Add ARM SHA256 asm implementation from Cryptogams (GH #840, PR #840)

Cryptogams is Andy Polyakov's project used to create high speed crypto algorithms and share them with other developers. Cryptogams  has a dual license. First is the OpenSSL license because Andy contributes to OpenSSL. Second is a BSD license for those who want a more permissive license.

Andy's implementation runs about 45% faster than C/C++ code. Testing on a 1 GHz Cortex-A7 shows Cryptograms at 17 cpb, and C++ at 30 cpb.

The integration instructions are documented at [Cryptogams SHA](https://wiki.openssl.org/index.php/Cryptogams_SHA) on the OpenSSL wiki.

1 files changed, 49 insertions, 1 deletions

diff --git a/sha.cpp b/sha.cpp
index 70b6d48a..c4662333 100644
--- a/sha.cpp
+++ b/sha.cpp
@@ -69,6 +69,11 @@ extern void SHA1_HashMultipleBlocks_ARMV8(word32 *state, const word32 *data, siz
 extern void SHA256_HashMultipleBlocks_ARMV8(word32 *state, const word32 *data, size_t length, ByteOrder order);

 #endif

 

+#if (CRYPTOGAMS_ARM_SHA256)

+extern "C" unsigned int CRYPTOGAMS_armcaps;

+extern "C" int sha256_block_data_order(word32* state, const word32 *data, size_t blocks);

+#endif

+

 #if CRYPTOPP_ARM_SHA512_AVAILABLE

 extern void SHA512_HashMultipleBlocks_ARMV8(word32 *state, const word32 *data, size_t length, ByteOrder order);

 #endif

@@ -309,7 +314,8 @@ size_t SHA1::HashMultipleBlocks(const word32 *input, size_t length)
         // for capabilities like ARMv7 and NEON. Storage is allocated in the

         // module. We still need to set CRYPTOGAMS_armcaps accordingly.

         // The Cryptogams code defines NEON as 1<<0; see ARMV7_NEON.

-        static unsigned int unused = CRYPTOGAMS_armcaps = HasNEON() ? (1<<0) : 0;

+        static const unsigned int unused = CRYPTOGAMS_armcaps = HasNEON() ? (1<<0) : 0;

+        CRYPTOPP_UNUSED(unused);

 

         sha1_block_data_order(m_state, input, length / SHA1::BLOCKSIZE);

         return length & (SHA1::BLOCKSIZE - 1);

@@ -428,6 +434,12 @@ std::string SHA256_AlgorithmProvider()
     if (HasSSE2())

         return "SSE2";

 #endif

+#if CRYPTOGAMS_ARM_SHA256

+    if (HasNEON())

+        return "NEON";

+    if (HasARMv7())

+        return "ARMv7";

+#endif

 #if CRYPTOPP_ARM_SHA2_AVAILABLE

     if (HasSHA2())

         return "ARMv8";

@@ -811,6 +823,14 @@ void SHA256::Transform(word32 *state, const word32 *data)
         return;

     }

 #endif

+#if CRYPTOGAMS_ARM_SHA256 && 0

+    // TODO: convert LE to BE and use Cryptogams code

+    if (HasARMv7())

+    {

+        sha256_block_data_order(state, data, 1);

+        return;

+    }

+#endif

 #if CRYPTOPP_ARM_SHA2_AVAILABLE

     if (HasSHA2())

     {

@@ -849,6 +869,20 @@ size_t SHA256::HashMultipleBlocks(const word32 *input, size_t length)
         return res;

     }

 #endif

+#if CRYPTOGAMS_ARM_SHA256

+    if (HasARMv7())

+    {

+        // The Cryptogams code uses a global variable named CRYPTOGAMS_armcaps

+        // for capabilities like ARMv7 and NEON. Storage is allocated in the

+        // module. We still need to set CRYPTOGAMS_armcaps accordingly.

+        // The Cryptogams code defines NEON as 1<<0; see ARMV7_NEON.

+        static const unsigned int unused = CRYPTOGAMS_armcaps = HasNEON() ? (1<<0) : 0;

+        CRYPTOPP_UNUSED(unused);

+

+        sha256_block_data_order(m_state, input, length / SHA256::BLOCKSIZE);

+        return length & (SHA256::BLOCKSIZE - 1);

+    }

+#endif

 #if CRYPTOPP_ARM_SHA2_AVAILABLE

     if (HasSHA2())

     {

@@ -905,6 +939,20 @@ size_t SHA224::HashMultipleBlocks(const word32 *input, size_t length)
         return res;

     }

 #endif

+#if CRYPTOGAMS_ARM_SHA256

+    if (HasARMv7())

+    {

+        // The Cryptogams code uses a global variable named CRYPTOGAMS_armcaps

+        // for capabilities like ARMv7 and NEON. Storage is allocated in the

+        // module. We still need to set CRYPTOGAMS_armcaps accordingly.

+        // The Cryptogams code defines NEON as 1<<0; see ARMV7_NEON.

+        static const unsigned int unused = CRYPTOGAMS_armcaps = HasNEON() ? (1<<0) : 0;

+        CRYPTOPP_UNUSED(unused);

+

+        sha256_block_data_order(m_state, input, length / SHA256::BLOCKSIZE);

+        return length & (SHA256::BLOCKSIZE - 1);

+    }

+#endif

 #if CRYPTOPP_ARM_SHA2_AVAILABLE

     if (HasSHA2())

     {