Hi,

An interesting video for this:

https://www.youtube.com/watch?v=gp_90-3R0pE

I think OpenBSD gets entropy from many sources - thisn't very much documented as far as I could see - such as time between interrupts, application memory allocation. In general a multi-threaded environment run by a clock running interrupts provides a good entropy that way. There surely is much more that feeds into the entropy.

You can have a look at getentropy() and arc4random().

Further it appears a simple usb device can't have that much entropy. Further arc4random polls out chunks of random to the various applications so as long as you're not running a constrained environment there's probably no issue (as long as you have several application threads running simultaneously).

Notice the random seed is regenerated at boot and shutdown so as long you've run once a certain amount of time and rebooted the computer once, it's got to a state where random generator shouldn't predictable since it bootstraps the entropy for next startup at shut-down every time.

Side note, for a server the source of entropy is plenty, network data are completely asynchronous so mixing them up with internal states probably make a very job.

I'd like to be able to help reading from your device, I haven't done that before sorry.

Jean-François


Le 10/07/2020 à 00:21, ken.hendrick...@l3harris.com a écrit :
I wrote:
How do I use a hardware random number generator to
continuously seed /dev/random with new truly random numbers?
--- Theo de Raadt wrote:
We consider these devices boring, because the kernel does a good enough job 
creating random.
randomness only has a bootstrap problem.  And these devices don't solve the 
bootstrap problem.
I'm thinking of headless servers, where randomness can ONLY come
from the network.  There is no keyboard, no mouse, etc.

I'm also thinking of first boot after install, when keys are generated.
I think that is what you mean by the bootstrap problem.

Thanks,
Ken

PS  I'm also thinking of very old hardware, without RDRAND in the CPU.
     Not to mention that you can't necessarily trust RDRAND!



CONFIDENTIALITY NOTICE: This email and any attachments are for the sole use of 
the intended recipient and may contain material that is proprietary, 
confidential, privileged or otherwise legally protected or restricted under 
applicable government laws. Any review, disclosure, distributing or other use 
without expressed permission of the sender is strictly prohibited. If you are 
not the intended recipient, please contact the sender and delete all copies 
without reading, printing, or saving.



Reply via email to