Hi, > On 3/13/21 12:32 AM, wangyugui wrote: > > HWMON_T_MIN is not common in NVMe SSD, so drop all of them in nvme-hwmon. > > > > HWMON_T_MAX is only common in NVMe SSD Composite sensor, so drop them in other sensors. > > > > Before this patch(SSD: PM1733): > > #sensors > > nvme-pci-4300 > > Adapter: PCI adapter > > Composite: +49.9°C (low = -273.1°C, high = +71.8°C) > > (crit = +84.8°C) > > Sensor 1: +47.9°C (low = -273.1°C, high = +65261.8°C) > > ERROR: Can't get value of subfeature temp3_min: I/O error > > ERROR: Can't get value of subfeature temp3_max: I/O error > > Sensor 2: +49.9°C (low = +0.0°C, high = +0.0°C) > > > > # cat /sys/class/nvme/nvme0/hwmon1/temp3_min > > cat: /sys/class/nvme/nvme0/hwmon1/temp3_min: Input/output error > > # cat /sys/class/nvme/nvme0/hwmon1/temp3_max > > cat: /sys/class/nvme/nvme0/hwmon1/temp3_max: Input/output error > > > > After this patch(SSD: PM1733): > > #sensors > > nvme-pci-4300 > > Adapter: PCI adapter > > Composite: +48.9°C (high = +71.8°C, crit = +84.8°C) > > Sensor 1: +46.9°C > > Sensor 2: +48.9°C > > > > Signed-off-by: missing. > > Either case, no. > > On one of my NVMEs, after setting the limits: > > nvme-pci-0100 > Adapter: PCI adapter > Composite: +29.9°C (low = -0.1°C, high = +76.8°C) > (crit = +78.8°C) > Sensor 1: +29.9°C (low = -0.1°C, high = +254.8°C) > Sensor 2: +37.9°C (low = -0.1°C, high = +254.8°C) high = +254.8°C is a real value or unconfigured value ? Best Regards Wang Yugui (wangyugui@xxxxxxxxxxxx) 2021/03/14 > That it doesn't work on yours doesn't mean it needs to be disabled for > all other NVMEs. Instead, we'll need to figure out how to correctly > identify that limits for the second subfeature sensor are not supported, > or what exactly the NVME complains about when trying to read the limits > for the second subsensor. One possible solution might be to call > nvme_get_temp_thresh() from nvme_hwmon_is_visible() and return 0 if the > call returns an error. > > Also, the low/high limit readings on the composite sensor and on > Sensor 1 only mean that limits are not configured. That is not a reason > to disable the limit attributes entirely. One could instead write useful > limits into those attributes.
