Default configuration of rng-tools in Arch adds no real entropy to /dev/random
bugs.archlinux.org
bugs.archlinux.org
edit: I understand the concern of RNGD_OPTS, but still think that if you're installing rng-tools you should likely know what you're doing with it. As stated by rng-tools:
"... Using the standard open()and read() system calls, you can read random data from the hardware RNG device. This data is NOT CHECKED by any fitness tests, and could potentially be bogus (if the hardware is faulty or has been tampered with). Data is only output if the hardware "has-data" flag is set, but nevertheless a security-conscious person would run fitness tests on the data before assuming it is truly random."
...so, in all reality the flag should be unset forcing the user to select a hardware, or other RNG device. /dev/random is likely a very bad idea as a default.
So is any simplified facility that caches data from places like random.org and provides it as a continuous stream of large pools of entropy? It'd be nice if a data source such as that could be used as a simplistic way to bootstrap the entropy pool on a new system.
With sshd, the user has to enable it (change startup settings or launch it manually) to become "less secure". By contrast, with rng-tools, she must change settings to become _more secure_.
Ever heard of the term "sane defaults"? To me, sane defaults are ones that by default, out of the box, keep the user secure. Decreasing the user's security, maybe to achieve increased functionality or ease of use, should require action on the part of the user. Why? Because history tells us that most users are lazy and will take little or no action. That includes action necessary to be secure.
>> Other than that, I am not strongly against changing the default configuration of rngd but I don't see the point and on the majority of systems there is no hardware entropy generation module, so at least this configuration always works.
>> This package is not a dependency of any other package. It also needs to be explicitly installed AND started/enabled so after installation the system is not any less secure.
The problem here is that to securely use entropy, an encrypted channel is needed, but it's hard to establish one without a source of entropy.
Mobile and personal computing devices need to have hardware entropy. It should be required.
I guess it's a good thing then that Intel added RdRand on IvyBridge. And doesn't TPM also have hardware RNG in it?
Sarcasm detected. Intentional or no?
Does rng-tools effectively do this?
dd if=/dev/urandom of=/dev/urandom bs=512 count=1
Or is someone doing the (more) sensible thing of saving the random state at shutdown and reloading it at startup?