boringssl/crypto/cpu-arm-linux.c

220 lines
5.6 KiB
C
Raw Normal View History

Rewrite ARM feature detection. This removes the thread-unsafe SIGILL-based detection and the multi-consumer-hostile CRYPTO_set_NEON_capable API. (Changing OPENSSL_armcap_P after initialization is likely to cause problems.) The right way to detect ARM features on Linux is getauxval. On aarch64, we should be able to rely on this, so use it straight. Split this out into its own file. The #ifdefs in the old cpu-arm.c meant it shared all but no code with its arm counterpart anyway. Unfortunately, various versions of Android have different missing APIs, so, on arm, we need a series of workarounds. Previously, we used a SIGILL fallback based on OpenSSL's logic, but this is inherently not thread-safe. (SIGILL also does not tell us if the OS knows how to save and restore NEON state.) Instead, base the behavior on Android NDK's cpu-features library, what Chromium currently uses with CRYPTO_set_NEON_capable: - Android before API level 20 does not provide getauxval. Where missing, we can read from /proc/self/auxv. - On some versions of Android, /proc/self/auxv is also not readable, so use /proc/cpuinfo's Features line. - Linux only advertises optional features in /proc/cpuinfo. ARMv8 makes NEON mandatory, so /proc/cpuinfo can't be used without additional effort. Finally, we must blacklist a particular chip because the NEON unit is broken (https://crbug.com/341598). Unfortunately, this means CRYPTO_library_init now depends on /proc being available, which will require some care with Chromium's sandbox. The simplest solution is to just call CRYPTO_library_init before entering the sandbox. It's worth noting that Chromium's current EnsureOpenSSLInit function already depends on /proc/cpuinfo to detect the broken CPU, by way of base::CPU. android_getCpuFeatures also interally depends on it. We were already relying on both of those being stateful and primed prior to entering the sandbox. BUG=chromium:589200 Change-Id: Ic5d1c341aab5a614eb129d8aa5ada2809edd6af8 Reviewed-on: https://boringssl-review.googlesource.com/7506 Reviewed-by: David Benjamin <davidben@google.com>
2016-03-01 22:35:47 +00:00
/* Copyright (c) 2016, Google Inc.
*
* Permission to use, copy, modify, and/or distribute this software for any
* purpose with or without fee is hereby granted, provided that the above
* copyright notice and this permission notice appear in all copies.
*
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY
* SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION
* OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN
* CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */
#include <openssl/cpu.h>
#if defined(OPENSSL_ARM) && !defined(OPENSSL_STATIC_ARMCAP)
Rewrite ARM feature detection. This removes the thread-unsafe SIGILL-based detection and the multi-consumer-hostile CRYPTO_set_NEON_capable API. (Changing OPENSSL_armcap_P after initialization is likely to cause problems.) The right way to detect ARM features on Linux is getauxval. On aarch64, we should be able to rely on this, so use it straight. Split this out into its own file. The #ifdefs in the old cpu-arm.c meant it shared all but no code with its arm counterpart anyway. Unfortunately, various versions of Android have different missing APIs, so, on arm, we need a series of workarounds. Previously, we used a SIGILL fallback based on OpenSSL's logic, but this is inherently not thread-safe. (SIGILL also does not tell us if the OS knows how to save and restore NEON state.) Instead, base the behavior on Android NDK's cpu-features library, what Chromium currently uses with CRYPTO_set_NEON_capable: - Android before API level 20 does not provide getauxval. Where missing, we can read from /proc/self/auxv. - On some versions of Android, /proc/self/auxv is also not readable, so use /proc/cpuinfo's Features line. - Linux only advertises optional features in /proc/cpuinfo. ARMv8 makes NEON mandatory, so /proc/cpuinfo can't be used without additional effort. Finally, we must blacklist a particular chip because the NEON unit is broken (https://crbug.com/341598). Unfortunately, this means CRYPTO_library_init now depends on /proc being available, which will require some care with Chromium's sandbox. The simplest solution is to just call CRYPTO_library_init before entering the sandbox. It's worth noting that Chromium's current EnsureOpenSSLInit function already depends on /proc/cpuinfo to detect the broken CPU, by way of base::CPU. android_getCpuFeatures also interally depends on it. We were already relying on both of those being stateful and primed prior to entering the sandbox. BUG=chromium:589200 Change-Id: Ic5d1c341aab5a614eb129d8aa5ada2809edd6af8 Reviewed-on: https://boringssl-review.googlesource.com/7506 Reviewed-by: David Benjamin <davidben@google.com>
2016-03-01 22:35:47 +00:00
#include <errno.h>
#include <fcntl.h>
#include <sys/types.h>
#include <unistd.h>
#include <openssl/arm_arch.h>
#include <openssl/buf.h>
#include <openssl/mem.h>
#include "cpu-arm-linux.h"
#define AT_HWCAP 16
#define AT_HWCAP2 26
// |getauxval| is not available on Android until API level 20. Link it as a weak
// symbol and use other methods as fallback.
unsigned long getauxval(unsigned long type) __attribute__((weak));
static int open_eintr(const char *path, int flags) {
int ret;
do {
ret = open(path, flags);
} while (ret < 0 && errno == EINTR);
return ret;
}
static ssize_t read_eintr(int fd, void *out, size_t len) {
ssize_t ret;
do {
ret = read(fd, out, len);
} while (ret < 0 && errno == EINTR);
return ret;
}
// read_full reads exactly |len| bytes from |fd| to |out|. On error or end of
// file, it returns zero.
static int read_full(int fd, void *out, size_t len) {
char *outp = out;
while (len > 0) {
ssize_t ret = read_eintr(fd, outp, len);
if (ret <= 0) {
return 0;
}
outp += ret;
len -= ret;
}
return 1;
}
// read_file opens |path| and reads until end-of-file. On success, it returns
// one and sets |*out_ptr| and |*out_len| to a newly-allocated buffer with the
// contents. Otherwise, it returns zero.
static int read_file(char **out_ptr, size_t *out_len, const char *path) {
int fd = open_eintr(path, O_RDONLY);
if (fd < 0) {
return 0;
}
static const size_t kReadSize = 1024;
int ret = 0;
size_t cap = kReadSize, len = 0;
char *buf = OPENSSL_malloc(cap);
if (buf == NULL) {
goto err;
}
for (;;) {
if (cap - len < kReadSize) {
size_t new_cap = cap * 2;
if (new_cap < cap) {
goto err;
}
char *new_buf = OPENSSL_realloc(buf, new_cap);
if (new_buf == NULL) {
goto err;
}
buf = new_buf;
cap = new_cap;
}
ssize_t bytes_read = read_eintr(fd, buf + len, kReadSize);
if (bytes_read < 0) {
goto err;
}
if (bytes_read == 0) {
break;
}
len += bytes_read;
}
*out_ptr = buf;
*out_len = len;
ret = 1;
buf = NULL;
err:
OPENSSL_free(buf);
close(fd);
return ret;
}
// getauxval_proc behaves like |getauxval| but reads from /proc/self/auxv.
static unsigned long getauxval_proc(unsigned long type) {
int fd = open_eintr("/proc/self/auxv", O_RDONLY);
if (fd < 0) {
return 0;
}
struct {
unsigned long tag;
unsigned long value;
} entry;
for (;;) {
if (!read_full(fd, &entry, sizeof(entry)) ||
(entry.tag == 0 && entry.value == 0)) {
break;
}
if (entry.tag == type) {
close(fd);
return entry.value;
}
}
close(fd);
return 0;
}
Rewrite ARM feature detection. This removes the thread-unsafe SIGILL-based detection and the multi-consumer-hostile CRYPTO_set_NEON_capable API. (Changing OPENSSL_armcap_P after initialization is likely to cause problems.) The right way to detect ARM features on Linux is getauxval. On aarch64, we should be able to rely on this, so use it straight. Split this out into its own file. The #ifdefs in the old cpu-arm.c meant it shared all but no code with its arm counterpart anyway. Unfortunately, various versions of Android have different missing APIs, so, on arm, we need a series of workarounds. Previously, we used a SIGILL fallback based on OpenSSL's logic, but this is inherently not thread-safe. (SIGILL also does not tell us if the OS knows how to save and restore NEON state.) Instead, base the behavior on Android NDK's cpu-features library, what Chromium currently uses with CRYPTO_set_NEON_capable: - Android before API level 20 does not provide getauxval. Where missing, we can read from /proc/self/auxv. - On some versions of Android, /proc/self/auxv is also not readable, so use /proc/cpuinfo's Features line. - Linux only advertises optional features in /proc/cpuinfo. ARMv8 makes NEON mandatory, so /proc/cpuinfo can't be used without additional effort. Finally, we must blacklist a particular chip because the NEON unit is broken (https://crbug.com/341598). Unfortunately, this means CRYPTO_library_init now depends on /proc being available, which will require some care with Chromium's sandbox. The simplest solution is to just call CRYPTO_library_init before entering the sandbox. It's worth noting that Chromium's current EnsureOpenSSLInit function already depends on /proc/cpuinfo to detect the broken CPU, by way of base::CPU. android_getCpuFeatures also interally depends on it. We were already relying on both of those being stateful and primed prior to entering the sandbox. BUG=chromium:589200 Change-Id: Ic5d1c341aab5a614eb129d8aa5ada2809edd6af8 Reviewed-on: https://boringssl-review.googlesource.com/7506 Reviewed-by: David Benjamin <davidben@google.com>
2016-03-01 22:35:47 +00:00
extern uint32_t OPENSSL_armcap_P;
static int g_has_broken_neon, g_needs_hwcap2_workaround;
Rewrite ARM feature detection. This removes the thread-unsafe SIGILL-based detection and the multi-consumer-hostile CRYPTO_set_NEON_capable API. (Changing OPENSSL_armcap_P after initialization is likely to cause problems.) The right way to detect ARM features on Linux is getauxval. On aarch64, we should be able to rely on this, so use it straight. Split this out into its own file. The #ifdefs in the old cpu-arm.c meant it shared all but no code with its arm counterpart anyway. Unfortunately, various versions of Android have different missing APIs, so, on arm, we need a series of workarounds. Previously, we used a SIGILL fallback based on OpenSSL's logic, but this is inherently not thread-safe. (SIGILL also does not tell us if the OS knows how to save and restore NEON state.) Instead, base the behavior on Android NDK's cpu-features library, what Chromium currently uses with CRYPTO_set_NEON_capable: - Android before API level 20 does not provide getauxval. Where missing, we can read from /proc/self/auxv. - On some versions of Android, /proc/self/auxv is also not readable, so use /proc/cpuinfo's Features line. - Linux only advertises optional features in /proc/cpuinfo. ARMv8 makes NEON mandatory, so /proc/cpuinfo can't be used without additional effort. Finally, we must blacklist a particular chip because the NEON unit is broken (https://crbug.com/341598). Unfortunately, this means CRYPTO_library_init now depends on /proc being available, which will require some care with Chromium's sandbox. The simplest solution is to just call CRYPTO_library_init before entering the sandbox. It's worth noting that Chromium's current EnsureOpenSSLInit function already depends on /proc/cpuinfo to detect the broken CPU, by way of base::CPU. android_getCpuFeatures also interally depends on it. We were already relying on both of those being stateful and primed prior to entering the sandbox. BUG=chromium:589200 Change-Id: Ic5d1c341aab5a614eb129d8aa5ada2809edd6af8 Reviewed-on: https://boringssl-review.googlesource.com/7506 Reviewed-by: David Benjamin <davidben@google.com>
2016-03-01 22:35:47 +00:00
void OPENSSL_cpuid_setup(void) {
char *cpuinfo_data;
size_t cpuinfo_len;
if (!read_file(&cpuinfo_data, &cpuinfo_len, "/proc/cpuinfo")) {
return;
}
STRING_PIECE cpuinfo;
cpuinfo.data = cpuinfo_data;
cpuinfo.len = cpuinfo_len;
// |getauxval| is not available on Android until API level 20. If it is
// unavailable, read from /proc/self/auxv as a fallback. This is unreadable
// on some versions of Android, so further fall back to /proc/cpuinfo.
//
// See
// https://android.googlesource.com/platform/ndk/+/882ac8f3392858991a0e1af33b4b7387ec856bd2
// and b/13679666 (Google-internal) for details.
Rewrite ARM feature detection. This removes the thread-unsafe SIGILL-based detection and the multi-consumer-hostile CRYPTO_set_NEON_capable API. (Changing OPENSSL_armcap_P after initialization is likely to cause problems.) The right way to detect ARM features on Linux is getauxval. On aarch64, we should be able to rely on this, so use it straight. Split this out into its own file. The #ifdefs in the old cpu-arm.c meant it shared all but no code with its arm counterpart anyway. Unfortunately, various versions of Android have different missing APIs, so, on arm, we need a series of workarounds. Previously, we used a SIGILL fallback based on OpenSSL's logic, but this is inherently not thread-safe. (SIGILL also does not tell us if the OS knows how to save and restore NEON state.) Instead, base the behavior on Android NDK's cpu-features library, what Chromium currently uses with CRYPTO_set_NEON_capable: - Android before API level 20 does not provide getauxval. Where missing, we can read from /proc/self/auxv. - On some versions of Android, /proc/self/auxv is also not readable, so use /proc/cpuinfo's Features line. - Linux only advertises optional features in /proc/cpuinfo. ARMv8 makes NEON mandatory, so /proc/cpuinfo can't be used without additional effort. Finally, we must blacklist a particular chip because the NEON unit is broken (https://crbug.com/341598). Unfortunately, this means CRYPTO_library_init now depends on /proc being available, which will require some care with Chromium's sandbox. The simplest solution is to just call CRYPTO_library_init before entering the sandbox. It's worth noting that Chromium's current EnsureOpenSSLInit function already depends on /proc/cpuinfo to detect the broken CPU, by way of base::CPU. android_getCpuFeatures also interally depends on it. We were already relying on both of those being stateful and primed prior to entering the sandbox. BUG=chromium:589200 Change-Id: Ic5d1c341aab5a614eb129d8aa5ada2809edd6af8 Reviewed-on: https://boringssl-review.googlesource.com/7506 Reviewed-by: David Benjamin <davidben@google.com>
2016-03-01 22:35:47 +00:00
unsigned long hwcap = 0;
if (getauxval != NULL) {
hwcap = getauxval(AT_HWCAP);
}
if (hwcap == 0) {
hwcap = getauxval_proc(AT_HWCAP);
}
if (hwcap == 0) {
hwcap = crypto_get_arm_hwcap_from_cpuinfo(&cpuinfo);
Rewrite ARM feature detection. This removes the thread-unsafe SIGILL-based detection and the multi-consumer-hostile CRYPTO_set_NEON_capable API. (Changing OPENSSL_armcap_P after initialization is likely to cause problems.) The right way to detect ARM features on Linux is getauxval. On aarch64, we should be able to rely on this, so use it straight. Split this out into its own file. The #ifdefs in the old cpu-arm.c meant it shared all but no code with its arm counterpart anyway. Unfortunately, various versions of Android have different missing APIs, so, on arm, we need a series of workarounds. Previously, we used a SIGILL fallback based on OpenSSL's logic, but this is inherently not thread-safe. (SIGILL also does not tell us if the OS knows how to save and restore NEON state.) Instead, base the behavior on Android NDK's cpu-features library, what Chromium currently uses with CRYPTO_set_NEON_capable: - Android before API level 20 does not provide getauxval. Where missing, we can read from /proc/self/auxv. - On some versions of Android, /proc/self/auxv is also not readable, so use /proc/cpuinfo's Features line. - Linux only advertises optional features in /proc/cpuinfo. ARMv8 makes NEON mandatory, so /proc/cpuinfo can't be used without additional effort. Finally, we must blacklist a particular chip because the NEON unit is broken (https://crbug.com/341598). Unfortunately, this means CRYPTO_library_init now depends on /proc being available, which will require some care with Chromium's sandbox. The simplest solution is to just call CRYPTO_library_init before entering the sandbox. It's worth noting that Chromium's current EnsureOpenSSLInit function already depends on /proc/cpuinfo to detect the broken CPU, by way of base::CPU. android_getCpuFeatures also interally depends on it. We were already relying on both of those being stateful and primed prior to entering the sandbox. BUG=chromium:589200 Change-Id: Ic5d1c341aab5a614eb129d8aa5ada2809edd6af8 Reviewed-on: https://boringssl-review.googlesource.com/7506 Reviewed-by: David Benjamin <davidben@google.com>
2016-03-01 22:35:47 +00:00
}
// Clear NEON support if known broken.
g_has_broken_neon = crypto_cpuinfo_has_broken_neon(&cpuinfo);
if (g_has_broken_neon) {
Rewrite ARM feature detection. This removes the thread-unsafe SIGILL-based detection and the multi-consumer-hostile CRYPTO_set_NEON_capable API. (Changing OPENSSL_armcap_P after initialization is likely to cause problems.) The right way to detect ARM features on Linux is getauxval. On aarch64, we should be able to rely on this, so use it straight. Split this out into its own file. The #ifdefs in the old cpu-arm.c meant it shared all but no code with its arm counterpart anyway. Unfortunately, various versions of Android have different missing APIs, so, on arm, we need a series of workarounds. Previously, we used a SIGILL fallback based on OpenSSL's logic, but this is inherently not thread-safe. (SIGILL also does not tell us if the OS knows how to save and restore NEON state.) Instead, base the behavior on Android NDK's cpu-features library, what Chromium currently uses with CRYPTO_set_NEON_capable: - Android before API level 20 does not provide getauxval. Where missing, we can read from /proc/self/auxv. - On some versions of Android, /proc/self/auxv is also not readable, so use /proc/cpuinfo's Features line. - Linux only advertises optional features in /proc/cpuinfo. ARMv8 makes NEON mandatory, so /proc/cpuinfo can't be used without additional effort. Finally, we must blacklist a particular chip because the NEON unit is broken (https://crbug.com/341598). Unfortunately, this means CRYPTO_library_init now depends on /proc being available, which will require some care with Chromium's sandbox. The simplest solution is to just call CRYPTO_library_init before entering the sandbox. It's worth noting that Chromium's current EnsureOpenSSLInit function already depends on /proc/cpuinfo to detect the broken CPU, by way of base::CPU. android_getCpuFeatures also interally depends on it. We were already relying on both of those being stateful and primed prior to entering the sandbox. BUG=chromium:589200 Change-Id: Ic5d1c341aab5a614eb129d8aa5ada2809edd6af8 Reviewed-on: https://boringssl-review.googlesource.com/7506 Reviewed-by: David Benjamin <davidben@google.com>
2016-03-01 22:35:47 +00:00
hwcap &= ~HWCAP_NEON;
}
// Matching OpenSSL, only report other features if NEON is present.
Rewrite ARM feature detection. This removes the thread-unsafe SIGILL-based detection and the multi-consumer-hostile CRYPTO_set_NEON_capable API. (Changing OPENSSL_armcap_P after initialization is likely to cause problems.) The right way to detect ARM features on Linux is getauxval. On aarch64, we should be able to rely on this, so use it straight. Split this out into its own file. The #ifdefs in the old cpu-arm.c meant it shared all but no code with its arm counterpart anyway. Unfortunately, various versions of Android have different missing APIs, so, on arm, we need a series of workarounds. Previously, we used a SIGILL fallback based on OpenSSL's logic, but this is inherently not thread-safe. (SIGILL also does not tell us if the OS knows how to save and restore NEON state.) Instead, base the behavior on Android NDK's cpu-features library, what Chromium currently uses with CRYPTO_set_NEON_capable: - Android before API level 20 does not provide getauxval. Where missing, we can read from /proc/self/auxv. - On some versions of Android, /proc/self/auxv is also not readable, so use /proc/cpuinfo's Features line. - Linux only advertises optional features in /proc/cpuinfo. ARMv8 makes NEON mandatory, so /proc/cpuinfo can't be used without additional effort. Finally, we must blacklist a particular chip because the NEON unit is broken (https://crbug.com/341598). Unfortunately, this means CRYPTO_library_init now depends on /proc being available, which will require some care with Chromium's sandbox. The simplest solution is to just call CRYPTO_library_init before entering the sandbox. It's worth noting that Chromium's current EnsureOpenSSLInit function already depends on /proc/cpuinfo to detect the broken CPU, by way of base::CPU. android_getCpuFeatures also interally depends on it. We were already relying on both of those being stateful and primed prior to entering the sandbox. BUG=chromium:589200 Change-Id: Ic5d1c341aab5a614eb129d8aa5ada2809edd6af8 Reviewed-on: https://boringssl-review.googlesource.com/7506 Reviewed-by: David Benjamin <davidben@google.com>
2016-03-01 22:35:47 +00:00
if (hwcap & HWCAP_NEON) {
OPENSSL_armcap_P |= ARMV7_NEON;
// Some ARMv8 Android devices don't expose AT_HWCAP2. Fall back to
// /proc/cpuinfo. See https://crbug.com/596156.
Work around Android devices without AT_HWCAP2. Some ARMv8 Android devices don't have AT_HWCAP2. This means, when running in 32-bit mode (ARM capability APIs on Linux are different between AArch32 and AArch64), we can't discover the various nice instructions. On a Nexus 6P, this gives a, uh, minor performance win when running in 32-bit mode. Before: Did 1085000 AES-128-GCM (16 bytes) seal operations in 1000003us (1084996.7 ops/sec): 17.4 MB/s Did 60000 AES-128-GCM (1350 bytes) seal operations in 1013416us (59205.7 ops/sec): 79.9 MB/s Did 11000 AES-128-GCM (8192 bytes) seal operations in 1019778us (10786.7 ops/sec): 88.4 MB/s Did 1009000 AES-256-GCM (16 bytes) seal operations in 1000650us (1008344.6 ops/sec): 16.1 MB/s Did 49000 AES-256-GCM (1350 bytes) seal operations in 1015698us (48242.7 ops/sec): 65.1 MB/s Did 9394 AES-256-GCM (8192 bytes) seal operations in 1071104us (8770.4 ops/sec): 71.8 MB/s Did 1557000 SHA-1 (16 bytes) operations in 1000317us (1556506.6 ops/sec): 24.9 MB/s Did 762000 SHA-1 (256 bytes) operations in 1000527us (761598.6 ops/sec): 195.0 MB/s Did 45000 SHA-1 (8192 bytes) operations in 1013773us (44388.6 ops/sec): 363.6 MB/s Did 1459000 SHA-256 (16 bytes) operations in 1000271us (1458604.7 ops/sec): 23.3 MB/s Did 538000 SHA-256 (256 bytes) operations in 1000990us (537467.9 ops/sec): 137.6 MB/s Did 26000 SHA-256 (8192 bytes) operations in 1008403us (25783.3 ops/sec): 211.2 MB/s After: Did 1890000 AES-128-GCM (16 bytes) seal operations in 1000068us (1889871.5 ops/sec): 30.2 MB/s Did 509000 AES-128-GCM (1350 bytes) seal operations in 1000112us (508943.0 ops/sec): 687.1 MB/s Did 110000 AES-128-GCM (8192 bytes) seal operations in 1007966us (109130.7 ops/sec): 894.0 MB/s Did 1960000 AES-256-GCM (16 bytes) seal operations in 1000303us (1959406.3 ops/sec): 31.4 MB/s Did 460000 AES-256-GCM (1350 bytes) seal operations in 1001873us (459140.0 ops/sec): 619.8 MB/s Did 97000 AES-256-GCM (8192 bytes) seal operations in 1005337us (96485.1 ops/sec): 790.4 MB/s Did 1927000 SHA-1 (16 bytes) operations in 1000429us (1926173.7 ops/sec): 30.8 MB/s Did 1151000 SHA-1 (256 bytes) operations in 1000425us (1150511.0 ops/sec): 294.5 MB/s Did 87000 SHA-1 (8192 bytes) operations in 1003089us (86732.1 ops/sec): 710.5 MB/s Did 2357390 SHA-256 (16 bytes) operations in 1000116us (2357116.6 ops/sec): 37.7 MB/s Did 1410000 SHA-256 (256 bytes) operations in 1000176us (1409751.9 ops/sec): 360.9 MB/s Did 101000 SHA-256 (8192 bytes) operations in 1007007us (100297.2 ops/sec): 821.6 MB/s BUG=chromium:596156 Change-Id: Iacc1f8d8a07e991d4615f2e12c5c54923fb31aa2 Reviewed-on: https://boringssl-review.googlesource.com/7507 Reviewed-by: David Benjamin <davidben@google.com>
2016-03-20 21:53:34 +00:00
unsigned long hwcap2 = 0;
Rewrite ARM feature detection. This removes the thread-unsafe SIGILL-based detection and the multi-consumer-hostile CRYPTO_set_NEON_capable API. (Changing OPENSSL_armcap_P after initialization is likely to cause problems.) The right way to detect ARM features on Linux is getauxval. On aarch64, we should be able to rely on this, so use it straight. Split this out into its own file. The #ifdefs in the old cpu-arm.c meant it shared all but no code with its arm counterpart anyway. Unfortunately, various versions of Android have different missing APIs, so, on arm, we need a series of workarounds. Previously, we used a SIGILL fallback based on OpenSSL's logic, but this is inherently not thread-safe. (SIGILL also does not tell us if the OS knows how to save and restore NEON state.) Instead, base the behavior on Android NDK's cpu-features library, what Chromium currently uses with CRYPTO_set_NEON_capable: - Android before API level 20 does not provide getauxval. Where missing, we can read from /proc/self/auxv. - On some versions of Android, /proc/self/auxv is also not readable, so use /proc/cpuinfo's Features line. - Linux only advertises optional features in /proc/cpuinfo. ARMv8 makes NEON mandatory, so /proc/cpuinfo can't be used without additional effort. Finally, we must blacklist a particular chip because the NEON unit is broken (https://crbug.com/341598). Unfortunately, this means CRYPTO_library_init now depends on /proc being available, which will require some care with Chromium's sandbox. The simplest solution is to just call CRYPTO_library_init before entering the sandbox. It's worth noting that Chromium's current EnsureOpenSSLInit function already depends on /proc/cpuinfo to detect the broken CPU, by way of base::CPU. android_getCpuFeatures also interally depends on it. We were already relying on both of those being stateful and primed prior to entering the sandbox. BUG=chromium:589200 Change-Id: Ic5d1c341aab5a614eb129d8aa5ada2809edd6af8 Reviewed-on: https://boringssl-review.googlesource.com/7506 Reviewed-by: David Benjamin <davidben@google.com>
2016-03-01 22:35:47 +00:00
if (getauxval != NULL) {
Work around Android devices without AT_HWCAP2. Some ARMv8 Android devices don't have AT_HWCAP2. This means, when running in 32-bit mode (ARM capability APIs on Linux are different between AArch32 and AArch64), we can't discover the various nice instructions. On a Nexus 6P, this gives a, uh, minor performance win when running in 32-bit mode. Before: Did 1085000 AES-128-GCM (16 bytes) seal operations in 1000003us (1084996.7 ops/sec): 17.4 MB/s Did 60000 AES-128-GCM (1350 bytes) seal operations in 1013416us (59205.7 ops/sec): 79.9 MB/s Did 11000 AES-128-GCM (8192 bytes) seal operations in 1019778us (10786.7 ops/sec): 88.4 MB/s Did 1009000 AES-256-GCM (16 bytes) seal operations in 1000650us (1008344.6 ops/sec): 16.1 MB/s Did 49000 AES-256-GCM (1350 bytes) seal operations in 1015698us (48242.7 ops/sec): 65.1 MB/s Did 9394 AES-256-GCM (8192 bytes) seal operations in 1071104us (8770.4 ops/sec): 71.8 MB/s Did 1557000 SHA-1 (16 bytes) operations in 1000317us (1556506.6 ops/sec): 24.9 MB/s Did 762000 SHA-1 (256 bytes) operations in 1000527us (761598.6 ops/sec): 195.0 MB/s Did 45000 SHA-1 (8192 bytes) operations in 1013773us (44388.6 ops/sec): 363.6 MB/s Did 1459000 SHA-256 (16 bytes) operations in 1000271us (1458604.7 ops/sec): 23.3 MB/s Did 538000 SHA-256 (256 bytes) operations in 1000990us (537467.9 ops/sec): 137.6 MB/s Did 26000 SHA-256 (8192 bytes) operations in 1008403us (25783.3 ops/sec): 211.2 MB/s After: Did 1890000 AES-128-GCM (16 bytes) seal operations in 1000068us (1889871.5 ops/sec): 30.2 MB/s Did 509000 AES-128-GCM (1350 bytes) seal operations in 1000112us (508943.0 ops/sec): 687.1 MB/s Did 110000 AES-128-GCM (8192 bytes) seal operations in 1007966us (109130.7 ops/sec): 894.0 MB/s Did 1960000 AES-256-GCM (16 bytes) seal operations in 1000303us (1959406.3 ops/sec): 31.4 MB/s Did 460000 AES-256-GCM (1350 bytes) seal operations in 1001873us (459140.0 ops/sec): 619.8 MB/s Did 97000 AES-256-GCM (8192 bytes) seal operations in 1005337us (96485.1 ops/sec): 790.4 MB/s Did 1927000 SHA-1 (16 bytes) operations in 1000429us (1926173.7 ops/sec): 30.8 MB/s Did 1151000 SHA-1 (256 bytes) operations in 1000425us (1150511.0 ops/sec): 294.5 MB/s Did 87000 SHA-1 (8192 bytes) operations in 1003089us (86732.1 ops/sec): 710.5 MB/s Did 2357390 SHA-256 (16 bytes) operations in 1000116us (2357116.6 ops/sec): 37.7 MB/s Did 1410000 SHA-256 (256 bytes) operations in 1000176us (1409751.9 ops/sec): 360.9 MB/s Did 101000 SHA-256 (8192 bytes) operations in 1007007us (100297.2 ops/sec): 821.6 MB/s BUG=chromium:596156 Change-Id: Iacc1f8d8a07e991d4615f2e12c5c54923fb31aa2 Reviewed-on: https://boringssl-review.googlesource.com/7507 Reviewed-by: David Benjamin <davidben@google.com>
2016-03-20 21:53:34 +00:00
hwcap2 = getauxval(AT_HWCAP2);
}
if (hwcap2 == 0) {
hwcap2 = crypto_get_arm_hwcap2_from_cpuinfo(&cpuinfo);
g_needs_hwcap2_workaround = hwcap2 != 0;
Work around Android devices without AT_HWCAP2. Some ARMv8 Android devices don't have AT_HWCAP2. This means, when running in 32-bit mode (ARM capability APIs on Linux are different between AArch32 and AArch64), we can't discover the various nice instructions. On a Nexus 6P, this gives a, uh, minor performance win when running in 32-bit mode. Before: Did 1085000 AES-128-GCM (16 bytes) seal operations in 1000003us (1084996.7 ops/sec): 17.4 MB/s Did 60000 AES-128-GCM (1350 bytes) seal operations in 1013416us (59205.7 ops/sec): 79.9 MB/s Did 11000 AES-128-GCM (8192 bytes) seal operations in 1019778us (10786.7 ops/sec): 88.4 MB/s Did 1009000 AES-256-GCM (16 bytes) seal operations in 1000650us (1008344.6 ops/sec): 16.1 MB/s Did 49000 AES-256-GCM (1350 bytes) seal operations in 1015698us (48242.7 ops/sec): 65.1 MB/s Did 9394 AES-256-GCM (8192 bytes) seal operations in 1071104us (8770.4 ops/sec): 71.8 MB/s Did 1557000 SHA-1 (16 bytes) operations in 1000317us (1556506.6 ops/sec): 24.9 MB/s Did 762000 SHA-1 (256 bytes) operations in 1000527us (761598.6 ops/sec): 195.0 MB/s Did 45000 SHA-1 (8192 bytes) operations in 1013773us (44388.6 ops/sec): 363.6 MB/s Did 1459000 SHA-256 (16 bytes) operations in 1000271us (1458604.7 ops/sec): 23.3 MB/s Did 538000 SHA-256 (256 bytes) operations in 1000990us (537467.9 ops/sec): 137.6 MB/s Did 26000 SHA-256 (8192 bytes) operations in 1008403us (25783.3 ops/sec): 211.2 MB/s After: Did 1890000 AES-128-GCM (16 bytes) seal operations in 1000068us (1889871.5 ops/sec): 30.2 MB/s Did 509000 AES-128-GCM (1350 bytes) seal operations in 1000112us (508943.0 ops/sec): 687.1 MB/s Did 110000 AES-128-GCM (8192 bytes) seal operations in 1007966us (109130.7 ops/sec): 894.0 MB/s Did 1960000 AES-256-GCM (16 bytes) seal operations in 1000303us (1959406.3 ops/sec): 31.4 MB/s Did 460000 AES-256-GCM (1350 bytes) seal operations in 1001873us (459140.0 ops/sec): 619.8 MB/s Did 97000 AES-256-GCM (8192 bytes) seal operations in 1005337us (96485.1 ops/sec): 790.4 MB/s Did 1927000 SHA-1 (16 bytes) operations in 1000429us (1926173.7 ops/sec): 30.8 MB/s Did 1151000 SHA-1 (256 bytes) operations in 1000425us (1150511.0 ops/sec): 294.5 MB/s Did 87000 SHA-1 (8192 bytes) operations in 1003089us (86732.1 ops/sec): 710.5 MB/s Did 2357390 SHA-256 (16 bytes) operations in 1000116us (2357116.6 ops/sec): 37.7 MB/s Did 1410000 SHA-256 (256 bytes) operations in 1000176us (1409751.9 ops/sec): 360.9 MB/s Did 101000 SHA-256 (8192 bytes) operations in 1007007us (100297.2 ops/sec): 821.6 MB/s BUG=chromium:596156 Change-Id: Iacc1f8d8a07e991d4615f2e12c5c54923fb31aa2 Reviewed-on: https://boringssl-review.googlesource.com/7507 Reviewed-by: David Benjamin <davidben@google.com>
2016-03-20 21:53:34 +00:00
}
Rewrite ARM feature detection. This removes the thread-unsafe SIGILL-based detection and the multi-consumer-hostile CRYPTO_set_NEON_capable API. (Changing OPENSSL_armcap_P after initialization is likely to cause problems.) The right way to detect ARM features on Linux is getauxval. On aarch64, we should be able to rely on this, so use it straight. Split this out into its own file. The #ifdefs in the old cpu-arm.c meant it shared all but no code with its arm counterpart anyway. Unfortunately, various versions of Android have different missing APIs, so, on arm, we need a series of workarounds. Previously, we used a SIGILL fallback based on OpenSSL's logic, but this is inherently not thread-safe. (SIGILL also does not tell us if the OS knows how to save and restore NEON state.) Instead, base the behavior on Android NDK's cpu-features library, what Chromium currently uses with CRYPTO_set_NEON_capable: - Android before API level 20 does not provide getauxval. Where missing, we can read from /proc/self/auxv. - On some versions of Android, /proc/self/auxv is also not readable, so use /proc/cpuinfo's Features line. - Linux only advertises optional features in /proc/cpuinfo. ARMv8 makes NEON mandatory, so /proc/cpuinfo can't be used without additional effort. Finally, we must blacklist a particular chip because the NEON unit is broken (https://crbug.com/341598). Unfortunately, this means CRYPTO_library_init now depends on /proc being available, which will require some care with Chromium's sandbox. The simplest solution is to just call CRYPTO_library_init before entering the sandbox. It's worth noting that Chromium's current EnsureOpenSSLInit function already depends on /proc/cpuinfo to detect the broken CPU, by way of base::CPU. android_getCpuFeatures also interally depends on it. We were already relying on both of those being stateful and primed prior to entering the sandbox. BUG=chromium:589200 Change-Id: Ic5d1c341aab5a614eb129d8aa5ada2809edd6af8 Reviewed-on: https://boringssl-review.googlesource.com/7506 Reviewed-by: David Benjamin <davidben@google.com>
2016-03-01 22:35:47 +00:00
Work around Android devices without AT_HWCAP2. Some ARMv8 Android devices don't have AT_HWCAP2. This means, when running in 32-bit mode (ARM capability APIs on Linux are different between AArch32 and AArch64), we can't discover the various nice instructions. On a Nexus 6P, this gives a, uh, minor performance win when running in 32-bit mode. Before: Did 1085000 AES-128-GCM (16 bytes) seal operations in 1000003us (1084996.7 ops/sec): 17.4 MB/s Did 60000 AES-128-GCM (1350 bytes) seal operations in 1013416us (59205.7 ops/sec): 79.9 MB/s Did 11000 AES-128-GCM (8192 bytes) seal operations in 1019778us (10786.7 ops/sec): 88.4 MB/s Did 1009000 AES-256-GCM (16 bytes) seal operations in 1000650us (1008344.6 ops/sec): 16.1 MB/s Did 49000 AES-256-GCM (1350 bytes) seal operations in 1015698us (48242.7 ops/sec): 65.1 MB/s Did 9394 AES-256-GCM (8192 bytes) seal operations in 1071104us (8770.4 ops/sec): 71.8 MB/s Did 1557000 SHA-1 (16 bytes) operations in 1000317us (1556506.6 ops/sec): 24.9 MB/s Did 762000 SHA-1 (256 bytes) operations in 1000527us (761598.6 ops/sec): 195.0 MB/s Did 45000 SHA-1 (8192 bytes) operations in 1013773us (44388.6 ops/sec): 363.6 MB/s Did 1459000 SHA-256 (16 bytes) operations in 1000271us (1458604.7 ops/sec): 23.3 MB/s Did 538000 SHA-256 (256 bytes) operations in 1000990us (537467.9 ops/sec): 137.6 MB/s Did 26000 SHA-256 (8192 bytes) operations in 1008403us (25783.3 ops/sec): 211.2 MB/s After: Did 1890000 AES-128-GCM (16 bytes) seal operations in 1000068us (1889871.5 ops/sec): 30.2 MB/s Did 509000 AES-128-GCM (1350 bytes) seal operations in 1000112us (508943.0 ops/sec): 687.1 MB/s Did 110000 AES-128-GCM (8192 bytes) seal operations in 1007966us (109130.7 ops/sec): 894.0 MB/s Did 1960000 AES-256-GCM (16 bytes) seal operations in 1000303us (1959406.3 ops/sec): 31.4 MB/s Did 460000 AES-256-GCM (1350 bytes) seal operations in 1001873us (459140.0 ops/sec): 619.8 MB/s Did 97000 AES-256-GCM (8192 bytes) seal operations in 1005337us (96485.1 ops/sec): 790.4 MB/s Did 1927000 SHA-1 (16 bytes) operations in 1000429us (1926173.7 ops/sec): 30.8 MB/s Did 1151000 SHA-1 (256 bytes) operations in 1000425us (1150511.0 ops/sec): 294.5 MB/s Did 87000 SHA-1 (8192 bytes) operations in 1003089us (86732.1 ops/sec): 710.5 MB/s Did 2357390 SHA-256 (16 bytes) operations in 1000116us (2357116.6 ops/sec): 37.7 MB/s Did 1410000 SHA-256 (256 bytes) operations in 1000176us (1409751.9 ops/sec): 360.9 MB/s Did 101000 SHA-256 (8192 bytes) operations in 1007007us (100297.2 ops/sec): 821.6 MB/s BUG=chromium:596156 Change-Id: Iacc1f8d8a07e991d4615f2e12c5c54923fb31aa2 Reviewed-on: https://boringssl-review.googlesource.com/7507 Reviewed-by: David Benjamin <davidben@google.com>
2016-03-20 21:53:34 +00:00
if (hwcap2 & HWCAP2_AES) {
OPENSSL_armcap_P |= ARMV8_AES;
}
if (hwcap2 & HWCAP2_PMULL) {
OPENSSL_armcap_P |= ARMV8_PMULL;
}
if (hwcap2 & HWCAP2_SHA1) {
OPENSSL_armcap_P |= ARMV8_SHA1;
}
if (hwcap2 & HWCAP2_SHA2) {
OPENSSL_armcap_P |= ARMV8_SHA256;
Rewrite ARM feature detection. This removes the thread-unsafe SIGILL-based detection and the multi-consumer-hostile CRYPTO_set_NEON_capable API. (Changing OPENSSL_armcap_P after initialization is likely to cause problems.) The right way to detect ARM features on Linux is getauxval. On aarch64, we should be able to rely on this, so use it straight. Split this out into its own file. The #ifdefs in the old cpu-arm.c meant it shared all but no code with its arm counterpart anyway. Unfortunately, various versions of Android have different missing APIs, so, on arm, we need a series of workarounds. Previously, we used a SIGILL fallback based on OpenSSL's logic, but this is inherently not thread-safe. (SIGILL also does not tell us if the OS knows how to save and restore NEON state.) Instead, base the behavior on Android NDK's cpu-features library, what Chromium currently uses with CRYPTO_set_NEON_capable: - Android before API level 20 does not provide getauxval. Where missing, we can read from /proc/self/auxv. - On some versions of Android, /proc/self/auxv is also not readable, so use /proc/cpuinfo's Features line. - Linux only advertises optional features in /proc/cpuinfo. ARMv8 makes NEON mandatory, so /proc/cpuinfo can't be used without additional effort. Finally, we must blacklist a particular chip because the NEON unit is broken (https://crbug.com/341598). Unfortunately, this means CRYPTO_library_init now depends on /proc being available, which will require some care with Chromium's sandbox. The simplest solution is to just call CRYPTO_library_init before entering the sandbox. It's worth noting that Chromium's current EnsureOpenSSLInit function already depends on /proc/cpuinfo to detect the broken CPU, by way of base::CPU. android_getCpuFeatures also interally depends on it. We were already relying on both of those being stateful and primed prior to entering the sandbox. BUG=chromium:589200 Change-Id: Ic5d1c341aab5a614eb129d8aa5ada2809edd6af8 Reviewed-on: https://boringssl-review.googlesource.com/7506 Reviewed-by: David Benjamin <davidben@google.com>
2016-03-01 22:35:47 +00:00
}
}
OPENSSL_free(cpuinfo_data);
}
int CRYPTO_has_broken_NEON(void) { return g_has_broken_neon; }
int CRYPTO_needs_hwcap2_workaround(void) { return g_needs_hwcap2_workaround; }
#endif // OPENSSL_ARM && !OPENSSL_STATIC_ARMCAP