aboutsummaryrefslogtreecommitdiffstats
path: root/arch/x86
diff options
context:
space:
mode:
Diffstat (limited to 'arch/x86')
-rw-r--r--arch/x86/Kconfig4
-rw-r--r--arch/x86/crypto/fpu.c1
-rw-r--r--arch/x86/ia32/ia32_aout.c1
-rw-r--r--arch/x86/ia32/ia32entry.S2
-rw-r--r--arch/x86/ia32/sys_ia32.c1
-rw-r--r--arch/x86/include/asm/amd_iommu_types.h3
-rw-r--r--arch/x86/include/asm/io.h1
-rw-r--r--arch/x86/include/asm/k8.h5
-rw-r--r--arch/x86/include/asm/lguest_hcall.h29
-rw-r--r--arch/x86/include/asm/mce.h8
-rw-r--r--arch/x86/include/asm/pgtable_32.h1
-rw-r--r--arch/x86/kernel/acpi/boot.c1
-rw-r--r--arch/x86/kernel/alternative.c1
-rw-r--r--arch/x86/kernel/amd_iommu.c22
-rw-r--r--arch/x86/kernel/amd_iommu_init.c50
-rw-r--r--arch/x86/kernel/apb_timer.c3
-rw-r--r--arch/x86/kernel/aperture_64.c15
-rw-r--r--arch/x86/kernel/apic/apic.c2
-rw-r--r--arch/x86/kernel/apic/es7000_32.c1
-rw-r--r--arch/x86/kernel/apic/io_apic.c4
-rw-r--r--arch/x86/kernel/apic/nmi.c1
-rw-r--r--arch/x86/kernel/apic/x2apic_uv_x.c1
-rw-r--r--arch/x86/kernel/bootflag.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/elanfreq.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/gx-suspmod.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/longrun.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/p4-clockmod.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/powernow-k6.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/powernow-k8.c3
-rw-r--r--arch/x86/kernel/cpu/cpufreq/speedstep-centrino.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/speedstep-ich.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/speedstep-lib.c1
-rw-r--r--arch/x86/kernel/cpu/cpufreq/speedstep-smi.c1
-rw-r--r--arch/x86/kernel/cpu/intel.c21
-rw-r--r--arch/x86/kernel/cpu/intel_cacheinfo.c4
-rw-r--r--arch/x86/kernel/cpu/mcheck/Makefile2
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce-apei.c138
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce-inject.c1
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce-internal.h23
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce.c80
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce_amd.c1
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce_intel.c1
-rw-r--r--arch/x86/kernel/cpu/mtrr/generic.c1
-rw-r--r--arch/x86/kernel/cpu/mtrr/if.c1
-rw-r--r--arch/x86/kernel/cpu/perf_event.c55
-rw-r--r--arch/x86/kernel/cpu/perf_event_amd.c80
-rw-r--r--arch/x86/kernel/cpu/perf_event_intel.c1
-rw-r--r--arch/x86/kernel/cpu/vmware.c2
-rw-r--r--arch/x86/kernel/cpuid.c1
-rw-r--r--arch/x86/kernel/crash.c6
-rw-r--r--arch/x86/kernel/crash_dump_32.c1
-rw-r--r--arch/x86/kernel/dumpstack.h13
-rw-r--r--arch/x86/kernel/e820.c24
-rw-r--r--arch/x86/kernel/hpet.c10
-rw-r--r--arch/x86/kernel/i387.c1
-rw-r--r--arch/x86/kernel/i8259.c1
-rw-r--r--arch/x86/kernel/irqinit.c1
-rw-r--r--arch/x86/kernel/k8.c2
-rw-r--r--arch/x86/kernel/kdebugfs.c1
-rw-r--r--arch/x86/kernel/kgdb.c2
-rw-r--r--arch/x86/kernel/kprobes.c27
-rw-r--r--arch/x86/kernel/ldt.c1
-rw-r--r--arch/x86/kernel/machine_kexec_64.c1
-rw-r--r--arch/x86/kernel/mca_32.c1
-rw-r--r--arch/x86/kernel/module.c1
-rw-r--r--arch/x86/kernel/mpparse.c4
-rw-r--r--arch/x86/kernel/msr.c1
-rw-r--r--arch/x86/kernel/pci-dma.c1
-rw-r--r--arch/x86/kernel/pci-gart_64.c4
-rw-r--r--arch/x86/kernel/pci-nommu.c1
-rw-r--r--arch/x86/kernel/process.c12
-rw-r--r--arch/x86/kernel/process_64.c4
-rw-r--r--arch/x86/kernel/ptrace.c1
-rw-r--r--arch/x86/kernel/setup.c15
-rw-r--r--arch/x86/kernel/smp.c1
-rw-r--r--arch/x86/kernel/smpboot.c5
-rw-r--r--arch/x86/kernel/tlb_uv.c1
-rw-r--r--arch/x86/kernel/uv_irq.c1
-rw-r--r--arch/x86/kernel/uv_time.c1
-rw-r--r--arch/x86/kernel/vmi_32.c1
-rw-r--r--arch/x86/kvm/i8254.c1
-rw-r--r--arch/x86/kvm/i8259.c1
-rw-r--r--arch/x86/kvm/lapic.c1
-rw-r--r--arch/x86/kvm/mmu.c12
-rw-r--r--arch/x86/kvm/svm.c34
-rw-r--r--arch/x86/kvm/vmx.c28
-rw-r--r--arch/x86/kvm/x86.c53
-rw-r--r--arch/x86/lguest/boot.c61
-rw-r--r--arch/x86/lguest/i386_head.S2
-rw-r--r--arch/x86/lib/rwsem_64.S2
-rw-r--r--arch/x86/mm/hugetlbpage.c1
-rw-r--r--arch/x86/mm/init.c1
-rw-r--r--arch/x86/mm/init_32.c2
-rw-r--r--arch/x86/mm/init_64.c1
-rw-r--r--arch/x86/mm/ioremap.c14
-rw-r--r--arch/x86/mm/kmmio.c1
-rw-r--r--arch/x86/mm/mmio-mod.c1
-rw-r--r--arch/x86/mm/pageattr.c2
-rw-r--r--arch/x86/mm/pat.c2
-rw-r--r--arch/x86/mm/pgtable.c1
-rw-r--r--arch/x86/mm/pgtable_32.c3
-rw-r--r--arch/x86/mm/srat_64.c3
-rw-r--r--arch/x86/pci/acpi.c85
-rw-r--r--arch/x86/pci/common.c1
-rw-r--r--arch/x86/pci/i386.c3
-rw-r--r--arch/x86/pci/irq.c1
-rw-r--r--arch/x86/pci/mmconfig-shared.c1
-rw-r--r--arch/x86/pci/mrst.c4
-rw-r--r--arch/x86/pci/pcbios.c1
-rw-r--r--arch/x86/power/hibernate_32.c1
-rw-r--r--arch/x86/power/hibernate_64.c1
-rw-r--r--arch/x86/power/hibernate_asm_32.S15
-rw-r--r--arch/x86/vdso/vma.c1
-rw-r--r--arch/x86/xen/debugfs.c1
-rw-r--r--arch/x86/xen/enlighten.c1
-rw-r--r--arch/x86/xen/mmu.c1
-rw-r--r--arch/x86/xen/smp.c1
-rw-r--r--arch/x86/xen/spinlock.c1
-rw-r--r--arch/x86/xen/time.c1
121 files changed, 778 insertions, 299 deletions
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 0eacb1ffb421..9458685902bd 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -1216,8 +1216,8 @@ config NUMA_EMU
1216 1216
1217config NODES_SHIFT 1217config NODES_SHIFT
1218 int "Maximum NUMA Nodes (as a power of 2)" if !MAXSMP 1218 int "Maximum NUMA Nodes (as a power of 2)" if !MAXSMP
1219 range 1 9 1219 range 1 10
1220 default "9" if MAXSMP 1220 default "10" if MAXSMP
1221 default "6" if X86_64 1221 default "6" if X86_64
1222 default "4" if X86_NUMAQ 1222 default "4" if X86_NUMAQ
1223 default "3" 1223 default "3"
diff --git a/arch/x86/crypto/fpu.c b/arch/x86/crypto/fpu.c
index daef6cd2b45d..1a8f8649c035 100644
--- a/arch/x86/crypto/fpu.c
+++ b/arch/x86/crypto/fpu.c
@@ -16,6 +16,7 @@
16#include <linux/init.h> 16#include <linux/init.h>
17#include <linux/kernel.h> 17#include <linux/kernel.h>
18#include <linux/module.h> 18#include <linux/module.h>
19#include <linux/slab.h>
19#include <asm/i387.h> 20#include <asm/i387.h>
20 21
21struct crypto_fpu_ctx { 22struct crypto_fpu_ctx {
diff --git a/arch/x86/ia32/ia32_aout.c b/arch/x86/ia32/ia32_aout.c
index 280c019cfad8..0350311906ae 100644
--- a/arch/x86/ia32/ia32_aout.c
+++ b/arch/x86/ia32/ia32_aout.c
@@ -21,7 +21,6 @@
21#include <linux/fcntl.h> 21#include <linux/fcntl.h>
22#include <linux/ptrace.h> 22#include <linux/ptrace.h>
23#include <linux/user.h> 23#include <linux/user.h>
24#include <linux/slab.h>
25#include <linux/binfmts.h> 24#include <linux/binfmts.h>
26#include <linux/personality.h> 25#include <linux/personality.h>
27#include <linux/init.h> 26#include <linux/init.h>
diff --git a/arch/x86/ia32/ia32entry.S b/arch/x86/ia32/ia32entry.S
index 59b4556a5b92..e790bc1fbfa3 100644
--- a/arch/x86/ia32/ia32entry.S
+++ b/arch/x86/ia32/ia32entry.S
@@ -626,7 +626,7 @@ ia32_sys_call_table:
626 .quad stub32_sigreturn 626 .quad stub32_sigreturn
627 .quad stub32_clone /* 120 */ 627 .quad stub32_clone /* 120 */
628 .quad sys_setdomainname 628 .quad sys_setdomainname
629 .quad sys_uname 629 .quad sys_newuname
630 .quad sys_modify_ldt 630 .quad sys_modify_ldt
631 .quad compat_sys_adjtimex 631 .quad compat_sys_adjtimex
632 .quad sys32_mprotect /* 125 */ 632 .quad sys32_mprotect /* 125 */
diff --git a/arch/x86/ia32/sys_ia32.c b/arch/x86/ia32/sys_ia32.c
index 74c35431b7d8..626be156d88d 100644
--- a/arch/x86/ia32/sys_ia32.c
+++ b/arch/x86/ia32/sys_ia32.c
@@ -40,6 +40,7 @@
40#include <linux/ptrace.h> 40#include <linux/ptrace.h>
41#include <linux/highuid.h> 41#include <linux/highuid.h>
42#include <linux/sysctl.h> 42#include <linux/sysctl.h>
43#include <linux/slab.h>
43#include <asm/mman.h> 44#include <asm/mman.h>
44#include <asm/types.h> 45#include <asm/types.h>
45#include <asm/uaccess.h> 46#include <asm/uaccess.h>
diff --git a/arch/x86/include/asm/amd_iommu_types.h b/arch/x86/include/asm/amd_iommu_types.h
index ba19ad4c47d0..86a0ff0aeac7 100644
--- a/arch/x86/include/asm/amd_iommu_types.h
+++ b/arch/x86/include/asm/amd_iommu_types.h
@@ -21,6 +21,7 @@
21#define _ASM_X86_AMD_IOMMU_TYPES_H 21#define _ASM_X86_AMD_IOMMU_TYPES_H
22 22
23#include <linux/types.h> 23#include <linux/types.h>
24#include <linux/mutex.h>
24#include <linux/list.h> 25#include <linux/list.h>
25#include <linux/spinlock.h> 26#include <linux/spinlock.h>
26 27
@@ -140,6 +141,7 @@
140 141
141/* constants to configure the command buffer */ 142/* constants to configure the command buffer */
142#define CMD_BUFFER_SIZE 8192 143#define CMD_BUFFER_SIZE 8192
144#define CMD_BUFFER_UNINITIALIZED 1
143#define CMD_BUFFER_ENTRIES 512 145#define CMD_BUFFER_ENTRIES 512
144#define MMIO_CMD_SIZE_SHIFT 56 146#define MMIO_CMD_SIZE_SHIFT 56
145#define MMIO_CMD_SIZE_512 (0x9ULL << MMIO_CMD_SIZE_SHIFT) 147#define MMIO_CMD_SIZE_512 (0x9ULL << MMIO_CMD_SIZE_SHIFT)
@@ -237,6 +239,7 @@ struct protection_domain {
237 struct list_head list; /* for list of all protection domains */ 239 struct list_head list; /* for list of all protection domains */
238 struct list_head dev_list; /* List of all devices in this domain */ 240 struct list_head dev_list; /* List of all devices in this domain */
239 spinlock_t lock; /* mostly used to lock the page table*/ 241 spinlock_t lock; /* mostly used to lock the page table*/
242 struct mutex api_lock; /* protect page tables in the iommu-api path */
240 u16 id; /* the domain id written to the device table */ 243 u16 id; /* the domain id written to the device table */
241 int mode; /* paging mode (0-6 levels) */ 244 int mode; /* paging mode (0-6 levels) */
242 u64 *pt_root; /* page table root pointer */ 245 u64 *pt_root; /* page table root pointer */
diff --git a/arch/x86/include/asm/io.h b/arch/x86/include/asm/io.h
index a1dcfa3ab17d..30a3e9776123 100644
--- a/arch/x86/include/asm/io.h
+++ b/arch/x86/include/asm/io.h
@@ -347,6 +347,7 @@ extern void __iomem *early_ioremap(resource_size_t phys_addr,
347extern void __iomem *early_memremap(resource_size_t phys_addr, 347extern void __iomem *early_memremap(resource_size_t phys_addr,
348 unsigned long size); 348 unsigned long size);
349extern void early_iounmap(void __iomem *addr, unsigned long size); 349extern void early_iounmap(void __iomem *addr, unsigned long size);
350extern void fixup_early_ioremap(void);
350 351
351#define IO_SPACE_LIMIT 0xffff 352#define IO_SPACE_LIMIT 0xffff
352 353
diff --git a/arch/x86/include/asm/k8.h b/arch/x86/include/asm/k8.h
index f70e60071fe8..af00bd1d2089 100644
--- a/arch/x86/include/asm/k8.h
+++ b/arch/x86/include/asm/k8.h
@@ -16,11 +16,16 @@ extern int k8_numa_init(unsigned long start_pfn, unsigned long end_pfn);
16extern int k8_scan_nodes(void); 16extern int k8_scan_nodes(void);
17 17
18#ifdef CONFIG_K8_NB 18#ifdef CONFIG_K8_NB
19extern int num_k8_northbridges;
20
19static inline struct pci_dev *node_to_k8_nb_misc(int node) 21static inline struct pci_dev *node_to_k8_nb_misc(int node)
20{ 22{
21 return (node < num_k8_northbridges) ? k8_northbridges[node] : NULL; 23 return (node < num_k8_northbridges) ? k8_northbridges[node] : NULL;
22} 24}
25
23#else 26#else
27#define num_k8_northbridges 0
28
24static inline struct pci_dev *node_to_k8_nb_misc(int node) 29static inline struct pci_dev *node_to_k8_nb_misc(int node)
25{ 30{
26 return NULL; 31 return NULL;
diff --git a/arch/x86/include/asm/lguest_hcall.h b/arch/x86/include/asm/lguest_hcall.h
index ba0eed8aa1a6..b60f2924c413 100644
--- a/arch/x86/include/asm/lguest_hcall.h
+++ b/arch/x86/include/asm/lguest_hcall.h
@@ -28,22 +28,39 @@
28 28
29#ifndef __ASSEMBLY__ 29#ifndef __ASSEMBLY__
30#include <asm/hw_irq.h> 30#include <asm/hw_irq.h>
31#include <asm/kvm_para.h>
32 31
33/*G:030 32/*G:030
34 * But first, how does our Guest contact the Host to ask for privileged 33 * But first, how does our Guest contact the Host to ask for privileged
35 * operations? There are two ways: the direct way is to make a "hypercall", 34 * operations? There are two ways: the direct way is to make a "hypercall",
36 * to make requests of the Host Itself. 35 * to make requests of the Host Itself.
37 * 36 *
38 * We use the KVM hypercall mechanism, though completely different hypercall 37 * Our hypercall mechanism uses the highest unused trap code (traps 32 and
39 * numbers. Seventeen hypercalls are available: the hypercall number is put in 38 * above are used by real hardware interrupts). Seventeen hypercalls are
40 * the %eax register, and the arguments (when required) are placed in %ebx, 39 * available: the hypercall number is put in the %eax register, and the
41 * %ecx, %edx and %esi. If a return value makes sense, it's returned in %eax. 40 * arguments (when required) are placed in %ebx, %ecx, %edx and %esi.
41 * If a return value makes sense, it's returned in %eax.
42 * 42 *
43 * Grossly invalid calls result in Sudden Death at the hands of the vengeful 43 * Grossly invalid calls result in Sudden Death at the hands of the vengeful
44 * Host, rather than returning failure. This reflects Winston Churchill's 44 * Host, rather than returning failure. This reflects Winston Churchill's
45 * definition of a gentleman: "someone who is only rude intentionally". 45 * definition of a gentleman: "someone who is only rude intentionally".
46:*/ 46 */
47static inline unsigned long
48hcall(unsigned long call,
49 unsigned long arg1, unsigned long arg2, unsigned long arg3,
50 unsigned long arg4)
51{
52 /* "int" is the Intel instruction to trigger a trap. */
53 asm volatile("int $" __stringify(LGUEST_TRAP_ENTRY)
54 /* The call in %eax (aka "a") might be overwritten */
55 : "=a"(call)
56 /* The arguments are in %eax, %ebx, %ecx, %edx & %esi */
57 : "a"(call), "b"(arg1), "c"(arg2), "d"(arg3), "S"(arg4)
58 /* "memory" means this might write somewhere in memory.
59 * This isn't true for all calls, but it's safe to tell
60 * gcc that it might happen so it doesn't get clever. */
61 : "memory");
62 return call;
63}
47 64
48/* Can't use our min() macro here: needs to be a constant */ 65/* Can't use our min() macro here: needs to be a constant */
49#define LGUEST_IRQS (NR_IRQS < 32 ? NR_IRQS: 32) 66#define LGUEST_IRQS (NR_IRQS < 32 ? NR_IRQS: 32)
diff --git a/arch/x86/include/asm/mce.h b/arch/x86/include/asm/mce.h
index 6c3fdd631ed3..f32a4301c4d4 100644
--- a/arch/x86/include/asm/mce.h
+++ b/arch/x86/include/asm/mce.h
@@ -225,5 +225,13 @@ extern void mcheck_intel_therm_init(void);
225static inline void mcheck_intel_therm_init(void) { } 225static inline void mcheck_intel_therm_init(void) { }
226#endif 226#endif
227 227
228/*
229 * Used by APEI to report memory error via /dev/mcelog
230 */
231
232struct cper_sec_mem_err;
233extern void apei_mce_report_mem_error(int corrected,
234 struct cper_sec_mem_err *mem_err);
235
228#endif /* __KERNEL__ */ 236#endif /* __KERNEL__ */
229#endif /* _ASM_X86_MCE_H */ 237#endif /* _ASM_X86_MCE_H */
diff --git a/arch/x86/include/asm/pgtable_32.h b/arch/x86/include/asm/pgtable_32.h
index 47339a1ac7b6..2984a25ff383 100644
--- a/arch/x86/include/asm/pgtable_32.h
+++ b/arch/x86/include/asm/pgtable_32.h
@@ -19,7 +19,6 @@
19#include <asm/paravirt.h> 19#include <asm/paravirt.h>
20 20
21#include <linux/bitops.h> 21#include <linux/bitops.h>
22#include <linux/slab.h>
23#include <linux/list.h> 22#include <linux/list.h>
24#include <linux/spinlock.h> 23#include <linux/spinlock.h>
25 24
diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c
index 0061ea263061..cd40aba6aa95 100644
--- a/arch/x86/kernel/acpi/boot.c
+++ b/arch/x86/kernel/acpi/boot.c
@@ -31,6 +31,7 @@
31#include <linux/module.h> 31#include <linux/module.h>
32#include <linux/dmi.h> 32#include <linux/dmi.h>
33#include <linux/irq.h> 33#include <linux/irq.h>
34#include <linux/slab.h>
34#include <linux/bootmem.h> 35#include <linux/bootmem.h>
35#include <linux/ioport.h> 36#include <linux/ioport.h>
36#include <linux/pci.h> 37#include <linux/pci.h>
diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c
index 3a4bf35c179b..1a160d5d44d0 100644
--- a/arch/x86/kernel/alternative.c
+++ b/arch/x86/kernel/alternative.c
@@ -8,6 +8,7 @@
8#include <linux/vmalloc.h> 8#include <linux/vmalloc.h>
9#include <linux/memory.h> 9#include <linux/memory.h>
10#include <linux/stop_machine.h> 10#include <linux/stop_machine.h>
11#include <linux/slab.h>
11#include <asm/alternative.h> 12#include <asm/alternative.h>
12#include <asm/sections.h> 13#include <asm/sections.h>
13#include <asm/pgtable.h> 14#include <asm/pgtable.h>
diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c
index adb0ba025702..f854d89b7edf 100644
--- a/arch/x86/kernel/amd_iommu.c
+++ b/arch/x86/kernel/amd_iommu.c
@@ -18,8 +18,8 @@
18 */ 18 */
19 19
20#include <linux/pci.h> 20#include <linux/pci.h>
21#include <linux/gfp.h>
22#include <linux/bitmap.h> 21#include <linux/bitmap.h>
22#include <linux/slab.h>
23#include <linux/debugfs.h> 23#include <linux/debugfs.h>
24#include <linux/scatterlist.h> 24#include <linux/scatterlist.h>
25#include <linux/dma-mapping.h> 25#include <linux/dma-mapping.h>
@@ -118,7 +118,7 @@ static bool check_device(struct device *dev)
118 return false; 118 return false;
119 119
120 /* No device or no PCI device */ 120 /* No device or no PCI device */
121 if (!dev || dev->bus != &pci_bus_type) 121 if (dev->bus != &pci_bus_type)
122 return false; 122 return false;
123 123
124 devid = get_device_id(dev); 124 devid = get_device_id(dev);
@@ -392,6 +392,7 @@ static int __iommu_queue_command(struct amd_iommu *iommu, struct iommu_cmd *cmd)
392 u32 tail, head; 392 u32 tail, head;
393 u8 *target; 393 u8 *target;
394 394
395 WARN_ON(iommu->cmd_buf_size & CMD_BUFFER_UNINITIALIZED);
395 tail = readl(iommu->mmio_base + MMIO_CMD_TAIL_OFFSET); 396 tail = readl(iommu->mmio_base + MMIO_CMD_TAIL_OFFSET);
396 target = iommu->cmd_buf + tail; 397 target = iommu->cmd_buf + tail;
397 memcpy_toio(target, cmd, sizeof(*cmd)); 398 memcpy_toio(target, cmd, sizeof(*cmd));
@@ -2186,7 +2187,7 @@ static void prealloc_protection_domains(void)
2186 struct dma_ops_domain *dma_dom; 2187 struct dma_ops_domain *dma_dom;
2187 u16 devid; 2188 u16 devid;
2188 2189
2189 while ((dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, dev)) != NULL) { 2190 for_each_pci_dev(dev) {
2190 2191
2191 /* Do we handle this device? */ 2192 /* Do we handle this device? */
2192 if (!check_device(&dev->dev)) 2193 if (!check_device(&dev->dev))
@@ -2298,7 +2299,7 @@ static void cleanup_domain(struct protection_domain *domain)
2298 list_for_each_entry_safe(dev_data, next, &domain->dev_list, list) { 2299 list_for_each_entry_safe(dev_data, next, &domain->dev_list, list) {
2299 struct device *dev = dev_data->dev; 2300 struct device *dev = dev_data->dev;
2300 2301
2301 do_detach(dev); 2302 __detach_device(dev);
2302 atomic_set(&dev_data->bind, 0); 2303 atomic_set(&dev_data->bind, 0);
2303 } 2304 }
2304 2305
@@ -2327,6 +2328,7 @@ static struct protection_domain *protection_domain_alloc(void)
2327 return NULL; 2328 return NULL;
2328 2329
2329 spin_lock_init(&domain->lock); 2330 spin_lock_init(&domain->lock);
2331 mutex_init(&domain->api_lock);
2330 domain->id = domain_id_alloc(); 2332 domain->id = domain_id_alloc();
2331 if (!domain->id) 2333 if (!domain->id)
2332 goto out_err; 2334 goto out_err;
@@ -2379,9 +2381,7 @@ static void amd_iommu_domain_destroy(struct iommu_domain *dom)
2379 2381
2380 free_pagetable(domain); 2382 free_pagetable(domain);
2381 2383
2382 domain_id_free(domain->id); 2384 protection_domain_free(domain);
2383
2384 kfree(domain);
2385 2385
2386 dom->priv = NULL; 2386 dom->priv = NULL;
2387} 2387}
@@ -2456,6 +2456,8 @@ static int amd_iommu_map_range(struct iommu_domain *dom,
2456 iova &= PAGE_MASK; 2456 iova &= PAGE_MASK;
2457 paddr &= PAGE_MASK; 2457 paddr &= PAGE_MASK;
2458 2458
2459 mutex_lock(&domain->api_lock);
2460
2459 for (i = 0; i < npages; ++i) { 2461 for (i = 0; i < npages; ++i) {
2460 ret = iommu_map_page(domain, iova, paddr, prot, PM_MAP_4k); 2462 ret = iommu_map_page(domain, iova, paddr, prot, PM_MAP_4k);
2461 if (ret) 2463 if (ret)
@@ -2465,6 +2467,8 @@ static int amd_iommu_map_range(struct iommu_domain *dom,
2465 paddr += PAGE_SIZE; 2467 paddr += PAGE_SIZE;
2466 } 2468 }
2467 2469
2470 mutex_unlock(&domain->api_lock);
2471
2468 return 0; 2472 return 0;
2469} 2473}
2470 2474
@@ -2477,12 +2481,16 @@ static void amd_iommu_unmap_range(struct iommu_domain *dom,
2477 2481
2478 iova &= PAGE_MASK; 2482 iova &= PAGE_MASK;
2479 2483
2484 mutex_lock(&domain->api_lock);
2485
2480 for (i = 0; i < npages; ++i) { 2486 for (i = 0; i < npages; ++i) {
2481 iommu_unmap_page(domain, iova, PM_MAP_4k); 2487 iommu_unmap_page(domain, iova, PM_MAP_4k);
2482 iova += PAGE_SIZE; 2488 iova += PAGE_SIZE;
2483 } 2489 }
2484 2490
2485 iommu_flush_tlb_pde(domain); 2491 iommu_flush_tlb_pde(domain);
2492
2493 mutex_unlock(&domain->api_lock);
2486} 2494}
2487 2495
2488static phys_addr_t amd_iommu_iova_to_phys(struct iommu_domain *dom, 2496static phys_addr_t amd_iommu_iova_to_phys(struct iommu_domain *dom,
diff --git a/arch/x86/kernel/amd_iommu_init.c b/arch/x86/kernel/amd_iommu_init.c
index 9dc91b431470..6360abf993d4 100644
--- a/arch/x86/kernel/amd_iommu_init.c
+++ b/arch/x86/kernel/amd_iommu_init.c
@@ -19,8 +19,8 @@
19 19
20#include <linux/pci.h> 20#include <linux/pci.h>
21#include <linux/acpi.h> 21#include <linux/acpi.h>
22#include <linux/gfp.h>
23#include <linux/list.h> 22#include <linux/list.h>
23#include <linux/slab.h>
24#include <linux/sysdev.h> 24#include <linux/sysdev.h>
25#include <linux/interrupt.h> 25#include <linux/interrupt.h>
26#include <linux/msi.h> 26#include <linux/msi.h>
@@ -138,9 +138,9 @@ int amd_iommus_present;
138bool amd_iommu_np_cache __read_mostly; 138bool amd_iommu_np_cache __read_mostly;
139 139
140/* 140/*
141 * Set to true if ACPI table parsing and hardware intialization went properly 141 * The ACPI table parsing functions set this variable on an error
142 */ 142 */
143static bool amd_iommu_initialized; 143static int __initdata amd_iommu_init_err;
144 144
145/* 145/*
146 * List of protection domains - used during resume 146 * List of protection domains - used during resume
@@ -391,9 +391,11 @@ static int __init find_last_devid_acpi(struct acpi_table_header *table)
391 */ 391 */
392 for (i = 0; i < table->length; ++i) 392 for (i = 0; i < table->length; ++i)
393 checksum += p[i]; 393 checksum += p[i];
394 if (checksum != 0) 394 if (checksum != 0) {
395 /* ACPI table corrupt */ 395 /* ACPI table corrupt */
396 return -ENODEV; 396 amd_iommu_init_err = -ENODEV;
397 return 0;
398 }
397 399
398 p += IVRS_HEADER_LENGTH; 400 p += IVRS_HEADER_LENGTH;
399 401
@@ -436,7 +438,7 @@ static u8 * __init alloc_command_buffer(struct amd_iommu *iommu)
436 if (cmd_buf == NULL) 438 if (cmd_buf == NULL)
437 return NULL; 439 return NULL;
438 440
439 iommu->cmd_buf_size = CMD_BUFFER_SIZE; 441 iommu->cmd_buf_size = CMD_BUFFER_SIZE | CMD_BUFFER_UNINITIALIZED;
440 442
441 return cmd_buf; 443 return cmd_buf;
442} 444}
@@ -472,12 +474,13 @@ static void iommu_enable_command_buffer(struct amd_iommu *iommu)
472 &entry, sizeof(entry)); 474 &entry, sizeof(entry));
473 475
474 amd_iommu_reset_cmd_buffer(iommu); 476 amd_iommu_reset_cmd_buffer(iommu);
477 iommu->cmd_buf_size &= ~(CMD_BUFFER_UNINITIALIZED);
475} 478}
476 479
477static void __init free_command_buffer(struct amd_iommu *iommu) 480static void __init free_command_buffer(struct amd_iommu *iommu)
478{ 481{
479 free_pages((unsigned long)iommu->cmd_buf, 482 free_pages((unsigned long)iommu->cmd_buf,
480 get_order(iommu->cmd_buf_size)); 483 get_order(iommu->cmd_buf_size & ~(CMD_BUFFER_UNINITIALIZED)));
481} 484}
482 485
483/* allocates the memory where the IOMMU will log its events to */ 486/* allocates the memory where the IOMMU will log its events to */
@@ -920,11 +923,16 @@ static int __init init_iommu_all(struct acpi_table_header *table)
920 h->mmio_phys); 923 h->mmio_phys);
921 924
922 iommu = kzalloc(sizeof(struct amd_iommu), GFP_KERNEL); 925 iommu = kzalloc(sizeof(struct amd_iommu), GFP_KERNEL);
923 if (iommu == NULL) 926 if (iommu == NULL) {
924 return -ENOMEM; 927 amd_iommu_init_err = -ENOMEM;
928 return 0;
929 }
930
925 ret = init_iommu_one(iommu, h); 931 ret = init_iommu_one(iommu, h);
926 if (ret) 932 if (ret) {
927 return ret; 933 amd_iommu_init_err = ret;
934 return 0;
935 }
928 break; 936 break;
929 default: 937 default:
930 break; 938 break;
@@ -934,8 +942,6 @@ static int __init init_iommu_all(struct acpi_table_header *table)
934 } 942 }
935 WARN_ON(p != end); 943 WARN_ON(p != end);
936 944
937 amd_iommu_initialized = true;
938
939 return 0; 945 return 0;
940} 946}
941 947
@@ -1211,6 +1217,10 @@ static int __init amd_iommu_init(void)
1211 if (acpi_table_parse("IVRS", find_last_devid_acpi) != 0) 1217 if (acpi_table_parse("IVRS", find_last_devid_acpi) != 0)
1212 return -ENODEV; 1218 return -ENODEV;
1213 1219
1220 ret = amd_iommu_init_err;
1221 if (ret)
1222 goto out;
1223
1214 dev_table_size = tbl_size(DEV_TABLE_ENTRY_SIZE); 1224 dev_table_size = tbl_size(DEV_TABLE_ENTRY_SIZE);
1215 alias_table_size = tbl_size(ALIAS_TABLE_ENTRY_SIZE); 1225 alias_table_size = tbl_size(ALIAS_TABLE_ENTRY_SIZE);
1216 rlookup_table_size = tbl_size(RLOOKUP_TABLE_ENTRY_SIZE); 1226 rlookup_table_size = tbl_size(RLOOKUP_TABLE_ENTRY_SIZE);
@@ -1270,12 +1280,19 @@ static int __init amd_iommu_init(void)
1270 if (acpi_table_parse("IVRS", init_iommu_all) != 0) 1280 if (acpi_table_parse("IVRS", init_iommu_all) != 0)
1271 goto free; 1281 goto free;
1272 1282
1273 if (!amd_iommu_initialized) 1283 if (amd_iommu_init_err) {
1284 ret = amd_iommu_init_err;
1274 goto free; 1285 goto free;
1286 }
1275 1287
1276 if (acpi_table_parse("IVRS", init_memory_definitions) != 0) 1288 if (acpi_table_parse("IVRS", init_memory_definitions) != 0)
1277 goto free; 1289 goto free;
1278 1290
1291 if (amd_iommu_init_err) {
1292 ret = amd_iommu_init_err;
1293 goto free;
1294 }
1295
1279 ret = sysdev_class_register(&amd_iommu_sysdev_class); 1296 ret = sysdev_class_register(&amd_iommu_sysdev_class);
1280 if (ret) 1297 if (ret)
1281 goto free; 1298 goto free;
@@ -1288,6 +1305,8 @@ static int __init amd_iommu_init(void)
1288 if (ret) 1305 if (ret)
1289 goto free; 1306 goto free;
1290 1307
1308 enable_iommus();
1309
1291 if (iommu_pass_through) 1310 if (iommu_pass_through)
1292 ret = amd_iommu_init_passthrough(); 1311 ret = amd_iommu_init_passthrough();
1293 else 1312 else
@@ -1300,8 +1319,6 @@ static int __init amd_iommu_init(void)
1300 1319
1301 amd_iommu_init_notifier(); 1320 amd_iommu_init_notifier();
1302 1321
1303 enable_iommus();
1304
1305 if (iommu_pass_through) 1322 if (iommu_pass_through)
1306 goto out; 1323 goto out;
1307 1324
@@ -1315,6 +1332,7 @@ out:
1315 return ret; 1332 return ret;
1316 1333
1317free: 1334free:
1335 disable_iommus();
1318 1336
1319 amd_iommu_uninit_devices(); 1337 amd_iommu_uninit_devices();
1320 1338
diff --git a/arch/x86/kernel/apb_timer.c b/arch/x86/kernel/apb_timer.c
index 4b7099526d2c..a35347501d36 100644
--- a/arch/x86/kernel/apb_timer.c
+++ b/arch/x86/kernel/apb_timer.c
@@ -33,6 +33,7 @@
33#include <linux/errno.h> 33#include <linux/errno.h>
34#include <linux/init.h> 34#include <linux/init.h>
35#include <linux/sysdev.h> 35#include <linux/sysdev.h>
36#include <linux/slab.h>
36#include <linux/pm.h> 37#include <linux/pm.h>
37#include <linux/pci.h> 38#include <linux/pci.h>
38#include <linux/sfi.h> 39#include <linux/sfi.h>
@@ -428,7 +429,7 @@ static int apbt_cpuhp_notify(struct notifier_block *n,
428 429
429static __init int apbt_late_init(void) 430static __init int apbt_late_init(void)
430{ 431{
431 if (disable_apbt_percpu) 432 if (disable_apbt_percpu || !apb_timer_block_enabled)
432 return 0; 433 return 0;
433 /* This notifier should be called after workqueue is ready */ 434 /* This notifier should be called after workqueue is ready */
434 hotcpu_notifier(apbt_cpuhp_notify, -20); 435 hotcpu_notifier(apbt_cpuhp_notify, -20);
diff --git a/arch/x86/kernel/aperture_64.c b/arch/x86/kernel/aperture_64.c
index 3704997e8b25..b5d8b0bcf235 100644
--- a/arch/x86/kernel/aperture_64.c
+++ b/arch/x86/kernel/aperture_64.c
@@ -393,6 +393,7 @@ void __init gart_iommu_hole_init(void)
393 for (i = 0; i < ARRAY_SIZE(bus_dev_ranges); i++) { 393 for (i = 0; i < ARRAY_SIZE(bus_dev_ranges); i++) {
394 int bus; 394 int bus;
395 int dev_base, dev_limit; 395 int dev_base, dev_limit;
396 u32 ctl;
396 397
397 bus = bus_dev_ranges[i].bus; 398 bus = bus_dev_ranges[i].bus;
398 dev_base = bus_dev_ranges[i].dev_base; 399 dev_base = bus_dev_ranges[i].dev_base;
@@ -406,7 +407,19 @@ void __init gart_iommu_hole_init(void)
406 gart_iommu_aperture = 1; 407 gart_iommu_aperture = 1;
407 x86_init.iommu.iommu_init = gart_iommu_init; 408 x86_init.iommu.iommu_init = gart_iommu_init;
408 409
409 aper_order = (read_pci_config(bus, slot, 3, AMD64_GARTAPERTURECTL) >> 1) & 7; 410 ctl = read_pci_config(bus, slot, 3,
411 AMD64_GARTAPERTURECTL);
412
413 /*
414 * Before we do anything else disable the GART. It may
415 * still be enabled if we boot into a crash-kernel here.
416 * Reconfiguring the GART while it is enabled could have
417 * unknown side-effects.
418 */
419 ctl &= ~GARTEN;
420 write_pci_config(bus, slot, 3, AMD64_GARTAPERTURECTL, ctl);
421
422 aper_order = (ctl >> 1) & 7;
410 aper_size = (32 * 1024 * 1024) << aper_order; 423 aper_size = (32 * 1024 * 1024) << aper_order;
411 aper_base = read_pci_config(bus, slot, 3, AMD64_GARTAPERTUREBASE) & 0x7fff; 424 aper_base = read_pci_config(bus, slot, 3, AMD64_GARTAPERTUREBASE) & 0x7fff;
412 aper_base <<= 25; 425 aper_base <<= 25;
diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c
index 00187f1fcfb7..e5a4a1e01618 100644
--- a/arch/x86/kernel/apic/apic.c
+++ b/arch/x86/kernel/apic/apic.c
@@ -1640,8 +1640,10 @@ int __init APIC_init_uniprocessor(void)
1640 } 1640 }
1641#endif 1641#endif
1642 1642
1643#ifndef CONFIG_SMP
1643 enable_IR_x2apic(); 1644 enable_IR_x2apic();
1644 default_setup_apic_routing(); 1645 default_setup_apic_routing();
1646#endif
1645 1647
1646 verify_local_APIC(); 1648 verify_local_APIC();
1647 connect_bsp_APIC(); 1649 connect_bsp_APIC();
diff --git a/arch/x86/kernel/apic/es7000_32.c b/arch/x86/kernel/apic/es7000_32.c
index dd2b5f264643..03ba1b895f5e 100644
--- a/arch/x86/kernel/apic/es7000_32.c
+++ b/arch/x86/kernel/apic/es7000_32.c
@@ -42,6 +42,7 @@
42#include <linux/errno.h> 42#include <linux/errno.h>
43#include <linux/acpi.h> 43#include <linux/acpi.h>
44#include <linux/init.h> 44#include <linux/init.h>
45#include <linux/gfp.h>
45#include <linux/nmi.h> 46#include <linux/nmi.h>
46#include <linux/smp.h> 47#include <linux/smp.h>
47#include <linux/io.h> 48#include <linux/io.h>
diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
index 463de9a858ad..eb2789c3f721 100644
--- a/arch/x86/kernel/apic/io_apic.c
+++ b/arch/x86/kernel/apic/io_apic.c
@@ -36,6 +36,7 @@
36#include <linux/freezer.h> 36#include <linux/freezer.h>
37#include <linux/kthread.h> 37#include <linux/kthread.h>
38#include <linux/jiffies.h> /* time_after() */ 38#include <linux/jiffies.h> /* time_after() */
39#include <linux/slab.h>
39#ifdef CONFIG_ACPI 40#ifdef CONFIG_ACPI
40#include <acpi/acpi_bus.h> 41#include <acpi/acpi_bus.h>
41#endif 42#endif
@@ -2544,6 +2545,9 @@ void irq_force_complete_move(int irq)
2544 struct irq_desc *desc = irq_to_desc(irq); 2545 struct irq_desc *desc = irq_to_desc(irq);
2545 struct irq_cfg *cfg = desc->chip_data; 2546 struct irq_cfg *cfg = desc->chip_data;
2546 2547
2548 if (!cfg)
2549 return;
2550
2547 __irq_complete_move(&desc, cfg->vector); 2551 __irq_complete_move(&desc, cfg->vector);
2548} 2552}
2549#else 2553#else
diff --git a/arch/x86/kernel/apic/nmi.c b/arch/x86/kernel/apic/nmi.c
index 8aa65adbd25d..1edaf15c0b8e 100644
--- a/arch/x86/kernel/apic/nmi.c
+++ b/arch/x86/kernel/apic/nmi.c
@@ -18,6 +18,7 @@
18#include <linux/delay.h> 18#include <linux/delay.h>
19#include <linux/interrupt.h> 19#include <linux/interrupt.h>
20#include <linux/module.h> 20#include <linux/module.h>
21#include <linux/slab.h>
21#include <linux/sysdev.h> 22#include <linux/sysdev.h>
22#include <linux/sysctl.h> 23#include <linux/sysctl.h>
23#include <linux/percpu.h> 24#include <linux/percpu.h>
diff --git a/arch/x86/kernel/apic/x2apic_uv_x.c b/arch/x86/kernel/apic/x2apic_uv_x.c
index 49dbeaef2a27..c085d52dbaf2 100644
--- a/arch/x86/kernel/apic/x2apic_uv_x.c
+++ b/arch/x86/kernel/apic/x2apic_uv_x.c
@@ -17,6 +17,7 @@
17#include <linux/ctype.h> 17#include <linux/ctype.h>
18#include <linux/sched.h> 18#include <linux/sched.h>
19#include <linux/timer.h> 19#include <linux/timer.h>
20#include <linux/slab.h>
20#include <linux/cpu.h> 21#include <linux/cpu.h>
21#include <linux/init.h> 22#include <linux/init.h>
22#include <linux/io.h> 23#include <linux/io.h>
diff --git a/arch/x86/kernel/bootflag.c b/arch/x86/kernel/bootflag.c
index 30f25a75fe28..5de7f4c56971 100644
--- a/arch/x86/kernel/bootflag.c
+++ b/arch/x86/kernel/bootflag.c
@@ -5,7 +5,6 @@
5#include <linux/kernel.h> 5#include <linux/kernel.h>
6#include <linux/init.h> 6#include <linux/init.h>
7#include <linux/string.h> 7#include <linux/string.h>
8#include <linux/slab.h>
9#include <linux/spinlock.h> 8#include <linux/spinlock.h>
10#include <linux/acpi.h> 9#include <linux/acpi.h>
11#include <asm/io.h> 10#include <asm/io.h>
diff --git a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
index 1b1920fa7c80..459168083b77 100644
--- a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
+++ b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
@@ -33,6 +33,7 @@
33#include <linux/cpufreq.h> 33#include <linux/cpufreq.h>
34#include <linux/compiler.h> 34#include <linux/compiler.h>
35#include <linux/dmi.h> 35#include <linux/dmi.h>
36#include <linux/slab.h>
36#include <trace/events/power.h> 37#include <trace/events/power.h>
37 38
38#include <linux/acpi.h> 39#include <linux/acpi.h>
diff --git a/arch/x86/kernel/cpu/cpufreq/elanfreq.c b/arch/x86/kernel/cpu/cpufreq/elanfreq.c
index 006b278b0d5d..c587db472a75 100644
--- a/arch/x86/kernel/cpu/cpufreq/elanfreq.c
+++ b/arch/x86/kernel/cpu/cpufreq/elanfreq.c
@@ -20,7 +20,6 @@
20#include <linux/module.h> 20#include <linux/module.h>
21#include <linux/init.h> 21#include <linux/init.h>
22 22
23#include <linux/slab.h>
24#include <linux/delay.h> 23#include <linux/delay.h>
25#include <linux/cpufreq.h> 24#include <linux/cpufreq.h>
26 25
diff --git a/arch/x86/kernel/cpu/cpufreq/gx-suspmod.c b/arch/x86/kernel/cpu/cpufreq/gx-suspmod.c
index ac27ec2264d5..16e3483be9e3 100644
--- a/arch/x86/kernel/cpu/cpufreq/gx-suspmod.c
+++ b/arch/x86/kernel/cpu/cpufreq/gx-suspmod.c
@@ -80,6 +80,7 @@
80#include <linux/cpufreq.h> 80#include <linux/cpufreq.h>
81#include <linux/pci.h> 81#include <linux/pci.h>
82#include <linux/errno.h> 82#include <linux/errno.h>
83#include <linux/slab.h>
83 84
84#include <asm/processor-cyrix.h> 85#include <asm/processor-cyrix.h>
85 86
diff --git a/arch/x86/kernel/cpu/cpufreq/longrun.c b/arch/x86/kernel/cpu/cpufreq/longrun.c
index da5f70fcb766..e7b559d74c52 100644
--- a/arch/x86/kernel/cpu/cpufreq/longrun.c
+++ b/arch/x86/kernel/cpu/cpufreq/longrun.c
@@ -9,7 +9,6 @@
9#include <linux/kernel.h> 9#include <linux/kernel.h>
10#include <linux/module.h> 10#include <linux/module.h>
11#include <linux/init.h> 11#include <linux/init.h>
12#include <linux/slab.h>
13#include <linux/cpufreq.h> 12#include <linux/cpufreq.h>
14#include <linux/timex.h> 13#include <linux/timex.h>
15 14
diff --git a/arch/x86/kernel/cpu/cpufreq/p4-clockmod.c b/arch/x86/kernel/cpu/cpufreq/p4-clockmod.c
index 869615193720..7b8a8ba67b07 100644
--- a/arch/x86/kernel/cpu/cpufreq/p4-clockmod.c
+++ b/arch/x86/kernel/cpu/cpufreq/p4-clockmod.c
@@ -25,7 +25,6 @@
25#include <linux/init.h> 25#include <linux/init.h>
26#include <linux/smp.h> 26#include <linux/smp.h>
27#include <linux/cpufreq.h> 27#include <linux/cpufreq.h>
28#include <linux/slab.h>
29#include <linux/cpumask.h> 28#include <linux/cpumask.h>
30#include <linux/timex.h> 29#include <linux/timex.h>
31 30
diff --git a/arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c b/arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c
index ff36d2979a90..ce7cde713e71 100644
--- a/arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c
+++ b/arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c
@@ -30,6 +30,7 @@
30#include <linux/sched.h> 30#include <linux/sched.h>
31#include <linux/cpufreq.h> 31#include <linux/cpufreq.h>
32#include <linux/compiler.h> 32#include <linux/compiler.h>
33#include <linux/slab.h>
33 34
34#include <linux/acpi.h> 35#include <linux/acpi.h>
35#include <linux/io.h> 36#include <linux/io.h>
diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k6.c b/arch/x86/kernel/cpu/cpufreq/powernow-k6.c
index cb01dac267d3..b3379d6a5c57 100644
--- a/arch/x86/kernel/cpu/cpufreq/powernow-k6.c
+++ b/arch/x86/kernel/cpu/cpufreq/powernow-k6.c
@@ -13,7 +13,6 @@
13#include <linux/init.h> 13#include <linux/init.h>
14#include <linux/cpufreq.h> 14#include <linux/cpufreq.h>
15#include <linux/ioport.h> 15#include <linux/ioport.h>
16#include <linux/slab.h>
17#include <linux/timex.h> 16#include <linux/timex.h>
18#include <linux/io.h> 17#include <linux/io.h>
19 18
diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
index d360b56e9825..b6215b9798e2 100644
--- a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
+++ b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
@@ -929,7 +929,8 @@ static int fill_powernow_table_pstate(struct powernow_k8_data *data,
929 powernow_table[i].index = index; 929 powernow_table[i].index = index;
930 930
931 /* Frequency may be rounded for these */ 931 /* Frequency may be rounded for these */
932 if (boot_cpu_data.x86 == 0x10 || boot_cpu_data.x86 == 0x11) { 932 if ((boot_cpu_data.x86 == 0x10 && boot_cpu_data.x86_model < 10)
933 || boot_cpu_data.x86 == 0x11) {
933 powernow_table[i].frequency = 934 powernow_table[i].frequency =
934 freq_from_fid_did(lo & 0x3f, (lo >> 6) & 7); 935 freq_from_fid_did(lo & 0x3f, (lo >> 6) & 7);
935 } else 936 } else
diff --git a/arch/x86/kernel/cpu/cpufreq/speedstep-centrino.c b/arch/x86/kernel/cpu/cpufreq/speedstep-centrino.c
index 8d672ef162ce..9b1ff37de46a 100644
--- a/arch/x86/kernel/cpu/cpufreq/speedstep-centrino.c
+++ b/arch/x86/kernel/cpu/cpufreq/speedstep-centrino.c
@@ -20,6 +20,7 @@
20#include <linux/sched.h> /* current */ 20#include <linux/sched.h> /* current */
21#include <linux/delay.h> 21#include <linux/delay.h>
22#include <linux/compiler.h> 22#include <linux/compiler.h>
23#include <linux/gfp.h>
23 24
24#include <asm/msr.h> 25#include <asm/msr.h>
25#include <asm/processor.h> 26#include <asm/processor.h>
diff --git a/arch/x86/kernel/cpu/cpufreq/speedstep-ich.c b/arch/x86/kernel/cpu/cpufreq/speedstep-ich.c
index 2ce8e0b5cc54..561758e95180 100644
--- a/arch/x86/kernel/cpu/cpufreq/speedstep-ich.c
+++ b/arch/x86/kernel/cpu/cpufreq/speedstep-ich.c
@@ -23,7 +23,6 @@
23#include <linux/init.h> 23#include <linux/init.h>
24#include <linux/cpufreq.h> 24#include <linux/cpufreq.h>
25#include <linux/pci.h> 25#include <linux/pci.h>
26#include <linux/slab.h>
27#include <linux/sched.h> 26#include <linux/sched.h>
28 27
29#include "speedstep-lib.h" 28#include "speedstep-lib.h"
diff --git a/arch/x86/kernel/cpu/cpufreq/speedstep-lib.c b/arch/x86/kernel/cpu/cpufreq/speedstep-lib.c
index ad0083abfa23..a94ec6be69fa 100644
--- a/arch/x86/kernel/cpu/cpufreq/speedstep-lib.c
+++ b/arch/x86/kernel/cpu/cpufreq/speedstep-lib.c
@@ -13,7 +13,6 @@
13#include <linux/moduleparam.h> 13#include <linux/moduleparam.h>
14#include <linux/init.h> 14#include <linux/init.h>
15#include <linux/cpufreq.h> 15#include <linux/cpufreq.h>
16#include <linux/slab.h>
17 16
18#include <asm/msr.h> 17#include <asm/msr.h>
19#include <asm/tsc.h> 18#include <asm/tsc.h>
diff --git a/arch/x86/kernel/cpu/cpufreq/speedstep-smi.c b/arch/x86/kernel/cpu/cpufreq/speedstep-smi.c
index 04d73c114e49..8abd869baabf 100644
--- a/arch/x86/kernel/cpu/cpufreq/speedstep-smi.c
+++ b/arch/x86/kernel/cpu/cpufreq/speedstep-smi.c
@@ -17,7 +17,6 @@
17#include <linux/moduleparam.h> 17#include <linux/moduleparam.h>
18#include <linux/init.h> 18#include <linux/init.h>
19#include <linux/cpufreq.h> 19#include <linux/cpufreq.h>
20#include <linux/slab.h>
21#include <linux/delay.h> 20#include <linux/delay.h>
22#include <linux/io.h> 21#include <linux/io.h>
23#include <asm/ist.h> 22#include <asm/ist.h>
diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c
index 7e1cca13af35..1366c7cfd483 100644
--- a/arch/x86/kernel/cpu/intel.c
+++ b/arch/x86/kernel/cpu/intel.c
@@ -47,6 +47,27 @@ static void __cpuinit early_init_intel(struct cpuinfo_x86 *c)
47 (c->x86 == 0x6 && c->x86_model >= 0x0e)) 47 (c->x86 == 0x6 && c->x86_model >= 0x0e))
48 set_cpu_cap(c, X86_FEATURE_CONSTANT_TSC); 48 set_cpu_cap(c, X86_FEATURE_CONSTANT_TSC);
49 49
50 /*
51 * Atom erratum AAE44/AAF40/AAG38/AAH41:
52 *
53 * A race condition between speculative fetches and invalidating
54 * a large page. This is worked around in microcode, but we
55 * need the microcode to have already been loaded... so if it is
56 * not, recommend a BIOS update and disable large pages.
57 */
58 if (c->x86 == 6 && c->x86_model == 0x1c && c->x86_mask <= 2) {
59 u32 ucode, junk;
60
61 wrmsr(MSR_IA32_UCODE_REV, 0, 0);
62 sync_core();
63 rdmsr(MSR_IA32_UCODE_REV, junk, ucode);
64
65 if (ucode < 0x20e) {
66 printk(KERN_WARNING "Atom PSE erratum detected, BIOS microcode update recommended\n");
67 clear_cpu_cap(c, X86_FEATURE_PSE);
68 }
69 }
70
50#ifdef CONFIG_X86_64 71#ifdef CONFIG_X86_64
51 set_cpu_cap(c, X86_FEATURE_SYSENTER32); 72 set_cpu_cap(c, X86_FEATURE_SYSENTER32);
52#else 73#else
diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c
index b3eeb66c0a51..95962a93f99a 100644
--- a/arch/x86/kernel/cpu/intel_cacheinfo.c
+++ b/arch/x86/kernel/cpu/intel_cacheinfo.c
@@ -340,6 +340,10 @@ amd_check_l3_disable(int index, struct _cpuid4_info_regs *this_leaf)
340 (boot_cpu_data.x86_mask < 0x1))) 340 (boot_cpu_data.x86_mask < 0x1)))
341 return; 341 return;
342 342
343 /* not in virtualized environments */
344 if (num_k8_northbridges == 0)
345 return;
346
343 this_leaf->can_disable = true; 347 this_leaf->can_disable = true;
344 this_leaf->l3_indices = amd_calc_l3_indices(); 348 this_leaf->l3_indices = amd_calc_l3_indices();
345} 349}
diff --git a/arch/x86/kernel/cpu/mcheck/Makefile b/arch/x86/kernel/cpu/mcheck/Makefile
index 4ac6d48fe11b..bb34b03af252 100644
--- a/arch/x86/kernel/cpu/mcheck/Makefile
+++ b/arch/x86/kernel/cpu/mcheck/Makefile
@@ -7,3 +7,5 @@ obj-$(CONFIG_X86_MCE_THRESHOLD) += threshold.o
7obj-$(CONFIG_X86_MCE_INJECT) += mce-inject.o 7obj-$(CONFIG_X86_MCE_INJECT) += mce-inject.o
8 8
9obj-$(CONFIG_X86_THERMAL_VECTOR) += therm_throt.o 9obj-$(CONFIG_X86_THERMAL_VECTOR) += therm_throt.o
10
11obj-$(CONFIG_ACPI_APEI) += mce-apei.o
diff --git a/arch/x86/kernel/cpu/mcheck/mce-apei.c b/arch/x86/kernel/cpu/mcheck/mce-apei.c
new file mode 100644
index 000000000000..745b54f9be89
--- /dev/null
+++ b/arch/x86/kernel/cpu/mcheck/mce-apei.c
@@ -0,0 +1,138 @@
1/*
2 * Bridge between MCE and APEI
3 *
4 * On some machine, corrected memory errors are reported via APEI
5 * generic hardware error source (GHES) instead of corrected Machine
6 * Check. These corrected memory errors can be reported to user space
7 * through /dev/mcelog via faking a corrected Machine Check, so that
8 * the error memory page can be offlined by /sbin/mcelog if the error
9 * count for one page is beyond the threshold.
10 *
11 * For fatal MCE, save MCE record into persistent storage via ERST, so
12 * that the MCE record can be logged after reboot via ERST.
13 *
14 * Copyright 2010 Intel Corp.
15 * Author: Huang Ying <ying.huang@intel.com>
16 *
17 * This program is free software; you can redistribute it and/or
18 * modify it under the terms of the GNU General Public License version
19 * 2 as published by the Free Software Foundation.
20 *
21 * This program is distributed in the hope that it will be useful,
22 * but WITHOUT ANY WARRANTY; without even the implied warranty of
23 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
24 * GNU General Public License for more details.
25 *
26 * You should have received a copy of the GNU General Public License
27 * along with this program; if not, write to the Free Software
28 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
29 */
30
31#include <linux/kernel.h>
32#include <linux/acpi.h>
33#include <linux/cper.h>
34#include <acpi/apei.h>
35#include <asm/mce.h>
36
37#include "mce-internal.h"
38
39void apei_mce_report_mem_error(int corrected, struct cper_sec_mem_err *mem_err)
40{
41 struct mce m;
42
43 /* Only corrected MC is reported */
44 if (!corrected)
45 return;
46
47 mce_setup(&m);
48 m.bank = 1;
49 /* Fake a memory read corrected error with unknown channel */
50 m.status = MCI_STATUS_VAL | MCI_STATUS_EN | MCI_STATUS_ADDRV | 0x9f;
51 m.addr = mem_err->physical_addr;
52 mce_log(&m);
53 mce_notify_irq();
54}
55EXPORT_SYMBOL_GPL(apei_mce_report_mem_error);
56
57#define CPER_CREATOR_MCE \
58 UUID_LE(0x75a574e3, 0x5052, 0x4b29, 0x8a, 0x8e, 0xbe, 0x2c, \
59 0x64, 0x90, 0xb8, 0x9d)
60#define CPER_SECTION_TYPE_MCE \
61 UUID_LE(0xfe08ffbe, 0x95e4, 0x4be7, 0xbc, 0x73, 0x40, 0x96, \
62 0x04, 0x4a, 0x38, 0xfc)
63
64/*
65 * CPER specification (in UEFI specification 2.3 appendix N) requires
66 * byte-packed.
67 */
68struct cper_mce_record {
69 struct cper_record_header hdr;
70 struct cper_section_descriptor sec_hdr;
71 struct mce mce;
72} __packed;
73
74int apei_write_mce(struct mce *m)
75{
76 struct cper_mce_record rcd;
77
78 memset(&rcd, 0, sizeof(rcd));
79 memcpy(rcd.hdr.signature, CPER_SIG_RECORD, CPER_SIG_SIZE);
80 rcd.hdr.revision = CPER_RECORD_REV;
81 rcd.hdr.signature_end = CPER_SIG_END;
82 rcd.hdr.section_count = 1;
83 rcd.hdr.error_severity = CPER_SER_FATAL;
84 /* timestamp, platform_id, partition_id are all invalid */
85 rcd.hdr.validation_bits = 0;
86 rcd.hdr.record_length = sizeof(rcd);
87 rcd.hdr.creator_id = CPER_CREATOR_MCE;
88 rcd.hdr.notification_type = CPER_NOTIFY_MCE;
89 rcd.hdr.record_id = cper_next_record_id();
90 rcd.hdr.flags = CPER_HW_ERROR_FLAGS_PREVERR;
91
92 rcd.sec_hdr.section_offset = (void *)&rcd.mce - (void *)&rcd;
93 rcd.sec_hdr.section_length = sizeof(rcd.mce);
94 rcd.sec_hdr.revision = CPER_SEC_REV;
95 /* fru_id and fru_text is invalid */
96 rcd.sec_hdr.validation_bits = 0;
97 rcd.sec_hdr.flags = CPER_SEC_PRIMARY;
98 rcd.sec_hdr.section_type = CPER_SECTION_TYPE_MCE;
99 rcd.sec_hdr.section_severity = CPER_SER_FATAL;
100
101 memcpy(&rcd.mce, m, sizeof(*m));
102
103 return erst_write(&rcd.hdr);
104}
105
106ssize_t apei_read_mce(struct mce *m, u64 *record_id)
107{
108 struct cper_mce_record rcd;
109 ssize_t len;
110
111 len = erst_read_next(&rcd.hdr, sizeof(rcd));
112 if (len <= 0)
113 return len;
114 /* Can not skip other records in storage via ERST unless clear them */
115 else if (len != sizeof(rcd) ||
116 uuid_le_cmp(rcd.hdr.creator_id, CPER_CREATOR_MCE)) {
117 if (printk_ratelimit())
118 pr_warning(
119 "MCE-APEI: Can not skip the unknown record in ERST");
120 return -EIO;
121 }
122
123 memcpy(m, &rcd.mce, sizeof(*m));
124 *record_id = rcd.hdr.record_id;
125
126 return sizeof(*m);
127}
128
129/* Check whether there is record in ERST */
130int apei_check_mce(void)
131{
132 return erst_get_record_count();
133}
134
135int apei_clear_mce(u64 record_id)
136{
137 return erst_clear(record_id);
138}
diff --git a/arch/x86/kernel/cpu/mcheck/mce-inject.c b/arch/x86/kernel/cpu/mcheck/mce-inject.c
index 73734baa50f2..e7dbde7bfedb 100644
--- a/arch/x86/kernel/cpu/mcheck/mce-inject.c
+++ b/arch/x86/kernel/cpu/mcheck/mce-inject.c
@@ -22,6 +22,7 @@
22#include <linux/kdebug.h> 22#include <linux/kdebug.h>
23#include <linux/cpu.h> 23#include <linux/cpu.h>
24#include <linux/sched.h> 24#include <linux/sched.h>
25#include <linux/gfp.h>
25#include <asm/mce.h> 26#include <asm/mce.h>
26#include <asm/apic.h> 27#include <asm/apic.h>
27 28
diff --git a/arch/x86/kernel/cpu/mcheck/mce-internal.h b/arch/x86/kernel/cpu/mcheck/mce-internal.h
index 32996f9fab67..fefcc69ee8b5 100644
--- a/arch/x86/kernel/cpu/mcheck/mce-internal.h
+++ b/arch/x86/kernel/cpu/mcheck/mce-internal.h
@@ -28,3 +28,26 @@ extern int mce_ser;
28 28
29extern struct mce_bank *mce_banks; 29extern struct mce_bank *mce_banks;
30 30
31#ifdef CONFIG_ACPI_APEI
32int apei_write_mce(struct mce *m);
33ssize_t apei_read_mce(struct mce *m, u64 *record_id);
34int apei_check_mce(void);
35int apei_clear_mce(u64 record_id);
36#else
37static inline int apei_write_mce(struct mce *m)
38{
39 return -EINVAL;
40}
41static inline ssize_t apei_read_mce(struct mce *m, u64 *record_id)
42{
43 return 0;
44}
45static inline int apei_check_mce(void)
46{
47 return 0;
48}
49static inline int apei_clear_mce(u64 record_id)
50{
51 return -EINVAL;
52}
53#endif
diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c
index 3ab9c886b613..09535ca9b9d7 100644
--- a/arch/x86/kernel/cpu/mcheck/mce.c
+++ b/arch/x86/kernel/cpu/mcheck/mce.c
@@ -26,6 +26,7 @@
26#include <linux/sched.h> 26#include <linux/sched.h>
27#include <linux/sysfs.h> 27#include <linux/sysfs.h>
28#include <linux/types.h> 28#include <linux/types.h>
29#include <linux/slab.h>
29#include <linux/init.h> 30#include <linux/init.h>
30#include <linux/kmod.h> 31#include <linux/kmod.h>
31#include <linux/poll.h> 32#include <linux/poll.h>
@@ -263,7 +264,7 @@ static void wait_for_panic(void)
263 264
264static void mce_panic(char *msg, struct mce *final, char *exp) 265static void mce_panic(char *msg, struct mce *final, char *exp)
265{ 266{
266 int i; 267 int i, apei_err = 0;
267 268
268 if (!fake_panic) { 269 if (!fake_panic) {
269 /* 270 /*
@@ -286,8 +287,11 @@ static void mce_panic(char *msg, struct mce *final, char *exp)
286 struct mce *m = &mcelog.entry[i]; 287 struct mce *m = &mcelog.entry[i];
287 if (!(m->status & MCI_STATUS_VAL)) 288 if (!(m->status & MCI_STATUS_VAL))
288 continue; 289 continue;
289 if (!(m->status & MCI_STATUS_UC)) 290 if (!(m->status & MCI_STATUS_UC)) {
290 print_mce(m); 291 print_mce(m);
292 if (!apei_err)
293 apei_err = apei_write_mce(m);
294 }
291 } 295 }
292 /* Now print uncorrected but with the final one last */ 296 /* Now print uncorrected but with the final one last */
293 for (i = 0; i < MCE_LOG_LEN; i++) { 297 for (i = 0; i < MCE_LOG_LEN; i++) {
@@ -296,11 +300,17 @@ static void mce_panic(char *msg, struct mce *final, char *exp)
296 continue; 300 continue;
297 if (!(m->status & MCI_STATUS_UC)) 301 if (!(m->status & MCI_STATUS_UC))
298 continue; 302 continue;
299 if (!final || memcmp(m, final, sizeof(struct mce))) 303 if (!final || memcmp(m, final, sizeof(struct mce))) {
300 print_mce(m); 304 print_mce(m);
305 if (!apei_err)
306 apei_err = apei_write_mce(m);
307 }
301 } 308 }
302 if (final) 309 if (final) {
303 print_mce(final); 310 print_mce(final);
311 if (!apei_err)
312 apei_err = apei_write_mce(final);
313 }
304 if (cpu_missing) 314 if (cpu_missing)
305 printk(KERN_EMERG "Some CPUs didn't answer in synchronization\n"); 315 printk(KERN_EMERG "Some CPUs didn't answer in synchronization\n");
306 print_mce_tail(); 316 print_mce_tail();
@@ -1492,6 +1502,43 @@ static void collect_tscs(void *data)
1492 rdtscll(cpu_tsc[smp_processor_id()]); 1502 rdtscll(cpu_tsc[smp_processor_id()]);
1493} 1503}
1494 1504
1505static int mce_apei_read_done;
1506
1507/* Collect MCE record of previous boot in persistent storage via APEI ERST. */
1508static int __mce_read_apei(char __user **ubuf, size_t usize)
1509{
1510 int rc;
1511 u64 record_id;
1512 struct mce m;
1513
1514 if (usize < sizeof(struct mce))
1515 return -EINVAL;
1516
1517 rc = apei_read_mce(&m, &record_id);
1518 /* Error or no more MCE record */
1519 if (rc <= 0) {
1520 mce_apei_read_done = 1;
1521 return rc;
1522 }
1523 rc = -EFAULT;
1524 if (copy_to_user(*ubuf, &m, sizeof(struct mce)))
1525 return rc;
1526 /*
1527 * In fact, we should have cleared the record after that has
1528 * been flushed to the disk or sent to network in
1529 * /sbin/mcelog, but we have no interface to support that now,
1530 * so just clear it to avoid duplication.
1531 */
1532 rc = apei_clear_mce(record_id);
1533 if (rc) {
1534 mce_apei_read_done = 1;
1535 return rc;
1536 }
1537 *ubuf += sizeof(struct mce);
1538
1539 return 0;
1540}
1541
1495static ssize_t mce_read(struct file *filp, char __user *ubuf, size_t usize, 1542static ssize_t mce_read(struct file *filp, char __user *ubuf, size_t usize,
1496 loff_t *off) 1543 loff_t *off)
1497{ 1544{
@@ -1505,15 +1552,19 @@ static ssize_t mce_read(struct file *filp, char __user *ubuf, size_t usize,
1505 return -ENOMEM; 1552 return -ENOMEM;
1506 1553
1507 mutex_lock(&mce_read_mutex); 1554 mutex_lock(&mce_read_mutex);
1555
1556 if (!mce_apei_read_done) {
1557 err = __mce_read_apei(&buf, usize);
1558 if (err || buf != ubuf)
1559 goto out;
1560 }
1561
1508 next = rcu_dereference_check_mce(mcelog.next); 1562 next = rcu_dereference_check_mce(mcelog.next);
1509 1563
1510 /* Only supports full reads right now */ 1564 /* Only supports full reads right now */
1511 if (*off != 0 || usize < MCE_LOG_LEN*sizeof(struct mce)) { 1565 err = -EINVAL;
1512 mutex_unlock(&mce_read_mutex); 1566 if (*off != 0 || usize < MCE_LOG_LEN*sizeof(struct mce))
1513 kfree(cpu_tsc); 1567 goto out;
1514
1515 return -EINVAL;
1516 }
1517 1568
1518 err = 0; 1569 err = 0;
1519 prev = 0; 1570 prev = 0;
@@ -1561,10 +1612,15 @@ timeout:
1561 memset(&mcelog.entry[i], 0, sizeof(struct mce)); 1612 memset(&mcelog.entry[i], 0, sizeof(struct mce));
1562 } 1613 }
1563 } 1614 }
1615
1616 if (err)
1617 err = -EFAULT;
1618
1619out:
1564 mutex_unlock(&mce_read_mutex); 1620 mutex_unlock(&mce_read_mutex);
1565 kfree(cpu_tsc); 1621 kfree(cpu_tsc);
1566 1622
1567 return err ? -EFAULT : buf - ubuf; 1623 return err ? err : buf - ubuf;
1568} 1624}
1569 1625
1570static unsigned int mce_poll(struct file *file, poll_table *wait) 1626static unsigned int mce_poll(struct file *file, poll_table *wait)
@@ -1572,6 +1628,8 @@ static unsigned int mce_poll(struct file *file, poll_table *wait)
1572 poll_wait(file, &mce_wait, wait); 1628 poll_wait(file, &mce_wait, wait);
1573 if (rcu_dereference_check_mce(mcelog.next)) 1629 if (rcu_dereference_check_mce(mcelog.next))
1574 return POLLIN | POLLRDNORM; 1630 return POLLIN | POLLRDNORM;
1631 if (!mce_apei_read_done && apei_check_mce())
1632 return POLLIN | POLLRDNORM;
1575 return 0; 1633 return 0;
1576} 1634}
1577 1635
diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd.c b/arch/x86/kernel/cpu/mcheck/mce_amd.c
index cda932ca3ade..224392d8fe8c 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_amd.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_amd.c
@@ -21,6 +21,7 @@
21#include <linux/errno.h> 21#include <linux/errno.h>
22#include <linux/sched.h> 22#include <linux/sched.h>
23#include <linux/sysfs.h> 23#include <linux/sysfs.h>
24#include <linux/slab.h>
24#include <linux/init.h> 25#include <linux/init.h>
25#include <linux/cpu.h> 26#include <linux/cpu.h>
26#include <linux/smp.h> 27#include <linux/smp.h>
diff --git a/arch/x86/kernel/cpu/mcheck/mce_intel.c b/arch/x86/kernel/cpu/mcheck/mce_intel.c
index d15df6e49bf0..62b48e40920a 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_intel.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_intel.c
@@ -5,6 +5,7 @@
5 * Author: Andi Kleen 5 * Author: Andi Kleen
6 */ 6 */
7 7
8#include <linux/gfp.h>
8#include <linux/init.h> 9#include <linux/init.h>
9#include <linux/interrupt.h> 10#include <linux/interrupt.h>
10#include <linux/percpu.h> 11#include <linux/percpu.h>
diff --git a/arch/x86/kernel/cpu/mtrr/generic.c b/arch/x86/kernel/cpu/mtrr/generic.c
index 9aa5dc76ff4a..fd31a441c61c 100644
--- a/arch/x86/kernel/cpu/mtrr/generic.c
+++ b/arch/x86/kernel/cpu/mtrr/generic.c
@@ -6,7 +6,6 @@
6 6
7#include <linux/module.h> 7#include <linux/module.h>
8#include <linux/init.h> 8#include <linux/init.h>
9#include <linux/slab.h>
10#include <linux/io.h> 9#include <linux/io.h>
11#include <linux/mm.h> 10#include <linux/mm.h>
12 11
diff --git a/arch/x86/kernel/cpu/mtrr/if.c b/arch/x86/kernel/cpu/mtrr/if.c
index e006e56f699c..79289632cb27 100644
--- a/arch/x86/kernel/cpu/mtrr/if.c
+++ b/arch/x86/kernel/cpu/mtrr/if.c
@@ -5,6 +5,7 @@
5#include <linux/module.h> 5#include <linux/module.h>
6#include <linux/ctype.h> 6#include <linux/ctype.h>
7#include <linux/string.h> 7#include <linux/string.h>
8#include <linux/slab.h>
8#include <linux/init.h> 9#include <linux/init.h>
9 10
10#define LINE_SIZE 80 11#define LINE_SIZE 80
diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
index 60398a0d947c..db5bdc8addf8 100644
--- a/arch/x86/kernel/cpu/perf_event.c
+++ b/arch/x86/kernel/cpu/perf_event.c
@@ -21,6 +21,7 @@
21#include <linux/kdebug.h> 21#include <linux/kdebug.h>
22#include <linux/sched.h> 22#include <linux/sched.h>
23#include <linux/uaccess.h> 23#include <linux/uaccess.h>
24#include <linux/slab.h>
24#include <linux/highmem.h> 25#include <linux/highmem.h>
25#include <linux/cpu.h> 26#include <linux/cpu.h>
26#include <linux/bitops.h> 27#include <linux/bitops.h>
@@ -28,6 +29,7 @@
28#include <asm/apic.h> 29#include <asm/apic.h>
29#include <asm/stacktrace.h> 30#include <asm/stacktrace.h>
30#include <asm/nmi.h> 31#include <asm/nmi.h>
32#include <asm/compat.h>
31 33
32static u64 perf_event_mask __read_mostly; 34static u64 perf_event_mask __read_mostly;
33 35
@@ -158,7 +160,7 @@ struct x86_pmu {
158 struct perf_event *event); 160 struct perf_event *event);
159 struct event_constraint *event_constraints; 161 struct event_constraint *event_constraints;
160 162
161 void (*cpu_prepare)(int cpu); 163 int (*cpu_prepare)(int cpu);
162 void (*cpu_starting)(int cpu); 164 void (*cpu_starting)(int cpu);
163 void (*cpu_dying)(int cpu); 165 void (*cpu_dying)(int cpu);
164 void (*cpu_dead)(int cpu); 166 void (*cpu_dead)(int cpu);
@@ -1333,11 +1335,12 @@ static int __cpuinit
1333x86_pmu_notifier(struct notifier_block *self, unsigned long action, void *hcpu) 1335x86_pmu_notifier(struct notifier_block *self, unsigned long action, void *hcpu)
1334{ 1336{
1335 unsigned int cpu = (long)hcpu; 1337 unsigned int cpu = (long)hcpu;
1338 int ret = NOTIFY_OK;
1336 1339
1337 switch (action & ~CPU_TASKS_FROZEN) { 1340 switch (action & ~CPU_TASKS_FROZEN) {
1338 case CPU_UP_PREPARE: 1341 case CPU_UP_PREPARE:
1339 if (x86_pmu.cpu_prepare) 1342 if (x86_pmu.cpu_prepare)
1340 x86_pmu.cpu_prepare(cpu); 1343 ret = x86_pmu.cpu_prepare(cpu);
1341 break; 1344 break;
1342 1345
1343 case CPU_STARTING: 1346 case CPU_STARTING:
@@ -1350,6 +1353,7 @@ x86_pmu_notifier(struct notifier_block *self, unsigned long action, void *hcpu)
1350 x86_pmu.cpu_dying(cpu); 1353 x86_pmu.cpu_dying(cpu);
1351 break; 1354 break;
1352 1355
1356 case CPU_UP_CANCELED:
1353 case CPU_DEAD: 1357 case CPU_DEAD:
1354 if (x86_pmu.cpu_dead) 1358 if (x86_pmu.cpu_dead)
1355 x86_pmu.cpu_dead(cpu); 1359 x86_pmu.cpu_dead(cpu);
@@ -1359,7 +1363,7 @@ x86_pmu_notifier(struct notifier_block *self, unsigned long action, void *hcpu)
1359 break; 1363 break;
1360 } 1364 }
1361 1365
1362 return NOTIFY_OK; 1366 return ret;
1363} 1367}
1364 1368
1365static void __init pmu_check_apic(void) 1369static void __init pmu_check_apic(void)
@@ -1628,14 +1632,42 @@ copy_from_user_nmi(void *to, const void __user *from, unsigned long n)
1628 return len; 1632 return len;
1629} 1633}
1630 1634
1631static int copy_stack_frame(const void __user *fp, struct stack_frame *frame) 1635#ifdef CONFIG_COMPAT
1636static inline int
1637perf_callchain_user32(struct pt_regs *regs, struct perf_callchain_entry *entry)
1632{ 1638{
1633 unsigned long bytes; 1639 /* 32-bit process in 64-bit kernel. */
1640 struct stack_frame_ia32 frame;
1641 const void __user *fp;
1634 1642
1635 bytes = copy_from_user_nmi(frame, fp, sizeof(*frame)); 1643 if (!test_thread_flag(TIF_IA32))
1644 return 0;
1645
1646 fp = compat_ptr(regs->bp);
1647 while (entry->nr < PERF_MAX_STACK_DEPTH) {
1648 unsigned long bytes;
1649 frame.next_frame = 0;
1650 frame.return_address = 0;
1651
1652 bytes = copy_from_user_nmi(&frame, fp, sizeof(frame));
1653 if (bytes != sizeof(frame))
1654 break;
1655
1656 if (fp < compat_ptr(regs->sp))
1657 break;
1636 1658
1637 return bytes == sizeof(*frame); 1659 callchain_store(entry, frame.return_address);
1660 fp = compat_ptr(frame.next_frame);
1661 }
1662 return 1;
1663}
1664#else
1665static inline int
1666perf_callchain_user32(struct pt_regs *regs, struct perf_callchain_entry *entry)
1667{
1668 return 0;
1638} 1669}
1670#endif
1639 1671
1640static void 1672static void
1641perf_callchain_user(struct pt_regs *regs, struct perf_callchain_entry *entry) 1673perf_callchain_user(struct pt_regs *regs, struct perf_callchain_entry *entry)
@@ -1651,11 +1683,16 @@ perf_callchain_user(struct pt_regs *regs, struct perf_callchain_entry *entry)
1651 callchain_store(entry, PERF_CONTEXT_USER); 1683 callchain_store(entry, PERF_CONTEXT_USER);
1652 callchain_store(entry, regs->ip); 1684 callchain_store(entry, regs->ip);
1653 1685
1686 if (perf_callchain_user32(regs, entry))
1687 return;
1688
1654 while (entry->nr < PERF_MAX_STACK_DEPTH) { 1689 while (entry->nr < PERF_MAX_STACK_DEPTH) {
1690 unsigned long bytes;
1655 frame.next_frame = NULL; 1691 frame.next_frame = NULL;
1656 frame.return_address = 0; 1692 frame.return_address = 0;
1657 1693
1658 if (!copy_stack_frame(fp, &frame)) 1694 bytes = copy_from_user_nmi(&frame, fp, sizeof(frame));
1695 if (bytes != sizeof(frame))
1659 break; 1696 break;
1660 1697
1661 if ((unsigned long)fp < regs->sp) 1698 if ((unsigned long)fp < regs->sp)
@@ -1702,7 +1739,6 @@ struct perf_callchain_entry *perf_callchain(struct pt_regs *regs)
1702 return entry; 1739 return entry;
1703} 1740}
1704 1741
1705#ifdef CONFIG_EVENT_TRACING
1706void perf_arch_fetch_caller_regs(struct pt_regs *regs, unsigned long ip, int skip) 1742void perf_arch_fetch_caller_regs(struct pt_regs *regs, unsigned long ip, int skip)
1707{ 1743{
1708 regs->ip = ip; 1744 regs->ip = ip;
@@ -1714,4 +1750,3 @@ void perf_arch_fetch_caller_regs(struct pt_regs *regs, unsigned long ip, int ski
1714 regs->cs = __KERNEL_CS; 1750 regs->cs = __KERNEL_CS;
1715 local_save_flags(regs->flags); 1751 local_save_flags(regs->flags);
1716} 1752}
1717#endif
diff --git a/arch/x86/kernel/cpu/perf_event_amd.c b/arch/x86/kernel/cpu/perf_event_amd.c
index b87e0b6970cb..db6f7d4056e1 100644
--- a/arch/x86/kernel/cpu/perf_event_amd.c
+++ b/arch/x86/kernel/cpu/perf_event_amd.c
@@ -137,6 +137,13 @@ static inline int amd_is_nb_event(struct hw_perf_event *hwc)
137 return (hwc->config & 0xe0) == 0xe0; 137 return (hwc->config & 0xe0) == 0xe0;
138} 138}
139 139
140static inline int amd_has_nb(struct cpu_hw_events *cpuc)
141{
142 struct amd_nb *nb = cpuc->amd_nb;
143
144 return nb && nb->nb_id != -1;
145}
146
140static void amd_put_event_constraints(struct cpu_hw_events *cpuc, 147static void amd_put_event_constraints(struct cpu_hw_events *cpuc,
141 struct perf_event *event) 148 struct perf_event *event)
142{ 149{
@@ -147,7 +154,7 @@ static void amd_put_event_constraints(struct cpu_hw_events *cpuc,
147 /* 154 /*
148 * only care about NB events 155 * only care about NB events
149 */ 156 */
150 if (!(nb && amd_is_nb_event(hwc))) 157 if (!(amd_has_nb(cpuc) && amd_is_nb_event(hwc)))
151 return; 158 return;
152 159
153 /* 160 /*
@@ -214,7 +221,7 @@ amd_get_event_constraints(struct cpu_hw_events *cpuc, struct perf_event *event)
214 /* 221 /*
215 * if not NB event or no NB, then no constraints 222 * if not NB event or no NB, then no constraints
216 */ 223 */
217 if (!(nb && amd_is_nb_event(hwc))) 224 if (!(amd_has_nb(cpuc) && amd_is_nb_event(hwc)))
218 return &unconstrained; 225 return &unconstrained;
219 226
220 /* 227 /*
@@ -293,51 +300,55 @@ static struct amd_nb *amd_alloc_nb(int cpu, int nb_id)
293 return nb; 300 return nb;
294} 301}
295 302
296static void amd_pmu_cpu_online(int cpu) 303static int amd_pmu_cpu_prepare(int cpu)
304{
305 struct cpu_hw_events *cpuc = &per_cpu(cpu_hw_events, cpu);
306
307 WARN_ON_ONCE(cpuc->amd_nb);
308
309 if (boot_cpu_data.x86_max_cores < 2)
310 return NOTIFY_OK;
311
312 cpuc->amd_nb = amd_alloc_nb(cpu, -1);
313 if (!cpuc->amd_nb)
314 return NOTIFY_BAD;
315
316 return NOTIFY_OK;
317}
318
319static void amd_pmu_cpu_starting(int cpu)
297{ 320{
298 struct cpu_hw_events *cpu1, *cpu2; 321 struct cpu_hw_events *cpuc = &per_cpu(cpu_hw_events, cpu);
299 struct amd_nb *nb = NULL; 322 struct amd_nb *nb;
300 int i, nb_id; 323 int i, nb_id;
301 324
302 if (boot_cpu_data.x86_max_cores < 2) 325 if (boot_cpu_data.x86_max_cores < 2)
303 return; 326 return;
304 327
305 /*
306 * function may be called too early in the
307 * boot process, in which case nb_id is bogus
308 */
309 nb_id = amd_get_nb_id(cpu); 328 nb_id = amd_get_nb_id(cpu);
310 if (nb_id == BAD_APICID) 329 WARN_ON_ONCE(nb_id == BAD_APICID);
311 return;
312
313 cpu1 = &per_cpu(cpu_hw_events, cpu);
314 cpu1->amd_nb = NULL;
315 330
316 raw_spin_lock(&amd_nb_lock); 331 raw_spin_lock(&amd_nb_lock);
317 332
318 for_each_online_cpu(i) { 333 for_each_online_cpu(i) {
319 cpu2 = &per_cpu(cpu_hw_events, i); 334 nb = per_cpu(cpu_hw_events, i).amd_nb;
320 nb = cpu2->amd_nb; 335 if (WARN_ON_ONCE(!nb))
321 if (!nb)
322 continue; 336 continue;
323 if (nb->nb_id == nb_id)
324 goto found;
325 }
326 337
327 nb = amd_alloc_nb(cpu, nb_id); 338 if (nb->nb_id == nb_id) {
328 if (!nb) { 339 kfree(cpuc->amd_nb);
329 pr_err("perf_events: failed NB allocation for CPU%d\n", cpu); 340 cpuc->amd_nb = nb;
330 raw_spin_unlock(&amd_nb_lock); 341 break;
331 return; 342 }
332 } 343 }
333found: 344
334 nb->refcnt++; 345 cpuc->amd_nb->nb_id = nb_id;
335 cpu1->amd_nb = nb; 346 cpuc->amd_nb->refcnt++;
336 347
337 raw_spin_unlock(&amd_nb_lock); 348 raw_spin_unlock(&amd_nb_lock);
338} 349}
339 350
340static void amd_pmu_cpu_offline(int cpu) 351static void amd_pmu_cpu_dead(int cpu)
341{ 352{
342 struct cpu_hw_events *cpuhw; 353 struct cpu_hw_events *cpuhw;
343 354
@@ -349,8 +360,10 @@ static void amd_pmu_cpu_offline(int cpu)
349 raw_spin_lock(&amd_nb_lock); 360 raw_spin_lock(&amd_nb_lock);
350 361
351 if (cpuhw->amd_nb) { 362 if (cpuhw->amd_nb) {
352 if (--cpuhw->amd_nb->refcnt == 0) 363 struct amd_nb *nb = cpuhw->amd_nb;
353 kfree(cpuhw->amd_nb); 364
365 if (nb->nb_id == -1 || --nb->refcnt == 0)
366 kfree(nb);
354 367
355 cpuhw->amd_nb = NULL; 368 cpuhw->amd_nb = NULL;
356 } 369 }
@@ -379,8 +392,9 @@ static __initconst struct x86_pmu amd_pmu = {
379 .get_event_constraints = amd_get_event_constraints, 392 .get_event_constraints = amd_get_event_constraints,
380 .put_event_constraints = amd_put_event_constraints, 393 .put_event_constraints = amd_put_event_constraints,
381 394
382 .cpu_prepare = amd_pmu_cpu_online, 395 .cpu_prepare = amd_pmu_cpu_prepare,
383 .cpu_dead = amd_pmu_cpu_offline, 396 .cpu_starting = amd_pmu_cpu_starting,
397 .cpu_dead = amd_pmu_cpu_dead,
384}; 398};
385 399
386static __init int amd_pmu_init(void) 400static __init int amd_pmu_init(void)
diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c
index 84bfde64a337..9c794ac87837 100644
--- a/arch/x86/kernel/cpu/perf_event_intel.c
+++ b/arch/x86/kernel/cpu/perf_event_intel.c
@@ -936,6 +936,7 @@ static __init int intel_pmu_init(void)
936 936
937 case 26: /* 45 nm nehalem, "Bloomfield" */ 937 case 26: /* 45 nm nehalem, "Bloomfield" */
938 case 30: /* 45 nm nehalem, "Lynnfield" */ 938 case 30: /* 45 nm nehalem, "Lynnfield" */
939 case 46: /* 45 nm nehalem-ex, "Beckton" */
939 memcpy(hw_cache_event_ids, nehalem_hw_cache_event_ids, 940 memcpy(hw_cache_event_ids, nehalem_hw_cache_event_ids,
940 sizeof(hw_cache_event_ids)); 941 sizeof(hw_cache_event_ids));
941 942
diff --git a/arch/x86/kernel/cpu/vmware.c b/arch/x86/kernel/cpu/vmware.c
index 1cbed97b59cf..dfdb4dba2320 100644
--- a/arch/x86/kernel/cpu/vmware.c
+++ b/arch/x86/kernel/cpu/vmware.c
@@ -22,6 +22,7 @@
22 */ 22 */
23 23
24#include <linux/dmi.h> 24#include <linux/dmi.h>
25#include <linux/module.h>
25#include <asm/div64.h> 26#include <asm/div64.h>
26#include <asm/vmware.h> 27#include <asm/vmware.h>
27#include <asm/x86_init.h> 28#include <asm/x86_init.h>
@@ -101,6 +102,7 @@ int vmware_platform(void)
101 102
102 return 0; 103 return 0;
103} 104}
105EXPORT_SYMBOL(vmware_platform);
104 106
105/* 107/*
106 * VMware hypervisor takes care of exporting a reliable TSC to the guest. 108 * VMware hypervisor takes care of exporting a reliable TSC to the guest.
diff --git a/arch/x86/kernel/cpuid.c b/arch/x86/kernel/cpuid.c
index 83e5e628de73..8b862d5900fe 100644
--- a/arch/x86/kernel/cpuid.c
+++ b/arch/x86/kernel/cpuid.c
@@ -40,6 +40,7 @@
40#include <linux/cpu.h> 40#include <linux/cpu.h>
41#include <linux/notifier.h> 41#include <linux/notifier.h>
42#include <linux/uaccess.h> 42#include <linux/uaccess.h>
43#include <linux/gfp.h>
43 44
44#include <asm/processor.h> 45#include <asm/processor.h>
45#include <asm/msr.h> 46#include <asm/msr.h>
diff --git a/arch/x86/kernel/crash.c b/arch/x86/kernel/crash.c
index a4849c10a77e..ebd4c51d096a 100644
--- a/arch/x86/kernel/crash.c
+++ b/arch/x86/kernel/crash.c
@@ -27,7 +27,6 @@
27#include <asm/cpu.h> 27#include <asm/cpu.h>
28#include <asm/reboot.h> 28#include <asm/reboot.h>
29#include <asm/virtext.h> 29#include <asm/virtext.h>
30#include <asm/x86_init.h>
31 30
32#if defined(CONFIG_SMP) && defined(CONFIG_X86_LOCAL_APIC) 31#if defined(CONFIG_SMP) && defined(CONFIG_X86_LOCAL_APIC)
33 32
@@ -103,10 +102,5 @@ void native_machine_crash_shutdown(struct pt_regs *regs)
103#ifdef CONFIG_HPET_TIMER 102#ifdef CONFIG_HPET_TIMER
104 hpet_disable(); 103 hpet_disable();
105#endif 104#endif
106
107#ifdef CONFIG_X86_64
108 x86_platform.iommu_shutdown();
109#endif
110
111 crash_save_cpu(regs, safe_smp_processor_id()); 105 crash_save_cpu(regs, safe_smp_processor_id());
112} 106}
diff --git a/arch/x86/kernel/crash_dump_32.c b/arch/x86/kernel/crash_dump_32.c
index cd97ce18c29d..67414550c3cc 100644
--- a/arch/x86/kernel/crash_dump_32.c
+++ b/arch/x86/kernel/crash_dump_32.c
@@ -5,6 +5,7 @@
5 * Copyright (C) IBM Corporation, 2004. All rights reserved 5 * Copyright (C) IBM Corporation, 2004. All rights reserved
6 */ 6 */
7 7
8#include <linux/slab.h>
8#include <linux/errno.h> 9#include <linux/errno.h>
9#include <linux/highmem.h> 10#include <linux/highmem.h>
10#include <linux/crash_dump.h> 11#include <linux/crash_dump.h>
diff --git a/arch/x86/kernel/dumpstack.h b/arch/x86/kernel/dumpstack.h
index 29e5f7c845b2..e1a93be4fd44 100644
--- a/arch/x86/kernel/dumpstack.h
+++ b/arch/x86/kernel/dumpstack.h
@@ -14,6 +14,8 @@
14#define get_bp(bp) asm("movq %%rbp, %0" : "=r" (bp) :) 14#define get_bp(bp) asm("movq %%rbp, %0" : "=r" (bp) :)
15#endif 15#endif
16 16
17#include <linux/uaccess.h>
18
17extern void 19extern void
18show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs, 20show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,
19 unsigned long *stack, unsigned long bp, char *log_lvl); 21 unsigned long *stack, unsigned long bp, char *log_lvl);
@@ -30,6 +32,11 @@ struct stack_frame {
30 unsigned long return_address; 32 unsigned long return_address;
31}; 33};
32 34
35struct stack_frame_ia32 {
36 u32 next_frame;
37 u32 return_address;
38};
39
33static inline unsigned long rewind_frame_pointer(int n) 40static inline unsigned long rewind_frame_pointer(int n)
34{ 41{
35 struct stack_frame *frame; 42 struct stack_frame *frame;
@@ -37,8 +44,10 @@ static inline unsigned long rewind_frame_pointer(int n)
37 get_bp(frame); 44 get_bp(frame);
38 45
39#ifdef CONFIG_FRAME_POINTER 46#ifdef CONFIG_FRAME_POINTER
40 while (n--) 47 while (n--) {
41 frame = frame->next_frame; 48 if (probe_kernel_address(&frame->next_frame, frame))
49 break;
50 }
42#endif 51#endif
43 52
44 return (unsigned long)frame; 53 return (unsigned long)frame;
diff --git a/arch/x86/kernel/e820.c b/arch/x86/kernel/e820.c
index 740b440fbd73..7bca3c6a02fb 100644
--- a/arch/x86/kernel/e820.c
+++ b/arch/x86/kernel/e820.c
@@ -519,29 +519,45 @@ u64 __init e820_remove_range(u64 start, u64 size, unsigned old_type,
519 printk(KERN_DEBUG "e820 remove range: %016Lx - %016Lx ", 519 printk(KERN_DEBUG "e820 remove range: %016Lx - %016Lx ",
520 (unsigned long long) start, 520 (unsigned long long) start,
521 (unsigned long long) end); 521 (unsigned long long) end);
522 e820_print_type(old_type); 522 if (checktype)
523 e820_print_type(old_type);
523 printk(KERN_CONT "\n"); 524 printk(KERN_CONT "\n");
524 525
525 for (i = 0; i < e820.nr_map; i++) { 526 for (i = 0; i < e820.nr_map; i++) {
526 struct e820entry *ei = &e820.map[i]; 527 struct e820entry *ei = &e820.map[i];
527 u64 final_start, final_end; 528 u64 final_start, final_end;
529 u64 ei_end;
528 530
529 if (checktype && ei->type != old_type) 531 if (checktype && ei->type != old_type)
530 continue; 532 continue;
533
534 ei_end = ei->addr + ei->size;
531 /* totally covered? */ 535 /* totally covered? */
532 if (ei->addr >= start && 536 if (ei->addr >= start && ei_end <= end) {
533 (ei->addr + ei->size) <= (start + size)) {
534 real_removed_size += ei->size; 537 real_removed_size += ei->size;
535 memset(ei, 0, sizeof(struct e820entry)); 538 memset(ei, 0, sizeof(struct e820entry));
536 continue; 539 continue;
537 } 540 }
541
542 /* new range is totally covered? */
543 if (ei->addr < start && ei_end > end) {
544 e820_add_region(end, ei_end - end, ei->type);
545 ei->size = start - ei->addr;
546 real_removed_size += size;
547 continue;
548 }
549
538 /* partially covered */ 550 /* partially covered */
539 final_start = max(start, ei->addr); 551 final_start = max(start, ei->addr);
540 final_end = min(start + size, ei->addr + ei->size); 552 final_end = min(end, ei_end);
541 if (final_start >= final_end) 553 if (final_start >= final_end)
542 continue; 554 continue;
543 real_removed_size += final_end - final_start; 555 real_removed_size += final_end - final_start;
544 556
557 /*
558 * left range could be head or tail, so need to update
559 * size at first.
560 */
545 ei->size -= final_end - final_start; 561 ei->size -= final_end - final_start;
546 if (ei->addr < final_start) 562 if (ei->addr < final_start)
547 continue; 563 continue;
diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c
index ee4fa1bfcb33..23b4ecdffa9b 100644
--- a/arch/x86/kernel/hpet.c
+++ b/arch/x86/kernel/hpet.c
@@ -4,6 +4,7 @@
4#include <linux/sysdev.h> 4#include <linux/sysdev.h>
5#include <linux/delay.h> 5#include <linux/delay.h>
6#include <linux/errno.h> 6#include <linux/errno.h>
7#include <linux/slab.h>
7#include <linux/hpet.h> 8#include <linux/hpet.h>
8#include <linux/init.h> 9#include <linux/init.h>
9#include <linux/cpu.h> 10#include <linux/cpu.h>
@@ -399,9 +400,15 @@ static int hpet_next_event(unsigned long delta,
399 * then we might have a real hardware problem. We can not do 400 * then we might have a real hardware problem. We can not do
400 * much about it here, but at least alert the user/admin with 401 * much about it here, but at least alert the user/admin with
401 * a prominent warning. 402 * a prominent warning.
403 * An erratum on some chipsets (ICH9,..), results in comparator read
404 * immediately following a write returning old value. Workaround
405 * for this is to read this value second time, when first
406 * read returns old value.
402 */ 407 */
403 WARN_ONCE(hpet_readl(HPET_Tn_CMP(timer)) != cnt, 408 if (unlikely((u32)hpet_readl(HPET_Tn_CMP(timer)) != cnt)) {
409 WARN_ONCE(hpet_readl(HPET_Tn_CMP(timer)) != cnt,
404 KERN_WARNING "hpet: compare register read back failed.\n"); 410 KERN_WARNING "hpet: compare register read back failed.\n");
411 }
405 412
406 return (s32)(hpet_readl(HPET_COUNTER) - cnt) >= 0 ? -ETIME : 0; 413 return (s32)(hpet_readl(HPET_COUNTER) - cnt) >= 0 ? -ETIME : 0;
407} 414}
@@ -1143,6 +1150,7 @@ int hpet_set_periodic_freq(unsigned long freq)
1143 do_div(clc, freq); 1150 do_div(clc, freq);
1144 clc >>= hpet_clockevent.shift; 1151 clc >>= hpet_clockevent.shift;
1145 hpet_pie_delta = clc; 1152 hpet_pie_delta = clc;
1153 hpet_pie_limit = 0;
1146 } 1154 }
1147 return 1; 1155 return 1;
1148} 1156}
diff --git a/arch/x86/kernel/i387.c b/arch/x86/kernel/i387.c
index c01a2b846d47..54c31c285488 100644
--- a/arch/x86/kernel/i387.c
+++ b/arch/x86/kernel/i387.c
@@ -8,6 +8,7 @@
8#include <linux/module.h> 8#include <linux/module.h>
9#include <linux/regset.h> 9#include <linux/regset.h>
10#include <linux/sched.h> 10#include <linux/sched.h>
11#include <linux/slab.h>
11 12
12#include <asm/sigcontext.h> 13#include <asm/sigcontext.h>
13#include <asm/processor.h> 14#include <asm/processor.h>
diff --git a/arch/x86/kernel/i8259.c b/arch/x86/kernel/i8259.c
index fb725ee15f55..7c9f02c130f3 100644
--- a/arch/x86/kernel/i8259.c
+++ b/arch/x86/kernel/i8259.c
@@ -5,7 +5,6 @@
5#include <linux/ioport.h> 5#include <linux/ioport.h>
6#include <linux/interrupt.h> 6#include <linux/interrupt.h>
7#include <linux/timex.h> 7#include <linux/timex.h>
8#include <linux/slab.h>
9#include <linux/random.h> 8#include <linux/random.h>
10#include <linux/init.h> 9#include <linux/init.h>
11#include <linux/kernel_stat.h> 10#include <linux/kernel_stat.h>
diff --git a/arch/x86/kernel/irqinit.c b/arch/x86/kernel/irqinit.c
index f01d390f9c5b..0ed2d300cd46 100644
--- a/arch/x86/kernel/irqinit.c
+++ b/arch/x86/kernel/irqinit.c
@@ -5,7 +5,6 @@
5#include <linux/ioport.h> 5#include <linux/ioport.h>
6#include <linux/interrupt.h> 6#include <linux/interrupt.h>
7#include <linux/timex.h> 7#include <linux/timex.h>
8#include <linux/slab.h>
9#include <linux/random.h> 8#include <linux/random.h>
10#include <linux/kprobes.h> 9#include <linux/kprobes.h>
11#include <linux/init.h> 10#include <linux/init.h>
diff --git a/arch/x86/kernel/k8.c b/arch/x86/kernel/k8.c
index 9b895464dd03..0f7bc20cfcde 100644
--- a/arch/x86/kernel/k8.c
+++ b/arch/x86/kernel/k8.c
@@ -2,8 +2,8 @@
2 * Shared support code for AMD K8 northbridges and derivates. 2 * Shared support code for AMD K8 northbridges and derivates.
3 * Copyright 2006 Andi Kleen, SUSE Labs. Subject to GPLv2. 3 * Copyright 2006 Andi Kleen, SUSE Labs. Subject to GPLv2.
4 */ 4 */
5#include <linux/gfp.h>
6#include <linux/types.h> 5#include <linux/types.h>
6#include <linux/slab.h>
7#include <linux/init.h> 7#include <linux/init.h>
8#include <linux/errno.h> 8#include <linux/errno.h>
9#include <linux/module.h> 9#include <linux/module.h>
diff --git a/arch/x86/kernel/kdebugfs.c b/arch/x86/kernel/kdebugfs.c
index e444357375ce..8afd9f321f10 100644
--- a/arch/x86/kernel/kdebugfs.c
+++ b/arch/x86/kernel/kdebugfs.c
@@ -9,6 +9,7 @@
9#include <linux/debugfs.h> 9#include <linux/debugfs.h>
10#include <linux/uaccess.h> 10#include <linux/uaccess.h>
11#include <linux/module.h> 11#include <linux/module.h>
12#include <linux/slab.h>
12#include <linux/init.h> 13#include <linux/init.h>
13#include <linux/stat.h> 14#include <linux/stat.h>
14#include <linux/io.h> 15#include <linux/io.h>
diff --git a/arch/x86/kernel/kgdb.c b/arch/x86/kernel/kgdb.c
index bfba6019d762..b2258ca91003 100644
--- a/arch/x86/kernel/kgdb.c
+++ b/arch/x86/kernel/kgdb.c
@@ -618,8 +618,8 @@ int kgdb_arch_init(void)
618 * portion of kgdb because this operation requires mutexs to 618 * portion of kgdb because this operation requires mutexs to
619 * complete. 619 * complete.
620 */ 620 */
621 hw_breakpoint_init(&attr);
621 attr.bp_addr = (unsigned long)kgdb_arch_init; 622 attr.bp_addr = (unsigned long)kgdb_arch_init;
622 attr.type = PERF_TYPE_BREAKPOINT;
623 attr.bp_len = HW_BREAKPOINT_LEN_1; 623 attr.bp_len = HW_BREAKPOINT_LEN_1;
624 attr.bp_type = HW_BREAKPOINT_W; 624 attr.bp_type = HW_BREAKPOINT_W;
625 attr.disabled = 1; 625 attr.disabled = 1;
diff --git a/arch/x86/kernel/kprobes.c b/arch/x86/kernel/kprobes.c
index b43bbaebe2c0..1658efdfb4e5 100644
--- a/arch/x86/kernel/kprobes.c
+++ b/arch/x86/kernel/kprobes.c
@@ -534,20 +534,6 @@ static int __kprobes kprobe_handler(struct pt_regs *regs)
534 struct kprobe_ctlblk *kcb; 534 struct kprobe_ctlblk *kcb;
535 535
536 addr = (kprobe_opcode_t *)(regs->ip - sizeof(kprobe_opcode_t)); 536 addr = (kprobe_opcode_t *)(regs->ip - sizeof(kprobe_opcode_t));
537 if (*addr != BREAKPOINT_INSTRUCTION) {
538 /*
539 * The breakpoint instruction was removed right
540 * after we hit it. Another cpu has removed
541 * either a probepoint or a debugger breakpoint
542 * at this address. In either case, no further
543 * handling of this interrupt is appropriate.
544 * Back up over the (now missing) int3 and run
545 * the original instruction.
546 */
547 regs->ip = (unsigned long)addr;
548 return 1;
549 }
550
551 /* 537 /*
552 * We don't want to be preempted for the entire 538 * We don't want to be preempted for the entire
553 * duration of kprobe processing. We conditionally 539 * duration of kprobe processing. We conditionally
@@ -579,6 +565,19 @@ static int __kprobes kprobe_handler(struct pt_regs *regs)
579 setup_singlestep(p, regs, kcb, 0); 565 setup_singlestep(p, regs, kcb, 0);
580 return 1; 566 return 1;
581 } 567 }
568 } else if (*addr != BREAKPOINT_INSTRUCTION) {
569 /*
570 * The breakpoint instruction was removed right
571 * after we hit it. Another cpu has removed
572 * either a probepoint or a debugger breakpoint
573 * at this address. In either case, no further
574 * handling of this interrupt is appropriate.
575 * Back up over the (now missing) int3 and run
576 * the original instruction.
577 */
578 regs->ip = (unsigned long)addr;
579 preempt_enable_no_resched();
580 return 1;
582 } else if (kprobe_running()) { 581 } else if (kprobe_running()) {
583 p = __get_cpu_var(current_kprobe); 582 p = __get_cpu_var(current_kprobe);
584 if (p->break_handler && p->break_handler(p, regs)) { 583 if (p->break_handler && p->break_handler(p, regs)) {
diff --git a/arch/x86/kernel/ldt.c b/arch/x86/kernel/ldt.c
index ec6ef60cbd17..ea697263b373 100644
--- a/arch/x86/kernel/ldt.c
+++ b/arch/x86/kernel/ldt.c
@@ -7,6 +7,7 @@
7 */ 7 */
8 8
9#include <linux/errno.h> 9#include <linux/errno.h>
10#include <linux/gfp.h>
10#include <linux/sched.h> 11#include <linux/sched.h>
11#include <linux/string.h> 12#include <linux/string.h>
12#include <linux/mm.h> 13#include <linux/mm.h>
diff --git a/arch/x86/kernel/machine_kexec_64.c b/arch/x86/kernel/machine_kexec_64.c
index 4a8bb82248ae..035c8c529181 100644
--- a/arch/x86/kernel/machine_kexec_64.c
+++ b/arch/x86/kernel/machine_kexec_64.c
@@ -9,6 +9,7 @@
9#include <linux/mm.h> 9#include <linux/mm.h>
10#include <linux/kexec.h> 10#include <linux/kexec.h>
11#include <linux/string.h> 11#include <linux/string.h>
12#include <linux/gfp.h>
12#include <linux/reboot.h> 13#include <linux/reboot.h>
13#include <linux/numa.h> 14#include <linux/numa.h>
14#include <linux/ftrace.h> 15#include <linux/ftrace.h>
diff --git a/arch/x86/kernel/mca_32.c b/arch/x86/kernel/mca_32.c
index 845d80ce1ef1..63eaf6596233 100644
--- a/arch/x86/kernel/mca_32.c
+++ b/arch/x86/kernel/mca_32.c
@@ -42,6 +42,7 @@
42#include <linux/kernel.h> 42#include <linux/kernel.h>
43#include <linux/mca.h> 43#include <linux/mca.h>
44#include <linux/kprobes.h> 44#include <linux/kprobes.h>
45#include <linux/slab.h>
45#include <asm/system.h> 46#include <asm/system.h>
46#include <asm/io.h> 47#include <asm/io.h>
47#include <linux/proc_fs.h> 48#include <linux/proc_fs.h>
diff --git a/arch/x86/kernel/module.c b/arch/x86/kernel/module.c
index 89f386f044e4..e0bc186d7501 100644
--- a/arch/x86/kernel/module.c
+++ b/arch/x86/kernel/module.c
@@ -23,6 +23,7 @@
23#include <linux/kernel.h> 23#include <linux/kernel.h>
24#include <linux/bug.h> 24#include <linux/bug.h>
25#include <linux/mm.h> 25#include <linux/mm.h>
26#include <linux/gfp.h>
26 27
27#include <asm/system.h> 28#include <asm/system.h>
28#include <asm/page.h> 29#include <asm/page.h>
diff --git a/arch/x86/kernel/mpparse.c b/arch/x86/kernel/mpparse.c
index a2c1edd2d3ac..e81030f71a8f 100644
--- a/arch/x86/kernel/mpparse.c
+++ b/arch/x86/kernel/mpparse.c
@@ -664,7 +664,7 @@ static void __init smp_reserve_memory(struct mpf_intel *mpf)
664{ 664{
665 unsigned long size = get_mpc_size(mpf->physptr); 665 unsigned long size = get_mpc_size(mpf->physptr);
666 666
667 reserve_early(mpf->physptr, mpf->physptr+size, "MP-table mpc"); 667 reserve_early_overlap_ok(mpf->physptr, mpf->physptr+size, "MP-table mpc");
668} 668}
669 669
670static int __init smp_scan_config(unsigned long base, unsigned long length) 670static int __init smp_scan_config(unsigned long base, unsigned long length)
@@ -693,7 +693,7 @@ static int __init smp_scan_config(unsigned long base, unsigned long length)
693 mpf, (u64)virt_to_phys(mpf)); 693 mpf, (u64)virt_to_phys(mpf));
694 694
695 mem = virt_to_phys(mpf); 695 mem = virt_to_phys(mpf);
696 reserve_early(mem, mem + sizeof(*mpf), "MP-table mpf"); 696 reserve_early_overlap_ok(mem, mem + sizeof(*mpf), "MP-table mpf");
697 if (mpf->physptr) 697 if (mpf->physptr)
698 smp_reserve_memory(mpf); 698 smp_reserve_memory(mpf);
699 699
diff --git a/arch/x86/kernel/msr.c b/arch/x86/kernel/msr.c
index 206735ac8cbd..4d4468e9f47c 100644
--- a/arch/x86/kernel/msr.c
+++ b/arch/x86/kernel/msr.c
@@ -37,6 +37,7 @@
37#include <linux/cpu.h> 37#include <linux/cpu.h>
38#include <linux/notifier.h> 38#include <linux/notifier.h>
39#include <linux/uaccess.h> 39#include <linux/uaccess.h>
40#include <linux/gfp.h>
40 41
41#include <asm/processor.h> 42#include <asm/processor.h>
42#include <asm/msr.h> 43#include <asm/msr.h>
diff --git a/arch/x86/kernel/pci-dma.c b/arch/x86/kernel/pci-dma.c
index a4ac764a6880..4b7e3d8b01dd 100644
--- a/arch/x86/kernel/pci-dma.c
+++ b/arch/x86/kernel/pci-dma.c
@@ -2,6 +2,7 @@
2#include <linux/dma-debug.h> 2#include <linux/dma-debug.h>
3#include <linux/dmar.h> 3#include <linux/dmar.h>
4#include <linux/bootmem.h> 4#include <linux/bootmem.h>
5#include <linux/gfp.h>
5#include <linux/pci.h> 6#include <linux/pci.h>
6#include <linux/kmemleak.h> 7#include <linux/kmemleak.h>
7 8
diff --git a/arch/x86/kernel/pci-gart_64.c b/arch/x86/kernel/pci-gart_64.c
index f3af115a573a..0f7f130caa67 100644
--- a/arch/x86/kernel/pci-gart_64.c
+++ b/arch/x86/kernel/pci-gart_64.c
@@ -29,6 +29,7 @@
29#include <linux/iommu-helper.h> 29#include <linux/iommu-helper.h>
30#include <linux/sysdev.h> 30#include <linux/sysdev.h>
31#include <linux/io.h> 31#include <linux/io.h>
32#include <linux/gfp.h>
32#include <asm/atomic.h> 33#include <asm/atomic.h>
33#include <asm/mtrr.h> 34#include <asm/mtrr.h>
34#include <asm/pgtable.h> 35#include <asm/pgtable.h>
@@ -564,6 +565,9 @@ static void enable_gart_translations(void)
564 565
565 enable_gart_translation(dev, __pa(agp_gatt_table)); 566 enable_gart_translation(dev, __pa(agp_gatt_table));
566 } 567 }
568
569 /* Flush the GART-TLB to remove stale entries */
570 k8_flush_garts();
567} 571}
568 572
569/* 573/*
diff --git a/arch/x86/kernel/pci-nommu.c b/arch/x86/kernel/pci-nommu.c
index 22be12b60a8f..3af4af810c07 100644
--- a/arch/x86/kernel/pci-nommu.c
+++ b/arch/x86/kernel/pci-nommu.c
@@ -4,6 +4,7 @@
4#include <linux/scatterlist.h> 4#include <linux/scatterlist.h>
5#include <linux/string.h> 5#include <linux/string.h>
6#include <linux/init.h> 6#include <linux/init.h>
7#include <linux/gfp.h>
7#include <linux/pci.h> 8#include <linux/pci.h>
8#include <linux/mm.h> 9#include <linux/mm.h>
9 10
diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c
index 28ad9f4d8b94..0415c3ef91b5 100644
--- a/arch/x86/kernel/process.c
+++ b/arch/x86/kernel/process.c
@@ -546,11 +546,13 @@ static int __cpuinit check_c1e_idle(const struct cpuinfo_x86 *c)
546 * check OSVW bit for CPUs that are not affected 546 * check OSVW bit for CPUs that are not affected
547 * by erratum #400 547 * by erratum #400
548 */ 548 */
549 rdmsrl(MSR_AMD64_OSVW_ID_LENGTH, val); 549 if (cpu_has(c, X86_FEATURE_OSVW)) {
550 if (val >= 2) { 550 rdmsrl(MSR_AMD64_OSVW_ID_LENGTH, val);
551 rdmsrl(MSR_AMD64_OSVW_STATUS, val); 551 if (val >= 2) {
552 if (!(val & BIT(1))) 552 rdmsrl(MSR_AMD64_OSVW_STATUS, val);
553 goto no_c1e_idle; 553 if (!(val & BIT(1)))
554 goto no_c1e_idle;
555 }
554 } 556 }
555 return 1; 557 return 1;
556 } 558 }
diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c
index dc9690b4c4cc..17cb3295cbf7 100644
--- a/arch/x86/kernel/process_64.c
+++ b/arch/x86/kernel/process_64.c
@@ -276,12 +276,12 @@ int copy_thread(unsigned long clone_flags, unsigned long sp,
276 276
277 set_tsk_thread_flag(p, TIF_FORK); 277 set_tsk_thread_flag(p, TIF_FORK);
278 278
279 p->thread.fs = me->thread.fs;
280 p->thread.gs = me->thread.gs;
281 p->thread.io_bitmap_ptr = NULL; 279 p->thread.io_bitmap_ptr = NULL;
282 280
283 savesegment(gs, p->thread.gsindex); 281 savesegment(gs, p->thread.gsindex);
282 p->thread.gs = p->thread.gsindex ? 0 : me->thread.gs;
284 savesegment(fs, p->thread.fsindex); 283 savesegment(fs, p->thread.fsindex);
284 p->thread.fs = p->thread.fsindex ? 0 : me->thread.fs;
285 savesegment(es, p->thread.es); 285 savesegment(es, p->thread.es);
286 savesegment(ds, p->thread.ds); 286 savesegment(ds, p->thread.ds);
287 287
diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c
index a503b1fd04e5..2e9b55027b7e 100644
--- a/arch/x86/kernel/ptrace.c
+++ b/arch/x86/kernel/ptrace.c
@@ -12,6 +12,7 @@
12#include <linux/mm.h> 12#include <linux/mm.h>
13#include <linux/smp.h> 13#include <linux/smp.h>
14#include <linux/errno.h> 14#include <linux/errno.h>
15#include <linux/slab.h>
15#include <linux/ptrace.h> 16#include <linux/ptrace.h>
16#include <linux/regset.h> 17#include <linux/regset.h>
17#include <linux/tracehook.h> 18#include <linux/tracehook.h>
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index d76e18570c60..c4851eff57b3 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -55,7 +55,6 @@
55#include <linux/stddef.h> 55#include <linux/stddef.h>
56#include <linux/unistd.h> 56#include <linux/unistd.h>
57#include <linux/ptrace.h> 57#include <linux/ptrace.h>
58#include <linux/slab.h>
59#include <linux/user.h> 58#include <linux/user.h>
60#include <linux/delay.h> 59#include <linux/delay.h>
61 60
@@ -608,6 +607,16 @@ static int __init setup_elfcorehdr(char *arg)
608early_param("elfcorehdr", setup_elfcorehdr); 607early_param("elfcorehdr", setup_elfcorehdr);
609#endif 608#endif
610 609
610static __init void reserve_ibft_region(void)
611{
612 unsigned long addr, size = 0;
613
614 addr = find_ibft_region(&size);
615
616 if (size)
617 reserve_early_overlap_ok(addr, addr + size, "ibft");
618}
619
611#ifdef CONFIG_X86_RESERVE_LOW_64K 620#ifdef CONFIG_X86_RESERVE_LOW_64K
612static int __init dmi_low_memory_corruption(const struct dmi_system_id *d) 621static int __init dmi_low_memory_corruption(const struct dmi_system_id *d)
613{ 622{
@@ -910,6 +919,8 @@ void __init setup_arch(char **cmdline_p)
910 */ 919 */
911 find_smp_config(); 920 find_smp_config();
912 921
922 reserve_ibft_region();
923
913 reserve_trampoline_memory(); 924 reserve_trampoline_memory();
914 925
915#ifdef CONFIG_ACPI_SLEEP 926#ifdef CONFIG_ACPI_SLEEP
@@ -977,8 +988,6 @@ void __init setup_arch(char **cmdline_p)
977 988
978 dma32_reserve_bootmem(); 989 dma32_reserve_bootmem();
979 990
980 reserve_ibft_region();
981
982#ifdef CONFIG_KVM_CLOCK 991#ifdef CONFIG_KVM_CLOCK
983 kvmclock_init(); 992 kvmclock_init();
984#endif 993#endif
diff --git a/arch/x86/kernel/smp.c b/arch/x86/kernel/smp.c
index ec1de97600e7..d801210945d6 100644
--- a/arch/x86/kernel/smp.c
+++ b/arch/x86/kernel/smp.c
@@ -21,6 +21,7 @@
21#include <linux/cache.h> 21#include <linux/cache.h>
22#include <linux/interrupt.h> 22#include <linux/interrupt.h>
23#include <linux/cpu.h> 23#include <linux/cpu.h>
24#include <linux/gfp.h>
24 25
25#include <asm/mtrr.h> 26#include <asm/mtrr.h>
26#include <asm/tlbflush.h> 27#include <asm/tlbflush.h>
diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
index 06d98ae5a802..763d815e27a0 100644
--- a/arch/x86/kernel/smpboot.c
+++ b/arch/x86/kernel/smpboot.c
@@ -49,6 +49,7 @@
49#include <linux/nmi.h> 49#include <linux/nmi.h>
50#include <linux/tboot.h> 50#include <linux/tboot.h>
51#include <linux/stackprotector.h> 51#include <linux/stackprotector.h>
52#include <linux/gfp.h>
52 53
53#include <asm/acpi.h> 54#include <asm/acpi.h>
54#include <asm/desc.h> 55#include <asm/desc.h>
@@ -242,8 +243,6 @@ static void __cpuinit smp_callin(void)
242 end_local_APIC_setup(); 243 end_local_APIC_setup();
243 map_cpu_to_logical_apicid(); 244 map_cpu_to_logical_apicid();
244 245
245 notify_cpu_starting(cpuid);
246
247 /* 246 /*
248 * Need to setup vector mappings before we enable interrupts. 247 * Need to setup vector mappings before we enable interrupts.
249 */ 248 */
@@ -264,6 +263,8 @@ static void __cpuinit smp_callin(void)
264 */ 263 */
265 smp_store_cpu_info(cpuid); 264 smp_store_cpu_info(cpuid);
266 265
266 notify_cpu_starting(cpuid);
267
267 /* 268 /*
268 * Allow the master to continue. 269 * Allow the master to continue.
269 */ 270 */
diff --git a/arch/x86/kernel/tlb_uv.c b/arch/x86/kernel/tlb_uv.c
index 364d015efebc..17b03dd3a6b5 100644
--- a/arch/x86/kernel/tlb_uv.c
+++ b/arch/x86/kernel/tlb_uv.c
@@ -9,6 +9,7 @@
9#include <linux/seq_file.h> 9#include <linux/seq_file.h>
10#include <linux/proc_fs.h> 10#include <linux/proc_fs.h>
11#include <linux/kernel.h> 11#include <linux/kernel.h>
12#include <linux/slab.h>
12 13
13#include <asm/mmu_context.h> 14#include <asm/mmu_context.h>
14#include <asm/uv/uv.h> 15#include <asm/uv/uv.h>
diff --git a/arch/x86/kernel/uv_irq.c b/arch/x86/kernel/uv_irq.c
index ece73d8e3240..1d40336b030a 100644
--- a/arch/x86/kernel/uv_irq.c
+++ b/arch/x86/kernel/uv_irq.c
@@ -10,6 +10,7 @@
10 10
11#include <linux/module.h> 11#include <linux/module.h>
12#include <linux/rbtree.h> 12#include <linux/rbtree.h>
13#include <linux/slab.h>
13#include <linux/irq.h> 14#include <linux/irq.h>
14 15
15#include <asm/apic.h> 16#include <asm/apic.h>
diff --git a/arch/x86/kernel/uv_time.c b/arch/x86/kernel/uv_time.c
index 2b75ef638dbc..56e421bc379b 100644
--- a/arch/x86/kernel/uv_time.c
+++ b/arch/x86/kernel/uv_time.c
@@ -19,6 +19,7 @@
19 * Copyright (c) Dimitri Sivanich 19 * Copyright (c) Dimitri Sivanich
20 */ 20 */
21#include <linux/clockchips.h> 21#include <linux/clockchips.h>
22#include <linux/slab.h>
22 23
23#include <asm/uv/uv_mmrs.h> 24#include <asm/uv/uv_mmrs.h>
24#include <asm/uv/uv_hub.h> 25#include <asm/uv/uv_hub.h>
diff --git a/arch/x86/kernel/vmi_32.c b/arch/x86/kernel/vmi_32.c
index 7dd599deca4a..ce9fbacb7526 100644
--- a/arch/x86/kernel/vmi_32.c
+++ b/arch/x86/kernel/vmi_32.c
@@ -28,6 +28,7 @@
28#include <linux/mm.h> 28#include <linux/mm.h>
29#include <linux/highmem.h> 29#include <linux/highmem.h>
30#include <linux/sched.h> 30#include <linux/sched.h>
31#include <linux/gfp.h>
31#include <asm/vmi.h> 32#include <asm/vmi.h>
32#include <asm/io.h> 33#include <asm/io.h>
33#include <asm/fixmap.h> 34#include <asm/fixmap.h>
diff --git a/arch/x86/kvm/i8254.c b/arch/x86/kvm/i8254.c
index 294698b6daff..0150affad25d 100644
--- a/arch/x86/kvm/i8254.c
+++ b/arch/x86/kvm/i8254.c
@@ -32,6 +32,7 @@
32#define pr_fmt(fmt) "pit: " fmt 32#define pr_fmt(fmt) "pit: " fmt
33 33
34#include <linux/kvm_host.h> 34#include <linux/kvm_host.h>
35#include <linux/slab.h>
35 36
36#include "irq.h" 37#include "irq.h"
37#include "i8254.h" 38#include "i8254.h"
diff --git a/arch/x86/kvm/i8259.c b/arch/x86/kvm/i8259.c
index 07771da85de5..a790fa128a9f 100644
--- a/arch/x86/kvm/i8259.c
+++ b/arch/x86/kvm/i8259.c
@@ -26,6 +26,7 @@
26 * Port from Qemu. 26 * Port from Qemu.
27 */ 27 */
28#include <linux/mm.h> 28#include <linux/mm.h>
29#include <linux/slab.h>
29#include <linux/bitops.h> 30#include <linux/bitops.h>
30#include "irq.h" 31#include "irq.h"
31 32
diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c
index 4b224f90087b..1eb7a4ae0c9c 100644
--- a/arch/x86/kvm/lapic.c
+++ b/arch/x86/kvm/lapic.c
@@ -26,6 +26,7 @@
26#include <linux/io.h> 26#include <linux/io.h>
27#include <linux/module.h> 27#include <linux/module.h>
28#include <linux/math64.h> 28#include <linux/math64.h>
29#include <linux/slab.h>
29#include <asm/processor.h> 30#include <asm/processor.h>
30#include <asm/msr.h> 31#include <asm/msr.h>
31#include <asm/page.h> 32#include <asm/page.h>
diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index 741373e8ca77..19a8906bcaa2 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -31,6 +31,7 @@
31#include <linux/hugetlb.h> 31#include <linux/hugetlb.h>
32#include <linux/compiler.h> 32#include <linux/compiler.h>
33#include <linux/srcu.h> 33#include <linux/srcu.h>
34#include <linux/slab.h>
34 35
35#include <asm/page.h> 36#include <asm/page.h>
36#include <asm/cmpxchg.h> 37#include <asm/cmpxchg.h>
@@ -1489,8 +1490,8 @@ static int mmu_zap_unsync_children(struct kvm *kvm,
1489 for_each_sp(pages, sp, parents, i) { 1490 for_each_sp(pages, sp, parents, i) {
1490 kvm_mmu_zap_page(kvm, sp); 1491 kvm_mmu_zap_page(kvm, sp);
1491 mmu_pages_clear_parents(&parents); 1492 mmu_pages_clear_parents(&parents);
1493 zapped++;
1492 } 1494 }
1493 zapped += pages.nr;
1494 kvm_mmu_pages_init(parent, &parents, &pages); 1495 kvm_mmu_pages_init(parent, &parents, &pages);
1495 } 1496 }
1496 1497
@@ -1541,14 +1542,16 @@ void kvm_mmu_change_mmu_pages(struct kvm *kvm, unsigned int kvm_nr_mmu_pages)
1541 */ 1542 */
1542 1543
1543 if (used_pages > kvm_nr_mmu_pages) { 1544 if (used_pages > kvm_nr_mmu_pages) {
1544 while (used_pages > kvm_nr_mmu_pages) { 1545 while (used_pages > kvm_nr_mmu_pages &&
1546 !list_empty(&kvm->arch.active_mmu_pages)) {
1545 struct kvm_mmu_page *page; 1547 struct kvm_mmu_page *page;
1546 1548
1547 page = container_of(kvm->arch.active_mmu_pages.prev, 1549 page = container_of(kvm->arch.active_mmu_pages.prev,
1548 struct kvm_mmu_page, link); 1550 struct kvm_mmu_page, link);
1549 kvm_mmu_zap_page(kvm, page); 1551 used_pages -= kvm_mmu_zap_page(kvm, page);
1550 used_pages--; 1552 used_pages--;
1551 } 1553 }
1554 kvm_nr_mmu_pages = used_pages;
1552 kvm->arch.n_free_mmu_pages = 0; 1555 kvm->arch.n_free_mmu_pages = 0;
1553 } 1556 }
1554 else 1557 else
@@ -1595,7 +1598,8 @@ static void mmu_unshadow(struct kvm *kvm, gfn_t gfn)
1595 && !sp->role.invalid) { 1598 && !sp->role.invalid) {
1596 pgprintk("%s: zap %lx %x\n", 1599 pgprintk("%s: zap %lx %x\n",
1597 __func__, gfn, sp->role.word); 1600 __func__, gfn, sp->role.word);
1598 kvm_mmu_zap_page(kvm, sp); 1601 if (kvm_mmu_zap_page(kvm, sp))
1602 nn = bucket->first;
1599 } 1603 }
1600 } 1604 }
1601} 1605}
diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
index 52f78dd03010..737361fcd503 100644
--- a/arch/x86/kvm/svm.c
+++ b/arch/x86/kvm/svm.c
@@ -26,6 +26,7 @@
26#include <linux/highmem.h> 26#include <linux/highmem.h>
27#include <linux/sched.h> 27#include <linux/sched.h>
28#include <linux/ftrace_event.h> 28#include <linux/ftrace_event.h>
29#include <linux/slab.h>
29 30
30#include <asm/desc.h> 31#include <asm/desc.h>
31 32
@@ -705,29 +706,28 @@ static struct kvm_vcpu *svm_create_vcpu(struct kvm *kvm, unsigned int id)
705 if (err) 706 if (err)
706 goto free_svm; 707 goto free_svm;
707 708
709 err = -ENOMEM;
708 page = alloc_page(GFP_KERNEL); 710 page = alloc_page(GFP_KERNEL);
709 if (!page) { 711 if (!page)
710 err = -ENOMEM;
711 goto uninit; 712 goto uninit;
712 }
713 713
714 err = -ENOMEM;
715 msrpm_pages = alloc_pages(GFP_KERNEL, MSRPM_ALLOC_ORDER); 714 msrpm_pages = alloc_pages(GFP_KERNEL, MSRPM_ALLOC_ORDER);
716 if (!msrpm_pages) 715 if (!msrpm_pages)
717 goto uninit; 716 goto free_page1;
718 717
719 nested_msrpm_pages = alloc_pages(GFP_KERNEL, MSRPM_ALLOC_ORDER); 718 nested_msrpm_pages = alloc_pages(GFP_KERNEL, MSRPM_ALLOC_ORDER);
720 if (!nested_msrpm_pages) 719 if (!nested_msrpm_pages)
721 goto uninit; 720 goto free_page2;
722
723 svm->msrpm = page_address(msrpm_pages);
724 svm_vcpu_init_msrpm(svm->msrpm);
725 721
726 hsave_page = alloc_page(GFP_KERNEL); 722 hsave_page = alloc_page(GFP_KERNEL);
727 if (!hsave_page) 723 if (!hsave_page)
728 goto uninit; 724 goto free_page3;
725
729 svm->nested.hsave = page_address(hsave_page); 726 svm->nested.hsave = page_address(hsave_page);
730 727
728 svm->msrpm = page_address(msrpm_pages);
729 svm_vcpu_init_msrpm(svm->msrpm);
730
731 svm->nested.msrpm = page_address(nested_msrpm_pages); 731 svm->nested.msrpm = page_address(nested_msrpm_pages);
732 732
733 svm->vmcb = page_address(page); 733 svm->vmcb = page_address(page);
@@ -743,6 +743,12 @@ static struct kvm_vcpu *svm_create_vcpu(struct kvm *kvm, unsigned int id)
743 743
744 return &svm->vcpu; 744 return &svm->vcpu;
745 745
746free_page3:
747 __free_pages(nested_msrpm_pages, MSRPM_ALLOC_ORDER);
748free_page2:
749 __free_pages(msrpm_pages, MSRPM_ALLOC_ORDER);
750free_page1:
751 __free_page(page);
746uninit: 752uninit:
747 kvm_vcpu_uninit(&svm->vcpu); 753 kvm_vcpu_uninit(&svm->vcpu);
748free_svm: 754free_svm:
@@ -2061,7 +2067,7 @@ static int cpuid_interception(struct vcpu_svm *svm)
2061static int iret_interception(struct vcpu_svm *svm) 2067static int iret_interception(struct vcpu_svm *svm)
2062{ 2068{
2063 ++svm->vcpu.stat.nmi_window_exits; 2069 ++svm->vcpu.stat.nmi_window_exits;
2064 svm->vmcb->control.intercept &= ~(1UL << INTERCEPT_IRET); 2070 svm->vmcb->control.intercept &= ~(1ULL << INTERCEPT_IRET);
2065 svm->vcpu.arch.hflags |= HF_IRET_MASK; 2071 svm->vcpu.arch.hflags |= HF_IRET_MASK;
2066 return 1; 2072 return 1;
2067} 2073}
@@ -2473,7 +2479,7 @@ static void svm_inject_nmi(struct kvm_vcpu *vcpu)
2473 2479
2474 svm->vmcb->control.event_inj = SVM_EVTINJ_VALID | SVM_EVTINJ_TYPE_NMI; 2480 svm->vmcb->control.event_inj = SVM_EVTINJ_VALID | SVM_EVTINJ_TYPE_NMI;
2475 vcpu->arch.hflags |= HF_NMI_MASK; 2481 vcpu->arch.hflags |= HF_NMI_MASK;
2476 svm->vmcb->control.intercept |= (1UL << INTERCEPT_IRET); 2482 svm->vmcb->control.intercept |= (1ULL << INTERCEPT_IRET);
2477 ++vcpu->stat.nmi_injections; 2483 ++vcpu->stat.nmi_injections;
2478} 2484}
2479 2485
@@ -2533,10 +2539,10 @@ static void svm_set_nmi_mask(struct kvm_vcpu *vcpu, bool masked)
2533 2539
2534 if (masked) { 2540 if (masked) {
2535 svm->vcpu.arch.hflags |= HF_NMI_MASK; 2541 svm->vcpu.arch.hflags |= HF_NMI_MASK;
2536 svm->vmcb->control.intercept |= (1UL << INTERCEPT_IRET); 2542 svm->vmcb->control.intercept |= (1ULL << INTERCEPT_IRET);
2537 } else { 2543 } else {
2538 svm->vcpu.arch.hflags &= ~HF_NMI_MASK; 2544 svm->vcpu.arch.hflags &= ~HF_NMI_MASK;
2539 svm->vmcb->control.intercept &= ~(1UL << INTERCEPT_IRET); 2545 svm->vmcb->control.intercept &= ~(1ULL << INTERCEPT_IRET);
2540 } 2546 }
2541} 2547}
2542 2548
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 14873b9f8430..2f8db0ec8ae4 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -26,6 +26,7 @@
26#include <linux/sched.h> 26#include <linux/sched.h>
27#include <linux/moduleparam.h> 27#include <linux/moduleparam.h>
28#include <linux/ftrace_event.h> 28#include <linux/ftrace_event.h>
29#include <linux/slab.h>
29#include "kvm_cache_regs.h" 30#include "kvm_cache_regs.h"
30#include "x86.h" 31#include "x86.h"
31 32
@@ -76,6 +77,8 @@ module_param(emulate_invalid_guest_state, bool, S_IRUGO);
76#define KVM_PMODE_VM_CR4_ALWAYS_ON (X86_CR4_PAE | X86_CR4_VMXE) 77#define KVM_PMODE_VM_CR4_ALWAYS_ON (X86_CR4_PAE | X86_CR4_VMXE)
77#define KVM_RMODE_VM_CR4_ALWAYS_ON (X86_CR4_VME | X86_CR4_PAE | X86_CR4_VMXE) 78#define KVM_RMODE_VM_CR4_ALWAYS_ON (X86_CR4_VME | X86_CR4_PAE | X86_CR4_VMXE)
78 79
80#define RMODE_GUEST_OWNED_EFLAGS_BITS (~(X86_EFLAGS_IOPL | X86_EFLAGS_VM))
81
79/* 82/*
80 * These 2 parameters are used to config the controls for Pause-Loop Exiting: 83 * These 2 parameters are used to config the controls for Pause-Loop Exiting:
81 * ple_gap: upper bound on the amount of time between two successive 84 * ple_gap: upper bound on the amount of time between two successive
@@ -130,7 +133,7 @@ struct vcpu_vmx {
130 } host_state; 133 } host_state;
131 struct { 134 struct {
132 int vm86_active; 135 int vm86_active;
133 u8 save_iopl; 136 ulong save_rflags;
134 struct kvm_save_segment { 137 struct kvm_save_segment {
135 u16 selector; 138 u16 selector;
136 unsigned long base; 139 unsigned long base;
@@ -817,18 +820,23 @@ static void vmx_fpu_deactivate(struct kvm_vcpu *vcpu)
817 820
818static unsigned long vmx_get_rflags(struct kvm_vcpu *vcpu) 821static unsigned long vmx_get_rflags(struct kvm_vcpu *vcpu)
819{ 822{
820 unsigned long rflags; 823 unsigned long rflags, save_rflags;
821 824
822 rflags = vmcs_readl(GUEST_RFLAGS); 825 rflags = vmcs_readl(GUEST_RFLAGS);
823 if (to_vmx(vcpu)->rmode.vm86_active) 826 if (to_vmx(vcpu)->rmode.vm86_active) {
824 rflags &= ~(unsigned long)(X86_EFLAGS_IOPL | X86_EFLAGS_VM); 827 rflags &= RMODE_GUEST_OWNED_EFLAGS_BITS;
828 save_rflags = to_vmx(vcpu)->rmode.save_rflags;
829 rflags |= save_rflags & ~RMODE_GUEST_OWNED_EFLAGS_BITS;
830 }
825 return rflags; 831 return rflags;
826} 832}
827 833
828static void vmx_set_rflags(struct kvm_vcpu *vcpu, unsigned long rflags) 834static void vmx_set_rflags(struct kvm_vcpu *vcpu, unsigned long rflags)
829{ 835{
830 if (to_vmx(vcpu)->rmode.vm86_active) 836 if (to_vmx(vcpu)->rmode.vm86_active) {
837 to_vmx(vcpu)->rmode.save_rflags = rflags;
831 rflags |= X86_EFLAGS_IOPL | X86_EFLAGS_VM; 838 rflags |= X86_EFLAGS_IOPL | X86_EFLAGS_VM;
839 }
832 vmcs_writel(GUEST_RFLAGS, rflags); 840 vmcs_writel(GUEST_RFLAGS, rflags);
833} 841}
834 842
@@ -1482,8 +1490,8 @@ static void enter_pmode(struct kvm_vcpu *vcpu)
1482 vmcs_write32(GUEST_TR_AR_BYTES, vmx->rmode.tr.ar); 1490 vmcs_write32(GUEST_TR_AR_BYTES, vmx->rmode.tr.ar);
1483 1491
1484 flags = vmcs_readl(GUEST_RFLAGS); 1492 flags = vmcs_readl(GUEST_RFLAGS);
1485 flags &= ~(X86_EFLAGS_IOPL | X86_EFLAGS_VM); 1493 flags &= RMODE_GUEST_OWNED_EFLAGS_BITS;
1486 flags |= (vmx->rmode.save_iopl << IOPL_SHIFT); 1494 flags |= vmx->rmode.save_rflags & ~RMODE_GUEST_OWNED_EFLAGS_BITS;
1487 vmcs_writel(GUEST_RFLAGS, flags); 1495 vmcs_writel(GUEST_RFLAGS, flags);
1488 1496
1489 vmcs_writel(GUEST_CR4, (vmcs_readl(GUEST_CR4) & ~X86_CR4_VME) | 1497 vmcs_writel(GUEST_CR4, (vmcs_readl(GUEST_CR4) & ~X86_CR4_VME) |
@@ -1556,8 +1564,7 @@ static void enter_rmode(struct kvm_vcpu *vcpu)
1556 vmcs_write32(GUEST_TR_AR_BYTES, 0x008b); 1564 vmcs_write32(GUEST_TR_AR_BYTES, 0x008b);
1557 1565
1558 flags = vmcs_readl(GUEST_RFLAGS); 1566 flags = vmcs_readl(GUEST_RFLAGS);
1559 vmx->rmode.save_iopl 1567 vmx->rmode.save_rflags = flags;
1560 = (flags & X86_EFLAGS_IOPL) >> IOPL_SHIFT;
1561 1568
1562 flags |= X86_EFLAGS_IOPL | X86_EFLAGS_VM; 1569 flags |= X86_EFLAGS_IOPL | X86_EFLAGS_VM;
1563 1570
@@ -2696,8 +2703,7 @@ static int vmx_nmi_allowed(struct kvm_vcpu *vcpu)
2696 return 0; 2703 return 0;
2697 2704
2698 return !(vmcs_read32(GUEST_INTERRUPTIBILITY_INFO) & 2705 return !(vmcs_read32(GUEST_INTERRUPTIBILITY_INFO) &
2699 (GUEST_INTR_STATE_STI | GUEST_INTR_STATE_MOV_SS | 2706 (GUEST_INTR_STATE_MOV_SS | GUEST_INTR_STATE_NMI));
2700 GUEST_INTR_STATE_NMI));
2701} 2707}
2702 2708
2703static bool vmx_get_nmi_mask(struct kvm_vcpu *vcpu) 2709static bool vmx_get_nmi_mask(struct kvm_vcpu *vcpu)
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index e46282a56565..c4f35b545c1d 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -39,6 +39,7 @@
39#include <linux/cpufreq.h> 39#include <linux/cpufreq.h>
40#include <linux/user-return-notifier.h> 40#include <linux/user-return-notifier.h>
41#include <linux/srcu.h> 41#include <linux/srcu.h>
42#include <linux/slab.h>
42#include <trace/events/kvm.h> 43#include <trace/events/kvm.h>
43#undef TRACE_INCLUDE_FILE 44#undef TRACE_INCLUDE_FILE
44#define CREATE_TRACE_POINTS 45#define CREATE_TRACE_POINTS
@@ -432,8 +433,6 @@ void kvm_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0)
432 433
433#ifdef CONFIG_X86_64 434#ifdef CONFIG_X86_64
434 if (cr0 & 0xffffffff00000000UL) { 435 if (cr0 & 0xffffffff00000000UL) {
435 printk(KERN_DEBUG "set_cr0: 0x%lx #GP, reserved bits 0x%lx\n",
436 cr0, kvm_read_cr0(vcpu));
437 kvm_inject_gp(vcpu, 0); 436 kvm_inject_gp(vcpu, 0);
438 return; 437 return;
439 } 438 }
@@ -442,14 +441,11 @@ void kvm_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0)
442 cr0 &= ~CR0_RESERVED_BITS; 441 cr0 &= ~CR0_RESERVED_BITS;
443 442
444 if ((cr0 & X86_CR0_NW) && !(cr0 & X86_CR0_CD)) { 443 if ((cr0 & X86_CR0_NW) && !(cr0 & X86_CR0_CD)) {
445 printk(KERN_DEBUG "set_cr0: #GP, CD == 0 && NW == 1\n");
446 kvm_inject_gp(vcpu, 0); 444 kvm_inject_gp(vcpu, 0);
447 return; 445 return;
448 } 446 }
449 447
450 if ((cr0 & X86_CR0_PG) && !(cr0 & X86_CR0_PE)) { 448 if ((cr0 & X86_CR0_PG) && !(cr0 & X86_CR0_PE)) {
451 printk(KERN_DEBUG "set_cr0: #GP, set PG flag "
452 "and a clear PE flag\n");
453 kvm_inject_gp(vcpu, 0); 449 kvm_inject_gp(vcpu, 0);
454 return; 450 return;
455 } 451 }
@@ -460,15 +456,11 @@ void kvm_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0)
460 int cs_db, cs_l; 456 int cs_db, cs_l;
461 457
462 if (!is_pae(vcpu)) { 458 if (!is_pae(vcpu)) {
463 printk(KERN_DEBUG "set_cr0: #GP, start paging "
464 "in long mode while PAE is disabled\n");
465 kvm_inject_gp(vcpu, 0); 459 kvm_inject_gp(vcpu, 0);
466 return; 460 return;
467 } 461 }
468 kvm_x86_ops->get_cs_db_l_bits(vcpu, &cs_db, &cs_l); 462 kvm_x86_ops->get_cs_db_l_bits(vcpu, &cs_db, &cs_l);
469 if (cs_l) { 463 if (cs_l) {
470 printk(KERN_DEBUG "set_cr0: #GP, start paging "
471 "in long mode while CS.L == 1\n");
472 kvm_inject_gp(vcpu, 0); 464 kvm_inject_gp(vcpu, 0);
473 return; 465 return;
474 466
@@ -476,8 +468,6 @@ void kvm_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0)
476 } else 468 } else
477#endif 469#endif
478 if (is_pae(vcpu) && !load_pdptrs(vcpu, vcpu->arch.cr3)) { 470 if (is_pae(vcpu) && !load_pdptrs(vcpu, vcpu->arch.cr3)) {
479 printk(KERN_DEBUG "set_cr0: #GP, pdptrs "
480 "reserved bits\n");
481 kvm_inject_gp(vcpu, 0); 471 kvm_inject_gp(vcpu, 0);
482 return; 472 return;
483 } 473 }
@@ -504,28 +494,23 @@ void kvm_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4)
504 unsigned long pdptr_bits = X86_CR4_PGE | X86_CR4_PSE | X86_CR4_PAE; 494 unsigned long pdptr_bits = X86_CR4_PGE | X86_CR4_PSE | X86_CR4_PAE;
505 495
506 if (cr4 & CR4_RESERVED_BITS) { 496 if (cr4 & CR4_RESERVED_BITS) {
507 printk(KERN_DEBUG "set_cr4: #GP, reserved bits\n");
508 kvm_inject_gp(vcpu, 0); 497 kvm_inject_gp(vcpu, 0);
509 return; 498 return;
510 } 499 }
511 500
512 if (is_long_mode(vcpu)) { 501 if (is_long_mode(vcpu)) {
513 if (!(cr4 & X86_CR4_PAE)) { 502 if (!(cr4 & X86_CR4_PAE)) {
514 printk(KERN_DEBUG "set_cr4: #GP, clearing PAE while "
515 "in long mode\n");
516 kvm_inject_gp(vcpu, 0); 503 kvm_inject_gp(vcpu, 0);
517 return; 504 return;
518 } 505 }
519 } else if (is_paging(vcpu) && (cr4 & X86_CR4_PAE) 506 } else if (is_paging(vcpu) && (cr4 & X86_CR4_PAE)
520 && ((cr4 ^ old_cr4) & pdptr_bits) 507 && ((cr4 ^ old_cr4) & pdptr_bits)
521 && !load_pdptrs(vcpu, vcpu->arch.cr3)) { 508 && !load_pdptrs(vcpu, vcpu->arch.cr3)) {
522 printk(KERN_DEBUG "set_cr4: #GP, pdptrs reserved bits\n");
523 kvm_inject_gp(vcpu, 0); 509 kvm_inject_gp(vcpu, 0);
524 return; 510 return;
525 } 511 }
526 512
527 if (cr4 & X86_CR4_VMXE) { 513 if (cr4 & X86_CR4_VMXE) {
528 printk(KERN_DEBUG "set_cr4: #GP, setting VMXE\n");
529 kvm_inject_gp(vcpu, 0); 514 kvm_inject_gp(vcpu, 0);
530 return; 515 return;
531 } 516 }
@@ -546,21 +531,16 @@ void kvm_set_cr3(struct kvm_vcpu *vcpu, unsigned long cr3)
546 531
547 if (is_long_mode(vcpu)) { 532 if (is_long_mode(vcpu)) {
548 if (cr3 & CR3_L_MODE_RESERVED_BITS) { 533 if (cr3 & CR3_L_MODE_RESERVED_BITS) {
549 printk(KERN_DEBUG "set_cr3: #GP, reserved bits\n");
550 kvm_inject_gp(vcpu, 0); 534 kvm_inject_gp(vcpu, 0);
551 return; 535 return;
552 } 536 }
553 } else { 537 } else {
554 if (is_pae(vcpu)) { 538 if (is_pae(vcpu)) {
555 if (cr3 & CR3_PAE_RESERVED_BITS) { 539 if (cr3 & CR3_PAE_RESERVED_BITS) {
556 printk(KERN_DEBUG
557 "set_cr3: #GP, reserved bits\n");
558 kvm_inject_gp(vcpu, 0); 540 kvm_inject_gp(vcpu, 0);
559 return; 541 return;
560 } 542 }
561 if (is_paging(vcpu) && !load_pdptrs(vcpu, cr3)) { 543 if (is_paging(vcpu) && !load_pdptrs(vcpu, cr3)) {
562 printk(KERN_DEBUG "set_cr3: #GP, pdptrs "
563 "reserved bits\n");
564 kvm_inject_gp(vcpu, 0); 544 kvm_inject_gp(vcpu, 0);
565 return; 545 return;
566 } 546 }
@@ -592,7 +572,6 @@ EXPORT_SYMBOL_GPL(kvm_set_cr3);
592void kvm_set_cr8(struct kvm_vcpu *vcpu, unsigned long cr8) 572void kvm_set_cr8(struct kvm_vcpu *vcpu, unsigned long cr8)
593{ 573{
594 if (cr8 & CR8_RESERVED_BITS) { 574 if (cr8 & CR8_RESERVED_BITS) {
595 printk(KERN_DEBUG "set_cr8: #GP, reserved bits 0x%lx\n", cr8);
596 kvm_inject_gp(vcpu, 0); 575 kvm_inject_gp(vcpu, 0);
597 return; 576 return;
598 } 577 }
@@ -648,15 +627,12 @@ static u32 emulated_msrs[] = {
648static void set_efer(struct kvm_vcpu *vcpu, u64 efer) 627static void set_efer(struct kvm_vcpu *vcpu, u64 efer)
649{ 628{
650 if (efer & efer_reserved_bits) { 629 if (efer & efer_reserved_bits) {
651 printk(KERN_DEBUG "set_efer: 0x%llx #GP, reserved bits\n",
652 efer);
653 kvm_inject_gp(vcpu, 0); 630 kvm_inject_gp(vcpu, 0);
654 return; 631 return;
655 } 632 }
656 633
657 if (is_paging(vcpu) 634 if (is_paging(vcpu)
658 && (vcpu->arch.efer & EFER_LME) != (efer & EFER_LME)) { 635 && (vcpu->arch.efer & EFER_LME) != (efer & EFER_LME)) {
659 printk(KERN_DEBUG "set_efer: #GP, change LME while paging\n");
660 kvm_inject_gp(vcpu, 0); 636 kvm_inject_gp(vcpu, 0);
661 return; 637 return;
662 } 638 }
@@ -666,7 +642,6 @@ static void set_efer(struct kvm_vcpu *vcpu, u64 efer)
666 642
667 feat = kvm_find_cpuid_entry(vcpu, 0x80000001, 0); 643 feat = kvm_find_cpuid_entry(vcpu, 0x80000001, 0);
668 if (!feat || !(feat->edx & bit(X86_FEATURE_FXSR_OPT))) { 644 if (!feat || !(feat->edx & bit(X86_FEATURE_FXSR_OPT))) {
669 printk(KERN_DEBUG "set_efer: #GP, enable FFXSR w/o CPUID capability\n");
670 kvm_inject_gp(vcpu, 0); 645 kvm_inject_gp(vcpu, 0);
671 return; 646 return;
672 } 647 }
@@ -677,7 +652,6 @@ static void set_efer(struct kvm_vcpu *vcpu, u64 efer)
677 652
678 feat = kvm_find_cpuid_entry(vcpu, 0x80000001, 0); 653 feat = kvm_find_cpuid_entry(vcpu, 0x80000001, 0);
679 if (!feat || !(feat->ecx & bit(X86_FEATURE_SVM))) { 654 if (!feat || !(feat->ecx & bit(X86_FEATURE_SVM))) {
680 printk(KERN_DEBUG "set_efer: #GP, enable SVM w/o SVM\n");
681 kvm_inject_gp(vcpu, 0); 655 kvm_inject_gp(vcpu, 0);
682 return; 656 return;
683 } 657 }
@@ -966,9 +940,13 @@ static int set_msr_mce(struct kvm_vcpu *vcpu, u32 msr, u64 data)
966 if (msr >= MSR_IA32_MC0_CTL && 940 if (msr >= MSR_IA32_MC0_CTL &&
967 msr < MSR_IA32_MC0_CTL + 4 * bank_num) { 941 msr < MSR_IA32_MC0_CTL + 4 * bank_num) {
968 u32 offset = msr - MSR_IA32_MC0_CTL; 942 u32 offset = msr - MSR_IA32_MC0_CTL;
969 /* only 0 or all 1s can be written to IA32_MCi_CTL */ 943 /* only 0 or all 1s can be written to IA32_MCi_CTL
944 * some Linux kernels though clear bit 10 in bank 4 to
945 * workaround a BIOS/GART TBL issue on AMD K8s, ignore
946 * this to avoid an uncatched #GP in the guest
947 */
970 if ((offset & 0x3) == 0 && 948 if ((offset & 0x3) == 0 &&
971 data != 0 && data != ~(u64)0) 949 data != 0 && (data | (1 << 10)) != ~(u64)0)
972 return -1; 950 return -1;
973 vcpu->arch.mce_banks[offset] = data; 951 vcpu->arch.mce_banks[offset] = data;
974 break; 952 break;
@@ -1734,6 +1712,7 @@ static int kvm_vcpu_ioctl_set_cpuid(struct kvm_vcpu *vcpu,
1734 if (copy_from_user(cpuid_entries, entries, 1712 if (copy_from_user(cpuid_entries, entries,
1735 cpuid->nent * sizeof(struct kvm_cpuid_entry))) 1713 cpuid->nent * sizeof(struct kvm_cpuid_entry)))
1736 goto out_free; 1714 goto out_free;
1715 vcpu_load(vcpu);
1737 for (i = 0; i < cpuid->nent; i++) { 1716 for (i = 0; i < cpuid->nent; i++) {
1738 vcpu->arch.cpuid_entries[i].function = cpuid_entries[i].function; 1717 vcpu->arch.cpuid_entries[i].function = cpuid_entries[i].function;
1739 vcpu->arch.cpuid_entries[i].eax = cpuid_entries[i].eax; 1718 vcpu->arch.cpuid_entries[i].eax = cpuid_entries[i].eax;
@@ -1751,6 +1730,7 @@ static int kvm_vcpu_ioctl_set_cpuid(struct kvm_vcpu *vcpu,
1751 r = 0; 1730 r = 0;
1752 kvm_apic_set_version(vcpu); 1731 kvm_apic_set_version(vcpu);
1753 kvm_x86_ops->cpuid_update(vcpu); 1732 kvm_x86_ops->cpuid_update(vcpu);
1733 vcpu_put(vcpu);
1754 1734
1755out_free: 1735out_free:
1756 vfree(cpuid_entries); 1736 vfree(cpuid_entries);
@@ -1771,9 +1751,11 @@ static int kvm_vcpu_ioctl_set_cpuid2(struct kvm_vcpu *vcpu,
1771 if (copy_from_user(&vcpu->arch.cpuid_entries, entries, 1751 if (copy_from_user(&vcpu->arch.cpuid_entries, entries,
1772 cpuid->nent * sizeof(struct kvm_cpuid_entry2))) 1752 cpuid->nent * sizeof(struct kvm_cpuid_entry2)))
1773 goto out; 1753 goto out;
1754 vcpu_load(vcpu);
1774 vcpu->arch.cpuid_nent = cpuid->nent; 1755 vcpu->arch.cpuid_nent = cpuid->nent;
1775 kvm_apic_set_version(vcpu); 1756 kvm_apic_set_version(vcpu);
1776 kvm_x86_ops->cpuid_update(vcpu); 1757 kvm_x86_ops->cpuid_update(vcpu);
1758 vcpu_put(vcpu);
1777 return 0; 1759 return 0;
1778 1760
1779out: 1761out:
@@ -2634,8 +2616,9 @@ static int kvm_vm_ioctl_reinject(struct kvm *kvm,
2634int kvm_vm_ioctl_get_dirty_log(struct kvm *kvm, 2616int kvm_vm_ioctl_get_dirty_log(struct kvm *kvm,
2635 struct kvm_dirty_log *log) 2617 struct kvm_dirty_log *log)
2636{ 2618{
2637 int r, n, i; 2619 int r, i;
2638 struct kvm_memory_slot *memslot; 2620 struct kvm_memory_slot *memslot;
2621 unsigned long n;
2639 unsigned long is_dirty = 0; 2622 unsigned long is_dirty = 0;
2640 unsigned long *dirty_bitmap = NULL; 2623 unsigned long *dirty_bitmap = NULL;
2641 2624
@@ -2650,7 +2633,7 @@ int kvm_vm_ioctl_get_dirty_log(struct kvm *kvm,
2650 if (!memslot->dirty_bitmap) 2633 if (!memslot->dirty_bitmap)
2651 goto out; 2634 goto out;
2652 2635
2653 n = ALIGN(memslot->npages, BITS_PER_LONG) / 8; 2636 n = kvm_dirty_bitmap_bytes(memslot);
2654 2637
2655 r = -ENOMEM; 2638 r = -ENOMEM;
2656 dirty_bitmap = vmalloc(n); 2639 dirty_bitmap = vmalloc(n);
@@ -4482,7 +4465,9 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
4482 kvm_set_cr8(vcpu, kvm_run->cr8); 4465 kvm_set_cr8(vcpu, kvm_run->cr8);
4483 4466
4484 if (vcpu->arch.pio.cur_count) { 4467 if (vcpu->arch.pio.cur_count) {
4468 vcpu->srcu_idx = srcu_read_lock(&vcpu->kvm->srcu);
4485 r = complete_pio(vcpu); 4469 r = complete_pio(vcpu);
4470 srcu_read_unlock(&vcpu->kvm->srcu, vcpu->srcu_idx);
4486 if (r) 4471 if (r)
4487 goto out; 4472 goto out;
4488 } 4473 }
@@ -5145,6 +5130,7 @@ int kvm_task_switch(struct kvm_vcpu *vcpu, u16 tss_selector, int reason)
5145 int ret = 0; 5130 int ret = 0;
5146 u32 old_tss_base = get_segment_base(vcpu, VCPU_SREG_TR); 5131 u32 old_tss_base = get_segment_base(vcpu, VCPU_SREG_TR);
5147 u16 old_tss_sel = get_segment_selector(vcpu, VCPU_SREG_TR); 5132 u16 old_tss_sel = get_segment_selector(vcpu, VCPU_SREG_TR);
5133 u32 desc_limit;
5148 5134
5149 old_tss_base = kvm_mmu_gva_to_gpa_write(vcpu, old_tss_base, NULL); 5135 old_tss_base = kvm_mmu_gva_to_gpa_write(vcpu, old_tss_base, NULL);
5150 5136
@@ -5167,7 +5153,10 @@ int kvm_task_switch(struct kvm_vcpu *vcpu, u16 tss_selector, int reason)
5167 } 5153 }
5168 } 5154 }
5169 5155
5170 if (!nseg_desc.p || get_desc_limit(&nseg_desc) < 0x67) { 5156 desc_limit = get_desc_limit(&nseg_desc);
5157 if (!nseg_desc.p ||
5158 ((desc_limit < 0x67 && (nseg_desc.type & 8)) ||
5159 desc_limit < 0x2b)) {
5171 kvm_queue_exception_e(vcpu, TS_VECTOR, tss_selector & 0xfffc); 5160 kvm_queue_exception_e(vcpu, TS_VECTOR, tss_selector & 0xfffc);
5172 return 1; 5161 return 1;
5173 } 5162 }
diff --git a/arch/x86/lguest/boot.c b/arch/x86/lguest/boot.c
index 7e59dc1d3fc2..2bdf628066bd 100644
--- a/arch/x86/lguest/boot.c
+++ b/arch/x86/lguest/boot.c
@@ -115,7 +115,7 @@ static void async_hcall(unsigned long call, unsigned long arg1,
115 local_irq_save(flags); 115 local_irq_save(flags);
116 if (lguest_data.hcall_status[next_call] != 0xFF) { 116 if (lguest_data.hcall_status[next_call] != 0xFF) {
117 /* Table full, so do normal hcall which will flush table. */ 117 /* Table full, so do normal hcall which will flush table. */
118 kvm_hypercall4(call, arg1, arg2, arg3, arg4); 118 hcall(call, arg1, arg2, arg3, arg4);
119 } else { 119 } else {
120 lguest_data.hcalls[next_call].arg0 = call; 120 lguest_data.hcalls[next_call].arg0 = call;
121 lguest_data.hcalls[next_call].arg1 = arg1; 121 lguest_data.hcalls[next_call].arg1 = arg1;
@@ -145,46 +145,45 @@ static void async_hcall(unsigned long call, unsigned long arg1,
145 * So, when we're in lazy mode, we call async_hcall() to store the call for 145 * So, when we're in lazy mode, we call async_hcall() to store the call for
146 * future processing: 146 * future processing:
147 */ 147 */
148static void lazy_hcall1(unsigned long call, 148static void lazy_hcall1(unsigned long call, unsigned long arg1)
149 unsigned long arg1)
150{ 149{
151 if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_NONE) 150 if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_NONE)
152 kvm_hypercall1(call, arg1); 151 hcall(call, arg1, 0, 0, 0);
153 else 152 else
154 async_hcall(call, arg1, 0, 0, 0); 153 async_hcall(call, arg1, 0, 0, 0);
155} 154}
156 155
157/* You can imagine what lazy_hcall2, 3 and 4 look like. :*/ 156/* You can imagine what lazy_hcall2, 3 and 4 look like. :*/
158static void lazy_hcall2(unsigned long call, 157static void lazy_hcall2(unsigned long call,
159 unsigned long arg1, 158 unsigned long arg1,
160 unsigned long arg2) 159 unsigned long arg2)
161{ 160{
162 if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_NONE) 161 if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_NONE)
163 kvm_hypercall2(call, arg1, arg2); 162 hcall(call, arg1, arg2, 0, 0);
164 else 163 else
165 async_hcall(call, arg1, arg2, 0, 0); 164 async_hcall(call, arg1, arg2, 0, 0);
166} 165}
167 166
168static void lazy_hcall3(unsigned long call, 167static void lazy_hcall3(unsigned long call,
169 unsigned long arg1, 168 unsigned long arg1,
170 unsigned long arg2, 169 unsigned long arg2,
171 unsigned long arg3) 170 unsigned long arg3)
172{ 171{
173 if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_NONE) 172 if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_NONE)
174 kvm_hypercall3(call, arg1, arg2, arg3); 173 hcall(call, arg1, arg2, arg3, 0);
175 else 174 else
176 async_hcall(call, arg1, arg2, arg3, 0); 175 async_hcall(call, arg1, arg2, arg3, 0);
177} 176}
178 177
179#ifdef CONFIG_X86_PAE 178#ifdef CONFIG_X86_PAE
180static void lazy_hcall4(unsigned long call, 179static void lazy_hcall4(unsigned long call,
181 unsigned long arg1, 180 unsigned long arg1,
182 unsigned long arg2, 181 unsigned long arg2,
183 unsigned long arg3, 182 unsigned long arg3,
184 unsigned long arg4) 183 unsigned long arg4)
185{ 184{
186 if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_NONE) 185 if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_NONE)
187 kvm_hypercall4(call, arg1, arg2, arg3, arg4); 186 hcall(call, arg1, arg2, arg3, arg4);
188 else 187 else
189 async_hcall(call, arg1, arg2, arg3, arg4); 188 async_hcall(call, arg1, arg2, arg3, arg4);
190} 189}
@@ -196,13 +195,13 @@ static void lazy_hcall4(unsigned long call,
196:*/ 195:*/
197static void lguest_leave_lazy_mmu_mode(void) 196static void lguest_leave_lazy_mmu_mode(void)
198{ 197{
199 kvm_hypercall0(LHCALL_FLUSH_ASYNC); 198 hcall(LHCALL_FLUSH_ASYNC, 0, 0, 0, 0);
200 paravirt_leave_lazy_mmu(); 199 paravirt_leave_lazy_mmu();
201} 200}
202 201
203static void lguest_end_context_switch(struct task_struct *next) 202static void lguest_end_context_switch(struct task_struct *next)
204{ 203{
205 kvm_hypercall0(LHCALL_FLUSH_ASYNC); 204 hcall(LHCALL_FLUSH_ASYNC, 0, 0, 0, 0);
206 paravirt_end_context_switch(next); 205 paravirt_end_context_switch(next);
207} 206}
208 207
@@ -286,7 +285,7 @@ static void lguest_write_idt_entry(gate_desc *dt,
286 /* Keep the local copy up to date. */ 285 /* Keep the local copy up to date. */
287 native_write_idt_entry(dt, entrynum, g); 286 native_write_idt_entry(dt, entrynum, g);
288 /* Tell Host about this new entry. */ 287 /* Tell Host about this new entry. */
289 kvm_hypercall3(LHCALL_LOAD_IDT_ENTRY, entrynum, desc[0], desc[1]); 288 hcall(LHCALL_LOAD_IDT_ENTRY, entrynum, desc[0], desc[1], 0);
290} 289}
291 290
292/* 291/*
@@ -300,7 +299,7 @@ static void lguest_load_idt(const struct desc_ptr *desc)
300 struct desc_struct *idt = (void *)desc->address; 299 struct desc_struct *idt = (void *)desc->address;
301 300
302 for (i = 0; i < (desc->size+1)/8; i++) 301 for (i = 0; i < (desc->size+1)/8; i++)
303 kvm_hypercall3(LHCALL_LOAD_IDT_ENTRY, i, idt[i].a, idt[i].b); 302 hcall(LHCALL_LOAD_IDT_ENTRY, i, idt[i].a, idt[i].b, 0);
304} 303}
305 304
306/* 305/*
@@ -321,7 +320,7 @@ static void lguest_load_gdt(const struct desc_ptr *desc)
321 struct desc_struct *gdt = (void *)desc->address; 320 struct desc_struct *gdt = (void *)desc->address;
322 321
323 for (i = 0; i < (desc->size+1)/8; i++) 322 for (i = 0; i < (desc->size+1)/8; i++)
324 kvm_hypercall3(LHCALL_LOAD_GDT_ENTRY, i, gdt[i].a, gdt[i].b); 323 hcall(LHCALL_LOAD_GDT_ENTRY, i, gdt[i].a, gdt[i].b, 0);
325} 324}
326 325
327/* 326/*
@@ -334,8 +333,8 @@ static void lguest_write_gdt_entry(struct desc_struct *dt, int entrynum,
334{ 333{
335 native_write_gdt_entry(dt, entrynum, desc, type); 334 native_write_gdt_entry(dt, entrynum, desc, type);
336 /* Tell Host about this new entry. */ 335 /* Tell Host about this new entry. */
337 kvm_hypercall3(LHCALL_LOAD_GDT_ENTRY, entrynum, 336 hcall(LHCALL_LOAD_GDT_ENTRY, entrynum,
338 dt[entrynum].a, dt[entrynum].b); 337 dt[entrynum].a, dt[entrynum].b, 0);
339} 338}
340 339
341/* 340/*
@@ -931,7 +930,7 @@ static int lguest_clockevent_set_next_event(unsigned long delta,
931 } 930 }
932 931
933 /* Please wake us this far in the future. */ 932 /* Please wake us this far in the future. */
934 kvm_hypercall1(LHCALL_SET_CLOCKEVENT, delta); 933 hcall(LHCALL_SET_CLOCKEVENT, delta, 0, 0, 0);
935 return 0; 934 return 0;
936} 935}
937 936
@@ -942,7 +941,7 @@ static void lguest_clockevent_set_mode(enum clock_event_mode mode,
942 case CLOCK_EVT_MODE_UNUSED: 941 case CLOCK_EVT_MODE_UNUSED:
943 case CLOCK_EVT_MODE_SHUTDOWN: 942 case CLOCK_EVT_MODE_SHUTDOWN:
944 /* A 0 argument shuts the clock down. */ 943 /* A 0 argument shuts the clock down. */
945 kvm_hypercall0(LHCALL_SET_CLOCKEVENT); 944 hcall(LHCALL_SET_CLOCKEVENT, 0, 0, 0, 0);
946 break; 945 break;
947 case CLOCK_EVT_MODE_ONESHOT: 946 case CLOCK_EVT_MODE_ONESHOT:
948 /* This is what we expect. */ 947 /* This is what we expect. */
@@ -1100,7 +1099,7 @@ static void set_lguest_basic_apic_ops(void)
1100/* STOP! Until an interrupt comes in. */ 1099/* STOP! Until an interrupt comes in. */
1101static void lguest_safe_halt(void) 1100static void lguest_safe_halt(void)
1102{ 1101{
1103 kvm_hypercall0(LHCALL_HALT); 1102 hcall(LHCALL_HALT, 0, 0, 0, 0);
1104} 1103}
1105 1104
1106/* 1105/*
@@ -1112,8 +1111,8 @@ static void lguest_safe_halt(void)
1112 */ 1111 */
1113static void lguest_power_off(void) 1112static void lguest_power_off(void)
1114{ 1113{
1115 kvm_hypercall2(LHCALL_SHUTDOWN, __pa("Power down"), 1114 hcall(LHCALL_SHUTDOWN, __pa("Power down"),
1116 LGUEST_SHUTDOWN_POWEROFF); 1115 LGUEST_SHUTDOWN_POWEROFF, 0, 0);
1117} 1116}
1118 1117
1119/* 1118/*
@@ -1123,7 +1122,7 @@ static void lguest_power_off(void)
1123 */ 1122 */
1124static int lguest_panic(struct notifier_block *nb, unsigned long l, void *p) 1123static int lguest_panic(struct notifier_block *nb, unsigned long l, void *p)
1125{ 1124{
1126 kvm_hypercall2(LHCALL_SHUTDOWN, __pa(p), LGUEST_SHUTDOWN_POWEROFF); 1125 hcall(LHCALL_SHUTDOWN, __pa(p), LGUEST_SHUTDOWN_POWEROFF, 0, 0);
1127 /* The hcall won't return, but to keep gcc happy, we're "done". */ 1126 /* The hcall won't return, but to keep gcc happy, we're "done". */
1128 return NOTIFY_DONE; 1127 return NOTIFY_DONE;
1129} 1128}
@@ -1162,7 +1161,7 @@ static __init int early_put_chars(u32 vtermno, const char *buf, int count)
1162 len = sizeof(scratch) - 1; 1161 len = sizeof(scratch) - 1;
1163 scratch[len] = '\0'; 1162 scratch[len] = '\0';
1164 memcpy(scratch, buf, len); 1163 memcpy(scratch, buf, len);
1165 kvm_hypercall1(LHCALL_NOTIFY, __pa(scratch)); 1164 hcall(LHCALL_NOTIFY, __pa(scratch), 0, 0, 0);
1166 1165
1167 /* This routine returns the number of bytes actually written. */ 1166 /* This routine returns the number of bytes actually written. */
1168 return len; 1167 return len;
@@ -1174,7 +1173,7 @@ static __init int early_put_chars(u32 vtermno, const char *buf, int count)
1174 */ 1173 */
1175static void lguest_restart(char *reason) 1174static void lguest_restart(char *reason)
1176{ 1175{
1177 kvm_hypercall2(LHCALL_SHUTDOWN, __pa(reason), LGUEST_SHUTDOWN_RESTART); 1176 hcall(LHCALL_SHUTDOWN, __pa(reason), LGUEST_SHUTDOWN_RESTART, 0, 0);
1178} 1177}
1179 1178
1180/*G:050 1179/*G:050
diff --git a/arch/x86/lguest/i386_head.S b/arch/x86/lguest/i386_head.S
index 27eac0faee48..4f420c2f2d55 100644
--- a/arch/x86/lguest/i386_head.S
+++ b/arch/x86/lguest/i386_head.S
@@ -32,7 +32,7 @@ ENTRY(lguest_entry)
32 */ 32 */
33 movl $LHCALL_LGUEST_INIT, %eax 33 movl $LHCALL_LGUEST_INIT, %eax
34 movl $lguest_data - __PAGE_OFFSET, %ebx 34 movl $lguest_data - __PAGE_OFFSET, %ebx
35 .byte 0x0f,0x01,0xc1 /* KVM_HYPERCALL */ 35 int $LGUEST_TRAP_ENTRY
36 36
37 /* Set up the initial stack so we can run C code. */ 37 /* Set up the initial stack so we can run C code. */
38 movl $(init_thread_union+THREAD_SIZE),%esp 38 movl $(init_thread_union+THREAD_SIZE),%esp
diff --git a/arch/x86/lib/rwsem_64.S b/arch/x86/lib/rwsem_64.S
index 15acecf0d7aa..41fcf00e49df 100644
--- a/arch/x86/lib/rwsem_64.S
+++ b/arch/x86/lib/rwsem_64.S
@@ -60,7 +60,7 @@ ENTRY(call_rwsem_down_write_failed)
60 ENDPROC(call_rwsem_down_write_failed) 60 ENDPROC(call_rwsem_down_write_failed)
61 61
62ENTRY(call_rwsem_wake) 62ENTRY(call_rwsem_wake)
63 decw %dx /* do nothing if still outstanding active readers */ 63 decl %edx /* do nothing if still outstanding active readers */
64 jnz 1f 64 jnz 1f
65 save_common_regs 65 save_common_regs
66 movq %rax,%rdi 66 movq %rax,%rdi
diff --git a/arch/x86/mm/hugetlbpage.c b/arch/x86/mm/hugetlbpage.c
index f46c340727b8..069ce7c37c01 100644
--- a/arch/x86/mm/hugetlbpage.c
+++ b/arch/x86/mm/hugetlbpage.c
@@ -9,7 +9,6 @@
9#include <linux/mm.h> 9#include <linux/mm.h>
10#include <linux/hugetlb.h> 10#include <linux/hugetlb.h>
11#include <linux/pagemap.h> 11#include <linux/pagemap.h>
12#include <linux/slab.h>
13#include <linux/err.h> 12#include <linux/err.h>
14#include <linux/sysctl.h> 13#include <linux/sysctl.h>
15#include <asm/mman.h> 14#include <asm/mman.h>
diff --git a/arch/x86/mm/init.c b/arch/x86/mm/init.c
index 452ee5b8f309..b278535b14aa 100644
--- a/arch/x86/mm/init.c
+++ b/arch/x86/mm/init.c
@@ -1,3 +1,4 @@
1#include <linux/gfp.h>
1#include <linux/initrd.h> 2#include <linux/initrd.h>
2#include <linux/ioport.h> 3#include <linux/ioport.h>
3#include <linux/swap.h> 4#include <linux/swap.h>
diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c
index 5cb3f0f54f47..bca79091b9d6 100644
--- a/arch/x86/mm/init_32.c
+++ b/arch/x86/mm/init_32.c
@@ -25,11 +25,11 @@
25#include <linux/pfn.h> 25#include <linux/pfn.h>
26#include <linux/poison.h> 26#include <linux/poison.h>
27#include <linux/bootmem.h> 27#include <linux/bootmem.h>
28#include <linux/slab.h>
29#include <linux/proc_fs.h> 28#include <linux/proc_fs.h>
30#include <linux/memory_hotplug.h> 29#include <linux/memory_hotplug.h>
31#include <linux/initrd.h> 30#include <linux/initrd.h>
32#include <linux/cpumask.h> 31#include <linux/cpumask.h>
32#include <linux/gfp.h>
33 33
34#include <asm/asm.h> 34#include <asm/asm.h>
35#include <asm/bios_ebda.h> 35#include <asm/bios_ebda.h>
diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c
index e9b040e1cde5..ee41bba315d1 100644
--- a/arch/x86/mm/init_64.c
+++ b/arch/x86/mm/init_64.c
@@ -29,6 +29,7 @@
29#include <linux/module.h> 29#include <linux/module.h>
30#include <linux/memory_hotplug.h> 30#include <linux/memory_hotplug.h>
31#include <linux/nmi.h> 31#include <linux/nmi.h>
32#include <linux/gfp.h>
32 33
33#include <asm/processor.h> 34#include <asm/processor.h>
34#include <asm/bios_ebda.h> 35#include <asm/bios_ebda.h>
diff --git a/arch/x86/mm/ioremap.c b/arch/x86/mm/ioremap.c
index 5eb1ba74a3a9..12e4d2d3c110 100644
--- a/arch/x86/mm/ioremap.c
+++ b/arch/x86/mm/ioremap.c
@@ -448,6 +448,20 @@ static inline void __init early_clear_fixmap(enum fixed_addresses idx)
448static void __iomem *prev_map[FIX_BTMAPS_SLOTS] __initdata; 448static void __iomem *prev_map[FIX_BTMAPS_SLOTS] __initdata;
449static unsigned long prev_size[FIX_BTMAPS_SLOTS] __initdata; 449static unsigned long prev_size[FIX_BTMAPS_SLOTS] __initdata;
450 450
451void __init fixup_early_ioremap(void)
452{
453 int i;
454
455 for (i = 0; i < FIX_BTMAPS_SLOTS; i++) {
456 if (prev_map[i]) {
457 WARN_ON(1);
458 break;
459 }
460 }
461
462 early_ioremap_init();
463}
464
451static int __init check_early_ioremap_leak(void) 465static int __init check_early_ioremap_leak(void)
452{ 466{
453 int count = 0; 467 int count = 0;
diff --git a/arch/x86/mm/kmmio.c b/arch/x86/mm/kmmio.c
index 536fb6823366..5d0e67fff1a6 100644
--- a/arch/x86/mm/kmmio.c
+++ b/arch/x86/mm/kmmio.c
@@ -21,6 +21,7 @@
21#include <linux/kdebug.h> 21#include <linux/kdebug.h>
22#include <linux/mutex.h> 22#include <linux/mutex.h>
23#include <linux/io.h> 23#include <linux/io.h>
24#include <linux/slab.h>
24#include <asm/cacheflush.h> 25#include <asm/cacheflush.h>
25#include <asm/tlbflush.h> 26#include <asm/tlbflush.h>
26#include <linux/errno.h> 27#include <linux/errno.h>
diff --git a/arch/x86/mm/mmio-mod.c b/arch/x86/mm/mmio-mod.c
index 34a3291ca103..3adff7dcc148 100644
--- a/arch/x86/mm/mmio-mod.c
+++ b/arch/x86/mm/mmio-mod.c
@@ -26,6 +26,7 @@
26 26
27#include <linux/module.h> 27#include <linux/module.h>
28#include <linux/debugfs.h> 28#include <linux/debugfs.h>
29#include <linux/slab.h>
29#include <linux/uaccess.h> 30#include <linux/uaccess.h>
30#include <linux/io.h> 31#include <linux/io.h>
31#include <linux/version.h> 32#include <linux/version.h>
diff --git a/arch/x86/mm/pageattr.c b/arch/x86/mm/pageattr.c
index cf07c26d9a4a..28195c350b97 100644
--- a/arch/x86/mm/pageattr.c
+++ b/arch/x86/mm/pageattr.c
@@ -6,13 +6,13 @@
6#include <linux/bootmem.h> 6#include <linux/bootmem.h>
7#include <linux/module.h> 7#include <linux/module.h>
8#include <linux/sched.h> 8#include <linux/sched.h>
9#include <linux/slab.h>
10#include <linux/mm.h> 9#include <linux/mm.h>
11#include <linux/interrupt.h> 10#include <linux/interrupt.h>
12#include <linux/seq_file.h> 11#include <linux/seq_file.h>
13#include <linux/debugfs.h> 12#include <linux/debugfs.h>
14#include <linux/pfn.h> 13#include <linux/pfn.h>
15#include <linux/percpu.h> 14#include <linux/percpu.h>
15#include <linux/gfp.h>
16 16
17#include <asm/e820.h> 17#include <asm/e820.h>
18#include <asm/processor.h> 18#include <asm/processor.h>
diff --git a/arch/x86/mm/pat.c b/arch/x86/mm/pat.c
index ae9648eb1c7f..edc8b95afc1a 100644
--- a/arch/x86/mm/pat.c
+++ b/arch/x86/mm/pat.c
@@ -12,7 +12,7 @@
12#include <linux/debugfs.h> 12#include <linux/debugfs.h>
13#include <linux/kernel.h> 13#include <linux/kernel.h>
14#include <linux/module.h> 14#include <linux/module.h>
15#include <linux/gfp.h> 15#include <linux/slab.h>
16#include <linux/mm.h> 16#include <linux/mm.h>
17#include <linux/fs.h> 17#include <linux/fs.h>
18#include <linux/rbtree.h> 18#include <linux/rbtree.h>
diff --git a/arch/x86/mm/pgtable.c b/arch/x86/mm/pgtable.c
index c9ba9deafe83..5c4ee422590e 100644
--- a/arch/x86/mm/pgtable.c
+++ b/arch/x86/mm/pgtable.c
@@ -1,4 +1,5 @@
1#include <linux/mm.h> 1#include <linux/mm.h>
2#include <linux/gfp.h>
2#include <asm/pgalloc.h> 3#include <asm/pgalloc.h>
3#include <asm/pgtable.h> 4#include <asm/pgtable.h>
4#include <asm/tlb.h> 5#include <asm/tlb.h>
diff --git a/arch/x86/mm/pgtable_32.c b/arch/x86/mm/pgtable_32.c
index 46c8834aedc0..792854003ed3 100644
--- a/arch/x86/mm/pgtable_32.c
+++ b/arch/x86/mm/pgtable_32.c
@@ -6,7 +6,6 @@
6#include <linux/swap.h> 6#include <linux/swap.h>
7#include <linux/smp.h> 7#include <linux/smp.h>
8#include <linux/highmem.h> 8#include <linux/highmem.h>
9#include <linux/slab.h>
10#include <linux/pagemap.h> 9#include <linux/pagemap.h>
11#include <linux/spinlock.h> 10#include <linux/spinlock.h>
12#include <linux/module.h> 11#include <linux/module.h>
@@ -19,6 +18,7 @@
19#include <asm/e820.h> 18#include <asm/e820.h>
20#include <asm/tlb.h> 19#include <asm/tlb.h>
21#include <asm/tlbflush.h> 20#include <asm/tlbflush.h>
21#include <asm/io.h>
22 22
23unsigned int __VMALLOC_RESERVE = 128 << 20; 23unsigned int __VMALLOC_RESERVE = 128 << 20;
24 24
@@ -129,6 +129,7 @@ static int __init parse_reservetop(char *arg)
129 129
130 address = memparse(arg, &arg); 130 address = memparse(arg, &arg);
131 reserve_top_address(address); 131 reserve_top_address(address);
132 fixup_early_ioremap();
132 return 0; 133 return 0;
133} 134}
134early_param("reservetop", parse_reservetop); 135early_param("reservetop", parse_reservetop);
diff --git a/arch/x86/mm/srat_64.c b/arch/x86/mm/srat_64.c
index 28c68762648f..38512d0c4742 100644
--- a/arch/x86/mm/srat_64.c
+++ b/arch/x86/mm/srat_64.c
@@ -461,7 +461,8 @@ void __init acpi_fake_nodes(const struct bootnode *fake_nodes, int num_nodes)
461 * node, it must now point to the fake node ID. 461 * node, it must now point to the fake node ID.
462 */ 462 */
463 for (j = 0; j < MAX_LOCAL_APIC; j++) 463 for (j = 0; j < MAX_LOCAL_APIC; j++)
464 if (apicid_to_node[j] == nid) 464 if (apicid_to_node[j] == nid &&
465 fake_apicid_to_node[j] == NUMA_NO_NODE)
465 fake_apicid_to_node[j] = i; 466 fake_apicid_to_node[j] = i;
466 } 467 }
467 for (i = 0; i < num_nodes; i++) 468 for (i = 0; i < num_nodes; i++)
diff --git a/arch/x86/pci/acpi.c b/arch/x86/pci/acpi.c
index 0b7882dbe784..9dcf43d7d0c0 100644
--- a/arch/x86/pci/acpi.c
+++ b/arch/x86/pci/acpi.c
@@ -3,6 +3,7 @@
3#include <linux/init.h> 3#include <linux/init.h>
4#include <linux/irq.h> 4#include <linux/irq.h>
5#include <linux/dmi.h> 5#include <linux/dmi.h>
6#include <linux/slab.h>
6#include <asm/numa.h> 7#include <asm/numa.h>
7#include <asm/pci_x86.h> 8#include <asm/pci_x86.h>
8 9
@@ -65,14 +66,44 @@ resource_to_addr(struct acpi_resource *resource,
65 struct acpi_resource_address64 *addr) 66 struct acpi_resource_address64 *addr)
66{ 67{
67 acpi_status status; 68 acpi_status status;
68 69 struct acpi_resource_memory24 *memory24;
69 status = acpi_resource_to_address64(resource, addr); 70 struct acpi_resource_memory32 *memory32;
70 if (ACPI_SUCCESS(status) && 71 struct acpi_resource_fixed_memory32 *fixed_memory32;
71 (addr->resource_type == ACPI_MEMORY_RANGE || 72
72 addr->resource_type == ACPI_IO_RANGE) && 73 memset(addr, 0, sizeof(*addr));
73 addr->address_length > 0 && 74 switch (resource->type) {
74 addr->producer_consumer == ACPI_PRODUCER) { 75 case ACPI_RESOURCE_TYPE_MEMORY24:
76 memory24 = &resource->data.memory24;
77 addr->resource_type = ACPI_MEMORY_RANGE;
78 addr->minimum = memory24->minimum;
79 addr->address_length = memory24->address_length;
80 addr->maximum = addr->minimum + addr->address_length - 1;
81 return AE_OK;
82 case ACPI_RESOURCE_TYPE_MEMORY32:
83 memory32 = &resource->data.memory32;
84 addr->resource_type = ACPI_MEMORY_RANGE;
85 addr->minimum = memory32->minimum;
86 addr->address_length = memory32->address_length;
87 addr->maximum = addr->minimum + addr->address_length - 1;
75 return AE_OK; 88 return AE_OK;
89 case ACPI_RESOURCE_TYPE_FIXED_MEMORY32:
90 fixed_memory32 = &resource->data.fixed_memory32;
91 addr->resource_type = ACPI_MEMORY_RANGE;
92 addr->minimum = fixed_memory32->address;
93 addr->address_length = fixed_memory32->address_length;
94 addr->maximum = addr->minimum + addr->address_length - 1;
95 return AE_OK;
96 case ACPI_RESOURCE_TYPE_ADDRESS16:
97 case ACPI_RESOURCE_TYPE_ADDRESS32:
98 case ACPI_RESOURCE_TYPE_ADDRESS64:
99 status = acpi_resource_to_address64(resource, addr);
100 if (ACPI_SUCCESS(status) &&
101 (addr->resource_type == ACPI_MEMORY_RANGE ||
102 addr->resource_type == ACPI_IO_RANGE) &&
103 addr->address_length > 0) {
104 return AE_OK;
105 }
106 break;
76 } 107 }
77 return AE_ERROR; 108 return AE_ERROR;
78} 109}
@@ -90,30 +121,6 @@ count_resource(struct acpi_resource *acpi_res, void *data)
90 return AE_OK; 121 return AE_OK;
91} 122}
92 123
93static void
94align_resource(struct acpi_device *bridge, struct resource *res)
95{
96 int align = (res->flags & IORESOURCE_MEM) ? 16 : 4;
97
98 /*
99 * Host bridge windows are not BARs, but the decoders on the PCI side
100 * that claim this address space have starting alignment and length
101 * constraints, so fix any obvious BIOS goofs.
102 */
103 if (!IS_ALIGNED(res->start, align)) {
104 dev_printk(KERN_DEBUG, &bridge->dev,
105 "host bridge window %pR invalid; "
106 "aligning start to %d-byte boundary\n", res, align);
107 res->start &= ~(align - 1);
108 }
109 if (!IS_ALIGNED(res->end + 1, align)) {
110 dev_printk(KERN_DEBUG, &bridge->dev,
111 "host bridge window %pR invalid; "
112 "aligning end to %d-byte boundary\n", res, align);
113 res->end = ALIGN(res->end, align) - 1;
114 }
115}
116
117static acpi_status 124static acpi_status
118setup_resource(struct acpi_resource *acpi_res, void *data) 125setup_resource(struct acpi_resource *acpi_res, void *data)
119{ 126{
@@ -123,7 +130,7 @@ setup_resource(struct acpi_resource *acpi_res, void *data)
123 acpi_status status; 130 acpi_status status;
124 unsigned long flags; 131 unsigned long flags;
125 struct resource *root, *conflict; 132 struct resource *root, *conflict;
126 u64 start, end, max_len; 133 u64 start, end;
127 134
128 status = resource_to_addr(acpi_res, &addr); 135 status = resource_to_addr(acpi_res, &addr);
129 if (!ACPI_SUCCESS(status)) 136 if (!ACPI_SUCCESS(status))
@@ -140,19 +147,8 @@ setup_resource(struct acpi_resource *acpi_res, void *data)
140 } else 147 } else
141 return AE_OK; 148 return AE_OK;
142 149
143 max_len = addr.maximum - addr.minimum + 1;
144 if (addr.address_length > max_len) {
145 dev_printk(KERN_DEBUG, &info->bridge->dev,
146 "host bridge window length %#llx doesn't fit in "
147 "%#llx-%#llx, trimming\n",
148 (unsigned long long) addr.address_length,
149 (unsigned long long) addr.minimum,
150 (unsigned long long) addr.maximum);
151 addr.address_length = max_len;
152 }
153
154 start = addr.minimum + addr.translation_offset; 150 start = addr.minimum + addr.translation_offset;
155 end = start + addr.address_length - 1; 151 end = addr.maximum + addr.translation_offset;
156 152
157 res = &info->res[info->res_num]; 153 res = &info->res[info->res_num];
158 res->name = info->name; 154 res->name = info->name;
@@ -160,7 +156,6 @@ setup_resource(struct acpi_resource *acpi_res, void *data)
160 res->start = start; 156 res->start = start;
161 res->end = end; 157 res->end = end;
162 res->child = NULL; 158 res->child = NULL;
163 align_resource(info->bridge, res);
164 159
165 if (!pci_use_crs) { 160 if (!pci_use_crs) {
166 dev_printk(KERN_DEBUG, &info->bridge->dev, 161 dev_printk(KERN_DEBUG, &info->bridge->dev,
diff --git a/arch/x86/pci/common.c b/arch/x86/pci/common.c
index 294e10cb11e1..cf2e93869c48 100644
--- a/arch/x86/pci/common.c
+++ b/arch/x86/pci/common.c
@@ -9,6 +9,7 @@
9#include <linux/ioport.h> 9#include <linux/ioport.h>
10#include <linux/init.h> 10#include <linux/init.h>
11#include <linux/dmi.h> 11#include <linux/dmi.h>
12#include <linux/slab.h>
12 13
13#include <asm/acpi.h> 14#include <asm/acpi.h>
14#include <asm/segment.h> 15#include <asm/segment.h>
diff --git a/arch/x86/pci/i386.c b/arch/x86/pci/i386.c
index 46fd43f79103..97da2ba9344b 100644
--- a/arch/x86/pci/i386.c
+++ b/arch/x86/pci/i386.c
@@ -72,6 +72,9 @@ pcibios_align_resource(void *data, const struct resource *res,
72 return start; 72 return start;
73 if (start & 0x300) 73 if (start & 0x300)
74 start = (start + 0x3ff) & ~0x3ff; 74 start = (start + 0x3ff) & ~0x3ff;
75 } else if (res->flags & IORESOURCE_MEM) {
76 if (start < BIOS_END)
77 start = BIOS_END;
75 } 78 }
76 return start; 79 return start;
77} 80}
diff --git a/arch/x86/pci/irq.c b/arch/x86/pci/irq.c
index 8b107521d24e..5d362b5ba06f 100644
--- a/arch/x86/pci/irq.c
+++ b/arch/x86/pci/irq.c
@@ -8,7 +8,6 @@
8#include <linux/kernel.h> 8#include <linux/kernel.h>
9#include <linux/pci.h> 9#include <linux/pci.h>
10#include <linux/init.h> 10#include <linux/init.h>
11#include <linux/slab.h>
12#include <linux/interrupt.h> 11#include <linux/interrupt.h>
13#include <linux/dmi.h> 12#include <linux/dmi.h>
14#include <linux/io.h> 13#include <linux/io.h>
diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c
index 8f3f9a50b1e0..39b9ebe8f886 100644
--- a/arch/x86/pci/mmconfig-shared.c
+++ b/arch/x86/pci/mmconfig-shared.c
@@ -16,6 +16,7 @@
16#include <linux/sfi_acpi.h> 16#include <linux/sfi_acpi.h>
17#include <linux/bitmap.h> 17#include <linux/bitmap.h>
18#include <linux/dmi.h> 18#include <linux/dmi.h>
19#include <linux/slab.h>
19#include <asm/e820.h> 20#include <asm/e820.h>
20#include <asm/pci_x86.h> 21#include <asm/pci_x86.h>
21#include <asm/acpi.h> 22#include <asm/acpi.h>
diff --git a/arch/x86/pci/mrst.c b/arch/x86/pci/mrst.c
index 8bf2fcb88d04..1cdc02cf8fa4 100644
--- a/arch/x86/pci/mrst.c
+++ b/arch/x86/pci/mrst.c
@@ -247,6 +247,10 @@ static void __devinit pci_fixed_bar_fixup(struct pci_dev *dev)
247 u32 size; 247 u32 size;
248 int i; 248 int i;
249 249
250 /* Must have extended configuration space */
251 if (dev->cfg_size < PCIE_CAP_OFFSET + 4)
252 return;
253
250 /* Fixup the BAR sizes for fixed BAR devices and make them unmoveable */ 254 /* Fixup the BAR sizes for fixed BAR devices and make them unmoveable */
251 offset = fixed_bar_cap(dev->bus, dev->devfn); 255 offset = fixed_bar_cap(dev->bus, dev->devfn);
252 if (!offset || PCI_DEVFN(2, 0) == dev->devfn || 256 if (!offset || PCI_DEVFN(2, 0) == dev->devfn ||
diff --git a/arch/x86/pci/pcbios.c b/arch/x86/pci/pcbios.c
index 1c975cc9839e..59a225c17b84 100644
--- a/arch/x86/pci/pcbios.c
+++ b/arch/x86/pci/pcbios.c
@@ -4,6 +4,7 @@
4 4
5#include <linux/pci.h> 5#include <linux/pci.h>
6#include <linux/init.h> 6#include <linux/init.h>
7#include <linux/slab.h>
7#include <linux/module.h> 8#include <linux/module.h>
8#include <linux/uaccess.h> 9#include <linux/uaccess.h>
9#include <asm/pci_x86.h> 10#include <asm/pci_x86.h>
diff --git a/arch/x86/power/hibernate_32.c b/arch/x86/power/hibernate_32.c
index 81197c62d5b3..3769079874d8 100644
--- a/arch/x86/power/hibernate_32.c
+++ b/arch/x86/power/hibernate_32.c
@@ -6,6 +6,7 @@
6 * Copyright (c) 2006 Rafael J. Wysocki <rjw@sisk.pl> 6 * Copyright (c) 2006 Rafael J. Wysocki <rjw@sisk.pl>
7 */ 7 */
8 8
9#include <linux/gfp.h>
9#include <linux/suspend.h> 10#include <linux/suspend.h>
10#include <linux/bootmem.h> 11#include <linux/bootmem.h>
11 12
diff --git a/arch/x86/power/hibernate_64.c b/arch/x86/power/hibernate_64.c
index 65fdc86e923f..d24f983ba1e5 100644
--- a/arch/x86/power/hibernate_64.c
+++ b/arch/x86/power/hibernate_64.c
@@ -8,6 +8,7 @@
8 * Copyright (c) 2001 Patrick Mochel <mochel@osdl.org> 8 * Copyright (c) 2001 Patrick Mochel <mochel@osdl.org>
9 */ 9 */
10 10
11#include <linux/gfp.h>
11#include <linux/smp.h> 12#include <linux/smp.h>
12#include <linux/suspend.h> 13#include <linux/suspend.h>
13#include <asm/proto.h> 14#include <asm/proto.h>
diff --git a/arch/x86/power/hibernate_asm_32.S b/arch/x86/power/hibernate_asm_32.S
index b641388d8286..ad47daeafa4e 100644
--- a/arch/x86/power/hibernate_asm_32.S
+++ b/arch/x86/power/hibernate_asm_32.S
@@ -27,10 +27,17 @@ ENTRY(swsusp_arch_suspend)
27 ret 27 ret
28 28
29ENTRY(restore_image) 29ENTRY(restore_image)
30 movl mmu_cr4_features, %ecx
30 movl resume_pg_dir, %eax 31 movl resume_pg_dir, %eax
31 subl $__PAGE_OFFSET, %eax 32 subl $__PAGE_OFFSET, %eax
32 movl %eax, %cr3 33 movl %eax, %cr3
33 34
35 jecxz 1f # cr4 Pentium and higher, skip if zero
36 andl $~(X86_CR4_PGE), %ecx
37 movl %ecx, %cr4; # turn off PGE
38 movl %cr3, %eax; # flush TLB
39 movl %eax, %cr3
401:
34 movl restore_pblist, %edx 41 movl restore_pblist, %edx
35 .p2align 4,,7 42 .p2align 4,,7
36 43
@@ -54,16 +61,8 @@ done:
54 movl $swapper_pg_dir, %eax 61 movl $swapper_pg_dir, %eax
55 subl $__PAGE_OFFSET, %eax 62 subl $__PAGE_OFFSET, %eax
56 movl %eax, %cr3 63 movl %eax, %cr3
57 /* Flush TLB, including "global" things (vmalloc) */
58 movl mmu_cr4_features, %ecx 64 movl mmu_cr4_features, %ecx
59 jecxz 1f # cr4 Pentium and higher, skip if zero 65 jecxz 1f # cr4 Pentium and higher, skip if zero
60 movl %ecx, %edx
61 andl $~(X86_CR4_PGE), %edx
62 movl %edx, %cr4; # turn off PGE
631:
64 movl %cr3, %eax; # flush TLB
65 movl %eax, %cr3
66 jecxz 1f # cr4 Pentium and higher, skip if zero
67 movl %ecx, %cr4; # turn PGE back on 66 movl %ecx, %cr4; # turn PGE back on
681: 671:
69 68
diff --git a/arch/x86/vdso/vma.c b/arch/x86/vdso/vma.c
index 21e1aeb9f3ea..ac74869b8140 100644
--- a/arch/x86/vdso/vma.c
+++ b/arch/x86/vdso/vma.c
@@ -6,6 +6,7 @@
6#include <linux/mm.h> 6#include <linux/mm.h>
7#include <linux/err.h> 7#include <linux/err.h>
8#include <linux/sched.h> 8#include <linux/sched.h>
9#include <linux/slab.h>
9#include <linux/init.h> 10#include <linux/init.h>
10#include <linux/random.h> 11#include <linux/random.h>
11#include <linux/elf.h> 12#include <linux/elf.h>
diff --git a/arch/x86/xen/debugfs.c b/arch/x86/xen/debugfs.c
index e133ce25e290..1304bcec8ee5 100644
--- a/arch/x86/xen/debugfs.c
+++ b/arch/x86/xen/debugfs.c
@@ -1,5 +1,6 @@
1#include <linux/init.h> 1#include <linux/init.h>
2#include <linux/debugfs.h> 2#include <linux/debugfs.h>
3#include <linux/slab.h>
3#include <linux/module.h> 4#include <linux/module.h>
4 5
5#include "debugfs.h" 6#include "debugfs.h"
diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c
index b607239c1ba8..65d8d79b46a8 100644
--- a/arch/x86/xen/enlighten.c
+++ b/arch/x86/xen/enlighten.c
@@ -28,6 +28,7 @@
28#include <linux/highmem.h> 28#include <linux/highmem.h>
29#include <linux/console.h> 29#include <linux/console.h>
30#include <linux/pci.h> 30#include <linux/pci.h>
31#include <linux/gfp.h>
31 32
32#include <xen/xen.h> 33#include <xen/xen.h>
33#include <xen/interface/xen.h> 34#include <xen/interface/xen.h>
diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c
index f9eb7de74f42..914f04695ce5 100644
--- a/arch/x86/xen/mmu.c
+++ b/arch/x86/xen/mmu.c
@@ -43,6 +43,7 @@
43#include <linux/debugfs.h> 43#include <linux/debugfs.h>
44#include <linux/bug.h> 44#include <linux/bug.h>
45#include <linux/module.h> 45#include <linux/module.h>
46#include <linux/gfp.h>
46 47
47#include <asm/pgtable.h> 48#include <asm/pgtable.h>
48#include <asm/tlbflush.h> 49#include <asm/tlbflush.h>
diff --git a/arch/x86/xen/smp.c b/arch/x86/xen/smp.c
index deafb65ef44e..a29693fd3138 100644
--- a/arch/x86/xen/smp.c
+++ b/arch/x86/xen/smp.c
@@ -14,6 +14,7 @@
14 */ 14 */
15#include <linux/sched.h> 15#include <linux/sched.h>
16#include <linux/err.h> 16#include <linux/err.h>
17#include <linux/slab.h>
17#include <linux/smp.h> 18#include <linux/smp.h>
18 19
19#include <asm/paravirt.h> 20#include <asm/paravirt.h>
diff --git a/arch/x86/xen/spinlock.c b/arch/x86/xen/spinlock.c
index 24ded31b5aec..e0500646585d 100644
--- a/arch/x86/xen/spinlock.c
+++ b/arch/x86/xen/spinlock.c
@@ -6,6 +6,7 @@
6#include <linux/spinlock.h> 6#include <linux/spinlock.h>
7#include <linux/debugfs.h> 7#include <linux/debugfs.h>
8#include <linux/log2.h> 8#include <linux/log2.h>
9#include <linux/gfp.h>
9 10
10#include <asm/paravirt.h> 11#include <asm/paravirt.h>
11 12
diff --git a/arch/x86/xen/time.c b/arch/x86/xen/time.c
index 0d3f07cd1b5f..32764b8880b5 100644
--- a/arch/x86/xen/time.c
+++ b/arch/x86/xen/time.c
@@ -13,6 +13,7 @@
13#include <linux/clockchips.h> 13#include <linux/clockchips.h>
14#include <linux/kernel_stat.h> 14#include <linux/kernel_stat.h>
15#include <linux/math64.h> 15#include <linux/math64.h>
16#include <linux/gfp.h>
16 17
17#include <asm/pvclock.h> 18#include <asm/pvclock.h>
18#include <asm/xen/hypervisor.h> 19#include <asm/xen/hypervisor.h>