00:47 <Dylanger> Spin 513 ordered 💦

00:53 <qzed> bamse: finally managed to get a log for that one oops

00:53 <qzed> https://gist.github.com/qzed/2288c651a017fcc016b524d3bac5a43e#file-ooop-dmesg-log

00:54 <qzed> something's going wrong in dp_display_request_irq(), which causes the probe to abort and that then causes the oops somehow

02:51 <bamse> qzed: seems like another issue with doing a full cleanup after a partital initialization

02:52 <qzed> yeah, somehow the IRQ doesn't get freed properly, and as it's not marked as shared it complains when the driver tries to request it the second time around

02:53 <qzed> at leas that's my curren theory

02:56 <bamse> i think we should refactor the whole thing to perform as much of the operations that can return EPROBE_DEFER already at probe()...before we bind and start initialize things

02:57 <bamse> the case with the aux-bus is still tricky, but i think it would be cleaner still

02:58 <qzed> makes sense, will probably be a bit of work but right now I feel like I'm trying to barely make one thing work to just find the next thing that breaks...

02:58 <bamse> i had some case the other day where next_bridge was -517, but somehow on the way back we lost that and would return some other error and not attempt the probe (and bind) again

02:59 <qzed> unfortunately I still don't quite have a good overview of everythign that's going on in there

02:59 <bamse> debugged it for 30 minutes and then it went away...

02:59 <bamse> we probe() all the pieces of mdss, then when all the components are in place we bind the components...and if the bind returns -517 we unbind everything and unprobe all the components

03:00 <bamse> so we bring up and tear down quite a bit of pieces by not resolving these external dependencies until bind()

03:00 <qzed> it's a bit annoying as I hoped I could just avoid this problem by building the panel in, but sometimes the driver still doesn't want to load (-EIO) and the whole thing defers again, bringing that issue up

03:00 <qzed> right

05:11 <Dylanger> Is the Acer Chromebook Spin 513 Cherry or Pumpkin?

05:12 <Dylanger> https://github.com/torvalds/linux/blob/master/arch/arm64/boot/dts/mediatek/mt8183-pumpkin.dts

05:12 <Dylanger> Looks like there's already a pumpkin DTS in mainline

13:14 <qzed> does anyone know what modules I need for USB/DP stuff?

13:15 <qzed> I think I should have everything implemented in the DT, but the display isn't recognized at all, so I guess I'm missing some config options

14:40 <bamse> qzed: are you trying to say "i'm connecting the cable, the type-c port manager notices that, but the displayport driver doesn't react"?

15:58 <qzed> bamse: something like that

15:58 <qzed> I have a hub with hdmi out (unfortunately no pure display-port adapter), which is connected

15:58 <qzed> but the display doesn't show up in debugfs nor is there any other indication that there's a display hooked up to it

15:59 <bamse> so you have a DP-x, but it's disconnected state?

16:02 <qzed> yeah, so I have one for eDP, which shows up, and 2 for usb-c

16:02 <qzed> and neither of the ones that should be usb-c is connected

16:04 <bamse> okay, so then what you're lacking is the notification from the port manager to the displayport driver about HPD events

16:05 <bamse> i have some patches on the list for that, but i discovered this week that they don't work

16:05 <bamse> i mean, they work...but then they stop working

16:06 <bamse> so i'm reworking that currently

16:06 <qzed> ah, okay, maybe I'll postpone trying to get that to work then... thought you'd already got it to work on the flex

16:07 <bamse> i have it working, but i don't have any patches that works well and applies to the latest kernel versions

16:07 <qzed> I'm trying to get most things semi-workable now and then maybe to a proper install on the internal SSD

16:07 <qzed> ah, got it

16:07 <qzed> I guess I'll try my luck with wifi again then

16:07 <bamse> dmitry did some refactoring wrt to how the dp driver deals with drm_connector and drm_bridge

16:08 <bamse> so i'm adapting to that

16:08 <qzed> ah, neat

20:01 <jenneron> macc24: Hi. Which Chrome OS devices do you have? i currently work on veyron chromebooks for pmOS, if you have one (except minnie and speedy), i would ask to test something if you don't mind

22:05 <steev> i always meant to get a speedy and never got around to it