Merge branches 'irq/sparseirq', 'x86/quirks' and 'x86/reboot' into cpus4096

We merge the irq/sparseirq, x86/quirks and x86/reboot trees into the
cpus4096 tree because the io-apic changes in the sparseirq change
conflict with the cpumask changes in the cpumask tree, and we
want to resolve those.

arch/x86/Kconfig

@@ -241,6 +241,16 @@ config X86_HAS_BOOT_CPU_ID
 	def_bool y
 	depends on X86_VOYAGER
 
+config SPARSE_IRQ
+	bool "Support sparse irq numbering"
+	depends on (PCI_MSI || HT_IRQ) && SMP
+	default y
+	help
+	  This enables support for sparse irq, esp for msi/msi-x. You may need
+	  if you have lots of cards supports msi-x installed.
+
+	  If you don't know what to do here, say Y.
+
 config X86_FIND_SMP_CONFIG
 	def_bool y
 	depends on X86_MPPARSE || X86_VOYAGER
@@ -468,10 +478,6 @@ config X86_CYCLONE_TIMER
 	def_bool y
 	depends on X86_GENERICARCH
 
-config ES7000_CLUSTERED_APIC
-	def_bool y
-	depends on SMP && X86_ES7000 && MPENTIUMIII
-
 source "arch/x86/Kconfig.cpu"
 
 config HPET_TIMER
@@ -1635,13 +1641,6 @@ config APM_ALLOW_INTS
 	  many of the newer IBM Thinkpads.  If you experience hangs when you
 	  suspend, try setting this to Y.  Otherwise, say N.
 
-config APM_REAL_MODE_POWER_OFF
-	bool "Use real mode APM BIOS call to power off"
-	help
-	  Use real mode APM BIOS calls to switch off the computer. This is
-	  a work-around for a number of buggy BIOSes. Switch this option on if
-	  your computer crashes instead of powering off properly.
-
 endif # APM
 
 source "arch/x86/kernel/cpu/cpufreq/Kconfig"

arch/x86/include/asm/apic.h

@@ -193,6 +193,7 @@ extern u8 setup_APIC_eilvt_ibs(u8 vector, u8 msg_type, u8 mask);
 static inline void lapic_shutdown(void) { }
 #define local_apic_timer_c2_ok		1
 static inline void init_apic_mappings(void) { }
+static inline void disable_local_APIC(void) { }
 
 #endif /* !CONFIG_X86_LOCAL_APIC */
 

arch/x86/include/asm/bigsmp/apic.h

@@ -24,8 +24,6 @@ static inline cpumask_t target_cpus(void)
 #define INT_DELIVERY_MODE	(dest_Fixed)
 #define INT_DEST_MODE		(0)    /* phys delivery to target proc */
 #define NO_BALANCE_IRQ		(0)
-#define WAKE_SECONDARY_VIA_INIT
-
 
 static inline unsigned long check_apicid_used(physid_mask_t bitmap, int apicid)
 {

arch/x86/include/asm/emergency-restart.h

@@ -8,7 +8,9 @@ enum reboot_type {
 	BOOT_BIOS = 'b',
 #endif
 	BOOT_ACPI = 'a',
-	BOOT_EFI = 'e'
+	BOOT_EFI = 'e',
+	BOOT_CF9 = 'p',
+	BOOT_CF9_COND = 'q',
 };
 
 extern enum reboot_type reboot_type;

arch/x86/include/asm/es7000/apic.h

@@ -9,31 +9,27 @@ static inline int apic_id_registered(void)
 	        return (1);
 }
 
-static inline cpumask_t target_cpus(void)
+static inline cpumask_t target_cpus_cluster(void)
 {
-#if defined CONFIG_ES7000_CLUSTERED_APIC
 	return CPU_MASK_ALL;
-#else
+}
+
+static inline cpumask_t target_cpus(void)
+{
 	return cpumask_of_cpu(smp_processor_id());
-#endif
 }
 
-#if defined CONFIG_ES7000_CLUSTERED_APIC
-#define APIC_DFR_VALUE		(APIC_DFR_CLUSTER)
-#define INT_DELIVERY_MODE	(dest_LowestPrio)
-#define INT_DEST_MODE		(1)    /* logical delivery broadcast to all procs */
-#define NO_BALANCE_IRQ		(1)
-#undef  WAKE_SECONDARY_VIA_INIT
-#define WAKE_SECONDARY_VIA_MIP
-#else
+#define APIC_DFR_VALUE_CLUSTER		(APIC_DFR_CLUSTER)
+#define INT_DELIVERY_MODE_CLUSTER	(dest_LowestPrio)
+#define INT_DEST_MODE_CLUSTER		(1) /* logical delivery broadcast to all procs */
+#define NO_BALANCE_IRQ_CLUSTER		(1)
+
 #define APIC_DFR_VALUE		(APIC_DFR_FLAT)
 #define INT_DELIVERY_MODE	(dest_Fixed)
 #define INT_DEST_MODE		(0)    /* phys delivery to target procs */
 #define NO_BALANCE_IRQ		(0)
 #undef  APIC_DEST_LOGICAL
 #define APIC_DEST_LOGICAL	0x0
-#define WAKE_SECONDARY_VIA_INIT
-#endif
 
 static inline unsigned long check_apicid_used(physid_mask_t bitmap, int apicid)
 {
@@ -60,6 +56,16 @@ static inline unsigned long calculate_ldr(int cpu)
  * an APIC.  See e.g. "AP-388 82489DX User's Manual" (Intel
  * document number 292116).  So here it goes...
  */
+static inline void init_apic_ldr_cluster(void)
+{
+	unsigned long val;
+	int cpu = smp_processor_id();
+
+	apic_write(APIC_DFR, APIC_DFR_VALUE_CLUSTER);
+	val = calculate_ldr(cpu);
+	apic_write(APIC_LDR, val);
+}
+
 static inline void init_apic_ldr(void)
 {
 	unsigned long val;
@@ -70,10 +76,6 @@ static inline void init_apic_ldr(void)
 	apic_write(APIC_LDR, val);
 }
 
-#ifndef CONFIG_X86_GENERICARCH
-extern void enable_apic_mode(void);
-#endif
-
 extern int apic_version [MAX_APICS];
 static inline void setup_apic_routing(void)
 {
@@ -144,7 +146,7 @@ static inline int check_phys_apicid_present(int cpu_physical_apicid)
 	return (1);
 }
 
-static inline unsigned int cpu_mask_to_apicid(cpumask_t cpumask)
+static inline unsigned int cpu_mask_to_apicid_cluster(cpumask_t cpumask)
 {
 	int num_bits_set;
 	int cpus_found = 0;
@@ -154,11 +156,7 @@ static inline unsigned int cpu_mask_to_apicid(cpumask_t cpumask)
 	num_bits_set = cpus_weight(cpumask);
 	/* Return id to all */
 	if (num_bits_set == NR_CPUS)
-#if defined CONFIG_ES7000_CLUSTERED_APIC
 		return 0xFF;
-#else
-		return cpu_to_logical_apicid(0);
-#endif
 	/*
 	 * The cpus in the mask must all be on the apic cluster.  If are not
 	 * on the same apicid cluster return default value of TARGET_CPUS.
@@ -171,11 +169,40 @@ static inline unsigned int cpu_mask_to_apicid(cpumask_t cpumask)
 			if (apicid_cluster(apicid) !=
 					apicid_cluster(new_apicid)){
 				printk ("%s: Not a valid mask!\n", __func__);
-#if defined CONFIG_ES7000_CLUSTERED_APIC
 				return 0xFF;
-#else
+			}
+			apicid = new_apicid;
+			cpus_found++;
+		}
+		cpu++;
+	}
+	return apicid;
+}
+
+static inline unsigned int cpu_mask_to_apicid(cpumask_t cpumask)
+{
+	int num_bits_set;
+	int cpus_found = 0;
+	int cpu;
+	int apicid;
+
+	num_bits_set = cpus_weight(cpumask);
+	/* Return id to all */
+	if (num_bits_set == NR_CPUS)
+		return cpu_to_logical_apicid(0);
+	/*
+	 * The cpus in the mask must all be on the apic cluster.  If are not
+	 * on the same apicid cluster return default value of TARGET_CPUS.
+	 */
+	cpu = first_cpu(cpumask);
+	apicid = cpu_to_logical_apicid(cpu);
+	while (cpus_found < num_bits_set) {
+		if (cpu_isset(cpu, cpumask)) {
+			int new_apicid = cpu_to_logical_apicid(cpu);
+			if (apicid_cluster(apicid) !=
+					apicid_cluster(new_apicid)){
+				printk ("%s: Not a valid mask!\n", __func__);
 				return cpu_to_logical_apicid(0);
-#endif
 			}
 			apicid = new_apicid;
 			cpus_found++;

arch/x86/include/asm/es7000/wakecpu.h

 #ifndef __ASM_ES7000_WAKECPU_H
 #define __ASM_ES7000_WAKECPU_H
 
-/*
- * This file copes with machines that wakeup secondary CPUs by the
- * INIT, INIT, STARTUP sequence.
- */
-
-#ifdef CONFIG_ES7000_CLUSTERED_APIC
-#define WAKE_SECONDARY_VIA_MIP
-#else
-#define WAKE_SECONDARY_VIA_INIT
-#endif
-
-#ifdef WAKE_SECONDARY_VIA_MIP
-extern int es7000_start_cpu(int cpu, unsigned long eip);
-static inline int
-wakeup_secondary_cpu(int phys_apicid, unsigned long start_eip)
-{
-	int boot_error = 0;
-	boot_error = es7000_start_cpu(phys_apicid, start_eip);
-	return boot_error;
-}
-#endif
-
-#define TRAMPOLINE_LOW phys_to_virt(0x467)
-#define TRAMPOLINE_HIGH phys_to_virt(0x469)
-
-#define boot_cpu_apicid boot_cpu_physical_apicid
+#define TRAMPOLINE_PHYS_LOW	0x467
+#define TRAMPOLINE_PHYS_HIGH	0x469
 
 static inline void wait_for_init_deassert(atomic_t *deassert)
 {
-#ifdef WAKE_SECONDARY_VIA_INIT
+#ifndef CONFIG_ES7000_CLUSTERED_APIC
 	while (!atomic_read(deassert))
 		cpu_relax();
 #endif
@@ -50,9 +26,12 @@ static inline void restore_NMI_vector(unsigned short *high, unsigned short *low)
 {
 }
 
-#define inquire_remote_apic(apicid) do {		\
-		if (apic_verbosity >= APIC_DEBUG)	\
-			__inquire_remote_apic(apicid);	\
-	} while (0)
+extern void __inquire_remote_apic(int apicid);
+
+static inline void inquire_remote_apic(int apicid)
+{
+	if (apic_verbosity >= APIC_DEBUG)
+		__inquire_remote_apic(apicid);
+}
 
 #endif /* __ASM_MACH_WAKECPU_H */

arch/x86/include/asm/genapic_32.h

@@ -2,6 +2,7 @@
 #define _ASM_X86_GENAPIC_32_H
 
 #include <asm/mpspec.h>
+#include <asm/atomic.h>
 
 /*
  * Generic APIC driver interface.
@@ -65,6 +66,14 @@ struct genapic {
 	void (*send_IPI_allbutself)(int vector);
 	void (*send_IPI_all)(int vector);
 #endif
+	int (*wakeup_cpu)(int apicid, unsigned long start_eip);
+	int trampoline_phys_low;
+	int trampoline_phys_high;
+	void (*wait_for_init_deassert)(atomic_t *deassert);
+	void (*smp_callin_clear_local_apic)(void);
+	void (*store_NMI_vector)(unsigned short *high, unsigned short *low);
+	void (*restore_NMI_vector)(unsigned short *high, unsigned short *low);
+	void (*inquire_remote_apic)(int apicid);
 };
 
 #define APICFUNC(x) .x = x,
@@ -105,16 +114,24 @@ struct genapic {
 	APICFUNC(get_apic_id)				\
 	.apic_id_mask = APIC_ID_MASK,			\
 	APICFUNC(cpu_mask_to_apicid)			\
-	APICFUNC(vector_allocation_domain)			\
+	APICFUNC(vector_allocation_domain)		\
 	APICFUNC(acpi_madt_oem_check)			\
 	IPIFUNC(send_IPI_mask)				\
 	IPIFUNC(send_IPI_allbutself)			\
 	IPIFUNC(send_IPI_all)				\
 	APICFUNC(enable_apic_mode)			\
 	APICFUNC(phys_pkg_id)				\
+	.trampoline_phys_low = TRAMPOLINE_PHYS_LOW,		\
+	.trampoline_phys_high = TRAMPOLINE_PHYS_HIGH,		\
+	APICFUNC(wait_for_init_deassert)		\
+	APICFUNC(smp_callin_clear_local_apic)		\
+	APICFUNC(store_NMI_vector)			\
+	APICFUNC(restore_NMI_vector)			\
+	APICFUNC(inquire_remote_apic)			\
 }
 
 extern struct genapic *genapic;
+extern void es7000_update_genapic_to_cluster(void);
 
 enum uv_system_type {UV_NONE, UV_LEGACY_APIC, UV_X2APIC, UV_NON_UNIQUE_APIC};
 #define get_uv_system_type()		UV_NONE

arch/x86/include/asm/genapic_64.h

@@ -32,6 +32,8 @@ struct genapic {
 	unsigned int (*get_apic_id)(unsigned long x);
 	unsigned long (*set_apic_id)(unsigned int id);
 	unsigned long apic_id_mask;
+	/* wakeup_secondary_cpu */
+	int (*wakeup_cpu)(int apicid, unsigned long start_eip);
 };
 
 extern struct genapic *genapic;

arch/x86/include/asm/io_apic.h

@@ -188,17 +188,14 @@ extern void restore_IO_APIC_setup(void);
 extern void reinit_intr_remapped_IO_APIC(int);
 #endif
 
-extern int probe_nr_irqs(void);
+extern void probe_nr_irqs_gsi(void);
 
 #else  /* !CONFIG_X86_IO_APIC */
 #define io_apic_assign_pci_irqs 0
 static const int timer_through_8259 = 0;
-static inline void ioapic_init_mappings(void) { }
+static inline void ioapic_init_mappings(void)	{ }
 
-static inline int probe_nr_irqs(void)
-{
-	return NR_IRQS;
-}
+static inline void probe_nr_irqs_gsi(void)	{ }
 #endif
 
 #endif /* _ASM_X86_IO_APIC_H */

arch/x86/include/asm/irq_vectors.h

@@ -101,12 +101,23 @@
 #define LAST_VM86_IRQ		15
 #define invalid_vm86_irq(irq)	((irq) < 3 || (irq) > 15)
 
+#define NR_IRQS_LEGACY		16
+
 #if defined(CONFIG_X86_IO_APIC) && !defined(CONFIG_X86_VOYAGER)
+
+#ifndef CONFIG_SPARSE_IRQ
 # if NR_CPUS < MAX_IO_APICS
 #  define NR_IRQS (NR_VECTORS + (32 * NR_CPUS))
 # else
 #  define NR_IRQS (NR_VECTORS + (32 * MAX_IO_APICS))
 # endif
+#else
+# if (8 * NR_CPUS) > (32 * MAX_IO_APICS)
+#  define NR_IRQS (NR_VECTORS + (8 * NR_CPUS))
+# else
+#  define NR_IRQS (NR_VECTORS + (32 * MAX_IO_APICS))
+# endif
+#endif
 
 #elif defined(CONFIG_X86_VOYAGER)
 

arch/x86/include/asm/mach-default/mach_apic.h

@@ -32,11 +32,13 @@ static inline cpumask_t target_cpus(void)
 #define vector_allocation_domain    (genapic->vector_allocation_domain)
 #define read_apic_id()  (GET_APIC_ID(apic_read(APIC_ID)))
 #define send_IPI_self (genapic->send_IPI_self)
+#define wakeup_secondary_cpu (genapic->wakeup_cpu)
 extern void setup_apic_routing(void);
 #else
 #define INT_DELIVERY_MODE dest_LowestPrio
 #define INT_DEST_MODE 1     /* logical delivery broadcast to all procs */
 #define TARGET_CPUS (target_cpus())
+#define wakeup_secondary_cpu wakeup_secondary_cpu_via_init
 /*
  * Set up the logical destination ID.
  *

arch/x86/include/asm/mach-default/mach_wakecpu.h

 #ifndef _ASM_X86_MACH_DEFAULT_MACH_WAKECPU_H
 #define _ASM_X86_MACH_DEFAULT_MACH_WAKECPU_H
 
-/* 
- * This file copes with machines that wakeup secondary CPUs by the
- * INIT, INIT, STARTUP sequence.
- */
-
-#define WAKE_SECONDARY_VIA_INIT
-
-#define TRAMPOLINE_LOW phys_to_virt(0x467)
-#define TRAMPOLINE_HIGH phys_to_virt(0x469)
-
-#define boot_cpu_apicid boot_cpu_physical_apicid
+#define TRAMPOLINE_PHYS_LOW (0x467)
+#define TRAMPOLINE_PHYS_HIGH (0x469)
 
 static inline void wait_for_init_deassert(atomic_t *deassert)
 {
@@ -33,9 +24,12 @@ static inline void restore_NMI_vector(unsigned short *high, unsigned short *low)
 {
 }
 
-#define inquire_remote_apic(apicid) do {		\
-		if (apic_verbosity >= APIC_DEBUG)	\
-			__inquire_remote_apic(apicid);	\
-	} while (0)
+extern void __inquire_remote_apic(int apicid);
+
+static inline void inquire_remote_apic(int apicid)
+{
+	if (apic_verbosity >= APIC_DEBUG)
+		__inquire_remote_apic(apicid);
+}
 
 #endif /* _ASM_X86_MACH_DEFAULT_MACH_WAKECPU_H */

arch/x86/include/asm/mach-default/smpboot_hooks.h

@@ -13,9 +13,11 @@ static inline void smpboot_setup_warm_reset_vector(unsigned long start_eip)
 	CMOS_WRITE(0xa, 0xf);
 	local_flush_tlb();
 	pr_debug("1.\n");
-	*((volatile unsigned short *) TRAMPOLINE_HIGH) = start_eip >> 4;
+	*((volatile unsigned short *)phys_to_virt(TRAMPOLINE_PHYS_HIGH)) =
+								 start_eip >> 4;
 	pr_debug("2.\n");
-	*((volatile unsigned short *) TRAMPOLINE_LOW) = start_eip & 0xf;
+	*((volatile unsigned short *)phys_to_virt(TRAMPOLINE_PHYS_LOW)) =
+							 start_eip & 0xf;
 	pr_debug("3.\n");
 }
 
@@ -32,7 +34,7 @@ static inline void smpboot_restore_warm_reset_vector(void)
 	 */
 	CMOS_WRITE(0, 0xf);
 
-	*((volatile long *) phys_to_virt(0x467)) = 0;
+	*((volatile long *)phys_to_virt(TRAMPOLINE_PHYS_LOW)) = 0;
 }
 
 static inline void __init smpboot_setup_io_apic(void)

arch/x86/include/asm/mach-generic/mach_apic.h

@@ -27,6 +27,7 @@
 #define vector_allocation_domain (genapic->vector_allocation_domain)
 #define enable_apic_mode (genapic->enable_apic_mode)
 #define phys_pkg_id (genapic->phys_pkg_id)
+#define wakeup_secondary_cpu (genapic->wakeup_cpu)
 
 extern void generic_bigsmp_probe(void);
 

arch/x86/include/asm/mach-generic/mach_wakecpu.h

+#ifndef _ASM_X86_MACH_GENERIC_MACH_WAKECPU_H
+#define _ASM_X86_MACH_GENERIC_MACH_WAKECPU_H
+
+#define TRAMPOLINE_PHYS_LOW (genapic->trampoline_phys_low)
+#define TRAMPOLINE_PHYS_HIGH (genapic->trampoline_phys_high)
+#define wait_for_init_deassert (genapic->wait_for_init_deassert)
+#define smp_callin_clear_local_apic (genapic->smp_callin_clear_local_apic)
+#define store_NMI_vector (genapic->store_NMI_vector)
+#define restore_NMI_vector (genapic->restore_NMI_vector)
+#define inquire_remote_apic (genapic->inquire_remote_apic)
+
+#endif /* _ASM_X86_MACH_GENERIC_MACH_APIC_H */

arch/x86/include/asm/numaq/wakecpu.h

@@ -3,12 +3,8 @@
 
 /* This file copes with machines that wakeup secondary CPUs by NMIs */
 
-#define WAKE_SECONDARY_VIA_NMI
-
-#define TRAMPOLINE_LOW phys_to_virt(0x8)
-#define TRAMPOLINE_HIGH phys_to_virt(0xa)
-
-#define boot_cpu_apicid boot_cpu_logical_apicid
+#define TRAMPOLINE_PHYS_LOW (0x8)
+#define TRAMPOLINE_PHYS_HIGH (0xa)
 
 /* We don't do anything here because we use NMI's to boot instead */
 static inline void wait_for_init_deassert(atomic_t *deassert)
@@ -27,17 +23,23 @@ static inline void smp_callin_clear_local_apic(void)
 static inline void store_NMI_vector(unsigned short *high, unsigned short *low)
 {
 	printk("Storing NMI vector\n");
-	*high = *((volatile unsigned short *) TRAMPOLINE_HIGH);
-	*low = *((volatile unsigned short *) TRAMPOLINE_LOW);
+	*high =
+	  *((volatile unsigned short *)phys_to_virt(TRAMPOLINE_PHYS_HIGH));
+	*low =
+	  *((volatile unsigned short *)phys_to_virt(TRAMPOLINE_PHYS_LOW));
 }
 
 static inline void restore_NMI_vector(unsigned short *high, unsigned short *low)
 {
 	printk("Restoring NMI vector\n");
-	*((volatile unsigned short *) TRAMPOLINE_HIGH) = *high;
-	*((volatile unsigned short *) TRAMPOLINE_LOW) = *low;
+	*((volatile unsigned short *)phys_to_virt(TRAMPOLINE_PHYS_HIGH)) =
+								 *high;
+	*((volatile unsigned short *)phys_to_virt(TRAMPOLINE_PHYS_LOW)) =
+								 *low;
 }
 
-#define inquire_remote_apic(apicid) {}
+static inline void inquire_remote_apic(int apicid)
+{
+}
 
 #endif /* __ASM_NUMAQ_WAKECPU_H */

arch/x86/include/asm/setup.h

@@ -16,6 +16,8 @@ static inline void visws_early_detect(void) { }
 static inline int is_visws_box(void) { return 0; }
 #endif
 
+extern int wakeup_secondary_cpu_via_nmi(int apicid, unsigned long start_eip);
+extern int wakeup_secondary_cpu_via_init(int apicid, unsigned long start_eip);
 /*
  * Any setup quirks to be performed?
  */
@@ -39,6 +41,7 @@ struct x86_quirks {
 	void (*smp_read_mpc_oem)(struct mp_config_oemtable *oemtable,
                                     unsigned short oemsize);
 	int (*setup_ioapic_ids)(void);
+	int (*update_genapic)(void);
 };
 
 extern struct x86_quirks *x86_quirks;

arch/x86/include/asm/system.h

@@ -314,6 +314,8 @@ extern void free_init_pages(char *what, unsigned long begin, unsigned long end);
 
 void default_idle(void);
 
+void stop_this_cpu(void *dummy);
+
 /*
  * Force strict CPU ordering.
  * And yes, this is required on UP too when we're talking

arch/x86/kernel/acpi/boot.c

@@ -1360,6 +1360,17 @@ static void __init acpi_process_madt(void)
 			disable_acpi();
 		}
 	}
+
+	/*
+	 * ACPI supports both logical (e.g. Hyper-Threading) and physical
+	 * processors, where MPS only supports physical.
+	 */
+	if (acpi_lapic && acpi_ioapic)
+		printk(KERN_INFO "Using ACPI (MADT) for SMP configuration "
+		       "information\n");
+	else if (acpi_lapic)
+		printk(KERN_INFO "Using ACPI for processor (LAPIC) "
+		       "configuration information\n");
 #endif
 	return;
 }

arch/x86/kernel/apm_32.c

@@ -391,11 +391,7 @@ static int power_off;
 #else
 static int power_off = 1;
 #endif
-#ifdef CONFIG_APM_REAL_MODE_POWER_OFF
-static int realmode_power_off = 1;
-#else
 static int realmode_power_off;
-#endif
 #ifdef CONFIG_APM_ALLOW_INTS
 static int allow_ints = 1;
 #else

arch/x86/kernel/es7000_32.c

@@ -38,8 +38,11 @@
 #include <asm/io.h>
 #include <asm/nmi.h>
 #include <asm/smp.h>
+#include <asm/atomic.h>
 #include <asm/apicdef.h>
 #include <mach_mpparse.h>
+#include <asm/genapic.h>
+#include <asm/setup.h>
 
 /*
  * ES7000 chipsets
@@ -161,6 +164,43 @@ es7000_rename_gsi(int ioapic, int gsi)
 	return gsi;
 }
 
+static int wakeup_secondary_cpu_via_mip(int cpu, unsigned long eip)
+{
+	unsigned long vect = 0, psaival = 0;
+
+	if (psai == NULL)
+		return -1;
+
+	vect = ((unsigned long)__pa(eip)/0x1000) << 16;
+	psaival = (0x1000000 | vect | cpu);
+
+	while (*psai & 0x1000000)
+		;
+
+	*psai = psaival;
+
+	return 0;
+}
+
+static void noop_wait_for_deassert(atomic_t *deassert_not_used)
+{
+}
+
+static int __init es7000_update_genapic(void)
+{
+	genapic->wakeup_cpu = wakeup_secondary_cpu_via_mip;
+
+	/* MPENTIUMIII */
+	if (boot_cpu_data.x86 == 6 &&
+	    (boot_cpu_data.x86_model >= 7 || boot_cpu_data.x86_model <= 11)) {
+		es7000_update_genapic_to_cluster();
+		genapic->wait_for_init_deassert = noop_wait_for_deassert;
+		genapic->wakeup_cpu = wakeup_secondary_cpu_via_mip;
+	}
+
+	return 0;
+}
+
 void __init
 setup_unisys(void)
 {
@@ -176,6 +216,8 @@ setup_unisys(void)
 	else
 		es7000_plat = ES7000_CLASSIC;
 	ioapic_renumber_irq = es7000_rename_gsi;
+
+	x86_quirks->update_genapic = es7000_update_genapic;
 }
 
 /*
@@ -317,26 +359,6 @@ es7000_mip_write(struct mip_reg *mip_reg)
 	return status;
 }
 
-int
-es7000_start_cpu(int cpu, unsigned long eip)
-{
-	unsigned long vect = 0, psaival = 0;
-
-	if (psai == NULL)
-		return -1;
-
-	vect = ((unsigned long)__pa(eip)/0x1000) << 16;
-	psaival = (0x1000000 | vect | cpu);
-
-	while (*psai & 0x1000000)
-                ;
-
-	*psai = psaival;
-
-	return 0;
-
-}
-
 void __init
 es7000_sw_apic(void)
 {

arch/x86/kernel/genapic_64.c

@@ -21,6 +21,7 @@
 #include <asm/smp.h>
 #include <asm/ipi.h>
 #include <asm/genapic.h>
+#include <asm/setup.h>
 
 extern struct genapic apic_flat;
 extern struct genapic apic_physflat;
@@ -53,6 +54,9 @@ void __init setup_apic_routing(void)
 			genapic = &apic_physflat;
 		printk(KERN_INFO "Setting APIC routing to %s\n", genapic->name);
 	}
+
+	if (x86_quirks->update_genapic)
+		x86_quirks->update_genapic();
 }
 
 /* Same for both flat and physical. */

arch/x86/kernel/irq.c

@@ -118,6 +118,9 @@ int show_interrupts(struct seq_file *p, void *v)
 	}
 
 	desc = irq_to_desc(i);
+	if (!desc)
+		return 0;
+
 	spin_lock_irqsave(&desc->lock, flags);
 #ifndef CONFIG_SMP
 	any_count = kstat_irqs(i);

arch/x86/kernel/irq_32.c

@@ -242,6 +242,8 @@ void fixup_irqs(cpumask_t map)
 	for_each_irq_desc(irq, desc) {
 		cpumask_t mask;
 
+		if (!desc)
+			continue;
 		if (irq == 2)
 			continue;
 

arch/x86/kernel/irq_64.c

@@ -94,6 +94,8 @@ void fixup_irqs(cpumask_t map)
 		int break_affinity = 0;
 		int set_affinity = 1;
 
+		if (!desc)
+			continue;
 		if (irq == 2)
 			continue;
 

arch/x86/kernel/irqinit_32.c

@@ -68,8 +68,7 @@ void __init init_ISA_irqs (void)
 	/*
 	 * 16 old-style INTA-cycle interrupts:
 	 */
-	for (i = 0; i < 16; i++) {
-		/* first time call this irq_desc */
+	for (i = 0; i < NR_IRQS_LEGACY; i++) {
 		struct irq_desc *desc = irq_to_desc(i);
 
 		desc->status = IRQ_DISABLED;

arch/x86/kernel/irqinit_64.c

@@ -142,8 +142,7 @@ void __init init_ISA_irqs(void)
 	init_bsp_APIC();
 	init_8259A(0);
 
-	for (i = 0; i < 16; i++) {
-		/* first time call this irq_desc */
+	for (i = 0; i < NR_IRQS_LEGACY; i++) {
 		struct irq_desc *desc = irq_to_desc(i);
 
 		desc->status = IRQ_DISABLED;

arch/x86/kernel/mpparse.c

@@ -586,26 +586,23 @@ static void __init __get_smp_config(unsigned int early)
 {
 	struct intel_mp_floating *mpf = mpf_found;
 
-	if (x86_quirks->mach_get_smp_config) {
-		if (x86_quirks->mach_get_smp_config(early))
-			return;
-	}
+	if (!mpf)
+		return;
+
 	if (acpi_lapic && early)
 		return;
+
 	/*
-	 * ACPI supports both logical (e.g. Hyper-Threading) and physical
-	 * processors, where MPS only supports physical.
+	 * MPS doesn't support hyperthreading, aka only have
+	 * thread 0 apic id in MPS table
 	 */
-	if (acpi_lapic && acpi_ioapic) {
-		printk(KERN_INFO "Using ACPI (MADT) for SMP configuration "
-		       "information\n");
+	if (acpi_lapic && acpi_ioapic)
 		return;
-	} else if (acpi_lapic)
-		printk(KERN_INFO "Using ACPI for processor (LAPIC) "
-		       "configuration information\n");
 
-	if (!mpf)
-		return;
+	if (x86_quirks->mach_get_smp_config) {
+		if (x86_quirks->mach_get_smp_config(early))
+			return;
+	}
 
 	printk(KERN_INFO "Intel MultiProcessor Specification v1.%d\n",
 	       mpf->mpf_specification);

arch/x86/kernel/numaq_32.c

@@ -31,7 +31,7 @@
 #include <asm/numaq.h>
 #include <asm/topology.h>
 #include <asm/processor.h>
-#include <asm/mpspec.h>
+#include <asm/genapic.h>
 #include <asm/e820.h>
 #include <asm/setup.h>
 
@@ -235,6 +235,13 @@ static int __init numaq_setup_ioapic_ids(void)
 	return 1;
 }
 
+static int __init numaq_update_genapic(void)
+{
+	genapic->wakeup_cpu = wakeup_secondary_cpu_via_nmi;
+
+	return 0;
+}
+
 static struct x86_quirks numaq_x86_quirks __initdata = {
 	.arch_pre_time_init	= numaq_pre_time_init,
 	.arch_time_init		= NULL,
@@ -250,6 +257,7 @@ static struct x86_quirks numaq_x86_quirks __initdata = {
 	.mpc_oem_pci_bus	= mpc_oem_pci_bus,
 	.smp_read_mpc_oem	= smp_read_mpc_oem,
 	.setup_ioapic_ids	= numaq_setup_ioapic_ids,
+	.update_genapic		= numaq_update_genapic,
 };
 
 void numaq_mps_oem_check(struct mp_config_table *mpc, char *oem,

arch/x86/kernel/process.c

@@ -9,6 +9,7 @@
 #include <linux/clockchips.h>
 #include <linux/ftrace.h>
 #include <asm/system.h>
+#include <asm/apic.h>
 
 unsigned long idle_halt;
 EXPORT_SYMBOL(idle_halt);
@@ -127,6 +128,21 @@ void default_idle(void)
 EXPORT_SYMBOL(default_idle);
 #endif
 
+void stop_this_cpu(void *dummy)
+{
+	local_irq_disable();
+	/*
+	 * Remove this CPU:
+	 */
+	cpu_clear(smp_processor_id(), cpu_online_map);
+	disable_local_APIC();
+
+	for (;;) {
+		if (hlt_works(smp_processor_id()))
+			halt();
+	}
+}
+
 static void do_nothing(void *unused)
 {
 }

arch/x86/kernel/reboot.c

@@ -36,7 +36,10 @@ int reboot_force;
 static int reboot_cpu = -1;
 #endif
 
-/* reboot=b[ios] | s[mp] | t[riple] | k[bd] | e[fi] [, [w]arm | [c]old]
+/* This is set by the PCI code if either type 1 or type 2 PCI is detected */
+bool port_cf9_safe = false;
+
+/* reboot=b[ios] | s[mp] | t[riple] | k[bd] | e[fi] [, [w]arm | [c]old] | p[ci]
    warm   Don't set the cold reboot flag
    cold   Set the cold reboot flag
    bios   Reboot by jumping through the BIOS (only for X86_32)
@@ -45,6 +48,7 @@ static int reboot_cpu = -1;
    kbd    Use the keyboard controller. cold reset (default)
    acpi   Use the RESET_REG in the FADT
    efi    Use efi reset_system runtime service
+   pci    Use the so-called "PCI reset register", CF9
    force  Avoid anything that could hang.
  */
 static int __init reboot_setup(char *str)
@@ -79,6 +83,7 @@ static int __init reboot_setup(char *str)
 		case 'k':
 		case 't':
 		case 'e':
+		case 'p':
 			reboot_type = *str;
 			break;
 
@@ -404,12 +409,27 @@ static void native_machine_emergency_restart(void)
 			reboot_type = BOOT_KBD;
 			break;
 
-
 		case BOOT_EFI:
 			if (efi_enabled)
-				efi.reset_system(reboot_mode ? EFI_RESET_WARM : EFI_RESET_COLD,
+				efi.reset_system(reboot_mode ?
+						 EFI_RESET_WARM :
+						 EFI_RESET_COLD,
 						 EFI_SUCCESS, 0, NULL);
+			reboot_type = BOOT_KBD;
+			break;
+
+		case BOOT_CF9:
+			port_cf9_safe = true;
+			/* fall through */
 
+		case BOOT_CF9_COND:
+			if (port_cf9_safe) {
+				u8 cf9 = inb(0xcf9) & ~6;
+				outb(cf9|2, 0xcf9); /* Request hard reset */
+				udelay(50);
+				outb(cf9|6, 0xcf9); /* Actually do the reset */
+				udelay(50);
+			}
 			reboot_type = BOOT_KBD;
 			break;
 		}
@@ -470,6 +490,11 @@ static void native_machine_restart(char *__unused)
 
 static void native_machine_halt(void)
 {
+	/* stop other cpus and apics */
+	machine_shutdown();
+
+	/* stop this cpu */
+	stop_this_cpu(NULL);
 }
 
 static void native_machine_power_off(void)

arch/x86/kernel/setup.c

@@ -583,7 +583,20 @@ static int __init setup_elfcorehdr(char *arg)
 early_param("elfcorehdr", setup_elfcorehdr);
 #endif
 
-static struct x86_quirks default_x86_quirks __initdata;
+static int __init default_update_genapic(void)
+{
+#ifdef CONFIG_X86_SMP
+# if defined(CONFIG_X86_GENERICARCH) || defined(CONFIG_X86_64)
+	genapic->wakeup_cpu = wakeup_secondary_cpu_via_init;
+# endif
+#endif
+
+	return 0;
+}
+
+static struct x86_quirks default_x86_quirks __initdata = {
+	.update_genapic         = default_update_genapic,
+};
 
 struct x86_quirks *x86_quirks __initdata = &default_x86_quirks;
 
@@ -1082,7 +1095,7 @@ void __init setup_arch(char **cmdline_p)
 	ioapic_init_mappings();
 
 	/* need to wait for io_apic is mapped */
-	nr_irqs = probe_nr_irqs();
+	probe_nr_irqs_gsi();
 
 	kvm_guest_init();
 

arch/x86/kernel/smp.c

@@ -140,19 +140,6 @@ void native_send_call_func_ipi(cpumask_t mask)
 		send_IPI_mask(mask, CALL_FUNCTION_VECTOR);
 }
 
-static void stop_this_cpu(void *dummy)
-{
-	local_irq_disable();
-	/*
-	 * Remove this CPU:
-	 */
-	cpu_clear(smp_processor_id(), cpu_online_map);
-	disable_local_APIC();
-	if (hlt_works(smp_processor_id()))
-		for (;;) halt();
-	for (;;);
-}
-
 /*
  * this function calls the 'stop' function on all other CPUs in the system.
  */

arch/x86/kernel/smpboot.c

@@ -62,6 +62,7 @@
 #include <asm/mtrr.h>
 #include <asm/vmi.h>
 #include <asm/genapic.h>
+#include <asm/setup.h>
 #include <linux/mc146818rtc.h>
 
 #include <mach_apic.h>
@@ -536,7 +537,7 @@ static void impress_friends(void)
 	pr_debug("Before bogocount - setting activated=1.\n");
 }
 
-static inline void __inquire_remote_apic(int apicid)
+void __inquire_remote_apic(int apicid)
 {
 	unsigned i, regs[] = { APIC_ID >> 4, APIC_LVR >> 4, APIC_SPIV >> 4 };
 	char *names[] = { "ID", "VERSION", "SPIV" };
@@ -575,14 +576,13 @@ static inline void __inquire_remote_apic(int apicid)
 	}
 }
 
-#ifdef WAKE_SECONDARY_VIA_NMI
 /*
  * Poke the other CPU in the eye via NMI to wake it up. Remember that the normal
  * INIT, INIT, STARTUP sequence will reset the chip hard for us, and this
  * won't ... remember to clear down the APIC, etc later.
  */
-static int __devinit
-wakeup_secondary_cpu(int logical_apicid, unsigned long start_eip)
+int __devinit
+wakeup_secondary_cpu_via_nmi(int logical_apicid, unsigned long start_eip)
 {
 	unsigned long send_status, accept_status = 0;
 	int maxlvt;
@@ -599,7 +599,7 @@ wakeup_secondary_cpu(int logical_apicid, unsigned long start_eip)
 	 * Give the other CPU some time to accept the IPI.
 	 */
 	udelay(200);
-	if (APIC_INTEGRATED(apic_version[phys_apicid])) {
+	if (APIC_INTEGRATED(apic_version[boot_cpu_physical_apicid])) {
 		maxlvt = lapic_get_maxlvt();
 		if (maxlvt > 3)			/* Due to the Pentium erratum 3AP.  */
 			apic_write(APIC_ESR, 0);
@@ -614,11 +614,9 @@ wakeup_secondary_cpu(int logical_apicid, unsigned long start_eip)
 
 	return (send_status | accept_status);
 }
-#endif	/* WAKE_SECONDARY_VIA_NMI */
 
-#ifdef WAKE_SECONDARY_VIA_INIT
-static int __devinit
-wakeup_secondary_cpu(int phys_apicid, unsigned long start_eip)
+int __devinit
+wakeup_secondary_cpu_via_init(int phys_apicid, unsigned long start_eip)
 {
 	unsigned long send_status, accept_status = 0;
 	int maxlvt, num_starts, j;
@@ -737,7 +735,6 @@ wakeup_secondary_cpu(int phys_apicid, unsigned long start_eip)
 
 	return (send_status | accept_status);
 }
-#endif	/* WAKE_SECONDARY_VIA_INIT */
 
 struct create_idle {
 	struct work_struct work;

arch/x86/mach-generic/bigsmp.c

@@ -17,6 +17,7 @@
 #include <asm/bigsmp/apic.h>
 #include <asm/bigsmp/ipi.h>
 #include <asm/mach-default/mach_mpparse.h>
+#include <asm/mach-default/mach_wakecpu.h>
 
 static int dmi_bigsmp; /* can be set by dmi scanners */
 

arch/x86/mach-generic/default.c

@@ -16,6 +16,7 @@
 #include <asm/mach-default/mach_apic.h>
 #include <asm/mach-default/mach_ipi.h>
 #include <asm/mach-default/mach_mpparse.h>
+#include <asm/mach-default/mach_wakecpu.h>
 
 /* should be called last. */
 static int probe_default(void)

arch/x86/mach-generic/es7000.c

@@ -16,7 +16,19 @@
 #include <asm/es7000/apic.h>
 #include <asm/es7000/ipi.h>
 #include <asm/es7000/mpparse.h>
-#include <asm/es7000/wakecpu.h>
+#include <asm/mach-default/mach_wakecpu.h>
+
+void __init es7000_update_genapic_to_cluster(void)
+{
+	genapic->target_cpus = target_cpus_cluster;
+	genapic->int_delivery_mode = INT_DELIVERY_MODE_CLUSTER;
+	genapic->int_dest_mode = INT_DEST_MODE_CLUSTER;
+	genapic->no_balance_irq = NO_BALANCE_IRQ_CLUSTER;
+
+	genapic->init_apic_ldr = init_apic_ldr_cluster;
+
+	genapic->cpu_mask_to_apicid = cpu_mask_to_apicid_cluster;
+}
 
 static int probe_es7000(void)
 {

arch/x86/mach-generic/probe.c

@@ -15,6 +15,7 @@
 #include <asm/mpspec.h>
 #include <asm/apicdef.h>
 #include <asm/genapic.h>
+#include <asm/setup.h>
 
 extern struct genapic apic_numaq;
 extern struct genapic apic_summit;
@@ -57,6 +58,9 @@ static int __init parse_apic(char *arg)
 		}
 	}
 
+	if (x86_quirks->update_genapic)
+		x86_quirks->update_genapic();
+
 	/* Parsed again by __setup for debug/verbose */
 	return 0;
 }
@@ -72,12 +76,15 @@ void __init generic_bigsmp_probe(void)
 	 * - we find more than 8 CPUs in acpi LAPIC listing with xAPIC support
 	 */
 
-	if (!cmdline_apic && genapic == &apic_default)
+	if (!cmdline_apic && genapic == &apic_default) {
 		if (apic_bigsmp.probe()) {
 			genapic = &apic_bigsmp;
+			if (x86_quirks->update_genapic)
+				x86_quirks->update_genapic();
 			printk(KERN_INFO "Overriding APIC driver with %s\n",
 			       genapic->name);
 		}
+	}
 #endif
 }
 
@@ -94,6 +101,9 @@ void __init generic_apic_probe(void)
 		/* Not visible without early console */
 		if (!apic_probe[i])
 			panic("Didn't find an APIC driver");
+
+		if (x86_quirks->update_genapic)
+			x86_quirks->update_genapic();
 	}
 	printk(KERN_INFO "Using APIC driver %s\n", genapic->name);
 }
@@ -108,6 +118,8 @@ int __init mps_oem_check(struct mp_config_table *mpc, char *oem,
 		if (apic_probe[i]->mps_oem_check(mpc, oem, productid)) {
 			if (!cmdline_apic) {
 				genapic = apic_probe[i];
+				if (x86_quirks->update_genapic)
+					x86_quirks->update_genapic();
 				printk(KERN_INFO "Switched to APIC driver `%s'.\n",
 				       genapic->name);
 			}
@@ -124,6 +136,8 @@ int __init acpi_madt_oem_check(char *oem_id, char *oem_table_id)
 		if (apic_probe[i]->acpi_madt_oem_check(oem_id, oem_table_id)) {
 			if (!cmdline_apic) {
 				genapic = apic_probe[i];
+				if (x86_quirks->update_genapic)
+					x86_quirks->update_genapic();
 				printk(KERN_INFO "Switched to APIC driver `%s'.\n",
 				       genapic->name);
 			}

arch/x86/mach-generic/summit.c

@@ -16,6 +16,7 @@
 #include <asm/summit/apic.h>
 #include <asm/summit/ipi.h>
 #include <asm/summit/mpparse.h>
+#include <asm/mach-default/mach_wakecpu.h>
 
 static int probe_summit(void)
 {

arch/x86/pci/direct.c

@@ -173,7 +173,7 @@ static int pci_conf2_write(unsigned int seg, unsigned int bus,
 
 #undef PCI_CONF2_ADDRESS
 
-static struct pci_raw_ops pci_direct_conf2 = {
+struct pci_raw_ops pci_direct_conf2 = {
 	.read =		pci_conf2_read,
 	.write =	pci_conf2_write,
 };
@@ -289,6 +289,7 @@ int __init pci_direct_probe(void)
 
 	if (pci_check_type1()) {
 		raw_pci_ops = &pci_direct_conf1;
+		port_cf9_safe = true;
 		return 1;
 	}
 	release_resource(region);
@@ -305,6 +306,7 @@ int __init pci_direct_probe(void)
 
 	if (pci_check_type2()) {
 		raw_pci_ops = &pci_direct_conf2;
+		port_cf9_safe = true;
 		return 2;
 	}
 

arch/x86/pci/pci.h

@@ -96,6 +96,7 @@ extern struct pci_raw_ops *raw_pci_ops;
 extern struct pci_raw_ops *raw_pci_ext_ops;
 
 extern struct pci_raw_ops pci_direct_conf1;
+extern bool port_cf9_safe;
 
 /* arch_initcall level */
 extern int pci_direct_probe(void);

drivers/char/random.c

@@ -558,23 +558,9 @@ struct timer_rand_state {
 	unsigned dont_count_entropy:1;
 };
 
-static struct timer_rand_state *irq_timer_state[NR_IRQS];
-
-static struct timer_rand_state *get_timer_rand_state(unsigned int irq)
-{
-	if (irq >= nr_irqs)
-		return NULL;
-
-	return irq_timer_state[irq];
-}
-
-static void set_timer_rand_state(unsigned int irq, struct timer_rand_state *state)
-{
-	if (irq >= nr_irqs)
-		return;
-
-	irq_timer_state[irq] = state;
-}
+#ifndef CONFIG_SPARSE_IRQ
+struct timer_rand_state *irq_timer_state[NR_IRQS];
+#endif
 
 static struct timer_rand_state input_timer_state;
 
@@ -933,8 +919,10 @@ void rand_initialize_irq(int irq)
 {
 	struct timer_rand_state *state;
 
+#ifndef CONFIG_SPARSE_IRQ
 	if (irq >= nr_irqs)
 		return;
+#endif
 
 	state = get_timer_rand_state(irq);
 

drivers/pci/intr_remapping.c

@@ -19,17 +19,75 @@ struct irq_2_iommu {
 	u8  irte_mask;
 };
 
-static struct irq_2_iommu irq_2_iommuX[NR_IRQS];
+#ifdef CONFIG_SPARSE_IRQ
+static struct irq_2_iommu *get_one_free_irq_2_iommu(int cpu)
+{
+	struct irq_2_iommu *iommu;
+	int node;
+
+	node = cpu_to_node(cpu);
+
+	iommu = kzalloc_node(sizeof(*iommu), GFP_ATOMIC, node);
+	printk(KERN_DEBUG "alloc irq_2_iommu on cpu %d node %d\n", cpu, node);
+
+	return iommu;
+}
 
 static struct irq_2_iommu *irq_2_iommu(unsigned int irq)
 {
-	return (irq < nr_irqs) ? irq_2_iommuX + irq : NULL;
+	struct irq_desc *desc;
+
+	desc = irq_to_desc(irq);
+
+	if (WARN_ON_ONCE(!desc))
+		return NULL;
+
+	return desc->irq_2_iommu;
+}
+
+static struct irq_2_iommu *irq_2_iommu_alloc_cpu(unsigned int irq, int cpu)
+{
+	struct irq_desc *desc;
+	struct irq_2_iommu *irq_iommu;
+
+	/*
+	 * alloc irq desc if not allocated already.
+	 */
+	desc = irq_to_desc_alloc_cpu(irq, cpu);
+	if (!desc) {
+		printk(KERN_INFO "can not get irq_desc for %d\n", irq);
+		return NULL;
+	}
+
+	irq_iommu = desc->irq_2_iommu;
+
+	if (!irq_iommu)
+		desc->irq_2_iommu = get_one_free_irq_2_iommu(cpu);
+
+	return desc->irq_2_iommu;
 }
 
+static struct irq_2_iommu *irq_2_iommu_alloc(unsigned int irq)
+{
+	return irq_2_iommu_alloc_cpu(irq, boot_cpu_id);
+}
+
+#else /* !CONFIG_SPARSE_IRQ */
+
+static struct irq_2_iommu irq_2_iommuX[NR_IRQS];
+
+static struct irq_2_iommu *irq_2_iommu(unsigned int irq)
+{
+	if (irq < nr_irqs)
+		return &irq_2_iommuX[irq];
+
+	return NULL;
+}
 static struct irq_2_iommu *irq_2_iommu_alloc(unsigned int irq)
 {
 	return irq_2_iommu(irq);
 }
+#endif
 
 static DEFINE_SPINLOCK(irq_2_ir_lock);
 
@@ -86,9 +144,11 @@ int alloc_irte(struct intel_iommu *iommu, int irq, u16 count)
 	if (!count)
 		return -1;
 
+#ifndef CONFIG_SPARSE_IRQ
 	/* protect irq_2_iommu_alloc later */
 	if (irq >= nr_irqs)
 		return -1;
+#endif
 
 	/*
 	 * start the IRTE search from index 0.
@@ -130,6 +190,12 @@ int alloc_irte(struct intel_iommu *iommu, int irq, u16 count)
 		table->base[i].present = 1;
 
 	irq_iommu = irq_2_iommu_alloc(irq);
+	if (!irq_iommu) {
+		spin_unlock(&irq_2_ir_lock);
+		printk(KERN_ERR "can't allocate irq_2_iommu\n");
+		return -1;
+	}
+
 	irq_iommu->iommu = iommu;
 	irq_iommu->irte_index =  index;
 	irq_iommu->sub_handle = 0;
@@ -177,6 +243,12 @@ int set_irte_irq(int irq, struct intel_iommu *iommu, u16 index, u16 subhandle)
 
 	irq_iommu = irq_2_iommu_alloc(irq);
 
+	if (!irq_iommu) {
+		spin_unlock(&irq_2_ir_lock);
+		printk(KERN_ERR "can't allocate irq_2_iommu\n");
+		return -1;
+	}
+
 	irq_iommu->iommu = iommu;
 	irq_iommu->irte_index = index;
 	irq_iommu->sub_handle = subhandle;

drivers/pci/msi.c

@@ -103,11 +103,11 @@ static void msix_set_enable(struct pci_dev *dev, int enable)
 	}
 }
 
-static void msix_flush_writes(unsigned int irq)
+static void msix_flush_writes(struct irq_desc *desc)
 {
 	struct msi_desc *entry;
 
-	entry = get_irq_msi(irq);
+	entry = get_irq_desc_msi(desc);
 	BUG_ON(!entry || !entry->dev);
 	switch (entry->msi_attrib.type) {
 	case PCI_CAP_ID_MSI:
@@ -135,11 +135,11 @@ static void msix_flush_writes(unsigned int irq)
  * Returns 1 if it succeeded in masking the interrupt and 0 if the device
  * doesn't support MSI masking.
  */
-static int msi_set_mask_bits(unsigned int irq, u32 mask, u32 flag)
+static int msi_set_mask_bits(struct irq_desc *desc, u32 mask, u32 flag)
 {
 	struct msi_desc *entry;
 
-	entry = get_irq_msi(irq);
+	entry = get_irq_desc_msi(desc);
 	BUG_ON(!entry || !entry->dev);
 	switch (entry->msi_attrib.type) {
 	case PCI_CAP_ID_MSI:
@@ -172,9 +172,9 @@ static int msi_set_mask_bits(unsigned int irq, u32 mask, u32 flag)
 	return 1;
 }
 
-void read_msi_msg(unsigned int irq, struct msi_msg *msg)
+void read_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg)
 {
-	struct msi_desc *entry = get_irq_msi(irq);
+	struct msi_desc *entry = get_irq_desc_msi(desc);
 	switch(entry->msi_attrib.type) {
 	case PCI_CAP_ID_MSI:
 	{
@@ -211,9 +211,16 @@ void read_msi_msg(unsigned int irq, struct msi_msg *msg)
 	}
 }
 
-void write_msi_msg(unsigned int irq, struct msi_msg *msg)
+void read_msi_msg(unsigned int irq, struct msi_msg *msg)
 {
-	struct msi_desc *entry = get_irq_msi(irq);
+	struct irq_desc *desc = irq_to_desc(irq);
+
+	read_msi_msg_desc(desc, msg);
+}
+
+void write_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg)
+{
+	struct msi_desc *entry = get_irq_desc_msi(desc);
 	switch (entry->msi_attrib.type) {
 	case PCI_CAP_ID_MSI:
 	{
@@ -252,21 +259,31 @@ void write_msi_msg(unsigned int irq, struct msi_msg *msg)
 	entry->msg = *msg;
 }
 
+void write_msi_msg(unsigned int irq, struct msi_msg *msg)
+{
+	struct irq_desc *desc = irq_to_desc(irq);
+
+	write_msi_msg_desc(desc, msg);
+}
+
 void mask_msi_irq(unsigned int irq)
 {
-	msi_set_mask_bits(irq, 1, 1);
-	msix_flush_writes(irq);
+	struct irq_desc *desc = irq_to_desc(irq);
+
+	msi_set_mask_bits(desc, 1, 1);
+	msix_flush_writes(desc);
 }
 
 void unmask_msi_irq(unsigned int irq)
 {
-	msi_set_mask_bits(irq, 1, 0);
-	msix_flush_writes(irq);
+	struct irq_desc *desc = irq_to_desc(irq);
+
+	msi_set_mask_bits(desc, 1, 0);
+	msix_flush_writes(desc);
 }
 
 static int msi_free_irqs(struct pci_dev* dev);
 
-
 static struct msi_desc* alloc_msi_entry(void)
 {
 	struct msi_desc *entry;
@@ -303,9 +320,11 @@ static void __pci_restore_msi_state(struct pci_dev *dev)
 	pci_intx_for_msi(dev, 0);
 	msi_set_enable(dev, 0);
 	write_msi_msg(dev->irq, &entry->msg);
-	if (entry->msi_attrib.maskbit)
-		msi_set_mask_bits(dev->irq, entry->msi_attrib.maskbits_mask,
+	if (entry->msi_attrib.maskbit) {
+		struct irq_desc *desc = irq_to_desc(dev->irq);
+		msi_set_mask_bits(desc, entry->msi_attrib.maskbits_mask,
 				  entry->msi_attrib.masked);
+	}
 
 	pci_read_config_word(dev, pos + PCI_MSI_FLAGS, &control);
 	control &= ~PCI_MSI_FLAGS_QSIZE;
@@ -327,8 +346,9 @@ static void __pci_restore_msix_state(struct pci_dev *dev)
 	msix_set_enable(dev, 0);
 
 	list_for_each_entry(entry, &dev->msi_list, list) {
+		struct irq_desc *desc = irq_to_desc(entry->irq);
 		write_msi_msg(entry->irq, &entry->msg);
-		msi_set_mask_bits(entry->irq, 1, entry->msi_attrib.masked);
+		msi_set_mask_bits(desc, 1, entry->msi_attrib.masked);
 	}
 
 	BUG_ON(list_empty(&dev->msi_list));
@@ -596,7 +616,8 @@ void pci_msi_shutdown(struct pci_dev* dev)
 	/* Return the the pci reset with msi irqs unmasked */
 	if (entry->msi_attrib.maskbit) {
 		u32 mask = entry->msi_attrib.maskbits_mask;
-		msi_set_mask_bits(dev->irq, mask, ~mask);
+		struct irq_desc *desc = irq_to_desc(dev->irq);
+		msi_set_mask_bits(desc, mask, ~mask);
 	}
 	if (!entry->dev || entry->msi_attrib.type != PCI_CAP_ID_MSI)
 		return;

drivers/xen/events.c

@@ -141,8 +141,12 @@ static void init_evtchn_cpu_bindings(void)
 	int i;
 
 	/* By default all event channels notify CPU#0. */
-	for_each_irq_desc(i, desc)
+	for_each_irq_desc(i, desc) {
+		if (!desc)
+			continue;
+
 		desc->affinity = cpumask_of_cpu(0);
+	}
 #endif
 
 	memset(cpu_evtchn, 0, sizeof(cpu_evtchn));
@@ -231,7 +235,7 @@ static int find_unbound_irq(void)
 	int irq;
 
 	/* Only allocate from dynirq range */
-	for_each_irq_nr(irq)
+	for (irq = 0; irq < nr_irqs; irq++)
 		if (irq_bindcount[irq] == 0)
 			break;
 
@@ -792,7 +796,7 @@ void xen_irq_resume(void)
 		mask_evtchn(evtchn);
 
 	/* No IRQ <-> event-channel mappings. */
-	for_each_irq_nr(irq)
+	for (irq = 0; irq < nr_irqs; irq++)
 		irq_info[irq].evtchn = 0; /* zap event-channel binding */
 
 	for (evtchn = 0; evtchn < NR_EVENT_CHANNELS; evtchn++)
@@ -824,7 +828,7 @@ void __init xen_init_IRQ(void)
 		mask_evtchn(i);
 
 	/* Dynamic IRQ space is currently unbound. Zero the refcnts. */
-	for_each_irq_nr(i)
+	for (i = 0; i < nr_irqs; i++)
 		irq_bindcount[i] = 0;
 
 	irq_ctx_init(smp_processor_id());

fs/proc/stat.c

@@ -27,6 +27,7 @@ static int show_stat(struct seq_file *p, void *v)
 	u64 sum = 0;
 	struct timespec boottime;
 	unsigned int per_irq_sum;
+	struct irq_desc *desc;
 
 	user = nice = system = idle = iowait =
 		irq = softirq = steal = cputime64_zero;
@@ -44,10 +45,14 @@ static int show_stat(struct seq_file *p, void *v)
 		softirq = cputime64_add(softirq, kstat_cpu(i).cpustat.softirq);
 		steal = cputime64_add(steal, kstat_cpu(i).cpustat.steal);
 		guest = cputime64_add(guest, kstat_cpu(i).cpustat.guest);
-
-		for_each_irq_nr(j)
+		for_each_irq_nr(j) {
+#ifdef CONFIG_SPARSE_IRQ
+			desc = irq_to_desc(j);
+			if (!desc)
+				continue;
+#endif
 			sum += kstat_irqs_cpu(j, i);
-
+		}
 		sum += arch_irq_stat_cpu(i);
 	}
 	sum += arch_irq_stat();
@@ -92,7 +97,13 @@ static int show_stat(struct seq_file *p, void *v)
 	/* sum again ? it could be updated? */
 	for_each_irq_nr(j) {
 		per_irq_sum = 0;
-
+#ifdef CONFIG_SPARSE_IRQ
+		desc = irq_to_desc(j);
+		if (!desc) {
+			seq_printf(p, " %u", per_irq_sum);
+			continue;
+		}
+#endif
 		for_each_possible_cpu(i)
 			per_irq_sum += kstat_irqs_cpu(j, i);
 

include/linux/Kbuild

@@ -313,6 +313,7 @@ unifdef-y += ptrace.h
 unifdef-y += qnx4_fs.h
 unifdef-y += quota.h
 unifdef-y += random.h
+unifdef-y += irqnr.h
 unifdef-y += reboot.h
 unifdef-y += reiserfs_fs.h
 unifdef-y += reiserfs_xattr.h

include/linux/interrupt.h

@@ -14,6 +14,8 @@
 #include <linux/irqflags.h>
 #include <linux/smp.h>
 #include <linux/percpu.h>
+#include <linux/irqnr.h>
+
 #include <asm/atomic.h>
 #include <asm/ptrace.h>
 #include <asm/system.h>

include/linux/irq.h

@@ -129,6 +129,8 @@ struct irq_chip {
 	const char	*typename;
 };
 
+struct timer_rand_state;
+struct irq_2_iommu;
 /**
  * struct irq_desc - interrupt descriptor
  * @irq:		interrupt number for this descriptor
@@ -154,6 +156,13 @@ struct irq_chip {
  */
 struct irq_desc {
 	unsigned int		irq;
+#ifdef CONFIG_SPARSE_IRQ
+	struct timer_rand_state *timer_rand_state;
+	unsigned int            *kstat_irqs;
+# ifdef CONFIG_INTR_REMAP
+	struct irq_2_iommu      *irq_2_iommu;
+# endif
+#endif
 	irq_flow_handler_t	handle_irq;
 	struct irq_chip		*chip;
 	struct msi_desc		*msi_desc;
@@ -181,14 +190,43 @@ struct irq_desc {
 	const char		*name;
 } ____cacheline_internodealigned_in_smp;
 
+extern void early_irq_init(void);
+extern void arch_early_irq_init(void);
+extern void arch_init_chip_data(struct irq_desc *desc, int cpu);
+extern void arch_init_copy_chip_data(struct irq_desc *old_desc,
+					struct irq_desc *desc, int cpu);
+extern void arch_free_chip_data(struct irq_desc *old_desc, struct irq_desc *desc);
 
+#ifndef CONFIG_SPARSE_IRQ
 extern struct irq_desc irq_desc[NR_IRQS];
 
 static inline struct irq_desc *irq_to_desc(unsigned int irq)
 {
-	return (irq < nr_irqs) ? irq_desc + irq : NULL;
+	return (irq < NR_IRQS) ? irq_desc + irq : NULL;
+}
+static inline struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu)
+{
+	return irq_to_desc(irq);
 }
 
+#else
+
+extern struct irq_desc *irq_to_desc(unsigned int irq);
+extern struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu);
+extern struct irq_desc *move_irq_desc(struct irq_desc *old_desc, int cpu);
+
+# define for_each_irq_desc(irq, desc)		\
+	for (irq = 0, desc = irq_to_desc(irq); irq < nr_irqs; irq++, desc = irq_to_desc(irq))
+# define for_each_irq_desc_reverse(irq, desc)                          \
+	for (irq = nr_irqs - 1, desc = irq_to_desc(irq); irq >= 0; irq--, desc = irq_to_desc(irq))
+
+#define kstat_irqs_this_cpu(DESC) \
+	((DESC)->kstat_irqs[smp_processor_id()])
+#define kstat_incr_irqs_this_cpu(irqno, DESC) \
+	((DESC)->kstat_irqs[smp_processor_id()]++)
+
+#endif
+
 /*
  * Migration helpers for obsolete names, they will go away:
  */
@@ -380,6 +418,11 @@ extern int set_irq_msi(unsigned int irq, struct msi_desc *entry);
 #define get_irq_data(irq)	(irq_to_desc(irq)->handler_data)
 #define get_irq_msi(irq)	(irq_to_desc(irq)->msi_desc)
 
+#define get_irq_desc_chip(desc)		((desc)->chip)
+#define get_irq_desc_chip_data(desc)	((desc)->chip_data)
+#define get_irq_desc_data(desc)		((desc)->handler_data)
+#define get_irq_desc_msi(desc)		((desc)->msi_desc)
+
 #endif /* CONFIG_GENERIC_HARDIRQS */
 
 #endif /* !CONFIG_S390 */

include/linux/irqnr.h

 #ifndef _LINUX_IRQNR_H
 #define _LINUX_IRQNR_H
 
+/*
+ * Generic irq_desc iterators:
+ */
+#ifdef __KERNEL__
+
 #ifndef CONFIG_GENERIC_HARDIRQS
 #include <asm/irq.h>
 # define nr_irqs		NR_IRQS
 
 # define for_each_irq_desc(irq, desc)		\
 	for (irq = 0; irq < nr_irqs; irq++)
+
+# define for_each_irq_desc_reverse(irq, desc)                          \
+	for (irq = nr_irqs - 1; irq >= 0; irq--)
 #else
+
 extern int nr_irqs;
 
+#ifndef CONFIG_SPARSE_IRQ
+
+struct irq_desc;
 # define for_each_irq_desc(irq, desc)		\
 	for (irq = 0, desc = irq_desc; irq < nr_irqs; irq++, desc++)
-
-# define for_each_irq_desc_reverse(irq, desc)				\
-	for (irq = nr_irqs - 1, desc = irq_desc + (nr_irqs - 1);	\
-	     irq >= 0; irq--, desc--)
+# define for_each_irq_desc_reverse(irq, desc)                          \
+	for (irq = nr_irqs - 1, desc = irq_desc + (nr_irqs - 1);        \
+	    irq >= 0; irq--, desc--)
+#endif
 #endif
 
-#define for_each_irq_nr(irq)			\
-	for (irq = 0; irq < nr_irqs; irq++)
+#define for_each_irq_nr(irq)                   \
+       for (irq = 0; irq < nr_irqs; irq++)
+
+#endif /* __KERNEL__ */
 
 #endif

include/linux/kernel_stat.h

@@ -28,7 +28,9 @@ struct cpu_usage_stat {
 
 struct kernel_stat {
 	struct cpu_usage_stat	cpustat;
-	unsigned int irqs[NR_IRQS];
+#ifndef CONFIG_SPARSE_IRQ
+       unsigned int irqs[NR_IRQS];
+#endif
 };
 
 DECLARE_PER_CPU(struct kernel_stat, kstat);
@@ -39,6 +41,10 @@ DECLARE_PER_CPU(struct kernel_stat, kstat);
 
 extern unsigned long long nr_context_switches(void);
 
+#ifndef CONFIG_SPARSE_IRQ
+#define kstat_irqs_this_cpu(irq) \
+	(kstat_this_cpu.irqs[irq])
+
 struct irq_desc;
 
 static inline void kstat_incr_irqs_this_cpu(unsigned int irq,
@@ -46,11 +52,17 @@ static inline void kstat_incr_irqs_this_cpu(unsigned int irq,
 {
 	kstat_this_cpu.irqs[irq]++;
 }
+#endif
+
 
+#ifndef CONFIG_SPARSE_IRQ
 static inline unsigned int kstat_irqs_cpu(unsigned int irq, int cpu)
 {
        return kstat_cpu(cpu).irqs[irq];
 }
+#else
+extern unsigned int kstat_irqs_cpu(unsigned int irq, int cpu);
+#endif
 
 /*
  * Number of interrupts per specific IRQ source, since bootup

include/linux/msi.h

@@ -10,8 +10,11 @@ struct msi_msg {
 };
 
 /* Helper functions */
+struct irq_desc;
 extern void mask_msi_irq(unsigned int irq);
 extern void unmask_msi_irq(unsigned int irq);
+extern void read_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg);
+extern void write_msi_msg_desc(struct irq_desc *desc, struct msi_msg *msg);
 extern void read_msi_msg(unsigned int irq, struct msi_msg *msg);
 extern void write_msi_msg(unsigned int irq, struct msi_msg *msg);
 

include/linux/random.h

@@ -8,6 +8,7 @@
 #define _LINUX_RANDOM_H
 
 #include <linux/ioctl.h>
+#include <linux/irqnr.h>
 
 /* ioctl()'s for the random number generator */
 
@@ -44,6 +45,56 @@ struct rand_pool_info {
 
 extern void rand_initialize_irq(int irq);
 
+struct timer_rand_state;
+#ifndef CONFIG_SPARSE_IRQ
+
+extern struct timer_rand_state *irq_timer_state[];
+
+static inline struct timer_rand_state *get_timer_rand_state(unsigned int irq)
+{
+	if (irq >= nr_irqs)
+		return NULL;
+
+	return irq_timer_state[irq];
+}
+
+static inline void set_timer_rand_state(unsigned int irq, struct timer_rand_state *state)
+{
+	if (irq >= nr_irqs)
+		return;
+
+	irq_timer_state[irq] = state;
+}
+
+#else
+
+#include <linux/irq.h>
+static inline struct timer_rand_state *get_timer_rand_state(unsigned int irq)
+{
+	struct irq_desc *desc;
+
+	desc = irq_to_desc(irq);
+
+	if (!desc)
+		return NULL;
+
+	return desc->timer_rand_state;
+}
+
+static inline void set_timer_rand_state(unsigned int irq, struct timer_rand_state *state)
+{
+	struct irq_desc *desc;
+
+	desc = irq_to_desc(irq);
+
+	if (!desc)
+		return;
+
+	desc->timer_rand_state = state;
+}
+#endif
+
+
 extern void add_input_randomness(unsigned int type, unsigned int code,
 				 unsigned int value);
 extern void add_interrupt_randomness(int irq);

init/main.c

@@ -540,6 +540,15 @@ void __init __weak thread_info_cache_init(void)
 {
 }
 
+void __init __weak arch_early_irq_init(void)
+{
+}
+
+void __init __weak early_irq_init(void)
+{
+	arch_early_irq_init();
+}
+
 asmlinkage void __init start_kernel(void)
 {
 	char * command_line;
@@ -604,6 +613,8 @@ asmlinkage void __init start_kernel(void)
 	sort_main_extable();
 	trap_init();
 	rcu_init();
+	/* init some links before init_ISA_irqs() */
+	early_irq_init();
 	init_IRQ();
 	pidhash_init();
 	init_timers();

kernel/irq/autoprobe.c

@@ -40,6 +40,9 @@ unsigned long probe_irq_on(void)
 	 * flush such a longstanding irq before considering it as spurious.
 	 */
 	for_each_irq_desc_reverse(i, desc) {
+		if (!desc)
+			continue;
+
 		spin_lock_irq(&desc->lock);
 		if (!desc->action && !(desc->status & IRQ_NOPROBE)) {
 			/*
@@ -68,6 +71,9 @@ unsigned long probe_irq_on(void)
 	 * happened in the previous stage, it may have masked itself)
 	 */
 	for_each_irq_desc_reverse(i, desc) {
+		if (!desc)
+			continue;
+
 		spin_lock_irq(&desc->lock);
 		if (!desc->action && !(desc->status & IRQ_NOPROBE)) {
 			desc->status |= IRQ_AUTODETECT | IRQ_WAITING;
@@ -86,6 +92,9 @@ unsigned long probe_irq_on(void)
 	 * Now filter out any obviously spurious interrupts
 	 */
 	for_each_irq_desc(i, desc) {
+		if (!desc)
+			continue;
+
 		spin_lock_irq(&desc->lock);
 		status = desc->status;
 
@@ -124,6 +133,9 @@ unsigned int probe_irq_mask(unsigned long val)
 	int i;
 
 	for_each_irq_desc(i, desc) {
+		if (!desc)
+			continue;
+
 		spin_lock_irq(&desc->lock);
 		status = desc->status;
 
@@ -166,6 +178,9 @@ int probe_irq_off(unsigned long val)
 	unsigned int status;
 
 	for_each_irq_desc(i, desc) {
+		if (!desc)
+			continue;
+
 		spin_lock_irq(&desc->lock);
 		status = desc->status;
 

kernel/irq/chip.c

@@ -24,9 +24,10 @@
  */
 void dynamic_irq_init(unsigned int irq)
 {
-	struct irq_desc *desc = irq_to_desc(irq);
+	struct irq_desc *desc;
 	unsigned long flags;
 
+	desc = irq_to_desc(irq);
 	if (!desc) {
 		WARN(1, KERN_ERR "Trying to initialize invalid IRQ%d\n", irq);
 		return;

kernel/irq/handle.c

@@ -15,9 +15,16 @@
 #include <linux/random.h>
 #include <linux/interrupt.h>
 #include <linux/kernel_stat.h>
+#include <linux/rculist.h>
+#include <linux/hash.h>
 
 #include "internals.h"
 
+/*
+ * lockdep: we want to handle all irq_desc locks as a single lock-class:
+ */
+static struct lock_class_key irq_desc_lock_class;
+
 /**
  * handle_bad_irq - handle spurious and unhandled irqs
  * @irq:       the interrupt number
@@ -49,6 +56,155 @@ void handle_bad_irq(unsigned int irq, struct irq_desc *desc)
 int nr_irqs = NR_IRQS;
 EXPORT_SYMBOL_GPL(nr_irqs);
 
+void __init __attribute__((weak)) arch_early_irq_init(void)
+{
+}
+
+#ifdef CONFIG_SPARSE_IRQ
+static struct irq_desc irq_desc_init = {
+	.irq	    = -1,
+	.status	    = IRQ_DISABLED,
+	.chip	    = &no_irq_chip,
+	.handle_irq = handle_bad_irq,
+	.depth      = 1,
+	.lock       = __SPIN_LOCK_UNLOCKED(irq_desc_init.lock),
+#ifdef CONFIG_SMP
+	.affinity   = CPU_MASK_ALL
+#endif
+};
+
+static void init_kstat_irqs(struct irq_desc *desc, int cpu, int nr)
+{
+	unsigned long bytes;
+	char *ptr;
+	int node;
+
+	/* Compute how many bytes we need per irq and allocate them */
+	bytes = nr * sizeof(unsigned int);
+
+	node = cpu_to_node(cpu);
+	ptr = kzalloc_node(bytes, GFP_ATOMIC, node);
+	printk(KERN_DEBUG "  alloc kstat_irqs on cpu %d node %d\n", cpu, node);
+
+	if (ptr)
+		desc->kstat_irqs = (unsigned int *)ptr;
+}
+
+void __attribute__((weak)) arch_init_chip_data(struct irq_desc *desc, int cpu)
+{
+}
+
+static void init_one_irq_desc(int irq, struct irq_desc *desc, int cpu)
+{
+	memcpy(desc, &irq_desc_init, sizeof(struct irq_desc));
+	desc->irq = irq;
+#ifdef CONFIG_SMP
+	desc->cpu = cpu;
+#endif
+	lockdep_set_class(&desc->lock, &irq_desc_lock_class);
+	init_kstat_irqs(desc, cpu, nr_cpu_ids);
+	if (!desc->kstat_irqs) {
+		printk(KERN_ERR "can not alloc kstat_irqs\n");
+		BUG_ON(1);
+	}
+	arch_init_chip_data(desc, cpu);
+}
+
+/*
+ * Protect the sparse_irqs:
+ */
+static DEFINE_SPINLOCK(sparse_irq_lock);
+
+struct irq_desc *irq_desc_ptrs[NR_IRQS] __read_mostly;
+
+static struct irq_desc irq_desc_legacy[NR_IRQS_LEGACY] __cacheline_aligned_in_smp = {
+	[0 ... NR_IRQS_LEGACY-1] = {
+		.irq	    = -1,
+		.status	    = IRQ_DISABLED,
+		.chip	    = &no_irq_chip,
+		.handle_irq = handle_bad_irq,
+		.depth	    = 1,
+		.lock	    = __SPIN_LOCK_UNLOCKED(irq_desc_init.lock),
+#ifdef CONFIG_SMP
+		.affinity   = CPU_MASK_ALL
+#endif
+	}
+};
+
+/* FIXME: use bootmem alloc ...*/
+static unsigned int kstat_irqs_legacy[NR_IRQS_LEGACY][NR_CPUS];
+
+void __init early_irq_init(void)
+{
+	struct irq_desc *desc;
+	int legacy_count;
+	int i;
+
+	desc = irq_desc_legacy;
+	legacy_count = ARRAY_SIZE(irq_desc_legacy);
+
+	for (i = 0; i < legacy_count; i++) {
+		desc[i].irq = i;
+		desc[i].kstat_irqs = kstat_irqs_legacy[i];
+
+		irq_desc_ptrs[i] = desc + i;
+	}
+
+	for (i = legacy_count; i < NR_IRQS; i++)
+		irq_desc_ptrs[i] = NULL;
+
+	arch_early_irq_init();
+}
+
+struct irq_desc *irq_to_desc(unsigned int irq)
+{
+	return (irq < NR_IRQS) ? irq_desc_ptrs[irq] : NULL;
+}
+
+struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu)
+{
+	struct irq_desc *desc;
+	unsigned long flags;
+	int node;
+
+	if (irq >= NR_IRQS) {
+		printk(KERN_WARNING "irq >= NR_IRQS in irq_to_desc_alloc: %d %d\n",
+				irq, NR_IRQS);
+		WARN_ON(1);
+		return NULL;
+	}
+
+	desc = irq_desc_ptrs[irq];
+	if (desc)
+		return desc;
+
+	spin_lock_irqsave(&sparse_irq_lock, flags);
+
+	/* We have to check it to avoid races with another CPU */
+	desc = irq_desc_ptrs[irq];
+	if (desc)
+		goto out_unlock;
+
+	node = cpu_to_node(cpu);
+	desc = kzalloc_node(sizeof(*desc), GFP_ATOMIC, node);
+	printk(KERN_DEBUG "  alloc irq_desc for %d on cpu %d node %d\n",
+		 irq, cpu, node);
+	if (!desc) {
+		printk(KERN_ERR "can not alloc irq_desc\n");
+		BUG_ON(1);
+	}
+	init_one_irq_desc(irq, desc, cpu);
+
+	irq_desc_ptrs[irq] = desc;
+
+out_unlock:
+	spin_unlock_irqrestore(&sparse_irq_lock, flags);
+
+	return desc;
+}
+
+#else
+
 struct irq_desc irq_desc[NR_IRQS] __cacheline_aligned_in_smp = {
 	[0 ... NR_IRQS-1] = {
 		.status = IRQ_DISABLED,
@@ -62,6 +218,8 @@ struct irq_desc irq_desc[NR_IRQS] __cacheline_aligned_in_smp = {
 	}
 };
 
+#endif
+
 /*
  * What should we do if we get a hw irq event on an illegal vector?
  * Each architecture has to answer this themself.
@@ -261,17 +419,28 @@ unsigned int __do_IRQ(unsigned int irq)
 
 
 #ifdef CONFIG_TRACE_IRQFLAGS
-/*
- * lockdep: we want to handle all irq_desc locks as a single lock-class:
- */
-static struct lock_class_key irq_desc_lock_class;
-
 void early_init_irq_lock_class(void)
 {
+#ifndef CONFIG_SPARSE_IRQ
 	struct irq_desc *desc;
 	int i;
 
-	for_each_irq_desc(i, desc)
+	for_each_irq_desc(i, desc) {
+		if (!desc)
+			continue;
+
 		lockdep_set_class(&desc->lock, &irq_desc_lock_class);
+	}
+#endif
+}
+#endif
+
+#ifdef CONFIG_SPARSE_IRQ
+unsigned int kstat_irqs_cpu(unsigned int irq, int cpu)
+{
+	struct irq_desc *desc = irq_to_desc(irq);
+	return desc->kstat_irqs[cpu];
 }
 #endif
+EXPORT_SYMBOL(kstat_irqs_cpu);
+

kernel/irq/proc.c

@@ -243,7 +243,11 @@ void init_irq_proc(void)
 	/*
 	 * Create entries for all existing IRQs.
 	 */
-	for_each_irq_desc(irq, desc)
+	for_each_irq_desc(irq, desc) {
+		if (!desc)
+			continue;
+
 		register_irq_proc(irq, desc);
+	}
 }
 

kernel/irq/spurious.c

@@ -91,6 +91,9 @@ static int misrouted_irq(int irq)
 	int i, ok = 0;
 
 	for_each_irq_desc(i, desc) {
+		if (!desc)
+			continue;
+
 		if (!i)
 			 continue;
 
@@ -112,6 +115,8 @@ static void poll_spurious_irqs(unsigned long dummy)
 	for_each_irq_desc(i, desc) {
 		unsigned int status;
 
+		if (!desc)
+			continue;
 		if (!i)
 			 continue;