[llvm] Enable logf128 constant folding for hosts with 128bit long double (PR #96287)

[Eli Friedman via llvm-commits]

================
@@ -9,18 +9,17 @@
 #ifndef LLVM_FLOAT128
 #define LLVM_FLOAT128
 
+#include <cmath>
+
 namespace llvm {
 
-#if defined(__clang__) && defined(__FLOAT128__) &&                             \
-    defined(__SIZEOF_INT128__) && !defined(__LONG_DOUBLE_IBM128__)
-#define HAS_IEE754_FLOAT128
-typedef __float128 float128;
-#elif defined(__FLOAT128__) && defined(__SIZEOF_INT128__) &&                   \
-    !defined(__LONG_DOUBLE_IBM128__) &&                                        \
-    (defined(__GNUC__) || defined(__GNUG__))
+#ifdef HAS_LOGF128
+#if (__LDBL_MANT_DIG__ == 113) && !defined(__LONG_DOUBLE_IBM128__) &&          \
----------------
efriedma-quic wrote:

(I was suggesting __LDBL_MANT_DIG__ with the old version that explicitly wrote out the type "long double".  It should be safe to assume that if logf128 returns a long double, then long double is actually an IEEE 128-bit float.)

https://github.com/llvm/llvm-project/pull/96287