#asahi-dev · 2023-01-23 06:52
marcanjannau: re the devm free, dunno if you saw Lina is adding non-devm functions for Rust. maybe you can just use that?#asahi-dev · 2023-01-23 06:49
marcanit's entirely possible existing hardware can do MST and it was just missing firmware support too#asahi-dev · 2023-01-23 06:49
marcanI assumed on 12.3 it just worked using tiling, virtually cutting the one output canvas into 2 pipes, which is not ideal. But you'd think under the hood the 2 pipes are independent and could reference completely different layers/surfaces, which would be a lot nicer to actually use them separately.#asahi-dev · 2023-01-23 06:48
marcanI also wonder if it means we might get to abuse the 2 pipes separately on existing hardware#asahi-dev · 2023-01-23 06:48
marcanMST would be interesting#asahi-dev · 2023-01-23 06:47
marcanof course you only get the top level out of that, individual types can have custom marshalling rules and that is not something strings will show#asahi-dev · 2023-01-23 06:46
marcanyeah, at least double check the prototype, it'll tell you if it's outright a new argument (or if it changed to an array or something)#asahi-dev · 2023-01-23 06:44
marcanwe can move to -re if you need to do more than that, but basically I'm not worried about looking into the DCP driver to get marshalling/interface info because it's just mandatory to make this work, effectively becoming part of the hardware interface#asahi-dev · 2023-01-23 06:43
marcanbut other than that yeah#asahi-dev · 2023-01-23 06:43
marcanyou'll need to 's/^_//' to get c++filt to like it#asahi-dev · 2023-01-23 06:41
marcanyou can just strings the driver and c++filt it to get the signatures for the calls IIRC (callbacks are a bit trickier)#asahi-dev · 2023-01-23 06:41
marcanhm, interesting#asahi-dev · 2023-01-23 06:41
marcanyes#asahi-dev · 2023-01-23 06:37
marcanBut I think I added that? :p#asahi-dev · 2023-01-23 06:37
marcanI meant idle as well#asahi-dev · 2023-01-23 03:22
marcanjannau: I'm also going to push some of lina's RTKit changes to asahi-soc and send a pull soonish. If you want to get your DCP-relevant changes in too, feel free to send those out and they can all go into the same batch :)#asahi-dev · 2023-01-22 16:38
marcanso I think systemd is probably triggering that immediately when the backlight device appears#asahi-dev · 2023-01-22 16:38
marcan[ 1.346562] apple-dcp 38bc00000.dcp: dcp_hotplug: connected: 1#asahi-dev · 2023-01-22 16:38
marcan[ 1.346558] apple-dcp 38bc00000.dcp: DCP is_main_display: 1#asahi-dev · 2023-01-22 16:38
marcan[ 1.346268] apple-dcp 38bc00000.dcp: DCP booted#asahi-dev · 2023-01-22 16:38
marcanthis all happens before:#asahi-dev · 2023-01-22 16:37
marcanfull trace: https://mrcn.st/p/W4Huoggy#asahi-dev · 2023-01-22 16:37
marcanor something like that#asahi-dev · 2023-01-22 16:37
marcanyeah but I think the issue is that there hasn't been a swap yet, so we don't have any atomic state#asahi-dev · 2023-01-22 16:36
marcananyway, not a big deal, just fyi#asahi-dev · 2023-01-22 16:36
marcanI'm guessing we might need to just cache backlight values somewhere before this stuff is ready?#asahi-dev · 2023-01-22 16:35
marcanthis happens before dcp_poweron even#asahi-dev · 2023-01-22 16:34
marcanWARN_ON(!crtc->state) is what is firing#asahi-dev · 2023-01-22 16:33
marcanbacklight itself works in KDE though#asahi-dev · 2023-01-22 16:33
marcanjannau: backlight restore by systemd fails#asahi-dev · 2023-01-22 16:33
marcan[ 1.326413] WARNING: CPU: 4 PID: 460 at drivers/gpu/drm/drm_atomic_state_helper.c:174 drm_atomic_helper_crtc_duplicate_state+0x58/0x74#asahi-dev · 2023-01-22 16:16
marcanbut anyway, doing a build with what we have now before going to sleep#asahi-dev · 2023-01-22 16:14
marcanwhether something else makes it in there (audio? DP?) depends on timing at that point, TBD#asahi-dev · 2023-01-22 16:14
marcani.e. to non-edge#asahi-dev · 2023-01-22 16:13
marcanand then when the M2 Pro/Max stuff is ready, release it together with promoting -edge stuff to stable#asahi-dev · 2023-01-22 16:12
marcanI'm thinking I might merge lina's new UAPI/rebase stuff plus rebase on the latest rc, and then after some smoke testing of that do a release to stable to see if we regress anything (which really should only affect -edge at that point)#asahi-dev · 2023-01-22 16:08
marcanok, I'll do a -dev release with this and m1n1#asahi-dev · 2023-01-22 16:06
marcanyeah, got it#asahi-dev · 2023-01-22 16:05
marcanmakes sense#asahi-dev · 2023-01-22 16:05
marcandcp_wait_ready() right#asahi-dev · 2023-01-22 16:05
marcanah, I guess you just start all DCPs in the callback when the components are ready?#asahi-dev · 2023-01-22 16:03
marcanjannau: how is the race solved in the end?#asahi-dev · 2023-01-22 16:01
marcanooh nice#asahi-dev · 2023-01-22 16:01
marcanjannau: what are your plans for 13.2?#asahi-dev · 2023-01-22 16:00
marcanthe previous one though, yeah (when I was missing the IRQ-safe patch)#asahi-dev · 2023-01-22 16:00
marcanoh this one wasn't causing any badness, the rpm code just re-incremented the count so it was just cosmetic#asahi-dev · 2023-01-22 15:54
marcanbut at least this works now#asahi-dev · 2023-01-22 15:54
marcanhonestly the mechanism I used there is kind of flaky and should probably be rethought#asahi-dev · 2023-01-22 15:53
marcanjannau: fixed now#asahi-dev · 2023-01-22 15:52
marcanwait I messed up