#asahi-dev · 2023-04-21 16:29
povikah, that was just something i introduced in https://github.com/AsahiLinux/linux/pull/112/commits/09789cd298aa90fb178958121c1f83aa424eb514#asahi-dev · 2023-04-21 16:29
povik16:57 < povik> though in my tree i have ps_sio_cpu for sio and ps_sio for the sio dart#asahi-dev · 2023-04-21 15:02
poviktransitional turning off of the sio or sio_cpu domains doesn't seem to matter before booting the coprocessor#asahi-dev · 2023-04-21 15:01
povikit won't be turned on again by the driver core before sio is probed?#asahi-dev · 2023-04-21 14:57
povikanyway can it be that issue then?#asahi-dev · 2023-04-21 14:57
povikwhich might be wrong actually since maybe it should be ps_sio_cpu for the sio dart per ADT#asahi-dev · 2023-04-21 14:57
povikthough in my tree i have ps_sio_cpu for sio and ps_sio for the sio dart#asahi-dev · 2023-04-21 14:55
povikha, i don't know that one#asahi-dev · 2023-04-21 14:44
povikunder a full OS install i see them maybe 1 in 5 boots#asahi-dev · 2023-04-21 14:43
povikhm, my starting position is that i don't see the failures at all under a development setup with everything built in#asahi-dev · 2023-04-21 14:33
povikyou mean, in that case it was unreliable, right?#asahi-dev · 2023-04-21 14:30
povikdoes compiling stuff as modules help?#asahi-dev · 2023-04-21 14:30
poviklet's see about reproducing the occasional boot failures#asahi-dev · 2023-04-21 14:30
povikso, sio confirmed booting on t6000 with existing loader code and in-kernel driver \o/#asahi-dev · 2023-04-16 15:17
povikyeah, maybe that pmdown_time i mentioned is what causes mute to happen later#asahi-dev · 2023-04-16 15:03
poviki will look at this more later#asahi-dev · 2023-04-16 15:03
povikmute can happen after hw_free so it's broken#asahi-dev · 2023-04-16 15:03
povikright#asahi-dev · 2023-04-16 15:01
povikerr, i though you wrote hw_params#asahi-dev · 2023-04-16 15:01
poviknot after prepare#asahi-dev · 2023-04-16 15:00
povikthough maybe pmdown_time messes up the ordering#asahi-dev · 2023-04-16 15:00
povikthat would still mean it's ok to turn off the clock there since hw_params happens *before* prepare#asahi-dev · 2023-04-16 14:59
povik16:55 <@marcan> hw_free is symmetrical to hw_params AIUI#asahi-dev · 2023-04-16 14:59
povikah, you are probably right#asahi-dev · 2023-04-16 14:59
povik16:55 <@marcan> shutdown is called when the device is closed#asahi-dev · 2023-04-16 14:49
povikunless the driver is unloaded#asahi-dev · 2023-04-16 14:48
povik* not sure 'shutdown' is ever called, usually#asahi-dev · 2023-04-16 14:48
povikalso i am not sure #asahi-dev · 2023-04-16 14:48
povikso that the way it was makes sense#asahi-dev · 2023-04-16 14:48
poviki thought hw_free is symmetrical to prepare, where we turn on the clocks#asahi-dev · 2023-04-16 14:48
povikmarcan: how did you check hw_free indeed happens before the mute?#asahi-dev · 2023-04-16 14:47
povikugh. tweaking asoc ordering...#asahi-dev · 2023-04-09 13:29
povik:p#asahi-dev · 2023-04-09 13:29
povikno, we don't#asahi-dev · 2023-04-09 13:29
povikmarcan: yes. we don't use them#asahi-dev · 2023-04-09 10:24
povikother than that there are holes in understanding of registers for the LEAP<->PDM glue but that does not block finishing the linux driver at this point#asahi-dev · 2023-04-09 10:24
povikrichyliu2: re microphones. it's pretty much figured out. there's been one puzzling thing about the LEAP processor that i finally cracked yesterday#asahi-dev · 2023-04-01 13:41
povikhaha, i think i have seen those#asahi-dev · 2023-04-01 12:36
povikand it fails earlier maybe becase the cable was plugged in at a different point in time?#asahi-dev · 2023-04-01 12:36
povikmy guess is the sio not crashing in this case is not the effect of dcp audio enabled, but because dcpext fails before it gets a chance to crash sio#asahi-dev · 2023-04-01 12:35
povik14:30 < jannau> with dcp and dcpext audio no crash, dcp works and dcpext fails with "aplay: set_params:1416: Unable to install hw params:"#asahi-dev · 2023-04-01 12:33
povikso only the audio-xmitter reference was missing#asahi-dev · 2023-04-01 12:33
povikthought you had to write that in#asahi-dev · 2023-04-01 12:33
poviki forgot i wrote the dpaudio1 node in DT#asahi-dev · 2023-04-01 12:29
povikwhile it's 0x64 for dpaudio0#asahi-dev · 2023-04-01 12:29
povikah, 0x66#asahi-dev · 2023-04-01 12:28
povikanyway there should be a dma 66 index, i think?#asahi-dev · 2023-04-01 12:28
povikcan i see it?#asahi-dev · 2023-04-01 12:28
poviki assume you added a dpaudio1 node#asahi-dev · 2023-04-01 12:25
povikthat's more like it :D