I don't think there is any conversion between D3 and D2. D3 comes from animal sources like oily fish, production in the skin following UVB exposure, or by supplements made from lanolin. D2 comes from plants, especially mushrooms exposed to UVB light.
Both of them are converted slowly to their 25 hydroxy form in the liver, but they remain distinct: 25 hydroxy-D3 and 25 hydroxy-D2. Those are the value measured in the blood, and can vary a lot.
The homeostatic tight regulation happens in the kidneys: The hydroxy forms are converted to 1,25 dihydroxy D3 and 1,25 dihydroxy D2. These are the active forms that bind to cellular Vitamin D receptors. D3 binds a little better, but D2 still does the job.
I have read a few papers that say some immune cells can make their own active 1,25 dihydroxy form out of the 25 hydroxy form that gets into the cell from the blood or intracellular space. That would obviously escape the tight regulation performed by the kidneys.
One reason I say there is no D2<->D3 conversion is because doctors can tell if patients with high 25 hydroxy-D are taking supplements (almost entirely D3) or being treated by a doctor with Vitamin D injections (all D2 in the US).
As for D2 being safe and D3 being dangerous, I think the original UK scare about Vitamin D toxicity was all about D2. Sunlight exposure tops out at about 70 ng/ml of D3, which does not seem to hurt anybody but people with rare Vitamin D-related diseases. Toxicity seems to start at about 150 ng/ml.
I'm sure you know a lot more about the control and role of Calcium in the body. I have not really looked at that at all.