_sp, ul_proto, saddr,
						      ul_tb_id, ipip_entry)) {
			mlxsw_sp_ipip_entry_demote_tunnel(mlxsw_sp, ipip_entry);
			return true;
		}
	}

	return false;
}

1705 1706 1707 1708 1709 1710 1711
static void mlxsw_sp_ipip_demote_tunnel_by_ul_netdev(struct mlxsw_sp *mlxsw_sp,
						     struct net_device *ul_dev)
{
	struct mlxsw_sp_ipip_entry *ipip_entry, *tmp;

	list_for_each_entry_safe(ipip_entry, tmp, &mlxsw_sp->router->ipip_list,
				 ipip_list_node) {
1712 1713
		struct net_device *ol_dev = ipip_entry->ol_dev;
		struct net_device *ipip_ul_dev;
1714

1715 1716 1717
		rcu_read_lock();
		ipip_ul_dev = __mlxsw_sp_ipip_netdev_ul_dev_get(ol_dev);
		rcu_read_unlock();
1718 1719 1720 1721 1722
		if (ipip_ul_dev == ul_dev)
			mlxsw_sp_ipip_entry_demote_tunnel(mlxsw_sp, ipip_entry);
	}
}

1723 1724 1725 1726
int mlxsw_sp_netdevice_ipip_ol_event(struct mlxsw_sp *mlxsw_sp,
				     struct net_device *ol_dev,
				     unsigned long event,
				     struct netdev_notifier_info *info)
1727
{
1728 1729
	struct netdev_notifier_changeupper_info *chup;
	struct netlink_ext_ack *extack;
1730
	int err = 0;
1731

1732
	mutex_lock(&mlxsw_sp->router->lock);
1733 1734
	switch (event) {
	case NETDEV_REGISTER:
1735 1736
		err = mlxsw_sp_netdevice_ipip_ol_reg_event(mlxsw_sp, ol_dev);
		break;
1737
	case NETDEV_UNREGISTER:
1738
		mlxsw_sp_netdevice_ipip_ol_unreg_event(mlxsw_sp, ol_dev);
1739
		break;
1740
	case NETDEV_UP:
1741
		mlxsw_sp_netdevice_ipip_ol_up_event(mlxsw_sp, ol_dev);
1742
		break;
1743
	case NETDEV_DOWN:
1744
		mlxsw_sp_netdevice_ipip_ol_down_event(mlxsw_sp, ol_dev);
1745
		break;
1746
	case NETDEV_CHANGEUPPER:
1747 1748 1749
		chup = container_of(info, typeof(*chup), info);
		extack = info->extack;
		if (netif_is_l3_master(chup->upper_dev))
1750 1751 1752 1753
			err = mlxsw_sp_netdevice_ipip_ol_vrf_event(mlxsw_sp,
								   ol_dev,
								   extack);
		break;
1754 1755
	case NETDEV_CHANGE:
		extack = info->extack;
1756 1757 1758
		err = mlxsw_sp_netdevice_ipip_ol_change_event(mlxsw_sp,
							      ol_dev, extack);
		break;
1759
	case NETDEV_CHANGEMTU:
1760 1761
		err = mlxsw_sp_netdevice_ipip_ol_update_mtu(mlxsw_sp, ol_dev);
		break;
1762
	}
1763 1764
	mutex_unlock(&mlxsw_sp->router->lock);
	return err;
1765 1766
}

1767 1768 1769 1770
static int
__mlxsw_sp_netdevice_ipip_ul_event(struct mlxsw_sp *mlxsw_sp,
				   struct mlxsw_sp_ipip_entry *ipip_entry,
				   struct net_device *ul_dev,
1771
				   bool *demote_this,
1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785
				   unsigned long event,
				   struct netdev_notifier_info *info)
{
	struct netdev_notifier_changeupper_info *chup;
	struct netlink_ext_ack *extack;

	switch (event) {
	case NETDEV_CHANGEUPPER:
		chup = container_of(info, typeof(*chup), info);
		extack = info->extack;
		if (netif_is_l3_master(chup->upper_dev))
			return mlxsw_sp_netdevice_ipip_ul_vrf_event(mlxsw_sp,
								    ipip_entry,
								    ul_dev,
1786
								    demote_this,
1787 1788
								    extack);
		break;
1789 1790 1791 1792 1793 1794 1795 1796

	case NETDEV_UP:
		return mlxsw_sp_netdevice_ipip_ul_up_event(mlxsw_sp, ipip_entry,
							   ul_dev);
	case NETDEV_DOWN:
		return mlxsw_sp_netdevice_ipip_ul_down_event(mlxsw_sp,
							     ipip_entry,
							     ul_dev);
1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807
	}
	return 0;
}

int
mlxsw_sp_netdevice_ipip_ul_event(struct mlxsw_sp *mlxsw_sp,
				 struct net_device *ul_dev,
				 unsigned long event,
				 struct netdev_notifier_info *info)
{
	struct mlxsw_sp_ipip_entry *ipip_entry = NULL;
1808
	int err = 0;
1809

1810
	mutex_lock(&mlxsw_sp->router->lock);
1811 1812 1813
	while ((ipip_entry = mlxsw_sp_ipip_entry_find_by_ul_dev(mlxsw_sp,
								ul_dev,
								ipip_entry))) {
1814 1815 1816
		struct mlxsw_sp_ipip_entry *prev;
		bool demote_this = false;

1817
		err = __mlxsw_sp_netdevice_ipip_ul_event(mlxsw_sp, ipip_entry,
1818 1819
							 ul_dev, &demote_this,
							 event, info);
1820 1821 1822
		if (err) {
			mlxsw_sp_ipip_demote_tunnel_by_ul_netdev(mlxsw_sp,
								 ul_dev);
1823
			break;
1824
		}
1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838

		if (demote_this) {
			if (list_is_first(&ipip_entry->ipip_list_node,
					  &mlxsw_sp->router->ipip_list))
				prev = NULL;
			else
				/* This can't be cached from previous iteration,
				 * because that entry could be gone now.
				 */
				prev = list_prev_entry(ipip_entry,
						       ipip_list_node);
			mlxsw_sp_ipip_entry_demote_tunnel(mlxsw_sp, ipip_entry);
			ipip_entry = prev;
		}
1839
	}
1840
	mutex_unlock(&mlxsw_sp->router->lock);
1841

1842
	return err;
1843 1844
}

1845 1846 1847 1848 1849 1850
int mlxsw_sp_router_nve_promote_decap(struct mlxsw_sp *mlxsw_sp, u32 ul_tb_id,
				      enum mlxsw_sp_l3proto ul_proto,
				      const union mlxsw_sp_l3addr *ul_sip,
				      u32 tunnel_index)
{
	enum mlxsw_sp_fib_entry_type type = MLXSW_SP_FIB_ENTRY_TYPE_TRAP;
1851
	struct mlxsw_sp_router *router = mlxsw_sp->router;
1852 1853 1854
	struct mlxsw_sp_fib_entry *fib_entry;
	int err;

1855 1856 1857 1858 1859 1860 1861 1862 1863
	if (WARN_ON_ONCE(router->nve_decap_config.valid))
		return -EINVAL;

	router->nve_decap_config.ul_tb_id = ul_tb_id;
	router->nve_decap_config.tunnel_index = tunnel_index;
	router->nve_decap_config.ul_proto = ul_proto;
	router->nve_decap_config.ul_sip = *ul_sip;
	router->nve_decap_config.valid = true;

1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892
	/* It is valid to create a tunnel with a local IP and only later
	 * assign this IP address to a local interface
	 */
	fib_entry = mlxsw_sp_router_ip2me_fib_entry_find(mlxsw_sp, ul_tb_id,
							 ul_proto, ul_sip,
							 type);
	if (!fib_entry)
		return 0;

	fib_entry->decap.tunnel_index = tunnel_index;
	fib_entry->type = MLXSW_SP_FIB_ENTRY_TYPE_NVE_DECAP;

	err = mlxsw_sp_fib_entry_update(mlxsw_sp, fib_entry);
	if (err)
		goto err_fib_entry_update;

	return 0;

err_fib_entry_update:
	fib_entry->type = MLXSW_SP_FIB_ENTRY_TYPE_TRAP;
	mlxsw_sp_fib_entry_update(mlxsw_sp, fib_entry);
	return err;
}

void mlxsw_sp_router_nve_demote_decap(struct mlxsw_sp *mlxsw_sp, u32 ul_tb_id,
				      enum mlxsw_sp_l3proto ul_proto,
				      const union mlxsw_sp_l3addr *ul_sip)
{
	enum mlxsw_sp_fib_entry_type type = MLXSW_SP_FIB_ENTRY_TYPE_NVE_DECAP;
1893
	struct mlxsw_sp_router *router = mlxsw_sp->router;
1894 1895
	struct mlxsw_sp_fib_entry *fib_entry;

1896 1897 1898 1899 1900
	if (WARN_ON_ONCE(!router->nve_decap_config.valid))
		return;

	router->nve_decap_config.valid = false;

1901 1902 1903 1904 1905 1906 1907 1908 1909 1910
	fib_entry = mlxsw_sp_router_ip2me_fib_entry_find(mlxsw_sp, ul_tb_id,
							 ul_proto, ul_sip,
							 type);
	if (!fib_entry)
		return;

	fib_entry->type = MLXSW_SP_FIB_ENTRY_TYPE_TRAP;
	mlxsw_sp_fib_entry_update(mlxsw_sp, fib_entry);
}

1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924
static bool mlxsw_sp_router_nve_is_decap(struct mlxsw_sp *mlxsw_sp,
					 u32 ul_tb_id,
					 enum mlxsw_sp_l3proto ul_proto,
					 const union mlxsw_sp_l3addr *ul_sip)
{
	struct mlxsw_sp_router *router = mlxsw_sp->router;

	return router->nve_decap_config.valid &&
	       router->nve_decap_config.ul_tb_id == ul_tb_id &&
	       router->nve_decap_config.ul_proto == ul_proto &&
	       !memcmp(&router->nve_decap_config.ul_sip, ul_sip,
		       sizeof(*ul_sip));
}

1925
struct mlxsw_sp_neigh_key {
1926
	struct neighbour *n;
1927 1928 1929
};

struct mlxsw_sp_neigh_entry {
1930
	struct list_head rif_list_node;
1931 1932 1933
	struct rhash_head ht_node;
	struct mlxsw_sp_neigh_key key;
	u16 rif;
1934
	bool connected;
1935
	unsigned char ha[ETH_ALEN];
1936 1937 1938
	struct list_head nexthop_list; /* list of nexthops using
					* this neigh entry
					*/
1939
	struct list_head nexthop_neighs_list_node;
1940 1941
	unsigned int counter_index;
	bool counter_valid;
1942 1943 1944 1945 1946 1947 1948 1949
};

static const struct rhashtable_params mlxsw_sp_neigh_ht_params = {
	.key_offset = offsetof(struct mlxsw_sp_neigh_entry, key),
	.head_offset = offsetof(struct mlxsw_sp_neigh_entry, ht_node),
	.key_len = sizeof(struct mlxsw_sp_neigh_key),
};

1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961
struct mlxsw_sp_neigh_entry *
mlxsw_sp_rif_neigh_next(struct mlxsw_sp_rif *rif,
			struct mlxsw_sp_neigh_entry *neigh_entry)
{
	if (!neigh_entry) {
		if (list_empty(&rif->neigh_list))
			return NULL;
		else
			return list_first_entry(&rif->neigh_list,
						typeof(*neigh_entry),
						rif_list_node);
	}
1962
	if (list_is_last(&neigh_entry->rif_list_node, &rif->neigh_list))
1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985
		return NULL;
	return list_next_entry(neigh_entry, rif_list_node);
}

int mlxsw_sp_neigh_entry_type(struct mlxsw_sp_neigh_entry *neigh_entry)
{
	return neigh_entry->key.n->tbl->family;
}

unsigned char *
mlxsw_sp_neigh_entry_ha(struct mlxsw_sp_neigh_entry *neigh_entry)
{
	return neigh_entry->ha;
}

u32 mlxsw_sp_neigh4_entry_dip(struct mlxsw_sp_neigh_entry *neigh_entry)
{
	struct neighbour *n;

	n = neigh_entry->key.n;
	return ntohl(*((__be32 *) n->primary_key));
}

1986 1987 1988 1989 1990 1991 1992 1993 1994
struct in6_addr *
mlxsw_sp_neigh6_entry_dip(struct mlxsw_sp_neigh_entry *neigh_entry)
{
	struct neighbour *n;

	n = neigh_entry->key.n;
	return (struct in6_addr *) &n->primary_key;
}

1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
int mlxsw_sp_neigh_counter_get(struct mlxsw_sp *mlxsw_sp,
			       struct mlxsw_sp_neigh_entry *neigh_entry,
			       u64 *p_counter)
{
	if (!neigh_entry->counter_valid)
		return -EINVAL;

	return mlxsw_sp_flow_counter_get(mlxsw_sp, neigh_entry->counter_index,
					 p_counter, NULL);
}

2006
static struct mlxsw_sp_neigh_entry *
2007 2008
mlxsw_sp_neigh_entry_alloc(struct mlxsw_sp *mlxsw_sp, struct neighbour *n,
			   u16 rif)
2009 2010 2011
{
	struct mlxsw_sp_neigh_entry *neigh_entry;

2012
	neigh_entry = kzalloc(sizeof(*neigh_entry), GFP_KERNEL);
2013 2014
	if (!neigh_entry)
		return NULL;
2015

2016
	neigh_entry->key.n = n;
2017
	neigh_entry->rif = rif;
2018
	INIT_LIST_HEAD(&neigh_entry->nexthop_list);
2019

2020 2021 2022
	return neigh_entry;
}

2023
static void mlxsw_sp_neigh_entry_free(struct mlxsw_sp_neigh_entry *neigh_entry)
2024 2025 2026 2027
{
	kfree(neigh_entry);
}

2028 2029 2030
static int
mlxsw_sp_neigh_entry_insert(struct mlxsw_sp *mlxsw_sp,
			    struct mlxsw_sp_neigh_entry *neigh_entry)
2031
{
2032
	return rhashtable_insert_fast(&mlxsw_sp->router->neigh_ht,
2033 2034 2035
				      &neigh_entry->ht_node,
				      mlxsw_sp_neigh_ht_params);
}
2036

2037 2038 2039 2040
static void
mlxsw_sp_neigh_entry_remove(struct mlxsw_sp *mlxsw_sp,
			    struct mlxsw_sp_neigh_entry *neigh_entry)
{
2041
	rhashtable_remove_fast(&mlxsw_sp->router->neigh_ht,
2042 2043
			       &neigh_entry->ht_node,
			       mlxsw_sp_neigh_ht_params);
2044 2045
}

2046
static bool
2047 2048
mlxsw_sp_neigh_counter_should_alloc(struct mlxsw_sp *mlxsw_sp,
				    struct mlxsw_sp_neigh_entry *neigh_entry)
2049 2050
{
	struct devlink *devlink;
2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063
	const char *table_name;

	switch (mlxsw_sp_neigh_entry_type(neigh_entry)) {
	case AF_INET:
		table_name = MLXSW_SP_DPIPE_TABLE_NAME_HOST4;
		break;
	case AF_INET6:
		table_name = MLXSW_SP_DPIPE_TABLE_NAME_HOST6;
		break;
	default:
		WARN_ON(1);
		return false;
	}
2064 2065

	devlink = priv_to_devlink(mlxsw_sp->core);
2066
	return devlink_dpipe_table_counter_enabled(devlink, table_name);
2067 2068 2069 2070 2071 2072
}

static void
mlxsw_sp_neigh_counter_alloc(struct mlxsw_sp *mlxsw_sp,
			     struct mlxsw_sp_neigh_entry *neigh_entry)
{
2073
	if (!mlxsw_sp_neigh_counter_should_alloc(mlxsw_sp, neigh_entry))
2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092
		return;

	if (mlxsw_sp_flow_counter_alloc(mlxsw_sp, &neigh_entry->counter_index))
		return;

	neigh_entry->counter_valid = true;
}

static void
mlxsw_sp_neigh_counter_free(struct mlxsw_sp *mlxsw_sp,
			    struct mlxsw_sp_neigh_entry *neigh_entry)
{
	if (!neigh_entry->counter_valid)
		return;
	mlxsw_sp_flow_counter_free(mlxsw_sp,
				   neigh_entry->counter_index);
	neigh_entry->counter_valid = false;
}

2093 2094
static struct mlxsw_sp_neigh_entry *
mlxsw_sp_neigh_entry_create(struct mlxsw_sp *mlxsw_sp, struct neighbour *n)
2095 2096
{
	struct mlxsw_sp_neigh_entry *neigh_entry;
2097
	struct mlxsw_sp_rif *rif;
2098 2099
	int err;

2100 2101
	rif = mlxsw_sp_rif_find_by_dev(mlxsw_sp, n->dev);
	if (!rif)
2102
		return ERR_PTR(-EINVAL);
2103

2104
	neigh_entry = mlxsw_sp_neigh_entry_alloc(mlxsw_sp, n, rif->rif_index);
2105
	if (!neigh_entry)
2106 2107
		return ERR_PTR(-ENOMEM);

2108 2109 2110
	err = mlxsw_sp_neigh_entry_insert(mlxsw_sp, neigh_entry);
	if (err)
		goto err_neigh_entry_insert;
2111

2112
	mlxsw_sp_neigh_counter_alloc(mlxsw_sp, neigh_entry);
2113
	list_add(&neigh_entry->rif_list_node, &rif->neigh_list);
2114

2115
	return neigh_entry;
2116 2117

err_neigh_entry_insert:
2118 2119
	mlxsw_sp_neigh_entry_free(neigh_entry);
	return ERR_PTR(err);
2120 2121
}

2122 2123 2124
static void
mlxsw_sp_neigh_entry_destroy(struct mlxsw_sp *mlxsw_sp,
			     struct mlxsw_sp_neigh_entry *neigh_entry)
2125
{
2126
	list_del(&neigh_entry->rif_list_node);
2127
	mlxsw_sp_neigh_counter_free(mlxsw_sp, neigh_entry);
2128 2129 2130
	mlxsw_sp_neigh_entry_remove(mlxsw_sp, neigh_entry);
	mlxsw_sp_neigh_entry_free(neigh_entry);
}
2131

2132 2133 2134 2135
static struct mlxsw_sp_neigh_entry *
mlxsw_sp_neigh_entry_lookup(struct mlxsw_sp *mlxsw_sp, struct neighbour *n)
{
	struct mlxsw_sp_neigh_key key;
2136

2137
	key.n = n;
2138
	return rhashtable_lookup_fast(&mlxsw_sp->router->neigh_ht,
2139
				      &key, mlxsw_sp_neigh_ht_params);
2140 2141
}

2142 2143 2144
static void
mlxsw_sp_router_neighs_update_interval_init(struct mlxsw_sp *mlxsw_sp)
{
2145
	unsigned long interval;
2146

2147
#if IS_ENABLED(CONFIG_IPV6)
2148 2149 2150
	interval = min_t(unsigned long,
			 NEIGH_VAR(&arp_tbl.parms, DELAY_PROBE_TIME),
			 NEIGH_VAR(&nd_tbl.parms, DELAY_PROBE_TIME));
2151 2152 2153
#else
	interval = NEIGH_VAR(&arp_tbl.parms, DELAY_PROBE_TIME);
#endif
2154
	mlxsw_sp->router->neighs_update.interval = jiffies_to_msecs(interval);
2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168
}

static void mlxsw_sp_router_neigh_ent_ipv4_process(struct mlxsw_sp *mlxsw_sp,
						   char *rauhtd_pl,
						   int ent_index)
{
	struct net_device *dev;
	struct neighbour *n;
	__be32 dipn;
	u32 dip;
	u16 rif;

	mlxsw_reg_rauhtd_ent_ipv4_unpack(rauhtd_pl, ent_index, &rif, &dip);

2169
	if (!mlxsw_sp->router->rifs[rif]) {
2170 2171 2172 2173 2174
		dev_err_ratelimited(mlxsw_sp->bus_info->dev, "Incorrect RIF in neighbour entry\n");
		return;
	}

	dipn = htonl(dip);
2175
	dev = mlxsw_sp->router->rifs[rif]->dev;
2176
	n = neigh_lookup(&arp_tbl, &dipn, dev);
2177
	if (!n)
2178 2179 2180 2181 2182 2183 2184
		return;

	netdev_dbg(dev, "Updating neighbour with IP=%pI4h\n", &dip);
	neigh_event_send(n, NULL);
	neigh_release(n);
}

2185
#if IS_ENABLED(CONFIG_IPV6)
2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204
static void mlxsw_sp_router_neigh_ent_ipv6_process(struct mlxsw_sp *mlxsw_sp,
						   char *rauhtd_pl,
						   int rec_index)
{
	struct net_device *dev;
	struct neighbour *n;
	struct in6_addr dip;
	u16 rif;

	mlxsw_reg_rauhtd_ent_ipv6_unpack(rauhtd_pl, rec_index, &rif,
					 (char *) &dip);

	if (!mlxsw_sp->router->rifs[rif]) {
		dev_err_ratelimited(mlxsw_sp->bus_info->dev, "Incorrect RIF in neighbour entry\n");
		return;
	}

	dev = mlxsw_sp->router->rifs[rif]->dev;
	n = neigh_lookup(&nd_tbl, &dip, dev);
2205
	if (!n)
2206 2207 2208 2209 2210 2211
		return;

	netdev_dbg(dev, "Updating neighbour with IP=%pI6c\n", &dip);
	neigh_event_send(n, NULL);
	neigh_release(n);
}
2212 2213 2214 2215 2216 2217 2218
#else
static void mlxsw_sp_router_neigh_ent_ipv6_process(struct mlxsw_sp *mlxsw_sp,
						   char *rauhtd_pl,
						   int rec_index)
{
}
#endif
2219

2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242
static void mlxsw_sp_router_neigh_rec_ipv4_process(struct mlxsw_sp *mlxsw_sp,
						   char *rauhtd_pl,
						   int rec_index)
{
	u8 num_entries;
	int i;

	num_entries = mlxsw_reg_rauhtd_ipv4_rec_num_entries_get(rauhtd_pl,
								rec_index);
	/* Hardware starts counting at 0, so add 1. */
	num_entries++;

	/* Each record consists of several neighbour entries. */
	for (i = 0; i < num_entries; i++) {
		int ent_index;

		ent_index = rec_index * MLXSW_REG_RAUHTD_IPV4_ENT_PER_REC + i;
		mlxsw_sp_router_neigh_ent_ipv4_process(mlxsw_sp, rauhtd_pl,
						       ent_index);
	}

}

2243 2244 2245 2246 2247 2248 2249 2250 2251
static void mlxsw_sp_router_neigh_rec_ipv6_process(struct mlxsw_sp *mlxsw_sp,
						   char *rauhtd_pl,
						   int rec_index)
{
	/* One record contains one entry. */
	mlxsw_sp_router_neigh_ent_ipv6_process(mlxsw_sp, rauhtd_pl,
					       rec_index);
}

2252 2253 2254 2255 2256 2257 2258 2259 2260
static void mlxsw_sp_router_neigh_rec_process(struct mlxsw_sp *mlxsw_sp,
					      char *rauhtd_pl, int rec_index)
{
	switch (mlxsw_reg_rauhtd_rec_type_get(rauhtd_pl, rec_index)) {
	case MLXSW_REG_RAUHTD_TYPE_IPV4:
		mlxsw_sp_router_neigh_rec_ipv4_process(mlxsw_sp, rauhtd_pl,
						       rec_index);
		break;
	case MLXSW_REG_RAUHTD_TYPE_IPV6:
2261 2262
		mlxsw_sp_router_neigh_rec_ipv6_process(mlxsw_sp, rauhtd_pl,
						       rec_index);
2263 2264 2265 2266
		break;
	}
}

2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286
static bool mlxsw_sp_router_rauhtd_is_full(char *rauhtd_pl)
{
	u8 num_rec, last_rec_index, num_entries;

	num_rec = mlxsw_reg_rauhtd_num_rec_get(rauhtd_pl);
	last_rec_index = num_rec - 1;

	if (num_rec < MLXSW_REG_RAUHTD_REC_MAX_NUM)
		return false;
	if (mlxsw_reg_rauhtd_rec_type_get(rauhtd_pl, last_rec_index) ==
	    MLXSW_REG_RAUHTD_TYPE_IPV6)
		return true;

	num_entries = mlxsw_reg_rauhtd_ipv4_rec_num_entries_get(rauhtd_pl,
								last_rec_index);
	if (++num_entries == MLXSW_REG_RAUHTD_IPV4_ENT_PER_REC)
		return true;
	return false;
}

2287 2288 2289 2290
static int
__mlxsw_sp_router_neighs_update_rauhtd(struct mlxsw_sp *mlxsw_sp,
				       char *rauhtd_pl,
				       enum mlxsw_reg_rauhtd_type type)
2291
{
2292 2293
	int i, num_rec;
	int err;
2294 2295

	rtnl_lock();
2296 2297
	/* Ensure the RIF we read from the device does not change mid-dump. */
	mutex_lock(&mlxsw_sp->router->lock);
2298
	do {
2299
		mlxsw_reg_rauhtd_pack(rauhtd_pl, type);
2300 2301 2302
		err = mlxsw_reg_query(mlxsw_sp->core, MLXSW_REG(rauhtd),
				      rauhtd_pl);
		if (err) {
P
Petr Machata 已提交
2303
			dev_err_ratelimited(mlxsw_sp->bus_info->dev, "Failed to dump neighbour table\n");
2304 2305 2306 2307 2308 2309
			break;
		}
		num_rec = mlxsw_reg_rauhtd_num_rec_get(rauhtd_pl);
		for (i = 0; i < num_rec; i++)
			mlxsw_sp_router_neigh_rec_process(mlxsw_sp, rauhtd_pl,
							  i);
2310
	} while (mlxsw_sp_router_rauhtd_is_full(rauhtd_pl));
2311
	mutex_unlock(&mlxsw_sp->router->lock);
2312 2313
	rtnl_unlock();

2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334
	return err;
}

static int mlxsw_sp_router_neighs_update_rauhtd(struct mlxsw_sp *mlxsw_sp)
{
	enum mlxsw_reg_rauhtd_type type;
	char *rauhtd_pl;
	int err;

	rauhtd_pl = kmalloc(MLXSW_REG_RAUHTD_LEN, GFP_KERNEL);
	if (!rauhtd_pl)
		return -ENOMEM;

	type = MLXSW_REG_RAUHTD_TYPE_IPV4;
	err = __mlxsw_sp_router_neighs_update_rauhtd(mlxsw_sp, rauhtd_pl, type);
	if (err)
		goto out;

	type = MLXSW_REG_RAUHTD_TYPE_IPV6;
	err = __mlxsw_sp_router_neighs_update_rauhtd(mlxsw_sp, rauhtd_pl, type);
out:
2335
	kfree(rauhtd_pl);
2336 2337 2338 2339 2340 2341 2342 2343
	return err;
}

static void mlxsw_sp_router_neighs_update_nh(struct mlxsw_sp *mlxsw_sp)
{
	struct mlxsw_sp_neigh_entry *neigh_entry;

	rtnl_lock();
2344
	mutex_lock(&mlxsw_sp->router->lock);
2345
	list_for_each_entry(neigh_entry, &mlxsw_sp->router->nexthop_neighs_list,
2346
			    nexthop_neighs_list_node)
2347 2348 2349
		/* If this neigh have nexthops, make the kernel think this neigh
		 * is active regardless of the traffic.
		 */
2350
		neigh_event_send(neigh_entry->key.n, NULL);
2351
	mutex_unlock(&mlxsw_sp->router->lock);
2352 2353 2354 2355 2356 2357
	rtnl_unlock();
}

static void
mlxsw_sp_router_neighs_update_work_schedule(struct mlxsw_sp *mlxsw_sp)
{
2358
	unsigned long interval = mlxsw_sp->router->neighs_update.interval;
2359

2360
	mlxsw_core_schedule_dw(&mlxsw_sp->router->neighs_update.dw,
2361 2362 2363 2364 2365
			       msecs_to_jiffies(interval));
}

static void mlxsw_sp_router_neighs_update_work(struct work_struct *work)
{
2366
	struct mlxsw_sp_router *router;
2367 2368
	int err;

2369 2370 2371
	router = container_of(work, struct mlxsw_sp_router,
			      neighs_update.dw.work);
	err = mlxsw_sp_router_neighs_update_rauhtd(router->mlxsw_sp);
2372
	if (err)
2373
		dev_err(router->mlxsw_sp->bus_info->dev, "Could not update kernel for neigh activity");
2374

2375
	mlxsw_sp_router_neighs_update_nh(router->mlxsw_sp);
2376

2377
	mlxsw_sp_router_neighs_update_work_schedule(router->mlxsw_sp);
2378 2379
}

2380 2381 2382
static void mlxsw_sp_router_probe_unresolved_nexthops(struct work_struct *work)
{
	struct mlxsw_sp_neigh_entry *neigh_entry;
2383
	struct mlxsw_sp_router *router;
2384

2385 2386
	router = container_of(work, struct mlxsw_sp_router,
			      nexthop_probe_dw.work);
2387 2388 2389 2390 2391 2392 2393
	/* Iterate over nexthop neighbours, find those who are unresolved and
	 * send arp on them. This solves the chicken-egg problem when
	 * the nexthop wouldn't get offloaded until the neighbor is resolved
	 * but it wouldn't get resolved ever in case traffic is flowing in HW
	 * using different nexthop.
	 */
	rtnl_lock();
2394
	mutex_lock(&router->lock);
2395
	list_for_each_entry(neigh_entry, &router->nexthop_neighs_list,
2396
			    nexthop_neighs_list_node)
2397
		if (!neigh_entry->connected)
2398
			neigh_event_send(neigh_entry->key.n, NULL);
2399
	mutex_unlock(&router->lock);
2400 2401
	rtnl_unlock();

2402
	mlxsw_core_schedule_dw(&router->nexthop_probe_dw,
2403 2404 2405
			       MLXSW_SP_UNRESOLVED_NH_PROBE_INTERVAL);
}

2406 2407 2408
static void
mlxsw_sp_nexthop_neigh_update(struct mlxsw_sp *mlxsw_sp,
			      struct mlxsw_sp_neigh_entry *neigh_entry,
2409
			      bool removing, bool dead);
2410

2411 2412 2413 2414 2415 2416
static enum mlxsw_reg_rauht_op mlxsw_sp_rauht_op(bool adding)
{
	return adding ? MLXSW_REG_RAUHT_OP_WRITE_ADD :
			MLXSW_REG_RAUHT_OP_WRITE_DELETE;
}

2417
static int
2418 2419 2420
mlxsw_sp_router_neigh_entry_op4(struct mlxsw_sp *mlxsw_sp,
				struct mlxsw_sp_neigh_entry *neigh_entry,
				enum mlxsw_reg_rauht_op op)
2421
{
2422
	struct neighbour *n = neigh_entry->key.n;
2423
	u32 dip = ntohl(*((__be32 *) n->primary_key));
2424
	char rauht_pl[MLXSW_REG_RAUHT_LEN];
2425 2426 2427

	mlxsw_reg_rauht_pack4(rauht_pl, op, neigh_entry->rif, neigh_entry->ha,
			      dip);
2428 2429 2430
	if (neigh_entry->counter_valid)
		mlxsw_reg_rauht_pack_counter(rauht_pl,
					     neigh_entry->counter_index);
2431
	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(rauht), rauht_pl);
2432 2433
}

2434
static int
2435 2436 2437 2438 2439 2440 2441 2442 2443 2444
mlxsw_sp_router_neigh_entry_op6(struct mlxsw_sp *mlxsw_sp,
				struct mlxsw_sp_neigh_entry *neigh_entry,
				enum mlxsw_reg_rauht_op op)
{
	struct neighbour *n = neigh_entry->key.n;
	char rauht_pl[MLXSW_REG_RAUHT_LEN];
	const char *dip = n->primary_key;

	mlxsw_reg_rauht_pack6(rauht_pl, op, neigh_entry->rif, neigh_entry->ha,
			      dip);
2445 2446 2447
	if (neigh_entry->counter_valid)
		mlxsw_reg_rauht_pack_counter(rauht_pl,
					     neigh_entry->counter_index);
2448
	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(rauht), rauht_pl);
2449 2450
}

2451
bool mlxsw_sp_neigh_ipv6_ignore(struct mlxsw_sp_neigh_entry *neigh_entry)
2452
{
2453 2454
	struct neighbour *n = neigh_entry->key.n;

2455 2456 2457 2458 2459 2460 2461 2462 2463 2464
	/* Packets with a link-local destination address are trapped
	 * after LPM lookup and never reach the neighbour table, so
	 * there is no need to program such neighbours to the device.
	 */
	if (ipv6_addr_type((struct in6_addr *) &n->primary_key) &
	    IPV6_ADDR_LINKLOCAL)
		return true;
	return false;
}

2465 2466 2467 2468 2469
static void
mlxsw_sp_neigh_entry_update(struct mlxsw_sp *mlxsw_sp,
			    struct mlxsw_sp_neigh_entry *neigh_entry,
			    bool adding)
{
2470 2471 2472
	enum mlxsw_reg_rauht_op op = mlxsw_sp_rauht_op(adding);
	int err;

2473 2474 2475
	if (!adding && !neigh_entry->connected)
		return;
	neigh_entry->connected = adding;
2476
	if (neigh_entry->key.n->tbl->family == AF_INET) {
2477 2478 2479 2480
		err = mlxsw_sp_router_neigh_entry_op4(mlxsw_sp, neigh_entry,
						      op);
		if (err)
			return;
2481
	} else if (neigh_entry->key.n->tbl->family == AF_INET6) {
2482
		if (mlxsw_sp_neigh_ipv6_ignore(neigh_entry))
2483
			return;
2484 2485 2486 2487
		err = mlxsw_sp_router_neigh_entry_op6(mlxsw_sp, neigh_entry,
						      op);
		if (err)
			return;
2488
	} else {
2489
		WARN_ON_ONCE(1);
2490
		return;
2491
	}
2492 2493 2494 2495 2496

	if (adding)
		neigh_entry->key.n->flags |= NTF_OFFLOADED;
	else
		neigh_entry->key.n->flags &= ~NTF_OFFLOADED;
2497 2498
}

2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510
void
mlxsw_sp_neigh_entry_counter_update(struct mlxsw_sp *mlxsw_sp,
				    struct mlxsw_sp_neigh_entry *neigh_entry,
				    bool adding)
{
	if (adding)
		mlxsw_sp_neigh_counter_alloc(mlxsw_sp, neigh_entry);
	else
		mlxsw_sp_neigh_counter_free(mlxsw_sp, neigh_entry);
	mlxsw_sp_neigh_entry_update(mlxsw_sp, neigh_entry, true);
}

2511
struct mlxsw_sp_netevent_work {
2512 2513 2514 2515 2516 2517 2518
	struct work_struct work;
	struct mlxsw_sp *mlxsw_sp;
	struct neighbour *n;
};

static void mlxsw_sp_router_neigh_event_work(struct work_struct *work)
{
2519 2520 2521
	struct mlxsw_sp_netevent_work *net_work =
		container_of(work, struct mlxsw_sp_netevent_work, work);
	struct mlxsw_sp *mlxsw_sp = net_work->mlxsw_sp;
2522
	struct mlxsw_sp_neigh_entry *neigh_entry;
2523
	struct neighbour *n = net_work->n;
2524
	unsigned char ha[ETH_ALEN];
2525
	bool entry_connected;
2526
	u8 nud_state, dead;
2527

2528 2529 2530 2531
	/* If these parameters are changed after we release the lock,
	 * then we are guaranteed to receive another event letting us
	 * know about it.
	 */
2532
	read_lock_bh(&n->lock);
2533
	memcpy(ha, n->ha, ETH_ALEN);
2534
	nud_state = n->nud_state;
2535
	dead = n->dead;
2536 2537
	read_unlock_bh(&n->lock);

2538
	rtnl_lock();
2539
	mutex_lock(&mlxsw_sp->router->lock);
2540 2541
	mlxsw_sp_span_respin(mlxsw_sp);

2542
	entry_connected = nud_state & NUD_VALID && !dead;
2543 2544 2545 2546 2547 2548 2549
	neigh_entry = mlxsw_sp_neigh_entry_lookup(mlxsw_sp, n);
	if (!entry_connected && !neigh_entry)
		goto out;
	if (!neigh_entry) {
		neigh_entry = mlxsw_sp_neigh_entry_create(mlxsw_sp, n);
		if (IS_ERR(neigh_entry))
			goto out;
2550 2551
	}

2552 2553
	memcpy(neigh_entry->ha, ha, ETH_ALEN);
	mlxsw_sp_neigh_entry_update(mlxsw_sp, neigh_entry, entry_connected);
2554 2555
	mlxsw_sp_nexthop_neigh_update(mlxsw_sp, neigh_entry, !entry_connected,
				      dead);
2556 2557 2558 2559 2560

	if (!neigh_entry->connected && list_empty(&neigh_entry->nexthop_list))
		mlxsw_sp_neigh_entry_destroy(mlxsw_sp, neigh_entry);

out:
2561
	mutex_unlock(&mlxsw_sp->router->lock);
2562
	rtnl_unlock();
2563
	neigh_release(n);
2564
	kfree(net_work);
2565 2566
}

2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578
static int mlxsw_sp_mp_hash_init(struct mlxsw_sp *mlxsw_sp);

static void mlxsw_sp_router_mp_hash_event_work(struct work_struct *work)
{
	struct mlxsw_sp_netevent_work *net_work =
		container_of(work, struct mlxsw_sp_netevent_work, work);
	struct mlxsw_sp *mlxsw_sp = net_work->mlxsw_sp;

	mlxsw_sp_mp_hash_init(mlxsw_sp);
	kfree(net_work);
}

2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590
static int __mlxsw_sp_router_init(struct mlxsw_sp *mlxsw_sp);

static void mlxsw_sp_router_update_priority_work(struct work_struct *work)
{
	struct mlxsw_sp_netevent_work *net_work =
		container_of(work, struct mlxsw_sp_netevent_work, work);
	struct mlxsw_sp *mlxsw_sp = net_work->mlxsw_sp;

	__mlxsw_sp_router_init(mlxsw_sp);
	kfree(net_work);
}

2591 2592 2593 2594 2595 2596 2597
static int mlxsw_sp_router_schedule_work(struct net *net,
					 struct notifier_block *nb,
					 void (*cb)(struct work_struct *))
{
	struct mlxsw_sp_netevent_work *net_work;
	struct mlxsw_sp_router *router;

2598 2599
	router = container_of(nb, struct mlxsw_sp_router, netevent_nb);
	if (!net_eq(net, mlxsw_sp_net(router->mlxsw_sp)))
2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611
		return NOTIFY_DONE;

	net_work = kzalloc(sizeof(*net_work), GFP_ATOMIC);
	if (!net_work)
		return NOTIFY_BAD;

	INIT_WORK(&net_work->work, cb);
	net_work->mlxsw_sp = router->mlxsw_sp;
	mlxsw_core_schedule_work(&net_work->work);
	return NOTIFY_DONE;
}

2612
static int mlxsw_sp_router_netevent_event(struct notifier_block *nb,
2613
					  unsigned long event, void *ptr)
2614
{
2615
	struct mlxsw_sp_netevent_work *net_work;
2616 2617 2618 2619
	struct mlxsw_sp_port *mlxsw_sp_port;
	struct mlxsw_sp *mlxsw_sp;
	unsigned long interval;
	struct neigh_parms *p;
2620
	struct neighbour *n;
2621 2622 2623 2624 2625 2626

	switch (event) {
	case NETEVENT_DELAY_PROBE_TIME_UPDATE:
		p = ptr;

		/* We don't care about changes in the default table. */
2627 2628
		if (!p->dev || (p->tbl->family != AF_INET &&
				p->tbl->family != AF_INET6))
2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639
			return NOTIFY_DONE;

		/* We are in atomic context and can't take RTNL mutex,
		 * so use RCU variant to walk the device chain.
		 */
		mlxsw_sp_port = mlxsw_sp_port_lower_dev_hold(p->dev);
		if (!mlxsw_sp_port)
			return NOTIFY_DONE;

		mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
		interval = jiffies_to_msecs(NEIGH_VAR(p, DELAY_PROBE_TIME));
2640
		mlxsw_sp->router->neighs_update.interval = interval;
2641 2642 2643

		mlxsw_sp_port_dev_put(mlxsw_sp_port);
		break;
2644 2645 2646
	case NETEVENT_NEIGH_UPDATE:
		n = ptr;

2647
		if (n->tbl->family != AF_INET && n->tbl->family != AF_INET6)
2648 2649
			return NOTIFY_DONE;

2650
		mlxsw_sp_port = mlxsw_sp_port_lower_dev_hold(n->dev);
2651 2652 2653
		if (!mlxsw_sp_port)
			return NOTIFY_DONE;

2654 2655
		net_work = kzalloc(sizeof(*net_work), GFP_ATOMIC);
		if (!net_work) {
2656
			mlxsw_sp_port_dev_put(mlxsw_sp_port);
2657
			return NOTIFY_BAD;
2658
		}
2659

2660 2661 2662
		INIT_WORK(&net_work->work, mlxsw_sp_router_neigh_event_work);
		net_work->mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
		net_work->n = n;
2663 2664 2665 2666 2667 2668

		/* Take a reference to ensure the neighbour won't be
		 * destructed until we drop the reference in delayed
		 * work.
		 */
		neigh_clone(n);
2669
		mlxsw_core_schedule_work(&net_work->work);
2670
		mlxsw_sp_port_dev_put(mlxsw_sp_port);
2671
		break;
2672
	case NETEVENT_IPV4_MPATH_HASH_UPDATE:
2673
	case NETEVENT_IPV6_MPATH_HASH_UPDATE:
2674 2675
		return mlxsw_sp_router_schedule_work(ptr, nb,
				mlxsw_sp_router_mp_hash_event_work);
2676

2677 2678 2679
	case NETEVENT_IPV4_FWD_UPDATE_PRIORITY_UPDATE:
		return mlxsw_sp_router_schedule_work(ptr, nb,
				mlxsw_sp_router_update_priority_work);
2680 2681 2682 2683 2684
	}

	return NOTIFY_DONE;
}

2685 2686
static int mlxsw_sp_neigh_init(struct mlxsw_sp *mlxsw_sp)
{
2687 2688
	int err;

2689
	err = rhashtable_init(&mlxsw_sp->router->neigh_ht,
2690 2691 2692 2693 2694 2695 2696 2697 2698
			      &mlxsw_sp_neigh_ht_params);
	if (err)
		return err;

	/* Initialize the polling interval according to the default
	 * table.
	 */
	mlxsw_sp_router_neighs_update_interval_init(mlxsw_sp);

2699
	/* Create the delayed works for the activity_update */
2700
	INIT_DELAYED_WORK(&mlxsw_sp->router->neighs_update.dw,
2701
			  mlxsw_sp_router_neighs_update_work);
2702
	INIT_DELAYED_WORK(&mlxsw_sp->router->nexthop_probe_dw,
2703
			  mlxsw_sp_router_probe_unresolved_nexthops);
2704 2705
	mlxsw_core_schedule_dw(&mlxsw_sp->router->neighs_update.dw, 0);
	mlxsw_core_schedule_dw(&mlxsw_sp->router->nexthop_probe_dw, 0);
2706
	return 0;
2707 2708 2709 2710
}

static void mlxsw_sp_neigh_fini(struct mlxsw_sp *mlxsw_sp)
{
2711 2712 2713
	cancel_delayed_work_sync(&mlxsw_sp->router->neighs_update.dw);
	cancel_delayed_work_sync(&mlxsw_sp->router->nexthop_probe_dw);
	rhashtable_destroy(&mlxsw_sp->router->neigh_ht);
2714 2715
}

2716
static void mlxsw_sp_neigh_rif_gone_sync(struct mlxsw_sp *mlxsw_sp,
2717
					 struct mlxsw_sp_rif *rif)
2718 2719 2720
{
	struct mlxsw_sp_neigh_entry *neigh_entry, *tmp;

2721
	list_for_each_entry_safe(neigh_entry, tmp, &rif->neigh_list,
2722 2723
				 rif_list_node) {
		mlxsw_sp_neigh_entry_update(mlxsw_sp, neigh_entry, false);
2724
		mlxsw_sp_neigh_entry_destroy(mlxsw_sp, neigh_entry);
2725
	}
2726 2727
}

2728 2729
enum mlxsw_sp_nexthop_type {
	MLXSW_SP_NEXTHOP_TYPE_ETH,
2730
	MLXSW_SP_NEXTHOP_TYPE_IPIP,
2731 2732
};

2733 2734 2735 2736
struct mlxsw_sp_nexthop_key {
	struct fib_nh *fib_nh;
};

2737 2738
struct mlxsw_sp_nexthop {
	struct list_head neigh_list_node; /* member of neigh entry list */
2739
	struct list_head rif_list_node;
2740
	struct list_head router_list_node;
2741 2742 2743
	struct mlxsw_sp_nexthop_group *nh_grp; /* pointer back to the group
						* this belongs to
						*/
2744 2745
	struct rhash_head ht_node;
	struct mlxsw_sp_nexthop_key key;
2746
	unsigned char gw_addr[sizeof(struct in6_addr)];
2747
	int ifindex;
2748
	int nh_weight;
2749 2750
	int norm_nh_weight;
	int num_adj_entries;
2751
	struct mlxsw_sp_rif *rif;
2752 2753 2754 2755 2756 2757 2758 2759 2760
	u8 should_offload:1, /* set indicates this neigh is connected and
			      * should be put to KVD linear area of this group.
			      */
	   offloaded:1, /* set in case the neigh is actually put into
			 * KVD linear area of this group.
			 */
	   update:1; /* set indicates that MAC of this neigh should be
		      * updated in HW
		      */
2761 2762 2763
	enum mlxsw_sp_nexthop_type type;
	union {
		struct mlxsw_sp_neigh_entry *neigh_entry;
2764
		struct mlxsw_sp_ipip_entry *ipip_entry;
2765
	};
2766 2767
	unsigned int counter_index;
	bool counter_valid;
2768 2769 2770
};

struct mlxsw_sp_nexthop_group {
2771
	void *priv;
2772
	struct rhash_head ht_node;
2773
	struct list_head fib_list; /* list of fib entries that use this group */
2774
	struct neigh_table *neigh_tbl;
2775 2776
	u8 adj_index_valid:1,
	   gateway:1; /* routes using the group use a gateway */
2777 2778 2779
	u32 adj_index;
	u16 ecmp_size;
	u16 count;
2780
	int sum_norm_weight;
2781
	struct mlxsw_sp_nexthop nexthops[0];
2782
#define nh_rif	nexthops[0].rif
2783 2784
};

2785 2786
void mlxsw_sp_nexthop_counter_alloc(struct mlxsw_sp *mlxsw_sp,
				    struct mlxsw_sp_nexthop *nh)
2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800
{
	struct devlink *devlink;

	devlink = priv_to_devlink(mlxsw_sp->core);
	if (!devlink_dpipe_table_counter_enabled(devlink,
						 MLXSW_SP_DPIPE_TABLE_NAME_ADJ))
		return;

	if (mlxsw_sp_flow_counter_alloc(mlxsw_sp, &nh->counter_index))
		return;

	nh->counter_valid = true;
}

2801 2802
void mlxsw_sp_nexthop_counter_free(struct mlxsw_sp *mlxsw_sp,
				   struct mlxsw_sp_nexthop *nh)
2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819
{
	if (!nh->counter_valid)
		return;
	mlxsw_sp_flow_counter_free(mlxsw_sp, nh->counter_index);
	nh->counter_valid = false;
}

int mlxsw_sp_nexthop_counter_get(struct mlxsw_sp *mlxsw_sp,
				 struct mlxsw_sp_nexthop *nh, u64 *p_counter)
{
	if (!nh->counter_valid)
		return -EINVAL;

	return mlxsw_sp_flow_counter_get(mlxsw_sp, nh->counter_index,
					 p_counter, NULL);
}

2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847
struct mlxsw_sp_nexthop *mlxsw_sp_nexthop_next(struct mlxsw_sp_router *router,
					       struct mlxsw_sp_nexthop *nh)
{
	if (!nh) {
		if (list_empty(&router->nexthop_list))
			return NULL;
		else
			return list_first_entry(&router->nexthop_list,
						typeof(*nh), router_list_node);
	}
	if (list_is_last(&nh->router_list_node, &router->nexthop_list))
		return NULL;
	return list_next_entry(nh, router_list_node);
}

bool mlxsw_sp_nexthop_offload(struct mlxsw_sp_nexthop *nh)
{
	return nh->offloaded;
}

unsigned char *mlxsw_sp_nexthop_ha(struct mlxsw_sp_nexthop *nh)
{
	if (!nh->offloaded)
		return NULL;
	return nh->neigh_entry->ha;
}

int mlxsw_sp_nexthop_indexes(struct mlxsw_sp_nexthop *nh, u32 *p_adj_index,
2848
			     u32 *p_adj_size, u32 *p_adj_hash_index)
2849 2850 2851 2852 2853 2854 2855 2856 2857
{
	struct mlxsw_sp_nexthop_group *nh_grp = nh->nh_grp;
	u32 adj_hash_index = 0;
	int i;

	if (!nh->offloaded || !nh_grp->adj_index_valid)
		return -EINVAL;

	*p_adj_index = nh_grp->adj_index;
2858
	*p_adj_size = nh_grp->ecmp_size;
2859 2860 2861 2862 2863 2864 2865

	for (i = 0; i < nh_grp->count; i++) {
		struct mlxsw_sp_nexthop *nh_iter = &nh_grp->nexthops[i];

		if (nh_iter == nh)
			break;
		if (nh_iter->offloaded)
2866
			adj_hash_index += nh_iter->num_adj_entries;
2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891
	}

	*p_adj_hash_index = adj_hash_index;
	return 0;
}

struct mlxsw_sp_rif *mlxsw_sp_nexthop_rif(struct mlxsw_sp_nexthop *nh)
{
	return nh->rif;
}

bool mlxsw_sp_nexthop_group_has_ipip(struct mlxsw_sp_nexthop *nh)
{
	struct mlxsw_sp_nexthop_group *nh_grp = nh->nh_grp;
	int i;

	for (i = 0; i < nh_grp->count; i++) {
		struct mlxsw_sp_nexthop *nh_iter = &nh_grp->nexthops[i];

		if (nh_iter->type == MLXSW_SP_NEXTHOP_TYPE_IPIP)
			return true;
	}
	return false;
}

2892 2893 2894 2895 2896 2897 2898
static struct fib_info *
mlxsw_sp_nexthop4_group_fi(const struct mlxsw_sp_nexthop_group *nh_grp)
{
	return nh_grp->priv;
}

struct mlxsw_sp_nexthop_group_cmp_arg {
2899 2900 2901 2902 2903
	enum mlxsw_sp_l3proto proto;
	union {
		struct fib_info *fi;
		struct mlxsw_sp_fib6_entry *fib6_entry;
	};
2904 2905
};

2906 2907
static bool
mlxsw_sp_nexthop6_group_has_nexthop(const struct mlxsw_sp_nexthop_group *nh_grp,
2908 2909
				    const struct in6_addr *gw, int ifindex,
				    int weight)
2910 2911 2912 2913 2914 2915 2916
{
	int i;

	for (i = 0; i < nh_grp->count; i++) {
		const struct mlxsw_sp_nexthop *nh;

		nh = &nh_grp->nexthops[i];
2917
		if (nh->ifindex == ifindex && nh->nh_weight == weight &&
2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934
		    ipv6_addr_equal(gw, (struct in6_addr *) nh->gw_addr))
			return true;
	}

	return false;
}

static bool
mlxsw_sp_nexthop6_group_cmp(const struct mlxsw_sp_nexthop_group *nh_grp,
			    const struct mlxsw_sp_fib6_entry *fib6_entry)
{
	struct mlxsw_sp_rt6 *mlxsw_sp_rt6;

	if (nh_grp->count != fib6_entry->nrt6)
		return false;

	list_for_each_entry(mlxsw_sp_rt6, &fib6_entry->rt6_list, list) {
2935
		struct fib6_nh *fib6_nh = mlxsw_sp_rt6->rt->fib6_nh;
2936
		struct in6_addr *gw;
2937
		int ifindex, weight;
2938

D
David Ahern 已提交
2939 2940 2941
		ifindex = fib6_nh->fib_nh_dev->ifindex;
		weight = fib6_nh->fib_nh_weight;
		gw = &fib6_nh->fib_nh_gw6;
2942 2943
		if (!mlxsw_sp_nexthop6_group_has_nexthop(nh_grp, gw, ifindex,
							 weight))
2944 2945 2946 2947 2948 2949
			return false;
	}

	return true;
}

2950 2951 2952 2953 2954 2955
static int
mlxsw_sp_nexthop_group_cmp(struct rhashtable_compare_arg *arg, const void *ptr)
{
	const struct mlxsw_sp_nexthop_group_cmp_arg *cmp_arg = arg->key;
	const struct mlxsw_sp_nexthop_group *nh_grp = ptr;

2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971
	switch (cmp_arg->proto) {
	case MLXSW_SP_L3_PROTO_IPV4:
		return cmp_arg->fi != mlxsw_sp_nexthop4_group_fi(nh_grp);
	case MLXSW_SP_L3_PROTO_IPV6:
		return !mlxsw_sp_nexthop6_group_cmp(nh_grp,
						    cmp_arg->fib6_entry);
	default:
		WARN_ON(1);
		return 1;
	}
}

static int
mlxsw_sp_nexthop_group_type(const struct mlxsw_sp_nexthop_group *nh_grp)
{
	return nh_grp->neigh_tbl->family;
2972 2973 2974 2975 2976
}

static u32 mlxsw_sp_nexthop_group_hash_obj(const void *data, u32 len, u32 seed)
{
	const struct mlxsw_sp_nexthop_group *nh_grp = data;
2977 2978 2979 2980
	const struct mlxsw_sp_nexthop *nh;
	struct fib_info *fi;
	unsigned int val;
	int i;
2981

2982 2983 2984 2985 2986 2987 2988 2989
	switch (mlxsw_sp_nexthop_group_type(nh_grp)) {
	case AF_INET:
		fi = mlxsw_sp_nexthop4_group_fi(nh_grp);
		return jhash(&fi, sizeof(fi), seed);
	case AF_INET6:
		val = nh_grp->count;
		for (i = 0; i < nh_grp->count; i++) {
			nh = &nh_grp->nexthops[i];
2990
			val ^= jhash(&nh->ifindex, sizeof(nh->ifindex), seed);
2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006
		}
		return jhash(&val, sizeof(val), seed);
	default:
		WARN_ON(1);
		return 0;
	}
}

static u32
mlxsw_sp_nexthop6_group_hash(struct mlxsw_sp_fib6_entry *fib6_entry, u32 seed)
{
	unsigned int val = fib6_entry->nrt6;
	struct mlxsw_sp_rt6 *mlxsw_sp_rt6;
	struct net_device *dev;

	list_for_each_entry(mlxsw_sp_rt6, &fib6_entry->rt6_list, list) {
3007
		dev = mlxsw_sp_rt6->rt->fib6_nh->fib_nh_dev;
3008
		val ^= jhash(&dev->ifindex, sizeof(dev->ifindex), seed);
3009 3010 3011
	}

	return jhash(&val, sizeof(val), seed);
3012 3013 3014 3015 3016 3017 3018
}

static u32
mlxsw_sp_nexthop_group_hash(const void *data, u32 len, u32 seed)
{
	const struct mlxsw_sp_nexthop_group_cmp_arg *cmp_arg = data;

3019 3020 3021 3022 3023 3024 3025 3026 3027
	switch (cmp_arg->proto) {
	case MLXSW_SP_L3_PROTO_IPV4:
		return jhash(&cmp_arg->fi, sizeof(cmp_arg->fi), seed);
	case MLXSW_SP_L3_PROTO_IPV6:
		return mlxsw_sp_nexthop6_group_hash(cmp_arg->fib6_entry, seed);
	default:
		WARN_ON(1);
		return 0;
	}
3028 3029
}

3030 3031
static const struct rhashtable_params mlxsw_sp_nexthop_group_ht_params = {
	.head_offset = offsetof(struct mlxsw_sp_nexthop_group, ht_node),
3032 3033 3034
	.hashfn	     = mlxsw_sp_nexthop_group_hash,
	.obj_hashfn  = mlxsw_sp_nexthop_group_hash_obj,
	.obj_cmpfn   = mlxsw_sp_nexthop_group_cmp,
3035 3036 3037 3038 3039
};

static int mlxsw_sp_nexthop_group_insert(struct mlxsw_sp *mlxsw_sp,
					 struct mlxsw_sp_nexthop_group *nh_grp)
{
3040 3041 3042 3043
	if (mlxsw_sp_nexthop_group_type(nh_grp) == AF_INET6 &&
	    !nh_grp->gateway)
		return 0;

3044
	return rhashtable_insert_fast(&mlxsw_sp->router->nexthop_group_ht,
3045 3046 3047 3048 3049 3050 3051
				      &nh_grp->ht_node,
				      mlxsw_sp_nexthop_group_ht_params);
}

static void mlxsw_sp_nexthop_group_remove(struct mlxsw_sp *mlxsw_sp,
					  struct mlxsw_sp_nexthop_group *nh_grp)
{
3052 3053 3054 3055
	if (mlxsw_sp_nexthop_group_type(nh_grp) == AF_INET6 &&
	    !nh_grp->gateway)
		return;

3056
	rhashtable_remove_fast(&mlxsw_sp->router->nexthop_group_ht,
3057 3058 3059 3060 3061
			       &nh_grp->ht_node,
			       mlxsw_sp_nexthop_group_ht_params);
}

static struct mlxsw_sp_nexthop_group *
3062 3063
mlxsw_sp_nexthop4_group_lookup(struct mlxsw_sp *mlxsw_sp,
			       struct fib_info *fi)
3064
{
3065 3066
	struct mlxsw_sp_nexthop_group_cmp_arg cmp_arg;

3067
	cmp_arg.proto = MLXSW_SP_L3_PROTO_IPV4;
3068 3069 3070
	cmp_arg.fi = fi;
	return rhashtable_lookup_fast(&mlxsw_sp->router->nexthop_group_ht,
				      &cmp_arg,
3071 3072 3073
				      mlxsw_sp_nexthop_group_ht_params);
}

3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086
static struct mlxsw_sp_nexthop_group *
mlxsw_sp_nexthop6_group_lookup(struct mlxsw_sp *mlxsw_sp,
			       struct mlxsw_sp_fib6_entry *fib6_entry)
{
	struct mlxsw_sp_nexthop_group_cmp_arg cmp_arg;

	cmp_arg.proto = MLXSW_SP_L3_PROTO_IPV6;
	cmp_arg.fib6_entry = fib6_entry;
	return rhashtable_lookup_fast(&mlxsw_sp->router->nexthop_group_ht,
				      &cmp_arg,
				      mlxsw_sp_nexthop_group_ht_params);
}

3087 3088 3089 3090 3091 3092 3093 3094 3095
static const struct rhashtable_params mlxsw_sp_nexthop_ht_params = {
	.key_offset = offsetof(struct mlxsw_sp_nexthop, key),
	.head_offset = offsetof(struct mlxsw_sp_nexthop, ht_node),
	.key_len = sizeof(struct mlxsw_sp_nexthop_key),
};

static int mlxsw_sp_nexthop_insert(struct mlxsw_sp *mlxsw_sp,
				   struct mlxsw_sp_nexthop *nh)
{
3096
	return rhashtable_insert_fast(&mlxsw_sp->router->nexthop_ht,
3097 3098 3099 3100 3101 3102
				      &nh->ht_node, mlxsw_sp_nexthop_ht_params);
}

static void mlxsw_sp_nexthop_remove(struct mlxsw_sp *mlxsw_sp,
				    struct mlxsw_sp_nexthop *nh)
{
3103
	rhashtable_remove_fast(&mlxsw_sp->router->nexthop_ht, &nh->ht_node,
3104 3105 3106
			       mlxsw_sp_nexthop_ht_params);
}

3107 3108 3109 3110
static struct mlxsw_sp_nexthop *
mlxsw_sp_nexthop_lookup(struct mlxsw_sp *mlxsw_sp,
			struct mlxsw_sp_nexthop_key key)
{
3111
	return rhashtable_lookup_fast(&mlxsw_sp->router->nexthop_ht, &key,
3112 3113 3114
				      mlxsw_sp_nexthop_ht_params);
}

3115
static int mlxsw_sp_adj_index_mass_update_vr(struct mlxsw_sp *mlxsw_sp,
3116
					     const struct mlxsw_sp_fib *fib,
3117 3118 3119 3120 3121 3122
					     u32 adj_index, u16 ecmp_size,
					     u32 new_adj_index,
					     u16 new_ecmp_size)
{
	char raleu_pl[MLXSW_REG_RALEU_LEN];

3123
	mlxsw_reg_raleu_pack(raleu_pl,
3124 3125
			     (enum mlxsw_reg_ralxx_protocol) fib->proto,
			     fib->vr->id, adj_index, ecmp_size, new_adj_index,
3126
			     new_ecmp_size);
3127 3128 3129 3130 3131 3132 3133 3134
	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(raleu), raleu_pl);
}

static int mlxsw_sp_adj_index_mass_update(struct mlxsw_sp *mlxsw_sp,
					  struct mlxsw_sp_nexthop_group *nh_grp,
					  u32 old_adj_index, u16 old_ecmp_size)
{
	struct mlxsw_sp_fib_entry *fib_entry;
3135
	struct mlxsw_sp_fib *fib = NULL;
3136 3137 3138
	int err;

	list_for_each_entry(fib_entry, &nh_grp->fib_list, nexthop_group_node) {
3139
		if (fib == fib_entry->fib_node->fib)
3140
			continue;
3141 3142
		fib = fib_entry->fib_node->fib;
		err = mlxsw_sp_adj_index_mass_update_vr(mlxsw_sp, fib,
3143 3144 3145 3146 3147 3148 3149 3150 3151 3152
							old_adj_index,
							old_ecmp_size,
							nh_grp->adj_index,
							nh_grp->ecmp_size);
		if (err)
			return err;
	}
	return 0;
}

3153 3154
static int __mlxsw_sp_nexthop_update(struct mlxsw_sp *mlxsw_sp, u32 adj_index,
				     struct mlxsw_sp_nexthop *nh)
3155 3156 3157 3158 3159
{
	struct mlxsw_sp_neigh_entry *neigh_entry = nh->neigh_entry;
	char ratr_pl[MLXSW_REG_RATR_LEN];

	mlxsw_reg_ratr_pack(ratr_pl, MLXSW_REG_RATR_OP_WRITE_WRITE_ENTRY,
3160 3161
			    true, MLXSW_REG_RATR_TYPE_ETHERNET,
			    adj_index, neigh_entry->rif);
3162
	mlxsw_reg_ratr_eth_entry_pack(ratr_pl, neigh_entry->ha);
3163 3164 3165 3166 3167
	if (nh->counter_valid)
		mlxsw_reg_ratr_counter_pack(ratr_pl, nh->counter_index, true);
	else
		mlxsw_reg_ratr_counter_pack(ratr_pl, 0, false);

3168 3169 3170
	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ratr), ratr_pl);
}

3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189
int mlxsw_sp_nexthop_update(struct mlxsw_sp *mlxsw_sp, u32 adj_index,
			    struct mlxsw_sp_nexthop *nh)
{
	int i;

	for (i = 0; i < nh->num_adj_entries; i++) {
		int err;

		err = __mlxsw_sp_nexthop_update(mlxsw_sp, adj_index + i, nh);
		if (err)
			return err;
	}

	return 0;
}

static int __mlxsw_sp_nexthop_ipip_update(struct mlxsw_sp *mlxsw_sp,
					  u32 adj_index,
					  struct mlxsw_sp_nexthop *nh)
3190 3191 3192 3193 3194 3195 3196
{
	const struct mlxsw_sp_ipip_ops *ipip_ops;

	ipip_ops = mlxsw_sp->router->ipip_ops_arr[nh->ipip_entry->ipipt];
	return ipip_ops->nexthop_update(mlxsw_sp, adj_index, nh->ipip_entry);
}

3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214
static int mlxsw_sp_nexthop_ipip_update(struct mlxsw_sp *mlxsw_sp,
					u32 adj_index,
					struct mlxsw_sp_nexthop *nh)
{
	int i;

	for (i = 0; i < nh->num_adj_entries; i++) {
		int err;

		err = __mlxsw_sp_nexthop_ipip_update(mlxsw_sp, adj_index + i,
						     nh);
		if (err)
			return err;
	}

	return 0;
}

3215
static int
3216 3217 3218
mlxsw_sp_nexthop_group_update(struct mlxsw_sp *mlxsw_sp,
			      struct mlxsw_sp_nexthop_group *nh_grp,
			      bool reallocate)
3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232
{
	u32 adj_index = nh_grp->adj_index; /* base */
	struct mlxsw_sp_nexthop *nh;
	int i;
	int err;

	for (i = 0; i < nh_grp->count; i++) {
		nh = &nh_grp->nexthops[i];

		if (!nh->should_offload) {
			nh->offloaded = 0;
			continue;
		}

3233
		if (nh->update || reallocate) {
3234 3235
			switch (nh->type) {
			case MLXSW_SP_NEXTHOP_TYPE_ETH:
3236
				err = mlxsw_sp_nexthop_update
3237 3238
					    (mlxsw_sp, adj_index, nh);
				break;
3239 3240 3241 3242
			case MLXSW_SP_NEXTHOP_TYPE_IPIP:
				err = mlxsw_sp_nexthop_ipip_update
					    (mlxsw_sp, adj_index, nh);
				break;
3243
			}
3244 3245 3246 3247 3248
			if (err)
				return err;
			nh->update = 0;
			nh->offloaded = 1;
		}
3249
		adj_index += nh->num_adj_entries;
3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268
	}
	return 0;
}

static int
mlxsw_sp_nexthop_fib_entries_update(struct mlxsw_sp *mlxsw_sp,
				    struct mlxsw_sp_nexthop_group *nh_grp)
{
	struct mlxsw_sp_fib_entry *fib_entry;
	int err;

	list_for_each_entry(fib_entry, &nh_grp->fib_list, nexthop_group_node) {
		err = mlxsw_sp_fib_entry_update(mlxsw_sp, fib_entry);
		if (err)
			return err;
	}
	return 0;
}

3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308
static void mlxsw_sp_adj_grp_size_round_up(u16 *p_adj_grp_size)
{
	/* Valid sizes for an adjacency group are:
	 * 1-64, 512, 1024, 2048 and 4096.
	 */
	if (*p_adj_grp_size <= 64)
		return;
	else if (*p_adj_grp_size <= 512)
		*p_adj_grp_size = 512;
	else if (*p_adj_grp_size <= 1024)
		*p_adj_grp_size = 1024;
	else if (*p_adj_grp_size <= 2048)
		*p_adj_grp_size = 2048;
	else
		*p_adj_grp_size = 4096;
}

static void mlxsw_sp_adj_grp_size_round_down(u16 *p_adj_grp_size,
					     unsigned int alloc_size)
{
	if (alloc_size >= 4096)
		*p_adj_grp_size = 4096;
	else if (alloc_size >= 2048)
		*p_adj_grp_size = 2048;
	else if (alloc_size >= 1024)
		*p_adj_grp_size = 1024;
	else if (alloc_size >= 512)
		*p_adj_grp_size = 512;
}

static int mlxsw_sp_fix_adj_grp_size(struct mlxsw_sp *mlxsw_sp,
				     u16 *p_adj_grp_size)
{
	unsigned int alloc_size;
	int err;

	/* Round up the requested group size to the next size supported
	 * by the device and make sure the request can be satisfied.
	 */
	mlxsw_sp_adj_grp_size_round_up(p_adj_grp_size);
3309 3310 3311
	err = mlxsw_sp_kvdl_alloc_count_query(mlxsw_sp,
					      MLXSW_SP_KVDL_ENTRY_TYPE_ADJ,
					      *p_adj_grp_size, &alloc_size);
3312 3313 3314 3315 3316 3317 3318 3319 3320 3321 3322
	if (err)
		return err;
	/* It is possible the allocation results in more allocated
	 * entries than requested. Try to use as much of them as
	 * possible.
	 */
	mlxsw_sp_adj_grp_size_round_down(p_adj_grp_size, alloc_size);

	return 0;
}

3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371
static void
mlxsw_sp_nexthop_group_normalize(struct mlxsw_sp_nexthop_group *nh_grp)
{
	int i, g = 0, sum_norm_weight = 0;
	struct mlxsw_sp_nexthop *nh;

	for (i = 0; i < nh_grp->count; i++) {
		nh = &nh_grp->nexthops[i];

		if (!nh->should_offload)
			continue;
		if (g > 0)
			g = gcd(nh->nh_weight, g);
		else
			g = nh->nh_weight;
	}

	for (i = 0; i < nh_grp->count; i++) {
		nh = &nh_grp->nexthops[i];

		if (!nh->should_offload)
			continue;
		nh->norm_nh_weight = nh->nh_weight / g;
		sum_norm_weight += nh->norm_nh_weight;
	}

	nh_grp->sum_norm_weight = sum_norm_weight;
}

static void
mlxsw_sp_nexthop_group_rebalance(struct mlxsw_sp_nexthop_group *nh_grp)
{
	int total = nh_grp->sum_norm_weight;
	u16 ecmp_size = nh_grp->ecmp_size;
	int i, weight = 0, lower_bound = 0;

	for (i = 0; i < nh_grp->count; i++) {
		struct mlxsw_sp_nexthop *nh = &nh_grp->nexthops[i];
		int upper_bound;

		if (!nh->should_offload)
			continue;
		weight += nh->norm_nh_weight;
		upper_bound = DIV_ROUND_CLOSEST(ecmp_size * weight, total);
		nh->num_adj_entries = upper_bound - lower_bound;
		lower_bound = upper_bound;
	}
}

3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438
static struct mlxsw_sp_nexthop *
mlxsw_sp_rt6_nexthop(struct mlxsw_sp_nexthop_group *nh_grp,
		     const struct mlxsw_sp_rt6 *mlxsw_sp_rt6);

static void
mlxsw_sp_nexthop4_group_offload_refresh(struct mlxsw_sp *mlxsw_sp,
					struct mlxsw_sp_nexthop_group *nh_grp)
{
	int i;

	for (i = 0; i < nh_grp->count; i++) {
		struct mlxsw_sp_nexthop *nh = &nh_grp->nexthops[i];

		if (nh->offloaded)
			nh->key.fib_nh->fib_nh_flags |= RTNH_F_OFFLOAD;
		else
			nh->key.fib_nh->fib_nh_flags &= ~RTNH_F_OFFLOAD;
	}
}

static void
__mlxsw_sp_nexthop6_group_offload_refresh(struct mlxsw_sp_nexthop_group *nh_grp,
					  struct mlxsw_sp_fib6_entry *fib6_entry)
{
	struct mlxsw_sp_rt6 *mlxsw_sp_rt6;

	list_for_each_entry(mlxsw_sp_rt6, &fib6_entry->rt6_list, list) {
		struct fib6_nh *fib6_nh = mlxsw_sp_rt6->rt->fib6_nh;
		struct mlxsw_sp_nexthop *nh;

		nh = mlxsw_sp_rt6_nexthop(nh_grp, mlxsw_sp_rt6);
		if (nh && nh->offloaded)
			fib6_nh->fib_nh_flags |= RTNH_F_OFFLOAD;
		else
			fib6_nh->fib_nh_flags &= ~RTNH_F_OFFLOAD;
	}
}

static void
mlxsw_sp_nexthop6_group_offload_refresh(struct mlxsw_sp *mlxsw_sp,
					struct mlxsw_sp_nexthop_group *nh_grp)
{
	struct mlxsw_sp_fib6_entry *fib6_entry;

	/* Unfortunately, in IPv6 the route and the nexthop are described by
	 * the same struct, so we need to iterate over all the routes using the
	 * nexthop group and set / clear the offload indication for them.
	 */
	list_for_each_entry(fib6_entry, &nh_grp->fib_list,
			    common.nexthop_group_node)
		__mlxsw_sp_nexthop6_group_offload_refresh(nh_grp, fib6_entry);
}

static void
mlxsw_sp_nexthop_group_offload_refresh(struct mlxsw_sp *mlxsw_sp,
				       struct mlxsw_sp_nexthop_group *nh_grp)
{
	switch (mlxsw_sp_nexthop_group_type(nh_grp)) {
	case AF_INET:
		mlxsw_sp_nexthop4_group_offload_refresh(mlxsw_sp, nh_grp);
		break;
	case AF_INET6:
		mlxsw_sp_nexthop6_group_offload_refresh(mlxsw_sp, nh_grp);
		break;
	}
}

3439 3440 3441 3442
static void
mlxsw_sp_nexthop_group_refresh(struct mlxsw_sp *mlxsw_sp,
			       struct mlxsw_sp_nexthop_group *nh_grp)
{
3443
	u16 ecmp_size, old_ecmp_size;
3444 3445 3446 3447 3448 3449 3450 3451
	struct mlxsw_sp_nexthop *nh;
	bool offload_change = false;
	u32 adj_index;
	bool old_adj_index_valid;
	u32 old_adj_index;
	int i;
	int err;

3452 3453 3454 3455 3456
	if (!nh_grp->gateway) {
		mlxsw_sp_nexthop_fib_entries_update(mlxsw_sp, nh_grp);
		return;
	}

3457 3458 3459
	for (i = 0; i < nh_grp->count; i++) {
		nh = &nh_grp->nexthops[i];

3460
		if (nh->should_offload != nh->offloaded) {
3461 3462 3463 3464 3465 3466 3467 3468 3469
			offload_change = true;
			if (nh->should_offload)
				nh->update = 1;
		}
	}
	if (!offload_change) {
		/* Nothing was added or removed, so no need to reallocate. Just
		 * update MAC on existing adjacency indexes.
		 */
3470
		err = mlxsw_sp_nexthop_group_update(mlxsw_sp, nh_grp, false);
3471 3472 3473 3474 3475 3476
		if (err) {
			dev_warn(mlxsw_sp->bus_info->dev, "Failed to update neigh MAC in adjacency table.\n");
			goto set_trap;
		}
		return;
	}
3477 3478
	mlxsw_sp_nexthop_group_normalize(nh_grp);
	if (!nh_grp->sum_norm_weight)
3479 3480 3481 3482 3483
		/* No neigh of this group is connected so we just set
		 * the trap and let everthing flow through kernel.
		 */
		goto set_trap;

3484
	ecmp_size = nh_grp->sum_norm_weight;
3485 3486 3487 3488 3489
	err = mlxsw_sp_fix_adj_grp_size(mlxsw_sp, &ecmp_size);
	if (err)
		/* No valid allocation size available. */
		goto set_trap;

3490 3491
	err = mlxsw_sp_kvdl_alloc(mlxsw_sp, MLXSW_SP_KVDL_ENTRY_TYPE_ADJ,
				  ecmp_size, &adj_index);
3492
	if (err) {
3493 3494 3495 3496 3497 3498 3499 3500 3501 3502 3503 3504
		/* We ran out of KVD linear space, just set the
		 * trap and let everything flow through kernel.
		 */
		dev_warn(mlxsw_sp->bus_info->dev, "Failed to allocate KVD linear area for nexthop group.\n");
		goto set_trap;
	}
	old_adj_index_valid = nh_grp->adj_index_valid;
	old_adj_index = nh_grp->adj_index;
	old_ecmp_size = nh_grp->ecmp_size;
	nh_grp->adj_index_valid = 1;
	nh_grp->adj_index = adj_index;
	nh_grp->ecmp_size = ecmp_size;
3505
	mlxsw_sp_nexthop_group_rebalance(nh_grp);
3506
	err = mlxsw_sp_nexthop_group_update(mlxsw_sp, nh_grp, true);
3507 3508 3509 3510 3511
	if (err) {
		dev_warn(mlxsw_sp->bus_info->dev, "Failed to update neigh MAC in adjacency table.\n");
		goto set_trap;
	}

3512 3513
	mlxsw_sp_nexthop_group_offload_refresh(mlxsw_sp, nh_grp);

3514 3515 3516 3517 3518 3519 3520 3521 3522 3523 3524 3525 3526 3527
	if (!old_adj_index_valid) {
		/* The trap was set for fib entries, so we have to call
		 * fib entry update to unset it and use adjacency index.
		 */
		err = mlxsw_sp_nexthop_fib_entries_update(mlxsw_sp, nh_grp);
		if (err) {
			dev_warn(mlxsw_sp->bus_info->dev, "Failed to add adjacency index to fib entries.\n");
			goto set_trap;
		}
		return;
	}

	err = mlxsw_sp_adj_index_mass_update(mlxsw_sp, nh_grp,
					     old_adj_index, old_ecmp_size);
3528
	mlxsw_sp_kvdl_free(mlxsw_sp, MLXSW_SP_KVDL_ENTRY_TYPE_ADJ,
3529
			   old_ecmp_size, old_adj_index);
3530 3531 3532 3533
	if (err) {
		dev_warn(mlxsw_sp->bus_info->dev, "Failed to mass-update adjacency index for nexthop group.\n");
		goto set_trap;
	}
3534

3535 3536 3537 3538 3539 3540 3541 3542 3543 3544 3545 3546
	return;

set_trap:
	old_adj_index_valid = nh_grp->adj_index_valid;
	nh_grp->adj_index_valid = 0;
	for (i = 0; i < nh_grp->count; i++) {
		nh = &nh_grp->nexthops[i];
		nh->offloaded = 0;
	}
	err = mlxsw_sp_nexthop_fib_entries_update(mlxsw_sp, nh_grp);
	if (err)
		dev_warn(mlxsw_sp->bus_info->dev, "Failed to set traps for fib entries.\n");
3547
	mlxsw_sp_nexthop_group_offload_refresh(mlxsw_sp, nh_grp);
3548
	if (old_adj_index_valid)
3549
		mlxsw_sp_kvdl_free(mlxsw_sp, MLXSW_SP_KVDL_ENTRY_TYPE_ADJ,
3550
				   nh_grp->ecmp_size, nh_grp->adj_index);
3551 3552 3553 3554 3555
}

static void __mlxsw_sp_nexthop_neigh_update(struct mlxsw_sp_nexthop *nh,
					    bool removing)
{
3556
	if (!removing)
3557
		nh->should_offload = 1;
3558
	else
3559 3560 3561 3562
		nh->should_offload = 0;
	nh->update = 1;
}

3563 3564 3565 3566 3567 3568 3569 3570 3571 3572 3573 3574 3575 3576 3577 3578 3579 3580 3581 3582 3583 3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 3612 3613 3614 3615
static int
mlxsw_sp_nexthop_dead_neigh_replace(struct mlxsw_sp *mlxsw_sp,
				    struct mlxsw_sp_neigh_entry *neigh_entry)
{
	struct neighbour *n, *old_n = neigh_entry->key.n;
	struct mlxsw_sp_nexthop *nh;
	bool entry_connected;
	u8 nud_state, dead;
	int err;

	nh = list_first_entry(&neigh_entry->nexthop_list,
			      struct mlxsw_sp_nexthop, neigh_list_node);

	n = neigh_lookup(nh->nh_grp->neigh_tbl, &nh->gw_addr, nh->rif->dev);
	if (!n) {
		n = neigh_create(nh->nh_grp->neigh_tbl, &nh->gw_addr,
				 nh->rif->dev);
		if (IS_ERR(n))
			return PTR_ERR(n);
		neigh_event_send(n, NULL);
	}

	mlxsw_sp_neigh_entry_remove(mlxsw_sp, neigh_entry);
	neigh_entry->key.n = n;
	err = mlxsw_sp_neigh_entry_insert(mlxsw_sp, neigh_entry);
	if (err)
		goto err_neigh_entry_insert;

	read_lock_bh(&n->lock);
	nud_state = n->nud_state;
	dead = n->dead;
	read_unlock_bh(&n->lock);
	entry_connected = nud_state & NUD_VALID && !dead;

	list_for_each_entry(nh, &neigh_entry->nexthop_list,
			    neigh_list_node) {
		neigh_release(old_n);
		neigh_clone(n);
		__mlxsw_sp_nexthop_neigh_update(nh, !entry_connected);
		mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh->nh_grp);
	}

	neigh_release(n);

	return 0;

err_neigh_entry_insert:
	neigh_entry->key.n = old_n;
	mlxsw_sp_neigh_entry_insert(mlxsw_sp, neigh_entry);
	neigh_release(n);
	return err;
}

3616 3617 3618
static void
mlxsw_sp_nexthop_neigh_update(struct mlxsw_sp *mlxsw_sp,
			      struct mlxsw_sp_neigh_entry *neigh_entry,
3619
			      bool removing, bool dead)
3620 3621 3622
{
	struct mlxsw_sp_nexthop *nh;

3623 3624 3625 3626 3627 3628 3629 3630 3631 3632 3633 3634 3635
	if (list_empty(&neigh_entry->nexthop_list))
		return;

	if (dead) {
		int err;

		err = mlxsw_sp_nexthop_dead_neigh_replace(mlxsw_sp,
							  neigh_entry);
		if (err)
			dev_err(mlxsw_sp->bus_info->dev, "Failed to replace dead neigh\n");
		return;
	}

3636 3637 3638 3639 3640 3641 3642
	list_for_each_entry(nh, &neigh_entry->nexthop_list,
			    neigh_list_node) {
		__mlxsw_sp_nexthop_neigh_update(nh, removing);
		mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh->nh_grp);
	}
}

3643
static void mlxsw_sp_nexthop_rif_init(struct mlxsw_sp_nexthop *nh,
3644
				      struct mlxsw_sp_rif *rif)
3645
{
3646
	if (nh->rif)
3647 3648
		return;

3649 3650
	nh->rif = rif;
	list_add(&nh->rif_list_node, &rif->nexthop_list);
3651 3652 3653 3654
}

static void mlxsw_sp_nexthop_rif_fini(struct mlxsw_sp_nexthop *nh)
{
3655
	if (!nh->rif)
3656 3657 3658
		return;

	list_del(&nh->rif_list_node);
3659
	nh->rif = NULL;
3660 3661
}

3662 3663
static int mlxsw_sp_nexthop_neigh_init(struct mlxsw_sp *mlxsw_sp,
				       struct mlxsw_sp_nexthop *nh)
3664 3665 3666
{
	struct mlxsw_sp_neigh_entry *neigh_entry;
	struct neighbour *n;
3667
	u8 nud_state, dead;
3668 3669
	int err;

3670
	if (!nh->nh_grp->gateway || nh->neigh_entry)
3671 3672
		return 0;

3673
	/* Take a reference of neigh here ensuring that neigh would
3674
	 * not be destructed before the nexthop entry is finished.
3675
	 * The reference is taken either in neigh_lookup() or
3676
	 * in neigh_create() in case n is not found.
3677
	 */
3678
	n = neigh_lookup(nh->nh_grp->neigh_tbl, &nh->gw_addr, nh->rif->dev);
3679
	if (!n) {
3680 3681
		n = neigh_create(nh->nh_grp->neigh_tbl, &nh->gw_addr,
				 nh->rif->dev);
3682 3683
		if (IS_ERR(n))
			return PTR_ERR(n);
3684
		neigh_event_send(n, NULL);
3685 3686 3687
	}
	neigh_entry = mlxsw_sp_neigh_entry_lookup(mlxsw_sp, n);
	if (!neigh_entry) {
3688 3689
		neigh_entry = mlxsw_sp_neigh_entry_create(mlxsw_sp, n);
		if (IS_ERR(neigh_entry)) {
3690 3691
			err = -EINVAL;
			goto err_neigh_entry_create;
3692
		}
3693
	}
3694 3695 3696 3697 3698 3699

	/* If that is the first nexthop connected to that neigh, add to
	 * nexthop_neighs_list
	 */
	if (list_empty(&neigh_entry->nexthop_list))
		list_add_tail(&neigh_entry->nexthop_neighs_list_node,
3700
			      &mlxsw_sp->router->nexthop_neighs_list);
3701

3702 3703 3704 3705
	nh->neigh_entry = neigh_entry;
	list_add_tail(&nh->neigh_list_node, &neigh_entry->nexthop_list);
	read_lock_bh(&n->lock);
	nud_state = n->nud_state;
3706
	dead = n->dead;
3707
	read_unlock_bh(&n->lock);
3708
	__mlxsw_sp_nexthop_neigh_update(nh, !(nud_state & NUD_VALID && !dead));
3709 3710

	return 0;
3711 3712 3713 3714

err_neigh_entry_create:
	neigh_release(n);
	return err;
3715 3716
}

3717 3718
static void mlxsw_sp_nexthop_neigh_fini(struct mlxsw_sp *mlxsw_sp,
					struct mlxsw_sp_nexthop *nh)
3719 3720
{
	struct mlxsw_sp_neigh_entry *neigh_entry = nh->neigh_entry;
3721
	struct neighbour *n;
3722

3723
	if (!neigh_entry)
3724 3725
		return;
	n = neigh_entry->key.n;
3726

3727
	__mlxsw_sp_nexthop_neigh_update(nh, true);
3728
	list_del(&nh->neigh_list_node);
3729
	nh->neigh_entry = NULL;
3730 3731 3732 3733

	/* If that is the last nexthop connected to that neigh, remove from
	 * nexthop_neighs_list
	 */
3734 3735
	if (list_empty(&neigh_entry->nexthop_list))
		list_del(&neigh_entry->nexthop_neighs_list_node);
3736

3737 3738 3739 3740
	if (!neigh_entry->connected && list_empty(&neigh_entry->nexthop_list))
		mlxsw_sp_neigh_entry_destroy(mlxsw_sp, neigh_entry);

	neigh_release(n);
3741
}
3742

3743 3744
static bool mlxsw_sp_ipip_netdev_ul_up(struct net_device *ol_dev)
{
3745 3746 3747 3748 3749 3750 3751
	struct net_device *ul_dev;
	bool is_up;

	rcu_read_lock();
	ul_dev = __mlxsw_sp_ipip_netdev_ul_dev_get(ol_dev);
	is_up = ul_dev ? (ul_dev->flags & IFF_UP) : true;
	rcu_read_unlock();
3752

3753
	return is_up;
3754 3755
}

3756 3757 3758
static void mlxsw_sp_nexthop_ipip_init(struct mlxsw_sp *mlxsw_sp,
				       struct mlxsw_sp_nexthop *nh,
				       struct mlxsw_sp_ipip_entry *ipip_entry)
3759
{
3760 3761
	bool removing;

3762
	if (!nh->nh_grp->gateway || nh->ipip_entry)
3763
		return;
3764

3765 3766
	nh->ipip_entry = ipip_entry;
	removing = !mlxsw_sp_ipip_netdev_ul_up(ipip_entry->ol_dev);
3767
	__mlxsw_sp_nexthop_neigh_update(nh, removing);
3768
	mlxsw_sp_nexthop_rif_init(nh, &ipip_entry->ol_lb->common);
3769 3770 3771 3772 3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786
}

static void mlxsw_sp_nexthop_ipip_fini(struct mlxsw_sp *mlxsw_sp,
				       struct mlxsw_sp_nexthop *nh)
{
	struct mlxsw_sp_ipip_entry *ipip_entry = nh->ipip_entry;

	if (!ipip_entry)
		return;

	__mlxsw_sp_nexthop_neigh_update(nh, true);
	nh->ipip_entry = NULL;
}

static bool mlxsw_sp_nexthop4_ipip_type(const struct mlxsw_sp *mlxsw_sp,
					const struct fib_nh *fib_nh,
					enum mlxsw_sp_ipip_type *p_ipipt)
{
D
David Ahern 已提交
3787
	struct net_device *dev = fib_nh->fib_nh_dev;
3788 3789 3790 3791 3792 3793

	return dev &&
	       fib_nh->nh_parent->fib_type == RTN_UNICAST &&
	       mlxsw_sp_netdev_ipip_type(mlxsw_sp, dev, p_ipipt);
}

3794 3795 3796 3797 3798 3799 3800 3801
static void mlxsw_sp_nexthop_type_fini(struct mlxsw_sp *mlxsw_sp,
				       struct mlxsw_sp_nexthop *nh)
{
	switch (nh->type) {
	case MLXSW_SP_NEXTHOP_TYPE_ETH:
		mlxsw_sp_nexthop_neigh_fini(mlxsw_sp, nh);
		mlxsw_sp_nexthop_rif_fini(nh);
		break;
3802
	case MLXSW_SP_NEXTHOP_TYPE_IPIP:
3803
		mlxsw_sp_nexthop_rif_fini(nh);
3804 3805
		mlxsw_sp_nexthop_ipip_fini(mlxsw_sp, nh);
		break;
3806 3807 3808 3809 3810 3811 3812
	}
}

static int mlxsw_sp_nexthop4_type_init(struct mlxsw_sp *mlxsw_sp,
				       struct mlxsw_sp_nexthop *nh,
				       struct fib_nh *fib_nh)
{
3813
	const struct mlxsw_sp_ipip_ops *ipip_ops;
D
David Ahern 已提交
3814
	struct net_device *dev = fib_nh->fib_nh_dev;
3815
	struct mlxsw_sp_ipip_entry *ipip_entry;
3816 3817 3818
	struct mlxsw_sp_rif *rif;
	int err;

3819 3820 3821 3822 3823 3824 3825 3826 3827
	ipip_entry = mlxsw_sp_ipip_entry_find_by_ol_dev(mlxsw_sp, dev);
	if (ipip_entry) {
		ipip_ops = mlxsw_sp->router->ipip_ops_arr[ipip_entry->ipipt];
		if (ipip_ops->can_offload(mlxsw_sp, dev,
					  MLXSW_SP_L3_PROTO_IPV4)) {
			nh->type = MLXSW_SP_NEXTHOP_TYPE_IPIP;
			mlxsw_sp_nexthop_ipip_init(mlxsw_sp, nh, ipip_entry);
			return 0;
		}
3828 3829
	}

3830 3831 3832 3833 3834 3835 3836 3837 3838 3839 3840 3841 3842 3843 3844 3845 3846 3847 3848 3849 3850 3851 3852
	nh->type = MLXSW_SP_NEXTHOP_TYPE_ETH;
	rif = mlxsw_sp_rif_find_by_dev(mlxsw_sp, dev);
	if (!rif)
		return 0;

	mlxsw_sp_nexthop_rif_init(nh, rif);
	err = mlxsw_sp_nexthop_neigh_init(mlxsw_sp, nh);
	if (err)
		goto err_neigh_init;

	return 0;

err_neigh_init:
	mlxsw_sp_nexthop_rif_fini(nh);
	return err;
}

static void mlxsw_sp_nexthop4_type_fini(struct mlxsw_sp *mlxsw_sp,
					struct mlxsw_sp_nexthop *nh)
{
	mlxsw_sp_nexthop_type_fini(mlxsw_sp, nh);
}

3853 3854 3855 3856
static int mlxsw_sp_nexthop4_init(struct mlxsw_sp *mlxsw_sp,
				  struct mlxsw_sp_nexthop_group *nh_grp,
				  struct mlxsw_sp_nexthop *nh,
				  struct fib_nh *fib_nh)
3857
{
D
David Ahern 已提交
3858
	struct net_device *dev = fib_nh->fib_nh_dev;
3859
	struct in_device *in_dev;
3860 3861 3862 3863
	int err;

	nh->nh_grp = nh_grp;
	nh->key.fib_nh = fib_nh;
3864
#ifdef CONFIG_IP_ROUTE_MULTIPATH
D
David Ahern 已提交
3865
	nh->nh_weight = fib_nh->fib_nh_weight;
3866 3867 3868
#else
	nh->nh_weight = 1;
#endif
D
David Ahern 已提交
3869
	memcpy(&nh->gw_addr, &fib_nh->fib_nh_gw4, sizeof(fib_nh->fib_nh_gw4));
3870 3871 3872 3873
	err = mlxsw_sp_nexthop_insert(mlxsw_sp, nh);
	if (err)
		return err;

3874
	mlxsw_sp_nexthop_counter_alloc(mlxsw_sp, nh);
3875 3876
	list_add_tail(&nh->router_list_node, &mlxsw_sp->router->nexthop_list);

3877 3878 3879
	if (!dev)
		return 0;

3880 3881
	rcu_read_lock();
	in_dev = __in_dev_get_rcu(dev);
3882
	if (in_dev && IN_DEV_IGNORE_ROUTES_WITH_LINKDOWN(in_dev) &&
3883 3884
	    fib_nh->fib_nh_flags & RTNH_F_LINKDOWN) {
		rcu_read_unlock();
3885
		return 0;
3886 3887
	}
	rcu_read_unlock();
3888

3889
	err = mlxsw_sp_nexthop4_type_init(mlxsw_sp, nh, fib_nh);
3890 3891 3892 3893 3894 3895 3896 3897 3898 3899
	if (err)
		goto err_nexthop_neigh_init;

	return 0;

err_nexthop_neigh_init:
	mlxsw_sp_nexthop_remove(mlxsw_sp, nh);
	return err;
}

3900 3901
static void mlxsw_sp_nexthop4_fini(struct mlxsw_sp *mlxsw_sp,
				   struct mlxsw_sp_nexthop *nh)
3902
{
3903
	mlxsw_sp_nexthop4_type_fini(mlxsw_sp, nh);
3904
	list_del(&nh->router_list_node);
3905
	mlxsw_sp_nexthop_counter_free(mlxsw_sp, nh);
3906
	mlxsw_sp_nexthop_remove(mlxsw_sp, nh);
3907 3908
}

3909 3910
static void mlxsw_sp_nexthop4_event(struct mlxsw_sp *mlxsw_sp,
				    unsigned long event, struct fib_nh *fib_nh)
3911 3912 3913 3914
{
	struct mlxsw_sp_nexthop_key key;
	struct mlxsw_sp_nexthop *nh;

3915
	if (mlxsw_sp->router->aborted)
3916 3917 3918 3919
		return;

	key.fib_nh = fib_nh;
	nh = mlxsw_sp_nexthop_lookup(mlxsw_sp, key);
3920
	if (!nh)
3921 3922 3923 3924
		return;

	switch (event) {
	case FIB_EVENT_NH_ADD:
3925
		mlxsw_sp_nexthop4_type_init(mlxsw_sp, nh, fib_nh);
3926 3927
		break;
	case FIB_EVENT_NH_DEL:
3928
		mlxsw_sp_nexthop4_type_fini(mlxsw_sp, nh);
3929 3930 3931 3932 3933 3934
		break;
	}

	mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh->nh_grp);
}

3935 3936 3937 3938
static void mlxsw_sp_nexthop_rif_update(struct mlxsw_sp *mlxsw_sp,
					struct mlxsw_sp_rif *rif)
{
	struct mlxsw_sp_nexthop *nh;
3939
	bool removing;
3940 3941

	list_for_each_entry(nh, &rif->nexthop_list, rif_list_node) {
3942 3943 3944 3945 3946 3947 3948 3949 3950 3951 3952 3953 3954
		switch (nh->type) {
		case MLXSW_SP_NEXTHOP_TYPE_ETH:
			removing = false;
			break;
		case MLXSW_SP_NEXTHOP_TYPE_IPIP:
			removing = !mlxsw_sp_ipip_netdev_ul_up(rif->dev);
			break;
		default:
			WARN_ON(1);
			continue;
		}

		__mlxsw_sp_nexthop_neigh_update(nh, removing);
3955 3956 3957 3958
		mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh->nh_grp);
	}
}

3959 3960 3961 3962 3963 3964 3965 3966 3967 3968 3969 3970
static void mlxsw_sp_nexthop_rif_migrate(struct mlxsw_sp *mlxsw_sp,
					 struct mlxsw_sp_rif *old_rif,
					 struct mlxsw_sp_rif *new_rif)
{
	struct mlxsw_sp_nexthop *nh;

	list_splice_init(&old_rif->nexthop_list, &new_rif->nexthop_list);
	list_for_each_entry(nh, &new_rif->nexthop_list, rif_list_node)
		nh->rif = new_rif;
	mlxsw_sp_nexthop_rif_update(mlxsw_sp, new_rif);
}

3971
static void mlxsw_sp_nexthop_rif_gone_sync(struct mlxsw_sp *mlxsw_sp,
3972
					   struct mlxsw_sp_rif *rif)
3973 3974 3975
{
	struct mlxsw_sp_nexthop *nh, *tmp;

3976
	list_for_each_entry_safe(nh, tmp, &rif->nexthop_list, rif_list_node) {
3977
		mlxsw_sp_nexthop_type_fini(mlxsw_sp, nh);
3978 3979 3980 3981
		mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh->nh_grp);
	}
}

3982
static bool mlxsw_sp_fi_is_gateway(const struct mlxsw_sp *mlxsw_sp,
3983
				   struct fib_info *fi)
3984
{
3985 3986 3987 3988
	const struct fib_nh *nh = fib_info_nh(fi, 0);

	return nh->fib_nh_scope == RT_SCOPE_LINK ||
	       mlxsw_sp_nexthop4_ipip_type(mlxsw_sp, nh, NULL);
3989 3990
}

3991
static struct mlxsw_sp_nexthop_group *
3992
mlxsw_sp_nexthop4_group_create(struct mlxsw_sp *mlxsw_sp, struct fib_info *fi)
3993
{
3994
	unsigned int nhs = fib_info_num_path(fi);
3995 3996 3997 3998 3999 4000
	struct mlxsw_sp_nexthop_group *nh_grp;
	struct mlxsw_sp_nexthop *nh;
	struct fib_nh *fib_nh;
	int i;
	int err;

4001
	nh_grp = kzalloc(struct_size(nh_grp, nexthops, nhs), GFP_KERNEL);
4002 4003
	if (!nh_grp)
		return ERR_PTR(-ENOMEM);
4004
	nh_grp->priv = fi;
4005
	INIT_LIST_HEAD(&nh_grp->fib_list);
4006 4007
	nh_grp->neigh_tbl = &arp_tbl;

4008
	nh_grp->gateway = mlxsw_sp_fi_is_gateway(mlxsw_sp, fi);
4009
	nh_grp->count = nhs;
4010
	fib_info_hold(fi);
4011 4012
	for (i = 0; i < nh_grp->count; i++) {
		nh = &nh_grp->nexthops[i];
4013
		fib_nh = fib_info_nh(fi, i);
4014
		err = mlxsw_sp_nexthop4_init(mlxsw_sp, nh_grp, nh, fib_nh);
4015
		if (err)
4016
			goto err_nexthop4_init;
4017
	}
4018 4019 4020
	err = mlxsw_sp_nexthop_group_insert(mlxsw_sp, nh_grp);
	if (err)
		goto err_nexthop_group_insert;
4021 4022 4023
	mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh_grp);
	return nh_grp;

4024
err_nexthop_group_insert:
4025
err_nexthop4_init:
4026 4027
	for (i--; i >= 0; i--) {
		nh = &nh_grp->nexthops[i];
4028
		mlxsw_sp_nexthop4_fini(mlxsw_sp, nh);
4029
	}
4030
	fib_info_put(fi);
4031 4032 4033 4034 4035
	kfree(nh_grp);
	return ERR_PTR(err);
}

static void
4036 4037
mlxsw_sp_nexthop4_group_destroy(struct mlxsw_sp *mlxsw_sp,
				struct mlxsw_sp_nexthop_group *nh_grp)
4038 4039 4040 4041
{
	struct mlxsw_sp_nexthop *nh;
	int i;

4042
	mlxsw_sp_nexthop_group_remove(mlxsw_sp, nh_grp);
4043 4044
	for (i = 0; i < nh_grp->count; i++) {
		nh = &nh_grp->nexthops[i];
4045
		mlxsw_sp_nexthop4_fini(mlxsw_sp, nh);
4046
	}
4047 4048
	mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh_grp);
	WARN_ON_ONCE(nh_grp->adj_index_valid);
4049
	fib_info_put(mlxsw_sp_nexthop4_group_fi(nh_grp));
4050 4051 4052
	kfree(nh_grp);
}

4053 4054 4055
static int mlxsw_sp_nexthop4_group_get(struct mlxsw_sp *mlxsw_sp,
				       struct mlxsw_sp_fib_entry *fib_entry,
				       struct fib_info *fi)
4056 4057 4058
{
	struct mlxsw_sp_nexthop_group *nh_grp;

4059
	nh_grp = mlxsw_sp_nexthop4_group_lookup(mlxsw_sp, fi);
4060
	if (!nh_grp) {
4061
		nh_grp = mlxsw_sp_nexthop4_group_create(mlxsw_sp, fi);
4062 4063 4064 4065 4066 4067 4068 4069
		if (IS_ERR(nh_grp))
			return PTR_ERR(nh_grp);
	}
	list_add_tail(&fib_entry->nexthop_group_node, &nh_grp->fib_list);
	fib_entry->nh_group = nh_grp;
	return 0;
}

4070 4071
static void mlxsw_sp_nexthop4_group_put(struct mlxsw_sp *mlxsw_sp,
					struct mlxsw_sp_fib_entry *fib_entry)
4072 4073 4074 4075 4076 4077
{
	struct mlxsw_sp_nexthop_group *nh_grp = fib_entry->nh_group;

	list_del(&fib_entry->nexthop_group_node);
	if (!list_empty(&nh_grp->fib_list))
		return;
4078
	mlxsw_sp_nexthop4_group_destroy(mlxsw_sp, nh_grp);
4079 4080
}

4081 4082 4083 4084 4085 4086 4087 4088 4089 4090
static bool
mlxsw_sp_fib4_entry_should_offload(const struct mlxsw_sp_fib_entry *fib_entry)
{
	struct mlxsw_sp_fib4_entry *fib4_entry;

	fib4_entry = container_of(fib_entry, struct mlxsw_sp_fib4_entry,
				  common);
	return !fib4_entry->tos;
}

4091 4092 4093 4094 4095
static bool
mlxsw_sp_fib_entry_should_offload(const struct mlxsw_sp_fib_entry *fib_entry)
{
	struct mlxsw_sp_nexthop_group *nh_group = fib_entry->nh_group;

4096 4097 4098 4099 4100 4101 4102 4103
	switch (fib_entry->fib_node->fib->proto) {
	case MLXSW_SP_L3_PROTO_IPV4:
		if (!mlxsw_sp_fib4_entry_should_offload(fib_entry))
			return false;
		break;
	case MLXSW_SP_L3_PROTO_IPV6:
		break;
	}
4104

4105 4106 4107 4108
	switch (fib_entry->type) {
	case MLXSW_SP_FIB_ENTRY_TYPE_REMOTE:
		return !!nh_group->adj_index_valid;
	case MLXSW_SP_FIB_ENTRY_TYPE_LOCAL:
4109
		return !!nh_group->nh_rif;
4110
	case MLXSW_SP_FIB_ENTRY_TYPE_BLACKHOLE:
4111
	case MLXSW_SP_FIB_ENTRY_TYPE_IPIP_DECAP:
4112
	case MLXSW_SP_FIB_ENTRY_TYPE_NVE_DECAP:
4113
		return true;
4114 4115 4116 4117 4118
	default:
		return false;
	}
}

4119 4120 4121 4122 4123 4124 4125 4126
static struct mlxsw_sp_nexthop *
mlxsw_sp_rt6_nexthop(struct mlxsw_sp_nexthop_group *nh_grp,
		     const struct mlxsw_sp_rt6 *mlxsw_sp_rt6)
{
	int i;

	for (i = 0; i < nh_grp->count; i++) {
		struct mlxsw_sp_nexthop *nh = &nh_grp->nexthops[i];
4127
		struct fib6_info *rt = mlxsw_sp_rt6->rt;
4128

4129
		if (nh->rif && nh->rif->dev == rt->fib6_nh->fib_nh_dev &&
4130
		    ipv6_addr_equal((const struct in6_addr *) &nh->gw_addr,
4131
				    &rt->fib6_nh->fib_nh_gw6))
4132 4133 4134 4135 4136 4137 4138
			return nh;
		continue;
	}

	return NULL;
}

4139
static void
4140 4141
mlxsw_sp_fib4_entry_hw_flags_set(struct mlxsw_sp *mlxsw_sp,
				 struct mlxsw_sp_fib_entry *fib_entry)
4142
{
4143 4144 4145 4146 4147 4148
	struct fib_info *fi = mlxsw_sp_nexthop4_group_fi(fib_entry->nh_group);
	u32 *p_dst = (u32 *) fib_entry->fib_node->key.addr;
	int dst_len = fib_entry->fib_node->key.prefix_len;
	struct mlxsw_sp_fib4_entry *fib4_entry;
	struct fib_rt_info fri;
	bool should_offload;
4149

4150 4151 4152 4153 4154 4155 4156 4157 4158 4159 4160 4161
	should_offload = mlxsw_sp_fib_entry_should_offload(fib_entry);
	fib4_entry = container_of(fib_entry, struct mlxsw_sp_fib4_entry,
				  common);
	fri.fi = fi;
	fri.tb_id = fib4_entry->tb_id;
	fri.dst = cpu_to_be32(*p_dst);
	fri.dst_len = dst_len;
	fri.tos = fib4_entry->tos;
	fri.type = fib4_entry->type;
	fri.offload = should_offload;
	fri.trap = !should_offload;
	fib_alias_hw_flags_set(mlxsw_sp_net(mlxsw_sp), &fri);
4162 4163 4164
}

static void
4165 4166
mlxsw_sp_fib4_entry_hw_flags_clear(struct mlxsw_sp *mlxsw_sp,
				   struct mlxsw_sp_fib_entry *fib_entry)
4167
{
4168 4169 4170 4171 4172
	struct fib_info *fi = mlxsw_sp_nexthop4_group_fi(fib_entry->nh_group);
	u32 *p_dst = (u32 *) fib_entry->fib_node->key.addr;
	int dst_len = fib_entry->fib_node->key.prefix_len;
	struct mlxsw_sp_fib4_entry *fib4_entry;
	struct fib_rt_info fri;
4173

4174 4175 4176 4177 4178 4179 4180 4181 4182 4183 4184
	fib4_entry = container_of(fib_entry, struct mlxsw_sp_fib4_entry,
				  common);
	fri.fi = fi;
	fri.tb_id = fib4_entry->tb_id;
	fri.dst = cpu_to_be32(*p_dst);
	fri.dst_len = dst_len;
	fri.tos = fib4_entry->tos;
	fri.type = fib4_entry->type;
	fri.offload = false;
	fri.trap = false;
	fib_alias_hw_flags_set(mlxsw_sp_net(mlxsw_sp), &fri);
4185 4186
}

4187
static void
4188 4189
mlxsw_sp_fib6_entry_hw_flags_set(struct mlxsw_sp *mlxsw_sp,
				 struct mlxsw_sp_fib_entry *fib_entry)
4190 4191 4192
{
	struct mlxsw_sp_fib6_entry *fib6_entry;
	struct mlxsw_sp_rt6 *mlxsw_sp_rt6;
4193 4194 4195
	bool should_offload;

	should_offload = mlxsw_sp_fib_entry_should_offload(fib_entry);
4196

4197 4198 4199
	/* In IPv6 a multipath route is represented using multiple routes, so
	 * we need to set the flags on all of them.
	 */
4200 4201
	fib6_entry = container_of(fib_entry, struct mlxsw_sp_fib6_entry,
				  common);
4202 4203 4204
	list_for_each_entry(mlxsw_sp_rt6, &fib6_entry->rt6_list, list)
		fib6_info_hw_flags_set(mlxsw_sp_rt6->rt, should_offload,
				       !should_offload);
4205 4206 4207
}

static void
4208 4209
mlxsw_sp_fib6_entry_hw_flags_clear(struct mlxsw_sp *mlxsw_sp,
				   struct mlxsw_sp_fib_entry *fib_entry)
4210 4211 4212 4213 4214 4215
{
	struct mlxsw_sp_fib6_entry *fib6_entry;
	struct mlxsw_sp_rt6 *mlxsw_sp_rt6;

	fib6_entry = container_of(fib_entry, struct mlxsw_sp_fib6_entry,
				  common);
4216 4217
	list_for_each_entry(mlxsw_sp_rt6, &fib6_entry->rt6_list, list)
		fib6_info_hw_flags_set(mlxsw_sp_rt6->rt, false, false);
4218 4219
}

4220 4221 4222
static void
mlxsw_sp_fib_entry_hw_flags_set(struct mlxsw_sp *mlxsw_sp,
				struct mlxsw_sp_fib_entry *fib_entry)
4223
{
4224
	switch (fib_entry->fib_node->fib->proto) {
4225
	case MLXSW_SP_L3_PROTO_IPV4:
4226
		mlxsw_sp_fib4_entry_hw_flags_set(mlxsw_sp, fib_entry);
4227 4228
		break;
	case MLXSW_SP_L3_PROTO_IPV6:
4229
		mlxsw_sp_fib6_entry_hw_flags_set(mlxsw_sp, fib_entry);
4230
		break;
4231 4232 4233 4234
	}
}

static void
4235 4236
mlxsw_sp_fib_entry_hw_flags_clear(struct mlxsw_sp *mlxsw_sp,
				  struct mlxsw_sp_fib_entry *fib_entry)
4237
{
4238
	switch (fib_entry->fib_node->fib->proto) {
4239
	case MLXSW_SP_L3_PROTO_IPV4:
4240
		mlxsw_sp_fib4_entry_hw_flags_clear(mlxsw_sp, fib_entry);
4241 4242
		break;
	case MLXSW_SP_L3_PROTO_IPV6:
4243
		mlxsw_sp_fib6_entry_hw_flags_clear(mlxsw_sp, fib_entry);
4244
		break;
4245 4246 4247 4248
	}
}

static void
4249 4250 4251
mlxsw_sp_fib_entry_hw_flags_refresh(struct mlxsw_sp *mlxsw_sp,
				    struct mlxsw_sp_fib_entry *fib_entry,
				    enum mlxsw_reg_ralue_op op)
4252 4253 4254
{
	switch (op) {
	case MLXSW_REG_RALUE_OP_WRITE_WRITE:
4255 4256 4257 4258 4259
		mlxsw_sp_fib_entry_hw_flags_set(mlxsw_sp, fib_entry);
		break;
	case MLXSW_REG_RALUE_OP_WRITE_DELETE:
		mlxsw_sp_fib_entry_hw_flags_clear(mlxsw_sp, fib_entry);
		break;
4260
	default:
4261
		break;
4262 4263 4264
	}
}

4265 4266 4267 4268
static void
mlxsw_sp_fib_entry_ralue_pack(char *ralue_pl,
			      const struct mlxsw_sp_fib_entry *fib_entry,
			      enum mlxsw_reg_ralue_op op)
4269
{
4270
	struct mlxsw_sp_fib *fib = fib_entry->fib_node->fib;
4271 4272 4273 4274 4275 4276 4277 4278 4279 4280 4281 4282 4283 4284 4285 4286 4287 4288 4289 4290
	enum mlxsw_reg_ralxx_protocol proto;
	u32 *p_dip;

	proto = (enum mlxsw_reg_ralxx_protocol) fib->proto;

	switch (fib->proto) {
	case MLXSW_SP_L3_PROTO_IPV4:
		p_dip = (u32 *) fib_entry->fib_node->key.addr;
		mlxsw_reg_ralue_pack4(ralue_pl, proto, op, fib->vr->id,
				      fib_entry->fib_node->key.prefix_len,
				      *p_dip);
		break;
	case MLXSW_SP_L3_PROTO_IPV6:
		mlxsw_reg_ralue_pack6(ralue_pl, proto, op, fib->vr->id,
				      fib_entry->fib_node->key.prefix_len,
				      fib_entry->fib_node->key.addr);
		break;
	}
}

4291 4292 4293 4294
static int mlxsw_sp_adj_discard_write(struct mlxsw_sp *mlxsw_sp, u16 rif_index)
{
	enum mlxsw_reg_ratr_trap_action trap_action;
	char ratr_pl[MLXSW_REG_RATR_LEN];
4295 4296 4297 4298 4299 4300 4301 4302 4303
	int err;

	if (mlxsw_sp->router->adj_discard_index_valid)
		return 0;

	err = mlxsw_sp_kvdl_alloc(mlxsw_sp, MLXSW_SP_KVDL_ENTRY_TYPE_ADJ, 1,
				  &mlxsw_sp->router->adj_discard_index);
	if (err)
		return err;
4304 4305 4306

	trap_action = MLXSW_REG_RATR_TRAP_ACTION_DISCARD_ERRORS;
	mlxsw_reg_ratr_pack(ratr_pl, MLXSW_REG_RATR_OP_WRITE_WRITE_ENTRY, true,
4307 4308
			    MLXSW_REG_RATR_TYPE_ETHERNET,
			    mlxsw_sp->router->adj_discard_index, rif_index);
4309
	mlxsw_reg_ratr_trap_action_set(ratr_pl, trap_action);
4310 4311 4312 4313 4314 4315 4316 4317 4318 4319 4320 4321
	err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ratr), ratr_pl);
	if (err)
		goto err_ratr_write;

	mlxsw_sp->router->adj_discard_index_valid = true;

	return 0;

err_ratr_write:
	mlxsw_sp_kvdl_free(mlxsw_sp, MLXSW_SP_KVDL_ENTRY_TYPE_ADJ, 1,
			   mlxsw_sp->router->adj_discard_index);
	return err;
4322 4323
}

4324 4325 4326 4327
static int mlxsw_sp_fib_entry_op_remote(struct mlxsw_sp *mlxsw_sp,
					struct mlxsw_sp_fib_entry *fib_entry,
					enum mlxsw_reg_ralue_op op)
{
4328
	struct mlxsw_sp_nexthop_group *nh_group = fib_entry->nh_group;
4329
	char ralue_pl[MLXSW_REG_RALUE_LEN];
4330 4331 4332 4333
	enum mlxsw_reg_ralue_trap_action trap_action;
	u16 trap_id = 0;
	u32 adjacency_index = 0;
	u16 ecmp_size = 0;
4334
	int err;
4335 4336 4337 4338 4339

	/* In case the nexthop group adjacency index is valid, use it
	 * with provided ECMP size. Otherwise, setup trap and pass
	 * traffic to kernel.
	 */
4340
	if (mlxsw_sp_fib_entry_should_offload(fib_entry)) {
4341 4342 4343
		trap_action = MLXSW_REG_RALUE_TRAP_ACTION_NOP;
		adjacency_index = fib_entry->nh_group->adj_index;
		ecmp_size = fib_entry->nh_group->ecmp_size;
4344 4345 4346 4347 4348 4349 4350 4351 4352
	} else if (!nh_group->adj_index_valid && nh_group->count &&
		   nh_group->nh_rif) {
		err = mlxsw_sp_adj_discard_write(mlxsw_sp,
						 nh_group->nh_rif->rif_index);
		if (err)
			return err;
		trap_action = MLXSW_REG_RALUE_TRAP_ACTION_NOP;
		adjacency_index = mlxsw_sp->router->adj_discard_index;
		ecmp_size = 1;
4353 4354 4355 4356 4357
	} else {
		trap_action = MLXSW_REG_RALUE_TRAP_ACTION_TRAP;
		trap_id = MLXSW_TRAP_ID_RTR_INGRESS0;
	}

4358
	mlxsw_sp_fib_entry_ralue_pack(ralue_pl, fib_entry, op);
4359 4360 4361 4362 4363
	mlxsw_reg_ralue_act_remote_pack(ralue_pl, trap_action, trap_id,
					adjacency_index, ecmp_size);
	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ralue), ralue_pl);
}

4364 4365 4366
static int mlxsw_sp_fib_entry_op_local(struct mlxsw_sp *mlxsw_sp,
				       struct mlxsw_sp_fib_entry *fib_entry,
				       enum mlxsw_reg_ralue_op op)
4367
{
4368
	struct mlxsw_sp_rif *rif = fib_entry->nh_group->nh_rif;
4369
	enum mlxsw_reg_ralue_trap_action trap_action;
4370
	char ralue_pl[MLXSW_REG_RALUE_LEN];
4371
	u16 trap_id = 0;
4372
	u16 rif_index = 0;
4373 4374 4375

	if (mlxsw_sp_fib_entry_should_offload(fib_entry)) {
		trap_action = MLXSW_REG_RALUE_TRAP_ACTION_NOP;
4376
		rif_index = rif->rif_index;
4377 4378 4379 4380
	} else {
		trap_action = MLXSW_REG_RALUE_TRAP_ACTION_TRAP;
		trap_id = MLXSW_TRAP_ID_RTR_INGRESS0;
	}
4381

4382
	mlxsw_sp_fib_entry_ralue_pack(ralue_pl, fib_entry, op);
4383 4384
	mlxsw_reg_ralue_act_local_pack(ralue_pl, trap_action, trap_id,
				       rif_index);
4385 4386 4387
	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ralue), ralue_pl);
}

4388 4389 4390
static int mlxsw_sp_fib_entry_op_trap(struct mlxsw_sp *mlxsw_sp,
				      struct mlxsw_sp_fib_entry *fib_entry,
				      enum mlxsw_reg_ralue_op op)
4391 4392 4393
{
	char ralue_pl[MLXSW_REG_RALUE_LEN];

4394
	mlxsw_sp_fib_entry_ralue_pack(ralue_pl, fib_entry, op);
4395 4396 4397 4398
	mlxsw_reg_ralue_act_ip2me_pack(ralue_pl);
	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ralue), ralue_pl);
}

4399 4400 4401 4402 4403 4404 4405 4406 4407 4408 4409 4410 4411
static int mlxsw_sp_fib_entry_op_blackhole(struct mlxsw_sp *mlxsw_sp,
					   struct mlxsw_sp_fib_entry *fib_entry,
					   enum mlxsw_reg_ralue_op op)
{
	enum mlxsw_reg_ralue_trap_action trap_action;
	char ralue_pl[MLXSW_REG_RALUE_LEN];

	trap_action = MLXSW_REG_RALUE_TRAP_ACTION_DISCARD_ERROR;
	mlxsw_sp_fib_entry_ralue_pack(ralue_pl, fib_entry, op);
	mlxsw_reg_ralue_act_local_pack(ralue_pl, trap_action, 0, 0);
	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ralue), ralue_pl);
}

4412 4413 4414 4415 4416 4417 4418 4419 4420 4421 4422 4423 4424 4425 4426 4427 4428
static int
mlxsw_sp_fib_entry_op_unreachable(struct mlxsw_sp *mlxsw_sp,
				  struct mlxsw_sp_fib_entry *fib_entry,
				  enum mlxsw_reg_ralue_op op)
{
	enum mlxsw_reg_ralue_trap_action trap_action;
	char ralue_pl[MLXSW_REG_RALUE_LEN];
	u16 trap_id;

	trap_action = MLXSW_REG_RALUE_TRAP_ACTION_TRAP;
	trap_id = MLXSW_TRAP_ID_RTR_INGRESS1;

	mlxsw_sp_fib_entry_ralue_pack(ralue_pl, fib_entry, op);
	mlxsw_reg_ralue_act_local_pack(ralue_pl, trap_action, trap_id, 0);
	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ralue), ralue_pl);
}

4429 4430 4431 4432 4433 4434 4435 4436 4437 4438 4439 4440 4441 4442 4443 4444
static int
mlxsw_sp_fib_entry_op_ipip_decap(struct mlxsw_sp *mlxsw_sp,
				 struct mlxsw_sp_fib_entry *fib_entry,
				 enum mlxsw_reg_ralue_op op)
{
	struct mlxsw_sp_ipip_entry *ipip_entry = fib_entry->decap.ipip_entry;
	const struct mlxsw_sp_ipip_ops *ipip_ops;

	if (WARN_ON(!ipip_entry))
		return -EINVAL;

	ipip_ops = mlxsw_sp->router->ipip_ops_arr[ipip_entry->ipipt];
	return ipip_ops->fib_entry_op(mlxsw_sp, ipip_entry, op,
				      fib_entry->decap.tunnel_index);
}

4445 4446 4447 4448 4449 4450 4451 4452 4453 4454 4455 4456
static int mlxsw_sp_fib_entry_op_nve_decap(struct mlxsw_sp *mlxsw_sp,
					   struct mlxsw_sp_fib_entry *fib_entry,
					   enum mlxsw_reg_ralue_op op)
{
	char ralue_pl[MLXSW_REG_RALUE_LEN];

	mlxsw_sp_fib_entry_ralue_pack(ralue_pl, fib_entry, op);
	mlxsw_reg_ralue_act_ip2me_tun_pack(ralue_pl,
					   fib_entry->decap.tunnel_index);
	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ralue), ralue_pl);
}

4457 4458 4459
static int __mlxsw_sp_fib_entry_op(struct mlxsw_sp *mlxsw_sp,
				   struct mlxsw_sp_fib_entry *fib_entry,
				   enum mlxsw_reg_ralue_op op)
4460 4461 4462
{
	switch (fib_entry->type) {
	case MLXSW_SP_FIB_ENTRY_TYPE_REMOTE:
4463
		return mlxsw_sp_fib_entry_op_remote(mlxsw_sp, fib_entry, op);
4464
	case MLXSW_SP_FIB_ENTRY_TYPE_LOCAL:
4465
		return mlxsw_sp_fib_entry_op_local(mlxsw_sp, fib_entry, op);
4466
	case MLXSW_SP_FIB_ENTRY_TYPE_TRAP:
4467
		return mlxsw_sp_fib_entry_op_trap(mlxsw_sp, fib_entry, op);
4468 4469
	case MLXSW_SP_FIB_ENTRY_TYPE_BLACKHOLE:
		return mlxsw_sp_fib_entry_op_blackhole(mlxsw_sp, fib_entry, op);
4470 4471 4472
	case MLXSW_SP_FIB_ENTRY_TYPE_UNREACHABLE:
		return mlxsw_sp_fib_entry_op_unreachable(mlxsw_sp, fib_entry,
							 op);
4473 4474 4475
	case MLXSW_SP_FIB_ENTRY_TYPE_IPIP_DECAP:
		return mlxsw_sp_fib_entry_op_ipip_decap(mlxsw_sp,
							fib_entry, op);
4476 4477
	case MLXSW_SP_FIB_ENTRY_TYPE_NVE_DECAP:
		return mlxsw_sp_fib_entry_op_nve_decap(mlxsw_sp, fib_entry, op);
4478 4479 4480 4481 4482 4483 4484 4485
	}
	return -EINVAL;
}

static int mlxsw_sp_fib_entry_op(struct mlxsw_sp *mlxsw_sp,
				 struct mlxsw_sp_fib_entry *fib_entry,
				 enum mlxsw_reg_ralue_op op)
{
4486
	int err = __mlxsw_sp_fib_entry_op(mlxsw_sp, fib_entry, op);
4487

4488 4489 4490 4491
	if (err)
		return err;

	mlxsw_sp_fib_entry_hw_flags_refresh(mlxsw_sp, fib_entry, op);
4492

4493
	return err;
4494 4495 4496 4497 4498
}

static int mlxsw_sp_fib_entry_update(struct mlxsw_sp *mlxsw_sp,
				     struct mlxsw_sp_fib_entry *fib_entry)
{
4499 4500
	return mlxsw_sp_fib_entry_op(mlxsw_sp, fib_entry,
				     MLXSW_REG_RALUE_OP_WRITE_WRITE);
4501 4502 4503 4504 4505 4506 4507 4508 4509 4510
}

static int mlxsw_sp_fib_entry_del(struct mlxsw_sp *mlxsw_sp,
				  struct mlxsw_sp_fib_entry *fib_entry)
{
	return mlxsw_sp_fib_entry_op(mlxsw_sp, fib_entry,
				     MLXSW_REG_RALUE_OP_WRITE_DELETE);
}

static int
4511 4512 4513
mlxsw_sp_fib4_entry_type_set(struct mlxsw_sp *mlxsw_sp,
			     const struct fib_entry_notifier_info *fen_info,
			     struct mlxsw_sp_fib_entry *fib_entry)
4514
{
4515
	struct net_device *dev = fib_info_nh(fen_info->fi, 0)->fib_nh_dev;
4516
	union mlxsw_sp_l3addr dip = { .addr4 = htonl(fen_info->dst) };
4517
	struct mlxsw_sp_router *router = mlxsw_sp->router;
4518
	u32 tb_id = mlxsw_sp_fix_tb_id(fen_info->tb_id);
4519
	struct mlxsw_sp_ipip_entry *ipip_entry;
4520
	struct fib_info *fi = fen_info->fi;
4521

4522 4523
	switch (fen_info->type) {
	case RTN_LOCAL:
4524 4525
		ipip_entry = mlxsw_sp_ipip_entry_find_by_decap(mlxsw_sp, dev,
						 MLXSW_SP_L3_PROTO_IPV4, dip);
4526
		if (ipip_entry && ipip_entry->ol_dev->flags & IFF_UP) {
4527 4528 4529 4530 4531
			fib_entry->type = MLXSW_SP_FIB_ENTRY_TYPE_IPIP_DECAP;
			return mlxsw_sp_fib_entry_decap_init(mlxsw_sp,
							     fib_entry,
							     ipip_entry);
		}
4532 4533 4534 4535
		if (mlxsw_sp_router_nve_is_decap(mlxsw_sp, tb_id,
						 MLXSW_SP_L3_PROTO_IPV4,
						 &dip)) {
			u32 tunnel_index;
4536

4537 4538
			tunnel_index = router->nve_decap_config.tunnel_index;
			fib_entry->decap.tunnel_index = tunnel_index;
4539 4540 4541
			fib_entry->type = MLXSW_SP_FIB_ENTRY_TYPE_NVE_DECAP;
			return 0;
		}
4542 4543
		/* fall through */
	case RTN_BROADCAST:
4544 4545
		fib_entry->type = MLXSW_SP_FIB_ENTRY_TYPE_TRAP;
		return 0;
4546 4547 4548
	case RTN_BLACKHOLE:
		fib_entry->type = MLXSW_SP_FIB_ENTRY_TYPE_BLACKHOLE;
		return 0;
4549 4550 4551 4552 4553 4554
	case RTN_UNREACHABLE: /* fall through */
	case RTN_PROHIBIT:
		/* Packets hitting these routes need to be trapped, but
		 * can do so with a lower priority than packets directed
		 * at the host, so use action type local instead of trap.
		 */
4555
		fib_entry->type = MLXSW_SP_FIB_ENTRY_TYPE_UNREACHABLE;
4556 4557
		return 0;
	case RTN_UNICAST:
4558
		if (mlxsw_sp_fi_is_gateway(mlxsw_sp, fi))
4559
			fib_entry->type = MLXSW_SP_FIB_ENTRY_TYPE_REMOTE;
4560 4561
		else
			fib_entry->type = MLXSW_SP_FIB_ENTRY_TYPE_LOCAL;
4562 4563 4564 4565
		return 0;
	default:
		return -EINVAL;
	}
4566 4567
}

4568 4569 4570 4571 4572 4573 4574 4575 4576 4577 4578 4579 4580
static void
mlxsw_sp_fib4_entry_type_unset(struct mlxsw_sp *mlxsw_sp,
			       struct mlxsw_sp_fib_entry *fib_entry)
{
	switch (fib_entry->type) {
	case MLXSW_SP_FIB_ENTRY_TYPE_IPIP_DECAP:
		mlxsw_sp_fib_entry_decap_fini(mlxsw_sp, fib_entry);
		break;
	default:
		break;
	}
}

4581
static struct mlxsw_sp_fib4_entry *
4582 4583 4584
mlxsw_sp_fib4_entry_create(struct mlxsw_sp *mlxsw_sp,
			   struct mlxsw_sp_fib_node *fib_node,
			   const struct fib_entry_notifier_info *fen_info)
4585
{
4586
	struct mlxsw_sp_fib4_entry *fib4_entry;
4587 4588 4589
	struct mlxsw_sp_fib_entry *fib_entry;
	int err;

4590 4591 4592 4593
	fib4_entry = kzalloc(sizeof(*fib4_entry), GFP_KERNEL);
	if (!fib4_entry)
		return ERR_PTR(-ENOMEM);
	fib_entry = &fib4_entry->common;
4594

4595
	err = mlxsw_sp_fib4_entry_type_set(mlxsw_sp, fen_info, fib_entry);
4596
	if (err)
4597
		goto err_fib4_entry_type_set;
4598

4599
	err = mlxsw_sp_nexthop4_group_get(mlxsw_sp, fib_entry, fen_info->fi);
4600
	if (err)
4601
		goto err_nexthop4_group_get;
4602

4603 4604 4605 4606
	fib4_entry->prio = fen_info->fi->fib_priority;
	fib4_entry->tb_id = fen_info->tb_id;
	fib4_entry->type = fen_info->type;
	fib4_entry->tos = fen_info->tos;
4607 4608 4609

	fib_entry->fib_node = fib_node;

4610
	return fib4_entry;
4611

4612
err_nexthop4_group_get:
4613
	mlxsw_sp_fib4_entry_type_unset(mlxsw_sp, fib_entry);
4614
err_fib4_entry_type_set:
4615
	kfree(fib4_entry);
4616 4617 4618
	return ERR_PTR(err);
}

4619
static void mlxsw_sp_fib4_entry_destroy(struct mlxsw_sp *mlxsw_sp,
4620
					struct mlxsw_sp_fib4_entry *fib4_entry)
4621
{
4622
	mlxsw_sp_nexthop4_group_put(mlxsw_sp, &fib4_entry->common);
4623
	mlxsw_sp_fib4_entry_type_unset(mlxsw_sp, &fib4_entry->common);
4624
	kfree(fib4_entry);
4625 4626
}

4627
static struct mlxsw_sp_fib4_entry *
4628 4629
mlxsw_sp_fib4_entry_lookup(struct mlxsw_sp *mlxsw_sp,
			   const struct fib_entry_notifier_info *fen_info)
4630
{
4631
	struct mlxsw_sp_fib4_entry *fib4_entry;
4632
	struct mlxsw_sp_fib_node *fib_node;
4633 4634 4635 4636 4637 4638 4639
	struct mlxsw_sp_fib *fib;
	struct mlxsw_sp_vr *vr;

	vr = mlxsw_sp_vr_find(mlxsw_sp, fen_info->tb_id);
	if (!vr)
		return NULL;
	fib = mlxsw_sp_vr_fib(vr, MLXSW_SP_L3_PROTO_IPV4);
4640

4641 4642 4643 4644
	fib_node = mlxsw_sp_fib_node_lookup(fib, &fen_info->dst,
					    sizeof(fen_info->dst),
					    fen_info->dst_len);
	if (!fib_node)
4645 4646
		return NULL;

4647 4648 4649 4650 4651 4652 4653 4654
	fib4_entry = container_of(fib_node->fib_entry,
				  struct mlxsw_sp_fib4_entry, common);
	if (fib4_entry->tb_id == fen_info->tb_id &&
	    fib4_entry->tos == fen_info->tos &&
	    fib4_entry->type == fen_info->type &&
	    mlxsw_sp_nexthop4_group_fi(fib4_entry->common.nh_group) ==
	    fen_info->fi)
		return fib4_entry;
4655 4656 4657 4658 4659 4660 4661 4662 4663 4664 4665 4666 4667 4668 4669 4670 4671 4672 4673 4674 4675 4676 4677 4678 4679 4680 4681 4682 4683 4684 4685 4686 4687 4688 4689 4690 4691 4692

	return NULL;
}

static const struct rhashtable_params mlxsw_sp_fib_ht_params = {
	.key_offset = offsetof(struct mlxsw_sp_fib_node, key),
	.head_offset = offsetof(struct mlxsw_sp_fib_node, ht_node),
	.key_len = sizeof(struct mlxsw_sp_fib_key),
	.automatic_shrinking = true,
};

static int mlxsw_sp_fib_node_insert(struct mlxsw_sp_fib *fib,
				    struct mlxsw_sp_fib_node *fib_node)
{
	return rhashtable_insert_fast(&fib->ht, &fib_node->ht_node,
				      mlxsw_sp_fib_ht_params);
}

static void mlxsw_sp_fib_node_remove(struct mlxsw_sp_fib *fib,
				     struct mlxsw_sp_fib_node *fib_node)
{
	rhashtable_remove_fast(&fib->ht, &fib_node->ht_node,
			       mlxsw_sp_fib_ht_params);
}

static struct mlxsw_sp_fib_node *
mlxsw_sp_fib_node_lookup(struct mlxsw_sp_fib *fib, const void *addr,
			 size_t addr_len, unsigned char prefix_len)
{
	struct mlxsw_sp_fib_key key;

	memset(&key, 0, sizeof(key));
	memcpy(key.addr, addr, addr_len);
	key.prefix_len = prefix_len;
	return rhashtable_lookup_fast(&fib->ht, &key, mlxsw_sp_fib_ht_params);
}

static struct mlxsw_sp_fib_node *
4693
mlxsw_sp_fib_node_create(struct mlxsw_sp_fib *fib, const void *addr,
4694 4695 4696 4697 4698 4699
			 size_t addr_len, unsigned char prefix_len)
{
	struct mlxsw_sp_fib_node *fib_node;

	fib_node = kzalloc(sizeof(*fib_node), GFP_KERNEL);
	if (!fib_node)
4700 4701
		return NULL;

4702
	list_add(&fib_node->list, &fib->node_list);
4703 4704 4705 4706 4707 4708 4709 4710 4711 4712 4713 4714
	memcpy(fib_node->key.addr, addr, addr_len);
	fib_node->key.prefix_len = prefix_len;

	return fib_node;
}

static void mlxsw_sp_fib_node_destroy(struct mlxsw_sp_fib_node *fib_node)
{
	list_del(&fib_node->list);
	kfree(fib_node);
}

4715 4716 4717
static int mlxsw_sp_fib_lpm_tree_link(struct mlxsw_sp *mlxsw_sp,
				      struct mlxsw_sp_fib_node *fib_node)
{
4718
	struct mlxsw_sp_prefix_usage req_prefix_usage;
4719
	struct mlxsw_sp_fib *fib = fib_node->fib;
4720 4721 4722
	struct mlxsw_sp_lpm_tree *lpm_tree;
	int err;

4723 4724 4725
	lpm_tree = mlxsw_sp->router->lpm.proto_trees[fib->proto];
	if (lpm_tree->prefix_ref_count[fib_node->key.prefix_len] != 0)
		goto out;
4726

4727 4728
	mlxsw_sp_prefix_usage_cpy(&req_prefix_usage, &lpm_tree->prefix_usage);
	mlxsw_sp_prefix_usage_set(&req_prefix_usage, fib_node->key.prefix_len);
4729 4730 4731 4732 4733 4734 4735
	lpm_tree = mlxsw_sp_lpm_tree_get(mlxsw_sp, &req_prefix_usage,
					 fib->proto);
	if (IS_ERR(lpm_tree))
		return PTR_ERR(lpm_tree);

	err = mlxsw_sp_vrs_lpm_tree_replace(mlxsw_sp, fib, lpm_tree);
	if (err)
4736
		goto err_lpm_tree_replace;
4737

4738 4739
out:
	lpm_tree->prefix_ref_count[fib_node->key.prefix_len]++;
4740
	return 0;
4741 4742 4743 4744

err_lpm_tree_replace:
	mlxsw_sp_lpm_tree_put(mlxsw_sp, lpm_tree);
	return err;
4745 4746 4747
}

static void mlxsw_sp_fib_lpm_tree_unlink(struct mlxsw_sp *mlxsw_sp,
4748
					 struct mlxsw_sp_fib_node *fib_node)
4749
{
4750 4751
	struct mlxsw_sp_lpm_tree *lpm_tree = fib_node->fib->lpm_tree;
	struct mlxsw_sp_prefix_usage req_prefix_usage;
4752
	struct mlxsw_sp_fib *fib = fib_node->fib;
4753
	int err;
4754

4755
	if (--lpm_tree->prefix_ref_count[fib_node->key.prefix_len] != 0)
4756
		return;
4757 4758
	/* Try to construct a new LPM tree from the current prefix usage
	 * minus the unused one. If we fail, continue using the old one.
4759
	 */
4760 4761 4762 4763 4764 4765 4766
	mlxsw_sp_prefix_usage_cpy(&req_prefix_usage, &lpm_tree->prefix_usage);
	mlxsw_sp_prefix_usage_clear(&req_prefix_usage,
				    fib_node->key.prefix_len);
	lpm_tree = mlxsw_sp_lpm_tree_get(mlxsw_sp, &req_prefix_usage,
					 fib->proto);
	if (IS_ERR(lpm_tree))
		return;
4767

4768 4769 4770
	err = mlxsw_sp_vrs_lpm_tree_replace(mlxsw_sp, fib, lpm_tree);
	if (err)
		goto err_lpm_tree_replace;
4771

4772
	return;
4773

4774 4775
err_lpm_tree_replace:
	mlxsw_sp_lpm_tree_put(mlxsw_sp, lpm_tree);
4776 4777
}

4778 4779 4780 4781 4782 4783 4784 4785 4786 4787 4788
static int mlxsw_sp_fib_node_init(struct mlxsw_sp *mlxsw_sp,
				  struct mlxsw_sp_fib_node *fib_node,
				  struct mlxsw_sp_fib *fib)
{
	int err;

	err = mlxsw_sp_fib_node_insert(fib, fib_node);
	if (err)
		return err;
	fib_node->fib = fib;

4789
	err = mlxsw_sp_fib_lpm_tree_link(mlxsw_sp, fib_node);
4790 4791
	if (err)
		goto err_fib_lpm_tree_link;
4792 4793 4794

	return 0;

4795
err_fib_lpm_tree_link:
4796 4797 4798 4799 4800 4801 4802 4803 4804 4805
	fib_node->fib = NULL;
	mlxsw_sp_fib_node_remove(fib, fib_node);
	return err;
}

static void mlxsw_sp_fib_node_fini(struct mlxsw_sp *mlxsw_sp,
				   struct mlxsw_sp_fib_node *fib_node)
{
	struct mlxsw_sp_fib *fib = fib_node->fib;

4806
	mlxsw_sp_fib_lpm_tree_unlink(mlxsw_sp, fib_node);
4807 4808 4809 4810
	fib_node->fib = NULL;
	mlxsw_sp_fib_node_remove(fib, fib_node);
}

4811
static struct mlxsw_sp_fib_node *
4812 4813 4814
mlxsw_sp_fib_node_get(struct mlxsw_sp *mlxsw_sp, u32 tb_id, const void *addr,
		      size_t addr_len, unsigned char prefix_len,
		      enum mlxsw_sp_l3proto proto)
4815
{
4816
	struct mlxsw_sp_fib_node *fib_node;
4817
	struct mlxsw_sp_fib *fib;
4818 4819 4820
	struct mlxsw_sp_vr *vr;
	int err;

4821
	vr = mlxsw_sp_vr_get(mlxsw_sp, tb_id, NULL);
4822 4823
	if (IS_ERR(vr))
		return ERR_CAST(vr);
4824
	fib = mlxsw_sp_vr_fib(vr, proto);
4825

4826
	fib_node = mlxsw_sp_fib_node_lookup(fib, addr, addr_len, prefix_len);
4827 4828
	if (fib_node)
		return fib_node;
4829

4830
	fib_node = mlxsw_sp_fib_node_create(fib, addr, addr_len, prefix_len);
4831 4832 4833
	if (!fib_node) {
		err = -ENOMEM;
		goto err_fib_node_create;
4834
	}
4835

4836 4837 4838 4839
	err = mlxsw_sp_fib_node_init(mlxsw_sp, fib_node, fib);
	if (err)
		goto err_fib_node_init;

4840 4841
	return fib_node;

4842 4843
err_fib_node_init:
	mlxsw_sp_fib_node_destroy(fib_node);
4844
err_fib_node_create:
4845
	mlxsw_sp_vr_put(mlxsw_sp, vr);
4846
	return ERR_PTR(err);
4847 4848
}

4849 4850
static void mlxsw_sp_fib_node_put(struct mlxsw_sp *mlxsw_sp,
				  struct mlxsw_sp_fib_node *fib_node)
4851
{
4852
	struct mlxsw_sp_vr *vr = fib_node->fib->vr;
4853

4854
	if (fib_node->fib_entry)
4855
		return;
4856
	mlxsw_sp_fib_node_fini(mlxsw_sp, fib_node);
4857
	mlxsw_sp_fib_node_destroy(fib_node);
4858
	mlxsw_sp_vr_put(mlxsw_sp, vr);
4859 4860
}

4861 4862
static int mlxsw_sp_fib_node_entry_link(struct mlxsw_sp *mlxsw_sp,
					struct mlxsw_sp_fib_entry *fib_entry)
4863
{
4864
	struct mlxsw_sp_fib_node *fib_node = fib_entry->fib_node;
4865 4866
	int err;

4867
	fib_node->fib_entry = fib_entry;
4868

4869
	err = mlxsw_sp_fib_entry_update(mlxsw_sp, fib_entry);
4870
	if (err)
4871
		goto err_fib_entry_update;
4872 4873 4874

	return 0;

4875 4876
err_fib_entry_update:
	fib_node->fib_entry = NULL;
4877 4878 4879 4880
	return err;
}

static void
4881 4882
mlxsw_sp_fib_node_entry_unlink(struct mlxsw_sp *mlxsw_sp,
			       struct mlxsw_sp_fib_entry *fib_entry)
4883
{
4884
	struct mlxsw_sp_fib_node *fib_node = fib_entry->fib_node;
4885

4886 4887
	mlxsw_sp_fib_entry_del(mlxsw_sp, fib_entry);
	fib_node->fib_entry = NULL;
4888 4889
}

4890 4891 4892 4893 4894 4895 4896 4897 4898 4899 4900 4901 4902 4903 4904 4905 4906
static bool mlxsw_sp_fib4_allow_replace(struct mlxsw_sp_fib4_entry *fib4_entry)
{
	struct mlxsw_sp_fib_node *fib_node = fib4_entry->common.fib_node;
	struct mlxsw_sp_fib4_entry *fib4_replaced;

	if (!fib_node->fib_entry)
		return true;

	fib4_replaced = container_of(fib_node->fib_entry,
				     struct mlxsw_sp_fib4_entry, common);
	if (fib4_entry->tb_id == RT_TABLE_MAIN &&
	    fib4_replaced->tb_id == RT_TABLE_LOCAL)
		return false;

	return true;
}

4907
static int
4908 4909
mlxsw_sp_router_fib4_replace(struct mlxsw_sp *mlxsw_sp,
			     const struct fib_entry_notifier_info *fen_info)
4910
{
4911 4912
	struct mlxsw_sp_fib4_entry *fib4_entry, *fib4_replaced;
	struct mlxsw_sp_fib_entry *replaced;
4913
	struct mlxsw_sp_fib_node *fib_node;
4914 4915
	int err;

4916
	if (mlxsw_sp->router->aborted)
4917 4918
		return 0;

4919 4920 4921 4922
	fib_node = mlxsw_sp_fib_node_get(mlxsw_sp, fen_info->tb_id,
					 &fen_info->dst, sizeof(fen_info->dst),
					 fen_info->dst_len,
					 MLXSW_SP_L3_PROTO_IPV4);
4923 4924 4925
	if (IS_ERR(fib_node)) {
		dev_warn(mlxsw_sp->bus_info->dev, "Failed to get FIB node\n");
		return PTR_ERR(fib_node);
4926
	}
4927

4928 4929
	fib4_entry = mlxsw_sp_fib4_entry_create(mlxsw_sp, fib_node, fen_info);
	if (IS_ERR(fib4_entry)) {
4930
		dev_warn(mlxsw_sp->bus_info->dev, "Failed to create FIB entry\n");
4931
		err = PTR_ERR(fib4_entry);
4932 4933
		goto err_fib4_entry_create;
	}
4934

4935 4936 4937 4938 4939 4940
	if (!mlxsw_sp_fib4_allow_replace(fib4_entry)) {
		mlxsw_sp_fib4_entry_destroy(mlxsw_sp, fib4_entry);
		mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
		return 0;
	}

4941
	replaced = fib_node->fib_entry;
4942
	err = mlxsw_sp_fib_node_entry_link(mlxsw_sp, &fib4_entry->common);
4943
	if (err) {
4944
		dev_warn(mlxsw_sp->bus_info->dev, "Failed to link FIB entry to node\n");
4945
		goto err_fib_node_entry_link;
4946
	}
4947

4948 4949 4950 4951
	/* Nothing to replace */
	if (!replaced)
		return 0;

4952
	mlxsw_sp_fib_entry_hw_flags_clear(mlxsw_sp, replaced);
4953 4954 4955
	fib4_replaced = container_of(replaced, struct mlxsw_sp_fib4_entry,
				     common);
	mlxsw_sp_fib4_entry_destroy(mlxsw_sp, fib4_replaced);
4956

4957 4958
	return 0;

4959
err_fib_node_entry_link:
4960
	fib_node->fib_entry = replaced;
4961
	mlxsw_sp_fib4_entry_destroy(mlxsw_sp, fib4_entry);
4962
err_fib4_entry_create:
4963
	mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
4964 4965 4966
	return err;
}

4967 4968
static void mlxsw_sp_router_fib4_del(struct mlxsw_sp *mlxsw_sp,
				     struct fib_entry_notifier_info *fen_info)
4969
{
4970
	struct mlxsw_sp_fib4_entry *fib4_entry;
4971
	struct mlxsw_sp_fib_node *fib_node;
4972

4973
	if (mlxsw_sp->router->aborted)
4974
		return;
4975

4976
	fib4_entry = mlxsw_sp_fib4_entry_lookup(mlxsw_sp, fen_info);
4977
	if (!fib4_entry)
4978
		return;
4979
	fib_node = fib4_entry->common.fib_node;
4980

4981
	mlxsw_sp_fib_node_entry_unlink(mlxsw_sp, &fib4_entry->common);
4982
	mlxsw_sp_fib4_entry_destroy(mlxsw_sp, fib4_entry);
4983
	mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
4984
}
4985

4986
static bool mlxsw_sp_fib6_rt_should_ignore(const struct fib6_info *rt)
4987 4988 4989 4990 4991
{
	/* Packets with link-local destination IP arriving to the router
	 * are trapped to the CPU, so no need to program specific routes
	 * for them.
	 */
4992
	if (ipv6_addr_type(&rt->fib6_dst.addr) & IPV6_ADDR_LINKLOCAL)
4993 4994 4995 4996 4997
		return true;

	/* Multicast routes aren't supported, so ignore them. Neighbour
	 * Discovery packets are specifically trapped.
	 */
4998
	if (ipv6_addr_type(&rt->fib6_dst.addr) & IPV6_ADDR_MULTICAST)
4999 5000 5001
		return true;

	/* Cloned routes are irrelevant in the forwarding path. */
5002
	if (rt->fib6_flags & RTF_CACHE)
5003 5004 5005 5006 5007
		return true;

	return false;
}

5008
static struct mlxsw_sp_rt6 *mlxsw_sp_rt6_create(struct fib6_info *rt)
5009 5010 5011 5012 5013 5014 5015 5016 5017 5018 5019 5020
{
	struct mlxsw_sp_rt6 *mlxsw_sp_rt6;

	mlxsw_sp_rt6 = kzalloc(sizeof(*mlxsw_sp_rt6), GFP_KERNEL);
	if (!mlxsw_sp_rt6)
		return ERR_PTR(-ENOMEM);

	/* In case of route replace, replaced route is deleted with
	 * no notification. Take reference to prevent accessing freed
	 * memory.
	 */
	mlxsw_sp_rt6->rt = rt;
5021
	fib6_info_hold(rt);
5022 5023 5024 5025 5026

	return mlxsw_sp_rt6;
}

#if IS_ENABLED(CONFIG_IPV6)
5027
static void mlxsw_sp_rt6_release(struct fib6_info *rt)
5028
{
5029
	fib6_info_release(rt);
5030 5031
}
#else
5032
static void mlxsw_sp_rt6_release(struct fib6_info *rt)
5033 5034 5035 5036 5037 5038
{
}
#endif

static void mlxsw_sp_rt6_destroy(struct mlxsw_sp_rt6 *mlxsw_sp_rt6)
{
5039 5040 5041
	struct fib6_nh *fib6_nh = mlxsw_sp_rt6->rt->fib6_nh;

	fib6_nh->fib_nh_flags &= ~RTNH_F_OFFLOAD;
5042 5043 5044 5045
	mlxsw_sp_rt6_release(mlxsw_sp_rt6->rt);
	kfree(mlxsw_sp_rt6);
}

5046
static struct fib6_info *
5047 5048 5049 5050 5051 5052 5053 5054
mlxsw_sp_fib6_entry_rt(const struct mlxsw_sp_fib6_entry *fib6_entry)
{
	return list_first_entry(&fib6_entry->rt6_list, struct mlxsw_sp_rt6,
				list)->rt;
}

static struct mlxsw_sp_rt6 *
mlxsw_sp_fib6_entry_rt_find(const struct mlxsw_sp_fib6_entry *fib6_entry,
5055
			    const struct fib6_info *rt)
5056 5057 5058 5059 5060 5061 5062 5063 5064 5065 5066
{
	struct mlxsw_sp_rt6 *mlxsw_sp_rt6;

	list_for_each_entry(mlxsw_sp_rt6, &fib6_entry->rt6_list, list) {
		if (mlxsw_sp_rt6->rt == rt)
			return mlxsw_sp_rt6;
	}

	return NULL;
}

5067
static bool mlxsw_sp_nexthop6_ipip_type(const struct mlxsw_sp *mlxsw_sp,
5068
					const struct fib6_info *rt,
5069 5070
					enum mlxsw_sp_ipip_type *ret)
{
5071 5072
	return rt->fib6_nh->fib_nh_dev &&
	       mlxsw_sp_netdev_ipip_type(mlxsw_sp, rt->fib6_nh->fib_nh_dev, ret);
5073 5074
}

5075 5076 5077
static int mlxsw_sp_nexthop6_type_init(struct mlxsw_sp *mlxsw_sp,
				       struct mlxsw_sp_nexthop_group *nh_grp,
				       struct mlxsw_sp_nexthop *nh,
5078
				       const struct fib6_info *rt)
5079
{
5080 5081
	const struct mlxsw_sp_ipip_ops *ipip_ops;
	struct mlxsw_sp_ipip_entry *ipip_entry;
5082
	struct net_device *dev = rt->fib6_nh->fib_nh_dev;
5083 5084 5085
	struct mlxsw_sp_rif *rif;
	int err;

5086 5087 5088 5089 5090 5091 5092 5093 5094
	ipip_entry = mlxsw_sp_ipip_entry_find_by_ol_dev(mlxsw_sp, dev);
	if (ipip_entry) {
		ipip_ops = mlxsw_sp->router->ipip_ops_arr[ipip_entry->ipipt];
		if (ipip_ops->can_offload(mlxsw_sp, dev,
					  MLXSW_SP_L3_PROTO_IPV6)) {
			nh->type = MLXSW_SP_NEXTHOP_TYPE_IPIP;
			mlxsw_sp_nexthop_ipip_init(mlxsw_sp, nh, ipip_entry);
			return 0;
		}
5095 5096
	}

5097
	nh->type = MLXSW_SP_NEXTHOP_TYPE_ETH;
5098 5099 5100 5101 5102 5103 5104 5105 5106 5107 5108 5109 5110 5111 5112 5113
	rif = mlxsw_sp_rif_find_by_dev(mlxsw_sp, dev);
	if (!rif)
		return 0;
	mlxsw_sp_nexthop_rif_init(nh, rif);

	err = mlxsw_sp_nexthop_neigh_init(mlxsw_sp, nh);
	if (err)
		goto err_nexthop_neigh_init;

	return 0;

err_nexthop_neigh_init:
	mlxsw_sp_nexthop_rif_fini(nh);
	return err;
}

5114 5115 5116 5117 5118 5119 5120 5121 5122
static void mlxsw_sp_nexthop6_type_fini(struct mlxsw_sp *mlxsw_sp,
					struct mlxsw_sp_nexthop *nh)
{
	mlxsw_sp_nexthop_type_fini(mlxsw_sp, nh);
}

static int mlxsw_sp_nexthop6_init(struct mlxsw_sp *mlxsw_sp,
				  struct mlxsw_sp_nexthop_group *nh_grp,
				  struct mlxsw_sp_nexthop *nh,
5123
				  const struct fib6_info *rt)
5124
{
5125
	struct net_device *dev = rt->fib6_nh->fib_nh_dev;
5126 5127

	nh->nh_grp = nh_grp;
5128 5129
	nh->nh_weight = rt->fib6_nh->fib_nh_weight;
	memcpy(&nh->gw_addr, &rt->fib6_nh->fib_nh_gw6, sizeof(nh->gw_addr));
5130
	mlxsw_sp_nexthop_counter_alloc(mlxsw_sp, nh);
5131

5132 5133
	list_add_tail(&nh->router_list_node, &mlxsw_sp->router->nexthop_list);

5134 5135 5136 5137 5138 5139 5140
	if (!dev)
		return 0;
	nh->ifindex = dev->ifindex;

	return mlxsw_sp_nexthop6_type_init(mlxsw_sp, nh_grp, nh, rt);
}

5141 5142 5143
static void mlxsw_sp_nexthop6_fini(struct mlxsw_sp *mlxsw_sp,
				   struct mlxsw_sp_nexthop *nh)
{
5144
	mlxsw_sp_nexthop6_type_fini(mlxsw_sp, nh);
5145
	list_del(&nh->router_list_node);
5146
	mlxsw_sp_nexthop_counter_free(mlxsw_sp, nh);
5147 5148
}

5149
static bool mlxsw_sp_rt6_is_gateway(const struct mlxsw_sp *mlxsw_sp,
5150
				    const struct fib6_info *rt)
5151
{
5152
	return rt->fib6_nh->fib_nh_gw_family ||
5153
	       mlxsw_sp_nexthop6_ipip_type(mlxsw_sp, rt, NULL);
5154 5155
}

5156 5157 5158 5159 5160 5161 5162 5163 5164 5165
static struct mlxsw_sp_nexthop_group *
mlxsw_sp_nexthop6_group_create(struct mlxsw_sp *mlxsw_sp,
			       struct mlxsw_sp_fib6_entry *fib6_entry)
{
	struct mlxsw_sp_nexthop_group *nh_grp;
	struct mlxsw_sp_rt6 *mlxsw_sp_rt6;
	struct mlxsw_sp_nexthop *nh;
	int i = 0;
	int err;

5166 5167
	nh_grp = kzalloc(struct_size(nh_grp, nexthops, fib6_entry->nrt6),
			 GFP_KERNEL);
5168 5169 5170 5171 5172 5173 5174 5175
	if (!nh_grp)
		return ERR_PTR(-ENOMEM);
	INIT_LIST_HEAD(&nh_grp->fib_list);
#if IS_ENABLED(CONFIG_IPV6)
	nh_grp->neigh_tbl = &nd_tbl;
#endif
	mlxsw_sp_rt6 = list_first_entry(&fib6_entry->rt6_list,
					struct mlxsw_sp_rt6, list);
5176
	nh_grp->gateway = mlxsw_sp_rt6_is_gateway(mlxsw_sp, mlxsw_sp_rt6->rt);
5177 5178
	nh_grp->count = fib6_entry->nrt6;
	for (i = 0; i < nh_grp->count; i++) {
5179
		struct fib6_info *rt = mlxsw_sp_rt6->rt;
5180 5181 5182 5183 5184 5185 5186

		nh = &nh_grp->nexthops[i];
		err = mlxsw_sp_nexthop6_init(mlxsw_sp, nh_grp, nh, rt);
		if (err)
			goto err_nexthop6_init;
		mlxsw_sp_rt6 = list_next_entry(mlxsw_sp_rt6, list);
	}
5187 5188 5189 5190 5191

	err = mlxsw_sp_nexthop_group_insert(mlxsw_sp, nh_grp);
	if (err)
		goto err_nexthop_group_insert;

5192 5193 5194
	mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh_grp);
	return nh_grp;

5195
err_nexthop_group_insert:
5196 5197 5198 5199 5200 5201 5202 5203 5204 5205 5206 5207 5208 5209 5210 5211
err_nexthop6_init:
	for (i--; i >= 0; i--) {
		nh = &nh_grp->nexthops[i];
		mlxsw_sp_nexthop6_fini(mlxsw_sp, nh);
	}
	kfree(nh_grp);
	return ERR_PTR(err);
}

static void
mlxsw_sp_nexthop6_group_destroy(struct mlxsw_sp *mlxsw_sp,
				struct mlxsw_sp_nexthop_group *nh_grp)
{
	struct mlxsw_sp_nexthop *nh;
	int i = nh_grp->count;

5212
	mlxsw_sp_nexthop_group_remove(mlxsw_sp, nh_grp);
5213 5214 5215 5216 5217 5218 5219 5220 5221 5222 5223 5224 5225 5226
	for (i--; i >= 0; i--) {
		nh = &nh_grp->nexthops[i];
		mlxsw_sp_nexthop6_fini(mlxsw_sp, nh);
	}
	mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh_grp);
	WARN_ON(nh_grp->adj_index_valid);
	kfree(nh_grp);
}

static int mlxsw_sp_nexthop6_group_get(struct mlxsw_sp *mlxsw_sp,
				       struct mlxsw_sp_fib6_entry *fib6_entry)
{
	struct mlxsw_sp_nexthop_group *nh_grp;

5227 5228 5229 5230 5231 5232
	nh_grp = mlxsw_sp_nexthop6_group_lookup(mlxsw_sp, fib6_entry);
	if (!nh_grp) {
		nh_grp = mlxsw_sp_nexthop6_group_create(mlxsw_sp, fib6_entry);
		if (IS_ERR(nh_grp))
			return PTR_ERR(nh_grp);
	}
5233 5234 5235 5236 5237

	list_add_tail(&fib6_entry->common.nexthop_group_node,
		      &nh_grp->fib_list);
	fib6_entry->common.nh_group = nh_grp;

5238 5239 5240 5241 5242
	/* The route and the nexthop are described by the same struct, so we
	 * need to the update the nexthop offload indication for the new route.
	 */
	__mlxsw_sp_nexthop6_group_offload_refresh(nh_grp, fib6_entry);

5243 5244 5245 5246 5247 5248 5249 5250 5251 5252 5253 5254 5255 5256 5257 5258 5259 5260 5261 5262 5263 5264 5265 5266 5267 5268 5269 5270 5271 5272 5273 5274
	return 0;
}

static void mlxsw_sp_nexthop6_group_put(struct mlxsw_sp *mlxsw_sp,
					struct mlxsw_sp_fib_entry *fib_entry)
{
	struct mlxsw_sp_nexthop_group *nh_grp = fib_entry->nh_group;

	list_del(&fib_entry->nexthop_group_node);
	if (!list_empty(&nh_grp->fib_list))
		return;
	mlxsw_sp_nexthop6_group_destroy(mlxsw_sp, nh_grp);
}

static int
mlxsw_sp_nexthop6_group_update(struct mlxsw_sp *mlxsw_sp,
			       struct mlxsw_sp_fib6_entry *fib6_entry)
{
	struct mlxsw_sp_nexthop_group *old_nh_grp = fib6_entry->common.nh_group;
	int err;

	fib6_entry->common.nh_group = NULL;
	list_del(&fib6_entry->common.nexthop_group_node);

	err = mlxsw_sp_nexthop6_group_get(mlxsw_sp, fib6_entry);
	if (err)
		goto err_nexthop6_group_get;

	/* In case this entry is offloaded, then the adjacency index
	 * currently associated with it in the device's table is that
	 * of the old group. Start using the new one instead.
	 */
5275
	err = mlxsw_sp_fib_entry_update(mlxsw_sp, &fib6_entry->common);
5276
	if (err)
5277
		goto err_fib_entry_update;
5278 5279 5280 5281 5282 5283

	if (list_empty(&old_nh_grp->fib_list))
		mlxsw_sp_nexthop6_group_destroy(mlxsw_sp, old_nh_grp);

	return 0;

5284
err_fib_entry_update:
5285 5286 5287 5288 5289 5290 5291 5292 5293 5294 5295
	mlxsw_sp_nexthop6_group_put(mlxsw_sp, &fib6_entry->common);
err_nexthop6_group_get:
	list_add_tail(&fib6_entry->common.nexthop_group_node,
		      &old_nh_grp->fib_list);
	fib6_entry->common.nh_group = old_nh_grp;
	return err;
}

static int
mlxsw_sp_fib6_entry_nexthop_add(struct mlxsw_sp *mlxsw_sp,
				struct mlxsw_sp_fib6_entry *fib6_entry,
5296
				struct fib6_info **rt_arr, unsigned int nrt6)
5297 5298
{
	struct mlxsw_sp_rt6 *mlxsw_sp_rt6;
5299
	int err, i;
5300

5301 5302 5303 5304 5305 5306
	for (i = 0; i < nrt6; i++) {
		mlxsw_sp_rt6 = mlxsw_sp_rt6_create(rt_arr[i]);
		if (IS_ERR(mlxsw_sp_rt6)) {
			err = PTR_ERR(mlxsw_sp_rt6);
			goto err_rt6_create;
		}
5307

5308 5309 5310
		list_add_tail(&mlxsw_sp_rt6->list, &fib6_entry->rt6_list);
		fib6_entry->nrt6++;
	}
5311 5312 5313 5314 5315 5316 5317 5318

	err = mlxsw_sp_nexthop6_group_update(mlxsw_sp, fib6_entry);
	if (err)
		goto err_nexthop6_group_update;

	return 0;

err_nexthop6_group_update:
5319 5320 5321 5322 5323 5324 5325 5326 5327
	i = nrt6;
err_rt6_create:
	for (i--; i >= 0; i--) {
		fib6_entry->nrt6--;
		mlxsw_sp_rt6 = list_last_entry(&fib6_entry->rt6_list,
					       struct mlxsw_sp_rt6, list);
		list_del(&mlxsw_sp_rt6->list);
		mlxsw_sp_rt6_destroy(mlxsw_sp_rt6);
	}
5328 5329 5330 5331 5332 5333
	return err;
}

static void
mlxsw_sp_fib6_entry_nexthop_del(struct mlxsw_sp *mlxsw_sp,
				struct mlxsw_sp_fib6_entry *fib6_entry,
5334
				struct fib6_info **rt_arr, unsigned int nrt6)
5335 5336
{
	struct mlxsw_sp_rt6 *mlxsw_sp_rt6;
5337
	int i;
5338

5339 5340 5341 5342 5343 5344 5345 5346 5347 5348
	for (i = 0; i < nrt6; i++) {
		mlxsw_sp_rt6 = mlxsw_sp_fib6_entry_rt_find(fib6_entry,
							   rt_arr[i]);
		if (WARN_ON_ONCE(!mlxsw_sp_rt6))
			continue;

		fib6_entry->nrt6--;
		list_del(&mlxsw_sp_rt6->list);
		mlxsw_sp_rt6_destroy(mlxsw_sp_rt6);
	}
5349 5350 5351 5352

	mlxsw_sp_nexthop6_group_update(mlxsw_sp, fib6_entry);
}

5353 5354
static void mlxsw_sp_fib6_entry_type_set(struct mlxsw_sp *mlxsw_sp,
					 struct mlxsw_sp_fib_entry *fib_entry,
5355
					 const struct fib6_info *rt)
5356 5357 5358 5359 5360 5361 5362
{
	/* Packets hitting RTF_REJECT routes need to be discarded by the
	 * stack. We can rely on their destination device not having a
	 * RIF (it's the loopback device) and can thus use action type
	 * local, which will cause them to be trapped with a lower
	 * priority than packets that need to be locally received.
	 */
5363
	if (rt->fib6_flags & (RTF_LOCAL | RTF_ANYCAST))
5364
		fib_entry->type = MLXSW_SP_FIB_ENTRY_TYPE_TRAP;
5365 5366
	else if (rt->fib6_type == RTN_BLACKHOLE)
		fib_entry->type = MLXSW_SP_FIB_ENTRY_TYPE_BLACKHOLE;
5367
	else if (rt->fib6_flags & RTF_REJECT)
5368
		fib_entry->type = MLXSW_SP_FIB_ENTRY_TYPE_UNREACHABLE;
5369
	else if (mlxsw_sp_rt6_is_gateway(mlxsw_sp, rt))
5370 5371 5372 5373 5374 5375 5376 5377 5378 5379 5380 5381 5382 5383 5384 5385 5386 5387 5388 5389 5390
		fib_entry->type = MLXSW_SP_FIB_ENTRY_TYPE_REMOTE;
	else
		fib_entry->type = MLXSW_SP_FIB_ENTRY_TYPE_LOCAL;
}

static void
mlxsw_sp_fib6_entry_rt_destroy_all(struct mlxsw_sp_fib6_entry *fib6_entry)
{
	struct mlxsw_sp_rt6 *mlxsw_sp_rt6, *tmp;

	list_for_each_entry_safe(mlxsw_sp_rt6, tmp, &fib6_entry->rt6_list,
				 list) {
		fib6_entry->nrt6--;
		list_del(&mlxsw_sp_rt6->list);
		mlxsw_sp_rt6_destroy(mlxsw_sp_rt6);
	}
}

static struct mlxsw_sp_fib6_entry *
mlxsw_sp_fib6_entry_create(struct mlxsw_sp *mlxsw_sp,
			   struct mlxsw_sp_fib_node *fib_node,
5391
			   struct fib6_info **rt_arr, unsigned int nrt6)
5392 5393 5394 5395
{
	struct mlxsw_sp_fib6_entry *fib6_entry;
	struct mlxsw_sp_fib_entry *fib_entry;
	struct mlxsw_sp_rt6 *mlxsw_sp_rt6;
5396
	int err, i;
5397 5398 5399 5400 5401 5402

	fib6_entry = kzalloc(sizeof(*fib6_entry), GFP_KERNEL);
	if (!fib6_entry)
		return ERR_PTR(-ENOMEM);
	fib_entry = &fib6_entry->common;

5403 5404 5405 5406 5407 5408 5409 5410 5411 5412
	INIT_LIST_HEAD(&fib6_entry->rt6_list);

	for (i = 0; i < nrt6; i++) {
		mlxsw_sp_rt6 = mlxsw_sp_rt6_create(rt_arr[i]);
		if (IS_ERR(mlxsw_sp_rt6)) {
			err = PTR_ERR(mlxsw_sp_rt6);
			goto err_rt6_create;
		}
		list_add_tail(&mlxsw_sp_rt6->list, &fib6_entry->rt6_list);
		fib6_entry->nrt6++;
5413 5414
	}

5415
	mlxsw_sp_fib6_entry_type_set(mlxsw_sp, fib_entry, rt_arr[0]);
5416 5417 5418 5419 5420 5421 5422 5423 5424 5425

	err = mlxsw_sp_nexthop6_group_get(mlxsw_sp, fib6_entry);
	if (err)
		goto err_nexthop6_group_get;

	fib_entry->fib_node = fib_node;

	return fib6_entry;

err_nexthop6_group_get:
5426
	i = nrt6;
5427
err_rt6_create:
5428 5429 5430 5431 5432 5433 5434
	for (i--; i >= 0; i--) {
		fib6_entry->nrt6--;
		mlxsw_sp_rt6 = list_last_entry(&fib6_entry->rt6_list,
					       struct mlxsw_sp_rt6, list);
		list_del(&mlxsw_sp_rt6->list);
		mlxsw_sp_rt6_destroy(mlxsw_sp_rt6);
	}
5435 5436 5437 5438 5439 5440 5441 5442 5443 5444 5445 5446 5447 5448 5449
	kfree(fib6_entry);
	return ERR_PTR(err);
}

static void mlxsw_sp_fib6_entry_destroy(struct mlxsw_sp *mlxsw_sp,
					struct mlxsw_sp_fib6_entry *fib6_entry)
{
	mlxsw_sp_nexthop6_group_put(mlxsw_sp, &fib6_entry->common);
	mlxsw_sp_fib6_entry_rt_destroy_all(fib6_entry);
	WARN_ON(fib6_entry->nrt6);
	kfree(fib6_entry);
}

static struct mlxsw_sp_fib6_entry *
mlxsw_sp_fib6_entry_lookup(struct mlxsw_sp *mlxsw_sp,
5450
			   const struct fib6_info *rt)
5451 5452 5453 5454
{
	struct mlxsw_sp_fib6_entry *fib6_entry;
	struct mlxsw_sp_fib_node *fib_node;
	struct mlxsw_sp_fib *fib;
5455
	struct fib6_info *cmp_rt;
5456 5457
	struct mlxsw_sp_vr *vr;

5458
	vr = mlxsw_sp_vr_find(mlxsw_sp, rt->fib6_table->tb6_id);
5459 5460 5461 5462
	if (!vr)
		return NULL;
	fib = mlxsw_sp_vr_fib(vr, MLXSW_SP_L3_PROTO_IPV6);

5463 5464 5465
	fib_node = mlxsw_sp_fib_node_lookup(fib, &rt->fib6_dst.addr,
					    sizeof(rt->fib6_dst.addr),
					    rt->fib6_dst.plen);
5466 5467 5468
	if (!fib_node)
		return NULL;

5469 5470 5471 5472 5473 5474 5475
	fib6_entry = container_of(fib_node->fib_entry,
				  struct mlxsw_sp_fib6_entry, common);
	cmp_rt = mlxsw_sp_fib6_entry_rt(fib6_entry);
	if (rt->fib6_table->tb6_id == cmp_rt->fib6_table->tb6_id &&
	    rt->fib6_metric == cmp_rt->fib6_metric &&
	    mlxsw_sp_fib6_entry_rt_find(fib6_entry, rt))
		return fib6_entry;
5476 5477 5478 5479

	return NULL;
}

5480 5481 5482 5483 5484 5485 5486 5487 5488 5489 5490 5491 5492 5493 5494 5495 5496 5497 5498 5499 5500
static bool mlxsw_sp_fib6_allow_replace(struct mlxsw_sp_fib6_entry *fib6_entry)
{
	struct mlxsw_sp_fib_node *fib_node = fib6_entry->common.fib_node;
	struct mlxsw_sp_fib6_entry *fib6_replaced;
	struct fib6_info *rt, *rt_replaced;

	if (!fib_node->fib_entry)
		return true;

	fib6_replaced = container_of(fib_node->fib_entry,
				     struct mlxsw_sp_fib6_entry,
				     common);
	rt = mlxsw_sp_fib6_entry_rt(fib6_entry);
	rt_replaced = mlxsw_sp_fib6_entry_rt(fib6_replaced);
	if (rt->fib6_table->tb6_id == RT_TABLE_MAIN &&
	    rt_replaced->fib6_table->tb6_id == RT_TABLE_LOCAL)
		return false;

	return true;
}

5501 5502 5503
static int mlxsw_sp_router_fib6_replace(struct mlxsw_sp *mlxsw_sp,
					struct fib6_info **rt_arr,
					unsigned int nrt6)
5504
{
5505 5506
	struct mlxsw_sp_fib6_entry *fib6_entry, *fib6_replaced;
	struct mlxsw_sp_fib_entry *replaced;
5507
	struct mlxsw_sp_fib_node *fib_node;
5508
	struct fib6_info *rt = rt_arr[0];
5509 5510 5511 5512 5513
	int err;

	if (mlxsw_sp->router->aborted)
		return 0;

5514
	if (rt->fib6_src.plen)
5515 5516
		return -EINVAL;

5517 5518 5519
	if (mlxsw_sp_fib6_rt_should_ignore(rt))
		return 0;

5520 5521 5522 5523
	fib_node = mlxsw_sp_fib_node_get(mlxsw_sp, rt->fib6_table->tb6_id,
					 &rt->fib6_dst.addr,
					 sizeof(rt->fib6_dst.addr),
					 rt->fib6_dst.plen,
5524 5525 5526 5527
					 MLXSW_SP_L3_PROTO_IPV6);
	if (IS_ERR(fib_node))
		return PTR_ERR(fib_node);

5528 5529
	fib6_entry = mlxsw_sp_fib6_entry_create(mlxsw_sp, fib_node, rt_arr,
						nrt6);
5530 5531 5532 5533 5534
	if (IS_ERR(fib6_entry)) {
		err = PTR_ERR(fib6_entry);
		goto err_fib6_entry_create;
	}

5535 5536 5537 5538 5539 5540
	if (!mlxsw_sp_fib6_allow_replace(fib6_entry)) {
		mlxsw_sp_fib6_entry_destroy(mlxsw_sp, fib6_entry);
		mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
		return 0;
	}

5541
	replaced = fib_node->fib_entry;
5542
	err = mlxsw_sp_fib_node_entry_link(mlxsw_sp, &fib6_entry->common);
5543
	if (err)
5544
		goto err_fib_node_entry_link;
5545

5546 5547 5548 5549
	/* Nothing to replace */
	if (!replaced)
		return 0;

5550
	mlxsw_sp_fib_entry_hw_flags_clear(mlxsw_sp, replaced);
5551 5552 5553
	fib6_replaced = container_of(replaced, struct mlxsw_sp_fib6_entry,
				     common);
	mlxsw_sp_fib6_entry_destroy(mlxsw_sp, fib6_replaced);
5554

5555 5556
	return 0;

5557
err_fib_node_entry_link:
5558
	fib_node->fib_entry = replaced;
5559 5560
	mlxsw_sp_fib6_entry_destroy(mlxsw_sp, fib6_entry);
err_fib6_entry_create:
5561 5562 5563 5564 5565 5566 5567 5568 5569 5570 5571 5572 5573 5574 5575 5576 5577 5578 5579 5580 5581 5582 5583 5584 5585 5586 5587 5588 5589 5590
	mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
	return err;
}

static int mlxsw_sp_router_fib6_append(struct mlxsw_sp *mlxsw_sp,
				       struct fib6_info **rt_arr,
				       unsigned int nrt6)
{
	struct mlxsw_sp_fib6_entry *fib6_entry;
	struct mlxsw_sp_fib_node *fib_node;
	struct fib6_info *rt = rt_arr[0];
	int err;

	if (mlxsw_sp->router->aborted)
		return 0;

	if (rt->fib6_src.plen)
		return -EINVAL;

	if (mlxsw_sp_fib6_rt_should_ignore(rt))
		return 0;

	fib_node = mlxsw_sp_fib_node_get(mlxsw_sp, rt->fib6_table->tb6_id,
					 &rt->fib6_dst.addr,
					 sizeof(rt->fib6_dst.addr),
					 rt->fib6_dst.plen,
					 MLXSW_SP_L3_PROTO_IPV6);
	if (IS_ERR(fib_node))
		return PTR_ERR(fib_node);

5591
	if (WARN_ON_ONCE(!fib_node->fib_entry)) {
5592 5593 5594 5595
		mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
		return -EINVAL;
	}

5596 5597
	fib6_entry = container_of(fib_node->fib_entry,
				  struct mlxsw_sp_fib6_entry, common);
5598 5599 5600 5601 5602 5603 5604
	err = mlxsw_sp_fib6_entry_nexthop_add(mlxsw_sp, fib6_entry, rt_arr,
					      nrt6);
	if (err)
		goto err_fib6_entry_nexthop_add;

	return 0;

5605 5606 5607 5608 5609 5610
err_fib6_entry_nexthop_add:
	mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
	return err;
}

static void mlxsw_sp_router_fib6_del(struct mlxsw_sp *mlxsw_sp,
5611 5612
				     struct fib6_info **rt_arr,
				     unsigned int nrt6)
5613 5614 5615
{
	struct mlxsw_sp_fib6_entry *fib6_entry;
	struct mlxsw_sp_fib_node *fib_node;
5616
	struct fib6_info *rt = rt_arr[0];
5617 5618 5619 5620 5621 5622 5623

	if (mlxsw_sp->router->aborted)
		return;

	if (mlxsw_sp_fib6_rt_should_ignore(rt))
		return;

5624 5625 5626 5627 5628
	/* Multipath routes are first added to the FIB trie and only then
	 * notified. If we vetoed the addition, we will get a delete
	 * notification for a route we do not have. Therefore, do not warn if
	 * route was not found.
	 */
5629
	fib6_entry = mlxsw_sp_fib6_entry_lookup(mlxsw_sp, rt);
5630
	if (!fib6_entry)
5631 5632
		return;

5633 5634
	/* If not all the nexthops are deleted, then only reduce the nexthop
	 * group.
5635
	 */
5636 5637 5638
	if (nrt6 != fib6_entry->nrt6) {
		mlxsw_sp_fib6_entry_nexthop_del(mlxsw_sp, fib6_entry, rt_arr,
						nrt6);
5639 5640 5641 5642 5643
		return;
	}

	fib_node = fib6_entry->common.fib_node;

5644
	mlxsw_sp_fib_node_entry_unlink(mlxsw_sp, &fib6_entry->common);
5645 5646 5647 5648
	mlxsw_sp_fib6_entry_destroy(mlxsw_sp, fib6_entry);
	mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
}

5649 5650 5651
static int __mlxsw_sp_router_set_abort_trap(struct mlxsw_sp *mlxsw_sp,
					    enum mlxsw_reg_ralxx_protocol proto,
					    u8 tree_id)
5652 5653 5654
{
	char ralta_pl[MLXSW_REG_RALTA_LEN];
	char ralst_pl[MLXSW_REG_RALST_LEN];
5655
	int i, err;
5656

5657
	mlxsw_reg_ralta_pack(ralta_pl, true, proto, tree_id);
5658 5659 5660 5661
	err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ralta), ralta_pl);
	if (err)
		return err;

5662
	mlxsw_reg_ralst_pack(ralst_pl, 0xff, tree_id);
5663 5664 5665 5666
	err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ralst), ralst_pl);
	if (err)
		return err;

5667
	for (i = 0; i < MLXSW_CORE_RES_GET(mlxsw_sp->core, MAX_VRS); i++) {
5668
		struct mlxsw_sp_vr *vr = &mlxsw_sp->router->vrs[i];
5669 5670
		char raltb_pl[MLXSW_REG_RALTB_LEN];
		char ralue_pl[MLXSW_REG_RALUE_LEN];
5671

5672
		mlxsw_reg_raltb_pack(raltb_pl, vr->id, proto, tree_id);
5673 5674 5675 5676 5677
		err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(raltb),
				      raltb_pl);
		if (err)
			return err;

5678 5679
		mlxsw_reg_ralue_pack(ralue_pl, proto,
				     MLXSW_REG_RALUE_OP_WRITE_WRITE, vr->id, 0);
5680 5681 5682 5683 5684 5685 5686 5687
		mlxsw_reg_ralue_act_ip2me_pack(ralue_pl);
		err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ralue),
				      ralue_pl);
		if (err)
			return err;
	}

	return 0;
5688 5689
}

5690 5691 5692
static struct mlxsw_sp_mr_table *
mlxsw_sp_router_fibmr_family_to_table(struct mlxsw_sp_vr *vr, int family)
{
5693
	if (family == RTNL_FAMILY_IPMR)
5694
		return vr->mr_table[MLXSW_SP_L3_PROTO_IPV4];
5695 5696
	else
		return vr->mr_table[MLXSW_SP_L3_PROTO_IPV6];
5697 5698
}

5699 5700 5701 5702
static int mlxsw_sp_router_fibmr_add(struct mlxsw_sp *mlxsw_sp,
				     struct mfc_entry_notifier_info *men_info,
				     bool replace)
{
5703
	struct mlxsw_sp_mr_table *mrt;
5704 5705 5706 5707 5708
	struct mlxsw_sp_vr *vr;

	if (mlxsw_sp->router->aborted)
		return 0;

5709
	vr = mlxsw_sp_vr_get(mlxsw_sp, men_info->tb_id, NULL);
5710 5711 5712
	if (IS_ERR(vr))
		return PTR_ERR(vr);

5713 5714
	mrt = mlxsw_sp_router_fibmr_family_to_table(vr, men_info->info.family);
	return mlxsw_sp_mr_route_add(mrt, men_info->mfc, replace);
5715 5716 5717 5718 5719
}

static void mlxsw_sp_router_fibmr_del(struct mlxsw_sp *mlxsw_sp,
				      struct mfc_entry_notifier_info *men_info)
{
5720
	struct mlxsw_sp_mr_table *mrt;
5721 5722 5723 5724 5725 5726 5727 5728 5729
	struct mlxsw_sp_vr *vr;

	if (mlxsw_sp->router->aborted)
		return;

	vr = mlxsw_sp_vr_find(mlxsw_sp, men_info->tb_id);
	if (WARN_ON(!vr))
		return;

5730 5731
	mrt = mlxsw_sp_router_fibmr_family_to_table(vr, men_info->info.family);
	mlxsw_sp_mr_route_del(mrt, men_info->mfc);
5732
	mlxsw_sp_vr_put(mlxsw_sp, vr);
5733 5734 5735 5736 5737 5738
}

static int
mlxsw_sp_router_fibmr_vif_add(struct mlxsw_sp *mlxsw_sp,
			      struct vif_entry_notifier_info *ven_info)
{
5739
	struct mlxsw_sp_mr_table *mrt;
5740 5741 5742 5743 5744 5745
	struct mlxsw_sp_rif *rif;
	struct mlxsw_sp_vr *vr;

	if (mlxsw_sp->router->aborted)
		return 0;

5746
	vr = mlxsw_sp_vr_get(mlxsw_sp, ven_info->tb_id, NULL);
5747 5748 5749
	if (IS_ERR(vr))
		return PTR_ERR(vr);

5750
	mrt = mlxsw_sp_router_fibmr_family_to_table(vr, ven_info->info.family);
5751
	rif = mlxsw_sp_rif_find_by_dev(mlxsw_sp, ven_info->dev);
5752
	return mlxsw_sp_mr_vif_add(mrt, ven_info->dev,
5753 5754 5755 5756 5757 5758 5759 5760
				   ven_info->vif_index,
				   ven_info->vif_flags, rif);
}

static void
mlxsw_sp_router_fibmr_vif_del(struct mlxsw_sp *mlxsw_sp,
			      struct vif_entry_notifier_info *ven_info)
{
5761
	struct mlxsw_sp_mr_table *mrt;
5762 5763 5764 5765 5766 5767 5768 5769 5770
	struct mlxsw_sp_vr *vr;

	if (mlxsw_sp->router->aborted)
		return;

	vr = mlxsw_sp_vr_find(mlxsw_sp, ven_info->tb_id);
	if (WARN_ON(!vr))
		return;

5771 5772
	mrt = mlxsw_sp_router_fibmr_family_to_table(vr, ven_info->info.family);
	mlxsw_sp_mr_vif_del(mrt, ven_info->vif_index);
5773
	mlxsw_sp_vr_put(mlxsw_sp, vr);
5774 5775
}

5776 5777 5778 5779 5780 5781 5782 5783 5784 5785
static int mlxsw_sp_router_set_abort_trap(struct mlxsw_sp *mlxsw_sp)
{
	enum mlxsw_reg_ralxx_protocol proto = MLXSW_REG_RALXX_PROTOCOL_IPV4;
	int err;

	err = __mlxsw_sp_router_set_abort_trap(mlxsw_sp, proto,
					       MLXSW_SP_LPM_TREE_MIN);
	if (err)
		return err;

5786 5787 5788 5789
	/* The multicast router code does not need an abort trap as by default,
	 * packets that don't match any routes are trapped to the CPU.
	 */

5790 5791 5792 5793 5794
	proto = MLXSW_REG_RALXX_PROTOCOL_IPV6;
	return __mlxsw_sp_router_set_abort_trap(mlxsw_sp, proto,
						MLXSW_SP_LPM_TREE_MIN + 1);
}

5795 5796 5797
static void mlxsw_sp_fib4_node_flush(struct mlxsw_sp *mlxsw_sp,
				     struct mlxsw_sp_fib_node *fib_node)
{
5798
	struct mlxsw_sp_fib4_entry *fib4_entry;
5799

5800 5801 5802 5803 5804
	fib4_entry = container_of(fib_node->fib_entry,
				  struct mlxsw_sp_fib4_entry, common);
	mlxsw_sp_fib_node_entry_unlink(mlxsw_sp, fib_node->fib_entry);
	mlxsw_sp_fib4_entry_destroy(mlxsw_sp, fib4_entry);
	mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
5805 5806
}

5807 5808 5809
static void mlxsw_sp_fib6_node_flush(struct mlxsw_sp *mlxsw_sp,
				     struct mlxsw_sp_fib_node *fib_node)
{
5810
	struct mlxsw_sp_fib6_entry *fib6_entry;
5811

5812 5813 5814 5815 5816
	fib6_entry = container_of(fib_node->fib_entry,
				  struct mlxsw_sp_fib6_entry, common);
	mlxsw_sp_fib_node_entry_unlink(mlxsw_sp, fib_node->fib_entry);
	mlxsw_sp_fib6_entry_destroy(mlxsw_sp, fib6_entry);
	mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
5817 5818
}

5819 5820 5821
static void mlxsw_sp_fib_node_flush(struct mlxsw_sp *mlxsw_sp,
				    struct mlxsw_sp_fib_node *fib_node)
{
5822
	switch (fib_node->fib->proto) {
5823 5824 5825 5826
	case MLXSW_SP_L3_PROTO_IPV4:
		mlxsw_sp_fib4_node_flush(mlxsw_sp, fib_node);
		break;
	case MLXSW_SP_L3_PROTO_IPV6:
5827
		mlxsw_sp_fib6_node_flush(mlxsw_sp, fib_node);
5828 5829 5830 5831
		break;
	}
}

5832 5833 5834
static void mlxsw_sp_vr_fib_flush(struct mlxsw_sp *mlxsw_sp,
				  struct mlxsw_sp_vr *vr,
				  enum mlxsw_sp_l3proto proto)
5835
{
5836
	struct mlxsw_sp_fib *fib = mlxsw_sp_vr_fib(vr, proto);
5837
	struct mlxsw_sp_fib_node *fib_node, *tmp;
5838 5839 5840 5841 5842 5843 5844 5845 5846 5847 5848 5849

	list_for_each_entry_safe(fib_node, tmp, &fib->node_list, list) {
		bool do_break = &tmp->list == &fib->node_list;

		mlxsw_sp_fib_node_flush(mlxsw_sp, fib_node);
		if (do_break)
			break;
	}
}

static void mlxsw_sp_router_fib_flush(struct mlxsw_sp *mlxsw_sp)
{
5850
	int i, j;
5851

J
Jiri Pirko 已提交
5852
	for (i = 0; i < MLXSW_CORE_RES_GET(mlxsw_sp->core, MAX_VRS); i++) {
5853
		struct mlxsw_sp_vr *vr = &mlxsw_sp->router->vrs[i];
5854

5855
		if (!mlxsw_sp_vr_is_used(vr))
5856
			continue;
5857

5858 5859
		for (j = 0; j < MLXSW_SP_L3_PROTO_MAX; j++)
			mlxsw_sp_mr_table_flush(vr->mr_table[j]);
5860
		mlxsw_sp_vr_fib_flush(mlxsw_sp, vr, MLXSW_SP_L3_PROTO_IPV4);
5861 5862 5863 5864 5865 5866 5867

		/* If virtual router was only used for IPv4, then it's no
		 * longer used.
		 */
		if (!mlxsw_sp_vr_is_used(vr))
			continue;
		mlxsw_sp_vr_fib_flush(mlxsw_sp, vr, MLXSW_SP_L3_PROTO_IPV6);
5868
	}
5869 5870 5871 5872 5873 5874 5875 5876 5877 5878

	/* After flushing all the routes, it is not possible anyone is still
	 * using the adjacency index that is discarding packets, so free it in
	 * case it was allocated.
	 */
	if (!mlxsw_sp->router->adj_discard_index_valid)
		return;
	mlxsw_sp_kvdl_free(mlxsw_sp, MLXSW_SP_KVDL_ENTRY_TYPE_ADJ, 1,
			   mlxsw_sp->router->adj_discard_index);
	mlxsw_sp->router->adj_discard_index_valid = false;
5879 5880
}

5881
static void mlxsw_sp_router_fib_abort(struct mlxsw_sp *mlxsw_sp)
5882 5883 5884
{
	int err;

5885
	if (mlxsw_sp->router->aborted)
5886 5887
		return;
	dev_warn(mlxsw_sp->bus_info->dev, "FIB abort triggered. Note that FIB entries are no longer being offloaded to this device.\n");
5888
	mlxsw_sp_router_fib_flush(mlxsw_sp);
5889
	mlxsw_sp->router->aborted = true;
5890 5891 5892 5893 5894
	err = mlxsw_sp_router_set_abort_trap(mlxsw_sp);
	if (err)
		dev_warn(mlxsw_sp->bus_info->dev, "Failed to set abort trap.\n");
}

5895 5896 5897 5898 5899
struct mlxsw_sp_fib6_event_work {
	struct fib6_info **rt_arr;
	unsigned int nrt6;
};

5900
struct mlxsw_sp_fib_event_work {
5901
	struct work_struct work;
5902
	union {
5903
		struct mlxsw_sp_fib6_event_work fib6_work;
5904
		struct fib_entry_notifier_info fen_info;
5905
		struct fib_rule_notifier_info fr_info;
5906
		struct fib_nh_notifier_info fnh_info;
5907 5908
		struct mfc_entry_notifier_info men_info;
		struct vif_entry_notifier_info ven_info;
5909
	};
5910 5911 5912 5913
	struct mlxsw_sp *mlxsw_sp;
	unsigned long event;
};

5914 5915 5916 5917 5918 5919 5920 5921 5922 5923 5924 5925 5926 5927 5928 5929 5930 5931 5932 5933 5934 5935 5936 5937 5938 5939 5940 5941 5942 5943 5944 5945 5946 5947 5948 5949 5950 5951 5952 5953 5954 5955 5956 5957 5958 5959 5960 5961
static int
mlxsw_sp_router_fib6_work_init(struct mlxsw_sp_fib6_event_work *fib6_work,
			       struct fib6_entry_notifier_info *fen6_info)
{
	struct fib6_info *rt = fen6_info->rt;
	struct fib6_info **rt_arr;
	struct fib6_info *iter;
	unsigned int nrt6;
	int i = 0;

	nrt6 = fen6_info->nsiblings + 1;

	rt_arr = kcalloc(nrt6, sizeof(struct fib6_info *), GFP_ATOMIC);
	if (!rt_arr)
		return -ENOMEM;

	fib6_work->rt_arr = rt_arr;
	fib6_work->nrt6 = nrt6;

	rt_arr[0] = rt;
	fib6_info_hold(rt);

	if (!fen6_info->nsiblings)
		return 0;

	list_for_each_entry(iter, &rt->fib6_siblings, fib6_siblings) {
		if (i == fen6_info->nsiblings)
			break;

		rt_arr[i + 1] = iter;
		fib6_info_hold(iter);
		i++;
	}
	WARN_ON_ONCE(i != fen6_info->nsiblings);

	return 0;
}

static void
mlxsw_sp_router_fib6_work_fini(struct mlxsw_sp_fib6_event_work *fib6_work)
{
	int i;

	for (i = 0; i < fib6_work->nrt6; i++)
		mlxsw_sp_rt6_release(fib6_work->rt_arr[i]);
	kfree(fib6_work->rt_arr);
}

5962
static void mlxsw_sp_router_fib4_event_work(struct work_struct *work)
5963
{
5964
	struct mlxsw_sp_fib_event_work *fib_work =
5965
		container_of(work, struct mlxsw_sp_fib_event_work, work);
5966
	struct mlxsw_sp *mlxsw_sp = fib_work->mlxsw_sp;
5967 5968
	int err;

5969
	rtnl_lock();
5970
	mutex_lock(&mlxsw_sp->router->lock);
5971 5972
	mlxsw_sp_span_respin(mlxsw_sp);

5973
	switch (fib_work->event) {
5974
	case FIB_EVENT_ENTRY_REPLACE:
5975 5976
		err = mlxsw_sp_router_fib4_replace(mlxsw_sp,
						   &fib_work->fen_info);
5977
		if (err)
5978
			mlxsw_sp_router_fib_abort(mlxsw_sp);
5979
		fib_info_put(fib_work->fen_info.fi);
5980
		break;
5981
	case FIB_EVENT_ENTRY_DEL:
5982 5983
		mlxsw_sp_router_fib4_del(mlxsw_sp, &fib_work->fen_info);
		fib_info_put(fib_work->fen_info.fi);
5984
		break;
5985 5986
	case FIB_EVENT_NH_ADD: /* fall through */
	case FIB_EVENT_NH_DEL:
5987 5988
		mlxsw_sp_nexthop4_event(mlxsw_sp, fib_work->event,
					fib_work->fnh_info.fib_nh);
5989 5990
		fib_info_put(fib_work->fnh_info.fib_nh->nh_parent);
		break;
5991
	}
5992
	mutex_unlock(&mlxsw_sp->router->lock);
5993 5994 5995 5996
	rtnl_unlock();
	kfree(fib_work);
}

5997 5998
static void mlxsw_sp_router_fib6_event_work(struct work_struct *work)
{
5999 6000 6001
	struct mlxsw_sp_fib_event_work *fib_work =
		container_of(work, struct mlxsw_sp_fib_event_work, work);
	struct mlxsw_sp *mlxsw_sp = fib_work->mlxsw_sp;
6002
	int err;
6003 6004

	rtnl_lock();
6005
	mutex_lock(&mlxsw_sp->router->lock);
6006 6007
	mlxsw_sp_span_respin(mlxsw_sp);

6008
	switch (fib_work->event) {
6009
	case FIB_EVENT_ENTRY_REPLACE:
6010 6011 6012
		err = mlxsw_sp_router_fib6_replace(mlxsw_sp,
						   fib_work->fib6_work.rt_arr,
						   fib_work->fib6_work.nrt6);
6013 6014
		if (err)
			mlxsw_sp_router_fib_abort(mlxsw_sp);
6015
		mlxsw_sp_router_fib6_work_fini(&fib_work->fib6_work);
6016
		break;
6017 6018 6019 6020 6021 6022 6023 6024
	case FIB_EVENT_ENTRY_APPEND:
		err = mlxsw_sp_router_fib6_append(mlxsw_sp,
						  fib_work->fib6_work.rt_arr,
						  fib_work->fib6_work.nrt6);
		if (err)
			mlxsw_sp_router_fib_abort(mlxsw_sp);
		mlxsw_sp_router_fib6_work_fini(&fib_work->fib6_work);
		break;
6025
	case FIB_EVENT_ENTRY_DEL:
6026
		mlxsw_sp_router_fib6_del(mlxsw_sp,
6027 6028
					 fib_work->fib6_work.rt_arr,
					 fib_work->fib6_work.nrt6);
6029
		mlxsw_sp_router_fib6_work_fini(&fib_work->fib6_work);
6030
		break;
6031
	}
6032
	mutex_unlock(&mlxsw_sp->router->lock);
6033 6034
	rtnl_unlock();
	kfree(fib_work);
6035 6036
}

6037 6038 6039 6040 6041 6042 6043 6044 6045
static void mlxsw_sp_router_fibmr_event_work(struct work_struct *work)
{
	struct mlxsw_sp_fib_event_work *fib_work =
		container_of(work, struct mlxsw_sp_fib_event_work, work);
	struct mlxsw_sp *mlxsw_sp = fib_work->mlxsw_sp;
	bool replace;
	int err;

	rtnl_lock();
6046
	mutex_lock(&mlxsw_sp->router->lock);
6047 6048 6049 6050 6051 6052 6053 6054 6055
	switch (fib_work->event) {
	case FIB_EVENT_ENTRY_REPLACE: /* fall through */
	case FIB_EVENT_ENTRY_ADD:
		replace = fib_work->event == FIB_EVENT_ENTRY_REPLACE;

		err = mlxsw_sp_router_fibmr_add(mlxsw_sp, &fib_work->men_info,
						replace);
		if (err)
			mlxsw_sp_router_fib_abort(mlxsw_sp);
Y
Yuval Mintz 已提交
6056
		mr_cache_put(fib_work->men_info.mfc);
6057 6058 6059
		break;
	case FIB_EVENT_ENTRY_DEL:
		mlxsw_sp_router_fibmr_del(mlxsw_sp, &fib_work->men_info);
Y
Yuval Mintz 已提交
6060
		mr_cache_put(fib_work->men_info.mfc);
6061 6062 6063 6064 6065 6066 6067 6068 6069 6070 6071 6072 6073 6074
		break;
	case FIB_EVENT_VIF_ADD:
		err = mlxsw_sp_router_fibmr_vif_add(mlxsw_sp,
						    &fib_work->ven_info);
		if (err)
			mlxsw_sp_router_fib_abort(mlxsw_sp);
		dev_put(fib_work->ven_info.dev);
		break;
	case FIB_EVENT_VIF_DEL:
		mlxsw_sp_router_fibmr_vif_del(mlxsw_sp,
					      &fib_work->ven_info);
		dev_put(fib_work->ven_info.dev);
		break;
	}
6075
	mutex_unlock(&mlxsw_sp->router->lock);
6076 6077 6078 6079
	rtnl_unlock();
	kfree(fib_work);
}

6080 6081 6082
static void mlxsw_sp_router_fib4_event(struct mlxsw_sp_fib_event_work *fib_work,
				       struct fib_notifier_info *info)
{
6083 6084 6085
	struct fib_entry_notifier_info *fen_info;
	struct fib_nh_notifier_info *fnh_info;

6086
	switch (fib_work->event) {
6087 6088
	case FIB_EVENT_ENTRY_REPLACE: /* fall through */
	case FIB_EVENT_ENTRY_DEL:
6089 6090 6091 6092
		fen_info = container_of(info, struct fib_entry_notifier_info,
					info);
		fib_work->fen_info = *fen_info;
		/* Take reference on fib_info to prevent it from being
6093 6094 6095 6096 6097 6098
		 * freed while work is queued. Release it afterwards.
		 */
		fib_info_hold(fib_work->fen_info.fi);
		break;
	case FIB_EVENT_NH_ADD: /* fall through */
	case FIB_EVENT_NH_DEL:
6099 6100 6101
		fnh_info = container_of(info, struct fib_nh_notifier_info,
					info);
		fib_work->fnh_info = *fnh_info;
6102 6103 6104 6105 6106
		fib_info_hold(fib_work->fnh_info.fib_nh->nh_parent);
		break;
	}
}

6107 6108
static int mlxsw_sp_router_fib6_event(struct mlxsw_sp_fib_event_work *fib_work,
				      struct fib_notifier_info *info)
6109
{
6110
	struct fib6_entry_notifier_info *fen6_info;
6111
	int err;
6112

6113
	switch (fib_work->event) {
6114
	case FIB_EVENT_ENTRY_REPLACE: /* fall through */
6115
	case FIB_EVENT_ENTRY_APPEND: /* fall through */
6116
	case FIB_EVENT_ENTRY_DEL:
6117 6118
		fen6_info = container_of(info, struct fib6_entry_notifier_info,
					 info);
6119 6120 6121 6122
		err = mlxsw_sp_router_fib6_work_init(&fib_work->fib6_work,
						     fen6_info);
		if (err)
			return err;
6123
		break;
6124
	}
6125 6126

	return 0;
6127 6128
}

6129 6130 6131 6132 6133 6134 6135 6136 6137
static void
mlxsw_sp_router_fibmr_event(struct mlxsw_sp_fib_event_work *fib_work,
			    struct fib_notifier_info *info)
{
	switch (fib_work->event) {
	case FIB_EVENT_ENTRY_REPLACE: /* fall through */
	case FIB_EVENT_ENTRY_ADD: /* fall through */
	case FIB_EVENT_ENTRY_DEL:
		memcpy(&fib_work->men_info, info, sizeof(fib_work->men_info));
Y
Yuval Mintz 已提交
6138
		mr_cache_hold(fib_work->men_info.mfc);
6139 6140 6141 6142 6143 6144
		break;
	case FIB_EVENT_VIF_ADD: /* fall through */
	case FIB_EVENT_VIF_DEL:
		memcpy(&fib_work->ven_info, info, sizeof(fib_work->ven_info));
		dev_hold(fib_work->ven_info.dev);
		break;
6145 6146 6147 6148 6149 6150 6151 6152 6153 6154 6155 6156 6157 6158 6159 6160 6161 6162 6163 6164 6165 6166
	}
}

static int mlxsw_sp_router_fib_rule_event(unsigned long event,
					  struct fib_notifier_info *info,
					  struct mlxsw_sp *mlxsw_sp)
{
	struct netlink_ext_ack *extack = info->extack;
	struct fib_rule_notifier_info *fr_info;
	struct fib_rule *rule;
	int err = 0;

	/* nothing to do at the moment */
	if (event == FIB_EVENT_RULE_DEL)
		return 0;

	if (mlxsw_sp->router->aborted)
		return 0;

	fr_info = container_of(info, struct fib_rule_notifier_info, info);
	rule = fr_info->rule;

6167
	/* Rule only affects locally generated traffic */
6168
	if (rule->iifindex == mlxsw_sp_net(mlxsw_sp)->loopback_dev->ifindex)
6169 6170
		return 0;

6171 6172 6173
	switch (info->family) {
	case AF_INET:
		if (!fib4_rule_default(rule) && !rule->l3mdev)
6174
			err = -EOPNOTSUPP;
6175 6176 6177
		break;
	case AF_INET6:
		if (!fib6_rule_default(rule) && !rule->l3mdev)
6178
			err = -EOPNOTSUPP;
6179 6180 6181
		break;
	case RTNL_FAMILY_IPMR:
		if (!ipmr_rule_default(rule) && !rule->l3mdev)
6182
			err = -EOPNOTSUPP;
6183
		break;
6184 6185
	case RTNL_FAMILY_IP6MR:
		if (!ip6mr_rule_default(rule) && !rule->l3mdev)
6186
			err = -EOPNOTSUPP;
6187
		break;
6188
	}
6189 6190

	if (err < 0)
6191
		NL_SET_ERR_MSG_MOD(extack, "FIB rules not supported");
6192 6193

	return err;
6194 6195
}

6196 6197 6198 6199 6200 6201
/* Called with rcu_read_lock() */
static int mlxsw_sp_router_fib_event(struct notifier_block *nb,
				     unsigned long event, void *ptr)
{
	struct mlxsw_sp_fib_event_work *fib_work;
	struct fib_notifier_info *info = ptr;
6202
	struct mlxsw_sp_router *router;
6203
	int err;
6204

6205
	if ((info->family != AF_INET && info->family != AF_INET6 &&
6206 6207
	     info->family != RTNL_FAMILY_IPMR &&
	     info->family != RTNL_FAMILY_IP6MR))
6208 6209
		return NOTIFY_DONE;

6210 6211 6212 6213 6214 6215 6216
	router = container_of(nb, struct mlxsw_sp_router, fib_nb);

	switch (event) {
	case FIB_EVENT_RULE_ADD: /* fall through */
	case FIB_EVENT_RULE_DEL:
		err = mlxsw_sp_router_fib_rule_event(event, info,
						     router->mlxsw_sp);
6217
		return notifier_from_errno(err);
6218
	case FIB_EVENT_ENTRY_ADD: /* fall through */
6219 6220
	case FIB_EVENT_ENTRY_REPLACE: /* fall through */
	case FIB_EVENT_ENTRY_APPEND:
6221 6222 6223 6224
		if (router->aborted) {
			NL_SET_ERR_MSG_MOD(info->extack, "FIB offload was aborted. Not configuring route");
			return notifier_from_errno(-EINVAL);
		}
6225 6226 6227 6228 6229 6230 6231
		if (info->family == AF_INET) {
			struct fib_entry_notifier_info *fen_info = ptr;

			if (fen_info->fi->fib_nh_is_v6) {
				NL_SET_ERR_MSG_MOD(info->extack, "IPv6 gateway with IPv4 route is not supported");
				return notifier_from_errno(-EINVAL);
			}
6232 6233 6234 6235 6236 6237 6238 6239 6240 6241 6242 6243 6244 6245
			if (fen_info->fi->nh) {
				NL_SET_ERR_MSG_MOD(info->extack, "IPv4 route with nexthop objects is not supported");
				return notifier_from_errno(-EINVAL);
			}
		} else if (info->family == AF_INET6) {
			struct fib6_entry_notifier_info *fen6_info;

			fen6_info = container_of(info,
						 struct fib6_entry_notifier_info,
						 info);
			if (fen6_info->rt->nh) {
				NL_SET_ERR_MSG_MOD(info->extack, "IPv6 route with nexthop objects is not supported");
				return notifier_from_errno(-EINVAL);
			}
6246
		}
6247
		break;
6248 6249
	}

6250 6251 6252 6253
	fib_work = kzalloc(sizeof(*fib_work), GFP_ATOMIC);
	if (WARN_ON(!fib_work))
		return NOTIFY_BAD;

6254
	fib_work->mlxsw_sp = router->mlxsw_sp;
6255 6256
	fib_work->event = event;

6257 6258 6259 6260
	switch (info->family) {
	case AF_INET:
		INIT_WORK(&fib_work->work, mlxsw_sp_router_fib4_event_work);
		mlxsw_sp_router_fib4_event(fib_work, info);
6261
		break;
6262 6263
	case AF_INET6:
		INIT_WORK(&fib_work->work, mlxsw_sp_router_fib6_event_work);
6264 6265 6266
		err = mlxsw_sp_router_fib6_event(fib_work, info);
		if (err)
			goto err_fib_event;
6267
		break;
6268
	case RTNL_FAMILY_IP6MR:
6269 6270 6271 6272
	case RTNL_FAMILY_IPMR:
		INIT_WORK(&fib_work->work, mlxsw_sp_router_fibmr_event_work);
		mlxsw_sp_router_fibmr_event(fib_work, info);
		break;
6273 6274
	}

6275
	mlxsw_core_schedule_work(&fib_work->work);
6276

6277
	return NOTIFY_DONE;
6278 6279 6280 6281

err_fib_event:
	kfree(fib_work);
	return NOTIFY_BAD;
6282 6283
}

6284
static struct mlxsw_sp_rif *
6285 6286 6287 6288 6289 6290
mlxsw_sp_rif_find_by_dev(const struct mlxsw_sp *mlxsw_sp,
			 const struct net_device *dev)
{
	int i;

	for (i = 0; i < MLXSW_CORE_RES_GET(mlxsw_sp->core, MAX_RIFS); i++)
6291 6292 6293
		if (mlxsw_sp->router->rifs[i] &&
		    mlxsw_sp->router->rifs[i]->dev == dev)
			return mlxsw_sp->router->rifs[i];
6294 6295 6296 6297

	return NULL;
}

6298 6299 6300 6301 6302 6303
bool mlxsw_sp_rif_exists(struct mlxsw_sp *mlxsw_sp,
			 const struct net_device *dev)
{
	return !!mlxsw_sp_rif_find_by_dev(mlxsw_sp, dev);
}

6304 6305 6306 6307 6308 6309 6310 6311 6312 6313 6314 6315 6316 6317 6318 6319 6320 6321 6322 6323 6324
u16 mlxsw_sp_rif_vid(struct mlxsw_sp *mlxsw_sp, const struct net_device *dev)
{
	struct mlxsw_sp_rif *rif;
	u16 vid = 0;

	rif = mlxsw_sp_rif_find_by_dev(mlxsw_sp, dev);
	if (!rif)
		goto out;

	/* We only return the VID for VLAN RIFs. Otherwise we return an
	 * invalid value (0).
	 */
	if (rif->ops->type != MLXSW_SP_RIF_TYPE_VLAN)
		goto out;

	vid = mlxsw_sp_fid_8021q_vid(rif->fid);

out:
	return vid;
}

6325 6326 6327 6328 6329 6330 6331
static int mlxsw_sp_router_rif_disable(struct mlxsw_sp *mlxsw_sp, u16 rif)
{
	char ritr_pl[MLXSW_REG_RITR_LEN];
	int err;

	mlxsw_reg_ritr_rif_pack(ritr_pl, rif);
	err = mlxsw_reg_query(mlxsw_sp->core, MLXSW_REG(ritr), ritr_pl);
6332
	if (err)
6333 6334 6335 6336 6337 6338 6339
		return err;

	mlxsw_reg_ritr_enable_set(ritr_pl, false);
	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ritr), ritr_pl);
}

static void mlxsw_sp_router_rif_gone_sync(struct mlxsw_sp *mlxsw_sp,
6340
					  struct mlxsw_sp_rif *rif)
6341
{
6342 6343 6344
	mlxsw_sp_router_rif_disable(mlxsw_sp, rif->rif_index);
	mlxsw_sp_nexthop_rif_gone_sync(mlxsw_sp, rif);
	mlxsw_sp_neigh_rif_gone_sync(mlxsw_sp, rif);
6345 6346
}

6347 6348 6349
static bool
mlxsw_sp_rif_should_config(struct mlxsw_sp_rif *rif, struct net_device *dev,
			   unsigned long event)
6350
{
6351 6352 6353 6354
	struct inet6_dev *inet6_dev;
	bool addr_list_empty = true;
	struct in_device *idev;

6355 6356
	switch (event) {
	case NETDEV_UP:
6357
		return rif == NULL;
6358
	case NETDEV_DOWN:
6359 6360
		rcu_read_lock();
		idev = __in_dev_get_rcu(dev);
6361 6362 6363 6364 6365 6366 6367
		if (idev && idev->ifa_list)
			addr_list_empty = false;

		inet6_dev = __in6_dev_get(dev);
		if (addr_list_empty && inet6_dev &&
		    !list_empty(&inet6_dev->addr_list))
			addr_list_empty = false;
6368
		rcu_read_unlock();
6369

6370 6371 6372 6373 6374 6375
		/* macvlans do not have a RIF, but rather piggy back on the
		 * RIF of their lower device.
		 */
		if (netif_is_macvlan(dev) && addr_list_empty)
			return true;

6376
		if (rif && addr_list_empty &&
6377
		    !netif_is_l3_slave(rif->dev))
6378 6379 6380 6381 6382 6383 6384 6385 6386 6387 6388
			return true;
		/* It is possible we already removed the RIF ourselves
		 * if it was assigned to a netdev that is now a bridge
		 * or LAG slave.
		 */
		return false;
	}

	return false;
}

6389 6390 6391 6392 6393 6394
static enum mlxsw_sp_rif_type
mlxsw_sp_dev_rif_type(const struct mlxsw_sp *mlxsw_sp,
		      const struct net_device *dev)
{
	enum mlxsw_sp_fid_type type;

6395 6396 6397 6398
	if (mlxsw_sp_netdev_ipip_type(mlxsw_sp, dev, NULL))
		return MLXSW_SP_RIF_TYPE_IPIP_LB;

	/* Otherwise RIF type is derived from the type of the underlying FID. */
6399 6400 6401 6402 6403 6404 6405 6406 6407 6408 6409 6410
	if (is_vlan_dev(dev) && netif_is_bridge_master(vlan_dev_real_dev(dev)))
		type = MLXSW_SP_FID_TYPE_8021Q;
	else if (netif_is_bridge_master(dev) && br_vlan_enabled(dev))
		type = MLXSW_SP_FID_TYPE_8021Q;
	else if (netif_is_bridge_master(dev))
		type = MLXSW_SP_FID_TYPE_8021D;
	else
		type = MLXSW_SP_FID_TYPE_RFID;

	return mlxsw_sp_fid_type_rif_type(mlxsw_sp, type);
}

6411
static int mlxsw_sp_rif_index_alloc(struct mlxsw_sp *mlxsw_sp, u16 *p_rif_index)
6412 6413 6414
{
	int i;

6415 6416 6417 6418 6419 6420
	for (i = 0; i < MLXSW_CORE_RES_GET(mlxsw_sp->core, MAX_RIFS); i++) {
		if (!mlxsw_sp->router->rifs[i]) {
			*p_rif_index = i;
			return 0;
		}
	}
6421

6422
	return -ENOBUFS;
6423 6424
}

6425 6426 6427
static struct mlxsw_sp_rif *mlxsw_sp_rif_alloc(size_t rif_size, u16 rif_index,
					       u16 vr_id,
					       struct net_device *l3_dev)
6428
{
6429
	struct mlxsw_sp_rif *rif;
6430

6431
	rif = kzalloc(rif_size, GFP_KERNEL);
6432
	if (!rif)
6433 6434
		return NULL;

6435 6436
	INIT_LIST_HEAD(&rif->nexthop_list);
	INIT_LIST_HEAD(&rif->neigh_list);
6437 6438 6439 6440 6441
	if (l3_dev) {
		ether_addr_copy(rif->addr, l3_dev->dev_addr);
		rif->mtu = l3_dev->mtu;
		rif->dev = l3_dev;
	}
6442 6443
	rif->vr_id = vr_id;
	rif->rif_index = rif_index;
6444

6445
	return rif;
6446 6447
}

6448 6449 6450 6451 6452 6453
struct mlxsw_sp_rif *mlxsw_sp_rif_by_index(const struct mlxsw_sp *mlxsw_sp,
					   u16 rif_index)
{
	return mlxsw_sp->router->rifs[rif_index];
}

6454 6455 6456 6457 6458
u16 mlxsw_sp_rif_index(const struct mlxsw_sp_rif *rif)
{
	return rif->rif_index;
}

6459 6460 6461 6462 6463 6464 6465
u16 mlxsw_sp_ipip_lb_rif_index(const struct mlxsw_sp_rif_ipip_lb *lb_rif)
{
	return lb_rif->common.rif_index;
}

u16 mlxsw_sp_ipip_lb_ul_vr_id(const struct mlxsw_sp_rif_ipip_lb *lb_rif)
{
6466 6467 6468 6469 6470 6471 6472 6473
	u32 ul_tb_id = mlxsw_sp_ipip_dev_ul_tb_id(lb_rif->common.dev);
	struct mlxsw_sp_vr *ul_vr;

	ul_vr = mlxsw_sp_vr_get(lb_rif->common.mlxsw_sp, ul_tb_id, NULL);
	if (WARN_ON(IS_ERR(ul_vr)))
		return 0;

	return ul_vr->id;
6474 6475
}

6476 6477 6478 6479 6480
u16 mlxsw_sp_ipip_lb_ul_rif_id(const struct mlxsw_sp_rif_ipip_lb *lb_rif)
{
	return lb_rif->ul_rif_id;
}

6481 6482 6483 6484 6485
int mlxsw_sp_rif_dev_ifindex(const struct mlxsw_sp_rif *rif)
{
	return rif->dev->ifindex;
}

6486 6487 6488 6489 6490
const struct net_device *mlxsw_sp_rif_dev(const struct mlxsw_sp_rif *rif)
{
	return rif->dev;
}

6491
static struct mlxsw_sp_rif *
6492
mlxsw_sp_rif_create(struct mlxsw_sp *mlxsw_sp,
6493 6494
		    const struct mlxsw_sp_rif_params *params,
		    struct netlink_ext_ack *extack)
6495
{
6496 6497
	u32 tb_id = l3mdev_fib_table(params->dev);
	const struct mlxsw_sp_rif_ops *ops;
6498
	struct mlxsw_sp_fid *fid = NULL;
6499
	enum mlxsw_sp_rif_type type;
6500
	struct mlxsw_sp_rif *rif;
6501 6502
	struct mlxsw_sp_vr *vr;
	u16 rif_index;
6503
	int i, err;
6504

6505
	type = mlxsw_sp_dev_rif_type(mlxsw_sp, params->dev);
6506
	ops = mlxsw_sp->rif_ops_arr[type];
6507

6508
	vr = mlxsw_sp_vr_get(mlxsw_sp, tb_id ? : RT_TABLE_MAIN, extack);
6509 6510
	if (IS_ERR(vr))
		return ERR_CAST(vr);
6511
	vr->rif_count++;
6512

6513
	err = mlxsw_sp_rif_index_alloc(mlxsw_sp, &rif_index);
6514
	if (err) {
6515
		NL_SET_ERR_MSG_MOD(extack, "Exceeded number of supported router interfaces");
6516
		goto err_rif_index_alloc;
6517
	}
6518

6519
	rif = mlxsw_sp_rif_alloc(ops->rif_size, rif_index, vr->id, params->dev);
6520 6521 6522 6523
	if (!rif) {
		err = -ENOMEM;
		goto err_rif_alloc;
	}
6524
	dev_hold(rif->dev);
6525
	mlxsw_sp->router->rifs[rif_index] = rif;
6526 6527
	rif->mlxsw_sp = mlxsw_sp;
	rif->ops = ops;
6528

6529
	if (ops->fid_get) {
6530
		fid = ops->fid_get(rif, extack);
6531 6532 6533 6534 6535
		if (IS_ERR(fid)) {
			err = PTR_ERR(fid);
			goto err_fid_get;
		}
		rif->fid = fid;
6536 6537
	}

6538 6539 6540 6541
	if (ops->setup)
		ops->setup(rif, params);

	err = ops->configure(rif);
6542
	if (err)
6543
		goto err_configure;
6544

6545 6546 6547 6548 6549
	for (i = 0; i < MLXSW_SP_L3_PROTO_MAX; i++) {
		err = mlxsw_sp_mr_rif_add(vr->mr_table[i], rif);
		if (err)
			goto err_mr_rif_add;
	}
6550

6551
	mlxsw_sp_rif_counters_alloc(rif);
6552

6553
	return rif;
6554

6555
err_mr_rif_add:
6556 6557
	for (i--; i >= 0; i--)
		mlxsw_sp_mr_rif_del(vr->mr_table[i], rif);
6558
	ops->deconfigure(rif);
6559
err_configure:
6560 6561
	if (fid)
		mlxsw_sp_fid_put(fid);
6562
err_fid_get:
6563
	mlxsw_sp->router->rifs[rif_index] = NULL;
6564
	dev_put(rif->dev);
6565 6566
	kfree(rif);
err_rif_alloc:
6567
err_rif_index_alloc:
6568
	vr->rif_count--;
6569
	mlxsw_sp_vr_put(mlxsw_sp, vr);
6570 6571 6572
	return ERR_PTR(err);
}

6573
static void mlxsw_sp_rif_destroy(struct mlxsw_sp_rif *rif)
6574
{
6575 6576
	const struct mlxsw_sp_rif_ops *ops = rif->ops;
	struct mlxsw_sp *mlxsw_sp = rif->mlxsw_sp;
6577
	struct mlxsw_sp_fid *fid = rif->fid;
6578
	struct mlxsw_sp_vr *vr;
6579
	int i;
6580

6581
	mlxsw_sp_router_rif_gone_sync(mlxsw_sp, rif);
6582
	vr = &mlxsw_sp->router->vrs[rif->vr_id];
6583

6584
	mlxsw_sp_rif_counters_free(rif);
6585 6586
	for (i = 0; i < MLXSW_SP_L3_PROTO_MAX; i++)
		mlxsw_sp_mr_rif_del(vr->mr_table[i], rif);
6587
	ops->deconfigure(rif);
6588 6589 6590
	if (fid)
		/* Loopback RIFs are not associated with a FID. */
		mlxsw_sp_fid_put(fid);
6591
	mlxsw_sp->router->rifs[rif->rif_index] = NULL;
6592
	dev_put(rif->dev);
6593
	kfree(rif);
6594
	vr->rif_count--;
6595
	mlxsw_sp_vr_put(mlxsw_sp, vr);
6596 6597
}

6598 6599 6600 6601 6602 6603 6604 6605 6606 6607 6608
void mlxsw_sp_rif_destroy_by_dev(struct mlxsw_sp *mlxsw_sp,
				 struct net_device *dev)
{
	struct mlxsw_sp_rif *rif;

	rif = mlxsw_sp_rif_find_by_dev(mlxsw_sp, dev);
	if (!rif)
		return;
	mlxsw_sp_rif_destroy(rif);
}

6609 6610 6611 6612 6613 6614 6615 6616 6617 6618 6619 6620 6621 6622
static void
mlxsw_sp_rif_subport_params_init(struct mlxsw_sp_rif_params *params,
				 struct mlxsw_sp_port_vlan *mlxsw_sp_port_vlan)
{
	struct mlxsw_sp_port *mlxsw_sp_port = mlxsw_sp_port_vlan->mlxsw_sp_port;

	params->vid = mlxsw_sp_port_vlan->vid;
	params->lag = mlxsw_sp_port->lagged;
	if (params->lag)
		params->lag_id = mlxsw_sp_port->lag_id;
	else
		params->system_port = mlxsw_sp_port->local_port;
}

6623 6624 6625 6626 6627 6628 6629 6630 6631 6632 6633 6634 6635 6636 6637 6638 6639 6640 6641 6642 6643 6644 6645 6646 6647 6648 6649 6650 6651 6652 6653 6654 6655 6656
static struct mlxsw_sp_rif_subport *
mlxsw_sp_rif_subport_rif(const struct mlxsw_sp_rif *rif)
{
	return container_of(rif, struct mlxsw_sp_rif_subport, common);
}

static struct mlxsw_sp_rif *
mlxsw_sp_rif_subport_get(struct mlxsw_sp *mlxsw_sp,
			 const struct mlxsw_sp_rif_params *params,
			 struct netlink_ext_ack *extack)
{
	struct mlxsw_sp_rif_subport *rif_subport;
	struct mlxsw_sp_rif *rif;

	rif = mlxsw_sp_rif_find_by_dev(mlxsw_sp, params->dev);
	if (!rif)
		return mlxsw_sp_rif_create(mlxsw_sp, params, extack);

	rif_subport = mlxsw_sp_rif_subport_rif(rif);
	refcount_inc(&rif_subport->ref_count);
	return rif;
}

static void mlxsw_sp_rif_subport_put(struct mlxsw_sp_rif *rif)
{
	struct mlxsw_sp_rif_subport *rif_subport;

	rif_subport = mlxsw_sp_rif_subport_rif(rif);
	if (!refcount_dec_and_test(&rif_subport->ref_count))
		return;

	mlxsw_sp_rif_destroy(rif);
}

6657
static int
6658
mlxsw_sp_port_vlan_router_join(struct mlxsw_sp_port_vlan *mlxsw_sp_port_vlan,
6659 6660
			       struct net_device *l3_dev,
			       struct netlink_ext_ack *extack)
6661
{
6662
	struct mlxsw_sp_port *mlxsw_sp_port = mlxsw_sp_port_vlan->mlxsw_sp_port;
6663
	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
6664 6665 6666
	struct mlxsw_sp_rif_params params = {
		.dev = l3_dev,
	};
6667
	u16 vid = mlxsw_sp_port_vlan->vid;
6668
	struct mlxsw_sp_rif *rif;
6669
	struct mlxsw_sp_fid *fid;
6670
	int err;
6671

6672 6673 6674 6675
	mlxsw_sp_rif_subport_params_init(&params, mlxsw_sp_port_vlan);
	rif = mlxsw_sp_rif_subport_get(mlxsw_sp, &params, extack);
	if (IS_ERR(rif))
		return PTR_ERR(rif);
6676

6677
	/* FID was already created, just take a reference */
6678
	fid = rif->ops->fid_get(rif, extack);
6679 6680 6681 6682
	err = mlxsw_sp_fid_port_vid_map(fid, mlxsw_sp_port, vid);
	if (err)
		goto err_fid_port_vid_map;

6683
	err = mlxsw_sp_port_vid_learning_set(mlxsw_sp_port, vid, false);
6684 6685 6686
	if (err)
		goto err_port_vid_learning_set;

6687
	err = mlxsw_sp_port_vid_stp_set(mlxsw_sp_port, vid,
6688 6689 6690 6691
					BR_STATE_FORWARDING);
	if (err)
		goto err_port_vid_stp_set;

6692
	mlxsw_sp_port_vlan->fid = fid;
6693 6694

	return 0;
6695 6696

err_port_vid_stp_set:
6697
	mlxsw_sp_port_vid_learning_set(mlxsw_sp_port, vid, true);
6698
err_port_vid_learning_set:
6699 6700 6701
	mlxsw_sp_fid_port_vid_unmap(fid, mlxsw_sp_port, vid);
err_fid_port_vid_map:
	mlxsw_sp_fid_put(fid);
6702
	mlxsw_sp_rif_subport_put(rif);
6703
	return err;
6704 6705
}

6706 6707
static void
__mlxsw_sp_port_vlan_router_leave(struct mlxsw_sp_port_vlan *mlxsw_sp_port_vlan)
6708
{
6709
	struct mlxsw_sp_port *mlxsw_sp_port = mlxsw_sp_port_vlan->mlxsw_sp_port;
6710
	struct mlxsw_sp_fid *fid = mlxsw_sp_port_vlan->fid;
6711
	struct mlxsw_sp_rif *rif = mlxsw_sp_fid_rif(fid);
6712 6713
	u16 vid = mlxsw_sp_port_vlan->vid;

6714 6715
	if (WARN_ON(mlxsw_sp_fid_type(fid) != MLXSW_SP_FID_TYPE_RFID))
		return;
6716

6717
	mlxsw_sp_port_vlan->fid = NULL;
6718 6719
	mlxsw_sp_port_vid_stp_set(mlxsw_sp_port, vid, BR_STATE_BLOCKING);
	mlxsw_sp_port_vid_learning_set(mlxsw_sp_port, vid, true);
6720 6721
	mlxsw_sp_fid_port_vid_unmap(fid, mlxsw_sp_port, vid);
	mlxsw_sp_fid_put(fid);
6722
	mlxsw_sp_rif_subport_put(rif);
6723 6724
}

6725 6726 6727 6728 6729 6730
void
mlxsw_sp_port_vlan_router_leave(struct mlxsw_sp_port_vlan *mlxsw_sp_port_vlan)
{
	__mlxsw_sp_port_vlan_router_leave(mlxsw_sp_port_vlan);
}

6731 6732
static int mlxsw_sp_inetaddr_port_vlan_event(struct net_device *l3_dev,
					     struct net_device *port_dev,
6733 6734
					     unsigned long event, u16 vid,
					     struct netlink_ext_ack *extack)
6735 6736
{
	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(port_dev);
6737
	struct mlxsw_sp_port_vlan *mlxsw_sp_port_vlan;
6738

6739
	mlxsw_sp_port_vlan = mlxsw_sp_port_vlan_find_by_vid(mlxsw_sp_port, vid);
6740 6741
	if (WARN_ON(!mlxsw_sp_port_vlan))
		return -EINVAL;
6742 6743 6744

	switch (event) {
	case NETDEV_UP:
6745
		return mlxsw_sp_port_vlan_router_join(mlxsw_sp_port_vlan,
6746
						      l3_dev, extack);
6747
	case NETDEV_DOWN:
6748
		__mlxsw_sp_port_vlan_router_leave(mlxsw_sp_port_vlan);
6749 6750 6751 6752 6753 6754 6755
		break;
	}

	return 0;
}

static int mlxsw_sp_inetaddr_port_event(struct net_device *port_dev,
6756 6757
					unsigned long event,
					struct netlink_ext_ack *extack)
6758
{
6759 6760 6761
	if (netif_is_bridge_port(port_dev) ||
	    netif_is_lag_port(port_dev) ||
	    netif_is_ovs_port(port_dev))
6762 6763
		return 0;

6764 6765
	return mlxsw_sp_inetaddr_port_vlan_event(port_dev, port_dev, event,
						 MLXSW_SP_DEFAULT_VID, extack);
6766 6767 6768 6769
}

static int __mlxsw_sp_inetaddr_lag_event(struct net_device *l3_dev,
					 struct net_device *lag_dev,
6770 6771
					 unsigned long event, u16 vid,
					 struct netlink_ext_ack *extack)
6772 6773 6774 6775 6776 6777 6778
{
	struct net_device *port_dev;
	struct list_head *iter;
	int err;

	netdev_for_each_lower_dev(lag_dev, port_dev, iter) {
		if (mlxsw_sp_port_dev_check(port_dev)) {
6779 6780
			err = mlxsw_sp_inetaddr_port_vlan_event(l3_dev,
								port_dev,
6781 6782
								event, vid,
								extack);
6783 6784 6785 6786 6787 6788 6789 6790 6791
			if (err)
				return err;
		}
	}

	return 0;
}

static int mlxsw_sp_inetaddr_lag_event(struct net_device *lag_dev,
6792 6793
				       unsigned long event,
				       struct netlink_ext_ack *extack)
6794 6795 6796 6797
{
	if (netif_is_bridge_port(lag_dev))
		return 0;

6798 6799
	return __mlxsw_sp_inetaddr_lag_event(lag_dev, lag_dev, event,
					     MLXSW_SP_DEFAULT_VID, extack);
6800 6801
}

6802 6803
static int mlxsw_sp_inetaddr_bridge_event(struct mlxsw_sp *mlxsw_sp,
					  struct net_device *l3_dev,
6804 6805
					  unsigned long event,
					  struct netlink_ext_ack *extack)
6806
{
6807 6808 6809
	struct mlxsw_sp_rif_params params = {
		.dev = l3_dev,
	};
6810
	struct mlxsw_sp_rif *rif;
6811 6812 6813

	switch (event) {
	case NETDEV_UP:
6814
		rif = mlxsw_sp_rif_create(mlxsw_sp, &params, extack);
6815 6816 6817
		if (IS_ERR(rif))
			return PTR_ERR(rif);
		break;
6818
	case NETDEV_DOWN:
6819
		rif = mlxsw_sp_rif_find_by_dev(mlxsw_sp, l3_dev);
6820
		mlxsw_sp_rif_destroy(rif);
6821 6822 6823 6824 6825 6826
		break;
	}

	return 0;
}

6827 6828
static int mlxsw_sp_inetaddr_vlan_event(struct mlxsw_sp *mlxsw_sp,
					struct net_device *vlan_dev,
6829 6830
					unsigned long event,
					struct netlink_ext_ack *extack)
6831 6832 6833 6834
{
	struct net_device *real_dev = vlan_dev_real_dev(vlan_dev);
	u16 vid = vlan_dev_vlan_id(vlan_dev);

6835 6836 6837
	if (netif_is_bridge_port(vlan_dev))
		return 0;

6838
	if (mlxsw_sp_port_dev_check(real_dev))
6839
		return mlxsw_sp_inetaddr_port_vlan_event(vlan_dev, real_dev,
6840
							 event, vid, extack);
6841 6842
	else if (netif_is_lag_master(real_dev))
		return __mlxsw_sp_inetaddr_lag_event(vlan_dev, real_dev, event,
6843
						     vid, extack);
6844
	else if (netif_is_bridge_master(real_dev) && br_vlan_enabled(real_dev))
6845 6846
		return mlxsw_sp_inetaddr_bridge_event(mlxsw_sp, vlan_dev, event,
						      extack);
6847 6848 6849 6850

	return 0;
}

6851 6852 6853 6854 6855 6856 6857 6858 6859 6860 6861 6862 6863 6864 6865 6866 6867 6868 6869 6870 6871 6872 6873 6874 6875 6876 6877 6878 6879 6880 6881 6882 6883 6884 6885 6886 6887 6888 6889 6890
static bool mlxsw_sp_rif_macvlan_is_vrrp4(const u8 *mac)
{
	u8 vrrp4[ETH_ALEN] = { 0x00, 0x00, 0x5e, 0x00, 0x01, 0x00 };
	u8 mask[ETH_ALEN] = { 0xff, 0xff, 0xff, 0xff, 0xff, 0x00 };

	return ether_addr_equal_masked(mac, vrrp4, mask);
}

static bool mlxsw_sp_rif_macvlan_is_vrrp6(const u8 *mac)
{
	u8 vrrp6[ETH_ALEN] = { 0x00, 0x00, 0x5e, 0x00, 0x02, 0x00 };
	u8 mask[ETH_ALEN] = { 0xff, 0xff, 0xff, 0xff, 0xff, 0x00 };

	return ether_addr_equal_masked(mac, vrrp6, mask);
}

static int mlxsw_sp_rif_vrrp_op(struct mlxsw_sp *mlxsw_sp, u16 rif_index,
				const u8 *mac, bool adding)
{
	char ritr_pl[MLXSW_REG_RITR_LEN];
	u8 vrrp_id = adding ? mac[5] : 0;
	int err;

	if (!mlxsw_sp_rif_macvlan_is_vrrp4(mac) &&
	    !mlxsw_sp_rif_macvlan_is_vrrp6(mac))
		return 0;

	mlxsw_reg_ritr_rif_pack(ritr_pl, rif_index);
	err = mlxsw_reg_query(mlxsw_sp->core, MLXSW_REG(ritr), ritr_pl);
	if (err)
		return err;

	if (mlxsw_sp_rif_macvlan_is_vrrp4(mac))
		mlxsw_reg_ritr_if_vrrp_id_ipv4_set(ritr_pl, vrrp_id);
	else
		mlxsw_reg_ritr_if_vrrp_id_ipv6_set(ritr_pl, vrrp_id);

	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ritr), ritr_pl);
}

6891 6892 6893 6894 6895 6896 6897 6898 6899 6900 6901 6902 6903 6904 6905 6906 6907 6908 6909
static int mlxsw_sp_rif_macvlan_add(struct mlxsw_sp *mlxsw_sp,
				    const struct net_device *macvlan_dev,
				    struct netlink_ext_ack *extack)
{
	struct macvlan_dev *vlan = netdev_priv(macvlan_dev);
	struct mlxsw_sp_rif *rif;
	int err;

	rif = mlxsw_sp_rif_find_by_dev(mlxsw_sp, vlan->lowerdev);
	if (!rif) {
		NL_SET_ERR_MSG_MOD(extack, "macvlan is only supported on top of router interfaces");
		return -EOPNOTSUPP;
	}

	err = mlxsw_sp_rif_fdb_op(mlxsw_sp, macvlan_dev->dev_addr,
				  mlxsw_sp_fid_index(rif->fid), true);
	if (err)
		return err;

6910 6911 6912 6913 6914
	err = mlxsw_sp_rif_vrrp_op(mlxsw_sp, rif->rif_index,
				   macvlan_dev->dev_addr, true);
	if (err)
		goto err_rif_vrrp_add;

6915 6916 6917 6918 6919 6920 6921
	/* Make sure the bridge driver does not have this MAC pointing at
	 * some other port.
	 */
	if (rif->ops->fdb_del)
		rif->ops->fdb_del(rif, macvlan_dev->dev_addr);

	return 0;
6922 6923 6924 6925 6926

err_rif_vrrp_add:
	mlxsw_sp_rif_fdb_op(mlxsw_sp, macvlan_dev->dev_addr,
			    mlxsw_sp_fid_index(rif->fid), false);
	return err;
6927 6928
}

6929 6930
static void __mlxsw_sp_rif_macvlan_del(struct mlxsw_sp *mlxsw_sp,
				       const struct net_device *macvlan_dev)
6931 6932 6933 6934 6935 6936 6937 6938 6939 6940
{
	struct macvlan_dev *vlan = netdev_priv(macvlan_dev);
	struct mlxsw_sp_rif *rif;

	rif = mlxsw_sp_rif_find_by_dev(mlxsw_sp, vlan->lowerdev);
	/* If we do not have a RIF, then we already took care of
	 * removing the macvlan's MAC during RIF deletion.
	 */
	if (!rif)
		return;
6941 6942
	mlxsw_sp_rif_vrrp_op(mlxsw_sp, rif->rif_index, macvlan_dev->dev_addr,
			     false);
6943 6944 6945 6946
	mlxsw_sp_rif_fdb_op(mlxsw_sp, macvlan_dev->dev_addr,
			    mlxsw_sp_fid_index(rif->fid), false);
}

6947 6948 6949 6950 6951 6952
void mlxsw_sp_rif_macvlan_del(struct mlxsw_sp *mlxsw_sp,
			      const struct net_device *macvlan_dev)
{
	__mlxsw_sp_rif_macvlan_del(mlxsw_sp, macvlan_dev);
}

6953 6954
static int mlxsw_sp_inetaddr_macvlan_event(struct mlxsw_sp *mlxsw_sp,
					   struct net_device *macvlan_dev,
6955 6956 6957 6958 6959 6960 6961
					   unsigned long event,
					   struct netlink_ext_ack *extack)
{
	switch (event) {
	case NETDEV_UP:
		return mlxsw_sp_rif_macvlan_add(mlxsw_sp, macvlan_dev, extack);
	case NETDEV_DOWN:
6962
		__mlxsw_sp_rif_macvlan_del(mlxsw_sp, macvlan_dev);
6963 6964 6965 6966 6967 6968
		break;
	}

	return 0;
}

6969 6970 6971 6972 6973 6974 6975 6976 6977 6978 6979
static int mlxsw_sp_router_port_check_rif_addr(struct mlxsw_sp *mlxsw_sp,
					       struct net_device *dev,
					       const unsigned char *dev_addr,
					       struct netlink_ext_ack *extack)
{
	struct mlxsw_sp_rif *rif;
	int i;

	/* A RIF is not created for macvlan netdevs. Their MAC is used to
	 * populate the FDB
	 */
6980
	if (netif_is_macvlan(dev) || netif_is_l3_master(dev))
6981 6982 6983 6984
		return 0;

	for (i = 0; i < MLXSW_CORE_RES_GET(mlxsw_sp->core, MAX_RIFS); i++) {
		rif = mlxsw_sp->router->rifs[i];
6985 6986 6987
		if (rif && rif->ops &&
		    rif->ops->type == MLXSW_SP_RIF_TYPE_IPIP_LB)
			continue;
6988
		if (rif && rif->dev && rif->dev != dev &&
6989 6990 6991 6992 6993 6994 6995 6996 6997 6998
		    !ether_addr_equal_masked(rif->dev->dev_addr, dev_addr,
					     mlxsw_sp->mac_mask)) {
			NL_SET_ERR_MSG_MOD(extack, "All router interface MAC addresses must have the same prefix");
			return -EINVAL;
		}
	}

	return 0;
}

6999 7000
static int __mlxsw_sp_inetaddr_event(struct mlxsw_sp *mlxsw_sp,
				     struct net_device *dev,
7001 7002
				     unsigned long event,
				     struct netlink_ext_ack *extack)
7003 7004
{
	if (mlxsw_sp_port_dev_check(dev))
7005
		return mlxsw_sp_inetaddr_port_event(dev, event, extack);
7006
	else if (netif_is_lag_master(dev))
7007
		return mlxsw_sp_inetaddr_lag_event(dev, event, extack);
7008
	else if (netif_is_bridge_master(dev))
7009 7010
		return mlxsw_sp_inetaddr_bridge_event(mlxsw_sp, dev, event,
						      extack);
7011
	else if (is_vlan_dev(dev))
7012 7013
		return mlxsw_sp_inetaddr_vlan_event(mlxsw_sp, dev, event,
						    extack);
7014
	else if (netif_is_macvlan(dev))
7015 7016
		return mlxsw_sp_inetaddr_macvlan_event(mlxsw_sp, dev, event,
						       extack);
7017 7018 7019 7020
	else
		return 0;
}

7021 7022
static int mlxsw_sp_inetaddr_event(struct notifier_block *nb,
				   unsigned long event, void *ptr)
7023 7024 7025
{
	struct in_ifaddr *ifa = (struct in_ifaddr *) ptr;
	struct net_device *dev = ifa->ifa_dev->dev;
7026
	struct mlxsw_sp_router *router;
7027
	struct mlxsw_sp_rif *rif;
7028 7029
	int err = 0;

7030 7031
	/* NETDEV_UP event is handled by mlxsw_sp_inetaddr_valid_event */
	if (event == NETDEV_UP)
7032
		return NOTIFY_DONE;
7033

7034
	router = container_of(nb, struct mlxsw_sp_router, inetaddr_nb);
7035
	mutex_lock(&router->lock);
7036
	rif = mlxsw_sp_rif_find_by_dev(router->mlxsw_sp, dev);
7037 7038 7039
	if (!mlxsw_sp_rif_should_config(rif, dev, event))
		goto out;

7040
	err = __mlxsw_sp_inetaddr_event(router->mlxsw_sp, dev, event, NULL);
7041
out:
7042
	mutex_unlock(&router->lock);
7043 7044 7045 7046 7047 7048 7049 7050 7051 7052 7053 7054
	return notifier_from_errno(err);
}

int mlxsw_sp_inetaddr_valid_event(struct notifier_block *unused,
				  unsigned long event, void *ptr)
{
	struct in_validator_info *ivi = (struct in_validator_info *) ptr;
	struct net_device *dev = ivi->ivi_dev->dev;
	struct mlxsw_sp *mlxsw_sp;
	struct mlxsw_sp_rif *rif;
	int err = 0;

7055 7056
	mlxsw_sp = mlxsw_sp_lower_get(dev);
	if (!mlxsw_sp)
7057
		return NOTIFY_DONE;
7058

7059
	mutex_lock(&mlxsw_sp->router->lock);
7060
	rif = mlxsw_sp_rif_find_by_dev(mlxsw_sp, dev);
7061
	if (!mlxsw_sp_rif_should_config(rif, dev, event))
7062 7063
		goto out;

7064 7065 7066 7067 7068
	err = mlxsw_sp_router_port_check_rif_addr(mlxsw_sp, dev, dev->dev_addr,
						  ivi->extack);
	if (err)
		goto out;

7069
	err = __mlxsw_sp_inetaddr_event(mlxsw_sp, dev, event, ivi->extack);
7070
out:
7071
	mutex_unlock(&mlxsw_sp->router->lock);
7072 7073 7074
	return notifier_from_errno(err);
}

7075 7076
struct mlxsw_sp_inet6addr_event_work {
	struct work_struct work;
7077
	struct mlxsw_sp *mlxsw_sp;
7078 7079 7080 7081 7082 7083 7084 7085
	struct net_device *dev;
	unsigned long event;
};

static void mlxsw_sp_inet6addr_event_work(struct work_struct *work)
{
	struct mlxsw_sp_inet6addr_event_work *inet6addr_work =
		container_of(work, struct mlxsw_sp_inet6addr_event_work, work);
7086
	struct mlxsw_sp *mlxsw_sp = inet6addr_work->mlxsw_sp;
7087 7088 7089 7090 7091
	struct net_device *dev = inet6addr_work->dev;
	unsigned long event = inet6addr_work->event;
	struct mlxsw_sp_rif *rif;

	rtnl_lock();
7092
	mutex_lock(&mlxsw_sp->router->lock);
7093 7094 7095 7096 7097

	rif = mlxsw_sp_rif_find_by_dev(mlxsw_sp, dev);
	if (!mlxsw_sp_rif_should_config(rif, dev, event))
		goto out;

7098
	__mlxsw_sp_inetaddr_event(mlxsw_sp, dev, event, NULL);
7099
out:
7100
	mutex_unlock(&mlxsw_sp->router->lock);
7101 7102 7103 7104 7105 7106
	rtnl_unlock();
	dev_put(dev);
	kfree(inet6addr_work);
}

/* Called with rcu_read_lock() */
7107 7108
static int mlxsw_sp_inet6addr_event(struct notifier_block *nb,
				    unsigned long event, void *ptr)
7109 7110 7111 7112
{
	struct inet6_ifaddr *if6 = (struct inet6_ifaddr *) ptr;
	struct mlxsw_sp_inet6addr_event_work *inet6addr_work;
	struct net_device *dev = if6->idev->dev;
7113
	struct mlxsw_sp_router *router;
7114

7115 7116 7117 7118
	/* NETDEV_UP event is handled by mlxsw_sp_inet6addr_valid_event */
	if (event == NETDEV_UP)
		return NOTIFY_DONE;

7119 7120 7121 7122
	inet6addr_work = kzalloc(sizeof(*inet6addr_work), GFP_ATOMIC);
	if (!inet6addr_work)
		return NOTIFY_BAD;

7123
	router = container_of(nb, struct mlxsw_sp_router, inet6addr_nb);
7124
	INIT_WORK(&inet6addr_work->work, mlxsw_sp_inet6addr_event_work);
7125
	inet6addr_work->mlxsw_sp = router->mlxsw_sp;
7126 7127 7128 7129 7130 7131 7132 7133
	inet6addr_work->dev = dev;
	inet6addr_work->event = event;
	dev_hold(dev);
	mlxsw_core_schedule_work(&inet6addr_work->work);

	return NOTIFY_DONE;
}

7134 7135 7136 7137 7138 7139 7140 7141 7142 7143 7144
int mlxsw_sp_inet6addr_valid_event(struct notifier_block *unused,
				   unsigned long event, void *ptr)
{
	struct in6_validator_info *i6vi = (struct in6_validator_info *) ptr;
	struct net_device *dev = i6vi->i6vi_dev->dev;
	struct mlxsw_sp *mlxsw_sp;
	struct mlxsw_sp_rif *rif;
	int err = 0;

	mlxsw_sp = mlxsw_sp_lower_get(dev);
	if (!mlxsw_sp)
7145
		return NOTIFY_DONE;
7146

7147
	mutex_lock(&mlxsw_sp->router->lock);
7148 7149 7150 7151
	rif = mlxsw_sp_rif_find_by_dev(mlxsw_sp, dev);
	if (!mlxsw_sp_rif_should_config(rif, dev, event))
		goto out;

7152 7153 7154 7155 7156
	err = mlxsw_sp_router_port_check_rif_addr(mlxsw_sp, dev, dev->dev_addr,
						  i6vi->extack);
	if (err)
		goto out;

7157
	err = __mlxsw_sp_inetaddr_event(mlxsw_sp, dev, event, i6vi->extack);
7158
out:
7159
	mutex_unlock(&mlxsw_sp->router->lock);
7160 7161 7162
	return notifier_from_errno(err);
}

7163
static int mlxsw_sp_rif_edit(struct mlxsw_sp *mlxsw_sp, u16 rif_index,
7164 7165 7166 7167 7168
			     const char *mac, int mtu)
{
	char ritr_pl[MLXSW_REG_RITR_LEN];
	int err;

7169
	mlxsw_reg_ritr_rif_pack(ritr_pl, rif_index);
7170 7171 7172 7173 7174 7175 7176 7177 7178 7179
	err = mlxsw_reg_query(mlxsw_sp->core, MLXSW_REG(ritr), ritr_pl);
	if (err)
		return err;

	mlxsw_reg_ritr_mtu_set(ritr_pl, mtu);
	mlxsw_reg_ritr_if_mac_memcpy_to(ritr_pl, mac);
	mlxsw_reg_ritr_op_set(ritr_pl, MLXSW_REG_RITR_RIF_CREATE);
	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ritr), ritr_pl);
}

7180 7181 7182
static int
mlxsw_sp_router_port_change_event(struct mlxsw_sp *mlxsw_sp,
				  struct mlxsw_sp_rif *rif)
7183
{
7184
	struct net_device *dev = rif->dev;
7185
	u16 fid_index;
7186 7187
	int err;

7188
	fid_index = mlxsw_sp_fid_index(rif->fid);
7189

7190
	err = mlxsw_sp_rif_fdb_op(mlxsw_sp, rif->addr, fid_index, false);
7191 7192 7193
	if (err)
		return err;

7194 7195
	err = mlxsw_sp_rif_edit(mlxsw_sp, rif->rif_index, dev->dev_addr,
				dev->mtu);
7196 7197 7198
	if (err)
		goto err_rif_edit;

7199
	err = mlxsw_sp_rif_fdb_op(mlxsw_sp, dev->dev_addr, fid_index, true);
7200 7201 7202
	if (err)
		goto err_rif_fdb_op;

7203 7204
	if (rif->mtu != dev->mtu) {
		struct mlxsw_sp_vr *vr;
7205
		int i;
7206 7207 7208 7209 7210 7211

		/* The RIF is relevant only to its mr_table instance, as unlike
		 * unicast routing, in multicast routing a RIF cannot be shared
		 * between several multicast routing tables.
		 */
		vr = &mlxsw_sp->router->vrs[rif->vr_id];
7212 7213 7214
		for (i = 0; i < MLXSW_SP_L3_PROTO_MAX; i++)
			mlxsw_sp_mr_rif_mtu_update(vr->mr_table[i],
						   rif, dev->mtu);
7215 7216
	}

7217 7218
	ether_addr_copy(rif->addr, dev->dev_addr);
	rif->mtu = dev->mtu;
7219

7220
	netdev_dbg(dev, "Updated RIF=%d\n", rif->rif_index);
7221 7222 7223 7224

	return 0;

err_rif_fdb_op:
7225
	mlxsw_sp_rif_edit(mlxsw_sp, rif->rif_index, rif->addr, rif->mtu);
7226
err_rif_edit:
7227
	mlxsw_sp_rif_fdb_op(mlxsw_sp, rif->addr, fid_index, true);
7228 7229 7230
	return err;
}

7231 7232 7233 7234 7235 7236 7237 7238 7239 7240
static int mlxsw_sp_router_port_pre_changeaddr_event(struct mlxsw_sp_rif *rif,
			    struct netdev_notifier_pre_changeaddr_info *info)
{
	struct netlink_ext_ack *extack;

	extack = netdev_notifier_info_to_extack(&info->info);
	return mlxsw_sp_router_port_check_rif_addr(rif->mlxsw_sp, rif->dev,
						   info->dev_addr, extack);
}

7241 7242 7243 7244 7245
int mlxsw_sp_netdevice_router_port_event(struct net_device *dev,
					 unsigned long event, void *ptr)
{
	struct mlxsw_sp *mlxsw_sp;
	struct mlxsw_sp_rif *rif;
7246
	int err = 0;
7247 7248 7249 7250 7251

	mlxsw_sp = mlxsw_sp_lower_get(dev);
	if (!mlxsw_sp)
		return 0;

7252
	mutex_lock(&mlxsw_sp->router->lock);
7253 7254
	rif = mlxsw_sp_rif_find_by_dev(mlxsw_sp, dev);
	if (!rif)
7255
		goto out;
7256 7257 7258 7259

	switch (event) {
	case NETDEV_CHANGEMTU: /* fall through */
	case NETDEV_CHANGEADDR:
7260 7261
		err = mlxsw_sp_router_port_change_event(mlxsw_sp, rif);
		break;
7262
	case NETDEV_PRE_CHANGEADDR:
7263 7264
		err = mlxsw_sp_router_port_pre_changeaddr_event(rif, ptr);
		break;
7265 7266
	}

7267 7268 7269
out:
	mutex_unlock(&mlxsw_sp->router->lock);
	return err;
7270 7271
}

7272
static int mlxsw_sp_port_vrf_join(struct mlxsw_sp *mlxsw_sp,
7273 7274
				  struct net_device *l3_dev,
				  struct netlink_ext_ack *extack)
7275
{
7276
	struct mlxsw_sp_rif *rif;
7277

7278 7279
	/* If netdev is already associated with a RIF, then we need to
	 * destroy it and create a new one with the new virtual router ID.
7280
	 */
7281 7282
	rif = mlxsw_sp_rif_find_by_dev(mlxsw_sp, l3_dev);
	if (rif)
7283 7284
		__mlxsw_sp_inetaddr_event(mlxsw_sp, l3_dev, NETDEV_DOWN,
					  extack);
7285

7286
	return __mlxsw_sp_inetaddr_event(mlxsw_sp, l3_dev, NETDEV_UP, extack);
7287 7288
}

7289 7290
static void mlxsw_sp_port_vrf_leave(struct mlxsw_sp *mlxsw_sp,
				    struct net_device *l3_dev)
7291
{
7292
	struct mlxsw_sp_rif *rif;
7293

7294 7295
	rif = mlxsw_sp_rif_find_by_dev(mlxsw_sp, l3_dev);
	if (!rif)
7296
		return;
7297
	__mlxsw_sp_inetaddr_event(mlxsw_sp, l3_dev, NETDEV_DOWN, NULL);
7298 7299
}

7300 7301
int mlxsw_sp_netdevice_vrf_event(struct net_device *l3_dev, unsigned long event,
				 struct netdev_notifier_changeupper_info *info)
7302
{
7303 7304
	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_lower_get(l3_dev);
	int err = 0;
7305

7306 7307 7308 7309
	/* We do not create a RIF for a macvlan, but only use it to
	 * direct more MAC addresses to the router.
	 */
	if (!mlxsw_sp || netif_is_macvlan(l3_dev))
7310
		return 0;
7311

7312
	mutex_lock(&mlxsw_sp->router->lock);
7313 7314
	switch (event) {
	case NETDEV_PRECHANGEUPPER:
7315
		break;
7316
	case NETDEV_CHANGEUPPER:
7317 7318 7319 7320 7321 7322
		if (info->linking) {
			struct netlink_ext_ack *extack;

			extack = netdev_notifier_info_to_extack(&info->info);
			err = mlxsw_sp_port_vrf_join(mlxsw_sp, l3_dev, extack);
		} else {
7323
			mlxsw_sp_port_vrf_leave(mlxsw_sp, l3_dev);
7324
		}
7325 7326
		break;
	}
7327
	mutex_unlock(&mlxsw_sp->router->lock);
7328

7329
	return err;
7330 7331
}

7332 7333 7334 7335 7336 7337 7338 7339 7340 7341 7342 7343 7344 7345 7346 7347 7348 7349 7350 7351 7352
static int __mlxsw_sp_rif_macvlan_flush(struct net_device *dev, void *data)
{
	struct mlxsw_sp_rif *rif = data;

	if (!netif_is_macvlan(dev))
		return 0;

	return mlxsw_sp_rif_fdb_op(rif->mlxsw_sp, dev->dev_addr,
				   mlxsw_sp_fid_index(rif->fid), false);
}

static int mlxsw_sp_rif_macvlan_flush(struct mlxsw_sp_rif *rif)
{
	if (!netif_is_macvlan_port(rif->dev))
		return 0;

	netdev_warn(rif->dev, "Router interface is deleted. Upper macvlans will not work\n");
	return netdev_walk_all_upper_dev_rcu(rif->dev,
					     __mlxsw_sp_rif_macvlan_flush, rif);
}

7353 7354 7355 7356 7357 7358
static void mlxsw_sp_rif_subport_setup(struct mlxsw_sp_rif *rif,
				       const struct mlxsw_sp_rif_params *params)
{
	struct mlxsw_sp_rif_subport *rif_subport;

	rif_subport = mlxsw_sp_rif_subport_rif(rif);
7359
	refcount_set(&rif_subport->ref_count, 1);
7360 7361 7362 7363
	rif_subport->vid = params->vid;
	rif_subport->lag = params->lag;
	if (params->lag)
		rif_subport->lag_id = params->lag_id;
7364
	else
7365 7366 7367 7368 7369 7370 7371 7372 7373 7374 7375
		rif_subport->system_port = params->system_port;
}

static int mlxsw_sp_rif_subport_op(struct mlxsw_sp_rif *rif, bool enable)
{
	struct mlxsw_sp *mlxsw_sp = rif->mlxsw_sp;
	struct mlxsw_sp_rif_subport *rif_subport;
	char ritr_pl[MLXSW_REG_RITR_LEN];

	rif_subport = mlxsw_sp_rif_subport_rif(rif);
	mlxsw_reg_ritr_pack(ritr_pl, enable, MLXSW_REG_RITR_SP_IF,
7376 7377
			    rif->rif_index, rif->vr_id, rif->dev->mtu);
	mlxsw_reg_ritr_mac_pack(ritr_pl, rif->dev->dev_addr);
7378 7379 7380 7381 7382 7383 7384 7385 7386 7387
	mlxsw_reg_ritr_sp_if_pack(ritr_pl, rif_subport->lag,
				  rif_subport->lag ? rif_subport->lag_id :
						     rif_subport->system_port,
				  rif_subport->vid);

	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ritr), ritr_pl);
}

static int mlxsw_sp_rif_subport_configure(struct mlxsw_sp_rif *rif)
{
7388 7389 7390 7391 7392 7393 7394 7395 7396 7397 7398 7399 7400 7401 7402 7403 7404
	int err;

	err = mlxsw_sp_rif_subport_op(rif, true);
	if (err)
		return err;

	err = mlxsw_sp_rif_fdb_op(rif->mlxsw_sp, rif->dev->dev_addr,
				  mlxsw_sp_fid_index(rif->fid), true);
	if (err)
		goto err_rif_fdb_op;

	mlxsw_sp_fid_rif_set(rif->fid, rif);
	return 0;

err_rif_fdb_op:
	mlxsw_sp_rif_subport_op(rif, false);
	return err;
7405 7406
}

7407 7408
static void mlxsw_sp_rif_subport_deconfigure(struct mlxsw_sp_rif *rif)
{
7409 7410 7411 7412 7413
	struct mlxsw_sp_fid *fid = rif->fid;

	mlxsw_sp_fid_rif_set(fid, NULL);
	mlxsw_sp_rif_fdb_op(rif->mlxsw_sp, rif->dev->dev_addr,
			    mlxsw_sp_fid_index(fid), false);
7414
	mlxsw_sp_rif_macvlan_flush(rif);
7415 7416 7417 7418
	mlxsw_sp_rif_subport_op(rif, false);
}

static struct mlxsw_sp_fid *
7419 7420
mlxsw_sp_rif_subport_fid_get(struct mlxsw_sp_rif *rif,
			     struct netlink_ext_ack *extack)
7421 7422 7423 7424 7425 7426 7427 7428 7429 7430 7431 7432 7433 7434 7435 7436 7437 7438 7439 7440 7441
{
	return mlxsw_sp_fid_rfid_get(rif->mlxsw_sp, rif->rif_index);
}

static const struct mlxsw_sp_rif_ops mlxsw_sp_rif_subport_ops = {
	.type			= MLXSW_SP_RIF_TYPE_SUBPORT,
	.rif_size		= sizeof(struct mlxsw_sp_rif_subport),
	.setup			= mlxsw_sp_rif_subport_setup,
	.configure		= mlxsw_sp_rif_subport_configure,
	.deconfigure		= mlxsw_sp_rif_subport_deconfigure,
	.fid_get		= mlxsw_sp_rif_subport_fid_get,
};

static int mlxsw_sp_rif_vlan_fid_op(struct mlxsw_sp_rif *rif,
				    enum mlxsw_reg_ritr_if_type type,
				    u16 vid_fid, bool enable)
{
	struct mlxsw_sp *mlxsw_sp = rif->mlxsw_sp;
	char ritr_pl[MLXSW_REG_RITR_LEN];

	mlxsw_reg_ritr_pack(ritr_pl, enable, type, rif->rif_index, rif->vr_id,
7442 7443
			    rif->dev->mtu);
	mlxsw_reg_ritr_mac_pack(ritr_pl, rif->dev->dev_addr);
7444 7445 7446 7447 7448
	mlxsw_reg_ritr_fid_set(ritr_pl, type, vid_fid);

	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ritr), ritr_pl);
}

7449
u8 mlxsw_sp_router_port(const struct mlxsw_sp *mlxsw_sp)
7450 7451 7452 7453 7454 7455 7456 7457 7458 7459 7460 7461 7462 7463
{
	return mlxsw_core_max_ports(mlxsw_sp->core) + 1;
}

static int mlxsw_sp_rif_vlan_configure(struct mlxsw_sp_rif *rif)
{
	struct mlxsw_sp *mlxsw_sp = rif->mlxsw_sp;
	u16 vid = mlxsw_sp_fid_8021q_vid(rif->fid);
	int err;

	err = mlxsw_sp_rif_vlan_fid_op(rif, MLXSW_REG_RITR_VLAN_IF, vid, true);
	if (err)
		return err;

7464 7465 7466 7467 7468
	err = mlxsw_sp_fid_flood_set(rif->fid, MLXSW_SP_FLOOD_TYPE_MC,
				     mlxsw_sp_router_port(mlxsw_sp), true);
	if (err)
		goto err_fid_mc_flood_set;

7469 7470 7471 7472 7473
	err = mlxsw_sp_fid_flood_set(rif->fid, MLXSW_SP_FLOOD_TYPE_BC,
				     mlxsw_sp_router_port(mlxsw_sp), true);
	if (err)
		goto err_fid_bc_flood_set;

7474 7475 7476 7477 7478 7479
	err = mlxsw_sp_rif_fdb_op(rif->mlxsw_sp, rif->dev->dev_addr,
				  mlxsw_sp_fid_index(rif->fid), true);
	if (err)
		goto err_rif_fdb_op;

	mlxsw_sp_fid_rif_set(rif->fid, rif);
7480 7481
	return 0;

7482 7483 7484
err_rif_fdb_op:
	mlxsw_sp_fid_flood_set(rif->fid, MLXSW_SP_FLOOD_TYPE_BC,
			       mlxsw_sp_router_port(mlxsw_sp), false);
7485
err_fid_bc_flood_set:
7486 7487 7488
	mlxsw_sp_fid_flood_set(rif->fid, MLXSW_SP_FLOOD_TYPE_MC,
			       mlxsw_sp_router_port(mlxsw_sp), false);
err_fid_mc_flood_set:
7489 7490 7491 7492 7493 7494 7495
	mlxsw_sp_rif_vlan_fid_op(rif, MLXSW_REG_RITR_VLAN_IF, vid, false);
	return err;
}

static void mlxsw_sp_rif_vlan_deconfigure(struct mlxsw_sp_rif *rif)
{
	u16 vid = mlxsw_sp_fid_8021q_vid(rif->fid);
7496 7497
	struct mlxsw_sp *mlxsw_sp = rif->mlxsw_sp;
	struct mlxsw_sp_fid *fid = rif->fid;
7498

7499 7500 7501
	mlxsw_sp_fid_rif_set(fid, NULL);
	mlxsw_sp_rif_fdb_op(rif->mlxsw_sp, rif->dev->dev_addr,
			    mlxsw_sp_fid_index(fid), false);
7502
	mlxsw_sp_rif_macvlan_flush(rif);
7503 7504
	mlxsw_sp_fid_flood_set(rif->fid, MLXSW_SP_FLOOD_TYPE_BC,
			       mlxsw_sp_router_port(mlxsw_sp), false);
7505 7506
	mlxsw_sp_fid_flood_set(rif->fid, MLXSW_SP_FLOOD_TYPE_MC,
			       mlxsw_sp_router_port(mlxsw_sp), false);
7507 7508 7509 7510
	mlxsw_sp_rif_vlan_fid_op(rif, MLXSW_REG_RITR_VLAN_IF, vid, false);
}

static struct mlxsw_sp_fid *
7511 7512
mlxsw_sp_rif_vlan_fid_get(struct mlxsw_sp_rif *rif,
			  struct netlink_ext_ack *extack)
7513
{
7514
	struct net_device *br_dev = rif->dev;
7515 7516 7517 7518 7519
	u16 vid;
	int err;

	if (is_vlan_dev(rif->dev)) {
		vid = vlan_dev_vlan_id(rif->dev);
7520 7521 7522
		br_dev = vlan_dev_real_dev(rif->dev);
		if (WARN_ON(!netif_is_bridge_master(br_dev)))
			return ERR_PTR(-EINVAL);
7523 7524
	} else {
		err = br_vlan_get_pvid(rif->dev, &vid);
7525
		if (err < 0 || !vid) {
7526
			NL_SET_ERR_MSG_MOD(extack, "Couldn't determine bridge PVID");
7527
			return ERR_PTR(-EINVAL);
7528 7529
		}
	}
7530

7531
	return mlxsw_sp_fid_8021q_get(rif->mlxsw_sp, vid);
7532 7533
}

7534 7535 7536 7537 7538 7539 7540 7541 7542 7543 7544 7545 7546 7547
static void mlxsw_sp_rif_vlan_fdb_del(struct mlxsw_sp_rif *rif, const char *mac)
{
	u16 vid = mlxsw_sp_fid_8021q_vid(rif->fid);
	struct switchdev_notifier_fdb_info info;
	struct net_device *br_dev;
	struct net_device *dev;

	br_dev = is_vlan_dev(rif->dev) ? vlan_dev_real_dev(rif->dev) : rif->dev;
	dev = br_fdb_find_port(br_dev, mac, vid);
	if (!dev)
		return;

	info.addr = mac;
	info.vid = vid;
7548 7549
	call_switchdev_notifiers(SWITCHDEV_FDB_DEL_TO_BRIDGE, dev, &info.info,
				 NULL);
7550 7551
}

7552 7553 7554 7555 7556 7557
static const struct mlxsw_sp_rif_ops mlxsw_sp_rif_vlan_ops = {
	.type			= MLXSW_SP_RIF_TYPE_VLAN,
	.rif_size		= sizeof(struct mlxsw_sp_rif),
	.configure		= mlxsw_sp_rif_vlan_configure,
	.deconfigure		= mlxsw_sp_rif_vlan_deconfigure,
	.fid_get		= mlxsw_sp_rif_vlan_fid_get,
7558
	.fdb_del		= mlxsw_sp_rif_vlan_fdb_del,
7559 7560 7561 7562 7563 7564 7565 7566 7567 7568 7569 7570 7571
};

static int mlxsw_sp_rif_fid_configure(struct mlxsw_sp_rif *rif)
{
	struct mlxsw_sp *mlxsw_sp = rif->mlxsw_sp;
	u16 fid_index = mlxsw_sp_fid_index(rif->fid);
	int err;

	err = mlxsw_sp_rif_vlan_fid_op(rif, MLXSW_REG_RITR_FID_IF, fid_index,
				       true);
	if (err)
		return err;

7572 7573 7574 7575 7576
	err = mlxsw_sp_fid_flood_set(rif->fid, MLXSW_SP_FLOOD_TYPE_MC,
				     mlxsw_sp_router_port(mlxsw_sp), true);
	if (err)
		goto err_fid_mc_flood_set;

7577 7578 7579 7580 7581
	err = mlxsw_sp_fid_flood_set(rif->fid, MLXSW_SP_FLOOD_TYPE_BC,
				     mlxsw_sp_router_port(mlxsw_sp), true);
	if (err)
		goto err_fid_bc_flood_set;

7582 7583 7584 7585 7586 7587
	err = mlxsw_sp_rif_fdb_op(rif->mlxsw_sp, rif->dev->dev_addr,
				  mlxsw_sp_fid_index(rif->fid), true);
	if (err)
		goto err_rif_fdb_op;

	mlxsw_sp_fid_rif_set(rif->fid, rif);
7588 7589
	return 0;

7590 7591 7592
err_rif_fdb_op:
	mlxsw_sp_fid_flood_set(rif->fid, MLXSW_SP_FLOOD_TYPE_BC,
			       mlxsw_sp_router_port(mlxsw_sp), false);
7593
err_fid_bc_flood_set:
7594 7595 7596
	mlxsw_sp_fid_flood_set(rif->fid, MLXSW_SP_FLOOD_TYPE_MC,
			       mlxsw_sp_router_port(mlxsw_sp), false);
err_fid_mc_flood_set:
7597 7598 7599 7600 7601 7602 7603
	mlxsw_sp_rif_vlan_fid_op(rif, MLXSW_REG_RITR_FID_IF, fid_index, false);
	return err;
}

static void mlxsw_sp_rif_fid_deconfigure(struct mlxsw_sp_rif *rif)
{
	u16 fid_index = mlxsw_sp_fid_index(rif->fid);
7604 7605
	struct mlxsw_sp *mlxsw_sp = rif->mlxsw_sp;
	struct mlxsw_sp_fid *fid = rif->fid;
7606

7607 7608 7609
	mlxsw_sp_fid_rif_set(fid, NULL);
	mlxsw_sp_rif_fdb_op(rif->mlxsw_sp, rif->dev->dev_addr,
			    mlxsw_sp_fid_index(fid), false);
7610
	mlxsw_sp_rif_macvlan_flush(rif);
7611 7612
	mlxsw_sp_fid_flood_set(rif->fid, MLXSW_SP_FLOOD_TYPE_BC,
			       mlxsw_sp_router_port(mlxsw_sp), false);
7613 7614
	mlxsw_sp_fid_flood_set(rif->fid, MLXSW_SP_FLOOD_TYPE_MC,
			       mlxsw_sp_router_port(mlxsw_sp), false);
7615 7616 7617 7618
	mlxsw_sp_rif_vlan_fid_op(rif, MLXSW_REG_RITR_FID_IF, fid_index, false);
}

static struct mlxsw_sp_fid *
7619 7620
mlxsw_sp_rif_fid_fid_get(struct mlxsw_sp_rif *rif,
			 struct netlink_ext_ack *extack)
7621
{
7622
	return mlxsw_sp_fid_8021d_get(rif->mlxsw_sp, rif->dev->ifindex);
7623 7624
}

7625 7626 7627 7628 7629 7630 7631 7632 7633 7634 7635
static void mlxsw_sp_rif_fid_fdb_del(struct mlxsw_sp_rif *rif, const char *mac)
{
	struct switchdev_notifier_fdb_info info;
	struct net_device *dev;

	dev = br_fdb_find_port(rif->dev, mac, 0);
	if (!dev)
		return;

	info.addr = mac;
	info.vid = 0;
7636 7637
	call_switchdev_notifiers(SWITCHDEV_FDB_DEL_TO_BRIDGE, dev, &info.info,
				 NULL);
7638 7639
}

7640 7641 7642 7643 7644 7645
static const struct mlxsw_sp_rif_ops mlxsw_sp_rif_fid_ops = {
	.type			= MLXSW_SP_RIF_TYPE_FID,
	.rif_size		= sizeof(struct mlxsw_sp_rif),
	.configure		= mlxsw_sp_rif_fid_configure,
	.deconfigure		= mlxsw_sp_rif_fid_deconfigure,
	.fid_get		= mlxsw_sp_rif_fid_fid_get,
7646
	.fdb_del		= mlxsw_sp_rif_fid_fdb_del,
7647 7648
};

7649 7650 7651 7652 7653 7654 7655 7656 7657
static const struct mlxsw_sp_rif_ops mlxsw_sp_rif_vlan_emu_ops = {
	.type			= MLXSW_SP_RIF_TYPE_VLAN,
	.rif_size		= sizeof(struct mlxsw_sp_rif),
	.configure		= mlxsw_sp_rif_fid_configure,
	.deconfigure		= mlxsw_sp_rif_fid_deconfigure,
	.fid_get		= mlxsw_sp_rif_vlan_fid_get,
	.fdb_del		= mlxsw_sp_rif_vlan_fdb_del,
};

7658 7659 7660 7661 7662 7663 7664 7665 7666 7667 7668 7669 7670 7671 7672 7673 7674 7675 7676 7677
static struct mlxsw_sp_rif_ipip_lb *
mlxsw_sp_rif_ipip_lb_rif(struct mlxsw_sp_rif *rif)
{
	return container_of(rif, struct mlxsw_sp_rif_ipip_lb, common);
}

static void
mlxsw_sp_rif_ipip_lb_setup(struct mlxsw_sp_rif *rif,
			   const struct mlxsw_sp_rif_params *params)
{
	struct mlxsw_sp_rif_params_ipip_lb *params_lb;
	struct mlxsw_sp_rif_ipip_lb *rif_lb;

	params_lb = container_of(params, struct mlxsw_sp_rif_params_ipip_lb,
				 common);
	rif_lb = mlxsw_sp_rif_ipip_lb_rif(rif);
	rif_lb->lb_config = params_lb->lb_config;
}

static int
7678
mlxsw_sp1_rif_ipip_lb_configure(struct mlxsw_sp_rif *rif)
7679 7680 7681 7682 7683 7684 7685
{
	struct mlxsw_sp_rif_ipip_lb *lb_rif = mlxsw_sp_rif_ipip_lb_rif(rif);
	u32 ul_tb_id = mlxsw_sp_ipip_dev_ul_tb_id(rif->dev);
	struct mlxsw_sp *mlxsw_sp = rif->mlxsw_sp;
	struct mlxsw_sp_vr *ul_vr;
	int err;

7686
	ul_vr = mlxsw_sp_vr_get(mlxsw_sp, ul_tb_id, NULL);
7687 7688 7689
	if (IS_ERR(ul_vr))
		return PTR_ERR(ul_vr);

7690
	err = mlxsw_sp_rif_ipip_lb_op(lb_rif, ul_vr->id, 0, true);
7691 7692 7693 7694
	if (err)
		goto err_loopback_op;

	lb_rif->ul_vr_id = ul_vr->id;
7695
	lb_rif->ul_rif_id = 0;
7696 7697 7698 7699
	++ul_vr->rif_count;
	return 0;

err_loopback_op:
7700
	mlxsw_sp_vr_put(mlxsw_sp, ul_vr);
7701 7702 7703
	return err;
}

7704
static void mlxsw_sp1_rif_ipip_lb_deconfigure(struct mlxsw_sp_rif *rif)
7705 7706 7707 7708 7709 7710
{
	struct mlxsw_sp_rif_ipip_lb *lb_rif = mlxsw_sp_rif_ipip_lb_rif(rif);
	struct mlxsw_sp *mlxsw_sp = rif->mlxsw_sp;
	struct mlxsw_sp_vr *ul_vr;

	ul_vr = &mlxsw_sp->router->vrs[lb_rif->ul_vr_id];
7711
	mlxsw_sp_rif_ipip_lb_op(lb_rif, ul_vr->id, 0, false);
7712 7713

	--ul_vr->rif_count;
7714
	mlxsw_sp_vr_put(mlxsw_sp, ul_vr);
7715 7716
}

7717
static const struct mlxsw_sp_rif_ops mlxsw_sp1_rif_ipip_lb_ops = {
7718 7719 7720
	.type			= MLXSW_SP_RIF_TYPE_IPIP_LB,
	.rif_size		= sizeof(struct mlxsw_sp_rif_ipip_lb),
	.setup                  = mlxsw_sp_rif_ipip_lb_setup,
7721 7722
	.configure		= mlxsw_sp1_rif_ipip_lb_configure,
	.deconfigure		= mlxsw_sp1_rif_ipip_lb_deconfigure,
7723 7724
};

7725
const struct mlxsw_sp_rif_ops *mlxsw_sp1_rif_ops_arr[] = {
7726
	[MLXSW_SP_RIF_TYPE_SUBPORT]	= &mlxsw_sp_rif_subport_ops,
7727
	[MLXSW_SP_RIF_TYPE_VLAN]	= &mlxsw_sp_rif_vlan_emu_ops,
7728
	[MLXSW_SP_RIF_TYPE_FID]		= &mlxsw_sp_rif_fid_ops,
7729 7730 7731
	[MLXSW_SP_RIF_TYPE_IPIP_LB]	= &mlxsw_sp1_rif_ipip_lb_ops,
};

7732 7733 7734 7735 7736 7737 7738 7739 7740 7741 7742 7743 7744 7745 7746 7747 7748 7749 7750 7751 7752 7753 7754 7755 7756 7757 7758 7759 7760 7761 7762 7763 7764 7765 7766 7767 7768 7769 7770 7771 7772 7773 7774 7775 7776 7777 7778 7779 7780 7781 7782 7783 7784 7785 7786 7787 7788 7789 7790 7791 7792 7793 7794 7795 7796 7797 7798 7799 7800 7801 7802 7803 7804 7805 7806 7807 7808 7809 7810 7811 7812 7813 7814 7815 7816 7817 7818 7819 7820 7821 7822 7823 7824 7825 7826 7827 7828 7829 7830 7831
static int
mlxsw_sp_rif_ipip_lb_ul_rif_op(struct mlxsw_sp_rif *ul_rif, bool enable)
{
	struct mlxsw_sp *mlxsw_sp = ul_rif->mlxsw_sp;
	char ritr_pl[MLXSW_REG_RITR_LEN];

	mlxsw_reg_ritr_pack(ritr_pl, enable, MLXSW_REG_RITR_LOOPBACK_IF,
			    ul_rif->rif_index, ul_rif->vr_id, IP_MAX_MTU);
	mlxsw_reg_ritr_loopback_protocol_set(ritr_pl,
					     MLXSW_REG_RITR_LOOPBACK_GENERIC);

	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ritr), ritr_pl);
}

static struct mlxsw_sp_rif *
mlxsw_sp_ul_rif_create(struct mlxsw_sp *mlxsw_sp, struct mlxsw_sp_vr *vr,
		       struct netlink_ext_ack *extack)
{
	struct mlxsw_sp_rif *ul_rif;
	u16 rif_index;
	int err;

	err = mlxsw_sp_rif_index_alloc(mlxsw_sp, &rif_index);
	if (err) {
		NL_SET_ERR_MSG_MOD(extack, "Exceeded number of supported router interfaces");
		return ERR_PTR(err);
	}

	ul_rif = mlxsw_sp_rif_alloc(sizeof(*ul_rif), rif_index, vr->id, NULL);
	if (!ul_rif)
		return ERR_PTR(-ENOMEM);

	mlxsw_sp->router->rifs[rif_index] = ul_rif;
	ul_rif->mlxsw_sp = mlxsw_sp;
	err = mlxsw_sp_rif_ipip_lb_ul_rif_op(ul_rif, true);
	if (err)
		goto ul_rif_op_err;

	return ul_rif;

ul_rif_op_err:
	mlxsw_sp->router->rifs[rif_index] = NULL;
	kfree(ul_rif);
	return ERR_PTR(err);
}

static void mlxsw_sp_ul_rif_destroy(struct mlxsw_sp_rif *ul_rif)
{
	struct mlxsw_sp *mlxsw_sp = ul_rif->mlxsw_sp;

	mlxsw_sp_rif_ipip_lb_ul_rif_op(ul_rif, false);
	mlxsw_sp->router->rifs[ul_rif->rif_index] = NULL;
	kfree(ul_rif);
}

static struct mlxsw_sp_rif *
mlxsw_sp_ul_rif_get(struct mlxsw_sp *mlxsw_sp, u32 tb_id,
		    struct netlink_ext_ack *extack)
{
	struct mlxsw_sp_vr *vr;
	int err;

	vr = mlxsw_sp_vr_get(mlxsw_sp, tb_id, extack);
	if (IS_ERR(vr))
		return ERR_CAST(vr);

	if (refcount_inc_not_zero(&vr->ul_rif_refcnt))
		return vr->ul_rif;

	vr->ul_rif = mlxsw_sp_ul_rif_create(mlxsw_sp, vr, extack);
	if (IS_ERR(vr->ul_rif)) {
		err = PTR_ERR(vr->ul_rif);
		goto err_ul_rif_create;
	}

	vr->rif_count++;
	refcount_set(&vr->ul_rif_refcnt, 1);

	return vr->ul_rif;

err_ul_rif_create:
	mlxsw_sp_vr_put(mlxsw_sp, vr);
	return ERR_PTR(err);
}

static void mlxsw_sp_ul_rif_put(struct mlxsw_sp_rif *ul_rif)
{
	struct mlxsw_sp *mlxsw_sp = ul_rif->mlxsw_sp;
	struct mlxsw_sp_vr *vr;

	vr = &mlxsw_sp->router->vrs[ul_rif->vr_id];

	if (!refcount_dec_and_test(&vr->ul_rif_refcnt))
		return;

	vr->rif_count--;
	mlxsw_sp_ul_rif_destroy(ul_rif);
	mlxsw_sp_vr_put(mlxsw_sp, vr);
}

7832 7833 7834 7835 7836 7837 7838 7839 7840 7841 7842 7843 7844 7845 7846 7847 7848 7849 7850 7851 7852 7853 7854 7855 7856 7857 7858 7859
int mlxsw_sp_router_ul_rif_get(struct mlxsw_sp *mlxsw_sp, u32 ul_tb_id,
			       u16 *ul_rif_index)
{
	struct mlxsw_sp_rif *ul_rif;

	ASSERT_RTNL();

	ul_rif = mlxsw_sp_ul_rif_get(mlxsw_sp, ul_tb_id, NULL);
	if (IS_ERR(ul_rif))
		return PTR_ERR(ul_rif);
	*ul_rif_index = ul_rif->rif_index;

	return 0;
}

void mlxsw_sp_router_ul_rif_put(struct mlxsw_sp *mlxsw_sp, u16 ul_rif_index)
{
	struct mlxsw_sp_rif *ul_rif;

	ASSERT_RTNL();

	ul_rif = mlxsw_sp->router->rifs[ul_rif_index];
	if (WARN_ON(!ul_rif))
		return;

	mlxsw_sp_ul_rif_put(ul_rif);
}

7860 7861 7862
static int
mlxsw_sp2_rif_ipip_lb_configure(struct mlxsw_sp_rif *rif)
{
7863 7864 7865 7866 7867 7868 7869 7870 7871 7872 7873 7874 7875 7876 7877 7878 7879
	struct mlxsw_sp_rif_ipip_lb *lb_rif = mlxsw_sp_rif_ipip_lb_rif(rif);
	u32 ul_tb_id = mlxsw_sp_ipip_dev_ul_tb_id(rif->dev);
	struct mlxsw_sp *mlxsw_sp = rif->mlxsw_sp;
	struct mlxsw_sp_rif *ul_rif;
	int err;

	ul_rif = mlxsw_sp_ul_rif_get(mlxsw_sp, ul_tb_id, NULL);
	if (IS_ERR(ul_rif))
		return PTR_ERR(ul_rif);

	err = mlxsw_sp_rif_ipip_lb_op(lb_rif, 0, ul_rif->rif_index, true);
	if (err)
		goto err_loopback_op;

	lb_rif->ul_vr_id = 0;
	lb_rif->ul_rif_id = ul_rif->rif_index;

7880
	return 0;
7881 7882 7883 7884

err_loopback_op:
	mlxsw_sp_ul_rif_put(ul_rif);
	return err;
7885 7886 7887 7888
}

static void mlxsw_sp2_rif_ipip_lb_deconfigure(struct mlxsw_sp_rif *rif)
{
7889 7890 7891 7892 7893 7894 7895
	struct mlxsw_sp_rif_ipip_lb *lb_rif = mlxsw_sp_rif_ipip_lb_rif(rif);
	struct mlxsw_sp *mlxsw_sp = rif->mlxsw_sp;
	struct mlxsw_sp_rif *ul_rif;

	ul_rif = mlxsw_sp_rif_by_index(mlxsw_sp, lb_rif->ul_rif_id);
	mlxsw_sp_rif_ipip_lb_op(lb_rif, 0, lb_rif->ul_rif_id, false);
	mlxsw_sp_ul_rif_put(ul_rif);
7896 7897 7898 7899 7900 7901 7902 7903 7904 7905 7906 7907 7908 7909 7910
}

static const struct mlxsw_sp_rif_ops mlxsw_sp2_rif_ipip_lb_ops = {
	.type			= MLXSW_SP_RIF_TYPE_IPIP_LB,
	.rif_size		= sizeof(struct mlxsw_sp_rif_ipip_lb),
	.setup                  = mlxsw_sp_rif_ipip_lb_setup,
	.configure		= mlxsw_sp2_rif_ipip_lb_configure,
	.deconfigure		= mlxsw_sp2_rif_ipip_lb_deconfigure,
};

const struct mlxsw_sp_rif_ops *mlxsw_sp2_rif_ops_arr[] = {
	[MLXSW_SP_RIF_TYPE_SUBPORT]	= &mlxsw_sp_rif_subport_ops,
	[MLXSW_SP_RIF_TYPE_VLAN]	= &mlxsw_sp_rif_vlan_emu_ops,
	[MLXSW_SP_RIF_TYPE_FID]		= &mlxsw_sp_rif_fid_ops,
	[MLXSW_SP_RIF_TYPE_IPIP_LB]	= &mlxsw_sp2_rif_ipip_lb_ops,
7911 7912
};

7913 7914 7915 7916 7917 7918 7919 7920 7921
static int mlxsw_sp_rifs_init(struct mlxsw_sp *mlxsw_sp)
{
	u64 max_rifs = MLXSW_CORE_RES_GET(mlxsw_sp->core, MAX_RIFS);

	mlxsw_sp->router->rifs = kcalloc(max_rifs,
					 sizeof(struct mlxsw_sp_rif *),
					 GFP_KERNEL);
	if (!mlxsw_sp->router->rifs)
		return -ENOMEM;
7922

7923 7924 7925 7926 7927 7928 7929 7930 7931 7932 7933 7934 7935
	return 0;
}

static void mlxsw_sp_rifs_fini(struct mlxsw_sp *mlxsw_sp)
{
	int i;

	for (i = 0; i < MLXSW_CORE_RES_GET(mlxsw_sp->core, MAX_RIFS); i++)
		WARN_ON_ONCE(mlxsw_sp->router->rifs[i]);

	kfree(mlxsw_sp->router->rifs);
}

7936 7937 7938 7939 7940 7941 7942 7943 7944
static int
mlxsw_sp_ipip_config_tigcr(struct mlxsw_sp *mlxsw_sp)
{
	char tigcr_pl[MLXSW_REG_TIGCR_LEN];

	mlxsw_reg_tigcr_pack(tigcr_pl, true, 0);
	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(tigcr), tigcr_pl);
}

7945 7946
static int mlxsw_sp_ipips_init(struct mlxsw_sp *mlxsw_sp)
{
7947 7948
	int err;

7949
	mlxsw_sp->router->ipip_ops_arr = mlxsw_sp_ipip_ops_arr;
7950
	INIT_LIST_HEAD(&mlxsw_sp->router->ipip_list);
7951 7952 7953 7954 7955 7956 7957 7958

	err = mlxsw_sp_ipip_ecn_encap_init(mlxsw_sp);
	if (err)
		return err;
	err = mlxsw_sp_ipip_ecn_decap_init(mlxsw_sp);
	if (err)
		return err;

7959
	return mlxsw_sp_ipip_config_tigcr(mlxsw_sp);
7960 7961 7962 7963
}

static void mlxsw_sp_ipips_fini(struct mlxsw_sp *mlxsw_sp)
{
7964
	WARN_ON(!list_empty(&mlxsw_sp->router->ipip_list));
7965 7966
}

7967 7968
static void mlxsw_sp_router_fib_dump_flush(struct notifier_block *nb)
{
7969
	struct mlxsw_sp_router *router;
7970 7971 7972 7973 7974 7975

	/* Flush pending FIB notifications and then flush the device's
	 * table before requesting another dump. The FIB notification
	 * block is unregistered, so no need to take RTNL.
	 */
	mlxsw_core_flush_owq();
7976 7977
	router = container_of(nb, struct mlxsw_sp_router, fib_nb);
	mlxsw_sp_router_fib_flush(router->mlxsw_sp);
7978 7979
}

7980 7981 7982 7983 7984 7985 7986 7987 7988 7989 7990
#ifdef CONFIG_IP_ROUTE_MULTIPATH
static void mlxsw_sp_mp_hash_header_set(char *recr2_pl, int header)
{
	mlxsw_reg_recr2_outer_header_enables_set(recr2_pl, header, true);
}

static void mlxsw_sp_mp_hash_field_set(char *recr2_pl, int field)
{
	mlxsw_reg_recr2_outer_header_fields_enable_set(recr2_pl, field, true);
}

7991
static void mlxsw_sp_mp4_hash_init(struct mlxsw_sp *mlxsw_sp, char *recr2_pl)
7992
{
7993 7994
	struct net *net = mlxsw_sp_net(mlxsw_sp);
	bool only_l3 = !net->ipv4.sysctl_fib_multipath_hash_policy;
7995 7996 7997 7998 7999 8000 8001 8002 8003 8004 8005 8006 8007 8008

	mlxsw_sp_mp_hash_header_set(recr2_pl,
				    MLXSW_REG_RECR2_IPV4_EN_NOT_TCP_NOT_UDP);
	mlxsw_sp_mp_hash_header_set(recr2_pl, MLXSW_REG_RECR2_IPV4_EN_TCP_UDP);
	mlxsw_reg_recr2_ipv4_sip_enable(recr2_pl);
	mlxsw_reg_recr2_ipv4_dip_enable(recr2_pl);
	if (only_l3)
		return;
	mlxsw_sp_mp_hash_header_set(recr2_pl, MLXSW_REG_RECR2_TCP_UDP_EN_IPV4);
	mlxsw_sp_mp_hash_field_set(recr2_pl, MLXSW_REG_RECR2_IPV4_PROTOCOL);
	mlxsw_sp_mp_hash_field_set(recr2_pl, MLXSW_REG_RECR2_TCP_UDP_SPORT);
	mlxsw_sp_mp_hash_field_set(recr2_pl, MLXSW_REG_RECR2_TCP_UDP_DPORT);
}

8009
static void mlxsw_sp_mp6_hash_init(struct mlxsw_sp *mlxsw_sp, char *recr2_pl)
8010
{
8011
	bool only_l3 = !ip6_multipath_hash_policy(mlxsw_sp_net(mlxsw_sp));
8012

8013 8014 8015 8016 8017 8018
	mlxsw_sp_mp_hash_header_set(recr2_pl,
				    MLXSW_REG_RECR2_IPV6_EN_NOT_TCP_NOT_UDP);
	mlxsw_sp_mp_hash_header_set(recr2_pl, MLXSW_REG_RECR2_IPV6_EN_TCP_UDP);
	mlxsw_reg_recr2_ipv6_sip_enable(recr2_pl);
	mlxsw_reg_recr2_ipv6_dip_enable(recr2_pl);
	mlxsw_sp_mp_hash_field_set(recr2_pl, MLXSW_REG_RECR2_IPV6_NEXT_HEADER);
8019 8020 8021 8022 8023 8024 8025 8026 8027 8028 8029
	if (only_l3) {
		mlxsw_sp_mp_hash_field_set(recr2_pl,
					   MLXSW_REG_RECR2_IPV6_FLOW_LABEL);
	} else {
		mlxsw_sp_mp_hash_header_set(recr2_pl,
					    MLXSW_REG_RECR2_TCP_UDP_EN_IPV6);
		mlxsw_sp_mp_hash_field_set(recr2_pl,
					   MLXSW_REG_RECR2_TCP_UDP_SPORT);
		mlxsw_sp_mp_hash_field_set(recr2_pl,
					   MLXSW_REG_RECR2_TCP_UDP_DPORT);
	}
8030 8031 8032 8033 8034 8035 8036
}

static int mlxsw_sp_mp_hash_init(struct mlxsw_sp *mlxsw_sp)
{
	char recr2_pl[MLXSW_REG_RECR2_LEN];
	u32 seed;

8037
	seed = jhash(mlxsw_sp->base_mac, sizeof(mlxsw_sp->base_mac), 0);
8038
	mlxsw_reg_recr2_pack(recr2_pl, seed);
8039 8040
	mlxsw_sp_mp4_hash_init(mlxsw_sp, recr2_pl);
	mlxsw_sp_mp6_hash_init(mlxsw_sp, recr2_pl);
8041 8042 8043 8044 8045 8046 8047 8048 8049 8050

	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(recr2), recr2_pl);
}
#else
static int mlxsw_sp_mp_hash_init(struct mlxsw_sp *mlxsw_sp)
{
	return 0;
}
#endif

8051 8052 8053 8054 8055 8056 8057 8058 8059 8060 8061 8062 8063 8064 8065 8066 8067 8068
static int mlxsw_sp_dscp_init(struct mlxsw_sp *mlxsw_sp)
{
	char rdpm_pl[MLXSW_REG_RDPM_LEN];
	unsigned int i;

	MLXSW_REG_ZERO(rdpm, rdpm_pl);

	/* HW is determining switch priority based on DSCP-bits, but the
	 * kernel is still doing that based on the ToS. Since there's a
	 * mismatch in bits we need to make sure to translate the right
	 * value ToS would observe, skipping the 2 least-significant ECN bits.
	 */
	for (i = 0; i < MLXSW_REG_RDPM_DSCP_ENTRY_REC_MAX_COUNT; i++)
		mlxsw_reg_rdpm_pack(rdpm_pl, i, rt_tos2priority(i << 2));

	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(rdpm), rdpm_pl);
}

8069 8070
static int __mlxsw_sp_router_init(struct mlxsw_sp *mlxsw_sp)
{
8071 8072
	struct net *net = mlxsw_sp_net(mlxsw_sp);
	bool usp = net->ipv4.sysctl_ip_fwd_update_priority;
8073 8074 8075 8076 8077 8078 8079 8080
	char rgcr_pl[MLXSW_REG_RGCR_LEN];
	u64 max_rifs;
	int err;

	if (!MLXSW_CORE_RES_VALID(mlxsw_sp->core, MAX_RIFS))
		return -EIO;
	max_rifs = MLXSW_CORE_RES_GET(mlxsw_sp->core, MAX_RIFS);

8081
	mlxsw_reg_rgcr_pack(rgcr_pl, true, true);
8082
	mlxsw_reg_rgcr_max_router_interfaces_set(rgcr_pl, max_rifs);
8083
	mlxsw_reg_rgcr_usp_set(rgcr_pl, usp);
8084 8085
	err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(rgcr), rgcr_pl);
	if (err)
8086
		return err;
8087 8088 8089 8090 8091 8092 8093
	return 0;
}

static void __mlxsw_sp_router_fini(struct mlxsw_sp *mlxsw_sp)
{
	char rgcr_pl[MLXSW_REG_RGCR_LEN];

8094
	mlxsw_reg_rgcr_pack(rgcr_pl, false, false);
8095 8096 8097
	mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(rgcr), rgcr_pl);
}

8098 8099
int mlxsw_sp_router_init(struct mlxsw_sp *mlxsw_sp,
			 struct netlink_ext_ack *extack)
8100
{
8101
	struct mlxsw_sp_router *router;
8102 8103
	int err;

8104 8105 8106
	router = kzalloc(sizeof(*mlxsw_sp->router), GFP_KERNEL);
	if (!router)
		return -ENOMEM;
8107
	mutex_init(&router->lock);
8108 8109 8110
	mlxsw_sp->router = router;
	router->mlxsw_sp = mlxsw_sp;

8111 8112 8113 8114 8115 8116 8117 8118 8119 8120
	router->inetaddr_nb.notifier_call = mlxsw_sp_inetaddr_event;
	err = register_inetaddr_notifier(&router->inetaddr_nb);
	if (err)
		goto err_register_inetaddr_notifier;

	router->inet6addr_nb.notifier_call = mlxsw_sp_inet6addr_event;
	err = register_inet6addr_notifier(&router->inet6addr_nb);
	if (err)
		goto err_register_inet6addr_notifier;

8121
	INIT_LIST_HEAD(&mlxsw_sp->router->nexthop_neighs_list);
8122 8123
	err = __mlxsw_sp_router_init(mlxsw_sp);
	if (err)
8124
		goto err_router_init;
8125

8126 8127 8128 8129
	err = mlxsw_sp_rifs_init(mlxsw_sp);
	if (err)
		goto err_rifs_init;

8130 8131 8132 8133
	err = mlxsw_sp_ipips_init(mlxsw_sp);
	if (err)
		goto err_ipips_init;

8134
	err = rhashtable_init(&mlxsw_sp->router->nexthop_ht,
8135 8136 8137 8138
			      &mlxsw_sp_nexthop_ht_params);
	if (err)
		goto err_nexthop_ht_init;

8139
	err = rhashtable_init(&mlxsw_sp->router->nexthop_group_ht,
8140 8141 8142 8143
			      &mlxsw_sp_nexthop_group_ht_params);
	if (err)
		goto err_nexthop_group_ht_init;

8144
	INIT_LIST_HEAD(&mlxsw_sp->router->nexthop_list);
8145 8146 8147 8148
	err = mlxsw_sp_lpm_init(mlxsw_sp);
	if (err)
		goto err_lpm_init;

8149 8150 8151 8152
	err = mlxsw_sp_mr_init(mlxsw_sp, &mlxsw_sp_mr_tcam_ops);
	if (err)
		goto err_mr_init;

8153 8154 8155 8156
	err = mlxsw_sp_vrs_init(mlxsw_sp);
	if (err)
		goto err_vrs_init;

8157
	err = mlxsw_sp_neigh_init(mlxsw_sp);
8158 8159 8160
	if (err)
		goto err_neigh_init;

8161 8162 8163 8164 8165 8166
	mlxsw_sp->router->netevent_nb.notifier_call =
		mlxsw_sp_router_netevent_event;
	err = register_netevent_notifier(&mlxsw_sp->router->netevent_nb);
	if (err)
		goto err_register_netevent_notifier;

8167 8168 8169 8170
	err = mlxsw_sp_mp_hash_init(mlxsw_sp);
	if (err)
		goto err_mp_hash_init;

8171 8172 8173 8174
	err = mlxsw_sp_dscp_init(mlxsw_sp);
	if (err)
		goto err_dscp_init;

8175
	mlxsw_sp->router->fib_nb.notifier_call = mlxsw_sp_router_fib_event;
8176 8177
	err = register_fib_notifier(mlxsw_sp_net(mlxsw_sp),
				    &mlxsw_sp->router->fib_nb,
8178
				    mlxsw_sp_router_fib_dump_flush, extack);
8179 8180 8181
	if (err)
		goto err_register_fib_notifier;

8182 8183
	return 0;

8184
err_register_fib_notifier:
8185
err_dscp_init:
8186
err_mp_hash_init:
8187 8188
	unregister_netevent_notifier(&mlxsw_sp->router->netevent_nb);
err_register_netevent_notifier:
8189
	mlxsw_sp_neigh_fini(mlxsw_sp);
8190 8191 8192
err_neigh_init:
	mlxsw_sp_vrs_fini(mlxsw_sp);
err_vrs_init:
8193 8194
	mlxsw_sp_mr_fini(mlxsw_sp);
err_mr_init:
8195 8196
	mlxsw_sp_lpm_fini(mlxsw_sp);
err_lpm_init:
8197
	rhashtable_destroy(&mlxsw_sp->router->nexthop_group_ht);
8198
err_nexthop_group_ht_init:
8199
	rhashtable_destroy(&mlxsw_sp->router->nexthop_ht);
8200
err_nexthop_ht_init:
8201 8202
	mlxsw_sp_ipips_fini(mlxsw_sp);
err_ipips_init:
8203 8204
	mlxsw_sp_rifs_fini(mlxsw_sp);
err_rifs_init:
8205
	__mlxsw_sp_router_fini(mlxsw_sp);
8206
err_router_init:
8207 8208 8209 8210
	unregister_inet6addr_notifier(&router->inet6addr_nb);
err_register_inet6addr_notifier:
	unregister_inetaddr_notifier(&router->inetaddr_nb);
err_register_inetaddr_notifier:
8211
	mutex_destroy(&mlxsw_sp->router->lock);
8212
	kfree(mlxsw_sp->router);
8213 8214 8215 8216 8217
	return err;
}

void mlxsw_sp_router_fini(struct mlxsw_sp *mlxsw_sp)
{
8218 8219
	unregister_fib_notifier(mlxsw_sp_net(mlxsw_sp),
				&mlxsw_sp->router->fib_nb);
8220
	unregister_netevent_notifier(&mlxsw_sp->router->netevent_nb);
8221 8222
	mlxsw_sp_neigh_fini(mlxsw_sp);
	mlxsw_sp_vrs_fini(mlxsw_sp);
8223
	mlxsw_sp_mr_fini(mlxsw_sp);
8224
	mlxsw_sp_lpm_fini(mlxsw_sp);
8225 8226
	rhashtable_destroy(&mlxsw_sp->router->nexthop_group_ht);
	rhashtable_destroy(&mlxsw_sp->router->nexthop_ht);
8227
	mlxsw_sp_ipips_fini(mlxsw_sp);
8228
	mlxsw_sp_rifs_fini(mlxsw_sp);
8229
	__mlxsw_sp_router_fini(mlxsw_sp);
8230 8231
	unregister_inet6addr_notifier(&mlxsw_sp->router->inet6addr_nb);
	unregister_inetaddr_notifier(&mlxsw_sp->router->inetaddr_nb);
8232
	mutex_destroy(&mlxsw_sp->router->lock);
8233
	kfree(mlxsw_sp->router);
8234
}