This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: bfd/archures.c and bfd_mach_ppc
On Sat, Jan 26, 2002 at 08:45:33AM -0500, Jimi X wrote:
> >>>>> "AM" == Alan Modra <amodra@bigpond.net.au> writes:
>
> AM> Hmm, I guess the real bug is in
> AM> the order in which the archs are scanned. cpu-powerpc.c ought to
> AM> look at bfd_mach_ppc _last_. Could you try the attached diff?
>
> yes, this solves that problem, thankyou.
>
> This is no biggie.. but..
> I wanted to create a cpu-powerpc.c 64-bit default definition that was
> not the powerpc_620 since someone might define a core register set for
> it. So I added:
> {
> 64, /* 64 bits in a word */
> 64, /* 64 bits in an address */
> 8, /* 8 bits in a byte */
> bfd_arch_powerpc,
> bfd_mach_ppc,
> "powerpc",
> "powerpc:common",
> 3,
> BFD_DEFAULT_TARGET_SIZE == 64, /* default for 64 bit target */
> powerpc_compatible,
> bfd_default_scan,
> &bfd_powerpc_archs[13]
> },
>
> (following you changes I made it the second last), unfortunately now
> that bfd_mach_ppc still is 0 the first entry becomes the default.
>
> so I added:
> .#define bfd_mach_ppc64 1
> to archures.c and used it in the entry above.
>
> is this cool with you?
Yes, seems reasonable to me, although you might want to change the
printable_name entry to "powerpc:common64". Geoff Keating is powerpc
binutils maintainer though, so he might want to have a say. Geoff,
the above is on top of
http://sources.redhat.com/ml/binutils/2002-01/msg00545.html
Any objections to the above idea?
--
Alan Modra
IBM OzLabs - Linux Technology Centre