Gentoo Archives: gentoo-hardened

From: "Peter S. Mazinger" <ps.m@×××.net>
To: gentoo-hardened@l.g.o
Subject: Re: [gentoo-hardened] TESTING kevquinn_overlays: amd64 hardened gcc - (bugzilla is down, i hope the mails are not)
Date: Mon, 08 Jan 2007 13:09:41
Message-Id: Pine.LNX.4.44.0701081351160.6783-100000@lnx.bridge.intra
In Reply to: Re: [gentoo-hardened] TESTING kevquinn_overlays: amd64 hardened gcc - (bugzilla is down, i hope the mails are not) by "Kevin F. Quinn"
1 On Sat, 6 Jan 2007, Kevin F. Quinn wrote:
2
3 [...]
4
5 > The cases internal to gcc/glibc are exceptions, in that the archives are
6 > used to build shared libraries (which use the lib*_nonshared.a archives
7 > which should always be PIC anyway) and those executables - which if we
8
9 lib*_nonshared.a does not have anything to do with this
10 libc_nonshared.a is used, because:
11 1. some objects have to go into each binary:
12 - __libc_csu_init/fini
13 - atexit
14 2. because glibc changed the stat interface and to make updates possible
15 they decided to put stat/fstat/lstat/fstatat/mknod/mknodat and their *64
16 counterparts into all binaries needing them
17 3. __stack_chk_fail_local to get rid of jump relocation on x86[_64] at
18 least (hardened/ssp)
19
20 haven't checked why __warn_memset_zero_len was added
21
22 for libpthread_nonshared.a, because:
23 1. pthread_atfork has to go into each binary
24
25 Note: for gcc libgcc.a is always -fPIC compiled (LIBGCC2_CFLAGS)
26
27 Peter
28
29 --
30 Peter S. Mazinger <ps dot m at gmx dot net> ID: 0xA5F059F2
31 Key fingerprint = 92A4 31E1 56BC 3D5A 2D08 BB6E C389 975E A5F0 59F2
32
33 --
34 gentoo-hardened@g.o mailing list