1 |
Hemmann, Volker Armin wrote: |
2 |
> and I get: |
3 |
> |
4 |
> strace gzip > /dev/null |
5 |
> execve("/bin/gzip", ["gzip"], [/* 63 vars */]) = 0 |
6 |
> <snip> |
7 |
> open("/lib/tls/libc.so.6", O_RDONLY) = 3 |
8 |
> <snip> |
9 |
> mmap(NULL, 2261000, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = |
10 |
> 0x2aaaaabc3000 |
11 |
> <snip> |
12 |
|
13 |
Thanks for the reply. So it seems you were right about Gentoo's |
14 |
versatility also showing in different addresses used for loading glibc. |
15 |
From your output i'm guessing you are using the linuxthreads-tls USE |
16 |
flag for glibc and this does seem to cause a different address for |
17 |
glibc. Is this the case or do you know what causes these differences? |
18 |
|
19 |
So if there are more of these things that can cause different addresses, |
20 |
that in itself would mean some sort of randomization. It would be |
21 |
interesting to see though how many variations can actually occur, i.e. |
22 |
if every different configuration does indeed cause different addresses. |
23 |
So if anyone else wants to post their findings we could see just how |
24 |
many variations exist, although maybe that would be helping the wrong |
25 |
sort of people, if you know what i mean... |
26 |
Alternatively, if someone knows the answer without checking on lots of |
27 |
different installs, i'd be glad to know. |
28 |
|
29 |
BTW, i also found out that "ldd /bin/gzip" will show you the addresses |
30 |
of the libs, no need to wade through the strace output. |
31 |
|
32 |
Thanks, |
33 |
Marco |
34 |
-- |
35 |
gentoo-amd64@g.o mailing list |