✨ Add CFU R4-type instructions

[stnolting]

This PR adds support for RISC-V R4-type instructions to the Custom Functions Module (CFU).

• User can now implement custom R4-type instructions that operate on three source registers (rs1, rs2, rs3).
• These new custom instruction type use the RISC-V custom-1 opcode.
• The CFU instruction "primitives" are reworked. Noe there are just 2 pre-defined primitives for utilizing custom CFU instructions:
  • neorv32_cfu_r3_instr(funct7, funct3, rs1, rs2)
  • neorv32_cfu_r4_instr(funct3, rs1, rs2, rs3)
• A compatibility wrapper for the outdated neorv32_cfu_cmd*() CFU instructions is added (for backwards-compatibility).
• The CFU hardware module and the according demo program are updated to provide a R4-type instruction example: multiply-add (rd <= rs1 * rs2 + rs3)
• Rework and update documentation (section Custom Functions Unit (CFU)).