The patch was intruduced by Jackub and approved by Richard Henderson Please add if patches clean or is not in a branch update. On another note can we setup a toolchain@ alias vs just a gcc-porting@ I would like to be added to this.
on the request of someone from #gentoo-ia64 : <solar> if current could you please add a note to http://bugs.gentoo.org/show_bug.cgi?id=36496 (Fix IA-64 -pie support) <solar> just saying that -pie -fPIE seems to be there but the Scrt1.S from your system seems to never of been compiled. (attempting to build a trivial C program gives: (%:~/t)- gcc -fpie -pie test.c /usr/lib/gcc-lib/ia64-unknown-linux-gnu/3.3.2/../../../../ia64-unknown-linux-gnu/bin/ld: cannot open Scrt1.o: No such file or directory collect2: ld returned 1 exit status )
http://sources.redhat.com/ml/binutils/2004-01/msg00356.html From: Jakub Jelinek <jakub@redhat.com> Reply-To: Jakub Jelinek <jakub@redhat.com> To: binutils@sources.redhat.com Subject: [PATCH] Fix -pie on IA-64 Date: Fri, 23 Jan 2004 11:33:33 +0100 Hi! PIEs don't work at all in binutils CVS since November. The problem is that linker script puts RELA section in wrong order: .rela.dyn .rela.IA_64.pltoff .rela.opd while it of course should be .rela.dyn (where .rela.opd is included here) .rela.IA_64.pltoff On IA-64, DT_RELA..DT_RELA+DT_RELASZ covers all allocated .rela sections (correct), but if .rela.IA_64.pltoff is in the middle (and thus DT_JMPREL..DT_JMPREL+DT_PLTRELSZ), the dynamic linker is certainly not able to cope with it (and relocates .rela.IA_64.pltoff relocs twice, first time as in non-lazy binding and second time adds l_addr to it as in preparation for lazy binding). Ok to commit? 2004-01-23 Jakub Jelinek <jakub@redhat.com> * emulparams/elf64_ia64.sh: Put .rela.opd into OTHER_GOT_RELOC_SECTIONS instead of OTHER_PLT_RELOC_SECTIONS.
IA-64 seems to have many other shared object handing problems. And nobody seems to want wants to jump on this.. Changing resolution to UPSTREAM