2016-10-25 01:02:26 +01:00
|
|
|
/* 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/base.h>
|
|
|
|
|
|
|
|
#include <stdio.h>
|
|
|
|
#include <string.h>
|
|
|
|
|
2017-06-09 01:04:08 +01:00
|
|
|
#include <gtest/gtest.h>
|
|
|
|
|
2016-10-25 01:02:26 +01:00
|
|
|
#include <openssl/bn.h>
|
2018-11-06 23:18:56 +00:00
|
|
|
#include <openssl/cpu.h>
|
|
|
|
#include <openssl/ec.h>
|
2016-10-25 01:02:26 +01:00
|
|
|
#include <openssl/mem.h>
|
2018-11-06 23:18:56 +00:00
|
|
|
#include <openssl/nid.h>
|
2016-10-25 01:02:26 +01:00
|
|
|
|
2018-11-06 23:18:56 +00:00
|
|
|
#include "internal.h"
|
2017-05-02 22:25:39 +01:00
|
|
|
#include "../bn/internal.h"
|
2017-08-08 23:19:07 +01:00
|
|
|
#include "../../internal.h"
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
#include "../../test/abi_test.h"
|
2017-05-02 22:25:39 +01:00
|
|
|
#include "../../test/file_test.h"
|
2017-06-09 01:04:08 +01:00
|
|
|
#include "../../test/test_util.h"
|
2016-10-25 01:02:26 +01:00
|
|
|
#include "p256-x86_64.h"
|
|
|
|
|
|
|
|
|
|
|
|
// Disable tests if BORINGSSL_SHARED_LIBRARY is defined. These tests need access
|
|
|
|
// to internal functions.
|
|
|
|
#if !defined(OPENSSL_NO_ASM) && defined(OPENSSL_X86_64) && \
|
|
|
|
!defined(OPENSSL_SMALL) && !defined(BORINGSSL_SHARED_LIBRARY)
|
|
|
|
|
2017-06-09 01:04:08 +01:00
|
|
|
TEST(P256_X86_64Test, SelectW5) {
|
2016-10-25 01:02:26 +01:00
|
|
|
// Fill a table with some garbage input.
|
2017-08-08 23:19:07 +01:00
|
|
|
alignas(64) P256_POINT table[16];
|
2016-10-25 01:02:26 +01:00
|
|
|
for (size_t i = 0; i < 16; i++) {
|
2016-12-13 06:07:13 +00:00
|
|
|
OPENSSL_memset(table[i].X, 3 * i, sizeof(table[i].X));
|
|
|
|
OPENSSL_memset(table[i].Y, 3 * i + 1, sizeof(table[i].Y));
|
|
|
|
OPENSSL_memset(table[i].Z, 3 * i + 2, sizeof(table[i].Z));
|
2016-10-25 01:02:26 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
for (int i = 0; i <= 16; i++) {
|
|
|
|
P256_POINT val;
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_select_w5, &val, table, i);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
|
|
|
P256_POINT expected;
|
|
|
|
if (i == 0) {
|
2016-12-13 06:07:13 +00:00
|
|
|
OPENSSL_memset(&expected, 0, sizeof(expected));
|
2016-10-25 01:02:26 +01:00
|
|
|
} else {
|
|
|
|
expected = table[i-1];
|
|
|
|
}
|
|
|
|
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_EQ(Bytes(reinterpret_cast<const char *>(&expected), sizeof(expected)),
|
|
|
|
Bytes(reinterpret_cast<const char *>(&val), sizeof(val)));
|
2016-10-25 01:02:26 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2017-06-09 01:04:08 +01:00
|
|
|
TEST(P256_X86_64Test, SelectW7) {
|
2016-10-25 01:02:26 +01:00
|
|
|
// Fill a table with some garbage input.
|
2017-08-08 23:19:07 +01:00
|
|
|
alignas(64) P256_POINT_AFFINE table[64];
|
2016-10-25 01:02:26 +01:00
|
|
|
for (size_t i = 0; i < 64; i++) {
|
2016-12-13 06:07:13 +00:00
|
|
|
OPENSSL_memset(table[i].X, 2 * i, sizeof(table[i].X));
|
|
|
|
OPENSSL_memset(table[i].Y, 2 * i + 1, sizeof(table[i].Y));
|
2016-10-25 01:02:26 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
for (int i = 0; i <= 64; i++) {
|
|
|
|
P256_POINT_AFFINE val;
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_select_w7, &val, table, i);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
|
|
|
P256_POINT_AFFINE expected;
|
|
|
|
if (i == 0) {
|
2016-12-13 06:07:13 +00:00
|
|
|
OPENSSL_memset(&expected, 0, sizeof(expected));
|
2016-10-25 01:02:26 +01:00
|
|
|
} else {
|
|
|
|
expected = table[i-1];
|
|
|
|
}
|
|
|
|
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_EQ(Bytes(reinterpret_cast<const char *>(&expected), sizeof(expected)),
|
|
|
|
Bytes(reinterpret_cast<const char *>(&val), sizeof(val)));
|
2016-10-25 01:02:26 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2018-11-06 23:18:56 +00:00
|
|
|
TEST(P256_X86_64Test, BEEU) {
|
|
|
|
if ((OPENSSL_ia32cap_P[1] & (1 << 28)) == 0) {
|
|
|
|
// No AVX support; cannot run the BEEU code.
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
bssl::UniquePtr<EC_GROUP> group(
|
|
|
|
EC_GROUP_new_by_curve_name(NID_X9_62_prime256v1));
|
|
|
|
ASSERT_TRUE(group);
|
|
|
|
|
|
|
|
BN_ULONG order_words[P256_LIMBS];
|
|
|
|
ASSERT_TRUE(
|
|
|
|
bn_copy_words(order_words, P256_LIMBS, EC_GROUP_get0_order(group.get())));
|
|
|
|
|
|
|
|
BN_ULONG in[P256_LIMBS], out[P256_LIMBS];
|
|
|
|
EC_SCALAR in_scalar, out_scalar, result;
|
|
|
|
OPENSSL_memset(in, 0, sizeof(in));
|
|
|
|
|
|
|
|
// Trying to find the inverse of zero should fail.
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
ASSERT_FALSE(CHECK_ABI(beeu_mod_inverse_vartime, out, in, order_words));
|
2018-11-06 23:18:56 +00:00
|
|
|
|
|
|
|
// kOneMont is 1, in Montgomery form.
|
|
|
|
static const BN_ULONG kOneMont[P256_LIMBS] = {
|
|
|
|
TOBN(0xc46353d, 0x039cdaaf), TOBN(0x43190552, 0x58e8617b),
|
|
|
|
0, 0xffffffff,
|
|
|
|
};
|
|
|
|
|
|
|
|
for (BN_ULONG i = 1; i < 2000; i++) {
|
|
|
|
SCOPED_TRACE(i);
|
|
|
|
|
|
|
|
in[0] = i;
|
|
|
|
if (i >= 1000) {
|
|
|
|
in[1] = i << 8;
|
|
|
|
in[2] = i << 32;
|
|
|
|
in[3] = i << 48;
|
|
|
|
} else {
|
|
|
|
in[1] = in[2] = in[3] = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
EXPECT_TRUE(bn_less_than_words(in, order_words, P256_LIMBS));
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
ASSERT_TRUE(CHECK_ABI(beeu_mod_inverse_vartime, out, in, order_words));
|
2018-11-06 23:18:56 +00:00
|
|
|
EXPECT_TRUE(bn_less_than_words(out, order_words, P256_LIMBS));
|
|
|
|
|
|
|
|
// Calculate out*in and confirm that it equals one, modulo the order.
|
|
|
|
OPENSSL_memcpy(in_scalar.bytes, in, sizeof(in));
|
|
|
|
OPENSSL_memcpy(out_scalar.bytes, out, sizeof(out));
|
|
|
|
ec_scalar_to_montgomery(group.get(), &in_scalar, &in_scalar);
|
|
|
|
ec_scalar_to_montgomery(group.get(), &out_scalar, &out_scalar);
|
|
|
|
ec_scalar_mul_montgomery(group.get(), &result, &in_scalar, &out_scalar);
|
|
|
|
|
|
|
|
EXPECT_EQ(0, OPENSSL_memcmp(kOneMont, &result, sizeof(kOneMont)));
|
|
|
|
|
|
|
|
// Invert the result and expect to get back to the original value.
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
ASSERT_TRUE(CHECK_ABI(beeu_mod_inverse_vartime, out, out, order_words));
|
2018-11-06 23:18:56 +00:00
|
|
|
EXPECT_EQ(0, OPENSSL_memcmp(in, out, sizeof(in)));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2016-10-25 01:02:26 +01:00
|
|
|
static bool GetFieldElement(FileTest *t, BN_ULONG out[P256_LIMBS],
|
|
|
|
const char *name) {
|
|
|
|
std::vector<uint8_t> bytes;
|
|
|
|
if (!t->GetBytes(&bytes, name)) {
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (bytes.size() != BN_BYTES * P256_LIMBS) {
|
2017-06-09 01:04:08 +01:00
|
|
|
ADD_FAILURE() << "Invalid length: " << name;
|
2016-10-25 01:02:26 +01:00
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
// |byte| contains bytes in big-endian while |out| should contain |BN_ULONG|s
|
|
|
|
// in little-endian.
|
2016-12-13 06:07:13 +00:00
|
|
|
OPENSSL_memset(out, 0, P256_LIMBS * sizeof(BN_ULONG));
|
2016-10-25 01:02:26 +01:00
|
|
|
for (size_t i = 0; i < bytes.size(); i++) {
|
|
|
|
out[P256_LIMBS - 1 - (i / BN_BYTES)] <<= 8;
|
|
|
|
out[P256_LIMBS - 1 - (i / BN_BYTES)] |= bytes[i];
|
|
|
|
}
|
|
|
|
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
static std::string FieldElementToString(const BN_ULONG a[P256_LIMBS]) {
|
|
|
|
std::string ret;
|
|
|
|
for (size_t i = P256_LIMBS-1; i < P256_LIMBS; i--) {
|
|
|
|
char buf[2 * BN_BYTES + 1];
|
|
|
|
BIO_snprintf(buf, sizeof(buf), BN_HEX_FMT2, a[i]);
|
|
|
|
ret += buf;
|
|
|
|
}
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2017-06-09 01:04:08 +01:00
|
|
|
static testing::AssertionResult ExpectFieldElementsEqual(
|
|
|
|
const char *expected_expr, const char *actual_expr,
|
|
|
|
const BN_ULONG expected[P256_LIMBS], const BN_ULONG actual[P256_LIMBS]) {
|
2016-12-13 06:07:13 +00:00
|
|
|
if (OPENSSL_memcmp(expected, actual, sizeof(BN_ULONG) * P256_LIMBS) == 0) {
|
2017-06-09 01:04:08 +01:00
|
|
|
return testing::AssertionSuccess();
|
2016-10-25 01:02:26 +01:00
|
|
|
}
|
|
|
|
|
2017-06-09 01:04:08 +01:00
|
|
|
return testing::AssertionFailure()
|
|
|
|
<< "Expected: " << FieldElementToString(expected) << " ("
|
|
|
|
<< expected_expr << ")\n"
|
|
|
|
<< "Actual: " << FieldElementToString(actual) << " (" << actual_expr
|
|
|
|
<< ")";
|
2016-10-25 01:02:26 +01:00
|
|
|
}
|
|
|
|
|
2017-06-09 01:04:08 +01:00
|
|
|
#define EXPECT_FIELD_ELEMENTS_EQUAL(a, b) \
|
|
|
|
EXPECT_PRED_FORMAT2(ExpectFieldElementsEqual, a, b)
|
|
|
|
|
2016-10-25 01:02:26 +01:00
|
|
|
static bool PointToAffine(P256_POINT_AFFINE *out, const P256_POINT *in) {
|
|
|
|
static const uint8_t kP[] = {
|
|
|
|
0xff, 0xff, 0xff, 0xff, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00,
|
|
|
|
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xff, 0xff,
|
|
|
|
0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
|
|
|
|
};
|
|
|
|
|
|
|
|
bssl::UniquePtr<BIGNUM> x(BN_new()), y(BN_new()), z(BN_new());
|
|
|
|
bssl::UniquePtr<BIGNUM> p(BN_bin2bn(kP, sizeof(kP), nullptr));
|
|
|
|
if (!x || !y || !z || !p ||
|
|
|
|
!bn_set_words(x.get(), in->X, P256_LIMBS) ||
|
|
|
|
!bn_set_words(y.get(), in->Y, P256_LIMBS) ||
|
|
|
|
!bn_set_words(z.get(), in->Z, P256_LIMBS)) {
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
// Coordinates must be fully-reduced.
|
2016-11-15 22:34:53 +00:00
|
|
|
if (BN_cmp(x.get(), p.get()) >= 0 ||
|
2016-10-25 01:02:26 +01:00
|
|
|
BN_cmp(y.get(), p.get()) >= 0 ||
|
|
|
|
BN_cmp(z.get(), p.get()) >= 0) {
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (BN_is_zero(z.get())) {
|
|
|
|
// The point at infinity is represented as (0, 0).
|
2018-01-20 21:51:54 +00:00
|
|
|
OPENSSL_memset(out, 0, sizeof(P256_POINT_AFFINE));
|
2016-10-25 01:02:26 +01:00
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
bssl::UniquePtr<BN_CTX> ctx(BN_CTX_new());
|
2018-01-23 22:03:26 +00:00
|
|
|
bssl::UniquePtr<BN_MONT_CTX> mont(
|
|
|
|
BN_MONT_CTX_new_for_modulus(p.get(), ctx.get()));
|
2016-10-25 01:02:26 +01:00
|
|
|
if (!ctx || !mont ||
|
|
|
|
// Invert Z.
|
|
|
|
!BN_from_montgomery(z.get(), z.get(), mont.get(), ctx.get()) ||
|
|
|
|
!BN_mod_inverse(z.get(), z.get(), p.get(), ctx.get()) ||
|
|
|
|
!BN_to_montgomery(z.get(), z.get(), mont.get(), ctx.get()) ||
|
|
|
|
// Convert (X, Y, Z) to (X/Z^2, Y/Z^3).
|
|
|
|
!BN_mod_mul_montgomery(x.get(), x.get(), z.get(), mont.get(),
|
|
|
|
ctx.get()) ||
|
|
|
|
!BN_mod_mul_montgomery(x.get(), x.get(), z.get(), mont.get(),
|
|
|
|
ctx.get()) ||
|
|
|
|
!BN_mod_mul_montgomery(y.get(), y.get(), z.get(), mont.get(),
|
|
|
|
ctx.get()) ||
|
|
|
|
!BN_mod_mul_montgomery(y.get(), y.get(), z.get(), mont.get(),
|
|
|
|
ctx.get()) ||
|
|
|
|
!BN_mod_mul_montgomery(y.get(), y.get(), z.get(), mont.get(),
|
2018-01-20 21:51:54 +00:00
|
|
|
ctx.get()) ||
|
|
|
|
!bn_copy_words(out->X, P256_LIMBS, x.get()) ||
|
|
|
|
!bn_copy_words(out->Y, P256_LIMBS, y.get())) {
|
2016-10-25 01:02:26 +01:00
|
|
|
return false;
|
|
|
|
}
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
2017-06-09 01:04:08 +01:00
|
|
|
static testing::AssertionResult ExpectPointsEqual(
|
|
|
|
const char *expected_expr, const char *actual_expr,
|
|
|
|
const P256_POINT_AFFINE *expected, const P256_POINT *actual) {
|
2016-10-25 01:02:26 +01:00
|
|
|
// There are multiple representations of the same |P256_POINT|, so convert to
|
|
|
|
// |P256_POINT_AFFINE| and compare.
|
|
|
|
P256_POINT_AFFINE affine;
|
2017-06-09 01:04:08 +01:00
|
|
|
if (!PointToAffine(&affine, actual)) {
|
|
|
|
return testing::AssertionFailure()
|
|
|
|
<< "Could not convert " << actual_expr << " to affine: ("
|
|
|
|
<< FieldElementToString(actual->X) << ", "
|
|
|
|
<< FieldElementToString(actual->Y) << ", "
|
|
|
|
<< FieldElementToString(actual->Z) << ")";
|
2016-10-25 01:02:26 +01:00
|
|
|
}
|
|
|
|
|
2016-12-13 06:07:13 +00:00
|
|
|
if (OPENSSL_memcmp(expected, &affine, sizeof(P256_POINT_AFFINE)) != 0) {
|
2017-06-09 01:04:08 +01:00
|
|
|
return testing::AssertionFailure()
|
|
|
|
<< "Expected: (" << FieldElementToString(expected->X) << ", "
|
|
|
|
<< FieldElementToString(expected->Y) << ") (" << expected_expr
|
|
|
|
<< "; affine)\n"
|
|
|
|
<< "Actual: (" << FieldElementToString(affine.X) << ", "
|
|
|
|
<< FieldElementToString(affine.Y) << ") (" << actual_expr << ")";
|
2016-10-25 01:02:26 +01:00
|
|
|
}
|
|
|
|
|
2017-06-09 01:04:08 +01:00
|
|
|
return testing::AssertionSuccess();
|
2016-10-25 01:02:26 +01:00
|
|
|
}
|
|
|
|
|
2017-06-09 01:04:08 +01:00
|
|
|
#define EXPECT_POINTS_EQUAL(a, b) EXPECT_PRED_FORMAT2(ExpectPointsEqual, a, b)
|
|
|
|
|
|
|
|
static void TestNegate(FileTest *t) {
|
2016-10-25 01:02:26 +01:00
|
|
|
BN_ULONG a[P256_LIMBS], b[P256_LIMBS];
|
2017-06-09 01:04:08 +01:00
|
|
|
ASSERT_TRUE(GetFieldElement(t, a, "A"));
|
|
|
|
ASSERT_TRUE(GetFieldElement(t, b, "B"));
|
2016-10-25 01:02:26 +01:00
|
|
|
|
|
|
|
// Test that -A = B.
|
|
|
|
BN_ULONG ret[P256_LIMBS];
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_neg, ret, a);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(b, ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
2016-12-13 06:07:13 +00:00
|
|
|
OPENSSL_memcpy(ret, a, sizeof(ret));
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_neg, ret, ret /* a */);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(b, ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
|
|
|
// Test that -B = A.
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_neg, ret, b);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(a, ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
2016-12-13 06:07:13 +00:00
|
|
|
OPENSSL_memcpy(ret, b, sizeof(ret));
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_neg, ret, ret /* b */);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(a, ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
}
|
|
|
|
|
2017-06-09 01:04:08 +01:00
|
|
|
static void TestMulMont(FileTest *t) {
|
2016-10-25 01:02:26 +01:00
|
|
|
BN_ULONG a[P256_LIMBS], b[P256_LIMBS], result[P256_LIMBS];
|
2017-06-09 01:04:08 +01:00
|
|
|
ASSERT_TRUE(GetFieldElement(t, a, "A"));
|
|
|
|
ASSERT_TRUE(GetFieldElement(t, b, "B"));
|
|
|
|
ASSERT_TRUE(GetFieldElement(t, result, "Result"));
|
2016-10-25 01:02:26 +01:00
|
|
|
|
|
|
|
BN_ULONG ret[P256_LIMBS];
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_mul_mont, ret, a, b);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(result, ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_mul_mont, ret, b, a);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(result, ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
2016-12-13 06:07:13 +00:00
|
|
|
OPENSSL_memcpy(ret, a, sizeof(ret));
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_mul_mont, ret, ret /* a */, b);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(result, ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
2016-12-13 06:07:13 +00:00
|
|
|
OPENSSL_memcpy(ret, a, sizeof(ret));
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_mul_mont, ret, b, ret);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(result, ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
2016-12-13 06:07:13 +00:00
|
|
|
OPENSSL_memcpy(ret, b, sizeof(ret));
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_mul_mont, ret, a, ret /* b */);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(result, ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
2016-12-13 06:07:13 +00:00
|
|
|
OPENSSL_memcpy(ret, b, sizeof(ret));
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_mul_mont, ret, ret /* b */, a);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(result, ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
2016-12-13 06:07:13 +00:00
|
|
|
if (OPENSSL_memcmp(a, b, sizeof(a)) == 0) {
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_sqr_mont, ret, a);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(result, ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
2016-12-13 06:07:13 +00:00
|
|
|
OPENSSL_memcpy(ret, a, sizeof(ret));
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_sqr_mont, ret, ret /* a */);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(result, ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2017-06-09 01:04:08 +01:00
|
|
|
static void TestFromMont(FileTest *t) {
|
2016-10-25 01:02:26 +01:00
|
|
|
BN_ULONG a[P256_LIMBS], result[P256_LIMBS];
|
2017-06-09 01:04:08 +01:00
|
|
|
ASSERT_TRUE(GetFieldElement(t, a, "A"));
|
|
|
|
ASSERT_TRUE(GetFieldElement(t, result, "Result"));
|
2016-10-25 01:02:26 +01:00
|
|
|
|
|
|
|
BN_ULONG ret[P256_LIMBS];
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_from_mont, ret, a);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(result, ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
2016-12-13 06:07:13 +00:00
|
|
|
OPENSSL_memcpy(ret, a, sizeof(ret));
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_from_mont, ret, ret /* a */);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(result, ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
}
|
|
|
|
|
2017-06-09 01:04:08 +01:00
|
|
|
static void TestPointAdd(FileTest *t) {
|
2016-10-25 01:02:26 +01:00
|
|
|
P256_POINT a, b;
|
|
|
|
P256_POINT_AFFINE result;
|
2017-06-09 01:04:08 +01:00
|
|
|
ASSERT_TRUE(GetFieldElement(t, a.X, "A.X"));
|
|
|
|
ASSERT_TRUE(GetFieldElement(t, a.Y, "A.Y"));
|
|
|
|
ASSERT_TRUE(GetFieldElement(t, a.Z, "A.Z"));
|
|
|
|
ASSERT_TRUE(GetFieldElement(t, b.X, "B.X"));
|
|
|
|
ASSERT_TRUE(GetFieldElement(t, b.Y, "B.Y"));
|
|
|
|
ASSERT_TRUE(GetFieldElement(t, b.Z, "B.Z"));
|
|
|
|
ASSERT_TRUE(GetFieldElement(t, result.X, "Result.X"));
|
|
|
|
ASSERT_TRUE(GetFieldElement(t, result.Y, "Result.Y"));
|
2016-10-25 01:02:26 +01:00
|
|
|
|
|
|
|
P256_POINT ret;
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_point_add, &ret, &a, &b);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_POINTS_EQUAL(&result, &ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_point_add, &ret, &b, &a);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_POINTS_EQUAL(&result, &ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
2016-12-13 06:07:13 +00:00
|
|
|
OPENSSL_memcpy(&ret, &a, sizeof(ret));
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_point_add, &ret, &ret /* a */, &b);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_POINTS_EQUAL(&result, &ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
2016-12-13 06:07:13 +00:00
|
|
|
OPENSSL_memcpy(&ret, &a, sizeof(ret));
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_point_add, &ret, &b, &ret /* a */);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_POINTS_EQUAL(&result, &ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
2016-12-13 06:07:13 +00:00
|
|
|
OPENSSL_memcpy(&ret, &b, sizeof(ret));
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_point_add, &ret, &a, &ret /* b */);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_POINTS_EQUAL(&result, &ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
2016-12-13 06:07:13 +00:00
|
|
|
OPENSSL_memcpy(&ret, &b, sizeof(ret));
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_point_add, &ret, &ret /* b */, &a);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_POINTS_EQUAL(&result, &ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
|
|
|
P256_POINT_AFFINE a_affine, b_affine, infinity;
|
2016-12-13 06:07:13 +00:00
|
|
|
OPENSSL_memset(&infinity, 0, sizeof(infinity));
|
2017-06-09 01:04:08 +01:00
|
|
|
ASSERT_TRUE(PointToAffine(&a_affine, &a));
|
|
|
|
ASSERT_TRUE(PointToAffine(&b_affine, &b));
|
2016-10-25 01:02:26 +01:00
|
|
|
|
|
|
|
// ecp_nistz256_point_add_affine does not work when a == b unless doubling the
|
|
|
|
// point at infinity.
|
2016-12-13 06:07:13 +00:00
|
|
|
if (OPENSSL_memcmp(&a_affine, &b_affine, sizeof(a_affine)) != 0 ||
|
|
|
|
OPENSSL_memcmp(&a_affine, &infinity, sizeof(a_affine)) == 0) {
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_point_add_affine, &ret, &a, &b_affine);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_POINTS_EQUAL(&result, &ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
2016-12-13 06:07:13 +00:00
|
|
|
OPENSSL_memcpy(&ret, &a, sizeof(ret));
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_point_add_affine, &ret, &ret /* a */, &b_affine);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_POINTS_EQUAL(&result, &ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_point_add_affine, &ret, &b, &a_affine);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_POINTS_EQUAL(&result, &ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
2016-12-13 06:07:13 +00:00
|
|
|
OPENSSL_memcpy(&ret, &b, sizeof(ret));
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_point_add_affine, &ret, &ret /* b */, &a_affine);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_POINTS_EQUAL(&result, &ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
}
|
|
|
|
|
2016-12-13 06:07:13 +00:00
|
|
|
if (OPENSSL_memcmp(&a, &b, sizeof(a)) == 0) {
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_point_double, &ret, &a);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_POINTS_EQUAL(&result, &ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
|
|
|
|
ret = a;
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_point_double, &ret, &ret /* a */);
|
2017-06-09 01:04:08 +01:00
|
|
|
EXPECT_POINTS_EQUAL(&result, &ret);
|
2016-10-25 01:02:26 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2018-04-21 06:20:15 +01:00
|
|
|
static void TestOrdMulMont(FileTest *t) {
|
|
|
|
// This test works on scalars rather than field elements, but the
|
|
|
|
// representation is the same.
|
|
|
|
BN_ULONG a[P256_LIMBS], b[P256_LIMBS], result[P256_LIMBS];
|
|
|
|
ASSERT_TRUE(GetFieldElement(t, a, "A"));
|
|
|
|
ASSERT_TRUE(GetFieldElement(t, b, "B"));
|
|
|
|
ASSERT_TRUE(GetFieldElement(t, result, "Result"));
|
|
|
|
|
|
|
|
BN_ULONG ret[P256_LIMBS];
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_ord_mul_mont, ret, a, b);
|
2018-04-21 06:20:15 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(result, ret);
|
|
|
|
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_ord_mul_mont, ret, b, a);
|
2018-04-21 06:20:15 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(result, ret);
|
|
|
|
|
|
|
|
OPENSSL_memcpy(ret, a, sizeof(ret));
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_ord_mul_mont, ret, ret /* a */, b);
|
2018-04-21 06:20:15 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(result, ret);
|
|
|
|
|
|
|
|
OPENSSL_memcpy(ret, a, sizeof(ret));
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_ord_mul_mont, ret, b, ret);
|
2018-04-21 06:20:15 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(result, ret);
|
|
|
|
|
|
|
|
OPENSSL_memcpy(ret, b, sizeof(ret));
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_ord_mul_mont, ret, a, ret /* b */);
|
2018-04-21 06:20:15 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(result, ret);
|
|
|
|
|
|
|
|
OPENSSL_memcpy(ret, b, sizeof(ret));
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_ord_mul_mont, ret, ret /* b */, a);
|
2018-04-21 06:20:15 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(result, ret);
|
|
|
|
|
|
|
|
if (OPENSSL_memcmp(a, b, sizeof(a)) == 0) {
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_ord_sqr_mont, ret, a, 1);
|
2018-04-21 06:20:15 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(result, ret);
|
|
|
|
|
|
|
|
OPENSSL_memcpy(ret, a, sizeof(ret));
|
Add an ABI testing framework.
Dear reader, I must apologize in advance. This CL contains the following:
- A new 256-line perlasm file with non-trivial perl bits and a dual-ABI
variadic function caller.
- C preprocessor gymnastics, with variadic macros and fun facts about
__VA_ARGS__'s behavior on empty argument lists.
- C++ template gymnastics, including variadic arguments, template
specialization, std::enable_if, and machinery to control template argument
deduction.
Enjoy.
This tests that our assembly functions correctly honor platform ABI
conventions. Right now this only tests callee-saved registers, but it should be
extendable to SEH/CFI unwind testing with single-step debugging APIs.
Register-checking does not involve anything funny and should be compatible with
SDE. (The future unwind testing is unlikely to be compatible.)
This CL adds support for x86_64 SysV and Win64 ABIs. ARM, AArch64, and x86 can
be added in the future. The testing is injected in two places. First, all the
assembly tests in p256-x86_64-test.cc are now instrumented. This is the
intended workflow and should capture all registers.
However, we currently do not unit-test our assembly much directly. We should do
that as follow-up work[0] but, in the meantime, I've also wrapped all of the GTest
main function in an ABI test. This is imperfect as ABI failures may be masked
by other stack frames, but it costs nothing[1] and is pretty reliable at
catching Win64 xmm register failures.
[0] An alternate strategy would be, in debug builds, unconditionally instrument
every assembly call in libcrypto. But the CHECK_ABI macro would be difficult to
replicate in pure C, and unwind testing may be too invasive for this. Still,
something to consider when we C++ libcrypto.
[1] When single-stepped unwind testing exists, it won't cost nothing. The
gtest_main.cc call will turn unwind testing off.
Change-Id: I6643b26445891fd46abfacac52bc024024c8d7f6
Reviewed-on: https://boringssl-review.googlesource.com/c/33764
Reviewed-by: Adam Langley <agl@google.com>
Reviewed-by: Adam Langley <alangley@gmail.com>
Commit-Queue: David Benjamin <davidben@google.com>
2018-12-16 00:58:43 +00:00
|
|
|
CHECK_ABI(ecp_nistz256_ord_sqr_mont, ret, ret /* a */, 1);
|
2018-04-21 06:20:15 +01:00
|
|
|
EXPECT_FIELD_ELEMENTS_EQUAL(result, ret);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2017-06-09 01:04:08 +01:00
|
|
|
TEST(P256_X86_64Test, TestVectors) {
|
|
|
|
return FileTestGTest("crypto/fipsmodule/ec/p256-x86_64_tests.txt",
|
|
|
|
[](FileTest *t) {
|
2016-10-25 01:02:26 +01:00
|
|
|
if (t->GetParameter() == "Negate") {
|
2017-06-09 01:04:08 +01:00
|
|
|
TestNegate(t);
|
|
|
|
} else if (t->GetParameter() == "MulMont") {
|
|
|
|
TestMulMont(t);
|
|
|
|
} else if (t->GetParameter() == "FromMont") {
|
|
|
|
TestFromMont(t);
|
|
|
|
} else if (t->GetParameter() == "PointAdd") {
|
|
|
|
TestPointAdd(t);
|
2018-04-21 06:20:15 +01:00
|
|
|
} else if (t->GetParameter() == "OrdMulMont") {
|
|
|
|
TestOrdMulMont(t);
|
2017-06-09 01:04:08 +01:00
|
|
|
} else {
|
|
|
|
FAIL() << "Unknown test type:" << t->GetParameter();
|
2016-10-25 01:02:26 +01:00
|
|
|
}
|
2017-06-09 01:04:08 +01:00
|
|
|
});
|
2016-10-25 01:02:26 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
#endif
|