Dispatch syscalls 0x79/0x7A to the existing SIF register handlers

[llesieur99]

What this fixes

Dispatcher.cpp has no branch for syscalls 0x79 and 0x7A, so sceSifInitCmd fell through to "Unimplemented PS2 syscall" and the game never reached the SIF path.

Measured on Ratchet & Clank (SCUS-97199), 140-second run with the ISO:

	before	after
[Syscall TODO] lines, all v1=0x7a	31 254 (7.8 MB of error stream)	0
SIF register path	never reached	sceSifGetReg, sceSifSetReg, sceSifSetDma lines appear

Why 0x79 and 0x7A are SifSetReg and SifGetReg

Read off the game's own disassembly, not guessed:

0x118B60: `li v1,0x79; syscall`  called with (0x80000000, value) then
                                 (0x80000001, address)  -> two arguments
0x118B70: `li v1,0x7A; syscall`  called with (0x80000000), (4), (2)
                                 -> one argument

Two arguments is the setter, one is the getter. The SDK numbering agrees — 118 SifDmaStat, 119 SifSetDma, 120 SifSetDChain, 121 SifSetReg, 122 SifGetReg — and the three lower neighbours are already dispatched a few lines above in the same switch. Bit 31 of the register number is the PS2 convention (0x80000000 | SIF register).

What changed

Both handlers already exist in Stubs/SIF.cpp and read a0/a1, which is the syscall convention, so only the two dispatch branches were missing. The change is 19 lines in one file, no other behaviour touched.

Verification

• Built locally against c5a9d02 (MSVC Release, exit 0) and run with the game ISO, 140 s.
• The run above is the A/B: same binary except this commit (rac1-port-1 → rac1-port-2).
• Scope limit, stated plainly: this fix does not make the game run. Our port still stops later, in SIF RPC initialisation; no frame is rendered. The claim here is only that the two syscalls are dispatched and the SIF path becomes reachable.

The syscall numbers were also cross-checked against the published PS2 SDK numbering, but the measurement above is what I would look at first.