Re: pkcs11 problem with Netscape

[Antonio Andres Espallardo]

Ok, now I have a buffer which is equal to the input data size. As I said 
in the post, when Netscape 7.2 (and 4.7) initializes a encrypt operation 
(RC2 Secret Key of 16 bytes). It's ok. After that it calls the 
C_EncryptUpdate method of the pkcs11 module with an input data of 8 
bytes. The method C_EncryptUpdate (which use openSSL functions to do the 
work: int result = EVP_CipherUpdate(ctx,pEncryptedPart,(int 
*)pulEncryptedPartLen,pPart,ulPartLen);) returns a encrypted block with 
a size of 8 bytes (which I think is the correct size). Well, that size 
not seems to be the hoped by Netscape, because it close the session (if 
all were ok it should call C_EncryptUpdate more times and ends with a 
call to C_EncryptFinal) wich opened to develope the C_Encrypt operation.
I don't understand what is happening. Thanks in advance.

Robert Relyea wrote:
Antonio Andres Espallardo wrote:
Hi. I'm developing a pkcs11 module, which offeir support for RC2 
algorithm and some algorithm more. The behavior of Netscape 7.2 when I 
make use of the pkcs11 module is like this:

1) Netscape creates an RC2 Secret Key of 16 bytes length
2) It inits an encrypt operation, calling the method C_EncryptInit
3) It calls the method C_EncryptUpdate of the pkcs11, but the out 
buffer is too small, consecuently the pkcs11 return 
CKR_BUFFER_TOO_SMALL. Well, now Netscape detroy the secret key object 
and finalizes, when it should retry with a bigger buffer.

How big of a buffer is NSS providing? The spec says for CKM_RC2_CBC you 
need a buffer which is equal to the input data size. (and input data 
should be a multiple of 8). What are you seeing from NSS?

bob
_______________________________________________
mozilla-crypto mailing list
mozilla-crypto@mozilla.org
http://mail.mozilla.org/listinfo/mozilla-crypto