diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2020-08-03 10:40:14 -0700 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2020-08-03 10:40:14 -0700 |
commit | ab5c60b79ab6cc50b39bbb21b2f9fb55af900b84 (patch) | |
tree | 71fa895fbf01e3b88f26cf257d9105f9d286b631 /arch/arm | |
parent | 5577416c39652d395a6045677f4f598564aba1cf (diff) | |
parent | 3cbfe80737c18ac6e635421ab676716a393d3074 (diff) |
Merge branch 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Pull crypto updates from Herbert Xu:
"API:
- Add support for allocating transforms on a specific NUMA Node
- Introduce the flag CRYPTO_ALG_ALLOCATES_MEMORY for storage users
Algorithms:
- Drop PMULL based ghash on arm64
- Fixes for building with clang on x86
- Add sha256 helper that does the digest in one go
- Add SP800-56A rev 3 validation checks to dh
Drivers:
- Permit users to specify NUMA node in hisilicon/zip
- Add support for i.MX6 in imx-rngc
- Add sa2ul crypto driver
- Add BA431 hwrng driver
- Add Ingenic JZ4780 and X1000 hwrng driver
- Spread IRQ affinity in inside-secure and marvell/cesa"
* 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6: (157 commits)
crypto: sa2ul - Fix inconsistent IS_ERR and PTR_ERR
hwrng: core - remove redundant initialization of variable ret
crypto: x86/curve25519 - Remove unused carry variables
crypto: ingenic - Add hardware RNG for Ingenic JZ4780 and X1000
dt-bindings: RNG: Add Ingenic RNG bindings.
crypto: caam/qi2 - add module alias
crypto: caam - add more RNG hw error codes
crypto: caam/jr - remove incorrect reference to caam_jr_register()
crypto: caam - silence .setkey in case of bad key length
crypto: caam/qi2 - create ahash shared descriptors only once
crypto: caam/qi2 - fix error reporting for caam_hash_alloc
crypto: caam - remove deadcode on 32-bit platforms
crypto: ccp - use generic power management
crypto: xts - Replace memcpy() invocation with simple assignment
crypto: marvell/cesa - irq balance
crypto: inside-secure - irq balance
crypto: ecc - SP800-56A rev 3 local public key validation
crypto: dh - SP800-56A rev 3 local public key validation
crypto: dh - check validity of Z before export
lib/mpi: Add mpi_sub_ui()
...
Diffstat (limited to 'arch/arm')
-rw-r--r-- | arch/arm/crypto/crc32-ce-core.S | 2 | ||||
-rw-r--r-- | arch/arm/crypto/ghash-ce-glue.c | 51 | ||||
-rw-r--r-- | arch/arm/crypto/sha1-armv4-large.S | 2 | ||||
-rw-r--r-- | arch/arm/crypto/sha256-armv4.pl | 2 | ||||
-rw-r--r-- | arch/arm/crypto/sha256-core.S_shipped | 2 | ||||
-rw-r--r-- | arch/arm/crypto/sha512-armv4.pl | 4 | ||||
-rw-r--r-- | arch/arm/crypto/sha512-core.S_shipped | 4 |
7 files changed, 32 insertions, 35 deletions
diff --git a/arch/arm/crypto/crc32-ce-core.S b/arch/arm/crypto/crc32-ce-core.S index 5cbd4a6fedad..3f13a76b9066 100644 --- a/arch/arm/crypto/crc32-ce-core.S +++ b/arch/arm/crypto/crc32-ce-core.S @@ -39,7 +39,7 @@ * CRC32 polynomial:0x04c11db7(BE)/0xEDB88320(LE) * PCLMULQDQ is a new instruction in Intel SSE4.2, the reference can be found * at: - * http://www.intel.com/products/processor/manuals/ + * https://www.intel.com/products/processor/manuals/ * Intel(R) 64 and IA-32 Architectures Software Developer's Manual * Volume 2B: Instruction Set Reference, N-Z * diff --git a/arch/arm/crypto/ghash-ce-glue.c b/arch/arm/crypto/ghash-ce-glue.c index a00fd329255f..f13401f3e669 100644 --- a/arch/arm/crypto/ghash-ce-glue.c +++ b/arch/arm/crypto/ghash-ce-glue.c @@ -16,6 +16,7 @@ #include <crypto/gf128mul.h> #include <linux/cpufeature.h> #include <linux/crypto.h> +#include <linux/jump_label.h> #include <linux/module.h> MODULE_DESCRIPTION("GHASH hash function using ARMv8 Crypto Extensions"); @@ -27,12 +28,8 @@ MODULE_ALIAS_CRYPTO("ghash"); #define GHASH_DIGEST_SIZE 16 struct ghash_key { - u64 h[2]; - u64 h2[2]; - u64 h3[2]; - u64 h4[2]; - be128 k; + u64 h[][2]; }; struct ghash_desc_ctx { @@ -46,16 +43,12 @@ struct ghash_async_ctx { }; asmlinkage void pmull_ghash_update_p64(int blocks, u64 dg[], const char *src, - struct ghash_key const *k, - const char *head); + u64 const h[][2], const char *head); asmlinkage void pmull_ghash_update_p8(int blocks, u64 dg[], const char *src, - struct ghash_key const *k, - const char *head); + u64 const h[][2], const char *head); -static void (*pmull_ghash_update)(int blocks, u64 dg[], const char *src, - struct ghash_key const *k, - const char *head); +static __ro_after_init DEFINE_STATIC_KEY_FALSE(use_p64); static int ghash_init(struct shash_desc *desc) { @@ -70,7 +63,10 @@ static void ghash_do_update(int blocks, u64 dg[], const char *src, { if (likely(crypto_simd_usable())) { kernel_neon_begin(); - pmull_ghash_update(blocks, dg, src, key, head); + if (static_branch_likely(&use_p64)) + pmull_ghash_update_p64(blocks, dg, src, key->h, head); + else + pmull_ghash_update_p8(blocks, dg, src, key->h, head); kernel_neon_end(); } else { be128 dst = { cpu_to_be64(dg[1]), cpu_to_be64(dg[0]) }; @@ -161,25 +157,26 @@ static int ghash_setkey(struct crypto_shash *tfm, const u8 *inkey, unsigned int keylen) { struct ghash_key *key = crypto_shash_ctx(tfm); - be128 h; if (keylen != GHASH_BLOCK_SIZE) return -EINVAL; /* needed for the fallback */ memcpy(&key->k, inkey, GHASH_BLOCK_SIZE); - ghash_reflect(key->h, &key->k); + ghash_reflect(key->h[0], &key->k); - h = key->k; - gf128mul_lle(&h, &key->k); - ghash_reflect(key->h2, &h); + if (static_branch_likely(&use_p64)) { + be128 h = key->k; - gf128mul_lle(&h, &key->k); - ghash_reflect(key->h3, &h); + gf128mul_lle(&h, &key->k); + ghash_reflect(key->h[1], &h); - gf128mul_lle(&h, &key->k); - ghash_reflect(key->h4, &h); + gf128mul_lle(&h, &key->k); + ghash_reflect(key->h[2], &h); + gf128mul_lle(&h, &key->k); + ghash_reflect(key->h[3], &h); + } return 0; } @@ -195,7 +192,7 @@ static struct shash_alg ghash_alg = { .base.cra_driver_name = "ghash-ce-sync", .base.cra_priority = 300 - 1, .base.cra_blocksize = GHASH_BLOCK_SIZE, - .base.cra_ctxsize = sizeof(struct ghash_key), + .base.cra_ctxsize = sizeof(struct ghash_key) + sizeof(u64[2]), .base.cra_module = THIS_MODULE, }; @@ -354,10 +351,10 @@ static int __init ghash_ce_mod_init(void) if (!(elf_hwcap & HWCAP_NEON)) return -ENODEV; - if (elf_hwcap2 & HWCAP2_PMULL) - pmull_ghash_update = pmull_ghash_update_p64; - else - pmull_ghash_update = pmull_ghash_update_p8; + if (elf_hwcap2 & HWCAP2_PMULL) { + ghash_alg.base.cra_ctxsize += 3 * sizeof(u64[2]); + static_branch_enable(&use_p64); + } err = crypto_register_shash(&ghash_alg); if (err) diff --git a/arch/arm/crypto/sha1-armv4-large.S b/arch/arm/crypto/sha1-armv4-large.S index f82cd8cf5a09..1c8b685149f2 100644 --- a/arch/arm/crypto/sha1-armv4-large.S +++ b/arch/arm/crypto/sha1-armv4-large.S @@ -13,7 +13,7 @@ @ Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL @ project. The module is, however, dual licensed under OpenSSL and @ CRYPTOGAMS licenses depending on where you obtain it. For further -@ details see http://www.openssl.org/~appro/cryptogams/. +@ details see https://www.openssl.org/~appro/cryptogams/. @ ==================================================================== @ sha1_block procedure for ARMv4. diff --git a/arch/arm/crypto/sha256-armv4.pl b/arch/arm/crypto/sha256-armv4.pl index a03cf4dfb781..9f96ff48e4a8 100644 --- a/arch/arm/crypto/sha256-armv4.pl +++ b/arch/arm/crypto/sha256-armv4.pl @@ -13,7 +13,7 @@ # Written by Andy Polyakov <appro@openssl.org> for the OpenSSL # project. The module is, however, dual licensed under OpenSSL and # CRYPTOGAMS licenses depending on where you obtain it. For further -# details see http://www.openssl.org/~appro/cryptogams/. +# details see https://www.openssl.org/~appro/cryptogams/. # ==================================================================== # SHA256 block procedure for ARMv4. May 2007. diff --git a/arch/arm/crypto/sha256-core.S_shipped b/arch/arm/crypto/sha256-core.S_shipped index 054aae0edfce..ea04b2ab0c33 100644 --- a/arch/arm/crypto/sha256-core.S_shipped +++ b/arch/arm/crypto/sha256-core.S_shipped @@ -12,7 +12,7 @@ @ Written by Andy Polyakov <appro@openssl.org> for the OpenSSL @ project. The module is, however, dual licensed under OpenSSL and @ CRYPTOGAMS licenses depending on where you obtain it. For further -@ details see http://www.openssl.org/~appro/cryptogams/. +@ details see https://www.openssl.org/~appro/cryptogams/. @ ==================================================================== @ SHA256 block procedure for ARMv4. May 2007. diff --git a/arch/arm/crypto/sha512-armv4.pl b/arch/arm/crypto/sha512-armv4.pl index 788c17b56ecc..69df68981acd 100644 --- a/arch/arm/crypto/sha512-armv4.pl +++ b/arch/arm/crypto/sha512-armv4.pl @@ -13,7 +13,7 @@ # Written by Andy Polyakov <appro@openssl.org> for the OpenSSL # project. The module is, however, dual licensed under OpenSSL and # CRYPTOGAMS licenses depending on where you obtain it. For further -# details see http://www.openssl.org/~appro/cryptogams/. +# details see https://www.openssl.org/~appro/cryptogams/. # ==================================================================== # SHA512 block procedure for ARMv4. September 2007. @@ -43,7 +43,7 @@ # terms it's 22.6 cycles per byte, which is disappointing result. # Technical writers asserted that 3-way S4 pipeline can sustain # multiple NEON instructions per cycle, but dual NEON issue could -# not be observed, see http://www.openssl.org/~appro/Snapdragon-S4.html +# not be observed, see https://www.openssl.org/~appro/Snapdragon-S4.html # for further details. On side note Cortex-A15 processes one byte in # 16 cycles. diff --git a/arch/arm/crypto/sha512-core.S_shipped b/arch/arm/crypto/sha512-core.S_shipped index 710ea309769e..cb147db5cbfe 100644 --- a/arch/arm/crypto/sha512-core.S_shipped +++ b/arch/arm/crypto/sha512-core.S_shipped @@ -12,7 +12,7 @@ @ Written by Andy Polyakov <appro@openssl.org> for the OpenSSL @ project. The module is, however, dual licensed under OpenSSL and @ CRYPTOGAMS licenses depending on where you obtain it. For further -@ details see http://www.openssl.org/~appro/cryptogams/. +@ details see https://www.openssl.org/~appro/cryptogams/. @ ==================================================================== @ SHA512 block procedure for ARMv4. September 2007. @@ -42,7 +42,7 @@ @ terms it's 22.6 cycles per byte, which is disappointing result. @ Technical writers asserted that 3-way S4 pipeline can sustain @ multiple NEON instructions per cycle, but dual NEON issue could -@ not be observed, see http://www.openssl.org/~appro/Snapdragon-S4.html +@ not be observed, see https://www.openssl.org/~appro/Snapdragon-S4.html @ for further details. On side note Cortex-A15 processes one byte in @ 16 cycles. |