Merge branch 'for-linus' of git://git390.marist.edu/pub/scm/linux-2.6

* 'for-linus' of git://git390.marist.edu/pub/scm/linux-2.6:
  [S390] mm: add ZONE_DMA to 31-bit config again
  [S390] mm: add page fault retry handling
  [S390] mm: handle kernel caused page fault oom situations
  [S390] delay: implement ndelay
  [S390] topology,sched: fix cpu_coregroup_mask/cpu_book_mask definitions
  [S390] hwsampler: allow cpu hotplug
  [S390] uaccess: turn __access_ok() into a define
  [S390] irq: merge irq.c and s390_ext.c
  [S390] irq: fix service signal external interrupt handling
  [S390] pfault: always enable service signal interrupt

20 files changed, 213 insertions, 183 deletions

diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig
index ff2d2371b2e9..9fab2aa9c2c8 100644
--- a/arch/s390/Kconfig
+++ b/arch/s390/Kconfig
@@ -2,7 +2,7 @@ config MMU
         def_bool y
 
 config ZONE_DMA
-        def_bool y if 64BIT
+        def_bool y
 
 config LOCKDEP_SUPPORT
         def_bool y
diff --git a/arch/s390/appldata/appldata_mem.c b/arch/s390/appldata/appldata_mem.c
index e43fe7537031..f7d3dc555bdb 100644
--- a/arch/s390/appldata/appldata_mem.c
+++ b/arch/s390/appldata/appldata_mem.c
@@ -92,9 +92,7 @@ static void appldata_get_mem_data(void *data)
         mem_data->pswpin     = ev[PSWPIN];
         mem_data->pswpout    = ev[PSWPOUT];
         mem_data->pgalloc    = ev[PGALLOC_NORMAL];
-#ifdef CONFIG_ZONE_DMA
         mem_data->pgalloc    += ev[PGALLOC_DMA];
-#endif
         mem_data->pgfault    = ev[PGFAULT];
         mem_data->pgmajfault = ev[PGMAJFAULT];
 
diff --git a/arch/s390/include/asm/delay.h b/arch/s390/include/asm/delay.h
index 8a096b83f51f..0e3b35f96be1 100644
--- a/arch/s390/include/asm/delay.h
+++ b/arch/s390/include/asm/delay.h
@@ -14,10 +14,12 @@
 #ifndef _S390_DELAY_H
 #define _S390_DELAY_H
 
-extern void __udelay(unsigned long long usecs);
-extern void udelay_simple(unsigned long long usecs);
-extern void __delay(unsigned long loops);
+void __ndelay(unsigned long long nsecs);
+void __udelay(unsigned long long usecs);
+void udelay_simple(unsigned long long usecs);
+void __delay(unsigned long loops);
 
+#define ndelay(n) __ndelay((unsigned long long) (n))
 #define udelay(n) __udelay((unsigned long long) (n))
 #define mdelay(n) __udelay((unsigned long long) (n) * 1000)
 
diff --git a/arch/s390/include/asm/irq.h b/arch/s390/include/asm/irq.h
index 1544b90bd6d6..ba7b01c726a3 100644
--- a/arch/s390/include/asm/irq.h
+++ b/arch/s390/include/asm/irq.h
@@ -2,6 +2,7 @@
 #define _ASM_IRQ_H
 
 #include <linux/hardirq.h>
+#include <linux/types.h>
 
 enum interruption_class {
         EXTERNAL_INTERRUPT,
@@ -31,4 +32,11 @@ enum interruption_class {
         NR_IRQS,
 };
 
+typedef void (*ext_int_handler_t)(unsigned int, 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);
+void service_subclass_irq_register(void);
+void service_subclass_irq_unregister(void);
+
 #endif /* _ASM_IRQ_H */
diff --git a/arch/s390/include/asm/s390_ext.h b/arch/s390/include/asm/s390_ext.h
deleted file mode 100644
index 080876d5f196..000000000000
--- a/arch/s390/include/asm/s390_ext.h
+++ /dev/null
@@ -1,17 +0,0 @@
-/*
- *    Copyright IBM Corp. 1999,2010
- *    Author(s): Holger Smolinski <Holger.Smolinski@de.ibm.com>,
- *               Martin Schwidefsky <schwidefsky@de.ibm.com>,
- */
-
-#ifndef _S390_EXTINT_H
-#define _S390_EXTINT_H
-
-#include <linux/types.h>
-
-typedef void (*ext_int_handler_t)(unsigned int, 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);
-
-#endif /* _S390_EXTINT_H */
diff --git a/arch/s390/include/asm/topology.h b/arch/s390/include/asm/topology.h
index c5338834ddbd..005d77d8ae2a 100644
--- a/arch/s390/include/asm/topology.h
+++ b/arch/s390/include/asm/topology.h
@@ -7,7 +7,7 @@
 extern unsigned char cpu_core_id[NR_CPUS];
 extern cpumask_t cpu_core_map[NR_CPUS];
 
-static inline const struct cpumask *cpu_coregroup_mask(unsigned int cpu)
+static inline const struct cpumask *cpu_coregroup_mask(int cpu)
 {
         return &cpu_core_map[cpu];
 }
@@ -21,7 +21,7 @@ static inline const struct cpumask *cpu_coregroup_mask(unsigned int cpu)
 extern unsigned char cpu_book_id[NR_CPUS];
 extern cpumask_t cpu_book_map[NR_CPUS];
 
-static inline const struct cpumask *cpu_book_mask(unsigned int cpu)
+static inline const struct cpumask *cpu_book_mask(int cpu)
 {
         return &cpu_book_map[cpu];
 }
diff --git a/arch/s390/include/asm/uaccess.h b/arch/s390/include/asm/uaccess.h
index 2d9ea11f919a..2b23885e81e9 100644
--- a/arch/s390/include/asm/uaccess.h
+++ b/arch/s390/include/asm/uaccess.h
@@ -49,12 +49,13 @@
 
 #define segment_eq(a,b) ((a).ar4 == (b).ar4)
 
+#define __access_ok(addr, size) \
+({                              \
+        __chk_user_ptr(addr);   \
+        1;                      \
+})
 
-static inline int __access_ok(const void __user *addr, unsigned long size)
-{
-        return 1;
-}
-#define access_ok(type,addr,size) __access_ok(addr,size)
+#define access_ok(type, addr, size) __access_ok(addr, size)
 
 /*
  * The exception table consists of pairs of addresses: the first is the
diff --git a/arch/s390/kernel/Makefile b/arch/s390/kernel/Makefile
index 5ff15dacb571..df3732249baa 100644
--- a/arch/s390/kernel/Makefile
+++ b/arch/s390/kernel/Makefile
@@ -20,10 +20,10 @@ CFLAGS_ptrace.o		+= -DUTS_MACHINE='"$(UTS_MACHINE)"'
 
 CFLAGS_sysinfo.o += -Iinclude/math-emu -Iarch/s390/math-emu -w
 
-obj-y   :=  bitmap.o traps.o time.o process.o base.o early.o setup.o \
-            processor.o sys_s390.o ptrace.o signal.o cpcmd.o ebcdic.o \
-            s390_ext.o debug.o irq.o ipl.o dis.o diag.o mem_detect.o \
-            vdso.o vtime.o sysinfo.o nmi.o sclp.o jump_label.o
+obj-y   :=  bitmap.o traps.o time.o process.o base.o early.o setup.o vtime.o \
+            processor.o sys_s390.o ptrace.o signal.o cpcmd.o ebcdic.o nmi.o \
+            debug.o irq.o ipl.o dis.o diag.o mem_detect.o sclp.o vdso.o \
+            sysinfo.o jump_label.o
 
 obj-y   += $(if $(CONFIG_64BIT),entry64.o,entry.o)
 obj-y   += $(if $(CONFIG_64BIT),reipl64.o,reipl.o)
diff --git a/arch/s390/kernel/dis.c b/arch/s390/kernel/dis.c
index 3d4a78fc1adc..1ca3d1d6a86c 100644
--- a/arch/s390/kernel/dis.c
+++ b/arch/s390/kernel/dis.c
@@ -30,9 +30,9 @@
 #include <asm/atomic.h>
 #include <asm/mathemu.h>
 #include <asm/cpcmd.h>
-#include <asm/s390_ext.h>
 #include <asm/lowcore.h>
 #include <asm/debug.h>
+#include <asm/irq.h>
 
 #ifndef CONFIG_64BIT
 #define ONELONG "%08lx: "
diff --git a/arch/s390/kernel/irq.c b/arch/s390/kernel/irq.c
index e204f9597aaf..e3264f6a9720 100644
--- a/arch/s390/kernel/irq.c
+++ b/arch/s390/kernel/irq.c
@@ -1,19 +1,28 @@
 /*
- *    Copyright IBM Corp. 2004,2010
- *    Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com),
- *               Thomas Spatzier (tspat@de.ibm.com)
+ *    Copyright IBM Corp. 2004,2011
+ *    Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>,
+ *               Holger Smolinski <Holger.Smolinski@de.ibm.com>,
+ *               Thomas Spatzier <tspat@de.ibm.com>,
  *
  * This file contains interrupt related functions.
  */
 
-#include <linux/module.h>
-#include <linux/kernel.h>
 #include <linux/kernel_stat.h>
 #include <linux/interrupt.h>
 #include <linux/seq_file.h>
-#include <linux/cpu.h>
 #include <linux/proc_fs.h>
 #include <linux/profile.h>
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/ftrace.h>
+#include <linux/errno.h>
+#include <linux/slab.h>
+#include <linux/cpu.h>
+#include <asm/irq_regs.h>
+#include <asm/cputime.h>
+#include <asm/lowcore.h>
+#include <asm/irq.h>
+#include "entry.h"
 
 struct irq_class {
         char *name;
@@ -82,8 +91,7 @@ int show_interrupts(struct seq_file *p, void *v)
  * For compatibilty only. S/390 specific setup of interrupts et al. is done
  * much later in init_channel_subsystem().
  */
-void __init
-init_IRQ(void)
+void __init init_IRQ(void)
 {
         /* nothing... */
 }
@@ -134,3 +142,116 @@ void init_irq_proc(void)
         create_prof_cpu_mask(root_irq_dir);
 }
 #endif
+
+/*
+ * ext_int_hash[index] is the start of the list for all external interrupts
+ * that hash to this index. With the current set of external interrupts
+ * (0x1202 external call, 0x1004 cpu timer, 0x2401 hwc console, 0x4000
+ * iucv and 0x2603 pfault) this is always the first element.
+ */
+
+struct ext_int_info {
+        struct ext_int_info *next;
+        ext_int_handler_t handler;
+        u16 code;
+};
+
+static struct ext_int_info *ext_int_hash[256];
+
+static inline int ext_hash(u16 code)
+{
+        return (code + (code >> 9)) & 0xff;
+}
+
+int register_external_interrupt(u16 code, ext_int_handler_t handler)
+{
+        struct ext_int_info *p;
+        int index;
+
+        p = kmalloc(sizeof(*p), GFP_ATOMIC);
+        if (!p)
+                return -ENOMEM;
+        p->code = code;
+        p->handler = handler;
+        index = ext_hash(code);
+        p->next = ext_int_hash[index];
+        ext_int_hash[index] = p;
+        return 0;
+}
+EXPORT_SYMBOL(register_external_interrupt);
+
+int unregister_external_interrupt(u16 code, ext_int_handler_t handler)
+{
+        struct ext_int_info *p, *q;
+        int index;
+
+        index = ext_hash(code);
+        q = NULL;
+        p = ext_int_hash[index];
+        while (p) {
+                if (p->code == code && p->handler == handler)
+                        break;
+                q = p;
+                p = p->next;
+        }
+        if (!p)
+                return -ENOENT;
+        if (q)
+                q->next = p->next;
+        else
+                ext_int_hash[index] = p->next;
+        kfree(p);
+        return 0;
+}
+EXPORT_SYMBOL(unregister_external_interrupt);
+
+void __irq_entry do_extint(struct pt_regs *regs, unsigned int ext_int_code,
+                           unsigned int param32, unsigned long param64)
+{
+        struct pt_regs *old_regs;
+        unsigned short code;
+        struct ext_int_info *p;
+        int index;
+
+        code = (unsigned short) ext_int_code;
+        old_regs = set_irq_regs(regs);
+        s390_idle_check(regs, S390_lowcore.int_clock,
+                        S390_lowcore.async_enter_timer);
+        irq_enter();
+        if (S390_lowcore.int_clock >= S390_lowcore.clock_comparator)
+                /* Serve timer interrupts first. */
+                clock_comparator_work();
+        kstat_cpu(smp_processor_id()).irqs[EXTERNAL_INTERRUPT]++;
+        if (code != 0x1004)
+                __get_cpu_var(s390_idle).nohz_delay = 1;
+        index = ext_hash(code);
+        for (p = ext_int_hash[index]; p; p = p->next) {
+                if (likely(p->code == code))
+                        p->handler(ext_int_code, param32, param64);
+        }
+        irq_exit();
+        set_irq_regs(old_regs);
+}
+
+static DEFINE_SPINLOCK(sc_irq_lock);
+static int sc_irq_refcount;
+
+void service_subclass_irq_register(void)
+{
+        spin_lock(&sc_irq_lock);
+        if (!sc_irq_refcount)
+                ctl_set_bit(0, 9);
+        sc_irq_refcount++;
+        spin_unlock(&sc_irq_lock);
+}
+EXPORT_SYMBOL(service_subclass_irq_register);
+
+void service_subclass_irq_unregister(void)
+{
+        spin_lock(&sc_irq_lock);
+        sc_irq_refcount--;
+        if (!sc_irq_refcount)
+                ctl_clear_bit(0, 9);
+        spin_unlock(&sc_irq_lock);
+}
+EXPORT_SYMBOL(service_subclass_irq_unregister);
diff --git a/arch/s390/kernel/s390_ext.c b/arch/s390/kernel/s390_ext.c
deleted file mode 100644
index 185029919c4d..000000000000
--- a/arch/s390/kernel/s390_ext.c
+++ /dev/null
@@ -1,108 +0,0 @@
-/*
- *    Copyright IBM Corp. 1999,2010
- *    Author(s): Holger Smolinski <Holger.Smolinski@de.ibm.com>,
- *               Martin Schwidefsky <schwidefsky@de.ibm.com>,
- */
-
-#include <linux/kernel_stat.h>
-#include <linux/interrupt.h>
-#include <linux/module.h>
-#include <linux/kernel.h>
-#include <linux/ftrace.h>
-#include <linux/errno.h>
-#include <linux/slab.h>
-#include <asm/s390_ext.h>
-#include <asm/irq_regs.h>
-#include <asm/cputime.h>
-#include <asm/lowcore.h>
-#include <asm/irq.h>
-#include "entry.h"
-
-struct ext_int_info {
-        struct ext_int_info *next;
-        ext_int_handler_t handler;
-        __u16 code;
-};
-
-/*
- * ext_int_hash[index] is the start of the list for all external interrupts
- * that hash to this index. With the current set of external interrupts 
- * (0x1202 external call, 0x1004 cpu timer, 0x2401 hwc console, 0x4000
- * iucv and 0x2603 pfault) this is always the first element. 
- */
-static struct ext_int_info *ext_int_hash[256];
-
-static inline int ext_hash(__u16 code)
-{
-        return (code + (code >> 9)) & 0xff;
-}
-
-int register_external_interrupt(__u16 code, ext_int_handler_t handler)
-{
-        struct ext_int_info *p;
-        int index;
-
-        p = kmalloc(sizeof(*p), GFP_ATOMIC);
-        if (!p)
-                return -ENOMEM;
-        p->code = code;
-        p->handler = handler;
-        index = ext_hash(code);
-        p->next = ext_int_hash[index];
-        ext_int_hash[index] = p;
-        return 0;
-}
-EXPORT_SYMBOL(register_external_interrupt);
-
-int unregister_external_interrupt(__u16 code, ext_int_handler_t handler)
-{
-        struct ext_int_info *p, *q;
-        int index;
-
-        index = ext_hash(code);
-        q = NULL;
-        p = ext_int_hash[index];
-        while (p) {
-                if (p->code == code && p->handler == handler)
-                        break;
-                q = p;
-                p = p->next;
-        }
-        if (!p)
-                return -ENOENT;
-        if (q)
-                q->next = p->next;
-        else
-                ext_int_hash[index] = p->next;
-        kfree(p);
-        return 0;
-}
-EXPORT_SYMBOL(unregister_external_interrupt);
-
-void __irq_entry do_extint(struct pt_regs *regs, unsigned int ext_int_code,
-                           unsigned int param32, unsigned long param64)
-{
-        struct pt_regs *old_regs;
-        unsigned short code;
-        struct ext_int_info *p;
-        int index;
-
-        code = (unsigned short) ext_int_code;
-        old_regs = set_irq_regs(regs);
-        s390_idle_check(regs, S390_lowcore.int_clock,
-                        S390_lowcore.async_enter_timer);
-        irq_enter();
-        if (S390_lowcore.int_clock >= S390_lowcore.clock_comparator)
-                /* Serve timer interrupts first. */
-                clock_comparator_work();
-        kstat_cpu(smp_processor_id()).irqs[EXTERNAL_INTERRUPT]++;
-        if (code != 0x1004)
-                __get_cpu_var(s390_idle).nohz_delay = 1;
-        index = ext_hash(code);
-        for (p = ext_int_hash[index]; p; p = p->next) {
-                if (likely(p->code == code))
-                        p->handler(ext_int_code, param32, param64);
-        }
-        irq_exit();
-        set_irq_regs(old_regs);
-}
diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c
index f8e85ecbc459..52420d2785b3 100644
--- a/arch/s390/kernel/smp.c
+++ b/arch/s390/kernel/smp.c
@@ -44,7 +44,6 @@
 #include <asm/sigp.h>
 #include <asm/pgalloc.h>
 #include <asm/irq.h>
-#include <asm/s390_ext.h>
 #include <asm/cpcmd.h>
 #include <asm/tlbflush.h>
 #include <asm/timer.h>
diff --git a/arch/s390/kernel/time.c b/arch/s390/kernel/time.c
index a59557f1fb5f..dff933065ab6 100644
--- a/arch/s390/kernel/time.c
+++ b/arch/s390/kernel/time.c
@@ -41,7 +41,6 @@
 #include <linux/kprobes.h>
 #include <asm/uaccess.h>
 #include <asm/delay.h>
-#include <asm/s390_ext.h>
 #include <asm/div64.h>
 #include <asm/vdso.h>
 #include <asm/irq.h>
diff --git a/arch/s390/kernel/topology.c b/arch/s390/kernel/topology.c
index 2eafb8c7a746..0cd340b72632 100644
--- a/arch/s390/kernel/topology.c
+++ b/arch/s390/kernel/topology.c
@@ -17,7 +17,6 @@
 #include <linux/smp.h>
 #include <linux/cpuset.h>
 #include <asm/delay.h>
-#include <asm/s390_ext.h>
 
 #define PTF_HORIZONTAL  (0UL)
 #define PTF_VERTICAL    (1UL)
diff --git a/arch/s390/kernel/traps.c b/arch/s390/kernel/traps.c
index b5a4a739b477..a65d2e82f61d 100644
--- a/arch/s390/kernel/traps.c
+++ b/arch/s390/kernel/traps.c
@@ -39,7 +39,6 @@
 #include <asm/atomic.h>
 #include <asm/mathemu.h>
 #include <asm/cpcmd.h>
-#include <asm/s390_ext.h>
 #include <asm/lowcore.h>
 #include <asm/debug.h>
 #include "entry.h"
diff --git a/arch/s390/kernel/vtime.c b/arch/s390/kernel/vtime.c
index 5e8ead4b4aba..2d6228f60cd6 100644
--- a/arch/s390/kernel/vtime.c
+++ b/arch/s390/kernel/vtime.c
@@ -22,10 +22,10 @@
 #include <linux/cpu.h>
 #include <linux/kprobes.h>
 
-#include <asm/s390_ext.h>
 #include <asm/timer.h>
 #include <asm/irq_regs.h>
 #include <asm/cputime.h>
+#include <asm/irq.h>
 
 static DEFINE_PER_CPU(struct vtimer_queue, virt_cpu_timer);
 
diff --git a/arch/s390/lib/delay.c b/arch/s390/lib/delay.c
index 0f53110e1d09..a65229d91c92 100644
--- a/arch/s390/lib/delay.c
+++ b/arch/s390/lib/delay.c
@@ -12,6 +12,7 @@
 #include <linux/module.h>
 #include <linux/irqflags.h>
 #include <linux/interrupt.h>
+#include <asm/div64.h>
 
 void __delay(unsigned long loops)
 {
@@ -116,3 +117,17 @@ void udelay_simple(unsigned long long usecs)
         while (get_clock() < end)
                 cpu_relax();
 }
+
+void __ndelay(unsigned long long nsecs)
+{
+        u64 end;
+
+        nsecs <<= 9;
+        do_div(nsecs, 125);
+        end = get_clock() + nsecs;
+        if (nsecs & ~0xfffUL)
+                __udelay(nsecs >> 12);
+        while (get_clock() < end)
+                barrier();
+}
+EXPORT_SYMBOL(__ndelay);
diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c
index a0f9e730f26a..fe103e891e7a 100644
--- a/arch/s390/mm/fault.c
+++ b/arch/s390/mm/fault.c
@@ -34,7 +34,7 @@
 #include <asm/asm-offsets.h>
 #include <asm/system.h>
 #include <asm/pgtable.h>
-#include <asm/s390_ext.h>
+#include <asm/irq.h>
 #include <asm/mmu_context.h>
 #include <asm/compat.h>
 #include "../kernel/entry.h"
@@ -245,9 +245,12 @@ static noinline void do_fault_error(struct pt_regs *regs, long int_code,
                 do_no_context(regs, int_code, trans_exc_code);
                 break;
         default: /* fault & VM_FAULT_ERROR */
-                if (fault & VM_FAULT_OOM)
-                        pagefault_out_of_memory();
-                else if (fault & VM_FAULT_SIGBUS) {
+                if (fault & VM_FAULT_OOM) {
+                        if (!(regs->psw.mask & PSW_MASK_PSTATE))
+                                do_no_context(regs, int_code, trans_exc_code);
+                        else
+                                pagefault_out_of_memory();
+                } else if (fault & VM_FAULT_SIGBUS) {
                         /* Kernel mode? Handle exceptions or die */
                         if (!(regs->psw.mask & PSW_MASK_PSTATE))
                                 do_no_context(regs, int_code, trans_exc_code);
@@ -277,7 +280,8 @@ static inline int do_exception(struct pt_regs *regs, int access,
         struct mm_struct *mm;
         struct vm_area_struct *vma;
         unsigned long address;
-        int fault, write;
+        unsigned int flags;
+        int fault;
 
         if (notify_page_fault(regs))
                 return 0;
@@ -296,6 +300,10 @@ static inline int do_exception(struct pt_regs *regs, int access,
 
         address = trans_exc_code & __FAIL_ADDR_MASK;
         perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address);
+        flags = FAULT_FLAG_ALLOW_RETRY;
+        if (access == VM_WRITE || (trans_exc_code & store_indication) == 0x400)
+                flags |= FAULT_FLAG_WRITE;
+retry:
         down_read(&mm->mmap_sem);
 
         fault = VM_FAULT_BADMAP;
@@ -325,21 +333,31 @@ static inline int do_exception(struct pt_regs *regs, int access,
          * make sure we exit gracefully rather than endlessly redo
          * the fault.
          */
-        write = (access == VM_WRITE ||
-                 (trans_exc_code & store_indication) == 0x400) ?
-                FAULT_FLAG_WRITE : 0;
-        fault = handle_mm_fault(mm, vma, address, write);
+        fault = handle_mm_fault(mm, vma, address, flags);
         if (unlikely(fault & VM_FAULT_ERROR))
                 goto out_up;
 
-        if (fault & VM_FAULT_MAJOR) {
-                tsk->maj_flt++;
-                perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0,
-                                     regs, address);
-        } else {
-                tsk->min_flt++;
-                perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0,
-                                     regs, address);
+        /*
+         * Major/minor page fault accounting is only done on the
+         * initial attempt. If we go through a retry, it is extremely
+         * likely that the page will be found in page cache at that point.
+         */
+        if (flags & FAULT_FLAG_ALLOW_RETRY) {
+                if (fault & VM_FAULT_MAJOR) {
+                        tsk->maj_flt++;
+                        perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0,
+                                      regs, address);
+                } else {
+                        tsk->min_flt++;
+                        perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0,
+                                      regs, address);
+                }
+                if (fault & VM_FAULT_RETRY) {
+                        /* Clear FAULT_FLAG_ALLOW_RETRY to avoid any risk
+                         * of starvation. */
+                        flags &= ~FAULT_FLAG_ALLOW_RETRY;
+                        goto retry;
+                }
         }
         /*
          * The instruction that caused the program check will
@@ -429,10 +447,9 @@ int __handle_fault(unsigned long uaddr, unsigned long pgm_int_code, int write)
         access = write ? VM_WRITE : VM_READ;
         fault = do_exception(&regs, access, uaddr | 2);
         if (unlikely(fault)) {
-                if (fault & VM_FAULT_OOM) {
-                        pagefault_out_of_memory();
-                        fault = 0;
-                } else if (fault & VM_FAULT_SIGBUS)
+                if (fault & VM_FAULT_OOM)
+                        return -EFAULT;
+                else if (fault & VM_FAULT_SIGBUS)
                         do_sigbus(&regs, pgm_int_code, uaddr);
         }
         return fault ? -EFAULT : 0;
@@ -485,7 +502,6 @@ int pfault_init(void)
                 "2:\n"
                 EX_TABLE(0b,1b)
                 : "=d" (rc) : "a" (&refbk), "m" (refbk) : "cc");
-        __ctl_set_bit(0, 9);
         return rc;
 }
 
@@ -500,7 +516,6 @@ void pfault_fini(void)
 
         if (!MACHINE_IS_VM || pfault_disable)
                 return;
-        __ctl_clear_bit(0,9);
         asm volatile(
                 "       diag    %0,0,0x258\n"
                 "0:\n"
@@ -615,6 +630,7 @@ static int __init pfault_irq_init(void)
         rc = pfault_init() == 0 ? 0 : -EOPNOTSUPP;
         if (rc)
                 goto out_pfault;
+        service_subclass_irq_register();
         hotcpu_notifier(pfault_cpu_notify, 0);
         return 0;
 
diff --git a/arch/s390/mm/init.c b/arch/s390/mm/init.c
index dfefc2171691..59b663109d90 100644
--- a/arch/s390/mm/init.c
+++ b/arch/s390/mm/init.c
@@ -119,9 +119,7 @@ void __init paging_init(void)
         sparse_memory_present_with_active_regions(MAX_NUMNODES);
         sparse_init();
         memset(max_zone_pfns, 0, sizeof(max_zone_pfns));
-#ifdef CONFIG_ZONE_DMA
         max_zone_pfns[ZONE_DMA] = PFN_DOWN(MAX_DMA_ADDRESS);
-#endif
         max_zone_pfns[ZONE_NORMAL] = max_low_pfn;
         free_area_init_nodes(max_zone_pfns);
         fault_init();
diff --git a/arch/s390/oprofile/hwsampler.c b/arch/s390/oprofile/hwsampler.c
index 053caa0fd276..4552ce40c81a 100644
--- a/arch/s390/oprofile/hwsampler.c
+++ b/arch/s390/oprofile/hwsampler.c
@@ -19,7 +19,7 @@
 #include <linux/oprofile.h>
 
 #include <asm/lowcore.h>
-#include <asm/s390_ext.h>
+#include <asm/irq.h>
 
 #include "hwsampler.h"
 
@@ -580,7 +580,7 @@ static int hws_cpu_callback(struct notifier_block *nfb,
 {
         /* We do not have sampler space available for all possible CPUs.
            All CPUs should be online when hw sampling is activated. */
-        return NOTIFY_BAD;
+        return (hws_state <= HWS_DEALLOCATED) ? NOTIFY_OK : NOTIFY_BAD;
 }
 
 static struct notifier_block hws_cpu_notifier = {