diff --git a/gdb/ChangeLog b/gdb/ChangeLog
index eaea262227..6812f96fca 100644
--- a/gdb/ChangeLog
+++ b/gdb/ChangeLog
@@ -1,3 +1,8 @@
+2018-12-07  Pedro Alves  <palves@redhat.com>
+
+	* dwarf2read.c (get_gdb_index_contents_from_section): Use
+	gdb::make_array_view.
+
 2018-12-05  Philippe Waroquiers  <philippe.waroquiers@skynet.be>
 
 	* language.c (_initialize_language): Fix leak by assigning
diff --git a/gdb/dwarf2read.c b/gdb/dwarf2read.c
index 8bfbd69394..30bc101f11 100644
--- a/gdb/dwarf2read.c
+++ b/gdb/dwarf2read.c
@@ -6189,7 +6189,13 @@ get_gdb_index_contents_from_section (objfile *obj, T *section_owner)
 
   dwarf2_read_section (obj, section);
 
-  return {section->buffer, section->size};
+  /* dwarf2_section_info::size is a bfd_size_type, while
+     gdb::array_view works with size_t.  On 32-bit hosts, with
+     --enable-64-bit-bfd, bfd_size_type is a 64-bit type, while size_t
+     is 32-bit.  So we need an explicit narrowing conversion here.
+     This is fine, because it's impossible to allocate or mmap an
+     array/buffer larger than what size_t can represent.  */
+  return gdb::make_array_view (section->buffer, section->size);
 }
 
 /* Lookup the index cache for the contents of the index associated to