Gentoo Archives: gentoo-hardened

From: "Javier Juan Martínez Cabezón" <tazok.id0@×××××.com>
To: gentoo-hardened@l.g.o
Subject: Re: [gentoo-hardened] Re: Hardened meeting summary 2010-05-16
Date: Tue, 18 May 2010 03:20:30
Message-Id: AANLkTilQJgIQ_NbukAQ7g228uDvuTB-Yifdwtb8jMHTd@mail.gmail.com
In Reply to: [gentoo-hardened] Re: Hardened meeting summary 2010-05-16 by Peter Hjalmarsson
1 AFAIK FORTIFY_SOURCE only works in fixed size buffers. To me ssp is a more
2 complete (and slightly different) approach, while FORTIFY_SOURCE checks the
3 existence of a buffer overflow directly, ssp does it by checking the
4 modification of the canary (indirect approach) but could get applied with
5 any kind of code since it's not limited to fixed size buffers. SSP to me is
6 really necessary
7
8 http://www.trl.ibm.com/projects/security/ssp/
9 http://gcc.gnu.org/ml/gcc-patches/2004-09/msg02055.html
10
11 2010/5/18 Peter Hjalmarsson <xake@×××××××××.net>
12
13 >
14 >
15 > I would say you're worrying too much.
16 >
17 > The important part in the toolchain equation is really PIE (and of
18 > course -z,now, relro and those other stuff people forgets about) to give
19 > you ASLR, and it is there in hardened gcc-4.3 in tree. SSP is also there
20 > to some extent because it is implemented in FORTIFY_SOURCE which is
21 > enabled in all of gentoo by default.
22 > So I would say that the extra part SSP from GCC is nice but not
23 > necessary.
24 >
25 > Regards
26 > Peter
27 >
28 >
29 >
30 >