Randomness with OneRNG and NeuG [LWN.net]

Randomness with OneRNG and NeuG

Posted Jul 1, 2015 18:42 UTC (Wed) by nix (subscriber, #2304)
Parent article: Randomness with OneRNG and NeuG

Given the existence of the AT_RANDOM entry in the ELF auxiliary vector (and its use in setting up stack guard canaries), there *is* actually a workload that will rapidly drain the entropy pool: exec() a lot. Just running a configure script will suffice...

to post comments

Randomness with OneRNG and NeuG

Posted Jul 2, 2015 15:46 UTC (Thu) by flussence (guest, #85566) [Link] (1 responses)

Maybe that'd be why my 64-bit compile runs feel so much slower than 32-bit... </crackpottheory>

Randomness with OneRNG and NeuG

Posted Jul 7, 2015 9:54 UTC (Tue) by nix (subscriber, #2304) [Link]

Nah. That's reduced cache efficiency from the increased alignment. 32-bit programs get AT_RANDOM too.

Randomness with OneRNG and NeuG

Posted Jul 2, 2015 23:07 UTC (Thu) by thestinger (guest, #91827) [Link] (1 responses)

There are also various random bases for ASLR generated on each exec(...).

Randomness with OneRNG and NeuG

Posted Jul 3, 2015 0:47 UTC (Fri) by spender (guest, #23067) [Link]

But those don't use get_random_bytes() (unlike the filling of AT_RANDOM) so they won't deplete entropy -- it's purely from AT_RANDOM.

-Brad