In case of driver probe and pm resume we need to check that the cryptocell
hardware reset cycle is completed. during the reset cycle that Cryptocell
provide read only access to the APB interface which allows to verify
through the CC registers that the reset is completed. Until reset
completion we assume that any write/crypto operation is blocked.
Signed-off-by: Ofir Drang <ofir.drang@arm.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
When calling debugfs functions, there is no need to ever check the
return value. The function can work or not, but the code logic should
never do something different based on this.
Also, there is no need to store the individual debugfs file names,
especially as the whole directiry is deleted at once, so remove the
unneeded structure entirely.
Cc: "Breno Leitão" <leitao@debian.org>
Cc: Nayna Jain <nayna@linux.ibm.com>
Cc: Paulo Flabiano Smorigo <pfsmorigo@gmail.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: linux-crypto@vger.kernel.org
Cc: linuxppc-dev@lists.ozlabs.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The CIFS code uses the sync skcipher API to invoke the ecb(arc4) skcipher,
of which only a single generic C code implementation exists. This means
that going through all the trouble of using scatterlists etc buys us
very little, and we're better off just invoking the arc4 library directly.
This also reverts commit 5f4b55699a ("CIFS: Fix BUG() in calc_seckey()"),
since it is no longer necessary to allocate sec_key on the heap.
Cc: linux-cifs@vger.kernel.org
Cc: Steve French <sfrench@samba.org>
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Acked-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The MPPE code uses the sync skcipher to invoke the ecb(arc4) skcipher,
of which only a single generic C code implementation exists. This means
that going through all the trouble of using scatterlists etc buys us
very little, and we're better off just invoking the arc4 library directly.
Note that the SHA1 shash used by this driver has several accelerated
implementations for various architectures, so retaining that part does
make sense.
Cc: linux-ppp@vger.kernel.org
Cc: Paul Mackerras <paulus@samba.org>
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
There are no remaining users of the cipher implementation, and there
are no meaningful ways in which the arc4 cipher can be combined with
templates other than ECB (and the way we do provide that combination
is highly dubious to begin with).
So let's drop the arc4 cipher altogether, and only keep the ecb(arc4)
skcipher, which is used in various places in the kernel.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The crypto API abstraction is not very useful for invoking ciphers
directly, especially in the case of arc4, which only has a generic
implementation in C. So let's invoke the library code directly.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The crypto API abstraction is not very useful for invoking ciphers
directly, especially in the case of arc4, which only has a generic
implementation in C. So let's invoke the library code directly.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The WEP code in the mac80211 subsystem currently uses the crypto
API to access the arc4 (RC4) cipher, which is overly complicated,
and doesn't really have an upside in this particular case, since
ciphers are always synchronous and therefore always implemented in
software. Given that we have no accelerated software implementations
either, it is much more straightforward to invoke a generic library
interface directly.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Refactor the core rc4 handling so we can move most users to a library
interface, permitting us to drop the cipher interface entirely in a
future patch. This is part of an effort to simplify the crypto API
and improve its robustness against incorrect use.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Below commit came with a typo in the CONFIG_ symbol, leading
to a permanently reduced max key size regarless of the driver
capabilities.
Reported-by: Horia Geantă <horia.geanta@nxp.com>
Fixes: b8fbdc2bc4 ("crypto: talitos - reduce max key size for SEC1")
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Reviewed-by: Horia Geantă <horia.geanta@nxp.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The CTR code comes from OpenSSL, where it does a 32-bit counter.
The kernel has a 128-bit counter. This difference has lead to
issues.
Document it.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Modify drivers to perform skcipher IV update using the crypto engine,
instead of performing the operation in SW.
Besides being more efficient, this also fixes IV update for CTR mode.
Output HW S/G table is appended with an entry pointing to the same
IV buffer used as input (which is now mapped BIDIRECTIONAL).
AS (Algorithm State) parameter of the OPERATION command is changed
from INIFINAL to INIT in descriptors used by ctr(aes), cbc(aes).
This is needed since in case FINAL bit is set, HW skips IV updating
in the Context Register for the last data block.
Signed-off-by: Horia Geantă <horia.geanta@nxp.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Currently, conversion of SW S/G table into HW S/G layout relies on
nents returned by sg_nents_for_len(sg, len).
However this leaves the possibility of HW S/G referencing more data
then needed: since buffer length in HW S/G entries is filled using
sg_dma_len(sg), the last entry in HW S/G table might have a length
that is bigger than needed for the crypto request.
This way of S/G table conversion is fine, unless after converting a table
more entries have to be appended to the HW S/G table.
In this case, crypto engine would access data from the S/G entry having
the incorrect length, instead of advancing in the S/G table.
This situation doesn't exist, but the upcoming implementation of
IV update for skcipher algorithms needs to add a S/G entry after
req->dst S/G (corresponding to output IV).
Signed-off-by: Horia Geantă <horia.geanta@nxp.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
BCM7211 features a RNG200 hardware random number generator block, add
support for this chip by matching the chip-specific compatible string.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
BCM7211 features a RNG200 block, document its compatible string.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Use devm_platform_ioremap_resource() to simplify the code a bit.
Signed-off-by: Fabio Estevam <festevam@gmail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Use devm_platform_ioremap_resource() to simplify the code a bit.
Signed-off-by: Fabio Estevam <festevam@gmail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch corrects the SPDX License Identifier style
in header file related to Crypto Drivers for Hisilicon
SEC Engine in Hip06 and Hip07.
For C header files Documentation/process/license-rules.rst
mandates C-like comments (opposed to C source files where
C++ style should be used)
Changes made by using a script provided by Joe Perches here:
https://lkml.org/lkml/2019/2/7/46
Suggested-by: Joe Perches <joe@perches.com>
Signed-off-by: Nishad Kamdar <nishadkamdar@gmail.com>
Acked-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
One of the more common cases of allocation size calculations is finding
the size of a structure that has a zero-sized array at the end, along
with memory for some number of elements for that array. For example:
struct qat_alg_buf_list {
...
struct qat_alg_buf bufers[];
} __packed __aligned(64);
Make use of the struct_size() helper instead of an open-coded version
in order to avoid any potential type mistakes.
So, replace the following form:
sizeof(struct qat_alg_buf_list) + ((1 + n) * sizeof(struct qat_alg_buf))
with:
struct_size(bufl, bufers, n + 1)
This code was detected with the help of Coccinelle.
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Add crypto node in device tree for CAAM support.
Noteworthy is that on 7ulp the interrupt line is shared
between the two job rings.
Signed-off-by: Iuliana Prodan <iuliana.prodan@nxp.com>
Signed-off-by: Franck LENORMAND <franck.lenormand@nxp.com>
Signed-off-by: Horia Geantă <horia.geanta@nxp.com>
Acked-by: Shawn Guo <shawnguo@kernel.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch corrects the SPDX License Identifier style
in header files related to Crypto Drivers for Cavium Nitrox
family CNN55XX devices.
For C header files Documentation/process/license-rules.rst
mandates C-like comments (opposed to C source files where
C++ style should be used)
Changes made by using a script provided by Joe Perches here:
https://lkml.org/lkml/2019/2/7/46
Suggested-by: Joe Perches <joe@perches.com>
Signed-off-by: Nishad Kamdar <nishadkamdar@gmail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Fix sparse warnings:
drivers/crypto/bcm/cipher.c:99:6: warning: symbol 'BCMHEADER' was not declared. Should it be static?
drivers/crypto/bcm/cipher.c:2096:6: warning: symbol 'spu_no_incr_hash' was not declared. Should it be static?
drivers/crypto/bcm/cipher.c:4823:5: warning: symbol 'bcm_spu_probe' was not declared. Should it be static?
drivers/crypto/bcm/cipher.c:4867:5: warning: symbol 'bcm_spu_remove' was not declared. Should it be static?
drivers/crypto/bcm/spu2.c:52:6: warning: symbol 'spu2_cipher_type_names' was not declared. Should it be static?
drivers/crypto/bcm/spu2.c:56:6: warning: symbol 'spu2_cipher_mode_names' was not declared. Should it be static?
drivers/crypto/bcm/spu2.c:60:6: warning: symbol 'spu2_hash_type_names' was not declared. Should it be static?
drivers/crypto/bcm/spu2.c:66:6: warning: symbol 'spu2_hash_mode_names' was not declared. Should it be static?
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Constify the ctx and iv arguments to crypto_chacha_init() and the
various chacha*_stream_xor() functions. This makes it clear that they
are not modified.
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>
- Use sg_init_one() instead of sg_init_table() then sg_set_buf().
- Remove unneeded calls to sg_init_table() prior to scatterwalk_ffwd().
- Simplify initializing the poly tail block.
- Simplify computing padlen.
This doesn't change any actual behavior.
Cc: Martin Willi <martin@strongswan.org>
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The 'chunksize' and 'walksize' properties of skcipher algorithms are
implementation details that users of the skcipher API should not be
looking at. So move their accessor functions from <crypto/skcipher.h>
to <crypto/internal/skcipher.h>.
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>
crypto_skcipher_encrypt() and crypto_skcipher_decrypt() have grown to be
more than a single indirect function call. They now also check whether
a key has been set, and with CONFIG_CRYPTO_STATS=y they also update the
crypto statistics. That can add up to a lot of bloat at every call
site. Moreover, these always involve a function call anyway, which
greatly limits the benefits of inlining.
So change them to be non-inline.
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>
crypto_aead_encrypt() and crypto_aead_decrypt() have grown to be more
than a single indirect function call. They now also check whether a key
has been set, the decryption side checks whether the input is at least
as long as the authentication tag length, and with CONFIG_CRYPTO_STATS=y
they also update the crypto statistics. That can add up to a lot of
bloat at every call site. Moreover, these always involve a function
call anyway, which greatly limits the benefits of inlining.
So change them to be non-inline.
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>
Since commit 944585a64f ("crypto: x86/aes-ni - remove special handling
of AES in PCBC mode"), the "__aes-aesni" internal cipher algorithm is no
longer used. So remove it too.
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>
Rewrite the skcipher API example, changing it to encrypt a buffer with
AES-256-XTS. This addresses various problems with the previous example:
- It requests a specific driver "cbc-aes-aesni", which is unusual.
Normally users ask for "cbc(aes)", not a specific driver.
- It encrypts only a single AES block. For the reader, that doesn't
clearly distinguish the "skcipher" API from the "cipher" API.
- Showing how to encrypt something with bare CBC is arguably a poor
choice of example, as it doesn't follow modern crypto trends. Now,
usually authenticated encryption is recommended, in which case the
user would use the AEAD API, not skcipher. Disk encryption is still a
legitimate use for skcipher, but for that usually XTS is recommended.
- Many other bugs and poor coding practices, such as not setting
CRYPTO_TFM_REQ_MAY_SLEEP, unnecessarily allocating a heap buffer for
the IV, unnecessary NULL checks, using a pointless wrapper struct, and
forgetting to set an error code in one case.
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>
Call cond_resched() after each fuzz test iteration. This avoids stall
warnings if fuzz_iterations is set very high for testing purposes.
While we're at it, also call cond_resched() after finishing testing each
test vector.
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 all algorithms explicitly set cra_driver_name, make it required
for algorithm registration and remove the code that generated a default
cra_driver_name.
Also add an explicit check that cra_name is set too, since that's
obviously required too, yet it didn't seem to be checked anywhere.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Most generic crypto algorithms declare a driver name ending in
"-generic". The rest don't declare a driver name and instead rely on
the crypto API automagically appending "-generic" upon registration.
Having multiple conventions is unnecessarily confusing and makes it
harder to grep for all generic algorithms in the kernel source tree.
But also, allowing NULL driver names is problematic because sometimes
people fail to set it, e.g. the case fixed by commit 4179803643
("crypto: cavium/zip - fix collision with generic cra_driver_name").
Of course, people can also incorrectly name their drivers "-generic".
But that's much easier to notice / grep for.
Therefore, let's make cra_driver_name mandatory. In preparation for
this, this patch makes all generic algorithms set cra_driver_name.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Clear the CRYPTO_TFM_REQ_MAY_SLEEP flag when the chacha20poly1305
operation is being continued from an async completion callback, since
sleeping may not be allowed in that context.
This is basically the same bug that was recently fixed in the xts and
lrw templates. But, it's always been broken in chacha20poly1305 too.
This was found using syzkaller in combination with the updated crypto
self-tests which actually test the MAY_SLEEP flag now.
Reproducer:
python -c 'import socket; socket.socket(socket.AF_ALG, 5, 0).bind(
("aead", "rfc7539(cryptd(chacha20-generic),poly1305-generic)"))'
Kernel output:
BUG: sleeping function called from invalid context at include/crypto/algapi.h:426
in_atomic(): 1, irqs_disabled(): 0, pid: 1001, name: kworker/2:2
[...]
CPU: 2 PID: 1001 Comm: kworker/2:2 Not tainted 5.2.0-rc2 #5
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014
Workqueue: crypto cryptd_queue_worker
Call Trace:
__dump_stack lib/dump_stack.c:77 [inline]
dump_stack+0x4d/0x6a lib/dump_stack.c:113
___might_sleep kernel/sched/core.c:6138 [inline]
___might_sleep.cold.19+0x8e/0x9f kernel/sched/core.c:6095
crypto_yield include/crypto/algapi.h:426 [inline]
crypto_hash_walk_done+0xd6/0x100 crypto/ahash.c:113
shash_ahash_update+0x41/0x60 crypto/shash.c:251
shash_async_update+0xd/0x10 crypto/shash.c:260
crypto_ahash_update include/crypto/hash.h:539 [inline]
poly_setkey+0xf6/0x130 crypto/chacha20poly1305.c:337
poly_init+0x51/0x60 crypto/chacha20poly1305.c:364
async_done_continue crypto/chacha20poly1305.c:78 [inline]
poly_genkey_done+0x15/0x30 crypto/chacha20poly1305.c:369
cryptd_skcipher_complete+0x29/0x70 crypto/cryptd.c:279
cryptd_skcipher_decrypt+0xcd/0x110 crypto/cryptd.c:339
cryptd_queue_worker+0x70/0xa0 crypto/cryptd.c:184
process_one_work+0x1ed/0x420 kernel/workqueue.c:2269
worker_thread+0x3e/0x3a0 kernel/workqueue.c:2415
kthread+0x11f/0x140 kernel/kthread.c:255
ret_from_fork+0x1f/0x30 arch/x86/entry/entry_64.S:352
Fixes: 71ebc4d1b2 ("crypto: chacha20poly1305 - Add a ChaCha20-Poly1305 AEAD construction, RFC7539")
Cc: <stable@vger.kernel.org> # v4.2+
Cc: Martin Willi <martin@strongswan.org>
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
If CRYPTO_DEV_ATMEL_ECC is set m, which select CRC16 to m,
while CRYPTO_DEV_ATMEL_SHA204A is set to y, building fails.
drivers/crypto/atmel-i2c.o: In function 'atmel_i2c_checksum':
atmel-i2c.c:(.text+0x16): undefined reference to 'crc16'
Add CRC16 dependency to CRYPTO_DEV_ATMEL_SHA204A
Reported-by: Hulk Robot <hulkci@huawei.com>
Fixes: da001fb651 ("crypto: atmel-i2c - add support for SHA204A random number generator")
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Disabled the check and set of 'mem' and 'emi_slow'
clocks, since these are not available for iMX7ULP.
Signed-off-by: Iuliana Prodan <iuliana.prodan@nxp.com>
Reviewed-by: Horia Geantă <horia.geanta@nxp.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The CAAM driver currently violates an undocumented and slightly
controversial requirement imposed by the crypto stack that a buffer
referred to by the request structure via its virtual address may not
be modified while any scatterlists passed via the same request
structure are mapped for inbound DMA.
This may result in errors like
alg: aead: decryption failed on test 1 for gcm_base(ctr-aes-caam,ghash-generic): ret=74
alg: aead: Failed to load transform for gcm(aes): -2
on non-cache coherent systems, due to the fact that the GCM driver
passes an IV buffer by virtual address which shares a cacheline with
the auth_tag buffer passed via a scatterlist, resulting in corruption
of the auth_tag when the IV is updated while the DMA mapping is live.
Since the IV that is returned to the caller is only valid for CBC mode,
and given that the in-kernel users of CBC (such as CTS) don't trigger the
same issue as the GCM driver, let's just disable the output IV generation
for all modes except CBC for the time being.
Fixes: 854b06f768 ("crypto: caam - properly set IV after {en,de}crypt")
Cc: Horia Geanta <horia.geanta@nxp.com>
Cc: Iuliana Prodan <iuliana.prodan@nxp.com>
Reported-by: Sascha Hauer <s.hauer@pengutronix.de>
Cc: <stable@vger.kernel.org>
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Horia Geanta <horia.geanta@nxp.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The param2 member in atmel_i2c_cmd is supposed to be little-endian
but was marked as u16. This patch changes it to a __le16 which
reveals a missing endian swap in atmel_i2c_init_read_cmd.
Another missing little-endian marking is also added in
atmel_i2c_checksum.
Fixes: 11105693fa ("crypto: atmel-ecc - introduce Microchip...")
Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Commit c778f96bf3 ("crypto: lrw - Optimize tweak computation")
incorrectly reduced the alignmask of LRW instances from
'__alignof__(u64) - 1' to '__alignof__(__be32) - 1'.
However, xor_tweak() and setkey() assume that the data and key,
respectively, are aligned to 'be128', which has u64 alignment.
Fix the alignmask to be at least '__alignof__(be128) - 1'.
Fixes: c778f96bf3 ("crypto: lrw - Optimize tweak computation")
Cc: <stable@vger.kernel.org> # v4.20+
Cc: Ondrej Mosnacek <omosnace@redhat.com>
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Changing ghash_mod_init() to be subsys_initcall made it start running
before the alignment fault handler has been installed on ARM. In kernel
builds where the keys in the ghash test vectors happened to be
misaligned in the kernel image, this exposed the longstanding bug that
ghash_setkey() is incorrectly casting the key buffer (which can have any
alignment) to be128 for passing to gf128mul_init_4k_lle().
Fix this by memcpy()ing the key to a temporary buffer.
Don't fix it by setting an alignmask on the algorithm instead because
that would unnecessarily force alignment of the data too.
Fixes: 2cdc6899a8 ("crypto: ghash - Add GHASH digest algorithm for GCM")
Reported-by: Peter Robinson <pbrobinson@gmail.com>
Cc: stable@vger.kernel.org
Signed-off-by: Eric Biggers <ebiggers@google.com>
Tested-by: Peter Robinson <pbrobinson@gmail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
xxhash is currently implemented as a self-contained module in /lib.
This patch enables that module to be used as part of the generic kernel
crypto framework. It adds a simple wrapper to the 64bit version.
I've also added test vectors (with help from Nick Terrell). The upstream
xxhash code is tested by running hashing operation on random 222 byte
data with seed values of 0 and a prime number. The upstream test
suite can be found at https://github.com/Cyan4973/xxHash/blob/cf46e0c/xxhsum.c#L664
Essentially hashing is run on data of length 0,1,14,222 with the
aforementioned seed values 0 and prime 2654435761. The particular random
222 byte string was provided to me by Nick Terrell by reading
/dev/random and the checksums were calculated by the upstream xxsum
utility with the following bash script:
dd if=/dev/random of=TEST_VECTOR bs=1 count=222
for a in 0 1; do
for l in 0 1 14 222; do
for s in 0 2654435761; do
echo algo $a length $l seed $s;
head -c $l TEST_VECTOR | ~/projects/kernel/xxHash/xxhsum -H$a -s$s
done
done
done
This produces output as follows:
algo 0 length 0 seed 0
02cc5d05 stdin
algo 0 length 0 seed 2654435761
02cc5d05 stdin
algo 0 length 1 seed 0
25201171 stdin
algo 0 length 1 seed 2654435761
25201171 stdin
algo 0 length 14 seed 0
c1d95975 stdin
algo 0 length 14 seed 2654435761
c1d95975 stdin
algo 0 length 222 seed 0
b38662a6 stdin
algo 0 length 222 seed 2654435761
b38662a6 stdin
algo 1 length 0 seed 0
ef46db3751d8e999 stdin
algo 1 length 0 seed 2654435761
ac75fda2929b17ef stdin
algo 1 length 1 seed 0
27c3f04c2881203a stdin
algo 1 length 1 seed 2654435761
4a15ed26415dfe4d stdin
algo 1 length 14 seed 0
3d33dc700231dfad stdin
algo 1 length 14 seed 2654435761
ea5f7ddef9a64f80 stdin
algo 1 length 222 seed 0
5f3d3c08ec2bef34 stdin
algo 1 length 222 seed 2654435761
6a9df59664c7ed62 stdin
algo 1 is xx64 variant, algo 0 is the 32 bit variant which is currently
not hooked up.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: Eric Biggers <ebiggers@kernel.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The Jitter RNG implementation is updated to comply with upstream version
2.1.2. The change covers the following aspects:
* Time variation measurement is conducted over the LFSR operation
instead of the XOR folding
* Invcation of stuck test during initialization
* Removal of the stirring functionality and the Von-Neumann
unbiaser as the LFSR using a primitive and irreducible polynomial
generates an identical distribution of random bits
This implementation was successfully used in FIPS 140-2 validations
as well as in German BSI evaluations.
This kernel implementation was tested as follows:
* The unchanged kernel code file jitterentropy.c is compiled as part
of user space application to generate raw unconditioned noise
data. That data is processed with the NIST SP800-90B non-IID test
tool to verify that the kernel code exhibits an equal amount of noise
as the upstream Jitter RNG version 2.1.2.
* Using AF_ALG with the libkcapi tool of kcapi-rng the Jitter RNG was
output tested with dieharder to verify that the output does not
exhibit statistical weaknesses. The following command was used:
kcapi-rng -n "jitterentropy_rng" -b 100000000000 | dieharder -a -g 200
* The unchanged kernel code file jitterentropy.c is compiled as part
of user space application to test the LFSR implementation. The
LFSR is injected a monotonically increasing counter as input and
the output is fed into dieharder to verify that the LFSR operation
does not exhibit statistical weaknesses.
* The patch was tested on the Muen separation kernel which returns
a more coarse time stamp to verify that the Jitter RNG does not cause
regressions with its initialization test considering that the Jitter
RNG depends on a high-resolution timer.
Tested-by: Reto Buerki <reet@codelabs.ch>
Signed-off-by: Stephan Mueller <smueller@chronox.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
For hash algorithms implemented using the "shash" algorithm type, test
both the ahash and shash APIs, not just the ahash API.
Testing the ahash API already tests the shash API indirectly, which is
normally good enough. However, there have been corner cases where there
have been shash bugs that don't get exposed through the ahash API. So,
update testmgr to test the shash API too.
This would have detected the arm64 SHA-1 and SHA-2 bugs for which fixes
were just sent out (https://patchwork.kernel.org/patch/10964843/ and
https://patchwork.kernel.org/patch/10965089/):
alg: shash: sha1-ce test failed (wrong result) on test vector 0, cfg="init+finup aligned buffer"
alg: shash: sha224-ce test failed (wrong result) on test vector 0, cfg="init+finup aligned buffer"
alg: shash: sha256-ce test failed (wrong result) on test vector 0, cfg="init+finup aligned buffer"
This also would have detected the bugs fixed by commit 307508d107
("crypto: crct10dif-generic - fix use via crypto_shash_digest()") and
commit dec3d0b107
("crypto: x86/crct10dif-pcl - fix use via crypto_shash_digest()").
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>
The sha256-ce finup implementation for ARM64 produces wrong digest
for empty input (len=0). Expected: the actual digest, result: initial
value of SHA internal state. The error is in sha256_ce_finup:
for empty data `finalize` will be 1, so the code is relying on
sha2_ce_transform to make the final round. However, in
sha256_base_do_update, the block function will not be called when
len == 0.
Fix it by setting finalize to 0 if data is empty.
Fixes: 03802f6a80 ("crypto: arm64/sha2-ce - move SHA-224/256 ARMv8 implementation to base layer")
Cc: stable@vger.kernel.org
Signed-off-by: Elena Petrova <lenaptr@google.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The sha1-ce finup implementation for ARM64 produces wrong digest
for empty input (len=0). Expected: da39a3ee..., result: 67452301...
(initial value of SHA internal state). The error is in sha1_ce_finup:
for empty data `finalize` will be 1, so the code is relying on
sha1_ce_transform to make the final round. However, in
sha1_base_do_update, the block function will not be called when
len == 0.
Fix it by setting finalize to 0 if data is empty.
Fixes: 07eb54d306 ("crypto: arm64/sha1-ce - move SHA-1 ARMv8 implementation to base layer")
Cc: stable@vger.kernel.org
Signed-off-by: Elena Petrova <lenaptr@google.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
For rsa and pkcs1pad, CAAM expects an input of modulus size.
For this we strip the leading zeros in case the size is more than modulus.
This commit avoids modifying the crypto request while stripping zeros from
input, to comply with the crypto API requirement. This is done by adding
a fixup input pointer and length.
Signed-off-by: Iuliana Prodan <iuliana.prodan@nxp.com>
Reviewed-by: Horia Geanta <horia.geanta@nxp.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The problem is with the input data size sent to CAAM for encrypt/decrypt.
Pkcs1pad is failing due to pkcs1 padding done in SW starting with0x01
instead of 0x00 0x01.
CAAM expects an input of modulus size. For this we strip the leading
zeros in case the size is more than modulus or pad the input with zeros
until the modulus size is reached.
Signed-off-by: Iuliana Prodan <iuliana.prodan@nxp.com>
Reviewed-by: Horia Geantă <horia.geanta@nxp.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
When performing a transformation the hardware is given result
descriptors to save the result data. Those result descriptors are
batched using a 'first' and a 'last' bit. There are cases were more
descriptors than needed are given to the engine, leading to the engine
only using some of them, and not setting the last bit on the last
descriptor we gave. This causes issues were the driver and the hardware
aren't in sync anymore about the number of result descriptors given (as
the driver do not give a pool of descriptor to use for any
transformation, but a pool of descriptors to use *per* transformation).
This patch fixes it by attaching the number of given result descriptors
to the requests, and by using this number instead of the 'last' bit
found on the descriptors to process them.
Signed-off-by: Antoine Tenart <antoine.tenart@bootlin.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Replace sg_nents_for_len by sg_nents when DMA mapping/unmapping buffers
and when looping over the SG entries. This fix cases where the SG
entries aren't used fully, which would in such cases led to using fewer
SG entries than needed (and thus the engine wouldn't have access to the
full input data and the result would be wrong).
Signed-off-by: Antoine Tenart <antoine.tenart@bootlin.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
