diff --git a/arch/ia64/Kconfig b/arch/ia64/Kconfig
index 56762d3c2a6a..8fa3faf5ef1b 100644
--- a/arch/ia64/Kconfig
+++ b/arch/ia64/Kconfig
@@ -156,6 +156,8 @@ config IA64_HP_ZX1_SWIOTLB
 
 config IA64_SGI_SN2
 	bool "SGI-SN2"
+	select NUMA
+	select ACPI_NUMA
 	help
 	  Selecting this option will optimize the kernel for use on sn2 based
 	  systems, but the resulting kernel binary will not run on other
diff --git a/arch/ia64/Makefile b/arch/ia64/Makefile
index b916ccfdef84..f1645c4f7039 100644
--- a/arch/ia64/Makefile
+++ b/arch/ia64/Makefile
@@ -11,6 +11,8 @@
 # Copyright (C) 1998-2004 by David Mosberger-Tang <davidm@hpl.hp.com>
 #
 
+KBUILD_DEFCONFIG := generic_defconfig
+
 NM := $(CROSS_COMPILE)nm -B
 READELF := $(CROSS_COMPILE)readelf
 
diff --git a/arch/ia64/defconfig b/arch/ia64/configs/generic_defconfig
similarity index 100%
rename from arch/ia64/defconfig
rename to arch/ia64/configs/generic_defconfig
diff --git a/arch/ia64/ia32/ia32_signal.c b/arch/ia64/ia32/ia32_signal.c
index 85e82f32e480..256a7faeda07 100644
--- a/arch/ia64/ia32/ia32_signal.c
+++ b/arch/ia64/ia32/ia32_signal.c
@@ -766,8 +766,19 @@ get_sigframe (struct k_sigaction *ka, struct pt_regs * regs, size_t frame_size)
 
 	/* This is the X/Open sanctioned signal stack switching.  */
 	if (ka->sa.sa_flags & SA_ONSTACK) {
-		if (!on_sig_stack(esp))
+		int onstack = sas_ss_flags(esp);
+
+		if (onstack == 0)
 			esp = current->sas_ss_sp + current->sas_ss_size;
+		else if (onstack == SS_ONSTACK) {
+			/*
+			 * If we are on the alternate signal stack and would
+			 * overflow it, don't. Return an always-bogus address
+			 * instead so we will die with SIGSEGV.
+			 */
+			if (!likely(on_sig_stack(esp - frame_size)))
+				return (void __user *) -1L;
+		}
 	}
 	/* Legacy stack switching not supported */
 
diff --git a/arch/ia64/kernel/iosapic.c b/arch/ia64/kernel/iosapic.c
index 398e2fd1cd25..7b3292282dea 100644
--- a/arch/ia64/kernel/iosapic.c
+++ b/arch/ia64/kernel/iosapic.c
@@ -345,7 +345,7 @@ iosapic_set_affinity (unsigned int irq, cpumask_t mask)
 	if (cpus_empty(mask))
 		return;
 
-	if (reassign_irq_vector(irq, first_cpu(mask)))
+	if (irq_prepare_move(irq, first_cpu(mask)))
 		return;
 
 	dest = cpu_physical_id(first_cpu(mask));
@@ -397,6 +397,7 @@ iosapic_end_level_irq (unsigned int irq)
 	struct iosapic_rte_info *rte;
 	int do_unmask_irq = 0;
 
+	irq_complete_move(irq);
 	if (unlikely(irq_desc[irq].status & IRQ_MOVE_PENDING)) {
 		do_unmask_irq = 1;
 		mask_irq(irq);
@@ -450,6 +451,7 @@ iosapic_ack_edge_irq (unsigned int irq)
 {
 	irq_desc_t *idesc = irq_desc + irq;
 
+	irq_complete_move(irq);
 	move_native_irq(irq);
 	/*
 	 * Once we have recorded IRQ_PENDING already, we can mask the
diff --git a/arch/ia64/kernel/irq_ia64.c b/arch/ia64/kernel/irq_ia64.c
index 0b52f19ed046..2b8cf6e85af4 100644
--- a/arch/ia64/kernel/irq_ia64.c
+++ b/arch/ia64/kernel/irq_ia64.c
@@ -260,6 +260,8 @@ void __setup_vector_irq(int cpu)
 }
 
 #if defined(CONFIG_SMP) && (defined(CONFIG_IA64_GENERIC) || defined(CONFIG_IA64_DIG))
+#define IA64_IRQ_MOVE_VECTOR	IA64_DEF_FIRST_DEVICE_VECTOR
+
 static enum vector_domain_type {
 	VECTOR_DOMAIN_NONE,
 	VECTOR_DOMAIN_PERCPU
@@ -272,6 +274,101 @@ static cpumask_t vector_allocation_domain(int cpu)
 	return CPU_MASK_ALL;
 }
 
+static int __irq_prepare_move(int irq, int cpu)
+{
+	struct irq_cfg *cfg = &irq_cfg[irq];
+	int vector;
+	cpumask_t domain;
+
+	if (cfg->move_in_progress || cfg->move_cleanup_count)
+		return -EBUSY;
+	if (cfg->vector == IRQ_VECTOR_UNASSIGNED || !cpu_online(cpu))
+		return -EINVAL;
+	if (cpu_isset(cpu, cfg->domain))
+		return 0;
+	domain = vector_allocation_domain(cpu);
+	vector = find_unassigned_vector(domain);
+	if (vector < 0)
+		return -ENOSPC;
+	cfg->move_in_progress = 1;
+	cfg->old_domain = cfg->domain;
+	cfg->vector = IRQ_VECTOR_UNASSIGNED;
+	cfg->domain = CPU_MASK_NONE;
+	BUG_ON(__bind_irq_vector(irq, vector, domain));
+	return 0;
+}
+
+int irq_prepare_move(int irq, int cpu)
+{
+	unsigned long flags;
+	int ret;
+
+	spin_lock_irqsave(&vector_lock, flags);
+	ret = __irq_prepare_move(irq, cpu);
+	spin_unlock_irqrestore(&vector_lock, flags);
+	return ret;
+}
+
+void irq_complete_move(unsigned irq)
+{
+	struct irq_cfg *cfg = &irq_cfg[irq];
+	cpumask_t cleanup_mask;
+	int i;
+
+	if (likely(!cfg->move_in_progress))
+		return;
+
+	if (unlikely(cpu_isset(smp_processor_id(), cfg->old_domain)))
+		return;
+
+	cpus_and(cleanup_mask, cfg->old_domain, cpu_online_map);
+	cfg->move_cleanup_count = cpus_weight(cleanup_mask);
+	for_each_cpu_mask(i, cleanup_mask)
+		platform_send_ipi(i, IA64_IRQ_MOVE_VECTOR, IA64_IPI_DM_INT, 0);
+	cfg->move_in_progress = 0;
+}
+
+static irqreturn_t smp_irq_move_cleanup_interrupt(int irq, void *dev_id)
+{
+	int me = smp_processor_id();
+	ia64_vector vector;
+	unsigned long flags;
+
+	for (vector = IA64_FIRST_DEVICE_VECTOR;
+	     vector < IA64_LAST_DEVICE_VECTOR; vector++) {
+		int irq;
+		struct irq_desc *desc;
+		struct irq_cfg *cfg;
+		irq = __get_cpu_var(vector_irq)[vector];
+		if (irq < 0)
+			continue;
+
+		desc = irq_desc + irq;
+		cfg = irq_cfg + irq;
+		spin_lock(&desc->lock);
+		if (!cfg->move_cleanup_count)
+			goto unlock;
+
+		if (!cpu_isset(me, cfg->old_domain))
+			goto unlock;
+
+		spin_lock_irqsave(&vector_lock, flags);
+		__get_cpu_var(vector_irq)[vector] = -1;
+		cpu_clear(me, vector_table[vector]);
+		spin_unlock_irqrestore(&vector_lock, flags);
+		cfg->move_cleanup_count--;
+	unlock:
+		spin_unlock(&desc->lock);
+	}
+	return IRQ_HANDLED;
+}
+
+static struct irqaction irq_move_irqaction = {
+	.handler =	smp_irq_move_cleanup_interrupt,
+	.flags =	IRQF_DISABLED,
+	.name =		"irq_move"
+};
+
 static int __init parse_vector_domain(char *arg)
 {
 	if (!arg)
@@ -303,36 +400,6 @@ void destroy_and_reserve_irq(unsigned int irq)
 	spin_unlock_irqrestore(&vector_lock, flags);
 }
 
-static int __reassign_irq_vector(int irq, int cpu)
-{
-	struct irq_cfg *cfg = &irq_cfg[irq];
-	int vector;
-	cpumask_t domain;
-
-	if (cfg->vector == IRQ_VECTOR_UNASSIGNED || !cpu_online(cpu))
-		return -EINVAL;
-	if (cpu_isset(cpu, cfg->domain))
-		return 0;
-	domain = vector_allocation_domain(cpu);
-	vector = find_unassigned_vector(domain);
-	if (vector < 0)
-		return -ENOSPC;
-	__clear_irq_vector(irq);
-	BUG_ON(__bind_irq_vector(irq, vector, domain));
-	return 0;
-}
-
-int reassign_irq_vector(int irq, int cpu)
-{
-	unsigned long flags;
-	int ret;
-
-	spin_lock_irqsave(&vector_lock, flags);
-	ret = __reassign_irq_vector(irq, cpu);
-	spin_unlock_irqrestore(&vector_lock, flags);
-	return ret;
-}
-
 /*
  * Dynamic irq allocate and deallocation for MSI
  */
@@ -578,6 +645,13 @@ init_IRQ (void)
 	register_percpu_irq(IA64_IPI_VECTOR, &ipi_irqaction);
 	register_percpu_irq(IA64_IPI_RESCHEDULE, &resched_irqaction);
 	register_percpu_irq(IA64_IPI_LOCAL_TLB_FLUSH, &tlb_irqaction);
+#if defined(CONFIG_IA64_GENERIC) || defined(CONFIG_IA64_DIG)
+	if (vector_domain_type != VECTOR_DOMAIN_NONE) {
+		BUG_ON(IA64_FIRST_DEVICE_VECTOR != IA64_IRQ_MOVE_VECTOR);
+		IA64_FIRST_DEVICE_VECTOR++;
+		register_percpu_irq(IA64_IRQ_MOVE_VECTOR, &irq_move_irqaction);
+	}
+#endif
 #endif
 #ifdef CONFIG_PERFMON
 	pfm_init_percpu();
diff --git a/arch/ia64/kernel/kprobes.c b/arch/ia64/kernel/kprobes.c
index b618487cdc85..615c3d2b6348 100644
--- a/arch/ia64/kernel/kprobes.c
+++ b/arch/ia64/kernel/kprobes.c
@@ -1001,6 +1001,11 @@ int __kprobes setjmp_pre_handler(struct kprobe *p, struct pt_regs *regs)
 	return 1;
 }
 
+/* ia64 does not need this */
+void __kprobes jprobe_return(void)
+{
+}
+
 int __kprobes longjmp_break_handler(struct kprobe *p, struct pt_regs *regs)
 {
 	struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
diff --git a/arch/ia64/kernel/msi_ia64.c b/arch/ia64/kernel/msi_ia64.c
index e86d02959794..60c6ef67ebb2 100644
--- a/arch/ia64/kernel/msi_ia64.c
+++ b/arch/ia64/kernel/msi_ia64.c
@@ -57,7 +57,7 @@ static void ia64_set_msi_irq_affinity(unsigned int irq, cpumask_t cpu_mask)
 	if (!cpu_online(cpu))
 		return;
 
-	if (reassign_irq_vector(irq, cpu))
+	if (irq_prepare_move(irq, cpu))
 		return;
 
 	read_msi_msg(irq, &msg);
@@ -119,6 +119,7 @@ void ia64_teardown_msi_irq(unsigned int irq)
 
 static void ia64_ack_msi_irq(unsigned int irq)
 {
+	irq_complete_move(irq);
 	move_native_irq(irq);
 	ia64_eoi();
 }
diff --git a/arch/ia64/kernel/sal.c b/arch/ia64/kernel/sal.c
index f44fe8412162..a3022dc48ef8 100644
--- a/arch/ia64/kernel/sal.c
+++ b/arch/ia64/kernel/sal.c
@@ -109,6 +109,13 @@ check_versions (struct ia64_sal_systab *systab)
 		sal_revision = SAL_VERSION_CODE(2, 8);
 		sal_version = SAL_VERSION_CODE(0, 0);
 	}
+
+	if (ia64_platform_is("sn2") && (sal_revision == SAL_VERSION_CODE(2, 9)))
+		/*
+		 * SGI Altix has hard-coded version 2.9 in their prom
+		 * but they actually implement 3.2, so let's fix it here.
+		 */
+		sal_revision = SAL_VERSION_CODE(3, 2);
 }
 
 static void __init
diff --git a/arch/ia64/kernel/signal.c b/arch/ia64/kernel/signal.c
index 309da3567bc8..5740296c35af 100644
--- a/arch/ia64/kernel/signal.c
+++ b/arch/ia64/kernel/signal.c
@@ -342,15 +342,33 @@ setup_frame (int sig, struct k_sigaction *ka, siginfo_t *info, sigset_t *set,
 
 	new_sp = scr->pt.r12;
 	tramp_addr = (unsigned long) __kernel_sigtramp;
-	if ((ka->sa.sa_flags & SA_ONSTACK) && sas_ss_flags(new_sp) == 0) {
-		new_sp = current->sas_ss_sp + current->sas_ss_size;
-		/*
-		 * We need to check for the register stack being on the signal stack
-		 * separately, because it's switched separately (memory stack is switched
-		 * in the kernel, register stack is switched in the signal trampoline).
-		 */
-		if (!rbs_on_sig_stack(scr->pt.ar_bspstore))
-			new_rbs = (current->sas_ss_sp + sizeof(long) - 1) & ~(sizeof(long) - 1);
+	if (ka->sa.sa_flags & SA_ONSTACK) {
+		int onstack = sas_ss_flags(new_sp);
+
+		if (onstack == 0) {
+			new_sp = current->sas_ss_sp + current->sas_ss_size;
+			/*
+			 * We need to check for the register stack being on the
+			 * signal stack separately, because it's switched
+			 * separately (memory stack is switched in the kernel,
+			 * register stack is switched in the signal trampoline).
+			 */
+			if (!rbs_on_sig_stack(scr->pt.ar_bspstore))
+				new_rbs = ALIGN(current->sas_ss_sp,
+						sizeof(long));
+		} else if (onstack == SS_ONSTACK) {
+			unsigned long check_sp;
+
+			/*
+			 * If we are on the alternate signal stack and would
+			 * overflow it, don't. Return an always-bogus address
+			 * instead so we will die with SIGSEGV.
+			 */
+			check_sp = (new_sp - sizeof(*frame)) & -STACK_ALIGN;
+			if (!likely(on_sig_stack(check_sp)))
+				return force_sigsegv_info(sig, (void __user *)
+							  check_sp);
+		}
 	}
 	frame = (void __user *) ((new_sp - sizeof(*frame)) & -STACK_ALIGN);
 
diff --git a/include/asm-ia64/Kbuild b/include/asm-ia64/Kbuild
index 4a1e48b9f403..eb24a3f47caa 100644
--- a/include/asm-ia64/Kbuild
+++ b/include/asm-ia64/Kbuild
@@ -3,7 +3,6 @@ include include/asm-generic/Kbuild.asm
 header-y += break.h
 header-y += fpu.h
 header-y += fpswa.h
-header-y += gcc_intrin.h
 header-y += ia64regs.h
 header-y += intel_intrin.h
 header-y += intrinsics.h
@@ -12,5 +11,6 @@ header-y += ptrace_offsets.h
 header-y += rse.h
 header-y += ucontext.h
 
+unifdef-y += gcc_intrin.h
 unifdef-y += perfmon.h
 unifdef-y += ustack.h
diff --git a/include/asm-ia64/hw_irq.h b/include/asm-ia64/hw_irq.h
index 7e6e3779670a..76366dc9c1a0 100644
--- a/include/asm-ia64/hw_irq.h
+++ b/include/asm-ia64/hw_irq.h
@@ -93,6 +93,9 @@ extern __u8 isa_irq_to_vector_map[16];
 struct irq_cfg {
 	ia64_vector vector;
 	cpumask_t domain;
+	cpumask_t old_domain;
+	unsigned move_cleanup_count;
+	u8 move_in_progress : 1;
 };
 extern spinlock_t vector_lock;
 extern struct irq_cfg irq_cfg[NR_IRQS];
@@ -106,12 +109,19 @@ extern int assign_irq_vector (int irq);	/* allocate a free vector */
 extern void free_irq_vector (int vector);
 extern int reserve_irq_vector (int vector);
 extern void __setup_vector_irq(int cpu);
-extern int reassign_irq_vector(int irq, int cpu);
 extern void ia64_send_ipi (int cpu, int vector, int delivery_mode, int redirect);
 extern void register_percpu_irq (ia64_vector vec, struct irqaction *action);
 extern int check_irq_used (int irq);
 extern void destroy_and_reserve_irq (unsigned int irq);
 
+#if defined(CONFIG_SMP) && (defined(CONFIG_IA64_GENERIC) || defined(CONFIG_IA64_DIG))
+extern int irq_prepare_move(int irq, int cpu);
+extern void irq_complete_move(unsigned int irq);
+#else
+static inline int irq_prepare_move(int irq, int cpu) { return 0; }
+static inline void irq_complete_move(unsigned int irq) {}
+#endif
+
 static inline void ia64_resend_irq(unsigned int vector)
 {
 	platform_send_ipi(smp_processor_id(), vector, IA64_IPI_DM_INT, 0);
diff --git a/include/asm-ia64/kprobes.h b/include/asm-ia64/kprobes.h
index adbaba14eb0a..8233b3a964c6 100644
--- a/include/asm-ia64/kprobes.h
+++ b/include/asm-ia64/kprobes.h
@@ -121,10 +121,6 @@ extern int kprobes_fault_handler(struct pt_regs *regs, int trapnr);
 extern int kprobe_exceptions_notify(struct notifier_block *self,
 				    unsigned long val, void *data);
 
-/* ia64 does not need this */
-static inline void jprobe_return(void)
-{
-}
 extern void invalidate_stacked_regs(void);
 extern void flush_register_stack(void);
 extern void arch_remove_kprobe(struct kprobe *p);
diff --git a/include/asm-ia64/sal.h b/include/asm-ia64/sal.h
index 2251118894ae..f4904db3b057 100644
--- a/include/asm-ia64/sal.h
+++ b/include/asm-ia64/sal.h
@@ -807,6 +807,10 @@ static inline s64
 ia64_sal_physical_id_info(u16 *splid)
 {
 	struct ia64_sal_retval isrv;
+
+	if (sal_revision < SAL_VERSION_CODE(3,2))
+		return -1;
+
 	SAL_CALL(isrv, SAL_PHYSICAL_ID_INFO, 0, 0, 0, 0, 0, 0, 0);
 	if (splid)
 		*splid = isrv.v0;