full-disclosure - ltrace bug

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <87ad8cog81.fsf@volatile.central.fluidsignal.com>
From: aroldan at debian.org (Andrés Roldán)
Subject: ltrace bug

This seems to be fixed in the latest version:

aroldan@...atile:~$ ltrace -V
ltrace version 0.3.30.
Copyright (C) 1997-2002 Juan Cespedes <cespedes@...ian.org>.
This is free software; see the GNU General Public Licence
version 2 or later for copying conditions.  There is NO warranty.

"Abhisek Datta" <abhisek@...uxmail.org> writes:

> 			*********************************
> 			   BFI Security Research Group
> 			*********************************
> 			     	    
> Vulnerability:
> ==============
> A heap based buffer overrun bug is identified in ltrace 'Library Call Tracer' utility version 0.3.10-12 which allows execution of arbitrarty code with root privilage by corrupting the heap.
>
> Affected Versions:
> =================
> So far now only tested version 0.3.10-12 on RedHat Linux 8 (2.4.18-14) seems to be vulnerable to this approach.
>
> Description:
> ============
> There seems to exists a heap based buffer overflow vulenrability in ltrace utility version 0.3.10-12. Based on my research so far, I think the bug is in static char *search_for_command(char * filename) function in options.c .
>
> [root@...alhost codes]# rpm -qa |grep ltrace
> ltrace-0.3.10-12
> [root@...alhost codes]# ltrace -V
> ltrace version 0.3.11.
> Copyright (C) 1997-1999 Juan Cespedes <cespedes@...ian.org>.
> This is free software; see the GNU General Public Licence
> version 2 or later for copying conditions.  There is NO warranty.
> [root@...alhost codes]# cp `which ltrace` ltrace
> [root@...alhost codes]# ltrace ./ltrace `perl -e 'print "A"x6000'`
> __libc_start_main(0x08049a40, 2, 0xbfff7f84, 0x08048f30, 0x08057644 <unfinished ...>
> getpid()                                          = 2487
> __cxa_atexit(0x08049cf0, 0, 0, 0x0805b860, 0x420aedf0) = 0
> signal(2, 0x08049c10)                             = NULL
> signal(15, 0x08049c10)                            = NULL
> getopt_long(2, 0xbfff7f84, "+dfiLSrthVCa:s:o:u:p:e:", 0x0805b060, 0xbfff6f08) = -1
> strchr("AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA"..., '/') = NULL
> getenv("PATH")                                    = "/usr/local/sbin:/usr/local/bin:/"...
> strchr("/usr/local/sbin:/usr/local/bin:/"..., ':') = ":/usr/local/bin:/sbin:/bin:/usr/"...
> strncpy(0x0805b880, "/usr/local/sbin", 15)        = 0x0805b880
> strcpy(0x0805b890, "AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA"... <unfinished ...>
> --- SIGSEGV (Segmentation fault) ---
> +++ killed by SIGSEGV +++
>
> ltrace 0.3.11 src
> ==================
> options.c :
> static char *search_for_command(char * filename)
> {
>         static char pathname[1024];
>         char *path;
>         int m, n;
>         if (strchr(filename, '/')) {
>                 return filename;
>         for (path = getenv("PATH"); path && *path; path += m) {
>                 if (strchr(path, ':')) {
>                         n = strchr(path, ':') - path;
>                         m = n + 1;
>                 } else {
>                         m = n = strlen(path);
>                 strncpy(pathname, path, n);     
>                 if (n && pathname[n - 1] != '/') {
>                         pathname[n++] = '/';
>                 strcpy(pathname + n, filename); /* Possible buffer overrun */
>                 if (!access(pathname, X_OK)) {
>                         return pathname;
>         return filename;
> }
>
> Seems like there is no bound checking for the length of filename which is passed as parameter to the function.Version 0.3.20 implements bound checking before  for(path = getenv("PATH"); path && *path; path += m) and hence cannot be exploited this way. The exploitation of this bug doesnt appear to be a simple one and conventional heap based overun exploitation techniques probably wont work. 
>
> by
> Abhisek Datta
> abhisek@...secure.com
>
> BFI Security Research Group
> India
> -- 
> ______________________________________________
> http://www.linuxmail.org/
> Now with e-mail forwarding for only US$5.95/yr
>
> Powered by Outblaze
>
> _______________________________________________
> Full-Disclosure - We believe in it.
> Charter: http://lists.netsys.com/full-disclosure-charter.html

-- 
Andres Roldan <aroldan@...ian.org>
GPG Key-ID: 0xB29396EB
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 188 bytes
Desc: not available
Url : http://lists.grok.org.uk/pipermail/full-disclosure/attachments/20031007/38185e29/attachment.bin