#asahi-dev · 2026-11-10 11:04
yukaok!#asahi-dev · 2026-11-10 11:04
integralpiloti havent sent it yet#asahi-dev · 2026-11-10 11:04
yukaok, will see what integralpilot replied now before adding my opinion#asahi-dev · 2026-11-10 11:03
jannauconfig symbol for 10gbps is AQTION, driver is called atlantic#asahi-dev · 2026-11-10 11:03
myriaAh, thanks!#asahi-dev · 2026-11-10 11:02
chaos_princessyuka: krzk's "Isn't this implied by the compatible?"#asahi-dev · 2026-11-10 11:02
chaos_princessthats 10g#asahi-dev · 2026-11-10 11:02
myriaWould AQUANTIA be needed at all? #asahi-dev · 2026-11-10 11:02
yukaintegralpilot I don't see any mails regarding stm subnode or has-stm property in the current dchid submission.#asahi-dev · 2026-11-10 11:01
jannauconfig symbol is iirc TIGON3#asahi-dev · 2026-11-10 11:01
jannautg3#asahi-dev · 2026-11-10 11:01
chaos_princesstg iirc, i can boot mine and check#asahi-dev · 2026-11-10 11:00
myriaHey, does anyone know what ethernet drivers the m1 mac mini needs? I can't seem to easily find anything#asahi-dev · 2026-11-10 11:00
yukaintegralpilot reply to what? I'm not in the loop#asahi-dev · 2026-11-10 11:00
integralpilotOK sounds fine then. i'll send reply now#asahi-dev · 2026-11-10 11:00
chaos_princessworst case you will have t6091-mtp-j9000 as the compatible#asahi-dev · 2026-11-10 10:59
integralpilotto use SoC-based compatibles#asahi-dev · 2026-11-10 10:58
integralpilotbut that's just speculation so maybe it's fine for now#asahi-dev · 2026-11-10 10:58
integralpilotthe only thing is, its not really the SoC that decides it, it's if they use the cheap trackpad or expensive trackpad. so in future there could be one macbooks with and without STM that use the same SoC#asahi-dev · 2026-11-10 10:57
jannauseparate compatible for a device which behaves differently looks like intended use of compatibles#asahi-dev · 2026-11-10 10:52
integralpiloti have my draft reply written but ill leave it until yuka confirms what they think about stm because that only really impacts Neo and i don't want to agree to something and it turns out we'd rather do it another way#asahi-dev · 2026-11-10 10:51
integralpilotyes, i probably should have so its clearer, but for v6 its looking like ill have to move it to a generic file somewhere anyway#asahi-dev · 2026-11-10 10:50
jannauI would have added it to the dt-bindings to make it clear that the bootloader might/will set it#asahi-dev · 2026-11-10 10:49
integralpilotoh, perfect then! i must have remembered wrong (or maybe when i looked at the code a while ago i misunderstood what setprop did). thanks#asahi-dev · 2026-11-10 10:48
jannauintegralpilot: m1n1 does not care either way. it does `fdt_setprop_u32(dt, node, "hid-country-code", ...)` which will insert the property if it does not exists#asahi-dev · 2026-11-10 10:43
integralpilotbut if its the nicer DT design, "it would be harder in the bootloader" isnt rly a counterargument so maybe we'll just have to do it#asahi-dev · 2026-11-10 10:43
integralpilotalso the other thing they want is make "0" the default country code and not have to specify it explicity. so then m1n1 would have to be modified to instead of searching for "hid-country-code" to change, it would have to insert it in the first place. i mean we _can_ do that, it's more annoying, thoughts?#asahi-dev · 2026-11-10 10:39
integralpilotbecause there's no A18 pro MTP in the current series, that would mean in v6 i will have to make it always wait for STM, and then when we add A18 pro seperate compatible, the driver can be updated for it to have that different behaviour#asahi-dev · 2026-11-10 10:38
integralpilots/what/wants#asahi-dev · 2026-11-10 10:38
integralpilotyuka: the dt maintainer whats to drop apple,has-stm and just determine it from A18 Pro compatible. is that OK with you or?#asahi-dev · 2026-11-10 10:29
jannauI've tagged asahi-7.3.3-2 please do upgrade immediately from asahi-7.3.3-1#asahi-dev · 2026-11-10 10:23
integralpilotlets move to #asahi because this is more user support rather than development work#asahi-dev · 2026-11-10 10:22
N_Tempus_04no a asahi linux vm with appel hypervisor framework on appel silicon#asahi-dev · 2026-11-10 10:19
jannauchadmed: see what I pushed to asahi-wip#asahi-dev · 2026-11-10 09:58
integralpilotN_Tempus_04: do you mean a VM on asahi linux host, or asahi linux in a VM on a different host (in which, there's no point, just make a normal mainline linux VM)?#asahi-dev · 2026-11-10 09:57
integralpilot6GHz first draft almost done. now need to find which devices in my house can actually be a 6GHz AP to test it#asahi-dev · 2026-11-10 09:51
N_Tempus_04how can i make a asahi linux vm#asahi-dev · 2026-11-10 09:20
jannauhmm no. it selects the clock of a BE port#asahi-dev · 2026-11-10 08:59
jannauchadmed: MCLK_SEL suggests that it is not a bit pattern but a number which matches `writel_relaxed(cl->no + 1, cl->base + REG_SYNCGEN_MCLK_SEL);` in mca_fe_enable_clocks(). you wouln't select multiple clocks. so `6 + first_port_cl_no + 1` selects the clock of the first port#asahi-dev · 2026-11-10 07:34
chadmedwhatever ill send a fix tomorrow#asahi-dev · 2026-11-10 07:33
chadmedwe set bit 3 only if the port is not 0? and for port 0 we set all bits 0-2? #asahi-dev · 2026-11-10 07:33
chadmedoffsetting the port by 7 makes no sense to me, the resultant bits that get set in the mclk_sel register then dont make sense#asahi-dev · 2026-11-10 07:32
chadmedyeah putting that 7 back in fixes everything#asahi-dev · 2026-11-10 07:19
chadmeddidnt hit it during testing when i first submitted this because its sample rate dependent#asahi-dev · 2026-11-10 07:18
chadmedjannau: elisaado: i have reproed weirdness with clock settings here, i think i can fix this easily#asahi-dev · 2026-11-10 07:03
yukasadly not#asahi-dev · 2026-11-10 07:03
yuka> let me see if any of the fixes for the A18 Pro also get me further on M5 Pro#asahi-dev · 2026-11-10 06:30
yukapossible#asahi-dev · 2026-11-10 06:27
nickchanmaybe the other port being entirely non operational is too confusing for apple to ship in an environment where user have access?#asahi-dev · 2026-11-10 06:26
nickchan(if you connect two macs together then obviously one of them has to be the host and other has to be device and the share disk function makes you connect macs like that)