This is a discussion on Re: DTLS and memory BIO - Openssl ; On Thursday 10 January 2008 9:22 pm, Justin Karneges wrote: [...snip...] > ask here about the state of DTLS and memory BIOs. > > If it is not possible, does anyone have a tip on how to make it possible? ...
On Thursday 10 January 2008 9:22 pm, Justin Karneges wrote:
> ask here about the state of DTLS and memory BIOs.
> If it is not possible, does anyone have a tip on how to make it possible?
> Maybe I could create a substitute for the dgram BIO that works on memory,
> but unlike the existing openssl memory BIO, mine could support all the
> DGRAM control commands?
=46rom reading the library and s_client code, it looks like the best way to=
OpenSSL DTLS is with BIO_dgram, a blocking socket, and with socket timeouts=
enabled. Is this true?
Does the OpenSSL DTLS implementation support packet retransmissions during =
handshake period? This is the only reason I suspect socket timeouts need t=
be enabled for proper functionality. Otherwise, you could get stuck in=20
BIO_dgram's recvfrom() indefinitely and OpenSSL wouldn't get a chance to=20
retransmit a lost UDP packet. Can anyone comment on this? Can I disable t=
timeouts once the handshake is complete?
I've made my own BIO_dgram substitute that replicates this blocking+timeout=
behavior but uses memory instead of a socket. I just want to be sure I'm o=
the right track.
OpenSSL Project http://www.openssl.org
User Support Mailing List firstname.lastname@example.org
Automated List Manager email@example.com