LibCryptParse

Soon to be a library for parsing linux' /proc/crypto file

Now collecting /proc/crypto files from the wild to analyze them and search for common patterns.

Contributing

For now, you can help by running ./collect.sh on your Linux machine. It should be run from the project root directory and it will save a copy of your /proc/crypto file into the samples/ subdirectory. You may run cryptsetup benchmark to load additional common ciphers/modes.

Observations

Below is a table of relationships between cipher type and which fields it may contain. A ✔ means the field is always present, a ✕ means that field is never present, a . means the field may or may not be present. It may also happen that a whole row is just empty, this happens if the supplied crypto file lacks a cipher of that type.

This table is generated by the ./parse.py script, which takes paths to crypto files as command line arguments. So you can run ./parse.py /proc/crypto to genarate this table for your system, running ./parse.py samples/*/* should yield the table below.

type/field	async	blocksize	chunksize	digestsize	driver	geniv	internal	ivsize	maxauthsize	max keysize	min keysize	module	name	priority	refcnt	seedsize	selftest	type	walksize
cipher	✕	✔	✕	✕	✔	✕	.	✕	✕	✔	✔	✔	✔	✔	.	✕	.	✔	✕
ablkcipher	✔	✔	✕	✕	✔	✔	.	✔	✕	✔	✔	✔	✔	✔	✔	✕	✔	✔	✕
akcipher	✕	✕	✕	✕	✔	✕	.	✕	✕	✕	✕	✔	✔	✔	✔	✕	✔	✔	✕
blkcipher	✕	✔	✕	✕	✔	✔	.	✔	✕	✔	✔	✔	✔	✔	✔	✕	✔	✔	✕
givcipher	✔	✔	✕	✕	✔	✔	✕	✔	✕	✔	✔	✔	✔	✔	✔	✕	✔	✔	✕
skcipher	✔	✔	✔	✕	✔	✕	✔	✔	✕	✔	✔	✔	✔	✔	✔	✕	✔	✔	✔
aead	✔	✔	✕	✕	✔	✔	.	✔	✔	✕	✕	✔	✔	✔	✔	✕	✔	✔	✕
nivaead	✔	✔	✕	✕	✔	✔	✕	✔	✔	✕	✕	✔	✔	✔	✔	✕	✔	✔	✕
ahash	✔	✔	✕	✔	✔	✕	.	✕	✕	✕	✕	✔	✔	✔	✔	✕	✔	✔	✕
shash	✕	✔	✕	✔	✔	✕	.	✕	✕	✕	✕	✔	✔	✔	✔	✕	✔	✔	✕
compression	✕	✕	✕	✕	✔	✕	.	✕	✕	✕	✕	✔	✔	✔	.	✕	.	✔	✕
digest	✕	✔	✕	✔	✔	✕	✕	✕	✕	✕	✕	✔	✔	✔	.	✕	✕	✔	✕
kpp	✕	✕	✕	✕	✔	✕	.	✕	✕	✕	✕	✔	✔	✔	✔	✕	✔	✔	✕
pcomp	✕	✕	✕	✕	✔	✕	.	✕	✕	✕	✕	✔	✔	✔	✔	✕	✔	✔	✕
scomp	✕	✕	✕	✕	✔	✕	.	✕	✕	✕	✕	✔	✔	✔	✔	✕	✔	✔	✕
rng	✕	✕	✕	✕	✔	✕	.	✕	✕	✕	✕	✔	✔	✔	✔	✔	✔	✔	✕