With damping set to 0, I've had some very odd periodic noise impulse artefacts turning up in the output of FreeVerbStereo, even with the input at 0 volume over sustained periods.

Anyway, i spotted something which might be irrelevant but seems to cure the issue when changed:

In the last line of the for( loop in the update() function it has:

outblockR->data[i] = sat16(outputL * 30, 0);

Should this be:
outblockR->data[i] = sat16(outputR * 30, 0);
Instead?

Otherwise it seems to throw most of the allpass filter calculations for the right output away?
In any case, changing this cured my weird self-oscillation issue. I have no idea if it's an actual issue with the code or i've understood even less of what the Freeverb code is doing than I thought I did.