On Sun, Jan 26, 2020 at 1:15 PM Joaquin de Andres <
address@hidden> wrote:
> > +struct CPUAVRState {
> > + uint32_t pc_w; /* 0x003fffff up to 22 bits */
> > +
> > + uint32_t sregC; /* 0x00000001 1 bit */
> > + uint32_t sregZ; /* 0x00000001 1 bit */
> > + uint32_t sregN; /* 0x00000001 1 bit */
> > + uint32_t sregV; /* 0x00000001 1 bit */
> > + uint32_t sregS; /* 0x00000001 1 bit */
> > + uint32_t sregH; /* 0x00000001 1 bit */
> > + uint32_t sregT; /* 0x00000001 1 bit */
> > + uint32_t sregI; /* 0x00000001 1 bit */
> > +
> > + uint32_t rampD; /* 0x00ff0000 8 bits */
> > + uint32_t rampX; /* 0x00ff0000 8 bits */
> > + uint32_t rampY; /* 0x00ff0000 8 bits */
> > + uint32_t rampZ; /* 0x00ff0000 8 bits */
> > + uint32_t eind; /* 0x00ff0000 8 bits */
> > +
> > + uint32_t r[NUMBER_OF_CPU_REGISTERS]; /* 8 bits each */
> > + uint32_t sp; /* 16 bits */
> > +
> > + uint32_t skip; /* if set skip instruction */
> > +
> > + uint64_t intsrc; /* interrupt sources */
> > + bool fullacc; /* CPU/MEM if true MEM only otherwise */
> > +
> > + uint32_t features;
> > +};
>
> Why not use 'uint8_t' for all sreg bits, ramps, eind and user registers,
> and bool for skip flag?
>
On the other hand, there are already appr. 25 features, and "uint32_t feature" won't be likely to be enough in near future. I suggested earlier "uint64_t" to be used.
Thanks,
Aleksandar