14:14 <mszyprow|home> wwilly: ping

14:15 <mszyprow|home> wwilly: i think I've found somthing related to the usb/lan issues on XU3. Could you check if removing all regulator-coupled-* properties from the exynos5422-odroid-core.dtsi fixes it?

15:11 <wwilly> mszyprow|home: hi

15:11 <wwilly> will check

15:12 <wwilly> I rafine the revert patch

15:12 <wwilly> I forgot to send

15:12 <wwilly> I check what I did

15:13 <wwilly> mszyprow|home: what's the problem you found? inpacting something else?

15:28 <wwilly> mszyprow|home: you mean regulator-coupled-* under buck2_reg: BUCK2 and buck3_reg: BUCK3 ?

15:40 <mszyprow|home> wwilly: yes

15:41 <mszyprow|home> wwilly: i managed to reproduce some usb issues on xu3 with usb pendrive

15:41 <mszyprow|home> wwilly: i wonder if this is the same you got

15:42 <mszyprow|home> wwilly: disabling regulator coupling seems to fix my issues

15:42 <mszyprow|home> wwilly: no idea why yet

15:43 <wwilly> ok

15:47 <wwilly> with which version you have issue?

16:07 <wwilly_> ping??

18:16 <mszyprow|home> wwilly: v5.6

19:01 <mszyprow|home> wwilly: but I've just checked and it is same for current linux-next (20200526)

19:22 <mszyprow|home> wwilly: alternatively you can change the regulator-min-microvolt to <925000> for buck3_reg (vdd_int)

19:22 <mszyprow|home> wwilly: both seems to be fixing my issue

19:22 <mszyprow|home> wwilly: I only need to figure out the bug in coupling code which lowers the voltage too much

19:23 <wwilly> I will check that tommorow, I was a bit busy right now

19:23 <mszyprow|home> wwilly: okay

19:23 <wwilly> but what I change just to have it working without full revert was https://pastebin.com/UM6Z3XSa

19:24 <mszyprow|home> wwilly: what about the issue with ondemand gouvernor and dmc devfreq?

19:25 <wwilly> I don't know why I was needing opp-shared, and keep operating-points-v2 = <&bus_fsys_apb_opp_table>; in bus_fsys

19:25 <mszyprow|home> wwilly: you mentioned that it waits for a paper acceptance

19:25 <wwilly> arf no, this is about lan crashing

19:26 <mszyprow|home> wwilly: i've checked the clocks once again and also copared your clock dump

19:26 <wwilly> yes, the paper is accepted, and I currently busy recording the virtual talk because of lockdown

19:26 <mszyprow|home> wwilly: any that change only affects devfreq

19:26 <mszyprow|home> wwilly: but it might be enough to change it in a way that the regulator balancing is done right

19:27 <wwilly> but, I will provide everything by next week, need to be properly ready release for the conference you see

19:27 <mszyprow|home> wwilly: but the source is in the regulator balancing code

19:27 <mszyprow|home> wwilly: one needs to trigger it somehow

19:27 <mszyprow|home> wwilly: your patch hides it

19:28 <wwilly> hide the problem you mean, without fixing it? so why it's running buy just hiding it??

19:29 <wwilly> ok, let's discuss tomorrow, family stuff, I'm already late :)

19:29 <wwilly> c u toms to discuss plenty

19:29 <mszyprow|home> wwilly: i will aswer that question once i find the bug :) now i just found what to change to really avoid it