Difference between revisions of "S1C88 Core"

From SublabWiki
Jump to: navigation, search
(Flag and Exception Register)
Line 119: Line 119:
| Binary Coded Decimal Mode (8-bit add\sub)
| Binary Coded Decimal Mode (8-bit add\sub)
| Division By Zero
| EX4
| 5
| 5
| Low-Mask Mode (8-bit add\sub)
| Low-Mask Mode (8-bit add\sub)
| (Enable Exceptions \ Game cube emulator only)
| 6
| 6

Latest revision as of 20:02, 12 July 2015

S1C88 Overview


The S1C88 is an 8-bit microcontroller with 16-bit operations (designed by Timex, now Epson).The processor provides numerous addressing modes with a 24bit addressing bus (with only 21bits mapped externally)

Epson S1C88 Core manual

Additionally, the Minx provides the capibility to handle up to 32 hardware enabled interrupts with delayed response capbility. Up to 128 interrupt vectors may be specified, allowing the remaining 96 for BIOS calls.

The CPU is clocked at 4.00mhz, although the processor operates on a 4 cycle data access period, leaving the system with a theoretical limit of 1MIPS.

Minx Register Mapping

The Minx operates with a handful of registers. The CPU is an amalgamation of Z80 like paradigms combined with an 8-bit microcontroller like bank system.

General Purpose Registers
8 Bit Registers (Low) 8 Bit Registers (Hi) 16 Bit Register Index Register
X Xi
Y Yi
Fixed Function Registers
Register Description
PC Program Cursor
V PC Bank Register
U V Delay Register
SP Stack Pointer
F Flag Register
E Exception Register

Since the program cursor is only 16 bits, it uses a special "delayed" register to account for the upper 8 bits of program access space. When PC has it's most significant bit set, the register V takes the place of the upper 8 bits, extending PC out to 23 bits in total. To prevent bank switch problems, V is "delayed" by the means of register U. After each branch instruction, the value of U is copied to register V implicitly, allowing for full 23bit jumps without special programming tricks or special functions.

The Minx also provides additional facilities to access 24bit addresses using registers. X and Y both provide 24bit addresses using the Xi and Yi register as their upper 8 bits.

Flag and Exception Register

Flag Mapping
Bit Flag F. Mne. Exception E. Mne.
0 Zero Z  ?? EX0
1 Carry C  ?? EX1
2 Overflow O  ?? EX2
3 Sign S  ?? EX3
4 Binary Coded Decimal Mode (8-bit add\sub) BCD
5 Low-Mask Mode (8-bit add\sub) NIBBLE
6 Interrupt Disable ID
7 Interrupt Branch IB

While the F register can, in some cases, be treated as a general purpose 8-bit register, the exception register however is not directly accessible by any conventional means. It is also to be noted that the exception trapping needs to be "enabled" by some means we've not discovered yet. Division by zero causes the system to hard lock, and the existence of this register is only known through the reverse engineering of Pokemon Channel's internal emulator. The lower 4 bits of both registers are used for branch conditions and carry chaining for arithmetic. The upper 4 bits are "control" registers.

The I register

Unlike X and Y, the upper 8 bits of the remaining addressing modes are not unique. The register I provides a bank extension to the these remaining 24 bit accesses: [HL], [I+$nnnn], and [N+$nn]. It is generally good practice to maintain I as $00 unless otherwise necessary.

The N indexed mode

The N Indexed mode is most useful for accessing register memory quickly. N provides the mid byte of a 24 bit addressing mode, and the $nn is an 8-bit immediate. In example. [N+$8A] would point to $208A (VPU_CNT) if N = $20 and I = $00. It is rare to see N with any value other than $20, but it is not entirely out of the question to see it change.