Speed up Linux crypto operations on the One A110 laptop with VIA Padlock

One Mini A110 subnotebook

OK, so I've been hacking on and testing my shiny new One A110 mini-laptop during the last few days and I must say I'm very happy with it. I'll write up some more details later (check the wiki if you're impatient), but today I want to highlight a very nice feature of this laptop (compared to, for instance, the Eee PC): The VIA C7-M ULV CPU in the laptop has VIA Padlock support.

VIA Padlock is a hardware feature in recent VIA CPUs which provides hardware-accelerated AES and SHA-1/SHA-256 support, among other things. This can be used in Linux (with the proper drivers and patches) to improve performance of dm-crypt, OpenSSL (and all programs using it), scp, sha1sum, OpenVPN, etc. etc.

I have written a quite extensive VIA Padlock HOWTO and benchmarks in the A110 wiki (but all of this will work on other systems which have VIA Padlock, too). To summarize, here are the most important benchmarks:

dm-crypt (256bit AES, cbc-essiv:sha256)

VIA Padlock dm-crypt benchmark

Without VIA Padlock support:

$ hdparm -tT /dev/mapper/hdc2_crypt
 Timing cached reads:   448 MB in  2.00 seconds = 223.47 MB/sec
 Timing buffered disk reads:   22 MB in  3.07 seconds =   7.17 MB/sec

With VIA Padlock support:

$ hdparm -tT /dev/mapper/hdc2_crypt
 Timing cached reads:   502 MB in  2.00 seconds = 250.41 MB/sec
 Timing buffered disk reads:   90 MB in  3.07 seconds =  29.36 MB/sec

The native speed of the SSD in the laptop is 31.01 MB/sec, so there is almost no performance penalty when using VIA Padlock.


VIA Padlock OpenSSL benchmark

OpenSSL speed benchmark, first line without Padlock, second line with Padlock enabled:

$ openssl speed -evp aes-256-cbc [-engine padlock]
type             16 bytes     64 bytes    256 bytes   1024 bytes   8192 bytes
aes-256-cbc       9187.18k    10572.28k    11054.32k    11179.36k    11218.02k
aes-256-cbc      47955.92k   150619.73k   325730.73k   458320.11k   520520.79k


VIA Padlock scp benchmark

Without VIA Padlock support:

$ scp -c aes256-cbc bigfile.dat localhost:/dev/null
bigfile.dat                100%  159MB   5.9MB/s   00:27

With VIA Padlock support:

$ scp -c aes256-cbc bigfile.dat localhost:/dev/null
bigfile.dat                100%  159MB  14.5MB/s   00:11


A real speed benchmark is pending (not measurable easily on 100MBit LAN, will try on a slower link), but as OpenVPN uses OpenSSL it should have roughly the same speedup iff you tell OpenVPN to use AES (it uses Blowfish per default).

Also, there's a measurable difference in CPU load while tranferring large files over OpenVPN: 8% CPU load with VIA Padlock (vs. 20% CPU load without VIA Padlock).

sha1sum / phe_sum

VIA Padlock sha1sum / phe_sum benchmark

phe_sum is a small C program which can be used as drop-in replacement for sha1sum (which doesn't support VIA Padlock yet). Quick benchmark:

sha1sum, without VIA Padlock:

$ time sha1sum bigfile.dat
real    0m6.511s
user    0m5.864s
sys     0m0.412s

phe_sum (with VIA Padlock support):

$ time ./phe_sum bigfile.dat
real    0m1.149s
user    0m0.704s
sys     0m0.424s

All in all VIA Padlock gives you a pretty impressive speedup for many crypto-using applications on Linux, which is especially useful on the A110 mini-laptop (think OpenVPN or scp for mobile usage, and dm-crypt for an encrypted SSD, of course).

Encrypted SMS Solutions?

Dear Lazyweb,

I was recently thinking about SMS encryption (you know, the short messages sent from your cell phone). Or MMS encryption for that matter. Are there any Free Software solutions to implement such a thing, based on well-known crypto primitives and proven implementations thereof?

From a quick glance I could not find anything usable, only lots of commercial, closed-source "solutions" which cost money and are basically crap.

I was thinking about hacking together a small Java application (so that most modern phones can run it) which basically asks for a password and pipes your text through AES before sending the SMS. On the receiver's side, you enter the same password and get to see the plaintext. That's pretty much it.

If you want something more elaborate you can use public-key crypto (basically embed GnuPG or similar into the application), but the above should be fine for most uses.

Anyways, I do not want to start implementing something like this if there are other, more mature projects out there which do the same...

Why you'd want to do this (other than simple paranoia, or the fact that governments and other institutions are spying on everyone these days, whether they're allowed to or not)? Well, one good reason for SMS encryption is when you're using some kind of SMS gateways, e.g. a website which gives you 2-3 free SMS per month if you sign up with them and give them tons of personal data. That alone is crappy enough, but you don't want their admins to be able to read and store all your SMSes in addition (which consist of simple plain-text, transmitted though HTTP in this case!). Neither should any local or remote scriptkiddy who knows how to use a sniffer be able to read your private SMSes.

Solution: Write the text in your favorite text editor, pipe it through aespipe (for example), cut'n'paste the result in the web form of the SMS service. The receiver does everthing backwards and you're done.

HOWTO: Encrypted USB thumb drives and (USB) hard disks using loop-AES

Yet another thing that has been on my TODO list for quite a while: encrypted USB thumb drives and/or encrypted external USB hard drives.

I have finally tried this over the weekend using loop-AES. This is very useful for securing your USB thumb drive contents in case you lose it or it gets stolen. Also, I use an external USB hard drive for backups (previously unencrypted). This is encryped now, too.

Here's a quick HOWTO:

  1. Get the loop-AES kernel patches, apply them, enable "AES encrypted loop device support" in "Device Drivers -> Block Devices -> Loopback device support", and recompile the kernel.
    I also enabled "loop encryption key scrubbing support" as it seems to promise higher security (can anybody confirm that?).
    If you're using the Debian kernel packages, apt-get install loop-aes-2.6-686 (or a similar package) should suffice.
  2. Get a loop-aes enabled losetup, mount etc.:
    apt-get install loop-aes-utils
  3. Securely delete the target partition: shred -n 1 -v /dev/sda3.
    Use -n 25 or higher if you want more security and have a few days time to wait for the thing to finish...
  4. Setup the loopback device: losetup -e aes256 -C 3 -S 'seed' /dev/loop0 /dev/sda3.

    • I used AES-256 as cipher, but others are possible.
    • The -C 3 means "run hashed password through 3000 iterations of AES-256 before using it for loop encryption. This consumes lots of CPU cycles at loop setup/mount time but not thereafter." (see losetup(8)). This is supposed to be more secure.
    • Using -S 'seed' (replace "seed" with a secret string like "g7sN4" or something) should make brute force attacks a bit harder. Don't forget the seed!
    • You'll be asked for a passphrase > 20 characters. Choose a good one. Don't forget it!
  5. Create the filesystem (I used ext3): mke2fs -j /dev/loop0
  6. Detach the loopback device: losetup -d /dev/loop0
  7. Add this to /etc/fstab:
    /dev/sda3 /mnt/crypted_sda3 ext3 noauto,loop=/dev/loop0,encryption=AES256,itercountk=3 0 0
  8. Mount the (now encrypted) partition by supplying the seed and entering the chosen password: mount -o pseed=seed /mnt/crypted_sda3
  9. Done. You can now copy stuff to /mnt/crypted_sda3 which will be encrypted automatically.

For a more detailed guide read the Encrypted-Root-Filesystem-HOWTO. A performance comparison of different ciphers is available, but in general I didn't notice too much of a slow-down because of the encryption...

Syndicate content