This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: Various bugs on powerpc-unknown-linux-gnu
- To: Geoffrey KEATING <geoffk at discus dot anu dot edu dot au>
- Subject: Re: Various bugs on powerpc-unknown-linux-gnu
- From: Jeffrey A Law <law at cygnus dot com>
- Date: Mon, 20 Oct 1997 14:19:48 -0600
- cc: egcs-bugs at cygnus dot com
- Reply-To: law at cygnus dot com
In message <199709050504.PAA24721@discus.anu.edu.au>you write:
> * config/rs6000/t-ppccomm: Use -msdata=none for crtstuff.
I've checked in this patch.
I'm not all that familiar with the ppc ports, so if someone one
wants to double check my analysis, please do so.
Lots of ppc/rs6000 ports use t-ppccomm, which had me a little
worried that this change could break something.
However, it doesn't appear that any of the other ppc/rs6000
ports use crt* files from gcc, so in theory we should be OK.
For reference this is the patch I installed.
* Fix crtbegin.o, etc.; otherwise, the heads/tails of the constructor
chains can end up in .sdata rather than in .ctor/.dtor , which is not
helpful.
--- config/rs6000/t-ppccomm~ Sat Aug 16 04:58:22 1997
+++ config/rs6000/t-ppccomm Thu Aug 21 00:00:47 1997
@@ -67,3 +67,7 @@
$(T)scrt0$(objext): scrt0.c
$(GCC_FOR_TARGET) $(GCC_CFLAGS) $(INCLUDES) $(MULTILIB_CFLAGS) -c scrt0
.c -o $(T)scrt0$(objext)
+
+# It is important that crtbegin.o, etc., aren't surprised by stuff in .sdata.
+CRTSTUFF_T_CFLAGS = -msdata=none
+CRTSTUFF_T_CFLAGS_S = -fpic -msdata=none