Better patch for the bug Samuel Neves described

src/arch_x86_64/p448.h

@@ -88,12 +88,22 @@ mask_t
 p448_is_zero (
     const p448_t *in
 );
-             
-static __inline__ void
+  
+static
+#ifndef GCC_HAS_A_BUG_SO_DONT_INLINE_FIELD_BIAS
+    __inline__
+#endif
+void
 p448_bias (
     p448_t *inout,
     int amount
-) __attribute__((unused,always_inline));
+)
+#ifdef GCC_HAS_A_BUG_SO_DONT_INLINE_FIELD_BIAS
+    __attribute__((unused,noinline,optimize("O1")))
+#else
+    __attribute__((unused,always_inline))
+#endif
+;
          
 void
 p448_mul (

src/scalarmul.c

@@ -1,6 +1,11 @@
 /* Copyright (c) 2014 Cryptography Research, Inc.
  * Released under the MIT License.  See LICENSE.txt for license information.
  */
+#if (defined(__GNUC__) && !defined(__clang__) && defined(__x86_64__))
+  /* This works around an apparent compiler bug in GCC, thanks Samuel Neves */
+  #define GCC_HAS_A_BUG_SO_DONT_INLINE_FIELD_BIAS
+#endif
+
 #include "word.h"
 
 #include <stdlib.h>
@@ -63,15 +68,7 @@ cond_negate_tw_pniels (
     cond_negate_tw_niels(&n->n, doNegate);
 }
 
-#if (defined(__GNUC__) && !defined(__clang__) && defined(__x86_64__) && !defined(__AVX2__))
-  /* This works around an apparent compiler bug in GCC, thanks Samuel Neves */
-  static void __attribute__((optimize("O1")))
-  #ifdef __OPTIMIZE_SIZE__
-    #warning "There's a bug in here somewhere with GCC -Os on non-AVX2 platforms"
-  #endif
-#else
-  static __inline__ void
-#endif
+static __inline__ void
 constant_time_lookup_tw_pniels (
     struct tw_pniels_t *out,
     const struct tw_pniels_t *in,