diff --git a/doc/go1.15.html b/doc/go1.15.html
index 15f56ee..93804a9 100644
--- a/doc/go1.15.html
+++ b/doc/go1.15.html
@@ -629,9 +629,9 @@
     </p>
 
     <p><!-- CL 220578 -->
-      It is now possible to call system calls that return floating
-      point values on windows/amd64.
-      TODO: Point to (write?) documentation on how to do this.
+      It is <a href="/pkg/syscall/#Proc.Call">now possible</a> to call
+      system calls that return floating point values
+      on <code>windows/amd64</code>.
     </p>
   </dd>
 </dl>
diff --git a/src/syscall/dll_windows.go b/src/syscall/dll_windows.go
index 819bc5e..c54feec 100644
--- a/src/syscall/dll_windows.go
+++ b/src/syscall/dll_windows.go
@@ -163,7 +163,15 @@
 // The returned error is always non-nil, constructed from the result of GetLastError.
 // Callers must inspect the primary return value to decide whether an error occurred
 // (according to the semantics of the specific function being called) before consulting
-// the error. The error will be guaranteed to contain syscall.Errno.
+// the error. The error always has type syscall.Errno.
+//
+// On amd64, Call can pass and return floating-point values. To pass
+// an argument x with C type "float", use
+// uintptr(math.Float32bits(x)). To pass an argument with C type
+// "double", use uintptr(math.Float64bits(x)). Floating-point return
+// values are returned in r2. The return value for C type "float" is
+// math.Float32frombits(uint32(r2)). For C type "double", it is
+// math.Float64frombits(uint64(r2)).
 func (p *Proc) Call(a ...uintptr) (r1, r2 uintptr, lastErr error) {
 	switch len(a) {
 	case 0: