I re-ran the NIST AES Validation Tests with 0.9.7 Beta3 Snapshot 20021027 which purports to have fixed the padding problem and so should handle 8-bit (1 octet) data blocks with AES-128 (and AES-192 and AES-256) in CFB Mode. The basic tests which all use 16 byte data blocks still work fine, but the Monte Carlo tests which use 8-bit data blocks still fail, i.e. the OpenSSL result does not match the predicted result, at even the first iteration. Should I manually turn off padding? I am using the sequence: EVP_CipherInit() + EVP_Cipher() for the first iteration, then just EVP_Cipher() for the following ones of the 999 iterations. Would it make any difference if I went down to EVP_CipherUpdate() instead of EVP_Cipher() with an EVP_CipherFinal() after the last iteration? I haven't seen any problems with stream encryption/decryption using AES CFB mode but then I am talking between my own machines, all with the same code. Chris Brook
______________________________________________________________________ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]