#asahi-dev · 2022-11-26 14:13
sventhe hard part is as usual to get the order between all the moving components right #asahi-dev · 2022-11-26 14:13
sventhat’s the easy part#asahi-dev · 2022-11-26 13:21
svenyes please!#asahi-dev · 2022-11-26 13:19
svenwhile usb specs are probably harmful to anyone who reads them and then has to live with that knowledge #asahi-dev · 2022-11-26 13:19
sven(Usb was more pain to be clear)#asahi-dev · 2022-11-26 13:18
sveneven the spec is readable by humans #asahi-dev · 2022-11-26 13:18
svennvme is like the opposite of usb#asahi-dev · 2022-11-26 13:18
svenusb obviously :D#asahi-dev · 2022-11-26 10:03
svenmaybe just needs some additional EPIC calls but i'd be surprised if it's very different#asahi-dev · 2022-11-26 10:02
svenso chances are if audio output over the mac mini's hdmi works it might also just work over the type-c ports :)#asahi-dev · 2022-11-26 07:57
svenshouldn’t type before coffee #asahi-dev · 2022-11-26 07:57
svenI meant the internal screen is hooked up via DP and the HDMI output on the mini goes through a converter #asahi-dev · 2022-11-26 07:56
svenuh. Never mind the last comment #asahi-dev · 2022-11-26 07:29
sveneven the internal MacBook screen is hooked up through a DP to hdmi converter even though it doesn’t even use type c #asahi-dev · 2022-11-26 07:28
svenpovik: almost nothing uses HDMI altmode. It’s always DP altmode plus some DP to hdmi converter #asahi-dev · 2022-11-25 22:06
svenif text is mapped via dart I don’t get why they have to lock it anymore #asahi-dev · 2022-11-25 22:06
svenweird.#asahi-dev · 2022-11-25 22:05
svenpovik: maybe it’s tongue-in-cheek :D#asahi-dev · 2022-11-25 22:05
svenand that dart is still locked?#asahi-dev · 2022-11-25 22:05
svenhuh, interesting. #asahi-dev · 2022-11-25 22:03
sven*our#asahi-dev · 2022-11-25 22:03
svenotherwise we could’ve just written out own fake-dma firmware :D#asahi-dev · 2022-11-25 22:03
svenok, that would’ve been really surprising :)#asahi-dev · 2022-11-25 22:02
svenit maps text via dart?#asahi-dev · 2022-11-25 21:42
svenand I think that’s all we support #asahi-dev · 2022-11-25 21:42
svenSMC has everything in its internal RAM as well#asahi-dev · 2022-11-25 21:42
svenand ANS has the data section mapped via SART#asahi-dev · 2022-11-25 21:42
svenfor DCP the PTEs already exist #asahi-dev · 2022-11-25 21:41
svenwhy would it?#asahi-dev · 2022-11-25 21:41
svenI think AOP is the exception because it probably has to work even when RAM doesn’t #asahi-dev · 2022-11-25 21:40
svenit’s only not used for DCP because that dart is locked and pre filled iirc #asahi-dev · 2022-11-25 21:39
svensegment ranges should be the source #asahi-dev · 2022-11-25 21:39
svendata section inside dart is pretty common #asahi-dev · 2022-11-25 20:45
svenyeah, probably #asahi-dev · 2022-11-25 20:44
svenI guess we could create a single threaded workqueue and defer it to there with a newly allocated work_struct each time but ugh :(#asahi-dev · 2022-11-25 20:43
svenlooks good to me#asahi-dev · 2022-11-25 20:43
svenso then we just have to find a way to fix it properly even when dwc3_set_mode is called from atomic context #asahi-dev · 2022-11-25 20:33
sven*mutex#asahi-dev · 2022-11-25 20:31
svenlooks like that additional muted also fixed the rare DP-altmode transition bugs I saw #asahi-dev · 2022-11-25 15:57
svenhttps://files.svenpeter.dev/41f708cd7022c9998b85fcde0e236ad93a9154e2f011fe5e6edc10b864b232fd_dwc_race.txt#asahi-dev · 2022-11-25 15:55
svenbut you then end up with PRTCAP = host and dwc3 bringing up gadget or vice versa#asahi-dev · 2022-11-25 15:54
svenbut that thing is incredibly right. I only managed to trigger it twice#asahi-dev · 2022-11-25 15:54
svenbut without it you can have one dwc->desired_dr_role value in dwc3_set_prtcap(dwc, dwc->desired_dr_role); and then three lines or so below a different one in switch (dwc->desired_dr_role)#asahi-dev · 2022-11-25 15:53
svenit's much easier to hit that race with the role switch reset quirk#asahi-dev · 2022-11-25 15:53
svenyou need a very quick A->B->A role change event#asahi-dev · 2022-11-25 15:48
sventhanks. fixing that race correctly is going to be annoying #asahi-dev · 2022-11-25 15:40
svennot 100% sure because reproducing it has always been flakey for me#asahi-dev · 2022-11-25 15:39
svenjannau: alright, looks like I can't reproduce the WARN anymore if I lock dwc->mutex around dwc->lock in dwc3_set_mode with that atcphy-WIP tree#asahi-dev · 2022-11-25 14:47
svennot sure if anything like that even exists and that also seems to be unrelated to cpufreq i think