>Hello penguinoids,
>
>I have wrote a small patch against 2.1.131-ac9 which implements the
>promised enhanced support for the new egcs/pgcc i386 target submodel
>options such as -mpentium, -mpentiumpro, -march=pentium and
>-march=pentiumpro.
>
>The patch should be pretty self-explanatory and touches:
>+++ linux-2.1.131-ac9-ma/Documentation/Configure.help Sun Dec 13 05:15:06
1998
>+++ linux-2.1.131-ac9-ma/arch/i386/Makefile Sun Dec 13 05:14:09 1998
>+++ linux-2.1.131-ac9-ma/arch/i386/config.in Sun Dec 13 05:14:21 1998
>
>
>
>Some questions left:
>
>1. Does gcc 2.8 support the aforementioned options?
>
Yes it does. BTW, GNU is recommending the -m{cpu name} not be used as it is
going to be fazed out in the future. The new style
is -mcpu=pentium, -mcpu=pentiumpro. (also -mcpu=586,686 if I remember
correctly)
And pegcs has -mamdk6 for the AMD K6, and I think another one for Cyrixes
too.
BTW, according to pgcc's web site to get the maximum, safe, optimizations
using their compiler you should use the -O6 option.
>2. What is the minimum egcs/pgcc version for these options?
>
The pentium and pentium pro options appeared in gcc 2.8.0, which is what
egcs is based on, and pgcc is an addon to egcs.
>3. What is the reason a 486 configured kernel will not run on the 386?
> It's evidently not the compiler as far as I can tell from my
> gcc-2.7.2.x/pgcc-1.1.1 documentation.
>
>
If I remember correctly it has something to do with the alignment in memory.
Not sure though.
>
>And *please*, do check the Configure.help if the contents are correct.
>There are probably some gcc/egcs/pgcc gurus out there which can comment
>on for-/backwards compatibility, particularly in context with the
>-march= option.
>
>Please Cc: your answers/comments.
>
>Happy compiling,
>
>--
>Matthias Andree
>
-
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/