#
#asahi-dev · 2026-05-15 12:19
tryhard[m]i saw dart-id = 6
#asahi-dev · 2026-05-15 12:18
chaos_princesswhy dart stream 6?
#asahi-dev · 2026-05-15 12:17
tryhard[m]https://paste.sh/4_L1uV0m#JkkPgbwctQDdJNr_Y9Zu8YHD
#asahi-dev · 2026-05-15 12:17
NickChan[m]^
#asahi-dev · 2026-05-15 12:17
chaos_princesssend your latest dts
#asahi-dev · 2026-05-15 12:16
tryhard[m]sent a code block: https://matrix.org/oftc/media/v1/media/download/AWsNON5JvLmJ6eyAhrUpMlTdA0UAeznlvV-hNdJE4kxchA_PBuqCZ1w7ZV1vQTvDm775MWXCdcXaFT6PMUhLXKVCeec6opXQAG1hdHJpeC5vcmcvbHVidGladUhZVlF0c0FIRWxDbXBNUER2
#asahi-dev · 2026-05-15 12:15
chaos_princessline?
#asahi-dev · 2026-05-15 12:14
tryhard[m]it's failing with the same error.
#asahi-dev · 2026-05-15 12:13
chaos_princessnames look ok
#asahi-dev · 2026-05-15 12:13
tryhard[m]yeah i mean the names
#asahi-dev · 2026-05-15 12:13
chaos_princessyours are not ascending
#asahi-dev · 2026-05-15 12:13
tryhard[m]NickChan[m]: i mean this
#asahi-dev · 2026-05-15 12:12
tryhard[m]basically the same?
#asahi-dev · 2026-05-15 12:12
chaos_princess^
#asahi-dev · 2026-05-15 12:12
NickChan[m]and for the interrupts put them in ascending order
#asahi-dev · 2026-05-15 12:12
NickChan[m]do "send-empty", "send-not-empty", "recv-empty", "recv-not-empty" for the names
#asahi-dev · 2026-05-15 12:11
tryhard[m]about the interrupt-names i'm not sure, i didn't include them at first but it complained that it couldn't find one of them so i put them in.
#asahi-dev · 2026-05-15 12:11
chaos_princessmacos and linux use different order
#asahi-dev · 2026-05-15 12:10
tryhard[m]chaos_princess: i took them from... (full message at <https://matrix.org/oftc/media/v1/media/download/AXY5yeeLp0Vbu6-S09o6ej_e6D-kXQOiV76vtJhr_QzB__gXngcWH9G_ghtwmyLXL3Bo8rylCjsEVOu_3W-5f9lCeec6Rz3wAG1hdHJpeC5vcmcveHlTb2xTdGphZEFrc1BDb2xNU29Od0xv>)
#asahi-dev · 2026-05-15 12:08
leiothere's https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31867 with rust bindings and things to write a mesa vulkan video driver.
#asahi-dev · 2026-05-15 12:08
chaos_princessinterrupts look wrong
#asahi-dev · 2026-05-15 12:06
tryhard[m]https://paste.sh/K0EHkBC1#YYwvq3QV9Y_yoXipMNHOOhcg
#asahi-dev · 2026-05-15 12:06
tryhard[m]i mean its failing again with error -62.
#asahi-dev · 2026-05-15 12:06
tryhard[m]actually i think i got the addresses wrong again lol
#asahi-dev · 2026-05-15 12:05
sofusleio: no, not all details. But its almost the same even tho my avd version is 4 (compared to 3)
#asahi-dev · 2026-05-15 11:57
chaos_princessshow your mbox node
#asahi-dev · 2026-05-15 11:57
JamesCalligeros[m]Be warned that the infrastructure around vk video is still rather immature so you may (will) end up writing mesa common code
#asahi-dev · 2026-05-15 11:56
tryhard[m]although i handle that in recv_message
#asahi-dev · 2026-05-15 11:55
tryhard[m]i get now like a spam log of RTKit: unknown management message: 0x0 (type: 0x00)
#asahi-dev · 2026-05-15 11:54
leiosofus: and you understand the details too?
#asahi-dev · 2026-05-15 11:54
tryhard[m]yeah that's the one i put now
#asahi-dev · 2026-05-15 11:52
NickChan[m]<tryhard[m]> "like u do this: pmp_mbox: mbox@2..." <- try 0x23BC08000 for pmp mailbox
#asahi-dev · 2026-05-15 11:52
chaos_princessgo download h264 conformance tests and start feeding them to those scripts :P
#asahi-dev · 2026-05-15 11:51
sofusleio: from what i understand the command stream seems stateless, what is there left to understand about it? It seems like eiln has figured out all the details
#asahi-dev · 2026-05-15 11:44
NickChan[m]always have been, though I think for t8103 it's just 0x2_0000_0000 for most things in /arm-io
#asahi-dev · 2026-05-15 11:40
tryhard[m]the addresses
#asahi-dev · 2026-05-15 11:40
tryhard[m]NickChan[m]: they're relative?
#asahi-dev · 2026-05-15 11:39
tryhard[m]chaos_princess: i mean yeah
#asahi-dev · 2026-05-15 11:39
NickChan[m]remember to take into account the ranges property of the parent node when dealing with the address
#asahi-dev · 2026-05-15 11:39
leio(but vulkan video is extensively documented in the specs)
#asahi-dev · 2026-05-15 11:38
leiosofus: that's the wrong angle of attack as I found out. We should start understanding the command stream, making sure it's fully stateless, maybe document it more, go from there
#asahi-dev · 2026-05-15 11:38
chaos_princessi guess? Don't you have an tracer that can confirm it?
#asahi-dev · 2026-05-15 11:37
sofusleio: do you know where can i read about vulkan video?
#asahi-dev · 2026-05-15 11:37
tryhard[m] reg = [Container:... (full message at <https://matrix.org/oftc/media/v1/media/download/AfKtfvxi8MsV67dxBtEz8aHP8Nvo6ZkSiUHq_BTJ-JGY3pxF0bC1ZaQbYtoQUkuLCT8MtM0kiz1z8E8HowYBxplCeec4YXsQAG1hdHJpeC5vcmcvcGVYa1pkQUZBT3JrbGNucVpXeWVYR2lh>)
#asahi-dev · 2026-05-15 11:35
chaos_princessquite likely
#asahi-dev · 2026-05-15 11:35
tryhard[m]when i do data.start(), i now get apple_pmp_v1 23bc00000.pmp: probe with driver apple_pmp_v1 failed with error -62. probably because i got the addresses wrong?
#asahi-dev · 2026-05-15 11:25
leiosofus: it isn't necessarily a bad fit, but if vulkan video works for it, then it's just nicer to do all of that in userspace instead in rust than C in v4l2.
#asahi-dev · 2026-05-15 11:13
sofuswhy is stateles v4l2 not a good fit?
#asahi-dev · 2026-05-15 11:13
sofusHi! I got avd to decode h264 on j416s using eiln's repo
#asahi-dev · 2026-05-15 11:10
tryhard[m]gotcha

plus ›