00:34 <bryanodonoghue> steev you need glink for data-role, orientation switching - I believe it all works in the linaro-laptop branch

00:35 <bryanodonoghue> spmi-pmic / tcpm will be for everything up go 83xx

00:35 <bryanodonoghue> or above

00:35 <steev> bryanodonoghue: hm... bamse had said that the phy needed something still, i'm not sure what linaro-laptop branch you're meaning though

00:35 <bryanodonoghue> sorry below

00:36 <bryanodonoghue> jhovold branch

00:36 <bryanodonoghue> https://github.com/jhovold/linux.git

00:37 <steev> as far as i know, orientation switching doesn't yet work for us on the thinkpad

00:37 <bryanodonoghue> ah right

00:37 <steev> definitely using his branch as a base

00:37 <bryanodonoghue> so bamse probably means the bringing of the QMP up and down for orientation switching

00:38 <bryanodonoghue> I think yes something like that is likely still necessary on 8280

00:38 <bryanodonoghue> so some version of lumag patch which I have in that series

00:38 <bryanodonoghue> but extended for 8280 and from the information we have here *glink

00:39 <bryanodonoghue> on 8250 a full phy reset is the downstream way

00:39 <bryanodonoghue> not sure how downstream does it for 8280/8350+

00:39 <bryanodonoghue> I guess the same way

01:59 <bamse> bryanodonoghue: right, except that i don't think "bring it down and up again" is the correct implementation

02:00 <bamse> bryanodonoghue: and no, we don't have a version of the orientation switching that works...i was hoping to bring in the old patch, but for some reason it didn't just work<tm>, and then y'all started hacking at it, so i didn't bother

08:28 <konradybcio> bamse: do we know what specifically causes it though? Maybe a full dump of the phy's reg range pre and post reset could show some magic register that we're missing that e.g. the bootloader could have set up differently for its usecases

13:56 <bamse> konradybcio: specifically causes it not to work? in my case it was probably because i didn't bring the phy back up properly...in some way

13:57 <konradybcio> yeah.. I'd expect that it may be brought up in some "simple" mode early on

13:58 <bamse> konradybcio: no, it's stated in the documentation that after modifying the signal routing properties (and a few other events), initialization needs to be performed again

13:58 <bamse> konradybcio: so it works the first time because we bring up the phy

13:59 <bamse> konradybcio: and what i'm saying above is that when i quickly ported the old patch that made orientation switching work, it didn't work...(it used to work)...so i must have screwed something up

13:59 <konradybcio> ah i misread the msg history