<jluthra>
pivi: i didn't know it by heart, but checking it in the CTRL_MMR region for am69, i see that AUDIO_EXT_REFCLK0/1_CTRL regs are at the same addresses as am62. That is REFCLK0_CTRL @ `0x1082e0` and REFCLK1_CTRL @ `0x1082e4`.
<jluthra>
Adding it inside the main syscon node for `0x100000` similar to AM62 should work in flipping the clock direction to soc->board
<jluthra>
I don't know why the registers are not part of the J784S4/AM69 TRM, I had to look up at internal documents
devarsh has quit [Quit: Connection closed for inactivity]
minash has quit [Remote host closed the connection]
minash has joined #linux-ti
<aradhya7>
pivi: we are changing how we support LVDS on AM62. OLDI TXes will be made independent. mainline eta is 6.10 at the moment.
<jluthra>
aradhya7: pivi is offline
pivi has joined #linux-ti
<pivi>
jluthra: thanks for looking into that. can you raise the issue of the missing documentation internally? we'd like to use such clocks for a design. wondering also who is going to do the required code changes ;-)
<pivi>
aradhya7: thanks for the update, looking forward to it. feel free to CC me (francesco@dolcini.it) on patches, so we can test it
mripard has joined #linux-ti
florian has joined #linux-ti
<rob_w>
is it considerable , that a proper timer from the arm side , is accessed directly by a pru , handle it and record its values in capture mode , into a dma for the arm to read later on ?
<rob_w>
hence accessing the timer registers and irq from within pru