aboutsummaryrefslogtreecommitdiffstats
path: root/arch/x86/kernel
diff options
context:
space:
mode:
Diffstat (limited to 'arch/x86/kernel')
-rw-r--r--arch/x86/kernel/acpi/boot.c3
-rw-r--r--arch/x86/kernel/acpi/sleep.c24
-rw-r--r--arch/x86/kernel/amd_iommu.c4
-rw-r--r--arch/x86/kernel/apic/apic.c2
-rw-r--r--arch/x86/kernel/apic/io_apic.c49
-rw-r--r--arch/x86/kernel/apic/numaq_32.c5
-rw-r--r--arch/x86/kernel/apic/x2apic_uv_x.c19
-rw-r--r--arch/x86/kernel/cpu/common.c2
-rw-r--r--arch/x86/kernel/cpu/cpufreq/longhaul.c2
-rw-r--r--arch/x86/kernel/cpu/intel.c6
-rw-r--r--arch/x86/kernel/cpu/intel_cacheinfo.c5
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce.c6
-rw-r--r--arch/x86/kernel/cpu/mtrr/cleanup.c51
-rw-r--r--arch/x86/kernel/cpu/perf_event.c2
-rw-r--r--arch/x86/kernel/entry_64.S4
-rw-r--r--arch/x86/kernel/ftrace.c17
-rw-r--r--arch/x86/kernel/head_32.S18
-rw-r--r--arch/x86/kernel/head_64.S3
-rw-r--r--arch/x86/kernel/hpet.c77
-rw-r--r--arch/x86/kernel/irq.c20
-rw-r--r--arch/x86/kernel/irqinit.c4
-rw-r--r--arch/x86/kernel/kprobes.c4
-rw-r--r--arch/x86/kernel/machine_kexec_32.c6
-rw-r--r--arch/x86/kernel/microcode_amd.c87
-rw-r--r--arch/x86/kernel/microcode_core.c8
-rw-r--r--arch/x86/kernel/mpparse.c44
-rw-r--r--arch/x86/kernel/process_64.c32
-rw-r--r--arch/x86/kernel/quirks.c9
-rw-r--r--arch/x86/kernel/reboot_fixups_32.c1
-rw-r--r--arch/x86/kernel/setup.c106
-rw-r--r--arch/x86/kernel/smpboot.c4
-rw-r--r--arch/x86/kernel/tlb_uv.c4
-rw-r--r--arch/x86/kernel/uv_time.c80
-rw-r--r--arch/x86/kernel/visws_quirks.c2
-rw-r--r--arch/x86/kernel/vmiclock_32.c2
-rw-r--r--arch/x86/kernel/vmlinux.lds.S38
-rw-r--r--arch/x86/kernel/vsyscall_64.c5
-rw-r--r--arch/x86/kernel/x86_init.c2
38 files changed, 463 insertions, 294 deletions
diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c
index 67e929b89875..fb1035cd9a6a 100644
--- a/arch/x86/kernel/acpi/boot.c
+++ b/arch/x86/kernel/acpi/boot.c
@@ -624,6 +624,7 @@ static int __init acpi_parse_hpet(struct acpi_table_header *table)
624 } 624 }
625 625
626 hpet_address = hpet_tbl->address.address; 626 hpet_address = hpet_tbl->address.address;
627 hpet_blockid = hpet_tbl->sequence;
627 628
628 /* 629 /*
629 * Some broken BIOSes advertise HPET at 0x0. We really do not 630 * Some broken BIOSes advertise HPET at 0x0. We really do not
@@ -1122,7 +1123,7 @@ static int __init acpi_parse_madt_ioapic_entries(void)
1122 if (!acpi_sci_override_gsi) 1123 if (!acpi_sci_override_gsi)
1123 acpi_sci_ioapic_setup(acpi_gbl_FADT.sci_interrupt, 0, 0); 1124 acpi_sci_ioapic_setup(acpi_gbl_FADT.sci_interrupt, 0, 0);
1124 1125
1125 /* Fill in identity legacy mapings where no override */ 1126 /* Fill in identity legacy mappings where no override */
1126 mp_config_acpi_legacy_irqs(); 1127 mp_config_acpi_legacy_irqs();
1127 1128
1128 count = 1129 count =
diff --git a/arch/x86/kernel/acpi/sleep.c b/arch/x86/kernel/acpi/sleep.c
index ca93638ba430..82e508677b91 100644
--- a/arch/x86/kernel/acpi/sleep.c
+++ b/arch/x86/kernel/acpi/sleep.c
@@ -78,12 +78,9 @@ int acpi_save_state_mem(void)
78#ifndef CONFIG_64BIT 78#ifndef CONFIG_64BIT
79 store_gdt((struct desc_ptr *)&header->pmode_gdt); 79 store_gdt((struct desc_ptr *)&header->pmode_gdt);
80 80
81 header->pmode_efer_low = nx_enabled; 81 if (rdmsr_safe(MSR_EFER, &header->pmode_efer_low,
82 if (header->pmode_efer_low & 1) { 82 &header->pmode_efer_high))
83 /* This is strange, why not save efer, always? */ 83 header->pmode_efer_low = header->pmode_efer_high = 0;
84 rdmsr(MSR_EFER, header->pmode_efer_low,
85 header->pmode_efer_high);
86 }
87#endif /* !CONFIG_64BIT */ 84#endif /* !CONFIG_64BIT */
88 85
89 header->pmode_cr0 = read_cr0(); 86 header->pmode_cr0 = read_cr0();
@@ -119,29 +116,32 @@ void acpi_restore_state_mem(void)
119 116
120 117
121/** 118/**
122 * acpi_reserve_bootmem - do _very_ early ACPI initialisation 119 * acpi_reserve_wakeup_memory - do _very_ early ACPI initialisation
123 * 120 *
124 * We allocate a page from the first 1MB of memory for the wakeup 121 * We allocate a page from the first 1MB of memory for the wakeup
125 * routine for when we come back from a sleep state. The 122 * routine for when we come back from a sleep state. The
126 * runtime allocator allows specification of <16MB pages, but not 123 * runtime allocator allows specification of <16MB pages, but not
127 * <1MB pages. 124 * <1MB pages.
128 */ 125 */
129void __init acpi_reserve_bootmem(void) 126void __init acpi_reserve_wakeup_memory(void)
130{ 127{
128 unsigned long mem;
129
131 if ((&wakeup_code_end - &wakeup_code_start) > WAKEUP_SIZE) { 130 if ((&wakeup_code_end - &wakeup_code_start) > WAKEUP_SIZE) {
132 printk(KERN_ERR 131 printk(KERN_ERR
133 "ACPI: Wakeup code way too big, S3 disabled.\n"); 132 "ACPI: Wakeup code way too big, S3 disabled.\n");
134 return; 133 return;
135 } 134 }
136 135
137 acpi_realmode = (unsigned long)alloc_bootmem_low(WAKEUP_SIZE); 136 mem = find_e820_area(0, 1<<20, WAKEUP_SIZE, PAGE_SIZE);
138 137
139 if (!acpi_realmode) { 138 if (mem == -1L) {
140 printk(KERN_ERR "ACPI: Cannot allocate lowmem, S3 disabled.\n"); 139 printk(KERN_ERR "ACPI: Cannot allocate lowmem, S3 disabled.\n");
141 return; 140 return;
142 } 141 }
143 142 acpi_realmode = (unsigned long) phys_to_virt(mem);
144 acpi_wakeup_address = virt_to_phys((void *)acpi_realmode); 143 acpi_wakeup_address = mem;
144 reserve_early(mem, mem + WAKEUP_SIZE, "ACPI WAKEUP");
145} 145}
146 146
147 147
diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c
index 32fb09102a13..1c0fb4d4ad55 100644
--- a/arch/x86/kernel/amd_iommu.c
+++ b/arch/x86/kernel/amd_iommu.c
@@ -1783,7 +1783,7 @@ retry:
1783 goto out; 1783 goto out;
1784 1784
1785 /* 1785 /*
1786 * aperture was sucessfully enlarged by 128 MB, try 1786 * aperture was successfully enlarged by 128 MB, try
1787 * allocation again 1787 * allocation again
1788 */ 1788 */
1789 goto retry; 1789 goto retry;
@@ -2490,7 +2490,7 @@ int __init amd_iommu_init_passthrough(void)
2490 struct pci_dev *dev = NULL; 2490 struct pci_dev *dev = NULL;
2491 u16 devid; 2491 u16 devid;
2492 2492
2493 /* allocate passthroug domain */ 2493 /* allocate passthrough domain */
2494 pt_domain = protection_domain_alloc(); 2494 pt_domain = protection_domain_alloc();
2495 if (!pt_domain) 2495 if (!pt_domain)
2496 return -ENOMEM; 2496 return -ENOMEM;
diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c
index ad8c75b9e453..efb2b9cd132c 100644
--- a/arch/x86/kernel/apic/apic.c
+++ b/arch/x86/kernel/apic/apic.c
@@ -647,7 +647,7 @@ static int __init calibrate_APIC_clock(void)
647 calibration_result = (delta * APIC_DIVISOR) / LAPIC_CAL_LOOPS; 647 calibration_result = (delta * APIC_DIVISOR) / LAPIC_CAL_LOOPS;
648 648
649 apic_printk(APIC_VERBOSE, "..... delta %ld\n", delta); 649 apic_printk(APIC_VERBOSE, "..... delta %ld\n", delta);
650 apic_printk(APIC_VERBOSE, "..... mult: %ld\n", lapic_clockevent.mult); 650 apic_printk(APIC_VERBOSE, "..... mult: %u\n", lapic_clockevent.mult);
651 apic_printk(APIC_VERBOSE, "..... calibration result: %u\n", 651 apic_printk(APIC_VERBOSE, "..... calibration result: %u\n",
652 calibration_result); 652 calibration_result);
653 653
diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
index c0b4468683f9..d5d498fbee4b 100644
--- a/arch/x86/kernel/apic/io_apic.c
+++ b/arch/x86/kernel/apic/io_apic.c
@@ -3267,7 +3267,8 @@ void destroy_irq(unsigned int irq)
3267 * MSI message composition 3267 * MSI message composition
3268 */ 3268 */
3269#ifdef CONFIG_PCI_MSI 3269#ifdef CONFIG_PCI_MSI
3270static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq, struct msi_msg *msg) 3270static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq,
3271 struct msi_msg *msg, u8 hpet_id)
3271{ 3272{
3272 struct irq_cfg *cfg; 3273 struct irq_cfg *cfg;
3273 int err; 3274 int err;
@@ -3301,7 +3302,10 @@ static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq, struct msi_ms
3301 irte.dest_id = IRTE_DEST(dest); 3302 irte.dest_id = IRTE_DEST(dest);
3302 3303
3303 /* Set source-id of interrupt request */ 3304 /* Set source-id of interrupt request */
3304 set_msi_sid(&irte, pdev); 3305 if (pdev)
3306 set_msi_sid(&irte, pdev);
3307 else
3308 set_hpet_sid(&irte, hpet_id);
3305 3309
3306 modify_irte(irq, &irte); 3310 modify_irte(irq, &irte);
3307 3311
@@ -3466,7 +3470,7 @@ static int setup_msi_irq(struct pci_dev *dev, struct msi_desc *msidesc, int irq)
3466 int ret; 3470 int ret;
3467 struct msi_msg msg; 3471 struct msi_msg msg;
3468 3472
3469 ret = msi_compose_msg(dev, irq, &msg); 3473 ret = msi_compose_msg(dev, irq, &msg, -1);
3470 if (ret < 0) 3474 if (ret < 0)
3471 return ret; 3475 return ret;
3472 3476
@@ -3599,7 +3603,7 @@ int arch_setup_dmar_msi(unsigned int irq)
3599 int ret; 3603 int ret;
3600 struct msi_msg msg; 3604 struct msi_msg msg;
3601 3605
3602 ret = msi_compose_msg(NULL, irq, &msg); 3606 ret = msi_compose_msg(NULL, irq, &msg, -1);
3603 if (ret < 0) 3607 if (ret < 0)
3604 return ret; 3608 return ret;
3605 dmar_msi_write(irq, &msg); 3609 dmar_msi_write(irq, &msg);
@@ -3639,6 +3643,19 @@ static int hpet_msi_set_affinity(unsigned int irq, const struct cpumask *mask)
3639 3643
3640#endif /* CONFIG_SMP */ 3644#endif /* CONFIG_SMP */
3641 3645
3646static struct irq_chip ir_hpet_msi_type = {
3647 .name = "IR-HPET_MSI",
3648 .unmask = hpet_msi_unmask,
3649 .mask = hpet_msi_mask,
3650#ifdef CONFIG_INTR_REMAP
3651 .ack = ir_ack_apic_edge,
3652#ifdef CONFIG_SMP
3653 .set_affinity = ir_set_msi_irq_affinity,
3654#endif
3655#endif
3656 .retrigger = ioapic_retrigger_irq,
3657};
3658
3642static struct irq_chip hpet_msi_type = { 3659static struct irq_chip hpet_msi_type = {
3643 .name = "HPET_MSI", 3660 .name = "HPET_MSI",
3644 .unmask = hpet_msi_unmask, 3661 .unmask = hpet_msi_unmask,
@@ -3650,20 +3667,36 @@ static struct irq_chip hpet_msi_type = {
3650 .retrigger = ioapic_retrigger_irq, 3667 .retrigger = ioapic_retrigger_irq,
3651}; 3668};
3652 3669
3653int arch_setup_hpet_msi(unsigned int irq) 3670int arch_setup_hpet_msi(unsigned int irq, unsigned int id)
3654{ 3671{
3655 int ret; 3672 int ret;
3656 struct msi_msg msg; 3673 struct msi_msg msg;
3657 struct irq_desc *desc = irq_to_desc(irq); 3674 struct irq_desc *desc = irq_to_desc(irq);
3658 3675
3659 ret = msi_compose_msg(NULL, irq, &msg); 3676 if (intr_remapping_enabled) {
3677 struct intel_iommu *iommu = map_hpet_to_ir(id);
3678 int index;
3679
3680 if (!iommu)
3681 return -1;
3682
3683 index = alloc_irte(iommu, irq, 1);
3684 if (index < 0)
3685 return -1;
3686 }
3687
3688 ret = msi_compose_msg(NULL, irq, &msg, id);
3660 if (ret < 0) 3689 if (ret < 0)
3661 return ret; 3690 return ret;
3662 3691
3663 hpet_msi_write(irq, &msg); 3692 hpet_msi_write(irq, &msg);
3664 desc->status |= IRQ_MOVE_PCNTXT; 3693 desc->status |= IRQ_MOVE_PCNTXT;
3665 set_irq_chip_and_handler_name(irq, &hpet_msi_type, handle_edge_irq, 3694 if (irq_remapped(irq))
3666 "edge"); 3695 set_irq_chip_and_handler_name(irq, &ir_hpet_msi_type,
3696 handle_edge_irq, "edge");
3697 else
3698 set_irq_chip_and_handler_name(irq, &hpet_msi_type,
3699 handle_edge_irq, "edge");
3667 3700
3668 return 0; 3701 return 0;
3669} 3702}
diff --git a/arch/x86/kernel/apic/numaq_32.c b/arch/x86/kernel/apic/numaq_32.c
index 07cdbdcd7a92..98c4665f251c 100644
--- a/arch/x86/kernel/apic/numaq_32.c
+++ b/arch/x86/kernel/apic/numaq_32.c
@@ -264,11 +264,6 @@ static void __init smp_read_mpc_oem(struct mpc_table *mpc)
264static __init void early_check_numaq(void) 264static __init void early_check_numaq(void)
265{ 265{
266 /* 266 /*
267 * Find possible boot-time SMP configuration:
268 */
269 early_find_smp_config();
270
271 /*
272 * get boot-time SMP configuration: 267 * get boot-time SMP configuration:
273 */ 268 */
274 if (smp_found_config) 269 if (smp_found_config)
diff --git a/arch/x86/kernel/apic/x2apic_uv_x.c b/arch/x86/kernel/apic/x2apic_uv_x.c
index 130c4b934877..b684bb303cbf 100644
--- a/arch/x86/kernel/apic/x2apic_uv_x.c
+++ b/arch/x86/kernel/apic/x2apic_uv_x.c
@@ -30,10 +30,22 @@
30#include <asm/apic.h> 30#include <asm/apic.h>
31#include <asm/ipi.h> 31#include <asm/ipi.h>
32#include <asm/smp.h> 32#include <asm/smp.h>
33#include <asm/x86_init.h>
33 34
34DEFINE_PER_CPU(int, x2apic_extra_bits); 35DEFINE_PER_CPU(int, x2apic_extra_bits);
35 36
36static enum uv_system_type uv_system_type; 37static enum uv_system_type uv_system_type;
38static u64 gru_start_paddr, gru_end_paddr;
39
40static inline bool is_GRU_range(u64 start, u64 end)
41{
42 return start >= gru_start_paddr && end <= gru_end_paddr;
43}
44
45static bool uv_is_untracked_pat_range(u64 start, u64 end)
46{
47 return is_ISA_range(start, end) || is_GRU_range(start, end);
48}
37 49
38static int early_get_nodeid(void) 50static int early_get_nodeid(void)
39{ 51{
@@ -49,6 +61,7 @@ static int early_get_nodeid(void)
49static int __init uv_acpi_madt_oem_check(char *oem_id, char *oem_table_id) 61static int __init uv_acpi_madt_oem_check(char *oem_id, char *oem_table_id)
50{ 62{
51 if (!strcmp(oem_id, "SGI")) { 63 if (!strcmp(oem_id, "SGI")) {
64 x86_platform.is_untracked_pat_range = uv_is_untracked_pat_range;
52 if (!strcmp(oem_table_id, "UVL")) 65 if (!strcmp(oem_table_id, "UVL"))
53 uv_system_type = UV_LEGACY_APIC; 66 uv_system_type = UV_LEGACY_APIC;
54 else if (!strcmp(oem_table_id, "UVX")) 67 else if (!strcmp(oem_table_id, "UVX"))
@@ -385,8 +398,12 @@ static __init void map_gru_high(int max_pnode)
385 int shift = UVH_RH_GAM_GRU_OVERLAY_CONFIG_MMR_BASE_SHFT; 398 int shift = UVH_RH_GAM_GRU_OVERLAY_CONFIG_MMR_BASE_SHFT;
386 399
387 gru.v = uv_read_local_mmr(UVH_RH_GAM_GRU_OVERLAY_CONFIG_MMR); 400 gru.v = uv_read_local_mmr(UVH_RH_GAM_GRU_OVERLAY_CONFIG_MMR);
388 if (gru.s.enable) 401 if (gru.s.enable) {
389 map_high("GRU", gru.s.base, shift, max_pnode, map_wb); 402 map_high("GRU", gru.s.base, shift, max_pnode, map_wb);
403 gru_start_paddr = ((u64)gru.s.base << shift);
404 gru_end_paddr = gru_start_paddr + (1UL << shift) * (max_pnode + 1);
405
406 }
390} 407}
391 408
392static __init void map_mmr_high(int max_pnode) 409static __init void map_mmr_high(int max_pnode)
diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
index a4ec8b647544..c1afa990a6c8 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -1136,7 +1136,7 @@ void __cpuinit cpu_init(void)
1136 wrmsrl(MSR_KERNEL_GS_BASE, 0); 1136 wrmsrl(MSR_KERNEL_GS_BASE, 0);
1137 barrier(); 1137 barrier();
1138 1138
1139 check_efer(); 1139 x86_configure_nx();
1140 if (cpu != 0) 1140 if (cpu != 0)
1141 enable_x2apic(); 1141 enable_x2apic();
1142 1142
diff --git a/arch/x86/kernel/cpu/cpufreq/longhaul.c b/arch/x86/kernel/cpu/cpufreq/longhaul.c
index cabd2fa3fc93..7e7eea4f8261 100644
--- a/arch/x86/kernel/cpu/cpufreq/longhaul.c
+++ b/arch/x86/kernel/cpu/cpufreq/longhaul.c
@@ -885,7 +885,7 @@ static int __init longhaul_cpu_init(struct cpufreq_policy *policy)
885 885
886 /* Find ACPI data for processor */ 886 /* Find ACPI data for processor */
887 acpi_walk_namespace(ACPI_TYPE_PROCESSOR, ACPI_ROOT_OBJECT, 887 acpi_walk_namespace(ACPI_TYPE_PROCESSOR, ACPI_ROOT_OBJECT,
888 ACPI_UINT32_MAX, &longhaul_walk_callback, 888 ACPI_UINT32_MAX, &longhaul_walk_callback, NULL,
889 NULL, (void *)&pr); 889 NULL, (void *)&pr);
890 890
891 /* Check ACPI support for C3 state */ 891 /* Check ACPI support for C3 state */
diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c
index 40e1835b35e8..c900b73f9224 100644
--- a/arch/x86/kernel/cpu/intel.c
+++ b/arch/x86/kernel/cpu/intel.c
@@ -263,8 +263,12 @@ static void __cpuinit srat_detect_node(struct cpuinfo_x86 *c)
263 /* Don't do the funky fallback heuristics the AMD version employs 263 /* Don't do the funky fallback heuristics the AMD version employs
264 for now. */ 264 for now. */
265 node = apicid_to_node[apicid]; 265 node = apicid_to_node[apicid];
266 if (node == NUMA_NO_NODE || !node_online(node)) 266 if (node == NUMA_NO_NODE)
267 node = first_node(node_online_map); 267 node = first_node(node_online_map);
268 else if (!node_online(node)) {
269 /* reuse the value from init_cpu_to_node() */
270 node = cpu_to_node(cpu);
271 }
268 numa_set_node(cpu, node); 272 numa_set_node(cpu, node);
269 273
270 printk(KERN_INFO "CPU %d/0x%x -> Node %d\n", cpu, apicid, node); 274 printk(KERN_INFO "CPU %d/0x%x -> Node %d\n", cpu, apicid, node);
diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c
index 0df4c2b7107f..6c40f6b5b340 100644
--- a/arch/x86/kernel/cpu/intel_cacheinfo.c
+++ b/arch/x86/kernel/cpu/intel_cacheinfo.c
@@ -94,7 +94,7 @@ static const struct _cache_table __cpuinitconst cache_table[] =
94 { 0xd1, LVL_3, 1024 }, /* 4-way set assoc, 64 byte line size */ 94 { 0xd1, LVL_3, 1024 }, /* 4-way set assoc, 64 byte line size */
95 { 0xd2, LVL_3, 2048 }, /* 4-way set assoc, 64 byte line size */ 95 { 0xd2, LVL_3, 2048 }, /* 4-way set assoc, 64 byte line size */
96 { 0xd6, LVL_3, 1024 }, /* 8-way set assoc, 64 byte line size */ 96 { 0xd6, LVL_3, 1024 }, /* 8-way set assoc, 64 byte line size */
97 { 0xd7, LVL_3, 2038 }, /* 8-way set assoc, 64 byte line size */ 97 { 0xd7, LVL_3, 2048 }, /* 8-way set assoc, 64 byte line size */
98 { 0xd8, LVL_3, 4096 }, /* 12-way set assoc, 64 byte line size */ 98 { 0xd8, LVL_3, 4096 }, /* 12-way set assoc, 64 byte line size */
99 { 0xdc, LVL_3, 2048 }, /* 12-way set assoc, 64 byte line size */ 99 { 0xdc, LVL_3, 2048 }, /* 12-way set assoc, 64 byte line size */
100 { 0xdd, LVL_3, 4096 }, /* 12-way set assoc, 64 byte line size */ 100 { 0xdd, LVL_3, 4096 }, /* 12-way set assoc, 64 byte line size */
@@ -102,6 +102,9 @@ static const struct _cache_table __cpuinitconst cache_table[] =
102 { 0xe2, LVL_3, 2048 }, /* 16-way set assoc, 64 byte line size */ 102 { 0xe2, LVL_3, 2048 }, /* 16-way set assoc, 64 byte line size */
103 { 0xe3, LVL_3, 4096 }, /* 16-way set assoc, 64 byte line size */ 103 { 0xe3, LVL_3, 4096 }, /* 16-way set assoc, 64 byte line size */
104 { 0xe4, LVL_3, 8192 }, /* 16-way set assoc, 64 byte line size */ 104 { 0xe4, LVL_3, 8192 }, /* 16-way set assoc, 64 byte line size */
105 { 0xea, LVL_3, 12288 }, /* 24-way set assoc, 64 byte line size */
106 { 0xeb, LVL_3, 18432 }, /* 24-way set assoc, 64 byte line size */
107 { 0xec, LVL_3, 24576 }, /* 24-way set assoc, 64 byte line size */
105 { 0x00, 0, 0} 108 { 0x00, 0, 0}
106}; 109};
107 110
diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c
index 0bcaa3875863..d7ebf25d10ed 100644
--- a/arch/x86/kernel/cpu/mcheck/mce.c
+++ b/arch/x86/kernel/cpu/mcheck/mce.c
@@ -2016,9 +2016,11 @@ mce_cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu)
2016 break; 2016 break;
2017 case CPU_DOWN_FAILED: 2017 case CPU_DOWN_FAILED:
2018 case CPU_DOWN_FAILED_FROZEN: 2018 case CPU_DOWN_FAILED_FROZEN:
2019 t->expires = round_jiffies(jiffies + 2019 if (!mce_ignore_ce && check_interval) {
2020 t->expires = round_jiffies(jiffies +
2020 __get_cpu_var(mce_next_interval)); 2021 __get_cpu_var(mce_next_interval));
2021 add_timer_on(t, cpu); 2022 add_timer_on(t, cpu);
2023 }
2022 smp_call_function_single(cpu, mce_reenable_cpu, &action, 1); 2024 smp_call_function_single(cpu, mce_reenable_cpu, &action, 1);
2023 break; 2025 break;
2024 case CPU_POST_DEAD: 2026 case CPU_POST_DEAD:
diff --git a/arch/x86/kernel/cpu/mtrr/cleanup.c b/arch/x86/kernel/cpu/mtrr/cleanup.c
index 73c86db5acbe..09b1698e0466 100644
--- a/arch/x86/kernel/cpu/mtrr/cleanup.c
+++ b/arch/x86/kernel/cpu/mtrr/cleanup.c
@@ -170,6 +170,41 @@ static int __init cmp_range(const void *x1, const void *x2)
170 return start1 - start2; 170 return start1 - start2;
171} 171}
172 172
173static int __init clean_sort_range(struct res_range *range, int az)
174{
175 int i, j, k = az - 1, nr_range = 0;
176
177 for (i = 0; i < k; i++) {
178 if (range[i].end)
179 continue;
180 for (j = k; j > i; j--) {
181 if (range[j].end) {
182 k = j;
183 break;
184 }
185 }
186 if (j == i)
187 break;
188 range[i].start = range[k].start;
189 range[i].end = range[k].end;
190 range[k].start = 0;
191 range[k].end = 0;
192 k--;
193 }
194 /* count it */
195 for (i = 0; i < az; i++) {
196 if (!range[i].end) {
197 nr_range = i;
198 break;
199 }
200 }
201
202 /* sort them */
203 sort(range, nr_range, sizeof(struct res_range), cmp_range, NULL);
204
205 return nr_range;
206}
207
173#define BIOS_BUG_MSG KERN_WARNING \ 208#define BIOS_BUG_MSG KERN_WARNING \
174 "WARNING: BIOS bug: VAR MTRR %d contains strange UC entry under 1M, check with your system vendor!\n" 209 "WARNING: BIOS bug: VAR MTRR %d contains strange UC entry under 1M, check with your system vendor!\n"
175 210
@@ -223,22 +258,18 @@ x86_get_mtrr_mem_range(struct res_range *range, int nr_range,
223 subtract_range(range, extra_remove_base, 258 subtract_range(range, extra_remove_base,
224 extra_remove_base + extra_remove_size - 1); 259 extra_remove_base + extra_remove_size - 1);
225 260
226 /* get new range num */
227 nr_range = 0;
228 for (i = 0; i < RANGE_NUM; i++) {
229 if (!range[i].end)
230 continue;
231 nr_range++;
232 }
233 if (debug_print) { 261 if (debug_print) {
234 printk(KERN_DEBUG "After UC checking\n"); 262 printk(KERN_DEBUG "After UC checking\n");
235 for (i = 0; i < nr_range; i++) 263 for (i = 0; i < RANGE_NUM; i++) {
264 if (!range[i].end)
265 continue;
236 printk(KERN_DEBUG "MTRR MAP PFN: %016lx - %016lx\n", 266 printk(KERN_DEBUG "MTRR MAP PFN: %016lx - %016lx\n",
237 range[i].start, range[i].end + 1); 267 range[i].start, range[i].end + 1);
268 }
238 } 269 }
239 270
240 /* sort the ranges */ 271 /* sort the ranges */
241 sort(range, nr_range, sizeof(struct res_range), cmp_range, NULL); 272 nr_range = clean_sort_range(range, RANGE_NUM);
242 if (debug_print) { 273 if (debug_print) {
243 printk(KERN_DEBUG "After sorting\n"); 274 printk(KERN_DEBUG "After sorting\n");
244 for (i = 0; i < nr_range; i++) 275 for (i = 0; i < nr_range; i++)
@@ -689,8 +720,6 @@ static int __init mtrr_need_cleanup(void)
689 continue; 720 continue;
690 if (!size) 721 if (!size)
691 type = MTRR_NUM_TYPES; 722 type = MTRR_NUM_TYPES;
692 if (type == MTRR_TYPE_WRPROT)
693 type = MTRR_TYPE_UNCACHABLE;
694 num[type]++; 723 num[type]++;
695 } 724 }
696 725
diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
index c1bbed1021d9..ab1a8a89b984 100644
--- a/arch/x86/kernel/cpu/perf_event.c
+++ b/arch/x86/kernel/cpu/perf_event.c
@@ -1286,7 +1286,7 @@ x86_perf_event_set_period(struct perf_event *event,
1286 return 0; 1286 return 0;
1287 1287
1288 /* 1288 /*
1289 * If we are way outside a reasoable range then just skip forward: 1289 * If we are way outside a reasonable range then just skip forward:
1290 */ 1290 */
1291 if (unlikely(left <= -period)) { 1291 if (unlikely(left <= -period)) {
1292 left = period; 1292 left = period;
diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S
index 4deb8fc849dd..63bca794c8f9 100644
--- a/arch/x86/kernel/entry_64.S
+++ b/arch/x86/kernel/entry_64.S
@@ -977,8 +977,8 @@ apicinterrupt UV_BAU_MESSAGE \
977#endif 977#endif
978apicinterrupt LOCAL_TIMER_VECTOR \ 978apicinterrupt LOCAL_TIMER_VECTOR \
979 apic_timer_interrupt smp_apic_timer_interrupt 979 apic_timer_interrupt smp_apic_timer_interrupt
980apicinterrupt GENERIC_INTERRUPT_VECTOR \ 980apicinterrupt X86_PLATFORM_IPI_VECTOR \
981 generic_interrupt smp_generic_interrupt 981 x86_platform_ipi smp_x86_platform_ipi
982 982
983#ifdef CONFIG_SMP 983#ifdef CONFIG_SMP
984apicinterrupt INVALIDATE_TLB_VECTOR_START+0 \ 984apicinterrupt INVALIDATE_TLB_VECTOR_START+0 \
diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c
index 5a1b9758fd62..309689245431 100644
--- a/arch/x86/kernel/ftrace.c
+++ b/arch/x86/kernel/ftrace.c
@@ -189,9 +189,26 @@ static void wait_for_nmi(void)
189 nmi_wait_count++; 189 nmi_wait_count++;
190} 190}
191 191
192static inline int
193within(unsigned long addr, unsigned long start, unsigned long end)
194{
195 return addr >= start && addr < end;
196}
197
192static int 198static int
193do_ftrace_mod_code(unsigned long ip, void *new_code) 199do_ftrace_mod_code(unsigned long ip, void *new_code)
194{ 200{
201 /*
202 * On x86_64, kernel text mappings are mapped read-only with
203 * CONFIG_DEBUG_RODATA. So we use the kernel identity mapping instead
204 * of the kernel text mapping to modify the kernel text.
205 *
206 * For 32bit kernels, these mappings are same and we can use
207 * kernel identity mapping to modify code.
208 */
209 if (within(ip, (unsigned long)_text, (unsigned long)_etext))
210 ip = (unsigned long)__va(__pa(ip));
211
195 mod_code_ip = (void *)ip; 212 mod_code_ip = (void *)ip;
196 mod_code_newcode = new_code; 213 mod_code_newcode = new_code;
197 214
diff --git a/arch/x86/kernel/head_32.S b/arch/x86/kernel/head_32.S
index 050c278481b1..7fd318bac59c 100644
--- a/arch/x86/kernel/head_32.S
+++ b/arch/x86/kernel/head_32.S
@@ -18,6 +18,8 @@
18#include <asm/asm-offsets.h> 18#include <asm/asm-offsets.h>
19#include <asm/setup.h> 19#include <asm/setup.h>
20#include <asm/processor-flags.h> 20#include <asm/processor-flags.h>
21#include <asm/msr-index.h>
22#include <asm/cpufeature.h>
21#include <asm/percpu.h> 23#include <asm/percpu.h>
22 24
23/* Physical address */ 25/* Physical address */
@@ -297,25 +299,27 @@ ENTRY(startup_32_smp)
297 orl %edx,%eax 299 orl %edx,%eax
298 movl %eax,%cr4 300 movl %eax,%cr4
299 301
300 btl $5, %eax # check if PAE is enabled 302 testb $X86_CR4_PAE, %al # check if PAE is enabled
301 jnc 6f 303 jz 6f
302 304
303 /* Check if extended functions are implemented */ 305 /* Check if extended functions are implemented */
304 movl $0x80000000, %eax 306 movl $0x80000000, %eax
305 cpuid 307 cpuid
306 cmpl $0x80000000, %eax 308 /* Value must be in the range 0x80000001 to 0x8000ffff */
307 jbe 6f 309 subl $0x80000001, %eax
310 cmpl $(0x8000ffff-0x80000001), %eax
311 ja 6f
308 mov $0x80000001, %eax 312 mov $0x80000001, %eax
309 cpuid 313 cpuid
310 /* Execute Disable bit supported? */ 314 /* Execute Disable bit supported? */
311 btl $20, %edx 315 btl $(X86_FEATURE_NX & 31), %edx
312 jnc 6f 316 jnc 6f
313 317
314 /* Setup EFER (Extended Feature Enable Register) */ 318 /* Setup EFER (Extended Feature Enable Register) */
315 movl $0xc0000080, %ecx 319 movl $MSR_EFER, %ecx
316 rdmsr 320 rdmsr
317 321
318 btsl $11, %eax 322 btsl $_EFER_NX, %eax
319 /* Make changes effective */ 323 /* Make changes effective */
320 wrmsr 324 wrmsr
321 325
diff --git a/arch/x86/kernel/head_64.S b/arch/x86/kernel/head_64.S
index 22db86a37643..2d8b5035371c 100644
--- a/arch/x86/kernel/head_64.S
+++ b/arch/x86/kernel/head_64.S
@@ -262,11 +262,11 @@ ENTRY(secondary_startup_64)
262 .quad x86_64_start_kernel 262 .quad x86_64_start_kernel
263 ENTRY(initial_gs) 263 ENTRY(initial_gs)
264 .quad INIT_PER_CPU_VAR(irq_stack_union) 264 .quad INIT_PER_CPU_VAR(irq_stack_union)
265 __FINITDATA
266 265
267 ENTRY(stack_start) 266 ENTRY(stack_start)
268 .quad init_thread_union+THREAD_SIZE-8 267 .quad init_thread_union+THREAD_SIZE-8
269 .word 0 268 .word 0
269 __FINITDATA
270 270
271bad_address: 271bad_address:
272 jmp bad_address 272 jmp bad_address
@@ -340,6 +340,7 @@ ENTRY(name)
340 i = i + 1 ; \ 340 i = i + 1 ; \
341 .endr 341 .endr
342 342
343 .data
343 /* 344 /*
344 * This default setting generates an ident mapping at address 0x100000 345 * This default setting generates an ident mapping at address 0x100000
345 * and a mapping for the kernel that precisely maps virtual address 346 * and a mapping for the kernel that precisely maps virtual address
diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c
index dedc2bddf7a5..ba6e65884603 100644
--- a/arch/x86/kernel/hpet.c
+++ b/arch/x86/kernel/hpet.c
@@ -33,6 +33,7 @@
33 * HPET address is set in acpi/boot.c, when an ACPI entry exists 33 * HPET address is set in acpi/boot.c, when an ACPI entry exists
34 */ 34 */
35unsigned long hpet_address; 35unsigned long hpet_address;
36u8 hpet_blockid; /* OS timer block num */
36#ifdef CONFIG_PCI_MSI 37#ifdef CONFIG_PCI_MSI
37static unsigned long hpet_num_timers; 38static unsigned long hpet_num_timers;
38#endif 39#endif
@@ -47,12 +48,12 @@ struct hpet_dev {
47 char name[10]; 48 char name[10];
48}; 49};
49 50
50unsigned long hpet_readl(unsigned long a) 51inline unsigned int hpet_readl(unsigned int a)
51{ 52{
52 return readl(hpet_virt_address + a); 53 return readl(hpet_virt_address + a);
53} 54}
54 55
55static inline void hpet_writel(unsigned long d, unsigned long a) 56static inline void hpet_writel(unsigned int d, unsigned int a)
56{ 57{
57 writel(d, hpet_virt_address + a); 58 writel(d, hpet_virt_address + a);
58} 59}
@@ -167,7 +168,7 @@ do { \
167 168
168static void hpet_reserve_msi_timers(struct hpet_data *hd); 169static void hpet_reserve_msi_timers(struct hpet_data *hd);
169 170
170static void hpet_reserve_platform_timers(unsigned long id) 171static void hpet_reserve_platform_timers(unsigned int id)
171{ 172{
172 struct hpet __iomem *hpet = hpet_virt_address; 173 struct hpet __iomem *hpet = hpet_virt_address;
173 struct hpet_timer __iomem *timer = &hpet->hpet_timers[2]; 174 struct hpet_timer __iomem *timer = &hpet->hpet_timers[2];
@@ -205,7 +206,7 @@ static void hpet_reserve_platform_timers(unsigned long id)
205 206
206} 207}
207#else 208#else
208static void hpet_reserve_platform_timers(unsigned long id) { } 209static void hpet_reserve_platform_timers(unsigned int id) { }
209#endif 210#endif
210 211
211/* 212/*
@@ -246,7 +247,7 @@ static void hpet_reset_counter(void)
246 247
247static void hpet_start_counter(void) 248static void hpet_start_counter(void)
248{ 249{
249 unsigned long cfg = hpet_readl(HPET_CFG); 250 unsigned int cfg = hpet_readl(HPET_CFG);
250 cfg |= HPET_CFG_ENABLE; 251 cfg |= HPET_CFG_ENABLE;
251 hpet_writel(cfg, HPET_CFG); 252 hpet_writel(cfg, HPET_CFG);
252} 253}
@@ -271,7 +272,7 @@ static void hpet_resume_counter(void)
271 272
272static void hpet_enable_legacy_int(void) 273static void hpet_enable_legacy_int(void)
273{ 274{
274 unsigned long cfg = hpet_readl(HPET_CFG); 275 unsigned int cfg = hpet_readl(HPET_CFG);
275 276
276 cfg |= HPET_CFG_LEGACY; 277 cfg |= HPET_CFG_LEGACY;
277 hpet_writel(cfg, HPET_CFG); 278 hpet_writel(cfg, HPET_CFG);
@@ -314,7 +315,7 @@ static int hpet_setup_msi_irq(unsigned int irq);
314static void hpet_set_mode(enum clock_event_mode mode, 315static void hpet_set_mode(enum clock_event_mode mode,
315 struct clock_event_device *evt, int timer) 316 struct clock_event_device *evt, int timer)
316{ 317{
317 unsigned long cfg, cmp, now; 318 unsigned int cfg, cmp, now;
318 uint64_t delta; 319 uint64_t delta;
319 320
320 switch (mode) { 321 switch (mode) {
@@ -323,7 +324,7 @@ static void hpet_set_mode(enum clock_event_mode mode,
323 delta = ((uint64_t)(NSEC_PER_SEC/HZ)) * evt->mult; 324 delta = ((uint64_t)(NSEC_PER_SEC/HZ)) * evt->mult;
324 delta >>= evt->shift; 325 delta >>= evt->shift;
325 now = hpet_readl(HPET_COUNTER); 326 now = hpet_readl(HPET_COUNTER);
326 cmp = now + (unsigned long) delta; 327 cmp = now + (unsigned int) delta;
327 cfg = hpet_readl(HPET_Tn_CFG(timer)); 328 cfg = hpet_readl(HPET_Tn_CFG(timer));
328 /* Make sure we use edge triggered interrupts */ 329 /* Make sure we use edge triggered interrupts */
329 cfg &= ~HPET_TN_LEVEL; 330 cfg &= ~HPET_TN_LEVEL;
@@ -339,7 +340,7 @@ static void hpet_set_mode(enum clock_event_mode mode,
339 * (See AMD-8111 HyperTransport I/O Hub Data Sheet, 340 * (See AMD-8111 HyperTransport I/O Hub Data Sheet,
340 * Publication # 24674) 341 * Publication # 24674)
341 */ 342 */
342 hpet_writel((unsigned long) delta, HPET_Tn_CMP(timer)); 343 hpet_writel((unsigned int) delta, HPET_Tn_CMP(timer));
343 hpet_start_counter(); 344 hpet_start_counter();
344 hpet_print_config(); 345 hpet_print_config();
345 break; 346 break;
@@ -383,13 +384,24 @@ static int hpet_next_event(unsigned long delta,
383 hpet_writel(cnt, HPET_Tn_CMP(timer)); 384 hpet_writel(cnt, HPET_Tn_CMP(timer));
384 385
385 /* 386 /*
386 * We need to read back the CMP register to make sure that 387 * We need to read back the CMP register on certain HPET
387 * what we wrote hit the chip before we compare it to the 388 * implementations (ATI chipsets) which seem to delay the
388 * counter. 389 * transfer of the compare register into the internal compare
390 * logic. With small deltas this might actually be too late as
391 * the counter could already be higher than the compare value
392 * at that point and we would wait for the next hpet interrupt
393 * forever. We found out that reading the CMP register back
394 * forces the transfer so we can rely on the comparison with
395 * the counter register below. If the read back from the
396 * compare register does not match the value we programmed
397 * then we might have a real hardware problem. We can not do
398 * much about it here, but at least alert the user/admin with
399 * a prominent warning.
389 */ 400 */
390 WARN_ON_ONCE((u32)hpet_readl(HPET_Tn_CMP(timer)) != cnt); 401 WARN_ONCE(hpet_readl(HPET_Tn_CMP(timer)) != cnt,
402 KERN_WARNING "hpet: compare register read back failed.\n");
391 403
392 return (s32)((u32)hpet_readl(HPET_COUNTER) - cnt) >= 0 ? -ETIME : 0; 404 return (s32)(hpet_readl(HPET_COUNTER) - cnt) >= 0 ? -ETIME : 0;
393} 405}
394 406
395static void hpet_legacy_set_mode(enum clock_event_mode mode, 407static void hpet_legacy_set_mode(enum clock_event_mode mode,
@@ -415,7 +427,7 @@ static struct hpet_dev *hpet_devs;
415void hpet_msi_unmask(unsigned int irq) 427void hpet_msi_unmask(unsigned int irq)
416{ 428{
417 struct hpet_dev *hdev = get_irq_data(irq); 429 struct hpet_dev *hdev = get_irq_data(irq);
418 unsigned long cfg; 430 unsigned int cfg;
419 431
420 /* unmask it */ 432 /* unmask it */
421 cfg = hpet_readl(HPET_Tn_CFG(hdev->num)); 433 cfg = hpet_readl(HPET_Tn_CFG(hdev->num));
@@ -425,7 +437,7 @@ void hpet_msi_unmask(unsigned int irq)
425 437
426void hpet_msi_mask(unsigned int irq) 438void hpet_msi_mask(unsigned int irq)
427{ 439{
428 unsigned long cfg; 440 unsigned int cfg;
429 struct hpet_dev *hdev = get_irq_data(irq); 441 struct hpet_dev *hdev = get_irq_data(irq);
430 442
431 /* mask it */ 443 /* mask it */
@@ -467,7 +479,7 @@ static int hpet_msi_next_event(unsigned long delta,
467 479
468static int hpet_setup_msi_irq(unsigned int irq) 480static int hpet_setup_msi_irq(unsigned int irq)
469{ 481{
470 if (arch_setup_hpet_msi(irq)) { 482 if (arch_setup_hpet_msi(irq, hpet_blockid)) {
471 destroy_irq(irq); 483 destroy_irq(irq);
472 return -EINVAL; 484 return -EINVAL;
473 } 485 }
@@ -584,6 +596,8 @@ static void hpet_msi_capability_lookup(unsigned int start_timer)
584 unsigned int num_timers_used = 0; 596 unsigned int num_timers_used = 0;
585 int i; 597 int i;
586 598
599 if (boot_cpu_has(X86_FEATURE_ARAT))
600 return;
587 id = hpet_readl(HPET_ID); 601 id = hpet_readl(HPET_ID);
588 602
589 num_timers = ((id & HPET_ID_NUMBER) >> HPET_ID_NUMBER_SHIFT); 603 num_timers = ((id & HPET_ID_NUMBER) >> HPET_ID_NUMBER_SHIFT);
@@ -598,7 +612,7 @@ static void hpet_msi_capability_lookup(unsigned int start_timer)
598 612
599 for (i = start_timer; i < num_timers - RESERVE_TIMERS; i++) { 613 for (i = start_timer; i < num_timers - RESERVE_TIMERS; i++) {
600 struct hpet_dev *hdev = &hpet_devs[num_timers_used]; 614 struct hpet_dev *hdev = &hpet_devs[num_timers_used];
601 unsigned long cfg = hpet_readl(HPET_Tn_CFG(i)); 615 unsigned int cfg = hpet_readl(HPET_Tn_CFG(i));
602 616
603 /* Only consider HPET timer with MSI support */ 617 /* Only consider HPET timer with MSI support */
604 if (!(cfg & HPET_TN_FSB_CAP)) 618 if (!(cfg & HPET_TN_FSB_CAP))
@@ -813,7 +827,7 @@ static int hpet_clocksource_register(void)
813 */ 827 */
814int __init hpet_enable(void) 828int __init hpet_enable(void)
815{ 829{
816 unsigned long id; 830 unsigned int id;
817 int i; 831 int i;
818 832
819 if (!is_hpet_capable()) 833 if (!is_hpet_capable())
@@ -872,10 +886,8 @@ int __init hpet_enable(void)
872 886
873 if (id & HPET_ID_LEGSUP) { 887 if (id & HPET_ID_LEGSUP) {
874 hpet_legacy_clockevent_register(); 888 hpet_legacy_clockevent_register();
875 hpet_msi_capability_lookup(2);
876 return 1; 889 return 1;
877 } 890 }
878 hpet_msi_capability_lookup(0);
879 return 0; 891 return 0;
880 892
881out_nohpet: 893out_nohpet:
@@ -908,9 +920,17 @@ static __init int hpet_late_init(void)
908 if (!hpet_virt_address) 920 if (!hpet_virt_address)
909 return -ENODEV; 921 return -ENODEV;
910 922
923 if (hpet_readl(HPET_ID) & HPET_ID_LEGSUP)
924 hpet_msi_capability_lookup(2);
925 else
926 hpet_msi_capability_lookup(0);
927
911 hpet_reserve_platform_timers(hpet_readl(HPET_ID)); 928 hpet_reserve_platform_timers(hpet_readl(HPET_ID));
912 hpet_print_config(); 929 hpet_print_config();
913 930
931 if (boot_cpu_has(X86_FEATURE_ARAT))
932 return 0;
933
914 for_each_online_cpu(cpu) { 934 for_each_online_cpu(cpu) {
915 hpet_cpuhp_notify(NULL, CPU_ONLINE, (void *)(long)cpu); 935 hpet_cpuhp_notify(NULL, CPU_ONLINE, (void *)(long)cpu);
916 } 936 }
@@ -925,7 +945,7 @@ fs_initcall(hpet_late_init);
925void hpet_disable(void) 945void hpet_disable(void)
926{ 946{
927 if (is_hpet_capable()) { 947 if (is_hpet_capable()) {
928 unsigned long cfg = hpet_readl(HPET_CFG); 948 unsigned int cfg = hpet_readl(HPET_CFG);
929 949
930 if (hpet_legacy_int_enabled) { 950 if (hpet_legacy_int_enabled) {
931 cfg &= ~HPET_CFG_LEGACY; 951 cfg &= ~HPET_CFG_LEGACY;
@@ -965,8 +985,8 @@ static int hpet_prev_update_sec;
965static struct rtc_time hpet_alarm_time; 985static struct rtc_time hpet_alarm_time;
966static unsigned long hpet_pie_count; 986static unsigned long hpet_pie_count;
967static u32 hpet_t1_cmp; 987static u32 hpet_t1_cmp;
968static unsigned long hpet_default_delta; 988static u32 hpet_default_delta;
969static unsigned long hpet_pie_delta; 989static u32 hpet_pie_delta;
970static unsigned long hpet_pie_limit; 990static unsigned long hpet_pie_limit;
971 991
972static rtc_irq_handler irq_handler; 992static rtc_irq_handler irq_handler;
@@ -1017,7 +1037,8 @@ EXPORT_SYMBOL_GPL(hpet_unregister_irq_handler);
1017 */ 1037 */
1018int hpet_rtc_timer_init(void) 1038int hpet_rtc_timer_init(void)
1019{ 1039{
1020 unsigned long cfg, cnt, delta, flags; 1040 unsigned int cfg, cnt, delta;
1041 unsigned long flags;
1021 1042
1022 if (!is_hpet_enabled()) 1043 if (!is_hpet_enabled())
1023 return 0; 1044 return 0;
@@ -1027,7 +1048,7 @@ int hpet_rtc_timer_init(void)
1027 1048
1028 clc = (uint64_t) hpet_clockevent.mult * NSEC_PER_SEC; 1049 clc = (uint64_t) hpet_clockevent.mult * NSEC_PER_SEC;
1029 clc >>= hpet_clockevent.shift + DEFAULT_RTC_SHIFT; 1050 clc >>= hpet_clockevent.shift + DEFAULT_RTC_SHIFT;
1030 hpet_default_delta = (unsigned long) clc; 1051 hpet_default_delta = clc;
1031 } 1052 }
1032 1053
1033 if (!(hpet_rtc_flags & RTC_PIE) || hpet_pie_limit) 1054 if (!(hpet_rtc_flags & RTC_PIE) || hpet_pie_limit)
@@ -1113,7 +1134,7 @@ int hpet_set_periodic_freq(unsigned long freq)
1113 clc = (uint64_t) hpet_clockevent.mult * NSEC_PER_SEC; 1134 clc = (uint64_t) hpet_clockevent.mult * NSEC_PER_SEC;
1114 do_div(clc, freq); 1135 do_div(clc, freq);
1115 clc >>= hpet_clockevent.shift; 1136 clc >>= hpet_clockevent.shift;
1116 hpet_pie_delta = (unsigned long) clc; 1137 hpet_pie_delta = clc;
1117 } 1138 }
1118 return 1; 1139 return 1;
1119} 1140}
@@ -1127,7 +1148,7 @@ EXPORT_SYMBOL_GPL(hpet_rtc_dropped_irq);
1127 1148
1128static void hpet_rtc_timer_reinit(void) 1149static void hpet_rtc_timer_reinit(void)
1129{ 1150{
1130 unsigned long cfg, delta; 1151 unsigned int cfg, delta;
1131 int lost_ints = -1; 1152 int lost_ints = -1;
1132 1153
1133 if (unlikely(!hpet_rtc_flags)) { 1154 if (unlikely(!hpet_rtc_flags)) {
diff --git a/arch/x86/kernel/irq.c b/arch/x86/kernel/irq.c
index fee6cc2b2079..664bcb7384ac 100644
--- a/arch/x86/kernel/irq.c
+++ b/arch/x86/kernel/irq.c
@@ -18,7 +18,7 @@
18atomic_t irq_err_count; 18atomic_t irq_err_count;
19 19
20/* Function pointer for generic interrupt vector handling */ 20/* Function pointer for generic interrupt vector handling */
21void (*generic_interrupt_extension)(void) = NULL; 21void (*x86_platform_ipi_callback)(void) = NULL;
22 22
23/* 23/*
24 * 'what should we do if we get a hw irq event on an illegal vector'. 24 * 'what should we do if we get a hw irq event on an illegal vector'.
@@ -72,10 +72,10 @@ static int show_other_interrupts(struct seq_file *p, int prec)
72 seq_printf(p, "%10u ", irq_stats(j)->apic_pending_irqs); 72 seq_printf(p, "%10u ", irq_stats(j)->apic_pending_irqs);
73 seq_printf(p, " Performance pending work\n"); 73 seq_printf(p, " Performance pending work\n");
74#endif 74#endif
75 if (generic_interrupt_extension) { 75 if (x86_platform_ipi_callback) {
76 seq_printf(p, "%*s: ", prec, "PLT"); 76 seq_printf(p, "%*s: ", prec, "PLT");
77 for_each_online_cpu(j) 77 for_each_online_cpu(j)
78 seq_printf(p, "%10u ", irq_stats(j)->generic_irqs); 78 seq_printf(p, "%10u ", irq_stats(j)->x86_platform_ipis);
79 seq_printf(p, " Platform interrupts\n"); 79 seq_printf(p, " Platform interrupts\n");
80 } 80 }
81#ifdef CONFIG_SMP 81#ifdef CONFIG_SMP
@@ -187,8 +187,8 @@ u64 arch_irq_stat_cpu(unsigned int cpu)
187 sum += irq_stats(cpu)->apic_perf_irqs; 187 sum += irq_stats(cpu)->apic_perf_irqs;
188 sum += irq_stats(cpu)->apic_pending_irqs; 188 sum += irq_stats(cpu)->apic_pending_irqs;
189#endif 189#endif
190 if (generic_interrupt_extension) 190 if (x86_platform_ipi_callback)
191 sum += irq_stats(cpu)->generic_irqs; 191 sum += irq_stats(cpu)->x86_platform_ipis;
192#ifdef CONFIG_SMP 192#ifdef CONFIG_SMP
193 sum += irq_stats(cpu)->irq_resched_count; 193 sum += irq_stats(cpu)->irq_resched_count;
194 sum += irq_stats(cpu)->irq_call_count; 194 sum += irq_stats(cpu)->irq_call_count;
@@ -251,9 +251,9 @@ unsigned int __irq_entry do_IRQ(struct pt_regs *regs)
251} 251}
252 252
253/* 253/*
254 * Handler for GENERIC_INTERRUPT_VECTOR. 254 * Handler for X86_PLATFORM_IPI_VECTOR.
255 */ 255 */
256void smp_generic_interrupt(struct pt_regs *regs) 256void smp_x86_platform_ipi(struct pt_regs *regs)
257{ 257{
258 struct pt_regs *old_regs = set_irq_regs(regs); 258 struct pt_regs *old_regs = set_irq_regs(regs);
259 259
@@ -263,10 +263,10 @@ void smp_generic_interrupt(struct pt_regs *regs)
263 263
264 irq_enter(); 264 irq_enter();
265 265
266 inc_irq_stat(generic_irqs); 266 inc_irq_stat(x86_platform_ipis);
267 267
268 if (generic_interrupt_extension) 268 if (x86_platform_ipi_callback)
269 generic_interrupt_extension(); 269 x86_platform_ipi_callback();
270 270
271 irq_exit(); 271 irq_exit();
272 272
diff --git a/arch/x86/kernel/irqinit.c b/arch/x86/kernel/irqinit.c
index 40f30773fb29..d5932226614f 100644
--- a/arch/x86/kernel/irqinit.c
+++ b/arch/x86/kernel/irqinit.c
@@ -200,8 +200,8 @@ static void __init apic_intr_init(void)
200 /* self generated IPI for local APIC timer */ 200 /* self generated IPI for local APIC timer */
201 alloc_intr_gate(LOCAL_TIMER_VECTOR, apic_timer_interrupt); 201 alloc_intr_gate(LOCAL_TIMER_VECTOR, apic_timer_interrupt);
202 202
203 /* generic IPI for platform specific use */ 203 /* IPI for X86 platform specific use */
204 alloc_intr_gate(GENERIC_INTERRUPT_VECTOR, generic_interrupt); 204 alloc_intr_gate(X86_PLATFORM_IPI_VECTOR, x86_platform_ipi);
205 205
206 /* IPI vectors for APIC spurious and error interrupts */ 206 /* IPI vectors for APIC spurious and error interrupts */
207 alloc_intr_gate(SPURIOUS_APIC_VECTOR, spurious_interrupt); 207 alloc_intr_gate(SPURIOUS_APIC_VECTOR, spurious_interrupt);
diff --git a/arch/x86/kernel/kprobes.c b/arch/x86/kernel/kprobes.c
index 1f3186ce213c..5b8c7505b3bc 100644
--- a/arch/x86/kernel/kprobes.c
+++ b/arch/x86/kernel/kprobes.c
@@ -481,7 +481,7 @@ static int __kprobes reenter_kprobe(struct kprobe *p, struct pt_regs *regs,
481 481
482/* 482/*
483 * Interrupts are disabled on entry as trap3 is an interrupt gate and they 483 * Interrupts are disabled on entry as trap3 is an interrupt gate and they
484 * remain disabled thorough out this function. 484 * remain disabled throughout this function.
485 */ 485 */
486static int __kprobes kprobe_handler(struct pt_regs *regs) 486static int __kprobes kprobe_handler(struct pt_regs *regs)
487{ 487{
@@ -818,7 +818,7 @@ no_change:
818 818
819/* 819/*
820 * Interrupts are disabled on entry as trap1 is an interrupt gate and they 820 * Interrupts are disabled on entry as trap1 is an interrupt gate and they
821 * remain disabled thoroughout this function. 821 * remain disabled throughout this function.
822 */ 822 */
823static int __kprobes post_kprobe_handler(struct pt_regs *regs) 823static int __kprobes post_kprobe_handler(struct pt_regs *regs)
824{ 824{
diff --git a/arch/x86/kernel/machine_kexec_32.c b/arch/x86/kernel/machine_kexec_32.c
index c843f8406da2..a3fa43ba5d3b 100644
--- a/arch/x86/kernel/machine_kexec_32.c
+++ b/arch/x86/kernel/machine_kexec_32.c
@@ -158,8 +158,7 @@ int machine_kexec_prepare(struct kimage *image)
158{ 158{
159 int error; 159 int error;
160 160
161 if (nx_enabled) 161 set_pages_x(image->control_code_page, 1);
162 set_pages_x(image->control_code_page, 1);
163 error = machine_kexec_alloc_page_tables(image); 162 error = machine_kexec_alloc_page_tables(image);
164 if (error) 163 if (error)
165 return error; 164 return error;
@@ -173,8 +172,7 @@ int machine_kexec_prepare(struct kimage *image)
173 */ 172 */
174void machine_kexec_cleanup(struct kimage *image) 173void machine_kexec_cleanup(struct kimage *image)
175{ 174{
176 if (nx_enabled) 175 set_pages_nx(image->control_code_page, 1);
177 set_pages_nx(image->control_code_page, 1);
178 machine_kexec_free_page_tables(image); 176 machine_kexec_free_page_tables(image);
179} 177}
180 178
diff --git a/arch/x86/kernel/microcode_amd.c b/arch/x86/kernel/microcode_amd.c
index f4c538b681ca..63123d902103 100644
--- a/arch/x86/kernel/microcode_amd.c
+++ b/arch/x86/kernel/microcode_amd.c
@@ -33,6 +33,9 @@ MODULE_LICENSE("GPL v2");
33#define UCODE_EQUIV_CPU_TABLE_TYPE 0x00000000 33#define UCODE_EQUIV_CPU_TABLE_TYPE 0x00000000
34#define UCODE_UCODE_TYPE 0x00000001 34#define UCODE_UCODE_TYPE 0x00000001
35 35
36const struct firmware *firmware;
37static int supported_cpu;
38
36struct equiv_cpu_entry { 39struct equiv_cpu_entry {
37 u32 installed_cpu; 40 u32 installed_cpu;
38 u32 fixed_errata_mask; 41 u32 fixed_errata_mask;
@@ -71,17 +74,14 @@ static struct equiv_cpu_entry *equiv_cpu_table;
71 74
72static int collect_cpu_info_amd(int cpu, struct cpu_signature *csig) 75static int collect_cpu_info_amd(int cpu, struct cpu_signature *csig)
73{ 76{
74 struct cpuinfo_x86 *c = &cpu_data(cpu);
75 u32 dummy; 77 u32 dummy;
76 78
77 memset(csig, 0, sizeof(*csig)); 79 if (!supported_cpu)
78 if (c->x86_vendor != X86_VENDOR_AMD || c->x86 < 0x10) {
79 printk(KERN_WARNING "microcode: CPU%d: AMD CPU family 0x%x not "
80 "supported\n", cpu, c->x86);
81 return -1; 80 return -1;
82 } 81
82 memset(csig, 0, sizeof(*csig));
83 rdmsr(MSR_AMD64_PATCH_LEVEL, csig->rev, dummy); 83 rdmsr(MSR_AMD64_PATCH_LEVEL, csig->rev, dummy);
84 printk(KERN_INFO "microcode: CPU%d: patch_level=0x%x\n", cpu, csig->rev); 84 pr_info("microcode: CPU%d: patch_level=0x%x\n", cpu, csig->rev);
85 return 0; 85 return 0;
86} 86}
87 87
@@ -103,22 +103,15 @@ static int get_matching_microcode(int cpu, void *mc, int rev)
103 i++; 103 i++;
104 } 104 }
105 105
106 if (!equiv_cpu_id) { 106 if (!equiv_cpu_id)
107 printk(KERN_WARNING "microcode: CPU%d: cpu revision "
108 "not listed in equivalent cpu table\n", cpu);
109 return 0; 107 return 0;
110 }
111 108
112 if (mc_header->processor_rev_id != equiv_cpu_id) { 109 if (mc_header->processor_rev_id != equiv_cpu_id)
113 printk(KERN_ERR "microcode: CPU%d: patch mismatch "
114 "(processor_rev_id: %x, equiv_cpu_id: %x)\n",
115 cpu, mc_header->processor_rev_id, equiv_cpu_id);
116 return 0; 110 return 0;
117 }
118 111
119 /* ucode might be chipset specific -- currently we don't support this */ 112 /* ucode might be chipset specific -- currently we don't support this */
120 if (mc_header->nb_dev_id || mc_header->sb_dev_id) { 113 if (mc_header->nb_dev_id || mc_header->sb_dev_id) {
121 printk(KERN_ERR "microcode: CPU%d: loading of chipset " 114 pr_err(KERN_ERR "microcode: CPU%d: loading of chipset "
122 "specific code not yet supported\n", cpu); 115 "specific code not yet supported\n", cpu);
123 return 0; 116 return 0;
124 } 117 }
@@ -148,14 +141,12 @@ static int apply_microcode_amd(int cpu)
148 141
149 /* check current patch id and patch's id for match */ 142 /* check current patch id and patch's id for match */
150 if (rev != mc_amd->hdr.patch_id) { 143 if (rev != mc_amd->hdr.patch_id) {
151 printk(KERN_ERR "microcode: CPU%d: update failed " 144 pr_err("microcode: CPU%d: update failed "
152 "(for patch_level=0x%x)\n", cpu, mc_amd->hdr.patch_id); 145 "(for patch_level=0x%x)\n", cpu, mc_amd->hdr.patch_id);
153 return -1; 146 return -1;
154 } 147 }
155 148
156 printk(KERN_INFO "microcode: CPU%d: updated (new patch_level=0x%x)\n", 149 pr_info("microcode: CPU%d: updated (new patch_level=0x%x)\n", cpu, rev);
157 cpu, rev);
158
159 uci->cpu_sig.rev = rev; 150 uci->cpu_sig.rev = rev;
160 151
161 return 0; 152 return 0;
@@ -178,18 +169,15 @@ get_next_ucode(const u8 *buf, unsigned int size, unsigned int *mc_size)
178 return NULL; 169 return NULL;
179 170
180 if (section_hdr[0] != UCODE_UCODE_TYPE) { 171 if (section_hdr[0] != UCODE_UCODE_TYPE) {
181 printk(KERN_ERR "microcode: error: invalid type field in " 172 pr_err("microcode: error: invalid type field in "
182 "container file section header\n"); 173 "container file section header\n");
183 return NULL; 174 return NULL;
184 } 175 }
185 176
186 total_size = (unsigned long) (section_hdr[4] + (section_hdr[5] << 8)); 177 total_size = (unsigned long) (section_hdr[4] + (section_hdr[5] << 8));
187 178
188 printk(KERN_DEBUG "microcode: size %u, total_size %u\n",
189 size, total_size);
190
191 if (total_size > size || total_size > UCODE_MAX_SIZE) { 179 if (total_size > size || total_size > UCODE_MAX_SIZE) {
192 printk(KERN_ERR "microcode: error: size mismatch\n"); 180 pr_err("microcode: error: size mismatch\n");
193 return NULL; 181 return NULL;
194 } 182 }
195 183
@@ -218,15 +206,14 @@ static int install_equiv_cpu_table(const u8 *buf)
218 size = buf_pos[2]; 206 size = buf_pos[2];
219 207
220 if (buf_pos[1] != UCODE_EQUIV_CPU_TABLE_TYPE || !size) { 208 if (buf_pos[1] != UCODE_EQUIV_CPU_TABLE_TYPE || !size) {
221 printk(KERN_ERR "microcode: error: invalid type field in " 209 pr_err("microcode: error: invalid type field in "
222 "container file section header\n"); 210 "container file section header\n");
223 return 0; 211 return 0;
224 } 212 }
225 213
226 equiv_cpu_table = (struct equiv_cpu_entry *) vmalloc(size); 214 equiv_cpu_table = (struct equiv_cpu_entry *) vmalloc(size);
227 if (!equiv_cpu_table) { 215 if (!equiv_cpu_table) {
228 printk(KERN_ERR "microcode: failed to allocate " 216 pr_err("microcode: failed to allocate equivalent CPU table\n");
229 "equivalent CPU table\n");
230 return 0; 217 return 0;
231 } 218 }
232 219
@@ -259,8 +246,7 @@ generic_load_microcode(int cpu, const u8 *data, size_t size)
259 246
260 offset = install_equiv_cpu_table(ucode_ptr); 247 offset = install_equiv_cpu_table(ucode_ptr);
261 if (!offset) { 248 if (!offset) {
262 printk(KERN_ERR "microcode: failed to create " 249 pr_err("microcode: failed to create equivalent cpu table\n");
263 "equivalent cpu table\n");
264 return UCODE_ERROR; 250 return UCODE_ERROR;
265 } 251 }
266 252
@@ -308,33 +294,27 @@ generic_load_microcode(int cpu, const u8 *data, size_t size)
308 294
309static enum ucode_state request_microcode_fw(int cpu, struct device *device) 295static enum ucode_state request_microcode_fw(int cpu, struct device *device)
310{ 296{
311 const char *fw_name = "amd-ucode/microcode_amd.bin";
312 const struct firmware *firmware;
313 enum ucode_state ret; 297 enum ucode_state ret;
314 298
315 if (request_firmware(&firmware, fw_name, device)) { 299 if (firmware == NULL)
316 printk(KERN_ERR "microcode: failed to load file %s\n", fw_name);
317 return UCODE_NFOUND; 300 return UCODE_NFOUND;
318 }
319 301
320 if (*(u32 *)firmware->data != UCODE_MAGIC) { 302 if (*(u32 *)firmware->data != UCODE_MAGIC) {
321 printk(KERN_ERR "microcode: invalid UCODE_MAGIC (0x%08x)\n", 303 pr_err("microcode: invalid UCODE_MAGIC (0x%08x)\n",
322 *(u32 *)firmware->data); 304 *(u32 *)firmware->data);
323 return UCODE_ERROR; 305 return UCODE_ERROR;
324 } 306 }
325 307
326 ret = generic_load_microcode(cpu, firmware->data, firmware->size); 308 ret = generic_load_microcode(cpu, firmware->data, firmware->size);
327 309
328 release_firmware(firmware);
329
330 return ret; 310 return ret;
331} 311}
332 312
333static enum ucode_state 313static enum ucode_state
334request_microcode_user(int cpu, const void __user *buf, size_t size) 314request_microcode_user(int cpu, const void __user *buf, size_t size)
335{ 315{
336 printk(KERN_INFO "microcode: AMD microcode update via " 316 pr_info("microcode: AMD microcode update via "
337 "/dev/cpu/microcode not supported\n"); 317 "/dev/cpu/microcode not supported\n");
338 return UCODE_ERROR; 318 return UCODE_ERROR;
339} 319}
340 320
@@ -346,7 +326,32 @@ static void microcode_fini_cpu_amd(int cpu)
346 uci->mc = NULL; 326 uci->mc = NULL;
347} 327}
348 328
329void init_microcode_amd(struct device *device)
330{
331 const char *fw_name = "amd-ucode/microcode_amd.bin";
332 struct cpuinfo_x86 *c = &boot_cpu_data;
333
334 WARN_ON(c->x86_vendor != X86_VENDOR_AMD);
335
336 if (c->x86 < 0x10) {
337 pr_warning("microcode: AMD CPU family 0x%x not supported\n",
338 c->x86);
339 return;
340 }
341 supported_cpu = 1;
342
343 if (request_firmware(&firmware, fw_name, device))
344 pr_err("microcode: failed to load file %s\n", fw_name);
345}
346
347void fini_microcode_amd(void)
348{
349 release_firmware(firmware);
350}
351
349static struct microcode_ops microcode_amd_ops = { 352static struct microcode_ops microcode_amd_ops = {
353 .init = init_microcode_amd,
354 .fini = fini_microcode_amd,
350 .request_microcode_user = request_microcode_user, 355 .request_microcode_user = request_microcode_user,
351 .request_microcode_fw = request_microcode_fw, 356 .request_microcode_fw = request_microcode_fw,
352 .collect_cpu_info = collect_cpu_info_amd, 357 .collect_cpu_info = collect_cpu_info_amd,
diff --git a/arch/x86/kernel/microcode_core.c b/arch/x86/kernel/microcode_core.c
index 2bcad3926edb..e68aae397869 100644
--- a/arch/x86/kernel/microcode_core.c
+++ b/arch/x86/kernel/microcode_core.c
@@ -391,7 +391,7 @@ static enum ucode_state microcode_update_cpu(int cpu)
391 struct ucode_cpu_info *uci = ucode_cpu_info + cpu; 391 struct ucode_cpu_info *uci = ucode_cpu_info + cpu;
392 enum ucode_state ustate; 392 enum ucode_state ustate;
393 393
394 if (uci->valid) 394 if (uci->valid && uci->mc)
395 ustate = microcode_resume_cpu(cpu); 395 ustate = microcode_resume_cpu(cpu);
396 else 396 else
397 ustate = microcode_init_cpu(cpu); 397 ustate = microcode_init_cpu(cpu);
@@ -518,6 +518,9 @@ static int __init microcode_init(void)
518 return PTR_ERR(microcode_pdev); 518 return PTR_ERR(microcode_pdev);
519 } 519 }
520 520
521 if (microcode_ops->init)
522 microcode_ops->init(&microcode_pdev->dev);
523
521 get_online_cpus(); 524 get_online_cpus();
522 mutex_lock(&microcode_mutex); 525 mutex_lock(&microcode_mutex);
523 526
@@ -561,6 +564,9 @@ static void __exit microcode_exit(void)
561 564
562 platform_device_unregister(microcode_pdev); 565 platform_device_unregister(microcode_pdev);
563 566
567 if (microcode_ops->fini)
568 microcode_ops->fini();
569
564 microcode_ops = NULL; 570 microcode_ops = NULL;
565 571
566 pr_info("Microcode Update Driver: v" MICROCODE_VERSION " removed.\n"); 572 pr_info("Microcode Update Driver: v" MICROCODE_VERSION " removed.\n");
diff --git a/arch/x86/kernel/mpparse.c b/arch/x86/kernel/mpparse.c
index 5be95ef4ffec..35a57c963df9 100644
--- a/arch/x86/kernel/mpparse.c
+++ b/arch/x86/kernel/mpparse.c
@@ -667,36 +667,18 @@ void __init default_get_smp_config(unsigned int early)
667 */ 667 */
668} 668}
669 669
670static void __init smp_reserve_bootmem(struct mpf_intel *mpf) 670static void __init smp_reserve_memory(struct mpf_intel *mpf)
671{ 671{
672 unsigned long size = get_mpc_size(mpf->physptr); 672 unsigned long size = get_mpc_size(mpf->physptr);
673#ifdef CONFIG_X86_32
674 /*
675 * We cannot access to MPC table to compute table size yet,
676 * as only few megabytes from the bottom is mapped now.
677 * PC-9800's MPC table places on the very last of physical
678 * memory; so that simply reserving PAGE_SIZE from mpf->physptr
679 * yields BUG() in reserve_bootmem.
680 * also need to make sure physptr is below than max_low_pfn
681 * we don't need reserve the area above max_low_pfn
682 */
683 unsigned long end = max_low_pfn * PAGE_SIZE;
684 673
685 if (mpf->physptr < end) { 674 reserve_early(mpf->physptr, mpf->physptr+size, "MP-table mpc");
686 if (mpf->physptr + size > end)
687 size = end - mpf->physptr;
688 reserve_bootmem_generic(mpf->physptr, size, BOOTMEM_DEFAULT);
689 }
690#else
691 reserve_bootmem_generic(mpf->physptr, size, BOOTMEM_DEFAULT);
692#endif
693} 675}
694 676
695static int __init smp_scan_config(unsigned long base, unsigned long length, 677static int __init smp_scan_config(unsigned long base, unsigned long length)
696 unsigned reserve)
697{ 678{
698 unsigned int *bp = phys_to_virt(base); 679 unsigned int *bp = phys_to_virt(base);
699 struct mpf_intel *mpf; 680 struct mpf_intel *mpf;
681 unsigned long mem;
700 682
701 apic_printk(APIC_VERBOSE, "Scan SMP from %p for %ld bytes.\n", 683 apic_printk(APIC_VERBOSE, "Scan SMP from %p for %ld bytes.\n",
702 bp, length); 684 bp, length);
@@ -717,12 +699,10 @@ static int __init smp_scan_config(unsigned long base, unsigned long length,
717 printk(KERN_INFO "found SMP MP-table at [%p] %llx\n", 699 printk(KERN_INFO "found SMP MP-table at [%p] %llx\n",
718 mpf, (u64)virt_to_phys(mpf)); 700 mpf, (u64)virt_to_phys(mpf));
719 701
720 if (!reserve) 702 mem = virt_to_phys(mpf);
721 return 1; 703 reserve_early(mem, mem + sizeof(*mpf), "MP-table mpf");
722 reserve_bootmem_generic(virt_to_phys(mpf), sizeof(*mpf),
723 BOOTMEM_DEFAULT);
724 if (mpf->physptr) 704 if (mpf->physptr)
725 smp_reserve_bootmem(mpf); 705 smp_reserve_memory(mpf);
726 706
727 return 1; 707 return 1;
728 } 708 }
@@ -732,7 +712,7 @@ static int __init smp_scan_config(unsigned long base, unsigned long length,
732 return 0; 712 return 0;
733} 713}
734 714
735void __init default_find_smp_config(unsigned int reserve) 715void __init default_find_smp_config(void)
736{ 716{
737 unsigned int address; 717 unsigned int address;
738 718
@@ -744,9 +724,9 @@ void __init default_find_smp_config(unsigned int reserve)
744 * 2) Scan the top 1K of base RAM 724 * 2) Scan the top 1K of base RAM
745 * 3) Scan the 64K of bios 725 * 3) Scan the 64K of bios
746 */ 726 */
747 if (smp_scan_config(0x0, 0x400, reserve) || 727 if (smp_scan_config(0x0, 0x400) ||
748 smp_scan_config(639 * 0x400, 0x400, reserve) || 728 smp_scan_config(639 * 0x400, 0x400) ||
749 smp_scan_config(0xF0000, 0x10000, reserve)) 729 smp_scan_config(0xF0000, 0x10000))
750 return; 730 return;
751 /* 731 /*
752 * If it is an SMP machine we should know now, unless the 732 * If it is an SMP machine we should know now, unless the
@@ -767,7 +747,7 @@ void __init default_find_smp_config(unsigned int reserve)
767 747
768 address = get_bios_ebda(); 748 address = get_bios_ebda();
769 if (address) 749 if (address)
770 smp_scan_config(address, 0x400, reserve); 750 smp_scan_config(address, 0x400);
771} 751}
772 752
773#ifdef CONFIG_X86_IO_APIC 753#ifdef CONFIG_X86_IO_APIC
diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c
index a98fe88fab64..c95c8f4e790a 100644
--- a/arch/x86/kernel/process_64.c
+++ b/arch/x86/kernel/process_64.c
@@ -349,26 +349,42 @@ out:
349 return err; 349 return err;
350} 350}
351 351
352void 352static void
353start_thread(struct pt_regs *regs, unsigned long new_ip, unsigned long new_sp) 353start_thread_common(struct pt_regs *regs, unsigned long new_ip,
354 unsigned long new_sp,
355 unsigned int _cs, unsigned int _ss, unsigned int _ds)
354{ 356{
355 loadsegment(fs, 0); 357 loadsegment(fs, 0);
356 loadsegment(es, 0); 358 loadsegment(es, _ds);
357 loadsegment(ds, 0); 359 loadsegment(ds, _ds);
358 load_gs_index(0); 360 load_gs_index(0);
359 regs->ip = new_ip; 361 regs->ip = new_ip;
360 regs->sp = new_sp; 362 regs->sp = new_sp;
361 percpu_write(old_rsp, new_sp); 363 percpu_write(old_rsp, new_sp);
362 regs->cs = __USER_CS; 364 regs->cs = _cs;
363 regs->ss = __USER_DS; 365 regs->ss = _ss;
364 regs->flags = 0x200; 366 regs->flags = X86_EFLAGS_IF;
365 set_fs(USER_DS); 367 set_fs(USER_DS);
366 /* 368 /*
367 * Free the old FP and other extended state 369 * Free the old FP and other extended state
368 */ 370 */
369 free_thread_xstate(current); 371 free_thread_xstate(current);
370} 372}
371EXPORT_SYMBOL_GPL(start_thread); 373
374void
375start_thread(struct pt_regs *regs, unsigned long new_ip, unsigned long new_sp)
376{
377 start_thread_common(regs, new_ip, new_sp,
378 __USER_CS, __USER_DS, 0);
379}
380
381#ifdef CONFIG_IA32_EMULATION
382void start_thread_ia32(struct pt_regs *regs, u32 new_ip, u32 new_sp)
383{
384 start_thread_common(regs, new_ip, new_sp,
385 __USER32_CS, __USER32_DS, __USER32_DS);
386}
387#endif
372 388
373/* 389/*
374 * switch_to(x,y) should switch tasks from x to y. 390 * switch_to(x,y) should switch tasks from x to y.
diff --git a/arch/x86/kernel/quirks.c b/arch/x86/kernel/quirks.c
index 6c3b2c6fd772..18093d7498f0 100644
--- a/arch/x86/kernel/quirks.c
+++ b/arch/x86/kernel/quirks.c
@@ -499,6 +499,7 @@ static void __init quirk_amd_nb_node(struct pci_dev *dev)
499{ 499{
500 struct pci_dev *nb_ht; 500 struct pci_dev *nb_ht;
501 unsigned int devfn; 501 unsigned int devfn;
502 u32 node;
502 u32 val; 503 u32 val;
503 504
504 devfn = PCI_DEVFN(PCI_SLOT(dev->devfn), 0); 505 devfn = PCI_DEVFN(PCI_SLOT(dev->devfn), 0);
@@ -507,7 +508,13 @@ static void __init quirk_amd_nb_node(struct pci_dev *dev)
507 return; 508 return;
508 509
509 pci_read_config_dword(nb_ht, 0x60, &val); 510 pci_read_config_dword(nb_ht, 0x60, &val);
510 set_dev_node(&dev->dev, val & 7); 511 node = val & 7;
512 /*
513 * Some hardware may return an invalid node ID,
514 * so check it first:
515 */
516 if (node_online(node))
517 set_dev_node(&dev->dev, node);
511 pci_dev_put(nb_ht); 518 pci_dev_put(nb_ht);
512} 519}
513 520
diff --git a/arch/x86/kernel/reboot_fixups_32.c b/arch/x86/kernel/reboot_fixups_32.c
index 61a837743fe5..201eab63b05f 100644
--- a/arch/x86/kernel/reboot_fixups_32.c
+++ b/arch/x86/kernel/reboot_fixups_32.c
@@ -80,6 +80,7 @@ void mach_reboot_fixups(void)
80 continue; 80 continue;
81 81
82 cur->reboot_fixup(dev); 82 cur->reboot_fixup(dev);
83 pci_dev_put(dev);
83 } 84 }
84} 85}
85 86
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 82e88cdda9bc..946a311a25c9 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -106,6 +106,7 @@
106#include <asm/percpu.h> 106#include <asm/percpu.h>
107#include <asm/topology.h> 107#include <asm/topology.h>
108#include <asm/apicdef.h> 108#include <asm/apicdef.h>
109#include <asm/k8.h>
109#ifdef CONFIG_X86_64 110#ifdef CONFIG_X86_64
110#include <asm/numa_64.h> 111#include <asm/numa_64.h>
111#endif 112#endif
@@ -487,42 +488,11 @@ static void __init reserve_early_setup_data(void)
487 488
488#ifdef CONFIG_KEXEC 489#ifdef CONFIG_KEXEC
489 490
490/**
491 * Reserve @size bytes of crashkernel memory at any suitable offset.
492 *
493 * @size: Size of the crashkernel memory to reserve.
494 * Returns the base address on success, and -1ULL on failure.
495 */
496static
497unsigned long long __init find_and_reserve_crashkernel(unsigned long long size)
498{
499 const unsigned long long alignment = 16<<20; /* 16M */
500 unsigned long long start = 0LL;
501
502 while (1) {
503 int ret;
504
505 start = find_e820_area(start, ULONG_MAX, size, alignment);
506 if (start == -1ULL)
507 return start;
508
509 /* try to reserve it */
510 ret = reserve_bootmem_generic(start, size, BOOTMEM_EXCLUSIVE);
511 if (ret >= 0)
512 return start;
513
514 start += alignment;
515 }
516}
517
518static inline unsigned long long get_total_mem(void) 491static inline unsigned long long get_total_mem(void)
519{ 492{
520 unsigned long long total; 493 unsigned long long total;
521 494
522 total = max_low_pfn - min_low_pfn; 495 total = max_pfn - min_low_pfn;
523#ifdef CONFIG_HIGHMEM
524 total += highend_pfn - highstart_pfn;
525#endif
526 496
527 return total << PAGE_SHIFT; 497 return total << PAGE_SHIFT;
528} 498}
@@ -542,21 +512,25 @@ static void __init reserve_crashkernel(void)
542 512
543 /* 0 means: find the address automatically */ 513 /* 0 means: find the address automatically */
544 if (crash_base <= 0) { 514 if (crash_base <= 0) {
545 crash_base = find_and_reserve_crashkernel(crash_size); 515 const unsigned long long alignment = 16<<20; /* 16M */
516
517 crash_base = find_e820_area(alignment, ULONG_MAX, crash_size,
518 alignment);
546 if (crash_base == -1ULL) { 519 if (crash_base == -1ULL) {
547 pr_info("crashkernel reservation failed. " 520 pr_info("crashkernel reservation failed - No suitable area found.\n");
548 "No suitable area found.\n");
549 return; 521 return;
550 } 522 }
551 } else { 523 } else {
552 ret = reserve_bootmem_generic(crash_base, crash_size, 524 unsigned long long start;
553 BOOTMEM_EXCLUSIVE); 525
554 if (ret < 0) { 526 start = find_e820_area(crash_base, ULONG_MAX, crash_size,
555 pr_info("crashkernel reservation failed - " 527 1<<20);
556 "memory is in use\n"); 528 if (start != crash_base) {
529 pr_info("crashkernel reservation failed - memory is in use.\n");
557 return; 530 return;
558 } 531 }
559 } 532 }
533 reserve_early(crash_base, crash_base + crash_size, "CRASH KERNEL");
560 534
561 printk(KERN_INFO "Reserving %ldMB of memory at %ldMB " 535 printk(KERN_INFO "Reserving %ldMB of memory at %ldMB "
562 "for crashkernel (System RAM: %ldMB)\n", 536 "for crashkernel (System RAM: %ldMB)\n",
@@ -699,6 +673,9 @@ static struct dmi_system_id __initdata bad_bios_dmi_table[] = {
699 673
700void __init setup_arch(char **cmdline_p) 674void __init setup_arch(char **cmdline_p)
701{ 675{
676 int acpi = 0;
677 int k8 = 0;
678
702#ifdef CONFIG_X86_32 679#ifdef CONFIG_X86_32
703 memcpy(&boot_cpu_data, &new_cpu_data, sizeof(new_cpu_data)); 680 memcpy(&boot_cpu_data, &new_cpu_data, sizeof(new_cpu_data));
704 visws_early_detect(); 681 visws_early_detect();
@@ -791,21 +768,18 @@ void __init setup_arch(char **cmdline_p)
791 strlcpy(command_line, boot_command_line, COMMAND_LINE_SIZE); 768 strlcpy(command_line, boot_command_line, COMMAND_LINE_SIZE);
792 *cmdline_p = command_line; 769 *cmdline_p = command_line;
793 770
794#ifdef CONFIG_X86_64
795 /* 771 /*
796 * Must call this twice: Once just to detect whether hardware doesn't 772 * x86_configure_nx() is called before parse_early_param() to detect
797 * support NX (so that the early EHCI debug console setup can safely 773 * whether hardware doesn't support NX (so that the early EHCI debug
798 * call set_fixmap(), and then again after parsing early parameters to 774 * console setup can safely call set_fixmap()). It may then be called
799 * honor the respective command line option. 775 * again from within noexec_setup() during parsing early parameters
776 * to honor the respective command line option.
800 */ 777 */
801 check_efer(); 778 x86_configure_nx();
802#endif
803 779
804 parse_early_param(); 780 parse_early_param();
805 781
806#ifdef CONFIG_X86_64 782 x86_report_nx();
807 check_efer();
808#endif
809 783
810 /* Must be before kernel pagetables are setup */ 784 /* Must be before kernel pagetables are setup */
811 vmi_activate(); 785 vmi_activate();
@@ -901,6 +875,13 @@ void __init setup_arch(char **cmdline_p)
901 875
902 reserve_brk(); 876 reserve_brk();
903 877
878#ifdef CONFIG_ACPI_SLEEP
879 /*
880 * Reserve low memory region for sleep support.
881 * even before init_memory_mapping
882 */
883 acpi_reserve_wakeup_memory();
884#endif
904 init_gbpages(); 885 init_gbpages();
905 886
906 /* max_pfn_mapped is updated here */ 887 /* max_pfn_mapped is updated here */
@@ -927,6 +908,8 @@ void __init setup_arch(char **cmdline_p)
927 908
928 reserve_initrd(); 909 reserve_initrd();
929 910
911 reserve_crashkernel();
912
930 vsmp_init(); 913 vsmp_init();
931 914
932 io_delay_init(); 915 io_delay_init();
@@ -938,27 +921,24 @@ void __init setup_arch(char **cmdline_p)
938 921
939 early_acpi_boot_init(); 922 early_acpi_boot_init();
940 923
924 /*
925 * Find and reserve possible boot-time SMP configuration:
926 */
927 find_smp_config();
928
941#ifdef CONFIG_ACPI_NUMA 929#ifdef CONFIG_ACPI_NUMA
942 /* 930 /*
943 * Parse SRAT to discover nodes. 931 * Parse SRAT to discover nodes.
944 */ 932 */
945 acpi_numa_init(); 933 acpi = acpi_numa_init();
946#endif 934#endif
947 935
948 initmem_init(0, max_pfn); 936#ifdef CONFIG_K8_NUMA
949 937 if (!acpi)
950#ifdef CONFIG_ACPI_SLEEP 938 k8 = !k8_numa_init(0, max_pfn);
951 /*
952 * Reserve low memory region for sleep support.
953 */
954 acpi_reserve_bootmem();
955#endif 939#endif
956 /*
957 * Find and reserve possible boot-time SMP configuration:
958 */
959 find_smp_config();
960 940
961 reserve_crashkernel(); 941 initmem_init(0, max_pfn, acpi, k8);
962 942
963#ifdef CONFIG_X86_64 943#ifdef CONFIG_X86_64
964 /* 944 /*
diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
index 324f2a44c221..29e6744f51e3 100644
--- a/arch/x86/kernel/smpboot.c
+++ b/arch/x86/kernel/smpboot.c
@@ -687,7 +687,7 @@ static int __cpuinit do_boot_cpu(int apicid, int cpu)
687 .done = COMPLETION_INITIALIZER_ONSTACK(c_idle.done), 687 .done = COMPLETION_INITIALIZER_ONSTACK(c_idle.done),
688 }; 688 };
689 689
690 INIT_WORK(&c_idle.work, do_fork_idle); 690 INIT_WORK_ON_STACK(&c_idle.work, do_fork_idle);
691 691
692 alternatives_smp_switch(1); 692 alternatives_smp_switch(1);
693 693
@@ -713,6 +713,7 @@ static int __cpuinit do_boot_cpu(int apicid, int cpu)
713 713
714 if (IS_ERR(c_idle.idle)) { 714 if (IS_ERR(c_idle.idle)) {
715 printk("failed fork for CPU %d\n", cpu); 715 printk("failed fork for CPU %d\n", cpu);
716 destroy_work_on_stack(&c_idle.work);
716 return PTR_ERR(c_idle.idle); 717 return PTR_ERR(c_idle.idle);
717 } 718 }
718 719
@@ -831,6 +832,7 @@ do_rest:
831 smpboot_restore_warm_reset_vector(); 832 smpboot_restore_warm_reset_vector();
832 } 833 }
833 834
835 destroy_work_on_stack(&c_idle.work);
834 return boot_error; 836 return boot_error;
835} 837}
836 838
diff --git a/arch/x86/kernel/tlb_uv.c b/arch/x86/kernel/tlb_uv.c
index 1740c85e24bb..364d015efebc 100644
--- a/arch/x86/kernel/tlb_uv.c
+++ b/arch/x86/kernel/tlb_uv.c
@@ -817,10 +817,8 @@ static int __init uv_init_blade(int blade)
817 */ 817 */
818 apicid = blade_to_first_apicid(blade); 818 apicid = blade_to_first_apicid(blade);
819 pa = uv_read_global_mmr64(pnode, UVH_BAU_DATA_CONFIG); 819 pa = uv_read_global_mmr64(pnode, UVH_BAU_DATA_CONFIG);
820 if ((pa & 0xff) != UV_BAU_MESSAGE) { 820 uv_write_global_mmr64(pnode, UVH_BAU_DATA_CONFIG,
821 uv_write_global_mmr64(pnode, UVH_BAU_DATA_CONFIG,
822 ((apicid << 32) | UV_BAU_MESSAGE)); 821 ((apicid << 32) | UV_BAU_MESSAGE));
823 }
824 return 0; 822 return 0;
825} 823}
826 824
diff --git a/arch/x86/kernel/uv_time.c b/arch/x86/kernel/uv_time.c
index 583f11d5c480..3c84aa001c11 100644
--- a/arch/x86/kernel/uv_time.c
+++ b/arch/x86/kernel/uv_time.c
@@ -74,7 +74,7 @@ struct uv_rtc_timer_head {
74 */ 74 */
75static struct uv_rtc_timer_head **blade_info __read_mostly; 75static struct uv_rtc_timer_head **blade_info __read_mostly;
76 76
77static int uv_rtc_enable; 77static int uv_rtc_evt_enable;
78 78
79/* 79/*
80 * Hardware interface routines 80 * Hardware interface routines
@@ -90,7 +90,7 @@ static void uv_rtc_send_IPI(int cpu)
90 pnode = uv_apicid_to_pnode(apicid); 90 pnode = uv_apicid_to_pnode(apicid);
91 val = (1UL << UVH_IPI_INT_SEND_SHFT) | 91 val = (1UL << UVH_IPI_INT_SEND_SHFT) |
92 (apicid << UVH_IPI_INT_APIC_ID_SHFT) | 92 (apicid << UVH_IPI_INT_APIC_ID_SHFT) |
93 (GENERIC_INTERRUPT_VECTOR << UVH_IPI_INT_VECTOR_SHFT); 93 (X86_PLATFORM_IPI_VECTOR << UVH_IPI_INT_VECTOR_SHFT);
94 94
95 uv_write_global_mmr64(pnode, UVH_IPI_INT, val); 95 uv_write_global_mmr64(pnode, UVH_IPI_INT, val);
96} 96}
@@ -115,7 +115,7 @@ static int uv_setup_intr(int cpu, u64 expires)
115 uv_write_global_mmr64(pnode, UVH_EVENT_OCCURRED0_ALIAS, 115 uv_write_global_mmr64(pnode, UVH_EVENT_OCCURRED0_ALIAS,
116 UVH_EVENT_OCCURRED0_RTC1_MASK); 116 UVH_EVENT_OCCURRED0_RTC1_MASK);
117 117
118 val = (GENERIC_INTERRUPT_VECTOR << UVH_RTC1_INT_CONFIG_VECTOR_SHFT) | 118 val = (X86_PLATFORM_IPI_VECTOR << UVH_RTC1_INT_CONFIG_VECTOR_SHFT) |
119 ((u64)cpu_physical_id(cpu) << UVH_RTC1_INT_CONFIG_APIC_ID_SHFT); 119 ((u64)cpu_physical_id(cpu) << UVH_RTC1_INT_CONFIG_APIC_ID_SHFT);
120 120
121 /* Set configuration */ 121 /* Set configuration */
@@ -123,7 +123,10 @@ static int uv_setup_intr(int cpu, u64 expires)
123 /* Initialize comparator value */ 123 /* Initialize comparator value */
124 uv_write_global_mmr64(pnode, UVH_INT_CMPB, expires); 124 uv_write_global_mmr64(pnode, UVH_INT_CMPB, expires);
125 125
126 return (expires < uv_read_rtc(NULL) && !uv_intr_pending(pnode)); 126 if (uv_read_rtc(NULL) <= expires)
127 return 0;
128
129 return !uv_intr_pending(pnode);
127} 130}
128 131
129/* 132/*
@@ -223,6 +226,7 @@ static int uv_rtc_set_timer(int cpu, u64 expires)
223 226
224 next_cpu = head->next_cpu; 227 next_cpu = head->next_cpu;
225 *t = expires; 228 *t = expires;
229
226 /* Will this one be next to go off? */ 230 /* Will this one be next to go off? */
227 if (next_cpu < 0 || bcpu == next_cpu || 231 if (next_cpu < 0 || bcpu == next_cpu ||
228 expires < head->cpu[next_cpu].expires) { 232 expires < head->cpu[next_cpu].expires) {
@@ -231,7 +235,7 @@ static int uv_rtc_set_timer(int cpu, u64 expires)
231 *t = ULLONG_MAX; 235 *t = ULLONG_MAX;
232 uv_rtc_find_next_timer(head, pnode); 236 uv_rtc_find_next_timer(head, pnode);
233 spin_unlock_irqrestore(&head->lock, flags); 237 spin_unlock_irqrestore(&head->lock, flags);
234 return 1; 238 return -ETIME;
235 } 239 }
236 } 240 }
237 241
@@ -244,7 +248,7 @@ static int uv_rtc_set_timer(int cpu, u64 expires)
244 * 248 *
245 * Returns 1 if this timer was pending. 249 * Returns 1 if this timer was pending.
246 */ 250 */
247static int uv_rtc_unset_timer(int cpu) 251static int uv_rtc_unset_timer(int cpu, int force)
248{ 252{
249 int pnode = uv_cpu_to_pnode(cpu); 253 int pnode = uv_cpu_to_pnode(cpu);
250 int bid = uv_cpu_to_blade_id(cpu); 254 int bid = uv_cpu_to_blade_id(cpu);
@@ -256,14 +260,15 @@ static int uv_rtc_unset_timer(int cpu)
256 260
257 spin_lock_irqsave(&head->lock, flags); 261 spin_lock_irqsave(&head->lock, flags);
258 262
259 if (head->next_cpu == bcpu && uv_read_rtc(NULL) >= *t) 263 if ((head->next_cpu == bcpu && uv_read_rtc(NULL) >= *t) || force)
260 rc = 1; 264 rc = 1;
261 265
262 *t = ULLONG_MAX; 266 if (rc) {
263 267 *t = ULLONG_MAX;
264 /* Was the hardware setup for this timer? */ 268 /* Was the hardware setup for this timer? */
265 if (head->next_cpu == bcpu) 269 if (head->next_cpu == bcpu)
266 uv_rtc_find_next_timer(head, pnode); 270 uv_rtc_find_next_timer(head, pnode);
271 }
267 272
268 spin_unlock_irqrestore(&head->lock, flags); 273 spin_unlock_irqrestore(&head->lock, flags);
269 274
@@ -310,32 +315,32 @@ static void uv_rtc_timer_setup(enum clock_event_mode mode,
310 break; 315 break;
311 case CLOCK_EVT_MODE_UNUSED: 316 case CLOCK_EVT_MODE_UNUSED:
312 case CLOCK_EVT_MODE_SHUTDOWN: 317 case CLOCK_EVT_MODE_SHUTDOWN:
313 uv_rtc_unset_timer(ced_cpu); 318 uv_rtc_unset_timer(ced_cpu, 1);
314 break; 319 break;
315 } 320 }
316} 321}
317 322
318static void uv_rtc_interrupt(void) 323static void uv_rtc_interrupt(void)
319{ 324{
320 struct clock_event_device *ced = &__get_cpu_var(cpu_ced);
321 int cpu = smp_processor_id(); 325 int cpu = smp_processor_id();
326 struct clock_event_device *ced = &per_cpu(cpu_ced, cpu);
322 327
323 if (!ced || !ced->event_handler) 328 if (!ced || !ced->event_handler)
324 return; 329 return;
325 330
326 if (uv_rtc_unset_timer(cpu) != 1) 331 if (uv_rtc_unset_timer(cpu, 0) != 1)
327 return; 332 return;
328 333
329 ced->event_handler(ced); 334 ced->event_handler(ced);
330} 335}
331 336
332static int __init uv_enable_rtc(char *str) 337static int __init uv_enable_evt_rtc(char *str)
333{ 338{
334 uv_rtc_enable = 1; 339 uv_rtc_evt_enable = 1;
335 340
336 return 1; 341 return 1;
337} 342}
338__setup("uvrtc", uv_enable_rtc); 343__setup("uvrtcevt", uv_enable_evt_rtc);
339 344
340static __init void uv_rtc_register_clockevents(struct work_struct *dummy) 345static __init void uv_rtc_register_clockevents(struct work_struct *dummy)
341{ 346{
@@ -350,27 +355,32 @@ static __init int uv_rtc_setup_clock(void)
350{ 355{
351 int rc; 356 int rc;
352 357
353 if (!uv_rtc_enable || !is_uv_system() || generic_interrupt_extension) 358 if (!is_uv_system())
354 return -ENODEV; 359 return -ENODEV;
355 360
356 generic_interrupt_extension = uv_rtc_interrupt;
357
358 clocksource_uv.mult = clocksource_hz2mult(sn_rtc_cycles_per_second, 361 clocksource_uv.mult = clocksource_hz2mult(sn_rtc_cycles_per_second,
359 clocksource_uv.shift); 362 clocksource_uv.shift);
360 363
364 /* If single blade, prefer tsc */
365 if (uv_num_possible_blades() == 1)
366 clocksource_uv.rating = 250;
367
361 rc = clocksource_register(&clocksource_uv); 368 rc = clocksource_register(&clocksource_uv);
362 if (rc) { 369 if (rc)
363 generic_interrupt_extension = NULL; 370 printk(KERN_INFO "UV RTC clocksource failed rc %d\n", rc);
371 else
372 printk(KERN_INFO "UV RTC clocksource registered freq %lu MHz\n",
373 sn_rtc_cycles_per_second/(unsigned long)1E6);
374
375 if (rc || !uv_rtc_evt_enable || x86_platform_ipi_callback)
364 return rc; 376 return rc;
365 }
366 377
367 /* Setup and register clockevents */ 378 /* Setup and register clockevents */
368 rc = uv_rtc_allocate_timers(); 379 rc = uv_rtc_allocate_timers();
369 if (rc) { 380 if (rc)
370 clocksource_unregister(&clocksource_uv); 381 goto error;
371 generic_interrupt_extension = NULL; 382
372 return rc; 383 x86_platform_ipi_callback = uv_rtc_interrupt;
373 }
374 384
375 clock_event_device_uv.mult = div_sc(sn_rtc_cycles_per_second, 385 clock_event_device_uv.mult = div_sc(sn_rtc_cycles_per_second,
376 NSEC_PER_SEC, clock_event_device_uv.shift); 386 NSEC_PER_SEC, clock_event_device_uv.shift);
@@ -383,11 +393,19 @@ static __init int uv_rtc_setup_clock(void)
383 393
384 rc = schedule_on_each_cpu(uv_rtc_register_clockevents); 394 rc = schedule_on_each_cpu(uv_rtc_register_clockevents);
385 if (rc) { 395 if (rc) {
386 clocksource_unregister(&clocksource_uv); 396 x86_platform_ipi_callback = NULL;
387 generic_interrupt_extension = NULL;
388 uv_rtc_deallocate_timers(); 397 uv_rtc_deallocate_timers();
398 goto error;
389 } 399 }
390 400
401 printk(KERN_INFO "UV RTC clockevents registered\n");
402
403 return 0;
404
405error:
406 clocksource_unregister(&clocksource_uv);
407 printk(KERN_INFO "UV RTC clockevents failed rc %d\n", rc);
408
391 return rc; 409 return rc;
392} 410}
393arch_initcall(uv_rtc_setup_clock); 411arch_initcall(uv_rtc_setup_clock);
diff --git a/arch/x86/kernel/visws_quirks.c b/arch/x86/kernel/visws_quirks.c
index abda6f53e71e..34a279a7471d 100644
--- a/arch/x86/kernel/visws_quirks.c
+++ b/arch/x86/kernel/visws_quirks.c
@@ -197,7 +197,7 @@ static void __init MP_processor_info(struct mpc_cpu *m)
197 apic_version[m->apicid] = ver; 197 apic_version[m->apicid] = ver;
198} 198}
199 199
200static void __init visws_find_smp_config(unsigned int reserve) 200static void __init visws_find_smp_config(void)
201{ 201{
202 struct mpc_cpu *mp = phys_to_virt(CO_CPU_TAB_PHYS); 202 struct mpc_cpu *mp = phys_to_virt(CO_CPU_TAB_PHYS);
203 unsigned short ncpus = readw(phys_to_virt(CO_CPU_NUM_PHYS)); 203 unsigned short ncpus = readw(phys_to_virt(CO_CPU_NUM_PHYS));
diff --git a/arch/x86/kernel/vmiclock_32.c b/arch/x86/kernel/vmiclock_32.c
index 611b9e2360d3..74c92bb194df 100644
--- a/arch/x86/kernel/vmiclock_32.c
+++ b/arch/x86/kernel/vmiclock_32.c
@@ -226,7 +226,7 @@ static void __devinit vmi_time_init_clockevent(void)
226 evt->min_delta_ns = clockevent_delta2ns(1, evt); 226 evt->min_delta_ns = clockevent_delta2ns(1, evt);
227 evt->cpumask = cpumask_of(cpu); 227 evt->cpumask = cpumask_of(cpu);
228 228
229 printk(KERN_WARNING "vmi: registering clock event %s. mult=%lu shift=%u\n", 229 printk(KERN_WARNING "vmi: registering clock event %s. mult=%u shift=%u\n",
230 evt->name, evt->mult, evt->shift); 230 evt->name, evt->mult, evt->shift);
231 clockevents_register_device(evt); 231 clockevents_register_device(evt);
232} 232}
diff --git a/arch/x86/kernel/vmlinux.lds.S b/arch/x86/kernel/vmlinux.lds.S
index 3c68fe2d46cf..f3f2104408d9 100644
--- a/arch/x86/kernel/vmlinux.lds.S
+++ b/arch/x86/kernel/vmlinux.lds.S
@@ -41,6 +41,32 @@ ENTRY(phys_startup_64)
41jiffies_64 = jiffies; 41jiffies_64 = jiffies;
42#endif 42#endif
43 43
44#if defined(CONFIG_X86_64) && defined(CONFIG_DEBUG_RODATA)
45/*
46 * On 64-bit, align RODATA to 2MB so that even with CONFIG_DEBUG_RODATA
47 * we retain large page mappings for boundaries spanning kernel text, rodata
48 * and data sections.
49 *
50 * However, kernel identity mappings will have different RWX permissions
51 * to the pages mapping to text and to the pages padding (which are freed) the
52 * text section. Hence kernel identity mappings will be broken to smaller
53 * pages. For 64-bit, kernel text and kernel identity mappings are different,
54 * so we can enable protection checks that come with CONFIG_DEBUG_RODATA,
55 * as well as retain 2MB large page mappings for kernel text.
56 */
57#define X64_ALIGN_DEBUG_RODATA_BEGIN . = ALIGN(HPAGE_SIZE);
58
59#define X64_ALIGN_DEBUG_RODATA_END \
60 . = ALIGN(HPAGE_SIZE); \
61 __end_rodata_hpage_align = .;
62
63#else
64
65#define X64_ALIGN_DEBUG_RODATA_BEGIN
66#define X64_ALIGN_DEBUG_RODATA_END
67
68#endif
69
44PHDRS { 70PHDRS {
45 text PT_LOAD FLAGS(5); /* R_E */ 71 text PT_LOAD FLAGS(5); /* R_E */
46 data PT_LOAD FLAGS(7); /* RWE */ 72 data PT_LOAD FLAGS(7); /* RWE */
@@ -90,7 +116,9 @@ SECTIONS
90 116
91 EXCEPTION_TABLE(16) :text = 0x9090 117 EXCEPTION_TABLE(16) :text = 0x9090
92 118
119 X64_ALIGN_DEBUG_RODATA_BEGIN
93 RO_DATA(PAGE_SIZE) 120 RO_DATA(PAGE_SIZE)
121 X64_ALIGN_DEBUG_RODATA_END
94 122
95 /* Data */ 123 /* Data */
96 .data : AT(ADDR(.data) - LOAD_OFFSET) { 124 .data : AT(ADDR(.data) - LOAD_OFFSET) {
@@ -107,13 +135,13 @@ SECTIONS
107 135
108 PAGE_ALIGNED_DATA(PAGE_SIZE) 136 PAGE_ALIGNED_DATA(PAGE_SIZE)
109 137
110 CACHELINE_ALIGNED_DATA(CONFIG_X86_L1_CACHE_BYTES) 138 CACHELINE_ALIGNED_DATA(L1_CACHE_BYTES)
111 139
112 DATA_DATA 140 DATA_DATA
113 CONSTRUCTORS 141 CONSTRUCTORS
114 142
115 /* rarely changed data like cpu maps */ 143 /* rarely changed data like cpu maps */
116 READ_MOSTLY_DATA(CONFIG_X86_INTERNODE_CACHE_BYTES) 144 READ_MOSTLY_DATA(INTERNODE_CACHE_BYTES)
117 145
118 /* End of data section */ 146 /* End of data section */
119 _edata = .; 147 _edata = .;
@@ -137,12 +165,12 @@ SECTIONS
137 *(.vsyscall_0) 165 *(.vsyscall_0)
138 } :user 166 } :user
139 167
140 . = ALIGN(CONFIG_X86_L1_CACHE_BYTES); 168 . = ALIGN(L1_CACHE_BYTES);
141 .vsyscall_fn : AT(VLOAD(.vsyscall_fn)) { 169 .vsyscall_fn : AT(VLOAD(.vsyscall_fn)) {
142 *(.vsyscall_fn) 170 *(.vsyscall_fn)
143 } 171 }
144 172
145 . = ALIGN(CONFIG_X86_L1_CACHE_BYTES); 173 . = ALIGN(L1_CACHE_BYTES);
146 .vsyscall_gtod_data : AT(VLOAD(.vsyscall_gtod_data)) { 174 .vsyscall_gtod_data : AT(VLOAD(.vsyscall_gtod_data)) {
147 *(.vsyscall_gtod_data) 175 *(.vsyscall_gtod_data)
148 } 176 }
@@ -166,7 +194,7 @@ SECTIONS
166 } 194 }
167 vgetcpu_mode = VVIRT(.vgetcpu_mode); 195 vgetcpu_mode = VVIRT(.vgetcpu_mode);
168 196
169 . = ALIGN(CONFIG_X86_L1_CACHE_BYTES); 197 . = ALIGN(L1_CACHE_BYTES);
170 .jiffies : AT(VLOAD(.jiffies)) { 198 .jiffies : AT(VLOAD(.jiffies)) {
171 *(.jiffies) 199 *(.jiffies)
172 } 200 }
diff --git a/arch/x86/kernel/vsyscall_64.c b/arch/x86/kernel/vsyscall_64.c
index e02d92d12bcd..9055e5872ff0 100644
--- a/arch/x86/kernel/vsyscall_64.c
+++ b/arch/x86/kernel/vsyscall_64.c
@@ -73,7 +73,8 @@ void update_vsyscall_tz(void)
73 write_sequnlock_irqrestore(&vsyscall_gtod_data.lock, flags); 73 write_sequnlock_irqrestore(&vsyscall_gtod_data.lock, flags);
74} 74}
75 75
76void update_vsyscall(struct timespec *wall_time, struct clocksource *clock) 76void update_vsyscall(struct timespec *wall_time, struct clocksource *clock,
77 u32 mult)
77{ 78{
78 unsigned long flags; 79 unsigned long flags;
79 80
@@ -82,7 +83,7 @@ void update_vsyscall(struct timespec *wall_time, struct clocksource *clock)
82 vsyscall_gtod_data.clock.vread = clock->vread; 83 vsyscall_gtod_data.clock.vread = clock->vread;
83 vsyscall_gtod_data.clock.cycle_last = clock->cycle_last; 84 vsyscall_gtod_data.clock.cycle_last = clock->cycle_last;
84 vsyscall_gtod_data.clock.mask = clock->mask; 85 vsyscall_gtod_data.clock.mask = clock->mask;
85 vsyscall_gtod_data.clock.mult = clock->mult; 86 vsyscall_gtod_data.clock.mult = mult;
86 vsyscall_gtod_data.clock.shift = clock->shift; 87 vsyscall_gtod_data.clock.shift = clock->shift;
87 vsyscall_gtod_data.wall_time_sec = wall_time->tv_sec; 88 vsyscall_gtod_data.wall_time_sec = wall_time->tv_sec;
88 vsyscall_gtod_data.wall_time_nsec = wall_time->tv_nsec; 89 vsyscall_gtod_data.wall_time_nsec = wall_time->tv_nsec;
diff --git a/arch/x86/kernel/x86_init.c b/arch/x86/kernel/x86_init.c
index d11c5ff7c65e..ccd179dec36e 100644
--- a/arch/x86/kernel/x86_init.c
+++ b/arch/x86/kernel/x86_init.c
@@ -13,6 +13,7 @@
13#include <asm/e820.h> 13#include <asm/e820.h>
14#include <asm/time.h> 14#include <asm/time.h>
15#include <asm/irq.h> 15#include <asm/irq.h>
16#include <asm/pat.h>
16#include <asm/tsc.h> 17#include <asm/tsc.h>
17#include <asm/iommu.h> 18#include <asm/iommu.h>
18 19
@@ -80,4 +81,5 @@ struct x86_platform_ops x86_platform = {
80 .get_wallclock = mach_get_cmos_time, 81 .get_wallclock = mach_get_cmos_time,
81 .set_wallclock = mach_set_rtc_mmss, 82 .set_wallclock = mach_set_rtc_mmss,
82 .iommu_shutdown = iommu_shutdown_noop, 83 .iommu_shutdown = iommu_shutdown_noop,
84 .is_untracked_pat_range = is_ISA_range,
83}; 85};
generated by cgit v1.2.2 (git 2.25.0) at 2025-01-04 01:31:25 -0500