On Sat, 10 Nov 2012, Iain Hibbert wrote: > On Sat, 10 Nov 2012, Robert Elz wrote: > > > Date: Fri, 9 Nov 2012 19:34:05 -0500 > > From: Thor Lancelot Simon <t...@panix.com> > > Message-ID: <20121110003405.ga4...@panix.com> > > > > | No, the cprng named "kernel" is used _inside_ the kernel. > > > > It hadn't occurred to me before that that name is significant. > > me neither > > The first one I get, is the "sysctl" one. It is produced during the > /etc/rc.d/sysdb script, probably by the dev_mkdb program (which uses cdb, > probably using arc4random which uses sysctl to load random data?) > > Since I have not updated my boot.cfg but random_seed=YES in the > /etc/defaults/rc.conf file, I guess the entropy file *is* being loaded, > but after it is required..
Well, strange.. I added "sysdb" to the random_seed BEFORE: line, and indeed get a message that the entropy file was loaded before the sysdb script runs.. but I still get the "sysctl" message during the sysdb script run. > > When I get two messages after a boot (which looks like it might be > > every time), one of them says "kernel" (seems to be the later one) > > and the other (the one soonest after a boot) says "sysctl". > > The second one I get is the "kernel" one, and this is produced during a > fetchmail run after I have logged in. I don't know why that would use > kernel entropy.. perhaps picking a random port number? Also, that is > *after* the entropy file was loaded.. is that not enough for the kernel? apparently not. I did some compilation, to hopefully generate a bit of entropy, then rebooted. I saw the "Loaded entropy file" message during the boot process, then logged in and root# rndctl -s 7073 bits mixed into pool 4096 bits currently stored in pool (max 4096) 2173 bits of entropy discarded due to full pool 804 hard-random bits generated 20156 pseudo-random bits generated plunky% fetchmail console> cprng kernel: WARNING pseudorandom rekeying. root# rndctl -s 7520 bits mixed into pool 4096 bits currently stored in pool (max 4096) 2364 bits of entropy discarded due to full pool 1060 hard-random bits generated 20156 pseudo-random bits generated ..why is my kernel complaining? regards, iain