From c72c5c41f7d3707e6b10bc24acbb73c0e06269b1 Mon Sep 17 00:00:00 2001
From: Pieter Noordhuis <pcnoordhuis@gmail.com>
Date: Tue, 9 Oct 2018 20:39:02 -0700
Subject: [PATCH] Fixes for SSE detection and propagation

On a machine without SSE4.1 the HAVE_SSE flag would still be set.
Because CFLAGS is set to include -msse4.2, the compiler happily
generates SSE 4.2 instructions. Running any resulting SSE-enabled
binary would then result in an illegal instruction error.

The HAVE_SSE check now checks for the presence of one of the SSE 4.1
instructions that is used in the SSE enabled convolutional decoder.
The check must run with -march=native to ensure it checks against the
host machine capabilities.

The HAVE_SSE definition is now propagated to downstream targets that
depend on libcorrect. This means they can now ifdef on HAVE_SSE to
decide whether or not to include libcorrect's SSE specific header.

Confirmed that the HAVE_SSE check now fails on an old machine without
SSE 4.1 (but with SSE 3 and SSSE 3).
---
 CMakeLists.txt | 21 +++++++++++++++++----
 1 file changed, 17 insertions(+), 4 deletions(-)

diff --git a/CMakeLists.txt b/CMakeLists.txt
index bef9687..f1c3022 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -3,6 +3,7 @@ project(Correct)
 include(CheckLibraryExists)
 include(CheckIncludeFiles)
 include(CheckCXXSourceCompiles)
+include(CMakePushCheckState)
 
 if(MSVC)
 set(LIBM "")
@@ -27,15 +28,24 @@ endif(MSVC)
 
 find_library(FEC fec)
 CHECK_LIBRARY_EXISTS(FEC dotprod "" HAVE_LIBFEC)
-check_cxx_source_compiles("
+
+if(NOT CMAKE_CROSSCOMPILING)
+  # Check if host machine can compile with SSE 4.1 intrinsic
+  cmake_push_check_state(RESET)
+  set(CMAKE_REQUIRED_DEFINITIONS -march=native)
+  check_cxx_source_compiles("
     #include <x86intrin.h>
     int main() {
-      __m128i vec;
+      __m128i a;
+      __m128i b;
+      __m128i c = _mm_min_epu16(a, b);
       return 0;
     }" HAVE_SSE)
+  cmake_pop_check_state()
+endif()
 
 if(HAVE_SSE)
-  set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -msse4.2")
+  set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -msse4.1")
 endif()
 
 set(CMAKE_CXX_VISIBILITY_PRESET hidden)
@@ -54,13 +64,16 @@ if(HAVE_SSE)
   set(correct_obj_files $<TARGET_OBJECTS:correct-reed-solomon> $<TARGET_OBJECTS:correct-convolutional> $<TARGET_OBJECTS:correct-convolutional-sse>)
   set(INSTALL_HEADERS ${INSTALL_HEADERS} ${PROJECT_BINARY_DIR}/include/correct-sse.h)
   add_custom_target(correct-sse-h ALL COMMAND ${CMAKE_COMMAND} -E copy ${PROJECT_SOURCE_DIR}/include/correct-sse.h ${PROJECT_BINARY_DIR}/include/correct-sse.h)
-  add_definitions(-DHAVE_SSE=1)
 else()
   set(correct_obj_files $<TARGET_OBJECTS:correct-reed-solomon> $<TARGET_OBJECTS:correct-convolutional>)
 endif()
 add_library(correct SHARED ${correct_obj_files})
 add_library(correct_static ${correct_obj_files})
 set_target_properties(correct_static PROPERTIES OUTPUT_NAME "correct")
+if(HAVE_SSE)
+  target_compile_definitions(correct PUBLIC HAVE_SSE=1)
+  target_compile_definitions(correct_static PUBLIC HAVE_SSE=1)
+endif()
 
 add_subdirectory(util)
 add_subdirectory(tests)