#asahi · 2021-10-11 09:39
marcanthe highperf ones are from the ADT; the lowperf ones are the boot values#asahi · 2021-10-11 09:35
marcanhttps://github.com/asahilinux/linux/tree/cpufreq/v1 if anyone wants to poke around before I send out an RFC#asahi · 2021-10-11 09:18
marcanyeah, I used that one#asahi · 2021-10-11 09:15
marcanmakes sense#asahi · 2021-10-11 09:14
marcankettenis: scalar types usually use maxItems 1 AIUI, though maybe it's not necessary for things with a type ref already?#asahi · 2021-10-11 09:13
marcanthe MCC driver I'm fixing up for submissing right now :)#asahi · 2021-10-11 09:09
marcanthat is how you do it#asahi · 2021-10-11 09:09
marcan(foo)#asahi · 2021-10-11 09:09
marcan description:#asahi · 2021-10-11 09:09
marcan maxItems: 1#asahi · 2021-10-11 09:09
marcan $ref: /schemas/types.yaml#/definitions/uint32#asahi · 2021-10-11 09:09
marcan apple,num-channels:#asahi · 2021-10-11 09:09
marcanyou need a $ref to the type#asahi · 2021-10-11 09:09
marcanj_ey: the error messages from the binding check kind of suck#asahi · 2021-10-11 09:05
marcan(e.g. for ANS2)#asahi · 2021-10-11 09:05
marcanyeah, we're already doing that#asahi · 2021-10-11 09:05
marcan(note that we don't have to have the same tree as the ADT, we know theirs is BS sometimes anyway)#asahi · 2021-10-11 09:05
marcanI think for most devices there would be one domain that is the device per se, and any other deps can be chained off of that#asahi · 2021-10-11 09:04
marcanwhat is your use case?#asahi · 2021-10-11 09:04
marcanthere isn't, not right now#asahi · 2021-10-11 09:04
marcanoh, I would not expect them to be ignored#asahi · 2021-10-11 09:03
marcan(which do support multiple deps properly)#asahi · 2021-10-11 09:03
marcanI would recommend using a single power domain for the audio devices, and instead representing extra dependencies in the power-controller nodes#asahi · 2021-10-11 09:03
marcanbecause the genpd framework doesn't know what to do with more than one#asahi · 2021-10-11 09:03
marcanyou need to explicitly instantiate them if there is more than one#asahi · 2021-10-11 09:02
marcanyes#asahi · 2021-10-11 09:02
marcanmultiple PDs require explicit device support#asahi · 2021-10-11 09:02
marcanah yes, you can't do that#asahi · 2021-10-10 17:55
marcanjust caused by the linux version then, probably?#asahi · 2021-10-10 17:54
marcanI see#asahi · 2021-10-10 17:54
marcanoh#asahi · 2021-10-10 17:54
marcanpovik: I2S probably depends on the SIO power domain, you'd better add that power-domain especially if you don't have console on UART (which would also keep it alive)#asahi · 2021-10-10 14:42
marcanyay, I think I fixed the i2c/hpm init fail in m1n1#asahi · 2021-10-10 14:31
marcanif it's supposed to work for outputs then I guess that makes sense#asahi · 2021-10-10 14:30
marcanor is it only used for inputs?#asahi · 2021-10-10 14:30
marcanis that function supposed to return the written value for outputs?#asahi · 2021-10-10 14:30
marcanthat looks weird#asahi · 2021-10-10 14:25
marcanyou don't *want* to update it when you read the pin#asahi · 2021-10-10 14:25
marcanthe cache is for writes#asahi · 2021-10-10 14:25
marcanalyssa: but the cache doesn't care about the read bit#asahi · 2021-10-10 13:17
marcanyou can just reply to the review email with your findings#asahi · 2021-10-10 13:15
marcanj_ey: if it doesn't implement caching on demand only then maybe it's not a good fit#asahi · 2021-10-10 13:02
marcanhopefully regmap can do that for you, otherwise it'd be quite broken#asahi · 2021-10-10 13:00
marcanthough this fabric is nGnRnE so writes stall too, but not as badly since they have no deps#asahi · 2021-10-10 13:00
marcan(which are slower since they stall)#asahi · 2021-10-10 13:00
marcanj_ey: in general caching is not bad, avoids reads#asahi · 2021-10-10 02:38
marcanthey're trying to buy ARM#asahi · 2021-10-10 02:37
marcanI sure hope that gets blocked#asahi · 2021-10-10 02:36
marcansimilarly, I'm pretty sure ARM can't stop you from releasing an ARM7 compatible CPU or something, though they might try#asahi · 2021-10-10 02:35
marcanwith SSE2