Introduce a flow gate control action and apply IEEE
Subsystems (13)
cover Introduce a flow gate control action and apply IEEE
This patch set is trying to intruduce a way to add tc flower offload for the enetc IEEE 802.1Qci (PSFP) function. There are four main feature parts in PSFP to implement the flow policing and filtering for ingress flow with IEEE 802.1Qci features. They are stream identify(this is defined in the P802.1cb exactly but needed by 802.1Qci), stream filtering, stream gate and flow metering.
The stream gate function is the important part in the features. But there is no compare actions in the qdisc filter part. The second patch introduce a ingress frame gate control flow action. tc create a gate action would provide a gate list to control class open/close state. when the gate open state, the flow could pass but not when gate state is close. The driver would repeat the gate list periodic. User also could assign a time point to start the gate list by the basetime parameter. if the basetime has passed current time, start time would calculate by the cycletime of the gate list. And it is introduce a software simulator gate control method.
The first patch is fix a flow offload can't provide dropped frame count number issue. This would be used for getting the hardware offload dropped frame.
The third patch is to adding the gate flow offloading.
The fourth patch is to add tc offload command in enetc. This is to control the on/off for the tc flower offloading.
Now the enetc driver would got the gate control list and filter mac address etc. So enetc would collected the parameters and create the stream identify entry and stream gate control entry. Then driver would create a stream filter entry by these inputs. Driver would maintain the flow chain list. The fifth patch implement the stream gate and stream filter and stream identify functions in driver by the tc flower actions and tc filter parameters.
The sixth patch extend the police action with max frame size parameter. This patch prepare for the PSFP per stream filtering by the frame size policing.
The seventh patch add the max frame size policing into the stream filtering function.
The eighth patch extend the police action with action index to the driver. So driver could know which hardware entry to police the rate and burst size.
The ninth patch add flow metering function in driver for the IEEE802.1Qci with the police action 'index'/'burst'/'rate_bytes_ps'.
The iproute2 test patch need to upload to:
git://git.kernel.org/pub/scm/network/iproute2/iproute2-next.git
There are some works still need to improve and I'd like your feedback:
- The gate action software simulator need to add admin/oper state machine. This state machine would keep the previous operation gate list before the new admin gate list start time arrived. Does it required for admin/oper in software implement?
- More PSFP flow metering parameters. flow metering is an optional function for a specific filter chain. Add more parameters this part would make the flow metering more completely. Flow metering privde a two rate (CIR, EIR) and two buckets (CBS, EBS) and mark flow three color(green, yellow, red). Current tc flower offload only provide "burst" and "rate_bytes_ps" in police action for driver. This patch set using these two parameters to set one bucket and one rate. Each flow metering entry own two sets buckets and two rate police. So the second rate/burst keep disable.
Po Liu (9):
net: qos offload add flow status with dropped count
net: qos: introduce a gate control flow action
net: schedule: add action gate offloading
net: enetc: add hw tc hw offload features for PSPF capability
net: enetc: add tc flower psfp offload driver
net: qos: add tc police offloading action with max frame size limit
net: enetc: add support max frame size for tc flower offload
net: qos: police action add index for tc flower offloading
net: enetc add tc flower offload flow metering policing action
drivers/net/ethernet/broadcom/bnxt/bnxt_tc.c | 2 +- .../ethernet/chelsio/cxgb4/cxgb4_tc_flower.c | 2 +- .../chelsio/cxgb4/cxgb4_tc_matchall.c | 2 +- drivers/net/ethernet/freescale/enetc/enetc.c | 34 +- drivers/net/ethernet/freescale/enetc/enetc.h | 86 ++ .../net/ethernet/freescale/enetc/enetc_hw.h | 183 +++ .../net/ethernet/freescale/enetc/enetc_pf.c | 6 + .../net/ethernet/freescale/enetc/enetc_qos.c | 1228 ++++++++++++++++- .../net/ethernet/mellanox/mlx5/core/en_tc.c | 4 +- .../ethernet/mellanox/mlxsw/spectrum_flower.c | 2 +- drivers/net/ethernet/mscc/ocelot_flower.c | 2 +- .../ethernet/netronome/nfp/flower/offload.c | 2 +- .../ethernet/netronome/nfp/flower/qos_conf.c | 2 +- include/net/act_api.h | 11 +- include/net/flow_offload.h | 17 +- include/net/pkt_cls.h | 5 +- include/net/tc_act/tc_gate.h | 169 +++ include/net/tc_act/tc_police.h | 10 + include/uapi/linux/pkt_cls.h | 1 + include/uapi/linux/tc_act/tc_gate.h | 47 + net/sched/Kconfig | 15 + net/sched/Makefile | 1 + net/sched/act_api.c | 12 +- net/sched/act_ct.c | 6 +- net/sched/act_gact.c | 7 +- net/sched/act_gate.c | 645 +++++++++ net/sched/act_mirred.c | 6 +- net/sched/act_police.c | 6 +- net/sched/act_vlan.c | 6 +- net/sched/cls_api.c | 35 + net/sched/cls_flower.c | 3 +- net/sched/cls_matchall.c | 3 +- 32 files changed, 2515 insertions(+), 45 deletions(-) create mode 100644 include/net/tc_act/tc_gate.h create mode 100644 include/uapi/linux/tc_act/tc_gate.h create mode 100644 net/sched/act_gate.c
Signature · 1 line
2.17.1
RFC 1/9 net: qos offload add flow status with dropped count
Add the hardware tc flower offloading with dropped frame counter for status update. action ops->stats_update only loaded by the tcf_exts_stats_update() and tcf_exts_stats_update() only loaded by matchall and tc flower hardware filter. But the stats_update only set the dropped count as default false in the ops->stats_update. This patch add the dropped counter to action stats update. Its dropped counter update by the hardware offloading driver.
This is changed by replacing the drop flag with dropped frame counter.
Driver side should update how many "packets" it filtered and how many "dropped" in those "packets".
--- drivers/net/ethernet/broadcom/bnxt/bnxt_tc.c | 2 +- drivers/net/ethernet/chelsio/cxgb4/cxgb4_tc_flower.c | 2 +- .../net/ethernet/chelsio/cxgb4/cxgb4_tc_matchall.c | 2 +- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 4 ++-- .../net/ethernet/mellanox/mlxsw/spectrum_flower.c | 2 +- drivers/net/ethernet/mscc/ocelot_flower.c | 2 +- drivers/net/ethernet/netronome/nfp/flower/offload.c | 2 +- drivers/net/ethernet/netronome/nfp/flower/qos_conf.c | 2 +- include/net/act_api.h | 11 ++++++----- include/net/flow_offload.h | 5 ++++- include/net/pkt_cls.h | 5 +++-- net/sched/act_api.c | 12 ++++++------ net/sched/act_ct.c | 6 +++--- net/sched/act_gact.c | 7 ++++--- net/sched/act_mirred.c | 6 +++--- net/sched/act_police.c | 6 +++--- net/sched/act_vlan.c | 6 +++--- net/sched/cls_flower.c | 3 ++- net/sched/cls_matchall.c | 3 ++- 19 files changed, 48 insertions(+), 40 deletions(-)
drivers/net/ethernet/broadcom/bnxt/bnxt_tc.c +1 -1
@@ -1634,7 +1634,7 @@ static int bnxt_tc_get_flow_stats(struct bnxt *bp,spin_unlock (&flow->stats_lock);flow_stats_update (&tc_flow_cmd->stats, stats.bytes, stats.packets, - lastused); + lastused,0 );return 0 ; }
drivers/net/ethernet/chelsio/cxgb4/cxgb4_tc_flower.c +1 -1
@@ -836,7 +836,7 @@ int cxgb4_tc_flower_stats(struct net_device *dev, ofld_stats->last_used = jiffies;flow_stats_update (&cls->stats, bytes - ofld_stats->byte_count, packets - ofld_stats->packet_count, - ofld_stats->last_used); + ofld_stats->last_used,0 ); ofld_stats->packet_count = packets; ofld_stats->byte_count = bytes;
drivers/net/ethernet/chelsio/cxgb4/cxgb4_tc_matchall.c +1 -1
@@ -353,7 +353,7 @@ int cxgb4_tc_matchall_stats(struct net_device *dev,flow_stats_update (&cls_matchall->stats, bytes - tc_port_matchall->ingress.bytes, packets - tc_port_matchall->ingress.packets, - tc_port_matchall->ingress.last_used); + tc_port_matchall->ingress.last_used,0 ); tc_port_matchall->ingress.packets = packets; tc_port_matchall->ingress.bytes = bytes;
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +2 -2
@@ -4091,7 +4091,7 @@ int mlx5e_stats_flower(struct net_device *dev, struct mlx5e_priv *priv, no_peer_counter:mlx5_devcom_release_peer_data (devcom,MLX5_DEVCOM_ESW_OFFLOADS ); out: -flow_stats_update (&f->stats, bytes, packets, lastuse); +flow_stats_update (&f->stats, bytes, packets, lastuse,0 );trace_mlx5e_stats_flower (f); errout:mlx5e_flow_put (priv, flow);
@@ -4199,7 +4199,7 @@ void mlx5e_tc_stats_matchall(struct mlx5e_priv *priv, dpkts = cur_stats.rx_packets - rpriv->prev_vf_vport_stats.rx_packets; dbytes = cur_stats.rx_bytes - rpriv->prev_vf_vport_stats.rx_bytes; rpriv->prev_vf_vport_stats = cur_stats; -flow_stats_update (&ma->stats, dpkts, dbytes, jiffies); +flow_stats_update (&ma->stats, dpkts, dbytes, jiffies,0 ); } static void mlx5e_tc_hairpin_update_dead_peer(struct mlx5e_priv *priv,
drivers/net/ethernet/mellanox/mlxsw/spectrum_flower.c +1 -1
@@ -565,7 +565,7 @@ int mlxsw_sp_flower_stats(struct mlxsw_sp *mlxsw_sp,if (err) goto err_rule_get_stats; -flow_stats_update (&f->stats, bytes, packets, lastuse); +flow_stats_update (&f->stats, bytes, packets, lastuse,0 );mlxsw_sp_acl_ruleset_put (mlxsw_sp, ruleset);return 0 ;
drivers/net/ethernet/mscc/ocelot_flower.c +1 -1
@@ -220,7 +220,7 @@ int ocelot_cls_flower_stats(struct ocelot *ocelot, int port,if (ret)return ret; -flow_stats_update (&f->stats,0x0 , ace.stats.pkts,0x0 ); +flow_stats_update (&f->stats,0x0 , ace.stats.pkts,0x0 ,0x0 );return 0 ; }EXPORT_SYMBOL_GPL (ocelot_cls_flower_stats);
drivers/net/ethernet/netronome/nfp/flower/offload.c +1 -1
@@ -1490,7 +1490,7 @@ nfp_flower_get_stats(struct nfp_app *app, struct net_device *netdev,nfp_flower_update_merge_stats (app, nfp_flow);flow_stats_update (&flow->stats, priv->stats[ctx_id].bytes, - priv->stats[ctx_id].pkts, priv->stats[ctx_id].used); + priv->stats[ctx_id].pkts, priv->stats[ctx_id].used,0 ); priv->stats[ctx_id].pkts =0 ; priv->stats[ctx_id].bytes =0 ;
drivers/net/ethernet/netronome/nfp/flower/qos_conf.c +1 -1
@@ -320,7 +320,7 @@ nfp_flower_stats_rate_limiter(struct nfp_app *app, struct net_device *netdev,spin_unlock_bh (&fl_priv->qos_stats_lock);flow_stats_update (&flow->stats, diff_bytes, diff_pkts, - repr_priv->qos_table.last_update); + repr_priv->qos_table.last_update,0 );return 0 ; }
include/net/act_api.h +6 -5
@@ -99,7 +99,7 @@ struct tc_action_ops { struct netlink_callback *, int, const struct tc_action_ops *, struct netlink_ext_ack *); -void (*stats_update)(struct tc_action *,u64 ,u32 ,u64 ,bool ); +void (*stats_update)(struct tc_action *,u64 ,u64 ,u64 ,u64 ,bool );size_t (*get_fill_size)(const struct tc_action *act);struct net_device *(*get_dev)(const struct tc_action *a,tc_action_priv_destructor *destructor);
@@ -225,8 +225,8 @@ static inline void tcf_action_inc_overlimit_qstats(struct tc_action *a)spin_unlock (&a->tcfa_lock); } -void tcf_action_update_stats (struct tc_action *a,u64 bytes,u32 packets, -bool drop,bool hw); +void tcf_action_update_stats (struct tc_action *a,u64 bytes,u64 packets, +u64 dropped,bool hw);int tcf_action_copy_stats (struct sk_buff *,struct tc_action *,int ); int tcf_action_check_ctrlact(int action, struct tcf_proto *tp,
@@ -237,13 +237,14 @@ struct tcf_chain *tcf_action_set_ctrlact(struct tc_action *a, int action,#endif /* CONFIG_NET_CLS_ACT */ static inline void tcf_action_stats_update (struct tc_action *a,u64 bytes, -u64 packets,u64 lastuse,bool hw) +u64 packets,u64 lastuse, +u64 dropped,bool hw) {#ifdef CONFIG_NET_CLS_ACT if (!a->ops->stats_update)return ; - a->ops->stats_update (a, bytes, packets, lastuse, hw); + a->ops->stats_update (a, bytes, packets, lastuse, dropped, hw);#endif }
include/net/flow_offload.h +4 -1
@@ -268,14 +268,17 @@ struct flow_stats {u64 pkts;u64 bytes;u64 lastused; +u64 dropped; };static inline void flow_stats_update (struct flow_stats *flow_stats, -u64 bytes,u64 pkts,u64 lastused) +u64 bytes,u64 pkts, +u64 lastused,u64 dropped) { flow_stats->pkts += pkts; flow_stats->bytes += bytes; flow_stats->lastused =max_t (u64, flow_stats->lastused, lastused); + flow_stats->dropped += dropped; }enum flow_block_command {
include/net/pkt_cls.h +3 -2
@@ -249,7 +249,7 @@ static inline void tcf_exts_put_net(struct tcf_exts *exts)static inline void tcf_exts_stats_update (const struct tcf_exts *exts, -u64 bytes,u64 packets,u64 lastuse) +u64 bytes,u64 packets,u64 lastuse,u64 dropped) {#ifdef CONFIG_NET_CLS_ACT int i;
@@ -259,7 +259,8 @@ tcf_exts_stats_update(const struct tcf_exts *exts,for (i =0 ; i < exts->nr_actions; i++) {struct tc_action *a = exts->actions[i]; -tcf_action_stats_update (a, bytes, packets, lastuse, true); +tcf_action_stats_update (a, bytes, packets, + lastuse, dropped, true); }preempt_enable ();
net/sched/act_api.c +6 -6
@@ -1017,14 +1017,13 @@ int tcf_action_init(struct net *net, struct tcf_proto *tp, struct nlattr *nla,return err; } -void tcf_action_update_stats (struct tc_action *a,u64 bytes,u32 packets, -bool drop,bool hw) +void tcf_action_update_stats (struct tc_action *a,u64 bytes,u64 packets, +u64 dropped,bool hw) {if (a->cpu_bstats) {_bstats_cpu_update (this_cpu_ptr (a->cpu_bstats), bytes, packets); -if (drop) -this_cpu_ptr (a->cpu_qstats)->drops += packets; +this_cpu_ptr (a->cpu_qstats)->drops += dropped;if (hw) _bstats_cpu_update(this_cpu_ptr(a->cpu_bstats_hw),
@@ -1033,8 +1032,9 @@ void tcf_action_update_stats(struct tc_action *a, u64 bytes, u32 packets, }_bstats_update (&a->tcfa_bstats, bytes, packets); -if (drop) - a->tcfa_qstats.drops += packets; + + a->tcfa_qstats.drops += dropped; +if (hw)_bstats_update (&a->tcfa_bstats_hw, bytes, packets); }
net/sched/act_ct.c +3 -3
@@ -1232,12 +1232,12 @@ static int tcf_ct_search(struct net *net, struct tc_action **a, u32 index)return tcf_idr_search (tn, a, index); } -static void tcf_stats_update (struct tc_action *a,u64 bytes,u32 packets, -u64 lastuse,bool hw) +static void tcf_stats_update (struct tc_action *a,u64 bytes,u64 packets, +u64 lastuse,u64 dropped,bool hw) {struct tcf_ct *c =to_ct (a); -tcf_action_update_stats (a, bytes, packets, false, hw); +tcf_action_update_stats (a, bytes, packets, dropped, hw); c->tcf_tm.lastuse =max_t (u64, c->tcf_tm.lastuse, lastuse); }
net/sched/act_gact.c +4 -3
@@ -171,14 +171,15 @@ static int tcf_gact_act(struct sk_buff *skb, const struct tc_action *a,return action; } -static void tcf_gact_stats_update (struct tc_action *a,u64 bytes,u32 packets, -u64 lastuse,bool hw) +static void tcf_gact_stats_update (struct tc_action *a,u64 bytes,u64 packets, +u64 lastuse,u64 dropped,bool hw) {struct tcf_gact *gact =to_gact (a);int action =READ_ONCE (gact->tcf_action);struct tcf_t *tm = &gact->tcf_tm; -tcf_action_update_stats (a, bytes, packets, action ==TC_ACT_SHOT , hw); +tcf_action_update_stats (a, bytes, packets, + (action ==TC_ACT_SHOT ) ? packets :0 , hw); tm->lastuse =max_t (u64, tm->lastuse, lastuse); }
net/sched/act_mirred.c +3 -3
@@ -314,13 +314,13 @@ static int tcf_mirred_act(struct sk_buff *skb, const struct tc_action *a,return retval; } -static void tcf_stats_update (struct tc_action *a,u64 bytes,u32 packets, -u64 lastuse,bool hw) +static void tcf_stats_update (struct tc_action *a,u64 bytes,u64 packets, +u64 lastuse,u64 dropped,bool hw) {struct tcf_mirred *m =to_mirred (a);struct tcf_t *tm = &m->tcf_tm; -tcf_action_update_stats (a, bytes, packets, false, hw); +tcf_action_update_stats (a, bytes, packets, dropped, hw); tm->lastuse =max_t (u64, tm->lastuse, lastuse); }
net/sched/act_police.c +3 -3
@@ -288,13 +288,13 @@ static void tcf_police_cleanup(struct tc_action *a) }static void tcf_police_stats_update (struct tc_action *a, -u64 bytes,u32 packets, -u64 lastuse,bool hw) +u64 bytes,u64 packets, +u64 lastuse,u64 dropped,bool hw) {struct tcf_police *police =to_police (a);struct tcf_t *tm = &police->tcf_tm; -tcf_action_update_stats (a, bytes, packets, false, hw); +tcf_action_update_stats (a, bytes, packets, dropped, hw); tm->lastuse =max_t (u64, tm->lastuse, lastuse); }
net/sched/act_vlan.c +3 -3
@@ -302,13 +302,13 @@ static int tcf_vlan_walker(struct net *net, struct sk_buff *skb,return tcf_generic_walker (tn, skb, cb, type, ops, extack); } -static void tcf_vlan_stats_update (struct tc_action *a,u64 bytes,u32 packets, -u64 lastuse,bool hw) +static void tcf_vlan_stats_update (struct tc_action *a,u64 bytes,u64 packets, +u64 lastuse,u64 dropped,bool hw) {struct tcf_vlan *v =to_vlan (a);struct tcf_t *tm = &v->tcf_tm; -tcf_action_update_stats (a, bytes, packets, false, hw); +tcf_action_update_stats (a, bytes, packets, dropped, hw); tm->lastuse =max_t (u64, tm->lastuse, lastuse); }
net/sched/cls_flower.c +2 -1
@@ -492,7 +492,8 @@ static void fl_hw_update_stats(struct tcf_proto *tp, struct cls_fl_filter *f,tcf_exts_stats_update (&f->exts, cls_flower.stats.bytes, cls_flower.stats.pkts, - cls_flower.stats.lastused); + cls_flower.stats.lastused, + cls_flower.stats.dropped); }static void __fl_put (struct cls_fl_filter *f)
net/sched/cls_matchall.c +2 -1
@@ -338,7 +338,8 @@ static void mall_stats_hw_filter(struct tcf_proto *tp,tc_setup_cb_call (block,TC_SETUP_CLSMATCHALL , &cls_mall, false, true);tcf_exts_stats_update (&head->exts, cls_mall.stats.bytes, - cls_mall.stats.pkts, cls_mall.stats.lastused); + cls_mall.stats.pkts, cls_mall.stats.lastused, + cls_mall.stats.dropped); } static int mall_dump(struct net *net, struct tcf_proto *tp, void *fh,
Signature · 1 line
2.17.1
RFC 2/9 net: qos: introduce a gate control flow action
In reply to Po LiuIntroduce a ingress frame gate control flow action. tc create a gate action would provide a gate list to control when open/close state. when the gate open state, the flow could pass but not when gate state is close. The driver would repeat the gate list cyclically. User also could assign a time point to start the gate list by the basetime parameter. if the basetime has passed current time, start time would calculate by the cycletime of the gate list.
The action gate behavior try to keep according to the IEEE 802.1Qci spec. For the software simulation, require the user input the clock type.
Below is the setting example in user space:
Quoted 3 lines
tc qdisc add dev eth0 ingress
tc filter add dev eth0 parent ffff: protocol ip \
flower src_ip 192.168.0.20 \ action gate index 2 \ sched-entry OPEN 200000000 -1 -1 \ sched-entry CLOSE 100000000 -1 -1
Quoted 1 line
tc chain del dev eth0 ingress chain 0
"sched-entry" follow the name taprio style. gate state is "OPEN"/"CLOSE". Follow the period nanosecond. Then next -1 is internal priority value means which ingress queue should put. "-1" means wildcard. The last value optional specifies the maximum number of MSDU octets that are permitted to pass the gate during the specified time interval.
NOTE: This software simulator version not separate the admin/operation state machine. Update setting would overwrite stop the previos setting and waiting new cycle start.
--- include/net/tc_act/tc_gate.h | 54 +++ include/uapi/linux/pkt_cls.h | 1 + include/uapi/linux/tc_act/tc_gate.h | 47 ++ net/sched/Kconfig | 15 + net/sched/Makefile | 1 + net/sched/act_gate.c | 645 ++++++++++++++++++++++++++++ 6 files changed, 763 insertions(+) create mode 100644 include/net/tc_act/tc_gate.h create mode 100644 include/uapi/linux/tc_act/tc_gate.h create mode 100644 net/sched/act_gate.c
include/net/tc_act/tc_gate.h new file +54 -0
@@ -0,0 +1,54 @@ +// SPDX-License-Identifier: (GPL-2.0+) +/* Copyright 2020 NXP */ + +#ifndef __NET_TC_gate_H +#define __NET_TC_gate_H + +#include <net/act_api.h> +#include <linux/tc_act/tc_gate.h> + +struct tcfg_gate_entry { +int index; +u8 gate_state; +u32 interval; +s32 ipv; +s32 maxoctets; +struct list_head list; +}; + +struct tcf_gate_params { +s32 tcfg_priority; +u64 tcfg_basetime; +u64 tcfg_cycletime; +u64 tcfg_cycletime_ext; +u32 tcfg_flags; +s32 tcfg_clockid; +size_t num_entries; +struct list_head entries; +}; + +#define GATE_ACT_GATE_OPEN BIT(0) +#define GATE_ACT_PENDING BIT(1) +struct gate_action { +struct tcf_gate_params param; +spinlock_t entry_lock; +u8 current_gate_status; +ktime_t current_close_time; +u32 current_entry_octets; +s32 current_max_octets; +struct tcfg_gate_entry __rcu *next_entry; +struct hrtimer hitimer; +enum tk_offsets tk_offset; +struct rcu_head rcu; +}; + +struct tcf_gate { +struct tc_action common; +struct gate_action __rcu *actg; +}; +#define to_gate (a) ((struct tcf_gate *)a) + +#define get_gate_param (act) ((struct tcf_gate_params *)act) +#define get_gate_action (p) ((struct gate_action *)p) + +#endif
include/uapi/linux/pkt_cls.h +1 -0
@@ -112,6 +112,7 @@ enum tca_id {TCA_ID_CTINFO ,TCA_ID_MPLS ,TCA_ID_CT , +TCA_ID_GATE ,/* other actions go here */ __TCA_ID_MAX = 255 };
include/uapi/linux/tc_act/tc_gate.h new file +47 -0
47 lines not on this page · raw message
net/sched/Kconfig +15 -0
@@ -981,6 +981,21 @@ config NET_ACT_CT To compile this code as a module, choose M here: the module will be called act_ct. +config NET_ACT_GATE + tristate "Frame gate list control tc action" + depends on NET_CLS_ACT + help + Say Y here to allow the control the ingress flow by the gate list + control. The frame policing by the time gate list control open/close + cycle time. The manipulation will simulate the IEEE 802.1Qci stream + gate control behavior. The action could be offload by the tc flower + to hardware driver which the hardware own the capability of IEEE + 802.1Qci. + + If unsure, say N. + To compile this code as a module, choose M here: the + module will be called act_gate. + config NET_IFE_SKBMARK tristate "Support to encoding decoding skb mark on IFE action" depends on NET_ACT_IFE
net/sched/Makefile +1 -0
@@ -30,6 +30,7 @@ obj-$(CONFIG_NET_IFE_SKBPRIO) += act_meta_skbprio.o obj-$(CONFIG_NET_IFE_SKBTCINDEX) += act_meta_skbtcindex.o obj-$(CONFIG_NET_ACT_TUNNEL_KEY)+= act_tunnel_key.o obj-$(CONFIG_NET_ACT_CT) += act_ct.o +obj-$(CONFIG_NET_ACT_GATE) += act_gate.o obj-$(CONFIG_NET_SCH_FIFO) += sch_fifo.o obj-$(CONFIG_NET_SCH_CBQ) += sch_cbq.o obj-$(CONFIG_NET_SCH_HTB) += sch_htb.o
net/sched/act_gate.c new file +645 -0
645 lines not on this page · raw message
Signature · 1 line
2.17.1
Jakub Kicinski
maintainerRe: [RFC,net-next 2/9] net: qos: introduce a gate control flow action
Quoted 41 lines · On Fri, 6 Mar 2020 20:56:00 +0800 Po Liu wrote:
Introduce a ingress frame gate control flow action. tc create a gate action would provide a gate list to control when open/close state. when the gate open state, the flow could pass but not when gate state is close. The driver would repeat the gate list cyclically. User also could assign a time point to start the gate list by the basetime parameter. if the basetime has passed current time, start time would calculate by the cycletime of the gate list.
The action gate behavior try to keep according to the IEEE 802.1Qci spec. For the software simulation, require the user input the clock type.Below is the setting example in user space:
Quoted 3 lines
tc qdisc add dev eth0 ingress
tc filter add dev eth0 parent ffff: protocol ip \
flower src_ip 192.168.0.20 \ action gate index 2 \ sched-entry OPEN 200000000 -1 -1 \ sched-entry CLOSE 100000000 -1 -1Quoted 1 line
tc chain del dev eth0 ingress chain 0
"sched-entry" follow the name taprio style. gate state is "OPEN"/"CLOSE". Follow the period nanosecond. Then next -1 is internal priority value means which ingress queue should put. "-1" means wildcard. The last value optional specifies the maximum number of MSDU octets that are permitted to pass the gate during the specified time interval.
NOTE: This software simulator version not separate the admin/operation state machine. Update setting would overwrite stop the previos setting and waiting new cycle start.
Signed-off-by Po Liu <[email protected]>net/sched/act_gate.c new file +0
-0
net/sched/act_gate.c line 1
Quoted 1 line
net/sched/act_gate.c new file +1
-0@@ -0,0 +1,645 @@ +// SPDX-License-Identifier: (GPL-2.0+)
Why the parenthesis?
Quoted 18 lines
net/sched/act_gate.c new file +18
-0+static const struct nla_policy entry_policy[TCA_GATE_ENTRY_MAX +1 ] = { + [TCA_GATE_ENTRY_INDEX ] = { .type =NLA_U32 }, + [TCA_GATE_ENTRY_GATE ] = { .type =NLA_FLAG }, + [TCA_GATE_ENTRY_INTERVAL ] = { .type =NLA_U32 }, + [TCA_GATE_ENTRY_IPV ] = { .type =NLA_S32 }, + [TCA_GATE_ENTRY_MAX_OCTETS ] = { .type =NLA_S32 }, +}; + +static const struct nla_policy gate_policy[TCA_GATE_MAX +1 ] = { + [TCA_GATE_PARMS ] = { .len =sizeof (struct tc_gate ) }, + [TCA_GATE_PRIORITY ] = { .type =NLA_S32 }, + [TCA_GATE_ENTRY_LIST ] = { .type =NLA_NESTED }, + [TCA_GATE_BASE_TIME ] = { .type =NLA_U64 }, + [TCA_GATE_CYCLE_TIME ] = { .type =NLA_U64 }, + [TCA_GATE_CYCLE_TIME_EXT ] = { .type =NLA_U64 }, + [TCA_GATE_FLAGS ] = { .type =NLA_U32 }, + [TCA_GATE_CLOCKID ] = { .type =NLA_S32 }, +};
net/sched/act_gate.c line 188
Quoted 10 lines
net/sched/act_gate.c new file +10
-0+ +static int fill_gate_entry (struct nlattr **tb,struct tcfg_gate_entry *entry, +struct netlink_ext_ack *extack) +{ +u32 interval =0 ; + +if (tb[TCA_GATE_ENTRY_GATE ]) + entry->gate_state =1 ; +else + entry->gate_state =0 ;
nla_get_flag()
Quoted 27 lines
net/sched/act_gate.c new file +26
-0+ +if (tb[TCA_GATE_ENTRY_INTERVAL ]) + interval =nla_get_u32 (tb[TCA_GATE_ENTRY_INTERVAL ]); + +if (interval ==0 ) { +NL_SET_ERR_MSG (extack,"Invalid interval for schedule entry" ); +return -EINVAL ; + } +static int parse_gate_list (struct nlattr *list, +struct tcf_gate_params *sched, +struct netlink_ext_ack *extack) +{ +struct nlattr *n; +int err, rem; +int i =0 ; + +if (!list) +return -EINVAL ; + +nla_for_each_nested (n, list, rem) { +struct tcfg_gate_entry *entry; + +if (nla_type (n) !=TCA_GATE_ONE_ENTRY ) { +NL_SET_ERR_MSG (extack,"Attribute isn't type 'entry'" ); +continue ; + }
net/sched/act_gate.c line 258
Quoted 10 lines
net/sched/act_gate.c new file +10
-0+ + entry =kzalloc (sizeof (*entry),GFP_KERNEL ); +if (!entry) { +NL_SET_ERR_MSG (extack,"Not enough memory for entry" ); +return -ENOMEM ; + } + + err =parse_gate_entry (n, entry, i, extack); +if (err <0 ) { +kfree (entry);
doesn't this leak previously added entries?
Quoted 10 lines
net/sched/act_gate.c new file +10
-0+return err; + } + +list_add_tail (&entry->list, &sched->entries); + i++; + } + + sched->num_entries = i; + +return i;
net/sched/act_gate.c
Quoted 10 lines
net/sched/act_gate.c new file +8
-0+} +static int parse_gate_entry (struct nlattr *n,struct tcfg_gate_entry *entry, +int index,struct netlink_ext_ack *extack) +{ +struct nlattr *tb[TCA_GATE_ENTRY_MAX +1 ] = { }; +int err; + + err = nla_parse_nested_deprecated(tb, TCA_GATE_ENTRY_MAX, n,
Please don't use the deprecated calls in new code.
Quoted 10 lines
net/sched/act_gate.c new file +10
-0+ entry_policy, NULL); +if (err <0 ) { +NL_SET_ERR_MSG (extack,"Could not parse nested entry" ); +return -EINVAL ; + } + + entry->index = index; + +return fill_gate_entry (tb, entry, extack); +}
RE: [EXT] Re: [RFC,net-next 2/9] net: qos: introduce a gate control flow action
Hi Jakub,
Quoted 62 lines
-----Original Message-----
From: Jakub Kicinski <[email protected]>
Sent: 2020年3月7日 3:11
To: Po Liu <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; Claudiu Manoil <[email protected]>; Vladimir Oltean <[email protected]>; Alexandru Marginean <[email protected]>; Xiaoliang Yang <[email protected]>; Roy Zang <[email protected]>; Mingkai Hu <[email protected]>; Jerry Huang <[email protected]>; Leo Li <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected];
[email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]
Subject: [EXT] Re: [RFC,net-next 2/9] net: qos: introduce a gate control flow actionCaution: EXT Email
Quoted 20 lines · On Fri, 6 Mar 2020 20:56:00 +0800 Po Liu wrote:
Introduce a ingress frame gate control flow action. tc create a gate action would provide a gate list to control when open/close state. when the gate open state, the flow could pass but not when gate state is close. The driver would repeat the gate list cyclically. User also could assign a time point to start the gate list by the basetime parameter. if the basetime has passed current time, start time would calculate by the cycletime of the gate list.
The action gate behavior try to keep according to the IEEE 802.1Qci spec. For the software simulation, require the user input the clock type.Below is the setting example in user space:
Quoted 3 lines
tc qdisc add dev eth0 ingress
tc filter add dev eth0 parent ffff: protocol ip \
flower src_ip 192.168.0.20 \ action gate index 2 \ sched-entry OPEN 200000000 -1 -1 \ sched-entry CLOSE 100000000 -1 -1Quoted 6 lines
Quoted 1 line
tc chain del dev eth0 ingress chain 0
"sched-entry" follow the name taprio style. gate state is "OPEN"/"CLOSE". Follow the period nanosecond. Then next -1 is internal priority value means which ingress queue should put. "-1" means wildcard. The last value optional specifies the maximum number of
MSDU
Quoted 4 lines
octets that are permitted to pass the gate during the specified time interval.
NOTE: This software simulator version not separate the
admin/operation
Quoted 4 lines
state machine. Update setting would overwrite stop the previos setting and waiting new cycle start.
Signed-off-by Po Liu <[email protected]>
Quoted 8 lines
net/sched/act_gate.c line 1
Quoted 6 lines
net/sched/act_gate.c new file mode renamed from net/sched/act_gate.c +0
-0100644 index 000000000000..c2c243ca028c
net/sched/act_gate.c new file +1
-0@@ -0,0 +1,645 @@ +// SPDX-License-Identifier: (GPL-2.0+) Why the parenthesis?
Will changed to GPL-2.0-or-later.
Quoted 23 lines
net/sched/act_gate.c line 2
Quoted 2 lines
net/sched/act_gate.c new file +1
-0+static const struct nla_policy entry_policy[TCA_GATE_ENTRY_MAX +1 ]= {
Quoted 20 lines
net/sched/act_gate.c new file +20
-0+ [TCA_GATE_ENTRY_INDEX] = { .type = NLA_U32 }, + [TCA_GATE_ENTRY_GATE] = { .type = NLA_FLAG }, + [TCA_GATE_ENTRY_INTERVAL] = { .type = NLA_U32 }, + [TCA_GATE_ENTRY_IPV] = { .type = NLA_S32 }, + [TCA_GATE_ENTRY_MAX_OCTETS] = { .type = NLA_S32 }, +}; + +static const struct nla_policy gate_policy[TCA_GATE_MAX +1 ] = { + [TCA_GATE_PARMS ] = { .len =sizeof (struct tc_gate ) }, + [TCA_GATE_PRIORITY ] = { .type =NLA_S32 }, + [TCA_GATE_ENTRY_LIST ] = { .type =NLA_NESTED }, + [TCA_GATE_BASE_TIME ] = { .type =NLA_U64 }, + [TCA_GATE_CYCLE_TIME ] = { .type =NLA_U64 }, + [TCA_GATE_CYCLE_TIME_EXT ] = { .type =NLA_U64 }, + [TCA_GATE_FLAGS ] = { .type =NLA_U32 }, + [TCA_GATE_CLOCKID ] = { .type =NLA_S32 }, +}; + +static int fill_gate_entry (struct nlattr **tb,struct tcfg_gate_entry *entry, +struct netlink_ext_ack *extack) {
Quoted 8 lines
net/sched/act_gate.c line 188
Quoted 6 lines
net/sched/act_gate.c new file +6
-0+u32 interval =0 ; + +if (tb[TCA_GATE_ENTRY_GATE ]) + entry->gate_state =1 ; +else + entry->gate_state =0 ;
nla_get_flag()
Ok, will parse by the nla_get_flag()
Quoted 31 lines
Quoted 31 lines
net/sched/act_gate.c new file +29
-0+ +if (tb[TCA_GATE_ENTRY_INTERVAL ]) + interval =nla_get_u32 (tb[TCA_GATE_ENTRY_INTERVAL ]); + +if (interval ==0 ) { +NL_SET_ERR_MSG (extack,"Invalid interval for schedule entry" ); +return -EINVAL ; + } +static int parse_gate_list (struct nlattr *list, +struct tcf_gate_params *sched, +struct netlink_ext_ack *extack) { +struct nlattr *n; +int err, rem; +int i =0 ; + +if (!list) +return -EINVAL ; + +nla_for_each_nested (n, list, rem) { +struct tcfg_gate_entry *entry; + +if (nla_type (n) !=TCA_GATE_ONE_ENTRY ) { +NL_SET_ERR_MSG (extack,"Attribute isn't type 'entry'" ); +continue ; + } + + entry =kzalloc (sizeof (*entry),GFP_KERNEL ); +if (!entry) { +NL_SET_ERR_MSG (extack,"Not enough memory for entry" );
Quoted 8 lines
net/sched/act_gate.c line 258
Quoted 6 lines
net/sched/act_gate.c new file +6
-0+return -ENOMEM ; + } + + err =parse_gate_entry (n, entry, i, extack); +if (err <0 ) { +kfree (entry);doesn't this leak previously added entries?
Ok, will fix it.
Quoted 15 lines
Quoted 5 lines
net/sched/act_gate.c new file +4
-0+return err; + } + +list_add_tail (&entry->list, &sched->entries);net/sched/act_gate.c
Quoted 10 lines
net/sched/act_gate.c new file +8
-0+ i++; + } + + sched->num_entries = i; + +return i; +} +static int parse_gate_entry(struct nlattr *n, struct tcfg_gate_entry
Quoted 8 lines
*entry,
net/sched/act_gate.c line 219
Quoted 5 lines
net/sched/act_gate.c new file +5
-0+ int index, struct netlink_ext_ack *extack) { + struct nlattr *tb[TCA_GATE_ENTRY_MAX + 1] = { }; +int err; + + err = nla_parse_nested_deprecated(tb, TCA_GATE_ENTRY_MAX, n,Please don't use the deprecated calls in new code.
Understand, will change to nla_parse_nested()
Quoted 10 lines
net/sched/act_gate.c
Quoted 10 lines
net/sched/act_gate.c new file +9
-0+ entry_policy, NULL); +if (err <0 ) { +NL_SET_ERR_MSG (extack,"Could not parse nested entry" ); +return -EINVAL ; + } + + entry->index = index; + +return fill_gate_entry (tb, entry, extack); }
Thanks!
Po Liu
Re: [RFC,net-next 2/9] net: qos: introduce a gate control flow action
In reply to Po LiuQuoted 12 lines · Po Liu
Introduce a ingress frame gate control flow action. tc create a gate action would provide a gate list to control when open/close state. when the gate open state, the flow could pass but not when gate state is close. The driver would repeat the gate list cyclically. User also could assign a time point to start the gate list by the basetime parameter. if the basetime has passed current time, start time would calculate by the cycletime of the gate list.
The action gate behavior try to keep according to the IEEE 802.1Qci spec. For the software simulation, require the user input the clock type.Below is the setting example in user space:
Quoted 9 lines
Quoted 3 lines
tc qdisc add dev eth0 ingress
tc filter add dev eth0 parent ffff: protocol ip \
flower src_ip 192.168.0.20 \ action gate index 2 \ sched-entry OPEN 200000000 -1 -1 \ sched-entry CLOSE 100000000 -1 -1Quoted 1 line
tc chain del dev eth0 ingress chain 0
A more complex example, showing how it would work for more filters would be nice.
For example, filters matching 3 different source IPs (even better, using MAC addresses so it would work when you add the 'skip_sw' flag), assigning different priorities to them, also showing how the "base-time" could be used.
Quoted 316 lines
"sched-entry" follow the name taprio style. gate state is "OPEN"/"CLOSE". Follow the period nanosecond. Then next -1 is internal priority value means which ingress queue should put. "-1" means wildcard. The last value optional specifies the maximum number of MSDU octets that are permitted to pass the gate during the specified time interval.
NOTE: This software simulator version not separate the admin/operation state machine. Update setting would overwrite stop the previos setting and waiting new cycle start.
Signed-off-by Po Liu <[email protected]>--- include/net/tc_act/tc_gate.h | 54 +++ include/uapi/linux/pkt_cls.h | 1 + include/uapi/linux/tc_act/tc_gate.h | 47 ++ net/sched/Kconfig | 15 + net/sched/Makefile | 1 + net/sched/act_gate.c | 645 ++++++++++++++++++++++++++++ 6 files changed, 763 insertions(+) create mode 100644 include/net/tc_act/tc_gate.h create mode 100644 include/uapi/linux/tc_act/tc_gate.h create mode 100644 net/sched/act_gate.cinclude/net/tc_act/tc_gate.h new file +54
-0@@ -0,0 +1,54 @@ +// SPDX-License-Identifier: (GPL-2.0+) +/* Copyright 2020 NXP */ + +#ifndef __NET_TC_gate_H +#define __NET_TC_gate_H + +#include <net/act_api.h> +#include <linux/tc_act/tc_gate.h> + +struct tcfg_gate_entry { +int index; +u8 gate_state; +u32 interval; +s32 ipv; +s32 maxoctets; +struct list_head list; +}; + +struct tcf_gate_params { +s32 tcfg_priority; +u64 tcfg_basetime; +u64 tcfg_cycletime; +u64 tcfg_cycletime_ext; +u32 tcfg_flags; +s32 tcfg_clockid; +size_t num_entries; +struct list_head entries; +}; + +#define GATE_ACT_GATE_OPEN BIT(0) +#define GATE_ACT_PENDING BIT(1) +struct gate_action { +struct tcf_gate_params param; +spinlock_t entry_lock; +u8 current_gate_status; +ktime_t current_close_time; +u32 current_entry_octets; +s32 current_max_octets; +struct tcfg_gate_entry __rcu *next_entry; +struct hrtimer hitimer; +enum tk_offsets tk_offset; +struct rcu_head rcu; +}; + +struct tcf_gate { +struct tc_action common; +struct gate_action __rcu *actg; +}; +#define to_gate (a) ((struct tcf_gate *)a) + +#define get_gate_param (act) ((struct tcf_gate_params *)act) +#define get_gate_action (p) ((struct gate_action *)p) + +#endif include/uapi/linux/pkt_cls.h +1
-0@@ -112,6 +112,7 @@ enum tca_id {TCA_ID_CTINFO ,TCA_ID_MPLS ,TCA_ID_CT , +TCA_ID_GATE ,/* other actions go here */ __TCA_ID_MAX = 255 };include/uapi/linux/tc_act/tc_gate.h new file +47
-0@@ -0,0 +1,47 @@ +// SPDX-License-Identifier: (GPL-2.0+) +/* Copyright 2020 NXP */ + +#ifndef __LINUX_TC_GATE_H +#define __LINUX_TC_GATE_H + +#include <linux/pkt_cls.h> + +struct tc_gate { +tc_gen ; +}; + +enum { +TCA_GATE_ENTRY_UNSPEC , +TCA_GATE_ENTRY_INDEX , +TCA_GATE_ENTRY_GATE , +TCA_GATE_ENTRY_INTERVAL , +TCA_GATE_ENTRY_IPV , +TCA_GATE_ENTRY_MAX_OCTETS , + __TCA_GATE_ENTRY_MAX, +}; +#define TCA_GATE_ENTRY_MAX (__TCA_GATE_ENTRY_MAX - 1) + +enum { +TCA_GATE_ONE_ENTRY_UNSPEC , +TCA_GATE_ONE_ENTRY , + __TCA_GATE_ONE_ENTRY_MAX, +}; +#define TCA_GATE_ONE_ENTRY_MAX (__TCA_GATE_ONE_ENTRY_MAX - 1) + +enum { +TCA_GATE_UNSPEC , +TCA_GATE_TM , +TCA_GATE_PARMS , +TCA_GATE_PAD , +TCA_GATE_PRIORITY , +TCA_GATE_ENTRY_LIST , +TCA_GATE_BASE_TIME , +TCA_GATE_CYCLE_TIME , +TCA_GATE_CYCLE_TIME_EXT , +TCA_GATE_FLAGS , +TCA_GATE_CLOCKID , + __TCA_GATE_MAX, +}; +#define TCA_GATE_MAX (__TCA_GATE_MAX - 1) + +#endif net/sched/Kconfig +15
-0@@ -981,6 +981,21 @@ config NET_ACT_CT To compile this code as a module, choose M here: the module will be called act_ct. +config NET_ACT_GATE + tristate "Frame gate list control tc action" + depends on NET_CLS_ACT + help + Say Y here to allow the control the ingress flow by the gate list + control. The frame policing by the time gate list control open/close + cycle time. The manipulation will simulate the IEEE 802.1Qci stream + gate control behavior. The action could be offload by the tc flower + to hardware driver which the hardware own the capability of IEEE + 802.1Qci. + + If unsure, say N. + To compile this code as a module, choose M here: the + module will be called act_gate. + config NET_IFE_SKBMARK tristate "Support to encoding decoding skb mark on IFE action" depends on NET_ACT_IFEnet/sched/Makefile +1
-0@@ -30,6 +30,7 @@ obj-$(CONFIG_NET_IFE_SKBPRIO) += act_meta_skbprio.o obj-$(CONFIG_NET_IFE_SKBTCINDEX) += act_meta_skbtcindex.o obj-$(CONFIG_NET_ACT_TUNNEL_KEY)+= act_tunnel_key.o obj-$(CONFIG_NET_ACT_CT) += act_ct.o +obj-$(CONFIG_NET_ACT_GATE) += act_gate.o obj-$(CONFIG_NET_SCH_FIFO) += sch_fifo.o obj-$(CONFIG_NET_SCH_CBQ) += sch_cbq.o obj-$(CONFIG_NET_SCH_HTB) += sch_htb.onet/sched/act_gate.c new file +122
-0@@ -0,0 +1,645 @@ +// SPDX-License-Identifier: (GPL-2.0+) +/* Copyright 2020 NXP */ + +#include <linux/module.h> +#include <linux/types.h> +#include <linux/kernel.h> +#include <linux/string.h> +#include <linux/errno.h> +#include <linux/skbuff.h> +#include <linux/rtnetlink.h> +#include <linux/init.h> +#include <linux/slab.h> +#include <net/act_api.h> +#include <net/netlink.h> +#include <net/pkt_cls.h> +#include <net/tc_act/tc_gate.h> + +static unsigned int gate_net_id; +static struct tc_action_ops act_gate_ops; + +static ktime_t gate_get_time (struct gate_action *gact) +{ +ktime_t mono =ktime_get (); + +switch (gact->tk_offset) { +case TK_OFFS_MAX : +return mono; +default : +return ktime_mono_to_any (mono, gact->tk_offset); + } + +return KTIME_MAX ; +} + +static int gate_get_start_time (struct gate_action *gact,ktime_t *start) +{ +struct tcf_gate_params *param =get_gate_param (gact); +ktime_t now, base, cycle; +u64 n; + + base =ns_to_ktime (param->tcfg_basetime); + now =gate_get_time (gact); + +if (ktime_after (base, now)) { + *start = base; +return 0 ; + } + + cycle = param->tcfg_cycletime; + +/* cycle time should not be zero */ +if (WARN_ON (!cycle)) +return -EFAULT ; + + n =div64_u64 (ktime_sub_ns (now, base), cycle); + *start =ktime_add_ns (base, (n +1 ) * cycle); +return 0 ; +} + +static void gate_start_timer (struct gate_action *gact,ktime_t start) +{ +ktime_t expires; + + expires =hrtimer_get_expires (&gact->hitimer); +if (expires ==0 ) + expires =KTIME_MAX ; + + start =min_t (ktime_t, start, expires); + +hrtimer_start (&gact->hitimer, start,HRTIMER_MODE_ABS ); +} + +static enum hrtimer_restart gate_timer_func (struct hrtimer *timer) +{ + struct gate_action *gact = container_of(timer, struct gate_action, + hitimer); +struct tcf_gate_params *p =get_gate_param (gact); +struct tcfg_gate_entry *next; +ktime_t close_time, now; + +spin_lock (&gact->entry_lock); + + next =rcu_dereference_protected (gact->next_entry, +lockdep_is_held (&gact->entry_lock)); + +/* cycle start, clear pending bit, clear total octets */ + gact->current_gate_status = next->gate_state ?GATE_ACT_GATE_OPEN :0 ; + gact->current_entry_octets =0 ; + gact->current_max_octets = next->maxoctets; + + gact->current_close_time =ktime_add_ns (gact->current_close_time, + next->interval); + + close_time = gact->current_close_time; + +if (list_is_last (&next->list, &p->entries)) + next =list_first_entry (&p->entries, + struct tcfg_gate_entry, list); +else + next =list_next_entry (next, list); + + now =gate_get_time (gact); + +if (ktime_after (now, close_time)) { +ktime_t cycle, base; +u64 n; + + cycle = p->tcfg_cycletime; + base =ns_to_ktime (p->tcfg_basetime); + n =div64_u64 (ktime_sub_ns (now, base), cycle); + close_time =ktime_add_ns (base, (n +1 ) * cycle); + } + +rcu_assign_pointer (gact->next_entry, next); +spin_unlock (&gact->entry_lock); + +hrtimer_set_expires (&gact->hitimer, close_time); + +return HRTIMER_RESTART ; +} + +static int tcf_gate_act(struct sk_buff *skb, const struct tc_action *a,
net/sched/act_gate.c line 132
Quoted 10 lines
net/sched/act_gate.c new file +10
-0+ struct tcf_result *res) +{ +struct tcf_gate *g =to_gate (a); +struct tcf_gate_params *p; +struct gate_action *gact; +int action; + +tcf_lastuse_update (&g->tcf_tm); +tcf_action_update_bstats (&g->common, skb); +
Please test this with lockdep enabled, I got the feeling that there's a missing rcu_read_lock() somewhere around here.
Quoted 259 lines
net/sched/act_gate.c new file +259
-0+ action =READ_ONCE (g->tcf_action); + gact =rcu_dereference_bh (g->actg); + p =get_gate_param (gact); + +if (gact->current_gate_status &GATE_ACT_PENDING ) +return action; + +if (!(gact->current_gate_status &GATE_ACT_GATE_OPEN )) + goto drop; + +spin_lock (&gact->entry_lock); +if (gact->current_max_octets >=0 ) { + gact->current_entry_octets +=qdisc_pkt_len (skb); +if (gact->current_entry_octets > gact->current_max_octets) { +tcf_action_inc_overlimit_qstats (&g->common); +spin_unlock (&gact->entry_lock); + goto drop; + } + } +spin_unlock (&gact->entry_lock); + +return action; +drop: +tcf_action_inc_drop_qstats (&g->common); + +return TC_ACT_SHOT ; +} + +static const struct nla_policy entry_policy[TCA_GATE_ENTRY_MAX +1 ] = { + [TCA_GATE_ENTRY_INDEX ] = { .type =NLA_U32 }, + [TCA_GATE_ENTRY_GATE ] = { .type =NLA_FLAG }, + [TCA_GATE_ENTRY_INTERVAL ] = { .type =NLA_U32 }, + [TCA_GATE_ENTRY_IPV ] = { .type =NLA_S32 }, + [TCA_GATE_ENTRY_MAX_OCTETS ] = { .type =NLA_S32 }, +}; + +static const struct nla_policy gate_policy[TCA_GATE_MAX +1 ] = { + [TCA_GATE_PARMS ] = { .len =sizeof (struct tc_gate ) }, + [TCA_GATE_PRIORITY ] = { .type =NLA_S32 }, + [TCA_GATE_ENTRY_LIST ] = { .type =NLA_NESTED }, + [TCA_GATE_BASE_TIME ] = { .type =NLA_U64 }, + [TCA_GATE_CYCLE_TIME ] = { .type =NLA_U64 }, + [TCA_GATE_CYCLE_TIME_EXT ] = { .type =NLA_U64 }, + [TCA_GATE_FLAGS ] = { .type =NLA_U32 }, + [TCA_GATE_CLOCKID ] = { .type =NLA_S32 }, +}; + +static int fill_gate_entry (struct nlattr **tb,struct tcfg_gate_entry *entry, +struct netlink_ext_ack *extack) +{ +u32 interval =0 ; + +if (tb[TCA_GATE_ENTRY_GATE ]) + entry->gate_state =1 ; +else + entry->gate_state =0 ; + +if (tb[TCA_GATE_ENTRY_INTERVAL ]) + interval =nla_get_u32 (tb[TCA_GATE_ENTRY_INTERVAL ]); + +if (interval ==0 ) { +NL_SET_ERR_MSG (extack,"Invalid interval for schedule entry" ); +return -EINVAL ; + } + + entry->interval = interval; + +if (tb[TCA_GATE_ENTRY_IPV ]) + entry->ipv =nla_get_s32 (tb[TCA_GATE_ENTRY_IPV ]); +else + entry->ipv =-1 ; + +if (tb[TCA_GATE_ENTRY_MAX_OCTETS ]) + entry->maxoctets =nla_get_s32 (tb[TCA_GATE_ENTRY_MAX_OCTETS ]); +else + entry->maxoctets =-1 ; + +return 0 ; +} + +static int parse_gate_entry (struct nlattr *n,struct tcfg_gate_entry *entry, +int index,struct netlink_ext_ack *extack) +{ +struct nlattr *tb[TCA_GATE_ENTRY_MAX +1 ] = { }; +int err; + + err =nla_parse_nested_deprecated (tb,TCA_GATE_ENTRY_MAX , n, + entry_policy,NULL ); +if (err <0 ) { +NL_SET_ERR_MSG (extack,"Could not parse nested entry" ); +return -EINVAL ; + } + + entry->index = index; + +return fill_gate_entry (tb, entry, extack); +} + +static int parse_gate_list (struct nlattr *list, +struct tcf_gate_params *sched, +struct netlink_ext_ack *extack) +{ +struct nlattr *n; +int err, rem; +int i =0 ; + +if (!list) +return -EINVAL ; + +nla_for_each_nested (n, list, rem) { +struct tcfg_gate_entry *entry; + +if (nla_type (n) !=TCA_GATE_ONE_ENTRY ) { +NL_SET_ERR_MSG (extack,"Attribute isn't type 'entry'" ); +continue ; + } + + entry =kzalloc (sizeof (*entry),GFP_KERNEL ); +if (!entry) { +NL_SET_ERR_MSG (extack,"Not enough memory for entry" ); +return -ENOMEM ; + } + + err =parse_gate_entry (n, entry, i, extack); +if (err <0 ) { +kfree (entry); +return err; + } + +list_add_tail (&entry->list, &sched->entries); + i++; + } + + sched->num_entries = i; + +return i; +} + +static int tcf_gate_init (struct net *net,struct nlattr *nla, +struct nlattr *est,struct tc_action **a, +int ovr,int bind,bool rtnl_held, +struct tcf_proto *tp,u32 flags, +struct netlink_ext_ack *extack) +{ +struct tc_action_net *tn =net_generic (net, gate_net_id); +enum tk_offsets tk_offset =TK_OFFS_TAI ; +struct nlattr *tb[TCA_GATE_MAX +1 ]; +struct tcf_chain *goto_ch =NULL ; +struct tcfg_gate_entry *next; +struct tcf_gate_params *p; +struct gate_action *gact; +s32 clockid =CLOCK_TAI ; +struct tc_gate *parm; +struct tcf_gate *g; +int ret =0 , err; +u64 basetime =0 ; +u32 gflags =0 ; +s32 prio =-1 ; +ktime_t start; +u32 index; + +if (!nla) +return -EINVAL ; + + err =nla_parse_nested_deprecated (tb,TCA_GATE_MAX , + nla, gate_policy,NULL ); +if (err <0 ) +return err; + +if (!tb[TCA_GATE_PARMS ]) +return -EINVAL ; + parm =nla_data (tb[TCA_GATE_PARMS ]); + index = parm->index; + err =tcf_idr_check_alloc (tn, &index, a, bind); +if (err <0 ) +return err; + +if (err && bind) +return 0 ; + +if (!err) { + ret =tcf_idr_create_from_flags (tn, index, est, a, + &act_gate_ops, bind, flags); +if (ret) { +tcf_idr_cleanup (tn, index); +return ret; + } + + ret =ACT_P_CREATED ; + }else if (!ovr) { +tcf_idr_release (*a, bind); +return -EEXIST ; + } + +if (tb[TCA_GATE_PRIORITY ]) + prio =nla_get_s32 (tb[TCA_GATE_PRIORITY ]); + +if (tb[TCA_GATE_BASE_TIME ]) + basetime =nla_get_u64 (tb[TCA_GATE_BASE_TIME ]); + +if (tb[TCA_GATE_FLAGS ]) + gflags =nla_get_u32 (tb[TCA_GATE_FLAGS ]); + +if (tb[TCA_GATE_CLOCKID ]) { + clockid =nla_get_s32 (tb[TCA_GATE_CLOCKID ]); +switch (clockid) { +case CLOCK_REALTIME : + tk_offset =TK_OFFS_REAL ; +break ; +case CLOCK_MONOTONIC : + tk_offset =TK_OFFS_MAX ; +break ; +case CLOCK_BOOTTIME : + tk_offset =TK_OFFS_BOOT ; +break ; +case CLOCK_TAI : + tk_offset =TK_OFFS_TAI ; +break ; +default : +NL_SET_ERR_MSG (extack,"Invalid 'clockid'" ); + goto release_idr; + } + } + + err =tcf_action_check_ctrlact (parm->action, tp, &goto_ch, extack); +if (err <0 ) + goto release_idr; + + g =to_gate (*a); + + gact =kzalloc (sizeof (*gact),GFP_KERNEL ); +if (!gact) { + err = -ENOMEM ; + goto put_chain; + } + + p =get_gate_param (gact); + +INIT_LIST_HEAD (&p->entries); +if (tb[TCA_GATE_ENTRY_LIST ]) { + err =parse_gate_list (tb[TCA_GATE_ENTRY_LIST ], p, extack); +if (err <0 ) + goto release_mem; + } + +if (tb[TCA_GATE_CYCLE_TIME ]) { + p->tcfg_cycletime =nla_get_u64 (tb[TCA_GATE_CYCLE_TIME ]); + }else { +struct tcfg_gate_entry *entry; +ktime_t cycle =0 ; + +list_for_each_entry (entry, &p->entries, list) + cycle =ktime_add_ns (cycle, entry->interval); + p->tcfg_cycletime = cycle; + } + +if (tb[TCA_GATE_CYCLE_TIME_EXT ]) + p->tcfg_cycletime_ext = +nla_get_u64 (tb[TCA_GATE_CYCLE_TIME_EXT ]);
net/sched/act_gate.c line 401
Quoted 10 lines
net/sched/act_gate.c new file +10
-0+ + p->tcfg_priority = prio; + p->tcfg_basetime = basetime; + p->tcfg_clockid = clockid; + p->tcfg_flags = gflags; + + gact->tk_offset = tk_offset; +spin_lock_init (&gact->entry_lock); +hrtimer_init (&gact->hitimer, clockid,HRTIMER_MODE_ABS ); + gact->hitimer.function = gate_timer_func;
Hm, having an hrtimer per filter seems kind dangerous, in the sense that it seems very hard to configure right.
But I don't have any alternative ideas for now.
Quoted 247 lines
net/sched/act_gate.c new file +244
-0+ + err =gate_get_start_time (gact, &start); +if (err <0 ) { +NL_SET_ERR_MSG (extack, +"Internal error: failed get start time" ); + goto release_mem; + } + + gact->current_close_time = start; + gact->current_gate_status =GATE_ACT_GATE_OPEN |GATE_ACT_PENDING ; + + next = list_first_entry(&p->entries, struct tcfg_gate_entry, list); +rcu_assign_pointer (gact->next_entry, next); + +gate_start_timer (gact, start); + +spin_lock_bh (&g->tcf_lock); + goto_ch =tcf_action_set_ctrlact (*a, parm->action, goto_ch); + gact =rcu_replace_pointer (g->actg, gact, +lockdep_is_held (&g->tcf_lock)); +spin_unlock_bh (&g->tcf_lock); + +if (goto_ch) +tcf_chain_put_by_act (goto_ch); +if (gact) +kfree_rcu (gact, rcu); + +if (ret ==ACT_P_CREATED ) +tcf_idr_insert (tn, *a); +return ret; + +release_mem: +kfree_rcu (gact, rcu); +put_chain: +if (goto_ch) +tcf_chain_put_by_act (goto_ch); +release_idr: +tcf_idr_release (*a, bind); +return err; +} + +static void tcf_gate_cleanup (struct tc_action *a) +{ +struct tcf_gate *g =to_gate (a); +struct tcfg_gate_entry *entry, *n; +struct tcf_gate_params *p; +struct gate_action *gact; + +spin_lock_bh (&g->tcf_lock); + gact =rcu_dereference_protected (g->actg, +lockdep_is_held (&g->tcf_lock)); +hrtimer_cancel (&gact->hitimer); + + p =get_gate_param (gact); +list_for_each_entry_safe (entry, n, &p->entries, list) { +list_del (&entry->list); +kfree (entry); + } +spin_unlock_bh (&g->tcf_lock); + +kfree_rcu (gact, rcu); +} + +static int dumping_entry (struct sk_buff *skb, +struct tcfg_gate_entry *entry) +{ +struct nlattr *item; + + item =nla_nest_start_noflag (skb,TCA_GATE_ONE_ENTRY ); +if (!item) +return -ENOSPC ; + +if (nla_put_u32 (skb,TCA_GATE_ENTRY_INDEX , entry->index)) + goto nla_put_failure; + +if (entry->gate_state &&nla_put_flag (skb,TCA_GATE_ENTRY_GATE )) + goto nla_put_failure; + +if (nla_put_u32 (skb,TCA_GATE_ENTRY_INTERVAL , entry->interval)) + goto nla_put_failure; + +if (nla_put_s32 (skb,TCA_GATE_ENTRY_MAX_OCTETS , entry->maxoctets)) + goto nla_put_failure; + +if (nla_put_s32 (skb,TCA_GATE_ENTRY_IPV , entry->ipv)) + goto nla_put_failure; + +return nla_nest_end (skb, item); + +nla_put_failure: +nla_nest_cancel (skb, item); +return -1 ; +} + +static int tcf_gate_dump (struct sk_buff *skb,struct tc_action *a, +int bind,int ref) +{ +unsigned char *b =skb_tail_pointer (skb); +struct tcf_gate *g =to_gate (a); +struct tc_gate opt = { + .index = g->tcf_index, + .refcnt =refcount_read (&g->tcf_refcnt) - ref, + .bindcnt =atomic_read (&g->tcf_bindcnt) - bind, + }; +struct tcfg_gate_entry *entry; +struct gate_action *gact; +struct tcf_gate_params *p; +struct nlattr *entry_list; +struct tcf_t t; + +spin_lock_bh (&g->tcf_lock); + opt.action = g->tcf_action; + gact =rcu_dereference_protected (g->actg, +lockdep_is_held (&g->tcf_lock)); + + p =get_gate_param (gact); + +if (nla_put (skb,TCA_GATE_PARMS ,sizeof (opt), &opt)) + goto nla_put_failure; + +if (nla_put_u64_64bit (skb,TCA_GATE_BASE_TIME , + p->tcfg_basetime,TCA_GATE_PAD )) + goto nla_put_failure; + +if (nla_put_u64_64bit (skb,TCA_GATE_CYCLE_TIME , + p->tcfg_cycletime,TCA_GATE_PAD )) + goto nla_put_failure; + +if (nla_put_u64_64bit (skb,TCA_GATE_CYCLE_TIME_EXT , + p->tcfg_cycletime_ext,TCA_GATE_PAD )) + goto nla_put_failure; + +if (nla_put_s32 (skb,TCA_GATE_CLOCKID , p->tcfg_clockid)) + goto nla_put_failure; + +if (nla_put_u32 (skb,TCA_GATE_FLAGS , p->tcfg_flags)) + goto nla_put_failure; + +if (nla_put_s32 (skb,TCA_GATE_PRIORITY , p->tcfg_priority)) + goto nla_put_failure; + + entry_list =nla_nest_start_noflag (skb,TCA_GATE_ENTRY_LIST ); +if (!entry_list) + goto nla_put_failure; + +list_for_each_entry (entry, &p->entries, list) { +if (dumping_entry (skb, entry) <0 ) + goto nla_put_failure; + } + +nla_nest_end (skb, entry_list); + +tcf_tm_dump (&t, &g->tcf_tm); +if (nla_put_64bit (skb,TCA_GATE_TM ,sizeof (t), &t,TCA_GATE_PAD )) + goto nla_put_failure; +spin_unlock_bh (&g->tcf_lock); + +return skb->len; + +nla_put_failure: +spin_unlock_bh (&g->tcf_lock); +nlmsg_trim (skb, b); +return -1 ; +} + +static int tcf_gate_walker (struct net *net,struct sk_buff *skb, +struct netlink_callback *cb,int type, +const struct tc_action_ops *ops, +struct netlink_ext_ack *extack) +{ +struct tc_action_net *tn =net_generic (net, gate_net_id); + +return tcf_generic_walker (tn, skb, cb, type, ops, extack); +} + +static void tcf_gate_stats_update (struct tc_action *a,u64 bytes,u64 packets, +u64 lastuse,u64 dropped,bool hw) +{ +struct tcf_gate *g =to_gate (a); +struct tcf_t *tm = &g->tcf_tm; + +tcf_action_update_stats (a, bytes, packets, dropped, hw); + tm->lastuse =max_t (u64, tm->lastuse, lastuse); +} + +static int tcf_gate_search (struct net *net,struct tc_action **a,u32 index) +{ +struct tc_action_net *tn =net_generic (net, gate_net_id); + +return tcf_idr_search (tn, a, index); +} + +static size_t tcf_gate_get_fill_size (const struct tc_action *act) +{ +return nla_total_size (sizeof (struct tc_gate )); +} + +static struct tc_action_ops act_gate_ops = { + .kind ="gate" , + .id =TCA_ID_GATE , + .owner =THIS_MODULE , + .act = tcf_gate_act, + .dump = tcf_gate_dump, + .init = tcf_gate_init, + .cleanup = tcf_gate_cleanup, + .walk = tcf_gate_walker, + .stats_update = tcf_gate_stats_update, + .get_fill_size = tcf_gate_get_fill_size, + .lookup = tcf_gate_search, + .size =sizeof (struct gate_action ), +}; + +static __net_init int gate_init_net (struct net *net) +{ +struct tc_action_net *tn =net_generic (net, gate_net_id); + +return tc_action_net_init (net, tn, &act_gate_ops); +} + +static void __net_exit gate_exit_net(struct list_head *net_list) +{ +tc_action_net_exit (net_list, gate_net_id); +} + +static struct pernet_operations gate_net_ops = { + .init = gate_init_net, + .exit_batch = gate_exit_net, + .id = &gate_net_id, + .size =sizeof (struct tc_action_net ), +}; + +static int __init gate_init_module (void ) +{ +return tcf_register_action (&act_gate_ops, &gate_net_ops); +} + +static void __exit gate_cleanup_module(void) +{ +tcf_unregister_action (&act_gate_ops, &gate_net_ops); +} + +module_init (gate_init_module); +module_exit (gate_cleanup_module); +MODULE_LICENSE ("GPL v2" );Signature · 1 line
2.17.1
Signature · 1 line
Vinicius
RE: [EXT] Re: [RFC,net-next 2/9] net: qos: introduce a gate control flow action
Hi Vinicius,
Quoted 13 lines
-----Original Message-----
From: Vinicius Costa Gomes <[email protected]>
Sent: 2020年3月13日 6:14Quoted 9 lines
Quoted 3 lines
tc qdisc add dev eth0 ingress
tc filter add dev eth0 parent ffff: protocol ip \
flower src_ip 192.168.0.20 \ action gate index 2 \ sched-entry OPEN 200000000 -1 -1 \ sched-entry CLOSE 100000000 -1 -1Quoted 1 line
tc chain del dev eth0 ingress chain 0
Quoted 7 lines
A more complex example, showing how it would work for more filters would be nice.
For example, filters matching 3 different source IPs (even better, using MAC addresses so it would work when you add the '
skip_sw' flag), assigning different priorities to them, also showing how the "base-time" could be used.
Ok, I would update with a more clear example.
Quoted 14 lines
net/sched/act_gate.c line 132
Quoted 11 lines
net/sched/act_gate.c new file +10
-0+static int tcf_gate_act (struct sk_buff *skb,const struct tc_action *a, +struct tcf_result *res) { +struct tcf_gate *g =to_gate (a); +struct tcf_gate_params *p; +struct gate_action *gact; +int action; + +tcf_lastuse_update (&g->tcf_tm); +tcf_action_update_bstats (&g->common, skb); +Please test this with lockdep enabled, I got the feeling that there's a missing
rcu_read_lock()somewhere around here.
I will add lock here.
Quoted 8 lines
net/sched/act_gate.c line 401
Quoted 5 lines
net/sched/act_gate.c new file +4
-0+ gact->tk_offset = tk_offset; +spin_lock_init (&gact->entry_lock); +hrtimer_init (&gact->hitimer, clockid,HRTIMER_MODE_ABS ); + gact->hitimer.function = gate_timer_func;Hm, having an hrtimer per filter seems kind dangerous, in the sense that it seems very hard to configure right.
Per gate action exactly.
Quoted 2 lines
But I don't have any alternative ideas for now.
Yes, hrtimer should only start when software gate simulator binding. But seems there is no way to link the skip_hw state with the gate action in current state. So hrtimer has to start at initial time when it is create.
Quoted 3 lines
--
Vinicius
Br,
Po Liu
Re: [RFC,net-next 2/9] net: qos: introduce a gate control flow action
In reply to Po LiuQuoted 14 lines · Po Liu
net/sched/act_gate.c new file +14
-0+static int parse_gate_list (struct nlattr *list, +struct tcf_gate_params *sched, +struct netlink_ext_ack *extack) +{ +struct nlattr *n; +int err, rem; +int i =0 ; + +if (!list) +return -EINVAL ; + +nla_for_each_nested (n, list, rem) { +struct tcfg_gate_entry *entry; +
net/sched/act_gate.c line 254
Quoted 10 lines
net/sched/act_gate.c new file +10
-0+if (nla_type (n) !=TCA_GATE_ONE_ENTRY ) { +NL_SET_ERR_MSG (extack,"Attribute isn't type 'entry'" ); +continue ; + } + + entry =kzalloc (sizeof (*entry),GFP_KERNEL ); +if (!entry) { +NL_SET_ERR_MSG (extack,"Not enough memory for entry" ); +return -ENOMEM ; + }
You need to free all previous allocated entries on this list when an error happens, right?
Quoted 170 lines
net/sched/act_gate.c new file +170
-0+ + err =parse_gate_entry (n, entry, i, extack); +if (err <0 ) { +kfree (entry); +return err; + } + +list_add_tail (&entry->list, &sched->entries); + i++; + } + + sched->num_entries = i; + +return i; +} + +static int tcf_gate_init (struct net *net,struct nlattr *nla, +struct nlattr *est,struct tc_action **a, +int ovr,int bind,bool rtnl_held, +struct tcf_proto *tp,u32 flags, +struct netlink_ext_ack *extack) +{ +struct tc_action_net *tn =net_generic (net, gate_net_id); +enum tk_offsets tk_offset =TK_OFFS_TAI ; +struct nlattr *tb[TCA_GATE_MAX +1 ]; +struct tcf_chain *goto_ch =NULL ; +struct tcfg_gate_entry *next; +struct tcf_gate_params *p; +struct gate_action *gact; +s32 clockid =CLOCK_TAI ; +struct tc_gate *parm; +struct tcf_gate *g; +int ret =0 , err; +u64 basetime =0 ; +u32 gflags =0 ; +s32 prio =-1 ; +ktime_t start; +u32 index; + +if (!nla) +return -EINVAL ; + + err =nla_parse_nested_deprecated (tb,TCA_GATE_MAX , + nla, gate_policy,NULL ); +if (err <0 ) +return err; + +if (!tb[TCA_GATE_PARMS ]) +return -EINVAL ; + parm =nla_data (tb[TCA_GATE_PARMS ]); + index = parm->index; + err =tcf_idr_check_alloc (tn, &index, a, bind); +if (err <0 ) +return err; + +if (err && bind) +return 0 ; + +if (!err) { + ret =tcf_idr_create_from_flags (tn, index, est, a, + &act_gate_ops, bind, flags); +if (ret) { +tcf_idr_cleanup (tn, index); +return ret; + } + + ret =ACT_P_CREATED ; + }else if (!ovr) { +tcf_idr_release (*a, bind); +return -EEXIST ; + } + +if (tb[TCA_GATE_PRIORITY ]) + prio =nla_get_s32 (tb[TCA_GATE_PRIORITY ]); + +if (tb[TCA_GATE_BASE_TIME ]) + basetime =nla_get_u64 (tb[TCA_GATE_BASE_TIME ]); + +if (tb[TCA_GATE_FLAGS ]) + gflags =nla_get_u32 (tb[TCA_GATE_FLAGS ]); + +if (tb[TCA_GATE_CLOCKID ]) { + clockid =nla_get_s32 (tb[TCA_GATE_CLOCKID ]); +switch (clockid) { +case CLOCK_REALTIME : + tk_offset =TK_OFFS_REAL ; +break ; +case CLOCK_MONOTONIC : + tk_offset =TK_OFFS_MAX ; +break ; +case CLOCK_BOOTTIME : + tk_offset =TK_OFFS_BOOT ; +break ; +case CLOCK_TAI : + tk_offset =TK_OFFS_TAI ; +break ; +default : +NL_SET_ERR_MSG (extack,"Invalid 'clockid'" ); + goto release_idr; + } + } + + err =tcf_action_check_ctrlact (parm->action, tp, &goto_ch, extack); +if (err <0 ) + goto release_idr; + + g =to_gate (*a); + + gact =kzalloc (sizeof (*gact),GFP_KERNEL ); +if (!gact) { + err = -ENOMEM ; + goto put_chain; + } + + p =get_gate_param (gact); + +INIT_LIST_HEAD (&p->entries); +if (tb[TCA_GATE_ENTRY_LIST ]) { + err =parse_gate_list (tb[TCA_GATE_ENTRY_LIST ], p, extack); +if (err <0 ) + goto release_mem; + } + +if (tb[TCA_GATE_CYCLE_TIME ]) { + p->tcfg_cycletime =nla_get_u64 (tb[TCA_GATE_CYCLE_TIME ]); + }else { +struct tcfg_gate_entry *entry; +ktime_t cycle =0 ; + +list_for_each_entry (entry, &p->entries, list) + cycle =ktime_add_ns (cycle, entry->interval); + p->tcfg_cycletime = cycle; + } + +if (tb[TCA_GATE_CYCLE_TIME_EXT ]) + p->tcfg_cycletime_ext = +nla_get_u64 (tb[TCA_GATE_CYCLE_TIME_EXT ]); + + p->tcfg_priority = prio; + p->tcfg_basetime = basetime; + p->tcfg_clockid = clockid; + p->tcfg_flags = gflags; + + gact->tk_offset = tk_offset; +spin_lock_init (&gact->entry_lock); +hrtimer_init (&gact->hitimer, clockid,HRTIMER_MODE_ABS ); + gact->hitimer.function = gate_timer_func; + + err =gate_get_start_time (gact, &start); +if (err <0 ) { +NL_SET_ERR_MSG (extack, +"Internal error: failed get start time" ); + goto release_mem; + } + + gact->current_close_time = start; + gact->current_gate_status =GATE_ACT_GATE_OPEN |GATE_ACT_PENDING ; + + next = list_first_entry(&p->entries, struct tcfg_gate_entry, list); +rcu_assign_pointer (gact->next_entry, next); + +gate_start_timer (gact, start); + +spin_lock_bh (&g->tcf_lock); + goto_ch =tcf_action_set_ctrlact (*a, parm->action, goto_ch); + gact =rcu_replace_pointer (g->actg, gact, +lockdep_is_held (&g->tcf_lock)); +spin_unlock_bh (&g->tcf_lock); + +if (goto_ch)
net/sched/act_gate.c line 434
Quoted 10 lines
net/sched/act_gate.c new file +10
-0+tcf_chain_put_by_act (goto_ch); +if (gact) +kfree_rcu (gact, rcu); + +if (ret ==ACT_P_CREATED ) +tcf_idr_insert (tn, *a); +return ret; + +release_mem: +kfree_rcu (gact, rcu);
No need to bother RCU for an error path, right?
Thanks.
RE: [EXT] Re: [RFC,net-next 2/9] net: qos: introduce a gate control flow action
Hi Cong,
Quoted 53 lines
-----Original Message-----
From: Cong Wang <[email protected]>
Sent: 2020年3月14日 2:37
To: Po Liu <[email protected]>
Cc: David Miller <[email protected]>; LKML <linux- [email protected]>; Linux Kernel Network Developers <[email protected]>; Vinicius Costa Gomes <[email protected]>; Claudiu Manoil <[email protected]>; Vladimir Oltean <[email protected]>; Alexandru Marginean <[email protected]>; Xiaoliang Yang <[email protected]>; Roy Zang <[email protected]>; Mingkai Hu <[email protected]>; Jerry Huang <[email protected]>; Leo Li <[email protected]>; [email protected]; [email protected]; Saeed Mahameed <[email protected]>; Leon Romanovsky <[email protected]>; Jiri Pirko <[email protected]>; Ido Schimmel <[email protected]>; [email protected]; [email protected]; Jakub Kicinski <[email protected]>; Jamal Hadi Salim <[email protected]>; John Hurley <[email protected]>; Simon Horman <[email protected]>; Pieter Jansen van Vuuren <[email protected]>; Pablo Neira Ayuso <[email protected]>; [email protected]; [email protected]; Murali Karicheri <[email protected]>; [email protected]; Jakub Kicinski <[email protected]>
Subject: [EXT] Re: [RFC,net-next 2/9] net: qos: introduce a gate control flow actionCaution: EXT Email
Quoted 11 lines · Po Liu
net/sched/act_gate.c new file +11
-0+static int parse_gate_list (struct nlattr *list, +struct tcf_gate_params *sched, +struct netlink_ext_ack *extack) { +struct nlattr *n; +int err, rem; +int i =0 ; + +if (!list) +return -EINVAL ; + +nla_for_each_nested (n, list, rem) {net/sched/act_gate.c
Quoted 10 lines
net/sched/act_gate.c new file +10
-0+struct tcfg_gate_entry *entry; + +if (nla_type (n) !=TCA_GATE_ONE_ENTRY ) { +NL_SET_ERR_MSG (extack,"Attribute isn't type 'entry'" ); +continue ; + } + + entry =kzalloc (sizeof (*entry),GFP_KERNEL ); +if (!entry) { + NL_SET_ERR_MSG(extack, "Not enough memory forentry");
Quoted 5 lines
Review
Quoted 2 lines
(quoted diff) +2
-0+ return -ENOMEM; + }You need to free all previous allocated entries on this list when an error happens, right?
Yes, just as Vinicius has mentioned. I'll fix it next version.
Quoted 176 lines
Quoted 176 lines
net/sched/act_gate.c new file +174
-0+ + err =parse_gate_entry (n, entry, i, extack); +if (err <0 ) { +kfree (entry); +return err; + } + +list_add_tail (&entry->list, &sched->entries); + i++; + } + + sched->num_entries = i; + +return i; +} + +static int tcf_gate_init (struct net *net,struct nlattr *nla, +struct nlattr *est,struct tc_action **a, +int ovr,int bind,bool rtnl_held, +struct tcf_proto *tp,u32 flags, +struct netlink_ext_ack *extack) { +struct tc_action_net *tn =net_generic (net, gate_net_id); +enum tk_offsets tk_offset =TK_OFFS_TAI ; +struct nlattr *tb[TCA_GATE_MAX +1 ]; +struct tcf_chain *goto_ch =NULL ; +struct tcfg_gate_entry *next; +struct tcf_gate_params *p; +struct gate_action *gact; +s32 clockid =CLOCK_TAI ; +struct tc_gate *parm; +struct tcf_gate *g; +int ret =0 , err; +u64 basetime =0 ; +u32 gflags =0 ; +s32 prio =-1 ; +ktime_t start; +u32 index; + +if (!nla) +return -EINVAL ; + + err =nla_parse_nested_deprecated (tb,TCA_GATE_MAX , + nla, gate_policy,NULL ); +if (err <0 ) +return err; + +if (!tb[TCA_GATE_PARMS ]) +return -EINVAL ; + parm =nla_data (tb[TCA_GATE_PARMS ]); + index = parm->index; + err =tcf_idr_check_alloc (tn, &index, a, bind); +if (err <0 ) +return err; + +if (err && bind) +return 0 ; + +if (!err) { + ret =tcf_idr_create_from_flags (tn, index, est, a, + &act_gate_ops, bind, flags); +if (ret) { +tcf_idr_cleanup (tn, index); +return ret; + } + + ret =ACT_P_CREATED ; + }else if (!ovr) { +tcf_idr_release (*a, bind); +return -EEXIST ; + } + +if (tb[TCA_GATE_PRIORITY ]) + prio =nla_get_s32 (tb[TCA_GATE_PRIORITY ]); + +if (tb[TCA_GATE_BASE_TIME ]) + basetime =nla_get_u64 (tb[TCA_GATE_BASE_TIME ]); + +if (tb[TCA_GATE_FLAGS ]) + gflags =nla_get_u32 (tb[TCA_GATE_FLAGS ]); + +if (tb[TCA_GATE_CLOCKID ]) { + clockid =nla_get_s32 (tb[TCA_GATE_CLOCKID ]); +switch (clockid) { +case CLOCK_REALTIME : + tk_offset =TK_OFFS_REAL ; +break ; +case CLOCK_MONOTONIC : + tk_offset =TK_OFFS_MAX ; +break ; +case CLOCK_BOOTTIME : + tk_offset =TK_OFFS_BOOT ; +break ; +case CLOCK_TAI : + tk_offset =TK_OFFS_TAI ; +break ; +default : +NL_SET_ERR_MSG (extack,"Invalid 'clockid'" ); + goto release_idr; + } + } + + err =tcf_action_check_ctrlact (parm->action, tp, &goto_ch, extack); +if (err <0 ) + goto release_idr; + + g =to_gate (*a); + + gact =kzalloc (sizeof (*gact),GFP_KERNEL ); +if (!gact) { + err = -ENOMEM ; + goto put_chain; + } + + p =get_gate_param (gact); + +INIT_LIST_HEAD (&p->entries); +if (tb[TCA_GATE_ENTRY_LIST ]) { + err =parse_gate_list (tb[TCA_GATE_ENTRY_LIST ], p, extack); +if (err <0 ) + goto release_mem; + } + +if (tb[TCA_GATE_CYCLE_TIME ]) { + p->tcfg_cycletime =nla_get_u64 (tb[TCA_GATE_CYCLE_TIME ]); + }else { +struct tcfg_gate_entry *entry; +ktime_t cycle =0 ; + +list_for_each_entry (entry, &p->entries, list) + cycle =ktime_add_ns (cycle, entry->interval); + p->tcfg_cycletime = cycle; + } + +if (tb[TCA_GATE_CYCLE_TIME_EXT ]) + p->tcfg_cycletime_ext = +nla_get_u64 (tb[TCA_GATE_CYCLE_TIME_EXT ]); + + p->tcfg_priority = prio; + p->tcfg_basetime = basetime; + p->tcfg_clockid = clockid; + p->tcfg_flags = gflags; + + gact->tk_offset = tk_offset; +spin_lock_init (&gact->entry_lock); +hrtimer_init (&gact->hitimer, clockid,HRTIMER_MODE_ABS ); + gact->hitimer.function = gate_timer_func; + + err =gate_get_start_time (gact, &start); +if (err <0 ) { +NL_SET_ERR_MSG (extack, +"Internal error: failed get start time" ); + goto release_mem; + } + + gact->current_close_time = start; + gact->current_gate_status =GATE_ACT_GATE_OPEN | +GATE_ACT_PENDING ; + + next = list_first_entry(&p->entries, struct tcfg_gate_entry, list); +rcu_assign_pointer (gact->next_entry, next); + +gate_start_timer (gact, start); + +spin_lock_bh (&g->tcf_lock); + goto_ch =tcf_action_set_ctrlact (*a, parm->action, goto_ch); + gact =rcu_replace_pointer (g->actg, gact, +lockdep_is_held (&g->tcf_lock)); +spin_unlock_bh (&g->tcf_lock); + +if (goto_ch) +tcf_chain_put_by_act (goto_ch); +if (gact) +kfree_rcu (gact, rcu); +
Quoted 8 lines
net/sched/act_gate.c line 434
Quoted 6 lines
net/sched/act_gate.c new file +6
-0+if (ret ==ACT_P_CREATED ) +tcf_idr_insert (tn, *a); +return ret; + +release_mem: +kfree_rcu (gact, rcu);No need to bother RCU for an error path, right?
Yes, the kfree_rcu() change to kfree() is proper.
Quoted 2 lines
Thanks.
Thanks!
Br,
Po Liu
RFC 3/9 net: schedule: add action gate offloading
In reply to Po LiuAdd the gate action to the flow action entry. Add the gate parameters to the tc_setup_flow_action() queueing to the entries of flow_action_entry array provide to the driver.
--- include/net/flow_offload.h | 10 +++ include/net/tc_act/tc_gate.h | 115 +++++++++++++++++++++++++++++++++++ net/sched/cls_api.c | 33 ++++++++++ 3 files changed, 158 insertions(+)
include/net/flow_offload.h +10 -0
@@ -138,6 +138,7 @@ enum flow_action_id {FLOW_ACTION_MPLS_PUSH ,FLOW_ACTION_MPLS_POP ,FLOW_ACTION_MPLS_MANGLE , +FLOW_ACTION_GATE , NUM_FLOW_ACTIONS, };
@@ -223,6 +224,15 @@ struct flow_action_entry {u8 bos;u8 ttl; } mpls_mangle; +struct { +u32 index; +s32 prio; +u64 basetime; +u64 cycletime; +u64 cycletimeext; +u32 num_entries; +struct action_gate_entry *entries; + } gate; };struct flow_action_cookie *cookie;/* user defined action cookie */ };
include/net/tc_act/tc_gate.h +115 -0
125 lines not on this page · raw message
net/sched/cls_api.c +33 -0
@@ -38,6 +38,7 @@#include <net/tc_act/tc_skbedit.h> #include <net/tc_act/tc_ct.h> #include <net/tc_act/tc_mpls.h> +#include <net/tc_act/tc_gate.h> #include <net/flow_offload.h> extern const struct nla_policy rtm_tca_policy[TCA_MAX +1 ];
@@ -3458,6 +3459,27 @@ static void tcf_sample_get_group(struct flow_action_entry *entry,#endif } +static void tcf_gate_entry_destructor (void *priv) +{ +struct action_gate_entry *oe = priv; + +kfree (oe); +} + +static int tcf_gate_get_entries (struct flow_action_entry *entry, +const struct tc_action *act) +{ + entry->gate.entries =tcf_gate_get_list (act); + +if (!entry->gate.entries) +return -EINVAL ; + + entry->destructor = tcf_gate_entry_destructor; + entry->destructor_priv = entry->gate.entries; + +return 0 ; +} +int tc_setup_flow_action (struct flow_action *flow_action,const struct tcf_exts *exts) {
@@ -3592,6 +3614,17 @@ int tc_setup_flow_action(struct flow_action *flow_action, }else if (is_tcf_skbedit_ptype (act)) { entry->id =FLOW_ACTION_PTYPE ; entry->ptype =tcf_skbedit_ptype (act); + }else if (is_tcf_gate (act)) { + entry->id =FLOW_ACTION_GATE ; + entry->gate.index =tcf_gate_index (act); + entry->gate.prio =tcf_gate_prio (act); + entry->gate.basetime =tcf_gate_basetime (act); + entry->gate.cycletime =tcf_gate_cycletime (act); + entry->gate.cycletimeext =tcf_gate_cycletimeext (act); + entry->gate.num_entries =tcf_gate_num_entries (act); + err =tcf_gate_get_entries (entry, act); +if (err) + goto err_out; }else { err = -EOPNOTSUPP ; goto err_out_locked;
Signature · 1 line
2.17.1
Jakub Kicinski
maintainerRe: [RFC,net-next 3/9] net: schedule: add action gate offloading
net/sched/cls_api.c static void tcf_sample_get_group(struct flow_action_entry *entry, · line 3478
Quoted 10 lines · On Fri, 6 Mar 2020 20:56:01 +0800 Po Liu wrote:
net/sched/cls_api.c +10
-0+static int tcf_gate_get_entries (struct flow_action_entry *entry, +const struct tc_action *act) +{ + entry->gate.entries =tcf_gate_get_list (act); + +if (!entry->gate.entries) +return -EINVAL ; + + entry->destructor = tcf_gate_entry_destructor; + entry->destructor_priv = entry->gate.entries;
What's this destructor stuff doing? I don't it being called.
Quoted 2 lines
net/sched/cls_api.c +2
-0+return 0 ; +}
Jakub Kicinski
maintainerRe: [RFC,net-next 3/9] net: schedule: add action gate offloading
Quoted 13 lines · On Fri, 6 Mar 2020 11:02:00 -0800 Jakub Kicinski wrote:
net/sched/cls_api.c static void tcf_sample_get_group(struct flow_action_entry *entry, · line 3478
Quoted 10 lines · On Fri, 6 Mar 2020 20:56:01 +0800 Po Liu wrote:
net/sched/cls_api.c +10
-0+static int tcf_gate_get_entries (struct flow_action_entry *entry, +const struct tc_action *act) +{ + entry->gate.entries =tcf_gate_get_list (act); + +if (!entry->gate.entries) +return -EINVAL ; + + entry->destructor = tcf_gate_entry_destructor; + entry->destructor_priv = entry->gate.entries;What's this destructor stuff doing? I don't it being called.
Ah, it's the action destructor, not something gate specific. Disregard.
RE: [EXT] Re: [RFC,net-next 3/9] net: schedule: add action gate offloading
Hi Jakub,
Quoted 22 lines
-----Original Message-----
From: Jakub Kicinski <[email protected]>
Sent: 2020年3月7日 3:19
To: Po Liu <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; Claudiu Manoil <[email protected]>; Vladimir Oltean <[email protected]>; Alexandru Marginean <[email protected]>; Xiaoliang Yang <[email protected]>; Roy Zang <[email protected]>; Mingkai Hu <[email protected]>; Jerry Huang <[email protected]>; Leo Li <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]
Subject: [EXT] Re: [RFC,net-next 3/9] net: schedule: add action gate offloading
Quoted 15 lines
Caution: EXT Email
Quoted 12 lines · On Fri, 6 Mar 2020 11:02:00 -0800 Jakub Kicinski wrote:
net/sched/cls_api.c static void tcf_sample_get_group(struct flow_action_entry *entry, · line 3478
Quoted 9 lines · On Fri, 6 Mar 2020 20:56:01 +0800 Po Liu wrote:
net/sched/cls_api.c +9
-0+static int tcf_gate_get_entries (struct flow_action_entry *entry, +const struct tc_action *act) { + entry->gate.entries =tcf_gate_get_list (act); + +if (!entry->gate.entries) +return -EINVAL ; + + entry->destructor = tcf_gate_entry_destructor; + entry->destructor_priv = entry->gate.entries;What's this destructor stuff doing? I don't it being called.
It prepare a gate list array parameters for offloading. So the driver side would not link the data with protocol side. Destructor would free the temporary gate list array.
Quoted 2 lines
Ah, it's the action destructor, not something gate specific. Disregard.
I understand here with actions are:
Each tc flower follow up with actions. Each action defined:
struct flow_action_entry {
enum flow_action_id id;
action_destr destructor;
void *destructor_priv;
union {
......
{}sample,
{}police,
{}gate,
}
}So for the destructor and destructor_priv are provided specific for the union action. So it is not gate specific. For mirror action, destructor_priv would be point to a mirror device data. I suppose it is for destroy the temporary data like it name. And after tc_setup_flow_action() loaded, the destructor function would be loaded by tc_cleanup_flow_action() to destroy and free the temporary data.
Code flow is :net/sched/cls_flower.c
static int fl_hw_replace_filter()
{
......
tc_setup_flow_action(); ---------------------------------------> assign action parameters (with the destructor and destructor_priv if the action needed)
......
tc_setup_cb_add() ----------------------------------------------> call the driver provide rules with actions datas for device
......
tc_cleanup_flow_action(&cls_flower.rule->action); ---> loading each action''s destructor(destructor_priv)
}So for each action would be with its private destructor and destructor_priv if the action needed, and then destroyed at tc_cleanup_flow_action().
Did I misunderstand anything?
Thanks!
Br,
Po Liu
RFC 4/9 net: enetc: add hw tc hw offload features for PSPF capability
In reply to Po LiuThis patch is to let ethtool enable/disable the tc flower offload features. Hardware enetc has the feature of PSFP which is for per-stream policing. When enable the tc hw offloading feature, driver would enable the IEEE 802.1Qci feature. It is only set the register enable bit for this feature not enable for any entry of per stream filtering and stream gate or stream identify but get how much capabilities for each features.
--- drivers/net/ethernet/freescale/enetc/enetc.c | 23 ++++++++ drivers/net/ethernet/freescale/enetc/enetc.h | 55 +++++++++++++++++++ .../net/ethernet/freescale/enetc/enetc_hw.h | 17 ++++++ .../net/ethernet/freescale/enetc/enetc_pf.c | 8 +++ 4 files changed, 103 insertions(+)
drivers/net/ethernet/freescale/enetc/enetc.c +23 -0
@@ -762,6 +762,9 @@ void enetc_get_si_caps(struct enetc_si *si)if (val &ENETC_SIPCAPR0_QBV ) si->hw_features |=ENETC_SI_F_QBV ; + +if (val &ENETC_SIPCAPR0_PSFP ) + si->hw_features |=ENETC_SI_F_PSFP ; }static int enetc_dma_alloc_bdr (struct enetc_bdr *r,size_t bd_size)
@@ -1566,6 +1569,23 @@ static int enetc_set_rss(struct net_device *ndev, int en)return 0 ; } +static int enetc_set_psfp (struct net_device *ndev,int en) +{ +struct enetc_ndev_priv *priv =netdev_priv (ndev); + +if (en) { + priv->active_offloads |=ENETC_F_QCI ; +enetc_get_max_cap (priv); +enetc_psfp_enable (&priv->si->hw); + }else { + priv->active_offloads &= ~ENETC_F_QCI ; +memset (&priv->psfp_cap,0 ,sizeof (struct psfp_cap )); +enetc_psfp_disable (&priv->si->hw); + } + +return 0 ; +} +int enetc_set_features (struct net_device *ndev,netdev_features_t features) {
@@ -1574,6 +1594,9 @@ int enetc_set_features(struct net_device *ndev,if (changed &NETIF_F_RXHASH )enetc_set_rss (ndev, !!(features &NETIF_F_RXHASH )); +if (changed &NETIF_F_HW_TC ) +enetc_set_psfp (ndev, !!(features &NETIF_F_HW_TC )); +return 0 ; }
drivers/net/ethernet/freescale/enetc/enetc.h +55 -0
88 lines not on this page · raw message
drivers/net/ethernet/freescale/enetc/enetc_hw.h +17 -0
@@ -19,6 +19,7 @@#define ENETC_SICTR1 0x1c#define ENETC_SIPCAPR0 0x20#define ENETC_SIPCAPR0_QBV BIT(4) +#define ENETC_SIPCAPR0_PSFP BIT(9)#define ENETC_SIPCAPR0_RSS BIT(8)#define ENETC_SIPCAPR1 0x24#define ENETC_SITGTGR 0x30
@@ -228,6 +229,15 @@ enum enetc_bdr_type {TX, RX};#define ENETC_PM0_IFM_RLP (BIT(5) | BIT(11))#define ENETC_PM0_IFM_RGAUTO (BIT(15) | ENETC_PMO_IFM_RG | BIT(1))#define ENETC_PM0_IFM_XGMII BIT(12) +#define ENETC_PSIDCAPR 0x1b08 +#define ENETC_PSIDCAPR_MSK GENMASK(15, 0) +#define ENETC_PSFCAPR 0x1b18 +#define ENETC_PSFCAPR_MSK GENMASK(15, 0) +#define ENETC_PSGCAPR 0x1b28 +#define ENETC_PSGCAPR_GCL_MSK GENMASK(18, 16) +#define ENETC_PSGCAPR_SGIT_MSK GENMASK(15, 0) +#define ENETC_PFMCAPR 0x1b38 +#define ENETC_PFMCAPR_MSK GENMASK(15, 0)/* MAC counters */ #define ENETC_PM0_REOCT 0x8100
@@ -624,3 +634,10 @@ struct enetc_cbd {/* Port time specific departure */ #define ENETC_PTCTSDR (n) (0x1210 + 4 * (n))#define ENETC_TSDE BIT(31) + +/* PSFP setting */ +#define ENETC_PPSFPMR 0x11b00 +#define ENETC_PPSFPMR_PSFPEN BIT(0) +#define ENETC_PPSFPMR_VS BIT(1) +#define ENETC_PPSFPMR_PVC BIT(2) +#define ENETC_PPSFPMR_PVZC BIT(3)
drivers/net/ethernet/freescale/enetc/enetc_pf.c +8 -0
@@ -740,6 +740,14 @@ static void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,if (si->hw_features &ENETC_SI_F_QBV ) priv->active_offloads |=ENETC_F_QBV ; +if (si->hw_features &ENETC_SI_F_PSFP ) { + priv->active_offloads |=ENETC_F_QCI ; + ndev->features |=NETIF_F_HW_TC ; + ndev->hw_features |=NETIF_F_HW_TC ; +enetc_get_max_cap (priv); +enetc_psfp_enable (&si->hw); + } +/* pick up primary MAC address from SI */ enetc_get_primary_mac_addr (&si->hw, ndev->dev_addr); }
Signature · 1 line
2.17.1
RFC 5/9 net: enetc: add tc flower psfp offload driver
In reply to Po LiuThis patch is to add tc flower offload for the enetc IEEE 802.1Qci(PSFP) function. There are four main feature parts to implement the flow policing and filtering for ingress flow with IEEE 802.1Qci features. They are stream identify(this is defined in the P802.1cb exactly but needed for 802.1Qci), stream filtering, stream gate and flow metering. Each function block includes many entries by index to assign parameters. So for one frame would be filtered by stream identify first, then flow into stream filter block by the same handle between stream identify and stream filtering. Then flow into stream gate control which assigned by the stream filtering entry. And then policing by the gate and limited by the max sdu in the filter block(optional). At last, policing by the flow metering block, index choosing at the fitering block. So you can see that each entry of block may link to many upper entries since they can be assigned same index means more streams want to share the same feature in the stream filtering or stream gate or flow metering.
To implement such features, each stream filtered by source/destination mac address, some stream maybe also plus the vlan id value would be treated as one flow chain. This would be identified by the chain_index which already in the tc filter concept. Driver would maintain this chain and also with gate modules. The stream filter entry create by the gate index and flow meter(optional) entry id and also one priority value. Offloading only transfer the gate action and flow filtering parameters. Driver would create (or search same gate id and flow meter id and
priority) one stream filter entry to set to the hardware. So stream
filtering do not need transfer by the action offloading. This architecture is same with tc filter and actions relationship. tc filter maintain the list for each flow feature by keys. And actions maintain by the action list.
Below showing a example commands by tc:
Quoted 3 lines
tc qdisc add dev eth0 ingress
ip link set eth0 address 10:00:80:00:00:00
tc filter add dev eth0 parent ffff: protocol ip chain 11 \
flower skip_sw dst_mac 10:00:80:00:00:00 \ action gate index 10 \ sched-entry OPEN 200000000 -1 -1 \ sched-entry CLOSE 100000000 -1 -1
(or just create one gate entry list to keep close)
That means to set the dst_mac 10:00:80:00:00:00 to index 11 of stream identify module. And the set the gate index 10 of stream gate module.
--- drivers/net/ethernet/freescale/enetc/enetc.c | 25 +- drivers/net/ethernet/freescale/enetc/enetc.h | 39 +- .../net/ethernet/freescale/enetc/enetc_hw.h | 142 +++ .../net/ethernet/freescale/enetc/enetc_pf.c | 4 +- .../net/ethernet/freescale/enetc/enetc_qos.c | 1078 ++++++++++++++++- 5 files changed, 1269 insertions(+), 19 deletions(-)
drivers/net/ethernet/freescale/enetc/enetc.c +17 -8
@@ -1520,6 +1520,8 @@ int enetc_setup_tc(struct net_device *ndev, enum tc_setup_type type,return enetc_setup_tc_cbs (ndev, type_data);case TC_SETUP_QDISC_ETF :return enetc_setup_tc_txtime (ndev, type_data); +case TC_SETUP_BLOCK : +return enetc_setup_tc_psfp (ndev, type_data);default :return -EOPNOTSUPP ; }
@@ -1572,17 +1574,23 @@ static int enetc_set_rss(struct net_device *ndev, int en)static int enetc_set_psfp (struct net_device *ndev,int en) {struct enetc_ndev_priv *priv =netdev_priv (ndev); +int err;if (en) { + err =enetc_psfp_enable (priv); +if (err) +return err; + priv->active_offloads |=ENETC_F_QCI ; -enetc_get_max_cap (priv); -enetc_psfp_enable (&priv->si->hw); - }else { - priv->active_offloads &= ~ENETC_F_QCI ; -memset (&priv->psfp_cap,0 ,sizeof (struct psfp_cap )); -enetc_psfp_disable (&priv->si->hw); +return 0 ; } + err =enetc_psfp_disable (priv); +if (err) +return err; + + priv->active_offloads &= ~ENETC_F_QCI ; +return 0 ; }
@@ -1590,14 +1598,15 @@ int enetc_set_features(struct net_device *ndev, netdev_features_t features) {netdev_features_t changed = ndev->features ^ features; +int err =0 ;if (changed &NETIF_F_RXHASH )enetc_set_rss (ndev, !!(features &NETIF_F_RXHASH ));if (changed &NETIF_F_HW_TC ) -enetc_set_psfp (ndev, !!(features &NETIF_F_HW_TC )); + err =enetc_set_psfp (ndev, !!(features &NETIF_F_HW_TC )); -return 0 ; +return err; }#ifdef CONFIG_FSL_ENETC_HW_TIMESTAMPING
drivers/net/ethernet/freescale/enetc/enetc.h +35 -4
@@ -269,6 +269,11 @@ int enetc_setup_tc_taprio(struct net_device *ndev, void *type_data);void enetc_sched_speed_set (struct net_device *ndev);int enetc_setup_tc_cbs (struct net_device *ndev,void *type_data);int enetc_setup_tc_txtime (struct net_device *ndev,void *type_data); +int enetc_setup_tc_block_cb (enum tc_setup_type type,void *type_data, +void *cb_priv); +int enetc_setup_tc_psfp (struct net_device *ndev,void *type_data); +int enetc_psfp_init (struct enetc_ndev_priv *priv); +int enetc_psfp_clean (struct enetc_ndev_priv *priv);static inline void enetc_get_max_cap (struct enetc_ndev_priv *priv) {
@@ -288,33 +293,59 @@ static inline void enetc_get_max_cap(struct enetc_ndev_priv *priv) priv->psfp_cap.max_psfp_meter = reg &ENETC_PFMCAPR_MSK ; } -static inline void enetc_psfp_enable (struct enetc_hw *hw) +static inline int enetc_psfp_enable (struct enetc_ndev_priv *priv) { +struct enetc_hw *hw = &priv->si->hw; +int err; + +enetc_get_max_cap (priv); + + err =enetc_psfp_init (priv); +if (err) +return err; +enetc_wr (hw,ENETC_PPSFPMR ,enetc_rd (hw,ENETC_PPSFPMR ) |ENETC_PPSFPMR_PSFPEN |ENETC_PPSFPMR_VS |ENETC_PPSFPMR_PVC |ENETC_PPSFPMR_PVZC ); + +return 0 ; } -static inline void enetc_psfp_disable (struct enetc_hw *hw) +static inline int enetc_psfp_disable (struct enetc_ndev_priv *priv) { +struct enetc_hw *hw = &priv->si->hw; +int err; + + err =enetc_psfp_clean (priv); +if (err) +return err; +enetc_wr (hw,ENETC_PPSFPMR ,enetc_rd (hw,ENETC_PPSFPMR ) & ~ENETC_PPSFPMR_PSFPEN & ~ENETC_PPSFPMR_VS & ~ENETC_PPSFPMR_PVC & ~ENETC_PPSFPMR_PVZC ); + +memset (&priv->psfp_cap,0 ,sizeof (struct psfp_cap )); + +return 0 ; } +#else #define enetc_setup_tc_taprio (ndev, type_data) -EOPNOTSUPP#define enetc_sched_speed_set (ndev) (void)0#define enetc_setup_tc_cbs (ndev, type_data) -EOPNOTSUPP#define enetc_setup_tc_txtime (ndev, type_data) -EOPNOTSUPP +#define enetc_setup_tc_psfp (ndev, type_data) -EOPNOTSUPP +#define enetc_setup_tc_block_cb NULL +#define enetc_get_max_cap (p) \ memset(&((p)->psfp_cap), 0, sizeof(struct psfp_cap)) -static inline void enetc_psfp_enable (struct enetc_hw *hw) +static inline int enetc_psfp_enable (struct enetc_hw *hw) {return 0 ; } -static inline void enetc_psfp_disable (struct enetc_hw *hw) +static inline int enetc_psfp_disable (struct enetc_hw *hw) {return 0 ; }
drivers/net/ethernet/freescale/enetc/enetc_hw.h +142 -0
161 lines not on this page · raw message
drivers/net/ethernet/freescale/enetc/enetc_pf.c +1 -3
@@ -740,12 +740,10 @@ static void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,if (si->hw_features &ENETC_SI_F_QBV ) priv->active_offloads |=ENETC_F_QBV ; -if (si->hw_features &ENETC_SI_F_PSFP ) { +if (si->hw_features &ENETC_SI_F_PSFP && !enetc_psfp_enable (priv)) { priv->active_offloads |=ENETC_F_QCI ; ndev->features |=NETIF_F_HW_TC ; ndev->hw_features |=NETIF_F_HW_TC ; -enetc_get_max_cap (priv); -enetc_psfp_enable (&si->hw); }/* pick up primary MAC address from SI */
drivers/net/ethernet/freescale/enetc/enetc_qos.c +1074 -4
1,096 lines not on this page · raw message
Signature · 1 line
2.17.1
Re: [RFC,net-next 5/9] net: enetc: add tc flower psfp offload driver
Quoted 338 lines · Po Liu
This patch is to add tc flower offload for the enetc IEEE 802.1Qci(PSFP) function. There are four main feature parts to implement the flow policing and filtering for ingress flow with IEEE 802.1Qci features. They are stream identify(this is defined in the P802.1cb exactly but needed for 802.1Qci), stream filtering, stream gate and flow metering. Each function block includes many entries by index to assign parameters. So for one frame would be filtered by stream identify first, then flow into stream filter block by the same handle between stream identify and stream filtering. Then flow into stream gate control which assigned by the stream filtering entry. And then policing by the gate and limited by the max sdu in the filter block(optional). At last, policing by the flow metering block, index choosing at the fitering block. So you can see that each entry of block may link to many upper entries since they can be assigned same index means more streams want to share the same feature in the stream filtering or stream gate or flow metering.
To implement such features, each stream filtered by source/destination mac address, some stream maybe also plus the vlan id value would be treated as one flow chain. This would be identified by thechain_indexwhich already in the tc filter concept. Driver would maintain this chain and also with gate modules. The stream filter entry create by the gate index and flow meter(optional) entry id and also one priority value. Offloading only transfer the gate action and flow filtering parameters. Driver would create (or search same gate id and flow meter id and
priority) one stream filter entry to set to the hardware. So stream
filtering do not need transfer by the action offloading. This architecture is same with tc filter and actions relationship. tc filter maintain the list for each flow feature by keys. And actions maintain by the action list.Below showing a example commands by tc:
Quoted 3 lines
tc qdisc add dev eth0 ingress
ip link set eth0 address 10:00:80:00:00:00
tc filter add dev eth0 parent ffff: protocol ip chain 11 \flower skip_sw dst_mac 10:00:80:00:00:00 \ action gate index 10 \ sched-entry OPEN 200000000 -1 -1 \ sched-entry CLOSE 100000000 -1 -1(or just create one gate entry list to keep close)
That means to set the
dst_mac10:00:80:00:00:00 to index 11 of stream identify module. And the set the gate index 10 of stream gate module.Signed-off-by Po Liu <[email protected]>--- drivers/net/ethernet/freescale/enetc/enetc.c | 25 +- drivers/net/ethernet/freescale/enetc/enetc.h | 39 +- .../net/ethernet/freescale/enetc/enetc_hw.h | 142 +++ .../net/ethernet/freescale/enetc/enetc_pf.c | 4 +- .../net/ethernet/freescale/enetc/enetc_qos.c | 1078 ++++++++++++++++- 5 files changed, 1269 insertions(+), 19 deletions(-)drivers/net/ethernet/freescale/enetc/enetc.c +17
-8@@ -1520,6 +1520,8 @@ int enetc_setup_tc(struct net_device *ndev, enum tc_setup_type type,return enetc_setup_tc_cbs (ndev, type_data);case TC_SETUP_QDISC_ETF :return enetc_setup_tc_txtime (ndev, type_data); +case TC_SETUP_BLOCK : +return enetc_setup_tc_psfp (ndev, type_data);default :return -EOPNOTSUPP ; }@@ -1572,17 +1574,23 @@ static int enetc_set_rss(struct net_device *ndev, int en)static int enetc_set_psfp (struct net_device *ndev,int en) {struct enetc_ndev_priv *priv =netdev_priv (ndev); +int err;if (en) { + err =enetc_psfp_enable (priv); +if (err) +return err; + priv->active_offloads |=ENETC_F_QCI ; -enetc_get_max_cap (priv); -enetc_psfp_enable (&priv->si->hw); - }else { - priv->active_offloads &= ~ENETC_F_QCI ; -memset (&priv->psfp_cap,0 ,sizeof (struct psfp_cap )); -enetc_psfp_disable (&priv->si->hw); +return 0 ; } + err =enetc_psfp_disable (priv); +if (err) +return err; + + priv->active_offloads &= ~ENETC_F_QCI ; +return 0 ; }@@ -1590,14 +1598,15 @@ int enetc_set_features(struct net_device *ndev, netdev_features_t features) {netdev_features_t changed = ndev->features ^ features; +int err =0 ;if (changed &NETIF_F_RXHASH )enetc_set_rss (ndev, !!(features &NETIF_F_RXHASH ));if (changed &NETIF_F_HW_TC ) -enetc_set_psfp (ndev, !!(features &NETIF_F_HW_TC )); + err =enetc_set_psfp (ndev, !!(features &NETIF_F_HW_TC )); -return 0 ; +return err; }#ifdef CONFIG_FSL_ENETC_HW_TIMESTAMPING drivers/net/ethernet/freescale/enetc/enetc.h +35
-4@@ -269,6 +269,11 @@ int enetc_setup_tc_taprio(struct net_device *ndev, void *type_data);void enetc_sched_speed_set (struct net_device *ndev);int enetc_setup_tc_cbs (struct net_device *ndev,void *type_data);int enetc_setup_tc_txtime (struct net_device *ndev,void *type_data); +int enetc_setup_tc_block_cb (enum tc_setup_type type,void *type_data, +void *cb_priv); +int enetc_setup_tc_psfp (struct net_device *ndev,void *type_data); +int enetc_psfp_init (struct enetc_ndev_priv *priv); +int enetc_psfp_clean (struct enetc_ndev_priv *priv);static inline void enetc_get_max_cap (struct enetc_ndev_priv *priv) {@@ -288,33 +293,59 @@ static inline void enetc_get_max_cap(struct enetc_ndev_priv *priv) priv->psfp_cap.max_psfp_meter = reg &ENETC_PFMCAPR_MSK ; } -static inline void enetc_psfp_enable (struct enetc_hw *hw) +static inline int enetc_psfp_enable (struct enetc_ndev_priv *priv) { +struct enetc_hw *hw = &priv->si->hw; +int err; + +enetc_get_max_cap (priv); + + err =enetc_psfp_init (priv); +if (err) +return err; +enetc_wr (hw,ENETC_PPSFPMR ,enetc_rd (hw,ENETC_PPSFPMR ) |ENETC_PPSFPMR_PSFPEN |ENETC_PPSFPMR_VS |ENETC_PPSFPMR_PVC |ENETC_PPSFPMR_PVZC ); + +return 0 ; } -static inline void enetc_psfp_disable (struct enetc_hw *hw) +static inline int enetc_psfp_disable (struct enetc_ndev_priv *priv) { +struct enetc_hw *hw = &priv->si->hw; +int err; + + err =enetc_psfp_clean (priv); +if (err) +return err; +enetc_wr (hw,ENETC_PPSFPMR ,enetc_rd (hw,ENETC_PPSFPMR ) & ~ENETC_PPSFPMR_PSFPEN & ~ENETC_PPSFPMR_VS & ~ENETC_PPSFPMR_PVC & ~ENETC_PPSFPMR_PVZC ); + +memset (&priv->psfp_cap,0 ,sizeof (struct psfp_cap )); + +return 0 ; } +#else #define enetc_setup_tc_taprio (ndev, type_data) -EOPNOTSUPP#define enetc_sched_speed_set (ndev) (void)0#define enetc_setup_tc_cbs (ndev, type_data) -EOPNOTSUPP#define enetc_setup_tc_txtime (ndev, type_data) -EOPNOTSUPP +#define enetc_setup_tc_psfp (ndev, type_data) -EOPNOTSUPP +#define enetc_setup_tc_block_cb NULL +#define enetc_get_max_cap (p) \ memset(&((p)->psfp_cap), 0, sizeof(struct psfp_cap)) -static inline void enetc_psfp_enable (struct enetc_hw *hw) +static inline int enetc_psfp_enable (struct enetc_hw *hw) {return 0 ; } -static inline void enetc_psfp_disable (struct enetc_hw *hw) +static inline int enetc_psfp_disable (struct enetc_hw *hw) {return 0 ; }drivers/net/ethernet/freescale/enetc/enetc_hw.h +130
-0@@ -570,6 +570,9 @@ enum bdcr_cmd_class {BDCR_CMD_RFS ,BDCR_CMD_PORT_GCL ,BDCR_CMD_RECV_CLASSIFIER , +BDCR_CMD_STREAM_IDENTIFY , +BDCR_CMD_STREAM_FILTER , +BDCR_CMD_STREAM_GCL , __BDCR_CMD_MAX_LEN, BDCR_CMD_MAX_LEN = __BDCR_CMD_MAX_LEN - 1, };@@ -601,13 +604,152 @@ struct tgs_gcl_data {struct gce entry[]; }; +/* class 7, command 0, Stream Identity Entry Configuration */ +struct streamid_conf { +__le32 stream_handle;/* init gate value */ +__le32 iports; +u8 id_type; +u8 oui[3 ]; +u8 res[3 ]; +u8 en; +}; + +#define ENETC_CBDR_SID_VID_MASK 0xfff +#define ENETC_CBDR_SID_VIDM BIT(12) +#define ENETC_CBDR_SID_TG_MASK 0xc000 +/* streamid_conf address point to this data space */ +struct streamid_data { +union { +u8 dmac[6 ]; +u8 smac[6 ]; + }; +u16 vid_vidm_tg; +}; + +#define ENETC_CBDR_SFI_PRI_MASK 0x7 +#define ENETC_CBDR_SFI_PRIM BIT(3) +#define ENETC_CBDR_SFI_BLOV BIT(4) +#define ENETC_CBDR_SFI_BLEN BIT(5) +#define ENETC_CBDR_SFI_MSDUEN BIT(6) +#define ENETC_CBDR_SFI_FMITEN BIT(7) +#define ENETC_CBDR_SFI_ENABLE BIT(7) +/* class 8, command 0, Stream Filter Instance, Short Format */ +struct sfi_conf { +__le32 stream_handle; +u8 multi; +u8 res[2 ]; +u8 sthm; +/* Max Service Data Unit or Flow Meter Instance Table index. +* Depending on the value of FLT this represents either Max +* Service Data Unit (max frame size) allowed by the filter +* entry or is an index into the Flow Meter Instance table +* index identifying the policer which will be used to police +* it. +*/ +__le16 fm_inst_table_index; +__le16 msdu; +__le16 sg_inst_table_index; +u8 res1[2 ]; +__le32 input_ports; +u8 res2[3 ]; +u8 en; +}; + +/* class 8, command 2 stream Filter Instance status query short format +* command no need structure define +* Stream Filter Instance Query Statistics Response data +*/ +struct sfi_counter_data { +u32 matchl; +u32 matchh; +u32 msdu_dropl; +u32 msdu_droph; +u32 stream_gate_dropl; +u32 stream_gate_droph; +u32 flow_meter_dropl; +u32 flow_meter_droph; +}; + +#define ENETC_CBDR_SGI_OIPV_MASK 0x7 +#define ENETC_CBDR_SGI_OIPV_EN BIT(3) +#define ENETC_CBDR_SGI_CGTST BIT(6) +#define ENETC_CBDR_SGI_OGTST BIT(7) +#define ENETC_CBDR_SGI_CFG_CHG BIT(1) +#define ENETC_CBDR_SGI_CFG_PND BIT(2) +#define ENETC_CBDR_SGI_OEX BIT(4) +#define ENETC_CBDR_SGI_OEXEN BIT(5) +#define ENETC_CBDR_SGI_IRX BIT(6) +#define ENETC_CBDR_SGI_IRXEN BIT(7) +#define ENETC_CBDR_SGI_ACLLEN_MASK 0x3 +#define ENETC_CBDR_SGI_OCLLEN_MASK 0xc +#define ENETC_CBDR_SGI_EN BIT(7) +/* class 9, command 0, Stream Gate Instance Table, Short Format +* class 9, command 2, Stream Gate Instance Table entry query write back +* Short Format +*/ +struct sgi_table { +u8 res[8 ]; +u8 oipv; +u8 res0[2 ]; +u8 ocgtst; +u8 res1[7 ]; +u8 gset; +u8 oacl_len; +u8 res2[2 ]; +u8 en; +}; + +#define ENETC_CBDR_SGI_AIPV_MASK 0x7 +#define ENETC_CBDR_SGI_AIPV_EN BIT(3) +#define ENETC_CBDR_SGI_AGTST BIT(7) + +/* class 9, command 1, Stream Gate Control List, Long Format */ +struct sgcl_conf { +u8 aipv; +u8 res[2 ]; +u8 agtst; +u8 res1[4 ]; +union { +struct { +u8 res2[4 ]; +u8 acl_len; +u8 res3[3 ]; + }; +u8 cct[8 ];/* Config change time */ + }; +}; + +#define ENETC_CBDR_SGL_IOMEN BIT(0) +#define ENETC_CBDR_SGL_IPVEN BIT(3) +#define ENETC_CBDR_SGL_GTST BIT(4) +#define ENETC_CBDR_SGL_IPV_MASK 0xe +/* Stream Gate Control List Entry */ +struct sgce { +u32 interval; +u8 msdu[3 ]; +u8 multi; +}; + +/* stream control list class 9 , cmd 1 data buffer */
drivers/net/ethernet/freescale/enetc/enetc_hw.h struct tgs_gcl_data { · line 743
Quoted 10 lines
drivers/net/ethernet/freescale/enetc/enetc_hw.h +8
-0+struct sgcl_data { +u32 btl; +u32 bth; +u32 ct; +u32 cte; +struct sgce sgcl[0 ]; +}; + struct enetc_cbd { union{
Wrong identation here. Should be fixed on a separated patch.
Quoted 48 lines
drivers/net/ethernet/freescale/enetc/enetc_hw.h +4
-0+struct sfi_conf sfi_conf; +struct sgi_table sgi_table;struct {__le32 addr[2 ];union {__le32 opt[4 ];struct tgs_gcl_conf gcl_conf; +struct streamid_conf sid_set; +struct sgcl_conf sgcl_conf; }; };/* Long format */ __le32 data[6 ];drivers/net/ethernet/freescale/enetc/enetc_pf.c +1
-3@@ -740,12 +740,10 @@ static void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,if (si->hw_features &ENETC_SI_F_QBV ) priv->active_offloads |=ENETC_F_QBV ; -if (si->hw_features &ENETC_SI_F_PSFP ) { +if (si->hw_features &ENETC_SI_F_PSFP && !enetc_psfp_enable (priv)) { priv->active_offloads |=ENETC_F_QCI ; ndev->features |=NETIF_F_HW_TC ; ndev->hw_features |=NETIF_F_HW_TC ; -enetc_get_max_cap (priv); -enetc_psfp_enable (&si->hw); }/* pick up primary MAC address from SI */ drivers/net/ethernet/freescale/enetc/enetc_qos.c +2
-1@@ -5,6 +5,8 @@#include <net/pkt_sched.h> #include <linux/math64.h> +#include <net/pkt_cls.h> +#include <net/tc_act/tc_gate.h> static u16 enetc_get_max_gcl_len (struct enetc_hw *hw) {@@ -108,13 +110,13 @@ static int enetc_setup_taprio(struct net_device *ndev, gcl_data->cte =cpu_to_le32 (admin_conf->cycle_time_extension);for (i =0 ; i < gcl_len; i++) { -struct tc_taprio_sched_entry *temp_entry;
drivers/net/ethernet/freescale/enetc/enetc_qos.c static int enetc_setup_taprio(struct net_device *ndev, · line 119
Quoted 10 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +4
-3+struct tc_taprio_sched_entry *to;struct gce *temp_gce = gce + i; - temp_entry = &admin_conf->entries[i]; + to = &admin_conf->entries[i]; - temp_gce->gate = (u8 )temp_entry->gate_mask; - temp_gce->period =cpu_to_le32 (temp_entry->interval); + temp_gce->gate = (u8 )to->gate_mask; + temp_gce->period =cpu_to_le32 (to->interval);
These changes seem unrelated to the series.
Quoted 55 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +48
-0} cbd.length =cpu_to_le16 (data_size);@@ -331,3 +333,1071 @@ int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data)return 0 ; } + +enum streamid_type { +STREAMID_TYPE_RESERVED =0 , +STREAMID_TYPE_NULL , +STREAMID_TYPE_SMAC , +}; + +enum streamid_vlan_tagged { +STREAMID_VLAN_RESERVED =0 , +STREAMID_VLAN_TAGGED , +STREAMID_VLAN_UNTAGGED , +STREAMID_VLAN_ALL , +}; + +#define ENETC_PSFP_WILDCARD -1 +#define HANDLE_OFFSET 100 + +enum forward_type { +FILTER_ACTION_TYPE_PSFP =BIT (0 ), +FILTER_ACTION_TYPE_ACL =BIT (1 ), +FILTER_ACTION_TYPE_BOTH =GENMASK (1 ,0 ), +}; + +/* This is for limit output type for input actions */ +struct actions_fwd { +u64 actions; +u64 keys;/* include the must needed keys */ +enum forward_type output; +}; + +struct psfp_streamfilter_counters { +u64 matching_frames_count; +u64 passing_frames_count; +u64 not_passing_frames_count; +u64 passing_sdu_count; +u64 not_passing_sdu_count; +u64 red_frames_count; +}; + +struct enetc_streamid { +u32 index; +union { +u8 src_mac[6 ]; +u8 dst_mac[6 ]; + }; +u8 filtertype; +u16 vid; +u8 tagged;
drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data) · line 393
Quoted 10 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +10
-0+s32 handle; +}; + +struct enetc_psfp_filter { +u32 index; +s32 handle; +s8 prio; +u32 gate_id; +s32 meter_id; +u32 refcount;
It may be more appropriate to use <linux/refcount.h> instead of rolling your own.
Quoted 39 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +39
-0+struct hlist_node node; +}; + +struct enetc_psfp_gate { +u32 index; +s8 init_ipv; +u64 basetime; +u64 cycletime; +u64 cycletimext; +u32 num_entries; +u32 refcount; +struct hlist_node node; +struct action_gate_entry entries[0 ]; +}; + +struct enetc_stream_filter { +struct enetc_streamid sid; +u32 sfi_index; +u32 sgi_index; +struct flow_stats stats; +struct hlist_node node; +}; + +struct enetc_psfp { +unsigned long dev_bitmap; +unsigned long *psfp_sfi_bitmap; +struct hlist_head stream_list; +struct hlist_head psfp_filter_list; +struct hlist_head psfp_gate_list; +spinlock_t psfp_lock;/* spinlock for the struct enetc_psfp r/w */ +}; + +struct actions_fwd enetc_act_fwd[] = { + { +BIT (FLOW_ACTION_GATE ), +BIT (FLOW_DISSECTOR_KEY_ETH_ADDRS ), +FILTER_ACTION_TYPE_PSFP + }, +/* example for ACL actions */
drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data) · line 442
Quoted 10 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +10
-0+ { +BIT (FLOW_ACTION_DROP ), +0 , + FILTER_ACTION_TYPE_ACL + } +}; + +static struct enetc_psfp epsfp = { + .psfp_sfi_bitmap =NULL , +};
Is it possible to have more than one these controllers in the same system? If it's possible to have more than one, this should be moved to live under a per-device struct.
Quoted 282 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +282
-0+ +static LIST_HEAD (enetc_block_cb_list); + +static inline int enetc_get_port (struct enetc_ndev_priv *priv) +{ +return priv->si->pdev->devfn &0x7 ; +} + +/* Stream Identity Entry Set Descriptor */ +static int enetc_streamid_hw_set (struct enetc_ndev_priv *priv, +struct enetc_streamid *sid, +u8 enable) +{ +struct enetc_cbd cbd = {.cmd =0 }; +struct streamid_data *si_data; +struct streamid_conf *si_conf; +u16 data_size; +dma_addr_t dma; +int err; + +if (sid->index >= priv->psfp_cap.max_streamid) +return -EINVAL ; + +if (sid->filtertype !=STREAMID_TYPE_NULL && + sid->filtertype !=STREAMID_TYPE_SMAC ) +return -EOPNOTSUPP ; + +/* Disable operation before enable */ + cbd.index =cpu_to_le16 ((u16 )sid->index); + cbd.cls =BDCR_CMD_STREAM_IDENTIFY ; + cbd.status_flags =0 ; + + data_size =sizeof (struct streamid_data ); + si_data =kzalloc (data_size, __GFP_DMA |GFP_KERNEL ); + cbd.length =cpu_to_le16 (data_size); + + dma =dma_map_single (&priv->si->pdev->dev, si_data, + data_size,DMA_FROM_DEVICE ); +if (dma_mapping_error (&priv->si->pdev->dev, dma)) { +netdev_err (priv->si->ndev,"DMA mapping failed!\n" ); +kfree (si_data); +return -ENOMEM ; + } + + cbd.addr[0 ] =lower_32_bits (dma); + cbd.addr[1 ] =upper_32_bits (dma); +memset (si_data->dmac,0xff ,ETH_ALEN ); + si_data->vid_vidm_tg = +cpu_to_le16 (ENETC_CBDR_SID_VID_MASK + + ((0x3 <<14 ) |ENETC_CBDR_SID_VIDM )); + + si_conf = &cbd.sid_set; +/* Only one port supported for one entry, set itself */ + si_conf->iports =1 <<enetc_get_port (priv); + si_conf->id_type =1 ; + si_conf->oui[2 ] =0x0 ; + si_conf->oui[1 ] =0x80 ; + si_conf->oui[0 ] =0xC2 ; + + err =enetc_send_cmd (priv->si, &cbd); +if (err) +return -EINVAL ; + +if (!enable) { +kfree (si_data); +return 0 ; + } + +/* Enable the entry overwrite again incase space flushed by hardware */ +memset (&cbd,0 ,sizeof (cbd)); + + cbd.index =cpu_to_le16 ((u16 )sid->index); + cbd.cmd =0 ; + cbd.cls =BDCR_CMD_STREAM_IDENTIFY ; + cbd.status_flags =0 ; + + si_conf->en =0x80 ; + si_conf->stream_handle =cpu_to_le32 (sid->handle); + si_conf->iports =1 <<enetc_get_port (priv); + si_conf->id_type = sid->filtertype; + si_conf->oui[2 ] =0x0 ; + si_conf->oui[1 ] =0x80 ; + si_conf->oui[0 ] =0xC2 ; + +memset (si_data,0 , data_size); + + cbd.length =cpu_to_le16 (data_size); + + cbd.addr[0 ] =lower_32_bits (dma); + cbd.addr[1 ] =upper_32_bits (dma); + +/* VIDM default to be 1. +* VID Match. If set (b1) then the VID must match, otherwise +* any VID is considered a match. VIDM setting is only used +* when TG is set to b01. +*/ +if (si_conf->id_type ==STREAMID_TYPE_NULL ) { +ether_addr_copy (si_data->dmac, sid->dst_mac); + si_data->vid_vidm_tg = +cpu_to_le16 ((sid->vid &ENETC_CBDR_SID_VID_MASK ) + + ((((u16)(sid->tagged) &0x3 ) <<14 ) + |ENETC_CBDR_SID_VIDM )); + }else if (si_conf->id_type ==STREAMID_TYPE_SMAC ) { +ether_addr_copy (si_data->smac, sid->src_mac); + si_data->vid_vidm_tg = +cpu_to_le16 ((sid->vid &ENETC_CBDR_SID_VID_MASK ) + + ((((u16)(sid->tagged) &0x3 ) <<14 ) + |ENETC_CBDR_SID_VIDM )); + } + + err =enetc_send_cmd (priv->si, &cbd); +kfree (si_data); + +return err; +} + +/* Stream Filter Instance Set Descriptor */ +static int enetc_streamfilter_hw_set (struct enetc_ndev_priv *priv, +struct enetc_psfp_filter *sfi, +u8 enable) +{ +struct enetc_cbd cbd = {.cmd =0 }; +struct sfi_conf *sfi_config; + + cbd.index =cpu_to_le16 (sfi->index); + cbd.cls =BDCR_CMD_STREAM_FILTER ; + cbd.status_flags =0x80 ; + cbd.length =cpu_to_le16 (1 ); + + sfi_config = &cbd.sfi_conf; +if (!enable) + goto exit; + + sfi_config->en =0x80 ; + +if (sfi->handle >=0 ) { + sfi_config->stream_handle = +cpu_to_le32 (sfi->handle); + sfi_config->sthm |=0x80 ; + } + + sfi_config->sg_inst_table_index =cpu_to_le16 (sfi->gate_id); + sfi_config->input_ports =1 <<enetc_get_port (priv); + +/* The priority value which may be matched against the +* frame’s priority value to determine a match for this entry. +*/ +if (sfi->prio >=0 ) + sfi_config->multi |= (sfi->prio &0x7 ) |0x8 ; + +/* Filter Type. Identifies the contents of the MSDU/FM_INST_INDEX +* field as being either an MSDU value or an index into the Flow +* Meter Instance table. +* TODO: no limit max sdu +*/ + +if (sfi->meter_id >=0 ) { + sfi_config->fm_inst_table_index =cpu_to_le16 (sfi->meter_id); + sfi_config->multi |=0x80 ; + } + +exit: +return enetc_send_cmd (priv->si, &cbd); +} + +static int enetc_streamcounter_hw_get (struct enetc_ndev_priv *priv, +u32 index, +struct psfp_streamfilter_counters *cnt) +{ +struct enetc_cbd cbd = { .cmd =2 }; +struct sfi_counter_data *data_buf; +dma_addr_t dma; +u16 data_size; +int err; + + cbd.index =cpu_to_le16 ((u16 )index); + cbd.cmd =2 ; + cbd.cls =BDCR_CMD_STREAM_FILTER ; + cbd.status_flags =0 ; + + data_size =sizeof (struct sfi_counter_data ); + data_buf =kzalloc (data_size, __GFP_DMA |GFP_KERNEL ); +if (!data_buf) +return -ENOMEM ; + + dma =dma_map_single (&priv->si->pdev->dev, data_buf, + data_size,DMA_FROM_DEVICE ); +if (dma_mapping_error (&priv->si->pdev->dev, dma)) { +netdev_err (priv->si->ndev,"DMA mapping failed!\n" ); + err = -ENOMEM ; + goto exit; + } + cbd.addr[0 ] =lower_32_bits (dma); + cbd.addr[1 ] =upper_32_bits (dma); + + cbd.length =cpu_to_le16 (data_size); + + err =enetc_send_cmd (priv->si, &cbd); +if (err) + goto exit; + + cnt->matching_frames_count = + ((u64 )le32_to_cpu (data_buf->matchh) <<32 ) + + data_buf->matchl; + + cnt->not_passing_sdu_count = + ((u64 )le32_to_cpu (data_buf->msdu_droph) <<32 ) + + data_buf->msdu_dropl; + + cnt->passing_sdu_count = cnt->matching_frames_count + - cnt->not_passing_sdu_count; + + cnt->not_passing_frames_count = + ((u64 )le32_to_cpu (data_buf->stream_gate_droph) <<32 ) + +le32_to_cpu (data_buf->stream_gate_dropl); + + cnt->passing_frames_count = cnt->matching_frames_count + - cnt->not_passing_sdu_count + - cnt->not_passing_frames_count; + + cnt->red_frames_count = + ((u64 )le32_to_cpu (data_buf->flow_meter_droph) <<32 ) + +le32_to_cpu (data_buf->flow_meter_dropl); + +exit: +kfree (data_buf); +return err; +} + +static int get_start_ns (struct enetc_ndev_priv *priv,u64 cycle,u64 *start) +{ +u64 now_lo, now_hi, now, n; + + now_lo =enetc_rd (&priv->si->hw,ENETC_SICTR0 ); + now_hi =enetc_rd (&priv->si->hw,ENETC_SICTR1 ); + now = now_lo | now_hi <<32 ; + +if (WARN_ON (!cycle)) +return -EFAULT ; + + n =div64_u64 (now, cycle); + + *start = (n +1 ) * cycle; + +return 0 ; +} + +/* Stream Gate Instance Set Descriptor */ +static int enetc_streamgate_hw_set (struct enetc_ndev_priv *priv, +struct enetc_psfp_gate *sgi, +u8 enable) +{ +struct enetc_cbd cbd = { .cmd =0 }; +struct sgi_table *sgi_config; +struct sgcl_conf *sgcl_config; +struct sgcl_data *sgcl_data; +struct sgce *sgce; +dma_addr_t dma; +u16 data_size; +int err, i; + + cbd.index =cpu_to_le16 (sgi->index); + cbd.cmd =0 ; + cbd.cls =BDCR_CMD_STREAM_GCL ; + cbd.status_flags =0x80 ; + +/* disable */ +if (!enable) +return enetc_send_cmd (priv->si, &cbd); + +/* enable */ + sgi_config = &cbd.sgi_table; + +/* Keep open before gate list start */ + sgi_config->ocgtst =0x80 ; + + sgi_config->oipv = (sgi->init_ipv <0 ) ? +0x0 : ((sgi->init_ipv &0x7 ) |0x8 ); + + sgi_config->en =0x80 ; + +/* Basic config */
drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data) · line 734
Quoted 10 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +10
-0+ err =enetc_send_cmd (priv->si, &cbd); +if (err) +return -EINVAL ; + +if (!sgi->num_entries) +return 0 ; + +if (sgi->num_entries > priv->psfp_cap.max_psfp_gatelist || + !sgi->cycletime) +return -EINVAL ;
You already check this when "sgi" is created, this check doesn't seem to be needed. If it's needed, it should be moved to the top of the function.
Quoted 137 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +137
-0+ +memset (&cbd,0 ,sizeof (cbd)); + + cbd.index =cpu_to_le16 (sgi->index); + cbd.cmd =1 ; + cbd.cls =BDCR_CMD_STREAM_GCL ; + cbd.status_flags =0 ; + + sgcl_config = &cbd.sgcl_conf; + + sgcl_config->acl_len = (sgi->num_entries -1 ) &0x3 ; + + data_size =struct_size (sgcl_data, sgcl, sgi->num_entries); + + sgcl_data =kzalloc (data_size, __GFP_DMA |GFP_KERNEL ); +if (!sgcl_data) +return -ENOMEM ; + + cbd.length =cpu_to_le16 (data_size); + + dma =dma_map_single (&priv->si->pdev->dev, + sgcl_data, data_size, +DMA_FROM_DEVICE ); +if (dma_mapping_error (&priv->si->pdev->dev, dma)) { +netdev_err (priv->si->ndev,"DMA mapping failed!\n" ); +kfree (sgcl_data); +return -ENOMEM ; + } + + cbd.addr[0 ] =lower_32_bits (dma); + cbd.addr[1 ] =upper_32_bits (dma); + + sgce = &sgcl_data->sgcl[0 ]; + + sgcl_config->agtst =0x80 ; + + sgcl_data->ct =cpu_to_le32 (sgi->cycletime); + sgcl_data->cte =cpu_to_le32 (sgi->cycletimext); + +if (sgi->init_ipv >=0 ) + sgcl_config->aipv = (sgi->init_ipv &0x7 ) |0x8 ; + +for (i =0 ; i < sgi->num_entries; i++) { +struct action_gate_entry *from = &sgi->entries[i]; +struct sgce *to = &sgce[i]; + +if (from->gate_state) + to->multi |=0x10 ; + +if (from->ipv >=0 ) + to->multi |= ((from->ipv &0x7 ) <<5 ) |0x08 ; + +if (from->maxoctets) + to->multi |=0x01 ; + + to->interval =cpu_to_le32 (from->interval); + to->msdu[0 ] = from->maxoctets &0xFF ; + to->msdu[1 ] = (from->maxoctets >>8 ) &0xFF ; + to->msdu[2 ] = (from->maxoctets >>16 ) &0xFF ; + } + +/* If basetime is 0, calculate start time */ +if (!sgi->basetime) { +u64 start; + + err =get_start_ns (priv, sgi->cycletime, &start); +if (err) + goto exit; + sgcl_data->btl =cpu_to_le32 (lower_32_bits (start)); + sgcl_data->bth =cpu_to_le32 (upper_32_bits (start)); + }else { +u32 hi, lo; + + hi =upper_32_bits (sgi->basetime); + lo =lower_32_bits (sgi->basetime); + sgcl_data->bth =cpu_to_le32 (hi); + sgcl_data->btl =cpu_to_le32 (lo); + } + + err =enetc_send_cmd (priv->si, &cbd); + +exit: +kfree (sgcl_data); + +return err; +} + +static struct enetc_stream_filter *enetc_get_stream_by_index (u32 index) +{ +struct enetc_stream_filter *f; + +hlist_for_each_entry (f, &epsfp.stream_list, node) +if (f->sid.index == index) +return f; + +return NULL ; +} + +static struct enetc_psfp_gate *enetc_get_gate_by_index (u32 index) +{ +struct enetc_psfp_gate *g; + +hlist_for_each_entry (g, &epsfp.psfp_gate_list, node) +if (g->index == index) +return g; + +return NULL ; +} + +static struct enetc_psfp_filter *enetc_get_filter_by_index (u32 index) +{ +struct enetc_psfp_filter *s; + +hlist_for_each_entry (s, &epsfp.psfp_filter_list, node) +if (s->index == index) +return s; + +return NULL ; +} + +static struct enetc_psfp_filter + *enetc_psfp_check_sfi (struct enetc_psfp_filter *sfi) +{ +struct enetc_psfp_filter *s; + +hlist_for_each_entry (s, &epsfp.psfp_filter_list, node) +if (s->gate_id == sfi->gate_id && + s->prio == sfi->prio && + s->meter_id == sfi->meter_id) +return s; + +return NULL ; +} + +static int enetc_get_free_index (struct enetc_ndev_priv *priv) +{ +u32 max_size = priv->psfp_cap.max_psfp_filter;
drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data) · line 881
Quoted 10 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +10
-0+unsigned long index; + + index =find_first_zero_bit (epsfp.psfp_sfi_bitmap, max_size); +if (index == max_size) +return -1 ; + +return index; +} + +static void reduce_ref_sfi (struct enetc_ndev_priv *priv,u32 index)
Please use the more usual idioms, something like "sfi_unref()" or "stream_filter_unref".
Quoted 241 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +241
-0+{ +struct enetc_psfp_filter *sfi; + + sfi =enetc_get_filter_by_index (index); +WARN_ON (!sfi); + sfi->refcount--; + +if (!sfi->refcount) { +enetc_streamfilter_hw_set (priv, sfi, false); +hlist_del (&sfi->node); +kfree (sfi); +clear_bit (sfi->index, epsfp.psfp_sfi_bitmap); + } +} + +static void reduce_ref_sgi (struct enetc_ndev_priv *priv,u32 index) +{ +struct enetc_psfp_gate *sgi; + + sgi =enetc_get_gate_by_index (index); +WARN_ON (!sgi); + sgi->refcount--; + +if (!sgi->refcount) { +enetc_streamgate_hw_set (priv, sgi, false); +hlist_del (&sgi->node); +kfree (sgi); + } +} + +static void remove_one_chain (struct enetc_ndev_priv *priv, +struct enetc_stream_filter *filter) +{ +reduce_ref_sgi (priv, filter->sgi_index); +reduce_ref_sfi (priv, filter->sfi_index); + +hlist_del (&filter->node); +kfree (filter); +} + +static int enetc_psfp_hw_set (struct enetc_ndev_priv *priv, +struct enetc_streamid *sid, +struct enetc_psfp_filter *sfi, +struct enetc_psfp_gate *sgi) +{ +int err; + + err =enetc_streamid_hw_set (priv, sid, true); +if (err) +return err; + +if (sfi) { + err =enetc_streamfilter_hw_set (priv, sfi, true); +if (err) + goto revert_sid; + } + + err =enetc_streamgate_hw_set (priv, sgi, true); +if (err) + goto revert_sfi; + +return 0 ; + +revert_sfi: +if (sfi && !sfi->refcount) +enetc_streamfilter_hw_set (priv, sfi, false); +revert_sid: +enetc_streamid_hw_set (priv, sid, false); +return err; +} + +struct actions_fwd *enetc_check_flow_actions (u64 acts,unsigned int inputkeys) +{ +int i; + +for (i =0 ; i <ARRAY_SIZE (enetc_act_fwd); i++) +if (acts == enetc_act_fwd[i].actions && + inputkeys & enetc_act_fwd[i].keys) +return &enetc_act_fwd[i]; + +return NULL ; +} + +static int enetc_psfp_parse_clsflower (struct enetc_ndev_priv *priv, +struct flow_cls_offload *f) +{ +struct flow_rule *rule =flow_cls_offload_flow_rule (f); +struct netlink_ext_ack *extack = f->common.extack; +struct enetc_stream_filter *filter, *old_filter; +struct enetc_psfp_filter *sfi, *old_sfi; +struct enetc_psfp_gate *sgi, *old_sgi; +struct flow_action_entry *entry; +struct action_gate_entry *e; +u8 sfi_overwrite =0 ; +int entries_size; +int i, err; + +if (f->common.chain_index >= priv->psfp_cap.max_streamid) { +NL_SET_ERR_MSG_MOD (extack,"No Stream identify resource!" ); +return -ENOSPC ; + } + +flow_action_for_each (i, entry, &rule->action) +if (entry->id ==FLOW_ACTION_GATE ) +break ; + +if (entry->id !=FLOW_ACTION_GATE ) +return -EINVAL ; + + filter =kzalloc (sizeof (*filter),GFP_KERNEL ); +if (!filter) +return -ENOMEM ; + + filter->sid.index = f->common.chain_index; + +if (flow_rule_match_key (rule,FLOW_DISSECTOR_KEY_ETH_ADDRS )) { +struct flow_match_eth_addrs match; + +flow_rule_match_eth_addrs (rule, &match); + +if (!is_zero_ether_addr (match.mask->dst)) { +ether_addr_copy (filter->sid.dst_mac, match.key->dst); + filter->sid.filtertype =STREAMID_TYPE_NULL ; + } + +if (!is_zero_ether_addr (match.mask->src)) { +ether_addr_copy (filter->sid.src_mac, match.key->src); + filter->sid.filtertype =STREAMID_TYPE_SMAC ; + } + }else { +NL_SET_ERR_MSG_MOD (extack,"Unsupported, must ETH_ADDRS" ); +return -EINVAL ; + } + +if (flow_rule_match_key (rule,FLOW_DISSECTOR_KEY_VLAN )) { +struct flow_match_vlan match; + +flow_rule_match_vlan (rule, &match); +if (match.mask->vlan_priority) { +if (match.mask->vlan_priority != + (VLAN_PRIO_MASK >>VLAN_PRIO_SHIFT )) { +NL_SET_ERR_MSG_MOD (extack,"Only full mask is supported for VLAN priority" ); + err = -EINVAL ; + goto free_filter; + } + } + +if (match.mask->vlan_tpid) { +if (match.mask->vlan_tpid !=VLAN_VID_MASK ) { +NL_SET_ERR_MSG_MOD (extack,"Only full mask is supported for VLAN id" ); + err = -EINVAL ; + goto free_filter; + } + + filter->sid.vid = match.key->vlan_tpid; +if (!filter->sid.vid) + filter->sid.tagged =STREAMID_VLAN_UNTAGGED ; +else + filter->sid.tagged =STREAMID_VLAN_TAGGED ; + } + }else { + filter->sid.tagged =STREAMID_VLAN_ALL ; + } + +/* parsing gate action */ +if (entry->gate.index >= priv->psfp_cap.max_psfp_gate) { +NL_SET_ERR_MSG_MOD (extack,"No Stream Gate resource!" ); + err = -ENOSPC ; + goto free_filter; + } + +if (entry->gate.num_entries >= priv->psfp_cap.max_psfp_gatelist) { +NL_SET_ERR_MSG_MOD (extack,"No Stream Gate resource!" ); + err = -ENOSPC ; + goto free_filter; + } + + entries_size =struct_size (sgi, entries, entry->gate.num_entries); + sgi =kzalloc (entries_size,GFP_KERNEL ); +if (!sgi) { + err = -ENOMEM ; + goto free_filter; + } + + sgi->index = entry->gate.index; + sgi->init_ipv = entry->gate.prio; + sgi->basetime = entry->gate.basetime; + sgi->cycletime = entry->gate.cycletime; + sgi->num_entries = entry->gate.num_entries; + + e = sgi->entries; +for (i =0 ; i < entry->gate.num_entries; i++) { + e[i].gate_state = entry->gate.entries[i].gate_state; + e[i].interval = entry->gate.entries[i].interval; + e[i].ipv = entry->gate.entries[i].ipv; + e[i].maxoctets = entry->gate.entries[i].maxoctets; + } + + filter->sgi_index = sgi->index; + + sfi =kzalloc (sizeof (*sfi),GFP_KERNEL ); +if (!sfi) { + err = -ENOMEM ; + goto free_gate; + } + + sfi->gate_id = sgi->index; + +/* flow meter not support yet */ + sfi->meter_id =ENETC_PSFP_WILDCARD ; + +/* prio ref the filter prio */ +if (f->common.prio && f->common.prio <=BIT (3 )) + sfi->prio = f->common.prio -1 ; +else + sfi->prio =ENETC_PSFP_WILDCARD ; + + old_sfi =enetc_psfp_check_sfi (sfi); +if (!old_sfi) { +int index; + + index =enetc_get_free_index (priv); +if (sfi->handle <0 ) { +NL_SET_ERR_MSG_MOD (extack,"No Stream Filter resource!" ); + err = -ENOSPC ; + goto free_sfi; + } + + sfi->index = index; + sfi->handle = index +HANDLE_OFFSET ; +/* Update the stream filter handle also */ + filter->sid.handle = sfi->handle; + filter->sfi_index = sfi->index; + sfi_overwrite =0 ; + }else { + filter->sfi_index = old_sfi->index; + filter->sid.handle = old_sfi->handle; + sfi_overwrite =1 ; + } + + err = enetc_psfp_hw_set(priv, &filter->sid,
drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data) · line 1132
Quoted 10 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +10
-0+ sfi_overwrite ? NULL : sfi, sgi); +if (err) + goto free_sfi; + +spin_lock (&epsfp.psfp_lock); + old_sgi =enetc_get_gate_by_index (filter->sgi_index); +if (old_sgi) { + sgi->refcount = old_sgi->refcount; +hlist_del (&old_sgi->node); +kfree (old_sgi);
I don't understand what you are trying to achieve here. But there should be a cleaner way.
Quoted 274 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +271
-0+ } + + sgi->refcount++; +hlist_add_head (&sgi->node, &epsfp.psfp_gate_list); + +if (!old_sfi) { +hlist_add_head (&sfi->node, &epsfp.psfp_filter_list); +set_bit (sfi->index, epsfp.psfp_sfi_bitmap); + sfi->refcount++; + }else { +kfree (sfi); + old_sfi->refcount++; + } + + old_filter =enetc_get_stream_by_index (filter->sid.index); +if (old_filter) +remove_one_chain (priv, old_filter); + + filter->stats.lastused = jiffies; +hlist_add_head (&filter->node, &epsfp.stream_list); + +spin_unlock (&epsfp.psfp_lock); + +return 0 ; + +free_sfi: +kfree (sfi); +free_gate: +kfree (sgi); +free_filter: +kfree (filter); + +return err; +} + +static int enetc_config_clsflower (struct enetc_ndev_priv *priv, +struct flow_cls_offload *cls_flower) +{ +struct flow_rule *rule =flow_cls_offload_flow_rule (cls_flower); +struct netlink_ext_ack *extack = cls_flower->common.extack; +struct flow_dissector *dissector = rule->match.dissector; +struct flow_action *action = &rule->action; +struct flow_action_entry *entry; +struct actions_fwd *fwd; +u64 actions =0 ; +int i, err; + +if (!flow_action_has_entries (action)) { +NL_SET_ERR_MSG_MOD (extack,"At least one action is needed" ); +return -EINVAL ; + } + +flow_action_for_each (i, entry, action) + actions |=BIT (entry->id); + + fwd =enetc_check_flow_actions (actions, dissector->used_keys); +if (!fwd) { +NL_SET_ERR_MSG_MOD (extack,"Unsupported filter type!" ); +return -EOPNOTSUPP ; + } + +if (fwd->output &FILTER_ACTION_TYPE_PSFP ) { + err =enetc_psfp_parse_clsflower (priv, cls_flower); +if (err) { +NL_SET_ERR_MSG_MOD (extack,"Invalid PSFP inputs" ); +return err; + } + }else { +NL_SET_ERR_MSG_MOD (extack,"Unsupported actions" ); +return -EOPNOTSUPP ; + } + +return 0 ; +} + +static int enetc_psfp_destroy_clsflower (struct enetc_ndev_priv *priv, +struct flow_cls_offload *f) +{ +struct enetc_stream_filter *filter; +struct netlink_ext_ack *extack = f->common.extack; +int err; + +if (f->common.chain_index >= priv->psfp_cap.max_streamid) { +NL_SET_ERR_MSG_MOD (extack,"No Stream identify resource!" ); +return -ENOSPC ; + } + + filter =enetc_get_stream_by_index (f->common.chain_index); +if (!filter) +return -EINVAL ; + + err =enetc_streamid_hw_set (priv, &filter->sid, false); +if (err) +return err; + +remove_one_chain (priv, filter); + +return 0 ; +} + +static int enetc_destroy_clsflower (struct enetc_ndev_priv *priv, +struct flow_cls_offload *f) +{ +return enetc_psfp_destroy_clsflower (priv, f); +} + +static int enetc_psfp_get_stats (struct enetc_ndev_priv *priv, +struct flow_cls_offload *f) +{ +struct psfp_streamfilter_counters counters = {}; +struct enetc_stream_filter *filter; +struct flow_stats stats = {}; +int err; + + filter =enetc_get_stream_by_index (f->common.chain_index); +if (!filter) +return -EINVAL ; + + err =enetc_streamcounter_hw_get (priv, filter->sfi_index, &counters); +if (err) +return -EINVAL ; + +spin_lock (&epsfp.psfp_lock); + stats.pkts = counters.matching_frames_count - filter->stats.pkts; + stats.dropped = counters.not_passing_frames_count - + filter->stats.dropped; + stats.lastused = filter->stats.lastused; + filter->stats.pkts += stats.pkts; + filter->stats.dropped += stats.dropped; + +spin_unlock (&epsfp.psfp_lock); + +flow_stats_update (&f->stats,0x0 , stats.pkts, stats.lastused, + stats.dropped); + +return 0 ; +} + +static int enetc_setup_tc_cls_flower (struct enetc_ndev_priv *priv, +struct flow_cls_offload *cls_flower) +{ +switch (cls_flower->command) { +case FLOW_CLS_REPLACE : +return enetc_config_clsflower (priv, cls_flower); +case FLOW_CLS_DESTROY : +return enetc_destroy_clsflower (priv, cls_flower); +case FLOW_CLS_STATS : +return enetc_psfp_get_stats (priv, cls_flower); +default : +return -EOPNOTSUPP ; + } +} + +static inline void clean_psfp_sfi_bitmap (void ) +{ +bitmap_free (epsfp.psfp_sfi_bitmap); + epsfp.psfp_sfi_bitmap =NULL ; +} + +static void clean_stream_list (void ) +{ +struct enetc_stream_filter *s; +struct hlist_node *tmp; + +hlist_for_each_entry_safe (s, tmp, &epsfp.stream_list, node) { +hlist_del (&s->node); +kfree (s); + } +} + +static void clean_sfi_list (void ) +{ +struct enetc_psfp_filter *sfi; +struct hlist_node *tmp; + +hlist_for_each_entry_safe (sfi, tmp, &epsfp.psfp_filter_list, node) { +hlist_del (&sfi->node); +kfree (sfi); + } +} + +static void clean_sgi_list (void ) +{ +struct enetc_psfp_gate *sgi; +struct hlist_node *tmp; + +hlist_for_each_entry_safe (sgi, tmp, &epsfp.psfp_gate_list, node) { +hlist_del (&sgi->node); +kfree (sgi); + } +} + +static void clean_psfp_all (void ) +{ +/* Disable all list nodes and free all memory */ +clean_sfi_list (); +clean_sgi_list (); +clean_stream_list (); + epsfp.dev_bitmap =0 ; +clean_psfp_sfi_bitmap (); +} + +int enetc_setup_tc_block_cb (enum tc_setup_type type,void *type_data, +void *cb_priv) +{ +struct net_device *ndev = cb_priv; + +if (!tc_can_offload (ndev)) +return -EOPNOTSUPP ; + +switch (type) { +case TC_SETUP_CLSFLOWER : +return enetc_setup_tc_cls_flower (netdev_priv (ndev), type_data); +default : +return -EOPNOTSUPP ; + } +} + +int enetc_psfp_init (struct enetc_ndev_priv *priv) +{ +if (epsfp.psfp_sfi_bitmap) +return 0 ; + + epsfp.psfp_sfi_bitmap =bitmap_zalloc (priv->psfp_cap.max_psfp_filter, +GFP_KERNEL ); +if (!epsfp.psfp_sfi_bitmap) +return -ENOMEM ; + +spin_lock_init (&epsfp.psfp_lock); + +if (list_empty (&enetc_block_cb_list)) + epsfp.dev_bitmap =0 ; + +return 0 ; +} + +int enetc_psfp_clean (struct enetc_ndev_priv *priv) +{ +if (!list_empty (&enetc_block_cb_list)) +return -EBUSY ; + +clean_psfp_all (); + +return 0 ; +} + +int enetc_setup_tc_psfp (struct net_device *ndev,void *type_data) +{ +struct enetc_ndev_priv *priv =netdev_priv (ndev); +struct flow_block_offload *f = type_data; +int err; + + err =flow_block_cb_setup_simple (f, &enetc_block_cb_list, + enetc_setup_tc_block_cb, + ndev, ndev, true); +if (err) +return err; + +switch (f->command) { +case FLOW_BLOCK_BIND : +set_bit (enetc_get_port (priv), &epsfp.dev_bitmap); +break ; +case FLOW_BLOCK_UNBIND : +clear_bit (enetc_get_port (priv), &epsfp.dev_bitmap); +if (!epsfp.dev_bitmap) +clean_psfp_all (); +break ; + } + +return 0 ; +}Signature · 1 line
2.17.1
Signature · 1 line
Vinicius
RE: [EXT] Re: [RFC,net-next 5/9] net: enetc: add tc flower psfp offload driver
Hi Vinicius
Quoted 8 lines
-----Original Message-----
From: Vinicius Costa Gomes <[email protected]>
Sent: 2020年3月13日 6:14Quoted 2 lines
struct enetc_cbd { union{Wrong identation here. Should be fixed on a separated patch.
Thanks.
Quoted 15 lines
Review static int enetc_setup_taprio(struct net_device
Quoted 4 lines
static u16
enetc_get_max_gcl_len(struct enetc_hw *hw){(quoted diff) +0
-0@@ -108,13 +110,13 @@ static int enetc_setup_taprio(struct net_device*ndev,
Quoted 10 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +2
-2gcl_data->cte =cpu_to_le32 (admin_conf->cycle_time_extension);for (i =0 ; i < gcl_len; i++) { -struct tc_taprio_sched_entry *temp_entry; +struct tc_taprio_sched_entry *to;struct gce *temp_gce = gce + i; - temp_entry = &admin_conf->entries[i]; + to = &admin_conf->entries[i];
Quoted 6 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c static int enetc_setup_taprio(struct net_device *ndev, · line 119
Quoted 4 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +2
-2- temp_gce->gate = (u8 )temp_entry->gate_mask; - temp_gce->period =cpu_to_le32 (temp_entry->interval); + temp_gce->gate = (u8 )to->gate_mask; + temp_gce->period =cpu_to_le32 (to->interval);These changes seem unrelated to the series.
Thanks, these should be mis operation by replace.
Quoted 62 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device
Quoted 5 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +0
-0} cbd.length =cpu_to_le16 (data_size);@@ -331,3 +333,1071 @@ int enetc_setup_tc_txtime(struct net_device*ndev, void
*type_data)Quoted 56 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +53
-0return 0 ; } + +enum streamid_type { +STREAMID_TYPE_RESERVED =0 , +STREAMID_TYPE_NULL , +STREAMID_TYPE_SMAC , +}; + +enum streamid_vlan_tagged { +STREAMID_VLAN_RESERVED =0 , +STREAMID_VLAN_TAGGED , +STREAMID_VLAN_UNTAGGED , +STREAMID_VLAN_ALL , +}; + +#define ENETC_PSFP_WILDCARD -1 +#define HANDLE_OFFSET 100 + +enum forward_type { +FILTER_ACTION_TYPE_PSFP =BIT (0 ), +FILTER_ACTION_TYPE_ACL =BIT (1 ), +FILTER_ACTION_TYPE_BOTH =GENMASK (1 ,0 ), +}; + +/* This is for limit output type for input actions */ +struct actions_fwd { +u64 actions; +u64 keys;/* include the must needed keys */ +enum forward_type output; +}; + +struct psfp_streamfilter_counters { +u64 matching_frames_count; +u64 passing_frames_count; +u64 not_passing_frames_count; +u64 passing_sdu_count; +u64 not_passing_sdu_count; +u64 red_frames_count; +}; + +struct enetc_streamid { +u32 index; +union { +u8 src_mac[6 ]; +u8 dst_mac[6 ]; + }; +u8 filtertype; +u16 vid; +u8 tagged; +s32 handle; +}; + +struct enetc_psfp_filter { +u32 index;
Quoted 8 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data) · line 393
Quoted 5 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +5
-0+s32 handle; +s8 prio; +u32 gate_id; +s32 meter_id; +u32 refcount;It may be more appropriate to use <
linux/refcount.h> instead of rolling your own.
Ok, will try. thanks.
Quoted 46 lines
Quoted 46 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +45
-0+struct hlist_node node; +}; + +struct enetc_psfp_gate { +u32 index; +s8 init_ipv; +u64 basetime; +u64 cycletime; +u64 cycletimext; +u32 num_entries; +u32 refcount; +struct hlist_node node; +struct action_gate_entry entries[0 ]; +}; + +struct enetc_stream_filter { +struct enetc_streamid sid; +u32 sfi_index; +u32 sgi_index; +struct flow_stats stats; +struct hlist_node node; +}; + +struct enetc_psfp { +unsigned long dev_bitmap; +unsigned long *psfp_sfi_bitmap; +struct hlist_head stream_list; +struct hlist_head psfp_filter_list; +struct hlist_head psfp_gate_list; +spinlock_t psfp_lock;/* spinlock for the struct enetc_psfp r/w */ +}; + +struct actions_fwd enetc_act_fwd[] = { + { +BIT (FLOW_ACTION_GATE ), +BIT (FLOW_DISSECTOR_KEY_ETH_ADDRS ), +FILTER_ACTION_TYPE_PSFP + }, +/* example for ACL actions */ + { +BIT (FLOW_ACTION_DROP ), +0 , +FILTER_ACTION_TYPE_ACL + } +};
Quoted 9 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data) · line 442
Quoted 4 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +4
-0+ +static struct enetc_psfp epsfp = { + .psfp_sfi_bitmap =NULL , +};Is it possible to have more than one these controllers in the same system? If it's possible to have more than one, this should be moved to live under a per-device struct.
There is only one Qci for one controller (here is one ENETC controller with multi ports). For each stream gate/flow meter entry could be assigned to any enetc port. So the enetc ports sharing this structure. psfp_sfi_bitmap means which entry is be occupied.
Quoted 288 lines
Quoted 288 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +288
-0+ +static LIST_HEAD (enetc_block_cb_list); + +static inline int enetc_get_port (struct enetc_ndev_priv *priv) +{ +return priv->si->pdev->devfn &0x7 ; +} + +/* Stream Identity Entry Set Descriptor */ +static int enetc_streamid_hw_set (struct enetc_ndev_priv *priv, +struct enetc_streamid *sid, +u8 enable) +{ +struct enetc_cbd cbd = {.cmd =0 }; +struct streamid_data *si_data; +struct streamid_conf *si_conf; +u16 data_size; +dma_addr_t dma; +int err; + +if (sid->index >= priv->psfp_cap.max_streamid) +return -EINVAL ; + +if (sid->filtertype !=STREAMID_TYPE_NULL && + sid->filtertype !=STREAMID_TYPE_SMAC ) +return -EOPNOTSUPP ; + +/* Disable operation before enable */ + cbd.index =cpu_to_le16 ((u16 )sid->index); + cbd.cls =BDCR_CMD_STREAM_IDENTIFY ; + cbd.status_flags =0 ; + + data_size =sizeof (struct streamid_data ); + si_data =kzalloc (data_size, __GFP_DMA |GFP_KERNEL ); + cbd.length =cpu_to_le16 (data_size); + + dma =dma_map_single (&priv->si->pdev->dev, si_data, + data_size,DMA_FROM_DEVICE ); +if (dma_mapping_error (&priv->si->pdev->dev, dma)) { +netdev_err (priv->si->ndev,"DMA mapping failed!\n" ); +kfree (si_data); +return -ENOMEM ; + } + + cbd.addr[0 ] =lower_32_bits (dma); + cbd.addr[1 ] =upper_32_bits (dma); +memset (si_data->dmac,0xff ,ETH_ALEN ); + si_data->vid_vidm_tg = +cpu_to_le16 (ENETC_CBDR_SID_VID_MASK + + ((0x3 <<14 ) |ENETC_CBDR_SID_VIDM )); + + si_conf = &cbd.sid_set; +/* Only one port supported for one entry, set itself */ + si_conf->iports =1 <<enetc_get_port (priv); + si_conf->id_type =1 ; + si_conf->oui[2 ] =0x0 ; + si_conf->oui[1 ] =0x80 ; + si_conf->oui[0 ] =0xC2 ; + + err =enetc_send_cmd (priv->si, &cbd); +if (err) +return -EINVAL ; + +if (!enable) { +kfree (si_data); +return 0 ; + } + +/* Enable the entry overwrite again incase space flushed by hardware */ +memset (&cbd,0 ,sizeof (cbd)); + + cbd.index =cpu_to_le16 ((u16 )sid->index); + cbd.cmd =0 ; + cbd.cls =BDCR_CMD_STREAM_IDENTIFY ; + cbd.status_flags =0 ; + + si_conf->en =0x80 ; + si_conf->stream_handle =cpu_to_le32 (sid->handle); + si_conf->iports =1 <<enetc_get_port (priv); + si_conf->id_type = sid->filtertype; + si_conf->oui[2 ] =0x0 ; + si_conf->oui[1 ] =0x80 ; + si_conf->oui[0 ] =0xC2 ; + +memset (si_data,0 , data_size); + + cbd.length =cpu_to_le16 (data_size); + + cbd.addr[0 ] =lower_32_bits (dma); + cbd.addr[1 ] =upper_32_bits (dma); + +/* VIDM default to be 1. +* VID Match. If set (b1) then the VID must match, otherwise +* any VID is considered a match. VIDM setting is only used +* when TG is set to b01. +*/ +if (si_conf->id_type ==STREAMID_TYPE_NULL ) { +ether_addr_copy (si_data->dmac, sid->dst_mac); + si_data->vid_vidm_tg = +cpu_to_le16 ((sid->vid &ENETC_CBDR_SID_VID_MASK ) + + ((((u16)(sid->tagged) &0x3 ) <<14 ) + |ENETC_CBDR_SID_VIDM )); + }else if (si_conf->id_type ==STREAMID_TYPE_SMAC ) { +ether_addr_copy (si_data->smac, sid->src_mac); + si_data->vid_vidm_tg = +cpu_to_le16 ((sid->vid &ENETC_CBDR_SID_VID_MASK ) + + ((((u16)(sid->tagged) &0x3 ) <<14 ) + |ENETC_CBDR_SID_VIDM )); + } + + err =enetc_send_cmd (priv->si, &cbd); +kfree (si_data); + +return err; +} + +/* Stream Filter Instance Set Descriptor */ +static int enetc_streamfilter_hw_set (struct enetc_ndev_priv *priv, +struct enetc_psfp_filter *sfi, +u8 enable) +{ +struct enetc_cbd cbd = {.cmd =0 }; +struct sfi_conf *sfi_config; + + cbd.index =cpu_to_le16 (sfi->index); + cbd.cls =BDCR_CMD_STREAM_FILTER ; + cbd.status_flags =0x80 ; + cbd.length =cpu_to_le16 (1 ); + + sfi_config = &cbd.sfi_conf; +if (!enable) + goto exit; + + sfi_config->en =0x80 ; + +if (sfi->handle >=0 ) { + sfi_config->stream_handle = +cpu_to_le32 (sfi->handle); + sfi_config->sthm |=0x80 ; + } + + sfi_config->sg_inst_table_index =cpu_to_le16 (sfi->gate_id); + sfi_config->input_ports =1 <<enetc_get_port (priv); + +/* The priority value which may be matched against the +* frame’s priority value to determine a match for this entry. +*/ +if (sfi->prio >=0 ) + sfi_config->multi |= (sfi->prio &0x7 ) |0x8 ; + +/* Filter Type. Identifies the contents of the MSDU/FM_INST_INDEX +* field as being either an MSDU value or an index into the Flow +* Meter Instance table. +* TODO: no limit max sdu +*/ + +if (sfi->meter_id >=0 ) { + sfi_config->fm_inst_table_index =cpu_to_le16 (sfi->meter_id); + sfi_config->multi |=0x80 ; + } + +exit: +return enetc_send_cmd (priv->si, &cbd); +} + +static int enetc_streamcounter_hw_get (struct enetc_ndev_priv *priv, +u32 index, +struct psfp_streamfilter_counters *cnt) +{ +struct enetc_cbd cbd = { .cmd =2 }; +struct sfi_counter_data *data_buf; +dma_addr_t dma; +u16 data_size; +int err; + + cbd.index =cpu_to_le16 ((u16 )index); + cbd.cmd =2 ; + cbd.cls =BDCR_CMD_STREAM_FILTER ; + cbd.status_flags =0 ; + + data_size =sizeof (struct sfi_counter_data ); + data_buf =kzalloc (data_size, __GFP_DMA |GFP_KERNEL ); +if (!data_buf) +return -ENOMEM ; + + dma =dma_map_single (&priv->si->pdev->dev, data_buf, + data_size,DMA_FROM_DEVICE ); +if (dma_mapping_error (&priv->si->pdev->dev, dma)) { +netdev_err (priv->si->ndev,"DMA mapping failed!\n" ); + err = -ENOMEM ; + goto exit; + } + cbd.addr[0 ] =lower_32_bits (dma); + cbd.addr[1 ] =upper_32_bits (dma); + + cbd.length =cpu_to_le16 (data_size); + + err =enetc_send_cmd (priv->si, &cbd); +if (err) + goto exit; + + cnt->matching_frames_count = + ((u64 )le32_to_cpu (data_buf->matchh) <<32 ) + + data_buf->matchl; + + cnt->not_passing_sdu_count = + ((u64 )le32_to_cpu (data_buf->msdu_droph) <<32 ) + + data_buf->msdu_dropl; + + cnt->passing_sdu_count = cnt->matching_frames_count + - cnt->not_passing_sdu_count; + + cnt->not_passing_frames_count = + ((u64 )le32_to_cpu (data_buf->stream_gate_droph) <<32 ) + +le32_to_cpu (data_buf->stream_gate_dropl); + + cnt->passing_frames_count = cnt->matching_frames_count + - cnt->not_passing_sdu_count + - cnt->not_passing_frames_count; + + cnt->red_frames_count = + ((u64 )le32_to_cpu (data_buf->flow_meter_droph) <<32 ) + +le32_to_cpu (data_buf->flow_meter_dropl); + +exit: +kfree (data_buf); +return err; +} + +static int get_start_ns (struct enetc_ndev_priv *priv,u64 cycle,u64 *start) +{ +u64 now_lo, now_hi, now, n; + + now_lo =enetc_rd (&priv->si->hw,ENETC_SICTR0 ); + now_hi =enetc_rd (&priv->si->hw,ENETC_SICTR1 ); + now = now_lo | now_hi <<32 ; + +if (WARN_ON (!cycle)) +return -EFAULT ; + + n =div64_u64 (now, cycle); + + *start = (n +1 ) * cycle; + +return 0 ; +} + +/* Stream Gate Instance Set Descriptor */ +static int enetc_streamgate_hw_set (struct enetc_ndev_priv *priv, +struct enetc_psfp_gate *sgi, +u8 enable) +{ +struct enetc_cbd cbd = { .cmd =0 }; +struct sgi_table *sgi_config; +struct sgcl_conf *sgcl_config; +struct sgcl_data *sgcl_data; +struct sgce *sgce; +dma_addr_t dma; +u16 data_size; +int err, i; + + cbd.index =cpu_to_le16 (sgi->index); + cbd.cmd =0 ; + cbd.cls =BDCR_CMD_STREAM_GCL ; + cbd.status_flags =0x80 ; + +/* disable */ +if (!enable) +return enetc_send_cmd (priv->si, &cbd); + +/* enable */ + sgi_config = &cbd.sgi_table; + +/* Keep open before gate list start */ + sgi_config->ocgtst =0x80 ; + + sgi_config->oipv = (sgi->init_ipv <0 ) ? +0x0 : ((sgi->init_ipv &0x7 ) |0x8 ); + + sgi_config->en =0x80 ; + +/* Basic config */ + err =enetc_send_cmd (priv->si, &cbd); +if (err) +return -EINVAL ; + +if (!sgi->num_entries) +return 0 ;
Quoted 8 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data) · line 734
Quoted 4 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +4
-0+ +if (sgi->num_entries > priv->psfp_cap.max_psfp_gatelist || + !sgi->cycletime) +return -EINVAL ;You already check this when "sgi" is created, this check doesn't seem to be needed. If it's needed, it should be moved to the top of the function.
Yes, should be move to upper after disable operation since sgi entries is not set when disable.
Quoted 143 lines
Quoted 143 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +142
-0+ +memset (&cbd,0 ,sizeof (cbd)); + + cbd.index =cpu_to_le16 (sgi->index); + cbd.cmd =1 ; + cbd.cls =BDCR_CMD_STREAM_GCL ; + cbd.status_flags =0 ; + + sgcl_config = &cbd.sgcl_conf; + + sgcl_config->acl_len = (sgi->num_entries -1 ) &0x3 ; + + data_size =struct_size (sgcl_data, sgcl, sgi->num_entries); + + sgcl_data =kzalloc (data_size, __GFP_DMA |GFP_KERNEL ); +if (!sgcl_data) +return -ENOMEM ; + + cbd.length =cpu_to_le16 (data_size); + + dma =dma_map_single (&priv->si->pdev->dev, + sgcl_data, data_size, +DMA_FROM_DEVICE ); +if (dma_mapping_error (&priv->si->pdev->dev, dma)) { +netdev_err (priv->si->ndev,"DMA mapping failed!\n" ); +kfree (sgcl_data); +return -ENOMEM ; + } + + cbd.addr[0 ] =lower_32_bits (dma); + cbd.addr[1 ] =upper_32_bits (dma); + + sgce = &sgcl_data->sgcl[0 ]; + + sgcl_config->agtst =0x80 ; + + sgcl_data->ct =cpu_to_le32 (sgi->cycletime); + sgcl_data->cte =cpu_to_le32 (sgi->cycletimext); + +if (sgi->init_ipv >=0 ) + sgcl_config->aipv = (sgi->init_ipv &0x7 ) |0x8 ; + +for (i =0 ; i < sgi->num_entries; i++) { +struct action_gate_entry *from = &sgi->entries[i]; +struct sgce *to = &sgce[i]; + +if (from->gate_state) + to->multi |=0x10 ; + +if (from->ipv >=0 ) + to->multi |= ((from->ipv &0x7 ) <<5 ) |0x08 ; + +if (from->maxoctets) + to->multi |=0x01 ; + + to->interval =cpu_to_le32 (from->interval); + to->msdu[0 ] = from->maxoctets &0xFF ; + to->msdu[1 ] = (from->maxoctets >>8 ) &0xFF ; + to->msdu[2 ] = (from->maxoctets >>16 ) &0xFF ; + } + +/* If basetime is 0, calculate start time */ +if (!sgi->basetime) { +u64 start; + + err =get_start_ns (priv, sgi->cycletime, &start); +if (err) + goto exit; + sgcl_data->btl =cpu_to_le32 (lower_32_bits (start)); + sgcl_data->bth =cpu_to_le32 (upper_32_bits (start)); + }else { +u32 hi, lo; + + hi =upper_32_bits (sgi->basetime); + lo =lower_32_bits (sgi->basetime); + sgcl_data->bth =cpu_to_le32 (hi); + sgcl_data->btl =cpu_to_le32 (lo); + } + + err =enetc_send_cmd (priv->si, &cbd); + +exit: +kfree (sgcl_data); + +return err; +} + +static struct enetc_stream_filter *enetc_get_stream_by_index (u32 index) +{ +struct enetc_stream_filter *f; + +hlist_for_each_entry (f, &epsfp.stream_list, node) +if (f->sid.index == index) +return f; + +return NULL ; +} + +static struct enetc_psfp_gate *enetc_get_gate_by_index (u32 index) +{ +struct enetc_psfp_gate *g; + +hlist_for_each_entry (g, &epsfp.psfp_gate_list, node) +if (g->index == index) +return g; + +return NULL ; +} + +static struct enetc_psfp_filter *enetc_get_filter_by_index (u32 index) +{ +struct enetc_psfp_filter *s; + +hlist_for_each_entry (s, &epsfp.psfp_filter_list, node) +if (s->index == index) +return s; + +return NULL ; +} + +static struct enetc_psfp_filter + *enetc_psfp_check_sfi (struct enetc_psfp_filter *sfi) +{ +struct enetc_psfp_filter *s; + +hlist_for_each_entry (s, &epsfp.psfp_filter_list, node) +if (s->gate_id == sfi->gate_id && + s->prio == sfi->prio && + s->meter_id == sfi->meter_id) +return s; + +return NULL ; +} + +static int enetc_get_free_index (struct enetc_ndev_priv *priv) +{ +u32 max_size = priv->psfp_cap.max_psfp_filter; +unsigned long index; + + index =find_first_zero_bit (epsfp.psfp_sfi_bitmap, max_size); +if (index == max_size) +return -1 ;
Quoted 8 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data) · line 881
Quoted 5 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +5
-0+ +return index; +} + +static void reduce_ref_sfi (struct enetc_ndev_priv *priv,u32 index)Please use the more usual idioms, something like "
sfi_unref()" or "stream_filter_unref".
Ok. make sense.
Quoted 251 lines
Quoted 63 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +62
-0+{ +struct enetc_psfp_filter *sfi; + + sfi =enetc_get_filter_by_index (index); +WARN_ON (!sfi); + sfi->refcount--; + +if (!sfi->refcount) { +enetc_streamfilter_hw_set (priv, sfi, false); +hlist_del (&sfi->node); +kfree (sfi); +clear_bit (sfi->index, epsfp.psfp_sfi_bitmap); + } +} + +static void reduce_ref_sgi (struct enetc_ndev_priv *priv,u32 index) +{ +struct enetc_psfp_gate *sgi; + + sgi =enetc_get_gate_by_index (index); +WARN_ON (!sgi); + sgi->refcount--; + +if (!sgi->refcount) { +enetc_streamgate_hw_set (priv, sgi, false); +hlist_del (&sgi->node); +kfree (sgi); + } +} + +static void remove_one_chain (struct enetc_ndev_priv *priv, +struct enetc_stream_filter *filter) +{ +reduce_ref_sgi (priv, filter->sgi_index); +reduce_ref_sfi (priv, filter->sfi_index); + +hlist_del (&filter->node); +kfree (filter); +} + +static int enetc_psfp_hw_set (struct enetc_ndev_priv *priv, +struct enetc_streamid *sid, +struct enetc_psfp_filter *sfi, +struct enetc_psfp_gate *sgi) +{ +int err; + + err =enetc_streamid_hw_set (priv, sid, true); +if (err) +return err; + +if (sfi) { + err =enetc_streamfilter_hw_set (priv, sfi, true); +if (err) + goto revert_sid; + } + + err =enetc_streamgate_hw_set (priv, sgi, true); +if (err) + goto revert_sfi; + +return 0 ;drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data)
Quoted 10 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +10
-0+ +revert_sfi: +if (sfi && !sfi->refcount) +enetc_streamfilter_hw_set (priv, sfi, false); +revert_sid: +enetc_streamid_hw_set (priv, sid, false); +return err; +} + +struct actions_fwd *enetc_check_flow_actions(u64 acts, unsigned intinputkeys)
Quoted 49 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +49
-0+{ +int i; + +for (i =0 ; i <ARRAY_SIZE (enetc_act_fwd); i++) +if (acts == enetc_act_fwd[i].actions && + inputkeys & enetc_act_fwd[i].keys) +return &enetc_act_fwd[i]; + +return NULL ; +} + +static int enetc_psfp_parse_clsflower (struct enetc_ndev_priv *priv, +struct flow_cls_offload *f) +{ +struct flow_rule *rule =flow_cls_offload_flow_rule (f); +struct netlink_ext_ack *extack = f->common.extack; +struct enetc_stream_filter *filter, *old_filter; +struct enetc_psfp_filter *sfi, *old_sfi; +struct enetc_psfp_gate *sgi, *old_sgi; +struct flow_action_entry *entry; +struct action_gate_entry *e; +u8 sfi_overwrite =0 ; +int entries_size; +int i, err; + +if (f->common.chain_index >= priv->psfp_cap.max_streamid) { +NL_SET_ERR_MSG_MOD (extack,"No Stream identify resource!" ); +return -ENOSPC ; + } + +flow_action_for_each (i, entry, &rule->action) +if (entry->id ==FLOW_ACTION_GATE ) +break ; + +if (entry->id !=FLOW_ACTION_GATE ) +return -EINVAL ; + + filter =kzalloc (sizeof (*filter),GFP_KERNEL ); +if (!filter) +return -ENOMEM ; + + filter->sid.index = f->common.chain_index; + +if (flow_rule_match_key (rule,FLOW_DISSECTOR_KEY_ETH_ADDRS )) { +struct flow_match_eth_addrs match; + +flow_rule_match_eth_addrs (rule, &match); + +if (!is_zero_ether_addr (match.mask->dst)) {drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data)
Quoted 10 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +10
-0+ether_addr_copy (filter->sid.dst_mac, match.key->dst); + filter->sid.filtertype =STREAMID_TYPE_NULL ; + } + +if (!is_zero_ether_addr (match.mask->src)) { +ether_addr_copy (filter->sid.src_mac, match.key->src); + filter->sid.filtertype =STREAMID_TYPE_SMAC ; + } + }else { + NL_SET_ERR_MSG_MOD(extack, "Unsupported, must
ETH_ADDRS");drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data)
Quoted 11 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +11
-0+return -EINVAL ; + } + +if (flow_rule_match_key (rule,FLOW_DISSECTOR_KEY_VLAN )) { +struct flow_match_vlan match; + +flow_rule_match_vlan (rule, &match); +if (match.mask->vlan_priority) { +if (match.mask->vlan_priority != + (VLAN_PRIO_MASK >>VLAN_PRIO_SHIFT )) { + NL_SET_ERR_MSG_MOD(extack, "Only full mask issupported for VLAN priority");
drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data)
Quoted 8 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +8
-0+ err = -EINVAL ; + goto free_filter; + } + } + +if (match.mask->vlan_tpid) { +if (match.mask->vlan_tpid !=VLAN_VID_MASK ) { + NL_SET_ERR_MSG_MOD(extack, "Only full mask issupported for VLAN id");
Quoted 96 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +96
-0+ err = -EINVAL ; + goto free_filter; + } + + filter->sid.vid = match.key->vlan_tpid; +if (!filter->sid.vid) + filter->sid.tagged =STREAMID_VLAN_UNTAGGED ; +else + filter->sid.tagged =STREAMID_VLAN_TAGGED ; + } + }else { + filter->sid.tagged =STREAMID_VLAN_ALL ; + } + +/* parsing gate action */ +if (entry->gate.index >= priv->psfp_cap.max_psfp_gate) { +NL_SET_ERR_MSG_MOD (extack,"No Stream Gate resource!" ); + err = -ENOSPC ; + goto free_filter; + } + +if (entry->gate.num_entries >= priv->psfp_cap.max_psfp_gatelist) { +NL_SET_ERR_MSG_MOD (extack,"No Stream Gate resource!" ); + err = -ENOSPC ; + goto free_filter; + } + + entries_size =struct_size (sgi, entries, entry->gate.num_entries); + sgi =kzalloc (entries_size,GFP_KERNEL ); +if (!sgi) { + err = -ENOMEM ; + goto free_filter; + } + + sgi->index = entry->gate.index; + sgi->init_ipv = entry->gate.prio; + sgi->basetime = entry->gate.basetime; + sgi->cycletime = entry->gate.cycletime; + sgi->num_entries = entry->gate.num_entries; + + e = sgi->entries; +for (i =0 ; i < entry->gate.num_entries; i++) { + e[i].gate_state = entry->gate.entries[i].gate_state; + e[i].interval = entry->gate.entries[i].interval; + e[i].ipv = entry->gate.entries[i].ipv; + e[i].maxoctets = entry->gate.entries[i].maxoctets; + } + + filter->sgi_index = sgi->index; + + sfi =kzalloc (sizeof (*sfi),GFP_KERNEL ); +if (!sfi) { + err = -ENOMEM ; + goto free_gate; + } + + sfi->gate_id = sgi->index; + +/* flow meter not support yet */ + sfi->meter_id =ENETC_PSFP_WILDCARD ; + +/* prio ref the filter prio */ +if (f->common.prio && f->common.prio <=BIT (3 )) + sfi->prio = f->common.prio -1 ; +else + sfi->prio =ENETC_PSFP_WILDCARD ; + + old_sfi =enetc_psfp_check_sfi (sfi); +if (!old_sfi) { +int index; + + index =enetc_get_free_index (priv); +if (sfi->handle <0 ) { +NL_SET_ERR_MSG_MOD (extack,"No Stream Filter resource!" ); + err = -ENOSPC ; + goto free_sfi; + } + + sfi->index = index; + sfi->handle = index +HANDLE_OFFSET ; +/* Update the stream filter handle also */ + filter->sid.handle = sfi->handle; + filter->sfi_index = sfi->index; + sfi_overwrite =0 ; + }else { + filter->sfi_index = old_sfi->index; + filter->sid.handle = old_sfi->handle; + sfi_overwrite =1 ; + } + + err =enetc_psfp_hw_set (priv, &filter->sid, + sfi_overwrite ?NULL : sfi, sgi); +if (err) + goto free_sfi; + +spin_lock (&epsfp.psfp_lock);
Quoted 8 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data) · line 1132
Quoted 5 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +5
-0+ old_sgi =enetc_get_gate_by_index (filter->sgi_index); +if (old_sgi) { + sgi->refcount = old_sgi->refcount; +hlist_del (&old_sgi->node); +kfree (old_sgi);I don't understand what you are trying to achieve here. But there should be a cleaner way.
The code here is to remove the old stream gate instance in list. What ever it is existed, will link a new list to replace the old setting but still need the bind count. Since the old filter may different with new gate instance, from user view, should know they are setting new one with same index, and would replaced by new gate setting. But old filter still working. I would check again for improvement if it can be more clear.
Quoted 270 lines
Quoted 270 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +269
-0+ } + + sgi->refcount++; +hlist_add_head (&sgi->node, &epsfp.psfp_gate_list); + +if (!old_sfi) { +hlist_add_head (&sfi->node, &epsfp.psfp_filter_list); +set_bit (sfi->index, epsfp.psfp_sfi_bitmap); + sfi->refcount++; + }else { +kfree (sfi); + old_sfi->refcount++; + } + + old_filter =enetc_get_stream_by_index (filter->sid.index); +if (old_filter) +remove_one_chain (priv, old_filter); + + filter->stats.lastused = jiffies; +hlist_add_head (&filter->node, &epsfp.stream_list); + +spin_unlock (&epsfp.psfp_lock); + +return 0 ; + +free_sfi: +kfree (sfi); +free_gate: +kfree (sgi); +free_filter: +kfree (filter); + +return err; +} + +static int enetc_config_clsflower (struct enetc_ndev_priv *priv, +struct flow_cls_offload *cls_flower) +{ +struct flow_rule *rule =flow_cls_offload_flow_rule (cls_flower); +struct netlink_ext_ack *extack = cls_flower->common.extack; +struct flow_dissector *dissector = rule->match.dissector; +struct flow_action *action = &rule->action; +struct flow_action_entry *entry; +struct actions_fwd *fwd; +u64 actions =0 ; +int i, err; + +if (!flow_action_has_entries (action)) { +NL_SET_ERR_MSG_MOD (extack,"At least one action is needed" ); +return -EINVAL ; + } + +flow_action_for_each (i, entry, action) + actions |=BIT (entry->id); + + fwd =enetc_check_flow_actions (actions, dissector->used_keys); +if (!fwd) { +NL_SET_ERR_MSG_MOD (extack,"Unsupported filter type!" ); +return -EOPNOTSUPP ; + } + +if (fwd->output &FILTER_ACTION_TYPE_PSFP ) { + err =enetc_psfp_parse_clsflower (priv, cls_flower); +if (err) { +NL_SET_ERR_MSG_MOD (extack,"Invalid PSFP inputs" ); +return err; + } + }else { +NL_SET_ERR_MSG_MOD (extack,"Unsupported actions" ); +return -EOPNOTSUPP ; + } + +return 0 ; +} + +static int enetc_psfp_destroy_clsflower (struct enetc_ndev_priv *priv, +struct flow_cls_offload *f) +{ +struct enetc_stream_filter *filter; +struct netlink_ext_ack *extack = f->common.extack; +int err; + +if (f->common.chain_index >= priv->psfp_cap.max_streamid) { +NL_SET_ERR_MSG_MOD (extack,"No Stream identify resource!" ); +return -ENOSPC ; + } + + filter =enetc_get_stream_by_index (f->common.chain_index); +if (!filter) +return -EINVAL ; + + err =enetc_streamid_hw_set (priv, &filter->sid, false); +if (err) +return err; + +remove_one_chain (priv, filter); + +return 0 ; +} + +static int enetc_destroy_clsflower (struct enetc_ndev_priv *priv, +struct flow_cls_offload *f) +{ +return enetc_psfp_destroy_clsflower (priv, f); +} + +static int enetc_psfp_get_stats (struct enetc_ndev_priv *priv, +struct flow_cls_offload *f) +{ +struct psfp_streamfilter_counters counters = {}; +struct enetc_stream_filter *filter; +struct flow_stats stats = {}; +int err; + + filter =enetc_get_stream_by_index (f->common.chain_index); +if (!filter) +return -EINVAL ; + + err =enetc_streamcounter_hw_get (priv, filter->sfi_index, &counters); +if (err) +return -EINVAL ; + +spin_lock (&epsfp.psfp_lock); + stats.pkts = counters.matching_frames_count - filter->stats.pkts; + stats.dropped = counters.not_passing_frames_count - + filter->stats.dropped; + stats.lastused = filter->stats.lastused; + filter->stats.pkts += stats.pkts; + filter->stats.dropped += stats.dropped; + +spin_unlock (&epsfp.psfp_lock); + +flow_stats_update (&f->stats,0x0 , stats.pkts, stats.lastused, + stats.dropped); + +return 0 ; +} + +static int enetc_setup_tc_cls_flower (struct enetc_ndev_priv *priv, +struct flow_cls_offload *cls_flower) +{ +switch (cls_flower->command) { +case FLOW_CLS_REPLACE : +return enetc_config_clsflower (priv, cls_flower); +case FLOW_CLS_DESTROY : +return enetc_destroy_clsflower (priv, cls_flower); +case FLOW_CLS_STATS : +return enetc_psfp_get_stats (priv, cls_flower); +default : +return -EOPNOTSUPP ; + } +} + +static inline void clean_psfp_sfi_bitmap (void ) +{ +bitmap_free (epsfp.psfp_sfi_bitmap); + epsfp.psfp_sfi_bitmap =NULL ; +} + +static void clean_stream_list (void ) +{ +struct enetc_stream_filter *s; +struct hlist_node *tmp; + +hlist_for_each_entry_safe (s, tmp, &epsfp.stream_list, node) { +hlist_del (&s->node); +kfree (s); + } +} + +static void clean_sfi_list (void ) +{ +struct enetc_psfp_filter *sfi; +struct hlist_node *tmp; + +hlist_for_each_entry_safe (sfi, tmp, &epsfp.psfp_filter_list, node) { +hlist_del (&sfi->node); +kfree (sfi); + } +} + +static void clean_sgi_list (void ) +{ +struct enetc_psfp_gate *sgi; +struct hlist_node *tmp; + +hlist_for_each_entry_safe (sgi, tmp, &epsfp.psfp_gate_list, node) { +hlist_del (&sgi->node); +kfree (sgi); + } +} + +static void clean_psfp_all (void ) +{ +/* Disable all list nodes and free all memory */ +clean_sfi_list (); +clean_sgi_list (); +clean_stream_list (); + epsfp.dev_bitmap =0 ; +clean_psfp_sfi_bitmap (); +} + +int enetc_setup_tc_block_cb (enum tc_setup_type type,void *type_data, +void *cb_priv) +{ +struct net_device *ndev = cb_priv; + +if (!tc_can_offload (ndev)) +return -EOPNOTSUPP ; + +switch (type) { +case TC_SETUP_CLSFLOWER : +return enetc_setup_tc_cls_flower (netdev_priv (ndev), type_data); +default : +return -EOPNOTSUPP ; + } +} + +int enetc_psfp_init (struct enetc_ndev_priv *priv) +{ +if (epsfp.psfp_sfi_bitmap) +return 0 ; + + epsfp.psfp_sfi_bitmap =bitmap_zalloc (priv->psfp_cap.max_psfp_filter, +GFP_KERNEL ); +if (!epsfp.psfp_sfi_bitmap) +return -ENOMEM ; + +spin_lock_init (&epsfp.psfp_lock); + +if (list_empty (&enetc_block_cb_list)) + epsfp.dev_bitmap =0 ; + +return 0 ; +} + +int enetc_psfp_clean (struct enetc_ndev_priv *priv) +{ +if (!list_empty (&enetc_block_cb_list)) +return -EBUSY ; + +clean_psfp_all (); + +return 0 ; +} + +int enetc_setup_tc_psfp (struct net_device *ndev,void *type_data) +{ +struct enetc_ndev_priv *priv =netdev_priv (ndev); +struct flow_block_offload *f = type_data; +int err; + + err =flow_block_cb_setup_simple (f, &enetc_block_cb_list, + enetc_setup_tc_block_cb, + ndev, ndev, true); +if (err) +return err; + +switch (f->command) { +case FLOW_BLOCK_BIND : +set_bit (enetc_get_port (priv), &epsfp.dev_bitmap); +break ; +case FLOW_BLOCK_UNBIND : +clear_bit (enetc_get_port (priv), &epsfp.dev_bitmap); +if (!epsfp.dev_bitmap) +clean_psfp_all (); +break ; + } +
Quoted 8 lines
Quoted 5 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +2
-1+return 0 ; +} --2.17.1
--
Vinicius
Br,
Po Liu
RFC 6/9 net: qos: add tc police offloading action with max frame size limit
In reply to Po LiuCurrent police offloading support the 'burst'' and 'rate_bytes_ps'. Some hardware own the capability to limit the frame size. If the frame size larger than the setting, the frame would be dropped. For the police action itself already accept the 'mtu' parameter in tc command. But not extend to tc flower offloading. So extend 'mtu' to tc flower offloading.
--- include/net/flow_offload.h | 1 + include/net/tc_act/tc_police.h | 10 ++++++++++ net/sched/cls_api.c | 1 + 3 files changed, 12 insertions(+)
include/net/flow_offload.h +1 -0
@@ -203,6 +203,7 @@ struct flow_action_entry {struct {/* FLOW_ACTION_POLICE */ s64 burst;u64 rate_bytes_ps; +u32 mtu; } police;struct {/* FLOW_ACTION_CT */ int action;
include/net/tc_act/tc_police.h +10 -0
@@ -69,4 +69,14 @@ static inline s64 tcf_police_tcfp_burst(const struct tc_action *act) return params->tcfp_burst; } +static inline u32 tcf_police_mtu (const struct tc_action *act) +{ +struct tcf_police *police =to_police (act); +struct tcf_police_params *params; + + params =rcu_dereference_protected (police->params, +lockdep_is_held (&police->tcf_lock)); + +return params->tcfp_mtu; +}#endif /* __NET_TC_POLICE_H */
net/sched/cls_api.c +1 -0
@@ -3583,6 +3583,7 @@ int tc_setup_flow_action(struct flow_action *flow_action, entry->police.burst =tcf_police_tcfp_burst (act); entry->police.rate_bytes_ps =tcf_police_rate_bytes_ps (act); + entry->police.mtu =tcf_police_mtu (act); }else if (is_tcf_ct (act)) { entry->id =FLOW_ACTION_CT ; entry->ct.action =tcf_ct_action (act);
Signature · 1 line
2.17.1
PATCH 1/4 net: qos: add tc police offloading action with max frame size limit
From: Po Liu <[email protected]>
Current police offloading support the 'burst'' and 'rate_bytes_ps'. Some hardware own the capability to limit the frame size. If the frame size larger than the setting, the frame would be dropped. For the police action itself already accept the 'mtu' parameter in tc command. But not extend to tc flower offloading. So extend 'mtu' to tc flower offloading.
---
continue the thread [email protected] for the police action offloading.
include/net/flow_offload.h | 1 + include/net/tc_act/tc_police.h | 10 ++++++++++ net/sched/cls_api.c | 1 + 3 files changed, 12 insertions(+)
include/net/flow_offload.h +1 -0
@@ -234,6 +234,7 @@ struct flow_action_entry {struct {/* FLOW_ACTION_POLICE */ s64 burst;u64 rate_bytes_ps; +u32 mtu; } police;struct {/* FLOW_ACTION_CT */ int action;
include/net/tc_act/tc_police.h +10 -0
@@ -69,4 +69,14 @@ static inline s64 tcf_police_tcfp_burst(const struct tc_action *act) return params->tcfp_burst; } +static inline u32 tcf_police_tcfp_mtu (const struct tc_action *act) +{ +struct tcf_police *police =to_police (act); +struct tcf_police_params *params; + + params =rcu_dereference_protected (police->params, +lockdep_is_held (&police->tcf_lock)); +return params->tcfp_mtu; +} +#endif /* __NET_TC_POLICE_H */
net/sched/cls_api.c +1 -0
@@ -3658,6 +3658,7 @@ int tc_setup_flow_action(struct flow_action *flow_action, entry->police.burst =tcf_police_tcfp_burst (act); entry->police.rate_bytes_ps =tcf_police_rate_bytes_ps (act); + entry->police.mtu =tcf_police_tcfp_mtu (act); }else if (is_tcf_ct (act)) { entry->id =FLOW_ACTION_CT ; entry->ct.action =tcf_ct_action (act);
Signature · 1 line
2.17.1
PATCH 2/4 net: enetc: add support max frame size for tc flower offload
From: Po Liu <[email protected]>
Base on the tc flower offload police action add max frame size by the parameter 'mtu'. Tc flower device driver working by the IEEE 802.1Qci stream filter can implement the max frame size filtering. Add it to the current hardware tc flower stearm filter driver.
--- .../net/ethernet/freescale/enetc/enetc_qos.c | 52 +++++++++++++------ 1 file changed, 36 insertions(+), 16 deletions(-)
drivers/net/ethernet/freescale/enetc/enetc_qos.c +36 -16
@@ -389,6 +389,7 @@ struct enetc_psfp_filter {u32 index;s32 handle;s8 prio; +u32 maxsdu;u32 gate_id;s32 meter_id;refcount_t refcount;
@@ -430,6 +431,12 @@ static struct actions_fwd enetc_act_fwd[] = {BIT (FLOW_DISSECTOR_KEY_ETH_ADDRS ),FILTER_ACTION_TYPE_PSFP }, + { +BIT (FLOW_ACTION_POLICE ) | +BIT (FLOW_ACTION_GATE ), +BIT (FLOW_DISSECTOR_KEY_ETH_ADDRS ), +FILTER_ACTION_TYPE_PSFP + },/* example for ACL actions */ { BIT(FLOW_ACTION_DROP),
@@ -594,8 +601,12 @@ static int enetc_streamfilter_hw_set(struct enetc_ndev_priv *priv,/* Filter Type. Identifies the contents of the MSDU/FM_INST_INDEX * field as being either an MSDU value or an index into the Flow * Meter Instance table. -* TODO: no limit max sdu */ +if (sfi->maxsdu) { + sfi_config->msdu = +cpu_to_le16 (sfi->maxsdu); + sfi_config->multi |=0x40 ; + }if (sfi->meter_id >=0 ) { sfi_config->fm_inst_table_index =cpu_to_le16 (sfi->meter_id);
@@ -872,6 +883,7 @@ static struct enetc_psfp_filterhlist_for_each_entry (s, &epsfp.psfp_filter_list, node)if (s->gate_id == sfi->gate_id && s->prio == sfi->prio && + s->maxsdu == sfi->maxsdu && s->meter_id == sfi->meter_id)return s;
@@ -979,6 +991,7 @@ static struct actions_fwd *enetc_check_flow_actions(u64 acts,static int enetc_psfp_parse_clsflower (struct enetc_ndev_priv *priv,struct flow_cls_offload *f) { +struct flow_action_entry *entryg =NULL , *entryp =NULL ;struct flow_rule *rule =flow_cls_offload_flow_rule (f);struct netlink_ext_ack *extack = f->common.extack;struct enetc_stream_filter *filter, *old_filter;
@@ -997,9 +1010,12 @@ static int enetc_psfp_parse_clsflower(struct enetc_ndev_priv *priv,flow_action_for_each (i, entry, &rule->action)if (entry->id ==FLOW_ACTION_GATE ) -break ; + entryg = entry; +else if (entry->id ==FLOW_ACTION_POLICE ) + entryp = entry; -if (entry->id !=FLOW_ACTION_GATE ) +/* Not support without gate action */ +if (!entryg)return -EINVAL ; filter =kzalloc (sizeof (*filter),GFP_KERNEL );
@@ -1079,19 +1095,19 @@ static int enetc_psfp_parse_clsflower(struct enetc_ndev_priv *priv, }/* parsing gate action */ -if (entry->gate.index >= priv->psfp_cap.max_psfp_gate) { +if (entryg->gate.index >= priv->psfp_cap.max_psfp_gate) {NL_SET_ERR_MSG_MOD (extack,"No Stream Gate resource!" ); err = -ENOSPC ; goto free_filter; } -if (entry->gate.num_entries >= priv->psfp_cap.max_psfp_gatelist) { +if (entryg->gate.num_entries >= priv->psfp_cap.max_psfp_gatelist) {NL_SET_ERR_MSG_MOD (extack,"No Stream Gate resource!" ); err = -ENOSPC ; goto free_filter; } - entries_size =struct_size (sgi, entries, entry->gate.num_entries); + entries_size =struct_size (sgi, entries, entryg->gate.num_entries); sgi =kzalloc (entries_size,GFP_KERNEL );if (!sgi) { err = -ENOMEM ;
@@ -1099,18 +1115,18 @@ static int enetc_psfp_parse_clsflower(struct enetc_ndev_priv *priv, }refcount_set (&sgi->refcount,1 ); - sgi->index = entry->gate.index; - sgi->init_ipv = entry->gate.prio; - sgi->basetime = entry->gate.basetime; - sgi->cycletime = entry->gate.cycletime; - sgi->num_entries = entry->gate.num_entries; + sgi->index = entryg->gate.index; + sgi->init_ipv = entryg->gate.prio; + sgi->basetime = entryg->gate.basetime; + sgi->cycletime = entryg->gate.cycletime; + sgi->num_entries = entryg->gate.num_entries; e = sgi->entries; -for (i =0 ; i < entry->gate.num_entries; i++) { - e[i].gate_state = entry->gate.entries[i].gate_state; - e[i].interval = entry->gate.entries[i].interval; - e[i].ipv = entry->gate.entries[i].ipv; - e[i].maxoctets = entry->gate.entries[i].maxoctets; +for (i =0 ; i < entryg->gate.num_entries; i++) { + e[i].gate_state = entryg->gate.entries[i].gate_state; + e[i].interval = entryg->gate.entries[i].interval; + e[i].ipv = entryg->gate.entries[i].ipv; + e[i].maxoctets = entryg->gate.entries[i].maxoctets; } filter->sgi_index = sgi->index;
@@ -1127,6 +1143,10 @@ static int enetc_psfp_parse_clsflower(struct enetc_ndev_priv *priv,/* flow meter not support yet */ sfi->meter_id =ENETC_PSFP_WILDCARD ; +/* Max frame size */ +if (entryp) + sfi->maxsdu = entryp->police.mtu; +/* prio ref the filter prio */ if (f->common.prio && f->common.prio <=BIT (3 )) sfi->prio = f->common.prio -1 ;
Signature · 1 line
2.17.1
PATCH 3/4 net: qos: police action add index for tc flower offloading
In reply to Po LiuFrom: Po Liu <[email protected]>
Hardware may own many entries for police flow. So that make one(or
multi) flow to be policed by one hardware entry. This patch add the
police action index provide to the driver side make it mapping the driver hardware entry index.
--- include/net/flow_offload.h | 1 + net/sched/cls_api.c | 1 + 2 files changed, 2 insertions(+)
include/net/flow_offload.h +1 -0
@@ -232,6 +232,7 @@ struct flow_action_entry {bool truncate; } sample;struct {/* FLOW_ACTION_POLICE */ +u32 index;s64 burst;u64 rate_bytes_ps;u32 mtu;
net/sched/cls_api.c +1 -0
@@ -3659,6 +3659,7 @@ int tc_setup_flow_action(struct flow_action *flow_action, entry->police.rate_bytes_ps =tcf_police_rate_bytes_ps (act); entry->police.mtu =tcf_police_tcfp_mtu (act); + entry->police.index = act->tcfa_index; }else if (is_tcf_ct (act)) { entry->id =FLOW_ACTION_CT ; entry->ct.action =tcf_ct_action (act);
Signature · 1 line
2.17.1
Re: [v1,net-next 3/4] net: qos: police action add index for tc flower offloading
Quoted 6 lines · Po Liu
From: Po Liu <[email protected]>
Hardware may own many entries for police flow. So that make one(or
multi) flow to be policed by one hardware entry. This patch add the
police action index provide to the driver side make it mapping the driver hardware entry index.
Maybe first mention that it is possible for multiple filters in software to share the same policer. Something like:
"
It is possible for several tc filters to share the same police action by specifying the action's index when installing the filters.
Propagate this index to device drivers through the flow offload intermediate representation, so that drivers could share a single hardware policer between multiple filters.
"
Quoted 34 lines
Signed-off-by Po Liu <[email protected]>--- include/net/flow_offload.h | 1 + net/sched/cls_api.c | 1 + 2 files changed, 2 insertions(+)include/net/flow_offload.h +1
-0@@ -232,6 +232,7 @@ struct flow_action_entry {bool truncate; } sample;struct {/* FLOW_ACTION_POLICE */ +u32 index;s64 burst;u64 rate_bytes_ps;u32 mtu;net/sched/cls_api.c +1
-0@@ -3659,6 +3659,7 @@ int tc_setup_flow_action(struct flow_action *flow_action, entry->police.rate_bytes_ps =tcf_police_rate_bytes_ps (act); entry->police.mtu =tcf_police_tcfp_mtu (act); + entry->police.index = act->tcfa_index; }else if (is_tcf_ct (act)) { entry->id =FLOW_ACTION_CT ; entry->ct.action =tcf_ct_action (act);Signature · 1 line
2.17.1
RE: [EXT] Re: [v1,net-next 3/4] net: qos: police action add index for tc flower offloading
Hi Ido,
Quoted 76 lines
-----Original Message-----
From: Ido Schimmel <[email protected]>
Sent: 2020年6月23日 15:10
To: Po Liu <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Claudiu Manoil <[email protected]>; Vladimir Oltean <[email protected]>; Alexandru Marginean <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]
Subject: [EXT] Re: [v1,net-next 3/4] net: qos: police action add index for tc flower offloadingCaution: EXT Email
Quoted 6 lines · Po Liu
From: Po Liu <[email protected]>
Hardware may own many entries for police flow. So that make one(or
multi) flow to be policed by one hardware entry. This patch add the
police action index provide to the driver side make it mapping the driver hardware entry index.Maybe first mention that it is possible for multiple filters in software to share the same policer. Something like:
"
It is possible for several tc filters to share the same police action by specifying the action's index when installing the filters.Propagate this index to device drivers through the flow offload intermediate representation, so that drivers could share a single hardware policer between multiple filters.
"net/sched/cls_api.c int tc_setup_flow_action(struct flow_action
Quoted 24 lines
Signed-off-by Po Liu <[email protected]>--- include/net/flow_offload.h | 1 + net/sched/cls_api.c | 1 + 2 files changed, 2 insertions(+)include/net/flow_offload.h +1
-0@@ -232,6 +232,7 @@ struct flow_action_entry {bool truncate; } sample;struct {/* FLOW_ACTION_POLICE */ +u32 index;s64 burst;u64 rate_bytes_ps;u32 mtu;net/sched/cls_api.c index renamed from net/sched/cls_api.c +0
-0
6aba7d5ba1ec..fdc4c89ca1fa100644net/sched/cls_api.c +0
-0@@ -3659,6 +3659,7 @@ int tc_setup_flow_action(struct flow_action
*flow_action,Quoted 10 lines
net/sched/cls_api.c +1
-1entry->police.rate_bytes_ps =tcf_police_rate_bytes_ps (act); entry->police.mtu =tcf_police_tcfp_mtu (act); + entry->police.index = act->tcfa_index; }else if (is_tcf_ct (act)) { entry->id =FLOW_ACTION_CT ; entry->ct.action =tcf_ct_action (act); --2.17.1
Jamal Hadi Salim
maintainerRe: [v1,net-next 3/4] net: qos: police action add index for tc flower offloading
In reply to Po LiuThis certainly brings an interesting point which i brought up earlier when Jiri was doing offloading of stats.
In this case the action index is being used as the offloaded policer index (note: there'd need to be a check whether the index is infact acceptable to the h/w etc unless there 2^32 meters available in the hardware).
My question: Is this any different from how stats are structured? In this case you can map the s/w action index to a h/w table index (of meters).
My comment then was: hardware i have encountered (and i pointed to P4 model as well) assumes an indexed table of stats.
cheers,
jamal
Quoted 38 lines · Po Liu
From: Po Liu <[email protected]>
Hardware may own many entries for police flow. So that make one(or
multi) flow to be policed by one hardware entry. This patch add the
police action index provide to the driver side make it mapping the driver hardware entry index.Signed-off-by Po Liu <[email protected]>--- include/net/flow_offload.h | 1 + net/sched/cls_api.c | 1 + 2 files changed, 2 insertions(+)include/net/flow_offload.h +1
-0@@ -232,6 +232,7 @@ struct flow_action_entry {bool truncate; } sample;struct {/* FLOW_ACTION_POLICE */ +u32 index;s64 burst;u64 rate_bytes_ps;u32 mtu;net/sched/cls_api.c +1
-0@@ -3659,6 +3659,7 @@ int tc_setup_flow_action(struct flow_action *flow_action, entry->police.rate_bytes_ps =tcf_police_rate_bytes_ps (act); entry->police.mtu =tcf_police_tcfp_mtu (act); + entry->police.index = act->tcfa_index; }else if (is_tcf_ct (act)) { entry->id =FLOW_ACTION_CT ; entry->ct.action =tcf_ct_action (act);
PATCH 4/4 net: enetc add tc flower offload flow metering policing action
In reply to Po LiuFrom: Po Liu <[email protected]>
Flow metering entries in IEEE 802.1Qci is an optional function for a flow filtering module. Flow metering is two rates two buckets and three color marker to policing the frames. This patch only enable one rate one bucket and in color blind mode. Flow metering instance are as specified in the algorithm in MEF 10.3 and in Bandwidth Profile Parameters. They are:
a) Flow meter instance identifier. An integer value identifying the flow meter instance. The patch use the police 'index' as thin value.
b) Committed Information Rate (CIR), in bits per second. This patch use the 'rate_bytes_ps' represent this value.
c) Committed Burst Size (CBS), in octets. This patch use the 'burst' represent this value.
d) Excess Information Rate (EIR), in bits per second.
e) Excess Burst Size per Bandwidth Profile Flow (EBS), in octets. And plus some other parameters. This patch set EIR/EBS default disable and color blind mode.
--- .../net/ethernet/freescale/enetc/enetc_hw.h | 24 +++ .../net/ethernet/freescale/enetc/enetc_qos.c | 160 ++++++++++++++++-- 2 files changed, 172 insertions(+), 12 deletions(-)
drivers/net/ethernet/freescale/enetc/enetc_hw.h +24 -0
@@ -570,6 +570,7 @@ enum bdcr_cmd_class {BDCR_CMD_STREAM_IDENTIFY ,BDCR_CMD_STREAM_FILTER ,BDCR_CMD_STREAM_GCL , +BDCR_CMD_FLOW_METER , __BDCR_CMD_MAX_LEN, BDCR_CMD_MAX_LEN = __BDCR_CMD_MAX_LEN - 1, };
@@ -736,10 +737,33 @@ struct sgcl_data {struct sgce sgcl[0 ]; }; +#define ENETC_CBDR_FMI_MR BIT(0) +#define ENETC_CBDR_FMI_MREN BIT(1) +#define ENETC_CBDR_FMI_DOY BIT(2) +#define ENETC_CBDR_FMI_CM BIT(3) +#define ENETC_CBDR_FMI_CF BIT(4) +#define ENETC_CBDR_FMI_NDOR BIT(5) +#define ENETC_CBDR_FMI_OALEN BIT(6) +#define ENETC_CBDR_FMI_IRFPP_MASK GENMASK(4, 0) + +/* class 10: command 0/1, Flow Meter Instance Set, short Format */ +struct fmi_conf { +__le32 cir; +__le32 cbs; +__le32 eir; +__le32 ebs; +u8 conf; +u8 res1; +u8 ir_fpp; +u8 res2[4 ]; +u8 en; +}; + struct enetc_cbd { union{ struct sfi_conf sfi_conf; struct sgi_table sgi_table; + struct fmi_conf fmi_conf; struct { __le32 addr[2]; union {
drivers/net/ethernet/freescale/enetc/enetc_qos.c +148 -12
252 lines not on this page · raw message
Signature · 1 line
2.17.1
kernel test robot
build botRe: [v1,net-next 4/4] net: enetc add tc flower offload flow metering policing action
Thank you for the patch! Yet something to improve:
Hi Po,
Thank you for the patch! Yet something to improve:
[auto build test ERROR on next-20200622]
[cannot apply to sparc-next/master linux/master linus/master ipvs/master v5.8-rc2 v5.8-rc1 v5.7 v5.8-rc2] [If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Po-Liu/net-qos-add-tc-police-offloading-action-with-max-frame-size-limit/20200623-153724 base: 27f11fea33608cbd321a97cbecfa2ef97dcc1821 config: parisc-allmodconfig (attached as .config) compiler: hppa-linux-gcc (GCC) 9.3.0
reproduce (this is a W=1 build):
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross ARCH=parisc
If you fix the issue, kindly add following tag as appropriate
All errors (new ones prefixed by >>, old ones prefixed by <<):
WARNING: modpost: missing MODULE_LICENSE() in lib/test_bits.o
>> ERROR: modpost: "__udivdi3" [drivers/net/ethernet/freescale/enetc/fsl-enetc-vf.ko] undefined! >> ERROR: modpost: "__divdi3" [drivers/net/ethernet/freescale/enetc/fsl-enetc-vf.ko] undefined! >> ERROR: modpost: "__udivdi3" [drivers/net/ethernet/freescale/enetc/fsl-enetc.ko] undefined! >> ERROR: modpost: "__divdi3" [drivers/net/ethernet/freescale/enetc/fsl-enetc.ko] undefined!
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org
kernel test robot
build botRe: [v1,net-next 4/4] net: enetc add tc flower offload flow metering policing action
In reply to Po LiuThank you for the patch! Yet something to improve:
Hi Po,
Thank you for the patch! Yet something to improve:
[auto build test ERROR on next-20200622]
[cannot apply to sparc-next/master linux/master linus/master ipvs/master v5.8-rc2 v5.8-rc1 v5.7 v5.8-rc2] [If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Po-Liu/net-qos-add-tc-police-offloading-action-with-max-frame-size-limit/20200623-153724 base: 27f11fea33608cbd321a97cbecfa2ef97dcc1821 config: i386-allyesconfig (attached as .config) compiler: gcc-9 (Debian 9.3.0-13) 9.3.0 reproduce (this is a W=1 build): # save the attached .config to linux build tree make W=1 ARCH=i386
If you fix the issue, kindly add following tag as appropriate
All errors (new ones prefixed by >>):
ld: drivers/net/ethernet/freescale/enetc/enetc_qos.o: in function `enetc_flowmeter_hw_set':
>> enetc_qos.c:(.text+0x66): undefined reference to `__udivdi3'
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org
PATCH v2 1/4 net: qos: add tc police offloading action with max frame size limit
In reply to Po LiuFrom: Po Liu <[email protected]>
Current police offloading support the 'burst'' and 'rate_bytes_ps'. Some hardware own the capability to limit the frame size. If the frame size larger than the setting, the frame would be dropped. For the police action itself already accept the 'mtu' parameter in tc command. But not extend to tc flower offloading. So extend 'mtu' to tc flower offloading.
---
v2:
-- No update.
include/net/flow_offload.h | 1 + include/net/tc_act/tc_police.h | 10 ++++++++++ net/sched/cls_api.c | 1 + 3 files changed, 12 insertions(+)
include/net/flow_offload.h +1 -0
@@ -234,6 +234,7 @@ struct flow_action_entry {struct {/* FLOW_ACTION_POLICE */ s64 burst;u64 rate_bytes_ps; +u32 mtu; } police;struct {/* FLOW_ACTION_CT */ int action;
include/net/tc_act/tc_police.h +10 -0
@@ -69,4 +69,14 @@ static inline s64 tcf_police_tcfp_burst(const struct tc_action *act) return params->tcfp_burst; } +static inline u32 tcf_police_tcfp_mtu (const struct tc_action *act) +{ +struct tcf_police *police =to_police (act); +struct tcf_police_params *params; + + params =rcu_dereference_protected (police->params, +lockdep_is_held (&police->tcf_lock)); +return params->tcfp_mtu; +} +#endif /* __NET_TC_POLICE_H */
net/sched/cls_api.c +1 -0
@@ -3658,6 +3658,7 @@ int tc_setup_flow_action(struct flow_action *flow_action, entry->police.burst =tcf_police_tcfp_burst (act); entry->police.rate_bytes_ps =tcf_police_rate_bytes_ps (act); + entry->police.mtu =tcf_police_tcfp_mtu (act); }else if (is_tcf_ct (act)) { entry->id =FLOW_ACTION_CT ; entry->ct.action =tcf_ct_action (act);
Signature · 1 line
2.17.1
PATCH v2 2/4 net: enetc: add support max frame size for tc flower offload
From: Po Liu <[email protected]>
Base on the tc flower offload police action add max frame size by the parameter 'mtu'. Tc flower device driver working by the IEEE 802.1Qci stream filter can implement the max frame size filtering. Add it to the current hardware tc flower stearm filter driver.
---
v2:
- No update.
.../net/ethernet/freescale/enetc/enetc_qos.c | 52 +++++++++++++------ 1 file changed, 36 insertions(+), 16 deletions(-)
drivers/net/ethernet/freescale/enetc/enetc_qos.c +36 -16
@@ -389,6 +389,7 @@ struct enetc_psfp_filter {u32 index;s32 handle;s8 prio; +u32 maxsdu;u32 gate_id;s32 meter_id;refcount_t refcount;
@@ -430,6 +431,12 @@ static struct actions_fwd enetc_act_fwd[] = {BIT (FLOW_DISSECTOR_KEY_ETH_ADDRS ),FILTER_ACTION_TYPE_PSFP }, + { +BIT (FLOW_ACTION_POLICE ) | +BIT (FLOW_ACTION_GATE ), +BIT (FLOW_DISSECTOR_KEY_ETH_ADDRS ), +FILTER_ACTION_TYPE_PSFP + },/* example for ACL actions */ { BIT(FLOW_ACTION_DROP),
@@ -594,8 +601,12 @@ static int enetc_streamfilter_hw_set(struct enetc_ndev_priv *priv,/* Filter Type. Identifies the contents of the MSDU/FM_INST_INDEX * field as being either an MSDU value or an index into the Flow * Meter Instance table. -* TODO: no limit max sdu */ +if (sfi->maxsdu) { + sfi_config->msdu = +cpu_to_le16 (sfi->maxsdu); + sfi_config->multi |=0x40 ; + }if (sfi->meter_id >=0 ) { sfi_config->fm_inst_table_index =cpu_to_le16 (sfi->meter_id);
@@ -872,6 +883,7 @@ static struct enetc_psfp_filterhlist_for_each_entry (s, &epsfp.psfp_filter_list, node)if (s->gate_id == sfi->gate_id && s->prio == sfi->prio && + s->maxsdu == sfi->maxsdu && s->meter_id == sfi->meter_id)return s;
@@ -979,6 +991,7 @@ static struct actions_fwd *enetc_check_flow_actions(u64 acts,static int enetc_psfp_parse_clsflower (struct enetc_ndev_priv *priv,struct flow_cls_offload *f) { +struct flow_action_entry *entryg =NULL , *entryp =NULL ;struct flow_rule *rule =flow_cls_offload_flow_rule (f);struct netlink_ext_ack *extack = f->common.extack;struct enetc_stream_filter *filter, *old_filter;
@@ -997,9 +1010,12 @@ static int enetc_psfp_parse_clsflower(struct enetc_ndev_priv *priv,flow_action_for_each (i, entry, &rule->action)if (entry->id ==FLOW_ACTION_GATE ) -break ; + entryg = entry; +else if (entry->id ==FLOW_ACTION_POLICE ) + entryp = entry; -if (entry->id !=FLOW_ACTION_GATE ) +/* Not support without gate action */ +if (!entryg)return -EINVAL ; filter =kzalloc (sizeof (*filter),GFP_KERNEL );
@@ -1079,19 +1095,19 @@ static int enetc_psfp_parse_clsflower(struct enetc_ndev_priv *priv, }/* parsing gate action */ -if (entry->gate.index >= priv->psfp_cap.max_psfp_gate) { +if (entryg->gate.index >= priv->psfp_cap.max_psfp_gate) {NL_SET_ERR_MSG_MOD (extack,"No Stream Gate resource!" ); err = -ENOSPC ; goto free_filter; } -if (entry->gate.num_entries >= priv->psfp_cap.max_psfp_gatelist) { +if (entryg->gate.num_entries >= priv->psfp_cap.max_psfp_gatelist) {NL_SET_ERR_MSG_MOD (extack,"No Stream Gate resource!" ); err = -ENOSPC ; goto free_filter; } - entries_size =struct_size (sgi, entries, entry->gate.num_entries); + entries_size =struct_size (sgi, entries, entryg->gate.num_entries); sgi =kzalloc (entries_size,GFP_KERNEL );if (!sgi) { err = -ENOMEM ;
@@ -1099,18 +1115,18 @@ static int enetc_psfp_parse_clsflower(struct enetc_ndev_priv *priv, }refcount_set (&sgi->refcount,1 ); - sgi->index = entry->gate.index; - sgi->init_ipv = entry->gate.prio; - sgi->basetime = entry->gate.basetime; - sgi->cycletime = entry->gate.cycletime; - sgi->num_entries = entry->gate.num_entries; + sgi->index = entryg->gate.index; + sgi->init_ipv = entryg->gate.prio; + sgi->basetime = entryg->gate.basetime; + sgi->cycletime = entryg->gate.cycletime; + sgi->num_entries = entryg->gate.num_entries; e = sgi->entries; -for (i =0 ; i < entry->gate.num_entries; i++) { - e[i].gate_state = entry->gate.entries[i].gate_state; - e[i].interval = entry->gate.entries[i].interval; - e[i].ipv = entry->gate.entries[i].ipv; - e[i].maxoctets = entry->gate.entries[i].maxoctets; +for (i =0 ; i < entryg->gate.num_entries; i++) { + e[i].gate_state = entryg->gate.entries[i].gate_state; + e[i].interval = entryg->gate.entries[i].interval; + e[i].ipv = entryg->gate.entries[i].ipv; + e[i].maxoctets = entryg->gate.entries[i].maxoctets; } filter->sgi_index = sgi->index;
@@ -1127,6 +1143,10 @@ static int enetc_psfp_parse_clsflower(struct enetc_ndev_priv *priv,/* flow meter not support yet */ sfi->meter_id =ENETC_PSFP_WILDCARD ; +/* Max frame size */ +if (entryp) + sfi->maxsdu = entryp->police.mtu; +/* prio ref the filter prio */ if (f->common.prio && f->common.prio <=BIT (3 )) sfi->prio = f->common.prio -1 ;
Signature · 1 line
2.17.1
Re: [v2,net-next 2/4] net: enetc: add support max frame size for tc flower offload
From: Po Liu <[email protected]>
Date: Wed, 24 Jun 2020 17:36:29 +0800
Quoted 8 lines
From: Po Liu <[email protected]>
Base on the tc flower offload police action add max frame size by the parameter 'mtu'. Tc flower device driver working by the IEEE 802.1Qci stream filter can implement the max frame size filtering. Add it to the current hardware tc flower stearm filter driver.
Signed-off-by Po Liu <[email protected]>
Applied.
PATCH v2 3/4 net: qos: police action add index for tc flower offloading
In reply to Po LiuFrom: Po Liu <[email protected]>
Hardware device may include more than one police entry. Specifying the action's index make it possible for several tc filters to share the same police action when installing the filters.
Propagate this index to device drivers through the flow offload intermediate representation, so that drivers could share a single hardware policer between multiple filters.
v1->v2 changes:
- Update the commit message suggest by Ido Schimmel <[email protected]>
--- include/net/flow_offload.h | 1 + net/sched/cls_api.c | 1 + 2 files changed, 2 insertions(+)
include/net/flow_offload.h +1 -0
@@ -232,6 +232,7 @@ struct flow_action_entry {bool truncate; } sample;struct {/* FLOW_ACTION_POLICE */ +u32 index;s64 burst;u64 rate_bytes_ps;u32 mtu;
net/sched/cls_api.c +1 -0
@@ -3659,6 +3659,7 @@ int tc_setup_flow_action(struct flow_action *flow_action, entry->police.rate_bytes_ps =tcf_police_rate_bytes_ps (act); entry->police.mtu =tcf_police_tcfp_mtu (act); + entry->police.index = act->tcfa_index; }else if (is_tcf_ct (act)) { entry->id =FLOW_ACTION_CT ; entry->ct.action =tcf_ct_action (act);
Signature · 1 line
2.17.1
Re: [v2,net-next 3/4] net: qos: police action add index for tc flower offloading
From: Po Liu <[email protected]>
Date: Wed, 24 Jun 2020 17:36:30 +0800
Quoted 6 lines
From: Po Liu <[email protected]>
Hardware device may include more than one police entry. Specifying the action's index make it possible for several tc filters to share the same police action when installing the filters.
Quoted 8 lines
Propagate this index to device drivers through the flow offload intermediate representation, so that drivers could share a single hardware policer between multiple filters.
v1->v2changes:
- Update the commit message suggest by Ido Schimmel <[email protected]>Signed-off-by Po Liu <[email protected]>
Applied.
PATCH v2 4/4 net: enetc add tc flower offload flow metering policing action
In reply to Po LiuFrom: Po Liu <[email protected]>
Flow metering entries in IEEE 802.1Qci is an optional function for a flow filtering module. Flow metering is two rates two buckets and three color marker to policing the frames. This patch only enable one rate one bucket and in color blind mode. Flow metering instance are as specified in the algorithm in MEF 10.3 and in Bandwidth Profile Parameters. They are:
a) Flow meter instance identifier. An integer value identifying the flow meter instance. The patch use the police 'index' as thin value.
b) Committed Information Rate (CIR), in bits per second. This patch use the 'rate_bytes_ps' represent this value.
c) Committed Burst Size (CBS), in octets. This patch use the 'burst' represent this value.
d) Excess Information Rate (EIR), in bits per second.
e) Excess Burst Size per Bandwidth Profile Flow (EBS), in octets. And plus some other parameters. This patch set EIR/EBS default disable and color blind mode.
v1->v2 changes:
- Use div_u64() as division replace the '/' report:
All errors (new ones prefixed by >>):
ld: drivers/net/ethernet/freescale/enetc/enetc_qos.o: in function `enetc_flowmeter_hw_set':
Quoted 1 line
Quoted 1 line
enetc_qos.c:(.text+0x66): undefined reference to `__udivdi3'
--- .../net/ethernet/freescale/enetc/enetc_hw.h | 24 +++ .../net/ethernet/freescale/enetc/enetc_qos.c | 160 ++++++++++++++++-- 2 files changed, 172 insertions(+), 12 deletions(-)
drivers/net/ethernet/freescale/enetc/enetc_hw.h +24 -0
@@ -570,6 +570,7 @@ enum bdcr_cmd_class {BDCR_CMD_STREAM_IDENTIFY ,BDCR_CMD_STREAM_FILTER ,BDCR_CMD_STREAM_GCL , +BDCR_CMD_FLOW_METER , __BDCR_CMD_MAX_LEN, BDCR_CMD_MAX_LEN = __BDCR_CMD_MAX_LEN - 1, };
@@ -736,10 +737,33 @@ struct sgcl_data {struct sgce sgcl[0 ]; }; +#define ENETC_CBDR_FMI_MR BIT(0) +#define ENETC_CBDR_FMI_MREN BIT(1) +#define ENETC_CBDR_FMI_DOY BIT(2) +#define ENETC_CBDR_FMI_CM BIT(3) +#define ENETC_CBDR_FMI_CF BIT(4) +#define ENETC_CBDR_FMI_NDOR BIT(5) +#define ENETC_CBDR_FMI_OALEN BIT(6) +#define ENETC_CBDR_FMI_IRFPP_MASK GENMASK(4, 0) + +/* class 10: command 0/1, Flow Meter Instance Set, short Format */ +struct fmi_conf { +__le32 cir; +__le32 cbs; +__le32 eir; +__le32 ebs; +u8 conf; +u8 res1; +u8 ir_fpp; +u8 res2[4 ]; +u8 en; +}; + struct enetc_cbd { union{ struct sfi_conf sfi_conf; struct sgi_table sgi_table; + struct fmi_conf fmi_conf; struct { __le32 addr[2]; union {
drivers/net/ethernet/freescale/enetc/enetc_qos.c +148 -12
252 lines not on this page · raw message
Signature · 1 line
2.17.1
Re: [v2,net-next 4/4] net: enetc add tc flower offload flow metering policing action
From: Po Liu <[email protected]>
Date: Wed, 24 Jun 2020 17:36:31 +0800
Quoted 23 lines
From: Po Liu <[email protected]>
Flow metering entries in IEEE 802.1Qci is an optional function for a flow filtering module. Flow metering is two rates two buckets and three color marker to policing the frames. This patch only enable one rate one bucket and in color blind mode. Flow metering instance are as specified in the algorithm in MEF 10.3 and in Bandwidth Profile Parameters. They are:
a) Flow meter instance identifier. An integer value identifying the flow meter instance. The patch use the police 'index' as thin value.
b) Committed Information Rate (CIR), in bits per second. This patch use the 'rate_bytes_ps' represent this value.
c) Committed Burst Size (CBS), in octets. This patch use the 'burst' represent this value.
d) Excess Information Rate (EIR), in bits per second.
e) Excess Burst Size per Bandwidth Profile Flow (EBS), in octets. And plus some other parameters. This patch set EIR/EBS default disable and color blind mode.
v1->v2changes:
- Usediv_u64()as division replace the '/' report:
Quoted 8 lines
Reported-by kernel test robot <[email protected]>All errors (new ones prefixed by >>):
ld:
drivers/net/ethernet/freescale/enetc/enetc_qos.o: in function `enetc_flowmeter_hw_set':Quoted 1 line
Quoted 1 line
enetc_qos.c:(.text+0x66): undefined reference to `__udivdi3'
Signed-off-by Po Liu <[email protected]>
Applied.
Re: [v2,net-next 1/4] net: qos: add tc police offloading action with max frame size limit
In reply to Po LiuFrom: Po Liu <[email protected]>
Date: Wed, 24 Jun 2020 17:36:28 +0800
Quoted 9 lines
From: Po Liu <[email protected]>
Current police offloading support the 'burst'' and '
rate_bytes_ps'. Some hardware own the capability to limit the frame size. If the frame size larger than the setting, the frame would be dropped. For the police action itself already accept the 'mtu' parameter in tc command. But not extend to tc flower offloading. So extend 'mtu' to tc flower offloading.Signed-off-by Po Liu <[email protected]>
Applied.
Re: [v1,net-next 1/4] net: qos: add tc police offloading action with max frame size limit
In reply to Po LiuQuoted 3 lines · Po Liu
From: Po Liu <[email protected]>
Current police offloading support the 'burst'' and '
rate_bytes_ps'. Some
s/support/supports/
s/'burst''/'burst'/
Quoted 3 lines
hardware own the capability to limit the frame size. If the frame size larger than the setting, the frame would be dropped. For the police action itself already accept the 'mtu' parameter in tc command. But not
s/accept/accepts/
Quoted 1 line
extend to tc flower offloading. So extend 'mtu' to tc flower offloading.
Throughout the submission you are always using the term 'flower offloading', but this has nothing to do with flower. Flower is the classifier, whereas you are extending police action which can be tied to any classifier.
Quoted 5 lines
Signed-off-by Po Liu <[email protected]>---
continue the thread [email protected] for the police action offloading.
For a patch set you need a cover letter (patch 0). It should include necessary background, motivation and overview of the patches. You can mention there that some of the patches were sent as RFC back in March and provide a link:
https://lore.kernel.org/netdev/[email protected]/
The code itself looks good to me.
Quoted 52 lines
include/net/flow_offload.h | 1 + include/net/tc_act/tc_police.h | 10 ++++++++++ net/sched/cls_api.c | 1 + 3 files changed, 12 insertions(+)include/net/flow_offload.h +1
-0@@ -234,6 +234,7 @@ struct flow_action_entry {struct {/* FLOW_ACTION_POLICE */ s64 burst;u64 rate_bytes_ps; +u32 mtu; } police;struct {/* FLOW_ACTION_CT */ int action;include/net/tc_act/tc_police.h +10
-0@@ -69,4 +69,14 @@ static inline s64 tcf_police_tcfp_burst(const struct tc_action *act) return params->tcfp_burst; } +static inline u32 tcf_police_tcfp_mtu (const struct tc_action *act) +{ +struct tcf_police *police =to_police (act); +struct tcf_police_params *params; + + params =rcu_dereference_protected (police->params, +lockdep_is_held (&police->tcf_lock)); +return params->tcfp_mtu; +} +#endif /* __NET_TC_POLICE_H */ net/sched/cls_api.c +1
-0@@ -3658,6 +3658,7 @@ int tc_setup_flow_action(struct flow_action *flow_action, entry->police.burst =tcf_police_tcfp_burst (act); entry->police.rate_bytes_ps =tcf_police_rate_bytes_ps (act); + entry->police.mtu =tcf_police_tcfp_mtu (act); }else if (is_tcf_ct (act)) { entry->id =FLOW_ACTION_CT ; entry->ct.action =tcf_ct_action (act);Signature · 1 line
2.17.1
RFC 7/9 net: enetc: add support max frame size for tc flower offload
In reply to Po LiuBase on the tc flower offload police action add max frame size by the parameter 'mtu'. Tc flower device driver working by the IEEE 802.1Qci stream filter can implement the max frame size filtering. Add it to the current hardware tc flower stearm filter driver.
The limitation is the police action must add burst size and rate per second. Driver ignore the parameter 'burst' size and 'rate_bytes_ps'. Next driver patch would make burst and rate to the flow metering function.
--- .../net/ethernet/freescale/enetc/enetc_qos.c | 52 +++++++++++++------ 1 file changed, 36 insertions(+), 16 deletions(-)
drivers/net/ethernet/freescale/enetc/enetc_qos.c +36 -16
@@ -388,6 +388,7 @@ struct enetc_psfp_filter {u32 index;s32 handle;s8 prio; +u32 maxsdu;u32 gate_id;s32 meter_id;u32 refcount;
@@ -429,6 +430,12 @@ struct actions_fwd enetc_act_fwd[] = {BIT (FLOW_DISSECTOR_KEY_ETH_ADDRS ),FILTER_ACTION_TYPE_PSFP }, + { +BIT (FLOW_ACTION_POLICE ) | +BIT (FLOW_ACTION_GATE ), +BIT (FLOW_DISSECTOR_KEY_ETH_ADDRS ), +FILTER_ACTION_TYPE_PSFP + },/* example for ACL actions */ { BIT(FLOW_ACTION_DROP),
@@ -593,8 +600,12 @@ static int enetc_streamfilter_hw_set(struct enetc_ndev_priv *priv,/* Filter Type. Identifies the contents of the MSDU/FM_INST_INDEX * field as being either an MSDU value or an index into the Flow * Meter Instance table. -* TODO: no limit max sdu */ +if (sfi->maxsdu) { + sfi_config->msdu = +cpu_to_le16 (sfi->maxsdu); + sfi_config->multi |=0x40 ; + }if (sfi->meter_id >=0 ) { sfi_config->fm_inst_table_index =cpu_to_le16 (sfi->meter_id);
@@ -860,6 +871,7 @@ static struct enetc_psfp_filterhlist_for_each_entry (s, &epsfp.psfp_filter_list, node)if (s->gate_id == sfi->gate_id && s->prio == sfi->prio && + s->maxsdu == sfi->maxsdu && s->meter_id == sfi->meter_id)return s;
@@ -965,6 +977,7 @@ struct actions_fwd *enetc_check_flow_actions(u64 acts, unsigned int inputkeys)static int enetc_psfp_parse_clsflower (struct enetc_ndev_priv *priv,struct flow_cls_offload *f) { +struct flow_action_entry *entryg =NULL , *entryp =NULL ;struct flow_rule *rule =flow_cls_offload_flow_rule (f);struct netlink_ext_ack *extack = f->common.extack;struct enetc_stream_filter *filter, *old_filter;
@@ -983,9 +996,12 @@ static int enetc_psfp_parse_clsflower(struct enetc_ndev_priv *priv,flow_action_for_each (i, entry, &rule->action)if (entry->id ==FLOW_ACTION_GATE ) -break ; + entryg = entry; +else if (entry->id ==FLOW_ACTION_POLICE ) + entryp = entry; -if (entry->id !=FLOW_ACTION_GATE ) +/* Not support without gate action */ +if (!entryg)return -EINVAL ; filter =kzalloc (sizeof (*filter),GFP_KERNEL );
@@ -1044,37 +1060,37 @@ static int enetc_psfp_parse_clsflower(struct enetc_ndev_priv *priv, }/* parsing gate action */ -if (entry->gate.index >= priv->psfp_cap.max_psfp_gate) { +if (entryg->gate.index >= priv->psfp_cap.max_psfp_gate) {NL_SET_ERR_MSG_MOD (extack,"No Stream Gate resource!" ); err = -ENOSPC ; goto free_filter; } -if (entry->gate.num_entries >= priv->psfp_cap.max_psfp_gatelist) { +if (entryg->gate.num_entries >= priv->psfp_cap.max_psfp_gatelist) {NL_SET_ERR_MSG_MOD (extack,"No Stream Gate resource!" ); err = -ENOSPC ; goto free_filter; } - entries_size =struct_size (sgi, entries, entry->gate.num_entries); + entries_size =struct_size (sgi, entries, entryg->gate.num_entries); sgi =kzalloc (entries_size,GFP_KERNEL );if (!sgi) { err = -ENOMEM ; goto free_filter; } - sgi->index = entry->gate.index; - sgi->init_ipv = entry->gate.prio; - sgi->basetime = entry->gate.basetime; - sgi->cycletime = entry->gate.cycletime; - sgi->num_entries = entry->gate.num_entries; + sgi->index = entryg->gate.index; + sgi->init_ipv = entryg->gate.prio; + sgi->basetime = entryg->gate.basetime; + sgi->cycletime = entryg->gate.cycletime; + sgi->num_entries = entryg->gate.num_entries; e = sgi->entries; -for (i =0 ; i < entry->gate.num_entries; i++) { - e[i].gate_state = entry->gate.entries[i].gate_state; - e[i].interval = entry->gate.entries[i].interval; - e[i].ipv = entry->gate.entries[i].ipv; - e[i].maxoctets = entry->gate.entries[i].maxoctets; +for (i =0 ; i < entryg->gate.num_entries; i++) { + e[i].gate_state = entryg->gate.entries[i].gate_state; + e[i].interval = entryg->gate.entries[i].interval; + e[i].ipv = entryg->gate.entries[i].ipv; + e[i].maxoctets = entryg->gate.entries[i].maxoctets; } filter->sgi_index = sgi->index;
@@ -1090,6 +1106,10 @@ static int enetc_psfp_parse_clsflower(struct enetc_ndev_priv *priv,/* flow meter not support yet */ sfi->meter_id =ENETC_PSFP_WILDCARD ; +/* Max frame size */ +if (entryp) + sfi->maxsdu = entryp->police.mtu; +/* prio ref the filter prio */ if (f->common.prio && f->common.prio <=BIT (3 )) sfi->prio = f->common.prio -1 ;
Signature · 1 line
2.17.1
RFC 8/9 net: qos: police action add index for tc flower offloading
In reply to Po LiuHardware may own many entries for police flow. So that make one(or
multi) flow to be policed by one hardware entry. This patch add the
police action index provide to the driver side make it mapping the driver hardware entry index.
--- include/net/flow_offload.h | 1 + net/sched/cls_api.c | 1 + 2 files changed, 2 insertions(+)
include/net/flow_offload.h +1 -0
@@ -201,6 +201,7 @@ struct flow_action_entry {bool truncate; } sample;struct {/* FLOW_ACTION_POLICE */ +u32 index;s64 burst;u64 rate_bytes_ps;u32 mtu;
net/sched/cls_api.c +1 -0
@@ -3584,6 +3584,7 @@ int tc_setup_flow_action(struct flow_action *flow_action, entry->police.rate_bytes_ps =tcf_police_rate_bytes_ps (act); entry->police.mtu =tcf_police_mtu (act); + entry->police.index = act->tcfa_index; }else if (is_tcf_ct (act)) { entry->id =FLOW_ACTION_CT ; entry->ct.action =tcf_ct_action (act);
Signature · 1 line
2.17.1
Re: [RFC,net-next 8/9] net: qos: police action add index for tc flower offloading
Quoted 6 lines · Po Liu
Hardware may own many entries for police flow. So that make one(or
multi) flow to be policed by one hardware entry. This patch add the
police action index provide to the driver side make it mapping the driver hardware entry index.Signed-off-by Po Liu <[email protected]>
Hi,
I started looking into tc-police offload in mlxsw and remembered your patch. Are you planning to formally submit it? I'm asking because in mlxsw it is also possible to share the same policer between multiple filters.
Thanks
Quoted 32 lines
--- include/net/flow_offload.h | 1 + net/sched/cls_api.c | 1 + 2 files changed, 2 insertions(+)include/net/flow_offload.h +1
-0@@ -201,6 +201,7 @@ struct flow_action_entry {bool truncate; } sample;struct {/* FLOW_ACTION_POLICE */ +u32 index;s64 burst;u64 rate_bytes_ps;u32 mtu;net/sched/cls_api.c +1
-0@@ -3584,6 +3584,7 @@ int tc_setup_flow_action(struct flow_action *flow_action, entry->police.rate_bytes_ps =tcf_police_rate_bytes_ps (act); entry->police.mtu =tcf_police_mtu (act); + entry->police.index = act->tcfa_index; }else if (is_tcf_ct (act)) { entry->id =FLOW_ACTION_CT ; entry->ct.action =tcf_ct_action (act);Signature · 1 line
2.17.1
RFC 9/9 net: enetc add tc flower offload flow metering policing action
In reply to Po LiuFlow metering entries in IEEE 802.1Qci is an optional function for a flow filtering module. Flow metering is two rates two buckets and three color marker to policing the frames. This patch only enable one rate one bucket and in color blind mode. Flow metering instance are as specified in the algorithm in MEF 10.3 and in Bandwidth Profile Parameters. They are:
a) Flow meter instance identifier. An integer value identifying the flow meter instance. The patch use the police 'index' as thin value.
b) Committed Information Rate (CIR), in bits per second. This patch use the 'rate_bytes_ps' represent this value.
c) Committed Burst Size (CBS), in octets. This patch use the 'burst' represent this value.
d) Excess Information Rate (EIR), in bits per second.
e) Excess Burst Size per Bandwidth Profile Flow (EBS), in octets. And plus some other parameters. This patch set EIR/EBS default disable and color blind mode.
--- .../net/ethernet/freescale/enetc/enetc_hw.h | 24 +++ .../net/ethernet/freescale/enetc/enetc_qos.c | 150 ++++++++++++++++-- 2 files changed, 164 insertions(+), 10 deletions(-)
drivers/net/ethernet/freescale/enetc/enetc_hw.h +24 -0
@@ -573,6 +573,7 @@ enum bdcr_cmd_class {BDCR_CMD_STREAM_IDENTIFY ,BDCR_CMD_STREAM_FILTER ,BDCR_CMD_STREAM_GCL , +BDCR_CMD_FLOW_METER , __BDCR_CMD_MAX_LEN, BDCR_CMD_MAX_LEN = __BDCR_CMD_MAX_LEN - 1, };
@@ -739,10 +740,33 @@ struct sgcl_data {struct sgce sgcl[0 ]; }; +#define ENETC_CBDR_FMI_MR BIT(0) +#define ENETC_CBDR_FMI_MREN BIT(1) +#define ENETC_CBDR_FMI_DOY BIT(2) +#define ENETC_CBDR_FMI_CM BIT(3) +#define ENETC_CBDR_FMI_CF BIT(4) +#define ENETC_CBDR_FMI_NDOR BIT(5) +#define ENETC_CBDR_FMI_OALEN BIT(6) +#define ENETC_CBDR_FMI_IRFPP_MASK GENMASK(4, 0) + +/* class 10: command 0/1, Flow Meter Instance Set, short Format */ +struct fmi_conf { +__le32 cir; +__le32 cbs; +__le32 eir; +__le32 ebs; +u8 conf; +u8 res1; +u8 ir_fpp; +u8 res2[4 ]; +u8 en; +}; + struct enetc_cbd { union{ struct sfi_conf sfi_conf; struct sgi_table sgi_table; + struct fmi_conf fmi_conf; struct { __le32 addr[2]; union {
drivers/net/ethernet/freescale/enetc/enetc_qos.c +140 -10
244 lines not on this page · raw message
Signature · 1 line
2.17.1
RFC 1/1 iproute2:tc:action: add a gate control action
In reply to Po LiuIntroduce a ingress frame gate control flow action. tc create a gate action would provide a gate list to control when open/close state. when the gate open state, the flow could pass but not when gate state is close. The driver would repeat the gate list cyclely. User also could assign a time point to start the gate list by the basetime parameter. if the basetime has passed current time, start time would calculate by the cycletime of the gate list.
The behavior try to keep according to the IEEE 802.1Qci spec. For the software emulation, require the user input the clock type.
Below is the setting software simulator example in user space:
Quoted 3 lines
tc qdisc add dev eth0 ingress
tc filter add dev eth0 parent ffff: protocol ip \
flower skip_hw src_ip 192.168.0.20 \
action gate index 2 \
sched-entry OPEN 200000000 -1 -1 \
sched-entry CLOSE 100000000 -1 -1Quoted 1 line
tc chain del dev eth0 ingress chain 0
"sched-entry" follow the name taprio style. gate state is "OPEN"/"CLOSE". Follow the period nanosecond. Then next -1 is internal priority value means which ingress queue should put. "-1" means wildcard. The last value optional specifies the maximum number of MSDU octets that are permitted to pass the gate during the specified time interval.
The command also support tc flower offload with device driver support:
Quoted 1 line
tc filter add dev eth0 parent ffff: protocol ip chain 14 \
flower skip_sw dst_mac 10:00:80:00:00:00 \ action gate index 12 sched-entry close 200000000 -1 -1
The iproute2 test patch need to upload to:
git://git.kernel.org/pub/scm/network/iproute2/iproute2-next.git
--- include/uapi/linux/tc_act/tc_gate.h | 47 +++ tc/Makefile | 1 + tc/m_gate.c | 483 ++++++++++++++++++++++++++++ 3 files changed, 531 insertions(+) create mode 100644 include/uapi/linux/tc_act/tc_gate.h create mode 100644 tc/m_gate.c
include/uapi/linux/tc_act/tc_gate.h new file +47 -0
@@ -0,0 +1,47 @@ +// SPDX-License-Identifier: (GPL-2.0+) +/* Copyright 2020 NXP */ + +#ifndef __LINUX_TC_GATE_H +#define __LINUX_TC_GATE_H + +#include <linux/pkt_cls.h> + +struct tc_gate { +tc_gen ; +}; + +enum { +TCA_GATE_ENTRY_UNSPEC , +TCA_GATE_ENTRY_INDEX , +TCA_GATE_ENTRY_GATE , +TCA_GATE_ENTRY_INTERVAL , +TCA_GATE_ENTRY_IPV , +TCA_GATE_ENTRY_MAX_OCTETS , + __TCA_GATE_ENTRY_MAX, +}; +#define TCA_GATE_ENTRY_MAX (__TCA_GATE_ENTRY_MAX - 1) + +enum { +TCA_GATE_ONE_ENTRY_UNSPEC , +TCA_GATE_ONE_ENTRY , + __TCA_GATE_ONE_ENTRY_MAX, +}; +#define TCA_GATE_ONE_ENTRY_MAX (__TCA_GATE_ONE_ENTRY_MAX - 1) + +enum { +TCA_GATE_UNSPEC , +TCA_GATE_TM , +TCA_GATE_PARMS , +TCA_GATE_PAD , +TCA_GATE_PRIORITY , +TCA_GATE_ENTRY_LIST , +TCA_GATE_BASE_TIME , +TCA_GATE_CYCLE_TIME , +TCA_GATE_CYCLE_TIME_EXT , +TCA_GATE_FLAGS , +TCA_GATE_CLOCKID , + __TCA_GATE_MAX, +}; +#define TCA_GATE_MAX (__TCA_GATE_MAX - 1) + +#endif
tc/Makefile +1 -0
@@ -54,6 +54,7 @@ TCMODULES += m_bpf.o TCMODULES += m_tunnel_key.o TCMODULES += m_sample.o TCMODULES += m_ct.o +TCMODULES += m_gate.o TCMODULES += p_ip.o TCMODULES += p_ip6.o TCMODULES += p_icmp.o
tc/m_gate.c new file +483 -0
483 lines not on this page · raw message
Signature · 1 line
2.17.1
cover Introduce a flow gate control action and apply IEEE
Changes from RFC:
0000: Reduce to 5 patches and remove the 4 max frame size offload and flow metering in the policing offload action, Only keep gate action offloading implementation.
0001: No changes.
0002:
- fix kfree lead ack by Jakub Kicinski and Cong Wang
- License fix from Jakub Kicinski and Stephen Hemminger
- Update example in commit acked by Vinicius Costa Gomes
- Fix the rcu protect in tcf_gate_act() acked by Vinicius
0003: No changes
0004: No changes
0005:
Acked by Vinicius Costa Gomes
- Use refcount kernel lib
- Update stream gate check code position
- Update reduce ref names more clear
iprout2 command patches:
0000: Update license expression and add gate id
0001: Add tc action gate man page
This patch set is trying to intruduce a way to add tc flower offload for the enetc IEEE 802.1Qci (PSFP) function. There are four main feature parts in PSFP to implement the flow policing and filtering for ingress flow with IEEE 802.1Qci features. They are stream identify(this is defined in the P802.1cb exactly but needed by 802.1Qci), stream filtering, stream gate and flow metering.
The stream gate function is the important part in the features. But there is no compare actions in the qdisc filter part. The second patch introduce a ingress frame gate control flow action. tc create a gate action would provide a gate list to control class open/close state. when the gate open state, the flow could pass but not when gate state is close. The driver would repeat the gate list periodic. User also could assign a time point to start the gate list by the basetime parameter. if the basetime has passed current time, start time would calculate by the cycletime of the gate list. And it is introduce a software simulator gate control method.
The first patch is fix a flow offload can't provide dropped frame count number issue. This would be used for getting the hardware offload dropped frame.
The third patch is to adding the gate flow offloading.
The fourth patch is to add tc offload command in enetc. This is to control the on/off for the tc flower offloading.
Now the enetc driver would got the gate control list and filter mac address etc. So enetc would collected the parameters and create the stream identify entry and stream gate control entry. Then driver would create a stream filter entry by these inputs. Driver would maintain the flow chain list. The fifth patch implement the stream gate and stream filter and stream identify functions in driver by the tc flower actions and tc filter parameters.
The iproute2 test patch need to upload to:
git://git.kernel.org/pub/scm/network/iproute2/iproute2-next.git
Po Liu (5):
net: qos offload add flow status with dropped count
net: qos: introduce a gate control flow action
net: schedule: add action gate offloading
net: enetc: add hw tc hw offload features for PSPF capability
net: enetc: add tc flower psfp offload driver
drivers/net/ethernet/broadcom/bnxt/bnxt_tc.c | 2 +- .../ethernet/chelsio/cxgb4/cxgb4_tc_flower.c | 2 +- .../chelsio/cxgb4/cxgb4_tc_matchall.c | 2 +- drivers/net/ethernet/freescale/enetc/enetc.c | 34 +- drivers/net/ethernet/freescale/enetc/enetc.h | 86 ++ .../net/ethernet/freescale/enetc/enetc_hw.h | 159 +++ .../net/ethernet/freescale/enetc/enetc_pf.c | 6 + .../net/ethernet/freescale/enetc/enetc_qos.c | 1074 +++++++++++++++++ .../net/ethernet/mellanox/mlx5/core/en_tc.c | 4 +- .../ethernet/mellanox/mlxsw/spectrum_flower.c | 2 +- drivers/net/ethernet/mscc/ocelot_flower.c | 2 +- .../ethernet/netronome/nfp/flower/offload.c | 2 +- .../ethernet/netronome/nfp/flower/qos_conf.c | 2 +- include/net/act_api.h | 11 +- include/net/flow_offload.h | 15 +- include/net/pkt_cls.h | 5 +- include/net/tc_act/tc_gate.h | 169 +++ include/uapi/linux/pkt_cls.h | 1 + include/uapi/linux/tc_act/tc_gate.h | 47 + net/sched/Kconfig | 15 + net/sched/Makefile | 1 + net/sched/act_api.c | 12 +- net/sched/act_ct.c | 6 +- net/sched/act_gact.c | 7 +- net/sched/act_gate.c | 647 ++++++++++ net/sched/act_mirred.c | 6 +- net/sched/act_police.c | 6 +- net/sched/act_vlan.c | 6 +- net/sched/cls_api.c | 33 + net/sched/cls_flower.c | 3 +- net/sched/cls_matchall.c | 3 +- 31 files changed, 2329 insertions(+), 41 deletions(-) create mode 100644 include/net/tc_act/tc_gate.h create mode 100644 include/uapi/linux/tc_act/tc_gate.h create mode 100644 net/sched/act_gate.c
Signature · 1 line
2.17.1
PATCH 1/5 net: qos offload add flow status with dropped count
Add the hardware tc flower offloading with dropped frame counter for status update. action ops->stats_update only loaded by the tcf_exts_stats_update() and tcf_exts_stats_update() only loaded by matchall and tc flower hardware filter. But the stats_update only set the dropped count as default false in the ops->stats_update. This patch add the dropped counter to action stats update. Its dropped counter update by the hardware offloading driver.
This is changed by replacing the drop flag with dropped frame counter.
Driver side should update how many "packets" it filtered and how many "dropped" in those "packets".
--- drivers/net/ethernet/broadcom/bnxt/bnxt_tc.c | 2 +- drivers/net/ethernet/chelsio/cxgb4/cxgb4_tc_flower.c | 2 +- .../net/ethernet/chelsio/cxgb4/cxgb4_tc_matchall.c | 2 +- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 4 ++-- .../net/ethernet/mellanox/mlxsw/spectrum_flower.c | 2 +- drivers/net/ethernet/mscc/ocelot_flower.c | 2 +- drivers/net/ethernet/netronome/nfp/flower/offload.c | 2 +- drivers/net/ethernet/netronome/nfp/flower/qos_conf.c | 2 +- include/net/act_api.h | 11 ++++++----- include/net/flow_offload.h | 5 ++++- include/net/pkt_cls.h | 5 +++-- net/sched/act_api.c | 12 ++++++------ net/sched/act_ct.c | 6 +++--- net/sched/act_gact.c | 7 ++++--- net/sched/act_mirred.c | 6 +++--- net/sched/act_police.c | 6 +++--- net/sched/act_vlan.c | 6 +++--- net/sched/cls_flower.c | 3 ++- net/sched/cls_matchall.c | 3 ++- 19 files changed, 48 insertions(+), 40 deletions(-)
drivers/net/ethernet/broadcom/bnxt/bnxt_tc.c +1 -1
@@ -1639,7 +1639,7 @@ static int bnxt_tc_get_flow_stats(struct bnxt *bp,spin_unlock (&flow->stats_lock);flow_stats_update (&tc_flow_cmd->stats, stats.bytes, stats.packets, - lastused); + lastused,0 );return 0 ; }
drivers/net/ethernet/chelsio/cxgb4/cxgb4_tc_flower.c +1 -1
@@ -903,7 +903,7 @@ int cxgb4_tc_flower_stats(struct net_device *dev, ofld_stats->last_used = jiffies;flow_stats_update (&cls->stats, bytes - ofld_stats->byte_count, packets - ofld_stats->packet_count, - ofld_stats->last_used); + ofld_stats->last_used,0 ); ofld_stats->packet_count = packets; ofld_stats->byte_count = bytes;
drivers/net/ethernet/chelsio/cxgb4/cxgb4_tc_matchall.c +1 -1
@@ -346,7 +346,7 @@ int cxgb4_tc_matchall_stats(struct net_device *dev,flow_stats_update (&cls_matchall->stats, bytes - tc_port_matchall->ingress.bytes, packets - tc_port_matchall->ingress.packets, - tc_port_matchall->ingress.last_used); + tc_port_matchall->ingress.last_used,0 ); tc_port_matchall->ingress.packets = packets; tc_port_matchall->ingress.bytes = bytes;
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +2 -2
@@ -4467,7 +4467,7 @@ int mlx5e_stats_flower(struct net_device *dev, struct mlx5e_priv *priv, no_peer_counter:mlx5_devcom_release_peer_data (devcom,MLX5_DEVCOM_ESW_OFFLOADS ); out: -flow_stats_update (&f->stats, bytes, packets, lastuse); +flow_stats_update (&f->stats, bytes, packets, lastuse,0 );trace_mlx5e_stats_flower (f); errout:mlx5e_flow_put (priv, flow);
@@ -4584,7 +4584,7 @@ void mlx5e_tc_stats_matchall(struct mlx5e_priv *priv, dpkts = cur_stats.rx_packets - rpriv->prev_vf_vport_stats.rx_packets; dbytes = cur_stats.rx_bytes - rpriv->prev_vf_vport_stats.rx_bytes; rpriv->prev_vf_vport_stats = cur_stats; -flow_stats_update (&ma->stats, dpkts, dbytes, jiffies); +flow_stats_update (&ma->stats, dpkts, dbytes, jiffies,0 ); } static void mlx5e_tc_hairpin_update_dead_peer(struct mlx5e_priv *priv,
drivers/net/ethernet/mellanox/mlxsw/spectrum_flower.c +1 -1
@@ -578,7 +578,7 @@ int mlxsw_sp_flower_stats(struct mlxsw_sp *mlxsw_sp,if (err) goto err_rule_get_stats; -flow_stats_update (&f->stats, bytes, packets, lastuse); +flow_stats_update (&f->stats, bytes, packets, lastuse,0 );mlxsw_sp_acl_ruleset_put (mlxsw_sp, ruleset);return 0 ;
drivers/net/ethernet/mscc/ocelot_flower.c +1 -1
@@ -224,7 +224,7 @@ int ocelot_cls_flower_stats(struct ocelot *ocelot, int port,if (ret)return ret; -flow_stats_update (&f->stats,0x0 , ace.stats.pkts,0x0 ); +flow_stats_update (&f->stats,0x0 , ace.stats.pkts,0x0 ,0x0 );return 0 ; }EXPORT_SYMBOL_GPL (ocelot_cls_flower_stats);
drivers/net/ethernet/netronome/nfp/flower/offload.c +1 -1
@@ -1490,7 +1490,7 @@ nfp_flower_get_stats(struct nfp_app *app, struct net_device *netdev,nfp_flower_update_merge_stats (app, nfp_flow);flow_stats_update (&flow->stats, priv->stats[ctx_id].bytes, - priv->stats[ctx_id].pkts, priv->stats[ctx_id].used); + priv->stats[ctx_id].pkts, priv->stats[ctx_id].used,0 ); priv->stats[ctx_id].pkts =0 ; priv->stats[ctx_id].bytes =0 ;
drivers/net/ethernet/netronome/nfp/flower/qos_conf.c +1 -1
@@ -320,7 +320,7 @@ nfp_flower_stats_rate_limiter(struct nfp_app *app, struct net_device *netdev,spin_unlock_bh (&fl_priv->qos_stats_lock);flow_stats_update (&flow->stats, diff_bytes, diff_pkts, - repr_priv->qos_table.last_update); + repr_priv->qos_table.last_update,0 );return 0 ; }
include/net/act_api.h +6 -5
@@ -103,7 +103,7 @@ struct tc_action_ops { struct netlink_callback *, int, const struct tc_action_ops *, struct netlink_ext_ack *); -void (*stats_update)(struct tc_action *,u64 ,u32 ,u64 ,bool ); +void (*stats_update)(struct tc_action *,u64 ,u64 ,u64 ,u64 ,bool );size_t (*get_fill_size)(const struct tc_action *act);struct net_device *(*get_dev)(const struct tc_action *a,tc_action_priv_destructor *destructor);
@@ -229,8 +229,8 @@ static inline void tcf_action_inc_overlimit_qstats(struct tc_action *a)spin_unlock (&a->tcfa_lock); } -void tcf_action_update_stats (struct tc_action *a,u64 bytes,u32 packets, -bool drop,bool hw); +void tcf_action_update_stats (struct tc_action *a,u64 bytes,u64 packets, +u64 dropped,bool hw);int tcf_action_copy_stats (struct sk_buff *,struct tc_action *,int ); int tcf_action_check_ctrlact(int action, struct tcf_proto *tp,
@@ -241,13 +241,14 @@ struct tcf_chain *tcf_action_set_ctrlact(struct tc_action *a, int action,#endif /* CONFIG_NET_CLS_ACT */ static inline void tcf_action_stats_update (struct tc_action *a,u64 bytes, -u64 packets,u64 lastuse,bool hw) +u64 packets,u64 lastuse, +u64 dropped,bool hw) {#ifdef CONFIG_NET_CLS_ACT if (!a->ops->stats_update)return ; - a->ops->stats_update (a, bytes, packets, lastuse, hw); + a->ops->stats_update (a, bytes, packets, lastuse, dropped, hw);#endif }
include/net/flow_offload.h +4 -1
@@ -369,14 +369,17 @@ struct flow_stats {u64 pkts;u64 bytes;u64 lastused; +u64 dropped; };static inline void flow_stats_update (struct flow_stats *flow_stats, -u64 bytes,u64 pkts,u64 lastused) +u64 bytes,u64 pkts, +u64 lastused,u64 dropped) { flow_stats->pkts += pkts; flow_stats->bytes += bytes; flow_stats->lastused =max_t (u64, flow_stats->lastused, lastused); + flow_stats->dropped += dropped; }enum flow_block_command {
include/net/pkt_cls.h +3 -2
@@ -262,7 +262,7 @@ static inline void tcf_exts_put_net(struct tcf_exts *exts)static inline void tcf_exts_stats_update (const struct tcf_exts *exts, -u64 bytes,u64 packets,u64 lastuse) +u64 bytes,u64 packets,u64 lastuse,u64 dropped) {#ifdef CONFIG_NET_CLS_ACT int i;
@@ -272,7 +272,8 @@ tcf_exts_stats_update(const struct tcf_exts *exts,for (i =0 ; i < exts->nr_actions; i++) {struct tc_action *a = exts->actions[i]; -tcf_action_stats_update (a, bytes, packets, lastuse, true); +tcf_action_stats_update (a, bytes, packets, + lastuse, dropped, true); }preempt_enable ();
net/sched/act_api.c +6 -6
@@ -1053,14 +1053,13 @@ int tcf_action_init(struct net *net, struct tcf_proto *tp, struct nlattr *nla,return err; } -void tcf_action_update_stats (struct tc_action *a,u64 bytes,u32 packets, -bool drop,bool hw) +void tcf_action_update_stats (struct tc_action *a,u64 bytes,u64 packets, +u64 dropped,bool hw) {if (a->cpu_bstats) {_bstats_cpu_update (this_cpu_ptr (a->cpu_bstats), bytes, packets); -if (drop) -this_cpu_ptr (a->cpu_qstats)->drops += packets; +this_cpu_ptr (a->cpu_qstats)->drops += dropped;if (hw) _bstats_cpu_update(this_cpu_ptr(a->cpu_bstats_hw),
@@ -1069,8 +1068,9 @@ void tcf_action_update_stats(struct tc_action *a, u64 bytes, u32 packets, }_bstats_update (&a->tcfa_bstats, bytes, packets); -if (drop) - a->tcfa_qstats.drops += packets; + + a->tcfa_qstats.drops += dropped; +if (hw)_bstats_update (&a->tcfa_bstats_hw, bytes, packets); }
net/sched/act_ct.c +3 -3
@@ -1445,12 +1445,12 @@ static int tcf_ct_search(struct net *net, struct tc_action **a, u32 index)return tcf_idr_search (tn, a, index); } -static void tcf_stats_update (struct tc_action *a,u64 bytes,u32 packets, -u64 lastuse,bool hw) +static void tcf_stats_update (struct tc_action *a,u64 bytes,u64 packets, +u64 lastuse,u64 dropped,bool hw) {struct tcf_ct *c =to_ct (a); -tcf_action_update_stats (a, bytes, packets, false, hw); +tcf_action_update_stats (a, bytes, packets, dropped, hw); c->tcf_tm.lastuse =max_t (u64, c->tcf_tm.lastuse, lastuse); }
net/sched/act_gact.c +4 -3
@@ -171,14 +171,15 @@ static int tcf_gact_act(struct sk_buff *skb, const struct tc_action *a,return action; } -static void tcf_gact_stats_update (struct tc_action *a,u64 bytes,u32 packets, -u64 lastuse,bool hw) +static void tcf_gact_stats_update (struct tc_action *a,u64 bytes,u64 packets, +u64 lastuse,u64 dropped,bool hw) {struct tcf_gact *gact =to_gact (a);int action =READ_ONCE (gact->tcf_action);struct tcf_t *tm = &gact->tcf_tm; -tcf_action_update_stats (a, bytes, packets, action ==TC_ACT_SHOT , hw); +tcf_action_update_stats (a, bytes, packets, + (action ==TC_ACT_SHOT ) ? packets :0 , hw); tm->lastuse =max_t (u64, tm->lastuse, lastuse); }
net/sched/act_mirred.c +3 -3
@@ -314,13 +314,13 @@ static int tcf_mirred_act(struct sk_buff *skb, const struct tc_action *a,return retval; } -static void tcf_stats_update (struct tc_action *a,u64 bytes,u32 packets, -u64 lastuse,bool hw) +static void tcf_stats_update (struct tc_action *a,u64 bytes,u64 packets, +u64 lastuse,u64 dropped,bool hw) {struct tcf_mirred *m =to_mirred (a);struct tcf_t *tm = &m->tcf_tm; -tcf_action_update_stats (a, bytes, packets, false, hw); +tcf_action_update_stats (a, bytes, packets, dropped, hw); tm->lastuse =max_t (u64, tm->lastuse, lastuse); }
net/sched/act_police.c +3 -3
@@ -288,13 +288,13 @@ static void tcf_police_cleanup(struct tc_action *a) }static void tcf_police_stats_update (struct tc_action *a, -u64 bytes,u32 packets, -u64 lastuse,bool hw) +u64 bytes,u64 packets, +u64 lastuse,u64 dropped,bool hw) {struct tcf_police *police =to_police (a);struct tcf_t *tm = &police->tcf_tm; -tcf_action_update_stats (a, bytes, packets, false, hw); +tcf_action_update_stats (a, bytes, packets, dropped, hw); tm->lastuse =max_t (u64, tm->lastuse, lastuse); }
net/sched/act_vlan.c +3 -3
@@ -302,13 +302,13 @@ static int tcf_vlan_walker(struct net *net, struct sk_buff *skb,return tcf_generic_walker (tn, skb, cb, type, ops, extack); } -static void tcf_vlan_stats_update (struct tc_action *a,u64 bytes,u32 packets, -u64 lastuse,bool hw) +static void tcf_vlan_stats_update (struct tc_action *a,u64 bytes,u64 packets, +u64 lastuse,u64 dropped,bool hw) {struct tcf_vlan *v =to_vlan (a);struct tcf_t *tm = &v->tcf_tm; -tcf_action_update_stats (a, bytes, packets, false, hw); +tcf_action_update_stats (a, bytes, packets, dropped, hw); tm->lastuse =max_t (u64, tm->lastuse, lastuse); }
net/sched/cls_flower.c +2 -1
@@ -492,7 +492,8 @@ static void fl_hw_update_stats(struct tcf_proto *tp, struct cls_fl_filter *f,tcf_exts_stats_update (&f->exts, cls_flower.stats.bytes, cls_flower.stats.pkts, - cls_flower.stats.lastused); + cls_flower.stats.lastused, + cls_flower.stats.dropped); }static void __fl_put (struct cls_fl_filter *f)
net/sched/cls_matchall.c +2 -1
@@ -338,7 +338,8 @@ static void mall_stats_hw_filter(struct tcf_proto *tp,tc_setup_cb_call (block,TC_SETUP_CLSMATCHALL , &cls_mall, false, true);tcf_exts_stats_update (&head->exts, cls_mall.stats.bytes, - cls_mall.stats.pkts, cls_mall.stats.lastused); + cls_mall.stats.pkts, cls_mall.stats.lastused, + cls_mall.stats.dropped); } static int mall_dump(struct net *net, struct tcf_proto *tp, void *fh,
Signature · 1 line
2.17.1
Jiri Pirko
maintainerRe: [v1,net-next 1/5] net: qos offload add flow status with dropped count
Quoted 8 lines · Tue, Mar 24, 2020 at 04:47:39AM CET, [email protected] wrote:
Add the hardware tc flower offloading with dropped frame counter for status update. action
ops->stats_updateonly loaded by thetcf_exts_stats_update()andtcf_exts_stats_update()only loaded by matchall and tc flower hardware filter. But thestats_updateonly set the dropped count as default false in theops->stats_update. This patch add the dropped counter to action stats update. Its dropped counter update by the hardware offloading driver.
This is changed by replacing the drop flag with dropped frame counter.
I just read this paragraph 3 times, I'm unable to decypher :(
Quoted 4 lines
Driver side should update how many "packets" it filtered and how many "dropped" in those "packets".
[...]
Quoted 12 lines
return action; }net/sched/act_gact.c +2
-2-static void tcf_gact_stats_update (struct tc_action *a,u64 bytes,u32 packets, -u64 lastuse,bool hw) +static void tcf_gact_stats_update (struct tc_action *a,u64 bytes,u64 packets, +u64 lastuse,u64 dropped,bool hw){ struct tcf_gact *gact = to_gact(a); int action = READ_ONCE(gact->tcf_action); struct tcf_t *tm = &gact->tcf_tm;
net/sched/act_gact.c static int tcf_gact_act(struct sk_buff *skb, const struct tc_action *a, · line 182
Quoted 3 lines
net/sched/act_gact.c +2
-1-tcf_action_update_stats (a, bytes, packets, action ==TC_ACT_SHOT , hw); +tcf_action_update_stats (a, bytes, packets, + (action ==TC_ACT_SHOT ) ? packets :0 , hw);
Avoid ()s here.
Quoted 3 lines
tm->lastuse = max_t(u64, tm->lastuse, lastuse); }
RE: [EXT] Re: [v1,net-next 1/5] net: qos offload add flow status with dropped count
Hi Jiri,
Quoted 32 lines
-----Original Message-----
From: Jiri Pirko <[email protected]>
Sent: 2020年3月24日 18:02
To: Po Liu <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; Claudiu Manoil <[email protected]>; Vladimir Oltean <[email protected]>; Alexandru Marginean <[email protected]>; Xiaoliang Yang <[email protected]>; Roy Zang <[email protected]>; Mingkai Hu <[email protected]>; Jerry Huang <[email protected]>; Leo Li <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]
Subject: [EXT] Re: [v1,net-next 1/5] net: qos offload add flow status with dropped countCaution: EXT Email
Quoted 8 lines · Tue, Mar 24, 2020 at 04:47:39AM CET, [email protected] wrote:
Add the hardware tc flower offloading with dropped frame counter for status update. action
ops->stats_updateonly loaded by thetcf_exts_stats_update()andtcf_exts_stats_update()only loaded by matchall and tc flower hardware filter. But thestats_updateonly set the dropped count as default false in theops->stats_update. This patch add the dropped counter to action stats update. Its dropped counter update by the hardware offloading driver.
This is changed by replacing the drop flag with dropped frame counter.
Quoted 1 line
I just read this paragraph 3 times, I'm unable to decypher :(
Sorry to make you confusing. I would make a clear description.
Before that, I just try to explain what the patch do here so you can provide more suggestion.
To get the stats in the tc flower offloading(by flag FLOW_CLS_STATS), it saves in the 'struct flow_stats' in the ' struct flow_cls_offload '. But the ' struct flow_stats ' only include the packages numbers. Some actions like policing also this 0002/0003 patch introduce stream gate action would produce dropped frames which is important for user evaluation. The packages number(pkts in struct flow_stats) and dropped number relation should be 'pkts' is how many frames were filtered, and 'dropped' number is how many frames dropped in those 'pkts'.
Eventually, the stats updated by the struct tc_action 's operation stats_update(). To implement add dropped number, then add parameter 'dropped' in the related functions: ops->stats_update/tcf_exts_stats_update() and also the tcf_action_update_stats(). There is a bool parameter 'drop' which is using now, can be understand the stats updating is for update for 'drop' count or not. But this flag is not useless as I checked in current kernel code(correct me if it is not). So replace the bool 'drop' flag with 'dropped' number in tcf_action_update_stats(). Make the tcf_action_update_stats() shows how many ' packets' updated and how many 'dropped' number in these 'packets'. Thanks!
Quoted 25 lines
Quoted 7 lines
Driver side should update how many "packets" it filtered and how many "dropped" in those "packets".
[...]
Quoted 4 lines
return action; }-static void tcf_gact_stats_update(struct
tc_action*a, u64 bytes, u32packets,
Review
Quoted 2 lines
(quoted diff) +1
-1- u64 lastuse, bool hw) +static void tcf_gact_stats_update(struct tc_action *a, u64 bytes, u64packets,
Quoted 6 lines
+ u64 lastuse, u64 dropped, bool hw)
{ struct tcf_gact *gact = to_gact(a); int action = READ_ONCE(gact->tcf_action); struct tcf_t *tm = &gact->tcf_tm;
Quoted 7 lines
Quoted 1 line
- tcf_action_update_stats(a, bytes, packets, action ==
TC_ACT_SHOT,hw);
Review
Quoted 3 lines
(quoted diff) +3
-0+ tcf_action_update_stats(a, bytes, packets, + (action == TC_ACT_SHOT) ? packets : 0, + hw);Avoid ()s here.
Ok.
Quoted 4 lines
Quoted 4 lines
tm->lastuse = max_t(u64, tm->lastuse, lastuse); }
Br,
Po Liu
PATCH 2/5 net: qos: introduce a gate control flow action
In reply to Po LiuIntroduce a ingress frame gate control flow action. tc create a gate action would provide a gate list to control when open/close state. when the gate open state, the flow could pass but not when gate state is close. The driver would repeat the gate list cyclically. User also could assign a time point to start the gate list by the basetime parameter. if the basetime has passed current time, start time would calculate by the cycletime of the gate list.
The action gate behavior try to keep according to the IEEE 802.1Qci spec. For the software simulation, require the user input the clock type.
Below is the setting example in user space. Tc filter a stream source ip address is 192.168.0.20 and gate action own two time slots. One is last 200ms gate open let frame pass another is last 100ms gate close let frames dropped. When the passed total frames over 8000000 bytes, it will dropped in one 200000000ns time slot.
Quoted 3 lines
tc qdisc add dev eth0 ingress
tc filter add dev eth0 parent ffff: protocol ip \
flower src_ip 192.168.0.20 \ action gate index 2 clockid CLOCK_TAI \ sched-entry OPEN 200000000 -1 8000000 \ sched-entry CLOSE 100000000 -1 -1
Quoted 1 line
tc chain del dev eth0 ingress chain 0
"sched-entry" follow the name taprio style. gate state is "OPEN"/"CLOSE". Follow the period nanosecond. Then next item is internal priority value means which ingress queue should put. "-1" means wildcard. The last value optional specifies the maximum number of MSDU octets that are permitted to pass the gate during the specified time interval.
Base-time is not set will be as 0 as default, as result start time would be ((N + 1) * cycletime) which is the minimal of future time.
Below example shows filtering a stream with destination mac address is 10:00:80:00:00:00 and ip type is ICMP, follow the action gate. The gate action would run with one close time slot which means always keep close. The time cycle is total 200000000ns. The base-time would calculate by:
1357000000000 + (N + 1) * cycletime
When the total value is the future time, it will be the start time. The cycletime here would be 200000000ns for this case.
Quoted 1 line
tc filter add dev eth0 parent ffff: protocol ip \
flower skip_hw ip_proto icmp dst_mac 10:00:80:00:00:00 \ action gate index 12 base-time 1357000000000 \ sched-entry CLOSE 200000000 -1 -1 \ clockid CLOCK_TAI
NOTE: This software simulator version not separate the admin/operation state machine. Update setting would overwrite stop the previos setting and waiting new cycle start.
---
Changes from RFC:
- fix kfree lead ack by Jakub Kicinski and Cong Wang
- License fix from Jakub Kicinski and Stephen Hemminger
- Update example in commit acked by Vinicius Costa Gomes
- Fix the rcu protect in tcf_gate_act() acked by Vinicius
include/net/tc_act/tc_gate.h | 54 +++ include/uapi/linux/pkt_cls.h | 1 + include/uapi/linux/tc_act/tc_gate.h | 47 ++ net/sched/Kconfig | 15 + net/sched/Makefile | 1 + net/sched/act_gate.c | 647 ++++++++++++++++++++++++++++ 6 files changed, 765 insertions(+) create mode 100644 include/net/tc_act/tc_gate.h create mode 100644 include/uapi/linux/tc_act/tc_gate.h create mode 100644 net/sched/act_gate.c
include/net/tc_act/tc_gate.h new file +54 -0
@@ -0,0 +1,54 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +/* Copyright 2020 NXP */ + +#ifndef __NET_TC_GATE_H +#define __NET_TC_GATE_H + +#include <net/act_api.h> +#include <linux/tc_act/tc_gate.h> + +struct tcfg_gate_entry { +int index; +u8 gate_state; +u32 interval; +s32 ipv; +s32 maxoctets; +struct list_head list; +}; + +struct tcf_gate_params { +s32 tcfg_priority; +u64 tcfg_basetime; +u64 tcfg_cycletime; +u64 tcfg_cycletime_ext; +u32 tcfg_flags; +s32 tcfg_clockid; +size_t num_entries; +struct list_head entries; +}; + +#define GATE_ACT_GATE_OPEN BIT(0) +#define GATE_ACT_PENDING BIT(1) +struct gate_action { +struct tcf_gate_params param; +spinlock_t entry_lock; +u8 current_gate_status; +ktime_t current_close_time; +u32 current_entry_octets; +s32 current_max_octets; +struct tcfg_gate_entry __rcu *next_entry; +struct hrtimer hitimer; +enum tk_offsets tk_offset; +struct rcu_head rcu; +}; + +struct tcf_gate { +struct tc_action common; +struct gate_action __rcu *actg; +}; +#define to_gate (a) ((struct tcf_gate *)a) + +#define get_gate_param (act) ((struct tcf_gate_params *)act) +#define get_gate_action (p) ((struct gate_action *)p) + +#endif
include/uapi/linux/pkt_cls.h +1 -0
@@ -134,6 +134,7 @@ enum tca_id {TCA_ID_CTINFO ,TCA_ID_MPLS ,TCA_ID_CT , +TCA_ID_GATE ,/* other actions go here */ __TCA_ID_MAX = 255 };
include/uapi/linux/tc_act/tc_gate.h new file +47 -0
47 lines not on this page · raw message
net/sched/Kconfig +15 -0
@@ -981,6 +981,21 @@ config NET_ACT_CT To compile this code as a module, choose M here: the module will be called act_ct. +config NET_ACT_GATE + tristate "Frame gate list control tc action" + depends on NET_CLS_ACT + help + Say Y here to allow the control the ingress flow by the gate list + control. The frame policing by the time gate list control open/close + cycle time. The manipulation will simulate the IEEE 802.1Qci stream + gate control behavior. The action could be offload by the tc flower + to hardware driver which the hardware own the capability of IEEE + 802.1Qci. + + If unsure, say N. + To compile this code as a module, choose M here: the + module will be called act_gate. + config NET_IFE_SKBMARK tristate "Support to encoding decoding skb mark on IFE action" depends on NET_ACT_IFE
net/sched/Makefile +1 -0
@@ -30,6 +30,7 @@ obj-$(CONFIG_NET_IFE_SKBPRIO) += act_meta_skbprio.o obj-$(CONFIG_NET_IFE_SKBTCINDEX) += act_meta_skbtcindex.o obj-$(CONFIG_NET_ACT_TUNNEL_KEY)+= act_tunnel_key.o obj-$(CONFIG_NET_ACT_CT) += act_ct.o +obj-$(CONFIG_NET_ACT_GATE) += act_gate.o obj-$(CONFIG_NET_SCH_FIFO) += sch_fifo.o obj-$(CONFIG_NET_SCH_CBQ) += sch_cbq.o obj-$(CONFIG_NET_SCH_HTB) += sch_htb.o
net/sched/act_gate.c new file +647 -0
647 lines not on this page · raw message
Signature · 1 line
2.17.1
Jiri Pirko
maintainerRe: [v1,net-next 2/5] net: qos: introduce a gate control flow action
Quoted 7 lines · Tue, Mar 24, 2020 at 04:47:40AM CET, [email protected] wrote:
Introduce a ingress frame gate control flow action. tc create a gate action would provide a gate list to control when open/close state. when the gate open state, the flow could pass but not when gate state is close. The driver would repeat the gate list cyclically. User also could assign a time point to start the gate list by the basetime parameter. if the basetime has passed current time, start time would calculate by the cycletime of the gate list.
Cannot decypher this either :/ Seriously, please make the patch descriptions readable.
Also, a sentence starts with capital letter.
Quoted 9 lines
The action gate behavior try to keep according to the IEEE 802.1Qci spec. For the software simulation, require the user input the clock type.
Below is the setting example in user space. Tc filter a stream source ip address is 192.168.0.20 and gate action own two time slots. One is last 200ms gate open let frame pass another is last 100ms gate close let frames dropped. When the passed total frames over 8000000 bytes, it will dropped in one 200000000ns time slot.
Quoted 7 lines
Quoted 3 lines
tc qdisc add dev eth0 ingress
tc filter add dev eth0 parent ffff: protocol ip \
flower src_ip 192.168.0.20 \ action gate index 2 clockid CLOCK_TAI \ sched-entry OPEN 200000000 -1 8000000 \ sched-entry CLOSE 100000000 -1 -1
The rest of the commands do not use capitals. Please lowercase these.
Quoted 32 lines
Quoted 2 lines
tc chain del dev eth0 ingress chain 0
"sched-entry" follow the name taprio style. gate state is "OPEN"/"CLOSE". Follow the period nanosecond. Then next item is internal priority value means which ingress queue should put. "-1" means wildcard. The last value optional specifies the maximum number of MSDU octets that are permitted to pass the gate during the specified time interval.
Base-time is not set will be as 0 as default, as result start time would be ((N + 1) * cycletime) which is the minimal of future time.Below example shows filtering a stream with destination mac address is 10:00:80:00:00:00 and ip type is ICMP, follow the action gate. The gate action would run with one close time slot which means always keep close. The time cycle is total 200000000ns. The base-time would calculate by:
1357000000000 + (N + 1) * cycletime
When the total value is the future time, it will be the start time. The cycletime here would be 200000000ns for this case.
Quoted 1 line
tc filter add dev eth0 parent ffff: protocol ip \
flower skip_hw ip_proto icmp dst_mac 10:00:80:00:00:00 \ action gate index 12 base-time 1357000000000 \ sched-entry CLOSE 200000000 -1 -1 \ clockid CLOCK_TAINOTE: This software simulator version not separate the admin/operation state machine. Update setting would overwrite stop the previos setting and waiting new cycle start.
[...]
net/sched/Kconfig config NET_ACT_CT · line 988
Quoted 13 lines
net/sched/Kconfig +5
-0@@ -981,6 +981,21 @@ config NET_ACT_CT To compile this code as a module, choose M here: the module will be called act_ct. +config NET_ACT_GATE + tristate "Frame gate list control tc action" + depends on NET_CLS_ACT + help + Say Y here to allow the control the ingress flow by the gate list
"to control"?
net/sched/Kconfig config NET_ACT_CT · line 989
Quoted 1 line
net/sched/Kconfig +1
-0+ control. The frame policing by the time gate list control open/close
Incomplete sentence.
net/sched/Kconfig config NET_ACT_CT · line 993
Quoted 4 lines
net/sched/Kconfig +4
-0+ cycle time. The manipulation will simulate the IEEE 802.1Qci stream + gate control behavior. The action could be offload by the tc flower + to hardware driver which the hardware own the capability of IEEE + 802.1Qci.
We do not mention offload for the other actions. I suggest to not to mention it here either.
Quoted 8 lines
net/sched/Kconfig +5
-0+ + If unsure, say N. + To compile this code as a module, choose M here: the + module will be called act_gate. +config
NET_IFE_SKBMARK
tristate "Support to encoding decoding skb mark on IFE action" depends onNET_ACT_IFE
[...]
RE: [EXT] Re: [v1,net-next 2/5] net: qos: introduce a gate control flow action
Hi Jiri,
I would update descriptions more clearly.
Quoted 31 lines
-----Original Message-----
From: Jiri Pirko <[email protected]>
Sent: 2020年3月24日 18:19
To: Po Liu <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; Claudiu Manoil <[email protected]>; Vladimir Oltean <[email protected]>; Alexandru Marginean <[email protected]>; Xiaoliang Yang <[email protected]>; Roy Zang <[email protected]>; Mingkai Hu <[email protected]>; Jerry Huang <[email protected]>; Leo Li <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]
Subject: [EXT] Re: [v1,net-next 2/5] net: qos: introduce a gate control flow actionCaution: EXT Email
Quoted 7 lines · Tue, Mar 24, 2020 at 04:47:40AM CET, [email protected] wrote:
Introduce a ingress frame gate control flow action. tc create a gate action would provide a gate list to control when open/close state. when the gate open state, the flow could pass but not when gate state is close. The driver would repeat the gate list cyclically. User also could assign a time point to start the gate list by the basetime parameter. if the basetime has passed current time, start time would calculate by the cycletime of the gate list.
Quoted 3 lines
Cannot decypher this either :/ Seriously, please make the patch descriptions readable.
Ok.
Quoted 15 lines
Also, a sentence starts with capital letter.
Quoted 11 lines
The action gate behavior try to keep according to the IEEE 802.1Qci spec. For the software simulation, require the user input the clock type.
Below is the setting example in user space. Tc filter a stream source ip address is 192.168.0.20 and gate action own two time slots. One is last 200ms gate open let frame pass another is last 100ms gate close let frames dropped. When the passed total frames over 8000000 bytes, it will dropped in one 200000000ns time slot.
Quoted 1 line
tc qdisc add dev eth0 ingress
Quoted 8 lines
Quoted 5 lines
Quoted 1 line
tc filter add dev eth0 parent ffff: protocol ip \
flower src_ip 192.168.0.20 \ action gate index 2 clockid CLOCK_TAI \ sched-entry OPEN 200000000 -1 8000000 \ sched-entry CLOSE 100000000 -1 -1The rest of the commands do not use capitals. Please lowercase these.
Ok.
Quoted 46 lines
Quoted 8 lines
Quoted 3 lines
tc chain del dev eth0 ingress chain 0
"sched-entry" follow the name taprio style. gate state is "OPEN"/"CLOSE". Follow the period nanosecond. Then next item is internal priority value means which ingress queue should put. "-1" means wildcard. The last value optional specifies the maximum number
of
Quoted 25 lines
MSDU octets that are permitted to pass the gate during the specified time interval.
Base-time is not set will be as 0 as default, as result start time would be ((N + 1) * cycletime) which is the minimal of future time.Below example shows filtering a stream with destination mac address is 10:00:80:00:00:00 and ip type is ICMP, follow the action gate. The gate action would run with one close time slot which means always keep close. The time cycle is total 200000000ns. The base-time would calculate by:
1357000000000 + (N + 1) * cycletime
When the total value is the future time, it will be the start time. The cycletime here would be 200000000ns for this case.
Quoted 1 line
tc filter add dev eth0 parent ffff: protocol ip \
flower skip_hw ip_proto icmp dst_mac 10:00:80:00:00:00 \ action gate index 12 base-time 1357000000000 \ sched-entry CLOSE 200000000 -1 -1 \ clockid CLOCK_TAINOTE: This software simulator version not separate the admin/operation state machine. Update setting would overwrite stop the previos setting and waiting new cycle start.
[...]
Quoted 8 lines
net/sched/Kconfig index renamed from net/sched/Kconfig +0
-0
bfbefb7bff9d..320471a0a21d100644net/sched/Kconfig +0
-0@@ -981,6 +981,21 @@ config NET_ACT_CT To compile this code as a module, choose M here: the module will be called act_ct.
Quoted 8 lines
net/sched/Kconfig config NET_ACT_CT
Quoted 6 lines
net/sched/Kconfig +6
-0+config NET_ACT_GATE + tristate "Frame gate list control tc action" + depends on NET_CLS_ACT + help + Say Y here to allow the control the ingress flow by the gate +list"to control"?
Ok.
Quoted 8 lines
net/sched/Kconfig config NET_ACT_CT · line 991
Quoted 4 lines
net/sched/Kconfig +2
-0+ control. The frame policing by the time gate list control + open/closeIncomplete sentence.
Quoted 7 lines
net/sched/Kconfig config NET_ACT_CT · line 993
Quoted 4 lines
net/sched/Kconfig +4
-0+ cycle time. The manipulation will simulate the IEEE 802.1Qci stream + gate control behavior. The action could be offload by the tc flower + to hardware driver which the hardware own the capability of IEEE + 802.1Qci.We do not mention offload for the other actions. I suggest to not to mention it here either.
Ok.
Quoted 12 lines
Quoted 10 lines
net/sched/Kconfig +5
-0+ + If unsure, say N. + To compile this code as a module, choose M here: the + module will be called act_gate. +config
NET_IFE_SKBMARK
tristate "Support to encoding decoding skb mark on IFE action" depends onNET_ACT_IFE[...]
Thanks!
Br,
Po Liu
PATCH 3/5 net: schedule: add action gate offloading
In reply to Po LiuAdd the gate action to the flow action entry. Add the gate parameters to the tc_setup_flow_action() queueing to the entries of flow_action_entry array provide to the driver.
--- include/net/flow_offload.h | 10 +++ include/net/tc_act/tc_gate.h | 115 +++++++++++++++++++++++++++++++++++ net/sched/cls_api.c | 33 ++++++++++ 3 files changed, 158 insertions(+)
include/net/flow_offload.h +10 -0
@@ -147,6 +147,7 @@ enum flow_action_id {FLOW_ACTION_MPLS_PUSH ,FLOW_ACTION_MPLS_POP ,FLOW_ACTION_MPLS_MANGLE , +FLOW_ACTION_GATE , NUM_FLOW_ACTIONS, };
@@ -254,6 +255,15 @@ struct flow_action_entry {u8 bos;u8 ttl; } mpls_mangle; +struct { +u32 index; +s32 prio; +u64 basetime; +u64 cycletime; +u64 cycletimeext; +u32 num_entries; +struct action_gate_entry *entries; + } gate; };struct flow_action_cookie *cookie;/* user defined action cookie */ };
include/net/tc_act/tc_gate.h +115 -0
125 lines not on this page · raw message
net/sched/cls_api.c +33 -0
@@ -39,6 +39,7 @@#include <net/tc_act/tc_skbedit.h> #include <net/tc_act/tc_ct.h> #include <net/tc_act/tc_mpls.h> +#include <net/tc_act/tc_gate.h> #include <net/flow_offload.h> extern const struct nla_policy rtm_tca_policy[TCA_MAX +1 ];
@@ -3522,6 +3523,27 @@ static void tcf_sample_get_group(struct flow_action_entry *entry,#endif } +static void tcf_gate_entry_destructor (void *priv) +{ +struct action_gate_entry *oe = priv; + +kfree (oe); +} + +static int tcf_gate_get_entries (struct flow_action_entry *entry, +const struct tc_action *act) +{ + entry->gate.entries =tcf_gate_get_list (act); + +if (!entry->gate.entries) +return -EINVAL ; + + entry->destructor = tcf_gate_entry_destructor; + entry->destructor_priv = entry->gate.entries; + +return 0 ; +} +int tc_setup_flow_action (struct flow_action *flow_action,const struct tcf_exts *exts) {
@@ -3668,6 +3690,17 @@ int tc_setup_flow_action(struct flow_action *flow_action, }else if (is_tcf_skbedit_priority (act)) { entry->id =FLOW_ACTION_PRIORITY ; entry->priority =tcf_skbedit_priority (act); + }else if (is_tcf_gate (act)) { + entry->id =FLOW_ACTION_GATE ; + entry->gate.index =tcf_gate_index (act); + entry->gate.prio =tcf_gate_prio (act); + entry->gate.basetime =tcf_gate_basetime (act); + entry->gate.cycletime =tcf_gate_cycletime (act); + entry->gate.cycletimeext =tcf_gate_cycletimeext (act); + entry->gate.num_entries =tcf_gate_num_entries (act); + err =tcf_gate_get_entries (entry, act); +if (err) + goto err_out; }else { err = -EOPNOTSUPP ; goto err_out_locked;
Signature · 1 line
2.17.1
PATCH 4/5 net: enetc: add hw tc hw offload features for PSPF capability
In reply to Po LiuThis patch is to let ethtool enable/disable the tc flower offload features. Hardware enetc has the feature of PSFP which is for per-stream policing. When enable the tc hw offloading feature, driver would enable the IEEE 802.1Qci feature. It is only set the register enable bit for this feature not enable for any entry of per stream filtering and stream gate or stream identify but get how much capabilities for each features.
--- drivers/net/ethernet/freescale/enetc/enetc.c | 23 ++++++++ drivers/net/ethernet/freescale/enetc/enetc.h | 55 +++++++++++++++++++ .../net/ethernet/freescale/enetc/enetc_hw.h | 17 ++++++ .../net/ethernet/freescale/enetc/enetc_pf.c | 8 +++ 4 files changed, 103 insertions(+)
drivers/net/ethernet/freescale/enetc/enetc.c +23 -0
@@ -756,6 +756,9 @@ void enetc_get_si_caps(struct enetc_si *si)if (val &ENETC_SIPCAPR0_QBV ) si->hw_features |=ENETC_SI_F_QBV ; + +if (val &ENETC_SIPCAPR0_PSFP ) + si->hw_features |=ENETC_SI_F_PSFP ; }static int enetc_dma_alloc_bdr (struct enetc_bdr *r,size_t bd_size)
@@ -1567,6 +1570,23 @@ static int enetc_set_rss(struct net_device *ndev, int en)return 0 ; } +static int enetc_set_psfp (struct net_device *ndev,int en) +{ +struct enetc_ndev_priv *priv =netdev_priv (ndev); + +if (en) { + priv->active_offloads |=ENETC_F_QCI ; +enetc_get_max_cap (priv); +enetc_psfp_enable (&priv->si->hw); + }else { + priv->active_offloads &= ~ENETC_F_QCI ; +memset (&priv->psfp_cap,0 ,sizeof (struct psfp_cap )); +enetc_psfp_disable (&priv->si->hw); + } + +return 0 ; +} +int enetc_set_features (struct net_device *ndev,netdev_features_t features) {
@@ -1575,6 +1595,9 @@ int enetc_set_features(struct net_device *ndev,if (changed &NETIF_F_RXHASH )enetc_set_rss (ndev, !!(features &NETIF_F_RXHASH )); +if (changed &NETIF_F_HW_TC ) +enetc_set_psfp (ndev, !!(features &NETIF_F_HW_TC )); +return 0 ; }
drivers/net/ethernet/freescale/enetc/enetc.h +55 -0
88 lines not on this page · raw message
drivers/net/ethernet/freescale/enetc/enetc_hw.h +17 -0
@@ -19,6 +19,7 @@#define ENETC_SICTR1 0x1c#define ENETC_SIPCAPR0 0x20#define ENETC_SIPCAPR0_QBV BIT(4) +#define ENETC_SIPCAPR0_PSFP BIT(9)#define ENETC_SIPCAPR0_RSS BIT(8)#define ENETC_SIPCAPR1 0x24#define ENETC_SITGTGR 0x30
@@ -228,6 +229,15 @@ enum enetc_bdr_type {TX, RX};#define ENETC_PM0_IFM_RLP (BIT(5) | BIT(11))#define ENETC_PM0_IFM_RGAUTO (BIT(15) | ENETC_PMO_IFM_RG | BIT(1))#define ENETC_PM0_IFM_XGMII BIT(12) +#define ENETC_PSIDCAPR 0x1b08 +#define ENETC_PSIDCAPR_MSK GENMASK(15, 0) +#define ENETC_PSFCAPR 0x1b18 +#define ENETC_PSFCAPR_MSK GENMASK(15, 0) +#define ENETC_PSGCAPR 0x1b28 +#define ENETC_PSGCAPR_GCL_MSK GENMASK(18, 16) +#define ENETC_PSGCAPR_SGIT_MSK GENMASK(15, 0) +#define ENETC_PFMCAPR 0x1b38 +#define ENETC_PFMCAPR_MSK GENMASK(15, 0)/* MAC counters */ #define ENETC_PM0_REOCT 0x8100
@@ -621,3 +631,10 @@ struct enetc_cbd {/* Port time specific departure */ #define ENETC_PTCTSDR (n) (0x1210 + 4 * (n))#define ENETC_TSDE BIT(31) + +/* PSFP setting */ +#define ENETC_PPSFPMR 0x11b00 +#define ENETC_PPSFPMR_PSFPEN BIT(0) +#define ENETC_PPSFPMR_VS BIT(1) +#define ENETC_PPSFPMR_PVC BIT(2) +#define ENETC_PPSFPMR_PVZC BIT(3)
drivers/net/ethernet/freescale/enetc/enetc_pf.c +8 -0
@@ -740,6 +740,14 @@ static void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,if (si->hw_features &ENETC_SI_F_QBV ) priv->active_offloads |=ENETC_F_QBV ; +if (si->hw_features &ENETC_SI_F_PSFP ) { + priv->active_offloads |=ENETC_F_QCI ; + ndev->features |=NETIF_F_HW_TC ; + ndev->hw_features |=NETIF_F_HW_TC ; +enetc_get_max_cap (priv); +enetc_psfp_enable (&si->hw); + } +/* pick up primary MAC address from SI */ enetc_get_primary_mac_addr (&si->hw, ndev->dev_addr); }
Signature · 1 line
2.17.1
kbuild test robot
build botRe: [v1,net-next 4/5] net: enetc: add hw tc hw offload features for PSPF capability
Thank you for the patch! Perhaps something to improve:
Hi Po,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on next-20200323]
[cannot apply to linux/master linus/master sparc-next/master ipvs/master v5.6-rc7 v5.6-rc6 v5.6-rc5 v5.6-rc7] [if your patch is applied to the wrong git tree, please drop us a note to help
improve the system. BTW, we also suggest to use '--base' option to specify the
base tree in git format-patch, please see https://stackoverflow.com/a/37406982]
url: https://github.com/0day-ci/linux/commits/Po-Liu/Introduce-a-flow-gate-control-action-and-apply-IEEE/20200324-121156 base: 5149100c3aebe5e640d6ff68e0b5e5a7eb8638e0 config: arm64-defconfig (attached as .config) compiler: aarch64-linux-gcc (GCC) 9.2.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
GCC_VERSION=9.2.0 make.cross ARCH=arm64
If you fix the issue, kindly add following tag
All warnings (new ones prefixed by >>):
In file included from drivers/net/ethernet/freescale/enetc/enetc_ptp.c:8:
drivers/net/ethernet/freescale/enetc/enetc.h: In function 'enetc_psfp_enable':
>> drivers/net/ethernet/freescale/enetc/enetc.h:345:9: warning: 'return' with a value, in function returning void [-Wreturn-type]
345 | return 0;
| ^
drivers/net/ethernet/freescale/enetc/enetc.h:343:20: note: declared here
343 | static inline void enetc_psfp_enable(struct enetc_hw *hw)
| ^~~~~~~~~~~~~~~~~
drivers/net/ethernet/freescale/enetc/enetc.h: In function 'enetc_psfp_disable':
drivers/net/ethernet/freescale/enetc/enetc.h:350:9: warning: 'return' with a value, in function returning void [-Wreturn-type]
350 | return 0;
| ^ drivers/net/ethernet/freescale/enetc/enetc.h:348:20: note: declared here
348 | static inline void enetc_psfp_disable(struct enetc_hw *hw)
| ^~~~~~~~~~~~~~~~~~
vim +/return +345 drivers/net/ethernet/freescale/enetc/enetc.h
328
329 static inline void enetc_psfp_disable(struct enetc_hw *hw)
330 {
331 enetc_wr(hw, ENETC_PPSFPMR, enetc_rd(hw, ENETC_PPSFPMR)
332 & ~ENETC_PPSFPMR_PSFPEN & ~ENETC_PPSFPMR_VS
333 & ~ENETC_PPSFPMR_PVC & ~ENETC_PPSFPMR_PVZC);
334 }
335 #else
336 #define enetc_setup_tc_taprio(ndev, type_data) -EOPNOTSUPP
337 #define enetc_sched_speed_set(ndev) (void)0
338 #define enetc_setup_tc_cbs(ndev, type_data) -EOPNOTSUPP
339 #define enetc_setup_tc_txtime(ndev, type_data) -EOPNOTSUPP
340 #define enetc_get_max_cap(p) \
341 memset(&((p)->psfp_cap), 0, sizeof(struct psfp_cap))
342
343 static inline void enetc_psfp_enable(struct enetc_hw *hw)
344 {
> 345 return 0;
346 }
347
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org
Jiri Pirko
maintainerRe: [v1,net-next 4/5] net: enetc: add hw tc hw offload features for PSPF capability
In reply to Po LiuTue, Mar 24, 2020 at 04:47:42AM CET, [email protected] wrote:
[...]
drivers/net/ethernet/freescale/enetc/enetc.h int enetc_setup_tc_taprio(struct net_device *ndev, void *type_data); · line 306
Quoted 8 lines
drivers/net/ethernet/freescale/enetc/enetc.h +4
-0@@ -289,9 +300,53 @@ int enetc_setup_tc_taprio(struct net_device *ndev, void *type_data);void enetc_sched_speed_set (struct net_device *ndev);int enetc_setup_tc_cbs (struct net_device *ndev,void *type_data);int enetc_setup_tc_txtime (struct net_device *ndev,void *type_data); + +static inline void enetc_get_max_cap (struct enetc_ndev_priv *priv) +{ +u32 reg =0 ;
Pointless init.
Quoted 9 lines
drivers/net/ethernet/freescale/enetc/enetc.h +9
-0+ + reg =enetc_port_rd (&priv->si->hw,ENETC_PSIDCAPR ); + priv->psfp_cap.max_streamid = reg &ENETC_PSIDCAPR_MSK ; +/* Port stream filter capability */ + reg =enetc_port_rd (&priv->si->hw,ENETC_PSFCAPR ); + priv->psfp_cap.max_psfp_filter = reg &ENETC_PSFCAPR_MSK ; +/* Port stream gate capability */ + reg =enetc_port_rd (&priv->si->hw,ENETC_PSGCAPR ); + priv->psfp_cap.max_psfp_gate = (reg &ENETC_PSGCAPR_SGIT_MSK );
drivers/net/ethernet/freescale/enetc/enetc.h int enetc_setup_tc_taprio(struct net_device *ndev, void *type_data); · line 325
Quoted 10 lines
drivers/net/ethernet/freescale/enetc/enetc.h +10
-0+ priv->psfp_cap.max_psfp_gatelist = (reg &ENETC_PSGCAPR_GCL_MSK ) >>16 ; +/* Port flow meter capability */ + reg =enetc_port_rd (&priv->si->hw,ENETC_PFMCAPR ); + priv->psfp_cap.max_psfp_meter = reg &ENETC_PFMCAPR_MSK ; +} + +static inline void enetc_psfp_enable (struct enetc_hw *hw) +{ +enetc_wr (hw,ENETC_PPSFPMR ,enetc_rd (hw,ENETC_PPSFPMR ) + | ENETC_PPSFPMR_PSFPEN | ENETC_PPSFPMR_VS
Hmm, I think it is better to have "|" at the end of the line".
drivers/net/ethernet/freescale/enetc/enetc.h int enetc_setup_tc_taprio(struct net_device *ndev, void *type_data); · line 332
Quoted 7 lines
drivers/net/ethernet/freescale/enetc/enetc.h +7
-0+ | ENETC_PPSFPMR_PVC | ENETC_PPSFPMR_PVZC); +} + +static inline void enetc_psfp_disable (struct enetc_hw *hw) +{ +enetc_wr (hw,ENETC_PPSFPMR ,enetc_rd (hw,ENETC_PPSFPMR ) + & ~ENETC_PPSFPMR_PSFPEN & ~ENETC_PPSFPMR_VS
Same here.
Quoted 7 lines
drivers/net/ethernet/freescale/enetc/enetc.h +2
-0+ & ~ENETC_PPSFPMR_PVC & ~ENETC_PPSFPMR_PVZC); +}#else #define enetc_setup_tc_taprio (ndev, type_data) -EOPNOTSUPP#define enetc_sched_speed_set (ndev) (void)0#define enetc_setup_tc_cbs (ndev, type_data) -EOPNOTSUPP#define enetc_setup_tc_txtime (ndev, type_data) -EOPNOTSUPP
[...]
PATCH 5/5 net: enetc: add tc flower psfp offload driver
In reply to Po LiuThis patch is to add tc flower offload for the enetc IEEE 802.1Qci(PSFP) function. There are four main feature parts to implement the flow policing and filtering for ingress flow with IEEE 802.1Qci features. They are stream identify(this is defined in the P802.1cb exactly but needed for 802.1Qci), stream filtering, stream gate and flow metering. Each function block includes many entries by index to assign parameters. So for one frame would be filtered by stream identify first, then flow into stream filter block by the same handle between stream identify and stream filtering. Then flow into stream gate control which assigned by the stream filtering entry. And then policing by the gate and limited by the max sdu in the filter block(optional). At last, policing by the flow metering block, index choosing at the fitering block. So you can see that each entry of block may link to many upper entries since they can be assigned same index means more streams want to share the same feature in the stream filtering or stream gate or flow metering.
To implement such features, each stream filtered by source/destination mac address, some stream maybe also plus the vlan id value would be treated as one flow chain. This would be identified by the chain_index which already in the tc filter concept. Driver would maintain this chain and also with gate modules. The stream filter entry create by the gate index and flow meter(optional) entry id and also one priority value. Offloading only transfer the gate action and flow filtering parameters. Driver would create (or search same gate id and flow meter id and
priority) one stream filter entry to set to the hardware. So stream
filtering do not need transfer by the action offloading. This architecture is same with tc filter and actions relationship. tc filter maintain the list for each flow feature by keys. And actions maintain by the action list.
Below showing a example commands by tc:
Quoted 3 lines
tc qdisc add dev eth0 ingress
ip link set eth0 address 10:00:80:00:00:00
tc filter add dev eth0 parent ffff: protocol ip chain 11 \
flower skip_sw dst_mac 10:00:80:00:00:00 \ action gate index 10 \ sched-entry OPEN 200000000 1 8000000 \ sched-entry CLOSE 100000000 -1 -1
Command means to set the dst_mac 10:00:80:00:00:00 to index 11 of stream identify module. And the set the gate index 10 of stream gate module. The gate list is keeping OPEN state 200ms, and through the frames to the ingress queue 1, and max octets are the 8Mbytes.
---
Changes from RFC:
Acked by Vinicius Costa Gomes
- Use refcount kernel lib
- Update stream gate check code position
- Update reduce ref names more clear
drivers/net/ethernet/freescale/enetc/enetc.c | 25 +- drivers/net/ethernet/freescale/enetc/enetc.h | 39 +- .../net/ethernet/freescale/enetc/enetc_hw.h | 142 +++ .../net/ethernet/freescale/enetc/enetc_pf.c | 4 +- .../net/ethernet/freescale/enetc/enetc_qos.c | 1074 +++++++++++++++++ 5 files changed, 1269 insertions(+), 15 deletions(-)
drivers/net/ethernet/freescale/enetc/enetc.c +17 -8
@@ -1521,6 +1521,8 @@ int enetc_setup_tc(struct net_device *ndev, enum tc_setup_type type,return enetc_setup_tc_cbs (ndev, type_data);case TC_SETUP_QDISC_ETF :return enetc_setup_tc_txtime (ndev, type_data); +case TC_SETUP_BLOCK : +return enetc_setup_tc_psfp (ndev, type_data);default :return -EOPNOTSUPP ; }
@@ -1573,17 +1575,23 @@ static int enetc_set_rss(struct net_device *ndev, int en)static int enetc_set_psfp (struct net_device *ndev,int en) {struct enetc_ndev_priv *priv =netdev_priv (ndev); +int err;if (en) { + err =enetc_psfp_enable (priv); +if (err) +return err; + priv->active_offloads |=ENETC_F_QCI ; -enetc_get_max_cap (priv); -enetc_psfp_enable (&priv->si->hw); - }else { - priv->active_offloads &= ~ENETC_F_QCI ; -memset (&priv->psfp_cap,0 ,sizeof (struct psfp_cap )); -enetc_psfp_disable (&priv->si->hw); +return 0 ; } + err =enetc_psfp_disable (priv); +if (err) +return err; + + priv->active_offloads &= ~ENETC_F_QCI ; +return 0 ; }
@@ -1591,14 +1599,15 @@ int enetc_set_features(struct net_device *ndev, netdev_features_t features) {netdev_features_t changed = ndev->features ^ features; +int err =0 ;if (changed &NETIF_F_RXHASH )enetc_set_rss (ndev, !!(features &NETIF_F_RXHASH ));if (changed &NETIF_F_HW_TC ) -enetc_set_psfp (ndev, !!(features &NETIF_F_HW_TC )); + err =enetc_set_psfp (ndev, !!(features &NETIF_F_HW_TC )); -return 0 ; +return err; }#ifdef CONFIG_FSL_ENETC_PTP_CLOCK
drivers/net/ethernet/freescale/enetc/enetc.h +35 -4
@@ -300,6 +300,11 @@ int enetc_setup_tc_taprio(struct net_device *ndev, void *type_data);void enetc_sched_speed_set (struct net_device *ndev);int enetc_setup_tc_cbs (struct net_device *ndev,void *type_data);int enetc_setup_tc_txtime (struct net_device *ndev,void *type_data); +int enetc_setup_tc_block_cb (enum tc_setup_type type,void *type_data, +void *cb_priv); +int enetc_setup_tc_psfp (struct net_device *ndev,void *type_data); +int enetc_psfp_init (struct enetc_ndev_priv *priv); +int enetc_psfp_clean (struct enetc_ndev_priv *priv);static inline void enetc_get_max_cap (struct enetc_ndev_priv *priv) {
@@ -319,33 +324,59 @@ static inline void enetc_get_max_cap(struct enetc_ndev_priv *priv) priv->psfp_cap.max_psfp_meter = reg &ENETC_PFMCAPR_MSK ; } -static inline void enetc_psfp_enable (struct enetc_hw *hw) +static inline int enetc_psfp_enable (struct enetc_ndev_priv *priv) { +struct enetc_hw *hw = &priv->si->hw; +int err; + +enetc_get_max_cap (priv); + + err =enetc_psfp_init (priv); +if (err) +return err; +enetc_wr (hw,ENETC_PPSFPMR ,enetc_rd (hw,ENETC_PPSFPMR ) |ENETC_PPSFPMR_PSFPEN |ENETC_PPSFPMR_VS |ENETC_PPSFPMR_PVC |ENETC_PPSFPMR_PVZC ); + +return 0 ; } -static inline void enetc_psfp_disable (struct enetc_hw *hw) +static inline int enetc_psfp_disable (struct enetc_ndev_priv *priv) { +struct enetc_hw *hw = &priv->si->hw; +int err; + + err =enetc_psfp_clean (priv); +if (err) +return err; +enetc_wr (hw,ENETC_PPSFPMR ,enetc_rd (hw,ENETC_PPSFPMR ) & ~ENETC_PPSFPMR_PSFPEN & ~ENETC_PPSFPMR_VS & ~ENETC_PPSFPMR_PVC & ~ENETC_PPSFPMR_PVZC ); + +memset (&priv->psfp_cap,0 ,sizeof (struct psfp_cap )); + +return 0 ; } +#else #define enetc_setup_tc_taprio (ndev, type_data) -EOPNOTSUPP#define enetc_sched_speed_set (ndev) (void)0#define enetc_setup_tc_cbs (ndev, type_data) -EOPNOTSUPP#define enetc_setup_tc_txtime (ndev, type_data) -EOPNOTSUPP +#define enetc_setup_tc_psfp (ndev, type_data) -EOPNOTSUPP +#define enetc_setup_tc_block_cb NULL +#define enetc_get_max_cap (p) \ memset(&((p)->psfp_cap), 0, sizeof(struct psfp_cap)) -static inline void enetc_psfp_enable (struct enetc_hw *hw) +static inline int enetc_psfp_enable (struct enetc_hw *hw) {return 0 ; } -static inline void enetc_psfp_disable (struct enetc_hw *hw) +static inline int enetc_psfp_disable (struct enetc_hw *hw) {return 0 ; }
drivers/net/ethernet/freescale/enetc/enetc_hw.h +142 -0
161 lines not on this page · raw message
drivers/net/ethernet/freescale/enetc/enetc_pf.c +1 -3
@@ -740,12 +740,10 @@ static void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,if (si->hw_features &ENETC_SI_F_QBV ) priv->active_offloads |=ENETC_F_QBV ; -if (si->hw_features &ENETC_SI_F_PSFP ) { +if (si->hw_features &ENETC_SI_F_PSFP && !enetc_psfp_enable (priv)) { priv->active_offloads |=ENETC_F_QCI ; ndev->features |=NETIF_F_HW_TC ; ndev->hw_features |=NETIF_F_HW_TC ; -enetc_get_max_cap (priv); -enetc_psfp_enable (&si->hw); }/* pick up primary MAC address from SI */
drivers/net/ethernet/freescale/enetc/enetc_qos.c +1074 -0
1,083 lines not on this page · raw message
Signature · 1 line
2.17.1
kbuild test robot
build botRe: [v1,net-next 5/5] net: enetc: add tc flower psfp offload driver
Thank you for the patch! Yet something to improve:
Hi Po,
Thank you for the patch! Yet something to improve:
[auto build test ERROR on next-20200323]
[cannot apply to linux/master linus/master sparc-next/master ipvs/master v5.6-rc7 v5.6-rc6 v5.6-rc5 v5.6-rc7] [if your patch is applied to the wrong git tree, please drop us a note to help
improve the system. BTW, we also suggest to use '--base' option to specify the
base tree in git format-patch, please see https://stackoverflow.com/a/37406982]
url: https://github.com/0day-ci/linux/commits/Po-Liu/Introduce-a-flow-gate-control-action-and-apply-IEEE/20200324-121156 base: 5149100c3aebe5e640d6ff68e0b5e5a7eb8638e0 config: arm64-defconfig (attached as .config) compiler: aarch64-linux-gcc (GCC) 9.2.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
GCC_VERSION=9.2.0 make.cross ARCH=arm64
If you fix the issue, kindly add following tag
All errors (new ones prefixed by >>):
drivers/net/ethernet/freescale/enetc/enetc_pf.c: In function 'enetc_pf_netdev_setup':
>> drivers/net/ethernet/freescale/enetc/enetc_pf.c:743:62: error: passing argument 1 of 'enetc_psfp_enable' from incompatible pointer type [-Werror=incompatible-pointer-types]
743 | if (si->hw_features & ENETC_SI_F_PSFP && !enetc_psfp_enable(priv)) {
| ^~~~
| |
| struct enetc_ndev_priv * In file included from drivers/net/ethernet/freescale/enetc/enetc_pf.h:4,
from drivers/net/ethernet/freescale/enetc/enetc_pf.c:8:
drivers/net/ethernet/freescale/enetc/enetc.h:374:54: note: expected 'struct enetc_hw *' but argument is of type 'struct enetc_ndev_priv *'
374 | static inline int enetc_psfp_enable(struct enetc_hw *hw)
| ~~~~~~~~~~~~~~~~~^~
cc1: some warnings being treated as errors
--
drivers/net/ethernet/freescale/enetc/enetc.c: In function 'enetc_set_psfp':
>> drivers/net/ethernet/freescale/enetc/enetc.c:1581:27: error: passing argument 1 of 'enetc_psfp_enable' from incompatible pointer type [-Werror=incompatible-pointer-types]
1581 | err = enetc_psfp_enable(priv);
| ^~~~
| |
| struct enetc_ndev_priv *
In file included from drivers/net/ethernet/freescale/enetc/enetc.c:4:
drivers/net/ethernet/freescale/enetc/enetc.h:374:54: note: expected 'struct enetc_hw *' but argument is of type 'struct enetc_ndev_priv *'
374 | static inline int enetc_psfp_enable(struct enetc_hw *hw)
| ~~~~~~~~~~~~~~~~~^~
>> drivers/net/ethernet/freescale/enetc/enetc.c:1589:27: error: passing argument 1 of 'enetc_psfp_disable' from incompatible pointer type [-Werror=incompatible-pointer-types]
1589 | err = enetc_psfp_disable(priv);
| ^~~~
| |
| struct enetc_ndev_priv * In file included from drivers/net/ethernet/freescale/enetc/enetc.c:4:
drivers/net/ethernet/freescale/enetc/enetc.h:379:55: note: expected 'struct enetc_hw *' but argument is of type 'struct enetc_ndev_priv *'
379 | static inline int enetc_psfp_disable(struct enetc_hw *hw)
| ~~~~~~~~~~~~~~~~~^~
cc1: some warnings being treated as errors
vim +/enetc_psfp_enable +743 drivers/net/ethernet/freescale/enetc/enetc_pf.c
703
704 static void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,
705 const struct net_device_ops *ndev_ops)
706 {
707 struct enetc_ndev_priv *priv = netdev_priv(ndev);
708
709 SET_NETDEV_DEV(ndev, &si->pdev->dev);
710 priv->ndev = ndev;
711 priv->si = si;
712 priv->dev = &si->pdev->dev;
713 si->ndev = ndev;
714
715 priv->msg_enable = (NETIF_MSG_WOL << 1) - 1;
716 ndev->netdev_ops = ndev_ops;
717 enetc_set_ethtool_ops(ndev);
718 ndev->watchdog_timeo = 5 * HZ;
719 ndev->max_mtu = ENETC_MAX_MTU;
720
721 ndev->hw_features = NETIF_F_SG | NETIF_F_RXCSUM | NETIF_F_HW_CSUM |
722 NETIF_F_HW_VLAN_CTAG_TX | NETIF_F_HW_VLAN_CTAG_RX |
723 NETIF_F_LOOPBACK;
724 ndev->features = NETIF_F_HIGHDMA | NETIF_F_SG |
725 NETIF_F_RXCSUM | NETIF_F_HW_CSUM |
726 NETIF_F_HW_VLAN_CTAG_TX |
727 NETIF_F_HW_VLAN_CTAG_RX |
728 NETIF_F_HW_VLAN_CTAG_FILTER;
729
730 if (si->num_rss)
731 ndev->hw_features |= NETIF_F_RXHASH;
732
733 if (si->errata & ENETC_ERR_TXCSUM) {
734 ndev->hw_features &= ~NETIF_F_HW_CSUM;
735 ndev->features &= ~NETIF_F_HW_CSUM;
736 }
737
738 ndev->priv_flags |= IFF_UNICAST_FLT;
739
740 if (si->hw_features & ENETC_SI_F_QBV)
741 priv->active_offloads |= ENETC_F_QBV;
742
> 743 if (si->hw_features & ENETC_SI_F_PSFP && !enetc_psfp_enable(priv)) { 744 priv->active_offloads |= ENETC_F_QCI;
745 ndev->features |= NETIF_F_HW_TC;
746 ndev->hw_features |= NETIF_F_HW_TC;
747 }
748
749 /* pick up primary MAC address from SI */
750 enetc_get_primary_mac_addr(&si->hw, ndev->dev_addr);
751 }
752
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org
PATCH 1/2 iproute2:tc:action: add a gate control action
In reply to Po LiuIntroduce a ingress frame gate control flow action. tc create a gate action would provide a gate list to control when open/close state. when the gate open state, the flow could pass but not when gate state is close. The driver would repeat the gate list cyclely. User also could assign a time point to start the gate list by the basetime parameter. if the basetime has passed current time, start time would calculate by the cycletime of the gate list.
The behavior try to keep according to the IEEE 802.1Qci spec. For the software emulation, require the user input the clock type.
Below is the setting example in user space. Tc filter a stream source ip address is 192.168.0.20 and gate action own two time slots. One is last 200ms gate open let frame pass another is last 100ms gate close let frames dropped.
Quoted 2 lines
tc qdisc add dev eth0 ingress
tc filter add dev eth0 parent ffff: protocol ip \
flower src_ip 192.168.0.20 \
action gate index 2 clockid CLOCK_TAI \
sched-entry OPEN 200000000 -1 -1 \
sched-entry CLOSE 100000000 -1 -1Quoted 1 line
tc chain del dev eth0 ingress chain 0
"sched-entry" follow the name taprio style. gate state is "OPEN"/"CLOSE". Follow the period nanosecond. Then next -1 is internal priority value means which ingress queue should put. "-1" means wildcard. The last value optional specifies the maximum number of MSDU octets that are permitted to pass the gate during the specified time interval.
Below example shows filtering a stream with destination mac address is 10:00:80:00:00:00 and ip type is ICMP, follow the action gate. The gate action would run with one close time slot which means always keep close. The time cycle is total 200000000ns. The base-time would calculate by:
1357000000000 + (N + 1) * cycletime
When the total value is the future time, it will be the start time. The cycletime here would be 200000000ns for this case.
Quoted 1 line
tc filter add dev eth0 parent ffff: protocol ip \
flower skip_hw ip_proto icmp dst_mac 10:00:80:00:00:00 \
action gate index 12 base-time 1357000000000 \
sched-entry CLOSE 200000000 -1 -1 \
clockid CLOCK_TAI--- include/uapi/linux/pkt_cls.h | 1 + include/uapi/linux/tc_act/tc_gate.h | 47 +++ tc/Makefile | 1 + tc/m_gate.c | 483 ++++++++++++++++++++++++++++ 4 files changed, 532 insertions(+) create mode 100644 include/uapi/linux/tc_act/tc_gate.h create mode 100644 tc/m_gate.c
include/uapi/linux/pkt_cls.h +1 -0
@@ -106,6 +106,7 @@ enum tca_id {TCA_ID_SAMPLE =TCA_ACT_SAMPLE ,TCA_ID_CTINFO ,TCA_ID_MPLS , +TCA_ID_GATE ,TCA_ID_CT ,/* other actions go here */ __TCA_ID_MAX =255
include/uapi/linux/tc_act/tc_gate.h new file +47 -0
47 lines not on this page · raw message
tc/Makefile +1 -0
@@ -54,6 +54,7 @@ TCMODULES += m_bpf.o TCMODULES += m_tunnel_key.o TCMODULES += m_sample.o TCMODULES += m_ct.o +TCMODULES += m_gate.o TCMODULES += p_ip.o TCMODULES += p_ip6.o TCMODULES += p_icmp.o
tc/m_gate.c new file +483 -0
483 lines not on this page · raw message
Signature · 1 line
2.17.1
Re: [v1,iproute2 1/2] iproute2:tc:action: add a gate control action
include/uapi/linux/pkt_cls.h enum tca_id { · line 112
Quoted 12 lines · On Tue, 24 Mar 2020 11:47:44 +0800 Po Liu <[email protected]> wrote:
include/uapi/linux/pkt_cls.h +1
-0@@ -106,6 +106,7 @@ enum tca_id {TCA_ID_SAMPLE =TCA_ACT_SAMPLE ,TCA_ID_CTINFO ,TCA_ID_MPLS , +TCA_ID_GATE ,TCA_ID_CT ,/* other actions go here */ __TCA_ID_MAX =255
All uapi headers need to come from checked in kernel.
This is an example of why, you have an out of date version because this version would have broken ABI.
This patch should be against iproute2-next since because it depends on net-next.
RE: [EXT] Re: [v1,iproute2 1/2] iproute2:tc:action: add a gate control action
Hi Stephen,
Quoted 40 lines
-----Original Message-----
From: Stephen Hemminger <[email protected]>
Sent: 2020年3月25日 5:59
To: Po Liu <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; Claudiu Manoil <[email protected]>; Vladimir Oltean <[email protected]>; Alexandru Marginean <[email protected]>; Xiaoliang Yang <[email protected]>; Roy Zang <[email protected]>; Mingkai Hu <[email protected]>; Jerry Huang <[email protected]>; Leo Li <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; David Ahern <[email protected]>
Subject: [EXT] Re: [v1,iproute2 1/2] iproute2:tc:action: add a gate control actionCaution: EXT Email
include/uapi/linux/pkt_cls.h enum tca_id { · line 112
Quoted 12 lines · On Tue, 24 Mar 2020 11:47:44 +0800 Po Liu <[email protected]> wrote:
(quoted diff) +0
-0
b/include/uapi/linux/pkt_cls.hindex a6aa466..7a047a9 100644include/uapi/linux/pkt_cls.h +1
-0@@ -106,6 +106,7 @@ enum tca_id {TCA_ID_SAMPLE =TCA_ACT_SAMPLE ,TCA_ID_CTINFO ,TCA_ID_MPLS , +TCA_ID_GATE ,TCA_ID_CT ,/* other actions go here */ __TCA_ID_MAX =255 All uapi headers need to come from checked in kernel.
Quoted 6 lines
This is an example of why, you have an out of date version because this version would have broken ABI.
This patch should be against iproute2-next since because it depends on net-next.
I'll keep up with the iproute2-next branch. Thanks!
Br,
Po Liu
PATCH 2/2 iproute2: add gate action man page
In reply to Po Liu--- man/man8/tc-gate.8 | 106 +++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 106 insertions(+) create mode 100644 man/man8/tc-gate.8
man/man8/tc-gate.8 new file +106 -0
@@ -0,0 +1,106 @@ +.TH GATE 8 "12 Mar 2020" "iproute2" "Linux" +.SH NAME +gate \- Stream Gate Action +.SH SYNOPSIS +.B tc " ... " action gate +.ti +8 +.B [ base-time +BASETIME ] +.B [ clockid +CLOCKID ] +.ti +8 +.B sched-entry +<gate state> <interval 1> <internal priority> <max octets> +.ti +8 +.B sched-entry +<gate state> <interval 2> <internal priority> <max octets> +.ti +8 +.B sched-entry +<gate state> <interval 3> <internal priority> <max octets> +.ti +8 +.B ...... +.ti +8 +.B sched-entry +<gate state> <interval N> <internal priority> <max octets> + +.SH DESCRIPTION +GATE action would provide a gate list to control when traffic keep +open/close state. when the gate open state, the flow could pass but +not when gate state is close. The driver would repeat the gate list +periodically. User also could assign a time point to start the gate +list by the basetime parameter. if the basetime has passed current +time, start time would calculate by the cycletime of the gate list. + +.SH PARAMETERS + +.TP +base-time +.br +Specifies the instant in nanoseconds, defining the time when the schedule +starts. If 'base-time' is a time in the past, the schedule will start at + +base-time + (N * cycle-time) + +where N is the smallest integer so the resulting time is greater than +"now", and "cycle-time" is the sum of all the intervals of the entries +in the schedule. Without base-time specified, will default to be 0. + +.TP +clockid +.br +Specifies the clock to be used by qdisc's internal timer for measuring +time and scheduling events. Not valid if using for offloading filter. +For example, tc filter command with +.B skip_sw parameter. + +.TP +sched-entry +.br +There may multiple +.B sched-entry +parameters in a single schedule. Each one has the format: + +sched-entry <gate state> <interval> <internal priority> <max octets> + +.br +<gate state> means gate states. 'OPEN' keep gate open, 'CLOSE' keep gate close. +.br +<interval> means how much nano seconds for this time slot. +.br +<internal priority> means internal priority value. Present of the +internal receiving queue for this stream. "-1" means wildcard. +.br +<max octets> means how many octets size for this time slot. Dropped +if overlimited. "-1" means wildcard. + +.SH EXAMPLES + +The following example shows tc filter frames source ip match to the +192.168.0.20 will be passed at offset time 0 last 200000000ns and will +be dropped at the offset time 200000000ns and last 100000000ns. Then +run the gate periodically. The schedule will start at instant 200000000000 +using the reference CLOCK_TAI. The schedule is composed of three entries +each of 300us duration. + +.EX +# tc filter add dev eth0 parent ffff: protocol ip \\ + flower skip_hw src_ip 192.168.0.20 \\ + action gate index 2 clockid CLOCK_TAI \\ + base-time 200000000000 \\ + sched-entry OPEN 200000000 -1 -1 \\ + sched-entry CLOSE 100000000 -1 -1 + +.EE + +Following is an example to filter a stream source mac match to the +10:00:80:00:00:00 will be dropped at any time. + +.EX +#tc filter add dev eth0 parent ffff: protocol ip chain 14 \\ + flower skip_sw dst_mac 10:00:80:00:00:00 \\ + action gate index 12 sched-entry close 200000000 -1 -1 + +.EE + +.SH AUTHORS +Po Liu <[email protected]>
Signature · 1 line
2.17.1
v2 cover Introduce a flow gate control action and apply IEEE
Changes from V1:
0000: Update description make it more clear
0001: Removed 'add update dropped stats' patch, will provide pull request as standalone patches.
0001: Update commit description make it more clear ack by Jiri Pirko. 0001: No changes
0003: Fix some code style ack by Jiri Pirko.
0004: Fix enetc_psfp_enable/disable parameter type ack by test robot
iprout2 command patches:
Not attach with these serial patches, will provide separate pull
request after kernel accept these patches.
Changes from RFC:
0000: Reduce to 5 patches and remove the 4 max frame size offload and flow metering in the policing offload action, Only keep gate action offloading implementation.
0001: No changes.
0002:
- fix kfree lead ack by Jakub Kicinski and Cong Wang
- License fix from Jakub Kicinski and Stephen Hemminger
- Update example in commit acked by Vinicius Costa Gomes
- Fix the rcu protect in tcf_gate_act() acked by Vinicius
0003: No changes
0004: No changes
0005:
Acked by Vinicius Costa Gomes
- Use refcount kernel lib
- Update stream gate check code position
- Update reduce ref names more clear
iprout2 command patches:
0000: Update license expression and add gate id
0001: Add tc action gate man page
--------------------------------------------------------------------
These patches add stream gate action policing in IEEE802.1Qci (Per-Stream Filtering and Policing) software support and hardware offload support in tc flower, and implement the stream identify, stream filtering and stream gate filtering action in the NXP ENETC ethernet driver. Per-Stream Filtering and Policing (PSFP) specifies flow policing and filtering for ingress flows, and has three main parts:
1. The stream filter instance table consists of an ordered list of
stream filters that determine the filtering and policing actions that are to be applied to frames received on a specific stream. The main elements are stream gate id, flow metering id and maximum SDU size.
2. The stream gate function setup a gate list to control ingress traffic
class open/close state. When the gate is running at open state, the flow could pass but dropped when gate state is running to close. User setup a bastime to tell gate when start running the entry list, then the hardware would periodiclly. There is no compare qdisc action support.
3. Flow metering is two rates two buckets and three-color marker to
policing the frames. Flow metering instance are as specified in the algorithm in MEF10.3. The most likely qdisc action is policing action.
The first patch introduces an ingress frame flow control gate action, for the point 2. The tc gate action maintains the open/close state gate list, allowing flows to pass when the gate is open. Each gate action may policing one or more qdisc filters. When the start time arrived, The driver would repeat the gate list periodiclly. User can assign a passed time, the driver would calculate a new future time by the cycletime of the gate list.
The 0002 patch introduces the gate flow hardware offloading.
The 0003 patch adds support control the on/off for the tc flower offloading by ethtool.
The 0004 patch implement the stream identify and stream filtering and stream gate filtering action in the NXP ENETC ethernet driver. Tc filter command provide filtering keys with MAC address and VLAN id. These keys would be set to stream identify instance entry. Stream gate instance entry would refer the gate action parameters. Stream filter instance entry would refer the stream gate index and assign a stream handle value matches to the stream identify instance.
Po Liu (4):
net: qos: introduce a gate control flow action
net: schedule: add action gate offloading
net: enetc: add hw tc hw offload features for PSPF capability
net: enetc: add tc flower psfp offload driver
drivers/net/ethernet/freescale/enetc/enetc.c | 34 +- drivers/net/ethernet/freescale/enetc/enetc.h | 86 ++ .../net/ethernet/freescale/enetc/enetc_hw.h | 159 +++ .../net/ethernet/freescale/enetc/enetc_pf.c | 6 + .../net/ethernet/freescale/enetc/enetc_qos.c | 1070 +++++++++++++++++ include/net/flow_offload.h | 10 + include/net/tc_act/tc_gate.h | 169 +++ include/uapi/linux/pkt_cls.h | 1 + include/uapi/linux/tc_act/tc_gate.h | 47 + net/sched/Kconfig | 13 + net/sched/Makefile | 1 + net/sched/act_gate.c | 647 ++++++++++ net/sched/cls_api.c | 33 + 13 files changed, 2275 insertions(+), 1 deletion(-) create mode 100644 include/net/tc_act/tc_gate.h create mode 100644 include/uapi/linux/tc_act/tc_gate.h create mode 100644 net/sched/act_gate.c
Signature · 1 line
2.17.1
PATCH v2 1/4 net: qos: introduce a gate control flow action
Introduce a ingress frame gate control flow action.
Tc gate action does the work like this:
Assume there is a gate allow specified ingress frames can be passed at specific time slot, and be dropped at specific time slot. Tc filter chooses the ingress frames, and tc gate action would specify what slot does these frames can be passed to device and what time slot would be dropped.
Tc gate action would provide an entry list to tell how much time gate keep open and how much time gate keep state close. Gate action also assign a start time to tell when the entry list start. Then driver would repeat the gate entry list cyclically.
For the software simulation, gate action requires the user assign a time clock type.
Below is the setting example in user space. Tc filter a stream source ip address is 192.168.0.20 and gate action own two time slots. One is last 200ms gate open let frame pass another is last 100ms gate close let frames dropped. When the frames have passed total frames over 8000000 bytes, frames will be dropped in one 200000000ns time slot.
Quoted 3 lines
tc qdisc add dev eth0 ingress
tc filter add dev eth0 parent ffff: protocol ip \
flower src_ip 192.168.0.20 \ action gate index 2 clockid CLOCK_TAI \ sched-entry open 200000000 -1 8000000 \ sched-entry close 100000000 -1 -1
Quoted 1 line
tc chain del dev eth0 ingress chain 0
"sched-entry" follow the name taprio style. Gate state is "open"/"close". Follow with period nanosecond. Then next item is internal priority value means which ingress queue should put. "-1" means wildcard. The last value optional specifies the maximum number of MSDU octets that are permitted to pass the gate during the specified time interval.
Base-time is not set will be 0 as default, as result start time would be ((N + 1) * cycletime) which is the minimal of future time.
Below example shows filtering a stream with destination mac address is 10:00:80:00:00:00 and ip type is ICMP, follow the action gate. The gate action would run with one close time slot which means always keep close. The time cycle is total 200000000ns. The base-time would calculate by:
1357000000000 + (N + 1) * cycletime
When the total value is the future time, it will be the start time. The cycletime here would be 200000000ns for this case.
Quoted 1 line
tc filter add dev eth0 parent ffff: protocol ip \
flower skip_hw ip_proto icmp dst_mac 10:00:80:00:00:00 \ action gate index 12 base-time 1357000000000 \ sched-entry close 200000000 -1 -1 \ clockid CLOCK_TAI
--- include/net/tc_act/tc_gate.h | 54 +++ include/uapi/linux/pkt_cls.h | 1 + include/uapi/linux/tc_act/tc_gate.h | 47 ++ net/sched/Kconfig | 13 + net/sched/Makefile | 1 + net/sched/act_gate.c | 647 ++++++++++++++++++++++++++++ 6 files changed, 763 insertions(+) create mode 100644 include/net/tc_act/tc_gate.h create mode 100644 include/uapi/linux/tc_act/tc_gate.h create mode 100644 net/sched/act_gate.c
include/net/tc_act/tc_gate.h new file +54 -0
@@ -0,0 +1,54 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +/* Copyright 2020 NXP */ + +#ifndef __NET_TC_GATE_H +#define __NET_TC_GATE_H + +#include <net/act_api.h> +#include <linux/tc_act/tc_gate.h> + +struct tcfg_gate_entry { +int index; +u8 gate_state; +u32 interval; +s32 ipv; +s32 maxoctets; +struct list_head list; +}; + +struct tcf_gate_params { +s32 tcfg_priority; +u64 tcfg_basetime; +u64 tcfg_cycletime; +u64 tcfg_cycletime_ext; +u32 tcfg_flags; +s32 tcfg_clockid; +size_t num_entries; +struct list_head entries; +}; + +#define GATE_ACT_GATE_OPEN BIT(0) +#define GATE_ACT_PENDING BIT(1) +struct gate_action { +struct tcf_gate_params param; +spinlock_t entry_lock; +u8 current_gate_status; +ktime_t current_close_time; +u32 current_entry_octets; +s32 current_max_octets; +struct tcfg_gate_entry __rcu *next_entry; +struct hrtimer hitimer; +enum tk_offsets tk_offset; +struct rcu_head rcu; +}; + +struct tcf_gate { +struct tc_action common; +struct gate_action __rcu *actg; +}; +#define to_gate (a) ((struct tcf_gate *)a) + +#define get_gate_param (act) ((struct tcf_gate_params *)act) +#define get_gate_action (p) ((struct gate_action *)p) + +#endif
include/uapi/linux/pkt_cls.h +1 -0
@@ -134,6 +134,7 @@ enum tca_id {TCA_ID_CTINFO ,TCA_ID_MPLS ,TCA_ID_CT , +TCA_ID_GATE ,/* other actions go here */ __TCA_ID_MAX = 255 };
include/uapi/linux/tc_act/tc_gate.h new file +47 -0
47 lines not on this page · raw message
net/sched/Kconfig +13 -0
@@ -981,6 +981,19 @@ config NET_ACT_CT To compile this code as a module, choose M here: the module will be called act_ct. +config NET_ACT_GATE + tristate "Frame gate entry list control tc action" + depends on NET_CLS_ACT + help + Say Y here to allow to control the ingress flow to be passed at + specific time slot and be dropped at other specific time slot by + the gate entry list. The manipulation will simulate the IEEE + 802.1Qci stream gate control behavior. + + If unsure, say N. + To compile this code as a module, choose M here: the + module will be called act_gate. + config NET_IFE_SKBMARK tristate "Support to encoding decoding skb mark on IFE action" depends on NET_ACT_IFE
net/sched/Makefile +1 -0
@@ -30,6 +30,7 @@ obj-$(CONFIG_NET_IFE_SKBPRIO) += act_meta_skbprio.o obj-$(CONFIG_NET_IFE_SKBTCINDEX) += act_meta_skbtcindex.o obj-$(CONFIG_NET_ACT_TUNNEL_KEY)+= act_tunnel_key.o obj-$(CONFIG_NET_ACT_CT) += act_ct.o +obj-$(CONFIG_NET_ACT_GATE) += act_gate.o obj-$(CONFIG_NET_SCH_FIFO) += sch_fifo.o obj-$(CONFIG_NET_SCH_CBQ) += sch_cbq.o obj-$(CONFIG_NET_SCH_HTB) += sch_htb.o
net/sched/act_gate.c new file +647 -0
647 lines not on this page · raw message
Signature · 1 line
2.17.1
PATCH v2 2/4 net: schedule: add action gate offloading
In reply to Po LiuAdd the gate action to the flow action entry. Add the gate parameters to the tc_setup_flow_action() queueing to the entries of flow_action_entry array provide to the driver.
--- include/net/flow_offload.h | 10 +++ include/net/tc_act/tc_gate.h | 115 +++++++++++++++++++++++++++++++++++ net/sched/cls_api.c | 33 ++++++++++ 3 files changed, 158 insertions(+)
include/net/flow_offload.h +10 -0
@@ -147,6 +147,7 @@ enum flow_action_id {FLOW_ACTION_MPLS_PUSH ,FLOW_ACTION_MPLS_POP ,FLOW_ACTION_MPLS_MANGLE , +FLOW_ACTION_GATE , NUM_FLOW_ACTIONS, };
@@ -255,6 +256,15 @@ struct flow_action_entry {u8 bos;u8 ttl; } mpls_mangle; +struct { +u32 index; +s32 prio; +u64 basetime; +u64 cycletime; +u64 cycletimeext; +u32 num_entries; +struct action_gate_entry *entries; + } gate; };struct flow_action_cookie *cookie;/* user defined action cookie */ };
include/net/tc_act/tc_gate.h +115 -0
125 lines not on this page · raw message
net/sched/cls_api.c +33 -0
@@ -39,6 +39,7 @@#include <net/tc_act/tc_skbedit.h> #include <net/tc_act/tc_ct.h> #include <net/tc_act/tc_mpls.h> +#include <net/tc_act/tc_gate.h> #include <net/flow_offload.h> extern const struct nla_policy rtm_tca_policy[TCA_MAX +1 ];
@@ -3522,6 +3523,27 @@ static void tcf_sample_get_group(struct flow_action_entry *entry,#endif } +static void tcf_gate_entry_destructor (void *priv) +{ +struct action_gate_entry *oe = priv; + +kfree (oe); +} + +static int tcf_gate_get_entries (struct flow_action_entry *entry, +const struct tc_action *act) +{ + entry->gate.entries =tcf_gate_get_list (act); + +if (!entry->gate.entries) +return -EINVAL ; + + entry->destructor = tcf_gate_entry_destructor; + entry->destructor_priv = entry->gate.entries; + +return 0 ; +} +int tc_setup_flow_action (struct flow_action *flow_action,const struct tcf_exts *exts) {
@@ -3668,6 +3690,17 @@ int tc_setup_flow_action(struct flow_action *flow_action, }else if (is_tcf_skbedit_priority (act)) { entry->id =FLOW_ACTION_PRIORITY ; entry->priority =tcf_skbedit_priority (act); + }else if (is_tcf_gate (act)) { + entry->id =FLOW_ACTION_GATE ; + entry->gate.index =tcf_gate_index (act); + entry->gate.prio =tcf_gate_prio (act); + entry->gate.basetime =tcf_gate_basetime (act); + entry->gate.cycletime =tcf_gate_cycletime (act); + entry->gate.cycletimeext =tcf_gate_cycletimeext (act); + entry->gate.num_entries =tcf_gate_num_entries (act); + err =tcf_gate_get_entries (entry, act); +if (err) + goto err_out; }else { err = -EOPNOTSUPP ; goto err_out_locked;
Signature · 1 line
2.17.1
PATCH v2 3/4 net: enetc: add hw tc hw offload features for PSPF capability
In reply to Po LiuThis patch is to let ethtool enable/disable the tc flower offload features. Hardware ENETC has the feature of PSFP which is for per-stream policing. When enable the tc hw offloading feature, driver would enable the IEEE 802.1Qci feature. It is only set the register enable bit for this feature not enable for any entry of per stream filtering and stream gate or stream identify but get how much capabilities for each feature.
--- drivers/net/ethernet/freescale/enetc/enetc.c | 23 +++++++++ drivers/net/ethernet/freescale/enetc/enetc.h | 48 +++++++++++++++++++ .../net/ethernet/freescale/enetc/enetc_hw.h | 17 +++++++ .../net/ethernet/freescale/enetc/enetc_pf.c | 8 ++++ 4 files changed, 96 insertions(+)
drivers/net/ethernet/freescale/enetc/enetc.c +23 -0
@@ -756,6 +756,9 @@ void enetc_get_si_caps(struct enetc_si *si)if (val &ENETC_SIPCAPR0_QBV ) si->hw_features |=ENETC_SI_F_QBV ; + +if (val &ENETC_SIPCAPR0_PSFP ) + si->hw_features |=ENETC_SI_F_PSFP ; }static int enetc_dma_alloc_bdr (struct enetc_bdr *r,size_t bd_size)
@@ -1567,6 +1570,23 @@ static int enetc_set_rss(struct net_device *ndev, int en)return 0 ; } +static int enetc_set_psfp (struct net_device *ndev,int en) +{ +struct enetc_ndev_priv *priv =netdev_priv (ndev); + +if (en) { + priv->active_offloads |=ENETC_F_QCI ; +enetc_get_max_cap (priv); +enetc_psfp_enable (&priv->si->hw); + }else { + priv->active_offloads &= ~ENETC_F_QCI ; +memset (&priv->psfp_cap,0 ,sizeof (struct psfp_cap )); +enetc_psfp_disable (&priv->si->hw); + } + +return 0 ; +} +int enetc_set_features (struct net_device *ndev,netdev_features_t features) {
@@ -1575,6 +1595,9 @@ int enetc_set_features(struct net_device *ndev,if (changed &NETIF_F_RXHASH )enetc_set_rss (ndev, !!(features &NETIF_F_RXHASH )); +if (changed &NETIF_F_HW_TC ) +enetc_set_psfp (ndev, !!(features &NETIF_F_HW_TC )); +return 0 ; }
drivers/net/ethernet/freescale/enetc/enetc.h +48 -0
81 lines not on this page · raw message
drivers/net/ethernet/freescale/enetc/enetc_hw.h +17 -0
@@ -19,6 +19,7 @@#define ENETC_SICTR1 0x1c#define ENETC_SIPCAPR0 0x20#define ENETC_SIPCAPR0_QBV BIT(4) +#define ENETC_SIPCAPR0_PSFP BIT(9)#define ENETC_SIPCAPR0_RSS BIT(8)#define ENETC_SIPCAPR1 0x24#define ENETC_SITGTGR 0x30
@@ -228,6 +229,15 @@ enum enetc_bdr_type {TX, RX};#define ENETC_PM0_IFM_RLP (BIT(5) | BIT(11))#define ENETC_PM0_IFM_RGAUTO (BIT(15) | ENETC_PMO_IFM_RG | BIT(1))#define ENETC_PM0_IFM_XGMII BIT(12) +#define ENETC_PSIDCAPR 0x1b08 +#define ENETC_PSIDCAPR_MSK GENMASK(15, 0) +#define ENETC_PSFCAPR 0x1b18 +#define ENETC_PSFCAPR_MSK GENMASK(15, 0) +#define ENETC_PSGCAPR 0x1b28 +#define ENETC_PSGCAPR_GCL_MSK GENMASK(18, 16) +#define ENETC_PSGCAPR_SGIT_MSK GENMASK(15, 0) +#define ENETC_PFMCAPR 0x1b38 +#define ENETC_PFMCAPR_MSK GENMASK(15, 0)/* MAC counters */ #define ENETC_PM0_REOCT 0x8100
@@ -621,3 +631,10 @@ struct enetc_cbd {/* Port time specific departure */ #define ENETC_PTCTSDR (n) (0x1210 + 4 * (n))#define ENETC_TSDE BIT(31) + +/* PSFP setting */ +#define ENETC_PPSFPMR 0x11b00 +#define ENETC_PPSFPMR_PSFPEN BIT(0) +#define ENETC_PPSFPMR_VS BIT(1) +#define ENETC_PPSFPMR_PVC BIT(2) +#define ENETC_PPSFPMR_PVZC BIT(3)
drivers/net/ethernet/freescale/enetc/enetc_pf.c +8 -0
@@ -739,6 +739,14 @@ static void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,if (si->hw_features &ENETC_SI_F_QBV ) priv->active_offloads |=ENETC_F_QBV ; +if (si->hw_features &ENETC_SI_F_PSFP ) { + priv->active_offloads |=ENETC_F_QCI ; + ndev->features |=NETIF_F_HW_TC ; + ndev->hw_features |=NETIF_F_HW_TC ; +enetc_get_max_cap (priv); +enetc_psfp_enable (&si->hw); + } +/* pick up primary MAC address from SI */ enetc_get_primary_mac_addr (&si->hw, ndev->dev_addr); }
Signature · 1 line
2.17.1
PATCH v2 4/4 net: enetc: add tc flower psfp offload driver
In reply to Po LiuThis patch is to add tc flower offload for the enetc IEEE 802.1Qci(PSFP) function. There are four main feature parts to implement the flow policing and filtering for ingress flow with IEEE 802.1Qci features. They are stream identify(this is defined in the P802.1cb exactly but needed for 802.1Qci), stream filtering, stream gate and flow metering. Each function block includes many entries by index to assign parameters. So for one frame would be filtered by stream identify first, then flow into stream filter block by the same handle between stream identify and stream filtering. Then flow into stream gate control which assigned by the stream filtering entry. And then policing by the gate and limited by the max sdu in the filter block(optional). At last, policing by the flow metering block, index choosing at the fitering block. So you can see that each entry of block may link to many upper entries since they can be assigned same index means more streams want to share the same feature in the stream filtering or stream gate or flow metering.
To implement such features, each stream filtered by source/destination mac address, some stream maybe also plus the vlan id value would be treated as one flow chain. This would be identified by the chain_index which already in the tc filter concept. Driver would maintain this chain and also with gate modules. The stream filter entry create by the gate index and flow meter(optional) entry id and also one priority value. Offloading only transfer the gate action and flow filtering parameters. Driver would create (or search same gate id and flow meter id and
priority) one stream filter entry to set to the hardware. So stream
filtering do not need transfer by the action offloading. This architecture is same with tc filter and actions relationship. tc filter maintain the list for each flow feature by keys. And actions maintain by the action list.
Below showing a example commands by tc:
Quoted 3 lines
tc qdisc add dev eth0 ingress
ip link set eth0 address 10:00:80:00:00:00
tc filter add dev eth0 parent ffff: protocol ip chain 11 \
flower skip_sw dst_mac 10:00:80:00:00:00 \ action gate index 10 \ sched-entry OPEN 200000000 1 8000000 \ sched-entry CLOSE 100000000 -1 -1
Command means to set the dst_mac 10:00:80:00:00:00 to index 11 of stream identify module. And the set the gate index 10 of stream gate module. The gate list is keeping OPEN state 200ms, and through the frames to the ingress queue 1, and max octets are the 8Mbytes.
--- drivers/net/ethernet/freescale/enetc/enetc.c | 25 +- drivers/net/ethernet/freescale/enetc/enetc.h | 46 +- .../net/ethernet/freescale/enetc/enetc_hw.h | 142 +++ .../net/ethernet/freescale/enetc/enetc_pf.c | 4 +- .../net/ethernet/freescale/enetc/enetc_qos.c | 1070 +++++++++++++++++ 5 files changed, 1272 insertions(+), 15 deletions(-)
drivers/net/ethernet/freescale/enetc/enetc.c +17 -8
@@ -1521,6 +1521,8 @@ int enetc_setup_tc(struct net_device *ndev, enum tc_setup_type type,return enetc_setup_tc_cbs (ndev, type_data);case TC_SETUP_QDISC_ETF :return enetc_setup_tc_txtime (ndev, type_data); +case TC_SETUP_BLOCK : +return enetc_setup_tc_psfp (ndev, type_data);default :return -EOPNOTSUPP ; }
@@ -1573,17 +1575,23 @@ static int enetc_set_rss(struct net_device *ndev, int en)static int enetc_set_psfp (struct net_device *ndev,int en) {struct enetc_ndev_priv *priv =netdev_priv (ndev); +int err;if (en) { + err =enetc_psfp_enable (priv); +if (err) +return err; + priv->active_offloads |=ENETC_F_QCI ; -enetc_get_max_cap (priv); -enetc_psfp_enable (&priv->si->hw); - }else { - priv->active_offloads &= ~ENETC_F_QCI ; -memset (&priv->psfp_cap,0 ,sizeof (struct psfp_cap )); -enetc_psfp_disable (&priv->si->hw); +return 0 ; } + err =enetc_psfp_disable (priv); +if (err) +return err; + + priv->active_offloads &= ~ENETC_F_QCI ; +return 0 ; }
@@ -1591,14 +1599,15 @@ int enetc_set_features(struct net_device *ndev, netdev_features_t features) {netdev_features_t changed = ndev->features ^ features; +int err =0 ;if (changed &NETIF_F_RXHASH )enetc_set_rss (ndev, !!(features &NETIF_F_RXHASH ));if (changed &NETIF_F_HW_TC ) -enetc_set_psfp (ndev, !!(features &NETIF_F_HW_TC )); + err =enetc_set_psfp (ndev, !!(features &NETIF_F_HW_TC )); -return 0 ; +return err; }#ifdef CONFIG_FSL_ENETC_PTP_CLOCK
drivers/net/ethernet/freescale/enetc/enetc.h +42 -4
75 lines not on this page · raw message
drivers/net/ethernet/freescale/enetc/enetc_hw.h +142 -0
161 lines not on this page · raw message
drivers/net/ethernet/freescale/enetc/enetc_pf.c +1 -3
@@ -739,12 +739,10 @@ static void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,if (si->hw_features &ENETC_SI_F_QBV ) priv->active_offloads |=ENETC_F_QBV ; -if (si->hw_features &ENETC_SI_F_PSFP ) { +if (si->hw_features &ENETC_SI_F_PSFP && !enetc_psfp_enable (priv)) { priv->active_offloads |=ENETC_F_QCI ; ndev->features |=NETIF_F_HW_TC ; ndev->hw_features |=NETIF_F_HW_TC ; -enetc_get_max_cap (priv); -enetc_psfp_enable (&si->hw); }/* pick up primary MAC address from SI */
drivers/net/ethernet/freescale/enetc/enetc_qos.c +1070 -0
1,079 lines not on this page · raw message
Signature · 1 line
2.17.1
David Miller
maintainernet-next is still closed, please resubmit when it opens back up.
Thank you.
v2 cover Introduce a flow gate control action and apply IEEE
In reply to Po LiuChanges from V1:
0000: Update description make it more clear
0001: Removed 'add update dropped stats' patch, will provide pull request as standalone patches.
0001: Update commit description make it more clear ack by Jiri Pirko. 0001: No changes
0003: Fix some code style ack by Jiri Pirko.
0004: Fix enetc_psfp_enable/disable parameter type ack by test robot
iprout2 command patches:
Not attach with these serial patches, will provide separate pull
request after kernel accept these patches.
Changes from RFC:
0000: Reduce to 5 patches and remove the 4 max frame size offload and flow metering in the policing offload action, Only keep gate action offloading implementation.
0001: No changes.
0002:
- fix kfree lead ack by Jakub Kicinski and Cong Wang
- License fix from Jakub Kicinski and Stephen Hemminger
- Update example in commit acked by Vinicius Costa Gomes
- Fix the rcu protect in tcf_gate_act() acked by Vinicius
0003: No changes
0004: No changes
0005:
Acked by Vinicius Costa Gomes
- Use refcount kernel lib
- Update stream gate check code position
- Update reduce ref names more clear
iprout2 command patches:
0000: Update license expression and add gate id
0001: Add tc action gate man page
--------------------------------------------------------------------
These patches add stream gate action policing in IEEE802.1Qci (Per-Stream Filtering and Policing) software support and hardware offload support in tc flower, and implement the stream identify, stream filtering and stream gate filtering action in the NXP ENETC ethernet driver. Per-Stream Filtering and Policing (PSFP) specifies flow policing and filtering for ingress flows, and has three main parts:
1. The stream filter instance table consists of an ordered list of
stream filters that determine the filtering and policing actions that are to be applied to frames received on a specific stream. The main elements are stream gate id, flow metering id and maximum SDU size.
2. The stream gate function setup a gate list to control ingress traffic
class open/close state. When the gate is running at open state, the flow could pass but dropped when gate state is running to close. User setup a bastime to tell gate when start running the entry list, then the hardware would periodiclly. There is no compare qdisc action support.
3. Flow metering is two rates two buckets and three-color marker to
policing the frames. Flow metering instance are as specified in the algorithm in MEF10.3. The most likely qdisc action is policing action.
The first patch introduces an ingress frame flow control gate action, for the point 2. The tc gate action maintains the open/close state gate list, allowing flows to pass when the gate is open. Each gate action may policing one or more qdisc filters. When the start time arrived, The driver would repeat the gate list periodiclly. User can assign a passed time, the driver would calculate a new future time by the cycletime of the gate list.
The 0002 patch introduces the gate flow hardware offloading.
The 0003 patch adds support control the on/off for the tc flower offloading by ethtool.
The 0004 patch implement the stream identify and stream filtering and stream gate filtering action in the NXP ENETC ethernet driver. Tc filter command provide filtering keys with MAC address and VLAN id. These keys would be set to stream identify instance entry. Stream gate instance entry would refer the gate action parameters. Stream filter instance entry would refer the stream gate index and assign a stream handle value matches to the stream identify instance.
Po Liu (4):
net: qos: introduce a gate control flow action
net: schedule: add action gate offloading
net: enetc: add hw tc hw offload features for PSPF capability
net: enetc: add tc flower psfp offload driver
drivers/net/ethernet/freescale/enetc/enetc.c | 34 +- drivers/net/ethernet/freescale/enetc/enetc.h | 86 ++ .../net/ethernet/freescale/enetc/enetc_hw.h | 159 +++ .../net/ethernet/freescale/enetc/enetc_pf.c | 6 + .../net/ethernet/freescale/enetc/enetc_qos.c | 1070 +++++++++++++++++ include/net/flow_offload.h | 10 + include/net/tc_act/tc_gate.h | 169 +++ include/uapi/linux/pkt_cls.h | 1 + include/uapi/linux/tc_act/tc_gate.h | 47 + net/sched/Kconfig | 13 + net/sched/Makefile | 1 + net/sched/act_gate.c | 647 ++++++++++ net/sched/cls_api.c | 33 + 13 files changed, 2275 insertions(+), 1 deletion(-) create mode 100644 include/net/tc_act/tc_gate.h create mode 100644 include/uapi/linux/tc_act/tc_gate.h create mode 100644 net/sched/act_gate.c
Signature · 1 line
2.17.1
PATCH v2 1/4 net: qos: introduce a gate control flow action
Introduce a ingress frame gate control flow action.
Tc gate action does the work like this:
Assume there is a gate allow specified ingress frames can be passed at specific time slot, and be dropped at specific time slot. Tc filter chooses the ingress frames, and tc gate action would specify what slot does these frames can be passed to device and what time slot would be dropped.
Tc gate action would provide an entry list to tell how much time gate keep open and how much time gate keep state close. Gate action also assign a start time to tell when the entry list start. Then driver would repeat the gate entry list cyclically.
For the software simulation, gate action requires the user assign a time clock type.
Below is the setting example in user space. Tc filter a stream source ip address is 192.168.0.20 and gate action own two time slots. One is last 200ms gate open let frame pass another is last 100ms gate close let frames dropped. When the frames have passed total frames over 8000000 bytes, frames will be dropped in one 200000000ns time slot.
Quoted 3 lines
tc qdisc add dev eth0 ingress
tc filter add dev eth0 parent ffff: protocol ip \
flower src_ip 192.168.0.20 \ action gate index 2 clockid CLOCK_TAI \ sched-entry open 200000000 -1 8000000 \ sched-entry close 100000000 -1 -1
Quoted 1 line
tc chain del dev eth0 ingress chain 0
"sched-entry" follow the name taprio style. Gate state is "open"/"close". Follow with period nanosecond. Then next item is internal priority value means which ingress queue should put. "-1" means wildcard. The last value optional specifies the maximum number of MSDU octets that are permitted to pass the gate during the specified time interval.
Base-time is not set will be 0 as default, as result start time would be ((N + 1) * cycletime) which is the minimal of future time.
Below example shows filtering a stream with destination mac address is 10:00:80:00:00:00 and ip type is ICMP, follow the action gate. The gate action would run with one close time slot which means always keep close. The time cycle is total 200000000ns. The base-time would calculate by:
1357000000000 + (N + 1) * cycletime
When the total value is the future time, it will be the start time. The cycletime here would be 200000000ns for this case.
Quoted 1 line
tc filter add dev eth0 parent ffff: protocol ip \
flower skip_hw ip_proto icmp dst_mac 10:00:80:00:00:00 \ action gate index 12 base-time 1357000000000 \ sched-entry close 200000000 -1 -1 \ clockid CLOCK_TAI
--- include/net/tc_act/tc_gate.h | 54 +++ include/uapi/linux/pkt_cls.h | 1 + include/uapi/linux/tc_act/tc_gate.h | 47 ++ net/sched/Kconfig | 13 + net/sched/Makefile | 1 + net/sched/act_gate.c | 647 ++++++++++++++++++++++++++++ 6 files changed, 763 insertions(+) create mode 100644 include/net/tc_act/tc_gate.h create mode 100644 include/uapi/linux/tc_act/tc_gate.h create mode 100644 net/sched/act_gate.c
include/net/tc_act/tc_gate.h new file +54 -0
@@ -0,0 +1,54 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +/* Copyright 2020 NXP */ + +#ifndef __NET_TC_GATE_H +#define __NET_TC_GATE_H + +#include <net/act_api.h> +#include <linux/tc_act/tc_gate.h> + +struct tcfg_gate_entry { +int index; +u8 gate_state; +u32 interval; +s32 ipv; +s32 maxoctets; +struct list_head list; +}; + +struct tcf_gate_params { +s32 tcfg_priority; +u64 tcfg_basetime; +u64 tcfg_cycletime; +u64 tcfg_cycletime_ext; +u32 tcfg_flags; +s32 tcfg_clockid; +size_t num_entries; +struct list_head entries; +}; + +#define GATE_ACT_GATE_OPEN BIT(0) +#define GATE_ACT_PENDING BIT(1) +struct gate_action { +struct tcf_gate_params param; +spinlock_t entry_lock; +u8 current_gate_status; +ktime_t current_close_time; +u32 current_entry_octets; +s32 current_max_octets; +struct tcfg_gate_entry __rcu *next_entry; +struct hrtimer hitimer; +enum tk_offsets tk_offset; +struct rcu_head rcu; +}; + +struct tcf_gate { +struct tc_action common; +struct gate_action __rcu *actg; +}; +#define to_gate (a) ((struct tcf_gate *)a) + +#define get_gate_param (act) ((struct tcf_gate_params *)act) +#define get_gate_action (p) ((struct gate_action *)p) + +#endif
include/uapi/linux/pkt_cls.h +1 -0
@@ -134,6 +134,7 @@ enum tca_id {TCA_ID_CTINFO ,TCA_ID_MPLS ,TCA_ID_CT , +TCA_ID_GATE ,/* other actions go here */ __TCA_ID_MAX = 255 };
include/uapi/linux/tc_act/tc_gate.h new file +47 -0
47 lines not on this page · raw message
net/sched/Kconfig +13 -0
@@ -981,6 +981,19 @@ config NET_ACT_CT To compile this code as a module, choose M here: the module will be called act_ct. +config NET_ACT_GATE + tristate "Frame gate entry list control tc action" + depends on NET_CLS_ACT + help + Say Y here to allow to control the ingress flow to be passed at + specific time slot and be dropped at other specific time slot by + the gate entry list. The manipulation will simulate the IEEE + 802.1Qci stream gate control behavior. + + If unsure, say N. + To compile this code as a module, choose M here: the + module will be called act_gate. + config NET_IFE_SKBMARK tristate "Support to encoding decoding skb mark on IFE action" depends on NET_ACT_IFE
net/sched/Makefile +1 -0
@@ -30,6 +30,7 @@ obj-$(CONFIG_NET_IFE_SKBPRIO) += act_meta_skbprio.o obj-$(CONFIG_NET_IFE_SKBTCINDEX) += act_meta_skbtcindex.o obj-$(CONFIG_NET_ACT_TUNNEL_KEY)+= act_tunnel_key.o obj-$(CONFIG_NET_ACT_CT) += act_ct.o +obj-$(CONFIG_NET_ACT_GATE) += act_gate.o obj-$(CONFIG_NET_SCH_FIFO) += sch_fifo.o obj-$(CONFIG_NET_SCH_CBQ) += sch_cbq.o obj-$(CONFIG_NET_SCH_HTB) += sch_htb.o
net/sched/act_gate.c new file +647 -0
647 lines not on this page · raw message
Signature · 1 line
2.17.1
PATCH v2 2/4 net: schedule: add action gate offloading
In reply to Po LiuAdd the gate action to the flow action entry. Add the gate parameters to the tc_setup_flow_action() queueing to the entries of flow_action_entry array provide to the driver.
--- include/net/flow_offload.h | 10 +++ include/net/tc_act/tc_gate.h | 115 +++++++++++++++++++++++++++++++++++ net/sched/cls_api.c | 33 ++++++++++ 3 files changed, 158 insertions(+)
include/net/flow_offload.h +10 -0
@@ -147,6 +147,7 @@ enum flow_action_id {FLOW_ACTION_MPLS_PUSH ,FLOW_ACTION_MPLS_POP ,FLOW_ACTION_MPLS_MANGLE , +FLOW_ACTION_GATE , NUM_FLOW_ACTIONS, };
@@ -255,6 +256,15 @@ struct flow_action_entry {u8 bos;u8 ttl; } mpls_mangle; +struct { +u32 index; +s32 prio; +u64 basetime; +u64 cycletime; +u64 cycletimeext; +u32 num_entries; +struct action_gate_entry *entries; + } gate; };struct flow_action_cookie *cookie;/* user defined action cookie */ };
include/net/tc_act/tc_gate.h +115 -0
125 lines not on this page · raw message
net/sched/cls_api.c +33 -0
@@ -39,6 +39,7 @@#include <net/tc_act/tc_skbedit.h> #include <net/tc_act/tc_ct.h> #include <net/tc_act/tc_mpls.h> +#include <net/tc_act/tc_gate.h> #include <net/flow_offload.h> extern const struct nla_policy rtm_tca_policy[TCA_MAX +1 ];
@@ -3522,6 +3523,27 @@ static void tcf_sample_get_group(struct flow_action_entry *entry,#endif } +static void tcf_gate_entry_destructor (void *priv) +{ +struct action_gate_entry *oe = priv; + +kfree (oe); +} + +static int tcf_gate_get_entries (struct flow_action_entry *entry, +const struct tc_action *act) +{ + entry->gate.entries =tcf_gate_get_list (act); + +if (!entry->gate.entries) +return -EINVAL ; + + entry->destructor = tcf_gate_entry_destructor; + entry->destructor_priv = entry->gate.entries; + +return 0 ; +} +int tc_setup_flow_action (struct flow_action *flow_action,const struct tcf_exts *exts) {
@@ -3668,6 +3690,17 @@ int tc_setup_flow_action(struct flow_action *flow_action, }else if (is_tcf_skbedit_priority (act)) { entry->id =FLOW_ACTION_PRIORITY ; entry->priority =tcf_skbedit_priority (act); + }else if (is_tcf_gate (act)) { + entry->id =FLOW_ACTION_GATE ; + entry->gate.index =tcf_gate_index (act); + entry->gate.prio =tcf_gate_prio (act); + entry->gate.basetime =tcf_gate_basetime (act); + entry->gate.cycletime =tcf_gate_cycletime (act); + entry->gate.cycletimeext =tcf_gate_cycletimeext (act); + entry->gate.num_entries =tcf_gate_num_entries (act); + err =tcf_gate_get_entries (entry, act); +if (err) + goto err_out; }else { err = -EOPNOTSUPP ; goto err_out_locked;
Signature · 1 line
2.17.1
PATCH v2 3/4 net: enetc: add hw tc hw offload features for PSPF capability
In reply to Po LiuThis patch is to let ethtool enable/disable the tc flower offload features. Hardware ENETC has the feature of PSFP which is for per-stream policing. When enable the tc hw offloading feature, driver would enable the IEEE 802.1Qci feature. It is only set the register enable bit for this feature not enable for any entry of per stream filtering and stream gate or stream identify but get how much capabilities for each feature.
--- drivers/net/ethernet/freescale/enetc/enetc.c | 23 +++++++++ drivers/net/ethernet/freescale/enetc/enetc.h | 48 +++++++++++++++++++ .../net/ethernet/freescale/enetc/enetc_hw.h | 17 +++++++ .../net/ethernet/freescale/enetc/enetc_pf.c | 8 ++++ 4 files changed, 96 insertions(+)
drivers/net/ethernet/freescale/enetc/enetc.c +23 -0
@@ -756,6 +756,9 @@ void enetc_get_si_caps(struct enetc_si *si)if (val &ENETC_SIPCAPR0_QBV ) si->hw_features |=ENETC_SI_F_QBV ; + +if (val &ENETC_SIPCAPR0_PSFP ) + si->hw_features |=ENETC_SI_F_PSFP ; }static int enetc_dma_alloc_bdr (struct enetc_bdr *r,size_t bd_size)
@@ -1567,6 +1570,23 @@ static int enetc_set_rss(struct net_device *ndev, int en)return 0 ; } +static int enetc_set_psfp (struct net_device *ndev,int en) +{ +struct enetc_ndev_priv *priv =netdev_priv (ndev); + +if (en) { + priv->active_offloads |=ENETC_F_QCI ; +enetc_get_max_cap (priv); +enetc_psfp_enable (&priv->si->hw); + }else { + priv->active_offloads &= ~ENETC_F_QCI ; +memset (&priv->psfp_cap,0 ,sizeof (struct psfp_cap )); +enetc_psfp_disable (&priv->si->hw); + } + +return 0 ; +} +int enetc_set_features (struct net_device *ndev,netdev_features_t features) {
@@ -1575,6 +1595,9 @@ int enetc_set_features(struct net_device *ndev,if (changed &NETIF_F_RXHASH )enetc_set_rss (ndev, !!(features &NETIF_F_RXHASH )); +if (changed &NETIF_F_HW_TC ) +enetc_set_psfp (ndev, !!(features &NETIF_F_HW_TC )); +return 0 ; }
drivers/net/ethernet/freescale/enetc/enetc.h +48 -0
81 lines not on this page · raw message
drivers/net/ethernet/freescale/enetc/enetc_hw.h +17 -0
@@ -19,6 +19,7 @@#define ENETC_SICTR1 0x1c#define ENETC_SIPCAPR0 0x20#define ENETC_SIPCAPR0_QBV BIT(4) +#define ENETC_SIPCAPR0_PSFP BIT(9)#define ENETC_SIPCAPR0_RSS BIT(8)#define ENETC_SIPCAPR1 0x24#define ENETC_SITGTGR 0x30
@@ -228,6 +229,15 @@ enum enetc_bdr_type {TX, RX};#define ENETC_PM0_IFM_RLP (BIT(5) | BIT(11))#define ENETC_PM0_IFM_RGAUTO (BIT(15) | ENETC_PMO_IFM_RG | BIT(1))#define ENETC_PM0_IFM_XGMII BIT(12) +#define ENETC_PSIDCAPR 0x1b08 +#define ENETC_PSIDCAPR_MSK GENMASK(15, 0) +#define ENETC_PSFCAPR 0x1b18 +#define ENETC_PSFCAPR_MSK GENMASK(15, 0) +#define ENETC_PSGCAPR 0x1b28 +#define ENETC_PSGCAPR_GCL_MSK GENMASK(18, 16) +#define ENETC_PSGCAPR_SGIT_MSK GENMASK(15, 0) +#define ENETC_PFMCAPR 0x1b38 +#define ENETC_PFMCAPR_MSK GENMASK(15, 0)/* MAC counters */ #define ENETC_PM0_REOCT 0x8100
@@ -621,3 +631,10 @@ struct enetc_cbd {/* Port time specific departure */ #define ENETC_PTCTSDR (n) (0x1210 + 4 * (n))#define ENETC_TSDE BIT(31) + +/* PSFP setting */ +#define ENETC_PPSFPMR 0x11b00 +#define ENETC_PPSFPMR_PSFPEN BIT(0) +#define ENETC_PPSFPMR_VS BIT(1) +#define ENETC_PPSFPMR_PVC BIT(2) +#define ENETC_PPSFPMR_PVZC BIT(3)
drivers/net/ethernet/freescale/enetc/enetc_pf.c +8 -0
@@ -739,6 +739,14 @@ static void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,if (si->hw_features &ENETC_SI_F_QBV ) priv->active_offloads |=ENETC_F_QBV ; +if (si->hw_features &ENETC_SI_F_PSFP ) { + priv->active_offloads |=ENETC_F_QCI ; + ndev->features |=NETIF_F_HW_TC ; + ndev->hw_features |=NETIF_F_HW_TC ; +enetc_get_max_cap (priv); +enetc_psfp_enable (&si->hw); + } +/* pick up primary MAC address from SI */ enetc_get_primary_mac_addr (&si->hw, ndev->dev_addr); }
Signature · 1 line
2.17.1
PATCH v2 4/4 net: enetc: add tc flower psfp offload driver
In reply to Po LiuThis patch is to add tc flower offload for the enetc IEEE 802.1Qci(PSFP) function. There are four main feature parts to implement the flow policing and filtering for ingress flow with IEEE 802.1Qci features. They are stream identify(this is defined in the P802.1cb exactly but needed for 802.1Qci), stream filtering, stream gate and flow metering. Each function block includes many entries by index to assign parameters. So for one frame would be filtered by stream identify first, then flow into stream filter block by the same handle between stream identify and stream filtering. Then flow into stream gate control which assigned by the stream filtering entry. And then policing by the gate and limited by the max sdu in the filter block(optional). At last, policing by the flow metering block, index choosing at the fitering block. So you can see that each entry of block may link to many upper entries since they can be assigned same index means more streams want to share the same feature in the stream filtering or stream gate or flow metering.
To implement such features, each stream filtered by source/destination mac address, some stream maybe also plus the vlan id value would be treated as one flow chain. This would be identified by the chain_index which already in the tc filter concept. Driver would maintain this chain and also with gate modules. The stream filter entry create by the gate index and flow meter(optional) entry id and also one priority value. Offloading only transfer the gate action and flow filtering parameters. Driver would create (or search same gate id and flow meter id and
priority) one stream filter entry to set to the hardware. So stream
filtering do not need transfer by the action offloading. This architecture is same with tc filter and actions relationship. tc filter maintain the list for each flow feature by keys. And actions maintain by the action list.
Below showing a example commands by tc:
Quoted 3 lines
tc qdisc add dev eth0 ingress
ip link set eth0 address 10:00:80:00:00:00
tc filter add dev eth0 parent ffff: protocol ip chain 11 \
flower skip_sw dst_mac 10:00:80:00:00:00 \ action gate index 10 \ sched-entry OPEN 200000000 1 8000000 \ sched-entry CLOSE 100000000 -1 -1
Command means to set the dst_mac 10:00:80:00:00:00 to index 11 of stream identify module. And the set the gate index 10 of stream gate module. The gate list is keeping OPEN state 200ms, and through the frames to the ingress queue 1, and max octets are the 8Mbytes.
--- drivers/net/ethernet/freescale/enetc/enetc.c | 25 +- drivers/net/ethernet/freescale/enetc/enetc.h | 46 +- .../net/ethernet/freescale/enetc/enetc_hw.h | 142 +++ .../net/ethernet/freescale/enetc/enetc_pf.c | 4 +- .../net/ethernet/freescale/enetc/enetc_qos.c | 1070 +++++++++++++++++ 5 files changed, 1272 insertions(+), 15 deletions(-)
drivers/net/ethernet/freescale/enetc/enetc.c +17 -8
@@ -1521,6 +1521,8 @@ int enetc_setup_tc(struct net_device *ndev, enum tc_setup_type type,return enetc_setup_tc_cbs (ndev, type_data);case TC_SETUP_QDISC_ETF :return enetc_setup_tc_txtime (ndev, type_data); +case TC_SETUP_BLOCK : +return enetc_setup_tc_psfp (ndev, type_data);default :return -EOPNOTSUPP ; }
@@ -1573,17 +1575,23 @@ static int enetc_set_rss(struct net_device *ndev, int en)static int enetc_set_psfp (struct net_device *ndev,int en) {struct enetc_ndev_priv *priv =netdev_priv (ndev); +int err;if (en) { + err =enetc_psfp_enable (priv); +if (err) +return err; + priv->active_offloads |=ENETC_F_QCI ; -enetc_get_max_cap (priv); -enetc_psfp_enable (&priv->si->hw); - }else { - priv->active_offloads &= ~ENETC_F_QCI ; -memset (&priv->psfp_cap,0 ,sizeof (struct psfp_cap )); -enetc_psfp_disable (&priv->si->hw); +return 0 ; } + err =enetc_psfp_disable (priv); +if (err) +return err; + + priv->active_offloads &= ~ENETC_F_QCI ; +return 0 ; }
@@ -1591,14 +1599,15 @@ int enetc_set_features(struct net_device *ndev, netdev_features_t features) {netdev_features_t changed = ndev->features ^ features; +int err =0 ;if (changed &NETIF_F_RXHASH )enetc_set_rss (ndev, !!(features &NETIF_F_RXHASH ));if (changed &NETIF_F_HW_TC ) -enetc_set_psfp (ndev, !!(features &NETIF_F_HW_TC )); + err =enetc_set_psfp (ndev, !!(features &NETIF_F_HW_TC )); -return 0 ; +return err; }#ifdef CONFIG_FSL_ENETC_PTP_CLOCK
drivers/net/ethernet/freescale/enetc/enetc.h +42 -4
75 lines not on this page · raw message
drivers/net/ethernet/freescale/enetc/enetc_hw.h +142 -0
161 lines not on this page · raw message
drivers/net/ethernet/freescale/enetc/enetc_pf.c +1 -3
@@ -739,12 +739,10 @@ static void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,if (si->hw_features &ENETC_SI_F_QBV ) priv->active_offloads |=ENETC_F_QBV ; -if (si->hw_features &ENETC_SI_F_PSFP ) { +if (si->hw_features &ENETC_SI_F_PSFP && !enetc_psfp_enable (priv)) { priv->active_offloads |=ENETC_F_QCI ; ndev->features |=NETIF_F_HW_TC ; ndev->hw_features |=NETIF_F_HW_TC ; -enetc_get_max_cap (priv); -enetc_psfp_enable (&si->hw); }/* pick up primary MAC address from SI */
drivers/net/ethernet/freescale/enetc/enetc_qos.c +1070 -0
1,079 lines not on this page · raw message
Signature · 1 line
2.17.1
Vladimir Oltean
maintainerRe: [ v2,net-next 4/4] net: enetc: add tc flower psfp offload driver
Hi Po,
Quoted 1,057 lines · Po Liu
This patch is to add tc flower offload for the enetc IEEE 802.1Qci(PSFP) function. There are four main feature parts to implement the flow policing and filtering for ingress flow with IEEE 802.1Qci features. They are stream identify(this is defined in the P802.1cb exactly but needed for 802.1Qci), stream filtering, stream gate and flow metering. Each function block includes many entries by index to assign parameters. So for one frame would be filtered by stream identify first, then flow into stream filter block by the same handle between stream identify and stream filtering. Then flow into stream gate control which assigned by the stream filtering entry. And then policing by the gate and limited by the max sdu in the filter block(optional). At last, policing by the flow metering block, index choosing at the fitering block. So you can see that each entry of block may link to many upper entries since they can be assigned same index means more streams want to share the same feature in the stream filtering or stream gate or flow metering.
To implement such features, each stream filtered by source/destination mac address, some stream maybe also plus the vlan id value would be treated as one flow chain. This would be identified by thechain_indexwhich already in the tc filter concept. Driver would maintain this chain and also with gate modules. The stream filter entry create by the gate index and flow meter(optional) entry id and also one priority value. Offloading only transfer the gate action and flow filtering parameters. Driver would create (or search same gate id and flow meter id and
priority) one stream filter entry to set to the hardware. So stream
filtering do not need transfer by the action offloading. This architecture is same with tc filter and actions relationship. tc filter maintain the list for each flow feature by keys. And actions maintain by the action list.Below showing a example commands by tc:
Quoted 3 lines
tc qdisc add dev eth0 ingress
ip link set eth0 address 10:00:80:00:00:00
tc filter add dev eth0 parent ffff: protocol ip chain 11 \flower skip_sw dst_mac 10:00:80:00:00:00 \ action gate index 10 \ sched-entry OPEN 200000000 1 8000000 \ sched-entry CLOSE 100000000 -1 -1Command means to set the
dst_mac10:00:80:00:00:00 to index 11 of stream identify module. And the set the gate index 10 of stream gate module. The gate list is keeping OPEN state 200ms, and through the frames to the ingress queue 1, and max octets are the 8Mbytes.Signed-off-by Po Liu <[email protected]>--- drivers/net/ethernet/freescale/enetc/enetc.c | 25 +- drivers/net/ethernet/freescale/enetc/enetc.h | 46 +- .../net/ethernet/freescale/enetc/enetc_hw.h | 142 +++ .../net/ethernet/freescale/enetc/enetc_pf.c | 4 +- .../net/ethernet/freescale/enetc/enetc_qos.c | 1070 +++++++++++++++++ 5 files changed, 1272 insertions(+), 15 deletions(-)drivers/net/ethernet/freescale/enetc/enetc.c +17
-8@@ -1521,6 +1521,8 @@ int enetc_setup_tc(struct net_device *ndev, enum tc_setup_type type,return enetc_setup_tc_cbs (ndev, type_data);case TC_SETUP_QDISC_ETF :return enetc_setup_tc_txtime (ndev, type_data); +case TC_SETUP_BLOCK : +return enetc_setup_tc_psfp (ndev, type_data);default :return -EOPNOTSUPP ; }@@ -1573,17 +1575,23 @@ static int enetc_set_rss(struct net_device *ndev, int en)static int enetc_set_psfp (struct net_device *ndev,int en) {struct enetc_ndev_priv *priv =netdev_priv (ndev); +int err;if (en) { + err =enetc_psfp_enable (priv); +if (err) +return err; + priv->active_offloads |=ENETC_F_QCI ; -enetc_get_max_cap (priv); -enetc_psfp_enable (&priv->si->hw); - }else { - priv->active_offloads &= ~ENETC_F_QCI ; -memset (&priv->psfp_cap,0 ,sizeof (struct psfp_cap )); -enetc_psfp_disable (&priv->si->hw); +return 0 ; } + err =enetc_psfp_disable (priv); +if (err) +return err; + + priv->active_offloads &= ~ENETC_F_QCI ; +return 0 ; }@@ -1591,14 +1599,15 @@ int enetc_set_features(struct net_device *ndev, netdev_features_t features) {netdev_features_t changed = ndev->features ^ features; +int err =0 ;if (changed &NETIF_F_RXHASH )enetc_set_rss (ndev, !!(features &NETIF_F_RXHASH ));if (changed &NETIF_F_HW_TC ) -enetc_set_psfp (ndev, !!(features &NETIF_F_HW_TC )); + err =enetc_set_psfp (ndev, !!(features &NETIF_F_HW_TC )); -return 0 ; +return err; }#ifdef CONFIG_FSL_ENETC_PTP_CLOCK drivers/net/ethernet/freescale/enetc/enetc.h +42
-4@@ -300,6 +300,11 @@ int enetc_setup_tc_taprio(struct net_device *ndev, void *type_data);void enetc_sched_speed_set (struct net_device *ndev);int enetc_setup_tc_cbs (struct net_device *ndev,void *type_data);int enetc_setup_tc_txtime (struct net_device *ndev,void *type_data); +int enetc_setup_tc_block_cb (enum tc_setup_type type,void *type_data, +void *cb_priv); +int enetc_setup_tc_psfp (struct net_device *ndev,void *type_data); +int enetc_psfp_init (struct enetc_ndev_priv *priv); +int enetc_psfp_clean (struct enetc_ndev_priv *priv);static inline void enetc_get_max_cap (struct enetc_ndev_priv *priv) {@@ -319,27 +324,60 @@ static inline void enetc_get_max_cap(struct enetc_ndev_priv *priv) priv->psfp_cap.max_psfp_meter = reg &ENETC_PFMCAPR_MSK ; } -static inline void enetc_psfp_enable (struct enetc_hw *hw) +static inline int enetc_psfp_enable (struct enetc_ndev_priv *priv) { +struct enetc_hw *hw = &priv->si->hw; +int err; + +enetc_get_max_cap (priv); + + err =enetc_psfp_init (priv); +if (err) +return err; +enetc_wr (hw,ENETC_PPSFPMR ,enetc_rd (hw,ENETC_PPSFPMR ) |ENETC_PPSFPMR_PSFPEN |ENETC_PPSFPMR_VS |ENETC_PPSFPMR_PVC |ENETC_PPSFPMR_PVZC ); + +return 0 ; } -static inline void enetc_psfp_disable (struct enetc_hw *hw) +static inline int enetc_psfp_disable (struct enetc_ndev_priv *priv) { +struct enetc_hw *hw = &priv->si->hw; +int err; + + err =enetc_psfp_clean (priv); +if (err) +return err; +enetc_wr (hw,ENETC_PPSFPMR ,enetc_rd (hw,ENETC_PPSFPMR ) & ~ENETC_PPSFPMR_PSFPEN & ~ENETC_PPSFPMR_VS & ~ENETC_PPSFPMR_PVC & ~ENETC_PPSFPMR_PVZC ); + +memset (&priv->psfp_cap,0 ,sizeof (struct psfp_cap )); + +return 0 ; } +#else #define enetc_setup_tc_taprio (ndev, type_data) -EOPNOTSUPP#define enetc_sched_speed_set (ndev) (void)0#define enetc_setup_tc_cbs (ndev, type_data) -EOPNOTSUPP#define enetc_setup_tc_txtime (ndev, type_data) -EOPNOTSUPP +#define enetc_setup_tc_psfp (ndev, type_data) -EOPNOTSUPP +#define enetc_setup_tc_block_cb NULL +#define enetc_get_max_cap (p) \ memset(&((p)->psfp_cap), 0, sizeof(struct psfp_cap)) -#define enetc_psfp_enable (hw) (void)0 -#define enetc_psfp_disable (hw) (void)0 +static inline int enetc_psfp_enable (struct enetc_ndev_priv *priv) +{ +return 0 ; +} + +static inline int enetc_psfp_disable (struct enetc_ndev_priv *priv) +{ +return 0 ; +}#endif drivers/net/ethernet/freescale/enetc/enetc_hw.h +142
-0@@ -567,6 +567,9 @@ enum bdcr_cmd_class {BDCR_CMD_RFS ,BDCR_CMD_PORT_GCL ,BDCR_CMD_RECV_CLASSIFIER , +BDCR_CMD_STREAM_IDENTIFY , +BDCR_CMD_STREAM_FILTER , +BDCR_CMD_STREAM_GCL , __BDCR_CMD_MAX_LEN, BDCR_CMD_MAX_LEN = __BDCR_CMD_MAX_LEN - 1, };@@ -598,13 +601,152 @@ struct tgs_gcl_data {struct gce entry[]; }; +/* class 7, command 0, Stream Identity Entry Configuration */ +struct streamid_conf { +__le32 stream_handle;/* init gate value */ +__le32 iports; +u8 id_type; +u8 oui[3 ]; +u8 res[3 ]; +u8 en; +}; + +#define ENETC_CBDR_SID_VID_MASK 0xfff +#define ENETC_CBDR_SID_VIDM BIT(12) +#define ENETC_CBDR_SID_TG_MASK 0xc000 +/* streamid_conf address point to this data space */ +struct streamid_data { +union { +u8 dmac[6 ]; +u8 smac[6 ]; + }; +u16 vid_vidm_tg; +}; + +#define ENETC_CBDR_SFI_PRI_MASK 0x7 +#define ENETC_CBDR_SFI_PRIM BIT(3) +#define ENETC_CBDR_SFI_BLOV BIT(4) +#define ENETC_CBDR_SFI_BLEN BIT(5) +#define ENETC_CBDR_SFI_MSDUEN BIT(6) +#define ENETC_CBDR_SFI_FMITEN BIT(7) +#define ENETC_CBDR_SFI_ENABLE BIT(7) +/* class 8, command 0, Stream Filter Instance, Short Format */ +struct sfi_conf { +__le32 stream_handle; +u8 multi; +u8 res[2 ]; +u8 sthm; +/* Max Service Data Unit or Flow Meter Instance Table index. +* Depending on the value of FLT this represents either Max +* Service Data Unit (max frame size) allowed by the filter +* entry or is an index into the Flow Meter Instance table +* index identifying the policer which will be used to police +* it. +*/ +__le16 fm_inst_table_index; +__le16 msdu; +__le16 sg_inst_table_index; +u8 res1[2 ]; +__le32 input_ports; +u8 res2[3 ]; +u8 en; +}; + +/* class 8, command 2 stream Filter Instance status query short format +* command no need structure define +* Stream Filter Instance Query Statistics Response data +*/ +struct sfi_counter_data { +u32 matchl; +u32 matchh; +u32 msdu_dropl; +u32 msdu_droph; +u32 stream_gate_dropl; +u32 stream_gate_droph; +u32 flow_meter_dropl; +u32 flow_meter_droph; +}; + +#define ENETC_CBDR_SGI_OIPV_MASK 0x7 +#define ENETC_CBDR_SGI_OIPV_EN BIT(3) +#define ENETC_CBDR_SGI_CGTST BIT(6) +#define ENETC_CBDR_SGI_OGTST BIT(7) +#define ENETC_CBDR_SGI_CFG_CHG BIT(1) +#define ENETC_CBDR_SGI_CFG_PND BIT(2) +#define ENETC_CBDR_SGI_OEX BIT(4) +#define ENETC_CBDR_SGI_OEXEN BIT(5) +#define ENETC_CBDR_SGI_IRX BIT(6) +#define ENETC_CBDR_SGI_IRXEN BIT(7) +#define ENETC_CBDR_SGI_ACLLEN_MASK 0x3 +#define ENETC_CBDR_SGI_OCLLEN_MASK 0xc +#define ENETC_CBDR_SGI_EN BIT(7) +/* class 9, command 0, Stream Gate Instance Table, Short Format +* class 9, command 2, Stream Gate Instance Table entry query write back +* Short Format +*/ +struct sgi_table { +u8 res[8 ]; +u8 oipv; +u8 res0[2 ]; +u8 ocgtst; +u8 res1[7 ]; +u8 gset; +u8 oacl_len; +u8 res2[2 ]; +u8 en; +}; + +#define ENETC_CBDR_SGI_AIPV_MASK 0x7 +#define ENETC_CBDR_SGI_AIPV_EN BIT(3) +#define ENETC_CBDR_SGI_AGTST BIT(7) + +/* class 9, command 1, Stream Gate Control List, Long Format */ +struct sgcl_conf { +u8 aipv; +u8 res[2 ]; +u8 agtst; +u8 res1[4 ]; +union { +struct { +u8 res2[4 ]; +u8 acl_len; +u8 res3[3 ]; + }; +u8 cct[8 ];/* Config change time */ + }; +}; + +#define ENETC_CBDR_SGL_IOMEN BIT(0) +#define ENETC_CBDR_SGL_IPVEN BIT(3) +#define ENETC_CBDR_SGL_GTST BIT(4) +#define ENETC_CBDR_SGL_IPV_MASK 0xe +/* Stream Gate Control List Entry */ +struct sgce { +u32 interval; +u8 msdu[3 ]; +u8 multi; +}; + +/* stream control list class 9 , cmd 1 data buffer */ +struct sgcl_data { +u32 btl; +u32 bth; +u32 ct; +u32 cte; +struct sgce sgcl[0 ]; +}; +struct enetc_cbd {union { +struct sfi_conf sfi_conf; +struct sgi_table sgi_table;struct {__le32 addr[2 ];union {__le32 opt[4 ];struct tgs_gcl_conf gcl_conf; +struct streamid_conf sid_set; +struct sgcl_conf sgcl_conf; }; };/* Long format */ __le32 data[6 ];drivers/net/ethernet/freescale/enetc/enetc_pf.c +1
-3@@ -739,12 +739,10 @@ static void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,if (si->hw_features &ENETC_SI_F_QBV ) priv->active_offloads |=ENETC_F_QBV ; -if (si->hw_features &ENETC_SI_F_PSFP ) { +if (si->hw_features &ENETC_SI_F_PSFP && !enetc_psfp_enable (priv)) { priv->active_offloads |=ENETC_F_QCI ; ndev->features |=NETIF_F_HW_TC ; ndev->hw_features |=NETIF_F_HW_TC ; -enetc_get_max_cap (priv); -enetc_psfp_enable (&si->hw); }/* pick up primary MAC address from SI */ drivers/net/ethernet/freescale/enetc/enetc_qos.c +661
-0@@ -5,6 +5,9 @@#include <net/pkt_sched.h> #include <linux/math64.h> +#include <linux/refcount.h> +#include <net/pkt_cls.h> +#include <net/tc_act/tc_gate.h> static u16 enetc_get_max_gcl_len (struct enetc_hw *hw) {@@ -331,3 +334,1070 @@ int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data)return 0 ; } + +enum streamid_type { +STREAMID_TYPE_RESERVED =0 , +STREAMID_TYPE_NULL , +STREAMID_TYPE_SMAC , +}; + +enum streamid_vlan_tagged { +STREAMID_VLAN_RESERVED =0 , +STREAMID_VLAN_TAGGED , +STREAMID_VLAN_UNTAGGED , +STREAMID_VLAN_ALL , +}; + +#define ENETC_PSFP_WILDCARD -1 +#define HANDLE_OFFSET 100 + +enum forward_type { +FILTER_ACTION_TYPE_PSFP =BIT (0 ), +FILTER_ACTION_TYPE_ACL =BIT (1 ), +FILTER_ACTION_TYPE_BOTH =GENMASK (1 ,0 ), +}; + +/* This is for limit output type for input actions */ +struct actions_fwd { +u64 actions; +u64 keys;/* include the must needed keys */ +enum forward_type output; +}; + +struct psfp_streamfilter_counters { +u64 matching_frames_count; +u64 passing_frames_count; +u64 not_passing_frames_count; +u64 passing_sdu_count; +u64 not_passing_sdu_count; +u64 red_frames_count; +}; + +struct enetc_streamid { +u32 index; +union { +u8 src_mac[6 ]; +u8 dst_mac[6 ]; + }; +u8 filtertype; +u16 vid; +u8 tagged; +s32 handle; +}; + +struct enetc_psfp_filter { +u32 index; +s32 handle; +s8 prio; +u32 gate_id; +s32 meter_id; +refcount_t refcount; +struct hlist_node node; +}; + +struct enetc_psfp_gate { +u32 index; +s8 init_ipv; +u64 basetime; +u64 cycletime; +u64 cycletimext; +u32 num_entries; +refcount_t refcount; +struct hlist_node node; +struct action_gate_entry entries[0 ]; +}; + +struct enetc_stream_filter { +struct enetc_streamid sid; +u32 sfi_index; +u32 sgi_index; +struct flow_stats stats; +struct hlist_node node; +}; + +struct enetc_psfp { +unsigned long dev_bitmap; +unsigned long *psfp_sfi_bitmap; +struct hlist_head stream_list; +struct hlist_head psfp_filter_list; +struct hlist_head psfp_gate_list; +spinlock_t psfp_lock;/* spinlock for the struct enetc_psfp r/w */ +}; + +struct actions_fwd enetc_act_fwd[] = { + { +BIT (FLOW_ACTION_GATE ), +BIT (FLOW_DISSECTOR_KEY_ETH_ADDRS ), +FILTER_ACTION_TYPE_PSFP + }, +/* example for ACL actions */ + { +BIT (FLOW_ACTION_DROP ), +0 , +FILTER_ACTION_TYPE_ACL + } +}; + +static struct enetc_psfp epsfp = { + .psfp_sfi_bitmap =NULL , +}; + +static LIST_HEAD (enetc_block_cb_list); + +static inline int enetc_get_port (struct enetc_ndev_priv *priv) +{ +return priv->si->pdev->devfn &0x7 ; +} + +/* Stream Identity Entry Set Descriptor */ +static int enetc_streamid_hw_set (struct enetc_ndev_priv *priv, +struct enetc_streamid *sid, +u8 enable) +{ +struct enetc_cbd cbd = {.cmd =0 }; +struct streamid_data *si_data; +struct streamid_conf *si_conf; +u16 data_size; +dma_addr_t dma; +int err; + +if (sid->index >= priv->psfp_cap.max_streamid) +return -EINVAL ; + +if (sid->filtertype !=STREAMID_TYPE_NULL && + sid->filtertype !=STREAMID_TYPE_SMAC ) +return -EOPNOTSUPP ; + +/* Disable operation before enable */ + cbd.index =cpu_to_le16 ((u16 )sid->index); + cbd.cls =BDCR_CMD_STREAM_IDENTIFY ; + cbd.status_flags =0 ; + + data_size =sizeof (struct streamid_data ); + si_data =kzalloc (data_size, __GFP_DMA |GFP_KERNEL ); + cbd.length =cpu_to_le16 (data_size); + + dma =dma_map_single (&priv->si->pdev->dev, si_data, + data_size,DMA_FROM_DEVICE ); +if (dma_mapping_error (&priv->si->pdev->dev, dma)) { +netdev_err (priv->si->ndev,"DMA mapping failed!\n" ); +kfree (si_data); +return -ENOMEM ; + } + + cbd.addr[0 ] =lower_32_bits (dma); + cbd.addr[1 ] =upper_32_bits (dma); +memset (si_data->dmac,0xff ,ETH_ALEN ); + si_data->vid_vidm_tg = +cpu_to_le16 (ENETC_CBDR_SID_VID_MASK + + ((0x3 <<14 ) |ENETC_CBDR_SID_VIDM )); + + si_conf = &cbd.sid_set; +/* Only one port supported for one entry, set itself */ + si_conf->iports =1 <<enetc_get_port (priv); + si_conf->id_type =1 ; + si_conf->oui[2 ] =0x0 ; + si_conf->oui[1 ] =0x80 ; + si_conf->oui[0 ] =0xC2 ; + + err =enetc_send_cmd (priv->si, &cbd); +if (err) +return -EINVAL ; + +if (!enable) { +kfree (si_data); +return 0 ; + } + +/* Enable the entry overwrite again incase space flushed by hardware */ +memset (&cbd,0 ,sizeof (cbd)); + + cbd.index =cpu_to_le16 ((u16 )sid->index); + cbd.cmd =0 ; + cbd.cls =BDCR_CMD_STREAM_IDENTIFY ; + cbd.status_flags =0 ; + + si_conf->en =0x80 ; + si_conf->stream_handle =cpu_to_le32 (sid->handle); + si_conf->iports =1 <<enetc_get_port (priv); + si_conf->id_type = sid->filtertype; + si_conf->oui[2 ] =0x0 ; + si_conf->oui[1 ] =0x80 ; + si_conf->oui[0 ] =0xC2 ; + +memset (si_data,0 , data_size); + + cbd.length =cpu_to_le16 (data_size); + + cbd.addr[0 ] =lower_32_bits (dma); + cbd.addr[1 ] =upper_32_bits (dma); + +/* VIDM default to be 1. +* VID Match. If set (b1) then the VID must match, otherwise +* any VID is considered a match. VIDM setting is only used +* when TG is set to b01. +*/ +if (si_conf->id_type ==STREAMID_TYPE_NULL ) { +ether_addr_copy (si_data->dmac, sid->dst_mac); + si_data->vid_vidm_tg = +cpu_to_le16 ((sid->vid &ENETC_CBDR_SID_VID_MASK ) + + ((((u16)(sid->tagged) &0x3 ) <<14 ) + |ENETC_CBDR_SID_VIDM )); + }else if (si_conf->id_type ==STREAMID_TYPE_SMAC ) { +ether_addr_copy (si_data->smac, sid->src_mac); + si_data->vid_vidm_tg = +cpu_to_le16 ((sid->vid &ENETC_CBDR_SID_VID_MASK ) + + ((((u16)(sid->tagged) &0x3 ) <<14 ) + |ENETC_CBDR_SID_VIDM )); + } + + err =enetc_send_cmd (priv->si, &cbd); +kfree (si_data); + +return err; +} + +/* Stream Filter Instance Set Descriptor */ +static int enetc_streamfilter_hw_set (struct enetc_ndev_priv *priv, +struct enetc_psfp_filter *sfi, +u8 enable) +{ +struct enetc_cbd cbd = {.cmd =0 }; +struct sfi_conf *sfi_config; + + cbd.index =cpu_to_le16 (sfi->index); + cbd.cls =BDCR_CMD_STREAM_FILTER ; + cbd.status_flags =0x80 ; + cbd.length =cpu_to_le16 (1 ); + + sfi_config = &cbd.sfi_conf; +if (!enable) + goto exit; + + sfi_config->en =0x80 ; + +if (sfi->handle >=0 ) { + sfi_config->stream_handle = +cpu_to_le32 (sfi->handle); + sfi_config->sthm |=0x80 ; + } + + sfi_config->sg_inst_table_index =cpu_to_le16 (sfi->gate_id); + sfi_config->input_ports =1 <<enetc_get_port (priv); + +/* The priority value which may be matched against the +* frame’s priority value to determine a match for this entry. +*/ +if (sfi->prio >=0 ) + sfi_config->multi |= (sfi->prio &0x7 ) |0x8 ; + +/* Filter Type. Identifies the contents of the MSDU/FM_INST_INDEX +* field as being either an MSDU value or an index into the Flow +* Meter Instance table. +* TODO: no limit max sdu +*/ + +if (sfi->meter_id >=0 ) { + sfi_config->fm_inst_table_index =cpu_to_le16 (sfi->meter_id); + sfi_config->multi |=0x80 ; + } + +exit: +return enetc_send_cmd (priv->si, &cbd); +} + +static int enetc_streamcounter_hw_get (struct enetc_ndev_priv *priv, +u32 index, +struct psfp_streamfilter_counters *cnt) +{ +struct enetc_cbd cbd = { .cmd =2 }; +struct sfi_counter_data *data_buf; +dma_addr_t dma; +u16 data_size; +int err; + + cbd.index =cpu_to_le16 ((u16 )index); + cbd.cmd =2 ; + cbd.cls =BDCR_CMD_STREAM_FILTER ; + cbd.status_flags =0 ; + + data_size =sizeof (struct sfi_counter_data ); + data_buf =kzalloc (data_size, __GFP_DMA |GFP_KERNEL ); +if (!data_buf) +return -ENOMEM ; + + dma =dma_map_single (&priv->si->pdev->dev, data_buf, + data_size,DMA_FROM_DEVICE ); +if (dma_mapping_error (&priv->si->pdev->dev, dma)) { +netdev_err (priv->si->ndev,"DMA mapping failed!\n" ); + err = -ENOMEM ; + goto exit; + } + cbd.addr[0 ] =lower_32_bits (dma); + cbd.addr[1 ] =upper_32_bits (dma); + + cbd.length =cpu_to_le16 (data_size); + + err =enetc_send_cmd (priv->si, &cbd); +if (err) + goto exit; + + cnt->matching_frames_count = + ((u64 )le32_to_cpu (data_buf->matchh) <<32 ) + + data_buf->matchl; + + cnt->not_passing_sdu_count = + ((u64 )le32_to_cpu (data_buf->msdu_droph) <<32 ) + + data_buf->msdu_dropl; + + cnt->passing_sdu_count = cnt->matching_frames_count + - cnt->not_passing_sdu_count; + + cnt->not_passing_frames_count = + ((u64 )le32_to_cpu (data_buf->stream_gate_droph) <<32 ) + +le32_to_cpu (data_buf->stream_gate_dropl); + + cnt->passing_frames_count = cnt->matching_frames_count + - cnt->not_passing_sdu_count + - cnt->not_passing_frames_count; + + cnt->red_frames_count = + ((u64 )le32_to_cpu (data_buf->flow_meter_droph) <<32 ) + +le32_to_cpu (data_buf->flow_meter_dropl); + +exit: +kfree (data_buf); +return err; +} + +static int get_start_ns (struct enetc_ndev_priv *priv,u64 cycle,u64 *start) +{ +u64 now_lo, now_hi, now, n; + + now_lo =enetc_rd (&priv->si->hw,ENETC_SICTR0 ); + now_hi =enetc_rd (&priv->si->hw,ENETC_SICTR1 ); + now = now_lo | now_hi <<32 ; + +if (WARN_ON (!cycle)) +return -EFAULT ; + + n =div64_u64 (now, cycle); + + *start = (n +1 ) * cycle; + +return 0 ; +} + +/* Stream Gate Instance Set Descriptor */ +static int enetc_streamgate_hw_set (struct enetc_ndev_priv *priv, +struct enetc_psfp_gate *sgi, +u8 enable) +{ +struct enetc_cbd cbd = { .cmd =0 }; +struct sgi_table *sgi_config; +struct sgcl_conf *sgcl_config; +struct sgcl_data *sgcl_data; +struct sgce *sgce; +dma_addr_t dma; +u16 data_size; +int err, i; + + cbd.index =cpu_to_le16 (sgi->index); + cbd.cmd =0 ; + cbd.cls =BDCR_CMD_STREAM_GCL ; + cbd.status_flags =0x80 ; + +/* disable */ +if (!enable) +return enetc_send_cmd (priv->si, &cbd); + +if (!sgi->num_entries) +return 0 ; + +if (sgi->num_entries > priv->psfp_cap.max_psfp_gatelist || + !sgi->cycletime) +return -EINVAL ; + +/* enable */ + sgi_config = &cbd.sgi_table; + +/* Keep open before gate list start */ + sgi_config->ocgtst =0x80 ; + + sgi_config->oipv = (sgi->init_ipv <0 ) ? +0x0 : ((sgi->init_ipv &0x7 ) |0x8 ); + + sgi_config->en =0x80 ; + +/* Basic config */ + err =enetc_send_cmd (priv->si, &cbd); +if (err) +return -EINVAL ; + +memset (&cbd,0 ,sizeof (cbd)); + + cbd.index =cpu_to_le16 (sgi->index); + cbd.cmd =1 ; + cbd.cls =BDCR_CMD_STREAM_GCL ; + cbd.status_flags =0 ; + + sgcl_config = &cbd.sgcl_conf; + + sgcl_config->acl_len = (sgi->num_entries -1 ) &0x3 ; + + data_size =struct_size (sgcl_data, sgcl, sgi->num_entries); + + sgcl_data =kzalloc (data_size, __GFP_DMA |GFP_KERNEL ); +if (!sgcl_data) +return -ENOMEM ; + + cbd.length =cpu_to_le16 (data_size); + + dma =dma_map_single (&priv->si->pdev->dev, + sgcl_data, data_size, +DMA_FROM_DEVICE ); +if (dma_mapping_error (&priv->si->pdev->dev, dma)) { +netdev_err (priv->si->ndev,"DMA mapping failed!\n" ); +kfree (sgcl_data); +return -ENOMEM ; + } + + cbd.addr[0 ] =lower_32_bits (dma); + cbd.addr[1 ] =upper_32_bits (dma); + + sgce = &sgcl_data->sgcl[0 ]; + + sgcl_config->agtst =0x80 ; + + sgcl_data->ct =cpu_to_le32 (sgi->cycletime); + sgcl_data->cte =cpu_to_le32 (sgi->cycletimext); + +if (sgi->init_ipv >=0 ) + sgcl_config->aipv = (sgi->init_ipv &0x7 ) |0x8 ; + +for (i =0 ; i < sgi->num_entries; i++) { +struct action_gate_entry *from = &sgi->entries[i]; +struct sgce *to = &sgce[i]; + +if (from->gate_state) + to->multi |=0x10 ; + +if (from->ipv >=0 ) + to->multi |= ((from->ipv &0x7 ) <<5 ) |0x08 ; + +if (from->maxoctets) + to->multi |=0x01 ; + + to->interval =cpu_to_le32 (from->interval); + to->msdu[0 ] = from->maxoctets &0xFF ; + to->msdu[1 ] = (from->maxoctets >>8 ) &0xFF ; + to->msdu[2 ] = (from->maxoctets >>16 ) &0xFF ; + } + +/* If basetime is 0, calculate start time */ +if (!sgi->basetime) { +u64 start; + + err =get_start_ns (priv, sgi->cycletime, &start); +if (err) + goto exit; + sgcl_data->btl =cpu_to_le32 (lower_32_bits (start)); + sgcl_data->bth =cpu_to_le32 (upper_32_bits (start)); + }else { +u32 hi, lo; + + hi =upper_32_bits (sgi->basetime); + lo =lower_32_bits (sgi->basetime); + sgcl_data->bth =cpu_to_le32 (hi); + sgcl_data->btl =cpu_to_le32 (lo); + } + + err =enetc_send_cmd (priv->si, &cbd); + +exit: +kfree (sgcl_data); + +return err; +} + +static struct enetc_stream_filter *enetc_get_stream_by_index (u32 index) +{ +struct enetc_stream_filter *f; + +hlist_for_each_entry (f, &epsfp.stream_list, node) +if (f->sid.index == index) +return f; + +return NULL ; +} + +static struct enetc_psfp_gate *enetc_get_gate_by_index (u32 index) +{ +struct enetc_psfp_gate *g; + +hlist_for_each_entry (g, &epsfp.psfp_gate_list, node) +if (g->index == index) +return g; + +return NULL ; +} + +static struct enetc_psfp_filter *enetc_get_filter_by_index (u32 index) +{ +struct enetc_psfp_filter *s; + +hlist_for_each_entry (s, &epsfp.psfp_filter_list, node) +if (s->index == index) +return s; + +return NULL ; +} + +static struct enetc_psfp_filter + *enetc_psfp_check_sfi (struct enetc_psfp_filter *sfi) +{ +struct enetc_psfp_filter *s; + +hlist_for_each_entry (s, &epsfp.psfp_filter_list, node) +if (s->gate_id == sfi->gate_id && + s->prio == sfi->prio && + s->meter_id == sfi->meter_id) +return s; + +return NULL ; +} + +static int enetc_get_free_index (struct enetc_ndev_priv *priv) +{ +u32 max_size = priv->psfp_cap.max_psfp_filter; +unsigned long index; + + index =find_first_zero_bit (epsfp.psfp_sfi_bitmap, max_size); +if (index == max_size) +return -1 ; + +return index; +} + +static void stream_filter_unref (struct enetc_ndev_priv *priv,u32 index) +{ +struct enetc_psfp_filter *sfi; +u8 z; + + sfi =enetc_get_filter_by_index (index); +WARN_ON (!sfi); + z =refcount_dec_and_test (&sfi->refcount); + +if (z) { +enetc_streamfilter_hw_set (priv, sfi, false); +hlist_del (&sfi->node); +kfree (sfi); +clear_bit (sfi->index, epsfp.psfp_sfi_bitmap); + } +} + +static void stream_gate_unref (struct enetc_ndev_priv *priv,u32 index) +{ +struct enetc_psfp_gate *sgi; +u8 z; + + sgi =enetc_get_gate_by_index (index); +WARN_ON (!sgi); + z =refcount_dec_and_test (&sgi->refcount); +if (z) { +enetc_streamgate_hw_set (priv, sgi, false); +hlist_del (&sgi->node); +kfree (sgi); + } +} + +static void remove_one_chain (struct enetc_ndev_priv *priv, +struct enetc_stream_filter *filter) +{ +stream_gate_unref (priv, filter->sgi_index); +stream_filter_unref (priv, filter->sfi_index); + +hlist_del (&filter->node); +kfree (filter); +} + +static int enetc_psfp_hw_set (struct enetc_ndev_priv *priv, +struct enetc_streamid *sid, +struct enetc_psfp_filter *sfi, +struct enetc_psfp_gate *sgi) +{ +int err; + + err =enetc_streamid_hw_set (priv, sid, true); +if (err) +return err; + +if (sfi) { + err =enetc_streamfilter_hw_set (priv, sfi, true); +if (err) + goto revert_sid; + } + + err =enetc_streamgate_hw_set (priv, sgi, true); +if (err) + goto revert_sfi; + +return 0 ; + +revert_sfi: +if (sfi) +enetc_streamfilter_hw_set (priv, sfi, false); +revert_sid: +enetc_streamid_hw_set (priv, sid, false); +return err; +} + +struct actions_fwd *enetc_check_flow_actions (u64 acts,unsigned int inputkeys) +{ +int i; + +for (i =0 ; i <ARRAY_SIZE (enetc_act_fwd); i++) +if (acts == enetc_act_fwd[i].actions && + inputkeys & enetc_act_fwd[i].keys) +return &enetc_act_fwd[i]; + +return NULL ; +} + +static int enetc_psfp_parse_clsflower (struct enetc_ndev_priv *priv, +struct flow_cls_offload *f) +{ +struct flow_rule *rule =flow_cls_offload_flow_rule (f); +struct netlink_ext_ack *extack = f->common.extack; +struct enetc_stream_filter *filter, *old_filter; +struct enetc_psfp_filter *sfi, *old_sfi; +struct enetc_psfp_gate *sgi, *old_sgi; +struct flow_action_entry *entry; +struct action_gate_entry *e; +u8 sfi_overwrite =0 ; +int entries_size; +int i, err; + +if (f->common.chain_index >= priv->psfp_cap.max_streamid) { +NL_SET_ERR_MSG_MOD (extack,"No Stream identify resource!" ); +return -ENOSPC ; + } + +flow_action_for_each (i, entry, &rule->action) +if (entry->id ==FLOW_ACTION_GATE ) +break ; + +if (entry->id !=FLOW_ACTION_GATE ) +return -EINVAL ; + + filter =kzalloc (sizeof (*filter),GFP_KERNEL ); +if (!filter)
drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data) · line 1004
Quoted 10 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +10
-0+return -ENOMEM ; + + filter->sid.index = f->common.chain_index; + +if (flow_rule_match_key (rule,FLOW_DISSECTOR_KEY_ETH_ADDRS )) { +struct flow_match_eth_addrs match; + +flow_rule_match_eth_addrs (rule, &match); + +if (!is_zero_ether_addr (match.mask->dst)) {
Does ENETC support masked matching on MAC address? If not, you should error out if the mask is not ff:ff:ff:ff:ff:ff.
Quoted 18 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +18
-0+ether_addr_copy (filter->sid.dst_mac, match.key->dst); + filter->sid.filtertype =STREAMID_TYPE_NULL ; + } + +if (!is_zero_ether_addr (match.mask->src)) { +ether_addr_copy (filter->sid.src_mac, match.key->src); + filter->sid.filtertype =STREAMID_TYPE_SMAC ; + } + }else { +NL_SET_ERR_MSG_MOD (extack,"Unsupported, must ETH_ADDRS" ); +return -EINVAL ; + } + +if (flow_rule_match_key (rule,FLOW_DISSECTOR_KEY_VLAN )) { +struct flow_match_vlan match; + +flow_rule_match_vlan (rule, &match); +if (match.mask->vlan_priority) {
drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data) · line 1032
Quoted 10 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +10
-0+if (match.mask->vlan_priority != + (VLAN_PRIO_MASK >>VLAN_PRIO_SHIFT )) { +NL_SET_ERR_MSG_MOD (extack,"Only full mask is supported for VLAN priority" ); + err = -EINVAL ; + goto free_filter; + } + } + +if (match.mask->vlan_tpid) { +if (match.mask->vlan_tpid !=VLAN_VID_MASK ) {
I'm pretty sure that vlan_tpid is the EtherType (0x8100, etc), and that you actually meant vlan_id.
Quoted 372 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +371
-1+NL_SET_ERR_MSG_MOD (extack,"Only full mask is supported for VLAN id" ); + err = -EINVAL ; + goto free_filter; + } + + filter->sid.vid = match.key->vlan_tpid; +if (!filter->sid.vid) + filter->sid.tagged =STREAMID_VLAN_UNTAGGED ; +else + filter->sid.tagged =STREAMID_VLAN_TAGGED ; + } + }else { + filter->sid.tagged =STREAMID_VLAN_ALL ; + } + +/* parsing gate action */ +if (entry->gate.index >= priv->psfp_cap.max_psfp_gate) { +NL_SET_ERR_MSG_MOD (extack,"No Stream Gate resource!" ); + err = -ENOSPC ; + goto free_filter; + } + +if (entry->gate.num_entries >= priv->psfp_cap.max_psfp_gatelist) { +NL_SET_ERR_MSG_MOD (extack,"No Stream Gate resource!" ); + err = -ENOSPC ; + goto free_filter; + } + + entries_size =struct_size (sgi, entries, entry->gate.num_entries); + sgi =kzalloc (entries_size,GFP_KERNEL ); +if (!sgi) { + err = -ENOMEM ; + goto free_filter; + } + +refcount_set (&sgi->refcount,1 ); + sgi->index = entry->gate.index; + sgi->init_ipv = entry->gate.prio; + sgi->basetime = entry->gate.basetime; + sgi->cycletime = entry->gate.cycletime; + sgi->num_entries = entry->gate.num_entries; + + e = sgi->entries; +for (i =0 ; i < entry->gate.num_entries; i++) { + e[i].gate_state = entry->gate.entries[i].gate_state; + e[i].interval = entry->gate.entries[i].interval; + e[i].ipv = entry->gate.entries[i].ipv; + e[i].maxoctets = entry->gate.entries[i].maxoctets; + } + + filter->sgi_index = sgi->index; + + sfi =kzalloc (sizeof (*sfi),GFP_KERNEL ); +if (!sfi) { + err = -ENOMEM ; + goto free_gate; + } + +refcount_set (&sfi->refcount,1 ); + sfi->gate_id = sgi->index; + +/* flow meter not support yet */ + sfi->meter_id =ENETC_PSFP_WILDCARD ; + +/* prio ref the filter prio */ +if (f->common.prio && f->common.prio <=BIT (3 )) + sfi->prio = f->common.prio -1 ; +else + sfi->prio =ENETC_PSFP_WILDCARD ; + + old_sfi =enetc_psfp_check_sfi (sfi); +if (!old_sfi) { +int index; + + index =enetc_get_free_index (priv); +if (sfi->handle <0 ) { +NL_SET_ERR_MSG_MOD (extack,"No Stream Filter resource!" ); + err = -ENOSPC ; + goto free_sfi; + } + + sfi->index = index; + sfi->handle = index +HANDLE_OFFSET ; +/* Update the stream filter handle also */ + filter->sid.handle = sfi->handle; + filter->sfi_index = sfi->index; + sfi_overwrite =0 ; + }else { + filter->sfi_index = old_sfi->index; + filter->sid.handle = old_sfi->handle; + sfi_overwrite =1 ; + } + + err =enetc_psfp_hw_set (priv, &filter->sid, + sfi_overwrite ?NULL : sfi, sgi); +if (err) + goto free_sfi; + +spin_lock (&epsfp.psfp_lock); +/* Remove the old node if exist and update with a new node */ + old_sgi =enetc_get_gate_by_index (filter->sgi_index); +if (old_sgi) { +refcount_set (&sgi->refcount, +refcount_read (&old_sgi->refcount) +1 ); +hlist_del (&old_sgi->node); +kfree (old_sgi); + } + +hlist_add_head (&sgi->node, &epsfp.psfp_gate_list); + +if (!old_sfi) { +hlist_add_head (&sfi->node, &epsfp.psfp_filter_list); +set_bit (sfi->index, epsfp.psfp_sfi_bitmap); + }else { +kfree (sfi); +refcount_inc (&old_sfi->refcount); + } + + old_filter =enetc_get_stream_by_index (filter->sid.index); +if (old_filter) +remove_one_chain (priv, old_filter); + + filter->stats.lastused = jiffies; +hlist_add_head (&filter->node, &epsfp.stream_list); + +spin_unlock (&epsfp.psfp_lock); + +return 0 ; + +free_sfi: +kfree (sfi); +free_gate: +kfree (sgi); +free_filter: +kfree (filter); + +return err; +} + +static int enetc_config_clsflower (struct enetc_ndev_priv *priv, +struct flow_cls_offload *cls_flower) +{ +struct flow_rule *rule =flow_cls_offload_flow_rule (cls_flower); +struct netlink_ext_ack *extack = cls_flower->common.extack; +struct flow_dissector *dissector = rule->match.dissector; +struct flow_action *action = &rule->action; +struct flow_action_entry *entry; +struct actions_fwd *fwd; +u64 actions =0 ; +int i, err; + +if (!flow_action_has_entries (action)) { +NL_SET_ERR_MSG_MOD (extack,"At least one action is needed" ); +return -EINVAL ; + } + +flow_action_for_each (i, entry, action) + actions |=BIT (entry->id); + + fwd =enetc_check_flow_actions (actions, dissector->used_keys); +if (!fwd) { +NL_SET_ERR_MSG_MOD (extack,"Unsupported filter type!" ); +return -EOPNOTSUPP ; + } + +if (fwd->output &FILTER_ACTION_TYPE_PSFP ) { + err =enetc_psfp_parse_clsflower (priv, cls_flower); +if (err) { +NL_SET_ERR_MSG_MOD (extack,"Invalid PSFP inputs" ); +return err; + } + }else { +NL_SET_ERR_MSG_MOD (extack,"Unsupported actions" ); +return -EOPNOTSUPP ; + } + +return 0 ; +} + +static int enetc_psfp_destroy_clsflower (struct enetc_ndev_priv *priv, +struct flow_cls_offload *f) +{ +struct enetc_stream_filter *filter; +struct netlink_ext_ack *extack = f->common.extack; +int err; + +if (f->common.chain_index >= priv->psfp_cap.max_streamid) { +NL_SET_ERR_MSG_MOD (extack,"No Stream identify resource!" ); +return -ENOSPC ; + } + + filter =enetc_get_stream_by_index (f->common.chain_index); +if (!filter) +return -EINVAL ; + + err =enetc_streamid_hw_set (priv, &filter->sid, false); +if (err) +return err; + +remove_one_chain (priv, filter); + +return 0 ; +} + +static int enetc_destroy_clsflower (struct enetc_ndev_priv *priv, +struct flow_cls_offload *f) +{ +return enetc_psfp_destroy_clsflower (priv, f); +} + +static int enetc_psfp_get_stats (struct enetc_ndev_priv *priv, +struct flow_cls_offload *f) +{ +struct psfp_streamfilter_counters counters = {}; +struct enetc_stream_filter *filter; +struct flow_stats stats = {}; +int err; + + filter =enetc_get_stream_by_index (f->common.chain_index); +if (!filter) +return -EINVAL ; + + err =enetc_streamcounter_hw_get (priv, filter->sfi_index, &counters); +if (err) +return -EINVAL ; + +spin_lock (&epsfp.psfp_lock); + stats.pkts = counters.matching_frames_count - filter->stats.pkts; + stats.lastused = filter->stats.lastused; + filter->stats.pkts += stats.pkts; +spin_unlock (&epsfp.psfp_lock); + +flow_stats_update (&f->stats,0x0 , stats.pkts, stats.lastused, +FLOW_ACTION_HW_STATS_DELAYED ); + +return 0 ; +} + +static int enetc_setup_tc_cls_flower (struct enetc_ndev_priv *priv, +struct flow_cls_offload *cls_flower) +{ +switch (cls_flower->command) { +case FLOW_CLS_REPLACE : +return enetc_config_clsflower (priv, cls_flower); +case FLOW_CLS_DESTROY : +return enetc_destroy_clsflower (priv, cls_flower); +case FLOW_CLS_STATS : +return enetc_psfp_get_stats (priv, cls_flower); +default : +return -EOPNOTSUPP ; + } +} + +static inline void clean_psfp_sfi_bitmap (void ) +{ +bitmap_free (epsfp.psfp_sfi_bitmap); + epsfp.psfp_sfi_bitmap =NULL ; +} + +static void clean_stream_list (void ) +{ +struct enetc_stream_filter *s; +struct hlist_node *tmp; + +hlist_for_each_entry_safe (s, tmp, &epsfp.stream_list, node) { +hlist_del (&s->node); +kfree (s); + } +} + +static void clean_sfi_list (void ) +{ +struct enetc_psfp_filter *sfi; +struct hlist_node *tmp; + +hlist_for_each_entry_safe (sfi, tmp, &epsfp.psfp_filter_list, node) { +hlist_del (&sfi->node); +kfree (sfi); + } +} + +static void clean_sgi_list (void ) +{ +struct enetc_psfp_gate *sgi; +struct hlist_node *tmp; + +hlist_for_each_entry_safe (sgi, tmp, &epsfp.psfp_gate_list, node) { +hlist_del (&sgi->node); +kfree (sgi); + } +} + +static void clean_psfp_all (void ) +{ +/* Disable all list nodes and free all memory */ +clean_sfi_list (); +clean_sgi_list (); +clean_stream_list (); + epsfp.dev_bitmap =0 ; +clean_psfp_sfi_bitmap (); +} + +int enetc_setup_tc_block_cb (enum tc_setup_type type,void *type_data, +void *cb_priv) +{ +struct net_device *ndev = cb_priv; + +if (!tc_can_offload (ndev)) +return -EOPNOTSUPP ; + +switch (type) { +case TC_SETUP_CLSFLOWER : +return enetc_setup_tc_cls_flower (netdev_priv (ndev), type_data); +default : +return -EOPNOTSUPP ; + } +} + +int enetc_psfp_init (struct enetc_ndev_priv *priv) +{ +if (epsfp.psfp_sfi_bitmap) +return 0 ; + + epsfp.psfp_sfi_bitmap =bitmap_zalloc (priv->psfp_cap.max_psfp_filter, +GFP_KERNEL ); +if (!epsfp.psfp_sfi_bitmap) +return -ENOMEM ; + +spin_lock_init (&epsfp.psfp_lock); + +if (list_empty (&enetc_block_cb_list)) + epsfp.dev_bitmap =0 ; + +return 0 ; +} + +int enetc_psfp_clean (struct enetc_ndev_priv *priv) +{ +if (!list_empty (&enetc_block_cb_list)) +return -EBUSY ; + +clean_psfp_all (); + +return 0 ; +} + +int enetc_setup_tc_psfp (struct net_device *ndev,void *type_data) +{ +struct enetc_ndev_priv *priv =netdev_priv (ndev); +struct flow_block_offload *f = type_data; +int err; + + err =flow_block_cb_setup_simple (f, &enetc_block_cb_list, + enetc_setup_tc_block_cb, + ndev, ndev, true); +if (err) +return err; + +switch (f->command) { +case FLOW_BLOCK_BIND : +set_bit (enetc_get_port (priv), &epsfp.dev_bitmap); +break ; +case FLOW_BLOCK_UNBIND : +clear_bit (enetc_get_port (priv), &epsfp.dev_bitmap); +if (!epsfp.dev_bitmap) +clean_psfp_all (); +break ; + } + +return 0 ; +} --
Quoted 2 lines
2.17.1
Thanks,
-Vladimir
RE: [EXT] Re: [ v2,net-next 4/4] net: enetc: add tc flower psfp offload driver
Hi Vladimir,
Quoted 30 lines
-----Original Message-----
From: Vladimir Oltean <[email protected]>
Sent: 2020年4月19日 6:53
To: Po Liu <[email protected]>
Cc: David S. Miller <[email protected]>; lkml <linux- [email protected]>; netdev <[email protected]>; Vinicius Costa Gomes <[email protected]>; Claudiu Manoil <[email protected]>; Vladimir Oltean <[email protected]>; Alexandru Marginean <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]; Jiri Pirko <[email protected]>; Ido Schimmel <[email protected]>; Alexandre Belloni <[email protected]>; Microchip Linux Driver Support <[email protected]>; Jakub Kicinski <[email protected]>; Jamal Hadi Salim <[email protected]>; Cong Wang <[email protected]>; [email protected]; [email protected]; [email protected]; Murali Karicheri <m- [email protected]>; Andre Guedes <[email protected]>; Stephen Hemminger <[email protected]>
Subject: [EXT] Re: [ v2,net-next 4/4] net: enetc: add tc flower psfp offload driverCaution: EXT Email
Hi Po,
Quoted 2 lines · Po Liu
+ if (
flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ETH_ADDRS)){
Quoted 8 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data) · line 1004
Quoted 5 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +5
-0+struct flow_match_eth_addrs match; + +flow_rule_match_eth_addrs (rule, &match); + +if (!is_zero_ether_addr (match.mask->dst)) {Does ENETC support masked matching on MAC address? If not, you should error out if the mask is not ff:ff:ff:ff:ff:ff.
I get it. Thanks.
Quoted 32 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data)
Quoted 11 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +10
-0+ether_addr_copy (filter->sid.dst_mac, match.key->dst); + filter->sid.filtertype =STREAMID_TYPE_NULL ; + } + +if (!is_zero_ether_addr (match.mask->src)) { +ether_addr_copy (filter->sid.src_mac, match.key->src); + filter->sid.filtertype =STREAMID_TYPE_SMAC ; + } + }else { + NL_SET_ERR_MSG_MOD(extack, "Unsupported, must
ETH_ADDRS");drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data)
Quoted 11 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +11
-0+return -EINVAL ; + } + +if (flow_rule_match_key (rule,FLOW_DISSECTOR_KEY_VLAN )) { +struct flow_match_vlan match; + +flow_rule_match_vlan (rule, &match); +if (match.mask->vlan_priority) { +if (match.mask->vlan_priority != + (VLAN_PRIO_MASK >>VLAN_PRIO_SHIFT )) { + NL_SET_ERR_MSG_MOD(extack, "Only full mask issupported for VLAN priority");
drivers/net/ethernet/freescale/enetc/enetc_qos.c int enetc_setup_tc_txtime(struct net_device *ndev, void *type_data) · line 1032
Quoted 7 lines
drivers/net/ethernet/freescale/enetc/enetc_qos.c +7
-0+ err = -EINVAL ; + goto free_filter; + } + } + +if (match.mask->vlan_tpid) { +if (match.mask->vlan_tpid !=VLAN_VID_MASK ) {
Quoted 3 lines
I'm pretty sure that
vlan_tpidis the EtherType (0x8100, etc), and that you actually meantvlan_id.
Yes, I'll correct it.
Quoted 6 lines
Quoted 3 lines
--
2.17.1Thanks,
-Vladimir
Br,
Po Liu
v3 cover Introduce a flow gate control action and apply IEEE
In reply to Po LiuChanges from V2:
0001: No changes.
0002: No changes.
0003: No changes.
0004: Fix the vlan id filter parameter and add reject src mac FF-FF-FF-FF-FF-FF filter in driver.
Changes from V1:
0000: Update description make it more clear
0001: Removed 'add update dropped stats' patch, will provide pull request as standalone patches.
0001: Update commit description make it more clear ack by Jiri Pirko. 0002: No changes
0003: Fix some code style ack by Jiri Pirko.
0004: Fix enetc_psfp_enable/disable parameter type ack by test robot
iprout2 command patches:
Not attach with these serial patches, will provide separate pull
request after kernel accept these patches.
Changes from RFC:
0000: Reduce to 5 patches and remove the 4 max frame size offload and flow metering in the policing offload action, Only keep gate action offloading implementation.
0001: No changes.
0002:
- fix kfree lead ack by Jakub Kicinski and Cong Wang
- License fix from Jakub Kicinski and Stephen Hemminger
- Update example in commit acked by Vinicius Costa Gomes
- Fix the rcu protect in tcf_gate_act() acked by Vinicius
0003: No changes
0004: No changes
0005:
Acked by Vinicius Costa Gomes
- Use refcount kernel lib
- Update stream gate check code position
- Update reduce ref names more clear
iprout2 command patches:
0000: Update license expression and add gate id
0001: Add tc action gate man page
--------------------------------------------------------------------
These patches add stream gate action policing in IEEE802.1Qci (Per-Stream Filtering and Policing) software support and hardware offload support in tc flower, and implement the stream identify, stream filtering and stream gate filtering action in the NXP ENETC ethernet driver. Per-Stream Filtering and Policing (PSFP) specifies flow policing and filtering for ingress flows, and has three main parts:
1. The stream filter instance table consists of an ordered list of
stream filters that determine the filtering and policing actions that are to be applied to frames received on a specific stream. The main elements are stream gate id, flow metering id and maximum SDU size.
2. The stream gate function setup a gate list to control ingress traffic
class open/close state. When the gate is running at open state, the flow could pass but dropped when gate state is running to close. User setup a bastime to tell gate when start running the entry list, then the hardware would periodiclly. There is no compare qdisc action support.
3. Flow metering is two rates two buckets and three-color marker to
policing the frames. Flow metering instance are as specified in the algorithm in MEF10.3. The most likely qdisc action is policing action.
The first patch introduces an ingress frame flow control gate action, for the point 2. The tc gate action maintains the open/close state gate list, allowing flows to pass when the gate is open. Each gate action may policing one or more qdisc filters. When the start time arrived, The driver would repeat the gate list periodiclly. User can assign a passed time, the driver would calculate a new future time by the cycletime of the gate list.
The 0002 patch introduces the gate flow hardware offloading.
The 0003 patch adds support control the on/off for the tc flower offloading by ethtool.
The 0004 patch implement the stream identify and stream filtering and stream gate filtering action in the NXP ENETC ethernet driver. Tc filter command provide filtering keys with MAC address and VLAN id. These keys would be set to stream identify instance entry. Stream gate instance entry would refer the gate action parameters. Stream filter instance entry would refer the stream gate index and assign a stream handle value matches to the stream identify instance.
Po Liu (4):
net: qos: introduce a gate control flow action
net: schedule: add action gate offloading
net: enetc: add hw tc hw offload features for PSPF capability
net: enetc: add tc flower psfp offload driver
drivers/net/ethernet/freescale/enetc/enetc.c | 34 +- drivers/net/ethernet/freescale/enetc/enetc.h | 86 ++ .../net/ethernet/freescale/enetc/enetc_hw.h | 159 +++ .../net/ethernet/freescale/enetc/enetc_pf.c | 6 + .../net/ethernet/freescale/enetc/enetc_qos.c | 1082 +++++++++++++++++ include/net/flow_offload.h | 10 + include/net/tc_act/tc_gate.h | 169 +++ include/uapi/linux/pkt_cls.h | 1 + include/uapi/linux/tc_act/tc_gate.h | 47 + net/sched/Kconfig | 13 + net/sched/Makefile | 1 + net/sched/act_gate.c | 647 ++++++++++ net/sched/cls_api.c | 33 + 13 files changed, 2287 insertions(+), 1 deletion(-) create mode 100644 include/net/tc_act/tc_gate.h create mode 100644 include/uapi/linux/tc_act/tc_gate.h create mode 100644 net/sched/act_gate.c
Signature · 1 line
2.17.1
PATCH v3 1/4 net: qos: introduce a gate control flow action
Introduce a ingress frame gate control flow action.
Tc gate action does the work like this:
Assume there is a gate allow specified ingress frames can be passed at specific time slot, and be dropped at specific time slot. Tc filter chooses the ingress frames, and tc gate action would specify what slot does these frames can be passed to device and what time slot would be dropped.
Tc gate action would provide an entry list to tell how much time gate keep open and how much time gate keep state close. Gate action also assign a start time to tell when the entry list start. Then driver would repeat the gate entry list cyclically.
For the software simulation, gate action requires the user assign a time clock type.
Below is the setting example in user space. Tc filter a stream source ip address is 192.168.0.20 and gate action own two time slots. One is last 200ms gate open let frame pass another is last 100ms gate close let frames dropped. When the frames have passed total frames over 8000000 bytes, frames will be dropped in one 200000000ns time slot.
Quoted 3 lines
tc qdisc add dev eth0 ingress
tc filter add dev eth0 parent ffff: protocol ip \
flower src_ip 192.168.0.20 \ action gate index 2 clockid CLOCK_TAI \ sched-entry open 200000000 -1 8000000 \ sched-entry close 100000000 -1 -1
Quoted 1 line
tc chain del dev eth0 ingress chain 0
"sched-entry" follow the name taprio style. Gate state is "open"/"close". Follow with period nanosecond. Then next item is internal priority value means which ingress queue should put. "-1" means wildcard. The last value optional specifies the maximum number of MSDU octets that are permitted to pass the gate during the specified time interval.
Base-time is not set will be 0 as default, as result start time would be ((N + 1) * cycletime) which is the minimal of future time.
Below example shows filtering a stream with destination mac address is 10:00:80:00:00:00 and ip type is ICMP, follow the action gate. The gate action would run with one close time slot which means always keep close. The time cycle is total 200000000ns. The base-time would calculate by:
1357000000000 + (N + 1) * cycletime
When the total value is the future time, it will be the start time. The cycletime here would be 200000000ns for this case.
Quoted 1 line
tc filter add dev eth0 parent ffff: protocol ip \
flower skip_hw ip_proto icmp dst_mac 10:00:80:00:00:00 \ action gate index 12 base-time 1357000000000 \ sched-entry close 200000000 -1 -1 \ clockid CLOCK_TAI
--- include/net/tc_act/tc_gate.h | 54 +++ include/uapi/linux/pkt_cls.h | 1 + include/uapi/linux/tc_act/tc_gate.h | 47 ++ net/sched/Kconfig | 13 + net/sched/Makefile | 1 + net/sched/act_gate.c | 647 ++++++++++++++++++++++++++++ 6 files changed, 763 insertions(+) create mode 100644 include/net/tc_act/tc_gate.h create mode 100644 include/uapi/linux/tc_act/tc_gate.h create mode 100644 net/sched/act_gate.c
include/net/tc_act/tc_gate.h new file +54 -0
@@ -0,0 +1,54 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +/* Copyright 2020 NXP */ + +#ifndef __NET_TC_GATE_H +#define __NET_TC_GATE_H + +#include <net/act_api.h> +#include <linux/tc_act/tc_gate.h> + +struct tcfg_gate_entry { +int index; +u8 gate_state; +u32 interval; +s32 ipv; +s32 maxoctets; +struct list_head list; +}; + +struct tcf_gate_params { +s32 tcfg_priority; +u64 tcfg_basetime; +u64 tcfg_cycletime; +u64 tcfg_cycletime_ext; +u32 tcfg_flags; +s32 tcfg_clockid; +size_t num_entries; +struct list_head entries; +}; + +#define GATE_ACT_GATE_OPEN BIT(0) +#define GATE_ACT_PENDING BIT(1) +struct gate_action { +struct tcf_gate_params param; +spinlock_t entry_lock; +u8 current_gate_status; +ktime_t current_close_time; +u32 current_entry_octets; +s32 current_max_octets; +struct tcfg_gate_entry __rcu *next_entry; +struct hrtimer hitimer; +enum tk_offsets tk_offset; +struct rcu_head rcu; +}; + +struct tcf_gate { +struct tc_action common; +struct gate_action __rcu *actg; +}; +#define to_gate (a) ((struct tcf_gate *)a) + +#define get_gate_param (act) ((struct tcf_gate_params *)act) +#define get_gate_action (p) ((struct gate_action *)p) + +#endif
include/uapi/linux/pkt_cls.h +1 -0
@@ -134,6 +134,7 @@ enum tca_id {TCA_ID_CTINFO ,TCA_ID_MPLS ,TCA_ID_CT , +TCA_ID_GATE ,/* other actions go here */ __TCA_ID_MAX = 255 };
include/uapi/linux/tc_act/tc_gate.h new file +47 -0
47 lines not on this page · raw message
net/sched/Kconfig +13 -0
@@ -981,6 +981,19 @@ config NET_ACT_CT To compile this code as a module, choose M here: the module will be called act_ct. +config NET_ACT_GATE + tristate "Frame gate entry list control tc action" + depends on NET_CLS_ACT + help + Say Y here to allow to control the ingress flow to be passed at + specific time slot and be dropped at other specific time slot by + the gate entry list. The manipulation will simulate the IEEE + 802.1Qci stream gate control behavior. + + If unsure, say N. + To compile this code as a module, choose M here: the + module will be called act_gate. + config NET_IFE_SKBMARK tristate "Support to encoding decoding skb mark on IFE action" depends on NET_ACT_IFE
net/sched/Makefile +1 -0
@@ -30,6 +30,7 @@ obj-$(CONFIG_NET_IFE_SKBPRIO) += act_meta_skbprio.o obj-$(CONFIG_NET_IFE_SKBTCINDEX) += act_meta_skbtcindex.o obj-$(CONFIG_NET_ACT_TUNNEL_KEY)+= act_tunnel_key.o obj-$(CONFIG_NET_ACT_CT) += act_ct.o +obj-$(CONFIG_NET_ACT_GATE) += act_gate.o obj-$(CONFIG_NET_SCH_FIFO) += sch_fifo.o obj-$(CONFIG_NET_SCH_CBQ) += sch_cbq.o obj-$(CONFIG_NET_SCH_HTB) += sch_htb.o
net/sched/act_gate.c new file +647 -0
647 lines not on this page · raw message
Signature · 1 line
2.17.1
Re: [v3,net-next 1/4] net: qos: introduce a gate control flow action
Hi Po,
Quoted 278 lines · Po Liu
Introduce a ingress frame gate control flow action.
Tc gate action does the work like this:
Assume there is a gate allow specified ingress frames can be passed at specific time slot, and be dropped at specific time slot. Tc filter chooses the ingress frames, and tc gate action would specify what slot does these frames can be passed to device and what time slot would be dropped.
Tc gate action would provide an entry list to tell how much time gate keep open and how much time gate keep state close. Gate action also assign a start time to tell when the entry list start. Then driver would repeat the gate entry list cyclically.
For the software simulation, gate action requires the user assign a time clock type.Below is the setting example in user space. Tc filter a stream source ip address is 192.168.0.20 and gate action own two time slots. One is last 200ms gate open let frame pass another is last 100ms gate close let frames dropped. When the frames have passed total frames over 8000000 bytes, frames will be dropped in one 200000000ns time slot.
Quoted 3 lines
tc qdisc add dev eth0 ingress
tc filter add dev eth0 parent ffff: protocol ip \
flower src_ip 192.168.0.20 \ action gate index 2 clockid CLOCK_TAI \ sched-entry open 200000000 -1 8000000 \ sched-entry close 100000000 -1 -1Quoted 1 line
tc chain del dev eth0 ingress chain 0
"sched-entry" follow the name taprio style. Gate state is "open"/"close". Follow with period nanosecond. Then next item is internal priority value means which ingress queue should put. "-1" means wildcard. The last value optional specifies the maximum number of MSDU octets that are permitted to pass the gate during the specified time interval.
Base-time is not set will be 0 as default, as result start time would be ((N + 1) * cycletime) which is the minimal of future time.Below example shows filtering a stream with destination mac address is 10:00:80:00:00:00 and ip type is ICMP, follow the action gate. The gate action would run with one close time slot which means always keep close. The time cycle is total 200000000ns. The base-time would calculate by:
1357000000000 + (N + 1) * cycletime
When the total value is the future time, it will be the start time. The cycletime here would be 200000000ns for this case.
Quoted 1 line
tc filter add dev eth0 parent ffff: protocol ip \
flower skip_hw ip_proto icmp dst_mac 10:00:80:00:00:00 \ action gate index 12 base-time 1357000000000 \ sched-entry close 200000000 -1 -1 \ clockid CLOCK_TAISigned-off-by Po Liu <[email protected]>--- include/net/tc_act/tc_gate.h | 54 +++ include/uapi/linux/pkt_cls.h | 1 + include/uapi/linux/tc_act/tc_gate.h | 47 ++ net/sched/Kconfig | 13 + net/sched/Makefile | 1 + net/sched/act_gate.c | 647 ++++++++++++++++++++++++++++ 6 files changed, 763 insertions(+) create mode 100644 include/net/tc_act/tc_gate.h create mode 100644 include/uapi/linux/tc_act/tc_gate.h create mode 100644 net/sched/act_gate.cinclude/net/tc_act/tc_gate.h new file +54
-0@@ -0,0 +1,54 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +/* Copyright 2020 NXP */ + +#ifndef __NET_TC_GATE_H +#define __NET_TC_GATE_H + +#include <net/act_api.h> +#include <linux/tc_act/tc_gate.h> + +struct tcfg_gate_entry { +int index; +u8 gate_state; +u32 interval; +s32 ipv; +s32 maxoctets; +struct list_head list; +}; + +struct tcf_gate_params { +s32 tcfg_priority; +u64 tcfg_basetime; +u64 tcfg_cycletime; +u64 tcfg_cycletime_ext; +u32 tcfg_flags; +s32 tcfg_clockid; +size_t num_entries; +struct list_head entries; +}; + +#define GATE_ACT_GATE_OPEN BIT(0) +#define GATE_ACT_PENDING BIT(1) +struct gate_action { +struct tcf_gate_params param; +spinlock_t entry_lock; +u8 current_gate_status; +ktime_t current_close_time; +u32 current_entry_octets; +s32 current_max_octets; +struct tcfg_gate_entry __rcu *next_entry; +struct hrtimer hitimer; +enum tk_offsets tk_offset; +struct rcu_head rcu; +}; + +struct tcf_gate { +struct tc_action common; +struct gate_action __rcu *actg; +}; +#define to_gate (a) ((struct tcf_gate *)a) + +#define get_gate_param (act) ((struct tcf_gate_params *)act) +#define get_gate_action (p) ((struct gate_action *)p) + +#endif include/uapi/linux/pkt_cls.h +1
-0@@ -134,6 +134,7 @@ enum tca_id {TCA_ID_CTINFO ,TCA_ID_MPLS ,TCA_ID_CT , +TCA_ID_GATE ,/* other actions go here */ __TCA_ID_MAX = 255 };include/uapi/linux/tc_act/tc_gate.h new file +47
-0@@ -0,0 +1,47 @@ +/* SPDX-License-Identifier: GPL-2.0+ WITH Linux-syscall-note */ +/* Copyright 2020 NXP */ + +#ifndef __LINUX_TC_GATE_H +#define __LINUX_TC_GATE_H + +#include <linux/pkt_cls.h> + +struct tc_gate { +tc_gen ; +}; + +enum { +TCA_GATE_ENTRY_UNSPEC , +TCA_GATE_ENTRY_INDEX , +TCA_GATE_ENTRY_GATE , +TCA_GATE_ENTRY_INTERVAL , +TCA_GATE_ENTRY_IPV , +TCA_GATE_ENTRY_MAX_OCTETS , + __TCA_GATE_ENTRY_MAX, +}; +#define TCA_GATE_ENTRY_MAX (__TCA_GATE_ENTRY_MAX - 1) + +enum { +TCA_GATE_ONE_ENTRY_UNSPEC , +TCA_GATE_ONE_ENTRY , + __TCA_GATE_ONE_ENTRY_MAX, +}; +#define TCA_GATE_ONE_ENTRY_MAX (__TCA_GATE_ONE_ENTRY_MAX - 1) + +enum { +TCA_GATE_UNSPEC , +TCA_GATE_TM , +TCA_GATE_PARMS , +TCA_GATE_PAD , +TCA_GATE_PRIORITY , +TCA_GATE_ENTRY_LIST , +TCA_GATE_BASE_TIME , +TCA_GATE_CYCLE_TIME , +TCA_GATE_CYCLE_TIME_EXT , +TCA_GATE_FLAGS , +TCA_GATE_CLOCKID , + __TCA_GATE_MAX, +}; +#define TCA_GATE_MAX (__TCA_GATE_MAX - 1) + +#endif net/sched/Kconfig +13
-0@@ -981,6 +981,19 @@ config NET_ACT_CT To compile this code as a module, choose M here: the module will be called act_ct. +config NET_ACT_GATE + tristate "Frame gate entry list control tc action" + depends on NET_CLS_ACT + help + Say Y here to allow to control the ingress flow to be passed at + specific time slot and be dropped at other specific time slot by + the gate entry list. The manipulation will simulate the IEEE + 802.1Qci stream gate control behavior. + + If unsure, say N. + To compile this code as a module, choose M here: the + module will be called act_gate. + config NET_IFE_SKBMARK tristate "Support to encoding decoding skb mark on IFE action" depends on NET_ACT_IFEnet/sched/Makefile +1
-0@@ -30,6 +30,7 @@ obj-$(CONFIG_NET_IFE_SKBPRIO) += act_meta_skbprio.o obj-$(CONFIG_NET_IFE_SKBTCINDEX) += act_meta_skbtcindex.o obj-$(CONFIG_NET_ACT_TUNNEL_KEY)+= act_tunnel_key.o obj-$(CONFIG_NET_ACT_CT) += act_ct.o +obj-$(CONFIG_NET_ACT_GATE) += act_gate.o obj-$(CONFIG_NET_SCH_FIFO) += sch_fifo.o obj-$(CONFIG_NET_SCH_CBQ) += sch_cbq.o obj-$(CONFIG_NET_SCH_HTB) += sch_htb.onet/sched/act_gate.c new file +43
-0@@ -0,0 +1,647 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* Copyright 2020 NXP */ + +#include <linux/module.h> +#include <linux/types.h> +#include <linux/kernel.h> +#include <linux/string.h> +#include <linux/errno.h> +#include <linux/skbuff.h> +#include <linux/rtnetlink.h> +#include <linux/init.h> +#include <linux/slab.h> +#include <net/act_api.h> +#include <net/netlink.h> +#include <net/pkt_cls.h> +#include <net/tc_act/tc_gate.h> + +static unsigned int gate_net_id; +static struct tc_action_ops act_gate_ops; + +static ktime_t gate_get_time (struct gate_action *gact) +{ +ktime_t mono =ktime_get (); + +switch (gact->tk_offset) { +case TK_OFFS_MAX : +return mono; +default : +return ktime_mono_to_any (mono, gact->tk_offset); + } + +return KTIME_MAX ; +} + +static int gate_get_start_time (struct gate_action *gact,ktime_t *start) +{ +struct tcf_gate_params *param =get_gate_param (gact); +ktime_t now, base, cycle; +u64 n; + + base =ns_to_ktime (param->tcfg_basetime); + now =gate_get_time (gact); +
net/sched/act_gate.c line 53
Quoted 10 lines
net/sched/act_gate.c new file +10
-0+if (ktime_after (base, now)) { + *start = base; +return 0 ; + } + + cycle = param->tcfg_cycletime; + +/* cycle time should not be zero */ +if (WARN_ON (!cycle)) +return -EFAULT ;
Looking at the init code it seems that this value can be set to 0 directly from netlink packet without further validation, which would allow user to trigger warning here.
Quoted 52 lines
net/sched/act_gate.c new file +52
-0+ + n =div64_u64 (ktime_sub_ns (now, base), cycle); + *start =ktime_add_ns (base, (n +1 ) * cycle); +return 0 ; +} + +static void gate_start_timer (struct gate_action *gact,ktime_t start) +{ +ktime_t expires; + + expires =hrtimer_get_expires (&gact->hitimer); +if (expires ==0 ) + expires =KTIME_MAX ; + + start =min_t (ktime_t, start, expires); + +hrtimer_start (&gact->hitimer, start,HRTIMER_MODE_ABS ); +} + +static enum hrtimer_restart gate_timer_func (struct hrtimer *timer) +{ + struct gate_action *gact = container_of(timer, struct gate_action, + hitimer); +struct tcf_gate_params *p =get_gate_param (gact); +struct tcfg_gate_entry *next; +ktime_t close_time, now; + +spin_lock (&gact->entry_lock); + + next =rcu_dereference_protected (gact->next_entry, +lockdep_is_held (&gact->entry_lock)); + +/* cycle start, clear pending bit, clear total octets */ + gact->current_gate_status = next->gate_state ?GATE_ACT_GATE_OPEN :0 ; + gact->current_entry_octets =0 ; + gact->current_max_octets = next->maxoctets; + + gact->current_close_time =ktime_add_ns (gact->current_close_time, + next->interval); + + close_time = gact->current_close_time; + +if (list_is_last (&next->list, &p->entries)) + next =list_first_entry (&p->entries, + struct tcfg_gate_entry, list); +else + next =list_next_entry (next, list); + + now =gate_get_time (gact); + +if (ktime_after (now, close_time)) { +ktime_t cycle, base;
net/sched/act_gate.c line 115
Quoted 10 lines
net/sched/act_gate.c new file +10
-0+u64 n; + + cycle = p->tcfg_cycletime; + base =ns_to_ktime (p->tcfg_basetime); + n =div64_u64 (ktime_sub_ns (now, base), cycle); + close_time =ktime_add_ns (base, (n +1 ) * cycle); + } + +rcu_assign_pointer (gact->next_entry, next); +spin_unlock (&gact->entry_lock);
I have couple of question about synchronization here:
- Why do you need next_entry to be rcu pointer? It is only assigned here with entry_lock protection and in init code before action is visible to concurrent users. I don't see any unlocked rcu-protected readers here that could benefit from it.
- Why create dedicated entry_lock instead of using already existing per-action tcf_lock?
Quoted 8 lines
net/sched/act_gate.c new file +8
-0+ +hrtimer_set_expires (&gact->hitimer, close_time); + +return HRTIMER_RESTART ; +} + +static int tcf_gate_act (struct sk_buff *skb,const struct tc_action *a, +struct tcf_result *res)
net/sched/act_gate.c line 133
Quoted 10 lines
net/sched/act_gate.c new file +10
-0+{ +struct tcf_gate *g =to_gate (a); +struct gate_action *gact; +int action; + +tcf_lastuse_update (&g->tcf_tm); +bstats_cpu_update (this_cpu_ptr (g->common.cpu_bstats), skb); + + action =READ_ONCE (g->tcf_action); +rcu_read_lock ();
Action fastpath is already rcu read lock protected, you don't need to manually obtain it.
net/sched/act_gate.c line 135
Quoted 2 lines
net/sched/act_gate.c new file +2
-0+ gact =rcu_dereference_bh (g->actg); +if (unlikely (gact->current_gate_status &GATE_ACT_PENDING )) {
Can't current_gate_status be concurrently modified by timer callback? This function doesn't use entry_lock to synchronize with timer.
net/sched/act_gate.c line 140
Quoted 5 lines
net/sched/act_gate.c new file +5
-0+rcu_read_unlock (); +return action; + } + +if (!(gact->current_gate_status &GATE_ACT_GATE_OPEN ))
...and here
net/sched/act_gate.c line 145
Quoted 5 lines
net/sched/act_gate.c new file +5
-0+ goto drop; + +if (gact->current_max_octets >=0 ) { + gact->current_entry_octets +=qdisc_pkt_len (skb); +if (gact->current_entry_octets > gact->current_max_octets) {
here also.
net/sched/act_gate.c line 146
Quoted 1 line
net/sched/act_gate.c new file +1
-0+qstats_overlimit_inc (this_cpu_ptr (g->common.cpu_qstats));
Please use tcf_action_inc_overlimit_qstats() and other wrappers for stats. Otherwise it will crash if user passes TCA_ACT_FLAGS_NO_PERCPU_STATS flag.
Quoted 273 lines
net/sched/act_gate.c new file +272
-0+ goto drop; + } + } +rcu_read_unlock (); + +return action; +drop: +rcu_read_unlock (); +qstats_drop_inc (this_cpu_ptr (g->common.cpu_qstats)); +return TC_ACT_SHOT ; +} + +static const struct nla_policy entry_policy[TCA_GATE_ENTRY_MAX +1 ] = { + [TCA_GATE_ENTRY_INDEX ] = { .type =NLA_U32 }, + [TCA_GATE_ENTRY_GATE ] = { .type =NLA_FLAG }, + [TCA_GATE_ENTRY_INTERVAL ] = { .type =NLA_U32 }, + [TCA_GATE_ENTRY_IPV ] = { .type =NLA_S32 }, + [TCA_GATE_ENTRY_MAX_OCTETS ] = { .type =NLA_S32 }, +}; + +static const struct nla_policy gate_policy[TCA_GATE_MAX +1 ] = { + [TCA_GATE_PARMS ] = { .len =sizeof (struct tc_gate ), + .type =NLA_EXACT_LEN }, + [TCA_GATE_PRIORITY ] = { .type =NLA_S32 }, + [TCA_GATE_ENTRY_LIST ] = { .type =NLA_NESTED }, + [TCA_GATE_BASE_TIME ] = { .type =NLA_U64 }, + [TCA_GATE_CYCLE_TIME ] = { .type =NLA_U64 }, + [TCA_GATE_CYCLE_TIME_EXT ] = { .type =NLA_U64 }, + [TCA_GATE_FLAGS ] = { .type =NLA_U32 }, + [TCA_GATE_CLOCKID ] = { .type =NLA_S32 }, +}; + +static int fill_gate_entry (struct nlattr **tb,struct tcfg_gate_entry *entry, +struct netlink_ext_ack *extack) +{ +u32 interval =0 ; + + entry->gate_state =nla_get_flag (tb[TCA_GATE_ENTRY_GATE ]); + +if (tb[TCA_GATE_ENTRY_INTERVAL ]) + interval =nla_get_u32 (tb[TCA_GATE_ENTRY_INTERVAL ]); + +if (interval ==0 ) { +NL_SET_ERR_MSG (extack,"Invalid interval for schedule entry" ); +return -EINVAL ; + } + + entry->interval = interval; + +if (tb[TCA_GATE_ENTRY_IPV ]) + entry->ipv =nla_get_s32 (tb[TCA_GATE_ENTRY_IPV ]); +else + entry->ipv =-1 ; + +if (tb[TCA_GATE_ENTRY_MAX_OCTETS ]) + entry->maxoctets =nla_get_s32 (tb[TCA_GATE_ENTRY_MAX_OCTETS ]); +else + entry->maxoctets =-1 ; + +return 0 ; +} + +static int parse_gate_entry (struct nlattr *n,struct tcfg_gate_entry *entry, +int index,struct netlink_ext_ack *extack) +{ +struct nlattr *tb[TCA_GATE_ENTRY_MAX +1 ] = { }; +int err; + err =nla_parse_nested (tb,TCA_GATE_ENTRY_MAX , n, entry_policy, extack); +if (err <0 ) { +NL_SET_ERR_MSG (extack,"Could not parse nested entry" ); +return -EINVAL ; + } + + entry->index = index; + +return fill_gate_entry (tb, entry, extack); +} + +static int parse_gate_list (struct nlattr *list_attr, +struct tcf_gate_params *sched, +struct netlink_ext_ack *extack) +{ +struct tcfg_gate_entry *entry, *e; +struct nlattr *n; +int err, rem; +int i =0 ; + +if (!list_attr) +return -EINVAL ; + +nla_for_each_nested (n, list_attr, rem) { +if (nla_type (n) !=TCA_GATE_ONE_ENTRY ) { +NL_SET_ERR_MSG (extack,"Attribute isn't type 'entry'" ); +continue ; + } + + entry =kzalloc (sizeof (*entry),GFP_KERNEL ); +if (!entry) { +NL_SET_ERR_MSG (extack,"Not enough memory for entry" ); + err = -ENOMEM ; + goto release_list; + } + + err =parse_gate_entry (n, entry, i, extack); +if (err <0 ) { +kfree (entry); + goto release_list; + } + +list_add_tail (&entry->list, &sched->entries); + i++; + } + + sched->num_entries = i; + +return i; + +release_list: +list_for_each_entry_safe (entry, e, &sched->entries, list) { +list_del (&entry->list); +kfree (entry); + } + +return err; +} + +static int tcf_gate_init (struct net *net,struct nlattr *nla, +struct nlattr *est,struct tc_action **a, +int ovr,int bind,bool rtnl_held, +struct tcf_proto *tp,u32 flags, +struct netlink_ext_ack *extack) +{ +struct tc_action_net *tn =net_generic (net, gate_net_id); +enum tk_offsets tk_offset =TK_OFFS_TAI ; +struct nlattr *tb[TCA_GATE_MAX +1 ]; +struct tcf_chain *goto_ch =NULL ; +struct tcfg_gate_entry *next; +struct tcf_gate_params *p; +struct gate_action *gact; +s32 clockid =CLOCK_TAI ; +struct tc_gate *parm; +struct tcf_gate *g; +int ret =0 , err; +u64 basetime =0 ; +u32 gflags =0 ; +s32 prio =-1 ; +ktime_t start; +u32 index; + +if (!nla) +return -EINVAL ; + + err =nla_parse_nested (tb,TCA_GATE_MAX , nla, gate_policy, extack); +if (err <0 ) +return err; + +if (!tb[TCA_GATE_PARMS ]) +return -EINVAL ; + parm =nla_data (tb[TCA_GATE_PARMS ]); + index = parm->index; + err =tcf_idr_check_alloc (tn, &index, a, bind); +if (err <0 ) +return err; + +if (err && bind) +return 0 ; + +if (!err) { + ret =tcf_idr_create_from_flags (tn, index, est, a, + &act_gate_ops, bind, flags); +if (ret) { +tcf_idr_cleanup (tn, index); +return ret; + } + + ret =ACT_P_CREATED ; + }else if (!ovr) { +tcf_idr_release (*a, bind); +return -EEXIST ; + } + +if (tb[TCA_GATE_PRIORITY ]) + prio =nla_get_s32 (tb[TCA_GATE_PRIORITY ]); + +if (tb[TCA_GATE_BASE_TIME ]) + basetime =nla_get_u64 (tb[TCA_GATE_BASE_TIME ]); + +if (tb[TCA_GATE_FLAGS ]) + gflags =nla_get_u32 (tb[TCA_GATE_FLAGS ]); + +if (tb[TCA_GATE_CLOCKID ]) { + clockid =nla_get_s32 (tb[TCA_GATE_CLOCKID ]); +switch (clockid) { +case CLOCK_REALTIME : + tk_offset =TK_OFFS_REAL ; +break ; +case CLOCK_MONOTONIC : + tk_offset =TK_OFFS_MAX ; +break ; +case CLOCK_BOOTTIME : + tk_offset =TK_OFFS_BOOT ; +break ; +case CLOCK_TAI : + tk_offset =TK_OFFS_TAI ; +break ; +default : +NL_SET_ERR_MSG (extack,"Invalid 'clockid'" ); + goto release_idr; + } + } + + err =tcf_action_check_ctrlact (parm->action, tp, &goto_ch, extack); +if (err <0 ) + goto release_idr; + + g =to_gate (*a); + + gact =kzalloc (sizeof (*gact),GFP_KERNEL ); +if (!gact) { + err = -ENOMEM ; + goto put_chain; + } + + p =get_gate_param (gact); + +INIT_LIST_HEAD (&p->entries); +if (tb[TCA_GATE_ENTRY_LIST ]) { + err =parse_gate_list (tb[TCA_GATE_ENTRY_LIST ], p, extack); +if (err <0 ) + goto release_mem; + } + +if (tb[TCA_GATE_CYCLE_TIME ]) { + p->tcfg_cycletime =nla_get_u64 (tb[TCA_GATE_CYCLE_TIME ]); + }else { +struct tcfg_gate_entry *entry; +ktime_t cycle =0 ; + +list_for_each_entry (entry, &p->entries, list) + cycle =ktime_add_ns (cycle, entry->interval); + p->tcfg_cycletime = cycle; + } + +if (tb[TCA_GATE_CYCLE_TIME_EXT ]) + p->tcfg_cycletime_ext = +nla_get_u64 (tb[TCA_GATE_CYCLE_TIME_EXT ]); + + p->tcfg_priority = prio; + p->tcfg_basetime = basetime; + p->tcfg_clockid = clockid; + p->tcfg_flags = gflags; + + gact->tk_offset = tk_offset; +spin_lock_init (&gact->entry_lock); +hrtimer_init (&gact->hitimer, clockid,HRTIMER_MODE_ABS ); + gact->hitimer.function = gate_timer_func; + + err =gate_get_start_time (gact, &start); +if (err <0 ) { +NL_SET_ERR_MSG (extack, +"Internal error: failed get start time" ); + goto release_mem; + } + + gact->current_close_time = start; + gact->current_gate_status =GATE_ACT_GATE_OPEN |GATE_ACT_PENDING ; + + next = list_first_entry(&p->entries, struct tcfg_gate_entry, list); +rcu_assign_pointer (gact->next_entry, next); + +gate_start_timer (gact, start); +
net/sched/act_gate.c line 429
Quoted 10 lines
net/sched/act_gate.c new file +10
-0+spin_lock_bh (&g->tcf_lock); + goto_ch =tcf_action_set_ctrlact (*a, parm->action, goto_ch); + gact =rcu_replace_pointer (g->actg, gact, +lockdep_is_held (&g->tcf_lock)); +spin_unlock_bh (&g->tcf_lock); + +if (goto_ch) +tcf_chain_put_by_act (goto_ch); +if (gact) +kfree_rcu (gact, rcu);
This leaks entries. For example, tunnel key action implements tunnel_key_release_params() helper that is used by both init and release code. I guess that would be the best approach here as well.
Quoted 218 lines
net/sched/act_gate.c new file +218
-0+ +if (ret ==ACT_P_CREATED ) +tcf_idr_insert (tn, *a); +return ret; + +release_mem: +kfree (gact); +put_chain: +if (goto_ch) +tcf_chain_put_by_act (goto_ch); +release_idr: +tcf_idr_release (*a, bind); +return err; +} + +static void tcf_gate_cleanup (struct tc_action *a) +{ +struct tcf_gate *g =to_gate (a); +struct tcfg_gate_entry *entry, *n; +struct tcf_gate_params *p; +struct gate_action *gact; + +spin_lock_bh (&g->tcf_lock); + gact =rcu_dereference_protected (g->actg, +lockdep_is_held (&g->tcf_lock)); +hrtimer_cancel (&gact->hitimer); + + p =get_gate_param (gact); +list_for_each_entry_safe (entry, n, &p->entries, list) { +list_del (&entry->list); +kfree (entry); + } +spin_unlock_bh (&g->tcf_lock); + +kfree_rcu (gact, rcu); +} + +static int dumping_entry (struct sk_buff *skb, +struct tcfg_gate_entry *entry) +{ +struct nlattr *item; + + item =nla_nest_start_noflag (skb,TCA_GATE_ONE_ENTRY ); +if (!item) +return -ENOSPC ; + +if (nla_put_u32 (skb,TCA_GATE_ENTRY_INDEX , entry->index)) + goto nla_put_failure; + +if (entry->gate_state &&nla_put_flag (skb,TCA_GATE_ENTRY_GATE )) + goto nla_put_failure; + +if (nla_put_u32 (skb,TCA_GATE_ENTRY_INTERVAL , entry->interval)) + goto nla_put_failure; + +if (nla_put_s32 (skb,TCA_GATE_ENTRY_MAX_OCTETS , entry->maxoctets)) + goto nla_put_failure; + +if (nla_put_s32 (skb,TCA_GATE_ENTRY_IPV , entry->ipv)) + goto nla_put_failure; + +return nla_nest_end (skb, item); + +nla_put_failure: +nla_nest_cancel (skb, item); +return -1 ; +} + +static int tcf_gate_dump (struct sk_buff *skb,struct tc_action *a, +int bind,int ref) +{ +unsigned char *b =skb_tail_pointer (skb); +struct tcf_gate *g =to_gate (a); +struct tc_gate opt = { + .index = g->tcf_index, + .refcnt =refcount_read (&g->tcf_refcnt) - ref, + .bindcnt =atomic_read (&g->tcf_bindcnt) - bind, + }; +struct tcfg_gate_entry *entry; +struct gate_action *gact; +struct tcf_gate_params *p; +struct nlattr *entry_list; +struct tcf_t t; + +spin_lock_bh (&g->tcf_lock); + opt.action = g->tcf_action; + gact =rcu_dereference_protected (g->actg, +lockdep_is_held (&g->tcf_lock)); + + p =get_gate_param (gact); + +if (nla_put (skb,TCA_GATE_PARMS ,sizeof (opt), &opt)) + goto nla_put_failure; + +if (nla_put_u64_64bit (skb,TCA_GATE_BASE_TIME , + p->tcfg_basetime,TCA_GATE_PAD )) + goto nla_put_failure; + +if (nla_put_u64_64bit (skb,TCA_GATE_CYCLE_TIME , + p->tcfg_cycletime,TCA_GATE_PAD )) + goto nla_put_failure; + +if (nla_put_u64_64bit (skb,TCA_GATE_CYCLE_TIME_EXT , + p->tcfg_cycletime_ext,TCA_GATE_PAD )) + goto nla_put_failure; + +if (nla_put_s32 (skb,TCA_GATE_CLOCKID , p->tcfg_clockid)) + goto nla_put_failure; + +if (nla_put_u32 (skb,TCA_GATE_FLAGS , p->tcfg_flags)) + goto nla_put_failure; + +if (nla_put_s32 (skb,TCA_GATE_PRIORITY , p->tcfg_priority)) + goto nla_put_failure; + + entry_list =nla_nest_start_noflag (skb,TCA_GATE_ENTRY_LIST ); +if (!entry_list) + goto nla_put_failure; + +list_for_each_entry (entry, &p->entries, list) { +if (dumping_entry (skb, entry) <0 ) + goto nla_put_failure; + } + +nla_nest_end (skb, entry_list); + +tcf_tm_dump (&t, &g->tcf_tm); +if (nla_put_64bit (skb,TCA_GATE_TM ,sizeof (t), &t,TCA_GATE_PAD )) + goto nla_put_failure; +spin_unlock_bh (&g->tcf_lock); + +return skb->len; + +nla_put_failure: +spin_unlock_bh (&g->tcf_lock); +nlmsg_trim (skb, b); +return -1 ; +} + +static int tcf_gate_walker (struct net *net,struct sk_buff *skb, +struct netlink_callback *cb,int type, +const struct tc_action_ops *ops, +struct netlink_ext_ack *extack) +{ +struct tc_action_net *tn =net_generic (net, gate_net_id); + +return tcf_generic_walker (tn, skb, cb, type, ops, extack); +} + +static void tcf_gate_stats_update (struct tc_action *a,u64 bytes,u32 packets, +u64 lastuse,bool hw) +{ +struct tcf_gate *g =to_gate (a); +struct tcf_t *tm = &g->tcf_tm; + +tcf_action_update_stats (a, bytes, packets, false, hw); + tm->lastuse =max_t (u64, tm->lastuse, lastuse); +} + +static int tcf_gate_search (struct net *net,struct tc_action **a,u32 index) +{ +struct tc_action_net *tn =net_generic (net, gate_net_id); + +return tcf_idr_search (tn, a, index); +} + +static size_t tcf_gate_get_fill_size (const struct tc_action *act) +{ +return nla_total_size (sizeof (struct tc_gate )); +} + +static struct tc_action_ops act_gate_ops = { + .kind ="gate" , + .id =TCA_ID_GATE , + .owner =THIS_MODULE , + .act = tcf_gate_act, + .dump = tcf_gate_dump, + .init = tcf_gate_init, + .cleanup = tcf_gate_cleanup, + .walk = tcf_gate_walker, + .stats_update = tcf_gate_stats_update, + .get_fill_size = tcf_gate_get_fill_size, + .lookup = tcf_gate_search, + .size =sizeof (struct gate_action ), +}; + +static __net_init int gate_init_net (struct net *net) +{ +struct tc_action_net *tn =net_generic (net, gate_net_id); + +return tc_action_net_init (net, tn, &act_gate_ops); +} + +static void __net_exit gate_exit_net(struct list_head *net_list) +{ +tc_action_net_exit (net_list, gate_net_id); +} + +static struct pernet_operations gate_net_ops = { + .init = gate_init_net, + .exit_batch = gate_exit_net, + .id = &gate_net_id, + .size =sizeof (struct tc_action_net ), +}; + +static int __init gate_init_module (void ) +{ +return tcf_register_action (&act_gate_ops, &gate_net_ops); +} + +static void __exit gate_cleanup_module(void) +{ +tcf_unregister_action (&act_gate_ops, &gate_net_ops); +} + +module_init (gate_init_module); +module_exit (gate_cleanup_module); +MODULE_LICENSE ("GPL v2" );
RE: [EXT] Re: [v3,net-next 1/4] net: qos: introduce a gate control flow action
Hi Vlad Buslov,
Quoted 310 lines
-----Original Message-----
From: Vlad Buslov <[email protected]>
Sent: 2020年4月22日 21:23
To: Po Liu <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; Claudiu Manoil <[email protected]>; Vladimir Oltean <[email protected]>; Alexandru Marginean <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]
Subject: [EXT] Re: [v3,net-next 1/4] net: qos: introduce a gate control flow actionCaution: EXT Email
Hi Po,
Quoted 39 lines · Po Liu
Introduce a ingress frame gate control flow action.
Tc gate action does the work like this:
Assume there is a gate allow specified ingress frames can be passed at specific time slot, and be dropped at specific time slot. Tc filter chooses the ingress frames, and tc gate action would specify what slot does these frames can be passed to device and what time slot would be dropped.
Tc gate action would provide an entry list to tell how much time gate keep open and how much time gate keep state close. Gate action also assign a start time to tell when the entry list start. Then driver would repeat the gate entry list cyclically.
For the software simulation, gate action requires the user assign a time clock type.Below is the setting example in user space. Tc filter a stream source ip address is 192.168.0.20 and gate action own two time slots. One is last 200ms gate open let frame pass another is last 100ms gate close let frames dropped. When the frames have passed total frames over 8000000 bytes, frames will be dropped in one 200000000ns time slot.
Quoted 3 lines
tc qdisc add dev eth0 ingress
tc filter add dev eth0 parent ffff: protocol ip \
flower src_ip 192.168.0.20 \ action gate index 2 clockid CLOCK_TAI \ sched-entry open 200000000 -1 8000000 \ sched-entry close 100000000 -1 -1Quoted 1 line
tc chain del dev eth0 ingress chain 0
"sched-entry" follow the name taprio style. Gate state is "open"/"close". Follow with period nanosecond. Then next item is internal priority value means which ingress queue should put. "-1" means wildcard. The last value optional specifies the maximum number of MSDU octets that are permitted to pass the gate during the specified time interval.
Base-time is not set will be 0 as default, as result start time would be ((N + 1) * cycletime) which is the minimal of future time.Quoted 3 lines
Below example shows filtering a stream with destination mac address is 10:00:80:00:00:00 and ip type is ICMP, follow the action gate. The gate action would run with one close time slot which means always keep
close.
Quoted 87 lines
The time cycle is total 200000000ns. The base-time would calculate by:
1357000000000 + (N + 1) * cycletime
When the total value is the future time, it will be the start time. The cycletime here would be 200000000ns for this case.
Quoted 1 line
tc filter add dev eth0 parent ffff: protocol ip \
flower skip_hw ip_proto icmp dst_mac 10:00:80:00:00:00 \ action gate index 12 base-time 1357000000000 \ sched-entry close 200000000 -1 -1 \ clockid CLOCK_TAISigned-off-by Po Liu <[email protected]>--- include/net/tc_act/tc_gate.h | 54 +++ include/uapi/linux/pkt_cls.h | 1 + include/uapi/linux/tc_act/tc_gate.h | 47 ++ net/sched/Kconfig | 13 + net/sched/Makefile | 1 + net/sched/act_gate.c | 647 ++++++++++++++++++++++++++++ 6 files changed, 763 insertions(+) create mode 100644 include/net/tc_act/tc_gate.h create mode 100644 include/uapi/linux/tc_act/tc_gate.h create mode 100644 net/sched/act_gate.c(quoted diff) +0
-0
b/include/net/tc_act/tc_gate.hnew file mode 100644 index 000000000000..b0ace55b2aaainclude/net/tc_act/tc_gate.h new file +54
-0@@ -0,0 +1,54 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +/* Copyright 2020 NXP */ + +#ifndef __NET_TC_GATE_H +#define __NET_TC_GATE_H + +#include <net/act_api.h> +#include <linux/tc_act/tc_gate.h> + +struct tcfg_gate_entry { +int index; +u8 gate_state; +u32 interval; +s32 ipv; +s32 maxoctets; +struct list_head list; +}; + +struct tcf_gate_params { +s32 tcfg_priority; +u64 tcfg_basetime; +u64 tcfg_cycletime; +u64 tcfg_cycletime_ext; +u32 tcfg_flags; +s32 tcfg_clockid; +size_t num_entries; +struct list_head entries; +}; + +#define GATE_ACT_GATE_OPEN BIT(0) +#define GATE_ACT_PENDING BIT(1) +struct gate_action { +struct tcf_gate_params param; +spinlock_t entry_lock; +u8 current_gate_status; +ktime_t current_close_time; +u32 current_entry_octets; +s32 current_max_octets; +struct tcfg_gate_entry __rcu *next_entry; +struct hrtimer hitimer; +enum tk_offsets tk_offset; +struct rcu_head rcu; +}; + +struct tcf_gate { +struct tc_action common; +struct gate_action __rcu *actg; +}; +#define to_gate (a) ((struct tcf_gate *)a) + +#define get_gate_param (act) ((struct tcf_gate_params *)act) #define +get_gate_action(p) ((struct gate_action *)p) + +#endif (quoted diff) +0
-0Quoted 1 line
b/include/uapi/linux/pkt_cls.hindex9f06d29cab70..fc672b232437100644
Quoted 36 lines
include/uapi/linux/pkt_cls.h +1
-0@@ -134,6 +134,7 @@ enum tca_id {TCA_ID_CTINFO ,TCA_ID_MPLS ,TCA_ID_CT , +TCA_ID_GATE ,/* other actions go here */ __TCA_ID_MAX = 255 };include/uapi/linux/tc_act/tc_gate.h new file +19
-0@@ -0,0 +1,47 @@ +/* SPDX-License-Identifier: GPL-2.0+ WITH Linux-syscall-note */ +/* Copyright 2020 NXP */ + +#ifndef __LINUX_TC_GATE_H +#define __LINUX_TC_GATE_H + +#include <linux/pkt_cls.h> + +struct tc_gate { +tc_gen ; +}; + +enum { +TCA_GATE_ENTRY_UNSPEC , +TCA_GATE_ENTRY_INDEX , +TCA_GATE_ENTRY_GATE , +TCA_GATE_ENTRY_INTERVAL , +TCA_GATE_ENTRY_IPV , + TCA_GATE_ENTRY_MAX_OCTETS,include/uapi/linux/tc_act/tc_gate.h line 29
Quoted 10 lines
include/uapi/linux/tc_act/tc_gate.h new file +10
-0+ __TCA_GATE_ENTRY_MAX, +}; +#define TCA_GATE_ENTRY_MAX (__TCA_GATE_ENTRY_MAX - 1) + +enum { +TCA_GATE_ONE_ENTRY_UNSPEC , +TCA_GATE_ONE_ENTRY , + __TCA_GATE_ONE_ENTRY_MAX, +}; +#define TCA_GATE_ONE_ENTRY_MAX (__TCA_GATE_ONE_ENTRY_MAX- 1)
net/sched/Makefile obj-$(CONFIG_NET_IFE_SKBPRIO) +=
Quoted 47 lines
include/uapi/linux/tc_act/tc_gate.h new file +18
-0+ +enum { +TCA_GATE_UNSPEC , +TCA_GATE_TM , +TCA_GATE_PARMS , +TCA_GATE_PAD , +TCA_GATE_PRIORITY , +TCA_GATE_ENTRY_LIST , +TCA_GATE_BASE_TIME , +TCA_GATE_CYCLE_TIME , +TCA_GATE_CYCLE_TIME_EXT , +TCA_GATE_FLAGS , +TCA_GATE_CLOCKID , + __TCA_GATE_MAX, +}; +#define TCA_GATE_MAX (__TCA_GATE_MAX - 1) + +#endif net/sched/Kconfig index renamed from net/sched/Kconfig +0
-0
bfbefb7bff9d..1314549c7567100644net/sched/Kconfig +13
-0@@ -981,6 +981,19 @@ config NET_ACT_CT To compile this code as a module, choose M here: the module will be called act_ct. +config NET_ACT_GATE + tristate "Frame gate entry list control tc action" + depends on NET_CLS_ACT + help + Say Y here to allow to control the ingress flow to be passed at + specific time slot and be dropped at other specific time slot by + the gate entry list. The manipulation will simulate the IEEE + 802.1Qci stream gate control behavior. + + If unsure, say N. + To compile this code as a module, choose M here: the + module will be called act_gate. + config NET_IFE_SKBMARK tristate "Support to encoding decoding skb mark on IFE action" depends on NET_ACT_IFEnet/sched/Makefile index renamed from net/sched/Makefile +0
-0
31c367a6cd09..66bbf9a98f9e100644net/sched/Makefile +0
-0@@ -30,6 +30,7 @@ obj-$(CONFIG_NET_IFE_SKBPRIO) +=act_meta_skbprio.o
Quoted 60 lines
net/sched/Makefile +1
-0obj-$(CONFIG_NET_IFE_SKBTCINDEX) += act_meta_skbtcindex.o obj-$(CONFIG_NET_ACT_TUNNEL_KEY)+= act_tunnel_key.o obj-$(CONFIG_NET_ACT_CT) += act_ct.o +obj-$(CONFIG_NET_ACT_GATE) += act_gate.o obj-$(CONFIG_NET_SCH_FIFO) += sch_fifo.o obj-$(CONFIG_NET_SCH_CBQ) += sch_cbq.o obj-$(CONFIG_NET_SCH_HTB) += sch_htb.onet/sched/act_gate.c new file mode renamed from net/sched/act_gate.c +0
-0100644 index 000000000000..e932f402b4f1
net/sched/act_gate.c new file +48
-0@@ -0,0 +1,647 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* Copyright 2020 NXP */ + +#include <linux/module.h> +#include <linux/types.h> +#include <linux/kernel.h> +#include <linux/string.h> +#include <linux/errno.h> +#include <linux/skbuff.h> +#include <linux/rtnetlink.h> +#include <linux/init.h> +#include <linux/slab.h> +#include <net/act_api.h> +#include <net/netlink.h> +#include <net/pkt_cls.h> +#include <net/tc_act/tc_gate.h> + +static unsigned int gate_net_id; +static struct tc_action_ops act_gate_ops; + +static ktime_t gate_get_time (struct gate_action *gact) { +ktime_t mono =ktime_get (); + +switch (gact->tk_offset) { +case TK_OFFS_MAX : +return mono; +default : +return ktime_mono_to_any (mono, gact->tk_offset); + } + +return KTIME_MAX ; +} + +static int gate_get_start_time (struct gate_action *gact,ktime_t +*start) { +struct tcf_gate_params *param =get_gate_param (gact); +ktime_t now, base, cycle; +u64 n; + + base =ns_to_ktime (param->tcfg_basetime); + now =gate_get_time (gact); + +if (ktime_after (base, now)) { + *start = base; +return 0 ; + } + + cycle = param->tcfg_cycletime;
Quoted 8 lines
net/sched/act_gate.c line 53
Quoted 4 lines
net/sched/act_gate.c new file +4
-0+ +/* cycle time should not be zero */ +if (WARN_ON (!cycle)) +return -EFAULT ;Looking at the init code it seems that this value can be set to 0 directly from netlink packet without further validation, which would allow user to trigger warning here.
Yes, will avoid at ahead point.
Quoted 66 lines
Quoted 24 lines
net/sched/act_gate.c new file +23
-0+ + n =div64_u64 (ktime_sub_ns (now, base), cycle); + *start =ktime_add_ns (base, (n +1 ) * cycle); +return 0 ; +} + +static void gate_start_timer (struct gate_action *gact,ktime_t start) +{ +ktime_t expires; + + expires =hrtimer_get_expires (&gact->hitimer); +if (expires ==0 ) + expires =KTIME_MAX ; + + start =min_t (ktime_t, start, expires); + +hrtimer_start (&gact->hitimer, start,HRTIMER_MODE_ABS ); } + +static enum hrtimer_restart gate_timer_func (struct hrtimer *timer) { + struct gate_action *gact = container_of(timer, struct gate_action, + hitimer); +struct tcf_gate_params *p =get_gate_param (gact); +struct tcfg_gate_entry *next;net/sched/act_gate.c
Quoted 10 lines
net/sched/act_gate.c new file +10
-0+ktime_t close_time, now; + +spin_lock (&gact->entry_lock); + + next =rcu_dereference_protected (gact->next_entry, + +lockdep_is_held (&gact->entry_lock)); + +/* cycle start, clear pending bit, clear total octets */ + gact->current_gate_status = next->gate_state ?
GATE_ACT_GATE_OPEN: 0;Quoted 18 lines
net/sched/act_gate.c new file +18
-0+ gact->current_entry_octets =0 ; + gact->current_max_octets = next->maxoctets; + + gact->current_close_time =ktime_add_ns (gact->current_close_time, + next->interval); + + close_time = gact->current_close_time; + +if (list_is_last (&next->list, &p->entries)) + next =list_first_entry (&p->entries, + struct tcfg_gate_entry, list); +else + next =list_next_entry (next, list); + + now =gate_get_time (gact); + +if (ktime_after (now, close_time)) { +ktime_t cycle, base;net/sched/act_gate.c line 115
Quoted 10 lines
net/sched/act_gate.c new file +10
-0+u64 n; + + cycle = p->tcfg_cycletime; + base =ns_to_ktime (p->tcfg_basetime); + n =div64_u64 (ktime_sub_ns (now, base), cycle); + close_time =ktime_add_ns (base, (n +1 ) * cycle); + } + +rcu_assign_pointer (gact->next_entry, next); +spin_unlock (&gact->entry_lock);I have couple of question about synchronization here:
Quoted 7 lines
- Why do you need
next_entryto be rcu pointer? It is only assigned here withentry_lockprotection and in init code before action is visible to concurrent users. I don't see any unlocked rcu-protected readers here that could benefit from it.- Why create dedicated
entry_lockinstead of using already existing per- actiontcf_lock?
Will try to use the tcf_lock for verification.
The thoughts came from that the timer period arrived then check through the list and then update next time would take much more time. Action function would be busy when traffic. So use a separate lock here for
Quoted 13 lines
Quoted 13 lines
net/sched/act_gate.c new file +12
-0+ +hrtimer_set_expires (&gact->hitimer, close_time); + +return HRTIMER_RESTART ; +} + +static int tcf_gate_act (struct sk_buff *skb,const struct tc_action *a, +struct tcf_result *res) { +struct tcf_gate *g =to_gate (a); +struct gate_action *gact; +int action; +
Quoted 8 lines
net/sched/act_gate.c line 133
Quoted 5 lines
net/sched/act_gate.c new file +5
-0+tcf_lastuse_update (&g->tcf_tm); +bstats_cpu_update (this_cpu_ptr (g->common.cpu_bstats), skb); + + action =READ_ONCE (g->tcf_action); +rcu_read_lock ();Action fastpath is already rcu read lock protected, you don't need to manually obtain it.
Will be removed.
Quoted 6 lines
net/sched/act_gate.c line 135
Quoted 3 lines
net/sched/act_gate.c new file +2
-0+ gact =rcu_dereference_bh (g->actg); +if (unlikely (gact->current_gate_status &GATE_ACT_PENDING )) {Can't
current_gate_statusbe concurrently modified by timer callback? This function doesn't useentry_lockto synchronize with timer.
Will try tcf_lock either.
Quoted 16 lines
net/sched/act_gate.c line 140
Quoted 6 lines
net/sched/act_gate.c new file +5
-0+rcu_read_unlock (); +return action; + } + +if (!(gact->current_gate_status &GATE_ACT_GATE_OPEN ))...and here
net/sched/act_gate.c
Quoted 6 lines
net/sched/act_gate.c new file +6
-0+ goto drop; + +if (gact->current_max_octets >=0 ) { + gact->current_entry_octets +=qdisc_pkt_len (skb); +if (gact->current_entry_octets > + gact->current_max_octets) {
Quoted 8 lines
here also.
net/sched/act_gate.c line 146
Quoted 2 lines
net/sched/act_gate.c new file +2
-0+ +qstats_overlimit_inc (this_cpu_ptr (g->common.cpu_qstats));Please use
tcf_action_inc_overlimit_qstats()and other wrappers for stats. Otherwise it will crash if user passesTCA_ACT_FLAGS_NO_PERCPU_STATSflag.
The tcf_action_inc_overlimit_qstats() can't show limit counts in tc show command. Is there anything need to do?
Quoted 280 lines
Quoted 4 lines
net/sched/act_gate.c new file +3
-0+ goto drop; + } + }net/sched/act_gate.c
Quoted 10 lines
net/sched/act_gate.c new file +10
-0+rcu_read_unlock (); + +return action; +drop: +rcu_read_unlock (); +qstats_drop_inc (this_cpu_ptr (g->common.cpu_qstats)); +return TC_ACT_SHOT ; +} + +static const struct nla_policy entry_policy[TCA_GATE_ENTRY_MAX +1 ]= {
Quoted 32 lines
net/sched/act_gate.c new file +32
-0+ [TCA_GATE_ENTRY_INDEX] = { .type = NLA_U32 }, + [TCA_GATE_ENTRY_GATE] = { .type = NLA_FLAG }, + [TCA_GATE_ENTRY_INTERVAL] = { .type = NLA_U32 }, + [TCA_GATE_ENTRY_IPV] = { .type = NLA_S32 }, + [TCA_GATE_ENTRY_MAX_OCTETS] = { .type = NLA_S32 }, +}; + +static const struct nla_policy gate_policy[TCA_GATE_MAX +1 ] = { + [TCA_GATE_PARMS ] = { .len =sizeof (struct tc_gate ), + .type =NLA_EXACT_LEN }, + [TCA_GATE_PRIORITY ] = { .type =NLA_S32 }, + [TCA_GATE_ENTRY_LIST ] = { .type =NLA_NESTED }, + [TCA_GATE_BASE_TIME ] = { .type =NLA_U64 }, + [TCA_GATE_CYCLE_TIME ] = { .type =NLA_U64 }, + [TCA_GATE_CYCLE_TIME_EXT ] = { .type =NLA_U64 }, + [TCA_GATE_FLAGS ] = { .type =NLA_U32 }, + [TCA_GATE_CLOCKID ] = { .type =NLA_S32 }, +}; + +static int fill_gate_entry (struct nlattr **tb,struct tcfg_gate_entry *entry, +struct netlink_ext_ack *extack) { +u32 interval =0 ; + + entry->gate_state =nla_get_flag (tb[TCA_GATE_ENTRY_GATE ]); + +if (tb[TCA_GATE_ENTRY_INTERVAL ]) + interval =nla_get_u32 (tb[TCA_GATE_ENTRY_INTERVAL ]); + +if (interval ==0 ) { +NL_SET_ERR_MSG (extack,"Invalid interval for schedule entry" ); +return -EINVAL ; + }net/sched/act_gate.c
Quoted 10 lines
net/sched/act_gate.c new file +10
-0+ + entry->interval = interval; + +if (tb[TCA_GATE_ENTRY_IPV ]) + entry->ipv =nla_get_s32 (tb[TCA_GATE_ENTRY_IPV ]); +else + entry->ipv =-1 ; + +if (tb[TCA_GATE_ENTRY_MAX_OCTETS ]) + entry->maxoctets =
nla_get_s32(tb[TCA_GATE_ENTRY_MAX_OCTETS]);net/sched/act_gate.c
Quoted 7 lines
net/sched/act_gate.c new file +7
-0+else + entry->maxoctets =-1 ; + +return 0 ; +} + +static int parse_gate_entry(struct nlattr *n, struct tcfg_gate_entry*entry,
net/sched/act_gate.c
Quoted 5 lines
net/sched/act_gate.c new file +4
-0+ int index, struct netlink_ext_ack *extack) { + struct nlattr *tb[TCA_GATE_ENTRY_MAX + 1] = { }; +int err; + err = nla_parse_nested(tb, TCA_GATE_ENTRY_MAX, n, entry_policy,extack);
Quoted 208 lines
net/sched/act_gate.c new file +208
-0+if (err <0 ) { +NL_SET_ERR_MSG (extack,"Could not parse nested entry" ); +return -EINVAL ; + } + + entry->index = index; + +return fill_gate_entry (tb, entry, extack); } + +static int parse_gate_list (struct nlattr *list_attr, +struct tcf_gate_params *sched, +struct netlink_ext_ack *extack) { +struct tcfg_gate_entry *entry, *e; +struct nlattr *n; +int err, rem; +int i =0 ; + +if (!list_attr) +return -EINVAL ; + +nla_for_each_nested (n, list_attr, rem) { +if (nla_type (n) !=TCA_GATE_ONE_ENTRY ) { +NL_SET_ERR_MSG (extack,"Attribute isn't type 'entry'" ); +continue ; + } + + entry =kzalloc (sizeof (*entry),GFP_KERNEL ); +if (!entry) { +NL_SET_ERR_MSG (extack,"Not enough memory for entry" ); + err = -ENOMEM ; + goto release_list; + } + + err =parse_gate_entry (n, entry, i, extack); +if (err <0 ) { +kfree (entry); + goto release_list; + } + +list_add_tail (&entry->list, &sched->entries); + i++; + } + + sched->num_entries = i; + +return i; + +release_list: +list_for_each_entry_safe (entry, e, &sched->entries, list) { +list_del (&entry->list); +kfree (entry); + } + +return err; +} + +static int tcf_gate_init (struct net *net,struct nlattr *nla, +struct nlattr *est,struct tc_action **a, +int ovr,int bind,bool rtnl_held, +struct tcf_proto *tp,u32 flags, +struct netlink_ext_ack *extack) { +struct tc_action_net *tn =net_generic (net, gate_net_id); +enum tk_offsets tk_offset =TK_OFFS_TAI ; +struct nlattr *tb[TCA_GATE_MAX +1 ]; +struct tcf_chain *goto_ch =NULL ; +struct tcfg_gate_entry *next; +struct tcf_gate_params *p; +struct gate_action *gact; +s32 clockid =CLOCK_TAI ; +struct tc_gate *parm; +struct tcf_gate *g; +int ret =0 , err; +u64 basetime =0 ; +u32 gflags =0 ; +s32 prio =-1 ; +ktime_t start; +u32 index; + +if (!nla) +return -EINVAL ; + + err =nla_parse_nested (tb,TCA_GATE_MAX , nla, gate_policy, extack); +if (err <0 ) +return err; + +if (!tb[TCA_GATE_PARMS ]) +return -EINVAL ; + parm =nla_data (tb[TCA_GATE_PARMS ]); + index = parm->index; + err =tcf_idr_check_alloc (tn, &index, a, bind); +if (err <0 ) +return err; + +if (err && bind) +return 0 ; + +if (!err) { + ret =tcf_idr_create_from_flags (tn, index, est, a, + &act_gate_ops, bind, flags); +if (ret) { +tcf_idr_cleanup (tn, index); +return ret; + } + + ret =ACT_P_CREATED ; + }else if (!ovr) { +tcf_idr_release (*a, bind); +return -EEXIST ; + } + +if (tb[TCA_GATE_PRIORITY ]) + prio =nla_get_s32 (tb[TCA_GATE_PRIORITY ]); + +if (tb[TCA_GATE_BASE_TIME ]) + basetime =nla_get_u64 (tb[TCA_GATE_BASE_TIME ]); + +if (tb[TCA_GATE_FLAGS ]) + gflags =nla_get_u32 (tb[TCA_GATE_FLAGS ]); + +if (tb[TCA_GATE_CLOCKID ]) { + clockid =nla_get_s32 (tb[TCA_GATE_CLOCKID ]); +switch (clockid) { +case CLOCK_REALTIME : + tk_offset =TK_OFFS_REAL ; +break ; +case CLOCK_MONOTONIC : + tk_offset =TK_OFFS_MAX ; +break ; +case CLOCK_BOOTTIME : + tk_offset =TK_OFFS_BOOT ; +break ; +case CLOCK_TAI : + tk_offset =TK_OFFS_TAI ; +break ; +default : +NL_SET_ERR_MSG (extack,"Invalid 'clockid'" ); + goto release_idr; + } + } + + err =tcf_action_check_ctrlact (parm->action, tp, &goto_ch, extack); +if (err <0 ) + goto release_idr; + + g =to_gate (*a); + + gact =kzalloc (sizeof (*gact),GFP_KERNEL ); +if (!gact) { + err = -ENOMEM ; + goto put_chain; + } + + p =get_gate_param (gact); + +INIT_LIST_HEAD (&p->entries); +if (tb[TCA_GATE_ENTRY_LIST ]) { + err =parse_gate_list (tb[TCA_GATE_ENTRY_LIST ], p, extack); +if (err <0 ) + goto release_mem; + } + +if (tb[TCA_GATE_CYCLE_TIME ]) { + p->tcfg_cycletime =nla_get_u64 (tb[TCA_GATE_CYCLE_TIME ]); + }else { +struct tcfg_gate_entry *entry; +ktime_t cycle =0 ; + +list_for_each_entry (entry, &p->entries, list) + cycle =ktime_add_ns (cycle, entry->interval); + p->tcfg_cycletime = cycle; + } + +if (tb[TCA_GATE_CYCLE_TIME_EXT ]) + p->tcfg_cycletime_ext = +nla_get_u64 (tb[TCA_GATE_CYCLE_TIME_EXT ]); + + p->tcfg_priority = prio; + p->tcfg_basetime = basetime; + p->tcfg_clockid = clockid; + p->tcfg_flags = gflags; + + gact->tk_offset = tk_offset; +spin_lock_init (&gact->entry_lock); +hrtimer_init (&gact->hitimer, clockid,HRTIMER_MODE_ABS ); + gact->hitimer.function = gate_timer_func; + + err =gate_get_start_time (gact, &start); +if (err <0 ) { +NL_SET_ERR_MSG (extack, +"Internal error: failed get start time" ); + goto release_mem; + } + + gact->current_close_time = start; + gact->current_gate_status =GATE_ACT_GATE_OPEN | +GATE_ACT_PENDING ; + + next = list_first_entry(&p->entries, struct tcfg_gate_entry, list); +rcu_assign_pointer (gact->next_entry, next); + +gate_start_timer (gact, start); + +spin_lock_bh (&g->tcf_lock); + goto_ch =tcf_action_set_ctrlact (*a, parm->action, goto_ch); + gact =rcu_replace_pointer (g->actg, gact, +lockdep_is_held (&g->tcf_lock)); +spin_unlock_bh (&g->tcf_lock); +
Quoted 9 lines
net/sched/act_gate.c line 429
Quoted 4 lines
net/sched/act_gate.c new file +4
-0+if (goto_ch) +tcf_chain_put_by_act (goto_ch); +if (gact) +kfree_rcu (gact, rcu);This leaks entries. For example, tunnel key action implements
tunnel_key_release_params()helper that is used by both init and release code. I guess that would be the best approach here as well.
Will refer the tunnel action. Thanks for pointing out.
Quoted 193 lines
Quoted 136 lines
net/sched/act_gate.c new file +54
-0+ +if (ret ==ACT_P_CREATED ) +tcf_idr_insert (tn, *a); +return ret; + +release_mem: +kfree (gact); +put_chain: +if (goto_ch) +tcf_chain_put_by_act (goto_ch); +release_idr: +tcf_idr_release (*a, bind); +return err; +} + +static void tcf_gate_cleanup (struct tc_action *a) { +struct tcf_gate *g =to_gate (a); +struct tcfg_gate_entry *entry, *n; +struct tcf_gate_params *p; +struct gate_action *gact; + +spin_lock_bh (&g->tcf_lock); + gact =rcu_dereference_protected (g->actg, +lockdep_is_held (&g->tcf_lock)); +hrtimer_cancel (&gact->hitimer); + + p =get_gate_param (gact); +list_for_each_entry_safe (entry, n, &p->entries, list) { +list_del (&entry->list); +kfree (entry); + } +spin_unlock_bh (&g->tcf_lock); + +kfree_rcu (gact, rcu); +} + +static int dumping_entry (struct sk_buff *skb, +struct tcfg_gate_entry *entry) { +struct nlattr *item; + + item =nla_nest_start_noflag (skb,TCA_GATE_ONE_ENTRY ); +if (!item) +return -ENOSPC ; + +if (nla_put_u32 (skb,TCA_GATE_ENTRY_INDEX , entry->index)) + goto nla_put_failure; + +if (entry->gate_state &&nla_put_flag (skb,TCA_GATE_ENTRY_GATE )) + goto nla_put_failure; + +if (nla_put_u32 (skb,TCA_GATE_ENTRY_INTERVAL , entry->interval)) + goto nla_put_failure; + + if (nla_put_s32(skb, TCA_GATE_ENTRY_MAX_OCTETS, entry-maxoctets))
net/sched/act_gate.c new file +81
-0+ goto nla_put_failure; + +if (nla_put_s32 (skb,TCA_GATE_ENTRY_IPV , entry->ipv)) + goto nla_put_failure; + +return nla_nest_end (skb, item); + +nla_put_failure: +nla_nest_cancel (skb, item); +return -1 ; +} + +static int tcf_gate_dump (struct sk_buff *skb,struct tc_action *a, +int bind,int ref) { +unsigned char *b =skb_tail_pointer (skb); +struct tcf_gate *g =to_gate (a); +struct tc_gate opt = { + .index = g->tcf_index, + .refcnt =refcount_read (&g->tcf_refcnt) - ref, + .bindcnt =atomic_read (&g->tcf_bindcnt) - bind, + }; +struct tcfg_gate_entry *entry; +struct gate_action *gact; +struct tcf_gate_params *p; +struct nlattr *entry_list; +struct tcf_t t; + +spin_lock_bh (&g->tcf_lock); + opt.action = g->tcf_action; + gact =rcu_dereference_protected (g->actg, +lockdep_is_held (&g->tcf_lock)); + + p =get_gate_param (gact); + +if (nla_put (skb,TCA_GATE_PARMS ,sizeof (opt), &opt)) + goto nla_put_failure; + +if (nla_put_u64_64bit (skb,TCA_GATE_BASE_TIME , + p->tcfg_basetime,TCA_GATE_PAD )) + goto nla_put_failure; + +if (nla_put_u64_64bit (skb,TCA_GATE_CYCLE_TIME , + p->tcfg_cycletime,TCA_GATE_PAD )) + goto nla_put_failure; + +if (nla_put_u64_64bit (skb,TCA_GATE_CYCLE_TIME_EXT , + p->tcfg_cycletime_ext,TCA_GATE_PAD )) + goto nla_put_failure; + +if (nla_put_s32 (skb,TCA_GATE_CLOCKID , p->tcfg_clockid)) + goto nla_put_failure; + +if (nla_put_u32 (skb,TCA_GATE_FLAGS , p->tcfg_flags)) + goto nla_put_failure; + +if (nla_put_s32 (skb,TCA_GATE_PRIORITY , p->tcfg_priority)) + goto nla_put_failure; + + entry_list =nla_nest_start_noflag (skb,TCA_GATE_ENTRY_LIST ); +if (!entry_list) + goto nla_put_failure; + +list_for_each_entry (entry, &p->entries, list) { +if (dumping_entry (skb, entry) <0 ) + goto nla_put_failure; + } + +nla_nest_end (skb, entry_list); + +tcf_tm_dump (&t, &g->tcf_tm); +if (nla_put_64bit (skb,TCA_GATE_TM ,sizeof (t), &t,TCA_GATE_PAD )) + goto nla_put_failure; +spin_unlock_bh (&g->tcf_lock); + +return skb->len; + +nla_put_failure: +spin_unlock_bh (&g->tcf_lock); +nlmsg_trim (skb, b); +return -1 ; +}net/sched/act_gate.c
Quoted 10 lines
net/sched/act_gate.c new file +10
-0+ +static int tcf_gate_walker (struct net *net,struct sk_buff *skb, +struct netlink_callback *cb,int type, +const struct tc_action_ops *ops, +struct netlink_ext_ack *extack) { +struct tc_action_net *tn =net_generic (net, gate_net_id); + +return tcf_generic_walker (tn, skb, cb, type, ops, extack); } + +static void tcf_gate_stats_update(struct tc_action *a, u64 bytes, u32packets,
Quoted 46 lines
net/sched/act_gate.c new file +46
-0+ u64 lastuse, bool hw) { +struct tcf_gate *g =to_gate (a); +struct tcf_t *tm = &g->tcf_tm; + +tcf_action_update_stats (a, bytes, packets, false, hw); + tm->lastuse =max_t (u64, tm->lastuse, lastuse); } + +static int tcf_gate_search (struct net *net,struct tc_action **a,u32 +index) { +struct tc_action_net *tn =net_generic (net, gate_net_id); + +return tcf_idr_search (tn, a, index); } + +static size_t tcf_gate_get_fill_size (const struct tc_action *act) { +return nla_total_size (sizeof (struct tc_gate )); } + +static struct tc_action_ops act_gate_ops = { + .kind ="gate" , + .id =TCA_ID_GATE , + .owner =THIS_MODULE , + .act = tcf_gate_act, + .dump = tcf_gate_dump, + .init = tcf_gate_init, + .cleanup = tcf_gate_cleanup, + .walk = tcf_gate_walker, + .stats_update = tcf_gate_stats_update, + .get_fill_size = tcf_gate_get_fill_size, + .lookup = tcf_gate_search, + .size =sizeof (struct gate_action ), +}; + +static __net_init int gate_init_net (struct net *net) { +struct tc_action_net *tn =net_generic (net, gate_net_id); + +return tc_action_net_init (net, tn, &act_gate_ops); } + +static void __net_exit gate_exit_net(struct list_head *net_list) { +tc_action_net_exit (net_list, gate_net_id); } + +static struct pernet_operations gate_net_ops = { + .init = gate_init_net, + .exit_batch = gate_exit_net, + .id = &gate_net_id, + .size =sizeof (struct tc_action_net ), }; + +static int __init gate_init_module (void ) {
Quoted 8 lines
net/sched/act_gate.c line 647
Quoted 8 lines
net/sched/act_gate.c new file +8
-0+return tcf_register_action (&act_gate_ops, &gate_net_ops); } + +static void __exit gate_cleanup_module(void) { +tcf_unregister_action (&act_gate_ops, &gate_net_ops); } + +module_init (gate_init_module); +module_exit (gate_cleanup_module); +MODULE_LICENSE ("GPL v2" );
Thanks a lot.
Br,
Po Liu