A puzzle: CAPZLOQ TEKNIQ 1.0
The cross-platform viral proof-of-concept in the news last week does indeed infect both Windows and Linux ELF binaries. At least it does on some kernels. Some tests show it doesn't work on the latest versions.
Hans-Werner Hilse is trying to puzzle out why. If anyone else wants to play with it and see if they can figure out why it is sometimes viral on Linux and sometimes not, drop me a note offlist.
Quoted 9 lines
The cross-platform viral proof-of-concept in the news last week does indeed infect both Windows and Linux ELF binaries. At least it does on some kernels. Some tests show it doesn't work on the latest versions.
Hans-Werner Hilse is trying to puzzle out why. If anyone else wants to play with it and see if they can figure out why it is sometimes viral on Linux and sometimes not, drop me a note offlist.
Ah, now I see the post on lwn.net. Well, already found something?
Jan Engelhardt
Quoted 10 lines
Quoted 7 lines
The cross-platform viral proof-of-concept in the news last week does indeed infect both Windows and Linux ELF binaries. At least it does on some kernels. Some tests show it doesn't work on the latest versions.
Hans-Werner Hilse is trying to puzzle out why. If anyone else wants to play with it and see if they can figure out why it is sometimes viral on Linux and sometimes not, drop me a note offlist.
From LWN/Newsforge:
--->2.6.15.4
[0804744d] open("E", O_RDWR) = 4
...
[0804747e] old_mmap(NULL, 28672, PROT_READ|PROT_WRITE, MAP_SHARED, 4, 0) = 0xb7fca000
--->2.6.16.2:
[0804744d] open("E", O_RDWR) = 4
...
[0804747e] old_mmap(NULL, 32768, PROT_READ|PROT_WRITE, MAP_SHARED, 1, 0) = -1 ENODEV (No such device)
Simple as that. open() returns fd 4, but old_mmap is called with fd 1, which is usually stdout. Looks to me like a userspace problem.
Jan Engelhardt
Signature · 3 lines
| Software Engineer and Linux/Unix Network Administrator | Alphagate Systems, http://alphagate.hopto.org/ | jengelh's site, http://jengelh.hopto.org/