On Monday 26 Nov 2018 at 12:36:31 (+0100), Daniel Lezcano wrote: > On 26/11/2018 12:09, Quentin Perret wrote: > > On Monday 26 Nov 2018 at 15:49:55 (+0530), Viresh Kumar wrote: > >> On 26-11-18, 11:08, Daniel Lezcano wrote: > >>> On 26/11/2018 10:52, Quentin Perret wrote: > >>>> Maybe you want to test 'if (!raw_capacity || cap_parsing_failed)' at the > >>>> top of topology_parse_cpu_capacity() ? > >>> > >>> I prefer to update the documentation, it makes more sense than adding > >>> more cumbersome tests in the current code. > >> > >> +1 > >> > >> Throwing an error and ignoring DT number completely for the capacity > >> are good enough in my opinion as well. > >> > >> And who cares for the platforms that can't even fill the DT properly :) > > > > Right, I think we all agree the case with a partially filled DT is > > broken. I don't actually care too much about the behaviour in this case, > > but it needs to be consistent with the doc. > > > > So, as long as you fix the doc, that change is fine by me :-) > > Ok what about the following change ? > > " > > If capacity-dmips-mhz is not specified or if the parsing fails, the > default capacity value will be computed against the highest frequency, > it will result most of the time on the same capacity value.
That "most of the time" sounds a bit odd no ? Maybe mention explicitly the case you're referring to (that is when all CPUs have the same max freq) ? > However on > some platform with different OPP set but the same micro-architecture, > the capacity will be scaled down for CPUs having lower frequencies. > > " Other than that LGTM. Thanks, Quentin
