Salsa20 SSE Windows build fix -- turns out you can't be as loose with SSE intrinsics in Visual Studio

node/Salsa20.cpp

@@ -144,10 +144,14 @@ void Salsa20::encrypt(const void *in,void *out,unsigned int bytes)
 		}
 
 #ifdef ZT_SALSA20_SSE
-		__m128i X0 = _state.v[0];
-		__m128i X1 = _state.v[1];
-		__m128i X2 = _state.v[2];
-		__m128i X3 = _state.v[3];
+		__m128i X0 = _mm_load_si128((const __m128i *)&(_state.v[0]));
+		__m128i X1 = _mm_load_si128((const __m128i *)&(_state.v[1]));
+		__m128i X2 = _mm_load_si128((const __m128i *)&(_state.v[2]));
+		__m128i X3 = _mm_load_si128((const __m128i *)&(_state.v[3]));
+		__m128i X0s = X0;
+		__m128i X1s = X1;
+		__m128i X2s = X2;
+		__m128i X3s = X3;
 
 		for (i=0;i<_roundsDiv2;++i) {
 			__m128i T = _mm_add_epi32(X0, X3);
@@ -185,10 +189,10 @@ void Salsa20::encrypt(const void *in,void *out,unsigned int bytes)
 			X3 = _mm_shuffle_epi32(X3, 0x93);
 		}
 
-		X0 = _mm_add_epi32(_state.v[0],X0);
-		X1 = _mm_add_epi32(_state.v[1],X1);
-		X2 = _mm_add_epi32(_state.v[2],X2);
-		X3 = _mm_add_epi32(_state.v[3],X3);
+		X0 = _mm_add_epi32(X0s,X0);
+		X1 = _mm_add_epi32(X1s,X1);
+		X2 = _mm_add_epi32(X2s,X2);
+		X3 = _mm_add_epi32(X3s,X3);
 
 		{
 			__m128i k02 = _mm_or_si128(_mm_slli_epi64(X0, 32), _mm_srli_epi64(X3, 32));
@@ -201,10 +205,10 @@ void Salsa20::encrypt(const void *in,void *out,unsigned int bytes)
 			const float *const mv = (const float *)m;
 			float *const cv = (float *)c;
 
-			_mm_storeu_ps(cv,_mm_xor_si128(_mm_unpackhi_epi64(k02,k20),_mm_loadu_ps(mv)));
-			_mm_storeu_ps(cv + 4,_mm_xor_si128(_mm_unpackhi_epi64(k13,k31),_mm_loadu_ps(mv + 4)));
-			_mm_storeu_ps(cv + 8,_mm_xor_si128(_mm_unpacklo_epi64(k20,k02),_mm_loadu_ps(mv + 8)));
-			_mm_storeu_ps(cv + 12,_mm_xor_si128(_mm_unpacklo_epi64(k31,k13),_mm_loadu_ps(mv + 12)));
+			_mm_storeu_ps(cv,_mm_castsi128_ps(_mm_xor_si128(_mm_unpackhi_epi64(k02,k20),_mm_castps_si128(_mm_loadu_ps(mv)))));
+			_mm_storeu_ps(cv + 4,_mm_castsi128_ps(_mm_xor_si128(_mm_unpackhi_epi64(k13,k31),_mm_castps_si128(_mm_loadu_ps(mv + 4)))));
+			_mm_storeu_ps(cv + 8,_mm_castsi128_ps(_mm_xor_si128(_mm_unpacklo_epi64(k20,k02),_mm_castps_si128(_mm_loadu_ps(mv + 8)))));
+			_mm_storeu_ps(cv + 12,_mm_castsi128_ps(_mm_xor_si128(_mm_unpacklo_epi64(k31,k13),_mm_castps_si128(_mm_loadu_ps(mv + 12)))));
 		}
 
 		if (!(++_state.i[8])) {

windows/ZeroTierOne/ZeroTierOne.vcxproj

@@ -40,6 +40,7 @@
     <ClCompile Include="..\..\node\PacketDecoder.cpp" />
     <ClCompile Include="..\..\node\Peer.cpp" />
     <ClCompile Include="..\..\node\Poly1305.cpp" />
+    <ClCompile Include="..\..\node\RoutingTable.cpp" />
     <ClCompile Include="..\..\node\Salsa20.cpp" />
     <ClCompile Include="..\..\node\Service.cpp" />
     <ClCompile Include="..\..\node\SHA512.cpp" />
@@ -51,7 +52,6 @@
     <ClCompile Include="..\..\node\Topology.cpp" />
     <ClCompile Include="..\..\node\UdpSocket.cpp" />
     <ClCompile Include="..\..\node\Utils.cpp" />
-    <ClCompile Include="..\..\node\WindowsEthernetTap.cpp" />
     <ClCompile Include="ServiceBase.cpp" />
     <ClCompile Include="ServiceInstaller.cpp" />
     <ClCompile Include="ZeroTierOneService.cpp" />
@@ -93,6 +93,7 @@
     <ClInclude Include="..\..\node\Path.hpp" />
     <ClInclude Include="..\..\node\Peer.hpp" />
     <ClInclude Include="..\..\node\Poly1305.hpp" />
+    <ClInclude Include="..\..\node\RoutingTable.hpp" />
     <ClInclude Include="..\..\node\RuntimeEnvironment.hpp" />
     <ClInclude Include="..\..\node\Salsa20.hpp" />
     <ClInclude Include="..\..\node\Service.hpp" />
@@ -108,7 +109,6 @@
     <ClInclude Include="..\..\node\Topology.hpp" />
     <ClInclude Include="..\..\node\UdpSocket.hpp" />
     <ClInclude Include="..\..\node\Utils.hpp" />
-    <ClInclude Include="..\..\node\WindowsEthernetTap.hpp" />
     <ClInclude Include="..\..\version.h" />
     <ClInclude Include="resource.h" />
     <ClInclude Include="ServiceBase.h" />
@@ -221,7 +221,7 @@
       <IntrinsicFunctions>true</IntrinsicFunctions>
       <SDLCheck>true</SDLCheck>
       <AdditionalIncludeDirectories>$(SolutionDir)\ext\bin\libcrypto\include</AdditionalIncludeDirectories>
-      <PreprocessorDefinitions>ZT_OFFICIAL_RELEASE;ZT_AUTO_UPDATE;%(PreprocessorDefinitions)</PreprocessorDefinitions>
+      <PreprocessorDefinitions>ZT_OFFICIAL_RELEASE;ZT_AUTO_UPDATE;ZT_SALSA20_SSE;%(PreprocessorDefinitions)</PreprocessorDefinitions>
       <RuntimeLibrary>MultiThreaded</RuntimeLibrary>
       <EnableEnhancedInstructionSet>NoExtensions</EnableEnhancedInstructionSet>
       <StringPooling>true</StringPooling>
@@ -245,7 +245,7 @@
       <IntrinsicFunctions>true</IntrinsicFunctions>
       <SDLCheck>true</SDLCheck>
       <AdditionalIncludeDirectories>$(SolutionDir)\ext\bin\libcrypto\include</AdditionalIncludeDirectories>
-      <PreprocessorDefinitions>ZT_OFFICIAL_RELEASE;ZT_AUTO_UPDATE;%(PreprocessorDefinitions)</PreprocessorDefinitions>
+      <PreprocessorDefinitions>ZT_OFFICIAL_RELEASE;ZT_AUTO_UPDATE;ZT_SALSA20_SSE;%(PreprocessorDefinitions)</PreprocessorDefinitions>
       <RuntimeLibrary>MultiThreaded</RuntimeLibrary>
       <EnableEnhancedInstructionSet>NotSet</EnableEnhancedInstructionSet>
       <StringPooling>true</StringPooling>

windows/ZeroTierOne/ZeroTierOne.vcxproj.filters

@@ -120,7 +120,7 @@
     <ClCompile Include="..\..\node\TcpSocket.cpp">
       <Filter>Source Files</Filter>
     </ClCompile>
-    <ClCompile Include="..\..\node\WindowsEthernetTap.cpp">
+    <ClCompile Include="..\..\node\RoutingTable.cpp">
       <Filter>Source Files</Filter>
     </ClCompile>
   </ItemGroup>
@@ -287,15 +287,15 @@
     <ClInclude Include="..\..\node\TcpSocket.hpp">
       <Filter>Header Files</Filter>
     </ClInclude>
-    <ClInclude Include="..\..\node\WindowsEthernetTap.hpp">
-      <Filter>Header Files</Filter>
-    </ClInclude>
     <ClInclude Include="..\..\node\AntiRecursion.hpp">
       <Filter>Header Files</Filter>
     </ClInclude>
     <ClInclude Include="..\..\version.h">
       <Filter>Header Files</Filter>
     </ClInclude>
+    <ClInclude Include="..\..\node\RoutingTable.hpp">
+      <Filter>Header Files</Filter>
+    </ClInclude>
   </ItemGroup>
   <ItemGroup>
     <ResourceCompile Include="ZeroTierOne.rc">