Mark sm4 and missing aes using protected keys which are indetical to
same algs with no HW protected keys as tested.
Signed-off-by: Gilad Ben-Yossef <gilad@benyossef.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The flags field in 'struct shash_desc' never actually does anything.
The only ostensibly supported flag is CRYPTO_TFM_REQ_MAY_SLEEP.
However, no shash algorithm ever sleeps, making this flag a no-op.
With this being the case, inevitably some users who can't sleep wrongly
pass MAY_SLEEP. These would all need to be fixed if any shash algorithm
actually started sleeping. For example, the shash_ahash_*() functions,
which wrap a shash algorithm with the ahash API, pass through MAY_SLEEP
from the ahash API to the shash API. However, the shash functions are
called under kmap_atomic(), so actually they're assumed to never sleep.
Even if it turns out that some users do need preemption points while
hashing large buffers, we could easily provide a helper function
crypto_shash_update_large() which divides the data into smaller chunks
and calls crypto_shash_update() and cond_resched() for each chunk. It's
not necessary to have a flag in 'struct shash_desc', nor is it necessary
to make individual shash algorithms aware of this at all.
Therefore, remove shash_desc::flags, and document that the
crypto_shash_*() functions can be called from any context.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
When the extra crypto self-tests are enabled, test each AEAD algorithm
against its generic implementation when one is available. This
involves: checking the algorithm properties for consistency, then
randomly generating test vectors using the generic implementation and
running them against the implementation under test. Both good and bad
inputs are tested.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
When the extra crypto self-tests are enabled, test each skcipher
algorithm against its generic implementation when one is available.
This involves: checking the algorithm properties for consistency, then
randomly generating test vectors using the generic implementation and
running them against the implementation under test. Both good and bad
inputs are tested.
This has already detected a bug in the skcipher_walk API, a bug in the
LRW template, and an inconsistency in the cts implementations.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
When the extra crypto self-tests are enabled, test each hash algorithm
against its generic implementation when one is available. This
involves: checking the algorithm properties for consistency, then
randomly generating test vectors using the generic implementation and
running them against the implementation under test. Both good and bad
inputs are tested.
This has already detected a bug in the x86 implementation of poly1305,
bugs in crct10dif, and an inconsistency in cbcmac.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Add some helper functions in preparation for fuzz testing algorithms
against their generic implementation.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
In preparation for fuzz testing algorithms against their generic
implementation, make error messages in testmgr identify test vectors by
name rather than index. Built-in test vectors are simply "named" by
their index in testmgr.h, as before. But (in later patches) generated
test vectors will be given more descriptive names to help developers
debug problems detected with them.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Update testmgr to support testing for specific errors from setkey() and
digest() for hashes; setkey() and encrypt()/decrypt() for skciphers and
ciphers; and setkey(), setauthsize(), and encrypt()/decrypt() for AEADs.
This is useful because algorithms usually restrict the lengths or format
of the message, key, and/or authentication tag in some way. And bad
inputs should be tested too, not just good inputs.
As part of this change, remove the ambiguously-named 'fail' flag and
replace it with 'setkey_error = -EINVAL' for the only test vector that
used it -- the DES weak key test vector. Note that this tightens the
test to require -EINVAL rather than any error code, but AFAICS this
won't cause any test failure.
Other than that, these new fields aren't set on any test vectors yet.
Later patches will do so.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Add testmgr test vectors for EC-RDSA algorithm for every of five
supported parameters (curves). Because there are no officially published
test vectors for the curves, the vectors are generated by gost-engine.
Signed-off-by: Vitaly Chikunov <vt@altlinux.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Some public key algorithms (like EC-DSA) keep in parameters field
important data such as digest and curve OIDs (possibly more for
different EC-DSA variants). Thus, just setting a public key (as
for RSA) is not enough.
Append parameters into the key stream for akcipher_set_{pub,priv}_key.
Appended data is: (u32) algo OID, (u32) parameters length, parameters
data.
This does not affect current akcipher API nor RSA ciphers (they could
ignore it). Idea of appending parameters to the key stream is by Herbert
Xu.
Cc: David Howells <dhowells@redhat.com>
Cc: Denis Kenzior <denkenz@gmail.com>
Cc: keyrings@vger.kernel.org
Signed-off-by: Vitaly Chikunov <vt@altlinux.org>
Reviewed-by: Denis Kenzior <denkenz@gmail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Previous akcipher .verify() just `decrypts' (using RSA encrypt which is
using public key) signature to uncover message hash, which was then
compared in upper level public_key_verify_signature() with the expected
hash value, which itself was never passed into verify().
This approach was incompatible with EC-DSA family of algorithms,
because, to verify a signature EC-DSA algorithm also needs a hash value
as input; then it's used (together with a signature divided into halves
`r||s') to produce a witness value, which is then compared with `r' to
determine if the signature is correct. Thus, for EC-DSA, nor
requirements of .verify() itself, nor its output expectations in
public_key_verify_signature() wasn't sufficient.
Make improved .verify() call which gets hash value as input and produce
complete signature check without any output besides status.
Now for the top level verification only crypto_akcipher_verify() needs
to be called and its return value inspected.
Make sure that `digest' is in kmalloc'd memory (in place of `output`) in
{public,tpm}_key_verify_signature() as insisted by Herbert Xu, and will
be changed in the following commit.
Cc: David Howells <dhowells@redhat.com>
Cc: keyrings@vger.kernel.org
Signed-off-by: Vitaly Chikunov <vt@altlinux.org>
Reviewed-by: Denis Kenzior <denkenz@gmail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Add a module parameter cryptomgr.panic_on_fail which causes the kernel
to panic if any crypto self-tests fail.
Use cases:
- More easily detect crypto self-test failures by boot testing,
e.g. on KernelCI.
- Get a bug report if syzkaller manages to use the template system to
instantiate an algorithm that fails its self-tests.
The command-line option "fips=1" already does this, but it also makes
other changes not wanted for general testing, such as disabling
"unapproved" algorithms. panic_on_fail just does what it says.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
All crypto API algorithms are supposed to support the case where they
are called in a context where SIMD instructions are unusable, e.g. IRQ
context on some architectures. However, this isn't tested for by the
self-tests, causing bugs to go undetected.
Now that all algorithms have been converted to use crypto_simd_usable(),
update the self-tests to test the no-SIMD case. First, a bool
testvec_config::nosimd is added. When set, the crypto operation is
executed with preemption disabled and with crypto_simd_usable() mocked
out to return false on the current CPU.
A bool test_sg_division::nosimd is also added. For hash algorithms it's
honored by the corresponding ->update(). By setting just a subset of
these bools, the case where some ->update()s are done in SIMD context
and some are done in no-SIMD context is also tested.
These bools are then randomly set by generate_random_testvec_config().
For now, all no-SIMD testing is limited to the extra crypto self-tests,
because it might be a bit too invasive for the regular self-tests.
But this could be changed later.
This has already found bugs in the arm64 AES-GCM and ChaCha algorithms.
This would have found some past bugs as well.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Acked-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
So that the no-SIMD fallback code can be tested by the crypto
self-tests, add a macro crypto_simd_usable() which wraps may_use_simd(),
but also returns false if the crypto self-tests have set a per-CPU bool
to disable SIMD in crypto code on the current CPU.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Now that all AEAD algorithms (that I have the hardware to test, at
least) have been fixed to not modify the user-provided aead_request,
remove the workaround from testmgr that reset aead_request::tfm after
each AEAD encryption/decryption.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Allow skcipher test vectors to declare the value the IV buffer should be
updated to at the end of the encryption or decryption operation.
(This check actually used to be supported in testmgr, but it was never
used and therefore got removed except for the AES-Keywrap special case.
But it will be used by CBC and CTR now, so re-add it.)
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Check that algorithms do not change the aead_request structure, as users
may rely on submitting the request again (e.g. after copying new data
into the same source buffer) without reinitializing everything.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Check that algorithms do not change the skcipher_request structure, as
users may rely on submitting the request again (e.g. after copying new
data into the same source buffer) without reinitializing everything.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Convert alg_test_hash() to use the new test framework, adding a list of
testvec_configs to test by default. When the extra self-tests are
enabled, randomly generated testvec_configs are tested as well.
This improves hash test coverage mainly because now all algorithms have
a variety of data layouts tested, whereas before each algorithm was
responsible for declaring its own chunked test cases which were often
missing or provided poor test coverage. The new code also tests both
the MAY_SLEEP and !MAY_SLEEP cases and buffers that cross pages.
This already found bugs in the hash walk code and in the arm32 and arm64
implementations of crct10dif.
I removed the hash chunked test vectors that were the same as
non-chunked ones, but left the ones that were unique.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Convert alg_test_aead() to use the new test framework, using the same
list of testvec_configs that skcipher testing uses.
This significantly improves AEAD test coverage mainly because previously
there was only very limited test coverage of the possible data layouts.
Now the data layouts to test are listed in one place for all algorithms
and optionally are also randomly generated. In fact, only one AEAD
algorithm (AES-GCM) even had a chunked test case before.
This already found bugs in all the AEGIS and MORUS implementations, the
x86 AES-GCM implementation, and the arm64 AES-CCM implementation.
I removed the AEAD chunked test vectors that were the same as
non-chunked ones, but left the ones that were unique.
Note: the rewritten test code allocates an aead_request just once per
algorithm rather than once per encryption/decryption, but some AEAD
algorithms incorrectly change the tfm pointer in the request. It's
nontrivial to fix these, so to move forward I'm temporarily working
around it by resetting the tfm pointer. But they'll need to be fixed.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Convert alg_test_skcipher() to use the new test framework, adding a list
of testvec_configs to test by default. When the extra self-tests are
enabled, randomly generated testvec_configs are tested as well.
This improves skcipher test coverage mainly because now all algorithms
have a variety of data layouts tested, whereas before each algorithm was
responsible for declaring its own chunked test cases which were often
missing or provided poor test coverage. The new code also tests both
the MAY_SLEEP and !MAY_SLEEP cases, different IV alignments, and buffers
that cross pages.
This has already found a bug in the arm64 ctr-aes-neonbs algorithm.
It would have easily found many past bugs.
I removed the skcipher chunked test vectors that were the same as
non-chunked ones, but left the ones that were unique.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Add functions that generate a random testvec_config, in preparation for
using it for randomized fuzz tests.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
To achieve more comprehensive crypto test coverage, I'd like to add fuzz
tests that use random data layouts and request flags.
To be most effective these tests should be part of testmgr, so they
automatically run on every algorithm registered with the crypto API.
However, they will take much longer to run than the current tests and
therefore will only really be intended to be run by developers, whereas
the current tests have a wider audience.
Therefore, add a new kconfig option CONFIG_CRYPTO_MANAGER_EXTRA_TESTS
that can be set by developers to enable these extra, expensive tests.
Similar to the regular tests, also add a module parameter
cryptomgr.noextratests to support disabling the tests.
Finally, another module parameter cryptomgr.fuzz_iterations is added to
control how many iterations the fuzz tests do. Note: for now setting
this to 0 will be equivalent to cryptomgr.noextratests=1. But I opted
for separate parameters to provide more flexibility to add other types
of tests under the "extra tests" category in the future.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Crypto algorithms must produce the same output for the same input
regardless of data layout, i.e. how the src and dst scatterlists are
divided into chunks and how each chunk is aligned. Request flags such
as CRYPTO_TFM_REQ_MAY_SLEEP must not affect the result either.
However, testing of this currently has many gaps. For example,
individual algorithms are responsible for providing their own chunked
test vectors. But many don't bother to do this or test only one or two
cases, providing poor test coverage. Also, other things such as
misaligned IVs and CRYPTO_TFM_REQ_MAY_SLEEP are never tested at all.
Test code is also duplicated between the chunked and non-chunked cases,
making it difficult to make other improvements.
To improve the situation, this patch series basically moves the chunk
descriptions into the testmgr itself so that they are shared by all
algorithms. However, it's done in an extensible way via a new struct
'testvec_config', which describes not just the scaled chunk lengths but
also all other aspects of the crypto operation besides the data itself
such as the buffer alignments, the request flags, whether the operation
is in-place or not, the IV alignment, and for hash algorithms when to
do each update() and when to use finup() vs. final() vs. digest().
Then, this patch series makes skcipher, aead, and hash algorithms be
tested against a list of default testvec_configs, replacing the current
test code. This improves overall test coverage, without reducing test
performance too much. Note that the test vectors themselves are not
changed, except for removing the chunk lists.
This series also adds randomized fuzz tests, enabled by a new kconfig
option intended for developer use only, where skcipher, aead, and hash
algorithms are tested against many randomly generated testvec_configs.
This provides much more comprehensive test coverage.
These improved tests have already exposed many bugs.
To start it off, this initial patch adds the testvec_config and various
helper functions that will be used by the skcipher, aead, and hash test
code that will be converted to use the new testvec_config framework.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The CRC32 is not a cryptographic hash algorithm,
so the FIPS restrictions should not apply to it.
(The CRC32C variant is already allowed.)
This CRC32 variant is used for in dm-crypt legacy TrueCrypt
IV implementation (tcw); detected by cryptsetup test suite
failure in FIPS mode.
Signed-off-by: Milan Broz <gmazyland@gmail.com>
Reviewed-by: Stephan Mueller <smueller@chronox.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Instantiating "cryptd(crc32c)" causes a crypto self-test failure because
the crypto_alloc_shash() in alg_test_crc32c() fails. This is because
cryptd(crc32c) is an ahash algorithm, not a shash algorithm; so it can
only be accessed through the ahash API, unlike shash algorithms which
can be accessed through both the ahash and shash APIs.
As the test is testing the shash descriptor format which is only
applicable to shash algorithms, skip it for ahash algorithms.
(Note that it's still important to fix crypto self-test failures even
for weird algorithm instantiations like cryptd(crc32c) that no one
would really use; in fips_enabled mode unprivileged users can use them
to panic the kernel, and also they prevent treating a crypto self-test
failure as a bug when fuzzing the kernel.)
Fixes: 8e3ee85e68 ("crypto: crc32c - Test descriptor context format")
Cc: stable@vger.kernel.org
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
CRYPTO_TFM_REQ_WEAK_KEY confuses newcomers to the crypto API because it
sounds like it is requesting a weak key. Actually, it is requesting
that weak keys be forbidden (for algorithms that have the notion of
"weak keys"; currently only DES and XTS do).
Also it is only one letter away from CRYPTO_TFM_RES_WEAK_KEY, with which
it can be easily confused. (This in fact happened in the UX500 driver,
though just in some debugging messages.)
Therefore, make the intent clear by renaming it to
CRYPTO_TFM_REQ_FORBID_WEAK_KEYS.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Currently testmgr has separate encryption and decryption test vectors
for AEADs. That's massively redundant, since usually the decryption
tests are identical to the encryption tests, just with the input/result
swapped. And for some algorithms it was forgotten to add decryption
test vectors, so for them currently only encryption is being tested.
Therefore, eliminate the redundancy by removing the AEAD decryption test
vectors and updating testmgr to test both AEAD encryption and decryption
using what used to be the encryption test vectors. Naming is adjusted
accordingly: each aead_testvec now has a 'ptext' (plaintext), 'plen'
(plaintext length), 'ctext' (ciphertext), and 'clen' (ciphertext length)
instead of an 'input', 'ilen', 'result', and 'rlen'. "Ciphertext" here
refers to the full ciphertext, including the authentication tag.
For now the scatterlist divisions are just given for the plaintext
length, not also the ciphertext length. For decryption, the last
scatterlist element is just extended by the authentication tag length.
In total, this removes over 5000 lines from testmgr.h, with no reduction
in test coverage since prior patches already copied the few unique
decryption test vectors into the encryption test vectors.
The testmgr.h portion of this patch was automatically generated using
the following awk script, except that I also manually updated the
definition of 'struct aead_testvec' and fixed the location of the
comment describing the AEGIS-128 test vectors.
BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER }
/^static const struct aead_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC }
/^static const struct aead_testvec.*_dec_/ { mode = DECVEC }
mode == ENCVEC {
sub(/\.input[[:space:]]*=/, ".ptext\t=")
sub(/\.result[[:space:]]*=/, ".ctext\t=")
sub(/\.ilen[[:space:]]*=/, ".plen\t=")
sub(/\.rlen[[:space:]]*=/, ".clen\t=")
print
}
mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER }
mode == OTHER { print }
mode == ENCVEC && /^};/ { mode = OTHER }
mode == DECVEC && /^};/ { mode = DECVEC_TAIL }
Note that git's default diff algorithm gets confused by the testmgr.h
portion of this patch, and reports too many lines added and removed.
It's better viewed with 'git diff --minimal' (or 'git show --minimal'),
which reports "2 files changed, 1235 insertions(+), 6491 deletions(-)".
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
In preparation for unifying the AEAD encryption and decryption test
vectors, skip AEAD test vectors with the 'novrfy' (verification failure
expected) flag set when testing encryption rather than decryption.
These test vectors only make sense for decryption.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The crc32c context is in CPU endianness, whereas the final digest is
little endian. alg_test_crc32c() got this mixed up. Fix it.
The test passes both before and after, but this patch fixes the
following sparse warning:
crypto/testmgr.c:1912:24: warning: cast to restricted __le32
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Before this, if akcipher_testvec have `public_key_vec' set to true
(i.e. having a public key) only sign/encrypt test is performed, but
verify/decrypt test is skipped.
With a public key we could do encrypt and verify, but to sign and decrypt
a private key is required.
This logic is correct for encrypt/decrypt tests (decrypt is skipped if
no private key). But incorrect for sign/verify tests - sign is performed
no matter if there is no private key, but verify is skipped if there is
a public key.
Rework `test_akcipher_one' to arrange tests properly depending on value
of `public_key_vec` and `siggen_sigver_test'.
No tests were missed since there is only one sign/verify test (which
have `siggen_sigver_test' set to true) and it has a private key, but
future tests could benefit from this improvement.
Signed-off-by: Vitaly Chikunov <vt@altlinux.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Add support for the Adiantum encryption mode. Adiantum was designed by
Paul Crowley and is specified by our paper:
Adiantum: length-preserving encryption for entry-level processors
(https://eprint.iacr.org/2018/720.pdf)
See our paper for full details; this patch only provides an overview.
Adiantum is a tweakable, length-preserving encryption mode designed for
fast and secure disk encryption, especially on CPUs without dedicated
crypto instructions. Adiantum encrypts each sector using the XChaCha12
stream cipher, two passes of an ε-almost-∆-universal (εA∆U) hash
function, and an invocation of the AES-256 block cipher on a single
16-byte block. On CPUs without AES instructions, Adiantum is much
faster than AES-XTS; for example, on ARM Cortex-A7, on 4096-byte sectors
Adiantum encryption is about 4 times faster than AES-256-XTS encryption,
and decryption about 5 times faster.
Adiantum is a specialization of the more general HBSH construction. Our
earlier proposal, HPolyC, was also a HBSH specialization, but it used a
different εA∆U hash function, one based on Poly1305 only. Adiantum's
εA∆U hash function, which is based primarily on the "NH" hash function
like that used in UMAC (RFC4418), is about twice as fast as HPolyC's;
consequently, Adiantum is about 20% faster than HPolyC.
This speed comes with no loss of security: Adiantum is provably just as
secure as HPolyC, in fact slightly *more* secure. Like HPolyC,
Adiantum's security is reducible to that of XChaCha12 and AES-256,
subject to a security bound. XChaCha12 itself has a security reduction
to ChaCha12. Therefore, one need not "trust" Adiantum; one need only
trust ChaCha12 and AES-256. Note that the εA∆U hash function is only
used for its proven combinatorical properties so cannot be "broken".
Adiantum is also a true wide-block encryption mode, so flipping any
plaintext bit in the sector scrambles the entire ciphertext, and vice
versa. No other such mode is available in the kernel currently; doing
the same with XTS scrambles only 16 bytes. Adiantum also supports
arbitrary-length tweaks and naturally supports any length input >= 16
bytes without needing "ciphertext stealing".
For the stream cipher, Adiantum uses XChaCha12 rather than XChaCha20 in
order to make encryption feasible on the widest range of devices.
Although the 20-round variant is quite popular, the best known attacks
on ChaCha are on only 7 rounds, so ChaCha12 still has a substantial
security margin; in fact, larger than AES-256's. 12-round Salsa20 is
also the eSTREAM recommendation. For the block cipher, Adiantum uses
AES-256, despite it having a lower security margin than XChaCha12 and
needing table lookups, due to AES's extensive adoption and analysis
making it the obvious first choice. Nevertheless, for flexibility this
patch also permits the "adiantum" template to be instantiated with
XChaCha20 and/or with an alternate block cipher.
We need Adiantum support in the kernel for use in dm-crypt and fscrypt,
where currently the only other suitable options are block cipher modes
such as AES-XTS. A big problem with this is that many low-end mobile
devices (e.g. Android Go phones sold primarily in developing countries,
as well as some smartwatches) still have CPUs that lack AES
instructions, e.g. ARM Cortex-A7. Sadly, AES-XTS encryption is much too
slow to be viable on these devices. We did find that some "lightweight"
block ciphers are fast enough, but these suffer from problems such as
not having much cryptanalysis or being too controversial.
The ChaCha stream cipher has excellent performance but is insecure to
use directly for disk encryption, since each sector's IV is reused each
time it is overwritten. Even restricting the threat model to offline
attacks only isn't enough, since modern flash storage devices don't
guarantee that "overwrites" are really overwrites, due to wear-leveling.
Adiantum avoids this problem by constructing a
"tweakable super-pseudorandom permutation"; this is the strongest
possible security model for length-preserving encryption.
Of course, storing random nonces along with the ciphertext would be the
ideal solution. But doing that with existing hardware and filesystems
runs into major practical problems; in most cases it would require data
journaling (like dm-integrity) which severely degrades performance.
Thus, for now length-preserving encryption is still needed.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Add a generic implementation of NHPoly1305, an ε-almost-∆-universal hash
function used in the Adiantum encryption mode.
CONFIG_NHPOLY1305 is not selectable by itself since there won't be any
real reason to enable it without also enabling Adiantum support.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Acked-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Now that the generic implementation of ChaCha20 has been refactored to
allow varying the number of rounds, add support for XChaCha12, which is
the XSalsa construction applied to ChaCha12. ChaCha12 is one of the
three ciphers specified by the original ChaCha paper
(https://cr.yp.to/chacha/chacha-20080128.pdf: "ChaCha, a variant of
Salsa20"), alongside ChaCha8 and ChaCha20. ChaCha12 is faster than
ChaCha20 but has a lower, but still large, security margin.
We need XChaCha12 support so that it can be used in the Adiantum
encryption mode, which enables disk/file encryption on low-end mobile
devices where AES-XTS is too slow as the CPUs lack AES instructions.
We'd prefer XChaCha20 (the more popular variant), but it's too slow on
some of our target devices, so at least in some cases we do need the
XChaCha12-based version. In more detail, the problem is that Adiantum
is still much slower than we're happy with, and encryption still has a
quite noticeable effect on the feel of low-end devices. Users and
vendors push back hard against encryption that degrades the user
experience, which always risks encryption being disabled entirely. So
we need to choose the fastest option that gives us a solid margin of
security, and here that's XChaCha12. The best known attack on ChaCha
breaks only 7 rounds and has 2^235 time complexity, so ChaCha12's
security margin is still better than AES-256's. Much has been learned
about cryptanalysis of ARX ciphers since Salsa20 was originally designed
in 2005, and it now seems we can be comfortable with a smaller number of
rounds. The eSTREAM project also suggests the 12-round version of
Salsa20 as providing the best balance among the different variants:
combining very good performance with a "comfortable margin of security".
Note that it would be trivial to add vanilla ChaCha12 in addition to
XChaCha12. However, it's unneeded for now and therefore is omitted.
As discussed in the patch that introduced XChaCha20 support, I
considered splitting the code into separate chacha-common, chacha20,
xchacha20, and xchacha12 modules, so that these algorithms could be
enabled/disabled independently. However, since nearly all the code is
shared anyway, I ultimately decided there would have been little benefit
to the added complexity.
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Acked-by: Martin Willi <martin@strongswan.org>
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Add support for the XChaCha20 stream cipher. XChaCha20 is the
application of the XSalsa20 construction
(https://cr.yp.to/snuffle/xsalsa-20081128.pdf) to ChaCha20 rather than
to Salsa20. XChaCha20 extends ChaCha20's nonce length from 64 bits (or
96 bits, depending on convention) to 192 bits, while provably retaining
ChaCha20's security. XChaCha20 uses the ChaCha20 permutation to map the
key and first 128 nonce bits to a 256-bit subkey. Then, it does the
ChaCha20 stream cipher with the subkey and remaining 64 bits of nonce.
We need XChaCha support in order to add support for the Adiantum
encryption mode. Note that to meet our performance requirements, we
actually plan to primarily use the variant XChaCha12. But we believe
it's wise to first add XChaCha20 as a baseline with a higher security
margin, in case there are any situations where it can be used.
Supporting both variants is straightforward.
Since XChaCha20's subkey differs for each request, XChaCha20 can't be a
template that wraps ChaCha20; that would require re-keying the
underlying ChaCha20 for every request, which wouldn't be thread-safe.
Instead, we make XChaCha20 its own top-level algorithm which calls the
ChaCha20 streaming implementation internally.
Similar to the existing ChaCha20 implementation, we define the IV to be
the nonce and stream position concatenated together. This allows users
to seek to any position in the stream.
I considered splitting the code into separate chacha20-common, chacha20,
and xchacha20 modules, so that chacha20 and xchacha20 could be
enabled/disabled independently. However, since nearly all the code is
shared anyway, I ultimately decided there would have been little benefit
to the added complexity of separate modules.
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Acked-by: Martin Willi <martin@strongswan.org>
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Add testmgr and tcrypt tests and vectors for Streebog hash function
from RFC 6986 and GOST R 34.11-2012, for HMAC-Streebog vectors are
from RFC 7836 and R 50.1.113-2016.
Cc: linux-integrity@vger.kernel.org
Signed-off-by: Vitaly Chikunov <vt@altlinux.org>
Acked-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
As per Sp800-38A addendum from Oct 2010[1], cts(cbc(aes)) is
allowed as a FIPS mode algorithm. Mark it as such.
[1] https://csrc.nist.gov/publications/detail/sp/800-38a/addendum/final
Signed-off-by: Gilad Ben-Yossef <gilad@benyossef.com>
Reviewed-by: Stephan Mueller <smueller@chronox.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
After allocation, output and decomp_output both point to memory chunks of
size COMP_BUF_SIZE. Then, only the first bytes are zeroed out using
sizeof(COMP_BUF_SIZE) as parameter to memset(), because
sizeof(COMP_BUF_SIZE) provides the size of the constant and not the size of
allocated memory.
Instead, the whole allocated memory is meant to be zeroed out. Use
COMP_BUF_SIZE as parameter to memset() directly in order to accomplish
this.
Fixes: 336073840a ("crypto: testmgr - Allow different compression results")
Signed-off-by: Michael Schupikov <michael@schupikov.de>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Add additional test vectors from "The SM4 Blockcipher Algorithm And Its
Modes Of Operations" draft-ribose-cfrg-sm4-10 and register cipher speed
tests for sm4.
Signed-off-by: Gilad Ben-Yossef <gilad@benyossef.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
These are unused, undesired, and have never actually been used by
anybody. The original authors of this code have changed their mind about
its inclusion. While originally proposed for disk encryption on low-end
devices, the idea was discarded [1] in favor of something else before
that could really get going. Therefore, this patch removes Speck.
[1] https://marc.info/?l=linux-crypto-vger&m=153359499015659
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Acked-by: Eric Biggers <ebiggers@google.com>
Cc: stable@vger.kernel.org
Acked-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The testmgr hash tests were testing init, digest, update and final
methods but not the finup method. Add a test for this one too.
While doing this, make sure we only run the partial tests once with
the digest tests and skip them with the final and finup tests since
they are the same.
Signed-off-by: Gilad Ben-Yossef <gilad@benyossef.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Remove the original version of the VMAC template that had the nonce
hardcoded to 0 and produced a digest with the wrong endianness. I'm
unsure whether this had users or not (there are no explicit in-kernel
references to it), but given that the hardcoded nonce made it wildly
insecure unless a unique key was used for each message, let's try
removing it and see if anyone complains.
Leave the new "vmac64" template that requires the nonce to be explicitly
specified as the first 16 bytes of data and uses the correct endianness
for the digest.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Currently the VMAC template uses a "nonce" hardcoded to 0, which makes
it insecure unless a unique key is set for every message. Also, the
endianness of the final digest is wrong: the implementation uses little
endian, but the VMAC specification has it as big endian, as do other
VMAC implementations such as the one in Crypto++.
Add a new VMAC template where the nonce is passed as the first 16 bytes
of data (similar to what is done for Poly1305's nonce), and the digest
is big endian. Call it "vmac64", since the old name of simply "vmac"
didn't clarify whether the implementation is of VMAC-64 or of VMAC-128
(which produce 64-bit and 128-bit digests respectively); so we fix the
naming ambiguity too.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Currently testmgr has separate encryption and decryption test vectors
for symmetric ciphers. That's massively redundant, since with few
exceptions (mostly mistakes, apparently), all decryption tests are
identical to the encryption tests, just with the input/result flipped.
Therefore, eliminate the redundancy by removing the decryption test
vectors and updating testmgr to test both encryption and decryption
using what used to be the encryption test vectors. Naming is adjusted
accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext'
(ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and
'rlen'. Note that it was always the case that 'ilen == rlen'.
AES keywrap ("kw(aes)") is special because its IV is generated by the
encryption. Previously this was handled by specifying 'iv_out' for
encryption and 'iv' for decryption. To make it work cleanly with only
one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a
boolean that indicates that the IV is generated by the encryption.
In total, this removes over 10000 lines from testmgr.h, with no
reduction in test coverage since prior patches already copied the few
unique decryption test vectors into the encryption test vectors.
This covers all algorithms that used 'struct cipher_testvec', e.g. any
block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or
keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD
tests, though we probably can eliminate a similar redundancy there too.
The testmgr.h portion of this patch was automatically generated using
the following awk script, with some slight manual fixups on top (updated
'struct cipher_testvec' definition, updated a few comments, and fixed up
the AES keywrap test vectors):
BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER }
/^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC }
/^static const struct cipher_testvec.*_dec_/ { mode = DECVEC }
mode == ENCVEC && !/\.ilen[[:space:]]*=/ {
sub(/\.input[[:space:]]*=$/, ".ptext =")
sub(/\.input[[:space:]]*=/, ".ptext\t=")
sub(/\.result[[:space:]]*=$/, ".ctext =")
sub(/\.result[[:space:]]*=/, ".ctext\t=")
sub(/\.rlen[[:space:]]*=/, ".len\t=")
print
}
mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER }
mode == OTHER { print }
mode == ENCVEC && /^};/ { mode = OTHER }
mode == DECVEC && /^};/ { mode = DECVEC_TAIL }
Note that git's default diff algorithm gets confused by the testmgr.h
portion of this patch, and reports too many lines added and removed.
It's better viewed with 'git diff --minimal' (or 'git show --minimal'),
which reports "2 files changed, 919 insertions(+), 11723 deletions(-)".
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Since testmgr uses a single tfm for all tests of each hash algorithm,
once a key is set the tfm won't be unkeyed anymore. But with crc32 and
crc32c, the key is really the "default initial state" and is optional;
those algorithms should have both keyed and unkeyed test vectors, to
verify that implementations use the correct default key.
Simply listing the unkeyed test vectors first isn't guaranteed to work
yet because testmgr makes multiple passes through the test vectors.
crc32c does have an unkeyed test vector listed first currently, but it
only works by chance because the last crc32c test vector happens to use
a key that is the same as the default key.
Therefore, teach testmgr to split hash test vectors into unkeyed and
keyed sections, and do all the unkeyed ones before the keyed ones.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The Blackfin CRC driver was removed by commit 9678a8dc53 ("crypto:
bfin_crc - remove blackfin CRC driver"), but it was forgotten to remove
the corresponding "hmac(crc32)" test vectors. I see no point in keeping
them since nothing else appears to implement or use "hmac(crc32)", which
isn't an algorithm that makes sense anyway because HMAC is meant to be
used with a cryptographically secure hash function, which CRC's are not.
Thus, remove the unneeded test vectors.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch adds test vectors for MORUS-640 and MORUS-1280. The test
vectors were generated using the reference implementation from
SUPERCOP (see code comments for more details).
Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Gilad Ben-Yossef