Allow 8 octets input signed extension function (64bit)

[emielsebastiaan]

Suggested changes (32bit -> 64bit)

1. In GP-0.5.0-eq:A.11: Allow 8 input octets for the signed extension function as a consequence of the change to a 64bit architecture. Direct evidence of this requirement in instructions 141 & 142.
2. In GP-0.5.0-eq:A.20: add mod2^32 to instruction 72.

[koute]

The limit of immediates being at most 4 bytes is a deliberate design decision. We don't want 8 byte immediates; it just makes recompilation into native code slower and more complex without any real benefit.

[emielsebastiaan]

@koute could you explain instruction 33 (GP-0.5.0) store_imm_u64.
If 8 byte immediates are not allowed (by design) how is this instruction different from instruction 32 store_imm_u32?

[koute]

8 byte physical (as encoded in the code blob) immediates are not allowed (except the load_imm_64 instruction). The immediate value (as encoded in the code blob) can be at most 4 bytes, but if necessary it is sign extended to full 64-bit. The difference between store_imm_u32 and store_imm_u64 is that store_imm_u32 will write the lower four bytes of the sign extended immediate to memory, and store_imm_u64 will write the full 8 bytes of the sign extended immediate.

[emielsebastiaan]

Thank you for the comments @koute
I changed this PR.
Two small suggestions remain.