[PATCH net-next 00/15] mlxsw: Preparations for nexthop objects support - part 1/2

[PATCH net-next 00/15] mlxsw: Preparations for nexthop objects support - part 1/2

[Ido Schimmel]

From: Ido Schimmel <idosch@nvidia.com>

This patch set contains small and non-functional changes aimed at making
it easier to support nexthop objects in mlxsw. Follow up patches can be
found here [1].

Patches #1-#4 add a type field to the nexthop group struct instead of
the existing protocol field. This will be used later on to add a nexthop
object type, which can contain both IPv4 and IPv6 nexthops.

Patches #5-#7 move the IPv4 FIB info pointer (i.e., 'struct fib_info')
from the nexthop group struct to the route. The pointer will not be
available when the nexthop group is a nexthop object, but it needs to be
accessible to routes regardless.

Patch #8 is the biggest change, but it is an entirely cosmetic change
and should therefore be easy to review. The motivation and the change
itself are explained in detail in the commit message.

Patches #9-#12 perform small changes so that two functions that are
currently split between IPv4 and IPv6 could be consolidated in patches
#13 and #14. The functions will be reused for nexthop objects.

Patch #15 removes an outdated comment.

[1] https://github.com/idosch/linux/tree/submit/nexthop_objects

Ido Schimmel (15):
  mlxsw: spectrum_router: Compare key with correct object type
  mlxsw: spectrum_router: Add nexthop group type field
  mlxsw: spectrum_router: Use nexthop group type in hash table key
  mlxsw: spectrum_router: Associate neighbour table with nexthop instead
    of group
  mlxsw: spectrum_router: Store FIB info in route
  mlxsw: spectrum_router: Remove unused field 'prio' from IPv4 FIB entry
    struct
  mlxsw: spectrum_router: Move IPv4 FIB info into a union in nexthop
    group struct
  mlxsw: spectrum_router: Split nexthop group configuration to a
    different struct
  mlxsw: spectrum_ipip: Remove overlay protocol from can_offload()
    callback
  mlxsw: spectrum_router: Pass nexthop netdev to
    mlxsw_sp_nexthop6_type_init()
  mlxsw: spectrum_router: Pass nexthop netdev to
    mlxsw_sp_nexthop4_type_init()
  mlxsw: spectrum_router: Remove unused argument from
    mlxsw_sp_nexthop6_type_init()
  mlxsw: spectrum_router: Consolidate mlxsw_sp_nexthop{4, 6}_type_init()
  mlxsw: spectrum_router: Consolidate mlxsw_sp_nexthop{4, 6}_type_fini()
  mlxsw: spectrum_router: Remove outdated comment

 .../ethernet/mellanox/mlxsw/spectrum_ipip.c   |   3 +-
 .../ethernet/mellanox/mlxsw/spectrum_ipip.h   |   3 +-
 .../ethernet/mellanox/mlxsw/spectrum_router.c | 580 +++++++++---------
 3 files changed, 300 insertions(+), 286 deletions(-)

-- 
2.28.0

[PATCH net-next 01/15] mlxsw: spectrum_router: Compare key with correct object type

[Ido Schimmel]

From: Ido Schimmel <idosch@nvidia.com>

When comparing a key with a nexthop group in rhastable's obj_cmpfn()
callback, make sure that the key and nexthop group are of the same type
(i.e., IPv4 / IPv6).

The bug is not currently visible because IPv6 nexthop groups do not
populate the FIB info pointer and IPv4 nexthop groups do not set the
ifindex for the individual nexthops.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
---
 .../ethernet/mellanox/mlxsw/spectrum_router.c    | 16 ++++++++++------
 1 file changed, 10 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index e692e5a39f6c..55ae16b03c44 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -3040,6 +3040,12 @@ mlxsw_sp_nexthop6_group_cmp(const struct mlxsw_sp_nexthop_group *nh_grp,
 	return true;
 }
 
+static int
+mlxsw_sp_nexthop_group_type(const struct mlxsw_sp_nexthop_group *nh_grp)
+{
+	return nh_grp->neigh_tbl->family;
+}
+
 static int
 mlxsw_sp_nexthop_group_cmp(struct rhashtable_compare_arg *arg, const void *ptr)
 {
@@ -3048,8 +3054,12 @@ mlxsw_sp_nexthop_group_cmp(struct rhashtable_compare_arg *arg, const void *ptr)
 
 	switch (cmp_arg->proto) {
 	case MLXSW_SP_L3_PROTO_IPV4:
+		if (mlxsw_sp_nexthop_group_type(nh_grp) != AF_INET)
+			return 1;
 		return cmp_arg->fi != mlxsw_sp_nexthop4_group_fi(nh_grp);
 	case MLXSW_SP_L3_PROTO_IPV6:
+		if (mlxsw_sp_nexthop_group_type(nh_grp) != AF_INET6)
+			return 1;
 		return !mlxsw_sp_nexthop6_group_cmp(nh_grp,
 						    cmp_arg->fib6_entry);
 	default:
@@ -3058,12 +3068,6 @@ mlxsw_sp_nexthop_group_cmp(struct rhashtable_compare_arg *arg, const void *ptr)
 	}
 }
 
-static int
-mlxsw_sp_nexthop_group_type(const struct mlxsw_sp_nexthop_group *nh_grp)
-{
-	return nh_grp->neigh_tbl->family;
-}
-
 static u32 mlxsw_sp_nexthop_group_hash_obj(const void *data, u32 len, u32 seed)
 {
 	const struct mlxsw_sp_nexthop_group *nh_grp = data;
-- 
2.28.0

[PATCH net-next 02/15] mlxsw: spectrum_router: Add nexthop group type field

[Ido Schimmel]

From: Ido Schimmel <idosch@nvidia.com>

Currently, the type (i.e., IPv4/IPv6) of the nexthop group is derived
from the neighbour table associated with the group.

This is problematic when nexthop objects are taken into account, as a
nexthop group object can contain both IPv4 and IPv6 nexthops.

Instead, add a new field that indicates the type of the group and
initialize it during the group's creation. Currently, the types are IPv4
('struct fib_info') and IPv6 ('struct fib6_info'). In the future another
type will be added for nexthop objects ('struct nexthop').

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
---
 .../ethernet/mellanox/mlxsw/spectrum_router.c | 34 ++++++++++---------
 1 file changed, 18 insertions(+), 16 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 55ae16b03c44..5b8363206325 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -2860,11 +2860,17 @@ struct mlxsw_sp_nexthop {
 	bool counter_valid;
 };
 
+enum mlxsw_sp_nexthop_group_type {
+	MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV4,
+	MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV6,
+};
+
 struct mlxsw_sp_nexthop_group {
 	void *priv;
 	struct rhash_head ht_node;
 	struct list_head fib_list; /* list of fib entries that use this group */
 	struct neigh_table *neigh_tbl;
+	enum mlxsw_sp_nexthop_group_type type;
 	u8 adj_index_valid:1,
 	   gateway:1; /* routes using the group use a gateway */
 	u32 adj_index;
@@ -3040,12 +3046,6 @@ mlxsw_sp_nexthop6_group_cmp(const struct mlxsw_sp_nexthop_group *nh_grp,
 	return true;
 }
 
-static int
-mlxsw_sp_nexthop_group_type(const struct mlxsw_sp_nexthop_group *nh_grp)
-{
-	return nh_grp->neigh_tbl->family;
-}
-
 static int
 mlxsw_sp_nexthop_group_cmp(struct rhashtable_compare_arg *arg, const void *ptr)
 {
@@ -3054,11 +3054,11 @@ mlxsw_sp_nexthop_group_cmp(struct rhashtable_compare_arg *arg, const void *ptr)
 
 	switch (cmp_arg->proto) {
 	case MLXSW_SP_L3_PROTO_IPV4:
-		if (mlxsw_sp_nexthop_group_type(nh_grp) != AF_INET)
+		if (nh_grp->type != MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV4)
 			return 1;
 		return cmp_arg->fi != mlxsw_sp_nexthop4_group_fi(nh_grp);
 	case MLXSW_SP_L3_PROTO_IPV6:
-		if (mlxsw_sp_nexthop_group_type(nh_grp) != AF_INET6)
+		if (nh_grp->type != MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV6)
 			return 1;
 		return !mlxsw_sp_nexthop6_group_cmp(nh_grp,
 						    cmp_arg->fib6_entry);
@@ -3076,11 +3076,11 @@ static u32 mlxsw_sp_nexthop_group_hash_obj(const void *data, u32 len, u32 seed)
 	unsigned int val;
 	int i;
 
-	switch (mlxsw_sp_nexthop_group_type(nh_grp)) {
-	case AF_INET:
+	switch (nh_grp->type) {
+	case MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV4:
 		fi = mlxsw_sp_nexthop4_group_fi(nh_grp);
 		return jhash(&fi, sizeof(fi), seed);
-	case AF_INET6:
+	case MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV6:
 		val = nh_grp->count;
 		for (i = 0; i < nh_grp->count; i++) {
 			nh = &nh_grp->nexthops[i];
@@ -3138,7 +3138,7 @@ static const struct rhashtable_params mlxsw_sp_nexthop_group_ht_params = {
 static int mlxsw_sp_nexthop_group_insert(struct mlxsw_sp *mlxsw_sp,
 					 struct mlxsw_sp_nexthop_group *nh_grp)
 {
-	if (mlxsw_sp_nexthop_group_type(nh_grp) == AF_INET6 &&
+	if (nh_grp->type == MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV6 &&
 	    !nh_grp->gateway)
 		return 0;
 
@@ -3150,7 +3150,7 @@ static int mlxsw_sp_nexthop_group_insert(struct mlxsw_sp *mlxsw_sp,
 static void mlxsw_sp_nexthop_group_remove(struct mlxsw_sp *mlxsw_sp,
 					  struct mlxsw_sp_nexthop_group *nh_grp)
 {
-	if (mlxsw_sp_nexthop_group_type(nh_grp) == AF_INET6 &&
+	if (nh_grp->type == MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV6 &&
 	    !nh_grp->gateway)
 		return;
 
@@ -3528,11 +3528,11 @@ 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:
+	switch (nh_grp->type) {
+	case MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV4:
 		mlxsw_sp_nexthop4_group_offload_refresh(mlxsw_sp, nh_grp);
 		break;
-	case AF_INET6:
+	case MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV6:
 		mlxsw_sp_nexthop6_group_offload_refresh(mlxsw_sp, nh_grp);
 		break;
 	}
@@ -4106,6 +4106,7 @@ mlxsw_sp_nexthop4_group_create(struct mlxsw_sp *mlxsw_sp, struct fib_info *fi)
 	nh_grp->priv = fi;
 	INIT_LIST_HEAD(&nh_grp->fib_list);
 	nh_grp->neigh_tbl = &arp_tbl;
+	nh_grp->type = MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV4;
 
 	nh_grp->gateway = mlxsw_sp_fi_is_gateway(mlxsw_sp, fi);
 	nh_grp->count = nhs;
@@ -5417,6 +5418,7 @@ mlxsw_sp_nexthop6_group_create(struct mlxsw_sp *mlxsw_sp,
 #if IS_ENABLED(CONFIG_IPV6)
 	nh_grp->neigh_tbl = &nd_tbl;
 #endif
+	nh_grp->type = MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV6;
 	mlxsw_sp_rt6 = list_first_entry(&fib6_entry->rt6_list,
 					struct mlxsw_sp_rt6, list);
 	nh_grp->gateway = mlxsw_sp_rt6_is_gateway(mlxsw_sp, mlxsw_sp_rt6->rt);
-- 
2.28.0

[PATCH net-next 03/15] mlxsw: spectrum_router: Use nexthop group type in hash table key

[Ido Schimmel]

From: Ido Schimmel <idosch@nvidia.com>

Both IPv4 and IPv6 nexthop groups are hashed in the same table. The
protocol field is used to indicate how the hash should be computed for
each group.

When nexthop group objects are supported, the hash will be computed for
them based on the nexthop identifier.

To differentiate between all the nexthop group types, encode the type of
the group in the key instead of the protocol.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
---
 .../ethernet/mellanox/mlxsw/spectrum_router.c | 25 +++++++++----------
 1 file changed, 12 insertions(+), 13 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 5b8363206325..5affe7f79a9a 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -2995,7 +2995,7 @@ mlxsw_sp_nexthop4_group_fi(const struct mlxsw_sp_nexthop_group *nh_grp)
 }
 
 struct mlxsw_sp_nexthop_group_cmp_arg {
-	enum mlxsw_sp_l3proto proto;
+	enum mlxsw_sp_nexthop_group_type type;
 	union {
 		struct fib_info *fi;
 		struct mlxsw_sp_fib6_entry *fib6_entry;
@@ -3052,14 +3052,13 @@ 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;
 
-	switch (cmp_arg->proto) {
-	case MLXSW_SP_L3_PROTO_IPV4:
-		if (nh_grp->type != MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV4)
-			return 1;
+	if (nh_grp->type != cmp_arg->type)
+		return 1;
+
+	switch (cmp_arg->type) {
+	case MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV4:
 		return cmp_arg->fi != mlxsw_sp_nexthop4_group_fi(nh_grp);
-	case MLXSW_SP_L3_PROTO_IPV6:
-		if (nh_grp->type != MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV6)
-			return 1;
+	case MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV6:
 		return !mlxsw_sp_nexthop6_group_cmp(nh_grp,
 						    cmp_arg->fib6_entry);
 	default:
@@ -3117,10 +3116,10 @@ mlxsw_sp_nexthop_group_hash(const void *data, u32 len, u32 seed)
 {
 	const struct mlxsw_sp_nexthop_group_cmp_arg *cmp_arg = data;
 
-	switch (cmp_arg->proto) {
-	case MLXSW_SP_L3_PROTO_IPV4:
+	switch (cmp_arg->type) {
+	case MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV4:
 		return jhash(&cmp_arg->fi, sizeof(cmp_arg->fi), seed);
-	case MLXSW_SP_L3_PROTO_IPV6:
+	case MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV6:
 		return mlxsw_sp_nexthop6_group_hash(cmp_arg->fib6_entry, seed);
 	default:
 		WARN_ON(1);
@@ -3165,7 +3164,7 @@ mlxsw_sp_nexthop4_group_lookup(struct mlxsw_sp *mlxsw_sp,
 {
 	struct mlxsw_sp_nexthop_group_cmp_arg cmp_arg;
 
-	cmp_arg.proto = MLXSW_SP_L3_PROTO_IPV4;
+	cmp_arg.type = MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV4;
 	cmp_arg.fi = fi;
 	return rhashtable_lookup_fast(&mlxsw_sp->router->nexthop_group_ht,
 				      &cmp_arg,
@@ -3178,7 +3177,7 @@ mlxsw_sp_nexthop6_group_lookup(struct mlxsw_sp *mlxsw_sp,
 {
 	struct mlxsw_sp_nexthop_group_cmp_arg cmp_arg;
 
-	cmp_arg.proto = MLXSW_SP_L3_PROTO_IPV6;
+	cmp_arg.type = MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV6;
 	cmp_arg.fib6_entry = fib6_entry;
 	return rhashtable_lookup_fast(&mlxsw_sp->router->nexthop_group_ht,
 				      &cmp_arg,
-- 
2.28.0

[PATCH net-next 04/15] mlxsw: spectrum_router: Associate neighbour table with nexthop instead of group

[Ido Schimmel]

From: Ido Schimmel <idosch@nvidia.com>

As explained in the previous patch, nexthop objects can have both IPv4
and IPv6 nexthops in the same group. Therefore, move the neighbour table
to be a property of the nexthop instead of the nexthop group.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
---
 .../ethernet/mellanox/mlxsw/spectrum_router.c | 20 +++++++++----------
 1 file changed, 9 insertions(+), 11 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 5affe7f79a9a..462ddab11c07 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -2835,6 +2835,7 @@ struct mlxsw_sp_nexthop {
 						* this belongs to
 						*/
 	struct rhash_head ht_node;
+	struct neigh_table *neigh_tbl;
 	struct mlxsw_sp_nexthop_key key;
 	unsigned char gw_addr[sizeof(struct in6_addr)];
 	int ifindex;
@@ -2869,7 +2870,6 @@ struct mlxsw_sp_nexthop_group {
 	void *priv;
 	struct rhash_head ht_node;
 	struct list_head fib_list; /* list of fib entries that use this group */
-	struct neigh_table *neigh_tbl;
 	enum mlxsw_sp_nexthop_group_type type;
 	u8 adj_index_valid:1,
 	   gateway:1; /* routes using the group use a gateway */
@@ -3674,10 +3674,9 @@ mlxsw_sp_nexthop_dead_neigh_replace(struct mlxsw_sp *mlxsw_sp,
 	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);
+	n = neigh_lookup(nh->neigh_tbl, &nh->gw_addr, nh->rif->dev);
 	if (!n) {
-		n = neigh_create(nh->nh_grp->neigh_tbl, &nh->gw_addr,
-				 nh->rif->dev);
+		n = neigh_create(nh->neigh_tbl, &nh->gw_addr, nh->rif->dev);
 		if (IS_ERR(n))
 			return PTR_ERR(n);
 		neigh_event_send(n, NULL);
@@ -3776,10 +3775,9 @@ static int mlxsw_sp_nexthop_neigh_init(struct mlxsw_sp *mlxsw_sp,
 	 * The reference is taken either in neigh_lookup() or
 	 * in neigh_create() in case n is not found.
 	 */
-	n = neigh_lookup(nh->nh_grp->neigh_tbl, &nh->gw_addr, nh->rif->dev);
+	n = neigh_lookup(nh->neigh_tbl, &nh->gw_addr, nh->rif->dev);
 	if (!n) {
-		n = neigh_create(nh->nh_grp->neigh_tbl, &nh->gw_addr,
-				 nh->rif->dev);
+		n = neigh_create(nh->neigh_tbl, &nh->gw_addr, nh->rif->dev);
 		if (IS_ERR(n))
 			return PTR_ERR(n);
 		neigh_event_send(n, NULL);
@@ -3968,6 +3966,7 @@ static int mlxsw_sp_nexthop4_init(struct mlxsw_sp *mlxsw_sp,
 	nh->nh_weight = 1;
 #endif
 	memcpy(&nh->gw_addr, &fib_nh->fib_nh_gw4, sizeof(fib_nh->fib_nh_gw4));
+	nh->neigh_tbl = &arp_tbl;
 	err = mlxsw_sp_nexthop_insert(mlxsw_sp, nh);
 	if (err)
 		return err;
@@ -4104,7 +4103,6 @@ mlxsw_sp_nexthop4_group_create(struct mlxsw_sp *mlxsw_sp, struct fib_info *fi)
 		return ERR_PTR(-ENOMEM);
 	nh_grp->priv = fi;
 	INIT_LIST_HEAD(&nh_grp->fib_list);
-	nh_grp->neigh_tbl = &arp_tbl;
 	nh_grp->type = MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV4;
 
 	nh_grp->gateway = mlxsw_sp_fi_is_gateway(mlxsw_sp, fi);
@@ -5373,6 +5371,9 @@ static int mlxsw_sp_nexthop6_init(struct mlxsw_sp *mlxsw_sp,
 	nh->nh_grp = nh_grp;
 	nh->nh_weight = rt->fib6_nh->fib_nh_weight;
 	memcpy(&nh->gw_addr, &rt->fib6_nh->fib_nh_gw6, sizeof(nh->gw_addr));
+#if IS_ENABLED(CONFIG_IPV6)
+	nh->neigh_tbl = &nd_tbl;
+#endif
 	mlxsw_sp_nexthop_counter_alloc(mlxsw_sp, nh);
 
 	list_add_tail(&nh->router_list_node, &mlxsw_sp->router->nexthop_list);
@@ -5414,9 +5415,6 @@ mlxsw_sp_nexthop6_group_create(struct mlxsw_sp *mlxsw_sp,
 	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
 	nh_grp->type = MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV6;
 	mlxsw_sp_rt6 = list_first_entry(&fib6_entry->rt6_list,
 					struct mlxsw_sp_rt6, list);
-- 
2.28.0

[PATCH net-next 05/15] mlxsw: spectrum_router: Store FIB info in route

[Ido Schimmel]

From: Ido Schimmel <idosch@nvidia.com>

When needed, IPv4 routes fetch the FIB info (i.e., 'struct fib_info')
from their associated nexthop group. This will not work when the nexthop
group represents a nexthop object (i.e., 'struct nexthop'), as it will
only have access to the nexthop's identifier.

Instead, store the FIB info in the route itself.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
---
 .../net/ethernet/mellanox/mlxsw/spectrum_router.c   | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 462ddab11c07..87acb2bbb6fe 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -431,6 +431,7 @@ struct mlxsw_sp_fib_entry {
 
 struct mlxsw_sp_fib4_entry {
 	struct mlxsw_sp_fib_entry common;
+	struct fib_info *fi;
 	u32 tb_id;
 	u32 prio;
 	u8 tos;
@@ -4240,7 +4241,6 @@ static void
 mlxsw_sp_fib4_entry_hw_flags_set(struct mlxsw_sp *mlxsw_sp,
 				 struct mlxsw_sp_fib_entry *fib_entry)
 {
-	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;
@@ -4250,7 +4250,7 @@ mlxsw_sp_fib4_entry_hw_flags_set(struct mlxsw_sp *mlxsw_sp,
 	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.fi = fib4_entry->fi;
 	fri.tb_id = fib4_entry->tb_id;
 	fri.dst = cpu_to_be32(*p_dst);
 	fri.dst_len = dst_len;
@@ -4265,7 +4265,6 @@ static void
 mlxsw_sp_fib4_entry_hw_flags_clear(struct mlxsw_sp *mlxsw_sp,
 				   struct mlxsw_sp_fib_entry *fib_entry)
 {
-	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;
@@ -4273,7 +4272,7 @@ mlxsw_sp_fib4_entry_hw_flags_clear(struct mlxsw_sp *mlxsw_sp,
 
 	fib4_entry = container_of(fib_entry, struct mlxsw_sp_fib4_entry,
 				  common);
-	fri.fi = fi;
+	fri.fi = fib4_entry->fi;
 	fri.tb_id = fib4_entry->tb_id;
 	fri.dst = cpu_to_be32(*p_dst);
 	fri.dst_len = dst_len;
@@ -4831,6 +4830,8 @@ mlxsw_sp_fib4_entry_create(struct mlxsw_sp *mlxsw_sp,
 	if (err)
 		goto err_nexthop4_group_get;
 
+	fib4_entry->fi = fen_info->fi;
+	fib_info_hold(fib4_entry->fi);
 	fib4_entry->prio = fen_info->fi->fib_priority;
 	fib4_entry->tb_id = fen_info->tb_id;
 	fib4_entry->type = fen_info->type;
@@ -4852,6 +4853,7 @@ mlxsw_sp_fib4_entry_create(struct mlxsw_sp *mlxsw_sp,
 static void mlxsw_sp_fib4_entry_destroy(struct mlxsw_sp *mlxsw_sp,
 					struct mlxsw_sp_fib4_entry *fib4_entry)
 {
+	fib_info_put(fib4_entry->fi);
 	mlxsw_sp_nexthop4_group_put(mlxsw_sp, &fib4_entry->common);
 	mlxsw_sp_fib4_entry_type_unset(mlxsw_sp, &fib4_entry->common);
 	mlxsw_sp_fib_entry_priv_put(fib4_entry->common.priv);
@@ -4883,8 +4885,7 @@ mlxsw_sp_fib4_entry_lookup(struct mlxsw_sp *mlxsw_sp,
 	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)
+	    fib4_entry->fi == fen_info->fi)
 		return fib4_entry;
 
 	return NULL;
-- 
2.28.0

[PATCH net-next 06/15] mlxsw: spectrum_router: Remove unused field 'prio' from IPv4 FIB entry struct

[Ido Schimmel]

From: Ido Schimmel <idosch@nvidia.com>

Not used anywhere.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 87acb2bbb6fe..bd4bf9316390 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -433,7 +433,6 @@ struct mlxsw_sp_fib4_entry {
 	struct mlxsw_sp_fib_entry common;
 	struct fib_info *fi;
 	u32 tb_id;
-	u32 prio;
 	u8 tos;
 	u8 type;
 };
@@ -4832,7 +4831,6 @@ mlxsw_sp_fib4_entry_create(struct mlxsw_sp *mlxsw_sp,
 
 	fib4_entry->fi = fen_info->fi;
 	fib_info_hold(fib4_entry->fi);
-	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;
-- 
2.28.0

[PATCH net-next 07/15] mlxsw: spectrum_router: Move IPv4 FIB info into a union in nexthop group struct

[Ido Schimmel]

From: Ido Schimmel <idosch@nvidia.com>

Instead of storing the FIB info as 'priv' when the nexthop group
represents an IPv4 nexthop group, simply store it as a FIB info with a
proper comment.

When nexthop objects are supported, this field will become a union with
the nexthop object's identifier.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
---
 .../ethernet/mellanox/mlxsw/spectrum_router.c | 20 +++++++++----------
 1 file changed, 9 insertions(+), 11 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index bd4bf9316390..6612951b6911 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -2867,7 +2867,11 @@ enum mlxsw_sp_nexthop_group_type {
 };
 
 struct mlxsw_sp_nexthop_group {
-	void *priv;
+	union {
+		struct {
+			struct fib_info *fi;
+		} ipv4;
+	};
 	struct rhash_head ht_node;
 	struct list_head fib_list; /* list of fib entries that use this group */
 	enum mlxsw_sp_nexthop_group_type type;
@@ -2988,12 +2992,6 @@ bool mlxsw_sp_nexthop_group_has_ipip(struct mlxsw_sp_nexthop *nh)
 	return false;
 }
 
-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 {
 	enum mlxsw_sp_nexthop_group_type type;
 	union {
@@ -3057,7 +3055,7 @@ mlxsw_sp_nexthop_group_cmp(struct rhashtable_compare_arg *arg, const void *ptr)
 
 	switch (cmp_arg->type) {
 	case MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV4:
-		return cmp_arg->fi != mlxsw_sp_nexthop4_group_fi(nh_grp);
+		return cmp_arg->fi != nh_grp->ipv4.fi;
 	case MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV6:
 		return !mlxsw_sp_nexthop6_group_cmp(nh_grp,
 						    cmp_arg->fib6_entry);
@@ -3077,7 +3075,7 @@ static u32 mlxsw_sp_nexthop_group_hash_obj(const void *data, u32 len, u32 seed)
 
 	switch (nh_grp->type) {
 	case MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV4:
-		fi = mlxsw_sp_nexthop4_group_fi(nh_grp);
+		fi = nh_grp->ipv4.fi;
 		return jhash(&fi, sizeof(fi), seed);
 	case MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV6:
 		val = nh_grp->count;
@@ -4101,12 +4099,12 @@ mlxsw_sp_nexthop4_group_create(struct mlxsw_sp *mlxsw_sp, struct fib_info *fi)
 	nh_grp = kzalloc(struct_size(nh_grp, nexthops, nhs), GFP_KERNEL);
 	if (!nh_grp)
 		return ERR_PTR(-ENOMEM);
-	nh_grp->priv = fi;
 	INIT_LIST_HEAD(&nh_grp->fib_list);
 	nh_grp->type = MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV4;
 
 	nh_grp->gateway = mlxsw_sp_fi_is_gateway(mlxsw_sp, fi);
 	nh_grp->count = nhs;
+	nh_grp->ipv4.fi = fi;
 	fib_info_hold(fi);
 	for (i = 0; i < nh_grp->count; i++) {
 		nh = &nh_grp->nexthops[i];
@@ -4146,7 +4144,7 @@ mlxsw_sp_nexthop4_group_destroy(struct mlxsw_sp *mlxsw_sp,
 	}
 	mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh_grp);
 	WARN_ON_ONCE(nh_grp->adj_index_valid);
-	fib_info_put(mlxsw_sp_nexthop4_group_fi(nh_grp));
+	fib_info_put(nh_grp->ipv4.fi);
 	kfree(nh_grp);
 }
 
-- 
2.28.0

[PATCH net-next 08/15] mlxsw: spectrum_router: Split nexthop group configuration to a different struct

[Ido Schimmel]

From: Ido Schimmel <idosch@nvidia.com>

Currently, the individual nexthops member in the group and attributes of
the group (e.g., its type) are stored in the same struct (i.e., 'struct
mlxsw_sp_nexthop_group'). This is fine since the individual nexthops
cannot change during the lifetime of the group.

With nexthop objects this is no longer the case. An existing nexthop
group can be replaced to use a new set of nexthops. Creating a new
struct whenever a group is replaced entails replacing the group pointer
of all the routes (i.e., 'struct mlxsw_sp_fib_entry') using the group.

Avoid this inefficient step by splitting the nexthop group configuration
to a different struct (i.e., 'struct mlxsw_sp_nexthop_group_info').
When a nexthop group is replaced a new group info struct is created and
the individual rotues do not need to be touched.

Illustration after the change:

  mlxsw_sp_fib_entry    mlxsw_sp_nexthop_group    mlxsw_sp_nexthop_group_info
+-------------------+  +----------------------+  +---------------------------+
| nh_group;         +--> nhgi;                +-->                           |
|                   |  |                      |  |                           |
+-------------------+  +----------------------+  +---------------------------+

No functional changes intended.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
---
 .../ethernet/mellanox/mlxsw/spectrum_router.c | 377 +++++++++++-------
 1 file changed, 228 insertions(+), 149 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 6612951b6911..3079be4bc5ec 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -352,6 +352,7 @@ enum mlxsw_sp_fib_entry_type {
 	MLXSW_SP_FIB_ENTRY_TYPE_NVE_DECAP,
 };
 
+struct mlxsw_sp_nexthop_group_info;
 struct mlxsw_sp_nexthop_group;
 struct mlxsw_sp_fib_entry;
 
@@ -2831,9 +2832,9 @@ struct mlxsw_sp_nexthop {
 	struct list_head neigh_list_node; /* member of neigh entry list */
 	struct list_head rif_list_node;
 	struct list_head router_list_node;
-	struct mlxsw_sp_nexthop_group *nh_grp; /* pointer back to the group
-						* this belongs to
-						*/
+	struct mlxsw_sp_nexthop_group_info *nhgi; /* pointer back to the group
+						   * this nexthop belongs to
+						   */
 	struct rhash_head ht_node;
 	struct neigh_table *neigh_tbl;
 	struct mlxsw_sp_nexthop_key key;
@@ -2866,23 +2867,28 @@ enum mlxsw_sp_nexthop_group_type {
 	MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV6,
 };
 
+struct mlxsw_sp_nexthop_group_info {
+	struct mlxsw_sp_nexthop_group *nh_grp;
+	u32 adj_index;
+	u16 ecmp_size;
+	u16 count;
+	int sum_norm_weight;
+	u8 adj_index_valid:1,
+	   gateway:1; /* routes using the group use a gateway */
+	struct mlxsw_sp_nexthop nexthops[0];
+#define nh_rif	nexthops[0].rif
+};
+
 struct mlxsw_sp_nexthop_group {
+	struct rhash_head ht_node;
+	struct list_head fib_list; /* list of fib entries that use this group */
 	union {
 		struct {
 			struct fib_info *fi;
 		} ipv4;
 	};
-	struct rhash_head ht_node;
-	struct list_head fib_list; /* list of fib entries that use this group */
+	struct mlxsw_sp_nexthop_group_info *nhgi;
 	enum mlxsw_sp_nexthop_group_type type;
-	u8 adj_index_valid:1,
-	   gateway:1; /* routes using the group use a gateway */
-	u32 adj_index;
-	u16 ecmp_size;
-	u16 count;
-	int sum_norm_weight;
-	struct mlxsw_sp_nexthop nexthops[0];
-#define nh_rif	nexthops[0].rif
 };
 
 void mlxsw_sp_nexthop_counter_alloc(struct mlxsw_sp *mlxsw_sp,
@@ -2950,18 +2956,18 @@ unsigned char *mlxsw_sp_nexthop_ha(struct mlxsw_sp_nexthop *nh)
 int mlxsw_sp_nexthop_indexes(struct mlxsw_sp_nexthop *nh, u32 *p_adj_index,
 			     u32 *p_adj_size, u32 *p_adj_hash_index)
 {
-	struct mlxsw_sp_nexthop_group *nh_grp = nh->nh_grp;
+	struct mlxsw_sp_nexthop_group_info *nhgi = nh->nhgi;
 	u32 adj_hash_index = 0;
 	int i;
 
-	if (!nh->offloaded || !nh_grp->adj_index_valid)
+	if (!nh->offloaded || !nhgi->adj_index_valid)
 		return -EINVAL;
 
-	*p_adj_index = nh_grp->adj_index;
-	*p_adj_size = nh_grp->ecmp_size;
+	*p_adj_index = nhgi->adj_index;
+	*p_adj_size = nhgi->ecmp_size;
 
-	for (i = 0; i < nh_grp->count; i++) {
-		struct mlxsw_sp_nexthop *nh_iter = &nh_grp->nexthops[i];
+	for (i = 0; i < nhgi->count; i++) {
+		struct mlxsw_sp_nexthop *nh_iter = &nhgi->nexthops[i];
 
 		if (nh_iter == nh)
 			break;
@@ -2980,11 +2986,11 @@ struct mlxsw_sp_rif *mlxsw_sp_nexthop_rif(struct mlxsw_sp_nexthop *nh)
 
 bool mlxsw_sp_nexthop_group_has_ipip(struct mlxsw_sp_nexthop *nh)
 {
-	struct mlxsw_sp_nexthop_group *nh_grp = nh->nh_grp;
+	struct mlxsw_sp_nexthop_group_info *nhgi = nh->nhgi;
 	int i;
 
-	for (i = 0; i < nh_grp->count; i++) {
-		struct mlxsw_sp_nexthop *nh_iter = &nh_grp->nexthops[i];
+	for (i = 0; i < nhgi->count; i++) {
+		struct mlxsw_sp_nexthop *nh_iter = &nhgi->nexthops[i];
 
 		if (nh_iter->type == MLXSW_SP_NEXTHOP_TYPE_IPIP)
 			return true;
@@ -3007,10 +3013,10 @@ mlxsw_sp_nexthop6_group_has_nexthop(const struct mlxsw_sp_nexthop_group *nh_grp,
 {
 	int i;
 
-	for (i = 0; i < nh_grp->count; i++) {
+	for (i = 0; i < nh_grp->nhgi->count; i++) {
 		const struct mlxsw_sp_nexthop *nh;
 
-		nh = &nh_grp->nexthops[i];
+		nh = &nh_grp->nhgi->nexthops[i];
 		if (nh->ifindex == ifindex && nh->nh_weight == weight &&
 		    ipv6_addr_equal(gw, (struct in6_addr *) nh->gw_addr))
 			return true;
@@ -3025,7 +3031,7 @@ mlxsw_sp_nexthop6_group_cmp(const struct mlxsw_sp_nexthop_group *nh_grp,
 {
 	struct mlxsw_sp_rt6 *mlxsw_sp_rt6;
 
-	if (nh_grp->count != fib6_entry->nrt6)
+	if (nh_grp->nhgi->count != fib6_entry->nrt6)
 		return false;
 
 	list_for_each_entry(mlxsw_sp_rt6, &fib6_entry->rt6_list, list) {
@@ -3078,9 +3084,9 @@ static u32 mlxsw_sp_nexthop_group_hash_obj(const void *data, u32 len, u32 seed)
 		fi = nh_grp->ipv4.fi;
 		return jhash(&fi, sizeof(fi), seed);
 	case MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV6:
-		val = nh_grp->count;
-		for (i = 0; i < nh_grp->count; i++) {
-			nh = &nh_grp->nexthops[i];
+		val = nh_grp->nhgi->count;
+		for (i = 0; i < nh_grp->nhgi->count; i++) {
+			nh = &nh_grp->nhgi->nexthops[i];
 			val ^= jhash(&nh->ifindex, sizeof(nh->ifindex), seed);
 			val ^= jhash(&nh->gw_addr, sizeof(nh->gw_addr), seed);
 		}
@@ -3136,7 +3142,7 @@ static int mlxsw_sp_nexthop_group_insert(struct mlxsw_sp *mlxsw_sp,
 					 struct mlxsw_sp_nexthop_group *nh_grp)
 {
 	if (nh_grp->type == MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV6 &&
-	    !nh_grp->gateway)
+	    !nh_grp->nhgi->gateway)
 		return 0;
 
 	return rhashtable_insert_fast(&mlxsw_sp->router->nexthop_group_ht,
@@ -3148,7 +3154,7 @@ static void mlxsw_sp_nexthop_group_remove(struct mlxsw_sp *mlxsw_sp,
 					  struct mlxsw_sp_nexthop_group *nh_grp)
 {
 	if (nh_grp->type == MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV6 &&
-	    !nh_grp->gateway)
+	    !nh_grp->nhgi->gateway)
 		return;
 
 	rhashtable_remove_fast(&mlxsw_sp->router->nexthop_group_ht,
@@ -3234,14 +3240,16 @@ static int mlxsw_sp_adj_index_mass_update(struct mlxsw_sp *mlxsw_sp,
 	int err;
 
 	list_for_each_entry(fib_entry, &nh_grp->fib_list, nexthop_group_node) {
+		struct mlxsw_sp_nexthop_group_info *nhgi = nh_grp->nhgi;
+
 		if (fib == fib_entry->fib_node->fib)
 			continue;
 		fib = fib_entry->fib_node->fib;
 		err = mlxsw_sp_adj_index_mass_update_vr(mlxsw_sp, fib,
 							old_adj_index,
 							old_ecmp_size,
-							nh_grp->adj_index,
-							nh_grp->ecmp_size);
+							nhgi->adj_index,
+							nhgi->ecmp_size);
 		if (err)
 			return err;
 	}
@@ -3312,15 +3320,15 @@ static int mlxsw_sp_nexthop_ipip_update(struct mlxsw_sp *mlxsw_sp,
 
 static int
 mlxsw_sp_nexthop_group_update(struct mlxsw_sp *mlxsw_sp,
-			      struct mlxsw_sp_nexthop_group *nh_grp,
+			      struct mlxsw_sp_nexthop_group_info *nhgi,
 			      bool reallocate)
 {
-	u32 adj_index = nh_grp->adj_index; /* base */
+	u32 adj_index = nhgi->adj_index; /* base */
 	struct mlxsw_sp_nexthop *nh;
 	int i;
 
-	for (i = 0; i < nh_grp->count; i++) {
-		nh = &nh_grp->nexthops[i];
+	for (i = 0; i < nhgi->count; i++) {
+		nh = &nhgi->nexthops[i];
 
 		if (!nh->should_offload) {
 			nh->offloaded = 0;
@@ -3420,13 +3428,13 @@ static int mlxsw_sp_fix_adj_grp_size(struct mlxsw_sp *mlxsw_sp,
 }
 
 static void
-mlxsw_sp_nexthop_group_normalize(struct mlxsw_sp_nexthop_group *nh_grp)
+mlxsw_sp_nexthop_group_normalize(struct mlxsw_sp_nexthop_group_info *nhgi)
 {
 	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];
+	for (i = 0; i < nhgi->count; i++) {
+		nh = &nhgi->nexthops[i];
 
 		if (!nh->should_offload)
 			continue;
@@ -3436,8 +3444,8 @@ mlxsw_sp_nexthop_group_normalize(struct mlxsw_sp_nexthop_group *nh_grp)
 			g = nh->nh_weight;
 	}
 
-	for (i = 0; i < nh_grp->count; i++) {
-		nh = &nh_grp->nexthops[i];
+	for (i = 0; i < nhgi->count; i++) {
+		nh = &nhgi->nexthops[i];
 
 		if (!nh->should_offload)
 			continue;
@@ -3445,18 +3453,18 @@ mlxsw_sp_nexthop_group_normalize(struct mlxsw_sp_nexthop_group *nh_grp)
 		sum_norm_weight += nh->norm_nh_weight;
 	}
 
-	nh_grp->sum_norm_weight = sum_norm_weight;
+	nhgi->sum_norm_weight = sum_norm_weight;
 }
 
 static void
-mlxsw_sp_nexthop_group_rebalance(struct mlxsw_sp_nexthop_group *nh_grp)
+mlxsw_sp_nexthop_group_rebalance(struct mlxsw_sp_nexthop_group_info *nhgi)
 {
-	int total = nh_grp->sum_norm_weight;
-	u16 ecmp_size = nh_grp->ecmp_size;
 	int i, weight = 0, lower_bound = 0;
+	int total = nhgi->sum_norm_weight;
+	u16 ecmp_size = nhgi->ecmp_size;
 
-	for (i = 0; i < nh_grp->count; i++) {
-		struct mlxsw_sp_nexthop *nh = &nh_grp->nexthops[i];
+	for (i = 0; i < nhgi->count; i++) {
+		struct mlxsw_sp_nexthop *nh = &nhgi->nexthops[i];
 		int upper_bound;
 
 		if (!nh->should_offload)
@@ -3478,8 +3486,8 @@ mlxsw_sp_nexthop4_group_offload_refresh(struct mlxsw_sp *mlxsw_sp,
 {
 	int i;
 
-	for (i = 0; i < nh_grp->count; i++) {
-		struct mlxsw_sp_nexthop *nh = &nh_grp->nexthops[i];
+	for (i = 0; i < nh_grp->nhgi->count; i++) {
+		struct mlxsw_sp_nexthop *nh = &nh_grp->nhgi->nexthops[i];
 
 		if (nh->offloaded)
 			nh->key.fib_nh->fib_nh_flags |= RTNH_F_OFFLOAD;
@@ -3539,6 +3547,7 @@ static void
 mlxsw_sp_nexthop_group_refresh(struct mlxsw_sp *mlxsw_sp,
 			       struct mlxsw_sp_nexthop_group *nh_grp)
 {
+	struct mlxsw_sp_nexthop_group_info *nhgi = nh_grp->nhgi;
 	u16 ecmp_size, old_ecmp_size;
 	struct mlxsw_sp_nexthop *nh;
 	bool offload_change = false;
@@ -3548,13 +3557,13 @@ mlxsw_sp_nexthop_group_refresh(struct mlxsw_sp *mlxsw_sp,
 	int i;
 	int err;
 
-	if (!nh_grp->gateway) {
+	if (!nhgi->gateway) {
 		mlxsw_sp_nexthop_fib_entries_update(mlxsw_sp, nh_grp);
 		return;
 	}
 
-	for (i = 0; i < nh_grp->count; i++) {
-		nh = &nh_grp->nexthops[i];
+	for (i = 0; i < nhgi->count; i++) {
+		nh = &nhgi->nexthops[i];
 
 		if (nh->should_offload != nh->offloaded) {
 			offload_change = true;
@@ -3566,21 +3575,21 @@ mlxsw_sp_nexthop_group_refresh(struct mlxsw_sp *mlxsw_sp,
 		/* Nothing was added or removed, so no need to reallocate. Just
 		 * update MAC on existing adjacency indexes.
 		 */
-		err = mlxsw_sp_nexthop_group_update(mlxsw_sp, nh_grp, false);
+		err = mlxsw_sp_nexthop_group_update(mlxsw_sp, nhgi, false);
 		if (err) {
 			dev_warn(mlxsw_sp->bus_info->dev, "Failed to update neigh MAC in adjacency table.\n");
 			goto set_trap;
 		}
 		return;
 	}
-	mlxsw_sp_nexthop_group_normalize(nh_grp);
-	if (!nh_grp->sum_norm_weight)
+	mlxsw_sp_nexthop_group_normalize(nhgi);
+	if (!nhgi->sum_norm_weight)
 		/* No neigh of this group is connected so we just set
 		 * the trap and let everthing flow through kernel.
 		 */
 		goto set_trap;
 
-	ecmp_size = nh_grp->sum_norm_weight;
+	ecmp_size = nhgi->sum_norm_weight;
 	err = mlxsw_sp_fix_adj_grp_size(mlxsw_sp, &ecmp_size);
 	if (err)
 		/* No valid allocation size available. */
@@ -3595,14 +3604,14 @@ mlxsw_sp_nexthop_group_refresh(struct mlxsw_sp *mlxsw_sp,
 		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;
-	mlxsw_sp_nexthop_group_rebalance(nh_grp);
-	err = mlxsw_sp_nexthop_group_update(mlxsw_sp, nh_grp, true);
+	old_adj_index_valid = nhgi->adj_index_valid;
+	old_adj_index = nhgi->adj_index;
+	old_ecmp_size = nhgi->ecmp_size;
+	nhgi->adj_index_valid = 1;
+	nhgi->adj_index = adj_index;
+	nhgi->ecmp_size = ecmp_size;
+	mlxsw_sp_nexthop_group_rebalance(nhgi);
+	err = mlxsw_sp_nexthop_group_update(mlxsw_sp, nhgi, true);
 	if (err) {
 		dev_warn(mlxsw_sp->bus_info->dev, "Failed to update neigh MAC in adjacency table.\n");
 		goto set_trap;
@@ -3634,10 +3643,10 @@ mlxsw_sp_nexthop_group_refresh(struct mlxsw_sp *mlxsw_sp,
 	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];
+	old_adj_index_valid = nhgi->adj_index_valid;
+	nhgi->adj_index_valid = 0;
+	for (i = 0; i < nhgi->count; i++) {
+		nh = &nhgi->nexthops[i];
 		nh->offloaded = 0;
 	}
 	err = mlxsw_sp_nexthop_fib_entries_update(mlxsw_sp, nh_grp);
@@ -3646,7 +3655,7 @@ mlxsw_sp_nexthop_group_refresh(struct mlxsw_sp *mlxsw_sp,
 	mlxsw_sp_nexthop_group_offload_refresh(mlxsw_sp, nh_grp);
 	if (old_adj_index_valid)
 		mlxsw_sp_kvdl_free(mlxsw_sp, MLXSW_SP_KVDL_ENTRY_TYPE_ADJ,
-				   nh_grp->ecmp_size, nh_grp->adj_index);
+				   nhgi->ecmp_size, nhgi->adj_index);
 }
 
 static void __mlxsw_sp_nexthop_neigh_update(struct mlxsw_sp_nexthop *nh,
@@ -3697,7 +3706,7 @@ mlxsw_sp_nexthop_dead_neigh_replace(struct mlxsw_sp *mlxsw_sp,
 		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);
+		mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh->nhgi->nh_grp);
 	}
 
 	neigh_release(n);
@@ -3734,7 +3743,7 @@ mlxsw_sp_nexthop_neigh_update(struct mlxsw_sp *mlxsw_sp,
 	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);
+		mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh->nhgi->nh_grp);
 	}
 }
 
@@ -3765,7 +3774,7 @@ static int mlxsw_sp_nexthop_neigh_init(struct mlxsw_sp *mlxsw_sp,
 	u8 nud_state, dead;
 	int err;
 
-	if (!nh->nh_grp->gateway || nh->neigh_entry)
+	if (!nh->nhgi->gateway || nh->neigh_entry)
 		return 0;
 
 	/* Take a reference of neigh here ensuring that neigh would
@@ -3856,7 +3865,7 @@ static void mlxsw_sp_nexthop_ipip_init(struct mlxsw_sp *mlxsw_sp,
 {
 	bool removing;
 
-	if (!nh->nh_grp->gateway || nh->ipip_entry)
+	if (!nh->nhgi->gateway || nh->ipip_entry)
 		return;
 
 	nh->ipip_entry = ipip_entry;
@@ -3956,7 +3965,7 @@ static int mlxsw_sp_nexthop4_init(struct mlxsw_sp *mlxsw_sp,
 	struct in_device *in_dev;
 	int err;
 
-	nh->nh_grp = nh_grp;
+	nh->nhgi = nh_grp->nhgi;
 	nh->key.fib_nh = fib_nh;
 #ifdef CONFIG_IP_ROUTE_MULTIPATH
 	nh->nh_weight = fib_nh->fib_nh_weight;
@@ -4027,7 +4036,7 @@ static void mlxsw_sp_nexthop4_event(struct mlxsw_sp *mlxsw_sp,
 		break;
 	}
 
-	mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh->nh_grp);
+	mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh->nhgi->nh_grp);
 }
 
 static void mlxsw_sp_nexthop_rif_update(struct mlxsw_sp *mlxsw_sp,
@@ -4050,7 +4059,7 @@ static void mlxsw_sp_nexthop_rif_update(struct mlxsw_sp *mlxsw_sp,
 		}
 
 		__mlxsw_sp_nexthop_neigh_update(nh, removing);
-		mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh->nh_grp);
+		mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh->nhgi->nh_grp);
 	}
 }
 
@@ -4073,7 +4082,7 @@ static void mlxsw_sp_nexthop_rif_gone_sync(struct mlxsw_sp *mlxsw_sp,
 
 	list_for_each_entry_safe(nh, tmp, &rif->nexthop_list, rif_list_node) {
 		mlxsw_sp_nexthop_type_fini(mlxsw_sp, nh);
-		mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh->nh_grp);
+		mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh->nhgi->nh_grp);
 	}
 }
 
@@ -4086,45 +4095,88 @@ static bool mlxsw_sp_fi_is_gateway(const struct mlxsw_sp *mlxsw_sp,
 	       mlxsw_sp_nexthop4_ipip_type(mlxsw_sp, nh, NULL);
 }
 
+static int
+mlxsw_sp_nexthop4_group_info_init(struct mlxsw_sp *mlxsw_sp,
+				  struct mlxsw_sp_nexthop_group *nh_grp)
+{
+	unsigned int nhs = fib_info_num_path(nh_grp->ipv4.fi);
+	struct mlxsw_sp_nexthop_group_info *nhgi;
+	struct mlxsw_sp_nexthop *nh;
+	int err, i;
+
+	nhgi = kzalloc(struct_size(nhgi, nexthops, nhs), GFP_KERNEL);
+	if (!nhgi)
+		return -ENOMEM;
+	nh_grp->nhgi = nhgi;
+	nhgi->nh_grp = nh_grp;
+	nhgi->gateway = mlxsw_sp_fi_is_gateway(mlxsw_sp, nh_grp->ipv4.fi);
+	nhgi->count = nhs;
+	for (i = 0; i < nhgi->count; i++) {
+		struct fib_nh *fib_nh;
+
+		nh = &nhgi->nexthops[i];
+		fib_nh = fib_info_nh(nh_grp->ipv4.fi, i);
+		err = mlxsw_sp_nexthop4_init(mlxsw_sp, nh_grp, nh, fib_nh);
+		if (err)
+			goto err_nexthop4_init;
+	}
+	mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh_grp);
+
+	return 0;
+
+err_nexthop4_init:
+	for (i--; i >= 0; i--) {
+		nh = &nhgi->nexthops[i];
+		mlxsw_sp_nexthop4_fini(mlxsw_sp, nh);
+	}
+	kfree(nhgi);
+	return err;
+}
+
+static void
+mlxsw_sp_nexthop4_group_info_fini(struct mlxsw_sp *mlxsw_sp,
+				  struct mlxsw_sp_nexthop_group *nh_grp)
+{
+	struct mlxsw_sp_nexthop_group_info *nhgi = nh_grp->nhgi;
+	int i;
+
+	for (i = nhgi->count - 1; i >= 0; i--) {
+		struct mlxsw_sp_nexthop *nh = &nhgi->nexthops[i];
+
+		mlxsw_sp_nexthop4_fini(mlxsw_sp, nh);
+	}
+	mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh_grp);
+	WARN_ON_ONCE(nhgi->adj_index_valid);
+	kfree(nhgi);
+}
+
 static struct mlxsw_sp_nexthop_group *
 mlxsw_sp_nexthop4_group_create(struct mlxsw_sp *mlxsw_sp, struct fib_info *fi)
 {
-	unsigned int nhs = fib_info_num_path(fi);
 	struct mlxsw_sp_nexthop_group *nh_grp;
-	struct mlxsw_sp_nexthop *nh;
-	struct fib_nh *fib_nh;
-	int i;
 	int err;
 
-	nh_grp = kzalloc(struct_size(nh_grp, nexthops, nhs), GFP_KERNEL);
+	nh_grp = kzalloc(sizeof(*nh_grp), GFP_KERNEL);
 	if (!nh_grp)
 		return ERR_PTR(-ENOMEM);
 	INIT_LIST_HEAD(&nh_grp->fib_list);
 	nh_grp->type = MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV4;
-
-	nh_grp->gateway = mlxsw_sp_fi_is_gateway(mlxsw_sp, fi);
-	nh_grp->count = nhs;
 	nh_grp->ipv4.fi = fi;
 	fib_info_hold(fi);
-	for (i = 0; i < nh_grp->count; i++) {
-		nh = &nh_grp->nexthops[i];
-		fib_nh = fib_info_nh(fi, i);
-		err = mlxsw_sp_nexthop4_init(mlxsw_sp, nh_grp, nh, fib_nh);
-		if (err)
-			goto err_nexthop4_init;
-	}
+
+	err = mlxsw_sp_nexthop4_group_info_init(mlxsw_sp, nh_grp);
+	if (err)
+		goto err_nexthop_group_info_init;
+
 	err = mlxsw_sp_nexthop_group_insert(mlxsw_sp, nh_grp);
 	if (err)
 		goto err_nexthop_group_insert;
-	mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh_grp);
+
 	return nh_grp;
 
 err_nexthop_group_insert:
-err_nexthop4_init:
-	for (i--; i >= 0; i--) {
-		nh = &nh_grp->nexthops[i];
-		mlxsw_sp_nexthop4_fini(mlxsw_sp, nh);
-	}
+	mlxsw_sp_nexthop4_group_info_fini(mlxsw_sp, nh_grp);
+err_nexthop_group_info_init:
 	fib_info_put(fi);
 	kfree(nh_grp);
 	return ERR_PTR(err);
@@ -4134,16 +4186,8 @@ static void
 mlxsw_sp_nexthop4_group_destroy(struct mlxsw_sp *mlxsw_sp,
 				struct mlxsw_sp_nexthop_group *nh_grp)
 {
-	struct mlxsw_sp_nexthop *nh;
-	int i;
-
 	mlxsw_sp_nexthop_group_remove(mlxsw_sp, nh_grp);
-	for (i = 0; i < nh_grp->count; i++) {
-		nh = &nh_grp->nexthops[i];
-		mlxsw_sp_nexthop4_fini(mlxsw_sp, nh);
-	}
-	mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh_grp);
-	WARN_ON_ONCE(nh_grp->adj_index_valid);
+	mlxsw_sp_nexthop4_group_info_fini(mlxsw_sp, nh_grp);
 	fib_info_put(nh_grp->ipv4.fi);
 	kfree(nh_grp);
 }
@@ -4202,9 +4246,9 @@ mlxsw_sp_fib_entry_should_offload(const struct mlxsw_sp_fib_entry *fib_entry)
 
 	switch (fib_entry->type) {
 	case MLXSW_SP_FIB_ENTRY_TYPE_REMOTE:
-		return !!nh_group->adj_index_valid;
+		return !!nh_group->nhgi->adj_index_valid;
 	case MLXSW_SP_FIB_ENTRY_TYPE_LOCAL:
-		return !!nh_group->nh_rif;
+		return !!nh_group->nhgi->nh_rif;
 	case MLXSW_SP_FIB_ENTRY_TYPE_BLACKHOLE:
 	case MLXSW_SP_FIB_ENTRY_TYPE_IPIP_DECAP:
 	case MLXSW_SP_FIB_ENTRY_TYPE_NVE_DECAP:
@@ -4220,8 +4264,8 @@ mlxsw_sp_rt6_nexthop(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];
+	for (i = 0; i < nh_grp->nhgi->count; i++) {
+		struct mlxsw_sp_nexthop *nh = &nh_grp->nhgi->nexthops[i];
 		struct fib6_info *rt = mlxsw_sp_rt6->rt;
 
 		if (nh->rif && nh->rif->dev == rt->fib6_nh->fib_nh_dev &&
@@ -4525,6 +4569,7 @@ static int mlxsw_sp_fib_entry_op_remote(struct mlxsw_sp *mlxsw_sp,
 {
 	const struct mlxsw_sp_router_ll_ops *ll_ops = fib_entry->fib_node->fib->ll_ops;
 	struct mlxsw_sp_nexthop_group *nh_group = fib_entry->nh_group;
+	struct mlxsw_sp_nexthop_group_info *nhgi = nh_group->nhgi;
 	enum mlxsw_reg_ralue_trap_action trap_action;
 	u16 trap_id = 0;
 	u32 adjacency_index = 0;
@@ -4537,12 +4582,11 @@ static int mlxsw_sp_fib_entry_op_remote(struct mlxsw_sp *mlxsw_sp,
 	 */
 	if (mlxsw_sp_fib_entry_should_offload(fib_entry)) {
 		trap_action = MLXSW_REG_RALUE_TRAP_ACTION_NOP;
-		adjacency_index = fib_entry->nh_group->adj_index;
-		ecmp_size = fib_entry->nh_group->ecmp_size;
-	} else if (!nh_group->adj_index_valid && nh_group->count &&
-		   nh_group->nh_rif) {
+		adjacency_index = nhgi->adj_index;
+		ecmp_size = nhgi->ecmp_size;
+	} else if (!nhgi->adj_index_valid && nhgi->count && nhgi->nh_rif) {
 		err = mlxsw_sp_adj_discard_write(mlxsw_sp,
-						 nh_group->nh_rif->rif_index);
+						 nhgi->nh_rif->rif_index);
 		if (err)
 			return err;
 		trap_action = MLXSW_REG_RALUE_TRAP_ACTION_NOP;
@@ -4565,7 +4609,7 @@ static int mlxsw_sp_fib_entry_op_local(struct mlxsw_sp *mlxsw_sp,
 				       enum mlxsw_sp_fib_entry_op op)
 {
 	const struct mlxsw_sp_router_ll_ops *ll_ops = fib_entry->fib_node->fib->ll_ops;
-	struct mlxsw_sp_rif *rif = fib_entry->nh_group->nh_rif;
+	struct mlxsw_sp_rif *rif = fib_entry->nh_group->nhgi->nh_rif;
 	enum mlxsw_reg_ralue_trap_action trap_action;
 	u16 trap_id = 0;
 	u16 rif_index = 0;
@@ -5365,7 +5409,7 @@ static int mlxsw_sp_nexthop6_init(struct mlxsw_sp *mlxsw_sp,
 {
 	struct net_device *dev = rt->fib6_nh->fib_nh_dev;
 
-	nh->nh_grp = nh_grp;
+	nh->nhgi = nh_grp->nhgi;
 	nh->nh_weight = rt->fib6_nh->fib_nh_weight;
 	memcpy(&nh->gw_addr, &rt->fib6_nh->fib_nh_gw6, sizeof(nh->gw_addr));
 #if IS_ENABLED(CONFIG_IPV6)
@@ -5397,49 +5441,92 @@ static bool mlxsw_sp_rt6_is_gateway(const struct mlxsw_sp *mlxsw_sp,
 	       mlxsw_sp_nexthop6_ipip_type(mlxsw_sp, rt, NULL);
 }
 
-static struct mlxsw_sp_nexthop_group *
-mlxsw_sp_nexthop6_group_create(struct mlxsw_sp *mlxsw_sp,
-			       struct mlxsw_sp_fib6_entry *fib6_entry)
+static int
+mlxsw_sp_nexthop6_group_info_init(struct mlxsw_sp *mlxsw_sp,
+				  struct mlxsw_sp_nexthop_group *nh_grp,
+				  struct mlxsw_sp_fib6_entry *fib6_entry)
 {
-	struct mlxsw_sp_nexthop_group *nh_grp;
+	struct mlxsw_sp_nexthop_group_info *nhgi;
 	struct mlxsw_sp_rt6 *mlxsw_sp_rt6;
 	struct mlxsw_sp_nexthop *nh;
-	int i = 0;
-	int err;
+	int err, i;
 
-	nh_grp = kzalloc(struct_size(nh_grp, nexthops, fib6_entry->nrt6),
-			 GFP_KERNEL);
-	if (!nh_grp)
-		return ERR_PTR(-ENOMEM);
-	INIT_LIST_HEAD(&nh_grp->fib_list);
-	nh_grp->type = MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV6;
+	nhgi = kzalloc(struct_size(nhgi, nexthops, fib6_entry->nrt6),
+		       GFP_KERNEL);
+	if (!nhgi)
+		return -ENOMEM;
+	nh_grp->nhgi = nhgi;
+	nhgi->nh_grp = nh_grp;
 	mlxsw_sp_rt6 = list_first_entry(&fib6_entry->rt6_list,
 					struct mlxsw_sp_rt6, list);
-	nh_grp->gateway = mlxsw_sp_rt6_is_gateway(mlxsw_sp, mlxsw_sp_rt6->rt);
-	nh_grp->count = fib6_entry->nrt6;
-	for (i = 0; i < nh_grp->count; i++) {
+	nhgi->gateway = mlxsw_sp_rt6_is_gateway(mlxsw_sp, mlxsw_sp_rt6->rt);
+	nhgi->count = fib6_entry->nrt6;
+	for (i = 0; i < nhgi->count; i++) {
 		struct fib6_info *rt = mlxsw_sp_rt6->rt;
 
-		nh = &nh_grp->nexthops[i];
+		nh = &nhgi->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);
 	}
+	nh_grp->nhgi = nhgi;
+	mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh_grp);
+
+	return 0;
+
+err_nexthop6_init:
+	for (i--; i >= 0; i--) {
+		nh = &nhgi->nexthops[i];
+		mlxsw_sp_nexthop6_fini(mlxsw_sp, nh);
+	}
+	kfree(nh_grp);
+	return err;
+}
+
+static void
+mlxsw_sp_nexthop6_group_info_fini(struct mlxsw_sp *mlxsw_sp,
+				  struct mlxsw_sp_nexthop_group *nh_grp)
+{
+	struct mlxsw_sp_nexthop_group_info *nhgi = nh_grp->nhgi;
+	int i;
+
+	for (i = nhgi->count - 1; i >= 0; i--) {
+		struct mlxsw_sp_nexthop *nh = &nhgi->nexthops[i];
+
+		mlxsw_sp_nexthop6_fini(mlxsw_sp, nh);
+	}
+	mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh_grp);
+	WARN_ON_ONCE(nhgi->adj_index_valid);
+	kfree(nhgi);
+}
+
+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;
+	int err;
+
+	nh_grp = kzalloc(sizeof(*nh_grp), GFP_KERNEL);
+	if (!nh_grp)
+		return ERR_PTR(-ENOMEM);
+	INIT_LIST_HEAD(&nh_grp->fib_list);
+	nh_grp->type = MLXSW_SP_NEXTHOP_GROUP_TYPE_IPV6;
+
+	err = mlxsw_sp_nexthop6_group_info_init(mlxsw_sp, nh_grp, fib6_entry);
+	if (err)
+		goto err_nexthop_group_info_init;
 
 	err = mlxsw_sp_nexthop_group_insert(mlxsw_sp, nh_grp);
 	if (err)
 		goto err_nexthop_group_insert;
 
-	mlxsw_sp_nexthop_group_refresh(mlxsw_sp, nh_grp);
 	return nh_grp;
 
 err_nexthop_group_insert:
-err_nexthop6_init:
-	for (i--; i >= 0; i--) {
-		nh = &nh_grp->nexthops[i];
-		mlxsw_sp_nexthop6_fini(mlxsw_sp, nh);
-	}
+	mlxsw_sp_nexthop6_group_info_fini(mlxsw_sp, nh_grp);
+err_nexthop_group_info_init:
 	kfree(nh_grp);
 	return ERR_PTR(err);
 }
@@ -5448,16 +5535,8 @@ 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;
-
 	mlxsw_sp_nexthop_group_remove(mlxsw_sp, nh_grp);
-	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);
+	mlxsw_sp_nexthop6_group_info_fini(mlxsw_sp, nh_grp);
 	kfree(nh_grp);
 }
 
-- 
2.28.0

[PATCH net-next 09/15] mlxsw: spectrum_ipip: Remove overlay protocol from can_offload() callback

[Ido Schimmel]

From: Ido Schimmel <idosch@nvidia.com>

The overlay protocol (i.e., IPv4/IPv6) that is being encapsulated has
no impact on whether a certain IP tunnel can be offloaded or not. Only
the underlay protocol matters.

Therefore, remove the unused overlay protocol parameter from the
callback.

This will later allow us to consolidate code paths between IPv4 and IPv6
code.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlxsw/spectrum_ipip.c  |  3 +--
 drivers/net/ethernet/mellanox/mlxsw/spectrum_ipip.h  |  3 +--
 .../net/ethernet/mellanox/mlxsw/spectrum_router.c    | 12 +++---------
 3 files changed, 5 insertions(+), 13 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_ipip.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_ipip.c
index ab2e0eb26c1a..089d99535f9e 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_ipip.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_ipip.c
@@ -233,8 +233,7 @@ static bool mlxsw_sp_ipip_tunnel_complete(enum mlxsw_sp_l3proto proto,
 }
 
 static bool mlxsw_sp_ipip_can_offload_gre4(const struct mlxsw_sp *mlxsw_sp,
-					   const struct net_device *ol_dev,
-					   enum mlxsw_sp_l3proto ol_proto)
+					   const struct net_device *ol_dev)
 {
 	struct ip_tunnel *tunnel = netdev_priv(ol_dev);
 	__be16 okflags = TUNNEL_KEY; /* We can't offload any other features. */
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_ipip.h b/drivers/net/ethernet/mellanox/mlxsw/spectrum_ipip.h
index 00448cbac639..d32702cb6ab4 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_ipip.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_ipip.h
@@ -43,8 +43,7 @@ struct mlxsw_sp_ipip_ops {
 			      struct mlxsw_sp_ipip_entry *ipip_entry);
 
 	bool (*can_offload)(const struct mlxsw_sp *mlxsw_sp,
-			    const struct net_device *ol_dev,
-			    enum mlxsw_sp_l3proto ol_proto);
+			    const struct net_device *ol_dev);
 
 	/* Return a configuration for creating an overlay loopback RIF. */
 	struct mlxsw_sp_rif_ipip_lb_config
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 3079be4bc5ec..7c7caea59d57 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -1453,11 +1453,7 @@ static bool mlxsw_sp_netdevice_ipip_can_offload(struct mlxsw_sp *mlxsw_sp,
 	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);
+	return ops->can_offload(mlxsw_sp, ol_dev);
 }
 
 static int mlxsw_sp_netdevice_ipip_ol_reg_event(struct mlxsw_sp *mlxsw_sp,
@@ -3925,8 +3921,7 @@ static int mlxsw_sp_nexthop4_type_init(struct mlxsw_sp *mlxsw_sp,
 	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)) {
+		if (ipip_ops->can_offload(mlxsw_sp, dev)) {
 			nh->type = MLXSW_SP_NEXTHOP_TYPE_IPIP;
 			mlxsw_sp_nexthop_ipip_init(mlxsw_sp, nh, ipip_entry);
 			return 0;
@@ -5371,8 +5366,7 @@ static int mlxsw_sp_nexthop6_type_init(struct mlxsw_sp *mlxsw_sp,
 	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)) {
+		if (ipip_ops->can_offload(mlxsw_sp, dev)) {
 			nh->type = MLXSW_SP_NEXTHOP_TYPE_IPIP;
 			mlxsw_sp_nexthop_ipip_init(mlxsw_sp, nh, ipip_entry);
 			return 0;
-- 
2.28.0

[PATCH net-next 10/15] mlxsw: spectrum_router: Pass nexthop netdev to mlxsw_sp_nexthop6_type_init()

[Ido Schimmel]

From: Ido Schimmel <idosch@nvidia.com>

Instead of passing the route and resolving the nexthop netdev from it,
pass the nexthop netdev directly.

This will later allow us to consolidate code paths between IPv4 and IPv6
code.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 7c7caea59d57..3bb93865f7b0 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -5355,11 +5355,10 @@ static bool mlxsw_sp_nexthop6_ipip_type(const struct mlxsw_sp *mlxsw_sp,
 static int mlxsw_sp_nexthop6_type_init(struct mlxsw_sp *mlxsw_sp,
 				       struct mlxsw_sp_nexthop_group *nh_grp,
 				       struct mlxsw_sp_nexthop *nh,
-				       const struct fib6_info *rt)
+				       const struct net_device *dev)
 {
 	const struct mlxsw_sp_ipip_ops *ipip_ops;
 	struct mlxsw_sp_ipip_entry *ipip_entry;
-	struct net_device *dev = rt->fib6_nh->fib_nh_dev;
 	struct mlxsw_sp_rif *rif;
 	int err;
 
@@ -5417,7 +5416,7 @@ static int mlxsw_sp_nexthop6_init(struct mlxsw_sp *mlxsw_sp,
 		return 0;
 	nh->ifindex = dev->ifindex;
 
-	return mlxsw_sp_nexthop6_type_init(mlxsw_sp, nh_grp, nh, rt);
+	return mlxsw_sp_nexthop6_type_init(mlxsw_sp, nh_grp, nh, dev);
 }
 
 static void mlxsw_sp_nexthop6_fini(struct mlxsw_sp *mlxsw_sp,
-- 
2.28.0

[PATCH net-next 11/15] mlxsw: spectrum_router: Pass nexthop netdev to mlxsw_sp_nexthop4_type_init()

[Ido Schimmel]

From: Ido Schimmel <idosch@nvidia.com>

Instead of passing the nexthop and resolving the nexthop netdev from it,
pass the nexthop netdev directly.

This will later allow us to consolidate code paths between IPv4 and IPv6
code.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 3bb93865f7b0..77be06b703bc 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -3910,10 +3910,9 @@ static void mlxsw_sp_nexthop_type_fini(struct mlxsw_sp *mlxsw_sp,
 
 static int mlxsw_sp_nexthop4_type_init(struct mlxsw_sp *mlxsw_sp,
 				       struct mlxsw_sp_nexthop *nh,
-				       struct fib_nh *fib_nh)
+				       const struct net_device *dev)
 {
 	const struct mlxsw_sp_ipip_ops *ipip_ops;
-	struct net_device *dev = fib_nh->fib_nh_dev;
 	struct mlxsw_sp_ipip_entry *ipip_entry;
 	struct mlxsw_sp_rif *rif;
 	int err;
@@ -3988,7 +3987,7 @@ static int mlxsw_sp_nexthop4_init(struct mlxsw_sp *mlxsw_sp,
 	}
 	rcu_read_unlock();
 
-	err = mlxsw_sp_nexthop4_type_init(mlxsw_sp, nh, fib_nh);
+	err = mlxsw_sp_nexthop4_type_init(mlxsw_sp, nh, dev);
 	if (err)
 		goto err_nexthop_neigh_init;
 
@@ -4024,7 +4023,7 @@ static void mlxsw_sp_nexthop4_event(struct mlxsw_sp *mlxsw_sp,
 
 	switch (event) {
 	case FIB_EVENT_NH_ADD:
-		mlxsw_sp_nexthop4_type_init(mlxsw_sp, nh, fib_nh);
+		mlxsw_sp_nexthop4_type_init(mlxsw_sp, nh, fib_nh->fib_nh_dev);
 		break;
 	case FIB_EVENT_NH_DEL:
 		mlxsw_sp_nexthop4_type_fini(mlxsw_sp, nh);
-- 
2.28.0

[PATCH net-next 12/15] mlxsw: spectrum_router: Remove unused argument from mlxsw_sp_nexthop6_type_init()

[Ido Schimmel]

From: Ido Schimmel <idosch@nvidia.com>

Remove it as it is unused.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 77be06b703bc..49b7b9294684 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -5352,7 +5352,6 @@ static bool mlxsw_sp_nexthop6_ipip_type(const struct mlxsw_sp *mlxsw_sp,
 }
 
 static int mlxsw_sp_nexthop6_type_init(struct mlxsw_sp *mlxsw_sp,
-				       struct mlxsw_sp_nexthop_group *nh_grp,
 				       struct mlxsw_sp_nexthop *nh,
 				       const struct net_device *dev)
 {
@@ -5415,7 +5414,7 @@ static int mlxsw_sp_nexthop6_init(struct mlxsw_sp *mlxsw_sp,
 		return 0;
 	nh->ifindex = dev->ifindex;
 
-	return mlxsw_sp_nexthop6_type_init(mlxsw_sp, nh_grp, nh, dev);
+	return mlxsw_sp_nexthop6_type_init(mlxsw_sp, nh, dev);
 }
 
 static void mlxsw_sp_nexthop6_fini(struct mlxsw_sp *mlxsw_sp,
-- 
2.28.0

[PATCH net-next 13/15] mlxsw: spectrum_router: Consolidate mlxsw_sp_nexthop{4, 6}_type_init()

[Ido Schimmel]

From: Ido Schimmel <idosch@nvidia.com>

The two functions are now identical, so consolidate them to
mlxsw_sp_nexthop_type_init().

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
---
 .../ethernet/mellanox/mlxsw/spectrum_router.c | 78 +++++--------------
 1 file changed, 21 insertions(+), 57 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 49b7b9294684..bbe07d74f2e8 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -3893,24 +3893,9 @@ static bool mlxsw_sp_nexthop4_ipip_type(const struct mlxsw_sp *mlxsw_sp,
 	       mlxsw_sp_netdev_ipip_type(mlxsw_sp, dev, p_ipipt);
 }
 
-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;
-	case MLXSW_SP_NEXTHOP_TYPE_IPIP:
-		mlxsw_sp_nexthop_rif_fini(nh);
-		mlxsw_sp_nexthop_ipip_fini(mlxsw_sp, nh);
-		break;
-	}
-}
-
-static int mlxsw_sp_nexthop4_type_init(struct mlxsw_sp *mlxsw_sp,
-				       struct mlxsw_sp_nexthop *nh,
-				       const struct net_device *dev)
+static int mlxsw_sp_nexthop_type_init(struct mlxsw_sp *mlxsw_sp,
+				      struct mlxsw_sp_nexthop *nh,
+				      const struct net_device *dev)
 {
 	const struct mlxsw_sp_ipip_ops *ipip_ops;
 	struct mlxsw_sp_ipip_entry *ipip_entry;
@@ -3944,6 +3929,21 @@ static int mlxsw_sp_nexthop4_type_init(struct mlxsw_sp *mlxsw_sp,
 	return err;
 }
 
+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;
+	case MLXSW_SP_NEXTHOP_TYPE_IPIP:
+		mlxsw_sp_nexthop_rif_fini(nh);
+		mlxsw_sp_nexthop_ipip_fini(mlxsw_sp, nh);
+		break;
+	}
+}
+
 static void mlxsw_sp_nexthop4_type_fini(struct mlxsw_sp *mlxsw_sp,
 					struct mlxsw_sp_nexthop *nh)
 {
@@ -3987,7 +3987,7 @@ static int mlxsw_sp_nexthop4_init(struct mlxsw_sp *mlxsw_sp,
 	}
 	rcu_read_unlock();
 
-	err = mlxsw_sp_nexthop4_type_init(mlxsw_sp, nh, dev);
+	err = mlxsw_sp_nexthop_type_init(mlxsw_sp, nh, dev);
 	if (err)
 		goto err_nexthop_neigh_init;
 
@@ -4023,7 +4023,7 @@ static void mlxsw_sp_nexthop4_event(struct mlxsw_sp *mlxsw_sp,
 
 	switch (event) {
 	case FIB_EVENT_NH_ADD:
-		mlxsw_sp_nexthop4_type_init(mlxsw_sp, nh, fib_nh->fib_nh_dev);
+		mlxsw_sp_nexthop_type_init(mlxsw_sp, nh, fib_nh->fib_nh_dev);
 		break;
 	case FIB_EVENT_NH_DEL:
 		mlxsw_sp_nexthop4_type_fini(mlxsw_sp, nh);
@@ -5351,42 +5351,6 @@ static bool mlxsw_sp_nexthop6_ipip_type(const struct mlxsw_sp *mlxsw_sp,
 	       mlxsw_sp_netdev_ipip_type(mlxsw_sp, rt->fib6_nh->fib_nh_dev, ret);
 }
 
-static int mlxsw_sp_nexthop6_type_init(struct mlxsw_sp *mlxsw_sp,
-				       struct mlxsw_sp_nexthop *nh,
-				       const struct net_device *dev)
-{
-	const struct mlxsw_sp_ipip_ops *ipip_ops;
-	struct mlxsw_sp_ipip_entry *ipip_entry;
-	struct mlxsw_sp_rif *rif;
-	int err;
-
-	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)) {
-			nh->type = MLXSW_SP_NEXTHOP_TYPE_IPIP;
-			mlxsw_sp_nexthop_ipip_init(mlxsw_sp, nh, ipip_entry);
-			return 0;
-		}
-	}
-
-	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_nexthop_neigh_init;
-
-	return 0;
-
-err_nexthop_neigh_init:
-	mlxsw_sp_nexthop_rif_fini(nh);
-	return err;
-}
-
 static void mlxsw_sp_nexthop6_type_fini(struct mlxsw_sp *mlxsw_sp,
 					struct mlxsw_sp_nexthop *nh)
 {
@@ -5414,7 +5378,7 @@ static int mlxsw_sp_nexthop6_init(struct mlxsw_sp *mlxsw_sp,
 		return 0;
 	nh->ifindex = dev->ifindex;
 
-	return mlxsw_sp_nexthop6_type_init(mlxsw_sp, nh, dev);
+	return mlxsw_sp_nexthop_type_init(mlxsw_sp, nh, dev);
 }
 
 static void mlxsw_sp_nexthop6_fini(struct mlxsw_sp *mlxsw_sp,
-- 
2.28.0

[PATCH net-next 14/15] mlxsw: spectrum_router: Consolidate mlxsw_sp_nexthop{4, 6}_type_fini()

[Ido Schimmel]

From: Ido Schimmel <idosch@nvidia.com>

The two functions are identical, so consolidate them to
mlxsw_sp_nexthop_type_fini().

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
---
 .../ethernet/mellanox/mlxsw/spectrum_router.c  | 18 +++---------------
 1 file changed, 3 insertions(+), 15 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index bbe07d74f2e8..ca9f7d06eab1 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -3944,12 +3944,6 @@ static void mlxsw_sp_nexthop_type_fini(struct mlxsw_sp *mlxsw_sp,
 	}
 }
 
-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);
-}
-
 static int mlxsw_sp_nexthop4_init(struct mlxsw_sp *mlxsw_sp,
 				  struct mlxsw_sp_nexthop_group *nh_grp,
 				  struct mlxsw_sp_nexthop *nh,
@@ -4001,7 +3995,7 @@ static int mlxsw_sp_nexthop4_init(struct mlxsw_sp *mlxsw_sp,
 static void mlxsw_sp_nexthop4_fini(struct mlxsw_sp *mlxsw_sp,
 				   struct mlxsw_sp_nexthop *nh)
 {
-	mlxsw_sp_nexthop4_type_fini(mlxsw_sp, nh);
+	mlxsw_sp_nexthop_type_fini(mlxsw_sp, nh);
 	list_del(&nh->router_list_node);
 	mlxsw_sp_nexthop_counter_free(mlxsw_sp, nh);
 	mlxsw_sp_nexthop_remove(mlxsw_sp, nh);
@@ -4026,7 +4020,7 @@ static void mlxsw_sp_nexthop4_event(struct mlxsw_sp *mlxsw_sp,
 		mlxsw_sp_nexthop_type_init(mlxsw_sp, nh, fib_nh->fib_nh_dev);
 		break;
 	case FIB_EVENT_NH_DEL:
-		mlxsw_sp_nexthop4_type_fini(mlxsw_sp, nh);
+		mlxsw_sp_nexthop_type_fini(mlxsw_sp, nh);
 		break;
 	}
 
@@ -5351,12 +5345,6 @@ static bool mlxsw_sp_nexthop6_ipip_type(const struct mlxsw_sp *mlxsw_sp,
 	       mlxsw_sp_netdev_ipip_type(mlxsw_sp, rt->fib6_nh->fib_nh_dev, ret);
 }
 
-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,
@@ -5384,7 +5372,7 @@ static int mlxsw_sp_nexthop6_init(struct mlxsw_sp *mlxsw_sp,
 static void mlxsw_sp_nexthop6_fini(struct mlxsw_sp *mlxsw_sp,
 				   struct mlxsw_sp_nexthop *nh)
 {
-	mlxsw_sp_nexthop6_type_fini(mlxsw_sp, nh);
+	mlxsw_sp_nexthop_type_fini(mlxsw_sp, nh);
 	list_del(&nh->router_list_node);
 	mlxsw_sp_nexthop_counter_free(mlxsw_sp, nh);
 }
-- 
2.28.0

[PATCH net-next 15/15] mlxsw: spectrum_router: Remove outdated comment

[Ido Schimmel]

From: Ido Schimmel <idosch@nvidia.com>

Since commit 21151f64a458 ("mlxsw: Add new FIB entry type for reject
routes") this comment is no longer correct. Remove it.

Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c | 6 ------
 1 file changed, 6 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index ca9f7d06eab1..a2e81ad5790f 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -5621,12 +5621,6 @@ static void mlxsw_sp_fib6_entry_type_set(struct mlxsw_sp *mlxsw_sp,
 					 struct mlxsw_sp_fib_entry *fib_entry,
 					 const struct fib6_info *rt)
 {
-	/* 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.
-	 */
 	if (rt->fib6_flags & (RTF_LOCAL | RTF_ANYCAST))
 		fib_entry->type = MLXSW_SP_FIB_ENTRY_TYPE_TRAP;
 	else if (rt->fib6_type == RTN_BLACKHOLE)
-- 
2.28.0

Re: [PATCH net-next 00/15] mlxsw: Preparations for nexthop objects support - part 1/2

[Jakub Kicinski]

On Fri, 13 Nov 2020 18:05:44 +0200 Ido Schimmel wrote:
> From: Ido Schimmel <idosch@nvidia.com>
> 
> This patch set contains small and non-functional changes aimed at making
> it easier to support nexthop objects in mlxsw. Follow up patches can be
> found here [1].
> 
> Patches #1-#4 add a type field to the nexthop group struct instead of
> the existing protocol field. This will be used later on to add a nexthop
> object type, which can contain both IPv4 and IPv6 nexthops.
> 
> Patches #5-#7 move the IPv4 FIB info pointer (i.e., 'struct fib_info')
> from the nexthop group struct to the route. The pointer will not be
> available when the nexthop group is a nexthop object, but it needs to be
> accessible to routes regardless.
> 
> Patch #8 is the biggest change, but it is an entirely cosmetic change
> and should therefore be easy to review. The motivation and the change
> itself are explained in detail in the commit message.
> 
> Patches #9-#12 perform small changes so that two functions that are
> currently split between IPv4 and IPv6 could be consolidated in patches
> #13 and #14. The functions will be reused for nexthop objects.
> 
> Patch #15 removes an outdated comment.

Pleasant read, applied, thanks!