Merge tag 'mlx5-updates-2017-08-24' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5-updates-2017-08-24

This series includes updates to mlx5 core driver.

From Gal and Saeed, three cleanup patches.
From Matan, Low level flow steering improvements and optimizations,
 - Use more efficient data structures for flow steering objects handling.
 - Add tracepoints to flow steering operations.
 - Overall these patches improve flow steering rule insertion rate by a
   factor of seven in large scales (~50K rules or more).

====================
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/mellanox/mlx5/core/Makefile

@@ -4,7 +4,8 @@ subdir-ccflags-y += -I$(src)
 mlx5_core-y :=	main.o cmd.o debugfs.o fw.o eq.o uar.o pagealloc.o \
 		health.o mcg.o cq.o srq.o alloc.o qp.o port.o mr.o pd.o \
 		mad.o transobj.o vport.o sriov.o fs_cmd.o fs_core.o \
-		fs_counters.o rl.o lag.o dev.o wq.o lib/gid.o
+		fs_counters.o rl.o lag.o dev.o wq.o lib/gid.o \
+		diag/fs_tracepoint.o
 
 mlx5_core-$(CONFIG_MLX5_ACCEL) += accel/ipsec.o
 
@@ -25,3 +26,5 @@ mlx5_core-$(CONFIG_MLX5_CORE_IPOIB) += ipoib/ipoib.o ipoib/ethtool.o
 
 mlx5_core-$(CONFIG_MLX5_EN_IPSEC) += en_accel/ipsec.o en_accel/ipsec_rxtx.o \
 		en_accel/ipsec_stats.o
+
+CFLAGS_tracepoint.o := -I$(src)

drivers/net/ethernet/mellanox/mlx5/core/diag/Makefile

+subdir-ccflags-y += -I$(src)/..

drivers/net/ethernet/mellanox/mlx5/core/diag/fs_tracepoint.c

+/*
+ * Copyright (c) 2017, Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#define CREATE_TRACE_POINTS
+
+#include "fs_tracepoint.h"
+#include <linux/stringify.h>
+
+#define DECLARE_MASK_VAL(type, name) struct {type m; type v; } name
+#define MASK_VAL(type, spec, name, mask, val, fld)	\
+		DECLARE_MASK_VAL(type, name) =		\
+			{.m = MLX5_GET(spec, mask, fld),\
+			 .v = MLX5_GET(spec, val, fld)}
+#define MASK_VAL_BE(type, spec, name, mask, val, fld)	\
+		    DECLARE_MASK_VAL(type, name) =	\
+			{.m = MLX5_GET_BE(type, spec, mask, fld),\
+			 .v = MLX5_GET_BE(type, spec, val, fld)}
+#define GET_MASKED_VAL(name) (name.m & name.v)
+
+#define GET_MASK_VAL(name, type, mask, val, fld)	\
+		(name.m = MLX5_GET(type, mask, fld),	\
+		 name.v = MLX5_GET(type, val, fld),	\
+		 name.m & name.v)
+#define PRINT_MASKED_VAL(name, p, format) {		\
+	if (name.m)			\
+		trace_seq_printf(p, __stringify(name) "=" format " ", name.v); \
+	}
+#define PRINT_MASKED_VALP(name, cast, p, format) {	\
+	if (name.m)			\
+		trace_seq_printf(p, __stringify(name) "=" format " ",	       \
+				 (cast)&name.v);\
+	}
+
+static void print_lyr_2_4_hdrs(struct trace_seq *p,
+			       const u32 *mask, const u32 *value)
+{
+#define MASK_VAL_L2(type, name, fld) \
+	MASK_VAL(type, fte_match_set_lyr_2_4, name, mask, value, fld)
+	DECLARE_MASK_VAL(u64, smac) = {
+		.m = MLX5_GET(fte_match_set_lyr_2_4, mask, smac_47_16) << 16 |
+		     MLX5_GET(fte_match_set_lyr_2_4, mask, smac_15_0),
+		.v = MLX5_GET(fte_match_set_lyr_2_4, value, smac_47_16) << 16 |
+		     MLX5_GET(fte_match_set_lyr_2_4, value, smac_15_0)};
+	DECLARE_MASK_VAL(u64, dmac) = {
+		.m = MLX5_GET(fte_match_set_lyr_2_4, mask, dmac_47_16) << 16 |
+		     MLX5_GET(fte_match_set_lyr_2_4, mask, dmac_15_0),
+		.v = MLX5_GET(fte_match_set_lyr_2_4, value, dmac_47_16) << 16 |
+		     MLX5_GET(fte_match_set_lyr_2_4, value, dmac_15_0)};
+	MASK_VAL_L2(u16, ethertype, ethertype);
+
+	PRINT_MASKED_VALP(smac, u8 *, p, "%pM");
+	PRINT_MASKED_VALP(dmac, u8 *, p, "%pM");
+	PRINT_MASKED_VAL(ethertype, p, "%04x");
+
+	if (ethertype.m == 0xffff) {
+		if (ethertype.v == ETH_P_IP) {
+#define MASK_VAL_L2_BE(type, name, fld) \
+	MASK_VAL_BE(type, fte_match_set_lyr_2_4, name, mask, value, fld)
+			MASK_VAL_L2_BE(u32, src_ipv4,
+				       src_ipv4_src_ipv6.ipv4_layout.ipv4);
+			MASK_VAL_L2_BE(u32, dst_ipv4,
+				       dst_ipv4_dst_ipv6.ipv4_layout.ipv4);
+
+			PRINT_MASKED_VALP(src_ipv4, typeof(&src_ipv4.v), p,
+					  "%pI4");
+			PRINT_MASKED_VALP(dst_ipv4, typeof(&dst_ipv4.v), p,
+					  "%pI4");
+		} else if (ethertype.v == ETH_P_IPV6) {
+			static const struct in6_addr full_ones = {
+				.in6_u.u6_addr32 = {htonl(0xffffffff),
+						    htonl(0xffffffff),
+						    htonl(0xffffffff),
+						    htonl(0xffffffff)},
+			};
+			DECLARE_MASK_VAL(struct in6_addr, src_ipv6);
+			DECLARE_MASK_VAL(struct in6_addr, dst_ipv6);
+
+			memcpy(src_ipv6.m.in6_u.u6_addr8,
+			       MLX5_ADDR_OF(fte_match_set_lyr_2_4, mask,
+					    src_ipv4_src_ipv6.ipv6_layout.ipv6),
+			       sizeof(src_ipv6.m));
+			memcpy(dst_ipv6.m.in6_u.u6_addr8,
+			       MLX5_ADDR_OF(fte_match_set_lyr_2_4, mask,
+					    dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+			       sizeof(dst_ipv6.m));
+			memcpy(src_ipv6.v.in6_u.u6_addr8,
+			       MLX5_ADDR_OF(fte_match_set_lyr_2_4, value,
+					    src_ipv4_src_ipv6.ipv6_layout.ipv6),
+			       sizeof(src_ipv6.v));
+			memcpy(dst_ipv6.v.in6_u.u6_addr8,
+			       MLX5_ADDR_OF(fte_match_set_lyr_2_4, value,
+					    dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+			       sizeof(dst_ipv6.v));
+
+			if (!memcmp(&src_ipv6.m, &full_ones, sizeof(full_ones)))
+				trace_seq_printf(p, "src_ipv6=%pI6 ",
+						 &src_ipv6.v);
+			if (!memcmp(&dst_ipv6.m, &full_ones, sizeof(full_ones)))
+				trace_seq_printf(p, "dst_ipv6=%pI6 ",
+						 &dst_ipv6.v);
+		}
+	}
+
+#define PRINT_MASKED_VAL_L2(type, name, fld, p, format) {\
+	MASK_VAL_L2(type, name, fld);		         \
+	PRINT_MASKED_VAL(name, p, format);		 \
+}
+
+	PRINT_MASKED_VAL_L2(u8, ip_protocol, ip_protocol, p, "%02x");
+	PRINT_MASKED_VAL_L2(u16, tcp_flags, tcp_flags, p, "%x");
+	PRINT_MASKED_VAL_L2(u16, tcp_sport, tcp_sport, p, "%u");
+	PRINT_MASKED_VAL_L2(u16, tcp_dport, tcp_dport, p, "%u");
+	PRINT_MASKED_VAL_L2(u16, udp_sport, udp_sport, p, "%u");
+	PRINT_MASKED_VAL_L2(u16, udp_dport, udp_dport, p, "%u");
+	PRINT_MASKED_VAL_L2(u16, first_vid, first_vid, p, "%04x");
+	PRINT_MASKED_VAL_L2(u8, first_prio, first_prio, p, "%x");
+	PRINT_MASKED_VAL_L2(u8, first_cfi, first_cfi, p, "%d");
+	PRINT_MASKED_VAL_L2(u8, ip_dscp, ip_dscp, p, "%02x");
+	PRINT_MASKED_VAL_L2(u8, ip_ecn, ip_ecn, p, "%x");
+	PRINT_MASKED_VAL_L2(u8, cvlan_tag, cvlan_tag, p, "%d");
+	PRINT_MASKED_VAL_L2(u8, svlan_tag, svlan_tag, p, "%d");
+	PRINT_MASKED_VAL_L2(u8, frag, frag, p, "%d");
+}
+
+static void print_misc_parameters_hdrs(struct trace_seq *p,
+				       const u32 *mask, const u32 *value)
+{
+#define MASK_VAL_MISC(type, name, fld) \
+	MASK_VAL(type, fte_match_set_misc, name, mask, value, fld)
+#define PRINT_MASKED_VAL_MISC(type, name, fld, p, format) {\
+	MASK_VAL_MISC(type, name, fld);			   \
+	PRINT_MASKED_VAL(name, p, format);		   \
+}
+	DECLARE_MASK_VAL(u64, gre_key) = {
+		.m = MLX5_GET(fte_match_set_misc, mask, gre_key_h) << 8 |
+		     MLX5_GET(fte_match_set_misc, mask, gre_key_l),
+		.v = MLX5_GET(fte_match_set_misc, value, gre_key_h) << 8 |
+		     MLX5_GET(fte_match_set_misc, value, gre_key_l)};
+
+	PRINT_MASKED_VAL(gre_key, p, "%llu");
+	PRINT_MASKED_VAL_MISC(u32, source_sqn, source_sqn, p, "%u");
+	PRINT_MASKED_VAL_MISC(u16, source_port, source_port, p, "%u");
+	PRINT_MASKED_VAL_MISC(u8, outer_second_prio, outer_second_prio,
+			      p, "%u");
+	PRINT_MASKED_VAL_MISC(u8, outer_second_cfi, outer_second_cfi, p, "%u");
+	PRINT_MASKED_VAL_MISC(u16, outer_second_vid, outer_second_vid, p, "%u");
+	PRINT_MASKED_VAL_MISC(u8, inner_second_prio, inner_second_prio,
+			      p, "%u");
+	PRINT_MASKED_VAL_MISC(u8, inner_second_cfi, inner_second_cfi, p, "%u");
+	PRINT_MASKED_VAL_MISC(u16, inner_second_vid, inner_second_vid, p, "%u");
+
+	PRINT_MASKED_VAL_MISC(u8, outer_second_cvlan_tag,
+			      outer_second_cvlan_tag, p, "%u");
+	PRINT_MASKED_VAL_MISC(u8, inner_second_cvlan_tag,
+			      inner_second_cvlan_tag, p, "%u");
+	PRINT_MASKED_VAL_MISC(u8, outer_second_svlan_tag,
+			      outer_second_svlan_tag, p, "%u");
+	PRINT_MASKED_VAL_MISC(u8, inner_second_svlan_tag,
+			      inner_second_svlan_tag, p, "%u");
+
+	PRINT_MASKED_VAL_MISC(u8, gre_protocol, gre_protocol, p, "%u");
+
+	PRINT_MASKED_VAL_MISC(u32, vxlan_vni, vxlan_vni, p, "%u");
+	PRINT_MASKED_VAL_MISC(u32, outer_ipv6_flow_label, outer_ipv6_flow_label,
+			      p, "%x");
+	PRINT_MASKED_VAL_MISC(u32, inner_ipv6_flow_label, inner_ipv6_flow_label,
+			      p, "%x");
+}
+
+const char *parse_fs_hdrs(struct trace_seq *p,
+			  u8 match_criteria_enable,
+			  const u32 *mask_outer,
+			  const u32 *mask_misc,
+			  const u32 *mask_inner,
+			  const u32 *value_outer,
+			  const u32 *value_misc,
+			  const u32 *value_inner)
+{
+	const char *ret = trace_seq_buffer_ptr(p);
+
+	if (match_criteria_enable &
+	    1 << MLX5_CREATE_FLOW_GROUP_IN_MATCH_CRITERIA_ENABLE_OUTER_HEADERS) {
+		trace_seq_printf(p, "[outer] ");
+		print_lyr_2_4_hdrs(p, mask_outer, value_outer);
+	}
+
+	if (match_criteria_enable &
+	    1 << MLX5_CREATE_FLOW_GROUP_IN_MATCH_CRITERIA_ENABLE_MISC_PARAMETERS) {
+		trace_seq_printf(p, "[misc] ");
+		print_misc_parameters_hdrs(p, mask_misc, value_misc);
+	}
+	if (match_criteria_enable &
+	    1 << MLX5_CREATE_FLOW_GROUP_IN_MATCH_CRITERIA_ENABLE_INNER_HEADERS) {
+		trace_seq_printf(p, "[inner] ");
+		print_lyr_2_4_hdrs(p, mask_inner, value_inner);
+	}
+	trace_seq_putc(p, 0);
+	return ret;
+}
+
+const char *parse_fs_dst(struct trace_seq *p,
+			 const struct mlx5_flow_destination *dst,
+			 u32 counter_id)
+{
+	const char *ret = trace_seq_buffer_ptr(p);
+
+	switch (dst->type) {
+	case MLX5_FLOW_DESTINATION_TYPE_VPORT:
+		trace_seq_printf(p, "vport=%u\n", dst->vport_num);
+		break;
+	case MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE:
+		trace_seq_printf(p, "ft=%p\n", dst->ft);
+		break;
+	case MLX5_FLOW_DESTINATION_TYPE_TIR:
+		trace_seq_printf(p, "tir=%u\n", dst->tir_num);
+		break;
+	case MLX5_FLOW_DESTINATION_TYPE_COUNTER:
+		trace_seq_printf(p, "counter_id=%u\n", counter_id);
+		break;
+	}
+
+	trace_seq_putc(p, 0);
+	return ret;
+}
+
+EXPORT_TRACEPOINT_SYMBOL(mlx5_fs_add_fg);
+EXPORT_TRACEPOINT_SYMBOL(mlx5_fs_del_fg);
+EXPORT_TRACEPOINT_SYMBOL(mlx5_fs_set_fte);
+EXPORT_TRACEPOINT_SYMBOL(mlx5_fs_del_fte);
+EXPORT_TRACEPOINT_SYMBOL(mlx5_fs_add_rule);
+EXPORT_TRACEPOINT_SYMBOL(mlx5_fs_del_rule);
+

drivers/net/ethernet/mellanox/mlx5/core/diag/fs_tracepoint.h

+/*
+ * Copyright (c) 2017, Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#if !defined(_MLX5_FS_TP_) || defined(TRACE_HEADER_MULTI_READ)
+#define _MLX5_FS_TP_
+
+#include <linux/tracepoint.h>
+#include <linux/trace_seq.h>
+#include "../fs_core.h"
+
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM mlx5
+
+#define __parse_fs_hdrs(match_criteria_enable, mouter, mmisc, minner, vouter, \
+			vinner, vmisc)					      \
+	parse_fs_hdrs(p, match_criteria_enable, mouter, mmisc, minner, vouter,\
+		      vinner, vmisc)
+
+const char *parse_fs_hdrs(struct trace_seq *p,
+			  u8 match_criteria_enable,
+			  const u32 *mask_outer,
+			  const u32 *mask_misc,
+			  const u32 *mask_inner,
+			  const u32 *value_outer,
+			  const u32 *value_misc,
+			  const u32 *value_inner);
+
+#define __parse_fs_dst(dst, counter_id) \
+	parse_fs_dst(p, (const struct mlx5_flow_destination *)dst, counter_id)
+
+const char *parse_fs_dst(struct trace_seq *p,
+			 const struct mlx5_flow_destination *dst,
+			 u32 counter_id);
+
+TRACE_EVENT(mlx5_fs_add_fg,
+	    TP_PROTO(const struct mlx5_flow_group *fg),
+	    TP_ARGS(fg),
+	    TP_STRUCT__entry(
+		__field(const struct mlx5_flow_group *, fg)
+		__field(const struct mlx5_flow_table *, ft)
+		__field(u32, start_index)
+		__field(u32, end_index)
+		__field(u32, id)
+		__field(u8, mask_enable)
+		__array(u32, mask_outer, MLX5_ST_SZ_DW(fte_match_set_lyr_2_4))
+		__array(u32, mask_inner, MLX5_ST_SZ_DW(fte_match_set_lyr_2_4))
+		__array(u32, mask_misc, MLX5_ST_SZ_DW(fte_match_set_misc))
+	    ),
+	    TP_fast_assign(
+			   __entry->fg = fg;
+			   fs_get_obj(__entry->ft, fg->node.parent);
+			   __entry->start_index = fg->start_index;
+			   __entry->end_index = fg->start_index + fg->max_ftes;
+			   __entry->id = fg->id;
+			   __entry->mask_enable = fg->mask.match_criteria_enable;
+			   memcpy(__entry->mask_outer,
+				  MLX5_ADDR_OF(fte_match_param,
+					       &fg->mask.match_criteria,
+					       outer_headers),
+				  sizeof(__entry->mask_outer));
+			   memcpy(__entry->mask_inner,
+				  MLX5_ADDR_OF(fte_match_param,
+					       &fg->mask.match_criteria,
+					       inner_headers),
+				  sizeof(__entry->mask_inner));
+			   memcpy(__entry->mask_misc,
+				  MLX5_ADDR_OF(fte_match_param,
+					       &fg->mask.match_criteria,
+					       misc_parameters),
+				  sizeof(__entry->mask_misc));
+
+	    ),
+	    TP_printk("fg=%p ft=%p id=%u start=%u end=%u bit_mask=%02x %s\n",
+		      __entry->fg, __entry->ft, __entry->id,
+		      __entry->start_index, __entry->end_index,
+		      __entry->mask_enable,
+		      __parse_fs_hdrs(__entry->mask_enable,
+				      __entry->mask_outer,
+				      __entry->mask_misc,
+				      __entry->mask_inner,
+				      __entry->mask_outer,
+				      __entry->mask_misc,
+				      __entry->mask_inner))
+	    );
+
+TRACE_EVENT(mlx5_fs_del_fg,
+	    TP_PROTO(const struct mlx5_flow_group *fg),
+	    TP_ARGS(fg),
+	    TP_STRUCT__entry(
+		__field(const struct mlx5_flow_group *, fg)
+		__field(u32, id)
+	    ),
+	    TP_fast_assign(
+			   __entry->fg = fg;
+			   __entry->id = fg->id;
+
+	    ),
+	    TP_printk("fg=%p id=%u\n",
+		      __entry->fg, __entry->id)
+	    );
+
+#define ACTION_FLAGS \
+	{MLX5_FLOW_CONTEXT_ACTION_ALLOW,	 "ALLOW"},\
+	{MLX5_FLOW_CONTEXT_ACTION_DROP,		 "DROP"},\
+	{MLX5_FLOW_CONTEXT_ACTION_FWD_DEST,	 "FWD"},\
+	{MLX5_FLOW_CONTEXT_ACTION_COUNT,	 "CNT"},\
+	{MLX5_FLOW_CONTEXT_ACTION_ENCAP,	 "ENCAP"},\
+	{MLX5_FLOW_CONTEXT_ACTION_DECAP,	 "DECAP"},\
+	{MLX5_FLOW_CONTEXT_ACTION_MOD_HDR,	 "MOD_HDR"},\
+	{MLX5_FLOW_CONTEXT_ACTION_FWD_NEXT_PRIO, "NEXT_PRIO"}
+
+TRACE_EVENT(mlx5_fs_set_fte,
+	    TP_PROTO(const struct fs_fte *fte, bool new_fte),
+	    TP_ARGS(fte, new_fte),
+	    TP_STRUCT__entry(
+		__field(const struct fs_fte *, fte)
+		__field(const struct mlx5_flow_group *, fg)
+		__field(u32, group_index)
+		__field(u32, index)
+		__field(u32, action)
+		__field(u32, flow_tag)
+		__field(u8,  mask_enable)
+		__field(bool, new_fte)
+		__array(u32, mask_outer, MLX5_ST_SZ_DW(fte_match_set_lyr_2_4))
+		__array(u32, mask_inner, MLX5_ST_SZ_DW(fte_match_set_lyr_2_4))
+		__array(u32, mask_misc, MLX5_ST_SZ_DW(fte_match_set_misc))
+		__array(u32, value_outer, MLX5_ST_SZ_DW(fte_match_set_lyr_2_4))
+		__array(u32, value_inner, MLX5_ST_SZ_DW(fte_match_set_lyr_2_4))
+		__array(u32, value_misc, MLX5_ST_SZ_DW(fte_match_set_misc))
+	    ),
+	    TP_fast_assign(
+			   __entry->fte = fte;
+			   __entry->new_fte = new_fte;
+			   fs_get_obj(__entry->fg, fte->node.parent);
+			   __entry->group_index = __entry->fg->id;
+			   __entry->index = fte->index;
+			   __entry->action = fte->action;
+			   __entry->mask_enable = __entry->fg->mask.match_criteria_enable;
+			   __entry->flow_tag = fte->flow_tag;
+			   memcpy(__entry->mask_outer,
+				  MLX5_ADDR_OF(fte_match_param,
+					       &__entry->fg->mask.match_criteria,
+					       outer_headers),
+				  sizeof(__entry->mask_outer));
+			   memcpy(__entry->mask_inner,
+				  MLX5_ADDR_OF(fte_match_param,
+					       &__entry->fg->mask.match_criteria,
+					       inner_headers),
+				  sizeof(__entry->mask_inner));
+			   memcpy(__entry->mask_misc,
+				  MLX5_ADDR_OF(fte_match_param,
+					       &__entry->fg->mask.match_criteria,
+					       misc_parameters),
+				  sizeof(__entry->mask_misc));
+			   memcpy(__entry->value_outer,
+				  MLX5_ADDR_OF(fte_match_param,
+					       &fte->val,
+					       outer_headers),
+				  sizeof(__entry->value_outer));
+			   memcpy(__entry->value_inner,
+				  MLX5_ADDR_OF(fte_match_param,
+					       &fte->val,
+					       inner_headers),
+				  sizeof(__entry->value_inner));
+			   memcpy(__entry->value_misc,
+				  MLX5_ADDR_OF(fte_match_param,
+					       &fte->val,
+					       misc_parameters),
+				  sizeof(__entry->value_misc));
+
+	    ),
+	    TP_printk("op=%s fte=%p fg=%p index=%u group_index=%u action=<%s> flow_tag=%x %s\n",
+		      __entry->new_fte ? "add" : "set",
+		      __entry->fte, __entry->fg, __entry->index,
+		      __entry->group_index, __print_flags(__entry->action, "|",
+							  ACTION_FLAGS),
+		      __entry->flow_tag,
+		      __parse_fs_hdrs(__entry->mask_enable,
+				      __entry->mask_outer,
+				      __entry->mask_misc,
+				      __entry->mask_inner,
+				      __entry->value_outer,
+				      __entry->value_misc,
+				      __entry->value_inner))
+	    );
+
+TRACE_EVENT(mlx5_fs_del_fte,
+	    TP_PROTO(const struct fs_fte *fte),
+	    TP_ARGS(fte),
+	    TP_STRUCT__entry(
+		__field(const struct fs_fte *, fte)
+		__field(u32, index)
+	    ),
+	    TP_fast_assign(
+			   __entry->fte = fte;
+			   __entry->index = fte->index;
+
+	    ),
+	    TP_printk("fte=%p index=%u\n",
+		      __entry->fte, __entry->index)
+	    );
+
+TRACE_EVENT(mlx5_fs_add_rule,
+	    TP_PROTO(const struct mlx5_flow_rule *rule),
+	    TP_ARGS(rule),
+	    TP_STRUCT__entry(
+		__field(const struct mlx5_flow_rule *, rule)
+		__field(const struct fs_fte *, fte)
+		__field(u32, sw_action)
+		__field(u32, index)
+		__field(u32, counter_id)
+		__array(u8, destination, sizeof(struct mlx5_flow_destination))
+	    ),
+	    TP_fast_assign(
+			   __entry->rule = rule;
+			   fs_get_obj(__entry->fte, rule->node.parent);
+			   __entry->index = __entry->fte->dests_size - 1;
+			   __entry->sw_action = rule->sw_action;
+			   memcpy(__entry->destination,
+				  &rule->dest_attr,
+				  sizeof(__entry->destination));
+			   if (rule->dest_attr.type & MLX5_FLOW_DESTINATION_TYPE_COUNTER &&
+			       rule->dest_attr.counter)
+				__entry->counter_id =
+				rule->dest_attr.counter->id;
+	    ),
+	    TP_printk("rule=%p fte=%p index=%u sw_action=<%s> [dst] %s\n",
+		      __entry->rule, __entry->fte, __entry->index,
+		      __print_flags(__entry->sw_action, "|", ACTION_FLAGS),
+		      __parse_fs_dst(__entry->destination, __entry->counter_id))
+	    );
+
+TRACE_EVENT(mlx5_fs_del_rule,
+	    TP_PROTO(const struct mlx5_flow_rule *rule),
+	    TP_ARGS(rule),
+	    TP_STRUCT__entry(
+		__field(const struct mlx5_flow_rule *, rule)
+		__field(const struct fs_fte *, fte)
+	    ),
+	    TP_fast_assign(
+			   __entry->rule = rule;
+			   fs_get_obj(__entry->fte, rule->node.parent);
+	    ),
+	    TP_printk("rule=%p fte=%p\n",
+		      __entry->rule, __entry->fte)
+	    );
+#endif
+
+#undef TRACE_INCLUDE_PATH
+#define TRACE_INCLUDE_PATH ./diag
+#undef TRACE_INCLUDE_FILE
+#define TRACE_INCLUDE_FILE fs_tracepoint
+#include <trace/define_trace.h>

drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c

@@ -254,8 +254,8 @@ static void mlx5e_fill_stats_strings(struct mlx5e_priv *priv, u8 *data)
 		       pcie_perf_stats_desc64[i].format);
 
 	for (i = 0; i < NUM_PCIE_PERF_STALL_COUNTERS(priv); i++)
-	 strcpy(data + (idx++) * ETH_GSTRING_LEN,
-		pcie_perf_stall_stats_desc[i].format);
+		strcpy(data + (idx++) * ETH_GSTRING_LEN,
+		       pcie_perf_stall_stats_desc[i].format);
 
 	for (prio = 0; prio < NUM_PPORT_PRIO; prio++) {
 		for (i = 0; i < NUM_PPORT_PER_PRIO_TRAFFIC_COUNTERS; i++)

drivers/net/ethernet/mellanox/mlx5/core/eq.c

@@ -188,8 +188,8 @@ static enum mlx5_dev_event port_subtype_event(u8 subtype)
 static void eq_update_ci(struct mlx5_eq *eq, int arm)
 {
 	__be32 __iomem *addr = eq->doorbell + (arm ? 0 : 2);
-
 	u32 val = (eq->cons_index & 0xffffff) | (eq->eqn << 24);
+
 	__raw_writel((__force u32)cpu_to_be32(val), addr);
 	/* We still want ordering, just not swabbing, so add a barrier */
 	mb();

drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c

@@ -263,7 +263,7 @@ static int mlx5_cmd_set_fte(struct mlx5_core_dev *dev,
 	MLX5_SET(flow_context, in_flow_context, modify_header_id, fte->modify_id);
 	in_match_value = MLX5_ADDR_OF(flow_context, in_flow_context,
 				      match_value);
-	memcpy(in_match_value, &fte->val, MLX5_ST_SZ_BYTES(fte_match_param));
+	memcpy(in_match_value, &fte->val, sizeof(fte->val));
 
 	in_dests = MLX5_ADDR_OF(flow_context, in_flow_context, destination);
 	if (fte->action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST) {

drivers/net/ethernet/mellanox/mlx5/core/fs_core.h

@@ -34,6 +34,7 @@
 #define _MLX5_FS_CORE_
 
 #include <linux/mlx5/fs.h>
+#include <linux/rhashtable.h>
 
 enum fs_node_type {
 	FS_TYPE_NAMESPACE,
@@ -118,6 +119,8 @@ struct mlx5_flow_table {
 	/* FWD rules that point on this flow table */
 	struct list_head		fwd_rules;
 	u32				flags;
+	struct ida			fte_allocator;
+	struct rhltable			fgs_hash;
 };
 
 struct mlx5_fc_cache {
@@ -143,10 +146,22 @@ struct mlx5_fc {
 	struct mlx5_fc_cache cache ____cacheline_aligned_in_smp;
 };
 
+#define MLX5_FTE_MATCH_PARAM_RESERVED	reserved_at_600
+/* Calculate the fte_match_param length and without the reserved length.
+ * Make sure the reserved field is the last.
+ */
+#define MLX5_ST_SZ_DW_MATCH_PARAM					    \
+	((MLX5_BYTE_OFF(fte_match_param, MLX5_FTE_MATCH_PARAM_RESERVED) / sizeof(u32)) + \
+	 BUILD_BUG_ON_ZERO(MLX5_ST_SZ_BYTES(fte_match_param) !=		     \
+			   MLX5_FLD_SZ_BYTES(fte_match_param,		     \
+					     MLX5_FTE_MATCH_PARAM_RESERVED) +\
+			   MLX5_BYTE_OFF(fte_match_param,		     \
+					 MLX5_FTE_MATCH_PARAM_RESERVED)))
+
 /* Type of children is mlx5_flow_rule */
 struct fs_fte {
 	struct fs_node			node;
-	u32				val[MLX5_ST_SZ_DW(fte_match_param)];
+	u32				val[MLX5_ST_SZ_DW_MATCH_PARAM];
 	u32				dests_size;
 	u32				flow_tag;
 	u32				index;
@@ -155,6 +170,7 @@ struct fs_fte {
 	u32				modify_id;
 	enum fs_fte_status		status;
 	struct mlx5_fc			*counter;
+	struct rhash_head		hash;
 };
 
 /* Type of children is mlx5_flow_table/namespace */
@@ -174,7 +190,7 @@ struct mlx5_flow_namespace {
 
 struct mlx5_flow_group_mask {
 	u8	match_criteria_enable;
-	u32	match_criteria[MLX5_ST_SZ_DW(fte_match_param)];
+	u32	match_criteria[MLX5_ST_SZ_DW_MATCH_PARAM];
 };
 
 /* Type of children is fs_fte */
@@ -183,8 +199,9 @@ struct mlx5_flow_group {
 	struct mlx5_flow_group_mask	mask;
 	u32				start_index;
 	u32				max_ftes;
-	u32				num_ftes;
 	u32				id;
+	struct rhashtable		ftes_hash;
+	struct rhlist_head		hash;
 };
 
 struct mlx5_flow_root_namespace {

include/linux/mlx5/device.h

@@ -48,7 +48,7 @@
 /* helper macros */
 #define __mlx5_nullp(typ) ((struct mlx5_ifc_##typ##_bits *)0)
 #define __mlx5_bit_sz(typ, fld) sizeof(__mlx5_nullp(typ)->fld)
-#define __mlx5_bit_off(typ, fld) ((unsigned)(unsigned long)(&(__mlx5_nullp(typ)->fld)))
+#define __mlx5_bit_off(typ, fld) (offsetof(struct mlx5_ifc_##typ##_bits, fld))
 #define __mlx5_dw_off(typ, fld) (__mlx5_bit_off(typ, fld) / 32)
 #define __mlx5_64_off(typ, fld) (__mlx5_bit_off(typ, fld) / 64)
 #define __mlx5_dw_bit_off(typ, fld) (32 - __mlx5_bit_sz(typ, fld) - (__mlx5_bit_off(typ, fld) & 0x1f))

include/linux/mlx5/driver.h

@@ -890,8 +890,6 @@ static inline void *mlx5_buf_offset(struct mlx5_buf *buf, int offset)
 		return buf->direct.buf + offset;
 }
 
-extern struct workqueue_struct *mlx5_core_wq;
-
 #define STRUCT_FIELD(header, field) \
 	.struct_offset_bytes = offsetof(struct ib_unpacked_ ## header, field),      \
 	.struct_size_bytes   = sizeof((struct ib_unpacked_ ## header *)0)->field