Re: [PATCH 2/3] Add basic Mips core tune config

[Richard Purdie <richard.purdie@linuxfoundation.org>]

On Tue, 2011-07-26 at 09:41 -0500, Mark Hatle wrote:
> Few quick items here:
> 
> On 7/26/11 7:44 AM, Richard Purdie wrote:
> > Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
> > ---
> >  meta/conf/machine/include/mips/arch-mips.inc |   64 +++++++++++++++++++++++++-
> >  meta/conf/machine/include/tune-mips32.inc    |   10 +++-
> >  2 files changed, 71 insertions(+), 3 deletions(-)
> > 
> > diff --git a/meta/conf/machine/include/mips/arch-mips.inc b/meta/conf/machine/include/mips/arch-mips.inc
> > index f7f4eed..071e6b5 100644
> > --- a/meta/conf/machine/include/mips/arch-mips.inc
> > +++ b/meta/conf/machine/include/mips/arch-mips.inc
> > @@ -1 +1,63 @@
> ...
> 
> > +# Floating point
> > +TUNEVALID[fpu-hard] = "Use hardware FPU"
> > +TUNE_CCARGS += "${@bb.utils.contains("TUNE_FEATURES", "fpu-hard", "-mhard-float", "-msoft-float", d)}"
> > +TARGET_FPU = "${@bb.utils.contains("TUNE_FEATURES", "fpu-hard", "", "soft", d)}"
> > +
> > +# Package naming
> > +MIPSPKGSFX_ENDIAN = "${@bb.utils.contains("TUNE_FEATURES", "bigendian", "", "el", d)}"
> > +MIPSPKGSFX_BYTE = "${@bb.utils.contains("TUNE_FEATURES", "n64" , "64", "", d)}"
> 
> n32 is also MIPS64.
> 
> so a:
> 
> MIPSPKGSFX_BYTE .= "${@bb.utils.contains("TUNE_FEATURES", "n32" , "64", "", d)}"
> 
> should fix the issue.
> 
> > +TUNE_ARCH = "mips${MIPSPKGSFX_BYTE}${MIPSPKGSFX_ENDIAN}"
> > +
> > +# Base tunes
> > +AVAILTUNES += "mips mips64-n32 mips64 mipsel mips64el-n32 mips64el mips-nf mips64-nf-n32 mips64-nf mipsel-nf mips64el-nf-n32 mips64el-nf"
> > +TUNE_FEATURES_tune-mips = "o32 bigendian fpu-hard"
> > +BASE_LIB_tune-mips = "lib"
> > +TUNE_FEATURES_tune-mips64-n32 = "n32 bigendian fpu-hard"
> > +BASE_LIB_tune-mips64-n32 = "lib32"
> > +TUNE_FEATURES_tune-mips64 = "n64 bigendian fpu-hard"
> > +BASE_LIB_tune-mips64 = "lib64"
> > +TUNE_FEATURES_tune-mipsel = "o32 fpu-hard"
> > +BASE_LIB_tune-mipsel = "lib"
> > +TUNE_FEATURES_tune-mips64el-n32 = "n32 fpu-hard"
> > +BASE_LIB_tune-mips64el-n32 = "lib32"
> > +TUNE_FEATURES_tune-mips64el = "n64 fpu-hard"
> > +BASE_LIB_tune-mips64el = "lib64"
> > +TUNE_FEATURES_tune-mips-nf = "o32 bigendian"
> > +BASE_LIB_tune-mips-nf = "lib"
> > +TUNE_FEATURES_tune-mips64-nf-n32 = "n32 bigendian"
> > +BASE_LIB_tune-mips64-nf-n32 = "lib32"
> > +TUNE_FEATURES_tune-mips64-nf = "n64 bigendian"
> > +BASE_LIB_tune-mips64-nf = "lib64"
> > +TUNE_FEATURES_tune-mipsel-nf = "o32"
> > +BASE_LIB_tune-mipsel-nf = "lib"
> > +TUNE_FEATURES_tune-mips64el-nf-n32 = "n32"
> > +BASE_LIB_tune-mips64el-nf-n32 = "lib32"
> > +TUNE_FEATURES_tune-mips64el-nf = "n64"
> > +BASE_LIB_tune-mips64el-nf = "lib64"
> > +
> > diff --git a/meta/conf/machine/include/tune-mips32.inc b/meta/conf/machine/include/tune-mips32.inc
> > index 28b0047..1f913df 100644
> > --- a/meta/conf/machine/include/tune-mips32.inc
> > +++ b/meta/conf/machine/include/tune-mips32.inc
> > @@ -1,4 +1,10 @@
> > +DEFAULTTUNE ?= "mips32"
> > +
> >  require conf/machine/include/mips/arch-mips.inc
> >  
> > -TUNE_CCARGS = "-march=mips32"
> > -TUNE_PKGARCH = "mips"
> > +TUNEVALID[mips32] = "Enable mips32 specific processor optimizations"
> 
> If tune-mips32 is used, then all of the n32 and n64 variants are not possible.
> Do we have a way to specify that?  Perhaps using the TUNECONFLICT?

Agreed, TUNE_CONFLICTS is the way to go. I've updated the patch on my
branch with this and the other tweak:

http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=rpurdie/ml4&id=2a31a7d38149b6bc972831964bfdeae7422c128c

Cheers,

Richard