#asahi · 2026-11-17 02:15
JamesCalligeros[m]yeah something like that#asahi · 2026-11-17 01:44
JamesCalligeros[m]https://copr.fedorainfracloud.org/coprs/g/asahi/kernel/build/11337164/#asahi · 2026-11-17 01:39
JamesCalligeros[m]7.3.3 got skipped for fedora iiuc#asahi · 2026-11-17 01:39
JamesCalligeros[m]there is not, and even if there was it would still take time to hit the repo as 7.3.4 was only tagged hours ago#asahi · 2026-11-17 01:37
JamesCalligeros[m]it should land in the fedora repos soon, otherwise youll have to build it yourself from source#asahi · 2026-11-09 11:48
JamesCalligeros[m]certainly looks worse on normal 1080p/1440p displays#asahi · 2026-11-09 11:48
JamesCalligeros[m]macos doesnt do any subpixel rendering but idk if that is a factor on the high res macbook screens#asahi · 2026-11-09 10:10
JamesCalligeros[m]/var/cache/distfiles says hello#asahi · 2026-11-09 09:40
JamesCalligeros[m]We have had eas for years. The efficiency gap is at idle/during suspend not during use#asahi · 2026-11-08 10:56
JamesCalligeros[m]cant have downtime if you never turn the thing on #asahi · 2026-11-08 10:52
JamesCalligeros[m]^#asahi · 2026-11-08 10:52
JamesCalligeros[m]you need to run tee as sudo #asahi · 2026-11-08 10:51
JamesCalligeros[m]echo -n ... | sudo tee is usually what i do for this#asahi · 2026-11-08 10:50
JamesCalligeros[m]schedutil should be the default on all production systems anyway#asahi · 2026-11-08 10:50
JamesCalligeros[m]did you try echo -n#asahi · 2026-11-08 10:46
JamesCalligeros[m]i can even push my hacked up freqbench fork if anyone really wants to take the easy way out#asahi · 2026-11-08 10:44
JamesCalligeros[m]the good news is that the "test bench" is just freqbench with smc pstr support patched in#asahi · 2026-11-08 10:43
JamesCalligeros[m]nope, i think we did capacity-dmips-mhz but that doesnt populate the energy model for eas #asahi · 2026-11-08 10:43
JamesCalligeros[m]we need to sort out opp-microwatt for m3+ and t602x but yeah#asahi · 2026-11-08 10:42
JamesCalligeros[m]im sure theres someone out there using some insane ebpf android scheduler and governor now that this is possible :P #asahi · 2026-11-08 10:41
JamesCalligeros[m]a few people asked about that quite some time ago, and i always just tell them to leave it on schedutil and forget about trying to hyperoptimise this#asahi · 2026-11-08 10:40
JamesCalligeros[m]i think ppd/tuned/etc do some other stuff on acpi platforms but i have not booted linux on an acpi platform for a very long time :P#asahi · 2026-11-08 01:31
JamesCalligeros[m]look up anylinuxfs#asahi · 2026-11-08 01:31
JamesCalligeros[m]no but you can do the opposite#asahi · 2026-11-08 01:28
JamesCalligeros[m]Soon(tm) #asahi · 2026-11-08 01:27
JamesCalligeros[m]that is 1 billion percent going to be a thunderbolt dock#asahi · 2026-11-07 08:20
JamesCalligeros[m]I removed it from main but we haven’t cut a release yet#asahi · 2026-11-07 08:19
JamesCalligeros[m]Yeah its broken/irrelevant to production systems#asahi · 2026-11-07 08:12
JamesCalligeros[m]And we got sick of it#asahi · 2026-11-07 08:12
JamesCalligeros[m]People were trying to enable them before we had the safety and eq stuff finalised, and then complaining that they sounded bad/damaged their mac#asahi · 2026-11-07 07:51
JamesCalligeros[m]Ignore the audio thing#asahi · 2026-11-03 23:14
JamesCalligeros[m]probably not#asahi · 2026-11-03 23:14
JamesCalligeros[m]why would i try writing to read keys...#asahi · 2026-11-03 23:13
JamesCalligeros[m]e.g. even on machines with the same soc, the cpu temperature sensors are all on different keys#asahi · 2026-11-03 23:13
JamesCalligeros[m]they should but... yknow... think different#asahi · 2026-11-03 23:12
JamesCalligeros[m]PSTR = Power - System Total - Real, PHPC = Power - Heat Pipe - Computed#asahi · 2026-11-03 23:12
JamesCalligeros[m]there are two types of power SMC key as far as i have gathered. some end in R which i assume means "real" and are accurate/measured, and some end in C which are sometimes obvious nonsense and so are obviously "computed" #asahi · 2026-11-03 23:08
JamesCalligeros[m]if you wanna go through the thousands of SMC keys, PRs are welcome :)#asahi · 2026-11-03 23:07
JamesCalligeros[m]not a thing as in either there are no sensors for it or we havent found the right smc keys#asahi · 2026-11-03 23:07
JamesCalligeros[m]smc exposes the total system draw on the PSTR key, which hwmon reports as "Total System Power"#asahi · 2026-11-03 23:07
JamesCalligeros[m]we dont have it wired up at all #asahi · 2026-11-03 23:07
JamesCalligeros[m] twb: youre not going to get per-component power use, thats not a thing on this platform #asahi · 2026-11-03 23:06
JamesCalligeros[m]yeah that sounds about right then#asahi · 2026-11-03 23:05
JamesCalligeros[m]which machine do you have#asahi · 2026-11-02 01:21
JamesCalligeros[m]Does powertop rely on acpi#asahi · 2026-10-30 02:05
JamesCalligeros[m]nah the kernel updates should run update-m1n1 for you automatically, you shouldnt have to do anything#asahi · 2026-10-30 02:03
JamesCalligeros[m]things it always is: dns, devicetrees#asahi · 2026-10-30 02:03
JamesCalligeros[m]it was probably devicetrees#asahi · 2026-10-30 01:58
JamesCalligeros[m]did your devicetrees not get updated or something#asahi · 2026-10-30 01:57
JamesCalligeros[m]no its still appledrm