[compiler-rt] [builtins] Support building the 128-bit float functions on ld80 platforms (PR #68132)

[Alexander Richardson via llvm-commits]

================
@@ -6,7 +6,9 @@
 #include <stdio.h>
 #include "fp_lib.h"
 
-#if defined(CRT_HAS_TF_MODE)
+// TODO: This test currently only works if long double is the same as fp_t.
+// On x86_64 we get undefined references to __trunctfxf2 and __extendxftf2
----------------
arichardson wrote:

There was a mismatch in the upper bits of the extended value in at least one test case, I will try to get the log later today.

https://github.com/llvm/llvm-project/pull/68132