| Message ID | 1456842290-7844-7-git-send-email-amir@vadai.me |
|---|---|
| State | Rejected, archived |
| Delegated to: | David Miller |
| Headers | show |
Tue, Mar 01, 2016 at 03:24:48PM CET, amir@vadai.me wrote: >Extend ndo_setup_tc() to support ingress tc offloading. Will be used by >later patches to offload tc flower filter. > >Feature is off by default and could be enabled by issuing: > # ethtool -K eth0 hw-tc-offload on > >Offloads flow table is dynamically created when first filter is >added. >Rules are saved in a hash table that is maintained by the consumer (for >example - the flower offload in the next patch). >When last filter is removed and no filters exist in the hash table, the >offload flow table is destroyed. <snip> >@@ -1880,6 +1883,17 @@ static int mlx5e_setup_tc(struct net_device *netdev, u8 tc) > static int mlx5e_ndo_setup_tc(struct net_device *dev, u32 handle, > __be16 proto, struct tc_to_netdev *tc) > { >+ struct mlx5e_priv *priv = netdev_priv(dev); >+ >+ if (TC_H_MAJ(handle) != TC_H_MAJ(TC_H_INGRESS)) >+ goto mqprio; >+ >+ switch (tc->type) { >+ default: >+ return -EINVAL; -EOPNOTSUPP would be better here perhaps? >+ } >+ >+mqprio: > if (handle != TC_H_ROOT || tc->type != TC_SETUP_MQPRIO) > return -EINVAL; > >@@ -1963,6 +1977,13 @@ static int mlx5e_set_features(struct net_device *netdev, > mlx5e_disable_vlan_filter(priv); > } > >+ if ((changes & NETIF_F_HW_TC) && !(features & NETIF_F_HW_TC) && >+ mlx5e_tc_num_filters(priv)) { >+ netdev_err(netdev, >+ "Active offloaded tc filters, can't turn hw_tc_offload off\n"); >+ return -EINVAL; This should not fail I believe. Just disable it in hw. I would even toss away the rules if necessary.
On Tue, Mar 1, 2016 at 4:24 PM, Amir Vadai <amir@vadai.me> wrote: > +#define FT_CAP(f) MLX5_CAP_FLOWTABLE(mdev, flow_table_properties_nic_receive.f) > + if (FT_CAP(flow_modify_en) && > + FT_CAP(modify_root) && > + FT_CAP(identified_miss_table_mode) && > + FT_CAP(flow_table_modify)) > + priv->netdev->hw_features |= NETIF_F_HW_TC; > + > netdev->features |= NETIF_F_HIGHDMA; > > netdev->priv_flags |= IFF_UNICAST_FLT; > > + mlx5e_tc_init(priv); This is not the place for this, We usually do internal data structure initialization after we create all HW resources in mlx5e_create_netdev Please see mlx5e_vxlan_init as example, and you already call mlx5e_tc_cleanup inside mlx5e_destroy_netdev, please move the mlx5e_tc_init to mlx5e_create_netdev after HW resources creation, > @@ -2558,6 +2588,7 @@ static void mlx5e_destroy_netdev(struct mlx5_core_dev *mdev, void *vpriv) > mlx5_core_dealloc_transport_domain(priv->mdev, priv->tdn); > mlx5_core_dealloc_pd(priv->mdev, priv->pdn); > mlx5_unmap_free_uar(priv->mdev, &priv->cq_uar); > + mlx5e_tc_cleanup(priv); I would suggest to move mlx5e_tc_init to be right after mlx5e_vxlan_init and mlx5e_tc_cleanup before mlx5e_vxlan_cleanup. > +struct mlx5_flow_rule *mlx5e_tc_add_flow(struct mlx5e_priv *priv, > + u32 *match_c, u32 *match_v, > + u32 action, u32 flow_tag) > +{ > + struct mlx5_flow_destination dest = { > + .type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE, > + {.ft = priv->fts.vlan.t}, > + }; > + struct mlx5_flow_rule *rule; > + bool table_created = false; > + > + if (IS_ERR_OR_NULL(priv->fts.tc.t)) { > + priv->fts.tc.t = > + mlx5_create_auto_grouped_flow_table(priv->fts.ns, 0, > + MLX5E_TC_FLOW_TABLE_NUM_ENTRIES, > + MLX5E_TC_FLOW_TABLE_NUM_GROUPS); > + if (IS_ERR(priv->fts.tc.t)) { > + netdev_err(priv->netdev, > + "Failed to create tc offload table\n"); > + return ERR_CAST(priv->fts.tc.t); Here priv->fts.tc.t will be invalid pointer and in your code you treat it as NULL in case of failure. > + } > + > + table_created = true; > + } > + > + rule = mlx5_add_flow_rule(priv->fts.tc.t, MLX5_MATCH_OUTER_HEADERS, > + match_c, match_v, > + action, flow_tag, > + action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST ? &dest : NULL); > + > + if (IS_ERR(rule) && table_created) { > + mlx5_destroy_flow_table(priv->fts.tc.t); > + priv->fts.tc.t = NULL; > + } > + > + return rule; > +} > + > +void mlx5e_tc_cleanup(struct mlx5e_priv *priv) > +{ > + struct mlx5e_tc_flow_table *tc = &priv->fts.tc; > + > + rhashtable_free_and_destroy(&tc->ht, _mlx5e_tc_del_flow, priv); > + > + if (priv->fts.tc.t) { priv->fts.tc.t will be invalid pointer and this test will pass in case mlx5_create_auto_grouped_flow_table had failed > + mlx5_destroy_flow_table(priv->fts.tc.t); > + priv->fts.tc.t = NULL; > + } > +}
On Tue, Mar 01, 2016 at 03:52:08PM +0100, Jiri Pirko wrote: > Tue, Mar 01, 2016 at 03:24:48PM CET, amir@vadai.me wrote: > >Extend ndo_setup_tc() to support ingress tc offloading. Will be used by > >later patches to offload tc flower filter. > > > >Feature is off by default and could be enabled by issuing: > > # ethtool -K eth0 hw-tc-offload on > > > >Offloads flow table is dynamically created when first filter is > >added. > >Rules are saved in a hash table that is maintained by the consumer (for > >example - the flower offload in the next patch). > >When last filter is removed and no filters exist in the hash table, the > >offload flow table is destroyed. > > <snip> > > >@@ -1880,6 +1883,17 @@ static int mlx5e_setup_tc(struct net_device *netdev, u8 tc) > > static int mlx5e_ndo_setup_tc(struct net_device *dev, u32 handle, > > __be16 proto, struct tc_to_netdev *tc) > > { > >+ struct mlx5e_priv *priv = netdev_priv(dev); > >+ > >+ if (TC_H_MAJ(handle) != TC_H_MAJ(TC_H_INGRESS)) > >+ goto mqprio; > >+ > >+ switch (tc->type) { > >+ default: > >+ return -EINVAL; > > -EOPNOTSUPP would be better here perhaps? > > > >+ } > >+ > >+mqprio: > > if (handle != TC_H_ROOT || tc->type != TC_SETUP_MQPRIO) > > return -EINVAL; > > > >@@ -1963,6 +1977,13 @@ static int mlx5e_set_features(struct net_device *netdev, > > mlx5e_disable_vlan_filter(priv); > > } > > > >+ if ((changes & NETIF_F_HW_TC) && !(features & NETIF_F_HW_TC) && > >+ mlx5e_tc_num_filters(priv)) { > >+ netdev_err(netdev, > >+ "Active offloaded tc filters, can't turn hw_tc_offload off\n"); > >+ return -EINVAL; > > This should not fail I believe. Just disable it in hw. I would even toss > away the rules if necessary. It depends on the answer regarding your comment on the previous patch. If we have the rule in both SW and HW, and remove it from the HW it is ok (although, currently I don't understand why would anyone want in both places). If the rule is processed by HW only - turning off this flag, will disable the offloaded rules - it might be misleading, so I prefered not to allow it and print a message. >
On Tue, Mar 01, 2016 at 05:59:59PM +0200, Saeed Mahameed wrote: > On Tue, Mar 1, 2016 at 4:24 PM, Amir Vadai <amir@vadai.me> wrote: > > +#define FT_CAP(f) MLX5_CAP_FLOWTABLE(mdev, flow_table_properties_nic_receive.f) > > + if (FT_CAP(flow_modify_en) && > > + FT_CAP(modify_root) && > > + FT_CAP(identified_miss_table_mode) && > > + FT_CAP(flow_table_modify)) > > + priv->netdev->hw_features |= NETIF_F_HW_TC; > > + > > netdev->features |= NETIF_F_HIGHDMA; > > > > netdev->priv_flags |= IFF_UNICAST_FLT; > > > > + mlx5e_tc_init(priv); > > This is not the place for this, We usually do internal data structure > initialization after we create all HW resources in > mlx5e_create_netdev > Please see mlx5e_vxlan_init as example, and you already call > mlx5e_tc_cleanup inside mlx5e_destroy_netdev, please move the > mlx5e_tc_init > to mlx5e_create_netdev after HW resources creation, ack > > > > @@ -2558,6 +2588,7 @@ static void mlx5e_destroy_netdev(struct mlx5_core_dev *mdev, void *vpriv) > > mlx5_core_dealloc_transport_domain(priv->mdev, priv->tdn); > > mlx5_core_dealloc_pd(priv->mdev, priv->pdn); > > mlx5_unmap_free_uar(priv->mdev, &priv->cq_uar); > > + mlx5e_tc_cleanup(priv); > > I would suggest to move mlx5e_tc_init to be right after > mlx5e_vxlan_init and mlx5e_tc_cleanup before mlx5e_vxlan_cleanup. ok > > > +struct mlx5_flow_rule *mlx5e_tc_add_flow(struct mlx5e_priv *priv, > > + u32 *match_c, u32 *match_v, > > + u32 action, u32 flow_tag) > > +{ > > + struct mlx5_flow_destination dest = { > > + .type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE, > > + {.ft = priv->fts.vlan.t}, > > + }; > > + struct mlx5_flow_rule *rule; > > + bool table_created = false; > > + > > + if (IS_ERR_OR_NULL(priv->fts.tc.t)) { > > + priv->fts.tc.t = > > + mlx5_create_auto_grouped_flow_table(priv->fts.ns, 0, > > + MLX5E_TC_FLOW_TABLE_NUM_ENTRIES, > > + MLX5E_TC_FLOW_TABLE_NUM_GROUPS); > > + if (IS_ERR(priv->fts.tc.t)) { > > + netdev_err(priv->netdev, > > + "Failed to create tc offload table\n"); > > + return ERR_CAST(priv->fts.tc.t); > > Here priv->fts.tc.t will be invalid pointer and in your code you treat > it as NULL in case of failure. > > > + } > > + > > + table_created = true; > > + } > > + > > + rule = mlx5_add_flow_rule(priv->fts.tc.t, MLX5_MATCH_OUTER_HEADERS, > > + match_c, match_v, > > + action, flow_tag, > > + action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST ? &dest : NULL); > > + > > + if (IS_ERR(rule) && table_created) { > > + mlx5_destroy_flow_table(priv->fts.tc.t); > > + priv->fts.tc.t = NULL; > > + } > > + > > + return rule; > > +} > > + > > > +void mlx5e_tc_cleanup(struct mlx5e_priv *priv) > > +{ > > + struct mlx5e_tc_flow_table *tc = &priv->fts.tc; > > + > > + rhashtable_free_and_destroy(&tc->ht, _mlx5e_tc_del_flow, priv); > > + > > + if (priv->fts.tc.t) { > > priv->fts.tc.t will be invalid pointer and this test will pass in case > mlx5_create_auto_grouped_flow_table had failed Yeh - should have used !IS_ERR_OR_NULL() or set it to NULL on failure above. Thanks, Amir > > > + mlx5_destroy_flow_table(priv->fts.tc.t); > > + priv->fts.tc.t = NULL; > > + } > > +}
On 16-03-01 09:00 AM, Amir Vadai wrote: > On Tue, Mar 01, 2016 at 03:52:08PM +0100, Jiri Pirko wrote: >> Tue, Mar 01, 2016 at 03:24:48PM CET, amir@vadai.me wrote: >>> Extend ndo_setup_tc() to support ingress tc offloading. Will be used by >>> later patches to offload tc flower filter. >>> >>> Feature is off by default and could be enabled by issuing: >>> # ethtool -K eth0 hw-tc-offload on >>> >>> Offloads flow table is dynamically created when first filter is >>> added. >>> Rules are saved in a hash table that is maintained by the consumer (for >>> example - the flower offload in the next patch). >>> When last filter is removed and no filters exist in the hash table, the >>> offload flow table is destroyed. >> >> <snip> >> >>> @@ -1880,6 +1883,17 @@ static int mlx5e_setup_tc(struct net_device *netdev, u8 tc) >>> static int mlx5e_ndo_setup_tc(struct net_device *dev, u32 handle, >>> __be16 proto, struct tc_to_netdev *tc) >>> { >>> + struct mlx5e_priv *priv = netdev_priv(dev); >>> + >>> + if (TC_H_MAJ(handle) != TC_H_MAJ(TC_H_INGRESS)) >>> + goto mqprio; >>> + >>> + switch (tc->type) { >>> + default: >>> + return -EINVAL; >> >> -EOPNOTSUPP would be better here perhaps? >> >> >>> + } >>> + >>> +mqprio: >>> if (handle != TC_H_ROOT || tc->type != TC_SETUP_MQPRIO) >>> return -EINVAL; >>> >>> @@ -1963,6 +1977,13 @@ static int mlx5e_set_features(struct net_device *netdev, >>> mlx5e_disable_vlan_filter(priv); >>> } >>> >>> + if ((changes & NETIF_F_HW_TC) && !(features & NETIF_F_HW_TC) && >>> + mlx5e_tc_num_filters(priv)) { >>> + netdev_err(netdev, >>> + "Active offloaded tc filters, can't turn hw_tc_offload off\n"); >>> + return -EINVAL; >> >> This should not fail I believe. Just disable it in hw. I would even toss >> away the rules if necessary. > It depends on the answer regarding your comment on the previous patch. > If we have the rule in both SW and HW, and remove it from the HW it is > ok (although, currently I don't understand why would anyone want in both > places). > If the rule is processed by HW only - turning off this flag, will > disable the offloaded rules - it might be misleading, so I prefered not > to allow it and print a message. When we get the HW only mode we will need to also flush the hardware representation in software as well as the hardware state.
On Tue, Mar 01, 2016 at 09:13:25AM -0800, John Fastabend wrote: > On 16-03-01 09:00 AM, Amir Vadai wrote: > > On Tue, Mar 01, 2016 at 03:52:08PM +0100, Jiri Pirko wrote: > >> Tue, Mar 01, 2016 at 03:24:48PM CET, amir@vadai.me wrote: > >>> Extend ndo_setup_tc() to support ingress tc offloading. Will be used by > >>> later patches to offload tc flower filter. > >>> > >>> Feature is off by default and could be enabled by issuing: > >>> # ethtool -K eth0 hw-tc-offload on > >>> > >>> Offloads flow table is dynamically created when first filter is > >>> added. > >>> Rules are saved in a hash table that is maintained by the consumer (for > >>> example - the flower offload in the next patch). > >>> When last filter is removed and no filters exist in the hash table, the > >>> offload flow table is destroyed. > >> > >> <snip> > >> > >>> @@ -1880,6 +1883,17 @@ static int mlx5e_setup_tc(struct net_device *netdev, u8 tc) > >>> static int mlx5e_ndo_setup_tc(struct net_device *dev, u32 handle, > >>> __be16 proto, struct tc_to_netdev *tc) > >>> { > >>> + struct mlx5e_priv *priv = netdev_priv(dev); > >>> + > >>> + if (TC_H_MAJ(handle) != TC_H_MAJ(TC_H_INGRESS)) > >>> + goto mqprio; > >>> + > >>> + switch (tc->type) { > >>> + default: > >>> + return -EINVAL; > >> > >> -EOPNOTSUPP would be better here perhaps? > >> > >> > >>> + } > >>> + > >>> +mqprio: > >>> if (handle != TC_H_ROOT || tc->type != TC_SETUP_MQPRIO) > >>> return -EINVAL; > >>> > >>> @@ -1963,6 +1977,13 @@ static int mlx5e_set_features(struct net_device *netdev, > >>> mlx5e_disable_vlan_filter(priv); > >>> } > >>> > >>> + if ((changes & NETIF_F_HW_TC) && !(features & NETIF_F_HW_TC) && > >>> + mlx5e_tc_num_filters(priv)) { > >>> + netdev_err(netdev, > >>> + "Active offloaded tc filters, can't turn hw_tc_offload off\n"); > >>> + return -EINVAL; > >> > >> This should not fail I believe. Just disable it in hw. I would even toss > >> away the rules if necessary. > > It depends on the answer regarding your comment on the previous patch. > > If we have the rule in both SW and HW, and remove it from the HW it is > > ok (although, currently I don't understand why would anyone want in both > > places). > > If the rule is processed by HW only - turning off this flag, will > > disable the offloaded rules - it might be misleading, so I prefered not > > to allow it and print a message. > > When we get the HW only mode we will need to also flush the hardware > representation in software as well as the hardware state. Yep, I do think that just failing the operation is the best appraoch. Will make the design simpler, and from the user point of view, less surprises. Jiri?
Wed, Mar 02, 2016 at 04:53:37PM CET, amir@vadai.me wrote: >On Tue, Mar 01, 2016 at 09:13:25AM -0800, John Fastabend wrote: >> On 16-03-01 09:00 AM, Amir Vadai wrote: >> > On Tue, Mar 01, 2016 at 03:52:08PM +0100, Jiri Pirko wrote: >> >> Tue, Mar 01, 2016 at 03:24:48PM CET, amir@vadai.me wrote: >> >>> Extend ndo_setup_tc() to support ingress tc offloading. Will be used by >> >>> later patches to offload tc flower filter. >> >>> >> >>> Feature is off by default and could be enabled by issuing: >> >>> # ethtool -K eth0 hw-tc-offload on >> >>> >> >>> Offloads flow table is dynamically created when first filter is >> >>> added. >> >>> Rules are saved in a hash table that is maintained by the consumer (for >> >>> example - the flower offload in the next patch). >> >>> When last filter is removed and no filters exist in the hash table, the >> >>> offload flow table is destroyed. >> >> >> >> <snip> >> >> >> >>> @@ -1880,6 +1883,17 @@ static int mlx5e_setup_tc(struct net_device *netdev, u8 tc) >> >>> static int mlx5e_ndo_setup_tc(struct net_device *dev, u32 handle, >> >>> __be16 proto, struct tc_to_netdev *tc) >> >>> { >> >>> + struct mlx5e_priv *priv = netdev_priv(dev); >> >>> + >> >>> + if (TC_H_MAJ(handle) != TC_H_MAJ(TC_H_INGRESS)) >> >>> + goto mqprio; >> >>> + >> >>> + switch (tc->type) { >> >>> + default: >> >>> + return -EINVAL; >> >> >> >> -EOPNOTSUPP would be better here perhaps? >> >> >> >> >> >>> + } >> >>> + >> >>> +mqprio: >> >>> if (handle != TC_H_ROOT || tc->type != TC_SETUP_MQPRIO) >> >>> return -EINVAL; >> >>> >> >>> @@ -1963,6 +1977,13 @@ static int mlx5e_set_features(struct net_device *netdev, >> >>> mlx5e_disable_vlan_filter(priv); >> >>> } >> >>> >> >>> + if ((changes & NETIF_F_HW_TC) && !(features & NETIF_F_HW_TC) && >> >>> + mlx5e_tc_num_filters(priv)) { >> >>> + netdev_err(netdev, >> >>> + "Active offloaded tc filters, can't turn hw_tc_offload off\n"); >> >>> + return -EINVAL; >> >> >> >> This should not fail I believe. Just disable it in hw. I would even toss >> >> away the rules if necessary. >> > It depends on the answer regarding your comment on the previous patch. >> > If we have the rule in both SW and HW, and remove it from the HW it is >> > ok (although, currently I don't understand why would anyone want in both >> > places). >> > If the rule is processed by HW only - turning off this flag, will >> > disable the offloaded rules - it might be misleading, so I prefered not >> > to allow it and print a message. >> >> When we get the HW only mode we will need to also flush the hardware >> representation in software as well as the hardware state. > >Yep, I do think that just failing the operation is the best appraoch. >Will make the design simpler, and from the user point of view, less >surprises. > >Jiri? I don't feel it is ok, but at the same time, it is probably the best solution for now. Other solutions would be too complicated.
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/Makefile index 11b592d..4fc45ee 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/Makefile +++ b/drivers/net/ethernet/mellanox/mlx5/core/Makefile @@ -6,6 +6,6 @@ mlx5_core-y := main.o cmd.o debugfs.o fw.o eq.o uar.o pagealloc.o \ mlx5_core-$(CONFIG_MLX5_CORE_EN) += wq.o eswitch.o \ en_main.o en_fs.o en_ethtool.o en_tx.o en_rx.o \ - en_txrx.o en_clock.o vxlan.o + en_txrx.o en_clock.o vxlan.o en_tc.o mlx5_core-$(CONFIG_MLX5_CORE_EN_DCB) += en_dcbnl.o diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index 1dca3dc..6571a25 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -43,6 +43,7 @@ #include <linux/mlx5/port.h> #include <linux/mlx5/vport.h> #include <linux/mlx5/transobj.h> +#include <linux/rhashtable.h> #include "wq.h" #include "mlx5_core.h" @@ -524,8 +525,16 @@ struct mlx5e_flow_table { struct mlx5_flow_group **g; }; +struct mlx5e_tc_flow_table { + struct mlx5_flow_table *t; + + struct rhashtable_params ht_params; + struct rhashtable ht; +}; + struct mlx5e_flow_tables { struct mlx5_flow_namespace *ns; + struct mlx5e_tc_flow_table tc; struct mlx5e_flow_table vlan; struct mlx5e_flow_table main; }; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 0d45f35..cb02b4c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -30,9 +30,12 @@ * SOFTWARE. */ +#include <net/tc_act/tc_gact.h> +#include <net/pkt_cls.h> #include <linux/mlx5/fs.h> #include <net/vxlan.h> #include "en.h" +#include "en_tc.h" #include "eswitch.h" #include "vxlan.h" @@ -1880,6 +1883,17 @@ static int mlx5e_setup_tc(struct net_device *netdev, u8 tc) static int mlx5e_ndo_setup_tc(struct net_device *dev, u32 handle, __be16 proto, struct tc_to_netdev *tc) { + struct mlx5e_priv *priv = netdev_priv(dev); + + if (TC_H_MAJ(handle) != TC_H_MAJ(TC_H_INGRESS)) + goto mqprio; + + switch (tc->type) { + default: + return -EINVAL; + } + +mqprio: if (handle != TC_H_ROOT || tc->type != TC_SETUP_MQPRIO) return -EINVAL; @@ -1963,6 +1977,13 @@ static int mlx5e_set_features(struct net_device *netdev, mlx5e_disable_vlan_filter(priv); } + if ((changes & NETIF_F_HW_TC) && !(features & NETIF_F_HW_TC) && + mlx5e_tc_num_filters(priv)) { + netdev_err(netdev, + "Active offloaded tc filters, can't turn hw_tc_offload off\n"); + return -EINVAL; + } + return err; } @@ -2361,10 +2382,18 @@ static void mlx5e_build_netdev(struct net_device *netdev) if (!priv->params.lro_en) netdev->features &= ~NETIF_F_LRO; +#define FT_CAP(f) MLX5_CAP_FLOWTABLE(mdev, flow_table_properties_nic_receive.f) + if (FT_CAP(flow_modify_en) && + FT_CAP(modify_root) && + FT_CAP(identified_miss_table_mode) && + FT_CAP(flow_table_modify)) + priv->netdev->hw_features |= NETIF_F_HW_TC; + netdev->features |= NETIF_F_HIGHDMA; netdev->priv_flags |= IFF_UNICAST_FLT; + mlx5e_tc_init(priv); mlx5e_set_netdev_dev_addr(netdev); } @@ -2531,6 +2560,7 @@ err_unmap_free_uar: mlx5_unmap_free_uar(mdev, &priv->cq_uar); err_free_netdev: + mlx5e_tc_cleanup(priv); free_netdev(netdev); return NULL; @@ -2558,6 +2588,7 @@ static void mlx5e_destroy_netdev(struct mlx5_core_dev *mdev, void *vpriv) mlx5_core_dealloc_transport_domain(priv->mdev, priv->tdn); mlx5_core_dealloc_pd(priv->mdev, priv->pdn); mlx5_unmap_free_uar(priv->mdev, &priv->cq_uar); + mlx5e_tc_cleanup(priv); free_netdev(netdev); } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c new file mode 100644 index 0000000..7d1c0a3 --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c @@ -0,0 +1,131 @@ +/* + * Copyright (c) 2016, Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include <linux/mlx5/fs.h> +#include <linux/mlx5/device.h> +#include <linux/rhashtable.h> +#include "en.h" +#include "en_tc.h" + +struct mlx5e_tc_flow { + struct rhash_head node; + u64 cookie; + struct mlx5_flow_rule *rule; +}; + +#define MLX5E_TC_FLOW_TABLE_NUM_ENTRIES 1024 +#define MLX5E_TC_FLOW_TABLE_NUM_GROUPS 4 + +struct mlx5_flow_rule *mlx5e_tc_add_flow(struct mlx5e_priv *priv, + u32 *match_c, u32 *match_v, + u32 action, u32 flow_tag) +{ + struct mlx5_flow_destination dest = { + .type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE, + {.ft = priv->fts.vlan.t}, + }; + struct mlx5_flow_rule *rule; + bool table_created = false; + + if (IS_ERR_OR_NULL(priv->fts.tc.t)) { + priv->fts.tc.t = + mlx5_create_auto_grouped_flow_table(priv->fts.ns, 0, + MLX5E_TC_FLOW_TABLE_NUM_ENTRIES, + MLX5E_TC_FLOW_TABLE_NUM_GROUPS); + if (IS_ERR(priv->fts.tc.t)) { + netdev_err(priv->netdev, + "Failed to create tc offload table\n"); + return ERR_CAST(priv->fts.tc.t); + } + + table_created = true; + } + + rule = mlx5_add_flow_rule(priv->fts.tc.t, MLX5_MATCH_OUTER_HEADERS, + match_c, match_v, + action, flow_tag, + action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST ? &dest : NULL); + + if (IS_ERR(rule) && table_created) { + mlx5_destroy_flow_table(priv->fts.tc.t); + priv->fts.tc.t = NULL; + } + + return rule; +} + +static void mlx5e_tc_del_flow(struct mlx5e_priv *priv, + struct mlx5_flow_rule *rule) +{ + mlx5_del_flow_rule(rule); + + if (!mlx5e_tc_num_filters(priv)) { + mlx5_destroy_flow_table(priv->fts.tc.t); + priv->fts.tc.t = NULL; + } +} + +static const struct rhashtable_params mlx5e_tc_flow_ht_params = { + .head_offset = offsetof(struct mlx5e_tc_flow, node), + .key_offset = offsetof(struct mlx5e_tc_flow, cookie), + .key_len = sizeof(((struct mlx5e_tc_flow *)0)->cookie), + .automatic_shrinking = true, +}; + +void mlx5e_tc_init(struct mlx5e_priv *priv) +{ + struct mlx5e_tc_flow_table *tc = &priv->fts.tc; + + tc->ht_params = mlx5e_tc_flow_ht_params; + rhashtable_init(&tc->ht, &tc->ht_params); +} + +static void _mlx5e_tc_del_flow(void *ptr, void *arg) +{ + struct mlx5e_tc_flow *flow = ptr; + struct mlx5e_priv *priv = arg; + + mlx5e_tc_del_flow(priv, flow->rule); + kfree(flow); +} + +void mlx5e_tc_cleanup(struct mlx5e_priv *priv) +{ + struct mlx5e_tc_flow_table *tc = &priv->fts.tc; + + rhashtable_free_and_destroy(&tc->ht, _mlx5e_tc_del_flow, priv); + + if (priv->fts.tc.t) { + mlx5_destroy_flow_table(priv->fts.tc.t); + priv->fts.tc.t = NULL; + } +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h new file mode 100644 index 0000000..8a0dc0d --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h @@ -0,0 +1,44 @@ +/* + * Copyright (c) 2016, Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#ifndef __MLX5_EN_TC_H__ +#define __MLX5_EN_TC_H__ + +void mlx5e_tc_init(struct mlx5e_priv *priv); +void mlx5e_tc_cleanup(struct mlx5e_priv *priv); + +static inline int mlx5e_tc_num_filters(struct mlx5e_priv *priv) +{ + return atomic_read(&priv->fts.tc.ht.nelems); +} + +#endif /* __MLX5_EN_TC_H__ */