[S390] use inline assembly contraints available with gcc 3.3.3

Drop support to compile the kernel with gcc versions older than 3.3.3.
This allows us to use the "Q" inline assembly contraint on some more
inline assemblies without duplicating a lot of complex code (e.g. __xchg
and __cmpxchg). The distinction for older gcc versions can be removed
which saves a few lines and simplifies the code.

Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

11 files changed, 226 insertions, 388 deletions

diff --git a/arch/s390/include/asm/atomic.h b/arch/s390/include/asm/atomic.h
index 2a113d6..451bfbb 100644
--- a/arch/s390/include/asm/atomic.h
+++ b/arch/s390/include/asm/atomic.h
@@ -18,8 +18,6 @@
 
 #define ATOMIC_INIT(i)  { (i) }
 
-#if __GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ > 2)
-
 #define __CS_LOOP(ptr, op_val, op_string) ({				\
 	int old_val, new_val;						\
 	asm volatile(							\
@@ -35,26 +33,6 @@
 	new_val;							\
 })
 
-#else /* __GNUC__ */
-
-#define __CS_LOOP(ptr, op_val, op_string) ({				\
-	int old_val, new_val;						\
-	asm volatile(							\
-		"	l	%0,0(%3)\n"				\
-		"0:	lr	%1,%0\n"				\
-		op_string "	%1,%4\n"				\
-		"	cs	%0,%1,0(%3)\n"				\
-		"	jl	0b"					\
-		: "=&d" (old_val), "=&d" (new_val),			\
-		  "=m" (((atomic_t *)(ptr))->counter)			\
-		: "a" (ptr), "d" (op_val),				\
-		  "m" (((atomic_t *)(ptr))->counter)			\
-		: "cc", "memory");					\
-	new_val;							\
-})
-
-#endif /* __GNUC__ */
-
 static inline int atomic_read(const atomic_t *v)
 {
 	barrier();
@@ -101,19 +79,11 @@ static inline void atomic_set_mask(unsigned long mask, atomic_t *v)
 
 static inline int atomic_cmpxchg(atomic_t *v, int old, int new)
 {
-#if __GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ > 2)
 	asm volatile(
 		"	cs	%0,%2,%1"
 		: "+d" (old), "=Q" (v->counter)
 		: "d" (new), "Q" (v->counter)
 		: "cc", "memory");
-#else /* __GNUC__ */
-	asm volatile(
-		"	cs	%0,%3,0(%2)"
-		: "+d" (old), "=m" (v->counter)
-		: "a" (v), "d" (new), "m" (v->counter)
-		: "cc", "memory");
-#endif /* __GNUC__ */
 	return old;
 }
 
@@ -140,8 +110,6 @@ static inline int atomic_add_unless(atomic_t *v, int a, int u)
 
 #ifdef CONFIG_64BIT
 
-#if __GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ > 2)
-
 #define __CSG_LOOP(ptr, op_val, op_string) ({				\
 	long long old_val, new_val;					\
 	asm volatile(							\
@@ -157,26 +125,6 @@ static inline int atomic_add_unless(atomic_t *v, int a, int u)
 	new_val;							\
 })
 
-#else /* __GNUC__ */
-
-#define __CSG_LOOP(ptr, op_val, op_string) ({				\
-	long long old_val, new_val;					\
-	asm volatile(							\
-		"	lg	%0,0(%3)\n"				\
-		"0:	lgr	%1,%0\n"				\
-		op_string "	%1,%4\n"				\
-		"	csg	%0,%1,0(%3)\n"				\
-		"	jl	0b"					\
-		: "=&d" (old_val), "=&d" (new_val),			\
-		  "=m" (((atomic_t *)(ptr))->counter)			\
-		: "a" (ptr), "d" (op_val),				\
-		  "m" (((atomic_t *)(ptr))->counter)			\
-		: "cc", "memory");					\
-	new_val;							\
-})
-
-#endif /* __GNUC__ */
-
 static inline long long atomic64_read(const atomic64_t *v)
 {
 	barrier();
@@ -214,19 +162,11 @@ static inline void atomic64_set_mask(unsigned long mask, atomic64_t *v)
 static inline long long atomic64_cmpxchg(atomic64_t *v,
 					     long long old, long long new)
 {
-#if __GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ > 2)
 	asm volatile(
 		"	csg	%0,%2,%1"
 		: "+d" (old), "=Q" (v->counter)
 		: "d" (new), "Q" (v->counter)
 		: "cc", "memory");
-#else /* __GNUC__ */
-	asm volatile(
-		"	csg	%0,%3,0(%2)"
-		: "+d" (old), "=m" (v->counter)
-		: "a" (v), "d" (new), "m" (v->counter)
-		: "cc", "memory");
-#endif /* __GNUC__ */
 	return old;
 }
 
@@ -243,10 +183,8 @@ static inline long long atomic64_read(const atomic64_t *v)
 	register_pair rp;
 
 	asm volatile(
-		"	lm	%0,%N0,0(%1)"
-		: "=&d" (rp)
-		: "a" (&v->counter), "m" (v->counter)
-		);
+		"	lm	%0,%N0,%1"
+		: "=&d" (rp) : "Q" (v->counter)	);
 	return rp.pair;
 }
 
@@ -255,10 +193,8 @@ static inline void atomic64_set(atomic64_t *v, long long i)
 	register_pair rp = {.pair = i};
 
 	asm volatile(
-		"	stm	%1,%N1,0(%2)"
-		: "=m" (v->counter)
-		: "d" (rp), "a" (&v->counter)
-		);
+		"	stm	%1,%N1,%0"
+		: "=Q" (v->counter) : "d" (rp) );
 }
 
 static inline long long atomic64_xchg(atomic64_t *v, long long new)
@@ -267,11 +203,11 @@ static inline long long atomic64_xchg(atomic64_t *v, long long new)
 	register_pair rp_old;
 
 	asm volatile(
-		"	lm	%0,%N0,0(%2)\n"
-		"0:	cds	%0,%3,0(%2)\n"
+		"	lm	%0,%N0,%1\n"
+		"0:	cds	%0,%2,%1\n"
 		"	jl	0b\n"
-		: "=&d" (rp_old), "+m" (v->counter)
-		: "a" (&v->counter), "d" (rp_new)
+		: "=&d" (rp_old), "=Q" (v->counter)
+		: "d" (rp_new), "Q" (v->counter)
 		: "cc");
 	return rp_old.pair;
 }
@@ -283,9 +219,9 @@ static inline long long atomic64_cmpxchg(atomic64_t *v,
 	register_pair rp_new = {.pair = new};
 
 	asm volatile(
-		"	cds	%0,%3,0(%2)"
-		: "+&d" (rp_old), "+m" (v->counter)
-		: "a" (&v->counter), "d" (rp_new)
+		"	cds	%0,%2,%1"
+		: "+&d" (rp_old), "=Q" (v->counter)
+		: "d" (rp_new), "Q" (v->counter)
 		: "cc");
 	return rp_old.pair;
 }
diff --git a/arch/s390/include/asm/bitops.h b/arch/s390/include/asm/bitops.h
index b30606f..2e05972 100644
--- a/arch/s390/include/asm/bitops.h
+++ b/arch/s390/include/asm/bitops.h
@@ -71,8 +71,6 @@ extern const char _sb_findmap[];
 #define __BITOPS_AND		"nr"
 #define __BITOPS_XOR		"xr"
 
-#if __GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ > 2)
-
 #define __BITOPS_LOOP(__old, __new, __addr, __val, __op_string)	\
 	asm volatile(						\
 		"	l	%0,%2\n"			\
@@ -85,22 +83,6 @@ extern const char _sb_findmap[];
 		: "d" (__val), "Q" (*(unsigned long *) __addr)	\
 		: "cc");
 
-#else /* __GNUC__ */
-
-#define __BITOPS_LOOP(__old, __new, __addr, __val, __op_string)	\
-	asm volatile(						\
-		"	l	%0,0(%4)\n"			\
-		"0:	lr	%1,%0\n"			\
-		__op_string "	%1,%3\n"			\
-		"	cs	%0,%1,0(%4)\n"			\
-		"	jl	0b"				\
-		: "=&d" (__old), "=&d" (__new),			\
-		  "=m" (*(unsigned long *) __addr)		\
-		: "d" (__val), "a" (__addr),			\
-		  "m" (*(unsigned long *) __addr) : "cc");
-
-#endif /* __GNUC__ */
-
 #else /* __s390x__ */
 
 #define __BITOPS_ALIGN		7
@@ -109,8 +91,6 @@ extern const char _sb_findmap[];
 #define __BITOPS_AND		"ngr"
 #define __BITOPS_XOR		"xgr"
 
-#if __GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ > 2)
-
 #define __BITOPS_LOOP(__old, __new, __addr, __val, __op_string)	\
 	asm volatile(						\
 		"	lg	%0,%2\n"			\
@@ -123,23 +103,6 @@ extern const char _sb_findmap[];
 		: "d" (__val), "Q" (*(unsigned long *) __addr)	\
 		: "cc");
 
-#else /* __GNUC__ */
-
-#define __BITOPS_LOOP(__old, __new, __addr, __val, __op_string)	\
-	asm volatile(						\
-		"	lg	%0,0(%4)\n"			\
-		"0:	lgr	%1,%0\n"			\
-		__op_string "	%1,%3\n"			\
-		"	csg	%0,%1,0(%4)\n"			\
-		"	jl	0b"				\
-		: "=&d" (__old), "=&d" (__new),			\
-		  "=m" (*(unsigned long *) __addr)		\
-		: "d" (__val), "a" (__addr),			\
-		  "m" (*(unsigned long *) __addr) : "cc");
-
-
-#endif /* __GNUC__ */
-
 #endif /* __s390x__ */
 
 #define __BITOPS_WORDS(bits) (((bits)+__BITOPS_WORDSIZE-1)/__BITOPS_WORDSIZE)
@@ -261,9 +224,8 @@ static inline void __set_bit(unsigned long nr, volatile unsigned long *ptr)
 
 	addr = (unsigned long) ptr + ((nr ^ (__BITOPS_WORDSIZE - 8)) >> 3);
 	asm volatile(
-		"	oc	0(1,%1),0(%2)"
-		: "=m" (*(char *) addr) : "a" (addr),
-		  "a" (_oi_bitmap + (nr & 7)), "m" (*(char *) addr) : "cc" );
+		"	oc	%O0(1,%R0),%1"
+		: "=Q" (*(char *) addr) : "Q" (_oi_bitmap[nr & 7]) : "cc" );
 }
 
 static inline void 
@@ -290,9 +252,8 @@ __clear_bit(unsigned long nr, volatile unsigned long *ptr)
 
 	addr = (unsigned long) ptr + ((nr ^ (__BITOPS_WORDSIZE - 8)) >> 3);
 	asm volatile(
-		"	nc	0(1,%1),0(%2)"
-		: "=m" (*(char *) addr)	: "a" (addr),
-		  "a" (_ni_bitmap + (nr & 7)), "m" (*(char *) addr) : "cc");
+		"	nc	%O0(1,%R0),%1"
+		: "=Q" (*(char *) addr) : "Q" (_ni_bitmap[nr & 7]) : "cc" );
 }
 
 static inline void 
@@ -318,9 +279,8 @@ static inline void __change_bit(unsigned long nr, volatile unsigned long *ptr)
 
 	addr = (unsigned long) ptr + ((nr ^ (__BITOPS_WORDSIZE - 8)) >> 3);
 	asm volatile(
-		"	xc	0(1,%1),0(%2)"
-		:  "=m" (*(char *) addr) : "a" (addr),
-		   "a" (_oi_bitmap + (nr & 7)), "m" (*(char *) addr) : "cc" );
+		"	xc	%O0(1,%R0),%1"
+		: "=Q" (*(char *) addr) : "Q" (_oi_bitmap[nr & 7]) : "cc" );
 }
 
 static inline void 
@@ -349,10 +309,9 @@ test_and_set_bit_simple(unsigned long nr, volatile unsigned long *ptr)
 	addr = (unsigned long) ptr + ((nr ^ (__BITOPS_WORDSIZE - 8)) >> 3);
 	ch = *(unsigned char *) addr;
 	asm volatile(
-		"	oc	0(1,%1),0(%2)"
-		: "=m" (*(char *) addr)
-		: "a" (addr), "a" (_oi_bitmap + (nr & 7)),
-		  "m" (*(char *) addr) : "cc", "memory");
+		"	oc	%O0(1,%R0),%1"
+		: "=Q" (*(char *) addr)	: "Q" (_oi_bitmap[nr & 7])
+		: "cc", "memory");
 	return (ch >> (nr & 7)) & 1;
 }
 #define __test_and_set_bit(X,Y)		test_and_set_bit_simple(X,Y)
@@ -369,10 +328,9 @@ test_and_clear_bit_simple(unsigned long nr, volatile unsigned long *ptr)
 	addr = (unsigned long) ptr + ((nr ^ (__BITOPS_WORDSIZE - 8)) >> 3);
 	ch = *(unsigned char *) addr;
 	asm volatile(
-		"	nc	0(1,%1),0(%2)"
-		: "=m" (*(char *) addr)
-		: "a" (addr), "a" (_ni_bitmap + (nr & 7)),
-		  "m" (*(char *) addr) : "cc", "memory");
+		"	nc	%O0(1,%R0),%1"
+		: "=Q" (*(char *) addr)	: "Q" (_ni_bitmap[nr & 7])
+		: "cc", "memory");
 	return (ch >> (nr & 7)) & 1;
 }
 #define __test_and_clear_bit(X,Y)	test_and_clear_bit_simple(X,Y)
@@ -389,10 +347,9 @@ test_and_change_bit_simple(unsigned long nr, volatile unsigned long *ptr)
 	addr = (unsigned long) ptr + ((nr ^ (__BITOPS_WORDSIZE - 8)) >> 3);
 	ch = *(unsigned char *) addr;
 	asm volatile(
-		"	xc	0(1,%1),0(%2)"
-		: "=m" (*(char *) addr)
-		: "a" (addr), "a" (_oi_bitmap + (nr & 7)),
-		  "m" (*(char *) addr) : "cc", "memory");
+		"	xc	%O0(1,%R0),%1"
+		: "=Q" (*(char *) addr)	: "Q" (_oi_bitmap[nr & 7])
+		: "cc", "memory");
 	return (ch >> (nr & 7)) & 1;
 }
 #define __test_and_change_bit(X,Y)	test_and_change_bit_simple(X,Y)
@@ -591,11 +548,11 @@ static inline unsigned long __load_ulong_le(const unsigned long *p,
 	p = (unsigned long *)((unsigned long) p + offset);
 #ifndef __s390x__
 	asm volatile(
-		"	ic	%0,0(%1)\n"
-		"	icm	%0,2,1(%1)\n"
-		"	icm	%0,4,2(%1)\n"
-		"	icm	%0,8,3(%1)"
-		: "=&d" (word) : "a" (p), "m" (*p) : "cc");
+		"	ic	%0,%O1(%R1)\n"
+		"	icm	%0,2,%O1+1(%R1)\n"
+		"	icm	%0,4,%O1+2(%R1)\n"
+		"	icm	%0,8,%O1+3(%R1)"
+		: "=&d" (word) : "Q" (*p) : "cc");
 #else
 	asm volatile(
 		"	lrvg	%0,%1"
diff --git a/arch/s390/include/asm/etr.h b/arch/s390/include/asm/etr.h
index 80ef58c..538e1b3 100644
--- a/arch/s390/include/asm/etr.h
+++ b/arch/s390/include/asm/etr.h
@@ -145,11 +145,11 @@ static inline int etr_setr(struct etr_eacr *ctrl)
 	int rc = -ENOSYS;
 
 	asm volatile(
-		"	.insn	s,0xb2160000,0(%2)\n"
+		"	.insn	s,0xb2160000,%1\n"
 		"0:	la	%0,0\n"
 		"1:\n"
 		EX_TABLE(0b,1b)
-		: "+d" (rc) : "m" (*ctrl), "a" (ctrl));
+		: "+d" (rc) : "Q" (*ctrl));
 	return rc;
 }
 
@@ -159,11 +159,11 @@ static inline int etr_stetr(struct etr_aib *aib)
 	int rc = -ENOSYS;
 
 	asm volatile(
-		"	.insn	s,0xb2170000,0(%2)\n"
+		"	.insn	s,0xb2170000,%1\n"
 		"0:	la	%0,0\n"
 		"1:\n"
 		EX_TABLE(0b,1b)
-		: "+d" (rc) : "m" (*aib), "a" (aib));
+		: "+d" (rc) : "Q" (*aib));
 	return rc;
 }
 
@@ -174,11 +174,11 @@ static inline int etr_steai(struct etr_aib *aib, unsigned int func)
 	int rc = -ENOSYS;
 
 	asm volatile(
-		"	.insn	s,0xb2b30000,0(%2)\n"
+		"	.insn	s,0xb2b30000,%1\n"
 		"0:	la	%0,0\n"
 		"1:\n"
 		EX_TABLE(0b,1b)
-		: "+d" (rc) : "m" (*aib), "a" (aib), "d" (reg0));
+		: "+d" (rc) : "Q" (*aib), "d" (reg0));
 	return rc;
 }
 
diff --git a/arch/s390/include/asm/irqflags.h b/arch/s390/include/asm/irqflags.h
index c2fb432..15b3ac2 100644
--- a/arch/s390/include/asm/irqflags.h
+++ b/arch/s390/include/asm/irqflags.h
@@ -8,8 +8,6 @@
 
 #include <linux/types.h>
 
-#if __GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ > 2)
-
 /* store then or system mask. */
 #define __raw_local_irq_stosm(__or)					\
 ({									\
@@ -36,40 +34,6 @@
 	asm volatile("ssm   %0" : : "Q" (__mask) : "memory");		\
 })
 
-#else /* __GNUC__ */
-
-/* store then or system mask. */
-#define __raw_local_irq_stosm(__or)					\
-({									\
-	unsigned long __mask;						\
-	asm volatile(							\
-		"	stosm	0(%1),%2"				\
-		: "=m" (__mask)						\
-		: "a" (&__mask), "i" (__or) : "memory");		\
-	__mask;								\
-})
-
-/* store then and system mask. */
-#define __raw_local_irq_stnsm(__and)					\
-({									\
-	unsigned long __mask;						\
-	asm volatile(							\
-		"	stnsm	0(%1),%2"				\
-		: "=m" (__mask)						\
-		: "a" (&__mask), "i" (__and) : "memory");		\
-	__mask;								\
-})
-
-/* set system mask. */
-#define __raw_local_irq_ssm(__mask)					\
-({									\
-	asm volatile(							\
-		"	ssm	0(%0)"					\
-		: : "a" (&__mask), "m" (__mask) : "memory");		\
-})
-
-#endif /* __GNUC__ */
-
 /* interrupt control.. */
 static inline unsigned long raw_local_irq_enable(void)
 {
diff --git a/arch/s390/include/asm/processor.h b/arch/s390/include/asm/processor.h
index b427154..73e2598 100644
--- a/arch/s390/include/asm/processor.h
+++ b/arch/s390/include/asm/processor.h
@@ -28,7 +28,7 @@
 
 static inline void get_cpu_id(struct cpuid *ptr)
 {
-	asm volatile("stidp 0(%1)" : "=m" (*ptr) : "a" (ptr));
+	asm volatile("stidp %0" : "=Q" (*ptr));
 }
 
 extern void s390_adjust_jiffies(void);
@@ -184,9 +184,9 @@ static inline void psw_set_key(unsigned int key)
 static inline void __load_psw(psw_t psw)
 {
 #ifndef __s390x__
-	asm volatile("lpsw  0(%0)" : : "a" (&psw), "m" (psw) : "cc");
+	asm volatile("lpsw  %0" : : "Q" (psw) : "cc");
 #else
-	asm volatile("lpswe 0(%0)" : : "a" (&psw), "m" (psw) : "cc");
+	asm volatile("lpswe %0" : : "Q" (psw) : "cc");
 #endif
 }
 
@@ -206,17 +206,17 @@ static inline void __load_psw_mask (unsigned long mask)
 	asm volatile(
 		"	basr	%0,0\n"
 		"0:	ahi	%0,1f-0b\n"
-		"	st	%0,4(%1)\n"
-		"	lpsw	0(%1)\n"
+		"	st	%0,%O1+4(%R1)\n"
+		"	lpsw	%1\n"
 		"1:"
-		: "=&d" (addr) : "a" (&psw), "m" (psw) : "memory", "cc");
+		: "=&d" (addr), "=Q" (psw) : "Q" (psw) : "memory", "cc");
 #else /* __s390x__ */
 	asm volatile(
 		"	larl	%0,1f\n"
-		"	stg	%0,8(%1)\n"
-		"	lpswe	0(%1)\n"
+		"	stg	%0,%O1+8(%R1)\n"
+		"	lpswe	%1\n"
 		"1:"
-		: "=&d" (addr) : "a" (&psw), "m" (psw) : "memory", "cc");
+		: "=&d" (addr), "=Q" (psw) : "Q" (psw) : "memory", "cc");
 #endif /* __s390x__ */
 }
  
diff --git a/arch/s390/include/asm/rwsem.h b/arch/s390/include/asm/rwsem.h
index 9d2a1797..423fdda 100644
--- a/arch/s390/include/asm/rwsem.h
+++ b/arch/s390/include/asm/rwsem.h
@@ -124,21 +124,21 @@ static inline void __down_read(struct rw_semaphore *sem)
 
 	asm volatile(
 #ifndef __s390x__
-		"	l	%0,0(%3)\n"
+		"	l	%0,%2\n"
 		"0:	lr	%1,%0\n"
-		"	ahi	%1,%5\n"
-		"	cs	%0,%1,0(%3)\n"
+		"	ahi	%1,%4\n"
+		"	cs	%0,%1,%2\n"
 		"	jl	0b"
 #else /* __s390x__ */
-		"	lg	%0,0(%3)\n"
+		"	lg	%0,%2\n"
 		"0:	lgr	%1,%0\n"
-		"	aghi	%1,%5\n"
-		"	csg	%0,%1,0(%3)\n"
+		"	aghi	%1,%4\n"
+		"	csg	%0,%1,%2\n"
 		"	jl	0b"
 #endif /* __s390x__ */
-		: "=&d" (old), "=&d" (new), "=m" (sem->count)
-		: "a" (&sem->count), "m" (sem->count),
-		  "i" (RWSEM_ACTIVE_READ_BIAS) : "cc", "memory");
+		: "=&d" (old), "=&d" (new), "=Q" (sem->count)
+		: "Q" (sem->count), "i" (RWSEM_ACTIVE_READ_BIAS)
+		: "cc", "memory");
 	if (old < 0)
 		rwsem_down_read_failed(sem);
 }
@@ -152,25 +152,25 @@ static inline int __down_read_trylock(struct rw_semaphore *sem)
 
 	asm volatile(
 #ifndef __s390x__
-		"	l	%0,0(%3)\n"
+		"	l	%0,%2\n"
 		"0:	ltr	%1,%0\n"
 		"	jm	1f\n"
-		"	ahi	%1,%5\n"
-		"	cs	%0,%1,0(%3)\n"
+		"	ahi	%1,%4\n"
+		"	cs	%0,%1,%2\n"
 		"	jl	0b\n"
 		"1:"
 #else /* __s390x__ */
-		"	lg	%0,0(%3)\n"
+		"	lg	%0,%2\n"
 		"0:	ltgr	%1,%0\n"
 		"	jm	1f\n"
-		"	aghi	%1,%5\n"
-		"	csg	%0,%1,0(%3)\n"
+		"	aghi	%1,%4\n"
+		"	csg	%0,%1,%2\n"
 		"	jl	0b\n"
 		"1:"
 #endif /* __s390x__ */
-		: "=&d" (old), "=&d" (new), "=m" (sem->count)
-		: "a" (&sem->count), "m" (sem->count),
-		  "i" (RWSEM_ACTIVE_READ_BIAS) : "cc", "memory");
+		: "=&d" (old), "=&d" (new), "=Q" (sem->count)
+		: "Q" (sem->count), "i" (RWSEM_ACTIVE_READ_BIAS)
+		: "cc", "memory");
 	return old >= 0 ? 1 : 0;
 }
 
@@ -184,20 +184,20 @@ static inline void __down_write_nested(struct rw_semaphore *sem, int subclass)
 	tmp = RWSEM_ACTIVE_WRITE_BIAS;
 	asm volatile(
 #ifndef __s390x__
-		"	l	%0,0(%3)\n"
+		"	l	%0,%2\n"
 		"0:	lr	%1,%0\n"
-		"	a	%1,%5\n"
-		"	cs	%0,%1,0(%3)\n"
+		"	a	%1,%4\n"
+		"	cs	%0,%1,%2\n"
 		"	jl	0b"
 #else /* __s390x__ */
-		"	lg	%0,0(%3)\n"
+		"	lg	%0,%2\n"
 		"0:	lgr	%1,%0\n"
-		"	ag	%1,%5\n"
-		"	csg	%0,%1,0(%3)\n"
+		"	ag	%1,%4\n"
+		"	csg	%0,%1,%2\n"
 		"	jl	0b"
 #endif /* __s390x__ */
-		: "=&d" (old), "=&d" (new), "=m" (sem->count)
-		: "a" (&sem->count), "m" (sem->count), "m" (tmp)
+		: "=&d" (old), "=&d" (new), "=Q" (sem->count)
+		: "Q" (sem->count), "m" (tmp)
 		: "cc", "memory");
 	if (old != 0)
 		rwsem_down_write_failed(sem);
@@ -217,22 +217,22 @@ static inline int __down_write_trylock(struct rw_semaphore *sem)
 
 	asm volatile(
 #ifndef __s390x__
-		"	l	%0,0(%2)\n"
+		"	l	%0,%1\n"
 		"0:	ltr	%0,%0\n"
 		"	jnz	1f\n"
-		"	cs	%0,%4,0(%2)\n"
+		"	cs	%0,%3,%1\n"
 		"	jl	0b\n"
 #else /* __s390x__ */
-		"	lg	%0,0(%2)\n"
+		"	lg	%0,%1\n"
 		"0:	ltgr	%0,%0\n"
 		"	jnz	1f\n"
-		"	csg	%0,%4,0(%2)\n"
+		"	csg	%0,%3,%1\n"
 		"	jl	0b\n"
 #endif /* __s390x__ */
 		"1:"
-		: "=&d" (old), "=m" (sem->count)
-		: "a" (&sem->count), "m" (sem->count),
-		  "d" (RWSEM_ACTIVE_WRITE_BIAS) : "cc", "memory");
+		: "=&d" (old), "=Q" (sem->count)
+		: "Q" (sem->count), "d" (RWSEM_ACTIVE_WRITE_BIAS)
+		: "cc", "memory");
 	return (old == RWSEM_UNLOCKED_VALUE) ? 1 : 0;
 }
 
@@ -245,21 +245,20 @@ static inline void __up_read(struct rw_semaphore *sem)
 
 	asm volatile(
 #ifndef __s390x__
-		"	l	%0,0(%3)\n"
+		"	l	%0,%2\n"
 		"0:	lr	%1,%0\n"
-		"	ahi	%1,%5\n"
-		"	cs	%0,%1,0(%3)\n"
+		"	ahi	%1,%4\n"
+		"	cs	%0,%1,%2\n"
 		"	jl	0b"
 #else /* __s390x__ */
-		"	lg	%0,0(%3)\n"
+		"	lg	%0,%2\n"
 		"0:	lgr	%1,%0\n"
-		"	aghi	%1,%5\n"
-		"	csg	%0,%1,0(%3)\n"
+		"	aghi	%1,%4\n"
+		"	csg	%0,%1,%2\n"
 		"	jl	0b"
 #endif /* __s390x__ */
-		: "=&d" (old), "=&d" (new), "=m" (sem->count)
-		: "a" (&sem->count), "m" (sem->count),
-		  "i" (-RWSEM_ACTIVE_READ_BIAS)
+		: "=&d" (old), "=&d" (new), "=Q" (sem->count)
+		: "Q" (sem->count), "i" (-RWSEM_ACTIVE_READ_BIAS)
 		: "cc", "memory");
 	if (new < 0)
 		if ((new & RWSEM_ACTIVE_MASK) == 0)
@@ -276,20 +275,20 @@ static inline void __up_write(struct rw_semaphore *sem)
 	tmp = -RWSEM_ACTIVE_WRITE_BIAS;
 	asm volatile(
 #ifndef __s390x__
-		"	l	%0,0(%3)\n"
+		"	l	%0,%2\n"
 		"0:	lr	%1,%0\n"
-		"	a	%1,%5\n"
-		"	cs	%0,%1,0(%3)\n"
+		"	a	%1,%4\n"
+		"	cs	%0,%1,%2\n"
 		"	jl	0b"
 #else /* __s390x__ */
-		"	lg	%0,0(%3)\n"
+		"	lg	%0,%2\n"
 		"0:	lgr	%1,%0\n"
-		"	ag	%1,%5\n"
-		"	csg	%0,%1,0(%3)\n"
+		"	ag	%1,%4\n"
+		"	csg	%0,%1,%2\n"
 		"	jl	0b"
 #endif /* __s390x__ */
-		: "=&d" (old), "=&d" (new), "=m" (sem->count)
-		: "a" (&sem->count), "m" (sem->count), "m" (tmp)
+		: "=&d" (old), "=&d" (new), "=Q" (sem->count)
+		: "Q" (sem->count), "m" (tmp)
 		: "cc", "memory");
 	if (new < 0)
 		if ((new & RWSEM_ACTIVE_MASK) == 0)
@@ -306,20 +305,20 @@ static inline void __downgrade_write(struct rw_semaphore *sem)
 	tmp = -RWSEM_WAITING_BIAS;
 	asm volatile(
 #ifndef __s390x__
-		"	l	%0,0(%3)\n"
+		"	l	%0,%2\n"
 		"0:	lr	%1,%0\n"
-		"	a	%1,%5\n"
-		"	cs	%0,%1,0(%3)\n"
+		"	a	%1,%4\n"
+		"	cs	%0,%1,%2\n"
 		"	jl	0b"
 #else /* __s390x__ */
-		"	lg	%0,0(%3)\n"
+		"	lg	%0,%2\n"
 		"0:	lgr	%1,%0\n"
-		"	ag	%1,%5\n"
-		"	csg	%0,%1,0(%3)\n"
+		"	ag	%1,%4\n"
+		"	csg	%0,%1,%2\n"
 		"	jl	0b"
 #endif /* __s390x__ */
-		: "=&d" (old), "=&d" (new), "=m" (sem->count)
-		: "a" (&sem->count), "m" (sem->count), "m" (tmp)
+		: "=&d" (old), "=&d" (new), "=Q" (sem->count)
+		: "Q" (sem->count), "m" (tmp)
 		: "cc", "memory");
 	if (new > 1)
 		rwsem_downgrade_wake(sem);
@@ -334,20 +333,20 @@ static inline void rwsem_atomic_add(long delta, struct rw_semaphore *sem)
 
 	asm volatile(
 #ifndef __s390x__
-		"	l	%0,0(%3)\n"
+		"	l	%0,%2\n"
 		"0:	lr	%1,%0\n"
-		"	ar	%1,%5\n"
-		"	cs	%0,%1,0(%3)\n"
+		"	ar	%1,%4\n"
+		"	cs	%0,%1,%2\n"
 		"	jl	0b"
 #else /* __s390x__ */
-		"	lg	%0,0(%3)\n"
+		"	lg	%0,%2\n"
 		"0:	lgr	%1,%0\n"
-		"	agr	%1,%5\n"
-		"	csg	%0,%1,0(%3)\n"
+		"	agr	%1,%4\n"
+		"	csg	%0,%1,%2\n"
 		"	jl	0b"
 #endif /* __s390x__ */
-		: "=&d" (old), "=&d" (new), "=m" (sem->count)
-		: "a" (&sem->count), "m" (sem->count), "d" (delta)
+		: "=&d" (old), "=&d" (new), "=Q" (sem->count)
+		: "Q" (sem->count), "d" (delta)
 		: "cc", "memory");
 }
 
@@ -360,20 +359,20 @@ static inline long rwsem_atomic_update(long delta, struct rw_semaphore *sem)
 
 	asm volatile(
 #ifndef __s390x__
-		"	l	%0,0(%3)\n"
+		"	l	%0,%2\n"
 		"0:	lr	%1,%0\n"
-		"	ar	%1,%5\n"
-		"	cs	%0,%1,0(%3)\n"
+		"	ar	%1,%4\n"
+		"	cs	%0,%1,%2\n"
 		"	jl	0b"
 #else /* __s390x__ */
-		"	lg	%0,0(%3)\n"
+		"	lg	%0,%2\n"
 		"0:	lgr	%1,%0\n"
-		"	agr	%1,%5\n"
-		"	csg	%0,%1,0(%3)\n"
+		"	agr	%1,%4\n"
+		"	csg	%0,%1,%2\n"
 		"	jl	0b"
 #endif /* __s390x__ */
-		: "=&d" (old), "=&d" (new), "=m" (sem->count)
-		: "a" (&sem->count), "m" (sem->count), "d" (delta)
+		: "=&d" (old), "=&d" (new), "=Q" (sem->count)
+		: "Q" (sem->count), "d" (delta)
 		: "cc", "memory");
 	return new;
 }
diff --git a/arch/s390/include/asm/spinlock.h b/arch/s390/include/asm/spinlock.h
index a587907..56612fc 100644
--- a/arch/s390/include/asm/spinlock.h
+++ b/arch/s390/include/asm/spinlock.h
@@ -13,8 +13,6 @@
 
 #include <linux/smp.h>
 
-#if __GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ > 2)
-
 static inline int
 _raw_compare_and_swap(volatile unsigned int *lock,
 		      unsigned int old, unsigned int new)
@@ -27,22 +25,6 @@ _raw_compare_and_swap(volatile unsigned int *lock,
 	return old;
 }
 
-#else /* __GNUC__ */
-
-static inline int
-_raw_compare_and_swap(volatile unsigned int *lock,
-		      unsigned int old, unsigned int new)
-{
-	asm volatile(
-		"	cs	%0,%3,0(%4)"
-		: "=d" (old), "=m" (*lock)
-		: "0" (old), "d" (new), "a" (lock), "m" (*lock)
-		: "cc", "memory" );
-	return old;
-}
-
-#endif /* __GNUC__ */
-
 /*
  * Simple spin lock operations.  There are two variants, one clears IRQ's
  * on the local processor, one does not.
diff --git a/arch/s390/include/asm/swab.h b/arch/s390/include/asm/swab.h
index eb18dc1..6bdee21 100644
--- a/arch/s390/include/asm/swab.h
+++ b/arch/s390/include/asm/swab.h
@@ -47,11 +47,11 @@ static inline __u32 __arch_swab32p(const __u32 *x)
 	
 	asm volatile(
 #ifndef __s390x__
-		"	icm	%0,8,3(%1)\n"
-		"	icm	%0,4,2(%1)\n"
-		"	icm	%0,2,1(%1)\n"
-		"	ic	%0,0(%1)"
-		: "=&d" (result) : "a" (x), "m" (*x) : "cc");
+		"	icm	%0,8,%O1+3(%R1)\n"
+		"	icm	%0,4,%O1+2(%R1)\n"
+		"	icm	%0,2,%O1+1(%R1)\n"
+		"	ic	%0,%1"
+		: "=&d" (result) : "Q" (*x) : "cc");
 #else /* __s390x__ */
 		"	lrv	%0,%1"
 		: "=d" (result) : "m" (*x));
@@ -77,9 +77,9 @@ static inline __u16 __arch_swab16p(const __u16 *x)
 	
 	asm volatile(
 #ifndef __s390x__
-		"	icm	%0,2,1(%1)\n"
-		"	ic	%0,0(%1)\n"
-		: "=&d" (result) : "a" (x), "m" (*x) : "cc");
+		"	icm	%0,2,%O+1(%R1)\n"
+		"	ic	%0,%1\n"
+		: "=&d" (result) : "Q" (*x) : "cc");
 #else /* __s390x__ */
 		"	lrvh	%0,%1"
 		: "=d" (result) : "m" (*x));
diff --git a/arch/s390/include/asm/system.h b/arch/s390/include/asm/system.h
index 379661d..67ee6c3 100644
--- a/arch/s390/include/asm/system.h
+++ b/arch/s390/include/asm/system.h
@@ -24,65 +24,65 @@ extern struct task_struct *__switch_to(void *, void *);
 static inline void save_fp_regs(s390_fp_regs *fpregs)
 {
 	asm volatile(
-		"	std	0,8(%1)\n"
-		"	std	2,24(%1)\n"
-		"	std	4,40(%1)\n"
-		"	std	6,56(%1)"
-		: "=m" (*fpregs) : "a" (fpregs), "m" (*fpregs) : "memory");
+		"	std	0,%O0+8(%R0)\n"
+		"	std	2,%O0+24(%R0)\n"
+		"	std	4,%O0+40(%R0)\n"
+		"	std	6,%O0+56(%R0)"
+		: "=Q" (*fpregs) : "Q" (*fpregs));
 	if (!MACHINE_HAS_IEEE)
 		return;
 	asm volatile(
-		"	stfpc	0(%1)\n"
-		"	std	1,16(%1)\n"
-		"	std	3,32(%1)\n"
-		"	std	5,48(%1)\n"
-		"	std	7,64(%1)\n"
-		"	std	8,72(%1)\n"
-		"	std	9,80(%1)\n"
-		"	std	10,88(%1)\n"
-		"	std	11,96(%1)\n"
-		"	std	12,104(%1)\n"
-		"	std	13,112(%1)\n"
-		"	std	14,120(%1)\n"
-		"	std	15,128(%1)\n"
-		: "=m" (*fpregs) : "a" (fpregs), "m" (*fpregs) : "memory");
+		"	stfpc	%0\n"
+		"	std	1,%O0+16(%R0)\n"
+		"	std	3,%O0+32(%R0)\n"
+		"	std	5,%O0+48(%R0)\n"
+		"	std	7,%O0+64(%R0)\n"
+		"	std	8,%O0+72(%R0)\n"
+		"	std	9,%O0+80(%R0)\n"
+		"	std	10,%O0+88(%R0)\n"
+		"	std	11,%O0+96(%R0)\n"
+		"	std	12,%O0+104(%R0)\n"
+		"	std	13,%O0+112(%R0)\n"
+		"	std	14,%O0+120(%R0)\n"
+		"	std	15,%O0+128(%R0)\n"
+		: "=Q" (*fpregs) : "Q" (*fpregs));
 }
 
 static inline void restore_fp_regs(s390_fp_regs *fpregs)
 {
 	asm volatile(
-		"	ld	0,8(%0)\n"
-		"	ld	2,24(%0)\n"
-		"	ld	4,40(%0)\n"
-		"	ld	6,56(%0)"
-		: : "a" (fpregs), "m" (*fpregs));
+		"	ld	0,%O0+8(%R0)\n"
+		"	ld	2,%O0+24(%R0)\n"
+		"	ld	4,%O0+40(%R0)\n"
+		"	ld	6,%O0+56(%R0)"
+		: : "Q" (*fpregs));
 	if (!MACHINE_HAS_IEEE)
 		return;
 	asm volatile(
-		"	lfpc	0(%0)\n"
-		"	ld	1,16(%0)\n"
-		"	ld	3,32(%0)\n"
-		"	ld	5,48(%0)\n"
-		"	ld	7,64(%0)\n"
-		"	ld	8,72(%0)\n"
-		"	ld	9,80(%0)\n"
-		"	ld	10,88(%0)\n"
-		"	ld	11,96(%0)\n"
-		"	ld	12,104(%0)\n"
-		"	ld	13,112(%0)\n"
-		"	ld	14,120(%0)\n"
-		"	ld	15,128(%0)\n"
-		: : "a" (fpregs), "m" (*fpregs));
+		"	lfpc	%0\n"
+		"	ld	1,%O0+16(%R0)\n"
+		"	ld	3,%O0+32(%R0)\n"
+		"	ld	5,%O0+48(%R0)\n"
+		"	ld	7,%O0+64(%R0)\n"
+		"	ld	8,%O0+72(%R0)\n"
+		"	ld	9,%O0+80(%R0)\n"
+		"	ld	10,%O0+88(%R0)\n"
+		"	ld	11,%O0+96(%R0)\n"
+		"	ld	12,%O0+104(%R0)\n"
+		"	ld	13,%O0+112(%R0)\n"
+		"	ld	14,%O0+120(%R0)\n"
+		"	ld	15,%O0+128(%R0)\n"
+		: : "Q" (*fpregs));
 }
 
 static inline void save_access_regs(unsigned int *acrs)
 {
-	asm volatile("stam 0,15,0(%0)" : : "a" (acrs) : "memory");
+	asm volatile("stam 0,15,%0" : "=Q" (*acrs));
 }
 
 static inline void restore_access_regs(unsigned int *acrs)
 {
-	asm volatile("lam 0,15,0(%0)" : : "a" (acrs));
+	asm volatile("lam 0,15,%0" : : "Q" (*acrs));
 }
 
 #define switch_to(prev,next,last) do {					     \
@@ -139,48 +139,48 @@ static inline unsigned long __xchg(unsigned long x, void * ptr, int size)
 		shift = (3 ^ (addr & 3)) << 3;
 		addr ^= addr & 3;
 		asm volatile(
-			"	l	%0,0(%4)\n"
+			"	l	%0,%4\n"
 			"0:	lr	0,%0\n"
 			"	nr	0,%3\n"
 			"	or	0,%2\n"
-			"	cs	%0,0,0(%4)\n"
+			"	cs	%0,0,%4\n"
 			"	jl	0b\n"
-			: "=&d" (old), "=m" (*(int *) addr)
-			: "d" (x << shift), "d" (~(255 << shift)), "a" (addr),
-			  "m" (*(int *) addr) : "memory", "cc", "0");
+			: "=&d" (old), "=Q" (*(int *) addr)
+			: "d" (x << shift), "d" (~(255 << shift)),
+			  "Q" (*(int *) addr) : "memory", "cc", "0");
 		return old >> shift;
 	case 2:
 		addr = (unsigned long) ptr;
 		shift = (2 ^ (addr & 2)) << 3;
 		addr ^= addr & 2;
 		asm volatile(
-			"	l	%0,0(%4)\n"
+			"	l	%0,%4\n"
 			"0:	lr	0,%0\n"
 			"	nr	0,%3\n"
 			"	or	0,%2\n"
-			"	cs	%0,0,0(%4)\n"
+			"	cs	%0,0,%4\n"
 			"	jl	0b\n"
-			: "=&d" (old), "=m" (*(int *) addr)
-			: "d" (x << shift), "d" (~(65535 << shift)), "a" (addr),
-			  "m" (*(int *) addr) : "memory", "cc", "0");
+			: "=&d" (old), "=Q" (*(int *) addr)
+			: "d" (x << shift), "d" (~(65535 << shift)),
+			  "Q" (*(int *) addr) : "memory", "cc", "0");
 		return old >> shift;
 	case 4:
 		asm volatile(
-			"	l	%0,0(%3)\n"
-			"0:	cs	%0,%2,0(%3)\n"
+			"	l	%0,%3\n"
+			"0:	cs	%0,%2,%3\n"
 			"	jl	0b\n"
-			: "=&d" (old), "=m" (*(int *) ptr)
-			: "d" (x), "a" (ptr), "m" (*(int *) ptr)
+			: "=&d" (old), "=Q" (*(int *) ptr)
+			: "d" (x), "Q" (*(int *) ptr)
 			: "memory", "cc");
 		return old;
 #ifdef __s390x__
 	case 8:
 		asm volatile(
-			"	lg	%0,0(%3)\n"
-			"0:	csg	%0,%2,0(%3)\n"
+			"	lg	%0,%3\n"
+			"0:	csg	%0,%2,%3\n"
 			"	jl	0b\n"
 			: "=&d" (old), "=m" (*(long *) ptr)
-			: "d" (x), "a" (ptr), "m" (*(long *) ptr)
+			: "d" (x), "Q" (*(long *) ptr)
 			: "memory", "cc");
 		return old;
 #endif /* __s390x__ */
@@ -215,20 +215,20 @@ __cmpxchg(volatile void *ptr, unsigned long old, unsigned long new, int size)
 		shift = (3 ^ (addr & 3)) << 3;
 		addr ^= addr & 3;
 		asm volatile(
-			"	l	%0,0(%4)\n"
+			"	l	%0,%2\n"
 			"0:	nr	%0,%5\n"
 			"	lr	%1,%0\n"
 			"	or	%0,%2\n"
 			"	or	%1,%3\n"
-			"	cs	%0,%1,0(%4)\n"
+			"	cs	%0,%1,%2\n"
 			"	jnl	1f\n"
 			"	xr	%1,%0\n"
 			"	nr	%1,%5\n"
 			"	jnz	0b\n"
 			"1:"
-			: "=&d" (prev), "=&d" (tmp)
-			: "d" (old << shift), "d" (new << shift), "a" (ptr),
-			  "d" (~(255 << shift))
+			: "=&d" (prev), "=&d" (tmp), "=Q" (*(int *) ptr)
+			: "d" (old << shift), "d" (new << shift),
+			  "d" (~(255 << shift)), "Q" (*(int *) ptr)
 			: "memory", "cc");
 		return prev >> shift;
 	case 2:
@@ -236,33 +236,35 @@ __cmpxchg(volatile void *ptr, unsigned long old, unsigned long new, int size)
 		shift = (2 ^ (addr & 2)) << 3;
 		addr ^= addr & 2;
 		asm volatile(
-			"	l	%0,0(%4)\n"
+			"	l	%0,%2\n"
 			"0:	nr	%0,%5\n"
 			"	lr	%1,%0\n"
 			"	or	%0,%2\n"
 			"	or	%1,%3\n"
-			"	cs	%0,%1,0(%4)\n"
+			"	cs	%0,%1,%2\n"
 			"	jnl	1f\n"
 			"	xr	%1,%0\n"
 			"	nr	%1,%5\n"
 			"	jnz	0b\n"
 			"1:"
-			: "=&d" (prev), "=&d" (tmp)
-			: "d" (old << shift), "d" (new << shift), "a" (ptr),
-			  "d" (~(65535 << shift))
+			: "=&d" (prev), "=&d" (tmp), "=Q" (*(int *) ptr)
+			: "d" (old << shift), "d" (new << shift),
+			  "d" (~(65535 << shift)), "Q" (*(int *) ptr)
 			: "memory", "cc");
 		return prev >> shift;
 	case 4:
 		asm volatile(
-			"	cs	%0,%2,0(%3)\n"
-			: "=&d" (prev) : "0" (old), "d" (new), "a" (ptr)
+			"	cs	%0,%3,%1\n"
+			: "=&d" (prev), "=Q" (*(int *) ptr)
+			: "0" (old), "d" (new), "Q" (*(int *) ptr)
 			: "memory", "cc");
 		return prev;
 #ifdef __s390x__
 	case 8:
 		asm volatile(
-			"	csg	%0,%2,0(%3)\n"
-			: "=&d" (prev) : "0" (old), "d" (new), "a" (ptr)
+			"	csg	%0,%3,%1\n"
+			: "=&d" (prev), "=Q" (*(long *) ptr)
+			: "0" (old), "d" (new), "Q" (*(long *) ptr)
 			: "memory", "cc");
 		return prev;
 #endif /* __s390x__ */
@@ -302,17 +304,17 @@ __cmpxchg(volatile void *ptr, unsigned long old, unsigned long new, int size)
 #define __ctl_load(array, low, high) ({				\
 	typedef struct { char _[sizeof(array)]; } addrtype;	\
 	asm volatile(						\
-		"	lctlg	%1,%2,0(%0)\n"			\
-		: : "a" (&array), "i" (low), "i" (high),	\
-		    "m" (*(addrtype *)(&array)));		\
+		"	lctlg	%1,%2,%0\n"			\
+		: : "Q" (*(addrtype *)(&array)),		\
+		    "i" (low), "i" (high));			\
 	})
 
 #define __ctl_store(array, low, high) ({			\
 	typedef struct { char _[sizeof(array)]; } addrtype;	\
 	asm volatile(						\
-		"	stctg	%2,%3,0(%1)\n"			\
-		: "=m" (*(addrtype *)(&array))			\
-		: "a" (&array), "i" (low), "i" (high));		\
+		"	stctg	%1,%2,%0\n"			\
+		: "=Q" (*(addrtype *)(&array))			\
+		: "i" (low), "i" (high));			\
 	})
 
 #else /* __s390x__ */
@@ -320,17 +322,17 @@ __cmpxchg(volatile void *ptr, unsigned long old, unsigned long new, int size)
 #define __ctl_load(array, low, high) ({				\
 	typedef struct { char _[sizeof(array)]; } addrtype;	\
 	asm volatile(						\
-		"	lctl	%1,%2,0(%0)\n"			\
-		: : "a" (&array), "i" (low), "i" (high),	\
-		    "m" (*(addrtype *)(&array)));		\
+		"	lctl	%1,%2,%0\n"			\
+		: : "Q" (*(addrtype *)(&array)),		\
+		    "i" (low), "i" (high));			\
 })
 
 #define __ctl_store(array, low, high) ({			\
 	typedef struct { char _[sizeof(array)]; } addrtype;	\
 	asm volatile(						\
-		"	stctl	%2,%3,0(%1)\n"			\
-		: "=m" (*(addrtype *)(&array))			\
-		: "a" (&array), "i" (low), "i" (high));		\
+		"	stctl	%1,%2,%0\n"			\
+		: "=Q" (*(addrtype *)(&array))			\
+		: "i" (low), "i" (high));			\
 	})
 
 #endif /* __s390x__ */
diff --git a/arch/s390/include/asm/timex.h b/arch/s390/include/asm/timex.h
index 68d9fea..f174bda 100644
--- a/arch/s390/include/asm/timex.h
+++ b/arch/s390/include/asm/timex.h
@@ -20,10 +20,10 @@ static inline int set_clock(__u64 time)
 	int cc;
 
 	asm volatile(
-		"   sck   0(%2)\n"
+		"   sck   %1\n"
 		"   ipm   %0\n"
 		"   srl   %0,28\n"
-		: "=d" (cc) : "m" (time), "a" (&time) : "cc");
+		: "=d" (cc) : "Q" (time) : "cc");
 	return cc;
 }
 
@@ -32,21 +32,21 @@ static inline int store_clock(__u64 *time)
 	int cc;
 
 	asm volatile(
-		"   stck  0(%2)\n"
+		"   stck  %1\n"
 		"   ipm   %0\n"
 		"   srl   %0,28\n"
-		: "=d" (cc), "=m" (*time) : "a" (time) : "cc");
+		: "=d" (cc), "=Q" (*time) : : "cc");
 	return cc;
 }
 
 static inline void set_clock_comparator(__u64 time)
 {
-	asm volatile("sckc 0(%1)" : : "m" (time), "a" (&time));
+	asm volatile("sckc %0" : : "Q" (time));
 }
 
 static inline void store_clock_comparator(__u64 *time)
 {
-	asm volatile("stckc 0(%1)" : "=m" (*time) : "a" (time));
+	asm volatile("stckc %0" : "=Q" (*time));
 }
 
 #define CLOCK_TICK_RATE	1193180 /* Underlying HZ */
@@ -57,11 +57,7 @@ static inline unsigned long long get_clock (void)
 {
 	unsigned long long clk;
 
-#if __GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ > 2)
 	asm volatile("stck %0" : "=Q" (clk) : : "cc");
-#else /* __GNUC__ */
-	asm volatile("stck 0(%1)" : "=m" (clk) : "a" (&clk) : "cc");
-#endif /* __GNUC__ */
 	return clk;
 }
 
@@ -69,13 +65,7 @@ static inline unsigned long long get_clock_xt(void)
 {
 	unsigned char clk[16];
 
-#if __GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ > 2)
 	asm volatile("stcke %0" : "=Q" (clk) : : "cc");
-#else /* __GNUC__ */
-	asm volatile("stcke 0(%1)" : "=m" (clk)
-				   : "a" (clk) : "cc");
-#endif /* __GNUC__ */
-
 	return *((unsigned long long *)&clk[1]);
 }
 
diff --git a/arch/s390/kernel/asm-offsets.c b/arch/s390/kernel/asm-offsets.c
index 63e4643..a5850a0 100644
--- a/arch/s390/kernel/asm-offsets.c
+++ b/arch/s390/kernel/asm-offsets.c
@@ -9,6 +9,14 @@
 #include <asm/vdso.h>
 #include <asm/sigp.h>
 
+/*
+ * Make sure that the compiler is new enough. We want a compiler that
+ * is known to work with the "Q" assembler constraint.
+ */
+#if __GNUC__ < 3 || (__GNUC__ == 3 && __GNUC_MINOR__ < 3)
+#error Your compiler is too old; please use version 3.3.3 or newer
+#endif
+
 int main(void)
 {
 	DEFINE(__THREAD_info, offsetof(struct task_struct, stack));