[oe] avr32-*-gdb requires modifying gdb.inc / gdb-cross.inc

Khem Raj raj.khem at gmail.com
Wed Apr 9 15:07:21 UTC 2008


Hi Geoffrey

I think it will be good to make this step arch specific to AVR only.
and best is to identify what changes are done in configure that needs it
to be regenerated. You can regenerate these and include them in the
patch itself instead of doing it at compile time.

Thanks

-Khem

On Wed, 2008-04-09 at 09:20 -0500, Geoffrey Wossum wrote:
> Hi all,
> 
> I currently have a BitBake recipe that can build gdb 6.4 for the AVR32, and 
> I'm about to make one that can build the gdb 6.7 using Atmel's newly released 
> patches for the AVR32.
> 
> According to this page, http://avr32linux.org/twiki/bin/view/Main/GDBPatches, 
> you must do some extra steps to build gdb if you patch it.  My experience 
> with Atmel's gdb 6.4 release was that even though they claim you don't have 
> to do these steps if you use the full source tarball, you have to anyway.
> 
> This requires modifying gdb.inc / gdb-cross.inc.  I think the modifications 
> are safe for all platforms, but I was hoping someone could comment before I 
> roll a patchset to submit.  
> 
> TIA,
> ---
> Geoffrey
> 
> 
> --- packages/gdb/gdb-cross.inc  2008-03-04 14:30:27.000000000 -0600
> +++ /home/geoff/lrs/playpaq/org.openembedded.dev/packages/gdb/gdb-cross.inc    
> 2008-04-09 09:09:50.000000000 -0500
> @@ -10,9 +10,38 @@ do_configure () {
>  # override this function to avoid the autoconf/automake/aclocal/autoheader
>  # calls for now
>         (cd ${S} && gnu-configize) || die "failure in running gnu-configize"
> +
> +
> +        # This stuff is required to build Atmel's patched gdb for the AVR32.
> +        # See http://avr32linux.org/twiki/bin/view/Main/GDBPatches
> +        (cd ${S}/bfd && autoreconf) || die "failure running autoreconf in bfd/"
> +        (cd ${S}/opcodes && autoreconf) || die "failure running autoreconf in opcodes/"
> +
>         oe_runconf
> +
> +
> +        # Also required according to the avr32linux GDBPatches page
> +        make configure-bfd || die "Error running configure-bfd"
> +        (cd bfd && make headers && cd "$OLDPWD") || \
> +                die "error running 'make headers' in bfd"
>  }
> 
>  do_stage () {
> +        # Left to fend for itself, "make install" will fail, saying that
> +        # makeinfo is missing.  I don't know if this is because my system
> +        # makeinfo isn't available inside the bitbake build, or if this is
> +        # due to the texinfo version detection bug that has plagued some
> +        # packages (see https://bugzilla.redhat.com/show_bug.cgi?id=345621).
> +        # Either way, we want to keep "make install"from wanting to run
> +        # makeinfo.  By copying over all the .info files before hand,
> +        # the files will exist in the build tree and have shiny new
> +        # timestamps, thus preventing "make install" from using makeinfo.
> +        cd ${S}
> +        for infofile in `find . -name "*.info"`; do
> +            cp "$infofile" ${B}/"$infofile"
> +        done
> +        cd "$OLDPWD"
> +
> +
>         oe_runmake install
>  }
> 
> 
> 
> 
> 
> _______________________________________________
> Openembedded-devel mailing list
> Openembedded-devel at lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel
-- 
Khem Raj <raj.khem at gmail.com>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://lists.openembedded.org/pipermail/openembedded-devel/attachments/20080409/4978d41b/attachment-0002.sig>


More information about the Openembedded-devel mailing list