Merge ssh://master.kernel.org/pub/scm/linux/kernel/git/tglx/linux-2.6-x86

* ssh://master.kernel.org/pub/scm/linux/kernel/git/tglx/linux-2.6-x86: (114 commits)
  x86: delete vsyscall files during make clean
  kbuild: fix typo SRCARCH in find_sources
  x86: fix kernel rebuild due to vsyscall fallout
  .gitignore update for x86 arch
  x86: unify include/asm/debugreg_32/64.h
  x86: unify include/asm/unwind_32/64.h
  x86: unify include/asm/types_32/64.h
  x86: unify include/asm/tlb_32/64.h
  x86: unify include/asm/siginfo_32/64.h
  x86: unify include/asm/bug_32/64.h
  x86: unify include/asm/mman_32/64.h
  x86: unify include/asm/agp_32/64.h
  x86: unify include/asm/kdebug_32/64.h
  x86: unify include/asm/ioctls_32/64.h
  x86: unify include/asm/floppy_32/64.h
  x86: apply missing DMA/OOM prevention to floppy_32.h
  x86: unify include/asm/cache_32/64.h
  x86: unify include/asm/cache_32/64.h
  x86: unify include/asm/dmi_32/64.h
  x86: unify include/asm/delay_32/64.h
  ...

arch/x86/kernel/mce_amd_64.c

@@ -237,6 +237,7 @@ asmlinkage void mce_threshold_interrupt(void)
 		}
 	}
 out:
+	add_pda(irq_threshold_count, 1);
 	irq_exit();
 }
 

arch/x86/kernel/mce_intel_64.c

@@ -26,6 +26,7 @@ asmlinkage void smp_thermal_interrupt(void)
 	if (therm_throt_process(msr_val & 1))
 		mce_log_therm_throt_event(smp_processor_id(), msr_val);
 
+	add_pda(irq_thermal_count, 1);
 	irq_exit();
 }
 

arch/x86/kernel/msr.c

@@ -133,7 +133,7 @@ static const struct file_operations msr_fops = {
 	.open = msr_open,
 };
 
-static int msr_device_create(int i)
+static int __cpuinit msr_device_create(int i)
 {
 	int err = 0;
 	struct device *dev;
@@ -144,7 +144,7 @@ static int msr_device_create(int i)
 	return err;
 }
 
-static int msr_class_cpu_callback(struct notifier_block *nfb,
+static int __cpuinit msr_class_cpu_callback(struct notifier_block *nfb,
 				unsigned long action, void *hcpu)
 {
 	unsigned int cpu = (unsigned long)hcpu;

arch/x86/kernel/pci-calgary_64.c

@@ -222,10 +222,10 @@ static inline unsigned int num_dma_pages(unsigned long dma, unsigned int dmalen)
 	return npages;
 }
 
-static inline int translate_phb(struct pci_dev* dev)
+static inline int translation_enabled(struct iommu_table *tbl)
 {
-	int disabled = bus_info[dev->bus->number].translation_disabled;
-	return !disabled;
+	/* only PHBs with translation enabled have an IOMMU table */
+	return (tbl != NULL);
 }
 
 static void iommu_range_reserve(struct iommu_table *tbl,
@@ -388,7 +388,7 @@ static void calgary_unmap_sg(struct device *dev,
 	struct scatterlist *s;
 	int i;
 
-	if (!translate_phb(to_pci_dev(dev)))
+	if (!translation_enabled(tbl))
 		return;
 
 	for_each_sg(sglist, s, nelems, i) {
@@ -428,7 +428,7 @@ static int calgary_map_sg(struct device *dev, struct scatterlist *sg,
 	unsigned long entry;
 	int i;
 
-	if (!translate_phb(to_pci_dev(dev)))
+	if (!translation_enabled(tbl))
 		return calgary_nontranslate_map_sg(dev, sg, nelems, direction);
 
 	for_each_sg(sg, s, nelems, i) {
@@ -474,7 +474,7 @@ static dma_addr_t calgary_map_single(struct device *dev, void *vaddr,
 	uaddr = (unsigned long)vaddr;
 	npages = num_dma_pages(uaddr, size);
 
-	if (translate_phb(to_pci_dev(dev)))
+	if (translation_enabled(tbl))
 		dma_handle = iommu_alloc(tbl, vaddr, npages, direction);
 	else
 		dma_handle = virt_to_bus(vaddr);
@@ -488,7 +488,7 @@ static void calgary_unmap_single(struct device *dev, dma_addr_t dma_handle,
 	struct iommu_table *tbl = find_iommu_table(dev);
 	unsigned int npages;
 
-	if (!translate_phb(to_pci_dev(dev)))
+	if (!translation_enabled(tbl))
 		return;
 
 	npages = num_dma_pages(dma_handle, size);
@@ -513,7 +513,7 @@ static void* calgary_alloc_coherent(struct device *dev, size_t size,
 		goto error;
 	memset(ret, 0, size);
 
-	if (translate_phb(to_pci_dev(dev))) {
+	if (translation_enabled(tbl)) {
 		/* set up tces to cover the allocated range */
 		mapping = iommu_alloc(tbl, ret, npages, DMA_BIDIRECTIONAL);
 		if (mapping == bad_dma_address)
@@ -1194,7 +1194,7 @@ static int __init calgary_init(void)
 {
 	int ret;
 	struct pci_dev *dev = NULL;
-	void *tce_space;
+	struct calgary_bus_info *info;
 
 	ret = calgary_locate_bbars();
 	if (ret)
@@ -1206,12 +1206,14 @@ static int __init calgary_init(void)
 			break;
 		if (!is_cal_pci_dev(dev->device))
 			continue;
-		if (!translate_phb(dev)) {
+
+		info = &bus_info[dev->bus->number];
+		if (info->translation_disabled) {
 			calgary_init_one_nontraslated(dev);
 			continue;
 		}
-		tce_space = bus_info[dev->bus->number].tce_space;
-		if (!tce_space && !translate_empty_slots)
+
+		if (!info->tce_space && !translate_empty_slots)
 			continue;
 
 		ret = calgary_init_one(dev);
@@ -1229,11 +1231,13 @@ static int __init calgary_init(void)
 			break;
 		if (!is_cal_pci_dev(dev->device))
 			continue;
-		if (!translate_phb(dev)) {
+
+		info = &bus_info[dev->bus->number];
+		if (info->translation_disabled) {
 			pci_dev_put(dev);
 			continue;
 		}
-		if (!bus_info[dev->bus->number].tce_space && !translate_empty_slots)
+		if (!info->tce_space && !translate_empty_slots)
 			continue;
 
 		calgary_disable_translation(dev);
@@ -1546,7 +1550,7 @@ static void __init calgary_fixup_one_tce_space(struct pci_dev *dev)
 static int __init calgary_fixup_tce_spaces(void)
 {
 	struct pci_dev *dev = NULL;
-	void *tce_space;
+	struct calgary_bus_info *info;
 
 	if (no_iommu || swiotlb || !calgary_detected)
 		return -ENODEV;
@@ -1559,11 +1563,12 @@ static int __init calgary_fixup_tce_spaces(void)
 			break;
 		if (!is_cal_pci_dev(dev->device))
 			continue;
-		if (!translate_phb(dev))
+
+		info = &bus_info[dev->bus->number];
+		if (info->translation_disabled)
 			continue;
 
-		tce_space = bus_info[dev->bus->number].tce_space;
-		if (!tce_space)
+		if (!info->tce_space)
 			continue;
 
 		calgary_fixup_one_tce_space(dev);

arch/x86/kernel/pci-dma_32.c

@@ -12,7 +12,6 @@
 #include <linux/string.h>
 #include <linux/pci.h>
 #include <linux/module.h>
-#include <linux/pci.h>
 #include <asm/io.h>
 
 struct dma_coherent_mem {

arch/x86/kernel/pci-dma_64.c

@@ -51,11 +51,9 @@ dma_alloc_pages(struct device *dev, gfp_t gfp, unsigned order)
 {
 	struct page *page;
 	int node;
-#ifdef CONFIG_PCI
-	if (dev->bus == &pci_bus_type)
-		node = pcibus_to_node(to_pci_dev(dev)->bus);
-	else
-#endif
+
+	node = dev_to_node(dev);
+	if (node == -1)
 		node = numa_node_id();
 
 	if (node < first_node(node_online_map))

arch/x86/kernel/pci-gart_64.c

@@ -8,6 +8,7 @@
  * See Documentation/DMA-mapping.txt for the interface specification.
  * 
  * Copyright 2002 Andi Kleen, SuSE Labs.
+ * Subject to the GNU General Public License v2 only.
  */
 
 #include <linux/types.h>
@@ -375,7 +376,8 @@ static inline int dma_map_cont(struct scatterlist *start, int nelems,
  * DMA map all entries in a scatterlist.
  * Merge chunks that have page aligned sizes into a continuous mapping. 
  */
-int gart_map_sg(struct device *dev, struct scatterlist *sg, int nents, int dir)
+static int gart_map_sg(struct device *dev, struct scatterlist *sg, int nents,
+			int dir)
 {
 	int i;
 	int out;

arch/x86/kernel/ptrace_32.c

@@ -165,7 +165,7 @@ static unsigned long convert_eip_to_linear(struct task_struct *child, struct pt_
 
 		seg &= ~7UL;
 
-		down(&child->mm->context.sem);
+		mutex_lock(&child->mm->context.lock);
 		if (unlikely((seg >> 3) >= child->mm->context.size))
 			addr = -1L; /* bogus selector, access would fault */
 		else {
@@ -179,7 +179,7 @@ static unsigned long convert_eip_to_linear(struct task_struct *child, struct pt_
 				addr &= 0xffff;
 			addr += base;
 		}
-		up(&child->mm->context.sem);
+		mutex_unlock(&child->mm->context.lock);
 	}
 	return addr;
 }

arch/x86/kernel/ptrace_64.c

@@ -103,7 +103,7 @@ unsigned long convert_rip_to_linear(struct task_struct *child, struct pt_regs *r
 
 		seg &= ~7UL;
 
-		down(&child->mm->context.sem);
+		mutex_lock(&child->mm->context.lock);
 		if (unlikely((seg >> 3) >= child->mm->context.size))
 			addr = -1L; /* bogus selector, access would fault */
 		else {
@@ -117,7 +117,7 @@ unsigned long convert_rip_to_linear(struct task_struct *child, struct pt_regs *r
 				addr &= 0xffff;
 			addr += base;
 		}
-		up(&child->mm->context.sem);
+		mutex_unlock(&child->mm->context.lock);
 	}
 
 	return addr;

arch/x86/kernel/setup_64.c

@@ -604,7 +604,7 @@ static void __cpuinit init_amd(struct cpuinfo_x86 *c)
 	level = cpuid_eax(1);
 	if (c->x86 == 15 && ((level >= 0x0f48 && level < 0x0f50) || level >= 0x0f58))
 		set_bit(X86_FEATURE_REP_GOOD, &c->x86_capability);
-	if (c->x86 == 0x10)
+	if (c->x86 == 0x10 || c->x86 == 0x11)
 		set_bit(X86_FEATURE_REP_GOOD, &c->x86_capability);
 
 	/* Enable workaround for FXSAVE leak */
@@ -968,7 +968,7 @@ static int show_cpuinfo(struct seq_file *m, void *v)
 	 * applications want to get the raw CPUID data, they should access
 	 * /dev/cpu/<cpu_nr>/cpuid instead.
 	 */
-	static char *x86_cap_flags[] = {
+	static const char *const x86_cap_flags[] = {
 		/* Intel-defined */
 	        "fpu", "vme", "de", "pse", "tsc", "msr", "pae", "mce",
 	        "cx8", "apic", NULL, "sep", "mtrr", "pge", "mca", "cmov",
@@ -1022,7 +1022,7 @@ static int show_cpuinfo(struct seq_file *m, void *v)
 		NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
 		NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
 	};
-	static char *x86_power_flags[] = { 
+	static const char *const x86_power_flags[] = {
 		"ts",	/* temperature sensor */
 		"fid",  /* frequency id control */
 		"vid",  /* voltage id control */

arch/x86/kernel/signal_32.c

@@ -385,7 +385,6 @@ static int setup_frame(int sig, struct k_sigaction *ka,
 	regs->edx = (unsigned long) 0;
 	regs->ecx = (unsigned long) 0;
 
-	set_fs(USER_DS);
 	regs->xds = __USER_DS;
 	regs->xes = __USER_DS;
 	regs->xss = __USER_DS;
@@ -479,7 +478,6 @@ static int setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info,
 	regs->edx = (unsigned long) &frame->info;
 	regs->ecx = (unsigned long) &frame->uc;
 
-	set_fs(USER_DS);
 	regs->xds = __USER_DS;
 	regs->xes = __USER_DS;
 	regs->xss = __USER_DS;

arch/x86/kernel/smp_32.c

@@ -342,6 +342,7 @@ fastcall void smp_invalidate_interrupt(struct pt_regs *regs)
 	smp_mb__after_clear_bit();
 out:
 	put_cpu_no_resched();
+	__get_cpu_var(irq_stat).irq_tlb_count++;
 }
 
 void native_flush_tlb_others(const cpumask_t *cpumaskp, struct mm_struct *mm,
@@ -640,6 +641,7 @@ static void native_smp_send_stop(void)
 fastcall void smp_reschedule_interrupt(struct pt_regs *regs)
 {
 	ack_APIC_irq();
+	__get_cpu_var(irq_stat).irq_resched_count++;
 }
 
 fastcall void smp_call_function_interrupt(struct pt_regs *regs)
@@ -660,6 +662,7 @@ fastcall void smp_call_function_interrupt(struct pt_regs *regs)
 	 */
 	irq_enter();
 	(*func)(info);
+	__get_cpu_var(irq_stat).irq_call_count++;
 	irq_exit();
 
 	if (wait) {
@@ -705,3 +708,10 @@ struct smp_ops smp_ops = {
 	.smp_send_reschedule = native_smp_send_reschedule,
 	.smp_call_function_mask = native_smp_call_function_mask,
 };
+
+int smp_call_function_mask(cpumask_t mask, void (*func) (void *info),
+			   void *info, int wait)
+{
+	return smp_ops.smp_call_function_mask(mask, func, info, wait);
+}
+EXPORT_SYMBOL(smp_call_function_mask);

arch/x86/kernel/smp_64.c

@@ -163,6 +163,7 @@ asmlinkage void smp_invalidate_interrupt(struct pt_regs *regs)
 out:
 	ack_APIC_irq();
 	cpu_clear(cpu, f->flush_cpumask);
+	add_pda(irq_tlb_count, 1);
 }
 
 static void flush_tlb_others(cpumask_t cpumask, struct mm_struct *mm,
@@ -493,6 +494,7 @@ void smp_send_stop(void)
 asmlinkage void smp_reschedule_interrupt(void)
 {
 	ack_APIC_irq();
+	add_pda(irq_resched_count, 1);
 }
 
 asmlinkage void smp_call_function_interrupt(void)
@@ -514,6 +516,7 @@ asmlinkage void smp_call_function_interrupt(void)
 	exit_idle();
 	irq_enter();
 	(*func)(info);
+	add_pda(irq_call_count, 1);
 	irq_exit();
 	if (wait) {
 		mb();

arch/x86/kernel/smpboot_32.c

@@ -102,8 +102,8 @@ u8 apicid_2_node[MAX_APICID];
  * Trampoline 80x86 program as an array.
  */
 
-extern unsigned char trampoline_data [];
-extern unsigned char trampoline_end  [];
+extern const unsigned char trampoline_data [];
+extern const unsigned char trampoline_end  [];
 static unsigned char *trampoline_base;
 static int trampoline_exec;
 
@@ -118,7 +118,7 @@ DEFINE_PER_CPU(int, cpu_state) = { 0 };
  * has made sure it's suitably aligned.
  */
 
-static unsigned long __devinit setup_trampoline(void)
+static unsigned long __cpuinit setup_trampoline(void)
 {
 	memcpy(trampoline_base, trampoline_data, trampoline_end - trampoline_data);
 	return virt_to_phys(trampoline_base);
@@ -1021,6 +1021,12 @@ static void __init smp_boot_cpus(unsigned int max_cpus)
 	if (!max_cpus) {
 		smp_found_config = 0;
 		printk(KERN_INFO "SMP mode deactivated, forcing use of dummy APIC emulation.\n");
+
+		if (nmi_watchdog == NMI_LOCAL_APIC) {
+			printk(KERN_INFO "activating minimal APIC for NMI watchdog use.\n");
+			connect_bsp_APIC();
+			setup_local_APIC();
+		}
 		smpboot_clear_io_apic_irqs();
 		phys_cpu_present_map = physid_mask_of_physid(0);
 		cpu_set(0, per_cpu(cpu_sibling_map, 0));

arch/x86/kernel/smpboot_64.c

@@ -102,8 +102,8 @@ EXPORT_PER_CPU_SYMBOL(cpu_core_map);
  * Trampoline 80x86 program as an array.
  */
 
-extern unsigned char trampoline_data[];
-extern unsigned char trampoline_end[];
+extern const unsigned char trampoline_data[];
+extern const unsigned char trampoline_end[];
 
 /* State of each CPU */
 DEFINE_PER_CPU(int, cpu_state) = { 0 };
@@ -695,7 +695,6 @@ static int __cpuinit do_boot_cpu(int cpu, int apicid)
 		cpu_clear(cpu, cpu_present_map);
 		cpu_clear(cpu, cpu_possible_map);
 		x86_cpu_to_apicid[cpu] = BAD_APICID;
-		x86_cpu_to_log_apicid[cpu] = BAD_APICID;
 		return -EIO;
 	}
 

arch/x86/kernel/stacktrace.c

@@ -33,7 +33,7 @@ static void save_stack_address(void *data, unsigned long addr)
 		trace->entries[trace->nr_entries++] = addr;
 }
 
-static struct stacktrace_ops save_stack_ops = {
+static const struct stacktrace_ops save_stack_ops = {
 	.warning = save_stack_warning,
 	.warning_symbol = save_stack_warning_symbol,
 	.stack = save_stack_stack,

arch/x86/kernel/tce_64.c

@@ -40,9 +40,9 @@ static inline void flush_tce(void* tceaddr)
 {
 	/* a single tce can't cross a cache line */
 	if (cpu_has_clflush)
-		asm volatile("clflush (%0)" :: "r" (tceaddr));
+		clflush(tceaddr);
 	else
-		asm volatile("wbinvd":::"memory");
+		wbinvd();
 }
 
 void tce_build(struct iommu_table *tbl, unsigned long index,

arch/x86/kernel/topology.c

@@ -44,15 +44,15 @@ int arch_register_cpu(int num)
 	 * Also certain PCI quirks require not to enable hotplug control
 	 * for all CPU's.
 	 */
-	if (num && enable_cpu_hotplug)
+#ifdef CONFIG_HOTPLUG_CPU
+	if (num)
 		cpu_devices[num].cpu.hotpluggable = 1;
+#endif
 
 	return register_cpu(&cpu_devices[num].cpu, num);
 }
 
 #ifdef CONFIG_HOTPLUG_CPU
-int enable_cpu_hotplug = 1;
-
 void arch_unregister_cpu(int num) {
 	return unregister_cpu(&cpu_devices[num].cpu);
 }

arch/x86/kernel/trampoline_32.S

@@ -36,11 +36,11 @@
 #include <asm/segment.h>
 #include <asm/page.h>
 
-.data
-
 /* We can free up trampoline after bootup if cpu hotplug is not supported. */
 #ifndef CONFIG_HOTPLUG_CPU
 .section ".init.data","aw",@progbits
+#else
+.section .rodata,"a",@progbits
 #endif
 
 .code16

arch/x86/kernel/trampoline_64.S

@@ -33,7 +33,12 @@
 #include <asm/msr.h>
 #include <asm/segment.h>
 
-.data
+/* We can free up trampoline after bootup if cpu hotplug is not supported. */
+#ifndef CONFIG_HOTPLUG_CPU
+.section .init.data, "aw", @progbits
+#else
+.section .rodata, "a", @progbits
+#endif
 
 .code16