[net-next PATCH v2 00/11] Add CN20K NIX and NPA contexts

[net-next PATCH v2 00/11] Add CN20K NIX and NPA contexts

[Subbaraya Sundeep]

The hardware contexts of blocks NIX and NPA in CN20K silicon are
different than that of previous silicons CN10K and CN9XK. This
patchset adds the new contexts of CN20K in AF and PF drivers.
A new mailbox for enqueuing contexts to hardware is added.

Patch 1 simplifies context writing and reading by using max context
size supported by hardware instead of using each context size.
Patch 2 and 3 adds NIX block contexts in AF driver and extends
debugfs to display those new contexts
Patch 4 and 5 adds NPA block contexts in AF driver and extends
debugfs to display those new contexts
Patch 6 omits NDC configuration since CN20K NPA does not use NDC
for caching its contexts
Patch 7 and 8 uses the new NIX and NPA contexts in PF/VF driver.
Patch 9, 10 and 11 are to support more bandwidth profiles present in
CN20K for RX ratelimiting and to display new profiles in debugfs

v2 changes:
 Fixed string fortifier warnings by padding structures

Linu Cherian (4):
  octeontx2-af: Add cn20k NPA block contexts
  octeontx2-af: Extend debugfs support for cn20k NPA
  octeontx2-af: Skip NDC operations for cn20k
  octeontx2-pf: Initialize cn20k specific aura and pool contexts

Subbaraya Sundeep (7):
  octeontx2-af: Simplify context writing and reading to hardware
  octeontx2-af: Add cn20k NIX block contexts
  octeontx2-af: Extend debugfs support for cn20k NIX
  octeontx2-pf: Initialize new NIX SQ context for cn20k
  octeontx2-af: Accommodate more bandwidth profiles for cn20k
  octeontx2-af: Display new bandwidth profiles too in debugfs
  octeontx2-pf: Use new bandwidth profiles in receive queue

 .../ethernet/marvell/octeontx2/af/Makefile    |   3 +-
 .../marvell/octeontx2/af/cn20k/debugfs.c      | 216 ++++++++++++
 .../marvell/octeontx2/af/cn20k/debugfs.h      |  28 ++
 .../ethernet/marvell/octeontx2/af/cn20k/nix.c |  20 ++
 .../ethernet/marvell/octeontx2/af/cn20k/npa.c |  21 ++
 .../marvell/octeontx2/af/cn20k/struct.h       | 326 ++++++++++++++++++
 .../net/ethernet/marvell/octeontx2/af/mbox.h  |  73 ++++
 .../net/ethernet/marvell/octeontx2/af/rvu.h   |   3 +
 .../marvell/octeontx2/af/rvu_debugfs.c        |  39 ++-
 .../ethernet/marvell/octeontx2/af/rvu_nix.c   |  69 ++--
 .../ethernet/marvell/octeontx2/af/rvu_npa.c   |  29 +-
 .../marvell/octeontx2/af/rvu_struct.h         |  13 +-
 .../ethernet/marvell/octeontx2/nic/cn10k.c    |  10 +
 .../ethernet/marvell/octeontx2/nic/cn20k.c    | 212 +++++++++++-
 .../marvell/octeontx2/nic/otx2_common.c       |  14 +
 .../marvell/octeontx2/nic/otx2_common.h       |  10 +
 16 files changed, 1032 insertions(+), 54 deletions(-)
 create mode 100644 drivers/net/ethernet/marvell/octeontx2/af/cn20k/debugfs.c
 create mode 100644 drivers/net/ethernet/marvell/octeontx2/af/cn20k/debugfs.h
 create mode 100644 drivers/net/ethernet/marvell/octeontx2/af/cn20k/nix.c
 create mode 100644 drivers/net/ethernet/marvell/octeontx2/af/cn20k/npa.c

-- 
2.34.1

[net-next PATCH v2 01/11] octeontx2-af: Simplify context writing and reading to hardware

[Subbaraya Sundeep]

Simplify NIX context reading and writing by using hardware
maximum context size instead of using individual sizes of
each context type.

Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com>
---
 .../ethernet/marvell/octeontx2/af/rvu_nix.c   | 46 ++++++++++---------
 .../marvell/octeontx2/af/rvu_struct.h         |  7 ++-
 2 files changed, 30 insertions(+), 23 deletions(-)

diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c b/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c
index bdf4d852c15d..48d44911b663 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c
@@ -17,6 +17,8 @@
 #include "lmac_common.h"
 #include "rvu_npc_hash.h"
 
+#define NIX_MAX_CTX_SIZE	128
+
 static void nix_free_tx_vtag_entries(struct rvu *rvu, u16 pcifunc);
 static int rvu_nix_get_bpid(struct rvu *rvu, struct nix_bp_cfg_req *req,
 			    int type, int chan_id);
@@ -1149,36 +1151,36 @@ static int rvu_nix_blk_aq_enq_inst(struct rvu *rvu, struct nix_hw *nix_hw,
 	case NIX_AQ_INSTOP_WRITE:
 		if (req->ctype == NIX_AQ_CTYPE_RQ)
 			memcpy(mask, &req->rq_mask,
-			       sizeof(struct nix_rq_ctx_s));
+			       NIX_MAX_CTX_SIZE);
 		else if (req->ctype == NIX_AQ_CTYPE_SQ)
 			memcpy(mask, &req->sq_mask,
-			       sizeof(struct nix_sq_ctx_s));
+			       NIX_MAX_CTX_SIZE);
 		else if (req->ctype == NIX_AQ_CTYPE_CQ)
 			memcpy(mask, &req->cq_mask,
-			       sizeof(struct nix_cq_ctx_s));
+			       NIX_MAX_CTX_SIZE);
 		else if (req->ctype == NIX_AQ_CTYPE_RSS)
 			memcpy(mask, &req->rss_mask,
-			       sizeof(struct nix_rsse_s));
+			       NIX_MAX_CTX_SIZE);
 		else if (req->ctype == NIX_AQ_CTYPE_MCE)
 			memcpy(mask, &req->mce_mask,
-			       sizeof(struct nix_rx_mce_s));
+			       NIX_MAX_CTX_SIZE);
 		else if (req->ctype == NIX_AQ_CTYPE_BANDPROF)
 			memcpy(mask, &req->prof_mask,
-			       sizeof(struct nix_bandprof_s));
+			       NIX_MAX_CTX_SIZE);
 		fallthrough;
 	case NIX_AQ_INSTOP_INIT:
 		if (req->ctype == NIX_AQ_CTYPE_RQ)
-			memcpy(ctx, &req->rq, sizeof(struct nix_rq_ctx_s));
+			memcpy(ctx, &req->rq, NIX_MAX_CTX_SIZE);
 		else if (req->ctype == NIX_AQ_CTYPE_SQ)
-			memcpy(ctx, &req->sq, sizeof(struct nix_sq_ctx_s));
+			memcpy(ctx, &req->sq, NIX_MAX_CTX_SIZE);
 		else if (req->ctype == NIX_AQ_CTYPE_CQ)
-			memcpy(ctx, &req->cq, sizeof(struct nix_cq_ctx_s));
+			memcpy(ctx, &req->cq, NIX_MAX_CTX_SIZE);
 		else if (req->ctype == NIX_AQ_CTYPE_RSS)
-			memcpy(ctx, &req->rss, sizeof(struct nix_rsse_s));
+			memcpy(ctx, &req->rss, NIX_MAX_CTX_SIZE);
 		else if (req->ctype == NIX_AQ_CTYPE_MCE)
-			memcpy(ctx, &req->mce, sizeof(struct nix_rx_mce_s));
+			memcpy(ctx, &req->mce, NIX_MAX_CTX_SIZE);
 		else if (req->ctype == NIX_AQ_CTYPE_BANDPROF)
-			memcpy(ctx, &req->prof, sizeof(struct nix_bandprof_s));
+			memcpy(ctx, &req->prof, NIX_MAX_CTX_SIZE);
 		break;
 	case NIX_AQ_INSTOP_NOP:
 	case NIX_AQ_INSTOP_READ:
@@ -1243,22 +1245,22 @@ static int rvu_nix_blk_aq_enq_inst(struct rvu *rvu, struct nix_hw *nix_hw,
 		if (req->op == NIX_AQ_INSTOP_READ) {
 			if (req->ctype == NIX_AQ_CTYPE_RQ)
 				memcpy(&rsp->rq, ctx,
-				       sizeof(struct nix_rq_ctx_s));
+				       NIX_MAX_CTX_SIZE);
 			else if (req->ctype == NIX_AQ_CTYPE_SQ)
 				memcpy(&rsp->sq, ctx,
-				       sizeof(struct nix_sq_ctx_s));
+				       NIX_MAX_CTX_SIZE);
 			else if (req->ctype == NIX_AQ_CTYPE_CQ)
 				memcpy(&rsp->cq, ctx,
-				       sizeof(struct nix_cq_ctx_s));
+				       NIX_MAX_CTX_SIZE);
 			else if (req->ctype == NIX_AQ_CTYPE_RSS)
 				memcpy(&rsp->rss, ctx,
-				       sizeof(struct nix_rsse_s));
+				       NIX_MAX_CTX_SIZE);
 			else if (req->ctype == NIX_AQ_CTYPE_MCE)
 				memcpy(&rsp->mce, ctx,
-				       sizeof(struct nix_rx_mce_s));
+				       NIX_MAX_CTX_SIZE);
 			else if (req->ctype == NIX_AQ_CTYPE_BANDPROF)
 				memcpy(&rsp->prof, ctx,
-				       sizeof(struct nix_bandprof_s));
+				       NIX_MAX_CTX_SIZE);
 		}
 	}
 
@@ -1289,8 +1291,8 @@ static int rvu_nix_verify_aq_ctx(struct rvu *rvu, struct nix_hw *nix_hw,
 	/* Make copy of original context & mask which are required
 	 * for resubmission
 	 */
-	memcpy(&aq_req.cq_mask, &req->cq_mask, sizeof(struct nix_cq_ctx_s));
-	memcpy(&aq_req.cq, &req->cq, sizeof(struct nix_cq_ctx_s));
+	memcpy(&aq_req.cq_mask, &req->cq_mask, NIX_MAX_CTX_SIZE);
+	memcpy(&aq_req.cq, &req->cq, NIX_MAX_CTX_SIZE);
 
 	/* exclude fields which HW can update */
 	aq_req.cq_mask.cq_err       = 0;
@@ -1309,7 +1311,7 @@ static int rvu_nix_verify_aq_ctx(struct rvu *rvu, struct nix_hw *nix_hw,
 	 * updated fields are masked out for request and response
 	 * comparison
 	 */
-	for (word = 0; word < sizeof(struct nix_cq_ctx_s) / sizeof(u64);
+	for (word = 0; word < NIX_MAX_CTX_SIZE / sizeof(u64);
 	     word++) {
 		*(u64 *)((u8 *)&aq_rsp.cq + word * 8) &=
 			(*(u64 *)((u8 *)&aq_req.cq_mask + word * 8));
@@ -1317,7 +1319,7 @@ static int rvu_nix_verify_aq_ctx(struct rvu *rvu, struct nix_hw *nix_hw,
 			(*(u64 *)((u8 *)&aq_req.cq_mask + word * 8));
 	}
 
-	if (memcmp(&aq_req.cq, &aq_rsp.cq, sizeof(struct nix_cq_ctx_s)))
+	if (memcmp(&aq_req.cq, &aq_rsp.cq, NIX_MAX_CTX_SIZE))
 		return NIX_AF_ERR_AQ_CTX_RETRY_WRITE;
 
 	return 0;
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_struct.h b/drivers/net/ethernet/marvell/octeontx2/af/rvu_struct.h
index 0596a3ac4c12..8a66f53a7658 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_struct.h
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_struct.h
@@ -370,6 +370,8 @@ struct nix_cq_ctx_s {
 	u64 qsize		: 4;
 	u64 cq_err_int		: 8;
 	u64 cq_err_int_ena	: 8;
+	/* Ensure all context sizes are minimum 128 bytes */
+	u64 padding[12];
 };
 
 /* CN10K NIX Receive queue context structure */
@@ -672,7 +674,8 @@ struct nix_sq_ctx_s {
 struct nix_rsse_s {
 	uint32_t rq			: 20;
 	uint32_t reserved_20_31		: 12;
-
+	/* Ensure all context sizes are minimum 128 bytes */
+	u64 padding[15];
 };
 
 /* NIX receive multicast/mirror entry structure */
@@ -684,6 +687,8 @@ struct nix_rx_mce_s {
 	uint64_t rsvd_31_24 : 8;
 	uint64_t pf_func    : 16;
 	uint64_t next       : 16;
+	/* Ensure all context sizes are minimum 128 bytes */
+	u64 padding[15];
 };
 
 enum nix_band_prof_layers {
-- 
2.34.1

[net-next PATCH v2 02/11] octeontx2-af: Add cn20k NIX block contexts

[Subbaraya Sundeep]

New CN20K silicon has NIX hardware context structures different from
previous silicons. Add NIX send and completion queue context
definitions for cn20k. Extend NIX context handling support to cn20k.

Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com>
---
 .../ethernet/marvell/octeontx2/af/Makefile    |   2 +-
 .../ethernet/marvell/octeontx2/af/cn20k/nix.c |  20 ++
 .../marvell/octeontx2/af/cn20k/struct.h       | 196 ++++++++++++++++++
 .../net/ethernet/marvell/octeontx2/af/mbox.h  |  38 ++++
 .../net/ethernet/marvell/octeontx2/af/rvu.h   |   3 +
 .../ethernet/marvell/octeontx2/af/rvu_nix.c   |  10 +-
 6 files changed, 266 insertions(+), 3 deletions(-)
 create mode 100644 drivers/net/ethernet/marvell/octeontx2/af/cn20k/nix.c

diff --git a/drivers/net/ethernet/marvell/octeontx2/af/Makefile b/drivers/net/ethernet/marvell/octeontx2/af/Makefile
index 532813d8d028..cb77b978eda5 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/Makefile
+++ b/drivers/net/ethernet/marvell/octeontx2/af/Makefile
@@ -12,4 +12,4 @@ rvu_af-y := cgx.o rvu.o rvu_cgx.o rvu_npa.o rvu_nix.o \
 		  rvu_reg.o rvu_npc.o rvu_debugfs.o ptp.o rvu_npc_fs.o \
 		  rvu_cpt.o rvu_devlink.o rpm.o rvu_cn10k.o rvu_switch.o \
 		  rvu_sdp.o rvu_npc_hash.o mcs.o mcs_rvu_if.o mcs_cnf10kb.o \
-		  rvu_rep.o cn20k/mbox_init.o
+		  rvu_rep.o cn20k/mbox_init.o cn20k/nix.o
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/cn20k/nix.c b/drivers/net/ethernet/marvell/octeontx2/af/cn20k/nix.c
new file mode 100644
index 000000000000..aa2016fd1bba
--- /dev/null
+++ b/drivers/net/ethernet/marvell/octeontx2/af/cn20k/nix.c
@@ -0,0 +1,20 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Marvell RVU Admin Function driver
+ *
+ * Copyright (C) 2024 Marvell.
+ *
+ */
+
+#include <linux/module.h>
+#include <linux/pci.h>
+
+#include "struct.h"
+#include "../rvu.h"
+
+int rvu_mbox_handler_nix_cn20k_aq_enq(struct rvu *rvu,
+				      struct nix_cn20k_aq_enq_req *req,
+				      struct nix_cn20k_aq_enq_rsp *rsp)
+{
+	return rvu_nix_aq_enq_inst(rvu, (struct nix_aq_enq_req *)req,
+				  (struct nix_aq_enq_rsp *)rsp);
+}
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/cn20k/struct.h b/drivers/net/ethernet/marvell/octeontx2/af/cn20k/struct.h
index 76ce3ec6da9c..92744ae44853 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/cn20k/struct.h
+++ b/drivers/net/ethernet/marvell/octeontx2/af/cn20k/struct.h
@@ -37,4 +37,200 @@ enum rvu_af_cn20k_int_vec_e {
 	RVU_AF_CN20K_INT_VEC_PFAF1_MBOX1	= 0x9,
 	RVU_AF_CN20K_INT_VEC_CNT		= 0xa,
 };
+
+struct nix_cn20k_sq_ctx_s {
+	u64 ena                         :  1; /* W0 */
+	u64 qint_idx                    :  6;
+	u64 substream                   : 20;
+	u64 sdp_mcast                   :  1;
+	u64 cq                          : 20;
+	u64 sqe_way_mask                : 16;
+	u64 smq                         : 11; /* W1 */
+	u64 cq_ena                      :  1;
+	u64 xoff                        :  1;
+	u64 sso_ena                     :  1;
+	u64 smq_rr_weight               : 14;
+	u64 default_chan                : 12;
+	u64 sqb_count                   : 16;
+	u64 reserved_120_120            :  1;
+	u64 smq_rr_count_lb             :  7;
+	u64 smq_rr_count_ub             : 25; /* W2 */
+	u64 sqb_aura                    : 20;
+	u64 sq_int                      :  8;
+	u64 sq_int_ena                  :  8;
+	u64 sqe_stype                   :  2;
+	u64 reserved_191_191            :  1;
+	u64 max_sqe_size                :  2; /* W3 */
+	u64 cq_limit                    :  8;
+	u64 lmt_dis                     :  1;
+	u64 mnq_dis                     :  1;
+	u64 smq_next_sq                 : 20;
+	u64 smq_lso_segnum              :  8;
+	u64 tail_offset                 :  6;
+	u64 smenq_offset                :  6;
+	u64 head_offset                 :  6;
+	u64 smenq_next_sqb_vld          :  1;
+	u64 smq_pend                    :  1;
+	u64 smq_next_sq_vld             :  1;
+	u64 reserved_253_255            :  3;
+	u64 next_sqb                    : 64; /* W4 */
+	u64 tail_sqb                    : 64; /* W5 */
+	u64 smenq_sqb                   : 64; /* W6 */
+	u64 smenq_next_sqb              : 64; /* W7 */
+	u64 head_sqb                    : 64; /* W8 */
+	u64 reserved_576_583            :  8; /* W9 */
+	u64 vfi_lso_total               : 18;
+	u64 vfi_lso_sizem1              :  3;
+	u64 vfi_lso_sb                  :  8;
+	u64 vfi_lso_mps                 : 14;
+	u64 vfi_lso_vlan0_ins_ena       :  1;
+	u64 vfi_lso_vlan1_ins_ena       :  1;
+	u64 vfi_lso_vld                 :  1;
+	u64 reserved_630_639            : 10;
+	u64 scm_lso_rem                 : 18; /* W10 */
+	u64 reserved_658_703            : 46;
+	u64 octs                        : 48; /* W11 */
+	u64 reserved_752_767            : 16;
+	u64 pkts                        : 48; /* W12 */
+	u64 reserved_816_831            : 16;
+	u64 aged_drop_octs              : 32; /* W13 */
+	u64 aged_drop_pkts              : 32;
+	u64 dropped_octs                : 48; /* W14 */
+	u64 reserved_944_959            : 16;
+	u64 dropped_pkts                : 48; /* W15 */
+	u64 reserved_1008_1023          : 16;
+};
+
+struct nix_cn20k_cq_ctx_s {
+	u64 base                        : 64; /* W0 */
+	u64 lbp_ena                     :  1; /* W1 */
+	u64 lbpid_low                   :  3;
+	u64 bp_ena                      :  1;
+	u64 lbpid_med                   :  3;
+	u64 bpid                        :  9;
+	u64 lbpid_high                  :  3;
+	u64 qint_idx                    :  7;
+	u64 cq_err                      :  1;
+	u64 cint_idx                    :  7;
+	u64 avg_con                     :  9;
+	u64 wrptr                       : 20;
+	u64 tail                        : 20; /* W2 */
+	u64 head                        : 20;
+	u64 avg_level                   :  8;
+	u64 update_time                 : 16;
+	u64 bp                          :  8; /* W3 */
+	u64 drop                        :  8;
+	u64 drop_ena                    :  1;
+	u64 ena                         :  1;
+	u64 cpt_drop_err_en             :  1;
+	u64 reserved_211_211            :  1;
+	u64 msh_dst                     : 11;
+	u64 msh_valid                   :  1;
+	u64 stash_thresh                :  4;
+	u64 lbp_frac                    :  4;
+	u64 caching                     :  1;
+	u64 stashing                    :  1;
+	u64 reserved_234_235            :  2;
+	u64 qsize                       :  4;
+	u64 cq_err_int                  :  8;
+	u64 cq_err_int_ena              :  8;
+	u64 bpid_ext                    :  2; /* W4 */
+	u64 reserved_258_259            :  2;
+	u64 lbpid_ext                   :  2;
+	u64 reserved_262_319            : 58;
+	u64 reserved_320_383            : 64; /* W5 */
+	u64 reserved_384_447            : 64; /* W6 */
+	u64 reserved_448_511            : 64; /* W7 */
+};
+
+struct nix_cn20k_rq_ctx_s {
+	u64 ena                         :  1;
+	u64 sso_ena                     :  1;
+	u64 ipsech_ena                  :  1;
+	u64 ena_wqwd                    :  1;
+	u64 cq                          : 20;
+	u64 reserved_24_34              : 11;
+	u64 port_il4_dis                :  1;
+	u64 port_ol4_dis                :  1;
+	u64 lenerr_dis                  :  1;
+	u64 csum_il4_dis                :  1;
+	u64 csum_ol4_dis                :  1;
+	u64 len_il4_dis                 :  1;
+	u64 len_il3_dis                 :  1;
+	u64 len_ol4_dis                 :  1;
+	u64 len_ol3_dis                 :  1;
+	u64 wqe_aura                    : 20;
+	u64 spb_aura                    : 20;
+	u64 lpb_aura                    : 20;
+	u64 sso_grp                     : 10;
+	u64 sso_tt                      :  2;
+	u64 pb_caching                  :  2;
+	u64 wqe_caching                 :  1;
+	u64 xqe_drop_ena                :  1;
+	u64 spb_drop_ena                :  1;
+	u64 lpb_drop_ena                :  1;
+	u64 pb_stashing                 :  1;
+	u64 ipsecd_drop_en              :  1;
+	u64 chi_ena                     :  1;
+	u64 reserved_125_127            :  3;
+	u64 band_prof_id_l              : 10;
+	u64 sso_fc_ena                  :  1;
+	u64 policer_ena                 :  1;
+	u64 spb_sizem1                  :  6;
+	u64 wqe_skip                    :  2;
+	u64 spb_high_sizem1             :  3;
+	u64 spb_ena                     :  1;
+	u64 lpb_sizem1                  : 12;
+	u64 first_skip                  :  7;
+	u64 reserved_171_171            :  1;
+	u64 later_skip                  :  6;
+	u64 xqe_imm_size                :  6;
+	u64 band_prof_id_h              :  4;
+	u64 reserved_188_189            :  2;
+	u64 xqe_imm_copy                :  1;
+	u64 xqe_hdr_split               :  1;
+	u64 xqe_drop                    :  8;
+	u64 xqe_pass                    :  8;
+	u64 wqe_pool_drop               :  8;
+	u64 wqe_pool_pass               :  8;
+	u64 spb_aura_drop               :  8;
+	u64 spb_aura_pass               :  8;
+	u64 spb_pool_drop               :  8;
+	u64 spb_pool_pass               :  8;
+	u64 lpb_aura_drop               :  8;
+	u64 lpb_aura_pass               :  8;
+	u64 lpb_pool_drop               :  8;
+	u64 lpb_pool_pass               :  8;
+	u64 reserved_288_291            :  4;
+	u64 rq_int                      :  8;
+	u64 rq_int_ena                  :  8;
+	u64 qint_idx                    :  7;
+	u64 reserved_315_319            :  5;
+	u64 ltag                        : 24;
+	u64 good_utag                   :  8;
+	u64 bad_utag                    :  8;
+	u64 flow_tagw                   :  6;
+	u64 ipsec_vwqe                  :  1;
+	u64 vwqe_ena                    :  1;
+	u64 vtime_wait                  :  8;
+	u64 max_vsize_exp               :  4;
+	u64 vwqe_skip                   :  2;
+	u64 reserved_382_383            :  2;
+	u64 octs                        : 48;
+	u64 reserved_432_447            : 16;
+	u64 pkts                        : 48;
+	u64 reserved_496_511            : 16;
+	u64 drop_octs                   : 48;
+	u64 reserved_560_575            : 16;
+	u64 drop_pkts                   : 48;
+	u64 reserved_624_639            : 16;
+	u64 re_pkts                     : 48;
+	u64 reserved_688_703            : 16;
+	u64 reserved_704_767            : 64;
+	u64 reserved_768_831            : 64;
+	u64 reserved_832_895            : 64;
+	u64 reserved_896_959            : 64;
+	u64 reserved_960_1023           : 64;
+};
+
 #endif
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/mbox.h b/drivers/net/ethernet/marvell/octeontx2/af/mbox.h
index 0bc0dc79868b..544efd4817c1 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/mbox.h
+++ b/drivers/net/ethernet/marvell/octeontx2/af/mbox.h
@@ -336,6 +336,8 @@ M(NIX_MCAST_GRP_UPDATE, 0x802d, nix_mcast_grp_update,				\
 				nix_mcast_grp_update_req,			\
 				nix_mcast_grp_update_rsp)			\
 M(NIX_LF_STATS, 0x802e, nix_lf_stats, nix_stats_req, nix_stats_rsp)	\
+M(NIX_CN20K_AQ_ENQ,	0x802f, nix_cn20k_aq_enq, nix_cn20k_aq_enq_req,		\
+				nix_cn20k_aq_enq_rsp)				\
 /* MCS mbox IDs (range 0xA000 - 0xBFFF) */					\
 M(MCS_ALLOC_RESOURCES,	0xa000, mcs_alloc_resources, mcs_alloc_rsrc_req,	\
 				mcs_alloc_rsrc_rsp)				\
@@ -935,6 +937,42 @@ struct nix_lf_free_req {
 	u64 flags;
 };
 
+/* CN20K NIX AQ enqueue msg */
+struct nix_cn20k_aq_enq_req {
+	struct mbox_msghdr hdr;
+	u32  qidx;
+	u8 ctype;
+	u8 op;
+	union {
+		struct nix_cn20k_rq_ctx_s rq;
+		struct nix_cn20k_sq_ctx_s sq;
+		struct nix_cn20k_cq_ctx_s cq;
+		struct nix_rsse_s   rss;
+		struct nix_rx_mce_s mce;
+		struct nix_bandprof_s prof;
+	};
+	union {
+		struct nix_cn20k_rq_ctx_s rq_mask;
+		struct nix_cn20k_sq_ctx_s sq_mask;
+		struct nix_cn20k_cq_ctx_s cq_mask;
+		struct nix_rsse_s   rss_mask;
+		struct nix_rx_mce_s mce_mask;
+		struct nix_bandprof_s prof_mask;
+	};
+};
+
+struct nix_cn20k_aq_enq_rsp {
+	struct mbox_msghdr hdr;
+	union {
+		struct nix_cn20k_rq_ctx_s rq;
+		struct nix_cn20k_sq_ctx_s sq;
+		struct nix_cn20k_cq_ctx_s cq;
+		struct nix_rsse_s   rss;
+		struct nix_rx_mce_s mce;
+		struct nix_bandprof_s prof;
+	};
+};
+
 /* CN10K NIX AQ enqueue msg */
 struct nix_cn10k_aq_enq_req {
 	struct mbox_msghdr hdr;
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu.h b/drivers/net/ethernet/marvell/octeontx2/af/rvu.h
index 7ee1fdeb5295..5a7487422bd5 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/rvu.h
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu.h
@@ -1017,6 +1017,9 @@ int rvu_nix_mcast_update_mcam_entry(struct rvu *rvu, u16 pcifunc,
 void rvu_nix_flr_free_bpids(struct rvu *rvu, u16 pcifunc);
 int rvu_alloc_cint_qint_mem(struct rvu *rvu, struct rvu_pfvf *pfvf,
 			    int blkaddr, int nixlf);
+int rvu_nix_aq_enq_inst(struct rvu *rvu, struct nix_aq_enq_req *req,
+			struct nix_aq_enq_rsp *rsp);
+
 /* NPC APIs */
 void rvu_npc_freemem(struct rvu *rvu);
 int rvu_npc_get_pkind(struct rvu *rvu, u16 pf);
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c b/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c
index 48d44911b663..47a64c4c5b7e 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c
@@ -1021,6 +1021,12 @@ static void nix_get_aq_req_smq(struct rvu *rvu, struct nix_aq_enq_req *req,
 {
 	struct nix_cn10k_aq_enq_req *aq_req;
 
+	if (is_cn20k(rvu->pdev)) {
+		*smq = ((struct nix_cn20k_aq_enq_req *)req)->sq.smq;
+		*smq_mask = ((struct nix_cn20k_aq_enq_req *)req)->sq_mask.smq;
+		return;
+	}
+
 	if (!is_rvu_otx2(rvu)) {
 		aq_req = (struct nix_cn10k_aq_enq_req *)req;
 		*smq = aq_req->sq.smq;
@@ -1325,8 +1331,8 @@ static int rvu_nix_verify_aq_ctx(struct rvu *rvu, struct nix_hw *nix_hw,
 	return 0;
 }
 
-static int rvu_nix_aq_enq_inst(struct rvu *rvu, struct nix_aq_enq_req *req,
-			       struct nix_aq_enq_rsp *rsp)
+int rvu_nix_aq_enq_inst(struct rvu *rvu, struct nix_aq_enq_req *req,
+			struct nix_aq_enq_rsp *rsp)
 {
 	struct nix_hw *nix_hw;
 	int err, retries = 5;
-- 
2.34.1

[net-next PATCH v2 03/11] octeontx2-af: Extend debugfs support for cn20k NIX

[Subbaraya Sundeep]

Extend debugfs to display CN20K NIX send, receive and
completion queue contexts.

Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com>
---
 .../ethernet/marvell/octeontx2/af/Makefile    |   2 +-
 .../marvell/octeontx2/af/cn20k/debugfs.c      | 132 ++++++++++++++++++
 .../marvell/octeontx2/af/cn20k/debugfs.h      |  24 ++++
 .../marvell/octeontx2/af/rvu_debugfs.c        |  17 +++
 4 files changed, 174 insertions(+), 1 deletion(-)
 create mode 100644 drivers/net/ethernet/marvell/octeontx2/af/cn20k/debugfs.c
 create mode 100644 drivers/net/ethernet/marvell/octeontx2/af/cn20k/debugfs.h

diff --git a/drivers/net/ethernet/marvell/octeontx2/af/Makefile b/drivers/net/ethernet/marvell/octeontx2/af/Makefile
index cb77b978eda5..57eeaa116116 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/Makefile
+++ b/drivers/net/ethernet/marvell/octeontx2/af/Makefile
@@ -12,4 +12,4 @@ rvu_af-y := cgx.o rvu.o rvu_cgx.o rvu_npa.o rvu_nix.o \
 		  rvu_reg.o rvu_npc.o rvu_debugfs.o ptp.o rvu_npc_fs.o \
 		  rvu_cpt.o rvu_devlink.o rpm.o rvu_cn10k.o rvu_switch.o \
 		  rvu_sdp.o rvu_npc_hash.o mcs.o mcs_rvu_if.o mcs_cnf10kb.o \
-		  rvu_rep.o cn20k/mbox_init.o cn20k/nix.o
+		  rvu_rep.o cn20k/mbox_init.o cn20k/nix.o cn20k/debugfs.o
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/cn20k/debugfs.c b/drivers/net/ethernet/marvell/octeontx2/af/cn20k/debugfs.c
new file mode 100644
index 000000000000..d39d8ea907ea
--- /dev/null
+++ b/drivers/net/ethernet/marvell/octeontx2/af/cn20k/debugfs.c
@@ -0,0 +1,132 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Marvell RVU Admin Function driver
+ *
+ * Copyright (C) 2024 Marvell.
+ *
+ */
+
+#include <linux/fs.h>
+#include <linux/debugfs.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+
+#include "struct.h"
+#include "debugfs.h"
+
+void print_nix_cn20k_sq_ctx(struct seq_file *m,
+			    struct nix_cn20k_sq_ctx_s *sq_ctx)
+{
+	seq_printf(m, "W0: ena \t\t\t%d\nW0: qint_idx \t\t\t%d\n",
+		   sq_ctx->ena, sq_ctx->qint_idx);
+	seq_printf(m, "W0: substream \t\t\t0x%03x\nW0: sdp_mcast \t\t\t%d\n",
+		   sq_ctx->substream, sq_ctx->sdp_mcast);
+	seq_printf(m, "W0: cq \t\t\t\t%d\nW0: sqe_way_mask \t\t%d\n\n",
+		   sq_ctx->cq, sq_ctx->sqe_way_mask);
+
+	seq_printf(m, "W1: smq \t\t\t%d\nW1: cq_ena \t\t\t%d\nW1: xoff\t\t\t%d\n",
+		   sq_ctx->smq, sq_ctx->cq_ena, sq_ctx->xoff);
+	seq_printf(m, "W1: sso_ena \t\t\t%d\nW1: smq_rr_weight\t\t%d\n",
+		   sq_ctx->sso_ena, sq_ctx->smq_rr_weight);
+	seq_printf(m, "W1: default_chan\t\t%d\nW1: sqb_count\t\t\t%d\n\n",
+		   sq_ctx->default_chan, sq_ctx->sqb_count);
+
+	seq_printf(m, "W1: smq_rr_count_lb \t\t%d\n", sq_ctx->smq_rr_count_lb);
+	seq_printf(m, "W2: smq_rr_count_ub \t\t%d\n", sq_ctx->smq_rr_count_ub);
+	seq_printf(m, "W2: sqb_aura \t\t\t%d\nW2: sq_int \t\t\t%d\n",
+		   sq_ctx->sqb_aura, sq_ctx->sq_int);
+	seq_printf(m, "W2: sq_int_ena \t\t\t%d\nW2: sqe_stype \t\t\t%d\n",
+		   sq_ctx->sq_int_ena, sq_ctx->sqe_stype);
+
+	seq_printf(m, "W3: max_sqe_size\t\t%d\nW3: cq_limit\t\t\t%d\n",
+		   sq_ctx->max_sqe_size, sq_ctx->cq_limit);
+	seq_printf(m, "W3: lmt_dis \t\t\t%d\nW3: mnq_dis \t\t\t%d\n",
+		   sq_ctx->lmt_dis, sq_ctx->mnq_dis);
+	seq_printf(m, "W3: smq_next_sq\t\t\t%d\nW3: smq_lso_segnum\t\t%d\n",
+		   sq_ctx->smq_next_sq, sq_ctx->smq_lso_segnum);
+	seq_printf(m, "W3: tail_offset \t\t%d\nW3: smenq_offset\t\t%d\n",
+		   sq_ctx->tail_offset, sq_ctx->smenq_offset);
+	seq_printf(m, "W3: head_offset\t\t\t%d\nW3: smenq_next_sqb_vld\t\t%d\n\n",
+		   sq_ctx->head_offset, sq_ctx->smenq_next_sqb_vld);
+
+	seq_printf(m, "W3: smq_next_sq_vld\t\t%d\nW3: smq_pend\t\t\t%d\n",
+		   sq_ctx->smq_next_sq_vld, sq_ctx->smq_pend);
+	seq_printf(m, "W4: next_sqb \t\t\t%llx\n\n", sq_ctx->next_sqb);
+	seq_printf(m, "W5: tail_sqb \t\t\t%llx\n\n", sq_ctx->tail_sqb);
+	seq_printf(m, "W6: smenq_sqb \t\t\t%llx\n\n", sq_ctx->smenq_sqb);
+	seq_printf(m, "W7: smenq_next_sqb \t\t%llx\n\n",
+		   sq_ctx->smenq_next_sqb);
+
+	seq_printf(m, "W8: head_sqb\t\t\t%llx\n\n", sq_ctx->head_sqb);
+
+	seq_printf(m, "W9: vfi_lso_total\t\t%d\n", sq_ctx->vfi_lso_total);
+	seq_printf(m, "W9: vfi_lso_sizem1\t\t%d\nW9: vfi_lso_sb\t\t\t%d\n",
+		   sq_ctx->vfi_lso_sizem1, sq_ctx->vfi_lso_sb);
+	seq_printf(m, "W9: vfi_lso_mps\t\t\t%d\nW9: vfi_lso_vlan0_ins_ena\t%d\n",
+		   sq_ctx->vfi_lso_mps, sq_ctx->vfi_lso_vlan0_ins_ena);
+	seq_printf(m, "W9: vfi_lso_vlan1_ins_ena\t%d\nW9: vfi_lso_vld \t\t%d\n\n",
+		   sq_ctx->vfi_lso_vld, sq_ctx->vfi_lso_vlan1_ins_ena);
+
+	seq_printf(m, "W10: scm_lso_rem \t\t%llu\n\n",
+		   (u64)sq_ctx->scm_lso_rem);
+	seq_printf(m, "W11: octs \t\t\t%llu\n\n", (u64)sq_ctx->octs);
+	seq_printf(m, "W12: pkts \t\t\t%llu\n\n", (u64)sq_ctx->pkts);
+	seq_printf(m, "W13: aged_drop_octs \t\t\t%llu\n\n", (u64)sq_ctx->aged_drop_octs);
+	seq_printf(m, "W13: aged_drop_pkts \t\t\t%llu\n\n", (u64)sq_ctx->aged_drop_pkts);
+	seq_printf(m, "W14: dropped_octs \t\t%llu\n\n",
+		   (u64)sq_ctx->dropped_octs);
+	seq_printf(m, "W15: dropped_pkts \t\t%llu\n\n",
+		   (u64)sq_ctx->dropped_pkts);
+}
+
+void print_nix_cn20k_cq_ctx(struct seq_file *m,
+			    struct nix_cn20k_aq_enq_rsp *rsp)
+{
+	struct nix_cn20k_cq_ctx_s *cq_ctx = &rsp->cq;
+
+	seq_printf(m, "W0: base \t\t\t%llx\n\n", cq_ctx->base);
+
+	seq_printf(m, "W1: wrptr \t\t\t%llx\n", (u64)cq_ctx->wrptr);
+	seq_printf(m, "W1: avg_con \t\t\t%d\nW1: cint_idx \t\t\t%d\n",
+		   cq_ctx->avg_con, cq_ctx->cint_idx);
+	seq_printf(m, "W1: cq_err \t\t\t%d\nW1: qint_idx \t\t\t%d\n",
+		   cq_ctx->cq_err, cq_ctx->qint_idx);
+	seq_printf(m, "W1: bpid \t\t\t%d\nW1: bp_ena \t\t\t%d\n\n",
+		   cq_ctx->bpid, cq_ctx->bp_ena);
+
+	seq_printf(m, "W1: lbpid_high \t\t\t0x%03x\n", cq_ctx->lbpid_high);
+	seq_printf(m, "W1: lbpid_med \t\t\t0x%03x\n", cq_ctx->lbpid_med);
+	seq_printf(m, "W1: lbpid_low \t\t\t0x%03x\n", cq_ctx->lbpid_low);
+	seq_printf(m, "(W1: lbpid) \t\t\t0x%03x\n",
+		   cq_ctx->lbpid_high << 6 | cq_ctx->lbpid_med << 3 |
+		   cq_ctx->lbpid_low);
+	seq_printf(m, "W1: lbp_ena \t\t\t\t%d\n\n", cq_ctx->lbp_ena);
+
+	seq_printf(m, "W2: update_time \t\t%d\nW2:avg_level \t\t\t%d\n",
+		   cq_ctx->update_time, cq_ctx->avg_level);
+	seq_printf(m, "W2: head \t\t\t%d\nW2:tail \t\t\t%d\n\n",
+		   cq_ctx->head, cq_ctx->tail);
+
+	seq_printf(m, "W3: cq_err_int_ena \t\t%d\nW3:cq_err_int \t\t\t%d\n",
+		   cq_ctx->cq_err_int_ena, cq_ctx->cq_err_int);
+	seq_printf(m, "W3: qsize \t\t\t%d\nW3:stashing \t\t\t%d\n",
+		   cq_ctx->qsize, cq_ctx->stashing);
+
+	seq_printf(m, "W3: caching \t\t\t%d\n", cq_ctx->caching);
+	seq_printf(m, "W3: lbp_frac \t\t\t%d\n", cq_ctx->lbp_frac);
+	seq_printf(m, "W3: stash_thresh \t\t\t%d\n",
+		   cq_ctx->stash_thresh);
+
+	seq_printf(m, "W3: msh_valid \t\t\t%d\nW3:msh_dst \t\t\t%d\n",
+		   cq_ctx->msh_valid, cq_ctx->msh_dst);
+
+	seq_printf(m, "W3: cpt_drop_err_en \t\t\t%d\n",
+		   cq_ctx->cpt_drop_err_en);
+	seq_printf(m, "W3: ena \t\t\t%d\n",
+		   cq_ctx->ena);
+	seq_printf(m, "W3: drop_ena \t\t\t%d\nW3: drop \t\t\t%d\n",
+		   cq_ctx->drop_ena, cq_ctx->drop);
+	seq_printf(m, "W3: bp \t\t\t\t%d\n\n", cq_ctx->bp);
+
+	seq_printf(m, "W4: lbpid_ext \t\t\t\t%d\n\n", cq_ctx->lbpid_ext);
+	seq_printf(m, "W4: bpid_ext \t\t\t\t%d\n\n", cq_ctx->bpid_ext);
+}
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/cn20k/debugfs.h b/drivers/net/ethernet/marvell/octeontx2/af/cn20k/debugfs.h
new file mode 100644
index 000000000000..9d3a98dc3000
--- /dev/null
+++ b/drivers/net/ethernet/marvell/octeontx2/af/cn20k/debugfs.h
@@ -0,0 +1,24 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Marvell OcteonTx2 CGX driver
+ *
+ * Copyright (C) 2024 Marvell.
+ *
+ */
+
+#ifndef DEBUFS_H
+#define DEBUFS_H
+
+#include <linux/fs.h>
+#include <linux/debugfs.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+
+#include "struct.h"
+#include "../mbox.h"
+
+void print_nix_cn20k_sq_ctx(struct seq_file *m,
+			    struct nix_cn20k_sq_ctx_s *sq_ctx);
+void print_nix_cn20k_cq_ctx(struct seq_file *m,
+			    struct nix_cn20k_aq_enq_rsp *rsp);
+
+#endif
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c b/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c
index 0c20642f81b9..3d7a4f923c04 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c
@@ -21,6 +21,8 @@
 #include "rvu_npc_hash.h"
 #include "mcs.h"
 
+#include "cn20k/debugfs.h"
+
 #define DEBUGFS_DIR_NAME "octeontx2"
 
 enum {
@@ -1944,10 +1946,16 @@ static void print_nix_sq_ctx(struct seq_file *m, struct nix_aq_enq_rsp *rsp)
 	struct nix_hw *nix_hw = m->private;
 	struct rvu *rvu = nix_hw->rvu;
 
+	if (is_cn20k(rvu->pdev)) {
+		print_nix_cn20k_sq_ctx(m, (struct nix_cn20k_sq_ctx_s *)sq_ctx);
+		return;
+	}
+
 	if (!is_rvu_otx2(rvu)) {
 		print_nix_cn10k_sq_ctx(m, (struct nix_cn10k_sq_ctx_s *)sq_ctx);
 		return;
 	}
+
 	seq_printf(m, "W0: sqe_way_mask \t\t%d\nW0: cq \t\t\t\t%d\n",
 		   sq_ctx->sqe_way_mask, sq_ctx->cq);
 	seq_printf(m, "W0: sdp_mcast \t\t\t%d\nW0: substream \t\t\t0x%03x\n",
@@ -2160,6 +2168,11 @@ static void print_nix_cq_ctx(struct seq_file *m, struct nix_aq_enq_rsp *rsp)
 	struct nix_hw *nix_hw = m->private;
 	struct rvu *rvu = nix_hw->rvu;
 
+	if (is_cn20k(rvu->pdev)) {
+		print_nix_cn20k_cq_ctx(m, (struct nix_cn20k_aq_enq_rsp *)rsp);
+		return;
+	}
+
 	seq_printf(m, "W0: base \t\t\t%llx\n\n", cq_ctx->base);
 
 	seq_printf(m, "W1: wrptr \t\t\t%llx\n", (u64)cq_ctx->wrptr);
@@ -2189,6 +2202,7 @@ static void print_nix_cq_ctx(struct seq_file *m, struct nix_aq_enq_rsp *rsp)
 		   cq_ctx->cq_err_int_ena, cq_ctx->cq_err_int);
 	seq_printf(m, "W3: qsize \t\t\t%d\nW3:caching \t\t\t%d\n",
 		   cq_ctx->qsize, cq_ctx->caching);
+
 	seq_printf(m, "W3: substream \t\t\t0x%03x\nW3: ena \t\t\t%d\n",
 		   cq_ctx->substream, cq_ctx->ena);
 	if (!is_rvu_otx2(rvu)) {
@@ -3791,6 +3805,9 @@ static void rvu_dbg_cpt_init(struct rvu *rvu, int blkaddr)
 
 static const char *rvu_get_dbg_dir_name(struct rvu *rvu)
 {
+	if (is_cn20k(rvu->pdev))
+		return "cn20k";
+
 	if (!is_rvu_otx2(rvu))
 		return "cn10k";
 	else
-- 
2.34.1

[net-next PATCH v2 04/11] octeontx2-af: Add cn20k NPA block contexts

[Subbaraya Sundeep]

From: Linu Cherian <lcherian@marvell.com>

New CN20K silicon has NPA hardware context structures different from
previous silicons. Add NPA aura and pool context definitions for cn20k.
Extend NPA context handling support to cn20k.

Signed-off-by: Linu Cherian <lcherian@marvell.com>
Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com>
---
 .../ethernet/marvell/octeontx2/af/Makefile    |   3 +-
 .../ethernet/marvell/octeontx2/af/cn20k/npa.c |  21 +++
 .../marvell/octeontx2/af/cn20k/struct.h       | 130 ++++++++++++++++++
 .../net/ethernet/marvell/octeontx2/af/mbox.h  |  35 +++++
 4 files changed, 188 insertions(+), 1 deletion(-)
 create mode 100644 drivers/net/ethernet/marvell/octeontx2/af/cn20k/npa.c

diff --git a/drivers/net/ethernet/marvell/octeontx2/af/Makefile b/drivers/net/ethernet/marvell/octeontx2/af/Makefile
index 57eeaa116116..244de500963e 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/Makefile
+++ b/drivers/net/ethernet/marvell/octeontx2/af/Makefile
@@ -12,4 +12,5 @@ rvu_af-y := cgx.o rvu.o rvu_cgx.o rvu_npa.o rvu_nix.o \
 		  rvu_reg.o rvu_npc.o rvu_debugfs.o ptp.o rvu_npc_fs.o \
 		  rvu_cpt.o rvu_devlink.o rpm.o rvu_cn10k.o rvu_switch.o \
 		  rvu_sdp.o rvu_npc_hash.o mcs.o mcs_rvu_if.o mcs_cnf10kb.o \
-		  rvu_rep.o cn20k/mbox_init.o cn20k/nix.o cn20k/debugfs.o
+		  rvu_rep.o cn20k/mbox_init.o cn20k/nix.o cn20k/debugfs.o \
+		  cn20k/npa.o
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/cn20k/npa.c b/drivers/net/ethernet/marvell/octeontx2/af/cn20k/npa.c
new file mode 100644
index 000000000000..fe8f926c8b75
--- /dev/null
+++ b/drivers/net/ethernet/marvell/octeontx2/af/cn20k/npa.c
@@ -0,0 +1,21 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Marvell RVU Admin Function driver
+ *
+ * Copyright (C) 2024 Marvell.
+ *
+ */
+
+#include <linux/module.h>
+#include <linux/pci.h>
+
+#include "struct.h"
+#include "../rvu.h"
+
+int rvu_mbox_handler_npa_cn20k_aq_enq(struct rvu *rvu,
+				      struct npa_cn20k_aq_enq_req *req,
+				      struct npa_cn20k_aq_enq_rsp *rsp)
+{
+	return rvu_npa_aq_enq_inst(rvu, (struct npa_aq_enq_req *)req,
+				   (struct npa_aq_enq_rsp *)rsp);
+}
+EXPORT_SYMBOL(rvu_mbox_handler_npa_cn20k_aq_enq);
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/cn20k/struct.h b/drivers/net/ethernet/marvell/octeontx2/af/cn20k/struct.h
index 92744ae44853..405ac85e6237 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/cn20k/struct.h
+++ b/drivers/net/ethernet/marvell/octeontx2/af/cn20k/struct.h
@@ -233,4 +233,134 @@ struct nix_cn20k_rq_ctx_s {
 	u64 reserved_960_1023           : 64;
 };
 
+struct npa_cn20k_aura_s {
+	u64 pool_addr;			/* W0 */
+	u64 ena                   : 1;  /* W1 */
+	u64 reserved_65           : 2;
+	u64 pool_caching          : 1;
+	u64 reserved_68           : 16;
+	u64 avg_con               : 9;
+	u64 reserved_93           : 1;
+	u64 pool_drop_ena         : 1;
+	u64 aura_drop_ena         : 1;
+	u64 bp_ena                : 1;
+	u64 reserved_97_103       : 7;
+	u64 aura_drop             : 8;
+	u64 shift                 : 6;
+	u64 reserved_118_119      : 2;
+	u64 avg_level             : 8;
+	u64 count                 : 36; /* W2 */
+	u64 reserved_164_167      : 4;
+	u64 bpid                  : 12;
+	u64 reserved_180_191      : 12;
+	u64 limit                 : 36; /* W3 */
+	u64 reserved_228_231      : 4;
+	u64 bp                    : 7;
+	u64 reserved_239_243      : 5;
+	u64 fc_ena                : 1;
+	u64 fc_up_crossing        : 1;
+	u64 fc_stype              : 2;
+	u64 fc_hyst_bits          : 4;
+	u64 reserved_252_255      : 4;
+	u64 fc_addr;			/* W4 */
+	u64 pool_drop             : 8;  /* W5 */
+	u64 update_time           : 16;
+	u64 err_int               : 8;
+	u64 err_int_ena           : 8;
+	u64 thresh_int            : 1;
+	u64 thresh_int_ena        : 1;
+	u64 thresh_up             : 1;
+	u64 reserved_363          : 1;
+	u64 thresh_qint_idx       : 7;
+	u64 reserved_371          : 1;
+	u64 err_qint_idx          : 7;
+	u64 reserved_379_383      : 5;
+	u64 thresh                : 36; /* W6*/
+	u64 rsvd_423_420          : 4;
+	u64 fc_msh_dst            : 11;
+	u64 reserved_435_438      : 4;
+	u64 op_dpc_ena            : 1;
+	u64 op_dpc_set            : 5;
+	u64 reserved_445_445      : 1;
+	u64 stream_ctx            : 1;
+	u64 unified_ctx           : 1;
+	u64 reserved_448_511;		/* W7 */
+};
+
+struct npa_cn20k_pool_s {
+	u64 stack_base;			/* W0 */
+	u64 ena                   : 1;
+	u64 nat_align             : 1;
+	u64 reserved_66_67        : 2;
+	u64 stack_caching         : 1;
+	u64 reserved_69_87        : 19;
+	u64 buf_offset            : 12;
+	u64 reserved_100_103      : 4;
+	u64 buf_size              : 12;
+	u64 reserved_116_119      : 4;
+	u64 ref_cnt_prof          : 3;
+	u64 reserved_123_127      : 5;
+	u64 stack_max_pages       : 32;
+	u64 stack_pages           : 32;
+	uint64_t bp_0             : 7;
+	uint64_t bp_1             : 7;
+	uint64_t bp_2             : 7;
+	uint64_t bp_3             : 7;
+	uint64_t bp_4             : 7;
+	uint64_t bp_5             : 7;
+	uint64_t bp_6             : 7;
+	uint64_t bp_7             : 7;
+	uint64_t bp_ena_0         : 1;
+	uint64_t bp_ena_1         : 1;
+	uint64_t bp_ena_2         : 1;
+	uint64_t bp_ena_3         : 1;
+	uint64_t bp_ena_4         : 1;
+	uint64_t bp_ena_5         : 1;
+	uint64_t bp_ena_6         : 1;
+	uint64_t bp_ena_7         : 1;
+	u64 stack_offset          : 4;
+	u64 reserved_260_263      : 4;
+	u64 shift                 : 6;
+	u64 reserved_270_271      : 2;
+	u64 avg_level             : 8;
+	u64 avg_con               : 9;
+	u64 fc_ena                : 1;
+	u64 fc_stype              : 2;
+	u64 fc_hyst_bits          : 4;
+	u64 fc_up_crossing        : 1;
+	u64 reserved_297_299      : 3;
+	u64 update_time           : 16;
+	u64 reserved_316_319      : 4;
+	u64 fc_addr;			/* W5 */
+	u64 ptr_start;			/* W6 */
+	u64 ptr_end;			/* W7 */
+	u64 bpid_0                : 12;
+	u64 reserved_524_535      : 12;
+	u64 err_int               : 8;
+	u64 err_int_ena           : 8;
+	u64 thresh_int            : 1;
+	u64 thresh_int_ena        : 1;
+	u64 thresh_up             : 1;
+	u64 reserved_555          : 1;
+	u64 thresh_qint_idx       : 7;
+	u64 reserved_563          : 1;
+	u64 err_qint_idx          : 7;
+	u64 reserved_571_575      : 5;
+	u64 thresh                : 36;
+	u64 rsvd_612_615	  : 4;
+	u64 fc_msh_dst		  : 11;
+	u64 reserved_627_630      : 4;
+	u64 op_dpc_ena            : 1;
+	u64 op_dpc_set            : 5;
+	u64 reserved_637_637      : 1;
+	u64 stream_ctx            : 1;
+	u64 reserved_639          : 1;
+	u64 reserved_640_703;		/* W10 */
+	u64 reserved_704_767;		/* W11 */
+	u64 reserved_768_831;		/* W12 */
+	u64 reserved_832_895;		/* W13 */
+	u64 reserved_896_959;		/* W14 */
+	u64 reserved_960_1023;		/* W15 */
+};
+
 #endif
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/mbox.h b/drivers/net/ethernet/marvell/octeontx2/af/mbox.h
index 544efd4817c1..682924da1d9e 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/mbox.h
+++ b/drivers/net/ethernet/marvell/octeontx2/af/mbox.h
@@ -203,6 +203,8 @@ M(NPA_LF_ALLOC,		0x400, npa_lf_alloc,				\
 M(NPA_LF_FREE,		0x401, npa_lf_free, msg_req, msg_rsp)		\
 M(NPA_AQ_ENQ,		0x402, npa_aq_enq, npa_aq_enq_req, npa_aq_enq_rsp)   \
 M(NPA_HWCTX_DISABLE,	0x403, npa_hwctx_disable, hwctx_disable_req, msg_rsp)\
+M(NPA_CN20K_AQ_ENQ,	0x404, npa_cn20k_aq_enq, npa_cn20k_aq_enq_req,	\
+				npa_cn20k_aq_enq_rsp)			\
 /* SSO/SSOW mbox IDs (range 0x600 - 0x7FF) */				\
 /* TIM mbox IDs (range 0x800 - 0x9FF) */				\
 /* CPT mbox IDs (range 0xA00 - 0xBFF) */				\
@@ -829,6 +831,39 @@ struct npa_aq_enq_rsp {
 	};
 };
 
+struct npa_cn20k_aq_enq_req {
+	struct mbox_msghdr hdr;
+	u32 aura_id;
+	u8 ctype;
+	u8 op;
+	union {
+		/* Valid when op == WRITE/INIT and ctype == AURA.
+		 * LF fills the pool_id in aura.pool_addr. AF will translate
+		 * the pool_id to pool context pointer.
+		 */
+		struct npa_cn20k_aura_s aura;
+		/* Valid when op == WRITE/INIT and ctype == POOL */
+		struct npa_cn20k_pool_s pool;
+	};
+	/* Mask data when op == WRITE (1=write, 0=don't write) */
+	union {
+		/* Valid when op == WRITE and ctype == AURA */
+		struct npa_cn20k_aura_s aura_mask;
+		/* Valid when op == WRITE and ctype == POOL */
+		struct npa_cn20k_pool_s pool_mask;
+	};
+};
+
+struct npa_cn20k_aq_enq_rsp {
+	struct mbox_msghdr hdr;
+	union {
+		/* Valid when op == READ and ctype == AURA */
+		struct npa_cn20k_aura_s aura;
+		/* Valid when op == READ and ctype == POOL */
+		struct npa_cn20k_pool_s pool;
+	};
+};
+
 /* Disable all contexts of type 'ctype' */
 struct hwctx_disable_req {
 	struct mbox_msghdr hdr;
-- 
2.34.1

[net-next PATCH v2 05/11] octeontx2-af: Extend debugfs support for cn20k NPA

[Subbaraya Sundeep]

From: Linu Cherian <lcherian@marvell.com>

Extend debugfs to display CN20K NPA aura and pool contexts.

Signed-off-by: Linu Cherian <lcherian@marvell.com>
Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com>
---
 .../marvell/octeontx2/af/cn20k/debugfs.c      | 84 +++++++++++++++++++
 .../marvell/octeontx2/af/cn20k/debugfs.h      |  4 +
 .../marvell/octeontx2/af/rvu_debugfs.c        | 10 +++
 3 files changed, 98 insertions(+)

diff --git a/drivers/net/ethernet/marvell/octeontx2/af/cn20k/debugfs.c b/drivers/net/ethernet/marvell/octeontx2/af/cn20k/debugfs.c
index d39d8ea907ea..1a41a241bc07 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/cn20k/debugfs.c
+++ b/drivers/net/ethernet/marvell/octeontx2/af/cn20k/debugfs.c
@@ -130,3 +130,87 @@ void print_nix_cn20k_cq_ctx(struct seq_file *m,
 	seq_printf(m, "W4: lbpid_ext \t\t\t\t%d\n\n", cq_ctx->lbpid_ext);
 	seq_printf(m, "W4: bpid_ext \t\t\t\t%d\n\n", cq_ctx->bpid_ext);
 }
+
+void print_npa_cn20k_aura_ctx(struct seq_file *m,
+			      struct npa_cn20k_aq_enq_rsp *rsp)
+{
+	struct npa_cn20k_aura_s *aura = &rsp->aura;
+
+	seq_printf(m, "W0: Pool addr\t\t%llx\n", aura->pool_addr);
+
+	seq_printf(m, "W1: ena\t\t\t%d\nW1: pool caching\t%d\n",
+		   aura->ena, aura->pool_caching);
+	seq_printf(m, "W1: avg con\t\t%d\n", aura->avg_con);
+	seq_printf(m, "W1: pool drop ena\t%d\nW1: aura drop ena\t%d\n",
+		   aura->pool_drop_ena, aura->aura_drop_ena);
+	seq_printf(m, "W1: bp_ena\t\t%d\nW1: aura drop\t\t%d\n",
+		   aura->bp_ena, aura->aura_drop);
+	seq_printf(m, "W1: aura shift\t\t%d\nW1: avg_level\t\t%d\n",
+		   aura->shift, aura->avg_level);
+
+	seq_printf(m, "W2: count\t\t%llu\nW2: nix_bpid\t\t%d\n",
+		   (u64)aura->count, aura->bpid);
+
+	seq_printf(m, "W3: limit\t\t%llu\nW3: bp\t\t\t%d\nW3: fc_ena\t\t%d\n",
+		   (u64)aura->limit, aura->bp, aura->fc_ena);
+
+	seq_printf(m, "W3: fc_up_crossing\t%d\nW3: fc_stype\t\t%d\n",
+		   aura->fc_up_crossing, aura->fc_stype);
+	seq_printf(m, "W3: fc_hyst_bits\t%d\n", aura->fc_hyst_bits);
+
+	seq_printf(m, "W4: fc_addr\t\t%llx\n", aura->fc_addr);
+
+	seq_printf(m, "W5: pool_drop\t\t%d\nW5: update_time\t\t%d\n",
+		   aura->pool_drop, aura->update_time);
+	seq_printf(m, "W5: err_int \t\t%d\nW5: err_int_ena\t\t%d\n",
+		   aura->err_int, aura->err_int_ena);
+	seq_printf(m, "W5: thresh_int\t\t%d\nW5: thresh_int_ena \t%d\n",
+		   aura->thresh_int, aura->thresh_int_ena);
+	seq_printf(m, "W5: thresh_up\t\t%d\nW5: thresh_qint_idx\t%d\n",
+		   aura->thresh_up, aura->thresh_qint_idx);
+	seq_printf(m, "W5: err_qint_idx \t%d\n", aura->err_qint_idx);
+
+	seq_printf(m, "W6: thresh\t\t%llu\n", (u64)aura->thresh);
+	seq_printf(m, "W6: fc_msh_dst\t\t%d\n", aura->fc_msh_dst);
+}
+
+void print_npa_cn20k_pool_ctx(struct seq_file *m,
+			      struct npa_cn20k_aq_enq_rsp *rsp)
+{
+	struct npa_cn20k_pool_s *pool = &rsp->pool;
+
+	seq_printf(m, "W0: Stack base\t\t%llx\n", pool->stack_base);
+
+	seq_printf(m, "W1: ena \t\t%d\nW1: nat_align \t\t%d\n",
+		   pool->ena, pool->nat_align);
+	seq_printf(m, "W1: stack_caching\t%d\n",
+		   pool->stack_caching);
+	seq_printf(m, "W1: buf_offset\t\t%d\nW1: buf_size\t\t%d\n",
+		   pool->buf_offset, pool->buf_size);
+
+	seq_printf(m, "W2: stack_max_pages \t%d\nW2: stack_pages\t\t%d\n",
+		   pool->stack_max_pages, pool->stack_pages);
+
+	seq_printf(m, "W4: stack_offset\t%d\nW4: shift\t\t%d\nW4: avg_level\t\t%d\n",
+		   pool->stack_offset, pool->shift, pool->avg_level);
+	seq_printf(m, "W4: avg_con \t\t%d\nW4: fc_ena\t\t%d\nW4: fc_stype\t\t%d\n",
+		   pool->avg_con, pool->fc_ena, pool->fc_stype);
+	seq_printf(m, "W4: fc_hyst_bits\t%d\nW4: fc_up_crossing\t%d\n",
+		   pool->fc_hyst_bits, pool->fc_up_crossing);
+	seq_printf(m, "W4: update_time\t\t%d\n", pool->update_time);
+
+	seq_printf(m, "W5: fc_addr\t\t%llx\n", pool->fc_addr);
+
+	seq_printf(m, "W6: ptr_start\t\t%llx\n", pool->ptr_start);
+
+	seq_printf(m, "W7: ptr_end\t\t%llx\n", pool->ptr_end);
+
+	seq_printf(m, "W8: err_int\t\t%d\nW8: err_int_ena\t\t%d\n",
+		   pool->err_int, pool->err_int_ena);
+	seq_printf(m, "W8: thresh_int\t\t%d\n", pool->thresh_int);
+	seq_printf(m, "W8: thresh_int_ena\t%d\nW8: thresh_up\t\t%d\n",
+		   pool->thresh_int_ena, pool->thresh_up);
+	seq_printf(m, "W8: thresh_qint_idx\t%d\nW8: err_qint_idx\t%d\n",
+		   pool->thresh_qint_idx, pool->err_qint_idx);
+	seq_printf(m, "W8: fc_msh_dst\t\t%d\n", pool->fc_msh_dst);
+}
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/cn20k/debugfs.h b/drivers/net/ethernet/marvell/octeontx2/af/cn20k/debugfs.h
index 9d3a98dc3000..a2e3a2cd6edb 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/cn20k/debugfs.h
+++ b/drivers/net/ethernet/marvell/octeontx2/af/cn20k/debugfs.h
@@ -20,5 +20,9 @@ void print_nix_cn20k_sq_ctx(struct seq_file *m,
 			    struct nix_cn20k_sq_ctx_s *sq_ctx);
 void print_nix_cn20k_cq_ctx(struct seq_file *m,
 			    struct nix_cn20k_aq_enq_rsp *rsp);
+void print_npa_cn20k_aura_ctx(struct seq_file *m,
+			      struct npa_cn20k_aq_enq_rsp *rsp);
+void print_npa_cn20k_pool_ctx(struct seq_file *m,
+			      struct npa_cn20k_aq_enq_rsp *rsp);
 
 #endif
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c b/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c
index 3d7a4f923c04..296012a2f3de 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c
@@ -1038,6 +1038,11 @@ static void print_npa_aura_ctx(struct seq_file *m, struct npa_aq_enq_rsp *rsp)
 	struct npa_aura_s *aura = &rsp->aura;
 	struct rvu *rvu = m->private;
 
+	if (is_cn20k(rvu->pdev)) {
+		print_npa_cn20k_aura_ctx(m, (struct npa_cn20k_aq_enq_rsp *)rsp);
+		return;
+	}
+
 	seq_printf(m, "W0: Pool addr\t\t%llx\n", aura->pool_addr);
 
 	seq_printf(m, "W1: ena\t\t\t%d\nW1: pool caching\t%d\n",
@@ -1086,6 +1091,11 @@ static void print_npa_pool_ctx(struct seq_file *m, struct npa_aq_enq_rsp *rsp)
 	struct npa_pool_s *pool = &rsp->pool;
 	struct rvu *rvu = m->private;
 
+	if (is_cn20k(rvu->pdev)) {
+		print_npa_cn20k_pool_ctx(m, (struct npa_cn20k_aq_enq_rsp *)rsp);
+		return;
+	}
+
 	seq_printf(m, "W0: Stack base\t\t%llx\n", pool->stack_base);
 
 	seq_printf(m, "W1: ena \t\t%d\nW1: nat_align \t\t%d\n",
-- 
2.34.1

[net-next PATCH v2 06/11] octeontx2-af: Skip NDC operations for cn20k

[Subbaraya Sundeep]

From: Linu Cherian <lcherian@marvell.com>

For cn20k, NPA block doesn't use the general purpose
NDC (Near Coprocessor Bus Data cache Unit) for caching,
hence skip the NDC related operations.
Also refactor NDC configuration code to a helper function.

Signed-off-by: Linu Cherian <lcherian@marvell.com>
Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com>
---
 .../marvell/octeontx2/af/rvu_debugfs.c        |  3 ++
 .../ethernet/marvell/octeontx2/af/rvu_npa.c   | 29 ++++++++++++++-----
 2 files changed, 24 insertions(+), 8 deletions(-)

diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c b/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c
index 296012a2f3de..7243592f0e43 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c
@@ -2743,6 +2743,9 @@ static void rvu_dbg_npa_init(struct rvu *rvu)
 			    &rvu_dbg_npa_aura_ctx_fops);
 	debugfs_create_file("pool_ctx", 0600, rvu->rvu_dbg.npa, rvu,
 			    &rvu_dbg_npa_pool_ctx_fops);
+
+	if (is_cn20k(rvu->pdev)) /* NDC not appliable for cn20k */
+		return;
 	debugfs_create_file("ndc_cache", 0600, rvu->rvu_dbg.npa, rvu,
 			    &rvu_dbg_npa_ndc_cache_fops);
 	debugfs_create_file("ndc_hits_miss", 0600, rvu->rvu_dbg.npa, rvu,
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_npa.c b/drivers/net/ethernet/marvell/octeontx2/af/rvu_npa.c
index 4f5ca5ab13a4..e2a33e46b48a 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_npa.c
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_npa.c
@@ -464,6 +464,23 @@ int rvu_mbox_handler_npa_lf_free(struct rvu *rvu, struct msg_req *req,
 	return 0;
 }
 
+static void npa_aq_ndc_config(struct rvu *rvu, struct rvu_block *block)
+{
+	u64 cfg;
+
+	if (is_cn20k(rvu->pdev)) /* NDC not applicable to cn20k */
+		return;
+
+	/* Do not bypass NDC cache */
+	cfg = rvu_read64(rvu, block->addr, NPA_AF_NDC_CFG);
+	cfg &= ~0x03DULL;
+#ifdef CONFIG_NDC_DIS_DYNAMIC_CACHING
+	/* Disable caching of stack pages */
+	cfg |= 0x10ULL;
+#endif
+	rvu_write64(rvu, block->addr, NPA_AF_NDC_CFG, cfg);
+}
+
 static int npa_aq_init(struct rvu *rvu, struct rvu_block *block)
 {
 	u64 cfg;
@@ -479,14 +496,7 @@ static int npa_aq_init(struct rvu *rvu, struct rvu_block *block)
 	rvu_write64(rvu, block->addr, NPA_AF_GEN_CFG, cfg);
 #endif
 
-	/* Do not bypass NDC cache */
-	cfg = rvu_read64(rvu, block->addr, NPA_AF_NDC_CFG);
-	cfg &= ~0x03DULL;
-#ifdef CONFIG_NDC_DIS_DYNAMIC_CACHING
-	/* Disable caching of stack pages */
-	cfg |= 0x10ULL;
-#endif
-	rvu_write64(rvu, block->addr, NPA_AF_NDC_CFG, cfg);
+	npa_aq_ndc_config(rvu, block);
 
 	/* For CN10K NPA BATCH DMA set 35 cache lines */
 	if (!is_rvu_otx2(rvu)) {
@@ -567,6 +577,9 @@ int rvu_ndc_fix_locked_cacheline(struct rvu *rvu, int blkaddr)
 	int bank, max_bank, line, max_line, err;
 	u64 reg, ndc_af_const;
 
+	if (is_cn20k(rvu->pdev)) /* NDC not applicable to cn20k */
+		return 0;
+
 	/* Set the ENABLE bit(63) to '0' */
 	reg = rvu_read64(rvu, blkaddr, NDC_AF_CAMS_RD_INTERVAL);
 	rvu_write64(rvu, blkaddr, NDC_AF_CAMS_RD_INTERVAL, reg & GENMASK_ULL(62, 0));
-- 
2.34.1

[net-next PATCH v2 07/11] octeontx2-pf: Initialize cn20k specific aura and pool contexts

[Subbaraya Sundeep]

From: Linu Cherian <lcherian@marvell.com>

With new CN20K NPA pool and aura contexts supported in AF
driver this patch modifies PF driver to use new NPA contexts.
Implement new hw_ops for intializing aura and pool contexts
for all the silicons.

Signed-off-by: Linu Cherian <lcherian@marvell.com>
Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com>
---
 .../ethernet/marvell/octeontx2/nic/cn10k.c    |   4 +
 .../ethernet/marvell/octeontx2/nic/cn20k.c    | 178 ++++++++++++++++--
 .../marvell/octeontx2/nic/otx2_common.c       |  14 ++
 .../marvell/octeontx2/nic/otx2_common.h       |  10 +
 4 files changed, 195 insertions(+), 11 deletions(-)

diff --git a/drivers/net/ethernet/marvell/octeontx2/nic/cn10k.c b/drivers/net/ethernet/marvell/octeontx2/nic/cn10k.c
index bec7d5b4d7cc..cab157aac251 100644
--- a/drivers/net/ethernet/marvell/octeontx2/nic/cn10k.c
+++ b/drivers/net/ethernet/marvell/octeontx2/nic/cn10k.c
@@ -15,6 +15,8 @@ static struct dev_hw_ops	otx2_hw_ops = {
 	.aura_freeptr = otx2_aura_freeptr,
 	.refill_pool_ptrs = otx2_refill_pool_ptrs,
 	.pfaf_mbox_intr_handler = otx2_pfaf_mbox_intr_handler,
+	.aura_aq_init = otx2_aura_aq_init,
+	.pool_aq_init = otx2_pool_aq_init,
 };
 
 static struct dev_hw_ops cn10k_hw_ops = {
@@ -23,6 +25,8 @@ static struct dev_hw_ops cn10k_hw_ops = {
 	.aura_freeptr = cn10k_aura_freeptr,
 	.refill_pool_ptrs = cn10k_refill_pool_ptrs,
 	.pfaf_mbox_intr_handler = otx2_pfaf_mbox_intr_handler,
+	.aura_aq_init = otx2_aura_aq_init,
+	.pool_aq_init = otx2_pool_aq_init,
 };
 
 void otx2_init_hw_ops(struct otx2_nic *pfvf)
diff --git a/drivers/net/ethernet/marvell/octeontx2/nic/cn20k.c b/drivers/net/ethernet/marvell/octeontx2/nic/cn20k.c
index ec8cde98076d..037548f36940 100644
--- a/drivers/net/ethernet/marvell/octeontx2/nic/cn20k.c
+++ b/drivers/net/ethernet/marvell/octeontx2/nic/cn20k.c
@@ -10,17 +10,6 @@
 #include "otx2_struct.h"
 #include "cn10k.h"
 
-static struct dev_hw_ops cn20k_hw_ops = {
-	.pfaf_mbox_intr_handler = cn20k_pfaf_mbox_intr_handler,
-	.vfaf_mbox_intr_handler = cn20k_vfaf_mbox_intr_handler,
-	.pfvf_mbox_intr_handler = cn20k_pfvf_mbox_intr_handler,
-};
-
-void cn20k_init(struct otx2_nic *pfvf)
-{
-	pfvf->hw_ops = &cn20k_hw_ops;
-}
-EXPORT_SYMBOL(cn20k_init);
 /* CN20K mbox AF => PFx irq handler */
 irqreturn_t cn20k_pfaf_mbox_intr_handler(int irq, void *pf_irq)
 {
@@ -250,3 +239,170 @@ int cn20k_register_pfvf_mbox_intr(struct otx2_nic *pf, int numvfs)
 
 	return 0;
 }
+
+#define RQ_BP_LVL_AURA   (255 - ((85 * 256) / 100)) /* BP when 85% is full */
+
+static int cn20k_aura_aq_init(struct otx2_nic *pfvf, int aura_id,
+			      int pool_id, int numptrs)
+{
+	struct npa_cn20k_aq_enq_req *aq;
+	struct otx2_pool *pool;
+	int err;
+
+	pool = &pfvf->qset.pool[pool_id];
+
+	/* Allocate memory for HW to update Aura count.
+	 * Alloc one cache line, so that it fits all FC_STYPE modes.
+	 */
+	if (!pool->fc_addr) {
+		err = qmem_alloc(pfvf->dev, &pool->fc_addr, 1, OTX2_ALIGN);
+		if (err)
+			return err;
+	}
+
+	/* Initialize this aura's context via AF */
+	aq = otx2_mbox_alloc_msg_npa_cn20k_aq_enq(&pfvf->mbox);
+	if (!aq) {
+		/* Shared mbox memory buffer is full, flush it and retry */
+		err = otx2_sync_mbox_msg(&pfvf->mbox);
+		if (err)
+			return err;
+		aq = otx2_mbox_alloc_msg_npa_cn20k_aq_enq(&pfvf->mbox);
+		if (!aq)
+			return -ENOMEM;
+	}
+
+	aq->aura_id = aura_id;
+
+	/* Will be filled by AF with correct pool context address */
+	aq->aura.pool_addr = pool_id;
+	aq->aura.pool_caching = 1;
+	aq->aura.shift = ilog2(numptrs) - 8;
+	aq->aura.count = numptrs;
+	aq->aura.limit = numptrs;
+	aq->aura.avg_level = 255;
+	aq->aura.ena = 1;
+	aq->aura.fc_ena = 1;
+	aq->aura.fc_addr = pool->fc_addr->iova;
+	aq->aura.fc_hyst_bits = 0; /* Store count on all updates */
+
+	/* Enable backpressure for RQ aura */
+	if (aura_id < pfvf->hw.rqpool_cnt && !is_otx2_lbkvf(pfvf->pdev)) {
+		aq->aura.bp_ena = 0;
+		/* If NIX1 LF is attached then specify NIX1_RX.
+		 *
+		 * Below NPA_AURA_S[BP_ENA] is set according to the
+		 * NPA_BPINTF_E enumeration given as:
+		 * 0x0 + a*0x1 where 'a' is 0 for NIX0_RX and 1 for NIX1_RX so
+		 * NIX0_RX is 0x0 + 0*0x1 = 0
+		 * NIX1_RX is 0x0 + 1*0x1 = 1
+		 * But in HRM it is given that
+		 * "NPA_AURA_S[BP_ENA](w1[33:32]) - Enable aura backpressure to
+		 * NIX-RX based on [BP] level. One bit per NIX-RX; index
+		 * enumerated by NPA_BPINTF_E."
+		 */
+		if (pfvf->nix_blkaddr == BLKADDR_NIX1)
+			aq->aura.bp_ena = 1;
+#ifdef CONFIG_DCB
+		aq->aura.bpid = pfvf->bpid[pfvf->queue_to_pfc_map[aura_id]];
+#else
+		aq->aura.bpid = pfvf->bpid[0];
+#endif
+
+		/* Set backpressure level for RQ's Aura */
+		aq->aura.bp = RQ_BP_LVL_AURA;
+	}
+
+	/* Fill AQ info */
+	aq->ctype = NPA_AQ_CTYPE_AURA;
+	aq->op = NPA_AQ_INSTOP_INIT;
+
+	return 0;
+}
+
+static int cn20k_pool_aq_init(struct otx2_nic *pfvf, u16 pool_id,
+			      int stack_pages, int numptrs, int buf_size,
+			      int type)
+{
+	struct page_pool_params pp_params = { 0 };
+	struct npa_cn20k_aq_enq_req *aq;
+	struct otx2_pool *pool;
+	int err, sz;
+
+	pool = &pfvf->qset.pool[pool_id];
+	/* Alloc memory for stack which is used to store buffer pointers */
+	err = qmem_alloc(pfvf->dev, &pool->stack,
+			 stack_pages, pfvf->hw.stack_pg_bytes);
+	if (err)
+		return err;
+
+	pool->rbsize = buf_size;
+
+	/* Initialize this pool's context via AF */
+	aq = otx2_mbox_alloc_msg_npa_cn20k_aq_enq(&pfvf->mbox);
+	if (!aq) {
+		/* Shared mbox memory buffer is full, flush it and retry */
+		err = otx2_sync_mbox_msg(&pfvf->mbox);
+		if (err) {
+			qmem_free(pfvf->dev, pool->stack);
+			return err;
+		}
+		aq = otx2_mbox_alloc_msg_npa_cn20k_aq_enq(&pfvf->mbox);
+		if (!aq) {
+			qmem_free(pfvf->dev, pool->stack);
+			return -ENOMEM;
+		}
+	}
+
+	aq->aura_id = pool_id;
+	aq->pool.stack_base = pool->stack->iova;
+	aq->pool.stack_caching = 1;
+	aq->pool.ena = 1;
+	aq->pool.buf_size = buf_size / 128;
+	aq->pool.stack_max_pages = stack_pages;
+	aq->pool.shift = ilog2(numptrs) - 8;
+	aq->pool.ptr_start = 0;
+	aq->pool.ptr_end = ~0ULL;
+
+	/* Fill AQ info */
+	aq->ctype = NPA_AQ_CTYPE_POOL;
+	aq->op = NPA_AQ_INSTOP_INIT;
+
+	if (type != AURA_NIX_RQ) {
+		pool->page_pool = NULL;
+		return 0;
+	}
+
+	sz = ALIGN(ALIGN(SKB_DATA_ALIGN(buf_size), OTX2_ALIGN), PAGE_SIZE);
+	pp_params.order = (sz / PAGE_SIZE) - 1;
+	pp_params.flags = PP_FLAG_DMA_MAP;
+	pp_params.pool_size = min(OTX2_PAGE_POOL_SZ, numptrs);
+	pp_params.nid = NUMA_NO_NODE;
+	pp_params.dev = pfvf->dev;
+	pp_params.dma_dir = DMA_FROM_DEVICE;
+	pool->page_pool = page_pool_create(&pp_params);
+	if (IS_ERR(pool->page_pool)) {
+		netdev_err(pfvf->netdev, "Creation of page pool failed\n");
+		return PTR_ERR(pool->page_pool);
+	}
+
+	return 0;
+}
+
+static struct dev_hw_ops cn20k_hw_ops = {
+	.pfaf_mbox_intr_handler = cn20k_pfaf_mbox_intr_handler,
+	.vfaf_mbox_intr_handler = cn20k_vfaf_mbox_intr_handler,
+	.pfvf_mbox_intr_handler = cn20k_pfvf_mbox_intr_handler,
+	.sq_aq_init = cn10k_sq_aq_init,
+	.sqe_flush = cn10k_sqe_flush,
+	.aura_freeptr = cn10k_aura_freeptr,
+	.refill_pool_ptrs = cn10k_refill_pool_ptrs,
+	.aura_aq_init = cn20k_aura_aq_init,
+	.pool_aq_init = cn20k_pool_aq_init,
+};
+
+void cn20k_init(struct otx2_nic *pfvf)
+{
+	pfvf->hw_ops = &cn20k_hw_ops;
+}
+EXPORT_SYMBOL(cn20k_init);
diff --git a/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.c b/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.c
index f674729124e6..ad57ded39bd1 100644
--- a/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.c
+++ b/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.c
@@ -1366,6 +1366,13 @@ void otx2_aura_pool_free(struct otx2_nic *pfvf)
 
 int otx2_aura_init(struct otx2_nic *pfvf, int aura_id,
 		   int pool_id, int numptrs)
+{
+	return pfvf->hw_ops->aura_aq_init(pfvf, aura_id, pool_id,
+					  numptrs);
+}
+
+int otx2_aura_aq_init(struct otx2_nic *pfvf, int aura_id,
+		      int pool_id, int numptrs)
 {
 	struct npa_aq_enq_req *aq;
 	struct otx2_pool *pool;
@@ -1443,6 +1450,13 @@ int otx2_aura_init(struct otx2_nic *pfvf, int aura_id,
 
 int otx2_pool_init(struct otx2_nic *pfvf, u16 pool_id,
 		   int stack_pages, int numptrs, int buf_size, int type)
+{
+	return pfvf->hw_ops->pool_aq_init(pfvf, pool_id, stack_pages, numptrs,
+					  buf_size, type);
+}
+
+int otx2_pool_aq_init(struct otx2_nic *pfvf, u16 pool_id,
+		      int stack_pages, int numptrs, int buf_size, int type)
 {
 	struct page_pool_params pp_params = { 0 };
 	struct xsk_buff_pool *xsk_pool;
diff --git a/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.h b/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.h
index e3765b73c434..b587ff931820 100644
--- a/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.h
+++ b/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.h
@@ -14,6 +14,7 @@
 #include <linux/net_tstamp.h>
 #include <linux/ptp_clock_kernel.h>
 #include <linux/timecounter.h>
+#include <linux/soc/marvell/silicons.h>
 #include <linux/soc/marvell/octeontx2/asm.h>
 #include <net/macsec.h>
 #include <net/pkt_cls.h>
@@ -374,6 +375,11 @@ struct dev_hw_ops {
 	irqreturn_t (*pfaf_mbox_intr_handler)(int irq, void *pf_irq);
 	irqreturn_t (*vfaf_mbox_intr_handler)(int irq, void *pf_irq);
 	irqreturn_t (*pfvf_mbox_intr_handler)(int irq, void *pf_irq);
+	int	(*aura_aq_init)(struct otx2_nic *pfvf, int aura_id,
+				int pool_id, int numptrs);
+	int	(*pool_aq_init)(struct otx2_nic *pfvf, u16 pool_id,
+				int stack_pages, int numptrs, int buf_size,
+				int type);
 };
 
 #define CN10K_MCS_SA_PER_SC	4
@@ -1058,6 +1064,10 @@ irqreturn_t otx2_cq_intr_handler(int irq, void *cq_irq);
 int otx2_rq_init(struct otx2_nic *pfvf, u16 qidx, u16 lpb_aura);
 int otx2_cq_init(struct otx2_nic *pfvf, u16 qidx);
 int otx2_set_hw_capabilities(struct otx2_nic *pfvf);
+int otx2_aura_aq_init(struct otx2_nic *pfvf, int aura_id,
+		      int pool_id, int numptrs);
+int otx2_pool_aq_init(struct otx2_nic *pfvf, u16 pool_id,
+		      int stack_pages, int numptrs, int buf_size, int type);
 
 /* RSS configuration APIs*/
 int otx2_rss_init(struct otx2_nic *pfvf);
-- 
2.34.1

[net-next PATCH v2 08/11] octeontx2-pf: Initialize new NIX SQ context for cn20k

[Subbaraya Sundeep]

cn20k has different NIX context for send queue hence use
the new cn20k mailbox to init SQ context.

Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com>
---
 .../ethernet/marvell/octeontx2/nic/cn20k.c    | 36 ++++++++++++++++++-
 1 file changed, 35 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/marvell/octeontx2/nic/cn20k.c b/drivers/net/ethernet/marvell/octeontx2/nic/cn20k.c
index 037548f36940..4f0afa5301b4 100644
--- a/drivers/net/ethernet/marvell/octeontx2/nic/cn20k.c
+++ b/drivers/net/ethernet/marvell/octeontx2/nic/cn20k.c
@@ -389,11 +389,45 @@ static int cn20k_pool_aq_init(struct otx2_nic *pfvf, u16 pool_id,
 	return 0;
 }
 
+static int cn20k_sq_aq_init(void *dev, u16 qidx, u8 chan_offset, u16 sqb_aura)
+{
+	struct nix_cn20k_aq_enq_req *aq;
+	struct otx2_nic *pfvf = dev;
+
+	/* Get memory to put this msg */
+	aq = otx2_mbox_alloc_msg_nix_cn20k_aq_enq(&pfvf->mbox);
+	if (!aq)
+		return -ENOMEM;
+
+	aq->sq.cq = pfvf->hw.rx_queues + qidx;
+	aq->sq.max_sqe_size = NIX_MAXSQESZ_W16; /* 128 byte */
+	aq->sq.cq_ena = 1;
+	aq->sq.ena = 1;
+	aq->sq.smq = otx2_get_smq_idx(pfvf, qidx);
+	aq->sq.smq_rr_weight = mtu_to_dwrr_weight(pfvf, pfvf->tx_max_pktlen);
+	aq->sq.default_chan = pfvf->hw.tx_chan_base + chan_offset;
+	aq->sq.sqe_stype = NIX_STYPE_STF; /* Cache SQB */
+	aq->sq.sqb_aura = sqb_aura;
+	aq->sq.sq_int_ena = NIX_SQINT_BITS;
+	aq->sq.qint_idx = 0;
+	/* Due pipelining impact minimum 2000 unused SQ CQE's
+	 * need to maintain to avoid CQ overflow.
+	 */
+	aq->sq.cq_limit = ((SEND_CQ_SKID * 256) / (pfvf->qset.sqe_cnt));
+
+	/* Fill AQ info */
+	aq->qidx = qidx;
+	aq->ctype = NIX_AQ_CTYPE_SQ;
+	aq->op = NIX_AQ_INSTOP_INIT;
+
+	return otx2_sync_mbox_msg(&pfvf->mbox);
+}
+
 static struct dev_hw_ops cn20k_hw_ops = {
 	.pfaf_mbox_intr_handler = cn20k_pfaf_mbox_intr_handler,
 	.vfaf_mbox_intr_handler = cn20k_vfaf_mbox_intr_handler,
 	.pfvf_mbox_intr_handler = cn20k_pfvf_mbox_intr_handler,
-	.sq_aq_init = cn10k_sq_aq_init,
+	.sq_aq_init = cn20k_sq_aq_init,
 	.sqe_flush = cn10k_sqe_flush,
 	.aura_freeptr = cn10k_aura_freeptr,
 	.refill_pool_ptrs = cn10k_refill_pool_ptrs,
-- 
2.34.1

[net-next PATCH v2 09/11] octeontx2-af: Accommodate more bandwidth profiles for cn20k

[Subbaraya Sundeep]

CN20K has 16k of leaf profiles, 2k of middle profiles and
256 of top profiles. This patch modifies existing receive
queue and bandwidth profile context structures to accommodate
additional profiles of cn20k.

Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com>
---
 drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c | 13 ++++++++-----
 .../net/ethernet/marvell/octeontx2/af/rvu_struct.h  |  6 ++++--
 2 files changed, 12 insertions(+), 7 deletions(-)

diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c b/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c
index 47a64c4c5b7e..ab3b8b6a9e43 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c
@@ -5858,7 +5858,7 @@ static int nix_verify_bandprof(struct nix_cn10k_aq_enq_req *req,
 		return -EINVAL;
 
 	ipolicer = &nix_hw->ipolicer[hi_layer];
-	prof_idx = req->prof.band_prof_id;
+	prof_idx =  req->prof.band_prof_id_h << 7 | req->prof.band_prof_id;
 	if (prof_idx >= ipolicer->band_prof.max ||
 	    ipolicer->pfvf_map[prof_idx] != pcifunc)
 		return -EINVAL;
@@ -6023,8 +6023,10 @@ static int nix_ipolicer_map_leaf_midprofs(struct rvu *rvu,
 	aq_req->op = NIX_AQ_INSTOP_WRITE;
 	aq_req->qidx = leaf_prof;
 
-	aq_req->prof.band_prof_id = mid_prof;
+	aq_req->prof.band_prof_id = mid_prof & 0x7F;
 	aq_req->prof_mask.band_prof_id = GENMASK(6, 0);
+	aq_req->prof.band_prof_id_h = mid_prof >> 7;
+	aq_req->prof_mask.band_prof_id_h = GENMASK(3, 0);
 	aq_req->prof.hl_en = 1;
 	aq_req->prof_mask.hl_en = 1;
 
@@ -6064,7 +6066,7 @@ int rvu_nix_setup_ratelimit_aggr(struct rvu *rvu, u16 pcifunc,
 		return 0;
 
 	/* Get the bandwidth profile ID mapped to this RQ */
-	leaf_prof = aq_rsp.rq.band_prof_id;
+	leaf_prof = aq_rsp.rq.band_prof_id_h << 10 | aq_rsp.rq.band_prof_id;
 
 	ipolicer = &nix_hw->ipolicer[BAND_PROF_LEAF_LAYER];
 	ipolicer->match_id[leaf_prof] = match_id;
@@ -6102,7 +6104,8 @@ int rvu_nix_setup_ratelimit_aggr(struct rvu *rvu, u16 pcifunc,
 		 * to different RQs and marked with same match_id
 		 * are rate limited in a aggregate fashion
 		 */
-		mid_prof = aq_rsp.prof.band_prof_id;
+		mid_prof = aq_rsp.prof.band_prof_id_h << 7 |
+			   aq_rsp.prof.band_prof_id;
 		rc = nix_ipolicer_map_leaf_midprofs(rvu, nix_hw,
 						    &aq_req, &aq_rsp,
 						    leaf_prof, mid_prof);
@@ -6224,7 +6227,7 @@ static void nix_clear_ratelimit_aggr(struct rvu *rvu, struct nix_hw *nix_hw,
 	if (!aq_rsp.prof.hl_en)
 		return;
 
-	mid_prof = aq_rsp.prof.band_prof_id;
+	mid_prof = aq_rsp.prof.band_prof_id_h << 7 | aq_rsp.prof.band_prof_id;
 	ipolicer = &nix_hw->ipolicer[BAND_PROF_MID_LAYER];
 	ipolicer->ref_count[mid_prof]--;
 	/* If ref_count is zero, free mid layer profile */
diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_struct.h b/drivers/net/ethernet/marvell/octeontx2/af/rvu_struct.h
index 8a66f53a7658..1060ef0caf9c 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_struct.h
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_struct.h
@@ -415,7 +415,8 @@ struct nix_cn10k_rq_ctx_s {
 	u64 rsvd_171		: 1;
 	u64 later_skip		: 6;
 	u64 xqe_imm_size	: 6;
-	u64 rsvd_189_184	: 6;
+	u64 band_prof_id_h	: 4;
+	u64 rsvd_189_188	: 2;
 	u64 xqe_imm_copy	: 1;
 	u64 xqe_hdr_split	: 1;
 	u64 xqe_drop		: 8; /* W3 */
@@ -741,7 +742,8 @@ struct nix_bandprof_s {
 	uint64_t rc_action                   :  2;
 	uint64_t meter_algo                  :  2;
 	uint64_t band_prof_id                :  7;
-	uint64_t reserved_111_118            :  8;
+	uint64_t band_prof_id_h              :  4;
+	uint64_t reserved_115_118            :  4;
 	uint64_t hl_en                       :  1;
 	uint64_t reserved_120_127            :  8;
 	uint64_t ts                          : 48; /* W2 */
-- 
2.34.1

[net-next PATCH v2 10/11] octeontx2-af: Display new bandwidth profiles too in debugfs

[Subbaraya Sundeep]

Consider the new profiles of cn20k too while displaying
bandwidth profile contexts in debugfs.

Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com>
---
 drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c b/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c
index 7243592f0e43..ed502ca0c927 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c
+++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_debugfs.c
@@ -2056,7 +2056,9 @@ static void print_nix_cn10k_rq_ctx(struct seq_file *m,
 	seq_printf(m, "W1: ipsecd_drop_ena \t\t%d\nW1: chi_ena \t\t\t%d\n\n",
 		   rq_ctx->ipsecd_drop_ena, rq_ctx->chi_ena);
 
-	seq_printf(m, "W2: band_prof_id \t\t%d\n", rq_ctx->band_prof_id);
+	seq_printf(m, "W2: band_prof_id \t\t%d\n",
+		   (u16)rq_ctx->band_prof_id_h << 10 | rq_ctx->band_prof_id);
+
 	seq_printf(m, "W2: policer_ena \t\t%d\n", rq_ctx->policer_ena);
 	seq_printf(m, "W2: spb_sizem1 \t\t\t%d\n", rq_ctx->spb_sizem1);
 	seq_printf(m, "W2: wqe_skip \t\t\t%d\nW2: sqb_ena \t\t\t%d\n",
@@ -2574,7 +2576,10 @@ static void print_band_prof_ctx(struct seq_file *m,
 		(prof->rc_action == 1) ? "DROP" : "RED";
 	seq_printf(m, "W1: rc_action\t\t%s\n", str);
 	seq_printf(m, "W1: meter_algo\t\t%d\n", prof->meter_algo);
-	seq_printf(m, "W1: band_prof_id\t%d\n", prof->band_prof_id);
+
+	seq_printf(m, "W1: band_prof_id\t%d\n",
+		   (u16)prof->band_prof_id_h << 7 | prof->band_prof_id);
+
 	seq_printf(m, "W1: hl_en\t\t%d\n", prof->hl_en);
 
 	seq_printf(m, "W2: ts\t\t\t%lld\n", (u64)prof->ts);
-- 
2.34.1

[net-next PATCH v2 11/11] octeontx2-pf: Use new bandwidth profiles in receive queue

[Subbaraya Sundeep]

Receive queue points to a bandwidth profile for rate limiting.
Since cn20k has additional bandwidth profiles use them
too while mapping receive queue to bandwidth profile.

Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com>
---
 drivers/net/ethernet/marvell/octeontx2/nic/cn10k.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/drivers/net/ethernet/marvell/octeontx2/nic/cn10k.c b/drivers/net/ethernet/marvell/octeontx2/nic/cn10k.c
index cab157aac251..3e1bf22cba69 100644
--- a/drivers/net/ethernet/marvell/octeontx2/nic/cn10k.c
+++ b/drivers/net/ethernet/marvell/octeontx2/nic/cn10k.c
@@ -341,6 +341,12 @@ int cn10k_map_unmap_rq_policer(struct otx2_nic *pfvf, int rq_idx,
 	aq->rq.band_prof_id = policer;
 	aq->rq_mask.band_prof_id = GENMASK(9, 0);
 
+	/* If policer id is greater than 1023 then it implies hardware supports
+	 * more leaf profiles. In that case use band_prof_id_h for 4 MSBs.
+	 */
+	aq->rq.band_prof_id_h = policer >> 10;
+	aq->rq_mask.band_prof_id_h = GENMASK(3, 0);
+
 	/* Fill AQ info */
 	aq->qidx = rq_idx;
 	aq->ctype = NIX_AQ_CTYPE_RQ;
-- 
2.34.1

Re: [net-next PATCH v2 01/11] octeontx2-af: Simplify context writing and reading to hardware

[Michal Swiatkowski]

On Tue, Jul 15, 2025 at 10:31:54PM +0530, Subbaraya Sundeep wrote:
> Simplify NIX context reading and writing by using hardware
> maximum context size instead of using individual sizes of
> each context type.
> 
> Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com>
> ---
>  .../ethernet/marvell/octeontx2/af/rvu_nix.c   | 46 ++++++++++---------
>  .../marvell/octeontx2/af/rvu_struct.h         |  7 ++-
>  2 files changed, 30 insertions(+), 23 deletions(-)
> 
> diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c b/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c
> index bdf4d852c15d..48d44911b663 100644
> --- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c
> +++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c
> @@ -17,6 +17,8 @@
>  #include "lmac_common.h"
>  #include "rvu_npc_hash.h"
>  
> +#define NIX_MAX_CTX_SIZE	128
> +

[...]

>  
>  	return 0;
> diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_struct.h b/drivers/net/ethernet/marvell/octeontx2/af/rvu_struct.h
> index 0596a3ac4c12..8a66f53a7658 100644
> --- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_struct.h
> +++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_struct.h
> @@ -370,6 +370,8 @@ struct nix_cq_ctx_s {
>  	u64 qsize		: 4;
>  	u64 cq_err_int		: 8;
>  	u64 cq_err_int_ena	: 8;
> +	/* Ensure all context sizes are minimum 128 bytes */
> +	u64 padding[12];
>  };
>  
>  /* CN10K NIX Receive queue context structure */
> @@ -672,7 +674,8 @@ struct nix_sq_ctx_s {
>  struct nix_rsse_s {
>  	uint32_t rq			: 20;
>  	uint32_t reserved_20_31		: 12;
> -
> +	/* Ensure all context sizes are minimum 128 bytes */
> +	u64 padding[15];
>  };
>  
>  /* NIX receive multicast/mirror entry structure */
> @@ -684,6 +687,8 @@ struct nix_rx_mce_s {
>  	uint64_t rsvd_31_24 : 8;
>  	uint64_t pf_func    : 16;
>  	uint64_t next       : 16;
> +	/* Ensure all context sizes are minimum 128 bytes */
> +	u64 padding[15];
>  };

To be sure that each used structures are correct size you can
use static assertion, sth like:
static_assert((NIC_MAX_CTX_SIZE) == sizeof(struct X))

Thanks

>  
>  enum nix_band_prof_layers {
> -- 
> 2.34.1
> 

Re: [net-next PATCH v2 01/11] octeontx2-af: Simplify context writing and reading to hardware

[Subbaraya Sundeep]

On 2025-07-16 at 09:24:07, Michal Swiatkowski (michal.swiatkowski@linux.intel.com) wrote:
> On Tue, Jul 15, 2025 at 10:31:54PM +0530, Subbaraya Sundeep wrote:
> > Simplify NIX context reading and writing by using hardware
> > maximum context size instead of using individual sizes of
> > each context type.
> > 
> > Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com>
> > ---
> >  .../ethernet/marvell/octeontx2/af/rvu_nix.c   | 46 ++++++++++---------
> >  .../marvell/octeontx2/af/rvu_struct.h         |  7 ++-
> >  2 files changed, 30 insertions(+), 23 deletions(-)
> > 
> > diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c b/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c
> > index bdf4d852c15d..48d44911b663 100644
> > --- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c
> > +++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c
> > @@ -17,6 +17,8 @@
> >  #include "lmac_common.h"
> >  #include "rvu_npc_hash.h"
> >  
> > +#define NIX_MAX_CTX_SIZE	128
> > +
> 
> [...]
> 
> >  
> >  	return 0;
> > diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_struct.h b/drivers/net/ethernet/marvell/octeontx2/af/rvu_struct.h
> > index 0596a3ac4c12..8a66f53a7658 100644
> > --- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_struct.h
> > +++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_struct.h
> > @@ -370,6 +370,8 @@ struct nix_cq_ctx_s {
> >  	u64 qsize		: 4;
> >  	u64 cq_err_int		: 8;
> >  	u64 cq_err_int_ena	: 8;
> > +	/* Ensure all context sizes are minimum 128 bytes */
> > +	u64 padding[12];
> >  };
> >  
> >  /* CN10K NIX Receive queue context structure */
> > @@ -672,7 +674,8 @@ struct nix_sq_ctx_s {
> >  struct nix_rsse_s {
> >  	uint32_t rq			: 20;
> >  	uint32_t reserved_20_31		: 12;
> > -
> > +	/* Ensure all context sizes are minimum 128 bytes */
> > +	u64 padding[15];
> >  };
> >  
> >  /* NIX receive multicast/mirror entry structure */
> > @@ -684,6 +687,8 @@ struct nix_rx_mce_s {
> >  	uint64_t rsvd_31_24 : 8;
> >  	uint64_t pf_func    : 16;
> >  	uint64_t next       : 16;
> > +	/* Ensure all context sizes are minimum 128 bytes */
> > +	u64 padding[15];
> >  };
> 
> To be sure that each used structures are correct size you can
> use static assertion, sth like:
> static_assert((NIC_MAX_CTX_SIZE) == sizeof(struct X))
> 
> Thanks
> 
Sure will do it.

Thanks,
Sundeep
> >  
> >  enum nix_band_prof_layers {
> > -- 
> > 2.34.1
> >