On Wed, Oct 13, 2021 at 07:17:14PM +0000, Souza, Jose wrote: > On Wed, 2021-10-13 at 12:32 +0300, Ville Syrjälä wrote: > > On Tue, Oct 12, 2021 at 06:00:46PM -0700, José Roberto de Souza wrote: > > > This memory frequency calculated is only used to check if it is zero, > > > what is not useful as it will never actually be zero. > > > > > > Also the calculation is wrong, we should be checking other bit to > > > select the appropriate frequency multiplier while this code is stuck > > > with a fixed multiplier. > > > > I don't think the alternate ref clock was ever used. > > At least I don't recall ever seeing it. > > > > The real problem with this is that IIRC this is just the last > > requested frequency. So on a system with SAGV this will > > change dynamically. > > > > > > > > So here dropping it as whole. > > > > We have a second copy of this in gen6_update_ring_freq(). Rather > > than removing one and leaving another potentially broken one behind we > > should probably just consolidate on a single implementation. > > gen6_update_ring_freq() is related to GPU frequency not memory, don't look related at all to me. > GPU, CPU and memory clocks are all needed there, at least on some platforms. I forget which ones did what exactly. -- Ville Syrjälä Intel
