Code review request: 7197245: Eliminate sun.security.ssl.JsseJce dependency on sun.security.ec

Vincent Ryan vincent.x.ryan at oracle.com
Mon Sep 17 07:01:56 PDT 2012


Thanks Mike.

I've filed a bug on that and will fix it separately from 7197245/7194075:

    7198901:  correct the field size check when decoding a point on ECC curve



On 12 Sep 2012, at 19:19, Michael StJohns wrote:

> Sorry - additional comment.  I think this one needs to be entered as a bug as it needs to be fixed both places.
> 
> In encodePoint at 426,427
> 
> int n = (data.length -1 ) / 2;  // fix component size calculation.
> if ( n != ((curve.getField().getFieldSize() + 7 ) >> 3)) { // has to be equal, not just greater.
> 
> Per x9.63 an encoded point is 1 byte of type followed by ceiling(log base 2 field size/8) bytes of x and same of y.  You can't trim or pad extra left zeros.
> 
> 
> At 12:22 PM 9/11/2012, Vincent Ryan wrote:
> 
>> Hello Brad,
>> 
>> Please review these changes to eliminate the dependency between the
>> SunJSSE provider and the ECParameters and NamedCurve classes in the
>> SunEC provider.
>> 
>> http://cr.openjdk.java.net/~vinnie/7197245/webrev.00/
>> 
>> Thanks.
> 
> 




More information about the security-dev mailing list