Re: 2.4.0-test9 + Winchip2/2A processor family == hang on boot

From: Mikael Pettersson (mikpe@csd.uu.se)
Date: Wed Oct 11 2000 - 05:48:23 EDT

  • Next message: Benjamin Herrenschmidt: "Re: __bad_udelay in 2.2.18pre15"

    On Wed, 11 Oct 2000, Frank de Lange wrote:

    >Solved the problem with Winchip not booting. It turns out they do not like the
    >code generated by gcc (gcc version 2.95.2 19991024 (release)) when the
    >-march=i686 flag is set. Changing this flag to -march=i586 makes to box boot
    >like it should.

    That makes sense. The WinChip C6/2/3 processors were designed to
    be compatible with the Intel Pentium and Pentium MMX processors.
    They are _not_ compatible with P6 cores at the application level;
    in particular, they do not have the conditional move instructions.

    "-march=i686" permits gcc to use i686-specific instructions. Seriously
    uncool if the processor doesn't have them.

    I don't know if WinChips benefit from i686-like scheduling heuristics,
    but perhaps "-mcpu=i686" would be appropriate for the Makefile
    (i.e. use the i386 instruction set but schedule etc. as for an i686).

    /Mikael
    -
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.kernel.org
    Please read the FAQ at http://www.tux.org/lkml/



    This archive was generated by hypermail 2b29 : Wed Oct 11 2000 - 05:54:04 EDT