1 |
...ok, once again not really, but i cant use that trick to get people's |
2 |
attention after today so i figured i might as well use it twice. :) |
3 |
|
4 |
recent gcc ebuilds have been patched to recognise an environment |
5 |
variable, GCC_SPECS, that sets which specs-file should be used. the gcc |
6 |
3.4.2-r2 ebuild also builds both hardened and non-hardened specs files |
7 |
for all users (though it doesnt make hardened the default specs file for |
8 |
non-hardened users). |
9 |
|
10 |
so, what does this mean? it means that everyone can now assist in fixing |
11 |
hardened toolchain related bugs in the packages they maintain without |
12 |
having to recompile gcc to get a hardened toolchain up. good stuff, eh? :) |
13 |
|
14 |
|
15 |
ayanami root # gcc main.c -o main ; file main |
16 |
main: ELF 64-bit LSB executable, AMD x86-64, version 1 (SYSV), for |
17 |
GNU/Linux 2.4.1, dynamically linked (uses shared libs), not stripped |
18 |
|
19 |
ayanami root # export |
20 |
GCC_SPECS=/usr/lib/gcc/x86_64-pc-linux-gnu/3.4.2/hardened.specs |
21 |
|
22 |
ayanami root # gcc main.c -o main ; file main |
23 |
main: ELF 64-bit LSB shared object, AMD x86-64, version 1 (SYSV), not |
24 |
stripped |
25 |
|
26 |
|
27 |
so now, for anyone interested, helping with hardened now requires the |
28 |
absolute minimum effort possible and shouldnt be a pain for devs who |
29 |
dont want to have a full hardened install. go team! *high-fives Rob |
30 |
Holland for writing the patch* |
31 |
|
32 |
patched ebuilds: |
33 |
gcc-3.3.4-r2 |
34 |
gcc-3.4.1-r3 |
35 |
gcc-3.4.2-r2 |
36 |
|
37 |
|
38 |
Travis Tilley |
39 |
Gentoo/AMD64 |
40 |
|
41 |
-- |
42 |
gentoo-dev@g.o mailing list |