09:10 <wwilly> krzk, are you here?

09:10 <wwilly> or anybody from samsung with the magical manual for 5422?

10:31 <krzk> wwilly: I am here

10:38 <wwilly> Hi saw a mail about a cci devfreq for mt8183 https://lkml.org/lkml/2019/1/29/58

10:38 <wwilly> I would like to know, is that the same for cci of exynos5422?

10:38 <wwilly> I that possible to control the frequency of the cci? If not, what is the frequency of the cci in the 5422?

10:46 <krzk> CCI on e5422 seems to be supplied by VDD_MIF regulator

10:47 <wwilly> means cci frequency follow the memory frequency?

10:50 <krzk> Probably yes... the driver for e5422 does not expose/define CCI clocks so they are not explicitly controlled by devfreq driver

10:54 <wwilly> when you say the driver, you mean the actual def in linux? and what does it say in you magic manual? do you think it's possible to upgrade the kernel for supporting that?

10:55 <krzk> There are also async-bridges whose frequency might be changed by devfreq but they look like outside of CCI (async bridges are between many blocks and interconnect bus but not every memory bus has CCI).

10:57 <wwilly> do you have any public info about this async-bridges?

10:57 <krzk> Simplifying there are 3 interconnect memory busses but only one of them has CCI (the one CPU-mem)

10:57 <krzk> wwilly: there is no public info at all about anything in Samsung :) which is the answer why Exynos community is so small...

10:57 <wwilly> ok :/

10:58 <wwilly> is that possible to opt to something somewhere to have those stuff?

10:58 <krzk> wwilly: you can take a look at 94aed538e032d82287023beb9c59ee80f03a1493 or fa87bd4360ab4244467571f4235ccb2b362fea24

10:58 <krzk> but that is not much