diff --git a/libkd/kdint_ttype_l.h b/libkd/kdint_ttype_l.h
index 475064c73..a40707472 100644
--- a/libkd/kdint_ttype_l.h
+++ b/libkd/kdint_ttype_l.h
@@ -5,15 +5,27 @@
 
 typedef u64 ttype;
 
+// https://stackoverflow.com/questions/16088282/is-there-a-128-bit-integer-in-gcc
+// gcc: 128-bit ints only available on 64-bit platforms, not 32-bit
+#ifdef __SIZEOF_INT128__
 // GCC only??
 typedef __int128           int128_t;
 typedef unsigned __int128 uint128_t;
 static const uint128_t UINT128_MAX = (uint128_t)((int128_t)(-1L));
-
 #define BIGTTYPE uint128_t
 #define BIGTTYPE_MAX UINT128_MAX
 typedef uint128_t bigttype;
 
+#else
+// Fall back to using just 64-bit types.  This *should* still work okay, because
+// we're careful to check the max possible value before using BIGT types; search
+// for "use_tmath" in the code.
+#define BIGTTYPE uint64_t
+#define BIGTTYPE_MAX UINT64_MAX
+typedef uint64_t bigttype;
+
+#endif
+
 #define TTYPE_INTEGER 1
 
 #define TTYPE_MIN 0