IBM Crypto Education Community

To perform a digital signature verify (CSNDDSV) with Rule 'PKCS-1.1' it is necessary to ASN.1 BER encode the used hash.
In appendix E, chapter "PKCS #1 Formats" of the ICSF Application Programmer's Guide some prefixes to be used are listed.
As far as I understood they refer to RFC 3447: PKCS #1: RSA Cryptography Specifications (http://tools.ietf.org/html/rfc3447#page-43).

It would be nice if anyone could explain the meaning of the different Bytes in the ASN.1 BER Prefix. For example using the hash-prefix provided in this sample.

INITIALIZE CSNDDSV-PARMS

MOVE 1                                         TO CSNDDSV-RULE-ARRAY-COUNT
MOVE 'PKCS-1.1'                                TO CSNDDSV-RULE

MOVE 64                                        TO CSNDDSV-PKA-PK-ID-LENGTH  
MOVE <Name of Public Key>                      TO CSNDDSV-PKA-PK-ID         

MOVE X'3031300D060960864801650304020105000420' TO WORK-HASH-PREFIX-SHA256
MOVE LENGTH OF WORK-HASH-PREFIX-SHA256         TO CSNDDSV-HASH-LENGTH
ADD  LENGTH OF <SHA-256 Hash>                  TO CSNDDSV-HASH-LENGTH
STRING WORK-HASH-PREFIX-SHA256                         
       <SHA-256 Hash>  DELIMITED BY SIZE     INTO CSNDDSV-HASH                                                                

MOVE LENGTH OF <Binary Signature>              TO CSNDDSV-SIGNATURE-LAENGE
MOVE <Binary Signature>                        TO CSNDDSV-SIGN-FIELD
 

CALL CSNDDSV USING
     CSNDDSV-RC,              
     CSNDDSV-RSN,             
     CSNDDSV-EXIT-DATA-LENGTH,
     CSNDDSV-EXIT-DATA,       
     CSNDDSV-RULE-ARRAY-COUNT,
     CSNDDSV-RULE-ARRAY,      
     CSNDDSV-PKA-PK-ID-LENGTH,
     CSNDDSV-PKA-PK-ID,       
     CSNDDSV-HASH-LENGTH,     
     CSNDDSV-HASH,            
     CSNDDSV-SIGN-FIELD-LENGTH,
     CSNDDSV-SIGN-FIELD

IF   CSNDDSV-RC NOT EQUALS ZERO ...-RC NOT EQUALS ZERO ...

Hi Olaf,

I'll try to explain based on your example... X'3031300D060960864801650304020105000420'

(Just FYI... this ASN.1 link is one of my favorite references... http://luca.ntop.org/Teaching/Appunti/asn1.html)

First, let's break up the DER encoding...

30 31
   30 0D
      06 09 60 86 48 01 65 03 04 02 01
      05 00
   04 20

A DER encoding generally has the format IDENTIFIER || LENGTH || CONTENTS

1. The first octet in the sample is the identifier, '30'x, which is '00110000'b. The last 5 bits represent the tag number which is '10'x. When you look up tag '10'x (see the link above) it is a SEQUENCE. Since it's a sequence, the content will be nested.
2. The second octet is the length which is '31'x which is 49 (decimal). The value is slightly misleading since the count of the number of octets in the sample is 19! I'll explain this a little later. ;-)
3. The third octet starts the content which turns out to be another sequence ('30'x).
   1. The identifier octet in the inner sequence is '06'x which is '00000110'b. So the tag number is '06'x which is an OBJECT IDENTIFIER (or OID).
   2. The length octet in the inner sequence is '09'x which is the length of the object identifier.
   3. The contents octet in the inner sequence is '608648016503040201' which represents id-sha256 (2.16.840.1.101.3.4.2.1) in BER encoding. [The ASN.1 link above explains how the OID is converted to DER encoding. This NIST page gives the formal OID name for crypto algorithms.]
   4. The next identifier octet in the inner sequence is '05'x which is '00000101'b which is NULL. It's following by a length octet of zero ('00'x). Now the inner sequence is complete. Notice that there were 13 octets matching the '0D'x length for the sequence.
4. The next identifier octet is '04'x which is an OCTET STRING. An OCTET STRING is just an arbitrary string of octets. The length identifier is '20'x which is 32 decimal.
5. But notice that the content octets are missing. That's probably why this example is considered a prefix. This prefix DER encoding would be prepended to the actual 32-byte SHA-256 hash. When that hash is added to the DER encoding, you'll have the full length ('31'x or 49 dec) defined in the first length octet.

Hopefully this helps but let me know if anything is unclear/confusing or if I misinterpreted the question.

P.S. It had been awhile since I looked at DER encodings! Good question. :-)

Thank you very much Eysha for the very clear explanation!

To memorize better I put your information into a table. Perhaps it might be interesting for the community:

Good idea! Great format!