Merge branch 'x86-mrst-for-linus' of...

Merge branch 'x86-mrst-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip

* 'x86-mrst-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (30 commits)
  x86, mrst: Fix whitespace breakage in apb_timer.c
  x86, mrst: Fix APB timer per cpu clockevent
  x86, mrst: Remove X86_MRST dependency on PCI_IOAPIC
  x86, olpc: Use pci subarch init for OLPC
  x86, pci: Add arch_init to x86_init abstraction
  x86, mrst: Add Kconfig dependencies for Moorestown
  x86, pci: Exclude Moorestown PCI code if CONFIG_X86_MRST=n
  x86, numaq: Make CONFIG_X86_NUMAQ depend on CONFIG_PCI
  x86, pci: Add sanity check for PCI fixed bar probing
  x86, legacy_irq: Remove duplicate vector assigment
  x86, legacy_irq: Remove left over nr_legacy_irqs
  x86, mrst: Platform clock setup code
  x86, apbt: Moorestown APB system timer driver
  x86, mrst: Add vrtc platform data setup code
  x86, mrst: Add platform timer info parsing code
  x86, mrst: Fill in PCI functions in x86_init layer
  x86, mrst: Add dummy legacy pic to platform setup
  x86/PCI: Moorestown PCI support
  x86, ioapic: Add dummy ioapic functions
  x86, ioapic: Early enable ioapic for timer irq
  ...

Fixed up semantic conflict of new clocksources due to commit
17622339 ("clocksource: add argument to resume callback").

Documentation/kernel-parameters.txt

@@ -2834,6 +2834,12 @@ and is between 256 and 4096 characters. It is defined in the file
 			default x2apic cluster mode on platforms
 			supporting x2apic.
 
+	x86_mrst_timer= [X86-32,APBT]
+			Choose timer option for x86 Moorestown MID platform.
+			Two valid options are apbt timer only and lapic timer
+			plus one apbt timer for broadcast timer.
+			x86_mrst_timer=apbt_only | lapic_and_apbt
+
 	xd=		[HW,XT] Original XT pre-IDE (RLL encoded) disks.
 	xd_geo=		See header of drivers/block/xd.c.
 

arch/x86/Kconfig

@@ -393,8 +393,12 @@ config X86_ELAN
 
 config X86_MRST
        bool "Moorestown MID platform"
+	depends on PCI
+	depends on PCI_GOANY
 	depends on X86_32
 	depends on X86_EXTENDED_PLATFORM
+	depends on X86_IO_APIC
+	select APB_TIMER
 	---help---
 	  Moorestown is Intel's Low Power Intel Architecture (LPIA) based Moblin
 	  Internet Device(MID) platform. Moorestown consists of two chips:
@@ -429,6 +433,7 @@ config X86_32_NON_STANDARD
 config X86_NUMAQ
 	bool "NUMAQ (IBM/Sequent)"
 	depends on X86_32_NON_STANDARD
+	depends on PCI
 	select NUMA
 	select X86_MPPARSE
 	---help---
@@ -629,6 +634,16 @@ config HPET_EMULATE_RTC
 	def_bool y
 	depends on HPET_TIMER && (RTC=y || RTC=m || RTC_DRV_CMOS=m || RTC_DRV_CMOS=y)
 
+config APB_TIMER
+       def_bool y if MRST
+       prompt "Langwell APB Timer Support" if X86_MRST
+       help
+         APB timer is the replacement for 8254, HPET on X86 MID platforms.
+         The APBT provides a stable time base on SMP
+         systems, unlike the TSC, but it is more expensive to access,
+         as it is off-chip. APB timers are always running regardless of CPU
+         C states, they are used as per CPU clockevent device when possible.
+
 # Mark as embedded because too many people got it wrong.
 # The code disables itself when not needed.
 config DMI

arch/x86/include/asm/apb_timer.h

+/*
+ * apb_timer.h: Driver for Langwell APB timer based on Synopsis DesignWare
+ *
+ * (C) Copyright 2009 Intel Corporation
+ * Author: Jacob Pan (jacob.jun.pan@intel.com)
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; version 2
+ * of the License.
+ *
+ * Note:
+ */
+
+#ifndef ASM_X86_APBT_H
+#define ASM_X86_APBT_H
+#include <linux/sfi.h>
+
+#ifdef CONFIG_APB_TIMER
+
+/* Langwell DW APB timer registers */
+#define APBTMR_N_LOAD_COUNT    0x00
+#define APBTMR_N_CURRENT_VALUE 0x04
+#define APBTMR_N_CONTROL       0x08
+#define APBTMR_N_EOI           0x0c
+#define APBTMR_N_INT_STATUS    0x10
+
+#define APBTMRS_INT_STATUS     0xa0
+#define APBTMRS_EOI            0xa4
+#define APBTMRS_RAW_INT_STATUS 0xa8
+#define APBTMRS_COMP_VERSION   0xac
+#define APBTMRS_REG_SIZE       0x14
+
+/* register bits */
+#define APBTMR_CONTROL_ENABLE  (1<<0)
+#define APBTMR_CONTROL_MODE_PERIODIC   (1<<1) /*1: periodic 0:free running */
+#define APBTMR_CONTROL_INT     (1<<2)
+
+/* default memory mapped register base */
+#define LNW_SCU_ADDR           0xFF100000
+#define LNW_EXT_TIMER_OFFSET   0x1B800
+#define APBT_DEFAULT_BASE      (LNW_SCU_ADDR+LNW_EXT_TIMER_OFFSET)
+#define LNW_EXT_TIMER_PGOFFSET         0x800
+
+/* APBT clock speed range from PCLK to fabric base, 25-100MHz */
+#define APBT_MAX_FREQ          50
+#define APBT_MIN_FREQ          1
+#define APBT_MMAP_SIZE         1024
+
+#define APBT_DEV_USED  1
+
+extern void apbt_time_init(void);
+extern struct clock_event_device *global_clock_event;
+extern unsigned long apbt_quick_calibrate(void);
+extern int arch_setup_apbt_irqs(int irq, int trigger, int mask, int cpu);
+extern void apbt_setup_secondary_clock(void);
+extern unsigned int boot_cpu_id;
+extern int disable_apbt_percpu;
+
+extern struct sfi_timer_table_entry *sfi_get_mtmr(int hint);
+extern void sfi_free_mtmr(struct sfi_timer_table_entry *mtmr);
+extern int sfi_mtimer_num;
+
+#else /* CONFIG_APB_TIMER */
+
+static inline unsigned long apbt_quick_calibrate(void) {return 0; }
+static inline void apbt_time_init(void) {return 0; }
+
+#endif
+#endif /* ASM_X86_APBT_H */

arch/x86/include/asm/hw_irq.h

@@ -53,13 +53,6 @@ extern void threshold_interrupt(void);
 extern void call_function_interrupt(void);
 extern void call_function_single_interrupt(void);
 
-/* PIC specific functions */
-extern void disable_8259A_irq(unsigned int irq);
-extern void enable_8259A_irq(unsigned int irq);
-extern int i8259A_irq_pending(unsigned int irq);
-extern void make_8259A_irq(unsigned int irq);
-extern void init_8259A(int aeoi);
-
 /* IOAPIC */
 #define IO_APIC_IRQ(x) (((x) >= NR_IRQS_LEGACY) || ((1<<(x)) & io_apic_irqs))
 extern unsigned long io_apic_irqs;

arch/x86/include/asm/i8259.h

@@ -26,11 +26,6 @@ extern unsigned int cached_irq_mask;
 
 extern raw_spinlock_t i8259A_lock;
 
-extern void init_8259A(int auto_eoi);
-extern void enable_8259A_irq(unsigned int irq);
-extern void disable_8259A_irq(unsigned int irq);
-extern unsigned int startup_8259A_irq(unsigned int irq);
-
 /* the PIC may need a careful delay on some platforms, hence specific calls */
 static inline unsigned char inb_pic(unsigned int port)
 {
@@ -57,7 +52,17 @@ static inline void outb_pic(unsigned char value, unsigned int port)
 
 extern struct irq_chip i8259A_chip;
 
-extern void mask_8259A(void);
-extern void unmask_8259A(void);
+struct legacy_pic {
+	int nr_legacy_irqs;
+	struct irq_chip *chip;
+	void (*mask_all)(void);
+	void (*restore_mask)(void);
+	void (*init)(int auto_eoi);
+	int (*irq_pending)(unsigned int irq);
+	void (*make_irq)(unsigned int irq);
+};
+
+extern struct legacy_pic *legacy_pic;
+extern struct legacy_pic null_legacy_pic;
 
 #endif /* _ASM_X86_I8259_H */

arch/x86/include/asm/io_apic.h

@@ -143,8 +143,6 @@ extern int noioapicreroute;
 /* 1 if the timer IRQ uses the '8259A Virtual Wire' mode */
 extern int timer_through_8259;
 
-extern void io_apic_disable_legacy(void);
-
 /*
  * If we use the IO-APIC for IRQ routing, disable automatic
  * assignment of PCI IRQ's.
@@ -189,6 +187,7 @@ extern struct mp_ioapic_gsi  mp_gsi_routing[];
 int mp_find_ioapic(int gsi);
 int mp_find_ioapic_pin(int ioapic, int gsi);
 void __init mp_register_ioapic(int id, u32 address, u32 gsi_base);
+extern void __init pre_init_apic_IRQ0(void);
 
 #else  /* !CONFIG_X86_IO_APIC */
 
@@ -198,7 +197,11 @@ static const int timer_through_8259 = 0;
 static inline void ioapic_init_mappings(void)	{ }
 static inline void ioapic_insert_resources(void) { }
 static inline void probe_nr_irqs_gsi(void)	{ }
+static inline int mp_find_ioapic(int gsi) { return 0; }
 
+struct io_apic_irq_attr;
+static inline int io_apic_set_pci_routing(struct device *dev, int irq,
+		 struct io_apic_irq_attr *irq_attr) { return 0; }
 #endif
 
 #endif /* _ASM_X86_IO_APIC_H */

arch/x86/include/asm/irq.h

@@ -48,6 +48,5 @@ extern DECLARE_BITMAP(used_vectors, NR_VECTORS);
 extern int vector_used_by_percpu_irq(unsigned int vector);
 
 extern void init_ISA_irqs(void);
-extern int nr_legacy_irqs;
 
 #endif /* _ASM_X86_IRQ_H */

arch/x86/include/asm/mrst.h

+/*
+ * mrst.h: Intel Moorestown platform specific setup code
+ *
+ * (C) Copyright 2009 Intel Corporation
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; version 2
+ * of the License.
+ */
+#ifndef _ASM_X86_MRST_H
+#define _ASM_X86_MRST_H
+extern int pci_mrst_init(void);
+int __init sfi_parse_mrtc(struct sfi_table_header *table);
+
+#define SFI_MTMR_MAX_NUM 8
+#define SFI_MRTC_MAX	8
+
+#endif /* _ASM_X86_MRST_H */

arch/x86/include/asm/numaq.h

@@ -30,6 +30,7 @@
 
 extern int found_numaq;
 extern int get_memcfg_numaq(void);
+extern int pci_numaq_init(void);
 
 extern void *xquad_portio;
 

arch/x86/include/asm/olpc.h

@@ -13,7 +13,6 @@ struct olpc_platform_t {
 
 #define OLPC_F_PRESENT		0x01
 #define OLPC_F_DCON		0x02
-#define OLPC_F_VSA		0x04
 
 #ifdef CONFIG_OLPC
 
@@ -50,18 +49,6 @@ static inline int olpc_has_dcon(void)
 	return (olpc_platform_info.flags & OLPC_F_DCON) ? 1 : 0;
 }
 
-/*
- * The VSA is software from AMD that typical Geode bioses will include.
- * It is used to emulate the PCI bus, VGA, etc.  OLPC's Open Firmware does
- * not include the VSA; instead, PCI is emulated by the kernel.
- *
- * The VSA is described further in arch/x86/pci/olpc.c.
- */
-static inline int olpc_has_vsa(void)
-{
-	return (olpc_platform_info.flags & OLPC_F_VSA) ? 1 : 0;
-}
-
 /*
  * The "Mass Production" version of OLPC's XO is identified as being model
  * C2.  During the prototype phase, the following models (in chronological
@@ -87,13 +74,10 @@ static inline int olpc_has_dcon(void)
 	return 0;
 }
 
-static inline int olpc_has_vsa(void)
-{
-	return 0;
-}
-
 #endif
 
+extern int pci_olpc_init(void);
+
 /* EC related functions */
 
 extern int olpc_ec_cmd(unsigned char cmd, unsigned char *inbuf, size_t inlen,

arch/x86/include/asm/pci.h

@@ -45,8 +45,15 @@ static inline int pci_proc_domain(struct pci_bus *bus)
 
 #ifdef CONFIG_PCI
 extern unsigned int pcibios_assign_all_busses(void);
+extern int pci_legacy_init(void);
+# ifdef CONFIG_ACPI
+#  define x86_default_pci_init pci_acpi_init
+# else
+#  define x86_default_pci_init pci_legacy_init
+# endif
 #else
-#define pcibios_assign_all_busses()	0
+# define pcibios_assign_all_busses()	0
+# define x86_default_pci_init		NULL
 #endif
 
 extern unsigned long pci_mem_start;

arch/x86/include/asm/pci_x86.h

@@ -83,7 +83,6 @@ struct irq_routing_table {
 
 extern unsigned int pcibios_irq_mask;
 
-extern int pcibios_scanned;
 extern spinlock_t pci_config_lock;
 
 extern int (*pcibios_enable_irq)(struct pci_dev *dev);
@@ -106,16 +105,15 @@ extern bool port_cf9_safe;
 extern int pci_direct_probe(void);
 extern void pci_direct_init(int type);
 extern void pci_pcbios_init(void);
-extern int pci_olpc_init(void);
 extern void __init dmi_check_pciprobe(void);
 extern void __init dmi_check_skip_isa_align(void);
 
 /* some common used subsys_initcalls */
 extern int __init pci_acpi_init(void);
-extern int __init pcibios_irq_init(void);
-extern int __init pci_visws_init(void);
-extern int __init pci_numaq_init(void);
+extern void __init pcibios_irq_init(void);
 extern int __init pcibios_init(void);
+extern int pci_legacy_init(void);
+extern void pcibios_fixup_irqs(void);
 
 /* pci-mmconfig.c */
 
@@ -183,3 +181,17 @@ static inline void mmio_config_writel(void __iomem *pos, u32 val)
 {
 	asm volatile("movl %%eax,(%1)" : : "a" (val), "r" (pos) : "memory");
 }
+
+#ifdef CONFIG_PCI
+# ifdef CONFIG_ACPI
+#  define x86_default_pci_init		pci_acpi_init
+# else
+#  define x86_default_pci_init		pci_legacy_init
+# endif
+# define x86_default_pci_init_irq	pcibios_irq_init
+# define x86_default_pci_fixup_irqs	pcibios_fixup_irqs
+#else
+# define x86_default_pci_init		NULL
+# define x86_default_pci_init_irq	NULL
+# define x86_default_pci_fixup_irqs	NULL
+#endif

arch/x86/include/asm/setup.h

@@ -37,10 +37,8 @@ void setup_bios_corruption_check(void);
 
 #ifdef CONFIG_X86_VISWS
 extern void visws_early_detect(void);
-extern int is_visws_box(void);
 #else
 static inline void visws_early_detect(void) { }
-static inline int is_visws_box(void) { return 0; }
 #endif
 
 extern unsigned long saved_video_mode;

arch/x86/include/asm/visws/cobalt.h

@@ -122,4 +122,6 @@ extern char visws_board_type;
 
 extern char visws_board_rev;
 
+extern int pci_visws_init(void);
+
 #endif /* _ASM_X86_VISWS_COBALT_H */

arch/x86/include/asm/x86_init.h

@@ -98,6 +98,20 @@ struct x86_init_iommu {
 	int (*iommu_init)(void);
 };
 
+/**
+ * struct x86_init_pci - platform specific pci init functions
+ * @arch_init:			platform specific pci arch init call
+ * @init:			platform specific pci subsystem init
+ * @init_irq:			platform specific pci irq init
+ * @fixup_irqs:			platform specific pci irq fixup
+ */
+struct x86_init_pci {
+	int (*arch_init)(void);
+	int (*init)(void);
+	void (*init_irq)(void);
+	void (*fixup_irqs)(void);
+};
+
 /**
  * struct x86_init_ops - functions for platform specific setup
  *
@@ -110,6 +124,7 @@ struct x86_init_ops {
 	struct x86_init_paging		paging;
 	struct x86_init_timers		timers;
 	struct x86_init_iommu		iommu;
+	struct x86_init_pci		pci;
 };
 
 /**

arch/x86/kernel/Makefile

@@ -87,6 +87,7 @@ obj-$(CONFIG_VM86)		+= vm86_32.o
 obj-$(CONFIG_EARLY_PRINTK)	+= early_printk.o
 
 obj-$(CONFIG_HPET_TIMER) 	+= hpet.o
+obj-$(CONFIG_APB_TIMER)		+= apb_timer.o
 
 obj-$(CONFIG_K8_NB)		+= k8.o
 obj-$(CONFIG_DEBUG_RODATA_TEST)	+= test_rodata.o

arch/x86/kernel/acpi/boot.c

@@ -35,6 +35,7 @@
 #include <linux/ioport.h>
 #include <linux/pci.h>
 
+#include <asm/pci_x86.h>
 #include <asm/pgtable.h>
 #include <asm/io_apic.h>
 #include <asm/apic.h>
@@ -1624,6 +1625,9 @@ int __init acpi_boot_init(void)
 
 	acpi_table_parse(ACPI_SIG_HPET, acpi_parse_hpet);
 
+	if (!acpi_noirq)
+		x86_init.pci.init = pci_acpi_init;
+
 	return 0;
 }
 

arch/x86/kernel/apb_timer.c

+/*
+ * apb_timer.c: Driver for Langwell APB timers
+ *
+ * (C) Copyright 2009 Intel Corporation
+ * Author: Jacob Pan (jacob.jun.pan@intel.com)
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; version 2
+ * of the License.
+ *
+ * Note:
+ * Langwell is the south complex of Intel Moorestown MID platform. There are
+ * eight external timers in total that can be used by the operating system.
+ * The timer information, such as frequency and addresses, is provided to the
+ * OS via SFI tables.
+ * Timer interrupts are routed via FW/HW emulated IOAPIC independently via
+ * individual redirection table entries (RTE).
+ * Unlike HPET, there is no master counter, therefore one of the timers are
+ * used as clocksource. The overall allocation looks like:
+ *  - timer 0 - NR_CPUs for per cpu timer
+ *  - one timer for clocksource
+ *  - one timer for watchdog driver.
+ * It is also worth notice that APB timer does not support true one-shot mode,
+ * free-running mode will be used here to emulate one-shot mode.
+ * APB timer can also be used as broadcast timer along with per cpu local APIC
+ * timer, but by default APB timer has higher rating than local APIC timers.
+ */
+
+#include <linux/clocksource.h>
+#include <linux/clockchips.h>
+#include <linux/delay.h>
+#include <linux/errno.h>
+#include <linux/init.h>
+#include <linux/sysdev.h>
+#include <linux/pm.h>
+#include <linux/pci.h>
+#include <linux/sfi.h>
+#include <linux/interrupt.h>
+#include <linux/cpu.h>
+#include <linux/irq.h>
+
+#include <asm/fixmap.h>
+#include <asm/apb_timer.h>
+
+#define APBT_MASK			CLOCKSOURCE_MASK(32)
+#define APBT_SHIFT			22
+#define APBT_CLOCKEVENT_RATING		150
+#define APBT_CLOCKSOURCE_RATING		250
+#define APBT_MIN_DELTA_USEC		200
+
+#define EVT_TO_APBT_DEV(evt) container_of(evt, struct apbt_dev, evt)
+#define APBT_CLOCKEVENT0_NUM   (0)
+#define APBT_CLOCKEVENT1_NUM   (1)
+#define APBT_CLOCKSOURCE_NUM   (2)
+
+static unsigned long apbt_address;
+static int apb_timer_block_enabled;
+static void __iomem *apbt_virt_address;
+static int phy_cs_timer_id;
+
+/*
+ * Common DW APB timer info
+ */
+static uint64_t apbt_freq;
+
+static void apbt_set_mode(enum clock_event_mode mode,
+			  struct clock_event_device *evt);
+static int apbt_next_event(unsigned long delta,
+			   struct clock_event_device *evt);
+static cycle_t apbt_read_clocksource(struct clocksource *cs);
+static void apbt_restart_clocksource(struct clocksource *cs);
+
+struct apbt_dev {
+	struct clock_event_device evt;
+	unsigned int num;
+	int cpu;
+	unsigned int irq;
+	unsigned int tick;
+	unsigned int count;
+	unsigned int flags;
+	char name[10];
+};
+
+int disable_apbt_percpu __cpuinitdata;
+
+static DEFINE_PER_CPU(struct apbt_dev, cpu_apbt_dev);
+
+#ifdef CONFIG_SMP
+static unsigned int apbt_num_timers_used;
+static struct apbt_dev *apbt_devs;
+#endif
+
+static	inline unsigned long apbt_readl_reg(unsigned long a)
+{
+	return readl(apbt_virt_address + a);
+}
+
+static inline void apbt_writel_reg(unsigned long d, unsigned long a)
+{
+	writel(d, apbt_virt_address + a);
+}
+
+static inline unsigned long apbt_readl(int n, unsigned long a)
+{
+	return readl(apbt_virt_address + a + n * APBTMRS_REG_SIZE);
+}
+
+static inline void apbt_writel(int n, unsigned long d, unsigned long a)
+{
+	writel(d, apbt_virt_address + a + n * APBTMRS_REG_SIZE);
+}
+
+static inline void apbt_set_mapping(void)
+{
+	struct sfi_timer_table_entry *mtmr;
+
+	if (apbt_virt_address) {
+		pr_debug("APBT base already mapped\n");
+		return;
+	}
+	mtmr = sfi_get_mtmr(APBT_CLOCKEVENT0_NUM);
+	if (mtmr == NULL) {
+		printk(KERN_ERR "Failed to get MTMR %d from SFI\n",
+		       APBT_CLOCKEVENT0_NUM);
+		return;
+	}
+	apbt_address = (unsigned long)mtmr->phys_addr;
+	if (!apbt_address) {
+		printk(KERN_WARNING "No timer base from SFI, use default\n");
+		apbt_address = APBT_DEFAULT_BASE;
+	}
+	apbt_virt_address = ioremap_nocache(apbt_address, APBT_MMAP_SIZE);
+	if (apbt_virt_address) {
+		pr_debug("Mapped APBT physical addr %p at virtual addr %p\n",\
+			 (void *)apbt_address, (void *)apbt_virt_address);
+	} else {
+		pr_debug("Failed mapping APBT phy address at %p\n",\
+			 (void *)apbt_address);
+		goto panic_noapbt;
+	}
+	apbt_freq = mtmr->freq_hz / USEC_PER_SEC;
+	sfi_free_mtmr(mtmr);
+
+	/* Now figure out the physical timer id for clocksource device */
+	mtmr = sfi_get_mtmr(APBT_CLOCKSOURCE_NUM);
+	if (mtmr == NULL)
+		goto panic_noapbt;
+
+	/* Now figure out the physical timer id */
+	phy_cs_timer_id = (unsigned int)(mtmr->phys_addr & 0xff)
+		/ APBTMRS_REG_SIZE;
+	pr_debug("Use timer %d for clocksource\n", phy_cs_timer_id);
+	return;
+
+panic_noapbt:
+	panic("Failed to setup APB system timer\n");
+
+}
+
+static inline void apbt_clear_mapping(void)
+{
+	iounmap(apbt_virt_address);
+	apbt_virt_address = NULL;
+}
+
+/*
+ * APBT timer interrupt enable / disable
+ */
+static inline int is_apbt_capable(void)
+{
+	return apbt_virt_address ? 1 : 0;
+}
+
+static struct clocksource clocksource_apbt = {
+	.name		= "apbt",
+	.rating		= APBT_CLOCKSOURCE_RATING,
+	.read		= apbt_read_clocksource,
+	.mask		= APBT_MASK,
+	.shift		= APBT_SHIFT,
+	.flags		= CLOCK_SOURCE_IS_CONTINUOUS,
+	.resume		= apbt_restart_clocksource,
+};
+
+/* boot APB clock event device */
+static struct clock_event_device apbt_clockevent = {
+	.name		= "apbt0",
+	.features	= CLOCK_EVT_FEAT_PERIODIC | CLOCK_EVT_FEAT_ONESHOT,
+	.set_mode	= apbt_set_mode,
+	.set_next_event = apbt_next_event,
+	.shift		= APBT_SHIFT,
+	.irq		= 0,
+	.rating		= APBT_CLOCKEVENT_RATING,
+};
+
+/*
+ * if user does not want to use per CPU apb timer, just give it a lower rating
+ * than local apic timer and skip the late per cpu timer init.
+ */
+static inline int __init setup_x86_mrst_timer(char *arg)
+{
+	if (!arg)
+		return -EINVAL;
+
+	if (strcmp("apbt_only", arg) == 0)
+		disable_apbt_percpu = 0;
+	else if (strcmp("lapic_and_apbt", arg) == 0)
+		disable_apbt_percpu = 1;
+	else {
+		pr_warning("X86 MRST timer option %s not recognised"
+			   " use x86_mrst_timer=apbt_only or lapic_and_apbt\n",
+			   arg);
+		return -EINVAL;
+	}
+	return 0;
+}
+__setup("x86_mrst_timer=", setup_x86_mrst_timer);
+
+/*
+ * start count down from 0xffff_ffff. this is done by toggling the enable bit
+ * then load initial load count to ~0.
+ */
+static void apbt_start_counter(int n)
+{
+	unsigned long ctrl = apbt_readl(n, APBTMR_N_CONTROL);
+
+	ctrl &= ~APBTMR_CONTROL_ENABLE;
+	apbt_writel(n, ctrl, APBTMR_N_CONTROL);
+	apbt_writel(n, ~0, APBTMR_N_LOAD_COUNT);
+	/* enable, mask interrupt */
+	ctrl &= ~APBTMR_CONTROL_MODE_PERIODIC;
+	ctrl |= (APBTMR_CONTROL_ENABLE | APBTMR_CONTROL_INT);
+	apbt_writel(n, ctrl, APBTMR_N_CONTROL);
+	/* read it once to get cached counter value initialized */
+	apbt_read_clocksource(&clocksource_apbt);
+}
+
+static irqreturn_t apbt_interrupt_handler(int irq, void *data)
+{
+	struct apbt_dev *dev = (struct apbt_dev *)data;
+	struct clock_event_device *aevt = &dev->evt;
+
+	if (!aevt->event_handler) {
+		printk(KERN_INFO "Spurious APBT timer interrupt on %d\n",
+		       dev->num);
+		return IRQ_NONE;
+	}
+	aevt->event_handler(aevt);
+	return IRQ_HANDLED;
+}
+
+static void apbt_restart_clocksource(struct clocksource *cs)
+{
+	apbt_start_counter(phy_cs_timer_id);
+}
+
+/* Setup IRQ routing via IOAPIC */
+#ifdef CONFIG_SMP
+static void apbt_setup_irq(struct apbt_dev *adev)
+{
+	struct irq_chip *chip;
+	struct irq_desc *desc;
+
+	/* timer0 irq has been setup early */
+	if (adev->irq == 0)
+		return;
+	desc = irq_to_desc(adev->irq);
+	chip = get_irq_chip(adev->irq);
+	disable_irq(adev->irq);
+	desc->status |= IRQ_MOVE_PCNTXT;
+	irq_set_affinity(adev->irq, cpumask_of(adev->cpu));
+	/* APB timer irqs are set up as mp_irqs, timer is edge triggerred */
+	set_irq_chip_and_handler_name(adev->irq, chip, handle_edge_irq, "edge");
+	enable_irq(adev->irq);
+	if (system_state == SYSTEM_BOOTING)
+		if (request_irq(adev->irq, apbt_interrupt_handler,
+				IRQF_TIMER | IRQF_DISABLED | IRQF_NOBALANCING,
+				adev->name, adev)) {
+			printk(KERN_ERR "Failed request IRQ for APBT%d\n",
+			       adev->num);
+		}
+}
+#endif
+
+static void apbt_enable_int(int n)
+{
+	unsigned long ctrl = apbt_readl(n, APBTMR_N_CONTROL);
+	/* clear pending intr */
+	apbt_readl(n, APBTMR_N_EOI);
+	ctrl &= ~APBTMR_CONTROL_INT;
+	apbt_writel(n, ctrl, APBTMR_N_CONTROL);
+}
+
+static void apbt_disable_int(int n)
+{
+	unsigned long ctrl = apbt_readl(n, APBTMR_N_CONTROL);
+
+	ctrl |= APBTMR_CONTROL_INT;
+	apbt_writel(n, ctrl, APBTMR_N_CONTROL);
+}
+
+
+static int __init apbt_clockevent_register(void)
+{
+	struct sfi_timer_table_entry *mtmr;
+	struct apbt_dev *adev = &__get_cpu_var(cpu_apbt_dev);
+
+	mtmr = sfi_get_mtmr(APBT_CLOCKEVENT0_NUM);
+	if (mtmr == NULL) {
+		printk(KERN_ERR "Failed to get MTMR %d from SFI\n",
+		       APBT_CLOCKEVENT0_NUM);
+		return -ENODEV;
+	}
+
+	/*
+	 * We need to calculate the scaled math multiplication factor for
+	 * nanosecond to apbt tick conversion.
+	 * mult = (nsec/cycle)*2^APBT_SHIFT
+	 */
+	apbt_clockevent.mult = div_sc((unsigned long) mtmr->freq_hz
+				      , NSEC_PER_SEC, APBT_SHIFT);
+
+	/* Calculate the min / max delta */
+	apbt_clockevent.max_delta_ns = clockevent_delta2ns(0x7FFFFFFF,
+							   &apbt_clockevent);
+	apbt_clockevent.min_delta_ns = clockevent_delta2ns(
+		APBT_MIN_DELTA_USEC*apbt_freq,
+		&apbt_clockevent);
+	/*
+	 * Start apbt with the boot cpu mask and make it
+	 * global if not used for per cpu timer.
+	 */
+	apbt_clockevent.cpumask = cpumask_of(smp_processor_id());
+	adev->num = smp_processor_id();
+	memcpy(&adev->evt, &apbt_clockevent, sizeof(struct clock_event_device));
+
+	if (disable_apbt_percpu) {
+		apbt_clockevent.rating = APBT_CLOCKEVENT_RATING - 100;
+		global_clock_event = &adev->evt;
+		printk(KERN_DEBUG "%s clockevent registered as global\n",
+		       global_clock_event->name);
+	}
+
+	if (request_irq(apbt_clockevent.irq, apbt_interrupt_handler,
+			IRQF_TIMER | IRQF_DISABLED | IRQF_NOBALANCING,
+			apbt_clockevent.name, adev)) {
+		printk(KERN_ERR "Failed request IRQ for APBT%d\n",
+		       apbt_clockevent.irq);
+	}
+
+	clockevents_register_device(&adev->evt);
+	/* Start APBT 0 interrupts */
+	apbt_enable_int(APBT_CLOCKEVENT0_NUM);
+
+	sfi_free_mtmr(mtmr);
+	return 0;
+}
+
+#ifdef CONFIG_SMP
+/* Should be called with per cpu */
+void apbt_setup_secondary_clock(void)
+{
+	struct apbt_dev *adev;
+	struct clock_event_device *aevt;
+	int cpu;
+
+	/* Don't register boot CPU clockevent */
+	cpu = smp_processor_id();
+	if (cpu == boot_cpu_id)
+		return;
+	/*
+	 * We need to calculate the scaled math multiplication factor for
+	 * nanosecond to apbt tick conversion.
+	 * mult = (nsec/cycle)*2^APBT_SHIFT
+	 */
+	printk(KERN_INFO "Init per CPU clockevent %d\n", cpu);
+	adev = &per_cpu(cpu_apbt_dev, cpu);
+	aevt = &adev->evt;
+
+	memcpy(aevt, &apbt_clockevent, sizeof(*aevt));
+	aevt->cpumask = cpumask_of(cpu);
+	aevt->name = adev->name;
+	aevt->mode = CLOCK_EVT_MODE_UNUSED;
+
+	printk(KERN_INFO "Registering CPU %d clockevent device %s, mask %08x\n",
+	       cpu, aevt->name, *(u32 *)aevt->cpumask);
+
+	apbt_setup_irq(adev);
+
+	clockevents_register_device(aevt);
+
+	apbt_enable_int(cpu);
+
+	return;
+}
+
+/*
+ * this notify handler process CPU hotplug events. in case of S0i3, nonboot
+ * cpus are disabled/enabled frequently, for performance reasons, we keep the
+ * per cpu timer irq registered so that we do need to do free_irq/request_irq.
+ *
+ * TODO: it might be more reliable to directly disable percpu clockevent device
+ * without the notifier chain. currently, cpu 0 may get interrupts from other
+ * cpu timers during the offline process due to the ordering of notification.
+ * the extra interrupt is harmless.
+ */
+static int apbt_cpuhp_notify(struct notifier_block *n,
+			     unsigned long action, void *hcpu)
+{
+	unsigned long cpu = (unsigned long)hcpu;
+	struct apbt_dev *adev = &per_cpu(cpu_apbt_dev, cpu);
+
+	switch (action & 0xf) {
+	case CPU_DEAD:
+		apbt_disable_int(cpu);
+		if (system_state == SYSTEM_RUNNING)
+			pr_debug("skipping APBT CPU %lu offline\n", cpu);
+		else if (adev) {
+			pr_debug("APBT clockevent for cpu %lu offline\n", cpu);
+			free_irq(adev->irq, adev);
+		}
+		break;
+	default:
+		pr_debug(KERN_INFO "APBT notified %lu, no action\n", action);
+	}
+	return NOTIFY_OK;
+}
+
+static __init int apbt_late_init(void)
+{
+	if (disable_apbt_percpu)
+		return 0;
+	/* This notifier should be called after workqueue is ready */
+	hotcpu_notifier(apbt_cpuhp_notify, -20);
+	return 0;
+}
+fs_initcall(apbt_late_init);
+#else
+
+void apbt_setup_secondary_clock(void) {}
+
+#endif /* CONFIG_SMP */
+
+static void apbt_set_mode(enum clock_event_mode mode,
+			  struct clock_event_device *evt)
+{
+	unsigned long ctrl;
+	uint64_t delta;
+	int timer_num;
+	struct apbt_dev *adev = EVT_TO_APBT_DEV(evt);
+
+	timer_num = adev->num;
+	pr_debug("%s CPU %d timer %d mode=%d\n",
+		 __func__, first_cpu(*evt->cpumask), timer_num, mode);
+
+	switch (mode) {
+	case CLOCK_EVT_MODE_PERIODIC:
+		delta = ((uint64_t)(NSEC_PER_SEC/HZ)) * apbt_clockevent.mult;
+		delta >>= apbt_clockevent.shift;
+		ctrl = apbt_readl(timer_num, APBTMR_N_CONTROL);
+		ctrl |= APBTMR_CONTROL_MODE_PERIODIC;
+		apbt_writel(timer_num, ctrl, APBTMR_N_CONTROL);
+		/*
+		 * DW APB p. 46, have to disable timer before load counter,
+		 * may cause sync problem.
+		 */
+		ctrl &= ~APBTMR_CONTROL_ENABLE;
+		apbt_writel(timer_num, ctrl, APBTMR_N_CONTROL);
+		udelay(1);
+		pr_debug("Setting clock period %d for HZ %d\n", (int)delta, HZ);
+		apbt_writel(timer_num, delta, APBTMR_N_LOAD_COUNT);
+		ctrl |= APBTMR_CONTROL_ENABLE;
+		apbt_writel(timer_num, ctrl, APBTMR_N_CONTROL);
+		break;
+		/* APB timer does not have one-shot mode, use free running mode */
+	case CLOCK_EVT_MODE_ONESHOT:
+		ctrl = apbt_readl(timer_num, APBTMR_N_CONTROL);
+		/*
+		 * set free running mode, this mode will let timer reload max
+		 * timeout which will give time (3min on 25MHz clock) to rearm
+		 * the next event, therefore emulate the one-shot mode.
+		 */
+		ctrl &= ~APBTMR_CONTROL_ENABLE;
+		ctrl &= ~APBTMR_CONTROL_MODE_PERIODIC;
+
+		apbt_writel(timer_num, ctrl, APBTMR_N_CONTROL);
+		/* write again to set free running mode */
+		apbt_writel(timer_num, ctrl, APBTMR_N_CONTROL);
+
+		/*
+		 * DW APB p. 46, load counter with all 1s before starting free
+		 * running mode.
+		 */
+		apbt_writel(timer_num, ~0, APBTMR_N_LOAD_COUNT);
+		ctrl &= ~APBTMR_CONTROL_INT;
+		ctrl |= APBTMR_CONTROL_ENABLE;
+		apbt_writel(timer_num, ctrl, APBTMR_N_CONTROL);
+		break;
+
+	case CLOCK_EVT_MODE_UNUSED:
+	case CLOCK_EVT_MODE_SHUTDOWN:
+		apbt_disable_int(timer_num);
+		ctrl = apbt_readl(timer_num, APBTMR_N_CONTROL);
+		ctrl &= ~APBTMR_CONTROL_ENABLE;
+		apbt_writel(timer_num, ctrl, APBTMR_N_CONTROL);
+		break;
+
+	case CLOCK_EVT_MODE_RESUME:
+		apbt_enable_int(timer_num);
+		break;
+	}
+}
+
+static int apbt_next_event(unsigned long delta,
+			   struct clock_event_device *evt)
+{
+	unsigned long ctrl;
+	int timer_num;
+
+	struct apbt_dev *adev = EVT_TO_APBT_DEV(evt);
+
+	timer_num = adev->num;
+	/* Disable timer */
+	ctrl = apbt_readl(timer_num, APBTMR_N_CONTROL);
+	ctrl &= ~APBTMR_CONTROL_ENABLE;
+	apbt_writel(timer_num, ctrl, APBTMR_N_CONTROL);
+	/* write new count */
+	apbt_writel(timer_num, delta, APBTMR_N_LOAD_COUNT);
+	ctrl |= APBTMR_CONTROL_ENABLE;
+	apbt_writel(timer_num, ctrl, APBTMR_N_CONTROL);
+	return 0;
+}
+
+/*
+ * APB timer clock is not in sync with pclk on Langwell, which translates to
+ * unreliable read value caused by sampling error. the error does not add up
+ * overtime and only happens when sampling a 0 as a 1 by mistake. so the time
+ * would go backwards. the following code is trying to prevent time traveling
+ * backwards. little bit paranoid.
+ */
+static cycle_t apbt_read_clocksource(struct clocksource *cs)
+{
+	unsigned long t0, t1, t2;
+	static unsigned long last_read;
+
+bad_count:
+	t1 = apbt_readl(phy_cs_timer_id,
+			APBTMR_N_CURRENT_VALUE);
+	t2 = apbt_readl(phy_cs_timer_id,
+			APBTMR_N_CURRENT_VALUE);
+	if (unlikely(t1 < t2)) {
+		pr_debug("APBT: read current count error %lx:%lx:%lx\n",
+			 t1, t2, t2 - t1);
+		goto bad_count;
+	}
+	/*
+	 * check against cached last read, makes sure time does not go back.
+	 * it could be a normal rollover but we will do tripple check anyway
+	 */
+	if (unlikely(t2 > last_read)) {
+		/* check if we have a normal rollover */
+		unsigned long raw_intr_status =
+			apbt_readl_reg(APBTMRS_RAW_INT_STATUS);
+		/*
+		 * cs timer interrupt is masked but raw intr bit is set if
+		 * rollover occurs. then we read EOI reg to clear it.
+		 */
+		if (raw_intr_status & (1 << phy_cs_timer_id)) {
+			apbt_readl(phy_cs_timer_id, APBTMR_N_EOI);
+			goto out;
+		}
+		pr_debug("APB CS going back %lx:%lx:%lx ",
+			 t2, last_read, t2 - last_read);
+bad_count_x3:
+		pr_debug(KERN_INFO "tripple check enforced\n");
+		t0 = apbt_readl(phy_cs_timer_id,
+				APBTMR_N_CURRENT_VALUE);
+		udelay(1);
+		t1 = apbt_readl(phy_cs_timer_id,
+				APBTMR_N_CURRENT_VALUE);
+		udelay(1);
+		t2 = apbt_readl(phy_cs_timer_id,
+				APBTMR_N_CURRENT_VALUE);
+		if ((t2 > t1) || (t1 > t0)) {
+			printk(KERN_ERR "Error: APB CS tripple check failed\n");
+			goto bad_count_x3;
+		}
+	}
+out:
+	last_read = t2;
+	return (cycle_t)~t2;
+}
+
+static int apbt_clocksource_register(void)
+{
+	u64 start, now;
+	cycle_t t1;
+
+	/* Start the counter, use timer 2 as source, timer 0/1 for event */
+	apbt_start_counter(phy_cs_timer_id);
+
+	/* Verify whether apbt counter works */
+	t1 = apbt_read_clocksource(&clocksource_apbt);
+	rdtscll(start);
+
+	/*
+	 * We don't know the TSC frequency yet, but waiting for
+	 * 200000 TSC cycles is safe:
+	 * 4 GHz == 50us
+	 * 1 GHz == 200us
+	 */
+	do {
+		rep_nop();
+		rdtscll(now);
+	} while ((now - start) < 200000UL);
+
+	/* APBT is the only always on clocksource, it has to work! */
+	if (t1 == apbt_read_clocksource(&clocksource_apbt))
+		panic("APBT counter not counting. APBT disabled\n");
+
+	/*
+	 * initialize and register APBT clocksource
+	 * convert that to ns/clock cycle
+	 * mult = (ns/c) * 2^APBT_SHIFT
+	 */
+	clocksource_apbt.mult = div_sc(MSEC_PER_SEC,
+				       (unsigned long) apbt_freq, APBT_SHIFT);
+	clocksource_register(&clocksource_apbt);
+
+	return 0;
+}
+
+/*
+ * Early setup the APBT timer, only use timer 0 for booting then switch to
+ * per CPU timer if possible.
+ * returns 1 if per cpu apbt is setup
+ * returns 0 if no per cpu apbt is chosen
+ * panic if set up failed, this is the only platform timer on Moorestown.
+ */
+void __init apbt_time_init(void)
+{
+#ifdef CONFIG_SMP
+	int i;
+	struct sfi_timer_table_entry *p_mtmr;
+	unsigned int percpu_timer;
+	struct apbt_dev *adev;
+#endif
+
+	if (apb_timer_block_enabled)
+		return;
+	apbt_set_mapping();
+	if (apbt_virt_address) {
+		pr_debug("Found APBT version 0x%lx\n",\
+			 apbt_readl_reg(APBTMRS_COMP_VERSION));
+	} else
+		goto out_noapbt;
+	/*
+	 * Read the frequency and check for a sane value, for ESL model
+	 * we extend the possible clock range to allow time scaling.
+	 */
+
+	if (apbt_freq < APBT_MIN_FREQ || apbt_freq > APBT_MAX_FREQ) {
+		pr_debug("APBT has invalid freq 0x%llx\n", apbt_freq);
+		goto out_noapbt;
+	}
+	if (apbt_clocksource_register()) {
+		pr_debug("APBT has failed to register clocksource\n");
+		goto out_noapbt;
+	}
+	if (!apbt_clockevent_register())
+		apb_timer_block_enabled = 1;
+	else {
+		pr_debug("APBT has failed to register clockevent\n");
+		goto out_noapbt;
+	}
+#ifdef CONFIG_SMP
+	/* kernel cmdline disable apb timer, so we will use lapic timers */
+	if (disable_apbt_percpu) {
+		printk(KERN_INFO "apbt: disabled per cpu timer\n");
+		return;
+	}
+	pr_debug("%s: %d CPUs online\n", __func__, num_online_cpus());
+	if (num_possible_cpus() <= sfi_mtimer_num) {
+		percpu_timer = 1;
+		apbt_num_timers_used = num_possible_cpus();
+	} else {
+		percpu_timer = 0;
+		apbt_num_timers_used = 1;
+		adev = &per_cpu(cpu_apbt_dev, 0);
+		adev->flags &= ~APBT_DEV_USED;
+	}
+	pr_debug("%s: %d APB timers used\n", __func__, apbt_num_timers_used);
+
+	/* here we set up per CPU timer data structure */
+	apbt_devs = kzalloc(sizeof(struct apbt_dev) * apbt_num_timers_used,
+			    GFP_KERNEL);
+	if (!apbt_devs) {
+		printk(KERN_ERR "Failed to allocate APB timer devices\n");
+		return;
+	}
+	for (i = 0; i < apbt_num_timers_used; i++) {
+		adev = &per_cpu(cpu_apbt_dev, i);
+		adev->num = i;
+		adev->cpu = i;
+		p_mtmr = sfi_get_mtmr(i);
+		if (p_mtmr) {
+			adev->tick = p_mtmr->freq_hz;
+			adev->irq = p_mtmr->irq;
+		} else
+			printk(KERN_ERR "Failed to get timer for cpu %d\n", i);
+		adev->count = 0;
+		sprintf(adev->name, "apbt%d", i);
+	}
+#endif
+
+	return;
+
+out_noapbt:
+	apbt_clear_mapping();
+	apb_timer_block_enabled = 0;
+	panic("failed to enable APB timer\n");
+}
+
+static inline void apbt_disable(int n)
+{
+	if (is_apbt_capable()) {
+		unsigned long ctrl =  apbt_readl(n, APBTMR_N_CONTROL);
+		ctrl &= ~APBTMR_CONTROL_ENABLE;
+		apbt_writel(n, ctrl, APBTMR_N_CONTROL);
+	}
+}
+
+/* called before apb_timer_enable, use early map */
+unsigned long apbt_quick_calibrate()
+{
+	int i, scale;
+	u64 old, new;
+	cycle_t t1, t2;
+	unsigned long khz = 0;
+	u32 loop, shift;
+
+	apbt_set_mapping();
+	apbt_start_counter(phy_cs_timer_id);
+
+	/* check if the timer can count down, otherwise return */
+	old = apbt_read_clocksource(&clocksource_apbt);
+	i = 10000;
+	while (--i) {
+		if (old != apbt_read_clocksource(&clocksource_apbt))
+			break;
+	}
+	if (!i)
+		goto failed;
+
+	/* count 16 ms */
+	loop = (apbt_freq * 1000) << 4;
+
+	/* restart the timer to ensure it won't get to 0 in the calibration */
+	apbt_start_counter(phy_cs_timer_id);
+
+	old = apbt_read_clocksource(&clocksource_apbt);
+	old += loop;
+
+	t1 = __native_read_tsc();
+
+	do {
+		new = apbt_read_clocksource(&clocksource_apbt);
+	} while (new < old);
+
+	t2 = __native_read_tsc();
+
+	shift = 5;
+	if (unlikely(loop >> shift == 0)) {
+		printk(KERN_INFO
+		       "APBT TSC calibration failed, not enough resolution\n");
+		return 0;
+	}
+	scale = (int)div_u64((t2 - t1), loop >> shift);
+	khz = (scale * apbt_freq * 1000) >> shift;
+	printk(KERN_INFO "TSC freq calculated by APB timer is %lu khz\n", khz);
+	return khz;
+failed:
+	return 0;
+}

arch/x86/kernel/apic/apic.c

@@ -1390,7 +1390,7 @@ void __init enable_IR_x2apic(void)
 	}
 
 	local_irq_save(flags);
-	mask_8259A();
+	legacy_pic->mask_all();
 	mask_IO_APIC_setup(ioapic_entries);
 
 	if (dmar_table_init_ret)
@@ -1422,7 +1422,7 @@ void __init enable_IR_x2apic(void)
 nox2apic:
 	if (!ret) /* IR enabling failed */
 		restore_IO_APIC_setup(ioapic_entries);
-	unmask_8259A();
+	legacy_pic->restore_mask();
 	local_irq_restore(flags);
 
 out:
@@ -2018,7 +2018,7 @@ static int lapic_resume(struct sys_device *dev)
 		}
 
 		mask_IO_APIC_setup(ioapic_entries);
-		mask_8259A();
+		legacy_pic->mask_all();
 	}
 
 	if (x2apic_mode)
@@ -2062,7 +2062,7 @@ static int lapic_resume(struct sys_device *dev)
 
 	if (intr_remapping_enabled) {
 		reenable_intr_remapping(x2apic_mode);
-		unmask_8259A();
+		legacy_pic->restore_mask();
 		restore_IO_APIC_setup(ioapic_entries);
 		free_ioapic_entries(ioapic_entries);
 	}

arch/x86/kernel/apic/io_apic.c

@@ -143,12 +143,6 @@ static struct irq_cfg irq_cfgx[NR_IRQS_LEGACY];
 static struct irq_cfg irq_cfgx[NR_IRQS];
 #endif
 
-void __init io_apic_disable_legacy(void)
-{
-	nr_legacy_irqs = 0;
-	nr_irqs_gsi = 0;
-}
-
 int __init arch_early_irq_init(void)
 {
 	struct irq_cfg *cfg;
@@ -157,6 +151,11 @@ int __init arch_early_irq_init(void)
 	int node;
 	int i;
 
+	if (!legacy_pic->nr_legacy_irqs) {
+		nr_irqs_gsi = 0;
+		io_apic_irqs = ~0UL;
+	}
+
 	cfg = irq_cfgx;
 	count = ARRAY_SIZE(irq_cfgx);
 	node= cpu_to_node(boot_cpu_id);
@@ -170,7 +169,7 @@ int __init arch_early_irq_init(void)
 		 * For legacy IRQ's, start with assigning irq0 to irq15 to
 		 * IRQ0_VECTOR to IRQ15_VECTOR on cpu 0.
 		 */
-		if (i < nr_legacy_irqs) {
+		if (i < legacy_pic->nr_legacy_irqs) {
 			cfg[i].vector = IRQ0_VECTOR + i;
 			cpumask_set_cpu(0, cfg[i].domain);
 		}
@@ -852,7 +851,7 @@ static int __init find_isa_irq_apic(int irq, int type)
  */
 static int EISA_ELCR(unsigned int irq)
 {
-	if (irq < nr_legacy_irqs) {
+	if (irq < legacy_pic->nr_legacy_irqs) {
 		unsigned int port = 0x4d0 + (irq >> 3);
 		return (inb(port) >> (irq & 7)) & 1;
 	}
@@ -1439,7 +1438,7 @@ static void setup_IO_APIC_irq(int apic_id, int pin, unsigned int irq, struct irq
 	 * controllers like 8259. Now that IO-APIC can handle this irq, update
 	 * the cfg->domain.
 	 */
-	if (irq < nr_legacy_irqs && cpumask_test_cpu(0, cfg->domain))
+	if (irq < legacy_pic->nr_legacy_irqs && cpumask_test_cpu(0, cfg->domain))
 		apic->vector_allocation_domain(0, cfg->domain);
 
 	if (assign_irq_vector(irq, cfg, apic->target_cpus()))
@@ -1463,8 +1462,8 @@ static void setup_IO_APIC_irq(int apic_id, int pin, unsigned int irq, struct irq
 	}
 
 	ioapic_register_intr(irq, desc, trigger);
-	if (irq < nr_legacy_irqs)
-		disable_8259A_irq(irq);
+	if (irq < legacy_pic->nr_legacy_irqs)
+		legacy_pic->chip->mask(irq);
 
 	ioapic_write_entry(apic_id, pin, entry);
 }
@@ -1873,7 +1872,7 @@ __apicdebuginit(void) print_PIC(void)
 	unsigned int v;
 	unsigned long flags;
 
-	if (!nr_legacy_irqs)
+	if (!legacy_pic->nr_legacy_irqs)
 		return;
 
 	printk(KERN_DEBUG "\nprinting PIC contents\n");
@@ -1957,7 +1956,7 @@ void __init enable_IO_APIC(void)
 		nr_ioapic_registers[apic] = reg_01.bits.entries+1;
 	}
 
-	if (!nr_legacy_irqs)
+	if (!legacy_pic->nr_legacy_irqs)
 		return;
 
 	for(apic = 0; apic < nr_ioapics; apic++) {
@@ -2014,7 +2013,7 @@ void disable_IO_APIC(void)
 	 */
 	clear_IO_APIC();
 
-	if (!nr_legacy_irqs)
+	if (!legacy_pic->nr_legacy_irqs)
 		return;
 
 	/*
@@ -2247,9 +2246,9 @@ static unsigned int startup_ioapic_irq(unsigned int irq)
 	struct irq_cfg *cfg;
 
 	raw_spin_lock_irqsave(&ioapic_lock, flags);
-	if (irq < nr_legacy_irqs) {
-		disable_8259A_irq(irq);
-		if (i8259A_irq_pending(irq))
+	if (irq < legacy_pic->nr_legacy_irqs) {
+		legacy_pic->chip->mask(irq);
+		if (legacy_pic->irq_pending(irq))
 			was_pending = 1;
 	}
 	cfg = irq_cfg(irq);
@@ -2782,8 +2781,8 @@ static inline void init_IO_APIC_traps(void)
 			 * so default to an old-fashioned 8259
 			 * interrupt if we can..
 			 */
-			if (irq < nr_legacy_irqs)
-				make_8259A_irq(irq);
+			if (irq < legacy_pic->nr_legacy_irqs)
+				legacy_pic->make_irq(irq);
 			else
 				/* Strange. Oh, well.. */
 				desc->chip = &no_irq_chip;
@@ -2940,7 +2939,7 @@ static inline void __init check_timer(void)
 	/*
 	 * get/set the timer IRQ vector:
 	 */
-	disable_8259A_irq(0);
+	legacy_pic->chip->mask(0);
 	assign_irq_vector(0, cfg, apic->target_cpus());
 
 	/*
@@ -2953,7 +2952,7 @@ static inline void __init check_timer(void)
 	 * automatically.
 	 */
 	apic_write(APIC_LVT0, APIC_LVT_MASKED | APIC_DM_EXTINT);
-	init_8259A(1);
+	legacy_pic->init(1);
 #ifdef CONFIG_X86_32
 	{
 		unsigned int ver;
@@ -3012,7 +3011,7 @@ static inline void __init check_timer(void)
 		if (timer_irq_works()) {
 			if (nmi_watchdog == NMI_IO_APIC) {
 				setup_nmi();
-				enable_8259A_irq(0);
+				legacy_pic->chip->unmask(0);
 			}
 			if (disable_timer_pin_1 > 0)
 				clear_IO_APIC_pin(0, pin1);
@@ -3035,14 +3034,14 @@ static inline void __init check_timer(void)
 		 */
 		replace_pin_at_irq_node(cfg, node, apic1, pin1, apic2, pin2);
 		setup_timer_IRQ0_pin(apic2, pin2, cfg->vector);
-		enable_8259A_irq(0);
+		legacy_pic->chip->unmask(0);
 		if (timer_irq_works()) {
 			apic_printk(APIC_QUIET, KERN_INFO "....... works.\n");
 			timer_through_8259 = 1;
 			if (nmi_watchdog == NMI_IO_APIC) {
-				disable_8259A_irq(0);
+				legacy_pic->chip->mask(0);
 				setup_nmi();
-				enable_8259A_irq(0);
+				legacy_pic->chip->unmask(0);
 			}
 			goto out;
 		}
@@ -3050,7 +3049,7 @@ static inline void __init check_timer(void)
 		 * Cleanup, just in case ...
 		 */
 		local_irq_disable();
-		disable_8259A_irq(0);
+		legacy_pic->chip->mask(0);
 		clear_IO_APIC_pin(apic2, pin2);
 		apic_printk(APIC_QUIET, KERN_INFO "....... failed.\n");
 	}
@@ -3069,22 +3068,22 @@ static inline void __init check_timer(void)
 
 	lapic_register_intr(0, desc);
 	apic_write(APIC_LVT0, APIC_DM_FIXED | cfg->vector);	/* Fixed mode */
-	enable_8259A_irq(0);
+	legacy_pic->chip->unmask(0);
 
 	if (timer_irq_works()) {
 		apic_printk(APIC_QUIET, KERN_INFO "..... works.\n");
 		goto out;
 	}
 	local_irq_disable();
-	disable_8259A_irq(0);
+	legacy_pic->chip->mask(0);
 	apic_write(APIC_LVT0, APIC_LVT_MASKED | APIC_DM_FIXED | cfg->vector);
 	apic_printk(APIC_QUIET, KERN_INFO "..... failed.\n");
 
 	apic_printk(APIC_QUIET, KERN_INFO
 		    "...trying to set up timer as ExtINT IRQ...\n");
 
-	init_8259A(0);
-	make_8259A_irq(0);
+	legacy_pic->init(0);
+	legacy_pic->make_irq(0);
 	apic_write(APIC_LVT0, APIC_DM_EXTINT);
 
 	unlock_ExtINT_logic();
@@ -3126,7 +3125,7 @@ void __init setup_IO_APIC(void)
 	/*
 	 * calling enable_IO_APIC() is moved to setup_local_APIC for BP
 	 */
-	io_apic_irqs = nr_legacy_irqs ? ~PIC_IRQS : ~0UL;
+	io_apic_irqs = legacy_pic->nr_legacy_irqs ? ~PIC_IRQS : ~0UL;
 
 	apic_printk(APIC_VERBOSE, "ENABLING IO-APIC IRQs\n");
 	/*
@@ -3137,7 +3136,7 @@ void __init setup_IO_APIC(void)
 	sync_Arb_IDs();
 	setup_IO_APIC_irqs();
 	init_IO_APIC_traps();
-	if (nr_legacy_irqs)
+	if (legacy_pic->nr_legacy_irqs)
 		check_timer();
 }
 
@@ -3928,7 +3927,7 @@ static int __io_apic_set_pci_routing(struct device *dev, int irq,
 	/*
 	 * IRQs < 16 are already in the irq_2_pin[] map
 	 */
-	if (irq >= nr_legacy_irqs) {
+	if (irq >= legacy_pic->nr_legacy_irqs) {
 		cfg = desc->chip_data;
 		if (add_pin_to_irq_node_nopanic(cfg, node, ioapic, pin)) {
 			printk(KERN_INFO "can not add pin %d for irq %d\n",
@@ -4302,3 +4301,24 @@ void __init mp_register_ioapic(int id, u32 address, u32 gsi_base)
 
 	nr_ioapics++;
 }
+
+/* Enable IOAPIC early just for system timer */
+void __init pre_init_apic_IRQ0(void)
+{
+	struct irq_cfg *cfg;
+	struct irq_desc *desc;
+
+	printk(KERN_INFO "Early APIC setup for system timer0\n");
+#ifndef CONFIG_SMP
+	phys_cpu_present_map = physid_mask_of_physid(boot_cpu_physical_apicid);
+#endif
+	desc = irq_to_desc_alloc_node(0, 0);
+
+	setup_local_APIC();
+
+	cfg = irq_cfg(0);
+	add_pin_to_irq_node(cfg, 0, 0, 0);
+	set_irq_chip_and_handler_name(0, &ioapic_chip, handle_edge_irq, "edge");
+
+	setup_IO_APIC_irq(0, 0, 0, desc, 0, 0);
+}

arch/x86/kernel/apic/nmi.c

@@ -177,7 +177,7 @@ int __init check_nmi_watchdog(void)
 error:
 	if (nmi_watchdog == NMI_IO_APIC) {
 		if (!timer_through_8259)
-			disable_8259A_irq(0);
+			legacy_pic->chip->mask(0);
 		on_each_cpu(__acpi_nmi_disable, NULL, 1);
 	}
 

arch/x86/kernel/apic/numaq_32.c

@@ -277,6 +277,7 @@ static __init void early_check_numaq(void)
 		x86_init.mpparse.mpc_oem_pci_bus = mpc_oem_pci_bus;
 		x86_init.mpparse.mpc_oem_bus_info = mpc_oem_bus_info;
 		x86_init.timers.tsc_pre_init = numaq_tsc_init;
+		x86_init.pci.init = pci_numaq_init;
 	}
 }
 

arch/x86/kernel/i8259.c

@@ -34,6 +34,12 @@
 static int i8259A_auto_eoi;
 DEFINE_RAW_SPINLOCK(i8259A_lock);
 static void mask_and_ack_8259A(unsigned int);
+static void mask_8259A(void);
+static void unmask_8259A(void);
+static void disable_8259A_irq(unsigned int irq);
+static void enable_8259A_irq(unsigned int irq);
+static void init_8259A(int auto_eoi);
+static int i8259A_irq_pending(unsigned int irq);
 
 struct irq_chip i8259A_chip = {
 	.name		= "XT-PIC",
@@ -63,7 +69,7 @@ unsigned int cached_irq_mask = 0xffff;
  */
 unsigned long io_apic_irqs;
 
-void disable_8259A_irq(unsigned int irq)
+static void disable_8259A_irq(unsigned int irq)
 {
 	unsigned int mask = 1 << irq;
 	unsigned long flags;
@@ -77,7 +83,7 @@ void disable_8259A_irq(unsigned int irq)
 	raw_spin_unlock_irqrestore(&i8259A_lock, flags);
 }
 
-void enable_8259A_irq(unsigned int irq)
+static void enable_8259A_irq(unsigned int irq)
 {
 	unsigned int mask = ~(1 << irq);
 	unsigned long flags;
@@ -91,7 +97,7 @@ void enable_8259A_irq(unsigned int irq)
 	raw_spin_unlock_irqrestore(&i8259A_lock, flags);
 }
 
-int i8259A_irq_pending(unsigned int irq)
+static int i8259A_irq_pending(unsigned int irq)
 {
 	unsigned int mask = 1<<irq;
 	unsigned long flags;
@@ -107,7 +113,7 @@ int i8259A_irq_pending(unsigned int irq)
 	return ret;
 }
 
-void make_8259A_irq(unsigned int irq)
+static void make_8259A_irq(unsigned int irq)
 {
 	disable_irq_nosync(irq);
 	io_apic_irqs &= ~(1<<irq);
@@ -281,7 +287,7 @@ static int __init i8259A_init_sysfs(void)
 
 device_initcall(i8259A_init_sysfs);
 
-void mask_8259A(void)
+static void mask_8259A(void)
 {
 	unsigned long flags;
 
@@ -293,7 +299,7 @@ void mask_8259A(void)
 	raw_spin_unlock_irqrestore(&i8259A_lock, flags);
 }
 
-void unmask_8259A(void)
+static void unmask_8259A(void)
 {
 	unsigned long flags;
 
@@ -305,7 +311,7 @@ void unmask_8259A(void)
 	raw_spin_unlock_irqrestore(&i8259A_lock, flags);
 }
 
-void init_8259A(int auto_eoi)
+static void init_8259A(int auto_eoi)
 {
 	unsigned long flags;
 
@@ -358,3 +364,47 @@ void init_8259A(int auto_eoi)
 
 	raw_spin_unlock_irqrestore(&i8259A_lock, flags);
 }
+
+/*
+ * make i8259 a driver so that we can select pic functions at run time. the goal
+ * is to make x86 binary compatible among pc compatible and non-pc compatible
+ * platforms, such as x86 MID.
+ */
+
+static void legacy_pic_noop(void) { };
+static void legacy_pic_uint_noop(unsigned int unused) { };
+static void legacy_pic_int_noop(int unused) { };
+
+static struct irq_chip dummy_pic_chip  = {
+	.name = "dummy pic",
+	.mask = legacy_pic_uint_noop,
+	.unmask = legacy_pic_uint_noop,
+	.disable = legacy_pic_uint_noop,
+	.mask_ack = legacy_pic_uint_noop,
+};
+static int legacy_pic_irq_pending_noop(unsigned int irq)
+{
+	return 0;
+}
+
+struct legacy_pic null_legacy_pic = {
+	.nr_legacy_irqs = 0,
+	.chip = &dummy_pic_chip,
+	.mask_all = legacy_pic_noop,
+	.restore_mask = legacy_pic_noop,
+	.init = legacy_pic_int_noop,
+	.irq_pending = legacy_pic_irq_pending_noop,
+	.make_irq = legacy_pic_uint_noop,
+};
+
+struct legacy_pic default_legacy_pic = {
+	.nr_legacy_irqs = NR_IRQS_LEGACY,
+	.chip  = &i8259A_chip,
+	.mask_all  = mask_8259A,
+	.restore_mask = unmask_8259A,
+	.init = init_8259A,
+	.irq_pending = i8259A_irq_pending,
+	.make_irq = make_8259A_irq,
+};
+
+struct legacy_pic *legacy_pic = &default_legacy_pic;

arch/x86/kernel/irqinit.c

@@ -99,9 +99,6 @@ int vector_used_by_percpu_irq(unsigned int vector)
 	return 0;
 }
 
-/* Number of legacy interrupts */
-int nr_legacy_irqs __read_mostly = NR_IRQS_LEGACY;
-
 void __init init_ISA_irqs(void)
 {
 	int i;
@@ -109,12 +106,12 @@ void __init init_ISA_irqs(void)
 #if defined(CONFIG_X86_64) || defined(CONFIG_X86_LOCAL_APIC)
 	init_bsp_APIC();
 #endif
-	init_8259A(0);
+	legacy_pic->init(0);
 
 	/*
 	 * 16 old-style INTA-cycle interrupts:
 	 */
-	for (i = 0; i < NR_IRQS_LEGACY; i++) {
+	for (i = 0; i < legacy_pic->nr_legacy_irqs; i++) {
 		struct irq_desc *desc = irq_to_desc(i);
 
 		desc->status = IRQ_DISABLED;
@@ -138,7 +135,7 @@ void __init init_IRQ(void)
 	 * then this vector space can be freed and re-used dynamically as the
 	 * irq's migrate etc.
 	 */
-	for (i = 0; i < nr_legacy_irqs; i++)
+	for (i = 0; i < legacy_pic->nr_legacy_irqs; i++)
 		per_cpu(vector_irq, 0)[IRQ0_VECTOR + i] = i;
 
 	x86_init.irqs.intr_init();

arch/x86/kernel/mrst.c

@@ -10,8 +10,211 @@
  * of the License.
  */
 #include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/sfi.h>
+#include <linux/irq.h>
+#include <linux/module.h>
 
 #include <asm/setup.h>
+#include <asm/mpspec_def.h>
+#include <asm/hw_irq.h>
+#include <asm/apic.h>
+#include <asm/io_apic.h>
+#include <asm/mrst.h>
+#include <asm/io.h>
+#include <asm/i8259.h>
+#include <asm/apb_timer.h>
+
+static u32 sfi_mtimer_usage[SFI_MTMR_MAX_NUM];
+static struct sfi_timer_table_entry sfi_mtimer_array[SFI_MTMR_MAX_NUM];
+int sfi_mtimer_num;
+
+struct sfi_rtc_table_entry sfi_mrtc_array[SFI_MRTC_MAX];
+EXPORT_SYMBOL_GPL(sfi_mrtc_array);
+int sfi_mrtc_num;
+
+static inline void assign_to_mp_irq(struct mpc_intsrc *m,
+				    struct mpc_intsrc *mp_irq)
+{
+	memcpy(mp_irq, m, sizeof(struct mpc_intsrc));
+}
+
+static inline int mp_irq_cmp(struct mpc_intsrc *mp_irq,
+				struct mpc_intsrc *m)
+{
+	return memcmp(mp_irq, m, sizeof(struct mpc_intsrc));
+}
+
+static void save_mp_irq(struct mpc_intsrc *m)
+{
+	int i;
+
+	for (i = 0; i < mp_irq_entries; i++) {
+		if (!mp_irq_cmp(&mp_irqs[i], m))
+			return;
+	}
+
+	assign_to_mp_irq(m, &mp_irqs[mp_irq_entries]);
+	if (++mp_irq_entries == MAX_IRQ_SOURCES)
+		panic("Max # of irq sources exceeded!!\n");
+}
+
+/* parse all the mtimer info to a static mtimer array */
+static int __init sfi_parse_mtmr(struct sfi_table_header *table)
+{
+	struct sfi_table_simple *sb;
+	struct sfi_timer_table_entry *pentry;
+	struct mpc_intsrc mp_irq;
+	int totallen;
+
+	sb = (struct sfi_table_simple *)table;
+	if (!sfi_mtimer_num) {
+		sfi_mtimer_num = SFI_GET_NUM_ENTRIES(sb,
+					struct sfi_timer_table_entry);
+		pentry = (struct sfi_timer_table_entry *) sb->pentry;
+		totallen = sfi_mtimer_num * sizeof(*pentry);
+		memcpy(sfi_mtimer_array, pentry, totallen);
+	}
+
+	printk(KERN_INFO "SFI: MTIMER info (num = %d):\n", sfi_mtimer_num);
+	pentry = sfi_mtimer_array;
+	for (totallen = 0; totallen < sfi_mtimer_num; totallen++, pentry++) {
+		printk(KERN_INFO "timer[%d]: paddr = 0x%08x, freq = %dHz,"
+			" irq = %d\n", totallen, (u32)pentry->phys_addr,
+			pentry->freq_hz, pentry->irq);
+			if (!pentry->irq)
+				continue;
+			mp_irq.type = MP_IOAPIC;
+			mp_irq.irqtype = mp_INT;
+/* triggering mode edge bit 2-3, active high polarity bit 0-1 */
+			mp_irq.irqflag = 5;
+			mp_irq.srcbus = 0;
+			mp_irq.srcbusirq = pentry->irq;	/* IRQ */
+			mp_irq.dstapic = MP_APIC_ALL;
+			mp_irq.dstirq = pentry->irq;
+			save_mp_irq(&mp_irq);
+	}
+
+	return 0;
+}
+
+struct sfi_timer_table_entry *sfi_get_mtmr(int hint)
+{
+	int i;
+	if (hint < sfi_mtimer_num) {
+		if (!sfi_mtimer_usage[hint]) {
+			pr_debug("hint taken for timer %d irq %d\n",\
+				hint, sfi_mtimer_array[hint].irq);
+			sfi_mtimer_usage[hint] = 1;
+			return &sfi_mtimer_array[hint];
+		}
+	}
+	/* take the first timer available */
+	for (i = 0; i < sfi_mtimer_num;) {
+		if (!sfi_mtimer_usage[i]) {
+			sfi_mtimer_usage[i] = 1;
+			return &sfi_mtimer_array[i];
+		}
+		i++;
+	}
+	return NULL;
+}
+
+void sfi_free_mtmr(struct sfi_timer_table_entry *mtmr)
+{
+	int i;
+	for (i = 0; i < sfi_mtimer_num;) {
+		if (mtmr->irq == sfi_mtimer_array[i].irq) {
+			sfi_mtimer_usage[i] = 0;
+			return;
+		}
+		i++;
+	}
+}
+
+/* parse all the mrtc info to a global mrtc array */
+int __init sfi_parse_mrtc(struct sfi_table_header *table)
+{
+	struct sfi_table_simple *sb;
+	struct sfi_rtc_table_entry *pentry;
+	struct mpc_intsrc mp_irq;
+
+	int totallen;
+
+	sb = (struct sfi_table_simple *)table;
+	if (!sfi_mrtc_num) {
+		sfi_mrtc_num = SFI_GET_NUM_ENTRIES(sb,
+						struct sfi_rtc_table_entry);
+		pentry = (struct sfi_rtc_table_entry *)sb->pentry;
+		totallen = sfi_mrtc_num * sizeof(*pentry);
+		memcpy(sfi_mrtc_array, pentry, totallen);
+	}
+
+	printk(KERN_INFO "SFI: RTC info (num = %d):\n", sfi_mrtc_num);
+	pentry = sfi_mrtc_array;
+	for (totallen = 0; totallen < sfi_mrtc_num; totallen++, pentry++) {
+		printk(KERN_INFO "RTC[%d]: paddr = 0x%08x, irq = %d\n",
+			totallen, (u32)pentry->phys_addr, pentry->irq);
+		mp_irq.type = MP_IOAPIC;
+		mp_irq.irqtype = mp_INT;
+		mp_irq.irqflag = 0;
+		mp_irq.srcbus = 0;
+		mp_irq.srcbusirq = pentry->irq;	/* IRQ */
+		mp_irq.dstapic = MP_APIC_ALL;
+		mp_irq.dstirq = pentry->irq;
+		save_mp_irq(&mp_irq);
+	}
+	return 0;
+}
+
+/*
+ * the secondary clock in Moorestown can be APBT or LAPIC clock, default to
+ * APBT but cmdline option can also override it.
+ */
+static void __cpuinit mrst_setup_secondary_clock(void)
+{
+	/* restore default lapic clock if disabled by cmdline */
+	if (disable_apbt_percpu)
+		return setup_secondary_APIC_clock();
+	apbt_setup_secondary_clock();
+}
+
+static unsigned long __init mrst_calibrate_tsc(void)
+{
+	unsigned long flags, fast_calibrate;
+
+	local_irq_save(flags);
+	fast_calibrate = apbt_quick_calibrate();
+	local_irq_restore(flags);
+
+	if (fast_calibrate)
+		return fast_calibrate;
+
+	return 0;
+}
+
+void __init mrst_time_init(void)
+{
+	sfi_table_parse(SFI_SIG_MTMR, NULL, NULL, sfi_parse_mtmr);
+	pre_init_apic_IRQ0();
+	apbt_time_init();
+}
+
+void __init mrst_rtc_init(void)
+{
+	sfi_table_parse(SFI_SIG_MRTC, NULL, NULL, sfi_parse_mrtc);
+}
+
+/*
+ * if we use per cpu apb timer, the bootclock already setup. if we use lapic
+ * timer and one apbt timer for broadcast, we need to set up lapic boot clock.
+ */
+static void __init mrst_setup_boot_clock(void)
+{
+	pr_info("%s: per cpu apbt flag %d \n", __func__, disable_apbt_percpu);
+	if (disable_apbt_percpu)
+		setup_boot_APIC_clock();
+};
 
 /*
  * Moorestown specific x86_init function overrides and early setup
@@ -21,4 +224,17 @@ void __init x86_mrst_early_setup(void)
 {
 	x86_init.resources.probe_roms = x86_init_noop;
 	x86_init.resources.reserve_resources = x86_init_noop;
+
+	x86_init.timers.timer_init = mrst_time_init;
+	x86_init.timers.setup_percpu_clockev = mrst_setup_boot_clock;
+
+	x86_init.irqs.pre_vector_init = x86_init_noop;
+
+	x86_cpuinit.setup_percpu_clockev = mrst_setup_secondary_clock;
+
+	x86_platform.calibrate_tsc = mrst_calibrate_tsc;
+	x86_init.pci.init = pci_mrst_init;
+	x86_init.pci.fixup_irqs = x86_init_noop;
+
+	legacy_pic = &null_legacy_pic;
 }

arch/x86/kernel/olpc.c

@@ -17,7 +17,9 @@
 #include <linux/spinlock.h>
 #include <linux/io.h>
 #include <linux/string.h>
+
 #include <asm/geode.h>
+#include <asm/setup.h>
 #include <asm/olpc.h>
 
 #ifdef CONFIG_OPEN_FIRMWARE
@@ -243,9 +245,11 @@ static int __init olpc_init(void)
 	olpc_ec_cmd(EC_FIRMWARE_REV, NULL, 0,
 			(unsigned char *) &olpc_platform_info.ecver, 1);
 
-	/* check to see if the VSA exists */
-	if (cs5535_has_vsa2())
-		olpc_platform_info.flags |= OLPC_F_VSA;
+#ifdef CONFIG_PCI_OLPC
+	/* If the VSA exists let it emulate PCI, if not emulate in kernel */
+	if (!cs5535_has_vsa2())
+		x86_init.pci.arch_init = pci_olpc_init;
+#endif
 
 	printk(KERN_INFO "OLPC board revision %s%X (EC=%x)\n",
 			((olpc_platform_info.boardrev & 0xf) < 8) ? "pre" : "",

arch/x86/kernel/smpboot.c

@@ -48,6 +48,7 @@
 #include <linux/err.h>
 #include <linux/nmi.h>
 #include <linux/tboot.h>
+#include <linux/stackprotector.h>
 
 #include <asm/acpi.h>
 #include <asm/desc.h>
@@ -67,6 +68,7 @@
 #include <linux/mc146818rtc.h>
 
 #include <asm/smpboot_hooks.h>
+#include <asm/i8259.h>
 
 #ifdef CONFIG_X86_32
 u8 apicid_2_node[MAX_APICID];
@@ -291,9 +293,9 @@ notrace static void __cpuinit start_secondary(void *unused)
 	check_tsc_sync_target();
 
 	if (nmi_watchdog == NMI_IO_APIC) {
-		disable_8259A_irq(0);
+		legacy_pic->chip->mask(0);
 		enable_NMI_through_LVT0();
-		enable_8259A_irq(0);
+		legacy_pic->chip->unmask(0);
 	}
 
 #ifdef CONFIG_X86_32
@@ -329,6 +331,9 @@ notrace static void __cpuinit start_secondary(void *unused)
 	/* enable local interrupts */
 	local_irq_enable();
 
+	/* to prevent fake stack check failure in clock setup */
+	boot_init_stack_canary();
+
 	x86_cpuinit.setup_percpu_clockev();
 
 	wmb();

arch/x86/kernel/visws_quirks.c

@@ -49,11 +49,6 @@ extern int no_broadcast;
 char visws_board_type	= -1;
 char visws_board_rev	= -1;
 
-int is_visws_box(void)
-{
-	return visws_board_type >= 0;
-}
-
 static void __init visws_time_init(void)
 {
 	printk(KERN_INFO "Starting Cobalt Timer system clock\n");
@@ -242,6 +237,8 @@ void __init visws_early_detect(void)
 	x86_init.irqs.pre_vector_init = visws_pre_intr_init;
 	x86_init.irqs.trap_init = visws_trap_init;
 	x86_init.timers.timer_init = visws_time_init;
+	x86_init.pci.init = pci_visws_init;
+	x86_init.pci.init_irq = x86_init_noop;
 
 	/*
 	 * Install reboot quirks:
@@ -508,7 +505,7 @@ static struct irq_chip cobalt_irq_type = {
  */
 static unsigned int startup_piix4_master_irq(unsigned int irq)
 {
-	init_8259A(0);
+	legacy_pic->init(0);
 
 	return startup_cobalt_irq(irq);
 }
@@ -532,9 +529,6 @@ static struct irq_chip piix4_master_irq_type = {
 
 static struct irq_chip piix4_virtual_irq_type = {
 	.name =		"PIIX4-virtual",
-	.shutdown =	disable_8259A_irq,
-	.enable =	enable_8259A_irq,
-	.disable =	disable_8259A_irq,
 };
 
 
@@ -609,7 +603,7 @@ static irqreturn_t piix4_master_intr(int irq, void *dev_id)
 		handle_IRQ_event(realirq, desc->action);
 
 	if (!(desc->status & IRQ_DISABLED))
-		enable_8259A_irq(realirq);
+		legacy_pic->chip->unmask(realirq);
 
 	return IRQ_HANDLED;
 
@@ -628,6 +622,12 @@ static struct irqaction cascade_action = {
 	.name =		"cascade",
 };
 
+static inline void set_piix4_virtual_irq_type(void)
+{
+	piix4_virtual_irq_type.shutdown = i8259A_chip.mask;
+	piix4_virtual_irq_type.enable =	i8259A_chip.unmask;
+	piix4_virtual_irq_type.disable = i8259A_chip.mask;
+}
 
 void init_VISWS_APIC_irqs(void)
 {
@@ -653,6 +653,7 @@ void init_VISWS_APIC_irqs(void)
 			desc->chip = &piix4_master_irq_type;
 		}
 		else if (i < CO_IRQ_APIC0) {
+			set_piix4_virtual_irq_type();
 			desc->chip = &piix4_virtual_irq_type;
 		}
 		else if (IS_CO_APIC(i)) {

arch/x86/kernel/x86_init.c

@@ -4,9 +4,11 @@
  *  For licencing details see kernel-base/COPYING
  */
 #include <linux/init.h>
+#include <linux/ioport.h>
 
 #include <asm/bios_ebda.h>
 #include <asm/paravirt.h>
+#include <asm/pci_x86.h>
 #include <asm/mpspec.h>
 #include <asm/setup.h>
 #include <asm/apic.h>
@@ -70,6 +72,12 @@ struct x86_init_ops x86_init __initdata = {
 	.iommu = {
 		.iommu_init		= iommu_init_noop,
 	},
+
+	.pci = {
+		.init			= x86_default_pci_init,
+		.init_irq		= x86_default_pci_init_irq,
+		.fixup_irqs		= x86_default_pci_fixup_irqs,
+	},
 };
 
 struct x86_cpuinit_ops x86_cpuinit __cpuinitdata = {

arch/x86/pci/Makefile

@@ -13,6 +13,8 @@ obj-$(CONFIG_X86_VISWS)		+= visws.o
 
 obj-$(CONFIG_X86_NUMAQ)		+= numaq_32.o
 
+obj-$(CONFIG_X86_MRST)		+= mrst.o
+
 obj-y				+= common.o early.o
 obj-y				+= amd_bus.o bus_numa.o
 

arch/x86/pci/acpi.c

@@ -298,17 +298,14 @@ int __init pci_acpi_init(void)
 {
 	struct pci_dev *dev = NULL;
 
-	if (pcibios_scanned)
-		return 0;
-
 	if (acpi_noirq)
-		return 0;
+		return -ENODEV;
 
 	printk(KERN_INFO "PCI: Using ACPI for IRQ routing\n");
 	acpi_irq_penalty_init();
-	pcibios_scanned++;
 	pcibios_enable_irq = acpi_pci_irq_enable;
 	pcibios_disable_irq = acpi_pci_irq_disable;
+	x86_init.pci.init_irq = x86_init_noop;
 
 	if (pci_routeirq) {
 		/*

arch/x86/pci/common.c

@@ -71,12 +71,6 @@ struct pci_ops pci_root_ops = {
 	.write = pci_write,
 };
 
-/*
- * legacy, numa, and acpi all want to call pcibios_scan_root
- * from their initcalls. This flag prevents that.
- */
-int pcibios_scanned;
-
 /*
  * This interrupt-safe spinlock protects all accesses to PCI
  * configuration space.

arch/x86/pci/init.c

 #include <linux/pci.h>
 #include <linux/init.h>
 #include <asm/pci_x86.h>
+#include <asm/x86_init.h>
 
 /* arch_initcall has too random ordering, so call the initializers
    in the right sequence from here. */
@@ -15,10 +16,9 @@ static __init int pci_arch_init(void)
 	if (!(pci_probe & PCI_PROBE_NOEARLY))
 		pci_mmcfg_early_init();
 
-#ifdef CONFIG_PCI_OLPC
-	if (!pci_olpc_init())
-		return 0;	/* skip additional checks if it's an XO */
-#endif
+	if (x86_init.pci.arch_init && !x86_init.pci.arch_init())
+		return 0;
+
 #ifdef CONFIG_PCI_BIOS
 	pci_pcbios_init();
 #endif

arch/x86/pci/irq.c

@@ -53,7 +53,7 @@ struct irq_router_handler {
 	int (*probe)(struct irq_router *r, struct pci_dev *router, u16 device);
 };
 
-int (*pcibios_enable_irq)(struct pci_dev *dev) = NULL;
+int (*pcibios_enable_irq)(struct pci_dev *dev) = pirq_enable_irq;
 void (*pcibios_disable_irq)(struct pci_dev *dev) = NULL;
 
 /*
@@ -1018,7 +1018,7 @@ static int pcibios_lookup_irq(struct pci_dev *dev, int assign)
 	return 1;
 }
 
-static void __init pcibios_fixup_irqs(void)
+void __init pcibios_fixup_irqs(void)
 {
 	struct pci_dev *dev = NULL;
 	u8 pin;
@@ -1112,12 +1112,12 @@ static struct dmi_system_id __initdata pciirq_dmi_table[] = {
 	{ }
 };
 
-int __init pcibios_irq_init(void)
+void __init pcibios_irq_init(void)
 {
 	DBG(KERN_DEBUG "PCI: IRQ init\n");
 
-	if (pcibios_enable_irq || raw_pci_ops == NULL)
-		return 0;
+	if (raw_pci_ops == NULL)
+		return;
 
 	dmi_check_system(pciirq_dmi_table);
 
@@ -1144,9 +1144,7 @@ int __init pcibios_irq_init(void)
 			pirq_table = NULL;
 	}
 
-	pcibios_enable_irq = pirq_enable_irq;
-
-	pcibios_fixup_irqs();
+	x86_init.pci.fixup_irqs();
 
 	if (io_apic_assign_pci_irqs && pci_routeirq) {
 		struct pci_dev *dev = NULL;
@@ -1159,8 +1157,6 @@ int __init pcibios_irq_init(void)
 		for_each_pci_dev(dev)
 			pirq_enable_irq(dev);
 	}
-
-	return 0;
 }
 
 static void pirq_penalize_isa_irq(int irq, int active)

arch/x86/pci/legacy.c

@@ -35,16 +35,13 @@ static void __devinit pcibios_fixup_peer_bridges(void)
 	}
 }
 
-static int __init pci_legacy_init(void)
+int __init pci_legacy_init(void)
 {
 	if (!raw_pci_ops) {
 		printk("PCI: System does not support PCI\n");
 		return 0;
 	}
 
-	if (pcibios_scanned++)
-		return 0;
-
 	printk("PCI: Probing PCI hardware\n");
 	pci_root_bus = pcibios_scan_root(0);
 	if (pci_root_bus)
@@ -55,18 +52,15 @@ static int __init pci_legacy_init(void)
 
 int __init pci_subsys_init(void)
 {
-#ifdef CONFIG_X86_NUMAQ
-	pci_numaq_init();
-#endif
-#ifdef CONFIG_ACPI
-	pci_acpi_init();
-#endif
-#ifdef CONFIG_X86_VISWS
-	pci_visws_init();
-#endif
+	/*
+	 * The init function returns an non zero value when
+	 * pci_legacy_init should be invoked.
+	 */
+	if (x86_init.pci.init())
 		pci_legacy_init();
+
 	pcibios_fixup_peer_bridges();
-	pcibios_irq_init();
+	x86_init.pci.init_irq();
 	pcibios_init();
 
 	return 0;

arch/x86/pci/mrst.c

+/*
+ * Moorestown PCI support
+ *   Copyright (c) 2008 Intel Corporation
+ *     Jesse Barnes <jesse.barnes@intel.com>
+ *
+ * Moorestown has an interesting PCI implementation:
+ *   - configuration space is memory mapped (as defined by MCFG)
+ *   - Lincroft devices also have a real, type 1 configuration space
+ *   - Early Lincroft silicon has a type 1 access bug that will cause
+ *     a hang if non-existent devices are accessed
+ *   - some devices have the "fixed BAR" capability, which means
+ *     they can't be relocated or modified; check for that during
+ *     BAR sizing
+ *
+ * So, we use the MCFG space for all reads and writes, but also send
+ * Lincroft writes to type 1 space.  But only read/write if the device
+ * actually exists, otherwise return all 1s for reads and bit bucket
+ * the writes.
+ */
+
+#include <linux/sched.h>
+#include <linux/pci.h>
+#include <linux/ioport.h>
+#include <linux/init.h>
+#include <linux/dmi.h>
+
+#include <asm/acpi.h>
+#include <asm/segment.h>
+#include <asm/io.h>
+#include <asm/smp.h>
+#include <asm/pci_x86.h>
+#include <asm/hw_irq.h>
+#include <asm/io_apic.h>
+
+#define PCIE_CAP_OFFSET	0x100
+
+/* Fixed BAR fields */
+#define PCIE_VNDR_CAP_ID_FIXED_BAR 0x00	/* Fixed BAR (TBD) */
+#define PCI_FIXED_BAR_0_SIZE	0x04
+#define PCI_FIXED_BAR_1_SIZE	0x08
+#define PCI_FIXED_BAR_2_SIZE	0x0c
+#define PCI_FIXED_BAR_3_SIZE	0x10
+#define PCI_FIXED_BAR_4_SIZE	0x14
+#define PCI_FIXED_BAR_5_SIZE	0x1c
+
+/**
+ * fixed_bar_cap - return the offset of the fixed BAR cap if found
+ * @bus: PCI bus
+ * @devfn: device in question
+ *
+ * Look for the fixed BAR cap on @bus and @devfn, returning its offset
+ * if found or 0 otherwise.
+ */
+static int fixed_bar_cap(struct pci_bus *bus, unsigned int devfn)
+{
+	int pos;
+	u32 pcie_cap = 0, cap_data;
+
+	pos = PCIE_CAP_OFFSET;
+
+	if (!raw_pci_ext_ops)
+		return 0;
+
+	while (pos) {
+		if (raw_pci_ext_ops->read(pci_domain_nr(bus), bus->number,
+					  devfn, pos, 4, &pcie_cap))
+			return 0;
+
+		if (pcie_cap == 0xffffffff)
+			return 0;
+
+		if (PCI_EXT_CAP_ID(pcie_cap) == PCI_EXT_CAP_ID_VNDR) {
+			raw_pci_ext_ops->read(pci_domain_nr(bus), bus->number,
+					      devfn, pos + 4, 4, &cap_data);
+			if ((cap_data & 0xffff) == PCIE_VNDR_CAP_ID_FIXED_BAR)
+				return pos;
+		}
+
+		pos = pcie_cap >> 20;
+	}
+
+	return 0;
+}
+
+static int pci_device_update_fixed(struct pci_bus *bus, unsigned int devfn,
+				   int reg, int len, u32 val, int offset)
+{
+	u32 size;
+	unsigned int domain, busnum;
+	int bar = (reg - PCI_BASE_ADDRESS_0) >> 2;
+
+	domain = pci_domain_nr(bus);
+	busnum = bus->number;
+
+	if (val == ~0 && len == 4) {
+		unsigned long decode;
+
+		raw_pci_ext_ops->read(domain, busnum, devfn,
+			       offset + 8 + (bar * 4), 4, &size);
+
+		/* Turn the size into a decode pattern for the sizing code */
+		if (size) {
+			decode = size - 1;
+			decode |= decode >> 1;
+			decode |= decode >> 2;
+			decode |= decode >> 4;
+			decode |= decode >> 8;
+			decode |= decode >> 16;
+			decode++;
+			decode = ~(decode - 1);
+		} else {
+			decode = ~0;
+		}
+
+		/*
+		 * If val is all ones, the core code is trying to size the reg,
+		 * so update the mmconfig space with the real size.
+		 *
+		 * Note: this assumes the fixed size we got is a power of two.
+		 */
+		return raw_pci_ext_ops->write(domain, busnum, devfn, reg, 4,
+				       decode);
+	}
+
+	/* This is some other kind of BAR write, so just do it. */
+	return raw_pci_ext_ops->write(domain, busnum, devfn, reg, len, val);
+}
+
+/**
+ * type1_access_ok - check whether to use type 1
+ * @bus: bus number
+ * @devfn: device & function in question
+ *
+ * If the bus is on a Lincroft chip and it exists, or is not on a Lincroft at
+ * all, the we can go ahead with any reads & writes.  If it's on a Lincroft,
+ * but doesn't exist, avoid the access altogether to keep the chip from
+ * hanging.
+ */
+static bool type1_access_ok(unsigned int bus, unsigned int devfn, int reg)
+{
+	/* This is a workaround for A0 LNC bug where PCI status register does
+	 * not have new CAP bit set. can not be written by SW either.
+	 *
+	 * PCI header type in real LNC indicates a single function device, this
+	 * will prevent probing other devices under the same function in PCI
+	 * shim. Therefore, use the header type in shim instead.
+	 */
+	if (reg >= 0x100 || reg == PCI_STATUS || reg == PCI_HEADER_TYPE)
+		return 0;
+	if (bus == 0 && (devfn == PCI_DEVFN(2, 0) || devfn == PCI_DEVFN(0, 0)))
+		return 1;
+	return 0; /* langwell on others */
+}
+
+static int pci_read(struct pci_bus *bus, unsigned int devfn, int where,
+		    int size, u32 *value)
+{
+	if (type1_access_ok(bus->number, devfn, where))
+		return pci_direct_conf1.read(pci_domain_nr(bus), bus->number,
+					devfn, where, size, value);
+	return raw_pci_ext_ops->read(pci_domain_nr(bus), bus->number,
+			      devfn, where, size, value);
+}
+
+static int pci_write(struct pci_bus *bus, unsigned int devfn, int where,
+		     int size, u32 value)
+{
+	int offset;
+
+	/* On MRST, there is no PCI ROM BAR, this will cause a subsequent read
+	 * to ROM BAR return 0 then being ignored.
+	 */
+	if (where == PCI_ROM_ADDRESS)
+		return 0;
+
+	/*
+	 * Devices with fixed BARs need special handling:
+	 *   - BAR sizing code will save, write ~0, read size, restore
+	 *   - so writes to fixed BARs need special handling
+	 *   - other writes to fixed BAR devices should go through mmconfig
+	 */
+	offset = fixed_bar_cap(bus, devfn);
+	if (offset &&
+	    (where >= PCI_BASE_ADDRESS_0 && where <= PCI_BASE_ADDRESS_5)) {
+		return pci_device_update_fixed(bus, devfn, where, size, value,
+					       offset);
+	}
+
+	/*
+	 * On Moorestown update both real & mmconfig space
+	 * Note: early Lincroft silicon can't handle type 1 accesses to
+	 *       non-existent devices, so just eat the write in that case.
+	 */
+	if (type1_access_ok(bus->number, devfn, where))
+		return pci_direct_conf1.write(pci_domain_nr(bus), bus->number,
+					      devfn, where, size, value);
+	return raw_pci_ext_ops->write(pci_domain_nr(bus), bus->number, devfn,
+			       where, size, value);
+}
+
+static int mrst_pci_irq_enable(struct pci_dev *dev)
+{
+	u8 pin;
+	struct io_apic_irq_attr irq_attr;
+
+	pci_read_config_byte(dev, PCI_INTERRUPT_PIN, &pin);
+
+	/* MRST only have IOAPIC, the PCI irq lines are 1:1 mapped to
+	 * IOAPIC RTE entries, so we just enable RTE for the device.
+	 */
+	irq_attr.ioapic = mp_find_ioapic(dev->irq);
+	irq_attr.ioapic_pin = dev->irq;
+	irq_attr.trigger = 1; /* level */
+	irq_attr.polarity = 1; /* active low */
+	io_apic_set_pci_routing(&dev->dev, dev->irq, &irq_attr);
+
+	return 0;
+}
+
+struct pci_ops pci_mrst_ops = {
+	.read = pci_read,
+	.write = pci_write,
+};
+
+/**
+ * pci_mrst_init - installs pci_mrst_ops
+ *
+ * Moorestown has an interesting PCI implementation (see above).
+ * Called when the early platform detection installs it.
+ */
+int __init pci_mrst_init(void)
+{
+	printk(KERN_INFO "Moorestown platform detected, using MRST PCI ops\n");
+	pci_mmcfg_late_init();
+	pcibios_enable_irq = mrst_pci_irq_enable;
+	pci_root_ops = pci_mrst_ops;
+	/* Continue with standard init */
+	return 1;
+}
+
+/*
+ * Langwell devices reside at fixed offsets, don't try to move them.
+ */
+static void __devinit pci_fixed_bar_fixup(struct pci_dev *dev)
+{
+	unsigned long offset;
+	u32 size;
+	int i;
+
+	/* Fixup the BAR sizes for fixed BAR devices and make them unmoveable */
+	offset = fixed_bar_cap(dev->bus, dev->devfn);
+	if (!offset || PCI_DEVFN(2, 0) == dev->devfn ||
+	    PCI_DEVFN(2, 2) == dev->devfn)
+		return;
+
+	for (i = 0; i < PCI_ROM_RESOURCE; i++) {
+		pci_read_config_dword(dev, offset + 8 + (i * 4), &size);
+		dev->resource[i].end = dev->resource[i].start + size - 1;
+		dev->resource[i].flags |= IORESOURCE_PCI_FIXED;
+	}
+}
+DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, PCI_ANY_ID, pci_fixed_bar_fixup);

arch/x86/pci/numaq_32.c

@@ -148,14 +148,8 @@ int __init pci_numaq_init(void)
 {
 	int quad;
 
-	if (!found_numaq)
-		return 0;
-
 	raw_pci_ops = &pci_direct_conf1_mq;
 
-	if (pcibios_scanned++)
-		return 0;
-
 	pci_root_bus = pcibios_scan_root(0);
 	if (pci_root_bus)
 		pci_bus_add_devices(pci_root_bus);

arch/x86/pci/olpc.c

@@ -304,9 +304,6 @@ static struct pci_raw_ops pci_olpc_conf = {
 
 int __init pci_olpc_init(void)
 {
-	if (!machine_is_olpc() || olpc_has_vsa())
-		return -ENODEV;
-
 	printk(KERN_INFO "PCI: Using configuration type OLPC\n");
 	raw_pci_ops = &pci_olpc_conf;
 	is_lx = is_geode_lx();

arch/x86/pci/visws.c

@@ -69,9 +69,6 @@ void __init pcibios_update_irq(struct pci_dev *dev, int irq)
 
 int __init pci_visws_init(void)
 {
-	if (!is_visws_box())
-		return -1;
-
 	pcibios_enable_irq = &pci_visws_enable_irq;
 	pcibios_disable_irq = &pci_visws_disable_irq;
 
@@ -90,5 +87,6 @@ int __init pci_visws_init(void)
 	pci_scan_bus_with_sysdata(pci_bus1);
 	pci_fixup_irqs(pci_common_swizzle, visws_map_irq);
 	pcibios_resource_survey();
-	return 0;
+	/* Request bus scan */
+	return 1;
 }

drivers/pci/pci.c

@@ -303,6 +303,49 @@ int pci_find_ext_capability(struct pci_dev *dev, int cap)
 }
 EXPORT_SYMBOL_GPL(pci_find_ext_capability);
 
+/**
+ * pci_bus_find_ext_capability - find an extended capability
+ * @bus:   the PCI bus to query
+ * @devfn: PCI device to query
+ * @cap:   capability code
+ *
+ * Like pci_find_ext_capability() but works for pci devices that do not have a
+ * pci_dev structure set up yet.
+ *
+ * Returns the address of the requested capability structure within the
+ * device's PCI configuration space or 0 in case the device does not
+ * support it.
+ */
+int pci_bus_find_ext_capability(struct pci_bus *bus, unsigned int devfn,
+				int cap)
+{
+	u32 header;
+	int ttl;
+	int pos = PCI_CFG_SPACE_SIZE;
+
+	/* minimum 8 bytes per capability */
+	ttl = (PCI_CFG_SPACE_EXP_SIZE - PCI_CFG_SPACE_SIZE) / 8;
+
+	if (!pci_bus_read_config_dword(bus, devfn, pos, &header))
+		return 0;
+	if (header == 0xffffffff || header == 0)
+		return 0;
+
+	while (ttl-- > 0) {
+		if (PCI_EXT_CAP_ID(header) == cap)
+			return pos;
+
+		pos = PCI_EXT_CAP_NEXT(header);
+		if (pos < PCI_CFG_SPACE_SIZE)
+			break;
+
+		if (!pci_bus_read_config_dword(bus, devfn, pos, &header))
+			break;
+	}
+
+	return 0;
+}
+
 static int __pci_find_next_ht_cap(struct pci_dev *dev, int pos, int ht_cap)
 {
 	int rc, ttl = PCI_FIND_CAP_TTL;

include/linux/pci.h

@@ -678,6 +678,8 @@ enum pci_lost_interrupt_reason pci_lost_interrupt(struct pci_dev *dev);
 int pci_find_capability(struct pci_dev *dev, int cap);
 int pci_find_next_capability(struct pci_dev *dev, u8 pos, int cap);
 int pci_find_ext_capability(struct pci_dev *dev, int cap);
+int pci_bus_find_ext_capability(struct pci_bus *bus, unsigned int devfn,
+				int cap);
 int pci_find_ht_capability(struct pci_dev *dev, int ht_cap);
 int pci_find_next_ht_capability(struct pci_dev *dev, int pos, int ht_cap);
 struct pci_bus *pci_find_next_bus(const struct pci_bus *from);

include/linux/pci_regs.h

@@ -507,6 +507,7 @@
 #define PCI_EXT_CAP_ID_VC	2
 #define PCI_EXT_CAP_ID_DSN	3
 #define PCI_EXT_CAP_ID_PWR	4
+#define PCI_EXT_CAP_ID_VNDR	11
 #define PCI_EXT_CAP_ID_ACS	13
 #define PCI_EXT_CAP_ID_ARI	14
 #define PCI_EXT_CAP_ID_ATS	15