Re: patch: p4tcc and speedstep cpufreq drivers

Chuck Swiger wrote:
> Maxim Sobolev wrote:
> 
>> Nate Lawson wrote:
> 
> [ ... ]
> 
>>> "On-Demand mode may be used at the same time Automatic mode is 
>>> enabled, however, if the system (tries to enable the TCC via 
>>> On-Demand mode[1]) at the same time (automatic mode is enabled[2]) 
>>> AND (a high temperature condition exists [3]), the duty cycle of the 
>>> automatic mode will override the duty cycle selected by the On-Demand 
>>> mode."
>>>
>>> Since automatic mode is set by the BIOS before we even boot, things 
>>> should be fine.
>>
>>
>> Well, this is quite tricky part of the spec. My reading is that the 
>> paragraph above applies only to situation if you are (trying to set 
>> on-demand mode [1]) when both (automatic mode is in effect [2]) *and* 
>> (high temperature condition already exists [3]), in that case 
>> automatic mode will win and override any manual settings.
> 
> 
> I suspect you'd read your paragraph with [1] and [2] joined together, 
> but they can be read seperately just as the spec can.  :-)
> 
>> However, in the case when you have on-demand mode already on and high 
>> temperature condition emerges it will have no effect on duty cycle 
>> until THERMTRIP# kicks in.
>>
>> That's in my view explains why there is big AND in the text above.
> 
> 
> I think the spec is advising developers who try to control TCC that if 
> PROCHOT gets asserted, the CPU may override the programmed settings in 
> favor of the automatic ones.  I don't think the spec is asserting that 
> the CPU is forbidden from reducing power usage if PROCHOT condition is 
> detected, regardless of what automatic mode is set to.
> 
> A reasonable processor would drop to a known minimal power usage state-- 
> hopefully one low enough to keep the CPU from completely overheating 
> even if a fan has failed-- if PROCHOT is seen.
> 
> [ It may also be the case that a CPU does not do so, in which case the 
> ACPI driver code ought to try to pay attention to PROCHOT and reduce 
> power consumption regardless and not just depend on CPU failsafes to 
> work.  If that is your position, well, I would agree with this.  :-) ]

Check the original spec, there is a note at the end of the respective 
paragraph (particularly the first sentence):

If automatic mode is disabled the processor will be operating out of 
specification and
cannot be guaranteed to provide reliable results. Regardless of enabling 
of the automatic
or On-Demand modes, in the event of a catastrophic cooling failure, the 
processor will
automatically shut down when the silicon has reached a temperature of 
approximately
135 °C. At this point the system bus signal THERMTRIP# will go active 
and stay active
until RESET# has been initiated. THERMTRIP# activation is independent of 
processor
activity and does not generate any bus cycles. If THERMTRIP# is 
asserted, processor
core voltage (Vcc) must be removed within the timeframe defined in Table 16.

-Maxim