Merge branch 'nexthop-add-support-for-nexthop-objects-offload'

Ido Schimmel says:

====================
nexthop: Add support for nexthop objects offload

This patch set adds support for nexthop objects offload with a dummy
implementation over netdevsim. mlxsw support will be added later.

The general idea is very similar to route offload in that notifications
are sent whenever nexthop objects are changed. A listener can veto the
change and the error will be communicated to user space with extack.

To keep listeners as simple as possible, they not only receive
notifications for the nexthop object that is changed, but also for all
the other objects affected by this change. For example, when a single
nexthop is replaced, a replace notification is sent for the single
nexthop, but also for all the nexthop groups this nexthop is member in.
This relieves listeners from the need to track such dependencies.

To simplify things further for listeners, the notification info does not
contain the raw nexthop data structures (e.g., 'struct nexthop'), but
less complex data structures into which the raw data structures are
parsed into.

Tested with a new selftest over netdevsim and with fib_nexthops.sh:

Tests passed: 164
Tests failed:   0

Patch set overview:

Patches #1-#4 introduce the aforementioned data structures and convert
existing listeners (i.e., the VXLAN driver) to use them.

Patches #5-#6 add a new RTNH_F_TRAP flag and the ability to set it and
RTNH_F_OFFLOAD on nexthops. This flag is used by netdevsim for testing
purposes and will also be used by mlxsw. These flags are consistent with
the existing RTM_F_OFFLOAD and RTM_F_TRAP flags.

Patches #7-#14 gradually add the new nexthop notifications.

Patches #15-#18 add a dummy implementation for nexthop offload over
netdevsim and a selftest to exercise both good and bad flows.

Changes since RFC [1]:

Patch #1: s/is_encap/has_encap/
Patch #3: Add a blank line in __nh_notifier_single_info_init()
Patch #5: Reword commit message
Patch #6: s/nexthop_hw_flags_set/nexthop_set_hw_flags/
Patch #7: Reword commit message
Patch #11: Allocate extack on the stack

Follow-up patch sets:

selftests: forwarding: Add nexthop objects tests
mlxsw: Preparations for nexthop objects support - part 1/2
mlxsw: Preparations for nexthop objects support - part 2/2
mlxsw: Add support for nexthop objects
mlxsw: Add support for blackhole nexthops
mlxsw: Update adjacency index more efficiently

[1] https://lore.kernel.org/netdev/20200908091037.2709823-1-idosch@idosch.org/
====================

Link: https://lore.kernel.org/r/20201104133040.1125369-1-idosch@idosch.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

1 files changed, 251 insertions, 14 deletions

diff --git a/drivers/net/netdevsim/fib.c b/drivers/net/netdevsim/fib.c
index deea17a0e79c..45d8a7790bd5 100644
--- a/drivers/net/netdevsim/fib.c
+++ b/drivers/net/netdevsim/fib.c
@@ -25,6 +25,7 @@
 #include <net/ip6_fib.h>
 #include <net/fib_rules.h>
 #include <net/net_namespace.h>
+#include <net/nexthop.h>
 
 #include "netdevsim.h"
 
@@ -42,9 +43,12 @@ struct nsim_fib_data {
 	struct notifier_block fib_nb;
 	struct nsim_per_fib_data ipv4;
 	struct nsim_per_fib_data ipv6;
+	struct nsim_fib_entry nexthops;
 	struct rhashtable fib_rt_ht;
 	struct list_head fib_rt_list;
 	spinlock_t fib_lock;	/* Protects hashtable, list and accounting */
+	struct notifier_block nexthop_nb;
+	struct rhashtable nexthop_ht;
 	struct devlink *devlink;
 };
 
@@ -86,6 +90,19 @@ static const struct rhashtable_params nsim_fib_rt_ht_params = {
 	.automatic_shrinking = true,
 };
 
+struct nsim_nexthop {
+	struct rhash_head ht_node;
+	u64 occ;
+	u32 id;
+};
+
+static const struct rhashtable_params nsim_nexthop_ht_params = {
+	.key_offset = offsetof(struct nsim_nexthop, id),
+	.head_offset = offsetof(struct nsim_nexthop, ht_node),
+	.key_len = sizeof(u32),
+	.automatic_shrinking = true,
+};
+
 u64 nsim_fib_get_val(struct nsim_fib_data *fib_data,
 		     enum nsim_resource_id res_id, bool max)
 {
@@ -104,6 +121,9 @@ u64 nsim_fib_get_val(struct nsim_fib_data *fib_data,
 	case NSIM_RESOURCE_IPV6_FIB_RULES:
 		entry = &fib_data->ipv6.rules;
 		break;
+	case NSIM_RESOURCE_NEXTHOPS:
+		entry = &fib_data->nexthops;
+		break;
 	default:
 		return 0;
 	}
@@ -129,6 +149,9 @@ static void nsim_fib_set_max(struct nsim_fib_data *fib_data,
 	case NSIM_RESOURCE_IPV6_FIB_RULES:
 		entry = &fib_data->ipv6.rules;
 		break;
+	case NSIM_RESOURCE_NEXTHOPS:
+		entry = &fib_data->nexthops;
+		break;
 	default:
 		WARN_ON(1);
 		return;
@@ -389,11 +412,6 @@ static int nsim_fib4_event(struct nsim_fib_data *data,
 
 	fen_info = container_of(info, struct fib_entry_notifier_info, info);
 
-	if (fen_info->fi->nh) {
-		NL_SET_ERR_MSG_MOD(info->extack, "IPv4 route with nexthop objects is not supported");
-		return 0;
-	}
-
 	switch (event) {
 	case FIB_EVENT_ENTRY_REPLACE:
 		err = nsim_fib4_rt_insert(data, fen_info);
@@ -704,11 +722,6 @@ static int nsim_fib6_event(struct nsim_fib_data *data,
 
 	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 0;
-	}
-
 	if (fen6_info->rt->fib6_src.plen) {
 		NL_SET_ERR_MSG_MOD(info->extack, "IPv6 source-specific route is not supported");
 		return 0;
@@ -838,6 +851,196 @@ static void nsim_fib_dump_inconsistent(struct notifier_block *nb)
 	data->ipv6.rules.num = 0ULL;
 }
 
+static struct nsim_nexthop *nsim_nexthop_create(struct nsim_fib_data *data,
+						struct nh_notifier_info *info)
+{
+	struct nsim_nexthop *nexthop;
+	u64 occ = 0;
+	int i;
+
+	nexthop = kzalloc(sizeof(*nexthop), GFP_KERNEL);
+	if (!nexthop)
+		return NULL;
+
+	nexthop->id = info->id;
+
+	/* Determine the number of nexthop entries the new nexthop will
+	 * occupy.
+	 */
+
+	if (!info->is_grp) {
+		occ = 1;
+		goto out;
+	}
+
+	for (i = 0; i < info->nh_grp->num_nh; i++)
+		occ += info->nh_grp->nh_entries[i].weight;
+
+out:
+	nexthop->occ = occ;
+	return nexthop;
+}
+
+static void nsim_nexthop_destroy(struct nsim_nexthop *nexthop)
+{
+	kfree(nexthop);
+}
+
+static int nsim_nexthop_account(struct nsim_fib_data *data, u64 occ,
+				bool add, struct netlink_ext_ack *extack)
+{
+	int err = 0;
+
+	if (add) {
+		if (data->nexthops.num + occ <= data->nexthops.max) {
+			data->nexthops.num += occ;
+		} else {
+			err = -ENOSPC;
+			NL_SET_ERR_MSG_MOD(extack, "Exceeded number of supported nexthops");
+		}
+	} else {
+		if (WARN_ON(occ > data->nexthops.num))
+			return -EINVAL;
+		data->nexthops.num -= occ;
+	}
+
+	return err;
+}
+
+static int nsim_nexthop_add(struct nsim_fib_data *data,
+			    struct nsim_nexthop *nexthop,
+			    struct netlink_ext_ack *extack)
+{
+	struct net *net = devlink_net(data->devlink);
+	int err;
+
+	err = nsim_nexthop_account(data, nexthop->occ, true, extack);
+	if (err)
+		return err;
+
+	err = rhashtable_insert_fast(&data->nexthop_ht, &nexthop->ht_node,
+				     nsim_nexthop_ht_params);
+	if (err) {
+		NL_SET_ERR_MSG_MOD(extack, "Failed to insert nexthop");
+		goto err_nexthop_dismiss;
+	}
+
+	nexthop_set_hw_flags(net, nexthop->id, false, true);
+
+	return 0;
+
+err_nexthop_dismiss:
+	nsim_nexthop_account(data, nexthop->occ, false, extack);
+	return err;
+}
+
+static int nsim_nexthop_replace(struct nsim_fib_data *data,
+				struct nsim_nexthop *nexthop,
+				struct nsim_nexthop *nexthop_old,
+				struct netlink_ext_ack *extack)
+{
+	struct net *net = devlink_net(data->devlink);
+	int err;
+
+	err = nsim_nexthop_account(data, nexthop->occ, true, extack);
+	if (err)
+		return err;
+
+	err = rhashtable_replace_fast(&data->nexthop_ht,
+				      &nexthop_old->ht_node, &nexthop->ht_node,
+				      nsim_nexthop_ht_params);
+	if (err) {
+		NL_SET_ERR_MSG_MOD(extack, "Failed to replace nexthop");
+		goto err_nexthop_dismiss;
+	}
+
+	nexthop_set_hw_flags(net, nexthop->id, false, true);
+	nsim_nexthop_account(data, nexthop_old->occ, false, extack);
+	nsim_nexthop_destroy(nexthop_old);
+
+	return 0;
+
+err_nexthop_dismiss:
+	nsim_nexthop_account(data, nexthop->occ, false, extack);
+	return err;
+}
+
+static int nsim_nexthop_insert(struct nsim_fib_data *data,
+			       struct nh_notifier_info *info)
+{
+	struct nsim_nexthop *nexthop, *nexthop_old;
+	int err;
+
+	nexthop = nsim_nexthop_create(data, info);
+	if (!nexthop)
+		return -ENOMEM;
+
+	nexthop_old = rhashtable_lookup_fast(&data->nexthop_ht, &info->id,
+					     nsim_nexthop_ht_params);
+	if (!nexthop_old)
+		err = nsim_nexthop_add(data, nexthop, info->extack);
+	else
+		err = nsim_nexthop_replace(data, nexthop, nexthop_old,
+					   info->extack);
+
+	if (err)
+		nsim_nexthop_destroy(nexthop);
+
+	return err;
+}
+
+static void nsim_nexthop_remove(struct nsim_fib_data *data,
+				struct nh_notifier_info *info)
+{
+	struct nsim_nexthop *nexthop;
+
+	nexthop = rhashtable_lookup_fast(&data->nexthop_ht, &info->id,
+					 nsim_nexthop_ht_params);
+	if (!nexthop)
+		return;
+
+	rhashtable_remove_fast(&data->nexthop_ht, &nexthop->ht_node,
+			       nsim_nexthop_ht_params);
+	nsim_nexthop_account(data, nexthop->occ, false, info->extack);
+	nsim_nexthop_destroy(nexthop);
+}
+
+static int nsim_nexthop_event_nb(struct notifier_block *nb, unsigned long event,
+				 void *ptr)
+{
+	struct nsim_fib_data *data = container_of(nb, struct nsim_fib_data,
+						  nexthop_nb);
+	struct nh_notifier_info *info = ptr;
+	int err = 0;
+
+	ASSERT_RTNL();
+
+	switch (event) {
+	case NEXTHOP_EVENT_REPLACE:
+		err = nsim_nexthop_insert(data, info);
+		break;
+	case NEXTHOP_EVENT_DEL:
+		nsim_nexthop_remove(data, info);
+		break;
+	default:
+		break;
+	}
+
+	return notifier_from_errno(err);
+}
+
+static void nsim_nexthop_free(void *ptr, void *arg)
+{
+	struct nsim_nexthop *nexthop = ptr;
+	struct nsim_fib_data *data = arg;
+	struct net *net;
+
+	net = devlink_net(data->devlink);
+	nexthop_set_hw_flags(net, nexthop->id, false, false);
+	nsim_nexthop_account(data, nexthop->occ, false, NULL);
+	nsim_nexthop_destroy(nexthop);
+}
+
 static u64 nsim_fib_ipv4_resource_occ_get(void *priv)
 {
 	struct nsim_fib_data *data = priv;
@@ -866,12 +1069,20 @@ static u64 nsim_fib_ipv6_rules_res_occ_get(void *priv)
 	return nsim_fib_get_val(data, NSIM_RESOURCE_IPV6_FIB_RULES, false);
 }
 
+static u64 nsim_fib_nexthops_res_occ_get(void *priv)
+{
+	struct nsim_fib_data *data = priv;
+
+	return nsim_fib_get_val(data, NSIM_RESOURCE_NEXTHOPS, false);
+}
+
 static void nsim_fib_set_max_all(struct nsim_fib_data *data,
 				 struct devlink *devlink)
 {
 	enum nsim_resource_id res_ids[] = {
 		NSIM_RESOURCE_IPV4_FIB, NSIM_RESOURCE_IPV4_FIB_RULES,
-		NSIM_RESOURCE_IPV6_FIB, NSIM_RESOURCE_IPV6_FIB_RULES
+		NSIM_RESOURCE_IPV6_FIB, NSIM_RESOURCE_IPV6_FIB_RULES,
+		NSIM_RESOURCE_NEXTHOPS,
 	};
 	int i;
 
@@ -897,20 +1108,32 @@ struct nsim_fib_data *nsim_fib_create(struct devlink *devlink,
 		return ERR_PTR(-ENOMEM);
 	data->devlink = devlink;
 
+	err = rhashtable_init(&data->nexthop_ht, &nsim_nexthop_ht_params);
+	if (err)
+		goto err_data_free;
+
 	spin_lock_init(&data->fib_lock);
 	INIT_LIST_HEAD(&data->fib_rt_list);
 	err = rhashtable_init(&data->fib_rt_ht, &nsim_fib_rt_ht_params);
 	if (err)
-		goto err_data_free;
+		goto err_rhashtable_nexthop_destroy;
 
 	nsim_fib_set_max_all(data, devlink);
 
+	data->nexthop_nb.notifier_call = nsim_nexthop_event_nb;
+	err = register_nexthop_notifier(devlink_net(devlink), &data->nexthop_nb,
+					extack);
+	if (err) {
+		pr_err("Failed to register nexthop notifier\n");
+		goto err_rhashtable_fib_destroy;
+	}
+
 	data->fib_nb.notifier_call = nsim_fib_event_nb;
 	err = register_fib_notifier(devlink_net(devlink), &data->fib_nb,
 				    nsim_fib_dump_inconsistent, extack);
 	if (err) {
 		pr_err("Failed to register fib notifier\n");
-		goto err_rhashtable_destroy;
+		goto err_nexthop_nb_unregister;
 	}
 
 	devlink_resource_occ_get_register(devlink,
@@ -929,11 +1152,20 @@ struct nsim_fib_data *nsim_fib_create(struct devlink *devlink,
 					  NSIM_RESOURCE_IPV6_FIB_RULES,
 					  nsim_fib_ipv6_rules_res_occ_get,
 					  data);
+	devlink_resource_occ_get_register(devlink,
+					  NSIM_RESOURCE_NEXTHOPS,
+					  nsim_fib_nexthops_res_occ_get,
+					  data);
 	return data;
 
-err_rhashtable_destroy:
+err_nexthop_nb_unregister:
+	unregister_nexthop_notifier(devlink_net(devlink), &data->nexthop_nb);
+err_rhashtable_fib_destroy:
 	rhashtable_free_and_destroy(&data->fib_rt_ht, nsim_fib_rt_free,
 				    data);
+err_rhashtable_nexthop_destroy:
+	rhashtable_free_and_destroy(&data->nexthop_ht, nsim_nexthop_free,
+				    data);
 err_data_free:
 	kfree(data);
 	return ERR_PTR(err);
@@ -942,6 +1174,8 @@ err_data_free:
 void nsim_fib_destroy(struct devlink *devlink, struct nsim_fib_data *data)
 {
 	devlink_resource_occ_get_unregister(devlink,
+					    NSIM_RESOURCE_NEXTHOPS);
+	devlink_resource_occ_get_unregister(devlink,
 					    NSIM_RESOURCE_IPV6_FIB_RULES);
 	devlink_resource_occ_get_unregister(devlink,
 					    NSIM_RESOURCE_IPV6_FIB);
@@ -950,8 +1184,11 @@ void nsim_fib_destroy(struct devlink *devlink, struct nsim_fib_data *data)
 	devlink_resource_occ_get_unregister(devlink,
 					    NSIM_RESOURCE_IPV4_FIB);
 	unregister_fib_notifier(devlink_net(devlink), &data->fib_nb);
+	unregister_nexthop_notifier(devlink_net(devlink), &data->nexthop_nb);
 	rhashtable_free_and_destroy(&data->fib_rt_ht, nsim_fib_rt_free,
 				    data);
+	rhashtable_free_and_destroy(&data->nexthop_ht, nsim_nexthop_free,
+				    data);
 	WARN_ON_ONCE(!list_empty(&data->fib_rt_list));
 	kfree(data);
 }