This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: MIPS libstdc++-v3 x86-x-mips-elf doesn't build
- To: "H . J . Lu" <hjl at lucon dot org>
- Subject: Re: MIPS libstdc++-v3 x86-x-mips-elf doesn't build
- From: Benjamin Kosnik <bkoz at redhat dot com>
- Date: Mon, 30 Jul 2001 18:46:45 -0700 (PDT)
- cc: Graham Stott <grahams at redhat dot com>, echristo at cygnus dot com, gcc-bugs at gcc dot gnu dot org, libstdc++ at gcc dot gnu dot org
Well, whatever.
I would prefer to have the mips port work like all the other ports.
Having some of the ports use this header file, and define _MIPS_ISA or
__mips, or whatever is bogus. There's got to be something consistent that
the configure scripts can use to figure out what the eventual target is.
Eric, I've cc'd you because I believe you are the mips maintainer. If you
can think of a way to do this, or can suggest a sane alternative, please
advise.
In the meantime, the configure.target bits that turn on mips atomicity.h
are being reverted. Sorry, HJ.
(In addition, the generic bits in mips/bits/atomicity.h have been excised.)
2001-07-30 Benjamin Kosnik <bkoz@redhat.com>
* configure.target: Remove mips from cpu table.
* config/cpu/mips/bits/atomicity.h: Remove generic
versions. Comment sgidefs include.
-benjamin
On Mon, 30 Jul 2001, H . J . Lu wrote:
> On Mon, Jul 30, 2001 at 10:45:39AM -0700, Benjamin Kosnik wrote:
> >
> > > All targets which define _MIPS_ISA have sgidefs.h.
> >
> > Can you find a way to differentiate between the two without _MIPS_ISA or
> > sgidefs.h?
>
> The whole idea of config/cpu/mips/bits/atomicity.h is to use the MIPS
> II instructions when gcc is generating code to run on MIPS II or above.
> The only reliable way I know of to detect it in the preprocessor is to
> check _MIPS_ISA. If a target doesn't define _MIPS_ISA, which requires
> sgidefs.h, it doesn't make any senses to use a mips specific
> atomicity.h since the only thing it does is to support _MIPS_ISA.
>
> BTW, some mips targets have
>
> %{mips4:-U__mips -D__mips=4 -D__mips64}
>
> But there is nothing for -mips2. If we want to add it for other mipss
> target, I'd suggest
>
> http://gcc.gnu.org/ml/gcc-patches/2001-07/msg01937.html
>
>
> >
> > Again, what is the intent here? Why can't something like the sparc
> > configuration process be used?
>
> You lost me here. I am not familiar with the sparc configuration
> process.
>
>
> H.J.
>