[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: glibc problem pinned down
From: |
Mark Kettenis |
Subject: |
Re: glibc problem pinned down |
Date: |
Wed, 18 Jul 2001 23:08:34 +0200 |
From: Roland McGrath <roland@gnu.org>
Date: Wed, 18 Jul 2001 16:31:16 -0400 (EDT)
> I noticed that crt1.o is in /lib, while crtbegin.o, crtbeginS.o, crtend.o
> and crtendS.o are in /lib/gcc-lib/i386-gnu/2.95.4. As such they seem to be
> part of the compile, while crt1.o is not. Is it okay to use the Linux
> versions of crtbegin/crtend, and just drop the bad crt1.o link?
Yes. Those files are part of the GCC installation, and they are generic to
any ELF/x86 platform.
Well, at least they should be. The Hurd versions of crtbeginS.o and
crtendS.o in GCC 2.95.x are a bit broken since they are not compiled
with -fpic. I believe it doesn't really matter, but it's just a bit
unoptimal since it keeps the kernel from sharing the pages that code
is in. The Linux versions of those files in GCC 2.95.x where x < 4
contain a bit of junk that keeps them from working with recent
versions of libc. But the Linux files from GCC 2.95.4 should be fine.