Merge tag 'mlx5-fixes-2022-11-21' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5 fixes 2022-11-21

This series provides bug fixes to mlx5 driver.

* tag 'mlx5-fixes-2022-11-21' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux:
  net/mlx5e: Fix possible race condition in macsec extended packet number update routine
  net/mlx5e: Fix MACsec update SecY
  net/mlx5e: Fix MACsec SA initialization routine
  net/mlx5e: Remove leftovers from old XSK queues enumeration
  net/mlx5e: Offload rule only when all encaps are valid
  net/mlx5e: Fix missing alignment in size of MTT/KLM entries
  net/mlx5: Fix sync reset event handler error flow
  net/mlx5: E-Switch, Set correctly vport destination
  net/mlx5: Lag, avoid lockdep warnings
  net/mlx5: Fix handling of entry refcount when command is not issued to FW
  net/mlx5: cmdif, Print info on any firmware cmd failure to tracepoint
  net/mlx5: SF: Fix probing active SFs during driver probe phase
  net/mlx5: Fix FW tracer timestamp calculation
  net/mlx5: Do not query pci info while pci disabled
====================

Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>

Author: kuba-moo

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

@@ -45,6 +45,8 @@
 #include "mlx5_core.h"
 #include "lib/eq.h"
 #include "lib/tout.h"
+#define CREATE_TRACE_POINTS
+#include "diag/cmd_tracepoint.h"
 
 enum {
 	CMD_IF_REV = 5,
@@ -785,27 +787,14 @@ EXPORT_SYMBOL(mlx5_cmd_out_err);
 static void cmd_status_print(struct mlx5_core_dev *dev, void *in, void *out)
 {
 	u16 opcode, op_mod;
-	u32 syndrome;
-	u8  status;
 	u16 uid;
-	int err;
-
-	syndrome = MLX5_GET(mbox_out, out, syndrome);
-	status = MLX5_GET(mbox_out, out, status);
 
 	opcode = MLX5_GET(mbox_in, in, opcode);
 	op_mod = MLX5_GET(mbox_in, in, op_mod);
 	uid    = MLX5_GET(mbox_in, in, uid);
 
-	err = cmd_status_to_err(status);
-
 	if (!uid && opcode != MLX5_CMD_OP_DESTROY_MKEY)
 		mlx5_cmd_out_err(dev, opcode, op_mod, out);
-	else
-		mlx5_core_dbg(dev,
-			"%s(0x%x) op_mod(0x%x) uid(%d) failed, status %s(0x%x), syndrome (0x%x), err(%d)\n",
-			mlx5_command_str(opcode), opcode, op_mod, uid,
-			cmd_status_str(status), status, syndrome, err);
 }
 
 int mlx5_cmd_check(struct mlx5_core_dev *dev, int err, void *in, void *out)
@@ -1016,14 +1005,14 @@ static void cmd_work_handler(struct work_struct *work)
 		cmd_ent_get(ent);
 	set_bit(MLX5_CMD_ENT_STATE_PENDING_COMP, &ent->state);
 
+	cmd_ent_get(ent); /* for the _real_ FW event on completion */
 	/* Skip sending command to fw if internal error */
 	if (mlx5_cmd_is_down(dev) || !opcode_allowed(&dev->cmd, ent->op)) {
 		ent->ret = -ENXIO;
 		mlx5_cmd_comp_handler(dev, 1ULL << ent->idx, true);
 		return;
 	}
 
-	cmd_ent_get(ent); /* for the _real_ FW event on completion */
 	/* ring doorbell after the descriptor is valid */
 	mlx5_core_dbg(dev, "writing 0x%x to command doorbell\n", 1 << ent->idx);
 	wmb();
@@ -1672,8 +1661,8 @@ static void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec, bool force
 				cmd_ent_put(ent); /* timeout work was canceled */
 
 			if (!forced || /* Real FW completion */
-			    pci_channel_offline(dev->pdev) || /* FW is inaccessible */
-			    dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR)
+			     mlx5_cmd_is_down(dev) || /* No real FW completion is expected */
+			     !opcode_allowed(cmd, ent->op))
 				cmd_ent_put(ent);
 
 			ent->ts2 = ktime_get_ns();
@@ -1892,6 +1881,16 @@ static int cmd_exec(struct mlx5_core_dev *dev, void *in, int in_size, void *out,
 	return err;
 }
 
+static void mlx5_cmd_err_trace(struct mlx5_core_dev *dev, u16 opcode, u16 op_mod, void *out)
+{
+	u32 syndrome = MLX5_GET(mbox_out, out, syndrome);
+	u8 status = MLX5_GET(mbox_out, out, status);
+
+	trace_mlx5_cmd(mlx5_command_str(opcode), opcode, op_mod,
+		       cmd_status_str(status), status, syndrome,
+		       cmd_status_to_err(status));
+}
+
 static void cmd_status_log(struct mlx5_core_dev *dev, u16 opcode, u8 status,
 			   u32 syndrome, int err)
 {
@@ -1914,16 +1913,18 @@ static void cmd_status_log(struct mlx5_core_dev *dev, u16 opcode, u8 status,
 }
 
 /* preserve -EREMOTEIO for outbox.status != OK, otherwise return err as is */
-static int cmd_status_err(struct mlx5_core_dev *dev, int err, u16 opcode, void *out)
+static int cmd_status_err(struct mlx5_core_dev *dev, int err, u16 opcode, u16 op_mod, void *out)
 {
 	u32 syndrome = MLX5_GET(mbox_out, out, syndrome);
 	u8 status = MLX5_GET(mbox_out, out, status);
 
 	if (err == -EREMOTEIO) /* -EREMOTEIO is preserved */
 		err = -EIO;
 
-	if (!err && status != MLX5_CMD_STAT_OK)
+	if (!err && status != MLX5_CMD_STAT_OK) {
 		err = -EREMOTEIO;
+		mlx5_cmd_err_trace(dev, opcode, op_mod, out);
+	}
 
 	cmd_status_log(dev, opcode, status, syndrome, err);
 	return err;
@@ -1951,9 +1952,9 @@ int mlx5_cmd_do(struct mlx5_core_dev *dev, void *in, int in_size, void *out, int
 {
 	int err = cmd_exec(dev, in, in_size, out, out_size, NULL, NULL, false);
 	u16 opcode = MLX5_GET(mbox_in, in, opcode);
+	u16 op_mod = MLX5_GET(mbox_in, in, op_mod);
 
-	err = cmd_status_err(dev, err, opcode, out);
-	return err;
+	return cmd_status_err(dev, err, opcode, op_mod, out);
 }
 EXPORT_SYMBOL(mlx5_cmd_do);
 
@@ -1997,8 +1998,9 @@ int mlx5_cmd_exec_polling(struct mlx5_core_dev *dev, void *in, int in_size,
 {
 	int err = cmd_exec(dev, in, in_size, out, out_size, NULL, NULL, true);
 	u16 opcode = MLX5_GET(mbox_in, in, opcode);
+	u16 op_mod = MLX5_GET(mbox_in, in, op_mod);
 
-	err = cmd_status_err(dev, err, opcode, out);
+	err = cmd_status_err(dev, err, opcode, op_mod, out);
 	return mlx5_cmd_check(dev, err, in, out);
 }
 EXPORT_SYMBOL(mlx5_cmd_exec_polling);
@@ -2034,7 +2036,7 @@ static void mlx5_cmd_exec_cb_handler(int status, void *_work)
 	struct mlx5_async_ctx *ctx;
 
 	ctx = work->ctx;
-	status = cmd_status_err(ctx->dev, status, work->opcode, work->out);
+	status = cmd_status_err(ctx->dev, status, work->opcode, work->op_mod, work->out);
 	work->user_callback(status, work);
 	if (atomic_dec_and_test(&ctx->num_inflight))
 		complete(&ctx->inflight_done);
@@ -2049,6 +2051,7 @@ int mlx5_cmd_exec_cb(struct mlx5_async_ctx *ctx, void *in, int in_size,
 	work->ctx = ctx;
 	work->user_callback = callback;
 	work->opcode = MLX5_GET(mbox_in, in, opcode);
+	work->op_mod = MLX5_GET(mbox_in, in, op_mod);
 	work->out = out;
 	if (WARN_ON(!atomic_inc_not_zero(&ctx->num_inflight)))
 		return -EIO;

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

@@ -0,0 +1,45 @@
+/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */
+/* Copyright (c) 2022, NVIDIA CORPORATION & AFFILIATES. All rights reserved. */
+
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM mlx5
+
+#if !defined(_MLX5_CMD_TP_H_) || defined(TRACE_HEADER_MULTI_READ)
+#define _MLX5_CMD_TP_H_
+
+#include <linux/tracepoint.h>
+#include <linux/trace_seq.h>
+
+TRACE_EVENT(mlx5_cmd,
+	    TP_PROTO(const char *command_str, u16 opcode, u16 op_mod,
+		     const char *status_str, u8 status, u32 syndrome, int err),
+	    TP_ARGS(command_str, opcode, op_mod, status_str, status, syndrome, err),
+	    TP_STRUCT__entry(__string(command_str, command_str)
+			     __field(u16, opcode)
+			     __field(u16, op_mod)
+			    __string(status_str, status_str)
+			    __field(u8, status)
+			    __field(u32, syndrome)
+			    __field(int, err)
+			    ),
+	    TP_fast_assign(__assign_str(command_str, command_str);
+			__entry->opcode = opcode;
+			__entry->op_mod = op_mod;
+			__assign_str(status_str, status_str);
+			__entry->status = status;
+			__entry->syndrome = syndrome;
+			__entry->err = err;
+	    ),
+	    TP_printk("%s(0x%x) op_mod(0x%x) failed, status %s(0x%x), syndrome (0x%x), err(%d)",
+		      __get_str(command_str), __entry->opcode, __entry->op_mod,
+		      __get_str(status_str), __entry->status, __entry->syndrome,
+		      __entry->err)
+);
+
+#endif /* _MLX5_CMD_TP_H_ */
+
+#undef TRACE_INCLUDE_PATH
+#define TRACE_INCLUDE_PATH ./diag
+#undef TRACE_INCLUDE_FILE
+#define TRACE_INCLUDE_FILE cmd_tracepoint
+#include <trace/define_trace.h>

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

@@ -638,7 +638,7 @@ static void mlx5_tracer_handle_timestamp_trace(struct mlx5_fw_tracer *tracer,
 			trace_timestamp = (timestamp_event.timestamp & MASK_52_7) |
 					  (str_frmt->timestamp & MASK_6_0);
 		else
-			trace_timestamp = ((timestamp_event.timestamp & MASK_52_7) - 1) |
+			trace_timestamp = ((timestamp_event.timestamp - 1) & MASK_52_7) |
 					  (str_frmt->timestamp & MASK_6_0);
 
 		mlx5_tracer_print_trace(str_frmt, dev, trace_timestamp);

drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c

@@ -224,15 +224,16 @@ void mlx5e_tc_encap_flows_del(struct mlx5e_priv *priv,
 	list_for_each_entry(flow, flow_list, tmp_list) {
 		if (!mlx5e_is_offloaded_flow(flow) || flow_flag_test(flow, SLOW))
 			continue;
-		spec = &flow->attr->parse_attr->spec;
-
-		/* update from encap rule to slow path rule */
-		rule = mlx5e_tc_offload_to_slow_path(esw, flow, spec);
 
 		attr = mlx5e_tc_get_encap_attr(flow);
 		esw_attr = attr->esw_attr;
 		/* mark the flow's encap dest as non-valid */
 		esw_attr->dests[flow->tmp_entry_index].flags &= ~MLX5_ESW_DEST_ENCAP_VALID;
+		esw_attr->dests[flow->tmp_entry_index].pkt_reformat = NULL;
+
+		/* update from encap rule to slow path rule */
+		spec = &flow->attr->parse_attr->spec;
+		rule = mlx5e_tc_offload_to_slow_path(esw, flow, spec);
 
 		if (IS_ERR(rule)) {
 			err = PTR_ERR(rule);
@@ -251,6 +252,7 @@ void mlx5e_tc_encap_flows_del(struct mlx5e_priv *priv,
 	/* we know that the encap is valid */
 	e->flags &= ~MLX5_ENCAP_ENTRY_VALID;
 	mlx5_packet_reformat_dealloc(priv->mdev, e->pkt_reformat);
+	e->pkt_reformat = NULL;
 }
 
 static void mlx5e_take_tmp_flow(struct mlx5e_tc_flow *flow,
@@ -762,8 +764,7 @@ int mlx5e_attach_encap(struct mlx5e_priv *priv,
 		       struct net_device *mirred_dev,
 		       int out_index,
 		       struct netlink_ext_ack *extack,
-		       struct net_device **encap_dev,
-		       bool *encap_valid)
+		       struct net_device **encap_dev)
 {
 	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
 	struct mlx5e_tc_flow_parse_attr *parse_attr;
@@ -878,9 +879,8 @@ int mlx5e_attach_encap(struct mlx5e_priv *priv,
 	if (e->flags & MLX5_ENCAP_ENTRY_VALID) {
 		attr->esw_attr->dests[out_index].pkt_reformat = e->pkt_reformat;
 		attr->esw_attr->dests[out_index].flags |= MLX5_ESW_DEST_ENCAP_VALID;
-		*encap_valid = true;
 	} else {
-		*encap_valid = false;
+		flow_flag_set(flow, SLOW);
 	}
 	mutex_unlock(&esw->offloads.encap_tbl_lock);
 

drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.h

@@ -17,8 +17,7 @@ int mlx5e_attach_encap(struct mlx5e_priv *priv,
 		       struct net_device *mirred_dev,
 		       int out_index,
 		       struct netlink_ext_ack *extack,
-		       struct net_device **encap_dev,
-		       bool *encap_valid);
+		       struct net_device **encap_dev);
 
 int mlx5e_attach_decap(struct mlx5e_priv *priv,
 		       struct mlx5e_tc_flow *flow,

drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c

@@ -368,15 +368,15 @@ static int mlx5e_macsec_init_sa(struct macsec_context *ctx,
 	obj_attrs.aso_pdn = macsec->aso.pdn;
 	obj_attrs.epn_state = sa->epn_state;
 
-	if (is_tx) {
-		obj_attrs.ssci = cpu_to_be32((__force u32)ctx->sa.tx_sa->ssci);
-		key = &ctx->sa.tx_sa->key;
-	} else {
-		obj_attrs.ssci = cpu_to_be32((__force u32)ctx->sa.rx_sa->ssci);
-		key = &ctx->sa.rx_sa->key;
+	key = (is_tx) ? &ctx->sa.tx_sa->key : &ctx->sa.rx_sa->key;
+
+	if (sa->epn_state.epn_enabled) {
+		obj_attrs.ssci = (is_tx) ? cpu_to_be32((__force u32)ctx->sa.tx_sa->ssci) :
+					   cpu_to_be32((__force u32)ctx->sa.rx_sa->ssci);
+
+		memcpy(&obj_attrs.salt, &key->salt, sizeof(key->salt));
 	}
 
-	memcpy(&obj_attrs.salt, &key->salt, sizeof(key->salt));
 	obj_attrs.replay_window = ctx->secy->replay_window;
 	obj_attrs.replay_protect = ctx->secy->replay_protect;
 
@@ -1155,7 +1155,7 @@ static int macsec_upd_secy_hw_address(struct macsec_context *ctx,
 				continue;
 
 			if (rx_sa->active) {
-				err = mlx5e_macsec_init_sa(ctx, rx_sa, false, false);
+				err = mlx5e_macsec_init_sa(ctx, rx_sa, true, false);
 				if (err)
 					goto out;
 			}
@@ -1536,6 +1536,8 @@ static void macsec_async_event(struct work_struct *work)
 
 	async_work = container_of(work, struct mlx5e_macsec_async_work, work);
 	macsec = async_work->macsec;
+	mutex_lock(&macsec->lock);
+
 	mdev = async_work->mdev;
 	obj_id = async_work->obj_id;
 	macsec_sa = get_macsec_tx_sa_from_obj_id(macsec, obj_id);
@@ -1557,6 +1559,7 @@ static void macsec_async_event(struct work_struct *work)
 
 out_async_work:
 	kfree(async_work);
+	mutex_unlock(&macsec->lock);
 }
 
 static int macsec_obj_change_event(struct notifier_block *nb, unsigned long event, void *data)

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

@@ -35,7 +35,6 @@
 #include "en.h"
 #include "en/port.h"
 #include "en/params.h"
-#include "en/xsk/pool.h"
 #include "en/ptp.h"
 #include "lib/clock.h"
 #include "en/fs_ethtool.h"
@@ -412,15 +411,8 @@ void mlx5e_ethtool_get_channels(struct mlx5e_priv *priv,
 				struct ethtool_channels *ch)
 {
 	mutex_lock(&priv->state_lock);
-
 	ch->max_combined   = priv->max_nch;
 	ch->combined_count = priv->channels.params.num_channels;
-	if (priv->xsk.refcnt) {
-		/* The upper half are XSK queues. */
-		ch->max_combined *= 2;
-		ch->combined_count *= 2;
-	}
-
 	mutex_unlock(&priv->state_lock);
 }
 
@@ -454,16 +446,6 @@ int mlx5e_ethtool_set_channels(struct mlx5e_priv *priv,
 
 	mutex_lock(&priv->state_lock);
 
-	/* Don't allow changing the number of channels if there is an active
-	 * XSK, because the numeration of the XSK and regular RQs will change.
-	 */
-	if (priv->xsk.refcnt) {
-		err = -EINVAL;
-		netdev_err(priv->netdev, "%s: AF_XDP is active, cannot change the number of channels\n",
-			   __func__);
-		goto out;
-	}
-
 	/* Don't allow changing the number of channels if HTB offload is active,
 	 * because the numeration of the QoS SQs will change, while per-queue
 	 * qdiscs are attached.

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

@@ -206,10 +206,11 @@ static void mlx5e_disable_blocking_events(struct mlx5e_priv *priv)
 static u16 mlx5e_mpwrq_umr_octowords(u32 entries, enum mlx5e_mpwrq_umr_mode umr_mode)
 {
 	u8 umr_entry_size = mlx5e_mpwrq_umr_entry_size(umr_mode);
+	u32 sz;
 
-	WARN_ON_ONCE(entries * umr_entry_size % MLX5_OCTWORD);
+	sz = ALIGN(entries * umr_entry_size, MLX5_UMR_MTT_ALIGNMENT);
 
-	return entries * umr_entry_size / MLX5_OCTWORD;
+	return sz / MLX5_OCTWORD;
 }
 
 static inline void mlx5e_build_umr_wqe(struct mlx5e_rq *rq,