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.c4
-rw-r--r--arch/x86/kernel/acpi/sleep.c2
-rw-r--r--arch/x86/kernel/alternative.c1
-rw-r--r--arch/x86/kernel/amd_iommu.c22
-rw-r--r--arch/x86/kernel/amd_iommu_init.c50
-rw-r--r--arch/x86/kernel/apb_timer.c3
-rw-r--r--arch/x86/kernel/aperture_64.c15
-rw-r--r--arch/x86/kernel/apic/apic.c2
-rw-r--r--arch/x86/kernel/apic/es7000_32.c1
-rw-r--r--arch/x86/kernel/apic/io_apic.c12
-rw-r--r--arch/x86/kernel/apic/nmi.c1
-rw-r--r--arch/x86/kernel/apic/x2apic_uv_x.c1
-rw-r--r--arch/x86/kernel/bootflag.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/elanfreq.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/gx-suspmod.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/longrun.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/p4-clockmod.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/powernow-k6.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/powernow-k8.c3
-rw-r--r--arch/x86/kernel/cpu/cpufreq/speedstep-centrino.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/speedstep-ich.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/speedstep-lib.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/speedstep-smi.c1
-rw-r--r--arch/x86/kernel/cpu/intel.c21
-rw-r--r--arch/x86/kernel/cpu/intel_cacheinfo.c4
-rw-r--r--arch/x86/kernel/cpu/mcheck/Makefile2
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce-apei.c138
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce-inject.c1
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce-internal.h23
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce.c80
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce_amd.c1
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce_intel.c1
-rw-r--r--arch/x86/kernel/cpu/mtrr/generic.c1
-rw-r--r--arch/x86/kernel/cpu/mtrr/if.c1
-rw-r--r--arch/x86/kernel/cpu/perf_event.c228
-rw-r--r--arch/x86/kernel/cpu/perf_event_amd.c142
-rw-r--r--arch/x86/kernel/cpu/perf_event_intel.c58
-rw-r--r--arch/x86/kernel/cpu/perf_event_p6.c10
-rw-r--r--arch/x86/kernel/cpu/vmware.c2
-rw-r--r--arch/x86/kernel/cpuid.c1
-rw-r--r--arch/x86/kernel/crash.c6
-rw-r--r--arch/x86/kernel/crash_dump_32.c1
-rw-r--r--arch/x86/kernel/dumpstack.h24
-rw-r--r--arch/x86/kernel/dumpstack_64.c4
-rw-r--r--arch/x86/kernel/e820.c24
-rw-r--r--arch/x86/kernel/head32.c4
-rw-r--r--arch/x86/kernel/head64.c3
-rw-r--r--arch/x86/kernel/hpet.c10
-rw-r--r--arch/x86/kernel/i387.c1
-rw-r--r--arch/x86/kernel/i8259.c1
-rw-r--r--arch/x86/kernel/irqinit.c23
-rw-r--r--arch/x86/kernel/k8.c2
-rw-r--r--arch/x86/kernel/kdebugfs.c1
-rw-r--r--arch/x86/kernel/kgdb.c2
-rw-r--r--arch/x86/kernel/kprobes.c27
-rw-r--r--arch/x86/kernel/ldt.c1
-rw-r--r--arch/x86/kernel/machine_kexec_64.c1
-rw-r--r--arch/x86/kernel/mca_32.c1
-rw-r--r--arch/x86/kernel/module.c1
-rw-r--r--arch/x86/kernel/mpparse.c4
-rw-r--r--arch/x86/kernel/msr.c1
-rw-r--r--arch/x86/kernel/pci-dma.c1
-rw-r--r--arch/x86/kernel/pci-gart_64.c4
-rw-r--r--arch/x86/kernel/pci-nommu.c1
-rw-r--r--arch/x86/kernel/process.c34
-rw-r--r--arch/x86/kernel/process_64.c4
-rw-r--r--arch/x86/kernel/ptrace.c1
-rw-r--r--arch/x86/kernel/setup.c25
-rw-r--r--arch/x86/kernel/smp.c1
-rw-r--r--arch/x86/kernel/smpboot.c7
-rw-r--r--arch/x86/kernel/tlb_uv.c1
-rw-r--r--arch/x86/kernel/uv_irq.c1
-rw-r--r--arch/x86/kernel/uv_time.c1
-rw-r--r--arch/x86/kernel/vmi_32.c1
-rw-r--r--arch/x86/kernel/vmlinux.lds.S2
77 files changed, 775 insertions, 294 deletions
diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c
index 63bcf39f8f24..b7023430d4dc 100644
--- a/arch/x86/kernel/acpi/boot.c
+++ b/arch/x86/kernel/acpi/boot.c
@@ -31,6 +31,7 @@
31#include <linux/module.h> 31#include <linux/module.h>
32#include <linux/dmi.h> 32#include <linux/dmi.h>
33#include <linux/irq.h> 33#include <linux/irq.h>
34#include <linux/slab.h>
34#include <linux/bootmem.h> 35#include <linux/bootmem.h>
35#include <linux/ioport.h> 36#include <linux/ioport.h>
36#include <linux/pci.h> 37#include <linux/pci.h>
@@ -489,6 +490,7 @@ int acpi_register_gsi(struct device *dev, u32 gsi, int trigger, int polarity)
489 * ACPI based hotplug support for CPU 490 * ACPI based hotplug support for CPU
490 */ 491 */
491#ifdef CONFIG_ACPI_HOTPLUG_CPU 492#ifdef CONFIG_ACPI_HOTPLUG_CPU
493#include <acpi/processor.h>
492 494
493static void acpi_map_cpu2node(acpi_handle handle, int cpu, int physid) 495static void acpi_map_cpu2node(acpi_handle handle, int cpu, int physid)
494{ 496{
@@ -566,6 +568,8 @@ static int __cpuinit _acpi_map_lsapic(acpi_handle handle, int *pcpu)
566 goto free_new_map; 568 goto free_new_map;
567 } 569 }
568 570
571 acpi_processor_set_pdc(handle);
572
569 cpu = cpumask_first(new_map); 573 cpu = cpumask_first(new_map);
570 acpi_map_cpu2node(handle, cpu, physid); 574 acpi_map_cpu2node(handle, cpu, physid);
571 575
diff --git a/arch/x86/kernel/acpi/sleep.c b/arch/x86/kernel/acpi/sleep.c
index f9961034e557..82e508677b91 100644
--- a/arch/x86/kernel/acpi/sleep.c
+++ b/arch/x86/kernel/acpi/sleep.c
@@ -162,8 +162,6 @@ static int __init acpi_sleep_setup(char *str)
162#endif 162#endif
163 if (strncmp(str, "old_ordering", 12) == 0) 163 if (strncmp(str, "old_ordering", 12) == 0)
164 acpi_old_suspend_ordering(); 164 acpi_old_suspend_ordering();
165 if (strncmp(str, "sci_force_enable", 16) == 0)
166 acpi_set_sci_en_on_resume();
167 str = strchr(str, ','); 165 str = strchr(str, ',');
168 if (str != NULL) 166 if (str != NULL)
169 str += strspn(str, ", \t"); 167 str += strspn(str, ", \t");
diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c
index 3a4bf35c179b..1a160d5d44d0 100644
--- a/arch/x86/kernel/alternative.c
+++ b/arch/x86/kernel/alternative.c
@@ -8,6 +8,7 @@
8#include <linux/vmalloc.h> 8#include <linux/vmalloc.h>
9#include <linux/memory.h> 9#include <linux/memory.h>
10#include <linux/stop_machine.h> 10#include <linux/stop_machine.h>
11#include <linux/slab.h>
11#include <asm/alternative.h> 12#include <asm/alternative.h>
12#include <asm/sections.h> 13#include <asm/sections.h>
13#include <asm/pgtable.h> 14#include <asm/pgtable.h>
diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c
index adb0ba025702..f854d89b7edf 100644
--- a/arch/x86/kernel/amd_iommu.c
+++ b/arch/x86/kernel/amd_iommu.c
@@ -18,8 +18,8 @@
18 */ 18 */
19 19
20#include <linux/pci.h> 20#include <linux/pci.h>
21#include <linux/gfp.h>
22#include <linux/bitmap.h> 21#include <linux/bitmap.h>
22#include <linux/slab.h>
23#include <linux/debugfs.h> 23#include <linux/debugfs.h>
24#include <linux/scatterlist.h> 24#include <linux/scatterlist.h>
25#include <linux/dma-mapping.h> 25#include <linux/dma-mapping.h>
@@ -118,7 +118,7 @@ static bool check_device(struct device *dev)
118 return false; 118 return false;
119 119
120 /* No device or no PCI device */ 120 /* No device or no PCI device */
121 if (!dev || dev->bus != &pci_bus_type) 121 if (dev->bus != &pci_bus_type)
122 return false; 122 return false;
123 123
124 devid = get_device_id(dev); 124 devid = get_device_id(dev);
@@ -392,6 +392,7 @@ static int __iommu_queue_command(struct amd_iommu *iommu, struct iommu_cmd *cmd)
392 u32 tail, head; 392 u32 tail, head;
393 u8 *target; 393 u8 *target;
394 394
395 WARN_ON(iommu->cmd_buf_size & CMD_BUFFER_UNINITIALIZED);
395 tail = readl(iommu->mmio_base + MMIO_CMD_TAIL_OFFSET); 396 tail = readl(iommu->mmio_base + MMIO_CMD_TAIL_OFFSET);
396 target = iommu->cmd_buf + tail; 397 target = iommu->cmd_buf + tail;
397 memcpy_toio(target, cmd, sizeof(*cmd)); 398 memcpy_toio(target, cmd, sizeof(*cmd));
@@ -2186,7 +2187,7 @@ static void prealloc_protection_domains(void)
2186 struct dma_ops_domain *dma_dom; 2187 struct dma_ops_domain *dma_dom;
2187 u16 devid; 2188 u16 devid;
2188 2189
2189 while ((dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, dev)) != NULL) { 2190 for_each_pci_dev(dev) {
2190 2191
2191 /* Do we handle this device? */ 2192 /* Do we handle this device? */
2192 if (!check_device(&dev->dev)) 2193 if (!check_device(&dev->dev))
@@ -2298,7 +2299,7 @@ static void cleanup_domain(struct protection_domain *domain)
2298 list_for_each_entry_safe(dev_data, next, &domain->dev_list, list) { 2299 list_for_each_entry_safe(dev_data, next, &domain->dev_list, list) {
2299 struct device *dev = dev_data->dev; 2300 struct device *dev = dev_data->dev;
2300 2301
2301 do_detach(dev); 2302 __detach_device(dev);
2302 atomic_set(&dev_data->bind, 0); 2303 atomic_set(&dev_data->bind, 0);
2303 } 2304 }
2304 2305
@@ -2327,6 +2328,7 @@ static struct protection_domain *protection_domain_alloc(void)
2327 return NULL; 2328 return NULL;
2328 2329
2329 spin_lock_init(&domain->lock); 2330 spin_lock_init(&domain->lock);
2331 mutex_init(&domain->api_lock);
2330 domain->id = domain_id_alloc(); 2332 domain->id = domain_id_alloc();
2331 if (!domain->id) 2333 if (!domain->id)
2332 goto out_err; 2334 goto out_err;
@@ -2379,9 +2381,7 @@ static void amd_iommu_domain_destroy(struct iommu_domain *dom)
2379 2381
2380 free_pagetable(domain); 2382 free_pagetable(domain);
2381 2383
2382 domain_id_free(domain->id); 2384 protection_domain_free(domain);
2383
2384 kfree(domain);
2385 2385
2386 dom->priv = NULL; 2386 dom->priv = NULL;
2387} 2387}
@@ -2456,6 +2456,8 @@ static int amd_iommu_map_range(struct iommu_domain *dom,
2456 iova &= PAGE_MASK; 2456 iova &= PAGE_MASK;
2457 paddr &= PAGE_MASK; 2457 paddr &= PAGE_MASK;
2458 2458
2459 mutex_lock(&domain->api_lock);
2460
2459 for (i = 0; i < npages; ++i) { 2461 for (i = 0; i < npages; ++i) {
2460 ret = iommu_map_page(domain, iova, paddr, prot, PM_MAP_4k); 2462 ret = iommu_map_page(domain, iova, paddr, prot, PM_MAP_4k);
2461 if (ret) 2463 if (ret)
@@ -2465,6 +2467,8 @@ static int amd_iommu_map_range(struct iommu_domain *dom,
2465 paddr += PAGE_SIZE; 2467 paddr += PAGE_SIZE;
2466 } 2468 }
2467 2469
2470 mutex_unlock(&domain->api_lock);
2471
2468 return 0; 2472 return 0;
2469} 2473}
2470 2474
@@ -2477,12 +2481,16 @@ static void amd_iommu_unmap_range(struct iommu_domain *dom,
2477 2481
2478 iova &= PAGE_MASK; 2482 iova &= PAGE_MASK;
2479 2483
2484 mutex_lock(&domain->api_lock);
2485
2480 for (i = 0; i < npages; ++i) { 2486 for (i = 0; i < npages; ++i) {
2481 iommu_unmap_page(domain, iova, PM_MAP_4k); 2487 iommu_unmap_page(domain, iova, PM_MAP_4k);
2482 iova += PAGE_SIZE; 2488 iova += PAGE_SIZE;
2483 } 2489 }
2484 2490
2485 iommu_flush_tlb_pde(domain); 2491 iommu_flush_tlb_pde(domain);
2492
2493 mutex_unlock(&domain->api_lock);
2486} 2494}
2487 2495
2488static phys_addr_t amd_iommu_iova_to_phys(struct iommu_domain *dom, 2496static phys_addr_t amd_iommu_iova_to_phys(struct iommu_domain *dom,
diff --git a/arch/x86/kernel/amd_iommu_init.c b/arch/x86/kernel/amd_iommu_init.c
index 9dc91b431470..6360abf993d4 100644
--- a/arch/x86/kernel/amd_iommu_init.c
+++ b/arch/x86/kernel/amd_iommu_init.c
@@ -19,8 +19,8 @@
19 19
20#include <linux/pci.h> 20#include <linux/pci.h>
21#include <linux/acpi.h> 21#include <linux/acpi.h>
22#include <linux/gfp.h>
23#include <linux/list.h> 22#include <linux/list.h>
23#include <linux/slab.h>
24#include <linux/sysdev.h> 24#include <linux/sysdev.h>
25#include <linux/interrupt.h> 25#include <linux/interrupt.h>
26#include <linux/msi.h> 26#include <linux/msi.h>
@@ -138,9 +138,9 @@ int amd_iommus_present;
138bool amd_iommu_np_cache __read_mostly; 138bool amd_iommu_np_cache __read_mostly;
139 139
140/* 140/*
141 * Set to true if ACPI table parsing and hardware intialization went properly 141 * The ACPI table parsing functions set this variable on an error
142 */ 142 */
143static bool amd_iommu_initialized; 143static int __initdata amd_iommu_init_err;
144 144
145/* 145/*
146 * List of protection domains - used during resume 146 * List of protection domains - used during resume
@@ -391,9 +391,11 @@ static int __init find_last_devid_acpi(struct acpi_table_header *table)
391 */ 391 */
392 for (i = 0; i < table->length; ++i) 392 for (i = 0; i < table->length; ++i)
393 checksum += p[i]; 393 checksum += p[i];
394 if (checksum != 0) 394 if (checksum != 0) {
395 /* ACPI table corrupt */ 395 /* ACPI table corrupt */
396 return -ENODEV; 396 amd_iommu_init_err = -ENODEV;
397 return 0;
398 }
397 399
398 p += IVRS_HEADER_LENGTH; 400 p += IVRS_HEADER_LENGTH;
399 401
@@ -436,7 +438,7 @@ static u8 * __init alloc_command_buffer(struct amd_iommu *iommu)
436 if (cmd_buf == NULL) 438 if (cmd_buf == NULL)
437 return NULL; 439 return NULL;
438 440
439 iommu->cmd_buf_size = CMD_BUFFER_SIZE; 441 iommu->cmd_buf_size = CMD_BUFFER_SIZE | CMD_BUFFER_UNINITIALIZED;
440 442
441 return cmd_buf; 443 return cmd_buf;
442} 444}
@@ -472,12 +474,13 @@ static void iommu_enable_command_buffer(struct amd_iommu *iommu)
472 &entry, sizeof(entry)); 474 &entry, sizeof(entry));
473 475
474 amd_iommu_reset_cmd_buffer(iommu); 476 amd_iommu_reset_cmd_buffer(iommu);
477 iommu->cmd_buf_size &= ~(CMD_BUFFER_UNINITIALIZED);
475} 478}
476 479
477static void __init free_command_buffer(struct amd_iommu *iommu) 480static void __init free_command_buffer(struct amd_iommu *iommu)
478{ 481{
479 free_pages((unsigned long)iommu->cmd_buf, 482 free_pages((unsigned long)iommu->cmd_buf,
480 get_order(iommu->cmd_buf_size)); 483 get_order(iommu->cmd_buf_size & ~(CMD_BUFFER_UNINITIALIZED)));
481} 484}
482 485
483/* allocates the memory where the IOMMU will log its events to */ 486/* allocates the memory where the IOMMU will log its events to */
@@ -920,11 +923,16 @@ static int __init init_iommu_all(struct acpi_table_header *table)
920 h->mmio_phys); 923 h->mmio_phys);
921 924
922 iommu = kzalloc(sizeof(struct amd_iommu), GFP_KERNEL); 925 iommu = kzalloc(sizeof(struct amd_iommu), GFP_KERNEL);
923 if (iommu == NULL) 926 if (iommu == NULL) {
924 return -ENOMEM; 927 amd_iommu_init_err = -ENOMEM;
928 return 0;
929 }
930
925 ret = init_iommu_one(iommu, h); 931 ret = init_iommu_one(iommu, h);
926 if (ret) 932 if (ret) {
927 return ret; 933 amd_iommu_init_err = ret;
934 return 0;
935 }
928 break; 936 break;
929 default: 937 default:
930 break; 938 break;
@@ -934,8 +942,6 @@ static int __init init_iommu_all(struct acpi_table_header *table)
934 } 942 }
935 WARN_ON(p != end); 943 WARN_ON(p != end);
936 944
937 amd_iommu_initialized = true;
938
939 return 0; 945 return 0;
940} 946}
941 947
@@ -1211,6 +1217,10 @@ static int __init amd_iommu_init(void)
1211 if (acpi_table_parse("IVRS", find_last_devid_acpi) != 0) 1217 if (acpi_table_parse("IVRS", find_last_devid_acpi) != 0)
1212 return -ENODEV; 1218 return -ENODEV;
1213 1219
1220 ret = amd_iommu_init_err;
1221 if (ret)
1222 goto out;
1223
1214 dev_table_size = tbl_size(DEV_TABLE_ENTRY_SIZE); 1224 dev_table_size = tbl_size(DEV_TABLE_ENTRY_SIZE);
1215 alias_table_size = tbl_size(ALIAS_TABLE_ENTRY_SIZE); 1225 alias_table_size = tbl_size(ALIAS_TABLE_ENTRY_SIZE);
1216 rlookup_table_size = tbl_size(RLOOKUP_TABLE_ENTRY_SIZE); 1226 rlookup_table_size = tbl_size(RLOOKUP_TABLE_ENTRY_SIZE);
@@ -1270,12 +1280,19 @@ static int __init amd_iommu_init(void)
1270 if (acpi_table_parse("IVRS", init_iommu_all) != 0) 1280 if (acpi_table_parse("IVRS", init_iommu_all) != 0)
1271 goto free; 1281 goto free;
1272 1282
1273 if (!amd_iommu_initialized) 1283 if (amd_iommu_init_err) {
1284 ret = amd_iommu_init_err;
1274 goto free; 1285 goto free;
1286 }
1275 1287
1276 if (acpi_table_parse("IVRS", init_memory_definitions) != 0) 1288 if (acpi_table_parse("IVRS", init_memory_definitions) != 0)
1277 goto free; 1289 goto free;
1278 1290
1291 if (amd_iommu_init_err) {
1292 ret = amd_iommu_init_err;
1293 goto free;
1294 }
1295
1279 ret = sysdev_class_register(&amd_iommu_sysdev_class); 1296 ret = sysdev_class_register(&amd_iommu_sysdev_class);
1280 if (ret) 1297 if (ret)
1281 goto free; 1298 goto free;
@@ -1288,6 +1305,8 @@ static int __init amd_iommu_init(void)
1288 if (ret) 1305 if (ret)
1289 goto free; 1306 goto free;
1290 1307
1308 enable_iommus();
1309
1291 if (iommu_pass_through) 1310 if (iommu_pass_through)
1292 ret = amd_iommu_init_passthrough(); 1311 ret = amd_iommu_init_passthrough();
1293 else 1312 else
@@ -1300,8 +1319,6 @@ static int __init amd_iommu_init(void)
1300 1319
1301 amd_iommu_init_notifier(); 1320 amd_iommu_init_notifier();
1302 1321
1303 enable_iommus();
1304
1305 if (iommu_pass_through) 1322 if (iommu_pass_through)
1306 goto out; 1323 goto out;
1307 1324
@@ -1315,6 +1332,7 @@ out:
1315 return ret; 1332 return ret;
1316 1333
1317free: 1334free:
1335 disable_iommus();
1318 1336
1319 amd_iommu_uninit_devices(); 1337 amd_iommu_uninit_devices();
1320 1338
diff --git a/arch/x86/kernel/apb_timer.c b/arch/x86/kernel/apb_timer.c
index 4b7099526d2c..a35347501d36 100644
--- a/arch/x86/kernel/apb_timer.c
+++ b/arch/x86/kernel/apb_timer.c
@@ -33,6 +33,7 @@
33#include <linux/errno.h> 33#include <linux/errno.h>
34#include <linux/init.h> 34#include <linux/init.h>
35#include <linux/sysdev.h> 35#include <linux/sysdev.h>
36#include <linux/slab.h>
36#include <linux/pm.h> 37#include <linux/pm.h>
37#include <linux/pci.h> 38#include <linux/pci.h>
38#include <linux/sfi.h> 39#include <linux/sfi.h>
@@ -428,7 +429,7 @@ static int apbt_cpuhp_notify(struct notifier_block *n,
428 429
429static __init int apbt_late_init(void) 430static __init int apbt_late_init(void)
430{ 431{
431 if (disable_apbt_percpu) 432 if (disable_apbt_percpu || !apb_timer_block_enabled)
432 return 0; 433 return 0;
433 /* This notifier should be called after workqueue is ready */ 434 /* This notifier should be called after workqueue is ready */
434 hotcpu_notifier(apbt_cpuhp_notify, -20); 435 hotcpu_notifier(apbt_cpuhp_notify, -20);
diff --git a/arch/x86/kernel/aperture_64.c b/arch/x86/kernel/aperture_64.c
index 3704997e8b25..b5d8b0bcf235 100644
--- a/arch/x86/kernel/aperture_64.c
+++ b/arch/x86/kernel/aperture_64.c
@@ -393,6 +393,7 @@ void __init gart_iommu_hole_init(void)
393 for (i = 0; i < ARRAY_SIZE(bus_dev_ranges); i++) { 393 for (i = 0; i < ARRAY_SIZE(bus_dev_ranges); i++) {
394 int bus; 394 int bus;
395 int dev_base, dev_limit; 395 int dev_base, dev_limit;
396 u32 ctl;
396 397
397 bus = bus_dev_ranges[i].bus; 398 bus = bus_dev_ranges[i].bus;
398 dev_base = bus_dev_ranges[i].dev_base; 399 dev_base = bus_dev_ranges[i].dev_base;
@@ -406,7 +407,19 @@ void __init gart_iommu_hole_init(void)
406 gart_iommu_aperture = 1; 407 gart_iommu_aperture = 1;
407 x86_init.iommu.iommu_init = gart_iommu_init; 408 x86_init.iommu.iommu_init = gart_iommu_init;
408 409
409 aper_order = (read_pci_config(bus, slot, 3, AMD64_GARTAPERTURECTL) >> 1) & 7; 410 ctl = read_pci_config(bus, slot, 3,
411 AMD64_GARTAPERTURECTL);
412
413 /*
414 * Before we do anything else disable the GART. It may
415 * still be enabled if we boot into a crash-kernel here.
416 * Reconfiguring the GART while it is enabled could have
417 * unknown side-effects.
418 */
419 ctl &= ~GARTEN;
420 write_pci_config(bus, slot, 3, AMD64_GARTAPERTURECTL, ctl);
421
422 aper_order = (ctl >> 1) & 7;
410 aper_size = (32 * 1024 * 1024) << aper_order; 423 aper_size = (32 * 1024 * 1024) << aper_order;
411 aper_base = read_pci_config(bus, slot, 3, AMD64_GARTAPERTUREBASE) & 0x7fff; 424 aper_base = read_pci_config(bus, slot, 3, AMD64_GARTAPERTUREBASE) & 0x7fff;
412 aper_base <<= 25; 425 aper_base <<= 25;
diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c
index 00187f1fcfb7..e5a4a1e01618 100644
--- a/arch/x86/kernel/apic/apic.c
+++ b/arch/x86/kernel/apic/apic.c
@@ -1640,8 +1640,10 @@ int __init APIC_init_uniprocessor(void)
1640 } 1640 }
1641#endif 1641#endif
1642 1642
1643#ifndef CONFIG_SMP
1643 enable_IR_x2apic(); 1644 enable_IR_x2apic();
1644 default_setup_apic_routing(); 1645 default_setup_apic_routing();
1646#endif
1645 1647
1646 verify_local_APIC(); 1648 verify_local_APIC();
1647 connect_bsp_APIC(); 1649 connect_bsp_APIC();
diff --git a/arch/x86/kernel/apic/es7000_32.c b/arch/x86/kernel/apic/es7000_32.c
index dd2b5f264643..03ba1b895f5e 100644
--- a/arch/x86/kernel/apic/es7000_32.c
+++ b/arch/x86/kernel/apic/es7000_32.c
@@ -42,6 +42,7 @@
42#include <linux/errno.h> 42#include <linux/errno.h>
43#include <linux/acpi.h> 43#include <linux/acpi.h>
44#include <linux/init.h> 44#include <linux/init.h>
45#include <linux/gfp.h>
45#include <linux/nmi.h> 46#include <linux/nmi.h>
46#include <linux/smp.h> 47#include <linux/smp.h>
47#include <linux/io.h> 48#include <linux/io.h>
diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
index e4e0ddcb1546..eb2789c3f721 100644
--- a/arch/x86/kernel/apic/io_apic.c
+++ b/arch/x86/kernel/apic/io_apic.c
@@ -36,6 +36,7 @@
36#include <linux/freezer.h> 36#include <linux/freezer.h>
37#include <linux/kthread.h> 37#include <linux/kthread.h>
38#include <linux/jiffies.h> /* time_after() */ 38#include <linux/jiffies.h> /* time_after() */
39#include <linux/slab.h>
39#ifdef CONFIG_ACPI 40#ifdef CONFIG_ACPI
40#include <acpi/acpi_bus.h> 41#include <acpi/acpi_bus.h>
41#endif 42#endif
@@ -1268,6 +1269,14 @@ void __setup_vector_irq(int cpu)
1268 /* Mark the inuse vectors */ 1269 /* Mark the inuse vectors */
1269 for_each_irq_desc(irq, desc) { 1270 for_each_irq_desc(irq, desc) {
1270 cfg = desc->chip_data; 1271 cfg = desc->chip_data;
1272
1273 /*
1274 * If it is a legacy IRQ handled by the legacy PIC, this cpu
1275 * will be part of the irq_cfg's domain.
1276 */
1277 if (irq < legacy_pic->nr_legacy_irqs && !IO_APIC_IRQ(irq))
1278 cpumask_set_cpu(cpu, cfg->domain);
1279
1271 if (!cpumask_test_cpu(cpu, cfg->domain)) 1280 if (!cpumask_test_cpu(cpu, cfg->domain))
1272 continue; 1281 continue;
1273 vector = cfg->vector; 1282 vector = cfg->vector;
@@ -2536,6 +2545,9 @@ void irq_force_complete_move(int irq)
2536 struct irq_desc *desc = irq_to_desc(irq); 2545 struct irq_desc *desc = irq_to_desc(irq);
2537 struct irq_cfg *cfg = desc->chip_data; 2546 struct irq_cfg *cfg = desc->chip_data;
2538 2547
2548 if (!cfg)
2549 return;
2550
2539 __irq_complete_move(&desc, cfg->vector); 2551 __irq_complete_move(&desc, cfg->vector);
2540} 2552}
2541#else 2553#else
diff --git a/arch/x86/kernel/apic/nmi.c b/arch/x86/kernel/apic/nmi.c
index 8aa65adbd25d..1edaf15c0b8e 100644
--- a/arch/x86/kernel/apic/nmi.c
+++ b/arch/x86/kernel/apic/nmi.c
@@ -18,6 +18,7 @@
18#include <linux/delay.h> 18#include <linux/delay.h>
19#include <linux/interrupt.h> 19#include <linux/interrupt.h>
20#include <linux/module.h> 20#include <linux/module.h>
21#include <linux/slab.h>
21#include <linux/sysdev.h> 22#include <linux/sysdev.h>
22#include <linux/sysctl.h> 23#include <linux/sysctl.h>
23#include <linux/percpu.h> 24#include <linux/percpu.h>
diff --git a/arch/x86/kernel/apic/x2apic_uv_x.c b/arch/x86/kernel/apic/x2apic_uv_x.c
index 49dbeaef2a27..c085d52dbaf2 100644
--- a/arch/x86/kernel/apic/x2apic_uv_x.c
+++ b/arch/x86/kernel/apic/x2apic_uv_x.c
@@ -17,6 +17,7 @@
17#include <linux/ctype.h> 17#include <linux/ctype.h>
18#include <linux/sched.h> 18#include <linux/sched.h>
19#include <linux/timer.h> 19#include <linux/timer.h>
20#include <linux/slab.h>
20#include <linux/cpu.h> 21#include <linux/cpu.h>
21#include <linux/init.h> 22#include <linux/init.h>
22#include <linux/io.h> 23#include <linux/io.h>
diff --git a/arch/x86/kernel/bootflag.c b/arch/x86/kernel/bootflag.c
index 30f25a75fe28..5de7f4c56971 100644
--- a/arch/x86/kernel/bootflag.c
+++ b/arch/x86/kernel/bootflag.c
@@ -5,7 +5,6 @@
5#include <linux/kernel.h> 5#include <linux/kernel.h>
6#include <linux/init.h> 6#include <linux/init.h>
7#include <linux/string.h> 7#include <linux/string.h>
8#include <linux/slab.h>
9#include <linux/spinlock.h> 8#include <linux/spinlock.h>
10#include <linux/acpi.h> 9#include <linux/acpi.h>
11#include <asm/io.h> 10#include <asm/io.h>
diff --git a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
index 1b1920fa7c80..459168083b77 100644
--- a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
+++ b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
@@ -33,6 +33,7 @@
33#include <linux/cpufreq.h> 33#include <linux/cpufreq.h>
34#include <linux/compiler.h> 34#include <linux/compiler.h>
35#include <linux/dmi.h> 35#include <linux/dmi.h>
36#include <linux/slab.h>
36#include <trace/events/power.h> 37#include <trace/events/power.h>
37 38
38#include <linux/acpi.h> 39#include <linux/acpi.h>
diff --git a/arch/x86/kernel/cpu/cpufreq/elanfreq.c b/arch/x86/kernel/cpu/cpufreq/elanfreq.c
index 006b278b0d5d..c587db472a75 100644
--- a/arch/x86/kernel/cpu/cpufreq/elanfreq.c
+++ b/arch/x86/kernel/cpu/cpufreq/elanfreq.c
@@ -20,7 +20,6 @@
20#include <linux/module.h> 20#include <linux/module.h>
21#include <linux/init.h> 21#include <linux/init.h>
22 22
23#include <linux/slab.h>
24#include <linux/delay.h> 23#include <linux/delay.h>
25#include <linux/cpufreq.h> 24#include <linux/cpufreq.h>
26 25
diff --git a/arch/x86/kernel/cpu/cpufreq/gx-suspmod.c b/arch/x86/kernel/cpu/cpufreq/gx-suspmod.c
index ac27ec2264d5..16e3483be9e3 100644
--- a/arch/x86/kernel/cpu/cpufreq/gx-suspmod.c
+++ b/arch/x86/kernel/cpu/cpufreq/gx-suspmod.c
@@ -80,6 +80,7 @@
80#include <linux/cpufreq.h> 80#include <linux/cpufreq.h>
81#include <linux/pci.h> 81#include <linux/pci.h>
82#include <linux/errno.h> 82#include <linux/errno.h>
83#include <linux/slab.h>
83 84
84#include <asm/processor-cyrix.h> 85#include <asm/processor-cyrix.h>
85 86
diff --git a/arch/x86/kernel/cpu/cpufreq/longrun.c b/arch/x86/kernel/cpu/cpufreq/longrun.c
index da5f70fcb766..e7b559d74c52 100644
--- a/arch/x86/kernel/cpu/cpufreq/longrun.c
+++ b/arch/x86/kernel/cpu/cpufreq/longrun.c
@@ -9,7 +9,6 @@
9#include <linux/kernel.h> 9#include <linux/kernel.h>
10#include <linux/module.h> 10#include <linux/module.h>
11#include <linux/init.h> 11#include <linux/init.h>
12#include <linux/slab.h>
13#include <linux/cpufreq.h> 12#include <linux/cpufreq.h>
14#include <linux/timex.h> 13#include <linux/timex.h>
15 14
diff --git a/arch/x86/kernel/cpu/cpufreq/p4-clockmod.c b/arch/x86/kernel/cpu/cpufreq/p4-clockmod.c
index 869615193720..7b8a8ba67b07 100644
--- a/arch/x86/kernel/cpu/cpufreq/p4-clockmod.c
+++ b/arch/x86/kernel/cpu/cpufreq/p4-clockmod.c
@@ -25,7 +25,6 @@
25#include <linux/init.h> 25#include <linux/init.h>
26#include <linux/smp.h> 26#include <linux/smp.h>
27#include <linux/cpufreq.h> 27#include <linux/cpufreq.h>
28#include <linux/slab.h>
29#include <linux/cpumask.h> 28#include <linux/cpumask.h>
30#include <linux/timex.h> 29#include <linux/timex.h>
31 30
diff --git a/arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c b/arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c
index ff36d2979a90..ce7cde713e71 100644
--- a/arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c
+++ b/arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c
@@ -30,6 +30,7 @@
30#include <linux/sched.h> 30#include <linux/sched.h>
31#include <linux/cpufreq.h> 31#include <linux/cpufreq.h>
32#include <linux/compiler.h> 32#include <linux/compiler.h>
33#include <linux/slab.h>
33 34
34#include <linux/acpi.h> 35#include <linux/acpi.h>
35#include <linux/io.h> 36#include <linux/io.h>
diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k6.c b/arch/x86/kernel/cpu/cpufreq/powernow-k6.c
index cb01dac267d3..b3379d6a5c57 100644
--- a/arch/x86/kernel/cpu/cpufreq/powernow-k6.c
+++ b/arch/x86/kernel/cpu/cpufreq/powernow-k6.c
@@ -13,7 +13,6 @@
13#include <linux/init.h> 13#include <linux/init.h>
14#include <linux/cpufreq.h> 14#include <linux/cpufreq.h>
15#include <linux/ioport.h> 15#include <linux/ioport.h>
16#include <linux/slab.h>
17#include <linux/timex.h> 16#include <linux/timex.h>
18#include <linux/io.h> 17#include <linux/io.h>
19 18
diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
index d360b56e9825..b6215b9798e2 100644
--- a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
+++ b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
@@ -929,7 +929,8 @@ static int fill_powernow_table_pstate(struct powernow_k8_data *data,
929 powernow_table[i].index = index; 929 powernow_table[i].index = index;
930 930
931 /* Frequency may be rounded for these */ 931 /* Frequency may be rounded for these */
932 if (boot_cpu_data.x86 == 0x10 || boot_cpu_data.x86 == 0x11) { 932 if ((boot_cpu_data.x86 == 0x10 && boot_cpu_data.x86_model < 10)
933 || boot_cpu_data.x86 == 0x11) {
933 powernow_table[i].frequency = 934 powernow_table[i].frequency =
934 freq_from_fid_did(lo & 0x3f, (lo >> 6) & 7); 935 freq_from_fid_did(lo & 0x3f, (lo >> 6) & 7);
935 } else 936 } else
diff --git a/arch/x86/kernel/cpu/cpufreq/speedstep-centrino.c b/arch/x86/kernel/cpu/cpufreq/speedstep-centrino.c
index 8d672ef162ce..9b1ff37de46a 100644
--- a/arch/x86/kernel/cpu/cpufreq/speedstep-centrino.c
+++ b/arch/x86/kernel/cpu/cpufreq/speedstep-centrino.c
@@ -20,6 +20,7 @@
20#include <linux/sched.h> /* current */ 20#include <linux/sched.h> /* current */
21#include <linux/delay.h> 21#include <linux/delay.h>
22#include <linux/compiler.h> 22#include <linux/compiler.h>
23#include <linux/gfp.h>
23 24
24#include <asm/msr.h> 25#include <asm/msr.h>
25#include <asm/processor.h> 26#include <asm/processor.h>
diff --git a/arch/x86/kernel/cpu/cpufreq/speedstep-ich.c b/arch/x86/kernel/cpu/cpufreq/speedstep-ich.c
index 2ce8e0b5cc54..561758e95180 100644
--- a/arch/x86/kernel/cpu/cpufreq/speedstep-ich.c
+++ b/arch/x86/kernel/cpu/cpufreq/speedstep-ich.c
@@ -23,7 +23,6 @@
23#include <linux/init.h> 23#include <linux/init.h>
24#include <linux/cpufreq.h> 24#include <linux/cpufreq.h>
25#include <linux/pci.h> 25#include <linux/pci.h>
26#include <linux/slab.h>
27#include <linux/sched.h> 26#include <linux/sched.h>
28 27
29#include "speedstep-lib.h" 28#include "speedstep-lib.h"
diff --git a/arch/x86/kernel/cpu/cpufreq/speedstep-lib.c b/arch/x86/kernel/cpu/cpufreq/speedstep-lib.c
index ad0083abfa23..a94ec6be69fa 100644
--- a/arch/x86/kernel/cpu/cpufreq/speedstep-lib.c
+++ b/arch/x86/kernel/cpu/cpufreq/speedstep-lib.c
@@ -13,7 +13,6 @@
13#include <linux/moduleparam.h> 13#include <linux/moduleparam.h>
14#include <linux/init.h> 14#include <linux/init.h>
15#include <linux/cpufreq.h> 15#include <linux/cpufreq.h>
16#include <linux/slab.h>
17 16
18#include <asm/msr.h> 17#include <asm/msr.h>
19#include <asm/tsc.h> 18#include <asm/tsc.h>
diff --git a/arch/x86/kernel/cpu/cpufreq/speedstep-smi.c b/arch/x86/kernel/cpu/cpufreq/speedstep-smi.c
index 04d73c114e49..8abd869baabf 100644
--- a/arch/x86/kernel/cpu/cpufreq/speedstep-smi.c
+++ b/arch/x86/kernel/cpu/cpufreq/speedstep-smi.c
@@ -17,7 +17,6 @@
17#include <linux/moduleparam.h> 17#include <linux/moduleparam.h>
18#include <linux/init.h> 18#include <linux/init.h>
19#include <linux/cpufreq.h> 19#include <linux/cpufreq.h>
20#include <linux/slab.h>
21#include <linux/delay.h> 20#include <linux/delay.h>
22#include <linux/io.h> 21#include <linux/io.h>
23#include <asm/ist.h> 22#include <asm/ist.h>
diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c
index 7e1cca13af35..1366c7cfd483 100644
--- a/arch/x86/kernel/cpu/intel.c
+++ b/arch/x86/kernel/cpu/intel.c
@@ -47,6 +47,27 @@ static void __cpuinit early_init_intel(struct cpuinfo_x86 *c)
47 (c->x86 == 0x6 && c->x86_model >= 0x0e)) 47 (c->x86 == 0x6 && c->x86_model >= 0x0e))
48 set_cpu_cap(c, X86_FEATURE_CONSTANT_TSC); 48 set_cpu_cap(c, X86_FEATURE_CONSTANT_TSC);
49 49
50 /*
51 * Atom erratum AAE44/AAF40/AAG38/AAH41:
52 *
53 * A race condition between speculative fetches and invalidating
54 * a large page. This is worked around in microcode, but we
55 * need the microcode to have already been loaded... so if it is
56 * not, recommend a BIOS update and disable large pages.
57 */
58 if (c->x86 == 6 && c->x86_model == 0x1c && c->x86_mask <= 2) {
59 u32 ucode, junk;
60
61 wrmsr(MSR_IA32_UCODE_REV, 0, 0);
62 sync_core();
63 rdmsr(MSR_IA32_UCODE_REV, junk, ucode);
64
65 if (ucode < 0x20e) {
66 printk(KERN_WARNING "Atom PSE erratum detected, BIOS microcode update recommended\n");
67 clear_cpu_cap(c, X86_FEATURE_PSE);
68 }
69 }
70
50#ifdef CONFIG_X86_64 71#ifdef CONFIG_X86_64
51 set_cpu_cap(c, X86_FEATURE_SYSENTER32); 72 set_cpu_cap(c, X86_FEATURE_SYSENTER32);
52#else 73#else
diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c
index b3eeb66c0a51..95962a93f99a 100644
--- a/arch/x86/kernel/cpu/intel_cacheinfo.c
+++ b/arch/x86/kernel/cpu/intel_cacheinfo.c
@@ -340,6 +340,10 @@ amd_check_l3_disable(int index, struct _cpuid4_info_regs *this_leaf)
340 (boot_cpu_data.x86_mask < 0x1))) 340 (boot_cpu_data.x86_mask < 0x1)))
341 return; 341 return;
342 342
343 /* not in virtualized environments */
344 if (num_k8_northbridges == 0)
345 return;
346
343 this_leaf->can_disable = true; 347 this_leaf->can_disable = true;
344 this_leaf->l3_indices = amd_calc_l3_indices(); 348 this_leaf->l3_indices = amd_calc_l3_indices();
345} 349}
diff --git a/arch/x86/kernel/cpu/mcheck/Makefile b/arch/x86/kernel/cpu/mcheck/Makefile
index 4ac6d48fe11b..bb34b03af252 100644
--- a/arch/x86/kernel/cpu/mcheck/Makefile
+++ b/arch/x86/kernel/cpu/mcheck/Makefile
@@ -7,3 +7,5 @@ obj-$(CONFIG_X86_MCE_THRESHOLD) += threshold.o
7obj-$(CONFIG_X86_MCE_INJECT) += mce-inject.o 7obj-$(CONFIG_X86_MCE_INJECT) += mce-inject.o
8 8
9obj-$(CONFIG_X86_THERMAL_VECTOR) += therm_throt.o 9obj-$(CONFIG_X86_THERMAL_VECTOR) += therm_throt.o
10
11obj-$(CONFIG_ACPI_APEI) += mce-apei.o
diff --git a/arch/x86/kernel/cpu/mcheck/mce-apei.c b/arch/x86/kernel/cpu/mcheck/mce-apei.c
new file mode 100644
index 000000000000..745b54f9be89
--- /dev/null
+++ b/arch/x86/kernel/cpu/mcheck/mce-apei.c
@@ -0,0 +1,138 @@
1/*
2 * Bridge between MCE and APEI
3 *
4 * On some machine, corrected memory errors are reported via APEI
5 * generic hardware error source (GHES) instead of corrected Machine
6 * Check. These corrected memory errors can be reported to user space
7 * through /dev/mcelog via faking a corrected Machine Check, so that
8 * the error memory page can be offlined by /sbin/mcelog if the error
9 * count for one page is beyond the threshold.
10 *
11 * For fatal MCE, save MCE record into persistent storage via ERST, so
12 * that the MCE record can be logged after reboot via ERST.
13 *
14 * Copyright 2010 Intel Corp.
15 * Author: Huang Ying <ying.huang@intel.com>
16 *
17 * This program is free software; you can redistribute it and/or
18 * modify it under the terms of the GNU General Public License version
19 * 2 as published by the Free Software Foundation.
20 *
21 * This program is distributed in the hope that it will be useful,
22 * but WITHOUT ANY WARRANTY; without even the implied warranty of
23 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
24 * GNU General Public License for more details.
25 *
26 * You should have received a copy of the GNU General Public License
27 * along with this program; if not, write to the Free Software
28 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
29 */
30
31#include <linux/kernel.h>
32#include <linux/acpi.h>
33#include <linux/cper.h>
34#include <acpi/apei.h>
35#include <asm/mce.h>
36
37#include "mce-internal.h"
38
39void apei_mce_report_mem_error(int corrected, struct cper_sec_mem_err *mem_err)
40{
41 struct mce m;
42
43 /* Only corrected MC is reported */
44 if (!corrected)
45 return;
46
47 mce_setup(&m);
48 m.bank = 1;
49 /* Fake a memory read corrected error with unknown channel */
50 m.status = MCI_STATUS_VAL | MCI_STATUS_EN | MCI_STATUS_ADDRV | 0x9f;
51 m.addr = mem_err->physical_addr;
52 mce_log(&m);
53 mce_notify_irq();
54}
55EXPORT_SYMBOL_GPL(apei_mce_report_mem_error);
56
57#define CPER_CREATOR_MCE \
58 UUID_LE(0x75a574e3, 0x5052, 0x4b29, 0x8a, 0x8e, 0xbe, 0x2c, \
59 0x64, 0x90, 0xb8, 0x9d)
60#define CPER_SECTION_TYPE_MCE \
61 UUID_LE(0xfe08ffbe, 0x95e4, 0x4be7, 0xbc, 0x73, 0x40, 0x96, \
62 0x04, 0x4a, 0x38, 0xfc)
63
64/*
65 * CPER specification (in UEFI specification 2.3 appendix N) requires
66 * byte-packed.
67 */
68struct cper_mce_record {
69 struct cper_record_header hdr;
70 struct cper_section_descriptor sec_hdr;
71 struct mce mce;
72} __packed;
73
74int apei_write_mce(struct mce *m)
75{
76 struct cper_mce_record rcd;
77
78 memset(&rcd, 0, sizeof(rcd));
79 memcpy(rcd.hdr.signature, CPER_SIG_RECORD, CPER_SIG_SIZE);
80 rcd.hdr.revision = CPER_RECORD_REV;
81 rcd.hdr.signature_end = CPER_SIG_END;
82 rcd.hdr.section_count = 1;
83 rcd.hdr.error_severity = CPER_SER_FATAL;
84 /* timestamp, platform_id, partition_id are all invalid */
85 rcd.hdr.validation_bits = 0;
86 rcd.hdr.record_length = sizeof(rcd);
87 rcd.hdr.creator_id = CPER_CREATOR_MCE;
88 rcd.hdr.notification_type = CPER_NOTIFY_MCE;
89 rcd.hdr.record_id = cper_next_record_id();
90 rcd.hdr.flags = CPER_HW_ERROR_FLAGS_PREVERR;
91
92 rcd.sec_hdr.section_offset = (void *)&rcd.mce - (void *)&rcd;
93 rcd.sec_hdr.section_length = sizeof(rcd.mce);
94 rcd.sec_hdr.revision = CPER_SEC_REV;
95 /* fru_id and fru_text is invalid */
96 rcd.sec_hdr.validation_bits = 0;
97 rcd.sec_hdr.flags = CPER_SEC_PRIMARY;
98 rcd.sec_hdr.section_type = CPER_SECTION_TYPE_MCE;
99 rcd.sec_hdr.section_severity = CPER_SER_FATAL;
100
101 memcpy(&rcd.mce, m, sizeof(*m));
102
103 return erst_write(&rcd.hdr);
104}
105
106ssize_t apei_read_mce(struct mce *m, u64 *record_id)
107{
108 struct cper_mce_record rcd;
109 ssize_t len;
110
111 len = erst_read_next(&rcd.hdr, sizeof(rcd));
112 if (len <= 0)
113 return len;
114 /* Can not skip other records in storage via ERST unless clear them */
115 else if (len != sizeof(rcd) ||
116 uuid_le_cmp(rcd.hdr.creator_id, CPER_CREATOR_MCE)) {
117 if (printk_ratelimit())
118 pr_warning(
119 "MCE-APEI: Can not skip the unknown record in ERST");
120 return -EIO;
121 }
122
123 memcpy(m, &rcd.mce, sizeof(*m));
124 *record_id = rcd.hdr.record_id;
125
126 return sizeof(*m);
127}
128
129/* Check whether there is record in ERST */
130int apei_check_mce(void)
131{
132 return erst_get_record_count();
133}
134
135int apei_clear_mce(u64 record_id)
136{
137 return erst_clear(record_id);
138}
diff --git a/arch/x86/kernel/cpu/mcheck/mce-inject.c b/arch/x86/kernel/cpu/mcheck/mce-inject.c
index 73734baa50f2..e7dbde7bfedb 100644
--- a/arch/x86/kernel/cpu/mcheck/mce-inject.c
+++ b/arch/x86/kernel/cpu/mcheck/mce-inject.c
@@ -22,6 +22,7 @@
22#include <linux/kdebug.h> 22#include <linux/kdebug.h>
23#include <linux/cpu.h> 23#include <linux/cpu.h>
24#include <linux/sched.h> 24#include <linux/sched.h>
25#include <linux/gfp.h>
25#include <asm/mce.h> 26#include <asm/mce.h>
26#include <asm/apic.h> 27#include <asm/apic.h>
27 28
diff --git a/arch/x86/kernel/cpu/mcheck/mce-internal.h b/arch/x86/kernel/cpu/mcheck/mce-internal.h
index 32996f9fab67..fefcc69ee8b5 100644
--- a/arch/x86/kernel/cpu/mcheck/mce-internal.h
+++ b/arch/x86/kernel/cpu/mcheck/mce-internal.h
@@ -28,3 +28,26 @@ extern int mce_ser;
28 28
29extern struct mce_bank *mce_banks; 29extern struct mce_bank *mce_banks;
30 30
31#ifdef CONFIG_ACPI_APEI
32int apei_write_mce(struct mce *m);
33ssize_t apei_read_mce(struct mce *m, u64 *record_id);
34int apei_check_mce(void);
35int apei_clear_mce(u64 record_id);
36#else
37static inline int apei_write_mce(struct mce *m)
38{
39 return -EINVAL;
40}
41static inline ssize_t apei_read_mce(struct mce *m, u64 *record_id)
42{
43 return 0;
44}
45static inline int apei_check_mce(void)
46{
47 return 0;
48}
49static inline int apei_clear_mce(u64 record_id)
50{
51 return -EINVAL;
52}
53#endif
diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c
index 3ab9c886b613..09535ca9b9d7 100644
--- a/arch/x86/kernel/cpu/mcheck/mce.c
+++ b/arch/x86/kernel/cpu/mcheck/mce.c
@@ -26,6 +26,7 @@
26#include <linux/sched.h> 26#include <linux/sched.h>
27#include <linux/sysfs.h> 27#include <linux/sysfs.h>
28#include <linux/types.h> 28#include <linux/types.h>
29#include <linux/slab.h>
29#include <linux/init.h> 30#include <linux/init.h>
30#include <linux/kmod.h> 31#include <linux/kmod.h>
31#include <linux/poll.h> 32#include <linux/poll.h>
@@ -263,7 +264,7 @@ static void wait_for_panic(void)
263 264
264static void mce_panic(char *msg, struct mce *final, char *exp) 265static void mce_panic(char *msg, struct mce *final, char *exp)
265{ 266{
266 int i; 267 int i, apei_err = 0;
267 268
268 if (!fake_panic) { 269 if (!fake_panic) {
269 /* 270 /*
@@ -286,8 +287,11 @@ static void mce_panic(char *msg, struct mce *final, char *exp)
286 struct mce *m = &mcelog.entry[i]; 287 struct mce *m = &mcelog.entry[i];
287 if (!(m->status & MCI_STATUS_VAL)) 288 if (!(m->status & MCI_STATUS_VAL))
288 continue; 289 continue;
289 if (!(m->status & MCI_STATUS_UC)) 290 if (!(m->status & MCI_STATUS_UC)) {
290 print_mce(m); 291 print_mce(m);
292 if (!apei_err)
293 apei_err = apei_write_mce(m);
294 }
291 } 295 }
292 /* Now print uncorrected but with the final one last */ 296 /* Now print uncorrected but with the final one last */
293 for (i = 0; i < MCE_LOG_LEN; i++) { 297 for (i = 0; i < MCE_LOG_LEN; i++) {
@@ -296,11 +300,17 @@ static void mce_panic(char *msg, struct mce *final, char *exp)
296 continue; 300 continue;
297 if (!(m->status & MCI_STATUS_UC)) 301 if (!(m->status & MCI_STATUS_UC))
298 continue; 302 continue;
299 if (!final || memcmp(m, final, sizeof(struct mce))) 303 if (!final || memcmp(m, final, sizeof(struct mce))) {
300 print_mce(m); 304 print_mce(m);
305 if (!apei_err)
306 apei_err = apei_write_mce(m);
307 }
301 } 308 }
302 if (final) 309 if (final) {
303 print_mce(final); 310 print_mce(final);
311 if (!apei_err)
312 apei_err = apei_write_mce(final);
313 }
304 if (cpu_missing) 314 if (cpu_missing)
305 printk(KERN_EMERG "Some CPUs didn't answer in synchronization\n"); 315 printk(KERN_EMERG "Some CPUs didn't answer in synchronization\n");
306 print_mce_tail(); 316 print_mce_tail();
@@ -1492,6 +1502,43 @@ static void collect_tscs(void *data)
1492 rdtscll(cpu_tsc[smp_processor_id()]); 1502 rdtscll(cpu_tsc[smp_processor_id()]);
1493} 1503}
1494 1504
1505static int mce_apei_read_done;
1506
1507/* Collect MCE record of previous boot in persistent storage via APEI ERST. */
1508static int __mce_read_apei(char __user **ubuf, size_t usize)
1509{
1510 int rc;
1511 u64 record_id;
1512 struct mce m;
1513
1514 if (usize < sizeof(struct mce))
1515 return -EINVAL;
1516
1517 rc = apei_read_mce(&m, &record_id);
1518 /* Error or no more MCE record */
1519 if (rc <= 0) {
1520 mce_apei_read_done = 1;
1521 return rc;
1522 }
1523 rc = -EFAULT;
1524 if (copy_to_user(*ubuf, &m, sizeof(struct mce)))
1525 return rc;
1526 /*
1527 * In fact, we should have cleared the record after that has
1528 * been flushed to the disk or sent to network in
1529 * /sbin/mcelog, but we have no interface to support that now,
1530 * so just clear it to avoid duplication.
1531 */
1532 rc = apei_clear_mce(record_id);
1533 if (rc) {
1534 mce_apei_read_done = 1;
1535 return rc;
1536 }
1537 *ubuf += sizeof(struct mce);
1538
1539 return 0;
1540}
1541
1495static ssize_t mce_read(struct file *filp, char __user *ubuf, size_t usize, 1542static ssize_t mce_read(struct file *filp, char __user *ubuf, size_t usize,
1496 loff_t *off) 1543 loff_t *off)
1497{ 1544{
@@ -1505,15 +1552,19 @@ static ssize_t mce_read(struct file *filp, char __user *ubuf, size_t usize,
1505 return -ENOMEM; 1552 return -ENOMEM;
1506 1553
1507 mutex_lock(&mce_read_mutex); 1554 mutex_lock(&mce_read_mutex);
1555
1556 if (!mce_apei_read_done) {
1557 err = __mce_read_apei(&buf, usize);
1558 if (err || buf != ubuf)
1559 goto out;
1560 }
1561
1508 next = rcu_dereference_check_mce(mcelog.next); 1562 next = rcu_dereference_check_mce(mcelog.next);
1509 1563
1510 /* Only supports full reads right now */ 1564 /* Only supports full reads right now */
1511 if (*off != 0 || usize < MCE_LOG_LEN*sizeof(struct mce)) { 1565 err = -EINVAL;
1512 mutex_unlock(&mce_read_mutex); 1566 if (*off != 0 || usize < MCE_LOG_LEN*sizeof(struct mce))
1513 kfree(cpu_tsc); 1567 goto out;
1514
1515 return -EINVAL;
1516 }
1517 1568
1518 err = 0; 1569 err = 0;
1519 prev = 0; 1570 prev = 0;
@@ -1561,10 +1612,15 @@ timeout:
1561 memset(&mcelog.entry[i], 0, sizeof(struct mce)); 1612 memset(&mcelog.entry[i], 0, sizeof(struct mce));
1562 } 1613 }
1563 } 1614 }
1615
1616 if (err)
1617 err = -EFAULT;
1618
1619out:
1564 mutex_unlock(&mce_read_mutex); 1620 mutex_unlock(&mce_read_mutex);
1565 kfree(cpu_tsc); 1621 kfree(cpu_tsc);
1566 1622
1567 return err ? -EFAULT : buf - ubuf; 1623 return err ? err : buf - ubuf;
1568} 1624}
1569 1625
1570static unsigned int mce_poll(struct file *file, poll_table *wait) 1626static unsigned int mce_poll(struct file *file, poll_table *wait)
@@ -1572,6 +1628,8 @@ static unsigned int mce_poll(struct file *file, poll_table *wait)
1572 poll_wait(file, &mce_wait, wait); 1628 poll_wait(file, &mce_wait, wait);
1573 if (rcu_dereference_check_mce(mcelog.next)) 1629 if (rcu_dereference_check_mce(mcelog.next))
1574 return POLLIN | POLLRDNORM; 1630 return POLLIN | POLLRDNORM;
1631 if (!mce_apei_read_done && apei_check_mce())
1632 return POLLIN | POLLRDNORM;
1575 return 0; 1633 return 0;
1576} 1634}
1577 1635
diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd.c b/arch/x86/kernel/cpu/mcheck/mce_amd.c
index cda932ca3ade..224392d8fe8c 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_amd.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_amd.c
@@ -21,6 +21,7 @@
21#include <linux/errno.h> 21#include <linux/errno.h>
22#include <linux/sched.h> 22#include <linux/sched.h>
23#include <linux/sysfs.h> 23#include <linux/sysfs.h>
24#include <linux/slab.h>
24#include <linux/init.h> 25#include <linux/init.h>
25#include <linux/cpu.h> 26#include <linux/cpu.h>
26#include <linux/smp.h> 27#include <linux/smp.h>
diff --git a/arch/x86/kernel/cpu/mcheck/mce_intel.c b/arch/x86/kernel/cpu/mcheck/mce_intel.c
index d15df6e49bf0..62b48e40920a 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_intel.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_intel.c
@@ -5,6 +5,7 @@
5 * Author: Andi Kleen 5 * Author: Andi Kleen
6 */ 6 */
7 7
8#include <linux/gfp.h>
8#include <linux/init.h> 9#include <linux/init.h>
9#include <linux/interrupt.h> 10#include <linux/interrupt.h>
10#include <linux/percpu.h> 11#include <linux/percpu.h>
diff --git a/arch/x86/kernel/cpu/mtrr/generic.c b/arch/x86/kernel/cpu/mtrr/generic.c
index 9aa5dc76ff4a..fd31a441c61c 100644
--- a/arch/x86/kernel/cpu/mtrr/generic.c
+++ b/arch/x86/kernel/cpu/mtrr/generic.c
@@ -6,7 +6,6 @@
6 6
7#include <linux/module.h> 7#include <linux/module.h>
8#include <linux/init.h> 8#include <linux/init.h>
9#include <linux/slab.h>
10#include <linux/io.h> 9#include <linux/io.h>
11#include <linux/mm.h> 10#include <linux/mm.h>
12 11
diff --git a/arch/x86/kernel/cpu/mtrr/if.c b/arch/x86/kernel/cpu/mtrr/if.c
index e006e56f699c..79289632cb27 100644
--- a/arch/x86/kernel/cpu/mtrr/if.c
+++ b/arch/x86/kernel/cpu/mtrr/if.c
@@ -5,6 +5,7 @@
5#include <linux/module.h> 5#include <linux/module.h>
6#include <linux/ctype.h> 6#include <linux/ctype.h>
7#include <linux/string.h> 7#include <linux/string.h>
8#include <linux/slab.h>
8#include <linux/init.h> 9#include <linux/init.h>
9 10
10#define LINE_SIZE 80 11#define LINE_SIZE 80
diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
index 42aafd11e170..db5bdc8addf8 100644
--- a/arch/x86/kernel/cpu/perf_event.c
+++ b/arch/x86/kernel/cpu/perf_event.c
@@ -21,6 +21,7 @@
21#include <linux/kdebug.h> 21#include <linux/kdebug.h>
22#include <linux/sched.h> 22#include <linux/sched.h>
23#include <linux/uaccess.h> 23#include <linux/uaccess.h>
24#include <linux/slab.h>
24#include <linux/highmem.h> 25#include <linux/highmem.h>
25#include <linux/cpu.h> 26#include <linux/cpu.h>
26#include <linux/bitops.h> 27#include <linux/bitops.h>
@@ -28,6 +29,7 @@
28#include <asm/apic.h> 29#include <asm/apic.h>
29#include <asm/stacktrace.h> 30#include <asm/stacktrace.h>
30#include <asm/nmi.h> 31#include <asm/nmi.h>
32#include <asm/compat.h>
31 33
32static u64 perf_event_mask __read_mostly; 34static u64 perf_event_mask __read_mostly;
33 35
@@ -133,8 +135,8 @@ struct x86_pmu {
133 int (*handle_irq)(struct pt_regs *); 135 int (*handle_irq)(struct pt_regs *);
134 void (*disable_all)(void); 136 void (*disable_all)(void);
135 void (*enable_all)(void); 137 void (*enable_all)(void);
136 void (*enable)(struct hw_perf_event *, int); 138 void (*enable)(struct perf_event *);
137 void (*disable)(struct hw_perf_event *, int); 139 void (*disable)(struct perf_event *);
138 unsigned eventsel; 140 unsigned eventsel;
139 unsigned perfctr; 141 unsigned perfctr;
140 u64 (*event_map)(int); 142 u64 (*event_map)(int);
@@ -157,6 +159,11 @@ struct x86_pmu {
157 void (*put_event_constraints)(struct cpu_hw_events *cpuc, 159 void (*put_event_constraints)(struct cpu_hw_events *cpuc,
158 struct perf_event *event); 160 struct perf_event *event);
159 struct event_constraint *event_constraints; 161 struct event_constraint *event_constraints;
162
163 int (*cpu_prepare)(int cpu);
164 void (*cpu_starting)(int cpu);
165 void (*cpu_dying)(int cpu);
166 void (*cpu_dead)(int cpu);
160}; 167};
161 168
162static struct x86_pmu x86_pmu __read_mostly; 169static struct x86_pmu x86_pmu __read_mostly;
@@ -165,8 +172,7 @@ static DEFINE_PER_CPU(struct cpu_hw_events, cpu_hw_events) = {
165 .enabled = 1, 172 .enabled = 1,
166}; 173};
167 174
168static int x86_perf_event_set_period(struct perf_event *event, 175static int x86_perf_event_set_period(struct perf_event *event);
169 struct hw_perf_event *hwc, int idx);
170 176
171/* 177/*
172 * Generalized hw caching related hw_event table, filled 178 * Generalized hw caching related hw_event table, filled
@@ -189,11 +195,12 @@ static u64 __read_mostly hw_cache_event_ids
189 * Returns the delta events processed. 195 * Returns the delta events processed.
190 */ 196 */
191static u64 197static u64
192x86_perf_event_update(struct perf_event *event, 198x86_perf_event_update(struct perf_event *event)
193 struct hw_perf_event *hwc, int idx)
194{ 199{
200 struct hw_perf_event *hwc = &event->hw;
195 int shift = 64 - x86_pmu.event_bits; 201 int shift = 64 - x86_pmu.event_bits;
196 u64 prev_raw_count, new_raw_count; 202 u64 prev_raw_count, new_raw_count;
203 int idx = hwc->idx;
197 s64 delta; 204 s64 delta;
198 205
199 if (idx == X86_PMC_IDX_FIXED_BTS) 206 if (idx == X86_PMC_IDX_FIXED_BTS)
@@ -293,7 +300,7 @@ static inline bool bts_available(void)
293 return x86_pmu.enable_bts != NULL; 300 return x86_pmu.enable_bts != NULL;
294} 301}
295 302
296static inline void init_debug_store_on_cpu(int cpu) 303static void init_debug_store_on_cpu(int cpu)
297{ 304{
298 struct debug_store *ds = per_cpu(cpu_hw_events, cpu).ds; 305 struct debug_store *ds = per_cpu(cpu_hw_events, cpu).ds;
299 306
@@ -305,7 +312,7 @@ static inline void init_debug_store_on_cpu(int cpu)
305 (u32)((u64)(unsigned long)ds >> 32)); 312 (u32)((u64)(unsigned long)ds >> 32));
306} 313}
307 314
308static inline void fini_debug_store_on_cpu(int cpu) 315static void fini_debug_store_on_cpu(int cpu)
309{ 316{
310 if (!per_cpu(cpu_hw_events, cpu).ds) 317 if (!per_cpu(cpu_hw_events, cpu).ds)
311 return; 318 return;
@@ -638,7 +645,7 @@ static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign)
638 if (test_bit(hwc->idx, used_mask)) 645 if (test_bit(hwc->idx, used_mask))
639 break; 646 break;
640 647
641 set_bit(hwc->idx, used_mask); 648 __set_bit(hwc->idx, used_mask);
642 if (assign) 649 if (assign)
643 assign[i] = hwc->idx; 650 assign[i] = hwc->idx;
644 } 651 }
@@ -687,7 +694,7 @@ static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign)
687 if (j == X86_PMC_IDX_MAX) 694 if (j == X86_PMC_IDX_MAX)
688 break; 695 break;
689 696
690 set_bit(j, used_mask); 697 __set_bit(j, used_mask);
691 698
692 if (assign) 699 if (assign)
693 assign[i] = j; 700 assign[i] = j;
@@ -780,6 +787,7 @@ static inline int match_prev_assignment(struct hw_perf_event *hwc,
780 hwc->last_tag == cpuc->tags[i]; 787 hwc->last_tag == cpuc->tags[i];
781} 788}
782 789
790static int x86_pmu_start(struct perf_event *event);
783static void x86_pmu_stop(struct perf_event *event); 791static void x86_pmu_stop(struct perf_event *event);
784 792
785void hw_perf_enable(void) 793void hw_perf_enable(void)
@@ -796,6 +804,7 @@ void hw_perf_enable(void)
796 return; 804 return;
797 805
798 if (cpuc->n_added) { 806 if (cpuc->n_added) {
807 int n_running = cpuc->n_events - cpuc->n_added;
799 /* 808 /*
800 * apply assignment obtained either from 809 * apply assignment obtained either from
801 * hw_perf_group_sched_in() or x86_pmu_enable() 810 * hw_perf_group_sched_in() or x86_pmu_enable()
@@ -803,8 +812,7 @@ void hw_perf_enable(void)
803 * step1: save events moving to new counters 812 * step1: save events moving to new counters
804 * step2: reprogram moved events into new counters 813 * step2: reprogram moved events into new counters
805 */ 814 */
806 for (i = 0; i < cpuc->n_events; i++) { 815 for (i = 0; i < n_running; i++) {
807
808 event = cpuc->event_list[i]; 816 event = cpuc->event_list[i];
809 hwc = &event->hw; 817 hwc = &event->hw;
810 818
@@ -819,29 +827,18 @@ void hw_perf_enable(void)
819 continue; 827 continue;
820 828
821 x86_pmu_stop(event); 829 x86_pmu_stop(event);
822
823 hwc->idx = -1;
824 } 830 }
825 831
826 for (i = 0; i < cpuc->n_events; i++) { 832 for (i = 0; i < cpuc->n_events; i++) {
827
828 event = cpuc->event_list[i]; 833 event = cpuc->event_list[i];
829 hwc = &event->hw; 834 hwc = &event->hw;
830 835
831 if (hwc->idx == -1) { 836 if (!match_prev_assignment(hwc, cpuc, i))
832 x86_assign_hw_event(event, cpuc, i); 837 x86_assign_hw_event(event, cpuc, i);
833 x86_perf_event_set_period(event, hwc, hwc->idx); 838 else if (i < n_running)
834 } 839 continue;
835 /*
836 * need to mark as active because x86_pmu_disable()
837 * clear active_mask and events[] yet it preserves
838 * idx
839 */
840 set_bit(hwc->idx, cpuc->active_mask);
841 cpuc->events[hwc->idx] = event;
842 840
843 x86_pmu.enable(hwc, hwc->idx); 841 x86_pmu_start(event);
844 perf_event_update_userpage(event);
845 } 842 }
846 cpuc->n_added = 0; 843 cpuc->n_added = 0;
847 perf_events_lapic_init(); 844 perf_events_lapic_init();
@@ -853,15 +850,16 @@ void hw_perf_enable(void)
853 x86_pmu.enable_all(); 850 x86_pmu.enable_all();
854} 851}
855 852
856static inline void __x86_pmu_enable_event(struct hw_perf_event *hwc, int idx) 853static inline void __x86_pmu_enable_event(struct hw_perf_event *hwc)
857{ 854{
858 (void)checking_wrmsrl(hwc->config_base + idx, 855 (void)checking_wrmsrl(hwc->config_base + hwc->idx,
859 hwc->config | ARCH_PERFMON_EVENTSEL_ENABLE); 856 hwc->config | ARCH_PERFMON_EVENTSEL_ENABLE);
860} 857}
861 858
862static inline void x86_pmu_disable_event(struct hw_perf_event *hwc, int idx) 859static inline void x86_pmu_disable_event(struct perf_event *event)
863{ 860{
864 (void)checking_wrmsrl(hwc->config_base + idx, hwc->config); 861 struct hw_perf_event *hwc = &event->hw;
862 (void)checking_wrmsrl(hwc->config_base + hwc->idx, hwc->config);
865} 863}
866 864
867static DEFINE_PER_CPU(u64 [X86_PMC_IDX_MAX], pmc_prev_left); 865static DEFINE_PER_CPU(u64 [X86_PMC_IDX_MAX], pmc_prev_left);
@@ -871,12 +869,12 @@ static DEFINE_PER_CPU(u64 [X86_PMC_IDX_MAX], pmc_prev_left);
871 * To be called with the event disabled in hw: 869 * To be called with the event disabled in hw:
872 */ 870 */
873static int 871static int
874x86_perf_event_set_period(struct perf_event *event, 872x86_perf_event_set_period(struct perf_event *event)
875 struct hw_perf_event *hwc, int idx)
876{ 873{
874 struct hw_perf_event *hwc = &event->hw;
877 s64 left = atomic64_read(&hwc->period_left); 875 s64 left = atomic64_read(&hwc->period_left);
878 s64 period = hwc->sample_period; 876 s64 period = hwc->sample_period;
879 int err, ret = 0; 877 int err, ret = 0, idx = hwc->idx;
880 878
881 if (idx == X86_PMC_IDX_FIXED_BTS) 879 if (idx == X86_PMC_IDX_FIXED_BTS)
882 return 0; 880 return 0;
@@ -922,11 +920,11 @@ x86_perf_event_set_period(struct perf_event *event,
922 return ret; 920 return ret;
923} 921}
924 922
925static void x86_pmu_enable_event(struct hw_perf_event *hwc, int idx) 923static void x86_pmu_enable_event(struct perf_event *event)
926{ 924{
927 struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); 925 struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
928 if (cpuc->enabled) 926 if (cpuc->enabled)
929 __x86_pmu_enable_event(hwc, idx); 927 __x86_pmu_enable_event(&event->hw);
930} 928}
931 929
932/* 930/*
@@ -962,34 +960,32 @@ static int x86_pmu_enable(struct perf_event *event)
962 memcpy(cpuc->assign, assign, n*sizeof(int)); 960 memcpy(cpuc->assign, assign, n*sizeof(int));
963 961
964 cpuc->n_events = n; 962 cpuc->n_events = n;
965 cpuc->n_added = n - n0; 963 cpuc->n_added += n - n0;
966 964
967 return 0; 965 return 0;
968} 966}
969 967
970static int x86_pmu_start(struct perf_event *event) 968static int x86_pmu_start(struct perf_event *event)
971{ 969{
972 struct hw_perf_event *hwc = &event->hw; 970 struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
971 int idx = event->hw.idx;
973 972
974 if (hwc->idx == -1) 973 if (idx == -1)
975 return -EAGAIN; 974 return -EAGAIN;
976 975
977 x86_perf_event_set_period(event, hwc, hwc->idx); 976 x86_perf_event_set_period(event);
978 x86_pmu.enable(hwc, hwc->idx); 977 cpuc->events[idx] = event;
978 __set_bit(idx, cpuc->active_mask);
979 x86_pmu.enable(event);
980 perf_event_update_userpage(event);
979 981
980 return 0; 982 return 0;
981} 983}
982 984
983static void x86_pmu_unthrottle(struct perf_event *event) 985static void x86_pmu_unthrottle(struct perf_event *event)
984{ 986{
985 struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); 987 int ret = x86_pmu_start(event);
986 struct hw_perf_event *hwc = &event->hw; 988 WARN_ON_ONCE(ret);
987
988 if (WARN_ON_ONCE(hwc->idx >= X86_PMC_IDX_MAX ||
989 cpuc->events[hwc->idx] != event))
990 return;
991
992 x86_pmu.enable(hwc, hwc->idx);
993} 989}
994 990
995void perf_event_print_debug(void) 991void perf_event_print_debug(void)
@@ -1049,18 +1045,16 @@ static void x86_pmu_stop(struct perf_event *event)
1049 struct hw_perf_event *hwc = &event->hw; 1045 struct hw_perf_event *hwc = &event->hw;
1050 int idx = hwc->idx; 1046 int idx = hwc->idx;
1051 1047
1052 /* 1048 if (!__test_and_clear_bit(idx, cpuc->active_mask))
1053 * Must be done before we disable, otherwise the nmi handler 1049 return;
1054 * could reenable again: 1050
1055 */ 1051 x86_pmu.disable(event);
1056 clear_bit(idx, cpuc->active_mask);
1057 x86_pmu.disable(hwc, idx);
1058 1052
1059 /* 1053 /*
1060 * Drain the remaining delta count out of a event 1054 * Drain the remaining delta count out of a event
1061 * that we are disabling: 1055 * that we are disabling:
1062 */ 1056 */
1063 x86_perf_event_update(event, hwc, idx); 1057 x86_perf_event_update(event);
1064 1058
1065 cpuc->events[idx] = NULL; 1059 cpuc->events[idx] = NULL;
1066} 1060}
@@ -1108,7 +1102,7 @@ static int x86_pmu_handle_irq(struct pt_regs *regs)
1108 event = cpuc->events[idx]; 1102 event = cpuc->events[idx];
1109 hwc = &event->hw; 1103 hwc = &event->hw;
1110 1104
1111 val = x86_perf_event_update(event, hwc, idx); 1105 val = x86_perf_event_update(event);
1112 if (val & (1ULL << (x86_pmu.event_bits - 1))) 1106 if (val & (1ULL << (x86_pmu.event_bits - 1)))
1113 continue; 1107 continue;
1114 1108
@@ -1118,11 +1112,11 @@ static int x86_pmu_handle_irq(struct pt_regs *regs)
1118 handled = 1; 1112 handled = 1;
1119 data.period = event->hw.last_period; 1113 data.period = event->hw.last_period;
1120 1114
1121 if (!x86_perf_event_set_period(event, hwc, idx)) 1115 if (!x86_perf_event_set_period(event))
1122 continue; 1116 continue;
1123 1117
1124 if (perf_event_overflow(event, 1, &data, regs)) 1118 if (perf_event_overflow(event, 1, &data, regs))
1125 x86_pmu.disable(hwc, idx); 1119 x86_pmu_stop(event);
1126 } 1120 }
1127 1121
1128 if (handled) 1122 if (handled)
@@ -1309,7 +1303,7 @@ int hw_perf_group_sched_in(struct perf_event *leader,
1309 memcpy(cpuc->assign, assign, n0*sizeof(int)); 1303 memcpy(cpuc->assign, assign, n0*sizeof(int));
1310 1304
1311 cpuc->n_events = n0; 1305 cpuc->n_events = n0;
1312 cpuc->n_added = n1; 1306 cpuc->n_added += n1;
1313 ctx->nr_active += n1; 1307 ctx->nr_active += n1;
1314 1308
1315 /* 1309 /*
@@ -1337,6 +1331,41 @@ undo:
1337#include "perf_event_p6.c" 1331#include "perf_event_p6.c"
1338#include "perf_event_intel.c" 1332#include "perf_event_intel.c"
1339 1333
1334static int __cpuinit
1335x86_pmu_notifier(struct notifier_block *self, unsigned long action, void *hcpu)
1336{
1337 unsigned int cpu = (long)hcpu;
1338 int ret = NOTIFY_OK;
1339
1340 switch (action & ~CPU_TASKS_FROZEN) {
1341 case CPU_UP_PREPARE:
1342 if (x86_pmu.cpu_prepare)
1343 ret = x86_pmu.cpu_prepare(cpu);
1344 break;
1345
1346 case CPU_STARTING:
1347 if (x86_pmu.cpu_starting)
1348 x86_pmu.cpu_starting(cpu);
1349 break;
1350
1351 case CPU_DYING:
1352 if (x86_pmu.cpu_dying)
1353 x86_pmu.cpu_dying(cpu);
1354 break;
1355
1356 case CPU_UP_CANCELED:
1357 case CPU_DEAD:
1358 if (x86_pmu.cpu_dead)
1359 x86_pmu.cpu_dead(cpu);
1360 break;
1361
1362 default:
1363 break;
1364 }
1365
1366 return ret;
1367}
1368
1340static void __init pmu_check_apic(void) 1369static void __init pmu_check_apic(void)
1341{ 1370{
1342 if (cpu_has_apic) 1371 if (cpu_has_apic)
@@ -1415,11 +1444,13 @@ void __init init_hw_perf_events(void)
1415 pr_info("... max period: %016Lx\n", x86_pmu.max_period); 1444 pr_info("... max period: %016Lx\n", x86_pmu.max_period);
1416 pr_info("... fixed-purpose events: %d\n", x86_pmu.num_events_fixed); 1445 pr_info("... fixed-purpose events: %d\n", x86_pmu.num_events_fixed);
1417 pr_info("... event mask: %016Lx\n", perf_event_mask); 1446 pr_info("... event mask: %016Lx\n", perf_event_mask);
1447
1448 perf_cpu_notifier(x86_pmu_notifier);
1418} 1449}
1419 1450
1420static inline void x86_pmu_read(struct perf_event *event) 1451static inline void x86_pmu_read(struct perf_event *event)
1421{ 1452{
1422 x86_perf_event_update(event, &event->hw, event->hw.idx); 1453 x86_perf_event_update(event);
1423} 1454}
1424 1455
1425static const struct pmu pmu = { 1456static const struct pmu pmu = {
@@ -1601,14 +1632,42 @@ copy_from_user_nmi(void *to, const void __user *from, unsigned long n)
1601 return len; 1632 return len;
1602} 1633}
1603 1634
1604static int copy_stack_frame(const void __user *fp, struct stack_frame *frame) 1635#ifdef CONFIG_COMPAT
1636static inline int
1637perf_callchain_user32(struct pt_regs *regs, struct perf_callchain_entry *entry)
1605{ 1638{
1606 unsigned long bytes; 1639 /* 32-bit process in 64-bit kernel. */
1640 struct stack_frame_ia32 frame;
1641 const void __user *fp;
1607 1642
1608 bytes = copy_from_user_nmi(frame, fp, sizeof(*frame)); 1643 if (!test_thread_flag(TIF_IA32))
1644 return 0;
1645
1646 fp = compat_ptr(regs->bp);
1647 while (entry->nr < PERF_MAX_STACK_DEPTH) {
1648 unsigned long bytes;
1649 frame.next_frame = 0;
1650 frame.return_address = 0;
1651
1652 bytes = copy_from_user_nmi(&frame, fp, sizeof(frame));
1653 if (bytes != sizeof(frame))
1654 break;
1609 1655
1610 return bytes == sizeof(*frame); 1656 if (fp < compat_ptr(regs->sp))
1657 break;
1658
1659 callchain_store(entry, frame.return_address);
1660 fp = compat_ptr(frame.next_frame);
1661 }
1662 return 1;
1611} 1663}
1664#else
1665static inline int
1666perf_callchain_user32(struct pt_regs *regs, struct perf_callchain_entry *entry)
1667{
1668 return 0;
1669}
1670#endif
1612 1671
1613static void 1672static void
1614perf_callchain_user(struct pt_regs *regs, struct perf_callchain_entry *entry) 1673perf_callchain_user(struct pt_regs *regs, struct perf_callchain_entry *entry)
@@ -1624,11 +1683,16 @@ perf_callchain_user(struct pt_regs *regs, struct perf_callchain_entry *entry)
1624 callchain_store(entry, PERF_CONTEXT_USER); 1683 callchain_store(entry, PERF_CONTEXT_USER);
1625 callchain_store(entry, regs->ip); 1684 callchain_store(entry, regs->ip);
1626 1685
1686 if (perf_callchain_user32(regs, entry))
1687 return;
1688
1627 while (entry->nr < PERF_MAX_STACK_DEPTH) { 1689 while (entry->nr < PERF_MAX_STACK_DEPTH) {
1690 unsigned long bytes;
1628 frame.next_frame = NULL; 1691 frame.next_frame = NULL;
1629 frame.return_address = 0; 1692 frame.return_address = 0;
1630 1693
1631 if (!copy_stack_frame(fp, &frame)) 1694 bytes = copy_from_user_nmi(&frame, fp, sizeof(frame));
1695 if (bytes != sizeof(frame))
1632 break; 1696 break;
1633 1697
1634 if ((unsigned long)fp < regs->sp) 1698 if ((unsigned long)fp < regs->sp)
@@ -1675,28 +1739,14 @@ struct perf_callchain_entry *perf_callchain(struct pt_regs *regs)
1675 return entry; 1739 return entry;
1676} 1740}
1677 1741
1678void hw_perf_event_setup_online(int cpu) 1742void perf_arch_fetch_caller_regs(struct pt_regs *regs, unsigned long ip, int skip)
1679{
1680 init_debug_store_on_cpu(cpu);
1681
1682 switch (boot_cpu_data.x86_vendor) {
1683 case X86_VENDOR_AMD:
1684 amd_pmu_cpu_online(cpu);
1685 break;
1686 default:
1687 return;
1688 }
1689}
1690
1691void hw_perf_event_setup_offline(int cpu)
1692{ 1743{
1693 init_debug_store_on_cpu(cpu); 1744 regs->ip = ip;
1694 1745 /*
1695 switch (boot_cpu_data.x86_vendor) { 1746 * perf_arch_fetch_caller_regs adds another call, we need to increment
1696 case X86_VENDOR_AMD: 1747 * the skip level
1697 amd_pmu_cpu_offline(cpu); 1748 */
1698 break; 1749 regs->bp = rewind_frame_pointer(skip + 1);
1699 default: 1750 regs->cs = __KERNEL_CS;
1700 return; 1751 local_save_flags(regs->flags);
1701 }
1702} 1752}
diff --git a/arch/x86/kernel/cpu/perf_event_amd.c b/arch/x86/kernel/cpu/perf_event_amd.c
index 8f3dbfda3c4f..db6f7d4056e1 100644
--- a/arch/x86/kernel/cpu/perf_event_amd.c
+++ b/arch/x86/kernel/cpu/perf_event_amd.c
@@ -137,6 +137,13 @@ static inline int amd_is_nb_event(struct hw_perf_event *hwc)
137 return (hwc->config & 0xe0) == 0xe0; 137 return (hwc->config & 0xe0) == 0xe0;
138} 138}
139 139
140static inline int amd_has_nb(struct cpu_hw_events *cpuc)
141{
142 struct amd_nb *nb = cpuc->amd_nb;
143
144 return nb && nb->nb_id != -1;
145}
146
140static void amd_put_event_constraints(struct cpu_hw_events *cpuc, 147static void amd_put_event_constraints(struct cpu_hw_events *cpuc,
141 struct perf_event *event) 148 struct perf_event *event)
142{ 149{
@@ -147,7 +154,7 @@ static void amd_put_event_constraints(struct cpu_hw_events *cpuc,
147 /* 154 /*
148 * only care about NB events 155 * only care about NB events
149 */ 156 */
150 if (!(nb && amd_is_nb_event(hwc))) 157 if (!(amd_has_nb(cpuc) && amd_is_nb_event(hwc)))
151 return; 158 return;
152 159
153 /* 160 /*
@@ -214,7 +221,7 @@ amd_get_event_constraints(struct cpu_hw_events *cpuc, struct perf_event *event)
214 /* 221 /*
215 * if not NB event or no NB, then no constraints 222 * if not NB event or no NB, then no constraints
216 */ 223 */
217 if (!(nb && amd_is_nb_event(hwc))) 224 if (!(amd_has_nb(cpuc) && amd_is_nb_event(hwc)))
218 return &unconstrained; 225 return &unconstrained;
219 226
220 /* 227 /*
@@ -271,28 +278,6 @@ done:
271 return &emptyconstraint; 278 return &emptyconstraint;
272} 279}
273 280
274static __initconst struct x86_pmu amd_pmu = {
275 .name = "AMD",
276 .handle_irq = x86_pmu_handle_irq,
277 .disable_all = x86_pmu_disable_all,
278 .enable_all = x86_pmu_enable_all,
279 .enable = x86_pmu_enable_event,
280 .disable = x86_pmu_disable_event,
281 .eventsel = MSR_K7_EVNTSEL0,
282 .perfctr = MSR_K7_PERFCTR0,
283 .event_map = amd_pmu_event_map,
284 .raw_event = amd_pmu_raw_event,
285 .max_events = ARRAY_SIZE(amd_perfmon_event_map),
286 .num_events = 4,
287 .event_bits = 48,
288 .event_mask = (1ULL << 48) - 1,
289 .apic = 1,
290 /* use highest bit to detect overflow */
291 .max_period = (1ULL << 47) - 1,
292 .get_event_constraints = amd_get_event_constraints,
293 .put_event_constraints = amd_put_event_constraints
294};
295
296static struct amd_nb *amd_alloc_nb(int cpu, int nb_id) 281static struct amd_nb *amd_alloc_nb(int cpu, int nb_id)
297{ 282{
298 struct amd_nb *nb; 283 struct amd_nb *nb;
@@ -309,57 +294,61 @@ static struct amd_nb *amd_alloc_nb(int cpu, int nb_id)
309 * initialize all possible NB constraints 294 * initialize all possible NB constraints
310 */ 295 */
311 for (i = 0; i < x86_pmu.num_events; i++) { 296 for (i = 0; i < x86_pmu.num_events; i++) {
312 set_bit(i, nb->event_constraints[i].idxmsk); 297 __set_bit(i, nb->event_constraints[i].idxmsk);
313 nb->event_constraints[i].weight = 1; 298 nb->event_constraints[i].weight = 1;
314 } 299 }
315 return nb; 300 return nb;
316} 301}
317 302
318static void amd_pmu_cpu_online(int cpu) 303static int amd_pmu_cpu_prepare(int cpu)
319{ 304{
320 struct cpu_hw_events *cpu1, *cpu2; 305 struct cpu_hw_events *cpuc = &per_cpu(cpu_hw_events, cpu);
321 struct amd_nb *nb = NULL; 306
307 WARN_ON_ONCE(cpuc->amd_nb);
308
309 if (boot_cpu_data.x86_max_cores < 2)
310 return NOTIFY_OK;
311
312 cpuc->amd_nb = amd_alloc_nb(cpu, -1);
313 if (!cpuc->amd_nb)
314 return NOTIFY_BAD;
315
316 return NOTIFY_OK;
317}
318
319static void amd_pmu_cpu_starting(int cpu)
320{
321 struct cpu_hw_events *cpuc = &per_cpu(cpu_hw_events, cpu);
322 struct amd_nb *nb;
322 int i, nb_id; 323 int i, nb_id;
323 324
324 if (boot_cpu_data.x86_max_cores < 2) 325 if (boot_cpu_data.x86_max_cores < 2)
325 return; 326 return;
326 327
327 /*
328 * function may be called too early in the
329 * boot process, in which case nb_id is bogus
330 */
331 nb_id = amd_get_nb_id(cpu); 328 nb_id = amd_get_nb_id(cpu);
332 if (nb_id == BAD_APICID) 329 WARN_ON_ONCE(nb_id == BAD_APICID);
333 return;
334
335 cpu1 = &per_cpu(cpu_hw_events, cpu);
336 cpu1->amd_nb = NULL;
337 330
338 raw_spin_lock(&amd_nb_lock); 331 raw_spin_lock(&amd_nb_lock);
339 332
340 for_each_online_cpu(i) { 333 for_each_online_cpu(i) {
341 cpu2 = &per_cpu(cpu_hw_events, i); 334 nb = per_cpu(cpu_hw_events, i).amd_nb;
342 nb = cpu2->amd_nb; 335 if (WARN_ON_ONCE(!nb))
343 if (!nb)
344 continue; 336 continue;
345 if (nb->nb_id == nb_id)
346 goto found;
347 }
348 337
349 nb = amd_alloc_nb(cpu, nb_id); 338 if (nb->nb_id == nb_id) {
350 if (!nb) { 339 kfree(cpuc->amd_nb);
351 pr_err("perf_events: failed NB allocation for CPU%d\n", cpu); 340 cpuc->amd_nb = nb;
352 raw_spin_unlock(&amd_nb_lock); 341 break;
353 return; 342 }
354 } 343 }
355found: 344
356 nb->refcnt++; 345 cpuc->amd_nb->nb_id = nb_id;
357 cpu1->amd_nb = nb; 346 cpuc->amd_nb->refcnt++;
358 347
359 raw_spin_unlock(&amd_nb_lock); 348 raw_spin_unlock(&amd_nb_lock);
360} 349}
361 350
362static void amd_pmu_cpu_offline(int cpu) 351static void amd_pmu_cpu_dead(int cpu)
363{ 352{
364 struct cpu_hw_events *cpuhw; 353 struct cpu_hw_events *cpuhw;
365 354
@@ -370,14 +359,44 @@ static void amd_pmu_cpu_offline(int cpu)
370 359
371 raw_spin_lock(&amd_nb_lock); 360 raw_spin_lock(&amd_nb_lock);
372 361
373 if (--cpuhw->amd_nb->refcnt == 0) 362 if (cpuhw->amd_nb) {
374 kfree(cpuhw->amd_nb); 363 struct amd_nb *nb = cpuhw->amd_nb;
364
365 if (nb->nb_id == -1 || --nb->refcnt == 0)
366 kfree(nb);
375 367
376 cpuhw->amd_nb = NULL; 368 cpuhw->amd_nb = NULL;
369 }
377 370
378 raw_spin_unlock(&amd_nb_lock); 371 raw_spin_unlock(&amd_nb_lock);
379} 372}
380 373
374static __initconst struct x86_pmu amd_pmu = {
375 .name = "AMD",
376 .handle_irq = x86_pmu_handle_irq,
377 .disable_all = x86_pmu_disable_all,
378 .enable_all = x86_pmu_enable_all,
379 .enable = x86_pmu_enable_event,
380 .disable = x86_pmu_disable_event,
381 .eventsel = MSR_K7_EVNTSEL0,
382 .perfctr = MSR_K7_PERFCTR0,
383 .event_map = amd_pmu_event_map,
384 .raw_event = amd_pmu_raw_event,
385 .max_events = ARRAY_SIZE(amd_perfmon_event_map),
386 .num_events = 4,
387 .event_bits = 48,
388 .event_mask = (1ULL << 48) - 1,
389 .apic = 1,
390 /* use highest bit to detect overflow */
391 .max_period = (1ULL << 47) - 1,
392 .get_event_constraints = amd_get_event_constraints,
393 .put_event_constraints = amd_put_event_constraints,
394
395 .cpu_prepare = amd_pmu_cpu_prepare,
396 .cpu_starting = amd_pmu_cpu_starting,
397 .cpu_dead = amd_pmu_cpu_dead,
398};
399
381static __init int amd_pmu_init(void) 400static __init int amd_pmu_init(void)
382{ 401{
383 /* Performance-monitoring supported from K7 and later: */ 402 /* Performance-monitoring supported from K7 and later: */
@@ -390,11 +409,6 @@ static __init int amd_pmu_init(void)
390 memcpy(hw_cache_event_ids, amd_hw_cache_event_ids, 409 memcpy(hw_cache_event_ids, amd_hw_cache_event_ids,
391 sizeof(hw_cache_event_ids)); 410 sizeof(hw_cache_event_ids));
392 411
393 /*
394 * explicitly initialize the boot cpu, other cpus will get
395 * the cpu hotplug callbacks from smp_init()
396 */
397 amd_pmu_cpu_online(smp_processor_id());
398 return 0; 412 return 0;
399} 413}
400 414
@@ -405,12 +419,4 @@ static int amd_pmu_init(void)
405 return 0; 419 return 0;
406} 420}
407 421
408static void amd_pmu_cpu_online(int cpu)
409{
410}
411
412static void amd_pmu_cpu_offline(int cpu)
413{
414}
415
416#endif 422#endif
diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c
index 44b60c852107..9c794ac87837 100644
--- a/arch/x86/kernel/cpu/perf_event_intel.c
+++ b/arch/x86/kernel/cpu/perf_event_intel.c
@@ -548,9 +548,9 @@ static inline void intel_pmu_ack_status(u64 ack)
548} 548}
549 549
550static inline void 550static inline void
551intel_pmu_disable_fixed(struct hw_perf_event *hwc, int __idx) 551intel_pmu_disable_fixed(struct hw_perf_event *hwc)
552{ 552{
553 int idx = __idx - X86_PMC_IDX_FIXED; 553 int idx = hwc->idx - X86_PMC_IDX_FIXED;
554 u64 ctrl_val, mask; 554 u64 ctrl_val, mask;
555 555
556 mask = 0xfULL << (idx * 4); 556 mask = 0xfULL << (idx * 4);
@@ -621,26 +621,28 @@ static void intel_pmu_drain_bts_buffer(void)
621} 621}
622 622
623static inline void 623static inline void
624intel_pmu_disable_event(struct hw_perf_event *hwc, int idx) 624intel_pmu_disable_event(struct perf_event *event)
625{ 625{
626 if (unlikely(idx == X86_PMC_IDX_FIXED_BTS)) { 626 struct hw_perf_event *hwc = &event->hw;
627
628 if (unlikely(hwc->idx == X86_PMC_IDX_FIXED_BTS)) {
627 intel_pmu_disable_bts(); 629 intel_pmu_disable_bts();
628 intel_pmu_drain_bts_buffer(); 630 intel_pmu_drain_bts_buffer();
629 return; 631 return;
630 } 632 }
631 633
632 if (unlikely(hwc->config_base == MSR_ARCH_PERFMON_FIXED_CTR_CTRL)) { 634 if (unlikely(hwc->config_base == MSR_ARCH_PERFMON_FIXED_CTR_CTRL)) {
633 intel_pmu_disable_fixed(hwc, idx); 635 intel_pmu_disable_fixed(hwc);
634 return; 636 return;
635 } 637 }
636 638
637 x86_pmu_disable_event(hwc, idx); 639 x86_pmu_disable_event(event);
638} 640}
639 641
640static inline void 642static inline void
641intel_pmu_enable_fixed(struct hw_perf_event *hwc, int __idx) 643intel_pmu_enable_fixed(struct hw_perf_event *hwc)
642{ 644{
643 int idx = __idx - X86_PMC_IDX_FIXED; 645 int idx = hwc->idx - X86_PMC_IDX_FIXED;
644 u64 ctrl_val, bits, mask; 646 u64 ctrl_val, bits, mask;
645 int err; 647 int err;
646 648
@@ -670,9 +672,11 @@ intel_pmu_enable_fixed(struct hw_perf_event *hwc, int __idx)
670 err = checking_wrmsrl(hwc->config_base, ctrl_val); 672 err = checking_wrmsrl(hwc->config_base, ctrl_val);
671} 673}
672 674
673static void intel_pmu_enable_event(struct hw_perf_event *hwc, int idx) 675static void intel_pmu_enable_event(struct perf_event *event)
674{ 676{
675 if (unlikely(idx == X86_PMC_IDX_FIXED_BTS)) { 677 struct hw_perf_event *hwc = &event->hw;
678
679 if (unlikely(hwc->idx == X86_PMC_IDX_FIXED_BTS)) {
676 if (!__get_cpu_var(cpu_hw_events).enabled) 680 if (!__get_cpu_var(cpu_hw_events).enabled)
677 return; 681 return;
678 682
@@ -681,11 +685,11 @@ static void intel_pmu_enable_event(struct hw_perf_event *hwc, int idx)
681 } 685 }
682 686
683 if (unlikely(hwc->config_base == MSR_ARCH_PERFMON_FIXED_CTR_CTRL)) { 687 if (unlikely(hwc->config_base == MSR_ARCH_PERFMON_FIXED_CTR_CTRL)) {
684 intel_pmu_enable_fixed(hwc, idx); 688 intel_pmu_enable_fixed(hwc);
685 return; 689 return;
686 } 690 }
687 691
688 __x86_pmu_enable_event(hwc, idx); 692 __x86_pmu_enable_event(hwc);
689} 693}
690 694
691/* 695/*
@@ -694,14 +698,8 @@ static void intel_pmu_enable_event(struct hw_perf_event *hwc, int idx)
694 */ 698 */
695static int intel_pmu_save_and_restart(struct perf_event *event) 699static int intel_pmu_save_and_restart(struct perf_event *event)
696{ 700{
697 struct hw_perf_event *hwc = &event->hw; 701 x86_perf_event_update(event);
698 int idx = hwc->idx; 702 return x86_perf_event_set_period(event);
699 int ret;
700
701 x86_perf_event_update(event, hwc, idx);
702 ret = x86_perf_event_set_period(event, hwc, idx);
703
704 return ret;
705} 703}
706 704
707static void intel_pmu_reset(void) 705static void intel_pmu_reset(void)
@@ -745,11 +743,11 @@ static int intel_pmu_handle_irq(struct pt_regs *regs)
745 743
746 cpuc = &__get_cpu_var(cpu_hw_events); 744 cpuc = &__get_cpu_var(cpu_hw_events);
747 745
748 perf_disable(); 746 intel_pmu_disable_all();
749 intel_pmu_drain_bts_buffer(); 747 intel_pmu_drain_bts_buffer();
750 status = intel_pmu_get_status(); 748 status = intel_pmu_get_status();
751 if (!status) { 749 if (!status) {
752 perf_enable(); 750 intel_pmu_enable_all();
753 return 0; 751 return 0;
754 } 752 }
755 753
@@ -759,8 +757,7 @@ again:
759 WARN_ONCE(1, "perfevents: irq loop stuck!\n"); 757 WARN_ONCE(1, "perfevents: irq loop stuck!\n");
760 perf_event_print_debug(); 758 perf_event_print_debug();
761 intel_pmu_reset(); 759 intel_pmu_reset();
762 perf_enable(); 760 goto done;
763 return 1;
764 } 761 }
765 762
766 inc_irq_stat(apic_perf_irqs); 763 inc_irq_stat(apic_perf_irqs);
@@ -768,7 +765,6 @@ again:
768 for_each_set_bit(bit, (unsigned long *)&status, X86_PMC_IDX_MAX) { 765 for_each_set_bit(bit, (unsigned long *)&status, X86_PMC_IDX_MAX) {
769 struct perf_event *event = cpuc->events[bit]; 766 struct perf_event *event = cpuc->events[bit];
770 767
771 clear_bit(bit, (unsigned long *) &status);
772 if (!test_bit(bit, cpuc->active_mask)) 768 if (!test_bit(bit, cpuc->active_mask))
773 continue; 769 continue;
774 770
@@ -778,7 +774,7 @@ again:
778 data.period = event->hw.last_period; 774 data.period = event->hw.last_period;
779 775
780 if (perf_event_overflow(event, 1, &data, regs)) 776 if (perf_event_overflow(event, 1, &data, regs))
781 intel_pmu_disable_event(&event->hw, bit); 777 x86_pmu_stop(event);
782 } 778 }
783 779
784 intel_pmu_ack_status(ack); 780 intel_pmu_ack_status(ack);
@@ -790,8 +786,8 @@ again:
790 if (status) 786 if (status)
791 goto again; 787 goto again;
792 788
793 perf_enable(); 789done:
794 790 intel_pmu_enable_all();
795 return 1; 791 return 1;
796} 792}
797 793
@@ -870,7 +866,10 @@ static __initconst struct x86_pmu intel_pmu = {
870 .max_period = (1ULL << 31) - 1, 866 .max_period = (1ULL << 31) - 1,
871 .enable_bts = intel_pmu_enable_bts, 867 .enable_bts = intel_pmu_enable_bts,
872 .disable_bts = intel_pmu_disable_bts, 868 .disable_bts = intel_pmu_disable_bts,
873 .get_event_constraints = intel_get_event_constraints 869 .get_event_constraints = intel_get_event_constraints,
870
871 .cpu_starting = init_debug_store_on_cpu,
872 .cpu_dying = fini_debug_store_on_cpu,
874}; 873};
875 874
876static __init int intel_pmu_init(void) 875static __init int intel_pmu_init(void)
@@ -937,6 +936,7 @@ static __init int intel_pmu_init(void)
937 936
938 case 26: /* 45 nm nehalem, "Bloomfield" */ 937 case 26: /* 45 nm nehalem, "Bloomfield" */
939 case 30: /* 45 nm nehalem, "Lynnfield" */ 938 case 30: /* 45 nm nehalem, "Lynnfield" */
939 case 46: /* 45 nm nehalem-ex, "Beckton" */
940 memcpy(hw_cache_event_ids, nehalem_hw_cache_event_ids, 940 memcpy(hw_cache_event_ids, nehalem_hw_cache_event_ids,
941 sizeof(hw_cache_event_ids)); 941 sizeof(hw_cache_event_ids));
942 942
diff --git a/arch/x86/kernel/cpu/perf_event_p6.c b/arch/x86/kernel/cpu/perf_event_p6.c
index a4e67b99d91c..a330485d14da 100644
--- a/arch/x86/kernel/cpu/perf_event_p6.c
+++ b/arch/x86/kernel/cpu/perf_event_p6.c
@@ -77,27 +77,29 @@ static void p6_pmu_enable_all(void)
77} 77}
78 78
79static inline void 79static inline void
80p6_pmu_disable_event(struct hw_perf_event *hwc, int idx) 80p6_pmu_disable_event(struct perf_event *event)
81{ 81{
82 struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); 82 struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
83 struct hw_perf_event *hwc = &event->hw;
83 u64 val = P6_NOP_EVENT; 84 u64 val = P6_NOP_EVENT;
84 85
85 if (cpuc->enabled) 86 if (cpuc->enabled)
86 val |= ARCH_PERFMON_EVENTSEL_ENABLE; 87 val |= ARCH_PERFMON_EVENTSEL_ENABLE;
87 88
88 (void)checking_wrmsrl(hwc->config_base + idx, val); 89 (void)checking_wrmsrl(hwc->config_base + hwc->idx, val);
89} 90}
90 91
91static void p6_pmu_enable_event(struct hw_perf_event *hwc, int idx) 92static void p6_pmu_enable_event(struct perf_event *event)
92{ 93{
93 struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); 94 struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
95 struct hw_perf_event *hwc = &event->hw;
94 u64 val; 96 u64 val;
95 97
96 val = hwc->config; 98 val = hwc->config;
97 if (cpuc->enabled) 99 if (cpuc->enabled)
98 val |= ARCH_PERFMON_EVENTSEL_ENABLE; 100 val |= ARCH_PERFMON_EVENTSEL_ENABLE;
99 101
100 (void)checking_wrmsrl(hwc->config_base + idx, val); 102 (void)checking_wrmsrl(hwc->config_base + hwc->idx, val);
101} 103}
102 104
103static __initconst struct x86_pmu p6_pmu = { 105static __initconst struct x86_pmu p6_pmu = {
diff --git a/arch/x86/kernel/cpu/vmware.c b/arch/x86/kernel/cpu/vmware.c
index 1cbed97b59cf..dfdb4dba2320 100644
--- a/arch/x86/kernel/cpu/vmware.c
+++ b/arch/x86/kernel/cpu/vmware.c
@@ -22,6 +22,7 @@
22 */ 22 */
23 23
24#include <linux/dmi.h> 24#include <linux/dmi.h>
25#include <linux/module.h>
25#include <asm/div64.h> 26#include <asm/div64.h>
26#include <asm/vmware.h> 27#include <asm/vmware.h>
27#include <asm/x86_init.h> 28#include <asm/x86_init.h>
@@ -101,6 +102,7 @@ int vmware_platform(void)
101 102
102 return 0; 103 return 0;
103} 104}
105EXPORT_SYMBOL(vmware_platform);
104 106
105/* 107/*
106 * VMware hypervisor takes care of exporting a reliable TSC to the guest. 108 * VMware hypervisor takes care of exporting a reliable TSC to the guest.
diff --git a/arch/x86/kernel/cpuid.c b/arch/x86/kernel/cpuid.c
index 83e5e628de73..8b862d5900fe 100644
--- a/arch/x86/kernel/cpuid.c
+++ b/arch/x86/kernel/cpuid.c
@@ -40,6 +40,7 @@
40#include <linux/cpu.h> 40#include <linux/cpu.h>
41#include <linux/notifier.h> 41#include <linux/notifier.h>
42#include <linux/uaccess.h> 42#include <linux/uaccess.h>
43#include <linux/gfp.h>
43 44
44#include <asm/processor.h> 45#include <asm/processor.h>
45#include <asm/msr.h> 46#include <asm/msr.h>
diff --git a/arch/x86/kernel/crash.c b/arch/x86/kernel/crash.c
index a4849c10a77e..ebd4c51d096a 100644
--- a/arch/x86/kernel/crash.c
+++ b/arch/x86/kernel/crash.c
@@ -27,7 +27,6 @@
27#include <asm/cpu.h> 27#include <asm/cpu.h>
28#include <asm/reboot.h> 28#include <asm/reboot.h>
29#include <asm/virtext.h> 29#include <asm/virtext.h>
30#include <asm/x86_init.h>
31 30
32#if defined(CONFIG_SMP) && defined(CONFIG_X86_LOCAL_APIC) 31#if defined(CONFIG_SMP) && defined(CONFIG_X86_LOCAL_APIC)
33 32
@@ -103,10 +102,5 @@ void native_machine_crash_shutdown(struct pt_regs *regs)
103#ifdef CONFIG_HPET_TIMER 102#ifdef CONFIG_HPET_TIMER
104 hpet_disable(); 103 hpet_disable();
105#endif 104#endif
106
107#ifdef CONFIG_X86_64
108 x86_platform.iommu_shutdown();
109#endif
110
111 crash_save_cpu(regs, safe_smp_processor_id()); 105 crash_save_cpu(regs, safe_smp_processor_id());
112} 106}
diff --git a/arch/x86/kernel/crash_dump_32.c b/arch/x86/kernel/crash_dump_32.c
index cd97ce18c29d..67414550c3cc 100644
--- a/arch/x86/kernel/crash_dump_32.c
+++ b/arch/x86/kernel/crash_dump_32.c
@@ -5,6 +5,7 @@
5 * Copyright (C) IBM Corporation, 2004. All rights reserved 5 * Copyright (C) IBM Corporation, 2004. All rights reserved
6 */ 6 */
7 7
8#include <linux/slab.h>
8#include <linux/errno.h> 9#include <linux/errno.h>
9#include <linux/highmem.h> 10#include <linux/highmem.h>
10#include <linux/crash_dump.h> 11#include <linux/crash_dump.h>
diff --git a/arch/x86/kernel/dumpstack.h b/arch/x86/kernel/dumpstack.h
index 4fd1420faffa..e1a93be4fd44 100644
--- a/arch/x86/kernel/dumpstack.h
+++ b/arch/x86/kernel/dumpstack.h
@@ -14,6 +14,8 @@
14#define get_bp(bp) asm("movq %%rbp, %0" : "=r" (bp) :) 14#define get_bp(bp) asm("movq %%rbp, %0" : "=r" (bp) :)
15#endif 15#endif
16 16
17#include <linux/uaccess.h>
18
17extern void 19extern void
18show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs, 20show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,
19 unsigned long *stack, unsigned long bp, char *log_lvl); 21 unsigned long *stack, unsigned long bp, char *log_lvl);
@@ -29,4 +31,26 @@ struct stack_frame {
29 struct stack_frame *next_frame; 31 struct stack_frame *next_frame;
30 unsigned long return_address; 32 unsigned long return_address;
31}; 33};
34
35struct stack_frame_ia32 {
36 u32 next_frame;
37 u32 return_address;
38};
39
40static inline unsigned long rewind_frame_pointer(int n)
41{
42 struct stack_frame *frame;
43
44 get_bp(frame);
45
46#ifdef CONFIG_FRAME_POINTER
47 while (n--) {
48 if (probe_kernel_address(&frame->next_frame, frame))
49 break;
50 }
32#endif 51#endif
52
53 return (unsigned long)frame;
54}
55
56#endif /* DUMPSTACK_H */
diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c
index d5e2a2ebb627..272c9f1f05f3 100644
--- a/arch/x86/kernel/dumpstack_64.c
+++ b/arch/x86/kernel/dumpstack_64.c
@@ -208,7 +208,7 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs,
208 if (in_irq_stack(stack, irq_stack, irq_stack_end)) { 208 if (in_irq_stack(stack, irq_stack, irq_stack_end)) {
209 if (ops->stack(data, "IRQ") < 0) 209 if (ops->stack(data, "IRQ") < 0)
210 break; 210 break;
211 bp = print_context_stack(tinfo, stack, bp, 211 bp = ops->walk_stack(tinfo, stack, bp,
212 ops, data, irq_stack_end, &graph); 212 ops, data, irq_stack_end, &graph);
213 /* 213 /*
214 * We link to the next stack (which would be 214 * We link to the next stack (which would be
@@ -229,7 +229,7 @@ void dump_trace(struct task_struct *task, struct pt_regs *regs,
229 /* 229 /*
230 * This handles the process stack: 230 * This handles the process stack:
231 */ 231 */
232 bp = print_context_stack(tinfo, stack, bp, ops, data, NULL, &graph); 232 bp = ops->walk_stack(tinfo, stack, bp, ops, data, NULL, &graph);
233 put_cpu(); 233 put_cpu();
234} 234}
235EXPORT_SYMBOL(dump_trace); 235EXPORT_SYMBOL(dump_trace);
diff --git a/arch/x86/kernel/e820.c b/arch/x86/kernel/e820.c
index 740b440fbd73..7bca3c6a02fb 100644
--- a/arch/x86/kernel/e820.c
+++ b/arch/x86/kernel/e820.c
@@ -519,29 +519,45 @@ u64 __init e820_remove_range(u64 start, u64 size, unsigned old_type,
519 printk(KERN_DEBUG "e820 remove range: %016Lx - %016Lx ", 519 printk(KERN_DEBUG "e820 remove range: %016Lx - %016Lx ",
520 (unsigned long long) start, 520 (unsigned long long) start,
521 (unsigned long long) end); 521 (unsigned long long) end);
522 e820_print_type(old_type); 522 if (checktype)
523 e820_print_type(old_type);
523 printk(KERN_CONT "\n"); 524 printk(KERN_CONT "\n");
524 525
525 for (i = 0; i < e820.nr_map; i++) { 526 for (i = 0; i < e820.nr_map; i++) {
526 struct e820entry *ei = &e820.map[i]; 527 struct e820entry *ei = &e820.map[i];
527 u64 final_start, final_end; 528 u64 final_start, final_end;
529 u64 ei_end;
528 530
529 if (checktype && ei->type != old_type) 531 if (checktype && ei->type != old_type)
530 continue; 532 continue;
533
534 ei_end = ei->addr + ei->size;
531 /* totally covered? */ 535 /* totally covered? */
532 if (ei->addr >= start && 536 if (ei->addr >= start && ei_end <= end) {
533 (ei->addr + ei->size) <= (start + size)) {
534 real_removed_size += ei->size; 537 real_removed_size += ei->size;
535 memset(ei, 0, sizeof(struct e820entry)); 538 memset(ei, 0, sizeof(struct e820entry));
536 continue; 539 continue;
537 } 540 }
541
542 /* new range is totally covered? */
543 if (ei->addr < start && ei_end > end) {
544 e820_add_region(end, ei_end - end, ei->type);
545 ei->size = start - ei->addr;
546 real_removed_size += size;
547 continue;
548 }
549
538 /* partially covered */ 550 /* partially covered */
539 final_start = max(start, ei->addr); 551 final_start = max(start, ei->addr);
540 final_end = min(start + size, ei->addr + ei->size); 552 final_end = min(end, ei_end);
541 if (final_start >= final_end) 553 if (final_start >= final_end)
542 continue; 554 continue;
543 real_removed_size += final_end - final_start; 555 real_removed_size += final_end - final_start;
544 556
557 /*
558 * left range could be head or tail, so need to update
559 * size at first.
560 */
545 ei->size -= final_end - final_start; 561 ei->size -= final_end - final_start;
546 if (ei->addr < final_start) 562 if (ei->addr < final_start)
547 continue; 563 continue;
diff --git a/arch/x86/kernel/head32.c b/arch/x86/kernel/head32.c
index adedeef1dedc..b2e246037392 100644
--- a/arch/x86/kernel/head32.c
+++ b/arch/x86/kernel/head32.c
@@ -7,6 +7,7 @@
7 7
8#include <linux/init.h> 8#include <linux/init.h>
9#include <linux/start_kernel.h> 9#include <linux/start_kernel.h>
10#include <linux/mm.h>
10 11
11#include <asm/setup.h> 12#include <asm/setup.h>
12#include <asm/sections.h> 13#include <asm/sections.h>
@@ -44,9 +45,10 @@ void __init i386_start_kernel(void)
44#ifdef CONFIG_BLK_DEV_INITRD 45#ifdef CONFIG_BLK_DEV_INITRD
45 /* Reserve INITRD */ 46 /* Reserve INITRD */
46 if (boot_params.hdr.type_of_loader && boot_params.hdr.ramdisk_image) { 47 if (boot_params.hdr.type_of_loader && boot_params.hdr.ramdisk_image) {
48 /* Assume only end is not page aligned */
47 u64 ramdisk_image = boot_params.hdr.ramdisk_image; 49 u64 ramdisk_image = boot_params.hdr.ramdisk_image;
48 u64 ramdisk_size = boot_params.hdr.ramdisk_size; 50 u64 ramdisk_size = boot_params.hdr.ramdisk_size;
49 u64 ramdisk_end = ramdisk_image + ramdisk_size; 51 u64 ramdisk_end = PAGE_ALIGN(ramdisk_image + ramdisk_size);
50 reserve_early(ramdisk_image, ramdisk_end, "RAMDISK"); 52 reserve_early(ramdisk_image, ramdisk_end, "RAMDISK");
51 } 53 }
52#endif 54#endif
diff --git a/arch/x86/kernel/head64.c b/arch/x86/kernel/head64.c
index b5a9896ca1e7..7147143fd614 100644
--- a/arch/x86/kernel/head64.c
+++ b/arch/x86/kernel/head64.c
@@ -103,9 +103,10 @@ void __init x86_64_start_reservations(char *real_mode_data)
103#ifdef CONFIG_BLK_DEV_INITRD 103#ifdef CONFIG_BLK_DEV_INITRD
104 /* Reserve INITRD */ 104 /* Reserve INITRD */
105 if (boot_params.hdr.type_of_loader && boot_params.hdr.ramdisk_image) { 105 if (boot_params.hdr.type_of_loader && boot_params.hdr.ramdisk_image) {
106 /* Assume only end is not page aligned */
106 unsigned long ramdisk_image = boot_params.hdr.ramdisk_image; 107 unsigned long ramdisk_image = boot_params.hdr.ramdisk_image;
107 unsigned long ramdisk_size = boot_params.hdr.ramdisk_size; 108 unsigned long ramdisk_size = boot_params.hdr.ramdisk_size;
108 unsigned long ramdisk_end = ramdisk_image + ramdisk_size; 109 unsigned long ramdisk_end = PAGE_ALIGN(ramdisk_image + ramdisk_size);
109 reserve_early(ramdisk_image, ramdisk_end, "RAMDISK"); 110 reserve_early(ramdisk_image, ramdisk_end, "RAMDISK");
110 } 111 }
111#endif 112#endif
diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c
index ee4fa1bfcb33..23b4ecdffa9b 100644
--- a/arch/x86/kernel/hpet.c
+++ b/arch/x86/kernel/hpet.c
@@ -4,6 +4,7 @@
4#include <linux/sysdev.h> 4#include <linux/sysdev.h>
5#include <linux/delay.h> 5#include <linux/delay.h>
6#include <linux/errno.h> 6#include <linux/errno.h>
7#include <linux/slab.h>
7#include <linux/hpet.h> 8#include <linux/hpet.h>
8#include <linux/init.h> 9#include <linux/init.h>
9#include <linux/cpu.h> 10#include <linux/cpu.h>
@@ -399,9 +400,15 @@ static int hpet_next_event(unsigned long delta,
399 * then we might have a real hardware problem. We can not do 400 * then we might have a real hardware problem. We can not do
400 * much about it here, but at least alert the user/admin with 401 * much about it here, but at least alert the user/admin with
401 * a prominent warning. 402 * a prominent warning.
403 * An erratum on some chipsets (ICH9,..), results in comparator read
404 * immediately following a write returning old value. Workaround
405 * for this is to read this value second time, when first
406 * read returns old value.
402 */ 407 */
403 WARN_ONCE(hpet_readl(HPET_Tn_CMP(timer)) != cnt, 408 if (unlikely((u32)hpet_readl(HPET_Tn_CMP(timer)) != cnt)) {
409 WARN_ONCE(hpet_readl(HPET_Tn_CMP(timer)) != cnt,
404 KERN_WARNING "hpet: compare register read back failed.\n"); 410 KERN_WARNING "hpet: compare register read back failed.\n");
411 }
405 412
406 return (s32)(hpet_readl(HPET_COUNTER) - cnt) >= 0 ? -ETIME : 0; 413 return (s32)(hpet_readl(HPET_COUNTER) - cnt) >= 0 ? -ETIME : 0;
407} 414}
@@ -1143,6 +1150,7 @@ int hpet_set_periodic_freq(unsigned long freq)
1143 do_div(clc, freq); 1150 do_div(clc, freq);
1144 clc >>= hpet_clockevent.shift; 1151 clc >>= hpet_clockevent.shift;
1145 hpet_pie_delta = clc; 1152 hpet_pie_delta = clc;
1153 hpet_pie_limit = 0;
1146 } 1154 }
1147 return 1; 1155 return 1;
1148} 1156}
diff --git a/arch/x86/kernel/i387.c b/arch/x86/kernel/i387.c
index c01a2b846d47..54c31c285488 100644
--- a/arch/x86/kernel/i387.c
+++ b/arch/x86/kernel/i387.c
@@ -8,6 +8,7 @@
8#include <linux/module.h> 8#include <linux/module.h>
9#include <linux/regset.h> 9#include <linux/regset.h>
10#include <linux/sched.h> 10#include <linux/sched.h>
11#include <linux/slab.h>
11 12
12#include <asm/sigcontext.h> 13#include <asm/sigcontext.h>
13#include <asm/processor.h> 14#include <asm/processor.h>
diff --git a/arch/x86/kernel/i8259.c b/arch/x86/kernel/i8259.c
index fb725ee15f55..7c9f02c130f3 100644
--- a/arch/x86/kernel/i8259.c
+++ b/arch/x86/kernel/i8259.c
@@ -5,7 +5,6 @@
5#include <linux/ioport.h> 5#include <linux/ioport.h>
6#include <linux/interrupt.h> 6#include <linux/interrupt.h>
7#include <linux/timex.h> 7#include <linux/timex.h>
8#include <linux/slab.h>
9#include <linux/random.h> 8#include <linux/random.h>
10#include <linux/init.h> 9#include <linux/init.h>
11#include <linux/kernel_stat.h> 10#include <linux/kernel_stat.h>
diff --git a/arch/x86/kernel/irqinit.c b/arch/x86/kernel/irqinit.c
index ef257fc2921b..0ed2d300cd46 100644
--- a/arch/x86/kernel/irqinit.c
+++ b/arch/x86/kernel/irqinit.c
@@ -5,7 +5,6 @@
5#include <linux/ioport.h> 5#include <linux/ioport.h>
6#include <linux/interrupt.h> 6#include <linux/interrupt.h>
7#include <linux/timex.h> 7#include <linux/timex.h>
8#include <linux/slab.h>
9#include <linux/random.h> 8#include <linux/random.h>
10#include <linux/kprobes.h> 9#include <linux/kprobes.h>
11#include <linux/init.h> 10#include <linux/init.h>
@@ -141,6 +140,28 @@ void __init init_IRQ(void)
141 x86_init.irqs.intr_init(); 140 x86_init.irqs.intr_init();
142} 141}
143 142
143/*
144 * Setup the vector to irq mappings.
145 */
146void setup_vector_irq(int cpu)
147{
148#ifndef CONFIG_X86_IO_APIC
149 int irq;
150
151 /*
152 * On most of the platforms, legacy PIC delivers the interrupts on the
153 * boot cpu. But there are certain platforms where PIC interrupts are
154 * delivered to multiple cpu's. If the legacy IRQ is handled by the
155 * legacy PIC, for the new cpu that is coming online, setup the static
156 * legacy vector to irq mapping:
157 */
158 for (irq = 0; irq < legacy_pic->nr_legacy_irqs; irq++)
159 per_cpu(vector_irq, cpu)[IRQ0_VECTOR + irq] = irq;
160#endif
161
162 __setup_vector_irq(cpu);
163}
164
144static void __init smp_intr_init(void) 165static void __init smp_intr_init(void)
145{ 166{
146#ifdef CONFIG_SMP 167#ifdef CONFIG_SMP
diff --git a/arch/x86/kernel/k8.c b/arch/x86/kernel/k8.c
index 9b895464dd03..0f7bc20cfcde 100644
--- a/arch/x86/kernel/k8.c
+++ b/arch/x86/kernel/k8.c
@@ -2,8 +2,8 @@
2 * Shared support code for AMD K8 northbridges and derivates. 2 * Shared support code for AMD K8 northbridges and derivates.
3 * Copyright 2006 Andi Kleen, SUSE Labs. Subject to GPLv2. 3 * Copyright 2006 Andi Kleen, SUSE Labs. Subject to GPLv2.
4 */ 4 */
5#include <linux/gfp.h>
6#include <linux/types.h> 5#include <linux/types.h>
6#include <linux/slab.h>
7#include <linux/init.h> 7#include <linux/init.h>
8#include <linux/errno.h> 8#include <linux/errno.h>
9#include <linux/module.h> 9#include <linux/module.h>
diff --git a/arch/x86/kernel/kdebugfs.c b/arch/x86/kernel/kdebugfs.c
index e444357375ce..8afd9f321f10 100644
--- a/arch/x86/kernel/kdebugfs.c
+++ b/arch/x86/kernel/kdebugfs.c
@@ -9,6 +9,7 @@
9#include <linux/debugfs.h> 9#include <linux/debugfs.h>
10#include <linux/uaccess.h> 10#include <linux/uaccess.h>
11#include <linux/module.h> 11#include <linux/module.h>
12#include <linux/slab.h>
12#include <linux/init.h> 13#include <linux/init.h>
13#include <linux/stat.h> 14#include <linux/stat.h>
14#include <linux/io.h> 15#include <linux/io.h>
diff --git a/arch/x86/kernel/kgdb.c b/arch/x86/kernel/kgdb.c
index bfba6019d762..b2258ca91003 100644
--- a/arch/x86/kernel/kgdb.c
+++ b/arch/x86/kernel/kgdb.c
@@ -618,8 +618,8 @@ int kgdb_arch_init(void)
618 * portion of kgdb because this operation requires mutexs to 618 * portion of kgdb because this operation requires mutexs to
619 * complete. 619 * complete.
620 */ 620 */
621 hw_breakpoint_init(&attr);
621 attr.bp_addr = (unsigned long)kgdb_arch_init; 622 attr.bp_addr = (unsigned long)kgdb_arch_init;
622 attr.type = PERF_TYPE_BREAKPOINT;
623 attr.bp_len = HW_BREAKPOINT_LEN_1; 623 attr.bp_len = HW_BREAKPOINT_LEN_1;
624 attr.bp_type = HW_BREAKPOINT_W; 624 attr.bp_type = HW_BREAKPOINT_W;
625 attr.disabled = 1; 625 attr.disabled = 1;
diff --git a/arch/x86/kernel/kprobes.c b/arch/x86/kernel/kprobes.c
index b43bbaebe2c0..1658efdfb4e5 100644
--- a/arch/x86/kernel/kprobes.c
+++ b/arch/x86/kernel/kprobes.c
@@ -534,20 +534,6 @@ static int __kprobes kprobe_handler(struct pt_regs *regs)
534 struct kprobe_ctlblk *kcb; 534 struct kprobe_ctlblk *kcb;
535 535
536 addr = (kprobe_opcode_t *)(regs->ip - sizeof(kprobe_opcode_t)); 536 addr = (kprobe_opcode_t *)(regs->ip - sizeof(kprobe_opcode_t));
537 if (*addr != BREAKPOINT_INSTRUCTION) {
538 /*
539 * The breakpoint instruction was removed right
540 * after we hit it. Another cpu has removed
541 * either a probepoint or a debugger breakpoint
542 * at this address. In either case, no further
543 * handling of this interrupt is appropriate.
544 * Back up over the (now missing) int3 and run
545 * the original instruction.
546 */
547 regs->ip = (unsigned long)addr;
548 return 1;
549 }
550
551 /* 537 /*
552 * We don't want to be preempted for the entire 538 * We don't want to be preempted for the entire
553 * duration of kprobe processing. We conditionally 539 * duration of kprobe processing. We conditionally
@@ -579,6 +565,19 @@ static int __kprobes kprobe_handler(struct pt_regs *regs)
579 setup_singlestep(p, regs, kcb, 0); 565 setup_singlestep(p, regs, kcb, 0);
580 return 1; 566 return 1;
581 } 567 }
568 } else if (*addr != BREAKPOINT_INSTRUCTION) {
569 /*
570 * The breakpoint instruction was removed right
571 * after we hit it. Another cpu has removed
572 * either a probepoint or a debugger breakpoint
573 * at this address. In either case, no further
574 * handling of this interrupt is appropriate.
575 * Back up over the (now missing) int3 and run
576 * the original instruction.
577 */
578 regs->ip = (unsigned long)addr;
579 preempt_enable_no_resched();
580 return 1;
582 } else if (kprobe_running()) { 581 } else if (kprobe_running()) {
583 p = __get_cpu_var(current_kprobe); 582 p = __get_cpu_var(current_kprobe);
584 if (p->break_handler && p->break_handler(p, regs)) { 583 if (p->break_handler && p->break_handler(p, regs)) {
diff --git a/arch/x86/kernel/ldt.c b/arch/x86/kernel/ldt.c
index ec6ef60cbd17..ea697263b373 100644
--- a/arch/x86/kernel/ldt.c
+++ b/arch/x86/kernel/ldt.c
@@ -7,6 +7,7 @@
7 */ 7 */
8 8
9#include <linux/errno.h> 9#include <linux/errno.h>
10#include <linux/gfp.h>
10#include <linux/sched.h> 11#include <linux/sched.h>
11#include <linux/string.h> 12#include <linux/string.h>
12#include <linux/mm.h> 13#include <linux/mm.h>
diff --git a/arch/x86/kernel/machine_kexec_64.c b/arch/x86/kernel/machine_kexec_64.c
index 4a8bb82248ae..035c8c529181 100644
--- a/arch/x86/kernel/machine_kexec_64.c
+++ b/arch/x86/kernel/machine_kexec_64.c
@@ -9,6 +9,7 @@
9#include <linux/mm.h> 9#include <linux/mm.h>
10#include <linux/kexec.h> 10#include <linux/kexec.h>
11#include <linux/string.h> 11#include <linux/string.h>
12#include <linux/gfp.h>
12#include <linux/reboot.h> 13#include <linux/reboot.h>
13#include <linux/numa.h> 14#include <linux/numa.h>
14#include <linux/ftrace.h> 15#include <linux/ftrace.h>
diff --git a/arch/x86/kernel/mca_32.c b/arch/x86/kernel/mca_32.c
index 845d80ce1ef1..63eaf6596233 100644
--- a/arch/x86/kernel/mca_32.c
+++ b/arch/x86/kernel/mca_32.c
@@ -42,6 +42,7 @@
42#include <linux/kernel.h> 42#include <linux/kernel.h>
43#include <linux/mca.h> 43#include <linux/mca.h>
44#include <linux/kprobes.h> 44#include <linux/kprobes.h>
45#include <linux/slab.h>
45#include <asm/system.h> 46#include <asm/system.h>
46#include <asm/io.h> 47#include <asm/io.h>
47#include <linux/proc_fs.h> 48#include <linux/proc_fs.h>
diff --git a/arch/x86/kernel/module.c b/arch/x86/kernel/module.c
index 89f386f044e4..e0bc186d7501 100644
--- a/arch/x86/kernel/module.c
+++ b/arch/x86/kernel/module.c
@@ -23,6 +23,7 @@
23#include <linux/kernel.h> 23#include <linux/kernel.h>
24#include <linux/bug.h> 24#include <linux/bug.h>
25#include <linux/mm.h> 25#include <linux/mm.h>
26#include <linux/gfp.h>
26 27
27#include <asm/system.h> 28#include <asm/system.h>
28#include <asm/page.h> 29#include <asm/page.h>
diff --git a/arch/x86/kernel/mpparse.c b/arch/x86/kernel/mpparse.c
index a2c1edd2d3ac..e81030f71a8f 100644
--- a/arch/x86/kernel/mpparse.c
+++ b/arch/x86/kernel/mpparse.c
@@ -664,7 +664,7 @@ static void __init smp_reserve_memory(struct mpf_intel *mpf)
664{ 664{
665 unsigned long size = get_mpc_size(mpf->physptr); 665 unsigned long size = get_mpc_size(mpf->physptr);
666 666
667 reserve_early(mpf->physptr, mpf->physptr+size, "MP-table mpc"); 667 reserve_early_overlap_ok(mpf->physptr, mpf->physptr+size, "MP-table mpc");
668} 668}
669 669
670static int __init smp_scan_config(unsigned long base, unsigned long length) 670static int __init smp_scan_config(unsigned long base, unsigned long length)
@@ -693,7 +693,7 @@ static int __init smp_scan_config(unsigned long base, unsigned long length)
693 mpf, (u64)virt_to_phys(mpf)); 693 mpf, (u64)virt_to_phys(mpf));
694 694
695 mem = virt_to_phys(mpf); 695 mem = virt_to_phys(mpf);
696 reserve_early(mem, mem + sizeof(*mpf), "MP-table mpf"); 696 reserve_early_overlap_ok(mem, mem + sizeof(*mpf), "MP-table mpf");
697 if (mpf->physptr) 697 if (mpf->physptr)
698 smp_reserve_memory(mpf); 698 smp_reserve_memory(mpf);
699 699
diff --git a/arch/x86/kernel/msr.c b/arch/x86/kernel/msr.c
index 206735ac8cbd..4d4468e9f47c 100644
--- a/arch/x86/kernel/msr.c
+++ b/arch/x86/kernel/msr.c
@@ -37,6 +37,7 @@
37#include <linux/cpu.h> 37#include <linux/cpu.h>
38#include <linux/notifier.h> 38#include <linux/notifier.h>
39#include <linux/uaccess.h> 39#include <linux/uaccess.h>
40#include <linux/gfp.h>
40 41
41#include <asm/processor.h> 42#include <asm/processor.h>
42#include <asm/msr.h> 43#include <asm/msr.h>
diff --git a/arch/x86/kernel/pci-dma.c b/arch/x86/kernel/pci-dma.c
index a4ac764a6880..4b7e3d8b01dd 100644
--- a/arch/x86/kernel/pci-dma.c
+++ b/arch/x86/kernel/pci-dma.c
@@ -2,6 +2,7 @@
2#include <linux/dma-debug.h> 2#include <linux/dma-debug.h>
3#include <linux/dmar.h> 3#include <linux/dmar.h>
4#include <linux/bootmem.h> 4#include <linux/bootmem.h>
5#include <linux/gfp.h>
5#include <linux/pci.h> 6#include <linux/pci.h>
6#include <linux/kmemleak.h> 7#include <linux/kmemleak.h>
7 8
diff --git a/arch/x86/kernel/pci-gart_64.c b/arch/x86/kernel/pci-gart_64.c
index f3af115a573a..0f7f130caa67 100644
--- a/arch/x86/kernel/pci-gart_64.c
+++ b/arch/x86/kernel/pci-gart_64.c
@@ -29,6 +29,7 @@
29#include <linux/iommu-helper.h> 29#include <linux/iommu-helper.h>
30#include <linux/sysdev.h> 30#include <linux/sysdev.h>
31#include <linux/io.h> 31#include <linux/io.h>
32#include <linux/gfp.h>
32#include <asm/atomic.h> 33#include <asm/atomic.h>
33#include <asm/mtrr.h> 34#include <asm/mtrr.h>
34#include <asm/pgtable.h> 35#include <asm/pgtable.h>
@@ -564,6 +565,9 @@ static void enable_gart_translations(void)
564 565
565 enable_gart_translation(dev, __pa(agp_gatt_table)); 566 enable_gart_translation(dev, __pa(agp_gatt_table));
566 } 567 }
568
569 /* Flush the GART-TLB to remove stale entries */
570 k8_flush_garts();
567} 571}
568 572
569/* 573/*
diff --git a/arch/x86/kernel/pci-nommu.c b/arch/x86/kernel/pci-nommu.c
index 22be12b60a8f..3af4af810c07 100644
--- a/arch/x86/kernel/pci-nommu.c
+++ b/arch/x86/kernel/pci-nommu.c
@@ -4,6 +4,7 @@
4#include <linux/scatterlist.h> 4#include <linux/scatterlist.h>
5#include <linux/string.h> 5#include <linux/string.h>
6#include <linux/init.h> 6#include <linux/init.h>
7#include <linux/gfp.h>
7#include <linux/pci.h> 8#include <linux/pci.h>
8#include <linux/mm.h> 9#include <linux/mm.h>
9 10
diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c
index ad9540676fcc..0415c3ef91b5 100644
--- a/arch/x86/kernel/process.c
+++ b/arch/x86/kernel/process.c
@@ -526,21 +526,39 @@ static int __cpuinit mwait_usable(const struct cpuinfo_x86 *c)
526} 526}
527 527
528/* 528/*
529 * Check for AMD CPUs, which have potentially C1E support 529 * Check for AMD CPUs, where APIC timer interrupt does not wake up CPU from C1e.
530 * For more information see
531 * - Erratum #400 for NPT family 0xf and family 0x10 CPUs
532 * - Erratum #365 for family 0x11 (not affected because C1e not in use)
530 */ 533 */
531static int __cpuinit check_c1e_idle(const struct cpuinfo_x86 *c) 534static int __cpuinit check_c1e_idle(const struct cpuinfo_x86 *c)
532{ 535{
536 u64 val;
533 if (c->x86_vendor != X86_VENDOR_AMD) 537 if (c->x86_vendor != X86_VENDOR_AMD)
534 return 0; 538 goto no_c1e_idle;
535
536 if (c->x86 < 0x0F)
537 return 0;
538 539
539 /* Family 0x0f models < rev F do not have C1E */ 540 /* Family 0x0f models < rev F do not have C1E */
540 if (c->x86 == 0x0f && c->x86_model < 0x40) 541 if (c->x86 == 0x0F && c->x86_model >= 0x40)
541 return 0; 542 return 1;
542 543
543 return 1; 544 if (c->x86 == 0x10) {
545 /*
546 * check OSVW bit for CPUs that are not affected
547 * by erratum #400
548 */
549 if (cpu_has(c, X86_FEATURE_OSVW)) {
550 rdmsrl(MSR_AMD64_OSVW_ID_LENGTH, val);
551 if (val >= 2) {
552 rdmsrl(MSR_AMD64_OSVW_STATUS, val);
553 if (!(val & BIT(1)))
554 goto no_c1e_idle;
555 }
556 }
557 return 1;
558 }
559
560no_c1e_idle:
561 return 0;
544} 562}
545 563
546static cpumask_var_t c1e_mask; 564static cpumask_var_t c1e_mask;
diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c
index dc9690b4c4cc..17cb3295cbf7 100644
--- a/arch/x86/kernel/process_64.c
+++ b/arch/x86/kernel/process_64.c
@@ -276,12 +276,12 @@ int copy_thread(unsigned long clone_flags, unsigned long sp,
276 276
277 set_tsk_thread_flag(p, TIF_FORK); 277 set_tsk_thread_flag(p, TIF_FORK);
278 278
279 p->thread.fs = me->thread.fs;
280 p->thread.gs = me->thread.gs;
281 p->thread.io_bitmap_ptr = NULL; 279 p->thread.io_bitmap_ptr = NULL;
282 280
283 savesegment(gs, p->thread.gsindex); 281 savesegment(gs, p->thread.gsindex);
282 p->thread.gs = p->thread.gsindex ? 0 : me->thread.gs;
284 savesegment(fs, p->thread.fsindex); 283 savesegment(fs, p->thread.fsindex);
284 p->thread.fs = p->thread.fsindex ? 0 : me->thread.fs;
285 savesegment(es, p->thread.es); 285 savesegment(es, p->thread.es);
286 savesegment(ds, p->thread.ds); 286 savesegment(ds, p->thread.ds);
287 287
diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c
index a503b1fd04e5..2e9b55027b7e 100644
--- a/arch/x86/kernel/ptrace.c
+++ b/arch/x86/kernel/ptrace.c
@@ -12,6 +12,7 @@
12#include <linux/mm.h> 12#include <linux/mm.h>
13#include <linux/smp.h> 13#include <linux/smp.h>
14#include <linux/errno.h> 14#include <linux/errno.h>
15#include <linux/slab.h>
15#include <linux/ptrace.h> 16#include <linux/ptrace.h>
16#include <linux/regset.h> 17#include <linux/regset.h>
17#include <linux/tracehook.h> 18#include <linux/tracehook.h>
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 5d7ba1a449bd..c4851eff57b3 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -55,7 +55,6 @@
55#include <linux/stddef.h> 55#include <linux/stddef.h>
56#include <linux/unistd.h> 56#include <linux/unistd.h>
57#include <linux/ptrace.h> 57#include <linux/ptrace.h>
58#include <linux/slab.h>
59#include <linux/user.h> 58#include <linux/user.h>
60#include <linux/delay.h> 59#include <linux/delay.h>
61 60
@@ -314,16 +313,17 @@ static void __init reserve_brk(void)
314#define MAX_MAP_CHUNK (NR_FIX_BTMAPS << PAGE_SHIFT) 313#define MAX_MAP_CHUNK (NR_FIX_BTMAPS << PAGE_SHIFT)
315static void __init relocate_initrd(void) 314static void __init relocate_initrd(void)
316{ 315{
317 316 /* Assume only end is not page aligned */
318 u64 ramdisk_image = boot_params.hdr.ramdisk_image; 317 u64 ramdisk_image = boot_params.hdr.ramdisk_image;
319 u64 ramdisk_size = boot_params.hdr.ramdisk_size; 318 u64 ramdisk_size = boot_params.hdr.ramdisk_size;
319 u64 area_size = PAGE_ALIGN(ramdisk_size);
320 u64 end_of_lowmem = max_low_pfn_mapped << PAGE_SHIFT; 320 u64 end_of_lowmem = max_low_pfn_mapped << PAGE_SHIFT;
321 u64 ramdisk_here; 321 u64 ramdisk_here;
322 unsigned long slop, clen, mapaddr; 322 unsigned long slop, clen, mapaddr;
323 char *p, *q; 323 char *p, *q;
324 324
325 /* We need to move the initrd down into lowmem */ 325 /* We need to move the initrd down into lowmem */
326 ramdisk_here = find_e820_area(0, end_of_lowmem, ramdisk_size, 326 ramdisk_here = find_e820_area(0, end_of_lowmem, area_size,
327 PAGE_SIZE); 327 PAGE_SIZE);
328 328
329 if (ramdisk_here == -1ULL) 329 if (ramdisk_here == -1ULL)
@@ -332,7 +332,7 @@ static void __init relocate_initrd(void)
332 332
333 /* Note: this includes all the lowmem currently occupied by 333 /* Note: this includes all the lowmem currently occupied by
334 the initrd, we rely on that fact to keep the data intact. */ 334 the initrd, we rely on that fact to keep the data intact. */
335 reserve_early(ramdisk_here, ramdisk_here + ramdisk_size, 335 reserve_early(ramdisk_here, ramdisk_here + area_size,
336 "NEW RAMDISK"); 336 "NEW RAMDISK");
337 initrd_start = ramdisk_here + PAGE_OFFSET; 337 initrd_start = ramdisk_here + PAGE_OFFSET;
338 initrd_end = initrd_start + ramdisk_size; 338 initrd_end = initrd_start + ramdisk_size;
@@ -376,9 +376,10 @@ static void __init relocate_initrd(void)
376 376
377static void __init reserve_initrd(void) 377static void __init reserve_initrd(void)
378{ 378{
379 /* Assume only end is not page aligned */
379 u64 ramdisk_image = boot_params.hdr.ramdisk_image; 380 u64 ramdisk_image = boot_params.hdr.ramdisk_image;
380 u64 ramdisk_size = boot_params.hdr.ramdisk_size; 381 u64 ramdisk_size = boot_params.hdr.ramdisk_size;
381 u64 ramdisk_end = ramdisk_image + ramdisk_size; 382 u64 ramdisk_end = PAGE_ALIGN(ramdisk_image + ramdisk_size);
382 u64 end_of_lowmem = max_low_pfn_mapped << PAGE_SHIFT; 383 u64 end_of_lowmem = max_low_pfn_mapped << PAGE_SHIFT;
383 384
384 if (!boot_params.hdr.type_of_loader || 385 if (!boot_params.hdr.type_of_loader ||
@@ -606,6 +607,16 @@ static int __init setup_elfcorehdr(char *arg)
606early_param("elfcorehdr", setup_elfcorehdr); 607early_param("elfcorehdr", setup_elfcorehdr);
607#endif 608#endif
608 609
610static __init void reserve_ibft_region(void)
611{
612 unsigned long addr, size = 0;
613
614 addr = find_ibft_region(&size);
615
616 if (size)
617 reserve_early_overlap_ok(addr, addr + size, "ibft");
618}
619
609#ifdef CONFIG_X86_RESERVE_LOW_64K 620#ifdef CONFIG_X86_RESERVE_LOW_64K
610static int __init dmi_low_memory_corruption(const struct dmi_system_id *d) 621static int __init dmi_low_memory_corruption(const struct dmi_system_id *d)
611{ 622{
@@ -908,6 +919,8 @@ void __init setup_arch(char **cmdline_p)
908 */ 919 */
909 find_smp_config(); 920 find_smp_config();
910 921
922 reserve_ibft_region();
923
911 reserve_trampoline_memory(); 924 reserve_trampoline_memory();
912 925
913#ifdef CONFIG_ACPI_SLEEP 926#ifdef CONFIG_ACPI_SLEEP
@@ -975,8 +988,6 @@ void __init setup_arch(char **cmdline_p)
975 988
976 dma32_reserve_bootmem(); 989 dma32_reserve_bootmem();
977 990
978 reserve_ibft_region();
979
980#ifdef CONFIG_KVM_CLOCK 991#ifdef CONFIG_KVM_CLOCK
981 kvmclock_init(); 992 kvmclock_init();
982#endif 993#endif
diff --git a/arch/x86/kernel/smp.c b/arch/x86/kernel/smp.c
index ec1de97600e7..d801210945d6 100644
--- a/arch/x86/kernel/smp.c
+++ b/arch/x86/kernel/smp.c
@@ -21,6 +21,7 @@
21#include <linux/cache.h> 21#include <linux/cache.h>
22#include <linux/interrupt.h> 22#include <linux/interrupt.h>
23#include <linux/cpu.h> 23#include <linux/cpu.h>
24#include <linux/gfp.h>
24 25
25#include <asm/mtrr.h> 26#include <asm/mtrr.h>
26#include <asm/tlbflush.h> 27#include <asm/tlbflush.h>
diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
index a02e80c3c54b..763d815e27a0 100644
--- a/arch/x86/kernel/smpboot.c
+++ b/arch/x86/kernel/smpboot.c
@@ -49,6 +49,7 @@
49#include <linux/nmi.h> 49#include <linux/nmi.h>
50#include <linux/tboot.h> 50#include <linux/tboot.h>
51#include <linux/stackprotector.h> 51#include <linux/stackprotector.h>
52#include <linux/gfp.h>
52 53
53#include <asm/acpi.h> 54#include <asm/acpi.h>
54#include <asm/desc.h> 55#include <asm/desc.h>
@@ -242,12 +243,10 @@ static void __cpuinit smp_callin(void)
242 end_local_APIC_setup(); 243 end_local_APIC_setup();
243 map_cpu_to_logical_apicid(); 244 map_cpu_to_logical_apicid();
244 245
245 notify_cpu_starting(cpuid);
246
247 /* 246 /*
248 * Need to setup vector mappings before we enable interrupts. 247 * Need to setup vector mappings before we enable interrupts.
249 */ 248 */
250 __setup_vector_irq(smp_processor_id()); 249 setup_vector_irq(smp_processor_id());
251 /* 250 /*
252 * Get our bogomips. 251 * Get our bogomips.
253 * 252 *
@@ -264,6 +263,8 @@ static void __cpuinit smp_callin(void)
264 */ 263 */
265 smp_store_cpu_info(cpuid); 264 smp_store_cpu_info(cpuid);
266 265
266 notify_cpu_starting(cpuid);
267
267 /* 268 /*
268 * Allow the master to continue. 269 * Allow the master to continue.
269 */ 270 */
diff --git a/arch/x86/kernel/tlb_uv.c b/arch/x86/kernel/tlb_uv.c
index 364d015efebc..17b03dd3a6b5 100644
--- a/arch/x86/kernel/tlb_uv.c
+++ b/arch/x86/kernel/tlb_uv.c
@@ -9,6 +9,7 @@
9#include <linux/seq_file.h> 9#include <linux/seq_file.h>
10#include <linux/proc_fs.h> 10#include <linux/proc_fs.h>
11#include <linux/kernel.h> 11#include <linux/kernel.h>
12#include <linux/slab.h>
12 13
13#include <asm/mmu_context.h> 14#include <asm/mmu_context.h>
14#include <asm/uv/uv.h> 15#include <asm/uv/uv.h>
diff --git a/arch/x86/kernel/uv_irq.c b/arch/x86/kernel/uv_irq.c
index ece73d8e3240..1d40336b030a 100644
--- a/arch/x86/kernel/uv_irq.c
+++ b/arch/x86/kernel/uv_irq.c
@@ -10,6 +10,7 @@
10 10
11#include <linux/module.h> 11#include <linux/module.h>
12#include <linux/rbtree.h> 12#include <linux/rbtree.h>
13#include <linux/slab.h>
13#include <linux/irq.h> 14#include <linux/irq.h>
14 15
15#include <asm/apic.h> 16#include <asm/apic.h>
diff --git a/arch/x86/kernel/uv_time.c b/arch/x86/kernel/uv_time.c
index 2b75ef638dbc..56e421bc379b 100644
--- a/arch/x86/kernel/uv_time.c
+++ b/arch/x86/kernel/uv_time.c
@@ -19,6 +19,7 @@
19 * Copyright (c) Dimitri Sivanich 19 * Copyright (c) Dimitri Sivanich
20 */ 20 */
21#include <linux/clockchips.h> 21#include <linux/clockchips.h>
22#include <linux/slab.h>
22 23
23#include <asm/uv/uv_mmrs.h> 24#include <asm/uv/uv_mmrs.h>
24#include <asm/uv/uv_hub.h> 25#include <asm/uv/uv_hub.h>
diff --git a/arch/x86/kernel/vmi_32.c b/arch/x86/kernel/vmi_32.c
index 7dd599deca4a..ce9fbacb7526 100644
--- a/arch/x86/kernel/vmi_32.c
+++ b/arch/x86/kernel/vmi_32.c
@@ -28,6 +28,7 @@
28#include <linux/mm.h> 28#include <linux/mm.h>
29#include <linux/highmem.h> 29#include <linux/highmem.h>
30#include <linux/sched.h> 30#include <linux/sched.h>
31#include <linux/gfp.h>
31#include <asm/vmi.h> 32#include <asm/vmi.h>
32#include <asm/io.h> 33#include <asm/io.h>
33#include <asm/fixmap.h> 34#include <asm/fixmap.h>
diff --git a/arch/x86/kernel/vmlinux.lds.S b/arch/x86/kernel/vmlinux.lds.S
index 44879df55696..2cc249718c46 100644
--- a/arch/x86/kernel/vmlinux.lds.S
+++ b/arch/x86/kernel/vmlinux.lds.S
@@ -291,8 +291,8 @@ SECTIONS
291 .smp_locks : AT(ADDR(.smp_locks) - LOAD_OFFSET) { 291 .smp_locks : AT(ADDR(.smp_locks) - LOAD_OFFSET) {
292 __smp_locks = .; 292 __smp_locks = .;
293 *(.smp_locks) 293 *(.smp_locks)
294 __smp_locks_end = .;
295 . = ALIGN(PAGE_SIZE); 294 . = ALIGN(PAGE_SIZE);
295 __smp_locks_end = .;
296 } 296 }
297 297
298#ifdef CONFIG_X86_64 298#ifdef CONFIG_X86_64
generated by cgit v1.2.2 (git 2.25.0) at 2025-01-12 23:01:53 -0500