aboutsummaryrefslogtreecommitdiffstats
path: root/arch/x86
diff options
context:
space:
mode:
authorIngo Molnar <mingo@elte.hu>2009-02-22 11:59:49 -0500
committerIngo Molnar <mingo@elte.hu>2009-02-22 11:59:49 -0500
commit2702e0a46c2d28da92b32c9b068ee1291fc0de35 (patch)
treec1e9868a4b91beeebd5f4e820588bc2ae8371911 /arch/x86
parent42bb8cc5e81028e217105299001070d57eb84ad7 (diff)
parentadfafefd104d840ee4461965f22624d77532675b (diff)
Merge branch 'linus' into timers/hpet
Diffstat (limited to 'arch/x86')
-rw-r--r--arch/x86/Kconfig11
-rw-r--r--arch/x86/Kconfig.cpu28
-rw-r--r--arch/x86/Kconfig.debug24
-rw-r--r--arch/x86/boot/video-vesa.c11
-rw-r--r--arch/x86/configs/i386_defconfig4
-rw-r--r--arch/x86/configs/x86_64_defconfig4
-rw-r--r--arch/x86/ia32/ia32entry.S8
-rw-r--r--arch/x86/include/asm/Kbuild1
-rw-r--r--arch/x86/include/asm/a.out-core.h4
-rw-r--r--arch/x86/include/asm/bitops.h14
-rw-r--r--arch/x86/include/asm/byteorder.h1
-rw-r--r--arch/x86/include/asm/cpufeature.h1
-rw-r--r--arch/x86/include/asm/dma-mapping.h4
-rw-r--r--arch/x86/include/asm/e820.h1
-rw-r--r--arch/x86/include/asm/io.h1
-rw-r--r--arch/x86/include/asm/kvm.h9
-rw-r--r--arch/x86/include/asm/mach-default/mach_wakecpu.h6
-rw-r--r--arch/x86/include/asm/math_emu.h29
-rw-r--r--arch/x86/include/asm/mce.h5
-rw-r--r--arch/x86/include/asm/mmzone_32.h2
-rw-r--r--arch/x86/include/asm/mmzone_64.h2
-rw-r--r--arch/x86/include/asm/mpspec.h6
-rw-r--r--arch/x86/include/asm/msr-index.h29
-rw-r--r--arch/x86/include/asm/mtrr.h11
-rw-r--r--arch/x86/include/asm/page.h1
-rw-r--r--arch/x86/include/asm/paravirt.h18
-rw-r--r--arch/x86/include/asm/pgalloc.h1
-rw-r--r--arch/x86/include/asm/pgtable.h45
-rw-r--r--arch/x86/include/asm/processor.h2
-rw-r--r--arch/x86/include/asm/ptrace-abi.h2
-rw-r--r--arch/x86/include/asm/sigcontext.h2
-rw-r--r--arch/x86/include/asm/sigcontext32.h2
-rw-r--r--arch/x86/include/asm/spinlock.h1
-rw-r--r--arch/x86/include/asm/swab.h2
-rw-r--r--arch/x86/include/asm/syscalls.h2
-rw-r--r--arch/x86/include/asm/timex.h13
-rw-r--r--arch/x86/include/asm/traps.h4
-rw-r--r--arch/x86/include/asm/xen/page.h2
-rw-r--r--arch/x86/kernel/acpi/boot.c23
-rw-r--r--arch/x86/kernel/acpi/sleep.c4
-rw-r--r--arch/x86/kernel/apic.c14
-rw-r--r--arch/x86/kernel/cpu/cpufreq/Kconfig11
-rw-r--r--arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c20
-rw-r--r--arch/x86/kernel/cpu/cpufreq/powernow-k8.c40
-rw-r--r--arch/x86/kernel/cpu/intel.c16
-rw-r--r--arch/x86/kernel/cpu/intel_cacheinfo.c15
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce_64.c7
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce_amd_64.c2
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce_intel_64.c4
-rw-r--r--arch/x86/kernel/cpu/mtrr/generic.c12
-rw-r--r--arch/x86/kernel/cpu/mtrr/main.c3
-rw-r--r--arch/x86/kernel/ds.c31
-rw-r--r--arch/x86/kernel/entry_32.S2
-rw-r--r--arch/x86/kernel/entry_64.S1
-rw-r--r--arch/x86/kernel/ftrace.c17
-rw-r--r--arch/x86/kernel/hpet.c17
-rw-r--r--arch/x86/kernel/i8237.c17
-rw-r--r--arch/x86/kernel/io_apic.c25
-rw-r--r--arch/x86/kernel/irqinit_32.c12
-rw-r--r--arch/x86/kernel/kprobes.c2
-rw-r--r--arch/x86/kernel/mpparse.c1
-rw-r--r--arch/x86/kernel/olpc.c2
-rw-r--r--arch/x86/kernel/paravirt.c26
-rw-r--r--arch/x86/kernel/pci-gart_64.c2
-rw-r--r--arch/x86/kernel/process.c6
-rw-r--r--arch/x86/kernel/process_32.c3
-rw-r--r--arch/x86/kernel/process_64.c9
-rw-r--r--arch/x86/kernel/ptrace.c16
-rw-r--r--arch/x86/kernel/setup.c2
-rw-r--r--arch/x86/kernel/setup_percpu.c2
-rw-r--r--arch/x86/kernel/signal.c11
-rw-r--r--arch/x86/kernel/syscall_table_32.S2
-rw-r--r--arch/x86/kernel/tlb_uv.c1
-rw-r--r--arch/x86/kernel/traps.c25
-rw-r--r--arch/x86/kernel/vmi_32.c13
-rw-r--r--arch/x86/kernel/vmiclock_32.c5
-rw-r--r--arch/x86/kvm/i8254.c2
-rw-r--r--arch/x86/kvm/irq.c7
-rw-r--r--arch/x86/kvm/irq.h1
-rw-r--r--arch/x86/kvm/lapic.c66
-rw-r--r--arch/x86/kvm/lapic.h2
-rw-r--r--arch/x86/kvm/mmu.c9
-rw-r--r--arch/x86/kvm/svm.c1
-rw-r--r--arch/x86/kvm/vmx.c5
-rw-r--r--arch/x86/kvm/x86.c10
-rw-r--r--arch/x86/lguest/boot.c4
-rw-r--r--arch/x86/lib/usercopy_32.c4
-rw-r--r--arch/x86/lib/usercopy_64.c4
-rw-r--r--arch/x86/mach-default/setup.c12
-rw-r--r--arch/x86/mach-voyager/setup.c12
-rw-r--r--arch/x86/mach-voyager/voyager_smp.c25
-rw-r--r--arch/x86/math-emu/fpu_entry.c6
-rw-r--r--arch/x86/math-emu/fpu_proto.h4
-rw-r--r--arch/x86/math-emu/fpu_system.h16
-rw-r--r--arch/x86/math-emu/get_address.c69
-rw-r--r--arch/x86/mm/fault.c10
-rw-r--r--arch/x86/mm/init_32.c48
-rw-r--r--arch/x86/mm/init_64.c2
-rw-r--r--arch/x86/mm/iomap_32.c10
-rw-r--r--arch/x86/mm/ioremap.c44
-rw-r--r--arch/x86/mm/numa_64.c2
-rw-r--r--arch/x86/mm/pageattr.c79
-rw-r--r--arch/x86/mm/pat.c103
-rw-r--r--arch/x86/pci/i386.c12
-rw-r--r--arch/x86/pci/irq.c1
-rw-r--r--arch/x86/scripts/strip-symbols1
-rw-r--r--arch/x86/xen/multicalls.h4
107 files changed, 773 insertions, 536 deletions
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 73f7fe8fd4d1..9c39095b33fc 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -1802,6 +1802,17 @@ config DMAR
1802 and include PCI device scope covered by these DMA 1802 and include PCI device scope covered by these DMA
1803 remapping devices. 1803 remapping devices.
1804 1804
1805config DMAR_DEFAULT_ON
1806 def_bool n
1807 prompt "Enable DMA Remapping Devices by default"
1808 depends on DMAR
1809 help
1810 Selecting this option will enable a DMAR device at boot time if
1811 one is found. If this option is not selected, DMAR support can
1812 be enabled by passing intel_iommu=on to the kernel. It is
1813 recommended you say N here while the DMAR code remains
1814 experimental.
1815
1805config DMAR_GFX_WA 1816config DMAR_GFX_WA
1806 def_bool y 1817 def_bool y
1807 prompt "Support for Graphics workaround" 1818 prompt "Support for Graphics workaround"
diff --git a/arch/x86/Kconfig.cpu b/arch/x86/Kconfig.cpu
index 8078955845ae..c98d52e82966 100644
--- a/arch/x86/Kconfig.cpu
+++ b/arch/x86/Kconfig.cpu
@@ -167,9 +167,9 @@ config MK7
167config MK8 167config MK8
168 bool "Opteron/Athlon64/Hammer/K8" 168 bool "Opteron/Athlon64/Hammer/K8"
169 help 169 help
170 Select this for an AMD Opteron or Athlon64 Hammer-family processor. Enables 170 Select this for an AMD Opteron or Athlon64 Hammer-family processor.
171 use of some extended instructions, and passes appropriate optimization 171 Enables use of some extended instructions, and passes appropriate
172 flags to GCC. 172 optimization flags to GCC.
173 173
174config MCRUSOE 174config MCRUSOE
175 bool "Crusoe" 175 bool "Crusoe"
@@ -256,9 +256,11 @@ config MPSC
256config MCORE2 256config MCORE2
257 bool "Core 2/newer Xeon" 257 bool "Core 2/newer Xeon"
258 help 258 help
259 Select this for Intel Core 2 and newer Core 2 Xeons (Xeon 51xx and 53xx) 259
260 CPUs. You can distinguish newer from older Xeons by the CPU family 260 Select this for Intel Core 2 and newer Core 2 Xeons (Xeon 51xx and
261 in /proc/cpuinfo. Newer ones have 6 and older ones 15 (not a typo) 261 53xx) CPUs. You can distinguish newer from older Xeons by the CPU
262 family in /proc/cpuinfo. Newer ones have 6 and older ones 15
263 (not a typo)
262 264
263config GENERIC_CPU 265config GENERIC_CPU
264 bool "Generic-x86-64" 266 bool "Generic-x86-64"
@@ -320,14 +322,14 @@ config X86_PPRO_FENCE
320 bool "PentiumPro memory ordering errata workaround" 322 bool "PentiumPro memory ordering errata workaround"
321 depends on M686 || M586MMX || M586TSC || M586 || M486 || M386 || MGEODEGX1 323 depends on M686 || M586MMX || M586TSC || M586 || M486 || M386 || MGEODEGX1
322 help 324 help
323 Old PentiumPro multiprocessor systems had errata that could cause memory 325 Old PentiumPro multiprocessor systems had errata that could cause
324 operations to violate the x86 ordering standard in rare cases. Enabling this 326 memory operations to violate the x86 ordering standard in rare cases.
325 option will attempt to work around some (but not all) occurances of 327 Enabling this option will attempt to work around some (but not all)
326 this problem, at the cost of much heavier spinlock and memory barrier 328 occurances of this problem, at the cost of much heavier spinlock and
327 operations. 329 memory barrier operations.
328 330
329 If unsure, say n here. Even distro kernels should think twice before enabling 331 If unsure, say n here. Even distro kernels should think twice before
330 this: there are few systems, and an unlikely bug. 332 enabling this: there are few systems, and an unlikely bug.
331 333
332config X86_F00F_BUG 334config X86_F00F_BUG
333 def_bool y 335 def_bool y
diff --git a/arch/x86/Kconfig.debug b/arch/x86/Kconfig.debug
index 10d6cc3fd052..e1983fa025d2 100644
--- a/arch/x86/Kconfig.debug
+++ b/arch/x86/Kconfig.debug
@@ -174,28 +174,8 @@ config IOMMU_LEAK
174 Add a simple leak tracer to the IOMMU code. This is useful when you 174 Add a simple leak tracer to the IOMMU code. This is useful when you
175 are debugging a buggy device driver that leaks IOMMU mappings. 175 are debugging a buggy device driver that leaks IOMMU mappings.
176 176
177config MMIOTRACE 177config HAVE_MMIOTRACE_SUPPORT
178 bool "Memory mapped IO tracing" 178 def_bool y
179 depends on DEBUG_KERNEL && PCI
180 select TRACING
181 help
182 Mmiotrace traces Memory Mapped I/O access and is meant for
183 debugging and reverse engineering. It is called from the ioremap
184 implementation and works via page faults. Tracing is disabled by
185 default and can be enabled at run-time.
186
187 See Documentation/tracers/mmiotrace.txt.
188 If you are not helping to develop drivers, say N.
189
190config MMIOTRACE_TEST
191 tristate "Test module for mmiotrace"
192 depends on MMIOTRACE && m
193 help
194 This is a dumb module for testing mmiotrace. It is very dangerous
195 as it will write garbage to IO memory starting at a given address.
196 However, it should be safe to use on e.g. unused portion of VRAM.
197
198 Say N, unless you absolutely know what you are doing.
199 179
200# 180#
201# IO delay types: 181# IO delay types:
diff --git a/arch/x86/boot/video-vesa.c b/arch/x86/boot/video-vesa.c
index 75115849af33..4a58c8ce3f69 100644
--- a/arch/x86/boot/video-vesa.c
+++ b/arch/x86/boot/video-vesa.c
@@ -269,9 +269,8 @@ void vesa_store_edid(void)
269 we genuinely have to assume all registers are destroyed here. */ 269 we genuinely have to assume all registers are destroyed here. */
270 270
271 asm("pushw %%es; movw %2,%%es; "INT10"; popw %%es" 271 asm("pushw %%es; movw %2,%%es; "INT10"; popw %%es"
272 : "+a" (ax), "+b" (bx) 272 : "+a" (ax), "+b" (bx), "+c" (cx), "+D" (di)
273 : "c" (cx), "D" (di) 273 : : "esi", "edx");
274 : "esi");
275 274
276 if (ax != 0x004f) 275 if (ax != 0x004f)
277 return; /* No EDID */ 276 return; /* No EDID */
@@ -285,9 +284,9 @@ void vesa_store_edid(void)
285 dx = 0; /* EDID block number */ 284 dx = 0; /* EDID block number */
286 di =(size_t) &boot_params.edid_info; /* (ES:)Pointer to block */ 285 di =(size_t) &boot_params.edid_info; /* (ES:)Pointer to block */
287 asm(INT10 286 asm(INT10
288 : "+a" (ax), "+b" (bx), "+d" (dx), "=m" (boot_params.edid_info) 287 : "+a" (ax), "+b" (bx), "+d" (dx), "=m" (boot_params.edid_info),
289 : "c" (cx), "D" (di) 288 "+c" (cx), "+D" (di)
290 : "esi"); 289 : : "esi");
291#endif /* CONFIG_FIRMWARE_EDID */ 290#endif /* CONFIG_FIRMWARE_EDID */
292} 291}
293 292
diff --git a/arch/x86/configs/i386_defconfig b/arch/x86/configs/i386_defconfig
index b30a08ed8eb4..edba00d98ac3 100644
--- a/arch/x86/configs/i386_defconfig
+++ b/arch/x86/configs/i386_defconfig
@@ -1331,8 +1331,8 @@ CONFIG_I2C_I801=y
1331# Miscellaneous I2C Chip support 1331# Miscellaneous I2C Chip support
1332# 1332#
1333# CONFIG_DS1682 is not set 1333# CONFIG_DS1682 is not set
1334# CONFIG_AT24 is not set 1334# CONFIG_EEPROM_AT24 is not set
1335# CONFIG_SENSORS_EEPROM is not set 1335# CONFIG_EEPROM_LEGACY is not set
1336# CONFIG_SENSORS_PCF8574 is not set 1336# CONFIG_SENSORS_PCF8574 is not set
1337# CONFIG_PCF8575 is not set 1337# CONFIG_PCF8575 is not set
1338# CONFIG_SENSORS_PCA9539 is not set 1338# CONFIG_SENSORS_PCA9539 is not set
diff --git a/arch/x86/configs/x86_64_defconfig b/arch/x86/configs/x86_64_defconfig
index 0e7dbc0a3e46..322dd2748fc9 100644
--- a/arch/x86/configs/x86_64_defconfig
+++ b/arch/x86/configs/x86_64_defconfig
@@ -1311,8 +1311,8 @@ CONFIG_I2C_I801=y
1311# Miscellaneous I2C Chip support 1311# Miscellaneous I2C Chip support
1312# 1312#
1313# CONFIG_DS1682 is not set 1313# CONFIG_DS1682 is not set
1314# CONFIG_AT24 is not set 1314# CONFIG_EEPROM_AT24 is not set
1315# CONFIG_SENSORS_EEPROM is not set 1315# CONFIG_EEPROM_LEGACY is not set
1316# CONFIG_SENSORS_PCF8574 is not set 1316# CONFIG_SENSORS_PCF8574 is not set
1317# CONFIG_PCF8575 is not set 1317# CONFIG_PCF8575 is not set
1318# CONFIG_SENSORS_PCA9539 is not set 1318# CONFIG_SENSORS_PCA9539 is not set
diff --git a/arch/x86/ia32/ia32entry.S b/arch/x86/ia32/ia32entry.S
index 256b00b61892..5a0d76dc56a4 100644
--- a/arch/x86/ia32/ia32entry.S
+++ b/arch/x86/ia32/ia32entry.S
@@ -418,9 +418,9 @@ ENTRY(ia32_syscall)
418 orl $TS_COMPAT,TI_status(%r10) 418 orl $TS_COMPAT,TI_status(%r10)
419 testl $_TIF_WORK_SYSCALL_ENTRY,TI_flags(%r10) 419 testl $_TIF_WORK_SYSCALL_ENTRY,TI_flags(%r10)
420 jnz ia32_tracesys 420 jnz ia32_tracesys
421ia32_do_syscall:
422 cmpl $(IA32_NR_syscalls-1),%eax 421 cmpl $(IA32_NR_syscalls-1),%eax
423 ja int_ret_from_sys_call /* ia32_tracesys has set RAX(%rsp) */ 422 ja ia32_badsys
423ia32_do_call:
424 IA32_ARG_FIXUP 424 IA32_ARG_FIXUP
425 call *ia32_sys_call_table(,%rax,8) # xxx: rip relative 425 call *ia32_sys_call_table(,%rax,8) # xxx: rip relative
426ia32_sysret: 426ia32_sysret:
@@ -435,7 +435,9 @@ ia32_tracesys:
435 call syscall_trace_enter 435 call syscall_trace_enter
436 LOAD_ARGS32 ARGOFFSET /* reload args from stack in case ptrace changed it */ 436 LOAD_ARGS32 ARGOFFSET /* reload args from stack in case ptrace changed it */
437 RESTORE_REST 437 RESTORE_REST
438 jmp ia32_do_syscall 438 cmpl $(IA32_NR_syscalls-1),%eax
439 ja int_ret_from_sys_call /* ia32_tracesys has set RAX(%rsp) */
440 jmp ia32_do_call
439END(ia32_syscall) 441END(ia32_syscall)
440 442
441ia32_badsys: 443ia32_badsys:
diff --git a/arch/x86/include/asm/Kbuild b/arch/x86/include/asm/Kbuild
index a9f8a814a1f7..4a8e80cdcfa5 100644
--- a/arch/x86/include/asm/Kbuild
+++ b/arch/x86/include/asm/Kbuild
@@ -22,4 +22,3 @@ unifdef-y += unistd_32.h
22unifdef-y += unistd_64.h 22unifdef-y += unistd_64.h
23unifdef-y += vm86.h 23unifdef-y += vm86.h
24unifdef-y += vsyscall.h 24unifdef-y += vsyscall.h
25unifdef-y += swab.h
diff --git a/arch/x86/include/asm/a.out-core.h b/arch/x86/include/asm/a.out-core.h
index 37822206083e..3c601f8224be 100644
--- a/arch/x86/include/asm/a.out-core.h
+++ b/arch/x86/include/asm/a.out-core.h
@@ -23,8 +23,6 @@
23 */ 23 */
24static inline void aout_dump_thread(struct pt_regs *regs, struct user *dump) 24static inline void aout_dump_thread(struct pt_regs *regs, struct user *dump)
25{ 25{
26 u16 gs;
27
28/* changed the size calculations - should hopefully work better. lbt */ 26/* changed the size calculations - should hopefully work better. lbt */
29 dump->magic = CMAGIC; 27 dump->magic = CMAGIC;
30 dump->start_code = 0; 28 dump->start_code = 0;
@@ -57,7 +55,7 @@ static inline void aout_dump_thread(struct pt_regs *regs, struct user *dump)
57 dump->regs.ds = (u16)regs->ds; 55 dump->regs.ds = (u16)regs->ds;
58 dump->regs.es = (u16)regs->es; 56 dump->regs.es = (u16)regs->es;
59 dump->regs.fs = (u16)regs->fs; 57 dump->regs.fs = (u16)regs->fs;
60 savesegment(gs, gs); 58 savesegment(gs, dump->regs.gs);
61 dump->regs.orig_ax = regs->orig_ax; 59 dump->regs.orig_ax = regs->orig_ax;
62 dump->regs.ip = regs->ip; 60 dump->regs.ip = regs->ip;
63 dump->regs.cs = (u16)regs->cs; 61 dump->regs.cs = (u16)regs->cs;
diff --git a/arch/x86/include/asm/bitops.h b/arch/x86/include/asm/bitops.h
index e02a359d2aa5..02b47a603fc8 100644
--- a/arch/x86/include/asm/bitops.h
+++ b/arch/x86/include/asm/bitops.h
@@ -3,6 +3,9 @@
3 3
4/* 4/*
5 * Copyright 1992, Linus Torvalds. 5 * Copyright 1992, Linus Torvalds.
6 *
7 * Note: inlines with more than a single statement should be marked
8 * __always_inline to avoid problems with older gcc's inlining heuristics.
6 */ 9 */
7 10
8#ifndef _LINUX_BITOPS_H 11#ifndef _LINUX_BITOPS_H
@@ -53,7 +56,8 @@
53 * Note that @nr may be almost arbitrarily large; this function is not 56 * Note that @nr may be almost arbitrarily large; this function is not
54 * restricted to acting on a single-word quantity. 57 * restricted to acting on a single-word quantity.
55 */ 58 */
56static inline void set_bit(unsigned int nr, volatile unsigned long *addr) 59static __always_inline void
60set_bit(unsigned int nr, volatile unsigned long *addr)
57{ 61{
58 if (IS_IMMEDIATE(nr)) { 62 if (IS_IMMEDIATE(nr)) {
59 asm volatile(LOCK_PREFIX "orb %1,%0" 63 asm volatile(LOCK_PREFIX "orb %1,%0"
@@ -90,7 +94,8 @@ static inline void __set_bit(int nr, volatile unsigned long *addr)
90 * you should call smp_mb__before_clear_bit() and/or smp_mb__after_clear_bit() 94 * you should call smp_mb__before_clear_bit() and/or smp_mb__after_clear_bit()
91 * in order to ensure changes are visible on other processors. 95 * in order to ensure changes are visible on other processors.
92 */ 96 */
93static inline void clear_bit(int nr, volatile unsigned long *addr) 97static __always_inline void
98clear_bit(int nr, volatile unsigned long *addr)
94{ 99{
95 if (IS_IMMEDIATE(nr)) { 100 if (IS_IMMEDIATE(nr)) {
96 asm volatile(LOCK_PREFIX "andb %1,%0" 101 asm volatile(LOCK_PREFIX "andb %1,%0"
@@ -204,7 +209,8 @@ static inline int test_and_set_bit(int nr, volatile unsigned long *addr)
204 * 209 *
205 * This is the same as test_and_set_bit on x86. 210 * This is the same as test_and_set_bit on x86.
206 */ 211 */
207static inline int test_and_set_bit_lock(int nr, volatile unsigned long *addr) 212static __always_inline int
213test_and_set_bit_lock(int nr, volatile unsigned long *addr)
208{ 214{
209 return test_and_set_bit(nr, addr); 215 return test_and_set_bit(nr, addr);
210} 216}
@@ -300,7 +306,7 @@ static inline int test_and_change_bit(int nr, volatile unsigned long *addr)
300 return oldbit; 306 return oldbit;
301} 307}
302 308
303static inline int constant_test_bit(unsigned int nr, const volatile unsigned long *addr) 309static __always_inline int constant_test_bit(unsigned int nr, const volatile unsigned long *addr)
304{ 310{
305 return ((1UL << (nr % BITS_PER_LONG)) & 311 return ((1UL << (nr % BITS_PER_LONG)) &
306 (((unsigned long *)addr)[nr / BITS_PER_LONG])) != 0; 312 (((unsigned long *)addr)[nr / BITS_PER_LONG])) != 0;
diff --git a/arch/x86/include/asm/byteorder.h b/arch/x86/include/asm/byteorder.h
index 7c49917e3d9d..b13a7a88f3eb 100644
--- a/arch/x86/include/asm/byteorder.h
+++ b/arch/x86/include/asm/byteorder.h
@@ -1,7 +1,6 @@
1#ifndef _ASM_X86_BYTEORDER_H 1#ifndef _ASM_X86_BYTEORDER_H
2#define _ASM_X86_BYTEORDER_H 2#define _ASM_X86_BYTEORDER_H
3 3
4#include <asm/swab.h>
5#include <linux/byteorder/little_endian.h> 4#include <linux/byteorder/little_endian.h>
6 5
7#endif /* _ASM_X86_BYTEORDER_H */ 6#endif /* _ASM_X86_BYTEORDER_H */
diff --git a/arch/x86/include/asm/cpufeature.h b/arch/x86/include/asm/cpufeature.h
index ea408dcba513..7301e60dc4a8 100644
--- a/arch/x86/include/asm/cpufeature.h
+++ b/arch/x86/include/asm/cpufeature.h
@@ -93,6 +93,7 @@
93#define X86_FEATURE_XTOPOLOGY (3*32+22) /* cpu topology enum extensions */ 93#define X86_FEATURE_XTOPOLOGY (3*32+22) /* cpu topology enum extensions */
94#define X86_FEATURE_TSC_RELIABLE (3*32+23) /* TSC is known to be reliable */ 94#define X86_FEATURE_TSC_RELIABLE (3*32+23) /* TSC is known to be reliable */
95#define X86_FEATURE_NONSTOP_TSC (3*32+24) /* TSC does not stop in C states */ 95#define X86_FEATURE_NONSTOP_TSC (3*32+24) /* TSC does not stop in C states */
96#define X86_FEATURE_CLFLUSH_MONITOR (3*32+25) /* "" clflush reqd with monitor */
96 97
97/* Intel-defined CPU features, CPUID level 0x00000001 (ecx), word 4 */ 98/* Intel-defined CPU features, CPUID level 0x00000001 (ecx), word 4 */
98#define X86_FEATURE_XMM3 (4*32+ 0) /* "pni" SSE-3 */ 99#define X86_FEATURE_XMM3 (4*32+ 0) /* "pni" SSE-3 */
diff --git a/arch/x86/include/asm/dma-mapping.h b/arch/x86/include/asm/dma-mapping.h
index 4035357f5b9d..132a134d12f2 100644
--- a/arch/x86/include/asm/dma-mapping.h
+++ b/arch/x86/include/asm/dma-mapping.h
@@ -2,8 +2,8 @@
2#define _ASM_X86_DMA_MAPPING_H 2#define _ASM_X86_DMA_MAPPING_H
3 3
4/* 4/*
5 * IOMMU interface. See Documentation/DMA-mapping.txt and DMA-API.txt for 5 * IOMMU interface. See Documentation/PCI/PCI-DMA-mapping.txt and
6 * documentation. 6 * Documentation/DMA-API.txt for documentation.
7 */ 7 */
8 8
9#include <linux/scatterlist.h> 9#include <linux/scatterlist.h>
diff --git a/arch/x86/include/asm/e820.h b/arch/x86/include/asm/e820.h
index 3d8ceddbd407..00d41ce4c844 100644
--- a/arch/x86/include/asm/e820.h
+++ b/arch/x86/include/asm/e820.h
@@ -49,6 +49,7 @@
49#define E820_RESERVED_KERN 128 49#define E820_RESERVED_KERN 128
50 50
51#ifndef __ASSEMBLY__ 51#ifndef __ASSEMBLY__
52#include <linux/types.h>
52struct e820entry { 53struct e820entry {
53 __u64 addr; /* start of memory segment */ 54 __u64 addr; /* start of memory segment */
54 __u64 size; /* size of memory segment */ 55 __u64 size; /* size of memory segment */
diff --git a/arch/x86/include/asm/io.h b/arch/x86/include/asm/io.h
index 05cfed4485fa..1dbbdf4be9b4 100644
--- a/arch/x86/include/asm/io.h
+++ b/arch/x86/include/asm/io.h
@@ -99,7 +99,6 @@ extern void __iomem *ioremap_wc(unsigned long offset, unsigned long size);
99 * A boot-time mapping is currently limited to at most 16 pages. 99 * A boot-time mapping is currently limited to at most 16 pages.
100 */ 100 */
101extern void early_ioremap_init(void); 101extern void early_ioremap_init(void);
102extern void early_ioremap_clear(void);
103extern void early_ioremap_reset(void); 102extern void early_ioremap_reset(void);
104extern void __iomem *early_ioremap(unsigned long offset, unsigned long size); 103extern void __iomem *early_ioremap(unsigned long offset, unsigned long size);
105extern void __iomem *early_memremap(unsigned long offset, unsigned long size); 104extern void __iomem *early_memremap(unsigned long offset, unsigned long size);
diff --git a/arch/x86/include/asm/kvm.h b/arch/x86/include/asm/kvm.h
index b95162af0bf6..886c9402ec45 100644
--- a/arch/x86/include/asm/kvm.h
+++ b/arch/x86/include/asm/kvm.h
@@ -6,9 +6,16 @@
6 * 6 *
7 */ 7 */
8 8
9#include <asm/types.h> 9#include <linux/types.h>
10#include <linux/ioctl.h> 10#include <linux/ioctl.h>
11 11
12/* Select x86 specific features in <linux/kvm.h> */
13#define __KVM_HAVE_PIT
14#define __KVM_HAVE_IOAPIC
15#define __KVM_HAVE_DEVICE_ASSIGNMENT
16#define __KVM_HAVE_MSI
17#define __KVM_HAVE_USER_NMI
18
12/* Architectural interrupt line count. */ 19/* Architectural interrupt line count. */
13#define KVM_NR_INTERRUPTS 256 20#define KVM_NR_INTERRUPTS 256
14 21
diff --git a/arch/x86/include/asm/mach-default/mach_wakecpu.h b/arch/x86/include/asm/mach-default/mach_wakecpu.h
index ceb013660146..89897a6a65b9 100644
--- a/arch/x86/include/asm/mach-default/mach_wakecpu.h
+++ b/arch/x86/include/asm/mach-default/mach_wakecpu.h
@@ -24,7 +24,13 @@ static inline void restore_NMI_vector(unsigned short *high, unsigned short *low)
24{ 24{
25} 25}
26 26
27#ifdef CONFIG_SMP
27extern void __inquire_remote_apic(int apicid); 28extern void __inquire_remote_apic(int apicid);
29#else /* CONFIG_SMP */
30static inline void __inquire_remote_apic(int apicid)
31{
32}
33#endif /* CONFIG_SMP */
28 34
29static inline void inquire_remote_apic(int apicid) 35static inline void inquire_remote_apic(int apicid)
30{ 36{
diff --git a/arch/x86/include/asm/math_emu.h b/arch/x86/include/asm/math_emu.h
index 5a65b107ad58..031f6266f425 100644
--- a/arch/x86/include/asm/math_emu.h
+++ b/arch/x86/include/asm/math_emu.h
@@ -1,31 +1,18 @@
1#ifndef _ASM_X86_MATH_EMU_H 1#ifndef _ASM_X86_MATH_EMU_H
2#define _ASM_X86_MATH_EMU_H 2#define _ASM_X86_MATH_EMU_H
3 3
4#include <asm/ptrace.h>
5#include <asm/vm86.h>
6
4/* This structure matches the layout of the data saved to the stack 7/* This structure matches the layout of the data saved to the stack
5 following a device-not-present interrupt, part of it saved 8 following a device-not-present interrupt, part of it saved
6 automatically by the 80386/80486. 9 automatically by the 80386/80486.
7 */ 10 */
8struct info { 11struct math_emu_info {
9 long ___orig_eip; 12 long ___orig_eip;
10 long ___ebx; 13 union {
11 long ___ecx; 14 struct pt_regs *regs;
12 long ___edx; 15 struct kernel_vm86_regs *vm86;
13 long ___esi; 16 };
14 long ___edi;
15 long ___ebp;
16 long ___eax;
17 long ___ds;
18 long ___es;
19 long ___fs;
20 long ___orig_eax;
21 long ___eip;
22 long ___cs;
23 long ___eflags;
24 long ___esp;
25 long ___ss;
26 long ___vm86_es; /* This and the following only in vm86 mode */
27 long ___vm86_ds;
28 long ___vm86_fs;
29 long ___vm86_gs;
30}; 17};
31#endif /* _ASM_X86_MATH_EMU_H */ 18#endif /* _ASM_X86_MATH_EMU_H */
diff --git a/arch/x86/include/asm/mce.h b/arch/x86/include/asm/mce.h
index 1d6e17c2f23a..32c6e17b960b 100644
--- a/arch/x86/include/asm/mce.h
+++ b/arch/x86/include/asm/mce.h
@@ -3,8 +3,8 @@
3 3
4#ifdef __x86_64__ 4#ifdef __x86_64__
5 5
6#include <linux/types.h>
6#include <asm/ioctls.h> 7#include <asm/ioctls.h>
7#include <asm/types.h>
8 8
9/* 9/*
10 * Machine Check support for x86 10 * Machine Check support for x86
@@ -115,8 +115,6 @@ extern int mce_notify_user(void);
115 115
116#endif /* !CONFIG_X86_32 */ 116#endif /* !CONFIG_X86_32 */
117 117
118
119
120#ifdef CONFIG_X86_MCE 118#ifdef CONFIG_X86_MCE
121extern void mcheck_init(struct cpuinfo_x86 *c); 119extern void mcheck_init(struct cpuinfo_x86 *c);
122#else 120#else
@@ -126,5 +124,4 @@ extern void stop_mce(void);
126extern void restart_mce(void); 124extern void restart_mce(void);
127 125
128#endif /* __KERNEL__ */ 126#endif /* __KERNEL__ */
129
130#endif /* _ASM_X86_MCE_H */ 127#endif /* _ASM_X86_MCE_H */
diff --git a/arch/x86/include/asm/mmzone_32.h b/arch/x86/include/asm/mmzone_32.h
index 07f1af494ca5..105fb90a0635 100644
--- a/arch/x86/include/asm/mmzone_32.h
+++ b/arch/x86/include/asm/mmzone_32.h
@@ -32,8 +32,6 @@ static inline void get_memcfg_numa(void)
32 get_memcfg_numa_flat(); 32 get_memcfg_numa_flat();
33} 33}
34 34
35extern int early_pfn_to_nid(unsigned long pfn);
36
37extern void resume_map_numa_kva(pgd_t *pgd); 35extern void resume_map_numa_kva(pgd_t *pgd);
38 36
39#else /* !CONFIG_NUMA */ 37#else /* !CONFIG_NUMA */
diff --git a/arch/x86/include/asm/mmzone_64.h b/arch/x86/include/asm/mmzone_64.h
index a5b3817d4b9e..a29f48c2a322 100644
--- a/arch/x86/include/asm/mmzone_64.h
+++ b/arch/x86/include/asm/mmzone_64.h
@@ -40,8 +40,6 @@ static inline __attribute__((pure)) int phys_to_nid(unsigned long addr)
40#define node_end_pfn(nid) (NODE_DATA(nid)->node_start_pfn + \ 40#define node_end_pfn(nid) (NODE_DATA(nid)->node_start_pfn + \
41 NODE_DATA(nid)->node_spanned_pages) 41 NODE_DATA(nid)->node_spanned_pages)
42 42
43extern int early_pfn_to_nid(unsigned long pfn);
44
45#ifdef CONFIG_NUMA_EMU 43#ifdef CONFIG_NUMA_EMU
46#define FAKE_NODE_MIN_SIZE (64 * 1024 * 1024) 44#define FAKE_NODE_MIN_SIZE (64 * 1024 * 1024)
47#define FAKE_NODE_MIN_HASH_MASK (~(FAKE_NODE_MIN_SIZE - 1UL)) 45#define FAKE_NODE_MIN_HASH_MASK (~(FAKE_NODE_MIN_SIZE - 1UL))
diff --git a/arch/x86/include/asm/mpspec.h b/arch/x86/include/asm/mpspec.h
index 62d14ce3cd00..bd22f2a3713f 100644
--- a/arch/x86/include/asm/mpspec.h
+++ b/arch/x86/include/asm/mpspec.h
@@ -60,6 +60,7 @@ extern void mp_override_legacy_irq(u8 bus_irq, u8 polarity, u8 trigger,
60 u32 gsi); 60 u32 gsi);
61extern void mp_config_acpi_legacy_irqs(void); 61extern void mp_config_acpi_legacy_irqs(void);
62extern int mp_register_gsi(u32 gsi, int edge_level, int active_high_low); 62extern int mp_register_gsi(u32 gsi, int edge_level, int active_high_low);
63extern int acpi_probe_gsi(void);
63#ifdef CONFIG_X86_IO_APIC 64#ifdef CONFIG_X86_IO_APIC
64extern int mp_config_acpi_gsi(unsigned char number, unsigned int devfn, u8 pin, 65extern int mp_config_acpi_gsi(unsigned char number, unsigned int devfn, u8 pin,
65 u32 gsi, int triggering, int polarity); 66 u32 gsi, int triggering, int polarity);
@@ -71,6 +72,11 @@ mp_config_acpi_gsi(unsigned char number, unsigned int devfn, u8 pin,
71 return 0; 72 return 0;
72} 73}
73#endif 74#endif
75#else /* !CONFIG_ACPI: */
76static inline int acpi_probe_gsi(void)
77{
78 return 0;
79}
74#endif /* CONFIG_ACPI */ 80#endif /* CONFIG_ACPI */
75 81
76#define PHYSID_ARRAY_SIZE BITS_TO_LONGS(MAX_APICS) 82#define PHYSID_ARRAY_SIZE BITS_TO_LONGS(MAX_APICS)
diff --git a/arch/x86/include/asm/msr-index.h b/arch/x86/include/asm/msr-index.h
index cb58643947b9..358acc59ae04 100644
--- a/arch/x86/include/asm/msr-index.h
+++ b/arch/x86/include/asm/msr-index.h
@@ -202,6 +202,35 @@
202#define MSR_IA32_THERM_STATUS 0x0000019c 202#define MSR_IA32_THERM_STATUS 0x0000019c
203#define MSR_IA32_MISC_ENABLE 0x000001a0 203#define MSR_IA32_MISC_ENABLE 0x000001a0
204 204
205/* MISC_ENABLE bits: architectural */
206#define MSR_IA32_MISC_ENABLE_FAST_STRING (1ULL << 0)
207#define MSR_IA32_MISC_ENABLE_TCC (1ULL << 1)
208#define MSR_IA32_MISC_ENABLE_EMON (1ULL << 7)
209#define MSR_IA32_MISC_ENABLE_BTS_UNAVAIL (1ULL << 11)
210#define MSR_IA32_MISC_ENABLE_PEBS_UNAVAIL (1ULL << 12)
211#define MSR_IA32_MISC_ENABLE_ENHANCED_SPEEDSTEP (1ULL << 16)
212#define MSR_IA32_MISC_ENABLE_MWAIT (1ULL << 18)
213#define MSR_IA32_MISC_ENABLE_LIMIT_CPUID (1ULL << 22)
214#define MSR_IA32_MISC_ENABLE_XTPR_DISABLE (1ULL << 23)
215#define MSR_IA32_MISC_ENABLE_XD_DISABLE (1ULL << 34)
216
217/* MISC_ENABLE bits: model-specific, meaning may vary from core to core */
218#define MSR_IA32_MISC_ENABLE_X87_COMPAT (1ULL << 2)
219#define MSR_IA32_MISC_ENABLE_TM1 (1ULL << 3)
220#define MSR_IA32_MISC_ENABLE_SPLIT_LOCK_DISABLE (1ULL << 4)
221#define MSR_IA32_MISC_ENABLE_L3CACHE_DISABLE (1ULL << 6)
222#define MSR_IA32_MISC_ENABLE_SUPPRESS_LOCK (1ULL << 8)
223#define MSR_IA32_MISC_ENABLE_PREFETCH_DISABLE (1ULL << 9)
224#define MSR_IA32_MISC_ENABLE_FERR (1ULL << 10)
225#define MSR_IA32_MISC_ENABLE_FERR_MULTIPLEX (1ULL << 10)
226#define MSR_IA32_MISC_ENABLE_TM2 (1ULL << 13)
227#define MSR_IA32_MISC_ENABLE_ADJ_PREF_DISABLE (1ULL << 19)
228#define MSR_IA32_MISC_ENABLE_SPEEDSTEP_LOCK (1ULL << 20)
229#define MSR_IA32_MISC_ENABLE_L1D_CONTEXT (1ULL << 24)
230#define MSR_IA32_MISC_ENABLE_DCU_PREF_DISABLE (1ULL << 37)
231#define MSR_IA32_MISC_ENABLE_TURBO_DISABLE (1ULL << 38)
232#define MSR_IA32_MISC_ENABLE_IP_PREF_DISABLE (1ULL << 39)
233
205/* Intel Model 6 */ 234/* Intel Model 6 */
206#define MSR_P6_EVNTSEL0 0x00000186 235#define MSR_P6_EVNTSEL0 0x00000186
207#define MSR_P6_EVNTSEL1 0x00000187 236#define MSR_P6_EVNTSEL1 0x00000187
diff --git a/arch/x86/include/asm/mtrr.h b/arch/x86/include/asm/mtrr.h
index cb988aab716d..a51ada8467de 100644
--- a/arch/x86/include/asm/mtrr.h
+++ b/arch/x86/include/asm/mtrr.h
@@ -23,6 +23,7 @@
23#ifndef _ASM_X86_MTRR_H 23#ifndef _ASM_X86_MTRR_H
24#define _ASM_X86_MTRR_H 24#define _ASM_X86_MTRR_H
25 25
26#include <linux/types.h>
26#include <linux/ioctl.h> 27#include <linux/ioctl.h>
27#include <linux/errno.h> 28#include <linux/errno.h>
28 29
@@ -58,15 +59,15 @@ struct mtrr_gentry {
58#endif /* !__i386__ */ 59#endif /* !__i386__ */
59 60
60struct mtrr_var_range { 61struct mtrr_var_range {
61 u32 base_lo; 62 __u32 base_lo;
62 u32 base_hi; 63 __u32 base_hi;
63 u32 mask_lo; 64 __u32 mask_lo;
64 u32 mask_hi; 65 __u32 mask_hi;
65}; 66};
66 67
67/* In the Intel processor's MTRR interface, the MTRR type is always held in 68/* In the Intel processor's MTRR interface, the MTRR type is always held in
68 an 8 bit field: */ 69 an 8 bit field: */
69typedef u8 mtrr_type; 70typedef __u8 mtrr_type;
70 71
71#define MTRR_NUM_FIXED_RANGES 88 72#define MTRR_NUM_FIXED_RANGES 88
72#define MTRR_MAX_VAR_RANGES 256 73#define MTRR_MAX_VAR_RANGES 256
diff --git a/arch/x86/include/asm/page.h b/arch/x86/include/asm/page.h
index e9873a2e8695..776579119a00 100644
--- a/arch/x86/include/asm/page.h
+++ b/arch/x86/include/asm/page.h
@@ -57,7 +57,6 @@ typedef struct { pgdval_t pgd; } pgd_t;
57typedef struct { pgprotval_t pgprot; } pgprot_t; 57typedef struct { pgprotval_t pgprot; } pgprot_t;
58 58
59extern int page_is_ram(unsigned long pagenr); 59extern int page_is_ram(unsigned long pagenr);
60extern int pagerange_is_ram(unsigned long start, unsigned long end);
61extern int devmem_is_allowed(unsigned long pagenr); 60extern int devmem_is_allowed(unsigned long pagenr);
62extern void map_devmem(unsigned long pfn, unsigned long size, 61extern void map_devmem(unsigned long pfn, unsigned long size,
63 pgprot_t vma_prot); 62 pgprot_t vma_prot);
diff --git a/arch/x86/include/asm/paravirt.h b/arch/x86/include/asm/paravirt.h
index ba3e2ff6aedc..e299287e8e33 100644
--- a/arch/x86/include/asm/paravirt.h
+++ b/arch/x86/include/asm/paravirt.h
@@ -1352,14 +1352,7 @@ static inline void arch_leave_lazy_cpu_mode(void)
1352 PVOP_VCALL0(pv_cpu_ops.lazy_mode.leave); 1352 PVOP_VCALL0(pv_cpu_ops.lazy_mode.leave);
1353} 1353}
1354 1354
1355static inline void arch_flush_lazy_cpu_mode(void) 1355void arch_flush_lazy_cpu_mode(void);
1356{
1357 if (unlikely(paravirt_get_lazy_mode() == PARAVIRT_LAZY_CPU)) {
1358 arch_leave_lazy_cpu_mode();
1359 arch_enter_lazy_cpu_mode();
1360 }
1361}
1362
1363 1356
1364#define __HAVE_ARCH_ENTER_LAZY_MMU_MODE 1357#define __HAVE_ARCH_ENTER_LAZY_MMU_MODE
1365static inline void arch_enter_lazy_mmu_mode(void) 1358static inline void arch_enter_lazy_mmu_mode(void)
@@ -1372,13 +1365,7 @@ static inline void arch_leave_lazy_mmu_mode(void)
1372 PVOP_VCALL0(pv_mmu_ops.lazy_mode.leave); 1365 PVOP_VCALL0(pv_mmu_ops.lazy_mode.leave);
1373} 1366}
1374 1367
1375static inline void arch_flush_lazy_mmu_mode(void) 1368void arch_flush_lazy_mmu_mode(void);
1376{
1377 if (unlikely(paravirt_get_lazy_mode() == PARAVIRT_LAZY_MMU)) {
1378 arch_leave_lazy_mmu_mode();
1379 arch_enter_lazy_mmu_mode();
1380 }
1381}
1382 1369
1383static inline void __set_fixmap(unsigned /* enum fixed_addresses */ idx, 1370static inline void __set_fixmap(unsigned /* enum fixed_addresses */ idx,
1384 unsigned long phys, pgprot_t flags) 1371 unsigned long phys, pgprot_t flags)
@@ -1402,6 +1389,7 @@ static inline int __raw_spin_is_contended(struct raw_spinlock *lock)
1402{ 1389{
1403 return PVOP_CALL1(int, pv_lock_ops.spin_is_contended, lock); 1390 return PVOP_CALL1(int, pv_lock_ops.spin_is_contended, lock);
1404} 1391}
1392#define __raw_spin_is_contended __raw_spin_is_contended
1405 1393
1406static __always_inline void __raw_spin_lock(struct raw_spinlock *lock) 1394static __always_inline void __raw_spin_lock(struct raw_spinlock *lock)
1407{ 1395{
diff --git a/arch/x86/include/asm/pgalloc.h b/arch/x86/include/asm/pgalloc.h
index cb7c151a8bff..dd14c54ac718 100644
--- a/arch/x86/include/asm/pgalloc.h
+++ b/arch/x86/include/asm/pgalloc.h
@@ -42,6 +42,7 @@ static inline void pte_free_kernel(struct mm_struct *mm, pte_t *pte)
42 42
43static inline void pte_free(struct mm_struct *mm, struct page *pte) 43static inline void pte_free(struct mm_struct *mm, struct page *pte)
44{ 44{
45 pgtable_page_dtor(pte);
45 __free_page(pte); 46 __free_page(pte);
46} 47}
47 48
diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h
index 83e69f4a37f0..4f5af8447d54 100644
--- a/arch/x86/include/asm/pgtable.h
+++ b/arch/x86/include/asm/pgtable.h
@@ -302,16 +302,30 @@ static inline pte_t pte_mkspecial(pte_t pte)
302 302
303extern pteval_t __supported_pte_mask; 303extern pteval_t __supported_pte_mask;
304 304
305/*
306 * Mask out unsupported bits in a present pgprot. Non-present pgprots
307 * can use those bits for other purposes, so leave them be.
308 */
309static inline pgprotval_t massage_pgprot(pgprot_t pgprot)
310{
311 pgprotval_t protval = pgprot_val(pgprot);
312
313 if (protval & _PAGE_PRESENT)
314 protval &= __supported_pte_mask;
315
316 return protval;
317}
318
305static inline pte_t pfn_pte(unsigned long page_nr, pgprot_t pgprot) 319static inline pte_t pfn_pte(unsigned long page_nr, pgprot_t pgprot)
306{ 320{
307 return __pte((((phys_addr_t)page_nr << PAGE_SHIFT) | 321 return __pte(((phys_addr_t)page_nr << PAGE_SHIFT) |
308 pgprot_val(pgprot)) & __supported_pte_mask); 322 massage_pgprot(pgprot));
309} 323}
310 324
311static inline pmd_t pfn_pmd(unsigned long page_nr, pgprot_t pgprot) 325static inline pmd_t pfn_pmd(unsigned long page_nr, pgprot_t pgprot)
312{ 326{
313 return __pmd((((phys_addr_t)page_nr << PAGE_SHIFT) | 327 return __pmd(((phys_addr_t)page_nr << PAGE_SHIFT) |
314 pgprot_val(pgprot)) & __supported_pte_mask); 328 massage_pgprot(pgprot));
315} 329}
316 330
317static inline pte_t pte_modify(pte_t pte, pgprot_t newprot) 331static inline pte_t pte_modify(pte_t pte, pgprot_t newprot)
@@ -323,7 +337,7 @@ static inline pte_t pte_modify(pte_t pte, pgprot_t newprot)
323 * the newprot (if present): 337 * the newprot (if present):
324 */ 338 */
325 val &= _PAGE_CHG_MASK; 339 val &= _PAGE_CHG_MASK;
326 val |= pgprot_val(newprot) & (~_PAGE_CHG_MASK) & __supported_pte_mask; 340 val |= massage_pgprot(newprot) & ~_PAGE_CHG_MASK;
327 341
328 return __pte(val); 342 return __pte(val);
329} 343}
@@ -339,7 +353,26 @@ static inline pgprot_t pgprot_modify(pgprot_t oldprot, pgprot_t newprot)
339 353
340#define pte_pgprot(x) __pgprot(pte_flags(x) & PTE_FLAGS_MASK) 354#define pte_pgprot(x) __pgprot(pte_flags(x) & PTE_FLAGS_MASK)
341 355
342#define canon_pgprot(p) __pgprot(pgprot_val(p) & __supported_pte_mask) 356#define canon_pgprot(p) __pgprot(massage_pgprot(p))
357
358static inline int is_new_memtype_allowed(unsigned long flags,
359 unsigned long new_flags)
360{
361 /*
362 * Certain new memtypes are not allowed with certain
363 * requested memtype:
364 * - request is uncached, return cannot be write-back
365 * - request is write-combine, return cannot be write-back
366 */
367 if ((flags == _PAGE_CACHE_UC_MINUS &&
368 new_flags == _PAGE_CACHE_WB) ||
369 (flags == _PAGE_CACHE_WC &&
370 new_flags == _PAGE_CACHE_WB)) {
371 return 0;
372 }
373
374 return 1;
375}
343 376
344#ifndef __ASSEMBLY__ 377#ifndef __ASSEMBLY__
345/* Indicate that x86 has its own track and untrack pfn vma functions */ 378/* Indicate that x86 has its own track and untrack pfn vma functions */
diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h
index 091cd8855f2e..3bfd5235a9eb 100644
--- a/arch/x86/include/asm/processor.h
+++ b/arch/x86/include/asm/processor.h
@@ -353,7 +353,7 @@ struct i387_soft_struct {
353 u8 no_update; 353 u8 no_update;
354 u8 rm; 354 u8 rm;
355 u8 alimit; 355 u8 alimit;
356 struct info *info; 356 struct math_emu_info *info;
357 u32 entry_eip; 357 u32 entry_eip;
358}; 358};
359 359
diff --git a/arch/x86/include/asm/ptrace-abi.h b/arch/x86/include/asm/ptrace-abi.h
index 25f1bb8fc626..8e0f8d199e05 100644
--- a/arch/x86/include/asm/ptrace-abi.h
+++ b/arch/x86/include/asm/ptrace-abi.h
@@ -83,7 +83,7 @@
83#ifdef CONFIG_X86_PTRACE_BTS 83#ifdef CONFIG_X86_PTRACE_BTS
84 84
85#ifndef __ASSEMBLY__ 85#ifndef __ASSEMBLY__
86#include <asm/types.h> 86#include <linux/types.h>
87 87
88/* configuration/status structure used in PTRACE_BTS_CONFIG and 88/* configuration/status structure used in PTRACE_BTS_CONFIG and
89 PTRACE_BTS_STATUS commands. 89 PTRACE_BTS_STATUS commands.
diff --git a/arch/x86/include/asm/sigcontext.h b/arch/x86/include/asm/sigcontext.h
index 0afcb5e58acc..ec666491aaa4 100644
--- a/arch/x86/include/asm/sigcontext.h
+++ b/arch/x86/include/asm/sigcontext.h
@@ -2,7 +2,7 @@
2#define _ASM_X86_SIGCONTEXT_H 2#define _ASM_X86_SIGCONTEXT_H
3 3
4#include <linux/compiler.h> 4#include <linux/compiler.h>
5#include <asm/types.h> 5#include <linux/types.h>
6 6
7#define FP_XSTATE_MAGIC1 0x46505853U 7#define FP_XSTATE_MAGIC1 0x46505853U
8#define FP_XSTATE_MAGIC2 0x46505845U 8#define FP_XSTATE_MAGIC2 0x46505845U
diff --git a/arch/x86/include/asm/sigcontext32.h b/arch/x86/include/asm/sigcontext32.h
index 6126188cf3a9..ad1478c4ae12 100644
--- a/arch/x86/include/asm/sigcontext32.h
+++ b/arch/x86/include/asm/sigcontext32.h
@@ -1,6 +1,8 @@
1#ifndef _ASM_X86_SIGCONTEXT32_H 1#ifndef _ASM_X86_SIGCONTEXT32_H
2#define _ASM_X86_SIGCONTEXT32_H 2#define _ASM_X86_SIGCONTEXT32_H
3 3
4#include <linux/types.h>
5
4/* signal context for 32bit programs. */ 6/* signal context for 32bit programs. */
5 7
6#define X86_FXSR_MAGIC 0x0000 8#define X86_FXSR_MAGIC 0x0000
diff --git a/arch/x86/include/asm/spinlock.h b/arch/x86/include/asm/spinlock.h
index d17c91981da2..8247e94ac6b1 100644
--- a/arch/x86/include/asm/spinlock.h
+++ b/arch/x86/include/asm/spinlock.h
@@ -245,6 +245,7 @@ static inline int __raw_spin_is_contended(raw_spinlock_t *lock)
245{ 245{
246 return __ticket_spin_is_contended(lock); 246 return __ticket_spin_is_contended(lock);
247} 247}
248#define __raw_spin_is_contended __raw_spin_is_contended
248 249
249static __always_inline void __raw_spin_lock(raw_spinlock_t *lock) 250static __always_inline void __raw_spin_lock(raw_spinlock_t *lock)
250{ 251{
diff --git a/arch/x86/include/asm/swab.h b/arch/x86/include/asm/swab.h
index 306d4178ffc9..557cd9f00661 100644
--- a/arch/x86/include/asm/swab.h
+++ b/arch/x86/include/asm/swab.h
@@ -1,7 +1,7 @@
1#ifndef _ASM_X86_SWAB_H 1#ifndef _ASM_X86_SWAB_H
2#define _ASM_X86_SWAB_H 2#define _ASM_X86_SWAB_H
3 3
4#include <asm/types.h> 4#include <linux/types.h>
5#include <linux/compiler.h> 5#include <linux/compiler.h>
6 6
7static inline __attribute_const__ __u32 __arch_swab32(__u32 val) 7static inline __attribute_const__ __u32 __arch_swab32(__u32 val)
diff --git a/arch/x86/include/asm/syscalls.h b/arch/x86/include/asm/syscalls.h
index 9c6797c3e56c..c0b0bda754ee 100644
--- a/arch/x86/include/asm/syscalls.h
+++ b/arch/x86/include/asm/syscalls.h
@@ -40,7 +40,7 @@ asmlinkage int sys_sigaction(int, const struct old_sigaction __user *,
40 struct old_sigaction __user *); 40 struct old_sigaction __user *);
41asmlinkage int sys_sigaltstack(unsigned long); 41asmlinkage int sys_sigaltstack(unsigned long);
42asmlinkage unsigned long sys_sigreturn(unsigned long); 42asmlinkage unsigned long sys_sigreturn(unsigned long);
43asmlinkage int sys_rt_sigreturn(struct pt_regs); 43asmlinkage int sys_rt_sigreturn(unsigned long);
44 44
45/* kernel/ioport.c */ 45/* kernel/ioport.c */
46asmlinkage long sys_iopl(unsigned long); 46asmlinkage long sys_iopl(unsigned long);
diff --git a/arch/x86/include/asm/timex.h b/arch/x86/include/asm/timex.h
index 1287dc1347d6..b5c9d45c981f 100644
--- a/arch/x86/include/asm/timex.h
+++ b/arch/x86/include/asm/timex.h
@@ -1,18 +1,13 @@
1/* x86 architecture timex specifications */
2#ifndef _ASM_X86_TIMEX_H 1#ifndef _ASM_X86_TIMEX_H
3#define _ASM_X86_TIMEX_H 2#define _ASM_X86_TIMEX_H
4 3
5#include <asm/processor.h> 4#include <asm/processor.h>
6#include <asm/tsc.h> 5#include <asm/tsc.h>
7 6
8#ifdef CONFIG_X86_ELAN 7/* The PIT ticks at this frequency (in HZ): */
9# define PIT_TICK_RATE 1189200 /* AMD Elan has different frequency! */ 8#define PIT_TICK_RATE 1193182
10#elif defined(CONFIG_X86_RDC321X) 9
11# define PIT_TICK_RATE 1041667 /* Underlying HZ for R8610 */ 10#define CLOCK_TICK_RATE PIT_TICK_RATE
12#else
13# define PIT_TICK_RATE 1193182 /* Underlying HZ */
14#endif
15#define CLOCK_TICK_RATE PIT_TICK_RATE
16 11
17#define ARCH_HAS_READ_CURRENT_TIMER 12#define ARCH_HAS_READ_CURRENT_TIMER
18 13
diff --git a/arch/x86/include/asm/traps.h b/arch/x86/include/asm/traps.h
index 2ee0a3bceedf..cf3bb053da0b 100644
--- a/arch/x86/include/asm/traps.h
+++ b/arch/x86/include/asm/traps.h
@@ -41,7 +41,7 @@ dotraplinkage void do_int3(struct pt_regs *, long);
41dotraplinkage void do_overflow(struct pt_regs *, long); 41dotraplinkage void do_overflow(struct pt_regs *, long);
42dotraplinkage void do_bounds(struct pt_regs *, long); 42dotraplinkage void do_bounds(struct pt_regs *, long);
43dotraplinkage void do_invalid_op(struct pt_regs *, long); 43dotraplinkage void do_invalid_op(struct pt_regs *, long);
44dotraplinkage void do_device_not_available(struct pt_regs *, long); 44dotraplinkage void do_device_not_available(struct pt_regs);
45dotraplinkage void do_coprocessor_segment_overrun(struct pt_regs *, long); 45dotraplinkage void do_coprocessor_segment_overrun(struct pt_regs *, long);
46dotraplinkage void do_invalid_TSS(struct pt_regs *, long); 46dotraplinkage void do_invalid_TSS(struct pt_regs *, long);
47dotraplinkage void do_segment_not_present(struct pt_regs *, long); 47dotraplinkage void do_segment_not_present(struct pt_regs *, long);
@@ -77,7 +77,7 @@ extern int panic_on_unrecovered_nmi;
77extern int kstack_depth_to_print; 77extern int kstack_depth_to_print;
78 78
79void math_error(void __user *); 79void math_error(void __user *);
80asmlinkage void math_emulate(long); 80void math_emulate(struct math_emu_info *);
81#ifdef CONFIG_X86_32 81#ifdef CONFIG_X86_32
82unsigned long patch_espfix_desc(unsigned long, unsigned long); 82unsigned long patch_espfix_desc(unsigned long, unsigned long);
83#else 83#else
diff --git a/arch/x86/include/asm/xen/page.h b/arch/x86/include/asm/xen/page.h
index 7ef617ef1df3..4bd990ee43df 100644
--- a/arch/x86/include/asm/xen/page.h
+++ b/arch/x86/include/asm/xen/page.h
@@ -137,7 +137,7 @@ static inline pte_t mfn_pte(unsigned long page_nr, pgprot_t pgprot)
137 pte_t pte; 137 pte_t pte;
138 138
139 pte.pte = ((phys_addr_t)page_nr << PAGE_SHIFT) | 139 pte.pte = ((phys_addr_t)page_nr << PAGE_SHIFT) |
140 (pgprot_val(pgprot) & __supported_pte_mask); 140 massage_pgprot(pgprot);
141 141
142 return pte; 142 return pte;
143} 143}
diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c
index d37593c2f438..7678f10c4568 100644
--- a/arch/x86/kernel/acpi/boot.c
+++ b/arch/x86/kernel/acpi/boot.c
@@ -973,6 +973,29 @@ void __init mp_register_ioapic(int id, u32 address, u32 gsi_base)
973 nr_ioapics++; 973 nr_ioapics++;
974} 974}
975 975
976int __init acpi_probe_gsi(void)
977{
978 int idx;
979 int gsi;
980 int max_gsi = 0;
981
982 if (acpi_disabled)
983 return 0;
984
985 if (!acpi_ioapic)
986 return 0;
987
988 max_gsi = 0;
989 for (idx = 0; idx < nr_ioapics; idx++) {
990 gsi = mp_ioapic_routing[idx].gsi_end;
991
992 if (gsi > max_gsi)
993 max_gsi = gsi;
994 }
995
996 return max_gsi + 1;
997}
998
976static void assign_to_mp_irq(struct mp_config_intsrc *m, 999static void assign_to_mp_irq(struct mp_config_intsrc *m,
977 struct mp_config_intsrc *mp_irq) 1000 struct mp_config_intsrc *mp_irq)
978{ 1001{
diff --git a/arch/x86/kernel/acpi/sleep.c b/arch/x86/kernel/acpi/sleep.c
index 707c1f6f95fa..a60c1f3bcb87 100644
--- a/arch/x86/kernel/acpi/sleep.c
+++ b/arch/x86/kernel/acpi/sleep.c
@@ -156,11 +156,11 @@ static int __init acpi_sleep_setup(char *str)
156#ifdef CONFIG_HIBERNATION 156#ifdef CONFIG_HIBERNATION
157 if (strncmp(str, "s4_nohwsig", 10) == 0) 157 if (strncmp(str, "s4_nohwsig", 10) == 0)
158 acpi_no_s4_hw_signature(); 158 acpi_no_s4_hw_signature();
159 if (strncmp(str, "s4_nonvs", 8) == 0)
160 acpi_s4_no_nvs();
159#endif 161#endif
160 if (strncmp(str, "old_ordering", 12) == 0) 162 if (strncmp(str, "old_ordering", 12) == 0)
161 acpi_old_suspend_ordering(); 163 acpi_old_suspend_ordering();
162 if (strncmp(str, "s4_nonvs", 8) == 0)
163 acpi_s4_no_nvs();
164 str = strchr(str, ','); 164 str = strchr(str, ',');
165 if (str != NULL) 165 if (str != NULL)
166 str += strspn(str, ", \t"); 166 str += strspn(str, ", \t");
diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c
index 566a08466b19..570f36e44e59 100644
--- a/arch/x86/kernel/apic.c
+++ b/arch/x86/kernel/apic.c
@@ -47,6 +47,7 @@
47#include <asm/proto.h> 47#include <asm/proto.h>
48#include <asm/apic.h> 48#include <asm/apic.h>
49#include <asm/i8259.h> 49#include <asm/i8259.h>
50#include <asm/smp.h>
50 51
51#include <mach_apic.h> 52#include <mach_apic.h>
52#include <mach_apicdef.h> 53#include <mach_apicdef.h>
@@ -861,7 +862,7 @@ void clear_local_APIC(void)
861 } 862 }
862 863
863 /* lets not touch this if we didn't frob it */ 864 /* lets not touch this if we didn't frob it */
864#if defined(CONFIG_X86_MCE_P4THERMAL) || defined(X86_MCE_INTEL) 865#if defined(CONFIG_X86_MCE_P4THERMAL) || defined(CONFIG_X86_MCE_INTEL)
865 if (maxlvt >= 5) { 866 if (maxlvt >= 5) {
866 v = apic_read(APIC_LVTTHMR); 867 v = apic_read(APIC_LVTTHMR);
867 apic_write(APIC_LVTTHMR, v | APIC_LVT_MASKED); 868 apic_write(APIC_LVTTHMR, v | APIC_LVT_MASKED);
@@ -894,6 +895,10 @@ void disable_local_APIC(void)
894{ 895{
895 unsigned int value; 896 unsigned int value;
896 897
898 /* APIC hasn't been mapped yet */
899 if (!apic_phys)
900 return;
901
897 clear_local_APIC(); 902 clear_local_APIC();
898 903
899 /* 904 /*
@@ -1431,7 +1436,7 @@ static int __init detect_init_APIC(void)
1431 switch (boot_cpu_data.x86_vendor) { 1436 switch (boot_cpu_data.x86_vendor) {
1432 case X86_VENDOR_AMD: 1437 case X86_VENDOR_AMD:
1433 if ((boot_cpu_data.x86 == 6 && boot_cpu_data.x86_model > 1) || 1438 if ((boot_cpu_data.x86 == 6 && boot_cpu_data.x86_model > 1) ||
1434 (boot_cpu_data.x86 == 15)) 1439 (boot_cpu_data.x86 >= 15))
1435 break; 1440 break;
1436 goto no_apic; 1441 goto no_apic;
1437 case X86_VENDOR_INTEL: 1442 case X86_VENDOR_INTEL:
@@ -1832,6 +1837,11 @@ void __cpuinit generic_processor_info(int apicid, int version)
1832 num_processors++; 1837 num_processors++;
1833 cpu = cpumask_next_zero(-1, cpu_present_mask); 1838 cpu = cpumask_next_zero(-1, cpu_present_mask);
1834 1839
1840 if (version != apic_version[boot_cpu_physical_apicid])
1841 WARN_ONCE(1,
1842 "ACPI: apic version mismatch, bootcpu: %x cpu %d: %x\n",
1843 apic_version[boot_cpu_physical_apicid], cpu, version);
1844
1835 physid_set(apicid, phys_cpu_present_map); 1845 physid_set(apicid, phys_cpu_present_map);
1836 if (apicid == boot_cpu_physical_apicid) { 1846 if (apicid == boot_cpu_physical_apicid) {
1837 /* 1847 /*
diff --git a/arch/x86/kernel/cpu/cpufreq/Kconfig b/arch/x86/kernel/cpu/cpufreq/Kconfig
index efae3b22a0ff..65792c2cc462 100644
--- a/arch/x86/kernel/cpu/cpufreq/Kconfig
+++ b/arch/x86/kernel/cpu/cpufreq/Kconfig
@@ -245,17 +245,6 @@ config X86_E_POWERSAVER
245 245
246comment "shared options" 246comment "shared options"
247 247
248config X86_ACPI_CPUFREQ_PROC_INTF
249 bool "/proc/acpi/processor/../performance interface (deprecated)"
250 depends on PROC_FS
251 depends on X86_ACPI_CPUFREQ || X86_POWERNOW_K7_ACPI || X86_POWERNOW_K8_ACPI
252 help
253 This enables the deprecated /proc/acpi/processor/../performance
254 interface. While it is helpful for debugging, the generic,
255 cross-architecture cpufreq interfaces should be used.
256
257 If in doubt, say N.
258
259config X86_SPEEDSTEP_LIB 248config X86_SPEEDSTEP_LIB
260 tristate 249 tristate
261 default (X86_SPEEDSTEP_ICH || X86_SPEEDSTEP_SMI || X86_P4_CLOCKMOD) 250 default (X86_SPEEDSTEP_ICH || X86_SPEEDSTEP_SMI || X86_P4_CLOCKMOD)
diff --git a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
index 06fcd8f9323c..4b1c319d30c3 100644
--- a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
+++ b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
@@ -145,7 +145,7 @@ typedef union {
145 145
146struct drv_cmd { 146struct drv_cmd {
147 unsigned int type; 147 unsigned int type;
148 cpumask_var_t mask; 148 const struct cpumask *mask;
149 drv_addr_union addr; 149 drv_addr_union addr;
150 u32 val; 150 u32 val;
151}; 151};
@@ -231,15 +231,9 @@ static u32 get_cur_val(const struct cpumask *mask)
231 return 0; 231 return 0;
232 } 232 }
233 233
234 if (unlikely(!alloc_cpumask_var(&cmd.mask, GFP_KERNEL))) 234 cmd.mask = mask;
235 return 0;
236
237 cpumask_copy(cmd.mask, mask);
238
239 drv_read(&cmd); 235 drv_read(&cmd);
240 236
241 free_cpumask_var(cmd.mask);
242
243 dprintk("get_cur_val = %u\n", cmd.val); 237 dprintk("get_cur_val = %u\n", cmd.val);
244 238
245 return cmd.val; 239 return cmd.val;
@@ -369,7 +363,7 @@ static unsigned int get_cur_freq_on_cpu(unsigned int cpu)
369 return freq; 363 return freq;
370} 364}
371 365
372static unsigned int check_freqs(const cpumask_t *mask, unsigned int freq, 366static unsigned int check_freqs(const struct cpumask *mask, unsigned int freq,
373 struct acpi_cpufreq_data *data) 367 struct acpi_cpufreq_data *data)
374{ 368{
375 unsigned int cur_freq; 369 unsigned int cur_freq;
@@ -404,9 +398,6 @@ static int acpi_cpufreq_target(struct cpufreq_policy *policy,
404 return -ENODEV; 398 return -ENODEV;
405 } 399 }
406 400
407 if (unlikely(!alloc_cpumask_var(&cmd.mask, GFP_KERNEL)))
408 return -ENOMEM;
409
410 perf = data->acpi_data; 401 perf = data->acpi_data;
411 result = cpufreq_frequency_table_target(policy, 402 result = cpufreq_frequency_table_target(policy,
412 data->freq_table, 403 data->freq_table,
@@ -451,9 +442,9 @@ static int acpi_cpufreq_target(struct cpufreq_policy *policy,
451 442
452 /* cpufreq holds the hotplug lock, so we are safe from here on */ 443 /* cpufreq holds the hotplug lock, so we are safe from here on */
453 if (policy->shared_type != CPUFREQ_SHARED_TYPE_ANY) 444 if (policy->shared_type != CPUFREQ_SHARED_TYPE_ANY)
454 cpumask_and(cmd.mask, cpu_online_mask, policy->cpus); 445 cmd.mask = policy->cpus;
455 else 446 else
456 cpumask_copy(cmd.mask, cpumask_of(policy->cpu)); 447 cmd.mask = cpumask_of(policy->cpu);
457 448
458 freqs.old = perf->states[perf->state].core_frequency * 1000; 449 freqs.old = perf->states[perf->state].core_frequency * 1000;
459 freqs.new = data->freq_table[next_state].frequency; 450 freqs.new = data->freq_table[next_state].frequency;
@@ -480,7 +471,6 @@ static int acpi_cpufreq_target(struct cpufreq_policy *policy,
480 perf->state = next_perf_state; 471 perf->state = next_perf_state;
481 472
482out: 473out:
483 free_cpumask_var(cmd.mask);
484 return result; 474 return result;
485} 475}
486 476
diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
index 5c28b37dea11..6428aa17b40e 100644
--- a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
+++ b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
@@ -939,10 +939,25 @@ static void powernow_k8_cpu_exit_acpi(struct powernow_k8_data *data)
939 free_cpumask_var(data->acpi_data.shared_cpu_map); 939 free_cpumask_var(data->acpi_data.shared_cpu_map);
940} 940}
941 941
942static int get_transition_latency(struct powernow_k8_data *data)
943{
944 int max_latency = 0;
945 int i;
946 for (i = 0; i < data->acpi_data.state_count; i++) {
947 int cur_latency = data->acpi_data.states[i].transition_latency
948 + data->acpi_data.states[i].bus_master_latency;
949 if (cur_latency > max_latency)
950 max_latency = cur_latency;
951 }
952 /* value in usecs, needs to be in nanoseconds */
953 return 1000 * max_latency;
954}
955
942#else 956#else
943static int powernow_k8_cpu_init_acpi(struct powernow_k8_data *data) { return -ENODEV; } 957static int powernow_k8_cpu_init_acpi(struct powernow_k8_data *data) { return -ENODEV; }
944static void powernow_k8_cpu_exit_acpi(struct powernow_k8_data *data) { return; } 958static void powernow_k8_cpu_exit_acpi(struct powernow_k8_data *data) { return; }
945static void powernow_k8_acpi_pst_values(struct powernow_k8_data *data, unsigned int index) { return; } 959static void powernow_k8_acpi_pst_values(struct powernow_k8_data *data, unsigned int index) { return; }
960static int get_transition_latency(struct powernow_k8_data *data) { return 0; }
946#endif /* CONFIG_X86_POWERNOW_K8_ACPI */ 961#endif /* CONFIG_X86_POWERNOW_K8_ACPI */
947 962
948/* Take a frequency, and issue the fid/vid transition command */ 963/* Take a frequency, and issue the fid/vid transition command */
@@ -1142,8 +1157,7 @@ static int __cpuinit powernowk8_cpu_init(struct cpufreq_policy *pol)
1142 data->cpu = pol->cpu; 1157 data->cpu = pol->cpu;
1143 data->currpstate = HW_PSTATE_INVALID; 1158 data->currpstate = HW_PSTATE_INVALID;
1144 1159
1145 rc = powernow_k8_cpu_init_acpi(data); 1160 if (powernow_k8_cpu_init_acpi(data)) {
1146 if (rc) {
1147 /* 1161 /*
1148 * Use the PSB BIOS structure. This is only availabe on 1162 * Use the PSB BIOS structure. This is only availabe on
1149 * an UP version, and is deprecated by AMD. 1163 * an UP version, and is deprecated by AMD.
@@ -1161,19 +1175,28 @@ static int __cpuinit powernowk8_cpu_init(struct cpufreq_policy *pol)
1161 "ACPI maintainers and complain to your BIOS " 1175 "ACPI maintainers and complain to your BIOS "
1162 "vendor.\n"); 1176 "vendor.\n");
1163#endif 1177#endif
1164 goto err_out; 1178 kfree(data);
1179 return -ENODEV;
1165 } 1180 }
1166 if (pol->cpu != 0) { 1181 if (pol->cpu != 0) {
1167 printk(KERN_ERR FW_BUG PFX "No ACPI _PSS objects for " 1182 printk(KERN_ERR FW_BUG PFX "No ACPI _PSS objects for "
1168 "CPU other than CPU0. Complain to your BIOS " 1183 "CPU other than CPU0. Complain to your BIOS "
1169 "vendor.\n"); 1184 "vendor.\n");
1170 goto err_out; 1185 kfree(data);
1186 return -ENODEV;
1171 } 1187 }
1172 rc = find_psb_table(data); 1188 rc = find_psb_table(data);
1173 if (rc) { 1189 if (rc) {
1174 goto err_out; 1190 kfree(data);
1191 return -ENODEV;
1175 } 1192 }
1176 } 1193 /* Take a crude guess here.
1194 * That guess was in microseconds, so multiply with 1000 */
1195 pol->cpuinfo.transition_latency = (
1196 ((data->rvo + 8) * data->vstable * VST_UNITS_20US) +
1197 ((1 << data->irt) * 30)) * 1000;
1198 } else /* ACPI _PSS objects available */
1199 pol->cpuinfo.transition_latency = get_transition_latency(data);
1177 1200
1178 /* only run on specific CPU from here on */ 1201 /* only run on specific CPU from here on */
1179 oldmask = current->cpus_allowed; 1202 oldmask = current->cpus_allowed;
@@ -1204,11 +1227,6 @@ static int __cpuinit powernowk8_cpu_init(struct cpufreq_policy *pol)
1204 cpumask_copy(pol->cpus, &per_cpu(cpu_core_map, pol->cpu)); 1227 cpumask_copy(pol->cpus, &per_cpu(cpu_core_map, pol->cpu));
1205 data->available_cores = pol->cpus; 1228 data->available_cores = pol->cpus;
1206 1229
1207 /* Take a crude guess here.
1208 * That guess was in microseconds, so multiply with 1000 */
1209 pol->cpuinfo.transition_latency = (((data->rvo + 8) * data->vstable * VST_UNITS_20US)
1210 + (3 * (1 << data->irt) * 10)) * 1000;
1211
1212 if (cpu_family == CPU_HW_PSTATE) 1230 if (cpu_family == CPU_HW_PSTATE)
1213 pol->cur = find_khz_freq_from_pstate(data->powernow_table, data->currpstate); 1231 pol->cur = find_khz_freq_from_pstate(data->powernow_table, data->currpstate);
1214 else 1232 else
diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c
index 8ea6929e974c..24ff26a38ade 100644
--- a/arch/x86/kernel/cpu/intel.c
+++ b/arch/x86/kernel/cpu/intel.c
@@ -29,6 +29,19 @@
29 29
30static void __cpuinit early_init_intel(struct cpuinfo_x86 *c) 30static void __cpuinit early_init_intel(struct cpuinfo_x86 *c)
31{ 31{
32 /* Unmask CPUID levels if masked: */
33 if (c->x86 > 6 || (c->x86 == 6 && c->x86_model >= 0xd)) {
34 u64 misc_enable;
35
36 rdmsrl(MSR_IA32_MISC_ENABLE, misc_enable);
37
38 if (misc_enable & MSR_IA32_MISC_ENABLE_LIMIT_CPUID) {
39 misc_enable &= ~MSR_IA32_MISC_ENABLE_LIMIT_CPUID;
40 wrmsrl(MSR_IA32_MISC_ENABLE, misc_enable);
41 c->cpuid_level = cpuid_eax(0);
42 }
43 }
44
32 if ((c->x86 == 0xf && c->x86_model >= 0x03) || 45 if ((c->x86 == 0xf && c->x86_model >= 0x03) ||
33 (c->x86 == 0x6 && c->x86_model >= 0x0e)) 46 (c->x86 == 0x6 && c->x86_model >= 0x0e))
34 set_cpu_cap(c, X86_FEATURE_CONSTANT_TSC); 47 set_cpu_cap(c, X86_FEATURE_CONSTANT_TSC);
@@ -278,6 +291,9 @@ static void __cpuinit init_intel(struct cpuinfo_x86 *c)
278 ds_init_intel(c); 291 ds_init_intel(c);
279 } 292 }
280 293
294 if (c->x86 == 6 && c->x86_model == 29 && cpu_has_clflush)
295 set_cpu_cap(c, X86_FEATURE_CLFLUSH_MONITOR);
296
281#ifdef CONFIG_X86_64 297#ifdef CONFIG_X86_64
282 if (c->x86 == 15) 298 if (c->x86 == 15)
283 c->x86_cache_alignment = c->x86_clflush_size * 2; 299 c->x86_cache_alignment = c->x86_clflush_size * 2;
diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c
index 48533d77be78..da299eb85fc0 100644
--- a/arch/x86/kernel/cpu/intel_cacheinfo.c
+++ b/arch/x86/kernel/cpu/intel_cacheinfo.c
@@ -36,8 +36,11 @@ static struct _cache_table cache_table[] __cpuinitdata =
36{ 36{
37 { 0x06, LVL_1_INST, 8 }, /* 4-way set assoc, 32 byte line size */ 37 { 0x06, LVL_1_INST, 8 }, /* 4-way set assoc, 32 byte line size */
38 { 0x08, LVL_1_INST, 16 }, /* 4-way set assoc, 32 byte line size */ 38 { 0x08, LVL_1_INST, 16 }, /* 4-way set assoc, 32 byte line size */
39 { 0x09, LVL_1_INST, 32 }, /* 4-way set assoc, 64 byte line size */
39 { 0x0a, LVL_1_DATA, 8 }, /* 2 way set assoc, 32 byte line size */ 40 { 0x0a, LVL_1_DATA, 8 }, /* 2 way set assoc, 32 byte line size */
40 { 0x0c, LVL_1_DATA, 16 }, /* 4-way set assoc, 32 byte line size */ 41 { 0x0c, LVL_1_DATA, 16 }, /* 4-way set assoc, 32 byte line size */
42 { 0x0d, LVL_1_DATA, 16 }, /* 4-way set assoc, 64 byte line size */
43 { 0x21, LVL_2, 256 }, /* 8-way set assoc, 64 byte line size */
41 { 0x22, LVL_3, 512 }, /* 4-way set assoc, sectored cache, 64 byte line size */ 44 { 0x22, LVL_3, 512 }, /* 4-way set assoc, sectored cache, 64 byte line size */
42 { 0x23, LVL_3, 1024 }, /* 8-way set assoc, sectored cache, 64 byte line size */ 45 { 0x23, LVL_3, 1024 }, /* 8-way set assoc, sectored cache, 64 byte line size */
43 { 0x25, LVL_3, 2048 }, /* 8-way set assoc, sectored cache, 64 byte line size */ 46 { 0x25, LVL_3, 2048 }, /* 8-way set assoc, sectored cache, 64 byte line size */
@@ -85,6 +88,18 @@ static struct _cache_table cache_table[] __cpuinitdata =
85 { 0x85, LVL_2, 2048 }, /* 8-way set assoc, 32 byte line size */ 88 { 0x85, LVL_2, 2048 }, /* 8-way set assoc, 32 byte line size */
86 { 0x86, LVL_2, 512 }, /* 4-way set assoc, 64 byte line size */ 89 { 0x86, LVL_2, 512 }, /* 4-way set assoc, 64 byte line size */
87 { 0x87, LVL_2, 1024 }, /* 8-way set assoc, 64 byte line size */ 90 { 0x87, LVL_2, 1024 }, /* 8-way set assoc, 64 byte line size */
91 { 0xd0, LVL_3, 512 }, /* 4-way set assoc, 64 byte line size */
92 { 0xd1, LVL_3, 1024 }, /* 4-way set assoc, 64 byte line size */
93 { 0xd2, LVL_3, 2048 }, /* 4-way set assoc, 64 byte line size */
94 { 0xd6, LVL_3, 1024 }, /* 8-way set assoc, 64 byte line size */
95 { 0xd7, LVL_3, 2038 }, /* 8-way set assoc, 64 byte line size */
96 { 0xd8, LVL_3, 4096 }, /* 12-way set assoc, 64 byte line size */
97 { 0xdc, LVL_3, 2048 }, /* 12-way set assoc, 64 byte line size */
98 { 0xdd, LVL_3, 4096 }, /* 12-way set assoc, 64 byte line size */
99 { 0xde, LVL_3, 8192 }, /* 12-way set assoc, 64 byte line size */
100 { 0xe2, LVL_3, 2048 }, /* 16-way set assoc, 64 byte line size */
101 { 0xe3, LVL_3, 4096 }, /* 16-way set assoc, 64 byte line size */
102 { 0xe4, LVL_3, 8192 }, /* 16-way set assoc, 64 byte line size */
88 { 0x00, 0, 0} 103 { 0x00, 0, 0}
89}; 104};
90 105
diff --git a/arch/x86/kernel/cpu/mcheck/mce_64.c b/arch/x86/kernel/cpu/mcheck/mce_64.c
index 1c838032fd37..fe79985ce0f2 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_64.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_64.c
@@ -295,11 +295,11 @@ void do_machine_check(struct pt_regs * regs, long error_code)
295 * If we know that the error was in user space, send a 295 * If we know that the error was in user space, send a
296 * SIGBUS. Otherwise, panic if tolerance is low. 296 * SIGBUS. Otherwise, panic if tolerance is low.
297 * 297 *
298 * do_exit() takes an awful lot of locks and has a slight 298 * force_sig() takes an awful lot of locks and has a slight
299 * risk of deadlocking. 299 * risk of deadlocking.
300 */ 300 */
301 if (user_space) { 301 if (user_space) {
302 do_exit(SIGBUS); 302 force_sig(SIGBUS, current);
303 } else if (panic_on_oops || tolerant < 2) { 303 } else if (panic_on_oops || tolerant < 2) {
304 mce_panic("Uncorrected machine check", 304 mce_panic("Uncorrected machine check",
305 &panicm, mcestart); 305 &panicm, mcestart);
@@ -490,7 +490,7 @@ static void __cpuinit mce_cpu_quirks(struct cpuinfo_x86 *c)
490 490
491} 491}
492 492
493static void __cpuinit mce_cpu_features(struct cpuinfo_x86 *c) 493static void mce_cpu_features(struct cpuinfo_x86 *c)
494{ 494{
495 switch (c->x86_vendor) { 495 switch (c->x86_vendor) {
496 case X86_VENDOR_INTEL: 496 case X86_VENDOR_INTEL:
@@ -734,6 +734,7 @@ __setup("mce=", mcheck_enable);
734static int mce_resume(struct sys_device *dev) 734static int mce_resume(struct sys_device *dev)
735{ 735{
736 mce_init(NULL); 736 mce_init(NULL);
737 mce_cpu_features(&current_cpu_data);
737 return 0; 738 return 0;
738} 739}
739 740
diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd_64.c b/arch/x86/kernel/cpu/mcheck/mce_amd_64.c
index 8ae8c4ff094d..f2ee0ae29bd6 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_amd_64.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_amd_64.c
@@ -121,7 +121,7 @@ static long threshold_restart_bank(void *_tr)
121} 121}
122 122
123/* cpu init entry point, called from mce.c with preempt off */ 123/* cpu init entry point, called from mce.c with preempt off */
124void __cpuinit mce_amd_feature_init(struct cpuinfo_x86 *c) 124void mce_amd_feature_init(struct cpuinfo_x86 *c)
125{ 125{
126 unsigned int bank, block; 126 unsigned int bank, block;
127 unsigned int cpu = smp_processor_id(); 127 unsigned int cpu = smp_processor_id();
diff --git a/arch/x86/kernel/cpu/mcheck/mce_intel_64.c b/arch/x86/kernel/cpu/mcheck/mce_intel_64.c
index 4b48f251fd39..f44c36624360 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_intel_64.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_intel_64.c
@@ -30,7 +30,7 @@ asmlinkage void smp_thermal_interrupt(void)
30 irq_exit(); 30 irq_exit();
31} 31}
32 32
33static void __cpuinit intel_init_thermal(struct cpuinfo_x86 *c) 33static void intel_init_thermal(struct cpuinfo_x86 *c)
34{ 34{
35 u32 l, h; 35 u32 l, h;
36 int tm2 = 0; 36 int tm2 = 0;
@@ -84,7 +84,7 @@ static void __cpuinit intel_init_thermal(struct cpuinfo_x86 *c)
84 return; 84 return;
85} 85}
86 86
87void __cpuinit mce_intel_feature_init(struct cpuinfo_x86 *c) 87void mce_intel_feature_init(struct cpuinfo_x86 *c)
88{ 88{
89 intel_init_thermal(c); 89 intel_init_thermal(c);
90} 90}
diff --git a/arch/x86/kernel/cpu/mtrr/generic.c b/arch/x86/kernel/cpu/mtrr/generic.c
index b59ddcc88cd8..0c0a455fe95c 100644
--- a/arch/x86/kernel/cpu/mtrr/generic.c
+++ b/arch/x86/kernel/cpu/mtrr/generic.c
@@ -33,11 +33,13 @@ u64 mtrr_tom2;
33struct mtrr_state_type mtrr_state = {}; 33struct mtrr_state_type mtrr_state = {};
34EXPORT_SYMBOL_GPL(mtrr_state); 34EXPORT_SYMBOL_GPL(mtrr_state);
35 35
36#undef MODULE_PARAM_PREFIX 36static int __initdata mtrr_show;
37#define MODULE_PARAM_PREFIX "mtrr." 37static int __init mtrr_debug(char *opt)
38 38{
39static int mtrr_show; 39 mtrr_show = 1;
40module_param_named(show, mtrr_show, bool, 0); 40 return 0;
41}
42early_param("mtrr.show", mtrr_debug);
41 43
42/* 44/*
43 * Returns the effective MTRR type for the region 45 * Returns the effective MTRR type for the region
diff --git a/arch/x86/kernel/cpu/mtrr/main.c b/arch/x86/kernel/cpu/mtrr/main.c
index d259e5d2e054..236a401b8259 100644
--- a/arch/x86/kernel/cpu/mtrr/main.c
+++ b/arch/x86/kernel/cpu/mtrr/main.c
@@ -1594,8 +1594,7 @@ int __init mtrr_trim_uncached_memory(unsigned long end_pfn)
1594 1594
1595 /* kvm/qemu doesn't have mtrr set right, don't trim them all */ 1595 /* kvm/qemu doesn't have mtrr set right, don't trim them all */
1596 if (!highest_pfn) { 1596 if (!highest_pfn) {
1597 WARN(!kvm_para_available(), KERN_WARNING 1597 printk(KERN_INFO "CPU MTRRs all blank - virtualized system.\n");
1598 "WARNING: strange, CPU MTRRs all blank?\n");
1599 return 0; 1598 return 0;
1600 } 1599 }
1601 1600
diff --git a/arch/x86/kernel/ds.c b/arch/x86/kernel/ds.c
index da91701a2348..169a120587be 100644
--- a/arch/x86/kernel/ds.c
+++ b/arch/x86/kernel/ds.c
@@ -15,8 +15,8 @@
15 * - buffer allocation (memory accounting) 15 * - buffer allocation (memory accounting)
16 * 16 *
17 * 17 *
18 * Copyright (C) 2007-2008 Intel Corporation. 18 * Copyright (C) 2007-2009 Intel Corporation.
19 * Markus Metzger <markus.t.metzger@intel.com>, 2007-2008 19 * Markus Metzger <markus.t.metzger@intel.com>, 2007-2009
20 */ 20 */
21 21
22 22
@@ -890,7 +890,7 @@ int ds_set_pebs_reset(struct pebs_tracer *tracer, u64 value)
890} 890}
891 891
892static const struct ds_configuration ds_cfg_netburst = { 892static const struct ds_configuration ds_cfg_netburst = {
893 .name = "netburst", 893 .name = "Netburst",
894 .ctl[dsf_bts] = (1 << 2) | (1 << 3), 894 .ctl[dsf_bts] = (1 << 2) | (1 << 3),
895 .ctl[dsf_bts_kernel] = (1 << 5), 895 .ctl[dsf_bts_kernel] = (1 << 5),
896 .ctl[dsf_bts_user] = (1 << 6), 896 .ctl[dsf_bts_user] = (1 << 6),
@@ -904,7 +904,7 @@ static const struct ds_configuration ds_cfg_netburst = {
904#endif 904#endif
905}; 905};
906static const struct ds_configuration ds_cfg_pentium_m = { 906static const struct ds_configuration ds_cfg_pentium_m = {
907 .name = "pentium m", 907 .name = "Pentium M",
908 .ctl[dsf_bts] = (1 << 6) | (1 << 7), 908 .ctl[dsf_bts] = (1 << 6) | (1 << 7),
909 909
910 .sizeof_field = sizeof(long), 910 .sizeof_field = sizeof(long),
@@ -915,8 +915,8 @@ static const struct ds_configuration ds_cfg_pentium_m = {
915 .sizeof_rec[ds_pebs] = sizeof(long) * 18, 915 .sizeof_rec[ds_pebs] = sizeof(long) * 18,
916#endif 916#endif
917}; 917};
918static const struct ds_configuration ds_cfg_core2 = { 918static const struct ds_configuration ds_cfg_core2_atom = {
919 .name = "core 2", 919 .name = "Core 2/Atom",
920 .ctl[dsf_bts] = (1 << 6) | (1 << 7), 920 .ctl[dsf_bts] = (1 << 6) | (1 << 7),
921 .ctl[dsf_bts_kernel] = (1 << 9), 921 .ctl[dsf_bts_kernel] = (1 << 9),
922 .ctl[dsf_bts_user] = (1 << 10), 922 .ctl[dsf_bts_user] = (1 << 10),
@@ -949,19 +949,22 @@ void __cpuinit ds_init_intel(struct cpuinfo_x86 *c)
949 switch (c->x86) { 949 switch (c->x86) {
950 case 0x6: 950 case 0x6:
951 switch (c->x86_model) { 951 switch (c->x86_model) {
952 case 0 ... 0xC: 952 case 0x9:
953 /* sorry, don't know about them */ 953 case 0xd: /* Pentium M */
954 break;
955 case 0xD:
956 case 0xE: /* Pentium M */
957 ds_configure(&ds_cfg_pentium_m); 954 ds_configure(&ds_cfg_pentium_m);
958 break; 955 break;
959 default: /* Core2, Atom, ... */ 956 case 0xf:
960 ds_configure(&ds_cfg_core2); 957 case 0x17: /* Core2 */
958 case 0x1c: /* Atom */
959 ds_configure(&ds_cfg_core2_atom);
960 break;
961 case 0x1a: /* i7 */
962 default:
963 /* sorry, don't know about them */
961 break; 964 break;
962 } 965 }
963 break; 966 break;
964 case 0xF: 967 case 0xf:
965 switch (c->x86_model) { 968 switch (c->x86_model) {
966 case 0x0: 969 case 0x0:
967 case 0x1: 970 case 0x1:
diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S
index d6f0490a7391..46469029e9d3 100644
--- a/arch/x86/kernel/entry_32.S
+++ b/arch/x86/kernel/entry_32.S
@@ -1203,7 +1203,6 @@ nmi_stack_correct:
1203 pushl %eax 1203 pushl %eax
1204 CFI_ADJUST_CFA_OFFSET 4 1204 CFI_ADJUST_CFA_OFFSET 4
1205 SAVE_ALL 1205 SAVE_ALL
1206 TRACE_IRQS_OFF
1207 xorl %edx,%edx # zero error code 1206 xorl %edx,%edx # zero error code
1208 movl %esp,%eax # pt_regs pointer 1207 movl %esp,%eax # pt_regs pointer
1209 call do_nmi 1208 call do_nmi
@@ -1244,7 +1243,6 @@ nmi_espfix_stack:
1244 pushl %eax 1243 pushl %eax
1245 CFI_ADJUST_CFA_OFFSET 4 1244 CFI_ADJUST_CFA_OFFSET 4
1246 SAVE_ALL 1245 SAVE_ALL
1247 TRACE_IRQS_OFF
1248 FIXUP_ESPFIX_STACK # %eax == %esp 1246 FIXUP_ESPFIX_STACK # %eax == %esp
1249 xorl %edx,%edx # zero error code 1247 xorl %edx,%edx # zero error code
1250 call do_nmi 1248 call do_nmi
diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S
index e28c7a987793..a1346217e43c 100644
--- a/arch/x86/kernel/entry_64.S
+++ b/arch/x86/kernel/entry_64.S
@@ -346,6 +346,7 @@ ENTRY(save_args)
346 popq_cfi %rax /* move return address... */ 346 popq_cfi %rax /* move return address... */
347 mov %gs:pda_irqstackptr,%rsp 347 mov %gs:pda_irqstackptr,%rsp
348 EMPTY_FRAME 0 348 EMPTY_FRAME 0
349 pushq_cfi %rbp /* backlink for unwinder */
349 pushq_cfi %rax /* ... to the new stack */ 350 pushq_cfi %rax /* ... to the new stack */
350 /* 351 /*
351 * We entered an interrupt context - irqs are off: 352 * We entered an interrupt context - irqs are off:
diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c
index 1b43086b097a..231bdd3c5b1c 100644
--- a/arch/x86/kernel/ftrace.c
+++ b/arch/x86/kernel/ftrace.c
@@ -488,20 +488,21 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr)
488 * ignore such a protection. 488 * ignore such a protection.
489 */ 489 */
490 asm volatile( 490 asm volatile(
491 "1: " _ASM_MOV " (%[parent_old]), %[old]\n" 491 "1: " _ASM_MOV " (%[parent]), %[old]\n"
492 "2: " _ASM_MOV " %[return_hooker], (%[parent_replaced])\n" 492 "2: " _ASM_MOV " %[return_hooker], (%[parent])\n"
493 " movl $0, %[faulted]\n" 493 " movl $0, %[faulted]\n"
494 "3:\n"
494 495
495 ".section .fixup, \"ax\"\n" 496 ".section .fixup, \"ax\"\n"
496 "3: movl $1, %[faulted]\n" 497 "4: movl $1, %[faulted]\n"
498 " jmp 3b\n"
497 ".previous\n" 499 ".previous\n"
498 500
499 _ASM_EXTABLE(1b, 3b) 501 _ASM_EXTABLE(1b, 4b)
500 _ASM_EXTABLE(2b, 3b) 502 _ASM_EXTABLE(2b, 4b)
501 503
502 : [parent_replaced] "=r" (parent), [old] "=r" (old), 504 : [old] "=r" (old), [faulted] "=r" (faulted)
503 [faulted] "=r" (faulted) 505 : [parent] "r" (parent), [return_hooker] "r" (return_hooker)
504 : [parent_old] "0" (parent), [return_hooker] "r" (return_hooker)
505 : "memory" 506 : "memory"
506 ); 507 );
507 508
diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c
index cd759ad90690..a00545fe5cdd 100644
--- a/arch/x86/kernel/hpet.c
+++ b/arch/x86/kernel/hpet.c
@@ -269,6 +269,8 @@ static void hpet_set_mode(enum clock_event_mode mode,
269 now = hpet_readl(HPET_COUNTER); 269 now = hpet_readl(HPET_COUNTER);
270 cmp = now + (unsigned long) delta; 270 cmp = now + (unsigned long) delta;
271 cfg = hpet_readl(HPET_Tn_CFG(timer)); 271 cfg = hpet_readl(HPET_Tn_CFG(timer));
272 /* Make sure we use edge triggered interrupts */
273 cfg &= ~HPET_TN_LEVEL;
272 cfg |= HPET_TN_ENABLE | HPET_TN_PERIODIC | 274 cfg |= HPET_TN_ENABLE | HPET_TN_PERIODIC |
273 HPET_TN_SETVAL | HPET_TN_32BIT; 275 HPET_TN_SETVAL | HPET_TN_32BIT;
274 hpet_writel(cfg, HPET_Tn_CFG(timer)); 276 hpet_writel(cfg, HPET_Tn_CFG(timer));
@@ -628,11 +630,12 @@ static int hpet_cpuhp_notify(struct notifier_block *n,
628 630
629 switch (action & 0xf) { 631 switch (action & 0xf) {
630 case CPU_ONLINE: 632 case CPU_ONLINE:
631 INIT_DELAYED_WORK(&work.work, hpet_work); 633 INIT_DELAYED_WORK_ON_STACK(&work.work, hpet_work);
632 init_completion(&work.complete); 634 init_completion(&work.complete);
633 /* FIXME: add schedule_work_on() */ 635 /* FIXME: add schedule_work_on() */
634 schedule_delayed_work_on(cpu, &work.work, 0); 636 schedule_delayed_work_on(cpu, &work.work, 0);
635 wait_for_completion(&work.complete); 637 wait_for_completion(&work.complete);
638 destroy_timer_on_stack(&work.work.timer);
636 break; 639 break;
637 case CPU_DEAD: 640 case CPU_DEAD:
638 if (hdev) { 641 if (hdev) {
@@ -896,7 +899,7 @@ static unsigned long hpet_rtc_flags;
896static int hpet_prev_update_sec; 899static int hpet_prev_update_sec;
897static struct rtc_time hpet_alarm_time; 900static struct rtc_time hpet_alarm_time;
898static unsigned long hpet_pie_count; 901static unsigned long hpet_pie_count;
899static unsigned long hpet_t1_cmp; 902static u32 hpet_t1_cmp;
900static unsigned long hpet_default_delta; 903static unsigned long hpet_default_delta;
901static unsigned long hpet_pie_delta; 904static unsigned long hpet_pie_delta;
902static unsigned long hpet_pie_limit; 905static unsigned long hpet_pie_limit;
@@ -904,6 +907,14 @@ static unsigned long hpet_pie_limit;
904static rtc_irq_handler irq_handler; 907static rtc_irq_handler irq_handler;
905 908
906/* 909/*
910 * Check that the hpet counter c1 is ahead of the c2
911 */
912static inline int hpet_cnt_ahead(u32 c1, u32 c2)
913{
914 return (s32)(c2 - c1) < 0;
915}
916
917/*
907 * Registers a IRQ handler. 918 * Registers a IRQ handler.
908 */ 919 */
909int hpet_register_irq_handler(rtc_irq_handler handler) 920int hpet_register_irq_handler(rtc_irq_handler handler)
@@ -1074,7 +1085,7 @@ static void hpet_rtc_timer_reinit(void)
1074 hpet_t1_cmp += delta; 1085 hpet_t1_cmp += delta;
1075 hpet_writel(hpet_t1_cmp, HPET_T1_CMP); 1086 hpet_writel(hpet_t1_cmp, HPET_T1_CMP);
1076 lost_ints++; 1087 lost_ints++;
1077 } while ((long)(hpet_readl(HPET_COUNTER) - hpet_t1_cmp) > 0); 1088 } while (!hpet_cnt_ahead(hpet_t1_cmp, hpet_readl(HPET_COUNTER)));
1078 1089
1079 if (lost_ints) { 1090 if (lost_ints) {
1080 if (hpet_rtc_flags & RTC_PIE) 1091 if (hpet_rtc_flags & RTC_PIE)
diff --git a/arch/x86/kernel/i8237.c b/arch/x86/kernel/i8237.c
index dbd6c1d1b638..b42ca694dc68 100644
--- a/arch/x86/kernel/i8237.c
+++ b/arch/x86/kernel/i8237.c
@@ -28,10 +28,10 @@ static int i8237A_resume(struct sys_device *dev)
28 28
29 flags = claim_dma_lock(); 29 flags = claim_dma_lock();
30 30
31 dma_outb(DMA1_RESET_REG, 0); 31 dma_outb(0, DMA1_RESET_REG);
32 dma_outb(DMA2_RESET_REG, 0); 32 dma_outb(0, DMA2_RESET_REG);
33 33
34 for (i = 0;i < 8;i++) { 34 for (i = 0; i < 8; i++) {
35 set_dma_addr(i, 0x000000); 35 set_dma_addr(i, 0x000000);
36 /* DMA count is a bit weird so this is not 0 */ 36 /* DMA count is a bit weird so this is not 0 */
37 set_dma_count(i, 1); 37 set_dma_count(i, 1);
@@ -51,14 +51,14 @@ static int i8237A_suspend(struct sys_device *dev, pm_message_t state)
51} 51}
52 52
53static struct sysdev_class i8237_sysdev_class = { 53static struct sysdev_class i8237_sysdev_class = {
54 .name = "i8237", 54 .name = "i8237",
55 .suspend = i8237A_suspend, 55 .suspend = i8237A_suspend,
56 .resume = i8237A_resume, 56 .resume = i8237A_resume,
57}; 57};
58 58
59static struct sys_device device_i8237A = { 59static struct sys_device device_i8237A = {
60 .id = 0, 60 .id = 0,
61 .cls = &i8237_sysdev_class, 61 .cls = &i8237_sysdev_class,
62}; 62};
63 63
64static int __init i8237A_init_sysfs(void) 64static int __init i8237A_init_sysfs(void)
@@ -68,5 +68,4 @@ static int __init i8237A_init_sysfs(void)
68 error = sysdev_register(&device_i8237A); 68 error = sysdev_register(&device_i8237A);
69 return error; 69 return error;
70} 70}
71
72device_initcall(i8237A_init_sysfs); 71device_initcall(i8237A_init_sysfs);
diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c
index 1c4a1302536c..bc7ac4da90d7 100644
--- a/arch/x86/kernel/io_apic.c
+++ b/arch/x86/kernel/io_apic.c
@@ -2528,14 +2528,15 @@ static void irq_complete_move(struct irq_desc **descp)
2528 2528
2529 vector = ~get_irq_regs()->orig_ax; 2529 vector = ~get_irq_regs()->orig_ax;
2530 me = smp_processor_id(); 2530 me = smp_processor_id();
2531
2532 if (vector == cfg->vector && cpumask_test_cpu(me, cfg->domain)) {
2531#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC 2533#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC
2532 *descp = desc = move_irq_desc(desc, me); 2534 *descp = desc = move_irq_desc(desc, me);
2533 /* get the new one */ 2535 /* get the new one */
2534 cfg = desc->chip_data; 2536 cfg = desc->chip_data;
2535#endif 2537#endif
2536
2537 if (vector == cfg->vector && cpumask_test_cpu(me, cfg->domain))
2538 send_cleanup_vector(cfg); 2538 send_cleanup_vector(cfg);
2539 }
2539} 2540}
2540#else 2541#else
2541static inline void irq_complete_move(struct irq_desc **descp) {} 2542static inline void irq_complete_move(struct irq_desc **descp) {}
@@ -3840,14 +3841,24 @@ int __init io_apic_get_redir_entries (int ioapic)
3840 3841
3841void __init probe_nr_irqs_gsi(void) 3842void __init probe_nr_irqs_gsi(void)
3842{ 3843{
3843 int idx;
3844 int nr = 0; 3844 int nr = 0;
3845 3845
3846 for (idx = 0; idx < nr_ioapics; idx++) 3846 nr = acpi_probe_gsi();
3847 nr += io_apic_get_redir_entries(idx) + 1; 3847 if (nr > nr_irqs_gsi) {
3848
3849 if (nr > nr_irqs_gsi)
3850 nr_irqs_gsi = nr; 3848 nr_irqs_gsi = nr;
3849 } else {
3850 /* for acpi=off or acpi is not compiled in */
3851 int idx;
3852
3853 nr = 0;
3854 for (idx = 0; idx < nr_ioapics; idx++)
3855 nr += io_apic_get_redir_entries(idx) + 1;
3856
3857 if (nr > nr_irqs_gsi)
3858 nr_irqs_gsi = nr;
3859 }
3860
3861 printk(KERN_DEBUG "nr_irqs_gsi: %d\n", nr_irqs_gsi);
3851} 3862}
3852 3863
3853/* -------------------------------------------------------------------------- 3864/* --------------------------------------------------------------------------
diff --git a/arch/x86/kernel/irqinit_32.c b/arch/x86/kernel/irqinit_32.c
index 1507ad4e674d..10a09c2f1828 100644
--- a/arch/x86/kernel/irqinit_32.c
+++ b/arch/x86/kernel/irqinit_32.c
@@ -78,15 +78,6 @@ void __init init_ISA_irqs(void)
78 } 78 }
79} 79}
80 80
81/*
82 * IRQ2 is cascade interrupt to second interrupt controller
83 */
84static struct irqaction irq2 = {
85 .handler = no_action,
86 .mask = CPU_MASK_NONE,
87 .name = "cascade",
88};
89
90DEFINE_PER_CPU(vector_irq_t, vector_irq) = { 81DEFINE_PER_CPU(vector_irq_t, vector_irq) = {
91 [0 ... IRQ0_VECTOR - 1] = -1, 82 [0 ... IRQ0_VECTOR - 1] = -1,
92 [IRQ0_VECTOR] = 0, 83 [IRQ0_VECTOR] = 0,
@@ -178,9 +169,6 @@ void __init native_init_IRQ(void)
178 alloc_intr_gate(THERMAL_APIC_VECTOR, thermal_interrupt); 169 alloc_intr_gate(THERMAL_APIC_VECTOR, thermal_interrupt);
179#endif 170#endif
180 171
181 if (!acpi_ioapic)
182 setup_irq(2, &irq2);
183
184 /* setup after call gates are initialised (usually add in 172 /* setup after call gates are initialised (usually add in
185 * the architecture specific gates) 173 * the architecture specific gates)
186 */ 174 */
diff --git a/arch/x86/kernel/kprobes.c b/arch/x86/kernel/kprobes.c
index 884d985b8b82..e948b28a5a9a 100644
--- a/arch/x86/kernel/kprobes.c
+++ b/arch/x86/kernel/kprobes.c
@@ -446,7 +446,7 @@ void __kprobes arch_prepare_kretprobe(struct kretprobe_instance *ri,
446static void __kprobes setup_singlestep(struct kprobe *p, struct pt_regs *regs, 446static void __kprobes setup_singlestep(struct kprobe *p, struct pt_regs *regs,
447 struct kprobe_ctlblk *kcb) 447 struct kprobe_ctlblk *kcb)
448{ 448{
449#if !defined(CONFIG_PREEMPT) || defined(CONFIG_PM) 449#if !defined(CONFIG_PREEMPT) || defined(CONFIG_FREEZER)
450 if (p->ainsn.boostable == 1 && !p->post_handler) { 450 if (p->ainsn.boostable == 1 && !p->post_handler) {
451 /* Boost up -- we can execute copied instructions directly */ 451 /* Boost up -- we can execute copied instructions directly */
452 reset_current_kprobe(); 452 reset_current_kprobe();
diff --git a/arch/x86/kernel/mpparse.c b/arch/x86/kernel/mpparse.c
index c0601c2848a1..a649a4ccad43 100644
--- a/arch/x86/kernel/mpparse.c
+++ b/arch/x86/kernel/mpparse.c
@@ -27,6 +27,7 @@
27#include <asm/e820.h> 27#include <asm/e820.h>
28#include <asm/trampoline.h> 28#include <asm/trampoline.h>
29#include <asm/setup.h> 29#include <asm/setup.h>
30#include <asm/smp.h>
30 31
31#include <mach_apic.h> 32#include <mach_apic.h>
32#ifdef CONFIG_X86_32 33#ifdef CONFIG_X86_32
diff --git a/arch/x86/kernel/olpc.c b/arch/x86/kernel/olpc.c
index 7a13fac63a1f..4006c522adc7 100644
--- a/arch/x86/kernel/olpc.c
+++ b/arch/x86/kernel/olpc.c
@@ -203,7 +203,7 @@ static void __init platform_detect(void)
203static void __init platform_detect(void) 203static void __init platform_detect(void)
204{ 204{
205 /* stopgap until OFW support is added to the kernel */ 205 /* stopgap until OFW support is added to the kernel */
206 olpc_platform_info.boardrev = 0xc2; 206 olpc_platform_info.boardrev = olpc_board(0xc2);
207} 207}
208#endif 208#endif
209 209
diff --git a/arch/x86/kernel/paravirt.c b/arch/x86/kernel/paravirt.c
index e4c8fb608873..c6520a4e85d4 100644
--- a/arch/x86/kernel/paravirt.c
+++ b/arch/x86/kernel/paravirt.c
@@ -268,6 +268,32 @@ enum paravirt_lazy_mode paravirt_get_lazy_mode(void)
268 return __get_cpu_var(paravirt_lazy_mode); 268 return __get_cpu_var(paravirt_lazy_mode);
269} 269}
270 270
271void arch_flush_lazy_mmu_mode(void)
272{
273 preempt_disable();
274
275 if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_MMU) {
276 WARN_ON(preempt_count() == 1);
277 arch_leave_lazy_mmu_mode();
278 arch_enter_lazy_mmu_mode();
279 }
280
281 preempt_enable();
282}
283
284void arch_flush_lazy_cpu_mode(void)
285{
286 preempt_disable();
287
288 if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_CPU) {
289 WARN_ON(preempt_count() == 1);
290 arch_leave_lazy_cpu_mode();
291 arch_enter_lazy_cpu_mode();
292 }
293
294 preempt_enable();
295}
296
271struct pv_info pv_info = { 297struct pv_info pv_info = {
272 .name = "bare hardware", 298 .name = "bare hardware",
273 .paravirt_enabled = 0, 299 .paravirt_enabled = 0,
diff --git a/arch/x86/kernel/pci-gart_64.c b/arch/x86/kernel/pci-gart_64.c
index 00c2bcd41463..d5768b1af080 100644
--- a/arch/x86/kernel/pci-gart_64.c
+++ b/arch/x86/kernel/pci-gart_64.c
@@ -5,7 +5,7 @@
5 * This allows to use PCI devices that only support 32bit addresses on systems 5 * This allows to use PCI devices that only support 32bit addresses on systems
6 * with more than 4GB. 6 * with more than 4GB.
7 * 7 *
8 * See Documentation/DMA-mapping.txt for the interface specification. 8 * See Documentation/PCI/PCI-DMA-mapping.txt for the interface specification.
9 * 9 *
10 * Copyright 2002 Andi Kleen, SuSE Labs. 10 * Copyright 2002 Andi Kleen, SuSE Labs.
11 * Subject to the GNU General Public License v2 only. 11 * Subject to the GNU General Public License v2 only.
diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c
index e68bb9e30864..6d12f7e37f8c 100644
--- a/arch/x86/kernel/process.c
+++ b/arch/x86/kernel/process.c
@@ -180,6 +180,9 @@ void mwait_idle_with_hints(unsigned long ax, unsigned long cx)
180 180
181 trace_power_start(&it, POWER_CSTATE, (ax>>4)+1); 181 trace_power_start(&it, POWER_CSTATE, (ax>>4)+1);
182 if (!need_resched()) { 182 if (!need_resched()) {
183 if (cpu_has(&current_cpu_data, X86_FEATURE_CLFLUSH_MONITOR))
184 clflush((void *)&current_thread_info()->flags);
185
183 __monitor((void *)&current_thread_info()->flags, 0, 0); 186 __monitor((void *)&current_thread_info()->flags, 0, 0);
184 smp_mb(); 187 smp_mb();
185 if (!need_resched()) 188 if (!need_resched())
@@ -194,6 +197,9 @@ static void mwait_idle(void)
194 struct power_trace it; 197 struct power_trace it;
195 if (!need_resched()) { 198 if (!need_resched()) {
196 trace_power_start(&it, POWER_CSTATE, 1); 199 trace_power_start(&it, POWER_CSTATE, 1);
200 if (cpu_has(&current_cpu_data, X86_FEATURE_CLFLUSH_MONITOR))
201 clflush((void *)&current_thread_info()->flags);
202
197 __monitor((void *)&current_thread_info()->flags, 0, 0); 203 __monitor((void *)&current_thread_info()->flags, 0, 0);
198 smp_mb(); 204 smp_mb();
199 if (!need_resched()) 205 if (!need_resched())
diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c
index a546f55c77b4..bd4da2af08ae 100644
--- a/arch/x86/kernel/process_32.c
+++ b/arch/x86/kernel/process_32.c
@@ -104,9 +104,6 @@ void cpu_idle(void)
104 check_pgt_cache(); 104 check_pgt_cache();
105 rmb(); 105 rmb();
106 106
107 if (rcu_pending(cpu))
108 rcu_check_callbacks(cpu, 0);
109
110 if (cpu_is_offline(cpu)) 107 if (cpu_is_offline(cpu))
111 play_dead(); 108 play_dead();
112 109
diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c
index 416fb9282f4f..85b4cb5c1980 100644
--- a/arch/x86/kernel/process_64.c
+++ b/arch/x86/kernel/process_64.c
@@ -40,6 +40,7 @@
40#include <linux/uaccess.h> 40#include <linux/uaccess.h>
41#include <linux/io.h> 41#include <linux/io.h>
42#include <linux/ftrace.h> 42#include <linux/ftrace.h>
43#include <linux/dmi.h>
43 44
44#include <asm/pgtable.h> 45#include <asm/pgtable.h>
45#include <asm/system.h> 46#include <asm/system.h>
@@ -151,14 +152,18 @@ void __show_regs(struct pt_regs *regs, int all)
151 unsigned long d0, d1, d2, d3, d6, d7; 152 unsigned long d0, d1, d2, d3, d6, d7;
152 unsigned int fsindex, gsindex; 153 unsigned int fsindex, gsindex;
153 unsigned int ds, cs, es; 154 unsigned int ds, cs, es;
155 const char *board;
154 156
155 printk("\n"); 157 printk("\n");
156 print_modules(); 158 print_modules();
157 printk(KERN_INFO "Pid: %d, comm: %.20s %s %s %.*s\n", 159 board = dmi_get_system_info(DMI_PRODUCT_NAME);
160 if (!board)
161 board = "";
162 printk(KERN_INFO "Pid: %d, comm: %.20s %s %s %.*s %s\n",
158 current->pid, current->comm, print_tainted(), 163 current->pid, current->comm, print_tainted(),
159 init_utsname()->release, 164 init_utsname()->release,
160 (int)strcspn(init_utsname()->version, " "), 165 (int)strcspn(init_utsname()->version, " "),
161 init_utsname()->version); 166 init_utsname()->version, board);
162 printk(KERN_INFO "RIP: %04lx:[<%016lx>] ", regs->cs & 0xffff, regs->ip); 167 printk(KERN_INFO "RIP: %04lx:[<%016lx>] ", regs->cs & 0xffff, regs->ip);
163 printk_address(regs->ip, 1); 168 printk_address(regs->ip, 1);
164 printk(KERN_INFO "RSP: %04lx:%016lx EFLAGS: %08lx\n", regs->ss, 169 printk(KERN_INFO "RSP: %04lx:%016lx EFLAGS: %08lx\n", regs->ss,
diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c
index 0a5df5f82fb9..5a4c23d89892 100644
--- a/arch/x86/kernel/ptrace.c
+++ b/arch/x86/kernel/ptrace.c
@@ -810,12 +810,16 @@ static void ptrace_bts_untrace(struct task_struct *child)
810 810
811static void ptrace_bts_detach(struct task_struct *child) 811static void ptrace_bts_detach(struct task_struct *child)
812{ 812{
813 if (unlikely(child->bts)) { 813 /*
814 ds_release_bts(child->bts); 814 * Ptrace_detach() races with ptrace_untrace() in case
815 child->bts = NULL; 815 * the child dies and is reaped by another thread.
816 816 *
817 ptrace_bts_free_buffer(child); 817 * We only do the memory accounting at this point and
818 } 818 * leave the buffer deallocation and the bts tracer
819 * release to ptrace_bts_untrace() which will be called
820 * later on with tasklist_lock held.
821 */
822 release_locked_buffer(child->bts_buffer, child->bts_size);
819} 823}
820#else 824#else
821static inline void ptrace_bts_fork(struct task_struct *tsk) {} 825static inline void ptrace_bts_fork(struct task_struct *tsk) {}
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index ae0d8042cf69..c461f6d69074 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -607,7 +607,7 @@ struct x86_quirks *x86_quirks __initdata = &default_x86_quirks;
607static int __init dmi_low_memory_corruption(const struct dmi_system_id *d) 607static int __init dmi_low_memory_corruption(const struct dmi_system_id *d)
608{ 608{
609 printk(KERN_NOTICE 609 printk(KERN_NOTICE
610 "%s detected: BIOS may corrupt low RAM, working it around.\n", 610 "%s detected: BIOS may corrupt low RAM, working around it.\n",
611 d->ident); 611 d->ident);
612 612
613 e820_update_range(0, 0x10000, E820_RAM, E820_RESERVED); 613 e820_update_range(0, 0x10000, E820_RAM, E820_RESERVED);
diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c
index 55c46074eba0..01161077a49c 100644
--- a/arch/x86/kernel/setup_percpu.c
+++ b/arch/x86/kernel/setup_percpu.c
@@ -136,7 +136,7 @@ static void __init setup_cpu_pda_map(void)
136#ifdef CONFIG_X86_64 136#ifdef CONFIG_X86_64
137 137
138/* correctly size the local cpu masks */ 138/* correctly size the local cpu masks */
139static void setup_cpu_local_masks(void) 139static void __init setup_cpu_local_masks(void)
140{ 140{
141 alloc_bootmem_cpumask_var(&cpu_initialized_mask); 141 alloc_bootmem_cpumask_var(&cpu_initialized_mask);
142 alloc_bootmem_cpumask_var(&cpu_callin_mask); 142 alloc_bootmem_cpumask_var(&cpu_callin_mask);
diff --git a/arch/x86/kernel/signal.c b/arch/x86/kernel/signal.c
index 89bb7668041d..df0587f24c54 100644
--- a/arch/x86/kernel/signal.c
+++ b/arch/x86/kernel/signal.c
@@ -632,9 +632,16 @@ badframe:
632} 632}
633 633
634#ifdef CONFIG_X86_32 634#ifdef CONFIG_X86_32
635asmlinkage int sys_rt_sigreturn(struct pt_regs regs) 635/*
636 * Note: do not pass in pt_regs directly as with tail-call optimization
637 * GCC will incorrectly stomp on the caller's frame and corrupt user-space
638 * register state:
639 */
640asmlinkage int sys_rt_sigreturn(unsigned long __unused)
636{ 641{
637 return do_rt_sigreturn(&regs); 642 struct pt_regs *regs = (struct pt_regs *)&__unused;
643
644 return do_rt_sigreturn(regs);
638} 645}
639#else /* !CONFIG_X86_32 */ 646#else /* !CONFIG_X86_32 */
640asmlinkage long sys_rt_sigreturn(struct pt_regs *regs) 647asmlinkage long sys_rt_sigreturn(struct pt_regs *regs)
diff --git a/arch/x86/kernel/syscall_table_32.S b/arch/x86/kernel/syscall_table_32.S
index d44395ff34c3..e2e86a08f31d 100644
--- a/arch/x86/kernel/syscall_table_32.S
+++ b/arch/x86/kernel/syscall_table_32.S
@@ -88,7 +88,7 @@ ENTRY(sys_call_table)
88 .long sys_uselib 88 .long sys_uselib
89 .long sys_swapon 89 .long sys_swapon
90 .long sys_reboot 90 .long sys_reboot
91 .long old_readdir 91 .long sys_old_readdir
92 .long old_mmap /* 90 */ 92 .long old_mmap /* 90 */
93 .long sys_munmap 93 .long sys_munmap
94 .long sys_truncate 94 .long sys_truncate
diff --git a/arch/x86/kernel/tlb_uv.c b/arch/x86/kernel/tlb_uv.c
index f885023167e0..6812b829ed83 100644
--- a/arch/x86/kernel/tlb_uv.c
+++ b/arch/x86/kernel/tlb_uv.c
@@ -200,6 +200,7 @@ static int uv_wait_completion(struct bau_desc *bau_desc,
200 destination_timeouts = 0; 200 destination_timeouts = 0;
201 } 201 }
202 } 202 }
203 cpu_relax();
203 } 204 }
204 return FLUSH_COMPLETE; 205 return FLUSH_COMPLETE;
205} 206}
diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
index 98c2d055284b..a9e7548e1790 100644
--- a/arch/x86/kernel/traps.c
+++ b/arch/x86/kernel/traps.c
@@ -99,6 +99,12 @@ static inline void preempt_conditional_sti(struct pt_regs *regs)
99 local_irq_enable(); 99 local_irq_enable();
100} 100}
101 101
102static inline void conditional_cli(struct pt_regs *regs)
103{
104 if (regs->flags & X86_EFLAGS_IF)
105 local_irq_disable();
106}
107
102static inline void preempt_conditional_cli(struct pt_regs *regs) 108static inline void preempt_conditional_cli(struct pt_regs *regs)
103{ 109{
104 if (regs->flags & X86_EFLAGS_IF) 110 if (regs->flags & X86_EFLAGS_IF)
@@ -626,8 +632,10 @@ clear_dr7:
626 632
627#ifdef CONFIG_X86_32 633#ifdef CONFIG_X86_32
628debug_vm86: 634debug_vm86:
635 /* reenable preemption: handle_vm86_trap() might sleep */
636 dec_preempt_count();
629 handle_vm86_trap((struct kernel_vm86_regs *) regs, error_code, 1); 637 handle_vm86_trap((struct kernel_vm86_regs *) regs, error_code, 1);
630 preempt_conditional_cli(regs); 638 conditional_cli(regs);
631 return; 639 return;
632#endif 640#endif
633 641
@@ -896,7 +904,7 @@ asmlinkage void math_state_restore(void)
896EXPORT_SYMBOL_GPL(math_state_restore); 904EXPORT_SYMBOL_GPL(math_state_restore);
897 905
898#ifndef CONFIG_MATH_EMULATION 906#ifndef CONFIG_MATH_EMULATION
899asmlinkage void math_emulate(long arg) 907void math_emulate(struct math_emu_info *info)
900{ 908{
901 printk(KERN_EMERG 909 printk(KERN_EMERG
902 "math-emulation not enabled and no coprocessor found.\n"); 910 "math-emulation not enabled and no coprocessor found.\n");
@@ -906,16 +914,19 @@ asmlinkage void math_emulate(long arg)
906} 914}
907#endif /* CONFIG_MATH_EMULATION */ 915#endif /* CONFIG_MATH_EMULATION */
908 916
909dotraplinkage void __kprobes 917dotraplinkage void __kprobes do_device_not_available(struct pt_regs regs)
910do_device_not_available(struct pt_regs *regs, long error)
911{ 918{
912#ifdef CONFIG_X86_32 919#ifdef CONFIG_X86_32
913 if (read_cr0() & X86_CR0_EM) { 920 if (read_cr0() & X86_CR0_EM) {
914 conditional_sti(regs); 921 struct math_emu_info info = { };
915 math_emulate(0); 922
923 conditional_sti(&regs);
924
925 info.regs = &regs;
926 math_emulate(&info);
916 } else { 927 } else {
917 math_state_restore(); /* interrupts still off */ 928 math_state_restore(); /* interrupts still off */
918 conditional_sti(regs); 929 conditional_sti(&regs);
919 } 930 }
920#else 931#else
921 math_state_restore(); 932 math_state_restore();
diff --git a/arch/x86/kernel/vmi_32.c b/arch/x86/kernel/vmi_32.c
index 23206ba16874..bef58b4982db 100644
--- a/arch/x86/kernel/vmi_32.c
+++ b/arch/x86/kernel/vmi_32.c
@@ -321,6 +321,16 @@ static void vmi_release_pmd(unsigned long pfn)
321} 321}
322 322
323/* 323/*
324 * We use the pgd_free hook for releasing the pgd page:
325 */
326static void vmi_pgd_free(struct mm_struct *mm, pgd_t *pgd)
327{
328 unsigned long pfn = __pa(pgd) >> PAGE_SHIFT;
329
330 vmi_ops.release_page(pfn, VMI_PAGE_L2);
331}
332
333/*
324 * Helper macros for MMU update flags. We can defer updates until a flush 334 * Helper macros for MMU update flags. We can defer updates until a flush
325 * or page invalidation only if the update is to the current address space 335 * or page invalidation only if the update is to the current address space
326 * (otherwise, there is no flush). We must check against init_mm, since 336 * (otherwise, there is no flush). We must check against init_mm, since
@@ -762,6 +772,7 @@ static inline int __init activate_vmi(void)
762 if (vmi_ops.release_page) { 772 if (vmi_ops.release_page) {
763 pv_mmu_ops.release_pte = vmi_release_pte; 773 pv_mmu_ops.release_pte = vmi_release_pte;
764 pv_mmu_ops.release_pmd = vmi_release_pmd; 774 pv_mmu_ops.release_pmd = vmi_release_pmd;
775 pv_mmu_ops.pgd_free = vmi_pgd_free;
765 } 776 }
766 777
767 /* Set linear is needed in all cases */ 778 /* Set linear is needed in all cases */
@@ -858,7 +869,7 @@ void __init vmi_init(void)
858#endif 869#endif
859} 870}
860 871
861void vmi_activate(void) 872void __init vmi_activate(void)
862{ 873{
863 unsigned long flags; 874 unsigned long flags;
864 875
diff --git a/arch/x86/kernel/vmiclock_32.c b/arch/x86/kernel/vmiclock_32.c
index c4c1f9e09402..bde106cae0a9 100644
--- a/arch/x86/kernel/vmiclock_32.c
+++ b/arch/x86/kernel/vmiclock_32.c
@@ -283,10 +283,13 @@ void __devinit vmi_time_ap_init(void)
283#endif 283#endif
284 284
285/** vmi clocksource */ 285/** vmi clocksource */
286static struct clocksource clocksource_vmi;
286 287
287static cycle_t read_real_cycles(void) 288static cycle_t read_real_cycles(void)
288{ 289{
289 return vmi_timer_ops.get_cycle_counter(VMI_CYCLES_REAL); 290 cycle_t ret = (cycle_t)vmi_timer_ops.get_cycle_counter(VMI_CYCLES_REAL);
291 return ret >= clocksource_vmi.cycle_last ?
292 ret : clocksource_vmi.cycle_last;
290} 293}
291 294
292static struct clocksource clocksource_vmi = { 295static struct clocksource clocksource_vmi = {
diff --git a/arch/x86/kvm/i8254.c b/arch/x86/kvm/i8254.c
index e665d1c623ca..72bd275a9b5c 100644
--- a/arch/x86/kvm/i8254.c
+++ b/arch/x86/kvm/i8254.c
@@ -207,7 +207,7 @@ static int __pit_timer_fn(struct kvm_kpit_state *ps)
207 hrtimer_add_expires_ns(&pt->timer, pt->period); 207 hrtimer_add_expires_ns(&pt->timer, pt->period);
208 pt->scheduled = hrtimer_get_expires_ns(&pt->timer); 208 pt->scheduled = hrtimer_get_expires_ns(&pt->timer);
209 if (pt->period) 209 if (pt->period)
210 ps->channels[0].count_load_time = hrtimer_get_expires(&pt->timer); 210 ps->channels[0].count_load_time = ktime_get();
211 211
212 return (pt->period == 0 ? 0 : 1); 212 return (pt->period == 0 ? 0 : 1);
213} 213}
diff --git a/arch/x86/kvm/irq.c b/arch/x86/kvm/irq.c
index c019b8edcdb7..cf17ed52f6fb 100644
--- a/arch/x86/kvm/irq.c
+++ b/arch/x86/kvm/irq.c
@@ -87,13 +87,6 @@ void kvm_inject_pending_timer_irqs(struct kvm_vcpu *vcpu)
87} 87}
88EXPORT_SYMBOL_GPL(kvm_inject_pending_timer_irqs); 88EXPORT_SYMBOL_GPL(kvm_inject_pending_timer_irqs);
89 89
90void kvm_timer_intr_post(struct kvm_vcpu *vcpu, int vec)
91{
92 kvm_apic_timer_intr_post(vcpu, vec);
93 /* TODO: PIT, RTC etc. */
94}
95EXPORT_SYMBOL_GPL(kvm_timer_intr_post);
96
97void __kvm_migrate_timers(struct kvm_vcpu *vcpu) 90void __kvm_migrate_timers(struct kvm_vcpu *vcpu)
98{ 91{
99 __kvm_migrate_apic_timer(vcpu); 92 __kvm_migrate_apic_timer(vcpu);
diff --git a/arch/x86/kvm/irq.h b/arch/x86/kvm/irq.h
index 2bf32a03ceec..82579ee538d0 100644
--- a/arch/x86/kvm/irq.h
+++ b/arch/x86/kvm/irq.h
@@ -89,7 +89,6 @@ static inline int irqchip_in_kernel(struct kvm *kvm)
89 89
90void kvm_pic_reset(struct kvm_kpic_state *s); 90void kvm_pic_reset(struct kvm_kpic_state *s);
91 91
92void kvm_timer_intr_post(struct kvm_vcpu *vcpu, int vec);
93void kvm_inject_pending_timer_irqs(struct kvm_vcpu *vcpu); 92void kvm_inject_pending_timer_irqs(struct kvm_vcpu *vcpu);
94void kvm_inject_apic_timer_irqs(struct kvm_vcpu *vcpu); 93void kvm_inject_apic_timer_irqs(struct kvm_vcpu *vcpu);
95void kvm_apic_nmi_wd_deliver(struct kvm_vcpu *vcpu); 94void kvm_apic_nmi_wd_deliver(struct kvm_vcpu *vcpu);
diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c
index afac68c0815c..f0b67f2cdd69 100644
--- a/arch/x86/kvm/lapic.c
+++ b/arch/x86/kvm/lapic.c
@@ -35,6 +35,12 @@
35#include "kvm_cache_regs.h" 35#include "kvm_cache_regs.h"
36#include "irq.h" 36#include "irq.h"
37 37
38#ifndef CONFIG_X86_64
39#define mod_64(x, y) ((x) - (y) * div64_u64(x, y))
40#else
41#define mod_64(x, y) ((x) % (y))
42#endif
43
38#define PRId64 "d" 44#define PRId64 "d"
39#define PRIx64 "llx" 45#define PRIx64 "llx"
40#define PRIu64 "u" 46#define PRIu64 "u"
@@ -511,52 +517,22 @@ static void apic_send_ipi(struct kvm_lapic *apic)
511 517
512static u32 apic_get_tmcct(struct kvm_lapic *apic) 518static u32 apic_get_tmcct(struct kvm_lapic *apic)
513{ 519{
514 u64 counter_passed; 520 ktime_t remaining;
515 ktime_t passed, now; 521 s64 ns;
516 u32 tmcct; 522 u32 tmcct;
517 523
518 ASSERT(apic != NULL); 524 ASSERT(apic != NULL);
519 525
520 now = apic->timer.dev.base->get_time();
521 tmcct = apic_get_reg(apic, APIC_TMICT);
522
523 /* if initial count is 0, current count should also be 0 */ 526 /* if initial count is 0, current count should also be 0 */
524 if (tmcct == 0) 527 if (apic_get_reg(apic, APIC_TMICT) == 0)
525 return 0; 528 return 0;
526 529
527 if (unlikely(ktime_to_ns(now) <= 530 remaining = hrtimer_expires_remaining(&apic->timer.dev);
528 ktime_to_ns(apic->timer.last_update))) { 531 if (ktime_to_ns(remaining) < 0)
529 /* Wrap around */ 532 remaining = ktime_set(0, 0);
530 passed = ktime_add(( { 533
531 (ktime_t) { 534 ns = mod_64(ktime_to_ns(remaining), apic->timer.period);
532 .tv64 = KTIME_MAX - 535 tmcct = div64_u64(ns, (APIC_BUS_CYCLE_NS * apic->timer.divide_count));
533 (apic->timer.last_update).tv64}; }
534 ), now);
535 apic_debug("time elapsed\n");
536 } else
537 passed = ktime_sub(now, apic->timer.last_update);
538
539 counter_passed = div64_u64(ktime_to_ns(passed),
540 (APIC_BUS_CYCLE_NS * apic->timer.divide_count));
541
542 if (counter_passed > tmcct) {
543 if (unlikely(!apic_lvtt_period(apic))) {
544 /* one-shot timers stick at 0 until reset */
545 tmcct = 0;
546 } else {
547 /*
548 * periodic timers reset to APIC_TMICT when they
549 * hit 0. The while loop simulates this happening N
550 * times. (counter_passed %= tmcct) would also work,
551 * but might be slower or not work on 32-bit??
552 */
553 while (counter_passed > tmcct)
554 counter_passed -= tmcct;
555 tmcct -= counter_passed;
556 }
557 } else {
558 tmcct -= counter_passed;
559 }
560 536
561 return tmcct; 537 return tmcct;
562} 538}
@@ -653,8 +629,6 @@ static void start_apic_timer(struct kvm_lapic *apic)
653{ 629{
654 ktime_t now = apic->timer.dev.base->get_time(); 630 ktime_t now = apic->timer.dev.base->get_time();
655 631
656 apic->timer.last_update = now;
657
658 apic->timer.period = apic_get_reg(apic, APIC_TMICT) * 632 apic->timer.period = apic_get_reg(apic, APIC_TMICT) *
659 APIC_BUS_CYCLE_NS * apic->timer.divide_count; 633 APIC_BUS_CYCLE_NS * apic->timer.divide_count;
660 atomic_set(&apic->timer.pending, 0); 634 atomic_set(&apic->timer.pending, 0);
@@ -1110,16 +1084,6 @@ void kvm_inject_apic_timer_irqs(struct kvm_vcpu *vcpu)
1110 } 1084 }
1111} 1085}
1112 1086
1113void kvm_apic_timer_intr_post(struct kvm_vcpu *vcpu, int vec)
1114{
1115 struct kvm_lapic *apic = vcpu->arch.apic;
1116
1117 if (apic && apic_lvt_vector(apic, APIC_LVTT) == vec)
1118 apic->timer.last_update = ktime_add_ns(
1119 apic->timer.last_update,
1120 apic->timer.period);
1121}
1122
1123int kvm_get_apic_interrupt(struct kvm_vcpu *vcpu) 1087int kvm_get_apic_interrupt(struct kvm_vcpu *vcpu)
1124{ 1088{
1125 int vector = kvm_apic_has_interrupt(vcpu); 1089 int vector = kvm_apic_has_interrupt(vcpu);
diff --git a/arch/x86/kvm/lapic.h b/arch/x86/kvm/lapic.h
index 81858881287e..45ab6ee71209 100644
--- a/arch/x86/kvm/lapic.h
+++ b/arch/x86/kvm/lapic.h
@@ -12,7 +12,6 @@ struct kvm_lapic {
12 atomic_t pending; 12 atomic_t pending;
13 s64 period; /* unit: ns */ 13 s64 period; /* unit: ns */
14 u32 divide_count; 14 u32 divide_count;
15 ktime_t last_update;
16 struct hrtimer dev; 15 struct hrtimer dev;
17 } timer; 16 } timer;
18 struct kvm_vcpu *vcpu; 17 struct kvm_vcpu *vcpu;
@@ -42,7 +41,6 @@ void kvm_set_apic_base(struct kvm_vcpu *vcpu, u64 data);
42void kvm_apic_post_state_restore(struct kvm_vcpu *vcpu); 41void kvm_apic_post_state_restore(struct kvm_vcpu *vcpu);
43int kvm_lapic_enabled(struct kvm_vcpu *vcpu); 42int kvm_lapic_enabled(struct kvm_vcpu *vcpu);
44int kvm_lapic_find_highest_irr(struct kvm_vcpu *vcpu); 43int kvm_lapic_find_highest_irr(struct kvm_vcpu *vcpu);
45void kvm_apic_timer_intr_post(struct kvm_vcpu *vcpu, int vec);
46 44
47void kvm_lapic_set_vapic_addr(struct kvm_vcpu *vcpu, gpa_t vapic_addr); 45void kvm_lapic_set_vapic_addr(struct kvm_vcpu *vcpu, gpa_t vapic_addr);
48void kvm_lapic_sync_from_vapic(struct kvm_vcpu *vcpu); 46void kvm_lapic_sync_from_vapic(struct kvm_vcpu *vcpu);
diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index 83f11c7474a1..2d4477c71473 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -1698,8 +1698,13 @@ static int set_spte(struct kvm_vcpu *vcpu, u64 *shadow_pte,
1698 if (largepage) 1698 if (largepage)
1699 spte |= PT_PAGE_SIZE_MASK; 1699 spte |= PT_PAGE_SIZE_MASK;
1700 if (mt_mask) { 1700 if (mt_mask) {
1701 mt_mask = get_memory_type(vcpu, gfn) << 1701 if (!kvm_is_mmio_pfn(pfn)) {
1702 kvm_x86_ops->get_mt_mask_shift(); 1702 mt_mask = get_memory_type(vcpu, gfn) <<
1703 kvm_x86_ops->get_mt_mask_shift();
1704 mt_mask |= VMX_EPT_IGMT_BIT;
1705 } else
1706 mt_mask = MTRR_TYPE_UNCACHABLE <<
1707 kvm_x86_ops->get_mt_mask_shift();
1703 spte |= mt_mask; 1708 spte |= mt_mask;
1704 } 1709 }
1705 1710
diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
index 1452851ae258..a9e769e4e251 100644
--- a/arch/x86/kvm/svm.c
+++ b/arch/x86/kvm/svm.c
@@ -1600,7 +1600,6 @@ static void svm_intr_assist(struct kvm_vcpu *vcpu)
1600 /* Okay, we can deliver the interrupt: grab it and update PIC state. */ 1600 /* Okay, we can deliver the interrupt: grab it and update PIC state. */
1601 intr_vector = kvm_cpu_get_interrupt(vcpu); 1601 intr_vector = kvm_cpu_get_interrupt(vcpu);
1602 svm_inject_irq(svm, intr_vector); 1602 svm_inject_irq(svm, intr_vector);
1603 kvm_timer_intr_post(vcpu, intr_vector);
1604out: 1603out:
1605 update_cr8_intercept(vcpu); 1604 update_cr8_intercept(vcpu);
1606} 1605}
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 6259d7467648..7611af576829 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -903,6 +903,7 @@ static int vmx_get_msr(struct kvm_vcpu *vcpu, u32 msr_index, u64 *pdata)
903 data = vmcs_readl(GUEST_SYSENTER_ESP); 903 data = vmcs_readl(GUEST_SYSENTER_ESP);
904 break; 904 break;
905 default: 905 default:
906 vmx_load_host_state(to_vmx(vcpu));
906 msr = find_msr_entry(to_vmx(vcpu), msr_index); 907 msr = find_msr_entry(to_vmx(vcpu), msr_index);
907 if (msr) { 908 if (msr) {
908 data = msr->data; 909 data = msr->data;
@@ -3285,7 +3286,6 @@ static void vmx_intr_assist(struct kvm_vcpu *vcpu)
3285 } 3286 }
3286 if (vcpu->arch.interrupt.pending) { 3287 if (vcpu->arch.interrupt.pending) {
3287 vmx_inject_irq(vcpu, vcpu->arch.interrupt.nr); 3288 vmx_inject_irq(vcpu, vcpu->arch.interrupt.nr);
3288 kvm_timer_intr_post(vcpu, vcpu->arch.interrupt.nr);
3289 if (kvm_cpu_has_interrupt(vcpu)) 3289 if (kvm_cpu_has_interrupt(vcpu))
3290 enable_irq_window(vcpu); 3290 enable_irq_window(vcpu);
3291 } 3291 }
@@ -3687,8 +3687,7 @@ static int __init vmx_init(void)
3687 if (vm_need_ept()) { 3687 if (vm_need_ept()) {
3688 bypass_guest_pf = 0; 3688 bypass_guest_pf = 0;
3689 kvm_mmu_set_base_ptes(VMX_EPT_READABLE_MASK | 3689 kvm_mmu_set_base_ptes(VMX_EPT_READABLE_MASK |
3690 VMX_EPT_WRITABLE_MASK | 3690 VMX_EPT_WRITABLE_MASK);
3691 VMX_EPT_IGMT_BIT);
3692 kvm_mmu_set_mask_ptes(0ull, 0ull, 0ull, 0ull, 3691 kvm_mmu_set_mask_ptes(0ull, 0ull, 0ull, 0ull,
3693 VMX_EPT_EXECUTABLE_MASK, 3692 VMX_EPT_EXECUTABLE_MASK,
3694 VMX_EPT_DEFAULT_MT << VMX_EPT_MT_EPTE_SHIFT); 3693 VMX_EPT_DEFAULT_MT << VMX_EPT_MT_EPTE_SHIFT);
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index cc17546a2406..758b7a155ae9 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -967,7 +967,6 @@ int kvm_dev_ioctl_check_extension(long ext)
967 case KVM_CAP_MMU_SHADOW_CACHE_CONTROL: 967 case KVM_CAP_MMU_SHADOW_CACHE_CONTROL:
968 case KVM_CAP_SET_TSS_ADDR: 968 case KVM_CAP_SET_TSS_ADDR:
969 case KVM_CAP_EXT_CPUID: 969 case KVM_CAP_EXT_CPUID:
970 case KVM_CAP_CLOCKSOURCE:
971 case KVM_CAP_PIT: 970 case KVM_CAP_PIT:
972 case KVM_CAP_NOP_IO_DELAY: 971 case KVM_CAP_NOP_IO_DELAY:
973 case KVM_CAP_MP_STATE: 972 case KVM_CAP_MP_STATE:
@@ -992,6 +991,9 @@ int kvm_dev_ioctl_check_extension(long ext)
992 case KVM_CAP_IOMMU: 991 case KVM_CAP_IOMMU:
993 r = iommu_found(); 992 r = iommu_found();
994 break; 993 break;
994 case KVM_CAP_CLOCKSOURCE:
995 r = boot_cpu_has(X86_FEATURE_CONSTANT_TSC);
996 break;
995 default: 997 default:
996 r = 0; 998 r = 0;
997 break; 999 break;
@@ -4127,9 +4129,13 @@ static void kvm_free_vcpus(struct kvm *kvm)
4127 4129
4128} 4130}
4129 4131
4130void kvm_arch_destroy_vm(struct kvm *kvm) 4132void kvm_arch_sync_events(struct kvm *kvm)
4131{ 4133{
4132 kvm_free_all_assigned_devices(kvm); 4134 kvm_free_all_assigned_devices(kvm);
4135}
4136
4137void kvm_arch_destroy_vm(struct kvm *kvm)
4138{
4133 kvm_iommu_unmap_guest(kvm); 4139 kvm_iommu_unmap_guest(kvm);
4134 kvm_free_pit(kvm); 4140 kvm_free_pit(kvm);
4135 kfree(kvm->arch.vpic); 4141 kfree(kvm->arch.vpic);
diff --git a/arch/x86/lguest/boot.c b/arch/x86/lguest/boot.c
index a7ed208f81e3..92f1c6f3e19d 100644
--- a/arch/x86/lguest/boot.c
+++ b/arch/x86/lguest/boot.c
@@ -931,7 +931,7 @@ static void lguest_restart(char *reason)
931 * that we can fit comfortably. 931 * that we can fit comfortably.
932 * 932 *
933 * First we need assembly templates of each of the patchable Guest operations, 933 * First we need assembly templates of each of the patchable Guest operations,
934 * and these are in lguest_asm.S. */ 934 * and these are in i386_head.S. */
935 935
936/*G:060 We construct a table from the assembler templates: */ 936/*G:060 We construct a table from the assembler templates: */
937static const struct lguest_insns 937static const struct lguest_insns
@@ -1093,7 +1093,7 @@ __init void lguest_init(void)
1093 acpi_ht = 0; 1093 acpi_ht = 0;
1094#endif 1094#endif
1095 1095
1096 /* We set the perferred console to "hvc". This is the "hypervisor 1096 /* We set the preferred console to "hvc". This is the "hypervisor
1097 * virtual console" driver written by the PowerPC people, which we also 1097 * virtual console" driver written by the PowerPC people, which we also
1098 * adapted for lguest's use. */ 1098 * adapted for lguest's use. */
1099 add_preferred_console("hvc", 0, NULL); 1099 add_preferred_console("hvc", 0, NULL);
diff --git a/arch/x86/lib/usercopy_32.c b/arch/x86/lib/usercopy_32.c
index 4a20b2f9a381..7c8ca91bb9ec 100644
--- a/arch/x86/lib/usercopy_32.c
+++ b/arch/x86/lib/usercopy_32.c
@@ -56,7 +56,7 @@ do { \
56 " jmp 2b\n" \ 56 " jmp 2b\n" \
57 ".previous\n" \ 57 ".previous\n" \
58 _ASM_EXTABLE(0b,3b) \ 58 _ASM_EXTABLE(0b,3b) \
59 : "=d"(res), "=c"(count), "=&a" (__d0), "=&S" (__d1), \ 59 : "=&d"(res), "=&c"(count), "=&a" (__d0), "=&S" (__d1), \
60 "=&D" (__d2) \ 60 "=&D" (__d2) \
61 : "i"(-EFAULT), "0"(count), "1"(count), "3"(src), "4"(dst) \ 61 : "i"(-EFAULT), "0"(count), "1"(count), "3"(src), "4"(dst) \
62 : "memory"); \ 62 : "memory"); \
@@ -218,7 +218,7 @@ long strnlen_user(const char __user *s, long n)
218 " .align 4\n" 218 " .align 4\n"
219 " .long 0b,2b\n" 219 " .long 0b,2b\n"
220 ".previous" 220 ".previous"
221 :"=r" (n), "=D" (s), "=a" (res), "=c" (tmp) 221 :"=&r" (n), "=&D" (s), "=&a" (res), "=&c" (tmp)
222 :"0" (n), "1" (s), "2" (0), "3" (mask) 222 :"0" (n), "1" (s), "2" (0), "3" (mask)
223 :"cc"); 223 :"cc");
224 return res & mask; 224 return res & mask;
diff --git a/arch/x86/lib/usercopy_64.c b/arch/x86/lib/usercopy_64.c
index 64d6c84e6353..ec13cb5f17ed 100644
--- a/arch/x86/lib/usercopy_64.c
+++ b/arch/x86/lib/usercopy_64.c
@@ -32,7 +32,7 @@ do { \
32 " jmp 2b\n" \ 32 " jmp 2b\n" \
33 ".previous\n" \ 33 ".previous\n" \
34 _ASM_EXTABLE(0b,3b) \ 34 _ASM_EXTABLE(0b,3b) \
35 : "=r"(res), "=c"(count), "=&a" (__d0), "=&S" (__d1), \ 35 : "=&r"(res), "=&c"(count), "=&a" (__d0), "=&S" (__d1), \
36 "=&D" (__d2) \ 36 "=&D" (__d2) \
37 : "i"(-EFAULT), "0"(count), "1"(count), "3"(src), "4"(dst) \ 37 : "i"(-EFAULT), "0"(count), "1"(count), "3"(src), "4"(dst) \
38 : "memory"); \ 38 : "memory"); \
@@ -86,7 +86,7 @@ unsigned long __clear_user(void __user *addr, unsigned long size)
86 ".previous\n" 86 ".previous\n"
87 _ASM_EXTABLE(0b,3b) 87 _ASM_EXTABLE(0b,3b)
88 _ASM_EXTABLE(1b,2b) 88 _ASM_EXTABLE(1b,2b)
89 : [size8] "=c"(size), [dst] "=&D" (__d0) 89 : [size8] "=&c"(size), [dst] "=&D" (__d0)
90 : [size1] "r"(size & 7), "[size8]" (size / 8), "[dst]"(addr), 90 : [size1] "r"(size & 7), "[size8]" (size / 8), "[dst]"(addr),
91 [zero] "r" (0UL), [eight] "r" (8UL)); 91 [zero] "r" (0UL), [eight] "r" (8UL));
92 return size; 92 return size;
diff --git a/arch/x86/mach-default/setup.c b/arch/x86/mach-default/setup.c
index df167f265622..a265a7c63190 100644
--- a/arch/x86/mach-default/setup.c
+++ b/arch/x86/mach-default/setup.c
@@ -38,6 +38,15 @@ void __init pre_intr_init_hook(void)
38 init_ISA_irqs(); 38 init_ISA_irqs();
39} 39}
40 40
41/*
42 * IRQ2 is cascade interrupt to second interrupt controller
43 */
44static struct irqaction irq2 = {
45 .handler = no_action,
46 .mask = CPU_MASK_NONE,
47 .name = "cascade",
48};
49
41/** 50/**
42 * intr_init_hook - post gate setup interrupt initialisation 51 * intr_init_hook - post gate setup interrupt initialisation
43 * 52 *
@@ -53,6 +62,9 @@ void __init intr_init_hook(void)
53 if (x86_quirks->arch_intr_init()) 62 if (x86_quirks->arch_intr_init())
54 return; 63 return;
55 } 64 }
65 if (!acpi_ioapic)
66 setup_irq(2, &irq2);
67
56} 68}
57 69
58/** 70/**
diff --git a/arch/x86/mach-voyager/setup.c b/arch/x86/mach-voyager/setup.c
index a580b9562e76..d914a7996a66 100644
--- a/arch/x86/mach-voyager/setup.c
+++ b/arch/x86/mach-voyager/setup.c
@@ -33,13 +33,23 @@ void __init intr_init_hook(void)
33 setup_irq(2, &irq2); 33 setup_irq(2, &irq2);
34} 34}
35 35
36void __init pre_setup_arch_hook(void) 36static void voyager_disable_tsc(void)
37{ 37{
38 /* Voyagers run their CPUs from independent clocks, so disable 38 /* Voyagers run their CPUs from independent clocks, so disable
39 * the TSC code because we can't sync them */ 39 * the TSC code because we can't sync them */
40 setup_clear_cpu_cap(X86_FEATURE_TSC); 40 setup_clear_cpu_cap(X86_FEATURE_TSC);
41} 41}
42 42
43void __init pre_setup_arch_hook(void)
44{
45 voyager_disable_tsc();
46}
47
48void __init pre_time_init_hook(void)
49{
50 voyager_disable_tsc();
51}
52
43void __init trap_init_hook(void) 53void __init trap_init_hook(void)
44{ 54{
45} 55}
diff --git a/arch/x86/mach-voyager/voyager_smp.c b/arch/x86/mach-voyager/voyager_smp.c
index 9840b7ec749a..7ffcdeec4631 100644
--- a/arch/x86/mach-voyager/voyager_smp.c
+++ b/arch/x86/mach-voyager/voyager_smp.c
@@ -81,7 +81,7 @@ static void enable_local_vic_irq(unsigned int irq);
81static void disable_local_vic_irq(unsigned int irq); 81static void disable_local_vic_irq(unsigned int irq);
82static void before_handle_vic_irq(unsigned int irq); 82static void before_handle_vic_irq(unsigned int irq);
83static void after_handle_vic_irq(unsigned int irq); 83static void after_handle_vic_irq(unsigned int irq);
84static void set_vic_irq_affinity(unsigned int irq, cpumask_t mask); 84static void set_vic_irq_affinity(unsigned int irq, const struct cpumask *mask);
85static void ack_vic_irq(unsigned int irq); 85static void ack_vic_irq(unsigned int irq);
86static void vic_enable_cpi(void); 86static void vic_enable_cpi(void);
87static void do_boot_cpu(__u8 cpuid); 87static void do_boot_cpu(__u8 cpuid);
@@ -211,8 +211,6 @@ static __u32 cpu_booted_map;
211static cpumask_t smp_commenced_mask = CPU_MASK_NONE; 211static cpumask_t smp_commenced_mask = CPU_MASK_NONE;
212 212
213/* This is for the new dynamic CPU boot code */ 213/* This is for the new dynamic CPU boot code */
214cpumask_t cpu_callin_map = CPU_MASK_NONE;
215cpumask_t cpu_callout_map = CPU_MASK_NONE;
216 214
217/* The per processor IRQ masks (these are usually kept in sync) */ 215/* The per processor IRQ masks (these are usually kept in sync) */
218static __u16 vic_irq_mask[NR_CPUS] __cacheline_aligned; 216static __u16 vic_irq_mask[NR_CPUS] __cacheline_aligned;
@@ -378,7 +376,7 @@ void __init find_smp_config(void)
378 cpus_addr(phys_cpu_present_map)[0] |= 376 cpus_addr(phys_cpu_present_map)[0] |=
379 voyager_extended_cmos_read(VOYAGER_PROCESSOR_PRESENT_MASK + 377 voyager_extended_cmos_read(VOYAGER_PROCESSOR_PRESENT_MASK +
380 3) << 24; 378 3) << 24;
381 cpu_possible_map = phys_cpu_present_map; 379 init_cpu_possible(&phys_cpu_present_map);
382 printk("VOYAGER SMP: phys_cpu_present_map = 0x%lx\n", 380 printk("VOYAGER SMP: phys_cpu_present_map = 0x%lx\n",
383 cpus_addr(phys_cpu_present_map)[0]); 381 cpus_addr(phys_cpu_present_map)[0]);
384 /* Here we set up the VIC to enable SMP */ 382 /* Here we set up the VIC to enable SMP */
@@ -1599,16 +1597,16 @@ static void after_handle_vic_irq(unsigned int irq)
1599 * change the mask and then do an interrupt enable CPI to re-enable on 1597 * change the mask and then do an interrupt enable CPI to re-enable on
1600 * the selected processors */ 1598 * the selected processors */
1601 1599
1602void set_vic_irq_affinity(unsigned int irq, cpumask_t mask) 1600void set_vic_irq_affinity(unsigned int irq, const struct cpumask *mask)
1603{ 1601{
1604 /* Only extended processors handle interrupts */ 1602 /* Only extended processors handle interrupts */
1605 unsigned long real_mask; 1603 unsigned long real_mask;
1606 unsigned long irq_mask = 1 << irq; 1604 unsigned long irq_mask = 1 << irq;
1607 int cpu; 1605 int cpu;
1608 1606
1609 real_mask = cpus_addr(mask)[0] & voyager_extended_vic_processors; 1607 real_mask = cpus_addr(*mask)[0] & voyager_extended_vic_processors;
1610 1608
1611 if (cpus_addr(mask)[0] == 0) 1609 if (cpus_addr(*mask)[0] == 0)
1612 /* can't have no CPUs to accept the interrupt -- extremely 1610 /* can't have no CPUs to accept the interrupt -- extremely
1613 * bad things will happen */ 1611 * bad things will happen */
1614 return; 1612 return;
@@ -1750,10 +1748,11 @@ static void __cpuinit voyager_smp_prepare_boot_cpu(void)
1750 init_gdt(smp_processor_id()); 1748 init_gdt(smp_processor_id());
1751 switch_to_new_gdt(); 1749 switch_to_new_gdt();
1752 1750
1753 cpu_set(smp_processor_id(), cpu_online_map); 1751 cpu_online_map = cpumask_of_cpu(smp_processor_id());
1754 cpu_set(smp_processor_id(), cpu_callout_map); 1752 cpu_callout_map = cpumask_of_cpu(smp_processor_id());
1755 cpu_set(smp_processor_id(), cpu_possible_map); 1753 cpu_callin_map = CPU_MASK_NONE;
1756 cpu_set(smp_processor_id(), cpu_present_map); 1754 cpu_present_map = cpumask_of_cpu(smp_processor_id());
1755
1757} 1756}
1758 1757
1759static int __cpuinit voyager_cpu_up(unsigned int cpu) 1758static int __cpuinit voyager_cpu_up(unsigned int cpu)
@@ -1783,9 +1782,9 @@ void __init smp_setup_processor_id(void)
1783 x86_write_percpu(cpu_number, hard_smp_processor_id()); 1782 x86_write_percpu(cpu_number, hard_smp_processor_id());
1784} 1783}
1785 1784
1786static void voyager_send_call_func(cpumask_t callmask) 1785static void voyager_send_call_func(const struct cpumask *callmask)
1787{ 1786{
1788 __u32 mask = cpus_addr(callmask)[0] & ~(1 << smp_processor_id()); 1787 __u32 mask = cpus_addr(*callmask)[0] & ~(1 << smp_processor_id());
1789 send_CPI(mask, VIC_CALL_FUNCTION_CPI); 1788 send_CPI(mask, VIC_CALL_FUNCTION_CPI);
1790} 1789}
1791 1790
diff --git a/arch/x86/math-emu/fpu_entry.c b/arch/x86/math-emu/fpu_entry.c
index c7b06feb139b..5d87f586f8d7 100644
--- a/arch/x86/math-emu/fpu_entry.c
+++ b/arch/x86/math-emu/fpu_entry.c
@@ -131,7 +131,7 @@ u_char emulating = 0;
131static int valid_prefix(u_char *Byte, u_char __user ** fpu_eip, 131static int valid_prefix(u_char *Byte, u_char __user ** fpu_eip,
132 overrides * override); 132 overrides * override);
133 133
134asmlinkage void math_emulate(long arg) 134void math_emulate(struct math_emu_info *info)
135{ 135{
136 u_char FPU_modrm, byte1; 136 u_char FPU_modrm, byte1;
137 unsigned short code; 137 unsigned short code;
@@ -161,7 +161,7 @@ asmlinkage void math_emulate(long arg)
161 RE_ENTRANT_CHECK_ON; 161 RE_ENTRANT_CHECK_ON;
162#endif /* RE_ENTRANT_CHECKING */ 162#endif /* RE_ENTRANT_CHECKING */
163 163
164 SETUP_DATA_AREA(arg); 164 FPU_info = info;
165 165
166 FPU_ORIG_EIP = FPU_EIP; 166 FPU_ORIG_EIP = FPU_EIP;
167 167
@@ -659,7 +659,7 @@ static int valid_prefix(u_char *Byte, u_char __user **fpu_eip,
659 } 659 }
660} 660}
661 661
662void math_abort(struct info *info, unsigned int signal) 662void math_abort(struct math_emu_info *info, unsigned int signal)
663{ 663{
664 FPU_EIP = FPU_ORIG_EIP; 664 FPU_EIP = FPU_ORIG_EIP;
665 current->thread.trap_no = 16; 665 current->thread.trap_no = 16;
diff --git a/arch/x86/math-emu/fpu_proto.h b/arch/x86/math-emu/fpu_proto.h
index aa49b6a0d850..9779df436b7d 100644
--- a/arch/x86/math-emu/fpu_proto.h
+++ b/arch/x86/math-emu/fpu_proto.h
@@ -51,8 +51,8 @@ extern void ffreep(void);
51extern void fst_i_(void); 51extern void fst_i_(void);
52extern void fstp_i(void); 52extern void fstp_i(void);
53/* fpu_entry.c */ 53/* fpu_entry.c */
54asmlinkage extern void math_emulate(long arg); 54extern void math_emulate(struct math_emu_info *info);
55extern void math_abort(struct info *info, unsigned int signal); 55extern void math_abort(struct math_emu_info *info, unsigned int signal);
56/* fpu_etc.c */ 56/* fpu_etc.c */
57extern void FPU_etc(void); 57extern void FPU_etc(void);
58/* fpu_tags.c */ 58/* fpu_tags.c */
diff --git a/arch/x86/math-emu/fpu_system.h b/arch/x86/math-emu/fpu_system.h
index 13488fa153e0..50fa0ec2c8a5 100644
--- a/arch/x86/math-emu/fpu_system.h
+++ b/arch/x86/math-emu/fpu_system.h
@@ -16,10 +16,6 @@
16#include <linux/kernel.h> 16#include <linux/kernel.h>
17#include <linux/mm.h> 17#include <linux/mm.h>
18 18
19/* This sets the pointer FPU_info to point to the argument part
20 of the stack frame of math_emulate() */
21#define SETUP_DATA_AREA(arg) FPU_info = (struct info *) &arg
22
23/* s is always from a cpu register, and the cpu does bounds checking 19/* s is always from a cpu register, and the cpu does bounds checking
24 * during register load --> no further bounds checks needed */ 20 * during register load --> no further bounds checks needed */
25#define LDT_DESCRIPTOR(s) (((struct desc_struct *)current->mm->context.ldt)[(s) >> 3]) 21#define LDT_DESCRIPTOR(s) (((struct desc_struct *)current->mm->context.ldt)[(s) >> 3])
@@ -38,12 +34,12 @@
38#define I387 (current->thread.xstate) 34#define I387 (current->thread.xstate)
39#define FPU_info (I387->soft.info) 35#define FPU_info (I387->soft.info)
40 36
41#define FPU_CS (*(unsigned short *) &(FPU_info->___cs)) 37#define FPU_CS (*(unsigned short *) &(FPU_info->regs->cs))
42#define FPU_SS (*(unsigned short *) &(FPU_info->___ss)) 38#define FPU_SS (*(unsigned short *) &(FPU_info->regs->ss))
43#define FPU_DS (*(unsigned short *) &(FPU_info->___ds)) 39#define FPU_DS (*(unsigned short *) &(FPU_info->regs->ds))
44#define FPU_EAX (FPU_info->___eax) 40#define FPU_EAX (FPU_info->regs->ax)
45#define FPU_EFLAGS (FPU_info->___eflags) 41#define FPU_EFLAGS (FPU_info->regs->flags)
46#define FPU_EIP (FPU_info->___eip) 42#define FPU_EIP (FPU_info->regs->ip)
47#define FPU_ORIG_EIP (FPU_info->___orig_eip) 43#define FPU_ORIG_EIP (FPU_info->___orig_eip)
48 44
49#define FPU_lookahead (I387->soft.lookahead) 45#define FPU_lookahead (I387->soft.lookahead)
diff --git a/arch/x86/math-emu/get_address.c b/arch/x86/math-emu/get_address.c
index d701e2b39e44..420b3b6e3915 100644
--- a/arch/x86/math-emu/get_address.c
+++ b/arch/x86/math-emu/get_address.c
@@ -29,46 +29,43 @@
29#define FPU_WRITE_BIT 0x10 29#define FPU_WRITE_BIT 0x10
30 30
31static int reg_offset[] = { 31static int reg_offset[] = {
32 offsetof(struct info, ___eax), 32 offsetof(struct pt_regs, ax),
33 offsetof(struct info, ___ecx), 33 offsetof(struct pt_regs, cx),
34 offsetof(struct info, ___edx), 34 offsetof(struct pt_regs, dx),
35 offsetof(struct info, ___ebx), 35 offsetof(struct pt_regs, bx),
36 offsetof(struct info, ___esp), 36 offsetof(struct pt_regs, sp),
37 offsetof(struct info, ___ebp), 37 offsetof(struct pt_regs, bp),
38 offsetof(struct info, ___esi), 38 offsetof(struct pt_regs, si),
39 offsetof(struct info, ___edi) 39 offsetof(struct pt_regs, di)
40}; 40};
41 41
42#define REG_(x) (*(long *)(reg_offset[(x)]+(u_char *) FPU_info)) 42#define REG_(x) (*(long *)(reg_offset[(x)] + (u_char *)FPU_info->regs))
43 43
44static int reg_offset_vm86[] = { 44static int reg_offset_vm86[] = {
45 offsetof(struct info, ___cs), 45 offsetof(struct pt_regs, cs),
46 offsetof(struct info, ___vm86_ds), 46 offsetof(struct kernel_vm86_regs, ds),
47 offsetof(struct info, ___vm86_es), 47 offsetof(struct kernel_vm86_regs, es),
48 offsetof(struct info, ___vm86_fs), 48 offsetof(struct kernel_vm86_regs, fs),
49 offsetof(struct info, ___vm86_gs), 49 offsetof(struct kernel_vm86_regs, gs),
50 offsetof(struct info, ___ss), 50 offsetof(struct pt_regs, ss),
51 offsetof(struct info, ___vm86_ds) 51 offsetof(struct kernel_vm86_regs, ds)
52}; 52};
53 53
54#define VM86_REG_(x) (*(unsigned short *) \ 54#define VM86_REG_(x) (*(unsigned short *) \
55 (reg_offset_vm86[((unsigned)x)]+(u_char *) FPU_info)) 55 (reg_offset_vm86[((unsigned)x)] + (u_char *)FPU_info->regs))
56
57/* This dummy, gs is not saved on the stack. */
58#define ___GS ___ds
59 56
60static int reg_offset_pm[] = { 57static int reg_offset_pm[] = {
61 offsetof(struct info, ___cs), 58 offsetof(struct pt_regs, cs),
62 offsetof(struct info, ___ds), 59 offsetof(struct pt_regs, ds),
63 offsetof(struct info, ___es), 60 offsetof(struct pt_regs, es),
64 offsetof(struct info, ___fs), 61 offsetof(struct pt_regs, fs),
65 offsetof(struct info, ___GS), 62 offsetof(struct pt_regs, ds), /* dummy, not saved on stack */
66 offsetof(struct info, ___ss), 63 offsetof(struct pt_regs, ss),
67 offsetof(struct info, ___ds) 64 offsetof(struct pt_regs, ds)
68}; 65};
69 66
70#define PM_REG_(x) (*(unsigned short *) \ 67#define PM_REG_(x) (*(unsigned short *) \
71 (reg_offset_pm[((unsigned)x)]+(u_char *) FPU_info)) 68 (reg_offset_pm[((unsigned)x)] + (u_char *)FPU_info->regs))
72 69
73/* Decode the SIB byte. This function assumes mod != 0 */ 70/* Decode the SIB byte. This function assumes mod != 0 */
74static int sib(int mod, unsigned long *fpu_eip) 71static int sib(int mod, unsigned long *fpu_eip)
@@ -349,34 +346,34 @@ void __user *FPU_get_address_16(u_char FPU_modrm, unsigned long *fpu_eip,
349 } 346 }
350 switch (rm) { 347 switch (rm) {
351 case 0: 348 case 0:
352 address += FPU_info->___ebx + FPU_info->___esi; 349 address += FPU_info->regs->bx + FPU_info->regs->si;
353 break; 350 break;
354 case 1: 351 case 1:
355 address += FPU_info->___ebx + FPU_info->___edi; 352 address += FPU_info->regs->bx + FPU_info->regs->di;
356 break; 353 break;
357 case 2: 354 case 2:
358 address += FPU_info->___ebp + FPU_info->___esi; 355 address += FPU_info->regs->bp + FPU_info->regs->si;
359 if (addr_modes.override.segment == PREFIX_DEFAULT) 356 if (addr_modes.override.segment == PREFIX_DEFAULT)
360 addr_modes.override.segment = PREFIX_SS_; 357 addr_modes.override.segment = PREFIX_SS_;
361 break; 358 break;
362 case 3: 359 case 3:
363 address += FPU_info->___ebp + FPU_info->___edi; 360 address += FPU_info->regs->bp + FPU_info->regs->di;
364 if (addr_modes.override.segment == PREFIX_DEFAULT) 361 if (addr_modes.override.segment == PREFIX_DEFAULT)
365 addr_modes.override.segment = PREFIX_SS_; 362 addr_modes.override.segment = PREFIX_SS_;
366 break; 363 break;
367 case 4: 364 case 4:
368 address += FPU_info->___esi; 365 address += FPU_info->regs->si;
369 break; 366 break;
370 case 5: 367 case 5:
371 address += FPU_info->___edi; 368 address += FPU_info->regs->di;
372 break; 369 break;
373 case 6: 370 case 6:
374 address += FPU_info->___ebp; 371 address += FPU_info->regs->bp;
375 if (addr_modes.override.segment == PREFIX_DEFAULT) 372 if (addr_modes.override.segment == PREFIX_DEFAULT)
376 addr_modes.override.segment = PREFIX_SS_; 373 addr_modes.override.segment = PREFIX_SS_;
377 break; 374 break;
378 case 7: 375 case 7:
379 address += FPU_info->___ebx; 376 address += FPU_info->regs->bx;
380 break; 377 break;
381 } 378 }
382 379
diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c
index 9e268b6b204e..c76ef1d701c9 100644
--- a/arch/x86/mm/fault.c
+++ b/arch/x86/mm/fault.c
@@ -534,7 +534,7 @@ static int vmalloc_fault(unsigned long address)
534 happen within a race in page table update. In the later 534 happen within a race in page table update. In the later
535 case just flush. */ 535 case just flush. */
536 536
537 pgd = pgd_offset(current->mm ?: &init_mm, address); 537 pgd = pgd_offset(current->active_mm, address);
538 pgd_ref = pgd_offset_k(address); 538 pgd_ref = pgd_offset_k(address);
539 if (pgd_none(*pgd_ref)) 539 if (pgd_none(*pgd_ref))
540 return -1; 540 return -1;
@@ -603,8 +603,6 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code)
603 603
604 si_code = SEGV_MAPERR; 604 si_code = SEGV_MAPERR;
605 605
606 if (notify_page_fault(regs))
607 return;
608 if (unlikely(kmmio_fault(regs, address))) 606 if (unlikely(kmmio_fault(regs, address)))
609 return; 607 return;
610 608
@@ -634,6 +632,9 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code)
634 if (spurious_fault(address, error_code)) 632 if (spurious_fault(address, error_code))
635 return; 633 return;
636 634
635 /* kprobes don't want to hook the spurious faults. */
636 if (notify_page_fault(regs))
637 return;
637 /* 638 /*
638 * Don't take the mm semaphore here. If we fixup a prefetch 639 * Don't take the mm semaphore here. If we fixup a prefetch
639 * fault we could otherwise deadlock. 640 * fault we could otherwise deadlock.
@@ -641,6 +642,9 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code)
641 goto bad_area_nosemaphore; 642 goto bad_area_nosemaphore;
642 } 643 }
643 644
645 /* kprobes don't want to hook the spurious faults. */
646 if (notify_page_fault(regs))
647 return;
644 648
645 /* 649 /*
646 * It's safe to allow irq's after cr2 has been saved and the 650 * It's safe to allow irq's after cr2 has been saved and the
diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c
index 88f1b10de3be..2cef05074413 100644
--- a/arch/x86/mm/init_32.c
+++ b/arch/x86/mm/init_32.c
@@ -138,6 +138,47 @@ static pte_t * __init one_page_table_init(pmd_t *pmd)
138 return pte_offset_kernel(pmd, 0); 138 return pte_offset_kernel(pmd, 0);
139} 139}
140 140
141static pte_t *__init page_table_kmap_check(pte_t *pte, pmd_t *pmd,
142 unsigned long vaddr, pte_t *lastpte)
143{
144#ifdef CONFIG_HIGHMEM
145 /*
146 * Something (early fixmap) may already have put a pte
147 * page here, which causes the page table allocation
148 * to become nonlinear. Attempt to fix it, and if it
149 * is still nonlinear then we have to bug.
150 */
151 int pmd_idx_kmap_begin = fix_to_virt(FIX_KMAP_END) >> PMD_SHIFT;
152 int pmd_idx_kmap_end = fix_to_virt(FIX_KMAP_BEGIN) >> PMD_SHIFT;
153
154 if (pmd_idx_kmap_begin != pmd_idx_kmap_end
155 && (vaddr >> PMD_SHIFT) >= pmd_idx_kmap_begin
156 && (vaddr >> PMD_SHIFT) <= pmd_idx_kmap_end
157 && ((__pa(pte) >> PAGE_SHIFT) < table_start
158 || (__pa(pte) >> PAGE_SHIFT) >= table_end)) {
159 pte_t *newpte;
160 int i;
161
162 BUG_ON(after_init_bootmem);
163 newpte = alloc_low_page();
164 for (i = 0; i < PTRS_PER_PTE; i++)
165 set_pte(newpte + i, pte[i]);
166
167 paravirt_alloc_pte(&init_mm, __pa(newpte) >> PAGE_SHIFT);
168 set_pmd(pmd, __pmd(__pa(newpte)|_PAGE_TABLE));
169 BUG_ON(newpte != pte_offset_kernel(pmd, 0));
170 __flush_tlb_all();
171
172 paravirt_release_pte(__pa(pte) >> PAGE_SHIFT);
173 pte = newpte;
174 }
175 BUG_ON(vaddr < fix_to_virt(FIX_KMAP_BEGIN - 1)
176 && vaddr > fix_to_virt(FIX_KMAP_END)
177 && lastpte && lastpte + PTRS_PER_PTE != pte);
178#endif
179 return pte;
180}
181
141/* 182/*
142 * This function initializes a certain range of kernel virtual memory 183 * This function initializes a certain range of kernel virtual memory
143 * with new bootmem page tables, everywhere page tables are missing in 184 * with new bootmem page tables, everywhere page tables are missing in
@@ -154,6 +195,7 @@ page_table_range_init(unsigned long start, unsigned long end, pgd_t *pgd_base)
154 unsigned long vaddr; 195 unsigned long vaddr;
155 pgd_t *pgd; 196 pgd_t *pgd;
156 pmd_t *pmd; 197 pmd_t *pmd;
198 pte_t *pte = NULL;
157 199
158 vaddr = start; 200 vaddr = start;
159 pgd_idx = pgd_index(vaddr); 201 pgd_idx = pgd_index(vaddr);
@@ -165,7 +207,8 @@ page_table_range_init(unsigned long start, unsigned long end, pgd_t *pgd_base)
165 pmd = pmd + pmd_index(vaddr); 207 pmd = pmd + pmd_index(vaddr);
166 for (; (pmd_idx < PTRS_PER_PMD) && (vaddr != end); 208 for (; (pmd_idx < PTRS_PER_PMD) && (vaddr != end);
167 pmd++, pmd_idx++) { 209 pmd++, pmd_idx++) {
168 one_page_table_init(pmd); 210 pte = page_table_kmap_check(one_page_table_init(pmd),
211 pmd, vaddr, pte);
169 212
170 vaddr += PMD_SIZE; 213 vaddr += PMD_SIZE;
171 } 214 }
@@ -508,7 +551,6 @@ static void __init early_ioremap_page_table_range_init(pgd_t *pgd_base)
508 * Fixed mappings, only the page table structure has to be 551 * Fixed mappings, only the page table structure has to be
509 * created - mappings will be set by set_fixmap(): 552 * created - mappings will be set by set_fixmap():
510 */ 553 */
511 early_ioremap_clear();
512 vaddr = __fix_to_virt(__end_of_fixed_addresses - 1) & PMD_MASK; 554 vaddr = __fix_to_virt(__end_of_fixed_addresses - 1) & PMD_MASK;
513 end = (FIXADDR_TOP + PMD_SIZE - 1) & PMD_MASK; 555 end = (FIXADDR_TOP + PMD_SIZE - 1) & PMD_MASK;
514 page_table_range_init(vaddr, end, pgd_base); 556 page_table_range_init(vaddr, end, pgd_base);
@@ -801,7 +843,7 @@ static void __init find_early_table_space(unsigned long end, int use_pse)
801 tables += PAGE_ALIGN(ptes * sizeof(pte_t)); 843 tables += PAGE_ALIGN(ptes * sizeof(pte_t));
802 844
803 /* for fixmap */ 845 /* for fixmap */
804 tables += PAGE_SIZE * 2; 846 tables += PAGE_ALIGN(__end_of_fixed_addresses * sizeof(pte_t));
805 847
806 /* 848 /*
807 * RED-PEN putting page tables only on node 0 could 849 * RED-PEN putting page tables only on node 0 could
diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c
index 23f68e77ad1f..e6d36b490250 100644
--- a/arch/x86/mm/init_64.c
+++ b/arch/x86/mm/init_64.c
@@ -596,7 +596,7 @@ static void __init init_gbpages(void)
596 direct_gbpages = 0; 596 direct_gbpages = 0;
597} 597}
598 598
599static unsigned long __init kernel_physical_mapping_init(unsigned long start, 599static unsigned long __meminit kernel_physical_mapping_init(unsigned long start,
600 unsigned long end, 600 unsigned long end,
601 unsigned long page_size_mask) 601 unsigned long page_size_mask)
602{ 602{
diff --git a/arch/x86/mm/iomap_32.c b/arch/x86/mm/iomap_32.c
index d0151d8ce452..ca53224fc56c 100644
--- a/arch/x86/mm/iomap_32.c
+++ b/arch/x86/mm/iomap_32.c
@@ -17,6 +17,7 @@
17 */ 17 */
18 18
19#include <asm/iomap.h> 19#include <asm/iomap.h>
20#include <asm/pat.h>
20#include <linux/module.h> 21#include <linux/module.h>
21 22
22/* Map 'pfn' using fixed map 'type' and protections 'prot' 23/* Map 'pfn' using fixed map 'type' and protections 'prot'
@@ -29,6 +30,15 @@ iomap_atomic_prot_pfn(unsigned long pfn, enum km_type type, pgprot_t prot)
29 30
30 pagefault_disable(); 31 pagefault_disable();
31 32
33 /*
34 * For non-PAT systems, promote PAGE_KERNEL_WC to PAGE_KERNEL_UC_MINUS.
35 * PAGE_KERNEL_WC maps to PWT, which translates to uncached if the
36 * MTRR is UC or WC. UC_MINUS gets the real intention, of the
37 * user, which is "WC if the MTRR is WC, UC if you can't do that."
38 */
39 if (!pat_enabled && pgprot_val(prot) == pgprot_val(PAGE_KERNEL_WC))
40 prot = PAGE_KERNEL_UC_MINUS;
41
32 idx = type + KM_TYPE_NR*smp_processor_id(); 42 idx = type + KM_TYPE_NR*smp_processor_id();
33 vaddr = __fix_to_virt(FIX_KMAP_BEGIN + idx); 43 vaddr = __fix_to_virt(FIX_KMAP_BEGIN + idx);
34 set_pte(kmap_pte-idx, pfn_pte(pfn, prot)); 44 set_pte(kmap_pte-idx, pfn_pte(pfn, prot));
diff --git a/arch/x86/mm/ioremap.c b/arch/x86/mm/ioremap.c
index bd85d42819e1..f45d5e29a72e 100644
--- a/arch/x86/mm/ioremap.c
+++ b/arch/x86/mm/ioremap.c
@@ -134,25 +134,6 @@ int page_is_ram(unsigned long pagenr)
134 return 0; 134 return 0;
135} 135}
136 136
137int pagerange_is_ram(unsigned long start, unsigned long end)
138{
139 int ram_page = 0, not_rampage = 0;
140 unsigned long page_nr;
141
142 for (page_nr = (start >> PAGE_SHIFT); page_nr < (end >> PAGE_SHIFT);
143 ++page_nr) {
144 if (page_is_ram(page_nr))
145 ram_page = 1;
146 else
147 not_rampage = 1;
148
149 if (ram_page == not_rampage)
150 return -1;
151 }
152
153 return ram_page;
154}
155
156/* 137/*
157 * Fix up the linear direct mapping of the kernel to avoid cache attribute 138 * Fix up the linear direct mapping of the kernel to avoid cache attribute
158 * conflicts. 139 * conflicts.
@@ -557,34 +538,9 @@ void __init early_ioremap_init(void)
557 } 538 }
558} 539}
559 540
560void __init early_ioremap_clear(void)
561{
562 pmd_t *pmd;
563
564 if (early_ioremap_debug)
565 printk(KERN_INFO "early_ioremap_clear()\n");
566
567 pmd = early_ioremap_pmd(fix_to_virt(FIX_BTMAP_BEGIN));
568 pmd_clear(pmd);
569 paravirt_release_pte(__pa(bm_pte) >> PAGE_SHIFT);
570 __flush_tlb_all();
571}
572
573void __init early_ioremap_reset(void) 541void __init early_ioremap_reset(void)
574{ 542{
575 enum fixed_addresses idx;
576 unsigned long addr, phys;
577 pte_t *pte;
578
579 after_paging_init = 1; 543 after_paging_init = 1;
580 for (idx = FIX_BTMAP_BEGIN; idx >= FIX_BTMAP_END; idx--) {
581 addr = fix_to_virt(idx);
582 pte = early_ioremap_pte(addr);
583 if (pte_present(*pte)) {
584 phys = pte_val(*pte) & PAGE_MASK;
585 set_fixmap(idx, phys);
586 }
587 }
588} 544}
589 545
590static void __init __early_set_fixmap(enum fixed_addresses idx, 546static void __init __early_set_fixmap(enum fixed_addresses idx,
diff --git a/arch/x86/mm/numa_64.c b/arch/x86/mm/numa_64.c
index 71a14f89f89e..f3516da035d1 100644
--- a/arch/x86/mm/numa_64.c
+++ b/arch/x86/mm/numa_64.c
@@ -145,7 +145,7 @@ int __init compute_hash_shift(struct bootnode *nodes, int numnodes,
145 return shift; 145 return shift;
146} 146}
147 147
148int early_pfn_to_nid(unsigned long pfn) 148int __meminit __early_pfn_to_nid(unsigned long pfn)
149{ 149{
150 return phys_to_nid(pfn << PAGE_SHIFT); 150 return phys_to_nid(pfn << PAGE_SHIFT);
151} 151}
diff --git a/arch/x86/mm/pageattr.c b/arch/x86/mm/pageattr.c
index e89d24815f26..7be47d1a97e4 100644
--- a/arch/x86/mm/pageattr.c
+++ b/arch/x86/mm/pageattr.c
@@ -508,18 +508,13 @@ static int split_large_page(pte_t *kpte, unsigned long address)
508#endif 508#endif
509 509
510 /* 510 /*
511 * Install the new, split up pagetable. Important details here: 511 * Install the new, split up pagetable.
512 * 512 *
513 * On Intel the NX bit of all levels must be cleared to make a 513 * We use the standard kernel pagetable protections for the new
514 * page executable. See section 4.13.2 of Intel 64 and IA-32 514 * pagetable protections, the actual ptes set above control the
515 * Architectures Software Developer's Manual). 515 * primary protection behavior:
516 *
517 * Mark the entry present. The current mapping might be
518 * set to not present, which we preserved above.
519 */ 516 */
520 ref_prot = pte_pgprot(pte_mkexec(pte_clrhuge(*kpte))); 517 __set_pmd_pte(kpte, address, mk_pte(base, __pgprot(_KERNPG_TABLE)));
521 pgprot_val(ref_prot) |= _PAGE_PRESENT;
522 __set_pmd_pte(kpte, address, mk_pte(base, ref_prot));
523 base = NULL; 518 base = NULL;
524 519
525out_unlock: 520out_unlock:
@@ -534,6 +529,36 @@ out_unlock:
534 return 0; 529 return 0;
535} 530}
536 531
532static int __cpa_process_fault(struct cpa_data *cpa, unsigned long vaddr,
533 int primary)
534{
535 /*
536 * Ignore all non primary paths.
537 */
538 if (!primary)
539 return 0;
540
541 /*
542 * Ignore the NULL PTE for kernel identity mapping, as it is expected
543 * to have holes.
544 * Also set numpages to '1' indicating that we processed cpa req for
545 * one virtual address page and its pfn. TBD: numpages can be set based
546 * on the initial value and the level returned by lookup_address().
547 */
548 if (within(vaddr, PAGE_OFFSET,
549 PAGE_OFFSET + (max_pfn_mapped << PAGE_SHIFT))) {
550 cpa->numpages = 1;
551 cpa->pfn = __pa(vaddr) >> PAGE_SHIFT;
552 return 0;
553 } else {
554 WARN(1, KERN_WARNING "CPA: called for zero pte. "
555 "vaddr = %lx cpa->vaddr = %lx\n", vaddr,
556 *cpa->vaddr);
557
558 return -EFAULT;
559 }
560}
561
537static int __change_page_attr(struct cpa_data *cpa, int primary) 562static int __change_page_attr(struct cpa_data *cpa, int primary)
538{ 563{
539 unsigned long address; 564 unsigned long address;
@@ -545,21 +570,14 @@ static int __change_page_attr(struct cpa_data *cpa, int primary)
545 address = cpa->vaddr[cpa->curpage]; 570 address = cpa->vaddr[cpa->curpage];
546 else 571 else
547 address = *cpa->vaddr; 572 address = *cpa->vaddr;
548
549repeat: 573repeat:
550 kpte = lookup_address(address, &level); 574 kpte = lookup_address(address, &level);
551 if (!kpte) 575 if (!kpte)
552 return 0; 576 return __cpa_process_fault(cpa, address, primary);
553 577
554 old_pte = *kpte; 578 old_pte = *kpte;
555 if (!pte_val(old_pte)) { 579 if (!pte_val(old_pte))
556 if (!primary) 580 return __cpa_process_fault(cpa, address, primary);
557 return 0;
558 WARN(1, KERN_WARNING "CPA: called for zero pte. "
559 "vaddr = %lx cpa->vaddr = %lx\n", address,
560 *cpa->vaddr);
561 return -EINVAL;
562 }
563 581
564 if (level == PG_LEVEL_4K) { 582 if (level == PG_LEVEL_4K) {
565 pte_t new_pte; 583 pte_t new_pte;
@@ -657,12 +675,7 @@ static int cpa_process_alias(struct cpa_data *cpa)
657 vaddr = *cpa->vaddr; 675 vaddr = *cpa->vaddr;
658 676
659 if (!(within(vaddr, PAGE_OFFSET, 677 if (!(within(vaddr, PAGE_OFFSET,
660 PAGE_OFFSET + (max_low_pfn_mapped << PAGE_SHIFT)) 678 PAGE_OFFSET + (max_pfn_mapped << PAGE_SHIFT)))) {
661#ifdef CONFIG_X86_64
662 || within(vaddr, PAGE_OFFSET + (1UL<<32),
663 PAGE_OFFSET + (max_pfn_mapped << PAGE_SHIFT))
664#endif
665 )) {
666 679
667 alias_cpa = *cpa; 680 alias_cpa = *cpa;
668 temp_cpa_vaddr = (unsigned long) __va(cpa->pfn << PAGE_SHIFT); 681 temp_cpa_vaddr = (unsigned long) __va(cpa->pfn << PAGE_SHIFT);
@@ -793,6 +806,13 @@ static int change_page_attr_set_clr(unsigned long *addr, int numpages,
793 806
794 vm_unmap_aliases(); 807 vm_unmap_aliases();
795 808
809 /*
810 * If we're called with lazy mmu updates enabled, the
811 * in-memory pte state may be stale. Flush pending updates to
812 * bring them up to date.
813 */
814 arch_flush_lazy_mmu_mode();
815
796 cpa.vaddr = addr; 816 cpa.vaddr = addr;
797 cpa.numpages = numpages; 817 cpa.numpages = numpages;
798 cpa.mask_set = mask_set; 818 cpa.mask_set = mask_set;
@@ -835,6 +855,13 @@ static int change_page_attr_set_clr(unsigned long *addr, int numpages,
835 } else 855 } else
836 cpa_flush_all(cache); 856 cpa_flush_all(cache);
837 857
858 /*
859 * If we've been called with lazy mmu updates enabled, then
860 * make sure that everything gets flushed out before we
861 * return.
862 */
863 arch_flush_lazy_mmu_mode();
864
838out: 865out:
839 return ret; 866 return ret;
840} 867}
diff --git a/arch/x86/mm/pat.c b/arch/x86/mm/pat.c
index 85cbd3cd3723..aebbf67a79d0 100644
--- a/arch/x86/mm/pat.c
+++ b/arch/x86/mm/pat.c
@@ -211,6 +211,33 @@ chk_conflict(struct memtype *new, struct memtype *entry, unsigned long *type)
211static struct memtype *cached_entry; 211static struct memtype *cached_entry;
212static u64 cached_start; 212static u64 cached_start;
213 213
214static int pat_pagerange_is_ram(unsigned long start, unsigned long end)
215{
216 int ram_page = 0, not_rampage = 0;
217 unsigned long page_nr;
218
219 for (page_nr = (start >> PAGE_SHIFT); page_nr < (end >> PAGE_SHIFT);
220 ++page_nr) {
221 /*
222 * For legacy reasons, physical address range in the legacy ISA
223 * region is tracked as non-RAM. This will allow users of
224 * /dev/mem to map portions of legacy ISA region, even when
225 * some of those portions are listed(or not even listed) with
226 * different e820 types(RAM/reserved/..)
227 */
228 if (page_nr >= (ISA_END_ADDRESS >> PAGE_SHIFT) &&
229 page_is_ram(page_nr))
230 ram_page = 1;
231 else
232 not_rampage = 1;
233
234 if (ram_page == not_rampage)
235 return -1;
236 }
237
238 return ram_page;
239}
240
214/* 241/*
215 * For RAM pages, mark the pages as non WB memory type using 242 * For RAM pages, mark the pages as non WB memory type using
216 * PageNonWB (PG_arch_1). We allow only one set_memory_uc() or 243 * PageNonWB (PG_arch_1). We allow only one set_memory_uc() or
@@ -333,9 +360,13 @@ int reserve_memtype(u64 start, u64 end, unsigned long req_type,
333 req_type & _PAGE_CACHE_MASK); 360 req_type & _PAGE_CACHE_MASK);
334 } 361 }
335 362
336 is_range_ram = pagerange_is_ram(start, end); 363 if (new_type)
364 *new_type = actual_type;
365
366 is_range_ram = pat_pagerange_is_ram(start, end);
337 if (is_range_ram == 1) 367 if (is_range_ram == 1)
338 return reserve_ram_pages_type(start, end, req_type, new_type); 368 return reserve_ram_pages_type(start, end, req_type,
369 new_type);
339 else if (is_range_ram < 0) 370 else if (is_range_ram < 0)
340 return -EINVAL; 371 return -EINVAL;
341 372
@@ -347,9 +378,6 @@ int reserve_memtype(u64 start, u64 end, unsigned long req_type,
347 new->end = end; 378 new->end = end;
348 new->type = actual_type; 379 new->type = actual_type;
349 380
350 if (new_type)
351 *new_type = actual_type;
352
353 spin_lock(&memtype_lock); 381 spin_lock(&memtype_lock);
354 382
355 if (cached_entry && start >= cached_start) 383 if (cached_entry && start >= cached_start)
@@ -437,7 +465,7 @@ int free_memtype(u64 start, u64 end)
437 if (is_ISA_range(start, end - 1)) 465 if (is_ISA_range(start, end - 1))
438 return 0; 466 return 0;
439 467
440 is_range_ram = pagerange_is_ram(start, end); 468 is_range_ram = pat_pagerange_is_ram(start, end);
441 if (is_range_ram == 1) 469 if (is_range_ram == 1)
442 return free_ram_pages_type(start, end); 470 return free_ram_pages_type(start, end);
443 else if (is_range_ram < 0) 471 else if (is_range_ram < 0)
@@ -601,39 +629,45 @@ void unmap_devmem(unsigned long pfn, unsigned long size, pgprot_t vma_prot)
601 * Reserved non RAM regions only and after successful reserve_memtype, 629 * Reserved non RAM regions only and after successful reserve_memtype,
602 * this func also keeps identity mapping (if any) in sync with this new prot. 630 * this func also keeps identity mapping (if any) in sync with this new prot.
603 */ 631 */
604static int reserve_pfn_range(u64 paddr, unsigned long size, pgprot_t vma_prot) 632static int reserve_pfn_range(u64 paddr, unsigned long size, pgprot_t *vma_prot,
633 int strict_prot)
605{ 634{
606 int is_ram = 0; 635 int is_ram = 0;
607 int id_sz, ret; 636 int id_sz, ret;
608 unsigned long flags; 637 unsigned long flags;
609 unsigned long want_flags = (pgprot_val(vma_prot) & _PAGE_CACHE_MASK); 638 unsigned long want_flags = (pgprot_val(*vma_prot) & _PAGE_CACHE_MASK);
610 639
611 is_ram = pagerange_is_ram(paddr, paddr + size); 640 is_ram = pat_pagerange_is_ram(paddr, paddr + size);
612 641
613 if (is_ram != 0) { 642 /*
614 /* 643 * reserve_pfn_range() doesn't support RAM pages.
615 * For mapping RAM pages, drivers need to call 644 */
616 * set_memory_[uc|wc|wb] directly, for reserve and free, before 645 if (is_ram != 0)
617 * setting up the PTE. 646 return -EINVAL;
618 */
619 WARN_ON_ONCE(1);
620 return 0;
621 }
622 647
623 ret = reserve_memtype(paddr, paddr + size, want_flags, &flags); 648 ret = reserve_memtype(paddr, paddr + size, want_flags, &flags);
624 if (ret) 649 if (ret)
625 return ret; 650 return ret;
626 651
627 if (flags != want_flags) { 652 if (flags != want_flags) {
628 free_memtype(paddr, paddr + size); 653 if (strict_prot || !is_new_memtype_allowed(want_flags, flags)) {
629 printk(KERN_ERR 654 free_memtype(paddr, paddr + size);
630 "%s:%d map pfn expected mapping type %s for %Lx-%Lx, got %s\n", 655 printk(KERN_ERR "%s:%d map pfn expected mapping type %s"
631 current->comm, current->pid, 656 " for %Lx-%Lx, got %s\n",
632 cattr_name(want_flags), 657 current->comm, current->pid,
633 (unsigned long long)paddr, 658 cattr_name(want_flags),
634 (unsigned long long)(paddr + size), 659 (unsigned long long)paddr,
635 cattr_name(flags)); 660 (unsigned long long)(paddr + size),
636 return -EINVAL; 661 cattr_name(flags));
662 return -EINVAL;
663 }
664 /*
665 * We allow returning different type than the one requested in
666 * non strict case.
667 */
668 *vma_prot = __pgprot((pgprot_val(*vma_prot) &
669 (~_PAGE_CACHE_MASK)) |
670 flags);
637 } 671 }
638 672
639 /* Need to keep identity mapping in sync */ 673 /* Need to keep identity mapping in sync */
@@ -666,7 +700,7 @@ static void free_pfn_range(u64 paddr, unsigned long size)
666{ 700{
667 int is_ram; 701 int is_ram;
668 702
669 is_ram = pagerange_is_ram(paddr, paddr + size); 703 is_ram = pat_pagerange_is_ram(paddr, paddr + size);
670 if (is_ram == 0) 704 if (is_ram == 0)
671 free_memtype(paddr, paddr + size); 705 free_memtype(paddr, paddr + size);
672} 706}
@@ -689,6 +723,7 @@ int track_pfn_vma_copy(struct vm_area_struct *vma)
689 unsigned long vma_start = vma->vm_start; 723 unsigned long vma_start = vma->vm_start;
690 unsigned long vma_end = vma->vm_end; 724 unsigned long vma_end = vma->vm_end;
691 unsigned long vma_size = vma_end - vma_start; 725 unsigned long vma_size = vma_end - vma_start;
726 pgprot_t pgprot;
692 727
693 if (!pat_enabled) 728 if (!pat_enabled)
694 return 0; 729 return 0;
@@ -702,7 +737,8 @@ int track_pfn_vma_copy(struct vm_area_struct *vma)
702 WARN_ON_ONCE(1); 737 WARN_ON_ONCE(1);
703 return -EINVAL; 738 return -EINVAL;
704 } 739 }
705 return reserve_pfn_range(paddr, vma_size, __pgprot(prot)); 740 pgprot = __pgprot(prot);
741 return reserve_pfn_range(paddr, vma_size, &pgprot, 1);
706 } 742 }
707 743
708 /* reserve entire vma page by page, using pfn and prot from pte */ 744 /* reserve entire vma page by page, using pfn and prot from pte */
@@ -710,7 +746,8 @@ int track_pfn_vma_copy(struct vm_area_struct *vma)
710 if (follow_phys(vma, vma_start + i, 0, &prot, &paddr)) 746 if (follow_phys(vma, vma_start + i, 0, &prot, &paddr))
711 continue; 747 continue;
712 748
713 retval = reserve_pfn_range(paddr, PAGE_SIZE, __pgprot(prot)); 749 pgprot = __pgprot(prot);
750 retval = reserve_pfn_range(paddr, PAGE_SIZE, &pgprot, 1);
714 if (retval) 751 if (retval)
715 goto cleanup_ret; 752 goto cleanup_ret;
716 } 753 }
@@ -741,7 +778,7 @@ cleanup_ret:
741 * Note that this function can be called with caller trying to map only a 778 * Note that this function can be called with caller trying to map only a
742 * subrange/page inside the vma. 779 * subrange/page inside the vma.
743 */ 780 */
744int track_pfn_vma_new(struct vm_area_struct *vma, pgprot_t prot, 781int track_pfn_vma_new(struct vm_area_struct *vma, pgprot_t *prot,
745 unsigned long pfn, unsigned long size) 782 unsigned long pfn, unsigned long size)
746{ 783{
747 int retval = 0; 784 int retval = 0;
@@ -758,14 +795,14 @@ int track_pfn_vma_new(struct vm_area_struct *vma, pgprot_t prot,
758 if (is_linear_pfn_mapping(vma)) { 795 if (is_linear_pfn_mapping(vma)) {
759 /* reserve the whole chunk starting from vm_pgoff */ 796 /* reserve the whole chunk starting from vm_pgoff */
760 paddr = (resource_size_t)vma->vm_pgoff << PAGE_SHIFT; 797 paddr = (resource_size_t)vma->vm_pgoff << PAGE_SHIFT;
761 return reserve_pfn_range(paddr, vma_size, prot); 798 return reserve_pfn_range(paddr, vma_size, prot, 0);
762 } 799 }
763 800
764 /* reserve page by page using pfn and size */ 801 /* reserve page by page using pfn and size */
765 base_paddr = (resource_size_t)pfn << PAGE_SHIFT; 802 base_paddr = (resource_size_t)pfn << PAGE_SHIFT;
766 for (i = 0; i < size; i += PAGE_SIZE) { 803 for (i = 0; i < size; i += PAGE_SIZE) {
767 paddr = base_paddr + i; 804 paddr = base_paddr + i;
768 retval = reserve_pfn_range(paddr, PAGE_SIZE, prot); 805 retval = reserve_pfn_range(paddr, PAGE_SIZE, prot, 0);
769 if (retval) 806 if (retval)
770 goto cleanup_ret; 807 goto cleanup_ret;
771 } 808 }
diff --git a/arch/x86/pci/i386.c b/arch/x86/pci/i386.c
index f884740da318..5ead808dd70c 100644
--- a/arch/x86/pci/i386.c
+++ b/arch/x86/pci/i386.c
@@ -314,17 +314,7 @@ int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
314 return retval; 314 return retval;
315 315
316 if (flags != new_flags) { 316 if (flags != new_flags) {
317 /* 317 if (!is_new_memtype_allowed(flags, new_flags)) {
318 * Do not fallback to certain memory types with certain
319 * requested type:
320 * - request is uncached, return cannot be write-back
321 * - request is uncached, return cannot be write-combine
322 * - request is write-combine, return cannot be write-back
323 */
324 if ((flags == _PAGE_CACHE_UC_MINUS &&
325 (new_flags == _PAGE_CACHE_WB)) ||
326 (flags == _PAGE_CACHE_WC &&
327 new_flags == _PAGE_CACHE_WB)) {
328 free_memtype(addr, addr+len); 318 free_memtype(addr, addr+len);
329 return -EINVAL; 319 return -EINVAL;
330 } 320 }
diff --git a/arch/x86/pci/irq.c b/arch/x86/pci/irq.c
index 4064345cf144..fecbce6e7d7c 100644
--- a/arch/x86/pci/irq.c
+++ b/arch/x86/pci/irq.c
@@ -572,6 +572,7 @@ static __init int intel_router_probe(struct irq_router *r, struct pci_dev *route
572 case PCI_DEVICE_ID_INTEL_ICH7_1: 572 case PCI_DEVICE_ID_INTEL_ICH7_1:
573 case PCI_DEVICE_ID_INTEL_ICH7_30: 573 case PCI_DEVICE_ID_INTEL_ICH7_30:
574 case PCI_DEVICE_ID_INTEL_ICH7_31: 574 case PCI_DEVICE_ID_INTEL_ICH7_31:
575 case PCI_DEVICE_ID_INTEL_TGP_LPC:
575 case PCI_DEVICE_ID_INTEL_ESB2_0: 576 case PCI_DEVICE_ID_INTEL_ESB2_0:
576 case PCI_DEVICE_ID_INTEL_ICH8_0: 577 case PCI_DEVICE_ID_INTEL_ICH8_0:
577 case PCI_DEVICE_ID_INTEL_ICH8_1: 578 case PCI_DEVICE_ID_INTEL_ICH8_1:
diff --git a/arch/x86/scripts/strip-symbols b/arch/x86/scripts/strip-symbols
deleted file mode 100644
index a2f1ccb827c7..000000000000
--- a/arch/x86/scripts/strip-symbols
+++ /dev/null
@@ -1 +0,0 @@
1__cpu_vendor_dev_X86_VENDOR_*
diff --git a/arch/x86/xen/multicalls.h b/arch/x86/xen/multicalls.h
index 858938241616..fa3e10725d98 100644
--- a/arch/x86/xen/multicalls.h
+++ b/arch/x86/xen/multicalls.h
@@ -19,8 +19,10 @@ DECLARE_PER_CPU(unsigned long, xen_mc_irq_flags);
19 paired with xen_mc_issue() */ 19 paired with xen_mc_issue() */
20static inline void xen_mc_batch(void) 20static inline void xen_mc_batch(void)
21{ 21{
22 unsigned long flags;
22 /* need to disable interrupts until this entry is complete */ 23 /* need to disable interrupts until this entry is complete */
23 local_irq_save(__get_cpu_var(xen_mc_irq_flags)); 24 local_irq_save(flags);
25 __get_cpu_var(xen_mc_irq_flags) = flags;
24} 26}
25 27
26static inline struct multicall_space xen_mc_entry(size_t args) 28static inline struct multicall_space xen_mc_entry(size_t args)
generated by cgit v1.2.2 (git 2.25.0) at 2024-12-02 06:43:22 -0500