x86/asm: 'Simplify' GEN_*_RMWcc() macros

Currently the GEN_*_RMWcc() macros include a return statement, which
pretty much mandates we directly wrap them in a (inline) function.

Macros with return statements are tricky and, as per the above, limit
use, so remove the return statement and make them
statement-expressions. This allows them to be used more widely.

Also, shuffle the arguments a bit. Place the @cc argument as 3rd, this
makes it consistent between UNARY and BINARY, but more importantly, it
makes the @arg0 argument last.

Since the @arg0 argument is now last, we can do CPP trickery and make
it an optional argument, simplifying the users; 17 out of 18
occurences do not need this argument.

Finally, change to asm symbolic names, instead of the numeric ordering
of operands, which allows us to get rid of __BINARY_RMWcc_ARG and get
cleaner code overall.

Signed-off-by: Peter Zijlstra (Intel) <[e-mail]>
Cc: [e-mail]
Cc: Linus Torvalds <[e-mail]>
Cc: Andy Lutomirski <[e-mail]>
Cc: Borislav Petkov <[e-mail]>
Cc: Brian Gerst <[e-mail]>
Cc: Denys Vlasenko <[e-mail]>
Cc: H. Peter Anvin <[e-mail]>
Cc: Linus Torvalds <[e-mail]>
Cc: Peter Zijlstra <[e-mail]>
Cc: Thomas Gleixner <[e-mail]>
Cc: Peter Zijlstra <[e-mail]>
Cc: Thomas Gleixner <[e-mail]>
Cc: [e-mail]
Cc: [e-mail]
Link: https://lkml.kernel.org/r/[e-mail]
Signed-off-by: Ingo Molnar <[e-mail]>

arch/x86/include/asm/atomic.h

···
  */
 static __always_inline bool arch_atomic_sub_and_test(int i, atomic_t *v)
 {
-	GEN_BINARY_RMWcc(LOCK_PREFIX "subl", v->counter, "er", i, "%0", e);
+	return GEN_BINARY_RMWcc(LOCK_PREFIX "subl", v->counter, e, "er", i);
 }
 #define arch_atomic_sub_and_test arch_atomic_sub_and_test
 
···
  */
 static __always_inline bool arch_atomic_dec_and_test(atomic_t *v)
 {
-	GEN_UNARY_RMWcc(LOCK_PREFIX "decl", v->counter, "%0", e);
+	return GEN_UNARY_RMWcc(LOCK_PREFIX "decl", v->counter, e);
 }
 #define arch_atomic_dec_and_test arch_atomic_dec_and_test
 
···
  */
 static __always_inline bool arch_atomic_inc_and_test(atomic_t *v)
 {
-	GEN_UNARY_RMWcc(LOCK_PREFIX "incl", v->counter, "%0", e);
+	return GEN_UNARY_RMWcc(LOCK_PREFIX "incl", v->counter, e);
 }
 #define arch_atomic_inc_and_test arch_atomic_inc_and_test
 
···
  */
 static __always_inline bool arch_atomic_add_negative(int i, atomic_t *v)
 {
-	GEN_BINARY_RMWcc(LOCK_PREFIX "addl", v->counter, "er", i, "%0", s);
+	return GEN_BINARY_RMWcc(LOCK_PREFIX "addl", v->counter, s, "er", i);
 }
 #define arch_atomic_add_negative arch_atomic_add_negative
 

arch/x86/include/asm/atomic64_64.h

···
  */
 static inline bool arch_atomic64_sub_and_test(long i, atomic64_t *v)
 {
-	GEN_BINARY_RMWcc(LOCK_PREFIX "subq", v->counter, "er", i, "%0", e);
+	return GEN_BINARY_RMWcc(LOCK_PREFIX "subq", v->counter, e, "er", i);
 }
 #define arch_atomic64_sub_and_test arch_atomic64_sub_and_test
 
···
  */
 static inline bool arch_atomic64_dec_and_test(atomic64_t *v)
 {
-	GEN_UNARY_RMWcc(LOCK_PREFIX "decq", v->counter, "%0", e);
+	return GEN_UNARY_RMWcc(LOCK_PREFIX "decq", v->counter, e);
 }
 #define arch_atomic64_dec_and_test arch_atomic64_dec_and_test
 
···
  */
 static inline bool arch_atomic64_inc_and_test(atomic64_t *v)
 {
-	GEN_UNARY_RMWcc(LOCK_PREFIX "incq", v->counter, "%0", e);
+	return GEN_UNARY_RMWcc(LOCK_PREFIX "incq", v->counter, e);
 }
 #define arch_atomic64_inc_and_test arch_atomic64_inc_and_test
 
···
  */
 static inline bool arch_atomic64_add_negative(long i, atomic64_t *v)
 {
-	GEN_BINARY_RMWcc(LOCK_PREFIX "addq", v->counter, "er", i, "%0", s);
+	return GEN_BINARY_RMWcc(LOCK_PREFIX "addq", v->counter, s, "er", i);
 }
 #define arch_atomic64_add_negative arch_atomic64_add_negative
 

arch/x86/include/asm/bitops.h

···
  */
 static __always_inline bool test_and_set_bit(long nr, volatile unsigned long *addr)
 {
-	GEN_BINARY_RMWcc(LOCK_PREFIX __ASM_SIZE(bts),
-	                 *addr, "Ir", nr, "%0", c);
+	return GEN_BINARY_RMWcc(LOCK_PREFIX __ASM_SIZE(bts), *addr, c, "Ir", nr);
 }
 
 /**
···
  */
 static __always_inline bool test_and_clear_bit(long nr, volatile unsigned long *addr)
 {
-	GEN_BINARY_RMWcc(LOCK_PREFIX __ASM_SIZE(btr),
-	                 *addr, "Ir", nr, "%0", c);
+	return GEN_BINARY_RMWcc(LOCK_PREFIX __ASM_SIZE(btr), *addr, c, "Ir", nr);
 }
 
 /**
···
  */
 static __always_inline bool test_and_change_bit(long nr, volatile unsigned long *addr)
 {
-	GEN_BINARY_RMWcc(LOCK_PREFIX __ASM_SIZE(btc),
-	                 *addr, "Ir", nr, "%0", c);
+	return GEN_BINARY_RMWcc(LOCK_PREFIX __ASM_SIZE(btc), *addr, c, "Ir", nr);
 }
 
 static __always_inline bool constant_test_bit(long nr, const volatile unsigned long *addr)

arch/x86/include/asm/local.h

···
  */
 static inline bool local_sub_and_test(long i, local_t *l)
 {
-	GEN_BINARY_RMWcc(_ASM_SUB, l->a.counter, "er", i, "%0", e);
+	return GEN_BINARY_RMWcc(_ASM_SUB, l->a.counter, e, "er", i);
 }
 
 /**
···
  */
 static inline bool local_dec_and_test(local_t *l)
 {
-	GEN_UNARY_RMWcc(_ASM_DEC, l->a.counter, "%0", e);
+	return GEN_UNARY_RMWcc(_ASM_DEC, l->a.counter, e);
 }
 
 /**
···
  */
 static inline bool local_inc_and_test(local_t *l)
 {
-	GEN_UNARY_RMWcc(_ASM_INC, l->a.counter, "%0", e);
+	return GEN_UNARY_RMWcc(_ASM_INC, l->a.counter, e);
 }
 
 /**
···
  */
 static inline bool local_add_negative(long i, local_t *l)
 {
-	GEN_BINARY_RMWcc(_ASM_ADD, l->a.counter, "er", i, "%0", s);
+	return GEN_BINARY_RMWcc(_ASM_ADD, l->a.counter, s, "er", i);
 }
 
 /**

arch/x86/include/asm/preempt.h

···
  */
 static __always_inline bool __preempt_count_dec_and_test(void)
 {
-	GEN_UNARY_RMWcc("decl", __preempt_count, __percpu_arg(0), e);
+	return GEN_UNARY_RMWcc("decl", __preempt_count, e, __percpu_arg([var]));
 }
 
 /*

arch/x86/include/asm/refcount.h

···
 static __always_inline __must_check
 bool refcount_sub_and_test(unsigned int i, refcount_t *r)
 {
-	GEN_BINARY_SUFFIXED_RMWcc(LOCK_PREFIX "subl",
-				  "REFCOUNT_CHECK_LT_ZERO counter=\"%0\"",
-				  r->refs.counter, "er", i, "%0", e, "cx");
+
+	return GEN_BINARY_SUFFIXED_RMWcc(LOCK_PREFIX "subl",
+					 "REFCOUNT_CHECK_LT_ZERO counter=\"%[var]\"",
+					 r->refs.counter, e, "er", i, "cx");
 }
 
 static __always_inline __must_check bool refcount_dec_and_test(refcount_t *r)
 {
-	GEN_UNARY_SUFFIXED_RMWcc(LOCK_PREFIX "decl",
-				 "REFCOUNT_CHECK_LT_ZERO counter=\"%0\"",
-				 r->refs.counter, "%0", e, "cx");
+	return GEN_UNARY_SUFFIXED_RMWcc(LOCK_PREFIX "decl",
+					"REFCOUNT_CHECK_LT_ZERO counter=\"%[var]\"",
+					r->refs.counter, e, "cx");
 }
 
 static __always_inline __must_check

arch/x86/include/asm/rmwcc.h

···
 #ifndef _ASM_X86_RMWcc
 #define _ASM_X86_RMWcc
 
+/* This counts to 12. Any more, it will return 13th argument. */
+#define __RMWcc_ARGS(_0, _1, _2, _3, _4, _5, _6, _7, _8, _9, _10, _11, _12, _n, X...) _n
+#define RMWcc_ARGS(X...) __RMWcc_ARGS(, ##X, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0)
+
+#define __RMWcc_CONCAT(a, b) a ## b
+#define RMWcc_CONCAT(a, b) __RMWcc_CONCAT(a, b)
+
 #define __CLOBBERS_MEM(clb...)	"memory", ## clb
 
 #if !defined(__GCC_ASM_FLAG_OUTPUTS__) && defined(CC_HAVE_ASM_GOTO)
 
 /* Use asm goto */
 
-#define __GEN_RMWcc(fullop, var, cc, clobbers, ...)			\
-do {									\
+#define __GEN_RMWcc(fullop, _var, cc, clobbers, ...)			\
+({									\
+	bool c = false;							\
 	asm_volatile_goto (fullop "; j" #cc " %l[cc_label]"		\
-			: : [counter] "m" (var), ## __VA_ARGS__		\
+			: : [var] "m" (_var), ## __VA_ARGS__		\
 			: clobbers : cc_label);				\
-	return 0;							\
-cc_label:								\
-	return 1;							\
-} while (0)
-
-#define __BINARY_RMWcc_ARG	" %1, "
-
+	if (0) {							\
+cc_label:	c = true;						\
+	}								\
+	c;								\
+})
 
 #else /* defined(__GCC_ASM_FLAG_OUTPUTS__) || !defined(CC_HAVE_ASM_GOTO) */
 
 /* Use flags output or a set instruction */
 
-#define __GEN_RMWcc(fullop, var, cc, clobbers, ...)			\
-do {									\
+#define __GEN_RMWcc(fullop, _var, cc, clobbers, ...)			\
+({									\
 	bool c;								\
 	asm volatile (fullop CC_SET(cc)					\
-			: [counter] "+m" (var), CC_OUT(cc) (c)		\
+			: [var] "+m" (_var), CC_OUT(cc) (c)		\
 			: __VA_ARGS__ : clobbers);			\
-	return c;							\
-} while (0)
-
-#define __BINARY_RMWcc_ARG	" %2, "
+	c;								\
+})
 
 #endif /* defined(__GCC_ASM_FLAG_OUTPUTS__) || !defined(CC_HAVE_ASM_GOTO) */
 
-#define GEN_UNARY_RMWcc(op, var, arg0, cc)				\
+#define GEN_UNARY_RMWcc_4(op, var, cc, arg0)				\
 	__GEN_RMWcc(op " " arg0, var, cc, __CLOBBERS_MEM())
 
-#define GEN_UNARY_SUFFIXED_RMWcc(op, suffix, var, arg0, cc, clobbers...)\
-	__GEN_RMWcc(op " " arg0 "\n\t" suffix, var, cc,			\
+#define GEN_UNARY_RMWcc_3(op, var, cc)					\
+	GEN_UNARY_RMWcc_4(op, var, cc, "%[var]")
+
+#define GEN_UNARY_RMWcc(X...) RMWcc_CONCAT(GEN_UNARY_RMWcc_, RMWcc_ARGS(X))(X)
+
+#define GEN_BINARY_RMWcc_6(op, var, cc, vcon, _val, arg0)		\
+	__GEN_RMWcc(op " %[val], " arg0, var, cc,			\
+		    __CLOBBERS_MEM(), [val] vcon (_val))
+
+#define GEN_BINARY_RMWcc_5(op, var, cc, vcon, val)			\
+	GEN_BINARY_RMWcc_6(op, var, cc, vcon, val, "%[var]")
+
+#define GEN_BINARY_RMWcc(X...) RMWcc_CONCAT(GEN_BINARY_RMWcc_, RMWcc_ARGS(X))(X)
+
+#define GEN_UNARY_SUFFIXED_RMWcc(op, suffix, var, cc, clobbers...)	\
+	__GEN_RMWcc(op " %[var]\n\t" suffix, var, cc,			\
 		    __CLOBBERS_MEM(clobbers))
 
-#define GEN_BINARY_RMWcc(op, var, vcon, val, arg0, cc)			\
-	__GEN_RMWcc(op __BINARY_RMWcc_ARG arg0, var, cc,		\
-		    __CLOBBERS_MEM(), vcon (val))
-
-#define GEN_BINARY_SUFFIXED_RMWcc(op, suffix, var, vcon, val, arg0, cc,	\
-				  clobbers...)				\
-	__GEN_RMWcc(op __BINARY_RMWcc_ARG arg0 "\n\t" suffix, var, cc,	\
-		    __CLOBBERS_MEM(clobbers), vcon (val))
+#define GEN_BINARY_SUFFIXED_RMWcc(op, suffix, var, cc, vcon, _val, clobbers...)\
+	__GEN_RMWcc(op " %[val], %[var]\n\t" suffix, var, cc,		\
+		    __CLOBBERS_MEM(clobbers), [val] vcon (_val))
 
 #endif /* _ASM_X86_RMWcc */