This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Patch to correct LINK_PIE_SPEC
- From: "Peter S. Mazinger" <ps dot m at gmx dot net>
- To: Jakub Jelinek <jakub at redhat dot com>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Mon, 15 Nov 2004 14:23:45 +0100 (CET)
- Subject: Re: Patch to correct LINK_PIE_SPEC
On Mon, 15 Nov 2004, Jakub Jelinek wrote:
> On Mon, Nov 15, 2004 at 01:52:10PM +0100, Peter S. Mazinger wrote:
> > > Why should it? GCC -pie option means:
> > > 1) pass -pie to the linker if it supports it
> > > 2) user appropriate startfiles and/or libraries for PIE generation
> >
> > yes, that's it, use appropiate startfiles (and this appropiate startfile
> > should provide also path for Scrt1.o because the normal startfile
> > definition won't be used)
>
> So, can you summarize why do you think your patch is an improvement
> over current situation?
> If you use gcc -shared -nostartfiles -nostdlibs, it doesn't mean gcc
> doesn't pass -shared to ld either, and -pie should certainly behave
> consistently with that.
there is quite a difference between -shared -nostdlib and -pie -nostdlib.
a library can be built like this if it does not use constructors (so they
do not need crti.o and crtn.o either), a -pie exec can't be built w/o
Scrt1.o, unless glibc would provide a version that works w/o crti.o/crtn.o.
Peter
--
Peter S. Mazinger <ps dot m at gmx dot net> ID: 0xA5F059F2
Key fingerprint = 92A4 31E1 56BC 3D5A 2D08 BB6E C389 975E A5F0 59F2