diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2023-11-02 16:15:30 -1000 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2023-11-02 16:15:30 -1000 |
commit | bc3012f4e3a9765de81f454cb8f9bb16aafc6ff5 (patch) | |
tree | 2c127c669218b8c74c843331e455372f88a6a848 /include/linux/crypto.h | |
parent | 6803bd7956ca8fc43069c2e42016f17f3c2fbf30 (diff) | |
parent | a312e07a65fb598ed239b940434392721385c722 (diff) |
Merge tag 'v6.7-p1' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Pull crypto updates from Herbert Xu:
"API:
- Add virtual-address based lskcipher interface
- Optimise ahash/shash performance in light of costly indirect calls
- Remove ahash alignmask attribute
Algorithms:
- Improve AES/XTS performance of 6-way unrolling for ppc
- Remove some uses of obsolete algorithms (md4, md5, sha1)
- Add FIPS 202 SHA-3 support in pkcs1pad
- Add fast path for single-page messages in adiantum
- Remove zlib-deflate
Drivers:
- Add support for S4 in meson RNG driver
- Add STM32MP13x support in stm32
- Add hwrng interface support in qcom-rng
- Add support for deflate algorithm in hisilicon/zip"
* tag 'v6.7-p1' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6: (283 commits)
crypto: adiantum - flush destination page before unmapping
crypto: testmgr - move pkcs1pad(rsa,sha3-*) to correct place
Documentation/module-signing.txt: bring up to date
module: enable automatic module signing with FIPS 202 SHA-3
crypto: asymmetric_keys - allow FIPS 202 SHA-3 signatures
crypto: rsa-pkcs1pad - Add FIPS 202 SHA-3 support
crypto: FIPS 202 SHA-3 register in hash info for IMA
x509: Add OIDs for FIPS 202 SHA-3 hash and signatures
crypto: ahash - optimize performance when wrapping shash
crypto: ahash - check for shash type instead of not ahash type
crypto: hash - move "ahash wrapping shash" functions to ahash.c
crypto: talitos - stop using crypto_ahash::init
crypto: chelsio - stop using crypto_ahash::init
crypto: ahash - improve file comment
crypto: ahash - remove struct ahash_request_priv
crypto: ahash - remove crypto_ahash_alignmask
crypto: gcm - stop using alignmask of ahash
crypto: chacha20poly1305 - stop using alignmask of ahash
crypto: ccm - stop using alignmask of ahash
net: ipv6: stop checking crypto_ahash_alignmask
...
Diffstat (limited to 'include/linux/crypto.h')
-rw-r--r-- | include/linux/crypto.h | 30 |
1 files changed, 15 insertions, 15 deletions
diff --git a/include/linux/crypto.h b/include/linux/crypto.h index 31f6fee0c36c..b164da5e129e 100644 --- a/include/linux/crypto.h +++ b/include/linux/crypto.h @@ -24,6 +24,7 @@ #define CRYPTO_ALG_TYPE_CIPHER 0x00000001 #define CRYPTO_ALG_TYPE_COMPRESS 0x00000002 #define CRYPTO_ALG_TYPE_AEAD 0x00000003 +#define CRYPTO_ALG_TYPE_LSKCIPHER 0x00000004 #define CRYPTO_ALG_TYPE_SKCIPHER 0x00000005 #define CRYPTO_ALG_TYPE_AKCIPHER 0x00000006 #define CRYPTO_ALG_TYPE_SIG 0x00000007 @@ -35,8 +36,6 @@ #define CRYPTO_ALG_TYPE_SHASH 0x0000000e #define CRYPTO_ALG_TYPE_AHASH 0x0000000f -#define CRYPTO_ALG_TYPE_HASH_MASK 0x0000000e -#define CRYPTO_ALG_TYPE_AHASH_MASK 0x0000000e #define CRYPTO_ALG_TYPE_ACOMPRESS_MASK 0x0000000e #define CRYPTO_ALG_LARVAL 0x00000010 @@ -111,7 +110,6 @@ * crypto_aead_walksize() (with the remainder going at the end), no chunk * can cross a page boundary or a scatterlist element boundary. * ahash: - * - The result buffer must be aligned to the algorithm's alignmask. * - crypto_ahash_finup() must not be used unless the algorithm implements * ->finup() natively. */ @@ -279,18 +277,20 @@ struct compress_alg { * @cra_ctxsize: Size of the operational context of the transformation. This * value informs the kernel crypto API about the memory size * needed to be allocated for the transformation context. - * @cra_alignmask: Alignment mask for the input and output data buffer. The data - * buffer containing the input data for the algorithm must be - * aligned to this alignment mask. The data buffer for the - * output data must be aligned to this alignment mask. Note that - * the Crypto API will do the re-alignment in software, but - * only under special conditions and there is a performance hit. - * The re-alignment happens at these occasions for different - * @cra_u types: cipher -- For both input data and output data - * buffer; ahash -- For output hash destination buf; shash -- - * For output hash destination buf. - * This is needed on hardware which is flawed by design and - * cannot pick data from arbitrary addresses. + * @cra_alignmask: For cipher, skcipher, lskcipher, and aead algorithms this is + * 1 less than the alignment, in bytes, that the algorithm + * implementation requires for input and output buffers. When + * the crypto API is invoked with buffers that are not aligned + * to this alignment, the crypto API automatically utilizes + * appropriately aligned temporary buffers to comply with what + * the algorithm needs. (For scatterlists this happens only if + * the algorithm uses the skcipher_walk helper functions.) This + * misalignment handling carries a performance penalty, so it is + * preferred that algorithms do not set a nonzero alignmask. + * Also, crypto API users may wish to allocate buffers aligned + * to the alignmask of the algorithm being used, in order to + * avoid the API having to realign them. Note: the alignmask is + * not supported for hash algorithms and is always 0 for them. * @cra_priority: Priority of this transformation implementation. In case * multiple transformations with same @cra_name are available to * the Crypto API, the kernel will use the one with highest |