merge

7 files changed, 39 insertions, 13 deletions

diff --git a/Eigen/src/Core/DenseStorage.h b/Eigen/src/Core/DenseStorage.h
index 3c02a1025..a8bb8a624 100644
--- a/Eigen/src/Core/DenseStorage.h
+++ b/Eigen/src/Core/DenseStorage.h
@@ -420,7 +420,7 @@ template<typename T, int _Options> class DenseStorage<T, Dynamic, Dynamic, Dynam
       if(size != m_rows*m_cols)
       {
         internal::conditional_aligned_delete_auto<T,(_Options&DontAlign)==0>(m_data, m_rows*m_cols);
-        if (size)
+        if (size>0) // >0 and not simply !=0 to let the compiler knows that size cannot be negative
           m_data = internal::conditional_aligned_new_auto<T,(_Options&DontAlign)==0>(size);
         else
           m_data = 0;
@@ -497,7 +497,7 @@ template<typename T, int _Rows, int _Options> class DenseStorage<T, Dynamic, _Ro
       if(size != _Rows*m_cols)
       {
         internal::conditional_aligned_delete_auto<T,(_Options&DontAlign)==0>(m_data, _Rows*m_cols);
-        if (size)
+        if (size>0) // >0 and not simply !=0 to let the compiler knows that size cannot be negative
           m_data = internal::conditional_aligned_new_auto<T,(_Options&DontAlign)==0>(size);
         else
           m_data = 0;
@@ -573,7 +573,7 @@ template<typename T, int _Cols, int _Options> class DenseStorage<T, Dynamic, Dyn
       if(size != m_rows*_Cols)
       {
         internal::conditional_aligned_delete_auto<T,(_Options&DontAlign)==0>(m_data, _Cols*m_rows);
-        if (size)
+        if (size>0) // >0 and not simply !=0 to let the compiler knows that size cannot be negative
           m_data = internal::conditional_aligned_new_auto<T,(_Options&DontAlign)==0>(size);
         else
           m_data = 0;
diff --git a/Eigen/src/Core/MathFunctions.h b/Eigen/src/Core/MathFunctions.h
index 72116e144..34dd15d85 100644
--- a/Eigen/src/Core/MathFunctions.h
+++ b/Eigen/src/Core/MathFunctions.h
@@ -1217,7 +1217,8 @@ inline int log2(int x)
 
 /** \returns the square root of \a x.
   *
-  * It is essentially equivalent to \code using std::sqrt; return sqrt(x); \endcode,
+  * It is essentially equivalent to
+  * \code using std::sqrt; return sqrt(x); \endcode
   * but slightly faster for float/double and some compilers (e.g., gcc), thanks to
   * specializations when SSE is enabled.
   *
diff --git a/Eigen/src/Core/arch/SSE/PacketMath.h b/Eigen/src/Core/arch/SSE/PacketMath.h
index 2e815c0c5..99d55d5e9 100755
--- a/Eigen/src/Core/arch/SSE/PacketMath.h
+++ b/Eigen/src/Core/arch/SSE/PacketMath.h
@@ -28,7 +28,7 @@ namespace internal {
 #endif
 #endif
 
-#if (defined EIGEN_VECTORIZE_AVX) && (EIGEN_COMP_GNUC_STRICT || EIGEN_COMP_MINGW) && (__GXX_ABI_VERSION < 1004)
+#if ((defined EIGEN_VECTORIZE_AVX) && (EIGEN_COMP_GNUC_STRICT || EIGEN_COMP_MINGW) && (__GXX_ABI_VERSION < 1004)) || EIGEN_OS_QNX
 // With GCC's default ABI version, a __m128 or __m256 are the same types and therefore we cannot
 // have overloads for both types without linking error.
 // One solution is to increase ABI version using -fabi-version=4 (or greater).
diff --git a/Eigen/src/Core/util/ConfigureVectorization.h b/Eigen/src/Core/util/ConfigureVectorization.h
index e75c7d89e..a2743624e 100644
--- a/Eigen/src/Core/util/ConfigureVectorization.h
+++ b/Eigen/src/Core/util/ConfigureVectorization.h
@@ -379,10 +379,12 @@
   #include <cuda_fp16.h>
 #endif
 
-#if defined(EIGEN_HIP_DEVICE_COMPILE)
-
+#if defined(EIGEN_HIPCC)
   #define EIGEN_VECTORIZE_GPU
   #include <hip/hip_vector_types.h>
+#endif
+
+#if defined(EIGEN_HIP_DEVICE_COMPILE)
 
   #define EIGEN_HAS_HIP_FP16
   #include <hip/hip_fp16.h>
diff --git a/Eigen/src/Core/util/IntegralConstant.h b/Eigen/src/Core/util/IntegralConstant.h
index bf99cd8ab..caeea232d 100644
--- a/Eigen/src/Core/util/IntegralConstant.h
+++ b/Eigen/src/Core/util/IntegralConstant.h
@@ -55,7 +55,9 @@ public:
   operator int() const { return value; }
   FixedInt() {}
   FixedInt( VariableAndFixedInt<N> other) {
-    EIGEN_ONLY_USED_FOR_DEBUG(other);
+    #ifndef EIGEN_INTERNAL_DEBUGGING
+    EIGEN_UNUSED_VARIABLE(other);
+    #endif
     eigen_internal_assert(int(other)==N);
   }
 
diff --git a/Eigen/src/Core/util/Memory.h b/Eigen/src/Core/util/Memory.h
index 9dd2e0252..6664770f3 100644
--- a/Eigen/src/Core/util/Memory.h
+++ b/Eigen/src/Core/util/Memory.h
@@ -96,10 +96,16 @@ inline void throw_std_bad_alloc()
 /** \internal Like malloc, but the returned pointer is guaranteed to be 16-byte aligned.
   * Fast, but wastes 16 additional bytes of memory. Does not throw any exception.
   */
-inline void* handmade_aligned_malloc(std::size_t size, std::size_t alignment = EIGEN_DEFAULT_ALIGN_BYTES)
+EIGEN_DEVICE_FUNC inline void* handmade_aligned_malloc(std::size_t size, std::size_t alignment = EIGEN_DEFAULT_ALIGN_BYTES)
 {
   eigen_assert(alignment >= sizeof(void*) && (alignment & -alignment) == alignment && "Alignment must be at least sizeof(void*) and a power of 2");
+  
+#if defined(EIGEN_HIP_DEVICE_COMPILE)
+  void *original = ::malloc(size+alignment);
+#else
   void *original = std::malloc(size+alignment);
+#endif
+  
   if (original == 0) return 0;
   void *aligned = reinterpret_cast<void*>((reinterpret_cast<std::size_t>(original) & ~(std::size_t(alignment-1))) + alignment);
   *(reinterpret_cast<void**>(aligned) - 1) = original;
@@ -107,9 +113,15 @@ inline void* handmade_aligned_malloc(std::size_t size, std::size_t alignment = E
 }
 
 /** \internal Frees memory allocated with handmade_aligned_malloc */
-inline void handmade_aligned_free(void *ptr)
+EIGEN_DEVICE_FUNC inline void handmade_aligned_free(void *ptr)
 {
-  if (ptr) std::free(*(reinterpret_cast<void**>(ptr) - 1));
+  if (ptr) {
+#if defined(EIGEN_HIP_DEVICE_COMPILE)
+    ::free(*(reinterpret_cast<void**>(ptr) - 1));
+#else
+    std::free(*(reinterpret_cast<void**>(ptr) - 1));
+#endif
+  }
 }
 
 /** \internal
@@ -872,6 +884,15 @@ public:
 
   ~aligned_allocator() {}
 
+  #if EIGEN_COMP_GNUC_STRICT && EIGEN_GNUC_AT_LEAST(7,0)
+  // In gcc std::allocator::max_size() is bugged making gcc triggers a warning:
+  // eigen/Eigen/src/Core/util/Memory.h:189:12: warning: argument 1 value '18446744073709551612' exceeds maximum object size 9223372036854775807
+  // See https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87544
+  size_type max_size() const {
+    return (std::numeric_limits<std::ptrdiff_t>::max)()/sizeof(T);
+  }
+  #endif
+
   pointer allocate(size_type num, const void* /*hint*/ = 0)
   {
     internal::check_size_for_overflow<T>(num);
diff --git a/Eigen/src/plugins/ReshapedMethods.h b/Eigen/src/plugins/ReshapedMethods.h
index 2466d24cb..538636ab5 100644
--- a/Eigen/src/plugins/ReshapedMethods.h
+++ b/Eigen/src/plugins/ReshapedMethods.h
@@ -105,7 +105,7 @@ EIGEN_DEVICE_FUNC
 inline Reshaped<EIGEN_RESHAPED_METHOD_CONST Derived,
                 internal::get_compiletime_reshape_size<NRowsType,NColsType,SizeAtCompileTime>::value,
                 internal::get_compiletime_reshape_size<NColsType,NRowsType,SizeAtCompileTime>::value,
-                Order==AutoOrder?Flags&RowMajorBit:Order>
+                (Order==AutoOrder?Flags&RowMajorBit:Order)>
 reshaped(NRowsType nRows, NColsType nCols) EIGEN_RESHAPED_METHOD_CONST
 {
   return Reshaped<EIGEN_RESHAPED_METHOD_CONST Derived,
@@ -128,7 +128,7 @@ reshaped() EIGEN_RESHAPED_METHOD_CONST
 
 template<int Order>
 EIGEN_DEVICE_FUNC
-inline Reshaped<EIGEN_RESHAPED_METHOD_CONST Derived, SizeAtCompileTime, 1, Order==AutoOrder?Flags&RowMajorBit:Order>
+inline Reshaped<EIGEN_RESHAPED_METHOD_CONST Derived, SizeAtCompileTime, 1, (Order==AutoOrder?Flags&RowMajorBit:Order)>
 reshaped() EIGEN_RESHAPED_METHOD_CONST
 {
   EIGEN_STATIC_ASSERT(Order==RowMajor || Order==ColMajor || Order==AutoOrder, INVALID_TEMPLATE_PARAMETER);