Hey
May some explains the following result about AES OFB (sounds like a BUG)?
I think that AES-128 block size is at 128 bits even for ECB or OFB mode but
openssl output is different.
INFO: Settings:
alg: aes-128-cbc
Algo Name: aes-128-cbc
EVP_CIPHER: NID: 01a3, Name: AES-128-CBC, Blk len: 16, key len: 16, iv len: 16,
flags: 2, mode: 2
INFO: Settings:
alg: aes-128-ofb
Algo Name: aes-128-ofb
EVP_CIPHER: NID: 01a4, Name: AES-128-OFB, Blk len: 1, key len: 16, iv len: 16,
flags: 4, mode: 4
INFO: Settings:
alg: aes-128-ecb
Algo Name: aes-128-ecb
EVP_CIPHER: NID: 01a2, Name: AES-128-ECB, Blk len: 16, key len: 16, iv len: 16,
flags: 1, mode: 1
here is the piece of code:
const EVP_CIPHER *evp_ciph = NULL;
evp_ciph = EVP_get_cipherbyname (alg_name);
fprintf (stdout, "Algo Name: %s\n", alg_name);
if (evp_ciph != NULL) {
fprintf (stdout, "EVP_CIPHER: NID: %04x, Name: %s, " \
"Blk len: %d, key len: %d, iv len: %d, " \
"flags: %lx, mode: %lx\n",
EVP_CIPHER_nid (evp_ciph), EVP_CIPHER_name (evp_ciph),
EVP_CIPHER_block_size (evp_ciph),
EVP_CIPHER_key_length (evp_ciph),
EVP_CIPHER_iv_length (evp_ciph),
EVP_CIPHER_flags (evp_ciph), EVP_CIPHER_mode
(evp_ciph));
thanks in advance
Fred
-----Original Message-----
From: Nils Larsch via RT [mailto:[EMAIL PROTECTED]
Sent: Thu 4/20/2006 3:44 PM
To: [EMAIL PROTECTED]
Cc: [email protected]
Subject: Re: [openssl.org #1318] [PATCH] AES-CFB1 and DES-CFB1 mode only
encrypts 1/8th of the source
Michael McDougall wrote:
> diff -ur openssl-SNAP-20060415/crypto/aes/aes_cfb.c
> openssl-SNAP-20060415.changed/crypto/aes/aes_cfb.c
> --- openssl-SNAP-20060415/crypto/aes/aes_cfb.c 2004-12-30
> 06:00:14.000000000 -0500
> +++ openssl-SNAP-20060415.changed/crypto/aes/aes_cfb.c 2006-04-17
> 22:18:11.000000000 -0400
> @@ -191,18 +191,20 @@
> }
>
> /* N.B. This expects the input to be packed, MS bit first */
> -void AES_cfb1_encrypt(const unsigned char *in, unsigned char *out,
> +void
> +AES_cfb1_encrypt(const unsigned char *in, unsigned char *out,
> const unsigned long length, const AES_KEY *key,
> unsigned char *ivec, int *num, const int enc)
> {
> unsigned int n;
> unsigned char c[1],d[1];
> + unsigned long length_in_bits = length * 8;
>
> assert(in && out && key && ivec && num);
> assert(*num == 0);
>
> memset(out,0,(length+7)/8);
> - for(n=0 ; n < length ; ++n)
> + for(n=0 ; n < length_in_bits ; ++n)
> {
> c[0]=(in[n/8]&(1 << (7-n%8))) ? 0x80 : 0;
> AES_cfbr_encrypt_block(c,d,1,key,ivec,enc);
the problem here is that that EVP layer specifies the length of the
input in bytes whereas this functions expects the length in bits
(although is not clearly specified somewhere).
> diff -ur openssl-SNAP-20060415/crypto/evp/e_des.c
> openssl-SNAP-20060415.changed/crypto/evp/e_des.c
> --- openssl-SNAP-20060415/crypto/evp/e_des.c 2004-03-28 13:00:11.000000000
> -0500
> +++ openssl-SNAP-20060415.changed/crypto/evp/e_des.c 2006-04-17
> 22:17:39.000000000 -0400
> @@ -109,8 +109,9 @@
> {
> unsigned int n;
> unsigned char c[1],d[1];
> + unsigned int inl_bits = inl * 8;
>
> - for(n=0 ; n < inl ; ++n)
> + for(n=0 ; n < inl_bits ; ++n)
> {
> c[0]=(in[n/8]&(1 << (7-n%8))) ? 0x80 : 0;
> DES_cfb_encrypt(c,d,1,1,ctx->cipher_data,(DES_cblock *)ctx->iv,
agree, with this one.
Cheers,
Nils
______________________________________________________________________
OpenSSL Project http://www.openssl.org
Development Mailing List [email protected]
Automated List Manager [EMAIL PROTECTED]
______________________________________________________________________
OpenSSL Project http://www.openssl.org
Development Mailing List [email protected]
Automated List Manager [EMAIL PROTECTED]
