CVS head do not bootstrap for powerpc

[Olaf Hering]

 On Thu, Jun 26, Jakub Jelinek wrote:

> On Thu, Jun 26, 2003 at 04:52:39PM +0930, Alan Modra wrote:
> > On Wed, Jun 25, 2003 at 08:47:30PM +0200, Olaf Hering wrote:
> > >  On Wed, Jun 25, Edmar Wienskoski wrote:
> > > 
> > > > collect2: ld terminated with signal 11 [Segmentation fault]
> > > 
> > > gcc bug, introduced after last friday triggered a binutils bug.
> > > The binutils bug was already fixed in the cvs. But the gcc bug was not.
> > [snip]
> > > /usr/powerpc64-suse-linux/bin/ld: .libs/strstream.o(.gnu.linkonce.t._ZTv0_n12_NSt10istrstreamD1Ev+0x10): R_PPC_PLTREL24 reloc against local symbol
> > 
> > See http://sources.redhat.com/ml/binutils/2003-06/msg00769.html
> > I wouldn't be surprised to find that the powerpc problem was introduced
> > by the same patch that caused the hppa problem.
> 
> I have bootstrapped/regtested the patch on gcc-3_3-rhl-branch for ia32,
> ia64, amd64, ppc32, ppc64, s390, s390x, so if it does not work on the trunk,
> it means rs6000.c output_mi_thunk lost code to handle binds_local_p symbols.
> Looking at the differences gcc-3_3-rhl-branch emits asm using fprintf
> directly while TRUNK creates rtl and then outputs it, so likely it is that
> change where it got lost.

The bug was introduced here:
cvs diff -pu -D 20030623 -D 20030624 

I see no rs6000 specific names there.

Gruss Olaf

-- 
USB is for mice, FireWire is for men!