Thanks Casey, but I am still getting same error !!
I have just enabled the debug on GNU classpath jar, so that you can
see the line number in stack trace.
Here is the new stack trace
gnu.javax.net.ssl.provider.AlertException: HANDSHAKE_FAILURE:
locally generated; FATAL
at
gnu.javax.net.ssl.provider.ServerHandshake.checkKeyExchange
(ServerHandshake.java:1184)
at
gnu.javax.net.ssl.provider.AbstractHandshake.getInputParams
(AbstractHandshake.java:330)
at gnu.javax.net.ssl.provider.SSLEngineImpl.unwrap
(SSLEngineImpl.java:560)
at javax.net.ssl.SSLEngine.unwrap(SSLEngine.java:566)
at gnu.javax.net.ssl.provider.SSLSocketImpl.doHandshake
(SSLSocketImpl.java:473)
at gnu.javax.net.ssl.provider.SSLSocketImpl
$SocketInputStream.read(SSLSocketImpl.java:150)
at sun.nio.cs.StreamDecoder$CharsetSD.readBytes
(StreamDecoder.java:411)
at sun.nio.cs.StreamDecoder$CharsetSD.implRead
(StreamDecoder.java:453)
at sun.nio.cs.StreamDecoder.read(StreamDecoder.java:183)
at java.io.InputStreamReader.read(InputStreamReader.java:167)
at java.io.BufferedReader.fill(BufferedReader.java:136)
at java.io.BufferedReader.read(BufferedReader.java:157)
at TLSDataListener.processData(TLSDataListener.java :26)
at TLSServer$2.run(TLSServer.java:197)
Caused by: javax.crypto.IllegalBlockSizeException : Data must not
be longer than 128 bytes
at com.sun.crypto.provider.RSACipher.a(DashoA12275)
at com.sun.crypto.provider.RSACipher.engineDoFinal
(DashoA12275)
at javax.crypto.Cipher.doFinal(DashoA12275)
at gnu.javax.net.ssl.provider.ServerHandshake
$RSAKeyExchange.implRun(ServerHandshake.java:1347)
at gnu.javax.net.ssl.provider.DelegatedTask.run
(DelegatedTask.java:68)
at gnu.javax.net.ssl.provider.SSLSocketImpl.doHandshake
(SSLSocketImpl.java :498)
... 9 more
thanks
chinmaya
On 3/23/07, Casey Marshall <address@hidden> wrote: On Mar 23,
2007, at 5:10 PM, Casey Marshall wrote:
> It looks like Sun's RSA requires that the input start with a byte
> 0, which the SSL spec doesn't require (the version of RSA in
> Classpath doesn't require this, either). I think the patch I've
> attached may help; I haven't tried this yet, however.
>
Actually, this likely won't fix this issue. It is more likely that
decrypting the session key is failing. My guess is this part of the
code (ServerHandshake.java:1334):
rsa.init(Cipher.DECRYPT_MODE, serverKey);
rsa.init(Cipher.DECRYPT_MODE, localCert);
is the problem. In Classpath, initializing an RSA cipher with two
keys (public and private) causes RSA to use "blinding," which helps
prevent timing attacks. You should try swapping the order of these
two statements, so they read:
rsa.init(Cipher.DECRYPT_MODE , localCert);
rsa.init(Cipher.DECRYPT_MODE, serverKey);