#asahi-dev · 2022-12-02 05:05
dsharshakovAnd yes, it might be useful to check whether hw doesn't give us obvious nonsense such as high-entropy data or constants (or other failure modes). We mustn't claim it's fine by measurements but raise an alarm#asahi-dev · 2022-12-02 05:03
dsharshakovSuspend can be a corner case for ALSA, btw. In PW we have quite some bugs of hardware not working after resume (some are caused by drivers though). So if daemon is dead after suspend (alsa bugged out, or device resumed badly and only gives 00s or otherwise garbage) it shouldn't be considered adequate #asahi-dev · 2022-12-02 05:00
dsharshakovchadmed: "how would that play out with suspend etc?" Block immediately after suspend and unlock when daemon thaws and starts working. We shouldn't let 1% chance (likely bc of user's own stupidity) to bite someone#asahi-dev · 2022-12-01 17:20
dsharshakovAnd then things go much complex since it will also need to keep track of user logins, seat and VT management etc...#asahi-dev · 2022-12-01 17:19
dsharshakovHowever, it's necessary to know how should it access the device and will it be root or not. Likely needs to be either root or a system daemon getting uinput access permission#asahi-dev · 2022-12-01 17:18
dsharshakovmarcan: yes, uinput + many decent things daemon is what we want likely. I wanted to discuss some of its working principles. It could be done with Rust and egui (just see it, appears to be a nice toolkit which can be shaped into a touchsceen-first UI for constrained space)#asahi-dev · 2022-12-01 11:13
dsharshakovthis type of quirk appears to be able to just discard the unwanted events for all libinput users: https://gitlab.freedesktop.org/libinput/libinput/-/blob/main/quirks/30-vendor-wacom.quirks#L28#asahi-dev · 2022-12-01 11:09
dsharshakovmaybe hwdb or libinput quirks can be used to disable the touchscreen input from being mapped, so touch part can be exposed as a touchscreen#asahi-dev · 2022-12-01 11:01
dsharshakov(and the USB driver could easily live in userspace as well)#asahi-dev · 2022-12-01 11:01
dsharshakovpovik: I mean if they're interested they know how to do a driver and then try out userspace stuff#asahi-dev · 2022-12-01 11:00
dsharshakovSome more integration than virtual keys is possible: volume + brightness controls (sliders) and global menu can be done relatively easily#asahi-dev · 2022-12-01 10:59
dsharshakovThe userspace part could be discussed in T2 chats, since apparently there's a knowledge on how to get the display working on those machines#asahi-dev · 2022-12-01 10:58
dsharshakov_jannau_: +1, first we need an understanding of the device to know how to hook it up in kernel#asahi-dev · 2022-12-01 10:57
dsharshakovMaybe the question should be postponed to when there's understanding of the display controller such as a m1n1 experiment, so that we see the properties better and know whether it's DRM or not#asahi-dev · 2022-12-01 10:55
dsharshakovHowever, someone might have different ideas and then compositor must use the touchbar, but just know about it and use wisely#asahi-dev · 2022-12-01 10:54
dsharshakovHowever, should touchbar support be in or stay away from compositors? I'm mostly for the latter approach for better cross-desktop support and easier app integration#asahi-dev · 2022-12-01 10:53
dsharshakovthey mustn't use non_desktop screens unless asked by a VR client, and the NON_VR flag will tell VR clients to avoid the bar as well#asahi-dev · 2022-12-01 10:52
dsharshakovMaybe a new flag? Non_desktop | non_vr = don't touch unless asked to?#asahi-dev · 2022-12-01 10:51
dsharshakovI wouldn't trust in compositors that much #asahi-dev · 2022-12-01 10:50
dsharshakovand t2linux likely have less work to do and somebody might devote their time to writing a Linux driver + discussing the concept#asahi-dev · 2022-12-01 10:49
dsharshakovYes, but existence of one means people discovered how to talk to bridgeOS to customize the touch bar. Maybe lack of Linux driver is just bc nobody cares#asahi-dev · 2022-12-01 10:48
dsharshakovI wanted to ask them whether someone can port that bootcamp driver to Linux to later try with my `auxdisplayd` approach#asahi-dev · 2022-12-01 10:47
dsharshakovProbably a FB driver should be sufficient to render the touchbar at its 60 Hz with plain pixel copies. Maybe it'd be great to co-ordinate with T2 devs to propose a common device api to expose touchbar display with#asahi-dev · 2022-12-01 10:44
dsharshakov"ok like vr stuff" <-- yes, for that#asahi-dev · 2022-12-01 10:44
dsharshakova drm driver would be difficult to prevent from being grabbed by the compositor (likely we'll have to either place our quasi-window there or ask for a DRM lease)#asahi-dev · 2022-12-01 10:43
dsharshakovThere's been a MIT-licensed BootCamp driver for TouchBar by someone, that just pushes pixels over usb#asahi-dev · 2022-12-01 10:42
dsharshakovLikely we can avoid drm, since that's unlikely to fit well. T2 touchbar is a framebuffer, not a proper display controller with swaps etc#asahi-dev · 2022-12-01 10:41
dsharshakovIt'd have a D-Bus (or socket) API for apps to register widgets with, as well as own widgets such as Fx (+esc) buttons or sth like that#asahi-dev · 2022-12-01 10:40
dsharshakovI think of writing a Rust-based touchbar service `auxdisplayd` for such cases as touchbar, ASUS ScreenPad or custom touchscreens makers want to attach#asahi-dev · 2022-12-01 10:39
dsharshakovChaosPrincess: hello! I wanted to reach out to either you or T2 people about touchbar userspace#asahi-dev · 2022-11-23 09:11
dsharshakovAlright, sorry for that. Sometimes having an outline of end product can help developing core as well, so I added those ideas to keep them around#asahi-dev · 2022-11-23 09:02
dsharshakovchadmed: if it could be wired up to TensorFlow PluggableDevice API, then tf apps will be able to make use of ANE even in case supported op count is low. Unified memory gives us possibilities to not care about moving data between ANE and CPU being slower than benefits from initial ops implementation #asahi-dev · 2022-11-23 08:54
dsharshakovAccel subsystem of Linux kernel I mean, read news about 6.2#asahi-dev · 2022-11-23 08:54
dsharshakovThat would depend on memory model#asahi-dev · 2022-11-23 08:49
dsharshakoveiln: what're your plans on UAPI to expose ANE to apps? How suitable does accel from 6.2 feel?#asahi-dev · 2022-11-23 08:48
dsharshakovBut that won't help organizing tons of bits branches however#asahi-dev · 2022-11-23 08:46
dsharshakovIt might be right time to have redundant mirrors of essential repository on g.fd.o#asahi-dev · 2022-11-23 08:01
dsharshakovOfc not a protection from malware running as root, but a measure against messing with sudo alsactl#asahi-dev · 2022-11-23 08:00
dsharshakovAnd after daemon shakes hands no one can take over hw gain control while it's running. Even root#asahi-dev · 2022-11-23 07:59
dsharshakovAlso ensure robust panic and hang handling#asahi-dev · 2022-11-23 07:59
dsharshakovYes, that is the proper approach when all the safety mechanisms are in place to make sure every single moment amps work there's strict observation to keep things safe. When no daemon is present, set low hardware volume caps unless it appears and starts monitoring. #asahi-dev · 2022-11-23 07:42
dsharshakovYep, will work on that! Need to port things to new wp 0.5 and fix validity issues#asahi-dev · 2022-11-23 07:41
dsharshakovpovik: my idea for enhanced speaker safety is to also have a panic handler disabling TAS2764 in case of kernel panic. I think admac will keep sending a stuck audio fragment in case cpu hangs#asahi-dev · 2022-11-23 07:39
dsharshakovGot it#asahi-dev · 2022-11-23 07:39
dsharshakovNeed better protection against fails#asahi-dev · 2022-11-23 07:38
dsharshakovNo way#asahi-dev · 2022-11-23 07:38
dsharshakovWhat if it get OOM killed?#asahi-dev · 2022-11-23 07:38
dsharshakov* or users messing with it as well#asahi-dev · 2022-11-23 07:37
dsharshakovI understand it's not to be in kernel. But userspace monitoring should be very robust and resilient against failures#asahi-dev · 2022-11-23 07:35
dsharshakovWhat would be the api for watchdog? Where would watchdog be?