Auto merge with /home/aegl/GIT/linus

16 files changed, 336 insertions, 104 deletions

diff --git a/arch/ia64/kernel/acpi.c b/arch/ia64/kernel/acpi.c
index a8e99c56a76..72dfd9e7de0 100644
--- a/arch/ia64/kernel/acpi.c
+++ b/arch/ia64/kernel/acpi.c
@@ -779,7 +779,7 @@ acpi_map_iosapic (acpi_handle handle, u32 depth, void *context, void **ret)
 	union acpi_object *obj;
 	struct acpi_table_iosapic *iosapic;
 	unsigned int gsi_base;
-	int node;
+	int pxm, node;
 
 	/* Only care about objects w/ a method that returns the MADT */
 	if (ACPI_FAILURE(acpi_evaluate_object(handle, "_MAT", NULL, &buffer)))
@@ -805,29 +805,16 @@ acpi_map_iosapic (acpi_handle handle, u32 depth, void *context, void **ret)
 	gsi_base = iosapic->global_irq_base;
 
 	acpi_os_free(buffer.pointer);
-	buffer.length = ACPI_ALLOCATE_BUFFER;
-	buffer.pointer = NULL;
 
 	/*
-	 * OK, it's an IOSAPIC MADT entry, look for a _PXM method to tell
+	 * OK, it's an IOSAPIC MADT entry, look for a _PXM value to tell
 	 * us which node to associate this with.
 	 */
-	if (ACPI_FAILURE(acpi_evaluate_object(handle, "_PXM", NULL, &buffer)))
-		return AE_OK;
-
-	if (!buffer.length || !buffer.pointer)
-		return AE_OK;
-
-	obj = buffer.pointer;
-
-	if (obj->type != ACPI_TYPE_INTEGER ||
-	    obj->integer.value >= MAX_PXM_DOMAINS) {
-		acpi_os_free(buffer.pointer);
+	pxm = acpi_get_pxm(handle);
+	if (pxm < 0)
 		return AE_OK;
-	}
 
-	node = pxm_to_nid_map[obj->integer.value];
-	acpi_os_free(buffer.pointer);
+	node = pxm_to_nid_map[pxm];
 
 	if (node >= MAX_NUMNODES || !node_online(node) ||
 	    cpus_empty(node_to_cpumask(node)))
diff --git a/arch/ia64/kernel/entry.S b/arch/ia64/kernel/entry.S
index 9353adc1895..8ebfcc0c813 100644
--- a/arch/ia64/kernel/entry.S
+++ b/arch/ia64/kernel/entry.S
@@ -777,7 +777,7 @@ GLOBAL_ENTRY(ia64_ret_from_ia32_execve)
 	st8.spill [r2]=r8	// store return value in slot for r8 and set unat bit
 	.mem.offset 8,0
 	st8.spill [r3]=r0	// clear error indication in slot for r10 and set unat bit
-END(ia64_ret_from_ia32_execve_syscall)
+END(ia64_ret_from_ia32_execve)
 	// fall through
 #endif /* CONFIG_IA32_SUPPORT */
 GLOBAL_ENTRY(ia64_leave_kernel)
@@ -1176,7 +1176,7 @@ ENTRY(notify_resume_user)
 	;;
 (pNonSys) mov out2=0				// out2==0 => not a syscall
 	.fframe 16
-	.spillpsp ar.unat, 16			// (note that offset is relative to psp+0x10!)
+	.spillsp ar.unat, 16
 	st8 [sp]=r9,-16				// allocate space for ar.unat and save it
 	st8 [out1]=loc1,-8			// save ar.pfs, out1=&sigscratch
 	.body
@@ -1202,7 +1202,7 @@ GLOBAL_ENTRY(sys_rt_sigsuspend)
 	adds out2=8,sp				// out2=&sigscratch->ar_pfs
 	;;
 	.fframe 16
-	.spillpsp ar.unat, 16			// (note that offset is relative to psp+0x10!)
+	.spillsp ar.unat, 16
 	st8 [sp]=r9,-16				// allocate space for ar.unat and save it
 	st8 [out2]=loc1,-8			// save ar.pfs, out2=&sigscratch
 	.body
diff --git a/arch/ia64/kernel/fsys.S b/arch/ia64/kernel/fsys.S
index f566ff43a38..7d7684a369d 100644
--- a/arch/ia64/kernel/fsys.S
+++ b/arch/ia64/kernel/fsys.S
@@ -460,9 +460,9 @@ EX(.fail_efault, ld8 r14=[r33])			// r14 <- *set
 	;;
 
 	st8 [r2]=r14				// update current->blocked with new mask
-	cmpxchg4.acq r14=[r9],r18,ar.ccv	// current->thread_info->flags <- r18
+	cmpxchg4.acq r8=[r9],r18,ar.ccv		// current->thread_info->flags <- r18
 	;;
-	cmp.ne p6,p0=r17,r14			// update failed?
+	cmp.ne p6,p0=r17,r8			// update failed?
 (p6)	br.cond.spnt.few 1b			// yes -> retry
 
 #ifdef CONFIG_SMP
diff --git a/arch/ia64/kernel/mca.c b/arch/ia64/kernel/mca.c
index 4d6c7b8f667..736e328b5e6 100644
--- a/arch/ia64/kernel/mca.c
+++ b/arch/ia64/kernel/mca.c
@@ -1103,8 +1103,6 @@ ia64_mca_cpe_int_caller(int cpe_irq, void *arg, struct pt_regs *ptregs)
 	return IRQ_HANDLED;
 }
 
-#endif /* CONFIG_ACPI */
-
 /*
  *  ia64_mca_cpe_poll
  *
@@ -1122,6 +1120,8 @@ ia64_mca_cpe_poll (unsigned long dummy)
 	platform_send_ipi(first_cpu(cpu_online_map), IA64_CPEP_VECTOR, IA64_IPI_DM_INT, 0);
 }
 
+#endif /* CONFIG_ACPI */
+
 /*
  * C portion of the OS INIT handler
  *
@@ -1390,8 +1390,7 @@ ia64_mca_init(void)
 	register_percpu_irq(IA64_MCA_WAKEUP_VECTOR, &mca_wkup_irqaction);
 
 #ifdef CONFIG_ACPI
-	/* Setup the CPEI/P vector and handler */
-	cpe_vector = acpi_request_vector(ACPI_INTERRUPT_CPEI);
+	/* Setup the CPEI/P handler */
 	register_percpu_irq(IA64_CPEP_VECTOR, &mca_cpep_irqaction);
 #endif
 
@@ -1436,6 +1435,7 @@ ia64_mca_late_init(void)
 
 #ifdef CONFIG_ACPI
 	/* Setup the CPEI/P vector and handler */
+	cpe_vector = acpi_request_vector(ACPI_INTERRUPT_CPEI);
 	init_timer(&cpe_poll_timer);
 	cpe_poll_timer.function = ia64_mca_cpe_poll;
 
diff --git a/arch/ia64/kernel/mca_drv.c b/arch/ia64/kernel/mca_drv.c
index ab478172c34..abc0113a821 100644
--- a/arch/ia64/kernel/mca_drv.c
+++ b/arch/ia64/kernel/mca_drv.c
@@ -132,8 +132,7 @@ mca_handler_bh(unsigned long paddr)
 	spin_unlock(&mca_bh_lock);
 
 	/* This process is about to be killed itself */
-	force_sig(SIGKILL, current);
-	schedule();
+	do_exit(SIGKILL);
 }
 
 /**
@@ -439,6 +438,7 @@ recover_from_read_error(slidx_table_t *slidx, peidx_table_t *peidx, pal_bus_chec
 			psr2 = (struct ia64_psr *)&pmsa->pmsa_ipsr;
 			psr2->cpl = 0;
 			psr2->ri  = 0;
+			psr2->i  = 0;
 
 			return 1;
 		}
diff --git a/arch/ia64/kernel/mca_drv_asm.S b/arch/ia64/kernel/mca_drv_asm.S
index bcfa05acc56..2d7e0217638 100644
--- a/arch/ia64/kernel/mca_drv_asm.S
+++ b/arch/ia64/kernel/mca_drv_asm.S
@@ -10,6 +10,7 @@
 
 #include <asm/asmmacro.h>
 #include <asm/processor.h>
+#include <asm/ptrace.h>
 
 GLOBAL_ENTRY(mca_handler_bhhook)
 	invala						// clear RSE ?
@@ -20,12 +21,21 @@ GLOBAL_ENTRY(mca_handler_bhhook)
 	;;						
 	alloc		r16=ar.pfs,0,2,1,0		// make a new frame
 	;;
+	mov		ar.rsc=0
+	;;
 	mov		r13=IA64_KR(CURRENT)		// current task pointer
 	;;
-	adds		r12=IA64_TASK_THREAD_KSP_OFFSET,r13
+	mov		r2=r13
+	;;
+	addl		r22=IA64_RBS_OFFSET,r2
+	;;
+	mov		ar.bspstore=r22
 	;;
-	ld8		r12=[r12]			// stack pointer
+	addl		sp=IA64_STK_OFFSET-IA64_PT_REGS_SIZE,r2
 	;;
+	adds		r2=IA64_TASK_THREAD_ON_USTACK_OFFSET,r13
+	;;
+	st1		[r2]=r0				// clear current->thread.on_ustack flag
 	mov		loc0=r16
 	movl		loc1=mca_handler_bh		// recovery C function
 	;;
@@ -34,7 +44,9 @@ GLOBAL_ENTRY(mca_handler_bhhook)
 	;;
 	mov		loc1=rp
 	;;
-	br.call.sptk.many    rp=b6			// not return ...
+	ssm		psr.i
+	;;
+	br.call.sptk.many    rp=b6			// does not return ...
 	;;
 	mov		ar.pfs=loc0
 	mov 		rp=loc1
diff --git a/arch/ia64/kernel/minstate.h b/arch/ia64/kernel/minstate.h
index 1dbc7b2497c..f6d8a010d99 100644
--- a/arch/ia64/kernel/minstate.h
+++ b/arch/ia64/kernel/minstate.h
@@ -41,7 +41,7 @@
 (pKStk) addl r3=THIS_CPU(ia64_mca_data),r3;;							\
 (pKStk) ld8 r3 = [r3];;										\
 (pKStk) addl r3=IA64_MCA_CPU_INIT_STACK_OFFSET,r3;;						\
-(pKStk) addl sp=IA64_STK_OFFSET-IA64_PT_REGS_SIZE,r3;						\
+(pKStk) addl r1=IA64_STK_OFFSET-IA64_PT_REGS_SIZE,r3;						\
 (pUStk)	mov ar.rsc=0;		/* set enforced lazy mode, pl 0, little-endian, loadrs=0 */	\
 (pUStk)	addl r22=IA64_RBS_OFFSET,r1;		/* compute base of register backing store */	\
 	;;											\
@@ -50,7 +50,6 @@
 (pUStk)	mov r23=ar.bspstore;				/* save ar.bspstore */			\
 (pUStk)	dep r22=-1,r22,61,3;			/* compute kernel virtual addr of RBS */	\
 	;;											\
-(pKStk) addl r1=-IA64_PT_REGS_SIZE,r1;		/* if in kernel mode, use sp (r12) */		\
 (pUStk)	mov ar.bspstore=r22;			/* switch to kernel RBS */			\
 	;;											\
 (pUStk)	mov r18=ar.bsp;										\
diff --git a/arch/ia64/kernel/module.c b/arch/ia64/kernel/module.c
index febc091c2f0..f1aca7cffd1 100644
--- a/arch/ia64/kernel/module.c
+++ b/arch/ia64/kernel/module.c
@@ -825,14 +825,16 @@ apply_relocate_add (Elf64_Shdr *sechdrs, const char *strtab, unsigned int symind
 		 * XXX Should have an arch-hook for running this after final section
 		 *     addresses have been selected...
 		 */
-		/* See if gp can cover the entire core module:  */
-		uint64_t gp = (uint64_t) mod->module_core + MAX_LTOFF / 2;
-		if (mod->core_size >= MAX_LTOFF)
+		uint64_t gp;
+		if (mod->core_size > MAX_LTOFF)
 			/*
 			 * This takes advantage of fact that SHF_ARCH_SMALL gets allocated
 			 * at the end of the module.
 			 */
-			gp = (uint64_t) mod->module_core + mod->core_size - MAX_LTOFF / 2;
+			gp = mod->core_size - MAX_LTOFF / 2;
+		else
+			gp = mod->core_size / 2;
+		gp = (uint64_t) mod->module_core + ((gp + 7) & -8);
 		mod->arch.gp = gp;
 		DEBUGP("%s: placing gp at 0x%lx\n", __FUNCTION__, gp);
 	}
diff --git a/arch/ia64/kernel/perfmon.c b/arch/ia64/kernel/perfmon.c
index 376fcbc3f8d..6407bff6bfd 100644
--- a/arch/ia64/kernel/perfmon.c
+++ b/arch/ia64/kernel/perfmon.c
@@ -11,7 +11,7 @@
  * Version Perfmon-2.x is a rewrite of perfmon-1.x
  * by Stephane Eranian, Hewlett Packard Co.
  *
- * Copyright (C) 1999-2003, 2005  Hewlett Packard Co
+ * Copyright (C) 1999-2005  Hewlett Packard Co
  *               Stephane Eranian <eranian@hpl.hp.com>
  *               David Mosberger-Tang <davidm@hpl.hp.com>
  *
@@ -497,6 +497,9 @@ typedef struct {
 static pfm_stats_t		pfm_stats[NR_CPUS];
 static pfm_session_t		pfm_sessions;	/* global sessions information */
 
+static spinlock_t pfm_alt_install_check = SPIN_LOCK_UNLOCKED;
+static pfm_intr_handler_desc_t  *pfm_alt_intr_handler;
+
 static struct proc_dir_entry 	*perfmon_dir;
 static pfm_uuid_t		pfm_null_uuid = {0,};
 
@@ -606,6 +609,7 @@ DEFINE_PER_CPU(unsigned long, pfm_syst_info);
 DEFINE_PER_CPU(struct task_struct *, pmu_owner);
 DEFINE_PER_CPU(pfm_context_t  *, pmu_ctx);
 DEFINE_PER_CPU(unsigned long, pmu_activation_number);
+EXPORT_PER_CPU_SYMBOL_GPL(pfm_syst_info);
 
 
 /* forward declaration */
@@ -1265,6 +1269,8 @@ out:
 }
 EXPORT_SYMBOL(pfm_unregister_buffer_fmt);
 
+extern void update_pal_halt_status(int);
+
 static int
 pfm_reserve_session(struct task_struct *task, int is_syswide, unsigned int cpu)
 {
@@ -1311,6 +1317,11 @@ pfm_reserve_session(struct task_struct *task, int is_syswide, unsigned int cpu)
 		is_syswide,
 		cpu));
 
+	/*
+	 * disable default_idle() to go to PAL_HALT
+	 */
+	update_pal_halt_status(0);
+
 	UNLOCK_PFS(flags);
 
 	return 0;
@@ -1318,7 +1329,7 @@ pfm_reserve_session(struct task_struct *task, int is_syswide, unsigned int cpu)
 error_conflict:
 	DPRINT(("system wide not possible, conflicting session [%d] on CPU%d\n",
   		pfm_sessions.pfs_sys_session[cpu]->pid,
-		smp_processor_id()));
+		cpu));
 abort:
 	UNLOCK_PFS(flags);
 
@@ -1366,6 +1377,12 @@ pfm_unreserve_session(pfm_context_t *ctx, int is_syswide, unsigned int cpu)
 		is_syswide,
 		cpu));
 
+	/*
+	 * if possible, enable default_idle() to go into PAL_HALT
+	 */
+	if (pfm_sessions.pfs_task_sessions == 0 && pfm_sessions.pfs_sys_sessions == 0)
+		update_pal_halt_status(1);
+
 	UNLOCK_PFS(flags);
 
 	return 0;
@@ -4202,7 +4219,7 @@ pfm_context_load(pfm_context_t *ctx, void *arg, int count, struct pt_regs *regs)
 		DPRINT(("cannot load to [%d], invalid ctx_state=%d\n",
 			req->load_pid,
 			ctx->ctx_state));
-		return -EINVAL;
+		return -EBUSY;
 	}
 
 	DPRINT(("load_pid [%d] using_dbreg=%d\n", req->load_pid, ctx->ctx_fl_using_dbreg));
@@ -4704,16 +4721,26 @@ recheck:
 	if (task == current || ctx->ctx_fl_system) return 0;
 
 	/*
-	 * if context is UNLOADED we are safe to go
+	 * we are monitoring another thread
 	 */
-	if (state == PFM_CTX_UNLOADED) return 0;
-
-	/*
-	 * no command can operate on a zombie context
-	 */
-	if (state == PFM_CTX_ZOMBIE) {
-		DPRINT(("cmd %d state zombie cannot operate on context\n", cmd));
-		return -EINVAL;
+	switch(state) {
+		case PFM_CTX_UNLOADED:
+			/*
+			 * if context is UNLOADED we are safe to go
+			 */
+			return 0;
+		case PFM_CTX_ZOMBIE:
+			/*
+			 * no command can operate on a zombie context
+			 */
+			DPRINT(("cmd %d state zombie cannot operate on context\n", cmd));
+			return -EINVAL;
+		case PFM_CTX_MASKED:
+			/*
+			 * PMU state has been saved to software even though
+			 * the thread may still be running.
+			 */
+			if (cmd != PFM_UNLOAD_CONTEXT) return 0;
 	}
 
 	/*
@@ -5532,26 +5559,32 @@ pfm_interrupt_handler(int irq, void *arg, struct pt_regs *regs)
 	int ret;
 
 	this_cpu = get_cpu();
-	min      = pfm_stats[this_cpu].pfm_ovfl_intr_cycles_min;
-	max      = pfm_stats[this_cpu].pfm_ovfl_intr_cycles_max;
+	if (likely(!pfm_alt_intr_handler)) {
+		min = pfm_stats[this_cpu].pfm_ovfl_intr_cycles_min;
+		max = pfm_stats[this_cpu].pfm_ovfl_intr_cycles_max;
 
-	start_cycles = ia64_get_itc();
+		start_cycles = ia64_get_itc();
 
-	ret = pfm_do_interrupt_handler(irq, arg, regs);
+		ret = pfm_do_interrupt_handler(irq, arg, regs);
 
-	total_cycles = ia64_get_itc();
+		total_cycles = ia64_get_itc();
 
-	/*
-	 * don't measure spurious interrupts
-	 */
-	if (likely(ret == 0)) {
-		total_cycles -= start_cycles;
+		/*
+		 * don't measure spurious interrupts
+		 */
+		if (likely(ret == 0)) {
+			total_cycles -= start_cycles;
 
-		if (total_cycles < min) pfm_stats[this_cpu].pfm_ovfl_intr_cycles_min = total_cycles;
-		if (total_cycles > max) pfm_stats[this_cpu].pfm_ovfl_intr_cycles_max = total_cycles;
+			if (total_cycles < min) pfm_stats[this_cpu].pfm_ovfl_intr_cycles_min = total_cycles;
+			if (total_cycles > max) pfm_stats[this_cpu].pfm_ovfl_intr_cycles_max = total_cycles;
 
-		pfm_stats[this_cpu].pfm_ovfl_intr_cycles += total_cycles;
+			pfm_stats[this_cpu].pfm_ovfl_intr_cycles += total_cycles;
+		}
 	}
+	else {
+		(*pfm_alt_intr_handler->handler)(irq, arg, regs);
+	}
+
 	put_cpu_no_resched();
 	return IRQ_HANDLED;
 }
@@ -6402,6 +6435,141 @@ static struct irqaction perfmon_irqaction = {
 	.name    = "perfmon"
 };
 
+static void
+pfm_alt_save_pmu_state(void *data)
+{
+	struct pt_regs *regs;
+
+	regs = ia64_task_regs(current);
+
+	DPRINT(("called\n"));
+
+	/*
+	 * should not be necessary but
+	 * let's take not risk
+	 */
+	pfm_clear_psr_up();
+	pfm_clear_psr_pp();
+	ia64_psr(regs)->pp = 0;
+
+	/*
+	 * This call is required
+	 * May cause a spurious interrupt on some processors
+	 */
+	pfm_freeze_pmu();
+
+	ia64_srlz_d();
+}
+
+void
+pfm_alt_restore_pmu_state(void *data)
+{
+	struct pt_regs *regs;
+
+	regs = ia64_task_regs(current);
+
+	DPRINT(("called\n"));
+
+	/*
+	 * put PMU back in state expected
+	 * by perfmon
+	 */
+	pfm_clear_psr_up();
+	pfm_clear_psr_pp();
+	ia64_psr(regs)->pp = 0;
+
+	/*
+	 * perfmon runs with PMU unfrozen at all times
+	 */
+	pfm_unfreeze_pmu();
+
+	ia64_srlz_d();
+}
+
+int
+pfm_install_alt_pmu_interrupt(pfm_intr_handler_desc_t *hdl)
+{
+	int ret, i;
+	int reserve_cpu;
+
+	/* some sanity checks */
+	if (hdl == NULL || hdl->handler == NULL) return -EINVAL;
+
+	/* do the easy test first */
+	if (pfm_alt_intr_handler) return -EBUSY;
+
+	/* one at a time in the install or remove, just fail the others */
+	if (!spin_trylock(&pfm_alt_install_check)) {
+		return -EBUSY;
+	}
+
+	/* reserve our session */
+	for_each_online_cpu(reserve_cpu) {
+		ret = pfm_reserve_session(NULL, 1, reserve_cpu);
+		if (ret) goto cleanup_reserve;
+	}
+
+	/* save the current system wide pmu states */
+	ret = on_each_cpu(pfm_alt_save_pmu_state, NULL, 0, 1);
+	if (ret) {
+		DPRINT(("on_each_cpu() failed: %d\n", ret));
+		goto cleanup_reserve;
+	}
+
+	/* officially change to the alternate interrupt handler */
+	pfm_alt_intr_handler = hdl;
+
+	spin_unlock(&pfm_alt_install_check);
+
+	return 0;
+
+cleanup_reserve:
+	for_each_online_cpu(i) {
+		/* don't unreserve more than we reserved */
+		if (i >= reserve_cpu) break;
+
+		pfm_unreserve_session(NULL, 1, i);
+	}
+
+	spin_unlock(&pfm_alt_install_check);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(pfm_install_alt_pmu_interrupt);
+
+int
+pfm_remove_alt_pmu_interrupt(pfm_intr_handler_desc_t *hdl)
+{
+	int i;
+	int ret;
+
+	if (hdl == NULL) return -EINVAL;
+
+	/* cannot remove someone else's handler! */
+	if (pfm_alt_intr_handler != hdl) return -EINVAL;
+
+	/* one at a time in the install or remove, just fail the others */
+	if (!spin_trylock(&pfm_alt_install_check)) {
+		return -EBUSY;
+	}
+
+	pfm_alt_intr_handler = NULL;
+
+	ret = on_each_cpu(pfm_alt_restore_pmu_state, NULL, 0, 1);
+	if (ret) {
+		DPRINT(("on_each_cpu() failed: %d\n", ret));
+	}
+
+	for_each_online_cpu(i) {
+		pfm_unreserve_session(NULL, 1, i);
+	}
+
+	spin_unlock(&pfm_alt_install_check);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(pfm_remove_alt_pmu_interrupt);
+
 /*
  * perfmon initialization routine, called from the initcall() table
  */
diff --git a/arch/ia64/kernel/process.c b/arch/ia64/kernel/process.c
index 7c43aea5f7f..ebb71f3d6d1 100644
--- a/arch/ia64/kernel/process.c
+++ b/arch/ia64/kernel/process.c
@@ -50,7 +50,7 @@
 #include "sigframe.h"
 
 void (*ia64_mark_idle)(int);
-static cpumask_t cpu_idle_map;
+static DEFINE_PER_CPU(unsigned int, cpu_idle_state);
 
 unsigned long boot_option_idle_override = 0;
 EXPORT_SYMBOL(boot_option_idle_override);
@@ -173,7 +173,9 @@ do_notify_resume_user (sigset_t *oldset, struct sigscratch *scr, long in_syscall
 		ia64_do_signal(oldset, scr, in_syscall);
 }
 
-static int pal_halt = 1;
+static int pal_halt        = 1;
+static int can_do_pal_halt = 1;
+
 static int __init nohalt_setup(char * str)
 {
 	pal_halt = 0;
@@ -181,16 +183,20 @@ static int __init nohalt_setup(char * str)
 }
 __setup("nohalt", nohalt_setup);
 
+void
+update_pal_halt_status(int status)
+{
+	can_do_pal_halt = pal_halt && status;
+}
+
 /*
  * We use this if we don't have any better idle routine..
  */
 void
 default_idle (void)
 {
-	unsigned long pmu_active = ia64_getreg(_IA64_REG_PSR) & (IA64_PSR_PP | IA64_PSR_UP);
-
 	while (!need_resched())
-		if (pal_halt && !pmu_active)
+		if (can_do_pal_halt)
 			safe_halt();
 		else
 			cpu_relax();
@@ -223,20 +229,31 @@ static inline void play_dead(void)
 }
 #endif /* CONFIG_HOTPLUG_CPU */
 
-
 void cpu_idle_wait(void)
 {
-        int cpu;
-        cpumask_t map;
+	unsigned int cpu, this_cpu = get_cpu();
+	cpumask_t map;
 
-        for_each_online_cpu(cpu)
-                cpu_set(cpu, cpu_idle_map);
+	set_cpus_allowed(current, cpumask_of_cpu(this_cpu));
+	put_cpu();
 
-        wmb();
-        do {
-                ssleep(1);
-                cpus_and(map, cpu_idle_map, cpu_online_map);
-        } while (!cpus_empty(map));
+	cpus_clear(map);
+	for_each_online_cpu(cpu) {
+		per_cpu(cpu_idle_state, cpu) = 1;
+		cpu_set(cpu, map);
+	}
+
+	__get_cpu_var(cpu_idle_state) = 0;
+
+	wmb();
+	do {
+		ssleep(1);
+		for_each_online_cpu(cpu) {
+			if (cpu_isset(cpu, map) && !per_cpu(cpu_idle_state, cpu))
+				cpu_clear(cpu, map);
+		}
+		cpus_and(map, map, cpu_online_map);
+	} while (!cpus_empty(map));
 }
 EXPORT_SYMBOL_GPL(cpu_idle_wait);
 
@@ -244,7 +261,6 @@ void __attribute__((noreturn))
 cpu_idle (void)
 {
 	void (*mark_idle)(int) = ia64_mark_idle;
-	int cpu = smp_processor_id();
 
 	/* endless idle loop with no priority at all */
 	while (1) {
@@ -255,12 +271,13 @@ cpu_idle (void)
 		while (!need_resched()) {
 			void (*idle)(void);
 
+			if (__get_cpu_var(cpu_idle_state))
+				__get_cpu_var(cpu_idle_state) = 0;
+
+			rmb();
 			if (mark_idle)
 				(*mark_idle)(1);
 
-			if (cpu_isset(cpu, cpu_idle_map))
-				cpu_clear(cpu, cpu_idle_map);
-			rmb();
 			idle = pm_idle;
 			if (!idle)
 				idle = default_idle;
diff --git a/arch/ia64/kernel/ptrace.c b/arch/ia64/kernel/ptrace.c
index 9e730c7bf0c..4c1d2f5442b 100644
--- a/arch/ia64/kernel/ptrace.c
+++ b/arch/ia64/kernel/ptrace.c
@@ -635,11 +635,17 @@ ia64_flush_fph (struct task_struct *task)
 {
 	struct ia64_psr *psr = ia64_psr(ia64_task_regs(task));
 
+	/*
+	 * Prevent migrating this task while
+	 * we're fiddling with the FPU state
+	 */
+	preempt_disable();
 	if (ia64_is_local_fpu_owner(task) && psr->mfh) {
 		psr->mfh = 0;
 		task->thread.flags |= IA64_THREAD_FPH_VALID;
 		ia64_save_fpu(&task->thread.fph[0]);
 	}
+	preempt_enable();
 }
 
 /*
@@ -692,16 +698,30 @@ convert_to_non_syscall (struct task_struct *child, struct pt_regs  *pt,
 			unsigned long cfm)
 {
 	struct unw_frame_info info, prev_info;
-	unsigned long ip, pr;
+	unsigned long ip, sp, pr;
 
 	unw_init_from_blocked_task(&info, child);
 	while (1) {
 		prev_info = info;
 		if (unw_unwind(&info) < 0)
 			return;
-		if (unw_get_rp(&info, &ip) < 0)
+
+		unw_get_sp(&info, &sp);
+		if ((long)((unsigned long)child + IA64_STK_OFFSET - sp)
+		    < IA64_PT_REGS_SIZE) {
+			dprintk("ptrace.%s: ran off the top of the kernel "
+				"stack\n", __FUNCTION__);
+			return;
+		}
+		if (unw_get_pr (&prev_info, &pr) < 0) {
+			unw_get_rp(&prev_info, &ip);
+			dprintk("ptrace.%s: failed to read "
+				"predicate register (ip=0x%lx)\n",
+				__FUNCTION__, ip);
 			return;
-		if (ip < FIXADDR_USER_END)
+		}
+		if (unw_is_intr_frame(&info)
+		    && (pr & (1UL << PRED_USER_STACK)))
 			break;
 	}
 
@@ -1616,20 +1636,25 @@ syscall_trace_enter (long arg0, long arg1, long arg2, long arg3,
 		     long arg4, long arg5, long arg6, long arg7,
 		     struct pt_regs regs)
 {
-	long syscall;
+	if (test_thread_flag(TIF_SYSCALL_TRACE) 
+	    && (current->ptrace & PT_PTRACED))
+		syscall_trace();
 
 	if (unlikely(current->audit_context)) {
-		if (IS_IA32_PROCESS(&regs))
+		long syscall;
+		int arch;
+
+		if (IS_IA32_PROCESS(&regs)) {
 			syscall = regs.r1;
-		else
+			arch = AUDIT_ARCH_I386;
+		} else {
 			syscall = regs.r15;
+			arch = AUDIT_ARCH_IA64;
+		}
 
-		audit_syscall_entry(current, syscall, arg0, arg1, arg2, arg3);
+		audit_syscall_entry(current, arch, syscall, arg0, arg1, arg2, arg3);
 	}
 
-	if (test_thread_flag(TIF_SYSCALL_TRACE)
-	    && (current->ptrace & PT_PTRACED))
-		syscall_trace();
 }
 
 /* "asmlinkage" so the input arguments are preserved... */
@@ -1640,7 +1665,7 @@ syscall_trace_leave (long arg0, long arg1, long arg2, long arg3,
 		     struct pt_regs regs)
 {
 	if (unlikely(current->audit_context))
-		audit_syscall_exit(current, regs.r8);
+		audit_syscall_exit(current, AUDITSC_RESULT(regs.r10), regs.r8);
 
 	if (test_thread_flag(TIF_SYSCALL_TRACE)
 	    && (current->ptrace & PT_PTRACED))
diff --git a/arch/ia64/kernel/setup.c b/arch/ia64/kernel/setup.c
index b7e6b4cb374..d14692e0920 100644
--- a/arch/ia64/kernel/setup.c
+++ b/arch/ia64/kernel/setup.c
@@ -720,7 +720,8 @@ cpu_init (void)
 	ia64_set_kr(IA64_KR_PT_BASE, __pa(ia64_imva(empty_zero_page)));
 
 	/*
-	 * Initialize default control register to defer all speculative faults.  The
+	 * Initialize default control register to defer speculative faults except
+	 * for those arising from TLB misses, which are not deferred.  The
 	 * kernel MUST NOT depend on a particular setting of these bits (in other words,
 	 * the kernel must have recovery code for all speculative accesses).  Turn on
 	 * dcr.lc as per recommendation by the architecture team.  Most IA-32 apps
diff --git a/arch/ia64/kernel/signal.c b/arch/ia64/kernel/signal.c
index 6891d86937d..499b7e5317c 100644
--- a/arch/ia64/kernel/signal.c
+++ b/arch/ia64/kernel/signal.c
@@ -224,7 +224,8 @@ ia64_rt_sigreturn (struct sigscratch *scr)
 	 * could be corrupted.
 	 */
 	retval = (long) &ia64_leave_kernel;
-	if (test_thread_flag(TIF_SYSCALL_TRACE))
+	if (test_thread_flag(TIF_SYSCALL_TRACE)
+	    || test_thread_flag(TIF_SYSCALL_AUDIT))
 		/*
 		 * strace expects to be notified after sigreturn returns even though the
 		 * context to which we return may not be in the middle of a syscall.
diff --git a/arch/ia64/kernel/smpboot.c b/arch/ia64/kernel/smpboot.c
index 0d5ee57c986..3865f088ffa 100644
--- a/arch/ia64/kernel/smpboot.c
+++ b/arch/ia64/kernel/smpboot.c
@@ -624,7 +624,7 @@ static struct {
 	__u16   thread_id;
 	__u16   proc_fixed_addr;
 	__u8    valid;
-}mt_info[NR_CPUS] __devinit;
+} mt_info[NR_CPUS] __devinitdata;
 
 #ifdef CONFIG_HOTPLUG_CPU
 static inline void
diff --git a/arch/ia64/kernel/sys_ia64.c b/arch/ia64/kernel/sys_ia64.c
index a8cf6d8a509..770fab37928 100644
--- a/arch/ia64/kernel/sys_ia64.c
+++ b/arch/ia64/kernel/sys_ia64.c
@@ -182,13 +182,6 @@ do_mmap2 (unsigned long addr, unsigned long len, int prot, int flags, int fd, un
 		}
 	}
 
-	/*
-	 * A zero mmap always succeeds in Linux, independent of whether or not the
-	 * remaining arguments are valid.
-	 */
-	if (len == 0)
-		goto out;
-
 	/* Careful about overflows.. */
 	len = PAGE_ALIGN(len);
 	if (!len || len > TASK_SIZE) {
diff --git a/arch/ia64/kernel/traps.c b/arch/ia64/kernel/traps.c
index e82ad78081b..1861173bd4f 100644
--- a/arch/ia64/kernel/traps.c
+++ b/arch/ia64/kernel/traps.c
@@ -111,6 +111,24 @@ ia64_bad_break (unsigned long break_num, struct pt_regs *regs)
 	siginfo_t siginfo;
 	int sig, code;
 
+	/* break.b always sets cr.iim to 0, which causes problems for
+	 * debuggers.  Get the real break number from the original instruction,
+	 * but only for kernel code.  User space break.b is left alone, to
+	 * preserve the existing behaviour.  All break codings have the same
+	 * format, so there is no need to check the slot type.
+	 */
+	if (break_num == 0 && !user_mode(regs)) {
+		struct ia64_psr *ipsr = ia64_psr(regs);
+		unsigned long *bundle = (unsigned long *)regs->cr_iip;
+		unsigned long slot;
+		switch (ipsr->ri) {
+		      case 0:  slot = (bundle[0] >>  5); break;
+		      case 1:  slot = (bundle[0] >> 46) | (bundle[1] << 18); break;
+		      default: slot = (bundle[1] >> 23); break;
+		}
+		break_num = ((slot >> 36 & 1) << 20) | (slot >> 6 & 0xfffff);
+	}
+
 	/* SIGILL, SIGFPE, SIGSEGV, and SIGBUS want these field initialized: */
 	siginfo.si_addr = (void __user *) (regs->cr_iip + ia64_psr(regs)->ri);
 	siginfo.si_imm = break_num;
@@ -202,13 +220,21 @@ disabled_fph_fault (struct pt_regs *regs)
 
 	/* first, grant user-level access to fph partition: */
 	psr->dfh = 0;
+
+	/*
+	 * Make sure that no other task gets in on this processor
+	 * while we're claiming the FPU
+	 */
+	preempt_disable();
 #ifndef CONFIG_SMP
 	{
 		struct task_struct *fpu_owner
 			= (struct task_struct *)ia64_get_kr(IA64_KR_FPU_OWNER);
 
-		if (ia64_is_local_fpu_owner(current))
+		if (ia64_is_local_fpu_owner(current)) {
+			preempt_enable_no_resched();
 			return;
+		}
 
 		if (fpu_owner)
 			ia64_flush_fph(fpu_owner);
@@ -226,6 +252,7 @@ disabled_fph_fault (struct pt_regs *regs)
 		 */
 		psr->mfh = 1;
 	}
+	preempt_enable_no_resched();
 }
 
 static inline int