OpenSSL 1.0.2 no-asm RC4_40 OOB read

[guidovranken]

Compile OpenSSL 1.0.2 without assembly

CC=clang ./config no-asm -fsanitize=address && make -j12

Then compile and run this file

#define CF_CHECK_EQ(expr, res) if ( (expr) != (res) ) { goto end; }
#define CF_CHECK_NE(expr, res) if ( (expr) == (res) ) { goto end; }

#include <openssl/cmac.h>

int main(void)
{
    const unsigned char input[6] = { 0 };
    const unsigned char key[1] = { 0 };
    CMAC_CTX* ctx = CMAC_CTX_new();
    const EVP_CIPHER* cipher = NULL;

    /* Initialize */
    {
        CF_CHECK_NE(cipher = EVP_rc4_40(), NULL);
        CF_CHECK_EQ(CMAC_Init(ctx, key, sizeof(key), cipher, NULL), 1);
    }

    CF_CHECK_EQ(CMAC_Update(ctx, input, sizeof(input)), 1);

end:
    return 0;
}

$ clang -fsanitize=address -I openssl-1.0.2r/include/ cmac-poc.c openssl-1.0.2r/libcrypto.a 
$ ./a.out 
=================================================================
==30930==ERROR: AddressSanitizer: global-buffer-overflow on address 0x0000006f24e0 at pc 0x000000525504 bp 0x7ffde649f590 sp 0x7ffde649f588
READ of size 8 at 0x0000006f24e0 thread T0
    #0 0x525503 in RC4 (/home/jhg/ossl-102-cmac-poc/a.out+0x525503)
    #1 0x5128f5 in rc4_cipher (/home/jhg/ossl-102-cmac-poc/a.out+0x5128f5)
    #2 0x51b28d in CMAC_Update (/home/jhg/ossl-102-cmac-poc/a.out+0x51b28d)
    #3 0x5127f2 in main (/home/jhg/ossl-102-cmac-poc/a.out+0x5127f2)
    #4 0x7efc41786b96 in __libc_start_main /build/glibc-OTsEL5/glibc-2.27/csu/../csu/libc-start.c:310
    #5 0x41a3a9 in _start (/home/jhg/ossl-102-cmac-poc/a.out+0x41a3a9)

0x0000006f24e6 is located 0 bytes to the right of global variable 'input' defined in 'cmac-poc.c:8:25' (0x6f24e0) of size 6
SUMMARY: AddressSanitizer: global-buffer-overflow (/home/jhg/ossl-102-cmac-poc/a.out+0x525503) in RC4
Shadow bytes around the buggy address:
  0x0000800d6440: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
  0x0000800d6450: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
  0x0000800d6460: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
  0x0000800d6470: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
  0x0000800d6480: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
=>0x0000800d6490: 00 00 00 00 00 00 00 00 00 00 00 00[06]f9 f9 f9
  0x0000800d64a0: f9 f9 f9 f9 01 f9 f9 f9 f9 f9 f9 f9 00 00 00 00
  0x0000800d64b0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
  0x0000800d64c0: 00 00 00 f9 f9 f9 f9 f9 00 00 00 00 00 00 00 00
  0x0000800d64d0: 00 00 00 f9 f9 f9 f9 f9 00 00 00 00 00 00 00 00
  0x0000800d64e0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
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
  Freed heap region:       fd
  Stack left redzone:      f1
  Stack mid redzone:       f2
  Stack right redzone:     f3
  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
==30930==ABORTING

[davidben]

Does CMAC with RC4 even make sense? It needs a block cipher. Moreover, the subkeys are computed with operations over GF(2^(block_size)), which means the scheme is parametrized by a suitable irreducible polynomial. SP 800-38B, section 5.3, defines polynomials for 64-bit (DES or 3DES) and 128-bit (AES) block ciphers.
https://nvlpubs.nist.gov/nistpubs/SpecialPublications/NIST.SP.800-38b.pdf

That corresponds to the bl == 16 check here, but that uses DES's polynomial for all other block size. OpenSSL should probably be making CMAC_Init fail here.
https://github.com/openssl/openssl/blob/master/crypto/cmac/cmac.c#L43

(To that end, there's no need to use EVP_MAX_BLOCK_LENGTH because the largest block size for which CMAC is currently defined is 16 bytes.)

[guidovranken]

@mattcaswell I suppose this won't be fixed because it's 1.0.2? Would you suggest I prevent performing CMAC RC4 in my fuzzer?

[mattcaswell]

Yes, assuming this only impacts 1.0.2 i don't think this will get fixed, so I suggest you prevent CMAC with RC4 for 1.0.2.

[levitte]

Please check if this still applicable on newer versions, such as 1.1.1 or master