Comment # 21
on bug 73338
from Chernovsky Oleg
(In reply to comment #20)
> Oleg, thanks a lot!
>
> Alex (or anyone else, particularly if you're from AMD): would it be possible
> to point Oleg to the relevant documentation if it exists, or at least find
> out whether it can be published in the near future (if it doesn't exist
> publicly yet), so that it doesn't have to reverse-engineered from scratch?
> Many thanks for any helpful info.
Actually I already come to some conclusions. First SMC writes and reads (before
SMC_MESSAGE) is just the "ci_check_smc_running" function (seems its analogue
presents in proprietary driver as well).
0x5c value should be some sort of command, so it can be defined with already
known ones in headers.
About other values things are not so clear. I suppose to get real fan speed we
need to somehow mask and shift data at smc index 0xc0300064.
I'll try to make other traces with different parameters to compare percentages
and values to the register reads and writes.
P.S. yep, some docs would be helpful, thanks