Initial commit: Linux kernel 4.19 for Unplugged OS

This commit is contained in:
ziv committed 2025-10-06 13:52:42 +00:00
commit d30235b04c
80519 files changed
+35800681

No files matched your search

+23
View File
@@ -0,0 +1,23 @@
Authenticated Encryption With Associated Data (AEAD) Algorithm Definitions
--------------------------------------------------------------------------
.. kernel-doc:: include/crypto/aead.h
:doc: Authenticated Encryption With Associated Data (AEAD) Cipher API
.. kernel-doc:: include/crypto/aead.h
:functions: aead_request aead_alg
Authenticated Encryption With Associated Data (AEAD) Cipher API
---------------------------------------------------------------
.. kernel-doc:: include/crypto/aead.h
:functions: crypto_alloc_aead crypto_free_aead crypto_aead_ivsize crypto_aead_authsize crypto_aead_blocksize crypto_aead_setkey crypto_aead_setauthsize crypto_aead_encrypt crypto_aead_decrypt
Asynchronous AEAD Request Handle
--------------------------------
.. kernel-doc:: include/crypto/aead.h
:doc: Asynchronous AEAD Request Handle
.. kernel-doc:: include/crypto/aead.h
:functions: crypto_aead_reqsize aead_request_set_tfm aead_request_alloc aead_request_free aead_request_set_callback aead_request_set_crypt aead_request_set_ad
+20
View File
@@ -0,0 +1,20 @@
Asymmetric Cipher Algorithm Definitions
---------------------------------------
.. kernel-doc:: include/crypto/akcipher.h
:functions: akcipher_alg akcipher_request
Asymmetric Cipher API
---------------------
.. kernel-doc:: include/crypto/akcipher.h
:doc: Generic Public Key API
.. kernel-doc:: include/crypto/akcipher.h
:functions: crypto_alloc_akcipher crypto_free_akcipher crypto_akcipher_set_pub_key crypto_akcipher_set_priv_key crypto_akcipher_maxsize crypto_akcipher_encrypt crypto_akcipher_decrypt crypto_akcipher_sign crypto_akcipher_verify
Asymmetric Cipher Request Handle
--------------------------------
.. kernel-doc:: include/crypto/akcipher.h
:functions: akcipher_request_alloc akcipher_request_free akcipher_request_set_callback akcipher_request_set_crypt
+35
View File
@@ -0,0 +1,35 @@
Message Digest Algorithm Definitions
------------------------------------
.. kernel-doc:: include/crypto/hash.h
:doc: Message Digest Algorithm Definitions
.. kernel-doc:: include/crypto/hash.h
:functions: hash_alg_common ahash_alg shash_alg
Asynchronous Message Digest API
-------------------------------
.. kernel-doc:: include/crypto/hash.h
:doc: Asynchronous Message Digest API
.. kernel-doc:: include/crypto/hash.h
:functions: crypto_alloc_ahash crypto_free_ahash crypto_ahash_init crypto_ahash_digestsize crypto_ahash_reqtfm crypto_ahash_reqsize crypto_ahash_statesize crypto_ahash_setkey crypto_ahash_finup crypto_ahash_final crypto_ahash_digest crypto_ahash_export crypto_ahash_import
Asynchronous Hash Request Handle
--------------------------------
.. kernel-doc:: include/crypto/hash.h
:doc: Asynchronous Hash Request Handle
.. kernel-doc:: include/crypto/hash.h
:functions: ahash_request_set_tfm ahash_request_alloc ahash_request_free ahash_request_set_callback ahash_request_set_crypt
Synchronous Message Digest API
------------------------------
.. kernel-doc:: include/crypto/hash.h
:doc: Synchronous Message Digest API
.. kernel-doc:: include/crypto/hash.h
:functions: crypto_alloc_shash crypto_free_shash crypto_shash_blocksize crypto_shash_digestsize crypto_shash_descsize crypto_shash_setkey crypto_shash_digest crypto_shash_export crypto_shash_import crypto_shash_init crypto_shash_update crypto_shash_final crypto_shash_finup
+250
View File
@@ -0,0 +1,250 @@
Scatterlist Cryptographic API
INTRODUCTION
The Scatterlist Crypto API takes page vectors (scatterlists) as
arguments, and works directly on pages. In some cases (e.g. ECB
mode ciphers), this will allow for pages to be encrypted in-place
with no copying.
One of the initial goals of this design was to readily support IPsec,
so that processing can be applied to paged skb's without the need
for linearization.
DETAILS
At the lowest level are algorithms, which register dynamically with the
API.
'Transforms' are user-instantiated objects, which maintain state, handle all
of the implementation logic (e.g. manipulating page vectors) and provide an
abstraction to the underlying algorithms. However, at the user
level they are very simple.
Conceptually, the API layering looks like this:
[transform api] (user interface)
[transform ops] (per-type logic glue e.g. cipher.c, compress.c)
[algorithm api] (for registering algorithms)
The idea is to make the user interface and algorithm registration API
very simple, while hiding the core logic from both. Many good ideas
from existing APIs such as Cryptoapi and Nettle have been adapted for this.
The API currently supports five main types of transforms: AEAD (Authenticated
Encryption with Associated Data), Block Ciphers, Ciphers, Compressors and
Hashes.
Please note that Block Ciphers is somewhat of a misnomer. It is in fact
meant to support all ciphers including stream ciphers. The difference
between Block Ciphers and Ciphers is that the latter operates on exactly
one block while the former can operate on an arbitrary amount of data,
subject to block size requirements (i.e., non-stream ciphers can only
process multiples of blocks).
Here's an example of how to use the API:
#include <crypto/hash.h>
#include <linux/err.h>
#include <linux/scatterlist.h>
struct scatterlist sg[2];
char result[128];
struct crypto_ahash *tfm;
struct ahash_request *req;
tfm = crypto_alloc_ahash("md5", 0, CRYPTO_ALG_ASYNC);
if (IS_ERR(tfm))
fail();
/* ... set up the scatterlists ... */
req = ahash_request_alloc(tfm, GFP_ATOMIC);
if (!req)
fail();
ahash_request_set_callback(req, 0, NULL, NULL);
ahash_request_set_crypt(req, sg, result, 2);
if (crypto_ahash_digest(req))
fail();
ahash_request_free(req);
crypto_free_ahash(tfm);
Many real examples are available in the regression test module (tcrypt.c).
DEVELOPER NOTES
Transforms may only be allocated in user context, and cryptographic
methods may only be called from softirq and user contexts. For
transforms with a setkey method it too should only be called from
user context.
When using the API for ciphers, performance will be optimal if each
scatterlist contains data which is a multiple of the cipher's block
size (typically 8 bytes). This prevents having to do any copying
across non-aligned page fragment boundaries.
ADDING NEW ALGORITHMS
When submitting a new algorithm for inclusion, a mandatory requirement
is that at least a few test vectors from known sources (preferably
standards) be included.
Converting existing well known code is preferred, as it is more likely
to have been reviewed and widely tested. If submitting code from LGPL
sources, please consider changing the license to GPL (see section 3 of
the LGPL).
Algorithms submitted must also be generally patent-free (e.g. IDEA
will not be included in the mainline until around 2011), and be based
on a recognized standard and/or have been subjected to appropriate
peer review.
Also check for any RFCs which may relate to the use of specific algorithms,
as well as general application notes such as RFC2451 ("The ESP CBC-Mode
Cipher Algorithms").
It's a good idea to avoid using lots of macros and use inlined functions
instead, as gcc does a good job with inlining, while excessive use of
macros can cause compilation problems on some platforms.
Also check the TODO list at the web site listed below to see what people
might already be working on.
BUGS
Send bug reports to:
linux-crypto@vger.kernel.org
Cc: Herbert Xu <herbert@gondor.apana.org.au>,
David S. Miller <davem@redhat.com>
FURTHER INFORMATION
For further patches and various updates, including the current TODO
list, see:
http://gondor.apana.org.au/~herbert/crypto/
AUTHORS
James Morris
David S. Miller
Herbert Xu
CREDITS
The following people provided invaluable feedback during the development
of the API:
Alexey Kuznetzov
Rusty Russell
Herbert Valerio Riedel
Jeff Garzik
Michael Richardson
Andrew Morton
Ingo Oeser
Christoph Hellwig
Portions of this API were derived from the following projects:
Kerneli Cryptoapi (http://www.kerneli.org/)
Alexander Kjeldaas
Herbert Valerio Riedel
Kyle McMartin
Jean-Luc Cooke
David Bryson
Clemens Fruhwirth
Tobias Ringstrom
Harald Welte
and;
Nettle (http://www.lysator.liu.se/~nisse/nettle/)
Niels Möller
Original developers of the crypto algorithms:
Dana L. How (DES)
Andrew Tridgell and Steve French (MD4)
Colin Plumb (MD5)
Steve Reid (SHA1)
Jean-Luc Cooke (SHA256, SHA384, SHA512)
Kazunori Miyazawa / USAGI (HMAC)
Matthew Skala (Twofish)
Dag Arne Osvik (Serpent)
Brian Gladman (AES)
Kartikey Mahendra Bhatt (CAST6)
Jon Oberheide (ARC4)
Jouni Malinen (Michael MIC)
NTT(Nippon Telegraph and Telephone Corporation) (Camellia)
SHA1 algorithm contributors:
Jean-Francois Dive
DES algorithm contributors:
Raimar Falke
Gisle Sælensminde
Niels Möller
Blowfish algorithm contributors:
Herbert Valerio Riedel
Kyle McMartin
Twofish algorithm contributors:
Werner Koch
Marc Mutz
SHA256/384/512 algorithm contributors:
Andrew McDonald
Kyle McMartin
Herbert Valerio Riedel
AES algorithm contributors:
Alexander Kjeldaas
Herbert Valerio Riedel
Kyle McMartin
Adam J. Richter
Fruhwirth Clemens (i586)
Linus Torvalds (i586)
CAST5 algorithm contributors:
Kartikey Mahendra Bhatt (original developers unknown, FSF copyright).
TEA/XTEA algorithm contributors:
Aaron Grothe
Michael Ringe
Khazad algorithm contributors:
Aaron Grothe
Whirlpool algorithm contributors:
Aaron Grothe
Jean-Luc Cooke
Anubis algorithm contributors:
Aaron Grothe
Tiger algorithm contributors:
Aaron Grothe
VIA PadLock contributors:
Michal Ludvig
Camellia algorithm contributors:
NTT(Nippon Telegraph and Telephone Corporation) (Camellia)
Generic scatterwalk code by Adam J. Richter <adam@yggdrasil.com>
Please send any credits updates or corrections to:
Herbert Xu <herbert@gondor.apana.org.au>
+38
View File
@@ -0,0 +1,38 @@
Key-agreement Protocol Primitives (KPP) Cipher Algorithm Definitions
--------------------------------------------------------------------
.. kernel-doc:: include/crypto/kpp.h
:functions: kpp_request crypto_kpp kpp_alg kpp_secret
Key-agreement Protocol Primitives (KPP) Cipher API
--------------------------------------------------
.. kernel-doc:: include/crypto/kpp.h
:doc: Generic Key-agreement Protocol Primitives API
.. kernel-doc:: include/crypto/kpp.h
:functions: crypto_alloc_kpp crypto_free_kpp crypto_kpp_set_secret crypto_kpp_generate_public_key crypto_kpp_compute_shared_secret crypto_kpp_maxsize
Key-agreement Protocol Primitives (KPP) Cipher Request Handle
-------------------------------------------------------------
.. kernel-doc:: include/crypto/kpp.h
:functions: kpp_request_alloc kpp_request_free kpp_request_set_callback kpp_request_set_input kpp_request_set_output
ECDH Helper Functions
---------------------
.. kernel-doc:: include/crypto/ecdh.h
:doc: ECDH Helper Functions
.. kernel-doc:: include/crypto/ecdh.h
:functions: ecdh crypto_ecdh_key_len crypto_ecdh_encode_key crypto_ecdh_decode_key
DH Helper Functions
-------------------
.. kernel-doc:: include/crypto/dh.h
:doc: DH Helper Functions
.. kernel-doc:: include/crypto/dh.h
:functions: dh crypto_dh_key_len crypto_dh_encode_key crypto_dh_decode_key
+14
View File
@@ -0,0 +1,14 @@
Random Number Algorithm Definitions
-----------------------------------
.. kernel-doc:: include/crypto/rng.h
:functions: rng_alg
Crypto API Random Number API
----------------------------
.. kernel-doc:: include/crypto/rng.h
:doc: Random number generator API
.. kernel-doc:: include/crypto/rng.h
:functions: crypto_alloc_rng crypto_rng_alg crypto_free_rng crypto_rng_generate crypto_rng_get_bytes crypto_rng_reset crypto_rng_seedsize
+210
View File
@@ -0,0 +1,210 @@
Code Examples
=============
Code Example For Symmetric Key Cipher Operation
-----------------------------------------------
::
/* tie all data structures together */
struct skcipher_def {
struct scatterlist sg;
struct crypto_skcipher *tfm;
struct skcipher_request *req;
struct crypto_wait wait;
};
/* Perform cipher operation */
static unsigned int test_skcipher_encdec(struct skcipher_def *sk,
int enc)
{
int rc;
if (enc)
rc = crypto_wait_req(crypto_skcipher_encrypt(sk->req), &sk->wait);
else
rc = crypto_wait_req(crypto_skcipher_decrypt(sk->req), &sk->wait);
if (rc)
pr_info("skcipher encrypt returned with result %d\n", rc);
return rc;
}
/* Initialize and trigger cipher operation */
static int test_skcipher(void)
{
struct skcipher_def sk;
struct crypto_skcipher *skcipher = NULL;
struct skcipher_request *req = NULL;
char *scratchpad = NULL;
char *ivdata = NULL;
unsigned char key[32];
int ret = -EFAULT;
skcipher = crypto_alloc_skcipher("cbc-aes-aesni", 0, 0);
if (IS_ERR(skcipher)) {
pr_info("could not allocate skcipher handle\n");
return PTR_ERR(skcipher);
}
req = skcipher_request_alloc(skcipher, GFP_KERNEL);
if (!req) {
pr_info("could not allocate skcipher request\n");
ret = -ENOMEM;
goto out;
}
skcipher_request_set_callback(req, CRYPTO_TFM_REQ_MAY_BACKLOG,
crypto_req_done,
&sk.wait);
/* AES 256 with random key */
get_random_bytes(&key, 32);
if (crypto_skcipher_setkey(skcipher, key, 32)) {
pr_info("key could not be set\n");
ret = -EAGAIN;
goto out;
}
/* IV will be random */
ivdata = kmalloc(16, GFP_KERNEL);
if (!ivdata) {
pr_info("could not allocate ivdata\n");
goto out;
}
get_random_bytes(ivdata, 16);
/* Input data will be random */
scratchpad = kmalloc(16, GFP_KERNEL);
if (!scratchpad) {
pr_info("could not allocate scratchpad\n");
goto out;
}
get_random_bytes(scratchpad, 16);
sk.tfm = skcipher;
sk.req = req;
/* We encrypt one block */
sg_init_one(&sk.sg, scratchpad, 16);
skcipher_request_set_crypt(req, &sk.sg, &sk.sg, 16, ivdata);
crypto_init_wait(&sk.wait);
/* encrypt data */
ret = test_skcipher_encdec(&sk, 1);
if (ret)
goto out;
pr_info("Encryption triggered successfully\n");
out:
if (skcipher)
crypto_free_skcipher(skcipher);
if (req)
skcipher_request_free(req);
if (ivdata)
kfree(ivdata);
if (scratchpad)
kfree(scratchpad);
return ret;
}
Code Example For Use of Operational State Memory With SHASH
-----------------------------------------------------------
::
struct sdesc {
struct shash_desc shash;
char ctx[];
};
static struct sdesc *init_sdesc(struct crypto_shash *alg)
{
struct sdesc *sdesc;
int size;
size = sizeof(struct shash_desc) + crypto_shash_descsize(alg);
sdesc = kmalloc(size, GFP_KERNEL);
if (!sdesc)
return ERR_PTR(-ENOMEM);
sdesc->shash.tfm = alg;
sdesc->shash.flags = 0x0;
return sdesc;
}
static int calc_hash(struct crypto_shash *alg,
const unsigned char *data, unsigned int datalen,
unsigned char *digest)
{
struct sdesc *sdesc;
int ret;
sdesc = init_sdesc(alg);
if (IS_ERR(sdesc)) {
pr_info("can't alloc sdesc\n");
return PTR_ERR(sdesc);
}
ret = crypto_shash_digest(&sdesc->shash, data, datalen, digest);
kfree(sdesc);
return ret;
}
static int test_hash(const unsigned char *data, unsigned int datalen,
unsigned char *digest)
{
struct crypto_shash *alg;
char *hash_alg_name = "sha1-padlock-nano";
int ret;
alg = crypto_alloc_shash(hash_alg_name, 0, 0);
if (IS_ERR(alg)) {
pr_info("can't alloc alg %s\n", hash_alg_name);
return PTR_ERR(alg);
}
ret = calc_hash(alg, data, datalen, digest);
crypto_free_shash(alg);
return ret;
}
Code Example For Random Number Generator Usage
----------------------------------------------
::
static int get_random_numbers(u8 *buf, unsigned int len)
{
struct crypto_rng *rng = NULL;
char *drbg = "drbg_nopr_sha256"; /* Hash DRBG with SHA-256, no PR */
int ret;
if (!buf || !len) {
pr_debug("No output buffer provided\n");
return -EINVAL;
}
rng = crypto_alloc_rng(drbg, 0, 0);
if (IS_ERR(rng)) {
pr_debug("could not allocate RNG handle for %s\n", drbg);
return PTR_ERR(rng);
}
ret = crypto_rng_get_bytes(rng, buf, len);
if (ret < 0)
pr_debug("generation of random numbers failed\n");
else if (ret == 0)
pr_debug("RNG returned no data");
else
pr_debug("RNG returned %d bytes of data\n", ret);
out:
crypto_free_rng(rng);
return ret;
}
+62
View File
@@ -0,0 +1,62 @@
Block Cipher Algorithm Definitions
----------------------------------
.. kernel-doc:: include/linux/crypto.h
:doc: Block Cipher Algorithm Definitions
.. kernel-doc:: include/linux/crypto.h
:functions: crypto_alg ablkcipher_alg blkcipher_alg cipher_alg
Symmetric Key Cipher API
------------------------
.. kernel-doc:: include/crypto/skcipher.h
:doc: Symmetric Key Cipher API
.. kernel-doc:: include/crypto/skcipher.h
:functions: crypto_alloc_skcipher crypto_free_skcipher crypto_has_skcipher crypto_skcipher_ivsize crypto_skcipher_blocksize crypto_skcipher_setkey crypto_skcipher_reqtfm crypto_skcipher_encrypt crypto_skcipher_decrypt
Symmetric Key Cipher Request Handle
-----------------------------------
.. kernel-doc:: include/crypto/skcipher.h
:doc: Symmetric Key Cipher Request Handle
.. kernel-doc:: include/crypto/skcipher.h
:functions: crypto_skcipher_reqsize skcipher_request_set_tfm skcipher_request_alloc skcipher_request_free skcipher_request_set_callback skcipher_request_set_crypt
Single Block Cipher API
-----------------------
.. kernel-doc:: include/linux/crypto.h
:doc: Single Block Cipher API
.. kernel-doc:: include/linux/crypto.h
:functions: crypto_alloc_cipher crypto_free_cipher crypto_has_cipher crypto_cipher_blocksize crypto_cipher_setkey crypto_cipher_encrypt_one crypto_cipher_decrypt_one
Asynchronous Block Cipher API - Deprecated
------------------------------------------
.. kernel-doc:: include/linux/crypto.h
:doc: Asynchronous Block Cipher API
.. kernel-doc:: include/linux/crypto.h
:functions: crypto_free_ablkcipher crypto_has_ablkcipher crypto_ablkcipher_ivsize crypto_ablkcipher_blocksize crypto_ablkcipher_setkey crypto_ablkcipher_reqtfm crypto_ablkcipher_encrypt crypto_ablkcipher_decrypt
Asynchronous Cipher Request Handle - Deprecated
-----------------------------------------------
.. kernel-doc:: include/linux/crypto.h
:doc: Asynchronous Cipher Request Handle
.. kernel-doc:: include/linux/crypto.h
:functions: crypto_ablkcipher_reqsize ablkcipher_request_set_tfm ablkcipher_request_alloc ablkcipher_request_free ablkcipher_request_set_callback ablkcipher_request_set_crypt
Synchronous Block Cipher API - Deprecated
-----------------------------------------
.. kernel-doc:: include/linux/crypto.h
:doc: Synchronous Block Cipher API
.. kernel-doc:: include/linux/crypto.h
:functions: crypto_alloc_blkcipher crypto_free_blkcipher crypto_has_blkcipher crypto_blkcipher_name crypto_blkcipher_ivsize crypto_blkcipher_blocksize crypto_blkcipher_setkey crypto_blkcipher_encrypt crypto_blkcipher_encrypt_iv crypto_blkcipher_decrypt crypto_blkcipher_decrypt_iv crypto_blkcipher_set_iv crypto_blkcipher_get_iv
+25
View File
@@ -0,0 +1,25 @@
Programming Interface
=====================
Please note that the kernel crypto API contains the AEAD givcrypt API
(crypto_aead_giv\* and aead_givcrypt\* function calls in
include/crypto/aead.h). This API is obsolete and will be removed in the
future. To obtain the functionality of an AEAD cipher with internal IV
generation, use the IV generator as a regular cipher. For example,
rfc4106(gcm(aes)) is the AEAD cipher with external IV generation and
seqniv(rfc4106(gcm(aes))) implies that the kernel crypto API generates
the IV. Different IV generators are available.
.. class:: toc-title
Table of contents
.. toctree::
:maxdepth: 2
api-skcipher
api-aead
api-digest
api-rng
api-akcipher
api-kpp
+441
View File
@@ -0,0 +1,441 @@
Kernel Crypto API Architecture
==============================
Cipher algorithm types
----------------------
The kernel crypto API provides different API calls for the following
cipher types:
- Symmetric ciphers
- AEAD ciphers
- Message digest, including keyed message digest
- Random number generation
- User space interface
Ciphers And Templates
---------------------
The kernel crypto API provides implementations of single block ciphers
and message digests. In addition, the kernel crypto API provides
numerous "templates" that can be used in conjunction with the single
block ciphers and message digests. Templates include all types of block
chaining mode, the HMAC mechanism, etc.
Single block ciphers and message digests can either be directly used by
a caller or invoked together with a template to form multi-block ciphers
or keyed message digests.
A single block cipher may even be called with multiple templates.
However, templates cannot be used without a single cipher.
See /proc/crypto and search for "name". For example:
- aes
- ecb(aes)
- cmac(aes)
- ccm(aes)
- rfc4106(gcm(aes))
- sha1
- hmac(sha1)
- authenc(hmac(sha1),cbc(aes))
In these examples, "aes" and "sha1" are the ciphers and all others are
the templates.
Synchronous And Asynchronous Operation
--------------------------------------
The kernel crypto API provides synchronous and asynchronous API
operations.
When using the synchronous API operation, the caller invokes a cipher
operation which is performed synchronously by the kernel crypto API.
That means, the caller waits until the cipher operation completes.
Therefore, the kernel crypto API calls work like regular function calls.
For synchronous operation, the set of API calls is small and
conceptually similar to any other crypto library.
Asynchronous operation is provided by the kernel crypto API which
implies that the invocation of a cipher operation will complete almost
instantly. That invocation triggers the cipher operation but it does not
signal its completion. Before invoking a cipher operation, the caller
must provide a callback function the kernel crypto API can invoke to
signal the completion of the cipher operation. Furthermore, the caller
must ensure it can handle such asynchronous events by applying
appropriate locking around its data. The kernel crypto API does not
perform any special serialization operation to protect the caller's data
integrity.
Crypto API Cipher References And Priority
-----------------------------------------
A cipher is referenced by the caller with a string. That string has the
following semantics:
::
template(single block cipher)
where "template" and "single block cipher" is the aforementioned
template and single block cipher, respectively. If applicable,
additional templates may enclose other templates, such as
::
template1(template2(single block cipher)))
The kernel crypto API may provide multiple implementations of a template
or a single block cipher. For example, AES on newer Intel hardware has
the following implementations: AES-NI, assembler implementation, or
straight C. Now, when using the string "aes" with the kernel crypto API,
which cipher implementation is used? The answer to that question is the
priority number assigned to each cipher implementation by the kernel
crypto API. When a caller uses the string to refer to a cipher during
initialization of a cipher handle, the kernel crypto API looks up all
implementations providing an implementation with that name and selects
the implementation with the highest priority.
Now, a caller may have the need to refer to a specific cipher
implementation and thus does not want to rely on the priority-based
selection. To accommodate this scenario, the kernel crypto API allows
the cipher implementation to register a unique name in addition to
common names. When using that unique name, a caller is therefore always
sure to refer to the intended cipher implementation.
The list of available ciphers is given in /proc/crypto. However, that
list does not specify all possible permutations of templates and
ciphers. Each block listed in /proc/crypto may contain the following
information -- if one of the components listed as follows are not
applicable to a cipher, it is not displayed:
- name: the generic name of the cipher that is subject to the
priority-based selection -- this name can be used by the cipher
allocation API calls (all names listed above are examples for such
generic names)
- driver: the unique name of the cipher -- this name can be used by the
cipher allocation API calls
- module: the kernel module providing the cipher implementation (or
"kernel" for statically linked ciphers)
- priority: the priority value of the cipher implementation
- refcnt: the reference count of the respective cipher (i.e. the number
of current consumers of this cipher)
- selftest: specification whether the self test for the cipher passed
- type:
- skcipher for symmetric key ciphers
- cipher for single block ciphers that may be used with an
additional template
- shash for synchronous message digest
- ahash for asynchronous message digest
- aead for AEAD cipher type
- compression for compression type transformations
- rng for random number generator
- givcipher for cipher with associated IV generator (see the geniv
entry below for the specification of the IV generator type used by
the cipher implementation)
- kpp for a Key-agreement Protocol Primitive (KPP) cipher such as
an ECDH or DH implementation
- blocksize: blocksize of cipher in bytes
- keysize: key size in bytes
- ivsize: IV size in bytes
- seedsize: required size of seed data for random number generator
- digestsize: output size of the message digest
- geniv: IV generation type:
- eseqiv for encrypted sequence number based IV generation
- seqiv for sequence number based IV generation
- chainiv for chain iv generation
- <builtin> is a marker that the cipher implements IV generation and
handling as it is specific to the given cipher
Key Sizes
---------
When allocating a cipher handle, the caller only specifies the cipher
type. Symmetric ciphers, however, typically support multiple key sizes
(e.g. AES-128 vs. AES-192 vs. AES-256). These key sizes are determined
with the length of the provided key. Thus, the kernel crypto API does
not provide a separate way to select the particular symmetric cipher key
size.
Cipher Allocation Type And Masks
--------------------------------
The different cipher handle allocation functions allow the specification
of a type and mask flag. Both parameters have the following meaning (and
are therefore not covered in the subsequent sections).
The type flag specifies the type of the cipher algorithm. The caller
usually provides a 0 when the caller wants the default handling.
Otherwise, the caller may provide the following selections which match
the aforementioned cipher types:
- CRYPTO_ALG_TYPE_CIPHER Single block cipher
- CRYPTO_ALG_TYPE_COMPRESS Compression
- CRYPTO_ALG_TYPE_AEAD Authenticated Encryption with Associated Data
(MAC)
- CRYPTO_ALG_TYPE_BLKCIPHER Synchronous multi-block cipher
- CRYPTO_ALG_TYPE_ABLKCIPHER Asynchronous multi-block cipher
- CRYPTO_ALG_TYPE_GIVCIPHER Asynchronous multi-block cipher packed
together with an IV generator (see geniv field in the /proc/crypto
listing for the known IV generators)
- CRYPTO_ALG_TYPE_KPP Key-agreement Protocol Primitive (KPP) such as
an ECDH or DH implementation
- CRYPTO_ALG_TYPE_DIGEST Raw message digest
- CRYPTO_ALG_TYPE_HASH Alias for CRYPTO_ALG_TYPE_DIGEST
- CRYPTO_ALG_TYPE_SHASH Synchronous multi-block hash
- CRYPTO_ALG_TYPE_AHASH Asynchronous multi-block hash
- CRYPTO_ALG_TYPE_RNG Random Number Generation
- CRYPTO_ALG_TYPE_AKCIPHER Asymmetric cipher
- CRYPTO_ALG_TYPE_PCOMPRESS Enhanced version of
CRYPTO_ALG_TYPE_COMPRESS allowing for segmented compression /
decompression instead of performing the operation on one segment
only. CRYPTO_ALG_TYPE_PCOMPRESS is intended to replace
CRYPTO_ALG_TYPE_COMPRESS once existing consumers are converted.
The mask flag restricts the type of cipher. The only allowed flag is
CRYPTO_ALG_ASYNC to restrict the cipher lookup function to
asynchronous ciphers. Usually, a caller provides a 0 for the mask flag.
When the caller provides a mask and type specification, the caller
limits the search the kernel crypto API can perform for a suitable
cipher implementation for the given cipher name. That means, even when a
caller uses a cipher name that exists during its initialization call,
the kernel crypto API may not select it due to the used type and mask
field.
Internal Structure of Kernel Crypto API
---------------------------------------
The kernel crypto API has an internal structure where a cipher
implementation may use many layers and indirections. This section shall
help to clarify how the kernel crypto API uses various components to
implement the complete cipher.
The following subsections explain the internal structure based on
existing cipher implementations. The first section addresses the most
complex scenario where all other scenarios form a logical subset.
Generic AEAD Cipher Structure
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The following ASCII art decomposes the kernel crypto API layers when
using the AEAD cipher with the automated IV generation. The shown
example is used by the IPSEC layer.
For other use cases of AEAD ciphers, the ASCII art applies as well, but
the caller may not use the AEAD cipher with a separate IV generator. In
this case, the caller must generate the IV.
The depicted example decomposes the AEAD cipher of GCM(AES) based on the
generic C implementations (gcm.c, aes-generic.c, ctr.c, ghash-generic.c,
seqiv.c). The generic implementation serves as an example showing the
complete logic of the kernel crypto API.
It is possible that some streamlined cipher implementations (like
AES-NI) provide implementations merging aspects which in the view of the
kernel crypto API cannot be decomposed into layers any more. In case of
the AES-NI implementation, the CTR mode, the GHASH implementation and
the AES cipher are all merged into one cipher implementation registered
with the kernel crypto API. In this case, the concept described by the
following ASCII art applies too. However, the decomposition of GCM into
the individual sub-components by the kernel crypto API is not done any
more.
Each block in the following ASCII art is an independent cipher instance
obtained from the kernel crypto API. Each block is accessed by the
caller or by other blocks using the API functions defined by the kernel
crypto API for the cipher implementation type.
The blocks below indicate the cipher type as well as the specific logic
implemented in the cipher.
The ASCII art picture also indicates the call structure, i.e. who calls
which component. The arrows point to the invoked block where the caller
uses the API applicable to the cipher type specified for the block.
::
kernel crypto API | IPSEC Layer
|
+-----------+ |
| | (1)
| aead | <----------------------------------- esp_output
| (seqiv) | ---+
+-----------+ |
| (2)
+-----------+ |
| | <--+ (2)
| aead | <----------------------------------- esp_input
| (gcm) | ------------+
+-----------+ |
| (3) | (5)
v v
+-----------+ +-----------+
| | | |
| skcipher | | ahash |
| (ctr) | ---+ | (ghash) |
+-----------+ | +-----------+
|
+-----------+ | (4)
| | <--+
| cipher |
| (aes) |
+-----------+
The following call sequence is applicable when the IPSEC layer triggers
an encryption operation with the esp_output function. During
configuration, the administrator set up the use of rfc4106(gcm(aes)) as
the cipher for ESP. The following call sequence is now depicted in the
ASCII art above:
1. esp_output() invokes crypto_aead_encrypt() to trigger an
encryption operation of the AEAD cipher with IV generator.
In case of GCM, the SEQIV implementation is registered as GIVCIPHER
in crypto_rfc4106_alloc().
The SEQIV performs its operation to generate an IV where the core
function is seqiv_geniv().
2. Now, SEQIV uses the AEAD API function calls to invoke the associated
AEAD cipher. In our case, during the instantiation of SEQIV, the
cipher handle for GCM is provided to SEQIV. This means that SEQIV
invokes AEAD cipher operations with the GCM cipher handle.
During instantiation of the GCM handle, the CTR(AES) and GHASH
ciphers are instantiated. The cipher handles for CTR(AES) and GHASH
are retained for later use.
The GCM implementation is responsible to invoke the CTR mode AES and
the GHASH cipher in the right manner to implement the GCM
specification.
3. The GCM AEAD cipher type implementation now invokes the SKCIPHER API
with the instantiated CTR(AES) cipher handle.
During instantiation of the CTR(AES) cipher, the CIPHER type
implementation of AES is instantiated. The cipher handle for AES is
retained.
That means that the SKCIPHER implementation of CTR(AES) only
implements the CTR block chaining mode. After performing the block
chaining operation, the CIPHER implementation of AES is invoked.
4. The SKCIPHER of CTR(AES) now invokes the CIPHER API with the AES
cipher handle to encrypt one block.
5. The GCM AEAD implementation also invokes the GHASH cipher
implementation via the AHASH API.
When the IPSEC layer triggers the esp_input() function, the same call
sequence is followed with the only difference that the operation starts
with step (2).
Generic Block Cipher Structure
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Generic block ciphers follow the same concept as depicted with the ASCII
art picture above.
For example, CBC(AES) is implemented with cbc.c, and aes-generic.c. The
ASCII art picture above applies as well with the difference that only
step (4) is used and the SKCIPHER block chaining mode is CBC.
Generic Keyed Message Digest Structure
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Keyed message digest implementations again follow the same concept as
depicted in the ASCII art picture above.
For example, HMAC(SHA256) is implemented with hmac.c and
sha256_generic.c. The following ASCII art illustrates the
implementation:
::
kernel crypto API | Caller
|
+-----------+ (1) |
| | <------------------ some_function
| ahash |
| (hmac) | ---+
+-----------+ |
| (2)
+-----------+ |
| | <--+
| shash |
| (sha256) |
+-----------+
The following call sequence is applicable when a caller triggers an HMAC
operation:
1. The AHASH API functions are invoked by the caller. The HMAC
implementation performs its operation as needed.
During initialization of the HMAC cipher, the SHASH cipher type of
SHA256 is instantiated. The cipher handle for the SHA256 instance is
retained.
At one time, the HMAC implementation requires a SHA256 operation
where the SHA256 cipher handle is used.
2. The HMAC instance now invokes the SHASH API with the SHA256 cipher
handle to calculate the message digest.
+413
View File
@@ -0,0 +1,413 @@
=============================================
ASYMMETRIC / PUBLIC-KEY CRYPTOGRAPHY KEY TYPE
=============================================
Contents:
- Overview.
- Key identification.
- Accessing asymmetric keys.
- Signature verification.
- Asymmetric key subtypes.
- Instantiation data parsers.
- Keyring link restrictions.
========
OVERVIEW
========
The "asymmetric" key type is designed to be a container for the keys used in
public-key cryptography, without imposing any particular restrictions on the
form or mechanism of the cryptography or form of the key.
The asymmetric key is given a subtype that defines what sort of data is
associated with the key and provides operations to describe and destroy it.
However, no requirement is made that the key data actually be stored in the
key.
A completely in-kernel key retention and operation subtype can be defined, but
it would also be possible to provide access to cryptographic hardware (such as
a TPM) that might be used to both retain the relevant key and perform
operations using that key. In such a case, the asymmetric key would then
merely be an interface to the TPM driver.
Also provided is the concept of a data parser. Data parsers are responsible
for extracting information from the blobs of data passed to the instantiation
function. The first data parser that recognises the blob gets to set the
subtype of the key and define the operations that can be done on that key.
A data parser may interpret the data blob as containing the bits representing a
key, or it may interpret it as a reference to a key held somewhere else in the
system (for example, a TPM).
==================
KEY IDENTIFICATION
==================
If a key is added with an empty name, the instantiation data parsers are given
the opportunity to pre-parse a key and to determine the description the key
should be given from the content of the key.
This can then be used to refer to the key, either by complete match or by
partial match. The key type may also use other criteria to refer to a key.
The asymmetric key type's match function can then perform a wider range of
comparisons than just the straightforward comparison of the description with
the criterion string:
(1) If the criterion string is of the form "id:<hexdigits>" then the match
function will examine a key's fingerprint to see if the hex digits given
after the "id:" match the tail. For instance:
keyctl search @s asymmetric id:5acc2142
will match a key with fingerprint:
1A00 2040 7601 7889 DE11 882C 3823 04AD 5ACC 2142
(2) If the criterion string is of the form "<subtype>:<hexdigits>" then the
match will match the ID as in (1), but with the added restriction that
only keys of the specified subtype (e.g. tpm) will be matched. For
instance:
keyctl search @s asymmetric tpm:5acc2142
Looking in /proc/keys, the last 8 hex digits of the key fingerprint are
displayed, along with the subtype:
1a39e171 I----- 1 perm 3f010000 0 0 asymmetric modsign.0: DSA 5acc2142 []
=========================
ACCESSING ASYMMETRIC KEYS
=========================
For general access to asymmetric keys from within the kernel, the following
inclusion is required:
#include <crypto/public_key.h>
This gives access to functions for dealing with asymmetric / public keys.
Three enums are defined there for representing public-key cryptography
algorithms:
enum pkey_algo
digest algorithms used by those:
enum pkey_hash_algo
and key identifier representations:
enum pkey_id_type
Note that the key type representation types are required because key
identifiers from different standards aren't necessarily compatible. For
instance, PGP generates key identifiers by hashing the key data plus some
PGP-specific metadata, whereas X.509 has arbitrary certificate identifiers.
The operations defined upon a key are:
(1) Signature verification.
Other operations are possible (such as encryption) with the same key data
required for verification, but not currently supported, and others
(eg. decryption and signature generation) require extra key data.
SIGNATURE VERIFICATION
----------------------
An operation is provided to perform cryptographic signature verification, using
an asymmetric key to provide or to provide access to the public key.
int verify_signature(const struct key *key,
const struct public_key_signature *sig);
The caller must have already obtained the key from some source and can then use
it to check the signature. The caller must have parsed the signature and
transferred the relevant bits to the structure pointed to by sig.
struct public_key_signature {
u8 *digest;
u8 digest_size;
enum pkey_hash_algo pkey_hash_algo : 8;
u8 nr_mpi;
union {
MPI mpi[2];
...
};
};
The algorithm used must be noted in sig->pkey_hash_algo, and all the MPIs that
make up the actual signature must be stored in sig->mpi[] and the count of MPIs
placed in sig->nr_mpi.
In addition, the data must have been digested by the caller and the resulting
hash must be pointed to by sig->digest and the size of the hash be placed in
sig->digest_size.
The function will return 0 upon success or -EKEYREJECTED if the signature
doesn't match.
The function may also return -ENOTSUPP if an unsupported public-key algorithm
or public-key/hash algorithm combination is specified or the key doesn't
support the operation; -EBADMSG or -ERANGE if some of the parameters have weird
data; or -ENOMEM if an allocation can't be performed. -EINVAL can be returned
if the key argument is the wrong type or is incompletely set up.
=======================
ASYMMETRIC KEY SUBTYPES
=======================
Asymmetric keys have a subtype that defines the set of operations that can be
performed on that key and that determines what data is attached as the key
payload. The payload format is entirely at the whim of the subtype.
The subtype is selected by the key data parser and the parser must initialise
the data required for it. The asymmetric key retains a reference on the
subtype module.
The subtype definition structure can be found in:
#include <keys/asymmetric-subtype.h>
and looks like the following:
struct asymmetric_key_subtype {
struct module *owner;
const char *name;
void (*describe)(const struct key *key, struct seq_file *m);
void (*destroy)(void *payload);
int (*verify_signature)(const struct key *key,
const struct public_key_signature *sig);
};
Asymmetric keys point to this with their payload[asym_subtype] member.
The owner and name fields should be set to the owning module and the name of
the subtype. Currently, the name is only used for print statements.
There are a number of operations defined by the subtype:
(1) describe().
Mandatory. This allows the subtype to display something in /proc/keys
against the key. For instance the name of the public key algorithm type
could be displayed. The key type will display the tail of the key
identity string after this.
(2) destroy().
Mandatory. This should free the memory associated with the key. The
asymmetric key will look after freeing the fingerprint and releasing the
reference on the subtype module.
(3) verify_signature().
Optional. These are the entry points for the key usage operations.
Currently there is only the one defined. If not set, the caller will be
given -ENOTSUPP. The subtype may do anything it likes to implement an
operation, including offloading to hardware.
==========================
INSTANTIATION DATA PARSERS
==========================
The asymmetric key type doesn't generally want to store or to deal with a raw
blob of data that holds the key data. It would have to parse it and error
check it each time it wanted to use it. Further, the contents of the blob may
have various checks that can be performed on it (eg. self-signatures, validity
dates) and may contain useful data about the key (identifiers, capabilities).
Also, the blob may represent a pointer to some hardware containing the key
rather than the key itself.
Examples of blob formats for which parsers could be implemented include:
- OpenPGP packet stream [RFC 4880].
- X.509 ASN.1 stream.
- Pointer to TPM key.
- Pointer to UEFI key.
During key instantiation each parser in the list is tried until one doesn't
return -EBADMSG.
The parser definition structure can be found in:
#include <keys/asymmetric-parser.h>
and looks like the following:
struct asymmetric_key_parser {
struct module *owner;
const char *name;
int (*parse)(struct key_preparsed_payload *prep);
};
The owner and name fields should be set to the owning module and the name of
the parser.
There is currently only a single operation defined by the parser, and it is
mandatory:
(1) parse().
This is called to preparse the key from the key creation and update paths.
In particular, it is called during the key creation _before_ a key is
allocated, and as such, is permitted to provide the key's description in
the case that the caller declines to do so.
The caller passes a pointer to the following struct with all of the fields
cleared, except for data, datalen and quotalen [see
Documentation/security/keys/core.rst].
struct key_preparsed_payload {
char *description;
void *payload[4];
const void *data;
size_t datalen;
size_t quotalen;
};
The instantiation data is in a blob pointed to by data and is datalen in
size. The parse() function is not permitted to change these two values at
all, and shouldn't change any of the other values _unless_ they are
recognise the blob format and will not return -EBADMSG to indicate it is
not theirs.
If the parser is happy with the blob, it should propose a description for
the key and attach it to ->description, ->payload[asym_subtype] should be
set to point to the subtype to be used, ->payload[asym_crypto] should be
set to point to the initialised data for that subtype,
->payload[asym_key_ids] should point to one or more hex fingerprints and
quotalen should be updated to indicate how much quota this key should
account for.
When clearing up, the data attached to ->payload[asym_key_ids] and
->description will be kfree()'d and the data attached to
->payload[asm_crypto] will be passed to the subtype's ->destroy() method
to be disposed of. A module reference for the subtype pointed to by
->payload[asym_subtype] will be put.
If the data format is not recognised, -EBADMSG should be returned. If it
is recognised, but the key cannot for some reason be set up, some other
negative error code should be returned. On success, 0 should be returned.
The key's fingerprint string may be partially matched upon. For a
public-key algorithm such as RSA and DSA this will likely be a printable
hex version of the key's fingerprint.
Functions are provided to register and unregister parsers:
int register_asymmetric_key_parser(struct asymmetric_key_parser *parser);
void unregister_asymmetric_key_parser(struct asymmetric_key_parser *subtype);
Parsers may not have the same name. The names are otherwise only used for
displaying in debugging messages.
=========================
KEYRING LINK RESTRICTIONS
=========================
Keyrings created from userspace using add_key can be configured to check the
signature of the key being linked. Keys without a valid signature are not
allowed to link.
Several restriction methods are available:
(1) Restrict using the kernel builtin trusted keyring
- Option string used with KEYCTL_RESTRICT_KEYRING:
- "builtin_trusted"
The kernel builtin trusted keyring will be searched for the signing key.
If the builtin trusted keyring is not configured, all links will be
rejected. The ca_keys kernel parameter also affects which keys are used
for signature verification.
(2) Restrict using the kernel builtin and secondary trusted keyrings
- Option string used with KEYCTL_RESTRICT_KEYRING:
- "builtin_and_secondary_trusted"
The kernel builtin and secondary trusted keyrings will be searched for the
signing key. If the secondary trusted keyring is not configured, this
restriction will behave like the "builtin_trusted" option. The ca_keys
kernel parameter also affects which keys are used for signature
verification.
(3) Restrict using a separate key or keyring
- Option string used with KEYCTL_RESTRICT_KEYRING:
- "key_or_keyring:<key or keyring serial number>[:chain]"
Whenever a key link is requested, the link will only succeed if the key
being linked is signed by one of the designated keys. This key may be
specified directly by providing a serial number for one asymmetric key, or
a group of keys may be searched for the signing key by providing the
serial number for a keyring.
When the "chain" option is provided at the end of the string, the keys
within the destination keyring will also be searched for signing keys.
This allows for verification of certificate chains by adding each
certificate in order (starting closest to the root) to a keyring. For
instance, one keyring can be populated with links to a set of root
certificates, with a separate, restricted keyring set up for each
certificate chain to be validated:
# Create and populate a keyring for root certificates
root_id=`keyctl add keyring root-certs "" @s`
keyctl padd asymmetric "" $root_id < root1.cert
keyctl padd asymmetric "" $root_id < root2.cert
# Create and restrict a keyring for the certificate chain
chain_id=`keyctl add keyring chain "" @s`
keyctl restrict_keyring $chain_id asymmetric key_or_keyring:$root_id:chain
# Attempt to add each certificate in the chain, starting with the
# certificate closest to the root.
keyctl padd asymmetric "" $chain_id < intermediateA.cert
keyctl padd asymmetric "" $chain_id < intermediateB.cert
keyctl padd asymmetric "" $chain_id < end-entity.cert
If the final end-entity certificate is successfully added to the "chain"
keyring, we can be certain that it has a valid signing chain going back to
one of the root certificates.
A single keyring can be used to verify a chain of signatures by
restricting the keyring after linking the root certificate:
# Create a keyring for the certificate chain and add the root
chain2_id=`keyctl add keyring chain2 "" @s`
keyctl padd asymmetric "" $chain2_id < root1.cert
# Restrict the keyring that already has root1.cert linked. The cert
# will remain linked by the keyring.
keyctl restrict_keyring $chain2_id asymmetric key_or_keyring:0:chain
# Attempt to add each certificate in the chain, starting with the
# certificate closest to the root.
keyctl padd asymmetric "" $chain2_id < intermediateA.cert
keyctl padd asymmetric "" $chain2_id < intermediateB.cert
keyctl padd asymmetric "" $chain2_id < end-entity.cert
If the final end-entity certificate is successfully added to the "chain2"
keyring, we can be certain that there is a valid signing chain going back
to the root certificate that was added before the keyring was restricted.
In all of these cases, if the signing key is found the signature of the key to
be linked will be verified using the signing key. The requested key is added
to the keyring only if the signature is successfully verified. -ENOKEY is
returned if the parent certificate could not be found, or -EKEYREJECTED is
returned if the signature check fails or the key is blacklisted. Other errors
may be returned if the signature check could not be performed.
+225
View File
@@ -0,0 +1,225 @@
Asynchronous Transfers/Transforms API
1 INTRODUCTION
2 GENEALOGY
3 USAGE
3.1 General format of the API
3.2 Supported operations
3.3 Descriptor management
3.4 When does the operation execute?
3.5 When does the operation complete?
3.6 Constraints
3.7 Example
4 DMAENGINE DRIVER DEVELOPER NOTES
4.1 Conformance points
4.2 "My application needs exclusive control of hardware channels"
5 SOURCE
---
1 INTRODUCTION
The async_tx API provides methods for describing a chain of asynchronous
bulk memory transfers/transforms with support for inter-transactional
dependencies. It is implemented as a dmaengine client that smooths over
the details of different hardware offload engine implementations. Code
that is written to the API can optimize for asynchronous operation and
the API will fit the chain of operations to the available offload
resources.
2 GENEALOGY
The API was initially designed to offload the memory copy and
xor-parity-calculations of the md-raid5 driver using the offload engines
present in the Intel(R) Xscale series of I/O processors. It also built
on the 'dmaengine' layer developed for offloading memory copies in the
network stack using Intel(R) I/OAT engines. The following design
features surfaced as a result:
1/ implicit synchronous path: users of the API do not need to know if
the platform they are running on has offload capabilities. The
operation will be offloaded when an engine is available and carried out
in software otherwise.
2/ cross channel dependency chains: the API allows a chain of dependent
operations to be submitted, like xor->copy->xor in the raid5 case. The
API automatically handles cases where the transition from one operation
to another implies a hardware channel switch.
3/ dmaengine extensions to support multiple clients and operation types
beyond 'memcpy'
3 USAGE
3.1 General format of the API:
struct dma_async_tx_descriptor *
async_<operation>(<op specific parameters>, struct async_submit ctl *submit)
3.2 Supported operations:
memcpy - memory copy between a source and a destination buffer
memset - fill a destination buffer with a byte value
xor - xor a series of source buffers and write the result to a
destination buffer
xor_val - xor a series of source buffers and set a flag if the
result is zero. The implementation attempts to prevent
writes to memory
pq - generate the p+q (raid6 syndrome) from a series of source buffers
pq_val - validate that a p and or q buffer are in sync with a given series of
sources
datap - (raid6_datap_recov) recover a raid6 data block and the p block
from the given sources
2data - (raid6_2data_recov) recover 2 raid6 data blocks from the given
sources
3.3 Descriptor management:
The return value is non-NULL and points to a 'descriptor' when the operation
has been queued to execute asynchronously. Descriptors are recycled
resources, under control of the offload engine driver, to be reused as
operations complete. When an application needs to submit a chain of
operations it must guarantee that the descriptor is not automatically recycled
before the dependency is submitted. This requires that all descriptors be
acknowledged by the application before the offload engine driver is allowed to
recycle (or free) the descriptor. A descriptor can be acked by one of the
following methods:
1/ setting the ASYNC_TX_ACK flag if no child operations are to be submitted
2/ submitting an unacknowledged descriptor as a dependency to another
async_tx call will implicitly set the acknowledged state.
3/ calling async_tx_ack() on the descriptor.
3.4 When does the operation execute?
Operations do not immediately issue after return from the
async_<operation> call. Offload engine drivers batch operations to
improve performance by reducing the number of mmio cycles needed to
manage the channel. Once a driver-specific threshold is met the driver
automatically issues pending operations. An application can force this
event by calling async_tx_issue_pending_all(). This operates on all
channels since the application has no knowledge of channel to operation
mapping.
3.5 When does the operation complete?
There are two methods for an application to learn about the completion
of an operation.
1/ Call dma_wait_for_async_tx(). This call causes the CPU to spin while
it polls for the completion of the operation. It handles dependency
chains and issuing pending operations.
2/ Specify a completion callback. The callback routine runs in tasklet
context if the offload engine driver supports interrupts, or it is
called in application context if the operation is carried out
synchronously in software. The callback can be set in the call to
async_<operation>, or when the application needs to submit a chain of
unknown length it can use the async_trigger_callback() routine to set a
completion interrupt/callback at the end of the chain.
3.6 Constraints:
1/ Calls to async_<operation> are not permitted in IRQ context. Other
contexts are permitted provided constraint #2 is not violated.
2/ Completion callback routines cannot submit new operations. This
results in recursion in the synchronous case and spin_locks being
acquired twice in the asynchronous case.
3.7 Example:
Perform a xor->copy->xor operation where each operation depends on the
result from the previous operation:
void callback(void *param)
{
struct completion *cmp = param;
complete(cmp);
}
void run_xor_copy_xor(struct page **xor_srcs,
int xor_src_cnt,
struct page *xor_dest,
size_t xor_len,
struct page *copy_src,
struct page *copy_dest,
size_t copy_len)
{
struct dma_async_tx_descriptor *tx;
addr_conv_t addr_conv[xor_src_cnt];
struct async_submit_ctl submit;
addr_conv_t addr_conv[NDISKS];
struct completion cmp;
init_async_submit(&submit, ASYNC_TX_XOR_DROP_DST, NULL, NULL, NULL,
addr_conv);
tx = async_xor(xor_dest, xor_srcs, 0, xor_src_cnt, xor_len, &submit)
submit->depend_tx = tx;
tx = async_memcpy(copy_dest, copy_src, 0, 0, copy_len, &submit);
init_completion(&cmp);
init_async_submit(&submit, ASYNC_TX_XOR_DROP_DST | ASYNC_TX_ACK, tx,
callback, &cmp, addr_conv);
tx = async_xor(xor_dest, xor_srcs, 0, xor_src_cnt, xor_len, &submit);
async_tx_issue_pending_all();
wait_for_completion(&cmp);
}
See include/linux/async_tx.h for more information on the flags. See the
ops_run_* and ops_complete_* routines in drivers/md/raid5.c for more
implementation examples.
4 DRIVER DEVELOPMENT NOTES
4.1 Conformance points:
There are a few conformance points required in dmaengine drivers to
accommodate assumptions made by applications using the async_tx API:
1/ Completion callbacks are expected to happen in tasklet context
2/ dma_async_tx_descriptor fields are never manipulated in IRQ context
3/ Use async_tx_run_dependencies() in the descriptor clean up path to
handle submission of dependent operations
4.2 "My application needs exclusive control of hardware channels"
Primarily this requirement arises from cases where a DMA engine driver
is being used to support device-to-memory operations. A channel that is
performing these operations cannot, for many platform specific reasons,
be shared. For these cases the dma_request_channel() interface is
provided.
The interface is:
struct dma_chan *dma_request_channel(dma_cap_mask_t mask,
dma_filter_fn filter_fn,
void *filter_param);
Where dma_filter_fn is defined as:
typedef bool (*dma_filter_fn)(struct dma_chan *chan, void *filter_param);
When the optional 'filter_fn' parameter is set to NULL
dma_request_channel simply returns the first channel that satisfies the
capability mask. Otherwise, when the mask parameter is insufficient for
specifying the necessary channel, the filter_fn routine can be used to
disposition the available channels in the system. The filter_fn routine
is called once for each free channel in the system. Upon seeing a
suitable channel filter_fn returns DMA_ACK which flags that channel to
be the return value from dma_request_channel. A channel allocated via
this interface is exclusive to the caller, until dma_release_channel()
is called.
The DMA_PRIVATE capability flag is used to tag dma devices that should
not be used by the general-purpose allocator. It can be set at
initialization time if it is known that a channel will always be
private. Alternatively, it is set when dma_request_channel() finds an
unused "public" channel.
A couple caveats to note when implementing a driver and consumer:
1/ Once a channel has been privately allocated it will no longer be
considered by the general-purpose allocator even after a call to
dma_release_channel().
2/ Since capabilities are specified at the device level a dma_device
with multiple channels will either have all channels public, or all
channels private.
5 SOURCE
include/linux/dmaengine.h: core header file for DMA drivers and api users
drivers/dma/dmaengine.c: offload engine channel management routines
drivers/dma/: location for offload engine drivers
include/linux/async_tx.h: core header file for the async_tx api
crypto/async_tx/async_tx.c: async_tx interface to dmaengine and common code
crypto/async_tx/async_memcpy.c: copy offload
crypto/async_tx/async_xor.c: xor and xor zero sum offload
+10
View File
@@ -0,0 +1,10 @@
# -*- coding: utf-8; mode: python -*-
project = 'Linux Kernel Crypto API'
tags.add("subproject")
latex_documents = [
('index', 'crypto-api.tex', 'Linux Kernel Crypto API manual',
'The kernel development community', 'manual'),
]
+50
View File
@@ -0,0 +1,50 @@
=============
CRYPTO ENGINE
=============
Overview
--------
The crypto engine API (CE), is a crypto queue manager.
Requirement
-----------
You have to put at start of your tfm_ctx the struct crypto_engine_ctx::
struct your_tfm_ctx {
struct crypto_engine_ctx enginectx;
...
};
Why: Since CE manage only crypto_async_request, it cannot know the underlying
request_type and so have access only on the TFM.
So using container_of for accessing __ctx is impossible.
Furthermore, the crypto engine cannot know the "struct your_tfm_ctx",
so it must assume that crypto_engine_ctx is at start of it.
Order of operations
-------------------
You have to obtain a struct crypto_engine via crypto_engine_alloc_init().
And start it via crypto_engine_start().
Before transferring any request, you have to fill the enginectx.
- prepare_request: (taking a function pointer) If you need to do some processing before doing the request
- unprepare_request: (taking a function pointer) Undoing what's done in prepare_request
- do_one_request: (taking a function pointer) Do encryption for current request
Note: that those three functions get the crypto_async_request associated with the received request.
So your need to get the original request via container_of(areq, struct yourrequesttype_request, base);
When your driver receive a crypto_request, you have to transfer it to
the cryptoengine via one of:
- crypto_transfer_ablkcipher_request_to_engine()
- crypto_transfer_aead_request_to_engine()
- crypto_transfer_akcipher_request_to_engine()
- crypto_transfer_hash_request_to_engine()
- crypto_transfer_skcipher_request_to_engine()
At the end of the request process, a call to one of the following function is needed:
- crypto_finalize_ablkcipher_request
- crypto_finalize_aead_request
- crypto_finalize_akcipher_request
- crypto_finalize_hash_request
- crypto_finalize_skcipher_request
+352
View File
@@ -0,0 +1,352 @@
Below is the original README file from the descore.shar package.
------------------------------------------------------------------------------
des - fast & portable DES encryption & decryption.
Copyright (C) 1992 Dana L. How
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU Library General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Library General Public License for more details.
You should have received a copy of the GNU Library General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
Author's address: how@isl.stanford.edu
$Id: README,v 1.15 1992/05/20 00:25:32 how E $
==>> To compile after untarring/unsharring, just `make' <<==
This package was designed with the following goals:
1. Highest possible encryption/decryption PERFORMANCE.
2. PORTABILITY to any byte-addressable host with a 32bit unsigned C type
3. Plug-compatible replacement for KERBEROS's low-level routines.
This second release includes a number of performance enhancements for
register-starved machines. My discussions with Richard Outerbridge,
71755.204@compuserve.com, sparked a number of these enhancements.
To more rapidly understand the code in this package, inspect desSmallFips.i
(created by typing `make') BEFORE you tackle desCode.h. The latter is set
up in a parameterized fashion so it can easily be modified by speed-daemon
hackers in pursuit of that last microsecond. You will find it more
illuminating to inspect one specific implementation,
and then move on to the common abstract skeleton with this one in mind.
performance comparison to other available des code which i could
compile on a SPARCStation 1 (cc -O4, gcc -O2):
this code (byte-order independent):
30us per encryption (options: 64k tables, no IP/FP)
33us per encryption (options: 64k tables, FIPS standard bit ordering)
45us per encryption (options: 2k tables, no IP/FP)
48us per encryption (options: 2k tables, FIPS standard bit ordering)
275us to set a new key (uses 1k of key tables)
this has the quickest encryption/decryption routines i've seen.
since i was interested in fast des filters rather than crypt(3)
and password cracking, i haven't really bothered yet to speed up
the key setting routine. also, i have no interest in re-implementing
all the other junk in the mit kerberos des library, so i've just
provided my routines with little stub interfaces so they can be
used as drop-in replacements with mit's code or any of the mit-
compatible packages below. (note that the first two timings above
are highly variable because of cache effects).
kerberos des replacement from australia (version 1.95):
53us per encryption (uses 2k of tables)
96us to set a new key (uses 2.25k of key tables)
so despite the author's inclusion of some of the performance
improvements i had suggested to him, this package's
encryption/decryption is still slower on the sparc and 68000.
more specifically, 19-40% slower on the 68020 and 11-35% slower
on the sparc, depending on the compiler;
in full gory detail (ALT_ECB is a libdes variant):
compiler machine desCore libdes ALT_ECB slower by
gcc 2.1 -O2 Sun 3/110 304 uS 369.5uS 461.8uS 22%
cc -O1 Sun 3/110 336 uS 436.6uS 399.3uS 19%
cc -O2 Sun 3/110 360 uS 532.4uS 505.1uS 40%
cc -O4 Sun 3/110 365 uS 532.3uS 505.3uS 38%
gcc 2.1 -O2 Sun 4/50 48 uS 53.4uS 57.5uS 11%
cc -O2 Sun 4/50 48 uS 64.6uS 64.7uS 35%
cc -O4 Sun 4/50 48 uS 64.7uS 64.9uS 35%
(my time measurements are not as accurate as his).
the comments in my first release of desCore on version 1.92:
68us per encryption (uses 2k of tables)
96us to set a new key (uses 2.25k of key tables)
this is a very nice package which implements the most important
of the optimizations which i did in my encryption routines.
it's a bit weak on common low-level optimizations which is why
it's 39%-106% slower. because he was interested in fast crypt(3) and
password-cracking applications, he also used the same ideas to
speed up the key-setting routines with impressive results.
(at some point i may do the same in my package). he also implements
the rest of the mit des library.
(code from eay@psych.psy.uq.oz.au via comp.sources.misc)
fast crypt(3) package from denmark:
the des routine here is buried inside a loop to do the
crypt function and i didn't feel like ripping it out and measuring
performance. his code takes 26 sparc instructions to compute one
des iteration; above, Quick (64k) takes 21 and Small (2k) takes 37.
he claims to use 280k of tables but the iteration calculation seems
to use only 128k. his tables and code are machine independent.
(code from glad@daimi.aau.dk via alt.sources or comp.sources.misc)
swedish reimplementation of Kerberos des library
108us per encryption (uses 34k worth of tables)
134us to set a new key (uses 32k of key tables to get this speed!)
the tables used seem to be machine-independent;
he seems to have included a lot of special case code
so that, e.g., `long' loads can be used instead of 4 `char' loads
when the machine's architecture allows it.
(code obtained from chalmers.se:pub/des)
crack 3.3c package from england:
as in crypt above, the des routine is buried in a loop. it's
also very modified for crypt. his iteration code uses 16k
of tables and appears to be slow.
(code obtained from aem@aber.ac.uk via alt.sources or comp.sources.misc)
``highly optimized'' and tweaked Kerberos/Athena code (byte-order dependent):
165us per encryption (uses 6k worth of tables)
478us to set a new key (uses <1k of key tables)
so despite the comments in this code, it was possible to get
faster code AND smaller tables, as well as making the tables
machine-independent.
(code obtained from prep.ai.mit.edu)
UC Berkeley code (depends on machine-endedness):
226us per encryption
10848us to set a new key
table sizes are unclear, but they don't look very small
(code obtained from wuarchive.wustl.edu)
motivation and history
a while ago i wanted some des routines and the routines documented on sun's
man pages either didn't exist or dumped core. i had heard of kerberos,
and knew that it used des, so i figured i'd use its routines. but once
i got it and looked at the code, it really set off a lot of pet peeves -
it was too convoluted, the code had been written without taking
advantage of the regular structure of operations such as IP, E, and FP
(i.e. the author didn't sit down and think before coding),
it was excessively slow, the author had attempted to clarify the code
by adding MORE statements to make the data movement more `consistent'
instead of simplifying his implementation and cutting down on all data
movement (in particular, his use of L1, R1, L2, R2), and it was full of
idiotic `tweaks' for particular machines which failed to deliver significant
speedups but which did obfuscate everything. so i took the test data
from his verification program and rewrote everything else.
a while later i ran across the great crypt(3) package mentioned above.
the fact that this guy was computing 2 sboxes per table lookup rather
than one (and using a MUCH larger table in the process) emboldened me to
do the same - it was a trivial change from which i had been scared away
by the larger table size. in his case he didn't realize you don't need to keep
the working data in TWO forms, one for easy use of half the sboxes in
indexing, the other for easy use of the other half; instead you can keep
it in the form for the first half and use a simple rotate to get the other
half. this means i have (almost) half the data manipulation and half
the table size. in fairness though he might be encoding something particular
to crypt(3) in his tables - i didn't check.
i'm glad that i implemented it the way i did, because this C version is
portable (the ifdef's are performance enhancements) and it is faster
than versions hand-written in assembly for the sparc!
porting notes
one thing i did not want to do was write an enormous mess
which depended on endedness and other machine quirks,
and which necessarily produced different code and different lookup tables
for different machines. see the kerberos code for an example
of what i didn't want to do; all their endedness-specific `optimizations'
obfuscate the code and in the end were slower than a simpler machine
independent approach. however, there are always some portability
considerations of some kind, and i have included some options
for varying numbers of register variables.
perhaps some will still regard the result as a mess!
1) i assume everything is byte addressable, although i don't actually
depend on the byte order, and that bytes are 8 bits.
i assume word pointers can be freely cast to and from char pointers.
note that 99% of C programs make these assumptions.
i always use unsigned char's if the high bit could be set.
2) the typedef `word' means a 32 bit unsigned integral type.
if `unsigned long' is not 32 bits, change the typedef in desCore.h.
i assume sizeof(word) == 4 EVERYWHERE.
the (worst-case) cost of my NOT doing endedness-specific optimizations
in the data loading and storing code surrounding the key iterations
is less than 12%. also, there is the added benefit that
the input and output work areas do not need to be word-aligned.
OPTIONAL performance optimizations
1) you should define one of `i386,' `vax,' `mc68000,' or `sparc,'
whichever one is closest to the capabilities of your machine.
see the start of desCode.h to see exactly what this selection implies.
note that if you select the wrong one, the des code will still work;
these are just performance tweaks.
2) for those with functional `asm' keywords: you should change the
ROR and ROL macros to use machine rotate instructions if you have them.
this will save 2 instructions and a temporary per use,
or about 32 to 40 instructions per en/decryption.
note that gcc is smart enough to translate the ROL/R macros into
machine rotates!
these optimizations are all rather persnickety, yet with them you should
be able to get performance equal to assembly-coding, except that:
1) with the lack of a bit rotate operator in C, rotates have to be synthesized
from shifts. so access to `asm' will speed things up if your machine
has rotates, as explained above in (3) (not necessary if you use gcc).
2) if your machine has less than 12 32-bit registers i doubt your compiler will
generate good code.
`i386' tries to configure the code for a 386 by only declaring 3 registers
(it appears that gcc can use ebx, esi and edi to hold register variables).
however, if you like assembly coding, the 386 does have 7 32-bit registers,
and if you use ALL of them, use `scaled by 8' address modes with displacement
and other tricks, you can get reasonable routines for DesQuickCore... with
about 250 instructions apiece. For DesSmall... it will help to rearrange
des_keymap, i.e., now the sbox # is the high part of the index and
the 6 bits of data is the low part; it helps to exchange these.
since i have no way to conveniently test it i have not provided my
shoehorned 386 version. note that with this release of desCore, gcc is able
to put everything in registers(!), and generate about 370 instructions apiece
for the DesQuickCore... routines!
coding notes
the en/decryption routines each use 6 necessary register variables,
with 4 being actively used at once during the inner iterations.
if you don't have 4 register variables get a new machine.
up to 8 more registers are used to hold constants in some configurations.
i assume that the use of a constant is more expensive than using a register:
a) additionally, i have tried to put the larger constants in registers.
registering priority was by the following:
anything more than 12 bits (bad for RISC and CISC)
greater than 127 in value (can't use movq or byte immediate on CISC)
9-127 (may not be able to use CISC shift immediate or add/sub quick),
1-8 were never registered, being the cheapest constants.
b) the compiler may be too stupid to realize table and table+256 should
be assigned to different constant registers and instead repetitively
do the arithmetic, so i assign these to explicit `m' register variables
when possible and helpful.
i assume that indexing is cheaper or equivalent to auto increment/decrement,
where the index is 7 bits unsigned or smaller.
this assumption is reversed for 68k and vax.
i assume that addresses can be cheaply formed from two registers,
or from a register and a small constant.
for the 68000, the `two registers and small offset' form is used sparingly.
all index scaling is done explicitly - no hidden shifts by log2(sizeof).
the code is written so that even a dumb compiler
should never need more than one hidden temporary,
increasing the chance that everything will fit in the registers.
KEEP THIS MORE SUBTLE POINT IN MIND IF YOU REWRITE ANYTHING.
(actually, there are some code fragments now which do require two temps,
but fixing it would either break the structure of the macros or
require declaring another temporary).
special efficient data format
bits are manipulated in this arrangement most of the time (S7 S5 S3 S1):
003130292827xxxx242322212019xxxx161514131211xxxx080706050403xxxx
(the x bits are still there, i'm just emphasizing where the S boxes are).
bits are rotated left 4 when computing S6 S4 S2 S0:
282726252423xxxx201918171615xxxx121110090807xxxx040302010031xxxx
the rightmost two bits are usually cleared so the lower byte can be used
as an index into an sbox mapping table. the next two x'd bits are set
to various values to access different parts of the tables.
how to use the routines
datatypes:
pointer to 8 byte area of type DesData
used to hold keys and input/output blocks to des.
pointer to 128 byte area of type DesKeys
used to hold full 768-bit key.
must be long-aligned.
DesQuickInit()
call this before using any other routine with `Quick' in its name.
it generates the special 64k table these routines need.
DesQuickDone()
frees this table
DesMethod(m, k)
m points to a 128byte block, k points to an 8 byte des key
which must have odd parity (or -1 is returned) and which must
not be a (semi-)weak key (or -2 is returned).
normally DesMethod() returns 0.
m is filled in from k so that when one of the routines below
is called with m, the routine will act like standard des
en/decryption with the key k. if you use DesMethod,
you supply a standard 56bit key; however, if you fill in
m yourself, you will get a 768bit key - but then it won't
be standard. it's 768bits not 1024 because the least significant
two bits of each byte are not used. note that these two bits
will be set to magic constants which speed up the encryption/decryption
on some machines. and yes, each byte controls
a specific sbox during a specific iteration.
you really shouldn't use the 768bit format directly; i should
provide a routine that converts 128 6-bit bytes (specified in
S-box mapping order or something) into the right format for you.
this would entail some byte concatenation and rotation.
Des{Small|Quick}{Fips|Core}{Encrypt|Decrypt}(d, m, s)
performs des on the 8 bytes at s into the 8 bytes at d. (d,s: char *).
uses m as a 768bit key as explained above.
the Encrypt|Decrypt choice is obvious.
Fips|Core determines whether a completely standard FIPS initial
and final permutation is done; if not, then the data is loaded
and stored in a nonstandard bit order (FIPS w/o IP/FP).
Fips slows down Quick by 10%, Small by 9%.
Small|Quick determines whether you use the normal routine
or the crazy quick one which gobbles up 64k more of memory.
Small is 50% slower then Quick, but Quick needs 32 times as much
memory. Quick is included for programs that do nothing but DES,
e.g., encryption filters, etc.
Getting it to compile on your machine
there are no machine-dependencies in the code (see porting),
except perhaps the `now()' macro in desTest.c.
ALL generated tables are machine independent.
you should edit the Makefile with the appropriate optimization flags
for your compiler (MAX optimization).
Speeding up kerberos (and/or its des library)
note that i have included a kerberos-compatible interface in desUtil.c
through the functions des_key_sched() and des_ecb_encrypt().
to use these with kerberos or kerberos-compatible code put desCore.a
ahead of the kerberos-compatible library on your linker's command line.
you should not need to #include desCore.h; just include the header
file provided with the kerberos library.
Other uses
the macros in desCode.h would be very useful for putting inline des
functions in more complicated encryption routines.
+255
View File
@@ -0,0 +1,255 @@
Developing Cipher Algorithms
============================
Registering And Unregistering Transformation
--------------------------------------------
There are three distinct types of registration functions in the Crypto
API. One is used to register a generic cryptographic transformation,
while the other two are specific to HASH transformations and
COMPRESSion. We will discuss the latter two in a separate chapter, here
we will only look at the generic ones.
Before discussing the register functions, the data structure to be
filled with each, struct crypto_alg, must be considered -- see below
for a description of this data structure.
The generic registration functions can be found in
include/linux/crypto.h and their definition can be seen below. The
former function registers a single transformation, while the latter
works on an array of transformation descriptions. The latter is useful
when registering transformations in bulk, for example when a driver
implements multiple transformations.
::
int crypto_register_alg(struct crypto_alg *alg);
int crypto_register_algs(struct crypto_alg *algs, int count);
The counterparts to those functions are listed below.
::
int crypto_unregister_alg(struct crypto_alg *alg);
int crypto_unregister_algs(struct crypto_alg *algs, int count);
Notice that both registration and unregistration functions do return a
value, so make sure to handle errors. A return code of zero implies
success. Any return code < 0 implies an error.
The bulk registration/unregistration functions register/unregister each
transformation in the given array of length count. They handle errors as
follows:
- crypto_register_algs() succeeds if and only if it successfully
registers all the given transformations. If an error occurs partway
through, then it rolls back successful registrations before returning
the error code. Note that if a driver needs to handle registration
errors for individual transformations, then it will need to use the
non-bulk function crypto_register_alg() instead.
- crypto_unregister_algs() tries to unregister all the given
transformations, continuing on error. It logs errors and always
returns zero.
Single-Block Symmetric Ciphers [CIPHER]
---------------------------------------
Example of transformations: aes, arc4, ...
This section describes the simplest of all transformation
implementations, that being the CIPHER type used for symmetric ciphers.
The CIPHER type is used for transformations which operate on exactly one
block at a time and there are no dependencies between blocks at all.
Registration specifics
~~~~~~~~~~~~~~~~~~~~~~
The registration of [CIPHER] algorithm is specific in that struct
crypto_alg field .cra_type is empty. The .cra_u.cipher has to be
filled in with proper callbacks to implement this transformation.
See struct cipher_alg below.
Cipher Definition With struct cipher_alg
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Struct cipher_alg defines a single block cipher.
Here are schematics of how these functions are called when operated from
other part of the kernel. Note that the .cia_setkey() call might happen
before or after any of these schematics happen, but must not happen
during any of these are in-flight.
::
KEY ---. PLAINTEXT ---.
v v
.cia_setkey() -> .cia_encrypt()
|
'-----> CIPHERTEXT
Please note that a pattern where .cia_setkey() is called multiple times
is also valid:
::
KEY1 --. PLAINTEXT1 --. KEY2 --. PLAINTEXT2 --.
v v v v
.cia_setkey() -> .cia_encrypt() -> .cia_setkey() -> .cia_encrypt()
| |
'---> CIPHERTEXT1 '---> CIPHERTEXT2
Multi-Block Ciphers
-------------------
Example of transformations: cbc(aes), ecb(arc4), ...
This section describes the multi-block cipher transformation
implementations. The multi-block ciphers are used for transformations
which operate on scatterlists of data supplied to the transformation
functions. They output the result into a scatterlist of data as well.
Registration Specifics
~~~~~~~~~~~~~~~~~~~~~~
The registration of multi-block cipher algorithms is one of the most
standard procedures throughout the crypto API.
Note, if a cipher implementation requires a proper alignment of data,
the caller should use the functions of crypto_skcipher_alignmask() to
identify a memory alignment mask. The kernel crypto API is able to
process requests that are unaligned. This implies, however, additional
overhead as the kernel crypto API needs to perform the realignment of
the data which may imply moving of data.
Cipher Definition With struct blkcipher_alg and ablkcipher_alg
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Struct blkcipher_alg defines a synchronous block cipher whereas struct
ablkcipher_alg defines an asynchronous block cipher.
Please refer to the single block cipher description for schematics of
the block cipher usage.
Specifics Of Asynchronous Multi-Block Cipher
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
There are a couple of specifics to the asynchronous interface.
First of all, some of the drivers will want to use the Generic
ScatterWalk in case the hardware needs to be fed separate chunks of the
scatterlist which contains the plaintext and will contain the
ciphertext. Please refer to the ScatterWalk interface offered by the
Linux kernel scatter / gather list implementation.
Hashing [HASH]
--------------
Example of transformations: crc32, md5, sha1, sha256,...
Registering And Unregistering The Transformation
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
There are multiple ways to register a HASH transformation, depending on
whether the transformation is synchronous [SHASH] or asynchronous
[AHASH] and the amount of HASH transformations we are registering. You
can find the prototypes defined in include/crypto/internal/hash.h:
::
int crypto_register_ahash(struct ahash_alg *alg);
int crypto_register_shash(struct shash_alg *alg);
int crypto_register_shashes(struct shash_alg *algs, int count);
The respective counterparts for unregistering the HASH transformation
are as follows:
::
int crypto_unregister_ahash(struct ahash_alg *alg);
int crypto_unregister_shash(struct shash_alg *alg);
int crypto_unregister_shashes(struct shash_alg *algs, int count);
Cipher Definition With struct shash_alg and ahash_alg
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Here are schematics of how these functions are called when operated from
other part of the kernel. Note that the .setkey() call might happen
before or after any of these schematics happen, but must not happen
during any of these are in-flight. Please note that calling .init()
followed immediately by .finish() is also a perfectly valid
transformation.
::
I) DATA -----------.
v
.init() -> .update() -> .final() ! .update() might not be called
^ | | at all in this scenario.
'----' '---> HASH
II) DATA -----------.-----------.
v v
.init() -> .update() -> .finup() ! .update() may not be called
^ | | at all in this scenario.
'----' '---> HASH
III) DATA -----------.
v
.digest() ! The entire process is handled
| by the .digest() call.
'---------------> HASH
Here is a schematic of how the .export()/.import() functions are called
when used from another part of the kernel.
::
KEY--. DATA--.
v v ! .update() may not be called
.setkey() -> .init() -> .update() -> .export() at all in this scenario.
^ | |
'-----' '--> PARTIAL_HASH
----------- other transformations happen here -----------
PARTIAL_HASH--. DATA1--.
v v
.import -> .update() -> .final() ! .update() may not be called
^ | | at all in this scenario.
'----' '--> HASH1
PARTIAL_HASH--. DATA2-.
v v
.import -> .finup()
|
'---------------> HASH2
Note that it is perfectly legal to "abandon" a request object:
- call .init() and then (as many times) .update()
- _not_ call any of .final(), .finup() or .export() at any point in future
In other words implementations should mind the resource allocation and clean-up.
No resources related to request objects should remain allocated after a call
to .init() or .update(), since there might be no chance to free them.
Specifics Of Asynchronous HASH Transformation
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Some of the drivers will want to use the Generic ScatterWalk in case the
implementation needs to be fed separate chunks of the scatterlist which
contains the input data. The buffer containing the resulting hash will
always be properly aligned to .cra_alignmask so there is no need to
worry about this.
+25
View File
@@ -0,0 +1,25 @@
=======================
Linux Kernel Crypto API
=======================
:Author: Stephan Mueller
:Author: Marek Vasut
This documentation outlines the Linux kernel crypto API with its
concepts, details about developing cipher implementations, employment of the API
for cryptographic use cases, as well as programming examples.
.. class:: toc-title
Table of contents
.. toctree::
:maxdepth: 2
intro
architecture
devel-algos
userspace-if
crypto_engine
api
api-samples
+74
View File
@@ -0,0 +1,74 @@
Kernel Crypto API Interface Specification
=========================================
Introduction
------------
The kernel crypto API offers a rich set of cryptographic ciphers as well
as other data transformation mechanisms and methods to invoke these.
This document contains a description of the API and provides example
code.
To understand and properly use the kernel crypto API a brief explanation
of its structure is given. Based on the architecture, the API can be
separated into different components. Following the architecture
specification, hints to developers of ciphers are provided. Pointers to
the API function call documentation are given at the end.
The kernel crypto API refers to all algorithms as "transformations".
Therefore, a cipher handle variable usually has the name "tfm". Besides
cryptographic operations, the kernel crypto API also knows compression
transformations and handles them the same way as ciphers.
The kernel crypto API serves the following entity types:
- consumers requesting cryptographic services
- data transformation implementations (typically ciphers) that can be
called by consumers using the kernel crypto API
This specification is intended for consumers of the kernel crypto API as
well as for developers implementing ciphers. This API specification,
however, does not discuss all API calls available to data transformation
implementations (i.e. implementations of ciphers and other
transformations (such as CRC or even compression algorithms) that can
register with the kernel crypto API).
Note: The terms "transformation" and cipher algorithm are used
interchangeably.
Terminology
-----------
The transformation implementation is an actual code or interface to
hardware which implements a certain transformation with precisely
defined behavior.
The transformation object (TFM) is an instance of a transformation
implementation. There can be multiple transformation objects associated
with a single transformation implementation. Each of those
transformation objects is held by a crypto API consumer or another
transformation. Transformation object is allocated when a crypto API
consumer requests a transformation implementation. The consumer is then
provided with a structure, which contains a transformation object (TFM).
The structure that contains transformation objects may also be referred
to as a "cipher handle". Such a cipher handle is always subject to the
following phases that are reflected in the API calls applicable to such
a cipher handle:
1. Initialization of a cipher handle.
2. Execution of all intended cipher operations applicable for the handle
where the cipher handle must be furnished to every API call.
3. Destruction of a cipher handle.
When using the initialization API calls, a cipher handle is created and
returned to the consumer. Therefore, please refer to all initialization
API calls that refer to the data structure type a consumer is expected
to receive and subsequently to use. The initialization API calls have
all the same naming conventions of crypto_alloc\*.
The transformation context is private data associated with the
transformation object.
+387
View File
@@ -0,0 +1,387 @@
User Space Interface
====================
Introduction
------------
The concepts of the kernel crypto API visible to kernel space is fully
applicable to the user space interface as well. Therefore, the kernel
crypto API high level discussion for the in-kernel use cases applies
here as well.
The major difference, however, is that user space can only act as a
consumer and never as a provider of a transformation or cipher
algorithm.
The following covers the user space interface exported by the kernel
crypto API. A working example of this description is libkcapi that can
be obtained from [1]. That library can be used by user space
applications that require cryptographic services from the kernel.
Some details of the in-kernel kernel crypto API aspects do not apply to
user space, however. This includes the difference between synchronous
and asynchronous invocations. The user space API call is fully
synchronous.
[1] http://www.chronox.de/libkcapi.html
User Space API General Remarks
------------------------------
The kernel crypto API is accessible from user space. Currently, the
following ciphers are accessible:
- Message digest including keyed message digest (HMAC, CMAC)
- Symmetric ciphers
- AEAD ciphers
- Random Number Generators
The interface is provided via socket type using the type AF_ALG. In
addition, the setsockopt option type is SOL_ALG. In case the user space
header files do not export these flags yet, use the following macros:
::
#ifndef AF_ALG
#define AF_ALG 38
#endif
#ifndef SOL_ALG
#define SOL_ALG 279
#endif
A cipher is accessed with the same name as done for the in-kernel API
calls. This includes the generic vs. unique naming schema for ciphers as
well as the enforcement of priorities for generic names.
To interact with the kernel crypto API, a socket must be created by the
user space application. User space invokes the cipher operation with the
send()/write() system call family. The result of the cipher operation is
obtained with the read()/recv() system call family.
The following API calls assume that the socket descriptor is already
opened by the user space application and discusses only the kernel
crypto API specific invocations.
To initialize the socket interface, the following sequence has to be
performed by the consumer:
1. Create a socket of type AF_ALG with the struct sockaddr_alg
parameter specified below for the different cipher types.
2. Invoke bind with the socket descriptor
3. Invoke accept with the socket descriptor. The accept system call
returns a new file descriptor that is to be used to interact with the
particular cipher instance. When invoking send/write or recv/read
system calls to send data to the kernel or obtain data from the
kernel, the file descriptor returned by accept must be used.
In-place Cipher operation
-------------------------
Just like the in-kernel operation of the kernel crypto API, the user
space interface allows the cipher operation in-place. That means that
the input buffer used for the send/write system call and the output
buffer used by the read/recv system call may be one and the same. This
is of particular interest for symmetric cipher operations where a
copying of the output data to its final destination can be avoided.
If a consumer on the other hand wants to maintain the plaintext and the
ciphertext in different memory locations, all a consumer needs to do is
to provide different memory pointers for the encryption and decryption
operation.
Message Digest API
------------------
The message digest type to be used for the cipher operation is selected
when invoking the bind syscall. bind requires the caller to provide a
filled struct sockaddr data structure. This data structure must be
filled as follows:
::
struct sockaddr_alg sa = {
.salg_family = AF_ALG,
.salg_type = "hash", /* this selects the hash logic in the kernel */
.salg_name = "sha1" /* this is the cipher name */
};
The salg_type value "hash" applies to message digests and keyed message
digests. Though, a keyed message digest is referenced by the appropriate
salg_name. Please see below for the setsockopt interface that explains
how the key can be set for a keyed message digest.
Using the send() system call, the application provides the data that
should be processed with the message digest. The send system call allows
the following flags to be specified:
- MSG_MORE: If this flag is set, the send system call acts like a
message digest update function where the final hash is not yet
calculated. If the flag is not set, the send system call calculates
the final message digest immediately.
With the recv() system call, the application can read the message digest
from the kernel crypto API. If the buffer is too small for the message
digest, the flag MSG_TRUNC is set by the kernel.
In order to set a message digest key, the calling application must use
the setsockopt() option of ALG_SET_KEY. If the key is not set the HMAC
operation is performed without the initial HMAC state change caused by
the key.
Symmetric Cipher API
--------------------
The operation is very similar to the message digest discussion. During
initialization, the struct sockaddr data structure must be filled as
follows:
::
struct sockaddr_alg sa = {
.salg_family = AF_ALG,
.salg_type = "skcipher", /* this selects the symmetric cipher */
.salg_name = "cbc(aes)" /* this is the cipher name */
};
Before data can be sent to the kernel using the write/send system call
family, the consumer must set the key. The key setting is described with
the setsockopt invocation below.
Using the sendmsg() system call, the application provides the data that
should be processed for encryption or decryption. In addition, the IV is
specified with the data structure provided by the sendmsg() system call.
The sendmsg system call parameter of struct msghdr is embedded into the
struct cmsghdr data structure. See recv(2) and cmsg(3) for more
information on how the cmsghdr data structure is used together with the
send/recv system call family. That cmsghdr data structure holds the
following information specified with a separate header instances:
- specification of the cipher operation type with one of these flags:
- ALG_OP_ENCRYPT - encryption of data
- ALG_OP_DECRYPT - decryption of data
- specification of the IV information marked with the flag ALG_SET_IV
The send system call family allows the following flag to be specified:
- MSG_MORE: If this flag is set, the send system call acts like a
cipher update function where more input data is expected with a
subsequent invocation of the send system call.
Note: The kernel reports -EINVAL for any unexpected data. The caller
must make sure that all data matches the constraints given in
/proc/crypto for the selected cipher.
With the recv() system call, the application can read the result of the
cipher operation from the kernel crypto API. The output buffer must be
at least as large as to hold all blocks of the encrypted or decrypted
data. If the output data size is smaller, only as many blocks are
returned that fit into that output buffer size.
AEAD Cipher API
---------------
The operation is very similar to the symmetric cipher discussion. During
initialization, the struct sockaddr data structure must be filled as
follows:
::
struct sockaddr_alg sa = {
.salg_family = AF_ALG,
.salg_type = "aead", /* this selects the symmetric cipher */
.salg_name = "gcm(aes)" /* this is the cipher name */
};
Before data can be sent to the kernel using the write/send system call
family, the consumer must set the key. The key setting is described with
the setsockopt invocation below.
In addition, before data can be sent to the kernel using the write/send
system call family, the consumer must set the authentication tag size.
To set the authentication tag size, the caller must use the setsockopt
invocation described below.
Using the sendmsg() system call, the application provides the data that
should be processed for encryption or decryption. In addition, the IV is
specified with the data structure provided by the sendmsg() system call.
The sendmsg system call parameter of struct msghdr is embedded into the
struct cmsghdr data structure. See recv(2) and cmsg(3) for more
information on how the cmsghdr data structure is used together with the
send/recv system call family. That cmsghdr data structure holds the
following information specified with a separate header instances:
- specification of the cipher operation type with one of these flags:
- ALG_OP_ENCRYPT - encryption of data
- ALG_OP_DECRYPT - decryption of data
- specification of the IV information marked with the flag ALG_SET_IV
- specification of the associated authentication data (AAD) with the
flag ALG_SET_AEAD_ASSOCLEN. The AAD is sent to the kernel together
with the plaintext / ciphertext. See below for the memory structure.
The send system call family allows the following flag to be specified:
- MSG_MORE: If this flag is set, the send system call acts like a
cipher update function where more input data is expected with a
subsequent invocation of the send system call.
Note: The kernel reports -EINVAL for any unexpected data. The caller
must make sure that all data matches the constraints given in
/proc/crypto for the selected cipher.
With the recv() system call, the application can read the result of the
cipher operation from the kernel crypto API. The output buffer must be
at least as large as defined with the memory structure below. If the
output data size is smaller, the cipher operation is not performed.
The authenticated decryption operation may indicate an integrity error.
Such breach in integrity is marked with the -EBADMSG error code.
AEAD Memory Structure
~~~~~~~~~~~~~~~~~~~~~
The AEAD cipher operates with the following information that is
communicated between user and kernel space as one data stream:
- plaintext or ciphertext
- associated authentication data (AAD)
- authentication tag
The sizes of the AAD and the authentication tag are provided with the
sendmsg and setsockopt calls (see there). As the kernel knows the size
of the entire data stream, the kernel is now able to calculate the right
offsets of the data components in the data stream.
The user space caller must arrange the aforementioned information in the
following order:
- AEAD encryption input: AAD \|\| plaintext
- AEAD decryption input: AAD \|\| ciphertext \|\| authentication tag
The output buffer the user space caller provides must be at least as
large to hold the following data:
- AEAD encryption output: ciphertext \|\| authentication tag
- AEAD decryption output: plaintext
Random Number Generator API
---------------------------
Again, the operation is very similar to the other APIs. During
initialization, the struct sockaddr data structure must be filled as
follows:
::
struct sockaddr_alg sa = {
.salg_family = AF_ALG,
.salg_type = "rng", /* this selects the symmetric cipher */
.salg_name = "drbg_nopr_sha256" /* this is the cipher name */
};
Depending on the RNG type, the RNG must be seeded. The seed is provided
using the setsockopt interface to set the key. For example, the
ansi_cprng requires a seed. The DRBGs do not require a seed, but may be
seeded.
Using the read()/recvmsg() system calls, random numbers can be obtained.
The kernel generates at most 128 bytes in one call. If user space
requires more data, multiple calls to read()/recvmsg() must be made.
WARNING: The user space caller may invoke the initially mentioned accept
system call multiple times. In this case, the returned file descriptors
have the same state.
Zero-Copy Interface
-------------------
In addition to the send/write/read/recv system call family, the AF_ALG
interface can be accessed with the zero-copy interface of
splice/vmsplice. As the name indicates, the kernel tries to avoid a copy
operation into kernel space.
The zero-copy operation requires data to be aligned at the page
boundary. Non-aligned data can be used as well, but may require more
operations of the kernel which would defeat the speed gains obtained
from the zero-copy interface.
The system-inherent limit for the size of one zero-copy operation is 16
pages. If more data is to be sent to AF_ALG, user space must slice the
input into segments with a maximum size of 16 pages.
Zero-copy can be used with the following code example (a complete
working example is provided with libkcapi):
::
int pipes[2];
pipe(pipes);
/* input data in iov */
vmsplice(pipes[1], iov, iovlen, SPLICE_F_GIFT);
/* opfd is the file descriptor returned from accept() system call */
splice(pipes[0], NULL, opfd, NULL, ret, 0);
read(opfd, out, outlen);
Setsockopt Interface
--------------------
In addition to the read/recv and send/write system call handling to send
and retrieve data subject to the cipher operation, a consumer also needs
to set the additional information for the cipher operation. This
additional information is set using the setsockopt system call that must
be invoked with the file descriptor of the open cipher (i.e. the file
descriptor returned by the accept system call).
Each setsockopt invocation must use the level SOL_ALG.
The setsockopt interface allows setting the following data using the
mentioned optname:
- ALG_SET_KEY -- Setting the key. Key setting is applicable to:
- the skcipher cipher type (symmetric ciphers)
- the hash cipher type (keyed message digests)
- the AEAD cipher type
- the RNG cipher type to provide the seed
- ALG_SET_AEAD_AUTHSIZE -- Setting the authentication tag size for
AEAD ciphers. For a encryption operation, the authentication tag of
the given size will be generated. For a decryption operation, the
provided ciphertext is assumed to contain an authentication tag of
the given size (see section about AEAD memory layout below).
User space API example
----------------------
Please see [1] for libkcapi which provides an easy-to-use wrapper around
the aforementioned Netlink kernel interface. [1] also contains a test
application that invokes all libkcapi API calls.
[1] http://www.chronox.de/libkcapi.html