Re: glibc-2.1.3, asm/elf.h and PPC kernels with AltiVec support

[Franz Sirl <Franz dot Sirl-kernel at lauterbach dot com>]

At 01:18 04.02.00 , Mark Kettenis wrote:
>    Date: Thu, 3 Feb 2000 15:04:38 -0800
>    From: Geoff Keating <geoffk@cygnus.com>
>
>    > From: Franz Sirl <Franz.Sirl-kernel@lauterbach.com>
>    > Date: Thu, 3 Feb 2000 22:19:51 +0100
>    > Cc: Andreas Tobler <toa@pop.agri.ch>
>
>    > >I think the correct fix would be a separate sys/elf.h for PPC, like 
> what
>    > >x86 and arm already do.
>
>    I think this is OK in principle.  We're trying to move away from using
>    kernel header files.
>
>But I'm not sure whether adding a seperate sys/elf.h for PPC is a good
>idea.  Why not add the necessary types directly to sys/procfs.h?  IMHO
>adding sys/elf.h for i386 was a mistake.  Simply copying the mistakes
>that were made for i386 to PPC doesn't seem very productive to me.
>
>The reason for the existence of sys/procfs.h is for the definitions
>for the layout of the special sections in ELF core-dumps.  The kernel
>uses names prefixed with elf_, but the kernel headers make them also
>available under the traditional SVR4 names.  In principle only these
>(thus without the elf_ prefix) are needed by GDB so one might choose
>to export these only.
>
>There is one caveat though.  The register sets used in core-dumps
>might differ from those used in signal contexts/user contexts.  This
>happens on Solaris and Linux/i386.  That's why prgregset_t and
>prfpregset_t exist.  On PPC the register sets seem to be identical so
>the `typedef gregset_t prgregset_t' in sys/procfs.h is probably right.

Well, including them in sys/procfs.h is fine for me, if that's the way to 
go. Probably the ARM people should revert their recently introduced 
sys/elf.h as well? At least that's what I used as a template :-) (I only 
look at x86 stuff as a last resort, other platforms are usually much cleaner).


>    > >If there are no objections, I'll put together such a file and post 
> it later
>    > >today with the corresponding changes to other files.
>
>    This seems to be missing a lot of stuff.  In particular, the
>    ELF_EXEC_PAGESIZE and ELF_CORE_COPY_REGS macros.
>
>Are these really meant to be used in userspace?  GDB doens't use
>them.  I doubt that they're really needed.

gdb builds fine for me, so it seems they are really not used. I couldn't 
find a single reference to these macros in the gdb sources too.

Franz.