The Manager in question is definitely the Expansion Bus Manager. I named the function "FIRSTFUNC" only because it sits at the very start of the Manager code.
This morning I have identified about half of the external entry points to the Manager (all via the AAF3 ExpansionBusDispatch trap, so not from the Start Manger). Check the repo. But my disassembly of the "FIRSTFUNC" still looks like this:
EXPANSIONBUSMGRFIRSTFUNC:
42160 4E56 0000 Link A6, #$0
42164 4EBA 742A Jsr (* + $742C)
42168 4EBA 5074 Jsr __g__EXPANSIONBUSMGR_VEC0110
4216C 4EBA 2E82 Jsr (* + $2E84)
42170 4E5E Unlk A6
42172 4E75 Rts
It calls three functions, all within the Expansion Bus Mgr. The use of otherwise-rare JSR instructions, and the superfluous LINK/UNLK pair, reveal this to be a compiled C function. Indeed, most of the Manager is in C.
Here is a brief overview of the three functions being called:
1. Offset $49590- last function in Manager
- makes NameRegistry calls
- also called from unexplored function at offset $47F62
- 4 strings: "pci", "device-type", "name", "device-tree"
2. Offset $471DE- vector $208c/$0110 (no known name)
- no other callers
- 1 string: "aapl,pci"
3. Offset $44FF0- no other callers
- 1 string: "driver-reg,AAPL,MacOS,PowerPC"