10 files changed, 123 insertions, 112 deletions

diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile
index 3db651fc8ec5..d679cb2c79b4 100644
--- a/arch/x86/kernel/Makefile
+++ b/arch/x86/kernel/Makefile
@@ -10,7 +10,7 @@ ifdef CONFIG_FTRACE
 # Do not profile debug and lowlevel utilities
 CFLAGS_REMOVE_tsc.o = -pg
 CFLAGS_REMOVE_rtc.o = -pg
-CFLAGS_REMOVE_paravirt.o = -pg
+CFLAGS_REMOVE_paravirt-spinlocks.o = -pg
 endif
 
 #
@@ -89,7 +89,7 @@ obj-$(CONFIG_DEBUG_NX_TEST)	+= test_nx.o
 obj-$(CONFIG_VMI)               += vmi_32.o vmiclock_32.o
 obj-$(CONFIG_KVM_GUEST)         += kvm.o
 obj-$(CONFIG_KVM_CLOCK)         += kvmclock.o
-obj-$(CONFIG_PARAVIRT)          += paravirt.o paravirt_patch_$(BITS).o
+obj-$(CONFIG_PARAVIRT)          += paravirt.o paravirt_patch_$(BITS).o paravirt-spinlocks.o
 obj-$(CONFIG_PARAVIRT_CLOCK)    += pvclock.o
 
 obj-$(CONFIG_PCSPKR_PLATFORM)   += pcspeaker.o
diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
index 8aab8517642e..1c7d39f0e89e 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -744,14 +744,3 @@ void __cpuinit cpu_init(void)
         mxcsr_feature_mask_init();
 }
 
-#ifdef CONFIG_HOTPLUG_CPU
-void __cpuinit cpu_uninit(void)
-{
-        int cpu = raw_smp_processor_id();
-        cpu_clear(cpu, cpu_initialized);
-
-        /* lazy TLB state */
-        per_cpu(cpu_tlbstate, cpu).state = 0;
-        per_cpu(cpu_tlbstate, cpu).active_mm = &init_mm;
-}
-#endif
diff --git a/arch/x86/kernel/ldt.c b/arch/x86/kernel/ldt.c
index b68e21f06f4f..6e388412a854 100644
--- a/arch/x86/kernel/ldt.c
+++ b/arch/x86/kernel/ldt.c
@@ -51,6 +51,8 @@ static int alloc_ldt(mm_context_t *pc, int mincount, int reload)
         memset(newldt + oldsize * LDT_ENTRY_SIZE, 0,
                (mincount - oldsize) * LDT_ENTRY_SIZE);
 
+        paravirt_alloc_ldt(newldt, mincount);
+
 #ifdef CONFIG_X86_64
         /* CHECKME: Do we really need this ? */
         wmb();
@@ -73,6 +75,7 @@ static int alloc_ldt(mm_context_t *pc, int mincount, int reload)
 #endif
         }
         if (oldsize) {
+                paravirt_free_ldt(oldldt, oldsize);
                 if (oldsize * LDT_ENTRY_SIZE > PAGE_SIZE)
                         vfree(oldldt);
                 else
@@ -84,10 +87,13 @@ static int alloc_ldt(mm_context_t *pc, int mincount, int reload)
 static inline int copy_ldt(mm_context_t *new, mm_context_t *old)
 {
         int err = alloc_ldt(new, old->size, 0);
+        int i;
 
         if (err < 0)
                 return err;
-        memcpy(new->ldt, old->ldt, old->size * LDT_ENTRY_SIZE);
+
+        for(i = 0; i < old->size; i++)
+                write_ldt_entry(new->ldt, i, old->ldt + i * LDT_ENTRY_SIZE);
         return 0;
 }
 
@@ -124,6 +130,7 @@ void destroy_context(struct mm_struct *mm)
                 if (mm == current->active_mm)
                         clear_LDT();
 #endif
+                paravirt_free_ldt(mm->context.ldt, mm->context.size);
                 if (mm->context.size * LDT_ENTRY_SIZE > PAGE_SIZE)
                         vfree(mm->context.ldt);
                 else
diff --git a/arch/x86/kernel/paravirt-spinlocks.c b/arch/x86/kernel/paravirt-spinlocks.c
new file mode 100644
index 000000000000..0e9f1982b1dd
--- /dev/null
+++ b/arch/x86/kernel/paravirt-spinlocks.c
@@ -0,0 +1,37 @@
+/*
+ * Split spinlock implementation out into its own file, so it can be
+ * compiled in a FTRACE-compatible way.
+ */
+#include <linux/spinlock.h>
+#include <linux/module.h>
+
+#include <asm/paravirt.h>
+
+static void default_spin_lock_flags(struct raw_spinlock *lock, unsigned long flags)
+{
+        __raw_spin_lock(lock);
+}
+
+struct pv_lock_ops pv_lock_ops = {
+#ifdef CONFIG_SMP
+        .spin_is_locked = __ticket_spin_is_locked,
+        .spin_is_contended = __ticket_spin_is_contended,
+
+        .spin_lock = __ticket_spin_lock,
+        .spin_lock_flags = default_spin_lock_flags,
+        .spin_trylock = __ticket_spin_trylock,
+        .spin_unlock = __ticket_spin_unlock,
+#endif
+};
+EXPORT_SYMBOL(pv_lock_ops);
+
+void __init paravirt_use_bytelocks(void)
+{
+#ifdef CONFIG_SMP
+        pv_lock_ops.spin_is_locked = __byte_spin_is_locked;
+        pv_lock_ops.spin_is_contended = __byte_spin_is_contended;
+        pv_lock_ops.spin_lock = __byte_spin_lock;
+        pv_lock_ops.spin_trylock = __byte_spin_trylock;
+        pv_lock_ops.spin_unlock = __byte_spin_unlock;
+#endif
+}
diff --git a/arch/x86/kernel/paravirt.c b/arch/x86/kernel/paravirt.c
index 300da17e61cb..7faea1817d05 100644
--- a/arch/x86/kernel/paravirt.c
+++ b/arch/x86/kernel/paravirt.c
@@ -268,17 +268,6 @@ enum paravirt_lazy_mode paravirt_get_lazy_mode(void)
         return __get_cpu_var(paravirt_lazy_mode);
 }
 
-void __init paravirt_use_bytelocks(void)
-{
-#ifdef CONFIG_SMP
-        pv_lock_ops.spin_is_locked = __byte_spin_is_locked;
-        pv_lock_ops.spin_is_contended = __byte_spin_is_contended;
-        pv_lock_ops.spin_lock = __byte_spin_lock;
-        pv_lock_ops.spin_trylock = __byte_spin_trylock;
-        pv_lock_ops.spin_unlock = __byte_spin_unlock;
-#endif
-}
-
 struct pv_info pv_info = {
         .name = "bare hardware",
         .paravirt_enabled = 0,
@@ -348,6 +337,10 @@ struct pv_cpu_ops pv_cpu_ops = {
         .write_ldt_entry = native_write_ldt_entry,
         .write_gdt_entry = native_write_gdt_entry,
         .write_idt_entry = native_write_idt_entry,
+
+        .alloc_ldt = paravirt_nop,
+        .free_ldt = paravirt_nop,
+
         .load_sp0 = native_load_sp0,
 
 #if defined(CONFIG_X86_32) || defined(CONFIG_IA32_EMULATION)
@@ -461,18 +454,6 @@ struct pv_mmu_ops pv_mmu_ops = {
         .set_fixmap = native_set_fixmap,
 };
 
-struct pv_lock_ops pv_lock_ops = {
-#ifdef CONFIG_SMP
-        .spin_is_locked = __ticket_spin_is_locked,
-        .spin_is_contended = __ticket_spin_is_contended,
-
-        .spin_lock = __ticket_spin_lock,
-        .spin_trylock = __ticket_spin_trylock,
-        .spin_unlock = __ticket_spin_unlock,
-#endif
-};
-EXPORT_SYMBOL(pv_lock_ops);
-
 EXPORT_SYMBOL_GPL(pv_time_ops);
 EXPORT_SYMBOL    (pv_cpu_ops);
 EXPORT_SYMBOL    (pv_mmu_ops);
diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c
index 3b7a1ddcc0bc..b76b38ff962b 100644
--- a/arch/x86/kernel/process_32.c
+++ b/arch/x86/kernel/process_32.c
@@ -72,46 +72,12 @@ unsigned long thread_saved_pc(struct task_struct *tsk)
         return ((unsigned long *)tsk->thread.sp)[3];
 }
 
-#ifdef CONFIG_HOTPLUG_CPU
-#include <asm/nmi.h>
-
-static void cpu_exit_clear(void)
-{
-        int cpu = raw_smp_processor_id();
-
-        idle_task_exit();
-
-        cpu_uninit();
-        irq_ctx_exit(cpu);
-
-        cpu_clear(cpu, cpu_callout_map);
-        cpu_clear(cpu, cpu_callin_map);
-
-        numa_remove_cpu(cpu);
-}
-
-/* We don't actually take CPU down, just spin without interrupts. */
-static inline void play_dead(void)
-{
-        /* This must be done before dead CPU ack */
-        cpu_exit_clear();
-        mb();
-        /* Ack it */
-        __get_cpu_var(cpu_state) = CPU_DEAD;
-
-        /*
-         * With physical CPU hotplug, we should halt the cpu
-         */
-        local_irq_disable();
-        /* mask all interrupts, flush any and all caches, and halt */
-        wbinvd_halt();
-}
-#else
+#ifndef CONFIG_SMP
 static inline void play_dead(void)
 {
         BUG();
 }
-#endif /* CONFIG_HOTPLUG_CPU */
+#endif
 
 /*
  * The idle thread. There's no useful work to be
diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c
index 71553b664e2a..ec27afa43d7e 100644
--- a/arch/x86/kernel/process_64.c
+++ b/arch/x86/kernel/process_64.c
@@ -85,28 +85,12 @@ void exit_idle(void)
         __exit_idle();
 }
 
-#ifdef CONFIG_HOTPLUG_CPU
-DECLARE_PER_CPU(int, cpu_state);
-
-#include <asm/nmi.h>
-/* We halt the CPU with physical CPU hotplug */
-static inline void play_dead(void)
-{
-        idle_task_exit();
-        mb();
-        /* Ack it */
-        __get_cpu_var(cpu_state) = CPU_DEAD;
-
-        local_irq_disable();
-        /* mask all interrupts, flush any and all caches, and halt */
-        wbinvd_halt();
-}
-#else
+#ifndef CONFIG_SMP
 static inline void play_dead(void)
 {
         BUG();
 }
-#endif /* CONFIG_HOTPLUG_CPU */
+#endif
 
 /*
  * The idle thread. There's no useful work to be
diff --git a/arch/x86/kernel/smp.c b/arch/x86/kernel/smp.c
index 361b7a4c640c..18f9b19f5f8f 100644
--- a/arch/x86/kernel/smp.c
+++ b/arch/x86/kernel/smp.c
@@ -214,12 +214,16 @@ void smp_call_function_single_interrupt(struct pt_regs *regs)
 struct smp_ops smp_ops = {
         .smp_prepare_boot_cpu = native_smp_prepare_boot_cpu,
         .smp_prepare_cpus = native_smp_prepare_cpus,
-        .cpu_up = native_cpu_up,
         .smp_cpus_done = native_smp_cpus_done,
 
         .smp_send_stop = native_smp_send_stop,
         .smp_send_reschedule = native_smp_send_reschedule,
 
+        .cpu_up = native_cpu_up,
+        .cpu_die = native_cpu_die,
+        .cpu_disable = native_cpu_disable,
+        .play_dead = native_play_dead,
+
         .send_call_func_ipi = native_send_call_func_ipi,
         .send_call_func_single_ipi = native_send_call_func_single_ipi,
 };
diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
index 7985c5b3f916..66b04e598817 100644
--- a/arch/x86/kernel/smpboot.c
+++ b/arch/x86/kernel/smpboot.c
@@ -1346,25 +1346,9 @@ static void __ref remove_cpu_from_maps(int cpu)
         numa_remove_cpu(cpu);
 }
 
-int __cpu_disable(void)
+void cpu_disable_common(void)
 {
         int cpu = smp_processor_id();
-
-        /*
-         * Perhaps use cpufreq to drop frequency, but that could go
-         * into generic code.
-         *
-         * We won't take down the boot processor on i386 due to some
-         * interrupts only being able to be serviced by the BSP.
-         * Especially so if we're not using an IOAPIC   -zwane
-         */
-        if (cpu == 0)
-                return -EBUSY;
-
-        if (nmi_watchdog == NMI_LOCAL_APIC)
-                stop_apic_nmi_watchdog(NULL);
-        clear_local_APIC();
-
         /*
          * HACK:
          * Allow any queued timer interrupts to get serviced
@@ -1382,10 +1366,32 @@ int __cpu_disable(void)
         remove_cpu_from_maps(cpu);
         unlock_vector_lock();
         fixup_irqs(cpu_online_map);
+}
+
+int native_cpu_disable(void)
+{
+        int cpu = smp_processor_id();
+
+        /*
+         * Perhaps use cpufreq to drop frequency, but that could go
+         * into generic code.
+         *
+         * We won't take down the boot processor on i386 due to some
+         * interrupts only being able to be serviced by the BSP.
+         * Especially so if we're not using an IOAPIC   -zwane
+         */
+        if (cpu == 0)
+                return -EBUSY;
+
+        if (nmi_watchdog == NMI_LOCAL_APIC)
+                stop_apic_nmi_watchdog(NULL);
+        clear_local_APIC();
+
+        cpu_disable_common();
         return 0;
 }
 
-void __cpu_die(unsigned int cpu)
+void native_cpu_die(unsigned int cpu)
 {
         /* We don't do anything here: idle task is faking death itself. */
         unsigned int i;
@@ -1402,15 +1408,44 @@ void __cpu_die(unsigned int cpu)
         }
         printk(KERN_ERR "CPU %u didn't die...\n", cpu);
 }
+
+void play_dead_common(void)
+{
+        idle_task_exit();
+        reset_lazy_tlbstate();
+        irq_ctx_exit(raw_smp_processor_id());
+
+        mb();
+        /* Ack it */
+        __get_cpu_var(cpu_state) = CPU_DEAD;
+
+        /*
+         * With physical CPU hotplug, we should halt the cpu
+         */
+        local_irq_disable();
+}
+
+void native_play_dead(void)
+{
+        play_dead_common();
+        wbinvd_halt();
+}
+
 #else /* ... !CONFIG_HOTPLUG_CPU */
-int __cpu_disable(void)
+int native_cpu_disable(void)
 {
         return -ENOSYS;
 }
 
-void __cpu_die(unsigned int cpu)
+void native_cpu_die(unsigned int cpu)
 {
         /* We said "no" in __cpu_disable */
         BUG();
 }
+
+void native_play_dead(void)
+{
+        BUG();
+}
+
 #endif
diff --git a/arch/x86/kernel/tlb_32.c b/arch/x86/kernel/tlb_32.c
index fec1ecedc9b7..e00534b33534 100644
--- a/arch/x86/kernel/tlb_32.c
+++ b/arch/x86/kernel/tlb_32.c
@@ -241,3 +241,11 @@ void flush_tlb_all(void)
         on_each_cpu(do_flush_tlb_all, NULL, 1);
 }
 
+void reset_lazy_tlbstate(void)
+{
+        int cpu = raw_smp_processor_id();
+
+        per_cpu(cpu_tlbstate, cpu).state = 0;
+        per_cpu(cpu_tlbstate, cpu).active_mm = &init_mm;
+}
+