]> xenbits.xensource.com Git - people/ssmith/netchannel2-pvops.bak.git/.git/commitdiff
Merge branch 'warnings/bug' into auto-warnings-next
authorIngo Molnar <mingo@elte.hu>
Wed, 6 May 2009 06:07:39 +0000 (08:07 +0200)
committerIngo Molnar <mingo@elte.hu>
Wed, 6 May 2009 06:07:39 +0000 (08:07 +0200)
16 files changed:
1  2 
arch/x86/kernel/apic/x2apic_uv_x.c
block/cfq-iosched.c
drivers/ata/pata_atiixp.c
drivers/hwmon/adt7473.c
drivers/hwmon/i5k_amb.c
drivers/memstick/core/mspro_block.c
drivers/net/sky2.c
fs/dcache.c
fs/udf/truncate.c
include/asm-generic/bug.h
kernel/power/snapshot.c
mm/bootmem.c
mm/mempolicy.c
net/packet/af_packet.c
net/rfkill/rfkill.c
net/sunrpc/svcauth_unix.c

index 2bda693529762b71f30ae2c26fa08a7692ca5cb5,0000000000000000000000000000000000000000..057ea232847b0c17346169b966d1352593502689
mode 100644,000000..100644
--- /dev/null
@@@ -1,669 -1,0 +1,670 @@@
-       BUG();
 +/*
 + * This file is subject to the terms and conditions of the GNU General Public
 + * License.  See the file "COPYING" in the main directory of this archive
 + * for more details.
 + *
 + * SGI UV APIC functions (note: not an Intel compatible APIC)
 + *
 + * Copyright (C) 2007-2008 Silicon Graphics, Inc. All rights reserved.
 + */
 +#include <linux/cpumask.h>
 +#include <linux/hardirq.h>
 +#include <linux/proc_fs.h>
 +#include <linux/threads.h>
 +#include <linux/kernel.h>
 +#include <linux/module.h>
 +#include <linux/string.h>
 +#include <linux/ctype.h>
 +#include <linux/sched.h>
 +#include <linux/timer.h>
 +#include <linux/cpu.h>
 +#include <linux/init.h>
 +#include <linux/io.h>
 +
 +#include <asm/uv/uv_mmrs.h>
 +#include <asm/uv/uv_hub.h>
 +#include <asm/current.h>
 +#include <asm/pgtable.h>
 +#include <asm/uv/bios.h>
 +#include <asm/uv/uv.h>
 +#include <asm/apic.h>
 +#include <asm/ipi.h>
 +#include <asm/smp.h>
 +
 +DEFINE_PER_CPU(int, x2apic_extra_bits);
 +
 +static enum uv_system_type uv_system_type;
 +
 +static int early_get_nodeid(void)
 +{
 +      union uvh_node_id_u node_id;
 +      unsigned long *mmr;
 +
 +      mmr = early_ioremap(UV_LOCAL_MMR_BASE | UVH_NODE_ID, sizeof(*mmr));
 +      node_id.v = *mmr;
 +      early_iounmap(mmr, sizeof(*mmr));
 +      return node_id.s.node_id;
 +}
 +
 +static int uv_acpi_madt_oem_check(char *oem_id, char *oem_table_id)
 +{
 +      if (!strcmp(oem_id, "SGI")) {
 +              if (!strcmp(oem_table_id, "UVL"))
 +                      uv_system_type = UV_LEGACY_APIC;
 +              else if (!strcmp(oem_table_id, "UVX"))
 +                      uv_system_type = UV_X2APIC;
 +              else if (!strcmp(oem_table_id, "UVH")) {
 +                      __get_cpu_var(x2apic_extra_bits) =
 +                              early_get_nodeid() << (UV_APIC_PNODE_SHIFT - 1);
 +                      uv_system_type = UV_NON_UNIQUE_APIC;
 +                      return 1;
 +              }
 +      }
 +      return 0;
 +}
 +
 +enum uv_system_type get_uv_system_type(void)
 +{
 +      return uv_system_type;
 +}
 +
 +int is_uv_system(void)
 +{
 +      return uv_system_type != UV_NONE;
 +}
 +EXPORT_SYMBOL_GPL(is_uv_system);
 +
 +DEFINE_PER_CPU(struct uv_hub_info_s, __uv_hub_info);
 +EXPORT_PER_CPU_SYMBOL_GPL(__uv_hub_info);
 +
 +struct uv_blade_info *uv_blade_info;
 +EXPORT_SYMBOL_GPL(uv_blade_info);
 +
 +short *uv_node_to_blade;
 +EXPORT_SYMBOL_GPL(uv_node_to_blade);
 +
 +short *uv_cpu_to_blade;
 +EXPORT_SYMBOL_GPL(uv_cpu_to_blade);
 +
 +short uv_possible_blades;
 +EXPORT_SYMBOL_GPL(uv_possible_blades);
 +
 +unsigned long sn_rtc_cycles_per_second;
 +EXPORT_SYMBOL(sn_rtc_cycles_per_second);
 +
 +/* Start with all IRQs pointing to boot CPU.  IRQ balancing will shift them. */
 +
 +static const struct cpumask *uv_target_cpus(void)
 +{
 +      return cpumask_of(0);
 +}
 +
 +static void uv_vector_allocation_domain(int cpu, struct cpumask *retmask)
 +{
 +      cpumask_clear(retmask);
 +      cpumask_set_cpu(cpu, retmask);
 +}
 +
 +static int uv_wakeup_secondary(int phys_apicid, unsigned long start_rip)
 +{
 +#ifdef CONFIG_SMP
 +      unsigned long val;
 +      int pnode;
 +
 +      pnode = uv_apicid_to_pnode(phys_apicid);
 +      val = (1UL << UVH_IPI_INT_SEND_SHFT) |
 +          (phys_apicid << UVH_IPI_INT_APIC_ID_SHFT) |
 +          ((start_rip << UVH_IPI_INT_VECTOR_SHFT) >> 12) |
 +          APIC_DM_INIT;
 +      uv_write_global_mmr64(pnode, UVH_IPI_INT, val);
 +      mdelay(10);
 +
 +      val = (1UL << UVH_IPI_INT_SEND_SHFT) |
 +          (phys_apicid << UVH_IPI_INT_APIC_ID_SHFT) |
 +          ((start_rip << UVH_IPI_INT_VECTOR_SHFT) >> 12) |
 +          APIC_DM_STARTUP;
 +      uv_write_global_mmr64(pnode, UVH_IPI_INT, val);
 +
 +      atomic_set(&init_deasserted, 1);
 +#endif
 +      return 0;
 +}
 +
 +static void uv_send_IPI_one(int cpu, int vector)
 +{
 +      unsigned long apicid;
 +      int pnode;
 +
 +      apicid = per_cpu(x86_cpu_to_apicid, cpu);
 +      pnode = uv_apicid_to_pnode(apicid);
 +      uv_hub_send_ipi(pnode, apicid, vector);
 +}
 +
 +static void uv_send_IPI_mask(const struct cpumask *mask, int vector)
 +{
 +      unsigned int cpu;
 +
 +      for_each_cpu(cpu, mask)
 +              uv_send_IPI_one(cpu, vector);
 +}
 +
 +static void uv_send_IPI_mask_allbutself(const struct cpumask *mask, int vector)
 +{
 +      unsigned int this_cpu = smp_processor_id();
 +      unsigned int cpu;
 +
 +      for_each_cpu(cpu, mask) {
 +              if (cpu != this_cpu)
 +                      uv_send_IPI_one(cpu, vector);
 +      }
 +}
 +
 +static void uv_send_IPI_allbutself(int vector)
 +{
 +      unsigned int this_cpu = smp_processor_id();
 +      unsigned int cpu;
 +
 +      for_each_online_cpu(cpu) {
 +              if (cpu != this_cpu)
 +                      uv_send_IPI_one(cpu, vector);
 +      }
 +}
 +
 +static void uv_send_IPI_all(int vector)
 +{
 +      uv_send_IPI_mask(cpu_online_mask, vector);
 +}
 +
 +static int uv_apic_id_registered(void)
 +{
 +      return 1;
 +}
 +
 +static void uv_init_apic_ldr(void)
 +{
 +}
 +
 +static unsigned int uv_cpu_mask_to_apicid(const struct cpumask *cpumask)
 +{
 +      /*
 +       * We're using fixed IRQ delivery, can only return one phys APIC ID.
 +       * May as well be the first.
 +       */
 +      int cpu = cpumask_first(cpumask);
 +
 +      if ((unsigned)cpu < nr_cpu_ids)
 +              return per_cpu(x86_cpu_to_apicid, cpu);
 +      else
 +              return BAD_APICID;
 +}
 +
 +static unsigned int
 +uv_cpu_mask_to_apicid_and(const struct cpumask *cpumask,
 +                        const struct cpumask *andmask)
 +{
 +      int cpu;
 +
 +      /*
 +       * We're using fixed IRQ delivery, can only return one phys APIC ID.
 +       * May as well be the first.
 +       */
 +      for_each_cpu_and(cpu, cpumask, andmask) {
 +              if (cpumask_test_cpu(cpu, cpu_online_mask))
 +                      break;
 +      }
 +      if (cpu < nr_cpu_ids)
 +              return per_cpu(x86_cpu_to_apicid, cpu);
 +
 +      return BAD_APICID;
 +}
 +
 +static unsigned int x2apic_get_apic_id(unsigned long x)
 +{
 +      unsigned int id;
 +
 +      WARN_ON(preemptible() && num_online_cpus() > 1);
 +      id = x | __get_cpu_var(x2apic_extra_bits);
 +
 +      return id;
 +}
 +
 +static unsigned long set_apic_id(unsigned int id)
 +{
 +      unsigned long x;
 +
 +      /* maskout x2apic_extra_bits ? */
 +      x = id;
 +      return x;
 +}
 +
 +static unsigned int uv_read_apic_id(void)
 +{
 +
 +      return x2apic_get_apic_id(apic_read(APIC_ID));
 +}
 +
 +static int uv_phys_pkg_id(int initial_apicid, int index_msb)
 +{
 +      return uv_read_apic_id() >> index_msb;
 +}
 +
 +static void uv_send_IPI_self(int vector)
 +{
 +      apic_write(APIC_SELF_IPI, vector);
 +}
 +
 +struct apic apic_x2apic_uv_x = {
 +
 +      .name                           = "UV large system",
 +      .probe                          = NULL,
 +      .acpi_madt_oem_check            = uv_acpi_madt_oem_check,
 +      .apic_id_registered             = uv_apic_id_registered,
 +
 +      .irq_delivery_mode              = dest_Fixed,
 +      .irq_dest_mode                  = 1, /* logical */
 +
 +      .target_cpus                    = uv_target_cpus,
 +      .disable_esr                    = 0,
 +      .dest_logical                   = APIC_DEST_LOGICAL,
 +      .check_apicid_used              = NULL,
 +      .check_apicid_present           = NULL,
 +
 +      .vector_allocation_domain       = uv_vector_allocation_domain,
 +      .init_apic_ldr                  = uv_init_apic_ldr,
 +
 +      .ioapic_phys_id_map             = NULL,
 +      .setup_apic_routing             = NULL,
 +      .multi_timer_check              = NULL,
 +      .apicid_to_node                 = NULL,
 +      .cpu_to_logical_apicid          = NULL,
 +      .cpu_present_to_apicid          = default_cpu_present_to_apicid,
 +      .apicid_to_cpu_present          = NULL,
 +      .setup_portio_remap             = NULL,
 +      .check_phys_apicid_present      = default_check_phys_apicid_present,
 +      .enable_apic_mode               = NULL,
 +      .phys_pkg_id                    = uv_phys_pkg_id,
 +      .mps_oem_check                  = NULL,
 +
 +      .get_apic_id                    = x2apic_get_apic_id,
 +      .set_apic_id                    = set_apic_id,
 +      .apic_id_mask                   = 0xFFFFFFFFu,
 +
 +      .cpu_mask_to_apicid             = uv_cpu_mask_to_apicid,
 +      .cpu_mask_to_apicid_and         = uv_cpu_mask_to_apicid_and,
 +
 +      .send_IPI_mask                  = uv_send_IPI_mask,
 +      .send_IPI_mask_allbutself       = uv_send_IPI_mask_allbutself,
 +      .send_IPI_allbutself            = uv_send_IPI_allbutself,
 +      .send_IPI_all                   = uv_send_IPI_all,
 +      .send_IPI_self                  = uv_send_IPI_self,
 +
 +      .wakeup_secondary_cpu           = uv_wakeup_secondary,
 +      .trampoline_phys_low            = DEFAULT_TRAMPOLINE_PHYS_LOW,
 +      .trampoline_phys_high           = DEFAULT_TRAMPOLINE_PHYS_HIGH,
 +      .wait_for_init_deassert         = NULL,
 +      .smp_callin_clear_local_apic    = NULL,
 +      .inquire_remote_apic            = NULL,
 +
 +      .read                           = native_apic_msr_read,
 +      .write                          = native_apic_msr_write,
 +      .icr_read                       = native_x2apic_icr_read,
 +      .icr_write                      = native_x2apic_icr_write,
 +      .wait_icr_idle                  = native_x2apic_wait_icr_idle,
 +      .safe_wait_icr_idle             = native_safe_x2apic_wait_icr_idle,
 +};
 +
 +static __cpuinit void set_x2apic_extra_bits(int pnode)
 +{
 +      __get_cpu_var(x2apic_extra_bits) = (pnode << 6);
 +}
 +
 +/*
 + * Called on boot cpu.
 + */
 +static __init int boot_pnode_to_blade(int pnode)
 +{
 +      int blade;
 +
 +      for (blade = 0; blade < uv_num_possible_blades(); blade++)
 +              if (pnode == uv_blade_info[blade].pnode)
 +                      return blade;
-       BUG();
++
++      panic("x2apic_uv: bad pnode!");
 +}
 +
 +struct redir_addr {
 +      unsigned long redirect;
 +      unsigned long alias;
 +};
 +
 +#define DEST_SHIFT UVH_RH_GAM_ALIAS210_REDIRECT_CONFIG_0_MMR_DEST_BASE_SHFT
 +
 +static __initdata struct redir_addr redir_addrs[] = {
 +      {UVH_RH_GAM_ALIAS210_REDIRECT_CONFIG_0_MMR, UVH_SI_ALIAS0_OVERLAY_CONFIG},
 +      {UVH_RH_GAM_ALIAS210_REDIRECT_CONFIG_1_MMR, UVH_SI_ALIAS1_OVERLAY_CONFIG},
 +      {UVH_RH_GAM_ALIAS210_REDIRECT_CONFIG_2_MMR, UVH_SI_ALIAS2_OVERLAY_CONFIG},
 +};
 +
 +static __init void get_lowmem_redirect(unsigned long *base, unsigned long *size)
 +{
 +      union uvh_si_alias0_overlay_config_u alias;
 +      union uvh_rh_gam_alias210_redirect_config_2_mmr_u redirect;
 +      int i;
 +
 +      for (i = 0; i < ARRAY_SIZE(redir_addrs); i++) {
 +              alias.v = uv_read_local_mmr(redir_addrs[i].alias);
 +              if (alias.s.base == 0) {
 +                      *size = (1UL << alias.s.m_alias);
 +                      redirect.v = uv_read_local_mmr(redir_addrs[i].redirect);
 +                      *base = (unsigned long)redirect.s.dest_base << DEST_SHIFT;
 +                      return;
 +              }
 +      }
++      panic("get_lowmem_redirect: no match!");
 +}
 +
 +static __init void map_low_mmrs(void)
 +{
 +      init_extra_mapping_uc(UV_GLOBAL_MMR32_BASE, UV_GLOBAL_MMR32_SIZE);
 +      init_extra_mapping_uc(UV_LOCAL_MMR_BASE, UV_LOCAL_MMR_SIZE);
 +}
 +
 +enum map_type {map_wb, map_uc};
 +
 +static __init void map_high(char *id, unsigned long base, int shift,
 +                          int max_pnode, enum map_type map_type)
 +{
 +      unsigned long bytes, paddr;
 +
 +      paddr = base << shift;
 +      bytes = (1UL << shift) * (max_pnode + 1);
 +      printk(KERN_INFO "UV: Map %s_HI 0x%lx - 0x%lx\n", id, paddr,
 +                                              paddr + bytes);
 +      if (map_type == map_uc)
 +              init_extra_mapping_uc(paddr, bytes);
 +      else
 +              init_extra_mapping_wb(paddr, bytes);
 +
 +}
 +static __init void map_gru_high(int max_pnode)
 +{
 +      union uvh_rh_gam_gru_overlay_config_mmr_u gru;
 +      int shift = UVH_RH_GAM_GRU_OVERLAY_CONFIG_MMR_BASE_SHFT;
 +
 +      gru.v = uv_read_local_mmr(UVH_RH_GAM_GRU_OVERLAY_CONFIG_MMR);
 +      if (gru.s.enable)
 +              map_high("GRU", gru.s.base, shift, max_pnode, map_wb);
 +}
 +
 +static __init void map_config_high(int max_pnode)
 +{
 +      union uvh_rh_gam_cfg_overlay_config_mmr_u cfg;
 +      int shift = UVH_RH_GAM_CFG_OVERLAY_CONFIG_MMR_BASE_SHFT;
 +
 +      cfg.v = uv_read_local_mmr(UVH_RH_GAM_CFG_OVERLAY_CONFIG_MMR);
 +      if (cfg.s.enable)
 +              map_high("CONFIG", cfg.s.base, shift, max_pnode, map_uc);
 +}
 +
 +static __init void map_mmr_high(int max_pnode)
 +{
 +      union uvh_rh_gam_mmr_overlay_config_mmr_u mmr;
 +      int shift = UVH_RH_GAM_MMR_OVERLAY_CONFIG_MMR_BASE_SHFT;
 +
 +      mmr.v = uv_read_local_mmr(UVH_RH_GAM_MMR_OVERLAY_CONFIG_MMR);
 +      if (mmr.s.enable)
 +              map_high("MMR", mmr.s.base, shift, max_pnode, map_uc);
 +}
 +
 +static __init void map_mmioh_high(int max_pnode)
 +{
 +      union uvh_rh_gam_mmioh_overlay_config_mmr_u mmioh;
 +      int shift = UVH_RH_GAM_MMIOH_OVERLAY_CONFIG_MMR_BASE_SHFT;
 +
 +      mmioh.v = uv_read_local_mmr(UVH_RH_GAM_MMIOH_OVERLAY_CONFIG_MMR);
 +      if (mmioh.s.enable)
 +              map_high("MMIOH", mmioh.s.base, shift, max_pnode, map_uc);
 +}
 +
 +static __init void uv_rtc_init(void)
 +{
 +      long status;
 +      u64 ticks_per_sec;
 +
 +      status = uv_bios_freq_base(BIOS_FREQ_BASE_REALTIME_CLOCK,
 +                                      &ticks_per_sec);
 +      if (status != BIOS_STATUS_SUCCESS || ticks_per_sec < 100000) {
 +              printk(KERN_WARNING
 +                      "unable to determine platform RTC clock frequency, "
 +                      "guessing.\n");
 +              /* BIOS gives wrong value for clock freq. so guess */
 +              sn_rtc_cycles_per_second = 1000000000000UL / 30000UL;
 +      } else
 +              sn_rtc_cycles_per_second = ticks_per_sec;
 +}
 +
 +/*
 + * percpu heartbeat timer
 + */
 +static void uv_heartbeat(unsigned long ignored)
 +{
 +      struct timer_list *timer = &uv_hub_info->scir.timer;
 +      unsigned char bits = uv_hub_info->scir.state;
 +
 +      /* flip heartbeat bit */
 +      bits ^= SCIR_CPU_HEARTBEAT;
 +
 +      /* is this cpu idle? */
 +      if (idle_cpu(raw_smp_processor_id()))
 +              bits &= ~SCIR_CPU_ACTIVITY;
 +      else
 +              bits |= SCIR_CPU_ACTIVITY;
 +
 +      /* update system controller interface reg */
 +      uv_set_scir_bits(bits);
 +
 +      /* enable next timer period */
 +      mod_timer(timer, jiffies + SCIR_CPU_HB_INTERVAL);
 +}
 +
 +static void __cpuinit uv_heartbeat_enable(int cpu)
 +{
 +      if (!uv_cpu_hub_info(cpu)->scir.enabled) {
 +              struct timer_list *timer = &uv_cpu_hub_info(cpu)->scir.timer;
 +
 +              uv_set_cpu_scir_bits(cpu, SCIR_CPU_HEARTBEAT|SCIR_CPU_ACTIVITY);
 +              setup_timer(timer, uv_heartbeat, cpu);
 +              timer->expires = jiffies + SCIR_CPU_HB_INTERVAL;
 +              add_timer_on(timer, cpu);
 +              uv_cpu_hub_info(cpu)->scir.enabled = 1;
 +      }
 +
 +      /* check boot cpu */
 +      if (!uv_cpu_hub_info(0)->scir.enabled)
 +              uv_heartbeat_enable(0);
 +}
 +
 +#ifdef CONFIG_HOTPLUG_CPU
 +static void __cpuinit uv_heartbeat_disable(int cpu)
 +{
 +      if (uv_cpu_hub_info(cpu)->scir.enabled) {
 +              uv_cpu_hub_info(cpu)->scir.enabled = 0;
 +              del_timer(&uv_cpu_hub_info(cpu)->scir.timer);
 +      }
 +      uv_set_cpu_scir_bits(cpu, 0xff);
 +}
 +
 +/*
 + * cpu hotplug notifier
 + */
 +static __cpuinit int uv_scir_cpu_notify(struct notifier_block *self,
 +                                     unsigned long action, void *hcpu)
 +{
 +      long cpu = (long)hcpu;
 +
 +      switch (action) {
 +      case CPU_ONLINE:
 +              uv_heartbeat_enable(cpu);
 +              break;
 +      case CPU_DOWN_PREPARE:
 +              uv_heartbeat_disable(cpu);
 +              break;
 +      default:
 +              break;
 +      }
 +      return NOTIFY_OK;
 +}
 +
 +static __init void uv_scir_register_cpu_notifier(void)
 +{
 +      hotcpu_notifier(uv_scir_cpu_notify, 0);
 +}
 +
 +#else /* !CONFIG_HOTPLUG_CPU */
 +
 +static __init void uv_scir_register_cpu_notifier(void)
 +{
 +}
 +
 +static __init int uv_init_heartbeat(void)
 +{
 +      int cpu;
 +
 +      if (is_uv_system())
 +              for_each_online_cpu(cpu)
 +                      uv_heartbeat_enable(cpu);
 +      return 0;
 +}
 +
 +late_initcall(uv_init_heartbeat);
 +
 +#endif /* !CONFIG_HOTPLUG_CPU */
 +
 +/*
 + * Called on each cpu to initialize the per_cpu UV data area.
 + * FIXME: hotplug not supported yet
 + */
 +void __cpuinit uv_cpu_init(void)
 +{
 +      /* CPU 0 initilization will be done via uv_system_init. */
 +      if (!uv_blade_info)
 +              return;
 +
 +      uv_blade_info[uv_numa_blade_id()].nr_online_cpus++;
 +
 +      if (get_uv_system_type() == UV_NON_UNIQUE_APIC)
 +              set_x2apic_extra_bits(uv_hub_info->pnode);
 +}
 +
 +
 +void __init uv_system_init(void)
 +{
 +      union uvh_si_addr_map_config_u m_n_config;
 +      union uvh_node_id_u node_id;
 +      unsigned long gnode_upper, lowmem_redir_base, lowmem_redir_size;
 +      int bytes, nid, cpu, lcpu, pnode, blade, i, j, m_val, n_val;
 +      int max_pnode = 0;
 +      unsigned long mmr_base, present, paddr;
 +      unsigned short pnode_mask;
 +
 +      map_low_mmrs();
 +
 +      m_n_config.v = uv_read_local_mmr(UVH_SI_ADDR_MAP_CONFIG);
 +      m_val = m_n_config.s.m_skt;
 +      n_val = m_n_config.s.n_skt;
 +      mmr_base =
 +          uv_read_local_mmr(UVH_RH_GAM_MMR_OVERLAY_CONFIG_MMR) &
 +          ~UV_MMR_ENABLE;
 +      printk(KERN_DEBUG "UV: global MMR base 0x%lx\n", mmr_base);
 +
 +      for(i = 0; i < UVH_NODE_PRESENT_TABLE_DEPTH; i++)
 +              uv_possible_blades +=
 +                hweight64(uv_read_local_mmr( UVH_NODE_PRESENT_TABLE + i * 8));
 +      printk(KERN_DEBUG "UV: Found %d blades\n", uv_num_possible_blades());
 +
 +      bytes = sizeof(struct uv_blade_info) * uv_num_possible_blades();
 +      uv_blade_info = kmalloc(bytes, GFP_KERNEL);
 +
 +      get_lowmem_redirect(&lowmem_redir_base, &lowmem_redir_size);
 +
 +      bytes = sizeof(uv_node_to_blade[0]) * num_possible_nodes();
 +      uv_node_to_blade = kmalloc(bytes, GFP_KERNEL);
 +      memset(uv_node_to_blade, 255, bytes);
 +
 +      bytes = sizeof(uv_cpu_to_blade[0]) * num_possible_cpus();
 +      uv_cpu_to_blade = kmalloc(bytes, GFP_KERNEL);
 +      memset(uv_cpu_to_blade, 255, bytes);
 +
 +      blade = 0;
 +      for (i = 0; i < UVH_NODE_PRESENT_TABLE_DEPTH; i++) {
 +              present = uv_read_local_mmr(UVH_NODE_PRESENT_TABLE + i * 8);
 +              for (j = 0; j < 64; j++) {
 +                      if (!test_bit(j, &present))
 +                              continue;
 +                      uv_blade_info[blade].pnode = (i * 64 + j);
 +                      uv_blade_info[blade].nr_possible_cpus = 0;
 +                      uv_blade_info[blade].nr_online_cpus = 0;
 +                      blade++;
 +              }
 +      }
 +
 +      pnode_mask = (1 << n_val) - 1;
 +      node_id.v = uv_read_local_mmr(UVH_NODE_ID);
 +      gnode_upper = (((unsigned long)node_id.s.node_id) &
 +                     ~((1 << n_val) - 1)) << m_val;
 +
 +      uv_bios_init();
 +      uv_bios_get_sn_info(0, &uv_type, &sn_partition_id,
 +                          &sn_coherency_id, &sn_region_size);
 +      uv_rtc_init();
 +
 +      for_each_present_cpu(cpu) {
 +              nid = cpu_to_node(cpu);
 +              pnode = uv_apicid_to_pnode(per_cpu(x86_cpu_to_apicid, cpu));
 +              blade = boot_pnode_to_blade(pnode);
 +              lcpu = uv_blade_info[blade].nr_possible_cpus;
 +              uv_blade_info[blade].nr_possible_cpus++;
 +
 +              uv_cpu_hub_info(cpu)->lowmem_remap_base = lowmem_redir_base;
 +              uv_cpu_hub_info(cpu)->lowmem_remap_top = lowmem_redir_size;
 +              uv_cpu_hub_info(cpu)->m_val = m_val;
 +              uv_cpu_hub_info(cpu)->n_val = m_val;
 +              uv_cpu_hub_info(cpu)->numa_blade_id = blade;
 +              uv_cpu_hub_info(cpu)->blade_processor_id = lcpu;
 +              uv_cpu_hub_info(cpu)->pnode = pnode;
 +              uv_cpu_hub_info(cpu)->pnode_mask = pnode_mask;
 +              uv_cpu_hub_info(cpu)->gpa_mask = (1 << (m_val + n_val)) - 1;
 +              uv_cpu_hub_info(cpu)->gnode_upper = gnode_upper;
 +              uv_cpu_hub_info(cpu)->global_mmr_base = mmr_base;
 +              uv_cpu_hub_info(cpu)->coherency_domain_number = sn_coherency_id;
 +              uv_cpu_hub_info(cpu)->scir.offset = SCIR_LOCAL_MMR_BASE + lcpu;
 +              uv_node_to_blade[nid] = blade;
 +              uv_cpu_to_blade[cpu] = blade;
 +              max_pnode = max(pnode, max_pnode);
 +
 +              printk(KERN_DEBUG "UV: cpu %d, apicid 0x%x, pnode %d, nid %d, "
 +                      "lcpu %d, blade %d\n",
 +                      cpu, per_cpu(x86_cpu_to_apicid, cpu), pnode, nid,
 +                      lcpu, blade);
 +      }
 +
 +      /* Add blade/pnode info for nodes without cpus */
 +      for_each_online_node(nid) {
 +              if (uv_node_to_blade[nid] >= 0)
 +                      continue;
 +              paddr = node_start_pfn(nid) << PAGE_SHIFT;
 +              paddr = uv_soc_phys_ram_to_gpa(paddr);
 +              pnode = (paddr >> m_val) & pnode_mask;
 +              blade = boot_pnode_to_blade(pnode);
 +              uv_node_to_blade[nid] = blade;
 +      }
 +
 +      map_gru_high(max_pnode);
 +      map_mmr_high(max_pnode);
 +      map_config_high(max_pnode);
 +      map_mmioh_high(max_pnode);
 +
 +      uv_cpu_init();
 +      uv_scir_register_cpu_notifier();
 +      proc_mkdir("sgi_uv", NULL);
 +}
Simple merge
Simple merge
Simple merge
Simple merge
Simple merge
Simple merge
diff --cc fs/dcache.c
Simple merge
index 225527cdc885591614e92ff6ff325c7fdd459229,08fd88c5d8d96e593c0dd7e58395c8f356699454..c17cbca466c014506ce3c4e38fc802399bcd9261
@@@ -83,11 -83,11 +83,11 @@@ void udf_truncate_tail_extent(struct in
                return;
  
        if (iinfo->i_alloc_type == ICBTAG_FLAG_AD_SHORT)
 -              adsize = sizeof(short_ad);
 +              adsize = sizeof(struct short_ad);
        else if (iinfo->i_alloc_type == ICBTAG_FLAG_AD_LONG)
 -              adsize = sizeof(long_ad);
 +              adsize = sizeof(struct long_ad);
        else
-               BUG();
+               panic("udf_truncate_tail_extent: unknown alloc type!");
  
        /* Find the last extent in the file */
        while ((netype = udf_next_aext(inode, &epos, &eloc, &elen, 1)) != -1) {
@@@ -210,11 -210,11 +210,11 @@@ void udf_truncate_extents(struct inode 
        struct udf_inode_info *iinfo = UDF_I(inode);
  
        if (iinfo->i_alloc_type == ICBTAG_FLAG_AD_SHORT)
 -              adsize = sizeof(short_ad);
 +              adsize = sizeof(struct short_ad);
        else if (iinfo->i_alloc_type == ICBTAG_FLAG_AD_LONG)
 -              adsize = sizeof(long_ad);
 +              adsize = sizeof(struct long_ad);
        else
-               BUG();
+               panic("udf_truncate_extents: unknown alloc type!");
  
        etype = inode_bmap(inode, first_block, &epos, &eloc, &elen, &offset);
        byte_offset = (offset << sb->s_blocksize_bits) +
Simple merge
Simple merge
diff --cc mm/bootmem.c
Simple merge
diff --cc mm/mempolicy.c
Simple merge
Simple merge
Simple merge
Simple merge