DOC: Add notes to explain constrains on overwrite_<>

scipy/linalg/blas.py

@@ -6,6 +6,24 @@
 
 .. versionadded:: 0.12.0
 
+.. note::
+
+   The common ``overwrite_<>`` option in many routines, allows the
+   input arrays to be overwritten to avoid extra memory allocation.
+   However this requires the array to satisfy two conditions
+   which are memory order and the data type to match exactly the
+   order and the type expected by the routine.
+
+   As an example, if you pass a double precision float array to any
+   ``S....`` routine which expects single precision arguments, f2py
+   will create an intermediate array to match the argument types and
+   overwriting will be performed on that intermediate array.
+
+   Similarly, if a C-contiguous array is passed, f2py will pass a
+   FORTRAN-contiguous array internally. Please make sure that these
+   details are satisfied. More information can be found in the f2py
+   documentation.
+
 .. warning::
 
    These functions do little to no error checking.

scipy/linalg/lapack.py

@@ -10,6 +10,24 @@
 
 .. versionadded:: 0.12.0
 
+.. note::
+
+    The common ``overwrite_<>`` option in many routines, allows the
+    input arrays to be overwritten to avoid extra memory allocation.
+    However this requires the array to satisfy two conditions
+    which are memory order and the data type to match exactly the
+    order and the type expected by the routine.
+
+    As an example, if you pass a double precision float array to any
+    ``S....`` routine which expects single precision arguments, f2py
+    will create an intermediate array to match the argument types and
+    overwriting will be performed on that intermediate array.
+
+    Similarly, if a C-contiguous array is passed, f2py will pass a
+    FORTRAN-contiguous array internally. Please make sure that these
+    details are satisfied. More information can be found in the f2py
+    documentation.
+
 .. warning::
 
    These functions do little to no error checking.