On Wed, 20 Nov 2013 10:15:29 -0800, Guenter Roeck wrote: > On Wed, Nov 20, 2013 at 07:06:41PM +0100, Jean Delvare wrote: > > Personally I'd just do the minimum to avoid returning an error. In > > other words I'd be fine returning values down to 6 degrees (for the > > Atom D510 at least). We know the value is wrong but it can be corrected > > in user-space, while if we clamp higher, it can no longer be corrected. > > Seems to me it would be much simpler to just return 0 if the valid bit is 0, > and not bother returning -EAGAIN in that case. After all, that is what > it boils down to, isn't it ? In the specific case of these Atom chips, sort of, yes (although I'd return maybe 5000 for continuity with the last known good value, but that's not so important in the end, plus there's no guarantee that the limit is the same for other Atom CPU models / samples.) However we can't really do that in general, as again there is no Intel documentation saying that valid == 0 can only mean "temperature too low". There may be other causes. Now I really can't remember if we ever had reports of -EAGAIN being returned transiently. If the error is always permanent then I'd agree that returning 0 (or any other arbitrarily low value) is no worse than returning -EAGAIN. Alternatively, as long as we have no clean way to return "lower than X but I don't know the exact value" to user-space, it might be more simple to just ignore the valid bit and always return the value. In ticket #2382 I found that values 1°C and 4°C would be returned instead of the error in that case, it's no worse than returning 0°C or an arbitrary 5°C, and it would come for absolutely no extra code. -- Jean Delvare _______________________________________________ lm-sensors mailing list lm-sensors@xxxxxxxxxxxxxx http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
