x86/alternatives: force inline stac() and clac()

In this case, we know better than the compiler.

gcc 4.7 (Debian Wheezy) chooses to create translation-unit-local functions
(even for non-debug builds) named stac() and clac(), and calls them.

$ objdump -d xen-syms | grep -c "<stac>:"
6

$ objdump -d xen-syms | grep -o "callq  [0-9a-f]\+ <stac>" | uniq -c
      5 callq  ffff82d0801166c9 <stac>
     20 callq  ffff82d08015ef99 <stac>
      4 callq  ffff82d080165169 <stac>
      8 callq  ffff82d080188cb9 <stac>
      3 callq  ffff82d080228779 <stac>
      4 callq  ffff82d08022c5c9 <stac>

Forcing always_inline removes these functions, and replaces each of the callqs
with the expected 3byte nops.

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>

diff --git a/xen/include/asm-x86/asm_defns.h b/xen/include/asm-x86/asm_defns.h
index f8ef1f84b4ee56eb6849fad0c7a0d0b626cfdfd2..1674c7c1431d6cbb7a36a086c67b4de96a4fc609 100644 (file)
--- a/xen/include/asm-x86/asm_defns.h
+++ b/xen/include/asm-x86/asm_defns.h
@@ -179,13 +179,13 @@ void ret_from_intr(void);
 #define ASM_STAC ASM_AC(STAC)
 #define ASM_CLAC ASM_AC(CLAC)
 #else
-static inline void clac(void)
+static always_inline void clac(void)
 {
     /* Note: a barrier is implicit in alternative() */
     alternative(ASM_NOP3, __stringify(__ASM_CLAC), X86_FEATURE_SMAP);
 }
 
-static inline void stac(void)
+static always_inline void stac(void)
 {
     /* Note: a barrier is implicit in alternative() */
     alternative(ASM_NOP3, __stringify(__ASM_STAC), X86_FEATURE_SMAP);