[lm-sensors] Patch v5 Merge Pkgtemp with coretemp
durgadoss.r at intel.com
Wed Apr 27 08:24:19 CEST 2011
Thanks for the comments.
Shall fix them and resubmit the patch.
But, I need clarifications on few things here.
> Functionality-wise, there is still a problem: If one core of a HT CPU
> is taken offline, the driver used to replace it with its sibling if
> it was online. This way the core's temperature was still displayed as
> long as at least one of the HT cores was online. This is no longer the case.
I knew it..but I was thinking to do it as a separate patch,
Once this gets acked. If we have to add this support, the following is
my way of doing it:
When a cpu with a core_id is offlined,
check whether there is any other cpu with same core_id(by looping
through for_each_online_cpu()). If so, get that online.
Please let me know if there are any alternatives.
> You should update Documentation/hwmon/coretemp to reflect your changes.
Should that be a separate patch or part of this patch ?
> > + /*
> > + * Initialize ttarget value. Eventually this will be
> > + * initialized with the value from MSR_IA32_THERM_INTERRUPT
> > + * register. If IA32_TEMPERATURE_TARGET is supported, this
> > + * value will be over written below.
> > + * To Do: Patch to initialize ttarget from MSR_IA32_THERM_INTERRUPT
> > + */
> Do you plan to submit a separate patch to address this ?
Yes. I already submitted a patch for this. Since we wanted any new patch
On coretemp should be done only after merging pkgtemp, I withheld that patch.
> > for_each_online_cpu(i)
> > - coretemp_device_add(i);
> > + get_core_online(i);
> This code should now be in the probe function, and be called before
> the hwmon device is registered.
Our probe function itself is called by coretemp_device_add()
which in turn is called by get_core_online(). So, If I move this code
to probe, not sure how the probe itself will be called. Hence, I feel
this code should be here only. Let me know your thoughts.
More information about the lm-sensors