07-25-2016 07:04 PM
Env: petalinux-v2016.1-final & Xilinx-ZC706-v2-16.1-final.bsp
When I boot the linux on the ZC706, the last message " random: nonblocking pool is initialized " may take several minutes. That's too slow.
How can I figure out the reason why that's too slow? And How can I speed up that?
BTW: Because my own app wouldn't work until that message show up, I want to fix this issue ASAP.
07-25-2016 10:41 PM
07-26-2016 01:58 AM
In the ug1156-petalinux-tools-workflow-tutorial.pdf, the Figure 2 shows the result I want as the following
But I couldn't find any description about what/how to do to speed up the nonblocking pool intializaiton.
Can you give me more clear clue such as which chapter or page?
10-14-2016 01:51 AM
It seems that I have the same problem in my application based on Avnet MMP Zynq board with Petalinux 2015.4.
Did you solved the problem about speed up the pool initialisation or did you find new information?
10-17-2016 07:54 AM
Are you trying to use getrandom(2)? This blocks until the non-blocking pool is initialized (it needs 128 bits of entropy). If you read /dev/urandom directly it won't block. It depends on what you need the randomness for - crypto keys? If you have a board with not much in the way of peripherals to generate hardware events then entropy will build slowly (like mine). If you have the ability you can save and re-use entropy across reboots (you can google for the procedure). It isn't practical in my case. Basically I just won't bother trying to create cryptographically strong keys on my target. If I wait on /dev/random it may as well be forever.
03-23-2018 05:56 AM - edited 03-23-2018 06:00 AM
Does anyone solved this problem? I currently have the same issue - tcf-agent blocks until "random: nonblocking pool is initialized" message appears, which takes about 2 minutes.
01-30-2020 09:51 PM
if you're using Qt, rebuild your qt library with the configure option