9

There is an option to enable or disable stack randomization but is there way to supply the ASLR algorithm with a seed to make it deterministic?

EDIT: The context is as follows: a memory bug I have pops up one out of ten executions under ASLR and cannot be reproduced with ASLR disabled. Providing a random seed to the ASLR algorithm would significantly help the debugging process.

Alex Zhi
  • 183
  • 5
  • You do realize that the R in ASLR stands for randomization, right? Making it deterministic seems a bit at odds with the goal of ASLR... – twalberg Jan 03 '17 at 20:46
  • @twalberg : just as most RANDOM functions have 2 versions, one which allows specifying a starting seed and one that doesn't (this producing truly random numbers), I can see the need for a developer wanting to have a deterministic path for debugging problems that seem like they might be related to ASLR. It would have been a better Q if the OP had indicated some research on the problem ;-/ Anyway, it's a New Year, Good luck to all! – shellter Jan 03 '17 at 23:16
  • @shellter Yes, and perhaps I'm just a bit too pedantic, but if an algorithm provides any means (starting seed or anything else) to provide the same stream more than once, I tend to think of it as at best pseudo-random. And in the case of ASLR, it would provide a means to completely defeat the security of the feature, such that there would be no point in even having it... – twalberg Jan 04 '17 at 02:04

0 Answers0