00:07 <valpackett> oh and the audio failure is due to the new shared gpio system it seems (with each woofer-tweeter pair sharing an lpass_tlmm gpio)

09:39 <konradybcio> Mohamed Mediouni: either the fastrpc repo or oneOf https://github.com/quic?q=hexagon&type=all&language=&sort= or https://github.com/qualcomm/qemu-hexagon-testing

09:40 <konradybcio> frankly I'm not sure, but the q6 folks actively poke at these repos so I would assume that's your best bet..

10:20 <mmediouni[m]> <konradybcio> "Mohamed Mediouni: either the..." <- fun random thing: that hexagon qemu repo doesn't even have dma. only the downstream one does. Ended up figuring out that v1 is supported on v68 but errors out if I mark bypass on for a VTCM destination...

10:21 <mmediouni[m]> downstream as in tarball in hexagon sdk

10:24 <mmediouni[m]> oh and is fastrpc known as broken on sc8280xp when using the dma_heap path? ended up filing https://github.com/qualcomm/fastrpc/issues/260 for that

10:25 <mmediouni[m]> it doesn't even pass a fastrpc_test for me with dma_heap enabled (note however that I have cma=512M as a boot arg if that matters) - but without it things do work...

10:26 <krzk> valpackett: thanks for the report on wsa884x EBUSY / shared gpios, good find :). brgl is already debugging/fixing it.

10:27 <brgl> yep, on it

10:27 <brgl> it's an unexpected interaction between explicitly shared GPIOs and reset-gpio doing implicit sharing

10:27 <brgl> at the same time