xtensa: add s32c1i-based atomic ops implementations

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
diff --git a/arch/xtensa/include/asm/atomic.h b/arch/xtensa/include/asm/atomic.h
index 24f50ca..c3f2891 100644
--- a/arch/xtensa/include/asm/atomic.h
+++ b/arch/xtensa/include/asm/atomic.h
@@ -66,19 +66,35 @@
  */
 static inline void atomic_add(int i, atomic_t * v)
 {
-    unsigned int vval;
+#if XCHAL_HAVE_S32C1I
+	unsigned long tmp;
+	int result;
 
-    __asm__ __volatile__(
-	"rsil    a15, "__stringify(LOCKLEVEL)"\n\t"
-	"l32i    %0, %2, 0              \n\t"
-	"add     %0, %0, %1             \n\t"
-	"s32i    %0, %2, 0              \n\t"
-	"wsr     a15, ps                \n\t"
-	"rsync                          \n"
-	: "=&a" (vval)
-	: "a" (i), "a" (v)
-	: "a15", "memory"
-	);
+	__asm__ __volatile__(
+			"1:     l32i    %1, %3, 0\n"
+			"       wsr     %1, scompare1\n"
+			"       add     %0, %1, %2\n"
+			"       s32c1i  %0, %3, 0\n"
+			"       bne     %0, %1, 1b\n"
+			: "=&a" (result), "=&a" (tmp)
+			: "a" (i), "a" (v)
+			: "memory"
+			);
+#else
+	unsigned int vval;
+
+	__asm__ __volatile__(
+			"       rsil    a15, "__stringify(LOCKLEVEL)"\n"
+			"       l32i    %0, %2, 0\n"
+			"       add     %0, %0, %1\n"
+			"       s32i    %0, %2, 0\n"
+			"       wsr     a15, ps\n"
+			"       rsync\n"
+			: "=&a" (vval)
+			: "a" (i), "a" (v)
+			: "a15", "memory"
+			);
+#endif
 }
 
 /**
@@ -90,19 +106,35 @@
  */
 static inline void atomic_sub(int i, atomic_t *v)
 {
-    unsigned int vval;
+#if XCHAL_HAVE_S32C1I
+	unsigned long tmp;
+	int result;
 
-    __asm__ __volatile__(
-	"rsil    a15, "__stringify(LOCKLEVEL)"\n\t"
-	"l32i    %0, %2, 0              \n\t"
-	"sub     %0, %0, %1             \n\t"
-	"s32i    %0, %2, 0              \n\t"
-	"wsr     a15, ps                \n\t"
-	"rsync                          \n"
-	: "=&a" (vval)
-	: "a" (i), "a" (v)
-	: "a15", "memory"
-	);
+	__asm__ __volatile__(
+			"1:     l32i    %1, %3, 0\n"
+			"       wsr     %1, scompare1\n"
+			"       sub     %0, %1, %2\n"
+			"       s32c1i  %0, %3, 0\n"
+			"       bne     %0, %1, 1b\n"
+			: "=&a" (result), "=&a" (tmp)
+			: "a" (i), "a" (v)
+			: "memory"
+			);
+#else
+	unsigned int vval;
+
+	__asm__ __volatile__(
+			"       rsil    a15, "__stringify(LOCKLEVEL)"\n"
+			"       l32i    %0, %2, 0\n"
+			"       sub     %0, %0, %1\n"
+			"       s32i    %0, %2, 0\n"
+			"       wsr     a15, ps\n"
+			"       rsync\n"
+			: "=&a" (vval)
+			: "a" (i), "a" (v)
+			: "a15", "memory"
+			);
+#endif
 }
 
 /*
@@ -111,40 +143,78 @@
 
 static inline int atomic_add_return(int i, atomic_t * v)
 {
-     unsigned int vval;
+#if XCHAL_HAVE_S32C1I
+	unsigned long tmp;
+	int result;
 
-    __asm__ __volatile__(
-	"rsil    a15,"__stringify(LOCKLEVEL)"\n\t"
-	"l32i    %0, %2, 0             \n\t"
-	"add     %0, %0, %1            \n\t"
-	"s32i    %0, %2, 0             \n\t"
-	"wsr     a15, ps               \n\t"
-	"rsync                         \n"
-	: "=&a" (vval)
-	: "a" (i), "a" (v)
-	: "a15", "memory"
-	);
+	__asm__ __volatile__(
+			"1:     l32i    %1, %3, 0\n"
+			"       wsr     %1, scompare1\n"
+			"       add     %0, %1, %2\n"
+			"       s32c1i  %0, %3, 0\n"
+			"       bne     %0, %1, 1b\n"
+			"       add     %0, %0, %2\n"
+			: "=&a" (result), "=&a" (tmp)
+			: "a" (i), "a" (v)
+			: "memory"
+			);
 
-    return vval;
+	return result;
+#else
+	unsigned int vval;
+
+	__asm__ __volatile__(
+			"       rsil    a15,"__stringify(LOCKLEVEL)"\n"
+			"       l32i    %0, %2, 0\n"
+			"       add     %0, %0, %1\n"
+			"       s32i    %0, %2, 0\n"
+			"       wsr     a15, ps\n"
+			"       rsync\n"
+			: "=&a" (vval)
+			: "a" (i), "a" (v)
+			: "a15", "memory"
+			);
+
+	return vval;
+#endif
 }
 
 static inline int atomic_sub_return(int i, atomic_t * v)
 {
-    unsigned int vval;
+#if XCHAL_HAVE_S32C1I
+	unsigned long tmp;
+	int result;
 
-    __asm__ __volatile__(
-	"rsil    a15,"__stringify(LOCKLEVEL)"\n\t"
-	"l32i    %0, %2, 0             \n\t"
-	"sub     %0, %0, %1            \n\t"
-	"s32i    %0, %2, 0             \n\t"
-	"wsr     a15, ps               \n\t"
-	"rsync                         \n"
-	: "=&a" (vval)
-	: "a" (i), "a" (v)
-	: "a15", "memory"
-	);
+	__asm__ __volatile__(
+			"1:     l32i    %1, %3, 0\n"
+			"       wsr     %1, scompare1\n"
+			"       sub     %0, %1, %2\n"
+			"       s32c1i  %0, %3, 0\n"
+			"       bne     %0, %1, 1b\n"
+			"       sub     %0, %0, %2\n"
+			: "=&a" (result), "=&a" (tmp)
+			: "a" (i), "a" (v)
+			: "memory"
+			);
 
-    return vval;
+	return result;
+#else
+	unsigned int vval;
+
+	__asm__ __volatile__(
+			"       rsil    a15,"__stringify(LOCKLEVEL)"\n"
+			"       l32i    %0, %2, 0\n"
+			"       sub     %0, %0, %1\n"
+			"       s32i    %0, %2, 0\n"
+			"       wsr     a15, ps\n"
+			"       rsync\n"
+			: "=&a" (vval)
+			: "a" (i), "a" (v)
+			: "a15", "memory"
+			);
+
+	return vval;
+#endif
 }
 
 /**
@@ -251,38 +321,70 @@
 
 static inline void atomic_clear_mask(unsigned int mask, atomic_t *v)
 {
-    unsigned int all_f = -1;
-    unsigned int vval;
+#if XCHAL_HAVE_S32C1I
+	unsigned long tmp;
+	int result;
 
-    __asm__ __volatile__(
-	"rsil    a15,"__stringify(LOCKLEVEL)"\n\t"
-	"l32i    %0, %2, 0             \n\t"
-	"xor     %1, %4, %3            \n\t"
-	"and     %0, %0, %4            \n\t"
-	"s32i    %0, %2, 0             \n\t"
-	"wsr     a15, ps               \n\t"
-	"rsync                         \n"
-	: "=&a" (vval), "=a" (mask)
-	: "a" (v), "a" (all_f), "1" (mask)
-	: "a15", "memory"
-	);
+	__asm__ __volatile__(
+			"1:     l32i    %1, %3, 0\n"
+			"       wsr     %1, scompare1\n"
+			"       and     %0, %1, %2\n"
+			"       s32c1i  %0, %3, 0\n"
+			"       bne     %0, %1, 1b\n"
+			: "=&a" (result), "=&a" (tmp)
+			: "a" (~mask), "a" (v)
+			: "memory"
+			);
+#else
+	unsigned int all_f = -1;
+	unsigned int vval;
+
+	__asm__ __volatile__(
+			"       rsil    a15,"__stringify(LOCKLEVEL)"\n"
+			"       l32i    %0, %2, 0\n"
+			"       xor     %1, %4, %3\n"
+			"       and     %0, %0, %4\n"
+			"       s32i    %0, %2, 0\n"
+			"       wsr     a15, ps\n"
+			"       rsync\n"
+			: "=&a" (vval), "=a" (mask)
+			: "a" (v), "a" (all_f), "1" (mask)
+			: "a15", "memory"
+			);
+#endif
 }
 
 static inline void atomic_set_mask(unsigned int mask, atomic_t *v)
 {
-    unsigned int vval;
+#if XCHAL_HAVE_S32C1I
+	unsigned long tmp;
+	int result;
 
-    __asm__ __volatile__(
-	"rsil    a15,"__stringify(LOCKLEVEL)"\n\t"
-	"l32i    %0, %2, 0             \n\t"
-	"or      %0, %0, %1            \n\t"
-	"s32i    %0, %2, 0             \n\t"
-	"wsr     a15, ps               \n\t"
-	"rsync                         \n"
-	: "=&a" (vval)
-	: "a" (mask), "a" (v)
-	: "a15", "memory"
-	);
+	__asm__ __volatile__(
+			"1:     l32i    %1, %3, 0\n"
+			"       wsr     %1, scompare1\n"
+			"       or      %0, %1, %2\n"
+			"       s32c1i  %0, %3, 0\n"
+			"       bne     %0, %1, 1b\n"
+			: "=&a" (result), "=&a" (tmp)
+			: "a" (mask), "a" (v)
+			: "memory"
+			);
+#else
+	unsigned int vval;
+
+	__asm__ __volatile__(
+			"       rsil    a15,"__stringify(LOCKLEVEL)"\n"
+			"       l32i    %0, %2, 0\n"
+			"       or      %0, %0, %1\n"
+			"       s32i    %0, %2, 0\n"
+			"       wsr     a15, ps\n"
+			"       rsync\n"
+			: "=&a" (vval)
+			: "a" (mask), "a" (v)
+			: "a15", "memory"
+			);
+#endif
 }
 
 /* Atomic operations are already serializing */
@@ -294,4 +396,3 @@
 #endif /* __KERNEL__ */
 
 #endif /* _XTENSA_ATOMIC_H */
-
diff --git a/arch/xtensa/include/asm/cmpxchg.h b/arch/xtensa/include/asm/cmpxchg.h
index 64dad04..25869a1 100644
--- a/arch/xtensa/include/asm/cmpxchg.h
+++ b/arch/xtensa/include/asm/cmpxchg.h
@@ -22,17 +22,30 @@
 static inline unsigned long
 __cmpxchg_u32(volatile int *p, int old, int new)
 {
-  __asm__ __volatile__("rsil    a15, "__stringify(LOCKLEVEL)"\n\t"
-		       "l32i    %0, %1, 0              \n\t"
-		       "bne	%0, %2, 1f             \n\t"
-		       "s32i    %3, %1, 0              \n\t"
-		       "1:                             \n\t"
-		       "wsr     a15, ps                \n\t"
-		       "rsync                          \n\t"
-		       : "=&a" (old)
-		       : "a" (p), "a" (old), "r" (new)
-		       : "a15", "memory");
-  return old;
+#if XCHAL_HAVE_S32C1I
+	__asm__ __volatile__(
+			"       wsr     %2, scompare1\n"
+			"       s32c1i  %0, %1, 0\n"
+			: "+a" (new)
+			: "a" (p), "a" (old)
+			: "memory"
+			);
+
+	return new;
+#else
+	__asm__ __volatile__(
+			"       rsil    a15, "__stringify(LOCKLEVEL)"\n"
+			"       l32i    %0, %1, 0\n"
+			"       bne     %0, %2, 1f\n"
+			"       s32i    %3, %1, 0\n"
+			"1:\n"
+			"       wsr     a15, ps\n"
+			"       rsync\n"
+			: "=&a" (old)
+			: "a" (p), "a" (old), "r" (new)
+			: "a15", "memory");
+	return old;
+#endif
 }
 /* This function doesn't exist, so you'll get a linker error
  * if something tries to do an invalid cmpxchg(). */
@@ -93,16 +106,32 @@
 
 static inline unsigned long xchg_u32(volatile int * m, unsigned long val)
 {
-  unsigned long tmp;
-  __asm__ __volatile__("rsil    a15, "__stringify(LOCKLEVEL)"\n\t"
-		       "l32i    %0, %1, 0              \n\t"
-		       "s32i    %2, %1, 0              \n\t"
-		       "wsr     a15, ps                \n\t"
-		       "rsync                          \n\t"
-		       : "=&a" (tmp)
-		       : "a" (m), "a" (val)
-		       : "a15", "memory");
-  return tmp;
+#if XCHAL_HAVE_S32C1I
+	unsigned long tmp, result;
+	__asm__ __volatile__(
+			"1:     l32i    %1, %2, 0\n"
+			"       mov     %0, %3\n"
+			"       wsr     %1, scompare1\n"
+			"       s32c1i  %0, %2, 0\n"
+			"       bne     %0, %1, 1b\n"
+			: "=&a" (result), "=&a" (tmp)
+			: "a" (m), "a" (val)
+			: "memory"
+			);
+	return result;
+#else
+	unsigned long tmp;
+	__asm__ __volatile__(
+			"       rsil    a15, "__stringify(LOCKLEVEL)"\n"
+			"       l32i    %0, %1, 0\n"
+			"       s32i    %2, %1, 0\n"
+			"       wsr     a15, ps\n"
+			"       rsync\n"
+			: "=&a" (tmp)
+			: "a" (m), "a" (val)
+			: "a15", "memory");
+	return tmp;
+#endif
 }
 
 #define xchg(ptr,x) ((__typeof__(*(ptr)))__xchg((unsigned long)(x),(ptr),sizeof(*(ptr))))