Merge bitcoin#28893: Fix SSE4.1-related issues

d440f13 crypto: Guard code with `ENABLE_SSE41` macro (Hennadii Stepanov)
6ec1ca7 build: Fix test for SSE4.1 intrinsics (Hennadii Stepanov)

Pull request description:

  1. Fix the test for SSE4.1 intrinsics during build system configuration, which currently can be false positive, for example, when `CXXFLAGS="-mno-sse4.1"` provided.

  This PR fixes the test by adding the `_mm_blend_epi16` SSE4.1 function used in our codebase.

  2. Guard `sha_x86_shani.cpp` code with `ENABLE_SSE41` macro as it uses the `_mm_blend_epi16` function from
  the SSE4.1 instruction set.

  It is possible that SHA-NI is enabled even when SSE4.1 is disabled, which causes compile errors in the master branch.

  Closes bitcoin#28864.

ACKs for top commit:
  sipa:
    utACK d440f13
  willcl-ark:
    tACK d440f13
  theuni:
    utACK d440f13

Tree-SHA512: a6e1e8c94e1b94874ff51846815ef445e6135cbdb01b08eb695b3548115f2340dd835ebe53673ae46a553fe6be4815e68d8642c34235dd7af5106c4b7c9ea6f3

configure.ac

@@ -482,11 +482,12 @@ TEMP_CXXFLAGS="$CXXFLAGS"
 CXXFLAGS="$SSE41_CXXFLAGS $CXXFLAGS"
 AC_MSG_CHECKING([for SSE4.1 intrinsics])
 AC_COMPILE_IFELSE([AC_LANG_PROGRAM([[
-    #include <stdint.h>
     #include <immintrin.h>
   ]],[[
-    __m128i l = _mm_set1_epi32(0);
-    return _mm_extract_epi32(l, 3);
+    __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);
   ]])],
  [ AC_MSG_RESULT([yes]); enable_sse41=yes; AC_DEFINE([ENABLE_SSE41], [1], [Define this symbol to build code that uses SSE4.1 intrinsics]) ],
  [ AC_MSG_RESULT([no])]

src/Makefile.am

@@ -51,15 +51,15 @@ LIBBITCOIN_CRYPTO = $(LIBBITCOIN_CRYPTO_BASE)
 if ENABLE_SSE41
 LIBBITCOIN_CRYPTO_SSE41 = crypto/libbitcoin_crypto_sse41.la
 LIBBITCOIN_CRYPTO += $(LIBBITCOIN_CRYPTO_SSE41)
+if ENABLE_X86_SHANI
+LIBBITCOIN_CRYPTO_X86_SHANI = crypto/libbitcoin_crypto_x86_shani.la
+LIBBITCOIN_CRYPTO += $(LIBBITCOIN_CRYPTO_X86_SHANI)
+endif
 endif
 if ENABLE_AVX2
 LIBBITCOIN_CRYPTO_AVX2 = crypto/libbitcoin_crypto_avx2.la
 LIBBITCOIN_CRYPTO += $(LIBBITCOIN_CRYPTO_AVX2)
 endif
-if ENABLE_X86_SHANI
-LIBBITCOIN_CRYPTO_X86_SHANI = crypto/libbitcoin_crypto_x86_shani.la
-LIBBITCOIN_CRYPTO += $(LIBBITCOIN_CRYPTO_X86_SHANI)
-endif
 if ENABLE_ARM_SHANI
 LIBBITCOIN_CRYPTO_ARM_SHANI = crypto/libbitcoin_crypto_arm_shani.la
 LIBBITCOIN_CRYPTO += $(LIBBITCOIN_CRYPTO_ARM_SHANI)
@@ -622,7 +622,7 @@ crypto_libbitcoin_crypto_x86_shani_la_LDFLAGS = $(AM_LDFLAGS) -static
 crypto_libbitcoin_crypto_x86_shani_la_CXXFLAGS = $(AM_CXXFLAGS) $(PIE_FLAGS) -static
 crypto_libbitcoin_crypto_x86_shani_la_CPPFLAGS = $(AM_CPPFLAGS)
 crypto_libbitcoin_crypto_x86_shani_la_CXXFLAGS += $(X86_SHANI_CXXFLAGS)
-crypto_libbitcoin_crypto_x86_shani_la_CPPFLAGS += -DENABLE_X86_SHANI
+crypto_libbitcoin_crypto_x86_shani_la_CPPFLAGS += -DENABLE_SSE41 -DENABLE_X86_SHANI
 crypto_libbitcoin_crypto_x86_shani_la_SOURCES = crypto/sha256_x86_shani.cpp
 
 # See explanation for -static in crypto_libbitcoin_crypto_base_la's LDFLAGS and

src/crypto/sha256.cpp

@@ -621,7 +621,7 @@ std::string SHA256AutoDetect(sha256_implementation::UseImplementation use_implem
         }
     }
 
-#if defined(ENABLE_X86_SHANI)
+#if defined(ENABLE_SSE41) && defined(ENABLE_X86_SHANI)
     if (have_x86_shani) {
         Transform = sha256_x86_shani::Transform;
         TransformD64 = TransformD64Wrapper<sha256_x86_shani::Transform>;

src/crypto/sha256_x86_shani.cpp

@@ -6,7 +6,7 @@
 // Written and placed in public domain by Jeffrey Walton.
 // Based on code from Intel, and by Sean Gulley for the miTLS project.
 
-#ifdef ENABLE_X86_SHANI
+#if defined(ENABLE_SSE41) && defined(ENABLE_X86_SHANI)
 
 #include <stdint.h>
 #include <immintrin.h>