On Saturday 16 January 2010 16:39:38 Elden Crom wrote: > A few questions: > I need C++ for Coldfire/uClinux, > a) Does it already work? > either in, cvs cvs.uclinux.org:/var/cvs repository? > or in the svn sources.blackfin.uclinux.org/svn/toolchain/trunk? > (perhaps it made the m68k tool chain work for C++ by side effect?) The > http://lists.uclibc.org/pipermail/uclibc/2009-April/042347.html thread > seems to have fixed it, for m68k using the bfin line?
i dont do any m68k work, but this patch is the only one in the Blackfin tree that hasnt been pushed to the uclinux.org cvs. hopefully it'll be merged soon and then we'll match exactly. > On a sort-of different note, mostly philosophical, why isn't elf2flt a > separate application that is run after ld? (and would complain if ld left > out the reloc tables (no -r?)) I would seem easier to add a flag to gcc > that says there is one more step to go, than to do the fake out the ld, > real-ld, ld.real stuff. And what file does gdb read? It must read the > output of the normal ld since it still has enough in it, so you need that > file output basically all of the time anyway. (This question may be better > answered as an RTFM....point me to it and I'll read it) ld-elf2flt calls ld with specially crafted linker flags and a dynamic linker script to create the .gdb file which is then given to elf2flt to process relocs and such to make the final FLAT. while the linker options could possibly be moved into gcc specs, it'd be harder to move the linker script pieces. and we'd still be left with external elf2flt. in reality, people would like to see elf2flt integrated right into binutils. the problem is that what we have now is stable/known/working and there really isnt anyone clamoring for binutils integration, let alone anyone actually working on it. -mike
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ uClinux-dev mailing list uClinux-dev@uclinux.org http://mailman.uclinux.org/mailman/listinfo/uclinux-dev This message was resent by uclinux-dev@uclinux.org To unsubscribe see: http://mailman.uclinux.org/mailman/options/uclinux-dev
