cross question on egcs pre-1.1 testing

[Ralf Corsepius]

Ralf Corsepius wrote:

> Ralf Corsepius wrote:
>
> > joel@OARcorp.com wrote:
> >
> > > On Thu, 27 Aug 1998, Ralf Corsepius wrote:
> > >
> > > > Jeffrey A Law wrote:
> > > >
> > > > >   In message < Pine.LNX.3.96.980824141127.12612m-100000@oar3remote.oarcorp.com >you write:
> > > > >
> > > > >   > I am in the process now of building all the RTEMS configurations using
> > > > >   > egcs-19980816 for C and C++.  If they all build without problems, I
> > > > >   > am thinking about doing it again and include the other languages.
> > > > > Cool.
> > > >
> > > > It will fail for the SH, while building for the SH3E:
> > >
> > > This is the only failure for any of the *-rtems configurations in
> > > egcs-19980816.
> >
> [..]
> >
> > IMO,  egcs/gcc's  backend for the SH contains some severe bugs. However, I've not yet been able
> > to build isolated test cases to demonstrate these bugs.
> [..]

In the meantime, I've been able to isolate one of these cases.

The code fragment below (stripped down from rtems' networking code) causes an "Internal compiler
error: program cc1 got fatal signal 11" error when being cross-compiled with -O2 or greater with
egcs-19980824 for an sh-rtems/sh-coff target on an i686-pc-linux-gnulibc1 host.

BTW: This fragment compiles with egcs-1.0.3a.

Ralf.

--
Ralf Corsepius
Forschungsinstitut fuer Anwendungsorientierte Wissensverarbeitung (FAW)
Helmholtzstr. 16, 89081 Ulm, Germany     Tel: +49/731/501-8690
mailto:corsepiu@faw.uni-ulm.de           FAX: +49/731/501-999
http://www.faw.uni-ulm.de