#asahi-dev · 2023-03-04 13:17
marcanbecause they are separate drivers#asahi-dev · 2023-03-04 13:17
marcan(panfrost, etnaviv, nouveau, and now asahi)#asahi-dev · 2023-03-04 13:17
marcanthere is precedent for calling community reverse engineered GPU drivers (both mesa and kernel side) by project names#asahi-dev · 2023-03-04 13:16
marcankettenis: good luck#asahi-dev · 2023-03-04 13:06
marcan^ I'll take that as that we should, if only to encourage the BSD guys to start supporting rust ;)#asahi-dev · 2023-03-04 13:04
marcanthere's always RiiR :D#asahi-dev · 2023-03-04 13:00
marcanand I would in fact consider it a requirement for upstreaming that we can prove we can handle that#asahi-dev · 2023-03-04 13:00
marcanso I would much prefer if we can support both versions side by side to prove that works#asahi-dev · 2023-03-04 12:59
marcanchadmed, jannau: re 13.2, I'm not opposed to forcing people to upgrade, buuut 1) it increases the releng workload, and 2) it's just kicking the can down the road, we *can't* drop firmware version support once upstreamed.#asahi-dev · 2023-02-26 16:04
marcanhttps://gitlab.freedesktop.org/asahi/ has been a thing for a while :)#asahi-dev · 2023-02-26 15:47
marcanI'll say when we officially switch to gitlab (probably once we have useful CI to make it worthwhile)#asahi-dev · 2023-02-26 15:47
marcanjannau: nah, still on github for now#asahi-dev · 2023-02-26 15:14
marcanjannau: FWIW I think we have a kernel release in the horizon (with all of the stuff Lina/Alyssa have been working on) so if DCP runtime-pm is still wonky I say we disable it for now.#asahi-dev · 2023-02-24 04:48
marcanthis is all for j314#asahi-dev · 2023-02-24 04:48
marcanso I don't think we want to cap lower#asahi-dev · 2023-02-24 04:47
marcanso we basically have ~8x power headroom give or take, which is not ridiculous when ~4x is clearly actually useful#asahi-dev · 2023-02-24 04:47
marcan(well, 11W peak for a sine, 22W is peak for a square/IWTLW)#asahi-dev · 2023-02-24 04:46
marcanbut the peak power the amps can put out is like 22W#asahi-dev · 2023-02-24 04:46
marcanmax steady-state for woofers is 1.3W, about 1.0W for tweeters#asahi-dev · 2023-02-24 04:45
marcanfor less ridiculous inputs, that means you can have some pretty loud everything and still sound crisp even if the tweeters are running at like 4x max steady-state power during HF transients or even more (that's 6dB free headroom)#asahi-dev · 2023-02-24 04:44
marcanthen they gradually come back after that part#asahi-dev · 2023-02-24 04:44
marcanyou can hear it in the IWTLW demo: the snares during the dubstep part of the song sound nice and crisp, but when the insane clipped pads come in, they overload and throttle#asahi-dev · 2023-02-24 04:43
marcanthe thing is though, especially for the tweeters, the headroom helps a *lot*#asahi-dev · 2023-02-24 04:43
marcanso that has to be the DVC level cap when no safety daemon is running#asahi-dev · 2023-02-24 04:43
marcanpovik: FWIW the safe cap is -14dB lower than the current one (at that volume level no speaker can overheat with worst-case input)#asahi-dev · 2023-02-23 16:39
marcanthe spi-hid stuff I'm more worried about#asahi-dev · 2023-02-23 16:39
marcanhost should be fine#asahi-dev · 2023-02-23 16:39
marcanthat was just the host though#asahi-dev · 2023-02-23 16:39
marcanyeah, it's been on the back burner forever#asahi-dev · 2023-02-23 16:10
marcanI mean the song was at 100%#asahi-dev · 2023-02-23 16:08
marcanj`ey: % of what?#asahi-dev · 2023-02-23 16:06
marcanI'll have to test without the 3dB pad to be really sure but I think this looks good#asahi-dev · 2023-02-23 16:06
marcanpassed the I Won The Loudness War at 0dBFS on every speaker test on j314 (with the kernel cap still 3dB lower than macOS)#asahi-dev · 2023-02-23 16:06
marcanother than that it works with the new daemon :) (still missing all the interlocks, sample rate solution, and catchup code though)#asahi-dev · 2023-02-23 16:05
marcanpovik: ivsense fix on sound/ivsense-fix (the sdout zero mask was getting cropped too much)#asahi-dev · 2023-02-18 13:24
marcanapcie just has this weird split where m1n1 sets some things up that get lost#asahi-dev · 2023-02-18 13:24
marcansven: I think it's just the phy calibration/fuses, I remember it breaking *sometimes*. if you are bringing everything up from Linux for usb4-pcie, then it's probably fine.#asahi-dev · 2023-02-18 01:36
marcanwe need save/restore or to actually move it to linux for that#asahi-dev · 2023-02-18 01:36
marcansven: because otherwise they lose the stuff m1n1 sets up IIRC#asahi-dev · 2023-02-17 08:03
marcanI guess that's what we get for being the first big.LITTLE system with a decent GPU :p (Intel's mess doesn't count)#asahi-dev · 2023-02-17 07:57
marcan(for compositors you want high *priority* but don't necessarily need high *performance*)#asahi-dev · 2023-02-17 07:56
marcanlike "we know kwin isn't a CPU hog, let it land on the e cores"#asahi-dev · 2023-02-17 07:56
marcanalso I think mesa has a quirks list, so it could have custom settings per app#asahi-dev · 2023-02-17 07:55
marcanif this gets implemented in mesa it probably should just detect no sync cases and go straight to that#asahi-dev · 2023-02-17 07:54
marcanyes, apps without vsync would go to 1024 and should#asahi-dev · 2023-02-17 07:54
marcanI wouldn't be surprised if that uclamp gives us >95% of the boost from pinning/performance on glmark, with much better power consumption#asahi-dev · 2023-02-17 07:53
marcanand we have fast switching so doing it often is no big deal#asahi-dev · 2023-02-17 07:53
marcanhere they would#asahi-dev · 2023-02-17 07:53
marcanin particular performance right now *sucks* because the CPUs don't drop down in frequency when idle#asahi-dev · 2023-02-17 07:52
marcanyes, but only for those apps