#asahi-dev · 2022-07-12 04:24
chadmedmarcan: yea ik it runs on boot on asahi but i havent gotten around to setting up an openrc service to do so on gentoo just yet so i probably should for when these fw upgrades roll out#asahi-dev · 2022-07-12 01:45
chadmedor would i be better off just writing a little openrc service that runs update-vendorfw on boot#asahi-dev · 2022-07-12 01:43
chadmedmarcan: does asahi-fwextract do anything different to the extant update-vendorfw script that would necessitate an ebuild?#asahi-dev · 2022-07-11 13:00
chadmedor is the boot picker an app inside the paired recovery#asahi-dev · 2022-07-11 13:00
chadmedit doesnt reboot if you do that#asahi-dev · 2022-07-11 12:59
chadmedhow does that work for booting into paired recovery tough #asahi-dev · 2022-07-11 06:41
chadmedcurrently it only supports getting information from DMI#asahi-dev · 2022-07-11 06:41
chadmedi kind of want to hack minimal DT support into kinfocenter's main view but i feel like it would be messy and not well received#asahi-dev · 2022-07-11 06:40
chadmedwell yeah i dont really care at this point, if the patches i sent to util-linux and kde get merged then thats good enough for me#asahi-dev · 2022-07-11 06:06
chadmedso it would be an absolute mess of case/esac and evil awk regex matching#asahi-dev · 2022-07-11 06:06
chadmedyoure not allowed to pipe anything and youre meant to avoid external utilities altogether... #asahi-dev · 2022-07-11 06:06
chadmedthe style rules for neofetch make it almost impossible to efficiently support proper identification of aarch64 parts anyway#asahi-dev · 2022-07-11 06:05
chadmedkf5solid and lscpu were easy #asahi-dev · 2022-07-11 06:04
chadmedbut its going to be too painful to fix so im not gonna bother#asahi-dev · 2022-07-11 06:04
chadmedheh true but "(10) @ 2.064 GHz" looks stupid and is wrong to boot#asahi-dev · 2022-07-11 05:55
chadmedits just an enormous shell script so it will require some pretty annoying changes to how it enumerates the cpus#asahi-dev · 2022-07-11 05:53
chadmedhmph neofetch is going to *suck* to make work properly with big.LITTLE systems#asahi-dev · 2022-07-09 14:42
chadmedthat seems more like an intel thing at this stage given their partnership with sifive#asahi-dev · 2022-07-09 14:40
chadmedunfortunately both neofetch and kinfocenter get their data from /proc/cpuinfo which is less than useful on aarch64 since it only gives you hex values out of MIDR_EL1 instead of human readable strings#asahi-dev · 2022-07-09 14:39
chadmedyeah of course, but if we could stop them being called cryptic bs based on apple's internal codenames for the socs then that would be nice which is my goal for this #asahi-dev · 2022-07-09 14:37
chadmedand given that it was only recently added, we can assume its still in flux and subject to change to improve the UX #asahi-dev · 2022-07-09 14:36
chadmedyeah but my gripe with it is that its gross and is a bad UX#asahi-dev · 2022-07-09 14:35
chadmedthat said i dont think apple is going to design (16C3)*3 cores any time soon #asahi-dev · 2022-07-09 14:33
chadmedyeah of course, im just concerned about ending up with an utterly enormous array for apple cores in lscpu-arm.c for util-linux, or worse yet running out of hex values for apple cores :P #asahi-dev · 2022-07-09 14:29
chadmedeverything has to be weird and backwards#asahi-dev · 2022-07-09 14:28
chadmedyeah i would assume theyd use REVIDR_EL1 instead of just incrementing MIDR_EL1 but this is apple we are talking about#asahi-dev · 2022-07-09 14:20
chadmedi will mull this over when i have less ethanol in my bloodstream :P#asahi-dev · 2022-07-09 14:20
chadmedkettenis: that sounds like a good idea though again thinking of the UX side of things it almost suggests that they are materially different cores#asahi-dev · 2022-07-09 05:06
chadmedespecially when people go to view information about their system in neofetch or kde settings and it says 20 m1 max firestorm cores instead of m1 ultra #asahi-dev · 2022-07-09 05:05
chadmedit probably makes more sense to go with just Max since the Ultra is 2x Max and the cores are identical parts but from a UX perspective that seems less than intuitive#asahi-dev · 2022-07-09 05:01
chadmedim adding missing parts to lscpu #asahi-dev · 2022-07-09 05:01
chadmedcant decide whether to call them "Icestorm (M1 Max/Ultra)" or just use "(M1 Max)"#asahi-dev · 2022-07-09 05:00
chadmedyeah cool just checking #asahi-dev · 2022-07-09 04:58
chadmedare they still 0x028 and 0x029? #asahi-dev · 2022-07-09 04:58
chadmedthe part IDs in MIDR_EL1 didn't change between the Max and the Ultra did they? #asahi-dev · 2022-07-04 09:36
chadmedheh have you ever tried to do it on a raspberry pi without distcc? that was not a very fun weekend for me#asahi-dev · 2022-07-04 09:34
chadmedthe only thing ive been defeated by with this setup before is chromium in form of qtwebengine, which i dont use anyway so this should be fine #asahi-dev · 2022-07-04 09:34
chadmed16gb, i shaved 16gb off the rootfs the asahi installer made and turned it into a swap though so im all good#asahi-dev · 2022-07-04 09:31
chadmedim bootstrapping clang on the mbp right now, all is going according to plan#asahi-dev · 2022-07-04 09:31
chadmedi had a feeling you might like something like this :)#asahi-dev · 2022-07-04 07:29
chadmedhttps://github.com/chadmed/asahi-gentoosupport if anyone's curious#asahi-dev · 2022-07-04 07:07
chadmedi guess the nice thing though is that we can install gentoo the "correct" way now #asahi-dev · 2022-07-04 07:07
chadmedtime required to repackage and boot a vanilla gento install cd: ~40 hours#asahi-dev · 2022-07-04 07:06
chadmedok im with sven packaging and distribution is the absolute worst oh my god#asahi-dev · 2022-07-03 07:34
chadmedthought so thanks #asahi-dev · 2022-07-03 07:32
chadmeddoes m1n1 still need imagemagick to build? im writing an ebuild for it #asahi-dev · 2022-07-01 01:15
chadmedaic2 #asahi-dev · 2022-06-28 11:57
chadmedyeah#asahi-dev · 2022-06-28 11:53
chadmedthe machine should be cooler and the battery life should be longer which is what matters anyway#asahi-dev · 2022-06-28 11:53
chadmedyou could A/B test with the old DT and compare battery life but its hard to create a reproducible workload that will sufficiently challenge the scheduler, its mostly an experience thing