Skip to content

crypto/sha256: Use pragmas to enforce necessary intrinsics for GCC and Clang #102

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 2 commits into
base: __branch_29
Choose a base branch
from
Open

crypto/sha256: Use pragmas to enforce necessary intrinsics for GCC and Clang #102

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
c244d63
crypto/sha256: Use pragmas to enforce necessary intrinsics for GCC an…
luke-jr Jul 28, 2018
4d9aff4
cmake: For intrinsics checks, use pragmas to explicitly enable them f…
luke-jr Aug 21, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
57 changes: 49 additions & 8 deletions cmake/introspection.cmake
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -164,58 +164,95 @@ if(NOT MSVC)

# Check for SSE4.1 intrinsics.
set(SSE41_CXXFLAGS -msse4.1)
check_cxx_source_compiles_with_flags("
check_cxx_source_compiles_with_flags([[
#include <immintrin.h>

#if defined(__clang__)
#pragma clang attribute push(__attribute__((__target__("sse4.1"))), apply_to = function)
__attribute__((__target__("arch=skylake-avx512")))
#elif defined(__GNUC__)
#pragma GCC target ("sse4.1")
#endif

int main()
{
__m128i a = _mm_set1_epi32(0);
__m128i b = _mm_set1_epi32(1);
__m128i r = _mm_blend_epi16(a, b, 0xFF);
return _mm_extract_epi32(r, 3);
}
" HAVE_SSE41

#if defined(__clang__)
#pragma clang attribute pop
#endif
]] HAVE_SSE41
CXXFLAGS ${SSE41_CXXFLAGS}
)
set(ENABLE_SSE41 ${HAVE_SSE41})

# Check for AVX2 intrinsics.
set(AVX2_CXXFLAGS -mavx -mavx2)
check_cxx_source_compiles_with_flags("
check_cxx_source_compiles_with_flags([[
#include <immintrin.h>

#if defined(__clang__)
#pragma clang attribute push(__attribute__((__target__("avx,avx2"))), apply_to = function)
#elif defined(__GNUC__)
#pragma GCC target ("avx,avx2")
#endif

int main()
{
__m256i l = _mm256_set1_epi32(0);
return _mm256_extract_epi32(l, 7);
}
" HAVE_AVX2

#if defined(__clang__)
#pragma clang attribute pop
#endif
]] HAVE_AVX2
CXXFLAGS ${AVX2_CXXFLAGS}
)
set(ENABLE_AVX2 ${HAVE_AVX2})

# Check for x86 SHA-NI intrinsics.
set(X86_SHANI_CXXFLAGS -msse4 -msha)
check_cxx_source_compiles_with_flags("
check_cxx_source_compiles_with_flags([[
#include <immintrin.h>

#if defined(__clang__)
#pragma clang attribute push(__attribute__((__target__("sse4,sse4.1,sha"))), apply_to = function)
#elif defined(__GNUC__)
#pragma GCC target ("sse4,sse4.1,sha")
#endif

int main()
{
__m128i i = _mm_set1_epi32(0);
__m128i j = _mm_set1_epi32(1);
__m128i k = _mm_set1_epi32(2);
return _mm_extract_epi32(_mm_sha256rnds2_epu32(i, j, k), 0);
}
" HAVE_X86_SHANI

#if defined(__clang__)
#pragma clang attribute pop
#endif
]] HAVE_X86_SHANI
CXXFLAGS ${X86_SHANI_CXXFLAGS}
)
set(ENABLE_X86_SHANI ${HAVE_X86_SHANI})

# Check for ARMv8 SHA-NI intrinsics.
set(ARM_SHANI_CXXFLAGS -march=armv8-a+crypto)
check_cxx_source_compiles_with_flags("
check_cxx_source_compiles_with_flags([[
#include <arm_neon.h>

#if defined(__clang__)
#pragma clang attribute push(__attribute__((__target__("armv8-a+crypto"))), apply_to = function)
#elif defined(__GNUC__)
#pragma GCC target ("armv8-a+crypto")
#endif

int main()
{
uint32x4_t a, b, c;
Expand All @@ -224,7 +261,11 @@ if(NOT MSVC)
vsha256su0q_u32(a, b);
vsha256su1q_u32(a, b, c);
}
" HAVE_ARM_SHANI

#if defined(__clang__)
#pragma clang attribute pop
#endif
]] HAVE_ARM_SHANI
CXXFLAGS ${ARM_SHANI_CXXFLAGS}
)
set(ENABLE_ARM_SHANI ${HAVE_ARM_SHANI})
Expand Down
10 changes: 10 additions & 0 deletions src/crypto/sha256_avx2.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,6 +10,12 @@
#include <attributes.h>
#include <crypto/common.h>

#if defined(__clang__)
#pragma clang attribute push(__attribute__((__target__("avx,avx2"))), apply_to = function)
#elif defined(__GNUC__)
#pragma GCC target ("avx,avx2")
#endif

namespace sha256d64_avx2 {
namespace {

Expand Down Expand Up @@ -326,4 +332,8 @@ void Transform_8way(unsigned char* out, const unsigned char* in)

}

#if defined(__clang__)
#pragma clang attribute pop
#endif

#endif
10 changes: 10 additions & 0 deletions src/crypto/sha256_sse41.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,6 +10,12 @@
#include <attributes.h>
#include <crypto/common.h>

#if defined(__clang__)
#pragma clang attribute push(__attribute__((__target__("sse4.1"))), apply_to = function)
#elif defined(__GNUC__)
#pragma GCC target ("sse4.1")
#endif

namespace sha256d64_sse41 {
namespace {

Expand Down Expand Up @@ -318,4 +324,8 @@ void Transform_4way(unsigned char* out, const unsigned char* in)

}

#if defined(__clang__)
#pragma clang attribute pop
#endif

#endif
10 changes: 10 additions & 0 deletions src/crypto/sha256_x86_shani.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,6 +13,12 @@

#include <attributes.h>

#if defined(__clang__)
#pragma clang attribute push(__attribute__((__target__("sse4,sse4.1,sha"))), apply_to = function)
#elif defined(__GNUC__)
#pragma GCC target ("sse4,sse4.1,sha")
#endif

namespace {

alignas(__m128i) const uint8_t MASK[16] = {0x03, 0x02, 0x01, 0x00, 0x07, 0x06, 0x05, 0x04, 0x0b, 0x0a, 0x09, 0x08, 0x0f, 0x0e, 0x0d, 0x0c};
Expand Down Expand Up @@ -355,4 +361,8 @@ void Transform_2way(unsigned char* out, const unsigned char* in)

}

#if defined(__clang__)
#pragma clang attribute pop
#endif

#endif
Loading