This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Recent change to mips_output_move


Eric Christopher <echristo@redhat.com> writes:
> > I personally find "move", "li" and "dli" more readable than things like
> > "or $4,$5,$0" and "addi $4,$0,0x1000".  Any chance of keeping them the
> > way they are?  After all, "move" always expands to a single insn, and we
> > already restrict ourselves to single-insn "li"s and "dli"s.
> 
> I don't think that'll be much of a problem. I suppose then we should
> change the .set nomacro warning to ignore those 3, since, really they're
> just aliases for readability.

As I understand it, .set nomacro only warns about multi-instruction
macros, not single-instruction ones.  We've been using move, li and
dli in delay slots for a long time.

In case it's any reassurance: for the last few months, I've been
testing with a version of gas in which all .set nomacro warnings
are hard errors.  I haven't seen it cause any problems.

> (Really should commit move_register to using or though...)

Why's that?

Richard


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]