diff --git a/include/asm-alpha/smp.h b/include/asm-alpha/smp.h
index 286e1d84..544c69a 100644
--- a/include/asm-alpha/smp.h
+++ b/include/asm-alpha/smp.h
@@ -47,12 +47,13 @@
 extern int smp_num_cpus;
 #define cpu_possible_map	cpu_present_map
 
-int smp_call_function_on_cpu(void (*func) (void *info), void *info,int retry, int wait, cpumask_t cpu);
+extern void arch_send_call_function_single_ipi(int cpu);
+extern void arch_send_call_function_ipi(cpumask_t mask);
 
 #else /* CONFIG_SMP */
 
 #define hard_smp_processor_id()		0
-#define smp_call_function_on_cpu(func,info,retry,wait,cpu)    ({ 0; })
+#define smp_call_function_on_cpu(func,info,wait,cpu)    ({ 0; })
 
 #endif /* CONFIG_SMP */
 
diff --git a/include/asm-arm/smp.h b/include/asm-arm/smp.h
index af99636..7fffa24 100644
--- a/include/asm-arm/smp.h
+++ b/include/asm-arm/smp.h
@@ -101,6 +101,9 @@
 extern int platform_cpu_kill(unsigned int cpu);
 extern void platform_cpu_enable(unsigned int cpu);
 
+extern void arch_send_call_function_single_ipi(int cpu);
+extern void arch_send_call_function_ipi(cpumask_t mask);
+
 /*
  * Local timer interrupt handling function (can be IPI'ed).
  */
diff --git a/include/asm-ia64/smp.h b/include/asm-ia64/smp.h
index ec5f355..27731e0 100644
--- a/include/asm-ia64/smp.h
+++ b/include/asm-ia64/smp.h
@@ -38,9 +38,6 @@
 	return lid.f.id << 8 | lid.f.eid;
 }
 
-extern int smp_call_function_mask(cpumask_t mask, void (*func)(void *),
-				  void *info, int wait);
-
 #define hard_smp_processor_id()		ia64_get_lid()
 
 #ifdef CONFIG_SMP
@@ -124,11 +121,12 @@
 extern void smp_do_timer (struct pt_regs *regs);
 
 extern void smp_send_reschedule (int cpu);
-extern void lock_ipi_calllock(void);
-extern void unlock_ipi_calllock(void);
 extern void identify_siblings (struct cpuinfo_ia64 *);
 extern int is_multithreading_enabled(void);
 
+extern void arch_send_call_function_single_ipi(int cpu);
+extern void arch_send_call_function_ipi(cpumask_t mask);
+
 #else /* CONFIG_SMP */
 
 #define cpu_logical_id(i)		0
diff --git a/include/asm-m32r/smp.h b/include/asm-m32r/smp.h
index 078e1a5..c5dd669 100644
--- a/include/asm-m32r/smp.h
+++ b/include/asm-m32r/smp.h
@@ -89,6 +89,9 @@
 extern void smp_send_timer(void);
 extern unsigned long send_IPI_mask_phys(cpumask_t, int, int);
 
+extern void arch_send_call_function_single_ipi(int cpu);
+extern void arch_send_call_function_ipi(cpumask_t mask);
+
 #endif	/* not __ASSEMBLY__ */
 
 #define NO_PROC_ID (0xff)	/* No processor magic marker */
@@ -104,6 +107,7 @@
 #define LOCAL_TIMER_IPI		(M32R_IRQ_IPI3-M32R_IRQ_IPI0)
 #define INVALIDATE_CACHE_IPI	(M32R_IRQ_IPI4-M32R_IRQ_IPI0)
 #define CPU_BOOT_IPI		(M32R_IRQ_IPI5-M32R_IRQ_IPI0)
+#define CALL_FUNC_SINGLE_IPI	(M32R_IRQ_IPI6-M32R_IRQ_IPI0)
 
 #define IPI_SHIFT	(0)
 #define NR_IPIS		(8)
diff --git a/include/asm-mips/smp.h b/include/asm-mips/smp.h
index 84fef1a..0ff5b52 100644
--- a/include/asm-mips/smp.h
+++ b/include/asm-mips/smp.h
@@ -35,16 +35,6 @@
 
 #define NO_PROC_ID	(-1)
 
-struct call_data_struct {
-	void		(*func)(void *);
-	void		*info;
-	atomic_t	started;
-	atomic_t	finished;
-	int		wait;
-};
-
-extern struct call_data_struct *call_data;
-
 #define SMP_RESCHEDULE_YOURSELF	0x1	/* XXX braindead */
 #define SMP_CALL_FUNCTION	0x2
 
@@ -67,4 +57,7 @@
 
 extern asmlinkage void smp_call_function_interrupt(void);
 
+extern void arch_send_call_function_single_ipi(int cpu);
+extern void arch_send_call_function_ipi(cpumask_t mask);
+
 #endif /* __ASM_SMP_H */
diff --git a/include/asm-parisc/smp.h b/include/asm-parisc/smp.h
index 306f495..398cdba 100644
--- a/include/asm-parisc/smp.h
+++ b/include/asm-parisc/smp.h
@@ -30,6 +30,9 @@
 extern void smp_send_reschedule(int cpu);
 extern void smp_send_all_nop(void);
 
+extern void arch_send_call_function_single_ipi(int cpu);
+extern void arch_send_call_function_ipi(cpumask_t mask);
+
 #endif /* !ASSEMBLY */
 
 /*
diff --git a/include/asm-powerpc/smp.h b/include/asm-powerpc/smp.h
index 505f35b..c663a1f 100644
--- a/include/asm-powerpc/smp.h
+++ b/include/asm-powerpc/smp.h
@@ -67,10 +67,7 @@
  * in /proc/interrupts will be wrong!!! --Troy */
 #define PPC_MSG_CALL_FUNCTION   0
 #define PPC_MSG_RESCHEDULE      1
-/* This is unused now */
-#if 0
-#define PPC_MSG_MIGRATE_TASK    2
-#endif
+#define PPC_MSG_CALL_FUNC_SINGLE	2
 #define PPC_MSG_DEBUGGER_BREAK  3
 
 void smp_init_iSeries(void);
@@ -117,6 +114,9 @@
 
 extern struct smp_ops_t *smp_ops;
 
+extern void arch_send_call_function_single_ipi(int cpu);
+extern void arch_send_call_function_ipi(cpumask_t mask);
+
 #endif /* __ASSEMBLY__ */
 
 #endif /* __KERNEL__ */
diff --git a/include/asm-sh/smp.h b/include/asm-sh/smp.h
index 9c8d34b0..593343c 100644
--- a/include/asm-sh/smp.h
+++ b/include/asm-sh/smp.h
@@ -26,18 +26,10 @@
 
 #define NO_PROC_ID	(-1)
 
-struct smp_fn_call_struct {
-	spinlock_t lock;
-	atomic_t   finished;
-	void (*fn)(void *);
-	void *data;
-};
-
-extern struct smp_fn_call_struct smp_fn_call;
-
 #define SMP_MSG_FUNCTION	0
 #define SMP_MSG_RESCHEDULE	1
-#define SMP_MSG_NR		2
+#define SMP_MSG_FUNCTION_SINGLE	2
+#define SMP_MSG_NR		3
 
 void plat_smp_setup(void);
 void plat_prepare_cpus(unsigned int max_cpus);
@@ -46,6 +38,8 @@
 void plat_send_ipi(unsigned int cpu, unsigned int message);
 int plat_register_ipi_handler(unsigned int message,
 			      void (*handler)(void *), void *arg);
+extern void arch_send_call_function_single_ipi(int cpu);
+extern void arch_send_call_function_ipi(cpumask_t mask);
 
 #else
 
diff --git a/include/asm-sparc/smp.h b/include/asm-sparc/smp.h
index e6d5615..b61e74b 100644
--- a/include/asm-sparc/smp.h
+++ b/include/asm-sparc/smp.h
@@ -72,7 +72,7 @@
 			   unsigned long arg3, unsigned long arg4, unsigned long arg5)
 { smp_cross_call(func, arg1, arg2, arg3, arg4, arg5); }
 
-static inline int smp_call_function(void (*func)(void *info), void *info, int nonatomic, int wait)
+static inline int smp_call_function(void (*func)(void *info), void *info, int wait)
 {
 	xc1((smpfunc_t)func, (unsigned long)info);
 	return 0;
diff --git a/include/asm-x86/hw_irq.h b/include/asm-x86/hw_irq.h
index 18f067c..77ba51d 100644
--- a/include/asm-x86/hw_irq.h
+++ b/include/asm-x86/hw_irq.h
@@ -48,6 +48,7 @@
 extern void threshold_interrupt(void);
 
 extern void call_function_interrupt(void);
+extern void call_function_single_interrupt(void);
 
 /* PIC specific functions */
 extern void disable_8259A_irq(unsigned int irq);
diff --git a/include/asm-x86/irq_vectors.h b/include/asm-x86/irq_vectors.h
index 0ac864e..90b1d1f 100644
--- a/include/asm-x86/irq_vectors.h
+++ b/include/asm-x86/irq_vectors.h
@@ -64,6 +64,7 @@
 # define INVALIDATE_TLB_VECTOR		0xfd
 # define RESCHEDULE_VECTOR		0xfc
 # define CALL_FUNCTION_VECTOR		0xfb
+# define CALL_FUNCTION_SINGLE_VECTOR	0xfa
 # define THERMAL_APIC_VECTOR		0xf0
 
 #else
@@ -72,6 +73,7 @@
 #define ERROR_APIC_VECTOR		0xfe
 #define RESCHEDULE_VECTOR		0xfd
 #define CALL_FUNCTION_VECTOR		0xfc
+#define CALL_FUNCTION_SINGLE_VECTOR	0xfb
 #define THERMAL_APIC_VECTOR		0xfa
 #define THRESHOLD_APIC_VECTOR		0xf9
 #define INVALIDATE_TLB_VECTOR_END	0xf7
@@ -143,6 +145,7 @@
 #define VIC_RESCHEDULE_CPI		4
 #define VIC_ENABLE_IRQ_CPI		5
 #define VIC_CALL_FUNCTION_CPI		6
+#define VIC_CALL_FUNCTION_SINGLE_CPI	7
 
 /* Now the QIC CPIs:  Since we don't need the two initial levels,
  * these are 2 less than the VIC CPIs */
@@ -152,9 +155,10 @@
 #define QIC_RESCHEDULE_CPI		(VIC_RESCHEDULE_CPI - QIC_CPI_OFFSET)
 #define QIC_ENABLE_IRQ_CPI		(VIC_ENABLE_IRQ_CPI - QIC_CPI_OFFSET)
 #define QIC_CALL_FUNCTION_CPI		(VIC_CALL_FUNCTION_CPI - QIC_CPI_OFFSET)
+#define QIC_CALL_FUNCTION_SINGLE_CPI	(VIC_CALL_FUNCTION_SINGLE_CPI - QIC_CPI_OFFSET)
 
 #define VIC_START_FAKE_CPI		VIC_TIMER_CPI
-#define VIC_END_FAKE_CPI		VIC_CALL_FUNCTION_CPI
+#define VIC_END_FAKE_CPI		VIC_CALL_FUNCTION_SINGLE_CPI
 
 /* this is the SYS_INT CPI. */
 #define VIC_SYS_INT			8
diff --git a/include/asm-x86/mach-default/entry_arch.h b/include/asm-x86/mach-default/entry_arch.h
index bc86146..9283b60 100644
--- a/include/asm-x86/mach-default/entry_arch.h
+++ b/include/asm-x86/mach-default/entry_arch.h
@@ -13,6 +13,7 @@
 BUILD_INTERRUPT(reschedule_interrupt,RESCHEDULE_VECTOR)
 BUILD_INTERRUPT(invalidate_interrupt,INVALIDATE_TLB_VECTOR)
 BUILD_INTERRUPT(call_function_interrupt,CALL_FUNCTION_VECTOR)
+BUILD_INTERRUPT(call_function_single_interrupt,CALL_FUNCTION_SINGLE_VECTOR)
 #endif
 
 /*
diff --git a/include/asm-x86/mach-visws/entry_arch.h b/include/asm-x86/mach-visws/entry_arch.h
index b183fa6..86be554 100644
--- a/include/asm-x86/mach-visws/entry_arch.h
+++ b/include/asm-x86/mach-visws/entry_arch.h
@@ -1,23 +1,5 @@
 /*
- * The following vectors are part of the Linux architecture, there
- * is no hardware IRQ pin equivalent for them, they are triggered
- * through the ICC by us (IPIs)
+ * VISWS uses the standard Linux entry points:
  */
-#ifdef CONFIG_X86_SMP
-BUILD_INTERRUPT(reschedule_interrupt,RESCHEDULE_VECTOR)
-BUILD_INTERRUPT(invalidate_interrupt,INVALIDATE_TLB_VECTOR)
-BUILD_INTERRUPT(call_function_interrupt,CALL_FUNCTION_VECTOR)
-#endif
 
-/*
- * every pentium local APIC has two 'local interrupts', with a
- * soft-definable vector attached to both interrupts, one of
- * which is a timer interrupt, the other one is error counter
- * overflow. Linux uses the local APIC timer interrupt to get
- * a much simpler SMP time architecture:
- */
-#ifdef CONFIG_X86_LOCAL_APIC
-BUILD_INTERRUPT(apic_timer_interrupt,LOCAL_TIMER_VECTOR)
-BUILD_INTERRUPT(error_interrupt,ERROR_APIC_VECTOR)
-BUILD_INTERRUPT(spurious_interrupt,SPURIOUS_APIC_VECTOR)
-#endif
+#include "../mach-default/entry_arch.h"
diff --git a/include/asm-x86/mach-voyager/entry_arch.h b/include/asm-x86/mach-voyager/entry_arch.h
index 4a1e1e8..ae52624 100644
--- a/include/asm-x86/mach-voyager/entry_arch.h
+++ b/include/asm-x86/mach-voyager/entry_arch.h
@@ -23,4 +23,4 @@
 BUILD_INTERRUPT(qic_reschedule_interrupt, QIC_RESCHEDULE_CPI);
 BUILD_INTERRUPT(qic_enable_irq_interrupt, QIC_ENABLE_IRQ_CPI);
 BUILD_INTERRUPT(qic_call_function_interrupt, QIC_CALL_FUNCTION_CPI);
-
+BUILD_INTERRUPT(qic_call_function_single_interrupt, QIC_CALL_FUNCTION_SINGLE_CPI);
diff --git a/include/asm-x86/smp.h b/include/asm-x86/smp.h
index 2e221f1..c2784b3 100644
--- a/include/asm-x86/smp.h
+++ b/include/asm-x86/smp.h
@@ -50,9 +50,9 @@
 
 	void (*smp_send_stop)(void);
 	void (*smp_send_reschedule)(int cpu);
-	int (*smp_call_function_mask)(cpumask_t mask,
-				      void (*func)(void *info), void *info,
-				      int wait);
+
+	void (*send_call_func_ipi)(cpumask_t mask);
+	void (*send_call_func_single_ipi)(int cpu);
 };
 
 /* Globals due to paravirt */
@@ -94,17 +94,22 @@
 	smp_ops.smp_send_reschedule(cpu);
 }
 
-static inline int smp_call_function_mask(cpumask_t mask,
-					 void (*func) (void *info), void *info,
-					 int wait)
+static inline void arch_send_call_function_single_ipi(int cpu)
 {
-	return smp_ops.smp_call_function_mask(mask, func, info, wait);
+	smp_ops.send_call_func_single_ipi(cpu);
+}
+
+static inline void arch_send_call_function_ipi(cpumask_t mask)
+{
+	smp_ops.send_call_func_ipi(mask);
 }
 
 void native_smp_prepare_boot_cpu(void);
 void native_smp_prepare_cpus(unsigned int max_cpus);
 void native_smp_cpus_done(unsigned int max_cpus);
 int native_cpu_up(unsigned int cpunum);
+void native_send_call_func_ipi(cpumask_t mask);
+void native_send_call_func_single_ipi(int cpu);
 
 extern int __cpu_disable(void);
 extern void __cpu_die(unsigned int cpu);
@@ -197,7 +202,5 @@
 extern void cpu_uninit(void);
 #endif
 
-extern void lock_ipi_call_lock(void);
-extern void unlock_ipi_call_lock(void);
 #endif /* __ASSEMBLY__ */
 #endif
diff --git a/include/asm-x86/xen/events.h b/include/asm-x86/xen/events.h
index 596312a..f8d57ea 100644
--- a/include/asm-x86/xen/events.h
+++ b/include/asm-x86/xen/events.h
@@ -4,6 +4,7 @@
 enum ipi_vector {
 	XEN_RESCHEDULE_VECTOR,
 	XEN_CALL_FUNCTION_VECTOR,
+	XEN_CALL_FUNCTION_SINGLE_VECTOR,
 
 	XEN_NR_IPIS,
 };
diff --git a/include/linux/smp.h b/include/linux/smp.h
index 55232cc..48262f8 100644
--- a/include/linux/smp.h
+++ b/include/linux/smp.h
@@ -7,9 +7,18 @@
  */
 
 #include <linux/errno.h>
+#include <linux/list.h>
+#include <linux/cpumask.h>
 
 extern void cpu_idle(void);
 
+struct call_single_data {
+	struct list_head list;
+	void (*func) (void *info);
+	void *info;
+	unsigned int flags;
+};
+
 #ifdef CONFIG_SMP
 
 #include <linux/preempt.h>
@@ -52,15 +61,34 @@
 /*
  * Call a function on all other processors
  */
-int smp_call_function(void(*func)(void *info), void *info, int retry, int wait);
-
+int smp_call_function(void(*func)(void *info), void *info, int wait);
+int smp_call_function_mask(cpumask_t mask, void(*func)(void *info), void *info,
+				int wait);
 int smp_call_function_single(int cpuid, void (*func) (void *info), void *info,
-				int retry, int wait);
+				int wait);
+void __smp_call_function_single(int cpuid, struct call_single_data *data);
+
+/*
+ * Generic and arch helpers
+ */
+#ifdef CONFIG_USE_GENERIC_SMP_HELPERS
+void generic_smp_call_function_single_interrupt(void);
+void generic_smp_call_function_interrupt(void);
+void init_call_single_data(void);
+void ipi_call_lock(void);
+void ipi_call_unlock(void);
+void ipi_call_lock_irq(void);
+void ipi_call_unlock_irq(void);
+#else
+static inline void init_call_single_data(void)
+{
+}
+#endif
 
 /*
  * Call a function on all processors
  */
-int on_each_cpu(void (*func) (void *info), void *info, int retry, int wait);
+int on_each_cpu(void (*func) (void *info), void *info, int wait);
 
 #define MSG_ALL_BUT_SELF	0x8000	/* Assume <32768 CPU's */
 #define MSG_ALL			0x8001
@@ -90,9 +118,9 @@
 {
 	return 0;
 }
-#define smp_call_function(func, info, retry, wait) \
+#define smp_call_function(func, info, wait) \
 			(up_smp_call_function(func, info))
-#define on_each_cpu(func,info,retry,wait)	\
+#define on_each_cpu(func,info,wait)		\
 	({					\
 		local_irq_disable();		\
 		func(info);			\
@@ -102,7 +130,7 @@
 static inline void smp_send_reschedule(int cpu) { }
 #define num_booting_cpus()			1
 #define smp_prepare_boot_cpu()			do {} while (0)
-#define smp_call_function_single(cpuid, func, info, retry, wait) \
+#define smp_call_function_single(cpuid, func, info, wait) \
 ({ \
 	WARN_ON(cpuid != 0);	\
 	local_irq_disable();	\
@@ -112,7 +140,9 @@
 })
 #define smp_call_function_mask(mask, func, info, wait) \
 			(up_smp_call_function(func, info))
-
+static inline void init_call_single_data(void)
+{
+}
 #endif /* !SMP */
 
 /*