mirror/openssl
2
0
Fork 0
mirror of https://github.com/openssl/openssl.git synced 2024-11-27 05:21:51 +08:00
Code Issues Packages Projects Releases Wiki Activity
5d70f11823
openssl/crypto/armcap.c
Tom Cosgrove 11adf9a75d Apply the AES-GCM unroll8 optimisation to Microsoft Azure Cobalt 100 
Performance improvements range from 18% to 32%.

Change-Id: Ifb89eeac3c0625a582a25ff07cf7f9c9ec8f5ba6

Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/23651)
2024-02-22 16:07:02 +01:00

446 lines
14 KiB
C
Raw Blame History

/*
* Copyright 2011-2023 The OpenSSL Project Authors. All Rights Reserved.
*
* Licensed under the Apache License 2.0 (the "License"). You may not use
* this file except in compliance with the License. You can obtain a copy
* in the file LICENSE in the source distribution or at
* https://www.openssl.org/source/license.html
*/
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <openssl/crypto.h>
#ifdef __APPLE__
#include <sys/sysctl.h>
#else
#include <setjmp.h>
#include <signal.h>
#endif
#include "internal/cryptlib.h"
#ifdef _WIN32
#include <windows.h>
#else
#include <unistd.h>
#endif
#include "arm_arch.h"
unsigned int OPENSSL_armcap_P = 0;
unsigned int OPENSSL_arm_midr = 0;
unsigned int OPENSSL_armv8_rsa_neonized = 0;
#ifdef _WIN32
void OPENSSL_cpuid_setup(void)
{
OPENSSL_armcap_P |= ARMV7_NEON;
OPENSSL_armv8_rsa_neonized = 1;
if (IsProcessorFeaturePresent(PF_ARM_V8_CRYPTO_INSTRUCTIONS_AVAILABLE)) {
// These are all covered by one call in Windows
OPENSSL_armcap_P |= ARMV8_AES;
OPENSSL_armcap_P |= ARMV8_PMULL;
OPENSSL_armcap_P |= ARMV8_SHA1;
OPENSSL_armcap_P |= ARMV8_SHA256;
}
}
uint32_t OPENSSL_rdtsc(void)
{
return 0;
}
#elif __ARM_MAX_ARCH__ < 7
void OPENSSL_cpuid_setup(void)
{
}
uint32_t OPENSSL_rdtsc(void)
{
return 0;
}
#else /* !_WIN32 && __ARM_MAX_ARCH__ >= 7 */
/* 3 ways of handling things here: __APPLE__, getauxval() or SIGILL detect */
/* First determine if getauxval() is available (OSSL_IMPLEMENT_GETAUXVAL) */
# if defined(__GNUC__) && __GNUC__>=2
void OPENSSL_cpuid_setup(void) __attribute__ ((constructor));
# endif
# if defined(__GLIBC__) && defined(__GLIBC_PREREQ)
# if __GLIBC_PREREQ(2, 16)
# include <sys/auxv.h>
# define OSSL_IMPLEMENT_GETAUXVAL
# endif
# elif defined(__ANDROID_API__)
/* see https://developer.android.google.cn/ndk/guides/cpu-features */
# if __ANDROID_API__ >= 18
# include <sys/auxv.h>
# define OSSL_IMPLEMENT_GETAUXVAL
# endif
# endif
# if defined(__FreeBSD__)
# include <sys/param.h>
# if __FreeBSD_version >= 1200000
# include <sys/auxv.h>
# define OSSL_IMPLEMENT_GETAUXVAL
static unsigned long getauxval(unsigned long key)
{
unsigned long val = 0ul;
if (elf_aux_info((int)key, &val, sizeof(val)) != 0)
return 0ul;
return val;
}
# endif
# endif
/*
* Android: according to https://developer.android.com/ndk/guides/cpu-features,
* getauxval is supported starting with API level 18
*/
# if defined(__ANDROID__) && defined(__ANDROID_API__) && __ANDROID_API__ >= 18
# include <sys/auxv.h>
# define OSSL_IMPLEMENT_GETAUXVAL
# endif
/*
* ARM puts the feature bits for Crypto Extensions in AT_HWCAP2, whereas
* AArch64 used AT_HWCAP.
*/
# ifndef AT_HWCAP
# define AT_HWCAP 16
# endif
# ifndef AT_HWCAP2
# define AT_HWCAP2 26
# endif
# if defined(__arm__) || defined (__arm)
# define OSSL_HWCAP AT_HWCAP
# define OSSL_HWCAP_NEON (1 << 12)
# define OSSL_HWCAP_CE AT_HWCAP2
# define OSSL_HWCAP_CE_AES (1 << 0)
# define OSSL_HWCAP_CE_PMULL (1 << 1)
# define OSSL_HWCAP_CE_SHA1 (1 << 2)
# define OSSL_HWCAP_CE_SHA256 (1 << 3)
# elif defined(__aarch64__)
# define OSSL_HWCAP AT_HWCAP
# define OSSL_HWCAP_NEON (1 << 1)
# define OSSL_HWCAP_CE AT_HWCAP
# define OSSL_HWCAP_CE_AES (1 << 3)
# define OSSL_HWCAP_CE_PMULL (1 << 4)
# define OSSL_HWCAP_CE_SHA1 (1 << 5)
# define OSSL_HWCAP_CE_SHA256 (1 << 6)
# define OSSL_HWCAP_CPUID (1 << 11)
# define OSSL_HWCAP_SHA3 (1 << 17)
# define OSSL_HWCAP_CE_SM3 (1 << 18)
# define OSSL_HWCAP_CE_SM4 (1 << 19)
# define OSSL_HWCAP_CE_SHA512 (1 << 21)
# define OSSL_HWCAP_SVE (1 << 22)
/* AT_HWCAP2 */
# define OSSL_HWCAP2 26
# define OSSL_HWCAP2_SVE2 (1 << 1)
# define OSSL_HWCAP2_RNG (1 << 16)
# endif
uint32_t _armv7_tick(void);
uint32_t OPENSSL_rdtsc(void)
{
if (OPENSSL_armcap_P & ARMV7_TICK)
return _armv7_tick();
else
return 0;
}
# ifdef __aarch64__
size_t OPENSSL_rndr_asm(unsigned char *buf, size_t len);
size_t OPENSSL_rndrrs_asm(unsigned char *buf, size_t len);
size_t OPENSSL_rndr_bytes(unsigned char *buf, size_t len);
size_t OPENSSL_rndrrs_bytes(unsigned char *buf, size_t len);
static size_t OPENSSL_rndr_wrapper(size_t (*func)(unsigned char *, size_t), unsigned char *buf, size_t len)
{
size_t buffer_size = 0;
int i;
for (i = 0; i < 8; i++) {
buffer_size = func(buf, len);
if (buffer_size == len)
break;
usleep(5000); /* 5000 microseconds (5 milliseconds) */
}
return buffer_size;
}
size_t OPENSSL_rndr_bytes(unsigned char *buf, size_t len)
{
return OPENSSL_rndr_wrapper(OPENSSL_rndr_asm, buf, len);
}
size_t OPENSSL_rndrrs_bytes(unsigned char *buf, size_t len)
{
return OPENSSL_rndr_wrapper(OPENSSL_rndrrs_asm, buf, len);
}
# endif
# if !defined(__APPLE__) && !defined(OSSL_IMPLEMENT_GETAUXVAL)
static sigset_t all_masked;
static sigjmp_buf ill_jmp;
static void ill_handler(int sig)
{
siglongjmp(ill_jmp, sig);
}
/*
* Following subroutines could have been inlined, but not all
* ARM compilers support inline assembler, and we'd then have to
* worry about the compiler optimising out the detection code...
*/
void _armv7_neon_probe(void);
void _armv8_aes_probe(void);
void _armv8_sha1_probe(void);
void _armv8_sha256_probe(void);
void _armv8_pmull_probe(void);
# ifdef __aarch64__
void _armv8_sm3_probe(void);
void _armv8_sm4_probe(void);
void _armv8_sha512_probe(void);
void _armv8_eor3_probe(void);
void _armv8_sve_probe(void);
void _armv8_sve2_probe(void);
void _armv8_rng_probe(void);
# endif
# endif /* !__APPLE__ && !OSSL_IMPLEMENT_GETAUXVAL */
/* We only call _armv8_cpuid_probe() if (OPENSSL_armcap_P & ARMV8_CPUID) != 0 */
unsigned int _armv8_cpuid_probe(void);
# if defined(__APPLE__)
/*
* Checks the specified integer sysctl, returning `value` if it's 1, otherwise returning 0.
*/
static unsigned int sysctl_query(const char *name, unsigned int value)
{
unsigned int sys_value = 0;
size_t len = sizeof(sys_value);
return (sysctlbyname(name, &sys_value, &len, NULL, 0) == 0 && sys_value == 1) ? value : 0;
}
# elif !defined(OSSL_IMPLEMENT_GETAUXVAL)
/*
* Calls a provided probe function, which may SIGILL. If it doesn't, return `value`, otherwise return 0.
*/
static unsigned int arm_probe_for(void (*probe)(void), volatile unsigned int value)
{
if (sigsetjmp(ill_jmp, 1) == 0) {
probe();
return value;
} else {
/* The probe function gave us SIGILL */
return 0;
}
}
# endif
void OPENSSL_cpuid_setup(void)
{
const char *e;
# if !defined(__APPLE__) && !defined(OSSL_IMPLEMENT_GETAUXVAL)
struct sigaction ill_oact, ill_act;
sigset_t oset;
# endif
static int trigger = 0;
if (trigger)
return;
trigger = 1;
OPENSSL_armcap_P = 0;
if ((e = getenv("OPENSSL_armcap"))) {
OPENSSL_armcap_P = (unsigned int)strtoul(e, NULL, 0);
return;
}
# if defined(__APPLE__)
# if !defined(__aarch64__)
/*
* Capability probing by catching SIGILL appears to be problematic
* on iOS. But since Apple universe is "monocultural", it's actually
* possible to simply set pre-defined processor capability mask.
*/
if (1) {
OPENSSL_armcap_P = ARMV7_NEON;
return;
}
# else
{
/*
* From
* https://github.com/llvm/llvm-project/blob/412237dcd07e5a2afbb1767858262a5f037149a3/llvm/lib/Target/AArch64/AArch64.td#L719
* all of these have been available on 64-bit Apple Silicon from the
* beginning (the A7).
*/
OPENSSL_armcap_P |= ARMV7_NEON | ARMV8_PMULL | ARMV8_AES | ARMV8_SHA1 | ARMV8_SHA256;
/* More recent extensions are indicated by sysctls */
OPENSSL_armcap_P |= sysctl_query("hw.optional.armv8_2_sha512", ARMV8_SHA512);
OPENSSL_armcap_P |= sysctl_query("hw.optional.armv8_2_sha3", ARMV8_SHA3);
if (OPENSSL_armcap_P & ARMV8_SHA3) {
char uarch[64];
size_t len = sizeof(uarch);
if ((sysctlbyname("machdep.cpu.brand_string", uarch, &len, NULL, 0) == 0) &&
((strncmp(uarch, "Apple M1", 8) == 0) ||
(strncmp(uarch, "Apple M2", 8) == 0) ||
(strncmp(uarch, "Apple M3", 8) == 0))) {
OPENSSL_armcap_P |= ARMV8_UNROLL8_EOR3;
OPENSSL_armcap_P |= ARMV8_HAVE_SHA3_AND_WORTH_USING;
}
}
}
# endif /* __aarch64__ */
# elif defined(OSSL_IMPLEMENT_GETAUXVAL)
if (getauxval(OSSL_HWCAP) & OSSL_HWCAP_NEON) {
unsigned long hwcap = getauxval(OSSL_HWCAP_CE);
OPENSSL_armcap_P |= ARMV7_NEON;
if (hwcap & OSSL_HWCAP_CE_AES)
OPENSSL_armcap_P |= ARMV8_AES;
if (hwcap & OSSL_HWCAP_CE_PMULL)
OPENSSL_armcap_P |= ARMV8_PMULL;
if (hwcap & OSSL_HWCAP_CE_SHA1)
OPENSSL_armcap_P |= ARMV8_SHA1;
if (hwcap & OSSL_HWCAP_CE_SHA256)
OPENSSL_armcap_P |= ARMV8_SHA256;
# ifdef __aarch64__
if (hwcap & OSSL_HWCAP_CE_SM4)
OPENSSL_armcap_P |= ARMV8_SM4;
if (hwcap & OSSL_HWCAP_CE_SHA512)
OPENSSL_armcap_P |= ARMV8_SHA512;
if (hwcap & OSSL_HWCAP_CPUID)
OPENSSL_armcap_P |= ARMV8_CPUID;
if (hwcap & OSSL_HWCAP_CE_SM3)
OPENSSL_armcap_P |= ARMV8_SM3;
if (hwcap & OSSL_HWCAP_SHA3)
OPENSSL_armcap_P |= ARMV8_SHA3;
# endif
}
# ifdef __aarch64__
if (getauxval(OSSL_HWCAP) & OSSL_HWCAP_SVE)
OPENSSL_armcap_P |= ARMV8_SVE;
if (getauxval(OSSL_HWCAP2) & OSSL_HWCAP2_SVE2)
OPENSSL_armcap_P |= ARMV8_SVE2;
if (getauxval(OSSL_HWCAP2) & OSSL_HWCAP2_RNG)
OPENSSL_armcap_P |= ARMV8_RNG;
# endif
# else /* !__APPLE__ && !OSSL_IMPLEMENT_GETAUXVAL */
/* If all else fails, do brute force SIGILL-based feature detection */
sigfillset(&all_masked);
sigdelset(&all_masked, SIGILL);
sigdelset(&all_masked, SIGTRAP);
sigdelset(&all_masked, SIGFPE);
sigdelset(&all_masked, SIGBUS);
sigdelset(&all_masked, SIGSEGV);
memset(&ill_act, 0, sizeof(ill_act));
ill_act.sa_handler = ill_handler;
ill_act.sa_mask = all_masked;
sigprocmask(SIG_SETMASK, &ill_act.sa_mask, &oset);
sigaction(SIGILL, &ill_act, &ill_oact);
OPENSSL_armcap_P |= arm_probe_for(_armv7_neon_probe, ARMV7_NEON);
if (OPENSSL_armcap_P & ARMV7_NEON) {
OPENSSL_armcap_P |= arm_probe_for(_armv8_pmull_probe, ARMV8_PMULL | ARMV8_AES);
if (!(OPENSSL_armcap_P & ARMV8_AES)) {
OPENSSL_armcap_P |= arm_probe_for(_armv8_aes_probe, ARMV8_AES);
}
OPENSSL_armcap_P |= arm_probe_for(_armv8_sha1_probe, ARMV8_SHA1);
OPENSSL_armcap_P |= arm_probe_for(_armv8_sha256_probe, ARMV8_SHA256);
# if defined(__aarch64__)
OPENSSL_armcap_P |= arm_probe_for(_armv8_sm3_probe, ARMV8_SM3);
OPENSSL_armcap_P |= arm_probe_for(_armv8_sm4_probe, ARMV8_SM4);
OPENSSL_armcap_P |= arm_probe_for(_armv8_sha512_probe, ARMV8_SHA512);
OPENSSL_armcap_P |= arm_probe_for(_armv8_eor3_probe, ARMV8_SHA3);
# endif
}
# ifdef __aarch64__
OPENSSL_armcap_P |= arm_probe_for(_armv8_sve_probe, ARMV8_SVE);
OPENSSL_armcap_P |= arm_probe_for(_armv8_sve2_probe, ARMV8_SVE2);
OPENSSL_armcap_P |= arm_probe_for(_armv8_rng_probe, ARMV8_RNG);
# endif
/*
* Probing for ARMV7_TICK is known to produce unreliable results,
* so we only use the feature when the user explicitly enables it
* with OPENSSL_armcap.
*/
sigaction(SIGILL, &ill_oact, NULL);
sigprocmask(SIG_SETMASK, &oset, NULL);
# endif /* __APPLE__, OSSL_IMPLEMENT_GETAUXVAL */
# ifdef __aarch64__
if (OPENSSL_armcap_P & ARMV8_CPUID)
OPENSSL_arm_midr = _armv8_cpuid_probe();
if ((MIDR_IS_CPU_MODEL(OPENSSL_arm_midr, ARM_CPU_IMP_ARM, ARM_CPU_PART_CORTEX_A72) ||
MIDR_IS_CPU_MODEL(OPENSSL_arm_midr, ARM_CPU_IMP_ARM, ARM_CPU_PART_N1)) &&
(OPENSSL_armcap_P & ARMV7_NEON)) {
OPENSSL_armv8_rsa_neonized = 1;
}
if ((MIDR_IS_CPU_MODEL(OPENSSL_arm_midr, ARM_CPU_IMP_ARM, ARM_CPU_PART_V1) ||
MIDR_IS_CPU_MODEL(OPENSSL_arm_midr, ARM_CPU_IMP_ARM, ARM_CPU_PART_N2) ||
MIDR_IS_CPU_MODEL(OPENSSL_arm_midr, ARM_CPU_IMP_MICROSOFT, MICROSOFT_CPU_PART_COBALT_100) ||
MIDR_IS_CPU_MODEL(OPENSSL_arm_midr, ARM_CPU_IMP_ARM, ARM_CPU_PART_V2)) &&
(OPENSSL_armcap_P & ARMV8_SHA3))
OPENSSL_armcap_P |= ARMV8_UNROLL8_EOR3;
if ((MIDR_IS_CPU_MODEL(OPENSSL_arm_midr, ARM_CPU_IMP_ARM, ARM_CPU_PART_V1) ||
MIDR_IS_CPU_MODEL(OPENSSL_arm_midr, ARM_CPU_IMP_ARM, ARM_CPU_PART_V2)) &&
(OPENSSL_armcap_P & ARMV8_SHA3))
OPENSSL_armcap_P |= ARMV8_UNROLL12_EOR3;
if ((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_ICESTORM) ||
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_ICESTORM_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_M1_ICESTORM_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_BLIZZARD) ||
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_BLIZZARD_PRO) ||
MIDR_IS_CPU_MODEL(OPENSSL_arm_midr, ARM_CPU_IMP_APPLE, APPLE_CPU_PART_M2_AVALANCHE_MAX) ||
MIDR_IS_CPU_MODEL(OPENSSL_arm_midr, ARM_CPU_IMP_APPLE, APPLE_CPU_PART_M2_BLIZZARD_MAX)) &&
(OPENSSL_armcap_P & ARMV8_SHA3))
OPENSSL_armcap_P |= ARMV8_HAVE_SHA3_AND_WORTH_USING;
# endif
}
#endif /* _WIN32, __ARM_MAX_ARCH__ >= 7 */
Reference in New Issue View Git Blame Copy Permalink