#asahi-dev · 2025-12-23 13:20
jannauwhat do we need to pass on as metadata?#asahi-dev · 2025-12-23 13:20
jannauyou could look at dcp's tracing#asahi-dev · 2025-12-23 13:19
jannauif you miss vblank it waits until the next one which is is usually more than 5s#asahi-dev · 2025-12-23 13:15
chadmedand even so theres seriously no way a page flip takes 5ms... right?#asahi-dev · 2025-12-23 13:15
chadmedi dont think the timing hack is the issue here. i tried handling cases up to 5ms between swap_start and now and nothing really helped#asahi-dev · 2025-12-23 12:22
sofusi think it only works on m2 devices#asahi-dev · 2025-12-23 12:19
sofushttps://github.com/AsahiLinux/linux/commit/3339f7a049f4ed359d757d44c0c976c6b8cfb402#asahi-dev · 2025-12-23 12:18
sofusHi, i made a small driver for the actuator, with it i can send commands that change the up/down haptics, trigger a haptic event and a "click handoff" event#asahi-dev · 2025-12-23 12:09
chadmedin more positive news it does seem like dcp supports the agx tiled formats after all, but i havent got any good ideas for setting the required metadata in a way that isnt awful #asahi-dev · 2025-12-23 11:49
chadmedyeah and kwin just reduced its own #asahi-dev · 2025-12-23 11:48
jannauthis is required because every compositor has its on guess on when exactly before the page flip it needs to commit the update. There are talks of adding a hint to the KMS interface so that the driver can communicate its guess#asahi-dev · 2025-12-23 11:38
chadmedthis is uhhh#asahi-dev · 2025-12-23 11:38
chadmedoh yeah i see the thing in dcp.c #asahi-dev · 2025-12-23 11:37
jannauDRM_CAP_ATOMIC_ASYNC_PAGE_FLIP might help (especially with cursor updates) but implementing it could be hard and require further dcp RE#asahi-dev · 2025-12-23 11:36
jannauthat's not how atomic commits work and shouldn't matter as compositors base their timing on the page flip event and not on when their commit returns#asahi-dev · 2025-12-23 11:36
mazfl0_id: I've just pushed a branch updated to SDK 2.2.0. on a sample of one, it seems to work fine.#asahi-dev · 2025-12-23 11:28
chadmedwe never send a vblank unless a swap fails and i cant see anything else time-sensitive userspace should care about#asahi-dev · 2025-12-23 11:27
chadmedive made it asynchronous so that atomic_flush notifies that its done after putting the iomfb swap request on a workqueue but that didnt help#asahi-dev · 2025-12-23 11:22
jannauI haven't seen that but I only tested briefly with firefox/testufo. kwin was fine but cosmic-comp didn't get 120Hz. we might need to adjust the hacky page flip time hackery#asahi-dev · 2025-12-23 10:08
chadmedthis does not happen on external monitors#asahi-dev · 2025-12-23 10:07
chadmedaquarium never goes above 60 and it takes ~10-15 seconds for the gears to get up to 120 fps#asahi-dev · 2025-12-23 10:06
chaos_princesshuh#asahi-dev · 2025-12-23 10:06
chadmedwebgl aquarium, glxgears, vkgears#asahi-dev · 2025-12-23 10:05
chaos_princessare you doing something or just sitting at desktop?#asahi-dev · 2025-12-23 10:04
chadmedive made atomic_flush non-blocking and it helps with pacing but kwin is still sitting at 60hz until it decides not to #asahi-dev · 2025-12-23 10:03
chadmedtheres still something up with the 14/16" machines not reliably running at 120 hz. kwin still wants to sit at 60 hz but will intermittently go up to 120#asahi-dev · 2025-12-23 08:21
jannauunrelated to the color space set for the surface. if the CTM applies post blending and dcp uses a differrent colorspace post blending it might not have the desired effect (but should not cause a green tint in any case)#asahi-dev · 2025-12-23 08:13
chadmedi havent tried explicitly setting a colourspace for our surfaces other than whatever "native" is#asahi-dev · 2025-12-23 08:10
jannauI don't test it often and have it disabled but in tests it always made the screen orange/red with higher color temperatures. I'm just not sure if it it the right amount since I'm unsure when and in which color space the ctm is applied#asahi-dev · 2025-12-23 08:06
chadmedit does this intermittently so we might not be the problem #asahi-dev · 2025-12-23 08:06
chadmedwell it should at the very minimum make my screen orange and not green :P#asahi-dev · 2025-12-23 07:49
jannauno. why would anyone turn that on if accurate colors are desired. I wanted to test if ctm has the intended effect but never did#asahi-dev · 2025-12-23 07:45
chadmedeven with ctm-based night light on?#asahi-dev · 2025-12-23 06:52
jannauthe mbp 14-inch display wasn't far off when I measured it (2023 / 2024) assuming I did not screw up the measurement and interpreted the results correctly#asahi-dev · 2025-12-23 00:33
chadmedthe colourimetry for the builtin displays is definitely wrong through and setting the right DisplayP3 profile does not fix it in the slightest #asahi-dev · 2025-12-23 00:32
chadmedi dont think we'll need one just to make dcp DTRT since it all happens in firmware and we just feed it constants to select a colour space and transfer function #asahi-dev · 2025-12-22 23:00
jannauI have one#asahi-dev · 2025-12-22 22:53
chaos_princesssomething tells me that expensing a colourimeter is in someone's future#asahi-dev · 2025-12-22 22:31
chadmedchanging how all of this behaves in teh driver currently is not going to be trivial though #asahi-dev · 2025-12-22 22:29
chadmednone of which we currently do#asahi-dev · 2025-12-22 22:29
chadmedwe also need to properly set the correct EOTF, colour space and transfer functions #asahi-dev · 2025-12-22 21:12
chaos_princesswhen dcp starts under linux it complains that it can't find a backlight service among other things. might be related#asahi-dev · 2025-12-22 21:11
oliverMaybe. I saw in a trace that the brightness was set to value larger then the max in the dcp driver. #asahi-dev · 2025-12-22 21:07
jannauI haven't looked into how the higher brightness for HDR is handled except that's likely not in the same way as we handle the backlight currently#asahi-dev · 2025-12-22 21:05
jannauoliver: we have information about the supported colorspaces. might be either somewhere in m1n1's dcp python code or I have notes#asahi-dev · 2025-12-22 20:49
oliverDid anyone look into hdr on the internal display in the past? Is there something written down? I see that gnome already uses the rbg2101010 format for the primary plane. I also see some bits in the dcp driver regarding color space and transfer function.#asahi-dev · 2025-12-22 16:22
fl0_idRight. I can try if it actually compiles a bit later. Flashing and testing would take a bit longer as I still need that machine on macos for a bit today....#asahi-dev · 2025-12-22 16:21
mazfl0_id: yup, got it. in my case, works means compiles, boots, and performs reliably. from past experience, it hasn't always been smooth sailing.#asahi-dev · 2025-12-22 16:19
fl0_idthx. (and just to clarify, by works I mean should work as it has the bumped cmake version. I didn't try all the versions)#asahi-dev · 2025-12-22 16:18
mazOK. I'll see if 2.0.0+ works with the current code. If it does, I'll pull that in.