#asahi-dev · 2024-05-26 08:57
LarstiQinteresting, now I'm interested in CPU heating. I know for LNG tanks the gradient is very non-linear and you can easily get 60℃ difference#asahi-dev · 2024-05-26 08:34
jannauthey differ quite bit iirc, too bad that we don't have a description of what they measure exactly#asahi-dev · 2024-05-26 08:32
jannauif there were 6 sensors per core. I did not find temp sensors for the efficiency cores#asahi-dev · 2024-05-26 08:29
LarstiQhow much can those differ in measurements?#asahi-dev · 2024-05-26 08:24
chadmedthe amount of data exposed by the smc is ridiculous and most of it is completely superfluous for our needs, e.g. having 6 temperature sensors per cpu core#asahi-dev · 2024-05-26 08:24
chadmedonly question really left is what do we actually want to expose#asahi-dev · 2024-05-26 08:18
chadmedhttps://github.com/chadmed/linux/commits/asahi-hwmon-merge/ mostly happy with this now as a starting point. gives some sensors on all supported devices. #asahi-dev · 2024-05-26 08:16
chadmedi squashed all the dt changes down because they were noisy and useless#asahi-dev · 2024-05-26 05:23
TellowKrinkleGenerally when people say "64-bit instruction set", they're referring to an instruction set that _can_ use 64-bit registers, not one that restricts you to always using the full 64 bits in every instruction#asahi-dev · 2024-05-25 10:56
chadmedheh#asahi-dev · 2024-05-25 10:50
jannauI wouldn't worry about the pro#asahi-dev · 2024-05-25 10:46
chadmedyeah that would probably work, again assuming theres not too much else that is divergent. im kinda worried that the mac pro will introduce even more nonsense that will necessitate a -desktop in between -common and -mini/-studio anyway #asahi-dev · 2024-05-25 10:42
jannauchadmed: I was thinking of hwmon-mini.dtsi / hwmon-studio.dtsi instead of desktop#asahi-dev · 2024-05-25 10:12
chadmedi think we need to play with more keys across more machines to have a better idea of what is actually common and what isnt. if its just a few useless/anciliary keys that arent shared then we can just not even bother with them #asahi-dev · 2024-05-25 10:12
chadmedi toyed with the idea of an hwmon-desktop.dtsi but since tw0p is on the minis but not the studios (and who even knows what is going on with the pro :p) that didnt really make a lot of sense to me. there doesnt really seem to be any common thread#asahi-dev · 2024-05-25 09:57
jannaumaybe we should do chassis specific includes? or just include key specific includes like '#include hwmon-temp-tw0p.dtsi`. that avoids duplicating the key info#asahi-dev · 2024-05-25 09:47
chadmedim not a fan of how fragmented the devicetree stuff needs to be but id rather deal with a bit of an ugly dt than have loads of "could not get key" errors polluting folks' dmesg #asahi-dev · 2024-05-25 09:47
chadmedhttps://github.com/chadmed/linux/tree/asahi-hwmon-next#asahi-dev · 2024-05-25 09:46
chadmedokay i think we're 99% of the way to something workable for hwmon if anyone has any feedback#asahi-dev · 2024-05-25 01:07
chadmedinteresting#asahi-dev · 2024-05-25 01:07
chadmedok yeah TW0P exists on the m2 pro mini#asahi-dev · 2024-05-25 00:49
chadmedthe set of common keys seems to be rather small. i doubt we can avoid sku-specific hwmon keys for very long#asahi-dev · 2024-05-25 00:48
chadmedboth j314 and j415 have it though#asahi-dev · 2024-05-25 00:48
chadmedi will install the branch on my server to confirm then. it definitely 100% does not exist on the studio though#asahi-dev · 2024-05-24 15:55
jannauchadmed: it exists here with reasonable values on m1,m2,m2 pro mac mini according to ./experiments/smc.py#asahi-dev · 2024-05-24 11:54
chadmedthere are missing keys across all machines that if we were looking at the dtsi files in a vacuum, we'd expect to be there#asahi-dev · 2024-05-24 11:54
chadmedfor context im mostly happy with where the driver is at but the devicetree stuff seems a bit scuffed #asahi-dev · 2024-05-24 11:40
chadmedit doesnt exist on m2 pro mini or m2 max studio #asahi-dev · 2024-05-24 11:40
chadmedjannau: do you recall off the top of your head if you ever found "TW0P" (wifi/bt module temp sensor) on any non-laptop machines#asahi-dev · 2024-05-23 09:07
chadmedboth sw cursor fallback and that ICC profile repaint issue have been fixed with kwin 6.0.5! #asahi-dev · 2024-05-18 14:36
tpw_rulesjannau: ok, so all users should use the asahi linux installer to update stage 1, and for those that don't, stage 2 is fixed with 1.4.14? #asahi-dev · 2024-05-18 14:16
mpsthanks for confirmation#asahi-dev · 2024-05-18 14:08
jannauyes#asahi-dev · 2024-05-18 10:50
mpsjannau: is this !392 fix merged in m1n1 1.4.14#asahi-dev · 2024-05-18 10:39
jannauno and that's off-topic here anyway#asahi-dev · 2024-05-18 10:34
z1d0v[m]<tpw_rules> "jannau: is m1n1 1.4.13 the..." <- Oh, so that might be the reason my booting failed process after an update now?#asahi-dev · 2024-05-18 05:58
jannautpw_rules: fix is merged/tagged and pushed to the installer#asahi-dev · 2024-05-18 05:24
jannautpw_rules: m1n1 1.4.13 works as stage2 we tested that. stage 1 fix is in https://github.com/AsahiLinux/m1n1/pull/392#asahi-dev · 2024-05-18 03:26
tpw_rulesis there an ETA on a stage 2 fix?#asahi-dev · 2024-05-18 03:25
tpw_rulesjannau: is m1n1 1.4.13 the broken on M1 one?#asahi-dev · 2024-05-17 19:52
flokliThanks ❤️#asahi-dev · 2024-05-17 16:24
jannauasahi-6.8.9-6 pushed to AsahiLinux/linux. containing a revert of a commit blamed for data corruption on dm-crypt partitions#asahi-dev · 2024-05-16 07:34
mazsven: yeah, I appreciate this is a thankless task...#asahi-dev · 2024-05-16 07:32
svenyeah, i really need to find the motivation to just rewrite that thing :/#asahi-dev · 2024-05-16 07:30
jannauto quote sven: "the pipehandler sequence is all wrong. I'm surprised usb works at all"#asahi-dev · 2024-05-16 07:30
jannauI think reverting 6d735722063a is (for now) required on all SOCs for usb3. at least the pipehandler lock message occurs with plain m2 as well#asahi-dev · 2024-05-16 07:27
mazanyhow, thanks a lot for the detective work!#asahi-dev · 2024-05-16 07:25
mazjannau: confirmed. hacking dwc3_power_off_all_roothub_ports() and reverting this patch results in a working setup again. now, what is so special about the M2 Pro that it requires all this?#asahi-dev · 2024-05-15 20:10
sventhe Bluetooth fixes have been merged rather quickly :)#asahi-dev · 2024-05-15 19:09
jannauthere are no broken builds. the gentoo patch set adds distro/gentoo Kconfig menus with a weird set of selects/depends on GCC_RANDSTRUCT config symbols and that explodes inside kconfig if you select RUST