3 files changed, 157 insertions, 237 deletions

diff --git a/kernel/arch/x86/platform/uv/uv_irq.c b/kernel/arch/x86/platform/uv/uv_irq.c
index 0ce673645..e1c24631a 100644
--- a/kernel/arch/x86/platform/uv/uv_irq.c
+++ b/kernel/arch/x86/platform/uv/uv_irq.c
@@ -13,22 +13,37 @@
 #include <linux/slab.h>
 #include <linux/irq.h>
 
+#include <asm/irqdomain.h>
 #include <asm/apic.h>
 #include <asm/uv/uv_irq.h>
 #include <asm/uv/uv_hub.h>
 
 /* MMR offset and pnode of hub sourcing interrupts for a given irq */
-struct uv_irq_2_mmr_pnode{
-	struct rb_node		list;
+struct uv_irq_2_mmr_pnode {
 	unsigned long		offset;
 	int			pnode;
-	int			irq;
 };
 
-static DEFINE_SPINLOCK(uv_irq_lock);
-static struct rb_root		uv_irq_root;
+static void uv_program_mmr(struct irq_cfg *cfg, struct uv_irq_2_mmr_pnode *info)
+{
+	unsigned long mmr_value;
+	struct uv_IO_APIC_route_entry *entry;
+
+	BUILD_BUG_ON(sizeof(struct uv_IO_APIC_route_entry) !=
+		     sizeof(unsigned long));
+
+	mmr_value = 0;
+	entry = (struct uv_IO_APIC_route_entry *)&mmr_value;
+	entry->vector		= cfg->vector;
+	entry->delivery_mode	= apic->irq_delivery_mode;
+	entry->dest_mode	= apic->irq_dest_mode;
+	entry->polarity		= 0;
+	entry->trigger		= 0;
+	entry->mask		= 0;
+	entry->dest		= cfg->dest_apicid;
 
-static int uv_set_irq_affinity(struct irq_data *, const struct cpumask *, bool);
+	uv_write_global_mmr64(info->pnode, info->offset, mmr_value);
+}
 
 static void uv_noop(struct irq_data *data) { }
 
@@ -37,6 +52,23 @@ static void uv_ack_apic(struct irq_data *data)
 	ack_APIC_irq();
 }
 
+static int
+uv_set_irq_affinity(struct irq_data *data, const struct cpumask *mask,
+		    bool force)
+{
+	struct irq_data *parent = data->parent_data;
+	struct irq_cfg *cfg = irqd_cfg(data);
+	int ret;
+
+	ret = parent->chip->irq_set_affinity(parent, mask, force);
+	if (ret >= 0) {
+		uv_program_mmr(cfg, data->chip_data);
+		send_cleanup_vector(cfg);
+	}
+
+	return ret;
+}
+
 static struct irq_chip uv_irq_chip = {
 	.name			= "UV-CORE",
 	.irq_mask		= uv_noop,
@@ -45,189 +77,99 @@ static struct irq_chip uv_irq_chip = {
 	.irq_set_affinity	= uv_set_irq_affinity,
 };
 
-/*
- * Add offset and pnode information of the hub sourcing interrupts to the
- * rb tree for a specific irq.
- */
-static int uv_set_irq_2_mmr_info(int irq, unsigned long offset, unsigned blade)
+static int uv_domain_alloc(struct irq_domain *domain, unsigned int virq,
+			   unsigned int nr_irqs, void *arg)
 {
-	struct rb_node **link = &uv_irq_root.rb_node;
-	struct rb_node *parent = NULL;
-	struct uv_irq_2_mmr_pnode *n;
-	struct uv_irq_2_mmr_pnode *e;
-	unsigned long irqflags;
-
-	n = kmalloc_node(sizeof(struct uv_irq_2_mmr_pnode), GFP_KERNEL,
-				uv_blade_to_memory_nid(blade));
-	if (!n)
+	struct uv_irq_2_mmr_pnode *chip_data;
+	struct irq_alloc_info *info = arg;
+	struct irq_data *irq_data = irq_domain_get_irq_data(domain, virq);
+	int ret;
+
+	if (nr_irqs > 1 || !info || info->type != X86_IRQ_ALLOC_TYPE_UV)
+		return -EINVAL;
+
+	chip_data = kmalloc_node(sizeof(*chip_data), GFP_KERNEL,
+				 irq_data_get_node(irq_data));
+	if (!chip_data)
 		return -ENOMEM;
 
-	n->irq = irq;
-	n->offset = offset;
-	n->pnode = uv_blade_to_pnode(blade);
-	spin_lock_irqsave(&uv_irq_lock, irqflags);
-	/* Find the right place in the rbtree: */
-	while (*link) {
-		parent = *link;
-		e = rb_entry(parent, struct uv_irq_2_mmr_pnode, list);
-
-		if (unlikely(irq == e->irq)) {
-			/* irq entry exists */
-			e->pnode = uv_blade_to_pnode(blade);
-			e->offset = offset;
-			spin_unlock_irqrestore(&uv_irq_lock, irqflags);
-			kfree(n);
-			return 0;
-		}
-
-		if (irq < e->irq)
-			link = &(*link)->rb_left;
+	ret = irq_domain_alloc_irqs_parent(domain, virq, nr_irqs, arg);
+	if (ret >= 0) {
+		if (info->uv_limit == UV_AFFINITY_CPU)
+			irq_set_status_flags(virq, IRQ_NO_BALANCING);
 		else
-			link = &(*link)->rb_right;
+			irq_set_status_flags(virq, IRQ_MOVE_PCNTXT);
+
+		chip_data->pnode = uv_blade_to_pnode(info->uv_blade);
+		chip_data->offset = info->uv_offset;
+		irq_domain_set_info(domain, virq, virq, &uv_irq_chip, chip_data,
+				    handle_percpu_irq, NULL, info->uv_name);
+	} else {
+		kfree(chip_data);
 	}
 
-	/* Insert the node into the rbtree. */
-	rb_link_node(&n->list, parent, link);
-	rb_insert_color(&n->list, &uv_irq_root);
-
-	spin_unlock_irqrestore(&uv_irq_lock, irqflags);
-	return 0;
+	return ret;
 }
 
-/* Retrieve offset and pnode information from the rb tree for a specific irq */
-int uv_irq_2_mmr_info(int irq, unsigned long *offset, int *pnode)
+static void uv_domain_free(struct irq_domain *domain, unsigned int virq,
+			   unsigned int nr_irqs)
 {
-	struct uv_irq_2_mmr_pnode *e;
-	struct rb_node *n;
-	unsigned long irqflags;
-
-	spin_lock_irqsave(&uv_irq_lock, irqflags);
-	n = uv_irq_root.rb_node;
-	while (n) {
-		e = rb_entry(n, struct uv_irq_2_mmr_pnode, list);
-
-		if (e->irq == irq) {
-			*offset = e->offset;
-			*pnode = e->pnode;
-			spin_unlock_irqrestore(&uv_irq_lock, irqflags);
-			return 0;
-		}
-
-		if (irq < e->irq)
-			n = n->rb_left;
-		else
-			n = n->rb_right;
-	}
-	spin_unlock_irqrestore(&uv_irq_lock, irqflags);
-	return -1;
+	struct irq_data *irq_data = irq_domain_get_irq_data(domain, virq);
+
+	BUG_ON(nr_irqs != 1);
+	kfree(irq_data->chip_data);
+	irq_clear_status_flags(virq, IRQ_MOVE_PCNTXT);
+	irq_clear_status_flags(virq, IRQ_NO_BALANCING);
+	irq_domain_free_irqs_top(domain, virq, nr_irqs);
 }
 
 /*
  * Re-target the irq to the specified CPU and enable the specified MMR located
  * on the specified blade to allow the sending of MSIs to the specified CPU.
  */
-static int
-arch_enable_uv_irq(char *irq_name, unsigned int irq, int cpu, int mmr_blade,
-		       unsigned long mmr_offset, int limit)
+static void uv_domain_activate(struct irq_domain *domain,
+			       struct irq_data *irq_data)
 {
-	const struct cpumask *eligible_cpu = cpumask_of(cpu);
-	struct irq_cfg *cfg = irq_cfg(irq);
-	unsigned long mmr_value;
-	struct uv_IO_APIC_route_entry *entry;
-	int mmr_pnode, err;
-	unsigned int dest;
-
-	BUILD_BUG_ON(sizeof(struct uv_IO_APIC_route_entry) !=
-			sizeof(unsigned long));
-
-	err = assign_irq_vector(irq, cfg, eligible_cpu);
-	if (err != 0)
-		return err;
-
-	err = apic->cpu_mask_to_apicid_and(eligible_cpu, eligible_cpu, &dest);
-	if (err != 0)
-		return err;
-
-	if (limit == UV_AFFINITY_CPU)
-		irq_set_status_flags(irq, IRQ_NO_BALANCING);
-	else
-		irq_set_status_flags(irq, IRQ_MOVE_PCNTXT);
-
-	irq_set_chip_and_handler_name(irq, &uv_irq_chip, handle_percpu_irq,
-				      irq_name);
-
-	mmr_value = 0;
-	entry = (struct uv_IO_APIC_route_entry *)&mmr_value;
-	entry->vector		= cfg->vector;
-	entry->delivery_mode	= apic->irq_delivery_mode;
-	entry->dest_mode	= apic->irq_dest_mode;
-	entry->polarity		= 0;
-	entry->trigger		= 0;
-	entry->mask		= 0;
-	entry->dest		= dest;
-
-	mmr_pnode = uv_blade_to_pnode(mmr_blade);
-	uv_write_global_mmr64(mmr_pnode, mmr_offset, mmr_value);
-
-	if (cfg->move_in_progress)
-		send_cleanup_vector(cfg);
-
-	return irq;
+	uv_program_mmr(irqd_cfg(irq_data), irq_data->chip_data);
 }
 
 /*
  * Disable the specified MMR located on the specified blade so that MSIs are
  * longer allowed to be sent.
  */
-static void arch_disable_uv_irq(int mmr_pnode, unsigned long mmr_offset)
+static void uv_domain_deactivate(struct irq_domain *domain,
+				 struct irq_data *irq_data)
 {
 	unsigned long mmr_value;
 	struct uv_IO_APIC_route_entry *entry;
 
-	BUILD_BUG_ON(sizeof(struct uv_IO_APIC_route_entry) !=
-			sizeof(unsigned long));
-
 	mmr_value = 0;
 	entry = (struct uv_IO_APIC_route_entry *)&mmr_value;
 	entry->mask = 1;
-
-	uv_write_global_mmr64(mmr_pnode, mmr_offset, mmr_value);
+	uv_program_mmr(irqd_cfg(irq_data), irq_data->chip_data);
 }
 
-static int
-uv_set_irq_affinity(struct irq_data *data, const struct cpumask *mask,
-		    bool force)
-{
-	struct irq_cfg *cfg = irqd_cfg(data);
-	unsigned int dest;
-	unsigned long mmr_value, mmr_offset;
-	struct uv_IO_APIC_route_entry *entry;
-	int mmr_pnode;
-
-	if (apic_set_affinity(data, mask, &dest))
-		return -1;
-
-	mmr_value = 0;
-	entry = (struct uv_IO_APIC_route_entry *)&mmr_value;
-
-	entry->vector		= cfg->vector;
-	entry->delivery_mode	= apic->irq_delivery_mode;
-	entry->dest_mode	= apic->irq_dest_mode;
-	entry->polarity		= 0;
-	entry->trigger		= 0;
-	entry->mask		= 0;
-	entry->dest		= dest;
-
-	/* Get previously stored MMR and pnode of hub sourcing interrupts */
-	if (uv_irq_2_mmr_info(data->irq, &mmr_offset, &mmr_pnode))
-		return -1;
-
-	uv_write_global_mmr64(mmr_pnode, mmr_offset, mmr_value);
+static const struct irq_domain_ops uv_domain_ops = {
+	.alloc		= uv_domain_alloc,
+	.free		= uv_domain_free,
+	.activate	= uv_domain_activate,
+	.deactivate	= uv_domain_deactivate,
+};
 
-	if (cfg->move_in_progress)
-		send_cleanup_vector(cfg);
+static struct irq_domain *uv_get_irq_domain(void)
+{
+	static struct irq_domain *uv_domain;
+	static DEFINE_MUTEX(uv_lock);
+
+	mutex_lock(&uv_lock);
+	if (uv_domain == NULL) {
+		uv_domain = irq_domain_add_tree(NULL, &uv_domain_ops, NULL);
+		if (uv_domain)
+			uv_domain->parent = x86_vector_domain;
+	}
+	mutex_unlock(&uv_lock);
 
-	return IRQ_SET_MASK_OK_NOCOPY;
+	return uv_domain;
 }
 
 /*
@@ -238,19 +180,21 @@ uv_set_irq_affinity(struct irq_data *data, const struct cpumask *mask,
 int uv_setup_irq(char *irq_name, int cpu, int mmr_blade,
 		 unsigned long mmr_offset, int limit)
 {
-	int ret, irq = irq_alloc_hwirq(uv_blade_to_memory_nid(mmr_blade));
+	struct irq_alloc_info info;
+	struct irq_domain *domain = uv_get_irq_domain();
 
-	if (!irq)
-		return -EBUSY;
+	if (!domain)
+		return -ENOMEM;
 
-	ret = arch_enable_uv_irq(irq_name, irq, cpu, mmr_blade, mmr_offset,
-		limit);
-	if (ret == irq)
-		uv_set_irq_2_mmr_info(irq, mmr_offset, mmr_blade);
-	else
-		irq_free_hwirq(irq);
+	init_irq_alloc_info(&info, cpumask_of(cpu));
+	info.type = X86_IRQ_ALLOC_TYPE_UV;
+	info.uv_limit = limit;
+	info.uv_blade = mmr_blade;
+	info.uv_offset = mmr_offset;
+	info.uv_name = irq_name;
 
-	return ret;
+	return irq_domain_alloc_irqs(domain, 1,
+				     uv_blade_to_memory_nid(mmr_blade), &info);
 }
 EXPORT_SYMBOL_GPL(uv_setup_irq);
 
@@ -263,26 +207,6 @@ EXPORT_SYMBOL_GPL(uv_setup_irq);
  */
 void uv_teardown_irq(unsigned int irq)
 {
-	struct uv_irq_2_mmr_pnode *e;
-	struct rb_node *n;
-	unsigned long irqflags;
-
-	spin_lock_irqsave(&uv_irq_lock, irqflags);
-	n = uv_irq_root.rb_node;
-	while (n) {
-		e = rb_entry(n, struct uv_irq_2_mmr_pnode, list);
-		if (e->irq == irq) {
-			arch_disable_uv_irq(e->pnode, e->offset);
-			rb_erase(n, &uv_irq_root);
-			kfree(e);
-			break;
-		}
-		if (irq < e->irq)
-			n = n->rb_left;
-		else
-			n = n->rb_right;
-	}
-	spin_unlock_irqrestore(&uv_irq_lock, irqflags);
-	irq_free_hwirq(irq);
+	irq_domain_free_irqs(irq, 1);
 }
 EXPORT_SYMBOL_GPL(uv_teardown_irq);
diff --git a/kernel/arch/x86/platform/uv/uv_nmi.c b/kernel/arch/x86/platform/uv/uv_nmi.c
index 7488cafab..327f21c3b 100644
--- a/kernel/arch/x86/platform/uv/uv_nmi.c
+++ b/kernel/arch/x86/platform/uv/uv_nmi.c
@@ -104,7 +104,7 @@ static int param_set_local64(const char *val, const struct kernel_param *kp)
 	return 0;
 }
 
-static struct kernel_param_ops param_ops_local64 = {
+static const struct kernel_param_ops param_ops_local64 = {
 	.get = param_get_local64,
 	.set = param_set_local64,
 };
@@ -376,38 +376,42 @@ static void uv_nmi_wait(int master)
 		atomic_read(&uv_nmi_cpus_in_nmi), num_online_cpus());
 }
 
+/* Dump Instruction Pointer header */
 static void uv_nmi_dump_cpu_ip_hdr(void)
 {
-	printk(KERN_DEFAULT
-		"\nUV: %4s %6s %-32s %s   (Note: PID 0 not listed)\n",
+	pr_info("\nUV: %4s %6s %-32s %s   (Note: PID 0 not listed)\n",
 		"CPU", "PID", "COMMAND", "IP");
 }
 
+/* Dump Instruction Pointer info */
 static void uv_nmi_dump_cpu_ip(int cpu, struct pt_regs *regs)
 {
-	printk(KERN_DEFAULT "UV: %4d %6d %-32.32s ",
-		cpu, current->pid, current->comm);
-
+	pr_info("UV: %4d %6d %-32.32s ", cpu, current->pid, current->comm);
 	printk_address(regs->ip);
 }
 
-/* Dump this cpu's state */
+/*
+ * Dump this CPU's state.  If action was set to "kdump" and the crash_kexec
+ * failed, then we provide "dump" as an alternate action.  Action "dump" now
+ * also includes the show "ips" (instruction pointers) action whereas the
+ * action "ips" only displays instruction pointers for the non-idle CPU's.
+ * This is an abbreviated form of the "ps" command.
+ */
 static void uv_nmi_dump_state_cpu(int cpu, struct pt_regs *regs)
 {
 	const char *dots = " ................................. ";
 
-	if (uv_nmi_action_is("ips")) {
-		if (cpu == 0)
-			uv_nmi_dump_cpu_ip_hdr();
+	if (cpu == 0)
+		uv_nmi_dump_cpu_ip_hdr();
 
-		if (current->pid != 0)
-			uv_nmi_dump_cpu_ip(cpu, regs);
+	if (current->pid != 0 || !uv_nmi_action_is("ips"))
+		uv_nmi_dump_cpu_ip(cpu, regs);
 
-	} else if (uv_nmi_action_is("dump")) {
-		printk(KERN_DEFAULT
-			"UV:%sNMI process trace for CPU %d\n", dots, cpu);
+	if (uv_nmi_action_is("dump")) {
+		pr_info("UV:%sNMI process trace for CPU %d\n", dots, cpu);
 		show_regs(regs);
 	}
+
 	this_cpu_write(uv_cpu_nmi.state, UV_NMI_STATE_DUMP_DONE);
 }
 
@@ -469,8 +473,7 @@ static void uv_nmi_dump_state(int cpu, struct pt_regs *regs, int master)
 				uv_nmi_trigger_dump(tcpu);
 		}
 		if (ignored)
-			printk(KERN_DEFAULT "UV: %d CPUs ignored NMI\n",
-				ignored);
+			pr_alert("UV: %d CPUs ignored NMI\n", ignored);
 
 		console_loglevel = saved_console_loglevel;
 		pr_alert("UV: process trace complete\n");
@@ -492,8 +495,9 @@ static void uv_nmi_touch_watchdogs(void)
 	touch_nmi_watchdog();
 }
 
-#if defined(CONFIG_KEXEC)
 static atomic_t uv_nmi_kexec_failed;
+
+#if defined(CONFIG_KEXEC_CORE)
 static void uv_nmi_kdump(int cpu, int master, struct pt_regs *regs)
 {
 	/* Call crash to dump system state */
@@ -502,10 +506,9 @@ static void uv_nmi_kdump(int cpu, int master, struct pt_regs *regs)
 		crash_kexec(regs);
 
 		pr_emerg("UV: crash_kexec unexpectedly returned, ");
+		atomic_set(&uv_nmi_kexec_failed, 1);
 		if (!kexec_crash_image) {
 			pr_cont("crash kernel not loaded\n");
-			atomic_set(&uv_nmi_kexec_failed, 1);
-			uv_nmi_sync_exit(1);
 			return;
 		}
 		pr_cont("kexec busy, stalling cpus while waiting\n");
@@ -514,18 +517,16 @@ static void uv_nmi_kdump(int cpu, int master, struct pt_regs *regs)
 	/* If crash exec fails the slaves should return, otherwise stall */
 	while (atomic_read(&uv_nmi_kexec_failed) == 0)
 		mdelay(10);
-
-	/* Crash kernel most likely not loaded, return in an orderly fashion */
-	uv_nmi_sync_exit(0);
 }
 
-#else /* !CONFIG_KEXEC */
+#else /* !CONFIG_KEXEC_CORE */
 static inline void uv_nmi_kdump(int cpu, int master, struct pt_regs *regs)
 {
 	if (master)
 		pr_err("UV: NMI kdump: KEXEC not supported in this kernel\n");
+	atomic_set(&uv_nmi_kexec_failed, 1);
 }
-#endif /* !CONFIG_KEXEC */
+#endif /* !CONFIG_KEXEC_CORE */
 
 #ifdef CONFIG_KGDB
 #ifdef CONFIG_KGDB_KDB
@@ -613,9 +614,14 @@ int uv_handle_nmi(unsigned int reason, struct pt_regs *regs)
 	master = (atomic_read(&uv_nmi_cpu) == cpu);
 
 	/* If NMI action is "kdump", then attempt to do it */
-	if (uv_nmi_action_is("kdump"))
+	if (uv_nmi_action_is("kdump")) {
 		uv_nmi_kdump(cpu, master, regs);
 
+		/* Unexpected return, revert action to "dump" */
+		if (master)
+			strncpy(uv_nmi_action, "dump", strlen(uv_nmi_action));
+	}
+
 	/* Pause as all cpus enter the NMI handler */
 	uv_nmi_wait(master);
 
@@ -640,6 +646,7 @@ int uv_handle_nmi(unsigned int reason, struct pt_regs *regs)
 		atomic_set(&uv_nmi_cpus_in_nmi, -1);
 		atomic_set(&uv_nmi_cpu, -1);
 		atomic_set(&uv_in_nmi, 0);
+		atomic_set(&uv_nmi_kexec_failed, 0);
 	}
 
 	uv_nmi_touch_watchdogs();
diff --git a/kernel/arch/x86/platform/uv/uv_time.c b/kernel/arch/x86/platform/uv/uv_time.c
index a718fe0d2..5e0b12262 100644
--- a/kernel/arch/x86/platform/uv/uv_time.c
+++ b/kernel/arch/x86/platform/uv/uv_time.c
@@ -32,8 +32,7 @@
 
 static cycle_t uv_read_rtc(struct clocksource *cs);
 static int uv_rtc_next_event(unsigned long, struct clock_event_device *);
-static void uv_rtc_timer_setup(enum clock_event_mode,
-				struct clock_event_device *);
+static int uv_rtc_shutdown(struct clock_event_device *evt);
 
 static struct clocksource clocksource_uv = {
 	.name		= RTC_NAME,
@@ -44,14 +43,14 @@ static struct clocksource clocksource_uv = {
 };
 
 static struct clock_event_device clock_event_device_uv = {
-	.name		= RTC_NAME,
-	.features	= CLOCK_EVT_FEAT_ONESHOT,
-	.shift		= 20,
-	.rating		= 400,
-	.irq		= -1,
-	.set_next_event	= uv_rtc_next_event,
-	.set_mode	= uv_rtc_timer_setup,
-	.event_handler	= NULL,
+	.name			= RTC_NAME,
+	.features		= CLOCK_EVT_FEAT_ONESHOT,
+	.shift			= 20,
+	.rating			= 400,
+	.irq			= -1,
+	.set_next_event		= uv_rtc_next_event,
+	.set_state_shutdown	= uv_rtc_shutdown,
+	.event_handler		= NULL,
 };
 
 static DEFINE_PER_CPU(struct clock_event_device, cpu_ced);
@@ -326,24 +325,14 @@ static int uv_rtc_next_event(unsigned long delta,
 }
 
 /*
- * Setup the RTC timer in oneshot mode
+ * Shutdown the RTC timer
  */
-static void uv_rtc_timer_setup(enum clock_event_mode mode,
-			       struct clock_event_device *evt)
+static int uv_rtc_shutdown(struct clock_event_device *evt)
 {
 	int ced_cpu = cpumask_first(evt->cpumask);
 
-	switch (mode) {
-	case CLOCK_EVT_MODE_PERIODIC:
-	case CLOCK_EVT_MODE_ONESHOT:
-	case CLOCK_EVT_MODE_RESUME:
-		/* Nothing to do here yet */
-		break;
-	case CLOCK_EVT_MODE_UNUSED:
-	case CLOCK_EVT_MODE_SHUTDOWN:
-		uv_rtc_unset_timer(ced_cpu, 1);
-		break;
-	}
+	uv_rtc_unset_timer(ced_cpu, 1);
+	return 0;
 }
 
 static void uv_rtc_interrupt(void)