sfx2000
Part of the Furniture
Noticed this the other day where the entropy pool was static at 256 (which is traditionally considered a low value for the pool)
Workarounds by the dev community here (@RMerlin @Voxel @john9527 ) included discussion around using haveged to keep the pool filled, as /dev/random is blocking if the pool empties... /dev/urandom is non-blocking, but some concerns about randomness of that pool).
Recently, the guy behind Wireguard, Jason A. Donenfeld (zx2c4), has made some wholescale changes to the rng included in kernel's 5.17 and later...
write up here... it looks like openwrt has backported the changes into the testing kernel over on master already...
Workarounds by the dev community here (@RMerlin @Voxel @john9527 ) included discussion around using haveged to keep the pool filled, as /dev/random is blocking if the pool empties... /dev/urandom is non-blocking, but some concerns about randomness of that pool).
Recently, the guy behind Wireguard, Jason A. Donenfeld (zx2c4), has made some wholescale changes to the rng included in kernel's 5.17 and later...
write up here... it looks like openwrt has backported the changes into the testing kernel over on master already...