Re: [avr-libc-dev] RFD: more avr-libc API changes

[Björn Haase]

Joerg Wunsch wrote on Freitag, 9. September 2005 23:39 :
> Our release policy says a major # bump is required iff an
> incompatibility with some GCC version arises.  This is currently not
> the case (but if I understood Björn right, that could happen in the
> near future).
In fact, I don't think so. The planned change in the main function (so that 
main() does no longer initialize the stack pointer) should not be a problem, 
IIUC, since the stack pointer is already initialized by the device-specific 
crt.?

One thing, however, that I think that we might want to change some day is the 
__tmp_reg__,__zero_reg__ issue. I assume that the original decision to use 
r0/r1 stems from times when there were no hardware multipliers. IMO it would 
be better to use r2:r3 instead so that one could, e.g., avoid all of those 
"clr __zero_reg__" in the multiplication operations. (At the same time one 
could start adding (use (reg:QI ZEROREG)) and (clobber (reg:QI TEMPREG)) for 
all the instructions that use these registers so that finally the prologue 
sequences have a possibility to find out if they need to save/restore them.)

While I cannot predict what the maintainers will judge to be useful in the 
future: My personal recommendation is just to have such a possible future 
re-numbering in mind when working on avr-libc.

Yours,

Bjoern