net: flow_offload: skip hw stats check for FLOW_ACTION_HW_STATS_DONT_CARE
Subsystems (5)
Pablo Neira Ayuso
maintainerPATCH v4 1/1 net: flow_offload: skip hw stats check for FLOW_ACTION_HW_STATS_DONT_CARE
This patch adds FLOW_ACTION_HW_STATS_DONT_CARE which tells the driver that the frontend does not need counters, this hw stats type request never fails. The FLOW_ACTION_HW_STATS_DISABLED type explicitly requests the driver to disable the stats, however, if the driver cannot disable counters, it bails out.
TCA_ACT_HW_STATS_* maintains the 1:1 mapping with FLOW_ACTION_HW_STATS_* except by disabled which is mapped to FLOW_ACTION_HW_STATS_DISABLED (this is 0 in tc). Add tc_act_hw_stats() to perform the mapping between TCA_ACT_HW_STATS_* and FLOW_ACTION_HW_STATS_*.
---
v4: Update mlxsw as Jiri prefers.
.../net/ethernet/mellanox/mlxsw/spectrum_flower.c | 3 ++- include/net/flow_offload.h | 9 ++++++++- net/sched/cls_api.c | 14 ++++++++++++-- 3 files changed, 22 insertions(+), 4 deletions(-)
drivers/net/ethernet/mellanox/mlxsw/spectrum_flower.c +2 -1
@@ -36,7 +36,8 @@ static int mlxsw_sp_flower_parse_actions(struct mlxsw_sp *mlxsw_sp, err =mlxsw_sp_acl_rulei_act_count (mlxsw_sp, rulei, extack);if (err)return err; - }else if (act->hw_stats !=FLOW_ACTION_HW_STATS_DISABLED ) { + }else if (act->hw_stats !=FLOW_ACTION_HW_STATS_DISABLED && + act->hw_stats !=FLOW_ACTION_HW_STATS_DONT_CARE ) {NL_SET_ERR_MSG_MOD (extack,"Unsupported action HW stats type" );return -EOPNOTSUPP ; }
include/net/flow_offload.h +8 -1
@@ -166,15 +166,18 @@ enum flow_action_mangle_base {enum flow_action_hw_stats_bit {FLOW_ACTION_HW_STATS_IMMEDIATE_BIT ,FLOW_ACTION_HW_STATS_DELAYED_BIT , +FLOW_ACTION_HW_STATS_DISABLED_BIT , };enum flow_action_hw_stats { -FLOW_ACTION_HW_STATS_DISABLED =0 , +FLOW_ACTION_HW_STATS_DONT_CARE =0 ,FLOW_ACTION_HW_STATS_IMMEDIATE =BIT (FLOW_ACTION_HW_STATS_IMMEDIATE_BIT ),FLOW_ACTION_HW_STATS_DELAYED =BIT (FLOW_ACTION_HW_STATS_DELAYED_BIT ),FLOW_ACTION_HW_STATS_ANY =FLOW_ACTION_HW_STATS_IMMEDIATE |FLOW_ACTION_HW_STATS_DELAYED , +FLOW_ACTION_HW_STATS_DISABLED = +BIT (FLOW_ACTION_HW_STATS_DISABLED_BIT ), };typedef void (*action_destr )(void *priv);
@@ -325,7 +328,11 @@ __flow_action_hw_stats_check(const struct flow_action *action,return true;if (!flow_action_mixed_hw_stats_check (action, extack))return false; + action_entry =flow_action_first_entry_get (action); +if (action_entry->hw_stats ==FLOW_ACTION_HW_STATS_DONT_CARE ) +return true; +if (!check_allow_bit && action_entry->hw_stats !=FLOW_ACTION_HW_STATS_ANY ) {NL_SET_ERR_MSG_MOD (extack,"Driver supports only default HW stats type \"any\"" );
net/sched/cls_api.c +12 -2
@@ -3523,6 +3523,16 @@ static void tcf_sample_get_group(struct flow_action_entry *entry,#endif } +static enum flow_action_hw_stats tc_act_hw_stats (u8 hw_stats) +{ +if (WARN_ON_ONCE (hw_stats >TCA_ACT_HW_STATS_ANY )) +return FLOW_ACTION_HW_STATS_DONT_CARE ; +else if (!hw_stats) +return FLOW_ACTION_HW_STATS_DISABLED ; + +return hw_stats; +} +int tc_setup_flow_action (struct flow_action *flow_action,const struct tcf_exts *exts) {
@@ -3546,7 +3556,7 @@ int tc_setup_flow_action(struct flow_action *flow_action,if (err) goto err_out_locked; - entry->hw_stats = act->hw_stats; + entry->hw_stats =tc_act_hw_stats (act->hw_stats);if (is_tcf_gact_ok (act)) { entry->id =FLOW_ACTION_ACCEPT ;
@@ -3614,7 +3624,7 @@ int tc_setup_flow_action(struct flow_action *flow_action, entry->mangle.mask =tcf_pedit_mask (act, k); entry->mangle.val =tcf_pedit_val (act, k); entry->mangle.offset =tcf_pedit_offset (act, k); - entry->hw_stats = act->hw_stats; + entry->hw_stats =tc_act_hw_stats (act->hw_stats); entry = &flow_action->entries[++j]; } }else if (is_tcf_csum (act)) {
Signature · 1 line
2.20.1
From: Pablo Neira Ayuso <[email protected]>
Date: Wed, 6 May 2020 20:34:50 +0200
Quoted 11 lines
This patch adds
FLOW_ACTION_HW_STATS_DONT_CAREwhich tells the driver that the frontend does not need counters, this hw stats type request never fails. TheFLOW_ACTION_HW_STATS_DISABLEDtype explicitly requests the driver to disable the stats, however, if the driver cannot disable counters, it bails out.TCA_ACT_HW_STATS_* maintains the 1:1 mapping with FLOW_ACTION_HW_STATS_* except by disabled which is mapped to
FLOW_ACTION_HW_STATS_DISABLED(this is 0 in tc). Addtc_act_hw_stats()to perform the mapping between TCA_ACT_HW_STATS_* and FLOW_ACTION_HW_STATS_*.
Quoted 4 lines
Fixes 319a1d19471e ("flow_offload: check for basic action hw stats type")Signed-off-by Pablo Neira Ayuso <[email protected]>---
v4: Update mlxsw as Jiri prefers.
Applied, thank you.
On 06/05/2020 19:34, Pablo Neira Ayuso wrote> - } else if (act->hw_stats != FLOW_ACTION_HW_STATS_DISABLED) {
drivers/net/ethernet/mellanox/mlxsw/spectrum_flower.c static int mlxsw_sp_flower_parse_actions(struct mlxsw_sp *mlxsw_sp, · line 43
Quoted 5 lines
drivers/net/ethernet/mellanox/mlxsw/spectrum_flower.c +2
-0+ }else if (act->hw_stats !=FLOW_ACTION_HW_STATS_DISABLED && + act->hw_stats !=FLOW_ACTION_HW_STATS_DONT_CARE ) {NL_SET_ERR_MSG_MOD (extack,"Unsupported action HW stats type" );return -EOPNOTSUPP ; }
FWIW my whole reason for suggesting DONT_CARE==0 in the first place
was so that drivers could just use it as a boolean, e.g.
if (act->hw_stats && !(act->hw_stats & FLOW_ACTION_HW_STATS_BLAH))
error("driver only supports BLAH stats");If you're not even doing that then the case for DONT_CARE == ~0 is
even stronger.
Sorry I wasn't quick enough on the draw to say this before v4 was
applied (I was waiting for an answer on the v2 thread; posting a Nack on v3 felt like it might come across as needlessly combative).
-ed
Pablo Neira Ayuso
maintainerQuoted 8 lines · Edward Cree
On 06/05/2020 19:34, Pablo Neira Ayuso wrote> - } else if (
act->hw_stats!=FLOW_ACTION_HW_STATS_DISABLED) {drivers/net/ethernet/mellanox/mlxsw/spectrum_flower.c static int mlxsw_sp_flower_parse_actions(struct mlxsw_sp *mlxsw_sp, · line 43
Quoted 5 lines
drivers/net/ethernet/mellanox/mlxsw/spectrum_flower.c +2
-0+ }else if (act->hw_stats !=FLOW_ACTION_HW_STATS_DISABLED && + act->hw_stats !=FLOW_ACTION_HW_STATS_DONT_CARE ) {NL_SET_ERR_MSG_MOD (extack,"Unsupported action HW stats type" );return -EOPNOTSUPP ; }FWIW my whole reason for suggesting DONT_CARE==0 in the first place
was so that drivers could just use it as a boolean, e.g.
Quoted 5 lines
if (act->hw_stats && !(act->hw_stats & FLOW_ACTION_HW_STATS_BLAH)) error("driver only supports BLAH stats");If you're not even doing that then the case for
DONT_CARE== ~0 is
even stronger.
Could you point to what driver might have any problem with this update?
Thank you.
Quoted 1 line · Pablo Neira Ayuso
Could you point to what driver might have any problem with this update?
Drivers *can* implement the API in this patch. It's just that the alternative API Jakub proposed would make for simpler driver code. I.e. I'm not saying it's bad, just that it could be made better. That's why I didn't hard-NACK it at any point.
I guess I should send the change I'm suggesting as a patch, rather than asking it of you — I'll try to get that done today. (Although I'm not sure if it's really 'net' material or if I should wait for David to merge net into net-next and make the patch against the latter — wdyt?)
-ed