Hello Roberto,
Keep in mind that more powerful macros can be written with CW assembler (not the inline version), than are available in C. The LMBD macro would appear sufficiently complex that a function may need to be used in lieu of a macro. Of course this will be somewhat slower than the assembly code macro, as will most code written in C.
It would appear that the LMBD macro takes the form it does so that the same number of execution cycles are required, independent of the return value. This may have to be foregone, as the C compiler will likely generate different assembly code than the macro. The equivalent C function would probably have the following form -
byte LMBD( byte val){ if (val >= 0x10) { if (val >= 0x40) { if (val >= 0x80) return 7; else return 6; } else { if (val >= 0x20) return 5; else return 4; } } else { if (val >= 0x04) { if (val >= 0x08) return 3; else return 2; } else { if (val >= 0x02) return 1; else return 0; } }}If you can tolerate significant variation in the execution cycles, depending on the output value, the following inline assembly code should produce a similar output to the macro.
byte LMBD( byte val){ __asm { LDA val; LDX #8;LOOP1: DECX; LSLA; BCC LOOP1; STX val; } return val;}After cursory examination, it would appear that the equivalent of INT_OBJ should be setup as a C structure for each interrupt source. INT_objtab would then become an array of structures.
Regards,
Mac
