* [patch net-next 09/15] mlxsw: Propagate extack down to register_fib_notifier()
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
In-Reply-To: <20190914064608.26799-1-jiri@resnulli.us>
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
* [patch net-next 05/15] net: fib_notifier: propagate extack down to the notifier block callback
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
In-Reply-To: <20190914064608.26799-1-jiri@resnulli.us>
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
* [patch net-next 08/15] mlxsw: Register port netdevices into net of core
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
In-Reply-To: <20190914064608.26799-1-jiri@resnulli.us>
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
* [patch net-next 07/15] mlxsw: spectrum: Take devlink net instead of init_net
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
In-Reply-To: <20190914064608.26799-1-jiri@resnulli.us>
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
* [patch net-next 06/15] net: devlink: export devlink net getter
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
In-Reply-To: <20190914064608.26799-1-jiri@resnulli.us>
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
* [patch net-next 04/15] mlxsw: spectrum_router: Don't rely on missing extack to symbolize dump
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
In-Reply-To: <20190914064608.26799-1-jiri@resnulli.us>
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
* [patch net-next 03/15] net: fib_notifier: propagate possible error during fib notifier registration
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
In-Reply-To: <20190914064608.26799-1-jiri@resnulli.us>
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
* [patch net-next 02/15] net: fib_notifier: make FIB notifier per-netns
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
In-Reply-To: <20190914064608.26799-1-jiri@resnulli.us>
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
* [patch net-next 01/15] netdevsim: change fib accounting and limitations to be per-device
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
In-Reply-To: <20190914064608.26799-1-jiri@resnulli.us>
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, ¶ms);
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, ¶ms);
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, ¶ms);
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, ¶ms);
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
* [patch net-next 00/15] devlink: allow devlink instances to change network namespace
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
* Re: [patch iproute2-next v4 0/2] devlink: couple forgotten flash patches
From: Jiri Pirko @ 2019-09-14 6:00 UTC (permalink / raw)
To: David Ahern; +Cc: netdev, stephen, jakub.kicinski, saeedm, mlxsw, f.fainelli
In-Reply-To: <2c201359-2fa4-b1e4-061b-64a53eb30920@gmail.com>
Fri, Sep 13, 2019 at 07:25:07PM CEST, dsahern@gmail.com wrote:
>On 9/12/19 12:29 PM, Jiri Pirko wrote:
>> From: Jiri Pirko <jiri@mellanox.com>
>>
>> I was under impression they are already merged, but apparently they are
>> not. I just rebased them on top of current iproute2 net-next tree.
>>
>
>they were not forgotten; they were dropped asking for changes.
>
>thread is here:
>https://lore.kernel.org/netdev/20190604134450.2839-3-jiri@resnulli.us/
Well not really. The path was discussed in the thread. However, that is
unrelated to the changes these patches do. The flashing itself is
already there and present. These patches only add status.
Did I missed something?
^ permalink raw reply
* [PATCH net-next 4/4] bnxt_en: Add a new BNXT_FW_RESET_STATE_POLL_FW_DOWN state.
From: Michael Chan @ 2019-09-14 4:01 UTC (permalink / raw)
To: davem; +Cc: netdev, vasundhara-v.volam
In-Reply-To: <1568433701-29000-1-git-send-email-michael.chan@broadcom.com>
From: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
This new state is required when firmware indicates that the error
recovery process requires polling for firmware state to be completely
down before initiating reset. For example, firmware may take some
time to collect the crash dump before it is down and ready to be
reset.
Signed-off-by: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
---
drivers/net/ethernet/broadcom/bnxt/bnxt.c | 48 +++++++++++++++++++++++++++----
drivers/net/ethernet/broadcom/bnxt/bnxt.h | 3 ++
2 files changed, 46 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index 58831dd..b4a8cf6 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -6947,6 +6947,8 @@ static int __bnxt_hwrm_func_qcaps(struct bnxt *bp)
bp->fw_cap |= BNXT_FW_CAP_EXT_STATS_SUPPORTED;
if (flags & FUNC_QCAPS_RESP_FLAGS_ERROR_RECOVERY_CAPABLE)
bp->fw_cap |= BNXT_FW_CAP_ERROR_RECOVERY;
+ if (flags & FUNC_QCAPS_RESP_FLAGS_ERR_RECOVER_RELOAD)
+ bp->fw_cap |= BNXT_FW_CAP_ERR_RECOVER_RELOAD;
bp->tx_push_thresh = 0;
if (flags & FUNC_QCAPS_RESP_FLAGS_PUSH_MODE_SUPPORTED)
@@ -10097,6 +10099,8 @@ static void bnxt_force_fw_reset(struct bnxt *bp)
wait_dsecs = fw_health->normal_func_wait_dsecs;
bp->fw_reset_state = BNXT_FW_RESET_STATE_ENABLE_DEV;
}
+
+ bp->fw_reset_min_dsecs = fw_health->post_reset_wait_dsecs;
bp->fw_reset_max_dsecs = fw_health->post_reset_max_wait_dsecs;
bnxt_queue_fw_reset_work(bp, wait_dsecs * HZ / 10);
}
@@ -10138,7 +10142,7 @@ void bnxt_fw_reset(struct bnxt *bp)
bnxt_rtnl_lock_sp(bp);
if (test_bit(BNXT_STATE_OPEN, &bp->state) &&
!test_bit(BNXT_STATE_IN_FW_RESET, &bp->state)) {
- int n = 0;
+ int n = 0, tmo;
set_bit(BNXT_STATE_IN_FW_RESET, &bp->state);
if (bp->pf.active_vfs &&
@@ -10161,8 +10165,14 @@ void bnxt_fw_reset(struct bnxt *bp)
goto fw_reset_exit;
}
bnxt_fw_reset_close(bp);
- bp->fw_reset_state = BNXT_FW_RESET_STATE_ENABLE_DEV;
- bnxt_queue_fw_reset_work(bp, bp->fw_reset_min_dsecs * HZ / 10);
+ if (bp->fw_cap & BNXT_FW_CAP_ERR_RECOVER_RELOAD) {
+ bp->fw_reset_state = BNXT_FW_RESET_STATE_POLL_FW_DOWN;
+ tmo = HZ / 10;
+ } else {
+ bp->fw_reset_state = BNXT_FW_RESET_STATE_ENABLE_DEV;
+ tmo = bp->fw_reset_min_dsecs * HZ / 10;
+ }
+ bnxt_queue_fw_reset_work(bp, tmo);
}
fw_reset_exit:
bnxt_rtnl_unlock_sp(bp);
@@ -10605,6 +10615,7 @@ static void bnxt_fw_reset_task(struct work_struct *work)
switch (bp->fw_reset_state) {
case BNXT_FW_RESET_STATE_POLL_VF: {
int n = bnxt_get_registered_vfs(bp);
+ int tmo;
if (n < 0) {
netdev_err(bp->dev, "Firmware reset aborted, subsequent func_qcfg cmd failed, rc = %d, %d msecs since reset timestamp\n",
@@ -10626,11 +10637,38 @@ static void bnxt_fw_reset_task(struct work_struct *work)
bp->fw_reset_timestamp = jiffies;
rtnl_lock();
bnxt_fw_reset_close(bp);
- bp->fw_reset_state = BNXT_FW_RESET_STATE_ENABLE_DEV;
+ if (bp->fw_cap & BNXT_FW_CAP_ERR_RECOVER_RELOAD) {
+ bp->fw_reset_state = BNXT_FW_RESET_STATE_POLL_FW_DOWN;
+ tmo = HZ / 10;
+ } else {
+ bp->fw_reset_state = BNXT_FW_RESET_STATE_ENABLE_DEV;
+ tmo = bp->fw_reset_min_dsecs * HZ / 10;
+ }
rtnl_unlock();
- bnxt_queue_fw_reset_work(bp, bp->fw_reset_min_dsecs * HZ / 10);
+ bnxt_queue_fw_reset_work(bp, tmo);
return;
}
+ case BNXT_FW_RESET_STATE_POLL_FW_DOWN: {
+ u32 val;
+
+ val = bnxt_fw_health_readl(bp, BNXT_FW_HEALTH_REG);
+ if (!(val & BNXT_FW_STATUS_SHUTDOWN) &&
+ !time_after(jiffies, bp->fw_reset_timestamp +
+ (bp->fw_reset_max_dsecs * HZ / 10))) {
+ bnxt_queue_fw_reset_work(bp, HZ / 5);
+ return;
+ }
+
+ if (!bp->fw_health->master) {
+ u32 wait_dsecs = bp->fw_health->normal_func_wait_dsecs;
+
+ bp->fw_reset_state = BNXT_FW_RESET_STATE_ENABLE_DEV;
+ bnxt_queue_fw_reset_work(bp, wait_dsecs * HZ / 10);
+ return;
+ }
+ bp->fw_reset_state = BNXT_FW_RESET_STATE_RESET_FW;
+ }
+ /* fall through */
case BNXT_FW_RESET_STATE_RESET_FW: {
u32 wait_dsecs = bp->fw_health->post_reset_wait_dsecs;
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.h b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
index 42a8a75..d333589 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
@@ -1398,6 +1398,7 @@ struct bnxt_fw_reporter_ctx {
#define BNXT_FW_HEALTH_WIN_MAP_OFF 8
#define BNXT_FW_STATUS_HEALTHY 0x8000
+#define BNXT_FW_STATUS_SHUTDOWN 0x100000
struct bnxt {
void __iomem *bar0;
@@ -1655,6 +1656,7 @@ struct bnxt {
#define BNXT_FW_CAP_CFA_RFS_RING_TBL_IDX 0x00010000
#define BNXT_FW_CAP_PCIE_STATS_SUPPORTED 0x00020000
#define BNXT_FW_CAP_EXT_STATS_SUPPORTED 0x00040000
+ #define BNXT_FW_CAP_ERR_RECOVER_RELOAD 0x00100000
#define BNXT_NEW_RM(bp) ((bp)->fw_cap & BNXT_FW_CAP_NEW_RM)
u32 hwrm_spec_code;
@@ -1744,6 +1746,7 @@ struct bnxt {
#define BNXT_FW_RESET_STATE_ENABLE_DEV 3
#define BNXT_FW_RESET_STATE_POLL_FW 4
#define BNXT_FW_RESET_STATE_OPENING 5
+#define BNXT_FW_RESET_STATE_POLL_FW_DOWN 6
u16 fw_reset_min_dsecs;
#define BNXT_DFLT_FW_RST_MIN_DSECS 20
--
2.5.1
^ permalink raw reply related
* [PATCH net-next 3/4] bnxt_en: Update firmware interface spec. to 1.10.0.100.
From: Michael Chan @ 2019-09-14 4:01 UTC (permalink / raw)
To: davem; +Cc: netdev, vasundhara-v.volam
In-Reply-To: <1568433701-29000-1-git-send-email-michael.chan@broadcom.com>
Some error recovery updates to the spec., among other minor changes.
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
---
drivers/net/ethernet/broadcom/bnxt/bnxt_hsi.h | 146 ++++++++++++++++++--------
1 file changed, 103 insertions(+), 43 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_hsi.h b/drivers/net/ethernet/broadcom/bnxt/bnxt_hsi.h
index 2cdef75..03b197e 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_hsi.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_hsi.h
@@ -44,11 +44,12 @@ struct hwrm_resp_hdr {
#define TLV_TYPE_ENGINE_CKV_IV 0x8003UL
#define TLV_TYPE_ENGINE_CKV_AUTH_TAG 0x8004UL
#define TLV_TYPE_ENGINE_CKV_CIPHERTEXT 0x8005UL
-#define TLV_TYPE_ENGINE_CKV_ALGORITHMS 0x8006UL
+#define TLV_TYPE_ENGINE_CKV_HOST_ALGORITHMS 0x8006UL
#define TLV_TYPE_ENGINE_CKV_HOST_ECC_PUBLIC_KEY 0x8007UL
#define TLV_TYPE_ENGINE_CKV_ECDSA_SIGNATURE 0x8008UL
-#define TLV_TYPE_ENGINE_CKV_SRT_ECC_PUBLIC_KEY 0x8009UL
-#define TLV_TYPE_LAST TLV_TYPE_ENGINE_CKV_SRT_ECC_PUBLIC_KEY
+#define TLV_TYPE_ENGINE_CKV_FW_ECC_PUBLIC_KEY 0x8009UL
+#define TLV_TYPE_ENGINE_CKV_FW_ALGORITHMS 0x800aUL
+#define TLV_TYPE_LAST TLV_TYPE_ENGINE_CKV_FW_ALGORITHMS
/* tlv (size:64b/8B) */
@@ -201,10 +202,16 @@ struct cmd_nums {
#define HWRM_PORT_QSTATS_EXT 0xb4UL
#define HWRM_PORT_PHY_MDIO_WRITE 0xb5UL
#define HWRM_PORT_PHY_MDIO_READ 0xb6UL
+ #define HWRM_PORT_PHY_MDIO_BUS_ACQUIRE 0xb7UL
+ #define HWRM_PORT_PHY_MDIO_BUS_RELEASE 0xb8UL
#define HWRM_FW_RESET 0xc0UL
#define HWRM_FW_QSTATUS 0xc1UL
#define HWRM_FW_HEALTH_CHECK 0xc2UL
#define HWRM_FW_SYNC 0xc3UL
+ #define HWRM_FW_STATE_BUFFER_QCAPS 0xc4UL
+ #define HWRM_FW_STATE_QUIESCE 0xc5UL
+ #define HWRM_FW_STATE_BACKUP 0xc6UL
+ #define HWRM_FW_STATE_RESTORE 0xc7UL
#define HWRM_FW_SET_TIME 0xc8UL
#define HWRM_FW_GET_TIME 0xc9UL
#define HWRM_FW_SET_STRUCTURED_DATA 0xcaUL
@@ -216,7 +223,10 @@ struct cmd_nums {
#define HWRM_FWD_ASYNC_EVENT_CMPL 0xd3UL
#define HWRM_OEM_CMD 0xd4UL
#define HWRM_PORT_PRBS_TEST 0xd5UL
+ #define HWRM_PORT_SFP_SIDEBAND_CFG 0xd6UL
+ #define HWRM_PORT_SFP_SIDEBAND_QCFG 0xd7UL
#define HWRM_TEMP_MONITOR_QUERY 0xe0UL
+ #define HWRM_REG_POWER_QUERY 0xe1UL
#define HWRM_WOL_FILTER_ALLOC 0xf0UL
#define HWRM_WOL_FILTER_FREE 0xf1UL
#define HWRM_WOL_FILTER_QCFG 0xf2UL
@@ -411,8 +421,8 @@ struct hwrm_err_output {
#define HWRM_VERSION_MAJOR 1
#define HWRM_VERSION_MINOR 10
#define HWRM_VERSION_UPDATE 0
-#define HWRM_VERSION_RSVD 89
-#define HWRM_VERSION_STR "1.10.0.89"
+#define HWRM_VERSION_RSVD 100
+#define HWRM_VERSION_STR "1.10.0.100"
/* hwrm_ver_get_input (size:192b/24B) */
struct hwrm_ver_get_input {
@@ -805,6 +815,37 @@ struct hwrm_async_event_cmpl_vf_cfg_change {
#define ASYNC_EVENT_CMPL_VF_CFG_CHANGE_EVENT_DATA1_TRUSTED_VF_CFG_CHANGE 0x10UL
};
+/* hwrm_async_event_cmpl_default_vnic_change (size:128b/16B) */
+struct hwrm_async_event_cmpl_default_vnic_change {
+ __le16 type;
+ #define ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_TYPE_MASK 0x3fUL
+ #define ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_TYPE_SFT 0
+ #define ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_TYPE_HWRM_ASYNC_EVENT 0x2eUL
+ #define ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_TYPE_LAST ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_TYPE_HWRM_ASYNC_EVENT
+ #define ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_UNUSED1_MASK 0xffc0UL
+ #define ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_UNUSED1_SFT 6
+ __le16 event_id;
+ #define ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_EVENT_ID_ALLOC_FREE_NOTIFICATION 0x35UL
+ #define ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_EVENT_ID_LAST ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_EVENT_ID_ALLOC_FREE_NOTIFICATION
+ __le32 event_data2;
+ u8 opaque_v;
+ #define ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_V 0x1UL
+ #define ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_OPAQUE_MASK 0xfeUL
+ #define ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_OPAQUE_SFT 1
+ u8 timestamp_lo;
+ __le16 timestamp_hi;
+ __le32 event_data1;
+ #define ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_EVENT_DATA1_DEF_VNIC_STATE_MASK 0x3UL
+ #define ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_EVENT_DATA1_DEF_VNIC_STATE_SFT 0
+ #define ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_EVENT_DATA1_DEF_VNIC_STATE_DEF_VNIC_ALLOC 0x1UL
+ #define ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_EVENT_DATA1_DEF_VNIC_STATE_DEF_VNIC_FREE 0x2UL
+ #define ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_EVENT_DATA1_DEF_VNIC_STATE_LAST ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_EVENT_DATA1_DEF_VNIC_STATE_DEF_VNIC_FREE
+ #define ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_EVENT_DATA1_PF_ID_MASK 0x3fcUL
+ #define ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_EVENT_DATA1_PF_ID_SFT 2
+ #define ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_EVENT_DATA1_VF_ID_MASK 0x3fffc00UL
+ #define ASYNC_EVENT_CMPL_DEFAULT_VNIC_CHANGE_EVENT_DATA1_VF_ID_SFT 10
+};
+
/* hwrm_async_event_cmpl_hw_flow_aged (size:128b/16B) */
struct hwrm_async_event_cmpl_hw_flow_aged {
__le16 type;
@@ -1047,31 +1088,33 @@ struct hwrm_func_qcaps_output {
__le16 fid;
__le16 port_id;
__le32 flags;
- #define FUNC_QCAPS_RESP_FLAGS_PUSH_MODE_SUPPORTED 0x1UL
- #define FUNC_QCAPS_RESP_FLAGS_GLOBAL_MSIX_AUTOMASKING 0x2UL
- #define FUNC_QCAPS_RESP_FLAGS_PTP_SUPPORTED 0x4UL
- #define FUNC_QCAPS_RESP_FLAGS_ROCE_V1_SUPPORTED 0x8UL
- #define FUNC_QCAPS_RESP_FLAGS_ROCE_V2_SUPPORTED 0x10UL
- #define FUNC_QCAPS_RESP_FLAGS_WOL_MAGICPKT_SUPPORTED 0x20UL
- #define FUNC_QCAPS_RESP_FLAGS_WOL_BMP_SUPPORTED 0x40UL
- #define FUNC_QCAPS_RESP_FLAGS_TX_RING_RL_SUPPORTED 0x80UL
- #define FUNC_QCAPS_RESP_FLAGS_TX_BW_CFG_SUPPORTED 0x100UL
- #define FUNC_QCAPS_RESP_FLAGS_VF_TX_RING_RL_SUPPORTED 0x200UL
- #define FUNC_QCAPS_RESP_FLAGS_VF_BW_CFG_SUPPORTED 0x400UL
- #define FUNC_QCAPS_RESP_FLAGS_STD_TX_RING_MODE_SUPPORTED 0x800UL
- #define FUNC_QCAPS_RESP_FLAGS_GENEVE_TUN_FLAGS_SUPPORTED 0x1000UL
- #define FUNC_QCAPS_RESP_FLAGS_NVGRE_TUN_FLAGS_SUPPORTED 0x2000UL
- #define FUNC_QCAPS_RESP_FLAGS_GRE_TUN_FLAGS_SUPPORTED 0x4000UL
- #define FUNC_QCAPS_RESP_FLAGS_MPLS_TUN_FLAGS_SUPPORTED 0x8000UL
- #define FUNC_QCAPS_RESP_FLAGS_PCIE_STATS_SUPPORTED 0x10000UL
- #define FUNC_QCAPS_RESP_FLAGS_ADOPTED_PF_SUPPORTED 0x20000UL
- #define FUNC_QCAPS_RESP_FLAGS_ADMIN_PF_SUPPORTED 0x40000UL
- #define FUNC_QCAPS_RESP_FLAGS_LINK_ADMIN_STATUS_SUPPORTED 0x80000UL
- #define FUNC_QCAPS_RESP_FLAGS_WCB_PUSH_MODE 0x100000UL
- #define FUNC_QCAPS_RESP_FLAGS_DYNAMIC_TX_RING_ALLOC 0x200000UL
- #define FUNC_QCAPS_RESP_FLAGS_HOT_RESET_CAPABLE 0x400000UL
- #define FUNC_QCAPS_RESP_FLAGS_ERROR_RECOVERY_CAPABLE 0x800000UL
- #define FUNC_QCAPS_RESP_FLAGS_EXT_STATS_SUPPORTED 0x1000000UL
+ #define FUNC_QCAPS_RESP_FLAGS_PUSH_MODE_SUPPORTED 0x1UL
+ #define FUNC_QCAPS_RESP_FLAGS_GLOBAL_MSIX_AUTOMASKING 0x2UL
+ #define FUNC_QCAPS_RESP_FLAGS_PTP_SUPPORTED 0x4UL
+ #define FUNC_QCAPS_RESP_FLAGS_ROCE_V1_SUPPORTED 0x8UL
+ #define FUNC_QCAPS_RESP_FLAGS_ROCE_V2_SUPPORTED 0x10UL
+ #define FUNC_QCAPS_RESP_FLAGS_WOL_MAGICPKT_SUPPORTED 0x20UL
+ #define FUNC_QCAPS_RESP_FLAGS_WOL_BMP_SUPPORTED 0x40UL
+ #define FUNC_QCAPS_RESP_FLAGS_TX_RING_RL_SUPPORTED 0x80UL
+ #define FUNC_QCAPS_RESP_FLAGS_TX_BW_CFG_SUPPORTED 0x100UL
+ #define FUNC_QCAPS_RESP_FLAGS_VF_TX_RING_RL_SUPPORTED 0x200UL
+ #define FUNC_QCAPS_RESP_FLAGS_VF_BW_CFG_SUPPORTED 0x400UL
+ #define FUNC_QCAPS_RESP_FLAGS_STD_TX_RING_MODE_SUPPORTED 0x800UL
+ #define FUNC_QCAPS_RESP_FLAGS_GENEVE_TUN_FLAGS_SUPPORTED 0x1000UL
+ #define FUNC_QCAPS_RESP_FLAGS_NVGRE_TUN_FLAGS_SUPPORTED 0x2000UL
+ #define FUNC_QCAPS_RESP_FLAGS_GRE_TUN_FLAGS_SUPPORTED 0x4000UL
+ #define FUNC_QCAPS_RESP_FLAGS_MPLS_TUN_FLAGS_SUPPORTED 0x8000UL
+ #define FUNC_QCAPS_RESP_FLAGS_PCIE_STATS_SUPPORTED 0x10000UL
+ #define FUNC_QCAPS_RESP_FLAGS_ADOPTED_PF_SUPPORTED 0x20000UL
+ #define FUNC_QCAPS_RESP_FLAGS_ADMIN_PF_SUPPORTED 0x40000UL
+ #define FUNC_QCAPS_RESP_FLAGS_LINK_ADMIN_STATUS_SUPPORTED 0x80000UL
+ #define FUNC_QCAPS_RESP_FLAGS_WCB_PUSH_MODE 0x100000UL
+ #define FUNC_QCAPS_RESP_FLAGS_DYNAMIC_TX_RING_ALLOC 0x200000UL
+ #define FUNC_QCAPS_RESP_FLAGS_HOT_RESET_CAPABLE 0x400000UL
+ #define FUNC_QCAPS_RESP_FLAGS_ERROR_RECOVERY_CAPABLE 0x800000UL
+ #define FUNC_QCAPS_RESP_FLAGS_EXT_STATS_SUPPORTED 0x1000000UL
+ #define FUNC_QCAPS_RESP_FLAGS_ERR_RECOVER_RELOAD 0x2000000UL
+ #define FUNC_QCAPS_RESP_FLAGS_NOTIFY_VF_DEF_VNIC_CHNG_SUPPORTED 0x4000000UL
u8 mac_address[6];
__le16 max_rsscos_ctx;
__le16 max_cmpl_rings;
@@ -1208,7 +1251,8 @@ struct hwrm_func_qcfg_output {
__le16 alloc_stat_ctx;
__le16 alloc_msix;
__le16 registered_vfs;
- u8 unused_1[3];
+ __le16 l2_doorbell_bar_size_kb;
+ u8 unused_1;
u8 always_1;
__le32 reset_addr_poll;
u8 unused_2[3];
@@ -1363,7 +1407,11 @@ struct hwrm_func_qstats_input {
__le16 target_id;
__le64 resp_addr;
__le16 fid;
- u8 unused_0[6];
+ u8 flags;
+ #define FUNC_QSTATS_REQ_FLAGS_UNUSED 0x0UL
+ #define FUNC_QSTATS_REQ_FLAGS_ROCE_ONLY 0x1UL
+ #define FUNC_QSTATS_REQ_FLAGS_LAST FUNC_QSTATS_REQ_FLAGS_ROCE_ONLY
+ u8 unused_0[5];
};
/* hwrm_func_qstats_output (size:1408b/176B) */
@@ -4714,7 +4762,7 @@ struct hwrm_vnic_free_output {
u8 valid;
};
-/* hwrm_vnic_cfg_input (size:320b/40B) */
+/* hwrm_vnic_cfg_input (size:384b/48B) */
struct hwrm_vnic_cfg_input {
__le16 req_type;
__le16 cmpl_ring;
@@ -4737,6 +4785,7 @@ struct hwrm_vnic_cfg_input {
#define VNIC_CFG_REQ_ENABLES_MRU 0x10UL
#define VNIC_CFG_REQ_ENABLES_DEFAULT_RX_RING_ID 0x20UL
#define VNIC_CFG_REQ_ENABLES_DEFAULT_CMPL_RING_ID 0x40UL
+ #define VNIC_CFG_REQ_ENABLES_QUEUE_ID 0x80UL
__le16 vnic_id;
__le16 dflt_ring_grp;
__le16 rss_rule;
@@ -4745,6 +4794,8 @@ struct hwrm_vnic_cfg_input {
__le16 mru;
__le16 default_rx_ring_id;
__le16 default_cmpl_ring_id;
+ __le16 queue_id;
+ u8 unused0[6];
};
/* hwrm_vnic_cfg_output (size:128b/16B) */
@@ -4785,6 +4836,7 @@ struct hwrm_vnic_qcaps_output {
#define VNIC_QCAPS_RESP_FLAGS_RSS_DFLT_CR_CAP 0x20UL
#define VNIC_QCAPS_RESP_FLAGS_ROCE_MIRRORING_CAPABLE_VNIC_CAP 0x40UL
#define VNIC_QCAPS_RESP_FLAGS_OUTERMOST_RSS_CAP 0x80UL
+ #define VNIC_QCAPS_RESP_FLAGS_COS_ASSIGNMENT_CAP 0x100UL
__le16 max_aggs_supported;
u8 unused_1[5];
u8 valid;
@@ -6794,15 +6846,16 @@ struct hwrm_fw_reset_input {
__le16 target_id;
__le64 resp_addr;
u8 embedded_proc_type;
- #define FW_RESET_REQ_EMBEDDED_PROC_TYPE_BOOT 0x0UL
- #define FW_RESET_REQ_EMBEDDED_PROC_TYPE_MGMT 0x1UL
- #define FW_RESET_REQ_EMBEDDED_PROC_TYPE_NETCTRL 0x2UL
- #define FW_RESET_REQ_EMBEDDED_PROC_TYPE_ROCE 0x3UL
- #define FW_RESET_REQ_EMBEDDED_PROC_TYPE_HOST 0x4UL
- #define FW_RESET_REQ_EMBEDDED_PROC_TYPE_AP 0x5UL
- #define FW_RESET_REQ_EMBEDDED_PROC_TYPE_CHIP 0x6UL
- #define FW_RESET_REQ_EMBEDDED_PROC_TYPE_HOST_RESOURCE_REINIT 0x7UL
- #define FW_RESET_REQ_EMBEDDED_PROC_TYPE_LAST FW_RESET_REQ_EMBEDDED_PROC_TYPE_HOST_RESOURCE_REINIT
+ #define FW_RESET_REQ_EMBEDDED_PROC_TYPE_BOOT 0x0UL
+ #define FW_RESET_REQ_EMBEDDED_PROC_TYPE_MGMT 0x1UL
+ #define FW_RESET_REQ_EMBEDDED_PROC_TYPE_NETCTRL 0x2UL
+ #define FW_RESET_REQ_EMBEDDED_PROC_TYPE_ROCE 0x3UL
+ #define FW_RESET_REQ_EMBEDDED_PROC_TYPE_HOST 0x4UL
+ #define FW_RESET_REQ_EMBEDDED_PROC_TYPE_AP 0x5UL
+ #define FW_RESET_REQ_EMBEDDED_PROC_TYPE_CHIP 0x6UL
+ #define FW_RESET_REQ_EMBEDDED_PROC_TYPE_HOST_RESOURCE_REINIT 0x7UL
+ #define FW_RESET_REQ_EMBEDDED_PROC_TYPE_IMPACTLESS_ACTIVATION 0x8UL
+ #define FW_RESET_REQ_EMBEDDED_PROC_TYPE_LAST FW_RESET_REQ_EMBEDDED_PROC_TYPE_IMPACTLESS_ACTIVATION
u8 selfrst_status;
#define FW_RESET_REQ_SELFRST_STATUS_SELFRSTNONE 0x0UL
#define FW_RESET_REQ_SELFRST_STATUS_SELFRSTASAP 0x1UL
@@ -7125,7 +7178,14 @@ struct hwrm_temp_monitor_query_output {
__le16 seq_id;
__le16 resp_len;
u8 temp;
- u8 unused_0[6];
+ u8 phy_temp;
+ u8 om_temp;
+ u8 flags;
+ #define TEMP_MONITOR_QUERY_RESP_FLAGS_TEMP_NOT_AVAILABLE 0x1UL
+ #define TEMP_MONITOR_QUERY_RESP_FLAGS_PHY_TEMP_NOT_AVAILABLE 0x2UL
+ #define TEMP_MONITOR_QUERY_RESP_FLAGS_OM_NOT_PRESENT 0x4UL
+ #define TEMP_MONITOR_QUERY_RESP_FLAGS_OM_TEMP_NOT_AVAILABLE 0x8UL
+ u8 unused_0[3];
u8 valid;
};
--
2.5.1
^ permalink raw reply related
* [PATCH net-next 2/4] bnxt_en: Increase timeout for HWRM_DBG_COREDUMP_XX commands
From: Michael Chan @ 2019-09-14 4:01 UTC (permalink / raw)
To: davem; +Cc: netdev, vasundhara-v.volam
In-Reply-To: <1568433701-29000-1-git-send-email-michael.chan@broadcom.com>
From: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
Firmware coredump messages take much longer than standard messages,
so increase the timeout accordingly.
Fixes: 6c5657d085ae ("bnxt_en: Add support for ethtool get dump.")
Signed-off-by: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
---
drivers/net/ethernet/broadcom/bnxt/bnxt.h | 1 +
drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c | 2 +-
2 files changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.h b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
index 333b0a8..42a8a75 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
@@ -648,6 +648,7 @@ struct nqe_cn {
#define SHORT_HWRM_CMD_TIMEOUT 20
#define HWRM_CMD_TIMEOUT (bp->hwrm_cmd_timeout)
#define HWRM_RESET_TIMEOUT ((HWRM_CMD_TIMEOUT) * 4)
+#define HWRM_COREDUMP_TIMEOUT ((HWRM_CMD_TIMEOUT) * 12)
#define HWRM_RESP_ERR_CODE_MASK 0xffff
#define HWRM_RESP_LEN_OFFSET 4
#define HWRM_RESP_LEN_MASK 0xffff0000
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
index 235265e..51c1404 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
@@ -3112,7 +3112,7 @@ static int bnxt_hwrm_dbg_coredump_initiate(struct bnxt *bp, u16 component_id,
req.component_id = cpu_to_le16(component_id);
req.segment_id = cpu_to_le16(segment_id);
- return hwrm_send_message(bp, &req, sizeof(req), HWRM_CMD_TIMEOUT);
+ return hwrm_send_message(bp, &req, sizeof(req), HWRM_COREDUMP_TIMEOUT);
}
static int bnxt_hwrm_dbg_coredump_retrieve(struct bnxt *bp, u16 component_id,
--
2.5.1
^ permalink raw reply related
* [PATCH net-next 1/4] bnxt_en: Don't proceed in .ndo_set_rx_mode() when device is not in open state.
From: Michael Chan @ 2019-09-14 4:01 UTC (permalink / raw)
To: davem; +Cc: netdev, vasundhara-v.volam
In-Reply-To: <1568433701-29000-1-git-send-email-michael.chan@broadcom.com>
Check the BNXT_STATE_OPEN flag instead of netif_running() in
bnxt_set_rx_mode(). If the driver is going through any reset, such
as firmware reset or even TX timeout, it may not be ready to set the RX
mode and may crash. The new rx mode settings will be picked up when
the device is opened again later.
Fixes: 230d1f0de754 ("bnxt_en: Handle firmware reset.")
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
---
drivers/net/ethernet/broadcom/bnxt/bnxt.c | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index 402d9f5..58831dd 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -9557,14 +9557,16 @@ static bool bnxt_uc_list_updated(struct bnxt *bp)
static void bnxt_set_rx_mode(struct net_device *dev)
{
struct bnxt *bp = netdev_priv(dev);
- struct bnxt_vnic_info *vnic = &bp->vnic_info[0];
- u32 mask = vnic->rx_mask;
+ struct bnxt_vnic_info *vnic;
bool mc_update = false;
bool uc_update;
+ u32 mask;
- if (!netif_running(dev))
+ if (!test_bit(BNXT_STATE_OPEN, &bp->state))
return;
+ vnic = &bp->vnic_info[0];
+ mask = vnic->rx_mask;
mask &= ~(CFA_L2_SET_RX_MASK_REQ_MASK_PROMISCUOUS |
CFA_L2_SET_RX_MASK_REQ_MASK_MCAST |
CFA_L2_SET_RX_MASK_REQ_MASK_ALL_MCAST |
--
2.5.1
^ permalink raw reply related
* [PATCH net-next 0/4] bnxt_en: error recovery follow-up patches.
From: Michael Chan @ 2019-09-14 4:01 UTC (permalink / raw)
To: davem; +Cc: netdev, vasundhara-v.volam
A follow-up patchset for the recently added health and error recovery
feature. The first fix is to prevent .ndo_set_rx_mode() from proceeding
when reset is in progress. The 2nd fix is for the firmware coredump
command. The 3rd and 4th patches update the error recovery process
slightly to add a state that polls and waits for the firmware to be down.
Michael Chan (2):
bnxt_en: Don't proceed in .ndo_set_rx_mode() when device is not in
open state.
bnxt_en: Update firmware interface spec. to 1.10.0.100.
Vasundhara Volam (2):
bnxt_en: Increase timeout for HWRM_DBG_COREDUMP_XX commands
bnxt_en: Add a new BNXT_FW_RESET_STATE_POLL_FW_DOWN state.
drivers/net/ethernet/broadcom/bnxt/bnxt.c | 56 +++++++--
drivers/net/ethernet/broadcom/bnxt/bnxt.h | 4 +
drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c | 2 +-
drivers/net/ethernet/broadcom/bnxt/bnxt_hsi.h | 146 +++++++++++++++-------
4 files changed, 156 insertions(+), 52 deletions(-)
--
2.5.1
^ permalink raw reply
* Re: SFP support with RGMII MAC via RGMII to SERDES/SGMII PHY?
From: Florian Fainelli @ 2019-09-14 3:31 UTC (permalink / raw)
To: George McCollister, netdev, Russell King, Andrew Lunn,
Heiner Kallweit
In-Reply-To: <CAFSKS=NmM9bPb0R_zoFN+9AuG=x6DUffTNXpLSNRAHuZz4ki-g@mail.gmail.com>
+Russell, Andrew, Heiner,
On 9/13/2019 9:44 AM, George McCollister wrote:
> Every example of phylink SFP support I've seen is using an Ethernet
> MAC with native SGMII.
> Can phylink facilitate support of Fiber and Copper SFP modules
> connected to an RGMII MAC if all of the following are true?
I don't think that use case has been presented before, but phylink
sounds like the tool that should help solve it. From your description
below, it sounds like all the pieces are there to support it. Is the
Ethernet MAC driver upstream?
>
> 1) The MAC is connected via RGMII to a transceiver/PHY (such as
> Marvell 88E1512) which then connects to the SFP via SERDER/SGMII. If
> you want to see a block diagram it's the first one here:
> https://www.marvell.com/transceivers/assets/Alaska_88E1512-001_product_brief.pdf
>
> 2) The 1G Ethernet driver has been converted to use phylink.
>
> 3) An I2C controller on the SoC is connected to the SFP cage.
>
> 4) TxFault, LOS and MOD-DEF0 are connected to GPIO on the SoC.
>
> 5) MDIO is connected to the intermediate PHY.
>
> Any thoughts on what might be missing to support this (if anything)
> would be appreciated--
Florian
^ permalink raw reply
* Re: [PATCH v2 net-next 2/7] net: dsa: Pass ndo_setup_tc slave callback to drivers
From: Florian Fainelli @ 2019-09-14 2:59 UTC (permalink / raw)
To: Vladimir Oltean, vivien.didelot, andrew, davem, vinicius.gomes,
vedang.patel, richardcochran
Cc: weifeng.voon, jiri, m-karicheri2, Jose.Abreu, ilias.apalodimas,
jhs, xiyou.wangcong, kurt.kanzenbach, joergen.andreasen, netdev
In-Reply-To: <20190914011802.1602-3-olteanv@gmail.com>
On 9/13/2019 6:17 PM, Vladimir Oltean wrote:
> DSA currently handles shared block filters (for the classifier-action
> qdisc) in the core due to what I believe are simply pragmatic reasons -
> hiding the complexity from drivers and offerring a simple API for port
> mirroring.
>
> Extend the dsa_slave_setup_tc function by passing all other qdisc
> offloads to the driver layer, where the driver may choose what it
> implements and how. DSA is simply a pass-through in this case.
>
> Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
> Acked-by: Kurt Kanzenbach <kurt@linutronix.de>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
--
Florian
^ permalink raw reply
* [PATCH v2 net-next 7/7] docs: net: dsa: sja1105: Add info about the time-aware scheduler
From: Vladimir Oltean @ 2019-09-14 1:18 UTC (permalink / raw)
To: f.fainelli, vivien.didelot, andrew, davem, vinicius.gomes,
vedang.patel, richardcochran
Cc: weifeng.voon, jiri, m-karicheri2, Jose.Abreu, ilias.apalodimas,
jhs, xiyou.wangcong, kurt.kanzenbach, joergen.andreasen, netdev,
Vladimir Oltean
In-Reply-To: <20190914011802.1602-1-olteanv@gmail.com>
While not an exhaustive usage tutorial, this describes the details
needed to build more complex scenarios.
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
---
Changes since v1:
- Patch is new.
Documentation/networking/dsa/sja1105.rst | 90 ++++++++++++++++++++++++
1 file changed, 90 insertions(+)
diff --git a/Documentation/networking/dsa/sja1105.rst b/Documentation/networking/dsa/sja1105.rst
index cb2858dece93..2eaa6edf9c5b 100644
--- a/Documentation/networking/dsa/sja1105.rst
+++ b/Documentation/networking/dsa/sja1105.rst
@@ -146,6 +146,96 @@ enslaves eth0 and eth1 (the DSA master of the switch ports). This is because in
this mode, the switch ports beneath br0 are not capable of regular traffic, and
are only used as a conduit for switchdev operations.
+Offloads
+========
+
+Time-aware scheduling
+---------------------
+
+The switch supports a variation of the enhancements for scheduled traffic
+specified in IEEE 802.1Q-2018 (formerly 802.1Qbv). This means it can be used to
+ensure deterministic latency for priority traffic that is sent in-band with its
+gate-open event in the network schedule.
+
+This capability can be managed through the tc-taprio offload ('flags 2'). The
+difference compared to the software implementation of taprio is that the latter
+would only be able to shape traffic originated from the CPU, but not
+autonomously forwarded flows.
+
+The device has 8 traffic classes, and maps incoming frames to one of them based
+on the VLAN PCP bits (if no VLAN is present, the port-based default is used).
+As described in the previous sections, depending on the value of
+``vlan_filtering``, the EtherType recognized by the switch as being VLAN can
+either be the typical 0x8100 or a custom value used internally by the driver
+for tagging. Therefore, the switch ignores the VLAN PCP if used in standalone
+or bridge mode with ``vlan_filtering=0``, as it will not recognize the 0x8100
+EtherType. In these modes, injecting into a particular TX queue can only be
+done by the DSA net devices, which populate the PCP field of the tagging header
+on egress. Using ``vlan_filtering=1``, the behavior is the other way around:
+offloaded flows can be steered to TX queues based on the VLAN PCP, but the DSA
+net devices are no longer able to do that. To inject frames into a hardware TX
+queue with VLAN awareness active, it is necessary to create a VLAN
+sub-interface on the DSA master port, and send normal (0x8100) VLAN-tagged
+towards the switch, with the VLAN PCP bits set appropriately.
+
+Management traffic (having DMAC 01-80-C2-xx-xx-xx or 01-19-1B-xx-xx-xx) is the
+notable exception: the switch always treats it with a fixed priority and
+disregards any VLAN PCP bits even if present. The traffic class for management
+traffic is configurable through ``CONFIG_NET_DSA_SJA1105_HOSTPRIO``, which by
+default has a value of 7 (highest priority).
+
+Below is an example of configuring a 500 us cyclic schedule on egress port
+``swp5``. The traffic class gate for management traffic (7) is open for 100 us,
+and the gates for all other traffic classes are open for 400 us::
+
+ #!/bin/bash
+
+ set -e -u -o pipefail
+
+ NSEC_PER_SEC="1000000000"
+
+ gatemask() {
+ local tc_list="$1"
+ local mask=0
+
+ for tc in ${tc_list}; do
+ mask=$((${mask} | (1 << ${tc})))
+ done
+
+ printf "%02x" ${mask}
+ }
+
+ if ! systemctl is-active --quiet ptp4l; then
+ echo "Please start the ptp4l service"
+ exit
+ fi
+
+ now=$(phc_ctl /dev/ptp1 get | gawk '/clock time is/ { print $5; }')
+ # Phase-align the base time to the start of the next second.
+ sec=$(echo "${now}" | gawk -F. '{ print $1; }')
+ base_time="$(((${sec} + 1) * ${NSEC_PER_SEC}))"
+
+ tc qdisc add dev swp5 parent root handle 100 taprio \
+ num_tc 8 \
+ map 0 1 2 3 5 6 7 \
+ queues 1@0 1@1 1@2 1@3 1@4 1@5 1@6 1@7 \
+ base-time ${base_time} \
+ sched-entry S $(gatemask 7) 100000 \
+ sched-entry S $(gatemask "0 1 2 3 4 5 6") 400000 \
+ flags 2
+
+It is possible to apply the tc-taprio offload on multiple egress ports. There
+are hardware restrictions related to the fact that no gate event may trigger
+simultaneously on two ports. The driver checks the consistency of the schedules
+against this restriction and errors out when appropriate. Schedule analysis is
+needed to avoid this, which is outside the scope of the document.
+
+At the moment, the time-aware scheduler can only be triggered based on a
+standalone clock and not based on PTP time. This means the base-time argument
+from tc-taprio is ignored and the schedule starts right away. It also means it
+is more difficult to phase-align the scheduler with the other devices in the
+network.
+
Device Tree bindings and board design
=====================================
--
2.17.1
^ permalink raw reply related
* [PATCH v2 net-next 6/7] net: dsa: sja1105: Configure the Time-Aware Scheduler via tc-taprio offload
From: Vladimir Oltean @ 2019-09-14 1:18 UTC (permalink / raw)
To: f.fainelli, vivien.didelot, andrew, davem, vinicius.gomes,
vedang.patel, richardcochran
Cc: weifeng.voon, jiri, m-karicheri2, Jose.Abreu, ilias.apalodimas,
jhs, xiyou.wangcong, kurt.kanzenbach, joergen.andreasen, netdev,
Vladimir Oltean
In-Reply-To: <20190914011802.1602-1-olteanv@gmail.com>
This qdisc offload is the closest thing to what the SJA1105 supports in
hardware for time-based egress shaping. The switch core really is built
around SAE AS6802/TTEthernet (a TTTech standard) but can be made to
operate similarly to IEEE 802.1Qbv with some constraints:
- The gate control list is a global list for all ports. There are 8
execution threads that iterate through this global list in parallel.
I don't know why 8, there are only 4 front-panel ports.
- Care must be taken by the user to make sure that two execution threads
never get to execute a GCL entry simultaneously. I created a O(n^4)
checker for this hardware limitation, prior to accepting a taprio
offload configuration as valid.
- The spec says that if a GCL entry's interval is shorter than the frame
length, you shouldn't send it (and end up in head-of-line blocking).
Well, this switch does anyway.
- The switch has no concept of ADMIN and OPER configurations. Because
it's so simple, the TAS settings are loaded through the static config
tables interface, so there isn't even place for any discussion about
'graceful switchover between ADMIN and OPER'. You just reset the
switch and upload a new OPER config.
- The switch accepts multiple time sources for the gate events. Right
now I am using the standalone clock source as opposed to PTP. So the
base time parameter doesn't really do much. Support for the PTP clock
source will be added in a future series.
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
---
Changes since v1:
- Adapted to the naming convention changes in 01/07 (taprio_get ->
taprio_offload_get, tas_config -> offload, etc).
Changes since RFC:
- Removed the sja1105_tas_config_work workqueue.
- Allocating memory with GFP_KERNEL.
- Made the ASCII art drawing fit in < 80 characters.
- Made most of the time-holding variables s64 instead of u64 (for fear
of them not holding the result of signed arithmetics properly).
drivers/net/dsa/sja1105/Kconfig | 8 +
drivers/net/dsa/sja1105/Makefile | 4 +
drivers/net/dsa/sja1105/sja1105.h | 6 +
drivers/net/dsa/sja1105/sja1105_main.c | 19 +-
drivers/net/dsa/sja1105/sja1105_tas.c | 419 +++++++++++++++++++++++++
drivers/net/dsa/sja1105/sja1105_tas.h | 44 +++
6 files changed, 499 insertions(+), 1 deletion(-)
create mode 100644 drivers/net/dsa/sja1105/sja1105_tas.c
create mode 100644 drivers/net/dsa/sja1105/sja1105_tas.h
diff --git a/drivers/net/dsa/sja1105/Kconfig b/drivers/net/dsa/sja1105/Kconfig
index e26666b1ecb9..4dc873e985e6 100644
--- a/drivers/net/dsa/sja1105/Kconfig
+++ b/drivers/net/dsa/sja1105/Kconfig
@@ -32,3 +32,11 @@ config NET_DSA_SJA1105_PTP
help
This enables support for timestamping and PTP clock manipulations in
the SJA1105 DSA driver.
+
+config NET_DSA_SJA1105_TAS
+ bool "Support for the Time-Aware Scheduler on NXP SJA1105"
+ depends on NET_DSA_SJA1105
+ help
+ This enables support for the TTEthernet-based egress scheduling
+ engine in the SJA1105 DSA driver, which is controlled using a
+ hardware offload of the tc-tqprio qdisc.
diff --git a/drivers/net/dsa/sja1105/Makefile b/drivers/net/dsa/sja1105/Makefile
index 4483113e6259..66161e874344 100644
--- a/drivers/net/dsa/sja1105/Makefile
+++ b/drivers/net/dsa/sja1105/Makefile
@@ -12,3 +12,7 @@ sja1105-objs := \
ifdef CONFIG_NET_DSA_SJA1105_PTP
sja1105-objs += sja1105_ptp.o
endif
+
+ifdef CONFIG_NET_DSA_SJA1105_TAS
+sja1105-objs += sja1105_tas.o
+endif
diff --git a/drivers/net/dsa/sja1105/sja1105.h b/drivers/net/dsa/sja1105/sja1105.h
index 78094db32622..e53e494c22e0 100644
--- a/drivers/net/dsa/sja1105/sja1105.h
+++ b/drivers/net/dsa/sja1105/sja1105.h
@@ -20,6 +20,8 @@
*/
#define SJA1105_AGEING_TIME_MS(ms) ((ms) / 10)
+#include "sja1105_tas.h"
+
/* Keeps the different addresses between E/T and P/Q/R/S */
struct sja1105_regs {
u64 device_id;
@@ -104,6 +106,7 @@ struct sja1105_private {
*/
struct mutex mgmt_lock;
struct sja1105_tagger_data tagger_data;
+ struct sja1105_tas_data tas_data;
};
#include "sja1105_dynamic_config.h"
@@ -120,6 +123,9 @@ typedef enum {
SPI_WRITE = 1,
} sja1105_spi_rw_mode_t;
+/* From sja1105_main.c */
+int sja1105_static_config_reload(struct sja1105_private *priv);
+
/* From sja1105_spi.c */
int sja1105_spi_send_packed_buf(const struct sja1105_private *priv,
sja1105_spi_rw_mode_t rw, u64 reg_addr,
diff --git a/drivers/net/dsa/sja1105/sja1105_main.c b/drivers/net/dsa/sja1105/sja1105_main.c
index 0b0205abc3d2..bb9ca761bb94 100644
--- a/drivers/net/dsa/sja1105/sja1105_main.c
+++ b/drivers/net/dsa/sja1105/sja1105_main.c
@@ -22,6 +22,7 @@
#include <linux/if_ether.h>
#include <linux/dsa/8021q.h>
#include "sja1105.h"
+#include "sja1105_tas.h"
static void sja1105_hw_reset(struct gpio_desc *gpio, unsigned int pulse_len,
unsigned int startup_delay)
@@ -1382,7 +1383,7 @@ static void sja1105_bridge_leave(struct dsa_switch *ds, int port,
* modify at runtime (currently only MAC) and restore them after uploading,
* such that this operation is relatively seamless.
*/
-static int sja1105_static_config_reload(struct sja1105_private *priv)
+int sja1105_static_config_reload(struct sja1105_private *priv)
{
struct sja1105_mac_config_entry *mac;
int speed_mbps[SJA1105_NUM_PORTS];
@@ -1727,6 +1728,7 @@ static void sja1105_teardown(struct dsa_switch *ds)
{
struct sja1105_private *priv = ds->priv;
+ sja1105_tas_teardown(priv);
cancel_work_sync(&priv->tagger_data.rxtstamp_work);
skb_queue_purge(&priv->tagger_data.skb_rxtstamp_queue);
sja1105_ptp_clock_unregister(priv);
@@ -2056,6 +2058,18 @@ static bool sja1105_port_txtstamp(struct dsa_switch *ds, int port,
return true;
}
+static int sja1105_port_setup_tc(struct dsa_switch *ds, int port,
+ enum tc_setup_type type,
+ void *type_data)
+{
+ switch (type) {
+ case TC_SETUP_QDISC_TAPRIO:
+ return sja1105_setup_tc_taprio(ds, port, type_data);
+ default:
+ return -EOPNOTSUPP;
+ }
+}
+
static const struct dsa_switch_ops sja1105_switch_ops = {
.get_tag_protocol = sja1105_get_tag_protocol,
.setup = sja1105_setup,
@@ -2088,6 +2102,7 @@ static const struct dsa_switch_ops sja1105_switch_ops = {
.port_hwtstamp_set = sja1105_hwtstamp_set,
.port_rxtstamp = sja1105_port_rxtstamp,
.port_txtstamp = sja1105_port_txtstamp,
+ .port_setup_tc = sja1105_port_setup_tc,
};
static int sja1105_check_device_id(struct sja1105_private *priv)
@@ -2197,6 +2212,8 @@ static int sja1105_probe(struct spi_device *spi)
}
mutex_init(&priv->mgmt_lock);
+ sja1105_tas_setup(priv);
+
return dsa_register_switch(priv->ds);
}
diff --git a/drivers/net/dsa/sja1105/sja1105_tas.c b/drivers/net/dsa/sja1105/sja1105_tas.c
new file mode 100644
index 000000000000..fd4fffecb901
--- /dev/null
+++ b/drivers/net/dsa/sja1105/sja1105_tas.c
@@ -0,0 +1,419 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2019, Vladimir Oltean <olteanv@gmail.com>
+ */
+#include "sja1105.h"
+
+#define SJA1105_TAS_CLKSRC_DISABLED 0
+#define SJA1105_TAS_CLKSRC_STANDALONE 1
+#define SJA1105_TAS_CLKSRC_AS6802 2
+#define SJA1105_TAS_CLKSRC_PTP 3
+#define SJA1105_GATE_MASK GENMASK_ULL(SJA1105_NUM_TC - 1, 0)
+#define SJA1105_TAS_MAX_DELTA BIT(19)
+
+/* This is not a preprocessor macro because the "ns" argument may or may not be
+ * s64 at caller side. This ensures it is properly type-cast before div_s64.
+ */
+static s64 ns_to_sja1105_delta(s64 ns)
+{
+ return div_s64(ns, 200);
+}
+
+/* Lo and behold: the egress scheduler from hell.
+ *
+ * At the hardware level, the Time-Aware Shaper holds a global linear arrray of
+ * all schedule entries for all ports. These are the Gate Control List (GCL)
+ * entries, let's call them "timeslots" for short. This linear array of
+ * timeslots is held in BLK_IDX_SCHEDULE.
+ *
+ * Then there are a maximum of 8 "execution threads" inside the switch, which
+ * iterate cyclically through the "schedule". Each "cycle" has an entry point
+ * and an exit point, both being timeslot indices in the schedule table. The
+ * hardware calls each cycle a "subschedule".
+ *
+ * Subschedule (cycle) i starts when
+ * ptpclkval >= ptpschtm + BLK_IDX_SCHEDULE_ENTRY_POINTS[i].delta.
+ *
+ * The hardware scheduler iterates BLK_IDX_SCHEDULE with a k ranging from
+ * k = BLK_IDX_SCHEDULE_ENTRY_POINTS[i].address to
+ * k = BLK_IDX_SCHEDULE_PARAMS.subscheind[i]
+ *
+ * For each schedule entry (timeslot) k, the engine executes the gate control
+ * list entry for the duration of BLK_IDX_SCHEDULE[k].delta.
+ *
+ * +---------+
+ * | | BLK_IDX_SCHEDULE_ENTRY_POINTS_PARAMS
+ * +---------+
+ * |
+ * +-----------------+
+ * | .actsubsch
+ * BLK_IDX_SCHEDULE_ENTRY_POINTS v
+ * +-------+-------+
+ * |cycle 0|cycle 1|
+ * +-------+-------+
+ * | | | |
+ * +----------------+ | | +-------------------------------------+
+ * | .subschindx | | .subschindx |
+ * | | +---------------+ |
+ * | .address | .address | |
+ * | | | |
+ * | | | |
+ * | BLK_IDX_SCHEDULE v v |
+ * | +-------+-------+-------+-------+-------+------+ |
+ * | |entry 0|entry 1|entry 2|entry 3|entry 4|entry5| |
+ * | +-------+-------+-------+-------+-------+------+ |
+ * | ^ ^ ^ ^ |
+ * | | | | | |
+ * | +-------------------------+ | | | |
+ * | | +-------------------------------+ | | |
+ * | | | +-------------------+ | |
+ * | | | | | |
+ * | +---------------------------------------------------------------+ |
+ * | |subscheind[0]<=subscheind[1]<=subscheind[2]<=...<=subscheind[7]| |
+ * | +---------------------------------------------------------------+ |
+ * | ^ ^ BLK_IDX_SCHEDULE_PARAMS |
+ * | | | |
+ * +--------+ +-------------------------------------------+
+ *
+ * In the above picture there are two subschedules (cycles):
+ *
+ * - cycle 0: iterates the schedule table from 0 to 2 (and back)
+ * - cycle 1: iterates the schedule table from 3 to 5 (and back)
+ *
+ * All other possible execution threads must be marked as unused by making
+ * their "subschedule end index" (subscheind) equal to the last valid
+ * subschedule's end index (in this case 5).
+ */
+static int sja1105_init_scheduling(struct sja1105_private *priv)
+{
+ struct sja1105_schedule_entry_points_entry *schedule_entry_points;
+ struct sja1105_schedule_entry_points_params_entry
+ *schedule_entry_points_params;
+ struct sja1105_schedule_params_entry *schedule_params;
+ struct sja1105_tas_data *tas_data = &priv->tas_data;
+ struct sja1105_schedule_entry *schedule;
+ struct sja1105_table *table;
+ int subscheind[8] = {0};
+ int schedule_start_idx;
+ s64 entry_point_delta;
+ int schedule_end_idx;
+ int num_entries = 0;
+ int num_cycles = 0;
+ int cycle = 0;
+ int i, k = 0;
+ int port;
+
+ /* Discard previous Schedule Table */
+ table = &priv->static_config.tables[BLK_IDX_SCHEDULE];
+ if (table->entry_count) {
+ kfree(table->entries);
+ table->entry_count = 0;
+ }
+
+ /* Discard previous Schedule Entry Points Parameters Table */
+ table = &priv->static_config.tables[BLK_IDX_SCHEDULE_ENTRY_POINTS_PARAMS];
+ if (table->entry_count) {
+ kfree(table->entries);
+ table->entry_count = 0;
+ }
+
+ /* Discard previous Schedule Parameters Table */
+ table = &priv->static_config.tables[BLK_IDX_SCHEDULE_PARAMS];
+ if (table->entry_count) {
+ kfree(table->entries);
+ table->entry_count = 0;
+ }
+
+ /* Discard previous Schedule Entry Points Table */
+ table = &priv->static_config.tables[BLK_IDX_SCHEDULE_ENTRY_POINTS];
+ if (table->entry_count) {
+ kfree(table->entries);
+ table->entry_count = 0;
+ }
+
+ /* Figure out the dimensioning of the problem */
+ for (port = 0; port < SJA1105_NUM_PORTS; port++) {
+ if (tas_data->offload[port]) {
+ num_entries += tas_data->offload[port]->num_entries;
+ num_cycles++;
+ }
+ }
+
+ /* Nothing to do */
+ if (!num_cycles)
+ return 0;
+
+ /* Pre-allocate space in the static config tables */
+
+ /* Schedule Table */
+ table = &priv->static_config.tables[BLK_IDX_SCHEDULE];
+ table->entries = kcalloc(num_entries, table->ops->unpacked_entry_size,
+ GFP_KERNEL);
+ if (!table->entries)
+ return -ENOMEM;
+ table->entry_count = num_entries;
+ schedule = table->entries;
+
+ /* Schedule Points Parameters Table */
+ table = &priv->static_config.tables[BLK_IDX_SCHEDULE_ENTRY_POINTS_PARAMS];
+ table->entries = kcalloc(SJA1105_MAX_SCHEDULE_ENTRY_POINTS_PARAMS_COUNT,
+ table->ops->unpacked_entry_size, GFP_KERNEL);
+ if (!table->entries)
+ return -ENOMEM;
+ table->entry_count = SJA1105_MAX_SCHEDULE_ENTRY_POINTS_PARAMS_COUNT;
+ schedule_entry_points_params = table->entries;
+
+ /* Schedule Parameters Table */
+ table = &priv->static_config.tables[BLK_IDX_SCHEDULE_PARAMS];
+ table->entries = kcalloc(SJA1105_MAX_SCHEDULE_PARAMS_COUNT,
+ table->ops->unpacked_entry_size, GFP_KERNEL);
+ if (!table->entries)
+ return -ENOMEM;
+ table->entry_count = SJA1105_MAX_SCHEDULE_PARAMS_COUNT;
+ schedule_params = table->entries;
+
+ /* Schedule Entry Points Table */
+ table = &priv->static_config.tables[BLK_IDX_SCHEDULE_ENTRY_POINTS];
+ table->entries = kcalloc(num_cycles, table->ops->unpacked_entry_size,
+ GFP_KERNEL);
+ if (!table->entries)
+ return -ENOMEM;
+ table->entry_count = num_cycles;
+ schedule_entry_points = table->entries;
+
+ /* Finally start populating the static config tables */
+ schedule_entry_points_params->clksrc = SJA1105_TAS_CLKSRC_STANDALONE;
+ schedule_entry_points_params->actsubsch = num_cycles - 1;
+
+ for (port = 0; port < SJA1105_NUM_PORTS; port++) {
+ const struct tc_taprio_qopt_offload *offload;
+
+ offload = tas_data->offload[port];
+ if (!offload)
+ continue;
+
+ schedule_start_idx = k;
+ schedule_end_idx = k + offload->num_entries - 1;
+ /* TODO this is only a relative base time for the subschedule
+ * (relative to PTPSCHTM). But as we're using standalone and
+ * not PTP clock as time reference, leave it like this for now.
+ * Later we'll have to enforce that all ports' base times are
+ * within SJA1105_TAS_MAX_DELTA 200ns cycles of one another.
+ */
+ entry_point_delta = ns_to_sja1105_delta(offload->base_time);
+
+ schedule_entry_points[cycle].subschindx = cycle;
+ schedule_entry_points[cycle].delta = entry_point_delta;
+ schedule_entry_points[cycle].address = schedule_start_idx;
+
+ for (i = cycle; i < 8; i++)
+ subscheind[i] = schedule_end_idx;
+
+ for (i = 0; i < offload->num_entries; i++, k++) {
+ s64 delta_ns = offload->entries[i].interval;
+
+ schedule[k].delta = ns_to_sja1105_delta(delta_ns);
+ schedule[k].destports = BIT(port);
+ schedule[k].resmedia_en = true;
+ schedule[k].resmedia = SJA1105_GATE_MASK &
+ ~offload->entries[i].gate_mask;
+ }
+ cycle++;
+ }
+
+ for (i = 0; i < 8; i++)
+ schedule_params->subscheind[i] = subscheind[i];
+
+ return 0;
+}
+
+/* Be there 2 port subschedules, each executing an arbitrary number of gate
+ * open/close events cyclically.
+ * None of those gate events must ever occur at the exact same time, otherwise
+ * the switch is known to act in exotically strange ways.
+ * However the hardware doesn't bother performing these integrity checks - the
+ * designers probably said "nah, let's leave that to the experts" - oh well,
+ * now we're the experts.
+ * So here we are with the task of validating whether the @new offload has any
+ * conflict with the already established TAS configuration in tas_data->offload.
+ * We already know the other ports are in harmony with one another, otherwise
+ * we wouldn't have saved them.
+ * Each gate event executes periodically, with a period of @cycle_time and a
+ * phase given by its cycle's @base_time plus its offset within the cycle
+ * (which in turn is given by the length of the events prior to it).
+ * There are two aspects to possible collisions:
+ * - Collisions within one cycle's (actually the longest cycle's) time frame.
+ * For that, we need to compare the cartesian product of each possible
+ * occurrence of each event within one cycle time.
+ * - Collisions in the future. Events may not collide within one cycle time,
+ * but if two port schedules don't have the same periodicity (aka the cycle
+ * times aren't multiples of one another), they surely will some time in the
+ * future (actually they will collide an infinite amount of times).
+ */
+static bool
+sja1105_tas_check_conflicts(struct sja1105_private *priv,
+ const struct tc_taprio_qopt_offload *new)
+{
+ struct sja1105_tas_data *tas_data = &priv->tas_data;
+ int port;
+
+ for (port = 0; port < SJA1105_NUM_PORTS; port++) {
+ const struct tc_taprio_qopt_offload *offload;
+ s64 max_cycle_time, min_cycle_time;
+ s64 delta1, delta2;
+ s64 rbt1, rbt2;
+ s64 stop_time;
+ s64 t1, t2;
+ int i, j;
+ s32 rem;
+
+ offload = tas_data->offload[port];
+
+ if (!offload)
+ continue;
+
+ /* Check if the two cycle times are multiples of one another.
+ * If they aren't, then they will surely collide.
+ */
+ max_cycle_time = max(offload->cycle_time, new->cycle_time);
+ min_cycle_time = min(offload->cycle_time, new->cycle_time);
+ div_s64_rem(max_cycle_time, min_cycle_time, &rem);
+ if (rem)
+ return true;
+
+ /* Calculate the "reduced" base time of each of the two cycles
+ * (transposed back as close to 0 as possible) by dividing to
+ * the cycle time.
+ */
+ div_s64_rem(offload->base_time, offload->cycle_time, &rem);
+ rbt1 = rem;
+
+ div_s64_rem(new->base_time, new->cycle_time, &rem);
+ rbt2 = rem;
+
+ stop_time = max_cycle_time + max(rbt1, rbt2);
+
+ /* delta1 is the relative base time of each GCL entry within
+ * the established ports' TAS config.
+ */
+ for (i = 0, delta1 = 0;
+ i < offload->num_entries;
+ delta1 += offload->entries[i].interval, i++) {
+
+ /* delta2 is the relative base time of each GCL entry
+ * within the newly added TAS config.
+ */
+ for (j = 0, delta2 = 0;
+ j < new->num_entries;
+ delta2 += new->entries[j].interval, j++) {
+
+ /* t1 follows all possible occurrences of the
+ * established ports' GCL entry i within the
+ * first cycle time.
+ */
+ for (t1 = rbt1 + delta1;
+ t1 <= stop_time;
+ t1 += offload->cycle_time) {
+
+ /* t2 follows all possible occurrences
+ * of the newly added GCL entry j
+ * within the first cycle time.
+ */
+ for (t2 = rbt2 + delta2;
+ t2 <= stop_time;
+ t2 += new->cycle_time) {
+
+ if (t1 == t2) {
+ dev_warn(priv->ds->dev,
+ "GCL entry %d collides with entry %d of port %d\n",
+ j, i, port);
+ return true;
+ }
+ }
+ }
+ }
+ }
+ }
+
+ return false;
+}
+
+int sja1105_setup_tc_taprio(struct dsa_switch *ds, int port,
+ struct tc_taprio_qopt_offload *offload)
+{
+ struct sja1105_private *priv = ds->priv;
+ struct sja1105_tas_data *tas_data = &priv->tas_data;
+ int rc, i;
+
+ /* Can't change an already configured port (must delete qdisc first).
+ * Can't delete the qdisc from an unconfigured port.
+ */
+ if (!!tas_data->offload[port] == offload->enable)
+ return -EINVAL;
+
+ if (!offload->enable) {
+ taprio_offload_free(tas_data->offload[port]);
+ tas_data->offload[port] = NULL;
+
+ rc = sja1105_init_scheduling(priv);
+ if (rc < 0)
+ return rc;
+
+ return sja1105_static_config_reload(priv);
+ }
+
+ /* The cycle time extension is the amount of time the last cycle from
+ * the old OPER needs to be extended in order to phase-align with the
+ * base time of the ADMIN when that becomes the new OPER.
+ * But of course our switch needs to be reset to switch-over between
+ * the ADMIN and the OPER configs - so much for a seamless transition.
+ * So don't add insult over injury and just say we don't support cycle
+ * time extension.
+ */
+ if (offload->cycle_time_extension)
+ return -ENOTSUPP;
+
+ if (!ns_to_sja1105_delta(offload->base_time)) {
+ dev_err(ds->dev, "A base time of zero is not hardware-allowed\n");
+ return -ERANGE;
+ }
+
+ for (i = 0; i < offload->num_entries; i++) {
+ s64 delta_ns = offload->entries[i].interval;
+ s64 delta_cycles = ns_to_sja1105_delta(delta_ns);
+ bool too_long, too_short;
+
+ too_long = (delta_cycles >= SJA1105_TAS_MAX_DELTA);
+ too_short = (delta_cycles == 0);
+ if (too_long || too_short) {
+ dev_err(priv->ds->dev,
+ "Interval %llu too %s for GCL entry %d\n",
+ delta_ns, too_long ? "long" : "short", i);
+ return -ERANGE;
+ }
+ }
+
+ if (sja1105_tas_check_conflicts(priv, offload))
+ return -ERANGE;
+
+ tas_data->offload[port] = taprio_offload_get(offload);
+
+ rc = sja1105_init_scheduling(priv);
+ if (rc < 0)
+ return rc;
+
+ return sja1105_static_config_reload(priv);
+}
+
+void sja1105_tas_setup(struct sja1105_private *priv)
+{
+}
+
+void sja1105_tas_teardown(struct sja1105_private *priv)
+{
+ struct sja1105_tas_data *tas_data = &priv->tas_data;
+ int port;
+
+ for (port = 0; port < SJA1105_NUM_PORTS; port++)
+ if (tas_data->offload[port])
+ taprio_offload_free(tas_data->offload[port]);
+}
diff --git a/drivers/net/dsa/sja1105/sja1105_tas.h b/drivers/net/dsa/sja1105/sja1105_tas.h
new file mode 100644
index 000000000000..a7dc03b876de
--- /dev/null
+++ b/drivers/net/dsa/sja1105/sja1105_tas.h
@@ -0,0 +1,44 @@
+/* SPDX-License-Identifier: GPL-2.0
+ * Copyright (c) 2019, Vladimir Oltean <olteanv@gmail.com>
+ */
+#ifndef _SJA1105_TAS_H
+#define _SJA1105_TAS_H
+
+#include <net/pkt_sched.h>
+
+#if IS_ENABLED(CONFIG_NET_DSA_SJA1105_TAS)
+
+struct sja1105_tas_data {
+ struct tc_taprio_qopt_offload *offload[SJA1105_NUM_PORTS];
+};
+
+struct sja1105_private;
+
+int sja1105_setup_tc_taprio(struct dsa_switch *ds, int port,
+ struct tc_taprio_qopt_offload *qopt);
+
+void sja1105_tas_setup(struct sja1105_private *priv);
+
+void sja1105_tas_teardown(struct sja1105_private *priv);
+
+#else
+
+/* C doesn't allow empty structures, bah! */
+struct sja1105_tas_data {
+ u8 dummy;
+};
+
+static inline int
+sja1105_setup_tc_taprio(struct dsa_switch *ds, int port,
+ struct tc_taprio_qopt_offload *qopt)
+{
+ return -EOPNOTSUPP;
+}
+
+static inline void sja1105_tas_setup(struct sja1105_private *priv) { }
+
+static inline void sja1105_tas_teardown(struct sja1105_private *priv) { }
+
+#endif /* IS_ENABLED(CONFIG_NET_DSA_SJA1105_TAS) */
+
+#endif /* _SJA1105_TAS_H */
--
2.17.1
^ permalink raw reply related
* [PATCH v2 net-next 5/7] net: dsa: sja1105: Make HOSTPRIO a kernel config
From: Vladimir Oltean @ 2019-09-14 1:18 UTC (permalink / raw)
To: f.fainelli, vivien.didelot, andrew, davem, vinicius.gomes,
vedang.patel, richardcochran
Cc: weifeng.voon, jiri, m-karicheri2, Jose.Abreu, ilias.apalodimas,
jhs, xiyou.wangcong, kurt.kanzenbach, joergen.andreasen, netdev,
Vladimir Oltean
In-Reply-To: <20190914011802.1602-1-olteanv@gmail.com>
Unfortunately with this hardware, there is no way to transmit in-band
QoS hints with management frames (i.e. VLAN PCP is ignored). The traffic
class for these is fixed in the static config (which in turn requires a
reset to change).
With the new ability to add time gates for individual traffic classes,
there is a real danger that the user might unknowingly turn off the
traffic class for PTP, BPDUs, LLDP etc.
So we need to manage this situation the best we can. There isn't any
knob in Linux for this, and changing it at runtime probably isn't worth
it either. So just make the setting loud enough by promoting it to a
Kconfig, which the user can customize to their particular setup.
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
---
Changes since v1:
- None.
Changes since RFC:
- None.
drivers/net/dsa/sja1105/Kconfig | 9 +++++++++
drivers/net/dsa/sja1105/sja1105_main.c | 2 +-
2 files changed, 10 insertions(+), 1 deletion(-)
diff --git a/drivers/net/dsa/sja1105/Kconfig b/drivers/net/dsa/sja1105/Kconfig
index 770134a66e48..e26666b1ecb9 100644
--- a/drivers/net/dsa/sja1105/Kconfig
+++ b/drivers/net/dsa/sja1105/Kconfig
@@ -17,6 +17,15 @@ tristate "NXP SJA1105 Ethernet switch family support"
- SJA1105R (Gen. 2, SGMII, No TT-Ethernet)
- SJA1105S (Gen. 2, SGMII, TT-Ethernet)
+config NET_DSA_SJA1105_HOSTPRIO
+ int "Traffic class for management traffic"
+ range 0 7
+ default 7
+ depends on NET_DSA_SJA1105
+ help
+ Configure the traffic class which will be used for management
+ (link-local) traffic sent and received over switch ports.
+
config NET_DSA_SJA1105_PTP
bool "Support for the PTP clock on the NXP SJA1105 Ethernet switch"
depends on NET_DSA_SJA1105
diff --git a/drivers/net/dsa/sja1105/sja1105_main.c b/drivers/net/dsa/sja1105/sja1105_main.c
index 108f62c27c28..0b0205abc3d2 100644
--- a/drivers/net/dsa/sja1105/sja1105_main.c
+++ b/drivers/net/dsa/sja1105/sja1105_main.c
@@ -387,7 +387,7 @@ static int sja1105_init_general_params(struct sja1105_private *priv)
/* Priority queue for link-local management frames
* (both ingress to and egress from CPU - PTP, STP etc)
*/
- .hostprio = 7,
+ .hostprio = CONFIG_NET_DSA_SJA1105_HOSTPRIO,
.mac_fltres1 = SJA1105_LINKLOCAL_FILTER_A,
.mac_flt1 = SJA1105_LINKLOCAL_FILTER_A_MASK,
.incl_srcpt1 = false,
--
2.17.1
^ permalink raw reply related
* [PATCH v2 net-next 3/7] net: dsa: sja1105: Add static config tables for scheduling
From: Vladimir Oltean @ 2019-09-14 1:17 UTC (permalink / raw)
To: f.fainelli, vivien.didelot, andrew, davem, vinicius.gomes,
vedang.patel, richardcochran
Cc: weifeng.voon, jiri, m-karicheri2, Jose.Abreu, ilias.apalodimas,
jhs, xiyou.wangcong, kurt.kanzenbach, joergen.andreasen, netdev,
Vladimir Oltean
In-Reply-To: <20190914011802.1602-1-olteanv@gmail.com>
In order to support tc-taprio offload, the TTEthernet egress scheduling
core registers must be made visible through the static interface.
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
---
Changes since v1:
- None.
Changes since RFC:
- None.
.../net/dsa/sja1105/sja1105_dynamic_config.c | 8 +
.../net/dsa/sja1105/sja1105_static_config.c | 167 ++++++++++++++++++
.../net/dsa/sja1105/sja1105_static_config.h | 48 ++++-
3 files changed, 222 insertions(+), 1 deletion(-)
diff --git a/drivers/net/dsa/sja1105/sja1105_dynamic_config.c b/drivers/net/dsa/sja1105/sja1105_dynamic_config.c
index 9988c9d18567..91da430045ff 100644
--- a/drivers/net/dsa/sja1105/sja1105_dynamic_config.c
+++ b/drivers/net/dsa/sja1105/sja1105_dynamic_config.c
@@ -488,6 +488,8 @@ sja1105et_general_params_entry_packing(void *buf, void *entry_ptr,
/* SJA1105E/T: First generation */
struct sja1105_dynamic_table_ops sja1105et_dyn_ops[BLK_IDX_MAX_DYN] = {
+ [BLK_IDX_SCHEDULE] = {0},
+ [BLK_IDX_SCHEDULE_ENTRY_POINTS] = {0},
[BLK_IDX_L2_LOOKUP] = {
.entry_packing = sja1105et_dyn_l2_lookup_entry_packing,
.cmd_packing = sja1105et_l2_lookup_cmd_packing,
@@ -529,6 +531,8 @@ struct sja1105_dynamic_table_ops sja1105et_dyn_ops[BLK_IDX_MAX_DYN] = {
.packed_size = SJA1105ET_SIZE_MAC_CONFIG_DYN_CMD,
.addr = 0x36,
},
+ [BLK_IDX_SCHEDULE_PARAMS] = {0},
+ [BLK_IDX_SCHEDULE_ENTRY_POINTS_PARAMS] = {0},
[BLK_IDX_L2_LOOKUP_PARAMS] = {
.entry_packing = sja1105et_l2_lookup_params_entry_packing,
.cmd_packing = sja1105et_l2_lookup_params_cmd_packing,
@@ -552,6 +556,8 @@ struct sja1105_dynamic_table_ops sja1105et_dyn_ops[BLK_IDX_MAX_DYN] = {
/* SJA1105P/Q/R/S: Second generation */
struct sja1105_dynamic_table_ops sja1105pqrs_dyn_ops[BLK_IDX_MAX_DYN] = {
+ [BLK_IDX_SCHEDULE] = {0},
+ [BLK_IDX_SCHEDULE_ENTRY_POINTS] = {0},
[BLK_IDX_L2_LOOKUP] = {
.entry_packing = sja1105pqrs_dyn_l2_lookup_entry_packing,
.cmd_packing = sja1105pqrs_l2_lookup_cmd_packing,
@@ -593,6 +599,8 @@ struct sja1105_dynamic_table_ops sja1105pqrs_dyn_ops[BLK_IDX_MAX_DYN] = {
.packed_size = SJA1105PQRS_SIZE_MAC_CONFIG_DYN_CMD,
.addr = 0x4B,
},
+ [BLK_IDX_SCHEDULE_PARAMS] = {0},
+ [BLK_IDX_SCHEDULE_ENTRY_POINTS_PARAMS] = {0},
[BLK_IDX_L2_LOOKUP_PARAMS] = {
.entry_packing = sja1105et_l2_lookup_params_entry_packing,
.cmd_packing = sja1105et_l2_lookup_params_cmd_packing,
diff --git a/drivers/net/dsa/sja1105/sja1105_static_config.c b/drivers/net/dsa/sja1105/sja1105_static_config.c
index b31c737dc560..0d03e13e9909 100644
--- a/drivers/net/dsa/sja1105/sja1105_static_config.c
+++ b/drivers/net/dsa/sja1105/sja1105_static_config.c
@@ -371,6 +371,63 @@ size_t sja1105pqrs_mac_config_entry_packing(void *buf, void *entry_ptr,
return size;
}
+static size_t
+sja1105_schedule_entry_points_params_entry_packing(void *buf, void *entry_ptr,
+ enum packing_op op)
+{
+ struct sja1105_schedule_entry_points_params_entry *entry = entry_ptr;
+ const size_t size = SJA1105_SIZE_SCHEDULE_ENTRY_POINTS_PARAMS_ENTRY;
+
+ sja1105_packing(buf, &entry->clksrc, 31, 30, size, op);
+ sja1105_packing(buf, &entry->actsubsch, 29, 27, size, op);
+ return size;
+}
+
+static size_t
+sja1105_schedule_entry_points_entry_packing(void *buf, void *entry_ptr,
+ enum packing_op op)
+{
+ struct sja1105_schedule_entry_points_entry *entry = entry_ptr;
+ const size_t size = SJA1105_SIZE_SCHEDULE_ENTRY_POINTS_ENTRY;
+
+ sja1105_packing(buf, &entry->subschindx, 31, 29, size, op);
+ sja1105_packing(buf, &entry->delta, 28, 11, size, op);
+ sja1105_packing(buf, &entry->address, 10, 1, size, op);
+ return size;
+}
+
+static size_t sja1105_schedule_params_entry_packing(void *buf, void *entry_ptr,
+ enum packing_op op)
+{
+ const size_t size = SJA1105_SIZE_SCHEDULE_PARAMS_ENTRY;
+ struct sja1105_schedule_params_entry *entry = entry_ptr;
+ int offset, i;
+
+ for (i = 0, offset = 16; i < 8; i++, offset += 10)
+ sja1105_packing(buf, &entry->subscheind[i],
+ offset + 9, offset + 0, size, op);
+ return size;
+}
+
+static size_t sja1105_schedule_entry_packing(void *buf, void *entry_ptr,
+ enum packing_op op)
+{
+ const size_t size = SJA1105_SIZE_SCHEDULE_ENTRY;
+ struct sja1105_schedule_entry *entry = entry_ptr;
+
+ sja1105_packing(buf, &entry->winstindex, 63, 54, size, op);
+ sja1105_packing(buf, &entry->winend, 53, 53, size, op);
+ sja1105_packing(buf, &entry->winst, 52, 52, size, op);
+ sja1105_packing(buf, &entry->destports, 51, 47, size, op);
+ sja1105_packing(buf, &entry->setvalid, 46, 46, size, op);
+ sja1105_packing(buf, &entry->txen, 45, 45, size, op);
+ sja1105_packing(buf, &entry->resmedia_en, 44, 44, size, op);
+ sja1105_packing(buf, &entry->resmedia, 43, 36, size, op);
+ sja1105_packing(buf, &entry->vlindex, 35, 26, size, op);
+ sja1105_packing(buf, &entry->delta, 25, 8, size, op);
+ return size;
+}
+
size_t sja1105_vlan_lookup_entry_packing(void *buf, void *entry_ptr,
enum packing_op op)
{
@@ -447,11 +504,15 @@ static void sja1105_table_write_crc(u8 *table_start, u8 *crc_ptr)
* before blindly indexing kernel memory with the blk_idx.
*/
static u64 blk_id_map[BLK_IDX_MAX] = {
+ [BLK_IDX_SCHEDULE] = BLKID_SCHEDULE,
+ [BLK_IDX_SCHEDULE_ENTRY_POINTS] = BLKID_SCHEDULE_ENTRY_POINTS,
[BLK_IDX_L2_LOOKUP] = BLKID_L2_LOOKUP,
[BLK_IDX_L2_POLICING] = BLKID_L2_POLICING,
[BLK_IDX_VLAN_LOOKUP] = BLKID_VLAN_LOOKUP,
[BLK_IDX_L2_FORWARDING] = BLKID_L2_FORWARDING,
[BLK_IDX_MAC_CONFIG] = BLKID_MAC_CONFIG,
+ [BLK_IDX_SCHEDULE_PARAMS] = BLKID_SCHEDULE_PARAMS,
+ [BLK_IDX_SCHEDULE_ENTRY_POINTS_PARAMS] = BLKID_SCHEDULE_ENTRY_POINTS_PARAMS,
[BLK_IDX_L2_LOOKUP_PARAMS] = BLKID_L2_LOOKUP_PARAMS,
[BLK_IDX_L2_FORWARDING_PARAMS] = BLKID_L2_FORWARDING_PARAMS,
[BLK_IDX_AVB_PARAMS] = BLKID_AVB_PARAMS,
@@ -461,6 +522,13 @@ static u64 blk_id_map[BLK_IDX_MAX] = {
const char *sja1105_static_config_error_msg[] = {
[SJA1105_CONFIG_OK] = "",
+ [SJA1105_TTETHERNET_NOT_SUPPORTED] =
+ "schedule-table present, but TTEthernet is "
+ "only supported on T and Q/S",
+ [SJA1105_INCORRECT_TTETHERNET_CONFIGURATION] =
+ "schedule-table present, but one of "
+ "schedule-entry-points-table, schedule-parameters-table or "
+ "schedule-entry-points-parameters table is empty",
[SJA1105_MISSING_L2_POLICING_TABLE] =
"l2-policing-table needs to have at least one entry",
[SJA1105_MISSING_L2_FORWARDING_TABLE] =
@@ -508,6 +576,21 @@ sja1105_static_config_check_valid(const struct sja1105_static_config *config)
#define IS_FULL(blk_idx) \
(tables[blk_idx].entry_count == tables[blk_idx].ops->max_entry_count)
+ if (tables[BLK_IDX_SCHEDULE].entry_count) {
+ if (config->device_id != SJA1105T_DEVICE_ID &&
+ config->device_id != SJA1105QS_DEVICE_ID)
+ return SJA1105_TTETHERNET_NOT_SUPPORTED;
+
+ if (tables[BLK_IDX_SCHEDULE_ENTRY_POINTS].entry_count == 0)
+ return SJA1105_INCORRECT_TTETHERNET_CONFIGURATION;
+
+ if (!IS_FULL(BLK_IDX_SCHEDULE_PARAMS))
+ return SJA1105_INCORRECT_TTETHERNET_CONFIGURATION;
+
+ if (!IS_FULL(BLK_IDX_SCHEDULE_ENTRY_POINTS_PARAMS))
+ return SJA1105_INCORRECT_TTETHERNET_CONFIGURATION;
+ }
+
if (tables[BLK_IDX_L2_POLICING].entry_count == 0)
return SJA1105_MISSING_L2_POLICING_TABLE;
@@ -614,6 +697,8 @@ sja1105_static_config_get_length(const struct sja1105_static_config *config)
/* SJA1105E: First generation, no TTEthernet */
struct sja1105_table_ops sja1105e_table_ops[BLK_IDX_MAX] = {
+ [BLK_IDX_SCHEDULE] = {0},
+ [BLK_IDX_SCHEDULE_ENTRY_POINTS] = {0},
[BLK_IDX_L2_LOOKUP] = {
.packing = sja1105et_l2_lookup_entry_packing,
.unpacked_entry_size = sizeof(struct sja1105_l2_lookup_entry),
@@ -644,6 +729,8 @@ struct sja1105_table_ops sja1105e_table_ops[BLK_IDX_MAX] = {
.packed_entry_size = SJA1105ET_SIZE_MAC_CONFIG_ENTRY,
.max_entry_count = SJA1105_MAX_MAC_CONFIG_COUNT,
},
+ [BLK_IDX_SCHEDULE_PARAMS] = {0},
+ [BLK_IDX_SCHEDULE_ENTRY_POINTS_PARAMS] = {0},
[BLK_IDX_L2_LOOKUP_PARAMS] = {
.packing = sja1105et_l2_lookup_params_entry_packing,
.unpacked_entry_size = sizeof(struct sja1105_l2_lookup_params_entry),
@@ -678,6 +765,18 @@ struct sja1105_table_ops sja1105e_table_ops[BLK_IDX_MAX] = {
/* SJA1105T: First generation, TTEthernet */
struct sja1105_table_ops sja1105t_table_ops[BLK_IDX_MAX] = {
+ [BLK_IDX_SCHEDULE] = {
+ .packing = sja1105_schedule_entry_packing,
+ .unpacked_entry_size = sizeof(struct sja1105_schedule_entry),
+ .packed_entry_size = SJA1105_SIZE_SCHEDULE_ENTRY,
+ .max_entry_count = SJA1105_MAX_SCHEDULE_COUNT,
+ },
+ [BLK_IDX_SCHEDULE_ENTRY_POINTS] = {
+ .packing = sja1105_schedule_entry_points_entry_packing,
+ .unpacked_entry_size = sizeof(struct sja1105_schedule_entry_points_entry),
+ .packed_entry_size = SJA1105_SIZE_SCHEDULE_ENTRY_POINTS_ENTRY,
+ .max_entry_count = SJA1105_MAX_SCHEDULE_ENTRY_POINTS_COUNT,
+ },
[BLK_IDX_L2_LOOKUP] = {
.packing = sja1105et_l2_lookup_entry_packing,
.unpacked_entry_size = sizeof(struct sja1105_l2_lookup_entry),
@@ -708,6 +807,18 @@ struct sja1105_table_ops sja1105t_table_ops[BLK_IDX_MAX] = {
.packed_entry_size = SJA1105ET_SIZE_MAC_CONFIG_ENTRY,
.max_entry_count = SJA1105_MAX_MAC_CONFIG_COUNT,
},
+ [BLK_IDX_SCHEDULE_PARAMS] = {
+ .packing = sja1105_schedule_params_entry_packing,
+ .unpacked_entry_size = sizeof(struct sja1105_schedule_params_entry),
+ .packed_entry_size = SJA1105_SIZE_SCHEDULE_PARAMS_ENTRY,
+ .max_entry_count = SJA1105_MAX_SCHEDULE_PARAMS_COUNT,
+ },
+ [BLK_IDX_SCHEDULE_ENTRY_POINTS_PARAMS] = {
+ .packing = sja1105_schedule_entry_points_params_entry_packing,
+ .unpacked_entry_size = sizeof(struct sja1105_schedule_entry_points_params_entry),
+ .packed_entry_size = SJA1105_SIZE_SCHEDULE_ENTRY_POINTS_PARAMS_ENTRY,
+ .max_entry_count = SJA1105_MAX_SCHEDULE_ENTRY_POINTS_PARAMS_COUNT,
+ },
[BLK_IDX_L2_LOOKUP_PARAMS] = {
.packing = sja1105et_l2_lookup_params_entry_packing,
.unpacked_entry_size = sizeof(struct sja1105_l2_lookup_params_entry),
@@ -742,6 +853,8 @@ struct sja1105_table_ops sja1105t_table_ops[BLK_IDX_MAX] = {
/* SJA1105P: Second generation, no TTEthernet, no SGMII */
struct sja1105_table_ops sja1105p_table_ops[BLK_IDX_MAX] = {
+ [BLK_IDX_SCHEDULE] = {0},
+ [BLK_IDX_SCHEDULE_ENTRY_POINTS] = {0},
[BLK_IDX_L2_LOOKUP] = {
.packing = sja1105pqrs_l2_lookup_entry_packing,
.unpacked_entry_size = sizeof(struct sja1105_l2_lookup_entry),
@@ -772,6 +885,8 @@ struct sja1105_table_ops sja1105p_table_ops[BLK_IDX_MAX] = {
.packed_entry_size = SJA1105PQRS_SIZE_MAC_CONFIG_ENTRY,
.max_entry_count = SJA1105_MAX_MAC_CONFIG_COUNT,
},
+ [BLK_IDX_SCHEDULE_PARAMS] = {0},
+ [BLK_IDX_SCHEDULE_ENTRY_POINTS_PARAMS] = {0},
[BLK_IDX_L2_LOOKUP_PARAMS] = {
.packing = sja1105pqrs_l2_lookup_params_entry_packing,
.unpacked_entry_size = sizeof(struct sja1105_l2_lookup_params_entry),
@@ -806,6 +921,18 @@ struct sja1105_table_ops sja1105p_table_ops[BLK_IDX_MAX] = {
/* SJA1105Q: Second generation, TTEthernet, no SGMII */
struct sja1105_table_ops sja1105q_table_ops[BLK_IDX_MAX] = {
+ [BLK_IDX_SCHEDULE] = {
+ .packing = sja1105_schedule_entry_packing,
+ .unpacked_entry_size = sizeof(struct sja1105_schedule_entry),
+ .packed_entry_size = SJA1105_SIZE_SCHEDULE_ENTRY,
+ .max_entry_count = SJA1105_MAX_SCHEDULE_COUNT,
+ },
+ [BLK_IDX_SCHEDULE_ENTRY_POINTS] = {
+ .packing = sja1105_schedule_entry_points_entry_packing,
+ .unpacked_entry_size = sizeof(struct sja1105_schedule_entry_points_entry),
+ .packed_entry_size = SJA1105_SIZE_SCHEDULE_ENTRY_POINTS_ENTRY,
+ .max_entry_count = SJA1105_MAX_SCHEDULE_ENTRY_POINTS_COUNT,
+ },
[BLK_IDX_L2_LOOKUP] = {
.packing = sja1105pqrs_l2_lookup_entry_packing,
.unpacked_entry_size = sizeof(struct sja1105_l2_lookup_entry),
@@ -836,6 +963,18 @@ struct sja1105_table_ops sja1105q_table_ops[BLK_IDX_MAX] = {
.packed_entry_size = SJA1105PQRS_SIZE_MAC_CONFIG_ENTRY,
.max_entry_count = SJA1105_MAX_MAC_CONFIG_COUNT,
},
+ [BLK_IDX_SCHEDULE_PARAMS] = {
+ .packing = sja1105_schedule_params_entry_packing,
+ .unpacked_entry_size = sizeof(struct sja1105_schedule_params_entry),
+ .packed_entry_size = SJA1105_SIZE_SCHEDULE_PARAMS_ENTRY,
+ .max_entry_count = SJA1105_MAX_SCHEDULE_PARAMS_COUNT,
+ },
+ [BLK_IDX_SCHEDULE_ENTRY_POINTS_PARAMS] = {
+ .packing = sja1105_schedule_entry_points_params_entry_packing,
+ .unpacked_entry_size = sizeof(struct sja1105_schedule_entry_points_params_entry),
+ .packed_entry_size = SJA1105_SIZE_SCHEDULE_ENTRY_POINTS_PARAMS_ENTRY,
+ .max_entry_count = SJA1105_MAX_SCHEDULE_ENTRY_POINTS_PARAMS_COUNT,
+ },
[BLK_IDX_L2_LOOKUP_PARAMS] = {
.packing = sja1105pqrs_l2_lookup_params_entry_packing,
.unpacked_entry_size = sizeof(struct sja1105_l2_lookup_params_entry),
@@ -870,6 +1009,8 @@ struct sja1105_table_ops sja1105q_table_ops[BLK_IDX_MAX] = {
/* SJA1105R: Second generation, no TTEthernet, SGMII */
struct sja1105_table_ops sja1105r_table_ops[BLK_IDX_MAX] = {
+ [BLK_IDX_SCHEDULE] = {0},
+ [BLK_IDX_SCHEDULE_ENTRY_POINTS] = {0},
[BLK_IDX_L2_LOOKUP] = {
.packing = sja1105pqrs_l2_lookup_entry_packing,
.unpacked_entry_size = sizeof(struct sja1105_l2_lookup_entry),
@@ -900,6 +1041,8 @@ struct sja1105_table_ops sja1105r_table_ops[BLK_IDX_MAX] = {
.packed_entry_size = SJA1105PQRS_SIZE_MAC_CONFIG_ENTRY,
.max_entry_count = SJA1105_MAX_MAC_CONFIG_COUNT,
},
+ [BLK_IDX_SCHEDULE_PARAMS] = {0},
+ [BLK_IDX_SCHEDULE_ENTRY_POINTS_PARAMS] = {0},
[BLK_IDX_L2_LOOKUP_PARAMS] = {
.packing = sja1105pqrs_l2_lookup_params_entry_packing,
.unpacked_entry_size = sizeof(struct sja1105_l2_lookup_params_entry),
@@ -934,6 +1077,18 @@ struct sja1105_table_ops sja1105r_table_ops[BLK_IDX_MAX] = {
/* SJA1105S: Second generation, TTEthernet, SGMII */
struct sja1105_table_ops sja1105s_table_ops[BLK_IDX_MAX] = {
+ [BLK_IDX_SCHEDULE] = {
+ .packing = sja1105_schedule_entry_packing,
+ .unpacked_entry_size = sizeof(struct sja1105_schedule_entry),
+ .packed_entry_size = SJA1105_SIZE_SCHEDULE_ENTRY,
+ .max_entry_count = SJA1105_MAX_SCHEDULE_COUNT,
+ },
+ [BLK_IDX_SCHEDULE_ENTRY_POINTS] = {
+ .packing = sja1105_schedule_entry_points_entry_packing,
+ .unpacked_entry_size = sizeof(struct sja1105_schedule_entry_points_entry),
+ .packed_entry_size = SJA1105_SIZE_SCHEDULE_ENTRY_POINTS_ENTRY,
+ .max_entry_count = SJA1105_MAX_SCHEDULE_ENTRY_POINTS_COUNT,
+ },
[BLK_IDX_L2_LOOKUP] = {
.packing = sja1105pqrs_l2_lookup_entry_packing,
.unpacked_entry_size = sizeof(struct sja1105_l2_lookup_entry),
@@ -964,6 +1119,18 @@ struct sja1105_table_ops sja1105s_table_ops[BLK_IDX_MAX] = {
.packed_entry_size = SJA1105PQRS_SIZE_MAC_CONFIG_ENTRY,
.max_entry_count = SJA1105_MAX_MAC_CONFIG_COUNT,
},
+ [BLK_IDX_SCHEDULE_PARAMS] = {
+ .packing = sja1105_schedule_params_entry_packing,
+ .unpacked_entry_size = sizeof(struct sja1105_schedule_params_entry),
+ .packed_entry_size = SJA1105_SIZE_SCHEDULE_PARAMS_ENTRY,
+ .max_entry_count = SJA1105_MAX_SCHEDULE_PARAMS_COUNT,
+ },
+ [BLK_IDX_SCHEDULE_ENTRY_POINTS_PARAMS] = {
+ .packing = sja1105_schedule_entry_points_params_entry_packing,
+ .unpacked_entry_size = sizeof(struct sja1105_schedule_entry_points_params_entry),
+ .packed_entry_size = SJA1105_SIZE_SCHEDULE_ENTRY_POINTS_PARAMS_ENTRY,
+ .max_entry_count = SJA1105_MAX_SCHEDULE_ENTRY_POINTS_PARAMS_COUNT,
+ },
[BLK_IDX_L2_LOOKUP_PARAMS] = {
.packing = sja1105pqrs_l2_lookup_params_entry_packing,
.unpacked_entry_size = sizeof(struct sja1105_l2_lookup_params_entry),
diff --git a/drivers/net/dsa/sja1105/sja1105_static_config.h b/drivers/net/dsa/sja1105/sja1105_static_config.h
index 684465fc0882..7f87022a2d61 100644
--- a/drivers/net/dsa/sja1105/sja1105_static_config.h
+++ b/drivers/net/dsa/sja1105/sja1105_static_config.h
@@ -11,11 +11,15 @@
#define SJA1105_SIZE_DEVICE_ID 4
#define SJA1105_SIZE_TABLE_HEADER 12
+#define SJA1105_SIZE_SCHEDULE_ENTRY 8
+#define SJA1105_SIZE_SCHEDULE_ENTRY_POINTS_ENTRY 4
#define SJA1105_SIZE_L2_POLICING_ENTRY 8
#define SJA1105_SIZE_VLAN_LOOKUP_ENTRY 8
#define SJA1105_SIZE_L2_FORWARDING_ENTRY 8
#define SJA1105_SIZE_L2_FORWARDING_PARAMS_ENTRY 12
#define SJA1105_SIZE_XMII_PARAMS_ENTRY 4
+#define SJA1105_SIZE_SCHEDULE_PARAMS_ENTRY 12
+#define SJA1105_SIZE_SCHEDULE_ENTRY_POINTS_PARAMS_ENTRY 4
#define SJA1105ET_SIZE_L2_LOOKUP_ENTRY 12
#define SJA1105ET_SIZE_MAC_CONFIG_ENTRY 28
#define SJA1105ET_SIZE_L2_LOOKUP_PARAMS_ENTRY 4
@@ -29,11 +33,15 @@
/* UM10944.pdf Page 11, Table 2. Configuration Blocks */
enum {
+ BLKID_SCHEDULE = 0x00,
+ BLKID_SCHEDULE_ENTRY_POINTS = 0x01,
BLKID_L2_LOOKUP = 0x05,
BLKID_L2_POLICING = 0x06,
BLKID_VLAN_LOOKUP = 0x07,
BLKID_L2_FORWARDING = 0x08,
BLKID_MAC_CONFIG = 0x09,
+ BLKID_SCHEDULE_PARAMS = 0x0A,
+ BLKID_SCHEDULE_ENTRY_POINTS_PARAMS = 0x0B,
BLKID_L2_LOOKUP_PARAMS = 0x0D,
BLKID_L2_FORWARDING_PARAMS = 0x0E,
BLKID_AVB_PARAMS = 0x10,
@@ -42,11 +50,15 @@ enum {
};
enum sja1105_blk_idx {
- BLK_IDX_L2_LOOKUP = 0,
+ BLK_IDX_SCHEDULE = 0,
+ BLK_IDX_SCHEDULE_ENTRY_POINTS,
+ BLK_IDX_L2_LOOKUP,
BLK_IDX_L2_POLICING,
BLK_IDX_VLAN_LOOKUP,
BLK_IDX_L2_FORWARDING,
BLK_IDX_MAC_CONFIG,
+ BLK_IDX_SCHEDULE_PARAMS,
+ BLK_IDX_SCHEDULE_ENTRY_POINTS_PARAMS,
BLK_IDX_L2_LOOKUP_PARAMS,
BLK_IDX_L2_FORWARDING_PARAMS,
BLK_IDX_AVB_PARAMS,
@@ -59,11 +71,15 @@ enum sja1105_blk_idx {
BLK_IDX_INVAL = -1,
};
+#define SJA1105_MAX_SCHEDULE_COUNT 1024
+#define SJA1105_MAX_SCHEDULE_ENTRY_POINTS_COUNT 2048
#define SJA1105_MAX_L2_LOOKUP_COUNT 1024
#define SJA1105_MAX_L2_POLICING_COUNT 45
#define SJA1105_MAX_VLAN_LOOKUP_COUNT 4096
#define SJA1105_MAX_L2_FORWARDING_COUNT 13
#define SJA1105_MAX_MAC_CONFIG_COUNT 5
+#define SJA1105_MAX_SCHEDULE_PARAMS_COUNT 1
+#define SJA1105_MAX_SCHEDULE_ENTRY_POINTS_PARAMS_COUNT 1
#define SJA1105_MAX_L2_LOOKUP_PARAMS_COUNT 1
#define SJA1105_MAX_L2_FORWARDING_PARAMS_COUNT 1
#define SJA1105_MAX_GENERAL_PARAMS_COUNT 1
@@ -83,6 +99,23 @@ enum sja1105_blk_idx {
#define SJA1105R_PART_NO 0x9A86
#define SJA1105S_PART_NO 0x9A87
+struct sja1105_schedule_entry {
+ u64 winstindex;
+ u64 winend;
+ u64 winst;
+ u64 destports;
+ u64 setvalid;
+ u64 txen;
+ u64 resmedia_en;
+ u64 resmedia;
+ u64 vlindex;
+ u64 delta;
+};
+
+struct sja1105_schedule_params_entry {
+ u64 subscheind[8];
+};
+
struct sja1105_general_params_entry {
u64 vllupformat;
u64 mirr_ptacu;
@@ -112,6 +145,17 @@ struct sja1105_general_params_entry {
u64 replay_port;
};
+struct sja1105_schedule_entry_points_entry {
+ u64 subschindx;
+ u64 delta;
+ u64 address;
+};
+
+struct sja1105_schedule_entry_points_params_entry {
+ u64 clksrc;
+ u64 actsubsch;
+};
+
struct sja1105_vlan_lookup_entry {
u64 ving_mirr;
u64 vegr_mirr;
@@ -256,6 +300,8 @@ sja1105_static_config_get_length(const struct sja1105_static_config *config);
typedef enum {
SJA1105_CONFIG_OK = 0,
+ SJA1105_TTETHERNET_NOT_SUPPORTED,
+ SJA1105_INCORRECT_TTETHERNET_CONFIGURATION,
SJA1105_MISSING_L2_POLICING_TABLE,
SJA1105_MISSING_L2_FORWARDING_TABLE,
SJA1105_MISSING_L2_FORWARDING_PARAMS_TABLE,
--
2.17.1
^ permalink raw reply related
* [PATCH v2 net-next 4/7] net: dsa: sja1105: Advertise the 8 TX queues
From: Vladimir Oltean @ 2019-09-14 1:17 UTC (permalink / raw)
To: f.fainelli, vivien.didelot, andrew, davem, vinicius.gomes,
vedang.patel, richardcochran
Cc: weifeng.voon, jiri, m-karicheri2, Jose.Abreu, ilias.apalodimas,
jhs, xiyou.wangcong, kurt.kanzenbach, joergen.andreasen, netdev,
Vladimir Oltean
In-Reply-To: <20190914011802.1602-1-olteanv@gmail.com>
This is a preparation patch for the tc-taprio offload (and potentially
for other future offloads such as tc-mqprio).
Instead of looking directly at skb->priority during xmit, let's get the
netdev queue and the queue-to-traffic-class mapping, and put the
resulting traffic class into the dsa_8021q PCP field. The switch is
configured with a 1-to-1 PCP-to-ingress-queue-to-egress-queue mapping
(see vlan_pmap in sja1105_main.c), so the effect is that we can inject
into a front-panel's egress traffic class through VLAN tagging from
Linux, completely transparently.
Unfortunately the switch doesn't look at the VLAN PCP in the case of
management traffic to/from the CPU (link-local frames at
01-80-C2-xx-xx-xx or 01-1B-19-xx-xx-xx) so we can't alter the
transmission queue of this type of traffic on a frame-by-frame basis. It
is only selected through the "hostprio" setting which ATM is harcoded in
the driver to 7.
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
---
Changes since v1:
- None, but the use of netdev_txq_to_tc is now finally correct after
adjusting the gate_mask meaning in the taprio offload structure.
Changes since RFC:
- None.
drivers/net/dsa/sja1105/sja1105_main.c | 7 ++++++-
net/dsa/tag_sja1105.c | 3 ++-
2 files changed, 8 insertions(+), 2 deletions(-)
diff --git a/drivers/net/dsa/sja1105/sja1105_main.c b/drivers/net/dsa/sja1105/sja1105_main.c
index d8cff0107ec4..108f62c27c28 100644
--- a/drivers/net/dsa/sja1105/sja1105_main.c
+++ b/drivers/net/dsa/sja1105/sja1105_main.c
@@ -384,7 +384,9 @@ static int sja1105_init_general_params(struct sja1105_private *priv)
/* Disallow dynamic changing of the mirror port */
.mirr_ptacu = 0,
.switchid = priv->ds->index,
- /* Priority queue for link-local frames trapped to CPU */
+ /* Priority queue for link-local management frames
+ * (both ingress to and egress from CPU - PTP, STP etc)
+ */
.hostprio = 7,
.mac_fltres1 = SJA1105_LINKLOCAL_FILTER_A,
.mac_flt1 = SJA1105_LINKLOCAL_FILTER_A_MASK,
@@ -1711,6 +1713,9 @@ static int sja1105_setup(struct dsa_switch *ds)
*/
ds->vlan_filtering_is_global = true;
+ /* Advertise the 8 egress queues */
+ ds->num_tx_queues = SJA1105_NUM_TC;
+
/* The DSA/switchdev model brings up switch ports in standalone mode by
* default, and that means vlan_filtering is 0 since they're not under
* a bridge, so it's safe to set up switch tagging at this time.
diff --git a/net/dsa/tag_sja1105.c b/net/dsa/tag_sja1105.c
index 47ee88163a9d..9c9aff3e52cf 100644
--- a/net/dsa/tag_sja1105.c
+++ b/net/dsa/tag_sja1105.c
@@ -89,7 +89,8 @@ static struct sk_buff *sja1105_xmit(struct sk_buff *skb,
struct dsa_port *dp = dsa_slave_to_port(netdev);
struct dsa_switch *ds = dp->ds;
u16 tx_vid = dsa_8021q_tx_vid(ds, dp->index);
- u8 pcp = skb->priority;
+ u16 queue_mapping = skb_get_queue_mapping(skb);
+ u8 pcp = netdev_txq_to_tc(netdev, queue_mapping);
/* Transmitting management traffic does not rely upon switch tagging,
* but instead SPI-installed management routes. Part 2 of this
--
2.17.1
^ permalink raw reply related
* [PATCH v2 net-next 1/7] taprio: Add support for hardware offloading
From: Vladimir Oltean @ 2019-09-14 1:17 UTC (permalink / raw)
To: f.fainelli, vivien.didelot, andrew, davem, vinicius.gomes,
vedang.patel, richardcochran
Cc: weifeng.voon, jiri, m-karicheri2, Jose.Abreu, ilias.apalodimas,
jhs, xiyou.wangcong, kurt.kanzenbach, joergen.andreasen, netdev,
Vladimir Oltean
In-Reply-To: <20190914011802.1602-1-olteanv@gmail.com>
From: Vinicius Costa Gomes <vinicius.gomes@intel.com>
This allows taprio to offload the schedule enforcement to capable
network cards, resulting in more precise windows and less CPU usage.
The gate mask acts on traffic classes (groups of queues of same
priority), as specified in IEEE 802.1Q-2018, and following the existing
taprio and mqprio semantics.
It is up to the driver to perform conversion between tc and individual
netdev queues if for some reason it needs to make that distinction.
Full offload is requested from the network interface by specifying
"flags 2" in the tc qdisc creation command, which in turn corresponds to
the TCA_TAPRIO_ATTR_FLAG_FULL_OFFLOAD bit.
The important detail here is the clockid which is implicitly /dev/ptpN
for full offload, and hence not configurable.
A reference counting API is added to support the use case where Ethernet
drivers need to keep the taprio offload structure locally (i.e. they are
a multi-port switch driver, and configuring a port depends on the
settings of other ports as well). The refcount_t variable is kept in a
private structure (__tc_taprio_qopt_offload) and not exposed to drivers.
In the future, the private structure might also be expanded with a
backpointer to taprio_sched *q, to implement the notification system
described in the patch (of when admin became oper, or an error occurred,
etc, so the offload can be monitored with 'tc qdisc show').
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Signed-off-by: Voon Weifeng <weifeng.voon@intel.com>
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
---
Changes since v1:
- Turned the next_sched hrtimer function into a simple
taprio_offload_config_changed function called synchronously (for now)
from taprio_enable_offload. But the idea is that the driver may have a
lot more means to figure out when the admin schedule is no longer
pending (perhaps even an interrupt), so leave an open window for
implementing a notification system from the driver.
- Made it an error to specify 'clockid' with full offload.
- Created a wrapper __tc_taprio_qopt_offload structure which holds the
refcount_t (for now) and maybe a backpointer to the qdisc_priv in the
future.
- Renamed taprio_get and taprio_free to taprio_offload_get and
taprio_offload_free. Renamed the "taprio" variable to "offload".
- Moved the reference counting helper implementations to sch_taprio.c.
- Removed the tc_mask_to_queue_mask manipulation done to the gate_mask
before passing it on to drivers. Instead of netdev queue gates, they
now see a mask of traffic class gates, which:
- They need to care about anyway, if they have a multi-queue device
and they need to configure the queue-to-tc hardware mapping.
- Makes no difference to them if the hardware makes no distinction
between queue and traffic class (there is only one egress queue per
tc, having a fixed priority). The sja1105 hw is in this situation.
Changes since RFC:
- Made the combination of FULL_OFFLOAD and TXTIME_ASSIST invalid.
- Made ndo_setup_tc be called from sleepable context.
- Added a taprio_alloc helper to avoid passing stack memory to drivers.
- Made taprio_disable_offload take the extack as well.
- Conditioned the setup of the software (and txtime-assisted)
implementation of taprio on there not being a full offload in place.
- Fixed a lockdep-related compilation bug.
include/linux/netdevice.h | 1 +
include/net/pkt_sched.h | 23 ++
include/uapi/linux/pkt_sched.h | 3 +-
net/sched/sch_taprio.c | 409 +++++++++++++++++++++++++++++----
4 files changed, 392 insertions(+), 44 deletions(-)
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index d7d5626002e9..9eda1c31d1f7 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -847,6 +847,7 @@ enum tc_setup_type {
TC_SETUP_QDISC_ETF,
TC_SETUP_ROOT_QDISC,
TC_SETUP_QDISC_GRED,
+ TC_SETUP_QDISC_TAPRIO,
};
/* These structures hold the attributes of bpf state that are being passed
diff --git a/include/net/pkt_sched.h b/include/net/pkt_sched.h
index a16fbe9a2a67..d1632979622e 100644
--- a/include/net/pkt_sched.h
+++ b/include/net/pkt_sched.h
@@ -161,4 +161,27 @@ struct tc_etf_qopt_offload {
s32 queue;
};
+struct tc_taprio_sched_entry {
+ u8 command; /* TC_TAPRIO_CMD_* */
+
+ /* The gate_mask in the offloading side refers to traffic classes */
+ u32 gate_mask;
+ u32 interval;
+};
+
+struct tc_taprio_qopt_offload {
+ u8 enable;
+ ktime_t base_time;
+ u64 cycle_time;
+ u64 cycle_time_extension;
+
+ size_t num_entries;
+ struct tc_taprio_sched_entry entries[0];
+};
+
+/* Reference counting */
+struct tc_taprio_qopt_offload *taprio_offload_get(struct tc_taprio_qopt_offload
+ *offload);
+void taprio_offload_free(struct tc_taprio_qopt_offload *offload);
+
#endif
diff --git a/include/uapi/linux/pkt_sched.h b/include/uapi/linux/pkt_sched.h
index 18f185299f47..5011259b8f67 100644
--- a/include/uapi/linux/pkt_sched.h
+++ b/include/uapi/linux/pkt_sched.h
@@ -1160,7 +1160,8 @@ enum {
* [TCA_TAPRIO_ATTR_SCHED_ENTRY_INTERVAL]
*/
-#define TCA_TAPRIO_ATTR_FLAG_TXTIME_ASSIST 0x1
+#define TCA_TAPRIO_ATTR_FLAG_TXTIME_ASSIST BIT(0)
+#define TCA_TAPRIO_ATTR_FLAG_FULL_OFFLOAD BIT(1)
enum {
TCA_TAPRIO_ATTR_UNSPEC,
diff --git a/net/sched/sch_taprio.c b/net/sched/sch_taprio.c
index 84b863e2bdbd..2f7b34205c82 100644
--- a/net/sched/sch_taprio.c
+++ b/net/sched/sch_taprio.c
@@ -29,8 +29,8 @@ static DEFINE_SPINLOCK(taprio_list_lock);
#define TAPRIO_ALL_GATES_OPEN -1
-#define FLAGS_VALID(flags) (!((flags) & ~TCA_TAPRIO_ATTR_FLAG_TXTIME_ASSIST))
#define TXTIME_ASSIST_IS_ENABLED(flags) ((flags) & TCA_TAPRIO_ATTR_FLAG_TXTIME_ASSIST)
+#define FULL_OFFLOAD_IS_ENABLED(flags) ((flags) & TCA_TAPRIO_ATTR_FLAG_FULL_OFFLOAD)
struct sched_entry {
struct list_head list;
@@ -75,9 +75,16 @@ struct taprio_sched {
struct sched_gate_list __rcu *admin_sched;
struct hrtimer advance_timer;
struct list_head taprio_list;
+ struct sk_buff *(*dequeue)(struct Qdisc *sch);
+ struct sk_buff *(*peek)(struct Qdisc *sch);
u32 txtime_delay;
};
+struct __tc_taprio_qopt_offload {
+ refcount_t users;
+ struct tc_taprio_qopt_offload offload;
+};
+
static ktime_t sched_base_time(const struct sched_gate_list *sched)
{
if (!sched)
@@ -268,6 +275,19 @@ static bool is_valid_interval(struct sk_buff *skb, struct Qdisc *sch)
return entry;
}
+static bool taprio_flags_valid(u32 flags)
+{
+ /* Make sure no other flag bits are set. */
+ if (flags & ~(TCA_TAPRIO_ATTR_FLAG_TXTIME_ASSIST |
+ TCA_TAPRIO_ATTR_FLAG_FULL_OFFLOAD))
+ return false;
+ /* txtime-assist and full offload are mutually exclusive */
+ if ((flags & TCA_TAPRIO_ATTR_FLAG_TXTIME_ASSIST) &&
+ (flags & TCA_TAPRIO_ATTR_FLAG_FULL_OFFLOAD))
+ return false;
+ return true;
+}
+
/* This returns the tstamp value set by TCP in terms of the set clock. */
static ktime_t get_tcp_tstamp(struct taprio_sched *q, struct sk_buff *skb)
{
@@ -417,7 +437,7 @@ static int taprio_enqueue(struct sk_buff *skb, struct Qdisc *sch,
return qdisc_enqueue(skb, child, to_free);
}
-static struct sk_buff *taprio_peek(struct Qdisc *sch)
+static struct sk_buff *taprio_peek_soft(struct Qdisc *sch)
{
struct taprio_sched *q = qdisc_priv(sch);
struct net_device *dev = qdisc_dev(sch);
@@ -461,6 +481,36 @@ static struct sk_buff *taprio_peek(struct Qdisc *sch)
return NULL;
}
+static struct sk_buff *taprio_peek_offload(struct Qdisc *sch)
+{
+ struct taprio_sched *q = qdisc_priv(sch);
+ struct net_device *dev = qdisc_dev(sch);
+ struct sk_buff *skb;
+ int i;
+
+ for (i = 0; i < dev->num_tx_queues; i++) {
+ struct Qdisc *child = q->qdiscs[i];
+
+ if (unlikely(!child))
+ continue;
+
+ skb = child->ops->peek(child);
+ if (!skb)
+ continue;
+
+ return skb;
+ }
+
+ return NULL;
+}
+
+static struct sk_buff *taprio_peek(struct Qdisc *sch)
+{
+ struct taprio_sched *q = qdisc_priv(sch);
+
+ return q->peek(sch);
+}
+
static void taprio_set_budget(struct taprio_sched *q, struct sched_entry *entry)
{
atomic_set(&entry->budget,
@@ -468,7 +518,7 @@ static void taprio_set_budget(struct taprio_sched *q, struct sched_entry *entry)
atomic64_read(&q->picos_per_byte)));
}
-static struct sk_buff *taprio_dequeue(struct Qdisc *sch)
+static struct sk_buff *taprio_dequeue_soft(struct Qdisc *sch)
{
struct taprio_sched *q = qdisc_priv(sch);
struct net_device *dev = qdisc_dev(sch);
@@ -550,6 +600,40 @@ static struct sk_buff *taprio_dequeue(struct Qdisc *sch)
return skb;
}
+static struct sk_buff *taprio_dequeue_offload(struct Qdisc *sch)
+{
+ struct taprio_sched *q = qdisc_priv(sch);
+ struct net_device *dev = qdisc_dev(sch);
+ struct sk_buff *skb;
+ int i;
+
+ for (i = 0; i < dev->num_tx_queues; i++) {
+ struct Qdisc *child = q->qdiscs[i];
+
+ if (unlikely(!child))
+ continue;
+
+ skb = child->ops->dequeue(child);
+ if (unlikely(!skb))
+ continue;
+
+ qdisc_bstats_update(sch, skb);
+ qdisc_qstats_backlog_dec(sch, skb);
+ sch->q.qlen--;
+
+ return skb;
+ }
+
+ return NULL;
+}
+
+static struct sk_buff *taprio_dequeue(struct Qdisc *sch)
+{
+ struct taprio_sched *q = qdisc_priv(sch);
+
+ return q->dequeue(sch);
+}
+
static bool should_restart_cycle(const struct sched_gate_list *oper,
const struct sched_entry *entry)
{
@@ -932,6 +1016,9 @@ static void taprio_start_sched(struct Qdisc *sch,
struct taprio_sched *q = qdisc_priv(sch);
ktime_t expires;
+ if (FULL_OFFLOAD_IS_ENABLED(q->flags))
+ return;
+
expires = hrtimer_get_expires(&q->advance_timer);
if (expires == 0)
expires = KTIME_MAX;
@@ -1011,6 +1098,254 @@ static void setup_txtime(struct taprio_sched *q,
}
}
+static struct tc_taprio_qopt_offload *taprio_offload_alloc(int num_entries)
+{
+ size_t size = sizeof(struct tc_taprio_sched_entry) * num_entries +
+ sizeof(struct __tc_taprio_qopt_offload);
+ struct __tc_taprio_qopt_offload *__offload;
+
+ __offload = kzalloc(size, GFP_KERNEL);
+ if (!__offload)
+ return NULL;
+
+ refcount_set(&__offload->users, 1);
+
+ return &__offload->offload;
+}
+
+struct tc_taprio_qopt_offload *taprio_offload_get(struct tc_taprio_qopt_offload
+ *offload)
+{
+ struct __tc_taprio_qopt_offload *__offload;
+
+ __offload = container_of(offload, struct __tc_taprio_qopt_offload,
+ offload);
+
+ refcount_inc(&__offload->users);
+
+ return offload;
+}
+EXPORT_SYMBOL_GPL(taprio_offload_get);
+
+void taprio_offload_free(struct tc_taprio_qopt_offload *offload)
+{
+ struct __tc_taprio_qopt_offload *__offload;
+
+ __offload = container_of(offload, struct __tc_taprio_qopt_offload,
+ offload);
+
+ if (!refcount_dec_and_test(&__offload->users))
+ return;
+
+ kfree(__offload);
+}
+EXPORT_SYMBOL_GPL(taprio_offload_free);
+
+/* The function will only serve to keep the pointers to the "oper" and "admin"
+ * schedules valid in relation to their base times, so when calling dump() the
+ * users looks at the right schedules.
+ * When using full offload, the admin configuration is promoted to oper at the
+ * base_time in the PHC time domain. But because the system time is not
+ * necessarily in sync with that, we can't just trigger a hrtimer to call
+ * switch_schedules at the right hardware time.
+ * At the moment we call this by hand right away from taprio, but in the future
+ * it will be useful to create a mechanism for drivers to notify taprio of the
+ * offload state (PENDING, ACTIVE, INACTIVE) so it can be visible in dump().
+ * This is left as TODO.
+ */
+void taprio_offload_config_changed(struct taprio_sched *q)
+{
+ struct sched_gate_list *oper, *admin;
+
+ spin_lock(&q->current_entry_lock);
+
+ oper = rcu_dereference_protected(q->oper_sched,
+ lockdep_is_held(&q->current_entry_lock));
+ admin = rcu_dereference_protected(q->admin_sched,
+ lockdep_is_held(&q->current_entry_lock));
+
+ switch_schedules(q, &admin, &oper);
+
+ spin_unlock(&q->current_entry_lock);
+}
+
+static void taprio_sched_to_offload(struct taprio_sched *q,
+ struct sched_gate_list *sched,
+ const struct tc_mqprio_qopt *mqprio,
+ struct tc_taprio_qopt_offload *offload)
+{
+ struct sched_entry *entry;
+ int i = 0;
+
+ offload->base_time = sched->base_time;
+ offload->cycle_time = sched->cycle_time;
+ offload->cycle_time_extension = sched->cycle_time_extension;
+
+ list_for_each_entry(entry, &sched->entries, list) {
+ struct tc_taprio_sched_entry *e = &offload->entries[i];
+
+ e->command = entry->command;
+ e->interval = entry->interval;
+ e->gate_mask = entry->gate_mask;
+ i++;
+ }
+
+ offload->num_entries = i;
+}
+
+static int taprio_enable_offload(struct net_device *dev,
+ struct tc_mqprio_qopt *mqprio,
+ struct taprio_sched *q,
+ struct sched_gate_list *sched,
+ struct netlink_ext_ack *extack)
+{
+ const struct net_device_ops *ops = dev->netdev_ops;
+ struct tc_taprio_qopt_offload *offload;
+ int err = 0;
+
+ if (!ops->ndo_setup_tc) {
+ NL_SET_ERR_MSG(extack,
+ "Device does not support taprio offload");
+ return -EOPNOTSUPP;
+ }
+
+ offload = taprio_offload_alloc(sched->num_entries);
+ if (!offload) {
+ NL_SET_ERR_MSG(extack,
+ "Not enough memory for enabling offload mode");
+ return -ENOMEM;
+ }
+ offload->enable = 1;
+ taprio_sched_to_offload(q, sched, mqprio, offload);
+
+ err = ops->ndo_setup_tc(dev, TC_SETUP_QDISC_TAPRIO, offload);
+ if (err < 0) {
+ NL_SET_ERR_MSG(extack,
+ "Device failed to setup taprio offload");
+ goto done;
+ }
+
+ taprio_offload_config_changed(q);
+
+done:
+ taprio_offload_free(offload);
+
+ return err;
+}
+
+static int taprio_disable_offload(struct net_device *dev,
+ struct taprio_sched *q,
+ struct netlink_ext_ack *extack)
+{
+ const struct net_device_ops *ops = dev->netdev_ops;
+ struct tc_taprio_qopt_offload *offload;
+ int err;
+
+ if (!FULL_OFFLOAD_IS_ENABLED(q->flags))
+ return 0;
+
+ if (!ops->ndo_setup_tc)
+ return -EOPNOTSUPP;
+
+ offload = taprio_offload_alloc(0);
+ if (!offload) {
+ NL_SET_ERR_MSG(extack,
+ "Not enough memory to disable offload mode");
+ return -ENOMEM;
+ }
+ offload->enable = 0;
+
+ err = ops->ndo_setup_tc(dev, TC_SETUP_QDISC_TAPRIO, offload);
+ if (err < 0) {
+ NL_SET_ERR_MSG(extack,
+ "Device failed to disable offload");
+ goto out;
+ }
+
+out:
+ taprio_offload_free(offload);
+
+ return err;
+}
+
+/* If full offload is enabled, the only possible clockid is the net device's
+ * PHC. For that reason, specifying a clockid through netlink is incorrect.
+ * For txtime-assist, it is implicitly assumed that the device's PHC is kept
+ * in sync with the specified clockid via a user space daemon such as phc2sys.
+ * For both software taprio and txtime-assist, the clockid is used for the
+ * hrtimer that advances the schedule and hence mandatory.
+ */
+static int taprio_parse_clockid(struct Qdisc *sch, struct nlattr **tb,
+ struct netlink_ext_ack *extack)
+{
+ struct taprio_sched *q = qdisc_priv(sch);
+ struct net_device *dev = qdisc_dev(sch);
+ int err = -EINVAL;
+
+ if (FULL_OFFLOAD_IS_ENABLED(q->flags)) {
+ const struct ethtool_ops *ops = dev->ethtool_ops;
+ struct ethtool_ts_info info = {
+ .cmd = ETHTOOL_GET_TS_INFO,
+ .phc_index = -1,
+ };
+
+ if (tb[TCA_TAPRIO_ATTR_SCHED_CLOCKID]) {
+ NL_SET_ERR_MSG(extack,
+ "The 'clockid' cannot be specified for full offload");
+ goto out;
+ }
+
+ if (ops && ops->get_ts_info)
+ err = ops->get_ts_info(dev, &info);
+
+ if (err || info.phc_index < 0) {
+ NL_SET_ERR_MSG(extack,
+ "Device does not have a PTP clock");
+ err = -ENOTSUPP;
+ goto out;
+ }
+ } else if (tb[TCA_TAPRIO_ATTR_SCHED_CLOCKID]) {
+ int clockid = nla_get_s32(tb[TCA_TAPRIO_ATTR_SCHED_CLOCKID]);
+
+ /* We only support static clockids and we don't allow
+ * for it to be modified after the first init.
+ */
+ if (clockid < 0 ||
+ (q->clockid != -1 && q->clockid != clockid)) {
+ NL_SET_ERR_MSG(extack,
+ "Changing the 'clockid' of a running schedule is not supported");
+ err = -ENOTSUPP;
+ goto out;
+ }
+
+ switch (clockid) {
+ case CLOCK_REALTIME:
+ q->tk_offset = TK_OFFS_REAL;
+ break;
+ case CLOCK_MONOTONIC:
+ q->tk_offset = TK_OFFS_MAX;
+ break;
+ case CLOCK_BOOTTIME:
+ q->tk_offset = TK_OFFS_BOOT;
+ break;
+ case CLOCK_TAI:
+ q->tk_offset = TK_OFFS_TAI;
+ break;
+ default:
+ NL_SET_ERR_MSG(extack, "Invalid 'clockid'");
+ err = -EINVAL;
+ goto out;
+ }
+
+ q->clockid = clockid;
+ } else {
+ NL_SET_ERR_MSG(extack, "Specifying a 'clockid' is mandatory");
+ goto out;
+ }
+out:
+ return err;
+}
+
static int taprio_change(struct Qdisc *sch, struct nlattr *opt,
struct netlink_ext_ack *extack)
{
@@ -1020,9 +1355,9 @@ static int taprio_change(struct Qdisc *sch, struct nlattr *opt,
struct net_device *dev = qdisc_dev(sch);
struct tc_mqprio_qopt *mqprio = NULL;
u32 taprio_flags = 0;
- int i, err, clockid;
unsigned long flags;
ktime_t start;
+ int i, err;
err = nla_parse_nested_deprecated(tb, TCA_TAPRIO_ATTR_MAX, opt,
taprio_policy, extack);
@@ -1038,7 +1373,7 @@ static int taprio_change(struct Qdisc *sch, struct nlattr *opt,
if (q->flags != 0 && q->flags != taprio_flags) {
NL_SET_ERR_MSG_MOD(extack, "Changing 'flags' of a running schedule is not supported");
return -EOPNOTSUPP;
- } else if (!FLAGS_VALID(taprio_flags)) {
+ } else if (!taprio_flags_valid(taprio_flags)) {
NL_SET_ERR_MSG_MOD(extack, "Specified 'flags' are not valid");
return -EINVAL;
}
@@ -1078,30 +1413,19 @@ static int taprio_change(struct Qdisc *sch, struct nlattr *opt,
goto free_sched;
}
- if (tb[TCA_TAPRIO_ATTR_SCHED_CLOCKID]) {
- clockid = nla_get_s32(tb[TCA_TAPRIO_ATTR_SCHED_CLOCKID]);
-
- /* We only support static clockids and we don't allow
- * for it to be modified after the first init.
- */
- if (clockid < 0 ||
- (q->clockid != -1 && q->clockid != clockid)) {
- NL_SET_ERR_MSG(extack, "Changing the 'clockid' of a running schedule is not supported");
- err = -ENOTSUPP;
- goto free_sched;
- }
-
- q->clockid = clockid;
- }
-
- if (q->clockid == -1 && !tb[TCA_TAPRIO_ATTR_SCHED_CLOCKID]) {
- NL_SET_ERR_MSG(extack, "Specifying a 'clockid' is mandatory");
- err = -EINVAL;
+ err = taprio_parse_clockid(sch, tb, extack);
+ if (err < 0)
goto free_sched;
- }
taprio_set_picos_per_byte(dev, q);
+ if (FULL_OFFLOAD_IS_ENABLED(taprio_flags))
+ err = taprio_enable_offload(dev, mqprio, q, new_admin, extack);
+ else
+ err = taprio_disable_offload(dev, q, extack);
+ if (err)
+ goto free_sched;
+
/* Protects against enqueue()/dequeue() */
spin_lock_bh(qdisc_lock(sch));
@@ -1116,6 +1440,7 @@ static int taprio_change(struct Qdisc *sch, struct nlattr *opt,
}
if (!TXTIME_ASSIST_IS_ENABLED(taprio_flags) &&
+ !FULL_OFFLOAD_IS_ENABLED(taprio_flags) &&
!hrtimer_active(&q->advance_timer)) {
hrtimer_init(&q->advance_timer, q->clockid, HRTIMER_MODE_ABS);
q->advance_timer.function = advance_sched;
@@ -1134,23 +1459,15 @@ static int taprio_change(struct Qdisc *sch, struct nlattr *opt,
mqprio->prio_tc_map[i]);
}
- switch (q->clockid) {
- case CLOCK_REALTIME:
- q->tk_offset = TK_OFFS_REAL;
- break;
- case CLOCK_MONOTONIC:
- q->tk_offset = TK_OFFS_MAX;
- break;
- case CLOCK_BOOTTIME:
- q->tk_offset = TK_OFFS_BOOT;
- break;
- case CLOCK_TAI:
- q->tk_offset = TK_OFFS_TAI;
- break;
- default:
- NL_SET_ERR_MSG(extack, "Invalid 'clockid'");
- err = -EINVAL;
- goto unlock;
+ if (FULL_OFFLOAD_IS_ENABLED(taprio_flags)) {
+ q->dequeue = taprio_dequeue_offload;
+ q->peek = taprio_peek_offload;
+ } else {
+ /* Be sure to always keep the function pointers
+ * in a consistent state.
+ */
+ q->dequeue = taprio_dequeue_soft;
+ q->peek = taprio_peek_soft;
}
err = taprio_get_start_time(sch, new_admin, &start);
@@ -1212,6 +1529,8 @@ static void taprio_destroy(struct Qdisc *sch)
hrtimer_cancel(&q->advance_timer);
+ taprio_disable_offload(dev, q, NULL);
+
if (q->qdiscs) {
for (i = 0; i < dev->num_tx_queues && q->qdiscs[i]; i++)
qdisc_put(q->qdiscs[i]);
@@ -1241,6 +1560,9 @@ static int taprio_init(struct Qdisc *sch, struct nlattr *opt,
hrtimer_init(&q->advance_timer, CLOCK_TAI, HRTIMER_MODE_ABS);
q->advance_timer.function = advance_sched;
+ q->dequeue = taprio_dequeue_soft;
+ q->peek = taprio_peek_soft;
+
q->root = sch;
/* We only support static clockids. Use an invalid value as default
@@ -1423,7 +1745,8 @@ static int taprio_dump(struct Qdisc *sch, struct sk_buff *skb)
if (nla_put(skb, TCA_TAPRIO_ATTR_PRIOMAP, sizeof(opt), &opt))
goto options_error;
- if (nla_put_s32(skb, TCA_TAPRIO_ATTR_SCHED_CLOCKID, q->clockid))
+ if (!FULL_OFFLOAD_IS_ENABLED(q->flags) &&
+ nla_put_s32(skb, TCA_TAPRIO_ATTR_SCHED_CLOCKID, q->clockid))
goto options_error;
if (q->flags && nla_put_u32(skb, TCA_TAPRIO_ATTR_FLAGS, q->flags))
--
2.17.1
^ permalink raw reply related
* [PATCH v2 net-next 2/7] net: dsa: Pass ndo_setup_tc slave callback to drivers
From: Vladimir Oltean @ 2019-09-14 1:17 UTC (permalink / raw)
To: f.fainelli, vivien.didelot, andrew, davem, vinicius.gomes,
vedang.patel, richardcochran
Cc: weifeng.voon, jiri, m-karicheri2, Jose.Abreu, ilias.apalodimas,
jhs, xiyou.wangcong, kurt.kanzenbach, joergen.andreasen, netdev,
Vladimir Oltean
In-Reply-To: <20190914011802.1602-1-olteanv@gmail.com>
DSA currently handles shared block filters (for the classifier-action
qdisc) in the core due to what I believe are simply pragmatic reasons -
hiding the complexity from drivers and offerring a simple API for port
mirroring.
Extend the dsa_slave_setup_tc function by passing all other qdisc
offloads to the driver layer, where the driver may choose what it
implements and how. DSA is simply a pass-through in this case.
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
Acked-by: Kurt Kanzenbach <kurt@linutronix.de>
---
Changes since v1:
- Added Kurt Kanzenbach's Acked-by.
Changes since RFC:
- Removed the unused declaration of struct tc_taprio_qopt_offload.
include/net/dsa.h | 2 ++
net/dsa/slave.c | 12 ++++++++----
2 files changed, 10 insertions(+), 4 deletions(-)
diff --git a/include/net/dsa.h b/include/net/dsa.h
index 96acb14ec1a8..541fb514e31d 100644
--- a/include/net/dsa.h
+++ b/include/net/dsa.h
@@ -515,6 +515,8 @@ struct dsa_switch_ops {
bool ingress);
void (*port_mirror_del)(struct dsa_switch *ds, int port,
struct dsa_mall_mirror_tc_entry *mirror);
+ int (*port_setup_tc)(struct dsa_switch *ds, int port,
+ enum tc_setup_type type, void *type_data);
/*
* Cross-chip operations
diff --git a/net/dsa/slave.c b/net/dsa/slave.c
index 9a88035517a6..75d58229a4bd 100644
--- a/net/dsa/slave.c
+++ b/net/dsa/slave.c
@@ -1035,12 +1035,16 @@ static int dsa_slave_setup_tc_block(struct net_device *dev,
static int dsa_slave_setup_tc(struct net_device *dev, enum tc_setup_type type,
void *type_data)
{
- switch (type) {
- case TC_SETUP_BLOCK:
+ struct dsa_port *dp = dsa_slave_to_port(dev);
+ struct dsa_switch *ds = dp->ds;
+
+ if (type == TC_SETUP_BLOCK)
return dsa_slave_setup_tc_block(dev, type_data);
- default:
+
+ if (!ds->ops->port_setup_tc)
return -EOPNOTSUPP;
- }
+
+ return ds->ops->port_setup_tc(ds, dp->index, type, type_data);
}
static void dsa_slave_get_stats64(struct net_device *dev,
--
2.17.1
^ permalink raw reply related
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox