On Tue, Sep 15, 2020 at 10:22:32PM -0700, Guenter Roeck wrote: > On 9/15/20 8:14 PM, Xu Yilun wrote: > > This patch adds hwmon functionality for Intel MAX 10 BMC chip. This BMC > > chip connects to a set of sensor chips to monitor current, voltage, > > thermal and power of different components on board. The BMC firmware is > > responsible for sensor data sampling and recording in shared registers. > > Host driver reads the sensor data from these shared registers and > > exposes them to users as hwmon interfaces. > > > > Signed-off-by: Xu Yilun <yilun.xu@xxxxxxxxx> > > Signed-off-by: Wu Hao <hao.wu@xxxxxxxxx> > > Signed-off-by: Matthew Gerlach <matthew.gerlach@xxxxxxxxxxxxxxx> > > Signed-off-by: Tom Rix <trix@xxxxxxxxxx> > > Is that really the Sign-off path, or are some of those reviewers ? Just wondering. The original version of the product driver is initiated by Hao. I made this upstream version. And during our internal procedures for this upstream version, Matthew & Tom had comments and also sent fix patches for it. So I added Signed-off for them. After reading Documentation/process/submitting-patches.rst, my understanding is that if people contribute the whole or part of the code, a Signed-off-by could be added. Is that right? > > > --- > > v2: add the Documentation > > refactor the code, provide static hwmon_channel_info > > remove Unnecessary hwmon-sysfs.h > > make the sensor data table const > > --- > > Documentation/hwmon/index.rst | 1 + > > Documentation/hwmon/intel-m10-bmc-hwmon.rst | 78 ++++++ > > drivers/hwmon/Kconfig | 11 + > > drivers/hwmon/Makefile | 1 + > > drivers/hwmon/intel-m10-bmc-hwmon.c | 355 ++++++++++++++++++++++++++++ > > 5 files changed, 446 insertions(+) > > create mode 100644 Documentation/hwmon/intel-m10-bmc-hwmon.rst > > create mode 100644 drivers/hwmon/intel-m10-bmc-hwmon.c > > > > diff --git a/Documentation/hwmon/index.rst b/Documentation/hwmon/index.rst > > index a926f1a..4bcb1a7 100644 > > --- a/Documentation/hwmon/index.rst > > +++ b/Documentation/hwmon/index.rst > > @@ -74,6 +74,7 @@ Hardware Monitoring Kernel Drivers > > ina209 > > ina2xx > > ina3221 > > + intel-m10-bmc-hwmon > > ir35221 > > ir38064 > > isl68137 > > diff --git a/Documentation/hwmon/intel-m10-bmc-hwmon.rst b/Documentation/hwmon/intel-m10-bmc-hwmon.rst > > new file mode 100644 > > index 0000000..3d148c6 > > --- /dev/null > > +++ b/Documentation/hwmon/intel-m10-bmc-hwmon.rst > > @@ -0,0 +1,78 @@ > > +.. SPDX-License-Identifier: GPL-2.0 > > + > > +Kernel driver intel-m10-bmc-hwmon > > +================================= > > + > > +Supported chips: > > + > > + * Intel MAX 10 BMC for Intel PAC N3000 > > + > > + Prefix: 'n3000bmc-hwmon' > > + > > +Author: Xu Yilun <yilun.xu@xxxxxxxxx> > > + > > + > > +Description > > +----------- > > + > > +This driver adds the temperature, voltage, current and power reading > > +support for the Intel MAX 10 Board Management Controller (BMC) chip. > > +The BMC chip is integrated in some Intel Programmable Acceleration > > +Cards (PAC). It connects to a set of sensor chips to monitor the > > +sensor data of different components on the board. The BMC firmware is > > +responsible for sensor data sampling and recording in shared > > +registers. The host driver reads the sensor data from these shared > > +registers and exposes them to users as hwmon interfaces. > > + > > +The BMC chip is implemented using the Intel MAX 10 CPLD. It could be > > +reprogramed to some variants in order to support different Intel > > +PACs. The driver is designed to be able to distinguish between the > > +variants, but now it only supports the BMC for Intel PAC N3000. > > + > > + > > +Sysfs attributes > > +---------------- > > + > > +The following attributes are supported: > > + > > +- Intel MAX 10 BMC for Intel PAC N3000: > > + > > +======================= ======================================================= > > +tempX_input Temperature of the component (specified by tempX_label) > > +tempX_max Temperature maximum setpoint of the component > > +tempX_crit Temperature critical setpoint of the component > > +tempX_max_hyst Hysteresis for temperature maximum of the component > > +tempX_crit_hyst Hysteresis for temperature critical of the component > > +temp1_label "Board Temperature" > > +temp2_label "FPGA Die Temperature" > > +temp3_label "QSFP0 Temperature" > > +temp4_label "QSFP1 Temperature" > > +temp5_label "Retimer A Temperature" > > +temp6_label "Retimer A SerDes Temperature" > > +temp7_label "Retimer B Temperature" > > +temp8_label "Retimer B SerDes Temperature" > > + > > +inX_input Measured voltage of the component (specified by > > + inX_label) > > +in0_label "QSFP0 Supply Voltage" > > +in1_label "QSFP1 Supply Voltage" > > +in2_label "FPGA Core Voltage" > > +in3_label "12V Backplane Voltage" > > +in4_label "1.2V Voltage" > > +in5_label "12V AUX Voltage" > > +in6_label "1.8V Voltage" > > +in7_label "3.3V Voltage" > > + > > +currX_input Measured current of the component (specified by > > + currX_label) > > +curr1_label "FPGA Core Current" > > +curr2_label "12V Backplane Current" > > +curr3_label "12V AUX Current" > > + > > +powerX_input Measured power of the component (specified by > > + powerX_label) > > +power1_label "Board Power" > > + > > +======================= ======================================================= > > + > > +All the attributes are read-only. > > diff --git a/drivers/hwmon/Kconfig b/drivers/hwmon/Kconfig > > index 8dc28b2..53af15c 100644 > > --- a/drivers/hwmon/Kconfig > > +++ b/drivers/hwmon/Kconfig > > @@ -2064,6 +2064,17 @@ config SENSORS_XGENE > > If you say yes here you get support for the temperature > > and power sensors for APM X-Gene SoC. > > > > +config SENSORS_INTEL_M10_BMC_HWMON > > + tristate "Intel MAX10 BMC Hardware Monitoring" > > + depends on MFD_INTEL_M10_BMC > > + help > > + This driver provides support for the hardware monitoring functionality > > + on Intel MAX10 BMC chip. > > + > > + This BMC Chip is used on Intel FPGA PCIe Acceleration Cards (PAC). Its > > + sensors monitor various telemetry data of different components on the > > + card, e.g. board temperature, FPGA core temperature/voltage/current. > > + > > if ACPI > > > > comment "ACPI drivers" > > diff --git a/drivers/hwmon/Makefile b/drivers/hwmon/Makefile > > index a8f4b35..ba5a25a 100644 > > --- a/drivers/hwmon/Makefile > > +++ b/drivers/hwmon/Makefile > > @@ -90,6 +90,7 @@ obj-$(CONFIG_SENSORS_IIO_HWMON) += iio_hwmon.o > > obj-$(CONFIG_SENSORS_INA209) += ina209.o > > obj-$(CONFIG_SENSORS_INA2XX) += ina2xx.o > > obj-$(CONFIG_SENSORS_INA3221) += ina3221.o > > +obj-$(CONFIG_SENSORS_INTEL_M10_BMC_HWMON) += intel-m10-bmc-hwmon.o > > obj-$(CONFIG_SENSORS_IT87) += it87.o > > obj-$(CONFIG_SENSORS_JC42) += jc42.o > > obj-$(CONFIG_SENSORS_K8TEMP) += k8temp.o > > diff --git a/drivers/hwmon/intel-m10-bmc-hwmon.c b/drivers/hwmon/intel-m10-bmc-hwmon.c > > new file mode 100644 > > index 0000000..ce73545 > > --- /dev/null > > +++ b/drivers/hwmon/intel-m10-bmc-hwmon.c > > @@ -0,0 +1,355 @@ > > +// SPDX-License-Identifier: GPL-2.0 > > +/* > > + * Intel MAX 10 BMC HWMON Driver > > + * > > + * Copyright (C) 2018-2020 Intel Corporation. All rights reserved. > > + * > > + */ > > +#include <linux/device.h> > > +#include <linux/hwmon.h> > > +#include <linux/mfd/intel-m10-bmc.h> > > +#include <linux/module.h> > > +#include <linux/mod_devicetable.h> > > +#include <linux/platform_device.h> > > + > > +struct m10bmc_sdata { > > + unsigned int reg_input; > > + unsigned int reg_max; > > + unsigned int reg_crit; > > + unsigned int reg_hyst; > > + unsigned int reg_min; > > + unsigned int multiplier; > > + const char *label; > > +}; > > + > > +struct m10bmc_hwmon_board_data { > > + const struct m10bmc_sdata *temp; > > + const struct m10bmc_sdata *in; > > + const struct m10bmc_sdata *curr; > > + const struct m10bmc_sdata *power; > > + const struct hwmon_channel_info **hinfo; > > +}; > > + > > +struct m10bmc_hwmon { > > + struct device *dev; > > + struct hwmon_chip_info chip; > > + char *hw_name; > > + struct intel_m10bmc *m10bmc; > > + struct m10bmc_hwmon_board_data *bdata; > > +}; > > + > > +static const struct m10bmc_sdata n3000bmc_temp_tbl[] = { > > + { 0x100, 0x104, 0x108, 0x10c, 0x0, 500, "Board Temperature" }, > > + { 0x110, 0x114, 0x118, 0x0, 0x0, 500, "FPGA Die Temperature" }, > > + { 0x11c, 0x124, 0x120, 0x0, 0x0, 500, "QSFP0 Temperature" }, > > + { 0x12c, 0x134, 0x130, 0x0, 0x0, 500, "QSFP1 Temperature" }, > > + { 0x168, 0x0, 0x0, 0x0, 0x0, 500, "Retimer A Temperature" }, > > + { 0x16c, 0x0, 0x0, 0x0, 0x0, 500, "Retimer A SerDes Temperature" }, > > + { 0x170, 0x0, 0x0, 0x0, 0x0, 500, "Retimer B Temperature" }, > > + { 0x174, 0x0, 0x0, 0x0, 0x0, 500, "Retimer B SerDes Temperature" }, > > +}; > > + > > +static const struct m10bmc_sdata n3000bmc_in_tbl[] = { > > + { 0x128, 0x0, 0x0, 0x0, 0x0, 1, "QSFP0 Supply Voltage" }, > > + { 0x138, 0x0, 0x0, 0x0, 0x0, 1, "QSFP1 Supply Voltage" }, > > + { 0x13c, 0x0, 0x0, 0x0, 0x0, 1, "FPGA Core Voltage" }, > > + { 0x144, 0x0, 0x0, 0x0, 0x0, 1, "12V Backplane Voltage" }, > > + { 0x14c, 0x0, 0x0, 0x0, 0x0, 1, "1.2V Voltage" }, > > + { 0x150, 0x0, 0x0, 0x0, 0x0, 1, "12V AUX Voltage" }, > > + { 0x158, 0x0, 0x0, 0x0, 0x0, 1, "1.8V Voltage" }, > > + { 0x15c, 0x0, 0x0, 0x0, 0x0, 1, "3.3V Voltage" }, > > +}; > > + > > +static const struct m10bmc_sdata n3000bmc_curr_tbl[] = { > > + { 0x140, 0x0, 0x0, 0x0, 0x0, 1, "FPGA Core Current" }, > > + { 0x148, 0x0, 0x0, 0x0, 0x0, 1, "12V Backplane Current" }, > > + { 0x154, 0x0, 0x0, 0x0, 0x0, 1, "12V AUX Current" }, > > +}; > > + > > +static const struct m10bmc_sdata n3000bmc_power_tbl[] = { > > + { 0x160, 0x0, 0x0, 0x0, 0x0, 1000, "Board Power" },> +}; > > + > > +static const struct hwmon_channel_info *n3000bmc_hinfo[] = { > > + HWMON_CHANNEL_INFO(temp, > > + HWMON_T_INPUT | HWMON_T_MAX | HWMON_T_MAX_HYST | > > + HWMON_T_CRIT | HWMON_T_CRIT_HYST | HWMON_T_LABEL, > > + HWMON_T_INPUT | HWMON_T_MAX | HWMON_T_CRIT | > > + HWMON_T_LABEL, > > + HWMON_T_INPUT | HWMON_T_MAX | HWMON_T_CRIT | > > + HWMON_T_LABEL, > > + HWMON_T_INPUT | HWMON_T_MAX | HWMON_T_CRIT | > > + HWMON_T_LABEL, > > + HWMON_T_INPUT | HWMON_T_LABEL, > > + HWMON_T_INPUT | HWMON_T_LABEL, > > + HWMON_T_INPUT | HWMON_T_LABEL, > > + HWMON_T_INPUT | HWMON_T_LABEL), > > + HWMON_CHANNEL_INFO(in, > > + HWMON_I_INPUT | HWMON_I_LABEL, > > + HWMON_I_INPUT | HWMON_I_LABEL, > > + HWMON_I_INPUT | HWMON_I_LABEL, > > + HWMON_I_INPUT | HWMON_I_LABEL, > > + HWMON_I_INPUT | HWMON_I_LABEL, > > + HWMON_I_INPUT | HWMON_I_LABEL, > > + HWMON_I_INPUT | HWMON_I_LABEL, > > + HWMON_I_INPUT | HWMON_I_LABEL), > > + HWMON_CHANNEL_INFO(curr, > > + HWMON_C_INPUT | HWMON_C_LABEL, > > + HWMON_C_INPUT | HWMON_C_LABEL, > > + HWMON_C_INPUT | HWMON_C_LABEL), > > + HWMON_CHANNEL_INFO(power, > > + HWMON_P_INPUT | HWMON_P_LABEL), > > + NULL > > +}; > > + > > +struct m10bmc_hwmon_board_data n3000bmc_hwmon_bdata = { > > static, and probably const Yes. > > > + .temp = n3000bmc_temp_tbl, > > + .in = n3000bmc_in_tbl, > > + .curr = n3000bmc_curr_tbl, > > + .power = n3000bmc_power_tbl, > > I would suggest to declare > const struct m10bmc_sdata *tables[hwmon_max]; > and initialize with > .tables[hwmon_temp] = n3000bmc_temp_tbl, > .tables[hwmon_in] = n3000bmc_in_tbl, > and so on. That makes the structure a bit larger, but simplifies > the operational code (see below). Yes my concern is also the data size. But since it simplifies the code I'll follow the change. > > > + .hinfo = n3000bmc_hinfo, > > +}; > > + > > +static umode_t > > +m10bmc_hwmon_is_visible(const void *data, enum hwmon_sensor_types type, > > + u32 attr, int channel) > > +{ > > + return 0444; > > +} > > + > > +static const struct m10bmc_sdata * > > +find_sensor_data(struct m10bmc_hwmon *hw, enum hwmon_sensor_types type, > > + int channel) > > +{ > > + const struct m10bmc_sdata *tbl; > > + > > + switch (type) { > > + case hwmon_temp: > > + tbl = hw->bdata->temp; > > + break; > > + case hwmon_in: > > + tbl = hw->bdata->in; > > + break; > > + case hwmon_curr: > > + tbl = hw->bdata->curr; > > + break; > > + case hwmon_power: > > + tbl = hw->bdata->power; > > + break; > > + default: > > + return ERR_PTR(-EOPNOTSUPP); > > + } > > Then you can use > tbl = hw->bdata->tables[type]; Yes. > > > + > > + if (!tbl) > > + return ERR_PTR(-EOPNOTSUPP); > > + > > + return &tbl[channel]; > > +} > > + > > +static int do_sensor_read(struct m10bmc_hwmon *hw, > > + const struct m10bmc_sdata *data, > > + unsigned int regoff, long *val) > > +{ > > + unsigned int regval; > > + int ret; > > + > > + ret = m10bmc_sys_read(hw->m10bmc, regoff, ®val); > > + if (ret) > > + return ret; > > + > > + /* > > + * BMC Firmware will return 0xdeadbeef if the sensor value is invalid > > + * at that time. This usually happens on sensor channels which connect > > + * to external pluggable modules, e.g. QSFP temperature and voltage. > > + * When the QSFP is unplugged from cage, driver will get 0xdeadbeef > > + * from their registers. > > + */ > > + if (regval == 0xdeadbeef) > > + return -EBUSY; > > Is this appropriate ? The description above would suggest differently. > -ENODATA, maybe ? Also, are those module hot pluggable ? If not, I could change to -ENODATA. Those modules are hot pluggable. QSFP is a compact, hot-pluggable optical module for ethernet connection. The real sensors are embedded in QSFP modules, not on the cages of the board. So at running time, if the QSFP is plugged in the cage, we have valid sensor data, if it is unplugged, we have 0xdeadbeef. > it may be more appropriate to detect the status in the in_visible function > and not create affected attributes in the first place. > > > + > > + *val = regval * data->multiplier; > > + > > + return 0; > > +} > > + > > +static int m10bmc_hwmon_read(struct device *dev, enum hwmon_sensor_types type, > > + u32 attr, int channel, long *val) > > +{ > > + struct m10bmc_hwmon *hw = dev_get_drvdata(dev); > > + unsigned int reg = 0, reg_hyst = 0; > > + const struct m10bmc_sdata *data; > > + long hyst, value; > > + int ret; > > + > > + data = find_sensor_data(hw, type, channel); > > + if (IS_ERR(data)) > > + return PTR_ERR(data); > > + > > + switch (type) { > > + case hwmon_temp: > > + switch (attr) { > > + case hwmon_temp_input: > > + reg = data->reg_input; > > + break; > > + case hwmon_temp_max_hyst: > > + reg_hyst = data->reg_hyst; > > + fallthrough; > > + case hwmon_temp_max: > > + reg = data->reg_max; > > + break; > > + case hwmon_temp_crit_hyst: > > + reg_hyst = data->reg_hyst; > > + fallthrough; > > + case hwmon_temp_crit: > > + reg = data->reg_crit; > > + break; > > + default: > > + return -EOPNOTSUPP; > > + } > > + break; > > + case hwmon_in: > > + switch (attr) { > > + case hwmon_in_input: > > + reg = data->reg_input; > > + break; > > + case hwmon_in_max: > > + reg = data->reg_max; > > + break; > > + case hwmon_in_crit: > > + reg = data->reg_crit; > > + break; > > + case hwmon_in_min: > > + reg = data->reg_min; > > + break; > > + default: > > + return -EOPNOTSUPP; > > + } > > + break; > > + case hwmon_curr: > > + switch (attr) { > > + case hwmon_curr_input: > > + reg = data->reg_input; > > + break; > > + case hwmon_curr_max: > > + reg = data->reg_max; > > + break; > > + case hwmon_curr_crit: > > + reg = data->reg_crit; > > + break; > > + default: > > + return -EOPNOTSUPP; > > + } > > + break; > > + case hwmon_power: > > + switch (attr) { > > + case hwmon_power_input: > > + reg = data->reg_input; > > + break; > > + default: > > + return -EOPNOTSUPP; > > + } > > + break; > > + default: > > + return -EOPNOTSUPP; > > + } > > + > > + if (!reg) > > + return -EOPNOTSUPP; > > + > > That can't happen with the code above. reg is always initialized. > All other cases already returned -EOPNOTSUPP. In the n3000bmc_xxx_tbl, we mark the reg as 0 if the attr is not supported, this should be aligned with the hwmon_channel_info. And it has to be ensured manually. If everything is fine, this can't happen. But if developers made mistake when inputing data, e.g. the HWMON_X_XXX bit is set but reg in table is 0, then we may got confusing value. This is the intention I added the check. > > > + ret = do_sensor_read(hw, data, reg, &value); > > + if (ret) > > + return ret; > > + > > + if (reg_hyst) { > > + ret = do_sensor_read(hw, data, reg_hyst, &hyst); > > + if (ret) > > + return ret; > > + > > + value -= hyst; > > Is that also correct for _min attributes ? Normally I'd assume that the hysteresis > for those is larger than the base attribute value. Mm.. the hysteresis attr value should be larger than min attr. But now the code expose no _min_hyst attr. So is it OK now? > > > + } > > + > > + *val = value; > > + > > + return ret; > > ret is always 0 here -> return 0; I'll change it. > > > +} > > + > > +static int m10bmc_hwmon_read_string(struct device *dev, > > + enum hwmon_sensor_types type, > > + u32 attr, int channel, const char **str) > > +{ > > + struct m10bmc_hwmon *hw = dev_get_drvdata(dev); > > + const struct m10bmc_sdata *data; > > + > > + data = find_sensor_data(hw, type, channel); > > + if (IS_ERR(data)) > > + return PTR_ERR(data); > > + > > + *str = data->label; > > + > > + return 0; > > +} > > + > > +static const struct hwmon_ops m10bmc_hwmon_ops = { > > + .is_visible = m10bmc_hwmon_is_visible, > > + .read = m10bmc_hwmon_read, > > + .read_string = m10bmc_hwmon_read_string, > > +}; > > + > > +static int m10bmc_hwmon_probe(struct platform_device *pdev) > > +{ > > + const struct platform_device_id *id = platform_get_device_id(pdev); > > + struct intel_m10bmc *m10bmc = dev_get_drvdata(pdev->dev.parent); > > + struct device *hwmon_dev, *dev = &pdev->dev; > > + struct m10bmc_hwmon *hw; > > + int i; > > + > > + if (!id || !id->driver_data) { > > That can not really happen. I see. We will not fall-back to driver name match if pdrv->id_table is provided. Thanks, Yilun > > > + dev_err(dev, "Failed to get board data\n"); > > + return -ENODEV; > > + } > > + > > + hw = devm_kzalloc(dev, sizeof(*hw), GFP_KERNEL); > > + if (!hw) > > + return -ENOMEM; > > + > > + hw->dev = dev; > > + hw->m10bmc = m10bmc; > > + hw->bdata = (struct m10bmc_hwmon_board_data *)id->driver_data; > > + > > + hw->chip.info = hw->bdata->hinfo; > > + hw->chip.ops = &m10bmc_hwmon_ops; > > + > > + hw->hw_name = devm_kstrdup(dev, id->name, GFP_KERNEL); > > + if (!hw->hw_name) > > + return -ENOMEM; > > + > > + for (i = 0; hw->hw_name[i]; i++) > > + if (hwmon_is_bad_char(hw->hw_name[i])) > > + hw->hw_name[i] = '_'; > > + > > + hwmon_dev = devm_hwmon_device_register_with_info(dev, hw->hw_name, > > + hw, &hw->chip, NULL); > > + return PTR_ERR_OR_ZERO(hwmon_dev); > > +} > > + > > +static const struct platform_device_id intel_m10bmc_hwmon_ids[] = { > > + { > > + .name = "n3000bmc-hwmon", > > + .driver_data = (unsigned long)&n3000bmc_hwmon_bdata, > > + }, > > + { } > > +}; > > + > > +static struct platform_driver intel_m10bmc_hwmon_driver = { > > + .probe = m10bmc_hwmon_probe, > > + .driver = { > > + .name = "intel-m10-bmc-hwmon", > > + }, > > + .id_table = intel_m10bmc_hwmon_ids, > > +}; > > +module_platform_driver(intel_m10bmc_hwmon_driver); > > + > > +MODULE_DEVICE_TABLE(platform, intel_m10bmc_hwmon_ids); > > +MODULE_AUTHOR("Intel Corporation"); > > +MODULE_DESCRIPTION("Intel MAX 10 BMC hardware monitor"); > > +MODULE_LICENSE("GPL"); > >