aboutsummaryrefslogtreecommitdiffstats
path: root/arch/x86
diff options
context:
space:
mode:
authorIngo Molnar <mingo@elte.hu>2009-10-17 03:58:25 -0400
committerIngo Molnar <mingo@elte.hu>2009-10-17 03:58:25 -0400
commitbb3c3e807140816b5f5fd4840473ee52a916ad4f (patch)
tree9e8a69d266a7df86ca16177eefffab4b4e910753 /arch/x86
parent595c36490deb49381dc51231a3d5e6b66786ed27 (diff)
parent012abeea669ea49636cf952d13298bb68654146a (diff)
Merge commit 'v2.6.32-rc5' into perf/probes
Conflicts: kernel/trace/trace_event_profile.c Merge reason: update to -rc5 and resolve conflict. Signed-off-by: Ingo Molnar <mingo@elte.hu>
Diffstat (limited to 'arch/x86')
-rw-r--r--arch/x86/Kconfig26
-rw-r--r--arch/x86/Kconfig.cpu3
-rw-r--r--arch/x86/Makefile4
-rw-r--r--arch/x86/boot/compressed/head_32.S3
-rw-r--r--arch/x86/boot/compressed/head_64.S3
-rw-r--r--arch/x86/boot/compressed/vmlinux.lds.S6
-rw-r--r--arch/x86/boot/install.sh4
-rw-r--r--arch/x86/ia32/ia32entry.S36
-rw-r--r--arch/x86/include/asm/cache.h4
-rw-r--r--arch/x86/include/asm/checksum_32.h3
-rw-r--r--arch/x86/include/asm/cmpxchg_32.h30
-rw-r--r--arch/x86/include/asm/kvm_host.h1
-rw-r--r--arch/x86/include/asm/mce.h2
-rw-r--r--arch/x86/include/asm/mmu_context.h6
-rw-r--r--arch/x86/include/asm/nmi.h3
-rw-r--r--arch/x86/include/asm/paravirt.h28
-rw-r--r--arch/x86/include/asm/paravirt_types.h10
-rw-r--r--arch/x86/include/asm/pci.h6
-rw-r--r--arch/x86/include/asm/pgtable_types.h1
-rw-r--r--arch/x86/include/asm/smp.h1
-rw-r--r--arch/x86/include/asm/topology.h10
-rw-r--r--arch/x86/kernel/acpi/cstate.c2
-rw-r--r--arch/x86/kernel/apic/io_apic.c7
-rw-r--r--arch/x86/kernel/apic/nmi.c4
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce-inject.c7
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce.c84
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce_intel.c1
-rw-r--r--arch/x86/kernel/cpu/mcheck/therm_throt.c67
-rw-r--r--arch/x86/kernel/cpu/mtrr/if.c17
-rw-r--r--arch/x86/kernel/cpu/perf_event.c3
-rw-r--r--arch/x86/kernel/dumpstack_32.c1
-rw-r--r--arch/x86/kernel/dumpstack_64.c1
-rw-r--r--arch/x86/kernel/e820.c4
-rw-r--r--arch/x86/kernel/early_printk.c10
-rw-r--r--arch/x86/kernel/head_32.S6
-rw-r--r--arch/x86/kernel/head_64.S4
-rw-r--r--arch/x86/kernel/i386_ksyms_32.c10
-rw-r--r--arch/x86/kernel/init_task.c5
-rw-r--r--arch/x86/kernel/irq.c4
-rw-r--r--arch/x86/kernel/ldt.c4
-rw-r--r--arch/x86/kernel/pci-dma.c4
-rw-r--r--arch/x86/kernel/pci-gart_64.c1
-rw-r--r--arch/x86/kernel/process.c6
-rw-r--r--arch/x86/kernel/reboot.c1
-rw-r--r--arch/x86/kernel/smpboot.c9
-rw-r--r--arch/x86/kernel/time.c4
-rw-r--r--arch/x86/kernel/trampoline.c12
-rw-r--r--arch/x86/kernel/trampoline_64.S4
-rw-r--r--arch/x86/kernel/traps.c7
-rw-r--r--arch/x86/kernel/tsc_sync.c2
-rw-r--r--arch/x86/kernel/vmi_32.c2
-rw-r--r--arch/x86/kernel/vmlinux.lds.S79
-rw-r--r--arch/x86/kernel/vsyscall_64.c10
-rw-r--r--arch/x86/kvm/lapic.c2
-rw-r--r--arch/x86/kvm/mmu.c84
-rw-r--r--arch/x86/kvm/paging_tmpl.h18
-rw-r--r--arch/x86/kvm/svm.c25
-rw-r--r--arch/x86/kvm/vmx.c2
-rw-r--r--arch/x86/kvm/x86.c2
-rw-r--r--arch/x86/lib/Makefile4
-rw-r--r--arch/x86/lib/cmpxchg8b_emu.S57
-rw-r--r--arch/x86/mm/Makefile3
-rw-r--r--arch/x86/mm/fault.c19
-rw-r--r--arch/x86/mm/init.c63
-rw-r--r--arch/x86/mm/pageattr.c1
-rw-r--r--arch/x86/mm/pat.c7
-rw-r--r--arch/x86/mm/setup_nx.c69
-rw-r--r--arch/x86/mm/tlb.c15
-rw-r--r--arch/x86/pci/common.c2
-rw-r--r--arch/x86/pci/i386.c2
-rw-r--r--arch/x86/vdso/Makefile2
-rw-r--r--arch/x86/xen/debugfs.c2
-rw-r--r--arch/x86/xen/enlighten.c10
-rw-r--r--arch/x86/xen/mmu.c4
74 files changed, 577 insertions, 388 deletions
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 93698794aa3a..07e01149e3bf 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -86,10 +86,6 @@ config STACKTRACE_SUPPORT
86config HAVE_LATENCYTOP_SUPPORT 86config HAVE_LATENCYTOP_SUPPORT
87 def_bool y 87 def_bool y
88 88
89config FAST_CMPXCHG_LOCAL
90 bool
91 default y
92
93config MMU 89config MMU
94 def_bool y 90 def_bool y
95 91
@@ -432,6 +428,17 @@ config X86_NUMAQ
432 of Flat Logical. You will need a new lynxer.elf file to flash your 428 of Flat Logical. You will need a new lynxer.elf file to flash your
433 firmware with - send email to <Martin.Bligh@us.ibm.com>. 429 firmware with - send email to <Martin.Bligh@us.ibm.com>.
434 430
431config X86_SUPPORTS_MEMORY_FAILURE
432 bool
433 # MCE code calls memory_failure():
434 depends on X86_MCE
435 # On 32-bit this adds too big of NODES_SHIFT and we run out of page flags:
436 depends on !X86_NUMAQ
437 # On 32-bit SPARSEMEM adds too big of SECTIONS_WIDTH:
438 depends on X86_64 || !SPARSEMEM
439 select ARCH_SUPPORTS_MEMORY_FAILURE
440 default y
441
435config X86_VISWS 442config X86_VISWS
436 bool "SGI 320/540 (Visual Workstation)" 443 bool "SGI 320/540 (Visual Workstation)"
437 depends on X86_32 && PCI && X86_MPPARSE && PCI_GODIRECT 444 depends on X86_32 && PCI && X86_MPPARSE && PCI_GODIRECT
@@ -484,7 +491,7 @@ if PARAVIRT_GUEST
484source "arch/x86/xen/Kconfig" 491source "arch/x86/xen/Kconfig"
485 492
486config VMI 493config VMI
487 bool "VMI Guest support" 494 bool "VMI Guest support (DEPRECATED)"
488 select PARAVIRT 495 select PARAVIRT
489 depends on X86_32 496 depends on X86_32
490 ---help--- 497 ---help---
@@ -493,6 +500,15 @@ config VMI
493 at the moment), by linking the kernel to a GPL-ed ROM module 500 at the moment), by linking the kernel to a GPL-ed ROM module
494 provided by the hypervisor. 501 provided by the hypervisor.
495 502
503 As of September 2009, VMware has started a phased retirement
504 of this feature from VMware's products. Please see
505 feature-removal-schedule.txt for details. If you are
506 planning to enable this option, please note that you cannot
507 live migrate a VMI enabled VM to a future VMware product,
508 which doesn't support VMI. So if you expect your kernel to
509 seamlessly migrate to newer VMware products, keep this
510 disabled.
511
496config KVM_CLOCK 512config KVM_CLOCK
497 bool "KVM paravirtualized clock" 513 bool "KVM paravirtualized clock"
498 select PARAVIRT 514 select PARAVIRT
diff --git a/arch/x86/Kconfig.cpu b/arch/x86/Kconfig.cpu
index 527519b8a9f9..f2824fb8c79c 100644
--- a/arch/x86/Kconfig.cpu
+++ b/arch/x86/Kconfig.cpu
@@ -400,7 +400,7 @@ config X86_TSC
400 400
401config X86_CMPXCHG64 401config X86_CMPXCHG64
402 def_bool y 402 def_bool y
403 depends on X86_PAE || X86_64 403 depends on X86_PAE || X86_64 || MCORE2 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MATOM
404 404
405# this should be set for all -march=.. options where the compiler 405# this should be set for all -march=.. options where the compiler
406# generates cmov. 406# generates cmov.
@@ -412,6 +412,7 @@ config X86_MINIMUM_CPU_FAMILY
412 int 412 int
413 default "64" if X86_64 413 default "64" if X86_64
414 default "6" if X86_32 && X86_P6_NOP 414 default "6" if X86_32 && X86_P6_NOP
415 default "5" if X86_32 && X86_CMPXCHG64
415 default "4" if X86_32 && (X86_XADD || X86_CMPXCHG || X86_BSWAP || X86_WP_WORKS_OK) 416 default "4" if X86_32 && (X86_XADD || X86_CMPXCHG || X86_BSWAP || X86_WP_WORKS_OK)
416 default "3" 417 default "3"
417 418
diff --git a/arch/x86/Makefile b/arch/x86/Makefile
index 4aefc034e9a0..ba7a6df4db92 100644
--- a/arch/x86/Makefile
+++ b/arch/x86/Makefile
@@ -182,8 +182,8 @@ archclean:
182define archhelp 182define archhelp
183 echo '* bzImage - Compressed kernel image (arch/x86/boot/bzImage)' 183 echo '* bzImage - Compressed kernel image (arch/x86/boot/bzImage)'
184 echo ' install - Install kernel using' 184 echo ' install - Install kernel using'
185 echo ' (your) ~/bin/installkernel or' 185 echo ' (your) ~/bin/$(INSTALLKERNEL) or'
186 echo ' (distribution) /sbin/installkernel or' 186 echo ' (distribution) /sbin/$(INSTALLKERNEL) or'
187 echo ' install to $$(INSTALL_PATH) and run lilo' 187 echo ' install to $$(INSTALL_PATH) and run lilo'
188 echo ' fdimage - Create 1.4MB boot floppy image (arch/x86/boot/fdimage)' 188 echo ' fdimage - Create 1.4MB boot floppy image (arch/x86/boot/fdimage)'
189 echo ' fdimage144 - Create 1.4MB boot floppy image (arch/x86/boot/fdimage)' 189 echo ' fdimage144 - Create 1.4MB boot floppy image (arch/x86/boot/fdimage)'
diff --git a/arch/x86/boot/compressed/head_32.S b/arch/x86/boot/compressed/head_32.S
index 75e4f001e706..f543b70ffae2 100644
--- a/arch/x86/boot/compressed/head_32.S
+++ b/arch/x86/boot/compressed/head_32.S
@@ -23,13 +23,14 @@
23 */ 23 */
24 .text 24 .text
25 25
26#include <linux/init.h>
26#include <linux/linkage.h> 27#include <linux/linkage.h>
27#include <asm/segment.h> 28#include <asm/segment.h>
28#include <asm/page_types.h> 29#include <asm/page_types.h>
29#include <asm/boot.h> 30#include <asm/boot.h>
30#include <asm/asm-offsets.h> 31#include <asm/asm-offsets.h>
31 32
32 .section ".text.head","ax",@progbits 33 __HEAD
33ENTRY(startup_32) 34ENTRY(startup_32)
34 cld 35 cld
35 /* 36 /*
diff --git a/arch/x86/boot/compressed/head_64.S b/arch/x86/boot/compressed/head_64.S
index f62c284db9eb..077e1b69198e 100644
--- a/arch/x86/boot/compressed/head_64.S
+++ b/arch/x86/boot/compressed/head_64.S
@@ -24,6 +24,7 @@
24 .code32 24 .code32
25 .text 25 .text
26 26
27#include <linux/init.h>
27#include <linux/linkage.h> 28#include <linux/linkage.h>
28#include <asm/segment.h> 29#include <asm/segment.h>
29#include <asm/pgtable_types.h> 30#include <asm/pgtable_types.h>
@@ -33,7 +34,7 @@
33#include <asm/processor-flags.h> 34#include <asm/processor-flags.h>
34#include <asm/asm-offsets.h> 35#include <asm/asm-offsets.h>
35 36
36 .section ".text.head" 37 __HEAD
37 .code32 38 .code32
38ENTRY(startup_32) 39ENTRY(startup_32)
39 cld 40 cld
diff --git a/arch/x86/boot/compressed/vmlinux.lds.S b/arch/x86/boot/compressed/vmlinux.lds.S
index cc353e1b3ffd..f4193bb48782 100644
--- a/arch/x86/boot/compressed/vmlinux.lds.S
+++ b/arch/x86/boot/compressed/vmlinux.lds.S
@@ -1,3 +1,5 @@
1#include <asm-generic/vmlinux.lds.h>
2
1OUTPUT_FORMAT(CONFIG_OUTPUT_FORMAT, CONFIG_OUTPUT_FORMAT, CONFIG_OUTPUT_FORMAT) 3OUTPUT_FORMAT(CONFIG_OUTPUT_FORMAT, CONFIG_OUTPUT_FORMAT, CONFIG_OUTPUT_FORMAT)
2 4
3#undef i386 5#undef i386
@@ -18,9 +20,9 @@ SECTIONS
18 * address 0. 20 * address 0.
19 */ 21 */
20 . = 0; 22 . = 0;
21 .text.head : { 23 .head.text : {
22 _head = . ; 24 _head = . ;
23 *(.text.head) 25 HEAD_TEXT
24 _ehead = . ; 26 _ehead = . ;
25 } 27 }
26 .rodata.compressed : { 28 .rodata.compressed : {
diff --git a/arch/x86/boot/install.sh b/arch/x86/boot/install.sh
index 8d60ee15dfd9..d13ec1c38640 100644
--- a/arch/x86/boot/install.sh
+++ b/arch/x86/boot/install.sh
@@ -33,8 +33,8 @@ verify "$3"
33 33
34# User may have a custom install script 34# User may have a custom install script
35 35
36if [ -x ~/bin/${CROSS_COMPILE}installkernel ]; then exec ~/bin/${CROSS_COMPILE}installkernel "$@"; fi 36if [ -x ~/bin/${INSTALLKERNEL} ]; then exec ~/bin/${INSTALLKERNEL} "$@"; fi
37if [ -x /sbin/${CROSS_COMPILE}installkernel ]; then exec /sbin/${CROSS_COMPILE}installkernel "$@"; fi 37if [ -x /sbin/${INSTALLKERNEL} ]; then exec /sbin/${INSTALLKERNEL} "$@"; fi
38 38
39# Default install - same as make zlilo 39# Default install - same as make zlilo
40 40
diff --git a/arch/x86/ia32/ia32entry.S b/arch/x86/ia32/ia32entry.S
index 74619c4f9fda..1733f9f65e82 100644
--- a/arch/x86/ia32/ia32entry.S
+++ b/arch/x86/ia32/ia32entry.S
@@ -21,8 +21,8 @@
21#define __AUDIT_ARCH_LE 0x40000000 21#define __AUDIT_ARCH_LE 0x40000000
22 22
23#ifndef CONFIG_AUDITSYSCALL 23#ifndef CONFIG_AUDITSYSCALL
24#define sysexit_audit int_ret_from_sys_call 24#define sysexit_audit ia32_ret_from_sys_call
25#define sysretl_audit int_ret_from_sys_call 25#define sysretl_audit ia32_ret_from_sys_call
26#endif 26#endif
27 27
28#define IA32_NR_syscalls ((ia32_syscall_end - ia32_sys_call_table)/8) 28#define IA32_NR_syscalls ((ia32_syscall_end - ia32_sys_call_table)/8)
@@ -39,12 +39,12 @@
39 .endm 39 .endm
40 40
41 /* clobbers %eax */ 41 /* clobbers %eax */
42 .macro CLEAR_RREGS _r9=rax 42 .macro CLEAR_RREGS offset=0, _r9=rax
43 xorl %eax,%eax 43 xorl %eax,%eax
44 movq %rax,R11(%rsp) 44 movq %rax,\offset+R11(%rsp)
45 movq %rax,R10(%rsp) 45 movq %rax,\offset+R10(%rsp)
46 movq %\_r9,R9(%rsp) 46 movq %\_r9,\offset+R9(%rsp)
47 movq %rax,R8(%rsp) 47 movq %rax,\offset+R8(%rsp)
48 .endm 48 .endm
49 49
50 /* 50 /*
@@ -172,6 +172,10 @@ sysexit_from_sys_call:
172 movl RIP-R11(%rsp),%edx /* User %eip */ 172 movl RIP-R11(%rsp),%edx /* User %eip */
173 CFI_REGISTER rip,rdx 173 CFI_REGISTER rip,rdx
174 RESTORE_ARGS 1,24,1,1,1,1 174 RESTORE_ARGS 1,24,1,1,1,1
175 xorq %r8,%r8
176 xorq %r9,%r9
177 xorq %r10,%r10
178 xorq %r11,%r11
175 popfq 179 popfq
176 CFI_ADJUST_CFA_OFFSET -8 180 CFI_ADJUST_CFA_OFFSET -8
177 /*CFI_RESTORE rflags*/ 181 /*CFI_RESTORE rflags*/
@@ -202,7 +206,7 @@ sysexit_from_sys_call:
202 206
203 .macro auditsys_exit exit,ebpsave=RBP 207 .macro auditsys_exit exit,ebpsave=RBP
204 testl $(_TIF_ALLWORK_MASK & ~_TIF_SYSCALL_AUDIT),TI_flags(%r10) 208 testl $(_TIF_ALLWORK_MASK & ~_TIF_SYSCALL_AUDIT),TI_flags(%r10)
205 jnz int_ret_from_sys_call 209 jnz ia32_ret_from_sys_call
206 TRACE_IRQS_ON 210 TRACE_IRQS_ON
207 sti 211 sti
208 movl %eax,%esi /* second arg, syscall return value */ 212 movl %eax,%esi /* second arg, syscall return value */
@@ -218,8 +222,9 @@ sysexit_from_sys_call:
218 cli 222 cli
219 TRACE_IRQS_OFF 223 TRACE_IRQS_OFF
220 testl %edi,TI_flags(%r10) 224 testl %edi,TI_flags(%r10)
221 jnz int_with_check 225 jz \exit
222 jmp \exit 226 CLEAR_RREGS -ARGOFFSET
227 jmp int_with_check
223 .endm 228 .endm
224 229
225sysenter_auditsys: 230sysenter_auditsys:
@@ -329,6 +334,9 @@ sysretl_from_sys_call:
329 CFI_REGISTER rip,rcx 334 CFI_REGISTER rip,rcx
330 movl EFLAGS-ARGOFFSET(%rsp),%r11d 335 movl EFLAGS-ARGOFFSET(%rsp),%r11d
331 /*CFI_REGISTER rflags,r11*/ 336 /*CFI_REGISTER rflags,r11*/
337 xorq %r10,%r10
338 xorq %r9,%r9
339 xorq %r8,%r8
332 TRACE_IRQS_ON 340 TRACE_IRQS_ON
333 movl RSP-ARGOFFSET(%rsp),%esp 341 movl RSP-ARGOFFSET(%rsp),%esp
334 CFI_RESTORE rsp 342 CFI_RESTORE rsp
@@ -353,7 +361,7 @@ cstar_tracesys:
353#endif 361#endif
354 xchgl %r9d,%ebp 362 xchgl %r9d,%ebp
355 SAVE_REST 363 SAVE_REST
356 CLEAR_RREGS r9 364 CLEAR_RREGS 0, r9
357 movq $-ENOSYS,RAX(%rsp) /* ptrace can change this for a bad syscall */ 365 movq $-ENOSYS,RAX(%rsp) /* ptrace can change this for a bad syscall */
358 movq %rsp,%rdi /* &pt_regs -> arg1 */ 366 movq %rsp,%rdi /* &pt_regs -> arg1 */
359 call syscall_trace_enter 367 call syscall_trace_enter
@@ -425,6 +433,8 @@ ia32_do_call:
425 call *ia32_sys_call_table(,%rax,8) # xxx: rip relative 433 call *ia32_sys_call_table(,%rax,8) # xxx: rip relative
426ia32_sysret: 434ia32_sysret:
427 movq %rax,RAX-ARGOFFSET(%rsp) 435 movq %rax,RAX-ARGOFFSET(%rsp)
436ia32_ret_from_sys_call:
437 CLEAR_RREGS -ARGOFFSET
428 jmp int_ret_from_sys_call 438 jmp int_ret_from_sys_call
429 439
430ia32_tracesys: 440ia32_tracesys:
@@ -442,8 +452,8 @@ END(ia32_syscall)
442 452
443ia32_badsys: 453ia32_badsys:
444 movq $0,ORIG_RAX-ARGOFFSET(%rsp) 454 movq $0,ORIG_RAX-ARGOFFSET(%rsp)
445 movq $-ENOSYS,RAX-ARGOFFSET(%rsp) 455 movq $-ENOSYS,%rax
446 jmp int_ret_from_sys_call 456 jmp ia32_sysret
447 457
448quiet_ni_syscall: 458quiet_ni_syscall:
449 movq $-ENOSYS,%rax 459 movq $-ENOSYS,%rax
diff --git a/arch/x86/include/asm/cache.h b/arch/x86/include/asm/cache.h
index 5d367caa0e36..549860d3be8f 100644
--- a/arch/x86/include/asm/cache.h
+++ b/arch/x86/include/asm/cache.h
@@ -1,6 +1,8 @@
1#ifndef _ASM_X86_CACHE_H 1#ifndef _ASM_X86_CACHE_H
2#define _ASM_X86_CACHE_H 2#define _ASM_X86_CACHE_H
3 3
4#include <linux/linkage.h>
5
4/* L1 cache line size */ 6/* L1 cache line size */
5#define L1_CACHE_SHIFT (CONFIG_X86_L1_CACHE_SHIFT) 7#define L1_CACHE_SHIFT (CONFIG_X86_L1_CACHE_SHIFT)
6#define L1_CACHE_BYTES (1 << L1_CACHE_SHIFT) 8#define L1_CACHE_BYTES (1 << L1_CACHE_SHIFT)
@@ -13,7 +15,7 @@
13#ifdef CONFIG_SMP 15#ifdef CONFIG_SMP
14#define __cacheline_aligned_in_smp \ 16#define __cacheline_aligned_in_smp \
15 __attribute__((__aligned__(1 << (INTERNODE_CACHE_SHIFT)))) \ 17 __attribute__((__aligned__(1 << (INTERNODE_CACHE_SHIFT)))) \
16 __attribute__((__section__(".data.page_aligned"))) 18 __page_aligned_data
17#endif 19#endif
18#endif 20#endif
19 21
diff --git a/arch/x86/include/asm/checksum_32.h b/arch/x86/include/asm/checksum_32.h
index 7c5ef8b14d92..46fc474fd819 100644
--- a/arch/x86/include/asm/checksum_32.h
+++ b/arch/x86/include/asm/checksum_32.h
@@ -161,7 +161,8 @@ static inline __sum16 csum_ipv6_magic(const struct in6_addr *saddr,
161 "adcl $0, %0 ;\n" 161 "adcl $0, %0 ;\n"
162 : "=&r" (sum) 162 : "=&r" (sum)
163 : "r" (saddr), "r" (daddr), 163 : "r" (saddr), "r" (daddr),
164 "r" (htonl(len)), "r" (htonl(proto)), "0" (sum)); 164 "r" (htonl(len)), "r" (htonl(proto)), "0" (sum)
165 : "memory");
165 166
166 return csum_fold(sum); 167 return csum_fold(sum);
167} 168}
diff --git a/arch/x86/include/asm/cmpxchg_32.h b/arch/x86/include/asm/cmpxchg_32.h
index 82ceb788a981..ee1931be6593 100644
--- a/arch/x86/include/asm/cmpxchg_32.h
+++ b/arch/x86/include/asm/cmpxchg_32.h
@@ -312,19 +312,23 @@ static inline unsigned long cmpxchg_386(volatile void *ptr, unsigned long old,
312 312
313extern unsigned long long cmpxchg_486_u64(volatile void *, u64, u64); 313extern unsigned long long cmpxchg_486_u64(volatile void *, u64, u64);
314 314
315#define cmpxchg64(ptr, o, n) \ 315#define cmpxchg64(ptr, o, n) \
316({ \ 316({ \
317 __typeof__(*(ptr)) __ret; \ 317 __typeof__(*(ptr)) __ret; \
318 if (likely(boot_cpu_data.x86 > 4)) \ 318 __typeof__(*(ptr)) __old = (o); \
319 __ret = (__typeof__(*(ptr)))__cmpxchg64((ptr), \ 319 __typeof__(*(ptr)) __new = (n); \
320 (unsigned long long)(o), \ 320 alternative_io("call cmpxchg8b_emu", \
321 (unsigned long long)(n)); \ 321 "lock; cmpxchg8b (%%esi)" , \
322 else \ 322 X86_FEATURE_CX8, \
323 __ret = (__typeof__(*(ptr)))cmpxchg_486_u64((ptr), \ 323 "=A" (__ret), \
324 (unsigned long long)(o), \ 324 "S" ((ptr)), "0" (__old), \
325 (unsigned long long)(n)); \ 325 "b" ((unsigned int)__new), \
326 __ret; \ 326 "c" ((unsigned int)(__new>>32)) \
327}) 327 : "memory"); \
328 __ret; })
329
330
331
328#define cmpxchg64_local(ptr, o, n) \ 332#define cmpxchg64_local(ptr, o, n) \
329({ \ 333({ \
330 __typeof__(*(ptr)) __ret; \ 334 __typeof__(*(ptr)) __ret; \
diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index 3be000435fad..d83892226f73 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -796,6 +796,7 @@ asmlinkage void kvm_handle_fault_on_reboot(void);
796#define KVM_ARCH_WANT_MMU_NOTIFIER 796#define KVM_ARCH_WANT_MMU_NOTIFIER
797int kvm_unmap_hva(struct kvm *kvm, unsigned long hva); 797int kvm_unmap_hva(struct kvm *kvm, unsigned long hva);
798int kvm_age_hva(struct kvm *kvm, unsigned long hva); 798int kvm_age_hva(struct kvm *kvm, unsigned long hva);
799void kvm_set_spte_hva(struct kvm *kvm, unsigned long hva, pte_t pte);
799int cpuid_maxphyaddr(struct kvm_vcpu *vcpu); 800int cpuid_maxphyaddr(struct kvm_vcpu *vcpu);
800int kvm_cpu_has_interrupt(struct kvm_vcpu *vcpu); 801int kvm_cpu_has_interrupt(struct kvm_vcpu *vcpu);
801int kvm_arch_interrupt_allowed(struct kvm_vcpu *vcpu); 802int kvm_arch_interrupt_allowed(struct kvm_vcpu *vcpu);
diff --git a/arch/x86/include/asm/mce.h b/arch/x86/include/asm/mce.h
index b608a64c5814..f1363b72364f 100644
--- a/arch/x86/include/asm/mce.h
+++ b/arch/x86/include/asm/mce.h
@@ -133,6 +133,8 @@ static inline void winchip_mcheck_init(struct cpuinfo_x86 *c) {}
133static inline void enable_p5_mce(void) {} 133static inline void enable_p5_mce(void) {}
134#endif 134#endif
135 135
136extern void (*x86_mce_decode_callback)(struct mce *m);
137
136void mce_setup(struct mce *m); 138void mce_setup(struct mce *m);
137void mce_log(struct mce *m); 139void mce_log(struct mce *m);
138DECLARE_PER_CPU(struct sys_device, mce_dev); 140DECLARE_PER_CPU(struct sys_device, mce_dev);
diff --git a/arch/x86/include/asm/mmu_context.h b/arch/x86/include/asm/mmu_context.h
index f923203dc39a..4a2d4e0c18d9 100644
--- a/arch/x86/include/asm/mmu_context.h
+++ b/arch/x86/include/asm/mmu_context.h
@@ -37,12 +37,12 @@ static inline void switch_mm(struct mm_struct *prev, struct mm_struct *next,
37 37
38 if (likely(prev != next)) { 38 if (likely(prev != next)) {
39 /* stop flush ipis for the previous mm */ 39 /* stop flush ipis for the previous mm */
40 cpu_clear(cpu, prev->cpu_vm_mask); 40 cpumask_clear_cpu(cpu, mm_cpumask(prev));
41#ifdef CONFIG_SMP 41#ifdef CONFIG_SMP
42 percpu_write(cpu_tlbstate.state, TLBSTATE_OK); 42 percpu_write(cpu_tlbstate.state, TLBSTATE_OK);
43 percpu_write(cpu_tlbstate.active_mm, next); 43 percpu_write(cpu_tlbstate.active_mm, next);
44#endif 44#endif
45 cpu_set(cpu, next->cpu_vm_mask); 45 cpumask_set_cpu(cpu, mm_cpumask(next));
46 46
47 /* Re-load page tables */ 47 /* Re-load page tables */
48 load_cr3(next->pgd); 48 load_cr3(next->pgd);
@@ -58,7 +58,7 @@ static inline void switch_mm(struct mm_struct *prev, struct mm_struct *next,
58 percpu_write(cpu_tlbstate.state, TLBSTATE_OK); 58 percpu_write(cpu_tlbstate.state, TLBSTATE_OK);
59 BUG_ON(percpu_read(cpu_tlbstate.active_mm) != next); 59 BUG_ON(percpu_read(cpu_tlbstate.active_mm) != next);
60 60
61 if (!cpu_test_and_set(cpu, next->cpu_vm_mask)) { 61 if (!cpumask_test_and_set_cpu(cpu, mm_cpumask(next))) {
62 /* We were in lazy tlb mode and leave_mm disabled 62 /* We were in lazy tlb mode and leave_mm disabled
63 * tlb flush IPI delivery. We must reload CR3 63 * tlb flush IPI delivery. We must reload CR3
64 * to make sure to use no freed page tables. 64 * to make sure to use no freed page tables.
diff --git a/arch/x86/include/asm/nmi.h b/arch/x86/include/asm/nmi.h
index e63cf7d441e1..139d4c1a33a7 100644
--- a/arch/x86/include/asm/nmi.h
+++ b/arch/x86/include/asm/nmi.h
@@ -40,8 +40,7 @@ extern unsigned int nmi_watchdog;
40#define NMI_INVALID 3 40#define NMI_INVALID 3
41 41
42struct ctl_table; 42struct ctl_table;
43struct file; 43extern int proc_nmi_enabled(struct ctl_table *, int ,
44extern int proc_nmi_enabled(struct ctl_table *, int , struct file *,
45 void __user *, size_t *, loff_t *); 44 void __user *, size_t *, loff_t *);
46extern int unknown_nmi_panic; 45extern int unknown_nmi_panic;
47 46
diff --git a/arch/x86/include/asm/paravirt.h b/arch/x86/include/asm/paravirt.h
index 8aebcc41041d..efb38994859c 100644
--- a/arch/x86/include/asm/paravirt.h
+++ b/arch/x86/include/asm/paravirt.h
@@ -840,42 +840,22 @@ static __always_inline void __raw_spin_unlock(struct raw_spinlock *lock)
840 840
841static inline unsigned long __raw_local_save_flags(void) 841static inline unsigned long __raw_local_save_flags(void)
842{ 842{
843 unsigned long f; 843 return PVOP_CALLEE0(unsigned long, pv_irq_ops.save_fl);
844
845 asm volatile(paravirt_alt(PARAVIRT_CALL)
846 : "=a"(f)
847 : paravirt_type(pv_irq_ops.save_fl),
848 paravirt_clobber(CLBR_EAX)
849 : "memory", "cc");
850 return f;
851} 844}
852 845
853static inline void raw_local_irq_restore(unsigned long f) 846static inline void raw_local_irq_restore(unsigned long f)
854{ 847{
855 asm volatile(paravirt_alt(PARAVIRT_CALL) 848 PVOP_VCALLEE1(pv_irq_ops.restore_fl, f);
856 : "=a"(f)
857 : PV_FLAGS_ARG(f),
858 paravirt_type(pv_irq_ops.restore_fl),
859 paravirt_clobber(CLBR_EAX)
860 : "memory", "cc");
861} 849}
862 850
863static inline void raw_local_irq_disable(void) 851static inline void raw_local_irq_disable(void)
864{ 852{
865 asm volatile(paravirt_alt(PARAVIRT_CALL) 853 PVOP_VCALLEE0(pv_irq_ops.irq_disable);
866 :
867 : paravirt_type(pv_irq_ops.irq_disable),
868 paravirt_clobber(CLBR_EAX)
869 : "memory", "eax", "cc");
870} 854}
871 855
872static inline void raw_local_irq_enable(void) 856static inline void raw_local_irq_enable(void)
873{ 857{
874 asm volatile(paravirt_alt(PARAVIRT_CALL) 858 PVOP_VCALLEE0(pv_irq_ops.irq_enable);
875 :
876 : paravirt_type(pv_irq_ops.irq_enable),
877 paravirt_clobber(CLBR_EAX)
878 : "memory", "eax", "cc");
879} 859}
880 860
881static inline unsigned long __raw_local_irq_save(void) 861static inline unsigned long __raw_local_irq_save(void)
diff --git a/arch/x86/include/asm/paravirt_types.h b/arch/x86/include/asm/paravirt_types.h
index dd0f5b32489d..9357473c8da0 100644
--- a/arch/x86/include/asm/paravirt_types.h
+++ b/arch/x86/include/asm/paravirt_types.h
@@ -494,10 +494,11 @@ int paravirt_disable_iospace(void);
494#define EXTRA_CLOBBERS 494#define EXTRA_CLOBBERS
495#define VEXTRA_CLOBBERS 495#define VEXTRA_CLOBBERS
496#else /* CONFIG_X86_64 */ 496#else /* CONFIG_X86_64 */
497/* [re]ax isn't an arg, but the return val */
497#define PVOP_VCALL_ARGS \ 498#define PVOP_VCALL_ARGS \
498 unsigned long __edi = __edi, __esi = __esi, \ 499 unsigned long __edi = __edi, __esi = __esi, \
499 __edx = __edx, __ecx = __ecx 500 __edx = __edx, __ecx = __ecx, __eax = __eax
500#define PVOP_CALL_ARGS PVOP_VCALL_ARGS, __eax 501#define PVOP_CALL_ARGS PVOP_VCALL_ARGS
501 502
502#define PVOP_CALL_ARG1(x) "D" ((unsigned long)(x)) 503#define PVOP_CALL_ARG1(x) "D" ((unsigned long)(x))
503#define PVOP_CALL_ARG2(x) "S" ((unsigned long)(x)) 504#define PVOP_CALL_ARG2(x) "S" ((unsigned long)(x))
@@ -509,6 +510,7 @@ int paravirt_disable_iospace(void);
509 "=c" (__ecx) 510 "=c" (__ecx)
510#define PVOP_CALL_CLOBBERS PVOP_VCALL_CLOBBERS, "=a" (__eax) 511#define PVOP_CALL_CLOBBERS PVOP_VCALL_CLOBBERS, "=a" (__eax)
511 512
513/* void functions are still allowed [re]ax for scratch */
512#define PVOP_VCALLEE_CLOBBERS "=a" (__eax) 514#define PVOP_VCALLEE_CLOBBERS "=a" (__eax)
513#define PVOP_CALLEE_CLOBBERS PVOP_VCALLEE_CLOBBERS 515#define PVOP_CALLEE_CLOBBERS PVOP_VCALLEE_CLOBBERS
514 516
@@ -583,8 +585,8 @@ int paravirt_disable_iospace(void);
583 VEXTRA_CLOBBERS, \ 585 VEXTRA_CLOBBERS, \
584 pre, post, ##__VA_ARGS__) 586 pre, post, ##__VA_ARGS__)
585 587
586#define __PVOP_VCALLEESAVE(rettype, op, pre, post, ...) \ 588#define __PVOP_VCALLEESAVE(op, pre, post, ...) \
587 ____PVOP_CALL(rettype, op.func, CLBR_RET_REG, \ 589 ____PVOP_VCALL(op.func, CLBR_RET_REG, \
588 PVOP_VCALLEE_CLOBBERS, , \ 590 PVOP_VCALLEE_CLOBBERS, , \
589 pre, post, ##__VA_ARGS__) 591 pre, post, ##__VA_ARGS__)
590 592
diff --git a/arch/x86/include/asm/pci.h b/arch/x86/include/asm/pci.h
index f76a162c082c..ada8c201d513 100644
--- a/arch/x86/include/asm/pci.h
+++ b/arch/x86/include/asm/pci.h
@@ -143,7 +143,11 @@ static inline int __pcibus_to_node(const struct pci_bus *bus)
143static inline const struct cpumask * 143static inline const struct cpumask *
144cpumask_of_pcibus(const struct pci_bus *bus) 144cpumask_of_pcibus(const struct pci_bus *bus)
145{ 145{
146 return cpumask_of_node(__pcibus_to_node(bus)); 146 int node;
147
148 node = __pcibus_to_node(bus);
149 return (node == -1) ? cpu_online_mask :
150 cpumask_of_node(node);
147} 151}
148#endif 152#endif
149 153
diff --git a/arch/x86/include/asm/pgtable_types.h b/arch/x86/include/asm/pgtable_types.h
index 7b467bf3c680..d1f4a760be23 100644
--- a/arch/x86/include/asm/pgtable_types.h
+++ b/arch/x86/include/asm/pgtable_types.h
@@ -277,6 +277,7 @@ static inline pteval_t pte_flags(pte_t pte)
277typedef struct page *pgtable_t; 277typedef struct page *pgtable_t;
278 278
279extern pteval_t __supported_pte_mask; 279extern pteval_t __supported_pte_mask;
280extern void set_nx(void);
280extern int nx_enabled; 281extern int nx_enabled;
281 282
282#define pgprot_writecombine pgprot_writecombine 283#define pgprot_writecombine pgprot_writecombine
diff --git a/arch/x86/include/asm/smp.h b/arch/x86/include/asm/smp.h
index 6a84ed166aec..1e796782cd7b 100644
--- a/arch/x86/include/asm/smp.h
+++ b/arch/x86/include/asm/smp.h
@@ -121,7 +121,6 @@ static inline void arch_send_call_function_single_ipi(int cpu)
121 smp_ops.send_call_func_single_ipi(cpu); 121 smp_ops.send_call_func_single_ipi(cpu);
122} 122}
123 123
124#define arch_send_call_function_ipi_mask arch_send_call_function_ipi_mask
125static inline void arch_send_call_function_ipi_mask(const struct cpumask *mask) 124static inline void arch_send_call_function_ipi_mask(const struct cpumask *mask)
126{ 125{
127 smp_ops.send_call_func_ipi(mask); 126 smp_ops.send_call_func_ipi(mask);
diff --git a/arch/x86/include/asm/topology.h b/arch/x86/include/asm/topology.h
index 6f0695d744bf..25a92842dd99 100644
--- a/arch/x86/include/asm/topology.h
+++ b/arch/x86/include/asm/topology.h
@@ -165,21 +165,11 @@ static inline int numa_node_id(void)
165 return 0; 165 return 0;
166} 166}
167 167
168static inline int cpu_to_node(int cpu)
169{
170 return 0;
171}
172
173static inline int early_cpu_to_node(int cpu) 168static inline int early_cpu_to_node(int cpu)
174{ 169{
175 return 0; 170 return 0;
176} 171}
177 172
178static inline const struct cpumask *cpumask_of_node(int node)
179{
180 return cpu_online_mask;
181}
182
183static inline void setup_node_to_cpumask_map(void) { } 173static inline void setup_node_to_cpumask_map(void) { }
184 174
185#endif 175#endif
diff --git a/arch/x86/kernel/acpi/cstate.c b/arch/x86/kernel/acpi/cstate.c
index 8c44c232efcb..59cdfa4686b2 100644
--- a/arch/x86/kernel/acpi/cstate.c
+++ b/arch/x86/kernel/acpi/cstate.c
@@ -48,7 +48,7 @@ void acpi_processor_power_init_bm_check(struct acpi_processor_flags *flags,
48 * P4, Core and beyond CPUs 48 * P4, Core and beyond CPUs
49 */ 49 */
50 if (c->x86_vendor == X86_VENDOR_INTEL && 50 if (c->x86_vendor == X86_VENDOR_INTEL &&
51 (c->x86 > 0x6 || (c->x86 == 6 && c->x86_model >= 14))) 51 (c->x86 > 0xf || (c->x86 == 6 && c->x86_model >= 14)))
52 flags->bm_control = 0; 52 flags->bm_control = 0;
53} 53}
54EXPORT_SYMBOL(acpi_processor_power_init_bm_check); 54EXPORT_SYMBOL(acpi_processor_power_init_bm_check);
diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
index 64970b9885f2..dc69f28489f5 100644
--- a/arch/x86/kernel/apic/io_apic.c
+++ b/arch/x86/kernel/apic/io_apic.c
@@ -227,17 +227,14 @@ static struct irq_cfg *get_one_free_irq_cfg(int node)
227 227
228 cfg = kzalloc_node(sizeof(*cfg), GFP_ATOMIC, node); 228 cfg = kzalloc_node(sizeof(*cfg), GFP_ATOMIC, node);
229 if (cfg) { 229 if (cfg) {
230 if (!alloc_cpumask_var_node(&cfg->domain, GFP_ATOMIC, node)) { 230 if (!zalloc_cpumask_var_node(&cfg->domain, GFP_ATOMIC, node)) {
231 kfree(cfg); 231 kfree(cfg);
232 cfg = NULL; 232 cfg = NULL;
233 } else if (!alloc_cpumask_var_node(&cfg->old_domain, 233 } else if (!zalloc_cpumask_var_node(&cfg->old_domain,
234 GFP_ATOMIC, node)) { 234 GFP_ATOMIC, node)) {
235 free_cpumask_var(cfg->domain); 235 free_cpumask_var(cfg->domain);
236 kfree(cfg); 236 kfree(cfg);
237 cfg = NULL; 237 cfg = NULL;
238 } else {
239 cpumask_clear(cfg->domain);
240 cpumask_clear(cfg->old_domain);
241 } 238 }
242 } 239 }
243 240
diff --git a/arch/x86/kernel/apic/nmi.c b/arch/x86/kernel/apic/nmi.c
index cb66a22d98ad..7ff61d6a188a 100644
--- a/arch/x86/kernel/apic/nmi.c
+++ b/arch/x86/kernel/apic/nmi.c
@@ -508,14 +508,14 @@ static int unknown_nmi_panic_callback(struct pt_regs *regs, int cpu)
508/* 508/*
509 * proc handler for /proc/sys/kernel/nmi 509 * proc handler for /proc/sys/kernel/nmi
510 */ 510 */
511int proc_nmi_enabled(struct ctl_table *table, int write, struct file *file, 511int proc_nmi_enabled(struct ctl_table *table, int write,
512 void __user *buffer, size_t *length, loff_t *ppos) 512 void __user *buffer, size_t *length, loff_t *ppos)
513{ 513{
514 int old_state; 514 int old_state;
515 515
516 nmi_watchdog_enabled = (atomic_read(&nmi_active) > 0) ? 1 : 0; 516 nmi_watchdog_enabled = (atomic_read(&nmi_active) > 0) ? 1 : 0;
517 old_state = nmi_watchdog_enabled; 517 old_state = nmi_watchdog_enabled;
518 proc_dointvec(table, write, file, buffer, length, ppos); 518 proc_dointvec(table, write, buffer, length, ppos);
519 if (!!old_state == !!nmi_watchdog_enabled) 519 if (!!old_state == !!nmi_watchdog_enabled)
520 return 0; 520 return 0;
521 521
diff --git a/arch/x86/kernel/cpu/mcheck/mce-inject.c b/arch/x86/kernel/cpu/mcheck/mce-inject.c
index 7029f0e2acad..472763d92098 100644
--- a/arch/x86/kernel/cpu/mcheck/mce-inject.c
+++ b/arch/x86/kernel/cpu/mcheck/mce-inject.c
@@ -98,8 +98,9 @@ static struct notifier_block mce_raise_nb = {
98}; 98};
99 99
100/* Inject mce on current CPU */ 100/* Inject mce on current CPU */
101static int raise_local(struct mce *m) 101static int raise_local(void)
102{ 102{
103 struct mce *m = &__get_cpu_var(injectm);
103 int context = MCJ_CTX(m->inject_flags); 104 int context = MCJ_CTX(m->inject_flags);
104 int ret = 0; 105 int ret = 0;
105 int cpu = m->extcpu; 106 int cpu = m->extcpu;
@@ -167,12 +168,12 @@ static void raise_mce(struct mce *m)
167 } 168 }
168 cpu_relax(); 169 cpu_relax();
169 } 170 }
170 raise_local(m); 171 raise_local();
171 put_cpu(); 172 put_cpu();
172 put_online_cpus(); 173 put_online_cpus();
173 } else 174 } else
174#endif 175#endif
175 raise_local(m); 176 raise_local();
176} 177}
177 178
178/* Error injection interface */ 179/* Error injection interface */
diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c
index 2f5aab26320e..b1598a9436d0 100644
--- a/arch/x86/kernel/cpu/mcheck/mce.c
+++ b/arch/x86/kernel/cpu/mcheck/mce.c
@@ -85,6 +85,18 @@ static DECLARE_WAIT_QUEUE_HEAD(mce_wait);
85static DEFINE_PER_CPU(struct mce, mces_seen); 85static DEFINE_PER_CPU(struct mce, mces_seen);
86static int cpu_missing; 86static int cpu_missing;
87 87
88static void default_decode_mce(struct mce *m)
89{
90 pr_emerg("No human readable MCE decoding support on this CPU type.\n");
91 pr_emerg("Run the message through 'mcelog --ascii' to decode.\n");
92}
93
94/*
95 * CPU/chipset specific EDAC code can register a callback here to print
96 * MCE errors in a human-readable form:
97 */
98void (*x86_mce_decode_callback)(struct mce *m) = default_decode_mce;
99EXPORT_SYMBOL(x86_mce_decode_callback);
88 100
89/* MCA banks polled by the period polling timer for corrected events */ 101/* MCA banks polled by the period polling timer for corrected events */
90DEFINE_PER_CPU(mce_banks_t, mce_poll_banks) = { 102DEFINE_PER_CPU(mce_banks_t, mce_poll_banks) = {
@@ -165,49 +177,46 @@ void mce_log(struct mce *mce)
165 set_bit(0, &mce_need_notify); 177 set_bit(0, &mce_need_notify);
166} 178}
167 179
168void __weak decode_mce(struct mce *m)
169{
170 return;
171}
172
173static void print_mce(struct mce *m) 180static void print_mce(struct mce *m)
174{ 181{
175 printk(KERN_EMERG 182 pr_emerg("CPU %d: Machine Check Exception: %16Lx Bank %d: %016Lx\n",
176 "CPU %d: Machine Check Exception: %16Lx Bank %d: %016Lx\n",
177 m->extcpu, m->mcgstatus, m->bank, m->status); 183 m->extcpu, m->mcgstatus, m->bank, m->status);
184
178 if (m->ip) { 185 if (m->ip) {
179 printk(KERN_EMERG "RIP%s %02x:<%016Lx> ", 186 pr_emerg("RIP%s %02x:<%016Lx> ",
180 !(m->mcgstatus & MCG_STATUS_EIPV) ? " !INEXACT!" : "", 187 !(m->mcgstatus & MCG_STATUS_EIPV) ? " !INEXACT!" : "",
181 m->cs, m->ip); 188 m->cs, m->ip);
189
182 if (m->cs == __KERNEL_CS) 190 if (m->cs == __KERNEL_CS)
183 print_symbol("{%s}", m->ip); 191 print_symbol("{%s}", m->ip);
184 printk(KERN_CONT "\n"); 192 pr_cont("\n");
185 } 193 }
186 printk(KERN_EMERG "TSC %llx ", m->tsc); 194
195 pr_emerg("TSC %llx ", m->tsc);
187 if (m->addr) 196 if (m->addr)
188 printk(KERN_CONT "ADDR %llx ", m->addr); 197 pr_cont("ADDR %llx ", m->addr);
189 if (m->misc) 198 if (m->misc)
190 printk(KERN_CONT "MISC %llx ", m->misc); 199 pr_cont("MISC %llx ", m->misc);
191 printk(KERN_CONT "\n"); 200
192 printk(KERN_EMERG "PROCESSOR %u:%x TIME %llu SOCKET %u APIC %x\n", 201 pr_cont("\n");
193 m->cpuvendor, m->cpuid, m->time, m->socketid, 202 pr_emerg("PROCESSOR %u:%x TIME %llu SOCKET %u APIC %x\n",
194 m->apicid); 203 m->cpuvendor, m->cpuid, m->time, m->socketid, m->apicid);
195 204
196 decode_mce(m); 205 /*
206 * Print out human-readable details about the MCE error,
207 * (if the CPU has an implementation for that):
208 */
209 x86_mce_decode_callback(m);
197} 210}
198 211
199static void print_mce_head(void) 212static void print_mce_head(void)
200{ 213{
201 printk(KERN_EMERG "\nHARDWARE ERROR\n"); 214 pr_emerg("\nHARDWARE ERROR\n");
202} 215}
203 216
204static void print_mce_tail(void) 217static void print_mce_tail(void)
205{ 218{
206 printk(KERN_EMERG "This is not a software problem!\n" 219 pr_emerg("This is not a software problem!\n");
207#if (!defined(CONFIG_EDAC) || !defined(CONFIG_CPU_SUP_AMD))
208 "Run through mcelog --ascii to decode and contact your hardware vendor\n"
209#endif
210 );
211} 220}
212 221
213#define PANIC_TIMEOUT 5 /* 5 seconds */ 222#define PANIC_TIMEOUT 5 /* 5 seconds */
@@ -221,6 +230,7 @@ static atomic_t mce_fake_paniced;
221static void wait_for_panic(void) 230static void wait_for_panic(void)
222{ 231{
223 long timeout = PANIC_TIMEOUT*USEC_PER_SEC; 232 long timeout = PANIC_TIMEOUT*USEC_PER_SEC;
233
224 preempt_disable(); 234 preempt_disable();
225 local_irq_enable(); 235 local_irq_enable();
226 while (timeout-- > 0) 236 while (timeout-- > 0)
@@ -288,6 +298,7 @@ static void mce_panic(char *msg, struct mce *final, char *exp)
288static int msr_to_offset(u32 msr) 298static int msr_to_offset(u32 msr)
289{ 299{
290 unsigned bank = __get_cpu_var(injectm.bank); 300 unsigned bank = __get_cpu_var(injectm.bank);
301
291 if (msr == rip_msr) 302 if (msr == rip_msr)
292 return offsetof(struct mce, ip); 303 return offsetof(struct mce, ip);
293 if (msr == MSR_IA32_MCx_STATUS(bank)) 304 if (msr == MSR_IA32_MCx_STATUS(bank))
@@ -305,13 +316,25 @@ static int msr_to_offset(u32 msr)
305static u64 mce_rdmsrl(u32 msr) 316static u64 mce_rdmsrl(u32 msr)
306{ 317{
307 u64 v; 318 u64 v;
319
308 if (__get_cpu_var(injectm).finished) { 320 if (__get_cpu_var(injectm).finished) {
309 int offset = msr_to_offset(msr); 321 int offset = msr_to_offset(msr);
322
310 if (offset < 0) 323 if (offset < 0)
311 return 0; 324 return 0;
312 return *(u64 *)((char *)&__get_cpu_var(injectm) + offset); 325 return *(u64 *)((char *)&__get_cpu_var(injectm) + offset);
313 } 326 }
314 rdmsrl(msr, v); 327
328 if (rdmsrl_safe(msr, &v)) {
329 WARN_ONCE(1, "mce: Unable to read msr %d!\n", msr);
330 /*
331 * Return zero in case the access faulted. This should
332 * not happen normally but can happen if the CPU does
333 * something weird, or if the code is buggy.
334 */
335 v = 0;
336 }
337
315 return v; 338 return v;
316} 339}
317 340
@@ -319,6 +342,7 @@ static void mce_wrmsrl(u32 msr, u64 v)
319{ 342{
320 if (__get_cpu_var(injectm).finished) { 343 if (__get_cpu_var(injectm).finished) {
321 int offset = msr_to_offset(msr); 344 int offset = msr_to_offset(msr);
345
322 if (offset >= 0) 346 if (offset >= 0)
323 *(u64 *)((char *)&__get_cpu_var(injectm) + offset) = v; 347 *(u64 *)((char *)&__get_cpu_var(injectm) + offset) = v;
324 return; 348 return;
@@ -415,7 +439,7 @@ static inline void mce_get_rip(struct mce *m, struct pt_regs *regs)
415 m->ip = mce_rdmsrl(rip_msr); 439 m->ip = mce_rdmsrl(rip_msr);
416} 440}
417 441
418#ifdef CONFIG_X86_LOCAL_APIC 442#ifdef CONFIG_X86_LOCAL_APIC
419/* 443/*
420 * Called after interrupts have been reenabled again 444 * Called after interrupts have been reenabled again
421 * when a MCE happened during an interrupts off region 445 * when a MCE happened during an interrupts off region
@@ -1172,6 +1196,7 @@ static int mce_banks_init(void)
1172 return -ENOMEM; 1196 return -ENOMEM;
1173 for (i = 0; i < banks; i++) { 1197 for (i = 0; i < banks; i++) {
1174 struct mce_bank *b = &mce_banks[i]; 1198 struct mce_bank *b = &mce_banks[i];
1199
1175 b->ctl = -1ULL; 1200 b->ctl = -1ULL;
1176 b->init = 1; 1201 b->init = 1;
1177 } 1202 }
@@ -1203,6 +1228,7 @@ static int __cpuinit mce_cap_init(void)
1203 banks = b; 1228 banks = b;
1204 if (!mce_banks) { 1229 if (!mce_banks) {
1205 int err = mce_banks_init(); 1230 int err = mce_banks_init();
1231
1206 if (err) 1232 if (err)
1207 return err; 1233 return err;
1208 } 1234 }
@@ -1237,6 +1263,7 @@ static void mce_init(void)
1237 1263
1238 for (i = 0; i < banks; i++) { 1264 for (i = 0; i < banks; i++) {
1239 struct mce_bank *b = &mce_banks[i]; 1265 struct mce_bank *b = &mce_banks[i];
1266
1240 if (!b->init) 1267 if (!b->init)
1241 continue; 1268 continue;
1242 wrmsrl(MSR_IA32_MCx_CTL(i), b->ctl); 1269 wrmsrl(MSR_IA32_MCx_CTL(i), b->ctl);
@@ -1626,6 +1653,7 @@ static int mce_disable(void)
1626 1653
1627 for (i = 0; i < banks; i++) { 1654 for (i = 0; i < banks; i++) {
1628 struct mce_bank *b = &mce_banks[i]; 1655 struct mce_bank *b = &mce_banks[i];
1656
1629 if (b->init) 1657 if (b->init)
1630 wrmsrl(MSR_IA32_MCx_CTL(i), 0); 1658 wrmsrl(MSR_IA32_MCx_CTL(i), 0);
1631 } 1659 }
@@ -1911,6 +1939,7 @@ static void mce_disable_cpu(void *h)
1911 cmci_clear(); 1939 cmci_clear();
1912 for (i = 0; i < banks; i++) { 1940 for (i = 0; i < banks; i++) {
1913 struct mce_bank *b = &mce_banks[i]; 1941 struct mce_bank *b = &mce_banks[i];
1942
1914 if (b->init) 1943 if (b->init)
1915 wrmsrl(MSR_IA32_MCx_CTL(i), 0); 1944 wrmsrl(MSR_IA32_MCx_CTL(i), 0);
1916 } 1945 }
@@ -1928,6 +1957,7 @@ static void mce_reenable_cpu(void *h)
1928 cmci_reenable(); 1957 cmci_reenable();
1929 for (i = 0; i < banks; i++) { 1958 for (i = 0; i < banks; i++) {
1930 struct mce_bank *b = &mce_banks[i]; 1959 struct mce_bank *b = &mce_banks[i];
1960
1931 if (b->init) 1961 if (b->init)
1932 wrmsrl(MSR_IA32_MCx_CTL(i), b->ctl); 1962 wrmsrl(MSR_IA32_MCx_CTL(i), b->ctl);
1933 } 1963 }
diff --git a/arch/x86/kernel/cpu/mcheck/mce_intel.c b/arch/x86/kernel/cpu/mcheck/mce_intel.c
index 889f665fe93d..7c785634af2b 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_intel.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_intel.c
@@ -8,6 +8,7 @@
8#include <linux/init.h> 8#include <linux/init.h>
9#include <linux/interrupt.h> 9#include <linux/interrupt.h>
10#include <linux/percpu.h> 10#include <linux/percpu.h>
11#include <linux/sched.h>
11#include <asm/apic.h> 12#include <asm/apic.h>
12#include <asm/processor.h> 13#include <asm/processor.h>
13#include <asm/msr.h> 14#include <asm/msr.h>
diff --git a/arch/x86/kernel/cpu/mcheck/therm_throt.c b/arch/x86/kernel/cpu/mcheck/therm_throt.c
index 63a56d147e4a..b3a1dba75330 100644
--- a/arch/x86/kernel/cpu/mcheck/therm_throt.c
+++ b/arch/x86/kernel/cpu/mcheck/therm_throt.c
@@ -34,20 +34,31 @@
34/* How long to wait between reporting thermal events */ 34/* How long to wait between reporting thermal events */
35#define CHECK_INTERVAL (300 * HZ) 35#define CHECK_INTERVAL (300 * HZ)
36 36
37static DEFINE_PER_CPU(__u64, next_check) = INITIAL_JIFFIES; 37/*
38static DEFINE_PER_CPU(unsigned long, thermal_throttle_count); 38 * Current thermal throttling state:
39static DEFINE_PER_CPU(bool, thermal_throttle_active); 39 */
40struct thermal_state {
41 bool is_throttled;
42
43 u64 next_check;
44 unsigned long throttle_count;
45 unsigned long last_throttle_count;
46};
47
48static DEFINE_PER_CPU(struct thermal_state, thermal_state);
40 49
41static atomic_t therm_throt_en = ATOMIC_INIT(0); 50static atomic_t therm_throt_en = ATOMIC_INIT(0);
42 51
43#ifdef CONFIG_SYSFS 52#ifdef CONFIG_SYSFS
44#define define_therm_throt_sysdev_one_ro(_name) \ 53#define define_therm_throt_sysdev_one_ro(_name) \
45 static SYSDEV_ATTR(_name, 0444, therm_throt_sysdev_show_##_name, NULL) 54 static SYSDEV_ATTR(_name, 0444, therm_throt_sysdev_show_##_name, NULL)
46 55
47#define define_therm_throt_sysdev_show_func(name) \ 56#define define_therm_throt_sysdev_show_func(name) \
48static ssize_t therm_throt_sysdev_show_##name(struct sys_device *dev, \ 57 \
49 struct sysdev_attribute *attr, \ 58static ssize_t therm_throt_sysdev_show_##name( \
50 char *buf) \ 59 struct sys_device *dev, \
60 struct sysdev_attribute *attr, \
61 char *buf) \
51{ \ 62{ \
52 unsigned int cpu = dev->id; \ 63 unsigned int cpu = dev->id; \
53 ssize_t ret; \ 64 ssize_t ret; \
@@ -55,7 +66,7 @@ static ssize_t therm_throt_sysdev_show_##name(struct sys_device *dev, \
55 preempt_disable(); /* CPU hotplug */ \ 66 preempt_disable(); /* CPU hotplug */ \
56 if (cpu_online(cpu)) \ 67 if (cpu_online(cpu)) \
57 ret = sprintf(buf, "%lu\n", \ 68 ret = sprintf(buf, "%lu\n", \
58 per_cpu(thermal_throttle_##name, cpu)); \ 69 per_cpu(thermal_state, cpu).name); \
59 else \ 70 else \
60 ret = 0; \ 71 ret = 0; \
61 preempt_enable(); \ 72 preempt_enable(); \
@@ -63,11 +74,11 @@ static ssize_t therm_throt_sysdev_show_##name(struct sys_device *dev, \
63 return ret; \ 74 return ret; \
64} 75}
65 76
66define_therm_throt_sysdev_show_func(count); 77define_therm_throt_sysdev_show_func(throttle_count);
67define_therm_throt_sysdev_one_ro(count); 78define_therm_throt_sysdev_one_ro(throttle_count);
68 79
69static struct attribute *thermal_throttle_attrs[] = { 80static struct attribute *thermal_throttle_attrs[] = {
70 &attr_count.attr, 81 &attr_throttle_count.attr,
71 NULL 82 NULL
72}; 83};
73 84
@@ -93,33 +104,39 @@ static struct attribute_group thermal_throttle_attr_group = {
93 * 1 : Event should be logged further, and a message has been 104 * 1 : Event should be logged further, and a message has been
94 * printed to the syslog. 105 * printed to the syslog.
95 */ 106 */
96static int therm_throt_process(int curr) 107static int therm_throt_process(bool is_throttled)
97{ 108{
98 unsigned int cpu = smp_processor_id(); 109 struct thermal_state *state;
99 __u64 tmp_jiffs = get_jiffies_64(); 110 unsigned int this_cpu;
100 bool was_throttled = __get_cpu_var(thermal_throttle_active); 111 bool was_throttled;
101 bool is_throttled = __get_cpu_var(thermal_throttle_active) = curr; 112 u64 now;
113
114 this_cpu = smp_processor_id();
115 now = get_jiffies_64();
116 state = &per_cpu(thermal_state, this_cpu);
117
118 was_throttled = state->is_throttled;
119 state->is_throttled = is_throttled;
102 120
103 if (is_throttled) 121 if (is_throttled)
104 __get_cpu_var(thermal_throttle_count)++; 122 state->throttle_count++;
105 123
106 if (!(was_throttled ^ is_throttled) && 124 if (time_before64(now, state->next_check) &&
107 time_before64(tmp_jiffs, __get_cpu_var(next_check))) 125 state->throttle_count != state->last_throttle_count)
108 return 0; 126 return 0;
109 127
110 __get_cpu_var(next_check) = tmp_jiffs + CHECK_INTERVAL; 128 state->next_check = now + CHECK_INTERVAL;
129 state->last_throttle_count = state->throttle_count;
111 130
112 /* if we just entered the thermal event */ 131 /* if we just entered the thermal event */
113 if (is_throttled) { 132 if (is_throttled) {
114 printk(KERN_CRIT "CPU%d: Temperature above threshold, " 133 printk(KERN_CRIT "CPU%d: Temperature above threshold, cpu clock throttled (total events = %lu)\n", this_cpu, state->throttle_count);
115 "cpu clock throttled (total events = %lu)\n",
116 cpu, __get_cpu_var(thermal_throttle_count));
117 134
118 add_taint(TAINT_MACHINE_CHECK); 135 add_taint(TAINT_MACHINE_CHECK);
119 return 1; 136 return 1;
120 } 137 }
121 if (was_throttled) { 138 if (was_throttled) {
122 printk(KERN_INFO "CPU%d: Temperature/speed normal\n", cpu); 139 printk(KERN_INFO "CPU%d: Temperature/speed normal\n", this_cpu);
123 return 1; 140 return 1;
124 } 141 }
125 142
@@ -213,7 +230,7 @@ static void intel_thermal_interrupt(void)
213 __u64 msr_val; 230 __u64 msr_val;
214 231
215 rdmsrl(MSR_IA32_THERM_STATUS, msr_val); 232 rdmsrl(MSR_IA32_THERM_STATUS, msr_val);
216 if (therm_throt_process(msr_val & THERM_STATUS_PROCHOT)) 233 if (therm_throt_process((msr_val & THERM_STATUS_PROCHOT) != 0))
217 mce_log_therm_throt_event(msr_val); 234 mce_log_therm_throt_event(msr_val);
218} 235}
219 236
diff --git a/arch/x86/kernel/cpu/mtrr/if.c b/arch/x86/kernel/cpu/mtrr/if.c
index f04e72527604..3c1b12d461d1 100644
--- a/arch/x86/kernel/cpu/mtrr/if.c
+++ b/arch/x86/kernel/cpu/mtrr/if.c
@@ -96,17 +96,24 @@ mtrr_write(struct file *file, const char __user *buf, size_t len, loff_t * ppos)
96 unsigned long long base, size; 96 unsigned long long base, size;
97 char *ptr; 97 char *ptr;
98 char line[LINE_SIZE]; 98 char line[LINE_SIZE];
99 int length;
99 size_t linelen; 100 size_t linelen;
100 101
101 if (!capable(CAP_SYS_ADMIN)) 102 if (!capable(CAP_SYS_ADMIN))
102 return -EPERM; 103 return -EPERM;
103 if (!len)
104 return -EINVAL;
105 104
106 memset(line, 0, LINE_SIZE); 105 memset(line, 0, LINE_SIZE);
107 if (len > LINE_SIZE) 106
108 len = LINE_SIZE; 107 length = len;
109 if (copy_from_user(line, buf, len - 1)) 108 length--;
109
110 if (length > LINE_SIZE - 1)
111 length = LINE_SIZE - 1;
112
113 if (length < 0)
114 return -EINVAL;
115
116 if (copy_from_user(line, buf, length))
110 return -EFAULT; 117 return -EFAULT;
111 118
112 linelen = strlen(line); 119 linelen = strlen(line);
diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
index a3c7adb06b78..b5801c311846 100644
--- a/arch/x86/kernel/cpu/perf_event.c
+++ b/arch/x86/kernel/cpu/perf_event.c
@@ -1790,6 +1790,9 @@ void smp_perf_pending_interrupt(struct pt_regs *regs)
1790void set_perf_event_pending(void) 1790void set_perf_event_pending(void)
1791{ 1791{
1792#ifdef CONFIG_X86_LOCAL_APIC 1792#ifdef CONFIG_X86_LOCAL_APIC
1793 if (!x86_pmu.apic || !x86_pmu_initialized())
1794 return;
1795
1793 apic->send_IPI_self(LOCAL_PENDING_VECTOR); 1796 apic->send_IPI_self(LOCAL_PENDING_VECTOR);
1794#endif 1797#endif
1795} 1798}
diff --git a/arch/x86/kernel/dumpstack_32.c b/arch/x86/kernel/dumpstack_32.c
index bca5fba91c9e..f7dd2a7c3bf4 100644
--- a/arch/x86/kernel/dumpstack_32.c
+++ b/arch/x86/kernel/dumpstack_32.c
@@ -5,7 +5,6 @@
5#include <linux/kallsyms.h> 5#include <linux/kallsyms.h>
6#include <linux/kprobes.h> 6#include <linux/kprobes.h>
7#include <linux/uaccess.h> 7#include <linux/uaccess.h>
8#include <linux/utsname.h>
9#include <linux/hardirq.h> 8#include <linux/hardirq.h>
10#include <linux/kdebug.h> 9#include <linux/kdebug.h>
11#include <linux/module.h> 10#include <linux/module.h>
diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c
index 54b0a3276766..a071e6be177e 100644
--- a/arch/x86/kernel/dumpstack_64.c
+++ b/arch/x86/kernel/dumpstack_64.c
@@ -5,7 +5,6 @@
5#include <linux/kallsyms.h> 5#include <linux/kallsyms.h>
6#include <linux/kprobes.h> 6#include <linux/kprobes.h>
7#include <linux/uaccess.h> 7#include <linux/uaccess.h>
8#include <linux/utsname.h>
9#include <linux/hardirq.h> 8#include <linux/hardirq.h>
10#include <linux/kdebug.h> 9#include <linux/kdebug.h>
11#include <linux/module.h> 10#include <linux/module.h>
diff --git a/arch/x86/kernel/e820.c b/arch/x86/kernel/e820.c
index 85419bb7d4ab..d17d482a04f4 100644
--- a/arch/x86/kernel/e820.c
+++ b/arch/x86/kernel/e820.c
@@ -1378,8 +1378,8 @@ static unsigned long ram_alignment(resource_size_t pos)
1378 if (mb < 16) 1378 if (mb < 16)
1379 return 1024*1024; 1379 return 1024*1024;
1380 1380
1381 /* To 32MB for anything above that */ 1381 /* To 64MB for anything above that */
1382 return 32*1024*1024; 1382 return 64*1024*1024;
1383} 1383}
1384 1384
1385#define MAX_RESOURCE_SIZE ((resource_size_t)-1) 1385#define MAX_RESOURCE_SIZE ((resource_size_t)-1)
diff --git a/arch/x86/kernel/early_printk.c b/arch/x86/kernel/early_printk.c
index 2acfd3fdc0cc..b9c830c12b4a 100644
--- a/arch/x86/kernel/early_printk.c
+++ b/arch/x86/kernel/early_printk.c
@@ -178,6 +178,11 @@ asmlinkage void early_printk(const char *fmt, ...)
178 178
179static inline void early_console_register(struct console *con, int keep_early) 179static inline void early_console_register(struct console *con, int keep_early)
180{ 180{
181 if (early_console->index != -1) {
182 printk(KERN_CRIT "ERROR: earlyprintk= %s already used\n",
183 con->name);
184 return;
185 }
181 early_console = con; 186 early_console = con;
182 if (keep_early) 187 if (keep_early)
183 early_console->flags &= ~CON_BOOT; 188 early_console->flags &= ~CON_BOOT;
@@ -201,8 +206,11 @@ static int __init setup_early_printk(char *buf)
201 206
202 while (*buf != '\0') { 207 while (*buf != '\0') {
203 if (!strncmp(buf, "serial", 6)) { 208 if (!strncmp(buf, "serial", 6)) {
204 early_serial_init(buf + 6); 209 buf += 6;
210 early_serial_init(buf);
205 early_console_register(&early_serial_console, keep); 211 early_console_register(&early_serial_console, keep);
212 if (!strncmp(buf, ",ttyS", 5))
213 buf += 5;
206 } 214 }
207 if (!strncmp(buf, "ttyS", 4)) { 215 if (!strncmp(buf, "ttyS", 4)) {
208 early_serial_init(buf + 4); 216 early_serial_init(buf + 4);
diff --git a/arch/x86/kernel/head_32.S b/arch/x86/kernel/head_32.S
index b766e8c7252d..050c278481b1 100644
--- a/arch/x86/kernel/head_32.S
+++ b/arch/x86/kernel/head_32.S
@@ -79,7 +79,7 @@ RESERVE_BRK(pagetables, INIT_MAP_SIZE)
79 * any particular GDT layout, because we load our own as soon as we 79 * any particular GDT layout, because we load our own as soon as we
80 * can. 80 * can.
81 */ 81 */
82.section .text.head,"ax",@progbits 82__HEAD
83ENTRY(startup_32) 83ENTRY(startup_32)
84 /* test KEEP_SEGMENTS flag to see if the bootloader is asking 84 /* test KEEP_SEGMENTS flag to see if the bootloader is asking
85 us to not reload segments */ 85 us to not reload segments */
@@ -608,7 +608,7 @@ ENTRY(initial_code)
608/* 608/*
609 * BSS section 609 * BSS section
610 */ 610 */
611.section ".bss.page_aligned","wa" 611__PAGE_ALIGNED_BSS
612 .align PAGE_SIZE_asm 612 .align PAGE_SIZE_asm
613#ifdef CONFIG_X86_PAE 613#ifdef CONFIG_X86_PAE
614swapper_pg_pmd: 614swapper_pg_pmd:
@@ -626,7 +626,7 @@ ENTRY(empty_zero_page)
626 * This starts the data section. 626 * This starts the data section.
627 */ 627 */
628#ifdef CONFIG_X86_PAE 628#ifdef CONFIG_X86_PAE
629.section ".data.page_aligned","wa" 629__PAGE_ALIGNED_DATA
630 /* Page-aligned for the benefit of paravirt? */ 630 /* Page-aligned for the benefit of paravirt? */
631 .align PAGE_SIZE_asm 631 .align PAGE_SIZE_asm
632ENTRY(swapper_pg_dir) 632ENTRY(swapper_pg_dir)
diff --git a/arch/x86/kernel/head_64.S b/arch/x86/kernel/head_64.S
index fa54f78e2a05..780cd928fcd5 100644
--- a/arch/x86/kernel/head_64.S
+++ b/arch/x86/kernel/head_64.S
@@ -40,7 +40,7 @@ L4_START_KERNEL = pgd_index(__START_KERNEL_map)
40L3_START_KERNEL = pud_index(__START_KERNEL_map) 40L3_START_KERNEL = pud_index(__START_KERNEL_map)
41 41
42 .text 42 .text
43 .section .text.head 43 __HEAD
44 .code64 44 .code64
45 .globl startup_64 45 .globl startup_64
46startup_64: 46startup_64:
@@ -418,7 +418,7 @@ ENTRY(phys_base)
418ENTRY(idt_table) 418ENTRY(idt_table)
419 .skip IDT_ENTRIES * 16 419 .skip IDT_ENTRIES * 16
420 420
421 .section .bss.page_aligned, "aw", @nobits 421 __PAGE_ALIGNED_BSS
422 .align PAGE_SIZE 422 .align PAGE_SIZE
423ENTRY(empty_zero_page) 423ENTRY(empty_zero_page)
424 .skip PAGE_SIZE 424 .skip PAGE_SIZE
diff --git a/arch/x86/kernel/i386_ksyms_32.c b/arch/x86/kernel/i386_ksyms_32.c
index 43cec6bdda63..9c3bd4a2050e 100644
--- a/arch/x86/kernel/i386_ksyms_32.c
+++ b/arch/x86/kernel/i386_ksyms_32.c
@@ -10,6 +10,16 @@
10EXPORT_SYMBOL(mcount); 10EXPORT_SYMBOL(mcount);
11#endif 11#endif
12 12
13/*
14 * Note, this is a prototype to get at the symbol for
15 * the export, but dont use it from C code, it is used
16 * by assembly code and is not using C calling convention!
17 */
18#ifndef CONFIG_X86_CMPXCHG64
19extern void cmpxchg8b_emu(void);
20EXPORT_SYMBOL(cmpxchg8b_emu);
21#endif
22
13/* Networking helper routines. */ 23/* Networking helper routines. */
14EXPORT_SYMBOL(csum_partial_copy_generic); 24EXPORT_SYMBOL(csum_partial_copy_generic);
15 25
diff --git a/arch/x86/kernel/init_task.c b/arch/x86/kernel/init_task.c
index 270ff83efc11..3a54dcb9cd0e 100644
--- a/arch/x86/kernel/init_task.c
+++ b/arch/x86/kernel/init_task.c
@@ -20,9 +20,8 @@ static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand);
20 * way process stacks are handled. This is done by having a special 20 * way process stacks are handled. This is done by having a special
21 * "init_task" linker map entry.. 21 * "init_task" linker map entry..
22 */ 22 */
23union thread_union init_thread_union 23union thread_union init_thread_union __init_task_data =
24 __attribute__((__section__(".data.init_task"))) = 24 { INIT_THREAD_INFO(init_task) };
25 { INIT_THREAD_INFO(init_task) };
26 25
27/* 26/*
28 * Initial task structure. 27 * Initial task structure.
diff --git a/arch/x86/kernel/irq.c b/arch/x86/kernel/irq.c
index 74656d1d4e30..04bbd5278568 100644
--- a/arch/x86/kernel/irq.c
+++ b/arch/x86/kernel/irq.c
@@ -63,10 +63,10 @@ static int show_other_interrupts(struct seq_file *p, int prec)
63 for_each_online_cpu(j) 63 for_each_online_cpu(j)
64 seq_printf(p, "%10u ", irq_stats(j)->irq_spurious_count); 64 seq_printf(p, "%10u ", irq_stats(j)->irq_spurious_count);
65 seq_printf(p, " Spurious interrupts\n"); 65 seq_printf(p, " Spurious interrupts\n");
66 seq_printf(p, "%*s: ", prec, "CNT"); 66 seq_printf(p, "%*s: ", prec, "PMI");
67 for_each_online_cpu(j) 67 for_each_online_cpu(j)
68 seq_printf(p, "%10u ", irq_stats(j)->apic_perf_irqs); 68 seq_printf(p, "%10u ", irq_stats(j)->apic_perf_irqs);
69 seq_printf(p, " Performance counter interrupts\n"); 69 seq_printf(p, " Performance monitoring interrupts\n");
70 seq_printf(p, "%*s: ", prec, "PND"); 70 seq_printf(p, "%*s: ", prec, "PND");
71 for_each_online_cpu(j) 71 for_each_online_cpu(j)
72 seq_printf(p, "%10u ", irq_stats(j)->apic_pending_irqs); 72 seq_printf(p, "%10u ", irq_stats(j)->apic_pending_irqs);
diff --git a/arch/x86/kernel/ldt.c b/arch/x86/kernel/ldt.c
index 71f1d99a635d..ec6ef60cbd17 100644
--- a/arch/x86/kernel/ldt.c
+++ b/arch/x86/kernel/ldt.c
@@ -67,8 +67,8 @@ static int alloc_ldt(mm_context_t *pc, int mincount, int reload)
67#ifdef CONFIG_SMP 67#ifdef CONFIG_SMP
68 preempt_disable(); 68 preempt_disable();
69 load_LDT(pc); 69 load_LDT(pc);
70 if (!cpus_equal(current->mm->cpu_vm_mask, 70 if (!cpumask_equal(mm_cpumask(current->mm),
71 cpumask_of_cpu(smp_processor_id()))) 71 cpumask_of(smp_processor_id())))
72 smp_call_function(flush_ldt, current->mm, 1); 72 smp_call_function(flush_ldt, current->mm, 1);
73 preempt_enable(); 73 preempt_enable();
74#else 74#else
diff --git a/arch/x86/kernel/pci-dma.c b/arch/x86/kernel/pci-dma.c
index 64b838eac18c..b2a71dca5642 100644
--- a/arch/x86/kernel/pci-dma.c
+++ b/arch/x86/kernel/pci-dma.c
@@ -35,7 +35,7 @@ int iommu_detected __read_mostly = 0;
35 35
36/* 36/*
37 * This variable becomes 1 if iommu=pt is passed on the kernel command line. 37 * This variable becomes 1 if iommu=pt is passed on the kernel command line.
38 * If this variable is 1, IOMMU implementations do no DMA ranslation for 38 * If this variable is 1, IOMMU implementations do no DMA translation for
39 * devices and allow every device to access to whole physical memory. This is 39 * devices and allow every device to access to whole physical memory. This is
40 * useful if a user want to use an IOMMU only for KVM device assignment to 40 * useful if a user want to use an IOMMU only for KVM device assignment to
41 * guests and not for driver dma translation. 41 * guests and not for driver dma translation.
@@ -311,7 +311,7 @@ void pci_iommu_shutdown(void)
311 amd_iommu_shutdown(); 311 amd_iommu_shutdown();
312} 312}
313/* Must execute after PCI subsystem */ 313/* Must execute after PCI subsystem */
314fs_initcall(pci_iommu_init); 314rootfs_initcall(pci_iommu_init);
315 315
316#ifdef CONFIG_PCI 316#ifdef CONFIG_PCI
317/* Many VIA bridges seem to corrupt data for DAC. Disable it here */ 317/* Many VIA bridges seem to corrupt data for DAC. Disable it here */
diff --git a/arch/x86/kernel/pci-gart_64.c b/arch/x86/kernel/pci-gart_64.c
index 98a827ee9ed7..a7f1b64f86e0 100644
--- a/arch/x86/kernel/pci-gart_64.c
+++ b/arch/x86/kernel/pci-gart_64.c
@@ -16,6 +16,7 @@
16#include <linux/agp_backend.h> 16#include <linux/agp_backend.h>
17#include <linux/init.h> 17#include <linux/init.h>
18#include <linux/mm.h> 18#include <linux/mm.h>
19#include <linux/sched.h>
19#include <linux/string.h> 20#include <linux/string.h>
20#include <linux/spinlock.h> 21#include <linux/spinlock.h>
21#include <linux/pci.h> 22#include <linux/pci.h>
diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c
index 847ab4160315..5284cd2b5776 100644
--- a/arch/x86/kernel/process.c
+++ b/arch/x86/kernel/process.c
@@ -555,10 +555,8 @@ void __cpuinit select_idle_routine(const struct cpuinfo_x86 *c)
555void __init init_c1e_mask(void) 555void __init init_c1e_mask(void)
556{ 556{
557 /* If we're using c1e_idle, we need to allocate c1e_mask. */ 557 /* If we're using c1e_idle, we need to allocate c1e_mask. */
558 if (pm_idle == c1e_idle) { 558 if (pm_idle == c1e_idle)
559 alloc_cpumask_var(&c1e_mask, GFP_KERNEL); 559 zalloc_cpumask_var(&c1e_mask, GFP_KERNEL);
560 cpumask_clear(c1e_mask);
561 }
562} 560}
563 561
564static int __init idle_setup(char *str) 562static int __init idle_setup(char *str)
diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c
index 27349f92a6d7..a1a3cdda06e1 100644
--- a/arch/x86/kernel/reboot.c
+++ b/arch/x86/kernel/reboot.c
@@ -4,6 +4,7 @@
4#include <linux/pm.h> 4#include <linux/pm.h>
5#include <linux/efi.h> 5#include <linux/efi.h>
6#include <linux/dmi.h> 6#include <linux/dmi.h>
7#include <linux/sched.h>
7#include <linux/tboot.h> 8#include <linux/tboot.h>
8#include <acpi/reboot.h> 9#include <acpi/reboot.h>
9#include <asm/io.h> 10#include <asm/io.h>
diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
index 09c5e077dff7..565ebc65920e 100644
--- a/arch/x86/kernel/smpboot.c
+++ b/arch/x86/kernel/smpboot.c
@@ -1059,12 +1059,9 @@ void __init native_smp_prepare_cpus(unsigned int max_cpus)
1059#endif 1059#endif
1060 current_thread_info()->cpu = 0; /* needed? */ 1060 current_thread_info()->cpu = 0; /* needed? */
1061 for_each_possible_cpu(i) { 1061 for_each_possible_cpu(i) {
1062 alloc_cpumask_var(&per_cpu(cpu_sibling_map, i), GFP_KERNEL); 1062 zalloc_cpumask_var(&per_cpu(cpu_sibling_map, i), GFP_KERNEL);
1063 alloc_cpumask_var(&per_cpu(cpu_core_map, i), GFP_KERNEL); 1063 zalloc_cpumask_var(&per_cpu(cpu_core_map, i), GFP_KERNEL);
1064 alloc_cpumask_var(&cpu_data(i).llc_shared_map, GFP_KERNEL); 1064 zalloc_cpumask_var(&cpu_data(i).llc_shared_map, GFP_KERNEL);
1065 cpumask_clear(per_cpu(cpu_core_map, i));
1066 cpumask_clear(per_cpu(cpu_sibling_map, i));
1067 cpumask_clear(cpu_data(i).llc_shared_map);
1068 } 1065 }
1069 set_cpu_sibling_map(0); 1066 set_cpu_sibling_map(0);
1070 1067
diff --git a/arch/x86/kernel/time.c b/arch/x86/kernel/time.c
index e293ac56c723..be2573448ed9 100644
--- a/arch/x86/kernel/time.c
+++ b/arch/x86/kernel/time.c
@@ -38,7 +38,8 @@ unsigned long profile_pc(struct pt_regs *regs)
38#ifdef CONFIG_FRAME_POINTER 38#ifdef CONFIG_FRAME_POINTER
39 return *(unsigned long *)(regs->bp + sizeof(long)); 39 return *(unsigned long *)(regs->bp + sizeof(long));
40#else 40#else
41 unsigned long *sp = (unsigned long *)regs->sp; 41 unsigned long *sp =
42 (unsigned long *)kernel_stack_pointer(regs);
42 /* 43 /*
43 * Return address is either directly at stack pointer 44 * Return address is either directly at stack pointer
44 * or above a saved flags. Eflags has bits 22-31 zero, 45 * or above a saved flags. Eflags has bits 22-31 zero,
@@ -93,7 +94,6 @@ static struct irqaction irq0 = {
93 94
94void __init setup_default_timer_irq(void) 95void __init setup_default_timer_irq(void)
95{ 96{
96 irq0.mask = cpumask_of_cpu(0);
97 setup_irq(0, &irq0); 97 setup_irq(0, &irq0);
98} 98}
99 99
diff --git a/arch/x86/kernel/trampoline.c b/arch/x86/kernel/trampoline.c
index 699f7eeb896a..cd022121cab6 100644
--- a/arch/x86/kernel/trampoline.c
+++ b/arch/x86/kernel/trampoline.c
@@ -3,8 +3,16 @@
3#include <asm/trampoline.h> 3#include <asm/trampoline.h>
4#include <asm/e820.h> 4#include <asm/e820.h>
5 5
6#if defined(CONFIG_X86_64) && defined(CONFIG_ACPI_SLEEP)
7#define __trampinit
8#define __trampinitdata
9#else
10#define __trampinit __cpuinit
11#define __trampinitdata __cpuinitdata
12#endif
13
6/* ready for x86_64 and x86 */ 14/* ready for x86_64 and x86 */
7unsigned char *__cpuinitdata trampoline_base = __va(TRAMPOLINE_BASE); 15unsigned char *__trampinitdata trampoline_base = __va(TRAMPOLINE_BASE);
8 16
9void __init reserve_trampoline_memory(void) 17void __init reserve_trampoline_memory(void)
10{ 18{
@@ -26,7 +34,7 @@ void __init reserve_trampoline_memory(void)
26 * bootstrap into the page concerned. The caller 34 * bootstrap into the page concerned. The caller
27 * has made sure it's suitably aligned. 35 * has made sure it's suitably aligned.
28 */ 36 */
29unsigned long __cpuinit setup_trampoline(void) 37unsigned long __trampinit setup_trampoline(void)
30{ 38{
31 memcpy(trampoline_base, trampoline_data, TRAMPOLINE_SIZE); 39 memcpy(trampoline_base, trampoline_data, TRAMPOLINE_SIZE);
32 return virt_to_phys(trampoline_base); 40 return virt_to_phys(trampoline_base);
diff --git a/arch/x86/kernel/trampoline_64.S b/arch/x86/kernel/trampoline_64.S
index 596d54c660a5..3af2dff58b21 100644
--- a/arch/x86/kernel/trampoline_64.S
+++ b/arch/x86/kernel/trampoline_64.S
@@ -32,8 +32,12 @@
32#include <asm/segment.h> 32#include <asm/segment.h>
33#include <asm/processor-flags.h> 33#include <asm/processor-flags.h>
34 34
35#ifdef CONFIG_ACPI_SLEEP
36.section .rodata, "a", @progbits
37#else
35/* We can free up the trampoline after bootup if cpu hotplug is not supported. */ 38/* We can free up the trampoline after bootup if cpu hotplug is not supported. */
36__CPUINITRODATA 39__CPUINITRODATA
40#endif
37.code16 41.code16
38 42
39ENTRY(trampoline_data) 43ENTRY(trampoline_data)
diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
index 9346e102338d..7e37dcee0cc3 100644
--- a/arch/x86/kernel/traps.c
+++ b/arch/x86/kernel/traps.c
@@ -14,7 +14,6 @@
14#include <linux/spinlock.h> 14#include <linux/spinlock.h>
15#include <linux/kprobes.h> 15#include <linux/kprobes.h>
16#include <linux/uaccess.h> 16#include <linux/uaccess.h>
17#include <linux/utsname.h>
18#include <linux/kdebug.h> 17#include <linux/kdebug.h>
19#include <linux/kernel.h> 18#include <linux/kernel.h>
20#include <linux/module.h> 19#include <linux/module.h>
@@ -73,11 +72,9 @@ char ignore_fpu_irq;
73 72
74/* 73/*
75 * The IDT has to be page-aligned to simplify the Pentium 74 * The IDT has to be page-aligned to simplify the Pentium
76 * F0 0F bug workaround.. We have a special link segment 75 * F0 0F bug workaround.
77 * for this.
78 */ 76 */
79gate_desc idt_table[NR_VECTORS] 77gate_desc idt_table[NR_VECTORS] __page_aligned_data = { { { { 0, 0 } } }, };
80 __attribute__((__section__(".data.idt"))) = { { { { 0, 0 } } }, };
81#endif 78#endif
82 79
83DECLARE_BITMAP(used_vectors, NR_VECTORS); 80DECLARE_BITMAP(used_vectors, NR_VECTORS);
diff --git a/arch/x86/kernel/tsc_sync.c b/arch/x86/kernel/tsc_sync.c
index 027b5b498993..f37930954d15 100644
--- a/arch/x86/kernel/tsc_sync.c
+++ b/arch/x86/kernel/tsc_sync.c
@@ -114,7 +114,7 @@ void __cpuinit check_tsc_sync_source(int cpu)
114 return; 114 return;
115 115
116 if (boot_cpu_has(X86_FEATURE_TSC_RELIABLE)) { 116 if (boot_cpu_has(X86_FEATURE_TSC_RELIABLE)) {
117 pr_info("Skipping synchronization checks as TSC is reliable.\n"); 117 printk_once(KERN_INFO "Skipping synchronization checks as TSC is reliable.\n");
118 return; 118 return;
119 } 119 }
120 120
diff --git a/arch/x86/kernel/vmi_32.c b/arch/x86/kernel/vmi_32.c
index 31e6f6cfe53e..d430e4c30193 100644
--- a/arch/x86/kernel/vmi_32.c
+++ b/arch/x86/kernel/vmi_32.c
@@ -648,7 +648,7 @@ static inline int __init activate_vmi(void)
648 648
649 pv_info.paravirt_enabled = 1; 649 pv_info.paravirt_enabled = 1;
650 pv_info.kernel_rpl = kernel_cs & SEGMENT_RPL_MASK; 650 pv_info.kernel_rpl = kernel_cs & SEGMENT_RPL_MASK;
651 pv_info.name = "vmi"; 651 pv_info.name = "vmi [deprecated]";
652 652
653 pv_init_ops.patch = vmi_patch; 653 pv_init_ops.patch = vmi_patch;
654 654
diff --git a/arch/x86/kernel/vmlinux.lds.S b/arch/x86/kernel/vmlinux.lds.S
index a46acccec38a..92929fb3f9fa 100644
--- a/arch/x86/kernel/vmlinux.lds.S
+++ b/arch/x86/kernel/vmlinux.lds.S
@@ -65,17 +65,11 @@ SECTIONS
65#endif 65#endif
66 66
67 /* Text and read-only data */ 67 /* Text and read-only data */
68
69 /* bootstrapping code */
70 .text.head : AT(ADDR(.text.head) - LOAD_OFFSET) {
71 _text = .;
72 *(.text.head)
73 } :text = 0x9090
74
75 /* The rest of the text */
76 .text : AT(ADDR(.text) - LOAD_OFFSET) { 68 .text : AT(ADDR(.text) - LOAD_OFFSET) {
69 _text = .;
70 /* bootstrapping code */
71 HEAD_TEXT
77#ifdef CONFIG_X86_32 72#ifdef CONFIG_X86_32
78 /* not really needed, already page aligned */
79 . = ALIGN(PAGE_SIZE); 73 . = ALIGN(PAGE_SIZE);
80 *(.text.page_aligned) 74 *(.text.page_aligned)
81#endif 75#endif
@@ -94,13 +88,7 @@ SECTIONS
94 88
95 NOTES :text :note 89 NOTES :text :note
96 90
97 /* Exception table */ 91 EXCEPTION_TABLE(16) :text = 0x9090
98 . = ALIGN(16);
99 __ex_table : AT(ADDR(__ex_table) - LOAD_OFFSET) {
100 __start___ex_table = .;
101 *(__ex_table)
102 __stop___ex_table = .;
103 } :text = 0x9090
104 92
105 RO_DATA(PAGE_SIZE) 93 RO_DATA(PAGE_SIZE)
106 94
@@ -118,7 +106,6 @@ SECTIONS
118#endif 106#endif
119 107
120 PAGE_ALIGNED_DATA(PAGE_SIZE) 108 PAGE_ALIGNED_DATA(PAGE_SIZE)
121 *(.data.idt)
122 109
123 CACHELINE_ALIGNED_DATA(CONFIG_X86_L1_CACHE_BYTES) 110 CACHELINE_ALIGNED_DATA(CONFIG_X86_L1_CACHE_BYTES)
124 111
@@ -135,24 +122,21 @@ SECTIONS
135#ifdef CONFIG_X86_64 122#ifdef CONFIG_X86_64
136 123
137#define VSYSCALL_ADDR (-10*1024*1024) 124#define VSYSCALL_ADDR (-10*1024*1024)
138#define VSYSCALL_PHYS_ADDR ((LOADADDR(.data) + SIZEOF(.data) + \
139 PAGE_SIZE - 1) & ~(PAGE_SIZE - 1))
140#define VSYSCALL_VIRT_ADDR ((ADDR(.data) + SIZEOF(.data) + \
141 PAGE_SIZE - 1) & ~(PAGE_SIZE - 1))
142 125
143#define VLOAD_OFFSET (VSYSCALL_ADDR - VSYSCALL_PHYS_ADDR) 126#define VLOAD_OFFSET (VSYSCALL_ADDR - __vsyscall_0 + LOAD_OFFSET)
144#define VLOAD(x) (ADDR(x) - VLOAD_OFFSET) 127#define VLOAD(x) (ADDR(x) - VLOAD_OFFSET)
145 128
146#define VVIRT_OFFSET (VSYSCALL_ADDR - VSYSCALL_VIRT_ADDR) 129#define VVIRT_OFFSET (VSYSCALL_ADDR - __vsyscall_0)
147#define VVIRT(x) (ADDR(x) - VVIRT_OFFSET) 130#define VVIRT(x) (ADDR(x) - VVIRT_OFFSET)
148 131
132 . = ALIGN(4096);
133 __vsyscall_0 = .;
134
149 . = VSYSCALL_ADDR; 135 . = VSYSCALL_ADDR;
150 .vsyscall_0 : AT(VSYSCALL_PHYS_ADDR) { 136 .vsyscall_0 : AT(VLOAD(.vsyscall_0)) {
151 *(.vsyscall_0) 137 *(.vsyscall_0)
152 } :user 138 } :user
153 139
154 __vsyscall_0 = VSYSCALL_VIRT_ADDR;
155
156 . = ALIGN(CONFIG_X86_L1_CACHE_BYTES); 140 . = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
157 .vsyscall_fn : AT(VLOAD(.vsyscall_fn)) { 141 .vsyscall_fn : AT(VLOAD(.vsyscall_fn)) {
158 *(.vsyscall_fn) 142 *(.vsyscall_fn)
@@ -192,11 +176,9 @@ SECTIONS
192 *(.vsyscall_3) 176 *(.vsyscall_3)
193 } 177 }
194 178
195 . = VSYSCALL_VIRT_ADDR + PAGE_SIZE; 179 . = __vsyscall_0 + PAGE_SIZE;
196 180
197#undef VSYSCALL_ADDR 181#undef VSYSCALL_ADDR
198#undef VSYSCALL_PHYS_ADDR
199#undef VSYSCALL_VIRT_ADDR
200#undef VLOAD_OFFSET 182#undef VLOAD_OFFSET
201#undef VLOAD 183#undef VLOAD
202#undef VVIRT_OFFSET 184#undef VVIRT_OFFSET
@@ -219,36 +201,12 @@ SECTIONS
219 PERCPU_VADDR(0, :percpu) 201 PERCPU_VADDR(0, :percpu)
220#endif 202#endif
221 203
222 .init.text : AT(ADDR(.init.text) - LOAD_OFFSET) { 204 INIT_TEXT_SECTION(PAGE_SIZE)
223 _sinittext = .;
224 INIT_TEXT
225 _einittext = .;
226 }
227#ifdef CONFIG_X86_64 205#ifdef CONFIG_X86_64
228 :init 206 :init
229#endif 207#endif
230 208
231 .init.data : AT(ADDR(.init.data) - LOAD_OFFSET) { 209 INIT_DATA_SECTION(16)
232 INIT_DATA
233 }
234
235 . = ALIGN(16);
236 .init.setup : AT(ADDR(.init.setup) - LOAD_OFFSET) {
237 __setup_start = .;
238 *(.init.setup)
239 __setup_end = .;
240 }
241 .initcall.init : AT(ADDR(.initcall.init) - LOAD_OFFSET) {
242 __initcall_start = .;
243 INITCALLS
244 __initcall_end = .;
245 }
246
247 .con_initcall.init : AT(ADDR(.con_initcall.init) - LOAD_OFFSET) {
248 __con_initcall_start = .;
249 *(.con_initcall.init)
250 __con_initcall_end = .;
251 }
252 210
253 .x86_cpu_dev.init : AT(ADDR(.x86_cpu_dev.init) - LOAD_OFFSET) { 211 .x86_cpu_dev.init : AT(ADDR(.x86_cpu_dev.init) - LOAD_OFFSET) {
254 __x86_cpu_dev_start = .; 212 __x86_cpu_dev_start = .;
@@ -256,8 +214,6 @@ SECTIONS
256 __x86_cpu_dev_end = .; 214 __x86_cpu_dev_end = .;
257 } 215 }
258 216
259 SECURITY_INIT
260
261 . = ALIGN(8); 217 . = ALIGN(8);
262 .parainstructions : AT(ADDR(.parainstructions) - LOAD_OFFSET) { 218 .parainstructions : AT(ADDR(.parainstructions) - LOAD_OFFSET) {
263 __parainstructions = .; 219 __parainstructions = .;
@@ -288,15 +244,6 @@ SECTIONS
288 EXIT_DATA 244 EXIT_DATA
289 } 245 }
290 246
291#ifdef CONFIG_BLK_DEV_INITRD
292 . = ALIGN(PAGE_SIZE);
293 .init.ramfs : AT(ADDR(.init.ramfs) - LOAD_OFFSET) {
294 __initramfs_start = .;
295 *(.init.ramfs)
296 __initramfs_end = .;
297 }
298#endif
299
300#if !defined(CONFIG_X86_64) || !defined(CONFIG_SMP) 247#if !defined(CONFIG_X86_64) || !defined(CONFIG_SMP)
301 PERCPU(PAGE_SIZE) 248 PERCPU(PAGE_SIZE)
302#endif 249#endif
diff --git a/arch/x86/kernel/vsyscall_64.c b/arch/x86/kernel/vsyscall_64.c
index cf53a78e2dcf..8cb4974ff599 100644
--- a/arch/x86/kernel/vsyscall_64.c
+++ b/arch/x86/kernel/vsyscall_64.c
@@ -228,19 +228,11 @@ static long __vsyscall(3) venosys_1(void)
228} 228}
229 229
230#ifdef CONFIG_SYSCTL 230#ifdef CONFIG_SYSCTL
231
232static int
233vsyscall_sysctl_change(ctl_table *ctl, int write, struct file * filp,
234 void __user *buffer, size_t *lenp, loff_t *ppos)
235{
236 return proc_dointvec(ctl, write, filp, buffer, lenp, ppos);
237}
238
239static ctl_table kernel_table2[] = { 231static ctl_table kernel_table2[] = {
240 { .procname = "vsyscall64", 232 { .procname = "vsyscall64",
241 .data = &vsyscall_gtod_data.sysctl_enabled, .maxlen = sizeof(int), 233 .data = &vsyscall_gtod_data.sysctl_enabled, .maxlen = sizeof(int),
242 .mode = 0644, 234 .mode = 0644,
243 .proc_handler = vsyscall_sysctl_change }, 235 .proc_handler = proc_dointvec },
244 {} 236 {}
245}; 237};
246 238
diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c
index 1ae5ceba7eb2..7024224f0fc8 100644
--- a/arch/x86/kvm/lapic.c
+++ b/arch/x86/kvm/lapic.c
@@ -664,7 +664,7 @@ static void start_apic_timer(struct kvm_lapic *apic)
664{ 664{
665 ktime_t now = apic->lapic_timer.timer.base->get_time(); 665 ktime_t now = apic->lapic_timer.timer.base->get_time();
666 666
667 apic->lapic_timer.period = apic_get_reg(apic, APIC_TMICT) * 667 apic->lapic_timer.period = (u64)apic_get_reg(apic, APIC_TMICT) *
668 APIC_BUS_CYCLE_NS * apic->divide_count; 668 APIC_BUS_CYCLE_NS * apic->divide_count;
669 atomic_set(&apic->lapic_timer.pending, 0); 669 atomic_set(&apic->lapic_timer.pending, 0);
670 670
diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index eca41ae9f453..685a4ffac8e6 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -156,6 +156,8 @@ module_param(oos_shadow, bool, 0644);
156#define CREATE_TRACE_POINTS 156#define CREATE_TRACE_POINTS
157#include "mmutrace.h" 157#include "mmutrace.h"
158 158
159#define SPTE_HOST_WRITEABLE (1ULL << PT_FIRST_AVAIL_BITS_SHIFT)
160
159#define SHADOW_PT_INDEX(addr, level) PT64_INDEX(addr, level) 161#define SHADOW_PT_INDEX(addr, level) PT64_INDEX(addr, level)
160 162
161struct kvm_rmap_desc { 163struct kvm_rmap_desc {
@@ -634,9 +636,7 @@ static void rmap_remove(struct kvm *kvm, u64 *spte)
634 if (*spte & shadow_accessed_mask) 636 if (*spte & shadow_accessed_mask)
635 kvm_set_pfn_accessed(pfn); 637 kvm_set_pfn_accessed(pfn);
636 if (is_writeble_pte(*spte)) 638 if (is_writeble_pte(*spte))
637 kvm_release_pfn_dirty(pfn); 639 kvm_set_pfn_dirty(pfn);
638 else
639 kvm_release_pfn_clean(pfn);
640 rmapp = gfn_to_rmap(kvm, sp->gfns[spte - sp->spt], sp->role.level); 640 rmapp = gfn_to_rmap(kvm, sp->gfns[spte - sp->spt], sp->role.level);
641 if (!*rmapp) { 641 if (!*rmapp) {
642 printk(KERN_ERR "rmap_remove: %p %llx 0->BUG\n", spte, *spte); 642 printk(KERN_ERR "rmap_remove: %p %llx 0->BUG\n", spte, *spte);
@@ -748,7 +748,7 @@ static int rmap_write_protect(struct kvm *kvm, u64 gfn)
748 return write_protected; 748 return write_protected;
749} 749}
750 750
751static int kvm_unmap_rmapp(struct kvm *kvm, unsigned long *rmapp) 751static int kvm_unmap_rmapp(struct kvm *kvm, unsigned long *rmapp, u64 data)
752{ 752{
753 u64 *spte; 753 u64 *spte;
754 int need_tlb_flush = 0; 754 int need_tlb_flush = 0;
@@ -763,8 +763,45 @@ static int kvm_unmap_rmapp(struct kvm *kvm, unsigned long *rmapp)
763 return need_tlb_flush; 763 return need_tlb_flush;
764} 764}
765 765
766static int kvm_handle_hva(struct kvm *kvm, unsigned long hva, 766static int kvm_set_pte_rmapp(struct kvm *kvm, unsigned long *rmapp, u64 data)
767 int (*handler)(struct kvm *kvm, unsigned long *rmapp)) 767{
768 int need_flush = 0;
769 u64 *spte, new_spte;
770 pte_t *ptep = (pte_t *)data;
771 pfn_t new_pfn;
772
773 WARN_ON(pte_huge(*ptep));
774 new_pfn = pte_pfn(*ptep);
775 spte = rmap_next(kvm, rmapp, NULL);
776 while (spte) {
777 BUG_ON(!is_shadow_present_pte(*spte));
778 rmap_printk("kvm_set_pte_rmapp: spte %p %llx\n", spte, *spte);
779 need_flush = 1;
780 if (pte_write(*ptep)) {
781 rmap_remove(kvm, spte);
782 __set_spte(spte, shadow_trap_nonpresent_pte);
783 spte = rmap_next(kvm, rmapp, NULL);
784 } else {
785 new_spte = *spte &~ (PT64_BASE_ADDR_MASK);
786 new_spte |= (u64)new_pfn << PAGE_SHIFT;
787
788 new_spte &= ~PT_WRITABLE_MASK;
789 new_spte &= ~SPTE_HOST_WRITEABLE;
790 if (is_writeble_pte(*spte))
791 kvm_set_pfn_dirty(spte_to_pfn(*spte));
792 __set_spte(spte, new_spte);
793 spte = rmap_next(kvm, rmapp, spte);
794 }
795 }
796 if (need_flush)
797 kvm_flush_remote_tlbs(kvm);
798
799 return 0;
800}
801
802static int kvm_handle_hva(struct kvm *kvm, unsigned long hva, u64 data,
803 int (*handler)(struct kvm *kvm, unsigned long *rmapp,
804 u64 data))
768{ 805{
769 int i, j; 806 int i, j;
770 int retval = 0; 807 int retval = 0;
@@ -786,13 +823,15 @@ static int kvm_handle_hva(struct kvm *kvm, unsigned long hva,
786 if (hva >= start && hva < end) { 823 if (hva >= start && hva < end) {
787 gfn_t gfn_offset = (hva - start) >> PAGE_SHIFT; 824 gfn_t gfn_offset = (hva - start) >> PAGE_SHIFT;
788 825
789 retval |= handler(kvm, &memslot->rmap[gfn_offset]); 826 retval |= handler(kvm, &memslot->rmap[gfn_offset],
827 data);
790 828
791 for (j = 0; j < KVM_NR_PAGE_SIZES - 1; ++j) { 829 for (j = 0; j < KVM_NR_PAGE_SIZES - 1; ++j) {
792 int idx = gfn_offset; 830 int idx = gfn_offset;
793 idx /= KVM_PAGES_PER_HPAGE(PT_DIRECTORY_LEVEL + j); 831 idx /= KVM_PAGES_PER_HPAGE(PT_DIRECTORY_LEVEL + j);
794 retval |= handler(kvm, 832 retval |= handler(kvm,
795 &memslot->lpage_info[j][idx].rmap_pde); 833 &memslot->lpage_info[j][idx].rmap_pde,
834 data);
796 } 835 }
797 } 836 }
798 } 837 }
@@ -802,10 +841,15 @@ static int kvm_handle_hva(struct kvm *kvm, unsigned long hva,
802 841
803int kvm_unmap_hva(struct kvm *kvm, unsigned long hva) 842int kvm_unmap_hva(struct kvm *kvm, unsigned long hva)
804{ 843{
805 return kvm_handle_hva(kvm, hva, kvm_unmap_rmapp); 844 return kvm_handle_hva(kvm, hva, 0, kvm_unmap_rmapp);
806} 845}
807 846
808static int kvm_age_rmapp(struct kvm *kvm, unsigned long *rmapp) 847void kvm_set_spte_hva(struct kvm *kvm, unsigned long hva, pte_t pte)
848{
849 kvm_handle_hva(kvm, hva, (u64)&pte, kvm_set_pte_rmapp);
850}
851
852static int kvm_age_rmapp(struct kvm *kvm, unsigned long *rmapp, u64 data)
809{ 853{
810 u64 *spte; 854 u64 *spte;
811 int young = 0; 855 int young = 0;
@@ -841,13 +885,13 @@ static void rmap_recycle(struct kvm_vcpu *vcpu, u64 *spte, gfn_t gfn)
841 gfn = unalias_gfn(vcpu->kvm, gfn); 885 gfn = unalias_gfn(vcpu->kvm, gfn);
842 rmapp = gfn_to_rmap(vcpu->kvm, gfn, sp->role.level); 886 rmapp = gfn_to_rmap(vcpu->kvm, gfn, sp->role.level);
843 887
844 kvm_unmap_rmapp(vcpu->kvm, rmapp); 888 kvm_unmap_rmapp(vcpu->kvm, rmapp, 0);
845 kvm_flush_remote_tlbs(vcpu->kvm); 889 kvm_flush_remote_tlbs(vcpu->kvm);
846} 890}
847 891
848int kvm_age_hva(struct kvm *kvm, unsigned long hva) 892int kvm_age_hva(struct kvm *kvm, unsigned long hva)
849{ 893{
850 return kvm_handle_hva(kvm, hva, kvm_age_rmapp); 894 return kvm_handle_hva(kvm, hva, 0, kvm_age_rmapp);
851} 895}
852 896
853#ifdef MMU_DEBUG 897#ifdef MMU_DEBUG
@@ -1756,7 +1800,7 @@ static int set_spte(struct kvm_vcpu *vcpu, u64 *sptep,
1756 unsigned pte_access, int user_fault, 1800 unsigned pte_access, int user_fault,
1757 int write_fault, int dirty, int level, 1801 int write_fault, int dirty, int level,
1758 gfn_t gfn, pfn_t pfn, bool speculative, 1802 gfn_t gfn, pfn_t pfn, bool speculative,
1759 bool can_unsync) 1803 bool can_unsync, bool reset_host_protection)
1760{ 1804{
1761 u64 spte; 1805 u64 spte;
1762 int ret = 0; 1806 int ret = 0;
@@ -1783,6 +1827,9 @@ static int set_spte(struct kvm_vcpu *vcpu, u64 *sptep,
1783 spte |= kvm_x86_ops->get_mt_mask(vcpu, gfn, 1827 spte |= kvm_x86_ops->get_mt_mask(vcpu, gfn,
1784 kvm_is_mmio_pfn(pfn)); 1828 kvm_is_mmio_pfn(pfn));
1785 1829
1830 if (reset_host_protection)
1831 spte |= SPTE_HOST_WRITEABLE;
1832
1786 spte |= (u64)pfn << PAGE_SHIFT; 1833 spte |= (u64)pfn << PAGE_SHIFT;
1787 1834
1788 if ((pte_access & ACC_WRITE_MASK) 1835 if ((pte_access & ACC_WRITE_MASK)
@@ -1828,7 +1875,8 @@ static void mmu_set_spte(struct kvm_vcpu *vcpu, u64 *sptep,
1828 unsigned pt_access, unsigned pte_access, 1875 unsigned pt_access, unsigned pte_access,
1829 int user_fault, int write_fault, int dirty, 1876 int user_fault, int write_fault, int dirty,
1830 int *ptwrite, int level, gfn_t gfn, 1877 int *ptwrite, int level, gfn_t gfn,
1831 pfn_t pfn, bool speculative) 1878 pfn_t pfn, bool speculative,
1879 bool reset_host_protection)
1832{ 1880{
1833 int was_rmapped = 0; 1881 int was_rmapped = 0;
1834 int was_writeble = is_writeble_pte(*sptep); 1882 int was_writeble = is_writeble_pte(*sptep);
@@ -1860,7 +1908,8 @@ static void mmu_set_spte(struct kvm_vcpu *vcpu, u64 *sptep,
1860 } 1908 }
1861 1909
1862 if (set_spte(vcpu, sptep, pte_access, user_fault, write_fault, 1910 if (set_spte(vcpu, sptep, pte_access, user_fault, write_fault,
1863 dirty, level, gfn, pfn, speculative, true)) { 1911 dirty, level, gfn, pfn, speculative, true,
1912 reset_host_protection)) {
1864 if (write_fault) 1913 if (write_fault)
1865 *ptwrite = 1; 1914 *ptwrite = 1;
1866 kvm_x86_ops->tlb_flush(vcpu); 1915 kvm_x86_ops->tlb_flush(vcpu);
@@ -1877,8 +1926,7 @@ static void mmu_set_spte(struct kvm_vcpu *vcpu, u64 *sptep,
1877 page_header_update_slot(vcpu->kvm, sptep, gfn); 1926 page_header_update_slot(vcpu->kvm, sptep, gfn);
1878 if (!was_rmapped) { 1927 if (!was_rmapped) {
1879 rmap_count = rmap_add(vcpu, sptep, gfn); 1928 rmap_count = rmap_add(vcpu, sptep, gfn);
1880 if (!is_rmap_spte(*sptep)) 1929 kvm_release_pfn_clean(pfn);
1881 kvm_release_pfn_clean(pfn);
1882 if (rmap_count > RMAP_RECYCLE_THRESHOLD) 1930 if (rmap_count > RMAP_RECYCLE_THRESHOLD)
1883 rmap_recycle(vcpu, sptep, gfn); 1931 rmap_recycle(vcpu, sptep, gfn);
1884 } else { 1932 } else {
@@ -1909,7 +1957,7 @@ static int __direct_map(struct kvm_vcpu *vcpu, gpa_t v, int write,
1909 if (iterator.level == level) { 1957 if (iterator.level == level) {
1910 mmu_set_spte(vcpu, iterator.sptep, ACC_ALL, ACC_ALL, 1958 mmu_set_spte(vcpu, iterator.sptep, ACC_ALL, ACC_ALL,
1911 0, write, 1, &pt_write, 1959 0, write, 1, &pt_write,
1912 level, gfn, pfn, false); 1960 level, gfn, pfn, false, true);
1913 ++vcpu->stat.pf_fixed; 1961 ++vcpu->stat.pf_fixed;
1914 break; 1962 break;
1915 } 1963 }
diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h
index d2fec9c12d22..72558f8ff3f5 100644
--- a/arch/x86/kvm/paging_tmpl.h
+++ b/arch/x86/kvm/paging_tmpl.h
@@ -273,9 +273,13 @@ static void FNAME(update_pte)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *page,
273 if (mmu_notifier_retry(vcpu, vcpu->arch.update_pte.mmu_seq)) 273 if (mmu_notifier_retry(vcpu, vcpu->arch.update_pte.mmu_seq))
274 return; 274 return;
275 kvm_get_pfn(pfn); 275 kvm_get_pfn(pfn);
276 /*
277 * we call mmu_set_spte() with reset_host_protection = true beacuse that
278 * vcpu->arch.update_pte.pfn was fetched from get_user_pages(write = 1).
279 */
276 mmu_set_spte(vcpu, spte, page->role.access, pte_access, 0, 0, 280 mmu_set_spte(vcpu, spte, page->role.access, pte_access, 0, 0,
277 gpte & PT_DIRTY_MASK, NULL, PT_PAGE_TABLE_LEVEL, 281 gpte & PT_DIRTY_MASK, NULL, PT_PAGE_TABLE_LEVEL,
278 gpte_to_gfn(gpte), pfn, true); 282 gpte_to_gfn(gpte), pfn, true, true);
279} 283}
280 284
281/* 285/*
@@ -308,7 +312,7 @@ static u64 *FNAME(fetch)(struct kvm_vcpu *vcpu, gva_t addr,
308 user_fault, write_fault, 312 user_fault, write_fault,
309 gw->ptes[gw->level-1] & PT_DIRTY_MASK, 313 gw->ptes[gw->level-1] & PT_DIRTY_MASK,
310 ptwrite, level, 314 ptwrite, level,
311 gw->gfn, pfn, false); 315 gw->gfn, pfn, false, true);
312 break; 316 break;
313 } 317 }
314 318
@@ -558,6 +562,7 @@ static void FNAME(prefetch_page)(struct kvm_vcpu *vcpu,
558static int FNAME(sync_page)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp) 562static int FNAME(sync_page)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp)
559{ 563{
560 int i, offset, nr_present; 564 int i, offset, nr_present;
565 bool reset_host_protection;
561 566
562 offset = nr_present = 0; 567 offset = nr_present = 0;
563 568
@@ -595,9 +600,16 @@ static int FNAME(sync_page)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp)
595 600
596 nr_present++; 601 nr_present++;
597 pte_access = sp->role.access & FNAME(gpte_access)(vcpu, gpte); 602 pte_access = sp->role.access & FNAME(gpte_access)(vcpu, gpte);
603 if (!(sp->spt[i] & SPTE_HOST_WRITEABLE)) {
604 pte_access &= ~ACC_WRITE_MASK;
605 reset_host_protection = 0;
606 } else {
607 reset_host_protection = 1;
608 }
598 set_spte(vcpu, &sp->spt[i], pte_access, 0, 0, 609 set_spte(vcpu, &sp->spt[i], pte_access, 0, 0,
599 is_dirty_gpte(gpte), PT_PAGE_TABLE_LEVEL, gfn, 610 is_dirty_gpte(gpte), PT_PAGE_TABLE_LEVEL, gfn,
600 spte_to_pfn(sp->spt[i]), true, false); 611 spte_to_pfn(sp->spt[i]), true, false,
612 reset_host_protection);
601 } 613 }
602 614
603 return !nr_present; 615 return !nr_present;
diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
index 944cc9c04b3c..c17404add91f 100644
--- a/arch/x86/kvm/svm.c
+++ b/arch/x86/kvm/svm.c
@@ -767,6 +767,8 @@ static void svm_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
767 rdtscll(tsc_this); 767 rdtscll(tsc_this);
768 delta = vcpu->arch.host_tsc - tsc_this; 768 delta = vcpu->arch.host_tsc - tsc_this;
769 svm->vmcb->control.tsc_offset += delta; 769 svm->vmcb->control.tsc_offset += delta;
770 if (is_nested(svm))
771 svm->nested.hsave->control.tsc_offset += delta;
770 vcpu->cpu = cpu; 772 vcpu->cpu = cpu;
771 kvm_migrate_timers(vcpu); 773 kvm_migrate_timers(vcpu);
772 svm->asid_generation = 0; 774 svm->asid_generation = 0;
@@ -2057,10 +2059,14 @@ static int svm_get_msr(struct kvm_vcpu *vcpu, unsigned ecx, u64 *data)
2057 2059
2058 switch (ecx) { 2060 switch (ecx) {
2059 case MSR_IA32_TSC: { 2061 case MSR_IA32_TSC: {
2060 u64 tsc; 2062 u64 tsc_offset;
2063
2064 if (is_nested(svm))
2065 tsc_offset = svm->nested.hsave->control.tsc_offset;
2066 else
2067 tsc_offset = svm->vmcb->control.tsc_offset;
2061 2068
2062 rdtscll(tsc); 2069 *data = tsc_offset + native_read_tsc();
2063 *data = svm->vmcb->control.tsc_offset + tsc;
2064 break; 2070 break;
2065 } 2071 }
2066 case MSR_K6_STAR: 2072 case MSR_K6_STAR:
@@ -2146,10 +2152,17 @@ static int svm_set_msr(struct kvm_vcpu *vcpu, unsigned ecx, u64 data)
2146 2152
2147 switch (ecx) { 2153 switch (ecx) {
2148 case MSR_IA32_TSC: { 2154 case MSR_IA32_TSC: {
2149 u64 tsc; 2155 u64 tsc_offset = data - native_read_tsc();
2156 u64 g_tsc_offset = 0;
2157
2158 if (is_nested(svm)) {
2159 g_tsc_offset = svm->vmcb->control.tsc_offset -
2160 svm->nested.hsave->control.tsc_offset;
2161 svm->nested.hsave->control.tsc_offset = tsc_offset;
2162 }
2163
2164 svm->vmcb->control.tsc_offset = tsc_offset + g_tsc_offset;
2150 2165
2151 rdtscll(tsc);
2152 svm->vmcb->control.tsc_offset = data - tsc;
2153 break; 2166 break;
2154 } 2167 }
2155 case MSR_K6_STAR: 2168 case MSR_K6_STAR:
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index f3812014bd0b..ed53b42caba1 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -709,7 +709,7 @@ static void vmx_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
709 if (vcpu->cpu != cpu) { 709 if (vcpu->cpu != cpu) {
710 vcpu_clear(vmx); 710 vcpu_clear(vmx);
711 kvm_migrate_timers(vcpu); 711 kvm_migrate_timers(vcpu);
712 vpid_sync_vcpu_all(vmx); 712 set_bit(KVM_REQ_TLB_FLUSH, &vcpu->requests);
713 local_irq_disable(); 713 local_irq_disable();
714 list_add(&vmx->local_vcpus_link, 714 list_add(&vmx->local_vcpus_link,
715 &per_cpu(vcpus_on_cpu, cpu)); 715 &per_cpu(vcpus_on_cpu, cpu));
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index be451ee44249..9b9695322f56 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -1591,6 +1591,8 @@ static int kvm_dev_ioctl_get_supported_cpuid(struct kvm_cpuid2 *cpuid,
1591 1591
1592 if (cpuid->nent < 1) 1592 if (cpuid->nent < 1)
1593 goto out; 1593 goto out;
1594 if (cpuid->nent > KVM_MAX_CPUID_ENTRIES)
1595 cpuid->nent = KVM_MAX_CPUID_ENTRIES;
1594 r = -ENOMEM; 1596 r = -ENOMEM;
1595 cpuid_entries = vmalloc(sizeof(struct kvm_cpuid_entry2) * cpuid->nent); 1597 cpuid_entries = vmalloc(sizeof(struct kvm_cpuid_entry2) * cpuid->nent);
1596 if (!cpuid_entries) 1598 if (!cpuid_entries)
diff --git a/arch/x86/lib/Makefile b/arch/x86/lib/Makefile
index 965026472c71..a2d6472895fb 100644
--- a/arch/x86/lib/Makefile
+++ b/arch/x86/lib/Makefile
@@ -29,7 +29,9 @@ ifeq ($(CONFIG_X86_32),y)
29 lib-y += checksum_32.o 29 lib-y += checksum_32.o
30 lib-y += strstr_32.o 30 lib-y += strstr_32.o
31 lib-y += semaphore_32.o string_32.o 31 lib-y += semaphore_32.o string_32.o
32 32ifneq ($(CONFIG_X86_CMPXCHG64),y)
33 lib-y += cmpxchg8b_emu.o
34endif
33 lib-$(CONFIG_X86_USE_3DNOW) += mmx_32.o 35 lib-$(CONFIG_X86_USE_3DNOW) += mmx_32.o
34else 36else
35 obj-y += io_64.o iomap_copy_64.o 37 obj-y += io_64.o iomap_copy_64.o
diff --git a/arch/x86/lib/cmpxchg8b_emu.S b/arch/x86/lib/cmpxchg8b_emu.S
new file mode 100644
index 000000000000..828cb710dec2
--- /dev/null
+++ b/arch/x86/lib/cmpxchg8b_emu.S
@@ -0,0 +1,57 @@
1/*
2 * This program is free software; you can redistribute it and/or
3 * modify it under the terms of the GNU General Public License
4 * as published by the Free Software Foundation; version 2
5 * of the License.
6 *
7 */
8
9#include <linux/linkage.h>
10#include <asm/alternative-asm.h>
11#include <asm/frame.h>
12#include <asm/dwarf2.h>
13
14
15.text
16
17/*
18 * Inputs:
19 * %esi : memory location to compare
20 * %eax : low 32 bits of old value
21 * %edx : high 32 bits of old value
22 * %ebx : low 32 bits of new value
23 * %ecx : high 32 bits of new value
24 */
25ENTRY(cmpxchg8b_emu)
26CFI_STARTPROC
27
28#
29# Emulate 'cmpxchg8b (%esi)' on UP except we don't
30# set the whole ZF thing (caller will just compare
31# eax:edx with the expected value)
32#
33cmpxchg8b_emu:
34 pushfl
35 cli
36
37 cmpl (%esi), %eax
38 jne not_same
39 cmpl 4(%esi), %edx
40 jne half_same
41
42 movl %ebx, (%esi)
43 movl %ecx, 4(%esi)
44
45 popfl
46 ret
47
48 not_same:
49 movl (%esi), %eax
50 half_same:
51 movl 4(%esi), %edx
52
53 popfl
54 ret
55
56CFI_ENDPROC
57ENDPROC(cmpxchg8b_emu)
diff --git a/arch/x86/mm/Makefile b/arch/x86/mm/Makefile
index 9b5a9f59a478..06630d26e56d 100644
--- a/arch/x86/mm/Makefile
+++ b/arch/x86/mm/Makefile
@@ -1,9 +1,10 @@
1obj-y := init.o init_$(BITS).o fault.o ioremap.o extable.o pageattr.o mmap.o \ 1obj-y := init.o init_$(BITS).o fault.o ioremap.o extable.o pageattr.o mmap.o \
2 pat.o pgtable.o physaddr.o gup.o 2 pat.o pgtable.o physaddr.o gup.o setup_nx.o
3 3
4# Make sure __phys_addr has no stackprotector 4# Make sure __phys_addr has no stackprotector
5nostackp := $(call cc-option, -fno-stack-protector) 5nostackp := $(call cc-option, -fno-stack-protector)
6CFLAGS_physaddr.o := $(nostackp) 6CFLAGS_physaddr.o := $(nostackp)
7CFLAGS_setup_nx.o := $(nostackp)
7 8
8obj-$(CONFIG_SMP) += tlb.o 9obj-$(CONFIG_SMP) += tlb.o
9 10
diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c
index 923ea3fb7037..8f4e2ac93928 100644
--- a/arch/x86/mm/fault.c
+++ b/arch/x86/mm/fault.c
@@ -168,6 +168,7 @@ force_sig_info_fault(int si_signo, int si_code, unsigned long address,
168 info.si_errno = 0; 168 info.si_errno = 0;
169 info.si_code = si_code; 169 info.si_code = si_code;
170 info.si_addr = (void __user *)address; 170 info.si_addr = (void __user *)address;
171 info.si_addr_lsb = si_code == BUS_MCEERR_AR ? PAGE_SHIFT : 0;
171 172
172 force_sig_info(si_signo, &info, tsk); 173 force_sig_info(si_signo, &info, tsk);
173} 174}
@@ -791,10 +792,12 @@ out_of_memory(struct pt_regs *regs, unsigned long error_code,
791} 792}
792 793
793static void 794static void
794do_sigbus(struct pt_regs *regs, unsigned long error_code, unsigned long address) 795do_sigbus(struct pt_regs *regs, unsigned long error_code, unsigned long address,
796 unsigned int fault)
795{ 797{
796 struct task_struct *tsk = current; 798 struct task_struct *tsk = current;
797 struct mm_struct *mm = tsk->mm; 799 struct mm_struct *mm = tsk->mm;
800 int code = BUS_ADRERR;
798 801
799 up_read(&mm->mmap_sem); 802 up_read(&mm->mmap_sem);
800 803
@@ -810,7 +813,15 @@ do_sigbus(struct pt_regs *regs, unsigned long error_code, unsigned long address)
810 tsk->thread.error_code = error_code; 813 tsk->thread.error_code = error_code;
811 tsk->thread.trap_no = 14; 814 tsk->thread.trap_no = 14;
812 815
813 force_sig_info_fault(SIGBUS, BUS_ADRERR, address, tsk); 816#ifdef CONFIG_MEMORY_FAILURE
817 if (fault & VM_FAULT_HWPOISON) {
818 printk(KERN_ERR
819 "MCE: Killing %s:%d due to hardware memory corruption fault at %lx\n",
820 tsk->comm, tsk->pid, address);
821 code = BUS_MCEERR_AR;
822 }
823#endif
824 force_sig_info_fault(SIGBUS, code, address, tsk);
814} 825}
815 826
816static noinline void 827static noinline void
@@ -820,8 +831,8 @@ mm_fault_error(struct pt_regs *regs, unsigned long error_code,
820 if (fault & VM_FAULT_OOM) { 831 if (fault & VM_FAULT_OOM) {
821 out_of_memory(regs, error_code, address); 832 out_of_memory(regs, error_code, address);
822 } else { 833 } else {
823 if (fault & VM_FAULT_SIGBUS) 834 if (fault & (VM_FAULT_SIGBUS|VM_FAULT_HWPOISON))
824 do_sigbus(regs, error_code, address); 835 do_sigbus(regs, error_code, address, fault);
825 else 836 else
826 BUG(); 837 BUG();
827 } 838 }
diff --git a/arch/x86/mm/init.c b/arch/x86/mm/init.c
index 0607119cef94..73ffd5536f62 100644
--- a/arch/x86/mm/init.c
+++ b/arch/x86/mm/init.c
@@ -28,69 +28,6 @@ int direct_gbpages
28#endif 28#endif
29; 29;
30 30
31int nx_enabled;
32
33#if defined(CONFIG_X86_64) || defined(CONFIG_X86_PAE)
34static int disable_nx __cpuinitdata;
35
36/*
37 * noexec = on|off
38 *
39 * Control non-executable mappings for processes.
40 *
41 * on Enable
42 * off Disable
43 */
44static int __init noexec_setup(char *str)
45{
46 if (!str)
47 return -EINVAL;
48 if (!strncmp(str, "on", 2)) {
49 __supported_pte_mask |= _PAGE_NX;
50 disable_nx = 0;
51 } else if (!strncmp(str, "off", 3)) {
52 disable_nx = 1;
53 __supported_pte_mask &= ~_PAGE_NX;
54 }
55 return 0;
56}
57early_param("noexec", noexec_setup);
58#endif
59
60#ifdef CONFIG_X86_PAE
61static void __init set_nx(void)
62{
63 unsigned int v[4], l, h;
64
65 if (cpu_has_pae && (cpuid_eax(0x80000000) > 0x80000001)) {
66 cpuid(0x80000001, &v[0], &v[1], &v[2], &v[3]);
67
68 if ((v[3] & (1 << 20)) && !disable_nx) {
69 rdmsr(MSR_EFER, l, h);
70 l |= EFER_NX;
71 wrmsr(MSR_EFER, l, h);
72 nx_enabled = 1;
73 __supported_pte_mask |= _PAGE_NX;
74 }
75 }
76}
77#else
78static inline void set_nx(void)
79{
80}
81#endif
82
83#ifdef CONFIG_X86_64
84void __cpuinit check_efer(void)
85{
86 unsigned long efer;
87
88 rdmsrl(MSR_EFER, efer);
89 if (!(efer & EFER_NX) || disable_nx)
90 __supported_pte_mask &= ~_PAGE_NX;
91}
92#endif
93
94static void __init find_early_table_space(unsigned long end, int use_pse, 31static void __init find_early_table_space(unsigned long end, int use_pse,
95 int use_gbpages) 32 int use_gbpages)
96{ 33{
diff --git a/arch/x86/mm/pageattr.c b/arch/x86/mm/pageattr.c
index 24952fdc7e40..dd38bfbefd1f 100644
--- a/arch/x86/mm/pageattr.c
+++ b/arch/x86/mm/pageattr.c
@@ -144,6 +144,7 @@ void clflush_cache_range(void *vaddr, unsigned int size)
144 144
145 mb(); 145 mb();
146} 146}
147EXPORT_SYMBOL_GPL(clflush_cache_range);
147 148
148static void __cpa_flush_all(void *arg) 149static void __cpa_flush_all(void *arg)
149{ 150{
diff --git a/arch/x86/mm/pat.c b/arch/x86/mm/pat.c
index 7257cf3decf9..e78cd0ec2bcf 100644
--- a/arch/x86/mm/pat.c
+++ b/arch/x86/mm/pat.c
@@ -81,6 +81,7 @@ enum {
81void pat_init(void) 81void pat_init(void)
82{ 82{
83 u64 pat; 83 u64 pat;
84 bool boot_cpu = !boot_pat_state;
84 85
85 if (!pat_enabled) 86 if (!pat_enabled)
86 return; 87 return;
@@ -122,8 +123,10 @@ void pat_init(void)
122 rdmsrl(MSR_IA32_CR_PAT, boot_pat_state); 123 rdmsrl(MSR_IA32_CR_PAT, boot_pat_state);
123 124
124 wrmsrl(MSR_IA32_CR_PAT, pat); 125 wrmsrl(MSR_IA32_CR_PAT, pat);
125 printk(KERN_INFO "x86 PAT enabled: cpu %d, old 0x%Lx, new 0x%Lx\n", 126
126 smp_processor_id(), boot_pat_state, pat); 127 if (boot_cpu)
128 printk(KERN_INFO "x86 PAT enabled: cpu %d, old 0x%Lx, new 0x%Lx\n",
129 smp_processor_id(), boot_pat_state, pat);
127} 130}
128 131
129#undef PAT 132#undef PAT
diff --git a/arch/x86/mm/setup_nx.c b/arch/x86/mm/setup_nx.c
new file mode 100644
index 000000000000..513d8ed5d2ec
--- /dev/null
+++ b/arch/x86/mm/setup_nx.c
@@ -0,0 +1,69 @@
1#include <linux/spinlock.h>
2#include <linux/errno.h>
3#include <linux/init.h>
4
5#include <asm/pgtable.h>
6
7int nx_enabled;
8
9#if defined(CONFIG_X86_64) || defined(CONFIG_X86_PAE)
10static int disable_nx __cpuinitdata;
11
12/*
13 * noexec = on|off
14 *
15 * Control non-executable mappings for processes.
16 *
17 * on Enable
18 * off Disable
19 */
20static int __init noexec_setup(char *str)
21{
22 if (!str)
23 return -EINVAL;
24 if (!strncmp(str, "on", 2)) {
25 __supported_pte_mask |= _PAGE_NX;
26 disable_nx = 0;
27 } else if (!strncmp(str, "off", 3)) {
28 disable_nx = 1;
29 __supported_pte_mask &= ~_PAGE_NX;
30 }
31 return 0;
32}
33early_param("noexec", noexec_setup);
34#endif
35
36#ifdef CONFIG_X86_PAE
37void __init set_nx(void)
38{
39 unsigned int v[4], l, h;
40
41 if (cpu_has_pae && (cpuid_eax(0x80000000) > 0x80000001)) {
42 cpuid(0x80000001, &v[0], &v[1], &v[2], &v[3]);
43
44 if ((v[3] & (1 << 20)) && !disable_nx) {
45 rdmsr(MSR_EFER, l, h);
46 l |= EFER_NX;
47 wrmsr(MSR_EFER, l, h);
48 nx_enabled = 1;
49 __supported_pte_mask |= _PAGE_NX;
50 }
51 }
52}
53#else
54void set_nx(void)
55{
56}
57#endif
58
59#ifdef CONFIG_X86_64
60void __cpuinit check_efer(void)
61{
62 unsigned long efer;
63
64 rdmsrl(MSR_EFER, efer);
65 if (!(efer & EFER_NX) || disable_nx)
66 __supported_pte_mask &= ~_PAGE_NX;
67}
68#endif
69
diff --git a/arch/x86/mm/tlb.c b/arch/x86/mm/tlb.c
index c814e144a3f0..36fe08eeb5c3 100644
--- a/arch/x86/mm/tlb.c
+++ b/arch/x86/mm/tlb.c
@@ -59,7 +59,8 @@ void leave_mm(int cpu)
59{ 59{
60 if (percpu_read(cpu_tlbstate.state) == TLBSTATE_OK) 60 if (percpu_read(cpu_tlbstate.state) == TLBSTATE_OK)
61 BUG(); 61 BUG();
62 cpu_clear(cpu, percpu_read(cpu_tlbstate.active_mm)->cpu_vm_mask); 62 cpumask_clear_cpu(cpu,
63 mm_cpumask(percpu_read(cpu_tlbstate.active_mm)));
63 load_cr3(swapper_pg_dir); 64 load_cr3(swapper_pg_dir);
64} 65}
65EXPORT_SYMBOL_GPL(leave_mm); 66EXPORT_SYMBOL_GPL(leave_mm);
@@ -234,8 +235,8 @@ void flush_tlb_current_task(void)
234 preempt_disable(); 235 preempt_disable();
235 236
236 local_flush_tlb(); 237 local_flush_tlb();
237 if (cpumask_any_but(&mm->cpu_vm_mask, smp_processor_id()) < nr_cpu_ids) 238 if (cpumask_any_but(mm_cpumask(mm), smp_processor_id()) < nr_cpu_ids)
238 flush_tlb_others(&mm->cpu_vm_mask, mm, TLB_FLUSH_ALL); 239 flush_tlb_others(mm_cpumask(mm), mm, TLB_FLUSH_ALL);
239 preempt_enable(); 240 preempt_enable();
240} 241}
241 242
@@ -249,8 +250,8 @@ void flush_tlb_mm(struct mm_struct *mm)
249 else 250 else
250 leave_mm(smp_processor_id()); 251 leave_mm(smp_processor_id());
251 } 252 }
252 if (cpumask_any_but(&mm->cpu_vm_mask, smp_processor_id()) < nr_cpu_ids) 253 if (cpumask_any_but(mm_cpumask(mm), smp_processor_id()) < nr_cpu_ids)
253 flush_tlb_others(&mm->cpu_vm_mask, mm, TLB_FLUSH_ALL); 254 flush_tlb_others(mm_cpumask(mm), mm, TLB_FLUSH_ALL);
254 255
255 preempt_enable(); 256 preempt_enable();
256} 257}
@@ -268,8 +269,8 @@ void flush_tlb_page(struct vm_area_struct *vma, unsigned long va)
268 leave_mm(smp_processor_id()); 269 leave_mm(smp_processor_id());
269 } 270 }
270 271
271 if (cpumask_any_but(&mm->cpu_vm_mask, smp_processor_id()) < nr_cpu_ids) 272 if (cpumask_any_but(mm_cpumask(mm), smp_processor_id()) < nr_cpu_ids)
272 flush_tlb_others(&mm->cpu_vm_mask, mm, va); 273 flush_tlb_others(mm_cpumask(mm), mm, va);
273 274
274 preempt_enable(); 275 preempt_enable();
275} 276}
diff --git a/arch/x86/pci/common.c b/arch/x86/pci/common.c
index 5db96d4304de..1331fcf26143 100644
--- a/arch/x86/pci/common.c
+++ b/arch/x86/pci/common.c
@@ -646,7 +646,7 @@ int get_mp_bus_to_node(int busnum)
646 646
647#else /* CONFIG_X86_32 */ 647#else /* CONFIG_X86_32 */
648 648
649static unsigned char mp_bus_to_node[BUS_NR] = { 649static int mp_bus_to_node[BUS_NR] = {
650 [0 ... BUS_NR - 1] = -1 650 [0 ... BUS_NR - 1] = -1
651}; 651};
652 652
diff --git a/arch/x86/pci/i386.c b/arch/x86/pci/i386.c
index 52e62e57fedd..b22d13b0c71d 100644
--- a/arch/x86/pci/i386.c
+++ b/arch/x86/pci/i386.c
@@ -266,7 +266,7 @@ void pcibios_set_master(struct pci_dev *dev)
266 pci_write_config_byte(dev, PCI_LATENCY_TIMER, lat); 266 pci_write_config_byte(dev, PCI_LATENCY_TIMER, lat);
267} 267}
268 268
269static struct vm_operations_struct pci_mmap_ops = { 269static const struct vm_operations_struct pci_mmap_ops = {
270 .access = generic_access_phys, 270 .access = generic_access_phys,
271}; 271};
272 272
diff --git a/arch/x86/vdso/Makefile b/arch/x86/vdso/Makefile
index 88112b49f02c..6b4ffedb93c9 100644
--- a/arch/x86/vdso/Makefile
+++ b/arch/x86/vdso/Makefile
@@ -122,7 +122,7 @@ quiet_cmd_vdso = VDSO $@
122 $(VDSO_LDFLAGS) $(VDSO_LDFLAGS_$(filter %.lds,$(^F))) \ 122 $(VDSO_LDFLAGS) $(VDSO_LDFLAGS_$(filter %.lds,$(^F))) \
123 -Wl,-T,$(filter %.lds,$^) $(filter %.o,$^) 123 -Wl,-T,$(filter %.lds,$^) $(filter %.o,$^)
124 124
125VDSO_LDFLAGS = -fPIC -shared $(call ld-option, -Wl$(comma)--hash-style=sysv) 125VDSO_LDFLAGS = -fPIC -shared $(call cc-ldoption, -Wl$(comma)--hash-style=sysv)
126GCOV_PROFILE := n 126GCOV_PROFILE := n
127 127
128# 128#
diff --git a/arch/x86/xen/debugfs.c b/arch/x86/xen/debugfs.c
index b53225d2cac3..e133ce25e290 100644
--- a/arch/x86/xen/debugfs.c
+++ b/arch/x86/xen/debugfs.c
@@ -100,7 +100,7 @@ static int xen_array_release(struct inode *inode, struct file *file)
100 return 0; 100 return 0;
101} 101}
102 102
103static struct file_operations u32_array_fops = { 103static const struct file_operations u32_array_fops = {
104 .owner = THIS_MODULE, 104 .owner = THIS_MODULE,
105 .open = u32_array_open, 105 .open = u32_array_open,
106 .release= xen_array_release, 106 .release= xen_array_release,
diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c
index 544eb7496531..3439616d69f1 100644
--- a/arch/x86/xen/enlighten.c
+++ b/arch/x86/xen/enlighten.c
@@ -1082,6 +1082,11 @@ asmlinkage void __init xen_start_kernel(void)
1082 1082
1083 __supported_pte_mask |= _PAGE_IOMAP; 1083 __supported_pte_mask |= _PAGE_IOMAP;
1084 1084
1085#ifdef CONFIG_X86_64
1086 /* Work out if we support NX */
1087 check_efer();
1088#endif
1089
1085 xen_setup_features(); 1090 xen_setup_features();
1086 1091
1087 /* Get mfn list */ 1092 /* Get mfn list */
@@ -1123,11 +1128,6 @@ asmlinkage void __init xen_start_kernel(void)
1123 1128
1124 pgd = (pgd_t *)xen_start_info->pt_base; 1129 pgd = (pgd_t *)xen_start_info->pt_base;
1125 1130
1126#ifdef CONFIG_X86_64
1127 /* Work out if we support NX */
1128 check_efer();
1129#endif
1130
1131 /* Don't do the full vcpu_info placement stuff until we have a 1131 /* Don't do the full vcpu_info placement stuff until we have a
1132 possible map and a non-dummy shared_info. */ 1132 possible map and a non-dummy shared_info. */
1133 per_cpu(xen_vcpu, 0) = &HYPERVISOR_shared_info->vcpu_info[0]; 1133 per_cpu(xen_vcpu, 0) = &HYPERVISOR_shared_info->vcpu_info[0];
diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c
index 093dd59b5385..3bf7b1d250ce 100644
--- a/arch/x86/xen/mmu.c
+++ b/arch/x86/xen/mmu.c
@@ -1165,14 +1165,14 @@ static void xen_drop_mm_ref(struct mm_struct *mm)
1165 /* Get the "official" set of cpus referring to our pagetable. */ 1165 /* Get the "official" set of cpus referring to our pagetable. */
1166 if (!alloc_cpumask_var(&mask, GFP_ATOMIC)) { 1166 if (!alloc_cpumask_var(&mask, GFP_ATOMIC)) {
1167 for_each_online_cpu(cpu) { 1167 for_each_online_cpu(cpu) {
1168 if (!cpumask_test_cpu(cpu, &mm->cpu_vm_mask) 1168 if (!cpumask_test_cpu(cpu, mm_cpumask(mm))
1169 && per_cpu(xen_current_cr3, cpu) != __pa(mm->pgd)) 1169 && per_cpu(xen_current_cr3, cpu) != __pa(mm->pgd))
1170 continue; 1170 continue;
1171 smp_call_function_single(cpu, drop_other_mm_ref, mm, 1); 1171 smp_call_function_single(cpu, drop_other_mm_ref, mm, 1);
1172 } 1172 }
1173 return; 1173 return;
1174 } 1174 }
1175 cpumask_copy(mask, &mm->cpu_vm_mask); 1175 cpumask_copy(mask, mm_cpumask(mm));
1176 1176
1177 /* It's possible that a vcpu may have a stale reference to our 1177 /* It's possible that a vcpu may have a stale reference to our
1178 cr3, because its in lazy mode, and it hasn't yet flushed 1178 cr3, because its in lazy mode, and it hasn't yet flushed
generated by cgit v1.2.2 (git 2.25.0) at 2025-01-25 07:23:20 -0500