i haven't tested the uart mode with this: hopefully it still works with that.
even with this, i get several hp-detect interrupts, which comes with its
own problem (i.e. pipewire restarts every second).
i didn't check the routing on those, but after applying this in full, 3.5mm-attached mic has way more volume, and i *think* this patch is what fixed that (mic bias)
the USB PHY which links RK3399 to the modem needed to be enabled, then the rest works OOTB
eg25-control works; modem-manager sees the modem, i haven't actually tried to get an IP address yet though
it might be getting further in the setup than before, but it's still not complete
eg25-control --power-on does show it sending the STATUS signal, but /dev/ttyUSB2 does not appear and modem-manager sees nothing.
eg25-manager likewise does not work
TODO: try lsusb?
combined with an out-of-tree ucm file, the audio device can actually make noise. no setup needed -- just run speaker-test after boot
this patch probably changes more than is strictly required: i can maybe reduce it over time
this is enough for the sound card to come online, visible in /dev/snd and also pipewire
however i don't *hear* anything. speaker-test -c 2 does not work
the codec is probed, and i think enumerated over i2c,
but `aplay -l` shows empty.
earlier version omitted one of the power supplies and the kernel driver
complained:
```
[ 33.506926] rt5640 1-001c: Device with ID register 0x0 is not rt5640/39
```
since this version of the patch doesn't have that error, i *think* the
driver read the device's id register over i2c, at least.
the device shows up at:
```
/sys/bus/platform/drivers/rk3x-i2c/ff110000.i2c/i2c-1/1-001c
```
but this is a static definition; it doesn't seem to say anything about
the runtime status of the device.
