Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux

Pull second set of s390 patches from Martin Schwidefsky:
 "The second part of Heikos uaccess rework, the page table walker for
  uaccess is now a thing of the past (yay!)

  The code change to fix the theoretical TLB flush problem allows us to
  add a TLB flush optimization for zEC12, this machine has new
  instructions that allow to do CPU local TLB flushes for single pages
  and for all pages of a specific address space.

  Plus the usual bug fixing and some more cleanup"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux:
  s390/uaccess: rework uaccess code - fix locking issues
  s390/mm,tlb: optimize TLB flushing for zEC12
  s390/mm,tlb: safeguard against speculative TLB creation
  s390/irq: Use defines for external interruption codes
  s390/irq: Add defines for external interruption codes
  s390/sclp: add timeout for queued requests
  kvm/s390: also set guest pages back to stable on kexec/kdump
  lcs: Add missing destroy_timer_on_stack()
  s390/tape: Add missing destroy_timer_on_stack()
  s390/tape: Use del_timer_sync()
  s390/3270: fix crash with multiple reset device requests
  s390/bitops,atomic: add missing memory barriers
  s390/zcrypt: add length check for aligned data to avoid overflow in msg-type 6

arch/s390/include/asm/atomic.h

@@ -15,23 +15,29 @@
 
 #include <linux/compiler.h>
 #include <linux/types.h>
+#include <asm/barrier.h>
 #include <asm/cmpxchg.h>
 
 #define ATOMIC_INIT(i)  { (i) }
 
+#define __ATOMIC_NO_BARRIER	"\n"
+
 #ifdef CONFIG_HAVE_MARCH_Z196_FEATURES
 
 #define __ATOMIC_OR	"lao"
 #define __ATOMIC_AND	"lan"
 #define __ATOMIC_ADD	"laa"
+#define __ATOMIC_BARRIER "bcr	14,0\n"
 
-#define __ATOMIC_LOOP(ptr, op_val, op_string)				\
+#define __ATOMIC_LOOP(ptr, op_val, op_string, __barrier)		\
 ({									\
 	int old_val;							\
 									\
 	typecheck(atomic_t *, ptr);					\
 	asm volatile(							\
+		__barrier						\
 		op_string "	%0,%2,%1\n"				\
+		__barrier						\
 		: "=d" (old_val), "+Q" ((ptr)->counter)			\
 		: "d" (op_val)						\
 		: "cc", "memory");					\
@@ -43,8 +49,9 @@
 #define __ATOMIC_OR	"or"
 #define __ATOMIC_AND	"nr"
 #define __ATOMIC_ADD	"ar"
+#define __ATOMIC_BARRIER "\n"
 
-#define __ATOMIC_LOOP(ptr, op_val, op_string)				\
+#define __ATOMIC_LOOP(ptr, op_val, op_string, __barrier)		\
 ({									\
 	int old_val, new_val;						\
 									\
@@ -82,7 +89,7 @@ static inline void atomic_set(atomic_t *v, int i)
 
 static inline int atomic_add_return(int i, atomic_t *v)
 {
-	return __ATOMIC_LOOP(v, i, __ATOMIC_ADD) + i;
+	return __ATOMIC_LOOP(v, i, __ATOMIC_ADD, __ATOMIC_BARRIER) + i;
 }
 
 static inline void atomic_add(int i, atomic_t *v)
@@ -94,12 +101,10 @@ static inline void atomic_add(int i, atomic_t *v)
 			: "+Q" (v->counter)
 			: "i" (i)
 			: "cc", "memory");
-	} else {
-		atomic_add_return(i, v);
+		return;
 	}
-#else
-	atomic_add_return(i, v);
 #endif
+	__ATOMIC_LOOP(v, i, __ATOMIC_ADD, __ATOMIC_NO_BARRIER);
 }
 
 #define atomic_add_negative(_i, _v)	(atomic_add_return(_i, _v) < 0)
@@ -115,12 +120,12 @@ static inline void atomic_add(int i, atomic_t *v)
 
 static inline void atomic_clear_mask(unsigned int mask, atomic_t *v)
 {
-	__ATOMIC_LOOP(v, ~mask, __ATOMIC_AND);
+	__ATOMIC_LOOP(v, ~mask, __ATOMIC_AND, __ATOMIC_NO_BARRIER);
 }
 
 static inline void atomic_set_mask(unsigned int mask, atomic_t *v)
 {
-	__ATOMIC_LOOP(v, mask, __ATOMIC_OR);
+	__ATOMIC_LOOP(v, mask, __ATOMIC_OR, __ATOMIC_NO_BARRIER);
 }
 
 #define atomic_xchg(v, new) (xchg(&((v)->counter), new))
@@ -157,19 +162,24 @@ static inline int __atomic_add_unless(atomic_t *v, int a, int u)
 
 #ifdef CONFIG_64BIT
 
+#define __ATOMIC64_NO_BARRIER	"\n"
+
 #ifdef CONFIG_HAVE_MARCH_Z196_FEATURES
 
 #define __ATOMIC64_OR	"laog"
 #define __ATOMIC64_AND	"lang"
 #define __ATOMIC64_ADD	"laag"
+#define __ATOMIC64_BARRIER "bcr	14,0\n"
 
-#define __ATOMIC64_LOOP(ptr, op_val, op_string)				\
+#define __ATOMIC64_LOOP(ptr, op_val, op_string, __barrier)		\
 ({									\
 	long long old_val;						\
 									\
 	typecheck(atomic64_t *, ptr);					\
 	asm volatile(							\
+		__barrier						\
 		op_string "	%0,%2,%1\n"				\
+		__barrier						\
 		: "=d" (old_val), "+Q" ((ptr)->counter)			\
 		: "d" (op_val)						\
 		: "cc", "memory");					\
@@ -181,8 +191,9 @@ static inline int __atomic_add_unless(atomic_t *v, int a, int u)
 #define __ATOMIC64_OR	"ogr"
 #define __ATOMIC64_AND	"ngr"
 #define __ATOMIC64_ADD	"agr"
+#define __ATOMIC64_BARRIER "\n"
 
-#define __ATOMIC64_LOOP(ptr, op_val, op_string)				\
+#define __ATOMIC64_LOOP(ptr, op_val, op_string, __barrier)		\
 ({									\
 	long long old_val, new_val;					\
 									\
@@ -220,17 +231,32 @@ static inline void atomic64_set(atomic64_t *v, long long i)
 
 static inline long long atomic64_add_return(long long i, atomic64_t *v)
 {
-	return __ATOMIC64_LOOP(v, i, __ATOMIC64_ADD) + i;
+	return __ATOMIC64_LOOP(v, i, __ATOMIC64_ADD, __ATOMIC64_BARRIER) + i;
+}
+
+static inline void atomic64_add(long long i, atomic64_t *v)
+{
+#ifdef CONFIG_HAVE_MARCH_Z196_FEATURES
+	if (__builtin_constant_p(i) && (i > -129) && (i < 128)) {
+		asm volatile(
+			"agsi	%0,%1\n"
+			: "+Q" (v->counter)
+			: "i" (i)
+			: "cc", "memory");
+		return;
+	}
+#endif
+	__ATOMIC64_LOOP(v, i, __ATOMIC64_ADD, __ATOMIC64_NO_BARRIER);
 }
 
 static inline void atomic64_clear_mask(unsigned long mask, atomic64_t *v)
 {
-	__ATOMIC64_LOOP(v, ~mask, __ATOMIC64_AND);
+	__ATOMIC64_LOOP(v, ~mask, __ATOMIC64_AND, __ATOMIC64_NO_BARRIER);
 }
 
 static inline void atomic64_set_mask(unsigned long mask, atomic64_t *v)
 {
-	__ATOMIC64_LOOP(v, mask, __ATOMIC64_OR);
+	__ATOMIC64_LOOP(v, mask, __ATOMIC64_OR, __ATOMIC64_NO_BARRIER);
 }
 
 #define atomic64_xchg(v, new) (xchg(&((v)->counter), new))
@@ -334,25 +360,13 @@ static inline void atomic64_clear_mask(unsigned long long mask, atomic64_t *v)
 	} while (atomic64_cmpxchg(v, old, new) != old);
 }
 
-#endif /* CONFIG_64BIT */
-
 static inline void atomic64_add(long long i, atomic64_t *v)
 {
-#ifdef CONFIG_HAVE_MARCH_Z196_FEATURES
-	if (__builtin_constant_p(i) && (i > -129) && (i < 128)) {
-		asm volatile(
-			"agsi	%0,%1\n"
-			: "+Q" (v->counter)
-			: "i" (i)
-			: "cc", "memory");
-	} else {
-		atomic64_add_return(i, v);
-	}
-#else
 	atomic64_add_return(i, v);
-#endif
 }
 
+#endif /* CONFIG_64BIT */
+
 static inline int atomic64_add_unless(atomic64_t *v, long long i, long long u)
 {
 	long long c, old;

arch/s390/include/asm/bitops.h

@@ -47,14 +47,18 @@
 
 #include <linux/typecheck.h>
 #include <linux/compiler.h>
+#include <asm/barrier.h>
+
+#define __BITOPS_NO_BARRIER	"\n"
 
 #ifndef CONFIG_64BIT
 
 #define __BITOPS_OR		"or"
 #define __BITOPS_AND		"nr"
 #define __BITOPS_XOR		"xr"
+#define __BITOPS_BARRIER	"\n"
 
-#define __BITOPS_LOOP(__addr, __val, __op_string)		\
+#define __BITOPS_LOOP(__addr, __val, __op_string, __barrier)	\
 ({								\
 	unsigned long __old, __new;				\
 								\
@@ -67,7 +71,7 @@
 		"	jl	0b"				\
 		: "=&d" (__old), "=&d" (__new), "+Q" (*(__addr))\
 		: "d" (__val)					\
-		: "cc");					\
+		: "cc", "memory");				\
 	__old;							\
 })
 
@@ -78,17 +82,20 @@
 #define __BITOPS_OR		"laog"
 #define __BITOPS_AND		"lang"
 #define __BITOPS_XOR		"laxg"
+#define __BITOPS_BARRIER	"bcr	14,0\n"
 
-#define __BITOPS_LOOP(__addr, __val, __op_string)		\
+#define __BITOPS_LOOP(__addr, __val, __op_string, __barrier)	\
 ({								\
 	unsigned long __old;					\
 								\
 	typecheck(unsigned long *, (__addr));			\
 	asm volatile(						\
+		__barrier					\
 		__op_string "	%0,%2,%1\n"			\
+		__barrier					\
 		: "=d" (__old),	"+Q" (*(__addr))		\
 		: "d" (__val)					\
-		: "cc");					\
+		: "cc", "memory");				\
 	__old;							\
 })
 
@@ -97,8 +104,9 @@
 #define __BITOPS_OR		"ogr"
 #define __BITOPS_AND		"ngr"
 #define __BITOPS_XOR		"xgr"
+#define __BITOPS_BARRIER	"\n"
 
-#define __BITOPS_LOOP(__addr, __val, __op_string)		\
+#define __BITOPS_LOOP(__addr, __val, __op_string, __barrier)	\
 ({								\
 	unsigned long __old, __new;				\
 								\
@@ -111,7 +119,7 @@
 		"	jl	0b"				\
 		: "=&d" (__old), "=&d" (__new), "+Q" (*(__addr))\
 		: "d" (__val)					\
-		: "cc");					\
+		: "cc", "memory");				\
 	__old;							\
 })
 
@@ -149,12 +157,12 @@ static inline void set_bit(unsigned long nr, volatile unsigned long *ptr)
 			"oi	%0,%b1\n"
 			: "+Q" (*caddr)
 			: "i" (1 << (nr & 7))
-			: "cc");
+			: "cc", "memory");
 		return;
 	}
 #endif
 	mask = 1UL << (nr & (BITS_PER_LONG - 1));
-	__BITOPS_LOOP(addr, mask, __BITOPS_OR);
+	__BITOPS_LOOP(addr, mask, __BITOPS_OR, __BITOPS_NO_BARRIER);
 }
 
 static inline void clear_bit(unsigned long nr, volatile unsigned long *ptr)
@@ -170,12 +178,12 @@ static inline void clear_bit(unsigned long nr, volatile unsigned long *ptr)
 			"ni	%0,%b1\n"
 			: "+Q" (*caddr)
 			: "i" (~(1 << (nr & 7)))
-			: "cc");
+			: "cc", "memory");
 		return;
 	}
 #endif
 	mask = ~(1UL << (nr & (BITS_PER_LONG - 1)));
-	__BITOPS_LOOP(addr, mask, __BITOPS_AND);
+	__BITOPS_LOOP(addr, mask, __BITOPS_AND, __BITOPS_NO_BARRIER);
 }
 
 static inline void change_bit(unsigned long nr, volatile unsigned long *ptr)
@@ -191,12 +199,12 @@ static inline void change_bit(unsigned long nr, volatile unsigned long *ptr)
 			"xi	%0,%b1\n"
 			: "+Q" (*caddr)
 			: "i" (1 << (nr & 7))
-			: "cc");
+			: "cc", "memory");
 		return;
 	}
 #endif
 	mask = 1UL << (nr & (BITS_PER_LONG - 1));
-	__BITOPS_LOOP(addr, mask, __BITOPS_XOR);
+	__BITOPS_LOOP(addr, mask, __BITOPS_XOR, __BITOPS_NO_BARRIER);
 }
 
 static inline int
@@ -206,8 +214,7 @@ test_and_set_bit(unsigned long nr, volatile unsigned long *ptr)
 	unsigned long old, mask;
 
 	mask = 1UL << (nr & (BITS_PER_LONG - 1));
-	old = __BITOPS_LOOP(addr, mask, __BITOPS_OR);
-	barrier();
+	old = __BITOPS_LOOP(addr, mask, __BITOPS_OR, __BITOPS_BARRIER);
 	return (old & mask) != 0;
 }
 
@@ -218,8 +225,7 @@ test_and_clear_bit(unsigned long nr, volatile unsigned long *ptr)
 	unsigned long old, mask;
 
 	mask = ~(1UL << (nr & (BITS_PER_LONG - 1)));
-	old = __BITOPS_LOOP(addr, mask, __BITOPS_AND);
-	barrier();
+	old = __BITOPS_LOOP(addr, mask, __BITOPS_AND, __BITOPS_BARRIER);
 	return (old & ~mask) != 0;
 }
 
@@ -230,8 +236,7 @@ test_and_change_bit(unsigned long nr, volatile unsigned long *ptr)
 	unsigned long old, mask;
 
 	mask = 1UL << (nr & (BITS_PER_LONG - 1));
-	old = __BITOPS_LOOP(addr, mask, __BITOPS_XOR);
-	barrier();
+	old = __BITOPS_LOOP(addr, mask, __BITOPS_XOR, __BITOPS_BARRIER);
 	return (old & mask) != 0;
 }
 

arch/s390/include/asm/futex.h

 #ifndef _ASM_S390_FUTEX_H
 #define _ASM_S390_FUTEX_H
 
-#include <linux/futex.h>
 #include <linux/uaccess.h>
+#include <linux/futex.h>
+#include <asm/mmu_context.h>
 #include <asm/errno.h>
 
-int futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr, u32 oldval, u32 newval);
-int __futex_atomic_op_inuser(int op, u32 __user *uaddr, int oparg, int *old);
+#define __futex_atomic_op(insn, ret, oldval, newval, uaddr, oparg)	\
+	asm volatile(							\
+		"   sacf  256\n"					\
+		"0: l     %1,0(%6)\n"					\
+		"1:"insn						\
+		"2: cs    %1,%2,0(%6)\n"				\
+		"3: jl    1b\n"						\
+		"   lhi   %0,0\n"					\
+		"4: sacf  768\n"					\
+		EX_TABLE(0b,4b) EX_TABLE(2b,4b) EX_TABLE(3b,4b)		\
+		: "=d" (ret), "=&d" (oldval), "=&d" (newval),		\
+		  "=m" (*uaddr)						\
+		: "0" (-EFAULT), "d" (oparg), "a" (uaddr),		\
+		  "m" (*uaddr) : "cc");
 
 static inline int futex_atomic_op_inuser(int encoded_op, u32 __user *uaddr)
 {
@@ -14,13 +27,37 @@ static inline int futex_atomic_op_inuser(int encoded_op, u32 __user *uaddr)
 	int cmp = (encoded_op >> 24) & 15;
 	int oparg = (encoded_op << 8) >> 20;
 	int cmparg = (encoded_op << 20) >> 20;
-	int oldval, ret;
+	int oldval = 0, newval, ret;
 
+	update_primary_asce(current);
 	if (encoded_op & (FUTEX_OP_OPARG_SHIFT << 28))
 		oparg = 1 << oparg;
 
 	pagefault_disable();
-	ret = __futex_atomic_op_inuser(op, uaddr, oparg, &oldval);
+	switch (op) {
+	case FUTEX_OP_SET:
+		__futex_atomic_op("lr %2,%5\n",
+				  ret, oldval, newval, uaddr, oparg);
+		break;
+	case FUTEX_OP_ADD:
+		__futex_atomic_op("lr %2,%1\nar %2,%5\n",
+				  ret, oldval, newval, uaddr, oparg);
+		break;
+	case FUTEX_OP_OR:
+		__futex_atomic_op("lr %2,%1\nor %2,%5\n",
+				  ret, oldval, newval, uaddr, oparg);
+		break;
+	case FUTEX_OP_ANDN:
+		__futex_atomic_op("lr %2,%1\nnr %2,%5\n",
+				  ret, oldval, newval, uaddr, oparg);
+		break;
+	case FUTEX_OP_XOR:
+		__futex_atomic_op("lr %2,%1\nxr %2,%5\n",
+				  ret, oldval, newval, uaddr, oparg);
+		break;
+	default:
+		ret = -ENOSYS;
+	}
 	pagefault_enable();
 
 	if (!ret) {
@@ -37,4 +74,23 @@ static inline int futex_atomic_op_inuser(int encoded_op, u32 __user *uaddr)
 	return ret;
 }
 
+static inline int futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr,
+						u32 oldval, u32 newval)
+{
+	int ret;
+
+	update_primary_asce(current);
+	asm volatile(
+		"   sacf 256\n"
+		"0: cs   %1,%4,0(%5)\n"
+		"1: la   %0,0\n"
+		"2: sacf 768\n"
+		EX_TABLE(0b,2b) EX_TABLE(1b,2b)
+		: "=d" (ret), "+d" (oldval), "=m" (*uaddr)
+		: "0" (-EFAULT), "d" (newval), "a" (uaddr), "m" (*uaddr)
+		: "cc", "memory");
+	*uval = oldval;
+	return ret;
+}
+
 #endif /* _ASM_S390_FUTEX_H */

arch/s390/include/asm/irq.h

@@ -16,6 +16,20 @@
 /* This number is used when no interrupt has been assigned */
 #define NO_IRQ		0
 
+/* External interruption codes */
+#define EXT_IRQ_INTERRUPT_KEY	0x0040
+#define EXT_IRQ_CLK_COMP	0x1004
+#define EXT_IRQ_CPU_TIMER	0x1005
+#define EXT_IRQ_WARNING_TRACK	0x1007
+#define EXT_IRQ_MALFUNC_ALERT	0x1200
+#define EXT_IRQ_EMERGENCY_SIG	0x1201
+#define EXT_IRQ_EXTERNAL_CALL	0x1202
+#define EXT_IRQ_TIMING_ALERT	0x1406
+#define EXT_IRQ_MEASURE_ALERT	0x1407
+#define EXT_IRQ_SERVICE_SIG	0x2401
+#define EXT_IRQ_CP_SERVICE	0x2603
+#define EXT_IRQ_IUCV		0x4000
+
 #ifndef __ASSEMBLY__
 
 #include <linux/hardirq.h>
@@ -77,8 +91,8 @@ struct ext_code {
 
 typedef void (*ext_int_handler_t)(struct ext_code, unsigned int, unsigned long);
 
-int register_external_interrupt(u16 code, ext_int_handler_t handler);
-int unregister_external_interrupt(u16 code, ext_int_handler_t handler);
+int register_external_irq(u16 code, ext_int_handler_t handler);
+int unregister_external_irq(u16 code, ext_int_handler_t handler);
 
 enum irq_subclass {
 	IRQ_SUBCLASS_MEASUREMENT_ALERT = 5,

arch/s390/include/asm/mmu.h

 #ifndef __MMU_H
 #define __MMU_H
 
+#include <linux/cpumask.h>
 #include <linux/errno.h>
 
 typedef struct {
+	cpumask_t cpu_attach_mask;
 	atomic_t attach_count;
 	unsigned int flush_mm;
 	spinlock_t list_lock;

arch/s390/include/asm/mmu_context.h

@@ -15,6 +15,7 @@
 static inline int init_new_context(struct task_struct *tsk,
 				   struct mm_struct *mm)
 {
+	cpumask_clear(&mm->context.cpu_attach_mask);
 	atomic_set(&mm->context.attach_count, 0);
 	mm->context.flush_mm = 0;
 	mm->context.asce_bits = _ASCE_TABLE_LENGTH | _ASCE_USER_BITS;
@@ -29,41 +30,61 @@ static inline int init_new_context(struct task_struct *tsk,
 
 #define destroy_context(mm)             do { } while (0)
 
-#ifndef CONFIG_64BIT
-#define LCTL_OPCODE "lctl"
-#else
-#define LCTL_OPCODE "lctlg"
-#endif
-
-static inline void update_mm(struct mm_struct *mm, struct task_struct *tsk)
+static inline void update_user_asce(struct mm_struct *mm, int load_primary)
 {
 	pgd_t *pgd = mm->pgd;
 
 	S390_lowcore.user_asce = mm->context.asce_bits | __pa(pgd);
-	/* Load primary space page table origin. */
-	asm volatile(LCTL_OPCODE" 1,1,%0\n" : : "m" (S390_lowcore.user_asce));
+	if (load_primary)
+		__ctl_load(S390_lowcore.user_asce, 1, 1);
 	set_fs(current->thread.mm_segment);
 }
 
+static inline void clear_user_asce(struct mm_struct *mm, int load_primary)
+{
+	S390_lowcore.user_asce = S390_lowcore.kernel_asce;
+
+	if (load_primary)
+		__ctl_load(S390_lowcore.user_asce, 1, 1);
+	__ctl_load(S390_lowcore.user_asce, 7, 7);
+}
+
+static inline void update_primary_asce(struct task_struct *tsk)
+{
+	unsigned long asce;
+
+	__ctl_store(asce, 1, 1);
+	if (asce != S390_lowcore.kernel_asce)
+		__ctl_load(S390_lowcore.kernel_asce, 1, 1);
+	set_tsk_thread_flag(tsk, TIF_ASCE);
+}
+
 static inline void switch_mm(struct mm_struct *prev, struct mm_struct *next,
 			     struct task_struct *tsk)
 {
 	int cpu = smp_processor_id();
 
+	update_primary_asce(tsk);
 	if (prev == next)
 		return;
+	if (MACHINE_HAS_TLB_LC)
+		cpumask_set_cpu(cpu, &next->context.cpu_attach_mask);
 	if (atomic_inc_return(&next->context.attach_count) >> 16) {
-		/* Delay update_mm until all TLB flushes are done. */
+		/* Delay update_user_asce until all TLB flushes are done. */
 		set_tsk_thread_flag(tsk, TIF_TLB_WAIT);
+		/* Clear old ASCE by loading the kernel ASCE. */
+		clear_user_asce(next, 0);
 	} else {
 		cpumask_set_cpu(cpu, mm_cpumask(next));
-		update_mm(next, tsk);
+		update_user_asce(next, 0);
 		if (next->context.flush_mm)
 			/* Flush pending TLBs */
 			__tlb_flush_mm(next);
 	}
 	atomic_dec(&prev->context.attach_count);
 	WARN_ON(atomic_read(&prev->context.attach_count) < 0);
+	if (MACHINE_HAS_TLB_LC)
+		cpumask_clear_cpu(cpu, &prev->context.cpu_attach_mask);
 }
 
 #define finish_arch_post_lock_switch finish_arch_post_lock_switch
@@ -80,7 +101,7 @@ static inline void finish_arch_post_lock_switch(void)
 		cpu_relax();
 
 	cpumask_set_cpu(smp_processor_id(), mm_cpumask(mm));
-	update_mm(mm, tsk);
+	update_user_asce(mm, 0);
 	if (mm->context.flush_mm)
 		__tlb_flush_mm(mm);
 	preempt_enable();

arch/s390/include/asm/pgtable.h

@@ -1070,12 +1070,35 @@ static inline void __ptep_ipte(unsigned long address, pte_t *ptep)
 		: "=m" (*ptep) : "m" (*ptep), "a" (pto), "a" (address));
 }
 
+static inline void __ptep_ipte_local(unsigned long address, pte_t *ptep)
+{
+	unsigned long pto = (unsigned long) ptep;
+
+#ifndef CONFIG_64BIT
+	/* pto in ESA mode must point to the start of the segment table */
+	pto &= 0x7ffffc00;
+#endif
+	/* Invalidation + local TLB flush for the pte */
+	asm volatile(
+		"	.insn rrf,0xb2210000,%2,%3,0,1"
+		: "=m" (*ptep) : "m" (*ptep), "a" (pto), "a" (address));
+}
+
 static inline void ptep_flush_direct(struct mm_struct *mm,
 				     unsigned long address, pte_t *ptep)
 {
+	int active, count;
+
 	if (pte_val(*ptep) & _PAGE_INVALID)
 		return;
+	active = (mm == current->active_mm) ? 1 : 0;
+	count = atomic_add_return(0x10000, &mm->context.attach_count);
+	if (MACHINE_HAS_TLB_LC && (count & 0xffff) <= active &&
+	    cpumask_equal(mm_cpumask(mm), cpumask_of(smp_processor_id())))
+		__ptep_ipte_local(address, ptep);
+	else
 		__ptep_ipte(address, ptep);
+	atomic_sub(0x10000, &mm->context.attach_count);
 }
 
 static inline void ptep_flush_lazy(struct mm_struct *mm,
@@ -1384,35 +1407,6 @@ static inline pmd_t *pmd_offset(pud_t *pud, unsigned long address)
 #define pte_offset_map(pmd, address) pte_offset_kernel(pmd, address)
 #define pte_unmap(pte) do { } while (0)
 
-static inline void __pmd_idte(unsigned long address, pmd_t *pmdp)
-{
-	unsigned long sto = (unsigned long) pmdp -
-			    pmd_index(address) * sizeof(pmd_t);
-
-	if (!(pmd_val(*pmdp) & _SEGMENT_ENTRY_INVALID)) {
-		asm volatile(
-			"	.insn	rrf,0xb98e0000,%2,%3,0,0"
-			: "=m" (*pmdp)
-			: "m" (*pmdp), "a" (sto),
-			  "a" ((address & HPAGE_MASK))
-			: "cc"
-		);
-	}
-}
-
-static inline void __pmd_csp(pmd_t *pmdp)
-{
-	register unsigned long reg2 asm("2") = pmd_val(*pmdp);
-	register unsigned long reg3 asm("3") = pmd_val(*pmdp) |
-					       _SEGMENT_ENTRY_INVALID;
-	register unsigned long reg4 asm("4") = ((unsigned long) pmdp) + 5;
-
-	asm volatile(
-		"	csp %1,%3"
-		: "=m" (*pmdp)
-		: "d" (reg2), "d" (reg3), "d" (reg4), "m" (*pmdp) : "cc");
-}
-
 #if defined(CONFIG_TRANSPARENT_HUGEPAGE) || defined(CONFIG_HUGETLB_PAGE)
 static inline unsigned long massage_pgprot_pmd(pgprot_t pgprot)
 {
@@ -1481,18 +1475,80 @@ static inline pmd_t pmd_mkwrite(pmd_t pmd)
 }
 #endif /* CONFIG_TRANSPARENT_HUGEPAGE || CONFIG_HUGETLB_PAGE */
 
+static inline void __pmdp_csp(pmd_t *pmdp)
+{
+	register unsigned long reg2 asm("2") = pmd_val(*pmdp);
+	register unsigned long reg3 asm("3") = pmd_val(*pmdp) |
+					       _SEGMENT_ENTRY_INVALID;
+	register unsigned long reg4 asm("4") = ((unsigned long) pmdp) + 5;
+
+	asm volatile(
+		"	csp %1,%3"
+		: "=m" (*pmdp)
+		: "d" (reg2), "d" (reg3), "d" (reg4), "m" (*pmdp) : "cc");
+}
+
+static inline void __pmdp_idte(unsigned long address, pmd_t *pmdp)
+{
+	unsigned long sto;
+
+	sto = (unsigned long) pmdp - pmd_index(address) * sizeof(pmd_t);
+	asm volatile(
+		"	.insn	rrf,0xb98e0000,%2,%3,0,0"
+		: "=m" (*pmdp)
+		: "m" (*pmdp), "a" (sto), "a" ((address & HPAGE_MASK))
+		: "cc" );
+}
+
+static inline void __pmdp_idte_local(unsigned long address, pmd_t *pmdp)
+{
+	unsigned long sto;
+
+	sto = (unsigned long) pmdp - pmd_index(address) * sizeof(pmd_t);
+	asm volatile(
+		"	.insn	rrf,0xb98e0000,%2,%3,0,1"
+		: "=m" (*pmdp)
+		: "m" (*pmdp), "a" (sto), "a" ((address & HPAGE_MASK))
+		: "cc" );
+}
+
+static inline void pmdp_flush_direct(struct mm_struct *mm,
+				     unsigned long address, pmd_t *pmdp)
+{
+	int active, count;
+
+	if (pmd_val(*pmdp) & _SEGMENT_ENTRY_INVALID)
+		return;
+	if (!MACHINE_HAS_IDTE) {
+		__pmdp_csp(pmdp);
+		return;
+	}
+	active = (mm == current->active_mm) ? 1 : 0;
+	count = atomic_add_return(0x10000, &mm->context.attach_count);
+	if (MACHINE_HAS_TLB_LC && (count & 0xffff) <= active &&
+	    cpumask_equal(mm_cpumask(mm), cpumask_of(smp_processor_id())))
+		__pmdp_idte_local(address, pmdp);
+	else
+		__pmdp_idte(address, pmdp);
+	atomic_sub(0x10000, &mm->context.attach_count);
+}
+
 static inline void pmdp_flush_lazy(struct mm_struct *mm,
 				   unsigned long address, pmd_t *pmdp)
 {
 	int active, count;
 
+	if (pmd_val(*pmdp) & _SEGMENT_ENTRY_INVALID)
+		return;
 	active = (mm == current->active_mm) ? 1 : 0;
 	count = atomic_add_return(0x10000, &mm->context.attach_count);
 	if ((count & 0xffff) <= active) {
 		pmd_val(*pmdp) |= _SEGMENT_ENTRY_INVALID;
 		mm->context.flush_mm = 1;
-	} else
-		__pmd_idte(address, pmdp);
+	} else if (MACHINE_HAS_IDTE)
+		__pmdp_idte(address, pmdp);
+	else
+		__pmdp_csp(pmdp);
 	atomic_sub(0x10000, &mm->context.attach_count);
 }
 
@@ -1545,7 +1601,7 @@ static inline int pmdp_test_and_clear_young(struct vm_area_struct *vma,
 	pmd_t pmd;
 
 	pmd = *pmdp;
-	__pmd_idte(address, pmdp);
+	pmdp_flush_direct(vma->vm_mm, address, pmdp);
 	*pmdp = pmd_mkold(pmd);
 	return pmd_young(pmd);
 }
@@ -1556,7 +1612,7 @@ static inline pmd_t pmdp_get_and_clear(struct mm_struct *mm,
 {
 	pmd_t pmd = *pmdp;
 
-	__pmd_idte(address, pmdp);
+	pmdp_flush_direct(mm, address, pmdp);
 	pmd_clear(pmdp);
 	return pmd;
 }
@@ -1572,7 +1628,7 @@ static inline pmd_t pmdp_clear_flush(struct vm_area_struct *vma,
 static inline void pmdp_invalidate(struct vm_area_struct *vma,
 				   unsigned long address, pmd_t *pmdp)
 {
-	__pmd_idte(address, pmdp);
+	pmdp_flush_direct(vma->vm_mm, address, pmdp);
 }
 
 #define __HAVE_ARCH_PMDP_SET_WRPROTECT
@@ -1582,7 +1638,7 @@ static inline void pmdp_set_wrprotect(struct mm_struct *mm,
 	pmd_t pmd = *pmdp;
 
 	if (pmd_write(pmd)) {
-		__pmd_idte(address, pmdp);
+		pmdp_flush_direct(mm, address, pmdp);
 		set_pmd_at(mm, address, pmdp, pmd_wrprotect(pmd));
 	}
 }

arch/s390/include/asm/setup.h

@@ -68,6 +68,7 @@ void create_mem_hole(struct mem_chunk mem_chunk[], unsigned long addr,
 #define MACHINE_FLAG_TOPOLOGY	(1UL << 14)
 #define MACHINE_FLAG_TE		(1UL << 15)
 #define MACHINE_FLAG_RRBM	(1UL << 16)
+#define MACHINE_FLAG_TLB_LC	(1UL << 17)
 
 #define MACHINE_IS_VM		(S390_lowcore.machine_flags & MACHINE_FLAG_VM)
 #define MACHINE_IS_KVM		(S390_lowcore.machine_flags & MACHINE_FLAG_KVM)
@@ -90,6 +91,7 @@ void create_mem_hole(struct mem_chunk mem_chunk[], unsigned long addr,
 #define MACHINE_HAS_TOPOLOGY	(0)
 #define MACHINE_HAS_TE		(0)
 #define MACHINE_HAS_RRBM	(0)
+#define MACHINE_HAS_TLB_LC	(0)
 #else /* CONFIG_64BIT */
 #define MACHINE_HAS_IEEE	(1)
 #define MACHINE_HAS_CSP		(1)
@@ -102,6 +104,7 @@ void create_mem_hole(struct mem_chunk mem_chunk[], unsigned long addr,
 #define MACHINE_HAS_TOPOLOGY	(S390_lowcore.machine_flags & MACHINE_FLAG_TOPOLOGY)
 #define MACHINE_HAS_TE		(S390_lowcore.machine_flags & MACHINE_FLAG_TE)
 #define MACHINE_HAS_RRBM	(S390_lowcore.machine_flags & MACHINE_FLAG_RRBM)
+#define MACHINE_HAS_TLB_LC	(S390_lowcore.machine_flags & MACHINE_FLAG_TLB_LC)
 #endif /* CONFIG_64BIT */
 
 /*

arch/s390/include/asm/switch_to.h

@@ -132,6 +132,7 @@ static inline void restore_access_regs(unsigned int *acrs)
 		update_cr_regs(next);					\
 	}								\
 	prev = __switch_to(prev,next);					\
+	update_primary_asce(current);					\
 } while (0)
 
 #define finish_arch_switch(prev) do {					     \

arch/s390/include/asm/thread_info.h

@@ -82,6 +82,7 @@ static inline struct thread_info *current_thread_info(void)
 #define TIF_SIGPENDING		2	/* signal pending */
 #define TIF_NEED_RESCHED	3	/* rescheduling necessary */
 #define TIF_TLB_WAIT		4	/* wait for TLB flush completion */
+#define TIF_ASCE		5	/* primary asce needs fixup / uaccess */
 #define TIF_PER_TRAP		6	/* deliver sigtrap on return to user */
 #define TIF_MCCK_PENDING	7	/* machine check handling is pending */
 #define TIF_SYSCALL_TRACE	8	/* syscall trace active */
@@ -99,6 +100,7 @@ static inline struct thread_info *current_thread_info(void)
 #define _TIF_SIGPENDING		(1<<TIF_SIGPENDING)
 #define _TIF_NEED_RESCHED	(1<<TIF_NEED_RESCHED)
 #define _TIF_TLB_WAIT		(1<<TIF_TLB_WAIT)
+#define _TIF_ASCE		(1<<TIF_ASCE)
 #define _TIF_PER_TRAP		(1<<TIF_PER_TRAP)
 #define _TIF_MCCK_PENDING	(1<<TIF_MCCK_PENDING)
 #define _TIF_SYSCALL_TRACE	(1<<TIF_SYSCALL_TRACE)

arch/s390/include/asm/tlb.h

@@ -57,8 +57,6 @@ static inline void tlb_gather_mmu(struct mmu_gather *tlb,
 	tlb->end = end;
 	tlb->fullmm = !(start | (end+1));
 	tlb->batch = NULL;
-	if (tlb->fullmm)
-		__tlb_flush_mm(mm);
 }
 
 static inline void tlb_flush_mmu(struct mmu_gather *tlb)
@@ -96,9 +94,7 @@ static inline void tlb_remove_page(struct mmu_gather *tlb, struct page *page)
 static inline void pte_free_tlb(struct mmu_gather *tlb, pgtable_t pte,
 				unsigned long address)
 {
-	if (!tlb->fullmm)
-		return page_table_free_rcu(tlb, (unsigned long *) pte);
-	page_table_free(tlb->mm, (unsigned long *) pte);
+	page_table_free_rcu(tlb, (unsigned long *) pte);
 }
 
 /*
@@ -114,9 +110,7 @@ static inline void pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd,
 #ifdef CONFIG_64BIT
 	if (tlb->mm->context.asce_limit <= (1UL << 31))
 		return;
-	if (!tlb->fullmm)
-		return tlb_remove_table(tlb, pmd);
-	crst_table_free(tlb->mm, (unsigned long *) pmd);
+	tlb_remove_table(tlb, pmd);
 #endif
 }
 
@@ -133,9 +127,7 @@ static inline void pud_free_tlb(struct mmu_gather *tlb, pud_t *pud,
 #ifdef CONFIG_64BIT
 	if (tlb->mm->context.asce_limit <= (1UL << 42))
 		return;
-	if (!tlb->fullmm)
-		return tlb_remove_table(tlb, pud);
-	crst_table_free(tlb->mm, (unsigned long *) pud);
+	tlb_remove_table(tlb, pud);
 #endif
 }
 

arch/s390/include/asm/tlbflush.h

@@ -7,19 +7,41 @@
 #include <asm/pgalloc.h>
 
 /*
- * Flush all tlb entries on the local cpu.
+ * Flush all TLB entries on the local CPU.
  */
 static inline void __tlb_flush_local(void)
 {
 	asm volatile("ptlb" : : : "memory");
 }
 
-#ifdef CONFIG_SMP
 /*
- * Flush all tlb entries on all cpus.
+ * Flush TLB entries for a specific ASCE on all CPUs
  */
+static inline void __tlb_flush_idte(unsigned long asce)
+{
+	/* Global TLB flush for the mm */
+	asm volatile(
+		"	.insn	rrf,0xb98e0000,0,%0,%1,0"
+		: : "a" (2048), "a" (asce) : "cc");
+}
+
+/*
+ * Flush TLB entries for a specific ASCE on the local CPU
+ */
+static inline void __tlb_flush_idte_local(unsigned long asce)
+{
+	/* Local TLB flush for the mm */
+	asm volatile(
+		"	.insn	rrf,0xb98e0000,0,%0,%1,1"
+		: : "a" (2048), "a" (asce) : "cc");
+}
+
+#ifdef CONFIG_SMP
 void smp_ptlb_all(void);
 
+/*
+ * Flush all TLB entries on all CPUs.
+ */
 static inline void __tlb_flush_global(void)
 {
 	register unsigned long reg2 asm("2");
@@ -42,36 +64,89 @@ static inline void __tlb_flush_global(void)
 		: : "d" (reg2), "d" (reg3), "d" (reg4), "m" (dummy) : "cc" );
 }
 
+/*
+ * Flush TLB entries for a specific mm on all CPUs (in case gmap is used
+ * this implicates multiple ASCEs!).
+ */
 static inline void __tlb_flush_full(struct mm_struct *mm)
 {
-	cpumask_t local_cpumask;
-
 	preempt_disable();
-	/*
-	 * If the process only ran on the local cpu, do a local flush.
-	 */
-	cpumask_copy(&local_cpumask, cpumask_of(smp_processor_id()));
-	if (cpumask_equal(mm_cpumask(mm), &local_cpumask))
+	atomic_add(0x10000, &mm->context.attach_count);
+	if (cpumask_equal(mm_cpumask(mm), cpumask_of(smp_processor_id()))) {
+		/* Local TLB flush */
 		__tlb_flush_local();
+	} else {
+		/* Global TLB flush */
+		__tlb_flush_global();
+		/* Reset TLB flush mask */
+		if (MACHINE_HAS_TLB_LC)
+			cpumask_copy(mm_cpumask(mm),
+				     &mm->context.cpu_attach_mask);
+	}
+	atomic_sub(0x10000, &mm->context.attach_count);
+	preempt_enable();
+}
+
+/*
+ * Flush TLB entries for a specific ASCE on all CPUs.
+ */
+static inline void __tlb_flush_asce(struct mm_struct *mm, unsigned long asce)
+{
+	int active, count;
+
+	preempt_disable();
+	active = (mm == current->active_mm) ? 1 : 0;
+	count = atomic_add_return(0x10000, &mm->context.attach_count);
+	if (MACHINE_HAS_TLB_LC && (count & 0xffff) <= active &&
+	    cpumask_equal(mm_cpumask(mm), cpumask_of(smp_processor_id()))) {
+		__tlb_flush_idte_local(asce);
+	} else {
+		if (MACHINE_HAS_IDTE)
+			__tlb_flush_idte(asce);
 		else
 			__tlb_flush_global();
+		/* Reset TLB flush mask */
+		if (MACHINE_HAS_TLB_LC)
+			cpumask_copy(mm_cpumask(mm),
+				     &mm->context.cpu_attach_mask);
+	}
+	atomic_sub(0x10000, &mm->context.attach_count);
 	preempt_enable();
 }
+
+static inline void __tlb_flush_kernel(void)
+{
+	if (MACHINE_HAS_IDTE)
+		__tlb_flush_idte((unsigned long) init_mm.pgd |
+				 init_mm.context.asce_bits);
+	else
+		__tlb_flush_global();
+}
 #else
-#define __tlb_flush_full(mm)	__tlb_flush_local()
 #define __tlb_flush_global()	__tlb_flush_local()
-#endif
+#define __tlb_flush_full(mm)	__tlb_flush_local()
 
 /*
- * Flush all tlb entries of a page table on all cpus.
+ * Flush TLB entries for a specific ASCE on all CPUs.
  */
-static inline void __tlb_flush_idte(unsigned long asce)
+static inline void __tlb_flush_asce(struct mm_struct *mm, unsigned long asce)
 {
-	asm volatile(
-		"	.insn	rrf,0xb98e0000,0,%0,%1,0"
-		: : "a" (2048), "a" (asce) : "cc" );
+	if (MACHINE_HAS_TLB_LC)
+		__tlb_flush_idte_local(asce);
+	else
+		__tlb_flush_local();
 }
 
+static inline void __tlb_flush_kernel(void)
+{
+	if (MACHINE_HAS_TLB_LC)
+		__tlb_flush_idte_local((unsigned long) init_mm.pgd |
+				       init_mm.context.asce_bits);
+	else
+		__tlb_flush_local();
+}
+#endif
+
 static inline void __tlb_flush_mm(struct mm_struct * mm)
 {
 	/*
@@ -80,7 +155,7 @@ static inline void __tlb_flush_mm(struct mm_struct * mm)
 	 * only ran on the local cpu.
 	 */
 	if (MACHINE_HAS_IDTE && list_empty(&mm->context.gmap_list))
-		__tlb_flush_idte((unsigned long) mm->pgd |
+		__tlb_flush_asce(mm, (unsigned long) mm->pgd |
 				 mm->context.asce_bits);
 	else
 		__tlb_flush_full(mm);
@@ -130,7 +205,7 @@ static inline void flush_tlb_range(struct vm_area_struct *vma,
 static inline void flush_tlb_kernel_range(unsigned long start,
 					  unsigned long end)
 {
-	__tlb_flush_mm(&init_mm);
+	__tlb_flush_kernel();
 }
 
 #endif /* _S390_TLBFLUSH_H */

arch/s390/include/asm/uaccess.h

@@ -92,8 +92,6 @@ static inline unsigned long extable_fixup(const struct exception_table_entry *x)
 #define ARCH_HAS_SORT_EXTABLE
 #define ARCH_HAS_SEARCH_EXTABLE
 
-int __handle_fault(unsigned long, unsigned long, int);
-
 /**
  * __copy_from_user: - Copy a block of data from user space, with less checking.
  * @to:   Destination address, in kernel space.

arch/s390/kernel/asm-offsets.c

@@ -136,6 +136,7 @@ int main(void)
 	DEFINE(__LC_RESTART_FN, offsetof(struct _lowcore, restart_fn));
 	DEFINE(__LC_RESTART_DATA, offsetof(struct _lowcore, restart_data));
 	DEFINE(__LC_RESTART_SOURCE, offsetof(struct _lowcore, restart_source));
+	DEFINE(__LC_KERNEL_ASCE, offsetof(struct _lowcore, kernel_asce));
 	DEFINE(__LC_USER_ASCE, offsetof(struct _lowcore, user_asce));
 	DEFINE(__LC_INT_CLOCK, offsetof(struct _lowcore, int_clock));
 	DEFINE(__LC_MCCK_CLOCK, offsetof(struct _lowcore, mcck_clock));

arch/s390/kernel/early.c

@@ -386,6 +386,8 @@ static __init void detect_machine_facilities(void)
 		S390_lowcore.machine_flags |= MACHINE_FLAG_TE;
 	if (test_facility(66))
 		S390_lowcore.machine_flags |= MACHINE_FLAG_RRBM;
+	if (test_facility(51))
+		S390_lowcore.machine_flags |= MACHINE_FLAG_TLB_LC;
 #endif
 }
 

arch/s390/kernel/entry.S

@@ -38,9 +38,9 @@ __PT_R14     =	__PT_GPRS + 56
 __PT_R15     =	__PT_GPRS + 60
 
 _TIF_WORK_SVC = (_TIF_SIGPENDING | _TIF_NOTIFY_RESUME | _TIF_NEED_RESCHED | \
-		 _TIF_MCCK_PENDING | _TIF_PER_TRAP )
+		 _TIF_MCCK_PENDING | _TIF_PER_TRAP | _TIF_ASCE)
 _TIF_WORK_INT = (_TIF_SIGPENDING | _TIF_NOTIFY_RESUME | _TIF_NEED_RESCHED | \
-		 _TIF_MCCK_PENDING)
+		 _TIF_MCCK_PENDING | _TIF_ASCE)
 _TIF_TRACE    = (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_AUDIT | _TIF_SECCOMP | \
 		 _TIF_SYSCALL_TRACEPOINT)
 _TIF_TRANSFER = (_TIF_MCCK_PENDING | _TIF_TLB_WAIT)
@@ -241,6 +241,8 @@ sysc_work:
 	jo	sysc_sigpending
 	tm	__TI_flags+3(%r12),_TIF_NOTIFY_RESUME
 	jo	sysc_notify_resume
+	tm	__TI_flags+3(%r12),_TIF_ASCE
+	jo	sysc_uaccess
 	j	sysc_return		# beware of critical section cleanup
 
 #
@@ -259,6 +261,14 @@ sysc_mcck_pending:
 	la	%r14,BASED(sysc_return)
 	br	%r1			# TIF bit will be cleared by handler
 
+#
+# _TIF_ASCE is set, load user space asce
+#
+sysc_uaccess:
+	ni	__TI_flags+3(%r12),255-_TIF_ASCE
+	lctl	%c1,%c1,__LC_USER_ASCE	# load primary asce
+	j	sysc_return
+
 #
 # _TIF_SIGPENDING is set, call do_signal
 #
@@ -522,6 +532,8 @@ io_work_tif:
 	jo	io_sigpending
 	tm	__TI_flags+3(%r12),_TIF_NOTIFY_RESUME
 	jo	io_notify_resume
+	tm	__TI_flags+3(%r12),_TIF_ASCE
+	jo	io_uaccess
 	j	io_return		# beware of critical section cleanup
 
 #
@@ -534,6 +546,14 @@ io_mcck_pending:
 	TRACE_IRQS_OFF
 	j	io_return
 
+#
+# _TIF_ASCE is set, load user space asce
+#
+io_uaccess:
+	ni	__TI_flags+3(%r12),255-_TIF_ASCE
+	lctl	%c1,%c1,__LC_USER_ASCE	# load primary asce
+	j	io_return
+
 #
 # _TIF_NEED_RESCHED is set, call schedule
 #

arch/s390/kernel/entry64.S

@@ -43,9 +43,9 @@ STACK_SIZE  = 1 << STACK_SHIFT
 STACK_INIT = STACK_SIZE - STACK_FRAME_OVERHEAD - __PT_SIZE
 
 _TIF_WORK_SVC = (_TIF_SIGPENDING | _TIF_NOTIFY_RESUME | _TIF_NEED_RESCHED | \
-		 _TIF_MCCK_PENDING | _TIF_PER_TRAP )
+		 _TIF_MCCK_PENDING | _TIF_PER_TRAP | _TIF_ASCE)
 _TIF_WORK_INT = (_TIF_SIGPENDING | _TIF_NOTIFY_RESUME | _TIF_NEED_RESCHED | \
-		 _TIF_MCCK_PENDING)
+		 _TIF_MCCK_PENDING | _TIF_ASCE)
 _TIF_TRACE    = (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_AUDIT | _TIF_SECCOMP | \
 		 _TIF_SYSCALL_TRACEPOINT)
 _TIF_TRANSFER = (_TIF_MCCK_PENDING | _TIF_TLB_WAIT)
@@ -275,6 +275,8 @@ sysc_work:
 	jo	sysc_sigpending
 	tm	__TI_flags+7(%r12),_TIF_NOTIFY_RESUME
 	jo	sysc_notify_resume
+	tm	__TI_flags+7(%r12),_TIF_ASCE
+	jo	sysc_uaccess
 	j	sysc_return		# beware of critical section cleanup
 
 #
@@ -291,6 +293,14 @@ sysc_mcck_pending:
 	larl	%r14,sysc_return
 	jg	s390_handle_mcck	# TIF bit will be cleared by handler
 
+#
+# _TIF_ASCE is set, load user space asce
+#
+sysc_uaccess:
+	ni	__TI_flags+7(%r12),255-_TIF_ASCE
+	lctlg	%c1,%c1,__LC_USER_ASCE		# load primary asce
+	j	sysc_return
+
 #
 # _TIF_SIGPENDING is set, call do_signal
 #
@@ -559,6 +569,8 @@ io_work_tif:
 	jo	io_sigpending
 	tm	__TI_flags+7(%r12),_TIF_NOTIFY_RESUME
 	jo	io_notify_resume
+	tm	__TI_flags+7(%r12),_TIF_ASCE
+	jo	io_uaccess
 	j	io_return		# beware of critical section cleanup
 
 #
@@ -570,6 +582,14 @@ io_mcck_pending:
 	TRACE_IRQS_OFF
 	j	io_return
 
+#
+# _TIF_ASCE is set, load user space asce
+#
+io_uaccess:
+	ni	__TI_flags+7(%r12),255-_TIF_ASCE
+	lctlg	%c1,%c1,__LC_USER_ASCE		# load primary asce
+	j	io_return
+
 #
 # _TIF_NEED_RESCHED is set, call schedule
 #

arch/s390/kernel/irq.c

@@ -207,7 +207,7 @@ static inline int ext_hash(u16 code)
 	return (code + (code >> 9)) & (ARRAY_SIZE(ext_int_hash) - 1);
 }
 
-int register_external_interrupt(u16 code, ext_int_handler_t handler)
+int register_external_irq(u16 code, ext_int_handler_t handler)
 {
 	struct ext_int_info *p;
 	unsigned long flags;
@@ -225,9 +225,9 @@ int register_external_interrupt(u16 code, ext_int_handler_t handler)
 	spin_unlock_irqrestore(&ext_int_hash_lock, flags);
 	return 0;
 }
-EXPORT_SYMBOL(register_external_interrupt);
+EXPORT_SYMBOL(register_external_irq);
 
-int unregister_external_interrupt(u16 code, ext_int_handler_t handler)
+int unregister_external_irq(u16 code, ext_int_handler_t handler)
 {
 	struct ext_int_info *p;
 	unsigned long flags;
@@ -243,7 +243,7 @@ int unregister_external_interrupt(u16 code, ext_int_handler_t handler)
 	spin_unlock_irqrestore(&ext_int_hash_lock, flags);
 	return 0;
 }
-EXPORT_SYMBOL(unregister_external_interrupt);
+EXPORT_SYMBOL(unregister_external_irq);
 
 static irqreturn_t do_ext_interrupt(int irq, void *dummy)
 {
@@ -253,7 +253,7 @@ static irqreturn_t do_ext_interrupt(int irq, void *dummy)
 	int index;
 
 	ext_code = *(struct ext_code *) &regs->int_code;
-	if (ext_code.code != 0x1004)
+	if (ext_code.code != EXT_IRQ_CLK_COMP)
 		__get_cpu_var(s390_idle).nohz_delay = 1;
 
 	index = ext_hash(ext_code.code);

arch/s390/kernel/perf_cpum_cf.c

@@ -673,7 +673,8 @@ static int __init cpumf_pmu_init(void)
 	ctl_clear_bit(0, 48);
 
 	/* register handler for measurement-alert interruptions */
-	rc = register_external_interrupt(0x1407, cpumf_measurement_alert);
+	rc = register_external_irq(EXT_IRQ_MEASURE_ALERT,
+				   cpumf_measurement_alert);
 	if (rc) {
 		pr_err("Registering for CPU-measurement alerts "
 		       "failed with rc=%i\n", rc);
@@ -684,7 +685,8 @@ static int __init cpumf_pmu_init(void)
 	rc = perf_pmu_register(&cpumf_pmu, "cpum_cf", PERF_TYPE_RAW);
 	if (rc) {
 		pr_err("Registering the cpum_cf PMU failed with rc=%i\n", rc);
-		unregister_external_interrupt(0x1407, cpumf_measurement_alert);
+		unregister_external_irq(EXT_IRQ_MEASURE_ALERT,
+					cpumf_measurement_alert);
 		goto out;
 	}
 	perf_cpu_notifier(cpumf_pmu_notifier);

arch/s390/kernel/perf_cpum_sf.c

@@ -1621,7 +1621,8 @@ static int __init init_cpum_sampling_pmu(void)
 		pr_err("Registering for s390dbf failed\n");
 	debug_register_view(sfdbg, &debug_sprintf_view);
 
-	err = register_external_interrupt(0x1407, cpumf_measurement_alert);
+	err = register_external_irq(EXT_IRQ_MEASURE_ALERT,
+				    cpumf_measurement_alert);
 	if (err) {
 		pr_cpumsf_err(RS_INIT_FAILURE_ALRT);
 		goto out;
@@ -1630,7 +1631,8 @@ static int __init init_cpum_sampling_pmu(void)
 	err = perf_pmu_register(&cpumf_sampling, "cpum_sf", PERF_TYPE_RAW);
 	if (err) {
 		pr_cpumsf_err(RS_INIT_FAILURE_PERF);
-		unregister_external_interrupt(0x1407, cpumf_measurement_alert);
+		unregister_external_irq(EXT_IRQ_MEASURE_ALERT,
+					cpumf_measurement_alert);
 		goto out;
 	}
 	perf_cpu_notifier(cpumf_pmu_notifier);

arch/s390/kernel/runtime_instr.c

@@ -138,7 +138,8 @@ static int __init runtime_instr_init(void)
 		return 0;
 
 	irq_subclass_register(IRQ_SUBCLASS_MEASUREMENT_ALERT);
-	rc = register_external_interrupt(0x1407, runtime_instr_int_handler);
+	rc = register_external_irq(EXT_IRQ_MEASURE_ALERT,
+				   runtime_instr_int_handler);
 	if (rc)
 		irq_subclass_unregister(IRQ_SUBCLASS_MEASUREMENT_ALERT);
 	else

arch/s390/kernel/sclp.S

@@ -9,6 +9,7 @@
  */
 
 #include <linux/linkage.h>
+#include <asm/irq.h>
 
 LC_EXT_NEW_PSW		= 0x58			# addr of ext int handler
 LC_EXT_NEW_PSW_64	= 0x1b0			# addr of ext int handler 64 bit
@@ -73,9 +74,9 @@ _sclp_wait_int:
 	lpsw	.LwaitpswS1-.LbaseS1(%r13)	# wait until interrupt
 .LwaitS1:
 	lh	%r7,LC_EXT_INT_CODE
-	chi	%r7,0x1004			# timeout?
+	chi	%r7,EXT_IRQ_CLK_COMP		# timeout?
 	je	.LtimeoutS1
-	chi	%r7,0x2401			# service int?
+	chi	%r7,EXT_IRQ_SERVICE_SIG		# service int?
 	jne	.LloopS1
 	sr	%r2,%r2
 	l	%r3,LC_EXT_INT_PARAM

arch/s390/kernel/smp.c

@@ -236,6 +236,9 @@ static void pcpu_prepare_secondary(struct pcpu *pcpu, int cpu)
 {
 	struct _lowcore *lc = pcpu->lowcore;
 
+	if (MACHINE_HAS_TLB_LC)
+		cpumask_set_cpu(cpu, &init_mm.context.cpu_attach_mask);
+	cpumask_set_cpu(cpu, mm_cpumask(&init_mm));
 	atomic_inc(&init_mm.context.attach_count);
 	lc->cpu_nr = cpu;
 	lc->percpu_offset = __per_cpu_offset[cpu];
@@ -760,6 +763,9 @@ void __cpu_die(unsigned int cpu)
 		cpu_relax();
 	pcpu_free_lowcore(pcpu);
 	atomic_dec(&init_mm.context.attach_count);
+	cpumask_clear_cpu(cpu, mm_cpumask(&init_mm));
+	if (MACHINE_HAS_TLB_LC)
+		cpumask_clear_cpu(cpu, &init_mm.context.cpu_attach_mask);
 }
 
 void __noreturn cpu_die(void)
@@ -785,10 +791,10 @@ void __init smp_fill_possible_mask(void)
 void __init smp_prepare_cpus(unsigned int max_cpus)
 {
 	/* request the 0x1201 emergency signal external interrupt */
-	if (register_external_interrupt(0x1201, do_ext_call_interrupt) != 0)
+	if (register_external_irq(EXT_IRQ_EMERGENCY_SIG, do_ext_call_interrupt))
 		panic("Couldn't request external interrupt 0x1201");
 	/* request the 0x1202 external call external interrupt */
-	if (register_external_interrupt(0x1202, do_ext_call_interrupt) != 0)
+	if (register_external_irq(EXT_IRQ_EXTERNAL_CALL, do_ext_call_interrupt))
 		panic("Couldn't request external interrupt 0x1202");
 	smp_detect_cpus();
 }

arch/s390/kernel/time.c

@@ -262,11 +262,11 @@ void __init time_init(void)
 	stp_reset();
 
 	/* request the clock comparator external interrupt */
-	if (register_external_interrupt(0x1004, clock_comparator_interrupt))
+	if (register_external_irq(EXT_IRQ_CLK_COMP, clock_comparator_interrupt))
 		panic("Couldn't request external interrupt 0x1004");
 
 	/* request the timing alert external interrupt */
-	if (register_external_interrupt(0x1406, timing_alert_interrupt))
+	if (register_external_irq(EXT_IRQ_TIMING_ALERT, timing_alert_interrupt))
 		panic("Couldn't request external interrupt 0x1406");
 
 	if (clocksource_register(&clocksource_tod) != 0)

arch/s390/kvm/diag.c

@@ -167,6 +167,10 @@ static int __diag_ipl_functions(struct kvm_vcpu *vcpu)
 
 	VCPU_EVENT(vcpu, 5, "diag ipl functions, subcode %lx", subcode);
 	switch (subcode) {
+	case 0:
+	case 1:
+		page_table_reset_pgste(current->mm, 0, TASK_SIZE);
+		return -EOPNOTSUPP;
 	case 3:
 		vcpu->run->s390_reset_flags = KVM_S390_RESET_CLEAR;
 		page_table_reset_pgste(current->mm, 0, TASK_SIZE);

arch/s390/lib/Makefile

@@ -2,7 +2,7 @@
 # Makefile for s390-specific library files..
 #
 
-lib-y += delay.o string.o uaccess_pt.o uaccess_mvcos.o find.o
+lib-y += delay.o string.o uaccess.o find.o
 obj-$(CONFIG_32BIT) += div64.o qrnnd.o ucmpdi2.o mem32.o
 obj-$(CONFIG_64BIT) += mem64.o
 lib-$(CONFIG_SMP) += spinlock.o

arch/s390/lib/uaccess.h

-/*
- *    Copyright IBM Corp. 2007
- *
- */
-
-#ifndef __ARCH_S390_LIB_UACCESS_H
-#define __ARCH_S390_LIB_UACCESS_H
-
-unsigned long copy_from_user_pt(void *to, const void __user *from, unsigned long n);
-unsigned long copy_to_user_pt(void __user *to, const void *from, unsigned long n);
-unsigned long copy_in_user_pt(void __user *to, const void __user *from, unsigned long n);
-unsigned long clear_user_pt(void __user *to, unsigned long n);
-unsigned long strnlen_user_pt(const char __user *src, unsigned long count);
-long strncpy_from_user_pt(char *dst, const char __user *src, long count);
-
-#endif /* __ARCH_S390_LIB_UACCESS_H */

arch/s390/mm/fault.c

@@ -106,21 +106,24 @@ void bust_spinlocks(int yes)
  * Returns the address space associated with the fault.
  * Returns 0 for kernel space and 1 for user space.
  */
-static inline int user_space_fault(unsigned long trans_exc_code)
+static inline int user_space_fault(struct pt_regs *regs)
 {
+	unsigned long trans_exc_code;
+
 	/*
 	 * The lowest two bits of the translation exception
 	 * identification indicate which paging table was used.
 	 */
-	trans_exc_code &= 3;
-	if (trans_exc_code == 2)
-		/* Access via secondary space, set_fs setting decides */
+	trans_exc_code = regs->int_parm_long & 3;
+	if (trans_exc_code == 3) /* home space -> kernel */
+		return 0;
+	if (user_mode(regs))
+		return 1;
+	if (trans_exc_code == 2) /* secondary space -> set_fs */
 		return current->thread.mm_segment.ar4;
-	/*
-	 * Access via primary space or access register is from user space
-	 * and access via home space is from the kernel.
-	 */
-	return trans_exc_code != 3;
+	if (current->flags & PF_VCPU)
+		return 1;
+	return 0;
 }
 
 static inline void report_user_fault(struct pt_regs *regs, long signr)
@@ -172,7 +175,7 @@ static noinline void do_no_context(struct pt_regs *regs)
 	 * terminate things with extreme prejudice.
 	 */
 	address = regs->int_parm_long & __FAIL_ADDR_MASK;
-	if (!user_space_fault(regs->int_parm_long))
+	if (!user_space_fault(regs))
 		printk(KERN_ALERT "Unable to handle kernel pointer dereference"
 		       " at virtual kernel address %p\n", (void *)address);
 	else
@@ -296,7 +299,7 @@ static inline int do_exception(struct pt_regs *regs, int access)
 	 * user context.
 	 */
 	fault = VM_FAULT_BADCONTEXT;
-	if (unlikely(!user_space_fault(trans_exc_code) || in_atomic() || !mm))
+	if (unlikely(!user_space_fault(regs) || in_atomic() || !mm))
 		goto out;
 
 	address = trans_exc_code & __FAIL_ADDR_MASK;
@@ -441,30 +444,6 @@ void __kprobes do_dat_exception(struct pt_regs *regs)
 		do_fault_error(regs, fault);
 }
 
-int __handle_fault(unsigned long uaddr, unsigned long pgm_int_code, int write)
-{
-	struct pt_regs regs;
-	int access, fault;
-
-	/* Emulate a uaccess fault from kernel mode. */
-	regs.psw.mask = PSW_KERNEL_BITS | PSW_MASK_DAT | PSW_MASK_MCHECK;
-	if (!irqs_disabled())
-		regs.psw.mask |= PSW_MASK_IO | PSW_MASK_EXT;
-	regs.psw.addr = (unsigned long) __builtin_return_address(0);
-	regs.psw.addr |= PSW_ADDR_AMODE;
-	regs.int_code = pgm_int_code;
-	regs.int_parm_long = (uaddr & PAGE_MASK) | 2;
-	access = write ? VM_WRITE : VM_READ;
-	fault = do_exception(&regs, access);
-	/*
-	 * Since the fault happened in kernel mode while performing a uaccess
-	 * all we need to do now is emulating a fixup in case "fault" is not
-	 * zero.
-	 * For the calling uaccess functions this results always in -EFAULT.
-	 */
-	return fault ? -EFAULT : 0;
-}
-
 #ifdef CONFIG_PFAULT 
 /*
  * 'pfault' pseudo page faults routines.
@@ -645,7 +624,7 @@ static int __init pfault_irq_init(void)
 {
 	int rc;
 
-	rc = register_external_interrupt(0x2603, pfault_interrupt);
+	rc = register_external_irq(EXT_IRQ_CP_SERVICE, pfault_interrupt);
 	if (rc)
 		goto out_extint;
 	rc = pfault_init() == 0 ? 0 : -EOPNOTSUPP;
@@ -656,7 +635,7 @@ static int __init pfault_irq_init(void)
 	return 0;
 
 out_pfault:
-	unregister_external_interrupt(0x2603, pfault_interrupt);
+	unregister_external_irq(EXT_IRQ_CP_SERVICE, pfault_interrupt);
 out_extint:
 	pfault_disable = 1;
 	return rc;

arch/s390/mm/hugetlbpage.c

@@ -123,10 +123,7 @@ pte_t huge_ptep_get_and_clear(struct mm_struct *mm,
 	pmd_t *pmdp = (pmd_t *) ptep;
 	pte_t pte = huge_ptep_get(ptep);
 
-	if (MACHINE_HAS_IDTE)
-		__pmd_idte(addr, pmdp);
-	else
-		__pmd_csp(pmdp);
+	pmdp_flush_direct(mm, addr, pmdp);
 	pmd_val(*pmdp) = _SEGMENT_ENTRY_EMPTY;
 	return pte;
 }

arch/s390/mm/init.c

@@ -124,8 +124,6 @@ void __init paging_init(void)
 	__ctl_load(S390_lowcore.kernel_asce, 13, 13);
 	arch_local_irq_restore(4UL << (BITS_PER_LONG - 8));
 
-	atomic_set(&init_mm.context.attach_count, 1);
-
 	sparse_memory_present_with_active_regions(MAX_NUMNODES);
 	sparse_init();
 	memset(max_zone_pfns, 0, sizeof(max_zone_pfns));
@@ -136,6 +134,11 @@ void __init paging_init(void)
 
 void __init mem_init(void)
 {
+	if (MACHINE_HAS_TLB_LC)
+		cpumask_set_cpu(0, &init_mm.context.cpu_attach_mask);
+	cpumask_set_cpu(0, mm_cpumask(&init_mm));
+	atomic_set(&init_mm.context.attach_count, 1);
+
         max_mapnr = max_low_pfn;
         high_memory = (void *) __va(max_low_pfn * PAGE_SIZE);
 

arch/s390/mm/pgtable.c

@@ -54,7 +54,7 @@ static void __crst_table_upgrade(void *arg)
 	struct mm_struct *mm = arg;
 
 	if (current->active_mm == mm)
-		update_mm(mm, current);
+		update_user_asce(mm, 1);
 	__tlb_flush_local();
 }
 
@@ -107,8 +107,10 @@ void crst_table_downgrade(struct mm_struct *mm, unsigned long limit)
 {
 	pgd_t *pgd;
 
-	if (current->active_mm == mm)
+	if (current->active_mm == mm) {
+		clear_user_asce(mm, 1);
 		__tlb_flush_mm(mm);
+	}
 	while (mm->context.asce_limit > limit) {
 		pgd = mm->pgd;
 		switch (pgd_val(*pgd) & _REGION_ENTRY_TYPE_MASK) {
@@ -132,7 +134,7 @@ void crst_table_downgrade(struct mm_struct *mm, unsigned long limit)
 		crst_table_free(mm, (unsigned long *) pgd);
 	}
 	if (current->active_mm == mm)
-		update_mm(mm, current);
+		update_user_asce(mm, 1);
 }
 #endif
 
@@ -198,7 +200,7 @@ static int gmap_unlink_segment(struct gmap *gmap, unsigned long *table)
 static void gmap_flush_tlb(struct gmap *gmap)
 {
 	if (MACHINE_HAS_IDTE)
-		__tlb_flush_idte((unsigned long) gmap->table |
+		__tlb_flush_asce(gmap->mm, (unsigned long) gmap->table |
 				 _ASCE_TYPE_REGION1);
 	else
 		__tlb_flush_global();
@@ -217,7 +219,7 @@ void gmap_free(struct gmap *gmap)
 
 	/* Flush tlb. */
 	if (MACHINE_HAS_IDTE)
-		__tlb_flush_idte((unsigned long) gmap->table |
+		__tlb_flush_asce(gmap->mm, (unsigned long) gmap->table |
 				 _ASCE_TYPE_REGION1);
 	else
 		__tlb_flush_global();

arch/s390/mm/vmem.c

@@ -138,7 +138,6 @@ static int vmem_add_mem(unsigned long start, unsigned long size, int ro)
 	}
 	ret = 0;
 out:
-	flush_tlb_kernel_range(start, end);
 	return ret;
 }
 
@@ -265,7 +264,6 @@ int __meminit vmemmap_populate(unsigned long start, unsigned long end, int node)
 	memset((void *)start, 0, end - start);
 	ret = 0;
 out:
-	flush_tlb_kernel_range(start, end);
 	return ret;
 }
 

arch/s390/oprofile/hwsampler.c

@@ -1033,7 +1033,7 @@ int hwsampler_setup(void)
 				max_sampler_rate = cb->qsi.max_sampl_rate;
 		}
 	}
-	register_external_interrupt(0x1407, hws_ext_handler);
+	register_external_irq(EXT_IRQ_MEASURE_ALERT, hws_ext_handler);
 
 	hws_state = HWS_DEALLOCATED;
 	rc = 0;
@@ -1068,7 +1068,7 @@ int hwsampler_shutdown(void)
 			hws_wq = NULL;
 		}
 
-		unregister_external_interrupt(0x1407, hws_ext_handler);
+		unregister_external_irq(EXT_IRQ_MEASURE_ALERT, hws_ext_handler);
 		hws_state = HWS_INIT;
 		rc = 0;
 	}

drivers/s390/block/dasd_diag.c

@@ -646,7 +646,7 @@ dasd_diag_init(void)
 	ASCEBC(dasd_diag_discipline.ebcname, 4);
 
 	irq_subclass_register(IRQ_SUBCLASS_SERVICE_SIGNAL);
-	register_external_interrupt(0x2603, dasd_ext_handler);
+	register_external_irq(EXT_IRQ_CP_SERVICE, dasd_ext_handler);
 	dasd_diag_discipline_pointer = &dasd_diag_discipline;
 	return 0;
 }
@@ -654,7 +654,7 @@ dasd_diag_init(void)
 static void __exit
 dasd_diag_cleanup(void)
 {
-	unregister_external_interrupt(0x2603, dasd_ext_handler);
+	unregister_external_irq(EXT_IRQ_CP_SERVICE, dasd_ext_handler);
 	irq_subclass_unregister(IRQ_SUBCLASS_SERVICE_SIGNAL);
 	dasd_diag_discipline_pointer = NULL;
 }

drivers/s390/char/raw3270.c

@@ -632,6 +632,8 @@ raw3270_reset_device_cb(struct raw3270_request *rq, void *data)
 		raw3270_size_device_done(rp);
 	} else
 		raw3270_writesf_readpart(rp);
+	memset(&rp->init_reset, 0, sizeof(rp->init_reset));
+	memset(&rp->init_data, 0, sizeof(rp->init_data));
 }
 
 static int
@@ -639,9 +641,10 @@ __raw3270_reset_device(struct raw3270 *rp)
 {
 	int rc;
 
+	/* Check if reset is already pending */
+	if (rp->init_reset.view)
+		return -EBUSY;
 	/* Store reset data stream to init_data/init_reset */
-	memset(&rp->init_reset, 0, sizeof(rp->init_reset));
-	memset(&rp->init_data, 0, sizeof(rp->init_data));
 	rp->init_data[0] = TW_KR;
 	rp->init_reset.ccw.cmd_code = TC_EWRITEA;
 	rp->init_reset.ccw.flags = CCW_FLAG_SLI;
@@ -850,7 +853,7 @@ raw3270_create_device(struct ccw_device *cdev)
 	char *ascebc;
 	int rc;
 
-	rp = kmalloc(sizeof(struct raw3270), GFP_KERNEL | GFP_DMA);
+	rp = kzalloc(sizeof(struct raw3270), GFP_KERNEL | GFP_DMA);
 	if (!rp)
 		return ERR_PTR(-ENOMEM);
 	ascebc = kmalloc(256, GFP_KERNEL);

drivers/s390/char/sclp.c

@@ -91,6 +91,9 @@ static struct sclp_req sclp_suspend_req;
 /* Timer for request retries. */
 static struct timer_list sclp_request_timer;
 
+/* Timer for queued requests. */
+static struct timer_list sclp_queue_timer;
+
 /* Internal state: is the driver initialized? */
 static volatile enum sclp_init_state_t {
 	sclp_init_state_uninitialized,
@@ -215,6 +218,76 @@ sclp_request_timeout(unsigned long data)
 	sclp_process_queue();
 }
 
+/*
+ * Returns the expire value in jiffies of the next pending request timeout,
+ * if any. Needs to be called with sclp_lock.
+ */
+static unsigned long __sclp_req_queue_find_next_timeout(void)
+{
+	unsigned long expires_next = 0;
+	struct sclp_req *req;
+
+	list_for_each_entry(req, &sclp_req_queue, list) {
+		if (!req->queue_expires)
+			continue;
+		if (!expires_next ||
+		   (time_before(req->queue_expires, expires_next)))
+				expires_next = req->queue_expires;
+	}
+	return expires_next;
+}
+
+/*
+ * Returns expired request, if any, and removes it from the list.
+ */
+static struct sclp_req *__sclp_req_queue_remove_expired_req(void)
+{
+	unsigned long flags, now;
+	struct sclp_req *req;
+
+	spin_lock_irqsave(&sclp_lock, flags);
+	now = jiffies;
+	/* Don't need list_for_each_safe because we break out after list_del */
+	list_for_each_entry(req, &sclp_req_queue, list) {
+		if (!req->queue_expires)
+			continue;
+		if (time_before_eq(req->queue_expires, now)) {
+			if (req->status == SCLP_REQ_QUEUED) {
+				req->status = SCLP_REQ_QUEUED_TIMEOUT;
+				list_del(&req->list);
+				goto out;
+			}
+		}
+	}
+	req = NULL;
+out:
+	spin_unlock_irqrestore(&sclp_lock, flags);
+	return req;
+}
+
+/*
+ * Timeout handler for queued requests. Removes request from list and
+ * invokes callback. This timer can be set per request in situations where
+ * waiting too long would be harmful to the system, e.g. during SE reboot.
+ */
+static void sclp_req_queue_timeout(unsigned long data)
+{
+	unsigned long flags, expires_next;
+	struct sclp_req *req;
+
+	do {
+		req = __sclp_req_queue_remove_expired_req();
+		if (req && req->callback)
+			req->callback(req, req->callback_data);
+	} while (req);
+
+	spin_lock_irqsave(&sclp_lock, flags);
+	expires_next = __sclp_req_queue_find_next_timeout();
+	if (expires_next)
+		mod_timer(&sclp_queue_timer, expires_next);
+	spin_unlock_irqrestore(&sclp_lock, flags);
+}
+
 /* Try to start a request. Return zero if the request was successfully
  * started or if it will be started at a later time. Return non-zero otherwise.
  * Called while sclp_lock is locked. */
@@ -317,6 +390,13 @@ sclp_add_request(struct sclp_req *req)
 	req->start_count = 0;
 	list_add_tail(&req->list, &sclp_req_queue);
 	rc = 0;
+	if (req->queue_timeout) {
+		req->queue_expires = jiffies + req->queue_timeout * HZ;
+		if (!timer_pending(&sclp_queue_timer) ||
+		    time_after(sclp_queue_timer.expires, req->queue_expires))
+			mod_timer(&sclp_queue_timer, req->queue_expires);
+	} else
+		req->queue_expires = 0;
 	/* Start if request is first in list */
 	if (sclp_running_state == sclp_running_state_idle &&
 	    req->list.prev == &sclp_req_queue) {
@@ -892,7 +972,7 @@ sclp_check_interface(void)
 
 	spin_lock_irqsave(&sclp_lock, flags);
 	/* Prepare init mask command */
-	rc = register_external_interrupt(0x2401, sclp_check_handler);
+	rc = register_external_irq(EXT_IRQ_SERVICE_SIG, sclp_check_handler);
 	if (rc) {
 		spin_unlock_irqrestore(&sclp_lock, flags);
 		return rc;
@@ -925,7 +1005,7 @@ sclp_check_interface(void)
 		} else
 			rc = -EBUSY;
 	}
-	unregister_external_interrupt(0x2401, sclp_check_handler);
+	unregister_external_irq(EXT_IRQ_SERVICE_SIG, sclp_check_handler);
 	spin_unlock_irqrestore(&sclp_lock, flags);
 	return rc;
 }
@@ -1113,6 +1193,8 @@ sclp_init(void)
 	INIT_LIST_HEAD(&sclp_reg_list);
 	list_add(&sclp_state_change_event.list, &sclp_reg_list);
 	init_timer(&sclp_request_timer);
+	init_timer(&sclp_queue_timer);
+	sclp_queue_timer.function = sclp_req_queue_timeout;
 	/* Check interface */
 	spin_unlock_irqrestore(&sclp_lock, flags);
 	rc = sclp_check_interface();
@@ -1124,7 +1206,7 @@ sclp_init(void)
 	if (rc)
 		goto fail_init_state_uninitialized;
 	/* Register interrupt handler */
-	rc = register_external_interrupt(0x2401, sclp_interrupt_handler);
+	rc = register_external_irq(EXT_IRQ_SERVICE_SIG, sclp_interrupt_handler);
 	if (rc)
 		goto fail_unregister_reboot_notifier;
 	sclp_init_state = sclp_init_state_initialized;

drivers/s390/char/sclp.h

@@ -133,6 +133,11 @@ struct sclp_req {
 	/* Callback that is called after reaching final status. */
 	void (*callback)(struct sclp_req *, void *data);
 	void *callback_data;
+	int queue_timeout;		/* request queue timeout (sec), set by
+					   caller of sclp_add_request(), if
+					   needed */
+	/* Internal fields */
+	unsigned long queue_expires;	/* request queue timeout (jiffies) */
 };
 
 #define SCLP_REQ_FILLED	  0x00	/* request is ready to be processed */
@@ -140,6 +145,9 @@ struct sclp_req {
 #define SCLP_REQ_RUNNING  0x02	/* request is currently running */
 #define SCLP_REQ_DONE	  0x03	/* request is completed successfully */
 #define SCLP_REQ_FAILED	  0x05	/* request is finally failed */
+#define SCLP_REQ_QUEUED_TIMEOUT 0x06	/* request on queue timed out */
+
+#define SCLP_QUEUE_INTERVAL 5	/* timeout interval for request queue */
 
 /* function pointers that a high level driver has to use for registration */
 /* of some routines it wants to be called from the low level driver */
@@ -173,6 +181,7 @@ int sclp_deactivate(void);
 int sclp_reactivate(void);
 int sclp_service_call(sclp_cmdw_t command, void *sccb);
 int sclp_sync_request(sclp_cmdw_t command, void *sccb);
+int sclp_sync_request_timeout(sclp_cmdw_t command, void *sccb, int timeout);
 
 int sclp_sdias_init(void);
 void sclp_sdias_exit(void);

drivers/s390/char/sclp_cmd.c

@@ -36,6 +36,11 @@ static void sclp_sync_callback(struct sclp_req *req, void *data)
 }
 
 int sclp_sync_request(sclp_cmdw_t cmd, void *sccb)
+{
+	return sclp_sync_request_timeout(cmd, sccb, 0);
+}
+
+int sclp_sync_request_timeout(sclp_cmdw_t cmd, void *sccb, int timeout)
 {
 	struct completion completion;
 	struct sclp_req *request;
@@ -44,6 +49,8 @@ int sclp_sync_request(sclp_cmdw_t cmd, void *sccb)
 	request = kzalloc(sizeof(*request), GFP_KERNEL);
 	if (!request)
 		return -ENOMEM;
+	if (timeout)
+		request->queue_timeout = timeout;
 	request->command = cmd;
 	request->sccb = sccb;
 	request->status = SCLP_REQ_FILLED;
@@ -110,7 +117,8 @@ int sclp_get_cpu_info(struct sclp_cpu_info *info)
 	if (!sccb)
 		return -ENOMEM;
 	sccb->header.length = sizeof(*sccb);
-	rc = sclp_sync_request(SCLP_CMDW_READ_CPU_INFO, sccb);
+	rc = sclp_sync_request_timeout(SCLP_CMDW_READ_CPU_INFO, sccb,
+				       SCLP_QUEUE_INTERVAL);
 	if (rc)
 		goto out;
 	if (sccb->header.response_code != 0x0010) {
@@ -144,7 +152,7 @@ static int do_cpu_configure(sclp_cmdw_t cmd)
 	if (!sccb)
 		return -ENOMEM;
 	sccb->header.length = sizeof(*sccb);
-	rc = sclp_sync_request(cmd, sccb);
+	rc = sclp_sync_request_timeout(cmd, sccb, SCLP_QUEUE_INTERVAL);
 	if (rc)
 		goto out;
 	switch (sccb->header.response_code) {
@@ -214,7 +222,7 @@ static int do_assign_storage(sclp_cmdw_t cmd, u16 rn)
 		return -ENOMEM;
 	sccb->header.length = PAGE_SIZE;
 	sccb->rn = rn;
-	rc = sclp_sync_request(cmd, sccb);
+	rc = sclp_sync_request_timeout(cmd, sccb, SCLP_QUEUE_INTERVAL);
 	if (rc)
 		goto out;
 	switch (sccb->header.response_code) {
@@ -269,7 +277,8 @@ static int sclp_attach_storage(u8 id)
 	if (!sccb)
 		return -ENOMEM;
 	sccb->header.length = PAGE_SIZE;
-	rc = sclp_sync_request(0x00080001 | id << 8, sccb);
+	rc = sclp_sync_request_timeout(0x00080001 | id << 8, sccb,
+				       SCLP_QUEUE_INTERVAL);
 	if (rc)
 		goto out;
 	switch (sccb->header.response_code) {

drivers/s390/char/tape_std.c

@@ -78,7 +78,8 @@ tape_std_assign(struct tape_device *device)
 
 	rc = tape_do_io_interruptible(device, request);
 
-	del_timer(&timeout);
+	del_timer_sync(&timeout);
+	destroy_timer_on_stack(&timeout);
 
 	if (rc != 0) {
 		DBF_EVENT(3, "%08x: assign failed - device might be busy\n",

drivers/s390/crypto/zcrypt_api.c

@@ -626,8 +626,8 @@ static long zcrypt_send_ep11_cprb(struct ep11_urb *xcrb)
 			return -ENOMEM;
 
 		if (copy_from_user(ep11_dev_list.targets,
-				   (struct ep11_target_dev *)xcrb->targets,
-				   xcrb->targets_num *
+				   (struct ep11_target_dev __force __user *)
+				   xcrb->targets, xcrb->targets_num *
 				   sizeof(struct ep11_target_dev)))
 			return -EFAULT;
 	}

drivers/s390/crypto/zcrypt_msgtype6.c

@@ -315,6 +315,10 @@ static int XCRB_msg_to_type6CPRB_msgX(struct zcrypt_device *zdev,
 	char *req_data = ap_msg->message + sizeof(struct type6_hdr) + rcblen;
 	char *function_code;
 
+	if (CEIL4(xcRB->request_control_blk_length) <
+			xcRB->request_control_blk_length)
+		return -EINVAL; /* overflow after alignment*/
+
 	/* length checks */
 	ap_msg->length = sizeof(struct type6_hdr) +
 		CEIL4(xcRB->request_control_blk_length) +
@@ -333,6 +337,10 @@ static int XCRB_msg_to_type6CPRB_msgX(struct zcrypt_device *zdev,
 		return -EINVAL;
 	}
 
+	if (CEIL4(xcRB->reply_control_blk_length) <
+			xcRB->reply_control_blk_length)
+		return -EINVAL; /* overflow after alignment*/
+
 	replylen = sizeof(struct type86_fmt2_msg) +
 		CEIL4(xcRB->reply_control_blk_length) +
 		xcRB->reply_data_length;
@@ -415,12 +423,18 @@ static int xcrb_msg_to_type6_ep11cprb_msgx(struct zcrypt_device *zdev,
 		unsigned int	dom_val;	/* domain id	   */
 	} __packed * payload_hdr;
 
+	if (CEIL4(xcRB->req_len) < xcRB->req_len)
+		return -EINVAL; /* overflow after alignment*/
+
 	/* length checks */
 	ap_msg->length = sizeof(struct type6_hdr) + xcRB->req_len;
 	if (CEIL4(xcRB->req_len) > MSGTYPE06_MAX_MSG_SIZE -
 				   (sizeof(struct type6_hdr)))
 		return -EINVAL;
 
+	if (CEIL4(xcRB->resp_len) < xcRB->resp_len)
+		return -EINVAL; /* overflow after alignment*/
+
 	if (CEIL4(xcRB->resp_len) > MSGTYPE06_MAX_MSG_SIZE -
 				    (sizeof(struct type86_fmt2_msg)))
 		return -EINVAL;
@@ -432,7 +446,7 @@ static int xcrb_msg_to_type6_ep11cprb_msgx(struct zcrypt_device *zdev,
 
 	/* Import CPRB data from the ioctl input parameter */
 	if (copy_from_user(&(msg->cprbx.cprb_len),
-			   (char *)xcRB->req, xcRB->req_len)) {
+			   (char __force __user *)xcRB->req, xcRB->req_len)) {
 		return -EFAULT;
 	}
 
@@ -645,7 +659,7 @@ static int convert_type86_ep11_xcrb(struct zcrypt_device *zdev,
 		return -EINVAL;
 
 	/* Copy response CPRB to user */
-	if (copy_to_user((char *)xcRB->resp,
+	if (copy_to_user((char __force __user *)xcRB->resp,
 			 data + msg->fmt2.offset1, msg->fmt2.count1))
 		return -EFAULT;
 	xcRB->resp_len = msg->fmt2.count1;

drivers/s390/kvm/kvm_virtio.c

@@ -477,7 +477,7 @@ static int __init kvm_devices_init(void)
 	INIT_WORK(&hotplug_work, hotplug_devices);
 
 	irq_subclass_register(IRQ_SUBCLASS_SERVICE_SIGNAL);
-	register_external_interrupt(0x2603, kvm_extint_handler);
+	register_external_irq(EXT_IRQ_CP_SERVICE, kvm_extint_handler);
 
 	scan_devices();
 	return 0;

drivers/s390/net/lcs.c

@@ -899,6 +899,7 @@ lcs_send_lancmd(struct lcs_card *card, struct lcs_buffer *buffer,
 	add_timer(&timer);
 	wait_event(reply->wait_q, reply->received);
 	del_timer_sync(&timer);
+	destroy_timer_on_stack(&timer);
 	LCS_DBF_TEXT_(4, trace, "rc:%d",reply->rc);
 	rc = reply->rc;
 	lcs_put_reply(reply);

net/iucv/iucv.c

@@ -2028,7 +2028,7 @@ static int __init iucv_init(void)
 	rc = iucv_query_maxconn();
 	if (rc)
 		goto out_ctl;
-	rc = register_external_interrupt(0x4000, iucv_external_interrupt);
+	rc = register_external_irq(EXT_IRQ_IUCV, iucv_external_interrupt);
 	if (rc)
 		goto out_ctl;
 	iucv_root = root_device_register("iucv");
@@ -2078,7 +2078,7 @@ static int __init iucv_init(void)
 
 	root_device_unregister(iucv_root);
 out_int:
-	unregister_external_interrupt(0x4000, iucv_external_interrupt);
+	unregister_external_irq(EXT_IRQ_IUCV, iucv_external_interrupt);
 out_ctl:
 	ctl_clear_bit(0, 1);
 out:
@@ -2109,7 +2109,7 @@ static void __exit iucv_exit(void)
 	cpu_notifier_register_done();
 	root_device_unregister(iucv_root);
 	bus_unregister(&iucv_bus);
-	unregister_external_interrupt(0x4000, iucv_external_interrupt);
+	unregister_external_irq(EXT_IRQ_IUCV, iucv_external_interrupt);
 }
 
 subsys_initcall(iucv_init);