[atomics.{ref,types}.float] Add missing \tcode

Author: tkoeppe

source/threads.tex

@@ -4103,11 +4103,11 @@
 If the result is not a representable value for its type\iref{expr.pre},
 the result is unspecified,
 but the operations otherwise have no undefined behavior.
-Atomic arithmetic operations on \placeholder{floating-point-type}
+Atomic arithmetic operations on \tcode{\placeholder{floating-point-type}}
 should conform to the \tcode{numeric_limits<\placeholder{floating-point-type}>}
 traits associated with the floating-point type\iref{limits.syn}.
 The floating-point environment\iref{cfenv}
-for atomic arithmetic operations on \placeholder{floating-point-type}
+for atomic arithmetic operations on \tcode{\placeholder{floating-point-type}}
 may be different than the calling thread's floating-point environment.
 The arithmetic rules of floating-point atomic modify-write operations
 may be different from operations on floating-point types or
@@ -5651,11 +5651,11 @@
 If the result is not a representable value for its type\iref{expr.pre}
 the result is unspecified,
 but the operations otherwise have no undefined behavior.
-Atomic arithmetic operations on \placeholder{floating-point-type}
+Atomic arithmetic operations on \tcode{\placeholder{floating-point-type}}
 should conform to the \tcode{numeric_limits<\placeholder{floating-point-type}>}
 traits associated with the floating-point type\iref{limits.syn}.
 The floating-point environment\iref{cfenv} for
-atomic arithmetic operations on \placeholder{floating-point-type}
+atomic arithmetic operations on \tcode{\placeholder{floating-point-type}}
 may be different than the calling thread's floating-point environment.
 The arithmetic rules of floating-point atomic modify-write operations
 may be different from operations on