GR8NET’s 2nd PSG is readable, and so is the one in 2nd gen OCM cores.
I don’t see a good reason for having a 2nd PSG be write-only (“just because MFR did it”). It must be readable in order to detect it. Without that ability it’s completely invisible to the software, and you can’t make sure your music and sound effects are played back the intended way. E.g. if you want to play SFX and music on different PSGs when a 2nd PSG is present, you can’t do this without detection ability.
In VGMPlay I try to detect it by reading port 12H, and if that fails I try to search slots for the MegaFlashROM SCC+ SD. If no 2nd PSG is found but a VGM does need a 2nd PSG for something I currently do output to port 10H “blindly”, but this inability to detect it does prevent VGMPlay from making the optimal playback device selection, and blocks certain future feature additions, so this fall-back behaviour may disappear in the future without mercy.
One can’t expect all software that supports a 2nd PSG to have complicated detection routines that need to be updated for every new device that has one. Especially not going to attempt to detect openMSX. Making the PSG readable is the proper, future-proof way. MFR gets an exception because it’s established for a long time now and not easy to update the FPGA.
C2 is write only I believe despite my pleas for read support so we can detect it properly… Something about FPGA space. I still hope read support will be added some time in the future.
For context, I found some old posts about PSG detection in VGMPlay here and here.
Latest patch for MoG for msx2, if it detects a scc and second psg both are enabled (I mean, the main menu at startup, it starts with it disabled if not detected but enabled if detected). Vgmplay latest version also, if it doesn't detect second psg on a a track that needs it (ie: choplifter arcade, vgmrips has it) the second psg will be isted with a question mark (not detected but trying to use it)
