On Mon, Aug 26, 2019 at 9:11 AM Jean Delvare <jdelvare@xxxxxxx> wrote: > > Hi all, > > I am investigating why "sensors" reports amdgpu's vddgfx as N/A on my > Radeon R5 240. Digging through the code I found that > si_dpm_read_sensor() is being called with idx = AMDGPU_PP_SENSOR_VDDGFX > which is *not* implemented by this function. So it's never going to > report any value. > > The situation is the same for kv_dpm_read_sensor() which also doesn't > support idx = AMDGPU_PP_SENSOR_VDDGFX. idx = AMDGPU_PP_SENSOR_GPU_POWER > is also not supported by either function while it is being called > (resulting in power1 being reported as N/A too). > > My question is whether this is something that can be implemented or not? It's certainly possible, but these chips are so old, we'd probably be hard pressed to find the relevant register/firmware interface to query this information at this point. > > If it can't be implemented because the hardware does not have these > sensors then I would prefer if the attributes are not created at all so > as to not clutter the output of "sensors". Probably best to add these cases to hwmon_attributes_visible() in amdgpu_pm.c. Alex > > Thanks, > -- > Jean Delvare > SUSE L3 Support > _______________________________________________ > amd-gfx mailing list > amd-gfx@xxxxxxxxxxxxxxxxxxxxx > https://lists.freedesktop.org/mailman/listinfo/amd-gfx _______________________________________________ amd-gfx mailing list amd-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/amd-gfx
