Messages in this thread | | | Date | Fri, 27 Jul 2012 22:39:38 -0400 | From | Theodore Ts'o <> | Subject | Re: [PATCH] random: mix in architectural randomness in extract_buf() |
| |
Ok, I'll add this patch to the random tree. I've modified the commit message a bit since the speed advertisement of RDRAND is rather pointless --- processes aren't generating session keys or long term keys at a high rate, and programs can't count on /dev/random being super fast and having unlimited entropy, since for most platforms and even most x86 CPU's deployed in service today, this isn't true --- and making your userspace program depond upon /dev/random in such a way that it only works on Ivy Bridge CPU's might be good for Intel from a vendor lock-in perspective, but it's really bad, non-portable programming style.
Also, in the future arch_get_random_long() will almost certainly be hooked up for other architectures, so putting an extended advertisement for RDRAND really isn't appropriate.
- Ted
commit d2e7c96af1e54b507ae2a6a7dd2baf588417a7e5 Author: H. Peter Anvin <hpa@linux.intel.com> Date: Fri Jul 27 22:26:08 2012 -0400
random: mix in architectural randomness in extract_buf()
Mix in any architectural randomness in extract_buf() instead of xfer_secondary_buf(). This allows us to mix in more architectural randomness, and it also makes xfer_secondary_buf() faster, moving a tiny bit of additional CPU overhead to process which is extracting the randomness.
[ Commit description modified by tytso to remove an extended advertisement for the RDRAND instruction. ]
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com> Acked-by: Ingo Molnar <mingo@kernel.org> Cc: DJ Johnston <dj.johnston@intel.com> Signed-off-by: Theodore Ts'o <tytso@mit.edu> Cc: stable@vger.kernel.org
| |