This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: sh-hms does not build on the release branch
- From: Ralf Corsepius <corsepiu at faw dot uni-ulm dot de>
- To: Daniel Jacobowitz <drow at mvista dot com>
- Cc: Joel Sherrill <joel dot sherrill at OARcorp dot com>, binutils at sources dot redhat dot com
- Date: 16 Feb 2002 04:25:38 +0100
- Subject: Re: sh-hms does not build on the release branch
- References: <3C6D5779.CF017424@OARcorp.com> <20020215140345.A569@nevyn.them.org>
Am Fre, 2002-02-15 um 20.03 schrieb Daniel Jacobowitz:
> On Fri, Feb 15, 2002 at 12:46:17PM -0600, Joel Sherrill wrote:
> >
> > Nothing weird on this failure. :)
> >
> > sh-coff, sh-elf, sh-rtems, sh-rtemself all build OK
> >
> > sh-hms fails trying to link gas with an undefined
> > reference to md_pcrel_from in fixup_segment() in
> > the file obj-coff.c:4397
> >
> > With a quick glance, I don't see anything special
> > for sh-hms not even a configure stanza. It must
> > be using some default sh-*-* that is not the same
> > as sh-coff. But my grep for sh-*-* shows it often
> > on the same line as sh-rtems so I am perplexed.
>
> Ah, I suspected it might not build. sh-hms confuses me. I see
> configure triplets for it being added in the changelogs, but I don't
> see when they disappeared again. It appears to be a COFF target, but
> unlike sh-elf and sh-coff bfd_gas is never set. That may be the
> problem....
sh-hms had been the sh-coff GNU-toolchain variant provided by/for
Hitachi (hms = Hitachi Micro Systems). IIRC [1], except for the name,
there had not been any differences to sh-coff in binutils and only very
few minor differences in gcc/newlib.
> Is sh-hms still a supported target?
Dunno.
Ralf
[1] I had used it, when getting started with SH1s ca. 4-5 years ago and
still have