*/
 struct hvlpevent_queue hvlpevent_queue __attribute__((__section__(".data")));
 
-static char *event_types[9] = {
+DEFINE_PER_CPU(unsigned long[HvLpEvent_Type_NumTypes], hvlpevent_counts);
+
+static char *event_types[HvLpEvent_Type_NumTypes] = {
        "Hypervisor\t\t",
        "Machine Facilities\t",
        "Session Manager\t",
 
 void process_hvlpevents(struct pt_regs *regs)
 {
-       unsigned numIntsProcessed = 0;
        struct HvLpEvent * nextLpEvent;
 
        /* If we have recursed, just return */
        for (;;) {
                nextLpEvent = get_next_hvlpevent();
                if ( nextLpEvent ) {
-                       ++numIntsProcessed;
-                       hvlpevent_queue.xLpIntCount++;
                        /* Call appropriate handler here, passing 
                         * a pointer to the LpEvent.  The handler
                         * must make a copy of the LpEvent if it
                         * here!
                         */
                        if ( nextLpEvent->xType < HvLpEvent_Type_NumTypes )
-                               hvlpevent_queue.xLpIntCountByType[nextLpEvent->xType]++;
+                               __get_cpu_var(hvlpevent_counts)[nextLpEvent->xType]++;
                        if ( nextLpEvent->xType < HvLpEvent_Type_NumTypes &&
                             lpEventHandler[nextLpEvent->xType] ) 
                                lpEventHandler[nextLpEvent->xType](nextLpEvent, regs);
        ItLpQueueInProcess = 0;
        mb();
        clear_inUse();
-
-       get_paca()->lpevent_count += numIntsProcessed;
 }
 
 static int set_spread_lpevents(char *str)
 
 static int proc_lpevents_show(struct seq_file *m, void *v)
 {
-       unsigned int i;
+       int cpu, i;
+       unsigned long sum;
+       static unsigned long cpu_totals[NR_CPUS];
+
+       /* FIXME: do we care that there's no locking here? */
+       sum = 0;
+       for_each_online_cpu(cpu) {
+               cpu_totals[cpu] = 0;
+               for (i = 0; i < HvLpEvent_Type_NumTypes; i++) {
+                       cpu_totals[cpu] += per_cpu(hvlpevent_counts, cpu)[i];
+               }
+               sum += cpu_totals[cpu];
+       }
 
        seq_printf(m, "LpEventQueue 0\n");
-       seq_printf(m, "  events processed:\t%lu\n",
-                  (unsigned long)hvlpevent_queue.xLpIntCount);
+       seq_printf(m, "  events processed:\t%lu\n", sum);
 
-       for (i = 0; i < 9; ++i)
-               seq_printf(m, "    %s %10lu\n", event_types[i],
-                          (unsigned long)hvlpevent_queue.xLpIntCountByType[i]);
+       for (i = 0; i < HvLpEvent_Type_NumTypes; ++i) {
+               sum = 0;
+               for_each_online_cpu(cpu) {
+                       sum += per_cpu(hvlpevent_counts, cpu)[i];
+               }
+
+               seq_printf(m, "    %s %10lu\n", event_types[i], sum);
+       }
 
        seq_printf(m, "\n  events processed by processor:\n");
 
-       for_each_online_cpu(i)
-               seq_printf(m, "    CPU%02d  %10u\n", i, paca[i].lpevent_count);
+       for_each_online_cpu(cpu) {
+               seq_printf(m, "    CPU%02d  %10lu\n", cpu, cpu_totals[cpu]);
+       }
 
        return 0;
 }