mlx5 updates 2022-07-28
Subsystems (5)
From: Saeed Mahameed <[email protected]>
This series adds misc updates to mlx5 driver.
For more information please see tag log below.
Please pull and let me know if there is any problem.
Thanks,
Saeed.
The following changes since commit 623cd87006983935de6c2ad8e2d50e68f1b7d6e7:
net: cdns,macb: use correct xlnx prefix for Xilinx (2022-07-28 13:08:53 +0200)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux.git tags/mlx5-updates-2022-07-28
for you to fetch changes up to 069448b2fd0a40b5252915dc10ae561c4b5c8c5a:
net/mlx5e: Move mlx5e_init_l2_addr to en_main (2022-07-28 13:55:30 -0700)
----------------------------------------------------------------
mlx5-updates-2022-07-28
Misc updates to mlx5 driver:
1) Gal corrects to use skb_tcp_all_headers on encapsulated skbs.
2) Roi Adds the support for offloading standalone police actions.
3) Lama did some refactoring to minimize code coupling with mlx5e_priv "god object" in some of the follows, and converts some of the objects to pointers to preserve on memory when these objects aren't needed. This is part one of two parts series.
----------------------------------------------------------------
Gal Pressman (1):
net/mlx5e: Fix wrong use of skb_tcp_all_headers() with encapsulation
Lama Kayal (9):
net/mlx5e: Convert mlx5e_tc_table member of mlx5e_flow_steering to pointer net/mlx5e: Make mlx5e_tc_table private
net/mlx5e: Allocate VLAN and TC for featured profiles only net/mlx5e: Convert mlx5e_flow_steering member of mlx5e_priv to pointer net/mlx5e: Report flow steering errors with mdev err report API net/mlx5e: Add mdev to flow_steering struct
net/mlx5e: Separate mlx5e_set_rx_mode_work and move caller to en_main net/mlx5e: Split en_fs ndo's and move to en_main
net/mlx5e: Move mlx5e_init_l2_addr to en_main
Roi Dayan (4):
net/mlx5e: TC, Allocate post meter ft per rule net/mlx5e: Add red and green counters for metering net/mlx5e: TC, Separate get/update/replace meter functions net/mlx5e: TC, Support tc action api for police
Yevgeny Kliteynik (1):
net/mlx5: DR, Add support for flow metering ASO
drivers/net/ethernet/mellanox/mlx5/core/en.h | 5 +- drivers/net/ethernet/mellanox/mlx5/core/en/fs.h | 44 +- .../mellanox/mlx5/core/en/fs_tt_redirect.c | 72 +-- drivers/net/ethernet/mellanox/mlx5/core/en/ptp.c | 8 +- .../net/ethernet/mellanox/mlx5/core/en/rep/tc.c | 117 ++++- .../ethernet/mellanox/mlx5/core/en/tc/act/act.h | 10 + .../ethernet/mellanox/mlx5/core/en/tc/act/goto.c | 2 +- .../ethernet/mellanox/mlx5/core/en/tc/act/police.c | 100 +++- .../net/ethernet/mellanox/mlx5/core/en/tc/meter.c | 189 +++++-- .../net/ethernet/mellanox/mlx5/core/en/tc/meter.h | 18 +- .../ethernet/mellanox/mlx5/core/en/tc/post_meter.c | 33 +- .../ethernet/mellanox/mlx5/core/en/tc/post_meter.h | 4 +- .../net/ethernet/mellanox/mlx5/core/en/tc_priv.h | 3 +- .../ethernet/mellanox/mlx5/core/en_accel/fs_tcp.c | 32 +- .../mellanox/mlx5/core/en_accel/ipsec_fs.c | 10 +- drivers/net/ethernet/mellanox/mlx5/core/en_arfs.c | 58 +-- drivers/net/ethernet/mellanox/mlx5/core/en_fs.c | 554 +++++++++++---------- .../ethernet/mellanox/mlx5/core/en_fs_ethtool.c | 24 +- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 63 ++- drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 34 +- drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 143 ++++-- drivers/net/ethernet/mellanox/mlx5/core/en_tc.h | 4 + drivers/net/ethernet/mellanox/mlx5/core/en_tx.c | 2 +- .../net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c | 20 +- .../mellanox/mlx5/core/steering/dr_action.c | 99 ++++ .../mellanox/mlx5/core/steering/dr_ste_v1.c | 56 +++ .../mellanox/mlx5/core/steering/dr_types.h | 17 + .../ethernet/mellanox/mlx5/core/steering/fs_dr.c | 21 + .../mellanox/mlx5/core/steering/mlx5_ifc_dr.h | 26 + .../ethernet/mellanox/mlx5/core/steering/mlx5dr.h | 8 + 30 files changed, 1265 insertions(+), 511 deletions(-)
PATCH 1/15 net/mlx5e: Fix wrong use of skb_tcp_all_headers() with encapsulation
From: Gal Pressman <[email protected]>
Use skb_inner_tcp_all_headers() instead of skb_tcp_all_headers() when transmitting an encapsulated packet in mlx5e_tx_get_gso_ihs().
--- drivers/net/ethernet/mellanox/mlx5/core/en_tx.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/mellanox/mlx5/core/en_tx.c +1 -1
@@ -152,7 +152,7 @@ mlx5e_tx_get_gso_ihs(struct mlx5e_txqsq *sq, struct sk_buff *skb, int *hopbyhop) *hopbyhop =0 ;if (skb->encapsulation) { - ihs =skb_tcp_all_headers (skb); + ihs =skb_inner_tcp_all_headers (skb); stats->tso_inner_packets++; stats->tso_inner_bytes += skb->len - ihs; }else {
Signature · 1 line
2.37.1
Re: [net-next 01/15] net/mlx5e: Fix wrong use of skb_tcp_all_headers() with encapsulation
This series was applied to netdev/net-next.git (master) by Saeed Mahameed <[email protected]>:
Hello:
This series was applied to netdev/net-next.git (master) by Saeed Mahameed <[email protected]>:
Quoted 5 lines · On Thu, 28 Jul 2022 13:57:14 -0700 you wrote:
From: Gal Pressman <[email protected]>
Use
skb_inner_tcp_all_headers()instead ofskb_tcp_all_headers()when transmitting an encapsulated packet inmlx5e_tx_get_gso_ihs().
Quoted 7 lines
Fixes 504148fedb85 ("net: add skb_[inner_]tcp_all_headers helpers")Cc Eric Dumazet <[email protected]>Signed-off-by Gal Pressman <[email protected]>Reviewed-by Tariq Toukan <[email protected]>Signed-off-by Saeed Mahameed <[email protected]>[...]
Here is the summary with links:
- [net-next,01/15] net/mlx5e: Fix wrong use of skb_tcp_all_headers() with encapsulation
https://git.kernel.org/netdev/net-next/c/ec082d31c161 - [net-next,02/15] net/mlx5: DR, Add support for flow metering ASO https://git.kernel.org/netdev/net-next/c/8920d92b8be6 - [net-next,03/15] net/mlx5e: TC, Allocate post meter ft per rule https://git.kernel.org/netdev/net-next/c/e5b1db27410e - [net-next,04/15] net/mlx5e: Add red and green counters for metering https://git.kernel.org/netdev/net-next/c/b50ce4350c10 - [net-next,05/15] net/mlx5e: TC, Separate get/update/replace meter functions https://git.kernel.org/netdev/net-next/c/f8e9d413a28a - [net-next,06/15] net/mlx5e: TC, Support tc action api for police https://git.kernel.org/netdev/net-next/c/7d1a5ce46e47 - [net-next,07/15] net/mlx5e: Convert mlx5e_tc_table member of mlx5e_flow_steering to pointer https://git.kernel.org/netdev/net-next/c/65f586c2730c - [net-next,08/15] net/mlx5e: Make mlx5e_tc_table private https://git.kernel.org/netdev/net-next/c/23bde065c3a2 - [net-next,09/15] net/mlx5e: Allocate VLAN and TC for featured profiles only https://git.kernel.org/netdev/net-next/c/454533aa87f4 - [net-next,10/15] net/mlx5e: Convert mlx5e_flow_steering member of mlx5e_priv to pointer https://git.kernel.org/netdev/net-next/c/af8bbf730068 - [net-next,11/15] net/mlx5e: Report flow steering errors with mdev err report API https://git.kernel.org/netdev/net-next/c/6a7bc5d0e1c3 - [net-next,12/15] net/mlx5e: Add mdev to flow_steering struct https://git.kernel.org/netdev/net-next/c/7bb7071568e3 - [net-next,13/15] net/mlx5e: Separate mlx5e_set_rx_mode_work and move caller to en_main https://git.kernel.org/netdev/net-next/c/5b031add2f94 - [net-next,14/15] net/mlx5e: Split en_fs ndo's and move to en_main https://git.kernel.org/netdev/net-next/c/a02c07ea5d0b - [net-next,15/15] net/mlx5e: Move mlx5e_init_l2_addr to en_main https://git.kernel.org/netdev/net-next/c/069448b2fd0a
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html
PATCH 2/15 net/mlx5: DR, Add support for flow metering ASO
In reply to Saeed MahameedFrom: Yevgeny Kliteynik <[email protected]>
Add support for ASO action of type flow metering on device that supports STEv1.
--- .../mellanox/mlx5/core/steering/dr_action.c | 99 +++++++++++++++++++ .../mellanox/mlx5/core/steering/dr_ste_v1.c | 56 +++++++++++ .../mellanox/mlx5/core/steering/dr_types.h | 17 ++++ .../mellanox/mlx5/core/steering/fs_dr.c | 21 ++++ .../mellanox/mlx5/core/steering/mlx5_ifc_dr.h | 26 +++++ .../mellanox/mlx5/core/steering/mlx5dr.h | 8 ++ 6 files changed, 227 insertions(+)
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_action.c +99 -0
@@ -22,6 +22,7 @@ enum dr_action_valid_state {DR_ACTION_STATE_PUSH_VLAN ,DR_ACTION_STATE_NON_TERM ,DR_ACTION_STATE_TERM , +DR_ACTION_STATE_ASO , DR_ACTION_STATE_MAX, };
@@ -42,6 +43,7 @@ static const char * const action_type_to_str[] = { [DR_ACTION_TYP_SAMPLER] = "DR_ACTION_TYP_SAMPLER", [DR_ACTION_TYP_INSERT_HDR] = "DR_ACTION_TYP_INSERT_HDR", [DR_ACTION_TYP_REMOVE_HDR] = "DR_ACTION_TYP_REMOVE_HDR", + [DR_ACTION_TYP_ASO_FLOW_METER] = "DR_ACTION_TYP_ASO_FLOW_METER", [DR_ACTION_TYP_MAX] = "DR_ACTION_UNKNOWN", };
@@ -71,6 +73,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_MODIFY_HDR] = DR_ACTION_STATE_MODIFY_HDR, [DR_ACTION_TYP_POP_VLAN] = DR_ACTION_STATE_POP_VLAN, [DR_ACTION_TYP_PUSH_VLAN] = DR_ACTION_STATE_PUSH_VLAN, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_DECAP] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM ,
@@ -85,6 +88,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_MODIFY_HDR] = DR_ACTION_STATE_MODIFY_HDR, [DR_ACTION_TYP_POP_VLAN] = DR_ACTION_STATE_POP_VLAN, [DR_ACTION_TYP_PUSH_VLAN] = DR_ACTION_STATE_PUSH_VLAN, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_ENCAP] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM ,
@@ -93,6 +97,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_SAMPLER] = DR_ACTION_STATE_TERM, [DR_ACTION_TYP_TAG] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_CTR] = DR_ACTION_STATE_ENCAP, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_MODIFY_HDR] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM ,
@@ -105,6 +110,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_L2_TO_TNL_L3] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_INSERT_HDR] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_PUSH_VLAN] = DR_ACTION_STATE_PUSH_VLAN, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_POP_VLAN] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM ,
@@ -118,6 +124,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_L2_TO_TNL_L2] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_L2_TO_TNL_L3] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_INSERT_HDR] = DR_ACTION_STATE_ENCAP, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_PUSH_VLAN] = { [DR_ACTION_TYP_QP ] =DR_ACTION_STATE_TERM ,
@@ -128,6 +135,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_PUSH_VLAN] = DR_ACTION_STATE_PUSH_VLAN, [DR_ACTION_TYP_L2_TO_TNL_L2] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_L2_TO_TNL_L3] = DR_ACTION_STATE_ENCAP, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_NON_TERM] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM ,
@@ -145,6 +153,13 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_MODIFY_HDR] = DR_ACTION_STATE_MODIFY_HDR, [DR_ACTION_TYP_POP_VLAN] = DR_ACTION_STATE_POP_VLAN, [DR_ACTION_TYP_PUSH_VLAN] = DR_ACTION_STATE_PUSH_VLAN, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , + }, + [DR_ACTION_STATE_ASO] = { + [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM , + [DR_ACTION_TYP_QP ] =DR_ACTION_STATE_TERM , + [DR_ACTION_TYP_FT ] =DR_ACTION_STATE_TERM , + [DR_ACTION_TYP_CTR ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_TERM ] = { [DR_ACTION_TYP_CTR ] =DR_ACTION_STATE_TERM ,
@@ -163,18 +178,21 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_MODIFY_HDR] = DR_ACTION_STATE_MODIFY_HDR, [DR_ACTION_TYP_PUSH_VLAN] = DR_ACTION_STATE_PUSH_VLAN, [DR_ACTION_TYP_POP_VLAN] = DR_ACTION_STATE_POP_VLAN, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_DECAP] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM , [DR_ACTION_TYP_FT ] =DR_ACTION_STATE_TERM , [DR_ACTION_TYP_SAMPLER ] =DR_ACTION_STATE_TERM , [DR_ACTION_TYP_CTR ] =DR_ACTION_STATE_DECAP , + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_ENCAP ] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM , [DR_ACTION_TYP_FT ] =DR_ACTION_STATE_TERM , [DR_ACTION_TYP_SAMPLER ] =DR_ACTION_STATE_TERM , [DR_ACTION_TYP_CTR ] =DR_ACTION_STATE_ENCAP , + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_MODIFY_HDR ] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM ,
@@ -185,6 +203,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_L2_TO_TNL_L3] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_INSERT_HDR] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_PUSH_VLAN] = DR_ACTION_STATE_PUSH_VLAN, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_POP_VLAN] = { [DR_ACTION_TYP_FT ] =DR_ACTION_STATE_TERM ,
@@ -196,6 +215,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_L2_TO_TNL_L2] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_L2_TO_TNL_L3] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_INSERT_HDR] = DR_ACTION_STATE_ENCAP, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_PUSH_VLAN] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM ,
@@ -206,6 +226,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_L2_TO_TNL_L2] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_L2_TO_TNL_L3] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_INSERT_HDR] = DR_ACTION_STATE_ENCAP, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_NON_TERM] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM ,
@@ -219,6 +240,16 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_MODIFY_HDR] = DR_ACTION_STATE_MODIFY_HDR, [DR_ACTION_TYP_PUSH_VLAN] = DR_ACTION_STATE_PUSH_VLAN, [DR_ACTION_TYP_POP_VLAN] = DR_ACTION_STATE_POP_VLAN, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , + }, + [DR_ACTION_STATE_ASO] = { + [DR_ACTION_TYP_L2_TO_TNL_L2 ] =DR_ACTION_STATE_ENCAP , + [DR_ACTION_TYP_L2_TO_TNL_L3 ] =DR_ACTION_STATE_ENCAP , + [DR_ACTION_TYP_MODIFY_HDR ] =DR_ACTION_STATE_MODIFY_HDR , + [DR_ACTION_TYP_PUSH_VLAN ] =DR_ACTION_STATE_PUSH_VLAN , + [DR_ACTION_TYP_CTR ] =DR_ACTION_STATE_ASO , + [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM , + [DR_ACTION_TYP_FT ] =DR_ACTION_STATE_TERM , }, [DR_ACTION_STATE_TERM ] = { [DR_ACTION_TYP_CTR ] =DR_ACTION_STATE_TERM ,
@@ -240,6 +271,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_PUSH_VLAN] = DR_ACTION_STATE_PUSH_VLAN, [DR_ACTION_TYP_POP_VLAN] = DR_ACTION_STATE_POP_VLAN, [DR_ACTION_TYP_VPORT] = DR_ACTION_STATE_TERM, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_DECAP] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM ,
@@ -253,6 +285,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_INSERT_HDR] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_POP_VLAN] = DR_ACTION_STATE_POP_VLAN, [DR_ACTION_TYP_PUSH_VLAN] = DR_ACTION_STATE_PUSH_VLAN, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_ENCAP] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM ,
@@ -261,6 +294,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_VPORT] = DR_ACTION_STATE_TERM, [DR_ACTION_TYP_SAMPLER] = DR_ACTION_STATE_TERM, [DR_ACTION_TYP_CTR] = DR_ACTION_STATE_ENCAP, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_MODIFY_HDR] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM ,
@@ -272,6 +306,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_L2_TO_TNL_L3] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_INSERT_HDR] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_PUSH_VLAN] = DR_ACTION_STATE_PUSH_VLAN, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_POP_VLAN] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM ,
@@ -284,6 +319,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_L2_TO_TNL_L2] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_L2_TO_TNL_L3] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_INSERT_HDR] = DR_ACTION_STATE_ENCAP, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_PUSH_VLAN] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM ,
@@ -296,6 +332,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_L2_TO_TNL_L2] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_L2_TO_TNL_L3] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_INSERT_HDR] = DR_ACTION_STATE_ENCAP, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_NON_TERM] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM ,
@@ -312,6 +349,13 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_POP_VLAN] = DR_ACTION_STATE_POP_VLAN, [DR_ACTION_TYP_PUSH_VLAN] = DR_ACTION_STATE_PUSH_VLAN, [DR_ACTION_TYP_VPORT] = DR_ACTION_STATE_TERM, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , + }, + [DR_ACTION_STATE_ASO] = { + [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM , + [DR_ACTION_TYP_FT ] =DR_ACTION_STATE_TERM , + [DR_ACTION_TYP_VPORT ] =DR_ACTION_STATE_TERM , + [DR_ACTION_TYP_CTR ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_TERM ] = { [DR_ACTION_TYP_CTR ] =DR_ACTION_STATE_TERM ,
@@ -331,6 +375,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_PUSH_VLAN] = DR_ACTION_STATE_PUSH_VLAN, [DR_ACTION_TYP_POP_VLAN] = DR_ACTION_STATE_POP_VLAN, [DR_ACTION_TYP_VPORT] = DR_ACTION_STATE_TERM, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_DECAP] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM ,
@@ -338,6 +383,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_CTR] = DR_ACTION_STATE_DECAP, [DR_ACTION_TYP_VPORT] = DR_ACTION_STATE_TERM, [DR_ACTION_TYP_SAMPLER] = DR_ACTION_STATE_TERM, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_ENCAP] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM ,
@@ -345,6 +391,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_CTR] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_SAMPLER] = DR_ACTION_STATE_TERM, [DR_ACTION_TYP_VPORT] = DR_ACTION_STATE_TERM, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_MODIFY_HDR] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM ,
@@ -356,6 +403,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_INSERT_HDR] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_PUSH_VLAN] = DR_ACTION_STATE_PUSH_VLAN, [DR_ACTION_TYP_VPORT] = DR_ACTION_STATE_TERM, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_POP_VLAN] = { [DR_ACTION_TYP_FT ] =DR_ACTION_STATE_TERM ,
@@ -368,6 +416,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_L2_TO_TNL_L3] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_INSERT_HDR] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_VPORT] = DR_ACTION_STATE_TERM, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_PUSH_VLAN] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM ,
@@ -379,6 +428,7 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_L2_TO_TNL_L3] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_INSERT_HDR] = DR_ACTION_STATE_ENCAP, [DR_ACTION_TYP_VPORT] = DR_ACTION_STATE_TERM, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_NON_TERM] = { [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM ,
@@ -393,6 +443,17 @@ next_action_state[DR_ACTION_DOMAIN_MAX][DR_ACTION_STATE_MAX][DR_ACTION_TYP_MAX] [DR_ACTION_TYP_PUSH_VLAN] = DR_ACTION_STATE_PUSH_VLAN, [DR_ACTION_TYP_POP_VLAN] = DR_ACTION_STATE_POP_VLAN, [DR_ACTION_TYP_VPORT] = DR_ACTION_STATE_TERM, + [DR_ACTION_TYP_ASO_FLOW_METER ] =DR_ACTION_STATE_ASO , + }, + [DR_ACTION_STATE_ASO] = { + [DR_ACTION_TYP_L2_TO_TNL_L2 ] =DR_ACTION_STATE_ENCAP , + [DR_ACTION_TYP_L2_TO_TNL_L3 ] =DR_ACTION_STATE_ENCAP , + [DR_ACTION_TYP_MODIFY_HDR ] =DR_ACTION_STATE_MODIFY_HDR , + [DR_ACTION_TYP_PUSH_VLAN ] =DR_ACTION_STATE_PUSH_VLAN , + [DR_ACTION_TYP_DROP ] =DR_ACTION_STATE_TERM , + [DR_ACTION_TYP_FT ] =DR_ACTION_STATE_TERM , + [DR_ACTION_TYP_VPORT ] =DR_ACTION_STATE_TERM , + [DR_ACTION_TYP_CTR ] =DR_ACTION_STATE_ASO , }, [DR_ACTION_STATE_TERM ] = { [DR_ACTION_TYP_CTR ] =DR_ACTION_STATE_TERM ,
@@ -738,6 +799,12 @@ int mlx5dr_actions_build_ste_arr(struct mlx5dr_matcher *matcher, attr.reformat.param_0 = action->reformat->param_0; attr.reformat.param_1 = action->reformat->param_1;break ; +case DR_ACTION_TYP_ASO_FLOW_METER : + attr.aso_flow_meter.obj_id = action->aso->obj_id; + attr.aso_flow_meter.offset = action->aso->offset; + attr.aso_flow_meter.dest_reg_id = action->aso->dest_reg_id; + attr.aso_flow_meter.init_color = action->aso->init_color; +break ;default :mlx5dr_err (dmn,"Unsupported action type %d\n" , action_type);return -EINVAL ;
@@ -798,6 +865,7 @@ static unsigned int action_size[DR_ACTION_TYP_MAX] = { [DR_ACTION_TYP_INSERT_HDR] = sizeof(struct mlx5dr_action_reformat), [DR_ACTION_TYP_REMOVE_HDR] = sizeof(struct mlx5dr_action_reformat), [DR_ACTION_TYP_SAMPLER] = sizeof(struct mlx5dr_action_sampler), + [DR_ACTION_TYP_ASO_FLOW_METER] = sizeof(struct mlx5dr_action_aso_flow_meter), }; static struct mlx5dr_action *
@@ -1830,6 +1898,34 @@ mlx5dr_action_create_dest_vport(struct mlx5dr_domain *dmn,return action; } +struct mlx5dr_action * +mlx5dr_action_create_aso (struct mlx5dr_domain *dmn,u32 obj_id, +u8 dest_reg_id,u8 aso_type, +u8 init_color,u8 meter_id) +{ +struct mlx5dr_action *action; + +if (aso_type !=MLX5_EXE_ASO_FLOW_METER ) +return NULL ; + +if (init_color >MLX5_FLOW_METER_COLOR_UNDEFINED ) +return NULL ; + + action =dr_action_create_generic (DR_ACTION_TYP_ASO_FLOW_METER ); +if (!action) +return NULL ; + + action->aso->obj_id = obj_id; + action->aso->offset = meter_id; + action->aso->dest_reg_id = dest_reg_id; + action->aso->init_color = init_color; + action->aso->dmn = dmn; + +refcount_inc (&dmn->refcount); + +return action; +} +int mlx5dr_action_destroy (struct mlx5dr_action *action) {if (WARN_ON_ONCE (refcount_read (&action->refcount) >1 ))
@@ -1881,6 +1977,9 @@ int mlx5dr_action_destroy(struct mlx5dr_action *action)case DR_ACTION_TYP_SAMPLER :refcount_dec (&action->sampler->dmn->refcount);break ; +case DR_ACTION_TYP_ASO_FLOW_METER : +refcount_dec (&action->aso->dmn->refcount); +break ;default :break ; }
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste_v1.c +56 -0
86 lines not on this page · raw message
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h +17 -0
@@ -127,6 +127,7 @@ enum mlx5dr_action_type {DR_ACTION_TYP_INSERT_HDR ,DR_ACTION_TYP_REMOVE_HDR ,DR_ACTION_TYP_SAMPLER , +DR_ACTION_TYP_ASO_FLOW_METER , DR_ACTION_TYP_MAX, };
@@ -271,6 +272,13 @@ struct mlx5dr_ste_actions_attr {int count;u32 headers[MLX5DR_MAX_VLANS ]; } vlans; + +struct { +u32 obj_id; +u32 offset; +u8 dest_reg_id; +u8 init_color; + } aso_flow_meter; }; void mlx5dr_ste_set_actions_rx(struct mlx5dr_ste_ctx *ste_ctx,
@@ -1035,6 +1043,14 @@ struct mlx5dr_rule_action_member {struct list_head list; }; +struct mlx5dr_action_aso_flow_meter { +struct mlx5dr_domain *dmn; +u32 obj_id; +u32 offset; +u8 dest_reg_id; +u8 init_color; +}; +struct mlx5dr_action {enum mlx5dr_action_type action_type;refcount_t refcount;
@@ -1049,6 +1065,7 @@ struct mlx5dr_action {struct mlx5dr_action_vport *vport;struct mlx5dr_action_push_vlan *push_vlan;struct mlx5dr_action_flow_tag *flow_tag; +struct mlx5dr_action_aso_flow_meter *aso; }; };
drivers/net/ethernet/mellanox/mlx5/core/steering/fs_dr.c +21 -0
@@ -500,6 +500,27 @@ static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns, } } +if (fte->action.action &MLX5_FLOW_CONTEXT_ACTION_EXECUTE_ASO ) { +if (fte->action.exe_aso.type !=MLX5_EXE_ASO_FLOW_METER ) { + err = -EOPNOTSUPP ; + goto free_actions; + } + + tmp_action = +mlx5dr_action_create_aso (domain, + fte->action.exe_aso.object_id, + fte->action.exe_aso.return_reg_id, + fte->action.exe_aso.type, + fte->action.exe_aso.flow_meter.init_color, + fte->action.exe_aso.flow_meter.meter_idx); +if (!tmp_action) { + err = -ENOMEM ; + goto free_actions; + } + fs_dr_actions[fs_dr_num_actions++] = tmp_action; + actions[num_actions++] = tmp_action; + } + params.match_sz = match_sz; params.match_buf = (u64 *)fte->val;if (num_term_actions ==1 ) {
drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5_ifc_dr.h +26 -0
@@ -574,4 +574,30 @@ struct mlx5_ifc_dr_action_hw_copy_bits { u8 reserved_at_38[0x8]; }; +enum { +MLX5DR_ASO_FLOW_METER_NUM_PER_OBJ =2 , +}; + +struct mlx5_ifc_ste_aso_flow_meter_action_bits { +u8 reserved_at_0[0xc ]; +u8 action[0x1 ]; +u8 initial_color[0x2 ]; +u8 line_id[0x1 ]; +}; + +struct mlx5_ifc_ste_double_action_aso_v1_bits { +u8 action_id[0x8 ]; +u8 aso_context_number[0x18 ]; + +u8 dest_reg_id[0x2 ]; +u8 change_ordering_tag[0x1 ]; +u8 aso_check_ordering[0x1 ]; +u8 aso_context_type[0x4 ]; +u8 reserved_at_28[0x8 ]; +union { +u8 aso_fields[0x10 ]; +struct mlx5_ifc_ste_aso_flow_meter_action_bits flow_meter; + }; +}; +#endif /* MLX5_IFC_DR_H */
drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5dr.h +8 -0
@@ -131,6 +131,14 @@ struct mlx5dr_action *mlx5dr_action_create_pop_vlan(void);struct mlx5dr_action *mlx5dr_action_create_push_vlan (struct mlx5dr_domain *domain,__be32 vlan_hdr); +struct mlx5dr_action * +mlx5dr_action_create_aso (struct mlx5dr_domain *dmn, +u32 obj_id, +u8 return_reg_id, +u8 aso_type, +u8 init_color, +u8 meter_id); +int mlx5dr_action_destroy (struct mlx5dr_action *action); static inline bool
Signature · 1 line
2.37.1
PATCH 3/15 net/mlx5e: TC, Allocate post meter ft per rule
In reply to Saeed MahameedFrom: Roi Dayan <[email protected]>
To support a TC police action notexceed counter and supporting actions other than drop/pipe there is a need to create separate ft and rules per rule and not to use a common one created on eswitch init.
--- .../ethernet/mellanox/mlx5/core/en/tc/meter.c | 19 +++----------- .../ethernet/mellanox/mlx5/core/en/tc/meter.h | 6 +++-- .../net/ethernet/mellanox/mlx5/core/en_tc.c | 26 +++++++++++++++++-- 3 files changed, 31 insertions(+), 20 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/tc/meter.c +3 -16
@@ -6,7 +6,6 @@#include "en/tc/post_act.h" #include "meter.h" #include "en/tc_priv.h" -#include "post_meter.h" #define MLX5_START_COLOR_SHIFT 28#define MLX5_METER_MODE_SHIFT 24
@@ -47,8 +46,6 @@ struct mlx5e_flow_meters {struct mlx5_core_dev *mdev;struct mlx5e_post_act *post_act; - -struct mlx5e_post_meter_priv *post_meter; }; static void
@@ -390,10 +387,10 @@ mlx5e_tc_meter_put(struct mlx5e_flow_meter_handle *meter)mutex_unlock (&flow_meters->sync_lock); } -struct mlx5_flow_table * -mlx5e_tc_meter_get_post_meter_ft (struct mlx5e_flow_meters *flow_meters) +enum mlx5_flow_namespace_type +mlx5e_tc_meter_get_namespace (struct mlx5e_flow_meters *flow_meters) { -return mlx5e_post_meter_get_ft (flow_meters->post_meter); +return flow_meters->ns_type; } struct mlx5e_flow_meters *
@@ -432,12 +429,6 @@ mlx5e_flow_meters_init(struct mlx5e_priv *priv, goto err_sq; } - flow_meters->post_meter =mlx5e_post_meter_init (priv, ns_type, post_act); -if (IS_ERR (flow_meters->post_meter)) { - err =PTR_ERR (flow_meters->post_meter); - goto err_post_meter; - } -mutex_init (&flow_meters->sync_lock);INIT_LIST_HEAD (&flow_meters->partial_list);INIT_LIST_HEAD (&flow_meters->full_list);
@@ -451,8 +442,6 @@ mlx5e_flow_meters_init(struct mlx5e_priv *priv,return flow_meters; -err_post_meter: -mlx5_aso_destroy (flow_meters->aso); err_sq:mlx5_core_dealloc_pd (mdev, flow_meters->pdn); err_out:
@@ -466,9 +455,7 @@ mlx5e_flow_meters_cleanup(struct mlx5e_flow_meters *flow_meters)if (IS_ERR_OR_NULL (flow_meters))return ; -mlx5e_post_meter_cleanup (flow_meters->post_meter);mlx5_aso_destroy (flow_meters->aso);mlx5_core_dealloc_pd (flow_meters->mdev, flow_meters->pdn); -kfree (flow_meters); }
drivers/net/ethernet/mellanox/mlx5/core/en/tc/meter.h +4 -2
@@ -4,6 +4,7 @@#ifndef __MLX5_EN_FLOW_METER_H__#define __MLX5_EN_FLOW_METER_H__ +struct mlx5e_post_meter_priv ;struct mlx5e_flow_meter_aso_obj ;struct mlx5e_flow_meters ;struct mlx5_flow_attr ;
@@ -35,6 +36,7 @@ struct mlx5e_flow_meter_handle {struct mlx5e_meter_attr {struct mlx5e_flow_meter_params params;struct mlx5e_flow_meter_handle *meter; +struct mlx5e_post_meter_priv *post_meter; };int
@@ -47,8 +49,8 @@ mlx5e_tc_meter_get(struct mlx5_core_dev *mdev, struct mlx5e_flow_meter_params *pvoid mlx5e_tc_meter_put (struct mlx5e_flow_meter_handle *meter); -struct mlx5_flow_table * -mlx5e_tc_meter_get_post_meter_ft (struct mlx5e_flow_meters *flow_meters); +enum mlx5_flow_namespace_type +mlx5e_tc_meter_get_namespace (struct mlx5e_flow_meters *flow_meters);struct mlx5e_flow_meters * mlx5e_flow_meters_init(struct mlx5e_priv *priv,
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +24 -2
@@ -356,6 +356,9 @@ static int mlx5e_tc_add_flow_meter(struct mlx5e_priv *priv, struct mlx5_flow_attr *attr) { + struct mlx5e_post_act *post_act = get_post_action(priv); +struct mlx5e_post_meter_priv *post_meter; +enum mlx5_flow_namespace_type ns_type;struct mlx5e_flow_meter_handle *meter; meter =mlx5e_tc_meter_get (priv->mdev, &attr->meter_attr.params);
@@ -364,11 +367,30 @@ mlx5e_tc_add_flow_meter(struct mlx5e_priv *priv,return PTR_ERR (meter); } + ns_type =mlx5e_tc_meter_get_namespace (meter->flow_meters); + post_meter =mlx5e_post_meter_init (priv, ns_type, post_act); +if (IS_ERR (post_meter)) { +mlx5_core_err (priv->mdev,"Failed to init post meter\n" ); + goto err_meter_init; + } + attr->meter_attr.meter = meter; - attr->dest_ft =mlx5e_tc_meter_get_post_meter_ft (meter->flow_meters); + attr->meter_attr.post_meter = post_meter; + attr->dest_ft =mlx5e_post_meter_get_ft (post_meter); attr->action |=MLX5_FLOW_CONTEXT_ACTION_FWD_DEST ;return 0 ; + +err_meter_init: +mlx5e_tc_meter_put (meter); +return PTR_ERR (post_meter); +} + +static void +mlx5e_tc_del_flow_meter (struct mlx5_flow_attr *attr) +{ +mlx5e_post_meter_cleanup (attr->meter_attr.post_meter); +mlx5e_tc_meter_put (attr->meter_attr.meter); } struct mlx5_flow_handle *
@@ -428,7 +450,7 @@ mlx5e_tc_rule_unoffload(struct mlx5e_priv *priv,mlx5_eswitch_del_offloaded_rule (esw, rule, attr);if (attr->meter_attr.meter) -mlx5e_tc_meter_put (attr->meter_attr.meter); +mlx5e_tc_del_flow_meter (attr); }int
Signature · 1 line
2.37.1
PATCH 4/15 net/mlx5e: Add red and green counters for metering
In reply to Saeed MahameedFrom: Roi Dayan <[email protected]>
Add red and green counters per meter instance. TC police action is implemented as a meter instance. The meter counters represent the police action notexceed/exceed counters.
TC rules using the same meter instance will use the same counters.
--- .../ethernet/mellanox/mlx5/core/en/tc/meter.c | 22 +++++++++++++ .../ethernet/mellanox/mlx5/core/en/tc/meter.h | 3 ++ .../mellanox/mlx5/core/en/tc/post_meter.c | 33 ++++++++++++------- .../mellanox/mlx5/core/en/tc/post_meter.h | 4 ++- .../net/ethernet/mellanox/mlx5/core/en_tc.c | 3 +- 5 files changed, 52 insertions(+), 13 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/tc/meter.c +22 -0
@@ -240,6 +240,7 @@ __mlx5e_flow_meter_alloc(struct mlx5e_flow_meters *flow_meters)struct mlx5_core_dev *mdev = flow_meters->mdev;struct mlx5e_flow_meter_aso_obj *meters_obj;struct mlx5e_flow_meter_handle *meter; +struct mlx5_fc *counter;int err, pos, total;u32 id;
@@ -247,6 +248,20 @@ __mlx5e_flow_meter_alloc(struct mlx5e_flow_meters *flow_meters)if (!meter)return ERR_PTR (-ENOMEM ); + counter =mlx5_fc_create (mdev, true); +if (IS_ERR (counter)) { + err =PTR_ERR (counter); + goto err_red_counter; + } + meter->red_counter = counter; + + counter =mlx5_fc_create (mdev, true); +if (IS_ERR (counter)) { + err =PTR_ERR (counter); + goto err_green_counter; + } + meter->green_counter = counter; + meters_obj =list_first_entry_or_null (&flow_meters->partial_list, struct mlx5e_flow_meter_aso_obj, entry);
@@ -292,6 +307,10 @@ __mlx5e_flow_meter_alloc(struct mlx5e_flow_meters *flow_meters) err_mem:mlx5e_flow_meter_destroy_aso_obj (mdev, id); err_create: +mlx5_fc_destroy (mdev, meter->green_counter); +err_green_counter: +mlx5_fc_destroy (mdev, meter->red_counter); +err_red_counter:kfree (meter);return ERR_PTR (err); }
@@ -304,6 +323,9 @@ __mlx5e_flow_meter_free(struct mlx5e_flow_meter_handle *meter)struct mlx5e_flow_meter_aso_obj *meters_obj;int n, pos; +mlx5_fc_destroy (mdev, meter->green_counter); +mlx5_fc_destroy (mdev, meter->red_counter); + meters_obj = meter->meters_obj; pos = (meter->obj_id - meters_obj->base_id) *2 + meter->idx;bitmap_clear (meters_obj->meters_map, pos,1 );
drivers/net/ethernet/mellanox/mlx5/core/en/tc/meter.h +3 -0
@@ -31,6 +31,9 @@ struct mlx5e_flow_meter_handle {int refcnt;struct hlist_node hlist;struct mlx5e_flow_meter_params params; + +struct mlx5_fc *green_counter; +struct mlx5_fc *red_counter; };struct mlx5e_meter_attr {
drivers/net/ethernet/mellanox/mlx5/core/en/tc/post_meter.c +22 -11
@@ -84,9 +84,11 @@ mlx5e_post_meter_fg_create(struct mlx5e_priv *priv,static int mlx5e_post_meter_rules_create (struct mlx5e_priv *priv,struct mlx5e_post_meter_priv *post_meter, -struct mlx5e_post_act *post_act) +struct mlx5e_post_act *post_act, +struct mlx5_fc *green_counter, +struct mlx5_fc *red_counter) { -struct mlx5_flow_destination dest = {}; +struct mlx5_flow_destination dest[2 ] = {};struct mlx5_flow_act flow_act = {};struct mlx5_flow_handle *rule;struct mlx5_flow_spec *spec;
@@ -98,10 +100,13 @@ mlx5e_post_meter_rules_create(struct mlx5e_priv *priv,mlx5e_tc_match_to_reg_match (spec,PACKET_COLOR_TO_REG ,MLX5_FLOW_METER_COLOR_RED ,MLX5_PACKET_COLOR_MASK ); - flow_act.action =MLX5_FLOW_CONTEXT_ACTION_DROP ; + flow_act.action =MLX5_FLOW_CONTEXT_ACTION_DROP | +MLX5_FLOW_CONTEXT_ACTION_COUNT ; flow_act.flags |=FLOW_ACT_IGNORE_FLOW_LEVEL ; + dest[0 ].type =MLX5_FLOW_DESTINATION_TYPE_COUNTER ; + dest[0 ].counter_id =mlx5_fc_id (red_counter); - rule =mlx5_add_flow_rules (post_meter->ft, spec, &flow_act,NULL ,0 ); + rule =mlx5_add_flow_rules (post_meter->ft, spec, &flow_act, dest,1 );if (IS_ERR (rule)) {mlx5_core_warn (priv->mdev,"Failed to create post_meter flow drop rule\n" ); err =PTR_ERR (rule);
@@ -111,11 +116,14 @@ mlx5e_post_meter_rules_create(struct mlx5e_priv *priv,mlx5e_tc_match_to_reg_match (spec,PACKET_COLOR_TO_REG ,MLX5_FLOW_METER_COLOR_GREEN ,MLX5_PACKET_COLOR_MASK ); - flow_act.action =MLX5_FLOW_CONTEXT_ACTION_FWD_DEST ; - dest.type =MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE ; - dest.ft =mlx5e_tc_post_act_get_ft (post_act); - - rule =mlx5_add_flow_rules (post_meter->ft, spec, &flow_act, &dest,1 ); + flow_act.action =MLX5_FLOW_CONTEXT_ACTION_FWD_DEST | +MLX5_FLOW_CONTEXT_ACTION_COUNT ; + dest[0 ].type =MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE ; + dest[0 ].ft =mlx5e_tc_post_act_get_ft (post_act); + dest[1 ].type =MLX5_FLOW_DESTINATION_TYPE_COUNTER ; + dest[1 ].counter_id =mlx5_fc_id (green_counter); + + rule =mlx5_add_flow_rules (post_meter->ft, spec, &flow_act, dest,2 );if (IS_ERR (rule)) {mlx5_core_warn (priv->mdev,"Failed to create post_meter flow fwd rule\n" ); err =PTR_ERR (rule);
@@ -155,7 +163,9 @@ mlx5e_post_meter_table_destroy(struct mlx5e_post_meter_priv *post_meter)struct mlx5e_post_meter_priv *mlx5e_post_meter_init (struct mlx5e_priv *priv,enum mlx5_flow_namespace_type ns_type, -struct mlx5e_post_act *post_act) +struct mlx5e_post_act *post_act, +struct mlx5_fc *green_counter, +struct mlx5_fc *red_counter) {struct mlx5e_post_meter_priv *post_meter;int err;
@@ -172,7 +182,8 @@ mlx5e_post_meter_init(struct mlx5e_priv *priv,if (err) goto err_fg; - err =mlx5e_post_meter_rules_create (priv, post_meter, post_act); + err =mlx5e_post_meter_rules_create (priv, post_meter, post_act, green_counter, + red_counter);if (err) goto err_rules;
drivers/net/ethernet/mellanox/mlx5/core/en/tc/post_meter.h +3 -1
@@ -20,7 +20,9 @@ mlx5e_post_meter_get_ft(struct mlx5e_post_meter_priv *post_meter);struct mlx5e_post_meter_priv *mlx5e_post_meter_init (struct mlx5e_priv *priv,enum mlx5_flow_namespace_type ns_type, -struct mlx5e_post_act *post_act); +struct mlx5e_post_act *post_act, +struct mlx5_fc *green_counter, +struct mlx5_fc *red_counter);void mlx5e_post_meter_cleanup (struct mlx5e_post_meter_priv *post_meter);
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +2 -1
@@ -368,7 +368,8 @@ mlx5e_tc_add_flow_meter(struct mlx5e_priv *priv, } ns_type =mlx5e_tc_meter_get_namespace (meter->flow_meters); - post_meter =mlx5e_post_meter_init (priv, ns_type, post_act); + post_meter =mlx5e_post_meter_init (priv, ns_type, post_act, meter->green_counter, + meter->red_counter);if (IS_ERR (post_meter)) {mlx5_core_err (priv->mdev,"Failed to init post meter\n" ); goto err_meter_init;
Signature · 1 line
2.37.1
PATCH 5/15 net/mlx5e: TC, Separate get/update/replace meter functions
In reply to Saeed MahameedFrom: Roi Dayan <[email protected]>
mlx5e_tc_meter_get() to get an existing meter. mlx5e_tc_meter_update() to update an existing meter without refcount. mlx5e_tc_meter_replace() to get/create a meter and update if needed.
--- .../ethernet/mellanox/mlx5/core/en/tc/meter.c | 132 ++++++++++++++---- .../ethernet/mellanox/mlx5/core/en/tc/meter.h | 5 + .../net/ethernet/mellanox/mlx5/core/en_tc.c | 2 +- 3 files changed, 112 insertions(+), 27 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/tc/meter.c +106 -26
@@ -344,69 +344,149 @@ __mlx5e_flow_meter_free(struct mlx5e_flow_meter_handle *meter)kfree (meter); } +static struct mlx5e_flow_meter_handle * +__mlx5e_tc_meter_get (struct mlx5e_flow_meters *flow_meters,u32 index) +{ +struct mlx5e_flow_meter_handle *meter; + +hash_for_each_possible (flow_meters->hashtbl, meter, hlist, index) +if (meter->params.index == index) + goto add_ref; + +return ERR_PTR (-ENOENT ); + +add_ref: + meter->refcnt++; + +return meter; +} +struct mlx5e_flow_meter_handle *mlx5e_tc_meter_get (struct mlx5_core_dev *mdev,struct mlx5e_flow_meter_params *params) {struct mlx5e_flow_meters *flow_meters;struct mlx5e_flow_meter_handle *meter; -int err; flow_meters =mlx5e_get_flow_meters (mdev);if (!flow_meters)return ERR_PTR (-EOPNOTSUPP );mutex_lock (&flow_meters->sync_lock); -hash_for_each_possible (flow_meters->hashtbl, meter, hlist, params->index) -if (meter->params.index == params->index) - goto add_ref; + meter =__mlx5e_tc_meter_get (flow_meters, params->index); +mutex_unlock (&flow_meters->sync_lock); - meter =__mlx5e_flow_meter_alloc (flow_meters); -if (IS_ERR (meter)) { - err =PTR_ERR (meter); - goto err_alloc; +return meter; +} + +static void +__mlx5e_tc_meter_put (struct mlx5e_flow_meter_handle *meter) +{ +if (--meter->refcnt ==0 ) { +hash_del (&meter->hlist); +__mlx5e_flow_meter_free (meter); } +} + +void +mlx5e_tc_meter_put (struct mlx5e_flow_meter_handle *meter) +{ +struct mlx5e_flow_meters *flow_meters = meter->flow_meters; + +mutex_lock (&flow_meters->sync_lock); +__mlx5e_tc_meter_put (meter); +mutex_unlock (&flow_meters->sync_lock); +} + +static struct mlx5e_flow_meter_handle * +mlx5e_tc_meter_alloc (struct mlx5e_flow_meters *flow_meters, +struct mlx5e_flow_meter_params *params) +{ +struct mlx5e_flow_meter_handle *meter; + + meter =__mlx5e_flow_meter_alloc (flow_meters); +if (IS_ERR (meter)) +return meter;hash_add (flow_meters->hashtbl, &meter->hlist, params->index); meter->params.index = params->index; - -add_ref: meter->refcnt++; +return meter; +} + +static int +__mlx5e_tc_meter_update (struct mlx5e_flow_meter_handle *meter, +struct mlx5e_flow_meter_params *params) +{ +struct mlx5_core_dev *mdev = meter->flow_meters->mdev; +int err =0 ; +if (meter->params.mode != params->mode || meter->params.rate != params->rate || meter->params.burst != params->burst) { err =mlx5e_tc_meter_modify (mdev, meter, params);if (err) - goto err_update; + goto out; meter->params.mode = params->mode; meter->params.rate = params->rate; meter->params.burst = params->burst; } -mutex_unlock (&flow_meters->sync_lock); -return meter; +out: +return err; +} -err_update: -if (--meter->refcnt ==0 ) { -hash_del (&meter->hlist); -__mlx5e_flow_meter_free (meter); - } -err_alloc: +int +mlx5e_tc_meter_update (struct mlx5e_flow_meter_handle *meter, +struct mlx5e_flow_meter_params *params) +{ +struct mlx5_core_dev *mdev = meter->flow_meters->mdev; +struct mlx5e_flow_meters *flow_meters; +int err; + + flow_meters =mlx5e_get_flow_meters (mdev); +if (!flow_meters) +return -EOPNOTSUPP ; + +mutex_lock (&flow_meters->sync_lock); + err =__mlx5e_tc_meter_update (meter, params);mutex_unlock (&flow_meters->sync_lock); -return ERR_PTR (err); +return err; } -void -mlx5e_tc_meter_put (struct mlx5e_flow_meter_handle *meter) +struct mlx5e_flow_meter_handle * +mlx5e_tc_meter_replace (struct mlx5_core_dev *mdev,struct mlx5e_flow_meter_params *params) { -struct mlx5e_flow_meters *flow_meters = meter->flow_meters; +struct mlx5e_flow_meters *flow_meters; +struct mlx5e_flow_meter_handle *meter; +int err; + + flow_meters =mlx5e_get_flow_meters (mdev); +if (!flow_meters) +return ERR_PTR (-EOPNOTSUPP );mutex_lock (&flow_meters->sync_lock); -if (--meter->refcnt ==0 ) { -hash_del (&meter->hlist); -__mlx5e_flow_meter_free (meter); + meter =__mlx5e_tc_meter_get (flow_meters, params->index); +if (IS_ERR (meter)) { + meter =mlx5e_tc_meter_alloc (flow_meters, params); +if (IS_ERR (meter)) { + err =PTR_ERR (meter); + goto err_get; + } } + + err =__mlx5e_tc_meter_update (meter, params); +if (err) + goto err_update; +mutex_unlock (&flow_meters->sync_lock); +return meter; + +err_update: +__mlx5e_tc_meter_put (meter); +err_get: +mutex_unlock (&flow_meters->sync_lock); +return ERR_PTR (err); }enum mlx5_flow_namespace_type
drivers/net/ethernet/mellanox/mlx5/core/en/tc/meter.h +5 -0
@@ -51,6 +51,11 @@ struct mlx5e_flow_meter_handle * mlx5e_tc_meter_get(struct mlx5_core_dev *mdev, struct mlx5e_flow_meter_params *params);void mlx5e_tc_meter_put (struct mlx5e_flow_meter_handle *meter); +int +mlx5e_tc_meter_update (struct mlx5e_flow_meter_handle *meter, +struct mlx5e_flow_meter_params *params); +struct mlx5e_flow_meter_handle * +mlx5e_tc_meter_replace (struct mlx5_core_dev *mdev,struct mlx5e_flow_meter_params *params);enum mlx5_flow_namespace_type mlx5e_tc_meter_get_namespace (struct mlx5e_flow_meters *flow_meters);
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +1 -1
@@ -361,7 +361,7 @@ mlx5e_tc_add_flow_meter(struct mlx5e_priv *priv,enum mlx5_flow_namespace_type ns_type;struct mlx5e_flow_meter_handle *meter; - meter =mlx5e_tc_meter_get (priv->mdev, &attr->meter_attr.params); + meter =mlx5e_tc_meter_replace (priv->mdev, &attr->meter_attr.params);if (IS_ERR (meter)) {mlx5_core_err (priv->mdev,"Failed to get flow meter\n" );return PTR_ERR (meter);
Signature · 1 line
2.37.1
PATCH 6/15 net/mlx5e: TC, Support tc action api for police
In reply to Saeed MahameedFrom: Roi Dayan <[email protected]>
Add support for tc action api for police. Offloading standalone police action without a tc rule and reporting stats.
--- .../ethernet/mellanox/mlx5/core/en/rep/tc.c | 117 +++++++++++++++++- .../mellanox/mlx5/core/en/tc/act/act.h | 10 ++ .../mellanox/mlx5/core/en/tc/act/police.c | 100 +++++++++++++-- .../ethernet/mellanox/mlx5/core/en/tc/meter.c | 16 +++ .../ethernet/mellanox/mlx5/core/en/tc/meter.h | 4 + 5 files changed, 239 insertions(+), 8 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.c +116 -1
@@ -21,6 +21,7 @@#include "en/tc/sample.h" #include "en_accel/ipsec_rxtx.h" #include "en/tc/int_port.h" +#include "en/tc/act/act.h" struct mlx5e_rep_indr_block_priv {struct net_device *netdev;
@@ -511,6 +512,120 @@ mlx5e_rep_indr_setup_block(struct net_device *netdev, struct Qdisc *sch,return 0 ; } +static int +mlx5e_rep_indr_replace_act (struct mlx5e_rep_priv *rpriv, +struct flow_offload_action *fl_act) + +{ +struct mlx5e_priv *priv =netdev_priv (rpriv->netdev); +struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; +enum mlx5_flow_namespace_type ns_type; +struct flow_action_entry *action; +struct mlx5e_tc_act *act; +bool add = false; +int i; + +/* There is no use case currently for more than one action (e.g. pedit). +* when there will be, need to handle cleaning multiple actions on err. +*/ +if (!flow_offload_has_one_action (&fl_act->action)) +return -EOPNOTSUPP ; + +if (esw && esw->mode ==MLX5_ESWITCH_OFFLOADS ) + ns_type =MLX5_FLOW_NAMESPACE_FDB ; +else + ns_type =MLX5_FLOW_NAMESPACE_KERNEL ; + +flow_action_for_each (i, action, &fl_act->action) { + act =mlx5e_tc_act_get (action->id, ns_type); +if (!act) +continue ; + +if (!act->offload_action) +continue ; + +if (!act->offload_action (priv, fl_act, action)) + add = true; + } + +return add ?0 : -EOPNOTSUPP ; +} + +static int +mlx5e_rep_indr_destroy_act (struct mlx5e_rep_priv *rpriv, +struct flow_offload_action *fl_act) +{ +struct mlx5e_priv *priv =netdev_priv (rpriv->netdev); +struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; +enum mlx5_flow_namespace_type ns_type; +struct mlx5e_tc_act *act; + +if (esw && esw->mode ==MLX5_ESWITCH_OFFLOADS ) + ns_type =MLX5_FLOW_NAMESPACE_FDB ; +else + ns_type =MLX5_FLOW_NAMESPACE_KERNEL ; + + act =mlx5e_tc_act_get (fl_act->id, ns_type); +if (!act || !act->destroy_action) +return -EOPNOTSUPP ; + +return act->destroy_action (priv, fl_act); +} + +static int +mlx5e_rep_indr_stats_act (struct mlx5e_rep_priv *rpriv, +struct flow_offload_action *fl_act) + +{ +struct mlx5e_priv *priv =netdev_priv (rpriv->netdev); +struct mlx5_eswitch *esw = priv->mdev->priv.eswitch; +enum mlx5_flow_namespace_type ns_type; +struct mlx5e_tc_act *act; + +if (esw && esw->mode ==MLX5_ESWITCH_OFFLOADS ) + ns_type =MLX5_FLOW_NAMESPACE_FDB ; +else + ns_type =MLX5_FLOW_NAMESPACE_KERNEL ; + + act =mlx5e_tc_act_get (fl_act->id, ns_type); +if (!act || !act->stats_action) +return -EOPNOTSUPP ; + +return act->stats_action (priv, fl_act); +} + +static int +mlx5e_rep_indr_setup_act (struct mlx5e_rep_priv *rpriv, +struct flow_offload_action *fl_act) +{ +switch (fl_act->command) { +case FLOW_ACT_REPLACE : +return mlx5e_rep_indr_replace_act (rpriv, fl_act); +case FLOW_ACT_DESTROY : +return mlx5e_rep_indr_destroy_act (rpriv, fl_act); +case FLOW_ACT_STATS : +return mlx5e_rep_indr_stats_act (rpriv, fl_act); +default : +return -EOPNOTSUPP ; + } +} + +static int +mlx5e_rep_indr_no_dev_setup (struct mlx5e_rep_priv *rpriv, +enum tc_setup_type type, +void *data) +{ +if (!data) +return -EOPNOTSUPP ; + +switch (type) { +case TC_SETUP_ACT : +return mlx5e_rep_indr_setup_act (rpriv, data); +default : +return -EOPNOTSUPP ; + } +} + static int mlx5e_rep_indr_setup_cb(struct net_device *netdev, struct Qdisc *sch, void *cb_priv, enum tc_setup_type type, void *type_data,
@@ -518,7 +633,7 @@ int mlx5e_rep_indr_setup_cb(struct net_device *netdev, struct Qdisc *sch, void * void (*cleanup)(struct flow_block_cb *block_cb)) {if (!netdev) -return -EOPNOTSUPP ; +return mlx5e_rep_indr_no_dev_setup (cb_priv, type, data);switch (type) {case TC_SETUP_BLOCK :
drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/act.h +10 -0
@@ -50,6 +50,16 @@ struct mlx5e_tc_act {bool (*is_multi_table_act)(struct mlx5e_priv *priv,const struct flow_action_entry *act,struct mlx5_flow_attr *attr); + +int (*offload_action)(struct mlx5e_priv *priv, +struct flow_offload_action *fl_act, +struct flow_action_entry *act); + +int (*destroy_action)(struct mlx5e_priv *priv, +struct flow_offload_action *fl_act); + +int (*stats_action)(struct mlx5e_priv *priv, +struct flow_offload_action *fl_act); };struct mlx5e_tc_flow_action {
drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/police.c +93 -7
121 lines not on this page · raw message
drivers/net/ethernet/mellanox/mlx5/core/en/tc/meter.c +16 -0
@@ -561,3 +561,19 @@ mlx5e_flow_meters_cleanup(struct mlx5e_flow_meters *flow_meters)mlx5_core_dealloc_pd (flow_meters->mdev, flow_meters->pdn);kfree (flow_meters); } + +void +mlx5e_tc_meter_get_stats (struct mlx5e_flow_meter_handle *meter, +u64 *bytes,u64 *packets,u64 *drops,u64 *lastuse) +{ +u64 bytes1, packets1, lastuse1; +u64 bytes2, packets2, lastuse2; + +mlx5_fc_query_cached (meter->green_counter, &bytes1, &packets1, &lastuse1); +mlx5_fc_query_cached (meter->red_counter, &bytes2, &packets2, &lastuse2); + + *bytes = bytes1 + bytes2; + *packets = packets1 + packets2; + *drops = packets2; + *lastuse =max_t (u64, lastuse1, lastuse2); +}
drivers/net/ethernet/mellanox/mlx5/core/en/tc/meter.h +4 -0
@@ -67,4 +67,8 @@ mlx5e_flow_meters_init(struct mlx5e_priv *priv,void mlx5e_flow_meters_cleanup (struct mlx5e_flow_meters *flow_meters); +void +mlx5e_tc_meter_get_stats (struct mlx5e_flow_meter_handle *meter, +u64 *bytes,u64 *packets,u64 *drops,u64 *lastuse); +#endif /* __MLX5_EN_FLOW_METER_H__ */
Signature · 1 line
2.37.1
Jakub Kicinski
maintainerRe: [net-next 06/15] net/mlx5e: TC, Support tc action api for police
Quoted 5 lines · On Thu, 28 Jul 2022 13:57:19 -0700 Saeed Mahameed wrote:
From: Roi Dayan <[email protected]>
Add support for tc action api for police.
Offloading standalone police action without
a tc rule and reporting stats.
Do you already support shared actions? I don't see anything later in the series that'd allow the binding of rules.
The metering in this series is for specific flower flows or the entire port?
Simon, Baowen, would you be willing to look thru these patches to make sure the action sharing works as expected?
Re: [net-next 06/15] net/mlx5e: TC, Support tc action api for police
Quoted 10 lines · Jakub Kicinski
Quoted 5 lines · On Thu, 28 Jul 2022 13:57:19 -0700 Saeed Mahameed wrote:
From: Roi Dayan <[email protected]>
Add support for tc action api for police.
Offloading standalone police action without
a tc rule and reporting stats.Do you already support shared actions? I don't see anything later in the series that'd allow the binding of rules.
Quoted 6 lines
The metering in this series is for specific flower flows or the entire port?
Simon, Baowen, would you be willing to look thru these patches to make sure the action sharing works as expected?
Certainly, we will review them.
Simon Horman
reviewerRe: [net-next 06/15] net/mlx5e: TC, Support tc action api for police
Quoted 11 lines · Simon Horman
Quoted 10 lines · Jakub Kicinski
Quoted 5 lines · On Thu, 28 Jul 2022 13:57:19 -0700 Saeed Mahameed wrote:
From: Roi Dayan <[email protected]>
Add support for tc action api for police.
Offloading standalone police action without
a tc rule and reporting stats.Do you already support shared actions? I don't see anything later in the series that'd allow the binding of rules.
Quoted 8 lines
Quoted 6 lines
The metering in this series is for specific flower flows or the entire port?
Simon, Baowen, would you be willing to look thru these patches to make sure the action sharing works as expected?
Certainly, we will review them.
Hi Jakub,
my reading of things is that the handling of offload of police (meter) actions in flower rules by the mlx5 driver is such that it can handle offloading actions by index - actions that it would now be possible to add to hardware with this patch in place.
My reasoning assumes that mlx5e_tc_add_flow_meter() is called to offload police (meter) actions in flower rules. And that it calls mlx5e_tc_meter_get(), which can find actions based on an index.
I could, however, be mistaken as I have so much knowledge of the mlx5 driver. And rather than dive deeper I wanted to respond as above - I am mindful of the point we are in the development cycle.
I would be happy to dive deeper into this as a mater of priority if desired.
Jakub Kicinski
maintainerRe: [net-next 06/15] net/mlx5e: TC, Support tc action api for police
Quoted 9 lines · On Fri, 29 Jul 2022 17:51:05 +0100 Simon Horman wrote:
my reading of things is that the handling of offload of police (meter) actions in flower rules by the mlx5 driver is such that it can handle offloading actions by index - actions that it would now be possible to add to hardware with this patch in place.
My reasoning assumes that
mlx5e_tc_add_flow_meter()is called to offload police (meter) actions in flower rules. And that it callsmlx5e_tc_meter_get(), which can find actions based on an index.
Quoted 6 lines
I could, however, be mistaken as I have so much knowledge of the mlx5 driver. And rather than dive deeper I wanted to respond as above - I am mindful of the point we are in the development cycle.
I would be happy to dive deeper into this as a mater of priority if desired.
Thank you! No very deep dives necessary from my perspective, just wanted for the authors of the action offload API to look over.
RE: [net-next 06/15] net/mlx5e: TC, Support tc action api for police
Quoted 16 lines
Subject: Re: [net-next 06/15] net/mlx5e: TC, Support tc action api for police
Quoted 13 lines · On Fri, 29 Jul 2022 17:51:05 +0100 Simon Horman wrote:
my reading of things is that the handling of offload of police (meter) actions in flower rules by the mlx5 driver is such that it can handle offloading actions by index - actions that it would now be possible to add to hardware with this patch in place.
My reasoning assumes that
mlx5e_tc_add_flow_meter()is called to offload police (meter) actions in flower rules. And that it callsmlx5e_tc_meter_get(), which can find actions based on an index.I could, however, be mistaken as I have so much knowledge of the mlx5 driver. And rather than dive deeper I wanted to respond as above - I am mindful of the point we are in the development cycle.
Quoted 5 lines
Quoted 2 lines
I would be happy to dive deeper into this as a mater of priority if desired.
Thank you! No very deep dives necessary from my perspective, just wanted for the authors of the action offload API to look over.
Hi Jakub, thanks for noticing us about this change, for this patch, it seems good to me.
I just have a tiny doubt, since Nvida just add a validation for a police offload, why it is not applied to this meter offload patch? Since I do not know much details about Nvdia meter implement, I guess not all the police action result is supported, so maybe validation is necessary.
Thanks
Re: [net-next 06/15] net/mlx5e: TC, Support tc action api for police
Quoted 19 lines · Baowen Zheng
Quoted 19 lines
Subject: Re: [net-next 06/15] net/mlx5e: TC, Support tc action api for police
Quoted 15 lines · On Fri, 29 Jul 2022 17:51:05 +0100 Simon Horman wrote:
my reading of things is that the handling of offload of police (meter) actions in flower rules by the mlx5 driver is such that it can handle offloading actions by index - actions that it would now be possible to add to hardware with this patch in place.
My reasoning assumes that
mlx5e_tc_add_flow_meter()is called to offload police (meter) actions in flower rules. And that it callsmlx5e_tc_meter_get(), which can find actions based on an index.I could, however, be mistaken as I have so much knowledge of the mlx5 driver. And rather than dive deeper I wanted to respond as above - I am mindful of the point we are in the development cycle.
I would be happy to dive deeper into this as a mater of priority if desired.
Quoted 7 lines
Quoted 2 lines
Thank you! No very deep dives necessary from my perspective, just wanted for the authors of the action offload API to look over.
Hi Jakub, thanks for noticing us about this change, for this patch, it seems good to me.
I just have a tiny doubt, since Nvida just add a validation for a police offload, why it is not applied to this meter offload patch? Since I do not know much details about Nvdia meter implement, I guess not all the police action result is supported, so maybe validation is necessary.Thanks
hi,
as Simon already mentioned. offloading action meter is done in mlx5e_tc_add_flow_meter() and adding a tc rule with meter will use mlx5e_tc_meter_get() to find the existing meter by index if exists.
as for Baowen Zhang comment. you are right. it seems I missed a validation part. since its merged I will send a fix commit.
thanks
PATCH 7/15 net/mlx5e: Convert mlx5e_tc_table member of mlx5e_flow_steering to pointer
In reply to Saeed MahameedFrom: Lama Kayal <[email protected]>
Make fs.tc be a pointer and allocate it dynamically.
Add mlx5e_priv pointer to mlx5e_tc_table, and thus get a work-around to accessing priv via tc when handling tc events inside mlx5e_tc_netdev_event.
--- .../net/ethernet/mellanox/mlx5/core/en/fs.h | 3 +- .../ethernet/mellanox/mlx5/core/en/tc_priv.h | 2 +- .../net/ethernet/mellanox/mlx5/core/en_fs.c | 11 +++- .../net/ethernet/mellanox/mlx5/core/en_tc.c | 61 +++++++++---------- 4 files changed, 43 insertions(+), 34 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/fs.h +2 -1
@@ -20,6 +20,7 @@ struct mlx5e_tc_table {* which is the nic tc root table. */ struct mutex t_lock; +struct mlx5e_priv *priv;struct mlx5_flow_table *t;struct mlx5_flow_table *miss_t;struct mlx5_fs_chains *chains;
@@ -169,7 +170,7 @@ struct mlx5e_flow_steering {#ifdef CONFIG_MLX5_EN_RXNFC struct mlx5e_ethtool_steering ethtool;#endif -struct mlx5e_tc_table tc; +struct mlx5e_tc_table *tc;struct mlx5e_promisc_table promisc;struct mlx5e_vlan_table *vlan;struct mlx5e_l2_table l2;
drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h +1 -1
@@ -11,7 +11,7 @@#define MLX5E_TC_MAX_SPLITS 1 -#define mlx5e_nic_chains (priv) ((priv)->fs.tc.chains) +#define mlx5e_nic_chains (priv) ((priv)->fs.tc->chains)enum {MLX5E_TC_FLOW_FLAG_INGRESS =MLX5E_TC_FLAG_INGRESS_BIT ,
drivers/net/ethernet/mellanox/mlx5/core/en_fs.c +10 -1
@@ -1346,12 +1346,21 @@ int mlx5e_fs_init(struct mlx5e_priv *priv) { priv->fs.vlan =kvzalloc (sizeof (*priv->fs.vlan),GFP_KERNEL );if (!priv->fs.vlan) -return -ENOMEM ; + goto err; + priv->fs.tc =kvzalloc (sizeof (*priv->fs.tc),GFP_KERNEL ); +if (!priv->fs.tc) + goto err_free_vlan;return 0 ; +err_free_vlan: +kvfree (priv->fs.vlan); + priv->fs.vlan =NULL ; +err: +return -ENOMEM ; }void mlx5e_fs_cleanup (struct mlx5e_priv *priv) { +kvfree (priv->fs.tc);kvfree (priv->fs.vlan); priv->fs.vlan =NULL ; }
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +30 -31
214 lines not on this page · raw message
Signature · 1 line
2.37.1
PATCH 8/15 net/mlx5e: Make mlx5e_tc_table private
In reply to Saeed MahameedFrom: Lama Kayal <[email protected]>
Move mlx5e_tc_table struct to en_tc.c thus make it private. Introduce allocation and deallocation functions as part of the tc API to allow this switch smoothly.
Convert mlx5e_nic_chain() macro to a function of en_tc.c.
--- .../net/ethernet/mellanox/mlx5/core/en/fs.h | 24 --------- .../mellanox/mlx5/core/en/tc/act/goto.c | 2 +- .../ethernet/mellanox/mlx5/core/en/tc_priv.h | 3 +- .../net/ethernet/mellanox/mlx5/core/en_fs.c | 9 ++-- .../net/ethernet/mellanox/mlx5/core/en_tc.c | 49 +++++++++++++++++-- .../net/ethernet/mellanox/mlx5/core/en_tc.h | 4 ++ 6 files changed, 59 insertions(+), 32 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/fs.h +0 -24
@@ -15,30 +15,6 @@ enum { MLX5E_TC_MISS_LEVEL, }; -struct mlx5e_tc_table { -/* Protects the dynamic assignment of the t parameter -* which is the nic tc root table. -*/ -struct mutex t_lock; -struct mlx5e_priv *priv; -struct mlx5_flow_table *t; -struct mlx5_flow_table *miss_t; -struct mlx5_fs_chains *chains; -struct mlx5e_post_act *post_act; - -struct rhashtable ht; - -struct mod_hdr_tbl mod_hdr; -struct mutex hairpin_tbl_lock;/* protects hairpin_tbl */ - DECLARE_HASHTABLE(hairpin_tbl, 8); - -struct notifier_block netdevice_nb; -struct netdev_net_notifier netdevice_nn; - -struct mlx5_tc_ct_priv *ct; -struct mapping_ctx *mapping; -}; -struct mlx5e_flow_table {int num_groups;struct mlx5_flow_table *t;
drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/goto.c +1 -1
@@ -21,7 +21,7 @@ validate_goto_chain(struct mlx5e_priv *priv,u32 max_chain; esw = priv->mdev->priv.eswitch; - chains = is_esw ?esw_chains (esw) :mlx5e_nic_chains (priv); + chains = is_esw ?esw_chains (esw) :mlx5e_nic_chains (priv->fs.tc); max_chain =mlx5_chains_get_chain_range (chains); reformat_and_fwd = is_esw ? MLX5_CAP_ESW_FLOWTABLE_FDB(priv->mdev, reformat_and_fwd_to_table) :
drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h +2 -1
@@ -11,7 +11,6 @@#define MLX5E_TC_MAX_SPLITS 1 -#define mlx5e_nic_chains (priv) ((priv)->fs.tc->chains)enum {MLX5E_TC_FLOW_FLAG_INGRESS =MLX5E_TC_FLAG_INGRESS_BIT ,
@@ -44,6 +43,8 @@ struct mlx5e_tc_flow_parse_attr {struct mlx5e_tc_act_parse_state parse_state; }; +struct mlx5_fs_chains *mlx5e_nic_chains (struct mlx5e_tc_table *tc); +/* Helper struct for accessing a struct containing list_head array. * Containing struct * |- Helper array
drivers/net/ethernet/mellanox/mlx5/core/en_fs.c +6 -3
@@ -38,6 +38,7 @@#include <linux/mlx5/mpfs.h> #include "en.h" #include "en_rep.h" +#include "en_tc.h" #include "lib/mpfs.h" #include "en/ptp.h"
@@ -1347,9 +1348,11 @@ int mlx5e_fs_init(struct mlx5e_priv *priv) priv->fs.vlan =kvzalloc (sizeof (*priv->fs.vlan),GFP_KERNEL );if (!priv->fs.vlan) goto err; - priv->fs.tc =kvzalloc (sizeof (*priv->fs.tc),GFP_KERNEL ); -if (!priv->fs.tc) + + priv->fs.tc =mlx5e_tc_table_alloc (); +if (IS_ERR (priv->fs.tc)) goto err_free_vlan; +return 0 ; err_free_vlan:kvfree (priv->fs.vlan);
@@ -1360,7 +1363,7 @@ int mlx5e_fs_init(struct mlx5e_priv *priv)void mlx5e_fs_cleanup (struct mlx5e_priv *priv) { -kvfree (priv->fs.tc); +mlx5e_tc_table_free (priv->fs.tc);kvfree (priv->fs.vlan); priv->fs.vlan =NULL ; }
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +46 -3
88 lines not on this page · raw message
drivers/net/ethernet/mellanox/mlx5/core/en_tc.h +4 -0
@@ -356,6 +356,8 @@ mlx5e_setup_tc_block_cb(enum tc_setup_type type, void *type_data, void *cb_priv)#endif #if IS_ENABLED (CONFIG_MLX5_CLS_ACT ) +struct mlx5e_tc_table *mlx5e_tc_table_alloc (void ); +void mlx5e_tc_table_free (struct mlx5e_tc_table *tc);static inline bool mlx5e_cqe_regb_chain (struct mlx5_cqe64 *cqe) {#if IS_ENABLED (CONFIG_NET_TC_SKB_EXT )
@@ -376,6 +378,8 @@ static inline bool mlx5e_cqe_regb_chain(struct mlx5_cqe64 *cqe)bool mlx5e_tc_update_skb (struct mlx5_cqe64 *cqe,struct sk_buff *skb);#else /* CONFIG_MLX5_CLS_ACT */ +static inline struct mlx5e_tc_table *mlx5e_tc_table_alloc (void ) {return NULL ; } +static inline void mlx5e_tc_table_free (struct mlx5e_tc_table *tc) {}static inline bool mlx5e_cqe_regb_chain (struct mlx5_cqe64 *cqe) {return false; } static inline bool
Signature · 1 line
2.37.1
PATCH 9/15 net/mlx5e: Allocate VLAN and TC for featured profiles only
In reply to Saeed MahameedFrom: Lama Kayal <[email protected]>
Introduce allocation and de-allocation functions for both flow steering VLAN and TC as part of fs API.
Add allocations of VLAN and TC as nic profile feature, such that fs_init() will allocate both VLAN and TC only if they're featured in the profile. VLAN and TC are relevant for nic_profile only.
--- drivers/net/ethernet/mellanox/mlx5/core/en.h | 2 + .../net/ethernet/mellanox/mlx5/core/en_fs.c | 52 +++++++++++++++---- .../net/ethernet/mellanox/mlx5/core/en_main.c | 4 +- 3 files changed, 46 insertions(+), 12 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en.h +2 -0
@@ -987,6 +987,8 @@ enum mlx5e_profile_feature {MLX5E_PROFILE_FEATURE_PTP_RX ,MLX5E_PROFILE_FEATURE_PTP_TX ,MLX5E_PROFILE_FEATURE_QOS_HTB , +MLX5E_PROFILE_FEATURE_FS_VLAN , + MLX5E_PROFILE_FEATURE_FS_TC, }; struct mlx5e_profile {
drivers/net/ethernet/mellanox/mlx5/core/en_fs.c +41 -11
68 lines not on this page · raw message
drivers/net/ethernet/mellanox/mlx5/core/en_main.c +3 -1
@@ -5242,7 +5242,9 @@ static const struct mlx5e_profile mlx5e_nic_profile = { .stats_grps_num = mlx5e_nic_stats_grps_num, .features =BIT (MLX5E_PROFILE_FEATURE_PTP_RX ) |BIT (MLX5E_PROFILE_FEATURE_PTP_TX ) | - BIT(MLX5E_PROFILE_FEATURE_QOS_HTB), +BIT (MLX5E_PROFILE_FEATURE_QOS_HTB ) | +BIT (MLX5E_PROFILE_FEATURE_FS_VLAN ) | + BIT(MLX5E_PROFILE_FEATURE_FS_TC), }; static int mlx5e_profile_max_num_channels(struct mlx5_core_dev *mdev,
Signature · 1 line
2.37.1
PATCH 10/15 net/mlx5e: Convert mlx5e_flow_steering member of mlx5e_priv to pointer
In reply to Saeed MahameedFrom: Lama Kayal <[email protected]>
Make mlx5e_flow_steering member of mlx5e_priv a pointer. Add dynamic allocation respectively.
Allocate fs for all profiles when initializing profile, symmetrically deallocate at profile cleanup.
--- drivers/net/ethernet/mellanox/mlx5/core/en.h | 2 +- .../net/ethernet/mellanox/mlx5/core/en/fs.h | 5 +- .../mellanox/mlx5/core/en/fs_tt_redirect.c | 72 +++--- .../net/ethernet/mellanox/mlx5/core/en/ptp.c | 8 +- .../mellanox/mlx5/core/en/tc/act/goto.c | 2 +- .../mellanox/mlx5/core/en_accel/fs_tcp.c | 32 +-- .../mellanox/mlx5/core/en_accel/ipsec_fs.c | 10 +- .../net/ethernet/mellanox/mlx5/core/en_arfs.c | 58 ++--- .../net/ethernet/mellanox/mlx5/core/en_fs.c | 226 +++++++++--------- .../mellanox/mlx5/core/en_fs_ethtool.c | 24 +- .../net/ethernet/mellanox/mlx5/core/en_main.c | 13 +- .../net/ethernet/mellanox/mlx5/core/en_rep.c | 31 ++- .../net/ethernet/mellanox/mlx5/core/en_tc.c | 60 ++--- .../ethernet/mellanox/mlx5/core/ipoib/ipoib.c | 19 +- 14 files changed, 298 insertions(+), 264 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en.h +1 -1
@@ -921,7 +921,7 @@ struct mlx5e_priv {struct mlx5e_rx_res *rx_res;u32 *tx_rates; -struct mlx5e_flow_steering fs; +struct mlx5e_flow_steering *fs;struct workqueue_struct *wq;struct work_struct update_carrier_work;
drivers/net/ethernet/mellanox/mlx5/core/en/fs.h +3 -2
@@ -137,6 +137,7 @@ static inline int mlx5e_arfs_disable(struct mlx5e_priv *priv) { return -EOPNOTSUstruct mlx5e_accel_fs_tcp ;#endif +struct mlx5e_profile ;struct mlx5e_fs_udp ;struct mlx5e_fs_any ;struct mlx5e_ptp_fs ;
@@ -177,8 +178,8 @@ void mlx5e_disable_cvlan_filter(struct mlx5e_priv *priv);int mlx5e_create_flow_steering (struct mlx5e_priv *priv);void mlx5e_destroy_flow_steering (struct mlx5e_priv *priv); -int mlx5e_fs_init (struct mlx5e_priv *priv); -void mlx5e_fs_cleanup (struct mlx5e_priv *priv); +struct mlx5e_flow_steering *mlx5e_fs_init (const struct mlx5e_profile *profile); +void mlx5e_fs_cleanup (struct mlx5e_flow_steering *fs);int mlx5e_add_vlan_trap (struct mlx5e_priv *priv,int trap_id,int tir_num);void mlx5e_remove_vlan_trap (struct mlx5e_priv *priv);
drivers/net/ethernet/mellanox/mlx5/core/en/fs_tt_redirect.c +36 -36
216 lines not on this page · raw message
drivers/net/ethernet/mellanox/mlx5/core/en/ptp.c +4 -4
@@ -624,7 +624,7 @@ static int mlx5e_ptp_set_state(struct mlx5e_ptp *c, struct mlx5e_params *params)static void mlx5e_ptp_rx_unset_fs (struct mlx5e_priv *priv) { -struct mlx5e_ptp_fs *ptp_fs = priv->fs.ptp_fs; +struct mlx5e_ptp_fs *ptp_fs = priv->fs->ptp_fs;if (!ptp_fs->valid)return ;
@@ -641,7 +641,7 @@ static void mlx5e_ptp_rx_unset_fs(struct mlx5e_priv *priv)static int mlx5e_ptp_rx_set_fs (struct mlx5e_priv *priv) {u32 tirn =mlx5e_rx_res_get_tirn_ptp (priv->rx_res); -struct mlx5e_ptp_fs *ptp_fs = priv->fs.ptp_fs; +struct mlx5e_ptp_fs *ptp_fs = priv->fs->ptp_fs;struct mlx5_flow_handle *rule;int err;
@@ -808,13 +808,13 @@ int mlx5e_ptp_alloc_rx_fs(struct mlx5e_priv *priv)if (!ptp_fs)return -ENOMEM ; - priv->fs.ptp_fs = ptp_fs; + priv->fs->ptp_fs = ptp_fs;return 0 ; }void mlx5e_ptp_free_rx_fs (struct mlx5e_priv *priv) { -struct mlx5e_ptp_fs *ptp_fs = priv->fs.ptp_fs; +struct mlx5e_ptp_fs *ptp_fs = priv->fs->ptp_fs;if (!mlx5e_profile_feature_cap (priv->profile,PTP_RX ))return ;
drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/goto.c +1 -1
@@ -21,7 +21,7 @@ validate_goto_chain(struct mlx5e_priv *priv,u32 max_chain; esw = priv->mdev->priv.eswitch; - chains = is_esw ?esw_chains (esw) :mlx5e_nic_chains (priv->fs.tc); + chains = is_esw ?esw_chains (esw) :mlx5e_nic_chains (priv->fs->tc); max_chain =mlx5_chains_get_chain_range (chains); reformat_and_fwd = is_esw ? MLX5_CAP_ESW_FLOWTABLE_FDB(priv->mdev, reformat_and_fwd_to_table) :
drivers/net/ethernet/mellanox/mlx5/core/en_accel/fs_tcp.c +16 -16
@@ -86,7 +86,7 @@ struct mlx5_flow_handle *mlx5e_accel_fs_add_sk(struct mlx5e_priv *priv,if (!spec)return ERR_PTR (-ENOMEM ); - fs_tcp = priv->fs.accel_tcp; + fs_tcp = priv->fs->accel_tcp; spec->match_criteria_enable =MLX5_MATCH_OUTER_HEADERS ;
@@ -158,10 +158,10 @@ static int accel_fs_tcp_add_default_rule(struct mlx5e_priv *priv,struct mlx5_flow_handle *rule;int err =0 ; - fs_tcp = priv->fs.accel_tcp; + fs_tcp = priv->fs->accel_tcp; accel_fs_t = &fs_tcp->tables[type]; - dest =mlx5_ttc_get_default_dest (priv->fs.ttc,fs_accel2tt (type)); + dest =mlx5_ttc_get_default_dest (priv->fs->ttc,fs_accel2tt (type)); rule =mlx5_add_flow_rules (accel_fs_t->t,NULL , &flow_act, &dest,1 );if (IS_ERR (rule)) { err =PTR_ERR (rule);
@@ -267,7 +267,7 @@ static int accel_fs_tcp_create_groups(struct mlx5e_flow_table *ft,static int accel_fs_tcp_create_table (struct mlx5e_priv *priv,enum accel_fs_tcp_type type) { -struct mlx5e_flow_table *ft = &priv->fs.accel_tcp->tables[type]; +struct mlx5e_flow_table *ft = &priv->fs->accel_tcp->tables[type];struct mlx5_flow_table_attr ft_attr = {};int err;
@@ -277,7 +277,7 @@ static int accel_fs_tcp_create_table(struct mlx5e_priv *priv, enum accel_fs_tcp_ ft_attr.level =MLX5E_ACCEL_FS_TCP_FT_LEVEL ; ft_attr.prio =MLX5E_NIC_PRIO ; - ft->t =mlx5_create_flow_table (priv->fs.ns, &ft_attr); + ft->t =mlx5_create_flow_table (priv->fs->ns, &ft_attr);if (IS_ERR (ft->t)) { err =PTR_ERR (ft->t); ft->t =NULL ;
@@ -307,7 +307,7 @@ static int accel_fs_tcp_disable(struct mlx5e_priv *priv)for (i =0 ; i <ACCEL_FS_TCP_NUM_TYPES ; i++) {/* Modify ttc rules destination to point back to the indir TIRs */ - err =mlx5_ttc_fwd_default_dest (priv->fs.ttc,fs_accel2tt (i)); + err =mlx5_ttc_fwd_default_dest (priv->fs->ttc,fs_accel2tt (i));if (err) { netdev_err(priv->netdev, "%s: modify ttc[%d] default destination failed, err(%d)\n",
@@ -326,10 +326,10 @@ static int accel_fs_tcp_enable(struct mlx5e_priv *priv) dest.type =MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE ;for (i =0 ; i <ACCEL_FS_TCP_NUM_TYPES ; i++) { - dest.ft = priv->fs.accel_tcp->tables[i].t; + dest.ft = priv->fs->accel_tcp->tables[i].t;/* Modify ttc rules destination to point on the accel_fs FTs */ - err =mlx5_ttc_fwd_dest (priv->fs.ttc,fs_accel2tt (i), &dest); + err =mlx5_ttc_fwd_dest (priv->fs->ttc,fs_accel2tt (i), &dest);if (err) { netdev_err(priv->netdev, "%s: modify ttc[%d] destination to accel failed, err(%d)\n",
@@ -344,7 +344,7 @@ static void accel_fs_tcp_destroy_table(struct mlx5e_priv *priv, int i) {struct mlx5e_accel_fs_tcp *fs_tcp; - fs_tcp = priv->fs.accel_tcp; + fs_tcp = priv->fs->accel_tcp;if (IS_ERR_OR_NULL (fs_tcp->tables[i].t))return ;
@@ -357,7 +357,7 @@ void mlx5e_accel_fs_tcp_destroy(struct mlx5e_priv *priv) {int i; -if (!priv->fs.accel_tcp) +if (!priv->fs->accel_tcp)return ;accel_fs_tcp_disable (priv);
@@ -365,8 +365,8 @@ void mlx5e_accel_fs_tcp_destroy(struct mlx5e_priv *priv)for (i =0 ; i <ACCEL_FS_TCP_NUM_TYPES ; i++)accel_fs_tcp_destroy_table (priv, i); -kfree (priv->fs.accel_tcp); - priv->fs.accel_tcp =NULL ; +kfree (priv->fs->accel_tcp); + priv->fs->accel_tcp =NULL ; }int mlx5e_accel_fs_tcp_create (struct mlx5e_priv *priv)
@@ -376,8 +376,8 @@ int mlx5e_accel_fs_tcp_create(struct mlx5e_priv *priv)if (!MLX5_CAP_FLOWTABLE_NIC_RX (priv->mdev, ft_field_support.outer_ip_version))return -EOPNOTSUPP ; - priv->fs.accel_tcp =kzalloc (sizeof (*priv->fs.accel_tcp),GFP_KERNEL ); -if (!priv->fs.accel_tcp) + priv->fs->accel_tcp =kzalloc (sizeof (*priv->fs->accel_tcp),GFP_KERNEL ); +if (!priv->fs->accel_tcp)return -ENOMEM ;for (i =0 ; i <ACCEL_FS_TCP_NUM_TYPES ; i++) {
@@ -396,7 +396,7 @@ int mlx5e_accel_fs_tcp_create(struct mlx5e_priv *priv)while (--i >=0 )accel_fs_tcp_destroy_table (priv, i); -kfree (priv->fs.accel_tcp); - priv->fs.accel_tcp =NULL ; +kfree (priv->fs->accel_tcp); + priv->fs->accel_tcp =NULL ;return err; }
drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c +5 -5
@@ -184,13 +184,13 @@ static int rx_create(struct mlx5e_priv *priv, enum accel_fs_esp_type type) fs_prot = &accel_esp->fs_prot[type]; fs_prot->default_dest = -mlx5_ttc_get_default_dest (priv->fs.ttc,fs_esp2tt (type)); +mlx5_ttc_get_default_dest (priv->fs->ttc,fs_esp2tt (type)); ft_attr.max_fte =1 ; ft_attr.autogroup.max_num_groups =1 ; ft_attr.level =MLX5E_ACCEL_FS_ESP_FT_ERR_LEVEL ; ft_attr.prio =MLX5E_NIC_PRIO ; - ft =mlx5_create_auto_grouped_flow_table (priv->fs.ns, &ft_attr); + ft =mlx5_create_auto_grouped_flow_table (priv->fs->ns, &ft_attr);if (IS_ERR (ft))return PTR_ERR (ft);
@@ -205,7 +205,7 @@ static int rx_create(struct mlx5e_priv *priv, enum accel_fs_esp_type type) ft_attr.prio =MLX5E_NIC_PRIO ; ft_attr.autogroup.num_reserved_entries =1 ; ft_attr.autogroup.max_num_groups =1 ; - ft =mlx5_create_auto_grouped_flow_table (priv->fs.ns, &ft_attr); + ft =mlx5_create_auto_grouped_flow_table (priv->fs->ns, &ft_attr);if (IS_ERR (ft)) { err =PTR_ERR (ft); goto err_fs_ft;
@@ -249,7 +249,7 @@ static int rx_ft_get(struct mlx5e_priv *priv, enum accel_fs_esp_type type)/* connect */ dest.type =MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE ; dest.ft = fs_prot->ft; -mlx5_ttc_fwd_dest (priv->fs.ttc,fs_esp2tt (type), &dest); +mlx5_ttc_fwd_dest (priv->fs->ttc,fs_esp2tt (type), &dest); skip: fs_prot->refcnt++;
@@ -271,7 +271,7 @@ static void rx_ft_put(struct mlx5e_priv *priv, enum accel_fs_esp_type type) goto out;/* disconnect */ -mlx5_ttc_fwd_default_dest (priv->fs.ttc,fs_esp2tt (type)); +mlx5_ttc_fwd_default_dest (priv->fs->ttc,fs_esp2tt (type));/* remove FT */ rx_destroy (priv, type);
drivers/net/ethernet/mellanox/mlx5/core/en_arfs.c +29 -29
167 lines not on this page · raw message
drivers/net/ethernet/mellanox/mlx5/core/en_fs.c +117 -109
604 lines not on this page · raw message
drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c +12 -12
@@ -81,18 +81,18 @@ static struct mlx5e_ethtool_table *get_flow_table(struct mlx5e_priv *priv,case UDP_V6_FLOW : max_tuples =ETHTOOL_NUM_L3_L4_FTS ; prio =MLX5E_ETHTOOL_L3_L4_PRIO + (max_tuples - num_tuples); - eth_ft = &priv->fs.ethtool.l3_l4_ft[prio]; + eth_ft = &priv->fs->ethtool.l3_l4_ft[prio];break ;case IP_USER_FLOW :case IPV6_USER_FLOW : max_tuples =ETHTOOL_NUM_L3_L4_FTS ; prio =MLX5E_ETHTOOL_L3_L4_PRIO + (max_tuples - num_tuples); - eth_ft = &priv->fs.ethtool.l3_l4_ft[prio]; + eth_ft = &priv->fs->ethtool.l3_l4_ft[prio];break ;case ETHER_FLOW : max_tuples =ETHTOOL_NUM_L2_FTS ; prio = max_tuples - num_tuples; - eth_ft = &priv->fs.ethtool.l2_ft[prio]; + eth_ft = &priv->fs->ethtool.l2_ft[prio]; prio +=MLX5E_ETHTOOL_L2_PRIO ;break ;default :
@@ -383,14 +383,14 @@ static void add_rule_to_list(struct mlx5e_priv *priv, struct mlx5e_ethtool_rule *rule) {struct mlx5e_ethtool_rule *iter; -struct list_head *head = &priv->fs.ethtool.rules; +struct list_head *head = &priv->fs->ethtool.rules; -list_for_each_entry (iter, &priv->fs.ethtool.rules, list) { +list_for_each_entry (iter, &priv->fs->ethtool.rules, list) {if (iter->flow_spec.location > rule->flow_spec.location)break ; head = &iter->list; } - priv->fs.ethtool.tot_num_rules++; + priv->fs->ethtool.tot_num_rules++;list_add (&rule->list, head); }
@@ -507,7 +507,7 @@ static void del_ethtool_rule(struct mlx5e_priv *priv,if (eth_rule->rss)mlx5e_rss_refcnt_dec (eth_rule->rss);list_del (ð_rule->list); - priv->fs.ethtool.tot_num_rules--; + priv->fs->ethtool.tot_num_rules--;put_flow_table (eth_rule->eth_ft);kfree (eth_rule); }
@@ -517,7 +517,7 @@ static struct mlx5e_ethtool_rule *find_ethtool_rule(struct mlx5e_priv *priv, {struct mlx5e_ethtool_rule *iter; -list_for_each_entry (iter, &priv->fs.ethtool.rules, list) { +list_for_each_entry (iter, &priv->fs->ethtool.rules, list) {if (iter->flow_spec.location == location)return iter; }
@@ -788,7 +788,7 @@ mlx5e_ethtool_get_flow(struct mlx5e_priv *priv,if (location <0 || location >=MAX_NUM_OF_ETHTOOL_RULES )return -EINVAL ; -list_for_each_entry (eth_rule, &priv->fs.ethtool.rules, list) { +list_for_each_entry (eth_rule, &priv->fs->ethtool.rules, list) {int index;if (eth_rule->flow_spec.location != location)
@@ -831,13 +831,13 @@ void mlx5e_ethtool_cleanup_steering(struct mlx5e_priv *priv)struct mlx5e_ethtool_rule *iter;struct mlx5e_ethtool_rule *temp; -list_for_each_entry_safe (iter, temp, &priv->fs.ethtool.rules, list) +list_for_each_entry_safe (iter, temp, &priv->fs->ethtool.rules, list)del_ethtool_rule (priv, iter); }void mlx5e_ethtool_init_steering (struct mlx5e_priv *priv) { -INIT_LIST_HEAD (&priv->fs.ethtool.rules); +INIT_LIST_HEAD (&priv->fs->ethtool.rules); }static int flow_type_to_traffic_type (u32 flow_type)
@@ -963,7 +963,7 @@ int mlx5e_ethtool_get_rxnfc(struct mlx5e_priv *priv,switch (info->cmd) {case ETHTOOL_GRXCLSRLCNT : - info->rule_cnt = priv->fs.ethtool.tot_num_rules; + info->rule_cnt = priv->fs->ethtool.tot_num_rules;break ;case ETHTOOL_GRXCLSRULE : err =mlx5e_ethtool_get_flow (priv, info, info->fs.location);
drivers/net/ethernet/mellanox/mlx5/core/en_main.c +8 -5
@@ -3888,8 +3888,8 @@ static netdev_features_t mlx5e_fix_features(struct net_device *netdev,mutex_lock (&priv->state_lock); params = &priv->channels.params; -if (!priv->fs.vlan || - !bitmap_empty (mlx5e_vlan_get_active_svlans (priv->fs.vlan),VLAN_N_VID )) { +if (!priv->fs->vlan || + !bitmap_empty (mlx5e_vlan_get_active_svlans (priv->fs->vlan),VLAN_N_VID )) {/* HW strips the outer C-tag header, this is a problem * for S-tag traffic. */
@@ -5012,6 +5012,7 @@ static int mlx5e_nic_init(struct mlx5_core_dev *mdev, struct net_device *netdev) {struct mlx5e_priv *priv =netdev_priv (netdev); +struct mlx5e_flow_steering *fs;int err;mlx5e_build_nic_params (priv, &priv->xsk, netdev->mtu);
@@ -5019,11 +5020,13 @@ static int mlx5e_nic_init(struct mlx5_core_dev *mdev,mlx5e_timestamp_init (priv); - err =mlx5e_fs_init (priv); -if (err) { + fs =mlx5e_fs_init (priv->profile); +if (!fs) { + err = -ENOMEM ;mlx5_core_err (mdev,"FS initialization failed, %d\n" , err);return err; } + priv->fs = fs; err =mlx5e_ipsec_init (priv);if (err)
@@ -5042,7 +5045,7 @@ static void mlx5e_nic_cleanup(struct mlx5e_priv *priv)mlx5e_health_destroy_reporters (priv);mlx5e_ktls_cleanup (priv);mlx5e_ipsec_cleanup (priv); -mlx5e_fs_cleanup (priv); +mlx5e_fs_cleanup (priv->fs); }static int mlx5e_init_nic_rx (struct mlx5e_priv *priv)
drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +21 -10
@@ -718,6 +718,7 @@ static int mlx5e_init_ul_rep(struct mlx5_core_dev *mdev,static void mlx5e_cleanup_rep (struct mlx5e_priv *priv) { +mlx5e_fs_cleanup (priv->fs);mlx5e_ipsec_cleanup (priv); }
@@ -728,8 +729,8 @@ static int mlx5e_create_rep_ttc_table(struct mlx5e_priv *priv)struct ttc_params ttc_params = {};int err; - priv->fs.ns =mlx5_get_flow_namespace (priv->mdev, -MLX5_FLOW_NAMESPACE_KERNEL ); + priv->fs->ns =mlx5_get_flow_namespace (priv->mdev, +MLX5_FLOW_NAMESPACE_KERNEL );/* The inner_ttc in the ttc params is intentionally not set */ mlx5e_set_ttc_params (priv, &ttc_params, false);
@@ -738,9 +739,9 @@ static int mlx5e_create_rep_ttc_table(struct mlx5e_priv *priv)/* To give uplik rep TTC a lower level for chaining from root ft */ ttc_params.ft_attr.level =MLX5E_TTC_FT_LEVEL +1 ; - priv->fs.ttc =mlx5_create_ttc_table (priv->mdev, &ttc_params); -if (IS_ERR (priv->fs.ttc)) { - err =PTR_ERR (priv->fs.ttc); + priv->fs->ttc =mlx5_create_ttc_table (priv->mdev, &ttc_params); +if (IS_ERR (priv->fs->ttc)) { + err =PTR_ERR (priv->fs->ttc);netdev_err (priv->netdev,"Failed to create rep ttc table, err=%d\n" , err);return err;
@@ -760,7 +761,7 @@ static int mlx5e_create_rep_root_ft(struct mlx5e_priv *priv)/* non uplik reps will skip any bypass tables and go directly to * their own ttc */ - rpriv->root_ft =mlx5_get_ttc_flow_table (priv->fs.ttc); + rpriv->root_ft =mlx5_get_ttc_flow_table (priv->fs->ttc);return 0 ; }
@@ -835,9 +836,17 @@ static int mlx5e_init_rep_rx(struct mlx5e_priv *priv)struct mlx5_core_dev *mdev = priv->mdev;int err; - priv->rx_res =mlx5e_rx_res_alloc (); -if (!priv->rx_res) + priv->fs =mlx5e_fs_init (priv->profile); +if (!priv->fs) { +netdev_err (priv->netdev,"FS allocation failed\n" );return -ENOMEM ; + } + + priv->rx_res =mlx5e_rx_res_alloc (); +if (!priv->rx_res) { + err = -ENOMEM ; + goto err_free_fs; + }mlx5e_init_l2_addr (priv);
@@ -873,13 +882,15 @@ static int mlx5e_init_rep_rx(struct mlx5e_priv *priv) err_destroy_root_ft:mlx5e_destroy_rep_root_ft (priv); err_destroy_ttc_table: -mlx5_destroy_ttc_table (priv->fs.ttc); +mlx5_destroy_ttc_table (priv->fs->ttc); err_destroy_rx_res:mlx5e_rx_res_destroy (priv->rx_res); err_close_drop_rq:mlx5e_close_drop_rq (&priv->drop_rq);mlx5e_rx_res_free (priv->rx_res); priv->rx_res =NULL ; +err_free_fs: +mlx5e_fs_cleanup (priv->fs);return err; }
@@ -888,7 +899,7 @@ static void mlx5e_cleanup_rep_rx(struct mlx5e_priv *priv)mlx5e_ethtool_cleanup_steering (priv);rep_vport_rx_rule_destroy (priv);mlx5e_destroy_rep_root_ft (priv); -mlx5_destroy_ttc_table (priv->fs.ttc); +mlx5_destroy_ttc_table (priv->fs->ttc);mlx5e_rx_res_destroy (priv->rx_res);mlx5e_close_drop_rq (&priv->drop_rq);mlx5e_rx_res_free (priv->rx_res);
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +30 -30
207 lines not on this page · raw message
drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c +15 -4
@@ -322,10 +322,10 @@ static int mlx5i_create_flow_steering(struct mlx5e_priv *priv) {int err; - priv->fs.ns =mlx5_get_flow_namespace (priv->mdev, + priv->fs->ns =mlx5_get_flow_namespace (priv->mdev,MLX5_FLOW_NAMESPACE_KERNEL ); -if (!priv->fs.ns) +if (!priv->fs->ns)return -EINVAL ; err =mlx5e_arfs_create_tables (priv);
@@ -364,9 +364,17 @@ static int mlx5i_init_rx(struct mlx5e_priv *priv)struct mlx5_core_dev *mdev = priv->mdev;int err; - priv->rx_res =mlx5e_rx_res_alloc (); -if (!priv->rx_res) + priv->fs =mlx5e_fs_init (priv->profile); +if (!priv->fs) { +netdev_err (priv->netdev,"FS allocation failed\n" );return -ENOMEM ; + } + + priv->rx_res =mlx5e_rx_res_alloc (); +if (!priv->rx_res) { + err = -ENOMEM ; + goto err_free_fs; + }mlx5e_create_q_counters (priv);
@@ -397,6 +405,8 @@ static int mlx5i_init_rx(struct mlx5e_priv *priv)mlx5e_destroy_q_counters (priv);mlx5e_rx_res_free (priv->rx_res); priv->rx_res =NULL ; +err_free_fs: +mlx5e_fs_cleanup (priv->fs);return err; }
@@ -408,6 +418,7 @@ static void mlx5i_cleanup_rx(struct mlx5e_priv *priv)mlx5e_destroy_q_counters (priv);mlx5e_rx_res_free (priv->rx_res); priv->rx_res =NULL ; +mlx5e_fs_cleanup (priv->fs); }/* The stats groups order is opposite to the update_stats() order calls */
Signature · 1 line
2.37.1
PATCH 11/15 net/mlx5e: Report flow steering errors with mdev err report API
In reply to Saeed MahameedFrom: Lama Kayal <[email protected]>
Let en_fs report errors via mdev error report API, aka mlx5_core_* macros, thus replace the netdev API reports.
This to minimize netdev coupling to the flow steering struct.
--- .../net/ethernet/mellanox/mlx5/core/en_fs.c | 70 +++++++++---------- 1 file changed, 34 insertions(+), 36 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en_fs.c +34 -36
@@ -135,7 +135,6 @@ struct mlx5_flow_table *mlx5e_vlan_get_flowtable(struct mlx5e_vlan_table *vlan)static int mlx5e_vport_context_update_vlans (struct mlx5e_priv *priv) { -struct net_device *ndev = priv->netdev;int max_list_size;int list_size;u16 *vlans;
@@ -150,9 +149,9 @@ static int mlx5e_vport_context_update_vlans(struct mlx5e_priv *priv) max_list_size =1 <<MLX5_CAP_GEN (priv->mdev, log_max_vlan_list);if (list_size > max_list_size) { -netdev_warn (ndev, -"netdev vlans list size (%d) > (%d) max vport list size, some vlans will be dropped\n" , - list_size, max_list_size); +mlx5_core_warn (priv->mdev, +"netdev vlans list size (%d) > (%d) max vport list size, some vlans will be dropped\n" , + list_size, max_list_size); list_size = max_list_size; }
@@ -169,8 +168,8 @@ static int mlx5e_vport_context_update_vlans(struct mlx5e_priv *priv) err =mlx5_modify_nic_vport_vlans (priv->mdev, vlans, list_size);if (err) -netdev_err (ndev,"Failed to modify vport vlans list err(%d)\n" , - err); +mlx5_core_err (priv->mdev,"Failed to modify vport vlans list err(%d)\n" , + err);kvfree (vlans);return err;
@@ -251,7 +250,7 @@ static int __mlx5e_add_vlan_rule(struct mlx5e_priv *priv,if (IS_ERR (*rule_p)) { err =PTR_ERR (*rule_p); *rule_p =NULL ; -netdev_err (priv->netdev,"%s: add rule failed\n" , __func__); +mlx5_core_err (priv->mdev,"%s: add rule failed\n" , __func__); }return err;
@@ -363,8 +362,8 @@ int mlx5e_add_vlan_trap(struct mlx5e_priv *priv, int trap_id, int tir_num)if (IS_ERR (rule)) { err =PTR_ERR (rule); priv->fs->vlan->trap_rule =NULL ; -netdev_err (priv->netdev,"%s: add VLAN trap rule failed, err %d\n" , - __func__, err); +mlx5_core_err (priv->mdev,"%s: add VLAN trap rule failed, err %d\n" , + __func__, err);return err; } priv->fs->vlan->trap_rule = rule;
@@ -389,8 +388,8 @@ int mlx5e_add_mac_trap(struct mlx5e_priv *priv, int trap_id, int tir_num)if (IS_ERR (rule)) { err =PTR_ERR (rule); priv->fs->l2.trap_rule =NULL ; -netdev_err (priv->netdev,"%s: add MAC trap rule failed, err %d\n" , - __func__, err); +mlx5_core_err (priv->mdev,"%s: add MAC trap rule failed, err %d\n" , + __func__, err);return err; } priv->fs->l2.trap_rule = rule;
@@ -565,8 +564,8 @@ static void mlx5e_execute_l2_action(struct mlx5e_priv *priv, }if (l2_err) -netdev_warn (priv->netdev,"MPFS, failed to %s mac %pM, err(%d)\n" , - action ==MLX5E_ACTION_ADD ?"add" :"del" , mac_addr, l2_err); +mlx5_core_warn (priv->mdev,"MPFS, failed to %s mac %pM, err(%d)\n" , + action ==MLX5E_ACTION_ADD ?"add" :"del" , mac_addr, l2_err); }static void mlx5e_sync_netdev_addr (struct mlx5e_priv *priv)
@@ -638,9 +637,9 @@ static void mlx5e_vport_context_update_addr_list(struct mlx5e_priv *priv, size++;if (size > max_size) { -netdev_warn (priv->netdev, -"netdev %s list size (%d) > (%d) max vport list size, some addresses will be dropped\n" , - is_uc ?"UC" :"MC" , size, max_size); +mlx5_core_warn (priv->mdev, +"mdev %s list size (%d) > (%d) max vport list size, some addresses will be dropped\n" , + is_uc ?"UC" :"MC" , size, max_size); size = max_size; }
@@ -656,9 +655,9 @@ static void mlx5e_vport_context_update_addr_list(struct mlx5e_priv *priv, err =mlx5_modify_nic_vport_mac_list (priv->mdev, list_type, addr_array, size); out:if (err) -netdev_err (priv->netdev, -"Failed to modify vport %s list err(%d)\n" , - is_uc ?"UC" :"MC" , err); +mlx5_core_err (priv->mdev, +"Failed to modify vport %s list err(%d)\n" , + is_uc ?"UC" :"MC" , err);kfree (addr_array); }
@@ -726,7 +725,7 @@ static int mlx5e_add_promisc_rule(struct mlx5e_priv *priv)if (IS_ERR (*rule_p)) { err =PTR_ERR (*rule_p); *rule_p =NULL ; -netdev_err (priv->netdev,"%s: add promiscuous rule failed\n" , __func__); +mlx5_core_err (priv->mdev,"%s: add promiscuous rule failed\n" , __func__); }kvfree (spec);return err;
@@ -746,7 +745,7 @@ static int mlx5e_create_promisc_table(struct mlx5e_priv *priv) ft->t =mlx5_create_auto_grouped_flow_table (priv->fs->ns, &ft_attr);if (IS_ERR (ft->t)) { err =PTR_ERR (ft->t); -netdev_err (priv->netdev,"fail to create promisc table err=%d\n" , err); +mlx5_core_err (priv->mdev,"fail to create promisc table err=%d\n" , err);return err; }
@@ -806,8 +805,8 @@ void mlx5e_set_rx_mode_work(struct work_struct *work)if (err) enable_promisc = false;if (!priv->channels.params.vlan_strip_disable && !err) -netdev_warn_once (ndev, -"S-tagged traffic will be dropped while C-tag vlan stripping is enabled\n" ); +mlx5_core_warn_once (priv->mdev, +"S-tagged traffic will be dropped while C-tag vlan stripping is enabled\n" ); }if (enable_allmulti)mlx5e_add_l2_flow_rule (priv, &ea->allmulti,MLX5E_ALLMULTI );
@@ -958,8 +957,8 @@ static int mlx5e_add_l2_flow_rule(struct mlx5e_priv *priv, ai->rule =mlx5_add_flow_rules (ft, spec, &flow_act, &dest,1 );if (IS_ERR (ai->rule)) { -netdev_err (priv->netdev,"%s: add l2 rule(mac:%pM) failed\n" , - __func__, mv_dmac); +mlx5_core_err (priv->mdev,"%s: add l2 rule(mac:%pM) failed\n" , + __func__, mv_dmac); err =PTR_ERR (ai->rule); ai->rule =NULL ; }
@@ -1276,37 +1275,36 @@ int mlx5e_create_flow_steering(struct mlx5e_priv *priv) err =mlx5e_arfs_create_tables (priv);if (err) { -netdev_err (priv->netdev,"Failed to create arfs tables, err=%d\n" , - err); +mlx5_core_err (priv->mdev,"Failed to create arfs tables, err=%d\n" , + err); priv->netdev->hw_features &= ~NETIF_F_NTUPLE ; } err =mlx5e_create_inner_ttc_table (priv);if (err) { -netdev_err (priv->netdev, -"Failed to create inner ttc table, err=%d\n" , - err); +mlx5_core_err (priv->mdev, +"Failed to create inner ttc table, err=%d\n" , err); goto err_destroy_arfs_tables; } err =mlx5e_create_ttc_table (priv);if (err) { -netdev_err (priv->netdev,"Failed to create ttc table, err=%d\n" , - err); +mlx5_core_err (priv->mdev,"Failed to create ttc table, err=%d\n" , + err); goto err_destroy_inner_ttc_table; } err =mlx5e_create_l2_table (priv);if (err) { -netdev_err (priv->netdev,"Failed to create l2 table, err=%d\n" , - err); +mlx5_core_err (priv->mdev,"Failed to create l2 table, err=%d\n" , + err); goto err_destroy_ttc_table; } err =mlx5e_create_vlan_table (priv);if (err) { -netdev_err (priv->netdev,"Failed to create vlan table, err=%d\n" , - err); +mlx5_core_err (priv->mdev,"Failed to create vlan table, err=%d\n" , + err); goto err_destroy_l2_table; }
Signature · 1 line
2.37.1
PATCH 12/15 net/mlx5e: Add mdev to flow_steering struct
In reply to Saeed MahameedFrom: Lama Kayal <[email protected]>
Make flow_steering struct contain mlx5_core_dev such that it becomes self contained and easier to decouple later on this series. Let its values be initialized in mlx5e_fs_init().
--- .../net/ethernet/mellanox/mlx5/core/en/fs.h | 4 +- .../net/ethernet/mellanox/mlx5/core/en_fs.c | 70 ++++++++++--------- .../net/ethernet/mellanox/mlx5/core/en_main.c | 2 +- .../net/ethernet/mellanox/mlx5/core/en_rep.c | 2 +- .../ethernet/mellanox/mlx5/core/ipoib/ipoib.c | 2 +- 5 files changed, 42 insertions(+), 38 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/fs.h +3 -1
@@ -143,6 +143,7 @@ struct mlx5e_fs_any;struct mlx5e_ptp_fs ;struct mlx5e_flow_steering { +struct mlx5_core_dev *mdev;struct mlx5_flow_namespace *ns;#ifdef CONFIG_MLX5_EN_RXNFC struct mlx5e_ethtool_steering ethtool;
@@ -178,7 +179,8 @@ void mlx5e_disable_cvlan_filter(struct mlx5e_priv *priv);int mlx5e_create_flow_steering (struct mlx5e_priv *priv);void mlx5e_destroy_flow_steering (struct mlx5e_priv *priv); -struct mlx5e_flow_steering *mlx5e_fs_init (const struct mlx5e_profile *profile); +struct mlx5e_flow_steering *mlx5e_fs_init (const struct mlx5e_profile *profile, +struct mlx5_core_dev *mdev);void mlx5e_fs_cleanup (struct mlx5e_flow_steering *fs);int mlx5e_add_vlan_trap (struct mlx5e_priv *priv,int trap_id,int tir_num);
drivers/net/ethernet/mellanox/mlx5/core/en_fs.c +36 -34
258 lines not on this page · raw message
drivers/net/ethernet/mellanox/mlx5/core/en_main.c +1 -1
@@ -5020,7 +5020,7 @@ static int mlx5e_nic_init(struct mlx5_core_dev *mdev,mlx5e_timestamp_init (priv); - fs =mlx5e_fs_init (priv->profile); + fs =mlx5e_fs_init (priv->profile, mdev);if (!fs) { err = -ENOMEM ;mlx5_core_err (mdev,"FS initialization failed, %d\n" , err);
drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +1 -1
@@ -836,7 +836,7 @@ static int mlx5e_init_rep_rx(struct mlx5e_priv *priv)struct mlx5_core_dev *mdev = priv->mdev;int err; - priv->fs =mlx5e_fs_init (priv->profile); + priv->fs =mlx5e_fs_init (priv->profile, mdev);if (!priv->fs) {netdev_err (priv->netdev,"FS allocation failed\n" );return -ENOMEM ;
drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c +1 -1
@@ -364,7 +364,7 @@ static int mlx5i_init_rx(struct mlx5e_priv *priv)struct mlx5_core_dev *mdev = priv->mdev;int err; - priv->fs =mlx5e_fs_init (priv->profile); + priv->fs =mlx5e_fs_init (priv->profile, mdev);if (!priv->fs) {netdev_err (priv->netdev,"FS allocation failed\n" );return -ENOMEM ;
Signature · 1 line
2.37.1
PATCH 13/15 net/mlx5e: Separate mlx5e_set_rx_mode_work and move caller to en_main
In reply to Saeed MahameedFrom: Lama Kayal <[email protected]>
Separate mlx5e_set_rx_mode into two, and move caller to en_main while keeping implementation in en_fs in the newly declared function mlx5e_fs_set_rx_mode. This; to minimize the coupling of flow_steering to priv.
Add a parallel boolean member vlan_strip_disable to mlx5e_flow_steering that's updated similarly as its identical in priv, thus making it possible to adjust the rx_mode work handler to current changes.
Also, add state_destroy boolean to mlx5e_flow_steering struct which replaces the old check : !test_bit(MLX5E_STATE_DESTROYING, &priv->state). This state member is updated accordingly prior to INIT_WORK(mlx5e_set_rx_mode_work), This is done for similar purposes as mentioned earlier and to minimize argument passings.
--- .../net/ethernet/mellanox/mlx5/core/en/fs.h | 7 +- .../net/ethernet/mellanox/mlx5/core/en_fs.c | 178 +++++++++--------- .../net/ethernet/mellanox/mlx5/core/en_main.c | 24 ++- .../net/ethernet/mellanox/mlx5/core/en_rep.c | 3 +- .../ethernet/mellanox/mlx5/core/ipoib/ipoib.c | 3 +- 5 files changed, 119 insertions(+), 96 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/fs.h +5 -2
@@ -143,6 +143,8 @@ struct mlx5e_fs_any;struct mlx5e_ptp_fs ;struct mlx5e_flow_steering { +bool state_destroy; +bool vlan_strip_disable;struct mlx5_core_dev *mdev;struct mlx5_flow_namespace *ns;#ifdef CONFIG_MLX5_EN_RXNFC
@@ -180,13 +182,14 @@ int mlx5e_create_flow_steering(struct mlx5e_priv *priv);void mlx5e_destroy_flow_steering (struct mlx5e_priv *priv);struct mlx5e_flow_steering *mlx5e_fs_init (const struct mlx5e_profile *profile, -struct mlx5_core_dev *mdev); +struct mlx5_core_dev *mdev, +bool state_destroy);void mlx5e_fs_cleanup (struct mlx5e_flow_steering *fs);int mlx5e_add_vlan_trap (struct mlx5e_priv *priv,int trap_id,int tir_num);void mlx5e_remove_vlan_trap (struct mlx5e_priv *priv);int mlx5e_add_mac_trap (struct mlx5e_priv *priv,int trap_id,int tir_num);void mlx5e_remove_mac_trap (struct mlx5e_priv *priv); - +void mlx5e_fs_set_rx_mode_work (struct mlx5e_flow_steering *fs,struct net_device *netdev);#endif /* __MLX5E_FLOW_STEER_H__ */
drivers/net/ethernet/mellanox/mlx5/core/en_fs.c +89 -89
408 lines not on this page · raw message
drivers/net/ethernet/mellanox/mlx5/core/en_main.c +21 -3
@@ -3777,14 +3777,17 @@ static int set_feature_rx_vlan(struct net_device *netdev, bool enable)mutex_lock (&priv->state_lock); + priv->fs->vlan_strip_disable = !enable; priv->channels.params.vlan_strip_disable = !enable; +if (!test_bit (MLX5E_STATE_OPENED , &priv->state)) goto unlock; err =mlx5e_modify_channels_vsd (&priv->channels, !enable); -if (err) +if (err) { + priv->fs->vlan_strip_disable = enable; priv->channels.params.vlan_strip_disable = enable; - + } unlock:mutex_unlock (&priv->state_lock);
@@ -5020,7 +5023,8 @@ static int mlx5e_nic_init(struct mlx5_core_dev *mdev,mlx5e_timestamp_init (priv); - fs =mlx5e_fs_init (priv->profile, mdev); + fs =mlx5e_fs_init (priv->profile, mdev, + !test_bit (MLX5E_STATE_DESTROYING , &priv->state));if (!fs) { err = -ENOMEM ;mlx5_core_err (mdev,"FS initialization failed, %d\n" , err);
@@ -5297,6 +5301,14 @@ int mlx5e_get_pf_num_tirs(struct mlx5_core_dev *mdev) +mlx5e_profile_max_num_channels (mdev, &mlx5e_nic_profile); } +void mlx5e_set_rx_mode_work (struct work_struct *work) +{ + struct mlx5e_priv *priv = container_of(work, struct mlx5e_priv, + set_rx_mode_work); + +return mlx5e_fs_set_rx_mode_work (priv->fs, priv->netdev); +} +/* mlx5e generic netdev management API (move to en_common.c) */ int mlx5e_priv_init(struct mlx5e_priv *priv, const struct mlx5e_profile *profile,
@@ -5474,6 +5486,8 @@ int mlx5e_attach_netdev(struct mlx5e_priv *priv)int err;clear_bit (MLX5E_STATE_DESTROYING , &priv->state); +if (priv->fs) + priv->fs->state_destroy = !test_bit (MLX5E_STATE_DESTROYING , &priv->state);/* max number of channels may have changed */ max_nch =mlx5e_calc_max_nch (priv->mdev, priv->netdev, profile);
@@ -5533,6 +5547,8 @@ int mlx5e_attach_netdev(struct mlx5e_priv *priv) out:mlx5e_reset_channels (priv->netdev);set_bit (MLX5E_STATE_DESTROYING , &priv->state); +if (priv->fs) + priv->fs->state_destroy = !test_bit (MLX5E_STATE_DESTROYING , &priv->state);cancel_work_sync (&priv->update_stats_work);return err; }
@@ -5542,6 +5558,8 @@ void mlx5e_detach_netdev(struct mlx5e_priv *priv)const struct mlx5e_profile *profile = priv->profile;set_bit (MLX5E_STATE_DESTROYING , &priv->state); +if (priv->fs) + priv->fs->state_destroy = !test_bit (MLX5E_STATE_DESTROYING , &priv->state);if (profile->disable) profile->disable (priv);
drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +2 -1
@@ -836,7 +836,8 @@ static int mlx5e_init_rep_rx(struct mlx5e_priv *priv)struct mlx5_core_dev *mdev = priv->mdev;int err; - priv->fs =mlx5e_fs_init (priv->profile, mdev); + priv->fs =mlx5e_fs_init (priv->profile, mdev, + !test_bit (MLX5E_STATE_DESTROYING , &priv->state));if (!priv->fs) {netdev_err (priv->netdev,"FS allocation failed\n" );return -ENOMEM ;
drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c +2 -1
@@ -364,7 +364,8 @@ static int mlx5i_init_rx(struct mlx5e_priv *priv)struct mlx5_core_dev *mdev = priv->mdev;int err; - priv->fs =mlx5e_fs_init (priv->profile, mdev); + priv->fs =mlx5e_fs_init (priv->profile, mdev, + !test_bit (MLX5E_STATE_DESTROYING , &priv->state));if (!priv->fs) {netdev_err (priv->netdev,"FS allocation failed\n" );return -ENOMEM ;
Signature · 1 line
2.37.1
PATCH 14/15 net/mlx5e: Split en_fs ndo's and move to en_main
In reply to Saeed MahameedFrom: Lama Kayal <[email protected]>
Add inner callee for ndo mlx5e_vlan_rx_add_vid and mlx5e_vlan_rx_kill_vid, to separate the priv usage from other flow steering flows.
Move wrapper ndo's into en_main, and split the rest of the functionality into a separate part inside en_fs.
--- .../net/ethernet/mellanox/mlx5/core/en/fs.h | 6 + .../net/ethernet/mellanox/mlx5/core/en_fs.c | 180 +++++++++--------- .../net/ethernet/mellanox/mlx5/core/en_main.c | 22 +++ 3 files changed, 120 insertions(+), 88 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/fs.h +6 -0
@@ -191,5 +191,11 @@ void mlx5e_remove_vlan_trap(struct mlx5e_priv *priv);int mlx5e_add_mac_trap (struct mlx5e_priv *priv,int trap_id,int tir_num);void mlx5e_remove_mac_trap (struct mlx5e_priv *priv);void mlx5e_fs_set_rx_mode_work (struct mlx5e_flow_steering *fs,struct net_device *netdev); +int mlx5e_fs_vlan_rx_add_vid (struct mlx5e_flow_steering *fs, +struct net_device *netdev, +__be16 proto,u16 vid); +int mlx5e_fs_vlan_rx_kill_vid (struct mlx5e_flow_steering *fs, +struct net_device *netdev, +__be16 proto,u16 vid);#endif /* __MLX5E_FLOW_STEER_H__ */
drivers/net/ethernet/mellanox/mlx5/core/en_fs.c +92 -88
413 lines not on this page · raw message
drivers/net/ethernet/mellanox/mlx5/core/en_main.c +22 -0
@@ -3794,6 +3794,28 @@ static int set_feature_rx_vlan(struct net_device *netdev, bool enable)return err; } +int mlx5e_vlan_rx_add_vid (struct net_device *dev,__be16 proto,u16 vid) +{ +struct mlx5e_priv *priv =netdev_priv (dev); +struct mlx5e_flow_steering *fs = priv->fs; + +if (mlx5e_is_uplink_rep (priv)) +return 0 ;/* no vlan table for uplink rep */ + +return mlx5e_fs_vlan_rx_add_vid (fs, dev, proto, vid); +} + +int mlx5e_vlan_rx_kill_vid (struct net_device *dev,__be16 proto,u16 vid) +{ +struct mlx5e_priv *priv =netdev_priv (dev); +struct mlx5e_flow_steering *fs = priv->fs; + +if (mlx5e_is_uplink_rep (priv)) +return 0 ;/* no vlan table for uplink rep */ + +return mlx5e_fs_vlan_rx_kill_vid (fs, dev, proto, vid); +} +#ifdef CONFIG_MLX5_EN_ARFS static int set_feature_arfs (struct net_device *netdev,bool enable) {
Signature · 1 line
2.37.1
PATCH 15/15 net/mlx5e: Move mlx5e_init_l2_addr to en_main
In reply to Saeed MahameedFrom: Lama Kayal <[email protected]>
Move the function declaration of mlx5e_init_l2_addr to en/fs.h, rename to mlx5e_fs_init_l2_addr to align with the fs API functions naming convention and let it take mlx5e_flow_steering as arguments while keeping implementation at en_fs.c file. This helps maintain a clean driver code and avoids unnecessary dependencies.
--- drivers/net/ethernet/mellanox/mlx5/core/en.h | 1 - drivers/net/ethernet/mellanox/mlx5/core/en/fs.h | 1 + drivers/net/ethernet/mellanox/mlx5/core/en_fs.c | 4 ++-- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 2 +- drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 2 +- 5 files changed, 5 insertions(+), 5 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en.h +0 -1
@@ -1024,7 +1024,6 @@ void mlx5e_shampo_dealloc_hd(struct mlx5e_rq *rq, u16 len, u16 start, bool closevoid mlx5e_get_stats (struct net_device *dev,struct rtnl_link_stats64 *stats);void mlx5e_fold_sw_stats64 (struct mlx5e_priv *priv,struct rtnl_link_stats64 *s); -void mlx5e_init_l2_addr (struct mlx5e_priv *priv);int mlx5e_self_test_num (struct mlx5e_priv *priv);int mlx5e_self_test_fill_strings (struct mlx5e_priv *priv,u8 *data); void mlx5e_self_test(struct net_device *ndev, struct ethtool_test *etest,
drivers/net/ethernet/mellanox/mlx5/core/en/fs.h +1 -0
@@ -197,5 +197,6 @@ int mlx5e_fs_vlan_rx_add_vid(struct mlx5e_flow_steering *fs,int mlx5e_fs_vlan_rx_kill_vid (struct mlx5e_flow_steering *fs,struct net_device *netdev,__be16 proto,u16 vid); +void mlx5e_fs_init_l2_addr (struct mlx5e_flow_steering *fs,struct net_device *netdev);#endif /* __MLX5E_FLOW_STEER_H__ */
drivers/net/ethernet/mellanox/mlx5/core/en_fs.c +2 -2
@@ -843,9 +843,9 @@ static void mlx5e_destroy_groups(struct mlx5e_flow_table *ft) ft->num_groups =0 ; } -void mlx5e_init_l2_addr (struct mlx5e_priv *priv) +void mlx5e_fs_init_l2_addr (struct mlx5e_flow_steering *fs,struct net_device *netdev) { -ether_addr_copy (priv->fs->l2.broadcast.addr, priv->netdev->broadcast); +ether_addr_copy (fs->l2.broadcast.addr, netdev->broadcast); }void mlx5e_destroy_flow_table (struct mlx5e_flow_table *ft)
drivers/net/ethernet/mellanox/mlx5/core/en_main.c +1 -1
@@ -5186,7 +5186,7 @@ static void mlx5e_nic_enable(struct mlx5e_priv *priv)struct net_device *netdev = priv->netdev;struct mlx5_core_dev *mdev = priv->mdev; -mlx5e_init_l2_addr (priv); +mlx5e_fs_init_l2_addr (priv->fs, netdev);/* Marking the link as currently not needed by the Driver */ if (!netif_running (netdev))
drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +1 -1
@@ -849,7 +849,7 @@ static int mlx5e_init_rep_rx(struct mlx5e_priv *priv) goto err_free_fs; } -mlx5e_init_l2_addr (priv); +mlx5e_fs_init_l2_addr (priv->fs, priv->netdev); err =mlx5e_open_drop_rq (priv, &priv->drop_rq);if (err) {
Signature · 1 line
2.37.1