This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: [PATCH] MIPS gas: Remove superfluous CPU_* checks
On Fri, 23 Apr 2004, Thiemo Seufer wrote:
> For .set mipsN we lose the hilo_interlocks optimization for SB1 if a
> MIPS IV or lower ISA is chosen.
>
> We lose gpr_interlocks optimizations for VR5400, VR5500 with .set mips1.
>
> We lose cop_interlocks optimizations for VR5400, VR5500, SB1 with
> .set mips[1-3]
>
> I think dropping the optimization when the ISA level was explicitly
> lowered is an improvement. We should probably generally set the arch
Well, for me that's the expected behavior -- I'd be first to change the
semantics thus if I were hit by a problem due to a different one.
> to the default value of the .set ISA.
Definitely.
--
+ Maciej W. Rozycki, Technical University of Gdansk, Poland +
+--------------------------------------------------------------+
+ e-mail: macro@ds2.pg.gda.pl, PGP key available +