19 files changed, 85 insertions, 141 deletions

diff --git a/arch/x86/kernel/acpi/cstate.c b/arch/x86/kernel/acpi/cstate.c
index 9220cf46aa10..c2502eb9aa83 100644
--- a/arch/x86/kernel/acpi/cstate.c
+++ b/arch/x86/kernel/acpi/cstate.c
@@ -73,7 +73,6 @@ int acpi_processor_ffh_cstate_probe(unsigned int cpu,
         struct cpuinfo_x86 *c = &cpu_data(cpu);
 
         cpumask_t saved_mask;
-        cpumask_of_cpu_ptr(new_mask, cpu);
         int retval;
         unsigned int eax, ebx, ecx, edx;
         unsigned int edx_part;
@@ -92,7 +91,7 @@ int acpi_processor_ffh_cstate_probe(unsigned int cpu,
 
         /* Make sure we are running on right CPU */
         saved_mask = current->cpus_allowed;
-        retval = set_cpus_allowed_ptr(current, new_mask);
+        retval = set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu));
         if (retval)
                 return -1;
 
diff --git a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
index ff2fff56f0a8..dd097b835839 100644
--- a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
+++ b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
@@ -200,12 +200,10 @@ static void drv_read(struct drv_cmd *cmd)
 static void drv_write(struct drv_cmd *cmd)
 {
         cpumask_t saved_mask = current->cpus_allowed;
-        cpumask_of_cpu_ptr_declare(cpu_mask);
         unsigned int i;
 
         for_each_cpu_mask_nr(i, cmd->mask) {
-                cpumask_of_cpu_ptr_next(cpu_mask, i);
-                set_cpus_allowed_ptr(current, cpu_mask);
+                set_cpus_allowed_ptr(current, &cpumask_of_cpu(i));
                 do_drv_write(cmd);
         }
 
@@ -269,12 +267,11 @@ static unsigned int get_measured_perf(unsigned int cpu)
         } aperf_cur, mperf_cur;
 
         cpumask_t saved_mask;
-        cpumask_of_cpu_ptr(cpu_mask, cpu);
         unsigned int perf_percent;
         unsigned int retval;
 
         saved_mask = current->cpus_allowed;
-        set_cpus_allowed_ptr(current, cpu_mask);
+        set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu));
         if (get_cpu() != cpu) {
                 /* We were not able to run on requested processor */
                 put_cpu();
@@ -340,7 +337,6 @@ static unsigned int get_measured_perf(unsigned int cpu)
 
 static unsigned int get_cur_freq_on_cpu(unsigned int cpu)
 {
-        cpumask_of_cpu_ptr(cpu_mask, cpu);
         struct acpi_cpufreq_data *data = per_cpu(drv_data, cpu);
         unsigned int freq;
         unsigned int cached_freq;
@@ -353,7 +349,7 @@ static unsigned int get_cur_freq_on_cpu(unsigned int cpu)
         }
 
         cached_freq = data->freq_table[data->acpi_data->state].frequency;
-        freq = extract_freq(get_cur_val(cpu_mask), data);
+        freq = extract_freq(get_cur_val(&cpumask_of_cpu(cpu)), data);
         if (freq != cached_freq) {
                 /*
                  * The dreaded BIOS frequency change behind our back.
diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
index 53c7b6936973..c45ca6d4dce1 100644
--- a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
+++ b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
@@ -479,12 +479,11 @@ static int core_voltage_post_transition(struct powernow_k8_data *data, u32 reqvi
 static int check_supported_cpu(unsigned int cpu)
 {
         cpumask_t oldmask;
-        cpumask_of_cpu_ptr(cpu_mask, cpu);
         u32 eax, ebx, ecx, edx;
         unsigned int rc = 0;
 
         oldmask = current->cpus_allowed;
-        set_cpus_allowed_ptr(current, cpu_mask);
+        set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu));
 
         if (smp_processor_id() != cpu) {
                 printk(KERN_ERR PFX "limiting to cpu %u failed\n", cpu);
@@ -1017,7 +1016,6 @@ static int transition_frequency_pstate(struct powernow_k8_data *data, unsigned i
 static int powernowk8_target(struct cpufreq_policy *pol, unsigned targfreq, unsigned relation)
 {
         cpumask_t oldmask;
-        cpumask_of_cpu_ptr(cpu_mask, pol->cpu);
         struct powernow_k8_data *data = per_cpu(powernow_data, pol->cpu);
         u32 checkfid;
         u32 checkvid;
@@ -1032,7 +1030,7 @@ static int powernowk8_target(struct cpufreq_policy *pol, unsigned targfreq, unsi
 
         /* only run on specific CPU from here on */
         oldmask = current->cpus_allowed;
-        set_cpus_allowed_ptr(current, cpu_mask);
+        set_cpus_allowed_ptr(current, &cpumask_of_cpu(pol->cpu));
 
         if (smp_processor_id() != pol->cpu) {
                 printk(KERN_ERR PFX "limiting to cpu %u failed\n", pol->cpu);
@@ -1107,7 +1105,6 @@ static int __cpuinit powernowk8_cpu_init(struct cpufreq_policy *pol)
 {
         struct powernow_k8_data *data;
         cpumask_t oldmask;
-        cpumask_of_cpu_ptr_declare(newmask);
         int rc;
 
         if (!cpu_online(pol->cpu))
@@ -1159,8 +1156,7 @@ static int __cpuinit powernowk8_cpu_init(struct cpufreq_policy *pol)
 
         /* only run on specific CPU from here on */
         oldmask = current->cpus_allowed;
-        cpumask_of_cpu_ptr_next(newmask, pol->cpu);
-        set_cpus_allowed_ptr(current, newmask);
+        set_cpus_allowed_ptr(current, &cpumask_of_cpu(pol->cpu));
 
         if (smp_processor_id() != pol->cpu) {
                 printk(KERN_ERR PFX "limiting to cpu %u failed\n", pol->cpu);
@@ -1182,7 +1178,7 @@ static int __cpuinit powernowk8_cpu_init(struct cpufreq_policy *pol)
         set_cpus_allowed_ptr(current, &oldmask);
 
         if (cpu_family == CPU_HW_PSTATE)
-                pol->cpus = *newmask;
+                pol->cpus = cpumask_of_cpu(pol->cpu);
         else
                 pol->cpus = per_cpu(cpu_core_map, pol->cpu);
         data->available_cores = &(pol->cpus);
@@ -1248,7 +1244,6 @@ static unsigned int powernowk8_get (unsigned int cpu)
 {
         struct powernow_k8_data *data;
         cpumask_t oldmask = current->cpus_allowed;
-        cpumask_of_cpu_ptr(newmask, cpu);
         unsigned int khz = 0;
         unsigned int first;
 
@@ -1258,7 +1253,7 @@ static unsigned int powernowk8_get (unsigned int cpu)
         if (!data)
                 return -EINVAL;
 
-        set_cpus_allowed_ptr(current, newmask);
+        set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu));
         if (smp_processor_id() != cpu) {
                 printk(KERN_ERR PFX
                         "limiting to CPU %d failed in powernowk8_get\n", cpu);
diff --git a/arch/x86/kernel/cpu/cpufreq/speedstep-centrino.c b/arch/x86/kernel/cpu/cpufreq/speedstep-centrino.c
index ca2ac13b7af2..15e13c01cc36 100644
--- a/arch/x86/kernel/cpu/cpufreq/speedstep-centrino.c
+++ b/arch/x86/kernel/cpu/cpufreq/speedstep-centrino.c
@@ -324,10 +324,9 @@ static unsigned int get_cur_freq(unsigned int cpu)
         unsigned l, h;
         unsigned clock_freq;
         cpumask_t saved_mask;
-        cpumask_of_cpu_ptr(new_mask, cpu);
 
         saved_mask = current->cpus_allowed;
-        set_cpus_allowed_ptr(current, new_mask);
+        set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu));
         if (smp_processor_id() != cpu)
                 return 0;
 
@@ -585,15 +584,12 @@ static int centrino_target (struct cpufreq_policy *policy,
                  * Best effort undo..
                  */
 
-                if (!cpus_empty(*covered_cpus)) {
-                        cpumask_of_cpu_ptr_declare(new_mask);
-
+                if (!cpus_empty(*covered_cpus))
                         for_each_cpu_mask_nr(j, *covered_cpus) {
-                                cpumask_of_cpu_ptr_next(new_mask, j);
-                                set_cpus_allowed_ptr(current, new_mask);
+                                set_cpus_allowed_ptr(current,
+                                                     &cpumask_of_cpu(j));
                                 wrmsr(MSR_IA32_PERF_CTL, oldmsr, h);
                         }
-                }
 
                 tmp = freqs.new;
                 freqs.new = freqs.old;
diff --git a/arch/x86/kernel/cpu/cpufreq/speedstep-ich.c b/arch/x86/kernel/cpu/cpufreq/speedstep-ich.c
index 2f3728dc24f6..191f7263c61d 100644
--- a/arch/x86/kernel/cpu/cpufreq/speedstep-ich.c
+++ b/arch/x86/kernel/cpu/cpufreq/speedstep-ich.c
@@ -244,8 +244,7 @@ static unsigned int _speedstep_get(const cpumask_t *cpus)
 
 static unsigned int speedstep_get(unsigned int cpu)
 {
-        cpumask_of_cpu_ptr(newmask, cpu);
-        return _speedstep_get(newmask);
+        return _speedstep_get(&cpumask_of_cpu(cpu));
 }
 
 /**
diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c
index 650d40f7912b..6b0a10b002f1 100644
--- a/arch/x86/kernel/cpu/intel_cacheinfo.c
+++ b/arch/x86/kernel/cpu/intel_cacheinfo.c
@@ -516,7 +516,6 @@ static int __cpuinit detect_cache_attributes(unsigned int cpu)
         unsigned long           j;
         int                     retval;
         cpumask_t               oldmask;
-        cpumask_of_cpu_ptr(newmask, cpu);
 
         if (num_cache_leaves == 0)
                 return -ENOENT;
@@ -527,7 +526,7 @@ static int __cpuinit detect_cache_attributes(unsigned int cpu)
                 return -ENOMEM;
 
         oldmask = current->cpus_allowed;
-        retval = set_cpus_allowed_ptr(current, newmask);
+        retval = set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu));
         if (retval)
                 goto out;
 
diff --git a/arch/x86/kernel/ldt.c b/arch/x86/kernel/ldt.c
index 3fee2aa50f3f..b68e21f06f4f 100644
--- a/arch/x86/kernel/ldt.c
+++ b/arch/x86/kernel/ldt.c
@@ -62,12 +62,10 @@ static int alloc_ldt(mm_context_t *pc, int mincount, int reload)
 
         if (reload) {
 #ifdef CONFIG_SMP
-                cpumask_of_cpu_ptr_declare(mask);
-
                 preempt_disable();
                 load_LDT(pc);
-                cpumask_of_cpu_ptr_next(mask, smp_processor_id());
-                if (!cpus_equal(current->mm->cpu_vm_mask, *mask))
+                if (!cpus_equal(current->mm->cpu_vm_mask,
+                                cpumask_of_cpu(smp_processor_id())))
                         smp_call_function(flush_ldt, current->mm, 1);
                 preempt_enable();
 #else
diff --git a/arch/x86/kernel/microcode.c b/arch/x86/kernel/microcode.c
index 6994c751590e..652fa5c38ebe 100644
--- a/arch/x86/kernel/microcode.c
+++ b/arch/x86/kernel/microcode.c
@@ -388,7 +388,6 @@ static int do_microcode_update (void)
         void *new_mc = NULL;
         int cpu;
         cpumask_t old;
-        cpumask_of_cpu_ptr_declare(newmask);
 
         old = current->cpus_allowed;
 
@@ -405,8 +404,7 @@ static int do_microcode_update (void)
 
                         if (!uci->valid)
                                 continue;
-                        cpumask_of_cpu_ptr_next(newmask, cpu);
-                        set_cpus_allowed_ptr(current, newmask);
+                        set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu));
                         error = get_maching_microcode(new_mc, cpu);
                         if (error < 0)
                                 goto out;
@@ -576,7 +574,6 @@ static int apply_microcode_check_cpu(int cpu)
         struct cpuinfo_x86 *c = &cpu_data(cpu);
         struct ucode_cpu_info *uci = ucode_cpu_info + cpu;
         cpumask_t old;
-        cpumask_of_cpu_ptr(newmask, cpu);
         unsigned int val[2];
         int err = 0;
 
@@ -585,7 +582,7 @@ static int apply_microcode_check_cpu(int cpu)
                 return 0;
 
         old = current->cpus_allowed;
-        set_cpus_allowed_ptr(current, newmask);
+        set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu));
 
         /* Check if the microcode we have in memory matches the CPU */
         if (c->x86_vendor != X86_VENDOR_INTEL || c->x86 < 6 ||
@@ -623,12 +620,11 @@ static int apply_microcode_check_cpu(int cpu)
 static void microcode_init_cpu(int cpu, int resume)
 {
         cpumask_t old;
-        cpumask_of_cpu_ptr(newmask, cpu);
         struct ucode_cpu_info *uci = ucode_cpu_info + cpu;
 
         old = current->cpus_allowed;
 
-        set_cpus_allowed_ptr(current, newmask);
+        set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu));
         mutex_lock(&microcode_mutex);
         collect_cpu_info(cpu);
         if (uci->valid && system_state == SYSTEM_RUNNING && !resume)
@@ -661,13 +657,10 @@ static ssize_t reload_store(struct sys_device *dev,
         if (end == buf)
                 return -EINVAL;
         if (val == 1) {
-                cpumask_t old;
-                cpumask_of_cpu_ptr(newmask, cpu);
-
-                old = current->cpus_allowed;
+                cpumask_t old = current->cpus_allowed;
 
                 get_online_cpus();
-                set_cpus_allowed_ptr(current, newmask);
+                set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu));
 
                 mutex_lock(&microcode_mutex);
                 if (uci->valid)
diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c
index 06a9f643817e..724adfc63cb9 100644
--- a/arch/x86/kernel/reboot.c
+++ b/arch/x86/kernel/reboot.c
@@ -414,25 +414,20 @@ void native_machine_shutdown(void)
 
         /* The boot cpu is always logical cpu 0 */
         int reboot_cpu_id = 0;
-        cpumask_of_cpu_ptr(newmask, reboot_cpu_id);
 
 #ifdef CONFIG_X86_32
         /* See if there has been given a command line override */
         if ((reboot_cpu != -1) && (reboot_cpu < NR_CPUS) &&
-                cpu_online(reboot_cpu)) {
+                cpu_online(reboot_cpu))
                 reboot_cpu_id = reboot_cpu;
-                cpumask_of_cpu_ptr_next(newmask, reboot_cpu_id);
-        }
 #endif
 
         /* Make certain the cpu I'm about to reboot on is online */
-        if (!cpu_online(reboot_cpu_id)) {
+        if (!cpu_online(reboot_cpu_id))
                 reboot_cpu_id = smp_processor_id();
-                cpumask_of_cpu_ptr_next(newmask, reboot_cpu_id);
-        }
 
         /* Make certain I only run on the appropriate processor */
-        set_cpus_allowed_ptr(current, newmask);
+        set_cpus_allowed_ptr(current, &cpumask_of_cpu(reboot_cpu_id));
 
         /* O.K Now that I'm on the appropriate processor,
          * stop all of the others.
diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c
index f7745f94c006..76e305e064f9 100644
--- a/arch/x86/kernel/setup_percpu.c
+++ b/arch/x86/kernel/setup_percpu.c
@@ -80,24 +80,6 @@ static void __init setup_per_cpu_maps(void)
 #endif
 }
 
-#ifdef CONFIG_HAVE_CPUMASK_OF_CPU_MAP
-cpumask_t *cpumask_of_cpu_map __read_mostly;
-EXPORT_SYMBOL(cpumask_of_cpu_map);
-
-/* requires nr_cpu_ids to be initialized */
-static void __init setup_cpumask_of_cpu(void)
-{
-        int i;
-
-        /* alloc_bootmem zeroes memory */
-        cpumask_of_cpu_map = alloc_bootmem_low(sizeof(cpumask_t) * nr_cpu_ids);
-        for (i = 0; i < nr_cpu_ids; i++)
-                cpu_set(i, cpumask_of_cpu_map[i]);
-}
-#else
-static inline void setup_cpumask_of_cpu(void) { }
-#endif
-
 #ifdef CONFIG_X86_32
 /*
  * Great future not-so-futuristic plan: make i386 and x86_64 do it
@@ -197,9 +179,6 @@ void __init setup_per_cpu_areas(void)
 
         /* Setup node to cpumask map */
         setup_node_to_cpumask_map();
-
-        /* Setup cpumask_of_cpu map */
-        setup_cpumask_of_cpu();
 }
 
 #endif
diff --git a/drivers/acpi/processor_throttling.c b/drivers/acpi/processor_throttling.c
index a2c3f9cfa549..a56fc6c4394b 100644
--- a/drivers/acpi/processor_throttling.c
+++ b/drivers/acpi/processor_throttling.c
@@ -827,7 +827,6 @@ static int acpi_processor_get_throttling_ptc(struct acpi_processor *pr)
 static int acpi_processor_get_throttling(struct acpi_processor *pr)
 {
         cpumask_t saved_mask;
-        cpumask_of_cpu_ptr_declare(new_mask);
         int ret;
 
         if (!pr)
@@ -839,8 +838,7 @@ static int acpi_processor_get_throttling(struct acpi_processor *pr)
          * Migrate task to the cpu pointed by pr.
          */
         saved_mask = current->cpus_allowed;
-        cpumask_of_cpu_ptr_next(new_mask, pr->id);
-        set_cpus_allowed_ptr(current, new_mask);
+        set_cpus_allowed_ptr(current, &cpumask_of_cpu(pr->id));
         ret = pr->throttling.acpi_processor_get_throttling(pr);
         /* restore the previous state */
         set_cpus_allowed_ptr(current, &saved_mask);
@@ -989,7 +987,6 @@ static int acpi_processor_set_throttling_ptc(struct acpi_processor *pr,
 int acpi_processor_set_throttling(struct acpi_processor *pr, int state)
 {
         cpumask_t saved_mask;
-        cpumask_of_cpu_ptr_declare(new_mask);
         int ret = 0;
         unsigned int i;
         struct acpi_processor *match_pr;
@@ -1028,8 +1025,7 @@ int acpi_processor_set_throttling(struct acpi_processor *pr, int state)
          * it can be called only for the cpu pointed by pr.
          */
         if (p_throttling->shared_type == DOMAIN_COORD_TYPE_SW_ANY) {
-                cpumask_of_cpu_ptr_next(new_mask, pr->id);
-                set_cpus_allowed_ptr(current, new_mask);
+                set_cpus_allowed_ptr(current, &cpumask_of_cpu(pr->id));
                 ret = p_throttling->acpi_processor_set_throttling(pr,
                                                 t_state.target_state);
         } else {
@@ -1060,8 +1056,7 @@ int acpi_processor_set_throttling(struct acpi_processor *pr, int state)
                                 continue;
                         }
                         t_state.cpu = i;
-                        cpumask_of_cpu_ptr_next(new_mask, i);
-                        set_cpus_allowed_ptr(current, new_mask);
+                        set_cpus_allowed_ptr(current, &cpumask_of_cpu(i));
                         ret = match_pr->throttling.
                                 acpi_processor_set_throttling(
                                 match_pr, t_state.target_state);
diff --git a/drivers/firmware/dcdbas.c b/drivers/firmware/dcdbas.c
index c66817e7717b..50a071f1c945 100644
--- a/drivers/firmware/dcdbas.c
+++ b/drivers/firmware/dcdbas.c
@@ -245,7 +245,6 @@ static ssize_t host_control_on_shutdown_store(struct device *dev,
 static int smi_request(struct smi_cmd *smi_cmd)
 {
         cpumask_t old_mask;
-        cpumask_of_cpu_ptr(new_mask, 0);
         int ret = 0;
 
         if (smi_cmd->magic != SMI_CMD_MAGIC) {
@@ -256,7 +255,7 @@ static int smi_request(struct smi_cmd *smi_cmd)
 
         /* SMI requires CPU 0 */
         old_mask = current->cpus_allowed;
-        set_cpus_allowed_ptr(current, new_mask);
+        set_cpus_allowed_ptr(current, &cpumask_of_cpu(0));
         if (smp_processor_id() != 0) {
                 dev_dbg(&dcdbas_pdev->dev, "%s: failed to get CPU 0\n",
                         __func__);
diff --git a/drivers/misc/sgi-xp/xpc_main.c b/drivers/misc/sgi-xp/xpc_main.c
index 579b01ff82d4..c3b4227f48a5 100644
--- a/drivers/misc/sgi-xp/xpc_main.c
+++ b/drivers/misc/sgi-xp/xpc_main.c
@@ -229,11 +229,10 @@ xpc_hb_checker(void *ignore)
         int last_IRQ_count = 0;
         int new_IRQ_count;
         int force_IRQ = 0;
-        cpumask_of_cpu_ptr(cpumask, XPC_HB_CHECK_CPU);
 
         /* this thread was marked active by xpc_hb_init() */
 
-        set_cpus_allowed_ptr(current, cpumask);
+        set_cpus_allowed_ptr(current, &cpumask_of_cpu(XPC_HB_CHECK_CPU));
 
         /* set our heartbeating to other partitions into motion */
         xpc_hb_check_timeout = jiffies + (xpc_hb_check_interval * HZ);
diff --git a/include/linux/cpumask.h b/include/linux/cpumask.h
index 1b5c98e7fef7..96d0509fb8d8 100644
--- a/include/linux/cpumask.h
+++ b/include/linux/cpumask.h
@@ -62,15 +62,7 @@
  * int next_cpu_nr(cpu, mask)           Next cpu past 'cpu', or nr_cpu_ids
  *
  * cpumask_t cpumask_of_cpu(cpu)        Return cpumask with bit 'cpu' set
- *ifdef CONFIG_HAS_CPUMASK_OF_CPU
- * cpumask_of_cpu_ptr_declare(v)        Declares cpumask_t *v
- * cpumask_of_cpu_ptr_next(v, cpu)      Sets v = &cpumask_of_cpu_map[cpu]
- * cpumask_of_cpu_ptr(v, cpu)           Combines above two operations
- *else
- * cpumask_of_cpu_ptr_declare(v)        Declares cpumask_t _v and *v = &_v
- * cpumask_of_cpu_ptr_next(v, cpu)      Sets _v = cpumask_of_cpu(cpu)
- * cpumask_of_cpu_ptr(v, cpu)           Combines above two operations
- *endif
+ *                                      (can be used as an lvalue)
  * CPU_MASK_ALL                         Initializer - all bits set
  * CPU_MASK_NONE                        Initializer - no bits set
  * unsigned long *cpus_addr(mask)       Array of unsigned long's in mask
@@ -273,37 +265,30 @@ static inline void __cpus_shift_left(cpumask_t *dstp,
         bitmap_shift_left(dstp->bits, srcp->bits, n, nbits);
 }
 
+/*
+ * Special-case data structure for "single bit set only" constant CPU masks.
+ *
+ * We pre-generate all the 64 (or 32) possible bit positions, with enough
+ * padding to the left and the right, and return the constant pointer
+ * appropriately offset.
+ */
+extern const unsigned long
+        cpu_bit_bitmap[BITS_PER_LONG+1][BITS_TO_LONGS(NR_CPUS)];
+
+static inline const cpumask_t *get_cpu_mask(unsigned int cpu)
+{
+        const unsigned long *p = cpu_bit_bitmap[1 + cpu % BITS_PER_LONG];
+        p -= cpu / BITS_PER_LONG;
+        return (const cpumask_t *)p;
+}
+
+/*
+ * In cases where we take the address of the cpumask immediately,
+ * gcc optimizes it out (it's a constant) and there's no huge stack
+ * variable created:
+ */
+#define cpumask_of_cpu(cpu) ({ *get_cpu_mask(cpu); })
 
-#ifdef CONFIG_HAVE_CPUMASK_OF_CPU_MAP
-extern cpumask_t *cpumask_of_cpu_map;
-#define cpumask_of_cpu(cpu)     (cpumask_of_cpu_map[cpu])
-#define cpumask_of_cpu_ptr(v, cpu)                                      \
-                const cpumask_t *v = &cpumask_of_cpu(cpu)
-#define cpumask_of_cpu_ptr_declare(v)                                   \
-                const cpumask_t *v
-#define cpumask_of_cpu_ptr_next(v, cpu)                                 \
-                                        v = &cpumask_of_cpu(cpu)
-#else
-#define cpumask_of_cpu(cpu)                                             \
-({                                                                      \
-        typeof(_unused_cpumask_arg_) m;                                 \
-        if (sizeof(m) == sizeof(unsigned long)) {                       \
-                m.bits[0] = 1UL<<(cpu);                                 \
-        } else {                                                        \
-                cpus_clear(m);                                          \
-                cpu_set((cpu), m);                                      \
-        }                                                               \
-        m;                                                              \
-})
-#define cpumask_of_cpu_ptr(v, cpu)                                      \
-                cpumask_t _##v = cpumask_of_cpu(cpu);                   \
-                const cpumask_t *v = &_##v
-#define cpumask_of_cpu_ptr_declare(v)                                   \
-                cpumask_t _##v;                                         \
-                const cpumask_t *v = &_##v
-#define cpumask_of_cpu_ptr_next(v, cpu)                                 \
-                                        _##v = cpumask_of_cpu(cpu)
-#endif
 
 #define CPU_MASK_LAST_WORD BITMAP_LAST_WORD_MASK(NR_CPUS)
 
diff --git a/kernel/cpu.c b/kernel/cpu.c
index 29510d68338a..e202a68d1cc1 100644
--- a/kernel/cpu.c
+++ b/kernel/cpu.c
@@ -455,3 +455,28 @@ out:
 #endif /* CONFIG_PM_SLEEP_SMP */
 
 #endif /* CONFIG_SMP */
+
+/*
+ * cpu_bit_bitmap[] is a special, "compressed" data structure that
+ * represents all NR_CPUS bits binary values of 1<<nr.
+ *
+ * It is used by cpumask_of_cpu() to get a constant address to a CPU
+ * mask value that has a single bit set only.
+ */
+
+/* cpu_bit_bitmap[0] is empty - so we can back into it */
+#define MASK_DECLARE_1(x)       [x+1][0] = 1UL << (x)
+#define MASK_DECLARE_2(x)       MASK_DECLARE_1(x), MASK_DECLARE_1(x+1)
+#define MASK_DECLARE_4(x)       MASK_DECLARE_2(x), MASK_DECLARE_2(x+2)
+#define MASK_DECLARE_8(x)       MASK_DECLARE_4(x), MASK_DECLARE_4(x+4)
+
+const unsigned long cpu_bit_bitmap[BITS_PER_LONG+1][BITS_TO_LONGS(NR_CPUS)] = {
+
+        MASK_DECLARE_8(0),      MASK_DECLARE_8(8),
+        MASK_DECLARE_8(16),     MASK_DECLARE_8(24),
+#if BITS_PER_LONG > 32
+        MASK_DECLARE_8(32),     MASK_DECLARE_8(40),
+        MASK_DECLARE_8(48),     MASK_DECLARE_8(56),
+#endif
+};
+EXPORT_SYMBOL_GPL(cpu_bit_bitmap);
diff --git a/kernel/time/tick-common.c b/kernel/time/tick-common.c
index bf43284d6855..80c4336f4188 100644
--- a/kernel/time/tick-common.c
+++ b/kernel/time/tick-common.c
@@ -196,12 +196,10 @@ static int tick_check_new_device(struct clock_event_device *newdev)
         struct tick_device *td;
         int cpu, ret = NOTIFY_OK;
         unsigned long flags;
-        cpumask_of_cpu_ptr_declare(cpumask);
 
         spin_lock_irqsave(&tick_device_lock, flags);
 
         cpu = smp_processor_id();
-        cpumask_of_cpu_ptr_next(cpumask, cpu);
         if (!cpu_isset(cpu, newdev->cpumask))
                 goto out_bc;
 
@@ -209,7 +207,7 @@ static int tick_check_new_device(struct clock_event_device *newdev)
         curdev = td->evtdev;
 
         /* cpu local device ? */
-        if (!cpus_equal(newdev->cpumask, *cpumask)) {
+        if (!cpus_equal(newdev->cpumask, cpumask_of_cpu(cpu))) {
 
                 /*
                  * If the cpu affinity of the device interrupt can not
@@ -222,7 +220,7 @@ static int tick_check_new_device(struct clock_event_device *newdev)
                  * If we have a cpu local device already, do not replace it
                  * by a non cpu local device
                  */
-                if (curdev && cpus_equal(curdev->cpumask, *cpumask))
+                if (curdev && cpus_equal(curdev->cpumask, cpumask_of_cpu(cpu)))
                         goto out_bc;
         }
 
@@ -254,7 +252,7 @@ static int tick_check_new_device(struct clock_event_device *newdev)
                 curdev = NULL;
         }
         clockevents_exchange_device(curdev, newdev);
-        tick_setup_device(td, newdev, cpu, cpumask);
+        tick_setup_device(td, newdev, cpu, &cpumask_of_cpu(cpu));
         if (newdev->features & CLOCK_EVT_FEAT_ONESHOT)
                 tick_oneshot_notify();
 
diff --git a/kernel/trace/trace_sysprof.c b/kernel/trace/trace_sysprof.c
index ce2d723c10e1..bb948e52ce20 100644
--- a/kernel/trace/trace_sysprof.c
+++ b/kernel/trace/trace_sysprof.c
@@ -213,9 +213,7 @@ static void start_stack_timers(void)
         int cpu;
 
         for_each_online_cpu(cpu) {
-                cpumask_of_cpu_ptr(new_mask, cpu);
-
-                set_cpus_allowed_ptr(current, new_mask);
+                set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu));
                 start_stack_timer(cpu);
         }
         set_cpus_allowed_ptr(current, &saved_mask);
diff --git a/lib/smp_processor_id.c b/lib/smp_processor_id.c
index c4381d9516f6..0f8fc22ed103 100644
--- a/lib/smp_processor_id.c
+++ b/lib/smp_processor_id.c
@@ -11,7 +11,6 @@ notrace unsigned int debug_smp_processor_id(void)
 {
         unsigned long preempt_count = preempt_count();
         int this_cpu = raw_smp_processor_id();
-        cpumask_of_cpu_ptr_declare(this_mask);
 
         if (likely(preempt_count))
                 goto out;
@@ -23,9 +22,7 @@ notrace unsigned int debug_smp_processor_id(void)
          * Kernel threads bound to a single CPU can safely use
          * smp_processor_id():
          */
-        cpumask_of_cpu_ptr_next(this_mask, this_cpu);
-
-        if (cpus_equal(current->cpus_allowed, *this_mask))
+        if (cpus_equal(current->cpus_allowed, cpumask_of_cpu(this_cpu)))
                 goto out;
 
         /*
diff --git a/net/sunrpc/svc.c b/net/sunrpc/svc.c
index 835d27413083..5a32cb7c4bb4 100644
--- a/net/sunrpc/svc.c
+++ b/net/sunrpc/svc.c
@@ -310,8 +310,7 @@ svc_pool_map_set_cpumask(struct task_struct *task, unsigned int pidx)
         switch (m->mode) {
         case SVC_POOL_PERCPU:
         {
-                cpumask_of_cpu_ptr(cpumask, node);
-                set_cpus_allowed_ptr(task, cpumask);
+                set_cpus_allowed_ptr(task, &cpumask_of_cpu(node));
                 break;
         }
         case SVC_POOL_PERNODE: