#asahi · 2026-11-13 13:43
tobias^:)#asahi · 2026-11-13 13:43
tobias^sorry, i just cleaned my keyboard #asahi · 2026-11-13 13:42
tobias^ö~234506ß7#asahi · 2026-09-14 23:24
tobias^Also I see active_count increasing when I unplug/plug the power supply, but no actual wakeup happens from suspended state.#asahi · 2026-09-14 23:23
tobias^Another question: Is there a reliable way to wake up the machine from sleep when a power supply is connected? /sys/class/power_supply/macsmc-ac/power/wakeup is set to "enabled"?#asahi · 2026-09-14 23:17
tobias^Another issue I noticed is that setting charge_control_start_threshold does not appear to have an effect. Its always charge_control_end_threshold - 5, except when its at 100, then its 100 as well.#asahi · 2026-09-14 23:11
tobias^Alright, thanks. :)#asahi · 2026-09-14 23:08
tobias^Although I re-installed this machine completely about 3 weeks ago, so probably not necessary.#asahi · 2026-09-14 23:08
tobias^That should be done before trying to upgrade Mac OS, right?#asahi · 2026-09-14 23:05
tobias^I have 7.1.13-402.asahi.fc44.aarch64+16k right now. Current enough?#asahi · 2026-09-14 23:01
tobias^Tahoe#asahi · 2026-09-14 23:01
tobias^Especially in combination with another bug that charges it to 100% sometimes despite setting the above.#asahi · 2026-09-14 23:00
tobias^That is... annoying. :(#asahi · 2026-09-14 22:56
tobias^I noticed that setting this actually *discharges* the battery when its above 80%, even with a external power source connected. Is that intentional?#asahi · 2026-09-14 22:56
tobias^echo 80 >/sys/class/power_supply/macsmc-battery/charge_control_end_threshold#asahi · 2026-08-30 23:58
tobias^It takes the currently loaded modules and puts it in dracut config in the correct order. After that, execute "dracut --force" and it puts everything in initramfs. Tested with bash and busybox. :)#asahi · 2026-08-30 23:54
tobias^l=""; while IFS=" " read -r m z z u z; do t="${u}"; l="${l}${m} "; while [ -n "${t}" ]; do i="${t%%,*}"; case "${i}" in "-") ;; *) l="${l}${i} "; ;; esac; case "${t}" in *,*) t="${t#*,}"; ;; *) t=""; ;; esac; done; done </proc/modules; r=""; for i in ${l}; do case "${r}" in *" ${i} "*) ;; *) r="${i} ${r}"; ;; esac; done; echo "force_drivers+=\" ${r}\"" >/etc/dracut.conf.d/00-force-drivers.conf;#asahi · 2026-08-30 23:54
tobias^By the way, I made a shell script to automate the initramfs module preloading with dracut. If anyone is interested, here it is:#asahi · 2026-08-30 23:45
tobias^In both cases the power consumption is lower than with just doing echo 0 >/sys/class/backlight/apple-panel-bl/brightness#asahi · 2026-08-30 23:38
tobias^niko[m]: It doesn't appear to make a difference on my machine. Same behavior and power consumption as with just the --blank 1.#asahi · 2026-08-30 07:06
tobias^TERM=linux setterm --blank 1 >/dev/tty1#asahi · 2026-08-30 07:06
tobias^found another 700mW (according to sensors output), kicks in after 1 minute#asahi · 2026-08-29 21:48
tobias^I'm repurposing my broken display M1 Air as a server. Are there any display-related power-savings available apart from "echo 0 >/sys/class/backlight/apple-panel-bl/brightness"?#asahi · 2026-08-29 10:59
tobias^Yes I'll revert to that for now.#asahi · 2026-08-29 10:57
tobias^I see. :( #asahi · 2026-08-29 10:45
tobias^Just tried "hostonly=no", but there are still missing modules like brcmfmac.#asahi · 2026-08-29 10:12
tobias^I would like to load all kernel modules required for a specific machine from within initramfs, so i can easily boot different userlands (e.g. Alpine Linux) with the root= parameter without worrying about modules there. My current approach is to boot into Fedora Minimal, and put the output of "lsmod" into the force_drivers list of dracut. Is there a better approach maybe?#asahi · 2026-08-28 04:06
tobias^maybe someone has done that already?#asahi · 2026-08-28 04:03
tobias^thinking about making a pendrive image that brings up all available network interfaces and launched an sshd with known root password in order to get back in#asahi · 2026-08-28 04:00
tobias^good#asahi · 2026-08-28 03:59
tobias^this can boot from usb when asahi is already installed, right?#asahi · 2026-08-28 03:29
tobias^looks like the screen on my m1 air just died :(#asahi · 2026-08-28 03:28
tobias^damn#asahi · 2026-08-24 13:38
tobias^i'll give it a try, thx :)#asahi · 2026-08-24 13:17
tobias^the usual slop generators suggest /proc/acpi too, but yeah thats not really a thing anymore.#asahi · 2026-08-24 12:44
tobias^thats why i would prefer to not rely on the event interface alone.#asahi · 2026-08-24 12:43
tobias^jannau: yesterday my macbook booted while the lid was closed. i'm not sure why this happened, but when it happens i don't get any close event and therefore massive battery drain...#asahi · 2026-08-24 12:21
tobias^sounds reasonable#asahi · 2026-08-24 12:19
tobias^ok, i guess i'll stick to my workaround for now. where would be the proper place in sysfs to put the lid sensor state in case i start an attempt to write a kernel module for this?#asahi · 2026-08-24 11:55
tobias^chaos_princess: not sure if this got buried, but do you have a suggestion about querying the lid state if the angle sensor is not available?#asahi · 2026-08-23 18:01
tobias^my current workaround is to always send the machine to sleep after acpid is started. #asahi · 2026-08-23 17:59
tobias^however, if for some reason the system boots while the lid is already closed, the lid close event is never triggered, causing the battery to be drained.#asahi · 2026-08-23 17:59
tobias^fyi: i'm running the asahi kernel from fedora with an alpine linux rootfs, and let acpid handle lid events to sleep the machine.#asahi · 2026-08-23 17:10
tobias^so no way to get lid state, just the "button" event?#asahi · 2026-08-23 17:07
tobias^its a macbook air m1 #asahi · 2026-08-23 17:06
tobias^i only see a single iio device, which has in_illuminance_input#asahi · 2026-08-23 17:02
tobias^chaos_princess: found an illuminance sensor in /sys/bus/iio, but not the lid angle. maybe i'm missing a kernel module or something?#asahi · 2026-08-23 16:38
tobias^according to the docs (https://asahilinux.org/docs/hw/soc/smc/) the current lid state can be queried with smc key MSLD. is this exposed somewhere in sysfs?