15:58 <aka_[m]> im experimenting around hfpll and i have question, how do we obtain requested rate in clk_hfpll_init_once ?

16:00 <aka_[m]> hmm

16:00 <aka_[m]> i could do it based on initial data

16:00 <aka_[m]> then set_rate will do anything else

16:02 <aka_[m]> ok i did some duckery there

16:02 <aka_[m]> i shouldn't reconfigure entire user_reg in set_rate

16:02 <aka_[m]> only toggle vco

16:03 <aka_[m]> or in init_once not even toggle vco, rather write stuff as-is

19:56 <aka_[m]> ok i think doing stuff dirty first and then working on unfolding stuff is better idea than risk crashing every consumer

22:40 <flto> bamse: was there a reason pil-squasher is trying to be clever and read chunks from the .mdt instead of always using the .b%02d file? (asking because it fails on certain cases because of that - silently generating incorrect output)

22:42 <bamse> flto: i have this naging feeling that some device i worked on has the signature duplicated in the .mdt and .b01, but only the prior is actually signed...

22:43 <bamse> flto: but i really would like to say "no"...

23:11 <lumag> flto, could you please send two first segments? like mdt / b00 /b01

23:11 <lumag> flto, or just the patch :D

23:33 <flto> lumag: https://github.com/andersson/pil-squasher/pull/8 attempted to make a patch that wouldn't break cases relying on loading hash from .mdt

23:34 <lumag> flto, ah, the newest format

23:34 <lumag> that makes sense