I'd like to ask the list if there's any reason the
arch/i386/Makefile includes "-m486" in the CFLAGS for the
586 and 686 configurations. Those of us using recent EGCS
compilers (for instance) like to take advantage of the Pentium
and Pentium Pro/II optimizations via "-mpentium -march=pentium"
and "-mpentiumpro -march=pentiumpro", respectively. It is my
understanding that some of these optimizations are enabled by
default, even if the compiler is not provided the "-mpentium"
or "-march=" flags. It's also my understanding that specifying
the "-m486" option will explictly disable any Pentium or
Pentium Pro/II optimizations. This is not good.
While it's not too inconvenient to open up the Makefile and
alter it to include my options, I wonder how many people
grab the latest EGCS, do a 3-stage compile with their specific
optimizations, and end up compiling a kernel for a 486 without
knowing it. :)
It would be convenient for me to put the "-mpentium
-march=pentium" flags in the 586 config (and likewise for
686), but I recognize that older versions of GCC might balk
at this. That's fine. I would like to suggest that the
-m486 be taken out of those blocks, though. Is this a bad
idea?
-- Shaw Terwilliger- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/