[PATCH] [COMPILER-RT] Implement __fixtfsi, __fixunstfsi

[Joerg Sonnenberger]

================
Comment at: lib/builtins/fixtfsi.c:31
@@ +30,3 @@
+    if ((unsigned int)exponent < significandBits) {
+        // value exceed INT_MAX, reuturn INT_MAX
+        if (exponent >= sizeof(int) * CHAR_BIT){
----------------
Typos, see comments below for return value.

================
Comment at: lib/builtins/fixtfsi.c:48
@@ +47,3 @@
+    else {
+        // value exceed INT_MAX, reuturn INT_MAX
+        if (exponent >= sizeof(int) * CHAR_BIT){
----------------
Typo. Doesn't agree with the comment above in that it provides a deterministic value?

================
Comment at: lib/builtins/fixtfsi.c:50
@@ +49,3 @@
+        if (exponent >= sizeof(int) * CHAR_BIT){
+            return sign * 0x7fffffff;
+        }
----------------
Shouldn't this be sign == 1 ? INT_MAX : INT_MIN, otherwise it would return INT_MIN + 1?

================
Comment at: lib/builtins/fixunstfsi.c:26
@@ +25,3 @@
+ *             value in long double is representable in su_int or is negative
+ *                 (no range checking performed)
+ */
----------------
Code does range check?

================
Comment at: lib/builtins/fixunstfsi.c:42
@@ +41,3 @@
+    if (e >= sizeof(su_int) * CHAR_BIT){
+        return 0xffffffff;
+    }
----------------
Typo in comment, please spell the constant as UINT_MAX too.

http://reviews.llvm.org/D2804