#
✕ nick:marcan
#asahi · 2022-09-25 07:37
marcanbluetail: asahi linux actually ships with a cut down version of my .zshrc
#asahi · 2022-09-23 05:21
marcanit's highly unlikely for these machines to die of silicon failure before they die of something else
#asahi · 2022-09-23 05:21
marcanit's just "using the CPU more". it's not going to have any more impact than if you used the CPU more for other reasons.
#asahi · 2022-09-19 11:36
marcansince hopefully they will think twice after being instructed to blow up their speakers
#asahi · 2022-09-19 11:35
marcan*and it also protects people from dumb tutorials, mostly
#asahi · 2022-09-19 11:35
marcanand i9t
#asahi · 2022-09-19 11:34
marcanpoint being if someone blows up their speakers after doing that, well, "told you so"
#asahi · 2022-09-19 11:34
marcanand the safety limits in macaudio are *not* enough yet
#asahi · 2022-09-18 16:32
marcanpushed
#asahi · 2022-09-17 08:13
marcanprobably best to move to #asahi-alt then, we can't help you with nonstandard setups
#asahi · 2022-09-17 08:05
marcanI don't know how Debian does it, on Arch the DTBs come from the kernel
#asahi · 2022-09-17 08:04
marcanjust set the display= argument
#asahi · 2022-09-17 08:03
marcanRy_Darcy: why are you rebuilding m1n1?
#asahi · 2022-09-17 08:03
marcanyou need to enable the drivers to get the jack to work anyway
#asahi · 2022-09-17 08:02
marcanbut that needs povik to implement IVSENSE first :)
#asahi · 2022-09-17 08:01
marcanthough I do plan to cook up my own safety models (always fun to beat apple at their own game)
#asahi · 2022-09-17 08:01
marcanfor those that don't, we could just pick something very conservative (audibly)
#asahi · 2022-09-17 08:01
marcanat least for models which have such safety model
#asahi · 2022-09-17 08:01
marcanre safe, now that we have the macOS speaker model data I think we can find a safe max volume. won't be loud, but it'll work.
#asahi · 2022-09-17 08:00
marcanit's already there, disabled by default in DTS
#asahi · 2022-09-17 07:54
marcanwith SBC it does work, with varying levels of dropouts, but wifi performance does consistently drop with BT enabled so I think it's doing *something* with coex
#asahi · 2022-09-17 07:54
marcanit completely kills LDAC earphones, but macOS doesn't support LDAC *at all*, so it's possible that is just not supported / a bandwidth issue with coex?
#asahi · 2022-09-17 07:53
marcanit's actually not entirely clear to what extent btcoex is broken
#asahi · 2022-09-17 07:52
marcanbut there's a good argument to be made that it's better than nothing
#asahi · 2022-09-17 07:52
marcanyeah, it's my only worry with pushing a shitty audio version, I'd have to make it *very* clear in the announcement that it's not representative
#asahi · 2022-09-17 07:51
marcan(just sending raw to woofers that is)
#asahi · 2022-09-17 07:51
marcanit's better than many other laptops, tbh...
#asahi · 2022-09-17 07:50
marcangood audio will take longer; good loud audio will take even longer ;)
#asahi · 2022-09-17 07:50
marcanwell, I think we can get crappy audio out pretty soon, I'm basically just waiting on the glitches that were reported in the latest kernel for that
#asahi · 2022-09-17 07:49
marcanhah
#asahi · 2022-09-17 07:49
marcanI intend to add something like /etc/m1n1.conf but it's not there yet
#asahi · 2022-09-17 07:48
marcankernel updates re-install m1n1. there is no persistent mechanism to configure custom m1n1 builds right now other than replacing the package.
#asahi · 2022-09-17 07:48
marcanOf course you coul also just add bare metal Asahi support, maybe once I fix brcmfmac monitor mode? ;)
#asahi · 2022-09-17 07:47
marcanJust a suggestion, because it does mark a significant difference vs. the x86 images, and the "direct access to hardware" etc points don't really apply
#asahi · 2022-09-17 07:46
marcan*supposed*, yeah... :)
#asahi · 2022-09-17 07:46
marcanI'd probably still add a note to the blurb above though, something like "Note: The ARM64 / Apple Silicon images are intended to be installed as a VM and will not work on most bare metal ARM64 platforms" perhaps?
#asahi · 2022-09-17 07:45
marcansteev: Thanks for making the Kali page changes. I think the wording now makes a lot more sense.
#asahi · 2022-09-16 04:22
marcanchild power domains actively refuse to power up without the parent
#asahi · 2022-09-16 04:22
marcanwe ran into that with ANS2, tried to be smart, turns out the ADT hierarchy was real and enforced by the hardware
#asahi · 2022-09-16 04:22
marcanbut if you do any reparenting, make sure it really can work
#asahi · 2022-09-16 04:21
marcanpovik: there are manual edits already, yeah
#asahi · 2022-09-15 18:40
marcansigh...
#asahi · 2022-09-15 18:13
marcanVMs under macOS on M1s do not virtualize an M1. If they did, they could run Asahi, and they couldn't run those Kali images.
#asahi · 2022-09-15 18:13
marcanThose are just generic ARM64 images, which run on generic ARM64 VMs under macOS on M1s (and on other platforms)
#asahi · 2022-09-15 18:13
marcanThe Kali naming is totally wrong, since there is absolutely nothing "Apple M1" about those images and Kali does not run on virtual Apple M1s at all.
#asahi · 2022-09-13 10:38
marcan(this isn't rocket science... I can't believe I still have to tell you these things...)
#asahi · 2022-09-13 10:38
marcanif you are building your kernels with suspend enabled, that's another mistake in your kernel .config for the endless list you've made so far :)
#asahi · 2022-09-13 10:38
marcanuserspace should detect missing kernel suspend support
#asahi · 2022-09-12 08:03
marcanlooking at it now
#asahi · 2022-09-11 09:35
marcan(that would be a bug in gnome-shell: doing something every frame even when idle is a major waste of power on any platform)

plus ›