#5 Linux Segmentation fault

closed-fixed
nobody
None
5
2009-10-04
2009-07-11
ezorro
No

Seg fault on LM hash

~/RT$ ./rcracki_mt -v -h 09752A3293831D17 lm_all-space#1-7_1_10000x41480657_distrrtgen[p][i]_20.rti
Debug: Couldn't read rcracki_mt.ini, continuing anyway.
Using 1 threads for pre-calculation and false alarm checking...
Found 1 rainbowtable files...

lm_all-space#1-7_1_10000x41480657_distrrtgen[p][i]_20.rti:
Debug: Saving 79992 bytes of memory for chainwalkset.
Debug: Allocated 37848019 bytes for index with filelen 37848019
Debug: Setting index to 0x00 in memory, 37848019 bytes
reading index... 37848019 bytes read, disk access time: 0.06 s
Debug: Allocated 331845256 bytes for 41480657 chains, filelen 331845256
Debug: Setting pChain to 0x00 in memory
reading table... 331845256 bytes read, disk access time: 0.36 s
verifying the file...
Segmentation fault
:~/RT$

Discussion

  • Daniël Niggebrugge

    hi, tnx for your bug report.

    could you give some more details?

    - version of rcracki_mt
    - linux version
    - 32 or 64bit
    - does it only crash with this table? or with every LM crack, or with all algorithms?

     
  • ezorro

    ezorro - 2009-07-11

    Hi,
    here what I did:
    1: I checked out the svn trunk and compiled it with make.
    2: copied the binary and charset to my rainbowtables dir.
    3: tried and got Segfault.
    4: tried other tables and got segfault too
    5: cross-checked md5sum on tables -> seems OK

    have not tried other algorithms. Will try NTLM later (download is still running)
    I'm using Linux 2.6.24-16-server #1 SMP Thu Apr 10 13:15:38 UTC 2008 x86_64 GNU/Linux
    Ubuntu Hardy Distribution

     
  • Daniël Niggebrugge

    i think it is caused by 64bit code, 64bit is untested. I'll try my best to locate the issue .
    If you could update from svn and try again? i added a very small message that says 'ok' after verifying a file, that way we can see if yours crashes at verification or at actually searching through the file.

    after that i can add some more specific debugging lines to try and find the exact cause of the crash

     
  • Marc-Andre Moreau

    I confirm that this bug happens when the executable has been compiled for 64 bit:

    aghaster@debian:~/rcracki_mt_0.6.2_src$ ./rcracki_mt -h E52CAC67419A9A22 4A3B108F3FA6CB6D -t 4 -o cracked.txt /media/archive/rainbow_tables/LM/
    Using 4 threads for pre-calculation and false alarm checking...
    Found 82 rainbowtable files...

    lm_all-space#1-7_0_10000x37601550_distrrtgen[p][i]_19.rti:
    reading index... 37068152 bytes read, disk access time: 0.02 s
    reading table... 300812400 bytes read, disk access time: 0.32 s
    verifying the file...
    Segmentation fault
    aghaster@debian:~/rcracki_mt_0.6.2_src$ make clean
    rm rcracki_mt
    aghaster@debian:~/rcracki_mt_0.6.2_src$ vim Makefile
    aghaster@debian:~/rcracki_mt_0.6.2_src$ make
    g++ *.cpp -m32 -lssl -lpthread -O3 -o rcracki_mt
    sha1.cpp:272: warning: this decimal constant is unsigned only in ISO C90
    aghaster@debian:~/rcracki_mt_0.6.2_src$ ./rcracki_mt -h E52CAC67419A9A22 4A3B108F3FA6CB6D -t 4 -o cracked.txt /media/archive/rainbow_tables/LM/
    Using 4 threads for pre-calculation and false alarm checking...
    Found 82 rainbowtable files...

    lm_all-space#1-7_0_10000x37601550_distrrtgen[p][i]_19.rti:
    reading index... 37068152 bytes read, disk access time: 0.04 s
    reading table... 268355464 bytes read, disk access time: 0.20 s
    verifying the file...
    searching for 1 hash...
    cryptanalysis time: 35.20 s
    reading table... 32456936 bytes read, disk access time: 0.02 s
    searching for 1 hash...
    cryptanalysis time: 0.04 s

    lm_all-space#1-7_0_10000x67108864_distrrtgen[p][i]_00.rti:
    reading index... 66157982 bytes read, disk access time: 0.10 s
    reading table... 268355464 bytes read, disk access time: 0.36 s
    verifying the file...
    searching for 1 hash...
    cryptanalysis time: 0.46 s
    reading table... 2

    I edited the Makefile to add the -m32 and it did it. My system is:
    Linux debian 2.6.30-1-amd64 #1 SMP Sat Aug 15 18:09:19 UTC 2009 x86_64 GNU/Linux

    64bit support would be nice, though.

     
  • Daniël Niggebrugge

    tnx for your confirmation mamoreau. i hope to get myself some 64-bit testing environment soon, until then 64 bit support just goes as far as 'it might work at your system'

     
  • James Nobis

    James Nobis - 2009-10-02

    neinbrucke,

    I know you can use qemu on a 64bit linux host to install and run emulated 32bit systems. I think this should work in reverse but I haven't tried it. I can probably give you a shell account on a 64bit OpenBSD system. Additionally, I could get you a shell on a 64bit linux install but it would be off my home DSL line.

     
  • James Nobis

    James Nobis - 2009-10-04
    • status: open --> closed-fixed
     
  • James Nobis

    James Nobis - 2009-10-04

    This segfault has been fixed in svn. Feel free to pull the latest code and see if you can confirm this fixes the issue for you as well.

     

Log in to post a comment.

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:





No, thanks