Re: [dm-crypt] couple of questions related to using crypto map

[Arno Wagner <arno@wagner.name>]

On Mon, Jun 29, 2020 at 02:13:21 CEST, Fourhundred Thecat wrote:
> Hello,
> 
> after some experiments, I came up with following 4 questions. I think
> that the answer to all of them is yes, but I would like to confirm this
> with the experts here:
> 
> 1)
> A crypto map aes-xts-plain64:sha512 with a random 512bit key is created
> over a block device. When zeroes are written to the device through the
> crypto map (encrypting), does it generate quality random data on the
> block device (comparable to /dev/urandom)?

Well, yes and no. Yes, if you just ise it once. No, as it gets
written to disk and that is it.
 
> 2)
> The same crypto map over a block device, but the block devices contains
> zeroes. When I read data, will it generate quality random data?
> (comparable to /dev/urandom or a random number generator?)

See above.

> 3)
> If yes should it be the fastest way to generate random data on a typical
> Linux computer, compared to reading /dev/urandom? (especially when
> having AES instruction support in the CPU)

No. /dev/urandom has gotten a lot faster. 

> 4)
> If the data obtained by reading from the zeroed device through that
> crypto map (aes-xts-plain64:sha512) is written back to the block device
> (using the same crypto map and key), will you get the original data? (in
> this example zeros). 

That is how disk encryption works, at least when you do not have
per-sector metadata and LUKS does not.

Regards,
Arno
-- 
Arno Wagner,     Dr. sc. techn., Dipl. Inform.,    Email: arno@wagner.name
GnuPG: ID: CB5D9718  FP: 12D6 C03B 1B30 33BB 13CF  B774 E35C 5FA1 CB5D 9718
----
A good decision is based on knowledge and not on numbers. -- Plato

If it's in the news, don't worry about it.  The very definition of 
"news" is "something that hardly ever happens." -- Bruce Schneier