#asahi-dev · 2026-07-17 18:05
yukaM4 Pro (t6040, j773s) irq num in the device tree is also wrong#asahi-dev · 2026-07-17 17:53
yuka✅️ wrote an experiment for determining dockchannel-uart irq num#asahi-dev · 2026-07-17 17:23
svenwe'll see, i have to finalize some tax related stuff before i can get money from that opencollective for hw, etc. anyway #asahi-dev · 2026-07-17 17:22
svenyeah, there should be some clock divider or something like that somewhere#asahi-dev · 2026-07-17 17:22
opticronyep, most of the esp32 chips have something along those lines that has a USB/UART bridge peripheral fused with a USB/JTAG peripheral#asahi-dev · 2026-07-17 17:21
yukathe dockchannel uart I'm currently working on is such an instance#asahi-dev · 2026-07-17 17:20
opticron(because there's no actual serial lines where the data is flowing, the peripheral just spits the data out as USB packets)#asahi-dev · 2026-07-17 17:19
opticronbut there are also UART->USB bridges embedded as peripherals on some parts that don't expose the clock dividers because that just doesn't make sense#asahi-dev · 2026-07-17 17:19
yukafair#asahi-dev · 2026-07-17 17:18
opticronwith every real UART peripheral I've ever touched#asahi-dev · 2026-07-17 17:18
yukaI don't really know at which granularity it works#asahi-dev · 2026-07-17 17:18
yukayou mean it can be adjusted in software (if it is baudrate)?#asahi-dev · 2026-07-17 17:17
opticronknowing exactly how off it is helps you figure out where the dividers are wrong sometimes#asahi-dev · 2026-07-17 17:17
yukabut unless we can actually fix the baudrate/whatever compatibility issue, it's a bit pointless, no?#asahi-dev · 2026-07-17 17:16
yukait's probably possible to figure that out#asahi-dev · 2026-07-17 17:16
yukaSo until now every time I needed a linux serial console, I commented out the samsung_tty init, and then it would work but then stuff wouldn't be set up correctly and I had to spam a lot of keys to get the console to refresh#asahi-dev · 2026-07-17 17:15
yukayes, but don't forget the linux-side init also changed and now breaks the serial on t8140#asahi-dev · 2026-07-17 17:06
sveni suspects it's just some weirdnesss with the baudrate or something similar#asahi-dev · 2026-07-17 17:05
sveni really have to get some of those devices with that broken debug-uart and hook them up to an oscilloscope #asahi-dev · 2026-07-17 17:05
yukaI need to clean it up a bit and add support for interrupt-based TX, but this is already very useful for these newer devices which have broken debug-uarts c:#asahi-dev · 2026-07-17 17:04
yukaand on t6050 too!#asahi-dev · 2026-07-17 17:00
yukadcuart as linux console works on t8140 🎉#asahi-dev · 2026-07-17 16:51
yukasomeone needs to run it on t602x too, since there is no interrupt specified in the adt#asahi-dev · 2026-07-17 16:51
yukaI will try to write a tool to find the correct interrupt num based on aic HWSTATE#asahi-dev · 2026-07-17 16:43
yukaThe t8140 one wasn't wrong in the adt, but I mistyped when copying it to the device tree#asahi-dev · 2026-07-17 16:33
yukafor reference, I meant 702 is the irq num for M5 Pro (t6050), not M5 (t8142). But since I don't have a base M5 device, I sometimes think of the M5 pro as the "M5 (class)" device ;)#asahi-dev · 2026-07-17 16:30
svenlol, i've said a few times that the ADT lies in here but I didn't expect it to be _that_ bad#asahi-dev · 2026-07-17 16:29
yukathis is going to be fun... validating each individual SoC's irq nums#asahi-dev · 2026-07-17 16:29
yukaadt says 360 but it's 702#asahi-dev · 2026-07-17 16:29
yukait's just wrong in the adt#asahi-dev · 2026-07-17 16:29
yukafor M5#asahi-dev · 2026-07-17 16:29
yukafound the irq num#asahi-dev · 2026-07-17 16:29
yukaOh!#asahi-dev · 2026-07-17 16:22
yukado we know of any situation where mac os actually uses the dockchannel uart with interrupts?#asahi-dev · 2026-07-17 16:21
yukathe config area differs in two bits, which I can't write#asahi-dev · 2026-07-17 16:20
yukawhere t8132 immediately throws interrupts once unmasked and t6050 I've never seen an interrupt being thrown#asahi-dev · 2026-07-17 16:20
yukaI diffed the mmio areas (that I know of) of dockchannel-uart on t8132 vs t6050#asahi-dev · 2026-07-17 16:17
Carlos_alright i will rebuild from the updated tree and continue#asahi-dev · 2026-07-17 16:16
Carlos_right#asahi-dev · 2026-07-17 16:16
chaos_princessby seeing that you mentioned j516, and then actually paying attention when reading the logs#asahi-dev · 2026-07-17 16:15
Carlos_how did you catch that i have an m3 pro instead btw?#asahi-dev · 2026-07-17 16:15
chaos_princessi force-pushed new m1n1 and kernel trees that assign the data that i got to the correct devices#asahi-dev · 2026-07-17 16:14
Carlos_chaos_princess: my mistake, yeah m3 pro. does that invalidate the tests you have me doing or still continue#asahi-dev · 2026-07-17 08:48
JamesCalligeros[m]It doesnt crash anymore though so that’s progress :p#asahi-dev · 2026-07-17 08:47
fl0_id_afkCarlos utm (or similar is easier) you might run into weird filesystem shenanigans otherwise (or don't use bind mounts)#asahi-dev · 2026-07-17 08:47
JamesCalligeros[m]Havent looked at dpavserv yet, edid copy does not work but that may be because the thing is wedged when it fails to create the dfb#asahi-dev · 2026-07-17 08:46
JamesCalligeros[m]Im hoping its just something silly. We get the timing elements etc back and then nothing happens. The failure to create the dfb is sus to me#asahi-dev · 2026-07-17 08:43
JamesCalligeros[m]Sweet#asahi-dev · 2026-07-17 08:43
jannauI was planning to do that anyway#asahi-dev · 2026-07-17 08:42
jannauchadmed: I can look after 14.8.3 dcp from middle of next week when I'm back from vacation