[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [f-cpu] Error in f-cpu_opcodes.m4 or in manual

> i see a good reason for putting some of the information in the opcode.
> we have to encode the "combine" and the mux flags and also

For me combine and mux are different instruction, that mean different
opcode. If I correctly count this : we have 2 bits for this part.

> have a form with 8-bit immediate data. Since this needs 2 bits more
> than the 6-bit register field, some informations has to be pushed back into
> the opcode field.

So it's logical to put mux/combine in the opcode more than the logic function
to execute (which need 3 bits). And it's more clear in the manual to split
mux/combine/logic in different instruction even if we finally decide that
the some of the most signifient bits from the opcode field are the same
for combine/mux/logic.

> Finally, the .m4 files are often  more up to date than the manual.
> This is not always true (now that CÚdric maintains it) but the ROP2
> operations have been redesigned, tested and refined several times
> so i'm confident in the code.

I didn't understant the end of your sentence, did you start coding the
decoder, because we speak about opcode ?

> But if at least i had time to read the new version of the manual.....

It will be great. The part that need to be read are the one that are not
green at the head of the page. I propose a deal, you read the manual, I read
your last snapshot ;-) (Is it the one from june ?)

To unsubscribe, send an e-mail to majordomo@seul.org with
unsubscribe f-cpu       in the body. http://f-cpu.seul.org/