I don’t believe we can drive these conclusions purely on what the M2 mini can do. It still would depend on how the eDP is hooked up internally. It might still be wired on the laptops in such a way that it bypasses the crossbar.
I assure you it's safe to draw such conclusions based only on M2 mini capabilities. I'm just going to have to draw out a block diagram, too many people aren't familiar with how this kind of thing works inside chips.
SERDES/PHY: SERializer/DESerializer. DP, USB3, and Thunderbolt (TBT) are all based on very high speed (multi-gigabit) serial links. A SERDES converts between a parallel data path at a slower clock speed and a serial data path at a much higher clock speed. For example, in PCI Express 1.0, serial line rate is 2.5 Gbps (2.5 GHz) and parallel line rate inside a chip is usually one of 156.250 MHz (16-bit datapath) or 312.500 MHz (8-bit datapath).
Together with line drivers on the serial side of the SERDES (responsible for matching the electrical signaling standards for the respective standards), equalizers, and other mixed-mode (analog+digital circuits), this block comprises the PHY or PHYsical Layer.
eDP: Embedded DisplayPort, a variant of DP designed for internal displays of notebooks and similar things. Mac Mini motherboards connect this to a DP-to-HDMI conversion chip for the HDMI port.
DCP: Display Coprocessor, a block in Apple SoCs which periodically reads display referesh data from a frame buffer somewhere in unified memory, formats it as DisplayPort packets, and sends it through a parallel link for internally transporting DP packets.
Crossbar Switch: Has several inputs and outputs, can route any input to any output. (But not two inputs to the same output.)
The dashed lines are the connections new to M2/M3 which did not exist in M1. They absolutely must exist in M2, otherwise there would be no way for DCP 1 to drive a display connected to one of the Type C PHYs. We know it can do that because you can connect two displays to the M2 Mini's two Type C ports.