Merge branch 'linus' into x86/x2apic

Conflicts:
	arch/x86/kernel/cpu/cyrix.c
	include/asm-x86/cpufeature.h

Signed-off-by: Ingo Molnar <mingo@elte.hu>

42 files changed, 590 insertions, 322 deletions

diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c
index 12e260e8fb2a..27ef365e757d 100644
--- a/arch/x86/kernel/acpi/boot.c
+++ b/arch/x86/kernel/acpi/boot.c
@@ -97,6 +97,8 @@ static u64 acpi_lapic_addr __initdata = APIC_DEFAULT_PHYS_BASE;
 #warning ACPI uses CMPXCHG, i486 and later hardware
 #endif
 
+static int acpi_mcfg_64bit_base_addr __initdata = FALSE;
+
 /* --------------------------------------------------------------------------
                               Boot-time Configuration
    -------------------------------------------------------------------------- */
@@ -158,6 +160,14 @@ char *__init __acpi_map_table(unsigned long phys, unsigned long size)
 struct acpi_mcfg_allocation *pci_mmcfg_config;
 int pci_mmcfg_config_num;
 
+static int __init acpi_mcfg_oem_check(struct acpi_table_mcfg *mcfg)
+{
+	if (!strcmp(mcfg->header.oem_id, "SGI"))
+		acpi_mcfg_64bit_base_addr = TRUE;
+
+	return 0;
+}
+
 int __init acpi_parse_mcfg(struct acpi_table_header *header)
 {
 	struct acpi_table_mcfg *mcfg;
@@ -190,8 +200,12 @@ int __init acpi_parse_mcfg(struct acpi_table_header *header)
 	}
 
 	memcpy(pci_mmcfg_config, &mcfg[1], config_size);
+
+	acpi_mcfg_oem_check(mcfg);
+
 	for (i = 0; i < pci_mmcfg_config_num; ++i) {
-		if (pci_mmcfg_config[i].address > 0xFFFFFFFF) {
+		if ((pci_mmcfg_config[i].address > 0xFFFFFFFF) &&
+		    !acpi_mcfg_64bit_base_addr) {
 			printk(KERN_ERR PREFIX
 			       "MMCONFIG not in low 4GB of memory\n");
 			kfree(pci_mmcfg_config);
diff --git a/arch/x86/kernel/acpi/sleep.c b/arch/x86/kernel/acpi/sleep.c
index fa2161d5003b..426e5d91b63a 100644
--- a/arch/x86/kernel/acpi/sleep.c
+++ b/arch/x86/kernel/acpi/sleep.c
@@ -20,7 +20,7 @@ unsigned long acpi_realmode_flags;
 /* address in low memory of the wakeup routine. */
 static unsigned long acpi_realmode;
 
-#ifdef CONFIG_64BIT
+#if defined(CONFIG_SMP) && defined(CONFIG_64BIT)
 static char temp_stack[10240];
 #endif
 
@@ -86,7 +86,7 @@ int acpi_save_state_mem(void)
 #endif /* !CONFIG_64BIT */
 
 	header->pmode_cr0 = read_cr0();
-	header->pmode_cr4 = read_cr4();
+	header->pmode_cr4 = read_cr4_safe();
 	header->realmode_flags = acpi_realmode_flags;
 	header->real_magic = 0x12345678;
 
diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c
index 22d7d050905d..69b4d060b21c 100644
--- a/arch/x86/kernel/amd_iommu.c
+++ b/arch/x86/kernel/amd_iommu.c
@@ -65,7 +65,7 @@ static int __iommu_queue_command(struct amd_iommu *iommu, struct iommu_cmd *cmd)
 	u8 *target;
 
 	tail = readl(iommu->mmio_base + MMIO_CMD_TAIL_OFFSET);
-	target = (iommu->cmd_buf + tail);
+	target = iommu->cmd_buf + tail;
 	memcpy_toio(target, cmd, sizeof(*cmd));
 	tail = (tail + sizeof(*cmd)) % iommu->cmd_buf_size;
 	head = readl(iommu->mmio_base + MMIO_CMD_HEAD_OFFSET);
@@ -101,16 +101,13 @@ static int iommu_queue_command(struct amd_iommu *iommu, struct iommu_cmd *cmd)
  */
 static int iommu_completion_wait(struct amd_iommu *iommu)
 {
-	int ret;
+	int ret, ready = 0;
+	unsigned status = 0;
 	struct iommu_cmd cmd;
-	volatile u64 ready = 0;
-	unsigned long ready_phys = virt_to_phys(&ready);
 	unsigned long i = 0;
 
 	memset(&cmd, 0, sizeof(cmd));
-	cmd.data[0] = LOW_U32(ready_phys) | CMD_COMPL_WAIT_STORE_MASK;
-	cmd.data[1] = upper_32_bits(ready_phys);
-	cmd.data[2] = 1; /* value written to 'ready' */
+	cmd.data[0] = CMD_COMPL_WAIT_INT_MASK;
 	CMD_SET_TYPE(&cmd, CMD_COMPL_WAIT);
 
 	iommu->need_sync = 0;
@@ -122,9 +119,15 @@ static int iommu_completion_wait(struct amd_iommu *iommu)
 
 	while (!ready && (i < EXIT_LOOP_COUNT)) {
 		++i;
-		cpu_relax();
+		/* wait for the bit to become one */
+		status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET);
+		ready = status & MMIO_STATUS_COM_WAIT_INT_MASK;
 	}
 
+	/* set bit back to zero */
+	status &= ~MMIO_STATUS_COM_WAIT_INT_MASK;
+	writel(status, iommu->mmio_base + MMIO_STATUS_OFFSET);
+
 	if (unlikely((i == EXIT_LOOP_COUNT) && printk_ratelimit()))
 		printk(KERN_WARNING "AMD IOMMU: Completion wait loop failed\n");
 
@@ -161,7 +164,7 @@ static int iommu_queue_inv_iommu_pages(struct amd_iommu *iommu,
 	address &= PAGE_MASK;
 	CMD_SET_TYPE(&cmd, CMD_INV_IOMMU_PAGES);
 	cmd.data[1] |= domid;
-	cmd.data[2] = LOW_U32(address);
+	cmd.data[2] = lower_32_bits(address);
 	cmd.data[3] = upper_32_bits(address);
 	if (s) /* size bit - we flush more than one 4kb page */
 		cmd.data[2] |= CMD_INV_IOMMU_PAGES_SIZE_MASK;
diff --git a/arch/x86/kernel/amd_iommu_init.c b/arch/x86/kernel/amd_iommu_init.c
index d9a9da597e79..a69cc0f52042 100644
--- a/arch/x86/kernel/amd_iommu_init.c
+++ b/arch/x86/kernel/amd_iommu_init.c
@@ -801,6 +801,21 @@ static int __init init_memory_definitions(struct acpi_table_header *table)
 }
 
 /*
+ * Init the device table to not allow DMA access for devices and
+ * suppress all page faults
+ */
+static void init_device_table(void)
+{
+	u16 devid;
+
+	for (devid = 0; devid <= amd_iommu_last_bdf; ++devid) {
+		set_dev_entry_bit(devid, DEV_ENTRY_VALID);
+		set_dev_entry_bit(devid, DEV_ENTRY_TRANSLATION);
+		set_dev_entry_bit(devid, DEV_ENTRY_NO_PAGE_FAULT);
+	}
+}
+
+/*
  * This function finally enables all IOMMUs found in the system after
  * they have been initialized
  */
@@ -931,6 +946,9 @@ int __init amd_iommu_init(void)
 	if (amd_iommu_pd_alloc_bitmap == NULL)
 		goto free;
 
+	/* init the device table */
+	init_device_table();
+
 	/*
 	 * let all alias entries point to itself
 	 */
@@ -954,15 +972,15 @@ int __init amd_iommu_init(void)
 	if (acpi_table_parse("IVRS", init_memory_definitions) != 0)
 		goto free;
 
-	ret = amd_iommu_init_dma_ops();
+	ret = sysdev_class_register(&amd_iommu_sysdev_class);
 	if (ret)
 		goto free;
 
-	ret = sysdev_class_register(&amd_iommu_sysdev_class);
+	ret = sysdev_register(&device_amd_iommu);
 	if (ret)
 		goto free;
 
-	ret = sysdev_register(&device_amd_iommu);
+	ret = amd_iommu_init_dma_ops();
 	if (ret)
 		goto free;
 
diff --git a/arch/x86/kernel/apic_32.c b/arch/x86/kernel/apic_32.c
index 8228222ec917..44cae65e32ef 100644
--- a/arch/x86/kernel/apic_32.c
+++ b/arch/x86/kernel/apic_32.c
@@ -1487,8 +1487,6 @@ void disconnect_bsp_APIC(int virt_wire_setup)
 	}
 }
 
-unsigned int __cpuinitdata maxcpus = NR_CPUS;
-
 void __cpuinit generic_processor_info(int apicid, int version)
 {
 	int cpu;
@@ -1515,12 +1513,6 @@ void __cpuinit generic_processor_info(int apicid, int version)
 		return;
 	}
 
-	if (num_processors >= maxcpus) {
-		printk(KERN_WARNING "WARNING: maxcpus limit of %i reached."
-			" Processor ignored.\n", maxcpus);
-		return;
-	}
-
 	num_processors++;
 	cpus_complement(tmp_map, cpu_present_map);
 	cpu = first_cpu(tmp_map);
diff --git a/arch/x86/kernel/apic_64.c b/arch/x86/kernel/apic_64.c
index cd63c0bc6180..b6256587f99e 100644
--- a/arch/x86/kernel/apic_64.c
+++ b/arch/x86/kernel/apic_64.c
@@ -97,7 +97,6 @@ static unsigned long apic_phys;
 
 unsigned long mp_lapic_addr;
 
-unsigned int __cpuinitdata maxcpus = NR_CPUS;
 /*
  * Get the LAPIC version
  */
@@ -1259,12 +1258,6 @@ void __cpuinit generic_processor_info(int apicid, int version)
 		return;
 	}
 
-	if (num_processors >= maxcpus) {
-		printk(KERN_WARNING "WARNING: maxcpus limit of %i reached."
-		       " Processor ignored.\n", maxcpus);
-		return;
-	}
-
 	num_processors++;
 	cpus_complement(tmp_map, cpu_present_map);
 	cpu = first_cpu(tmp_map);
diff --git a/arch/x86/kernel/cpu/addon_cpuid_features.c b/arch/x86/kernel/cpu/addon_cpuid_features.c
index a2d1a545767d..0d9c993aa93e 100644
--- a/arch/x86/kernel/cpu/addon_cpuid_features.c
+++ b/arch/x86/kernel/cpu/addon_cpuid_features.c
@@ -144,9 +144,22 @@ void __cpuinit validate_pat_support(struct cpuinfo_x86 *c)
 
 	switch (c->x86_vendor) {
 	case X86_VENDOR_INTEL:
-		if (c->x86 == 0xF || (c->x86 == 6 && c->x86_model >= 15))
+		/*
+		 * There is a known erratum on Pentium III and Core Solo
+		 * and Core Duo CPUs.
+		 * " Page with PAT set to WC while associated MTRR is UC
+		 *   may consolidate to UC "
+		 * Because of this erratum, it is better to stick with
+		 * setting WC in MTRR rather than using PAT on these CPUs.
+		 *
+		 * Enable PAT WC only on P4, Core 2 or later CPUs.
+		 */
+		if (c->x86 > 0x6 || (c->x86 == 6 && c->x86_model >= 15))
 			return;
-		break;
+
+		pat_disable("PAT WC disabled due to known CPU erratum.");
+		return;
+
 	case X86_VENDOR_AMD:
 	case X86_VENDOR_CENTAUR:
 	case X86_VENDOR_TRANSMETA:
diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
index 4e7271999a74..84bb395038d8 100644
--- a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
+++ b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
@@ -737,63 +737,44 @@ static int find_psb_table(struct powernow_k8_data *data)
 #ifdef CONFIG_X86_POWERNOW_K8_ACPI
 static void powernow_k8_acpi_pst_values(struct powernow_k8_data *data, unsigned int index)
 {
-	if (!data->acpi_data->state_count || (cpu_family == CPU_HW_PSTATE))
+	if (!data->acpi_data.state_count || (cpu_family == CPU_HW_PSTATE))
 		return;
 
-	data->irt = (data->acpi_data->states[index].control >> IRT_SHIFT) & IRT_MASK;
-	data->rvo = (data->acpi_data->states[index].control >> RVO_SHIFT) & RVO_MASK;
-	data->exttype = (data->acpi_data->states[index].control >> EXT_TYPE_SHIFT) & EXT_TYPE_MASK;
-	data->plllock = (data->acpi_data->states[index].control >> PLL_L_SHIFT) & PLL_L_MASK;
-	data->vidmvs = 1 << ((data->acpi_data->states[index].control >> MVS_SHIFT) & MVS_MASK);
-	data->vstable = (data->acpi_data->states[index].control >> VST_SHIFT) & VST_MASK;
-}
-
-
-static struct acpi_processor_performance *acpi_perf_data;
-static int preregister_valid;
-
-static int powernow_k8_cpu_preinit_acpi(void)
-{
-	acpi_perf_data = alloc_percpu(struct acpi_processor_performance);
-	if (!acpi_perf_data)
-		return -ENODEV;
-
-	if (acpi_processor_preregister_performance(acpi_perf_data))
-		return -ENODEV;
-	else
-		preregister_valid = 1;
-	return 0;
+	data->irt = (data->acpi_data.states[index].control >> IRT_SHIFT) & IRT_MASK;
+	data->rvo = (data->acpi_data.states[index].control >> RVO_SHIFT) & RVO_MASK;
+	data->exttype = (data->acpi_data.states[index].control >> EXT_TYPE_SHIFT) & EXT_TYPE_MASK;
+	data->plllock = (data->acpi_data.states[index].control >> PLL_L_SHIFT) & PLL_L_MASK;
+	data->vidmvs = 1 << ((data->acpi_data.states[index].control >> MVS_SHIFT) & MVS_MASK);
+	data->vstable = (data->acpi_data.states[index].control >> VST_SHIFT) & VST_MASK;
 }
 
 static int powernow_k8_cpu_init_acpi(struct powernow_k8_data *data)
 {
 	struct cpufreq_frequency_table *powernow_table;
 	int ret_val;
-	int cpu = 0;
 
-	data->acpi_data = percpu_ptr(acpi_perf_data, cpu);
-	if (acpi_processor_register_performance(data->acpi_data, data->cpu)) {
+	if (acpi_processor_register_performance(&data->acpi_data, data->cpu)) {
 		dprintk("register performance failed: bad ACPI data\n");
 		return -EIO;
 	}
 
 	/* verify the data contained in the ACPI structures */
-	if (data->acpi_data->state_count <= 1) {
+	if (data->acpi_data.state_count <= 1) {
 		dprintk("No ACPI P-States\n");
 		goto err_out;
 	}
 
-	if ((data->acpi_data->control_register.space_id != ACPI_ADR_SPACE_FIXED_HARDWARE) ||
-		(data->acpi_data->status_register.space_id != ACPI_ADR_SPACE_FIXED_HARDWARE)) {
+	if ((data->acpi_data.control_register.space_id != ACPI_ADR_SPACE_FIXED_HARDWARE) ||
+		(data->acpi_data.status_register.space_id != ACPI_ADR_SPACE_FIXED_HARDWARE)) {
 		dprintk("Invalid control/status registers (%x - %x)\n",
-			data->acpi_data->control_register.space_id,
-			data->acpi_data->status_register.space_id);
+			data->acpi_data.control_register.space_id,
+			data->acpi_data.status_register.space_id);
 		goto err_out;
 	}
 
 	/* fill in data->powernow_table */
 	powernow_table = kmalloc((sizeof(struct cpufreq_frequency_table)
-		* (data->acpi_data->state_count + 1)), GFP_KERNEL);
+		* (data->acpi_data.state_count + 1)), GFP_KERNEL);
 	if (!powernow_table) {
 		dprintk("powernow_table memory alloc failure\n");
 		goto err_out;
@@ -806,12 +787,12 @@ static int powernow_k8_cpu_init_acpi(struct powernow_k8_data *data)
 	if (ret_val)
 		goto err_out_mem;
 
-	powernow_table[data->acpi_data->state_count].frequency = CPUFREQ_TABLE_END;
-	powernow_table[data->acpi_data->state_count].index = 0;
+	powernow_table[data->acpi_data.state_count].frequency = CPUFREQ_TABLE_END;
+	powernow_table[data->acpi_data.state_count].index = 0;
 	data->powernow_table = powernow_table;
 
 	/* fill in data */
-	data->numps = data->acpi_data->state_count;
+	data->numps = data->acpi_data.state_count;
 	if (first_cpu(per_cpu(cpu_core_map, data->cpu)) == data->cpu)
 		print_basics(data);
 	powernow_k8_acpi_pst_values(data, 0);
@@ -819,31 +800,16 @@ static int powernow_k8_cpu_init_acpi(struct powernow_k8_data *data)
 	/* notify BIOS that we exist */
 	acpi_processor_notify_smm(THIS_MODULE);
 
-	/* determine affinity, from ACPI if available */
-	if (preregister_valid) {
-		if ((data->acpi_data->shared_type == CPUFREQ_SHARED_TYPE_ALL) ||
-			(data->acpi_data->shared_type == CPUFREQ_SHARED_TYPE_ANY))
-			data->starting_core_affinity = data->acpi_data->shared_cpu_map;
-		else
-			data->starting_core_affinity = cpumask_of_cpu(data->cpu);
-	} else {
-	/* best guess from family if not */
-		if (cpu_family == CPU_HW_PSTATE)
-			data->starting_core_affinity = cpumask_of_cpu(data->cpu);
-		else
-			data->starting_core_affinity = per_cpu(cpu_core_map, data->cpu);
-	}
-
 	return 0;
 
 err_out_mem:
 	kfree(powernow_table);
 
 err_out:
-	acpi_processor_unregister_performance(data->acpi_data, data->cpu);
+	acpi_processor_unregister_performance(&data->acpi_data, data->cpu);
 
 	/* data->acpi_data.state_count informs us at ->exit() whether ACPI was used */
-	data->acpi_data->state_count = 0;
+	data->acpi_data.state_count = 0;
 
 	return -ENODEV;
 }
@@ -855,10 +821,10 @@ static int fill_powernow_table_pstate(struct powernow_k8_data *data, struct cpuf
 	rdmsr(MSR_PSTATE_CUR_LIMIT, hi, lo);
 	data->max_hw_pstate = (hi & HW_PSTATE_MAX_MASK) >> HW_PSTATE_MAX_SHIFT;
 
-	for (i = 0; i < data->acpi_data->state_count; i++) {
+	for (i = 0; i < data->acpi_data.state_count; i++) {
 		u32 index;
 
-		index = data->acpi_data->states[i].control & HW_PSTATE_MASK;
+		index = data->acpi_data.states[i].control & HW_PSTATE_MASK;
 		if (index > data->max_hw_pstate) {
 			printk(KERN_ERR PFX "invalid pstate %d - bad value %d.\n", i, index);
 			printk(KERN_ERR PFX "Please report to BIOS manufacturer\n");
@@ -874,7 +840,7 @@ static int fill_powernow_table_pstate(struct powernow_k8_data *data, struct cpuf
 
 		powernow_table[i].index = index;
 
-		powernow_table[i].frequency = data->acpi_data->states[i].core_frequency * 1000;
+		powernow_table[i].frequency = data->acpi_data.states[i].core_frequency * 1000;
 	}
 	return 0;
 }
@@ -883,16 +849,16 @@ static int fill_powernow_table_fidvid(struct powernow_k8_data *data, struct cpuf
 {
 	int i;
 	int cntlofreq = 0;
-	for (i = 0; i < data->acpi_data->state_count; i++) {
+	for (i = 0; i < data->acpi_data.state_count; i++) {
 		u32 fid;
 		u32 vid;
 
 		if (data->exttype) {
-			fid = data->acpi_data->states[i].status & EXT_FID_MASK;
-			vid = (data->acpi_data->states[i].status >> VID_SHIFT) & EXT_VID_MASK;
+			fid = data->acpi_data.states[i].status & EXT_FID_MASK;
+			vid = (data->acpi_data.states[i].status >> VID_SHIFT) & EXT_VID_MASK;
 		} else {
-			fid = data->acpi_data->states[i].control & FID_MASK;
-			vid = (data->acpi_data->states[i].control >> VID_SHIFT) & VID_MASK;
+			fid = data->acpi_data.states[i].control & FID_MASK;
+			vid = (data->acpi_data.states[i].control >> VID_SHIFT) & VID_MASK;
 		}
 
 		dprintk("   %d : fid 0x%x, vid 0x%x\n", i, fid, vid);
@@ -933,10 +899,10 @@ static int fill_powernow_table_fidvid(struct powernow_k8_data *data, struct cpuf
 				cntlofreq = i;
 		}
 
-		if (powernow_table[i].frequency != (data->acpi_data->states[i].core_frequency * 1000)) {
+		if (powernow_table[i].frequency != (data->acpi_data.states[i].core_frequency * 1000)) {
 			printk(KERN_INFO PFX "invalid freq entries %u kHz vs. %u kHz\n",
 				powernow_table[i].frequency,
-				(unsigned int) (data->acpi_data->states[i].core_frequency * 1000));
+				(unsigned int) (data->acpi_data.states[i].core_frequency * 1000));
 			powernow_table[i].frequency = CPUFREQ_ENTRY_INVALID;
 			continue;
 		}
@@ -946,12 +912,11 @@ static int fill_powernow_table_fidvid(struct powernow_k8_data *data, struct cpuf
 
 static void powernow_k8_cpu_exit_acpi(struct powernow_k8_data *data)
 {
-	if (data->acpi_data->state_count)
-		acpi_processor_unregister_performance(data->acpi_data, data->cpu);
+	if (data->acpi_data.state_count)
+		acpi_processor_unregister_performance(&data->acpi_data, data->cpu);
 }
 
 #else
-static int powernow_k8_cpu_preinit_acpi(void) { return -ENODEV; }
 static int powernow_k8_cpu_init_acpi(struct powernow_k8_data *data) { return -ENODEV; }
 static void powernow_k8_cpu_exit_acpi(struct powernow_k8_data *data) { return; }
 static void powernow_k8_acpi_pst_values(struct powernow_k8_data *data, unsigned int index) { return; }
@@ -1136,7 +1101,7 @@ static int powernowk8_verify(struct cpufreq_policy *pol)
 static int __cpuinit powernowk8_cpu_init(struct cpufreq_policy *pol)
 {
 	struct powernow_k8_data *data;
-	cpumask_t oldmask = CPU_MASK_ALL;
+	cpumask_t oldmask;
 	int rc;
 
 	if (!cpu_online(pol->cpu))
@@ -1209,7 +1174,10 @@ static int __cpuinit powernowk8_cpu_init(struct cpufreq_policy *pol)
 	/* run on any CPU again */
 	set_cpus_allowed_ptr(current, &oldmask);
 
-	pol->cpus = data->starting_core_affinity;
+	if (cpu_family == CPU_HW_PSTATE)
+		pol->cpus = cpumask_of_cpu(pol->cpu);
+	else
+		pol->cpus = per_cpu(cpu_core_map, pol->cpu);
 	data->available_cores = &(pol->cpus);
 
 	/* Take a crude guess here.
@@ -1332,7 +1300,6 @@ static int __cpuinit powernowk8_init(void)
 	}
 
 	if (supported_cpus == num_online_cpus()) {
-		powernow_k8_cpu_preinit_acpi();
 		printk(KERN_INFO PFX "Found %d %s "
 			"processors (%d cpu cores) (" VERSION ")\n",
 			num_online_nodes(),
@@ -1349,10 +1316,6 @@ static void __exit powernowk8_exit(void)
 	dprintk("exit\n");
 
 	cpufreq_unregister_driver(&cpufreq_amd64_driver);
-
-#ifdef CONFIG_X86_POWERNOW_K8_ACPI
-	free_percpu(acpi_perf_data);
-#endif
 }
 
 MODULE_AUTHOR("Paul Devriendt <paul.devriendt@amd.com> and Mark Langsdorf <mark.langsdorf@amd.com>");
diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k8.h b/arch/x86/kernel/cpu/cpufreq/powernow-k8.h
index a62612cd4be8..ab48cfed4d96 100644
--- a/arch/x86/kernel/cpu/cpufreq/powernow-k8.h
+++ b/arch/x86/kernel/cpu/cpufreq/powernow-k8.h
@@ -33,13 +33,12 @@ struct powernow_k8_data {
 #ifdef CONFIG_X86_POWERNOW_K8_ACPI
 	/* the acpi table needs to be kept. it's only available if ACPI was
 	 * used to determine valid frequency/vid/fid states */
-	struct acpi_processor_performance *acpi_data;
+	struct acpi_processor_performance acpi_data;
 #endif
 	/* we need to keep track of associated cores, but let cpufreq
 	 * handle hotplug events - so just point at cpufreq pol->cpus
 	 * structure */
 	cpumask_t *available_cores;
-	cpumask_t starting_core_affinity;
 };
 
 
diff --git a/arch/x86/kernel/cpu/cyrix.c b/arch/x86/kernel/cpu/cyrix.c
index db5868cd2443..ada50505a5c8 100644
--- a/arch/x86/kernel/cpu/cyrix.c
+++ b/arch/x86/kernel/cpu/cyrix.c
@@ -134,23 +134,6 @@ static void __cpuinit set_cx86_memwb(void)
 	setCx86_old(CX86_CCR2, getCx86_old(CX86_CCR2) | 0x14);
 }
 
-static void __cpuinit set_cx86_inc(void)
-{
-	unsigned char ccr3;
-
-	printk(KERN_INFO "Enable Incrementor on Cyrix/NSC processor.\n");
-
-	ccr3 = getCx86(CX86_CCR3);
-	setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* enable MAPEN */
-	/* PCR1 -- Performance Control */
-	/* Incrementor on, whatever that is */
-	setCx86_old(CX86_PCR1, getCx86_old(CX86_PCR1) | 0x02);
-	/* PCR0 -- Performance Control */
-	/* Incrementor Margin 10 */
-	setCx86_old(CX86_PCR0, getCx86_old(CX86_PCR0) | 0x04);
-	setCx86(CX86_CCR3, ccr3);	/* disable MAPEN */
-}
-
 /*
  *	Configure later MediaGX and/or Geode processor.
  */
@@ -174,7 +157,6 @@ static void __cpuinit geode_configure(void)
 
 	set_cx86_memwb();
 	set_cx86_reorder();
-	set_cx86_inc();
 
 	local_irq_restore(flags);
 }
diff --git a/arch/x86/kernel/cpu/mcheck/mce_64.c b/arch/x86/kernel/cpu/mcheck/mce_64.c
index 65a339678ece..726a5fcdf341 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_64.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_64.c
@@ -759,6 +759,7 @@ static struct sysdev_class mce_sysclass = {
 };
 
 DEFINE_PER_CPU(struct sys_device, device_mce);
+void (*threshold_cpu_callback)(unsigned long action, unsigned int cpu) __cpuinitdata;
 
 /* Why are there no generic functions for this? */
 #define ACCESSOR(name, var, start) \
@@ -883,9 +884,13 @@ static int __cpuinit mce_cpu_callback(struct notifier_block *nfb,
 	case CPU_ONLINE:
 	case CPU_ONLINE_FROZEN:
 		mce_create_device(cpu);
+		if (threshold_cpu_callback)
+			threshold_cpu_callback(action, cpu);
 		break;
 	case CPU_DEAD:
 	case CPU_DEAD_FROZEN:
+		if (threshold_cpu_callback)
+			threshold_cpu_callback(action, cpu);
 		mce_remove_device(cpu);
 		break;
 	}
diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd_64.c b/arch/x86/kernel/cpu/mcheck/mce_amd_64.c
index 88736cadbaa6..5eb390a4b2e9 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_amd_64.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_amd_64.c
@@ -628,6 +628,7 @@ static void threshold_remove_bank(unsigned int cpu, int bank)
 	deallocate_threshold_block(cpu, bank);
 
 free_out:
+	kobject_del(b->kobj);
 	kobject_put(b->kobj);
 	kfree(b);
 	per_cpu(threshold_banks, cpu)[bank] = NULL;
@@ -645,14 +646,11 @@ static void threshold_remove_device(unsigned int cpu)
 }
 
 /* get notified when a cpu comes on/off */
-static int __cpuinit threshold_cpu_callback(struct notifier_block *nfb,
-					    unsigned long action, void *hcpu)
+static void __cpuinit amd_64_threshold_cpu_callback(unsigned long action,
+						     unsigned int cpu)
 {
-	/* cpu was unsigned int to begin with */
-	unsigned int cpu = (unsigned long)hcpu;
-
 	if (cpu >= NR_CPUS)
-		goto out;
+		return;
 
 	switch (action) {
 	case CPU_ONLINE:
@@ -666,14 +664,8 @@ static int __cpuinit threshold_cpu_callback(struct notifier_block *nfb,
 	default:
 		break;
 	}
-      out:
-	return NOTIFY_OK;
 }
 
-static struct notifier_block threshold_cpu_notifier __cpuinitdata = {
-	.notifier_call = threshold_cpu_callback,
-};
-
 static __init int threshold_init_device(void)
 {
 	unsigned lcpu = 0;
@@ -684,7 +676,7 @@ static __init int threshold_init_device(void)
 		if (err)
 			return err;
 	}
-	register_hotcpu_notifier(&threshold_cpu_notifier);
+	threshold_cpu_callback = amd_64_threshold_cpu_callback;
 	return 0;
 }
 
diff --git a/arch/x86/kernel/cpu/mtrr/generic.c b/arch/x86/kernel/cpu/mtrr/generic.c
index 509bd3d9eacd..cb7d3b6a80eb 100644
--- a/arch/x86/kernel/cpu/mtrr/generic.c
+++ b/arch/x86/kernel/cpu/mtrr/generic.c
@@ -379,6 +379,7 @@ static void generic_get_mtrr(unsigned int reg, unsigned long *base,
 			     unsigned long *size, mtrr_type *type)
 {
 	unsigned int mask_lo, mask_hi, base_lo, base_hi;
+	unsigned int tmp, hi;
 
 	rdmsr(MTRRphysMask_MSR(reg), mask_lo, mask_hi);
 	if ((mask_lo & 0x800) == 0) {
@@ -392,8 +393,23 @@ static void generic_get_mtrr(unsigned int reg, unsigned long *base,
 	rdmsr(MTRRphysBase_MSR(reg), base_lo, base_hi);
 
 	/* Work out the shifted address mask. */
-	mask_lo = size_or_mask | mask_hi << (32 - PAGE_SHIFT)
-	    | mask_lo >> PAGE_SHIFT;
+	tmp = mask_hi << (32 - PAGE_SHIFT) | mask_lo >> PAGE_SHIFT;
+	mask_lo = size_or_mask | tmp;
+	/* Expand tmp with high bits to all 1s*/
+	hi = fls(tmp);
+	if (hi > 0) {
+		tmp |= ~((1<<(hi - 1)) - 1);
+
+		if (tmp != mask_lo) {
+			static int once = 1;
+
+			if (once) {
+				printk(KERN_INFO "mtrr: your BIOS has set up an incorrect mask, fixing it up.\n");
+				once = 0;
+			}
+			mask_lo = tmp;
+		}
+	}
 
 	/* This works correctly if size is a power of two, i.e. a
 	   contiguous range. */
diff --git a/arch/x86/kernel/cpu/mtrr/main.c b/arch/x86/kernel/cpu/mtrr/main.c
index 6f23969c8faf..b117d7f8a564 100644
--- a/arch/x86/kernel/cpu/mtrr/main.c
+++ b/arch/x86/kernel/cpu/mtrr/main.c
@@ -1496,11 +1496,8 @@ int __init mtrr_trim_uncached_memory(unsigned long end_pfn)
 
 	/* kvm/qemu doesn't have mtrr set right, don't trim them all */
 	if (!highest_pfn) {
-		if (!kvm_para_available()) {
-			printk(KERN_WARNING
+		WARN(!kvm_para_available(), KERN_WARNING
 				"WARNING: strange, CPU MTRRs all blank?\n");
-			WARN_ON(1);
-		}
 		return 0;
 	}
 
diff --git a/arch/x86/kernel/cpu/perfctr-watchdog.c b/arch/x86/kernel/cpu/perfctr-watchdog.c
index de7439f82b92..05cc22dbd4ff 100644
--- a/arch/x86/kernel/cpu/perfctr-watchdog.c
+++ b/arch/x86/kernel/cpu/perfctr-watchdog.c
@@ -478,7 +478,13 @@ static int setup_p4_watchdog(unsigned nmi_hz)
 		perfctr_msr = MSR_P4_IQ_PERFCTR1;
 		evntsel_msr = MSR_P4_CRU_ESCR0;
 		cccr_msr = MSR_P4_IQ_CCCR1;
-		cccr_val = P4_CCCR_OVF_PMI1 | P4_CCCR_ESCR_SELECT(4);
+
+		/* Pentium 4 D processors don't support P4_CCCR_OVF_PMI1 */
+		if (boot_cpu_data.x86_model == 4 && boot_cpu_data.x86_mask == 4)
+			cccr_val = P4_CCCR_OVF_PMI0;
+		else
+			cccr_val = P4_CCCR_OVF_PMI1;
+		cccr_val |= P4_CCCR_ESCR_SELECT(4);
 	}
 
 	evntsel = P4_ESCR_EVENT_SELECT(0x3F)
diff --git a/arch/x86/kernel/cpuid.c b/arch/x86/kernel/cpuid.c
index 14b11b3be31c..8e9cd6a8ec12 100644
--- a/arch/x86/kernel/cpuid.c
+++ b/arch/x86/kernel/cpuid.c
@@ -89,6 +89,8 @@ static ssize_t cpuid_read(struct file *file, char __user *buf,
 	struct cpuid_regs cmd;
 	int cpu = iminor(file->f_path.dentry->d_inode);
 	u64 pos = *ppos;
+	ssize_t bytes = 0;
+	int err = 0;
 
 	if (count % 16)
 		return -EINVAL;	/* Invalid chunk size */
@@ -96,14 +98,19 @@ static ssize_t cpuid_read(struct file *file, char __user *buf,
 	for (; count; count -= 16) {
 		cmd.eax = pos;
 		cmd.ecx = pos >> 32;
-		smp_call_function_single(cpu, cpuid_smp_cpuid, &cmd, 1);
-		if (copy_to_user(tmp, &cmd, 16))
-			return -EFAULT;
+		err = smp_call_function_single(cpu, cpuid_smp_cpuid, &cmd, 1);
+		if (err)
+			break;
+		if (copy_to_user(tmp, &cmd, 16)) {
+			err = -EFAULT;
+			break;
+		}
 		tmp += 16;
+		bytes += 16;
 		*ppos = ++pos;
 	}
 
-	return tmp - buf;
+	return bytes ? bytes : err;
 }
 
 static int cpuid_open(struct inode *inode, struct file *file)
diff --git a/arch/x86/kernel/efi_32.c b/arch/x86/kernel/efi_32.c
index 4b63c8e1f13b..5cab48ee61a4 100644
--- a/arch/x86/kernel/efi_32.c
+++ b/arch/x86/kernel/efi_32.c
@@ -53,7 +53,7 @@ void efi_call_phys_prelog(void)
 	 * directory. If I have PAE, I just need to duplicate one entry in
 	 * page directory.
 	 */
-	cr4 = read_cr4();
+	cr4 = read_cr4_safe();
 
 	if (cr4 & X86_CR4_PAE) {
 		efi_bak_pg_dir_pointer[0].pgd =
@@ -91,7 +91,7 @@ void efi_call_phys_epilog(void)
 	gdt_descr.size = GDT_SIZE - 1;
 	load_gdt(&gdt_descr);
 
-	cr4 = read_cr4();
+	cr4 = read_cr4_safe();
 
 	if (cr4 & X86_CR4_PAE) {
 		swapper_pg_dir[pgd_index(0)].pgd =
diff --git a/arch/x86/kernel/genx2apic_uv_x.c b/arch/x86/kernel/genx2apic_uv_x.c
index 16a93ed7baf1..ae2ffc8a400c 100644
--- a/arch/x86/kernel/genx2apic_uv_x.c
+++ b/arch/x86/kernel/genx2apic_uv_x.c
@@ -286,7 +286,7 @@ static __init void map_low_mmrs(void)
 
 enum map_type {map_wb, map_uc};
 
-static void map_high(char *id, unsigned long base, int shift, enum map_type map_type)
+static __init void map_high(char *id, unsigned long base, int shift, enum map_type map_type)
 {
 	unsigned long bytes, paddr;
 
@@ -357,7 +357,9 @@ static __init void uv_rtc_init(void)
 		sn_rtc_cycles_per_second = ticks_per_sec;
 }
 
-static __init void uv_system_init(void)
+static bool uv_system_inited;
+
+void __init uv_system_init(void)
 {
 	union uvh_si_addr_map_config_u m_n_config;
 	union uvh_node_id_u node_id;
@@ -447,6 +449,7 @@ static __init void uv_system_init(void)
 	map_mmr_high(max_pnode);
 	map_config_high(max_pnode);
 	map_mmioh_high(max_pnode);
+	uv_system_inited = true;
 }
 
 /*
@@ -455,8 +458,7 @@ static __init void uv_system_init(void)
  */
 void __cpuinit uv_cpu_init(void)
 {
-	if (!uv_node_to_blade)
-		uv_system_init();
+	BUG_ON(!uv_system_inited);
 
 	uv_blade_info[uv_numa_blade_id()].nr_online_cpus++;
 
diff --git a/arch/x86/kernel/head64.c b/arch/x86/kernel/head64.c
index 1b318e903bf6..9bfc4d72fb2e 100644
--- a/arch/x86/kernel/head64.c
+++ b/arch/x86/kernel/head64.c
@@ -88,6 +88,7 @@ void __init x86_64_start_kernel(char * real_mode_data)
 	BUILD_BUG_ON(!(MODULES_VADDR > __START_KERNEL));
 	BUILD_BUG_ON(!(((MODULES_END - 1) & PGDIR_MASK) ==
 				(__START_KERNEL & PGDIR_MASK)));
+	BUILD_BUG_ON(__fix_to_virt(__end_of_fixed_addresses) <= MODULES_END);
 
 	/* clear bss before set_intr_gate with early_idt_handler */
 	clear_bss();
diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c
index ad2b15a1334d..59fd3b6b1303 100644
--- a/arch/x86/kernel/hpet.c
+++ b/arch/x86/kernel/hpet.c
@@ -359,6 +359,7 @@ static int hpet_clocksource_register(void)
 int __init hpet_enable(void)
 {
 	unsigned long id;
+	int i;
 
 	if (!is_hpet_capable())
 		return 0;
@@ -369,6 +370,29 @@ int __init hpet_enable(void)
 	 * Read the period and check for a sane value:
 	 */
 	hpet_period = hpet_readl(HPET_PERIOD);
+
+	/*
+	 * AMD SB700 based systems with spread spectrum enabled use a
+	 * SMM based HPET emulation to provide proper frequency
+	 * setting. The SMM code is initialized with the first HPET
+	 * register access and takes some time to complete. During
+	 * this time the config register reads 0xffffffff. We check
+	 * for max. 1000 loops whether the config register reads a non
+	 * 0xffffffff value to make sure that HPET is up and running
+	 * before we go further. A counting loop is safe, as the HPET
+	 * access takes thousands of CPU cycles. On non SB700 based
+	 * machines this check is only done once and has no side
+	 * effects.
+	 */
+	for (i = 0; hpet_readl(HPET_CFG) == 0xFFFFFFFF; i++) {
+		if (i == 1000) {
+			printk(KERN_WARNING
+			       "HPET config register value = 0xFFFFFFFF. "
+			       "Disabling HPET\n");
+			goto out_nohpet;
+		}
+	}
+
 	if (hpet_period < HPET_MIN_PERIOD || hpet_period > HPET_MAX_PERIOD)
 		goto out_nohpet;
 
diff --git a/arch/x86/kernel/machine_kexec_32.c b/arch/x86/kernel/machine_kexec_32.c
index 9fe478d98406..0732adba05ca 100644
--- a/arch/x86/kernel/machine_kexec_32.c
+++ b/arch/x86/kernel/machine_kexec_32.c
@@ -12,6 +12,7 @@
 #include <linux/init.h>
 #include <linux/numa.h>
 #include <linux/ftrace.h>
+#include <linux/suspend.h>
 
 #include <asm/pgtable.h>
 #include <asm/pgalloc.h>
@@ -78,7 +79,7 @@ static void load_segments(void)
 /*
  * A architecture hook called to validate the
  * proposed image and prepare the control pages
- * as needed.  The pages for KEXEC_CONTROL_CODE_SIZE
+ * as needed.  The pages for KEXEC_CONTROL_PAGE_SIZE
  * have been allocated, but the segments have yet
  * been copied into the kernel.
  *
@@ -113,6 +114,7 @@ void machine_kexec(struct kimage *image)
 {
 	unsigned long page_list[PAGES_NR];
 	void *control_page;
+	int save_ftrace_enabled;
 	asmlinkage unsigned long
 		(*relocate_kernel_ptr)(unsigned long indirection_page,
 				       unsigned long control_page,
@@ -120,7 +122,12 @@ void machine_kexec(struct kimage *image)
 				       unsigned int has_pae,
 				       unsigned int preserve_context);
 
-	tracer_disable();
+#ifdef CONFIG_KEXEC_JUMP
+	if (kexec_image->preserve_context)
+		save_processor_state();
+#endif
+
+	save_ftrace_enabled = __ftrace_enabled_save();
 
 	/* Interrupts aren't acceptable while we reboot */
 	local_irq_disable();
@@ -138,7 +145,7 @@ void machine_kexec(struct kimage *image)
 	}
 
 	control_page = page_address(image->control_code_page);
-	memcpy(control_page, relocate_kernel, PAGE_SIZE/2);
+	memcpy(control_page, relocate_kernel, KEXEC_CONTROL_CODE_MAX_SIZE);
 
 	relocate_kernel_ptr = control_page;
 	page_list[PA_CONTROL_PAGE] = __pa(control_page);
@@ -178,6 +185,13 @@ void machine_kexec(struct kimage *image)
 					   (unsigned long)page_list,
 					   image->start, cpu_has_pae,
 					   image->preserve_context);
+
+#ifdef CONFIG_KEXEC_JUMP
+	if (kexec_image->preserve_context)
+		restore_processor_state();
+#endif
+
+	__ftrace_enabled_restore(save_ftrace_enabled);
 }
 
 void arch_crash_save_vmcoreinfo(void)
diff --git a/arch/x86/kernel/mfgpt_32.c b/arch/x86/kernel/mfgpt_32.c
index 07c0f828f488..3b599518c322 100644
--- a/arch/x86/kernel/mfgpt_32.c
+++ b/arch/x86/kernel/mfgpt_32.c
@@ -33,6 +33,8 @@
 #include <linux/module.h>
 #include <asm/geode.h>
 
+#define MFGPT_DEFAULT_IRQ	7
+
 static struct mfgpt_timer_t {
 	unsigned int avail:1;
 } mfgpt_timers[MFGPT_MAX_TIMERS];
@@ -157,29 +159,48 @@ int geode_mfgpt_toggle_event(int timer, int cmp, int event, int enable)
 }
 EXPORT_SYMBOL_GPL(geode_mfgpt_toggle_event);
 
-int geode_mfgpt_set_irq(int timer, int cmp, int irq, int enable)
+int geode_mfgpt_set_irq(int timer, int cmp, int *irq, int enable)
 {
-	u32 val, dummy;
-	int offset;
+	u32 zsel, lpc, dummy;
+	int shift;
 
 	if (timer < 0 || timer >= MFGPT_MAX_TIMERS)
 		return -EIO;
 
-	if (geode_mfgpt_toggle_event(timer, cmp, MFGPT_EVENT_IRQ, enable))
+	/*
+	 * Unfortunately, MFGPTs come in pairs sharing their IRQ lines. If VSA
+	 * is using the same CMP of the timer's Siamese twin, the IRQ is set to
+	 * 2, and we mustn't use nor change it.
+	 * XXX: Likewise, 2 Linux drivers might clash if the 2nd overwrites the
+	 * IRQ of the 1st. This can only happen if forcing an IRQ, calling this
+	 * with *irq==0 is safe. Currently there _are_ no 2 drivers.
+	 */
+	rdmsr(MSR_PIC_ZSEL_LOW, zsel, dummy);
+	shift = ((cmp == MFGPT_CMP1 ? 0 : 4) + timer % 4) * 4;
+	if (((zsel >> shift) & 0xF) == 2)
 		return -EIO;
 
-	rdmsr(MSR_PIC_ZSEL_LOW, val, dummy);
+	/* Choose IRQ: if none supplied, keep IRQ already set or use default */
+	if (!*irq)
+		*irq = (zsel >> shift) & 0xF;
+	if (!*irq)
+		*irq = MFGPT_DEFAULT_IRQ;
 
-	offset = (timer % 4) * 4;
-
-	val &= ~((0xF << offset) | (0xF << (offset + 16)));
+	/* Can't use IRQ if it's 0 (=disabled), 2, or routed to LPC */
+	if (*irq < 1 || *irq == 2 || *irq > 15)
+		return -EIO;
+	rdmsr(MSR_PIC_IRQM_LPC, lpc, dummy);
+	if (lpc & (1 << *irq))
+		return -EIO;
 
+	/* All chosen and checked - go for it */
+	if (geode_mfgpt_toggle_event(timer, cmp, MFGPT_EVENT_IRQ, enable))
+		return -EIO;
 	if (enable) {
-		val |= (irq & 0x0F) << (offset);
-		val |= (irq & 0x0F) << (offset + 16);
+		zsel = (zsel & ~(0xF << shift)) | (*irq << shift);
+		wrmsr(MSR_PIC_ZSEL_LOW, zsel, dummy);
 	}
 
-	wrmsr(MSR_PIC_ZSEL_LOW, val, dummy);
 	return 0;
 }
 
@@ -242,7 +263,7 @@ EXPORT_SYMBOL_GPL(geode_mfgpt_alloc_timer);
 static unsigned int mfgpt_tick_mode = CLOCK_EVT_MODE_SHUTDOWN;
 static u16 mfgpt_event_clock;
 
-static int irq = 7;
+static int irq;
 static int __init mfgpt_setup(char *str)
 {
 	get_option(&str, &irq);
@@ -346,7 +367,7 @@ int __init mfgpt_timer_setup(void)
 	mfgpt_event_clock = timer;
 
 	/* Set up the IRQ on the MFGPT side */
-	if (geode_mfgpt_setup_irq(mfgpt_event_clock, MFGPT_CMP2, irq)) {
+	if (geode_mfgpt_setup_irq(mfgpt_event_clock, MFGPT_CMP2, &irq)) {
 		printk(KERN_ERR "mfgpt-timer:  Could not set up IRQ %d\n", irq);
 		return -EIO;
 	}
@@ -374,13 +395,14 @@ int __init mfgpt_timer_setup(void)
 			&mfgpt_clockevent);
 
 	printk(KERN_INFO
-	       "mfgpt-timer:  registering the MFGPT timer as a clock event.\n");
+	       "mfgpt-timer:  Registering MFGPT timer %d as a clock event, using IRQ %d\n",
+	       timer, irq);
 	clockevents_register_device(&mfgpt_clockevent);
 
 	return 0;
 
 err:
-	geode_mfgpt_release_irq(mfgpt_event_clock, MFGPT_CMP2, irq);
+	geode_mfgpt_release_irq(mfgpt_event_clock, MFGPT_CMP2, &irq);
 	printk(KERN_ERR
 	       "mfgpt-timer:  Unable to set up the MFGPT clock source\n");
 	return -EIO;
diff --git a/arch/x86/kernel/mmconf-fam10h_64.c b/arch/x86/kernel/mmconf-fam10h_64.c
index fdfdc550b366..efc2f361fe85 100644
--- a/arch/x86/kernel/mmconf-fam10h_64.c
+++ b/arch/x86/kernel/mmconf-fam10h_64.c
@@ -238,7 +238,7 @@ static struct dmi_system_id __devinitdata mmconf_dmi_table[] = {
 	{}
 };
 
-void __init check_enable_amd_mmconf_dmi(void)
+void __cpuinit check_enable_amd_mmconf_dmi(void)
 {
 	dmi_check_system(mmconf_dmi_table);
 }
diff --git a/arch/x86/kernel/mpparse.c b/arch/x86/kernel/mpparse.c
index e5d23675bb7c..f98f4e1dba09 100644
--- a/arch/x86/kernel/mpparse.c
+++ b/arch/x86/kernel/mpparse.c
@@ -49,7 +49,7 @@ static int __init mpf_checksum(unsigned char *mp, int len)
 	return sum & 0xFF;
 }
 
-static void __cpuinit MP_processor_info(struct mpc_config_processor *m)
+static void __init MP_processor_info(struct mpc_config_processor *m)
 {
 	int apicid;
 	char *bootup_cpu = "";
@@ -486,7 +486,7 @@ static void __init construct_default_ioirq_mptable(int mpc_default_type)
 }
 
 
-static void construct_ioapic_table(int mpc_default_type)
+static void __init construct_ioapic_table(int mpc_default_type)
 {
 	struct mpc_config_ioapic ioapic;
 	struct mpc_config_bus bus;
@@ -531,7 +531,7 @@ static void construct_ioapic_table(int mpc_default_type)
 	construct_default_ioirq_mptable(mpc_default_type);
 }
 #else
-static inline void construct_ioapic_table(int mpc_default_type) { }
+static inline void __init construct_ioapic_table(int mpc_default_type) { }
 #endif
 
 static inline void __init construct_default_ISA_mptable(int mpc_default_type)
diff --git a/arch/x86/kernel/msr.c b/arch/x86/kernel/msr.c
index 9fd809552447..2e2af5d18191 100644
--- a/arch/x86/kernel/msr.c
+++ b/arch/x86/kernel/msr.c
@@ -72,21 +72,28 @@ static ssize_t msr_read(struct file *file, char __user *buf,
 	u32 data[2];
 	u32 reg = *ppos;
 	int cpu = iminor(file->f_path.dentry->d_inode);
-	int err;
+	int err = 0;
+	ssize_t bytes = 0;
 
 	if (count % 8)
 		return -EINVAL;	/* Invalid chunk size */
 
 	for (; count; count -= 8) {
 		err = rdmsr_safe_on_cpu(cpu, reg, &data[0], &data[1]);
-		if (err)
-			return -EIO;
-		if (copy_to_user(tmp, &data, 8))
-			return -EFAULT;
+		if (err) {
+			if (err == -EFAULT) /* Fix idiotic error code */
+				err = -EIO;
+			break;
+		}
+		if (copy_to_user(tmp, &data, 8)) {
+			err = -EFAULT;
+			break;
+		}
 		tmp += 2;
+		bytes += 8;
 	}
 
-	return ((char __user *)tmp) - buf;
+	return bytes ? bytes : err;
 }
 
 static ssize_t msr_write(struct file *file, const char __user *buf,
@@ -96,21 +103,28 @@ static ssize_t msr_write(struct file *file, const char __user *buf,
 	u32 data[2];
 	u32 reg = *ppos;
 	int cpu = iminor(file->f_path.dentry->d_inode);
-	int err;
+	int err = 0;
+	ssize_t bytes = 0;
 
 	if (count % 8)
 		return -EINVAL;	/* Invalid chunk size */
 
 	for (; count; count -= 8) {
-		if (copy_from_user(&data, tmp, 8))
-			return -EFAULT;
+		if (copy_from_user(&data, tmp, 8)) {
+			err = -EFAULT;
+			break;
+		}
 		err = wrmsr_safe_on_cpu(cpu, reg, data[0], data[1]);
-		if (err)
-			return -EIO;
+		if (err) {
+			if (err == -EFAULT) /* Fix idiotic error code */
+				err = -EIO;
+			break;
+		}
 		tmp += 2;
+		bytes += 8;
 	}
 
-	return ((char __user *)tmp) - buf;
+	return bytes ? bytes : err;
 }
 
 static int msr_open(struct inode *inode, struct file *file)
@@ -131,7 +145,7 @@ static int msr_open(struct inode *inode, struct file *file)
 		ret = -EIO;	/* MSR not supported */
 out:
 	unlock_kernel();
-	return 0;
+	return ret;
 }
 
 /*
diff --git a/arch/x86/kernel/nmi.c b/arch/x86/kernel/nmi.c
index ac6d51222e7d..abb78a2cc4ad 100644
--- a/arch/x86/kernel/nmi.c
+++ b/arch/x86/kernel/nmi.c
@@ -114,6 +114,23 @@ static __init void nmi_cpu_busy(void *data)
 }
 #endif
 
+static void report_broken_nmi(int cpu, int *prev_nmi_count)
+{
+	printk(KERN_CONT "\n");
+
+	printk(KERN_WARNING
+		"WARNING: CPU#%d: NMI appears to be stuck (%d->%d)!\n",
+			cpu, prev_nmi_count[cpu], get_nmi_count(cpu));
+
+	printk(KERN_WARNING
+		"Please report this to bugzilla.kernel.org,\n");
+	printk(KERN_WARNING
+		"and attach the output of the 'dmesg' command.\n");
+
+	per_cpu(wd_enabled, cpu) = 0;
+	atomic_dec(&nmi_active);
+}
+
 int __init check_nmi_watchdog(void)
 {
 	unsigned int *prev_nmi_count;
@@ -141,15 +158,8 @@ int __init check_nmi_watchdog(void)
 	for_each_online_cpu(cpu) {
 		if (!per_cpu(wd_enabled, cpu))
 			continue;
-		if (get_nmi_count(cpu) - prev_nmi_count[cpu] <= 5) {
-			printk(KERN_WARNING "WARNING: CPU#%d: NMI "
-				"appears to be stuck (%d->%d)!\n",
-				cpu,
-				prev_nmi_count[cpu],
-				get_nmi_count(cpu));
-			per_cpu(wd_enabled, cpu) = 0;
-			atomic_dec(&nmi_active);
-		}
+		if (get_nmi_count(cpu) - prev_nmi_count[cpu] <= 5)
+			report_broken_nmi(cpu, prev_nmi_count);
 	}
 	endflag = 1;
 	if (!atomic_read(&nmi_active)) {
diff --git a/arch/x86/kernel/numaq_32.c b/arch/x86/kernel/numaq_32.c
index b8c45610b20a..eecc8c18f010 100644
--- a/arch/x86/kernel/numaq_32.c
+++ b/arch/x86/kernel/numaq_32.c
@@ -73,7 +73,7 @@ static void __init smp_dump_qct(void)
 }
 
 
-void __init numaq_tsc_disable(void)
+void __cpuinit numaq_tsc_disable(void)
 {
 	if (!found_numaq)
 		return;
diff --git a/arch/x86/kernel/paravirt.c b/arch/x86/kernel/paravirt.c
index 5744789a78f4..4090cd6f8436 100644
--- a/arch/x86/kernel/paravirt.c
+++ b/arch/x86/kernel/paravirt.c
@@ -469,7 +469,7 @@ struct pv_lock_ops pv_lock_ops = {
 	.spin_unlock = __ticket_spin_unlock,
 #endif
 };
-EXPORT_SYMBOL_GPL(pv_lock_ops);
+EXPORT_SYMBOL(pv_lock_ops);
 
 EXPORT_SYMBOL_GPL(pv_time_ops);
 EXPORT_SYMBOL    (pv_cpu_ops);
diff --git a/arch/x86/kernel/pci-calgary_64.c b/arch/x86/kernel/pci-calgary_64.c
index 02d19328525d..dcdac6c826e9 100644
--- a/arch/x86/kernel/pci-calgary_64.c
+++ b/arch/x86/kernel/pci-calgary_64.c
@@ -343,9 +343,8 @@ static void iommu_free(struct iommu_table *tbl, dma_addr_t dma_addr,
 	/* were we called with bad_dma_address? */
 	badend = bad_dma_address + (EMERGENCY_PAGES * PAGE_SIZE);
 	if (unlikely((dma_addr >= bad_dma_address) && (dma_addr < badend))) {
-		printk(KERN_ERR "Calgary: driver tried unmapping bad DMA "
+		WARN(1, KERN_ERR "Calgary: driver tried unmapping bad DMA "
 		       "address 0x%Lx\n", dma_addr);
-		WARN_ON(1);
 		return;
 	}
 
@@ -1269,13 +1268,15 @@ static inline int __init determine_tce_table_size(u64 ram)
 static int __init build_detail_arrays(void)
 {
 	unsigned long ptr;
-	int i, scal_detail_size, rio_detail_size;
+	unsigned numnodes, i;
+	int scal_detail_size, rio_detail_size;
 
-	if (rio_table_hdr->num_scal_dev > MAX_NUMNODES){
+	numnodes = rio_table_hdr->num_scal_dev;
+	if (numnodes > MAX_NUMNODES){
 		printk(KERN_WARNING
 			"Calgary: MAX_NUMNODES too low! Defined as %d, "
 			"but system has %d nodes.\n",
-			MAX_NUMNODES, rio_table_hdr->num_scal_dev);
+			MAX_NUMNODES, numnodes);
 		return -ENODEV;
 	}
 
@@ -1296,8 +1297,7 @@ static int __init build_detail_arrays(void)
 	}
 
 	ptr = ((unsigned long)rio_table_hdr) + 3;
-	for (i = 0; i < rio_table_hdr->num_scal_dev;
-		    i++, ptr += scal_detail_size)
+	for (i = 0; i < numnodes; i++, ptr += scal_detail_size)
 		scal_devs[i] = (struct scal_detail *)ptr;
 
 	for (i = 0; i < rio_table_hdr->num_rio_dev;
@@ -1350,7 +1350,7 @@ static void calgary_init_bitmap_from_tce_table(struct iommu_table *tbl)
  * Function for kdump case. Get the tce tables from first kernel
  * by reading the contents of the base adress register of calgary iommu
  */
-static void get_tce_space_from_tar(void)
+static void __init get_tce_space_from_tar(void)
 {
 	int bus;
 	void __iomem *target;
diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c
index 53bc653ed5ca..3b7a1ddcc0bc 100644
--- a/arch/x86/kernel/process_32.c
+++ b/arch/x86/kernel/process_32.c
@@ -95,7 +95,6 @@ static inline void play_dead(void)
 {
 	/* This must be done before dead CPU ack */
 	cpu_exit_clear();
-	wbinvd();
 	mb();
 	/* Ack it */
 	__get_cpu_var(cpu_state) = CPU_DEAD;
@@ -104,8 +103,8 @@ static inline void play_dead(void)
 	 * With physical CPU hotplug, we should halt the cpu
 	 */
 	local_irq_disable();
-	while (1)
-		halt();
+	/* mask all interrupts, flush any and all caches, and halt */
+	wbinvd_halt();
 }
 #else
 static inline void play_dead(void)
diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c
index 3fb62a7d9a16..71553b664e2a 100644
--- a/arch/x86/kernel/process_64.c
+++ b/arch/x86/kernel/process_64.c
@@ -93,14 +93,13 @@ DECLARE_PER_CPU(int, cpu_state);
 static inline void play_dead(void)
 {
 	idle_task_exit();
-	wbinvd();
 	mb();
 	/* Ack it */
 	__get_cpu_var(cpu_state) = CPU_DEAD;
 
 	local_irq_disable();
-	while (1)
-		halt();
+	/* mask all interrupts, flush any and all caches, and halt */
+	wbinvd_halt();
 }
 #else
 static inline void play_dead(void)
diff --git a/arch/x86/kernel/relocate_kernel_32.S b/arch/x86/kernel/relocate_kernel_32.S
index 703310a99023..6f50664b2ba5 100644
--- a/arch/x86/kernel/relocate_kernel_32.S
+++ b/arch/x86/kernel/relocate_kernel_32.S
@@ -20,10 +20,11 @@
 #define PAGE_ATTR (_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED | _PAGE_DIRTY)
 #define PAE_PGD_ATTR (_PAGE_PRESENT)
 
-/* control_page + PAGE_SIZE/2 ~ control_page + PAGE_SIZE * 3/4 are
- * used to save some data for jumping back
+/* control_page + KEXEC_CONTROL_CODE_MAX_SIZE
+ * ~ control_page + PAGE_SIZE are used as data storage and stack for
+ * jumping back
  */
-#define DATA(offset)		(PAGE_SIZE/2+(offset))
+#define DATA(offset)		(KEXEC_CONTROL_CODE_MAX_SIZE+(offset))
 
 /* Minimal CPU state */
 #define ESP			DATA(0x0)
@@ -376,3 +377,6 @@ swap_pages:
 	popl	%ebx
 	popl	%ebp
 	ret
+
+	.globl kexec_control_code_size
+.set kexec_control_code_size, . - relocate_kernel
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 59f07e14d083..673f12cf6eb0 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -445,7 +445,7 @@ static void __init reserve_early_setup_data(void)
  * @size: Size of the crashkernel memory to reserve.
  * Returns the base address on success, and -1ULL on failure.
  */
-unsigned long long find_and_reserve_crashkernel(unsigned long long size)
+unsigned long long __init find_and_reserve_crashkernel(unsigned long long size)
 {
 	const unsigned long long alignment = 16<<20; 	/* 16M */
 	unsigned long long start = 0LL;
@@ -604,14 +604,6 @@ void __init setup_arch(char **cmdline_p)
 	early_cpu_init();
 	early_ioremap_init();
 
-#if defined(CONFIG_VMI) && defined(CONFIG_X86_32)
-	/*
-	 * Must be before kernel pagetables are setup
-	 * or fixmap area is touched.
-	 */
-	vmi_init();
-#endif
-
 	ROOT_DEV = old_decode_dev(boot_params.hdr.root_dev);
 	screen_info = boot_params.screen_info;
 	edid_info = boot_params.edid_info;
@@ -678,6 +670,14 @@ void __init setup_arch(char **cmdline_p)
 
 	parse_early_param();
 
+#if defined(CONFIG_VMI) && defined(CONFIG_X86_32)
+	/*
+	 * Must be before kernel pagetables are setup
+	 * or fixmap area is touched.
+	 */
+	vmi_init();
+#endif
+
 	/* after early param, so could get panic from serial */
 	reserve_early_setup_data();
 
diff --git a/arch/x86/kernel/signal_64.c b/arch/x86/kernel/signal_64.c
index b45ef8ddd651..ca316b5b742c 100644
--- a/arch/x86/kernel/signal_64.c
+++ b/arch/x86/kernel/signal_64.c
@@ -104,7 +104,16 @@ static inline int restore_i387(struct _fpstate __user *buf)
 		clts();
 		task_thread_info(current)->status |= TS_USEDFPU;
 	}
-	return restore_fpu_checking((__force struct i387_fxsave_struct *)buf);
+	err = restore_fpu_checking((__force struct i387_fxsave_struct *)buf);
+	if (unlikely(err)) {
+		/*
+		 * Encountered an error while doing the restore from the
+		 * user buffer, clear the fpu state.
+		 */
+		clear_fpu(tsk);
+		clear_used_math();
+	}
+	return err;
 }
 
 /*
diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
index 04f78ab51b45..b25097c599a5 100644
--- a/arch/x86/kernel/smpboot.c
+++ b/arch/x86/kernel/smpboot.c
@@ -749,6 +749,14 @@ static void __cpuinit do_fork_idle(struct work_struct *work)
 }
 
 #ifdef CONFIG_X86_64
+
+/* __ref because it's safe to call free_bootmem when after_bootmem == 0. */
+static void __ref free_bootmem_pda(struct x8664_pda *oldpda)
+{
+	if (!after_bootmem)
+		free_bootmem((unsigned long)oldpda, sizeof(*oldpda));
+}
+
 /*
  * Allocate node local memory for the AP pda.
  *
@@ -777,8 +785,7 @@ int __cpuinit get_local_pda(int cpu)
 
 	if (oldpda) {
 		memcpy(newpda, oldpda, size);
-		if (!after_bootmem)
-			free_bootmem((unsigned long)oldpda, size);
+		free_bootmem_pda(oldpda);
 	}
 
 	newpda->in_bootmem = 0;
@@ -987,17 +994,7 @@ int __cpuinit native_cpu_up(unsigned int cpu)
 	flush_tlb_all();
 	low_mappings = 1;
 
-#ifdef CONFIG_X86_PC
-	if (def_to_bigsmp && apicid > 8) {
-		printk(KERN_WARNING
-			"More than 8 CPUs detected - skipping them.\n"
-			"Use CONFIG_X86_GENERICARCH and CONFIG_X86_BIGSMP.\n");
-		err = -1;
-	} else
-		err = do_boot_cpu(apicid, cpu);
-#else
 	err = do_boot_cpu(apicid, cpu);
-#endif
 
 	zap_low_mappings();
 	low_mappings = 0;
@@ -1051,6 +1048,34 @@ static __init void disable_smp(void)
 static int __init smp_sanity_check(unsigned max_cpus)
 {
 	preempt_disable();
+
+#if defined(CONFIG_X86_PC) && defined(CONFIG_X86_32)
+	if (def_to_bigsmp && nr_cpu_ids > 8) {
+		unsigned int cpu;
+		unsigned nr;
+
+		printk(KERN_WARNING
+		       "More than 8 CPUs detected - skipping them.\n"
+		       "Use CONFIG_X86_GENERICARCH and CONFIG_X86_BIGSMP.\n");
+
+		nr = 0;
+		for_each_present_cpu(cpu) {
+			if (nr >= 8)
+				cpu_clear(cpu, cpu_present_map);
+			nr++;
+		}
+
+		nr = 0;
+		for_each_possible_cpu(cpu) {
+			if (nr >= 8)
+				cpu_clear(cpu, cpu_possible_map);
+			nr++;
+		}
+
+		nr_cpu_ids = 8;
+	}
+#endif
+
 	if (!physid_isset(hard_smp_processor_id(), phys_cpu_present_map)) {
 		printk(KERN_WARNING "weird, boot CPU (#%d) not listed"
 				    "by the BIOS.\n", hard_smp_processor_id());
@@ -1196,6 +1221,9 @@ void __init native_smp_prepare_cpus(unsigned int max_cpus)
 	printk(KERN_INFO "CPU%d: ", 0);
 	print_cpu_info(&cpu_data(0));
 	setup_boot_clock();
+
+	if (is_uv_system())
+		uv_system_init();
 out:
 	preempt_enable();
 }
@@ -1386,17 +1414,3 @@ void __cpu_die(unsigned int cpu)
 	BUG();
 }
 #endif
-
-/*
- * If the BIOS enumerates physical processors before logical,
- * maxcpus=N at enumeration-time can be used to disable HT.
- */
-static int __init parse_maxcpus(char *arg)
-{
-	extern unsigned int maxcpus;
-
-	if (arg)
-		maxcpus = simple_strtoul(arg, NULL, 0);
-	return 0;
-}
-early_param("maxcpus", parse_maxcpus);
diff --git a/arch/x86/kernel/smpcommon.c b/arch/x86/kernel/smpcommon.c
index 99941b37eca0..397e309839dd 100644
--- a/arch/x86/kernel/smpcommon.c
+++ b/arch/x86/kernel/smpcommon.c
@@ -8,18 +8,21 @@
 DEFINE_PER_CPU(unsigned long, this_cpu_off);
 EXPORT_PER_CPU_SYMBOL(this_cpu_off);
 
-/* Initialize the CPU's GDT.  This is either the boot CPU doing itself
-   (still using the master per-cpu area), or a CPU doing it for a
-   secondary which will soon come up. */
+/*
+ * Initialize the CPU's GDT.  This is either the boot CPU doing itself
+ * (still using the master per-cpu area), or a CPU doing it for a
+ * secondary which will soon come up.
+ */
 __cpuinit void init_gdt(int cpu)
 {
-	struct desc_struct *gdt = get_cpu_gdt_table(cpu);
+	struct desc_struct gdt;
 
-	pack_descriptor(&gdt[GDT_ENTRY_PERCPU],
-			__per_cpu_offset[cpu], 0xFFFFF,
+	pack_descriptor(&gdt, __per_cpu_offset[cpu], 0xFFFFF,
 			0x2 | DESCTYPE_S, 0x8);
+	gdt.s = 1;
 
-	gdt[GDT_ENTRY_PERCPU].s = 1;
+	write_gdt_entry(get_cpu_gdt_table(cpu),
+			GDT_ENTRY_PERCPU, &gdt, DESCTYPE_S);
 
 	per_cpu(this_cpu_off, cpu) = __per_cpu_offset[cpu];
 	per_cpu(cpu_number, cpu) = cpu;
diff --git a/arch/x86/kernel/tlb_uv.c b/arch/x86/kernel/tlb_uv.c
index d0fbb7712ab0..8b8c0d6640fa 100644
--- a/arch/x86/kernel/tlb_uv.c
+++ b/arch/x86/kernel/tlb_uv.c
@@ -17,6 +17,7 @@
 #include <asm/genapic.h>
 #include <asm/idle.h>
 #include <asm/tsc.h>
+#include <asm/irq_vectors.h>
 
 #include <mach_apic.h>
 
@@ -783,7 +784,7 @@ static int __init uv_bau_init(void)
 		uv_init_blade(blade, node, cur_cpu);
 		cur_cpu += uv_blade_nr_possible_cpus(blade);
 	}
-	set_intr_gate(UV_BAU_MESSAGE, uv_bau_message_intr1);
+	alloc_intr_gate(UV_BAU_MESSAGE, uv_bau_message_intr1);
 	uv_enable_timeouts();
 
 	return 0;
diff --git a/arch/x86/kernel/traps_64.c b/arch/x86/kernel/traps_64.c
index 3f18d73f420c..513caaca7115 100644
--- a/arch/x86/kernel/traps_64.c
+++ b/arch/x86/kernel/traps_64.c
@@ -1131,7 +1131,14 @@ asmlinkage void math_state_restore(void)
 	}
 
 	clts();				/* Allow maths ops (or we recurse) */
-	restore_fpu_checking(&me->thread.xstate->fxsave);
+	/*
+	 * Paranoid restore. send a SIGSEGV if we fail to restore the state.
+	 */
+	if (unlikely(restore_fpu_checking(&me->thread.xstate->fxsave))) {
+		stts();
+		force_sig(SIGSEGV, me);
+		return;
+	}
 	task_thread_info(me)->status |= TS_USEDFPU;
 	me->fpu_counter++;
 }
diff --git a/arch/x86/kernel/tsc.c b/arch/x86/kernel/tsc.c
index 7603c0553909..346cae5ac423 100644
--- a/arch/x86/kernel/tsc.c
+++ b/arch/x86/kernel/tsc.c
@@ -104,7 +104,7 @@ __setup("notsc", notsc_setup);
 /*
  * Read TSC and the reference counters. Take care of SMI disturbance
  */
-static u64 __init tsc_read_refs(u64 *pm, u64 *hpet)
+static u64 tsc_read_refs(u64 *pm, u64 *hpet)
 {
 	u64 t1, t2;
 	int i;
@@ -122,80 +122,217 @@ static u64 __init tsc_read_refs(u64 *pm, u64 *hpet)
 	return ULLONG_MAX;
 }
 
-/**
- * native_calibrate_tsc - calibrate the tsc on boot
+/*
+ * Try to calibrate the TSC against the Programmable
+ * Interrupt Timer and return the frequency of the TSC
+ * in kHz.
+ *
+ * Return ULONG_MAX on failure to calibrate.
  */
-unsigned long native_calibrate_tsc(void)
+static unsigned long pit_calibrate_tsc(void)
 {
-	unsigned long flags;
-	u64 tsc1, tsc2, tr1, tr2, delta, pm1, pm2, hpet1, hpet2;
-	int hpet = is_hpet_enabled();
-	unsigned int tsc_khz_val = 0;
-
-	local_irq_save(flags);
-
-	tsc1 = tsc_read_refs(&pm1, hpet ? &hpet1 : NULL);
+	u64 tsc, t1, t2, delta;
+	unsigned long tscmin, tscmax;
+	int pitcnt;
 
+	/* Set the Gate high, disable speaker */
 	outb((inb(0x61) & ~0x02) | 0x01, 0x61);
 
+	/*
+	 * Setup CTC channel 2* for mode 0, (interrupt on terminal
+	 * count mode), binary count. Set the latch register to 50ms
+	 * (LSB then MSB) to begin countdown.
+	 */
 	outb(0xb0, 0x43);
 	outb((CLOCK_TICK_RATE / (1000 / 50)) & 0xff, 0x42);
 	outb((CLOCK_TICK_RATE / (1000 / 50)) >> 8, 0x42);
-	tr1 = get_cycles();
-	while ((inb(0x61) & 0x20) == 0);
-	tr2 = get_cycles();
 
-	tsc2 = tsc_read_refs(&pm2, hpet ? &hpet2 : NULL);
+	tsc = t1 = t2 = get_cycles();
 
-	local_irq_restore(flags);
+	pitcnt = 0;
+	tscmax = 0;
+	tscmin = ULONG_MAX;
+	while ((inb(0x61) & 0x20) == 0) {
+		t2 = get_cycles();
+		delta = t2 - tsc;
+		tsc = t2;
+		if ((unsigned long) delta < tscmin)
+			tscmin = (unsigned int) delta;
+		if ((unsigned long) delta > tscmax)
+			tscmax = (unsigned int) delta;
+		pitcnt++;
+	}
 
 	/*
-	 * Preset the result with the raw and inaccurate PIT
-	 * calibration value
+	 * Sanity checks:
+	 *
+	 * If we were not able to read the PIT more than 5000
+	 * times, then we have been hit by a massive SMI
+	 *
+	 * If the maximum is 10 times larger than the minimum,
+	 * then we got hit by an SMI as well.
 	 */
-	delta = (tr2 - tr1);
+	if (pitcnt < 5000 || tscmax > 10 * tscmin)
+		return ULONG_MAX;
+
+	/* Calculate the PIT value */
+	delta = t2 - t1;
 	do_div(delta, 50);
-	tsc_khz_val = delta;
+	return delta;
+}
+
+
+/**
+ * native_calibrate_tsc - calibrate the tsc on boot
+ */
+unsigned long native_calibrate_tsc(void)
+{
+	u64 tsc1, tsc2, delta, pm1, pm2, hpet1, hpet2;
+	unsigned long tsc_pit_min = ULONG_MAX, tsc_ref_min = ULONG_MAX;
+	unsigned long flags;
+	int hpet = is_hpet_enabled(), i;
+
+	/*
+	 * Run 5 calibration loops to get the lowest frequency value
+	 * (the best estimate). We use two different calibration modes
+	 * here:
+	 *
+	 * 1) PIT loop. We set the PIT Channel 2 to oneshot mode and
+	 * load a timeout of 50ms. We read the time right after we
+	 * started the timer and wait until the PIT count down reaches
+	 * zero. In each wait loop iteration we read the TSC and check
+	 * the delta to the previous read. We keep track of the min
+	 * and max values of that delta. The delta is mostly defined
+	 * by the IO time of the PIT access, so we can detect when a
+	 * SMI/SMM disturbance happend between the two reads. If the
+	 * maximum time is significantly larger than the minimum time,
+	 * then we discard the result and have another try.
+	 *
+	 * 2) Reference counter. If available we use the HPET or the
+	 * PMTIMER as a reference to check the sanity of that value.
+	 * We use separate TSC readouts and check inside of the
+	 * reference read for a SMI/SMM disturbance. We dicard
+	 * disturbed values here as well. We do that around the PIT
+	 * calibration delay loop as we have to wait for a certain
+	 * amount of time anyway.
+	 */
+	for (i = 0; i < 5; i++) {
+		unsigned long tsc_pit_khz;
+
+		/*
+		 * Read the start value and the reference count of
+		 * hpet/pmtimer when available. Then do the PIT
+		 * calibration, which will take at least 50ms, and
+		 * read the end value.
+		 */
+		local_irq_save(flags);
+		tsc1 = tsc_read_refs(&pm1, hpet ? &hpet1 : NULL);
+		tsc_pit_khz = pit_calibrate_tsc();
+		tsc2 = tsc_read_refs(&pm2, hpet ? &hpet2 : NULL);
+		local_irq_restore(flags);
+
+		/* Pick the lowest PIT TSC calibration so far */
+		tsc_pit_min = min(tsc_pit_min, tsc_pit_khz);
+
+		/* hpet or pmtimer available ? */
+		if (!hpet && !pm1 && !pm2)
+			continue;
+
+		/* Check, whether the sampling was disturbed by an SMI */
+		if (tsc1 == ULLONG_MAX || tsc2 == ULLONG_MAX)
+			continue;
+
+		tsc2 = (tsc2 - tsc1) * 1000000LL;
+
+		if (hpet) {
+			if (hpet2 < hpet1)
+				hpet2 += 0x100000000ULL;
+			hpet2 -= hpet1;
+			tsc1 = ((u64)hpet2 * hpet_readl(HPET_PERIOD));
+			do_div(tsc1, 1000000);
+		} else {
+			if (pm2 < pm1)
+				pm2 += (u64)ACPI_PM_OVRRUN;
+			pm2 -= pm1;
+			tsc1 = pm2 * 1000000000LL;
+			do_div(tsc1, PMTMR_TICKS_PER_SEC);
+		}
+
+		do_div(tsc2, tsc1);
+		tsc_ref_min = min(tsc_ref_min, (unsigned long) tsc2);
+	}
+
+	/*
+	 * Now check the results.
+	 */
+	if (tsc_pit_min == ULONG_MAX) {
+		/* PIT gave no useful value */
+		printk(KERN_WARNING "TSC: PIT calibration failed due to "
+		       "SMI disturbance.\n");
+
+		/* We don't have an alternative source, disable TSC */
+		if (!hpet && !pm1 && !pm2) {
+			printk("TSC: No reference (HPET/PMTIMER) available\n");
+			return 0;
+		}
+
+		/* The alternative source failed as well, disable TSC */
+		if (tsc_ref_min == ULONG_MAX) {
+			printk(KERN_WARNING "TSC: HPET/PMTIMER calibration "
+			       "failed due to SMI disturbance.\n");
+			return 0;
+		}
+
+		/* Use the alternative source */
+		printk(KERN_INFO "TSC: using %s reference calibration\n",
+		       hpet ? "HPET" : "PMTIMER");
+
+		return tsc_ref_min;
+	}
 
-	/* hpet or pmtimer available ? */
+	/* We don't have an alternative source, use the PIT calibration value */
 	if (!hpet && !pm1 && !pm2) {
-		printk(KERN_INFO "TSC calibrated against PIT\n");
-		goto out;
+		printk(KERN_INFO "TSC: Using PIT calibration value\n");
+		return tsc_pit_min;
 	}
 
-	/* Check, whether the sampling was disturbed by an SMI */
-	if (tsc1 == ULLONG_MAX || tsc2 == ULLONG_MAX) {
-		printk(KERN_WARNING "TSC calibration disturbed by SMI, "
-				"using PIT calibration result\n");
-		goto out;
+	/* The alternative source failed, use the PIT calibration value */
+	if (tsc_ref_min == ULONG_MAX) {
+		printk(KERN_WARNING "TSC: HPET/PMTIMER calibration failed due "
+		       "to SMI disturbance. Using PIT calibration\n");
+		return tsc_pit_min;
 	}
 
-	tsc2 = (tsc2 - tsc1) * 1000000LL;
-
-	if (hpet) {
-		printk(KERN_INFO "TSC calibrated against HPET\n");
-		if (hpet2 < hpet1)
-			hpet2 += 0x100000000ULL;
-		hpet2 -= hpet1;
-		tsc1 = ((u64)hpet2 * hpet_readl(HPET_PERIOD));
-		do_div(tsc1, 1000000);
-	} else {
-		printk(KERN_INFO "TSC calibrated against PM_TIMER\n");
-		if (pm2 < pm1)
-			pm2 += (u64)ACPI_PM_OVRRUN;
-		pm2 -= pm1;
-		tsc1 = pm2 * 1000000000LL;
-		do_div(tsc1, PMTMR_TICKS_PER_SEC);
+	/* Check the reference deviation */
+	delta = ((u64) tsc_pit_min) * 100;
+	do_div(delta, tsc_ref_min);
+
+	/*
+	 * If both calibration results are inside a 5% window, the we
+	 * use the lower frequency of those as it is probably the
+	 * closest estimate.
+	 */
+	if (delta >= 95 && delta <= 105) {
+		printk(KERN_INFO "TSC: PIT calibration confirmed by %s.\n",
+		       hpet ? "HPET" : "PMTIMER");
+		printk(KERN_INFO "TSC: using %s calibration value\n",
+		       tsc_pit_min <= tsc_ref_min ? "PIT" :
+		       hpet ? "HPET" : "PMTIMER");
+		return tsc_pit_min <= tsc_ref_min ? tsc_pit_min : tsc_ref_min;
 	}
 
-	do_div(tsc2, tsc1);
-	tsc_khz_val = tsc2;
+	printk(KERN_WARNING "TSC: PIT calibration deviates from %s: %lu %lu.\n",
+	       hpet ? "HPET" : "PMTIMER", tsc_pit_min, tsc_ref_min);
 
-out:
-	return tsc_khz_val;
+	/*
+	 * The calibration values differ too much. In doubt, we use
+	 * the PIT value as we know that there are PMTIMERs around
+	 * running at double speed.
+	 */
+	printk(KERN_INFO "TSC: Using PIT calibration value\n");
+	return tsc_pit_min;
 }
 
-
 #ifdef CONFIG_X86_32
 /* Only called from the Powernow K7 cpu freq driver */
 int recalibrate_cpu_khz(void)
@@ -314,7 +451,7 @@ static int time_cpufreq_notifier(struct notifier_block *nb, unsigned long val,
 			mark_tsc_unstable("cpufreq changes");
 	}
 
-	set_cyc2ns_scale(tsc_khz_ref, freq->cpu);
+	set_cyc2ns_scale(tsc_khz, freq->cpu);
 
 	return 0;
 }
@@ -325,6 +462,10 @@ static struct notifier_block time_cpufreq_notifier_block = {
 
 static int __init cpufreq_tsc(void)
 {
+	if (!cpu_has_tsc)
+		return 0;
+	if (boot_cpu_has(X86_FEATURE_CONSTANT_TSC))
+		return 0;
 	cpufreq_register_notifier(&time_cpufreq_notifier_block,
 				CPUFREQ_TRANSITION_NOTIFIER);
 	return 0;
diff --git a/arch/x86/kernel/tsc_sync.c b/arch/x86/kernel/tsc_sync.c
index 0577825cf89b..9ffb01c31c40 100644
--- a/arch/x86/kernel/tsc_sync.c
+++ b/arch/x86/kernel/tsc_sync.c
@@ -88,11 +88,9 @@ static __cpuinit void check_tsc_warp(void)
 			__raw_spin_unlock(&sync_lock);
 		}
 	}
-	if (!(now-start)) {
-		printk("Warning: zero tsc calibration delta: %Ld [max: %Ld]\n",
+	WARN(!(now-start),
+		"Warning: zero tsc calibration delta: %Ld [max: %Ld]\n",
 			now-start, end-start);
-		WARN_ON(1);
-	}
 }
 
 /*
diff --git a/arch/x86/kernel/visws_quirks.c b/arch/x86/kernel/visws_quirks.c
index 41e01b145c48..594ef47f0a63 100644
--- a/arch/x86/kernel/visws_quirks.c
+++ b/arch/x86/kernel/visws_quirks.c
@@ -184,8 +184,6 @@ static int __init visws_get_smp_config(unsigned int early)
 	return 1;
 }
 
-extern unsigned int __cpuinitdata maxcpus;
-
 /*
  * The Visual Workstation is Intel MP compliant in the hardware
  * sense, but it doesn't have a BIOS(-configuration table).
@@ -244,8 +242,8 @@ static int __init visws_find_smp_config(unsigned int reserve)
 		ncpus = CO_CPU_MAX;
 	}
 
-	if (ncpus > maxcpus)
-		ncpus = maxcpus;
+	if (ncpus > setup_max_cpus)
+		ncpus = setup_max_cpus;
 
 #ifdef CONFIG_X86_LOCAL_APIC
 	smp_found_config = 1;
diff --git a/arch/x86/kernel/vmlinux_32.lds.S b/arch/x86/kernel/vmlinux_32.lds.S
index cdb2363697d2..af5bdad84604 100644
--- a/arch/x86/kernel/vmlinux_32.lds.S
+++ b/arch/x86/kernel/vmlinux_32.lds.S
@@ -209,3 +209,11 @@ SECTIONS
 
   DWARF_DEBUG
 }
+
+#ifdef CONFIG_KEXEC
+/* Link time checks */
+#include <asm/kexec.h>
+
+ASSERT(kexec_control_code_size <= KEXEC_CONTROL_CODE_MAX_SIZE,
+       "kexec control code size is too big")
+#endif