Commit 45fee8ed authored by Moshe Shemesh's avatar Moshe Shemesh Committed by Saeed Mahameed
Browse files

net/mlx5: Add clarification on sync reset failure



In case devlink reload action fw_activate failed in sync reset stage,
use the new MFRL field reset_state to find why it failed and share this
clarification with the user.

Signed-off-by: default avatarMoshe Shemesh <moshe@nvidia.com>
Signed-off-by: default avatarSaeed Mahameed <saeedm@nvidia.com>
parent 72fb3b60
Loading
Loading
Loading
Loading
+3 −7
Original line number Diff line number Diff line
@@ -100,15 +100,11 @@ static int mlx5_devlink_reload_fw_activate(struct devlink *devlink, struct netli
	}

	net_port_alive = !!(reset_type & MLX5_MFRL_REG_RESET_TYPE_NET_PORT_ALIVE);
	err = mlx5_fw_reset_set_reset_sync(dev, net_port_alive);
	err = mlx5_fw_reset_set_reset_sync(dev, net_port_alive, extack);
	if (err)
		goto out;

	err = mlx5_fw_reset_wait_reset_done(dev);
out:
	if (err)
		NL_SET_ERR_MSG_MOD(extack, "FW activate command failed");
		return err;

	return mlx5_fw_reset_wait_reset_done(dev);
}

static int mlx5_devlink_trigger_fw_live_patch(struct devlink *devlink,
+50 −7
Original line number Diff line number Diff line
@@ -57,7 +57,8 @@ static int mlx5_reg_mfrl_set(struct mlx5_core_dev *dev, u8 reset_level,
	return mlx5_core_access_reg(dev, in, sizeof(in), out, sizeof(out), MLX5_REG_MFRL, 0, 1);
}

static int mlx5_reg_mfrl_query(struct mlx5_core_dev *dev, u8 *reset_level, u8 *reset_type)
static int mlx5_reg_mfrl_query(struct mlx5_core_dev *dev, u8 *reset_level,
			       u8 *reset_type, u8 *reset_state)
{
	u32 out[MLX5_ST_SZ_DW(mfrl_reg)] = {};
	u32 in[MLX5_ST_SZ_DW(mfrl_reg)] = {};
@@ -71,25 +72,67 @@ static int mlx5_reg_mfrl_query(struct mlx5_core_dev *dev, u8 *reset_level, u8 *r
		*reset_level = MLX5_GET(mfrl_reg, out, reset_level);
	if (reset_type)
		*reset_type = MLX5_GET(mfrl_reg, out, reset_type);
	if (reset_state)
		*reset_state = MLX5_GET(mfrl_reg, out, reset_state);

	return 0;
}

int mlx5_fw_reset_query(struct mlx5_core_dev *dev, u8 *reset_level, u8 *reset_type)
{
	return mlx5_reg_mfrl_query(dev, reset_level, reset_type);
	return mlx5_reg_mfrl_query(dev, reset_level, reset_type, NULL);
}

int mlx5_fw_reset_set_reset_sync(struct mlx5_core_dev *dev, u8 reset_type_sel)
static int mlx5_fw_reset_get_reset_state_err(struct mlx5_core_dev *dev,
					     struct netlink_ext_ack *extack)
{
	u8 reset_state;

	if (mlx5_reg_mfrl_query(dev, NULL, NULL, &reset_state))
		goto out;

	switch (reset_state) {
	case MLX5_MFRL_REG_RESET_STATE_IN_NEGOTIATION:
	case MLX5_MFRL_REG_RESET_STATE_RESET_IN_PROGRESS:
		NL_SET_ERR_MSG_MOD(extack, "Sync reset was already triggered");
		return -EBUSY;
	case MLX5_MFRL_REG_RESET_STATE_TIMEOUT:
		NL_SET_ERR_MSG_MOD(extack, "Sync reset got timeout");
		return -ETIMEDOUT;
	case MLX5_MFRL_REG_RESET_STATE_NACK:
		NL_SET_ERR_MSG_MOD(extack, "One of the hosts disabled reset");
		return -EPERM;
	}

out:
	NL_SET_ERR_MSG_MOD(extack, "Sync reset failed");
	return -EIO;
}

int mlx5_fw_reset_set_reset_sync(struct mlx5_core_dev *dev, u8 reset_type_sel,
				 struct netlink_ext_ack *extack)
{
	struct mlx5_fw_reset *fw_reset = dev->priv.fw_reset;
	u32 out[MLX5_ST_SZ_DW(mfrl_reg)] = {};
	u32 in[MLX5_ST_SZ_DW(mfrl_reg)] = {};
	int err;

	set_bit(MLX5_FW_RESET_FLAGS_PENDING_COMP, &fw_reset->reset_flags);
	err = mlx5_reg_mfrl_set(dev, MLX5_MFRL_REG_RESET_LEVEL3, reset_type_sel, 0, true);
	if (err)

	MLX5_SET(mfrl_reg, in, reset_level, MLX5_MFRL_REG_RESET_LEVEL3);
	MLX5_SET(mfrl_reg, in, rst_type_sel, reset_type_sel);
	MLX5_SET(mfrl_reg, in, pci_sync_for_fw_update_start, 1);
	err = mlx5_access_reg(dev, in, sizeof(in), out, sizeof(out),
			      MLX5_REG_MFRL, 0, 1, false);
	if (!err)
		return 0;

	clear_bit(MLX5_FW_RESET_FLAGS_PENDING_COMP, &fw_reset->reset_flags);
	return err;
	if (err == -EREMOTEIO && MLX5_CAP_MCAM_FEATURE(dev, reset_state))
		return mlx5_fw_reset_get_reset_state_err(dev, extack);

	NL_SET_ERR_MSG_MOD(extack, "Sync reset command failed");
	return mlx5_cmd_check(dev, err, in, out);
}

int mlx5_fw_reset_set_live_patch(struct mlx5_core_dev *dev)
+2 −1
Original line number Diff line number Diff line
@@ -9,7 +9,8 @@
void mlx5_fw_reset_enable_remote_dev_reset_set(struct mlx5_core_dev *dev, bool enable);
bool mlx5_fw_reset_enable_remote_dev_reset_get(struct mlx5_core_dev *dev);
int mlx5_fw_reset_query(struct mlx5_core_dev *dev, u8 *reset_level, u8 *reset_type);
int mlx5_fw_reset_set_reset_sync(struct mlx5_core_dev *dev, u8 reset_type_sel);
int mlx5_fw_reset_set_reset_sync(struct mlx5_core_dev *dev, u8 reset_type_sel,
				 struct netlink_ext_ack *extack);
int mlx5_fw_reset_set_live_patch(struct mlx5_core_dev *dev);

int mlx5_fw_reset_wait_reset_done(struct mlx5_core_dev *dev);
+16 −4
Original line number Diff line number Diff line
@@ -33,9 +33,10 @@
#include <linux/mlx5/port.h>
#include "mlx5_core.h"

int mlx5_core_access_reg(struct mlx5_core_dev *dev, void *data_in,
			 int size_in, void *data_out, int size_out,
			 u16 reg_id, int arg, int write)
/* calling with verbose false will not print error to log */
int mlx5_access_reg(struct mlx5_core_dev *dev, void *data_in, int size_in,
		    void *data_out, int size_out, u16 reg_id, int arg,
		    int write, bool verbose)
{
	int outlen = MLX5_ST_SZ_BYTES(access_register_out) + size_out;
	int inlen = MLX5_ST_SZ_BYTES(access_register_in) + size_in;
@@ -57,7 +58,9 @@ int mlx5_core_access_reg(struct mlx5_core_dev *dev, void *data_in,
	MLX5_SET(access_register_in, in, argument, arg);
	MLX5_SET(access_register_in, in, register_id, reg_id);

	err = mlx5_cmd_exec(dev, in, inlen, out, outlen);
	err = mlx5_cmd_do(dev, in, inlen, out, outlen);
	if (verbose)
		err = mlx5_cmd_check(dev, err, in, out);
	if (err)
		goto out;

@@ -69,6 +72,15 @@ int mlx5_core_access_reg(struct mlx5_core_dev *dev, void *data_in,
	kvfree(in);
	return err;
}
EXPORT_SYMBOL_GPL(mlx5_access_reg);

int mlx5_core_access_reg(struct mlx5_core_dev *dev, void *data_in,
			 int size_in, void *data_out, int size_out,
			 u16 reg_id, int arg, int write)
{
	return mlx5_access_reg(dev, data_in, size_in, data_out, size_out,
			       reg_id, arg, write, true);
}
EXPORT_SYMBOL_GPL(mlx5_core_access_reg);

int mlx5_query_pcam_reg(struct mlx5_core_dev *dev, u32 *pcam, u8 feature_group,
+3 −0
Original line number Diff line number Diff line
@@ -1031,6 +1031,9 @@ int mlx5_core_detach_mcg(struct mlx5_core_dev *dev, union ib_gid *mgid, u32 qpn)

void mlx5_qp_debugfs_init(struct mlx5_core_dev *dev);
void mlx5_qp_debugfs_cleanup(struct mlx5_core_dev *dev);
int mlx5_access_reg(struct mlx5_core_dev *dev, void *data_in, int size_in,
		    void *data_out, int size_out, u16 reg_id, int arg,
		    int write, bool verbose);
int mlx5_core_access_reg(struct mlx5_core_dev *dev, void *data_in,
			 int size_in, void *data_out, int size_out,
			 u16 reg_num, int arg, int write);