On Thu, Nov 17, 2022 at 04:54:03PM +0530, Viresh Kumar wrote: > On 17-11-22, 10:19, Sudeep Holla wrote: > > Why do you need the above 3 changes if the below(4/4) will ensure > > cpufreq_get(cpu) returns the clock frequency. I was expecting to drop the > > whole "confusing" clock bindings and the unnecessary clock provider. > > > > Can't we just use cpufreq_get(cpu) ? > > https://lore.kernel.org/lkml/cover.1657695140.git.viresh.kumar@xxxxxxxxxx/ > > The basic idea (need) here was to fix the DT and let the CPU nodes have clock > related properties, which are missing currently. > > The context can be seen in the above thread. Thanks for the link. Sorry I still don't get the complete picture. Who are the consumers of these clock nodes if not cpufreq itself. I am going to guess, so other device(like inter-connect) with phandle into CPU device perhaps ? Also I assume it will have phandle to non-CPU device and hence we need generic device clock solution. Sorry for the noise, but I still find having both clocks and qcom,freq-domain property is quite confusing but I am fine as I understand it bit better now. -- Regards, Sudeep
