Merge commit 'v2.6.33-rc5' into next

18 files changed, 93 insertions, 107 deletions

diff --git a/arch/sparc/kernel/entry.S b/arch/sparc/kernel/entry.S
index ec9c7bc67d21..1504df8ddf70 100644
--- a/arch/sparc/kernel/entry.S
+++ b/arch/sparc/kernel/entry.S
@@ -1294,7 +1294,7 @@ linux_sparc_syscall:
         sethi   %hi(PSR_SYSCALL), %l4
         or      %l0, %l4, %l0
         /* Direct access to user regs, must faster. */
-        cmp     %g1, NR_SYSCALLS
+        cmp     %g1, NR_syscalls
         bgeu    linux_sparc_ni_syscall
          sll    %g1, 2, %l4
         ld      [%l7 + %l4], %l7
diff --git a/arch/sparc/kernel/ftrace.c b/arch/sparc/kernel/ftrace.c
index d3b1a3076569..29973daa9930 100644
--- a/arch/sparc/kernel/ftrace.c
+++ b/arch/sparc/kernel/ftrace.c
@@ -4,6 +4,7 @@
 #include <linux/percpu.h>
 #include <linux/init.h>
 #include <linux/list.h>
+#include <trace/syscall.h>
 
 #include <asm/ftrace.h>
 
@@ -91,3 +92,13 @@ int __init ftrace_dyn_arch_init(void *data)
 }
 #endif
 
+#ifdef CONFIG_FTRACE_SYSCALLS
+
+extern unsigned int sys_call_table[];
+
+unsigned long __init arch_syscall_addr(int nr)
+{
+        return (unsigned long)sys_call_table[nr];
+}
+
+#endif
diff --git a/arch/sparc/kernel/iommu.c b/arch/sparc/kernel/iommu.c
index 7690cc219ecc..5fad94950e76 100644
--- a/arch/sparc/kernel/iommu.c
+++ b/arch/sparc/kernel/iommu.c
@@ -11,6 +11,7 @@
 #include <linux/dma-mapping.h>
 #include <linux/errno.h>
 #include <linux/iommu-helper.h>
+#include <linux/bitmap.h>
 
 #ifdef CONFIG_PCI
 #include <linux/pci.h>
@@ -169,7 +170,7 @@ void iommu_range_free(struct iommu *iommu, dma_addr_t dma_addr, unsigned long np
 
         entry = (dma_addr - iommu->page_table_map_base) >> IO_PAGE_SHIFT;
 
-        iommu_area_free(arena->map, entry, npages);
+        bitmap_clear(arena->map, entry, npages);
 }
 
 int iommu_table_init(struct iommu *iommu, int tsbsize,
diff --git a/arch/sparc/kernel/irq_64.c b/arch/sparc/kernel/irq_64.c
index ce996f97855f..8d6882bb480a 100644
--- a/arch/sparc/kernel/irq_64.c
+++ b/arch/sparc/kernel/irq_64.c
@@ -176,7 +176,7 @@ int show_interrupts(struct seq_file *p, void *v)
         }
 
         if (i < NR_IRQS) {
-                spin_lock_irqsave(&irq_desc[i].lock, flags);
+                raw_spin_lock_irqsave(&irq_desc[i].lock, flags);
                 action = irq_desc[i].action;
                 if (!action)
                         goto skip;
@@ -195,7 +195,7 @@ int show_interrupts(struct seq_file *p, void *v)
 
                 seq_putc(p, '\n');
 skip:
-                spin_unlock_irqrestore(&irq_desc[i].lock, flags);
+                raw_spin_unlock_irqrestore(&irq_desc[i].lock, flags);
         } else if (i == NR_IRQS) {
                 seq_printf(p, "NMI: ");
                 for_each_online_cpu(j)
@@ -785,14 +785,14 @@ void fixup_irqs(void)
         for (irq = 0; irq < NR_IRQS; irq++) {
                 unsigned long flags;
 
-                spin_lock_irqsave(&irq_desc[irq].lock, flags);
+                raw_spin_lock_irqsave(&irq_desc[irq].lock, flags);
                 if (irq_desc[irq].action &&
                     !(irq_desc[irq].status & IRQ_PER_CPU)) {
                         if (irq_desc[irq].chip->set_affinity)
                                 irq_desc[irq].chip->set_affinity(irq,
                                         irq_desc[irq].affinity);
                 }
-                spin_unlock_irqrestore(&irq_desc[irq].lock, flags);
+                raw_spin_unlock_irqrestore(&irq_desc[irq].lock, flags);
         }
 
         tick_ops->disable_irq();
diff --git a/arch/sparc/kernel/kprobes.c b/arch/sparc/kernel/kprobes.c
index 3bc6527c95af..6716584e48ab 100644
--- a/arch/sparc/kernel/kprobes.c
+++ b/arch/sparc/kernel/kprobes.c
@@ -46,6 +46,9 @@ struct kretprobe_blackpoint kretprobe_blacklist[] = {{NULL, NULL}};
 
 int __kprobes arch_prepare_kprobe(struct kprobe *p)
 {
+        if ((unsigned long) p->addr & 0x3UL)
+                return -EILSEQ;
+
         p->ainsn.insn[0] = *p->addr;
         flushi(&p->ainsn.insn[0]);
 
diff --git a/arch/sparc/kernel/ldc.c b/arch/sparc/kernel/ldc.c
index cb3c72c45aab..df39a0f0d27a 100644
--- a/arch/sparc/kernel/ldc.c
+++ b/arch/sparc/kernel/ldc.c
@@ -14,6 +14,7 @@
 #include <linux/interrupt.h>
 #include <linux/list.h>
 #include <linux/init.h>
+#include <linux/bitmap.h>
 
 #include <asm/hypervisor.h>
 #include <asm/iommu.h>
@@ -1242,13 +1243,13 @@ int ldc_bind(struct ldc_channel *lp, const char *name)
         snprintf(lp->tx_irq_name, LDC_IRQ_NAME_MAX, "%s TX", name);
 
         err = request_irq(lp->cfg.rx_irq, ldc_rx,
-                          IRQF_SAMPLE_RANDOM | IRQF_DISABLED | IRQF_SHARED,
+                          IRQF_SAMPLE_RANDOM | IRQF_DISABLED,
                           lp->rx_irq_name, lp);
         if (err)
                 return err;
 
         err = request_irq(lp->cfg.tx_irq, ldc_tx,
-                          IRQF_SAMPLE_RANDOM | IRQF_DISABLED | IRQF_SHARED,
+                          IRQF_SAMPLE_RANDOM | IRQF_DISABLED,
                           lp->tx_irq_name, lp);
         if (err) {
                 free_irq(lp->cfg.rx_irq, lp);
@@ -1875,7 +1876,7 @@ EXPORT_SYMBOL(ldc_read);
 static long arena_alloc(struct ldc_iommu *iommu, unsigned long npages)
 {
         struct iommu_arena *arena = &iommu->arena;
-        unsigned long n, i, start, end, limit;
+        unsigned long n, start, end, limit;
         int pass;
 
         limit = arena->limit;
@@ -1883,7 +1884,7 @@ static long arena_alloc(struct ldc_iommu *iommu, unsigned long npages)
         pass = 0;
 
 again:
-        n = find_next_zero_bit(arena->map, limit, start);
+        n = bitmap_find_next_zero_area(arena->map, limit, start, npages, 0);
         end = n + npages;
         if (unlikely(end >= limit)) {
                 if (likely(pass < 1)) {
@@ -1896,16 +1897,7 @@ again:
                         return -1;
                 }
         }
-
-        for (i = n; i < end; i++) {
-                if (test_bit(i, arena->map)) {
-                        start = i + 1;
-                        goto again;
-                }
-        }
-
-        for (i = n; i < end; i++)
-                __set_bit(i, arena->map);
+        bitmap_set(arena->map, n, npages);
 
         arena->hint = end;
 
diff --git a/arch/sparc/kernel/mdesc.c b/arch/sparc/kernel/mdesc.c
index 938da19dc065..cdc91d919e93 100644
--- a/arch/sparc/kernel/mdesc.c
+++ b/arch/sparc/kernel/mdesc.c
@@ -10,6 +10,7 @@
 #include <linux/slab.h>
 #include <linux/mm.h>
 #include <linux/miscdevice.h>
+#include <linux/bootmem.h>
 
 #include <asm/cpudata.h>
 #include <asm/hypervisor.h>
@@ -108,25 +109,15 @@ static struct mdesc_handle * __init mdesc_lmb_alloc(unsigned int mdesc_size)
 
 static void mdesc_lmb_free(struct mdesc_handle *hp)
 {
-        unsigned int alloc_size, handle_size = hp->handle_size;
-        unsigned long start, end;
+        unsigned int alloc_size;
+        unsigned long start;
 
         BUG_ON(atomic_read(&hp->refcnt) != 0);
         BUG_ON(!list_empty(&hp->list));
 
-        alloc_size = PAGE_ALIGN(handle_size);
-
-        start = (unsigned long) hp;
-        end = start + alloc_size;
-
-        while (start < end) {
-                struct page *p;
-
-                p = virt_to_page(start);
-                ClearPageReserved(p);
-                __free_page(p);
-                start += PAGE_SIZE;
-        }
+        alloc_size = PAGE_ALIGN(hp->handle_size);
+        start = __pa(hp);
+        free_bootmem_late(start, alloc_size);
 }
 
 static struct mdesc_mem_ops lmb_mdesc_ops = {
diff --git a/arch/sparc/kernel/nmi.c b/arch/sparc/kernel/nmi.c
index f30f4a1ead23..d242a7340541 100644
--- a/arch/sparc/kernel/nmi.c
+++ b/arch/sparc/kernel/nmi.c
@@ -96,7 +96,6 @@ notrace __kprobes void perfctr_irq(int irq, struct pt_regs *regs)
         int cpu = smp_processor_id();
 
         clear_softint(1 << irq);
-        pcr_ops->write(PCR_PIC_PRIV);
 
         local_cpu_data().__nmi_count++;
 
@@ -105,6 +104,8 @@ notrace __kprobes void perfctr_irq(int irq, struct pt_regs *regs)
         if (notify_die(DIE_NMI, "nmi", regs, 0,
                        pt_regs_trap_type(regs), SIGINT) == NOTIFY_STOP)
                 touched = 1;
+        else
+                pcr_ops->write(PCR_PIC_PRIV);
 
         sum = kstat_irqs_cpu(0, cpu);
         if (__get_cpu_var(nmi_touch)) {
diff --git a/arch/sparc/kernel/of_device_64.c b/arch/sparc/kernel/of_device_64.c
index 881947e59e95..0a6f2d1798d1 100644
--- a/arch/sparc/kernel/of_device_64.c
+++ b/arch/sparc/kernel/of_device_64.c
@@ -104,9 +104,19 @@ static int of_bus_pci_map(u32 *addr, const u32 *range,
         int i;
 
         /* Check address type match */
-        if ((addr[0] ^ range[0]) & 0x03000000)
-                return -EINVAL;
+        if (!((addr[0] ^ range[0]) & 0x03000000))
+                goto type_match;
+
+        /* Special exception, we can map a 64-bit address into
+         * a 32-bit range.
+         */
+        if ((addr[0] & 0x03000000) == 0x03000000 &&
+            (range[0] & 0x03000000) == 0x02000000)
+                goto type_match;
+
+        return -EINVAL;
 
+type_match:
         if (of_out_of_range(addr + 1, range + 1, range + na + pna,
                             na - 1, ns))
                 return -EINVAL;
diff --git a/arch/sparc/kernel/pci.c b/arch/sparc/kernel/pci.c
index b85374f7cf94..539e83f8e087 100644
--- a/arch/sparc/kernel/pci.c
+++ b/arch/sparc/kernel/pci.c
@@ -1064,7 +1064,6 @@ int pci64_dma_supported(struct pci_dev *pdev, u64 device_mask)
 
         return (device_mask & dma_addr_mask) == dma_addr_mask;
 }
-EXPORT_SYMBOL(pci_dma_supported);
 
 void pci_resource_to_user(const struct pci_dev *pdev, int bar,
                           const struct resource *rp, resource_size_t *start,
diff --git a/arch/sparc/kernel/perf_event.c b/arch/sparc/kernel/perf_event.c
index fa5936e1c3b9..198fb4e79ba2 100644
--- a/arch/sparc/kernel/perf_event.c
+++ b/arch/sparc/kernel/perf_event.c
@@ -986,6 +986,17 @@ static int __kprobes perf_event_nmi_handler(struct notifier_block *self,
         data.addr = 0;
 
         cpuc = &__get_cpu_var(cpu_hw_events);
+
+        /* If the PMU has the TOE IRQ enable bits, we need to do a
+         * dummy write to the %pcr to clear the overflow bits and thus
+         * the interrupt.
+         *
+         * Do this before we peek at the counters to determine
+         * overflow so we don't lose any events.
+         */
+        if (sparc_pmu->irq_bit)
+                pcr_ops->write(cpuc->pcr);
+
         for (idx = 0; idx < MAX_HWEVENTS; idx++) {
                 struct perf_event *event = cpuc->events[idx];
                 struct hw_perf_event *hwc;
diff --git a/arch/sparc/kernel/ptrace_64.c b/arch/sparc/kernel/ptrace_64.c
index 4ae91dc2feb9..2f6524d1a817 100644
--- a/arch/sparc/kernel/ptrace_64.c
+++ b/arch/sparc/kernel/ptrace_64.c
@@ -23,6 +23,7 @@
 #include <linux/signal.h>
 #include <linux/regset.h>
 #include <linux/tracehook.h>
+#include <trace/syscall.h>
 #include <linux/compat.h>
 #include <linux/elf.h>
 
@@ -37,6 +38,9 @@
 #include <asm/cpudata.h>
 #include <asm/cacheflush.h>
 
+#define CREATE_TRACE_POINTS
+#include <trace/events/syscalls.h>
+
 #include "entry.h"
 
 /* #define ALLOW_INIT_TRACING */
@@ -1059,6 +1063,9 @@ asmlinkage int syscall_trace_enter(struct pt_regs *regs)
         if (test_thread_flag(TIF_SYSCALL_TRACE))
                 ret = tracehook_report_syscall_entry(regs);
 
+        if (unlikely(test_thread_flag(TIF_SYSCALL_TRACEPOINT)))
+                trace_sys_enter(regs, regs->u_regs[UREG_G1]);
+
         if (unlikely(current->audit_context) && !ret)
                 audit_syscall_entry((test_thread_flag(TIF_32BIT) ?
                                      AUDIT_ARCH_SPARC :
@@ -1084,6 +1091,9 @@ asmlinkage void syscall_trace_leave(struct pt_regs *regs)
                 audit_syscall_exit(result, regs->u_regs[UREG_I0]);
         }
 
+        if (unlikely(test_thread_flag(TIF_SYSCALL_TRACEPOINT)))
+                trace_sys_exit(regs, regs->u_regs[UREG_G1]);
+
         if (test_thread_flag(TIF_SYSCALL_TRACE))
                 tracehook_report_syscall_exit(regs, 0);
 }
diff --git a/arch/sparc/kernel/sparc_ksyms_64.c b/arch/sparc/kernel/sparc_ksyms_64.c
index 0f26066a08d9..372ad59c4cba 100644
--- a/arch/sparc/kernel/sparc_ksyms_64.c
+++ b/arch/sparc/kernel/sparc_ksyms_64.c
@@ -38,17 +38,5 @@ EXPORT_SYMBOL(sun4v_niagara_setperf);
 EXPORT_SYMBOL(sun4v_niagara2_getperf);
 EXPORT_SYMBOL(sun4v_niagara2_setperf);
 
-#ifdef CONFIG_PCI
-/* inline functions in asm/pci_64.h */
-EXPORT_SYMBOL(pci_alloc_consistent);
-EXPORT_SYMBOL(pci_free_consistent);
-EXPORT_SYMBOL(pci_map_single);
-EXPORT_SYMBOL(pci_unmap_single);
-EXPORT_SYMBOL(pci_map_sg);
-EXPORT_SYMBOL(pci_unmap_sg);
-EXPORT_SYMBOL(pci_dma_sync_single_for_cpu);
-EXPORT_SYMBOL(pci_dma_sync_sg_for_cpu);
-#endif
-
 /* Exporting a symbol from /init/main.c */
 EXPORT_SYMBOL(saved_command_line);
diff --git a/arch/sparc/kernel/syscalls.S b/arch/sparc/kernel/syscalls.S
index d150c2aa98d2..dc4a458f74dc 100644
--- a/arch/sparc/kernel/syscalls.S
+++ b/arch/sparc/kernel/syscalls.S
@@ -62,7 +62,7 @@ sys32_rt_sigreturn:
 #endif
         .align  32
 1:      ldx     [%g6 + TI_FLAGS], %l5
-        andcc   %l5, (_TIF_SYSCALL_TRACE|_TIF_SECCOMP|_TIF_SYSCALL_AUDIT), %g0
+        andcc   %l5, (_TIF_SYSCALL_TRACE|_TIF_SECCOMP|_TIF_SYSCALL_AUDIT|_TIF_SYSCALL_TRACEPOINT), %g0
         be,pt   %icc, rtrap
          nop
         call    syscall_trace_leave
@@ -187,7 +187,7 @@ linux_syscall_trace:
         .globl  linux_sparc_syscall32
 linux_sparc_syscall32:
         /* Direct access to user regs, much faster. */
-        cmp     %g1, NR_SYSCALLS                        ! IEU1  Group
+        cmp     %g1, NR_syscalls                        ! IEU1  Group
         bgeu,pn %xcc, linux_sparc_ni_syscall            ! CTI
          srl    %i0, 0, %o0                             ! IEU0
         sll     %g1, 2, %l4                             ! IEU0  Group
@@ -198,7 +198,7 @@ linux_sparc_syscall32:
 
         srl     %i5, 0, %o5                             ! IEU1
         srl     %i2, 0, %o2                             ! IEU0  Group
-        andcc   %l0, (_TIF_SYSCALL_TRACE|_TIF_SECCOMP|_TIF_SYSCALL_AUDIT), %g0
+        andcc   %l0, (_TIF_SYSCALL_TRACE|_TIF_SECCOMP|_TIF_SYSCALL_AUDIT|_TIF_SYSCALL_TRACEPOINT), %g0
         bne,pn  %icc, linux_syscall_trace32             ! CTI
          mov    %i0, %l5                                ! IEU1
         call    %l7                                     ! CTI   Group brk forced
@@ -210,7 +210,7 @@ linux_sparc_syscall32:
         .globl  linux_sparc_syscall
 linux_sparc_syscall:
         /* Direct access to user regs, much faster. */
-        cmp     %g1, NR_SYSCALLS                        ! IEU1  Group
+        cmp     %g1, NR_syscalls                        ! IEU1  Group
         bgeu,pn %xcc, linux_sparc_ni_syscall            ! CTI
          mov    %i0, %o0                                ! IEU0
         sll     %g1, 2, %l4                             ! IEU0  Group
@@ -221,7 +221,7 @@ linux_sparc_syscall:
 
         mov     %i3, %o3                                ! IEU1
         mov     %i4, %o4                                ! IEU0  Group
-        andcc   %l0, (_TIF_SYSCALL_TRACE|_TIF_SECCOMP|_TIF_SYSCALL_AUDIT), %g0
+        andcc   %l0, (_TIF_SYSCALL_TRACE|_TIF_SECCOMP|_TIF_SYSCALL_AUDIT|_TIF_SYSCALL_TRACEPOINT), %g0
         bne,pn  %icc, linux_syscall_trace               ! CTI   Group
          mov    %i0, %l5                                ! IEU0
 2:      call    %l7                                     ! CTI   Group brk forced
@@ -245,7 +245,7 @@ ret_sys_call:
 
         cmp     %o0, -ERESTART_RESTARTBLOCK
         bgeu,pn %xcc, 1f
-         andcc  %l0, (_TIF_SYSCALL_TRACE|_TIF_SECCOMP|_TIF_SYSCALL_AUDIT), %l6
+         andcc  %l0, (_TIF_SYSCALL_TRACE|_TIF_SECCOMP|_TIF_SYSCALL_AUDIT|_TIF_SYSCALL_TRACEPOINT), %l6
 80:
         /* System call success, clear Carry condition code. */
         andn    %g3, %g2, %g3
@@ -260,7 +260,7 @@ ret_sys_call:
         /* System call failure, set Carry condition code.
          * Also, get abs(errno) to return to the process.
          */
-        andcc   %l0, (_TIF_SYSCALL_TRACE|_TIF_SECCOMP|_TIF_SYSCALL_AUDIT), %l6  
+        andcc   %l0, (_TIF_SYSCALL_TRACE|_TIF_SECCOMP|_TIF_SYSCALL_AUDIT|_TIF_SYSCALL_TRACEPOINT), %l6  
         sub     %g0, %o0, %o0
         or      %g3, %g2, %g3
         stx     %o0, [%sp + PTREGS_OFF + PT_V9_I0]
diff --git a/arch/sparc/kernel/time_64.c b/arch/sparc/kernel/time_64.c
index 63f73ae8a892..67e165102885 100644
--- a/arch/sparc/kernel/time_64.c
+++ b/arch/sparc/kernel/time_64.c
@@ -774,26 +774,9 @@ void __devinit setup_sparc64_timer(void)
 static struct clocksource clocksource_tick = {
         .rating         = 100,
         .mask           = CLOCKSOURCE_MASK(64),
-        .shift          = 16,
         .flags          = CLOCK_SOURCE_IS_CONTINUOUS,
 };
 
-static void __init setup_clockevent_multiplier(unsigned long hz)
-{
-        unsigned long mult, shift = 32;
-
-        while (1) {
-                mult = div_sc(hz, NSEC_PER_SEC, shift);
-                if (mult && (mult >> 32UL) == 0UL)
-                        break;
-
-                shift--;
-        }
-
-        sparc64_clockevent.shift = shift;
-        sparc64_clockevent.mult = mult;
-}
-
 static unsigned long tb_ticks_per_usec __read_mostly;
 
 void __delay(unsigned long loops)
@@ -828,9 +811,7 @@ void __init time_init(void)
                 clocksource_hz2mult(freq, SPARC64_NSEC_PER_CYC_SHIFT);
 
         clocksource_tick.name = tick_ops->name;
-        clocksource_tick.mult =
-                clocksource_hz2mult(freq,
-                                    clocksource_tick.shift);
+        clocksource_calc_mult_shift(&clocksource_tick, freq, 4);
         clocksource_tick.read = clocksource_tick_read;
 
         printk("clocksource: mult[%x] shift[%d]\n",
@@ -839,15 +820,14 @@ void __init time_init(void)
         clocksource_register(&clocksource_tick);
 
         sparc64_clockevent.name = tick_ops->name;
-
-        setup_clockevent_multiplier(freq);
+        clockevents_calc_mult_shift(&sparc64_clockevent, freq, 4);
 
         sparc64_clockevent.max_delta_ns =
                 clockevent_delta2ns(0x7fffffffffffffffUL, &sparc64_clockevent);
         sparc64_clockevent.min_delta_ns =
                 clockevent_delta2ns(0xF, &sparc64_clockevent);
 
-        printk("clockevent: mult[%ux] shift[%d]\n",
+        printk("clockevent: mult[%x] shift[%d]\n",
                sparc64_clockevent.mult, sparc64_clockevent.shift);
 
         setup_sparc64_timer();
diff --git a/arch/sparc/kernel/unaligned_32.c b/arch/sparc/kernel/unaligned_32.c
index 6b1e6cde6fff..f8514e291e15 100644
--- a/arch/sparc/kernel/unaligned_32.c
+++ b/arch/sparc/kernel/unaligned_32.c
@@ -17,8 +17,7 @@
 #include <asm/uaccess.h>
 #include <linux/smp.h>
 #include <linux/smp_lock.h>
-
-/* #define DEBUG_MNA */
+#include <linux/perf_event.h>
 
 enum direction {
         load,    /* ld, ldd, ldh, ldsh */
@@ -29,12 +28,6 @@ enum direction {
         invalid,
 };
 
-#ifdef DEBUG_MNA
-static char *dirstrings[] = {
-  "load", "store", "both", "fpload", "fpstore", "invalid"
-};
-#endif
-
 static inline enum direction decode_direction(unsigned int insn)
 {
         unsigned long tmp = (insn >> 21) & 1;
@@ -255,10 +248,7 @@ asmlinkage void kernel_unaligned_trap(struct pt_regs *regs, unsigned int insn)
                 unsigned long addr = compute_effective_address(regs, insn);
                 int err;
 
-#ifdef DEBUG_MNA
-                printk("KMNA: pc=%08lx [dir=%s addr=%08lx size=%d] retpc[%08lx]\n",
-                       regs->pc, dirstrings[dir], addr, size, regs->u_regs[UREG_RETPC]);
-#endif
+                perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0, regs, addr);
                 switch (dir) {
                 case load:
                         err = do_int_load(fetch_reg_addr(((insn>>25)&0x1f),
@@ -350,6 +340,7 @@ asmlinkage void user_unaligned_trap(struct pt_regs *regs, unsigned int insn)
                 }
 
                 addr = compute_effective_address(regs, insn);
+                perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0, regs, addr);
                 switch(dir) {
                 case load:
                         err = do_int_load(fetch_reg_addr(((insn>>25)&0x1f),
diff --git a/arch/sparc/kernel/unaligned_64.c b/arch/sparc/kernel/unaligned_64.c
index 379209982a07..378ca82b9ccc 100644
--- a/arch/sparc/kernel/unaligned_64.c
+++ b/arch/sparc/kernel/unaligned_64.c
@@ -20,10 +20,9 @@
 #include <asm/uaccess.h>
 #include <linux/smp.h>
 #include <linux/bitops.h>
+#include <linux/perf_event.h>
 #include <asm/fpumacro.h>
 
-/* #define DEBUG_MNA */
-
 enum direction {
         load,    /* ld, ldd, ldh, ldsh */
         store,   /* st, std, sth, stsh */
@@ -33,12 +32,6 @@ enum direction {
         invalid,
 };
 
-#ifdef DEBUG_MNA
-static char *dirstrings[] = {
-  "load", "store", "both", "fpload", "fpstore", "invalid"
-};
-#endif
-
 static inline enum direction decode_direction(unsigned int insn)
 {
         unsigned long tmp = (insn >> 21) & 1;
@@ -327,12 +320,7 @@ asmlinkage void kernel_unaligned_trap(struct pt_regs *regs, unsigned int insn)
 
                 addr = compute_effective_address(regs, insn,
                                                  ((insn >> 25) & 0x1f));
-#ifdef DEBUG_MNA
-                printk("KMNA: pc=%016lx [dir=%s addr=%016lx size=%d] "
-                       "retpc[%016lx]\n",
-                       regs->tpc, dirstrings[dir], addr, size,
-                       regs->u_regs[UREG_RETPC]);
-#endif
+                perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0, regs, addr);
                 switch (asi) {
                 case ASI_NL:
                 case ASI_AIUPL:
@@ -399,6 +387,7 @@ int handle_popc(u32 insn, struct pt_regs *regs)
         int ret, i, rd = ((insn >> 25) & 0x1f);
         int from_kernel = (regs->tstate & TSTATE_PRIV) != 0;
                                 
+        perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, 0);
         if (insn & 0x2000) {
                 maybe_flush_windows(0, 0, rd, from_kernel);
                 value = sign_extend_imm13(insn);
@@ -445,6 +434,8 @@ int handle_ldf_stq(u32 insn, struct pt_regs *regs)
         int asi = decode_asi(insn, regs);
         int flag = (freg < 32) ? FPRS_DL : FPRS_DU;
 
+        perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, 0);
+
         save_and_clear_fpu();
         current_thread_info()->xfsr[0] &= ~0x1c000;
         if (freg & 3) {
@@ -566,6 +557,8 @@ void handle_ld_nf(u32 insn, struct pt_regs *regs)
         int from_kernel = (regs->tstate & TSTATE_PRIV) != 0;
         unsigned long *reg;
                                 
+        perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, 0);
+
         maybe_flush_windows(0, 0, rd, from_kernel);
         reg = fetch_reg_addr(rd, regs);
         if (from_kernel || rd < 16) {
@@ -596,6 +589,7 @@ void handle_lddfmna(struct pt_regs *regs, unsigned long sfar, unsigned long sfsr
 
         if (tstate & TSTATE_PRIV)
                 die_if_kernel("lddfmna from kernel", regs);
+        perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0, regs, sfar);
         if (test_thread_flag(TIF_32BIT))
                 pc = (u32)pc;
         if (get_user(insn, (u32 __user *) pc) != -EFAULT) {
@@ -657,6 +651,7 @@ void handle_stdfmna(struct pt_regs *regs, unsigned long sfar, unsigned long sfsr
 
         if (tstate & TSTATE_PRIV)
                 die_if_kernel("stdfmna from kernel", regs);
+        perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0, regs, sfar);
         if (test_thread_flag(TIF_32BIT))
                 pc = (u32)pc;
         if (get_user(insn, (u32 __user *) pc) != -EFAULT) {
diff --git a/arch/sparc/kernel/visemul.c b/arch/sparc/kernel/visemul.c
index d231cbd5c526..9dfd2ebcb157 100644
--- a/arch/sparc/kernel/visemul.c
+++ b/arch/sparc/kernel/visemul.c
@@ -5,6 +5,7 @@
 #include <linux/kernel.h>
 #include <linux/errno.h>
 #include <linux/thread_info.h>
+#include <linux/perf_event.h>
 
 #include <asm/ptrace.h>
 #include <asm/pstate.h>
@@ -801,6 +802,8 @@ int vis_emul(struct pt_regs *regs, unsigned int insn)
 
         BUG_ON(regs->tstate & TSTATE_PRIV);
 
+        perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, 0);
+
         if (test_thread_flag(TIF_32BIT))
                 pc = (u32)pc;