On Mon, Feb 25, 2019 at 03:53:09PM +0000, Ben Hutchings wrote:
The major input into the new seed file contents is the old seed file contents.
Yes, I'd just drop the seed file once used, then have a scheduled job write a new one at some point in the future if the random quality is high enough. If you reboot twice in a row the second boot won't get seeded, but that's better than a package that introduces potentially insecure random seeding by default. Maybe add a non-default option to allow seed reuse with a lot of warnings, but don't do it by default.