netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [patch net-next 00/15] devlink: allow devlink instances to change network namespace
@ 2019-09-14  6:45 Jiri Pirko
  2019-09-14  6:45 ` [patch net-next 01/15] netdevsim: change fib accounting and limitations to be per-device Jiri Pirko
                   ` (17 more replies)
  0 siblings, 18 replies; 34+ messages in thread
From: Jiri Pirko @ 2019-09-14  6:45 UTC (permalink / raw)
  To: netdev
  Cc: davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

From: Jiri Pirko <jiri@mellanox.com>

Devlink from the beginning counts with network namespaces, but the
instances has been fixed to init_net.

Implement change of network namespace as part of "devlink reload"
procedure like this:

$ ip netns add testns1
$ devlink/devlink dev reload netdevsim/netdevsim10 netns testns1

This command reloads device "netdevsim10" into network
namespace "testns1".

Note that "devlink reload" reinstantiates driver objects, effectively it
reloads the driver instance, including possible hw reset etc. Newly
created netdevices respect the network namespace of the parent devlink
instance and according to that, they are created in target network
namespace.

Driver is able to refuse to be reloaded into different namespace. That
is the case of mlx4 right now.

FIB entries and rules are replayed during FIB notifier registration
which is triggered during reload (driver instance init). FIB notifier
is also registered to the target network namespace, that allows user
to use netdevsim devlink resources to setup per-namespace limits of FIB
entries and FIB rules. In fact, with multiple netdevsim instances
in each network namespace, user might setup different limits.
This maintains and extends current netdevsim resources behaviour.

Patch 1 prepares netdevsim code for the follow-up changes in the
patchset. It does not change the behaviour, only moves pet-init_netns
accounting to netdevsim instance, which is also in init_netns.

Patches 2-5 prepare the FIB notifier making it per-netns and to behave
correctly upon error conditions.

Patch 6 just exports a devlink_net helper so it can be used in drivers.

Patches 7-9 do preparations in mlxsw driver.

Patches 10-13 do preparations in netdevsim driver, namely patch 12
implements proper devlink reload where the driver instance objects are
actually re-created as they should be.

Patch 14 actually implements the possibility to reload into a different
network namespace.

Patch 15 adds needed selftests for devlink reload into namespace for
netdevsim driver.

Jiri Pirko (15):
  netdevsim: change fib accounting and limitations to be per-device
  net: fib_notifier: make FIB notifier per-netns
  net: fib_notifier: propagate possible error during fib notifier
    registration
  mlxsw: spectrum_router: Don't rely on missing extack to symbolize dump
  net: fib_notifier: propagate extack down to the notifier block
    callback
  net: devlink: export devlink net getter
  mlxsw: spectrum: Take devlink net instead of init_net
  mlxsw: Register port netdevices into net of core
  mlxsw: Propagate extack down to register_fib_notifier()
  netdevsim: add all ports in nsim_dev_create() and del them in
    destroy()
  netdevsim: implement proper devlink reload
  netdevsim: register port netdevices into net of device
  netdevsim: take devlink net instead of init_net
  net: devlink: allow to change namespaces during reload
  selftests: netdevsim: add tests for devlink reload with resources

 drivers/net/ethernet/mellanox/mlx4/main.c     |   4 +
 .../net/ethernet/mellanox/mlx5/core/lag_mp.c  |   9 +-
 drivers/net/ethernet/mellanox/mlxsw/core.c    |  13 +-
 drivers/net/ethernet/mellanox/mlxsw/core.h    |  12 +-
 drivers/net/ethernet/mellanox/mlxsw/i2c.c     |   2 +-
 drivers/net/ethernet/mellanox/mlxsw/minimal.c |   4 +-
 drivers/net/ethernet/mellanox/mlxsw/pci.c     |   2 +-
 .../net/ethernet/mellanox/mlxsw/spectrum.c    |  16 +-
 .../net/ethernet/mellanox/mlxsw/spectrum.h    |   9 +-
 .../ethernet/mellanox/mlxsw/spectrum_nve.c    |   2 +-
 .../ethernet/mellanox/mlxsw/spectrum_router.c |  57 ++-
 .../mellanox/mlxsw/spectrum_switchdev.c       |   2 +-
 .../net/ethernet/mellanox/mlxsw/switchib.c    |   3 +-
 .../net/ethernet/mellanox/mlxsw/switchx2.c    |   4 +-
 drivers/net/ethernet/rocker/rocker_main.c     |   9 +-
 drivers/net/netdevsim/dev.c                   | 341 +++++++++---------
 drivers/net/netdevsim/fib.c                   | 175 +++++----
 drivers/net/netdevsim/netdev.c                |  10 +-
 drivers/net/netdevsim/netdevsim.h             |  15 +-
 include/linux/mroute_base.h                   |  26 +-
 include/net/devlink.h                         |   1 +
 include/net/fib_notifier.h                    |  12 +-
 include/net/fib_rules.h                       |   3 +-
 include/net/ip6_fib.h                         |  11 +-
 include/net/ip_fib.h                          |  11 +-
 include/uapi/linux/devlink.h                  |   4 +
 net/core/devlink.c                            | 158 +++++++-
 net/core/fib_notifier.c                       |  95 +++--
 net/core/fib_rules.c                          |  23 +-
 net/ipv4/fib_notifier.c                       |  13 +-
 net/ipv4/fib_rules.c                          |   5 +-
 net/ipv4/fib_trie.c                           |  44 ++-
 net/ipv4/ipmr.c                               |  13 +-
 net/ipv4/ipmr_base.c                          |  30 +-
 net/ipv6/fib6_notifier.c                      |  11 +-
 net/ipv6/fib6_rules.c                         |   5 +-
 net/ipv6/ip6_fib.c                            |  50 ++-
 net/ipv6/ip6mr.c                              |  13 +-
 .../drivers/net/netdevsim/devlink.sh          | 120 +++++-
 39 files changed, 848 insertions(+), 489 deletions(-)

-- 
2.21.0


^ permalink raw reply	[flat|nested] 34+ messages in thread

* [patch net-next 01/15] netdevsim: change fib accounting and limitations to be per-device
  2019-09-14  6:45 [patch net-next 00/15] devlink: allow devlink instances to change network namespace Jiri Pirko
@ 2019-09-14  6:45 ` Jiri Pirko
  2019-09-14  6:45 ` [patch net-next 02/15] net: fib_notifier: make FIB notifier per-netns Jiri Pirko
                   ` (16 subsequent siblings)
  17 siblings, 0 replies; 34+ messages in thread
From: Jiri Pirko @ 2019-09-14  6:45 UTC (permalink / raw)
  To: netdev
  Cc: davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

From: Jiri Pirko <jiri@mellanox.com>

Currently, the accounting is done per-namespace. However, devlink
instance is always in init_net namespace for now, so only the accounting
related to init_net is used. Limitations set using devlink resources
are only considered for init_net. nsim_devlink_net() always
returns init_net always.

Make the accounting per-device. This brings no functional change.
Per-device accounting has the same values as per-net.
For a single netdevsim instance, the behaviour is exactly the same
as before. When multiple netdevsim instances are created, each
can have different limits.

This is in prepare to implement proper devlink netns support. After
that, the devlink instance which would exist in particular netns would
account and limit that netns.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 drivers/net/netdevsim/dev.c       |  79 ++++------------
 drivers/net/netdevsim/fib.c       | 144 +++++++++++++++++++-----------
 drivers/net/netdevsim/netdev.c    |   9 +-
 drivers/net/netdevsim/netdevsim.h |  10 ++-
 4 files changed, 114 insertions(+), 128 deletions(-)

diff --git a/drivers/net/netdevsim/dev.c b/drivers/net/netdevsim/dev.c
index 7fba7b271a57..d623501de3ea 100644
--- a/drivers/net/netdevsim/dev.c
+++ b/drivers/net/netdevsim/dev.c
@@ -123,39 +123,6 @@ static void nsim_dev_port_debugfs_exit(struct nsim_dev_port *nsim_dev_port)
 	debugfs_remove_recursive(nsim_dev_port->ddir);
 }
 
-static struct net *nsim_devlink_net(struct devlink *devlink)
-{
-	return &init_net;
-}
-
-static u64 nsim_dev_ipv4_fib_resource_occ_get(void *priv)
-{
-	struct net *net = priv;
-
-	return nsim_fib_get_val(net, NSIM_RESOURCE_IPV4_FIB, false);
-}
-
-static u64 nsim_dev_ipv4_fib_rules_res_occ_get(void *priv)
-{
-	struct net *net = priv;
-
-	return nsim_fib_get_val(net, NSIM_RESOURCE_IPV4_FIB_RULES, false);
-}
-
-static u64 nsim_dev_ipv6_fib_resource_occ_get(void *priv)
-{
-	struct net *net = priv;
-
-	return nsim_fib_get_val(net, NSIM_RESOURCE_IPV6_FIB, false);
-}
-
-static u64 nsim_dev_ipv6_fib_rules_res_occ_get(void *priv)
-{
-	struct net *net = priv;
-
-	return nsim_fib_get_val(net, NSIM_RESOURCE_IPV6_FIB_RULES, false);
-}
-
 static int nsim_dev_resources_register(struct devlink *devlink)
 {
 	struct devlink_resource_size_params params = {
@@ -163,9 +130,7 @@ static int nsim_dev_resources_register(struct devlink *devlink)
 		.size_granularity = 1,
 		.unit = DEVLINK_RESOURCE_UNIT_ENTRY
 	};
-	struct net *net = nsim_devlink_net(devlink);
 	int err;
-	u64 n;
 
 	/* Resources for IPv4 */
 	err = devlink_resource_register(devlink, "IPv4", (u64)-1,
@@ -177,8 +142,7 @@ static int nsim_dev_resources_register(struct devlink *devlink)
 		goto out;
 	}
 
-	n = nsim_fib_get_val(net, NSIM_RESOURCE_IPV4_FIB, true);
-	err = devlink_resource_register(devlink, "fib", n,
+	err = devlink_resource_register(devlink, "fib", (u64)-1,
 					NSIM_RESOURCE_IPV4_FIB,
 					NSIM_RESOURCE_IPV4, &params);
 	if (err) {
@@ -186,8 +150,7 @@ static int nsim_dev_resources_register(struct devlink *devlink)
 		return err;
 	}
 
-	n = nsim_fib_get_val(net, NSIM_RESOURCE_IPV4_FIB_RULES, true);
-	err = devlink_resource_register(devlink, "fib-rules", n,
+	err = devlink_resource_register(devlink, "fib-rules", (u64)-1,
 					NSIM_RESOURCE_IPV4_FIB_RULES,
 					NSIM_RESOURCE_IPV4, &params);
 	if (err) {
@@ -205,8 +168,7 @@ static int nsim_dev_resources_register(struct devlink *devlink)
 		goto out;
 	}
 
-	n = nsim_fib_get_val(net, NSIM_RESOURCE_IPV6_FIB, true);
-	err = devlink_resource_register(devlink, "fib", n,
+	err = devlink_resource_register(devlink, "fib", (u64)-1,
 					NSIM_RESOURCE_IPV6_FIB,
 					NSIM_RESOURCE_IPV6, &params);
 	if (err) {
@@ -214,8 +176,7 @@ static int nsim_dev_resources_register(struct devlink *devlink)
 		return err;
 	}
 
-	n = nsim_fib_get_val(net, NSIM_RESOURCE_IPV6_FIB_RULES, true);
-	err = devlink_resource_register(devlink, "fib-rules", n,
+	err = devlink_resource_register(devlink, "fib-rules", (u64)-1,
 					NSIM_RESOURCE_IPV6_FIB_RULES,
 					NSIM_RESOURCE_IPV6, &params);
 	if (err) {
@@ -223,22 +184,6 @@ static int nsim_dev_resources_register(struct devlink *devlink)
 		return err;
 	}
 
-	devlink_resource_occ_get_register(devlink,
-					  NSIM_RESOURCE_IPV4_FIB,
-					  nsim_dev_ipv4_fib_resource_occ_get,
-					  net);
-	devlink_resource_occ_get_register(devlink,
-					  NSIM_RESOURCE_IPV4_FIB_RULES,
-					  nsim_dev_ipv4_fib_rules_res_occ_get,
-					  net);
-	devlink_resource_occ_get_register(devlink,
-					  NSIM_RESOURCE_IPV6_FIB,
-					  nsim_dev_ipv6_fib_resource_occ_get,
-					  net);
-	devlink_resource_occ_get_register(devlink,
-					  NSIM_RESOURCE_IPV6_FIB_RULES,
-					  nsim_dev_ipv6_fib_rules_res_occ_get,
-					  net);
 out:
 	return err;
 }
@@ -530,11 +475,11 @@ static int nsim_dev_reload_down(struct devlink *devlink,
 static int nsim_dev_reload_up(struct devlink *devlink,
 			      struct netlink_ext_ack *extack)
 {
+	struct nsim_dev *nsim_dev = devlink_priv(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
 	};
-	struct net *net = nsim_devlink_net(devlink);
 	int i;
 
 	for (i = 0; i < ARRAY_SIZE(res_ids); ++i) {
@@ -543,7 +488,8 @@ static int nsim_dev_reload_up(struct devlink *devlink,
 
 		err = devlink_resource_size_get(devlink, res_ids[i], &val);
 		if (!err) {
-			err = nsim_fib_set_max(net, res_ids[i], val, extack);
+			err = nsim_fib_set_max(nsim_dev->fib_data,
+					       res_ids[i], val, extack);
 			if (err)
 				return err;
 		}
@@ -678,9 +624,15 @@ nsim_dev_create(struct nsim_bus_dev *nsim_bus_dev, unsigned int port_count)
 	if (err)
 		goto err_devlink_free;
 
+	nsim_dev->fib_data = nsim_fib_create(devlink);
+	if (IS_ERR(nsim_dev->fib_data)) {
+		err = PTR_ERR(nsim_dev->fib_data);
+		goto err_resources_unregister;
+	}
+
 	err = devlink_register(devlink, &nsim_bus_dev->dev);
 	if (err)
-		goto err_resources_unregister;
+		goto err_fib_destroy;
 
 	err = devlink_params_register(devlink, nsim_devlink_params,
 				      ARRAY_SIZE(nsim_devlink_params));
@@ -718,6 +670,8 @@ nsim_dev_create(struct nsim_bus_dev *nsim_bus_dev, unsigned int port_count)
 				  ARRAY_SIZE(nsim_devlink_params));
 err_dl_unregister:
 	devlink_unregister(devlink);
+err_fib_destroy:
+	nsim_fib_destroy(devlink, nsim_dev->fib_data);
 err_resources_unregister:
 	devlink_resources_unregister(devlink, NULL);
 err_devlink_free:
@@ -736,6 +690,7 @@ static void nsim_dev_destroy(struct nsim_dev *nsim_dev)
 	devlink_params_unregister(devlink, nsim_devlink_params,
 				  ARRAY_SIZE(nsim_devlink_params));
 	devlink_unregister(devlink);
+	nsim_fib_destroy(devlink, nsim_dev->fib_data);
 	devlink_resources_unregister(devlink, NULL);
 	mutex_destroy(&nsim_dev->port_list_lock);
 	devlink_free(devlink);
diff --git a/drivers/net/netdevsim/fib.c b/drivers/net/netdevsim/fib.c
index f61d094746c0..7de17e42d77a 100644
--- a/drivers/net/netdevsim/fib.c
+++ b/drivers/net/netdevsim/fib.c
@@ -18,7 +18,7 @@
 #include <net/ip_fib.h>
 #include <net/ip6_fib.h>
 #include <net/fib_rules.h>
-#include <net/netns/generic.h>
+#include <net/net_namespace.h>
 
 #include "netdevsim.h"
 
@@ -33,15 +33,14 @@ struct nsim_per_fib_data {
 };
 
 struct nsim_fib_data {
+	struct notifier_block fib_nb;
 	struct nsim_per_fib_data ipv4;
 	struct nsim_per_fib_data ipv6;
 };
 
-static unsigned int nsim_fib_net_id;
-
-u64 nsim_fib_get_val(struct net *net, enum nsim_resource_id res_id, bool max)
+u64 nsim_fib_get_val(struct nsim_fib_data *fib_data,
+		     enum nsim_resource_id res_id, bool max)
 {
-	struct nsim_fib_data *fib_data = net_generic(net, nsim_fib_net_id);
 	struct nsim_fib_entry *entry;
 
 	switch (res_id) {
@@ -64,10 +63,10 @@ u64 nsim_fib_get_val(struct net *net, enum nsim_resource_id res_id, bool max)
 	return max ? entry->max : entry->num;
 }
 
-int nsim_fib_set_max(struct net *net, enum nsim_resource_id res_id, u64 val,
+int nsim_fib_set_max(struct nsim_fib_data *fib_data,
+		     enum nsim_resource_id res_id, u64 val,
 		     struct netlink_ext_ack *extack)
 {
-	struct nsim_fib_data *fib_data = net_generic(net, nsim_fib_net_id);
 	struct nsim_fib_entry *entry;
 	int err = 0;
 
@@ -120,9 +119,9 @@ static int nsim_fib_rule_account(struct nsim_fib_entry *entry, bool add,
 	return err;
 }
 
-static int nsim_fib_rule_event(struct fib_notifier_info *info, bool add)
+static int nsim_fib_rule_event(struct nsim_fib_data *data,
+			       struct fib_notifier_info *info, bool add)
 {
-	struct nsim_fib_data *data = net_generic(info->net, nsim_fib_net_id);
 	struct netlink_ext_ack *extack = info->extack;
 	int err = 0;
 
@@ -157,9 +156,9 @@ static int nsim_fib_account(struct nsim_fib_entry *entry, bool add,
 	return err;
 }
 
-static int nsim_fib_event(struct fib_notifier_info *info, bool add)
+static int nsim_fib_event(struct nsim_fib_data *data,
+			  struct fib_notifier_info *info, bool add)
 {
-	struct nsim_fib_data *data = net_generic(info->net, nsim_fib_net_id);
 	struct netlink_ext_ack *extack = info->extack;
 	int err = 0;
 
@@ -178,18 +177,25 @@ static int nsim_fib_event(struct fib_notifier_info *info, bool add)
 static int nsim_fib_event_nb(struct notifier_block *nb, unsigned long event,
 			     void *ptr)
 {
+	struct nsim_fib_data *data = container_of(nb, struct nsim_fib_data,
+						  fib_nb);
 	struct fib_notifier_info *info = ptr;
 	int err = 0;
 
+	if (!net_eq(info->net, &init_net))
+		return NOTIFY_DONE;
+
 	switch (event) {
 	case FIB_EVENT_RULE_ADD: /* fall through */
 	case FIB_EVENT_RULE_DEL:
-		err = nsim_fib_rule_event(info, event == FIB_EVENT_RULE_ADD);
+		err = nsim_fib_rule_event(data, info,
+					  event == FIB_EVENT_RULE_ADD);
 		break;
 
 	case FIB_EVENT_ENTRY_ADD:  /* fall through */
 	case FIB_EVENT_ENTRY_DEL:
-		err = nsim_fib_event(info, event == FIB_EVENT_ENTRY_ADD);
+		err = nsim_fib_event(data, info,
+				     event == FIB_EVENT_ENTRY_ADD);
 		break;
 	}
 
@@ -199,68 +205,98 @@ static int nsim_fib_event_nb(struct notifier_block *nb, unsigned long event,
 /* inconsistent dump, trying again */
 static void nsim_fib_dump_inconsistent(struct notifier_block *nb)
 {
-	struct nsim_fib_data *data;
-	struct net *net;
+	struct nsim_fib_data *data = container_of(nb, struct nsim_fib_data,
+						  fib_nb);
 
-	rcu_read_lock();
-	for_each_net_rcu(net) {
-		data = net_generic(net, nsim_fib_net_id);
+	data->ipv4.fib.num = 0ULL;
+	data->ipv4.rules.num = 0ULL;
+	data->ipv6.fib.num = 0ULL;
+	data->ipv6.rules.num = 0ULL;
+}
 
-		data->ipv4.fib.num = 0ULL;
-		data->ipv4.rules.num = 0ULL;
+static u64 nsim_fib_ipv4_resource_occ_get(void *priv)
+{
+	struct nsim_fib_data *data = priv;
 
-		data->ipv6.fib.num = 0ULL;
-		data->ipv6.rules.num = 0ULL;
-	}
-	rcu_read_unlock();
+	return nsim_fib_get_val(data, NSIM_RESOURCE_IPV4_FIB, false);
 }
 
-static struct notifier_block nsim_fib_nb = {
-	.notifier_call = nsim_fib_event_nb,
-};
-
-/* Initialize per network namespace state */
-static int __net_init nsim_fib_netns_init(struct net *net)
+static u64 nsim_fib_ipv4_rules_res_occ_get(void *priv)
 {
-	struct nsim_fib_data *data = net_generic(net, nsim_fib_net_id);
+	struct nsim_fib_data *data = priv;
 
-	data->ipv4.fib.max = (u64)-1;
-	data->ipv4.rules.max = (u64)-1;
+	return nsim_fib_get_val(data, NSIM_RESOURCE_IPV4_FIB_RULES, false);
+}
 
-	data->ipv6.fib.max = (u64)-1;
-	data->ipv6.rules.max = (u64)-1;
+static u64 nsim_fib_ipv6_resource_occ_get(void *priv)
+{
+	struct nsim_fib_data *data = priv;
 
-	return 0;
+	return nsim_fib_get_val(data, NSIM_RESOURCE_IPV6_FIB, false);
 }
 
-static struct pernet_operations nsim_fib_net_ops = {
-	.init = nsim_fib_netns_init,
-	.id   = &nsim_fib_net_id,
-	.size = sizeof(struct nsim_fib_data),
-};
-
-void nsim_fib_exit(void)
+static u64 nsim_fib_ipv6_rules_res_occ_get(void *priv)
 {
-	unregister_pernet_subsys(&nsim_fib_net_ops);
-	unregister_fib_notifier(&nsim_fib_nb);
+	struct nsim_fib_data *data = priv;
+
+	return nsim_fib_get_val(data, NSIM_RESOURCE_IPV6_FIB_RULES, false);
 }
 
-int nsim_fib_init(void)
+struct nsim_fib_data *nsim_fib_create(struct devlink *devlink)
 {
+	struct nsim_fib_data *data;
 	int err;
 
-	err = register_pernet_subsys(&nsim_fib_net_ops);
-	if (err < 0) {
-		pr_err("Failed to register pernet subsystem\n");
-		goto err_out;
-	}
+	data = kzalloc(sizeof(*data), GFP_KERNEL);
+	if (!data)
+		return ERR_PTR(-ENOMEM);
 
-	err = register_fib_notifier(&nsim_fib_nb, nsim_fib_dump_inconsistent);
-	if (err < 0) {
+	data->ipv4.fib.max = (u64)-1;
+	data->ipv4.rules.max = (u64)-1;
+
+	data->ipv6.fib.max = (u64)-1;
+	data->ipv6.rules.max = (u64)-1;
+
+	data->fib_nb.notifier_call = nsim_fib_event_nb;
+	err = register_fib_notifier(&data->fib_nb, nsim_fib_dump_inconsistent);
+	if (err) {
 		pr_err("Failed to register fib notifier\n");
 		goto err_out;
 	}
 
+	devlink_resource_occ_get_register(devlink,
+					  NSIM_RESOURCE_IPV4_FIB,
+					  nsim_fib_ipv4_resource_occ_get,
+					  data);
+	devlink_resource_occ_get_register(devlink,
+					  NSIM_RESOURCE_IPV4_FIB_RULES,
+					  nsim_fib_ipv4_rules_res_occ_get,
+					  data);
+	devlink_resource_occ_get_register(devlink,
+					  NSIM_RESOURCE_IPV6_FIB,
+					  nsim_fib_ipv6_resource_occ_get,
+					  data);
+	devlink_resource_occ_get_register(devlink,
+					  NSIM_RESOURCE_IPV6_FIB_RULES,
+					  nsim_fib_ipv6_rules_res_occ_get,
+					  data);
+	return data;
+
 err_out:
-	return err;
+	kfree(data);
+	return ERR_PTR(err);
+}
+
+void nsim_fib_destroy(struct devlink *devlink, struct nsim_fib_data *data)
+{
+	devlink_resource_occ_get_unregister(devlink,
+					    NSIM_RESOURCE_IPV6_FIB_RULES);
+	devlink_resource_occ_get_unregister(devlink,
+					    NSIM_RESOURCE_IPV6_FIB);
+	devlink_resource_occ_get_unregister(devlink,
+					    NSIM_RESOURCE_IPV4_FIB_RULES);
+	devlink_resource_occ_get_unregister(devlink,
+					    NSIM_RESOURCE_IPV4_FIB);
+	unregister_fib_notifier(&data->fib_nb);
+	kfree(data);
 }
diff --git a/drivers/net/netdevsim/netdev.c b/drivers/net/netdevsim/netdev.c
index 55f57f76d01b..0740940f41b1 100644
--- a/drivers/net/netdevsim/netdev.c
+++ b/drivers/net/netdevsim/netdev.c
@@ -357,18 +357,12 @@ static int __init nsim_module_init(void)
 	if (err)
 		goto err_dev_exit;
 
-	err = nsim_fib_init();
-	if (err)
-		goto err_bus_exit;
-
 	err = rtnl_link_register(&nsim_link_ops);
 	if (err)
-		goto err_fib_exit;
+		goto err_bus_exit;
 
 	return 0;
 
-err_fib_exit:
-	nsim_fib_exit();
 err_bus_exit:
 	nsim_bus_exit();
 err_dev_exit:
@@ -379,7 +373,6 @@ static int __init nsim_module_init(void)
 static void __exit nsim_module_exit(void)
 {
 	rtnl_link_unregister(&nsim_link_ops);
-	nsim_fib_exit();
 	nsim_bus_exit();
 	nsim_dev_exit();
 }
diff --git a/drivers/net/netdevsim/netdevsim.h b/drivers/net/netdevsim/netdevsim.h
index 66bf13765ad0..ac506cf253b6 100644
--- a/drivers/net/netdevsim/netdevsim.h
+++ b/drivers/net/netdevsim/netdevsim.h
@@ -173,10 +173,12 @@ int nsim_dev_port_add(struct nsim_bus_dev *nsim_bus_dev,
 int nsim_dev_port_del(struct nsim_bus_dev *nsim_bus_dev,
 		      unsigned int port_index);
 
-int nsim_fib_init(void);
-void nsim_fib_exit(void);
-u64 nsim_fib_get_val(struct net *net, enum nsim_resource_id res_id, bool max);
-int nsim_fib_set_max(struct net *net, enum nsim_resource_id res_id, u64 val,
+struct nsim_fib_data *nsim_fib_create(struct devlink *devlink);
+void nsim_fib_destroy(struct devlink *devlink, struct nsim_fib_data *data);
+u64 nsim_fib_get_val(struct nsim_fib_data *fib_data,
+		     enum nsim_resource_id res_id, bool max);
+int nsim_fib_set_max(struct nsim_fib_data *fib_data,
+		     enum nsim_resource_id res_id, u64 val,
 		     struct netlink_ext_ack *extack);
 
 #if IS_ENABLED(CONFIG_XFRM_OFFLOAD)
-- 
2.21.0


^ permalink raw reply related	[flat|nested] 34+ messages in thread

* [patch net-next 02/15] net: fib_notifier: make FIB notifier per-netns
  2019-09-14  6:45 [patch net-next 00/15] devlink: allow devlink instances to change network namespace Jiri Pirko
  2019-09-14  6:45 ` [patch net-next 01/15] netdevsim: change fib accounting and limitations to be per-device Jiri Pirko
@ 2019-09-14  6:45 ` Jiri Pirko
  2019-09-15  8:06   ` Ido Schimmel
  2019-09-15 20:05   ` David Ahern
  2019-09-14  6:45 ` [patch net-next 03/15] net: fib_notifier: propagate possible error during fib notifier registration Jiri Pirko
                   ` (15 subsequent siblings)
  17 siblings, 2 replies; 34+ messages in thread
From: Jiri Pirko @ 2019-09-14  6:45 UTC (permalink / raw)
  To: netdev
  Cc: davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

From: Jiri Pirko <jiri@mellanox.com>

Currently all users of FIB notifier only cares about events in init_net.
Later in this patchset, users get interested in other namespaces too.
However, for every registered block user is interested only about one
namespace. Make the FIB notifier registration per-netns and avoid
unnecessary calls of notifier block for other namespaces.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 .../net/ethernet/mellanox/mlx5/core/lag_mp.c  |  7 +-
 .../ethernet/mellanox/mlxsw/spectrum_router.c |  7 +-
 drivers/net/ethernet/rocker/rocker_main.c     |  9 +-
 drivers/net/netdevsim/fib.c                   |  8 +-
 include/linux/mroute_base.h                   |  8 +-
 include/net/fib_notifier.h                    |  6 +-
 include/net/ip6_fib.h                         |  2 +-
 include/net/ip_fib.h                          |  2 +-
 net/core/fib_notifier.c                       | 85 +++++++++----------
 net/core/fib_rules.c                          |  7 +-
 net/ipv4/fib_notifier.c                       |  4 +-
 net/ipv4/fib_trie.c                           | 17 ++--
 net/ipv4/ipmr_base.c                          |  4 +-
 net/ipv6/fib6_notifier.c                      |  4 +-
 net/ipv6/ip6_fib.c                            |  6 +-
 15 files changed, 77 insertions(+), 99 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c b/drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c
index 5d20d615663e..fe0cc969cf94 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c
@@ -248,9 +248,6 @@ static int mlx5_lag_fib_event(struct notifier_block *nb,
 	struct net_device *fib_dev;
 	struct fib_info *fi;
 
-	if (!net_eq(info->net, &init_net))
-		return NOTIFY_DONE;
-
 	if (info->family != AF_INET)
 		return NOTIFY_DONE;
 
@@ -311,7 +308,7 @@ int mlx5_lag_mp_init(struct mlx5_lag *ldev)
 		return 0;
 
 	mp->fib_nb.notifier_call = mlx5_lag_fib_event;
-	err = register_fib_notifier(&mp->fib_nb,
+	err = register_fib_notifier(&init_net, &mp->fib_nb,
 				    mlx5_lag_fib_event_flush);
 	if (err)
 		mp->fib_nb.notifier_call = NULL;
@@ -326,6 +323,6 @@ void mlx5_lag_mp_cleanup(struct mlx5_lag *ldev)
 	if (!mp->fib_nb.notifier_call)
 		return;
 
-	unregister_fib_notifier(&mp->fib_nb);
+	unregister_fib_notifier(&init_net, &mp->fib_nb);
 	mp->fib_nb.notifier_call = NULL;
 }
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index a330b369e899..6d78d2002335 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -6250,8 +6250,7 @@ static int mlxsw_sp_router_fib_event(struct notifier_block *nb,
 	struct mlxsw_sp_router *router;
 	int err;
 
-	if (!net_eq(info->net, &init_net) ||
-	    (info->family != AF_INET && info->family != AF_INET6 &&
+	if ((info->family != AF_INET && info->family != AF_INET6 &&
 	     info->family != RTNL_FAMILY_IPMR &&
 	     info->family != RTNL_FAMILY_IP6MR))
 		return NOTIFY_DONE;
@@ -8155,7 +8154,7 @@ int mlxsw_sp_router_init(struct mlxsw_sp *mlxsw_sp)
 		goto err_dscp_init;
 
 	mlxsw_sp->router->fib_nb.notifier_call = mlxsw_sp_router_fib_event;
-	err = register_fib_notifier(&mlxsw_sp->router->fib_nb,
+	err = register_fib_notifier(&init_net, &mlxsw_sp->router->fib_nb,
 				    mlxsw_sp_router_fib_dump_flush);
 	if (err)
 		goto err_register_fib_notifier;
@@ -8195,7 +8194,7 @@ int mlxsw_sp_router_init(struct mlxsw_sp *mlxsw_sp)
 
 void mlxsw_sp_router_fini(struct mlxsw_sp *mlxsw_sp)
 {
-	unregister_fib_notifier(&mlxsw_sp->router->fib_nb);
+	unregister_fib_notifier(&init_net, &mlxsw_sp->router->fib_nb);
 	unregister_netevent_notifier(&mlxsw_sp->router->netevent_nb);
 	mlxsw_sp_neigh_fini(mlxsw_sp);
 	mlxsw_sp_vrs_fini(mlxsw_sp);
diff --git a/drivers/net/ethernet/rocker/rocker_main.c b/drivers/net/ethernet/rocker/rocker_main.c
index 786b158bd305..e54f6341a785 100644
--- a/drivers/net/ethernet/rocker/rocker_main.c
+++ b/drivers/net/ethernet/rocker/rocker_main.c
@@ -2189,9 +2189,6 @@ static int rocker_router_fib_event(struct notifier_block *nb,
 	struct rocker_fib_event_work *fib_work;
 	struct fib_notifier_info *info = ptr;
 
-	if (!net_eq(info->net, &init_net))
-		return NOTIFY_DONE;
-
 	if (info->family != AF_INET)
 		return NOTIFY_DONE;
 
@@ -2994,7 +2991,7 @@ static int rocker_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	 * the device, so no need to pass a callback.
 	 */
 	rocker->fib_nb.notifier_call = rocker_router_fib_event;
-	err = register_fib_notifier(&rocker->fib_nb, NULL);
+	err = register_fib_notifier(&init_net, &rocker->fib_nb, NULL);
 	if (err)
 		goto err_register_fib_notifier;
 
@@ -3021,7 +3018,7 @@ static int rocker_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 err_register_switchdev_blocking_notifier:
 	unregister_switchdev_notifier(&rocker_switchdev_notifier);
 err_register_switchdev_notifier:
-	unregister_fib_notifier(&rocker->fib_nb);
+	unregister_fib_notifier(&init_net, &rocker->fib_nb);
 err_register_fib_notifier:
 	rocker_remove_ports(rocker);
 err_probe_ports:
@@ -3057,7 +3054,7 @@ static void rocker_remove(struct pci_dev *pdev)
 	unregister_switchdev_blocking_notifier(nb);
 
 	unregister_switchdev_notifier(&rocker_switchdev_notifier);
-	unregister_fib_notifier(&rocker->fib_nb);
+	unregister_fib_notifier(&init_net, &rocker->fib_nb);
 	rocker_remove_ports(rocker);
 	rocker_write32(rocker, CONTROL, ROCKER_CONTROL_RESET);
 	destroy_workqueue(rocker->rocker_owq);
diff --git a/drivers/net/netdevsim/fib.c b/drivers/net/netdevsim/fib.c
index 7de17e42d77a..01ee9cc54605 100644
--- a/drivers/net/netdevsim/fib.c
+++ b/drivers/net/netdevsim/fib.c
@@ -182,9 +182,6 @@ static int nsim_fib_event_nb(struct notifier_block *nb, unsigned long event,
 	struct fib_notifier_info *info = ptr;
 	int err = 0;
 
-	if (!net_eq(info->net, &init_net))
-		return NOTIFY_DONE;
-
 	switch (event) {
 	case FIB_EVENT_RULE_ADD: /* fall through */
 	case FIB_EVENT_RULE_DEL:
@@ -258,7 +255,8 @@ struct nsim_fib_data *nsim_fib_create(struct devlink *devlink)
 	data->ipv6.rules.max = (u64)-1;
 
 	data->fib_nb.notifier_call = nsim_fib_event_nb;
-	err = register_fib_notifier(&data->fib_nb, nsim_fib_dump_inconsistent);
+	err = register_fib_notifier(&init_net, &data->fib_nb,
+				    nsim_fib_dump_inconsistent);
 	if (err) {
 		pr_err("Failed to register fib notifier\n");
 		goto err_out;
@@ -297,6 +295,6 @@ void nsim_fib_destroy(struct devlink *devlink, struct nsim_fib_data *data)
 					    NSIM_RESOURCE_IPV4_FIB_RULES);
 	devlink_resource_occ_get_unregister(devlink,
 					    NSIM_RESOURCE_IPV4_FIB);
-	unregister_fib_notifier(&data->fib_nb);
+	unregister_fib_notifier(&init_net, &data->fib_nb);
 	kfree(data);
 }
diff --git a/include/linux/mroute_base.h b/include/linux/mroute_base.h
index 34de06b426ef..d68c2cd014c3 100644
--- a/include/linux/mroute_base.h
+++ b/include/linux/mroute_base.h
@@ -47,7 +47,6 @@ struct vif_entry_notifier_info {
 };
 
 static inline int mr_call_vif_notifier(struct notifier_block *nb,
-				       struct net *net,
 				       unsigned short family,
 				       enum fib_event_type event_type,
 				       struct vif_device *vif,
@@ -56,7 +55,6 @@ static inline int mr_call_vif_notifier(struct notifier_block *nb,
 	struct vif_entry_notifier_info info = {
 		.info = {
 			.family = family,
-			.net = net,
 		},
 		.dev = vif->dev,
 		.vif_index = vif_index,
@@ -64,7 +62,7 @@ static inline int mr_call_vif_notifier(struct notifier_block *nb,
 		.tb_id = tb_id,
 	};
 
-	return call_fib_notifier(nb, net, event_type, &info.info);
+	return call_fib_notifier(nb, event_type, &info.info);
 }
 
 static inline int mr_call_vif_notifiers(struct net *net,
@@ -173,7 +171,6 @@ struct mfc_entry_notifier_info {
 };
 
 static inline int mr_call_mfc_notifier(struct notifier_block *nb,
-				       struct net *net,
 				       unsigned short family,
 				       enum fib_event_type event_type,
 				       struct mr_mfc *mfc, u32 tb_id)
@@ -181,13 +178,12 @@ static inline int mr_call_mfc_notifier(struct notifier_block *nb,
 	struct mfc_entry_notifier_info info = {
 		.info = {
 			.family = family,
-			.net = net,
 		},
 		.mfc = mfc,
 		.tb_id = tb_id
 	};
 
-	return call_fib_notifier(nb, net, event_type, &info.info);
+	return call_fib_notifier(nb, event_type, &info.info);
 }
 
 static inline int mr_call_mfc_notifiers(struct net *net,
diff --git a/include/net/fib_notifier.h b/include/net/fib_notifier.h
index c49d7bfb5c30..54bb8565efc3 100644
--- a/include/net/fib_notifier.h
+++ b/include/net/fib_notifier.h
@@ -35,14 +35,14 @@ struct fib_notifier_ops {
 	struct rcu_head rcu;
 };
 
-int call_fib_notifier(struct notifier_block *nb, struct net *net,
+int call_fib_notifier(struct notifier_block *nb,
 		      enum fib_event_type event_type,
 		      struct fib_notifier_info *info);
 int call_fib_notifiers(struct net *net, enum fib_event_type event_type,
 		       struct fib_notifier_info *info);
-int register_fib_notifier(struct notifier_block *nb,
+int register_fib_notifier(struct net *net, struct notifier_block *nb,
 			  void (*cb)(struct notifier_block *nb));
-int unregister_fib_notifier(struct notifier_block *nb);
+int unregister_fib_notifier(struct net *net, struct notifier_block *nb);
 struct fib_notifier_ops *
 fib_notifier_ops_register(const struct fib_notifier_ops *tmpl, struct net *net);
 void fib_notifier_ops_unregister(struct fib_notifier_ops *ops);
diff --git a/include/net/ip6_fib.h b/include/net/ip6_fib.h
index 4b5656c71abc..14e9fca0e326 100644
--- a/include/net/ip6_fib.h
+++ b/include/net/ip6_fib.h
@@ -478,7 +478,7 @@ struct ipv6_route_iter {
 
 extern const struct seq_operations ipv6_route_seq_ops;
 
-int call_fib6_notifier(struct notifier_block *nb, struct net *net,
+int call_fib6_notifier(struct notifier_block *nb,
 		       enum fib_event_type event_type,
 		       struct fib_notifier_info *info);
 int call_fib6_notifiers(struct net *net, enum fib_event_type event_type,
diff --git a/include/net/ip_fib.h b/include/net/ip_fib.h
index 4c81846ccce8..4cec9ecaa95e 100644
--- a/include/net/ip_fib.h
+++ b/include/net/ip_fib.h
@@ -219,7 +219,7 @@ struct fib_nh_notifier_info {
 	struct fib_nh *fib_nh;
 };
 
-int call_fib4_notifier(struct notifier_block *nb, struct net *net,
+int call_fib4_notifier(struct notifier_block *nb,
 		       enum fib_event_type event_type,
 		       struct fib_notifier_info *info);
 int call_fib4_notifiers(struct net *net, enum fib_event_type event_type,
diff --git a/net/core/fib_notifier.c b/net/core/fib_notifier.c
index 470a606d5e8d..b965f3c0ec9a 100644
--- a/net/core/fib_notifier.c
+++ b/net/core/fib_notifier.c
@@ -12,17 +12,15 @@ static unsigned int fib_notifier_net_id;
 
 struct fib_notifier_net {
 	struct list_head fib_notifier_ops;
+	struct atomic_notifier_head fib_chain;
 };
 
-static ATOMIC_NOTIFIER_HEAD(fib_chain);
-
-int call_fib_notifier(struct notifier_block *nb, struct net *net,
+int call_fib_notifier(struct notifier_block *nb,
 		      enum fib_event_type event_type,
 		      struct fib_notifier_info *info)
 {
 	int err;
 
-	info->net = net;
 	err = nb->notifier_call(nb, event_type, info);
 	return notifier_to_errno(err);
 }
@@ -31,35 +29,29 @@ EXPORT_SYMBOL(call_fib_notifier);
 int call_fib_notifiers(struct net *net, enum fib_event_type event_type,
 		       struct fib_notifier_info *info)
 {
+	struct fib_notifier_net *fn_net = net_generic(net, fib_notifier_net_id);
 	int err;
 
-	info->net = net;
-	err = atomic_notifier_call_chain(&fib_chain, event_type, info);
+	err = atomic_notifier_call_chain(&fn_net->fib_chain, event_type, info);
 	return notifier_to_errno(err);
 }
 EXPORT_SYMBOL(call_fib_notifiers);
 
-static unsigned int fib_seq_sum(void)
+static unsigned int fib_seq_sum(struct net *net)
 {
-	struct fib_notifier_net *fn_net;
+	struct fib_notifier_net *fn_net = net_generic(net, fib_notifier_net_id);
 	struct fib_notifier_ops *ops;
 	unsigned int fib_seq = 0;
-	struct net *net;
 
 	rtnl_lock();
-	down_read(&net_rwsem);
-	for_each_net(net) {
-		fn_net = net_generic(net, fib_notifier_net_id);
-		rcu_read_lock();
-		list_for_each_entry_rcu(ops, &fn_net->fib_notifier_ops, list) {
-			if (!try_module_get(ops->owner))
-				continue;
-			fib_seq += ops->fib_seq_read(net);
-			module_put(ops->owner);
-		}
-		rcu_read_unlock();
+	rcu_read_lock();
+	list_for_each_entry_rcu(ops, &fn_net->fib_notifier_ops, list) {
+		if (!try_module_get(ops->owner))
+			continue;
+		fib_seq += ops->fib_seq_read(net);
+		module_put(ops->owner);
 	}
-	up_read(&net_rwsem);
+	rcu_read_unlock();
 	rtnl_unlock();
 
 	return fib_seq;
@@ -69,7 +61,9 @@ static int fib_net_dump(struct net *net, struct notifier_block *nb)
 {
 	struct fib_notifier_net *fn_net = net_generic(net, fib_notifier_net_id);
 	struct fib_notifier_ops *ops;
+	int err = 0;
 
+	rcu_read_lock();
 	list_for_each_entry_rcu(ops, &fn_net->fib_notifier_ops, list) {
 		int err;
 
@@ -78,59 +72,57 @@ static int fib_net_dump(struct net *net, struct notifier_block *nb)
 		err = ops->fib_dump(net, nb);
 		module_put(ops->owner);
 		if (err)
-			return err;
+			goto unlock;
 	}
 
-	return 0;
+unlock:
+	rcu_read_unlock();
+
+	return err;
 }
 
-static bool fib_dump_is_consistent(struct notifier_block *nb,
+static bool fib_dump_is_consistent(struct net *net, struct notifier_block *nb,
 				   void (*cb)(struct notifier_block *nb),
 				   unsigned int fib_seq)
 {
-	atomic_notifier_chain_register(&fib_chain, nb);
-	if (fib_seq == fib_seq_sum())
+	struct fib_notifier_net *fn_net = net_generic(net, fib_notifier_net_id);
+
+	atomic_notifier_chain_register(&fn_net->fib_chain, nb);
+	if (fib_seq == fib_seq_sum(net))
 		return true;
-	atomic_notifier_chain_unregister(&fib_chain, nb);
+	atomic_notifier_chain_unregister(&fn_net->fib_chain, nb);
 	if (cb)
 		cb(nb);
 	return false;
 }
 
 #define FIB_DUMP_MAX_RETRIES 5
-int register_fib_notifier(struct notifier_block *nb,
+int register_fib_notifier(struct net *net, struct notifier_block *nb,
 			  void (*cb)(struct notifier_block *nb))
 {
 	int retries = 0;
 	int err;
 
 	do {
-		unsigned int fib_seq = fib_seq_sum();
-		struct net *net;
-
-		rcu_read_lock();
-		for_each_net_rcu(net) {
-			err = fib_net_dump(net, nb);
-			if (err)
-				goto err_fib_net_dump;
-		}
-		rcu_read_unlock();
-
-		if (fib_dump_is_consistent(nb, cb, fib_seq))
+		unsigned int fib_seq = fib_seq_sum(net);
+
+		err = fib_net_dump(net, nb);
+		if (err)
+			return err;
+
+		if (fib_dump_is_consistent(net, nb, cb, fib_seq))
 			return 0;
 	} while (++retries < FIB_DUMP_MAX_RETRIES);
 
 	return -EBUSY;
-
-err_fib_net_dump:
-	rcu_read_unlock();
-	return err;
 }
 EXPORT_SYMBOL(register_fib_notifier);
 
-int unregister_fib_notifier(struct notifier_block *nb)
+int unregister_fib_notifier(struct net *net, struct notifier_block *nb)
 {
-	return atomic_notifier_chain_unregister(&fib_chain, nb);
+	struct fib_notifier_net *fn_net = net_generic(net, fib_notifier_net_id);
+
+	return atomic_notifier_chain_unregister(&fn_net->fib_chain, nb);
 }
 EXPORT_SYMBOL(unregister_fib_notifier);
 
@@ -181,6 +173,7 @@ static int __net_init fib_notifier_net_init(struct net *net)
 	struct fib_notifier_net *fn_net = net_generic(net, fib_notifier_net_id);
 
 	INIT_LIST_HEAD(&fn_net->fib_notifier_ops);
+	ATOMIC_INIT_NOTIFIER_HEAD(&fn_net->fib_chain);
 	return 0;
 }
 
diff --git a/net/core/fib_rules.c b/net/core/fib_rules.c
index dd220ce7ca7a..28cbf07102bc 100644
--- a/net/core/fib_rules.c
+++ b/net/core/fib_rules.c
@@ -321,7 +321,7 @@ int fib_rules_lookup(struct fib_rules_ops *ops, struct flowi *fl,
 }
 EXPORT_SYMBOL_GPL(fib_rules_lookup);
 
-static int call_fib_rule_notifier(struct notifier_block *nb, struct net *net,
+static int call_fib_rule_notifier(struct notifier_block *nb,
 				  enum fib_event_type event_type,
 				  struct fib_rule *rule, int family)
 {
@@ -330,7 +330,7 @@ static int call_fib_rule_notifier(struct notifier_block *nb, struct net *net,
 		.rule = rule,
 	};
 
-	return call_fib_notifier(nb, net, event_type, &info.info);
+	return call_fib_notifier(nb, event_type, &info.info);
 }
 
 static int call_fib_rule_notifiers(struct net *net,
@@ -359,8 +359,7 @@ int fib_rules_dump(struct net *net, struct notifier_block *nb, int family)
 	if (!ops)
 		return -EAFNOSUPPORT;
 	list_for_each_entry_rcu(rule, &ops->rules_list, list)
-		call_fib_rule_notifier(nb, net, FIB_EVENT_RULE_ADD, rule,
-				       family);
+		call_fib_rule_notifier(nb, FIB_EVENT_RULE_ADD, rule, family);
 	rules_ops_put(ops);
 
 	return 0;
diff --git a/net/ipv4/fib_notifier.c b/net/ipv4/fib_notifier.c
index b804ccbdb241..1a128c1346fb 100644
--- a/net/ipv4/fib_notifier.c
+++ b/net/ipv4/fib_notifier.c
@@ -9,12 +9,12 @@
 #include <net/netns/ipv4.h>
 #include <net/ip_fib.h>
 
-int call_fib4_notifier(struct notifier_block *nb, struct net *net,
+int call_fib4_notifier(struct notifier_block *nb,
 		       enum fib_event_type event_type,
 		       struct fib_notifier_info *info)
 {
 	info->family = AF_INET;
-	return call_fib_notifier(nb, net, event_type, info);
+	return call_fib_notifier(nb, event_type, info);
 }
 
 int call_fib4_notifiers(struct net *net, enum fib_event_type event_type,
diff --git a/net/ipv4/fib_trie.c b/net/ipv4/fib_trie.c
index 1ab2fb6bb37d..5b600b2a2aa3 100644
--- a/net/ipv4/fib_trie.c
+++ b/net/ipv4/fib_trie.c
@@ -74,7 +74,7 @@
 #include <trace/events/fib.h>
 #include "fib_lookup.h"
 
-static int call_fib_entry_notifier(struct notifier_block *nb, struct net *net,
+static int call_fib_entry_notifier(struct notifier_block *nb,
 				   enum fib_event_type event_type, u32 dst,
 				   int dst_len, struct fib_alias *fa)
 {
@@ -86,7 +86,7 @@ static int call_fib_entry_notifier(struct notifier_block *nb, struct net *net,
 		.type = fa->fa_type,
 		.tb_id = fa->tb_id,
 	};
-	return call_fib4_notifier(nb, net, event_type, &info.info);
+	return call_fib4_notifier(nb, event_type, &info.info);
 }
 
 static int call_fib_entry_notifiers(struct net *net,
@@ -2015,8 +2015,8 @@ void fib_info_notify_update(struct net *net, struct nl_info *info)
 	}
 }
 
-static void fib_leaf_notify(struct net *net, struct key_vector *l,
-			    struct fib_table *tb, struct notifier_block *nb)
+static void fib_leaf_notify(struct key_vector *l, struct fib_table *tb,
+			    struct notifier_block *nb)
 {
 	struct fib_alias *fa;
 
@@ -2032,20 +2032,19 @@ static void fib_leaf_notify(struct net *net, struct key_vector *l,
 		if (tb->tb_id != fa->tb_id)
 			continue;
 
-		call_fib_entry_notifier(nb, net, FIB_EVENT_ENTRY_ADD, l->key,
+		call_fib_entry_notifier(nb, FIB_EVENT_ENTRY_ADD, l->key,
 					KEYLENGTH - fa->fa_slen, fa);
 	}
 }
 
-static void fib_table_notify(struct net *net, struct fib_table *tb,
-			     struct notifier_block *nb)
+static void fib_table_notify(struct fib_table *tb, struct notifier_block *nb)
 {
 	struct trie *t = (struct trie *)tb->tb_data;
 	struct key_vector *l, *tp = t->kv;
 	t_key key = 0;
 
 	while ((l = leaf_walk_rcu(&tp, key)) != NULL) {
-		fib_leaf_notify(net, l, tb, nb);
+		fib_leaf_notify(l, tb, nb);
 
 		key = l->key + 1;
 		/* stop in case of wrap around */
@@ -2063,7 +2062,7 @@ void fib_notify(struct net *net, struct notifier_block *nb)
 		struct fib_table *tb;
 
 		hlist_for_each_entry_rcu(tb, head, tb_hlist)
-			fib_table_notify(net, tb, nb);
+			fib_table_notify(tb, nb);
 	}
 }
 
diff --git a/net/ipv4/ipmr_base.c b/net/ipv4/ipmr_base.c
index ea48bd15a575..4dcc3214e3cc 100644
--- a/net/ipv4/ipmr_base.c
+++ b/net/ipv4/ipmr_base.c
@@ -409,7 +409,7 @@ int mr_dump(struct net *net, struct notifier_block *nb, unsigned short family,
 			if (!v->dev)
 				continue;
 
-			mr_call_vif_notifier(nb, net, family,
+			mr_call_vif_notifier(nb, family,
 					     FIB_EVENT_VIF_ADD,
 					     v, vifi, mrt->id);
 		}
@@ -417,7 +417,7 @@ int mr_dump(struct net *net, struct notifier_block *nb, unsigned short family,
 
 		/* Notify on table MFC entries */
 		list_for_each_entry_rcu(mfc, &mrt->mfc_cache_list, list)
-			mr_call_mfc_notifier(nb, net, family,
+			mr_call_mfc_notifier(nb, family,
 					     FIB_EVENT_ENTRY_ADD,
 					     mfc, mrt->id);
 	}
diff --git a/net/ipv6/fib6_notifier.c b/net/ipv6/fib6_notifier.c
index 05f82baaa99e..4fe79296999a 100644
--- a/net/ipv6/fib6_notifier.c
+++ b/net/ipv6/fib6_notifier.c
@@ -7,12 +7,12 @@
 #include <net/netns/ipv6.h>
 #include <net/ip6_fib.h>
 
-int call_fib6_notifier(struct notifier_block *nb, struct net *net,
+int call_fib6_notifier(struct notifier_block *nb,
 		       enum fib_event_type event_type,
 		       struct fib_notifier_info *info)
 {
 	info->family = AF_INET6;
-	return call_fib_notifier(nb, net, event_type, info);
+	return call_fib_notifier(nb, event_type, info);
 }
 
 int call_fib6_notifiers(struct net *net, enum fib_event_type event_type,
diff --git a/net/ipv6/ip6_fib.c b/net/ipv6/ip6_fib.c
index 87f47bc55c5e..a6d500023457 100644
--- a/net/ipv6/ip6_fib.c
+++ b/net/ipv6/ip6_fib.c
@@ -357,7 +357,7 @@ unsigned int fib6_tables_seq_read(struct net *net)
 	return fib_seq;
 }
 
-static int call_fib6_entry_notifier(struct notifier_block *nb, struct net *net,
+static int call_fib6_entry_notifier(struct notifier_block *nb,
 				    enum fib_event_type event_type,
 				    struct fib6_info *rt)
 {
@@ -365,7 +365,7 @@ static int call_fib6_entry_notifier(struct notifier_block *nb, struct net *net,
 		.rt = rt,
 	};
 
-	return call_fib6_notifier(nb, net, event_type, &info.info);
+	return call_fib6_notifier(nb, event_type, &info.info);
 }
 
 int call_fib6_entry_notifiers(struct net *net,
@@ -407,7 +407,7 @@ static void fib6_rt_dump(struct fib6_info *rt, struct fib6_dump_arg *arg)
 {
 	if (rt == arg->net->ipv6.fib6_null_entry)
 		return;
-	call_fib6_entry_notifier(arg->nb, arg->net, FIB_EVENT_ENTRY_ADD, rt);
+	call_fib6_entry_notifier(arg->nb, FIB_EVENT_ENTRY_ADD, rt);
 }
 
 static int fib6_node_dump(struct fib6_walker *w)
-- 
2.21.0


^ permalink raw reply related	[flat|nested] 34+ messages in thread

* [patch net-next 03/15] net: fib_notifier: propagate possible error during fib notifier registration
  2019-09-14  6:45 [patch net-next 00/15] devlink: allow devlink instances to change network namespace Jiri Pirko
  2019-09-14  6:45 ` [patch net-next 01/15] netdevsim: change fib accounting and limitations to be per-device Jiri Pirko
  2019-09-14  6:45 ` [patch net-next 02/15] net: fib_notifier: make FIB notifier per-netns Jiri Pirko
@ 2019-09-14  6:45 ` Jiri Pirko
  2019-09-15  8:17   ` Ido Schimmel
  2019-09-14  6:45 ` [patch net-next 04/15] mlxsw: spectrum_router: Don't rely on missing extack to symbolize dump Jiri Pirko
                   ` (14 subsequent siblings)
  17 siblings, 1 reply; 34+ messages in thread
From: Jiri Pirko @ 2019-09-14  6:45 UTC (permalink / raw)
  To: netdev
  Cc: davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

From: Jiri Pirko <jiri@mellanox.com>

Unlike events for registered notifier, during the registration, the
errors that happened for the block being registered are not propagated
up to the caller. For fib rules, this is already present, but not for
fib entries. So make sure the error is propagated for those as well.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 include/net/ip_fib.h    |  2 +-
 net/core/fib_notifier.c |  2 --
 net/core/fib_rules.c    | 11 ++++++++---
 net/ipv4/fib_notifier.c |  4 +---
 net/ipv4/fib_trie.c     | 31 ++++++++++++++++++++++---------
 net/ipv4/ipmr_base.c    | 22 +++++++++++++++-------
 net/ipv6/ip6_fib.c      | 36 ++++++++++++++++++++++++------------
 7 files changed, 71 insertions(+), 37 deletions(-)

diff --git a/include/net/ip_fib.h b/include/net/ip_fib.h
index 4cec9ecaa95e..caae0fa610aa 100644
--- a/include/net/ip_fib.h
+++ b/include/net/ip_fib.h
@@ -229,7 +229,7 @@ int __net_init fib4_notifier_init(struct net *net);
 void __net_exit fib4_notifier_exit(struct net *net);
 
 void fib_info_notify_update(struct net *net, struct nl_info *info);
-void fib_notify(struct net *net, struct notifier_block *nb);
+int fib_notify(struct net *net, struct notifier_block *nb);
 
 struct fib_table {
 	struct hlist_node	tb_hlist;
diff --git a/net/core/fib_notifier.c b/net/core/fib_notifier.c
index b965f3c0ec9a..fbd029425638 100644
--- a/net/core/fib_notifier.c
+++ b/net/core/fib_notifier.c
@@ -65,8 +65,6 @@ static int fib_net_dump(struct net *net, struct notifier_block *nb)
 
 	rcu_read_lock();
 	list_for_each_entry_rcu(ops, &fn_net->fib_notifier_ops, list) {
-		int err;
-
 		if (!try_module_get(ops->owner))
 			continue;
 		err = ops->fib_dump(net, nb);
diff --git a/net/core/fib_rules.c b/net/core/fib_rules.c
index 28cbf07102bc..592d8aef90e3 100644
--- a/net/core/fib_rules.c
+++ b/net/core/fib_rules.c
@@ -354,15 +354,20 @@ int fib_rules_dump(struct net *net, struct notifier_block *nb, int family)
 {
 	struct fib_rules_ops *ops;
 	struct fib_rule *rule;
+	int err = 0;
 
 	ops = lookup_rules_ops(net, family);
 	if (!ops)
 		return -EAFNOSUPPORT;
-	list_for_each_entry_rcu(rule, &ops->rules_list, list)
-		call_fib_rule_notifier(nb, FIB_EVENT_RULE_ADD, rule, family);
+	list_for_each_entry_rcu(rule, &ops->rules_list, list) {
+		err = call_fib_rule_notifier(nb, FIB_EVENT_RULE_ADD,
+					     rule, family);
+		if (err)
+			break;
+	}
 	rules_ops_put(ops);
 
-	return 0;
+	return err;
 }
 EXPORT_SYMBOL_GPL(fib_rules_dump);
 
diff --git a/net/ipv4/fib_notifier.c b/net/ipv4/fib_notifier.c
index 1a128c1346fb..0c57f68a9340 100644
--- a/net/ipv4/fib_notifier.c
+++ b/net/ipv4/fib_notifier.c
@@ -42,9 +42,7 @@ static int fib4_dump(struct net *net, struct notifier_block *nb)
 	if (err)
 		return err;
 
-	fib_notify(net, nb);
-
-	return 0;
+	return fib_notify(net, nb);
 }
 
 static const struct fib_notifier_ops fib4_notifier_ops_template = {
diff --git a/net/ipv4/fib_trie.c b/net/ipv4/fib_trie.c
index 5b600b2a2aa3..568e59423773 100644
--- a/net/ipv4/fib_trie.c
+++ b/net/ipv4/fib_trie.c
@@ -2015,10 +2015,11 @@ void fib_info_notify_update(struct net *net, struct nl_info *info)
 	}
 }
 
-static void fib_leaf_notify(struct key_vector *l, struct fib_table *tb,
-			    struct notifier_block *nb)
+static int fib_leaf_notify(struct key_vector *l, struct fib_table *tb,
+			   struct notifier_block *nb)
 {
 	struct fib_alias *fa;
+	int err;
 
 	hlist_for_each_entry_rcu(fa, &l->leaf, fa_list) {
 		struct fib_info *fi = fa->fa_info;
@@ -2032,38 +2033,50 @@ static void fib_leaf_notify(struct key_vector *l, struct fib_table *tb,
 		if (tb->tb_id != fa->tb_id)
 			continue;
 
-		call_fib_entry_notifier(nb, FIB_EVENT_ENTRY_ADD, l->key,
-					KEYLENGTH - fa->fa_slen, fa);
+		err = call_fib_entry_notifier(nb, FIB_EVENT_ENTRY_ADD, l->key,
+					      KEYLENGTH - fa->fa_slen, fa);
+		if (err)
+			return err;
 	}
+	return 0;
 }
 
-static void fib_table_notify(struct fib_table *tb, struct notifier_block *nb)
+static int fib_table_notify(struct fib_table *tb, struct notifier_block *nb)
 {
 	struct trie *t = (struct trie *)tb->tb_data;
 	struct key_vector *l, *tp = t->kv;
 	t_key key = 0;
+	int err;
 
 	while ((l = leaf_walk_rcu(&tp, key)) != NULL) {
-		fib_leaf_notify(l, tb, nb);
+		err = fib_leaf_notify(l, tb, nb);
+		if (err)
+			return err;
 
 		key = l->key + 1;
 		/* stop in case of wrap around */
 		if (key < l->key)
 			break;
 	}
+	return 0;
 }
 
-void fib_notify(struct net *net, struct notifier_block *nb)
+int fib_notify(struct net *net, struct notifier_block *nb)
 {
 	unsigned int h;
+	int err;
 
 	for (h = 0; h < FIB_TABLE_HASHSZ; h++) {
 		struct hlist_head *head = &net->ipv4.fib_table_hash[h];
 		struct fib_table *tb;
 
-		hlist_for_each_entry_rcu(tb, head, tb_hlist)
-			fib_table_notify(tb, nb);
+		hlist_for_each_entry_rcu(tb, head, tb_hlist) {
+			err = fib_table_notify(tb, nb);
+			if (err)
+				return err;
+		}
 	}
+	return 0;
 }
 
 static void __trie_free_rcu(struct rcu_head *head)
diff --git a/net/ipv4/ipmr_base.c b/net/ipv4/ipmr_base.c
index 4dcc3214e3cc..c4e23c2a0d5c 100644
--- a/net/ipv4/ipmr_base.c
+++ b/net/ipv4/ipmr_base.c
@@ -409,17 +409,25 @@ int mr_dump(struct net *net, struct notifier_block *nb, unsigned short family,
 			if (!v->dev)
 				continue;
 
-			mr_call_vif_notifier(nb, family,
-					     FIB_EVENT_VIF_ADD,
-					     v, vifi, mrt->id);
+			err = mr_call_vif_notifier(nb, family,
+						   FIB_EVENT_VIF_ADD,
+						   v, vifi, mrt->id);
+			if (err)
+				break;
 		}
 		read_unlock(mrt_lock);
 
+		if (err)
+			return err;
+
 		/* Notify on table MFC entries */
-		list_for_each_entry_rcu(mfc, &mrt->mfc_cache_list, list)
-			mr_call_mfc_notifier(nb, family,
-					     FIB_EVENT_ENTRY_ADD,
-					     mfc, mrt->id);
+		list_for_each_entry_rcu(mfc, &mrt->mfc_cache_list, list) {
+			err = mr_call_mfc_notifier(nb, family,
+						   FIB_EVENT_ENTRY_ADD,
+						   mfc, mrt->id);
+			if (err)
+				return err;
+		}
 	}
 
 	return 0;
diff --git a/net/ipv6/ip6_fib.c b/net/ipv6/ip6_fib.c
index a6d500023457..4acca152843a 100644
--- a/net/ipv6/ip6_fib.c
+++ b/net/ipv6/ip6_fib.c
@@ -403,30 +403,37 @@ struct fib6_dump_arg {
 	struct notifier_block *nb;
 };
 
-static void fib6_rt_dump(struct fib6_info *rt, struct fib6_dump_arg *arg)
+static int fib6_rt_dump(struct fib6_info *rt, struct fib6_dump_arg *arg)
 {
 	if (rt == arg->net->ipv6.fib6_null_entry)
-		return;
-	call_fib6_entry_notifier(arg->nb, FIB_EVENT_ENTRY_ADD, rt);
+		return 0;
+	return call_fib6_entry_notifier(arg->nb, FIB_EVENT_ENTRY_ADD, rt);
 }
 
 static int fib6_node_dump(struct fib6_walker *w)
 {
 	struct fib6_info *rt;
+	int err = 0;
 
-	for_each_fib6_walker_rt(w)
-		fib6_rt_dump(rt, w->args);
+	for_each_fib6_walker_rt(w) {
+		err = fib6_rt_dump(rt, w->args);
+		if (err)
+			break;
+	}
 	w->leaf = NULL;
-	return 0;
+	return err;
 }
 
-static void fib6_table_dump(struct net *net, struct fib6_table *tb,
-			    struct fib6_walker *w)
+static int fib6_table_dump(struct net *net, struct fib6_table *tb,
+			   struct fib6_walker *w)
 {
+	int err;
+
 	w->root = &tb->tb6_root;
 	spin_lock_bh(&tb->tb6_lock);
-	fib6_walk(net, w);
+	err = fib6_walk(net, w);
 	spin_unlock_bh(&tb->tb6_lock);
+	return err;
 }
 
 /* Called with rcu_read_lock() */
@@ -435,6 +442,7 @@ int fib6_tables_dump(struct net *net, struct notifier_block *nb)
 	struct fib6_dump_arg arg;
 	struct fib6_walker *w;
 	unsigned int h;
+	int err = 0;
 
 	w = kzalloc(sizeof(*w), GFP_ATOMIC);
 	if (!w)
@@ -449,13 +457,17 @@ int fib6_tables_dump(struct net *net, struct notifier_block *nb)
 		struct hlist_head *head = &net->ipv6.fib_table_hash[h];
 		struct fib6_table *tb;
 
-		hlist_for_each_entry_rcu(tb, head, tb6_hlist)
-			fib6_table_dump(net, tb, w);
+		hlist_for_each_entry_rcu(tb, head, tb6_hlist) {
+			err = fib6_table_dump(net, tb, w);
+			if (err < 0)
+				goto out;
+		}
 	}
 
+out:
 	kfree(w);
 
-	return 0;
+	return err;
 }
 
 static int fib6_dump_node(struct fib6_walker *w)
-- 
2.21.0


^ permalink raw reply related	[flat|nested] 34+ messages in thread

* [patch net-next 04/15] mlxsw: spectrum_router: Don't rely on missing extack to symbolize dump
  2019-09-14  6:45 [patch net-next 00/15] devlink: allow devlink instances to change network namespace Jiri Pirko
                   ` (2 preceding siblings ...)
  2019-09-14  6:45 ` [patch net-next 03/15] net: fib_notifier: propagate possible error during fib notifier registration Jiri Pirko
@ 2019-09-14  6:45 ` Jiri Pirko
  2019-09-14  6:45 ` [patch net-next 05/15] net: fib_notifier: propagate extack down to the notifier block callback Jiri Pirko
                   ` (13 subsequent siblings)
  17 siblings, 0 replies; 34+ messages in thread
From: Jiri Pirko @ 2019-09-14  6:45 UTC (permalink / raw)
  To: netdev
  Cc: davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

From: Jiri Pirko <jiri@mellanox.com>

Currently if info->extack is NULL, mlxsw assumes that the event came
down from dump. Originally, the dump did not propagate the return value
back to the original caller (fib_notifier_register()). However, that is
now happening. So benefit from this and push the error up if it happened.
Remove rule cases in work handlers that are now dead code.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 .../ethernet/mellanox/mlxsw/spectrum_router.c | 22 +------------------
 1 file changed, 1 insertion(+), 21 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 6d78d2002335..43dbe25674f7 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -6019,12 +6019,6 @@ static void mlxsw_sp_router_fib4_event_work(struct work_struct *work)
 		mlxsw_sp_router_fib4_del(mlxsw_sp, &fib_work->fen_info);
 		fib_info_put(fib_work->fen_info.fi);
 		break;
-	case FIB_EVENT_RULE_ADD:
-		/* if we get here, a rule was added that we do not support.
-		 * just do the fib_abort
-		 */
-		mlxsw_sp_router_fib_abort(mlxsw_sp);
-		break;
 	case FIB_EVENT_NH_ADD: /* fall through */
 	case FIB_EVENT_NH_DEL:
 		mlxsw_sp_nexthop4_event(mlxsw_sp, fib_work->event,
@@ -6065,12 +6059,6 @@ static void mlxsw_sp_router_fib6_event_work(struct work_struct *work)
 					 fib_work->fib6_work.nrt6);
 		mlxsw_sp_router_fib6_work_fini(&fib_work->fib6_work);
 		break;
-	case FIB_EVENT_RULE_ADD:
-		/* if we get here, a rule was added that we do not support.
-		 * just do the fib_abort
-		 */
-		mlxsw_sp_router_fib_abort(mlxsw_sp);
-		break;
 	}
 	rtnl_unlock();
 	kfree(fib_work);
@@ -6112,12 +6100,6 @@ static void mlxsw_sp_router_fibmr_event_work(struct work_struct *work)
 					      &fib_work->ven_info);
 		dev_put(fib_work->ven_info.dev);
 		break;
-	case FIB_EVENT_RULE_ADD:
-		/* if we get here, a rule was added that we do not support.
-		 * just do the fib_abort
-		 */
-		mlxsw_sp_router_fib_abort(mlxsw_sp);
-		break;
 	}
 	rtnl_unlock();
 	kfree(fib_work);
@@ -6262,9 +6244,7 @@ static int mlxsw_sp_router_fib_event(struct notifier_block *nb,
 	case FIB_EVENT_RULE_DEL:
 		err = mlxsw_sp_router_fib_rule_event(event, info,
 						     router->mlxsw_sp);
-		if (!err || info->extack)
-			return notifier_from_errno(err);
-		break;
+		return notifier_from_errno(err);
 	case FIB_EVENT_ENTRY_ADD:
 	case FIB_EVENT_ENTRY_REPLACE: /* fall through */
 	case FIB_EVENT_ENTRY_APPEND:  /* fall through */
-- 
2.21.0


^ permalink raw reply related	[flat|nested] 34+ messages in thread

* [patch net-next 05/15] net: fib_notifier: propagate extack down to the notifier block callback
  2019-09-14  6:45 [patch net-next 00/15] devlink: allow devlink instances to change network namespace Jiri Pirko
                   ` (3 preceding siblings ...)
  2019-09-14  6:45 ` [patch net-next 04/15] mlxsw: spectrum_router: Don't rely on missing extack to symbolize dump Jiri Pirko
@ 2019-09-14  6:45 ` Jiri Pirko
  2019-09-14  6:45 ` [patch net-next 06/15] net: devlink: export devlink net getter Jiri Pirko
                   ` (12 subsequent siblings)
  17 siblings, 0 replies; 34+ messages in thread
From: Jiri Pirko @ 2019-09-14  6:45 UTC (permalink / raw)
  To: netdev
  Cc: davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

From: Jiri Pirko <jiri@mellanox.com>

Since errors are propagated all the way up to the caller, propagate
possible extack of the caller all the way down to the notifier block
callback.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 .../net/ethernet/mellanox/mlx5/core/lag_mp.c  |  2 +-
 .../ethernet/mellanox/mlxsw/spectrum_router.c |  2 +-
 drivers/net/ethernet/rocker/rocker_main.c     |  2 +-
 drivers/net/netdevsim/fib.c                   |  2 +-
 include/linux/mroute_base.h                   | 18 +++++++++++------
 include/net/fib_notifier.h                    |  6 ++++--
 include/net/fib_rules.h                       |  3 ++-
 include/net/ip6_fib.h                         |  9 ++++++---
 include/net/ip_fib.h                          |  9 ++++++---
 net/core/fib_notifier.c                       | 10 ++++++----
 net/core/fib_rules.c                          |  9 ++++++---
 net/ipv4/fib_notifier.c                       |  7 ++++---
 net/ipv4/fib_rules.c                          |  5 +++--
 net/ipv4/fib_trie.c                           | 20 ++++++++++++-------
 net/ipv4/ipmr.c                               | 13 +++++++-----
 net/ipv4/ipmr_base.c                          | 12 ++++++-----
 net/ipv6/fib6_notifier.c                      |  7 ++++---
 net/ipv6/fib6_rules.c                         |  5 +++--
 net/ipv6/ip6_fib.c                            | 12 ++++++++---
 net/ipv6/ip6mr.c                              | 13 +++++++-----
 20 files changed, 105 insertions(+), 61 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c b/drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c
index fe0cc969cf94..13e2944b1274 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c
@@ -309,7 +309,7 @@ int mlx5_lag_mp_init(struct mlx5_lag *ldev)
 
 	mp->fib_nb.notifier_call = mlx5_lag_fib_event;
 	err = register_fib_notifier(&init_net, &mp->fib_nb,
-				    mlx5_lag_fib_event_flush);
+				    mlx5_lag_fib_event_flush, NULL);
 	if (err)
 		mp->fib_nb.notifier_call = NULL;
 
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 43dbe25674f7..5d2cdb9d7d16 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -8135,7 +8135,7 @@ int mlxsw_sp_router_init(struct mlxsw_sp *mlxsw_sp)
 
 	mlxsw_sp->router->fib_nb.notifier_call = mlxsw_sp_router_fib_event;
 	err = register_fib_notifier(&init_net, &mlxsw_sp->router->fib_nb,
-				    mlxsw_sp_router_fib_dump_flush);
+				    mlxsw_sp_router_fib_dump_flush, NULL);
 	if (err)
 		goto err_register_fib_notifier;
 
diff --git a/drivers/net/ethernet/rocker/rocker_main.c b/drivers/net/ethernet/rocker/rocker_main.c
index e54f6341a785..bc4f951315da 100644
--- a/drivers/net/ethernet/rocker/rocker_main.c
+++ b/drivers/net/ethernet/rocker/rocker_main.c
@@ -2991,7 +2991,7 @@ static int rocker_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	 * the device, so no need to pass a callback.
 	 */
 	rocker->fib_nb.notifier_call = rocker_router_fib_event;
-	err = register_fib_notifier(&init_net, &rocker->fib_nb, NULL);
+	err = register_fib_notifier(&init_net, &rocker->fib_nb, NULL, NULL);
 	if (err)
 		goto err_register_fib_notifier;
 
diff --git a/drivers/net/netdevsim/fib.c b/drivers/net/netdevsim/fib.c
index 01ee9cc54605..d2aeac0f4c2c 100644
--- a/drivers/net/netdevsim/fib.c
+++ b/drivers/net/netdevsim/fib.c
@@ -256,7 +256,7 @@ struct nsim_fib_data *nsim_fib_create(struct devlink *devlink)
 
 	data->fib_nb.notifier_call = nsim_fib_event_nb;
 	err = register_fib_notifier(&init_net, &data->fib_nb,
-				    nsim_fib_dump_inconsistent);
+				    nsim_fib_dump_inconsistent, NULL);
 	if (err) {
 		pr_err("Failed to register fib notifier\n");
 		goto err_out;
diff --git a/include/linux/mroute_base.h b/include/linux/mroute_base.h
index d68c2cd014c3..0010d98635ec 100644
--- a/include/linux/mroute_base.h
+++ b/include/linux/mroute_base.h
@@ -50,11 +50,13 @@ static inline int mr_call_vif_notifier(struct notifier_block *nb,
 				       unsigned short family,
 				       enum fib_event_type event_type,
 				       struct vif_device *vif,
-				       unsigned short vif_index, u32 tb_id)
+				       unsigned short vif_index, u32 tb_id,
+				       struct netlink_ext_ack *extack)
 {
 	struct vif_entry_notifier_info info = {
 		.info = {
 			.family = family,
+			.extack = extack,
 		},
 		.dev = vif->dev,
 		.vif_index = vif_index,
@@ -173,11 +175,13 @@ struct mfc_entry_notifier_info {
 static inline int mr_call_mfc_notifier(struct notifier_block *nb,
 				       unsigned short family,
 				       enum fib_event_type event_type,
-				       struct mr_mfc *mfc, u32 tb_id)
+				       struct mr_mfc *mfc, u32 tb_id,
+				       struct netlink_ext_ack *extack)
 {
 	struct mfc_entry_notifier_info info = {
 		.info = {
 			.family = family,
+			.extack = extack,
 		},
 		.mfc = mfc,
 		.tb_id = tb_id
@@ -297,10 +301,11 @@ int mr_rtm_dumproute(struct sk_buff *skb, struct netlink_callback *cb,
 
 int mr_dump(struct net *net, struct notifier_block *nb, unsigned short family,
 	    int (*rules_dump)(struct net *net,
-			      struct notifier_block *nb),
+			      struct notifier_block *nb,
+			      struct netlink_ext_ack *extack),
 	    struct mr_table *(*mr_iter)(struct net *net,
 					struct mr_table *mrt),
-	    rwlock_t *mrt_lock);
+	    rwlock_t *mrt_lock, struct netlink_ext_ack *extack);
 #else
 static inline void vif_device_init(struct vif_device *v,
 				   struct net_device *dev,
@@ -351,10 +356,11 @@ mr_rtm_dumproute(struct sk_buff *skb, struct netlink_callback *cb,
 static inline int mr_dump(struct net *net, struct notifier_block *nb,
 			  unsigned short family,
 			  int (*rules_dump)(struct net *net,
-					    struct notifier_block *nb),
+					    struct notifier_block *nb,
+					    struct netlink_ext_ack *extack),
 			  struct mr_table *(*mr_iter)(struct net *net,
 						      struct mr_table *mrt),
-			  rwlock_t *mrt_lock)
+			  rwlock_t *mrt_lock, struct netlink_ext_ack *extack)
 {
 	return -EINVAL;
 }
diff --git a/include/net/fib_notifier.h b/include/net/fib_notifier.h
index 54bb8565efc3..d78930c05772 100644
--- a/include/net/fib_notifier.h
+++ b/include/net/fib_notifier.h
@@ -30,7 +30,8 @@ struct fib_notifier_ops {
 	int family;
 	struct list_head list;
 	unsigned int (*fib_seq_read)(struct net *net);
-	int (*fib_dump)(struct net *net, struct notifier_block *nb);
+	int (*fib_dump)(struct net *net, struct notifier_block *nb,
+			struct netlink_ext_ack *extack);
 	struct module *owner;
 	struct rcu_head rcu;
 };
@@ -41,7 +42,8 @@ int call_fib_notifier(struct notifier_block *nb,
 int call_fib_notifiers(struct net *net, enum fib_event_type event_type,
 		       struct fib_notifier_info *info);
 int register_fib_notifier(struct net *net, struct notifier_block *nb,
-			  void (*cb)(struct notifier_block *nb));
+			  void (*cb)(struct notifier_block *nb),
+			  struct netlink_ext_ack *extack);
 int unregister_fib_notifier(struct net *net, struct notifier_block *nb);
 struct fib_notifier_ops *
 fib_notifier_ops_register(const struct fib_notifier_ops *tmpl, struct net *net);
diff --git a/include/net/fib_rules.h b/include/net/fib_rules.h
index 20dcadd8eed9..54e227e6b06a 100644
--- a/include/net/fib_rules.h
+++ b/include/net/fib_rules.h
@@ -194,7 +194,8 @@ int fib_rules_lookup(struct fib_rules_ops *, struct flowi *, int flags,
 int fib_default_rule_add(struct fib_rules_ops *, u32 pref, u32 table,
 			 u32 flags);
 bool fib_rule_matchall(const struct fib_rule *rule);
-int fib_rules_dump(struct net *net, struct notifier_block *nb, int family);
+int fib_rules_dump(struct net *net, struct notifier_block *nb, int family,
+		   struct netlink_ext_ack *extack);
 unsigned int fib_rules_seq_read(struct net *net, int family);
 
 int fib_nl_newrule(struct sk_buff *skb, struct nlmsghdr *nlh,
diff --git a/include/net/ip6_fib.h b/include/net/ip6_fib.h
index 14e9fca0e326..5d1615463138 100644
--- a/include/net/ip6_fib.h
+++ b/include/net/ip6_fib.h
@@ -488,7 +488,8 @@ int __net_init fib6_notifier_init(struct net *net);
 void __net_exit fib6_notifier_exit(struct net *net);
 
 unsigned int fib6_tables_seq_read(struct net *net);
-int fib6_tables_dump(struct net *net, struct notifier_block *nb);
+int fib6_tables_dump(struct net *net, struct notifier_block *nb,
+		     struct netlink_ext_ack *extack);
 
 void fib6_update_sernum(struct net *net, struct fib6_info *rt);
 void fib6_update_sernum_upto_root(struct net *net, struct fib6_info *rt);
@@ -504,7 +505,8 @@ static inline bool fib6_metric_locked(struct fib6_info *f6i, int metric)
 int fib6_rules_init(void);
 void fib6_rules_cleanup(void);
 bool fib6_rule_default(const struct fib_rule *rule);
-int fib6_rules_dump(struct net *net, struct notifier_block *nb);
+int fib6_rules_dump(struct net *net, struct notifier_block *nb,
+		    struct netlink_ext_ack *extack);
 unsigned int fib6_rules_seq_read(struct net *net);
 
 static inline bool fib6_rules_early_flow_dissect(struct net *net,
@@ -537,7 +539,8 @@ static inline bool fib6_rule_default(const struct fib_rule *rule)
 {
 	return true;
 }
-static inline int fib6_rules_dump(struct net *net, struct notifier_block *nb)
+static inline int fib6_rules_dump(struct net *net, struct notifier_block *nb,
+				  struct netlink_ext_ack *extack)
 {
 	return 0;
 }
diff --git a/include/net/ip_fib.h b/include/net/ip_fib.h
index caae0fa610aa..55b3f830ff99 100644
--- a/include/net/ip_fib.h
+++ b/include/net/ip_fib.h
@@ -229,7 +229,8 @@ int __net_init fib4_notifier_init(struct net *net);
 void __net_exit fib4_notifier_exit(struct net *net);
 
 void fib_info_notify_update(struct net *net, struct nl_info *info);
-int fib_notify(struct net *net, struct notifier_block *nb);
+int fib_notify(struct net *net, struct notifier_block *nb,
+	       struct netlink_ext_ack *extack);
 
 struct fib_table {
 	struct hlist_node	tb_hlist;
@@ -315,7 +316,8 @@ static inline bool fib4_rule_default(const struct fib_rule *rule)
 	return true;
 }
 
-static inline int fib4_rules_dump(struct net *net, struct notifier_block *nb)
+static inline int fib4_rules_dump(struct net *net, struct notifier_block *nb,
+				  struct netlink_ext_ack *extack)
 {
 	return 0;
 }
@@ -377,7 +379,8 @@ static inline int fib_lookup(struct net *net, struct flowi4 *flp,
 }
 
 bool fib4_rule_default(const struct fib_rule *rule);
-int fib4_rules_dump(struct net *net, struct notifier_block *nb);
+int fib4_rules_dump(struct net *net, struct notifier_block *nb,
+		    struct netlink_ext_ack *extack);
 unsigned int fib4_rules_seq_read(struct net *net);
 
 static inline bool fib4_rules_early_flow_dissect(struct net *net,
diff --git a/net/core/fib_notifier.c b/net/core/fib_notifier.c
index fbd029425638..fc96259807b6 100644
--- a/net/core/fib_notifier.c
+++ b/net/core/fib_notifier.c
@@ -57,7 +57,8 @@ static unsigned int fib_seq_sum(struct net *net)
 	return fib_seq;
 }
 
-static int fib_net_dump(struct net *net, struct notifier_block *nb)
+static int fib_net_dump(struct net *net, struct notifier_block *nb,
+			struct netlink_ext_ack *extack)
 {
 	struct fib_notifier_net *fn_net = net_generic(net, fib_notifier_net_id);
 	struct fib_notifier_ops *ops;
@@ -67,7 +68,7 @@ static int fib_net_dump(struct net *net, struct notifier_block *nb)
 	list_for_each_entry_rcu(ops, &fn_net->fib_notifier_ops, list) {
 		if (!try_module_get(ops->owner))
 			continue;
-		err = ops->fib_dump(net, nb);
+		err = ops->fib_dump(net, nb, extack);
 		module_put(ops->owner);
 		if (err)
 			goto unlock;
@@ -96,7 +97,8 @@ static bool fib_dump_is_consistent(struct net *net, struct notifier_block *nb,
 
 #define FIB_DUMP_MAX_RETRIES 5
 int register_fib_notifier(struct net *net, struct notifier_block *nb,
-			  void (*cb)(struct notifier_block *nb))
+			  void (*cb)(struct notifier_block *nb),
+			  struct netlink_ext_ack *extack)
 {
 	int retries = 0;
 	int err;
@@ -104,7 +106,7 @@ int register_fib_notifier(struct net *net, struct notifier_block *nb,
 	do {
 		unsigned int fib_seq = fib_seq_sum(net);
 
-		err = fib_net_dump(net, nb);
+		err = fib_net_dump(net, nb, extack);
 		if (err)
 			return err;
 
diff --git a/net/core/fib_rules.c b/net/core/fib_rules.c
index 592d8aef90e3..3e7e15278c46 100644
--- a/net/core/fib_rules.c
+++ b/net/core/fib_rules.c
@@ -323,10 +323,12 @@ EXPORT_SYMBOL_GPL(fib_rules_lookup);
 
 static int call_fib_rule_notifier(struct notifier_block *nb,
 				  enum fib_event_type event_type,
-				  struct fib_rule *rule, int family)
+				  struct fib_rule *rule, int family,
+				  struct netlink_ext_ack *extack)
 {
 	struct fib_rule_notifier_info info = {
 		.info.family = family,
+		.info.extack = extack,
 		.rule = rule,
 	};
 
@@ -350,7 +352,8 @@ static int call_fib_rule_notifiers(struct net *net,
 }
 
 /* Called with rcu_read_lock() */
-int fib_rules_dump(struct net *net, struct notifier_block *nb, int family)
+int fib_rules_dump(struct net *net, struct notifier_block *nb, int family,
+		   struct netlink_ext_ack *extack)
 {
 	struct fib_rules_ops *ops;
 	struct fib_rule *rule;
@@ -361,7 +364,7 @@ int fib_rules_dump(struct net *net, struct notifier_block *nb, int family)
 		return -EAFNOSUPPORT;
 	list_for_each_entry_rcu(rule, &ops->rules_list, list) {
 		err = call_fib_rule_notifier(nb, FIB_EVENT_RULE_ADD,
-					     rule, family);
+					     rule, family, extack);
 		if (err)
 			break;
 	}
diff --git a/net/ipv4/fib_notifier.c b/net/ipv4/fib_notifier.c
index 0c57f68a9340..0c28bd469a68 100644
--- a/net/ipv4/fib_notifier.c
+++ b/net/ipv4/fib_notifier.c
@@ -34,15 +34,16 @@ static unsigned int fib4_seq_read(struct net *net)
 	return net->ipv4.fib_seq + fib4_rules_seq_read(net);
 }
 
-static int fib4_dump(struct net *net, struct notifier_block *nb)
+static int fib4_dump(struct net *net, struct notifier_block *nb,
+		     struct netlink_ext_ack *extack)
 {
 	int err;
 
-	err = fib4_rules_dump(net, nb);
+	err = fib4_rules_dump(net, nb, extack);
 	if (err)
 		return err;
 
-	return fib_notify(net, nb);
+	return fib_notify(net, nb, extack);
 }
 
 static const struct fib_notifier_ops fib4_notifier_ops_template = {
diff --git a/net/ipv4/fib_rules.c b/net/ipv4/fib_rules.c
index b43a7ba5c6a4..f99e3bac5cab 100644
--- a/net/ipv4/fib_rules.c
+++ b/net/ipv4/fib_rules.c
@@ -65,9 +65,10 @@ bool fib4_rule_default(const struct fib_rule *rule)
 }
 EXPORT_SYMBOL_GPL(fib4_rule_default);
 
-int fib4_rules_dump(struct net *net, struct notifier_block *nb)
+int fib4_rules_dump(struct net *net, struct notifier_block *nb,
+		    struct netlink_ext_ack *extack)
 {
-	return fib_rules_dump(net, nb, AF_INET);
+	return fib_rules_dump(net, nb, AF_INET, extack);
 }
 
 unsigned int fib4_rules_seq_read(struct net *net)
diff --git a/net/ipv4/fib_trie.c b/net/ipv4/fib_trie.c
index 568e59423773..b9df9c09b84e 100644
--- a/net/ipv4/fib_trie.c
+++ b/net/ipv4/fib_trie.c
@@ -76,9 +76,11 @@
 
 static int call_fib_entry_notifier(struct notifier_block *nb,
 				   enum fib_event_type event_type, u32 dst,
-				   int dst_len, struct fib_alias *fa)
+				   int dst_len, struct fib_alias *fa,
+				   struct netlink_ext_ack *extack)
 {
 	struct fib_entry_notifier_info info = {
+		.info.extack = extack,
 		.dst = dst,
 		.dst_len = dst_len,
 		.fi = fa->fa_info,
@@ -2016,7 +2018,8 @@ void fib_info_notify_update(struct net *net, struct nl_info *info)
 }
 
 static int fib_leaf_notify(struct key_vector *l, struct fib_table *tb,
-			   struct notifier_block *nb)
+			   struct notifier_block *nb,
+			   struct netlink_ext_ack *extack)
 {
 	struct fib_alias *fa;
 	int err;
@@ -2034,14 +2037,16 @@ static int fib_leaf_notify(struct key_vector *l, struct fib_table *tb,
 			continue;
 
 		err = call_fib_entry_notifier(nb, FIB_EVENT_ENTRY_ADD, l->key,
-					      KEYLENGTH - fa->fa_slen, fa);
+					      KEYLENGTH - fa->fa_slen,
+					      fa, extack);
 		if (err)
 			return err;
 	}
 	return 0;
 }
 
-static int fib_table_notify(struct fib_table *tb, struct notifier_block *nb)
+static int fib_table_notify(struct fib_table *tb, struct notifier_block *nb,
+			    struct netlink_ext_ack *extack)
 {
 	struct trie *t = (struct trie *)tb->tb_data;
 	struct key_vector *l, *tp = t->kv;
@@ -2049,7 +2054,7 @@ static int fib_table_notify(struct fib_table *tb, struct notifier_block *nb)
 	int err;
 
 	while ((l = leaf_walk_rcu(&tp, key)) != NULL) {
-		err = fib_leaf_notify(l, tb, nb);
+		err = fib_leaf_notify(l, tb, nb, extack);
 		if (err)
 			return err;
 
@@ -2061,7 +2066,8 @@ static int fib_table_notify(struct fib_table *tb, struct notifier_block *nb)
 	return 0;
 }
 
-int fib_notify(struct net *net, struct notifier_block *nb)
+int fib_notify(struct net *net, struct notifier_block *nb,
+	       struct netlink_ext_ack *extack)
 {
 	unsigned int h;
 	int err;
@@ -2071,7 +2077,7 @@ int fib_notify(struct net *net, struct notifier_block *nb)
 		struct fib_table *tb;
 
 		hlist_for_each_entry_rcu(tb, head, tb_hlist) {
-			err = fib_table_notify(tb, nb);
+			err = fib_table_notify(tb, nb, extack);
 			if (err)
 				return err;
 		}
diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c
index 313470f6bb14..051f365b64d2 100644
--- a/net/ipv4/ipmr.c
+++ b/net/ipv4/ipmr.c
@@ -278,9 +278,10 @@ static void __net_exit ipmr_rules_exit(struct net *net)
 	rtnl_unlock();
 }
 
-static int ipmr_rules_dump(struct net *net, struct notifier_block *nb)
+static int ipmr_rules_dump(struct net *net, struct notifier_block *nb,
+			   struct netlink_ext_ack *extack)
 {
-	return fib_rules_dump(net, nb, RTNL_FAMILY_IPMR);
+	return fib_rules_dump(net, nb, RTNL_FAMILY_IPMR, extack);
 }
 
 static unsigned int ipmr_rules_seq_read(struct net *net)
@@ -336,7 +337,8 @@ static void __net_exit ipmr_rules_exit(struct net *net)
 	rtnl_unlock();
 }
 
-static int ipmr_rules_dump(struct net *net, struct notifier_block *nb)
+static int ipmr_rules_dump(struct net *net, struct notifier_block *nb,
+			   struct netlink_ext_ack *extack)
 {
 	return 0;
 }
@@ -3040,10 +3042,11 @@ static unsigned int ipmr_seq_read(struct net *net)
 	return net->ipv4.ipmr_seq + ipmr_rules_seq_read(net);
 }
 
-static int ipmr_dump(struct net *net, struct notifier_block *nb)
+static int ipmr_dump(struct net *net, struct notifier_block *nb,
+		     struct netlink_ext_ack *extack)
 {
 	return mr_dump(net, nb, RTNL_FAMILY_IPMR, ipmr_rules_dump,
-		       ipmr_mr_table_iter, &mrt_lock);
+		       ipmr_mr_table_iter, &mrt_lock, extack);
 }
 
 static const struct fib_notifier_ops ipmr_notifier_ops_template = {
diff --git a/net/ipv4/ipmr_base.c b/net/ipv4/ipmr_base.c
index c4e23c2a0d5c..aa8738a91210 100644
--- a/net/ipv4/ipmr_base.c
+++ b/net/ipv4/ipmr_base.c
@@ -386,15 +386,17 @@ EXPORT_SYMBOL(mr_rtm_dumproute);
 
 int mr_dump(struct net *net, struct notifier_block *nb, unsigned short family,
 	    int (*rules_dump)(struct net *net,
-			      struct notifier_block *nb),
+			      struct notifier_block *nb,
+			      struct netlink_ext_ack *extack),
 	    struct mr_table *(*mr_iter)(struct net *net,
 					struct mr_table *mrt),
-	    rwlock_t *mrt_lock)
+	    rwlock_t *mrt_lock,
+	    struct netlink_ext_ack *extack)
 {
 	struct mr_table *mrt;
 	int err;
 
-	err = rules_dump(net, nb);
+	err = rules_dump(net, nb, extack);
 	if (err)
 		return err;
 
@@ -411,7 +413,7 @@ int mr_dump(struct net *net, struct notifier_block *nb, unsigned short family,
 
 			err = mr_call_vif_notifier(nb, family,
 						   FIB_EVENT_VIF_ADD,
-						   v, vifi, mrt->id);
+						   v, vifi, mrt->id, extack);
 			if (err)
 				break;
 		}
@@ -424,7 +426,7 @@ int mr_dump(struct net *net, struct notifier_block *nb, unsigned short family,
 		list_for_each_entry_rcu(mfc, &mrt->mfc_cache_list, list) {
 			err = mr_call_mfc_notifier(nb, family,
 						   FIB_EVENT_ENTRY_ADD,
-						   mfc, mrt->id);
+						   mfc, mrt->id, extack);
 			if (err)
 				return err;
 		}
diff --git a/net/ipv6/fib6_notifier.c b/net/ipv6/fib6_notifier.c
index 4fe79296999a..f87ae33e1d01 100644
--- a/net/ipv6/fib6_notifier.c
+++ b/net/ipv6/fib6_notifier.c
@@ -27,15 +27,16 @@ static unsigned int fib6_seq_read(struct net *net)
 	return fib6_tables_seq_read(net) + fib6_rules_seq_read(net);
 }
 
-static int fib6_dump(struct net *net, struct notifier_block *nb)
+static int fib6_dump(struct net *net, struct notifier_block *nb,
+		     struct netlink_ext_ack *extack)
 {
 	int err;
 
-	err = fib6_rules_dump(net, nb);
+	err = fib6_rules_dump(net, nb, extack);
 	if (err)
 		return err;
 
-	return fib6_tables_dump(net, nb);
+	return fib6_tables_dump(net, nb, extack);
 }
 
 static const struct fib_notifier_ops fib6_notifier_ops_template = {
diff --git a/net/ipv6/fib6_rules.c b/net/ipv6/fib6_rules.c
index d22b6c140f23..04748de03b95 100644
--- a/net/ipv6/fib6_rules.c
+++ b/net/ipv6/fib6_rules.c
@@ -47,9 +47,10 @@ bool fib6_rule_default(const struct fib_rule *rule)
 }
 EXPORT_SYMBOL_GPL(fib6_rule_default);
 
-int fib6_rules_dump(struct net *net, struct notifier_block *nb)
+int fib6_rules_dump(struct net *net, struct notifier_block *nb,
+		    struct netlink_ext_ack *extack)
 {
-	return fib_rules_dump(net, nb, AF_INET6);
+	return fib_rules_dump(net, nb, AF_INET6, extack);
 }
 
 unsigned int fib6_rules_seq_read(struct net *net)
diff --git a/net/ipv6/ip6_fib.c b/net/ipv6/ip6_fib.c
index 4acca152843a..ecb1dc23af03 100644
--- a/net/ipv6/ip6_fib.c
+++ b/net/ipv6/ip6_fib.c
@@ -359,9 +359,11 @@ unsigned int fib6_tables_seq_read(struct net *net)
 
 static int call_fib6_entry_notifier(struct notifier_block *nb,
 				    enum fib_event_type event_type,
-				    struct fib6_info *rt)
+				    struct fib6_info *rt,
+				    struct netlink_ext_ack *extack)
 {
 	struct fib6_entry_notifier_info info = {
+		.info.extack = extack,
 		.rt = rt,
 	};
 
@@ -401,13 +403,15 @@ int call_fib6_multipath_entry_notifiers(struct net *net,
 struct fib6_dump_arg {
 	struct net *net;
 	struct notifier_block *nb;
+	struct netlink_ext_ack *extack;
 };
 
 static int fib6_rt_dump(struct fib6_info *rt, struct fib6_dump_arg *arg)
 {
 	if (rt == arg->net->ipv6.fib6_null_entry)
 		return 0;
-	return call_fib6_entry_notifier(arg->nb, FIB_EVENT_ENTRY_ADD, rt);
+	return call_fib6_entry_notifier(arg->nb, FIB_EVENT_ENTRY_ADD,
+					rt, arg->extack);
 }
 
 static int fib6_node_dump(struct fib6_walker *w)
@@ -437,7 +441,8 @@ static int fib6_table_dump(struct net *net, struct fib6_table *tb,
 }
 
 /* Called with rcu_read_lock() */
-int fib6_tables_dump(struct net *net, struct notifier_block *nb)
+int fib6_tables_dump(struct net *net, struct notifier_block *nb,
+		     struct netlink_ext_ack *extack)
 {
 	struct fib6_dump_arg arg;
 	struct fib6_walker *w;
@@ -451,6 +456,7 @@ int fib6_tables_dump(struct net *net, struct notifier_block *nb)
 	w->func = fib6_node_dump;
 	arg.net = net;
 	arg.nb = nb;
+	arg.extack = extack;
 	w->args = &arg;
 
 	for (h = 0; h < FIB6_TABLE_HASHSZ; h++) {
diff --git a/net/ipv6/ip6mr.c b/net/ipv6/ip6mr.c
index 857a89ad4d6c..bfa49ff70531 100644
--- a/net/ipv6/ip6mr.c
+++ b/net/ipv6/ip6mr.c
@@ -265,9 +265,10 @@ static void __net_exit ip6mr_rules_exit(struct net *net)
 	rtnl_unlock();
 }
 
-static int ip6mr_rules_dump(struct net *net, struct notifier_block *nb)
+static int ip6mr_rules_dump(struct net *net, struct notifier_block *nb,
+			    struct netlink_ext_ack *extack)
 {
-	return fib_rules_dump(net, nb, RTNL_FAMILY_IP6MR);
+	return fib_rules_dump(net, nb, RTNL_FAMILY_IP6MR, extack);
 }
 
 static unsigned int ip6mr_rules_seq_read(struct net *net)
@@ -324,7 +325,8 @@ static void __net_exit ip6mr_rules_exit(struct net *net)
 	rtnl_unlock();
 }
 
-static int ip6mr_rules_dump(struct net *net, struct notifier_block *nb)
+static int ip6mr_rules_dump(struct net *net, struct notifier_block *nb,
+			    struct netlink_ext_ack *extack)
 {
 	return 0;
 }
@@ -1256,10 +1258,11 @@ static unsigned int ip6mr_seq_read(struct net *net)
 	return net->ipv6.ipmr_seq + ip6mr_rules_seq_read(net);
 }
 
-static int ip6mr_dump(struct net *net, struct notifier_block *nb)
+static int ip6mr_dump(struct net *net, struct notifier_block *nb,
+		      struct netlink_ext_ack *extack)
 {
 	return mr_dump(net, nb, RTNL_FAMILY_IP6MR, ip6mr_rules_dump,
-		       ip6mr_mr_table_iter, &mrt_lock);
+		       ip6mr_mr_table_iter, &mrt_lock, extack);
 }
 
 static struct notifier_block ip6_mr_notifier = {
-- 
2.21.0


^ permalink raw reply related	[flat|nested] 34+ messages in thread

* [patch net-next 06/15] net: devlink: export devlink net getter
  2019-09-14  6:45 [patch net-next 00/15] devlink: allow devlink instances to change network namespace Jiri Pirko
                   ` (4 preceding siblings ...)
  2019-09-14  6:45 ` [patch net-next 05/15] net: fib_notifier: propagate extack down to the notifier block callback Jiri Pirko
@ 2019-09-14  6:45 ` Jiri Pirko
  2019-09-14  6:46 ` [patch net-next 07/15] mlxsw: spectrum: Take devlink net instead of init_net Jiri Pirko
                   ` (11 subsequent siblings)
  17 siblings, 0 replies; 34+ messages in thread
From: Jiri Pirko @ 2019-09-14  6:45 UTC (permalink / raw)
  To: netdev
  Cc: davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

From: Jiri Pirko <jiri@mellanox.com>

Allow drivers to get net struct for devlink instance.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 include/net/devlink.h | 1 +
 net/core/devlink.c    | 3 ++-
 2 files changed, 3 insertions(+), 1 deletion(-)

diff --git a/include/net/devlink.h b/include/net/devlink.h
index 23e4b65ec9df..5ac2be0f0857 100644
--- a/include/net/devlink.h
+++ b/include/net/devlink.h
@@ -771,6 +771,7 @@ static inline struct devlink *netdev_to_devlink(struct net_device *dev)
 
 struct ib_device;
 
+struct net *devlink_net(const struct devlink *devlink);
 struct devlink *devlink_alloc(const struct devlink_ops *ops, size_t priv_size);
 int devlink_register(struct devlink *devlink, struct device *dev);
 void devlink_unregister(struct devlink *devlink);
diff --git a/net/core/devlink.c b/net/core/devlink.c
index e48680efe54a..362cbbcca225 100644
--- a/net/core/devlink.c
+++ b/net/core/devlink.c
@@ -95,10 +95,11 @@ static LIST_HEAD(devlink_list);
  */
 static DEFINE_MUTEX(devlink_mutex);
 
-static struct net *devlink_net(const struct devlink *devlink)
+struct net *devlink_net(const struct devlink *devlink)
 {
 	return read_pnet(&devlink->_net);
 }
+EXPORT_SYMBOL_GPL(devlink_net);
 
 static void devlink_net_set(struct devlink *devlink, struct net *net)
 {
-- 
2.21.0


^ permalink raw reply related	[flat|nested] 34+ messages in thread

* [patch net-next 07/15] mlxsw: spectrum: Take devlink net instead of init_net
  2019-09-14  6:45 [patch net-next 00/15] devlink: allow devlink instances to change network namespace Jiri Pirko
                   ` (5 preceding siblings ...)
  2019-09-14  6:45 ` [patch net-next 06/15] net: devlink: export devlink net getter Jiri Pirko
@ 2019-09-14  6:46 ` Jiri Pirko
  2019-09-14  6:46 ` [patch net-next 08/15] mlxsw: Register port netdevices into net of core Jiri Pirko
                   ` (10 subsequent siblings)
  17 siblings, 0 replies; 34+ messages in thread
From: Jiri Pirko @ 2019-09-14  6:46 UTC (permalink / raw)
  To: netdev
  Cc: davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

From: Jiri Pirko <jiri@mellanox.com>

Follow-up patch is going to allow to reload devlink instance into
different network namespace, so use devlink_net() helper instead
of init_net.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlxsw/core.h    |  6 +++++
 .../net/ethernet/mellanox/mlxsw/spectrum.h    |  6 +++++
 .../ethernet/mellanox/mlxsw/spectrum_nve.c    |  2 +-
 .../ethernet/mellanox/mlxsw/spectrum_router.c | 27 +++++++++++--------
 .../mellanox/mlxsw/spectrum_switchdev.c       |  2 +-
 5 files changed, 30 insertions(+), 13 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/core.h b/drivers/net/ethernet/mellanox/mlxsw/core.h
index b65a17d49e43..693b3c5ab355 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/core.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/core.h
@@ -11,6 +11,7 @@
 #include <linux/types.h>
 #include <linux/skbuff.h>
 #include <linux/workqueue.h>
+#include <linux/net_namespace.h>
 #include <net/devlink.h>
 
 #include "trap.h"
@@ -345,6 +346,11 @@ u64 mlxsw_core_res_get(struct mlxsw_core *mlxsw_core,
 #define MLXSW_CORE_RES_GET(mlxsw_core, short_res_id)			\
 	mlxsw_core_res_get(mlxsw_core, MLXSW_RES_ID_##short_res_id)
 
+static inline struct net *mlxsw_core_net(struct mlxsw_core *mlxsw_core)
+{
+	return devlink_net(priv_to_devlink(mlxsw_core));
+}
+
 #define MLXSW_BUS_F_TXRX	BIT(0)
 #define MLXSW_BUS_F_RESET	BIT(1)
 
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
index b2a0028b1694..f58d45e770cd 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
@@ -14,6 +14,7 @@
 #include <linux/dcbnl.h>
 #include <linux/in6.h>
 #include <linux/notifier.h>
+#include <linux/net_namespace.h>
 #include <net/psample.h>
 #include <net/pkt_cls.h>
 #include <net/red.h>
@@ -982,4 +983,9 @@ int mlxsw_sp_trap_action_set(struct mlxsw_core *mlxsw_core,
 int mlxsw_sp_trap_group_init(struct mlxsw_core *mlxsw_core,
 			     const struct devlink_trap_group *group);
 
+static inline struct net *mlxsw_sp_net(struct mlxsw_sp *mlxsw_sp)
+{
+	return mlxsw_core_net(mlxsw_sp->core);
+}
+
 #endif
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c
index 17f334b46c40..2153bcc4b585 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c
@@ -870,7 +870,7 @@ void mlxsw_sp_nve_fid_disable(struct mlxsw_sp *mlxsw_sp,
 		    mlxsw_sp_fid_vni(fid, &vni)))
 		goto out;
 
-	nve_dev = dev_get_by_index(&init_net, nve_ifindex);
+	nve_dev = dev_get_by_index(mlxsw_sp_net(mlxsw_sp), nve_ifindex);
 	if (!nve_dev)
 		goto out;
 
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 5d2cdb9d7d16..a1c06889178c 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -16,6 +16,7 @@
 #include <linux/if_macvlan.h>
 #include <linux/refcount.h>
 #include <linux/jhash.h>
+#include <linux/net_namespace.h>
 #include <net/netevent.h>
 #include <net/neighbour.h>
 #include <net/arp.h>
@@ -2551,14 +2552,14 @@ static int mlxsw_sp_router_schedule_work(struct net *net,
 	struct mlxsw_sp_netevent_work *net_work;
 	struct mlxsw_sp_router *router;
 
-	if (!net_eq(net, &init_net))
+	router = container_of(nb, struct mlxsw_sp_router, netevent_nb);
+	if (!net_eq(net, mlxsw_sp_net(router->mlxsw_sp)))
 		return NOTIFY_DONE;
 
 	net_work = kzalloc(sizeof(*net_work), GFP_ATOMIC);
 	if (!net_work)
 		return NOTIFY_BAD;
 
-	router = container_of(nb, struct mlxsw_sp_router, netevent_nb);
 	INIT_WORK(&net_work->work, cb);
 	net_work->mlxsw_sp = router->mlxsw_sp;
 	mlxsw_core_schedule_work(&net_work->work);
@@ -7953,9 +7954,10 @@ 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);
 }
 
-static void mlxsw_sp_mp4_hash_init(char *recr2_pl)
+static void mlxsw_sp_mp4_hash_init(struct mlxsw_sp *mlxsw_sp, char *recr2_pl)
 {
-	bool only_l3 = !init_net.ipv4.sysctl_fib_multipath_hash_policy;
+	struct net *net = mlxsw_sp_net(mlxsw_sp);
+	bool only_l3 = !net->ipv4.sysctl_fib_multipath_hash_policy;
 
 	mlxsw_sp_mp_hash_header_set(recr2_pl,
 				    MLXSW_REG_RECR2_IPV4_EN_NOT_TCP_NOT_UDP);
@@ -7970,9 +7972,9 @@ static void mlxsw_sp_mp4_hash_init(char *recr2_pl)
 	mlxsw_sp_mp_hash_field_set(recr2_pl, MLXSW_REG_RECR2_TCP_UDP_DPORT);
 }
 
-static void mlxsw_sp_mp6_hash_init(char *recr2_pl)
+static void mlxsw_sp_mp6_hash_init(struct mlxsw_sp *mlxsw_sp, char *recr2_pl)
 {
-	bool only_l3 = !ip6_multipath_hash_policy(&init_net);
+	bool only_l3 = !ip6_multipath_hash_policy(mlxsw_sp_net(mlxsw_sp));
 
 	mlxsw_sp_mp_hash_header_set(recr2_pl,
 				    MLXSW_REG_RECR2_IPV6_EN_NOT_TCP_NOT_UDP);
@@ -8000,8 +8002,8 @@ static int mlxsw_sp_mp_hash_init(struct mlxsw_sp *mlxsw_sp)
 
 	seed = jhash(mlxsw_sp->base_mac, sizeof(mlxsw_sp->base_mac), 0);
 	mlxsw_reg_recr2_pack(recr2_pl, seed);
-	mlxsw_sp_mp4_hash_init(recr2_pl);
-	mlxsw_sp_mp6_hash_init(recr2_pl);
+	mlxsw_sp_mp4_hash_init(mlxsw_sp, recr2_pl);
+	mlxsw_sp_mp6_hash_init(mlxsw_sp, recr2_pl);
 
 	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(recr2), recr2_pl);
 }
@@ -8032,7 +8034,8 @@ static int mlxsw_sp_dscp_init(struct mlxsw_sp *mlxsw_sp)
 
 static int __mlxsw_sp_router_init(struct mlxsw_sp *mlxsw_sp)
 {
-	bool usp = init_net.ipv4.sysctl_ip_fwd_update_priority;
+	struct net *net = mlxsw_sp_net(mlxsw_sp);
+	bool usp = net->ipv4.sysctl_ip_fwd_update_priority;
 	char rgcr_pl[MLXSW_REG_RGCR_LEN];
 	u64 max_rifs;
 	int err;
@@ -8134,7 +8137,8 @@ int mlxsw_sp_router_init(struct mlxsw_sp *mlxsw_sp)
 		goto err_dscp_init;
 
 	mlxsw_sp->router->fib_nb.notifier_call = mlxsw_sp_router_fib_event;
-	err = register_fib_notifier(&init_net, &mlxsw_sp->router->fib_nb,
+	err = register_fib_notifier(mlxsw_sp_net(mlxsw_sp),
+				    &mlxsw_sp->router->fib_nb,
 				    mlxsw_sp_router_fib_dump_flush, NULL);
 	if (err)
 		goto err_register_fib_notifier;
@@ -8174,7 +8178,8 @@ int mlxsw_sp_router_init(struct mlxsw_sp *mlxsw_sp)
 
 void mlxsw_sp_router_fini(struct mlxsw_sp *mlxsw_sp)
 {
-	unregister_fib_notifier(&init_net, &mlxsw_sp->router->fib_nb);
+	unregister_fib_notifier(mlxsw_sp_net(mlxsw_sp),
+				&mlxsw_sp->router->fib_nb);
 	unregister_netevent_notifier(&mlxsw_sp->router->netevent_nb);
 	mlxsw_sp_neigh_fini(mlxsw_sp);
 	mlxsw_sp_vrs_fini(mlxsw_sp);
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
index 5ecb45118400..a3af171c6358 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
@@ -2591,7 +2591,7 @@ __mlxsw_sp_fdb_notify_mac_uc_tunnel_process(struct mlxsw_sp *mlxsw_sp,
 	if (err)
 		return err;
 
-	dev = __dev_get_by_index(&init_net, nve_ifindex);
+	dev = __dev_get_by_index(mlxsw_sp_net(mlxsw_sp), nve_ifindex);
 	if (!dev)
 		return -EINVAL;
 	*nve_dev = dev;
-- 
2.21.0


^ permalink raw reply related	[flat|nested] 34+ messages in thread

* [patch net-next 08/15] mlxsw: Register port netdevices into net of core
  2019-09-14  6:45 [patch net-next 00/15] devlink: allow devlink instances to change network namespace Jiri Pirko
                   ` (6 preceding siblings ...)
  2019-09-14  6:46 ` [patch net-next 07/15] mlxsw: spectrum: Take devlink net instead of init_net Jiri Pirko
@ 2019-09-14  6:46 ` Jiri Pirko
  2019-09-15  8:37   ` Ido Schimmel
  2019-09-14  6:46 ` [patch net-next 09/15] mlxsw: Propagate extack down to register_fib_notifier() Jiri Pirko
                   ` (9 subsequent siblings)
  17 siblings, 1 reply; 34+ messages in thread
From: Jiri Pirko @ 2019-09-14  6:46 UTC (permalink / raw)
  To: netdev
  Cc: davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

From: Jiri Pirko <jiri@mellanox.com>

When creating netdevices for ports, put then under network namespace
that the core/parent devlink belongs to.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlxsw/minimal.c  | 1 +
 drivers/net/ethernet/mellanox/mlxsw/spectrum.c | 1 +
 drivers/net/ethernet/mellanox/mlxsw/switchx2.c | 1 +
 3 files changed, 3 insertions(+)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/minimal.c b/drivers/net/ethernet/mellanox/mlxsw/minimal.c
index 471b0ca6d69a..cee16ad58307 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/minimal.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/minimal.c
@@ -172,6 +172,7 @@ mlxsw_m_port_create(struct mlxsw_m *mlxsw_m, u8 local_port, u8 module)
 	}
 
 	SET_NETDEV_DEV(dev, mlxsw_m->bus_info->dev);
+	dev_net_set(dev, mlxsw_core_net(mlxsw_m->core));
 	mlxsw_m_port = netdev_priv(dev);
 	mlxsw_m_port->dev = dev;
 	mlxsw_m_port->mlxsw_m = mlxsw_m;
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
index 91e4792bb7e7..92b37b806dc1 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
@@ -3635,6 +3635,7 @@ static int mlxsw_sp_port_create(struct mlxsw_sp *mlxsw_sp, u8 local_port,
 		goto err_alloc_etherdev;
 	}
 	SET_NETDEV_DEV(dev, mlxsw_sp->bus_info->dev);
+	dev_net_set(dev, mlxsw_sp_net(mlxsw_sp));
 	mlxsw_sp_port = netdev_priv(dev);
 	mlxsw_sp_port->dev = dev;
 	mlxsw_sp_port->mlxsw_sp = mlxsw_sp;
diff --git a/drivers/net/ethernet/mellanox/mlxsw/switchx2.c b/drivers/net/ethernet/mellanox/mlxsw/switchx2.c
index 1c14c051ee52..a4d09392a8d7 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/switchx2.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/switchx2.c
@@ -992,6 +992,7 @@ static int __mlxsw_sx_port_eth_create(struct mlxsw_sx *mlxsw_sx, u8 local_port,
 	if (!dev)
 		return -ENOMEM;
 	SET_NETDEV_DEV(dev, mlxsw_sx->bus_info->dev);
+	dev_net_set(dev, mlxsw_core_net(mlxsw_sx->core));
 	mlxsw_sx_port = netdev_priv(dev);
 	mlxsw_sx_port->dev = dev;
 	mlxsw_sx_port->mlxsw_sx = mlxsw_sx;
-- 
2.21.0


^ permalink raw reply related	[flat|nested] 34+ messages in thread

* [patch net-next 09/15] mlxsw: Propagate extack down to register_fib_notifier()
  2019-09-14  6:45 [patch net-next 00/15] devlink: allow devlink instances to change network namespace Jiri Pirko
                   ` (7 preceding siblings ...)
  2019-09-14  6:46 ` [patch net-next 08/15] mlxsw: Register port netdevices into net of core Jiri Pirko
@ 2019-09-14  6:46 ` Jiri Pirko
  2019-09-15  8:39   ` Ido Schimmel
  2019-09-14  6:46 ` [patch net-next 10/15] netdevsim: add all ports in nsim_dev_create() and del them in destroy() Jiri Pirko
                   ` (8 subsequent siblings)
  17 siblings, 1 reply; 34+ messages in thread
From: Jiri Pirko @ 2019-09-14  6:46 UTC (permalink / raw)
  To: netdev
  Cc: davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

From: Jiri Pirko <jiri@mellanox.com>

During the devlink reaload the extack is present, so propagate it all
the way down to register_fib_notifier() call in spectrum_router.c.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlxsw/core.c        | 13 ++++++++-----
 drivers/net/ethernet/mellanox/mlxsw/core.h        |  6 ++++--
 drivers/net/ethernet/mellanox/mlxsw/i2c.c         |  2 +-
 drivers/net/ethernet/mellanox/mlxsw/minimal.c     |  3 ++-
 drivers/net/ethernet/mellanox/mlxsw/pci.c         |  2 +-
 drivers/net/ethernet/mellanox/mlxsw/spectrum.c    | 15 +++++++++------
 drivers/net/ethernet/mellanox/mlxsw/spectrum.h    |  3 ++-
 .../net/ethernet/mellanox/mlxsw/spectrum_router.c |  5 +++--
 drivers/net/ethernet/mellanox/mlxsw/switchib.c    |  3 ++-
 drivers/net/ethernet/mellanox/mlxsw/switchx2.c    |  3 ++-
 10 files changed, 34 insertions(+), 21 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/core.c b/drivers/net/ethernet/mellanox/mlxsw/core.c
index 3fa96076e8a5..2bec677318d9 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/core.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/core.c
@@ -1005,7 +1005,7 @@ mlxsw_devlink_core_bus_device_reload_up(struct devlink *devlink,
 	return mlxsw_core_bus_device_register(mlxsw_core->bus_info,
 					      mlxsw_core->bus,
 					      mlxsw_core->bus_priv, true,
-					      devlink);
+					      devlink, extack);
 }
 
 static int mlxsw_devlink_flash_update(struct devlink *devlink,
@@ -1098,7 +1098,8 @@ static int
 __mlxsw_core_bus_device_register(const struct mlxsw_bus_info *mlxsw_bus_info,
 				 const struct mlxsw_bus *mlxsw_bus,
 				 void *bus_priv, bool reload,
-				 struct devlink *devlink)
+				 struct devlink *devlink,
+				 struct netlink_ext_ack *extack)
 {
 	const char *device_kind = mlxsw_bus_info->device_kind;
 	struct mlxsw_core *mlxsw_core;
@@ -1172,7 +1173,7 @@ __mlxsw_core_bus_device_register(const struct mlxsw_bus_info *mlxsw_bus_info,
 	}
 
 	if (mlxsw_driver->init) {
-		err = mlxsw_driver->init(mlxsw_core, mlxsw_bus_info);
+		err = mlxsw_driver->init(mlxsw_core, mlxsw_bus_info, extack);
 		if (err)
 			goto err_driver_init;
 	}
@@ -1223,14 +1224,16 @@ __mlxsw_core_bus_device_register(const struct mlxsw_bus_info *mlxsw_bus_info,
 int mlxsw_core_bus_device_register(const struct mlxsw_bus_info *mlxsw_bus_info,
 				   const struct mlxsw_bus *mlxsw_bus,
 				   void *bus_priv, bool reload,
-				   struct devlink *devlink)
+				   struct devlink *devlink,
+				   struct netlink_ext_ack *extack)
 {
 	bool called_again = false;
 	int err;
 
 again:
 	err = __mlxsw_core_bus_device_register(mlxsw_bus_info, mlxsw_bus,
-					       bus_priv, reload, devlink);
+					       bus_priv, reload,
+					       devlink, extack);
 	/* -EAGAIN is returned in case the FW was updated. FW needs
 	 * a reset, so lets try to call __mlxsw_core_bus_device_register()
 	 * again.
diff --git a/drivers/net/ethernet/mellanox/mlxsw/core.h b/drivers/net/ethernet/mellanox/mlxsw/core.h
index 693b3c5ab355..361fcdf780c3 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/core.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/core.h
@@ -37,7 +37,8 @@ void mlxsw_core_driver_unregister(struct mlxsw_driver *mlxsw_driver);
 int mlxsw_core_bus_device_register(const struct mlxsw_bus_info *mlxsw_bus_info,
 				   const struct mlxsw_bus *mlxsw_bus,
 				   void *bus_priv, bool reload,
-				   struct devlink *devlink);
+				   struct devlink *devlink,
+				   struct netlink_ext_ack *extack);
 void mlxsw_core_bus_device_unregister(struct mlxsw_core *mlxsw_core, bool reload);
 
 struct mlxsw_tx_info {
@@ -248,7 +249,8 @@ struct mlxsw_driver {
 	const char *kind;
 	size_t priv_size;
 	int (*init)(struct mlxsw_core *mlxsw_core,
-		    const struct mlxsw_bus_info *mlxsw_bus_info);
+		    const struct mlxsw_bus_info *mlxsw_bus_info,
+		    struct netlink_ext_ack *extack);
 	void (*fini)(struct mlxsw_core *mlxsw_core);
 	int (*basic_trap_groups_set)(struct mlxsw_core *mlxsw_core);
 	int (*port_type_set)(struct mlxsw_core *mlxsw_core, u8 local_port,
diff --git a/drivers/net/ethernet/mellanox/mlxsw/i2c.c b/drivers/net/ethernet/mellanox/mlxsw/i2c.c
index 95f408d0e103..34566eb62c47 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/i2c.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/i2c.c
@@ -640,7 +640,7 @@ static int mlxsw_i2c_probe(struct i2c_client *client,
 
 	err = mlxsw_core_bus_device_register(&mlxsw_i2c->bus_info,
 					     &mlxsw_i2c_bus, mlxsw_i2c, false,
-					     NULL);
+					     NULL, NULL);
 	if (err) {
 		dev_err(&client->dev, "Fail to register core bus\n");
 		return err;
diff --git a/drivers/net/ethernet/mellanox/mlxsw/minimal.c b/drivers/net/ethernet/mellanox/mlxsw/minimal.c
index cee16ad58307..5edd8de57a24 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/minimal.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/minimal.c
@@ -327,7 +327,8 @@ static void mlxsw_m_ports_remove(struct mlxsw_m *mlxsw_m)
 }
 
 static int mlxsw_m_init(struct mlxsw_core *mlxsw_core,
-			const struct mlxsw_bus_info *mlxsw_bus_info)
+			const struct mlxsw_bus_info *mlxsw_bus_info,
+			struct netlink_ext_ack *extack)
 {
 	struct mlxsw_m *mlxsw_m = mlxsw_core_driver_priv(mlxsw_core);
 	int err;
diff --git a/drivers/net/ethernet/mellanox/mlxsw/pci.c b/drivers/net/ethernet/mellanox/mlxsw/pci.c
index 615455a21567..4ac2f5c16adf 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/pci.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/pci.c
@@ -1790,7 +1790,7 @@ static int mlxsw_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 
 	err = mlxsw_core_bus_device_register(&mlxsw_pci->bus_info,
 					     &mlxsw_pci_bus, mlxsw_pci, false,
-					     NULL);
+					     NULL, NULL);
 	if (err) {
 		dev_err(&pdev->dev, "cannot register bus device\n");
 		goto err_bus_device_register;
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
index 92b37b806dc1..87e06e718646 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
@@ -4684,7 +4684,8 @@ static int mlxsw_sp_netdevice_event(struct notifier_block *unused,
 				    unsigned long event, void *ptr);
 
 static int mlxsw_sp_init(struct mlxsw_core *mlxsw_core,
-			 const struct mlxsw_bus_info *mlxsw_bus_info)
+			 const struct mlxsw_bus_info *mlxsw_bus_info,
+			 struct netlink_ext_ack *extack)
 {
 	struct mlxsw_sp *mlxsw_sp = mlxsw_core_driver_priv(mlxsw_core);
 	int err;
@@ -4777,7 +4778,7 @@ static int mlxsw_sp_init(struct mlxsw_core *mlxsw_core,
 		goto err_acl_init;
 	}
 
-	err = mlxsw_sp_router_init(mlxsw_sp);
+	err = mlxsw_sp_router_init(mlxsw_sp, extack);
 	if (err) {
 		dev_err(mlxsw_sp->bus_info->dev, "Failed to initialize router\n");
 		goto err_router_init;
@@ -4870,7 +4871,8 @@ static int mlxsw_sp_init(struct mlxsw_core *mlxsw_core,
 }
 
 static int mlxsw_sp1_init(struct mlxsw_core *mlxsw_core,
-			  const struct mlxsw_bus_info *mlxsw_bus_info)
+			  const struct mlxsw_bus_info *mlxsw_bus_info,
+			  struct netlink_ext_ack *extack)
 {
 	struct mlxsw_sp *mlxsw_sp = mlxsw_core_driver_priv(mlxsw_core);
 
@@ -4890,11 +4892,12 @@ static int mlxsw_sp1_init(struct mlxsw_core *mlxsw_core,
 	mlxsw_sp->listeners = mlxsw_sp1_listener;
 	mlxsw_sp->listeners_count = ARRAY_SIZE(mlxsw_sp1_listener);
 
-	return mlxsw_sp_init(mlxsw_core, mlxsw_bus_info);
+	return mlxsw_sp_init(mlxsw_core, mlxsw_bus_info, extack);
 }
 
 static int mlxsw_sp2_init(struct mlxsw_core *mlxsw_core,
-			  const struct mlxsw_bus_info *mlxsw_bus_info)
+			  const struct mlxsw_bus_info *mlxsw_bus_info,
+			  struct netlink_ext_ack *extack)
 {
 	struct mlxsw_sp *mlxsw_sp = mlxsw_core_driver_priv(mlxsw_core);
 
@@ -4910,7 +4913,7 @@ static int mlxsw_sp2_init(struct mlxsw_core *mlxsw_core,
 	mlxsw_sp->port_type_speed_ops = &mlxsw_sp2_port_type_speed_ops;
 	mlxsw_sp->ptp_ops = &mlxsw_sp2_ptp_ops;
 
-	return mlxsw_sp_init(mlxsw_core, mlxsw_bus_info);
+	return mlxsw_sp_init(mlxsw_core, mlxsw_bus_info, extack);
 }
 
 static void mlxsw_sp_fini(struct mlxsw_core *mlxsw_core)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
index f58d45e770cd..8f99d70d6b8b 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
@@ -525,7 +525,8 @@ union mlxsw_sp_l3addr {
 	struct in6_addr addr6;
 };
 
-int mlxsw_sp_router_init(struct mlxsw_sp *mlxsw_sp);
+int mlxsw_sp_router_init(struct mlxsw_sp *mlxsw_sp,
+			 struct netlink_ext_ack *extack);
 void mlxsw_sp_router_fini(struct mlxsw_sp *mlxsw_sp);
 int mlxsw_sp_netdevice_router_port_event(struct net_device *dev,
 					 unsigned long event, void *ptr);
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index a1c06889178c..308526ed16ba 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -8061,7 +8061,8 @@ static void __mlxsw_sp_router_fini(struct mlxsw_sp *mlxsw_sp)
 	mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(rgcr), rgcr_pl);
 }
 
-int mlxsw_sp_router_init(struct mlxsw_sp *mlxsw_sp)
+int mlxsw_sp_router_init(struct mlxsw_sp *mlxsw_sp,
+			 struct netlink_ext_ack *extack)
 {
 	struct mlxsw_sp_router *router;
 	int err;
@@ -8139,7 +8140,7 @@ int mlxsw_sp_router_init(struct mlxsw_sp *mlxsw_sp)
 	mlxsw_sp->router->fib_nb.notifier_call = mlxsw_sp_router_fib_event;
 	err = register_fib_notifier(mlxsw_sp_net(mlxsw_sp),
 				    &mlxsw_sp->router->fib_nb,
-				    mlxsw_sp_router_fib_dump_flush, NULL);
+				    mlxsw_sp_router_fib_dump_flush, extack);
 	if (err)
 		goto err_register_fib_notifier;
 
diff --git a/drivers/net/ethernet/mellanox/mlxsw/switchib.c b/drivers/net/ethernet/mellanox/mlxsw/switchib.c
index 0d9356b3f65d..4ff1e623aa76 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/switchib.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/switchib.c
@@ -446,7 +446,8 @@ static int mlxsw_sib_basic_trap_groups_set(struct mlxsw_core *mlxsw_core)
 }
 
 static int mlxsw_sib_init(struct mlxsw_core *mlxsw_core,
-			  const struct mlxsw_bus_info *mlxsw_bus_info)
+			  const struct mlxsw_bus_info *mlxsw_bus_info,
+			  struct netlink_ext_ack *extack)
 {
 	struct mlxsw_sib *mlxsw_sib = mlxsw_core_driver_priv(mlxsw_core);
 	int err;
diff --git a/drivers/net/ethernet/mellanox/mlxsw/switchx2.c b/drivers/net/ethernet/mellanox/mlxsw/switchx2.c
index a4d09392a8d7..de6cb22f68b1 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/switchx2.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/switchx2.c
@@ -1564,7 +1564,8 @@ static int mlxsw_sx_basic_trap_groups_set(struct mlxsw_core *mlxsw_core)
 }
 
 static int mlxsw_sx_init(struct mlxsw_core *mlxsw_core,
-			 const struct mlxsw_bus_info *mlxsw_bus_info)
+			 const struct mlxsw_bus_info *mlxsw_bus_info,
+			 struct netlink_ext_ack *extack)
 {
 	struct mlxsw_sx *mlxsw_sx = mlxsw_core_driver_priv(mlxsw_core);
 	int err;
-- 
2.21.0


^ permalink raw reply related	[flat|nested] 34+ messages in thread

* [patch net-next 10/15] netdevsim: add all ports in nsim_dev_create() and del them in destroy()
  2019-09-14  6:45 [patch net-next 00/15] devlink: allow devlink instances to change network namespace Jiri Pirko
                   ` (8 preceding siblings ...)
  2019-09-14  6:46 ` [patch net-next 09/15] mlxsw: Propagate extack down to register_fib_notifier() Jiri Pirko
@ 2019-09-14  6:46 ` Jiri Pirko
  2019-09-14  6:46 ` [patch net-next 11/15] netdevsim: implement proper devlink reload Jiri Pirko
                   ` (7 subsequent siblings)
  17 siblings, 0 replies; 34+ messages in thread
From: Jiri Pirko @ 2019-09-14  6:46 UTC (permalink / raw)
  To: netdev
  Cc: davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

From: Jiri Pirko <jiri@mellanox.com>

Currently the probe/remove function does this separately. Put the
addition an deletion of ports into nsim_dev_create() and
nsim_dev_destroy().

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 drivers/net/netdevsim/dev.c | 175 +++++++++++++++++++-----------------
 1 file changed, 93 insertions(+), 82 deletions(-)

diff --git a/drivers/net/netdevsim/dev.c b/drivers/net/netdevsim/dev.c
index d623501de3ea..65e02b933aa3 100644
--- a/drivers/net/netdevsim/dev.c
+++ b/drivers/net/netdevsim/dev.c
@@ -600,8 +600,92 @@ static const struct devlink_ops nsim_dev_devlink_ops = {
 #define NSIM_DEV_MAX_MACS_DEFAULT 32
 #define NSIM_DEV_TEST1_DEFAULT true
 
+static int __nsim_dev_port_add(struct nsim_dev *nsim_dev,
+			       unsigned int port_index)
+{
+	struct nsim_dev_port *nsim_dev_port;
+	struct devlink_port *devlink_port;
+	int err;
+
+	nsim_dev_port = kzalloc(sizeof(*nsim_dev_port), GFP_KERNEL);
+	if (!nsim_dev_port)
+		return -ENOMEM;
+	nsim_dev_port->port_index = port_index;
+
+	devlink_port = &nsim_dev_port->devlink_port;
+	devlink_port_attrs_set(devlink_port, DEVLINK_PORT_FLAVOUR_PHYSICAL,
+			       port_index + 1, 0, 0,
+			       nsim_dev->switch_id.id,
+			       nsim_dev->switch_id.id_len);
+	err = devlink_port_register(priv_to_devlink(nsim_dev), devlink_port,
+				    port_index);
+	if (err)
+		goto err_port_free;
+
+	err = nsim_dev_port_debugfs_init(nsim_dev, nsim_dev_port);
+	if (err)
+		goto err_dl_port_unregister;
+
+	nsim_dev_port->ns = nsim_create(nsim_dev, nsim_dev_port);
+	if (IS_ERR(nsim_dev_port->ns)) {
+		err = PTR_ERR(nsim_dev_port->ns);
+		goto err_port_debugfs_exit;
+	}
+
+	devlink_port_type_eth_set(devlink_port, nsim_dev_port->ns->netdev);
+	list_add(&nsim_dev_port->list, &nsim_dev->port_list);
+
+	return 0;
+
+err_port_debugfs_exit:
+	nsim_dev_port_debugfs_exit(nsim_dev_port);
+err_dl_port_unregister:
+	devlink_port_unregister(devlink_port);
+err_port_free:
+	kfree(nsim_dev_port);
+	return err;
+}
+
+static void __nsim_dev_port_del(struct nsim_dev_port *nsim_dev_port)
+{
+	struct devlink_port *devlink_port = &nsim_dev_port->devlink_port;
+
+	list_del(&nsim_dev_port->list);
+	devlink_port_type_clear(devlink_port);
+	nsim_destroy(nsim_dev_port->ns);
+	nsim_dev_port_debugfs_exit(nsim_dev_port);
+	devlink_port_unregister(devlink_port);
+	kfree(nsim_dev_port);
+}
+
+static void nsim_dev_port_del_all(struct nsim_dev *nsim_dev)
+{
+	struct nsim_dev_port *nsim_dev_port, *tmp;
+
+	list_for_each_entry_safe(nsim_dev_port, tmp,
+				 &nsim_dev->port_list, list)
+		__nsim_dev_port_del(nsim_dev_port);
+}
+
+static int nsim_dev_port_add_all(struct nsim_dev *nsim_dev,
+				 unsigned int port_count)
+{
+	int i, err;
+
+	for (i = 0; i < port_count; i++) {
+		err = __nsim_dev_port_add(nsim_dev, i);
+		if (err)
+			goto err_port_del_all;
+	}
+	return 0;
+
+err_port_del_all:
+	nsim_dev_port_del_all(nsim_dev);
+	return err;
+}
+
 static struct nsim_dev *
-nsim_dev_create(struct nsim_bus_dev *nsim_bus_dev, unsigned int port_count)
+nsim_dev_create(struct nsim_bus_dev *nsim_bus_dev)
 {
 	struct nsim_dev *nsim_dev;
 	struct devlink *devlink;
@@ -656,9 +740,15 @@ nsim_dev_create(struct nsim_bus_dev *nsim_bus_dev, unsigned int port_count)
 	if (err)
 		goto err_debugfs_exit;
 
+	err = nsim_dev_port_add_all(nsim_dev, nsim_bus_dev->port_count);
+	if (err)
+		goto err_bpf_dev_exit;
+
 	devlink_params_publish(devlink);
 	return nsim_dev;
 
+err_bpf_dev_exit:
+	nsim_bpf_dev_exit(nsim_dev);
 err_debugfs_exit:
 	nsim_dev_debugfs_exit(nsim_dev);
 err_traps_exit:
@@ -683,6 +773,7 @@ static void nsim_dev_destroy(struct nsim_dev *nsim_dev)
 {
 	struct devlink *devlink = priv_to_devlink(nsim_dev);
 
+	nsim_dev_port_del_all(nsim_dev);
 	nsim_bpf_dev_exit(nsim_dev);
 	nsim_dev_debugfs_exit(nsim_dev);
 	nsim_dev_traps_exit(devlink);
@@ -696,102 +787,22 @@ static void nsim_dev_destroy(struct nsim_dev *nsim_dev)
 	devlink_free(devlink);
 }
 
-static int __nsim_dev_port_add(struct nsim_dev *nsim_dev,
-			       unsigned int port_index)
-{
-	struct nsim_dev_port *nsim_dev_port;
-	struct devlink_port *devlink_port;
-	int err;
-
-	nsim_dev_port = kzalloc(sizeof(*nsim_dev_port), GFP_KERNEL);
-	if (!nsim_dev_port)
-		return -ENOMEM;
-	nsim_dev_port->port_index = port_index;
-
-	devlink_port = &nsim_dev_port->devlink_port;
-	devlink_port_attrs_set(devlink_port, DEVLINK_PORT_FLAVOUR_PHYSICAL,
-			       port_index + 1, 0, 0,
-			       nsim_dev->switch_id.id,
-			       nsim_dev->switch_id.id_len);
-	err = devlink_port_register(priv_to_devlink(nsim_dev), devlink_port,
-				    port_index);
-	if (err)
-		goto err_port_free;
-
-	err = nsim_dev_port_debugfs_init(nsim_dev, nsim_dev_port);
-	if (err)
-		goto err_dl_port_unregister;
-
-	nsim_dev_port->ns = nsim_create(nsim_dev, nsim_dev_port);
-	if (IS_ERR(nsim_dev_port->ns)) {
-		err = PTR_ERR(nsim_dev_port->ns);
-		goto err_port_debugfs_exit;
-	}
-
-	devlink_port_type_eth_set(devlink_port, nsim_dev_port->ns->netdev);
-	list_add(&nsim_dev_port->list, &nsim_dev->port_list);
-
-	return 0;
-
-err_port_debugfs_exit:
-	nsim_dev_port_debugfs_exit(nsim_dev_port);
-err_dl_port_unregister:
-	devlink_port_unregister(devlink_port);
-err_port_free:
-	kfree(nsim_dev_port);
-	return err;
-}
-
-static void __nsim_dev_port_del(struct nsim_dev_port *nsim_dev_port)
-{
-	struct devlink_port *devlink_port = &nsim_dev_port->devlink_port;
-
-	list_del(&nsim_dev_port->list);
-	devlink_port_type_clear(devlink_port);
-	nsim_destroy(nsim_dev_port->ns);
-	nsim_dev_port_debugfs_exit(nsim_dev_port);
-	devlink_port_unregister(devlink_port);
-	kfree(nsim_dev_port);
-}
-
-static void nsim_dev_port_del_all(struct nsim_dev *nsim_dev)
-{
-	struct nsim_dev_port *nsim_dev_port, *tmp;
-
-	list_for_each_entry_safe(nsim_dev_port, tmp,
-				 &nsim_dev->port_list, list)
-		__nsim_dev_port_del(nsim_dev_port);
-}
-
 int nsim_dev_probe(struct nsim_bus_dev *nsim_bus_dev)
 {
 	struct nsim_dev *nsim_dev;
-	int i;
-	int err;
 
-	nsim_dev = nsim_dev_create(nsim_bus_dev, nsim_bus_dev->port_count);
+	nsim_dev = nsim_dev_create(nsim_bus_dev);
 	if (IS_ERR(nsim_dev))
 		return PTR_ERR(nsim_dev);
 	dev_set_drvdata(&nsim_bus_dev->dev, nsim_dev);
 
-	for (i = 0; i < nsim_bus_dev->port_count; i++) {
-		err = __nsim_dev_port_add(nsim_dev, i);
-		if (err)
-			goto err_port_del_all;
-	}
 	return 0;
-
-err_port_del_all:
-	nsim_dev_port_del_all(nsim_dev);
-	nsim_dev_destroy(nsim_dev);
-	return err;
 }
 
 void nsim_dev_remove(struct nsim_bus_dev *nsim_bus_dev)
 {
 	struct nsim_dev *nsim_dev = dev_get_drvdata(&nsim_bus_dev->dev);
 
-	nsim_dev_port_del_all(nsim_dev);
 	nsim_dev_destroy(nsim_dev);
 }
 
-- 
2.21.0


^ permalink raw reply related	[flat|nested] 34+ messages in thread

* [patch net-next 11/15] netdevsim: implement proper devlink reload
  2019-09-14  6:45 [patch net-next 00/15] devlink: allow devlink instances to change network namespace Jiri Pirko
                   ` (9 preceding siblings ...)
  2019-09-14  6:46 ` [patch net-next 10/15] netdevsim: add all ports in nsim_dev_create() and del them in destroy() Jiri Pirko
@ 2019-09-14  6:46 ` Jiri Pirko
  2019-09-14  6:46 ` [patch net-next 12/15] netdevsim: register port netdevices into net of device Jiri Pirko
                   ` (6 subsequent siblings)
  17 siblings, 0 replies; 34+ messages in thread
From: Jiri Pirko @ 2019-09-14  6:46 UTC (permalink / raw)
  To: netdev
  Cc: davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

From: Jiri Pirko <jiri@mellanox.com>

During devlink reload, all driver objects should be reinstantiated with
the exception of devlink instance and devlink resources and params.
Move existing devlink_resource_size_get() calls into fib_create() just
before fib notifier is registered. Also, make sure that extack is
propagated down to fib_notifier_register() call.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 drivers/net/netdevsim/dev.c       | 137 +++++++++++++++++-------------
 drivers/net/netdevsim/fib.c       |  53 ++++++------
 drivers/net/netdevsim/netdevsim.h |   8 +-
 3 files changed, 109 insertions(+), 89 deletions(-)

diff --git a/drivers/net/netdevsim/dev.c b/drivers/net/netdevsim/dev.c
index 65e02b933aa3..ad376b443a34 100644
--- a/drivers/net/netdevsim/dev.c
+++ b/drivers/net/netdevsim/dev.c
@@ -466,37 +466,28 @@ static void nsim_dev_traps_exit(struct devlink *devlink)
 	kfree(nsim_dev->trap_data);
 }
 
+static struct nsim_dev *
+nsim_dev_create(struct nsim_bus_dev *nsim_bus_dev, struct nsim_dev *nsim_dev,
+		struct netlink_ext_ack *extack);
+static void nsim_dev_destroy(struct nsim_dev *nsim_dev, bool reload);
+
 static int nsim_dev_reload_down(struct devlink *devlink,
 				struct netlink_ext_ack *extack)
 {
+	struct nsim_dev *nsim_dev = devlink_priv(devlink);
+
+	nsim_dev_destroy(nsim_dev, true);
 	return 0;
 }
 
 static int nsim_dev_reload_up(struct devlink *devlink,
 			      struct netlink_ext_ack *extack)
+
 {
 	struct nsim_dev *nsim_dev = devlink_priv(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
-	};
-	int i;
-
-	for (i = 0; i < ARRAY_SIZE(res_ids); ++i) {
-		int err;
-		u64 val;
-
-		err = devlink_resource_size_get(devlink, res_ids[i], &val);
-		if (!err) {
-			err = nsim_fib_set_max(nsim_dev->fib_data,
-					       res_ids[i], val, extack);
-			if (err)
-				return err;
-		}
-	}
-	nsim_devlink_param_load_driverinit_values(devlink);
 
-	return 0;
+	nsim_dev = nsim_dev_create(nsim_dev->nsim_bus_dev, nsim_dev, extack);
+	return PTR_ERR_OR_ZERO(nsim_dev);
 }
 
 #define NSIM_DEV_FLASH_SIZE 500000
@@ -685,15 +676,21 @@ static int nsim_dev_port_add_all(struct nsim_dev *nsim_dev,
 }
 
 static struct nsim_dev *
-nsim_dev_create(struct nsim_bus_dev *nsim_bus_dev)
+nsim_dev_create(struct nsim_bus_dev *nsim_bus_dev, struct nsim_dev *nsim_dev,
+		struct netlink_ext_ack *extack)
 {
-	struct nsim_dev *nsim_dev;
+	bool reload = !!nsim_dev;
 	struct devlink *devlink;
 	int err;
 
-	devlink = devlink_alloc(&nsim_dev_devlink_ops, sizeof(*nsim_dev));
-	if (!devlink)
-		return ERR_PTR(-ENOMEM);
+	if (!reload) {
+		devlink = devlink_alloc(&nsim_dev_devlink_ops,
+					sizeof(*nsim_dev));
+		if (!devlink)
+			return ERR_PTR(-ENOMEM);
+	} else {
+		devlink = priv_to_devlink(nsim_dev);
+	}
 	nsim_dev = devlink_priv(devlink);
 	nsim_dev->nsim_bus_dev = nsim_bus_dev;
 	nsim_dev->switch_id.id_len = sizeof(nsim_dev->switch_id.id);
@@ -701,28 +698,35 @@ nsim_dev_create(struct nsim_bus_dev *nsim_bus_dev)
 	INIT_LIST_HEAD(&nsim_dev->port_list);
 	mutex_init(&nsim_dev->port_list_lock);
 	nsim_dev->fw_update_status = true;
-	nsim_dev->max_macs = NSIM_DEV_MAX_MACS_DEFAULT;
-	nsim_dev->test1 = NSIM_DEV_TEST1_DEFAULT;
 
-	err = nsim_dev_resources_register(devlink);
-	if (err)
-		goto err_devlink_free;
+	if (!reload) {
+		err = nsim_dev_resources_register(devlink);
+		if (err)
+			goto err_devlink_free;
+	}
 
-	nsim_dev->fib_data = nsim_fib_create(devlink);
+	nsim_dev->fib_data = nsim_fib_create(devlink, extack);
 	if (IS_ERR(nsim_dev->fib_data)) {
 		err = PTR_ERR(nsim_dev->fib_data);
 		goto err_resources_unregister;
 	}
 
-	err = devlink_register(devlink, &nsim_bus_dev->dev);
-	if (err)
-		goto err_fib_destroy;
+	if (!reload) {
+		nsim_dev->max_macs = NSIM_DEV_MAX_MACS_DEFAULT;
+		nsim_dev->test1 = NSIM_DEV_TEST1_DEFAULT;
 
-	err = devlink_params_register(devlink, nsim_devlink_params,
-				      ARRAY_SIZE(nsim_devlink_params));
-	if (err)
-		goto err_dl_unregister;
-	nsim_devlink_set_params_init_values(nsim_dev, devlink);
+		err = devlink_register(devlink, &nsim_bus_dev->dev);
+		if (err)
+			goto err_fib_destroy;
+
+		err = devlink_params_register(devlink, nsim_devlink_params,
+					      ARRAY_SIZE(nsim_devlink_params));
+		if (err)
+			goto err_dl_unregister;
+		nsim_devlink_set_params_init_values(nsim_dev, devlink);
+	} else {
+		nsim_devlink_param_load_driverinit_values(devlink);
+	}
 
 	err = nsim_dev_dummy_region_init(nsim_dev, devlink);
 	if (err)
@@ -744,7 +748,8 @@ nsim_dev_create(struct nsim_bus_dev *nsim_bus_dev)
 	if (err)
 		goto err_bpf_dev_exit;
 
-	devlink_params_publish(devlink);
+	if (reload)
+		devlink_params_publish(devlink);
 	return nsim_dev;
 
 err_bpf_dev_exit:
@@ -756,42 +761,54 @@ nsim_dev_create(struct nsim_bus_dev *nsim_bus_dev)
 err_dummy_region_exit:
 	nsim_dev_dummy_region_exit(nsim_dev);
 err_params_unregister:
-	devlink_params_unregister(devlink, nsim_devlink_params,
-				  ARRAY_SIZE(nsim_devlink_params));
+	if (!reload) {
+		devlink_params_unregister(devlink, nsim_devlink_params,
+					  ARRAY_SIZE(nsim_devlink_params));
 err_dl_unregister:
-	devlink_unregister(devlink);
+		devlink_unregister(devlink);
+	}
 err_fib_destroy:
 	nsim_fib_destroy(devlink, nsim_dev->fib_data);
 err_resources_unregister:
-	devlink_resources_unregister(devlink, NULL);
+	if (!reload) {
+		devlink_resources_unregister(devlink, NULL);
 err_devlink_free:
-	devlink_free(devlink);
+		devlink_free(devlink);
+	}
 	return ERR_PTR(err);
 }
 
-static void nsim_dev_destroy(struct nsim_dev *nsim_dev)
+static void nsim_dev_destroy(struct nsim_dev *nsim_dev, bool reload)
 {
 	struct devlink *devlink = priv_to_devlink(nsim_dev);
 
-	nsim_dev_port_del_all(nsim_dev);
-	nsim_bpf_dev_exit(nsim_dev);
-	nsim_dev_debugfs_exit(nsim_dev);
-	nsim_dev_traps_exit(devlink);
-	nsim_dev_dummy_region_exit(nsim_dev);
-	devlink_params_unregister(devlink, nsim_devlink_params,
-				  ARRAY_SIZE(nsim_devlink_params));
-	devlink_unregister(devlink);
-	nsim_fib_destroy(devlink, nsim_dev->fib_data);
-	devlink_resources_unregister(devlink, NULL);
-	mutex_destroy(&nsim_dev->port_list_lock);
-	devlink_free(devlink);
+	if (!devlink_is_reload_failed(devlink)) {
+		nsim_dev_port_del_all(nsim_dev);
+		nsim_bpf_dev_exit(nsim_dev);
+		nsim_dev_debugfs_exit(nsim_dev);
+		nsim_dev_traps_exit(devlink);
+		nsim_dev_dummy_region_exit(nsim_dev);
+		mutex_destroy(&nsim_dev->port_list_lock);
+	}
+	if (!reload) {
+		devlink_params_unregister(devlink, nsim_devlink_params,
+					  ARRAY_SIZE(nsim_devlink_params));
+		devlink_unregister(devlink);
+	}
+	if (!devlink_is_reload_failed(devlink))
+		nsim_fib_destroy(devlink, nsim_dev->fib_data);
+	if (!reload) {
+		devlink_resources_unregister(devlink, NULL);
+		mutex_destroy(&nsim_dev->port_list_lock);
+		devlink_free(devlink);
+	}
 }
 
 int nsim_dev_probe(struct nsim_bus_dev *nsim_bus_dev)
 {
 	struct nsim_dev *nsim_dev;
 
-	nsim_dev = nsim_dev_create(nsim_bus_dev);
+	nsim_dev = nsim_dev_create(nsim_bus_dev, NULL, NULL);
 	if (IS_ERR(nsim_dev))
 		return PTR_ERR(nsim_dev);
 	dev_set_drvdata(&nsim_bus_dev->dev, nsim_dev);
@@ -803,7 +820,7 @@ void nsim_dev_remove(struct nsim_bus_dev *nsim_bus_dev)
 {
 	struct nsim_dev *nsim_dev = dev_get_drvdata(&nsim_bus_dev->dev);
 
-	nsim_dev_destroy(nsim_dev);
+	nsim_dev_destroy(nsim_dev, false);
 }
 
 static struct nsim_dev_port *
diff --git a/drivers/net/netdevsim/fib.c b/drivers/net/netdevsim/fib.c
index d2aeac0f4c2c..fdc682f3a09a 100644
--- a/drivers/net/netdevsim/fib.c
+++ b/drivers/net/netdevsim/fib.c
@@ -63,12 +63,10 @@ u64 nsim_fib_get_val(struct nsim_fib_data *fib_data,
 	return max ? entry->max : entry->num;
 }
 
-int nsim_fib_set_max(struct nsim_fib_data *fib_data,
-		     enum nsim_resource_id res_id, u64 val,
-		     struct netlink_ext_ack *extack)
+static void nsim_fib_set_max(struct nsim_fib_data *fib_data,
+			     enum nsim_resource_id res_id, u64 val)
 {
 	struct nsim_fib_entry *entry;
-	int err = 0;
 
 	switch (res_id) {
 	case NSIM_RESOURCE_IPV4_FIB:
@@ -84,20 +82,10 @@ int nsim_fib_set_max(struct nsim_fib_data *fib_data,
 		entry = &fib_data->ipv6.rules;
 		break;
 	default:
-		return 0;
-	}
-
-	/* not allowing a new max to be less than curren occupancy
-	 * --> no means of evicting entries
-	 */
-	if (val < entry->num) {
-		NL_SET_ERR_MSG_MOD(extack, "New size is less than current occupancy");
-		err = -EINVAL;
-	} else {
-		entry->max = val;
+		WARN_ON(1);
+		return;
 	}
-
-	return err;
+	entry->max = val;
 }
 
 static int nsim_fib_rule_account(struct nsim_fib_entry *entry, bool add,
@@ -239,7 +227,28 @@ static u64 nsim_fib_ipv6_rules_res_occ_get(void *priv)
 	return nsim_fib_get_val(data, NSIM_RESOURCE_IPV6_FIB_RULES, false);
 }
 
-struct nsim_fib_data *nsim_fib_create(struct devlink *devlink)
+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
+	};
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(res_ids); i++) {
+		int err;
+		u64 val;
+
+		err = devlink_resource_size_get(devlink, res_ids[i], &val);
+		if (err)
+			val = (u64) -1;
+		nsim_fib_set_max(data, res_ids[i], val);
+	}
+}
+
+struct nsim_fib_data *nsim_fib_create(struct devlink *devlink,
+				      struct netlink_ext_ack *extack)
 {
 	struct nsim_fib_data *data;
 	int err;
@@ -248,15 +257,11 @@ struct nsim_fib_data *nsim_fib_create(struct devlink *devlink)
 	if (!data)
 		return ERR_PTR(-ENOMEM);
 
-	data->ipv4.fib.max = (u64)-1;
-	data->ipv4.rules.max = (u64)-1;
-
-	data->ipv6.fib.max = (u64)-1;
-	data->ipv6.rules.max = (u64)-1;
+	nsim_fib_set_max_all(data, devlink);
 
 	data->fib_nb.notifier_call = nsim_fib_event_nb;
 	err = register_fib_notifier(&init_net, &data->fib_nb,
-				    nsim_fib_dump_inconsistent, NULL);
+				    nsim_fib_dump_inconsistent, extack);
 	if (err) {
 		pr_err("Failed to register fib notifier\n");
 		goto err_out;
diff --git a/drivers/net/netdevsim/netdevsim.h b/drivers/net/netdevsim/netdevsim.h
index ac506cf253b6..702d951fe160 100644
--- a/drivers/net/netdevsim/netdevsim.h
+++ b/drivers/net/netdevsim/netdevsim.h
@@ -173,13 +173,11 @@ int nsim_dev_port_add(struct nsim_bus_dev *nsim_bus_dev,
 int nsim_dev_port_del(struct nsim_bus_dev *nsim_bus_dev,
 		      unsigned int port_index);
 
-struct nsim_fib_data *nsim_fib_create(struct devlink *devlink);
-void nsim_fib_destroy(struct devlink *devlink, struct nsim_fib_data *data);
+struct nsim_fib_data *nsim_fib_create(struct devlink *devlink,
+				      struct netlink_ext_ack *extack);
+void nsim_fib_destroy(struct devlink *devlink, struct nsim_fib_data *fib_data);
 u64 nsim_fib_get_val(struct nsim_fib_data *fib_data,
 		     enum nsim_resource_id res_id, bool max);
-int nsim_fib_set_max(struct nsim_fib_data *fib_data,
-		     enum nsim_resource_id res_id, u64 val,
-		     struct netlink_ext_ack *extack);
 
 #if IS_ENABLED(CONFIG_XFRM_OFFLOAD)
 void nsim_ipsec_init(struct netdevsim *ns);
-- 
2.21.0


^ permalink raw reply related	[flat|nested] 34+ messages in thread

* [patch net-next 12/15] netdevsim: register port netdevices into net of device
  2019-09-14  6:45 [patch net-next 00/15] devlink: allow devlink instances to change network namespace Jiri Pirko
                   ` (10 preceding siblings ...)
  2019-09-14  6:46 ` [patch net-next 11/15] netdevsim: implement proper devlink reload Jiri Pirko
@ 2019-09-14  6:46 ` Jiri Pirko
  2019-09-14  6:46 ` [patch net-next 13/15] netdevsim: take devlink net instead of init_net Jiri Pirko
                   ` (5 subsequent siblings)
  17 siblings, 0 replies; 34+ messages in thread
From: Jiri Pirko @ 2019-09-14  6:46 UTC (permalink / raw)
  To: netdev
  Cc: davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

From: Jiri Pirko <jiri@mellanox.com>

Register newly created port netdevice into net namespace
that the parent device belongs to.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 drivers/net/netdevsim/netdev.c    | 1 +
 drivers/net/netdevsim/netdevsim.h | 5 +++++
 2 files changed, 6 insertions(+)

diff --git a/drivers/net/netdevsim/netdev.c b/drivers/net/netdevsim/netdev.c
index 0740940f41b1..2908e0a0d6e1 100644
--- a/drivers/net/netdevsim/netdev.c
+++ b/drivers/net/netdevsim/netdev.c
@@ -290,6 +290,7 @@ nsim_create(struct nsim_dev *nsim_dev, struct nsim_dev_port *nsim_dev_port)
 	if (!dev)
 		return ERR_PTR(-ENOMEM);
 
+	dev_net_set(dev, nsim_dev_net(nsim_dev));
 	ns = netdev_priv(dev);
 	ns->netdev = dev;
 	ns->nsim_dev = nsim_dev;
diff --git a/drivers/net/netdevsim/netdevsim.h b/drivers/net/netdevsim/netdevsim.h
index 702d951fe160..198ca31cec94 100644
--- a/drivers/net/netdevsim/netdevsim.h
+++ b/drivers/net/netdevsim/netdevsim.h
@@ -164,6 +164,11 @@ struct nsim_dev {
 	struct devlink_region *dummy_region;
 };
 
+static inline struct net *nsim_dev_net(struct nsim_dev *nsim_dev)
+{
+	return devlink_net(priv_to_devlink(nsim_dev));
+}
+
 int nsim_dev_init(void);
 void nsim_dev_exit(void);
 int nsim_dev_probe(struct nsim_bus_dev *nsim_bus_dev);
-- 
2.21.0


^ permalink raw reply related	[flat|nested] 34+ messages in thread

* [patch net-next 13/15] netdevsim: take devlink net instead of init_net
  2019-09-14  6:45 [patch net-next 00/15] devlink: allow devlink instances to change network namespace Jiri Pirko
                   ` (11 preceding siblings ...)
  2019-09-14  6:46 ` [patch net-next 12/15] netdevsim: register port netdevices into net of device Jiri Pirko
@ 2019-09-14  6:46 ` Jiri Pirko
  2019-09-14  6:46 ` [patch net-next 14/15] net: devlink: allow to change namespaces during reload Jiri Pirko
                   ` (4 subsequent siblings)
  17 siblings, 0 replies; 34+ messages in thread
From: Jiri Pirko @ 2019-09-14  6:46 UTC (permalink / raw)
  To: netdev
  Cc: davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

From: Jiri Pirko <jiri@mellanox.com>

Follow-up patch is going to allow to reload devlink instance into
different network namespace, so use devlink_net() helper instead
of init_net.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 drivers/net/netdevsim/fib.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/netdevsim/fib.c b/drivers/net/netdevsim/fib.c
index fdc682f3a09a..13540dee7364 100644
--- a/drivers/net/netdevsim/fib.c
+++ b/drivers/net/netdevsim/fib.c
@@ -260,7 +260,7 @@ struct nsim_fib_data *nsim_fib_create(struct devlink *devlink,
 	nsim_fib_set_max_all(data, devlink);
 
 	data->fib_nb.notifier_call = nsim_fib_event_nb;
-	err = register_fib_notifier(&init_net, &data->fib_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");
@@ -300,6 +300,6 @@ void nsim_fib_destroy(struct devlink *devlink, struct nsim_fib_data *data)
 					    NSIM_RESOURCE_IPV4_FIB_RULES);
 	devlink_resource_occ_get_unregister(devlink,
 					    NSIM_RESOURCE_IPV4_FIB);
-	unregister_fib_notifier(&init_net, &data->fib_nb);
+	unregister_fib_notifier(devlink_net(devlink), &data->fib_nb);
 	kfree(data);
 }
-- 
2.21.0


^ permalink raw reply related	[flat|nested] 34+ messages in thread

* [patch net-next 14/15] net: devlink: allow to change namespaces during reload
  2019-09-14  6:45 [patch net-next 00/15] devlink: allow devlink instances to change network namespace Jiri Pirko
                   ` (12 preceding siblings ...)
  2019-09-14  6:46 ` [patch net-next 13/15] netdevsim: take devlink net instead of init_net Jiri Pirko
@ 2019-09-14  6:46 ` Jiri Pirko
  2019-09-15  8:58   ` Ido Schimmel
  2019-09-14  6:46 ` [patch net-next 15/15] selftests: netdevsim: add tests for devlink reload with resources Jiri Pirko
                   ` (3 subsequent siblings)
  17 siblings, 1 reply; 34+ messages in thread
From: Jiri Pirko @ 2019-09-14  6:46 UTC (permalink / raw)
  To: netdev
  Cc: davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

From: Jiri Pirko <jiri@mellanox.com>

All devlink instances are created in init_net and stay there for a
lifetime. Allow user to be able to move devlink instances into
namespaces during devlink reload operation. That ensures proper
re-instantiation of driver objects, including netdevices.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx4/main.c |   4 +
 include/uapi/linux/devlink.h              |   4 +
 net/core/devlink.c                        | 155 ++++++++++++++++++++--
 3 files changed, 155 insertions(+), 8 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c
index ef3f3d06ff1e..989d0882aaa9 100644
--- a/drivers/net/ethernet/mellanox/mlx4/main.c
+++ b/drivers/net/ethernet/mellanox/mlx4/main.c
@@ -3942,6 +3942,10 @@ static int mlx4_devlink_reload_down(struct devlink *devlink,
 	struct mlx4_dev *dev = &priv->dev;
 	struct mlx4_dev_persistent *persist = dev->persist;
 
+	if (!net_eq(devlink_net(devlink), &init_net)) {
+		NL_SET_ERR_MSG_MOD(extack, "Namespace change is not supported");
+		return -EOPNOTSUPP;
+	}
 	if (persist->num_vfs)
 		mlx4_warn(persist->dev, "Reload performed on PF, will cause reset on operating Virtual Functions\n");
 	mlx4_restart_one_down(persist->pdev);
diff --git a/include/uapi/linux/devlink.h b/include/uapi/linux/devlink.h
index 580b7a2e40e1..b558ea88b766 100644
--- a/include/uapi/linux/devlink.h
+++ b/include/uapi/linux/devlink.h
@@ -421,6 +421,10 @@ enum devlink_attr {
 
 	DEVLINK_ATTR_RELOAD_FAILED,			/* u8 0 or 1 */
 
+	DEVLINK_ATTR_NETNS_FD,			/* u32 */
+	DEVLINK_ATTR_NETNS_PID,			/* u32 */
+	DEVLINK_ATTR_NETNS_ID,			/* u32 */
+
 	/* add new attributes above here, update the policy in devlink.c */
 
 	__DEVLINK_ATTR_MAX,
diff --git a/net/core/devlink.c b/net/core/devlink.c
index 362cbbcca225..2a5db95cce3c 100644
--- a/net/core/devlink.c
+++ b/net/core/devlink.c
@@ -435,8 +435,16 @@ static void devlink_nl_post_doit(const struct genl_ops *ops,
 {
 	struct devlink *devlink;
 
-	devlink = devlink_get_from_info(info);
-	if (~ops->internal_flags & DEVLINK_NL_FLAG_NO_LOCK)
+	/* When devlink changes netns, it would not be found
+	 * by devlink_get_from_info(). So try if it is stored first.
+	 */
+	if (ops->internal_flags & DEVLINK_NL_FLAG_NEED_DEVLINK) {
+		devlink = info->user_ptr[0];
+	} else {
+		devlink = devlink_get_from_info(info);
+		WARN_ON(IS_ERR(devlink));
+	}
+	if (!IS_ERR(devlink) && ~ops->internal_flags & DEVLINK_NL_FLAG_NO_LOCK)
 		mutex_unlock(&devlink->lock);
 	mutex_unlock(&devlink_mutex);
 }
@@ -2675,6 +2683,73 @@ devlink_resources_validate(struct devlink *devlink,
 	return err;
 }
 
+static struct net *devlink_netns_get(struct sk_buff *skb,
+				     struct devlink *devlink,
+				     struct genl_info *info)
+{
+	struct nlattr *netns_pid_attr = info->attrs[DEVLINK_ATTR_NETNS_PID];
+	struct nlattr *netns_fd_attr = info->attrs[DEVLINK_ATTR_NETNS_FD];
+	struct nlattr *netns_id_attr = info->attrs[DEVLINK_ATTR_NETNS_ID];
+	struct net *net;
+
+	if (!!netns_pid_attr + !!netns_fd_attr + !!netns_id_attr > 1) {
+		NL_SET_ERR_MSG(info->extack, "multiple netns identifying attributes specified");
+		return ERR_PTR(-EINVAL);
+	}
+
+	if (netns_pid_attr) {
+		net = get_net_ns_by_pid(nla_get_u32(netns_pid_attr));
+	} else if (netns_fd_attr) {
+		net = get_net_ns_by_fd(nla_get_u32(netns_fd_attr));
+	} else if (netns_id_attr) {
+		net = get_net_ns_by_id(sock_net(skb->sk),
+				       nla_get_u32(netns_id_attr));
+		if (!net)
+			net = ERR_PTR(-EINVAL);
+	} else {
+		WARN_ON(1);
+		net = ERR_PTR(-EINVAL);
+	}
+	if (IS_ERR(net)) {
+		NL_SET_ERR_MSG(info->extack, "Unknown network namespace");
+		return ERR_PTR(-EINVAL);
+	}
+	if (!netlink_ns_capable(skb, net->user_ns, CAP_NET_ADMIN)) {
+		put_net(net);
+		return ERR_PTR(-EPERM);
+	}
+	return net;
+}
+
+static void devlink_param_notify(struct devlink *devlink,
+				 unsigned int port_index,
+				 struct devlink_param_item *param_item,
+				 enum devlink_command cmd);
+
+static void devlink_reload_netns_change(struct devlink *devlink,
+					struct net *dest_net)
+{
+	struct devlink_param_item *param_item;
+
+	/* Userspace needs to be notified about devlink objects
+	 * removed from original and entering new network namespace.
+	 * The rest of the devlink objects are re-created during
+	 * reload process so the notifications are generated separatelly.
+	 */
+
+	list_for_each_entry(param_item, &devlink->param_list, list)
+		devlink_param_notify(devlink, 0, param_item,
+				     DEVLINK_CMD_PARAM_DEL);
+	devlink_notify(devlink, DEVLINK_CMD_DEL);
+
+	devlink_net_set(devlink, dest_net);
+
+	devlink_notify(devlink, DEVLINK_CMD_NEW);
+	list_for_each_entry(param_item, &devlink->param_list, list)
+		devlink_param_notify(devlink, 0, param_item,
+				     DEVLINK_CMD_PARAM_NEW);
+}
+
 static bool devlink_reload_supported(struct devlink *devlink)
 {
 	return devlink->ops->reload_down && devlink->ops->reload_up;
@@ -2695,9 +2770,27 @@ bool devlink_is_reload_failed(const struct devlink *devlink)
 }
 EXPORT_SYMBOL_GPL(devlink_is_reload_failed);
 
+static int devlink_reload(struct devlink *devlink, struct net *dest_net,
+			  struct netlink_ext_ack *extack)
+{
+	int err;
+
+	err = devlink->ops->reload_down(devlink, extack);
+	if (err)
+		return err;
+
+	if (dest_net && !net_eq(dest_net, devlink_net(devlink)))
+		devlink_reload_netns_change(devlink, dest_net);
+
+	err = devlink->ops->reload_up(devlink, extack);
+	devlink_reload_failed_set(devlink, !!err);
+	return err;
+}
+
 static int devlink_nl_cmd_reload(struct sk_buff *skb, struct genl_info *info)
 {
 	struct devlink *devlink = info->user_ptr[0];
+	struct net *dest_net = NULL;
 	int err;
 
 	if (!devlink_reload_supported(devlink))
@@ -2708,11 +2801,20 @@ static int devlink_nl_cmd_reload(struct sk_buff *skb, struct genl_info *info)
 		NL_SET_ERR_MSG_MOD(info->extack, "resources size validation failed");
 		return err;
 	}
-	err = devlink->ops->reload_down(devlink, info->extack);
-	if (err)
-		return err;
-	err = devlink->ops->reload_up(devlink, info->extack);
-	devlink_reload_failed_set(devlink, !!err);
+
+	if (info->attrs[DEVLINK_ATTR_NETNS_PID] ||
+	    info->attrs[DEVLINK_ATTR_NETNS_FD] ||
+	    info->attrs[DEVLINK_ATTR_NETNS_ID]) {
+		dest_net = devlink_netns_get(skb, devlink, info);
+		if (IS_ERR(dest_net))
+			return PTR_ERR(dest_net);
+	}
+
+	err = devlink_reload(devlink, dest_net, info->extack);
+
+	if (dest_net)
+		put_net(dest_net);
+
 	return err;
 }
 
@@ -5794,6 +5896,9 @@ static const struct nla_policy devlink_nl_policy[DEVLINK_ATTR_MAX + 1] = {
 	[DEVLINK_ATTR_TRAP_NAME] = { .type = NLA_NUL_STRING },
 	[DEVLINK_ATTR_TRAP_ACTION] = { .type = NLA_U8 },
 	[DEVLINK_ATTR_TRAP_GROUP_NAME] = { .type = NLA_NUL_STRING },
+	[DEVLINK_ATTR_NETNS_PID] = { .type = NLA_U32 },
+	[DEVLINK_ATTR_NETNS_FD] = { .type = NLA_U32 },
+	[DEVLINK_ATTR_NETNS_ID] = { .type = NLA_U32 },
 };
 
 static const struct genl_ops devlink_nl_ops[] = {
@@ -8061,9 +8166,43 @@ int devlink_compat_switch_id_get(struct net_device *dev,
 	return 0;
 }
 
+static void __net_exit devlink_pernet_pre_exit(struct net *net)
+{
+	struct devlink *devlink;
+	int err;
+
+	/* In case network namespace is getting destroyed, reload
+	 * all devlink instances from this namespace into init_net.
+	 */
+	mutex_lock(&devlink_mutex);
+	list_for_each_entry(devlink, &devlink_list, list) {
+		if (net_eq(devlink_net(devlink), net)) {
+			if (WARN_ON(!devlink_reload_supported(devlink)))
+				continue;
+			err = devlink_reload(devlink, &init_net, NULL);
+			if (err)
+				pr_warn("Failed to reload devlink instance into init_net\n");
+		}
+	}
+	mutex_unlock(&devlink_mutex);
+}
+
+static struct pernet_operations devlink_pernet_ops __net_initdata = {
+	.pre_exit = devlink_pernet_pre_exit,
+};
+
 static int __init devlink_init(void)
 {
-	return genl_register_family(&devlink_nl_family);
+	int err;
+
+	err = genl_register_family(&devlink_nl_family);
+	if (err)
+		goto out;
+	err = register_pernet_subsys(&devlink_pernet_ops);
+
+out:
+	WARN_ON(err);
+	return err;
 }
 
 subsys_initcall(devlink_init);
-- 
2.21.0


^ permalink raw reply related	[flat|nested] 34+ messages in thread

* [patch net-next 15/15] selftests: netdevsim: add tests for devlink reload with resources
  2019-09-14  6:45 [patch net-next 00/15] devlink: allow devlink instances to change network namespace Jiri Pirko
                   ` (13 preceding siblings ...)
  2019-09-14  6:46 ` [patch net-next 14/15] net: devlink: allow to change namespaces during reload Jiri Pirko
@ 2019-09-14  6:46 ` Jiri Pirko
  2019-09-14  6:57 ` [patch iproute2-next 1/2] devlink: introduce cmdline option to switch to a different namespace Jiri Pirko
                   ` (2 subsequent siblings)
  17 siblings, 0 replies; 34+ messages in thread
From: Jiri Pirko @ 2019-09-14  6:46 UTC (permalink / raw)
  To: netdev
  Cc: davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

From: Jiri Pirko <jiri@mellanox.com>

Add couple of tests for devlink reload testing and also resource
limitations testing, along with devlink reload.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
 .../drivers/net/netdevsim/devlink.sh          | 120 +++++++++++++++++-
 1 file changed, 119 insertions(+), 1 deletion(-)

diff --git a/tools/testing/selftests/drivers/net/netdevsim/devlink.sh b/tools/testing/selftests/drivers/net/netdevsim/devlink.sh
index 115837355eaf..69af99bd562b 100755
--- a/tools/testing/selftests/drivers/net/netdevsim/devlink.sh
+++ b/tools/testing/selftests/drivers/net/netdevsim/devlink.sh
@@ -3,7 +3,8 @@
 
 lib_dir=$(dirname $0)/../../../net/forwarding
 
-ALL_TESTS="fw_flash_test params_test regions_test"
+ALL_TESTS="fw_flash_test params_test regions_test reload_test \
+	   netns_reload_test resource_test"
 NUM_NETIFS=0
 source $lib_dir/lib.sh
 
@@ -142,6 +143,123 @@ regions_test()
 	log_test "regions test"
 }
 
+reload_test()
+{
+	RET=0
+
+	devlink dev reload $DL_HANDLE
+	check_err $? "Failed to reload"
+
+	log_test "reload test"
+}
+
+netns_reload_test()
+{
+	RET=0
+
+	ip netns add testns1
+	check_err $? "Failed add netns \"testns1\""
+	ip netns add testns2
+	check_err $? "Failed add netns \"testns2\""
+
+	devlink dev reload $DL_HANDLE netns testns1
+	check_err $? "Failed to reload into netns \"testns1\""
+
+	devlink -N testns1 dev reload $DL_HANDLE netns testns2
+	check_err $? "Failed to reload from netns \"testns1\" into netns \"testns2\""
+
+	ip netns del testns2
+	ip netns del testns1
+
+	log_test "netns reload test"
+}
+
+DUMMYDEV="dummytest"
+
+res_val_get()
+{
+	local netns=$1
+	local parentname=$2
+	local name=$3
+	local type=$4
+
+	cmd_jq "devlink -N $netns resource show $DL_HANDLE -j" \
+	       ".[][][] | select(.name == \"$parentname\").resources[] \
+	        | select(.name == \"$name\").$type"
+}
+
+resource_test()
+{
+	RET=0
+
+	ip netns add testns1
+	check_err $? "Failed add netns \"testns1\""
+	ip netns add testns2
+	check_err $? "Failed add netns \"testns2\""
+
+	devlink dev reload $DL_HANDLE netns testns1
+	check_err $? "Failed to reload into netns \"testns1\""
+
+	# Create dummy dev to add the address and routes on.
+
+	ip -n testns1 link add name $DUMMYDEV type dummy
+	check_err $? "Failed create dummy device"
+	ip -n testns1 link set $DUMMYDEV up
+	check_err $? "Failed bring up dummy device"
+	ip -n testns1 a a 192.0.1.1/24 dev $DUMMYDEV
+	check_err $? "Failed add an IP address to dummy device"
+
+	local occ=$(res_val_get testns1 IPv4 fib occ)
+	local limit=$((occ+1))
+
+	# Set fib size limit to handle one another route only.
+
+	devlink -N testns1 resource set $DL_HANDLE path IPv4/fib size $limit
+	check_err $? "Failed to set IPv4/fib resource size"
+	local size_new=$(res_val_get testns1 IPv4 fib size_new)
+	[ "$size_new" -eq "$limit" ]
+	check_err $? "Unexpected \"size_new\" value (got $size_new, expected $limit)"
+
+	devlink -N testns1 dev reload $DL_HANDLE
+	check_err $? "Failed to reload"
+	local size=$(res_val_get testns1 IPv4 fib size)
+	[ "$size" -eq "$limit" ]
+	check_err $? "Unexpected \"size\" value (got $size, expected $limit)"
+
+	# Insert 2 routes, the first is going to be inserted,
+	# the second is expected to fail to be inserted.
+
+	ip -n testns1 r a 192.0.2.0/24 via 192.0.1.2
+	check_err $? "Failed to add route"
+
+	ip -n testns1 r a 192.0.3.0/24 via 192.0.1.2
+	check_fail $? "Unexpected successful route add over limit"
+
+	# Now create another dummy in second network namespace and
+	# insert two routes. That is over the limit of the netdevsim
+	# instance in the first namespace. Move the netdevsim instance
+	# into the second namespace and expect it to fail.
+
+	ip -n testns2 link add name $DUMMYDEV type dummy
+	check_err $? "Failed create dummy device"
+	ip -n testns2 link set $DUMMYDEV up
+	check_err $? "Failed bring up dummy device"
+	ip -n testns2 a a 192.0.1.1/24 dev $DUMMYDEV
+	check_err $? "Failed add an IP address to dummy device"
+	ip -n testns2 r a 192.0.2.0/24 via 192.0.1.2
+	check_err $? "Failed to add route"
+	ip -n testns2 r a 192.0.3.0/24 via 192.0.1.2
+	check_err $? "Failed to add route"
+
+	devlink -N testns1 dev reload $DL_HANDLE netns testns2
+	check_fail $? "Unexpected successful reload from netns \"testns1\" into netns \"testns2\""
+
+	ip netns del testns2
+	ip netns del testns1
+
+	log_test "resource test"
+}
+
 setup_prepare()
 {
 	modprobe netdevsim
-- 
2.21.0


^ permalink raw reply related	[flat|nested] 34+ messages in thread

* [patch iproute2-next 1/2] devlink: introduce cmdline option to switch to a different namespace
  2019-09-14  6:45 [patch net-next 00/15] devlink: allow devlink instances to change network namespace Jiri Pirko
                   ` (14 preceding siblings ...)
  2019-09-14  6:46 ` [patch net-next 15/15] selftests: netdevsim: add tests for devlink reload with resources Jiri Pirko
@ 2019-09-14  6:57 ` Jiri Pirko
  2019-09-15 18:01   ` David Ahern
  2019-09-14  6:57 ` [patch iproute2-next 2/2] devlink: extend reload command to add support for network namespace change Jiri Pirko
  2019-09-16  7:01 ` [patch net-next 00/15] devlink: allow devlink instances to change network namespace David Miller
  17 siblings, 1 reply; 34+ messages in thread
From: Jiri Pirko @ 2019-09-14  6:57 UTC (permalink / raw)
  To: netdev
  Cc: davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

From: Jiri Pirko <jiri@mellanox.com>

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
v3->v4:
- rebased on top of trap patches
---
 devlink/devlink.c  | 12 ++++++++++--
 man/man8/devlink.8 |  4 ++++
 2 files changed, 14 insertions(+), 2 deletions(-)

diff --git a/devlink/devlink.c b/devlink/devlink.c
index a1be8528c3c9..8020d76dd7f7 100644
--- a/devlink/devlink.c
+++ b/devlink/devlink.c
@@ -31,6 +31,7 @@
 #include "mnlg.h"
 #include "json_writer.h"
 #include "utils.h"
+#include "namespace.h"
 
 #define ESWITCH_MODE_LEGACY "legacy"
 #define ESWITCH_MODE_SWITCHDEV "switchdev"
@@ -6748,7 +6749,7 @@ static int cmd_trap(struct dl *dl)
 static void help(void)
 {
 	pr_err("Usage: devlink [ OPTIONS ] OBJECT { COMMAND | help }\n"
-	       "       devlink [ -f[orce] ] -b[atch] filename\n"
+	       "       devlink [ -f[orce] ] -b[atch] filename -N[etns] netnsname\n"
 	       "where  OBJECT := { dev | port | sb | monitor | dpipe | resource | region | health | trap }\n"
 	       "       OPTIONS := { -V[ersion] | -n[o-nice-names] | -j[son] | -p[retty] | -v[erbose] -s[tatistics] }\n");
 }
@@ -6898,6 +6899,7 @@ int main(int argc, char **argv)
 		{ "pretty",		no_argument,		NULL, 'p' },
 		{ "verbose",		no_argument,		NULL, 'v' },
 		{ "statistics",		no_argument,		NULL, 's' },
+		{ "Netns",		required_argument,	NULL, 'N' },
 		{ NULL, 0, NULL, 0 }
 	};
 	const char *batch_file = NULL;
@@ -6913,7 +6915,7 @@ int main(int argc, char **argv)
 		return EXIT_FAILURE;
 	}
 
-	while ((opt = getopt_long(argc, argv, "Vfb:njpvs",
+	while ((opt = getopt_long(argc, argv, "Vfb:njpvsN:",
 				  long_options, NULL)) >= 0) {
 
 		switch (opt) {
@@ -6942,6 +6944,12 @@ int main(int argc, char **argv)
 		case 's':
 			dl->stats = true;
 			break;
+		case 'N':
+			if (netns_switch(optarg)) {
+				ret = EXIT_FAILURE;
+				goto dl_free;
+			}
+			break;
 		default:
 			pr_err("Unknown option.\n");
 			help();
diff --git a/man/man8/devlink.8 b/man/man8/devlink.8
index 12d489440a3d..7f4eda568081 100644
--- a/man/man8/devlink.8
+++ b/man/man8/devlink.8
@@ -55,6 +55,10 @@ Turn on verbose output.
 .BR "\-s" , " --statistics"
 Output statistics.
 
+.TP
+.BR "\-N", " \-Netns " <NETNSNAME>
+Switches to the specified network namespace.
+
 .SS
 .I OBJECT
 
-- 
2.21.0


^ permalink raw reply related	[flat|nested] 34+ messages in thread

* [patch iproute2-next 2/2] devlink: extend reload command to add support for network namespace change
  2019-09-14  6:45 [patch net-next 00/15] devlink: allow devlink instances to change network namespace Jiri Pirko
                   ` (15 preceding siblings ...)
  2019-09-14  6:57 ` [patch iproute2-next 1/2] devlink: introduce cmdline option to switch to a different namespace Jiri Pirko
@ 2019-09-14  6:57 ` Jiri Pirko
  2019-09-15  7:16   ` Ido Schimmel
  2019-09-16  7:01 ` [patch net-next 00/15] devlink: allow devlink instances to change network namespace David Miller
  17 siblings, 1 reply; 34+ messages in thread
From: Jiri Pirko @ 2019-09-14  6:57 UTC (permalink / raw)
  To: netdev
  Cc: davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

From: Jiri Pirko <jiri@mellanox.com>

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
v3->v4:
- rebased on top of trap patches
- moved netns change to reload command instead of set
---
 devlink/devlink.c            | 31 +++++++++++++++++++++++++++----
 include/uapi/linux/devlink.h |  4 ++++
 man/man8/devlink-dev.8       | 12 ++++++++++++
 3 files changed, 43 insertions(+), 4 deletions(-)

diff --git a/devlink/devlink.c b/devlink/devlink.c
index 8020d76dd7f7..6a28a7aa58a4 100644
--- a/devlink/devlink.c
+++ b/devlink/devlink.c
@@ -237,6 +237,7 @@ static void ifname_map_free(struct ifname_map *ifname_map)
 #define DL_OPT_TRAP_NAME		BIT(29)
 #define DL_OPT_TRAP_ACTION		BIT(30)
 #define DL_OPT_TRAP_GROUP_NAME		BIT(31)
+#define DL_OPT_NETNS	BIT(32)
 
 struct dl_opts {
 	uint64_t present; /* flags of present items */
@@ -276,6 +277,8 @@ struct dl_opts {
 	const char *trap_name;
 	const char *trap_group_name;
 	enum devlink_trap_action trap_action;
+	bool netns_is_pid;
+	uint32_t netns;
 };
 
 struct dl {
@@ -1412,6 +1415,22 @@ static int dl_argv_parse(struct dl *dl, uint64_t o_required,
 			if (err)
 				return err;
 			o_found |= DL_OPT_TRAP_ACTION;
+		} else if (dl_argv_match(dl, "netns") &&
+			(o_all & DL_OPT_NETNS)) {
+			const char *netns_str;
+
+			dl_arg_inc(dl);
+			err = dl_argv_str(dl, &netns_str);
+			if (err)
+				return err;
+			opts->netns = netns_get_fd(netns_str);
+			if (opts->netns < 0) {
+				err = dl_argv_uint32_t(dl, &opts->netns);
+				if (err)
+					return err;
+				opts->netns_is_pid = true;
+			}
+			o_found |= DL_OPT_NETNS;
 		} else {
 			pr_err("Unknown option \"%s\"\n", dl_argv(dl));
 			return -EINVAL;
@@ -1534,7 +1553,11 @@ static void dl_opts_put(struct nlmsghdr *nlh, struct dl *dl)
 	if (opts->present & DL_OPT_TRAP_ACTION)
 		mnl_attr_put_u8(nlh, DEVLINK_ATTR_TRAP_ACTION,
 				opts->trap_action);
-
+	if (opts->present & DL_OPT_NETNS)
+		mnl_attr_put_u32(nlh,
+				 opts->netns_is_pid ? DEVLINK_ATTR_NETNS_PID :
+						      DEVLINK_ATTR_NETNS_FD,
+				 opts->netns);
 }
 
 static int dl_argv_parse_put(struct nlmsghdr *nlh, struct dl *dl,
@@ -1595,7 +1618,7 @@ static void cmd_dev_help(void)
 	pr_err("       devlink dev eswitch show DEV\n");
 	pr_err("       devlink dev param set DEV name PARAMETER value VALUE cmode { permanent | driverinit | runtime }\n");
 	pr_err("       devlink dev param show [DEV name PARAMETER]\n");
-	pr_err("       devlink dev reload DEV\n");
+	pr_err("       devlink dev reload DEV [ netns { PID | NAME | ID } ]\n");
 	pr_err("       devlink dev info [ DEV ]\n");
 	pr_err("       devlink dev flash DEV file PATH [ component NAME ]\n");
 }
@@ -2671,7 +2694,7 @@ static int cmd_dev_show(struct dl *dl)
 
 static void cmd_dev_reload_help(void)
 {
-	pr_err("Usage: devlink dev reload [ DEV ]\n");
+	pr_err("Usage: devlink dev reload DEV [ netns { PID | NAME | ID } ]\n");
 }
 
 static int cmd_dev_reload(struct dl *dl)
@@ -2687,7 +2710,7 @@ static int cmd_dev_reload(struct dl *dl)
 	nlh = mnlg_msg_prepare(dl->nlg, DEVLINK_CMD_RELOAD,
 			       NLM_F_REQUEST | NLM_F_ACK);
 
-	err = dl_argv_parse_put(nlh, dl, DL_OPT_HANDLE, 0);
+	err = dl_argv_parse_put(nlh, dl, DL_OPT_HANDLE, DL_OPT_NETNS);
 	if (err)
 		return err;
 
diff --git a/include/uapi/linux/devlink.h b/include/uapi/linux/devlink.h
index d63cf9723f57..f2608cfc9706 100644
--- a/include/uapi/linux/devlink.h
+++ b/include/uapi/linux/devlink.h
@@ -412,6 +412,10 @@ enum devlink_attr {
 
 	DEVLINK_ATTR_RELOAD_FAILED,			/* u8 0 or 1 */
 
+	DEVLINK_ATTR_NETNS_FD,			/* u32 */
+	DEVLINK_ATTR_NETNS_PID,			/* u32 */
+	DEVLINK_ATTR_NETNS_ID,			/* u32 */
+
 	/* add new attributes above here, update the policy in devlink.c */
 
 	__DEVLINK_ATTR_MAX,
diff --git a/man/man8/devlink-dev.8 b/man/man8/devlink-dev.8
index 1804463b2321..0e1a5523fa7b 100644
--- a/man/man8/devlink-dev.8
+++ b/man/man8/devlink-dev.8
@@ -25,6 +25,13 @@ devlink-dev \- devlink device configuration
 .ti -8
 .B devlink dev help
 
+.ti -8
+.BR "devlink dev set"
+.IR DEV
+.RI "[ "
+.BI "netns { " PID " | " NAME " | " ID " }
+.RI "]"
+
 .ti -8
 .BR "devlink dev eswitch set"
 .IR DEV
@@ -92,6 +99,11 @@ Format is:
 .in +2
 BUS_NAME/BUS_ADDRESS
 
+.SS devlink dev set  - sets devlink device attributes
+
+.TP
+.BI "netns { " PID " | " NAME " | " ID " }
+
 .SS devlink dev eswitch show - display devlink device eswitch attributes
 .SS devlink dev eswitch set  - sets devlink device eswitch attributes
 
-- 
2.21.0


^ permalink raw reply related	[flat|nested] 34+ messages in thread

* Re: [patch iproute2-next 2/2] devlink: extend reload command to add support for network namespace change
  2019-09-14  6:57 ` [patch iproute2-next 2/2] devlink: extend reload command to add support for network namespace change Jiri Pirko
@ 2019-09-15  7:16   ` Ido Schimmel
  2019-09-15  9:44     ` Jiri Pirko
  0 siblings, 1 reply; 34+ messages in thread
From: Ido Schimmel @ 2019-09-15  7:16 UTC (permalink / raw)
  To: Jiri Pirko
  Cc: netdev, davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm,
	kuznet, yoshfuji, shuah, mlxsw

On Sat, Sep 14, 2019 at 08:57:57AM +0200, Jiri Pirko wrote:
> diff --git a/man/man8/devlink-dev.8 b/man/man8/devlink-dev.8
> index 1804463b2321..0e1a5523fa7b 100644
> --- a/man/man8/devlink-dev.8
> +++ b/man/man8/devlink-dev.8
> @@ -25,6 +25,13 @@ devlink-dev \- devlink device configuration
>  .ti -8
>  .B devlink dev help
>  
> +.ti -8
> +.BR "devlink dev set"
> +.IR DEV
> +.RI "[ "
> +.BI "netns { " PID " | " NAME " | " ID " }
> +.RI "]"
> +
>  .ti -8
>  .BR "devlink dev eswitch set"
>  .IR DEV
> @@ -92,6 +99,11 @@ Format is:
>  .in +2
>  BUS_NAME/BUS_ADDRESS
>  
> +.SS devlink dev set  - sets devlink device attributes
> +
> +.TP
> +.BI "netns { " PID " | " NAME " | " ID " }

This looks like leftover from previous version?

> +
>  .SS devlink dev eswitch show - display devlink device eswitch attributes
>  .SS devlink dev eswitch set  - sets devlink device eswitch attributes
>  
> -- 
> 2.21.0
> 

^ permalink raw reply	[flat|nested] 34+ messages in thread

* Re: [patch net-next 02/15] net: fib_notifier: make FIB notifier per-netns
  2019-09-14  6:45 ` [patch net-next 02/15] net: fib_notifier: make FIB notifier per-netns Jiri Pirko
@ 2019-09-15  8:06   ` Ido Schimmel
  2019-09-15  9:37     ` Jiri Pirko
  2019-09-15 20:05   ` David Ahern
  1 sibling, 1 reply; 34+ messages in thread
From: Ido Schimmel @ 2019-09-15  8:06 UTC (permalink / raw)
  To: Jiri Pirko
  Cc: netdev, davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm,
	kuznet, yoshfuji, shuah, mlxsw

On Sat, Sep 14, 2019 at 08:45:55AM +0200, Jiri Pirko wrote:
> From: Jiri Pirko <jiri@mellanox.com>
> 
> Currently all users of FIB notifier only cares about events in init_net.

s/cares/care/

> Later in this patchset, users get interested in other namespaces too.
> However, for every registered block user is interested only about one
> namespace. Make the FIB notifier registration per-netns and avoid
> unnecessary calls of notifier block for other namespaces.

...

> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c b/drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c
> index 5d20d615663e..fe0cc969cf94 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c
> @@ -248,9 +248,6 @@ static int mlx5_lag_fib_event(struct notifier_block *nb,
>  	struct net_device *fib_dev;
>  	struct fib_info *fi;
>  
> -	if (!net_eq(info->net, &init_net))
> -		return NOTIFY_DONE;

I don't see anymore uses of 'info->net'. Can it be removed from 'struct
fib_notifier_info' ?

^ permalink raw reply	[flat|nested] 34+ messages in thread

* Re: [patch net-next 03/15] net: fib_notifier: propagate possible error during fib notifier registration
  2019-09-14  6:45 ` [patch net-next 03/15] net: fib_notifier: propagate possible error during fib notifier registration Jiri Pirko
@ 2019-09-15  8:17   ` Ido Schimmel
  2019-09-15  9:41     ` Jiri Pirko
  0 siblings, 1 reply; 34+ messages in thread
From: Ido Schimmel @ 2019-09-15  8:17 UTC (permalink / raw)
  To: Jiri Pirko
  Cc: netdev, davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm,
	kuznet, yoshfuji, shuah, mlxsw

On Sat, Sep 14, 2019 at 08:45:56AM +0200, Jiri Pirko wrote:
> From: Jiri Pirko <jiri@mellanox.com>
> 
> Unlike events for registered notifier, during the registration, the
> errors that happened for the block being registered are not propagated
> up to the caller. For fib rules, this is already present, but not for

What do you mean by "already present" ? You added it below for rules as
well...

> fib entries. So make sure the error is propagated for those as well.
> 
> Signed-off-by: Jiri Pirko <jiri@mellanox.com>
> ---
>  include/net/ip_fib.h    |  2 +-
>  net/core/fib_notifier.c |  2 --
>  net/core/fib_rules.c    | 11 ++++++++---
>  net/ipv4/fib_notifier.c |  4 +---
>  net/ipv4/fib_trie.c     | 31 ++++++++++++++++++++++---------
>  net/ipv4/ipmr_base.c    | 22 +++++++++++++++-------
>  net/ipv6/ip6_fib.c      | 36 ++++++++++++++++++++++++------------
>  7 files changed, 71 insertions(+), 37 deletions(-)
> 
> diff --git a/include/net/ip_fib.h b/include/net/ip_fib.h
> index 4cec9ecaa95e..caae0fa610aa 100644
> --- a/include/net/ip_fib.h
> +++ b/include/net/ip_fib.h
> @@ -229,7 +229,7 @@ int __net_init fib4_notifier_init(struct net *net);
>  void __net_exit fib4_notifier_exit(struct net *net);
>  
>  void fib_info_notify_update(struct net *net, struct nl_info *info);
> -void fib_notify(struct net *net, struct notifier_block *nb);
> +int fib_notify(struct net *net, struct notifier_block *nb);
>  
>  struct fib_table {
>  	struct hlist_node	tb_hlist;
> diff --git a/net/core/fib_notifier.c b/net/core/fib_notifier.c
> index b965f3c0ec9a..fbd029425638 100644
> --- a/net/core/fib_notifier.c
> +++ b/net/core/fib_notifier.c
> @@ -65,8 +65,6 @@ static int fib_net_dump(struct net *net, struct notifier_block *nb)
>  
>  	rcu_read_lock();
>  	list_for_each_entry_rcu(ops, &fn_net->fib_notifier_ops, list) {
> -		int err;

Looks like this should have been removed in previous patch

> -
>  		if (!try_module_get(ops->owner))
>  			continue;
>  		err = ops->fib_dump(net, nb);
> diff --git a/net/core/fib_rules.c b/net/core/fib_rules.c
> index 28cbf07102bc..592d8aef90e3 100644
> --- a/net/core/fib_rules.c
> +++ b/net/core/fib_rules.c
> @@ -354,15 +354,20 @@ int fib_rules_dump(struct net *net, struct notifier_block *nb, int family)
>  {
>  	struct fib_rules_ops *ops;
>  	struct fib_rule *rule;
> +	int err = 0;
>  
>  	ops = lookup_rules_ops(net, family);
>  	if (!ops)
>  		return -EAFNOSUPPORT;
> -	list_for_each_entry_rcu(rule, &ops->rules_list, list)
> -		call_fib_rule_notifier(nb, FIB_EVENT_RULE_ADD, rule, family);
> +	list_for_each_entry_rcu(rule, &ops->rules_list, list) {
> +		err = call_fib_rule_notifier(nb, FIB_EVENT_RULE_ADD,
> +					     rule, family);

Here you add it for rules

> +		if (err)
> +			break;
> +	}
>  	rules_ops_put(ops);
>  
> -	return 0;
> +	return err;
>  }
>  EXPORT_SYMBOL_GPL(fib_rules_dump);

^ permalink raw reply	[flat|nested] 34+ messages in thread

* Re: [patch net-next 08/15] mlxsw: Register port netdevices into net of core
  2019-09-14  6:46 ` [patch net-next 08/15] mlxsw: Register port netdevices into net of core Jiri Pirko
@ 2019-09-15  8:37   ` Ido Schimmel
  0 siblings, 0 replies; 34+ messages in thread
From: Ido Schimmel @ 2019-09-15  8:37 UTC (permalink / raw)
  To: Jiri Pirko
  Cc: netdev, davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm,
	kuznet, yoshfuji, shuah, mlxsw

On Sat, Sep 14, 2019 at 08:46:01AM +0200, Jiri Pirko wrote:
> From: Jiri Pirko <jiri@mellanox.com>
> 
> When creating netdevices for ports, put then under network namespace

s/then/them/

> that the core/parent devlink belongs to.

^ permalink raw reply	[flat|nested] 34+ messages in thread

* Re: [patch net-next 09/15] mlxsw: Propagate extack down to register_fib_notifier()
  2019-09-14  6:46 ` [patch net-next 09/15] mlxsw: Propagate extack down to register_fib_notifier() Jiri Pirko
@ 2019-09-15  8:39   ` Ido Schimmel
  0 siblings, 0 replies; 34+ messages in thread
From: Ido Schimmel @ 2019-09-15  8:39 UTC (permalink / raw)
  To: Jiri Pirko
  Cc: netdev, davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm,
	kuznet, yoshfuji, shuah, mlxsw

On Sat, Sep 14, 2019 at 08:46:02AM +0200, Jiri Pirko wrote:
> From: Jiri Pirko <jiri@mellanox.com>
> 
> During the devlink reaload the extack is present, so propagate it all

s/reaload/reload/

> the way down to register_fib_notifier() call in spectrum_router.c.

^ permalink raw reply	[flat|nested] 34+ messages in thread

* Re: [patch net-next 14/15] net: devlink: allow to change namespaces during reload
  2019-09-14  6:46 ` [patch net-next 14/15] net: devlink: allow to change namespaces during reload Jiri Pirko
@ 2019-09-15  8:58   ` Ido Schimmel
  2019-09-15  9:43     ` Jiri Pirko
  0 siblings, 1 reply; 34+ messages in thread
From: Ido Schimmel @ 2019-09-15  8:58 UTC (permalink / raw)
  To: Jiri Pirko
  Cc: netdev, davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm,
	kuznet, yoshfuji, shuah, mlxsw

On Sat, Sep 14, 2019 at 08:46:07AM +0200, Jiri Pirko wrote:
> From: Jiri Pirko <jiri@mellanox.com>
> 
> All devlink instances are created in init_net and stay there for a
> lifetime. Allow user to be able to move devlink instances into
> namespaces during devlink reload operation. That ensures proper
> re-instantiation of driver objects, including netdevices.
> 
> Signed-off-by: Jiri Pirko <jiri@mellanox.com>
> ---
>  drivers/net/ethernet/mellanox/mlx4/main.c |   4 +
>  include/uapi/linux/devlink.h              |   4 +
>  net/core/devlink.c                        | 155 ++++++++++++++++++++--
>  3 files changed, 155 insertions(+), 8 deletions(-)
> 
> diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c
> index ef3f3d06ff1e..989d0882aaa9 100644
> --- a/drivers/net/ethernet/mellanox/mlx4/main.c
> +++ b/drivers/net/ethernet/mellanox/mlx4/main.c
> @@ -3942,6 +3942,10 @@ static int mlx4_devlink_reload_down(struct devlink *devlink,
>  	struct mlx4_dev *dev = &priv->dev;
>  	struct mlx4_dev_persistent *persist = dev->persist;
>  
> +	if (!net_eq(devlink_net(devlink), &init_net)) {
> +		NL_SET_ERR_MSG_MOD(extack, "Namespace change is not supported");
> +		return -EOPNOTSUPP;
> +	}

Are you sure that this actually works? I see that you first invoke
reload_down(), then set the new namespace, then invoke reload_up().

So shouldn't this check be done in reload_up() callback instead?

>  	if (persist->num_vfs)
>  		mlx4_warn(persist->dev, "Reload performed on PF, will cause reset on operating Virtual Functions\n");
>  	mlx4_restart_one_down(persist->pdev);
> diff --git a/include/uapi/linux/devlink.h b/include/uapi/linux/devlink.h
> index 580b7a2e40e1..b558ea88b766 100644
> --- a/include/uapi/linux/devlink.h
> +++ b/include/uapi/linux/devlink.h
> @@ -421,6 +421,10 @@ enum devlink_attr {
>  
>  	DEVLINK_ATTR_RELOAD_FAILED,			/* u8 0 or 1 */
>  
> +	DEVLINK_ATTR_NETNS_FD,			/* u32 */
> +	DEVLINK_ATTR_NETNS_PID,			/* u32 */
> +	DEVLINK_ATTR_NETNS_ID,			/* u32 */
> +
>  	/* add new attributes above here, update the policy in devlink.c */
>  
>  	__DEVLINK_ATTR_MAX,
> diff --git a/net/core/devlink.c b/net/core/devlink.c
> index 362cbbcca225..2a5db95cce3c 100644
> --- a/net/core/devlink.c
> +++ b/net/core/devlink.c
> @@ -435,8 +435,16 @@ static void devlink_nl_post_doit(const struct genl_ops *ops,
>  {
>  	struct devlink *devlink;
>  
> -	devlink = devlink_get_from_info(info);
> -	if (~ops->internal_flags & DEVLINK_NL_FLAG_NO_LOCK)
> +	/* When devlink changes netns, it would not be found
> +	 * by devlink_get_from_info(). So try if it is stored first.
> +	 */
> +	if (ops->internal_flags & DEVLINK_NL_FLAG_NEED_DEVLINK) {
> +		devlink = info->user_ptr[0];
> +	} else {
> +		devlink = devlink_get_from_info(info);
> +		WARN_ON(IS_ERR(devlink));
> +	}
> +	if (!IS_ERR(devlink) && ~ops->internal_flags & DEVLINK_NL_FLAG_NO_LOCK)
>  		mutex_unlock(&devlink->lock);
>  	mutex_unlock(&devlink_mutex);
>  }
> @@ -2675,6 +2683,73 @@ devlink_resources_validate(struct devlink *devlink,
>  	return err;
>  }
>  
> +static struct net *devlink_netns_get(struct sk_buff *skb,
> +				     struct devlink *devlink,
> +				     struct genl_info *info)
> +{
> +	struct nlattr *netns_pid_attr = info->attrs[DEVLINK_ATTR_NETNS_PID];
> +	struct nlattr *netns_fd_attr = info->attrs[DEVLINK_ATTR_NETNS_FD];
> +	struct nlattr *netns_id_attr = info->attrs[DEVLINK_ATTR_NETNS_ID];
> +	struct net *net;
> +
> +	if (!!netns_pid_attr + !!netns_fd_attr + !!netns_id_attr > 1) {
> +		NL_SET_ERR_MSG(info->extack, "multiple netns identifying attributes specified");
> +		return ERR_PTR(-EINVAL);
> +	}
> +
> +	if (netns_pid_attr) {
> +		net = get_net_ns_by_pid(nla_get_u32(netns_pid_attr));
> +	} else if (netns_fd_attr) {
> +		net = get_net_ns_by_fd(nla_get_u32(netns_fd_attr));
> +	} else if (netns_id_attr) {
> +		net = get_net_ns_by_id(sock_net(skb->sk),
> +				       nla_get_u32(netns_id_attr));
> +		if (!net)
> +			net = ERR_PTR(-EINVAL);
> +	} else {
> +		WARN_ON(1);
> +		net = ERR_PTR(-EINVAL);
> +	}
> +	if (IS_ERR(net)) {
> +		NL_SET_ERR_MSG(info->extack, "Unknown network namespace");
> +		return ERR_PTR(-EINVAL);
> +	}
> +	if (!netlink_ns_capable(skb, net->user_ns, CAP_NET_ADMIN)) {
> +		put_net(net);
> +		return ERR_PTR(-EPERM);
> +	}
> +	return net;
> +}
> +
> +static void devlink_param_notify(struct devlink *devlink,
> +				 unsigned int port_index,
> +				 struct devlink_param_item *param_item,
> +				 enum devlink_command cmd);
> +
> +static void devlink_reload_netns_change(struct devlink *devlink,
> +					struct net *dest_net)
> +{
> +	struct devlink_param_item *param_item;
> +
> +	/* Userspace needs to be notified about devlink objects
> +	 * removed from original and entering new network namespace.
> +	 * The rest of the devlink objects are re-created during
> +	 * reload process so the notifications are generated separatelly.
> +	 */
> +
> +	list_for_each_entry(param_item, &devlink->param_list, list)
> +		devlink_param_notify(devlink, 0, param_item,
> +				     DEVLINK_CMD_PARAM_DEL);
> +	devlink_notify(devlink, DEVLINK_CMD_DEL);
> +
> +	devlink_net_set(devlink, dest_net);
> +
> +	devlink_notify(devlink, DEVLINK_CMD_NEW);
> +	list_for_each_entry(param_item, &devlink->param_list, list)
> +		devlink_param_notify(devlink, 0, param_item,
> +				     DEVLINK_CMD_PARAM_NEW);
> +}
> +
>  static bool devlink_reload_supported(struct devlink *devlink)
>  {
>  	return devlink->ops->reload_down && devlink->ops->reload_up;
> @@ -2695,9 +2770,27 @@ bool devlink_is_reload_failed(const struct devlink *devlink)
>  }
>  EXPORT_SYMBOL_GPL(devlink_is_reload_failed);
>  
> +static int devlink_reload(struct devlink *devlink, struct net *dest_net,
> +			  struct netlink_ext_ack *extack)
> +{
> +	int err;
> +
> +	err = devlink->ops->reload_down(devlink, extack);
> +	if (err)
> +		return err;
> +
> +	if (dest_net && !net_eq(dest_net, devlink_net(devlink)))
> +		devlink_reload_netns_change(devlink, dest_net);
> +
> +	err = devlink->ops->reload_up(devlink, extack);
> +	devlink_reload_failed_set(devlink, !!err);
> +	return err;
> +}
> +
>  static int devlink_nl_cmd_reload(struct sk_buff *skb, struct genl_info *info)
>  {
>  	struct devlink *devlink = info->user_ptr[0];
> +	struct net *dest_net = NULL;
>  	int err;
>  
>  	if (!devlink_reload_supported(devlink))
> @@ -2708,11 +2801,20 @@ static int devlink_nl_cmd_reload(struct sk_buff *skb, struct genl_info *info)
>  		NL_SET_ERR_MSG_MOD(info->extack, "resources size validation failed");
>  		return err;
>  	}
> -	err = devlink->ops->reload_down(devlink, info->extack);
> -	if (err)
> -		return err;
> -	err = devlink->ops->reload_up(devlink, info->extack);
> -	devlink_reload_failed_set(devlink, !!err);
> +
> +	if (info->attrs[DEVLINK_ATTR_NETNS_PID] ||
> +	    info->attrs[DEVLINK_ATTR_NETNS_FD] ||
> +	    info->attrs[DEVLINK_ATTR_NETNS_ID]) {
> +		dest_net = devlink_netns_get(skb, devlink, info);

Hmm, you're never using 'devlink' there, so I guess you can drop it.

> +		if (IS_ERR(dest_net))
> +			return PTR_ERR(dest_net);
> +	}
> +
> +	err = devlink_reload(devlink, dest_net, info->extack);
> +
> +	if (dest_net)
> +		put_net(dest_net);
> +
>  	return err;
>  }
>  
> @@ -5794,6 +5896,9 @@ static const struct nla_policy devlink_nl_policy[DEVLINK_ATTR_MAX + 1] = {
>  	[DEVLINK_ATTR_TRAP_NAME] = { .type = NLA_NUL_STRING },
>  	[DEVLINK_ATTR_TRAP_ACTION] = { .type = NLA_U8 },
>  	[DEVLINK_ATTR_TRAP_GROUP_NAME] = { .type = NLA_NUL_STRING },
> +	[DEVLINK_ATTR_NETNS_PID] = { .type = NLA_U32 },
> +	[DEVLINK_ATTR_NETNS_FD] = { .type = NLA_U32 },
> +	[DEVLINK_ATTR_NETNS_ID] = { .type = NLA_U32 },
>  };
>  
>  static const struct genl_ops devlink_nl_ops[] = {
> @@ -8061,9 +8166,43 @@ int devlink_compat_switch_id_get(struct net_device *dev,
>  	return 0;
>  }
>  
> +static void __net_exit devlink_pernet_pre_exit(struct net *net)
> +{
> +	struct devlink *devlink;
> +	int err;
> +
> +	/* In case network namespace is getting destroyed, reload
> +	 * all devlink instances from this namespace into init_net.
> +	 */
> +	mutex_lock(&devlink_mutex);
> +	list_for_each_entry(devlink, &devlink_list, list) {
> +		if (net_eq(devlink_net(devlink), net)) {
> +			if (WARN_ON(!devlink_reload_supported(devlink)))
> +				continue;
> +			err = devlink_reload(devlink, &init_net, NULL);
> +			if (err)
> +				pr_warn("Failed to reload devlink instance into init_net\n");
> +		}
> +	}
> +	mutex_unlock(&devlink_mutex);
> +}
> +
> +static struct pernet_operations devlink_pernet_ops __net_initdata = {
> +	.pre_exit = devlink_pernet_pre_exit,
> +};
> +
>  static int __init devlink_init(void)
>  {
> -	return genl_register_family(&devlink_nl_family);
> +	int err;
> +
> +	err = genl_register_family(&devlink_nl_family);
> +	if (err)
> +		goto out;
> +	err = register_pernet_subsys(&devlink_pernet_ops);
> +
> +out:
> +	WARN_ON(err);
> +	return err;
>  }
>  
>  subsys_initcall(devlink_init);
> -- 
> 2.21.0
> 

^ permalink raw reply	[flat|nested] 34+ messages in thread

* Re: [patch net-next 02/15] net: fib_notifier: make FIB notifier per-netns
  2019-09-15  8:06   ` Ido Schimmel
@ 2019-09-15  9:37     ` Jiri Pirko
  0 siblings, 0 replies; 34+ messages in thread
From: Jiri Pirko @ 2019-09-15  9:37 UTC (permalink / raw)
  To: Ido Schimmel
  Cc: netdev, davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm,
	kuznet, yoshfuji, shuah, mlxsw

Sun, Sep 15, 2019 at 10:06:02AM CEST, idosch@idosch.org wrote:
>On Sat, Sep 14, 2019 at 08:45:55AM +0200, Jiri Pirko wrote:
>> From: Jiri Pirko <jiri@mellanox.com>
>> 
>> Currently all users of FIB notifier only cares about events in init_net.
>
>s/cares/care/

ok


>
>> Later in this patchset, users get interested in other namespaces too.
>> However, for every registered block user is interested only about one
>> namespace. Make the FIB notifier registration per-netns and avoid
>> unnecessary calls of notifier block for other namespaces.
>
>...
>
>> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c b/drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c
>> index 5d20d615663e..fe0cc969cf94 100644
>> --- a/drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c
>> +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c
>> @@ -248,9 +248,6 @@ static int mlx5_lag_fib_event(struct notifier_block *nb,
>>  	struct net_device *fib_dev;
>>  	struct fib_info *fi;
>>  
>> -	if (!net_eq(info->net, &init_net))
>> -		return NOTIFY_DONE;
>
>I don't see anymore uses of 'info->net'. Can it be removed from 'struct
>fib_notifier_info' ?

correct. I missed that.


^ permalink raw reply	[flat|nested] 34+ messages in thread

* Re: [patch net-next 03/15] net: fib_notifier: propagate possible error during fib notifier registration
  2019-09-15  8:17   ` Ido Schimmel
@ 2019-09-15  9:41     ` Jiri Pirko
  0 siblings, 0 replies; 34+ messages in thread
From: Jiri Pirko @ 2019-09-15  9:41 UTC (permalink / raw)
  To: Ido Schimmel
  Cc: netdev, davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm,
	kuznet, yoshfuji, shuah, mlxsw

Sun, Sep 15, 2019 at 10:17:46AM CEST, idosch@idosch.org wrote:
>On Sat, Sep 14, 2019 at 08:45:56AM +0200, Jiri Pirko wrote:
>> From: Jiri Pirko <jiri@mellanox.com>
>> 
>> Unlike events for registered notifier, during the registration, the
>> errors that happened for the block being registered are not propagated
>> up to the caller. For fib rules, this is already present, but not for
>
>What do you mean by "already present" ? You added it below for rules as
>well...

Right, will fix.


>
>> fib entries. So make sure the error is propagated for those as well.
>> 
>> Signed-off-by: Jiri Pirko <jiri@mellanox.com>
>> ---
>>  include/net/ip_fib.h    |  2 +-
>>  net/core/fib_notifier.c |  2 --
>>  net/core/fib_rules.c    | 11 ++++++++---
>>  net/ipv4/fib_notifier.c |  4 +---
>>  net/ipv4/fib_trie.c     | 31 ++++++++++++++++++++++---------
>>  net/ipv4/ipmr_base.c    | 22 +++++++++++++++-------
>>  net/ipv6/ip6_fib.c      | 36 ++++++++++++++++++++++++------------
>>  7 files changed, 71 insertions(+), 37 deletions(-)
>> 
>> diff --git a/include/net/ip_fib.h b/include/net/ip_fib.h
>> index 4cec9ecaa95e..caae0fa610aa 100644
>> --- a/include/net/ip_fib.h
>> +++ b/include/net/ip_fib.h
>> @@ -229,7 +229,7 @@ int __net_init fib4_notifier_init(struct net *net);
>>  void __net_exit fib4_notifier_exit(struct net *net);
>>  
>>  void fib_info_notify_update(struct net *net, struct nl_info *info);
>> -void fib_notify(struct net *net, struct notifier_block *nb);
>> +int fib_notify(struct net *net, struct notifier_block *nb);
>>  
>>  struct fib_table {
>>  	struct hlist_node	tb_hlist;
>> diff --git a/net/core/fib_notifier.c b/net/core/fib_notifier.c
>> index b965f3c0ec9a..fbd029425638 100644
>> --- a/net/core/fib_notifier.c
>> +++ b/net/core/fib_notifier.c
>> @@ -65,8 +65,6 @@ static int fib_net_dump(struct net *net, struct notifier_block *nb)
>>  
>>  	rcu_read_lock();
>>  	list_for_each_entry_rcu(ops, &fn_net->fib_notifier_ops, list) {
>> -		int err;
>
>Looks like this should have been removed in previous patch

Correct, will move.


>
>> -
>>  		if (!try_module_get(ops->owner))
>>  			continue;
>>  		err = ops->fib_dump(net, nb);
>> diff --git a/net/core/fib_rules.c b/net/core/fib_rules.c
>> index 28cbf07102bc..592d8aef90e3 100644
>> --- a/net/core/fib_rules.c
>> +++ b/net/core/fib_rules.c
>> @@ -354,15 +354,20 @@ int fib_rules_dump(struct net *net, struct notifier_block *nb, int family)
>>  {
>>  	struct fib_rules_ops *ops;
>>  	struct fib_rule *rule;
>> +	int err = 0;
>>  
>>  	ops = lookup_rules_ops(net, family);
>>  	if (!ops)
>>  		return -EAFNOSUPPORT;
>> -	list_for_each_entry_rcu(rule, &ops->rules_list, list)
>> -		call_fib_rule_notifier(nb, FIB_EVENT_RULE_ADD, rule, family);
>> +	list_for_each_entry_rcu(rule, &ops->rules_list, list) {
>> +		err = call_fib_rule_notifier(nb, FIB_EVENT_RULE_ADD,
>> +					     rule, family);
>
>Here you add it for rules
>
>> +		if (err)
>> +			break;
>> +	}
>>  	rules_ops_put(ops);
>>  
>> -	return 0;
>> +	return err;
>>  }
>>  EXPORT_SYMBOL_GPL(fib_rules_dump);

^ permalink raw reply	[flat|nested] 34+ messages in thread

* Re: [patch net-next 14/15] net: devlink: allow to change namespaces during reload
  2019-09-15  8:58   ` Ido Schimmel
@ 2019-09-15  9:43     ` Jiri Pirko
  0 siblings, 0 replies; 34+ messages in thread
From: Jiri Pirko @ 2019-09-15  9:43 UTC (permalink / raw)
  To: Ido Schimmel
  Cc: netdev, davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm,
	kuznet, yoshfuji, shuah, mlxsw

Sun, Sep 15, 2019 at 10:58:29AM CEST, idosch@idosch.org wrote:
>On Sat, Sep 14, 2019 at 08:46:07AM +0200, Jiri Pirko wrote:
>> From: Jiri Pirko <jiri@mellanox.com>
>> 
>> All devlink instances are created in init_net and stay there for a
>> lifetime. Allow user to be able to move devlink instances into
>> namespaces during devlink reload operation. That ensures proper
>> re-instantiation of driver objects, including netdevices.
>> 
>> Signed-off-by: Jiri Pirko <jiri@mellanox.com>
>> ---
>>  drivers/net/ethernet/mellanox/mlx4/main.c |   4 +
>>  include/uapi/linux/devlink.h              |   4 +
>>  net/core/devlink.c                        | 155 ++++++++++++++++++++--
>>  3 files changed, 155 insertions(+), 8 deletions(-)
>> 
>> diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c
>> index ef3f3d06ff1e..989d0882aaa9 100644
>> --- a/drivers/net/ethernet/mellanox/mlx4/main.c
>> +++ b/drivers/net/ethernet/mellanox/mlx4/main.c
>> @@ -3942,6 +3942,10 @@ static int mlx4_devlink_reload_down(struct devlink *devlink,
>>  	struct mlx4_dev *dev = &priv->dev;
>>  	struct mlx4_dev_persistent *persist = dev->persist;
>>  
>> +	if (!net_eq(devlink_net(devlink), &init_net)) {
>> +		NL_SET_ERR_MSG_MOD(extack, "Namespace change is not supported");
>> +		return -EOPNOTSUPP;
>> +	}
>
>Are you sure that this actually works? I see that you first invoke
>reload_down(), then set the new namespace, then invoke reload_up().
>
>So shouldn't this check be done in reload_up() callback instead?

Correct, need to fix this. But I need to do this in down phase so the
objects are not removed.


>
>>  	if (persist->num_vfs)
>>  		mlx4_warn(persist->dev, "Reload performed on PF, will cause reset on operating Virtual Functions\n");
>>  	mlx4_restart_one_down(persist->pdev);
>> diff --git a/include/uapi/linux/devlink.h b/include/uapi/linux/devlink.h
>> index 580b7a2e40e1..b558ea88b766 100644
>> --- a/include/uapi/linux/devlink.h
>> +++ b/include/uapi/linux/devlink.h
>> @@ -421,6 +421,10 @@ enum devlink_attr {
>>  
>>  	DEVLINK_ATTR_RELOAD_FAILED,			/* u8 0 or 1 */
>>  
>> +	DEVLINK_ATTR_NETNS_FD,			/* u32 */
>> +	DEVLINK_ATTR_NETNS_PID,			/* u32 */
>> +	DEVLINK_ATTR_NETNS_ID,			/* u32 */
>> +
>>  	/* add new attributes above here, update the policy in devlink.c */
>>  
>>  	__DEVLINK_ATTR_MAX,
>> diff --git a/net/core/devlink.c b/net/core/devlink.c
>> index 362cbbcca225..2a5db95cce3c 100644
>> --- a/net/core/devlink.c
>> +++ b/net/core/devlink.c
>> @@ -435,8 +435,16 @@ static void devlink_nl_post_doit(const struct genl_ops *ops,
>>  {
>>  	struct devlink *devlink;
>>  
>> -	devlink = devlink_get_from_info(info);
>> -	if (~ops->internal_flags & DEVLINK_NL_FLAG_NO_LOCK)
>> +	/* When devlink changes netns, it would not be found
>> +	 * by devlink_get_from_info(). So try if it is stored first.
>> +	 */
>> +	if (ops->internal_flags & DEVLINK_NL_FLAG_NEED_DEVLINK) {
>> +		devlink = info->user_ptr[0];
>> +	} else {
>> +		devlink = devlink_get_from_info(info);
>> +		WARN_ON(IS_ERR(devlink));
>> +	}
>> +	if (!IS_ERR(devlink) && ~ops->internal_flags & DEVLINK_NL_FLAG_NO_LOCK)
>>  		mutex_unlock(&devlink->lock);
>>  	mutex_unlock(&devlink_mutex);
>>  }
>> @@ -2675,6 +2683,73 @@ devlink_resources_validate(struct devlink *devlink,
>>  	return err;
>>  }
>>  
>> +static struct net *devlink_netns_get(struct sk_buff *skb,
>> +				     struct devlink *devlink,
>> +				     struct genl_info *info)
>> +{
>> +	struct nlattr *netns_pid_attr = info->attrs[DEVLINK_ATTR_NETNS_PID];
>> +	struct nlattr *netns_fd_attr = info->attrs[DEVLINK_ATTR_NETNS_FD];
>> +	struct nlattr *netns_id_attr = info->attrs[DEVLINK_ATTR_NETNS_ID];
>> +	struct net *net;
>> +
>> +	if (!!netns_pid_attr + !!netns_fd_attr + !!netns_id_attr > 1) {
>> +		NL_SET_ERR_MSG(info->extack, "multiple netns identifying attributes specified");
>> +		return ERR_PTR(-EINVAL);
>> +	}
>> +
>> +	if (netns_pid_attr) {
>> +		net = get_net_ns_by_pid(nla_get_u32(netns_pid_attr));
>> +	} else if (netns_fd_attr) {
>> +		net = get_net_ns_by_fd(nla_get_u32(netns_fd_attr));
>> +	} else if (netns_id_attr) {
>> +		net = get_net_ns_by_id(sock_net(skb->sk),
>> +				       nla_get_u32(netns_id_attr));
>> +		if (!net)
>> +			net = ERR_PTR(-EINVAL);
>> +	} else {
>> +		WARN_ON(1);
>> +		net = ERR_PTR(-EINVAL);
>> +	}
>> +	if (IS_ERR(net)) {
>> +		NL_SET_ERR_MSG(info->extack, "Unknown network namespace");
>> +		return ERR_PTR(-EINVAL);
>> +	}
>> +	if (!netlink_ns_capable(skb, net->user_ns, CAP_NET_ADMIN)) {
>> +		put_net(net);
>> +		return ERR_PTR(-EPERM);
>> +	}
>> +	return net;
>> +}
>> +
>> +static void devlink_param_notify(struct devlink *devlink,
>> +				 unsigned int port_index,
>> +				 struct devlink_param_item *param_item,
>> +				 enum devlink_command cmd);
>> +
>> +static void devlink_reload_netns_change(struct devlink *devlink,
>> +					struct net *dest_net)
>> +{
>> +	struct devlink_param_item *param_item;
>> +
>> +	/* Userspace needs to be notified about devlink objects
>> +	 * removed from original and entering new network namespace.
>> +	 * The rest of the devlink objects are re-created during
>> +	 * reload process so the notifications are generated separatelly.
>> +	 */
>> +
>> +	list_for_each_entry(param_item, &devlink->param_list, list)
>> +		devlink_param_notify(devlink, 0, param_item,
>> +				     DEVLINK_CMD_PARAM_DEL);
>> +	devlink_notify(devlink, DEVLINK_CMD_DEL);
>> +
>> +	devlink_net_set(devlink, dest_net);
>> +
>> +	devlink_notify(devlink, DEVLINK_CMD_NEW);
>> +	list_for_each_entry(param_item, &devlink->param_list, list)
>> +		devlink_param_notify(devlink, 0, param_item,
>> +				     DEVLINK_CMD_PARAM_NEW);
>> +}
>> +
>>  static bool devlink_reload_supported(struct devlink *devlink)
>>  {
>>  	return devlink->ops->reload_down && devlink->ops->reload_up;
>> @@ -2695,9 +2770,27 @@ bool devlink_is_reload_failed(const struct devlink *devlink)
>>  }
>>  EXPORT_SYMBOL_GPL(devlink_is_reload_failed);
>>  
>> +static int devlink_reload(struct devlink *devlink, struct net *dest_net,
>> +			  struct netlink_ext_ack *extack)
>> +{
>> +	int err;
>> +
>> +	err = devlink->ops->reload_down(devlink, extack);
>> +	if (err)
>> +		return err;
>> +
>> +	if (dest_net && !net_eq(dest_net, devlink_net(devlink)))
>> +		devlink_reload_netns_change(devlink, dest_net);
>> +
>> +	err = devlink->ops->reload_up(devlink, extack);
>> +	devlink_reload_failed_set(devlink, !!err);
>> +	return err;
>> +}
>> +
>>  static int devlink_nl_cmd_reload(struct sk_buff *skb, struct genl_info *info)
>>  {
>>  	struct devlink *devlink = info->user_ptr[0];
>> +	struct net *dest_net = NULL;
>>  	int err;
>>  
>>  	if (!devlink_reload_supported(devlink))
>> @@ -2708,11 +2801,20 @@ static int devlink_nl_cmd_reload(struct sk_buff *skb, struct genl_info *info)
>>  		NL_SET_ERR_MSG_MOD(info->extack, "resources size validation failed");
>>  		return err;
>>  	}
>> -	err = devlink->ops->reload_down(devlink, info->extack);
>> -	if (err)
>> -		return err;
>> -	err = devlink->ops->reload_up(devlink, info->extack);
>> -	devlink_reload_failed_set(devlink, !!err);
>> +
>> +	if (info->attrs[DEVLINK_ATTR_NETNS_PID] ||
>> +	    info->attrs[DEVLINK_ATTR_NETNS_FD] ||
>> +	    info->attrs[DEVLINK_ATTR_NETNS_ID]) {
>> +		dest_net = devlink_netns_get(skb, devlink, info);
>
>Hmm, you're never using 'devlink' there, so I guess you can drop it.

Right, will do.


>
>> +		if (IS_ERR(dest_net))
>> +			return PTR_ERR(dest_net);
>> +	}
>> +
>> +	err = devlink_reload(devlink, dest_net, info->extack);
>> +
>> +	if (dest_net)
>> +		put_net(dest_net);
>> +
>>  	return err;
>>  }
>>  
>> @@ -5794,6 +5896,9 @@ static const struct nla_policy devlink_nl_policy[DEVLINK_ATTR_MAX + 1] = {
>>  	[DEVLINK_ATTR_TRAP_NAME] = { .type = NLA_NUL_STRING },
>>  	[DEVLINK_ATTR_TRAP_ACTION] = { .type = NLA_U8 },
>>  	[DEVLINK_ATTR_TRAP_GROUP_NAME] = { .type = NLA_NUL_STRING },
>> +	[DEVLINK_ATTR_NETNS_PID] = { .type = NLA_U32 },
>> +	[DEVLINK_ATTR_NETNS_FD] = { .type = NLA_U32 },
>> +	[DEVLINK_ATTR_NETNS_ID] = { .type = NLA_U32 },
>>  };
>>  
>>  static const struct genl_ops devlink_nl_ops[] = {
>> @@ -8061,9 +8166,43 @@ int devlink_compat_switch_id_get(struct net_device *dev,
>>  	return 0;
>>  }
>>  
>> +static void __net_exit devlink_pernet_pre_exit(struct net *net)
>> +{
>> +	struct devlink *devlink;
>> +	int err;
>> +
>> +	/* In case network namespace is getting destroyed, reload
>> +	 * all devlink instances from this namespace into init_net.
>> +	 */
>> +	mutex_lock(&devlink_mutex);
>> +	list_for_each_entry(devlink, &devlink_list, list) {
>> +		if (net_eq(devlink_net(devlink), net)) {
>> +			if (WARN_ON(!devlink_reload_supported(devlink)))
>> +				continue;
>> +			err = devlink_reload(devlink, &init_net, NULL);
>> +			if (err)
>> +				pr_warn("Failed to reload devlink instance into init_net\n");
>> +		}
>> +	}
>> +	mutex_unlock(&devlink_mutex);
>> +}
>> +
>> +static struct pernet_operations devlink_pernet_ops __net_initdata = {
>> +	.pre_exit = devlink_pernet_pre_exit,
>> +};
>> +
>>  static int __init devlink_init(void)
>>  {
>> -	return genl_register_family(&devlink_nl_family);
>> +	int err;
>> +
>> +	err = genl_register_family(&devlink_nl_family);
>> +	if (err)
>> +		goto out;
>> +	err = register_pernet_subsys(&devlink_pernet_ops);
>> +
>> +out:
>> +	WARN_ON(err);
>> +	return err;
>>  }
>>  
>>  subsys_initcall(devlink_init);
>> -- 
>> 2.21.0
>> 

^ permalink raw reply	[flat|nested] 34+ messages in thread

* Re: [patch iproute2-next 2/2] devlink: extend reload command to add support for network namespace change
  2019-09-15  7:16   ` Ido Schimmel
@ 2019-09-15  9:44     ` Jiri Pirko
  0 siblings, 0 replies; 34+ messages in thread
From: Jiri Pirko @ 2019-09-15  9:44 UTC (permalink / raw)
  To: Ido Schimmel
  Cc: netdev, davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm,
	kuznet, yoshfuji, shuah, mlxsw

Sun, Sep 15, 2019 at 09:16:39AM CEST, idosch@idosch.org wrote:
>On Sat, Sep 14, 2019 at 08:57:57AM +0200, Jiri Pirko wrote:
>> diff --git a/man/man8/devlink-dev.8 b/man/man8/devlink-dev.8
>> index 1804463b2321..0e1a5523fa7b 100644
>> --- a/man/man8/devlink-dev.8
>> +++ b/man/man8/devlink-dev.8
>> @@ -25,6 +25,13 @@ devlink-dev \- devlink device configuration
>>  .ti -8
>>  .B devlink dev help
>>  
>> +.ti -8
>> +.BR "devlink dev set"
>> +.IR DEV
>> +.RI "[ "
>> +.BI "netns { " PID " | " NAME " | " ID " }
>> +.RI "]"
>> +
>>  .ti -8
>>  .BR "devlink dev eswitch set"
>>  .IR DEV
>> @@ -92,6 +99,11 @@ Format is:
>>  .in +2
>>  BUS_NAME/BUS_ADDRESS
>>  
>> +.SS devlink dev set  - sets devlink device attributes
>> +
>> +.TP
>> +.BI "netns { " PID " | " NAME " | " ID " }
>
>This looks like leftover from previous version?

Will fix. Thanks!

>
>> +
>>  .SS devlink dev eswitch show - display devlink device eswitch attributes
>>  .SS devlink dev eswitch set  - sets devlink device eswitch attributes
>>  
>> -- 
>> 2.21.0
>> 

^ permalink raw reply	[flat|nested] 34+ messages in thread

* Re: [patch iproute2-next 1/2] devlink: introduce cmdline option to switch to a different namespace
  2019-09-14  6:57 ` [patch iproute2-next 1/2] devlink: introduce cmdline option to switch to a different namespace Jiri Pirko
@ 2019-09-15 18:01   ` David Ahern
  0 siblings, 0 replies; 34+ messages in thread
From: David Ahern @ 2019-09-15 18:01 UTC (permalink / raw)
  To: Jiri Pirko, netdev
  Cc: davem, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

On 9/14/19 12:57 AM, Jiri Pirko wrote:
> From: Jiri Pirko <jiri@mellanox.com>
> 

Both of these patches have no commit logs. Please add - e.g., why -N
versus -n that other commands use.

> Signed-off-by: Jiri Pirko <jiri@mellanox.com>
> ---
> v3->v4:
> - rebased on top of trap patches
> ---
>  devlink/devlink.c  | 12 ++++++++++--
>  man/man8/devlink.8 |  4 ++++
>  2 files changed, 14 insertions(+), 2 deletions(-)
> 




^ permalink raw reply	[flat|nested] 34+ messages in thread

* Re: [patch net-next 02/15] net: fib_notifier: make FIB notifier per-netns
  2019-09-14  6:45 ` [patch net-next 02/15] net: fib_notifier: make FIB notifier per-netns Jiri Pirko
  2019-09-15  8:06   ` Ido Schimmel
@ 2019-09-15 20:05   ` David Ahern
  2019-09-16  5:38     ` Jiri Pirko
  1 sibling, 1 reply; 34+ messages in thread
From: David Ahern @ 2019-09-15 20:05 UTC (permalink / raw)
  To: Jiri Pirko, netdev
  Cc: davem, idosch, jakub.kicinski, tariqt, saeedm, kuznet, yoshfuji,
	shuah, mlxsw

On 9/14/19 12:45 AM, Jiri Pirko wrote:
>  #define FIB_DUMP_MAX_RETRIES 5
> -int register_fib_notifier(struct notifier_block *nb,
> +int register_fib_notifier(struct net *net, struct notifier_block *nb,
>  			  void (*cb)(struct notifier_block *nb))
>  {
>  	int retries = 0;
>  	int err;
>  
>  	do {
> -		unsigned int fib_seq = fib_seq_sum();
> -		struct net *net;
> -
> -		rcu_read_lock();
> -		for_each_net_rcu(net) {
> -			err = fib_net_dump(net, nb);
> -			if (err)
> -				goto err_fib_net_dump;
> -		}
> -		rcu_read_unlock();
> -
> -		if (fib_dump_is_consistent(nb, cb, fib_seq))
> +		unsigned int fib_seq = fib_seq_sum(net);
> +
> +		err = fib_net_dump(net, nb);
> +		if (err)
> +			return err;
> +
> +		if (fib_dump_is_consistent(net, nb, cb, fib_seq))
>  			return 0;
>  	} while (++retries < FIB_DUMP_MAX_RETRIES);

This is still more complicated than it needs to be. Why lump all
fib_notifier_ops into 1 dump when they are separate databases with
separate seq numbers? Just dump them 1 at a time and retry that 1
database as needed.

ie., This:
    list_for_each_entry_rcu(ops, &net->fib_notifier_ops, list) {
should be in register_fib_notifier and not fib_net_dump.

as it stands you are potentially replaying way more than is needed when
a dump is inconsistent.


>  
>  	return -EBUSY;
> -
> -err_fib_net_dump:
> -	rcu_read_unlock();
> -	return err;
>  }
>  EXPORT_SYMBOL(register_fib_notifier);
>  
> -int unregister_fib_notifier(struct notifier_block *nb)
> +int unregister_fib_notifier(struct net *net, struct notifier_block *nb)
>  {
> -	return atomic_notifier_chain_unregister(&fib_chain, nb);
> +	struct fib_notifier_net *fn_net = net_generic(net, fib_notifier_net_id);
> +
> +	return atomic_notifier_chain_unregister(&fn_net->fib_chain, nb);
>  }
>  EXPORT_SYMBOL(unregister_fib_notifier);
>  






^ permalink raw reply	[flat|nested] 34+ messages in thread

* Re: [patch net-next 02/15] net: fib_notifier: make FIB notifier per-netns
  2019-09-15 20:05   ` David Ahern
@ 2019-09-16  5:38     ` Jiri Pirko
  2019-09-16 14:54       ` David Ahern
  0 siblings, 1 reply; 34+ messages in thread
From: Jiri Pirko @ 2019-09-16  5:38 UTC (permalink / raw)
  To: David Ahern
  Cc: netdev, davem, idosch, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

Sun, Sep 15, 2019 at 10:05:47PM CEST, dsahern@gmail.com wrote:
>On 9/14/19 12:45 AM, Jiri Pirko wrote:
>>  #define FIB_DUMP_MAX_RETRIES 5
>> -int register_fib_notifier(struct notifier_block *nb,
>> +int register_fib_notifier(struct net *net, struct notifier_block *nb,
>>  			  void (*cb)(struct notifier_block *nb))
>>  {
>>  	int retries = 0;
>>  	int err;
>>  
>>  	do {
>> -		unsigned int fib_seq = fib_seq_sum();
>> -		struct net *net;
>> -
>> -		rcu_read_lock();
>> -		for_each_net_rcu(net) {
>> -			err = fib_net_dump(net, nb);
>> -			if (err)
>> -				goto err_fib_net_dump;
>> -		}
>> -		rcu_read_unlock();
>> -
>> -		if (fib_dump_is_consistent(nb, cb, fib_seq))
>> +		unsigned int fib_seq = fib_seq_sum(net);
>> +
>> +		err = fib_net_dump(net, nb);
>> +		if (err)
>> +			return err;
>> +
>> +		if (fib_dump_is_consistent(net, nb, cb, fib_seq))
>>  			return 0;
>>  	} while (++retries < FIB_DUMP_MAX_RETRIES);
>
>This is still more complicated than it needs to be. Why lump all
>fib_notifier_ops into 1 dump when they are separate databases with
>separate seq numbers? Just dump them 1 at a time and retry that 1
>database as needed.

Well I think that what you describe is out of scope of this patch. It is
another optimization of fib_notifier. The aim of this patchset is not
optimization of fib_notifier, but devlink netns change. This patchset is
just a dependency.

Can't we do optimization in another patchset? I already struggled to
keep this one within 15-patch limit.

Thanks

>
>ie., This:
>    list_for_each_entry_rcu(ops, &net->fib_notifier_ops, list) {
>should be in register_fib_notifier and not fib_net_dump.
>
>as it stands you are potentially replaying way more than is needed when
>a dump is inconsistent.
>
>
>>  
>>  	return -EBUSY;
>> -
>> -err_fib_net_dump:
>> -	rcu_read_unlock();
>> -	return err;
>>  }
>>  EXPORT_SYMBOL(register_fib_notifier);
>>  
>> -int unregister_fib_notifier(struct notifier_block *nb)
>> +int unregister_fib_notifier(struct net *net, struct notifier_block *nb)
>>  {
>> -	return atomic_notifier_chain_unregister(&fib_chain, nb);
>> +	struct fib_notifier_net *fn_net = net_generic(net, fib_notifier_net_id);
>> +
>> +	return atomic_notifier_chain_unregister(&fn_net->fib_chain, nb);
>>  }
>>  EXPORT_SYMBOL(unregister_fib_notifier);
>>  
>
>
>
>
>

^ permalink raw reply	[flat|nested] 34+ messages in thread

* Re: [patch net-next 00/15] devlink: allow devlink instances to change network namespace
  2019-09-14  6:45 [patch net-next 00/15] devlink: allow devlink instances to change network namespace Jiri Pirko
                   ` (16 preceding siblings ...)
  2019-09-14  6:57 ` [patch iproute2-next 2/2] devlink: extend reload command to add support for network namespace change Jiri Pirko
@ 2019-09-16  7:01 ` David Miller
  2019-09-16  7:09   ` Jiri Pirko
  17 siblings, 1 reply; 34+ messages in thread
From: David Miller @ 2019-09-16  7:01 UTC (permalink / raw)
  To: jiri
  Cc: netdev, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw


Jiri, this has to wait until the next merge window sorry.

^ permalink raw reply	[flat|nested] 34+ messages in thread

* Re: [patch net-next 00/15] devlink: allow devlink instances to change network namespace
  2019-09-16  7:01 ` [patch net-next 00/15] devlink: allow devlink instances to change network namespace David Miller
@ 2019-09-16  7:09   ` Jiri Pirko
  0 siblings, 0 replies; 34+ messages in thread
From: Jiri Pirko @ 2019-09-16  7:09 UTC (permalink / raw)
  To: David Miller
  Cc: netdev, idosch, dsahern, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

Mon, Sep 16, 2019 at 09:01:11AM CEST, davem@davemloft.net wrote:
>
>Jiri, this has to wait until the next merge window sorry.

Sure, no worries :)

^ permalink raw reply	[flat|nested] 34+ messages in thread

* Re: [patch net-next 02/15] net: fib_notifier: make FIB notifier per-netns
  2019-09-16  5:38     ` Jiri Pirko
@ 2019-09-16 14:54       ` David Ahern
  0 siblings, 0 replies; 34+ messages in thread
From: David Ahern @ 2019-09-16 14:54 UTC (permalink / raw)
  To: Jiri Pirko
  Cc: netdev, davem, idosch, jakub.kicinski, tariqt, saeedm, kuznet,
	yoshfuji, shuah, mlxsw

On 9/15/19 11:38 PM, Jiri Pirko wrote:
>> This is still more complicated than it needs to be. Why lump all
>> fib_notifier_ops into 1 dump when they are separate databases with
>> separate seq numbers? Just dump them 1 at a time and retry that 1
>> database as needed.
> 
> Well I think that what you describe is out of scope of this patch. It is
> another optimization of fib_notifier. The aim of this patchset is not
> optimization of fib_notifier, but devlink netns change. This patchset is
> just a dependency.
> 
> Can't we do optimization in another patchset? I already struggled to
> keep this one within 15-patch limit.
> 

sure, but it seems to me it is less work to churn this code all at once
and you are already doing the testing. The fib notifier changes can
always be done as a prep set.


^ permalink raw reply	[flat|nested] 34+ messages in thread

end of thread, other threads:[~2019-09-16 14:54 UTC | newest]

Thread overview: 34+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-09-14  6:45 [patch net-next 00/15] devlink: allow devlink instances to change network namespace Jiri Pirko
2019-09-14  6:45 ` [patch net-next 01/15] netdevsim: change fib accounting and limitations to be per-device Jiri Pirko
2019-09-14  6:45 ` [patch net-next 02/15] net: fib_notifier: make FIB notifier per-netns Jiri Pirko
2019-09-15  8:06   ` Ido Schimmel
2019-09-15  9:37     ` Jiri Pirko
2019-09-15 20:05   ` David Ahern
2019-09-16  5:38     ` Jiri Pirko
2019-09-16 14:54       ` David Ahern
2019-09-14  6:45 ` [patch net-next 03/15] net: fib_notifier: propagate possible error during fib notifier registration Jiri Pirko
2019-09-15  8:17   ` Ido Schimmel
2019-09-15  9:41     ` Jiri Pirko
2019-09-14  6:45 ` [patch net-next 04/15] mlxsw: spectrum_router: Don't rely on missing extack to symbolize dump Jiri Pirko
2019-09-14  6:45 ` [patch net-next 05/15] net: fib_notifier: propagate extack down to the notifier block callback Jiri Pirko
2019-09-14  6:45 ` [patch net-next 06/15] net: devlink: export devlink net getter Jiri Pirko
2019-09-14  6:46 ` [patch net-next 07/15] mlxsw: spectrum: Take devlink net instead of init_net Jiri Pirko
2019-09-14  6:46 ` [patch net-next 08/15] mlxsw: Register port netdevices into net of core Jiri Pirko
2019-09-15  8:37   ` Ido Schimmel
2019-09-14  6:46 ` [patch net-next 09/15] mlxsw: Propagate extack down to register_fib_notifier() Jiri Pirko
2019-09-15  8:39   ` Ido Schimmel
2019-09-14  6:46 ` [patch net-next 10/15] netdevsim: add all ports in nsim_dev_create() and del them in destroy() Jiri Pirko
2019-09-14  6:46 ` [patch net-next 11/15] netdevsim: implement proper devlink reload Jiri Pirko
2019-09-14  6:46 ` [patch net-next 12/15] netdevsim: register port netdevices into net of device Jiri Pirko
2019-09-14  6:46 ` [patch net-next 13/15] netdevsim: take devlink net instead of init_net Jiri Pirko
2019-09-14  6:46 ` [patch net-next 14/15] net: devlink: allow to change namespaces during reload Jiri Pirko
2019-09-15  8:58   ` Ido Schimmel
2019-09-15  9:43     ` Jiri Pirko
2019-09-14  6:46 ` [patch net-next 15/15] selftests: netdevsim: add tests for devlink reload with resources Jiri Pirko
2019-09-14  6:57 ` [patch iproute2-next 1/2] devlink: introduce cmdline option to switch to a different namespace Jiri Pirko
2019-09-15 18:01   ` David Ahern
2019-09-14  6:57 ` [patch iproute2-next 2/2] devlink: extend reload command to add support for network namespace change Jiri Pirko
2019-09-15  7:16   ` Ido Schimmel
2019-09-15  9:44     ` Jiri Pirko
2019-09-16  7:01 ` [patch net-next 00/15] devlink: allow devlink instances to change network namespace David Miller
2019-09-16  7:09   ` Jiri Pirko

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).