buzzmarshall has quit [Quit: Konversation terminated!]
vagrantc has quit [Quit: leaving]
tdebrouw has joined #linux-amlogic
<narmstrong>
tortoise: usually boards are based on reference designs, so you should try to boot it with meson-g12a-u200.dtb
<narmstrong>
but before dump the eMMC and try to decompile the vendor DT
<narmstrong>
then boot the board and dump the dmesg, /proc/interrupts, /sys/kernel/debug/gpio /sys/kernel/debug/pinctrl* to figure out which are the GPIOs and pins used
<narmstrong>
then you can copy the radxa-zero and modify it for your board
tolszak has joined #linux-amlogic
chewitt has joined #linux-amlogic
<narmstrong>
tolszak: ok I think pulsaudio is doing weird stuff
<tolszak>
narmstrong: I use pulseaudio, I thought it does weird stuff by design :)
<narmstrong>
tolszak: it tries to use PCM1, which is not configured with the LibreELEC soundconfig, PCM0 will output on HDMI & Analog
<narmstrong>
I see `hdmi-audio-codec hdmi-audio-codec.1.auto: Not able to map channels to speakers`, this means the EDID ELD is not correct for my monitor
<tolszak>
narmstrong: I just worked for me so I didn't dig it more. I've just noticed a lot of this fe.dai-link-1 dmesg messages
<narmstrong>
yep PA seems to try evevy PCM on each sound playback attempt
<narmstrong>
`pactl info` gives me correct infos
<narmstrong>
the card and profiles are correctly detected
<narmstrong>
Oh I know, the card definition sets axg-sound-card.pcm.iec958.0 on pcm1
<narmstrong>
but the N2 doesn't have spdif
<tolszak>
narmstrong: patch for axg-sound-card then?
<narmstrong>
tolszak: nop, it's the cons to have a generic multiplatform alsa sound card definition
<tolszak>
narmstrong: I mean patch for the file for hardkernel-odroidn2plus
<narmstrong>
no but you can add a different 0001-WIP-add-confs-for-gx-sound-card-and-axg-sound-card.patch without the SPDIF card only for odroid-n2
<narmstrong>
same for 0001-WIP-add-confs-for-gx-sound-card-and-axg-sound-card.patch
<narmstrong>
it should be simpler to have the axg-sound-card.conf & gx-sound-card.conf as files instead of patches
<narmstrong>
so we can add custom ones for boards
<narmstrong>
the patch would simple add the files to src/conf/cards/Makefile.am
<narmstrong>
indeed if I remove the iec958 card from /usr/share/alsa/cards/axg-sound-card.conf the error goes away
<narmstrong>
tolszak: do you want me to add these as comments in the PR ?
<tolszak>
narmstrong: Yes please, I'm currently distracted with something else
<narmstrong>
ack
buzzmarshall has joined #linux-amlogic
<chewitt>
@nar
<chewitt>
@narmstrong @tolszak the gx-sound-card conf is fine to include iec958 content because the hint in the conf ensures it's only used when device-tree has the S/PDIF nodes
<chewitt>
this doesn't work on the axg-sound-card conf, so it's always seen as present (even when it isn't) and this trips up pulse and other things
<narmstrong>
Ok this confirms what we see
<chewitt>
this means for AXG cards we need to do some level of runtime detection of what's in device-tree to present the right conf .. which will suck
<narmstrong>
But it’s only warnings, but we can add variant files for boarsds very easily with yocto
<chewitt>
for yocto and such where people are probably building for a specific board it's manageable
<chewitt>
for distros like LE/Armbian trying to support many boards from a single core image it's a royal pain in the arse
<chewitt>
alsa ucm2 stuff may be the way forwards, but that's rather sparsely documented
<chewitt>
and probably depends on per-card mappings to a config again
<chewitt>
per-board, not per-audio-card
jernej has joined #linux-amlogic
tolszak has quit [Ping timeout: 240 seconds]
elastic_1 has quit [Ping timeout: 240 seconds]
elastic_1 has joined #linux-amlogic
vagrantc has joined #linux-amlogic
mkorpershoek has quit [Quit: Connection closed for inactivity]