linux-stable/drivers/net/ethernet/mellanox/mlxsw/core_env.c

1480 lines
41 KiB
C
Raw Normal View History

// SPDX-License-Identifier: BSD-3-Clause OR GPL-2.0
/* Copyright (c) 2018 Mellanox Technologies. All rights reserved */
#include <linux/kernel.h>
#include <linux/err.h>
#include <linux/ethtool.h>
#include <linux/sfp.h>
#include <linux/mutex.h>
#include "core.h"
#include "core_env.h"
#include "item.h"
#include "reg.h"
struct mlxsw_env_module_info {
u64 module_overheat_counter;
bool is_overheat;
int num_ports_mapped;
int num_ports_up;
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
enum ethtool_module_power_mode_policy power_mode_policy;
enum mlxsw_reg_pmtm_module_type type;
};
struct mlxsw_env_line_card {
u8 module_count;
bool active;
struct mlxsw_env_module_info module_info[];
};
struct mlxsw_env {
struct mlxsw_core *core;
const struct mlxsw_bus_info *bus_info;
u8 max_module_count; /* Maximum number of modules per-slot. */
u8 num_of_slots; /* Including the main board. */
u8 max_eeprom_len; /* Maximum module EEPROM transaction length. */
struct mutex line_cards_lock; /* Protects line cards. */
struct mlxsw_env_line_card *line_cards[];
};
static bool __mlxsw_env_linecard_is_active(struct mlxsw_env *mlxsw_env,
u8 slot_index)
{
return mlxsw_env->line_cards[slot_index]->active;
}
static bool mlxsw_env_linecard_is_active(struct mlxsw_env *mlxsw_env,
u8 slot_index)
{
bool active;
mutex_lock(&mlxsw_env->line_cards_lock);
active = __mlxsw_env_linecard_is_active(mlxsw_env, slot_index);
mutex_unlock(&mlxsw_env->line_cards_lock);
return active;
}
static struct
mlxsw_env_module_info *mlxsw_env_module_info_get(struct mlxsw_core *mlxsw_core,
u8 slot_index, u8 module)
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(mlxsw_core);
return &mlxsw_env->line_cards[slot_index]->module_info[module];
}
static int __mlxsw_env_validate_module_type(struct mlxsw_core *core,
u8 slot_index, u8 module)
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(core);
struct mlxsw_env_module_info *module_info;
int err;
if (!__mlxsw_env_linecard_is_active(mlxsw_env, slot_index))
return 0;
module_info = mlxsw_env_module_info_get(core, slot_index, module);
switch (module_info->type) {
case MLXSW_REG_PMTM_MODULE_TYPE_TWISTED_PAIR:
err = -EINVAL;
break;
default:
err = 0;
}
return err;
}
static int mlxsw_env_validate_module_type(struct mlxsw_core *core,
u8 slot_index, u8 module)
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(core);
int err;
mutex_lock(&mlxsw_env->line_cards_lock);
err = __mlxsw_env_validate_module_type(core, slot_index, module);
mutex_unlock(&mlxsw_env->line_cards_lock);
return err;
}
static int
mlxsw_env_validate_cable_ident(struct mlxsw_core *core, u8 slot_index, int id,
bool *qsfp, bool *cmis)
{
char mcia_pl[MLXSW_REG_MCIA_LEN];
char *eeprom_tmp;
u8 ident;
int err;
err = mlxsw_env_validate_module_type(core, slot_index, id);
if (err)
return err;
mlxsw_reg_mcia_pack(mcia_pl, slot_index, id,
MLXSW_REG_MCIA_PAGE0_LO_OFF, 0, 1,
MLXSW_REG_MCIA_I2C_ADDR_LOW);
err = mlxsw_reg_query(core, MLXSW_REG(mcia), mcia_pl);
if (err)
return err;
eeprom_tmp = mlxsw_reg_mcia_eeprom_data(mcia_pl);
ident = eeprom_tmp[0];
*cmis = false;
switch (ident) {
case MLXSW_REG_MCIA_EEPROM_MODULE_INFO_ID_SFP:
*qsfp = false;
break;
case MLXSW_REG_MCIA_EEPROM_MODULE_INFO_ID_QSFP:
case MLXSW_REG_MCIA_EEPROM_MODULE_INFO_ID_QSFP_PLUS:
case MLXSW_REG_MCIA_EEPROM_MODULE_INFO_ID_QSFP28:
*qsfp = true;
break;
case MLXSW_REG_MCIA_EEPROM_MODULE_INFO_ID_QSFP_DD:
case MLXSW_REG_MCIA_EEPROM_MODULE_INFO_ID_OSFP:
*qsfp = true;
*cmis = true;
break;
default:
return -EINVAL;
}
return 0;
}
static int
mlxsw_env_query_module_eeprom(struct mlxsw_core *mlxsw_core, u8 slot_index,
int module, u16 offset, u16 size, void *data,
bool qsfp, unsigned int *p_read_size)
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(mlxsw_core);
char mcia_pl[MLXSW_REG_MCIA_LEN];
char *eeprom_tmp;
u16 i2c_addr;
u8 page = 0;
int status;
int err;
size = min_t(u16, size, mlxsw_env->max_eeprom_len);
if (offset < MLXSW_REG_MCIA_EEPROM_PAGE_LENGTH &&
offset + size > MLXSW_REG_MCIA_EEPROM_PAGE_LENGTH)
/* Cross pages read, read until offset 256 in low page */
size = MLXSW_REG_MCIA_EEPROM_PAGE_LENGTH - offset;
i2c_addr = MLXSW_REG_MCIA_I2C_ADDR_LOW;
if (offset >= MLXSW_REG_MCIA_EEPROM_PAGE_LENGTH) {
if (qsfp) {
/* When reading upper pages 1, 2 and 3 the offset
* starts at 128. Please refer to "QSFP+ Memory Map"
mlxsw: core: Add ethtool support for QSFP-DD transceivers The Quad Small Form Factor Pluggable Double Density (QSFP-DD) hardware specification defines a form factor that supports up to 400 Gbps in aggregate over an 8x50-Gbps electrical interface. The QSFP-DD supports both optical and copper interfaces. Implementation is based on Common Management Interface Specification; Rev 4.0 May 8, 2019. Table 8-2 "Identifier and Status Summary (Lower Page)" from this spec defines "Id and Status" fields located at offsets 00h - 02h. Bit 2 at offset 02h ("Flat_mem") specifies QSFP EEPROM memory mode, which could be "upper memory flat" or "paged". Flat memory mode is coded "1", and indicates that only page 00h is implemented in EEPROM. Paged memory is coded "0" and indicates that pages 00h, 01h, 02h, 10h and 11h are implemented. Pages 10h and 11h are currently not supported by the driver. "Flat" memory mode is used for the passive copper transceivers. For this type only page 00h (256 bytes) is available. "Paged" memory is used for the optical transceivers. For this type pages 00h (256 bytes), 01h (128 bytes) and 02h (128 bytes) are available. Upper page 01h contains static advertising field, while upper page 02h contains the module-defined thresholds and lane-specific monitors. Extend enumerator 'mlxsw_reg_mcia_eeprom_module_info_id' with additional field 'MLXSW_REG_MCIA_EEPROM_MODULE_INFO_TYPE_ID'. This field is used to indicate for QSFP-DD transceiver type which memory mode is to be used. Expose 256 bytes buffer for QSFP-DD passive copper transceiver and 512 bytes buffer for optical. Signed-off-by: Vadim Pasternak <vadimp@mellanox.com> Signed-off-by: Ido Schimmel <idosch@mellanox.com> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-07-28 10:20:15 +00:00
* figure in SFF-8436 specification and to "CMIS Module
* Memory Map" figure in CMIS specification for
* graphical depiction.
*/
page = MLXSW_REG_MCIA_PAGE_GET(offset);
offset -= MLXSW_REG_MCIA_EEPROM_UP_PAGE_LENGTH * page;
if (offset + size > MLXSW_REG_MCIA_EEPROM_PAGE_LENGTH)
size = MLXSW_REG_MCIA_EEPROM_PAGE_LENGTH - offset;
} else {
/* When reading upper pages 1, 2 and 3 the offset
* starts at 0 and I2C high address is used. Please refer
* to "Memory Organization" figure in SFF-8472
* specification for graphical depiction.
*/
i2c_addr = MLXSW_REG_MCIA_I2C_ADDR_HIGH;
offset -= MLXSW_REG_MCIA_EEPROM_PAGE_LENGTH;
}
}
mlxsw_reg_mcia_pack(mcia_pl, slot_index, module, page, offset, size,
i2c_addr);
err = mlxsw_reg_query(mlxsw_core, MLXSW_REG(mcia), mcia_pl);
if (err)
return err;
status = mlxsw_reg_mcia_status_get(mcia_pl);
if (status)
return -EIO;
eeprom_tmp = mlxsw_reg_mcia_eeprom_data(mcia_pl);
memcpy(data, eeprom_tmp, size);
*p_read_size = size;
return 0;
}
int
mlxsw_env_module_temp_thresholds_get(struct mlxsw_core *core, u8 slot_index,
int module, int off, int *temp)
{
unsigned int module_temp, module_crit, module_emerg;
union {
u8 buf[MLXSW_REG_MCIA_TH_ITEM_SIZE];
u16 temp;
} temp_thresh;
char mcia_pl[MLXSW_REG_MCIA_LEN] = {0};
char mtmp_pl[MLXSW_REG_MTMP_LEN];
char *eeprom_tmp;
bool qsfp, cmis;
int page;
int err;
mlxsw_reg_mtmp_pack(mtmp_pl, slot_index,
MLXSW_REG_MTMP_MODULE_INDEX_MIN + module, false,
false);
err = mlxsw_reg_query(core, MLXSW_REG(mtmp), mtmp_pl);
if (err)
return err;
mlxsw_reg_mtmp_unpack(mtmp_pl, &module_temp, NULL, &module_crit,
&module_emerg, NULL);
if (!module_temp) {
*temp = 0;
return 0;
}
/* Validate if threshold reading is available through MTMP register,
* otherwise fallback to read through MCIA.
*/
if (module_emerg) {
*temp = off == SFP_TEMP_HIGH_WARN ? module_crit : module_emerg;
return 0;
}
/* Read Free Side Device Temperature Thresholds from page 03h
* (MSB at lower byte address).
* Bytes:
* 128-129 - Temp High Alarm (SFP_TEMP_HIGH_ALARM);
* 130-131 - Temp Low Alarm (SFP_TEMP_LOW_ALARM);
* 132-133 - Temp High Warning (SFP_TEMP_HIGH_WARN);
* 134-135 - Temp Low Warning (SFP_TEMP_LOW_WARN);
*/
/* Validate module identifier value. */
err = mlxsw_env_validate_cable_ident(core, slot_index, module, &qsfp,
&cmis);
if (err)
return err;
if (qsfp) {
/* For QSFP/CMIS module-defined thresholds are located in page
* 02h, otherwise in page 03h.
*/
if (cmis)
page = MLXSW_REG_MCIA_TH_PAGE_CMIS_NUM;
else
page = MLXSW_REG_MCIA_TH_PAGE_NUM;
mlxsw_reg_mcia_pack(mcia_pl, slot_index, module, page,
MLXSW_REG_MCIA_TH_PAGE_OFF + off,
MLXSW_REG_MCIA_TH_ITEM_SIZE,
MLXSW_REG_MCIA_I2C_ADDR_LOW);
} else {
mlxsw_reg_mcia_pack(mcia_pl, slot_index, module,
MLXSW_REG_MCIA_PAGE0_LO,
off, MLXSW_REG_MCIA_TH_ITEM_SIZE,
MLXSW_REG_MCIA_I2C_ADDR_HIGH);
}
err = mlxsw_reg_query(core, MLXSW_REG(mcia), mcia_pl);
if (err)
return err;
eeprom_tmp = mlxsw_reg_mcia_eeprom_data(mcia_pl);
memcpy(temp_thresh.buf, eeprom_tmp, MLXSW_REG_MCIA_TH_ITEM_SIZE);
*temp = temp_thresh.temp * 1000;
return 0;
}
int mlxsw_env_get_module_info(struct net_device *netdev,
struct mlxsw_core *mlxsw_core, u8 slot_index,
int module, struct ethtool_modinfo *modinfo)
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(mlxsw_core);
u8 module_info[MLXSW_REG_MCIA_EEPROM_MODULE_INFO_SIZE];
u16 offset = MLXSW_REG_MCIA_EEPROM_MODULE_INFO_SIZE;
u8 module_rev_id, module_id, diag_mon;
unsigned int read_size;
int err;
if (!mlxsw_env_linecard_is_active(mlxsw_env, slot_index)) {
netdev_err(netdev, "Cannot read EEPROM of module on an inactive line card\n");
return -EIO;
}
err = mlxsw_env_validate_module_type(mlxsw_core, slot_index, module);
if (err) {
netdev_err(netdev,
"EEPROM is not equipped on port module type");
return err;
}
err = mlxsw_env_query_module_eeprom(mlxsw_core, slot_index, module, 0,
offset, module_info, false,
&read_size);
if (err)
return err;
if (read_size < offset)
return -EIO;
module_rev_id = module_info[MLXSW_REG_MCIA_EEPROM_MODULE_INFO_REV_ID];
module_id = module_info[MLXSW_REG_MCIA_EEPROM_MODULE_INFO_ID];
switch (module_id) {
case MLXSW_REG_MCIA_EEPROM_MODULE_INFO_ID_QSFP:
modinfo->type = ETH_MODULE_SFF_8436;
modinfo->eeprom_len = ETH_MODULE_SFF_8436_MAX_LEN;
break;
case MLXSW_REG_MCIA_EEPROM_MODULE_INFO_ID_QSFP_PLUS:
case MLXSW_REG_MCIA_EEPROM_MODULE_INFO_ID_QSFP28:
if (module_id == MLXSW_REG_MCIA_EEPROM_MODULE_INFO_ID_QSFP28 ||
module_rev_id >=
MLXSW_REG_MCIA_EEPROM_MODULE_INFO_REV_ID_8636) {
modinfo->type = ETH_MODULE_SFF_8636;
modinfo->eeprom_len = ETH_MODULE_SFF_8636_MAX_LEN;
} else {
modinfo->type = ETH_MODULE_SFF_8436;
modinfo->eeprom_len = ETH_MODULE_SFF_8436_MAX_LEN;
}
break;
case MLXSW_REG_MCIA_EEPROM_MODULE_INFO_ID_SFP:
/* Verify if transceiver provides diagnostic monitoring page */
err = mlxsw_env_query_module_eeprom(mlxsw_core, slot_index,
module, SFP_DIAGMON, 1,
&diag_mon, false,
&read_size);
if (err)
return err;
if (read_size < 1)
return -EIO;
modinfo->type = ETH_MODULE_SFF_8472;
if (diag_mon)
modinfo->eeprom_len = ETH_MODULE_SFF_8472_LEN;
else
modinfo->eeprom_len = ETH_MODULE_SFF_8472_LEN / 2;
break;
mlxsw: core: Add ethtool support for QSFP-DD transceivers The Quad Small Form Factor Pluggable Double Density (QSFP-DD) hardware specification defines a form factor that supports up to 400 Gbps in aggregate over an 8x50-Gbps electrical interface. The QSFP-DD supports both optical and copper interfaces. Implementation is based on Common Management Interface Specification; Rev 4.0 May 8, 2019. Table 8-2 "Identifier and Status Summary (Lower Page)" from this spec defines "Id and Status" fields located at offsets 00h - 02h. Bit 2 at offset 02h ("Flat_mem") specifies QSFP EEPROM memory mode, which could be "upper memory flat" or "paged". Flat memory mode is coded "1", and indicates that only page 00h is implemented in EEPROM. Paged memory is coded "0" and indicates that pages 00h, 01h, 02h, 10h and 11h are implemented. Pages 10h and 11h are currently not supported by the driver. "Flat" memory mode is used for the passive copper transceivers. For this type only page 00h (256 bytes) is available. "Paged" memory is used for the optical transceivers. For this type pages 00h (256 bytes), 01h (128 bytes) and 02h (128 bytes) are available. Upper page 01h contains static advertising field, while upper page 02h contains the module-defined thresholds and lane-specific monitors. Extend enumerator 'mlxsw_reg_mcia_eeprom_module_info_id' with additional field 'MLXSW_REG_MCIA_EEPROM_MODULE_INFO_TYPE_ID'. This field is used to indicate for QSFP-DD transceiver type which memory mode is to be used. Expose 256 bytes buffer for QSFP-DD passive copper transceiver and 512 bytes buffer for optical. Signed-off-by: Vadim Pasternak <vadimp@mellanox.com> Signed-off-by: Ido Schimmel <idosch@mellanox.com> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-07-28 10:20:15 +00:00
case MLXSW_REG_MCIA_EEPROM_MODULE_INFO_ID_QSFP_DD:
case MLXSW_REG_MCIA_EEPROM_MODULE_INFO_ID_OSFP:
mlxsw: core: Add ethtool support for QSFP-DD transceivers The Quad Small Form Factor Pluggable Double Density (QSFP-DD) hardware specification defines a form factor that supports up to 400 Gbps in aggregate over an 8x50-Gbps electrical interface. The QSFP-DD supports both optical and copper interfaces. Implementation is based on Common Management Interface Specification; Rev 4.0 May 8, 2019. Table 8-2 "Identifier and Status Summary (Lower Page)" from this spec defines "Id and Status" fields located at offsets 00h - 02h. Bit 2 at offset 02h ("Flat_mem") specifies QSFP EEPROM memory mode, which could be "upper memory flat" or "paged". Flat memory mode is coded "1", and indicates that only page 00h is implemented in EEPROM. Paged memory is coded "0" and indicates that pages 00h, 01h, 02h, 10h and 11h are implemented. Pages 10h and 11h are currently not supported by the driver. "Flat" memory mode is used for the passive copper transceivers. For this type only page 00h (256 bytes) is available. "Paged" memory is used for the optical transceivers. For this type pages 00h (256 bytes), 01h (128 bytes) and 02h (128 bytes) are available. Upper page 01h contains static advertising field, while upper page 02h contains the module-defined thresholds and lane-specific monitors. Extend enumerator 'mlxsw_reg_mcia_eeprom_module_info_id' with additional field 'MLXSW_REG_MCIA_EEPROM_MODULE_INFO_TYPE_ID'. This field is used to indicate for QSFP-DD transceiver type which memory mode is to be used. Expose 256 bytes buffer for QSFP-DD passive copper transceiver and 512 bytes buffer for optical. Signed-off-by: Vadim Pasternak <vadimp@mellanox.com> Signed-off-by: Ido Schimmel <idosch@mellanox.com> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-07-28 10:20:15 +00:00
/* Use SFF_8636 as base type. ethtool should recognize specific
* type through the identifier value.
*/
modinfo->type = ETH_MODULE_SFF_8636;
/* Verify if module EEPROM is a flat memory. In case of flat
* memory only page 00h (0-255 bytes) can be read. Otherwise
* upper pages 01h and 02h can also be read. Upper pages 10h
* and 11h are currently not supported by the driver.
*/
if (module_info[MLXSW_REG_MCIA_EEPROM_MODULE_INFO_TYPE_ID] &
MLXSW_REG_MCIA_EEPROM_CMIS_FLAT_MEMORY)
modinfo->eeprom_len = ETH_MODULE_SFF_8636_LEN;
else
modinfo->eeprom_len = ETH_MODULE_SFF_8472_LEN;
break;
default:
return -EINVAL;
}
return 0;
}
EXPORT_SYMBOL(mlxsw_env_get_module_info);
int mlxsw_env_get_module_eeprom(struct net_device *netdev,
struct mlxsw_core *mlxsw_core, u8 slot_index,
int module, struct ethtool_eeprom *ee,
u8 *data)
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(mlxsw_core);
int offset = ee->offset;
unsigned int read_size;
bool qsfp, cmis;
int i = 0;
int err;
if (!ee->len)
return -EINVAL;
if (!mlxsw_env_linecard_is_active(mlxsw_env, slot_index)) {
netdev_err(netdev, "Cannot read EEPROM of module on an inactive line card\n");
return -EIO;
}
memset(data, 0, ee->len);
/* Validate module identifier value. */
err = mlxsw_env_validate_cable_ident(mlxsw_core, slot_index, module,
&qsfp, &cmis);
if (err)
return err;
while (i < ee->len) {
err = mlxsw_env_query_module_eeprom(mlxsw_core, slot_index,
module, offset,
ee->len - i, data + i,
qsfp, &read_size);
if (err) {
netdev_err(netdev, "Eeprom query failed\n");
return err;
}
i += read_size;
offset += read_size;
}
return 0;
}
EXPORT_SYMBOL(mlxsw_env_get_module_eeprom);
static int mlxsw_env_mcia_status_process(const char *mcia_pl,
struct netlink_ext_ack *extack)
{
u8 status = mlxsw_reg_mcia_status_get(mcia_pl);
switch (status) {
case MLXSW_REG_MCIA_STATUS_GOOD:
return 0;
case MLXSW_REG_MCIA_STATUS_NO_EEPROM_MODULE:
NL_SET_ERR_MSG_MOD(extack, "No response from module's EEPROM");
return -EIO;
case MLXSW_REG_MCIA_STATUS_MODULE_NOT_SUPPORTED:
NL_SET_ERR_MSG_MOD(extack, "Module type not supported by the device");
return -EOPNOTSUPP;
case MLXSW_REG_MCIA_STATUS_MODULE_NOT_CONNECTED:
NL_SET_ERR_MSG_MOD(extack, "No module present indication");
return -EIO;
case MLXSW_REG_MCIA_STATUS_I2C_ERROR:
NL_SET_ERR_MSG_MOD(extack, "Error occurred while trying to access module's EEPROM using I2C");
return -EIO;
case MLXSW_REG_MCIA_STATUS_MODULE_DISABLED:
NL_SET_ERR_MSG_MOD(extack, "Module is disabled");
return -EIO;
default:
NL_SET_ERR_MSG_MOD(extack, "Unknown error");
return -EIO;
}
}
int
mlxsw_env_get_module_eeprom_by_page(struct mlxsw_core *mlxsw_core,
u8 slot_index, u8 module,
const struct ethtool_module_eeprom *page,
struct netlink_ext_ack *extack)
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(mlxsw_core);
u32 bytes_read = 0;
u16 device_addr;
int err;
if (!mlxsw_env_linecard_is_active(mlxsw_env, slot_index)) {
NL_SET_ERR_MSG_MOD(extack,
"Cannot read EEPROM of module on an inactive line card");
return -EIO;
}
err = mlxsw_env_validate_module_type(mlxsw_core, slot_index, module);
if (err) {
NL_SET_ERR_MSG_MOD(extack, "EEPROM is not equipped on port module type");
return err;
}
/* Offset cannot be larger than 2 * ETH_MODULE_EEPROM_PAGE_LEN */
device_addr = page->offset;
while (bytes_read < page->length) {
char mcia_pl[MLXSW_REG_MCIA_LEN];
char *eeprom_tmp;
u8 size;
size = min_t(u8, page->length - bytes_read,
mlxsw_env->max_eeprom_len);
mlxsw_reg_mcia_pack(mcia_pl, slot_index, module, page->page,
device_addr + bytes_read, size,
page->i2c_address);
mlxsw_reg_mcia_bank_number_set(mcia_pl, page->bank);
err = mlxsw_reg_query(mlxsw_core, MLXSW_REG(mcia), mcia_pl);
if (err) {
NL_SET_ERR_MSG_MOD(extack, "Failed to access module's EEPROM");
return err;
}
err = mlxsw_env_mcia_status_process(mcia_pl, extack);
if (err)
return err;
eeprom_tmp = mlxsw_reg_mcia_eeprom_data(mcia_pl);
memcpy(page->data + bytes_read, eeprom_tmp, size);
bytes_read += size;
}
return bytes_read;
}
EXPORT_SYMBOL(mlxsw_env_get_module_eeprom_by_page);
static int mlxsw_env_module_reset(struct mlxsw_core *mlxsw_core, u8 slot_index,
u8 module)
mlxsw: Add support for transceiver modules reset Implement support for ethtool_ops::reset in order to reset transceiver modules. The module backing the netdev is reset when the 'ETH_RESET_PHY' flag is set. After a successful reset, the flag is cleared by the driver and other flags are ignored. This is in accordance with the interface documentation: "The reset() operation must clear the flags for the components which were actually reset. On successful return, the flags indicate the components which were not reset, either because they do not exist in the hardware or because they cannot be reset independently. The driver must never reset any components that were not requested." Reset is useful in order to allow a module to transition out of a fault state. From section 6.3.2.12 in CMIS 5.0: "Except for a power cycle, the only exit path from the ModuleFault state is to perform a module reset by taking an action that causes the ResetS transition signal to become TRUE (see Table 6-11)". An error is returned when the netdev is administratively up: # ip link set dev swp11 up # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Cannot issue ETHTOOL_RESET: Invalid argument # ip link set dev swp11 down # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Components reset: 0x40 An error is returned when the module is shared by multiple ports (split ports) and the "phy-shared" flag is not set: # devlink port split swp11 count 4 # ethtool --reset swp11s0 phy ETHTOOL_RESET 0x40 Cannot issue ETHTOOL_RESET: Invalid argument # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Components reset: 0x400000 # devlink port unsplit swp11s0 # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Components reset: 0x40 An error is also returned when one of the ports using the module is administratively up: # devlink port split swp11 count 4 # ip link set dev swp11s1 up # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Cannot issue ETHTOOL_RESET: Invalid argument # ip link set dev swp11s1 down # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Components reset: 0x400000 Reset is performed by writing to the "rst" bit of the PMAOS register, which instructs the firmware to assert the reset signal connected to the module for a fixed amount of time. Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-09-15 10:13:14 +00:00
{
char pmaos_pl[MLXSW_REG_PMAOS_LEN];
mlxsw_reg_pmaos_pack(pmaos_pl, slot_index, module);
mlxsw: Add support for transceiver modules reset Implement support for ethtool_ops::reset in order to reset transceiver modules. The module backing the netdev is reset when the 'ETH_RESET_PHY' flag is set. After a successful reset, the flag is cleared by the driver and other flags are ignored. This is in accordance with the interface documentation: "The reset() operation must clear the flags for the components which were actually reset. On successful return, the flags indicate the components which were not reset, either because they do not exist in the hardware or because they cannot be reset independently. The driver must never reset any components that were not requested." Reset is useful in order to allow a module to transition out of a fault state. From section 6.3.2.12 in CMIS 5.0: "Except for a power cycle, the only exit path from the ModuleFault state is to perform a module reset by taking an action that causes the ResetS transition signal to become TRUE (see Table 6-11)". An error is returned when the netdev is administratively up: # ip link set dev swp11 up # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Cannot issue ETHTOOL_RESET: Invalid argument # ip link set dev swp11 down # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Components reset: 0x40 An error is returned when the module is shared by multiple ports (split ports) and the "phy-shared" flag is not set: # devlink port split swp11 count 4 # ethtool --reset swp11s0 phy ETHTOOL_RESET 0x40 Cannot issue ETHTOOL_RESET: Invalid argument # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Components reset: 0x400000 # devlink port unsplit swp11s0 # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Components reset: 0x40 An error is also returned when one of the ports using the module is administratively up: # devlink port split swp11 count 4 # ip link set dev swp11s1 up # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Cannot issue ETHTOOL_RESET: Invalid argument # ip link set dev swp11s1 down # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Components reset: 0x400000 Reset is performed by writing to the "rst" bit of the PMAOS register, which instructs the firmware to assert the reset signal connected to the module for a fixed amount of time. Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-09-15 10:13:14 +00:00
mlxsw_reg_pmaos_rst_set(pmaos_pl, true);
return mlxsw_reg_write(mlxsw_core, MLXSW_REG(pmaos), pmaos_pl);
}
int mlxsw_env_reset_module(struct net_device *netdev,
struct mlxsw_core *mlxsw_core, u8 slot_index,
u8 module, u32 *flags)
mlxsw: Add support for transceiver modules reset Implement support for ethtool_ops::reset in order to reset transceiver modules. The module backing the netdev is reset when the 'ETH_RESET_PHY' flag is set. After a successful reset, the flag is cleared by the driver and other flags are ignored. This is in accordance with the interface documentation: "The reset() operation must clear the flags for the components which were actually reset. On successful return, the flags indicate the components which were not reset, either because they do not exist in the hardware or because they cannot be reset independently. The driver must never reset any components that were not requested." Reset is useful in order to allow a module to transition out of a fault state. From section 6.3.2.12 in CMIS 5.0: "Except for a power cycle, the only exit path from the ModuleFault state is to perform a module reset by taking an action that causes the ResetS transition signal to become TRUE (see Table 6-11)". An error is returned when the netdev is administratively up: # ip link set dev swp11 up # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Cannot issue ETHTOOL_RESET: Invalid argument # ip link set dev swp11 down # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Components reset: 0x40 An error is returned when the module is shared by multiple ports (split ports) and the "phy-shared" flag is not set: # devlink port split swp11 count 4 # ethtool --reset swp11s0 phy ETHTOOL_RESET 0x40 Cannot issue ETHTOOL_RESET: Invalid argument # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Components reset: 0x400000 # devlink port unsplit swp11s0 # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Components reset: 0x40 An error is also returned when one of the ports using the module is administratively up: # devlink port split swp11 count 4 # ip link set dev swp11s1 up # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Cannot issue ETHTOOL_RESET: Invalid argument # ip link set dev swp11s1 down # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Components reset: 0x400000 Reset is performed by writing to the "rst" bit of the PMAOS register, which instructs the firmware to assert the reset signal connected to the module for a fixed amount of time. Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-09-15 10:13:14 +00:00
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(mlxsw_core);
struct mlxsw_env_module_info *module_info;
mlxsw: Add support for transceiver modules reset Implement support for ethtool_ops::reset in order to reset transceiver modules. The module backing the netdev is reset when the 'ETH_RESET_PHY' flag is set. After a successful reset, the flag is cleared by the driver and other flags are ignored. This is in accordance with the interface documentation: "The reset() operation must clear the flags for the components which were actually reset. On successful return, the flags indicate the components which were not reset, either because they do not exist in the hardware or because they cannot be reset independently. The driver must never reset any components that were not requested." Reset is useful in order to allow a module to transition out of a fault state. From section 6.3.2.12 in CMIS 5.0: "Except for a power cycle, the only exit path from the ModuleFault state is to perform a module reset by taking an action that causes the ResetS transition signal to become TRUE (see Table 6-11)". An error is returned when the netdev is administratively up: # ip link set dev swp11 up # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Cannot issue ETHTOOL_RESET: Invalid argument # ip link set dev swp11 down # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Components reset: 0x40 An error is returned when the module is shared by multiple ports (split ports) and the "phy-shared" flag is not set: # devlink port split swp11 count 4 # ethtool --reset swp11s0 phy ETHTOOL_RESET 0x40 Cannot issue ETHTOOL_RESET: Invalid argument # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Components reset: 0x400000 # devlink port unsplit swp11s0 # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Components reset: 0x40 An error is also returned when one of the ports using the module is administratively up: # devlink port split swp11 count 4 # ip link set dev swp11s1 up # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Cannot issue ETHTOOL_RESET: Invalid argument # ip link set dev swp11s1 down # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Components reset: 0x400000 Reset is performed by writing to the "rst" bit of the PMAOS register, which instructs the firmware to assert the reset signal connected to the module for a fixed amount of time. Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-09-15 10:13:14 +00:00
u32 req = *flags;
int err;
if (!(req & ETH_RESET_PHY) &&
!(req & (ETH_RESET_PHY << ETH_RESET_SHARED_SHIFT)))
return 0;
if (!mlxsw_env_linecard_is_active(mlxsw_env, slot_index)) {
netdev_err(netdev, "Cannot reset module on an inactive line card\n");
return -EIO;
}
mutex_lock(&mlxsw_env->line_cards_lock);
mlxsw: Add support for transceiver modules reset Implement support for ethtool_ops::reset in order to reset transceiver modules. The module backing the netdev is reset when the 'ETH_RESET_PHY' flag is set. After a successful reset, the flag is cleared by the driver and other flags are ignored. This is in accordance with the interface documentation: "The reset() operation must clear the flags for the components which were actually reset. On successful return, the flags indicate the components which were not reset, either because they do not exist in the hardware or because they cannot be reset independently. The driver must never reset any components that were not requested." Reset is useful in order to allow a module to transition out of a fault state. From section 6.3.2.12 in CMIS 5.0: "Except for a power cycle, the only exit path from the ModuleFault state is to perform a module reset by taking an action that causes the ResetS transition signal to become TRUE (see Table 6-11)". An error is returned when the netdev is administratively up: # ip link set dev swp11 up # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Cannot issue ETHTOOL_RESET: Invalid argument # ip link set dev swp11 down # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Components reset: 0x40 An error is returned when the module is shared by multiple ports (split ports) and the "phy-shared" flag is not set: # devlink port split swp11 count 4 # ethtool --reset swp11s0 phy ETHTOOL_RESET 0x40 Cannot issue ETHTOOL_RESET: Invalid argument # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Components reset: 0x400000 # devlink port unsplit swp11s0 # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Components reset: 0x40 An error is also returned when one of the ports using the module is administratively up: # devlink port split swp11 count 4 # ip link set dev swp11s1 up # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Cannot issue ETHTOOL_RESET: Invalid argument # ip link set dev swp11s1 down # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Components reset: 0x400000 Reset is performed by writing to the "rst" bit of the PMAOS register, which instructs the firmware to assert the reset signal connected to the module for a fixed amount of time. Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-09-15 10:13:14 +00:00
err = __mlxsw_env_validate_module_type(mlxsw_core, slot_index, module);
if (err) {
netdev_err(netdev, "Reset module is not supported on port module type\n");
goto out;
}
module_info = mlxsw_env_module_info_get(mlxsw_core, slot_index, module);
if (module_info->num_ports_up) {
mlxsw: Add support for transceiver modules reset Implement support for ethtool_ops::reset in order to reset transceiver modules. The module backing the netdev is reset when the 'ETH_RESET_PHY' flag is set. After a successful reset, the flag is cleared by the driver and other flags are ignored. This is in accordance with the interface documentation: "The reset() operation must clear the flags for the components which were actually reset. On successful return, the flags indicate the components which were not reset, either because they do not exist in the hardware or because they cannot be reset independently. The driver must never reset any components that were not requested." Reset is useful in order to allow a module to transition out of a fault state. From section 6.3.2.12 in CMIS 5.0: "Except for a power cycle, the only exit path from the ModuleFault state is to perform a module reset by taking an action that causes the ResetS transition signal to become TRUE (see Table 6-11)". An error is returned when the netdev is administratively up: # ip link set dev swp11 up # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Cannot issue ETHTOOL_RESET: Invalid argument # ip link set dev swp11 down # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Components reset: 0x40 An error is returned when the module is shared by multiple ports (split ports) and the "phy-shared" flag is not set: # devlink port split swp11 count 4 # ethtool --reset swp11s0 phy ETHTOOL_RESET 0x40 Cannot issue ETHTOOL_RESET: Invalid argument # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Components reset: 0x400000 # devlink port unsplit swp11s0 # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Components reset: 0x40 An error is also returned when one of the ports using the module is administratively up: # devlink port split swp11 count 4 # ip link set dev swp11s1 up # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Cannot issue ETHTOOL_RESET: Invalid argument # ip link set dev swp11s1 down # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Components reset: 0x400000 Reset is performed by writing to the "rst" bit of the PMAOS register, which instructs the firmware to assert the reset signal connected to the module for a fixed amount of time. Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-09-15 10:13:14 +00:00
netdev_err(netdev, "Cannot reset module when ports using it are administratively up\n");
err = -EINVAL;
goto out;
}
if (module_info->num_ports_mapped > 1 &&
mlxsw: Add support for transceiver modules reset Implement support for ethtool_ops::reset in order to reset transceiver modules. The module backing the netdev is reset when the 'ETH_RESET_PHY' flag is set. After a successful reset, the flag is cleared by the driver and other flags are ignored. This is in accordance with the interface documentation: "The reset() operation must clear the flags for the components which were actually reset. On successful return, the flags indicate the components which were not reset, either because they do not exist in the hardware or because they cannot be reset independently. The driver must never reset any components that were not requested." Reset is useful in order to allow a module to transition out of a fault state. From section 6.3.2.12 in CMIS 5.0: "Except for a power cycle, the only exit path from the ModuleFault state is to perform a module reset by taking an action that causes the ResetS transition signal to become TRUE (see Table 6-11)". An error is returned when the netdev is administratively up: # ip link set dev swp11 up # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Cannot issue ETHTOOL_RESET: Invalid argument # ip link set dev swp11 down # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Components reset: 0x40 An error is returned when the module is shared by multiple ports (split ports) and the "phy-shared" flag is not set: # devlink port split swp11 count 4 # ethtool --reset swp11s0 phy ETHTOOL_RESET 0x40 Cannot issue ETHTOOL_RESET: Invalid argument # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Components reset: 0x400000 # devlink port unsplit swp11s0 # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Components reset: 0x40 An error is also returned when one of the ports using the module is administratively up: # devlink port split swp11 count 4 # ip link set dev swp11s1 up # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Cannot issue ETHTOOL_RESET: Invalid argument # ip link set dev swp11s1 down # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Components reset: 0x400000 Reset is performed by writing to the "rst" bit of the PMAOS register, which instructs the firmware to assert the reset signal connected to the module for a fixed amount of time. Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-09-15 10:13:14 +00:00
!(req & (ETH_RESET_PHY << ETH_RESET_SHARED_SHIFT))) {
netdev_err(netdev, "Cannot reset module without \"phy-shared\" flag when shared by multiple ports\n");
err = -EINVAL;
goto out;
}
err = mlxsw_env_module_reset(mlxsw_core, slot_index, module);
mlxsw: Add support for transceiver modules reset Implement support for ethtool_ops::reset in order to reset transceiver modules. The module backing the netdev is reset when the 'ETH_RESET_PHY' flag is set. After a successful reset, the flag is cleared by the driver and other flags are ignored. This is in accordance with the interface documentation: "The reset() operation must clear the flags for the components which were actually reset. On successful return, the flags indicate the components which were not reset, either because they do not exist in the hardware or because they cannot be reset independently. The driver must never reset any components that were not requested." Reset is useful in order to allow a module to transition out of a fault state. From section 6.3.2.12 in CMIS 5.0: "Except for a power cycle, the only exit path from the ModuleFault state is to perform a module reset by taking an action that causes the ResetS transition signal to become TRUE (see Table 6-11)". An error is returned when the netdev is administratively up: # ip link set dev swp11 up # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Cannot issue ETHTOOL_RESET: Invalid argument # ip link set dev swp11 down # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Components reset: 0x40 An error is returned when the module is shared by multiple ports (split ports) and the "phy-shared" flag is not set: # devlink port split swp11 count 4 # ethtool --reset swp11s0 phy ETHTOOL_RESET 0x40 Cannot issue ETHTOOL_RESET: Invalid argument # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Components reset: 0x400000 # devlink port unsplit swp11s0 # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Components reset: 0x40 An error is also returned when one of the ports using the module is administratively up: # devlink port split swp11 count 4 # ip link set dev swp11s1 up # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Cannot issue ETHTOOL_RESET: Invalid argument # ip link set dev swp11s1 down # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Components reset: 0x400000 Reset is performed by writing to the "rst" bit of the PMAOS register, which instructs the firmware to assert the reset signal connected to the module for a fixed amount of time. Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-09-15 10:13:14 +00:00
if (err) {
netdev_err(netdev, "Failed to reset module\n");
goto out;
}
*flags &= ~(ETH_RESET_PHY | (ETH_RESET_PHY << ETH_RESET_SHARED_SHIFT));
out:
mutex_unlock(&mlxsw_env->line_cards_lock);
mlxsw: Add support for transceiver modules reset Implement support for ethtool_ops::reset in order to reset transceiver modules. The module backing the netdev is reset when the 'ETH_RESET_PHY' flag is set. After a successful reset, the flag is cleared by the driver and other flags are ignored. This is in accordance with the interface documentation: "The reset() operation must clear the flags for the components which were actually reset. On successful return, the flags indicate the components which were not reset, either because they do not exist in the hardware or because they cannot be reset independently. The driver must never reset any components that were not requested." Reset is useful in order to allow a module to transition out of a fault state. From section 6.3.2.12 in CMIS 5.0: "Except for a power cycle, the only exit path from the ModuleFault state is to perform a module reset by taking an action that causes the ResetS transition signal to become TRUE (see Table 6-11)". An error is returned when the netdev is administratively up: # ip link set dev swp11 up # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Cannot issue ETHTOOL_RESET: Invalid argument # ip link set dev swp11 down # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Components reset: 0x40 An error is returned when the module is shared by multiple ports (split ports) and the "phy-shared" flag is not set: # devlink port split swp11 count 4 # ethtool --reset swp11s0 phy ETHTOOL_RESET 0x40 Cannot issue ETHTOOL_RESET: Invalid argument # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Components reset: 0x400000 # devlink port unsplit swp11s0 # ethtool --reset swp11 phy ETHTOOL_RESET 0x40 Components reset: 0x40 An error is also returned when one of the ports using the module is administratively up: # devlink port split swp11 count 4 # ip link set dev swp11s1 up # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Cannot issue ETHTOOL_RESET: Invalid argument # ip link set dev swp11s1 down # ethtool --reset swp11s0 phy-shared ETHTOOL_RESET 0x400000 Components reset: 0x400000 Reset is performed by writing to the "rst" bit of the PMAOS register, which instructs the firmware to assert the reset signal connected to the module for a fixed amount of time. Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2021-09-15 10:13:14 +00:00
return err;
}
EXPORT_SYMBOL(mlxsw_env_reset_module);
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
int
mlxsw_env_get_module_power_mode(struct mlxsw_core *mlxsw_core, u8 slot_index,
u8 module,
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
struct ethtool_module_power_mode_params *params,
struct netlink_ext_ack *extack)
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(mlxsw_core);
struct mlxsw_env_module_info *module_info;
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
char mcion_pl[MLXSW_REG_MCION_LEN];
u32 status_bits;
int err = 0;
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
mutex_lock(&mlxsw_env->line_cards_lock);
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
err = __mlxsw_env_validate_module_type(mlxsw_core, slot_index, module);
if (err) {
NL_SET_ERR_MSG_MOD(extack, "Power mode is not supported on port module type");
goto out;
}
module_info = mlxsw_env_module_info_get(mlxsw_core, slot_index, module);
params->policy = module_info->power_mode_policy;
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
/* Avoid accessing an inactive line card, as it will result in an error. */
if (!__mlxsw_env_linecard_is_active(mlxsw_env, slot_index))
goto out;
mlxsw_reg_mcion_pack(mcion_pl, slot_index, module);
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
err = mlxsw_reg_query(mlxsw_core, MLXSW_REG(mcion), mcion_pl);
if (err) {
NL_SET_ERR_MSG_MOD(extack, "Failed to retrieve module's power mode");
goto out;
}
status_bits = mlxsw_reg_mcion_module_status_bits_get(mcion_pl);
if (!(status_bits & MLXSW_REG_MCION_MODULE_STATUS_BITS_PRESENT_MASK))
goto out;
if (status_bits & MLXSW_REG_MCION_MODULE_STATUS_BITS_LOW_POWER_MASK)
params->mode = ETHTOOL_MODULE_POWER_MODE_LOW;
else
params->mode = ETHTOOL_MODULE_POWER_MODE_HIGH;
out:
mutex_unlock(&mlxsw_env->line_cards_lock);
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
return err;
}
EXPORT_SYMBOL(mlxsw_env_get_module_power_mode);
static int mlxsw_env_module_enable_set(struct mlxsw_core *mlxsw_core,
u8 slot_index, u8 module, bool enable)
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
{
enum mlxsw_reg_pmaos_admin_status admin_status;
char pmaos_pl[MLXSW_REG_PMAOS_LEN];
mlxsw_reg_pmaos_pack(pmaos_pl, slot_index, module);
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
admin_status = enable ? MLXSW_REG_PMAOS_ADMIN_STATUS_ENABLED :
MLXSW_REG_PMAOS_ADMIN_STATUS_DISABLED;
mlxsw_reg_pmaos_admin_status_set(pmaos_pl, admin_status);
mlxsw_reg_pmaos_ase_set(pmaos_pl, true);
return mlxsw_reg_write(mlxsw_core, MLXSW_REG(pmaos), pmaos_pl);
}
static int mlxsw_env_module_low_power_set(struct mlxsw_core *mlxsw_core,
u8 slot_index, u8 module,
bool low_power)
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
{
u16 eeprom_override_mask, eeprom_override;
char pmmp_pl[MLXSW_REG_PMMP_LEN];
mlxsw_reg_pmmp_pack(pmmp_pl, slot_index, module);
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
mlxsw_reg_pmmp_sticky_set(pmmp_pl, true);
/* Mask all the bits except low power mode. */
eeprom_override_mask = ~MLXSW_REG_PMMP_EEPROM_OVERRIDE_LOW_POWER_MASK;
mlxsw_reg_pmmp_eeprom_override_mask_set(pmmp_pl, eeprom_override_mask);
eeprom_override = low_power ? MLXSW_REG_PMMP_EEPROM_OVERRIDE_LOW_POWER_MASK :
0;
mlxsw_reg_pmmp_eeprom_override_set(pmmp_pl, eeprom_override);
return mlxsw_reg_write(mlxsw_core, MLXSW_REG(pmmp), pmmp_pl);
}
static int __mlxsw_env_set_module_power_mode(struct mlxsw_core *mlxsw_core,
u8 slot_index, u8 module,
bool low_power,
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
struct netlink_ext_ack *extack)
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(mlxsw_core);
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
int err;
/* Avoid accessing an inactive line card, as it will result in an error.
* Cached configuration will be applied by mlxsw_env_got_active() when
* line card becomes active.
*/
if (!__mlxsw_env_linecard_is_active(mlxsw_env, slot_index))
return 0;
err = mlxsw_env_module_enable_set(mlxsw_core, slot_index, module, false);
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
if (err) {
NL_SET_ERR_MSG_MOD(extack, "Failed to disable module");
return err;
}
err = mlxsw_env_module_low_power_set(mlxsw_core, slot_index, module,
low_power);
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
if (err) {
NL_SET_ERR_MSG_MOD(extack, "Failed to set module's power mode");
goto err_module_low_power_set;
}
err = mlxsw_env_module_enable_set(mlxsw_core, slot_index, module, true);
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
if (err) {
NL_SET_ERR_MSG_MOD(extack, "Failed to enable module");
goto err_module_enable_set;
}
return 0;
err_module_enable_set:
mlxsw_env_module_low_power_set(mlxsw_core, slot_index, module,
!low_power);
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
err_module_low_power_set:
mlxsw_env_module_enable_set(mlxsw_core, slot_index, module, true);
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
return err;
}
static int
mlxsw_env_set_module_power_mode_apply(struct mlxsw_core *mlxsw_core,
u8 slot_index, u8 module,
enum ethtool_module_power_mode_policy policy,
struct netlink_ext_ack *extack)
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
{
struct mlxsw_env_module_info *module_info;
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
bool low_power;
int err = 0;
err = __mlxsw_env_validate_module_type(mlxsw_core, slot_index, module);
if (err) {
NL_SET_ERR_MSG_MOD(extack,
"Power mode set is not supported on port module type");
goto out;
}
module_info = mlxsw_env_module_info_get(mlxsw_core, slot_index, module);
if (module_info->power_mode_policy == policy)
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
goto out;
/* If any ports are up, we are already in high power mode. */
if (module_info->num_ports_up)
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
goto out_set_policy;
low_power = policy == ETHTOOL_MODULE_POWER_MODE_POLICY_AUTO;
err = __mlxsw_env_set_module_power_mode(mlxsw_core, slot_index, module,
low_power, extack);
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
if (err)
goto out;
out_set_policy:
module_info->power_mode_policy = policy;
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
out:
return err;
}
int
mlxsw_env_set_module_power_mode(struct mlxsw_core *mlxsw_core, u8 slot_index,
u8 module,
enum ethtool_module_power_mode_policy policy,
struct netlink_ext_ack *extack)
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(mlxsw_core);
int err;
if (policy != ETHTOOL_MODULE_POWER_MODE_POLICY_HIGH &&
policy != ETHTOOL_MODULE_POWER_MODE_POLICY_AUTO) {
NL_SET_ERR_MSG_MOD(extack, "Unsupported power mode policy");
return -EOPNOTSUPP;
}
mutex_lock(&mlxsw_env->line_cards_lock);
err = mlxsw_env_set_module_power_mode_apply(mlxsw_core, slot_index,
module, policy, extack);
mutex_unlock(&mlxsw_env->line_cards_lock);
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
return err;
}
EXPORT_SYMBOL(mlxsw_env_set_module_power_mode);
static int mlxsw_env_module_has_temp_sensor(struct mlxsw_core *mlxsw_core,
u8 slot_index, u8 module,
bool *p_has_temp_sensor)
{
char mtbr_pl[MLXSW_REG_MTBR_LEN];
u16 temp;
int err;
mlxsw_reg_mtbr_pack(mtbr_pl, slot_index,
MLXSW_REG_MTBR_BASE_MODULE_INDEX + module, 1);
err = mlxsw_reg_query(mlxsw_core, MLXSW_REG(mtbr), mtbr_pl);
if (err)
return err;
mlxsw_reg_mtbr_temp_unpack(mtbr_pl, 0, &temp, NULL);
switch (temp) {
case MLXSW_REG_MTBR_BAD_SENS_INFO:
case MLXSW_REG_MTBR_NO_CONN:
case MLXSW_REG_MTBR_NO_TEMP_SENS:
case MLXSW_REG_MTBR_INDEX_NA:
*p_has_temp_sensor = false;
break;
default:
*p_has_temp_sensor = temp ? true : false;
}
return 0;
}
static int
mlxsw_env_temp_event_set(struct mlxsw_core *mlxsw_core, u8 slot_index,
u16 sensor_index, bool enable)
{
char mtmp_pl[MLXSW_REG_MTMP_LEN] = {0};
enum mlxsw_reg_mtmp_tee tee;
int err, threshold_hi;
mlxsw_reg_mtmp_slot_index_set(mtmp_pl, slot_index);
mlxsw_reg_mtmp_sensor_index_set(mtmp_pl, sensor_index);
err = mlxsw_reg_query(mlxsw_core, MLXSW_REG(mtmp), mtmp_pl);
if (err)
return err;
if (enable) {
err = mlxsw_env_module_temp_thresholds_get(mlxsw_core,
slot_index,
sensor_index -
MLXSW_REG_MTMP_MODULE_INDEX_MIN,
SFP_TEMP_HIGH_WARN,
&threshold_hi);
/* In case it is not possible to query the module's threshold,
* use the default value.
*/
if (err)
threshold_hi = MLXSW_REG_MTMP_THRESH_HI;
else
/* mlxsw_env_module_temp_thresholds_get() multiplies
* Celsius degrees by 1000 whereas MTMP expects
* temperature in 0.125 Celsius degrees units.
* Convert threshold_hi to correct units.
*/
threshold_hi = threshold_hi / 1000 * 8;
mlxsw_reg_mtmp_temperature_threshold_hi_set(mtmp_pl, threshold_hi);
mlxsw_reg_mtmp_temperature_threshold_lo_set(mtmp_pl, threshold_hi -
MLXSW_REG_MTMP_HYSTERESIS_TEMP);
}
tee = enable ? MLXSW_REG_MTMP_TEE_GENERATE_EVENT : MLXSW_REG_MTMP_TEE_NO_EVENT;
mlxsw_reg_mtmp_tee_set(mtmp_pl, tee);
return mlxsw_reg_write(mlxsw_core, MLXSW_REG(mtmp), mtmp_pl);
}
static int mlxsw_env_module_temp_event_enable(struct mlxsw_core *mlxsw_core,
u8 slot_index)
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(mlxsw_core);
int i, err, sensor_index;
bool has_temp_sensor;
for (i = 0; i < mlxsw_env->line_cards[slot_index]->module_count; i++) {
err = mlxsw_env_module_has_temp_sensor(mlxsw_core, slot_index,
i, &has_temp_sensor);
if (err)
return err;
if (!has_temp_sensor)
continue;
sensor_index = i + MLXSW_REG_MTMP_MODULE_INDEX_MIN;
err = mlxsw_env_temp_event_set(mlxsw_core, slot_index,
sensor_index, true);
if (err)
return err;
}
return 0;
}
struct mlxsw_env_module_temp_warn_event {
struct mlxsw_env *mlxsw_env;
char mtwe_pl[MLXSW_REG_MTWE_LEN];
struct work_struct work;
};
static void mlxsw_env_mtwe_event_work(struct work_struct *work)
{
struct mlxsw_env_module_temp_warn_event *event;
struct mlxsw_env_module_info *module_info;
struct mlxsw_env *mlxsw_env;
int i, sensor_warning;
bool is_overheat;
event = container_of(work, struct mlxsw_env_module_temp_warn_event,
work);
mlxsw_env = event->mlxsw_env;
for (i = 0; i < mlxsw_env->max_module_count; i++) {
/* 64-127 of sensor_index are mapped to the port modules
* sequentially (module 0 is mapped to sensor_index 64,
* module 1 to sensor_index 65 and so on)
*/
sensor_warning =
mlxsw_reg_mtwe_sensor_warning_get(event->mtwe_pl,
i + MLXSW_REG_MTMP_MODULE_INDEX_MIN);
mutex_lock(&mlxsw_env->line_cards_lock);
/* MTWE only supports main board. */
module_info = mlxsw_env_module_info_get(mlxsw_env->core, 0, i);
is_overheat = module_info->is_overheat;
if ((is_overheat && sensor_warning) ||
(!is_overheat && !sensor_warning)) {
/* Current state is "warning" and MTWE still reports
* warning OR current state in "no warning" and MTWE
* does not report warning.
*/
mutex_unlock(&mlxsw_env->line_cards_lock);
continue;
} else if (is_overheat && !sensor_warning) {
/* MTWE reports "no warning", turn is_overheat off.
*/
module_info->is_overheat = false;
mutex_unlock(&mlxsw_env->line_cards_lock);
} else {
/* Current state is "no warning" and MTWE reports
* "warning", increase the counter and turn is_overheat
* on.
*/
module_info->is_overheat = true;
module_info->module_overheat_counter++;
mutex_unlock(&mlxsw_env->line_cards_lock);
}
}
kfree(event);
}
static void
mlxsw_env_mtwe_listener_func(const struct mlxsw_reg_info *reg, char *mtwe_pl,
void *priv)
{
struct mlxsw_env_module_temp_warn_event *event;
struct mlxsw_env *mlxsw_env = priv;
event = kmalloc(sizeof(*event), GFP_ATOMIC);
if (!event)
return;
event->mlxsw_env = mlxsw_env;
memcpy(event->mtwe_pl, mtwe_pl, MLXSW_REG_MTWE_LEN);
INIT_WORK(&event->work, mlxsw_env_mtwe_event_work);
mlxsw_core_schedule_work(&event->work);
}
static const struct mlxsw_listener mlxsw_env_temp_warn_listener =
MLXSW_CORE_EVENTL(mlxsw_env_mtwe_listener_func, MTWE);
static int mlxsw_env_temp_warn_event_register(struct mlxsw_core *mlxsw_core)
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(mlxsw_core);
return mlxsw_core_trap_register(mlxsw_core,
&mlxsw_env_temp_warn_listener,
mlxsw_env);
}
static void mlxsw_env_temp_warn_event_unregister(struct mlxsw_env *mlxsw_env)
{
mlxsw_core_trap_unregister(mlxsw_env->core,
&mlxsw_env_temp_warn_listener, mlxsw_env);
}
struct mlxsw_env_module_plug_unplug_event {
struct mlxsw_env *mlxsw_env;
u8 slot_index;
u8 module;
struct work_struct work;
};
static void mlxsw_env_pmpe_event_work(struct work_struct *work)
{
struct mlxsw_env_module_plug_unplug_event *event;
struct mlxsw_env_module_info *module_info;
struct mlxsw_env *mlxsw_env;
bool has_temp_sensor;
u16 sensor_index;
int err;
event = container_of(work, struct mlxsw_env_module_plug_unplug_event,
work);
mlxsw_env = event->mlxsw_env;
mutex_lock(&mlxsw_env->line_cards_lock);
module_info = mlxsw_env_module_info_get(mlxsw_env->core,
event->slot_index,
event->module);
module_info->is_overheat = false;
mutex_unlock(&mlxsw_env->line_cards_lock);
err = mlxsw_env_module_has_temp_sensor(mlxsw_env->core,
event->slot_index,
event->module,
&has_temp_sensor);
/* Do not disable events on modules without sensors or faulty sensors
* because FW returns errors.
*/
if (err)
goto out;
if (!has_temp_sensor)
goto out;
sensor_index = event->module + MLXSW_REG_MTMP_MODULE_INDEX_MIN;
mlxsw_env_temp_event_set(mlxsw_env->core, event->slot_index,
sensor_index, true);
out:
kfree(event);
}
static void
mlxsw_env_pmpe_listener_func(const struct mlxsw_reg_info *reg, char *pmpe_pl,
void *priv)
{
u8 slot_index = mlxsw_reg_pmpe_slot_index_get(pmpe_pl);
struct mlxsw_env_module_plug_unplug_event *event;
enum mlxsw_reg_pmpe_module_status module_status;
u8 module = mlxsw_reg_pmpe_module_get(pmpe_pl);
struct mlxsw_env *mlxsw_env = priv;
if (WARN_ON_ONCE(module >= mlxsw_env->max_module_count ||
slot_index >= mlxsw_env->num_of_slots))
return;
module_status = mlxsw_reg_pmpe_module_status_get(pmpe_pl);
if (module_status != MLXSW_REG_PMPE_MODULE_STATUS_PLUGGED_ENABLED)
return;
event = kmalloc(sizeof(*event), GFP_ATOMIC);
if (!event)
return;
event->mlxsw_env = mlxsw_env;
event->slot_index = slot_index;
event->module = module;
INIT_WORK(&event->work, mlxsw_env_pmpe_event_work);
mlxsw_core_schedule_work(&event->work);
}
static const struct mlxsw_listener mlxsw_env_module_plug_listener =
MLXSW_CORE_EVENTL(mlxsw_env_pmpe_listener_func, PMPE);
static int
mlxsw_env_module_plug_event_register(struct mlxsw_core *mlxsw_core)
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(mlxsw_core);
return mlxsw_core_trap_register(mlxsw_core,
&mlxsw_env_module_plug_listener,
mlxsw_env);
}
static void
mlxsw_env_module_plug_event_unregister(struct mlxsw_env *mlxsw_env)
{
mlxsw_core_trap_unregister(mlxsw_env->core,
&mlxsw_env_module_plug_listener,
mlxsw_env);
}
static int
mlxsw_env_module_oper_state_event_enable(struct mlxsw_core *mlxsw_core,
u8 slot_index)
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(mlxsw_core);
int i, err;
for (i = 0; i < mlxsw_env->line_cards[slot_index]->module_count; i++) {
char pmaos_pl[MLXSW_REG_PMAOS_LEN];
mlxsw_reg_pmaos_pack(pmaos_pl, slot_index, i);
mlxsw_reg_pmaos_e_set(pmaos_pl,
MLXSW_REG_PMAOS_E_GENERATE_EVENT);
mlxsw_reg_pmaos_ee_set(pmaos_pl, true);
err = mlxsw_reg_write(mlxsw_core, MLXSW_REG(pmaos), pmaos_pl);
if (err)
return err;
}
return 0;
}
int
mlxsw_env_module_overheat_counter_get(struct mlxsw_core *mlxsw_core, u8 slot_index,
u8 module, u64 *p_counter)
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(mlxsw_core);
struct mlxsw_env_module_info *module_info;
mutex_lock(&mlxsw_env->line_cards_lock);
module_info = mlxsw_env_module_info_get(mlxsw_core, slot_index, module);
*p_counter = module_info->module_overheat_counter;
mutex_unlock(&mlxsw_env->line_cards_lock);
return 0;
}
EXPORT_SYMBOL(mlxsw_env_module_overheat_counter_get);
void mlxsw_env_module_port_map(struct mlxsw_core *mlxsw_core, u8 slot_index,
u8 module)
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(mlxsw_core);
struct mlxsw_env_module_info *module_info;
mutex_lock(&mlxsw_env->line_cards_lock);
module_info = mlxsw_env_module_info_get(mlxsw_core, slot_index, module);
module_info->num_ports_mapped++;
mutex_unlock(&mlxsw_env->line_cards_lock);
}
EXPORT_SYMBOL(mlxsw_env_module_port_map);
void mlxsw_env_module_port_unmap(struct mlxsw_core *mlxsw_core, u8 slot_index,
u8 module)
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(mlxsw_core);
struct mlxsw_env_module_info *module_info;
mutex_lock(&mlxsw_env->line_cards_lock);
module_info = mlxsw_env_module_info_get(mlxsw_core, slot_index, module);
module_info->num_ports_mapped--;
mutex_unlock(&mlxsw_env->line_cards_lock);
}
EXPORT_SYMBOL(mlxsw_env_module_port_unmap);
int mlxsw_env_module_port_up(struct mlxsw_core *mlxsw_core, u8 slot_index,
u8 module)
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(mlxsw_core);
struct mlxsw_env_module_info *module_info;
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
int err = 0;
mutex_lock(&mlxsw_env->line_cards_lock);
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
module_info = mlxsw_env_module_info_get(mlxsw_core, slot_index, module);
if (module_info->power_mode_policy !=
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
ETHTOOL_MODULE_POWER_MODE_POLICY_AUTO)
goto out_inc;
if (module_info->num_ports_up != 0)
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
goto out_inc;
/* Transition to high power mode following first port using the module
* being put administratively up.
*/
err = __mlxsw_env_set_module_power_mode(mlxsw_core, slot_index, module,
false, NULL);
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
if (err)
goto out_unlock;
out_inc:
module_info->num_ports_up++;
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
out_unlock:
mutex_unlock(&mlxsw_env->line_cards_lock);
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
return err;
}
EXPORT_SYMBOL(mlxsw_env_module_port_up);
void mlxsw_env_module_port_down(struct mlxsw_core *mlxsw_core, u8 slot_index,
u8 module)
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(mlxsw_core);
struct mlxsw_env_module_info *module_info;
mutex_lock(&mlxsw_env->line_cards_lock);
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
module_info = mlxsw_env_module_info_get(mlxsw_core, slot_index, module);
module_info->num_ports_up--;
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
if (module_info->power_mode_policy !=
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
ETHTOOL_MODULE_POWER_MODE_POLICY_AUTO)
goto out_unlock;
if (module_info->num_ports_up != 0)
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
goto out_unlock;
/* Transition to low power mode following last port using the module
* being put administratively down.
*/
__mlxsw_env_set_module_power_mode(mlxsw_core, slot_index, module, true,
NULL);
mlxsw: Add ability to control transceiver modules' power mode Implement support for ethtool_ops::.get_module_power_mode and ethtool_ops::set_module_power_mode. The get operation is implemented using the Management Cable IO and Notifications (MCION) register that reports the operational power mode of the module and its presence. In case a module is not present, its operational power mode is not reported to ethtool and user space. If not set before, the power mode policy is reported as "high", which is the default on Mellanox systems. The set operation is implemented using the Port Module Memory Map Properties (PMMP) register. The register instructs the device's firmware to transition a plugged-in module to / out of low power mode by writing to its memory map. When the power mode policy is set to 'auto', a module will not transition to low power mode as long as any ports using it are administratively up. Example: # devlink port split swp11 count 4 # ethtool --set-module swp11s0 power-mode-policy auto $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low # ip link set dev swp11s0 up # ip link set dev swp11s1 up $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s1 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode high # ip link set dev swp11s0 down $ ethtool --show-module swp11s0 Module parameters for swp11s0: power-mode-policy auto power-mode low Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2021-10-06 10:46:45 +00:00
out_unlock:
mutex_unlock(&mlxsw_env->line_cards_lock);
}
EXPORT_SYMBOL(mlxsw_env_module_port_down);
static int mlxsw_env_line_cards_alloc(struct mlxsw_env *env)
{
struct mlxsw_env_module_info *module_info;
int i, j;
for (i = 0; i < env->num_of_slots; i++) {
env->line_cards[i] = kzalloc(struct_size(env->line_cards[i],
module_info,
env->max_module_count),
GFP_KERNEL);
if (!env->line_cards[i])
goto kzalloc_err;
/* Firmware defaults to high power mode policy where modules
* are transitioned to high power mode following plug-in.
*/
for (j = 0; j < env->max_module_count; j++) {
module_info = &env->line_cards[i]->module_info[j];
module_info->power_mode_policy =
ETHTOOL_MODULE_POWER_MODE_POLICY_HIGH;
}
}
return 0;
kzalloc_err:
for (i--; i >= 0; i--)
kfree(env->line_cards[i]);
return -ENOMEM;
}
static void mlxsw_env_line_cards_free(struct mlxsw_env *env)
{
int i = env->num_of_slots;
for (i--; i >= 0; i--)
kfree(env->line_cards[i]);
}
static int
mlxsw_env_module_event_enable(struct mlxsw_env *mlxsw_env, u8 slot_index)
{
int err;
err = mlxsw_env_module_oper_state_event_enable(mlxsw_env->core,
slot_index);
if (err)
return err;
err = mlxsw_env_module_temp_event_enable(mlxsw_env->core, slot_index);
if (err)
return err;
return 0;
}
static void
mlxsw_env_module_event_disable(struct mlxsw_env *mlxsw_env, u8 slot_index)
{
}
static int
mlxsw_env_module_type_set(struct mlxsw_core *mlxsw_core, u8 slot_index)
{
struct mlxsw_env *mlxsw_env = mlxsw_core_env(mlxsw_core);
int i;
for (i = 0; i < mlxsw_env->line_cards[slot_index]->module_count; i++) {
struct mlxsw_env_module_info *module_info;
char pmtm_pl[MLXSW_REG_PMTM_LEN];
int err;
mlxsw_reg_pmtm_pack(pmtm_pl, slot_index, i);
err = mlxsw_reg_query(mlxsw_core, MLXSW_REG(pmtm), pmtm_pl);
if (err)
return err;
module_info = mlxsw_env_module_info_get(mlxsw_core, slot_index,
i);
module_info->type = mlxsw_reg_pmtm_module_type_get(pmtm_pl);
}
return 0;
}
static void
mlxsw_env_linecard_modules_power_mode_apply(struct mlxsw_core *mlxsw_core,
struct mlxsw_env *env,
u8 slot_index)
{
int i;
for (i = 0; i < env->line_cards[slot_index]->module_count; i++) {
enum ethtool_module_power_mode_policy policy;
struct mlxsw_env_module_info *module_info;
struct netlink_ext_ack extack;
int err;
module_info = &env->line_cards[slot_index]->module_info[i];
policy = module_info->power_mode_policy;
err = mlxsw_env_set_module_power_mode_apply(mlxsw_core,
slot_index, i,
policy, &extack);
if (err)
dev_err(env->bus_info->dev, "%s\n", extack._msg);
}
}
static void
mlxsw_env_got_active(struct mlxsw_core *mlxsw_core, u8 slot_index, void *priv)
{
struct mlxsw_env *mlxsw_env = priv;
char mgpir_pl[MLXSW_REG_MGPIR_LEN];
int err;
mutex_lock(&mlxsw_env->line_cards_lock);
if (__mlxsw_env_linecard_is_active(mlxsw_env, slot_index))
goto out_unlock;
mlxsw_reg_mgpir_pack(mgpir_pl, slot_index);
err = mlxsw_reg_query(mlxsw_env->core, MLXSW_REG(mgpir), mgpir_pl);
if (err)
goto out_unlock;
mlxsw_reg_mgpir_unpack(mgpir_pl, NULL, NULL, NULL,
&mlxsw_env->line_cards[slot_index]->module_count,
NULL);
err = mlxsw_env_module_event_enable(mlxsw_env, slot_index);
if (err) {
dev_err(mlxsw_env->bus_info->dev, "Failed to enable port module events for line card in slot %d\n",
slot_index);
goto err_mlxsw_env_module_event_enable;
}
err = mlxsw_env_module_type_set(mlxsw_env->core, slot_index);
if (err) {
dev_err(mlxsw_env->bus_info->dev, "Failed to set modules' type for line card in slot %d\n",
slot_index);
goto err_type_set;
}
mlxsw_env->line_cards[slot_index]->active = true;
/* Apply power mode policy. */
mlxsw_env_linecard_modules_power_mode_apply(mlxsw_core, mlxsw_env,
slot_index);
mutex_unlock(&mlxsw_env->line_cards_lock);
return;
err_type_set:
mlxsw_env_module_event_disable(mlxsw_env, slot_index);
err_mlxsw_env_module_event_enable:
out_unlock:
mutex_unlock(&mlxsw_env->line_cards_lock);
}
static void
mlxsw_env_got_inactive(struct mlxsw_core *mlxsw_core, u8 slot_index,
void *priv)
{
struct mlxsw_env *mlxsw_env = priv;
mutex_lock(&mlxsw_env->line_cards_lock);
if (!__mlxsw_env_linecard_is_active(mlxsw_env, slot_index))
goto out_unlock;
mlxsw_env->line_cards[slot_index]->active = false;
mlxsw_env_module_event_disable(mlxsw_env, slot_index);
mlxsw_env->line_cards[slot_index]->module_count = 0;
out_unlock:
mutex_unlock(&mlxsw_env->line_cards_lock);
}
static struct mlxsw_linecards_event_ops mlxsw_env_event_ops = {
.got_active = mlxsw_env_got_active,
.got_inactive = mlxsw_env_got_inactive,
};
static void mlxsw_env_max_module_eeprom_len_query(struct mlxsw_env *mlxsw_env)
{
char mcam_pl[MLXSW_REG_MCAM_LEN];
bool mcia_128b_supported = false;
int err;
mlxsw_reg_mcam_pack(mcam_pl,
MLXSW_REG_MCAM_FEATURE_GROUP_ENHANCED_FEATURES);
err = mlxsw_reg_query(mlxsw_env->core, MLXSW_REG(mcam), mcam_pl);
if (!err)
mlxsw_reg_mcam_unpack(mcam_pl, MLXSW_REG_MCAM_MCIA_128B,
&mcia_128b_supported);
mlxsw_env->max_eeprom_len = mcia_128b_supported ? 128 : 48;
}
int mlxsw_env_init(struct mlxsw_core *mlxsw_core,
const struct mlxsw_bus_info *bus_info,
struct mlxsw_env **p_env)
{
u8 module_count, num_of_slots, max_module_count;
char mgpir_pl[MLXSW_REG_MGPIR_LEN];
struct mlxsw_env *env;
int err;
mlxsw_reg_mgpir_pack(mgpir_pl, 0);
err = mlxsw_reg_query(mlxsw_core, MLXSW_REG(mgpir), mgpir_pl);
if (err)
return err;
mlxsw_reg_mgpir_unpack(mgpir_pl, NULL, NULL, NULL, &module_count,
&num_of_slots);
/* If the system is modular, get the maximum number of modules per-slot.
* Otherwise, get the maximum number of modules on the main board.
*/
max_module_count = num_of_slots ?
mlxsw_reg_mgpir_max_modules_per_slot_get(mgpir_pl) :
module_count;
env = kzalloc(struct_size(env, line_cards, num_of_slots + 1),
GFP_KERNEL);
if (!env)
return -ENOMEM;
env->core = mlxsw_core;
env->bus_info = bus_info;
env->num_of_slots = num_of_slots + 1;
env->max_module_count = max_module_count;
err = mlxsw_env_line_cards_alloc(env);
if (err)
goto err_mlxsw_env_line_cards_alloc;
mutex_init(&env->line_cards_lock);
*p_env = env;
err = mlxsw_linecards_event_ops_register(env->core,
&mlxsw_env_event_ops, env);
if (err)
goto err_linecards_event_ops_register;
err = mlxsw_env_temp_warn_event_register(mlxsw_core);
if (err)
goto err_temp_warn_event_register;
err = mlxsw_env_module_plug_event_register(mlxsw_core);
if (err)
goto err_module_plug_event_register;
/* Set 'module_count' only for main board. Actual count for line card
* is to be set after line card is activated.
*/
env->line_cards[0]->module_count = num_of_slots ? 0 : module_count;
/* Enable events only for main board. Line card events are to be
* configured only after line card is activated. Before that, access to
* modules on line cards is not allowed.
*/
err = mlxsw_env_module_event_enable(env, 0);
if (err)
goto err_mlxsw_env_module_event_enable;
err = mlxsw_env_module_type_set(mlxsw_core, 0);
if (err)
goto err_type_set;
mlxsw_env_max_module_eeprom_len_query(env);
env->line_cards[0]->active = true;
return 0;
err_type_set:
mlxsw_env_module_event_disable(env, 0);
err_mlxsw_env_module_event_enable:
mlxsw_env_module_plug_event_unregister(env);
err_module_plug_event_register:
mlxsw_env_temp_warn_event_unregister(env);
err_temp_warn_event_register:
mlxsw_linecards_event_ops_unregister(env->core,
&mlxsw_env_event_ops, env);
err_linecards_event_ops_register:
mutex_destroy(&env->line_cards_lock);
mlxsw_env_line_cards_free(env);
err_mlxsw_env_line_cards_alloc:
kfree(env);
return err;
}
void mlxsw_env_fini(struct mlxsw_env *env)
{
env->line_cards[0]->active = false;
mlxsw_env_module_event_disable(env, 0);
mlxsw_env_module_plug_event_unregister(env);
/* Make sure there is no more event work scheduled. */
mlxsw_core_flush_owq();
mlxsw_env_temp_warn_event_unregister(env);
mlxsw_linecards_event_ops_unregister(env->core,
&mlxsw_env_event_ops, env);
mutex_destroy(&env->line_cards_lock);
mlxsw_env_line_cards_free(env);
kfree(env);
}