Skip to contentSkip to conversation

Please Pull Mellanox Shared Code

GIT PULL Leon Romanovsky 5 participants 9 messages Jun 13, 2016 – Jun 14, 2016 2 messages not in the archive

Subsystems (5)
lore
Share & tools
lore
#0 lore raw

Hi Doug and David,

This is Mellanox mlx5_core shared code for both net-next and RDMA trees for 4.8 kernel cycle.

This mlx5_ifc update introduces following ConnectX-4 features.

* netdev part:
- MLX5_CQ_PERIOD_NUM_MODES for adaptive moderation support
- QoS rate limiting
- SQ context rate limiting
- Auto negotiation fields in PTYS register
- Source SQN field in flow table entry match structure
- DCBX parameters

* RDMA part:
- New XRQ opcodes, commands and capabilities layout
- Extend Q counters definition to support IB.

Thank you,
Saeed and Leon.

The following changes since commit af8c34ce6ae32addda3788d54a7e340cad22516b:

Linux 4.7-rc2 (2016-06-05 14:31:26 -0700)

are available in the git repository at:

[email protected]:pub/scm/linux/kernel/git/leon/linux-rdma tags/shared

for you to fetch changes up to 7486216b3a0bd26375b17b2cc168a311106cea70:

{net,IB}/mlx5: mlx5_ifc updates (2016-06-10 13:29:14 +0300)

----------------------------------------------------------------
Saeed Mahameed (1):

      {net,IB}/mlx5: mlx5_ifc updates
 include/linux/mlx5/mlx5_ifc.h | 275 ++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 263 insertions(+), 12 deletions(-)
#1 lore raw
Quoted 15 lines · Leon Romanovsky

Hi Doug and David,

This is Mellanox mlx5_core shared code for both net-next and RDMA trees for 4.8 kernel cycle.

This mlx5_ifc update introduces following ConnectX-4 features.

* netdev part:
- MLX5_CQ_PERIOD_NUM_MODES for adaptive moderation support
- QoS rate limiting
- SQ context rate limiting
- Auto negotiation fields in PTYS register
- Source SQN field in flow table entry match structure
- DCBX parameters

Quoted 3 lines

* RDMA part:
- New XRQ opcodes, commands and capabilities layout
- Extend Q counters definition to support IB.

Leon, any reason not to send the actual patch over the list as any other submission we do? please send it over.

Or.

#4 lore raw
In reply to a message not in the archive
Quoted 5 lines · David Miller

A bare pull request does not work.

You must post the actual patches so that they can be reviewed properly.

Sorry, It was first time attempt to get the sense for the proper procedure.

I will send the patch as a reply to this pull request and my hope that it will be fine with you.

Thanks.

#5 lore raw
In reply to Leon Romanovsky
Quoted 32 lines · Leon Romanovsky

Hi Doug and David,

This is Mellanox mlx5_core shared code for both net-next and RDMA trees for 4.8 kernel cycle.

This mlx5_ifc update introduces following ConnectX-4 features.

* netdev part:
- MLX5_CQ_PERIOD_NUM_MODES for adaptive moderation support
- QoS rate limiting
- SQ context rate limiting
- Auto negotiation fields in PTYS register
- Source SQN field in flow table entry match structure
- DCBX parameters

* RDMA part:
- New XRQ opcodes, commands and capabilities layout
- Extend Q counters definition to support IB.

Thank you,
Saeed and Leon.

The following changes since commit af8c34ce6ae32addda3788d54a7e340cad22516b:

Linux 4.7-rc2 (2016-06-05 14:31:26 -0700)

are available in the git repository at:

[email protected]:pub/scm/linux/kernel/git/leon/linux-rdma tags/shared

for you to fetch changes up to 7486216b3a0bd26375b17b2cc168a311106cea70:

Quoted 8 lines

{net,IB}/mlx5: mlx5_ifc updates (2016-06-10 13:29:14 +0300)

----------------------------------------------------------------
Saeed Mahameed (1):

      {net,IB}/mlx5: mlx5_ifc updates
 include/linux/mlx5/mlx5_ifc.h | 275 ++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 263 insertions(+), 12 deletions(-)

From 7486216b3a0bd26375b17b2cc168a311106cea70 Mon Sep 17 00:00:00 2001
From: Saeed Mahameed <[email protected]>
Date: Thu, 9 Jun 2016 15:11:34 +0300
Subject: [PATCH] {net,IB}/mlx5: mlx5_ifc updates

Introducing mlx5_ifc updates for upcoming ConnectX-4 features.

Needed bits and hardware structures for mlx5e netdev:
- MLX5_CQ_PERIOD_NUM_MODES for adaptive moderation support
- QoS rate limiting
- SQ context rate limiting
- Auto negotiation fields in PTYS register
- Source SQN field in flow table entry match structure
- DCBX parameters

Needed bits and hardware structures for IB:
- New XRQ opcodes, commands and capabilities layout
- Extend q counters definition to support IB.

---
 include/linux/mlx5/mlx5_ifc.h | 275 ++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 263 insertions(+), 12 deletions(-)
include/linux/mlx5/mlx5_ifc.h +263 -12
index 9a05cd7..209add9 100644
@@ -123,6 +123,10 @@ enum {
 	MLX5_CMD_OP_DRAIN_DCT                     = 0x712,
 	MLX5_CMD_OP_QUERY_DCT                     = 0x713,
 	MLX5_CMD_OP_ARM_DCT_FOR_KEY_VIOLATION     = 0x714,
+	MLX5_CMD_OP_CREATE_XRQ                    = 0x717,
+	MLX5_CMD_OP_DESTROY_XRQ                   = 0x718,
+	MLX5_CMD_OP_QUERY_XRQ                     = 0x719,
+	MLX5_CMD_OP_ARM_XRQ                       = 0x71a,
 	MLX5_CMD_OP_QUERY_VPORT_STATE             = 0x750,
 	MLX5_CMD_OP_MODIFY_VPORT_STATE            = 0x751,
 	MLX5_CMD_OP_QUERY_ESW_VPORT_CONTEXT       = 0x752,
@@ -139,6 +143,8 @@ enum {
 	MLX5_CMD_OP_ALLOC_Q_COUNTER               = 0x771,
 	MLX5_CMD_OP_DEALLOC_Q_COUNTER             = 0x772,
 	MLX5_CMD_OP_QUERY_Q_COUNTER               = 0x773,
+	MLX5_CMD_OP_SET_RATE_LIMIT                = 0x780,
+	MLX5_CMD_OP_QUERY_RATE_LIMIT              = 0x781,
 	MLX5_CMD_OP_ALLOC_PD                      = 0x800,
 	MLX5_CMD_OP_DEALLOC_PD                    = 0x801,
 	MLX5_CMD_OP_ALLOC_UAR                     = 0x802,
@@ -361,7 +367,8 @@ struct mlx5_ifc_fte_match_set_lyr_2_4_bits {
 };
 
 struct mlx5_ifc_fte_match_set_misc_bits {
-	u8         reserved_at_0[0x20];
+	u8         reserved_at_0[0x8];
+	u8         source_sqn[0x18];
 
 	u8         reserved_at_20[0x10];
 	u8         source_port[0x10];
@@ -505,6 +512,17 @@ struct mlx5_ifc_e_switch_cap_bits {
 	u8         reserved_at_20[0x7e0];
 };
 
+struct mlx5_ifc_qos_cap_bits {
+	u8         packet_pacing[0x1];
+	u8         reserved_0[0x1f];
+	u8         reserved_1[0x20];
+	u8         packet_pacing_max_rate[0x20];
+	u8         packet_pacing_min_rate[0x20];
+	u8         reserved_2[0x10];
+	u8         packet_pacing_rate_table_size[0x10];
+	u8         reserved_3[0x760];
+};
+
 struct mlx5_ifc_per_protocol_networking_offload_caps_bits {
 	u8         csum_cap[0x1];
 	u8         vlan_cap[0x1];
@@ -744,7 +762,8 @@ struct mlx5_ifc_cmd_hca_cap_bits {
 
 	u8         out_of_seq_cnt[0x1];
 	u8         vport_counters[0x1];
-	u8         reserved_at_182[0x4];
+	u8         retransmission_q_counters[0x1];
+	u8         reserved_at_183[0x3];
 	u8         max_qp_cnt[0xa];
 	u8         pkey_table_size[0x10];
 
@@ -771,7 +790,9 @@ struct mlx5_ifc_cmd_hca_cap_bits {
 	u8         log_max_msg[0x5];
 	u8         reserved_at_1c8[0x4];
 	u8         max_tc[0x4];
-	u8         reserved_at_1d0[0x6];
+	u8         reserved_at_1d0[0x1];
+	u8         dcbx[0x1];
+	u8         reserved_at_1d2[0x4];
 	u8         rol_s[0x1];
 	u8         rol_g[0x1];
 	u8         reserved_at_1d8[0x1];
@@ -803,7 +824,7 @@ struct mlx5_ifc_cmd_hca_cap_bits {
 	u8         tph[0x1];
 	u8         rf[0x1];
 	u8         dct[0x1];
-	u8         reserved_at_21b[0x1];
+	u8         qos[0x1];
 	u8         eth_net_offloads[0x1];
 	u8         roce[0x1];
 	u8         atomic[0x1];
@@ -929,7 +950,15 @@ struct mlx5_ifc_cmd_hca_cap_bits {
 	u8         cqe_compression_timeout[0x10];
 	u8         cqe_compression_max_num[0x10];
 
-	u8         reserved_at_5e0[0x220];
+	u8         reserved_at_5e0[0x10];
+	u8         tag_matching[0x1];
+	u8         rndv_offload_rc[0x1];
+	u8         rndv_offload_dc[0x1];
+	u8         log_tag_matching_list_sz[0x5];
+	u8         reserved_at_5e8[0x3];
+	u8         log_max_xrq[0x5];
+
+	u8         reserved_at_5f0[0x200];
 };
 
 enum mlx5_flow_destination_type {
@@ -1967,7 +1996,7 @@ struct mlx5_ifc_qpc_bits {
 
 	u8         reserved_at_560[0x5];
 	u8         rq_type[0x3];
-	u8         srqn_rmpn[0x18];
+	u8         srqn_rmpn_xrqn[0x18];
 
 	u8         reserved_at_580[0x8];
 	u8         rmsn[0x18];
@@ -2018,6 +2047,7 @@ union mlx5_ifc_hca_cap_union_bits {
 	struct mlx5_ifc_flow_table_eswitch_cap_bits flow_table_eswitch_cap;
 	struct mlx5_ifc_e_switch_cap_bits e_switch_cap;
 	struct mlx5_ifc_vector_calc_cap_bits vector_calc_cap;
+	struct mlx5_ifc_qos_cap_bits qos_cap;
 	u8         reserved_at_0[0x8000];
 };
 
@@ -2244,8 +2274,9 @@ struct mlx5_ifc_sqc_bits {
 	u8         reserved_at_40[0x8];
 	u8         cqn[0x18];
 
-	u8         reserved_at_60[0xa0];
+	u8         reserved_at_60[0x90];
 
+	u8         packet_pacing_rate_limit_index[0x10];
 	u8         tis_lst_sz[0x10];
 	u8         reserved_at_110[0x10];
 
@@ -2593,7 +2624,7 @@ struct mlx5_ifc_dctc_bits {
 	u8         reserved_at_98[0x8];
 
 	u8         reserved_at_a0[0x8];
-	u8         srqn[0x18];
+	u8         srqn_xrqn[0x18];
 
 	u8         reserved_at_c0[0x8];
 	u8         pd[0x18];
@@ -2645,6 +2676,7 @@ enum {
 enum {
 	MLX5_CQ_PERIOD_MODE_START_FROM_EQE = 0x0,
 	MLX5_CQ_PERIOD_MODE_START_FROM_CQE = 0x1,
+	MLX5_CQ_PERIOD_NUM_MODES
 };
 
 struct mlx5_ifc_cqc_bits {
@@ -2722,6 +2754,54 @@ struct mlx5_ifc_query_adapter_param_block_bits {
 	u8         vsd_contd_psid[16][0x8];
 };
 
+enum {
+	MLX5_XRQC_STATE_GOOD   = 0x0,
+	MLX5_XRQC_STATE_ERROR  = 0x1,
+};
+
+enum {
+	MLX5_XRQC_TOPOLOGY_NO_SPECIAL_TOPOLOGY = 0x0,
+	MLX5_XRQC_TOPOLOGY_TAG_MATCHING        = 0x1,
+};
+
+enum {
+	MLX5_XRQC_OFFLOAD_RNDV = 0x1,
+};
+
+struct mlx5_ifc_tag_matching_topology_context_bits {
+	u8         log_matching_list_sz[0x4];
+	u8         reserved_at_4[0xc];
+	u8         append_next_index[0x10];
+
+	u8         sw_phase_cnt[0x10];
+	u8         hw_phase_cnt[0x10];
+
+	u8         reserved_at_40[0x40];
+};
+
+struct mlx5_ifc_xrqc_bits {
+	u8         state[0x4];
+	u8         rlkey[0x1];
+	u8         reserved_at_5[0xf];
+	u8         topology[0x4];
+	u8         reserved_at_18[0x4];
+	u8         offload[0x4];
+
+	u8         reserved_at_20[0x8];
+	u8         user_index[0x18];
+
+	u8         reserved_at_40[0x8];
+	u8         cqn[0x18];
+
+	u8         reserved_at_60[0xa0];
+
+	struct mlx5_ifc_tag_matching_topology_context_bits tag_matching_topology_context;
+
+	u8         reserved_at_180[0x180];
+
+	struct mlx5_ifc_wq_bits wq;
+};
+
 union mlx5_ifc_modify_field_select_resize_field_select_auto_bits {
 	struct mlx5_ifc_modify_field_select_bits modify_field_select;
 	struct mlx5_ifc_resize_field_select_bits resize_field_select;
@@ -3144,6 +3224,30 @@ struct mlx5_ifc_rst2init_qp_in_bits {
 	u8         reserved_at_800[0x80];
 };
 
+struct mlx5_ifc_query_xrq_out_bits {
+	u8         status[0x8];
+	u8         reserved_at_8[0x18];
+
+	u8         syndrome[0x20];
+
+	u8         reserved_at_40[0x40];
+
+	struct mlx5_ifc_xrqc_bits xrq_context;
+};
+
+struct mlx5_ifc_query_xrq_in_bits {
+	u8         opcode[0x10];
+	u8         reserved_at_10[0x10];
+
+	u8         reserved_at_20[0x10];
+	u8         op_mod[0x10];
+
+	u8         reserved_at_40[0x8];
+	u8         xrqn[0x18];
+
+	u8         reserved_at_60[0x20];
+};
+
 struct mlx5_ifc_query_xrc_srq_out_bits {
 	u8         status[0x8];
 	u8         reserved_at_8[0x18];
@@ -3547,7 +3651,27 @@ struct mlx5_ifc_query_q_counter_out_bits {
 
 	u8         out_of_sequence[0x20];
 
-	u8         reserved_at_1e0[0x620];
+	u8         reserved_at_1e0[0x20];
+
+	u8         duplicate_request[0x20];
+
+	u8         reserved_at_220[0x20];
+
+	u8         rnr_nak_retry_err[0x20];
+
+	u8         reserved_at_260[0x20];
+
+	u8         packet_seq_err[0x20];
+
+	u8         reserved_at_2a0[0x20];
+
+	u8         implied_nak_seq_err[0x20];
+
+	u8         reserved_at_2e0[0x20];
+
+	u8         local_ack_timeout_err[0x20];
+
+	u8         reserved_at_320[0x4e0];
 };
 
 struct mlx5_ifc_query_q_counter_in_bits {
@@ -4998,6 +5122,28 @@ struct mlx5_ifc_detach_from_mcg_in_bits {
 	u8         multicast_gid[16][0x8];
 };
 
+struct mlx5_ifc_destroy_xrq_out_bits {
+	u8         status[0x8];
+	u8         reserved_at_8[0x18];
+
+	u8         syndrome[0x20];
+
+	u8         reserved_at_40[0x40];
+};
+
+struct mlx5_ifc_destroy_xrq_in_bits {
+	u8         opcode[0x10];
+	u8         reserved_at_10[0x10];
+
+	u8         reserved_at_20[0x10];
+	u8         op_mod[0x10];
+
+	u8         reserved_at_40[0x8];
+	u8         xrqn[0x18];
+
+	u8         reserved_at_60[0x20];
+};
+
 struct mlx5_ifc_destroy_xrc_srq_out_bits {
 	u8         status[0x8];
 	u8         reserved_at_8[0x18];
@@ -5583,6 +5729,30 @@ struct mlx5_ifc_dealloc_flow_counter_in_bits {
 	u8         reserved_at_60[0x20];
 };
 
+struct mlx5_ifc_create_xrq_out_bits {
+	u8         status[0x8];
+	u8         reserved_at_8[0x18];
+
+	u8         syndrome[0x20];
+
+	u8         reserved_at_40[0x8];
+	u8         xrqn[0x18];
+
+	u8         reserved_at_60[0x20];
+};
+
+struct mlx5_ifc_create_xrq_in_bits {
+	u8         opcode[0x10];
+	u8         reserved_at_10[0x10];
+
+	u8         reserved_at_20[0x10];
+	u8         op_mod[0x10];
+
+	u8         reserved_at_40[0x40];
+
+	struct mlx5_ifc_xrqc_bits xrq_context;
+};
+
 struct mlx5_ifc_create_xrc_srq_out_bits {
 	u8         status[0x8];
 	u8         reserved_at_8[0x18];
@@ -6124,6 +6294,29 @@ struct mlx5_ifc_attach_to_mcg_in_bits {
 	u8         multicast_gid[16][0x8];
 };
 
+struct mlx5_ifc_arm_xrq_out_bits {
+	u8         status[0x8];
+	u8         reserved_at_8[0x18];
+
+	u8         syndrome[0x20];
+
+	u8         reserved_at_40[0x40];
+};
+
+struct mlx5_ifc_arm_xrq_in_bits {
+	u8         opcode[0x10];
+	u8         reserved_at_10[0x10];
+
+	u8         reserved_at_20[0x10];
+	u8         op_mod[0x10];
+
+	u8         reserved_at_40[0x8];
+	u8         xrqn[0x18];
+
+	u8         reserved_at_60[0x10];
+	u8         lwm[0x10];
+};
+
 struct mlx5_ifc_arm_xrc_srq_out_bits {
 	u8         status[0x8];
 	u8         reserved_at_8[0x18];
@@ -6161,7 +6354,8 @@ struct mlx5_ifc_arm_rq_out_bits {
 };
 
 enum {
-	MLX5_ARM_RQ_IN_OP_MOD_SRQ_  = 0x1,
+	MLX5_ARM_RQ_IN_OP_MOD_SRQ = 0x1,
+	MLX5_ARM_RQ_IN_OP_MOD_XRQ = 0x2,
 };
 
 struct mlx5_ifc_arm_rq_in_bits {
@@ -6354,6 +6548,30 @@ struct mlx5_ifc_add_vxlan_udp_dport_in_bits {
 	u8         vxlan_udp_port[0x10];
 };
 
+struct mlx5_ifc_set_rate_limit_out_bits {
+	u8         status[0x8];
+	u8         reserved_at_8[0x18];
+
+	u8         syndrome[0x20];
+
+	u8         reserved_at_40[0x40];
+};
+
+struct mlx5_ifc_set_rate_limit_in_bits {
+	u8         opcode[0x10];
+	u8         reserved_at_10[0x10];
+
+	u8         reserved_at_20[0x10];
+	u8         op_mod[0x10];
+
+	u8         reserved_at_40[0x10];
+	u8         rate_limit_index[0x10];
+
+	u8         reserved_at_60[0x20];
+
+	u8         rate_limit[0x20];
+};
+
 struct mlx5_ifc_access_register_out_bits {
 	u8         status[0x8];
 	u8         reserved_at_8[0x18];
@@ -6478,12 +6696,15 @@ struct mlx5_ifc_pude_reg_bits {
 };
 
 struct mlx5_ifc_ptys_reg_bits {
-	u8         reserved_at_0[0x8];
+	u8         an_disable_cap[0x1];
+	u8         an_disable_admin[0x1];
+	u8         reserved_at_2[0x6];
 	u8         local_port[0x8];
 	u8         reserved_at_10[0xd];
 	u8         proto_mask[0x3];
 
-	u8         reserved_at_20[0x40];
+	u8         an_status[0x4];
+	u8         reserved_at_24[0x3c];
 
 	u8         eth_proto_capability[0x20];
 
@@ -7444,4 +7665,34 @@ struct mlx5_ifc_mcia_reg_bits {
 	u8         dword_11[0x20];
 };
 
+struct mlx5_ifc_dcbx_param_bits {
+	u8         dcbx_cee_cap[0x1];
+	u8         dcbx_ieee_cap[0x1];
+	u8         dcbx_standby_cap[0x1];
+	u8         reserved_at_0[0x5];
+	u8         port_number[0x8];
+	u8         reserved_at_10[0xa];
+	u8         max_application_table_size[6];
+	u8         reserved_at_20[0x15];
+	u8         version_oper[0x3];
+	u8         reserved_at_38[5];
+	u8         version_admin[0x3];
+	u8         willing_admin[0x1];
+	u8         reserved_at_41[0x3];
+	u8         pfc_cap_oper[0x4];
+	u8         reserved_at_48[0x4];
+	u8         pfc_cap_admin[0x4];
+	u8         reserved_at_50[0x4];
+	u8         num_of_tc_oper[0x4];
+	u8         reserved_at_58[0x4];
+	u8         num_of_tc_admin[0x4];
+	u8         remote_willing[0x1];
+	u8         reserved_at_61[3];
+	u8         remote_pfc_cap[4];
+	u8         reserved_at_68[0x14];
+	u8         remote_num_of_tc[0x4];
+	u8         reserved_at_80[0x18];
+	u8         error[0x8];
+	u8         reserved_at_a0[0x160];
+};
 #endif /* MLX5_IFC_H */
Signature · 1 line
2.1.4
maintainer
#6 lore raw

Please, just stop right now.

We're not asking for the a huge patch covering the entire pull request.

We're asking for you to post a patch series, one per commit, and as numbered, fresh, new mailing list patch postings. As if you were submitting these changes for the very first time to the list.

Please ask a colleague for how to do this properly before you spam the list again with more unnecessary and improperly formatted submissions.

Thank.

#8 lore raw
In reply to a message not in the archive
Quoted 6 lines · David Miller

Please, just stop right now.

We're not asking for the a huge patch covering the entire pull request.

Quoted 6 lines

We're asking for you to post a patch series, one per commit, and as numbered, fresh, new mailing list patch postings. As if you were submitting these changes for the very first time to the list.

Please ask a colleague for how to do this properly before you spam the list again with more unnecessary and improperly formatted submissions.

Dave,

He's doing what Linus requested from Mellanox. They were instructed to identify all of the coming changes in the next release where patches sent through your tree and patches sent through my tree would have conflicts. They were then instructed to make git commits that merged those changes and have both you and I pull them (please note "pull", not git am patches) so that we have the exact same commit hashes in our trees and git will do the right thing when your tree and my tree are merged in the next merge window. As it turns out the conflicts are mainly in this firmware offset definition file. They just put all of the needed changes in there in one commit. I doubt they have individual patches here. This was a custom made patch who's sole purpose is to combine all the otherwise possibly conflicting changes in one place. I'm sure they could split it out into different commits, but they wouldn't be full commits, just the portion that applies to this firmware file.

--
Doug Ledford <[email protected]>
GPG Key ID: 0E572FDD

#9 lore raw
Quoted 13 lines · Doug Ledford

Dave,

He's doing what Linus requested from Mellanox. They were instructed to identify all of the coming changes in the next release where patches sent through your tree and patches sent through my tree would have conflicts. They were then instructed to make git commits that merged those changes and have both you and I pull them (please note "pull", not git am patches) so that we have the exact same commit hashes in our trees and git will do the right thing when your tree and my tree are merged in the next merge window. As it turns out the conflicts are mainly in this firmware offset definition file. They just put all of the needed changes in there in one commit. I doubt they have individual patches here.

Doug,

The main comment was that the submission lacked posting of the actual changes inline as patch/es along with a cover letter that gives the background plus the URL for git pulling of this. In this case (4.8 shared code), the submission only touches the mlx5_ifc.c FW file and indeed it makes sense to have only one patch that covers it all, in prev/future cases we had/would have that in multiple patches.

Leon re-submitted that later today in the proper way

Or.

Quoted 4 lines

This was a custom made patch who's sole purpose is to
combine all the otherwise possibly conflicting changes in one place. I'm sure they could split it out into different commits, but they wouldn't be full commits, just the portion that applies to this firmware file

#10 lore raw
In reply to Doug Ledford
Quoted 16 lines · Doug Ledford
Quoted 6 lines · David Miller

Please, just stop right now.

We're not asking for the a huge patch covering the entire pull request.

Quoted 6 lines

We're asking for you to post a patch series, one per commit, and as numbered, fresh, new mailing list patch postings. As if you were submitting these changes for the very first time to the list.

Please ask a colleague for how to do this properly before you spam the list again with more unnecessary and improperly formatted submissions.

Dave,

Quoted 15 lines

He's doing what Linus requested from Mellanox. They were instructed to identify all of the coming changes in the next release where patches sent through your tree and patches sent through my tree would have conflicts. They were then instructed to make git commits that merged those changes and have both you and I pull them (please note "pull", not git am patches) so that we have the exact same commit hashes in our trees and git will do the right thing when your tree and my tree are merged in the next merge window. As it turns out the conflicts are mainly in this firmware offset definition file. They just put all of the needed changes in there in one commit. I doubt they have individual patches here. This was a custom made patch who's sole purpose is to combine all the otherwise possibly conflicting changes in one place. I'm sure they could split it out into different commits, but they wouldn't be full commits, just the portion that applies to this firmware file.

Thanks Doug,

In addition to your accurate description. We don't have separate commits for this file (mlx5_ifc.h) for many reasons: auto generated file, one logical change (enable FW features), no helpful for bisect e.t.c.

My goal was to simplify the maintainers' work by sending minimal possible pull request with easy for review information.

I also resent the pull request [1] as a cover letter together with patch itself [2].

[1] http://lkml.kernel.org/r/1465900500-1578-1-git-send-email-leonro+()+mellanox+!+com
[2] http://lkml.kernel.org/r/1465900500-1578-2-git-send-email-leonro+()+mellanox+!+com

Thanks.

Quoted 5 lines

--
Doug Ledford <[email protected]>
GPG Key ID: 0E572FDD