SSL_read buffering problem?

Hi, Does the SSL layer have its own read buffering? Our application has a standard non-blocking socket loop where we wait for a read event, read 8k, and go back to wait for more. If we have more than 8k then we have multiple events until the data is complete. On extending the application to

Re: SSL_read buffering problem?

Thanks Lutz, And further reading on the archive and google has also enlightened me on this. I'm actually almost surprised that this has worked at all, but I guess it depends on the relative sizes of the SSL block and my 8k block. I don't want to sound too ignorant of SSL (although I am), but

Re: -verify option for s_client

The depth should be greater than or equal to the depth of the CA chain. i.e. if you have: Certificate Intermediate CA 1 Intermediate CA 2 Root CA ... you need a depth of 4. More specifically if the Root CA is not found within your verify depth, then the verification will

RE: On-the-fly self generated certs for network application

Just use Anonymous Diffie-Hellman if you don't need certificates. All you need to do is change the cipher set. G. "Marco Cunha" [EMAIL PROTECTED] on 11/01/2001 11:51:54 Please respond to [EMAIL PROTECTED] To: [EMAIL PROTECTED] cc:(bcc: George Shaw/EMEA/Viewlocity) Subject: RE:

Re: ADH + certs on same SSLCTX ??

Hi, If you are allowing anon-DH, then you are permitting certificate-less SSL, which is a potential security hole, unless your implementation is not worried about that. Our implementation needed to enable this, and our solution was to make this configurable to the end user, then it is their

Different meanings/effects of verify depth?

Hi, I've been investigating the way that verify depth works in OpenSSL, and I've managed to confuse myself over what it actually means. It seems to be used in two places, with different meanings and/or effects. 1. In the verify_callback example in s_cb.c. Here the verify depth is set as a