Control: tag -1 wontfix On Fri, Apr 25, 2014 at 08:56:27AM +0200, Michael Prokop wrote: > Colin, what's your take on this? I'd be willing to work on this if > there's any chance to get it merged.
I'm generally extremely reluctant to do this kind of thing because, while it may be useful in some cases, there's a significant risk that it will cause keys to be generated in environments where not much entropy is available (see e.g. https://factorable.net/weakkeys12.extended.pdf). The current arrangements make it much more probable that keys will be generated after the system has been up for a while and has accumulated a decent amount of entropy, and I'm much happier with that. Of course any other random package installed in a cloud guest or a live CD, say, can arrange to start something before sshd that runs "ssh-keygen -A". But I would be much happier with that kind of thing living *outside* the OpenSSH packaging, because firstly I don't have to be responsible for the consequences :-), and secondly it stands at least some chance of being able to be more aware of the context in which it will be run and take steps to mitigate it, such as by making context-specific arrangements for the entropy pool to be well-populated before generating keys: the kind of thing that things outside the OpenSSH packaging are in a much better position to decide. In the case of a cloud guest, for example, it may be more appropriate to do the key generation from the host, or at least make sure that entropy is fed in from the host *before* running ssh-keygen. So, for me, I'm afraid this bug is wontfix in OpenSSH itself because I would much rather be safe by default, but you're welcome to arrange for it to be done somewhere else where you can understand the problems and arrange to avoid them as much as possible. Cheers, -- Colin Watson [cjwat...@debian.org] -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
