#
✕ nick:sven
#asahi-dev · 2022-11-01 13:23
svenyeah, I can’t reproduce it here on the mini either
#asahi-dev · 2022-11-01 13:20
svenamarioguy: imho the tps thing is a bug in the tps driver, I’d send out a v4
#asahi-dev · 2022-11-01 12:05
svenlooks like 144a96f7f96e412c3367f51cf2c57f52e3f1110d fixes my confusion. so essentially atcphy just needs "mode-switch" to support all altmodes
#asahi-dev · 2022-11-01 11:30
svenI actually need the hotplug callback for dcp now and soon also for thunderbolt
#asahi-dev · 2022-11-01 11:30
sventhat's why I just hacked it in atcphy-wip but now i'm at a point where that won't work nicely any longer :D
#asahi-dev · 2022-11-01 11:30
sventhe typec_mux stuff is a bit weird and i'm still confused by it. last time I looked at it it looked like there was a different mux for each mode or something
#asahi-dev · 2022-11-01 11:29
sveni think the last mails are pretty much "just use the type c connector" + they seem to introduce that mode-switch property
#asahi-dev · 2022-11-01 11:22
svenoh.. maybe I should read the entire thread and not just the original patch
#asahi-dev · 2022-11-01 11:21
svenand a graph connection to the DRM connector and then we can just use that oob_hotplug mechanism that already exists
#asahi-dev · 2022-11-01 11:21
svenjannau: hrm, i'm a bit confused by that switch stuff. there's already that typec connector binding, the typec mux mess and pretty much all that's missing is some property for the atcphy node that indicates it can mux dp-altmode,tbt, etc.
#asahi-dev · 2022-11-01 09:48
svenI finally can’t procrastinate looking at the typec subsystem anymore
#asahi-dev · 2022-11-01 09:47
svenjannau: have you already thought about how to represent multiple drm connectors in the dt?
#asahi-dev · 2022-11-01 08:36
svenusbms is simple enough, xhci will be a pain though :D
#asahi-dev · 2022-10-31 18:13
svenalright, seems to work reliable with delays at the correct positions now. at least for HBR.
#asahi-dev · 2022-10-31 15:25
svenpretty sure it's missing some important delays somewhere. just adding a mdelay(10) to every single register write seems to help a lot
#asahi-dev · 2022-10-31 15:24
svenIf I boot like 10 times it only works maybe 7 times
#asahi-dev · 2022-10-31 15:23
svensometimes
#asahi-dev · 2022-10-31 15:20
svennow it sometimes works :/
#asahi-dev · 2022-10-31 15:19
svenaaand it was something at least silly. clear vs. set in the crossbar for one register that's CLK_N_EN instead of CLK_EN.
#asahi-dev · 2022-10-31 13:40
sventhis is going to be something silly
#asahi-dev · 2022-10-31 13:40
svenlol, if I run the macos trace before I boot into linux and still run the atcphy.c it still works
#asahi-dev · 2022-10-31 13:38
svenhrm, let me check if there's something strange there
#asahi-dev · 2022-10-31 13:36
svenyes
#asahi-dev · 2022-10-31 13:36
sveni'm starting to suspect it's something very dumb
#asahi-dev · 2022-10-31 13:36
svenand random delays don't seem to fix anything.
#asahi-dev · 2022-10-31 13:36
sven*hv traces
#asahi-dev · 2022-10-31 13:35
sveni've traced both and diffed the hv looks and both do the same amount of reads
#asahi-dev · 2022-10-31 13:35
sveni'm pretty sure I can account for all macos reads
#asahi-dev · 2022-10-31 13:04
sveni wonder if there's some delay() missing somewhere that I just don't see in the MMIO trace but replaying it with python happens to be slow enough :(
#asahi-dev · 2022-10-31 13:04
svenurgh.. so the replayed xnu trace and my atcphy.c do the same pokes but the first one works and the latter one doesn't.
#asahi-dev · 2022-10-29 12:08
svenfive fixed magic MMIO pokes later and it’s still not working :(
#asahi-dev · 2022-10-29 06:17
svenI haven’t been able to reproduce it here but I think jannau can reproduce it reliably
#asahi-dev · 2022-10-28 15:23
svenI assume it somehow notices that the connection isn't stable and that's how it reacts
#asahi-dev · 2022-10-28 15:23
svenI think it happens when anything with the link rate PHY setup goes wrong
#asahi-dev · 2022-10-28 15:17
svenbah... atcphy.c seems to get stuck at color_mode = -1 for all modes again. wonder what I broke this time
#asahi-dev · 2022-10-28 15:12
svenusually I use the latest rc as a base for my submissions
#asahi-dev · 2022-10-28 15:11
svenjust make sure to e.g. not include the m2 parts already since I don't think those device trees are upstream yet ;)
#asahi-dev · 2022-10-28 15:10
sveni'd probably just send it then
#asahi-dev · 2022-10-28 15:09
svencan't hurt to send it to the mailing lists once you're confident enough that it works and/or had a few people test it
#asahi-dev · 2022-10-28 15:06
svenin this case it didn't probe because I added links to the phy and crossbar and messed up my probe code
#asahi-dev · 2022-10-28 14:54
sven[ 11.687257] dcp_platform_shutdown+0x14/0x20
#asahi-dev · 2022-10-28 14:54
sven[ 11.686318] iomfb_shutdown+0x40/0x118
#asahi-dev · 2022-10-28 14:54
sven[ 11.685447] dcp_push+0xd4/0x3b0
#asahi-dev · 2022-10-28 14:54
sven[ 11.684656] Call trace:
#asahi-dev · 2022-10-28 14:54
sven[ 11.646816] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000
#asahi-dev · 2022-10-28 14:53
svenjannau: if display-subsystem fails to probe but dcp itself probensthe kernel panics when you shutdown
#asahi-dev · 2022-10-27 15:13
svenah, true. forgot about that again
#asahi-dev · 2022-10-27 15:13
svenadds another month to the release date
#asahi-dev · 2022-10-27 15:12
svenfinally another one!
#asahi-dev · 2022-10-27 15:11
svenjust sent bluetooth v4 since v3 silently disappeared from patchwork...

plus ›