s390/raid6: convert to use standard fpu_*() inline assemblies

Move the s390 specific raid6 inline assemblies, make them generic, and
reuse them to implement the raid6 gen/xor implementation.
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>

arch/s390/include/asm/fpu-insn.h

@@ -108,6 +108,14 @@ static __always_inline void fpu_stfpc(unsigned int *fpc)
 		     : "memory");
 }
 
+static __always_inline void fpu_vab(u8 v1, u8 v2, u8 v3)
+{
+	asm volatile("VAB	%[v1],%[v2],%[v3]"
+		     :
+		     : [v1] "I" (v1), [v2] "I" (v2), [v3] "I" (v3)
+		     : "memory");
+}
+
 static __always_inline void fpu_vcksm(u8 v1, u8 v2, u8 v3)
 {
 	asm volatile("VCKSM	%[v1],%[v2],%[v3]"
@@ -116,6 +124,14 @@ static __always_inline void fpu_vcksm(u8 v1, u8 v2, u8 v3)
 		     : "memory");
 }
 
+static __always_inline void fpu_vesravb(u8 v1, u8 v2, u8 v3)
+{
+	asm volatile("VESRAVB	%[v1],%[v2],%[v3]"
+		     :
+		     : [v1] "I" (v1), [v2] "I" (v2), [v3] "I" (v3)
+		     : "memory");
+}
+
 #ifdef CONFIG_CC_IS_CLANG
 
 static __always_inline void fpu_vl(u8 v1, const void *vxr)
@@ -231,6 +247,14 @@ static __always_inline void fpu_vll(u8 v1, u32 index, const void *vxr)
 
 #endif /* CONFIG_CC_IS_CLANG */
 
+static __always_inline void fpu_vlr(u8 v1, u8 v2)
+{
+	asm volatile("VLR	%[v1],%[v2]"
+		     :
+		     : [v1] "I" (v1), [v2] "I" (v2)
+		     : "memory");
+}
+
 static __always_inline void fpu_vlvgf(u8 v, u32 val, u16 index)
 {
 	asm volatile("VLVGF	%[v],%[val],%[index]"
@@ -239,6 +263,22 @@ static __always_inline void fpu_vlvgf(u8 v, u32 val, u16 index)
 		     : "memory");
 }
 
+static __always_inline void fpu_vn(u8 v1, u8 v2, u8 v3)
+{
+	asm volatile("VN	%[v1],%[v2],%[v3]"
+		     :
+		     : [v1] "I" (v1), [v2] "I" (v2), [v3] "I" (v3)
+		     : "memory");
+}
+
+static __always_inline void fpu_vrepib(u8 v1, s16 i2)
+{
+	asm volatile("VREPIB	%[v1],%[i2]"
+		     :
+		     : [v1] "I" (v1), [i2] "K" (i2)
+		     : "memory");
+}
+
 #ifdef CONFIG_CC_IS_CLANG
 
 static __always_inline void fpu_vst(u8 v1, const void *vxr)
@@ -335,6 +375,14 @@ static __always_inline void fpu_vstl(u8 v1, u32 index, const void *vxr)
 
 #endif /* CONFIG_CC_IS_CLANG */
 
+static __always_inline void fpu_vx(u8 v1, u8 v2, u8 v3)
+{
+	asm volatile("VX	%[v1],%[v2],%[v3]"
+		     :
+		     : [v1] "I" (v1), [v2] "I" (v2), [v3] "I" (v3)
+		     : "memory");
+}
+
 static __always_inline void fpu_vzero(u8 v)
 {
 	asm volatile("VZERO	%[v]"

lib/raid6/s390vx.uc

@@ -16,10 +16,10 @@
 
 #define NSIZE 16
 
-static inline void LOAD_CONST(void)
+static __always_inline void LOAD_CONST(void)
 {
-	asm volatile("VREPIB %v24,7");
-	asm volatile("VREPIB %v25,0x1d");
+	fpu_vrepib(24, 0x07);
+	fpu_vrepib(25, 0x1d);
 }
 
 /*
@@ -27,10 +27,7 @@ static inline void LOAD_CONST(void)
  * vector register y left by 1 bit and stores the result in
  * vector register x.
  */
-static inline void SHLBYTE(int x, int y)
-{
-	asm volatile ("VAB %0,%1,%1" : : "i" (x), "i" (y));
-}
+#define SHLBYTE(x, y)		fpu_vab(x, y, y)
 
 /*
  * For each of the 16 bytes in the vector register y the MASK()
@@ -38,45 +35,13 @@ static inline void SHLBYTE(int x, int y)
  * or 0x00 if the high bit is 0. The result is stored in vector
  * register x.
  */
-static inline void MASK(int x, int y)
-{
-	asm volatile ("VESRAVB	%0,%1,24" : : "i" (x), "i" (y));
-}
-
-static inline void AND(int x, int y, int z)
-{
-	asm volatile ("VN %0,%1,%2" : : "i" (x), "i" (y), "i" (z));
-}
-
-static inline void XOR(int x, int y, int z)
-{
-	asm volatile ("VX %0,%1,%2" : : "i" (x), "i" (y), "i" (z));
-}
+#define MASK(x, y)		fpu_vesravb(x, y, 24)
 
-static inline void LOAD_DATA(int x, u8 *ptr)
-{
-	typedef struct { u8 _[16 * $#]; } addrtype;
-	register addrtype *__ptr asm("1") = (addrtype *) ptr;
-
-	asm volatile ("VLM %2,%3,0,%1"
-		      : : "m" (*__ptr), "a" (__ptr), "i" (x),
-			  "i" (x + $# - 1));
-}
-
-static inline void STORE_DATA(int x, u8 *ptr)
-{
-	typedef struct { u8 _[16 * $#]; } addrtype;
-	register addrtype *__ptr asm("1") = (addrtype *) ptr;
-
-	asm volatile ("VSTM %2,%3,0,1"
-		      : "=m" (*__ptr) : "a" (__ptr), "i" (x),
-			"i" (x + $# - 1));
-}
-
-static inline void COPY_VEC(int x, int y)
-{
-	asm volatile ("VLR %0,%1" : : "i" (x), "i" (y));
-}
+#define AND(x, y, z)		fpu_vn(x, y, z)
+#define XOR(x, y, z)		fpu_vx(x, y, z)
+#define LOAD_DATA(x, ptr)	fpu_vlm(x, x + $# - 1, ptr)
+#define STORE_DATA(x, ptr)	fpu_vstm(x, x + $# - 1, ptr)
+#define COPY_VEC(x, y)		fpu_vlr(x, y)
 
 static void raid6_s390vx$#_gen_syndrome(int disks, size_t bytes, void **ptrs)
 {