[onerng talk] /dev/random/urandom question - wireguard related

Paul Campbell paul at taniwha.com
Tue Jul 21 21:34:43 BST 2020

On Wednesday, 22 July 2020 8:26:18 AM NZST alexandre wrote:
> but i faced that urandom  /proc/sys/kernel/random/urandom_min_reseed_secs
> not being honored on arch linux a long time ago and basically stopped using
> the onerng at that time.
> However i just tested on arch linux with kernel , 5.4.51-1-lts and it looks
> like it is working again . Could you confirm if im wrong as i dont know all
> the internals like you  besides the led is blinking part 

Thanks that's an interesting data point - part of what's making this new 
release difficult is collecting an appropriate collection of new kernels, 
installing them into virtual machines and testing them one by one .....

If it is blinking (and it's a result of feeding /dev/urandom and /dev/random) 
then we're probably OK at least in the short term.

One of the things I've tried very hard to do (and so far succeeded at) is to 
create a software release that contains NO compiled binaries, it will happily 
load and run on x86s, ARMs, RISC-Vs etc Sadly one of the changes here may be 
that I need to build a kernel driver to support OneRNG, if that turns out to 
be true it will make software releases a lots less forgiving of different 
kernel and CPU versions


More information about the Discuss mailing list