diff options
| -rw-r--r-- | Documentation/devicetree/bindings/thermal/rockchip-thermal.txt | 3 | ||||
| -rw-r--r-- | drivers/thermal/Kconfig | 4 | ||||
| -rw-r--r-- | drivers/thermal/db8500_thermal.c | 1 | ||||
| -rw-r--r-- | drivers/thermal/devfreq_cooling.c | 5 | ||||
| -rw-r--r-- | drivers/thermal/int340x_thermal/int3400_thermal.c | 2 | ||||
| -rw-r--r-- | drivers/thermal/intel_pch_thermal.c | 64 | ||||
| -rw-r--r-- | drivers/thermal/max77620_thermal.c | 1 | ||||
| -rw-r--r-- | drivers/thermal/qcom-spmi-temp-alarm.c | 6 | ||||
| -rw-r--r-- | drivers/thermal/rockchip_thermal.c | 7 | ||||
| -rw-r--r-- | drivers/thermal/tango_thermal.c | 1 | ||||
| -rw-r--r-- | drivers/thermal/thermal_hwmon.c | 4 | ||||
| -rw-r--r-- | drivers/thermal/ti-soc-thermal/ti-bandgap.c | 5 | ||||
| -rw-r--r-- | drivers/thermal/x86_pkg_temp_thermal.c | 587 | ||||
| -rw-r--r-- | include/linux/devfreq_cooling.h | 9 |
14 files changed, 314 insertions, 385 deletions
diff --git a/Documentation/devicetree/bindings/thermal/rockchip-thermal.txt b/Documentation/devicetree/bindings/thermal/rockchip-thermal.txt index 08efe6bc2193..43003aec94bd 100644 --- a/Documentation/devicetree/bindings/thermal/rockchip-thermal.txt +++ b/Documentation/devicetree/bindings/thermal/rockchip-thermal.txt @@ -22,10 +22,13 @@ Required properties: TSADC controller. - pinctrl-2 : The "sleep" pinctrl state, it will be in for suspend. - #thermal-sensor-cells : Should be 1. See ./thermal.txt for a description. + +Optional properties: - rockchip,hw-tshut-temp : The hardware-controlled shutdown temperature value. - rockchip,hw-tshut-mode : The hardware-controlled shutdown mode 0:CRU 1:GPIO. - rockchip,hw-tshut-polarity : The hardware-controlled active polarity 0:LOW 1:HIGH. +- rockchip,grf : The phandle of the syscon node for the general register file. Exiample: tsadc: tsadc@ff280000 { diff --git a/drivers/thermal/Kconfig b/drivers/thermal/Kconfig index a13541bdc726..c2c056cc7ea5 100644 --- a/drivers/thermal/Kconfig +++ b/drivers/thermal/Kconfig @@ -177,8 +177,10 @@ config THERMAL_EMULATION config HISI_THERMAL tristate "Hisilicon thermal driver" - depends on (ARCH_HISI && CPU_THERMAL && OF) || COMPILE_TEST + depends on ARCH_HISI || COMPILE_TEST depends on HAS_IOMEM + depends on OF + default y help Enable this to plug hisilicon's thermal sensor driver into the Linux thermal framework. cpufreq is used as the cooling device to throttle diff --git a/drivers/thermal/db8500_thermal.c b/drivers/thermal/db8500_thermal.c index e776cea80cfc..f491faf16592 100644 --- a/drivers/thermal/db8500_thermal.c +++ b/drivers/thermal/db8500_thermal.c @@ -512,6 +512,7 @@ static const struct of_device_id db8500_thermal_match[] = { { .compatible = "stericsson,db8500-thermal" }, {}, }; +MODULE_DEVICE_TABLE(of, db8500_thermal_match); #endif static struct platform_driver db8500_thermal_driver = { diff --git a/drivers/thermal/devfreq_cooling.c b/drivers/thermal/devfreq_cooling.c index 81631b110e17..5a737fd5f1aa 100644 --- a/drivers/thermal/devfreq_cooling.c +++ b/drivers/thermal/devfreq_cooling.c @@ -238,7 +238,7 @@ get_static_power(struct devfreq_cooling_device *dfc, unsigned long freq) return 0; } - return dfc->power_ops->get_static_power(voltage); + return dfc->power_ops->get_static_power(df, voltage); } /** @@ -262,7 +262,8 @@ get_dynamic_power(struct devfreq_cooling_device *dfc, unsigned long freq, struct devfreq_cooling_power *dfc_power = dfc->power_ops; if (dfc_power->get_dynamic_power) - return dfc_power->get_dynamic_power(freq, voltage); + return dfc_power->get_dynamic_power(dfc->devfreq, freq, + voltage); freq_mhz = freq / 1000000; power = (u64)dfc_power->dyn_power_coeff * freq_mhz * voltage * voltage; diff --git a/drivers/thermal/int340x_thermal/int3400_thermal.c b/drivers/thermal/int340x_thermal/int3400_thermal.c index 5836e5554433..9413c4abf0b9 100644 --- a/drivers/thermal/int340x_thermal/int3400_thermal.c +++ b/drivers/thermal/int340x_thermal/int3400_thermal.c @@ -96,7 +96,7 @@ static ssize_t current_uuid_store(struct device *dev, return -EINVAL; } -static DEVICE_ATTR(current_uuid, 0644, current_uuid_show, current_uuid_store); +static DEVICE_ATTR_RW(current_uuid); static DEVICE_ATTR_RO(available_uuids); static struct attribute *uuid_attrs[] = { &dev_attr_available_uuids.attr, diff --git a/drivers/thermal/intel_pch_thermal.c b/drivers/thermal/intel_pch_thermal.c index 19bf2028e508..2b49e8d0fe9e 100644 --- a/drivers/thermal/intel_pch_thermal.c +++ b/drivers/thermal/intel_pch_thermal.c @@ -29,6 +29,7 @@ #define PCH_THERMAL_DID_HSW_2 0x8C24 /* Haswell PCH */ #define PCH_THERMAL_DID_WPT 0x9CA4 /* Wildcat Point */ #define PCH_THERMAL_DID_SKL 0x9D31 /* Skylake PCH */ +#define PCH_THERMAL_DID_SKL_H 0xA131 /* Skylake PCH 100 series */ /* Wildcat Point-LP PCH Thermal registers */ #define WPT_TEMP 0x0000 /* Temperature */ @@ -273,37 +274,44 @@ static struct thermal_zone_device_ops tzd_ops = { .get_trip_temp = pch_get_trip_temp, }; +enum board_ids { + board_hsw, + board_wpt, + board_skl, +}; + +static const struct board_info { + const char *name; + const struct pch_dev_ops *ops; +} board_info[] = { + [board_hsw] = { + .name = "pch_haswell", + .ops = &pch_dev_ops_wpt, + }, + [board_wpt] = { + .name = "pch_wildcat_point", + .ops = &pch_dev_ops_wpt, + }, + [board_skl] = { + .name = "pch_skylake", + .ops = &pch_dev_ops_wpt, + }, +}; static int intel_pch_thermal_probe(struct pci_dev *pdev, const struct pci_device_id *id) { + enum board_ids board_id = id->driver_data; + const struct board_info *bi = &board_info[board_id]; struct pch_thermal_device *ptd; int err; int nr_trips; - char *dev_name; ptd = devm_kzalloc(&pdev->dev, sizeof(*ptd), GFP_KERNEL); if (!ptd) return -ENOMEM; - switch (pdev->device) { - case PCH_THERMAL_DID_WPT: - ptd->ops = &pch_dev_ops_wpt; - dev_name = "pch_wildcat_point"; - break; - case PCH_THERMAL_DID_SKL: - ptd->ops = &pch_dev_ops_wpt; - dev_name = "pch_skylake"; - break; - case PCH_THERMAL_DID_HSW_1: - case PCH_THERMAL_DID_HSW_2: - ptd->ops = &pch_dev_ops_wpt; - dev_name = "pch_haswell"; - break; - default: - dev_err(&pdev->dev, "unknown pch thermal device\n"); - return -ENODEV; - } + ptd->ops = bi->ops; pci_set_drvdata(pdev, ptd); ptd->pdev = pdev; @@ -331,11 +339,11 @@ static int intel_pch_thermal_probe(struct pci_dev *pdev, if (err) goto error_cleanup; - ptd->tzd = thermal_zone_device_register(dev_name, nr_trips, 0, ptd, + ptd->tzd = thermal_zone_device_register(bi->name, nr_trips, 0, ptd, &tzd_ops, NULL, 0, 0); if (IS_ERR(ptd->tzd)) { dev_err(&pdev->dev, "Failed to register thermal zone %s\n", - dev_name); + bi->name); err = PTR_ERR(ptd->tzd); goto error_cleanup; } @@ -380,10 +388,16 @@ static int intel_pch_thermal_resume(struct device *device) } static struct pci_device_id intel_pch_thermal_id[] = { - { PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCH_THERMAL_DID_WPT) }, - { PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCH_THERMAL_DID_SKL) }, - { PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCH_THERMAL_DID_HSW_1) }, - { PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCH_THERMAL_DID_HSW_2) }, + { PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCH_THERMAL_DID_HSW_1), + .driver_data = board_hsw, }, + { PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCH_THERMAL_DID_HSW_2), + .driver_data = board_hsw, }, + { PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCH_THERMAL_DID_WPT), + .driver_data = board_wpt, }, + { PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCH_THERMAL_DID_SKL), + .driver_data = board_skl, }, + { PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCH_THERMAL_DID_SKL_H), + .driver_data = board_skl, }, { 0, }, }; MODULE_DEVICE_TABLE(pci, intel_pch_thermal_id); diff --git a/drivers/thermal/max77620_thermal.c b/drivers/thermal/max77620_thermal.c index 83905ff46e40..e9a1fe342760 100644 --- a/drivers/thermal/max77620_thermal.c +++ b/drivers/thermal/max77620_thermal.c @@ -149,6 +149,7 @@ static struct platform_device_id max77620_thermal_devtype[] = { { .name = "max77620-thermal", }, {}, }; +MODULE_DEVICE_TABLE(platform, max77620_thermal_devtype); static struct platform_driver max77620_thermal_driver = { .driver = { diff --git a/drivers/thermal/qcom-spmi-temp-alarm.c b/drivers/thermal/qcom-spmi-temp-alarm.c index 819c6d5d7aa7..f50241962ad2 100644 --- a/drivers/thermal/qcom-spmi-temp-alarm.c +++ b/drivers/thermal/qcom-spmi-temp-alarm.c @@ -200,7 +200,7 @@ static int qpnp_tm_probe(struct platform_device *pdev) struct qpnp_tm_chip *chip; struct device_node *node; u8 type, subtype; - u32 res[2]; + u32 res; int ret, irq; node = pdev->dev.of_node; @@ -215,7 +215,7 @@ static int qpnp_tm_probe(struct platform_device *pdev) if (!chip->map) return -ENXIO; - ret = of_property_read_u32_array(node, "reg", res, 2); + ret = of_property_read_u32(node, "reg", &res); if (ret < 0) return ret; @@ -228,7 +228,7 @@ static int qpnp_tm_probe(struct platform_device *pdev) if (PTR_ERR(chip->adc) == -EPROBE_DEFER) return PTR_ERR(chip->adc); - chip->base = res[0]; + chip->base = res; ret = qpnp_tm_read(chip, QPNP_TM_REG_TYPE, &type); if (ret < 0) { diff --git a/drivers/thermal/rockchip_thermal.c b/drivers/thermal/rockchip_thermal.c index e227a9f0acf7..b811b0fb61b1 100644 --- a/drivers/thermal/rockchip_thermal.c +++ b/drivers/thermal/rockchip_thermal.c @@ -524,11 +524,6 @@ static void rk_tsadcv2_initialize(struct regmap *grf, void __iomem *regs, regs + TSADCV2_AUTO_PERIOD_HT); writel_relaxed(TSADCV2_HIGHT_TSHUT_DEBOUNCE_COUNT, regs + TSADCV2_HIGHT_TSHUT_DEBOUNCE); - - if (IS_ERR(grf)) { - pr_warn("%s: Missing rockchip,grf property\n", __func__); - return; - } } /** @@ -971,6 +966,8 @@ static int rockchip_configure_from_dt(struct device *dev, * need this property. */ thermal->grf = syscon_regmap_lookup_by_phandle(np, "rockchip,grf"); + if (IS_ERR(thermal->grf)) + dev_warn(dev, "Missing rockchip,grf property\n"); return 0; } diff --git a/drivers/thermal/tango_thermal.c b/drivers/thermal/tango_thermal.c index 201304aeafeb..4e67795cb6ce 100644 --- a/drivers/thermal/tango_thermal.c +++ b/drivers/thermal/tango_thermal.c @@ -107,6 +107,7 @@ static const struct of_device_id tango_sensor_ids[] = { }, { /* sentinel */ } }; +MODULE_DEVICE_TABLE(of, tango_sensor_ids); static struct platform_driver tango_thermal_driver = { .probe = tango_thermal_probe, diff --git a/drivers/thermal/thermal_hwmon.c b/drivers/thermal/thermal_hwmon.c index c41c7742903a..541af5946203 100644 --- a/drivers/thermal/thermal_hwmon.c +++ b/drivers/thermal/thermal_hwmon.c @@ -64,7 +64,7 @@ name_show(struct device *dev, struct device_attribute *attr, char *buf) struct thermal_hwmon_device *hwmon = dev_get_drvdata(dev); return sprintf(buf, "%s\n", hwmon->type); } -static DEVICE_ATTR(name, 0444, name_show, NULL); +static DEVICE_ATTR_RO(name); static ssize_t temp_input_show(struct device *dev, struct device_attribute *attr, char *buf) @@ -98,7 +98,7 @@ temp_crit_show(struct device *dev, struct device_attribute *attr, char *buf) int temperature; int ret; - ret = tz->ops->get_trip_temp(tz, 0, &temperature); + ret = tz->ops->get_crit_temp(tz, &temperature); if (ret) return ret; diff --git a/drivers/thermal/ti-soc-thermal/ti-bandgap.c b/drivers/thermal/ti-soc-thermal/ti-bandgap.c index 06ea9766a70a..ba9c302454fb 100644 --- a/drivers/thermal/ti-soc-thermal/ti-bandgap.c +++ b/drivers/thermal/ti-soc-thermal/ti-bandgap.c @@ -1298,7 +1298,7 @@ int ti_bandgap_probe(struct platform_device *pdev) if (IS_ERR(bgp->div_clk)) { dev_err(&pdev->dev, "failed to request div_ts_ck clock ref\n"); ret = PTR_ERR(bgp->div_clk); - goto free_irqs; + goto put_fclock; } for (i = 0; i < bgp->conf->sensor_count; i++) { @@ -1430,8 +1430,9 @@ disable_clk: if (TI_BANDGAP_HAS(bgp, CLK_CTRL)) clk_disable_unprepare(bgp->fclock); put_clks: - clk_put(bgp->fclock); clk_put(bgp->div_clk); +put_fclock: + clk_put(bgp->fclock); free_irqs: if (TI_BANDGAP_HAS(bgp, TSHUT)) { free_irq(gpio_to_irq(bgp->tshut_gpio), NULL); diff --git a/drivers/thermal/x86_pkg_temp_thermal.c b/drivers/thermal/x86_pkg_temp_thermal.c index 95f4c1bcdb4c..d93eee2f101b 100644 --- a/drivers/thermal/x86_pkg_temp_thermal.c +++ b/drivers/thermal/x86_pkg_temp_thermal.c @@ -54,37 +54,33 @@ MODULE_PARM_DESC(notify_delay_ms, * is some wrong values returned by cpuid for number of thresholds. */ #define MAX_NUMBER_OF_TRIPS 2 -/* Limit number of package temp zones */ -#define MAX_PKG_TEMP_ZONE_IDS 256 - -struct phy_dev_entry { - struct list_head list; - u16 phys_proc_id; - u16 first_cpu; - u32 tj_max; - int ref_cnt; - u32 start_pkg_therm_low; - u32 start_pkg_therm_high; - struct thermal_zone_device *tzone; + +struct pkg_device { + int cpu; + bool work_scheduled; + u32 tj_max; + u32 msr_pkg_therm_low; + u32 msr_pkg_therm_high; + struct delayed_work work; + struct thermal_zone_device *tzone; + struct cpumask cpumask; }; static struct thermal_zone_params pkg_temp_tz_params = { .no_hwmon = true, }; -/* List maintaining number of package instances */ -static LIST_HEAD(phy_dev_list); -static DEFINE_MUTEX(phy_dev_list_mutex); - -/* Interrupt to work function schedule queue */ -static DEFINE_PER_CPU(struct delayed_work, pkg_temp_thermal_threshold_work); +/* Keep track of how many package pointers we allocated in init() */ +static int max_packages __read_mostly; +/* Array of package pointers */ +static struct pkg_device **packages; +/* Serializes interrupt notification, work and hotplug */ +static DEFINE_SPINLOCK(pkg_temp_lock); +/* Protects zone operation in the work function against hotplug removal */ +static DEFINE_MUTEX(thermal_zone_mutex); -/* To track if the work is already scheduled on a package */ -static u8 *pkg_work_scheduled; - -/* Spin lock to prevent races with pkg_work_scheduled */ -static spinlock_t pkg_work_lock; -static u16 max_phy_id; +/* The dynamically assigned cpu hotplug state for module_exit() */ +static enum cpuhp_state pkg_thermal_hp_state __read_mostly; /* Debug counters to show using debugfs */ static struct dentry *debugfs; @@ -116,22 +112,20 @@ err_out: return -ENOENT; } -static struct phy_dev_entry - *pkg_temp_thermal_get_phy_entry(unsigned int cpu) +/* + * Protection: + * + * - cpu hotplug: Read serialized by cpu hotplug lock + * Write must hold pkg_temp_lock + * + * - Other callsites: Must hold pkg_temp_lock + */ +static struct pkg_device *pkg_temp_thermal_get_dev(unsigned int cpu) { - u16 phys_proc_id = topology_physical_package_id(cpu); - struct phy_dev_entry *phy_ptr; - - mutex_lock(&phy_dev_list_mutex); - - list_for_each_entry(phy_ptr, &phy_dev_list, list) - if (phy_ptr->phys_proc_id == phys_proc_id) { - mutex_unlock(&phy_dev_list_mutex); - return phy_ptr; - } - - mutex_unlock(&phy_dev_list_mutex); + int pkgid = topology_logical_package_id(cpu); + if (pkgid >= 0 && pkgid < max_packages) + return packages[pkgid]; return NULL; } @@ -141,61 +135,44 @@ static struct phy_dev_entry */ static int get_tj_max(int cpu, u32 *tj_max) { - u32 eax, edx; - u32 val; + u32 eax, edx, val; int err; err = rdmsr_safe_on_cpu(cpu, MSR_IA32_TEMPERATURE_TARGET, &eax, &edx); if (err) - goto err_ret; - else { - val = (eax >> 16) & 0xff; - if (val) - *tj_max = val * 1000; - else { - err = -EINVAL; - goto err_ret; - } - } + return err; - return 0; -err_ret: - *tj_max = 0; - return err; + val = (eax >> 16) & 0xff; + *tj_max = val * 1000; + + return val ? 0 : -EINVAL; } static int sys_get_curr_temp(struct thermal_zone_device *tzd, int *temp) { + struct pkg_device *pkgdev = tzd->devdata; u32 eax, edx; - struct phy_dev_entry *phy_dev_entry; - phy_dev_entry = tzd->devdata; - rdmsr_on_cpu(phy_dev_entry->first_cpu, MSR_IA32_PACKAGE_THERM_STATUS, - &eax, &edx); + rdmsr_on_cpu(pkgdev->cpu, MSR_IA32_PACKAGE_THERM_STATUS, &eax, &edx); if (eax & 0x80000000) { - *temp = phy_dev_entry->tj_max - - ((eax >> 16) & 0x7f) * 1000; + *temp = pkgdev->tj_max - ((eax >> 16) & 0x7f) * 1000; pr_debug("sys_get_curr_temp %d\n", *temp); return 0; } - return -EINVAL; } static int sys_get_trip_temp(struct thermal_zone_device *tzd, - int trip, int *temp) + int trip, int *temp) { - u32 eax, edx; - struct phy_dev_entry *phy_dev_entry; - u32 mask, shift; + struct pkg_device *pkgdev = tzd->devdata; unsigned long thres_reg_value; + u32 mask, shift, eax, edx; int ret; if (trip >= MAX_NUMBER_OF_TRIPS) return -EINVAL; - phy_dev_entry = tzd->devdata; - if (trip) { mask = THERM_MASK_THRESHOLD1; shift = THERM_SHIFT_THRESHOLD1; @@ -204,14 +181,14 @@ static int sys_get_trip_temp(struct thermal_zone_device *tzd, shift = THERM_SHIFT_THRESHOLD0; } - ret = rdmsr_on_cpu(phy_dev_entry->first_cpu, - MSR_IA32_PACKAGE_THERM_INTERRUPT, &eax, &edx); + ret = rdmsr_on_cpu(pkgdev->cpu, MSR_IA32_PACKAGE_THERM_INTERRUPT, + &eax, &edx); if (ret < 0) - return -EINVAL; + return ret; thres_reg_value = (eax & mask) >> shift; if (thres_reg_value) - *temp = phy_dev_entry->tj_max - thres_reg_value * 1000; + *temp = pkgdev->tj_max - thres_reg_value * 1000; else *temp = 0; pr_debug("sys_get_trip_temp %d\n", *temp); @@ -219,24 +196,20 @@ static int sys_get_trip_temp(struct thermal_zone_device *tzd, return 0; } -static int sys_set_trip_temp(struct thermal_zone_device *tzd, int trip, - int temp) +static int +sys_set_trip_temp(struct thermal_zone_device *tzd, int trip, int temp) { - u32 l, h; - struct phy_dev_entry *phy_dev_entry; - u32 mask, shift, intr; + struct pkg_device *pkgdev = tzd->devdata; + u32 l, h, mask, shift, intr; int ret; - phy_dev_entry = tzd->devdata; - - if (trip >= MAX_NUMBER_OF_TRIPS || temp >= phy_dev_entry->tj_max) + if (trip >= MAX_NUMBER_OF_TRIPS || temp >= pkgdev->tj_max) return -EINVAL; - ret = rdmsr_on_cpu(phy_dev_entry->first_cpu, - MSR_IA32_PACKAGE_THERM_INTERRUPT, - &l, &h); + ret = rdmsr_on_cpu(pkgdev->cpu, MSR_IA32_PACKAGE_THERM_INTERRUPT, + &l, &h); if (ret < 0) - return -EINVAL; + return ret; if (trip) { mask = THERM_MASK_THRESHOLD1; @@ -252,24 +225,20 @@ static int sys_set_trip_temp(struct thermal_zone_device *tzd, int trip, * When users space sets a trip temperature == 0, which is indication * that, it is no longer interested in receiving notifications. */ - if (!temp) + if (!temp) { l &= ~intr; - else { - l |= (phy_dev_entry->tj_max - temp)/1000 << shift; + } else { + l |= (pkgdev->tj_max - temp)/1000 << shift; l |= intr; } - return wrmsr_on_cpu(phy_dev_entry->first_cpu, - MSR_IA32_PACKAGE_THERM_INTERRUPT, - l, h); + return wrmsr_on_cpu(pkgdev->cpu, MSR_IA32_PACKAGE_THERM_INTERRUPT, l, h); } -static int sys_get_trip_type(struct thermal_zone_device *thermal, - int trip, enum thermal_trip_type *type) +static int sys_get_trip_type(struct thermal_zone_device *thermal, int trip, + enum thermal_trip_type *type) { - *type = THERMAL_TRIP_PASSIVE; - return 0; } @@ -281,7 +250,7 @@ static struct thermal_zone_device_ops tzone_ops = { .set_trip_temp = sys_set_trip_temp, }; -static bool pkg_temp_thermal_platform_thermal_rate_control(void) +static bool pkg_thermal_rate_control(void) { return true; } @@ -289,8 +258,8 @@ static bool pkg_temp_thermal_platform_thermal_rate_control(void) /* Enable threshold interrupt on local package/cpu */ static inline void enable_pkg_thres_interrupt(void) { - u32 l, h; u8 thres_0, thres_1; + u32 l, h; rdmsr(MSR_IA32_PACKAGE_THERM_INTERRUPT, l, h); /* only enable/disable if it had valid threshold value */ @@ -307,271 +276,232 @@ static inline void enable_pkg_thres_interrupt(void) static inline void disable_pkg_thres_interrupt(void) { u32 l, h; + rdmsr(MSR_IA32_PACKAGE_THERM_INTERRUPT, l, h); - wrmsr(MSR_IA32_PACKAGE_THERM_INTERRUPT, - l & (~THERM_INT_THRESHOLD0_ENABLE) & - (~THERM_INT_THRESHOLD1_ENABLE), h); + + l &= ~(THERM_INT_THRESHOLD0_ENABLE | THERM_INT_THRESHOLD1_ENABLE); + wrmsr(MSR_IA32_PACKAGE_THERM_INTERRUPT, l, h); } static void pkg_temp_thermal_threshold_work_fn(struct work_struct *work) { - __u64 msr_val; + struct thermal_zone_device *tzone = NULL; int cpu = smp_processor_id(); - int phy_id = topology_physical_package_id(cpu); - struct phy_dev_entry *phdev = pkg_temp_thermal_get_phy_entry(cpu); - bool notify = false; - unsigned long flags; - - if (!phdev) - return; + struct pkg_device *pkgdev; + u64 msr_val, wr_val; - spin_lock_irqsave(&pkg_work_lock, flags); + mutex_lock(&thermal_zone_mutex); + spin_lock_irq(&pkg_temp_lock); ++pkg_work_cnt; - if (unlikely(phy_id > max_phy_id)) { - spin_unlock_irqrestore(&pkg_work_lock, flags); + + pkgdev = pkg_temp_thermal_get_dev(cpu); + if (!pkgdev) { + spin_unlock_irq(&pkg_temp_lock); + mutex_unlock(&thermal_zone_mutex); return; } - pkg_work_scheduled[phy_id] = 0; - spin_unlock_irqrestore(&pkg_work_lock, flags); + pkgdev->work_scheduled = false; - enable_pkg_thres_interrupt(); rdmsrl(MSR_IA32_PACKAGE_THERM_STATUS, msr_val); - if (msr_val & THERM_LOG_THRESHOLD0) { - wrmsrl(MSR_IA32_PACKAGE_THERM_STATUS, - msr_val & ~THERM_LOG_THRESHOLD0); - notify = true; - } - if (msr_val & THERM_LOG_THRESHOLD1) { - wrmsrl(MSR_IA32_PACKAGE_THERM_STATUS, - msr_val & ~THERM_LOG_THRESHOLD1); - notify = true; - } - if (notify) { - pr_debug("thermal_zone_device_update\n"); - thermal_zone_device_update(phdev->tzone, - THERMAL_EVENT_UNSPECIFIED); + wr_val = msr_val & ~(THERM_LOG_THRESHOLD0 | THERM_LOG_THRESHOLD1); + if (wr_val != msr_val) { + wrmsrl(MSR_IA32_PACKAGE_THERM_STATUS, wr_val); + tzone = pkgdev->tzone; } + + enable_pkg_thres_interrupt(); + spin_unlock_irq(&pkg_temp_lock); + + /* + * If tzone is not NULL, then thermal_zone_mutex will prevent the + * concurrent removal in the cpu offline callback. + */ + if (tzone) + thermal_zone_device_update(tzone, THERMAL_EVENT_UNSPECIFIED); + + mutex_unlock(&thermal_zone_mutex); } -static int pkg_temp_thermal_platform_thermal_notify(__u64 msr_val) +static void pkg_thermal_schedule_work(int cpu, struct delayed_work *work) +{ + unsigned long ms = msecs_to_jiffies(notify_delay_ms); + + schedule_delayed_work_on(cpu, work, ms); +} + +static int pkg_thermal_notify(u64 msr_val) { - unsigned long flags; int cpu = smp_processor_id(); - int phy_id = topology_physical_package_id(cpu); + struct pkg_device *pkgdev; + unsigned long flags; - /* - * When a package is in interrupted state, all CPU's in that package - * are in the same interrupt state. So scheduling on any one CPU in - * the package is enough and simply return for others. - */ - spin_lock_irqsave(&pkg_work_lock, flags); + spin_lock_irqsave(&pkg_temp_lock, flags); ++pkg_interrupt_cnt; - if (unlikely(phy_id > max_phy_id) || unlikely(!pkg_work_scheduled) || - pkg_work_scheduled[phy_id]) { - disable_pkg_thres_interrupt(); - spin_unlock_irqrestore(&pkg_work_lock, flags); - return -EINVAL; - } - pkg_work_scheduled[phy_id] = 1; - spin_unlock_irqrestore(&pkg_work_lock, flags); disable_pkg_thres_interrupt(); - schedule_delayed_work_on(cpu, - &per_cpu(pkg_temp_thermal_threshold_work, cpu), - msecs_to_jiffies(notify_delay_ms)); - return 0; -} -static int find_siblings_cpu(int cpu) -{ - int i; - int id = topology_physical_package_id(cpu); - - for_each_online_cpu(i) - if (i != cpu && topology_physical_package_id(i) == id) - return i; + /* Work is per package, so scheduling it once is enough. */ + pkgdev = pkg_temp_thermal_get_dev(cpu); + if (pkgdev && !pkgdev->work_scheduled) { + pkgdev->work_scheduled = true; + pkg_thermal_schedule_work(pkgdev->cpu, &pkgdev->work); + } + spin_unlock_irqrestore(&pkg_temp_lock, flags); return 0; } static int pkg_temp_thermal_device_add(unsigned int cpu) { - int err; - u32 tj_max; - struct phy_dev_entry *phy_dev_entry; - int thres_count; - u32 eax, ebx, ecx, edx; - u8 *temp; - unsigned long flags; + int pkgid = topology_logical_package_id(cpu); + u32 tj_max, eax, ebx, ecx, edx; + struct pkg_device *pkgdev; + int thres_count, err; + + if (pkgid >= max_packages) + return -ENOMEM; cpuid(6, &eax, &ebx, &ecx, &edx); thres_count = ebx & 0x07; if (!thres_count) return -ENODEV; - if (topology_physical_package_id(cpu) > MAX_PKG_TEMP_ZONE_IDS) - return -ENODEV; - thres_count = clamp_val(thres_count, 0, MAX_NUMBER_OF_TRIPS); err = get_tj_max(cpu, &tj_max); if (err) - goto err_ret; - - mutex_lock(&phy_dev_list_mutex); + return err; - phy_dev_entry = kzalloc(sizeof(*phy_dev_entry), GFP_KERNEL); - if (!phy_dev_entry) { - err = -ENOMEM; - goto err_ret_unlock; - } + pkgdev = kzalloc(sizeof(*pkgdev), GFP_KERNEL); + if (!pkgdev) + return -ENOMEM; - spin_lock_irqsave(&pkg_work_lock, flags); - if (topology_physical_package_id(cpu) > max_phy_id) - max_phy_id = topology_physical_package_id(cpu); - temp = krealloc(pkg_work_scheduled, - (max_phy_id+1) * sizeof(u8), GFP_ATOMIC); - if (!temp) { - spin_unlock_irqrestore(&pkg_work_lock, flags); - err = -ENOMEM; - goto err_ret_free; - } - pkg_work_scheduled = temp; - pkg_work_scheduled[topology_physical_package_id(cpu)] = 0; - spin_unlock_irqrestore(&pkg_work_lock, flags); - - phy_dev_entry->phys_proc_id = topology_physical_package_id(cpu); - phy_dev_entry->first_cpu = cpu; - phy_dev_entry->tj_max = tj_max; - phy_dev_entry->ref_cnt = 1; - phy_dev_entry->tzone = thermal_zone_device_register("x86_pkg_temp", + INIT_DELAYED_WORK(&pkgdev->work, pkg_temp_thermal_threshold_work_fn); + pkgdev->cpu = cpu; + pkgdev->tj_max = tj_max; + pkgdev->tzone = thermal_zone_device_register("x86_pkg_temp", thres_count, - (thres_count == MAX_NUMBER_OF_TRIPS) ? - 0x03 : 0x01, - phy_dev_entry, &tzone_ops, &pkg_temp_tz_params, 0, 0); - if (IS_ERR(phy_dev_entry->tzone)) { - err = PTR_ERR(phy_dev_entry->tzone); - goto err_ret_free; + (thres_count == MAX_NUMBER_OF_TRIPS) ? 0x03 : 0x01, + pkgdev, &tzone_ops, &pkg_temp_tz_params, 0, 0); + if (IS_ERR(pkgdev->tzone)) { + err = PTR_ERR(pkgdev->tzone); + kfree(pkgdev); + return err; } /* Store MSR value for package thermal interrupt, to restore at exit */ - rdmsr_on_cpu(cpu, MSR_IA32_PACKAGE_THERM_INTERRUPT, - &phy_dev_entry->start_pkg_therm_low, - &phy_dev_entry->start_pkg_therm_high); - - list_add_tail(&phy_dev_entry->list, &phy_dev_list); - pr_debug("pkg_temp_thermal_device_add :phy_id %d cpu %d\n", - phy_dev_entry->phys_proc_id, cpu); - - mutex_unlock(&phy_dev_list_mutex); + rdmsr(MSR_IA32_PACKAGE_THERM_INTERRUPT, pkgdev->msr_pkg_therm_low, + pkgdev->msr_pkg_therm_high); + cpumask_set_cpu(cpu, &pkgdev->cpumask); + spin_lock_irq(&pkg_temp_lock); + packages[pkgid] = pkgdev; + spin_unlock_irq(&pkg_temp_lock); return 0; - -err_ret_free: - kfree(phy_dev_entry); -err_ret_unlock: - mutex_unlock(&phy_dev_list_mutex); - -err_ret: - return err; } -static int pkg_temp_thermal_device_remove(unsigned int cpu) +static int pkg_thermal_cpu_offline(unsigned int cpu) { - struct phy_dev_entry *n; - u16 phys_proc_id = topology_physical_package_id(cpu); - struct phy_dev_entry *phdev = - pkg_temp_thermal_get_phy_entry(cpu); + struct pkg_device *pkgdev = pkg_temp_thermal_get_dev(cpu); + bool lastcpu, was_target; + int target; - if (!phdev) - return -ENODEV; + if (!pkgdev) + return 0; - mutex_lock(&phy_dev_list_mutex); - /* If we are loosing the first cpu for this package, we need change */ - if (phdev->first_cpu == cpu) { - phdev->first_cpu = find_siblings_cpu(cpu); - pr_debug("thermal_device_remove: first cpu switched %d\n", - phdev->first_cpu); + target = cpumask_any_but(&pkgdev->cpumask, cpu); + cpumask_clear_cpu(cpu, &pkgdev->cpumask); + lastcpu = target >= nr_cpu_ids; + /* + * Remove the sysfs files, if this is the last cpu in the package + * before doing further cleanups. + */ + if (lastcpu) { + struct thermal_zone_device *tzone = pkgdev->tzone; + + /* + * We must protect against a work function calling + * thermal_zone_update, after/while unregister. We null out + * the pointer under the zone mutex, so the worker function + * won't try to call. + */ + mutex_lock(&thermal_zone_mutex); + pkgdev->tzone = NULL; + mutex_unlock(&thermal_zone_mutex); + + thermal_zone_device_unregister(tzone); } + + /* Protect against work and interrupts */ + spin_lock_irq(&pkg_temp_lock); + /* - * It is possible that no siblings left as this was the last cpu - * going offline. We don't need to worry about this assignment - * as the phydev entry will be removed in this case and - * thermal zone is removed. - */ - --phdev->ref_cnt; - pr_debug("thermal_device_remove: pkg: %d cpu %d ref_cnt %d\n", - phys_proc_id, cpu, phdev->ref_cnt); - if (!phdev->ref_cnt) - list_for_each_entry_safe(phdev, n, &phy_dev_list, list) { - if (phdev->phys_proc_id == phys_proc_id) { - thermal_zone_device_unregister(phdev->tzone); - list_del(&phdev->list); - kfree(phdev); - break; - } - } - mutex_unlock(&phy_dev_list_mutex); + * Check whether this cpu was the current target and store the new + * one. When we drop the lock, then the interrupt notify function + * will see the new target. + */ + was_target = pkgdev->cpu == cpu; + pkgdev->cpu = target; - return 0; -} + /* + * If this is the last CPU in the package remove the package + * reference from the array and restore the interrupt MSR. When we + * drop the lock neither the interrupt notify function nor the + * worker will see the package anymore. + */ + if (lastcpu) { + packages[topology_logical_package_id(cpu)] = NULL; + /* After this point nothing touches the MSR anymore. */ + wrmsr(MSR_IA32_PACKAGE_THERM_INTERRUPT, + pkgdev->msr_pkg_therm_low, pkgdev->msr_pkg_therm_high); + } -static int get_core_online(unsigned int cpu) -{ - struct cpuinfo_x86 *c = &cpu_data(cpu); - struct phy_dev_entry *phdev = pkg_temp_thermal_get_phy_entry(cpu); - - /* Check if there is already an instance for this package */ - if (!phdev) { - if (!cpu_has(c, X86_FEATURE_DTHERM) || - !cpu_has(c, X86_FEATURE_PTS)) - return -ENODEV; - if (pkg_temp_thermal_device_add(cpu)) - return -ENODEV; - } else { - mutex_lock(&phy_dev_list_mutex); - ++phdev->ref_cnt; - pr_debug("get_core_online: cpu %d ref_cnt %d\n", - cpu, phdev->ref_cnt); - mutex_unlock(&phy_dev_list_mutex); + /* + * Check whether there is work scheduled and whether the work is + * targeted at the outgoing CPU. + */ + if (pkgdev->work_scheduled && was_target) { + /* + * To cancel the work we need to drop the lock, otherwise + * we might deadlock if the work needs to be flushed. + */ + spin_unlock_irq(&pkg_temp_lock); + cancel_delayed_work_sync(&pkgdev->work); + spin_lock_irq(&pkg_temp_lock); + /* + * If this is not the last cpu in the package and the work + * did not run after we dropped the lock above, then we + * need to reschedule the work, otherwise the interrupt + * stays disabled forever. + */ + if (!lastcpu && pkgdev->work_scheduled) + pkg_thermal_schedule_work(target, &pkgdev->work); } - INIT_DELAYED_WORK(&per_cpu(pkg_temp_thermal_threshold_work, cpu), - pkg_temp_thermal_threshold_work_fn); - pr_debug("get_core_online: cpu %d successful\n", cpu); + spin_unlock_irq(&pkg_temp_lock); + /* Final cleanup if this is the last cpu */ + if (lastcpu) + kfree(pkgdev); return 0; } -static void put_core_offline(unsigned int cpu) +static int pkg_thermal_cpu_online(unsigned int cpu) { - if (!pkg_temp_thermal_device_remove(cpu)) - cancel_delayed_work_sync( - &per_cpu(pkg_temp_thermal_threshold_work, cpu)); + struct pkg_device *pkgdev = pkg_temp_thermal_get_dev(cpu); + struct cpuinfo_x86 *c = &cpu_data(cpu); - pr_debug("put_core_offline: cpu %d\n", cpu); -} + /* Paranoia check */ + if (!cpu_has(c, X86_FEATURE_DTHERM) || !cpu_has(c, X86_FEATURE_PTS)) + return -ENODEV; -static int pkg_temp_thermal_cpu_callback(struct notifier_block *nfb, - unsigned long action, void *hcpu) -{ - unsigned int cpu = (unsigned long) hcpu; - - switch (action & ~CPU_TASKS_FROZEN) { - case CPU_ONLINE: - case CPU_DOWN_FAILED: - get_core_online(cpu); - break; - case CPU_DOWN_PREPARE: - put_core_offline(cpu); - break; + /* If the package exists, nothing to do */ + if (pkgdev) { + cpumask_set_cpu(cpu, &pkgdev->cpumask); + return 0; } - return NOTIFY_OK; + return pkg_temp_thermal_device_add(cpu); } -static struct notifier_block pkg_temp_thermal_notifier __refdata = { - .notifier_call = pkg_temp_thermal_cpu_callback, -}; - static const struct x86_cpu_id __initconst pkg_temp_thermal_ids[] = { { X86_VENDOR_INTEL, X86_FAMILY_ANY, X86_MODEL_ANY, X86_FEATURE_PTS }, {} @@ -580,71 +510,46 @@ MODULE_DEVICE_TABLE(x86cpu, pkg_temp_thermal_ids); static int __init pkg_temp_thermal_init(void) { - int i; + int ret; if (!x86_match_cpu(pkg_temp_thermal_ids)) return -ENODEV; - spin_lock_init(&pkg_work_lock); - platform_thermal_package_notify = - pkg_temp_thermal_platform_thermal_notify; - platform_thermal_package_rate_control = - pkg_temp_thermal_platform_thermal_rate_control; + max_packages = topology_max_packages(); + packages = kzalloc(max_packages * sizeof(struct pkg_device *), GFP_KERNEL); + if (!packages) + return -ENOMEM; - cpu_notifier_register_begin(); - for_each_online_cpu(i) - if (get_core_online(i)) - goto err_ret; - __register_hotcpu_notifier(&pkg_temp_thermal_notifier); - cpu_notifier_register_done(); + ret = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "thermal/x86_pkg:online", + pkg_thermal_cpu_online, pkg_thermal_cpu_offline); + if (ret < 0) + goto err; - pkg_temp_debugfs_init(); /* Don't care if fails */ + /* Store the state for module exit */ + pkg_thermal_hp_state = ret; - return 0; + platform_thermal_package_notify = pkg_thermal_notify; + platform_thermal_package_rate_control = pkg_thermal_rate_control; -err_ret: - for_each_online_cpu(i) - put_core_offline(i); - cpu_notifier_register_done(); - kfree(pkg_work_scheduled); - platform_thermal_package_notify = NULL; - platform_thermal_package_rate_control = NULL; + /* Don't care if it fails */ + pkg_temp_debugfs_init(); + return 0; - return -ENODEV; +err: + kfree(packages); + return ret; } +module_init(pkg_temp_thermal_init) static void __exit pkg_temp_thermal_exit(void) { - struct phy_dev_entry *phdev, *n; - int i; - - cpu_notifier_register_begin(); - __unregister_hotcpu_notifier(&pkg_temp_thermal_notifier); - mutex_lock(&phy_dev_list_mutex); - list_for_each_entry_safe(phdev, n, &phy_dev_list, list) { - /* Retore old MSR value for package thermal interrupt */ - wrmsr_on_cpu(phdev->first_cpu, - MSR_IA32_PACKAGE_THERM_INTERRUPT, - phdev->start_pkg_therm_low, - phdev->start_pkg_therm_high); - thermal_zone_device_unregister(phdev->tzone); - list_del(&phdev->list); - kfree(phdev); - } - mutex_unlock(&phy_dev_list_mutex); platform_thermal_package_notify = NULL; platform_thermal_package_rate_control = NULL; - for_each_online_cpu(i) - cancel_delayed_work_sync( - &per_cpu(pkg_temp_thermal_threshold_work, i)); - cpu_notifier_register_done(); - - kfree(pkg_work_scheduled); + cpuhp_remove_state(pkg_thermal_hp_state); debugfs_remove_recursive(debugfs); + kfree(packages); } - -module_init(pkg_temp_thermal_init) module_exit(pkg_temp_thermal_exit) MODULE_DESCRIPTION("X86 PKG TEMP Thermal Driver"); diff --git a/include/linux/devfreq_cooling.h b/include/linux/devfreq_cooling.h index 7adf6cc4b305..c35d0c0e0ada 100644 --- a/include/linux/devfreq_cooling.h +++ b/include/linux/devfreq_cooling.h @@ -20,7 +20,6 @@ #include <linux/devfreq.h> #include <linux/thermal.h> -#ifdef CONFIG_DEVFREQ_THERMAL /** * struct devfreq_cooling_power - Devfreq cooling power ops @@ -37,12 +36,16 @@ * @dyn_power_coeff * frequency * voltage^2 */ struct devfreq_cooling_power { - unsigned long (*get_static_power)(unsigned long voltage); - unsigned long (*get_dynamic_power)(unsigned long freq, + unsigned long (*get_static_power)(struct devfreq *devfreq, + unsigned long voltage); + unsigned long (*get_dynamic_power)(struct devfreq *devfreq, + unsigned long freq, unsigned long voltage); unsigned long dyn_power_coeff; }; +#ifdef CONFIG_DEVFREQ_THERMAL + struct thermal_cooling_device * of_devfreq_cooling_register_power(struct device_node *np, struct devfreq *df, struct devfreq_cooling_power *dfc_power); |