БЕГЕМОТИКОВ МОЖНО! Войти !bnw Сегодня Клубы

Из битемаджего [chan] privacy:
Truecrypt certainly is better than notusing encryption at all. If you want whole disk encryption look into dm-crypt, else veracrypt/gostcrypt or GPG will do fine.

If you really are concerned about backdoors that much I like to show you the following:

Ever heard of intrinsics? It's about using hardware-implemented functions. To see what intrinsics your processor supports run "cat /proc/cpuinfo" (on unix-systems).

Intel processer povide intrinsics for AES and if you have a modern computer with an intel cpu, you are sure to have those intrinsics. And openssl happily uses these intrinsics per default. Here is a minor problem:

If you know how AES works, you know that encryption and decryption take the same amount of work. When using the openssl functions for AES encryption (not the ones which use intrinsics!!!) however, encryption can take roughly twice as long as decryption. now let's look at the intrinsic: If you benchmark those functions you'll notice that encryption takes waaayyy too long compared to decryption.

Of course that doesn't proove that there's a backdoor, but if you have sensitive information whiches security is in your hands, you should keep the above information in mind.

If you use veracrypt, you will have options available to disable intrinsics. To my knowledge, dm-crypt uses intrinsics per defualt. Feel free to run cryptsetup benchmark.

If you want to be sure that there is no hardware backdoor, I suggest you use an uncommon cipher such as serpent that have no intrinsics available. Keep in mind that doesn't make https and gpg magically stop using AES. I will yet have to find out myself how to disable intrinsics in dmcrypt and openssl and would be most grateful if someone on BM could post the answer.

KISS:
If you handle very sensitive stuff, make sure you do not use intrinsics. And don't forget to airgap your machine.
For personal use, veracrypt (and in my point of view truecrypt too, correct me if I'm wrong) are most suitable and I enjoy the acceleration provided by intrinsics.
Though to make it harder for NSA to spy on you and learn something new, feel free to do some custom stuff such as a hardware number generator or a seperately encrypted container for your personal keys.

Here are my sources:
http://www.ct.de/cs1403154
https://en.wikipedia.org/wiki/AES_instruction_set
https://wiki.archlinux.org/index.php/Dm-crypt/Device_encryption
[feel free to add more, such as sources direct from openssl]
[maybe someone could provide more details about backdoors in instruction sets]

Рекомендовали: @minoru
#MC8UIV / @ninesigns / 3581 день назад

Этот дебил реально считает, что можно так изменить детерминированный алгоритм так, чтобы он выдавал уязвимый криптотекст, и чтобы при этом оригинальный алгоритм дешифрования получал на выходе правильный plaintext? Или я что-то пропустила и AES - недетерминированный алгоритм?
#MC8UIV/TND / @etw / 3581 день назад
@etw Короче, сравнить софтовую и хардовую версии надобно.
#MC8UIV/0AV / @sanovach --> #MC8UIV/TND / 3580 дней назад
Чо, реально дольше работают? У меня Core i3 мобильный, первого поколения, не могу у себя погонять.
#MC8UIV/KMY / @minoru / 3580 дней назад
ipv6 ready BnW для ведрофона BnW на Реформале Викивач Котятки

Цоперайт © 2010-2016 @stiletto.