ssh: ssh keep alive

[alexandrejbr]

This PR aims at introducing the equivalent of openssh ClientAliveInterval/ClientAliveCountMax (https://man.openbsd.org/sshd_config#ClientAliveCountMax) and ServerAliveInterval/ServerAliveCountMax

There's 2 notable differences:

1. In openssh "The default is 0, indicating that these messages will not be sent to the client.", but in this implementation infinity is used instead of 0 and ssh_options checks that a positive integer is presented;
2. Keep-alive messages can't be sent during renegotiation, but since this feature acts as a keep-alive and a timeout, an equivalent timeout is established for the renegotiation procedure if alive is enable. This is implemented with a timeout called renegotiation_alive

[github-actions]

CT Test Results

Tests are running... https://github.com/erlang/otp/actions/runs/12057006879

Results for commit 82454d3

To speed up review, make sure that you have read Contributing to Erlang/OTP and that all checks pass.

See the TESTING and DEVELOPMENT HowTo guides for details about how to run test locally.

Artifacts

• No CT logs found
• No HTML docs found
• No Windows Installer found

// Erlang/OTP Github Action Bot

[alexandrejbr]

@u3s this a feature I promised long time ago, if you are still interested in it please continue reading :)

I know you are short on time, but if you could give me some tips on how to create some tests for this would be appreciated.

I wanted to add tests at least these 3 scenarios:

• Normal scenario client and server send alive messages (This one I have one idea how to test which would be using the ssh_dbg module, ssh_dbg:on([ssh_messages]) reveals the sending of the messages but it's kind of ugly to use a regex to get the keep alive message);
• Timeout scenario. I would like that the server or client become not responsive and I would like to check that the connections is terminated (this one we have tested before suspending the openssh process but it's not so nice);
• The renegotiation scenario in which the during the renegotiation the peer becomes unreachable and the timeout should have effect.