#asahi-dev · 2023-08-29 11:55
svennot very familiar with u-boot but it seems to be related to getting the right EFI partition when multiple OSes are installed#asahi-dev · 2023-08-29 11:53
sventhat commit is probably there for a reason, blindly reverting it isn't a good idea#asahi-dev · 2023-08-27 22:13
sven*stops#asahi-dev · 2023-08-27 22:13
svenSMC was the easiest coproc to „shut down“ because it never fully shuts down but just stop communicating and because it manages its own buffers inside its SRAM anyway #asahi-dev · 2023-08-27 22:12
sveniirc SMC shutdown is just that usual sleep message with state 0x10 or so#asahi-dev · 2023-08-27 22:11
svennice!#asahi-dev · 2023-08-26 22:00
svenwhat helped was dumping the entire ringbuffer at every command and reply #asahi-dev · 2023-08-26 22:00
sveni think it was because the Python code (or the C code) used uninitialized memory which messed something up or something weird like that #asahi-dev · 2023-08-26 21:58
svenI remember having an issue like that when I first brought up dptx #asahi-dev · 2023-08-23 19:02
svenjust speculation though :)#asahi-dev · 2023-08-23 19:02
svenpresumably so that they can use it everywhere #asahi-dev · 2023-08-23 19:01
sveniirc it also contains some core pcie stuff that they just reimplemented for sone reason #asahi-dev · 2023-08-23 19:01
svenit looks a bit like that #asahi-dev · 2023-08-23 14:19
sveniirc the only problem with bcmhd is that they invented their own hardware abstraction layers which is why they couldn't upstream it#asahi-dev · 2023-08-23 13:48
svenyup, and that's more recent than some of the stuff inside the upstream driver 🙃#asahi-dev · 2023-08-23 13:46
svenand the source should be in android trees#asahi-dev · 2023-08-23 13:46
sventhat's their android driver#asahi-dev · 2023-08-23 13:46
svenbcmdhd, yes#asahi-dev · 2023-08-23 08:56
svenlet's see if it keeps working :)#asahi-dev · 2023-08-23 08:56
sventhe soldering also doesn't look bad and they probably reflowed the board anyway instead of hand soldering it so dunno#asahi-dev · 2023-08-23 08:55
sventhough I got these usb c ones from lcsc as well which I think is where jlcpcb sources everything as well#asahi-dev · 2023-08-23 08:54
sveneither shitty connectors or bad solder joints I guess :/#asahi-dev · 2023-08-23 07:07
svenmaz: okay, so I removed both usb connectors now, cleaned the pads and soldered down new ones and it seems to be working now #asahi-dev · 2023-08-22 20:31
svenso far that wasn’t required ;)#asahi-dev · 2023-08-22 20:21
sventhat’s why I switched mine to atcphy-YYYYMMDD :D#asahi-dev · 2023-08-21 14:33
svenbecause the dptx stuff seems to be the same then #asahi-dev · 2023-08-21 14:33
svenI wonder if that also works with atcphy then #asahi-dev · 2023-08-20 18:01
svenokay, this makes no sense. How did I break usb3 by _removing_ thunderbolt patches?!#asahi-dev · 2023-08-20 15:32
sventhat too#asahi-dev · 2023-08-20 15:32
svenyeah, not sure how to see arrogance in that reply#asahi-dev · 2023-08-20 15:05
svennah, no bridges#asahi-dev · 2023-08-20 15:01
svenlet me check again to make sure#asahi-dev · 2023-08-20 15:00
svenI didn't see D+ or D- bridged to anything else#asahi-dev · 2023-08-20 15:00
sveni checked for stuff like that#asahi-dev · 2023-08-20 14:56
svenno idea, was about to ask maz :D#asahi-dev · 2023-08-20 14:56
sveni only checked for obvious shorts between d+/d-, if gnd was connected, etc.#asahi-dev · 2023-08-20 14:55
sveni've cleaned up the connectors and i don't think there's anything else in the signal path on the pcb#asahi-dev · 2023-08-20 14:54
svenyeah, i'm surprised it falls back to full speed as well#asahi-dev · 2023-08-20 14:54
svenprobably makes sense to try the test points first :D#asahi-dev · 2023-08-20 14:51
svenhm.. i could try cutting the traces and bypassing the pcb with a twisted pair. if it still fails at high speed then it's the connectors. if it works then it's something else on the pcb#asahi-dev · 2023-08-20 12:44
svencutting traces was the first plan before i realized i can just sacrifice a cable instead#asahi-dev · 2023-08-20 12:44
sveni split the cable off before#asahi-dev · 2023-08-20 12:44
sventhe pcb isn't in the signal path at all for d+/d- now#asahi-dev · 2023-08-20 12:38
svenmaz: it's been stable the entire time now. pretty sure something on the pcb is the issue fwiw#asahi-dev · 2023-08-20 10:11
svenwould be interesting to see what's actually going on#asahi-dev · 2023-08-20 10:10
svenI unfortunately don't have a scope and even if I had one I doubt it would be fast enough to actually look at high speed D+/D- or capture eye diagrams#asahi-dev · 2023-08-20 10:03
svenseems to work reliably now. i guess something on the pcb degrades the signal just enough for high-speed to sometimes fail