aboutsummaryrefslogtreecommitdiffstats
path: root/arch/x86
diff options
context:
space:
mode:
Diffstat (limited to 'arch/x86')
-rw-r--r--arch/x86/Kconfig17
-rw-r--r--arch/x86/Kconfig.cpu24
-rw-r--r--arch/x86/boot/compressed/.gitignore2
-rw-r--r--arch/x86/boot/tty.c2
-rw-r--r--arch/x86/include/asm/acpi.h1
-rw-r--r--arch/x86/include/asm/amd_iommu_types.h24
-rw-r--r--arch/x86/include/asm/cpufeature.h2
-rw-r--r--arch/x86/include/asm/dma-mapping.h10
-rw-r--r--arch/x86/include/asm/ds.h6
-rw-r--r--arch/x86/include/asm/es7000/wakecpu.h9
-rw-r--r--arch/x86/include/asm/fixmap.h4
-rw-r--r--arch/x86/include/asm/fixmap_32.h4
-rw-r--r--arch/x86/include/asm/ftrace.h4
-rw-r--r--arch/x86/include/asm/highmem.h5
-rw-r--r--arch/x86/include/asm/io.h6
-rw-r--r--arch/x86/include/asm/iomap.h30
-rw-r--r--arch/x86/include/asm/iommu.h2
-rw-r--r--arch/x86/include/asm/irq_vectors.h20
-rw-r--r--arch/x86/include/asm/kvm_host.h3
-rw-r--r--arch/x86/include/asm/mach-default/mach_wakecpu.h9
-rw-r--r--arch/x86/include/asm/mmzone_32.h4
-rw-r--r--arch/x86/include/asm/msr.h2
-rw-r--r--arch/x86/include/asm/pci_64.h14
-rw-r--r--arch/x86/include/asm/pgtable-3level.h4
-rw-r--r--arch/x86/include/asm/ptrace.h2
-rw-r--r--arch/x86/include/asm/smp.h6
-rw-r--r--arch/x86/include/asm/topology.h9
-rw-r--r--arch/x86/include/asm/tsc.h8
-rw-r--r--arch/x86/include/asm/uaccess_64.h2
-rw-r--r--arch/x86/include/asm/unistd_64.h4
-rw-r--r--arch/x86/include/asm/uv/uv_hub.h1
-rw-r--r--arch/x86/include/asm/voyager.h1
-rw-r--r--arch/x86/kernel/Makefile5
-rw-r--r--arch/x86/kernel/acpi/boot.c1
-rw-r--r--arch/x86/kernel/amd_iommu.c58
-rw-r--r--arch/x86/kernel/amd_iommu_init.c6
-rw-r--r--arch/x86/kernel/apic.c2
-rw-r--r--arch/x86/kernel/cpu/addon_cpuid_features.c2
-rw-r--r--arch/x86/kernel/cpu/common.c7
-rw-r--r--arch/x86/kernel/cpu/cpufreq/powernow-k8.c18
-rw-r--r--arch/x86/kernel/cpu/cpufreq/powernow-k8.h17
-rw-r--r--arch/x86/kernel/ds.c81
-rw-r--r--arch/x86/kernel/e820.c8
-rw-r--r--arch/x86/kernel/early-quirks.c18
-rw-r--r--arch/x86/kernel/entry_32.S4
-rw-r--r--arch/x86/kernel/entry_64.S4
-rw-r--r--arch/x86/kernel/es7000_32.c9
-rw-r--r--arch/x86/kernel/ftrace.c50
-rw-r--r--arch/x86/kernel/genx2apic_uv_x.c7
-rw-r--r--arch/x86/kernel/hpet.c4
-rw-r--r--arch/x86/kernel/i386_ksyms_32.c2
-rw-r--r--arch/x86/kernel/i387.c2
-rw-r--r--arch/x86/kernel/io_apic.c60
-rw-r--r--arch/x86/kernel/k8.c1
-rw-r--r--arch/x86/kernel/kvmclock.c2
-rw-r--r--arch/x86/kernel/machine_kexec_32.c5
-rw-r--r--arch/x86/kernel/microcode_amd.c2
-rw-r--r--arch/x86/kernel/microcode_core.c4
-rw-r--r--arch/x86/kernel/mpparse.c3
-rw-r--r--arch/x86/kernel/paravirt-spinlocks.c3
-rw-r--r--arch/x86/kernel/pci-calgary_64.c2
-rw-r--r--arch/x86/kernel/pci-dma.c16
-rw-r--r--arch/x86/kernel/pci-gart_64.c4
-rw-r--r--arch/x86/kernel/pci-swiotlb_64.c14
-rw-r--r--arch/x86/kernel/reboot.c15
-rw-r--r--arch/x86/kernel/setup.c2
-rw-r--r--arch/x86/kernel/tlb_32.c6
-rw-r--r--arch/x86/kernel/tlb_64.c5
-rw-r--r--arch/x86/kernel/tsc.c12
-rw-r--r--arch/x86/kernel/tsc_sync.c4
-rw-r--r--arch/x86/kernel/vsmp_64.c2
-rw-r--r--arch/x86/kernel/x8664_ksyms_64.c2
-rw-r--r--arch/x86/kernel/xsave.c2
-rw-r--r--arch/x86/kvm/Kconfig2
-rw-r--r--arch/x86/kvm/i8254.c13
-rw-r--r--arch/x86/kvm/i8254.h1
-rw-r--r--arch/x86/kvm/mmu.c5
-rw-r--r--arch/x86/kvm/paging_tmpl.h1
-rw-r--r--arch/x86/kvm/vmx.c7
-rw-r--r--arch/x86/kvm/vmx.h1
-rw-r--r--arch/x86/kvm/x86.c6
-rw-r--r--arch/x86/lguest/boot.c32
-rw-r--r--arch/x86/mach-voyager/setup.c2
-rw-r--r--arch/x86/mach-voyager/voyager_smp.c30
-rw-r--r--arch/x86/mm/Makefile2
-rw-r--r--arch/x86/mm/gup.c2
-rw-r--r--arch/x86/mm/init_32.c3
-rw-r--r--arch/x86/mm/init_64.c75
-rw-r--r--arch/x86/mm/iomap_32.c59
-rw-r--r--arch/x86/mm/ioremap.c22
-rw-r--r--arch/x86/mm/numa_32.c35
-rw-r--r--arch/x86/mm/pageattr.c8
-rw-r--r--arch/x86/mm/pat.c4
-rw-r--r--arch/x86/oprofile/nmi_int.c5
-rw-r--r--arch/x86/oprofile/op_model_ppro.c15
-rw-r--r--arch/x86/pci/fixup.c25
-rw-r--r--arch/x86/power/hibernate_32.c4
-rw-r--r--arch/x86/xen/Makefile2
-rw-r--r--arch/x86/xen/enlighten.c5
-rw-r--r--arch/x86/xen/mmu.c50
-rw-r--r--arch/x86/xen/smp.c2
-rw-r--r--arch/x86/xen/xen-ops.h2
102 files changed, 733 insertions, 399 deletions
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index bacac556b189..e795b5ba25b2 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -28,7 +28,7 @@ config X86
28 select HAVE_KRETPROBES 28 select HAVE_KRETPROBES
29 select HAVE_FTRACE_MCOUNT_RECORD 29 select HAVE_FTRACE_MCOUNT_RECORD
30 select HAVE_DYNAMIC_FTRACE 30 select HAVE_DYNAMIC_FTRACE
31 select HAVE_FTRACE 31 select HAVE_FUNCTION_TRACER
32 select HAVE_KVM if ((X86_32 && !X86_VOYAGER && !X86_VISWS && !X86_NUMAQ) || X86_64) 32 select HAVE_KVM if ((X86_32 && !X86_VOYAGER && !X86_VISWS && !X86_NUMAQ) || X86_64)
33 select HAVE_ARCH_KGDB if !X86_VOYAGER 33 select HAVE_ARCH_KGDB if !X86_VOYAGER
34 select HAVE_ARCH_TRACEHOOK 34 select HAVE_ARCH_TRACEHOOK
@@ -167,9 +167,12 @@ config GENERIC_PENDING_IRQ
167config X86_SMP 167config X86_SMP
168 bool 168 bool
169 depends on SMP && ((X86_32 && !X86_VOYAGER) || X86_64) 169 depends on SMP && ((X86_32 && !X86_VOYAGER) || X86_64)
170 select USE_GENERIC_SMP_HELPERS
171 default y 170 default y
172 171
172config USE_GENERIC_SMP_HELPERS
173 def_bool y
174 depends on SMP
175
173config X86_32_SMP 176config X86_32_SMP
174 def_bool y 177 def_bool y
175 depends on X86_32 && SMP 178 depends on X86_32 && SMP
@@ -231,6 +234,10 @@ config SMP
231 234
232 If you don't know what to do here, say N. 235 If you don't know what to do here, say N.
233 236
237config X86_HAS_BOOT_CPU_ID
238 def_bool y
239 depends on X86_VOYAGER
240
234config X86_FIND_SMP_CONFIG 241config X86_FIND_SMP_CONFIG
235 def_bool y 242 def_bool y
236 depends on X86_MPPARSE || X86_VOYAGER 243 depends on X86_MPPARSE || X86_VOYAGER
@@ -1494,7 +1501,7 @@ config HAVE_ARCH_EARLY_PFN_TO_NID
1494 def_bool X86_64 1501 def_bool X86_64
1495 depends on NUMA 1502 depends on NUMA
1496 1503
1497menu "Power management options" 1504menu "Power management and ACPI options"
1498 depends on !X86_VOYAGER 1505 depends on !X86_VOYAGER
1499 1506
1500config ARCH_HIBERNATION_HEADER 1507config ARCH_HIBERNATION_HEADER
@@ -1894,6 +1901,10 @@ config SYSVIPC_COMPAT
1894endmenu 1901endmenu
1895 1902
1896 1903
1904config HAVE_ATOMIC_IOMAP
1905 def_bool y
1906 depends on X86_32
1907
1897source "net/Kconfig" 1908source "net/Kconfig"
1898 1909
1899source "drivers/Kconfig" 1910source "drivers/Kconfig"
diff --git a/arch/x86/Kconfig.cpu b/arch/x86/Kconfig.cpu
index 0b7c4a3f0651..b815664fe370 100644
--- a/arch/x86/Kconfig.cpu
+++ b/arch/x86/Kconfig.cpu
@@ -513,19 +513,19 @@ config CPU_SUP_UMC_32
513 If unsure, say N. 513 If unsure, say N.
514 514
515config X86_DS 515config X86_DS
516 bool "Debug Store support" 516 def_bool X86_PTRACE_BTS
517 default y 517 depends on X86_DEBUGCTLMSR
518 help
519 Add support for Debug Store.
520 This allows the kernel to provide a memory buffer to the hardware
521 to store various profiling and tracing events.
522 518
523config X86_PTRACE_BTS 519config X86_PTRACE_BTS
524 bool "ptrace interface to Branch Trace Store" 520 bool "Branch Trace Store"
525 default y 521 default y
526 depends on (X86_DS && X86_DEBUGCTLMSR) 522 depends on X86_DEBUGCTLMSR
527 help 523 help
528 Add a ptrace interface to allow collecting an execution trace 524 This adds a ptrace interface to the hardware's branch trace store.
529 of the traced task. 525
530 This collects control flow changes in a (cyclic) buffer and allows 526 Debuggers may use it to collect an execution trace of the debugged
531 debuggers to fill in the gaps and show an execution trace of the debuggee. 527 application in order to answer the question 'how did I get here?'.
528 Debuggers may trace user mode as well as kernel mode.
529
530 Say Y unless there is no application development on this machine
531 and you want to save a small amount of code size.
diff --git a/arch/x86/boot/compressed/.gitignore b/arch/x86/boot/compressed/.gitignore
index be0ed065249b..63eff3b04d01 100644
--- a/arch/x86/boot/compressed/.gitignore
+++ b/arch/x86/boot/compressed/.gitignore
@@ -1 +1,3 @@
1relocs 1relocs
2vmlinux.bin.all
3vmlinux.relocs
diff --git a/arch/x86/boot/tty.c b/arch/x86/boot/tty.c
index 0be77b39328a..7e8e8b25f5f6 100644
--- a/arch/x86/boot/tty.c
+++ b/arch/x86/boot/tty.c
@@ -74,7 +74,7 @@ static int kbd_pending(void)
74{ 74{
75 u8 pending; 75 u8 pending;
76 asm volatile("int $0x16; setnz %0" 76 asm volatile("int $0x16; setnz %0"
77 : "=rm" (pending) 77 : "=qm" (pending)
78 : "a" (0x0100)); 78 : "a" (0x0100));
79 return pending; 79 return pending;
80} 80}
diff --git a/arch/x86/include/asm/acpi.h b/arch/x86/include/asm/acpi.h
index 8d676d8ecde9..9830681446ad 100644
--- a/arch/x86/include/asm/acpi.h
+++ b/arch/x86/include/asm/acpi.h
@@ -113,7 +113,6 @@ static inline void acpi_disable_pci(void)
113 acpi_pci_disabled = 1; 113 acpi_pci_disabled = 1;
114 acpi_noirq_set(); 114 acpi_noirq_set();
115} 115}
116extern int acpi_irq_balance_set(char *str);
117 116
118/* routines for saving/restoring kernel state */ 117/* routines for saving/restoring kernel state */
119extern int acpi_save_state_mem(void); 118extern int acpi_save_state_mem(void);
diff --git a/arch/x86/include/asm/amd_iommu_types.h b/arch/x86/include/asm/amd_iommu_types.h
index 1a30c0440c6b..ac302a2fa339 100644
--- a/arch/x86/include/asm/amd_iommu_types.h
+++ b/arch/x86/include/asm/amd_iommu_types.h
@@ -251,13 +251,6 @@ struct amd_iommu {
251 /* Pointer to PCI device of this IOMMU */ 251 /* Pointer to PCI device of this IOMMU */
252 struct pci_dev *dev; 252 struct pci_dev *dev;
253 253
254 /*
255 * Capability pointer. There could be more than one IOMMU per PCI
256 * device function if there are more than one AMD IOMMU capability
257 * pointers.
258 */
259 u16 cap_ptr;
260
261 /* physical address of MMIO space */ 254 /* physical address of MMIO space */
262 u64 mmio_phys; 255 u64 mmio_phys;
263 /* virtual address of MMIO space */ 256 /* virtual address of MMIO space */
@@ -266,6 +259,13 @@ struct amd_iommu {
266 /* capabilities of that IOMMU read from ACPI */ 259 /* capabilities of that IOMMU read from ACPI */
267 u32 cap; 260 u32 cap;
268 261
262 /*
263 * Capability pointer. There could be more than one IOMMU per PCI
264 * device function if there are more than one AMD IOMMU capability
265 * pointers.
266 */
267 u16 cap_ptr;
268
269 /* pci domain of this IOMMU */ 269 /* pci domain of this IOMMU */
270 u16 pci_seg; 270 u16 pci_seg;
271 271
@@ -284,19 +284,19 @@ struct amd_iommu {
284 /* size of command buffer */ 284 /* size of command buffer */
285 u32 cmd_buf_size; 285 u32 cmd_buf_size;
286 286
287 /* event buffer virtual address */
288 u8 *evt_buf;
289 /* size of event buffer */ 287 /* size of event buffer */
290 u32 evt_buf_size; 288 u32 evt_buf_size;
289 /* event buffer virtual address */
290 u8 *evt_buf;
291 /* MSI number for event interrupt */ 291 /* MSI number for event interrupt */
292 u16 evt_msi_num; 292 u16 evt_msi_num;
293 293
294 /* if one, we need to send a completion wait command */
295 int need_sync;
296
297 /* true if interrupts for this IOMMU are already enabled */ 294 /* true if interrupts for this IOMMU are already enabled */
298 bool int_enabled; 295 bool int_enabled;
299 296
297 /* if one, we need to send a completion wait command */
298 int need_sync;
299
300 /* default dma_ops domain for that IOMMU */ 300 /* default dma_ops domain for that IOMMU */
301 struct dma_ops_domain *default_dom; 301 struct dma_ops_domain *default_dom;
302}; 302};
diff --git a/arch/x86/include/asm/cpufeature.h b/arch/x86/include/asm/cpufeature.h
index f73e95d75b45..cfdf8c2c5c31 100644
--- a/arch/x86/include/asm/cpufeature.h
+++ b/arch/x86/include/asm/cpufeature.h
@@ -91,7 +91,7 @@
91#define X86_FEATURE_11AP (3*32+19) /* "" Bad local APIC aka 11AP */ 91#define X86_FEATURE_11AP (3*32+19) /* "" Bad local APIC aka 11AP */
92#define X86_FEATURE_NOPL (3*32+20) /* The NOPL (0F 1F) instructions */ 92#define X86_FEATURE_NOPL (3*32+20) /* The NOPL (0F 1F) instructions */
93#define X86_FEATURE_AMDC1E (3*32+21) /* AMD C1E detected */ 93#define X86_FEATURE_AMDC1E (3*32+21) /* AMD C1E detected */
94#define X86_FEATURE_XTOPOLOGY (3*32+21) /* cpu topology enum extensions */ 94#define X86_FEATURE_XTOPOLOGY (3*32+22) /* cpu topology enum extensions */
95 95
96/* Intel-defined CPU features, CPUID level 0x00000001 (ecx), word 4 */ 96/* Intel-defined CPU features, CPUID level 0x00000001 (ecx), word 4 */
97#define X86_FEATURE_XMM3 (4*32+ 0) /* "pni" SSE-3 */ 97#define X86_FEATURE_XMM3 (4*32+ 0) /* "pni" SSE-3 */
diff --git a/arch/x86/include/asm/dma-mapping.h b/arch/x86/include/asm/dma-mapping.h
index 4a5397bfce27..097794ff6b79 100644
--- a/arch/x86/include/asm/dma-mapping.h
+++ b/arch/x86/include/asm/dma-mapping.h
@@ -71,15 +71,13 @@ static inline struct dma_mapping_ops *get_dma_ops(struct device *dev)
71/* Make sure we keep the same behaviour */ 71/* Make sure we keep the same behaviour */
72static inline int dma_mapping_error(struct device *dev, dma_addr_t dma_addr) 72static inline int dma_mapping_error(struct device *dev, dma_addr_t dma_addr)
73{ 73{
74#ifdef CONFIG_X86_32 74#ifdef CONFIG_X86_64
75 return 0;
76#else
77 struct dma_mapping_ops *ops = get_dma_ops(dev); 75 struct dma_mapping_ops *ops = get_dma_ops(dev);
78 if (ops->mapping_error) 76 if (ops->mapping_error)
79 return ops->mapping_error(dev, dma_addr); 77 return ops->mapping_error(dev, dma_addr);
80 78
81 return (dma_addr == bad_dma_address);
82#endif 79#endif
80 return (dma_addr == bad_dma_address);
83} 81}
84 82
85#define dma_alloc_noncoherent(d, s, h, f) dma_alloc_coherent(d, s, h, f) 83#define dma_alloc_noncoherent(d, s, h, f) dma_alloc_coherent(d, s, h, f)
@@ -255,9 +253,11 @@ static inline unsigned long dma_alloc_coherent_mask(struct device *dev,
255 253
256static inline gfp_t dma_alloc_coherent_gfp_flags(struct device *dev, gfp_t gfp) 254static inline gfp_t dma_alloc_coherent_gfp_flags(struct device *dev, gfp_t gfp)
257{ 255{
258#ifdef CONFIG_X86_64
259 unsigned long dma_mask = dma_alloc_coherent_mask(dev, gfp); 256 unsigned long dma_mask = dma_alloc_coherent_mask(dev, gfp);
260 257
258 if (dma_mask <= DMA_24BIT_MASK)
259 gfp |= GFP_DMA;
260#ifdef CONFIG_X86_64
261 if (dma_mask <= DMA_32BIT_MASK && !(gfp & GFP_DMA)) 261 if (dma_mask <= DMA_32BIT_MASK && !(gfp & GFP_DMA))
262 gfp |= GFP_DMA32; 262 gfp |= GFP_DMA32;
263#endif 263#endif
diff --git a/arch/x86/include/asm/ds.h b/arch/x86/include/asm/ds.h
index 72c5a190bf48..a95008457ea4 100644
--- a/arch/x86/include/asm/ds.h
+++ b/arch/x86/include/asm/ds.h
@@ -23,12 +23,13 @@
23#ifndef _ASM_X86_DS_H 23#ifndef _ASM_X86_DS_H
24#define _ASM_X86_DS_H 24#define _ASM_X86_DS_H
25 25
26#ifdef CONFIG_X86_DS
27 26
28#include <linux/types.h> 27#include <linux/types.h>
29#include <linux/init.h> 28#include <linux/init.h>
30 29
31 30
31#ifdef CONFIG_X86_DS
32
32struct task_struct; 33struct task_struct;
33 34
34/* 35/*
@@ -232,7 +233,8 @@ extern void ds_free(struct ds_context *context);
232 233
233#else /* CONFIG_X86_DS */ 234#else /* CONFIG_X86_DS */
234 235
235#define ds_init_intel(config) do {} while (0) 236struct cpuinfo_x86;
237static inline void __cpuinit ds_init_intel(struct cpuinfo_x86 *ignored) {}
236 238
237#endif /* CONFIG_X86_DS */ 239#endif /* CONFIG_X86_DS */
238#endif /* _ASM_X86_DS_H */ 240#endif /* _ASM_X86_DS_H */
diff --git a/arch/x86/include/asm/es7000/wakecpu.h b/arch/x86/include/asm/es7000/wakecpu.h
index 3ffc5a7bf667..398493461913 100644
--- a/arch/x86/include/asm/es7000/wakecpu.h
+++ b/arch/x86/include/asm/es7000/wakecpu.h
@@ -50,10 +50,9 @@ static inline void restore_NMI_vector(unsigned short *high, unsigned short *low)
50{ 50{
51} 51}
52 52
53#if APIC_DEBUG 53#define inquire_remote_apic(apicid) do { \
54 #define inquire_remote_apic(apicid) __inquire_remote_apic(apicid) 54 if (apic_verbosity >= APIC_DEBUG) \
55#else 55 __inquire_remote_apic(apicid); \
56 #define inquire_remote_apic(apicid) {} 56 } while (0)
57#endif
58 57
59#endif /* __ASM_MACH_WAKECPU_H */ 58#endif /* __ASM_MACH_WAKECPU_H */
diff --git a/arch/x86/include/asm/fixmap.h b/arch/x86/include/asm/fixmap.h
index 8668a94f850e..23696d44a0af 100644
--- a/arch/x86/include/asm/fixmap.h
+++ b/arch/x86/include/asm/fixmap.h
@@ -9,6 +9,10 @@
9 9
10extern int fixmaps_set; 10extern int fixmaps_set;
11 11
12extern pte_t *kmap_pte;
13extern pgprot_t kmap_prot;
14extern pte_t *pkmap_page_table;
15
12void __native_set_fixmap(enum fixed_addresses idx, pte_t pte); 16void __native_set_fixmap(enum fixed_addresses idx, pte_t pte);
13void native_set_fixmap(enum fixed_addresses idx, 17void native_set_fixmap(enum fixed_addresses idx,
14 unsigned long phys, pgprot_t flags); 18 unsigned long phys, pgprot_t flags);
diff --git a/arch/x86/include/asm/fixmap_32.h b/arch/x86/include/asm/fixmap_32.h
index 09f29ab5c139..c7115c1d7217 100644
--- a/arch/x86/include/asm/fixmap_32.h
+++ b/arch/x86/include/asm/fixmap_32.h
@@ -28,10 +28,8 @@ extern unsigned long __FIXADDR_TOP;
28#include <asm/acpi.h> 28#include <asm/acpi.h>
29#include <asm/apicdef.h> 29#include <asm/apicdef.h>
30#include <asm/page.h> 30#include <asm/page.h>
31#ifdef CONFIG_HIGHMEM
32#include <linux/threads.h> 31#include <linux/threads.h>
33#include <asm/kmap_types.h> 32#include <asm/kmap_types.h>
34#endif
35 33
36/* 34/*
37 * Here we define all the compile-time 'special' virtual 35 * Here we define all the compile-time 'special' virtual
@@ -75,10 +73,8 @@ enum fixed_addresses {
75#ifdef CONFIG_X86_CYCLONE_TIMER 73#ifdef CONFIG_X86_CYCLONE_TIMER
76 FIX_CYCLONE_TIMER, /*cyclone timer register*/ 74 FIX_CYCLONE_TIMER, /*cyclone timer register*/
77#endif 75#endif
78#ifdef CONFIG_HIGHMEM
79 FIX_KMAP_BEGIN, /* reserved pte's for temporary kernel mappings */ 76 FIX_KMAP_BEGIN, /* reserved pte's for temporary kernel mappings */
80 FIX_KMAP_END = FIX_KMAP_BEGIN+(KM_TYPE_NR*NR_CPUS)-1, 77 FIX_KMAP_END = FIX_KMAP_BEGIN+(KM_TYPE_NR*NR_CPUS)-1,
81#endif
82#ifdef CONFIG_PCI_MMCONFIG 78#ifdef CONFIG_PCI_MMCONFIG
83 FIX_PCIE_MCFG, 79 FIX_PCIE_MCFG,
84#endif 80#endif
diff --git a/arch/x86/include/asm/ftrace.h b/arch/x86/include/asm/ftrace.h
index 47f7e65e6c1d..9e8bc29b8b17 100644
--- a/arch/x86/include/asm/ftrace.h
+++ b/arch/x86/include/asm/ftrace.h
@@ -1,7 +1,7 @@
1#ifndef _ASM_X86_FTRACE_H 1#ifndef _ASM_X86_FTRACE_H
2#define _ASM_X86_FTRACE_H 2#define _ASM_X86_FTRACE_H
3 3
4#ifdef CONFIG_FTRACE 4#ifdef CONFIG_FUNCTION_TRACER
5#define MCOUNT_ADDR ((long)(mcount)) 5#define MCOUNT_ADDR ((long)(mcount))
6#define MCOUNT_INSN_SIZE 5 /* sizeof mcount call */ 6#define MCOUNT_INSN_SIZE 5 /* sizeof mcount call */
7 7
@@ -19,6 +19,6 @@ static inline unsigned long ftrace_call_adjust(unsigned long addr)
19} 19}
20#endif 20#endif
21 21
22#endif /* CONFIG_FTRACE */ 22#endif /* CONFIG_FUNCTION_TRACER */
23 23
24#endif /* _ASM_X86_FTRACE_H */ 24#endif /* _ASM_X86_FTRACE_H */
diff --git a/arch/x86/include/asm/highmem.h b/arch/x86/include/asm/highmem.h
index a3b3b7c3027b..bf9276bea660 100644
--- a/arch/x86/include/asm/highmem.h
+++ b/arch/x86/include/asm/highmem.h
@@ -25,14 +25,11 @@
25#include <asm/kmap_types.h> 25#include <asm/kmap_types.h>
26#include <asm/tlbflush.h> 26#include <asm/tlbflush.h>
27#include <asm/paravirt.h> 27#include <asm/paravirt.h>
28#include <asm/fixmap.h>
28 29
29/* declarations for highmem.c */ 30/* declarations for highmem.c */
30extern unsigned long highstart_pfn, highend_pfn; 31extern unsigned long highstart_pfn, highend_pfn;
31 32
32extern pte_t *kmap_pte;
33extern pgprot_t kmap_prot;
34extern pte_t *pkmap_page_table;
35
36/* 33/*
37 * Right now we initialize only a single pte table. It can be extended 34 * Right now we initialize only a single pte table. It can be extended
38 * easily, subsequent pte tables have to be allocated in one physical 35 * easily, subsequent pte tables have to be allocated in one physical
diff --git a/arch/x86/include/asm/io.h b/arch/x86/include/asm/io.h
index 5618a103f395..ac2abc88cd95 100644
--- a/arch/x86/include/asm/io.h
+++ b/arch/x86/include/asm/io.h
@@ -82,9 +82,9 @@ extern void __iomem *ioremap_wc(unsigned long offset, unsigned long size);
82extern void early_ioremap_init(void); 82extern void early_ioremap_init(void);
83extern void early_ioremap_clear(void); 83extern void early_ioremap_clear(void);
84extern void early_ioremap_reset(void); 84extern void early_ioremap_reset(void);
85extern void *early_ioremap(unsigned long offset, unsigned long size); 85extern void __iomem *early_ioremap(unsigned long offset, unsigned long size);
86extern void *early_memremap(unsigned long offset, unsigned long size); 86extern void __iomem *early_memremap(unsigned long offset, unsigned long size);
87extern void early_iounmap(void *addr, unsigned long size); 87extern void early_iounmap(void __iomem *addr, unsigned long size);
88extern void __iomem *fix_ioremap(unsigned idx, unsigned long phys); 88extern void __iomem *fix_ioremap(unsigned idx, unsigned long phys);
89 89
90 90
diff --git a/arch/x86/include/asm/iomap.h b/arch/x86/include/asm/iomap.h
new file mode 100644
index 000000000000..c1f06289b14b
--- /dev/null
+++ b/arch/x86/include/asm/iomap.h
@@ -0,0 +1,30 @@
1/*
2 * Copyright © 2008 Ingo Molnar
3 *
4 * This program is free software; you can redistribute it and/or modify
5 * it under the terms of the GNU General Public License as published by
6 * the Free Software Foundation; either version 2 of the License, or
7 * (at your option) any later version.
8 *
9 * This program is distributed in the hope that it will be useful, but
10 * WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 * General Public License for more details.
13 *
14 * You should have received a copy of the GNU General Public License along
15 * with this program; if not, write to the Free Software Foundation, Inc.,
16 * 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
17 */
18
19#include <linux/fs.h>
20#include <linux/mm.h>
21#include <linux/uaccess.h>
22#include <asm/cacheflush.h>
23#include <asm/pgtable.h>
24#include <asm/tlbflush.h>
25
26void *
27iomap_atomic_prot_pfn(unsigned long pfn, enum km_type type, pgprot_t prot);
28
29void
30iounmap_atomic(void *kvaddr, enum km_type type);
diff --git a/arch/x86/include/asm/iommu.h b/arch/x86/include/asm/iommu.h
index 98e28ea8cd16..0b500c5b6446 100644
--- a/arch/x86/include/asm/iommu.h
+++ b/arch/x86/include/asm/iommu.h
@@ -6,8 +6,6 @@ extern void no_iommu_init(void);
6extern struct dma_mapping_ops nommu_dma_ops; 6extern struct dma_mapping_ops nommu_dma_ops;
7extern int force_iommu, no_iommu; 7extern int force_iommu, no_iommu;
8extern int iommu_detected; 8extern int iommu_detected;
9extern int dmar_disabled;
10extern int forbid_dac;
11 9
12extern unsigned long iommu_nr_pages(unsigned long addr, unsigned long len); 10extern unsigned long iommu_nr_pages(unsigned long addr, unsigned long len);
13 11
diff --git a/arch/x86/include/asm/irq_vectors.h b/arch/x86/include/asm/irq_vectors.h
index d843ed0e9b2e..0005adb0f941 100644
--- a/arch/x86/include/asm/irq_vectors.h
+++ b/arch/x86/include/asm/irq_vectors.h
@@ -101,30 +101,22 @@
101#define LAST_VM86_IRQ 15 101#define LAST_VM86_IRQ 15
102#define invalid_vm86_irq(irq) ((irq) < 3 || (irq) > 15) 102#define invalid_vm86_irq(irq) ((irq) < 3 || (irq) > 15)
103 103
104#ifdef CONFIG_X86_64 104#if defined(CONFIG_X86_IO_APIC) && !defined(CONFIG_X86_VOYAGER)
105# if NR_CPUS < MAX_IO_APICS 105# if NR_CPUS < MAX_IO_APICS
106# define NR_IRQS (NR_VECTORS + (32 * NR_CPUS)) 106# define NR_IRQS (NR_VECTORS + (32 * NR_CPUS))
107# else 107# else
108# define NR_IRQS (NR_VECTORS + (32 * MAX_IO_APICS)) 108# define NR_IRQS (NR_VECTORS + (32 * MAX_IO_APICS))
109# endif 109# endif
110 110
111#elif !defined(CONFIG_X86_VOYAGER) 111#elif defined(CONFIG_X86_VOYAGER)
112 112
113# if defined(CONFIG_X86_IO_APIC) || defined(CONFIG_PARAVIRT) || defined(CONFIG_X86_VISWS) 113# define NR_IRQS 224
114
115# define NR_IRQS 224
116
117# else /* IO_APIC || PARAVIRT */
118
119# define NR_IRQS 16
120
121# endif
122 114
123#else /* !VISWS && !VOYAGER */ 115#else /* IO_APIC || VOYAGER */
124 116
125# define NR_IRQS 224 117# define NR_IRQS 16
126 118
127#endif /* VISWS */ 119#endif
128 120
129/* Voyager specific defines */ 121/* Voyager specific defines */
130/* These define the CPIs we use in linux */ 122/* These define the CPIs we use in linux */
diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index 65679d006337..8346be87cfa1 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -364,6 +364,9 @@ struct kvm_arch{
364 364
365 struct page *ept_identity_pagetable; 365 struct page *ept_identity_pagetable;
366 bool ept_identity_pagetable_done; 366 bool ept_identity_pagetable_done;
367
368 unsigned long irq_sources_bitmap;
369 unsigned long irq_states[KVM_IOAPIC_NUM_PINS];
367}; 370};
368 371
369struct kvm_vm_stat { 372struct kvm_vm_stat {
diff --git a/arch/x86/include/asm/mach-default/mach_wakecpu.h b/arch/x86/include/asm/mach-default/mach_wakecpu.h
index d5c0b826a4ff..9d80db91e992 100644
--- a/arch/x86/include/asm/mach-default/mach_wakecpu.h
+++ b/arch/x86/include/asm/mach-default/mach_wakecpu.h
@@ -33,10 +33,9 @@ static inline void restore_NMI_vector(unsigned short *high, unsigned short *low)
33{ 33{
34} 34}
35 35
36#if APIC_DEBUG 36#define inquire_remote_apic(apicid) do { \
37 #define inquire_remote_apic(apicid) __inquire_remote_apic(apicid) 37 if (apic_verbosity >= APIC_DEBUG) \
38#else 38 __inquire_remote_apic(apicid); \
39 #define inquire_remote_apic(apicid) {} 39 } while (0)
40#endif
41 40
42#endif /* _ASM_X86_MACH_DEFAULT_MACH_WAKECPU_H */ 41#endif /* _ASM_X86_MACH_DEFAULT_MACH_WAKECPU_H */
diff --git a/arch/x86/include/asm/mmzone_32.h b/arch/x86/include/asm/mmzone_32.h
index 485bdf059ffb..07f1af494ca5 100644
--- a/arch/x86/include/asm/mmzone_32.h
+++ b/arch/x86/include/asm/mmzone_32.h
@@ -34,10 +34,14 @@ static inline void get_memcfg_numa(void)
34 34
35extern int early_pfn_to_nid(unsigned long pfn); 35extern int early_pfn_to_nid(unsigned long pfn);
36 36
37extern void resume_map_numa_kva(pgd_t *pgd);
38
37#else /* !CONFIG_NUMA */ 39#else /* !CONFIG_NUMA */
38 40
39#define get_memcfg_numa get_memcfg_numa_flat 41#define get_memcfg_numa get_memcfg_numa_flat
40 42
43static inline void resume_map_numa_kva(pgd_t *pgd) {}
44
41#endif /* CONFIG_NUMA */ 45#endif /* CONFIG_NUMA */
42 46
43#ifdef CONFIG_DISCONTIGMEM 47#ifdef CONFIG_DISCONTIGMEM
diff --git a/arch/x86/include/asm/msr.h b/arch/x86/include/asm/msr.h
index 478a9245aae1..397efa375581 100644
--- a/arch/x86/include/asm/msr.h
+++ b/arch/x86/include/asm/msr.h
@@ -108,9 +108,7 @@ static __always_inline unsigned long long __native_read_tsc(void)
108{ 108{
109 DECLARE_ARGS(val, low, high); 109 DECLARE_ARGS(val, low, high);
110 110
111 rdtsc_barrier();
112 asm volatile("rdtsc" : EAX_EDX_RET(val, low, high)); 111 asm volatile("rdtsc" : EAX_EDX_RET(val, low, high));
113 rdtsc_barrier();
114 112
115 return EAX_EDX_VAL(val, low, high); 113 return EAX_EDX_VAL(val, low, high);
116} 114}
diff --git a/arch/x86/include/asm/pci_64.h b/arch/x86/include/asm/pci_64.h
index 5b28995d664e..d02d936840a3 100644
--- a/arch/x86/include/asm/pci_64.h
+++ b/arch/x86/include/asm/pci_64.h
@@ -34,8 +34,6 @@ extern void pci_iommu_alloc(void);
34 */ 34 */
35#define PCI_DMA_BUS_IS_PHYS (dma_ops->is_phys) 35#define PCI_DMA_BUS_IS_PHYS (dma_ops->is_phys)
36 36
37#if defined(CONFIG_GART_IOMMU) || defined(CONFIG_CALGARY_IOMMU)
38
39#define DECLARE_PCI_UNMAP_ADDR(ADDR_NAME) \ 37#define DECLARE_PCI_UNMAP_ADDR(ADDR_NAME) \
40 dma_addr_t ADDR_NAME; 38 dma_addr_t ADDR_NAME;
41#define DECLARE_PCI_UNMAP_LEN(LEN_NAME) \ 39#define DECLARE_PCI_UNMAP_LEN(LEN_NAME) \
@@ -49,18 +47,6 @@ extern void pci_iommu_alloc(void);
49#define pci_unmap_len_set(PTR, LEN_NAME, VAL) \ 47#define pci_unmap_len_set(PTR, LEN_NAME, VAL) \
50 (((PTR)->LEN_NAME) = (VAL)) 48 (((PTR)->LEN_NAME) = (VAL))
51 49
52#else
53/* No IOMMU */
54
55#define DECLARE_PCI_UNMAP_ADDR(ADDR_NAME)
56#define DECLARE_PCI_UNMAP_LEN(LEN_NAME)
57#define pci_unmap_addr(PTR, ADDR_NAME) (0)
58#define pci_unmap_addr_set(PTR, ADDR_NAME, VAL) do { } while (0)
59#define pci_unmap_len(PTR, LEN_NAME) (0)
60#define pci_unmap_len_set(PTR, LEN_NAME, VAL) do { } while (0)
61
62#endif
63
64#endif /* __KERNEL__ */ 50#endif /* __KERNEL__ */
65 51
66#endif /* _ASM_X86_PCI_64_H */ 52#endif /* _ASM_X86_PCI_64_H */
diff --git a/arch/x86/include/asm/pgtable-3level.h b/arch/x86/include/asm/pgtable-3level.h
index fb16cec702e4..52597aeadfff 100644
--- a/arch/x86/include/asm/pgtable-3level.h
+++ b/arch/x86/include/asm/pgtable-3level.h
@@ -120,13 +120,13 @@ static inline void pud_clear(pud_t *pudp)
120 write_cr3(pgd); 120 write_cr3(pgd);
121} 121}
122 122
123#define pud_page(pud) ((struct page *) __va(pud_val(pud) & PTE_PFN_MASK)) 123#define pud_page(pud) pfn_to_page(pud_val(pud) >> PAGE_SHIFT)
124 124
125#define pud_page_vaddr(pud) ((unsigned long) __va(pud_val(pud) & PTE_PFN_MASK)) 125#define pud_page_vaddr(pud) ((unsigned long) __va(pud_val(pud) & PTE_PFN_MASK))
126 126
127 127
128/* Find an entry in the second-level page table.. */ 128/* Find an entry in the second-level page table.. */
129#define pmd_offset(pud, address) ((pmd_t *)pud_page(*(pud)) + \ 129#define pmd_offset(pud, address) ((pmd_t *)pud_page_vaddr(*(pud)) + \
130 pmd_index(address)) 130 pmd_index(address))
131 131
132#ifdef CONFIG_SMP 132#ifdef CONFIG_SMP
diff --git a/arch/x86/include/asm/ptrace.h b/arch/x86/include/asm/ptrace.h
index d1531c8480b7..eefb0594b058 100644
--- a/arch/x86/include/asm/ptrace.h
+++ b/arch/x86/include/asm/ptrace.h
@@ -271,8 +271,6 @@ extern int do_get_thread_area(struct task_struct *p, int idx,
271extern int do_set_thread_area(struct task_struct *p, int idx, 271extern int do_set_thread_area(struct task_struct *p, int idx,
272 struct user_desc __user *info, int can_allocate); 272 struct user_desc __user *info, int can_allocate);
273 273
274#define __ARCH_WANT_COMPAT_SYS_PTRACE
275
276#endif /* __KERNEL__ */ 274#endif /* __KERNEL__ */
277 275
278#endif /* !__ASSEMBLY__ */ 276#endif /* !__ASSEMBLY__ */
diff --git a/arch/x86/include/asm/smp.h b/arch/x86/include/asm/smp.h
index 2766021aef80..d12811ce51d9 100644
--- a/arch/x86/include/asm/smp.h
+++ b/arch/x86/include/asm/smp.h
@@ -225,5 +225,11 @@ static inline int hard_smp_processor_id(void)
225 225
226#endif /* CONFIG_X86_LOCAL_APIC */ 226#endif /* CONFIG_X86_LOCAL_APIC */
227 227
228#ifdef CONFIG_X86_HAS_BOOT_CPU_ID
229extern unsigned char boot_cpu_id;
230#else
231#define boot_cpu_id 0
232#endif
233
228#endif /* __ASSEMBLY__ */ 234#endif /* __ASSEMBLY__ */
229#endif /* _ASM_X86_SMP_H */ 235#endif /* _ASM_X86_SMP_H */
diff --git a/arch/x86/include/asm/topology.h b/arch/x86/include/asm/topology.h
index 90ac7718469a..ff386ff50ed7 100644
--- a/arch/x86/include/asm/topology.h
+++ b/arch/x86/include/asm/topology.h
@@ -154,7 +154,7 @@ extern unsigned long node_remap_size[];
154 154
155#endif 155#endif
156 156
157/* sched_domains SD_NODE_INIT for NUMAQ machines */ 157/* sched_domains SD_NODE_INIT for NUMA machines */
158#define SD_NODE_INIT (struct sched_domain) { \ 158#define SD_NODE_INIT (struct sched_domain) { \
159 .min_interval = 8, \ 159 .min_interval = 8, \
160 .max_interval = 32, \ 160 .max_interval = 32, \
@@ -169,8 +169,9 @@ extern unsigned long node_remap_size[];
169 .flags = SD_LOAD_BALANCE \ 169 .flags = SD_LOAD_BALANCE \
170 | SD_BALANCE_EXEC \ 170 | SD_BALANCE_EXEC \
171 | SD_BALANCE_FORK \ 171 | SD_BALANCE_FORK \
172 | SD_SERIALIZE \ 172 | SD_WAKE_AFFINE \
173 | SD_WAKE_BALANCE, \ 173 | SD_WAKE_BALANCE \
174 | SD_SERIALIZE, \
174 .last_balance = jiffies, \ 175 .last_balance = jiffies, \
175 .balance_interval = 1, \ 176 .balance_interval = 1, \
176} 177}
@@ -238,7 +239,7 @@ struct pci_bus;
238void set_pci_bus_resources_arch_default(struct pci_bus *b); 239void set_pci_bus_resources_arch_default(struct pci_bus *b);
239 240
240#ifdef CONFIG_SMP 241#ifdef CONFIG_SMP
241#define mc_capable() (boot_cpu_data.x86_max_cores > 1) 242#define mc_capable() (cpus_weight(per_cpu(cpu_core_map, 0)) != nr_cpu_ids)
242#define smt_capable() (smp_num_siblings > 1) 243#define smt_capable() (smp_num_siblings > 1)
243#endif 244#endif
244 245
diff --git a/arch/x86/include/asm/tsc.h b/arch/x86/include/asm/tsc.h
index 38ae163cc91b..9cd83a8e40d5 100644
--- a/arch/x86/include/asm/tsc.h
+++ b/arch/x86/include/asm/tsc.h
@@ -34,6 +34,8 @@ static inline cycles_t get_cycles(void)
34 34
35static __always_inline cycles_t vget_cycles(void) 35static __always_inline cycles_t vget_cycles(void)
36{ 36{
37 cycles_t cycles;
38
37 /* 39 /*
38 * We only do VDSOs on TSC capable CPUs, so this shouldnt 40 * We only do VDSOs on TSC capable CPUs, so this shouldnt
39 * access boot_cpu_data (which is not VDSO-safe): 41 * access boot_cpu_data (which is not VDSO-safe):
@@ -42,7 +44,11 @@ static __always_inline cycles_t vget_cycles(void)
42 if (!cpu_has_tsc) 44 if (!cpu_has_tsc)
43 return 0; 45 return 0;
44#endif 46#endif
45 return (cycles_t)__native_read_tsc(); 47 rdtsc_barrier();
48 cycles = (cycles_t)__native_read_tsc();
49 rdtsc_barrier();
50
51 return cycles;
46} 52}
47 53
48extern void tsc_init(void); 54extern void tsc_init(void);
diff --git a/arch/x86/include/asm/uaccess_64.h b/arch/x86/include/asm/uaccess_64.h
index 664f15280f14..f8cfd00db450 100644
--- a/arch/x86/include/asm/uaccess_64.h
+++ b/arch/x86/include/asm/uaccess_64.h
@@ -46,7 +46,7 @@ int __copy_from_user(void *dst, const void __user *src, unsigned size)
46 return ret; 46 return ret;
47 case 10: 47 case 10:
48 __get_user_asm(*(u64 *)dst, (u64 __user *)src, 48 __get_user_asm(*(u64 *)dst, (u64 __user *)src,
49 ret, "q", "", "=r", 16); 49 ret, "q", "", "=r", 10);
50 if (unlikely(ret)) 50 if (unlikely(ret))
51 return ret; 51 return ret;
52 __get_user_asm(*(u16 *)(8 + (char *)dst), 52 __get_user_asm(*(u16 *)(8 + (char *)dst),
diff --git a/arch/x86/include/asm/unistd_64.h b/arch/x86/include/asm/unistd_64.h
index 834b2c1d89fb..d2e415e6666f 100644
--- a/arch/x86/include/asm/unistd_64.h
+++ b/arch/x86/include/asm/unistd_64.h
@@ -639,8 +639,8 @@ __SYSCALL(__NR_fallocate, sys_fallocate)
639__SYSCALL(__NR_timerfd_settime, sys_timerfd_settime) 639__SYSCALL(__NR_timerfd_settime, sys_timerfd_settime)
640#define __NR_timerfd_gettime 287 640#define __NR_timerfd_gettime 287
641__SYSCALL(__NR_timerfd_gettime, sys_timerfd_gettime) 641__SYSCALL(__NR_timerfd_gettime, sys_timerfd_gettime)
642#define __NR_paccept 288 642#define __NR_accept4 288
643__SYSCALL(__NR_paccept, sys_paccept) 643__SYSCALL(__NR_accept4, sys_accept4)
644#define __NR_signalfd4 289 644#define __NR_signalfd4 289
645__SYSCALL(__NR_signalfd4, sys_signalfd4) 645__SYSCALL(__NR_signalfd4, sys_signalfd4)
646#define __NR_eventfd2 290 646#define __NR_eventfd2 290
diff --git a/arch/x86/include/asm/uv/uv_hub.h b/arch/x86/include/asm/uv/uv_hub.h
index c6ad93e315c8..7a5782610b2b 100644
--- a/arch/x86/include/asm/uv/uv_hub.h
+++ b/arch/x86/include/asm/uv/uv_hub.h
@@ -13,6 +13,7 @@
13 13
14#include <linux/numa.h> 14#include <linux/numa.h>
15#include <linux/percpu.h> 15#include <linux/percpu.h>
16#include <linux/timer.h>
16#include <asm/types.h> 17#include <asm/types.h>
17#include <asm/percpu.h> 18#include <asm/percpu.h>
18 19
diff --git a/arch/x86/include/asm/voyager.h b/arch/x86/include/asm/voyager.h
index 9c811d2e6f91..b3e647307625 100644
--- a/arch/x86/include/asm/voyager.h
+++ b/arch/x86/include/asm/voyager.h
@@ -520,6 +520,7 @@ extern void voyager_restart(void);
520extern void voyager_cat_power_off(void); 520extern void voyager_cat_power_off(void);
521extern void voyager_cat_do_common_interrupt(void); 521extern void voyager_cat_do_common_interrupt(void);
522extern void voyager_handle_nmi(void); 522extern void voyager_handle_nmi(void);
523extern void voyager_smp_intr_init(void);
523/* Commands for the following are */ 524/* Commands for the following are */
524#define VOYAGER_PSI_READ 0 525#define VOYAGER_PSI_READ 0
525#define VOYAGER_PSI_WRITE 1 526#define VOYAGER_PSI_WRITE 1
diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile
index d7e5a58ee22f..b62a7667828e 100644
--- a/arch/x86/kernel/Makefile
+++ b/arch/x86/kernel/Makefile
@@ -6,11 +6,12 @@ extra-y := head_$(BITS).o head$(BITS).o head.o init_task.o vmlinu
6 6
7CPPFLAGS_vmlinux.lds += -U$(UTS_MACHINE) 7CPPFLAGS_vmlinux.lds += -U$(UTS_MACHINE)
8 8
9ifdef CONFIG_FTRACE 9ifdef CONFIG_FUNCTION_TRACER
10# Do not profile debug and lowlevel utilities 10# Do not profile debug and lowlevel utilities
11CFLAGS_REMOVE_tsc.o = -pg 11CFLAGS_REMOVE_tsc.o = -pg
12CFLAGS_REMOVE_rtc.o = -pg 12CFLAGS_REMOVE_rtc.o = -pg
13CFLAGS_REMOVE_paravirt-spinlocks.o = -pg 13CFLAGS_REMOVE_paravirt-spinlocks.o = -pg
14CFLAGS_REMOVE_ftrace.o = -pg
14endif 15endif
15 16
16# 17#
@@ -40,7 +41,7 @@ obj-$(CONFIG_X86_TRAMPOLINE) += trampoline.o
40obj-y += process.o 41obj-y += process.o
41obj-y += i387.o xsave.o 42obj-y += i387.o xsave.o
42obj-y += ptrace.o 43obj-y += ptrace.o
43obj-y += ds.o 44obj-$(CONFIG_X86_DS) += ds.o
44obj-$(CONFIG_X86_32) += tls.o 45obj-$(CONFIG_X86_32) += tls.o
45obj-$(CONFIG_IA32_EMULATION) += tls.o 46obj-$(CONFIG_IA32_EMULATION) += tls.o
46obj-y += step.o 47obj-y += step.o
diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c
index 8c1f76abae9e..4c51a2f8fd31 100644
--- a/arch/x86/kernel/acpi/boot.c
+++ b/arch/x86/kernel/acpi/boot.c
@@ -1343,7 +1343,6 @@ static void __init acpi_process_madt(void)
1343 error = acpi_parse_madt_ioapic_entries(); 1343 error = acpi_parse_madt_ioapic_entries();
1344 if (!error) { 1344 if (!error) {
1345 acpi_irq_model = ACPI_IRQ_MODEL_IOAPIC; 1345 acpi_irq_model = ACPI_IRQ_MODEL_IOAPIC;
1346 acpi_irq_balance_set(NULL);
1347 acpi_ioapic = 1; 1346 acpi_ioapic = 1;
1348 1347
1349 smp_found_config = 1; 1348 smp_found_config = 1;
diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c
index a8fd9ebdc8e2..a7b6dec6fc3f 100644
--- a/arch/x86/kernel/amd_iommu.c
+++ b/arch/x86/kernel/amd_iommu.c
@@ -50,7 +50,7 @@ static int dma_ops_unity_map(struct dma_ops_domain *dma_dom,
50/* returns !0 if the IOMMU is caching non-present entries in its TLB */ 50/* returns !0 if the IOMMU is caching non-present entries in its TLB */
51static int iommu_has_npcache(struct amd_iommu *iommu) 51static int iommu_has_npcache(struct amd_iommu *iommu)
52{ 52{
53 return iommu->cap & IOMMU_CAP_NPCACHE; 53 return iommu->cap & (1UL << IOMMU_CAP_NPCACHE);
54} 54}
55 55
56/**************************************************************************** 56/****************************************************************************
@@ -187,6 +187,8 @@ static int iommu_queue_command(struct amd_iommu *iommu, struct iommu_cmd *cmd)
187 187
188 spin_lock_irqsave(&iommu->lock, flags); 188 spin_lock_irqsave(&iommu->lock, flags);
189 ret = __iommu_queue_command(iommu, cmd); 189 ret = __iommu_queue_command(iommu, cmd);
190 if (!ret)
191 iommu->need_sync = 1;
190 spin_unlock_irqrestore(&iommu->lock, flags); 192 spin_unlock_irqrestore(&iommu->lock, flags);
191 193
192 return ret; 194 return ret;
@@ -210,10 +212,13 @@ static int iommu_completion_wait(struct amd_iommu *iommu)
210 cmd.data[0] = CMD_COMPL_WAIT_INT_MASK; 212 cmd.data[0] = CMD_COMPL_WAIT_INT_MASK;
211 CMD_SET_TYPE(&cmd, CMD_COMPL_WAIT); 213 CMD_SET_TYPE(&cmd, CMD_COMPL_WAIT);
212 214
213 iommu->need_sync = 0;
214
215 spin_lock_irqsave(&iommu->lock, flags); 215 spin_lock_irqsave(&iommu->lock, flags);
216 216
217 if (!iommu->need_sync)
218 goto out;
219
220 iommu->need_sync = 0;
221
217 ret = __iommu_queue_command(iommu, &cmd); 222 ret = __iommu_queue_command(iommu, &cmd);
218 223
219 if (ret) 224 if (ret)
@@ -254,8 +259,6 @@ static int iommu_queue_inv_dev_entry(struct amd_iommu *iommu, u16 devid)
254 259
255 ret = iommu_queue_command(iommu, &cmd); 260 ret = iommu_queue_command(iommu, &cmd);
256 261
257 iommu->need_sync = 1;
258
259 return ret; 262 return ret;
260} 263}
261 264
@@ -281,8 +284,6 @@ static int iommu_queue_inv_iommu_pages(struct amd_iommu *iommu,
281 284
282 ret = iommu_queue_command(iommu, &cmd); 285 ret = iommu_queue_command(iommu, &cmd);
283 286
284 iommu->need_sync = 1;
285
286 return ret; 287 return ret;
287} 288}
288 289
@@ -343,7 +344,7 @@ static int iommu_map(struct protection_domain *dom,
343 u64 __pte, *pte, *page; 344 u64 __pte, *pte, *page;
344 345
345 bus_addr = PAGE_ALIGN(bus_addr); 346 bus_addr = PAGE_ALIGN(bus_addr);
346 phys_addr = PAGE_ALIGN(bus_addr); 347 phys_addr = PAGE_ALIGN(phys_addr);
347 348
348 /* only support 512GB address spaces for now */ 349 /* only support 512GB address spaces for now */
349 if (bus_addr > IOMMU_MAP_SIZE_L3 || !(prot & IOMMU_PROT_MASK)) 350 if (bus_addr > IOMMU_MAP_SIZE_L3 || !(prot & IOMMU_PROT_MASK))
@@ -536,6 +537,9 @@ static void dma_ops_free_addresses(struct dma_ops_domain *dom,
536{ 537{
537 address >>= PAGE_SHIFT; 538 address >>= PAGE_SHIFT;
538 iommu_area_free(dom->bitmap, address, pages); 539 iommu_area_free(dom->bitmap, address, pages);
540
541 if (address >= dom->next_bit)
542 dom->need_flush = true;
539} 543}
540 544
541/**************************************************************************** 545/****************************************************************************
@@ -596,7 +600,7 @@ static void dma_ops_free_pagetable(struct dma_ops_domain *dma_dom)
596 continue; 600 continue;
597 601
598 p2 = IOMMU_PTE_PAGE(p1[i]); 602 p2 = IOMMU_PTE_PAGE(p1[i]);
599 for (j = 0; j < 512; ++i) { 603 for (j = 0; j < 512; ++j) {
600 if (!IOMMU_PTE_PRESENT(p2[j])) 604 if (!IOMMU_PTE_PRESENT(p2[j]))
601 continue; 605 continue;
602 p3 = IOMMU_PTE_PAGE(p2[j]); 606 p3 = IOMMU_PTE_PAGE(p2[j]);
@@ -759,8 +763,6 @@ static void set_device_domain(struct amd_iommu *iommu,
759 write_unlock_irqrestore(&amd_iommu_devtable_lock, flags); 763 write_unlock_irqrestore(&amd_iommu_devtable_lock, flags);
760 764
761 iommu_queue_inv_dev_entry(iommu, devid); 765 iommu_queue_inv_dev_entry(iommu, devid);
762
763 iommu->need_sync = 1;
764} 766}
765 767
766/***************************************************************************** 768/*****************************************************************************
@@ -855,6 +857,9 @@ static int get_device_resources(struct device *dev,
855 print_devid(_bdf, 1); 857 print_devid(_bdf, 1);
856 } 858 }
857 859
860 if (domain_for_device(_bdf) == NULL)
861 set_device_domain(*iommu, *domain, _bdf);
862
858 return 1; 863 return 1;
859} 864}
860 865
@@ -905,7 +910,7 @@ static void dma_ops_domain_unmap(struct amd_iommu *iommu,
905 if (address >= dom->aperture_size) 910 if (address >= dom->aperture_size)
906 return; 911 return;
907 912
908 WARN_ON(address & 0xfffULL || address > dom->aperture_size); 913 WARN_ON(address & ~PAGE_MASK || address >= dom->aperture_size);
909 914
910 pte = dom->pte_pages[IOMMU_PTE_L1_INDEX(address)]; 915 pte = dom->pte_pages[IOMMU_PTE_L1_INDEX(address)];
911 pte += IOMMU_PTE_L0_INDEX(address); 916 pte += IOMMU_PTE_L0_INDEX(address);
@@ -917,8 +922,8 @@ static void dma_ops_domain_unmap(struct amd_iommu *iommu,
917 922
918/* 923/*
919 * This function contains common code for mapping of a physically 924 * This function contains common code for mapping of a physically
920 * contiguous memory region into DMA address space. It is uses by all 925 * contiguous memory region into DMA address space. It is used by all
921 * mapping functions provided by this IOMMU driver. 926 * mapping functions provided with this IOMMU driver.
922 * Must be called with the domain lock held. 927 * Must be called with the domain lock held.
923 */ 928 */
924static dma_addr_t __map_single(struct device *dev, 929static dma_addr_t __map_single(struct device *dev,
@@ -978,7 +983,8 @@ static void __unmap_single(struct amd_iommu *iommu,
978 dma_addr_t i, start; 983 dma_addr_t i, start;
979 unsigned int pages; 984 unsigned int pages;
980 985
981 if ((dma_addr == 0) || (dma_addr + size > dma_dom->aperture_size)) 986 if ((dma_addr == bad_dma_address) ||
987 (dma_addr + size > dma_dom->aperture_size))
982 return; 988 return;
983 989
984 pages = iommu_num_pages(dma_addr, size, PAGE_SIZE); 990 pages = iommu_num_pages(dma_addr, size, PAGE_SIZE);
@@ -992,8 +998,10 @@ static void __unmap_single(struct amd_iommu *iommu,
992 998
993 dma_ops_free_addresses(dma_dom, dma_addr, pages); 999 dma_ops_free_addresses(dma_dom, dma_addr, pages);
994 1000
995 if (amd_iommu_unmap_flush) 1001 if (amd_iommu_unmap_flush || dma_dom->need_flush) {
996 iommu_flush_pages(iommu, dma_dom->domain.id, dma_addr, size); 1002 iommu_flush_pages(iommu, dma_dom->domain.id, dma_addr, size);
1003 dma_dom->need_flush = false;
1004 }
997} 1005}
998 1006
999/* 1007/*
@@ -1026,8 +1034,7 @@ static dma_addr_t map_single(struct device *dev, phys_addr_t paddr,
1026 if (addr == bad_dma_address) 1034 if (addr == bad_dma_address)
1027 goto out; 1035 goto out;
1028 1036
1029 if (unlikely(iommu->need_sync)) 1037 iommu_completion_wait(iommu);
1030 iommu_completion_wait(iommu);
1031 1038
1032out: 1039out:
1033 spin_unlock_irqrestore(&domain->lock, flags); 1040 spin_unlock_irqrestore(&domain->lock, flags);
@@ -1055,8 +1062,7 @@ static void unmap_single(struct device *dev, dma_addr_t dma_addr,
1055 1062
1056 __unmap_single(iommu, domain->priv, dma_addr, size, dir); 1063 __unmap_single(iommu, domain->priv, dma_addr, size, dir);
1057 1064
1058 if (unlikely(iommu->need_sync)) 1065 iommu_completion_wait(iommu);
1059 iommu_completion_wait(iommu);
1060 1066
1061 spin_unlock_irqrestore(&domain->lock, flags); 1067 spin_unlock_irqrestore(&domain->lock, flags);
1062} 1068}
@@ -1122,8 +1128,7 @@ static int map_sg(struct device *dev, struct scatterlist *sglist,
1122 goto unmap; 1128 goto unmap;
1123 } 1129 }
1124 1130
1125 if (unlikely(iommu->need_sync)) 1131 iommu_completion_wait(iommu);
1126 iommu_completion_wait(iommu);
1127 1132
1128out: 1133out:
1129 spin_unlock_irqrestore(&domain->lock, flags); 1134 spin_unlock_irqrestore(&domain->lock, flags);
@@ -1168,8 +1173,7 @@ static void unmap_sg(struct device *dev, struct scatterlist *sglist,
1168 s->dma_address = s->dma_length = 0; 1173 s->dma_address = s->dma_length = 0;
1169 } 1174 }
1170 1175
1171 if (unlikely(iommu->need_sync)) 1176 iommu_completion_wait(iommu);
1172 iommu_completion_wait(iommu);
1173 1177
1174 spin_unlock_irqrestore(&domain->lock, flags); 1178 spin_unlock_irqrestore(&domain->lock, flags);
1175} 1179}
@@ -1220,8 +1224,7 @@ static void *alloc_coherent(struct device *dev, size_t size,
1220 goto out; 1224 goto out;
1221 } 1225 }
1222 1226
1223 if (unlikely(iommu->need_sync)) 1227 iommu_completion_wait(iommu);
1224 iommu_completion_wait(iommu);
1225 1228
1226out: 1229out:
1227 spin_unlock_irqrestore(&domain->lock, flags); 1230 spin_unlock_irqrestore(&domain->lock, flags);
@@ -1252,8 +1255,7 @@ static void free_coherent(struct device *dev, size_t size,
1252 1255
1253 __unmap_single(iommu, domain->priv, dma_addr, size, DMA_BIDIRECTIONAL); 1256 __unmap_single(iommu, domain->priv, dma_addr, size, DMA_BIDIRECTIONAL);
1254 1257
1255 if (unlikely(iommu->need_sync)) 1258 iommu_completion_wait(iommu);
1256 iommu_completion_wait(iommu);
1257 1259
1258 spin_unlock_irqrestore(&domain->lock, flags); 1260 spin_unlock_irqrestore(&domain->lock, flags);
1259 1261
diff --git a/arch/x86/kernel/amd_iommu_init.c b/arch/x86/kernel/amd_iommu_init.c
index 0cdcda35a05f..30ae2701b3df 100644
--- a/arch/x86/kernel/amd_iommu_init.c
+++ b/arch/x86/kernel/amd_iommu_init.c
@@ -121,7 +121,7 @@ u16 amd_iommu_last_bdf; /* largest PCI device id we have
121LIST_HEAD(amd_iommu_unity_map); /* a list of required unity mappings 121LIST_HEAD(amd_iommu_unity_map); /* a list of required unity mappings
122 we find in ACPI */ 122 we find in ACPI */
123unsigned amd_iommu_aperture_order = 26; /* size of aperture in power of 2 */ 123unsigned amd_iommu_aperture_order = 26; /* size of aperture in power of 2 */
124int amd_iommu_isolate; /* if 1, device isolation is enabled */ 124int amd_iommu_isolate = 1; /* if 1, device isolation is enabled */
125bool amd_iommu_unmap_flush; /* if true, flush on every unmap */ 125bool amd_iommu_unmap_flush; /* if true, flush on every unmap */
126 126
127LIST_HEAD(amd_iommu_list); /* list of all AMD IOMMUs in the 127LIST_HEAD(amd_iommu_list); /* list of all AMD IOMMUs in the
@@ -1213,7 +1213,9 @@ static int __init parse_amd_iommu_options(char *str)
1213 for (; *str; ++str) { 1213 for (; *str; ++str) {
1214 if (strncmp(str, "isolate", 7) == 0) 1214 if (strncmp(str, "isolate", 7) == 0)
1215 amd_iommu_isolate = 1; 1215 amd_iommu_isolate = 1;
1216 if (strncmp(str, "fullflush", 11) == 0) 1216 if (strncmp(str, "share", 5) == 0)
1217 amd_iommu_isolate = 0;
1218 if (strncmp(str, "fullflush", 9) == 0)
1217 amd_iommu_unmap_flush = true; 1219 amd_iommu_unmap_flush = true;
1218 } 1220 }
1219 1221
diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c
index 04a7f960bbc0..16f94879b525 100644
--- a/arch/x86/kernel/apic.c
+++ b/arch/x86/kernel/apic.c
@@ -1315,7 +1315,7 @@ void enable_x2apic(void)
1315 } 1315 }
1316} 1316}
1317 1317
1318void enable_IR_x2apic(void) 1318void __init enable_IR_x2apic(void)
1319{ 1319{
1320#ifdef CONFIG_INTR_REMAP 1320#ifdef CONFIG_INTR_REMAP
1321 int ret; 1321 int ret;
diff --git a/arch/x86/kernel/cpu/addon_cpuid_features.c b/arch/x86/kernel/cpu/addon_cpuid_features.c
index 0d9c993aa93e..ef8f831af823 100644
--- a/arch/x86/kernel/cpu/addon_cpuid_features.c
+++ b/arch/x86/kernel/cpu/addon_cpuid_features.c
@@ -69,7 +69,7 @@ void __cpuinit init_scattered_cpuid_features(struct cpuinfo_x86 *c)
69 */ 69 */
70void __cpuinit detect_extended_topology(struct cpuinfo_x86 *c) 70void __cpuinit detect_extended_topology(struct cpuinfo_x86 *c)
71{ 71{
72#ifdef CONFIG_SMP 72#ifdef CONFIG_X86_SMP
73 unsigned int eax, ebx, ecx, edx, sub_index; 73 unsigned int eax, ebx, ecx, edx, sub_index;
74 unsigned int ht_mask_width, core_plus_mask_width; 74 unsigned int ht_mask_width, core_plus_mask_width;
75 unsigned int core_select_mask, core_level_siblings; 75 unsigned int core_select_mask, core_level_siblings;
diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
index 25581dcb280e..b9c9ea0217a9 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -20,6 +20,7 @@
20#include <asm/pat.h> 20#include <asm/pat.h>
21#include <asm/asm.h> 21#include <asm/asm.h>
22#include <asm/numa.h> 22#include <asm/numa.h>
23#include <asm/smp.h>
23#ifdef CONFIG_X86_LOCAL_APIC 24#ifdef CONFIG_X86_LOCAL_APIC
24#include <asm/mpspec.h> 25#include <asm/mpspec.h>
25#include <asm/apic.h> 26#include <asm/apic.h>
@@ -549,6 +550,10 @@ static void __init early_identify_cpu(struct cpuinfo_x86 *c)
549 this_cpu->c_early_init(c); 550 this_cpu->c_early_init(c);
550 551
551 validate_pat_support(c); 552 validate_pat_support(c);
553
554#ifdef CONFIG_SMP
555 c->cpu_index = boot_cpu_id;
556#endif
552} 557}
553 558
554void __init early_cpu_init(void) 559void __init early_cpu_init(void)
@@ -1134,7 +1139,7 @@ void __cpuinit cpu_init(void)
1134 /* 1139 /*
1135 * Boot processor to setup the FP and extended state context info. 1140 * Boot processor to setup the FP and extended state context info.
1136 */ 1141 */
1137 if (!smp_processor_id()) 1142 if (smp_processor_id() == boot_cpu_id)
1138 init_thread_xstate(); 1143 init_thread_xstate();
1139 1144
1140 xsave_init(); 1145 xsave_init();
diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
index d3dcd58b87cd..7f05f44b97e9 100644
--- a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
+++ b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
@@ -115,9 +115,20 @@ static int query_current_values_with_pending_wait(struct powernow_k8_data *data)
115 u32 i = 0; 115 u32 i = 0;
116 116
117 if (cpu_family == CPU_HW_PSTATE) { 117 if (cpu_family == CPU_HW_PSTATE) {
118 rdmsr(MSR_PSTATE_STATUS, lo, hi); 118 if (data->currpstate == HW_PSTATE_INVALID) {
119 i = lo & HW_PSTATE_MASK; 119 /* read (initial) hw pstate if not yet set */
120 data->currpstate = i; 120 rdmsr(MSR_PSTATE_STATUS, lo, hi);
121 i = lo & HW_PSTATE_MASK;
122
123 /*
124 * a workaround for family 11h erratum 311 might cause
125 * an "out-of-range Pstate if the core is in Pstate-0
126 */
127 if (i >= data->numps)
128 data->currpstate = HW_PSTATE_0;
129 else
130 data->currpstate = i;
131 }
121 return 0; 132 return 0;
122 } 133 }
123 do { 134 do {
@@ -1121,6 +1132,7 @@ static int __cpuinit powernowk8_cpu_init(struct cpufreq_policy *pol)
1121 } 1132 }
1122 1133
1123 data->cpu = pol->cpu; 1134 data->cpu = pol->cpu;
1135 data->currpstate = HW_PSTATE_INVALID;
1124 1136
1125 if (powernow_k8_cpu_init_acpi(data)) { 1137 if (powernow_k8_cpu_init_acpi(data)) {
1126 /* 1138 /*
diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k8.h b/arch/x86/kernel/cpu/cpufreq/powernow-k8.h
index ab48cfed4d96..65cfb5d7f77f 100644
--- a/arch/x86/kernel/cpu/cpufreq/powernow-k8.h
+++ b/arch/x86/kernel/cpu/cpufreq/powernow-k8.h
@@ -5,6 +5,19 @@
5 * http://www.gnu.org/licenses/gpl.html 5 * http://www.gnu.org/licenses/gpl.html
6 */ 6 */
7 7
8
9enum pstate {
10 HW_PSTATE_INVALID = 0xff,
11 HW_PSTATE_0 = 0,
12 HW_PSTATE_1 = 1,
13 HW_PSTATE_2 = 2,
14 HW_PSTATE_3 = 3,
15 HW_PSTATE_4 = 4,
16 HW_PSTATE_5 = 5,
17 HW_PSTATE_6 = 6,
18 HW_PSTATE_7 = 7,
19};
20
8struct powernow_k8_data { 21struct powernow_k8_data {
9 unsigned int cpu; 22 unsigned int cpu;
10 23
@@ -23,7 +36,9 @@ struct powernow_k8_data {
23 u32 exttype; /* extended interface = 1 */ 36 u32 exttype; /* extended interface = 1 */
24 37
25 /* keep track of the current fid / vid or pstate */ 38 /* keep track of the current fid / vid or pstate */
26 u32 currvid, currfid, currpstate; 39 u32 currvid;
40 u32 currfid;
41 enum pstate currpstate;
27 42
28 /* the powernow_table includes all frequency and vid/fid pairings: 43 /* the powernow_table includes all frequency and vid/fid pairings:
29 * fid are the lower 8 bits of the index, vid are the upper 8 bits. 44 * fid are the lower 8 bits of the index, vid are the upper 8 bits.
diff --git a/arch/x86/kernel/ds.c b/arch/x86/kernel/ds.c
index 2b69994fd3a8..a2d1176c38ee 100644
--- a/arch/x86/kernel/ds.c
+++ b/arch/x86/kernel/ds.c
@@ -21,8 +21,6 @@
21 */ 21 */
22 22
23 23
24#ifdef CONFIG_X86_DS
25
26#include <asm/ds.h> 24#include <asm/ds.h>
27 25
28#include <linux/errno.h> 26#include <linux/errno.h>
@@ -211,14 +209,15 @@ static DEFINE_PER_CPU(struct ds_context *, system_context);
211static inline struct ds_context *ds_get_context(struct task_struct *task) 209static inline struct ds_context *ds_get_context(struct task_struct *task)
212{ 210{
213 struct ds_context *context; 211 struct ds_context *context;
212 unsigned long irq;
214 213
215 spin_lock(&ds_lock); 214 spin_lock_irqsave(&ds_lock, irq);
216 215
217 context = (task ? task->thread.ds_ctx : this_system_context); 216 context = (task ? task->thread.ds_ctx : this_system_context);
218 if (context) 217 if (context)
219 context->count++; 218 context->count++;
220 219
221 spin_unlock(&ds_lock); 220 spin_unlock_irqrestore(&ds_lock, irq);
222 221
223 return context; 222 return context;
224} 223}
@@ -226,18 +225,16 @@ static inline struct ds_context *ds_get_context(struct task_struct *task)
226/* 225/*
227 * Same as ds_get_context, but allocates the context and it's DS 226 * Same as ds_get_context, but allocates the context and it's DS
228 * structure, if necessary; returns NULL; if out of memory. 227 * structure, if necessary; returns NULL; if out of memory.
229 *
230 * pre: requires ds_lock to be held
231 */ 228 */
232static inline struct ds_context *ds_alloc_context(struct task_struct *task) 229static inline struct ds_context *ds_alloc_context(struct task_struct *task)
233{ 230{
234 struct ds_context **p_context = 231 struct ds_context **p_context =
235 (task ? &task->thread.ds_ctx : &this_system_context); 232 (task ? &task->thread.ds_ctx : &this_system_context);
236 struct ds_context *context = *p_context; 233 struct ds_context *context = *p_context;
234 unsigned long irq;
237 235
238 if (!context) { 236 if (!context) {
239 context = kzalloc(sizeof(*context), GFP_KERNEL); 237 context = kzalloc(sizeof(*context), GFP_KERNEL);
240
241 if (!context) 238 if (!context)
242 return NULL; 239 return NULL;
243 240
@@ -247,18 +244,27 @@ static inline struct ds_context *ds_alloc_context(struct task_struct *task)
247 return NULL; 244 return NULL;
248 } 245 }
249 246
250 *p_context = context; 247 spin_lock_irqsave(&ds_lock, irq);
251 248
252 context->this = p_context; 249 if (*p_context) {
253 context->task = task; 250 kfree(context->ds);
251 kfree(context);
252
253 context = *p_context;
254 } else {
255 *p_context = context;
254 256
255 if (task) 257 context->this = p_context;
256 set_tsk_thread_flag(task, TIF_DS_AREA_MSR); 258 context->task = task;
257 259
258 if (!task || (task == current)) 260 if (task)
259 wrmsr(MSR_IA32_DS_AREA, (unsigned long)context->ds, 0); 261 set_tsk_thread_flag(task, TIF_DS_AREA_MSR);
260 262
261 get_tracer(task); 263 if (!task || (task == current))
264 wrmsrl(MSR_IA32_DS_AREA,
265 (unsigned long)context->ds);
266 }
267 spin_unlock_irqrestore(&ds_lock, irq);
262 } 268 }
263 269
264 context->count++; 270 context->count++;
@@ -272,10 +278,12 @@ static inline struct ds_context *ds_alloc_context(struct task_struct *task)
272 */ 278 */
273static inline void ds_put_context(struct ds_context *context) 279static inline void ds_put_context(struct ds_context *context)
274{ 280{
281 unsigned long irq;
282
275 if (!context) 283 if (!context)
276 return; 284 return;
277 285
278 spin_lock(&ds_lock); 286 spin_lock_irqsave(&ds_lock, irq);
279 287
280 if (--context->count) 288 if (--context->count)
281 goto out; 289 goto out;
@@ -297,7 +305,7 @@ static inline void ds_put_context(struct ds_context *context)
297 kfree(context->ds); 305 kfree(context->ds);
298 kfree(context); 306 kfree(context);
299 out: 307 out:
300 spin_unlock(&ds_lock); 308 spin_unlock_irqrestore(&ds_lock, irq);
301} 309}
302 310
303 311
@@ -368,6 +376,7 @@ static int ds_request(struct task_struct *task, void *base, size_t size,
368 struct ds_context *context; 376 struct ds_context *context;
369 unsigned long buffer, adj; 377 unsigned long buffer, adj;
370 const unsigned long alignment = (1 << 3); 378 const unsigned long alignment = (1 << 3);
379 unsigned long irq;
371 int error = 0; 380 int error = 0;
372 381
373 if (!ds_cfg.sizeof_ds) 382 if (!ds_cfg.sizeof_ds)
@@ -382,25 +391,27 @@ static int ds_request(struct task_struct *task, void *base, size_t size,
382 return -EOPNOTSUPP; 391 return -EOPNOTSUPP;
383 392
384 393
385 spin_lock(&ds_lock);
386
387 if (!check_tracer(task))
388 return -EPERM;
389
390 error = -ENOMEM;
391 context = ds_alloc_context(task); 394 context = ds_alloc_context(task);
392 if (!context) 395 if (!context)
396 return -ENOMEM;
397
398 spin_lock_irqsave(&ds_lock, irq);
399
400 error = -EPERM;
401 if (!check_tracer(task))
393 goto out_unlock; 402 goto out_unlock;
394 403
404 get_tracer(task);
405
395 error = -EALREADY; 406 error = -EALREADY;
396 if (context->owner[qual] == current) 407 if (context->owner[qual] == current)
397 goto out_unlock; 408 goto out_put_tracer;
398 error = -EPERM; 409 error = -EPERM;
399 if (context->owner[qual] != NULL) 410 if (context->owner[qual] != NULL)
400 goto out_unlock; 411 goto out_put_tracer;
401 context->owner[qual] = current; 412 context->owner[qual] = current;
402 413
403 spin_unlock(&ds_lock); 414 spin_unlock_irqrestore(&ds_lock, irq);
404 415
405 416
406 error = -ENOMEM; 417 error = -ENOMEM;
@@ -448,10 +459,17 @@ static int ds_request(struct task_struct *task, void *base, size_t size,
448 out_release: 459 out_release:
449 context->owner[qual] = NULL; 460 context->owner[qual] = NULL;
450 ds_put_context(context); 461 ds_put_context(context);
462 put_tracer(task);
463 return error;
464
465 out_put_tracer:
466 spin_unlock_irqrestore(&ds_lock, irq);
467 ds_put_context(context);
468 put_tracer(task);
451 return error; 469 return error;
452 470
453 out_unlock: 471 out_unlock:
454 spin_unlock(&ds_lock); 472 spin_unlock_irqrestore(&ds_lock, irq);
455 ds_put_context(context); 473 ds_put_context(context);
456 return error; 474 return error;
457} 475}
@@ -801,13 +819,21 @@ static const struct ds_configuration ds_cfg_var = {
801 .sizeof_ds = sizeof(long) * 12, 819 .sizeof_ds = sizeof(long) * 12,
802 .sizeof_field = sizeof(long), 820 .sizeof_field = sizeof(long),
803 .sizeof_rec[ds_bts] = sizeof(long) * 3, 821 .sizeof_rec[ds_bts] = sizeof(long) * 3,
822#ifdef __i386__
804 .sizeof_rec[ds_pebs] = sizeof(long) * 10 823 .sizeof_rec[ds_pebs] = sizeof(long) * 10
824#else
825 .sizeof_rec[ds_pebs] = sizeof(long) * 18
826#endif
805}; 827};
806static const struct ds_configuration ds_cfg_64 = { 828static const struct ds_configuration ds_cfg_64 = {
807 .sizeof_ds = 8 * 12, 829 .sizeof_ds = 8 * 12,
808 .sizeof_field = 8, 830 .sizeof_field = 8,
809 .sizeof_rec[ds_bts] = 8 * 3, 831 .sizeof_rec[ds_bts] = 8 * 3,
832#ifdef __i386__
810 .sizeof_rec[ds_pebs] = 8 * 10 833 .sizeof_rec[ds_pebs] = 8 * 10
834#else
835 .sizeof_rec[ds_pebs] = 8 * 18
836#endif
811}; 837};
812 838
813static inline void 839static inline void
@@ -861,4 +887,3 @@ void ds_free(struct ds_context *context)
861 while (leftovers--) 887 while (leftovers--)
862 ds_put_context(context); 888 ds_put_context(context);
863} 889}
864#endif /* CONFIG_X86_DS */
diff --git a/arch/x86/kernel/e820.c b/arch/x86/kernel/e820.c
index ce97bf3bed12..7aafeb5263ef 100644
--- a/arch/x86/kernel/e820.c
+++ b/arch/x86/kernel/e820.c
@@ -1290,15 +1290,17 @@ void __init e820_reserve_resources(void)
1290 res->start = e820.map[i].addr; 1290 res->start = e820.map[i].addr;
1291 res->end = end; 1291 res->end = end;
1292 1292
1293 res->flags = IORESOURCE_MEM | IORESOURCE_BUSY; 1293 res->flags = IORESOURCE_MEM;
1294 1294
1295 /* 1295 /*
1296 * don't register the region that could be conflicted with 1296 * don't register the region that could be conflicted with
1297 * pci device BAR resource and insert them later in 1297 * pci device BAR resource and insert them later in
1298 * pcibios_resource_survey() 1298 * pcibios_resource_survey()
1299 */ 1299 */
1300 if (e820.map[i].type != E820_RESERVED || res->start < (1ULL<<20)) 1300 if (e820.map[i].type != E820_RESERVED || res->start < (1ULL<<20)) {
1301 res->flags |= IORESOURCE_BUSY;
1301 insert_resource(&iomem_resource, res); 1302 insert_resource(&iomem_resource, res);
1303 }
1302 res++; 1304 res++;
1303 } 1305 }
1304 1306
@@ -1318,7 +1320,7 @@ void __init e820_reserve_resources_late(void)
1318 res = e820_res; 1320 res = e820_res;
1319 for (i = 0; i < e820.nr_map; i++) { 1321 for (i = 0; i < e820.nr_map; i++) {
1320 if (!res->parent && res->end) 1322 if (!res->parent && res->end)
1321 reserve_region_with_split(&iomem_resource, res->start, res->end, res->name); 1323 insert_resource_expand_to_fit(&iomem_resource, res);
1322 res++; 1324 res++;
1323 } 1325 }
1324} 1326}
diff --git a/arch/x86/kernel/early-quirks.c b/arch/x86/kernel/early-quirks.c
index 3ce029ffaa55..1b894b72c0f5 100644
--- a/arch/x86/kernel/early-quirks.c
+++ b/arch/x86/kernel/early-quirks.c
@@ -188,20 +188,6 @@ static void __init ati_bugs_contd(int num, int slot, int func)
188} 188}
189#endif 189#endif
190 190
191#ifdef CONFIG_DMAR
192static void __init intel_g33_dmar(int num, int slot, int func)
193{
194 struct acpi_table_header *dmar_tbl;
195 acpi_status status;
196
197 status = acpi_get_table(ACPI_SIG_DMAR, 0, &dmar_tbl);
198 if (ACPI_SUCCESS(status)) {
199 printk(KERN_INFO "BIOS BUG: DMAR advertised on Intel G31/G33 chipset -- ignoring\n");
200 dmar_disabled = 1;
201 }
202}
203#endif
204
205#define QFLAG_APPLY_ONCE 0x1 191#define QFLAG_APPLY_ONCE 0x1
206#define QFLAG_APPLIED 0x2 192#define QFLAG_APPLIED 0x2
207#define QFLAG_DONE (QFLAG_APPLY_ONCE|QFLAG_APPLIED) 193#define QFLAG_DONE (QFLAG_APPLY_ONCE|QFLAG_APPLIED)
@@ -225,10 +211,6 @@ static struct chipset early_qrk[] __initdata = {
225 PCI_CLASS_SERIAL_SMBUS, PCI_ANY_ID, 0, ati_bugs }, 211 PCI_CLASS_SERIAL_SMBUS, PCI_ANY_ID, 0, ati_bugs },
226 { PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_SBX00_SMBUS, 212 { PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_SBX00_SMBUS,
227 PCI_CLASS_SERIAL_SMBUS, PCI_ANY_ID, 0, ati_bugs_contd }, 213 PCI_CLASS_SERIAL_SMBUS, PCI_ANY_ID, 0, ati_bugs_contd },
228#ifdef CONFIG_DMAR
229 { PCI_VENDOR_ID_INTEL, 0x29c0,
230 PCI_CLASS_BRIDGE_HOST, PCI_ANY_ID, 0, intel_g33_dmar },
231#endif
232 {} 214 {}
233}; 215};
234 216
diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S
index dd65143941a8..28b597ef9ca1 100644
--- a/arch/x86/kernel/entry_32.S
+++ b/arch/x86/kernel/entry_32.S
@@ -1149,7 +1149,7 @@ ENDPROC(xen_failsafe_callback)
1149 1149
1150#endif /* CONFIG_XEN */ 1150#endif /* CONFIG_XEN */
1151 1151
1152#ifdef CONFIG_FTRACE 1152#ifdef CONFIG_FUNCTION_TRACER
1153#ifdef CONFIG_DYNAMIC_FTRACE 1153#ifdef CONFIG_DYNAMIC_FTRACE
1154 1154
1155ENTRY(mcount) 1155ENTRY(mcount)
@@ -1204,7 +1204,7 @@ trace:
1204 jmp ftrace_stub 1204 jmp ftrace_stub
1205END(mcount) 1205END(mcount)
1206#endif /* CONFIG_DYNAMIC_FTRACE */ 1206#endif /* CONFIG_DYNAMIC_FTRACE */
1207#endif /* CONFIG_FTRACE */ 1207#endif /* CONFIG_FUNCTION_TRACER */
1208 1208
1209.section .rodata,"a" 1209.section .rodata,"a"
1210#include "syscall_table_32.S" 1210#include "syscall_table_32.S"
diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S
index 09e7145484c5..b86f332c96a6 100644
--- a/arch/x86/kernel/entry_64.S
+++ b/arch/x86/kernel/entry_64.S
@@ -61,7 +61,7 @@
61 61
62 .code64 62 .code64
63 63
64#ifdef CONFIG_FTRACE 64#ifdef CONFIG_FUNCTION_TRACER
65#ifdef CONFIG_DYNAMIC_FTRACE 65#ifdef CONFIG_DYNAMIC_FTRACE
66ENTRY(mcount) 66ENTRY(mcount)
67 retq 67 retq
@@ -138,7 +138,7 @@ trace:
138 jmp ftrace_stub 138 jmp ftrace_stub
139END(mcount) 139END(mcount)
140#endif /* CONFIG_DYNAMIC_FTRACE */ 140#endif /* CONFIG_DYNAMIC_FTRACE */
141#endif /* CONFIG_FTRACE */ 141#endif /* CONFIG_FUNCTION_TRACER */
142 142
143#ifndef CONFIG_PREEMPT 143#ifndef CONFIG_PREEMPT
144#define retint_kernel retint_restore_args 144#define retint_kernel retint_restore_args
diff --git a/arch/x86/kernel/es7000_32.c b/arch/x86/kernel/es7000_32.c
index f454c78fcef6..0aa2c443d600 100644
--- a/arch/x86/kernel/es7000_32.c
+++ b/arch/x86/kernel/es7000_32.c
@@ -250,31 +250,24 @@ int __init find_unisys_acpi_oem_table(unsigned long *oem_addr)
250{ 250{
251 struct acpi_table_header *header = NULL; 251 struct acpi_table_header *header = NULL;
252 int i = 0; 252 int i = 0;
253 acpi_size tbl_size;
254 253
255 while (ACPI_SUCCESS(acpi_get_table_with_size("OEM1", i++, &header, &tbl_size))) { 254 while (ACPI_SUCCESS(acpi_get_table("OEM1", i++, &header))) {
256 if (!memcmp((char *) &header->oem_id, "UNISYS", 6)) { 255 if (!memcmp((char *) &header->oem_id, "UNISYS", 6)) {
257 struct oem_table *t = (struct oem_table *)header; 256 struct oem_table *t = (struct oem_table *)header;
258 257
259 oem_addrX = t->OEMTableAddr; 258 oem_addrX = t->OEMTableAddr;
260 oem_size = t->OEMTableSize; 259 oem_size = t->OEMTableSize;
261 early_acpi_os_unmap_memory(header, tbl_size);
262 260
263 *oem_addr = (unsigned long)__acpi_map_table(oem_addrX, 261 *oem_addr = (unsigned long)__acpi_map_table(oem_addrX,
264 oem_size); 262 oem_size);
265 return 0; 263 return 0;
266 } 264 }
267 early_acpi_os_unmap_memory(header, tbl_size);
268 } 265 }
269 return -1; 266 return -1;
270} 267}
271 268
272void __init unmap_unisys_acpi_oem_table(unsigned long oem_addr) 269void __init unmap_unisys_acpi_oem_table(unsigned long oem_addr)
273{ 270{
274 if (!oem_addr)
275 return;
276
277 __acpi_unmap_table((char *)oem_addr, oem_size);
278} 271}
279#endif 272#endif
280 273
diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c
index d073d981a730..50ea0ac8c9bf 100644
--- a/arch/x86/kernel/ftrace.c
+++ b/arch/x86/kernel/ftrace.c
@@ -21,8 +21,7 @@
21#include <asm/nops.h> 21#include <asm/nops.h>
22 22
23 23
24/* Long is fine, even if it is only 4 bytes ;-) */ 24static unsigned char ftrace_nop[MCOUNT_INSN_SIZE];
25static unsigned long *ftrace_nop;
26 25
27union ftrace_code_union { 26union ftrace_code_union {
28 char code[MCOUNT_INSN_SIZE]; 27 char code[MCOUNT_INSN_SIZE];
@@ -33,17 +32,17 @@ union ftrace_code_union {
33}; 32};
34 33
35 34
36static int notrace ftrace_calc_offset(long ip, long addr) 35static int ftrace_calc_offset(long ip, long addr)
37{ 36{
38 return (int)(addr - ip); 37 return (int)(addr - ip);
39} 38}
40 39
41notrace unsigned char *ftrace_nop_replace(void) 40unsigned char *ftrace_nop_replace(void)
42{ 41{
43 return (char *)ftrace_nop; 42 return ftrace_nop;
44} 43}
45 44
46notrace unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr) 45unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr)
47{ 46{
48 static union ftrace_code_union calc; 47 static union ftrace_code_union calc;
49 48
@@ -57,7 +56,7 @@ notrace unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr)
57 return calc.code; 56 return calc.code;
58} 57}
59 58
60notrace int 59int
61ftrace_modify_code(unsigned long ip, unsigned char *old_code, 60ftrace_modify_code(unsigned long ip, unsigned char *old_code,
62 unsigned char *new_code) 61 unsigned char *new_code)
63{ 62{
@@ -66,26 +65,31 @@ ftrace_modify_code(unsigned long ip, unsigned char *old_code,
66 /* 65 /*
67 * Note: Due to modules and __init, code can 66 * Note: Due to modules and __init, code can
68 * disappear and change, we need to protect against faulting 67 * disappear and change, we need to protect against faulting
69 * as well as code changing. 68 * as well as code changing. We do this by using the
69 * probe_kernel_* functions.
70 * 70 *
71 * No real locking needed, this code is run through 71 * No real locking needed, this code is run through
72 * kstop_machine, or before SMP starts. 72 * kstop_machine, or before SMP starts.
73 */ 73 */
74 if (__copy_from_user_inatomic(replaced, (char __user *)ip, MCOUNT_INSN_SIZE))
75 return 1;
76 74
75 /* read the text we want to modify */
76 if (probe_kernel_read(replaced, (void *)ip, MCOUNT_INSN_SIZE))
77 return -EFAULT;
78
79 /* Make sure it is what we expect it to be */
77 if (memcmp(replaced, old_code, MCOUNT_INSN_SIZE) != 0) 80 if (memcmp(replaced, old_code, MCOUNT_INSN_SIZE) != 0)
78 return 2; 81 return -EINVAL;
79 82
80 WARN_ON_ONCE(__copy_to_user_inatomic((char __user *)ip, new_code, 83 /* replace the text with the new text */
81 MCOUNT_INSN_SIZE)); 84 if (probe_kernel_write((void *)ip, new_code, MCOUNT_INSN_SIZE))
85 return -EPERM;
82 86
83 sync_core(); 87 sync_core();
84 88
85 return 0; 89 return 0;
86} 90}
87 91
88notrace int ftrace_update_ftrace_func(ftrace_func_t func) 92int ftrace_update_ftrace_func(ftrace_func_t func)
89{ 93{
90 unsigned long ip = (unsigned long)(&ftrace_call); 94 unsigned long ip = (unsigned long)(&ftrace_call);
91 unsigned char old[MCOUNT_INSN_SIZE], *new; 95 unsigned char old[MCOUNT_INSN_SIZE], *new;
@@ -98,13 +102,6 @@ notrace int ftrace_update_ftrace_func(ftrace_func_t func)
98 return ret; 102 return ret;
99} 103}
100 104
101notrace int ftrace_mcount_set(unsigned long *data)
102{
103 /* mcount is initialized as a nop */
104 *data = 0;
105 return 0;
106}
107
108int __init ftrace_dyn_arch_init(void *data) 105int __init ftrace_dyn_arch_init(void *data)
109{ 106{
110 extern const unsigned char ftrace_test_p6nop[]; 107 extern const unsigned char ftrace_test_p6nop[];
@@ -127,9 +124,6 @@ int __init ftrace_dyn_arch_init(void *data)
127 * TODO: check the cpuid to determine the best nop. 124 * TODO: check the cpuid to determine the best nop.
128 */ 125 */
129 asm volatile ( 126 asm volatile (
130 "jmp ftrace_test_jmp\n"
131 /* This code needs to stay around */
132 ".section .text, \"ax\"\n"
133 "ftrace_test_jmp:" 127 "ftrace_test_jmp:"
134 "jmp ftrace_test_p6nop\n" 128 "jmp ftrace_test_p6nop\n"
135 "nop\n" 129 "nop\n"
@@ -140,8 +134,6 @@ int __init ftrace_dyn_arch_init(void *data)
140 "jmp 1f\n" 134 "jmp 1f\n"
141 "ftrace_test_nop5:" 135 "ftrace_test_nop5:"
142 ".byte 0x66,0x66,0x66,0x66,0x90\n" 136 ".byte 0x66,0x66,0x66,0x66,0x90\n"
143 "jmp 1f\n"
144 ".previous\n"
145 "1:" 137 "1:"
146 ".section .fixup, \"ax\"\n" 138 ".section .fixup, \"ax\"\n"
147 "2: movl $1, %0\n" 139 "2: movl $1, %0\n"
@@ -156,15 +148,15 @@ int __init ftrace_dyn_arch_init(void *data)
156 switch (faulted) { 148 switch (faulted) {
157 case 0: 149 case 0:
158 pr_info("ftrace: converting mcount calls to 0f 1f 44 00 00\n"); 150 pr_info("ftrace: converting mcount calls to 0f 1f 44 00 00\n");
159 ftrace_nop = (unsigned long *)ftrace_test_p6nop; 151 memcpy(ftrace_nop, ftrace_test_p6nop, MCOUNT_INSN_SIZE);
160 break; 152 break;
161 case 1: 153 case 1:
162 pr_info("ftrace: converting mcount calls to 66 66 66 66 90\n"); 154 pr_info("ftrace: converting mcount calls to 66 66 66 66 90\n");
163 ftrace_nop = (unsigned long *)ftrace_test_nop5; 155 memcpy(ftrace_nop, ftrace_test_nop5, MCOUNT_INSN_SIZE);
164 break; 156 break;
165 case 2: 157 case 2:
166 pr_info("ftrace: converting mcount calls to jmp . + 5\n"); 158 pr_info("ftrace: converting mcount calls to jmp . + 5\n");
167 ftrace_nop = (unsigned long *)ftrace_test_jmp; 159 memcpy(ftrace_nop, ftrace_test_jmp, MCOUNT_INSN_SIZE);
168 break; 160 break;
169 } 161 }
170 162
diff --git a/arch/x86/kernel/genx2apic_uv_x.c b/arch/x86/kernel/genx2apic_uv_x.c
index 680a06557c5e..2c7dbdb98278 100644
--- a/arch/x86/kernel/genx2apic_uv_x.c
+++ b/arch/x86/kernel/genx2apic_uv_x.c
@@ -15,7 +15,6 @@
15#include <linux/ctype.h> 15#include <linux/ctype.h>
16#include <linux/init.h> 16#include <linux/init.h>
17#include <linux/sched.h> 17#include <linux/sched.h>
18#include <linux/bootmem.h>
19#include <linux/module.h> 18#include <linux/module.h>
20#include <linux/hardirq.h> 19#include <linux/hardirq.h>
21#include <asm/smp.h> 20#include <asm/smp.h>
@@ -398,16 +397,16 @@ void __init uv_system_init(void)
398 printk(KERN_DEBUG "UV: Found %d blades\n", uv_num_possible_blades()); 397 printk(KERN_DEBUG "UV: Found %d blades\n", uv_num_possible_blades());
399 398
400 bytes = sizeof(struct uv_blade_info) * uv_num_possible_blades(); 399 bytes = sizeof(struct uv_blade_info) * uv_num_possible_blades();
401 uv_blade_info = alloc_bootmem_pages(bytes); 400 uv_blade_info = kmalloc(bytes, GFP_KERNEL);
402 401
403 get_lowmem_redirect(&lowmem_redir_base, &lowmem_redir_size); 402 get_lowmem_redirect(&lowmem_redir_base, &lowmem_redir_size);
404 403
405 bytes = sizeof(uv_node_to_blade[0]) * num_possible_nodes(); 404 bytes = sizeof(uv_node_to_blade[0]) * num_possible_nodes();
406 uv_node_to_blade = alloc_bootmem_pages(bytes); 405 uv_node_to_blade = kmalloc(bytes, GFP_KERNEL);
407 memset(uv_node_to_blade, 255, bytes); 406 memset(uv_node_to_blade, 255, bytes);
408 407
409 bytes = sizeof(uv_cpu_to_blade[0]) * num_possible_cpus(); 408 bytes = sizeof(uv_cpu_to_blade[0]) * num_possible_cpus();
410 uv_cpu_to_blade = alloc_bootmem_pages(bytes); 409 uv_cpu_to_blade = kmalloc(bytes, GFP_KERNEL);
411 memset(uv_cpu_to_blade, 255, bytes); 410 memset(uv_cpu_to_blade, 255, bytes);
412 411
413 blade = 0; 412 blade = 0;
diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c
index 77017e834cf7..067d8de913f6 100644
--- a/arch/x86/kernel/hpet.c
+++ b/arch/x86/kernel/hpet.c
@@ -322,7 +322,7 @@ static int hpet_next_event(unsigned long delta,
322 * what we wrote hit the chip before we compare it to the 322 * what we wrote hit the chip before we compare it to the
323 * counter. 323 * counter.
324 */ 324 */
325 WARN_ON((u32)hpet_readl(HPET_T0_CMP) != cnt); 325 WARN_ON_ONCE((u32)hpet_readl(HPET_Tn_CMP(timer)) != cnt);
326 326
327 return (s32)((u32)hpet_readl(HPET_COUNTER) - cnt) >= 0 ? -ETIME : 0; 327 return (s32)((u32)hpet_readl(HPET_COUNTER) - cnt) >= 0 ? -ETIME : 0;
328} 328}
@@ -445,7 +445,7 @@ static int hpet_setup_irq(struct hpet_dev *dev)
445{ 445{
446 446
447 if (request_irq(dev->irq, hpet_interrupt_handler, 447 if (request_irq(dev->irq, hpet_interrupt_handler,
448 IRQF_SHARED|IRQF_NOBALANCING, dev->name, dev)) 448 IRQF_DISABLED|IRQF_NOBALANCING, dev->name, dev))
449 return -1; 449 return -1;
450 450
451 disable_irq(dev->irq); 451 disable_irq(dev->irq);
diff --git a/arch/x86/kernel/i386_ksyms_32.c b/arch/x86/kernel/i386_ksyms_32.c
index dd7ebee446af..43cec6bdda63 100644
--- a/arch/x86/kernel/i386_ksyms_32.c
+++ b/arch/x86/kernel/i386_ksyms_32.c
@@ -5,7 +5,7 @@
5#include <asm/desc.h> 5#include <asm/desc.h>
6#include <asm/ftrace.h> 6#include <asm/ftrace.h>
7 7
8#ifdef CONFIG_FTRACE 8#ifdef CONFIG_FUNCTION_TRACER
9/* mcount is defined in assembly */ 9/* mcount is defined in assembly */
10EXPORT_SYMBOL(mcount); 10EXPORT_SYMBOL(mcount);
11#endif 11#endif
diff --git a/arch/x86/kernel/i387.c b/arch/x86/kernel/i387.c
index 1f20608d4ca8..b0f61f0dcd0a 100644
--- a/arch/x86/kernel/i387.c
+++ b/arch/x86/kernel/i387.c
@@ -58,7 +58,7 @@ void __cpuinit mxcsr_feature_mask_init(void)
58 stts(); 58 stts();
59} 59}
60 60
61void __init init_thread_xstate(void) 61void __cpuinit init_thread_xstate(void)
62{ 62{
63 if (!HAVE_HWFP) { 63 if (!HAVE_HWFP) {
64 xstate_size = sizeof(struct i387_soft_struct); 64 xstate_size = sizeof(struct i387_soft_struct);
diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c
index b764d7429c61..9043251210fb 100644
--- a/arch/x86/kernel/io_apic.c
+++ b/arch/x86/kernel/io_apic.c
@@ -1140,6 +1140,20 @@ static void __clear_irq_vector(int irq)
1140 1140
1141 cfg->vector = 0; 1141 cfg->vector = 0;
1142 cpus_clear(cfg->domain); 1142 cpus_clear(cfg->domain);
1143
1144 if (likely(!cfg->move_in_progress))
1145 return;
1146 cpus_and(mask, cfg->old_domain, cpu_online_map);
1147 for_each_cpu_mask_nr(cpu, mask) {
1148 for (vector = FIRST_EXTERNAL_VECTOR; vector < NR_VECTORS;
1149 vector++) {
1150 if (per_cpu(vector_irq, cpu)[vector] != irq)
1151 continue;
1152 per_cpu(vector_irq, cpu)[vector] = -1;
1153 break;
1154 }
1155 }
1156 cfg->move_in_progress = 0;
1143} 1157}
1144 1158
1145void __setup_vector_irq(int cpu) 1159void __setup_vector_irq(int cpu)
@@ -3594,25 +3608,7 @@ int __init io_apic_get_redir_entries (int ioapic)
3594 3608
3595int __init probe_nr_irqs(void) 3609int __init probe_nr_irqs(void)
3596{ 3610{
3597 int idx; 3611 return NR_IRQS;
3598 int nr = 0;
3599#ifndef CONFIG_XEN
3600 int nr_min = 32;
3601#else
3602 int nr_min = NR_IRQS;
3603#endif
3604
3605 for (idx = 0; idx < nr_ioapics; idx++)
3606 nr += io_apic_get_redir_entries(idx) + 1;
3607
3608 /* double it for hotplug and msi and nmi */
3609 nr <<= 1;
3610
3611 /* something wrong ? */
3612 if (nr < nr_min)
3613 nr = nr_min;
3614
3615 return nr;
3616} 3612}
3617 3613
3618/* -------------------------------------------------------------------------- 3614/* --------------------------------------------------------------------------
@@ -3759,7 +3755,9 @@ int acpi_get_override_irq(int bus_irq, int *trigger, int *polarity)
3759void __init setup_ioapic_dest(void) 3755void __init setup_ioapic_dest(void)
3760{ 3756{
3761 int pin, ioapic, irq, irq_entry; 3757 int pin, ioapic, irq, irq_entry;
3758 struct irq_desc *desc;
3762 struct irq_cfg *cfg; 3759 struct irq_cfg *cfg;
3760 cpumask_t mask;
3763 3761
3764 if (skip_ioapic_setup == 1) 3762 if (skip_ioapic_setup == 1)
3765 return; 3763 return;
@@ -3776,16 +3774,30 @@ void __init setup_ioapic_dest(void)
3776 * cpu is online. 3774 * cpu is online.
3777 */ 3775 */
3778 cfg = irq_cfg(irq); 3776 cfg = irq_cfg(irq);
3779 if (!cfg->vector) 3777 if (!cfg->vector) {
3780 setup_IO_APIC_irq(ioapic, pin, irq, 3778 setup_IO_APIC_irq(ioapic, pin, irq,
3781 irq_trigger(irq_entry), 3779 irq_trigger(irq_entry),
3782 irq_polarity(irq_entry)); 3780 irq_polarity(irq_entry));
3781 continue;
3782
3783 }
3784
3785 /*
3786 * Honour affinities which have been set in early boot
3787 */
3788 desc = irq_to_desc(irq);
3789 if (desc->status &
3790 (IRQ_NO_BALANCING | IRQ_AFFINITY_SET))
3791 mask = desc->affinity;
3792 else
3793 mask = TARGET_CPUS;
3794
3783#ifdef CONFIG_INTR_REMAP 3795#ifdef CONFIG_INTR_REMAP
3784 else if (intr_remapping_enabled) 3796 if (intr_remapping_enabled)
3785 set_ir_ioapic_affinity_irq(irq, TARGET_CPUS); 3797 set_ir_ioapic_affinity_irq(irq, mask);
3786#endif
3787 else 3798 else
3788 set_ioapic_affinity_irq(irq, TARGET_CPUS); 3799#endif
3800 set_ioapic_affinity_irq(irq, mask);
3789 } 3801 }
3790 3802
3791 } 3803 }
diff --git a/arch/x86/kernel/k8.c b/arch/x86/kernel/k8.c
index 304d8bad6559..cbc4332a77b2 100644
--- a/arch/x86/kernel/k8.c
+++ b/arch/x86/kernel/k8.c
@@ -18,7 +18,6 @@ static u32 *flush_words;
18struct pci_device_id k8_nb_ids[] = { 18struct pci_device_id k8_nb_ids[] = {
19 { PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_K8_NB_MISC) }, 19 { PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_K8_NB_MISC) },
20 { PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_10H_NB_MISC) }, 20 { PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_10H_NB_MISC) },
21 { PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_11H_NB_MISC) },
22 {} 21 {}
23}; 22};
24EXPORT_SYMBOL(k8_nb_ids); 23EXPORT_SYMBOL(k8_nb_ids);
diff --git a/arch/x86/kernel/kvmclock.c b/arch/x86/kernel/kvmclock.c
index 774ac4991568..e169ae9b6a62 100644
--- a/arch/x86/kernel/kvmclock.c
+++ b/arch/x86/kernel/kvmclock.c
@@ -128,7 +128,7 @@ static int kvm_register_clock(char *txt)
128} 128}
129 129
130#ifdef CONFIG_X86_LOCAL_APIC 130#ifdef CONFIG_X86_LOCAL_APIC
131static void kvm_setup_secondary_clock(void) 131static void __cpuinit kvm_setup_secondary_clock(void)
132{ 132{
133 /* 133 /*
134 * Now that the first cpu already had this clocksource initialized, 134 * Now that the first cpu already had this clocksource initialized,
diff --git a/arch/x86/kernel/machine_kexec_32.c b/arch/x86/kernel/machine_kexec_32.c
index 0732adba05ca..7a385746509a 100644
--- a/arch/x86/kernel/machine_kexec_32.c
+++ b/arch/x86/kernel/machine_kexec_32.c
@@ -162,7 +162,10 @@ void machine_kexec(struct kimage *image)
162 page_list[VA_PTE_0] = (unsigned long)kexec_pte0; 162 page_list[VA_PTE_0] = (unsigned long)kexec_pte0;
163 page_list[PA_PTE_1] = __pa(kexec_pte1); 163 page_list[PA_PTE_1] = __pa(kexec_pte1);
164 page_list[VA_PTE_1] = (unsigned long)kexec_pte1; 164 page_list[VA_PTE_1] = (unsigned long)kexec_pte1;
165 page_list[PA_SWAP_PAGE] = (page_to_pfn(image->swap_page) << PAGE_SHIFT); 165
166 if (image->type == KEXEC_TYPE_DEFAULT)
167 page_list[PA_SWAP_PAGE] = (page_to_pfn(image->swap_page)
168 << PAGE_SHIFT);
166 169
167 /* The segment registers are funny things, they have both a 170 /* The segment registers are funny things, they have both a
168 * visible and an invisible part. Whenever the visible part is 171 * visible and an invisible part. Whenever the visible part is
diff --git a/arch/x86/kernel/microcode_amd.c b/arch/x86/kernel/microcode_amd.c
index 7a1f8eeac2c7..5f8e5d75a254 100644
--- a/arch/x86/kernel/microcode_amd.c
+++ b/arch/x86/kernel/microcode_amd.c
@@ -39,7 +39,7 @@
39#include <asm/microcode.h> 39#include <asm/microcode.h>
40 40
41MODULE_DESCRIPTION("AMD Microcode Update Driver"); 41MODULE_DESCRIPTION("AMD Microcode Update Driver");
42MODULE_AUTHOR("Peter Oruba <peter.oruba@amd.com>"); 42MODULE_AUTHOR("Peter Oruba");
43MODULE_LICENSE("GPL v2"); 43MODULE_LICENSE("GPL v2");
44 44
45#define UCODE_MAGIC 0x00414d44 45#define UCODE_MAGIC 0x00414d44
diff --git a/arch/x86/kernel/microcode_core.c b/arch/x86/kernel/microcode_core.c
index 936d8d55f230..82fb2809ce32 100644
--- a/arch/x86/kernel/microcode_core.c
+++ b/arch/x86/kernel/microcode_core.c
@@ -480,8 +480,8 @@ static int __init microcode_init(void)
480 480
481 printk(KERN_INFO 481 printk(KERN_INFO
482 "Microcode Update Driver: v" MICROCODE_VERSION 482 "Microcode Update Driver: v" MICROCODE_VERSION
483 " <tigran@aivazian.fsnet.co.uk>" 483 " <tigran@aivazian.fsnet.co.uk>,"
484 " <peter.oruba@amd.com>\n"); 484 " Peter Oruba\n");
485 485
486 return 0; 486 return 0;
487} 487}
diff --git a/arch/x86/kernel/mpparse.c b/arch/x86/kernel/mpparse.c
index f98f4e1dba09..0f4c1fd5a1f4 100644
--- a/arch/x86/kernel/mpparse.c
+++ b/arch/x86/kernel/mpparse.c
@@ -604,6 +604,9 @@ static void __init __get_smp_config(unsigned int early)
604 printk(KERN_INFO "Using ACPI for processor (LAPIC) " 604 printk(KERN_INFO "Using ACPI for processor (LAPIC) "
605 "configuration information\n"); 605 "configuration information\n");
606 606
607 if (!mpf)
608 return;
609
607 printk(KERN_INFO "Intel MultiProcessor Specification v1.%d\n", 610 printk(KERN_INFO "Intel MultiProcessor Specification v1.%d\n",
608 mpf->mpf_specification); 611 mpf->mpf_specification);
609#if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86_32) 612#if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86_32)
diff --git a/arch/x86/kernel/paravirt-spinlocks.c b/arch/x86/kernel/paravirt-spinlocks.c
index 0e9f1982b1dd..95777b0faa73 100644
--- a/arch/x86/kernel/paravirt-spinlocks.c
+++ b/arch/x86/kernel/paravirt-spinlocks.c
@@ -7,7 +7,8 @@
7 7
8#include <asm/paravirt.h> 8#include <asm/paravirt.h>
9 9
10static void default_spin_lock_flags(struct raw_spinlock *lock, unsigned long flags) 10static inline void
11default_spin_lock_flags(raw_spinlock_t *lock, unsigned long flags)
11{ 12{
12 __raw_spin_lock(lock); 13 __raw_spin_lock(lock);
13} 14}
diff --git a/arch/x86/kernel/pci-calgary_64.c b/arch/x86/kernel/pci-calgary_64.c
index e1e731d78f38..d28bbdc35e4e 100644
--- a/arch/x86/kernel/pci-calgary_64.c
+++ b/arch/x86/kernel/pci-calgary_64.c
@@ -1567,7 +1567,7 @@ static int __init calgary_parse_options(char *p)
1567 ++p; 1567 ++p;
1568 if (*p == '\0') 1568 if (*p == '\0')
1569 break; 1569 break;
1570 bridge = simple_strtol(p, &endp, 0); 1570 bridge = simple_strtoul(p, &endp, 0);
1571 if (p == endp) 1571 if (p == endp)
1572 break; 1572 break;
1573 1573
diff --git a/arch/x86/kernel/pci-dma.c b/arch/x86/kernel/pci-dma.c
index 1972266e8ba5..192624820217 100644
--- a/arch/x86/kernel/pci-dma.c
+++ b/arch/x86/kernel/pci-dma.c
@@ -9,6 +9,8 @@
9#include <asm/calgary.h> 9#include <asm/calgary.h>
10#include <asm/amd_iommu.h> 10#include <asm/amd_iommu.h>
11 11
12static int forbid_dac __read_mostly;
13
12struct dma_mapping_ops *dma_ops; 14struct dma_mapping_ops *dma_ops;
13EXPORT_SYMBOL(dma_ops); 15EXPORT_SYMBOL(dma_ops);
14 16
@@ -291,3 +293,17 @@ void pci_iommu_shutdown(void)
291} 293}
292/* Must execute after PCI subsystem */ 294/* Must execute after PCI subsystem */
293fs_initcall(pci_iommu_init); 295fs_initcall(pci_iommu_init);
296
297#ifdef CONFIG_PCI
298/* Many VIA bridges seem to corrupt data for DAC. Disable it here */
299
300static __devinit void via_no_dac(struct pci_dev *dev)
301{
302 if ((dev->class >> 8) == PCI_CLASS_BRIDGE_PCI && forbid_dac == 0) {
303 printk(KERN_INFO "PCI: VIA PCI bridge detected."
304 "Disabling DAC.\n");
305 forbid_dac = 1;
306 }
307}
308DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_VIA, PCI_ANY_ID, via_no_dac);
309#endif
diff --git a/arch/x86/kernel/pci-gart_64.c b/arch/x86/kernel/pci-gart_64.c
index e3f75bbcedea..ba7ad83e20a8 100644
--- a/arch/x86/kernel/pci-gart_64.c
+++ b/arch/x86/kernel/pci-gart_64.c
@@ -123,6 +123,8 @@ static void free_iommu(unsigned long offset, int size)
123 123
124 spin_lock_irqsave(&iommu_bitmap_lock, flags); 124 spin_lock_irqsave(&iommu_bitmap_lock, flags);
125 iommu_area_free(iommu_gart_bitmap, offset, size); 125 iommu_area_free(iommu_gart_bitmap, offset, size);
126 if (offset >= next_bit)
127 next_bit = offset + size;
126 spin_unlock_irqrestore(&iommu_bitmap_lock, flags); 128 spin_unlock_irqrestore(&iommu_bitmap_lock, flags);
127} 129}
128 130
@@ -744,7 +746,7 @@ void __init gart_iommu_init(void)
744 long i; 746 long i;
745 747
746 if (cache_k8_northbridges() < 0 || num_k8_northbridges == 0) { 748 if (cache_k8_northbridges() < 0 || num_k8_northbridges == 0) {
747 printk(KERN_INFO "PCI-GART: No AMD northbridge found.\n"); 749 printk(KERN_INFO "PCI-GART: No AMD GART found.\n");
748 return; 750 return;
749 } 751 }
750 752
diff --git a/arch/x86/kernel/pci-swiotlb_64.c b/arch/x86/kernel/pci-swiotlb_64.c
index c4ce0332759e..3c539d111abb 100644
--- a/arch/x86/kernel/pci-swiotlb_64.c
+++ b/arch/x86/kernel/pci-swiotlb_64.c
@@ -18,9 +18,21 @@ swiotlb_map_single_phys(struct device *hwdev, phys_addr_t paddr, size_t size,
18 return swiotlb_map_single(hwdev, phys_to_virt(paddr), size, direction); 18 return swiotlb_map_single(hwdev, phys_to_virt(paddr), size, direction);
19} 19}
20 20
21static void *x86_swiotlb_alloc_coherent(struct device *hwdev, size_t size,
22 dma_addr_t *dma_handle, gfp_t flags)
23{
24 void *vaddr;
25
26 vaddr = dma_generic_alloc_coherent(hwdev, size, dma_handle, flags);
27 if (vaddr)
28 return vaddr;
29
30 return swiotlb_alloc_coherent(hwdev, size, dma_handle, flags);
31}
32
21struct dma_mapping_ops swiotlb_dma_ops = { 33struct dma_mapping_ops swiotlb_dma_ops = {
22 .mapping_error = swiotlb_dma_mapping_error, 34 .mapping_error = swiotlb_dma_mapping_error,
23 .alloc_coherent = swiotlb_alloc_coherent, 35 .alloc_coherent = x86_swiotlb_alloc_coherent,
24 .free_coherent = swiotlb_free_coherent, 36 .free_coherent = swiotlb_free_coherent,
25 .map_single = swiotlb_map_single_phys, 37 .map_single = swiotlb_map_single_phys,
26 .unmap_single = swiotlb_unmap_single, 38 .unmap_single = swiotlb_unmap_single,
diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c
index f4c93f1cfc19..cc5a2545dd41 100644
--- a/arch/x86/kernel/reboot.c
+++ b/arch/x86/kernel/reboot.c
@@ -29,11 +29,7 @@ EXPORT_SYMBOL(pm_power_off);
29 29
30static const struct desc_ptr no_idt = {}; 30static const struct desc_ptr no_idt = {};
31static int reboot_mode; 31static int reboot_mode;
32/* 32enum reboot_type reboot_type = BOOT_KBD;
33 * Keyboard reset and triple fault may result in INIT, not RESET, which
34 * doesn't work when we're in vmx root mode. Try ACPI first.
35 */
36enum reboot_type reboot_type = BOOT_ACPI;
37int reboot_force; 33int reboot_force;
38 34
39#if defined(CONFIG_X86_32) && defined(CONFIG_SMP) 35#if defined(CONFIG_X86_32) && defined(CONFIG_SMP)
@@ -173,6 +169,15 @@ static struct dmi_system_id __initdata reboot_dmi_table[] = {
173 DMI_MATCH(DMI_BOARD_NAME, "0KW626"), 169 DMI_MATCH(DMI_BOARD_NAME, "0KW626"),
174 }, 170 },
175 }, 171 },
172 { /* Handle problems with rebooting on Dell Optiplex 330 with 0KP561 */
173 .callback = set_bios_reboot,
174 .ident = "Dell OptiPlex 330",
175 .matches = {
176 DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
177 DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex 330"),
178 DMI_MATCH(DMI_BOARD_NAME, "0KP561"),
179 },
180 },
176 { /* Handle problems with rebooting on Dell 2400's */ 181 { /* Handle problems with rebooting on Dell 2400's */
177 .callback = set_bios_reboot, 182 .callback = set_bios_reboot,
178 .ident = "Dell PowerEdge 2400", 183 .ident = "Dell PowerEdge 2400",
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 0fa6790c1dd3..9d5674f7b6cc 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -764,7 +764,7 @@ static struct dmi_system_id __initdata bad_bios_dmi_table[] = {
764 .callback = dmi_low_memory_corruption, 764 .callback = dmi_low_memory_corruption,
765 .ident = "Phoenix BIOS", 765 .ident = "Phoenix BIOS",
766 .matches = { 766 .matches = {
767 DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies, LTD"), 767 DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies"),
768 }, 768 },
769 }, 769 },
770#endif 770#endif
diff --git a/arch/x86/kernel/tlb_32.c b/arch/x86/kernel/tlb_32.c
index e00534b33534..f4049f3513b6 100644
--- a/arch/x86/kernel/tlb_32.c
+++ b/arch/x86/kernel/tlb_32.c
@@ -154,6 +154,12 @@ void native_flush_tlb_others(const cpumask_t *cpumaskp, struct mm_struct *mm,
154 flush_mm = mm; 154 flush_mm = mm;
155 flush_va = va; 155 flush_va = va;
156 cpus_or(flush_cpumask, cpumask, flush_cpumask); 156 cpus_or(flush_cpumask, cpumask, flush_cpumask);
157
158 /*
159 * Make the above memory operations globally visible before
160 * sending the IPI.
161 */
162 smp_mb();
157 /* 163 /*
158 * We have to send the IPI only to 164 * We have to send the IPI only to
159 * CPUs affected. 165 * CPUs affected.
diff --git a/arch/x86/kernel/tlb_64.c b/arch/x86/kernel/tlb_64.c
index dcbf7a1159ea..8f919ca69494 100644
--- a/arch/x86/kernel/tlb_64.c
+++ b/arch/x86/kernel/tlb_64.c
@@ -183,6 +183,11 @@ void native_flush_tlb_others(const cpumask_t *cpumaskp, struct mm_struct *mm,
183 cpus_or(f->flush_cpumask, cpumask, f->flush_cpumask); 183 cpus_or(f->flush_cpumask, cpumask, f->flush_cpumask);
184 184
185 /* 185 /*
186 * Make the above memory operations globally visible before
187 * sending the IPI.
188 */
189 smp_mb();
190 /*
186 * We have to send the IPI only to 191 * We have to send the IPI only to
187 * CPUs affected. 192 * CPUs affected.
188 */ 193 */
diff --git a/arch/x86/kernel/tsc.c b/arch/x86/kernel/tsc.c
index 161bb850fc47..424093b157d3 100644
--- a/arch/x86/kernel/tsc.c
+++ b/arch/x86/kernel/tsc.c
@@ -55,7 +55,7 @@ u64 native_sched_clock(void)
55 rdtscll(this_offset); 55 rdtscll(this_offset);
56 56
57 /* return the value in ns */ 57 /* return the value in ns */
58 return cycles_2_ns(this_offset); 58 return __cycles_2_ns(this_offset);
59} 59}
60 60
61/* We need to define a real function for sched_clock, to override the 61/* We need to define a real function for sched_clock, to override the
@@ -759,7 +759,7 @@ __cpuinit int unsynchronized_tsc(void)
759 if (!cpu_has_tsc || tsc_unstable) 759 if (!cpu_has_tsc || tsc_unstable)
760 return 1; 760 return 1;
761 761
762#ifdef CONFIG_SMP 762#ifdef CONFIG_X86_SMP
763 if (apic_is_clustered_box()) 763 if (apic_is_clustered_box())
764 return 1; 764 return 1;
765#endif 765#endif
@@ -813,10 +813,6 @@ void __init tsc_init(void)
813 cpu_khz = calibrate_cpu(); 813 cpu_khz = calibrate_cpu();
814#endif 814#endif
815 815
816 lpj = ((u64)tsc_khz * 1000);
817 do_div(lpj, HZ);
818 lpj_fine = lpj;
819
820 printk("Detected %lu.%03lu MHz processor.\n", 816 printk("Detected %lu.%03lu MHz processor.\n",
821 (unsigned long)cpu_khz / 1000, 817 (unsigned long)cpu_khz / 1000,
822 (unsigned long)cpu_khz % 1000); 818 (unsigned long)cpu_khz % 1000);
@@ -836,6 +832,10 @@ void __init tsc_init(void)
836 /* now allow native_sched_clock() to use rdtsc */ 832 /* now allow native_sched_clock() to use rdtsc */
837 tsc_disabled = 0; 833 tsc_disabled = 0;
838 834
835 lpj = ((u64)tsc_khz * 1000);
836 do_div(lpj, HZ);
837 lpj_fine = lpj;
838
839 use_tsc_delay(); 839 use_tsc_delay();
840 /* Check and install the TSC clocksource */ 840 /* Check and install the TSC clocksource */
841 dmi_check_system(bad_tsc_dmi_table); 841 dmi_check_system(bad_tsc_dmi_table);
diff --git a/arch/x86/kernel/tsc_sync.c b/arch/x86/kernel/tsc_sync.c
index 9ffb01c31c40..1c0dfbca87c1 100644
--- a/arch/x86/kernel/tsc_sync.c
+++ b/arch/x86/kernel/tsc_sync.c
@@ -46,7 +46,9 @@ static __cpuinit void check_tsc_warp(void)
46 cycles_t start, now, prev, end; 46 cycles_t start, now, prev, end;
47 int i; 47 int i;
48 48
49 rdtsc_barrier();
49 start = get_cycles(); 50 start = get_cycles();
51 rdtsc_barrier();
50 /* 52 /*
51 * The measurement runs for 20 msecs: 53 * The measurement runs for 20 msecs:
52 */ 54 */
@@ -61,7 +63,9 @@ static __cpuinit void check_tsc_warp(void)
61 */ 63 */
62 __raw_spin_lock(&sync_lock); 64 __raw_spin_lock(&sync_lock);
63 prev = last_tsc; 65 prev = last_tsc;
66 rdtsc_barrier();
64 now = get_cycles(); 67 now = get_cycles();
68 rdtsc_barrier();
65 last_tsc = now; 69 last_tsc = now;
66 __raw_spin_unlock(&sync_lock); 70 __raw_spin_unlock(&sync_lock);
67 71
diff --git a/arch/x86/kernel/vsmp_64.c b/arch/x86/kernel/vsmp_64.c
index 7766d36983fc..a688f3bfaec2 100644
--- a/arch/x86/kernel/vsmp_64.c
+++ b/arch/x86/kernel/vsmp_64.c
@@ -78,7 +78,7 @@ static unsigned __init_or_module vsmp_patch(u8 type, u16 clobbers, void *ibuf,
78 78
79static void __init set_vsmp_pv_ops(void) 79static void __init set_vsmp_pv_ops(void)
80{ 80{
81 void *address; 81 void __iomem *address;
82 unsigned int cap, ctl, cfg; 82 unsigned int cap, ctl, cfg;
83 83
84 /* set vSMP magic bits to indicate vSMP capable kernel */ 84 /* set vSMP magic bits to indicate vSMP capable kernel */
diff --git a/arch/x86/kernel/x8664_ksyms_64.c b/arch/x86/kernel/x8664_ksyms_64.c
index b545f371b5f5..695e426aa354 100644
--- a/arch/x86/kernel/x8664_ksyms_64.c
+++ b/arch/x86/kernel/x8664_ksyms_64.c
@@ -12,7 +12,7 @@
12#include <asm/desc.h> 12#include <asm/desc.h>
13#include <asm/ftrace.h> 13#include <asm/ftrace.h>
14 14
15#ifdef CONFIG_FTRACE 15#ifdef CONFIG_FUNCTION_TRACER
16/* mcount is defined in assembly */ 16/* mcount is defined in assembly */
17EXPORT_SYMBOL(mcount); 17EXPORT_SYMBOL(mcount);
18#endif 18#endif
diff --git a/arch/x86/kernel/xsave.c b/arch/x86/kernel/xsave.c
index b13acb75e822..15c3e6999182 100644
--- a/arch/x86/kernel/xsave.c
+++ b/arch/x86/kernel/xsave.c
@@ -310,7 +310,7 @@ static void __init setup_xstate_init(void)
310/* 310/*
311 * Enable and initialize the xsave feature. 311 * Enable and initialize the xsave feature.
312 */ 312 */
313void __init xsave_cntxt_init(void) 313void __ref xsave_cntxt_init(void)
314{ 314{
315 unsigned int eax, ebx, ecx, edx; 315 unsigned int eax, ebx, ecx, edx;
316 316
diff --git a/arch/x86/kvm/Kconfig b/arch/x86/kvm/Kconfig
index ce3251ce5504..b81125f0bdee 100644
--- a/arch/x86/kvm/Kconfig
+++ b/arch/x86/kvm/Kconfig
@@ -20,6 +20,8 @@ if VIRTUALIZATION
20config KVM 20config KVM
21 tristate "Kernel-based Virtual Machine (KVM) support" 21 tristate "Kernel-based Virtual Machine (KVM) support"
22 depends on HAVE_KVM 22 depends on HAVE_KVM
23 # for device assignment:
24 depends on PCI
23 select PREEMPT_NOTIFIERS 25 select PREEMPT_NOTIFIERS
24 select MMU_NOTIFIER 26 select MMU_NOTIFIER
25 select ANON_INODES 27 select ANON_INODES
diff --git a/arch/x86/kvm/i8254.c b/arch/x86/kvm/i8254.c
index 11c6725fb798..59ebd37ad79e 100644
--- a/arch/x86/kvm/i8254.c
+++ b/arch/x86/kvm/i8254.c
@@ -545,6 +545,14 @@ struct kvm_pit *kvm_create_pit(struct kvm *kvm)
545 if (!pit) 545 if (!pit)
546 return NULL; 546 return NULL;
547 547
548 mutex_lock(&kvm->lock);
549 pit->irq_source_id = kvm_request_irq_source_id(kvm);
550 mutex_unlock(&kvm->lock);
551 if (pit->irq_source_id < 0) {
552 kfree(pit);
553 return NULL;
554 }
555
548 mutex_init(&pit->pit_state.lock); 556 mutex_init(&pit->pit_state.lock);
549 mutex_lock(&pit->pit_state.lock); 557 mutex_lock(&pit->pit_state.lock);
550 spin_lock_init(&pit->pit_state.inject_lock); 558 spin_lock_init(&pit->pit_state.inject_lock);
@@ -587,6 +595,7 @@ void kvm_free_pit(struct kvm *kvm)
587 mutex_lock(&kvm->arch.vpit->pit_state.lock); 595 mutex_lock(&kvm->arch.vpit->pit_state.lock);
588 timer = &kvm->arch.vpit->pit_state.pit_timer.timer; 596 timer = &kvm->arch.vpit->pit_state.pit_timer.timer;
589 hrtimer_cancel(timer); 597 hrtimer_cancel(timer);
598 kvm_free_irq_source_id(kvm, kvm->arch.vpit->irq_source_id);
590 mutex_unlock(&kvm->arch.vpit->pit_state.lock); 599 mutex_unlock(&kvm->arch.vpit->pit_state.lock);
591 kfree(kvm->arch.vpit); 600 kfree(kvm->arch.vpit);
592 } 601 }
@@ -595,8 +604,8 @@ void kvm_free_pit(struct kvm *kvm)
595static void __inject_pit_timer_intr(struct kvm *kvm) 604static void __inject_pit_timer_intr(struct kvm *kvm)
596{ 605{
597 mutex_lock(&kvm->lock); 606 mutex_lock(&kvm->lock);
598 kvm_set_irq(kvm, 0, 1); 607 kvm_set_irq(kvm, kvm->arch.vpit->irq_source_id, 0, 1);
599 kvm_set_irq(kvm, 0, 0); 608 kvm_set_irq(kvm, kvm->arch.vpit->irq_source_id, 0, 0);
600 mutex_unlock(&kvm->lock); 609 mutex_unlock(&kvm->lock);
601} 610}
602 611
diff --git a/arch/x86/kvm/i8254.h b/arch/x86/kvm/i8254.h
index e436d4983aa1..4178022b97aa 100644
--- a/arch/x86/kvm/i8254.h
+++ b/arch/x86/kvm/i8254.h
@@ -44,6 +44,7 @@ struct kvm_pit {
44 struct kvm_io_device speaker_dev; 44 struct kvm_io_device speaker_dev;
45 struct kvm *kvm; 45 struct kvm *kvm;
46 struct kvm_kpit_state pit_state; 46 struct kvm_kpit_state pit_state;
47 int irq_source_id;
47}; 48};
48 49
49#define KVM_PIT_BASE_ADDRESS 0x40 50#define KVM_PIT_BASE_ADDRESS 0x40
diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index 99c239c5c0ac..410ddbc1aa2e 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -314,7 +314,7 @@ static int mmu_topup_memory_caches(struct kvm_vcpu *vcpu)
314 if (r) 314 if (r)
315 goto out; 315 goto out;
316 r = mmu_topup_memory_cache(&vcpu->arch.mmu_rmap_desc_cache, 316 r = mmu_topup_memory_cache(&vcpu->arch.mmu_rmap_desc_cache,
317 rmap_desc_cache, 1); 317 rmap_desc_cache, 4);
318 if (r) 318 if (r)
319 goto out; 319 goto out;
320 r = mmu_topup_memory_cache_page(&vcpu->arch.mmu_page_cache, 8); 320 r = mmu_topup_memory_cache_page(&vcpu->arch.mmu_page_cache, 8);
@@ -1038,13 +1038,13 @@ static int kvm_sync_page(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp)
1038 } 1038 }
1039 1039
1040 rmap_write_protect(vcpu->kvm, sp->gfn); 1040 rmap_write_protect(vcpu->kvm, sp->gfn);
1041 kvm_unlink_unsync_page(vcpu->kvm, sp);
1041 if (vcpu->arch.mmu.sync_page(vcpu, sp)) { 1042 if (vcpu->arch.mmu.sync_page(vcpu, sp)) {
1042 kvm_mmu_zap_page(vcpu->kvm, sp); 1043 kvm_mmu_zap_page(vcpu->kvm, sp);
1043 return 1; 1044 return 1;
1044 } 1045 }
1045 1046
1046 kvm_mmu_flush_tlb(vcpu); 1047 kvm_mmu_flush_tlb(vcpu);
1047 kvm_unlink_unsync_page(vcpu->kvm, sp);
1048 return 0; 1048 return 0;
1049} 1049}
1050 1050
@@ -2634,6 +2634,7 @@ static int kvm_pv_mmu_write(struct kvm_vcpu *vcpu,
2634static int kvm_pv_mmu_flush_tlb(struct kvm_vcpu *vcpu) 2634static int kvm_pv_mmu_flush_tlb(struct kvm_vcpu *vcpu)
2635{ 2635{
2636 kvm_x86_ops->tlb_flush(vcpu); 2636 kvm_x86_ops->tlb_flush(vcpu);
2637 set_bit(KVM_REQ_MMU_SYNC, &vcpu->requests);
2637 return 1; 2638 return 1;
2638} 2639}
2639 2640
diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h
index 613ec9aa674a..84eee43bbe74 100644
--- a/arch/x86/kvm/paging_tmpl.h
+++ b/arch/x86/kvm/paging_tmpl.h
@@ -331,6 +331,7 @@ static int FNAME(shadow_walk_entry)(struct kvm_shadow_walk *_sw,
331 r = kvm_read_guest_atomic(vcpu->kvm, gw->pte_gpa[level - 2], 331 r = kvm_read_guest_atomic(vcpu->kvm, gw->pte_gpa[level - 2],
332 &curr_pte, sizeof(curr_pte)); 332 &curr_pte, sizeof(curr_pte));
333 if (r || curr_pte != gw->ptes[level - 2]) { 333 if (r || curr_pte != gw->ptes[level - 2]) {
334 kvm_mmu_put_page(shadow_page, sptep);
334 kvm_release_pfn_clean(sw->pfn); 335 kvm_release_pfn_clean(sw->pfn);
335 sw->sptep = NULL; 336 sw->sptep = NULL;
336 return 1; 337 return 1;
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 2643b430d83a..a4018b01e1f9 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -3149,7 +3149,9 @@ static void vmx_intr_assist(struct kvm_vcpu *vcpu)
3149 3149
3150 if (cpu_has_virtual_nmis()) { 3150 if (cpu_has_virtual_nmis()) {
3151 if (vcpu->arch.nmi_pending && !vcpu->arch.nmi_injected) { 3151 if (vcpu->arch.nmi_pending && !vcpu->arch.nmi_injected) {
3152 if (vmx_nmi_enabled(vcpu)) { 3152 if (vcpu->arch.interrupt.pending) {
3153 enable_nmi_window(vcpu);
3154 } else if (vmx_nmi_enabled(vcpu)) {
3153 vcpu->arch.nmi_pending = false; 3155 vcpu->arch.nmi_pending = false;
3154 vcpu->arch.nmi_injected = true; 3156 vcpu->arch.nmi_injected = true;
3155 } else { 3157 } else {
@@ -3564,7 +3566,8 @@ static int __init vmx_init(void)
3564 bypass_guest_pf = 0; 3566 bypass_guest_pf = 0;
3565 kvm_mmu_set_base_ptes(VMX_EPT_READABLE_MASK | 3567 kvm_mmu_set_base_ptes(VMX_EPT_READABLE_MASK |
3566 VMX_EPT_WRITABLE_MASK | 3568 VMX_EPT_WRITABLE_MASK |
3567 VMX_EPT_DEFAULT_MT << VMX_EPT_MT_EPTE_SHIFT); 3569 VMX_EPT_DEFAULT_MT << VMX_EPT_MT_EPTE_SHIFT |
3570 VMX_EPT_IGMT_BIT);
3568 kvm_mmu_set_mask_ptes(0ull, 0ull, 0ull, 0ull, 3571 kvm_mmu_set_mask_ptes(0ull, 0ull, 0ull, 0ull,
3569 VMX_EPT_EXECUTABLE_MASK); 3572 VMX_EPT_EXECUTABLE_MASK);
3570 kvm_enable_tdp(); 3573 kvm_enable_tdp();
diff --git a/arch/x86/kvm/vmx.h b/arch/x86/kvm/vmx.h
index 3e010d21fdd7..ec5edc339da6 100644
--- a/arch/x86/kvm/vmx.h
+++ b/arch/x86/kvm/vmx.h
@@ -352,6 +352,7 @@ enum vmcs_field {
352#define VMX_EPT_READABLE_MASK 0x1ull 352#define VMX_EPT_READABLE_MASK 0x1ull
353#define VMX_EPT_WRITABLE_MASK 0x2ull 353#define VMX_EPT_WRITABLE_MASK 0x2ull
354#define VMX_EPT_EXECUTABLE_MASK 0x4ull 354#define VMX_EPT_EXECUTABLE_MASK 0x4ull
355#define VMX_EPT_IGMT_BIT (1ull << 6)
355 356
356#define VMX_EPT_IDENTITY_PAGETABLE_ADDR 0xfffbc000ul 357#define VMX_EPT_IDENTITY_PAGETABLE_ADDR 0xfffbc000ul
357 358
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 4f0677d1eae8..f1f8ff2f1fa2 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -1742,7 +1742,8 @@ long kvm_arch_vm_ioctl(struct file *filp,
1742 goto out; 1742 goto out;
1743 if (irqchip_in_kernel(kvm)) { 1743 if (irqchip_in_kernel(kvm)) {
1744 mutex_lock(&kvm->lock); 1744 mutex_lock(&kvm->lock);
1745 kvm_set_irq(kvm, irq_event.irq, irq_event.level); 1745 kvm_set_irq(kvm, KVM_USERSPACE_IRQ_SOURCE_ID,
1746 irq_event.irq, irq_event.level);
1746 mutex_unlock(&kvm->lock); 1747 mutex_unlock(&kvm->lock);
1747 r = 0; 1748 r = 0;
1748 } 1749 }
@@ -4013,6 +4014,9 @@ struct kvm *kvm_arch_create_vm(void)
4013 INIT_LIST_HEAD(&kvm->arch.active_mmu_pages); 4014 INIT_LIST_HEAD(&kvm->arch.active_mmu_pages);
4014 INIT_LIST_HEAD(&kvm->arch.assigned_dev_head); 4015 INIT_LIST_HEAD(&kvm->arch.assigned_dev_head);
4015 4016
4017 /* Reserve bit 0 of irq_sources_bitmap for userspace irq source */
4018 set_bit(KVM_USERSPACE_IRQ_SOURCE_ID, &kvm->arch.irq_sources_bitmap);
4019
4016 return kvm; 4020 return kvm;
4017} 4021}
4018 4022
diff --git a/arch/x86/lguest/boot.c b/arch/x86/lguest/boot.c
index 48ee4f9435f4..a5d8e1ace1cf 100644
--- a/arch/x86/lguest/boot.c
+++ b/arch/x86/lguest/boot.c
@@ -367,10 +367,9 @@ static void lguest_cpuid(unsigned int *ax, unsigned int *bx,
367 * lazily after a task switch, and Linux uses that gratefully, but wouldn't a 367 * lazily after a task switch, and Linux uses that gratefully, but wouldn't a
368 * name like "FPUTRAP bit" be a little less cryptic? 368 * name like "FPUTRAP bit" be a little less cryptic?
369 * 369 *
370 * We store cr0 (and cr3) locally, because the Host never changes it. The 370 * We store cr0 locally because the Host never changes it. The Guest sometimes
371 * Guest sometimes wants to read it and we'd prefer not to bother the Host 371 * wants to read it and we'd prefer not to bother the Host unnecessarily. */
372 * unnecessarily. */ 372static unsigned long current_cr0;
373static unsigned long current_cr0, current_cr3;
374static void lguest_write_cr0(unsigned long val) 373static void lguest_write_cr0(unsigned long val)
375{ 374{
376 lazy_hcall(LHCALL_TS, val & X86_CR0_TS, 0, 0); 375 lazy_hcall(LHCALL_TS, val & X86_CR0_TS, 0, 0);
@@ -399,17 +398,23 @@ static unsigned long lguest_read_cr2(void)
399 return lguest_data.cr2; 398 return lguest_data.cr2;
400} 399}
401 400
401/* See lguest_set_pte() below. */
402static bool cr3_changed = false;
403
402/* cr3 is the current toplevel pagetable page: the principle is the same as 404/* cr3 is the current toplevel pagetable page: the principle is the same as
403 * cr0. Keep a local copy, and tell the Host when it changes. */ 405 * cr0. Keep a local copy, and tell the Host when it changes. The only
406 * difference is that our local copy is in lguest_data because the Host needs
407 * to set it upon our initial hypercall. */
404static void lguest_write_cr3(unsigned long cr3) 408static void lguest_write_cr3(unsigned long cr3)
405{ 409{
410 lguest_data.pgdir = cr3;
406 lazy_hcall(LHCALL_NEW_PGTABLE, cr3, 0, 0); 411 lazy_hcall(LHCALL_NEW_PGTABLE, cr3, 0, 0);
407 current_cr3 = cr3; 412 cr3_changed = true;
408} 413}
409 414
410static unsigned long lguest_read_cr3(void) 415static unsigned long lguest_read_cr3(void)
411{ 416{
412 return current_cr3; 417 return lguest_data.pgdir;
413} 418}
414 419
415/* cr4 is used to enable and disable PGE, but we don't care. */ 420/* cr4 is used to enable and disable PGE, but we don't care. */
@@ -498,13 +503,13 @@ static void lguest_set_pmd(pmd_t *pmdp, pmd_t pmdval)
498 * to forget all of them. Fortunately, this is very rare. 503 * to forget all of them. Fortunately, this is very rare.
499 * 504 *
500 * ... except in early boot when the kernel sets up the initial pagetables, 505 * ... except in early boot when the kernel sets up the initial pagetables,
501 * which makes booting astonishingly slow. So we don't even tell the Host 506 * which makes booting astonishingly slow: 1.83 seconds! So we don't even tell
502 * anything changed until we've done the first page table switch. */ 507 * the Host anything changed until we've done the first page table switch,
508 * which brings boot back to 0.25 seconds. */
503static void lguest_set_pte(pte_t *ptep, pte_t pteval) 509static void lguest_set_pte(pte_t *ptep, pte_t pteval)
504{ 510{
505 *ptep = pteval; 511 *ptep = pteval;
506 /* Don't bother with hypercall before initial setup. */ 512 if (cr3_changed)
507 if (current_cr3)
508 lazy_hcall(LHCALL_FLUSH_TLB, 1, 0, 0); 513 lazy_hcall(LHCALL_FLUSH_TLB, 1, 0, 0);
509} 514}
510 515
@@ -521,7 +526,7 @@ static void lguest_set_pte(pte_t *ptep, pte_t pteval)
521static void lguest_flush_tlb_single(unsigned long addr) 526static void lguest_flush_tlb_single(unsigned long addr)
522{ 527{
523 /* Simply set it to zero: if it was not, it will fault back in. */ 528 /* Simply set it to zero: if it was not, it will fault back in. */
524 lazy_hcall(LHCALL_SET_PTE, current_cr3, addr, 0); 529 lazy_hcall(LHCALL_SET_PTE, lguest_data.pgdir, addr, 0);
525} 530}
526 531
527/* This is what happens after the Guest has removed a large number of entries. 532/* This is what happens after the Guest has removed a large number of entries.
@@ -581,6 +586,9 @@ static void __init lguest_init_IRQ(void)
581 586
582 for (i = 0; i < LGUEST_IRQS; i++) { 587 for (i = 0; i < LGUEST_IRQS; i++) {
583 int vector = FIRST_EXTERNAL_VECTOR + i; 588 int vector = FIRST_EXTERNAL_VECTOR + i;
589 /* Some systems map "vectors" to interrupts weirdly. Lguest has
590 * a straightforward 1 to 1 mapping, so force that here. */
591 __get_cpu_var(vector_irq)[vector] = i;
584 if (vector != SYSCALL_VECTOR) { 592 if (vector != SYSCALL_VECTOR) {
585 set_intr_gate(vector, interrupt[vector]); 593 set_intr_gate(vector, interrupt[vector]);
586 set_irq_chip_and_handler_name(i, &lguest_irq_controller, 594 set_irq_chip_and_handler_name(i, &lguest_irq_controller,
diff --git a/arch/x86/mach-voyager/setup.c b/arch/x86/mach-voyager/setup.c
index 6bbdd633864c..a580b9562e76 100644
--- a/arch/x86/mach-voyager/setup.c
+++ b/arch/x86/mach-voyager/setup.c
@@ -27,7 +27,7 @@ static struct irqaction irq2 = {
27void __init intr_init_hook(void) 27void __init intr_init_hook(void)
28{ 28{
29#ifdef CONFIG_SMP 29#ifdef CONFIG_SMP
30 smp_intr_init(); 30 voyager_smp_intr_init();
31#endif 31#endif
32 32
33 setup_irq(2, &irq2); 33 setup_irq(2, &irq2);
diff --git a/arch/x86/mach-voyager/voyager_smp.c b/arch/x86/mach-voyager/voyager_smp.c
index 0f6e8a6523ae..52145007bd7e 100644
--- a/arch/x86/mach-voyager/voyager_smp.c
+++ b/arch/x86/mach-voyager/voyager_smp.c
@@ -7,6 +7,7 @@
7 * This file provides all the same external entries as smp.c but uses 7 * This file provides all the same external entries as smp.c but uses
8 * the voyager hal to provide the functionality 8 * the voyager hal to provide the functionality
9 */ 9 */
10#include <linux/cpu.h>
10#include <linux/module.h> 11#include <linux/module.h>
11#include <linux/mm.h> 12#include <linux/mm.h>
12#include <linux/kernel_stat.h> 13#include <linux/kernel_stat.h>
@@ -90,6 +91,7 @@ static void ack_vic_irq(unsigned int irq);
90static void vic_enable_cpi(void); 91static void vic_enable_cpi(void);
91static void do_boot_cpu(__u8 cpuid); 92static void do_boot_cpu(__u8 cpuid);
92static void do_quad_bootstrap(void); 93static void do_quad_bootstrap(void);
94static void initialize_secondary(void);
93 95
94int hard_smp_processor_id(void); 96int hard_smp_processor_id(void);
95int safe_smp_processor_id(void); 97int safe_smp_processor_id(void);
@@ -344,6 +346,12 @@ static void do_quad_bootstrap(void)
344 } 346 }
345} 347}
346 348
349void prefill_possible_map(void)
350{
351 /* This is empty on voyager because we need a much
352 * earlier detection which is done in find_smp_config */
353}
354
347/* Set up all the basic stuff: read the SMP config and make all the 355/* Set up all the basic stuff: read the SMP config and make all the
348 * SMP information reflect only the boot cpu. All others will be 356 * SMP information reflect only the boot cpu. All others will be
349 * brought on-line later. */ 357 * brought on-line later. */
@@ -413,6 +421,7 @@ void __init smp_store_cpu_info(int id)
413 struct cpuinfo_x86 *c = &cpu_data(id); 421 struct cpuinfo_x86 *c = &cpu_data(id);
414 422
415 *c = boot_cpu_data; 423 *c = boot_cpu_data;
424 c->cpu_index = id;
416 425
417 identify_secondary_cpu(c); 426 identify_secondary_cpu(c);
418} 427}
@@ -650,6 +659,8 @@ void __init smp_boot_cpus(void)
650 smp_tune_scheduling(); 659 smp_tune_scheduling();
651 */ 660 */
652 smp_store_cpu_info(boot_cpu_id); 661 smp_store_cpu_info(boot_cpu_id);
662 /* setup the jump vector */
663 initial_code = (unsigned long)initialize_secondary;
653 printk("CPU%d: ", boot_cpu_id); 664 printk("CPU%d: ", boot_cpu_id);
654 print_cpu_info(&cpu_data(boot_cpu_id)); 665 print_cpu_info(&cpu_data(boot_cpu_id));
655 666
@@ -702,7 +713,7 @@ void __init smp_boot_cpus(void)
702 713
703/* Reload the secondary CPUs task structure (this function does not 714/* Reload the secondary CPUs task structure (this function does not
704 * return ) */ 715 * return ) */
705void __init initialize_secondary(void) 716static void __init initialize_secondary(void)
706{ 717{
707#if 0 718#if 0
708 // AC kernels only 719 // AC kernels only
@@ -1248,7 +1259,7 @@ static void handle_vic_irq(unsigned int irq, struct irq_desc *desc)
1248#define QIC_SET_GATE(cpi, vector) \ 1259#define QIC_SET_GATE(cpi, vector) \
1249 set_intr_gate((cpi) + QIC_DEFAULT_CPI_BASE, (vector)) 1260 set_intr_gate((cpi) + QIC_DEFAULT_CPI_BASE, (vector))
1250 1261
1251void __init smp_intr_init(void) 1262void __init voyager_smp_intr_init(void)
1252{ 1263{
1253 int i; 1264 int i;
1254 1265
@@ -1780,6 +1791,17 @@ void __init smp_setup_processor_id(void)
1780 x86_write_percpu(cpu_number, hard_smp_processor_id()); 1791 x86_write_percpu(cpu_number, hard_smp_processor_id());
1781} 1792}
1782 1793
1794static void voyager_send_call_func(cpumask_t callmask)
1795{
1796 __u32 mask = cpus_addr(callmask)[0] & ~(1 << smp_processor_id());
1797 send_CPI(mask, VIC_CALL_FUNCTION_CPI);
1798}
1799
1800static void voyager_send_call_func_single(int cpu)
1801{
1802 send_CPI(1 << cpu, VIC_CALL_FUNCTION_SINGLE_CPI);
1803}
1804
1783struct smp_ops smp_ops = { 1805struct smp_ops smp_ops = {
1784 .smp_prepare_boot_cpu = voyager_smp_prepare_boot_cpu, 1806 .smp_prepare_boot_cpu = voyager_smp_prepare_boot_cpu,
1785 .smp_prepare_cpus = voyager_smp_prepare_cpus, 1807 .smp_prepare_cpus = voyager_smp_prepare_cpus,
@@ -1789,6 +1811,6 @@ struct smp_ops smp_ops = {
1789 .smp_send_stop = voyager_smp_send_stop, 1811 .smp_send_stop = voyager_smp_send_stop,
1790 .smp_send_reschedule = voyager_smp_send_reschedule, 1812 .smp_send_reschedule = voyager_smp_send_reschedule,
1791 1813
1792 .send_call_func_ipi = native_send_call_func_ipi, 1814 .send_call_func_ipi = voyager_send_call_func,
1793 .send_call_func_single_ipi = native_send_call_func_single_ipi, 1815 .send_call_func_single_ipi = voyager_send_call_func_single,
1794}; 1816};
diff --git a/arch/x86/mm/Makefile b/arch/x86/mm/Makefile
index 59f89b434b45..fea4565ff576 100644
--- a/arch/x86/mm/Makefile
+++ b/arch/x86/mm/Makefile
@@ -1,7 +1,7 @@
1obj-y := init_$(BITS).o fault.o ioremap.o extable.o pageattr.o mmap.o \ 1obj-y := init_$(BITS).o fault.o ioremap.o extable.o pageattr.o mmap.o \
2 pat.o pgtable.o gup.o 2 pat.o pgtable.o gup.o
3 3
4obj-$(CONFIG_X86_32) += pgtable_32.o 4obj-$(CONFIG_X86_32) += pgtable_32.o iomap_32.o
5 5
6obj-$(CONFIG_HUGETLB_PAGE) += hugetlbpage.o 6obj-$(CONFIG_HUGETLB_PAGE) += hugetlbpage.o
7obj-$(CONFIG_X86_PTDUMP) += dump_pagetables.o 7obj-$(CONFIG_X86_PTDUMP) += dump_pagetables.o
diff --git a/arch/x86/mm/gup.c b/arch/x86/mm/gup.c
index 4ba373c5b8c8..be54176e9eb2 100644
--- a/arch/x86/mm/gup.c
+++ b/arch/x86/mm/gup.c
@@ -233,7 +233,7 @@ int get_user_pages_fast(unsigned long start, int nr_pages, int write,
233 len = (unsigned long) nr_pages << PAGE_SHIFT; 233 len = (unsigned long) nr_pages << PAGE_SHIFT;
234 end = start + len; 234 end = start + len;
235 if (unlikely(!access_ok(write ? VERIFY_WRITE : VERIFY_READ, 235 if (unlikely(!access_ok(write ? VERIFY_WRITE : VERIFY_READ,
236 start, len))) 236 (void __user *)start, len)))
237 goto slow_irqon; 237 goto slow_irqon;
238 238
239 /* 239 /*
diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c
index 8396868e82c5..c483f4242079 100644
--- a/arch/x86/mm/init_32.c
+++ b/arch/x86/mm/init_32.c
@@ -334,7 +334,6 @@ int devmem_is_allowed(unsigned long pagenr)
334 return 0; 334 return 0;
335} 335}
336 336
337#ifdef CONFIG_HIGHMEM
338pte_t *kmap_pte; 337pte_t *kmap_pte;
339pgprot_t kmap_prot; 338pgprot_t kmap_prot;
340 339
@@ -357,6 +356,7 @@ static void __init kmap_init(void)
357 kmap_prot = PAGE_KERNEL; 356 kmap_prot = PAGE_KERNEL;
358} 357}
359 358
359#ifdef CONFIG_HIGHMEM
360static void __init permanent_kmaps_init(pgd_t *pgd_base) 360static void __init permanent_kmaps_init(pgd_t *pgd_base)
361{ 361{
362 unsigned long vaddr; 362 unsigned long vaddr;
@@ -436,7 +436,6 @@ static void __init set_highmem_pages_init(void)
436#endif /* !CONFIG_NUMA */ 436#endif /* !CONFIG_NUMA */
437 437
438#else 438#else
439# define kmap_init() do { } while (0)
440# define permanent_kmaps_init(pgd_base) do { } while (0) 439# define permanent_kmaps_init(pgd_base) do { } while (0)
441# define set_highmem_pages_init() do { } while (0) 440# define set_highmem_pages_init() do { } while (0)
442#endif /* CONFIG_HIGHMEM */ 441#endif /* CONFIG_HIGHMEM */
diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c
index b8e461d49412..9db01db6e3cd 100644
--- a/arch/x86/mm/init_64.c
+++ b/arch/x86/mm/init_64.c
@@ -350,8 +350,10 @@ phys_pte_init(pte_t *pte_page, unsigned long addr, unsigned long end,
350 * pagetable pages as RO. So assume someone who pre-setup 350 * pagetable pages as RO. So assume someone who pre-setup
351 * these mappings are more intelligent. 351 * these mappings are more intelligent.
352 */ 352 */
353 if (pte_val(*pte)) 353 if (pte_val(*pte)) {
354 pages++;
354 continue; 355 continue;
356 }
355 357
356 if (0) 358 if (0)
357 printk(" pte=%p addr=%lx pte=%016lx\n", 359 printk(" pte=%p addr=%lx pte=%016lx\n",
@@ -418,8 +420,10 @@ phys_pmd_init(pmd_t *pmd_page, unsigned long address, unsigned long end,
418 * not differ with respect to page frame and 420 * not differ with respect to page frame and
419 * attributes. 421 * attributes.
420 */ 422 */
421 if (page_size_mask & (1 << PG_LEVEL_2M)) 423 if (page_size_mask & (1 << PG_LEVEL_2M)) {
424 pages++;
422 continue; 425 continue;
426 }
423 new_prot = pte_pgprot(pte_clrhuge(*(pte_t *)pmd)); 427 new_prot = pte_pgprot(pte_clrhuge(*(pte_t *)pmd));
424 } 428 }
425 429
@@ -499,8 +503,10 @@ phys_pud_init(pud_t *pud_page, unsigned long addr, unsigned long end,
499 * not differ with respect to page frame and 503 * not differ with respect to page frame and
500 * attributes. 504 * attributes.
501 */ 505 */
502 if (page_size_mask & (1 << PG_LEVEL_1G)) 506 if (page_size_mask & (1 << PG_LEVEL_1G)) {
507 pages++;
503 continue; 508 continue;
509 }
504 prot = pte_pgprot(pte_clrhuge(*(pte_t *)pud)); 510 prot = pte_pgprot(pte_clrhuge(*(pte_t *)pud));
505 } 511 }
506 512
@@ -665,12 +671,13 @@ unsigned long __init_refok init_memory_mapping(unsigned long start,
665 unsigned long last_map_addr = 0; 671 unsigned long last_map_addr = 0;
666 unsigned long page_size_mask = 0; 672 unsigned long page_size_mask = 0;
667 unsigned long start_pfn, end_pfn; 673 unsigned long start_pfn, end_pfn;
674 unsigned long pos;
668 675
669 struct map_range mr[NR_RANGE_MR]; 676 struct map_range mr[NR_RANGE_MR];
670 int nr_range, i; 677 int nr_range, i;
671 int use_pse, use_gbpages; 678 int use_pse, use_gbpages;
672 679
673 printk(KERN_INFO "init_memory_mapping\n"); 680 printk(KERN_INFO "init_memory_mapping: %016lx-%016lx\n", start, end);
674 681
675 /* 682 /*
676 * Find space for the kernel direct mapping tables. 683 * Find space for the kernel direct mapping tables.
@@ -704,35 +711,50 @@ unsigned long __init_refok init_memory_mapping(unsigned long start,
704 711
705 /* head if not big page alignment ?*/ 712 /* head if not big page alignment ?*/
706 start_pfn = start >> PAGE_SHIFT; 713 start_pfn = start >> PAGE_SHIFT;
707 end_pfn = ((start + (PMD_SIZE - 1)) >> PMD_SHIFT) 714 pos = start_pfn << PAGE_SHIFT;
715 end_pfn = ((pos + (PMD_SIZE - 1)) >> PMD_SHIFT)
708 << (PMD_SHIFT - PAGE_SHIFT); 716 << (PMD_SHIFT - PAGE_SHIFT);
709 nr_range = save_mr(mr, nr_range, start_pfn, end_pfn, 0); 717 if (start_pfn < end_pfn) {
718 nr_range = save_mr(mr, nr_range, start_pfn, end_pfn, 0);
719 pos = end_pfn << PAGE_SHIFT;
720 }
710 721
711 /* big page (2M) range*/ 722 /* big page (2M) range*/
712 start_pfn = ((start + (PMD_SIZE - 1))>>PMD_SHIFT) 723 start_pfn = ((pos + (PMD_SIZE - 1))>>PMD_SHIFT)
713 << (PMD_SHIFT - PAGE_SHIFT); 724 << (PMD_SHIFT - PAGE_SHIFT);
714 end_pfn = ((start + (PUD_SIZE - 1))>>PUD_SHIFT) 725 end_pfn = ((pos + (PUD_SIZE - 1))>>PUD_SHIFT)
715 << (PUD_SHIFT - PAGE_SHIFT); 726 << (PUD_SHIFT - PAGE_SHIFT);
716 if (end_pfn > ((end>>PUD_SHIFT)<<(PUD_SHIFT - PAGE_SHIFT))) 727 if (end_pfn > ((end>>PMD_SHIFT)<<(PMD_SHIFT - PAGE_SHIFT)))
717 end_pfn = ((end>>PUD_SHIFT)<<(PUD_SHIFT - PAGE_SHIFT)); 728 end_pfn = ((end>>PMD_SHIFT)<<(PMD_SHIFT - PAGE_SHIFT));
718 nr_range = save_mr(mr, nr_range, start_pfn, end_pfn, 729 if (start_pfn < end_pfn) {
719 page_size_mask & (1<<PG_LEVEL_2M)); 730 nr_range = save_mr(mr, nr_range, start_pfn, end_pfn,
731 page_size_mask & (1<<PG_LEVEL_2M));
732 pos = end_pfn << PAGE_SHIFT;
733 }
720 734
721 /* big page (1G) range */ 735 /* big page (1G) range */
722 start_pfn = end_pfn; 736 start_pfn = ((pos + (PUD_SIZE - 1))>>PUD_SHIFT)
723 end_pfn = (end>>PUD_SHIFT) << (PUD_SHIFT - PAGE_SHIFT); 737 << (PUD_SHIFT - PAGE_SHIFT);
724 nr_range = save_mr(mr, nr_range, start_pfn, end_pfn, 738 end_pfn = (end >> PUD_SHIFT) << (PUD_SHIFT - PAGE_SHIFT);
739 if (start_pfn < end_pfn) {
740 nr_range = save_mr(mr, nr_range, start_pfn, end_pfn,
725 page_size_mask & 741 page_size_mask &
726 ((1<<PG_LEVEL_2M)|(1<<PG_LEVEL_1G))); 742 ((1<<PG_LEVEL_2M)|(1<<PG_LEVEL_1G)));
743 pos = end_pfn << PAGE_SHIFT;
744 }
727 745
728 /* tail is not big page (1G) alignment */ 746 /* tail is not big page (1G) alignment */
729 start_pfn = end_pfn; 747 start_pfn = ((pos + (PMD_SIZE - 1))>>PMD_SHIFT)
730 end_pfn = (end>>PMD_SHIFT) << (PMD_SHIFT - PAGE_SHIFT); 748 << (PMD_SHIFT - PAGE_SHIFT);
731 nr_range = save_mr(mr, nr_range, start_pfn, end_pfn, 749 end_pfn = (end >> PMD_SHIFT) << (PMD_SHIFT - PAGE_SHIFT);
732 page_size_mask & (1<<PG_LEVEL_2M)); 750 if (start_pfn < end_pfn) {
751 nr_range = save_mr(mr, nr_range, start_pfn, end_pfn,
752 page_size_mask & (1<<PG_LEVEL_2M));
753 pos = end_pfn << PAGE_SHIFT;
754 }
733 755
734 /* tail is not big page (2M) alignment */ 756 /* tail is not big page (2M) alignment */
735 start_pfn = end_pfn; 757 start_pfn = pos>>PAGE_SHIFT;
736 end_pfn = end>>PAGE_SHIFT; 758 end_pfn = end>>PAGE_SHIFT;
737 nr_range = save_mr(mr, nr_range, start_pfn, end_pfn, 0); 759 nr_range = save_mr(mr, nr_range, start_pfn, end_pfn, 0);
738 760
@@ -831,12 +853,12 @@ int arch_add_memory(int nid, u64 start, u64 size)
831 unsigned long nr_pages = size >> PAGE_SHIFT; 853 unsigned long nr_pages = size >> PAGE_SHIFT;
832 int ret; 854 int ret;
833 855
834 last_mapped_pfn = init_memory_mapping(start, start + size-1); 856 last_mapped_pfn = init_memory_mapping(start, start + size);
835 if (last_mapped_pfn > max_pfn_mapped) 857 if (last_mapped_pfn > max_pfn_mapped)
836 max_pfn_mapped = last_mapped_pfn; 858 max_pfn_mapped = last_mapped_pfn;
837 859
838 ret = __add_pages(zone, start_pfn, nr_pages); 860 ret = __add_pages(zone, start_pfn, nr_pages);
839 WARN_ON(1); 861 WARN_ON_ONCE(ret);
840 862
841 return ret; 863 return ret;
842} 864}
@@ -878,6 +900,7 @@ static struct kcore_list kcore_mem, kcore_vmalloc, kcore_kernel,
878void __init mem_init(void) 900void __init mem_init(void)
879{ 901{
880 long codesize, reservedpages, datasize, initsize; 902 long codesize, reservedpages, datasize, initsize;
903 unsigned long absent_pages;
881 904
882 start_periodic_check_for_corruption(); 905 start_periodic_check_for_corruption();
883 906
@@ -893,8 +916,9 @@ void __init mem_init(void)
893#else 916#else
894 totalram_pages = free_all_bootmem(); 917 totalram_pages = free_all_bootmem();
895#endif 918#endif
896 reservedpages = max_pfn - totalram_pages - 919
897 absent_pages_in_range(0, max_pfn); 920 absent_pages = absent_pages_in_range(0, max_pfn);
921 reservedpages = max_pfn - totalram_pages - absent_pages;
898 after_bootmem = 1; 922 after_bootmem = 1;
899 923
900 codesize = (unsigned long) &_etext - (unsigned long) &_text; 924 codesize = (unsigned long) &_etext - (unsigned long) &_text;
@@ -911,10 +935,11 @@ void __init mem_init(void)
911 VSYSCALL_END - VSYSCALL_START); 935 VSYSCALL_END - VSYSCALL_START);
912 936
913 printk(KERN_INFO "Memory: %luk/%luk available (%ldk kernel code, " 937 printk(KERN_INFO "Memory: %luk/%luk available (%ldk kernel code, "
914 "%ldk reserved, %ldk data, %ldk init)\n", 938 "%ldk absent, %ldk reserved, %ldk data, %ldk init)\n",
915 (unsigned long) nr_free_pages() << (PAGE_SHIFT-10), 939 (unsigned long) nr_free_pages() << (PAGE_SHIFT-10),
916 max_pfn << (PAGE_SHIFT-10), 940 max_pfn << (PAGE_SHIFT-10),
917 codesize >> 10, 941 codesize >> 10,
942 absent_pages << (PAGE_SHIFT-10),
918 reservedpages << (PAGE_SHIFT-10), 943 reservedpages << (PAGE_SHIFT-10),
919 datasize >> 10, 944 datasize >> 10,
920 initsize >> 10); 945 initsize >> 10);
diff --git a/arch/x86/mm/iomap_32.c b/arch/x86/mm/iomap_32.c
new file mode 100644
index 000000000000..d0151d8ce452
--- /dev/null
+++ b/arch/x86/mm/iomap_32.c
@@ -0,0 +1,59 @@
1/*
2 * Copyright © 2008 Ingo Molnar
3 *
4 * This program is free software; you can redistribute it and/or modify
5 * it under the terms of the GNU General Public License as published by
6 * the Free Software Foundation; either version 2 of the License, or
7 * (at your option) any later version.
8 *
9 * This program is distributed in the hope that it will be useful, but
10 * WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 * General Public License for more details.
13 *
14 * You should have received a copy of the GNU General Public License along
15 * with this program; if not, write to the Free Software Foundation, Inc.,
16 * 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
17 */
18
19#include <asm/iomap.h>
20#include <linux/module.h>
21
22/* Map 'pfn' using fixed map 'type' and protections 'prot'
23 */
24void *
25iomap_atomic_prot_pfn(unsigned long pfn, enum km_type type, pgprot_t prot)
26{
27 enum fixed_addresses idx;
28 unsigned long vaddr;
29
30 pagefault_disable();
31
32 idx = type + KM_TYPE_NR*smp_processor_id();
33 vaddr = __fix_to_virt(FIX_KMAP_BEGIN + idx);
34 set_pte(kmap_pte-idx, pfn_pte(pfn, prot));
35 arch_flush_lazy_mmu_mode();
36
37 return (void*) vaddr;
38}
39EXPORT_SYMBOL_GPL(iomap_atomic_prot_pfn);
40
41void
42iounmap_atomic(void *kvaddr, enum km_type type)
43{
44 unsigned long vaddr = (unsigned long) kvaddr & PAGE_MASK;
45 enum fixed_addresses idx = type + KM_TYPE_NR*smp_processor_id();
46
47 /*
48 * Force other mappings to Oops if they'll try to access this pte
49 * without first remap it. Keeping stale mappings around is a bad idea
50 * also, in case the page changes cacheability attributes or becomes
51 * a protected page in a hypervisor.
52 */
53 if (vaddr == __fix_to_virt(FIX_KMAP_BEGIN+idx))
54 kpte_clear_flush(kmap_pte-idx, vaddr);
55
56 arch_flush_lazy_mmu_mode();
57 pagefault_enable();
58}
59EXPORT_SYMBOL_GPL(iounmap_atomic);
diff --git a/arch/x86/mm/ioremap.c b/arch/x86/mm/ioremap.c
index ae71e11eb3e5..d4c4307ff3e0 100644
--- a/arch/x86/mm/ioremap.c
+++ b/arch/x86/mm/ioremap.c
@@ -387,7 +387,7 @@ static void __iomem *ioremap_default(resource_size_t phys_addr,
387 unsigned long size) 387 unsigned long size)
388{ 388{
389 unsigned long flags; 389 unsigned long flags;
390 void *ret; 390 void __iomem *ret;
391 int err; 391 int err;
392 392
393 /* 393 /*
@@ -399,11 +399,11 @@ static void __iomem *ioremap_default(resource_size_t phys_addr,
399 if (err < 0) 399 if (err < 0)
400 return NULL; 400 return NULL;
401 401
402 ret = (void *) __ioremap_caller(phys_addr, size, flags, 402 ret = __ioremap_caller(phys_addr, size, flags,
403 __builtin_return_address(0)); 403 __builtin_return_address(0));
404 404
405 free_memtype(phys_addr, phys_addr + size); 405 free_memtype(phys_addr, phys_addr + size);
406 return (void __iomem *)ret; 406 return ret;
407} 407}
408 408
409void __iomem *ioremap_prot(resource_size_t phys_addr, unsigned long size, 409void __iomem *ioremap_prot(resource_size_t phys_addr, unsigned long size,
@@ -622,7 +622,7 @@ static inline void __init early_clear_fixmap(enum fixed_addresses idx)
622 __early_set_fixmap(idx, 0, __pgprot(0)); 622 __early_set_fixmap(idx, 0, __pgprot(0));
623} 623}
624 624
625static void *prev_map[FIX_BTMAPS_SLOTS] __initdata; 625static void __iomem *prev_map[FIX_BTMAPS_SLOTS] __initdata;
626static unsigned long prev_size[FIX_BTMAPS_SLOTS] __initdata; 626static unsigned long prev_size[FIX_BTMAPS_SLOTS] __initdata;
627static int __init check_early_ioremap_leak(void) 627static int __init check_early_ioremap_leak(void)
628{ 628{
@@ -645,7 +645,7 @@ static int __init check_early_ioremap_leak(void)
645} 645}
646late_initcall(check_early_ioremap_leak); 646late_initcall(check_early_ioremap_leak);
647 647
648static void __init *__early_ioremap(unsigned long phys_addr, unsigned long size, pgprot_t prot) 648static void __init __iomem *__early_ioremap(unsigned long phys_addr, unsigned long size, pgprot_t prot)
649{ 649{
650 unsigned long offset, last_addr; 650 unsigned long offset, last_addr;
651 unsigned int nrpages; 651 unsigned int nrpages;
@@ -713,23 +713,23 @@ static void __init *__early_ioremap(unsigned long phys_addr, unsigned long size,
713 if (early_ioremap_debug) 713 if (early_ioremap_debug)
714 printk(KERN_CONT "%08lx + %08lx\n", offset, fix_to_virt(idx0)); 714 printk(KERN_CONT "%08lx + %08lx\n", offset, fix_to_virt(idx0));
715 715
716 prev_map[slot] = (void *) (offset + fix_to_virt(idx0)); 716 prev_map[slot] = (void __iomem *)(offset + fix_to_virt(idx0));
717 return prev_map[slot]; 717 return prev_map[slot];
718} 718}
719 719
720/* Remap an IO device */ 720/* Remap an IO device */
721void __init *early_ioremap(unsigned long phys_addr, unsigned long size) 721void __init __iomem *early_ioremap(unsigned long phys_addr, unsigned long size)
722{ 722{
723 return __early_ioremap(phys_addr, size, PAGE_KERNEL_IO); 723 return __early_ioremap(phys_addr, size, PAGE_KERNEL_IO);
724} 724}
725 725
726/* Remap memory */ 726/* Remap memory */
727void __init *early_memremap(unsigned long phys_addr, unsigned long size) 727void __init __iomem *early_memremap(unsigned long phys_addr, unsigned long size)
728{ 728{
729 return __early_ioremap(phys_addr, size, PAGE_KERNEL); 729 return __early_ioremap(phys_addr, size, PAGE_KERNEL);
730} 730}
731 731
732void __init early_iounmap(void *addr, unsigned long size) 732void __init early_iounmap(void __iomem *addr, unsigned long size)
733{ 733{
734 unsigned long virt_addr; 734 unsigned long virt_addr;
735 unsigned long offset; 735 unsigned long offset;
@@ -779,7 +779,7 @@ void __init early_iounmap(void *addr, unsigned long size)
779 --idx; 779 --idx;
780 --nrpages; 780 --nrpages;
781 } 781 }
782 prev_map[slot] = 0; 782 prev_map[slot] = NULL;
783} 783}
784 784
785void __this_fixmap_does_not_exist(void) 785void __this_fixmap_does_not_exist(void)
diff --git a/arch/x86/mm/numa_32.c b/arch/x86/mm/numa_32.c
index 847c164725f4..8518c678d83f 100644
--- a/arch/x86/mm/numa_32.c
+++ b/arch/x86/mm/numa_32.c
@@ -222,6 +222,41 @@ static void __init remap_numa_kva(void)
222 } 222 }
223} 223}
224 224
225#ifdef CONFIG_HIBERNATION
226/**
227 * resume_map_numa_kva - add KVA mapping to the temporary page tables created
228 * during resume from hibernation
229 * @pgd_base - temporary resume page directory
230 */
231void resume_map_numa_kva(pgd_t *pgd_base)
232{
233 int node;
234
235 for_each_online_node(node) {
236 unsigned long start_va, start_pfn, size, pfn;
237
238 start_va = (unsigned long)node_remap_start_vaddr[node];
239 start_pfn = node_remap_start_pfn[node];
240 size = node_remap_size[node];
241
242 printk(KERN_DEBUG "%s: node %d\n", __FUNCTION__, node);
243
244 for (pfn = 0; pfn < size; pfn += PTRS_PER_PTE) {
245 unsigned long vaddr = start_va + (pfn << PAGE_SHIFT);
246 pgd_t *pgd = pgd_base + pgd_index(vaddr);
247 pud_t *pud = pud_offset(pgd, vaddr);
248 pmd_t *pmd = pmd_offset(pud, vaddr);
249
250 set_pmd(pmd, pfn_pmd(start_pfn + pfn,
251 PAGE_KERNEL_LARGE_EXEC));
252
253 printk(KERN_DEBUG "%s: %08lx -> pfn %08lx\n",
254 __FUNCTION__, vaddr, start_pfn + pfn);
255 }
256 }
257}
258#endif
259
225static unsigned long calculate_numa_remap_pages(void) 260static unsigned long calculate_numa_remap_pages(void)
226{ 261{
227 int nid; 262 int nid;
diff --git a/arch/x86/mm/pageattr.c b/arch/x86/mm/pageattr.c
index f1dc1b75d166..e89d24815f26 100644
--- a/arch/x86/mm/pageattr.c
+++ b/arch/x86/mm/pageattr.c
@@ -67,18 +67,18 @@ static void split_page_count(int level)
67 67
68void arch_report_meminfo(struct seq_file *m) 68void arch_report_meminfo(struct seq_file *m)
69{ 69{
70 seq_printf(m, "DirectMap4k: %8lu kB\n", 70 seq_printf(m, "DirectMap4k: %8lu kB\n",
71 direct_pages_count[PG_LEVEL_4K] << 2); 71 direct_pages_count[PG_LEVEL_4K] << 2);
72#if defined(CONFIG_X86_64) || defined(CONFIG_X86_PAE) 72#if defined(CONFIG_X86_64) || defined(CONFIG_X86_PAE)
73 seq_printf(m, "DirectMap2M: %8lu kB\n", 73 seq_printf(m, "DirectMap2M: %8lu kB\n",
74 direct_pages_count[PG_LEVEL_2M] << 11); 74 direct_pages_count[PG_LEVEL_2M] << 11);
75#else 75#else
76 seq_printf(m, "DirectMap4M: %8lu kB\n", 76 seq_printf(m, "DirectMap4M: %8lu kB\n",
77 direct_pages_count[PG_LEVEL_2M] << 12); 77 direct_pages_count[PG_LEVEL_2M] << 12);
78#endif 78#endif
79#ifdef CONFIG_X86_64 79#ifdef CONFIG_X86_64
80 if (direct_gbpages) 80 if (direct_gbpages)
81 seq_printf(m, "DirectMap1G: %8lu kB\n", 81 seq_printf(m, "DirectMap1G: %8lu kB\n",
82 direct_pages_count[PG_LEVEL_1G] << 20); 82 direct_pages_count[PG_LEVEL_1G] << 20);
83#endif 83#endif
84} 84}
diff --git a/arch/x86/mm/pat.c b/arch/x86/mm/pat.c
index 738fd0f24958..eb1bf000d12e 100644
--- a/arch/x86/mm/pat.c
+++ b/arch/x86/mm/pat.c
@@ -481,12 +481,16 @@ static inline int range_is_allowed(unsigned long pfn, unsigned long size)
481 return 1; 481 return 1;
482} 482}
483#else 483#else
484/* This check is needed to avoid cache aliasing when PAT is enabled */
484static inline int range_is_allowed(unsigned long pfn, unsigned long size) 485static inline int range_is_allowed(unsigned long pfn, unsigned long size)
485{ 486{
486 u64 from = ((u64)pfn) << PAGE_SHIFT; 487 u64 from = ((u64)pfn) << PAGE_SHIFT;
487 u64 to = from + size; 488 u64 to = from + size;
488 u64 cursor = from; 489 u64 cursor = from;
489 490
491 if (!pat_enabled)
492 return 1;
493
490 while (cursor < to) { 494 while (cursor < to) {
491 if (!devmem_is_allowed(pfn)) { 495 if (!devmem_is_allowed(pfn)) {
492 printk(KERN_INFO 496 printk(KERN_INFO
diff --git a/arch/x86/oprofile/nmi_int.c b/arch/x86/oprofile/nmi_int.c
index 022cd41ea9b4..202864ad49a7 100644
--- a/arch/x86/oprofile/nmi_int.c
+++ b/arch/x86/oprofile/nmi_int.c
@@ -401,14 +401,13 @@ static int __init ppro_init(char **cpu_type)
401 *cpu_type = "i386/pii"; 401 *cpu_type = "i386/pii";
402 break; 402 break;
403 case 6 ... 8: 403 case 6 ... 8:
404 case 10 ... 11:
404 *cpu_type = "i386/piii"; 405 *cpu_type = "i386/piii";
405 break; 406 break;
406 case 9: 407 case 9:
408 case 13:
407 *cpu_type = "i386/p6_mobile"; 409 *cpu_type = "i386/p6_mobile";
408 break; 410 break;
409 case 10 ... 13:
410 *cpu_type = "i386/p6";
411 break;
412 case 14: 411 case 14:
413 *cpu_type = "i386/core"; 412 *cpu_type = "i386/core";
414 break; 413 break;
diff --git a/arch/x86/oprofile/op_model_ppro.c b/arch/x86/oprofile/op_model_ppro.c
index 0620d6d45f7d..e9f80c744cf3 100644
--- a/arch/x86/oprofile/op_model_ppro.c
+++ b/arch/x86/oprofile/op_model_ppro.c
@@ -27,8 +27,7 @@ static int num_counters = 2;
27static int counter_width = 32; 27static int counter_width = 32;
28 28
29#define CTR_IS_RESERVED(msrs, c) (msrs->counters[(c)].addr ? 1 : 0) 29#define CTR_IS_RESERVED(msrs, c) (msrs->counters[(c)].addr ? 1 : 0)
30#define CTR_READ(l, h, msrs, c) do {rdmsr(msrs->counters[(c)].addr, (l), (h)); } while (0) 30#define CTR_OVERFLOWED(n) (!((n) & (1ULL<<(counter_width-1))))
31#define CTR_OVERFLOWED(n) (!((n) & (1U<<(counter_width-1))))
32 31
33#define CTRL_IS_RESERVED(msrs, c) (msrs->controls[(c)].addr ? 1 : 0) 32#define CTRL_IS_RESERVED(msrs, c) (msrs->controls[(c)].addr ? 1 : 0)
34#define CTRL_READ(l, h, msrs, c) do {rdmsr((msrs->controls[(c)].addr), (l), (h)); } while (0) 33#define CTRL_READ(l, h, msrs, c) do {rdmsr((msrs->controls[(c)].addr), (l), (h)); } while (0)
@@ -70,7 +69,7 @@ static void ppro_setup_ctrs(struct op_msrs const * const msrs)
70 int i; 69 int i;
71 70
72 if (!reset_value) { 71 if (!reset_value) {
73 reset_value = kmalloc(sizeof(unsigned) * num_counters, 72 reset_value = kmalloc(sizeof(reset_value[0]) * num_counters,
74 GFP_ATOMIC); 73 GFP_ATOMIC);
75 if (!reset_value) 74 if (!reset_value)
76 return; 75 return;
@@ -124,14 +123,14 @@ static void ppro_setup_ctrs(struct op_msrs const * const msrs)
124static int ppro_check_ctrs(struct pt_regs * const regs, 123static int ppro_check_ctrs(struct pt_regs * const regs,
125 struct op_msrs const * const msrs) 124 struct op_msrs const * const msrs)
126{ 125{
127 unsigned int low, high; 126 u64 val;
128 int i; 127 int i;
129 128
130 for (i = 0 ; i < num_counters; ++i) { 129 for (i = 0 ; i < num_counters; ++i) {
131 if (!reset_value[i]) 130 if (!reset_value[i])
132 continue; 131 continue;
133 CTR_READ(low, high, msrs, i); 132 rdmsrl(msrs->counters[i].addr, val);
134 if (CTR_OVERFLOWED(low)) { 133 if (CTR_OVERFLOWED(val)) {
135 oprofile_add_sample(regs, i); 134 oprofile_add_sample(regs, i);
136 wrmsrl(msrs->counters[i].addr, -reset_value[i]); 135 wrmsrl(msrs->counters[i].addr, -reset_value[i]);
137 } 136 }
@@ -157,6 +156,8 @@ static void ppro_start(struct op_msrs const * const msrs)
157 unsigned int low, high; 156 unsigned int low, high;
158 int i; 157 int i;
159 158
159 if (!reset_value)
160 return;
160 for (i = 0; i < num_counters; ++i) { 161 for (i = 0; i < num_counters; ++i) {
161 if (reset_value[i]) { 162 if (reset_value[i]) {
162 CTRL_READ(low, high, msrs, i); 163 CTRL_READ(low, high, msrs, i);
@@ -172,6 +173,8 @@ static void ppro_stop(struct op_msrs const * const msrs)
172 unsigned int low, high; 173 unsigned int low, high;
173 int i; 174 int i;
174 175
176 if (!reset_value)
177 return;
175 for (i = 0; i < num_counters; ++i) { 178 for (i = 0; i < num_counters; ++i) {
176 if (!reset_value[i]) 179 if (!reset_value[i])
177 continue; 180 continue;
diff --git a/arch/x86/pci/fixup.c b/arch/x86/pci/fixup.c
index 3c27a809393b..2051dc96b8e9 100644
--- a/arch/x86/pci/fixup.c
+++ b/arch/x86/pci/fixup.c
@@ -496,21 +496,24 @@ DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_SIEMENS, 0x0015,
496 pci_siemens_interrupt_controller); 496 pci_siemens_interrupt_controller);
497 497
498/* 498/*
499 * Regular PCI devices have 256 bytes, but AMD Family 10h Opteron ext config 499 * Regular PCI devices have 256 bytes, but AMD Family 10h/11h CPUs have
500 * have 4096 bytes. Even if the device is capable, that doesn't mean we can 500 * 4096 bytes configuration space for each function of their processor
501 * access it. Maybe we don't have a way to generate extended config space 501 * configuration space.
502 * accesses. So check it
503 */ 502 */
504static void fam10h_pci_cfg_space_size(struct pci_dev *dev) 503static void amd_cpu_pci_cfg_space_size(struct pci_dev *dev)
505{ 504{
506 dev->cfg_size = pci_cfg_space_size_ext(dev); 505 dev->cfg_size = pci_cfg_space_size_ext(dev);
507} 506}
508 507DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_AMD, 0x1200, amd_cpu_pci_cfg_space_size);
509DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_AMD, 0x1200, fam10h_pci_cfg_space_size); 508DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_AMD, 0x1201, amd_cpu_pci_cfg_space_size);
510DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_AMD, 0x1201, fam10h_pci_cfg_space_size); 509DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_AMD, 0x1202, amd_cpu_pci_cfg_space_size);
511DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_AMD, 0x1202, fam10h_pci_cfg_space_size); 510DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_AMD, 0x1203, amd_cpu_pci_cfg_space_size);
512DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_AMD, 0x1203, fam10h_pci_cfg_space_size); 511DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_AMD, 0x1204, amd_cpu_pci_cfg_space_size);
513DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_AMD, 0x1204, fam10h_pci_cfg_space_size); 512DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_AMD, 0x1300, amd_cpu_pci_cfg_space_size);
513DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_AMD, 0x1301, amd_cpu_pci_cfg_space_size);
514DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_AMD, 0x1302, amd_cpu_pci_cfg_space_size);
515DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_AMD, 0x1303, amd_cpu_pci_cfg_space_size);
516DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_AMD, 0x1304, amd_cpu_pci_cfg_space_size);
514 517
515/* 518/*
516 * SB600: Disable BAR1 on device 14.0 to avoid HPET resources from 519 * SB600: Disable BAR1 on device 14.0 to avoid HPET resources from
diff --git a/arch/x86/power/hibernate_32.c b/arch/x86/power/hibernate_32.c
index f2b6e3f11bfc..81197c62d5b3 100644
--- a/arch/x86/power/hibernate_32.c
+++ b/arch/x86/power/hibernate_32.c
@@ -12,6 +12,7 @@
12#include <asm/system.h> 12#include <asm/system.h>
13#include <asm/page.h> 13#include <asm/page.h>
14#include <asm/pgtable.h> 14#include <asm/pgtable.h>
15#include <asm/mmzone.h>
15 16
16/* Defined in hibernate_asm_32.S */ 17/* Defined in hibernate_asm_32.S */
17extern int restore_image(void); 18extern int restore_image(void);
@@ -127,6 +128,9 @@ static int resume_physical_mapping_init(pgd_t *pgd_base)
127 } 128 }
128 } 129 }
129 } 130 }
131
132 resume_map_numa_kva(pgd_base);
133
130 return 0; 134 return 0;
131} 135}
132 136
diff --git a/arch/x86/xen/Makefile b/arch/x86/xen/Makefile
index 313947940a1a..6dcefba7836f 100644
--- a/arch/x86/xen/Makefile
+++ b/arch/x86/xen/Makefile
@@ -1,4 +1,4 @@
1ifdef CONFIG_FTRACE 1ifdef CONFIG_FUNCTION_TRACER
2# Do not profile debug and lowlevel utilities 2# Do not profile debug and lowlevel utilities
3CFLAGS_REMOVE_spinlock.o = -pg 3CFLAGS_REMOVE_spinlock.o = -pg
4CFLAGS_REMOVE_time.o = -pg 4CFLAGS_REMOVE_time.o = -pg
diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c
index b61534c7a4c4..5e4686d70f62 100644
--- a/arch/x86/xen/enlighten.c
+++ b/arch/x86/xen/enlighten.c
@@ -863,15 +863,16 @@ static void xen_alloc_ptpage(struct mm_struct *mm, unsigned long pfn, unsigned l
863 if (PagePinned(virt_to_page(mm->pgd))) { 863 if (PagePinned(virt_to_page(mm->pgd))) {
864 SetPagePinned(page); 864 SetPagePinned(page);
865 865
866 vm_unmap_aliases();
866 if (!PageHighMem(page)) { 867 if (!PageHighMem(page)) {
867 make_lowmem_page_readonly(__va(PFN_PHYS((unsigned long)pfn))); 868 make_lowmem_page_readonly(__va(PFN_PHYS((unsigned long)pfn)));
868 if (level == PT_PTE && USE_SPLIT_PTLOCKS) 869 if (level == PT_PTE && USE_SPLIT_PTLOCKS)
869 pin_pagetable_pfn(MMUEXT_PIN_L1_TABLE, pfn); 870 pin_pagetable_pfn(MMUEXT_PIN_L1_TABLE, pfn);
870 } else 871 } else {
871 /* make sure there are no stray mappings of 872 /* make sure there are no stray mappings of
872 this page */ 873 this page */
873 kmap_flush_unused(); 874 kmap_flush_unused();
874 vm_unmap_aliases(); 875 }
875 } 876 }
876} 877}
877 878
diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c
index d4d52f5a1cf7..636ef4caa52d 100644
--- a/arch/x86/xen/mmu.c
+++ b/arch/x86/xen/mmu.c
@@ -246,11 +246,21 @@ xmaddr_t arbitrary_virt_to_machine(void *vaddr)
246{ 246{
247 unsigned long address = (unsigned long)vaddr; 247 unsigned long address = (unsigned long)vaddr;
248 unsigned int level; 248 unsigned int level;
249 pte_t *pte = lookup_address(address, &level); 249 pte_t *pte;
250 unsigned offset = address & ~PAGE_MASK; 250 unsigned offset;
251 251
252 BUG_ON(pte == NULL); 252 /*
253 * if the PFN is in the linear mapped vaddr range, we can just use
254 * the (quick) virt_to_machine() p2m lookup
255 */
256 if (virt_addr_valid(vaddr))
257 return virt_to_machine(vaddr);
253 258
259 /* otherwise we have to do a (slower) full page-table walk */
260
261 pte = lookup_address(address, &level);
262 BUG_ON(pte == NULL);
263 offset = address & ~PAGE_MASK;
254 return XMADDR(((phys_addr_t)pte_mfn(*pte) << PAGE_SHIFT) + offset); 264 return XMADDR(((phys_addr_t)pte_mfn(*pte) << PAGE_SHIFT) + offset);
255} 265}
256 266
@@ -410,7 +420,7 @@ void xen_ptep_modify_prot_commit(struct mm_struct *mm, unsigned long addr,
410 420
411 xen_mc_batch(); 421 xen_mc_batch();
412 422
413 u.ptr = virt_to_machine(ptep).maddr | MMU_PT_UPDATE_PRESERVE_AD; 423 u.ptr = arbitrary_virt_to_machine(ptep).maddr | MMU_PT_UPDATE_PRESERVE_AD;
414 u.val = pte_val_ma(pte); 424 u.val = pte_val_ma(pte);
415 xen_extend_mmu_update(&u); 425 xen_extend_mmu_update(&u);
416 426
@@ -651,12 +661,11 @@ void xen_set_pgd(pgd_t *ptr, pgd_t val)
651 * For 64-bit, we must skip the Xen hole in the middle of the address 661 * For 64-bit, we must skip the Xen hole in the middle of the address
652 * space, just after the big x86-64 virtual hole. 662 * space, just after the big x86-64 virtual hole.
653 */ 663 */
654static int xen_pgd_walk(struct mm_struct *mm, 664static int __xen_pgd_walk(struct mm_struct *mm, pgd_t *pgd,
655 int (*func)(struct mm_struct *mm, struct page *, 665 int (*func)(struct mm_struct *mm, struct page *,
656 enum pt_level), 666 enum pt_level),
657 unsigned long limit) 667 unsigned long limit)
658{ 668{
659 pgd_t *pgd = mm->pgd;
660 int flush = 0; 669 int flush = 0;
661 unsigned hole_low, hole_high; 670 unsigned hole_low, hole_high;
662 unsigned pgdidx_limit, pudidx_limit, pmdidx_limit; 671 unsigned pgdidx_limit, pudidx_limit, pmdidx_limit;
@@ -743,6 +752,14 @@ out:
743 return flush; 752 return flush;
744} 753}
745 754
755static int xen_pgd_walk(struct mm_struct *mm,
756 int (*func)(struct mm_struct *mm, struct page *,
757 enum pt_level),
758 unsigned long limit)
759{
760 return __xen_pgd_walk(mm, mm->pgd, func, limit);
761}
762
746/* If we're using split pte locks, then take the page's lock and 763/* If we're using split pte locks, then take the page's lock and
747 return a pointer to it. Otherwise return NULL. */ 764 return a pointer to it. Otherwise return NULL. */
748static spinlock_t *xen_pte_lock(struct page *page, struct mm_struct *mm) 765static spinlock_t *xen_pte_lock(struct page *page, struct mm_struct *mm)
@@ -840,13 +857,16 @@ static int xen_pin_page(struct mm_struct *mm, struct page *page,
840 read-only, and can be pinned. */ 857 read-only, and can be pinned. */
841static void __xen_pgd_pin(struct mm_struct *mm, pgd_t *pgd) 858static void __xen_pgd_pin(struct mm_struct *mm, pgd_t *pgd)
842{ 859{
860 vm_unmap_aliases();
861
843 xen_mc_batch(); 862 xen_mc_batch();
844 863
845 if (xen_pgd_walk(mm, xen_pin_page, USER_LIMIT)) { 864 if (__xen_pgd_walk(mm, pgd, xen_pin_page, USER_LIMIT)) {
846 /* re-enable interrupts for kmap_flush_unused */ 865 /* re-enable interrupts for flushing */
847 xen_mc_issue(0); 866 xen_mc_issue(0);
867
848 kmap_flush_unused(); 868 kmap_flush_unused();
849 vm_unmap_aliases(); 869
850 xen_mc_batch(); 870 xen_mc_batch();
851 } 871 }
852 872
@@ -864,7 +884,7 @@ static void __xen_pgd_pin(struct mm_struct *mm, pgd_t *pgd)
864#else /* CONFIG_X86_32 */ 884#else /* CONFIG_X86_32 */
865#ifdef CONFIG_X86_PAE 885#ifdef CONFIG_X86_PAE
866 /* Need to make sure unshared kernel PMD is pinnable */ 886 /* Need to make sure unshared kernel PMD is pinnable */
867 xen_pin_page(mm, virt_to_page(pgd_page(pgd[pgd_index(TASK_SIZE)])), 887 xen_pin_page(mm, pgd_page(pgd[pgd_index(TASK_SIZE)]),
868 PT_PMD); 888 PT_PMD);
869#endif 889#endif
870 xen_do_pin(MMUEXT_PIN_L3_TABLE, PFN_DOWN(__pa(pgd))); 890 xen_do_pin(MMUEXT_PIN_L3_TABLE, PFN_DOWN(__pa(pgd)));
@@ -981,11 +1001,11 @@ static void __xen_pgd_unpin(struct mm_struct *mm, pgd_t *pgd)
981 1001
982#ifdef CONFIG_X86_PAE 1002#ifdef CONFIG_X86_PAE
983 /* Need to make sure unshared kernel PMD is unpinned */ 1003 /* Need to make sure unshared kernel PMD is unpinned */
984 xen_unpin_page(mm, virt_to_page(pgd_page(pgd[pgd_index(TASK_SIZE)])), 1004 xen_unpin_page(mm, pgd_page(pgd[pgd_index(TASK_SIZE)]),
985 PT_PMD); 1005 PT_PMD);
986#endif 1006#endif
987 1007
988 xen_pgd_walk(mm, xen_unpin_page, USER_LIMIT); 1008 __xen_pgd_walk(mm, pgd, xen_unpin_page, USER_LIMIT);
989 1009
990 xen_mc_issue(0); 1010 xen_mc_issue(0);
991} 1011}
diff --git a/arch/x86/xen/smp.c b/arch/x86/xen/smp.c
index d77da613b1d2..acd9b6705e02 100644
--- a/arch/x86/xen/smp.c
+++ b/arch/x86/xen/smp.c
@@ -362,7 +362,7 @@ static void xen_cpu_die(unsigned int cpu)
362 alternatives_smp_switch(0); 362 alternatives_smp_switch(0);
363} 363}
364 364
365static void xen_play_dead(void) 365static void __cpuinit xen_play_dead(void) /* used only with CPU_HOTPLUG */
366{ 366{
367 play_dead_common(); 367 play_dead_common();
368 HYPERVISOR_vcpu_op(VCPUOP_down, smp_processor_id(), NULL); 368 HYPERVISOR_vcpu_op(VCPUOP_down, smp_processor_id(), NULL);
diff --git a/arch/x86/xen/xen-ops.h b/arch/x86/xen/xen-ops.h
index d7422dc2a55c..9e1afae8461f 100644
--- a/arch/x86/xen/xen-ops.h
+++ b/arch/x86/xen/xen-ops.h
@@ -49,7 +49,7 @@ bool xen_vcpu_stolen(int vcpu);
49 49
50void xen_mark_init_mm_pinned(void); 50void xen_mark_init_mm_pinned(void);
51 51
52void __init xen_setup_vcpu_info_placement(void); 52void xen_setup_vcpu_info_placement(void);
53 53
54#ifdef CONFIG_SMP 54#ifdef CONFIG_SMP
55void xen_smp_init(void); 55void xen_smp_init(void);
generated by cgit v1.2.2 (git 2.25.0) at 2024-12-12 14:17:48 -0500