include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit...
[safe/jmp/linux-2.6] / arch / s390 / kernel / smp.c
index 5a445b1..e4d98de 100644 (file)
@@ -1,7 +1,7 @@
 /*
  *  arch/s390/kernel/smp.c
  *
- *    Copyright IBM Corp. 1999,2007
+ *    Copyright IBM Corp. 1999, 2009
  *    Author(s): Denis Joseph Barrow (djbarrow@de.ibm.com,barrow_dj@yahoo.com),
  *              Martin Schwidefsky (schwidefsky@de.ibm.com)
  *              Heiko Carstens (heiko.carstens@de.ibm.com)
@@ -20,6 +20,9 @@
  * cpu_number_map in other architectures.
  */
 
+#define KMSG_COMPONENT "cpu"
+#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt
+
 #include <linux/module.h>
 #include <linux/init.h>
 #include <linux/mm.h>
 #include <linux/delay.h>
 #include <linux/cache.h>
 #include <linux/interrupt.h>
+#include <linux/irqflags.h>
 #include <linux/cpu.h>
 #include <linux/timex.h>
 #include <linux/bootmem.h>
+#include <linux/slab.h>
+#include <asm/asm-offsets.h>
 #include <asm/ipl.h>
 #include <asm/setup.h>
 #include <asm/sigp.h>
 #include <asm/timer.h>
 #include <asm/lowcore.h>
 #include <asm/sclp.h>
+#include <asm/cputime.h>
+#include <asm/vdso.h>
 #include <asm/cpu.h>
+#include "entry.h"
 
-/*
- * An array with a pointer the lowcore of every CPU.
- */
-struct _lowcore *lowcore_ptr[NR_CPUS];
-EXPORT_SYMBOL(lowcore_ptr);
-
-cpumask_t cpu_online_map = CPU_MASK_NONE;
-EXPORT_SYMBOL(cpu_online_map);
-
-cpumask_t cpu_possible_map = CPU_MASK_ALL;
-EXPORT_SYMBOL(cpu_possible_map);
+/* logical cpu to cpu address */
+unsigned short __cpu_logical_map[NR_CPUS];
 
 static struct task_struct *current_set[NR_CPUS];
 
@@ -74,166 +74,61 @@ static int cpu_management;
 
 static DEFINE_PER_CPU(struct cpu, cpu_devices);
 
-static void smp_ext_bitcall(int, ec_bit_sig);
-
-/*
- * Structure and data for __smp_call_function_map(). This is designed to
- * minimise static memory requirements. It also looks cleaner.
- */
-static DEFINE_SPINLOCK(call_lock);
-
-struct call_data_struct {
-       void (*func) (void *info);
-       void *info;
-       cpumask_t started;
-       cpumask_t finished;
-       int wait;
-};
+static void smp_ext_bitcall(int, int);
 
-static struct call_data_struct *call_data;
-
-/*
- * 'Call function' interrupt callback
- */
-static void do_call_function(void)
-{
-       void (*func) (void *info) = call_data->func;
-       void *info = call_data->info;
-       int wait = call_data->wait;
-
-       cpu_set(smp_processor_id(), call_data->started);
-       (*func)(info);
-       if (wait)
-               cpu_set(smp_processor_id(), call_data->finished);;
-}
-
-static void __smp_call_function_map(void (*func) (void *info), void *info,
-                                   int nonatomic, int wait, cpumask_t map)
+static int raw_cpu_stopped(int cpu)
 {
-       struct call_data_struct data;
-       int cpu, local = 0;
-
-       /*
-        * Can deadlock when interrupts are disabled or if in wrong context.
-        */
-       WARN_ON(irqs_disabled() || in_irq());
-
-       /*
-        * Check for local function call. We have to have the same call order
-        * as in on_each_cpu() because of machine_restart_smp().
-        */
-       if (cpu_isset(smp_processor_id(), map)) {
-               local = 1;
-               cpu_clear(smp_processor_id(), map);
-       }
-
-       cpus_and(map, map, cpu_online_map);
-       if (cpus_empty(map))
-               goto out;
-
-       data.func = func;
-       data.info = info;
-       data.started = CPU_MASK_NONE;
-       data.wait = wait;
-       if (wait)
-               data.finished = CPU_MASK_NONE;
-
-       spin_lock(&call_lock);
-       call_data = &data;
+       u32 status;
 
-       for_each_cpu_mask(cpu, map)
-               smp_ext_bitcall(cpu, ec_call_function);
-
-       /* Wait for response */
-       while (!cpus_equal(map, data.started))
-               cpu_relax();
-       if (wait)
-               while (!cpus_equal(map, data.finished))
-                       cpu_relax();
-       spin_unlock(&call_lock);
-out:
-       if (local) {
-               local_irq_disable();
-               func(info);
-               local_irq_enable();
+       switch (raw_sigp_ps(&status, 0, cpu, sigp_sense)) {
+       case sigp_status_stored:
+               /* Check for stopped and check stop state */
+               if (status & 0x50)
+                       return 1;
+               break;
+       default:
+               break;
        }
-}
-
-/*
- * smp_call_function:
- * @func: the function to run; this must be fast and non-blocking
- * @info: an arbitrary pointer to pass to the function
- * @nonatomic: unused
- * @wait: if true, wait (atomically) until function has completed on other CPUs
- *
- * Run a function on all other CPUs.
- *
- * You must not call this function with disabled interrupts, from a
- * hardware interrupt handler or from a bottom half.
- */
-int smp_call_function(void (*func) (void *info), void *info, int nonatomic,
-                     int wait)
-{
-       cpumask_t map;
-
-       preempt_disable();
-       map = cpu_online_map;
-       cpu_clear(smp_processor_id(), map);
-       __smp_call_function_map(func, info, nonatomic, wait, map);
-       preempt_enable();
        return 0;
 }
-EXPORT_SYMBOL(smp_call_function);
 
-/*
- * smp_call_function_single:
- * @cpu: the CPU where func should run
- * @func: the function to run; this must be fast and non-blocking
- * @info: an arbitrary pointer to pass to the function
- * @nonatomic: unused
- * @wait: if true, wait (atomically) until function has completed on other CPUs
- *
- * Run a function on one processor.
- *
- * You must not call this function with disabled interrupts, from a
- * hardware interrupt handler or from a bottom half.
- */
-int smp_call_function_single(int cpu, void (*func) (void *info), void *info,
-                            int nonatomic, int wait)
+static inline int cpu_stopped(int cpu)
 {
-       preempt_disable();
-       __smp_call_function_map(func, info, nonatomic, wait,
-                               cpumask_of_cpu(cpu));
-       preempt_enable();
-       return 0;
+       return raw_cpu_stopped(cpu_logical_map(cpu));
 }
-EXPORT_SYMBOL(smp_call_function_single);
-
-/**
- * smp_call_function_mask(): Run a function on a set of other CPUs.
- * @mask: The set of cpus to run on.  Must not include the current cpu.
- * @func: The function to run. This must be fast and non-blocking.
- * @info: An arbitrary pointer to pass to the function.
- * @wait: If true, wait (atomically) until function has completed on other CPUs.
- *
- * Returns 0 on success, else a negative status code.
- *
- * If @wait is true, then returns once @func has returned; otherwise
- * it returns just before the target cpu calls @func.
- *
- * You must not call this function with disabled interrupts or from a
- * hardware interrupt handler or from a bottom half handler.
- */
-int smp_call_function_mask(cpumask_t mask, void (*func)(void *), void *info,
-                          int wait)
+
+void smp_switch_to_ipl_cpu(void (*func)(void *), void *data)
 {
-       preempt_disable();
-       cpu_clear(smp_processor_id(), mask);
-       __smp_call_function_map(func, info, 0, wait, mask);
-       preempt_enable();
-       return 0;
+       struct _lowcore *lc, *current_lc;
+       struct stack_frame *sf;
+       struct pt_regs *regs;
+       unsigned long sp;
+
+       if (smp_processor_id() == 0)
+               func(data);
+       __load_psw_mask(PSW_BASE_BITS | PSW_DEFAULT_KEY);
+       /* Disable lowcore protection */
+       __ctl_clear_bit(0, 28);
+       current_lc = lowcore_ptr[smp_processor_id()];
+       lc = lowcore_ptr[0];
+       if (!lc)
+               lc = current_lc;
+       lc->restart_psw.mask = PSW_BASE_BITS | PSW_DEFAULT_KEY;
+       lc->restart_psw.addr = PSW_ADDR_AMODE | (unsigned long) smp_restart_cpu;
+       if (!cpu_online(0))
+               smp_switch_to_cpu(func, data, 0, stap(), __cpu_logical_map[0]);
+       while (sigp(0, sigp_stop_and_store_status) == sigp_busy)
+               cpu_relax();
+       sp = lc->panic_stack;
+       sp -= sizeof(struct pt_regs);
+       regs = (struct pt_regs *) sp;
+       memcpy(&regs->gprs, &current_lc->gpregs_save_area, sizeof(regs->gprs));
+       regs->psw = lc->psw_save_area;
+       sp -= STACK_FRAME_OVERHEAD;
+       sf = (struct stack_frame *) sp;
+       sf->back_chain = regs->gprs[15];
+       smp_switch_to_cpu(func, data, sp, stap(), __cpu_logical_map[0]);
 }
-EXPORT_SYMBOL(smp_call_function_mask);
 
 void smp_send_stop(void)
 {
@@ -241,19 +136,17 @@ void smp_send_stop(void)
 
        /* Disable all interrupts/machine checks */
        __load_psw_mask(psw_kernel_bits & ~PSW_MASK_MCHECK);
-
-       /* write magic number to zero page (absolute 0) */
-       lowcore_ptr[smp_processor_id()]->panic_magic = __PANIC_MAGIC;
+       trace_hardirqs_off();
 
        /* stop all processors */
        for_each_online_cpu(cpu) {
                if (cpu == smp_processor_id())
                        continue;
                do {
-                       rc = signal_processor(cpu, sigp_stop);
+                       rc = sigp(cpu, sigp_stop);
                } while (rc == sigp_busy);
 
-               while (!smp_cpu_not_running(cpu))
+               while (!cpu_stopped(cpu))
                        cpu_relax();
        }
 }
@@ -276,35 +169,51 @@ static void do_ext_call_interrupt(__u16 code)
        bits = xchg(&S390_lowcore.ext_call_fast, 0);
 
        if (test_bit(ec_call_function, &bits))
-               do_call_function();
+               generic_smp_call_function_interrupt();
+
+       if (test_bit(ec_call_function_single, &bits))
+               generic_smp_call_function_single_interrupt();
 }
 
 /*
  * Send an external call sigp to another cpu and return without waiting
  * for its completion.
  */
-static void smp_ext_bitcall(int cpu, ec_bit_sig sig)
+static void smp_ext_bitcall(int cpu, int sig)
 {
        /*
         * Set signaling bit in lowcore of target cpu and kick it
         */
        set_bit(sig, (unsigned long *) &lowcore_ptr[cpu]->ext_call_fast);
-       while (signal_processor(cpu, sigp_emergency_signal) == sigp_busy)
+       while (sigp(cpu, sigp_emergency_signal) == sigp_busy)
                udelay(10);
 }
 
+void arch_send_call_function_ipi_mask(const struct cpumask *mask)
+{
+       int cpu;
+
+       for_each_cpu(cpu, mask)
+               smp_ext_bitcall(cpu, ec_call_function);
+}
+
+void arch_send_call_function_single_ipi(int cpu)
+{
+       smp_ext_bitcall(cpu, ec_call_function_single);
+}
+
 #ifndef CONFIG_64BIT
 /*
  * this function sends a 'purge tlb' signal to another CPU.
  */
-void smp_ptlb_callback(void *info)
+static void smp_ptlb_callback(void *info)
 {
        __tlb_flush_local();
 }
 
 void smp_ptlb_all(void)
 {
-       on_each_cpu(smp_ptlb_callback, NULL, 0, 1);
+       on_each_cpu(smp_ptlb_callback, NULL, 1);
 }
 EXPORT_SYMBOL(smp_ptlb_all);
 #endif /* ! CONFIG_64BIT */
@@ -352,7 +261,7 @@ void smp_ctl_set_bit(int cr, int bit)
        memset(&parms.orvals, 0, sizeof(parms.orvals));
        memset(&parms.andvals, 0xff, sizeof(parms.andvals));
        parms.orvals[cr] = 1 << bit;
-       on_each_cpu(smp_ctl_bit_callback, &parms, 0, 1);
+       on_each_cpu(smp_ctl_bit_callback, &parms, 1);
 }
 EXPORT_SYMBOL(smp_ctl_set_bit);
 
@@ -366,72 +275,37 @@ void smp_ctl_clear_bit(int cr, int bit)
        memset(&parms.orvals, 0, sizeof(parms.orvals));
        memset(&parms.andvals, 0xff, sizeof(parms.andvals));
        parms.andvals[cr] = ~(1L << bit);
-       on_each_cpu(smp_ctl_bit_callback, &parms, 0, 1);
+       on_each_cpu(smp_ctl_bit_callback, &parms, 1);
 }
 EXPORT_SYMBOL(smp_ctl_clear_bit);
 
-/*
- * In early ipl state a temp. logically cpu number is needed, so the sigp
- * functions can be used to sense other cpus. Since NR_CPUS is >= 2 on
- * CONFIG_SMP and the ipl cpu is logical cpu 0, it must be 1.
- */
-#define CPU_INIT_NO    1
-
-#if defined(CONFIG_ZFCPDUMP) || defined(CONFIG_ZFCPDUMP_MODULE)
-
-/*
- * zfcpdump_prefix_array holds prefix registers for the following scenario:
- * 64 bit zfcpdump kernel and 31 bit kernel which is to be dumped. We have to
- * save its prefix registers, since they get lost, when switching from 31 bit
- * to 64 bit.
- */
-unsigned int zfcpdump_prefix_array[NR_CPUS + 1] \
-       __attribute__((__section__(".data")));
+#ifdef CONFIG_ZFCPDUMP
 
 static void __init smp_get_save_area(unsigned int cpu, unsigned int phy_cpu)
 {
        if (ipl_info.type != IPL_TYPE_FCP_DUMP)
                return;
        if (cpu >= NR_CPUS) {
-               printk(KERN_WARNING "Registers for cpu %i not saved since dump "
-                      "kernel was compiled with NR_CPUS=%i\n", cpu, NR_CPUS);
+               pr_warning("CPU %i exceeds the maximum %i and is excluded from "
+                          "the dump\n", cpu, NR_CPUS - 1);
                return;
        }
-       zfcpdump_save_areas[cpu] = kmalloc(sizeof(union save_area), GFP_KERNEL);
-       __cpu_logical_map[CPU_INIT_NO] = (__u16) phy_cpu;
-       while (signal_processor(CPU_INIT_NO, sigp_stop_and_store_status) ==
-              sigp_busy)
+       zfcpdump_save_areas[cpu] = kmalloc(sizeof(struct save_area), GFP_KERNEL);
+       while (raw_sigp(phy_cpu, sigp_stop_and_store_status) == sigp_busy)
                cpu_relax();
-       memcpy(zfcpdump_save_areas[cpu],
-              (void *)(unsigned long) store_prefix() + SAVE_AREA_BASE,
-              SAVE_AREA_SIZE);
-#ifdef CONFIG_64BIT
-       /* copy original prefix register */
-       zfcpdump_save_areas[cpu]->s390x.pref_reg = zfcpdump_prefix_array[cpu];
-#endif
+       memcpy_real(zfcpdump_save_areas[cpu],
+                   (void *)(unsigned long) store_prefix() + SAVE_AREA_BASE,
+                   sizeof(struct save_area));
 }
 
-union save_area *zfcpdump_save_areas[NR_CPUS + 1];
+struct save_area *zfcpdump_save_areas[NR_CPUS + 1];
 EXPORT_SYMBOL_GPL(zfcpdump_save_areas);
 
 #else
 
 static inline void smp_get_save_area(unsigned int cpu, unsigned int phy_cpu) { }
 
-#endif /* CONFIG_ZFCPDUMP || CONFIG_ZFCPDUMP_MODULE */
-
-static int cpu_stopped(int cpu)
-{
-       __u32 status;
-
-       /* Check for stopped state */
-       if (signal_processor_ps(&status, 0, cpu, sigp_sense) ==
-           sigp_status_stored) {
-               if (status & 0x40)
-                       return 1;
-       }
-       return 0;
-}
+#endif /* CONFIG_ZFCPDUMP */
 
 static int cpu_known(int cpu_id)
 {
@@ -448,10 +322,10 @@ static int smp_rescan_cpus_sigp(cpumask_t avail)
 {
        int cpu_id, logical_cpu;
 
-       logical_cpu = first_cpu(avail);
-       if (logical_cpu == NR_CPUS)
+       logical_cpu = cpumask_first(&avail);
+       if (logical_cpu >= nr_cpu_ids)
                return 0;
-       for (cpu_id = 0; cpu_id <= 65535; cpu_id++) {
+       for (cpu_id = 0; cpu_id <= MAX_CPU_ADDRESS; cpu_id++) {
                if (cpu_known(cpu_id))
                        continue;
                __cpu_logical_map[logical_cpu] = cpu_id;
@@ -460,8 +334,8 @@ static int smp_rescan_cpus_sigp(cpumask_t avail)
                        continue;
                cpu_set(logical_cpu, cpu_present_map);
                smp_cpu_state[logical_cpu] = CPU_STATE_CONFIGURED;
-               logical_cpu = next_cpu(logical_cpu, avail);
-               if (logical_cpu == NR_CPUS)
+               logical_cpu = cpumask_next(logical_cpu, &avail);
+               if (logical_cpu >= nr_cpu_ids)
                        break;
        }
        return 0;
@@ -473,8 +347,8 @@ static int smp_rescan_cpus_sclp(cpumask_t avail)
        int cpu_id, logical_cpu, cpu;
        int rc;
 
-       logical_cpu = first_cpu(avail);
-       if (logical_cpu == NR_CPUS)
+       logical_cpu = cpumask_first(&avail);
+       if (logical_cpu >= nr_cpu_ids)
                return 0;
        info = kmalloc(sizeof(*info), GFP_KERNEL);
        if (!info)
@@ -495,8 +369,8 @@ static int smp_rescan_cpus_sclp(cpumask_t avail)
                        smp_cpu_state[logical_cpu] = CPU_STATE_STANDBY;
                else
                        smp_cpu_state[logical_cpu] = CPU_STATE_CONFIGURED;
-               logical_cpu = next_cpu(logical_cpu, avail);
-               if (logical_cpu == NR_CPUS)
+               logical_cpu = cpumask_next(logical_cpu, &avail);
+               if (logical_cpu >= nr_cpu_ids)
                        break;
        }
 out:
@@ -504,7 +378,7 @@ out:
        return rc;
 }
 
-static int smp_rescan_cpus(void)
+static int __smp_rescan_cpus(void)
 {
        cpumask_t avail;
 
@@ -523,18 +397,17 @@ static void __init smp_detect_cpus(void)
 
        c_cpus = 1;
        s_cpus = 0;
-       boot_cpu_addr = S390_lowcore.cpu_data.cpu_addr;
+       boot_cpu_addr = __cpu_logical_map[0];
        info = kmalloc(sizeof(*info), GFP_KERNEL);
        if (!info)
                panic("smp_detect_cpus failed to allocate memory\n");
        /* Use sigp detection algorithm if sclp doesn't work. */
        if (sclp_get_cpu_info(info)) {
                smp_use_sigp_detection = 1;
-               for (cpu = 0; cpu <= 65535; cpu++) {
+               for (cpu = 0; cpu <= MAX_CPU_ADDRESS; cpu++) {
                        if (cpu == boot_cpu_addr)
                                continue;
-                       __cpu_logical_map[CPU_INIT_NO] = cpu;
-                       if (!cpu_stopped(CPU_INIT_NO))
+                       if (!raw_cpu_stopped(cpu))
                                continue;
                        smp_get_save_area(c_cpus, cpu);
                        c_cpus++;
@@ -557,8 +430,7 @@ static void __init smp_detect_cpus(void)
                cpu_addr = info->cpu[cpu].address;
                if (cpu_addr == boot_cpu_addr)
                        continue;
-               __cpu_logical_map[CPU_INIT_NO] = cpu_addr;
-               if (!cpu_stopped(CPU_INIT_NO)) {
+               if (!raw_cpu_stopped(cpu_addr)) {
                        s_cpus++;
                        continue;
                }
@@ -567,9 +439,9 @@ static void __init smp_detect_cpus(void)
        }
 out:
        kfree(info);
-       printk(KERN_INFO "CPUs: %d configured, %d standby\n", c_cpus, s_cpus);
+       pr_info("%d configured CPUs, %d standby CPUs\n", c_cpus, s_cpus);
        get_online_cpus();
-       smp_rescan_cpus();
+       __smp_rescan_cpus();
        put_online_cpus();
 }
 
@@ -583,19 +455,21 @@ int __cpuinit start_secondary(void *cpuvoid)
        preempt_disable();
        /* Enable TOD clock interrupts on the secondary cpu. */
        init_cpu_timer();
-#ifdef CONFIG_VIRT_TIMER
        /* Enable cpu timer interrupts on the secondary cpu. */
        init_cpu_vtimer();
-#endif
        /* Enable pfault pseudo page faults on this cpu. */
        pfault_init();
 
+       /* call cpu notifiers */
+       notify_cpu_starting(smp_processor_id());
        /* Mark this cpu as online */
+       ipi_call_lock();
        cpu_set(smp_processor_id(), cpu_online_map);
+       ipi_call_unlock();
        /* Switch on interrupts */
        local_irq_enable();
        /* Print info about this processor */
-       print_cpu_info(&S390_lowcore.cpu_data);
+       print_cpu_info();
        /* cpu_idle will call schedule for us */
        cpu_idle();
        return 0;
@@ -613,17 +487,14 @@ static void __init smp_create_idle(unsigned int cpu)
        if (IS_ERR(p))
                panic("failed fork for CPU %u: %li", cpu, PTR_ERR(p));
        current_set[cpu] = p;
-       spin_lock_init(&(&per_cpu(s390_idle, cpu))->lock);
 }
 
 static int __cpuinit smp_alloc_lowcore(int cpu)
 {
        unsigned long async_stack, panic_stack;
        struct _lowcore *lowcore;
-       int lc_order;
 
-       lc_order = sizeof(long) == 8 ? 1 : 0;
-       lowcore = (void *) __get_free_pages(GFP_KERNEL | GFP_DMA, lc_order);
+       lowcore = (void *) __get_free_pages(GFP_KERNEL | GFP_DMA, LC_ORDER);
        if (!lowcore)
                return -ENOMEM;
        async_stack = __get_free_pages(GFP_KERNEL, ASYNC_ORDER);
@@ -641,63 +512,64 @@ static int __cpuinit smp_alloc_lowcore(int cpu)
 
                save_area = get_zeroed_page(GFP_KERNEL);
                if (!save_area)
-                       goto out_save_area;
+                       goto out;
                lowcore->extended_save_area_addr = (u32) save_area;
        }
+#else
+       if (vdso_alloc_per_cpu(cpu, lowcore))
+               goto out;
 #endif
        lowcore_ptr[cpu] = lowcore;
        return 0;
 
-#ifndef CONFIG_64BIT
-out_save_area:
-       free_page(panic_stack);
-#endif
 out:
+       free_page(panic_stack);
        free_pages(async_stack, ASYNC_ORDER);
-       free_pages((unsigned long) lowcore, lc_order);
+       free_pages((unsigned long) lowcore, LC_ORDER);
        return -ENOMEM;
 }
 
-#ifdef CONFIG_HOTPLUG_CPU
 static void smp_free_lowcore(int cpu)
 {
        struct _lowcore *lowcore;
-       int lc_order;
 
-       lc_order = sizeof(long) == 8 ? 1 : 0;
        lowcore = lowcore_ptr[cpu];
 #ifndef CONFIG_64BIT
        if (MACHINE_HAS_IEEE)
                free_page((unsigned long) lowcore->extended_save_area_addr);
+#else
+       vdso_free_per_cpu(cpu, lowcore);
 #endif
        free_page(lowcore->panic_stack - PAGE_SIZE);
        free_pages(lowcore->async_stack - ASYNC_SIZE, ASYNC_ORDER);
-       free_pages((unsigned long) lowcore, lc_order);
+       free_pages((unsigned long) lowcore, LC_ORDER);
        lowcore_ptr[cpu] = NULL;
 }
-#endif /* CONFIG_HOTPLUG_CPU */
 
 /* Upping and downing of CPUs */
 int __cpuinit __cpu_up(unsigned int cpu)
 {
-       struct task_struct *idle;
        struct _lowcore *cpu_lowcore;
+       struct task_struct *idle;
        struct stack_frame *sf;
-       sigp_ccode ccode;
+       u32 lowcore;
+       int ccode;
 
        if (smp_cpu_state[cpu] != CPU_STATE_CONFIGURED)
                return -EIO;
        if (smp_alloc_lowcore(cpu))
                return -ENOMEM;
-
-       ccode = signal_processor_p((__u32)(unsigned long)(lowcore_ptr[cpu]),
-                                  cpu, sigp_set_prefix);
-       if (ccode) {
-               printk("sigp_set_prefix failed for cpu %d "
-                      "with condition code %d\n",
-                      (int) cpu, (int) ccode);
-               return -EIO;
-       }
+       do {
+               ccode = sigp(cpu, sigp_initial_cpu_reset);
+               if (ccode == sigp_busy)
+                       udelay(10);
+               if (ccode == sigp_not_operational)
+                       goto err_out;
+       } while (ccode == sigp_busy);
+
+       lowcore = (u32)(unsigned long)lowcore_ptr[cpu];
+       while (sigp_p(lowcore, cpu, sigp_set_prefix) == sigp_busy)
+               udelay(10);
 
        idle = current_set[cpu];
        cpu_lowcore = lowcore_ptr[cpu];
@@ -710,23 +582,28 @@ int __cpuinit __cpu_up(unsigned int cpu)
        memset(sf, 0, sizeof(struct stack_frame));
        sf->gprs[9] = (unsigned long) sf;
        cpu_lowcore->save_area[15] = (unsigned long) sf;
-       __ctl_store(cpu_lowcore->cregs_save_area[0], 0, 15);
+       __ctl_store(cpu_lowcore->cregs_save_area, 0, 15);
        asm volatile(
                "       stam    0,15,0(%0)"
                : : "a" (&cpu_lowcore->access_regs_save_area) : "memory");
        cpu_lowcore->percpu_offset = __per_cpu_offset[cpu];
        cpu_lowcore->current_task = (unsigned long) idle;
-       cpu_lowcore->cpu_data.cpu_nr = cpu;
+       cpu_lowcore->cpu_nr = cpu;
        cpu_lowcore->kernel_asce = S390_lowcore.kernel_asce;
-       cpu_lowcore->ipl_device = S390_lowcore.ipl_device;
+       cpu_lowcore->machine_flags = S390_lowcore.machine_flags;
+       cpu_lowcore->ftrace_func = S390_lowcore.ftrace_func;
        eieio();
 
-       while (signal_processor(cpu, sigp_restart) == sigp_busy)
+       while (sigp(cpu, sigp_restart) == sigp_busy)
                udelay(10);
 
        while (!cpu_online(cpu))
                cpu_relax();
        return 0;
+
+err_out:
+       smp_free_lowcore(cpu);
+       return -EIO;
 }
 
 static int __init setup_possible_cpus(char *s)
@@ -734,9 +611,9 @@ static int __init setup_possible_cpus(char *s)
        int pcpus, cpu;
 
        pcpus = simple_strtoul(s, NULL, 0);
-       cpu_possible_map = cpumask_of_cpu(0);
-       for (cpu = 1; cpu < pcpus && cpu < NR_CPUS; cpu++)
-               cpu_set(cpu, cpu_possible_map);
+       init_cpu_possible(cpumask_of(0));
+       for (cpu = 1; cpu < pcpus && cpu < nr_cpu_ids; cpu++)
+               set_cpu_possible(cpu, true);
        return 0;
 }
 early_param("possible_cpus", setup_possible_cpus);
@@ -777,17 +654,19 @@ int __cpu_disable(void)
 void __cpu_die(unsigned int cpu)
 {
        /* Wait until target cpu is down */
-       while (!smp_cpu_not_running(cpu))
+       while (!cpu_stopped(cpu))
                cpu_relax();
+       while (sigp_p(0, cpu, sigp_set_prefix) == sigp_busy)
+               udelay(10);
        smp_free_lowcore(cpu);
-       printk(KERN_INFO "Processor %d spun down\n", cpu);
+       pr_info("Processor %d stopped\n", cpu);
 }
 
 void cpu_die(void)
 {
        idle_task_exit();
-       signal_processor(smp_processor_id(), sigp_stop);
-       BUG();
+       while (sigp(smp_processor_id(), sigp_stop) == sigp_busy)
+               cpu_relax();
        for (;;);
 }
 
@@ -801,20 +680,19 @@ void __init smp_prepare_cpus(unsigned int max_cpus)
        unsigned long async_stack, panic_stack;
        struct _lowcore *lowcore;
        unsigned int cpu;
-       int lc_order;
 
        smp_detect_cpus();
 
        /* request the 0x1201 emergency signal external interrupt */
        if (register_external_interrupt(0x1201, do_ext_call_interrupt) != 0)
                panic("Couldn't request external interrupt 0x1201");
-       print_cpu_info(&S390_lowcore.cpu_data);
+       print_cpu_info();
 
        /* Reallocate current lowcore, but keep its contents. */
-       lc_order = sizeof(long) == 8 ? 1 : 0;
-       lowcore = (void *) __get_free_pages(GFP_KERNEL | GFP_DMA, lc_order);
+       lowcore = (void *) __get_free_pages(GFP_KERNEL | GFP_DMA, LC_ORDER);
        panic_stack = __get_free_page(GFP_KERNEL);
        async_stack = __get_free_pages(GFP_KERNEL, ASYNC_ORDER);
+       BUG_ON(!lowcore || !panic_stack || !async_stack);
 #ifndef CONFIG_64BIT
        if (MACHINE_HAS_IEEE)
                save_area = get_zeroed_page(GFP_KERNEL);
@@ -832,6 +710,10 @@ void __init smp_prepare_cpus(unsigned int max_cpus)
        set_prefix((u32)(unsigned long) lowcore);
        local_mcck_enable();
        local_irq_enable();
+#ifdef CONFIG_64BIT
+       if (vdso_alloc_per_cpu(smp_processor_id(), &S390_lowcore))
+               BUG();
+#endif
        for_each_possible_cpu(cpu)
                if (cpu != smp_processor_id())
                        smp_create_idle(cpu);
@@ -848,13 +730,18 @@ void __init smp_prepare_boot_cpu(void)
        current_set[0] = current;
        smp_cpu_state[0] = CPU_STATE_CONFIGURED;
        smp_cpu_polarization[0] = POLARIZATION_UNKNWN;
-       spin_lock_init(&(&__get_cpu_var(s390_idle))->lock);
 }
 
 void __init smp_cpus_done(unsigned int max_cpus)
 {
 }
 
+void __init smp_setup_processor_id(void)
+{
+       S390_lowcore.cpu_nr = 0;
+       __cpu_logical_map[0] = stap();
+}
+
 /*
  * the frequency of the profiling timer can be changed
  * by writing a multiplier value into /proc/profile.
@@ -867,7 +754,8 @@ int setup_profiling_timer(unsigned int multiplier)
 }
 
 #ifdef CONFIG_HOTPLUG_CPU
-static ssize_t cpu_configure_show(struct sys_device *dev, char *buf)
+static ssize_t cpu_configure_show(struct sys_device *dev,
+                               struct sysdev_attribute *attr, char *buf)
 {
        ssize_t count;
 
@@ -877,8 +765,9 @@ static ssize_t cpu_configure_show(struct sys_device *dev, char *buf)
        return count;
 }
 
-static ssize_t cpu_configure_store(struct sys_device *dev, const char *buf,
-                                  size_t count)
+static ssize_t cpu_configure_store(struct sys_device *dev,
+                                 struct sysdev_attribute *attr,
+                                 const char *buf, size_t count)
 {
        int cpu = dev->id;
        int val, rc;
@@ -889,10 +778,11 @@ static ssize_t cpu_configure_store(struct sys_device *dev, const char *buf,
        if (val != 0 && val != 1)
                return -EINVAL;
 
-       mutex_lock(&smp_cpu_state_mutex);
        get_online_cpus();
+       mutex_lock(&smp_cpu_state_mutex);
        rc = -EBUSY;
-       if (cpu_online(cpu))
+       /* disallow configuration changes of online cpus and cpu 0 */
+       if (cpu_online(cpu) || cpu == 0)
                goto out;
        rc = 0;
        switch (val) {
@@ -918,14 +808,15 @@ static ssize_t cpu_configure_store(struct sys_device *dev, const char *buf,
                break;
        }
 out:
-       put_online_cpus();
        mutex_unlock(&smp_cpu_state_mutex);
+       put_online_cpus();
        return rc ? rc : count;
 }
 static SYSDEV_ATTR(configure, 0644, cpu_configure_show, cpu_configure_store);
 #endif /* CONFIG_HOTPLUG_CPU */
 
-static ssize_t cpu_polarization_show(struct sys_device *dev, char *buf)
+static ssize_t cpu_polarization_show(struct sys_device *dev,
+                                    struct sysdev_attribute *attr, char *buf)
 {
        int cpu = dev->id;
        ssize_t count;
@@ -953,7 +844,8 @@ static ssize_t cpu_polarization_show(struct sys_device *dev, char *buf)
 }
 static SYSDEV_ATTR(polarization, 0444, cpu_polarization_show, NULL);
 
-static ssize_t show_cpu_address(struct sys_device *dev, char *buf)
+static ssize_t show_cpu_address(struct sys_device *dev,
+                               struct sysdev_attribute *attr, char *buf)
 {
        return sprintf(buf, "%d\n", __cpu_logical_map[dev->id]);
 }
@@ -973,7 +865,8 @@ static struct attribute_group cpu_common_attr_group = {
        .attrs = cpu_common_attrs,
 };
 
-static ssize_t show_capability(struct sys_device *dev, char *buf)
+static ssize_t show_capability(struct sys_device *dev,
+                               struct sysdev_attribute *attr, char *buf)
 {
        unsigned int capability;
        int rc;
@@ -985,34 +878,51 @@ static ssize_t show_capability(struct sys_device *dev, char *buf)
 }
 static SYSDEV_ATTR(capability, 0444, show_capability, NULL);
 
-static ssize_t show_idle_count(struct sys_device *dev, char *buf)
+static ssize_t show_idle_count(struct sys_device *dev,
+                               struct sysdev_attribute *attr, char *buf)
 {
        struct s390_idle_data *idle;
        unsigned long long idle_count;
+       unsigned int sequence;
 
        idle = &per_cpu(s390_idle, dev->id);
-       spin_lock_irq(&idle->lock);
+repeat:
+       sequence = idle->sequence;
+       smp_rmb();
+       if (sequence & 1)
+               goto repeat;
        idle_count = idle->idle_count;
-       spin_unlock_irq(&idle->lock);
+       if (idle->idle_enter)
+               idle_count++;
+       smp_rmb();
+       if (idle->sequence != sequence)
+               goto repeat;
        return sprintf(buf, "%llu\n", idle_count);
 }
 static SYSDEV_ATTR(idle_count, 0444, show_idle_count, NULL);
 
-static ssize_t show_idle_time(struct sys_device *dev, char *buf)
+static ssize_t show_idle_time(struct sys_device *dev,
+                               struct sysdev_attribute *attr, char *buf)
 {
        struct s390_idle_data *idle;
-       unsigned long long new_time;
+       unsigned long long now, idle_time, idle_enter;
+       unsigned int sequence;
 
        idle = &per_cpu(s390_idle, dev->id);
-       spin_lock_irq(&idle->lock);
-       if (idle->in_idle) {
-               new_time = get_clock();
-               idle->idle_time += new_time - idle->idle_enter;
-               idle->idle_enter = new_time;
-       }
-       new_time = idle->idle_time;
-       spin_unlock_irq(&idle->lock);
-       return sprintf(buf, "%llu\n", new_time >> 12);
+       now = get_clock();
+repeat:
+       sequence = idle->sequence;
+       smp_rmb();
+       if (sequence & 1)
+               goto repeat;
+       idle_time = idle->idle_time;
+       idle_enter = idle->idle_enter;
+       if (idle_enter != 0ULL && idle_enter < now)
+               idle_time += now - idle_enter;
+       smp_rmb();
+       if (idle->sequence != sequence)
+               goto repeat;
+       return sprintf(buf, "%llu\n", idle_time >> 12);
 }
 static SYSDEV_ATTR(idle_time_us, 0444, show_idle_time, NULL);
 
@@ -1039,11 +949,7 @@ static int __cpuinit smp_cpu_notify(struct notifier_block *self,
        case CPU_ONLINE:
        case CPU_ONLINE_FROZEN:
                idle = &per_cpu(s390_idle, cpu);
-               spin_lock_irq(&idle->lock);
-               idle->idle_enter = 0;
-               idle->idle_time = 0;
-               idle->idle_count = 0;
-               spin_unlock_irq(&idle->lock);
+               memset(idle, 0, sizeof(struct s390_idle_data));
                if (sysfs_create_group(&s->kobj, &cpu_online_attr_group))
                        return NOTIFY_BAD;
                break;
@@ -1087,17 +993,17 @@ out:
 }
 
 #ifdef CONFIG_HOTPLUG_CPU
-static ssize_t __ref rescan_store(struct sys_device *dev,
-                                 const char *buf, size_t count)
+
+int __ref smp_rescan_cpus(void)
 {
        cpumask_t newcpus;
        int cpu;
        int rc;
 
-       mutex_lock(&smp_cpu_state_mutex);
        get_online_cpus();
+       mutex_lock(&smp_cpu_state_mutex);
        newcpus = cpu_present_map;
-       rc = smp_rescan_cpus();
+       rc = __smp_rescan_cpus();
        if (rc)
                goto out;
        cpus_andnot(newcpus, cpu_present_map, newcpus);
@@ -1108,16 +1014,29 @@ static ssize_t __ref rescan_store(struct sys_device *dev,
        }
        rc = 0;
 out:
-       put_online_cpus();
        mutex_unlock(&smp_cpu_state_mutex);
+       put_online_cpus();
        if (!cpus_empty(newcpus))
                topology_schedule_update();
+       return rc;
+}
+
+static ssize_t __ref rescan_store(struct sysdev_class *class,
+                                 struct sysdev_class_attribute *attr,
+                                 const char *buf,
+                                 size_t count)
+{
+       int rc;
+
+       rc = smp_rescan_cpus();
        return rc ? rc : count;
 }
-static SYSDEV_ATTR(rescan, 0200, NULL, rescan_store);
+static SYSDEV_CLASS_ATTR(rescan, 0200, NULL, rescan_store);
 #endif /* CONFIG_HOTPLUG_CPU */
 
-static ssize_t dispatching_show(struct sys_device *dev, char *buf)
+static ssize_t dispatching_show(struct sysdev_class *class,
+                               struct sysdev_class_attribute *attr,
+                               char *buf)
 {
        ssize_t count;
 
@@ -1127,7 +1046,9 @@ static ssize_t dispatching_show(struct sys_device *dev, char *buf)
        return count;
 }
 
-static ssize_t dispatching_store(struct sys_device *dev, const char *buf,
+static ssize_t dispatching_store(struct sysdev_class *dev,
+                                struct sysdev_class_attribute *attr,
+                                const char *buf,
                                 size_t count)
 {
        int val, rc;
@@ -1138,19 +1059,20 @@ static ssize_t dispatching_store(struct sys_device *dev, const char *buf,
        if (val != 0 && val != 1)
                return -EINVAL;
        rc = 0;
-       mutex_lock(&smp_cpu_state_mutex);
        get_online_cpus();
+       mutex_lock(&smp_cpu_state_mutex);
        if (cpu_management == val)
                goto out;
        rc = topology_set_cpu_management(val);
        if (!rc)
                cpu_management = val;
 out:
-       put_online_cpus();
        mutex_unlock(&smp_cpu_state_mutex);
+       put_online_cpus();
        return rc ? rc : count;
 }
-static SYSDEV_ATTR(dispatching, 0644, dispatching_show, dispatching_store);
+static SYSDEV_CLASS_ATTR(dispatching, 0644, dispatching_show,
+                        dispatching_store);
 
 static int __init topology_init(void)
 {
@@ -1160,13 +1082,11 @@ static int __init topology_init(void)
        register_cpu_notifier(&smp_cpu_nb);
 
 #ifdef CONFIG_HOTPLUG_CPU
-       rc = sysfs_create_file(&cpu_sysdev_class.kset.kobj,
-                              &attr_rescan.attr);
+       rc = sysdev_class_create_file(&cpu_sysdev_class, &attr_rescan);
        if (rc)
                return rc;
 #endif
-       rc = sysfs_create_file(&cpu_sysdev_class.kset.kobj,
-                              &attr_dispatching.attr);
+       rc = sysdev_class_create_file(&cpu_sysdev_class, &attr_dispatching);
        if (rc)
                return rc;
        for_each_present_cpu(cpu) {