diff options
| author | sdlyyxy <sdlyyxy@icloud.com> | 2023-07-07 19:26:10 +0800 |
|---|---|---|
| committer | Pauli <pauli@openssl.org> | 2023-07-21 10:19:19 +1000 |
| commit | f6484de23df5f04e1f9fa8418e942c45c1a65578 (patch) | |
| tree | 487e136cf7d68e78ae7f14ba2b442300001570d4 /providers | |
| parent | 0c85bcbaeabe3a695831bec44ab87964725a51a6 (diff) | |
Enable ARMv8.2 accelerated SHA3 on compatible Apple CPUs
The hardware-assisted ARMv8.2 implementation is already in keccak1600-armv8.pl.
It is not called because the author mentioned that it's not actually obvious
that it will provide performance improvements. The test on Apple M1 Firestorm
shows that the ARMv8.2 implementation could improve about 36% for large blocks.
So let's enable ARMv8.2 accelerated SHA3 on Apple CPU family.
Fixes #21380
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21398)
Diffstat (limited to 'providers')
| -rw-r--r-- | providers/implementations/digests/sha3_prov.c | 59 |
1 files changed, 59 insertions, 0 deletions
diff --git a/providers/implementations/digests/sha3_prov.c b/providers/implementations/digests/sha3_prov.c index 825d3249fa..1348d0e06a 100644 --- a/providers/implementations/digests/sha3_prov.c +++ b/providers/implementations/digests/sha3_prov.c @@ -249,6 +249,65 @@ static PROV_SHA3_METHOD kmac_s390x_md = } else { \ ctx->meth = sha3_generic_md; \ } +#elif defined(__aarch64__) +# include "arm_arch.h" + +static sha3_absorb_fn armsha3_sha3_absorb; + +size_t SHA3_absorb_cext(uint64_t A[5][5], const unsigned char *inp, size_t len, + size_t r); +/*- + * Hardware-assisted ARMv8.2 SHA3 extension version of the absorb() + */ +static size_t armsha3_sha3_absorb(void *vctx, const void *inp, size_t len) +{ + KECCAK1600_CTX *ctx = vctx; + + return SHA3_absorb_cext(ctx->A, inp, len, ctx->block_size); +} + +static PROV_SHA3_METHOD sha3_ARMSHA3_md = +{ + armsha3_sha3_absorb, + generic_sha3_final +}; +/* Detection on Apple operating systems */ +# if defined(__APPLE__) +# define ARM_SHA3_CAPABLE (OPENSSL_armcap_P & ARMV8_SHA3) +# define SHA3_SET_MD(uname, typ) \ + if (ARM_SHA3_CAPABLE) { \ + ctx->meth = sha3_ARMSHA3_md; \ + } else { \ + ctx->meth = sha3_generic_md; \ + } +# define KMAC_SET_MD(bitlen) \ + if (ARM_SHA3_CAPABLE) { \ + ctx->meth = sha3_ARMSHA3_md; \ + } else { \ + ctx->meth = sha3_generic_md; \ + } +/* Detection on other operating systems */ +# else +# define ARM_HAS_FASTER_SHA3 \ + (MIDR_IS_CPU_MODEL(OPENSSL_arm_midr, ARM_CPU_IMP_APPLE, APPLE_CPU_PART_M1_FIRESTORM) ||\ + MIDR_IS_CPU_MODEL(OPENSSL_arm_midr, ARM_CPU_IMP_APPLE, APPLE_CPU_PART_M1_FIRESTORM_PRO) ||\ + MIDR_IS_CPU_MODEL(OPENSSL_arm_midr, ARM_CPU_IMP_APPLE, APPLE_CPU_PART_M1_FIRESTORM_MAX) ||\ + MIDR_IS_CPU_MODEL(OPENSSL_arm_midr, ARM_CPU_IMP_APPLE, APPLE_CPU_PART_M2_AVALANCHE) ||\ + MIDR_IS_CPU_MODEL(OPENSSL_arm_midr, ARM_CPU_IMP_APPLE, APPLE_CPU_PART_M2_AVALANCHE_PRO) ||\ + MIDR_IS_CPU_MODEL(OPENSSL_arm_midr, ARM_CPU_IMP_APPLE, APPLE_CPU_PART_M2_AVALANCHE_MAX)) +# define SHA3_SET_MD(uname, typ) \ + if (ARM_HAS_FASTER_SHA3) { \ + ctx->meth = sha3_ARMSHA3_md; \ + } else { \ + ctx->meth = sha3_generic_md; \ + } +# define KMAC_SET_MD(bitlen) \ + if (ARM_HAS_FASTER_SHA3) { \ + ctx->meth = sha3_ARMSHA3_md; \ + } else { \ + ctx->meth = sha3_generic_md; \ + } +# endif /* APPLE */ #else # define SHA3_SET_MD(uname, typ) ctx->meth = sha3_generic_md; # define KMAC_SET_MD(bitlen) ctx->meth = sha3_generic_md; |