#asahi · 2025-04-05 22:03
chaos_princessidk right now, i only measured idle power consumption#asahi · 2025-04-05 22:02
chaos_princessand yes, pmp does save power just by being turned on, well, except for the whole "crashing dcp" bit#asahi · 2025-04-05 22:00
chaos_princessthis changed on m1pro+, those have "pmp v2", while m1 has pmp v1 that has a different init sequence and does get sent messages over rtkit#asahi · 2025-04-05 21:58
chaos_princessBut the pmp io regions are no longer touched after initialization completes, the "dashboard" and other stuff is in pmgr#asahi · 2025-04-05 21:57
chaos_princesschadmed: no, there is something called "pmp dashboard" that needs to be poked on ip block power state transitions#asahi · 2025-04-05 21:09
chaos_princessthough it might be related, not sure#asahi · 2025-04-05 21:09
chaos_princessjannau: no, iirc that is different#asahi · 2025-04-05 20:36
chaos_princesson t600x, it is somehow related to io-range 41 in pmgr#asahi · 2025-04-05 20:33
chaos_princessor in other words - make this not crash dcp https://github.com/WhatAmISupposedToPutHere/linux/commit/5b5e2e3c2db62f672788b084c3dd4b1796070c02#asahi · 2025-04-05 20:33
chaos_princesssure, figure out how blocks like dcp and such communicate power states to pmp and do fabric bandwidth control. #asahi · 2025-04-04 22:26
chaos_princess"generic" swe is also kinda very different from what we do#asahi · 2025-04-04 22:04
chaos_princessi do not think there is such a link, you just kinda have to know :P#asahi · 2025-04-04 22:03
chaos_princessknown issue, wait for muvm next muvm release that contains this fix https://github.com/AsahiLinux/muvm/pull/162#asahi · 2025-04-04 21:23
chaos_princessbut also in general, everything involving sleep and/or power management is not as good as it can be#asahi · 2025-04-04 21:22
chaos_princessit is probably fine? We do "phone-style" sleep where cpu is still on, it is just in a very deep power down state#asahi · 2025-04-04 21:16
chaos_princessthat too#asahi · 2025-04-04 21:15
chaos_princessah, you mean on desktops#asahi · 2025-04-04 21:13
chaos_princessi think the issue is that kb+trackpad gets fully powered down when in our sleep state, and macos enters some low power state instead#asahi · 2025-04-04 21:11
chaos_princessnot supported (yet?)#asahi · 2025-04-04 14:19
chaos_princesssounds like it#asahi · 2025-04-04 14:09
chaos_princessno#asahi · 2025-04-04 14:07
chaos_princessiboot -> checks if button is pressed (no) -> tries to load from nvram variable -> if fails - tries to load recovery -> if fails goes to dfu#asahi · 2025-04-04 14:01
chaos_princessthis is not like on efi systems where the boot selector is a part of boot process and "continues" into the os, here it is a slimmed-down macos install, that reboots into what you picked#asahi · 2025-04-04 13:59
chaos_princessbecause at that point you are running iBoot, and boot selector is an app running under full macos#asahi · 2025-04-04 13:05
chaos_princessyes, it loops through partitions, but at that point you are actually running a tiny macos#asahi · 2025-04-04 13:03
chaos_princess(probably, but i am not brave enough to check)#asahi · 2025-04-04 13:02
chaos_princessif it does not exist or corrupted, it boots from system recovery, and _that_ one is order-sensitive#asahi · 2025-04-04 13:02
chaos_princessthere is a nvram variable that contains the uuid of partition to boot from#asahi · 2025-04-04 13:02
chaos_princessoh, you mean iBoot->m1n1 stage 1?#asahi · 2025-04-04 13:01
chaos_princessu-boot is embedded in m1n1 stage 2#asahi · 2025-04-04 13:01
chaos_princessfor asahi stuff, m1n1 chainload just looks up partition by uuid, and loads stage2 from it using fat32#asahi · 2025-04-04 12:59
chaos_princesswell, depending on your os install, there might be hardcoded partition numbers somewhere, but from what i remember, fedora does it right and uses uuids only#asahi · 2025-04-04 12:48
chaos_princessorder does not matter, partitions should refer to each other by uuids#asahi · 2025-04-04 11:34
chaos_princesscreate a live usb by copying your rootfs there, boot from it, use gparted to move partitions around#asahi · 2025-04-04 11:34
chaos_princesswith difficulty#asahi · 2025-04-03 15:52
chaos_princesspre-t2 is basically just an slightly weird intel machine, most distros should just be fine#asahi · 2025-04-03 15:51
chaos_princessthat sounds pre-t2, so, probably no?#asahi · 2025-04-03 15:50
chaos_princessthere is a separate project for intel macs, they will probably be more helpful https://t2linux.org/#asahi · 2025-04-03 15:46
chaos_princessdepends? if it has fmm, i think it will ask for a password shortly after#asahi · 2025-04-03 10:44
chaos_princessanything suspicious in dmesg?#asahi · 2025-04-03 08:24
chaos_princesscan i ask a pretty stupid question first? does this combo of monitor+cable work under macos?#asahi · 2025-04-03 07:57
chaos_princessyou are using the actual hdmi port, right?#asahi · 2025-04-03 07:56
chaos_princesswell, it should work then#asahi · 2025-04-03 07:56
chaos_princesspaste output of `cat /sys/firmware/devicetree/base/chosen/asahi,os-fw-version`#asahi · 2025-04-03 07:53
chaos_princess13 inch or 14/16?#asahi · 2025-03-31 11:11
chaos_princessits an ancient platform that is only interesting for it's historical value, and it has no relevance for this project#asahi · 2025-03-31 11:10
chaos_princess?#asahi · 2025-03-31 11:09
chaos_princesshttps://asahilinux.org/docs/platform/introduction/ #asahi · 2025-03-31 11:08
chaos_princesshttps://asahilinux.org/docs/platform/security/#asahi · 2025-03-31 10:50
chaos_princessthere is also the bit that we can't do this on a technical level even if we hypothetically wanted to. "what if someone wipes macos and uses it as a linux box only" was a design consideration when making fmm/the rest of the secure boot system