Recent events have inspired me to take a closer look at the /dev/random
driver and make a number of changes to improve it. The changes reflect
relatively minor changes across the entire /dev/random architecture,
including:
* how we collect entropy on non-x86 architectures
* optimizing the CPU overhead when we collect entropy
* improving our entropy accumulation estimates
* small changes to the mixing function suggested by researchers who have
been analyzing the /dev/random driver
I would appreciate folks taking a look at the changes and making any
comments, asking questions, etc.
Cheers!
- Ted
H. Peter Anvin (3):
random: Statically compute poolbitshift, poolbytes, poolbits
random: Allow fractional bits to be tracked
random: Account for entropy loss due to overwrites
Theodore Ts'o (9):
random: run random_int_secret_init() run after all late_initcalls
random: fix the tracepoint for get_random_bytes(_arch)
random: optimize spinlock use in add_device_randomness()
random: allow architectures to optionally define random_get_entropy()
random: mix in architectural randomness earlier in extract_buf()
random: optimize the entropy_store structure
random: cap the rate which the /dev/urandom pool gets reseeded
random: speed up the fast_mix function by a factor of four
random: adjust the generator polynomials in the mixing function
slightly
drivers/char/random.c | 419 +++++++++++++++++++++++++++---------------
include/linux/random.h | 1 +
include/linux/timex.h | 17 ++
include/trace/events/random.h | 33 +++-
init/main.c | 2 +
5 files changed, 321 insertions(+), 151 deletions(-)
--
1.7.12.rc0.22.gcdd159b
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
