aboutsummaryrefslogtreecommitdiffstats
path: root/arch/x86
diff options
context:
space:
mode:
Diffstat (limited to 'arch/x86')
-rw-r--r--arch/x86/Makefile9
-rw-r--r--arch/x86/boot/Makefile4
-rw-r--r--arch/x86/boot/compressed/misc.c2
-rw-r--r--arch/x86/include/asm/hpet.h1
-rw-r--r--arch/x86/include/asm/hugetlb.h1
-rw-r--r--arch/x86/include/asm/kvm_host.h2
-rw-r--r--arch/x86/include/asm/page_64_types.h2
-rw-r--r--arch/x86/include/uapi/asm/msr-index.h2
-rw-r--r--arch/x86/kernel/acpi/sleep.c2
-rw-r--r--arch/x86/kernel/apic/io_apic.c7
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce.c4
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce_intel.c18
-rw-r--r--arch/x86/kernel/cpu/mcheck/therm_throt.c4
-rw-r--r--arch/x86/kernel/cpu/mcheck/threshold.c4
-rw-r--r--arch/x86/kernel/cpu/perf_event_intel.c1
-rw-r--r--arch/x86/kernel/cpu/perf_event_intel_rapl.c46
-rw-r--r--arch/x86/kernel/cpu/rdrand.c1
-rw-r--r--arch/x86/kernel/early-quirks.c18
-rw-r--r--arch/x86/kernel/head32.c2
-rw-r--r--arch/x86/kernel/head64.c2
-rw-r--r--arch/x86/kernel/hpet.c2
-rw-r--r--arch/x86/kernel/kprobes/core.c16
-rw-r--r--arch/x86/kernel/ldt.c4
-rw-r--r--arch/x86/kernel/process_64.c2
-rw-r--r--arch/x86/kernel/reboot.c82
-rw-r--r--arch/x86/kernel/smp.c2
-rw-r--r--arch/x86/kernel/traps.c6
-rw-r--r--arch/x86/kernel/vsmp_64.c17
-rw-r--r--arch/x86/kernel/vsyscall_gtod.c2
-rw-r--r--arch/x86/kvm/cpuid.c2
-rw-r--r--arch/x86/kvm/cpuid.h8
-rw-r--r--arch/x86/kvm/mmu.c38
-rw-r--r--arch/x86/kvm/mmu.h44
-rw-r--r--arch/x86/kvm/paging_tmpl.h2
-rw-r--r--arch/x86/kvm/vmx.c71
-rw-r--r--arch/x86/kvm/x86.c18
-rw-r--r--arch/x86/lguest/boot.c4
-rw-r--r--arch/x86/lib/msr.c2
-rw-r--r--arch/x86/math-emu/errors.c16
-rw-r--r--arch/x86/net/bpf_jit_comp.c2
-rw-r--r--arch/x86/platform/efi/early_printk.c83
-rw-r--r--arch/x86/platform/olpc/olpc-xo1-pm.c2
-rw-r--r--arch/x86/power/hibernate_64.c2
-rw-r--r--arch/x86/syscalls/Makefile2
-rw-r--r--arch/x86/syscalls/syscall_32.tbl1
-rw-r--r--arch/x86/tools/Makefile2
-rw-r--r--arch/x86/vdso/vdso-layout.lds.S19
-rw-r--r--arch/x86/vdso/vdso32-setup.c8
-rw-r--r--arch/x86/xen/enlighten.c2
-rw-r--r--arch/x86/xen/irq.c6
-rw-r--r--arch/x86/xen/smp.c3
-rw-r--r--arch/x86/xen/spinlock.c5
-rw-r--r--arch/x86/xen/xen-asm_32.S25
53 files changed, 443 insertions, 189 deletions
diff --git a/arch/x86/Makefile b/arch/x86/Makefile
index 602f57e590b5..33f71b01fd22 100644
--- a/arch/x86/Makefile
+++ b/arch/x86/Makefile
@@ -79,11 +79,14 @@ else
79 UTS_MACHINE := x86_64 79 UTS_MACHINE := x86_64
80 CHECKFLAGS += -D__x86_64__ -m64 80 CHECKFLAGS += -D__x86_64__ -m64
81 81
82 biarch := -m64
82 KBUILD_AFLAGS += -m64 83 KBUILD_AFLAGS += -m64
83 KBUILD_CFLAGS += -m64 84 KBUILD_CFLAGS += -m64
84 85
85 # Don't autogenerate traditional x87, MMX or SSE instructions 86 # Don't autogenerate traditional x87, MMX or SSE instructions
86 KBUILD_CFLAGS += -mno-mmx -mno-sse -mno-80387 -mno-fp-ret-in-387 87 KBUILD_CFLAGS += -mno-mmx -mno-sse
88 KBUILD_CFLAGS += $(call cc-option,-mno-80387)
89 KBUILD_CFLAGS += $(call cc-option,-mno-fp-ret-in-387)
87 90
88 # Use -mpreferred-stack-boundary=3 if supported. 91 # Use -mpreferred-stack-boundary=3 if supported.
89 KBUILD_CFLAGS += $(call cc-option,-mpreferred-stack-boundary=3) 92 KBUILD_CFLAGS += $(call cc-option,-mpreferred-stack-boundary=3)
@@ -250,8 +253,8 @@ archclean:
250PHONY += kvmconfig 253PHONY += kvmconfig
251kvmconfig: 254kvmconfig:
252 $(if $(wildcard $(objtree)/.config),, $(error You need an existing .config for this target)) 255 $(if $(wildcard $(objtree)/.config),, $(error You need an existing .config for this target))
253 $(Q)$(CONFIG_SHELL) $(srctree)/scripts/kconfig/merge_config.sh -m -O $(objtree) $(objtree)/.config arch/x86/configs/kvm_guest.config 256 $(Q)$(CONFIG_SHELL) $(srctree)/scripts/kconfig/merge_config.sh -m -O $(objtree) $(objtree)/.config $(srctree)/arch/x86/configs/kvm_guest.config
254 $(Q)yes "" | $(MAKE) oldconfig 257 $(Q)yes "" | $(MAKE) -f $(srctree)/Makefile oldconfig
255 258
256define archhelp 259define archhelp
257 echo '* bzImage - Compressed kernel image (arch/x86/boot/bzImage)' 260 echo '* bzImage - Compressed kernel image (arch/x86/boot/bzImage)'
diff --git a/arch/x86/boot/Makefile b/arch/x86/boot/Makefile
index abb9eba61b50..dbe8dd2fe247 100644
--- a/arch/x86/boot/Makefile
+++ b/arch/x86/boot/Makefile
@@ -71,7 +71,7 @@ $(obj)/vmlinux.bin: $(obj)/compressed/vmlinux FORCE
71 71
72SETUP_OBJS = $(addprefix $(obj)/,$(setup-y)) 72SETUP_OBJS = $(addprefix $(obj)/,$(setup-y))
73 73
74sed-voffset := -e 's/^\([0-9a-fA-F]*\) . \(_text\|_end\)$$/\#define VO_\2 0x\1/p' 74sed-voffset := -e 's/^\([0-9a-fA-F]*\) [ABCDGRSTVW] \(_text\|_end\)$$/\#define VO_\2 0x\1/p'
75 75
76quiet_cmd_voffset = VOFFSET $@ 76quiet_cmd_voffset = VOFFSET $@
77 cmd_voffset = $(NM) $< | sed -n $(sed-voffset) > $@ 77 cmd_voffset = $(NM) $< | sed -n $(sed-voffset) > $@
@@ -80,7 +80,7 @@ targets += voffset.h
80$(obj)/voffset.h: vmlinux FORCE 80$(obj)/voffset.h: vmlinux FORCE
81 $(call if_changed,voffset) 81 $(call if_changed,voffset)
82 82
83sed-zoffset := -e 's/^\([0-9a-fA-F]*\) . \(startup_32\|startup_64\|efi32_stub_entry\|efi64_stub_entry\|efi_pe_entry\|input_data\|_end\|z_.*\)$$/\#define ZO_\2 0x\1/p' 83sed-zoffset := -e 's/^\([0-9a-fA-F]*\) [ABCDGRSTVW] \(startup_32\|startup_64\|efi32_stub_entry\|efi64_stub_entry\|efi_pe_entry\|input_data\|_end\|z_.*\)$$/\#define ZO_\2 0x\1/p'
84 84
85quiet_cmd_zoffset = ZOFFSET $@ 85quiet_cmd_zoffset = ZOFFSET $@
86 cmd_zoffset = $(NM) $< | sed -n $(sed-zoffset) > $@ 86 cmd_zoffset = $(NM) $< | sed -n $(sed-zoffset) > $@
diff --git a/arch/x86/boot/compressed/misc.c b/arch/x86/boot/compressed/misc.c
index 17684615374b..57ab74df7eea 100644
--- a/arch/x86/boot/compressed/misc.c
+++ b/arch/x86/boot/compressed/misc.c
@@ -354,7 +354,7 @@ static void parse_elf(void *output)
354 free(phdrs); 354 free(phdrs);
355} 355}
356 356
357asmlinkage void *decompress_kernel(void *rmode, memptr heap, 357asmlinkage __visible void *decompress_kernel(void *rmode, memptr heap,
358 unsigned char *input_data, 358 unsigned char *input_data,
359 unsigned long input_len, 359 unsigned long input_len,
360 unsigned char *output, 360 unsigned char *output,
diff --git a/arch/x86/include/asm/hpet.h b/arch/x86/include/asm/hpet.h
index b18df579c0e9..36f7125945e3 100644
--- a/arch/x86/include/asm/hpet.h
+++ b/arch/x86/include/asm/hpet.h
@@ -63,6 +63,7 @@
63/* hpet memory map physical address */ 63/* hpet memory map physical address */
64extern unsigned long hpet_address; 64extern unsigned long hpet_address;
65extern unsigned long force_hpet_address; 65extern unsigned long force_hpet_address;
66extern int boot_hpet_disable;
66extern u8 hpet_blockid; 67extern u8 hpet_blockid;
67extern int hpet_force_user; 68extern int hpet_force_user;
68extern u8 hpet_msi_disable; 69extern u8 hpet_msi_disable;
diff --git a/arch/x86/include/asm/hugetlb.h b/arch/x86/include/asm/hugetlb.h
index a8091216963b..68c05398bba9 100644
--- a/arch/x86/include/asm/hugetlb.h
+++ b/arch/x86/include/asm/hugetlb.h
@@ -52,6 +52,7 @@ static inline pte_t huge_ptep_get_and_clear(struct mm_struct *mm,
52static inline void huge_ptep_clear_flush(struct vm_area_struct *vma, 52static inline void huge_ptep_clear_flush(struct vm_area_struct *vma,
53 unsigned long addr, pte_t *ptep) 53 unsigned long addr, pte_t *ptep)
54{ 54{
55 ptep_clear_flush(vma, addr, ptep);
55} 56}
56 57
57static inline int huge_pte_none(pte_t pte) 58static inline int huge_pte_none(pte_t pte)
diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index fcaf9c961265..7de069afb382 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -60,7 +60,7 @@
60 | X86_CR4_PSE | X86_CR4_PAE | X86_CR4_MCE \ 60 | X86_CR4_PSE | X86_CR4_PAE | X86_CR4_MCE \
61 | X86_CR4_PGE | X86_CR4_PCE | X86_CR4_OSFXSR | X86_CR4_PCIDE \ 61 | X86_CR4_PGE | X86_CR4_PCE | X86_CR4_OSFXSR | X86_CR4_PCIDE \
62 | X86_CR4_OSXSAVE | X86_CR4_SMEP | X86_CR4_FSGSBASE \ 62 | X86_CR4_OSXSAVE | X86_CR4_SMEP | X86_CR4_FSGSBASE \
63 | X86_CR4_OSXMMEXCPT | X86_CR4_VMXE)) 63 | X86_CR4_OSXMMEXCPT | X86_CR4_VMXE | X86_CR4_SMAP))
64 64
65#define CR8_RESERVED_BITS (~(unsigned long)X86_CR8_TPR) 65#define CR8_RESERVED_BITS (~(unsigned long)X86_CR8_TPR)
66 66
diff --git a/arch/x86/include/asm/page_64_types.h b/arch/x86/include/asm/page_64_types.h
index 8de6d9cf3b95..678205195ae1 100644
--- a/arch/x86/include/asm/page_64_types.h
+++ b/arch/x86/include/asm/page_64_types.h
@@ -1,7 +1,7 @@
1#ifndef _ASM_X86_PAGE_64_DEFS_H 1#ifndef _ASM_X86_PAGE_64_DEFS_H
2#define _ASM_X86_PAGE_64_DEFS_H 2#define _ASM_X86_PAGE_64_DEFS_H
3 3
4#define THREAD_SIZE_ORDER 1 4#define THREAD_SIZE_ORDER 2
5#define THREAD_SIZE (PAGE_SIZE << THREAD_SIZE_ORDER) 5#define THREAD_SIZE (PAGE_SIZE << THREAD_SIZE_ORDER)
6#define CURRENT_MASK (~(THREAD_SIZE - 1)) 6#define CURRENT_MASK (~(THREAD_SIZE - 1))
7 7
diff --git a/arch/x86/include/uapi/asm/msr-index.h b/arch/x86/include/uapi/asm/msr-index.h
index c827ace3121b..fcf2b3ae1bf0 100644
--- a/arch/x86/include/uapi/asm/msr-index.h
+++ b/arch/x86/include/uapi/asm/msr-index.h
@@ -384,7 +384,7 @@
384#define MSR_IA32_MISC_ENABLE_MWAIT_BIT 18 384#define MSR_IA32_MISC_ENABLE_MWAIT_BIT 18
385#define MSR_IA32_MISC_ENABLE_MWAIT (1ULL << MSR_IA32_MISC_ENABLE_MWAIT_BIT) 385#define MSR_IA32_MISC_ENABLE_MWAIT (1ULL << MSR_IA32_MISC_ENABLE_MWAIT_BIT)
386#define MSR_IA32_MISC_ENABLE_LIMIT_CPUID_BIT 22 386#define MSR_IA32_MISC_ENABLE_LIMIT_CPUID_BIT 22
387#define MSR_IA32_MISC_ENABLE_LIMIT_CPUID (1ULL << MSR_IA32_MISC_ENABLE_LIMIT_CPUID_BIT); 387#define MSR_IA32_MISC_ENABLE_LIMIT_CPUID (1ULL << MSR_IA32_MISC_ENABLE_LIMIT_CPUID_BIT)
388#define MSR_IA32_MISC_ENABLE_XTPR_DISABLE_BIT 23 388#define MSR_IA32_MISC_ENABLE_XTPR_DISABLE_BIT 23
389#define MSR_IA32_MISC_ENABLE_XTPR_DISABLE (1ULL << MSR_IA32_MISC_ENABLE_XTPR_DISABLE_BIT) 389#define MSR_IA32_MISC_ENABLE_XTPR_DISABLE (1ULL << MSR_IA32_MISC_ENABLE_XTPR_DISABLE_BIT)
390#define MSR_IA32_MISC_ENABLE_XD_DISABLE_BIT 34 390#define MSR_IA32_MISC_ENABLE_XD_DISABLE_BIT 34
diff --git a/arch/x86/kernel/acpi/sleep.c b/arch/x86/kernel/acpi/sleep.c
index 3a2ae4c88948..31368207837c 100644
--- a/arch/x86/kernel/acpi/sleep.c
+++ b/arch/x86/kernel/acpi/sleep.c
@@ -31,7 +31,7 @@ static char temp_stack[4096];
31 * 31 *
32 * Wrapper around acpi_enter_sleep_state() to be called by assmebly. 32 * Wrapper around acpi_enter_sleep_state() to be called by assmebly.
33 */ 33 */
34acpi_status asmlinkage x86_acpi_enter_sleep_state(u8 state) 34acpi_status asmlinkage __visible x86_acpi_enter_sleep_state(u8 state)
35{ 35{
36 return acpi_enter_sleep_state(state); 36 return acpi_enter_sleep_state(state);
37} 37}
diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
index 6ad4658de705..992060e09897 100644
--- a/arch/x86/kernel/apic/io_apic.c
+++ b/arch/x86/kernel/apic/io_apic.c
@@ -2189,7 +2189,7 @@ void send_cleanup_vector(struct irq_cfg *cfg)
2189 cfg->move_in_progress = 0; 2189 cfg->move_in_progress = 0;
2190} 2190}
2191 2191
2192asmlinkage void smp_irq_move_cleanup_interrupt(void) 2192asmlinkage __visible void smp_irq_move_cleanup_interrupt(void)
2193{ 2193{
2194 unsigned vector, me; 2194 unsigned vector, me;
2195 2195
@@ -3425,6 +3425,11 @@ int get_nr_irqs_gsi(void)
3425 return nr_irqs_gsi; 3425 return nr_irqs_gsi;
3426} 3426}
3427 3427
3428unsigned int arch_dynirq_lower_bound(unsigned int from)
3429{
3430 return from < nr_irqs_gsi ? nr_irqs_gsi : from;
3431}
3432
3428int __init arch_probe_nr_irqs(void) 3433int __init arch_probe_nr_irqs(void)
3429{ 3434{
3430 int nr; 3435 int nr;
diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c
index eeee23ff75ef..68317c80de7f 100644
--- a/arch/x86/kernel/cpu/mcheck/mce.c
+++ b/arch/x86/kernel/cpu/mcheck/mce.c
@@ -598,7 +598,6 @@ void machine_check_poll(enum mcp_flags flags, mce_banks_t *b)
598{ 598{
599 struct mce m; 599 struct mce m;
600 int i; 600 int i;
601 unsigned long *v;
602 601
603 this_cpu_inc(mce_poll_count); 602 this_cpu_inc(mce_poll_count);
604 603
@@ -618,8 +617,7 @@ void machine_check_poll(enum mcp_flags flags, mce_banks_t *b)
618 if (!(m.status & MCI_STATUS_VAL)) 617 if (!(m.status & MCI_STATUS_VAL))
619 continue; 618 continue;
620 619
621 v = &get_cpu_var(mce_polled_error); 620 this_cpu_write(mce_polled_error, 1);
622 set_bit(0, v);
623 /* 621 /*
624 * Uncorrected or signalled events are handled by the exception 622 * Uncorrected or signalled events are handled by the exception
625 * handler when it is enabled, so don't process those here. 623 * handler when it is enabled, so don't process those here.
diff --git a/arch/x86/kernel/cpu/mcheck/mce_intel.c b/arch/x86/kernel/cpu/mcheck/mce_intel.c
index 3bdb95ae8c43..9a316b21df8b 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_intel.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_intel.c
@@ -42,7 +42,7 @@ static DEFINE_PER_CPU(mce_banks_t, mce_banks_owned);
42 * cmci_discover_lock protects against parallel discovery attempts 42 * cmci_discover_lock protects against parallel discovery attempts
43 * which could race against each other. 43 * which could race against each other.
44 */ 44 */
45static DEFINE_RAW_SPINLOCK(cmci_discover_lock); 45static DEFINE_SPINLOCK(cmci_discover_lock);
46 46
47#define CMCI_THRESHOLD 1 47#define CMCI_THRESHOLD 1
48#define CMCI_POLL_INTERVAL (30 * HZ) 48#define CMCI_POLL_INTERVAL (30 * HZ)
@@ -144,14 +144,14 @@ static void cmci_storm_disable_banks(void)
144 int bank; 144 int bank;
145 u64 val; 145 u64 val;
146 146
147 raw_spin_lock_irqsave(&cmci_discover_lock, flags); 147 spin_lock_irqsave(&cmci_discover_lock, flags);
148 owned = __get_cpu_var(mce_banks_owned); 148 owned = __get_cpu_var(mce_banks_owned);
149 for_each_set_bit(bank, owned, MAX_NR_BANKS) { 149 for_each_set_bit(bank, owned, MAX_NR_BANKS) {
150 rdmsrl(MSR_IA32_MCx_CTL2(bank), val); 150 rdmsrl(MSR_IA32_MCx_CTL2(bank), val);
151 val &= ~MCI_CTL2_CMCI_EN; 151 val &= ~MCI_CTL2_CMCI_EN;
152 wrmsrl(MSR_IA32_MCx_CTL2(bank), val); 152 wrmsrl(MSR_IA32_MCx_CTL2(bank), val);
153 } 153 }
154 raw_spin_unlock_irqrestore(&cmci_discover_lock, flags); 154 spin_unlock_irqrestore(&cmci_discover_lock, flags);
155} 155}
156 156
157static bool cmci_storm_detect(void) 157static bool cmci_storm_detect(void)
@@ -211,7 +211,7 @@ static void cmci_discover(int banks)
211 int i; 211 int i;
212 int bios_wrong_thresh = 0; 212 int bios_wrong_thresh = 0;
213 213
214 raw_spin_lock_irqsave(&cmci_discover_lock, flags); 214 spin_lock_irqsave(&cmci_discover_lock, flags);
215 for (i = 0; i < banks; i++) { 215 for (i = 0; i < banks; i++) {
216 u64 val; 216 u64 val;
217 int bios_zero_thresh = 0; 217 int bios_zero_thresh = 0;
@@ -266,7 +266,7 @@ static void cmci_discover(int banks)
266 WARN_ON(!test_bit(i, __get_cpu_var(mce_poll_banks))); 266 WARN_ON(!test_bit(i, __get_cpu_var(mce_poll_banks)));
267 } 267 }
268 } 268 }
269 raw_spin_unlock_irqrestore(&cmci_discover_lock, flags); 269 spin_unlock_irqrestore(&cmci_discover_lock, flags);
270 if (mca_cfg.bios_cmci_threshold && bios_wrong_thresh) { 270 if (mca_cfg.bios_cmci_threshold && bios_wrong_thresh) {
271 pr_info_once( 271 pr_info_once(
272 "bios_cmci_threshold: Some banks do not have valid thresholds set\n"); 272 "bios_cmci_threshold: Some banks do not have valid thresholds set\n");
@@ -316,10 +316,10 @@ void cmci_clear(void)
316 316
317 if (!cmci_supported(&banks)) 317 if (!cmci_supported(&banks))
318 return; 318 return;
319 raw_spin_lock_irqsave(&cmci_discover_lock, flags); 319 spin_lock_irqsave(&cmci_discover_lock, flags);
320 for (i = 0; i < banks; i++) 320 for (i = 0; i < banks; i++)
321 __cmci_disable_bank(i); 321 __cmci_disable_bank(i);
322 raw_spin_unlock_irqrestore(&cmci_discover_lock, flags); 322 spin_unlock_irqrestore(&cmci_discover_lock, flags);
323} 323}
324 324
325static void cmci_rediscover_work_func(void *arg) 325static void cmci_rediscover_work_func(void *arg)
@@ -360,9 +360,9 @@ void cmci_disable_bank(int bank)
360 if (!cmci_supported(&banks)) 360 if (!cmci_supported(&banks))
361 return; 361 return;
362 362
363 raw_spin_lock_irqsave(&cmci_discover_lock, flags); 363 spin_lock_irqsave(&cmci_discover_lock, flags);
364 __cmci_disable_bank(bank); 364 __cmci_disable_bank(bank);
365 raw_spin_unlock_irqrestore(&cmci_discover_lock, flags); 365 spin_unlock_irqrestore(&cmci_discover_lock, flags);
366} 366}
367 367
368static void intel_init_cmci(void) 368static void intel_init_cmci(void)
diff --git a/arch/x86/kernel/cpu/mcheck/therm_throt.c b/arch/x86/kernel/cpu/mcheck/therm_throt.c
index d921b7ee6595..36a1bb6d1ee0 100644
--- a/arch/x86/kernel/cpu/mcheck/therm_throt.c
+++ b/arch/x86/kernel/cpu/mcheck/therm_throt.c
@@ -429,14 +429,14 @@ static inline void __smp_thermal_interrupt(void)
429 smp_thermal_vector(); 429 smp_thermal_vector();
430} 430}
431 431
432asmlinkage void smp_thermal_interrupt(struct pt_regs *regs) 432asmlinkage __visible void smp_thermal_interrupt(struct pt_regs *regs)
433{ 433{
434 entering_irq(); 434 entering_irq();
435 __smp_thermal_interrupt(); 435 __smp_thermal_interrupt();
436 exiting_ack_irq(); 436 exiting_ack_irq();
437} 437}
438 438
439asmlinkage void smp_trace_thermal_interrupt(struct pt_regs *regs) 439asmlinkage __visible void smp_trace_thermal_interrupt(struct pt_regs *regs)
440{ 440{
441 entering_irq(); 441 entering_irq();
442 trace_thermal_apic_entry(THERMAL_APIC_VECTOR); 442 trace_thermal_apic_entry(THERMAL_APIC_VECTOR);
diff --git a/arch/x86/kernel/cpu/mcheck/threshold.c b/arch/x86/kernel/cpu/mcheck/threshold.c
index fe6b1c86645b..7245980186ee 100644
--- a/arch/x86/kernel/cpu/mcheck/threshold.c
+++ b/arch/x86/kernel/cpu/mcheck/threshold.c
@@ -24,14 +24,14 @@ static inline void __smp_threshold_interrupt(void)
24 mce_threshold_vector(); 24 mce_threshold_vector();
25} 25}
26 26
27asmlinkage void smp_threshold_interrupt(void) 27asmlinkage __visible void smp_threshold_interrupt(void)
28{ 28{
29 entering_irq(); 29 entering_irq();
30 __smp_threshold_interrupt(); 30 __smp_threshold_interrupt();
31 exiting_ack_irq(); 31 exiting_ack_irq();
32} 32}
33 33
34asmlinkage void smp_trace_threshold_interrupt(void) 34asmlinkage __visible void smp_trace_threshold_interrupt(void)
35{ 35{
36 entering_irq(); 36 entering_irq();
37 trace_threshold_apic_entry(THRESHOLD_APIC_VECTOR); 37 trace_threshold_apic_entry(THRESHOLD_APIC_VECTOR);
diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c
index aa333d966886..adb02aa62af5 100644
--- a/arch/x86/kernel/cpu/perf_event_intel.c
+++ b/arch/x86/kernel/cpu/perf_event_intel.c
@@ -169,7 +169,6 @@ static struct event_constraint intel_slm_event_constraints[] __read_mostly =
169{ 169{
170 FIXED_EVENT_CONSTRAINT(0x00c0, 0), /* INST_RETIRED.ANY */ 170 FIXED_EVENT_CONSTRAINT(0x00c0, 0), /* INST_RETIRED.ANY */
171 FIXED_EVENT_CONSTRAINT(0x003c, 1), /* CPU_CLK_UNHALTED.CORE */ 171 FIXED_EVENT_CONSTRAINT(0x003c, 1), /* CPU_CLK_UNHALTED.CORE */
172 FIXED_EVENT_CONSTRAINT(0x013c, 2), /* CPU_CLK_UNHALTED.REF */
173 FIXED_EVENT_CONSTRAINT(0x0300, 2), /* pseudo CPU_CLK_UNHALTED.REF */ 172 FIXED_EVENT_CONSTRAINT(0x0300, 2), /* pseudo CPU_CLK_UNHALTED.REF */
174 EVENT_CONSTRAINT_END 173 EVENT_CONSTRAINT_END
175}; 174};
diff --git a/arch/x86/kernel/cpu/perf_event_intel_rapl.c b/arch/x86/kernel/cpu/perf_event_intel_rapl.c
index 059218ed5208..619f7699487a 100644
--- a/arch/x86/kernel/cpu/perf_event_intel_rapl.c
+++ b/arch/x86/kernel/cpu/perf_event_intel_rapl.c
@@ -59,7 +59,7 @@
59#define INTEL_RAPL_PKG 0x2 /* pseudo-encoding */ 59#define INTEL_RAPL_PKG 0x2 /* pseudo-encoding */
60#define RAPL_IDX_RAM_NRG_STAT 2 /* DRAM */ 60#define RAPL_IDX_RAM_NRG_STAT 2 /* DRAM */
61#define INTEL_RAPL_RAM 0x3 /* pseudo-encoding */ 61#define INTEL_RAPL_RAM 0x3 /* pseudo-encoding */
62#define RAPL_IDX_PP1_NRG_STAT 3 /* DRAM */ 62#define RAPL_IDX_PP1_NRG_STAT 3 /* gpu */
63#define INTEL_RAPL_PP1 0x4 /* pseudo-encoding */ 63#define INTEL_RAPL_PP1 0x4 /* pseudo-encoding */
64 64
65/* Clients have PP0, PKG */ 65/* Clients have PP0, PKG */
@@ -72,6 +72,12 @@
72 1<<RAPL_IDX_PKG_NRG_STAT|\ 72 1<<RAPL_IDX_PKG_NRG_STAT|\
73 1<<RAPL_IDX_RAM_NRG_STAT) 73 1<<RAPL_IDX_RAM_NRG_STAT)
74 74
75/* Servers have PP0, PKG, RAM, PP1 */
76#define RAPL_IDX_HSW (1<<RAPL_IDX_PP0_NRG_STAT|\
77 1<<RAPL_IDX_PKG_NRG_STAT|\
78 1<<RAPL_IDX_RAM_NRG_STAT|\
79 1<<RAPL_IDX_PP1_NRG_STAT)
80
75/* 81/*
76 * event code: LSB 8 bits, passed in attr->config 82 * event code: LSB 8 bits, passed in attr->config
77 * any other bit is reserved 83 * any other bit is reserved
@@ -425,6 +431,24 @@ static struct attribute *rapl_events_cln_attr[] = {
425 NULL, 431 NULL,
426}; 432};
427 433
434static struct attribute *rapl_events_hsw_attr[] = {
435 EVENT_PTR(rapl_cores),
436 EVENT_PTR(rapl_pkg),
437 EVENT_PTR(rapl_gpu),
438 EVENT_PTR(rapl_ram),
439
440 EVENT_PTR(rapl_cores_unit),
441 EVENT_PTR(rapl_pkg_unit),
442 EVENT_PTR(rapl_gpu_unit),
443 EVENT_PTR(rapl_ram_unit),
444
445 EVENT_PTR(rapl_cores_scale),
446 EVENT_PTR(rapl_pkg_scale),
447 EVENT_PTR(rapl_gpu_scale),
448 EVENT_PTR(rapl_ram_scale),
449 NULL,
450};
451
428static struct attribute_group rapl_pmu_events_group = { 452static struct attribute_group rapl_pmu_events_group = {
429 .name = "events", 453 .name = "events",
430 .attrs = NULL, /* patched at runtime */ 454 .attrs = NULL, /* patched at runtime */
@@ -511,6 +535,7 @@ static int rapl_cpu_prepare(int cpu)
511 struct rapl_pmu *pmu = per_cpu(rapl_pmu, cpu); 535 struct rapl_pmu *pmu = per_cpu(rapl_pmu, cpu);
512 int phys_id = topology_physical_package_id(cpu); 536 int phys_id = topology_physical_package_id(cpu);
513 u64 ms; 537 u64 ms;
538 u64 msr_rapl_power_unit_bits;
514 539
515 if (pmu) 540 if (pmu)
516 return 0; 541 return 0;
@@ -518,6 +543,10 @@ static int rapl_cpu_prepare(int cpu)
518 if (phys_id < 0) 543 if (phys_id < 0)
519 return -1; 544 return -1;
520 545
546 /* protect rdmsrl() to handle virtualization */
547 if (rdmsrl_safe(MSR_RAPL_POWER_UNIT, &msr_rapl_power_unit_bits))
548 return -1;
549
521 pmu = kzalloc_node(sizeof(*pmu), GFP_KERNEL, cpu_to_node(cpu)); 550 pmu = kzalloc_node(sizeof(*pmu), GFP_KERNEL, cpu_to_node(cpu));
522 if (!pmu) 551 if (!pmu)
523 return -1; 552 return -1;
@@ -531,8 +560,7 @@ static int rapl_cpu_prepare(int cpu)
531 * 560 *
532 * we cache in local PMU instance 561 * we cache in local PMU instance
533 */ 562 */
534 rdmsrl(MSR_RAPL_POWER_UNIT, pmu->hw_unit); 563 pmu->hw_unit = (msr_rapl_power_unit_bits >> 8) & 0x1FULL;
535 pmu->hw_unit = (pmu->hw_unit >> 8) & 0x1FULL;
536 pmu->pmu = &rapl_pmu_class; 564 pmu->pmu = &rapl_pmu_class;
537 565
538 /* 566 /*
@@ -631,11 +659,14 @@ static int __init rapl_pmu_init(void)
631 switch (boot_cpu_data.x86_model) { 659 switch (boot_cpu_data.x86_model) {
632 case 42: /* Sandy Bridge */ 660 case 42: /* Sandy Bridge */
633 case 58: /* Ivy Bridge */ 661 case 58: /* Ivy Bridge */
634 case 60: /* Haswell */
635 case 69: /* Haswell-Celeron */
636 rapl_cntr_mask = RAPL_IDX_CLN; 662 rapl_cntr_mask = RAPL_IDX_CLN;
637 rapl_pmu_events_group.attrs = rapl_events_cln_attr; 663 rapl_pmu_events_group.attrs = rapl_events_cln_attr;
638 break; 664 break;
665 case 60: /* Haswell */
666 case 69: /* Haswell-Celeron */
667 rapl_cntr_mask = RAPL_IDX_HSW;
668 rapl_pmu_events_group.attrs = rapl_events_hsw_attr;
669 break;
639 case 45: /* Sandy Bridge-EP */ 670 case 45: /* Sandy Bridge-EP */
640 case 62: /* IvyTown */ 671 case 62: /* IvyTown */
641 rapl_cntr_mask = RAPL_IDX_SRV; 672 rapl_cntr_mask = RAPL_IDX_SRV;
@@ -650,7 +681,9 @@ static int __init rapl_pmu_init(void)
650 cpu_notifier_register_begin(); 681 cpu_notifier_register_begin();
651 682
652 for_each_online_cpu(cpu) { 683 for_each_online_cpu(cpu) {
653 rapl_cpu_prepare(cpu); 684 ret = rapl_cpu_prepare(cpu);
685 if (ret)
686 goto out;
654 rapl_cpu_init(cpu); 687 rapl_cpu_init(cpu);
655 } 688 }
656 689
@@ -673,6 +706,7 @@ static int __init rapl_pmu_init(void)
673 hweight32(rapl_cntr_mask), 706 hweight32(rapl_cntr_mask),
674 ktime_to_ms(pmu->timer_interval)); 707 ktime_to_ms(pmu->timer_interval));
675 708
709out:
676 cpu_notifier_register_done(); 710 cpu_notifier_register_done();
677 711
678 return 0; 712 return 0;
diff --git a/arch/x86/kernel/cpu/rdrand.c b/arch/x86/kernel/cpu/rdrand.c
index 384df5105fbc..136ac74dee82 100644
--- a/arch/x86/kernel/cpu/rdrand.c
+++ b/arch/x86/kernel/cpu/rdrand.c
@@ -27,6 +27,7 @@
27static int __init x86_rdrand_setup(char *s) 27static int __init x86_rdrand_setup(char *s)
28{ 28{
29 setup_clear_cpu_cap(X86_FEATURE_RDRAND); 29 setup_clear_cpu_cap(X86_FEATURE_RDRAND);
30 setup_clear_cpu_cap(X86_FEATURE_RDSEED);
30 return 1; 31 return 1;
31} 32}
32__setup("nordrand", x86_rdrand_setup); 33__setup("nordrand", x86_rdrand_setup);
diff --git a/arch/x86/kernel/early-quirks.c b/arch/x86/kernel/early-quirks.c
index b0cc3809723d..6cda0baeac9d 100644
--- a/arch/x86/kernel/early-quirks.c
+++ b/arch/x86/kernel/early-quirks.c
@@ -17,6 +17,7 @@
17#include <asm/dma.h> 17#include <asm/dma.h>
18#include <asm/io_apic.h> 18#include <asm/io_apic.h>
19#include <asm/apic.h> 19#include <asm/apic.h>
20#include <asm/hpet.h>
20#include <asm/iommu.h> 21#include <asm/iommu.h>
21#include <asm/gart.h> 22#include <asm/gart.h>
22#include <asm/irq_remapping.h> 23#include <asm/irq_remapping.h>
@@ -240,7 +241,7 @@ static u32 __init intel_stolen_base(int num, int slot, int func, size_t stolen_s
240 return base; 241 return base;
241} 242}
242 243
243#define KB(x) ((x) * 1024) 244#define KB(x) ((x) * 1024UL)
244#define MB(x) (KB (KB (x))) 245#define MB(x) (KB (KB (x)))
245#define GB(x) (MB (KB (x))) 246#define GB(x) (MB (KB (x)))
246 247
@@ -530,6 +531,15 @@ static void __init intel_graphics_stolen(int num, int slot, int func)
530 } 531 }
531} 532}
532 533
534static void __init force_disable_hpet(int num, int slot, int func)
535{
536#ifdef CONFIG_HPET_TIMER
537 boot_hpet_disable = 1;
538 pr_info("x86/hpet: Will disable the HPET for this platform because it's not reliable\n");
539#endif
540}
541
542
533#define QFLAG_APPLY_ONCE 0x1 543#define QFLAG_APPLY_ONCE 0x1
534#define QFLAG_APPLIED 0x2 544#define QFLAG_APPLIED 0x2
535#define QFLAG_DONE (QFLAG_APPLY_ONCE|QFLAG_APPLIED) 545#define QFLAG_DONE (QFLAG_APPLY_ONCE|QFLAG_APPLIED)
@@ -567,6 +577,12 @@ static struct chipset early_qrk[] __initdata = {
567 PCI_BASE_CLASS_BRIDGE, 0, intel_remapping_check }, 577 PCI_BASE_CLASS_BRIDGE, 0, intel_remapping_check },
568 { PCI_VENDOR_ID_INTEL, PCI_ANY_ID, PCI_CLASS_DISPLAY_VGA, PCI_ANY_ID, 578 { PCI_VENDOR_ID_INTEL, PCI_ANY_ID, PCI_CLASS_DISPLAY_VGA, PCI_ANY_ID,
569 QFLAG_APPLY_ONCE, intel_graphics_stolen }, 579 QFLAG_APPLY_ONCE, intel_graphics_stolen },
580 /*
581 * HPET on current version of Baytrail platform has accuracy
582 * problems, disable it for now:
583 */
584 { PCI_VENDOR_ID_INTEL, 0x0f00,
585 PCI_CLASS_BRIDGE_HOST, PCI_ANY_ID, 0, force_disable_hpet},
570 {} 586 {}
571}; 587};
572 588
diff --git a/arch/x86/kernel/head32.c b/arch/x86/kernel/head32.c
index c61a14a4a310..d6c1b9836995 100644
--- a/arch/x86/kernel/head32.c
+++ b/arch/x86/kernel/head32.c
@@ -29,7 +29,7 @@ static void __init i386_default_early_setup(void)
29 reserve_ebda_region(); 29 reserve_ebda_region();
30} 30}
31 31
32asmlinkage void __init i386_start_kernel(void) 32asmlinkage __visible void __init i386_start_kernel(void)
33{ 33{
34 sanitize_boot_params(&boot_params); 34 sanitize_boot_params(&boot_params);
35 35
diff --git a/arch/x86/kernel/head64.c b/arch/x86/kernel/head64.c
index 85126ccbdf6b..068054f4bf20 100644
--- a/arch/x86/kernel/head64.c
+++ b/arch/x86/kernel/head64.c
@@ -137,7 +137,7 @@ static void __init copy_bootdata(char *real_mode_data)
137 } 137 }
138} 138}
139 139
140asmlinkage void __init x86_64_start_kernel(char * real_mode_data) 140asmlinkage __visible void __init x86_64_start_kernel(char * real_mode_data)
141{ 141{
142 int i; 142 int i;
143 143
diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c
index 8d80ae011603..4177bfbc80b0 100644
--- a/arch/x86/kernel/hpet.c
+++ b/arch/x86/kernel/hpet.c
@@ -88,7 +88,7 @@ static inline void hpet_clear_mapping(void)
88/* 88/*
89 * HPET command line enable / disable 89 * HPET command line enable / disable
90 */ 90 */
91static int boot_hpet_disable; 91int boot_hpet_disable;
92int hpet_force_user; 92int hpet_force_user;
93static int hpet_verbose; 93static int hpet_verbose;
94 94
diff --git a/arch/x86/kernel/kprobes/core.c b/arch/x86/kernel/kprobes/core.c
index 79a3f9682871..61b17dc2c277 100644
--- a/arch/x86/kernel/kprobes/core.c
+++ b/arch/x86/kernel/kprobes/core.c
@@ -897,9 +897,10 @@ int __kprobes kprobe_fault_handler(struct pt_regs *regs, int trapnr)
897 struct kprobe *cur = kprobe_running(); 897 struct kprobe *cur = kprobe_running();
898 struct kprobe_ctlblk *kcb = get_kprobe_ctlblk(); 898 struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
899 899
900 switch (kcb->kprobe_status) { 900 if (unlikely(regs->ip == (unsigned long)cur->ainsn.insn)) {
901 case KPROBE_HIT_SS: 901 /* This must happen on single-stepping */
902 case KPROBE_REENTER: 902 WARN_ON(kcb->kprobe_status != KPROBE_HIT_SS &&
903 kcb->kprobe_status != KPROBE_REENTER);
903 /* 904 /*
904 * We are here because the instruction being single 905 * We are here because the instruction being single
905 * stepped caused a page fault. We reset the current 906 * stepped caused a page fault. We reset the current
@@ -914,9 +915,8 @@ int __kprobes kprobe_fault_handler(struct pt_regs *regs, int trapnr)
914 else 915 else
915 reset_current_kprobe(); 916 reset_current_kprobe();
916 preempt_enable_no_resched(); 917 preempt_enable_no_resched();
917 break; 918 } else if (kcb->kprobe_status == KPROBE_HIT_ACTIVE ||
918 case KPROBE_HIT_ACTIVE: 919 kcb->kprobe_status == KPROBE_HIT_SSDONE) {
919 case KPROBE_HIT_SSDONE:
920 /* 920 /*
921 * We increment the nmissed count for accounting, 921 * We increment the nmissed count for accounting,
922 * we can also use npre/npostfault count for accounting 922 * we can also use npre/npostfault count for accounting
@@ -945,10 +945,8 @@ int __kprobes kprobe_fault_handler(struct pt_regs *regs, int trapnr)
945 * fixup routine could not handle it, 945 * fixup routine could not handle it,
946 * Let do_page_fault() fix it. 946 * Let do_page_fault() fix it.
947 */ 947 */
948 break;
949 default:
950 break;
951 } 948 }
949
952 return 0; 950 return 0;
953} 951}
954 952
diff --git a/arch/x86/kernel/ldt.c b/arch/x86/kernel/ldt.c
index af1d14a9ebda..dcbbaa165bde 100644
--- a/arch/x86/kernel/ldt.c
+++ b/arch/x86/kernel/ldt.c
@@ -20,6 +20,8 @@
20#include <asm/mmu_context.h> 20#include <asm/mmu_context.h>
21#include <asm/syscalls.h> 21#include <asm/syscalls.h>
22 22
23int sysctl_ldt16 = 0;
24
23#ifdef CONFIG_SMP 25#ifdef CONFIG_SMP
24static void flush_ldt(void *current_mm) 26static void flush_ldt(void *current_mm)
25{ 27{
@@ -234,7 +236,7 @@ static int write_ldt(void __user *ptr, unsigned long bytecount, int oldmode)
234 * IRET leaking the high bits of the kernel stack address. 236 * IRET leaking the high bits of the kernel stack address.
235 */ 237 */
236#ifdef CONFIG_X86_64 238#ifdef CONFIG_X86_64
237 if (!ldt_info.seg_32bit) { 239 if (!ldt_info.seg_32bit && !sysctl_ldt16) {
238 error = -EINVAL; 240 error = -EINVAL;
239 goto out_unlock; 241 goto out_unlock;
240 } 242 }
diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c
index 9c0280f93d05..898d077617a9 100644
--- a/arch/x86/kernel/process_64.c
+++ b/arch/x86/kernel/process_64.c
@@ -52,7 +52,7 @@
52 52
53asmlinkage extern void ret_from_fork(void); 53asmlinkage extern void ret_from_fork(void);
54 54
55asmlinkage DEFINE_PER_CPU(unsigned long, old_rsp); 55__visible DEFINE_PER_CPU(unsigned long, old_rsp);
56 56
57/* Prints also some state that isn't saved in the pt_regs */ 57/* Prints also some state that isn't saved in the pt_regs */
58void __show_regs(struct pt_regs *regs, int all) 58void __show_regs(struct pt_regs *regs, int all)
diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c
index 654b46574b91..52b1157c53eb 100644
--- a/arch/x86/kernel/reboot.c
+++ b/arch/x86/kernel/reboot.c
@@ -114,8 +114,8 @@ EXPORT_SYMBOL(machine_real_restart);
114 */ 114 */
115static int __init set_pci_reboot(const struct dmi_system_id *d) 115static int __init set_pci_reboot(const struct dmi_system_id *d)
116{ 116{
117 if (reboot_type != BOOT_CF9) { 117 if (reboot_type != BOOT_CF9_FORCE) {
118 reboot_type = BOOT_CF9; 118 reboot_type = BOOT_CF9_FORCE;
119 pr_info("%s series board detected. Selecting %s-method for reboots.\n", 119 pr_info("%s series board detected. Selecting %s-method for reboots.\n",
120 d->ident, "PCI"); 120 d->ident, "PCI");
121 } 121 }
@@ -191,6 +191,16 @@ static struct dmi_system_id __initdata reboot_dmi_table[] = {
191 }, 191 },
192 }, 192 },
193 193
194 /* Certec */
195 { /* Handle problems with rebooting on Certec BPC600 */
196 .callback = set_pci_reboot,
197 .ident = "Certec BPC600",
198 .matches = {
199 DMI_MATCH(DMI_SYS_VENDOR, "Certec"),
200 DMI_MATCH(DMI_PRODUCT_NAME, "BPC600"),
201 },
202 },
203
194 /* Dell */ 204 /* Dell */
195 { /* Handle problems with rebooting on Dell DXP061 */ 205 { /* Handle problems with rebooting on Dell DXP061 */
196 .callback = set_bios_reboot, 206 .callback = set_bios_reboot,
@@ -458,20 +468,23 @@ void __attribute__((weak)) mach_reboot_fixups(void)
458} 468}
459 469
460/* 470/*
461 * Windows compatible x86 hardware expects the following on reboot: 471 * To the best of our knowledge Windows compatible x86 hardware expects
472 * the following on reboot:
462 * 473 *
463 * 1) If the FADT has the ACPI reboot register flag set, try it 474 * 1) If the FADT has the ACPI reboot register flag set, try it
464 * 2) If still alive, write to the keyboard controller 475 * 2) If still alive, write to the keyboard controller
465 * 3) If still alive, write to the ACPI reboot register again 476 * 3) If still alive, write to the ACPI reboot register again
466 * 4) If still alive, write to the keyboard controller again 477 * 4) If still alive, write to the keyboard controller again
467 * 5) If still alive, call the EFI runtime service to reboot 478 * 5) If still alive, call the EFI runtime service to reboot
468 * 6) If still alive, write to the PCI IO port 0xCF9 to reboot 479 * 6) If no EFI runtime service, call the BIOS to do a reboot
469 * 7) If still alive, inform BIOS to do a proper reboot 480 *
481 * We default to following the same pattern. We also have
482 * two other reboot methods: 'triple fault' and 'PCI', which
483 * can be triggered via the reboot= kernel boot option or
484 * via quirks.
470 * 485 *
471 * If the machine is still alive at this stage, it gives up. We default to 486 * This means that this function can never return, it can misbehave
472 * following the same pattern, except that if we're still alive after (7) we'll 487 * by not rebooting properly and hanging.
473 * try to force a triple fault and then cycle between hitting the keyboard
474 * controller and doing that
475 */ 488 */
476static void native_machine_emergency_restart(void) 489static void native_machine_emergency_restart(void)
477{ 490{
@@ -492,6 +505,11 @@ static void native_machine_emergency_restart(void)
492 for (;;) { 505 for (;;) {
493 /* Could also try the reset bit in the Hammer NB */ 506 /* Could also try the reset bit in the Hammer NB */
494 switch (reboot_type) { 507 switch (reboot_type) {
508 case BOOT_ACPI:
509 acpi_reboot();
510 reboot_type = BOOT_KBD;
511 break;
512
495 case BOOT_KBD: 513 case BOOT_KBD:
496 mach_reboot_fixups(); /* For board specific fixups */ 514 mach_reboot_fixups(); /* For board specific fixups */
497 515
@@ -509,43 +527,29 @@ static void native_machine_emergency_restart(void)
509 } 527 }
510 break; 528 break;
511 529
512 case BOOT_TRIPLE:
513 load_idt(&no_idt);
514 __asm__ __volatile__("int3");
515
516 /* We're probably dead after this, but... */
517 reboot_type = BOOT_KBD;
518 break;
519
520 case BOOT_BIOS:
521 machine_real_restart(MRR_BIOS);
522
523 /* We're probably dead after this, but... */
524 reboot_type = BOOT_TRIPLE;
525 break;
526
527 case BOOT_ACPI:
528 acpi_reboot();
529 reboot_type = BOOT_KBD;
530 break;
531
532 case BOOT_EFI: 530 case BOOT_EFI:
533 if (efi_enabled(EFI_RUNTIME_SERVICES)) 531 if (efi_enabled(EFI_RUNTIME_SERVICES))
534 efi.reset_system(reboot_mode == REBOOT_WARM ? 532 efi.reset_system(reboot_mode == REBOOT_WARM ?
535 EFI_RESET_WARM : 533 EFI_RESET_WARM :
536 EFI_RESET_COLD, 534 EFI_RESET_COLD,
537 EFI_SUCCESS, 0, NULL); 535 EFI_SUCCESS, 0, NULL);
538 reboot_type = BOOT_CF9_COND; 536 reboot_type = BOOT_BIOS;
537 break;
538
539 case BOOT_BIOS:
540 machine_real_restart(MRR_BIOS);
541
542 /* We're probably dead after this, but... */
543 reboot_type = BOOT_CF9_SAFE;
539 break; 544 break;
540 545
541 case BOOT_CF9: 546 case BOOT_CF9_FORCE:
542 port_cf9_safe = true; 547 port_cf9_safe = true;
543 /* Fall through */ 548 /* Fall through */
544 549
545 case BOOT_CF9_COND: 550 case BOOT_CF9_SAFE:
546 if (port_cf9_safe) { 551 if (port_cf9_safe) {
547 u8 reboot_code = reboot_mode == REBOOT_WARM ? 552 u8 reboot_code = reboot_mode == REBOOT_WARM ? 0x06 : 0x0E;
548 0x06 : 0x0E;
549 u8 cf9 = inb(0xcf9) & ~reboot_code; 553 u8 cf9 = inb(0xcf9) & ~reboot_code;
550 outb(cf9|2, 0xcf9); /* Request hard reset */ 554 outb(cf9|2, 0xcf9); /* Request hard reset */
551 udelay(50); 555 udelay(50);
@@ -553,7 +557,15 @@ static void native_machine_emergency_restart(void)
553 outb(cf9|reboot_code, 0xcf9); 557 outb(cf9|reboot_code, 0xcf9);
554 udelay(50); 558 udelay(50);
555 } 559 }
556 reboot_type = BOOT_BIOS; 560 reboot_type = BOOT_TRIPLE;
561 break;
562
563 case BOOT_TRIPLE:
564 load_idt(&no_idt);
565 __asm__ __volatile__("int3");
566
567 /* We're probably dead after this, but... */
568 reboot_type = BOOT_KBD;
557 break; 569 break;
558 } 570 }
559 } 571 }
diff --git a/arch/x86/kernel/smp.c b/arch/x86/kernel/smp.c
index 7c3a5a61f2e4..be8e1bde07aa 100644
--- a/arch/x86/kernel/smp.c
+++ b/arch/x86/kernel/smp.c
@@ -168,7 +168,7 @@ static int smp_stop_nmi_callback(unsigned int val, struct pt_regs *regs)
168 * this function calls the 'stop' function on all other CPUs in the system. 168 * this function calls the 'stop' function on all other CPUs in the system.
169 */ 169 */
170 170
171asmlinkage void smp_reboot_interrupt(void) 171asmlinkage __visible void smp_reboot_interrupt(void)
172{ 172{
173 ack_APIC_irq(); 173 ack_APIC_irq();
174 irq_enter(); 174 irq_enter();
diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
index 57409f6b8c62..f73b5d435bdc 100644
--- a/arch/x86/kernel/traps.c
+++ b/arch/x86/kernel/traps.c
@@ -357,7 +357,7 @@ exit:
357 * for scheduling or signal handling. The actual stack switch is done in 357 * for scheduling or signal handling. The actual stack switch is done in
358 * entry.S 358 * entry.S
359 */ 359 */
360asmlinkage __kprobes struct pt_regs *sync_regs(struct pt_regs *eregs) 360asmlinkage __visible __kprobes struct pt_regs *sync_regs(struct pt_regs *eregs)
361{ 361{
362 struct pt_regs *regs = eregs; 362 struct pt_regs *regs = eregs;
363 /* Did already sync */ 363 /* Did already sync */
@@ -601,11 +601,11 @@ do_spurious_interrupt_bug(struct pt_regs *regs, long error_code)
601#endif 601#endif
602} 602}
603 603
604asmlinkage void __attribute__((weak)) smp_thermal_interrupt(void) 604asmlinkage __visible void __attribute__((weak)) smp_thermal_interrupt(void)
605{ 605{
606} 606}
607 607
608asmlinkage void __attribute__((weak)) smp_threshold_interrupt(void) 608asmlinkage __visible void __attribute__((weak)) smp_threshold_interrupt(void)
609{ 609{
610} 610}
611 611
diff --git a/arch/x86/kernel/vsmp_64.c b/arch/x86/kernel/vsmp_64.c
index f6584a90aba3..b99b9ad8540c 100644
--- a/arch/x86/kernel/vsmp_64.c
+++ b/arch/x86/kernel/vsmp_64.c
@@ -26,6 +26,9 @@
26 26
27#define TOPOLOGY_REGISTER_OFFSET 0x10 27#define TOPOLOGY_REGISTER_OFFSET 0x10
28 28
29/* Flag below is initialized once during vSMP PCI initialization. */
30static int irq_routing_comply = 1;
31
29#if defined CONFIG_PCI && defined CONFIG_PARAVIRT 32#if defined CONFIG_PCI && defined CONFIG_PARAVIRT
30/* 33/*
31 * Interrupt control on vSMPowered systems: 34 * Interrupt control on vSMPowered systems:
@@ -33,7 +36,7 @@
33 * and vice versa. 36 * and vice versa.
34 */ 37 */
35 38
36asmlinkage unsigned long vsmp_save_fl(void) 39asmlinkage __visible unsigned long vsmp_save_fl(void)
37{ 40{
38 unsigned long flags = native_save_fl(); 41 unsigned long flags = native_save_fl();
39 42
@@ -53,7 +56,7 @@ __visible void vsmp_restore_fl(unsigned long flags)
53} 56}
54PV_CALLEE_SAVE_REGS_THUNK(vsmp_restore_fl); 57PV_CALLEE_SAVE_REGS_THUNK(vsmp_restore_fl);
55 58
56asmlinkage void vsmp_irq_disable(void) 59asmlinkage __visible void vsmp_irq_disable(void)
57{ 60{
58 unsigned long flags = native_save_fl(); 61 unsigned long flags = native_save_fl();
59 62
@@ -61,7 +64,7 @@ asmlinkage void vsmp_irq_disable(void)
61} 64}
62PV_CALLEE_SAVE_REGS_THUNK(vsmp_irq_disable); 65PV_CALLEE_SAVE_REGS_THUNK(vsmp_irq_disable);
63 66
64asmlinkage void vsmp_irq_enable(void) 67asmlinkage __visible void vsmp_irq_enable(void)
65{ 68{
66 unsigned long flags = native_save_fl(); 69 unsigned long flags = native_save_fl();
67 70
@@ -101,6 +104,10 @@ static void __init set_vsmp_pv_ops(void)
101#ifdef CONFIG_SMP 104#ifdef CONFIG_SMP
102 if (cap & ctl & BIT(8)) { 105 if (cap & ctl & BIT(8)) {
103 ctl &= ~BIT(8); 106 ctl &= ~BIT(8);
107
108 /* Interrupt routing set to ignore */
109 irq_routing_comply = 0;
110
104#ifdef CONFIG_PROC_FS 111#ifdef CONFIG_PROC_FS
105 /* Don't let users change irq affinity via procfs */ 112 /* Don't let users change irq affinity via procfs */
106 no_irq_affinity = 1; 113 no_irq_affinity = 1;
@@ -218,7 +225,9 @@ static void vsmp_apic_post_init(void)
218{ 225{
219 /* need to update phys_pkg_id */ 226 /* need to update phys_pkg_id */
220 apic->phys_pkg_id = apicid_phys_pkg_id; 227 apic->phys_pkg_id = apicid_phys_pkg_id;
221 apic->vector_allocation_domain = fill_vector_allocation_domain; 228
229 if (!irq_routing_comply)
230 apic->vector_allocation_domain = fill_vector_allocation_domain;
222} 231}
223 232
224void __init vsmp_init(void) 233void __init vsmp_init(void)
diff --git a/arch/x86/kernel/vsyscall_gtod.c b/arch/x86/kernel/vsyscall_gtod.c
index f9c6e56e14b5..9531fbb123ba 100644
--- a/arch/x86/kernel/vsyscall_gtod.c
+++ b/arch/x86/kernel/vsyscall_gtod.c
@@ -43,7 +43,7 @@ void update_vsyscall(struct timekeeper *tk)
43 vdata->monotonic_time_sec = tk->xtime_sec 43 vdata->monotonic_time_sec = tk->xtime_sec
44 + tk->wall_to_monotonic.tv_sec; 44 + tk->wall_to_monotonic.tv_sec;
45 vdata->monotonic_time_snsec = tk->xtime_nsec 45 vdata->monotonic_time_snsec = tk->xtime_nsec
46 + (tk->wall_to_monotonic.tv_nsec 46 + ((u64)tk->wall_to_monotonic.tv_nsec
47 << tk->shift); 47 << tk->shift);
48 while (vdata->monotonic_time_snsec >= 48 while (vdata->monotonic_time_snsec >=
49 (((u64)NSEC_PER_SEC) << tk->shift)) { 49 (((u64)NSEC_PER_SEC) << tk->shift)) {
diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
index bea60671ef8a..f47a104a749c 100644
--- a/arch/x86/kvm/cpuid.c
+++ b/arch/x86/kvm/cpuid.c
@@ -308,7 +308,7 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
308 const u32 kvm_supported_word9_x86_features = 308 const u32 kvm_supported_word9_x86_features =
309 F(FSGSBASE) | F(BMI1) | F(HLE) | F(AVX2) | F(SMEP) | 309 F(FSGSBASE) | F(BMI1) | F(HLE) | F(AVX2) | F(SMEP) |
310 F(BMI2) | F(ERMS) | f_invpcid | F(RTM) | f_mpx | F(RDSEED) | 310 F(BMI2) | F(ERMS) | f_invpcid | F(RTM) | f_mpx | F(RDSEED) |
311 F(ADX); 311 F(ADX) | F(SMAP);
312 312
313 /* all calls to cpuid_count() should be made on the same cpu */ 313 /* all calls to cpuid_count() should be made on the same cpu */
314 get_cpu(); 314 get_cpu();
diff --git a/arch/x86/kvm/cpuid.h b/arch/x86/kvm/cpuid.h
index a2a1bb7ed8c1..eeecbed26ac7 100644
--- a/arch/x86/kvm/cpuid.h
+++ b/arch/x86/kvm/cpuid.h
@@ -48,6 +48,14 @@ static inline bool guest_cpuid_has_smep(struct kvm_vcpu *vcpu)
48 return best && (best->ebx & bit(X86_FEATURE_SMEP)); 48 return best && (best->ebx & bit(X86_FEATURE_SMEP));
49} 49}
50 50
51static inline bool guest_cpuid_has_smap(struct kvm_vcpu *vcpu)
52{
53 struct kvm_cpuid_entry2 *best;
54
55 best = kvm_find_cpuid_entry(vcpu, 7, 0);
56 return best && (best->ebx & bit(X86_FEATURE_SMAP));
57}
58
51static inline bool guest_cpuid_has_fsgsbase(struct kvm_vcpu *vcpu) 59static inline bool guest_cpuid_has_fsgsbase(struct kvm_vcpu *vcpu)
52{ 60{
53 struct kvm_cpuid_entry2 *best; 61 struct kvm_cpuid_entry2 *best;
diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index f5704d9e5ddc..813d31038b93 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -3601,20 +3601,27 @@ static void reset_rsvds_bits_mask_ept(struct kvm_vcpu *vcpu,
3601 } 3601 }
3602} 3602}
3603 3603
3604static void update_permission_bitmask(struct kvm_vcpu *vcpu, 3604void update_permission_bitmask(struct kvm_vcpu *vcpu,
3605 struct kvm_mmu *mmu, bool ept) 3605 struct kvm_mmu *mmu, bool ept)
3606{ 3606{
3607 unsigned bit, byte, pfec; 3607 unsigned bit, byte, pfec;
3608 u8 map; 3608 u8 map;
3609 bool fault, x, w, u, wf, uf, ff, smep; 3609 bool fault, x, w, u, wf, uf, ff, smapf, cr4_smap, cr4_smep, smap = 0;
3610 3610
3611 smep = kvm_read_cr4_bits(vcpu, X86_CR4_SMEP); 3611 cr4_smep = kvm_read_cr4_bits(vcpu, X86_CR4_SMEP);
3612 cr4_smap = kvm_read_cr4_bits(vcpu, X86_CR4_SMAP);
3612 for (byte = 0; byte < ARRAY_SIZE(mmu->permissions); ++byte) { 3613 for (byte = 0; byte < ARRAY_SIZE(mmu->permissions); ++byte) {
3613 pfec = byte << 1; 3614 pfec = byte << 1;
3614 map = 0; 3615 map = 0;
3615 wf = pfec & PFERR_WRITE_MASK; 3616 wf = pfec & PFERR_WRITE_MASK;
3616 uf = pfec & PFERR_USER_MASK; 3617 uf = pfec & PFERR_USER_MASK;
3617 ff = pfec & PFERR_FETCH_MASK; 3618 ff = pfec & PFERR_FETCH_MASK;
3619 /*
3620 * PFERR_RSVD_MASK bit is set in PFEC if the access is not
3621 * subject to SMAP restrictions, and cleared otherwise. The
3622 * bit is only meaningful if the SMAP bit is set in CR4.
3623 */
3624 smapf = !(pfec & PFERR_RSVD_MASK);
3618 for (bit = 0; bit < 8; ++bit) { 3625 for (bit = 0; bit < 8; ++bit) {
3619 x = bit & ACC_EXEC_MASK; 3626 x = bit & ACC_EXEC_MASK;
3620 w = bit & ACC_WRITE_MASK; 3627 w = bit & ACC_WRITE_MASK;
@@ -3626,12 +3633,33 @@ static void update_permission_bitmask(struct kvm_vcpu *vcpu,
3626 /* Allow supervisor writes if !cr0.wp */ 3633 /* Allow supervisor writes if !cr0.wp */
3627 w |= !is_write_protection(vcpu) && !uf; 3634 w |= !is_write_protection(vcpu) && !uf;
3628 /* Disallow supervisor fetches of user code if cr4.smep */ 3635 /* Disallow supervisor fetches of user code if cr4.smep */
3629 x &= !(smep && u && !uf); 3636 x &= !(cr4_smep && u && !uf);
3637
3638 /*
3639 * SMAP:kernel-mode data accesses from user-mode
3640 * mappings should fault. A fault is considered
3641 * as a SMAP violation if all of the following
3642 * conditions are ture:
3643 * - X86_CR4_SMAP is set in CR4
3644 * - An user page is accessed
3645 * - Page fault in kernel mode
3646 * - if CPL = 3 or X86_EFLAGS_AC is clear
3647 *
3648 * Here, we cover the first three conditions.
3649 * The fourth is computed dynamically in
3650 * permission_fault() and is in smapf.
3651 *
3652 * Also, SMAP does not affect instruction
3653 * fetches, add the !ff check here to make it
3654 * clearer.
3655 */
3656 smap = cr4_smap && u && !uf && !ff;
3630 } else 3657 } else
3631 /* Not really needed: no U/S accesses on ept */ 3658 /* Not really needed: no U/S accesses on ept */
3632 u = 1; 3659 u = 1;
3633 3660
3634 fault = (ff && !x) || (uf && !u) || (wf && !w); 3661 fault = (ff && !x) || (uf && !u) || (wf && !w) ||
3662 (smapf && smap);
3635 map |= fault << bit; 3663 map |= fault << bit;
3636 } 3664 }
3637 mmu->permissions[byte] = map; 3665 mmu->permissions[byte] = map;
diff --git a/arch/x86/kvm/mmu.h b/arch/x86/kvm/mmu.h
index 292615274358..3842e70bdb7c 100644
--- a/arch/x86/kvm/mmu.h
+++ b/arch/x86/kvm/mmu.h
@@ -44,11 +44,17 @@
44#define PT_DIRECTORY_LEVEL 2 44#define PT_DIRECTORY_LEVEL 2
45#define PT_PAGE_TABLE_LEVEL 1 45#define PT_PAGE_TABLE_LEVEL 1
46 46
47#define PFERR_PRESENT_MASK (1U << 0) 47#define PFERR_PRESENT_BIT 0
48#define PFERR_WRITE_MASK (1U << 1) 48#define PFERR_WRITE_BIT 1
49#define PFERR_USER_MASK (1U << 2) 49#define PFERR_USER_BIT 2
50#define PFERR_RSVD_MASK (1U << 3) 50#define PFERR_RSVD_BIT 3
51#define PFERR_FETCH_MASK (1U << 4) 51#define PFERR_FETCH_BIT 4
52
53#define PFERR_PRESENT_MASK (1U << PFERR_PRESENT_BIT)
54#define PFERR_WRITE_MASK (1U << PFERR_WRITE_BIT)
55#define PFERR_USER_MASK (1U << PFERR_USER_BIT)
56#define PFERR_RSVD_MASK (1U << PFERR_RSVD_BIT)
57#define PFERR_FETCH_MASK (1U << PFERR_FETCH_BIT)
52 58
53int kvm_mmu_get_spte_hierarchy(struct kvm_vcpu *vcpu, u64 addr, u64 sptes[4]); 59int kvm_mmu_get_spte_hierarchy(struct kvm_vcpu *vcpu, u64 addr, u64 sptes[4]);
54void kvm_mmu_set_mmio_spte_mask(u64 mmio_mask); 60void kvm_mmu_set_mmio_spte_mask(u64 mmio_mask);
@@ -73,6 +79,8 @@ int handle_mmio_page_fault_common(struct kvm_vcpu *vcpu, u64 addr, bool direct);
73void kvm_init_shadow_mmu(struct kvm_vcpu *vcpu, struct kvm_mmu *context); 79void kvm_init_shadow_mmu(struct kvm_vcpu *vcpu, struct kvm_mmu *context);
74void kvm_init_shadow_ept_mmu(struct kvm_vcpu *vcpu, struct kvm_mmu *context, 80void kvm_init_shadow_ept_mmu(struct kvm_vcpu *vcpu, struct kvm_mmu *context,
75 bool execonly); 81 bool execonly);
82void update_permission_bitmask(struct kvm_vcpu *vcpu, struct kvm_mmu *mmu,
83 bool ept);
76 84
77static inline unsigned int kvm_mmu_available_pages(struct kvm *kvm) 85static inline unsigned int kvm_mmu_available_pages(struct kvm *kvm)
78{ 86{
@@ -110,10 +118,30 @@ static inline bool is_write_protection(struct kvm_vcpu *vcpu)
110 * Will a fault with a given page-fault error code (pfec) cause a permission 118 * Will a fault with a given page-fault error code (pfec) cause a permission
111 * fault with the given access (in ACC_* format)? 119 * fault with the given access (in ACC_* format)?
112 */ 120 */
113static inline bool permission_fault(struct kvm_mmu *mmu, unsigned pte_access, 121static inline bool permission_fault(struct kvm_vcpu *vcpu, struct kvm_mmu *mmu,
114 unsigned pfec) 122 unsigned pte_access, unsigned pfec)
115{ 123{
116 return (mmu->permissions[pfec >> 1] >> pte_access) & 1; 124 int cpl = kvm_x86_ops->get_cpl(vcpu);
125 unsigned long rflags = kvm_x86_ops->get_rflags(vcpu);
126
127 /*
128 * If CPL < 3, SMAP prevention are disabled if EFLAGS.AC = 1.
129 *
130 * If CPL = 3, SMAP applies to all supervisor-mode data accesses
131 * (these are implicit supervisor accesses) regardless of the value
132 * of EFLAGS.AC.
133 *
134 * This computes (cpl < 3) && (rflags & X86_EFLAGS_AC), leaving
135 * the result in X86_EFLAGS_AC. We then insert it in place of
136 * the PFERR_RSVD_MASK bit; this bit will always be zero in pfec,
137 * but it will be one in index if SMAP checks are being overridden.
138 * It is important to keep this branchless.
139 */
140 unsigned long smap = (cpl - 3) & (rflags & X86_EFLAGS_AC);
141 int index = (pfec >> 1) +
142 (smap >> (X86_EFLAGS_AC_BIT - PFERR_RSVD_BIT + 1));
143
144 return (mmu->permissions[index] >> pte_access) & 1;
117} 145}
118 146
119void kvm_mmu_invalidate_zap_all_pages(struct kvm *kvm); 147void kvm_mmu_invalidate_zap_all_pages(struct kvm *kvm);
diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h
index b1e6c1bf68d3..123efd3ec29f 100644
--- a/arch/x86/kvm/paging_tmpl.h
+++ b/arch/x86/kvm/paging_tmpl.h
@@ -353,7 +353,7 @@ retry_walk:
353 walker->ptes[walker->level - 1] = pte; 353 walker->ptes[walker->level - 1] = pte;
354 } while (!is_last_gpte(mmu, walker->level, pte)); 354 } while (!is_last_gpte(mmu, walker->level, pte));
355 355
356 if (unlikely(permission_fault(mmu, pte_access, access))) { 356 if (unlikely(permission_fault(vcpu, mmu, pte_access, access))) {
357 errcode |= PFERR_PRESENT_MASK; 357 errcode |= PFERR_PRESENT_MASK;
358 goto error; 358 goto error;
359 } 359 }
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 1320e0f8e611..138ceffc6377 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -503,7 +503,7 @@ static inline struct vcpu_vmx *to_vmx(struct kvm_vcpu *vcpu)
503 [number##_HIGH] = VMCS12_OFFSET(name)+4 503 [number##_HIGH] = VMCS12_OFFSET(name)+4
504 504
505 505
506static const unsigned long shadow_read_only_fields[] = { 506static unsigned long shadow_read_only_fields[] = {
507 /* 507 /*
508 * We do NOT shadow fields that are modified when L0 508 * We do NOT shadow fields that are modified when L0
509 * traps and emulates any vmx instruction (e.g. VMPTRLD, 509 * traps and emulates any vmx instruction (e.g. VMPTRLD,
@@ -526,10 +526,10 @@ static const unsigned long shadow_read_only_fields[] = {
526 GUEST_LINEAR_ADDRESS, 526 GUEST_LINEAR_ADDRESS,
527 GUEST_PHYSICAL_ADDRESS 527 GUEST_PHYSICAL_ADDRESS
528}; 528};
529static const int max_shadow_read_only_fields = 529static int max_shadow_read_only_fields =
530 ARRAY_SIZE(shadow_read_only_fields); 530 ARRAY_SIZE(shadow_read_only_fields);
531 531
532static const unsigned long shadow_read_write_fields[] = { 532static unsigned long shadow_read_write_fields[] = {
533 GUEST_RIP, 533 GUEST_RIP,
534 GUEST_RSP, 534 GUEST_RSP,
535 GUEST_CR0, 535 GUEST_CR0,
@@ -558,7 +558,7 @@ static const unsigned long shadow_read_write_fields[] = {
558 HOST_FS_SELECTOR, 558 HOST_FS_SELECTOR,
559 HOST_GS_SELECTOR 559 HOST_GS_SELECTOR
560}; 560};
561static const int max_shadow_read_write_fields = 561static int max_shadow_read_write_fields =
562 ARRAY_SIZE(shadow_read_write_fields); 562 ARRAY_SIZE(shadow_read_write_fields);
563 563
564static const unsigned short vmcs_field_to_offset_table[] = { 564static const unsigned short vmcs_field_to_offset_table[] = {
@@ -3009,6 +3009,41 @@ static void free_kvm_area(void)
3009 } 3009 }
3010} 3010}
3011 3011
3012static void init_vmcs_shadow_fields(void)
3013{
3014 int i, j;
3015
3016 /* No checks for read only fields yet */
3017
3018 for (i = j = 0; i < max_shadow_read_write_fields; i++) {
3019 switch (shadow_read_write_fields[i]) {
3020 case GUEST_BNDCFGS:
3021 if (!vmx_mpx_supported())
3022 continue;
3023 break;
3024 default:
3025 break;
3026 }
3027
3028 if (j < i)
3029 shadow_read_write_fields[j] =
3030 shadow_read_write_fields[i];
3031 j++;
3032 }
3033 max_shadow_read_write_fields = j;
3034
3035 /* shadowed fields guest access without vmexit */
3036 for (i = 0; i < max_shadow_read_write_fields; i++) {
3037 clear_bit(shadow_read_write_fields[i],
3038 vmx_vmwrite_bitmap);
3039 clear_bit(shadow_read_write_fields[i],
3040 vmx_vmread_bitmap);
3041 }
3042 for (i = 0; i < max_shadow_read_only_fields; i++)
3043 clear_bit(shadow_read_only_fields[i],
3044 vmx_vmread_bitmap);
3045}
3046
3012static __init int alloc_kvm_area(void) 3047static __init int alloc_kvm_area(void)
3013{ 3048{
3014 int cpu; 3049 int cpu;
@@ -3039,6 +3074,8 @@ static __init int hardware_setup(void)
3039 enable_vpid = 0; 3074 enable_vpid = 0;
3040 if (!cpu_has_vmx_shadow_vmcs()) 3075 if (!cpu_has_vmx_shadow_vmcs())
3041 enable_shadow_vmcs = 0; 3076 enable_shadow_vmcs = 0;
3077 if (enable_shadow_vmcs)
3078 init_vmcs_shadow_fields();
3042 3079
3043 if (!cpu_has_vmx_ept() || 3080 if (!cpu_has_vmx_ept() ||
3044 !cpu_has_vmx_ept_4levels()) { 3081 !cpu_has_vmx_ept_4levels()) {
@@ -3484,13 +3521,14 @@ static int vmx_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4)
3484 hw_cr4 &= ~X86_CR4_PAE; 3521 hw_cr4 &= ~X86_CR4_PAE;
3485 hw_cr4 |= X86_CR4_PSE; 3522 hw_cr4 |= X86_CR4_PSE;
3486 /* 3523 /*
3487 * SMEP is disabled if CPU is in non-paging mode in 3524 * SMEP/SMAP is disabled if CPU is in non-paging mode
3488 * hardware. However KVM always uses paging mode to 3525 * in hardware. However KVM always uses paging mode to
3489 * emulate guest non-paging mode with TDP. 3526 * emulate guest non-paging mode with TDP.
3490 * To emulate this behavior, SMEP needs to be manually 3527 * To emulate this behavior, SMEP/SMAP needs to be
3491 * disabled when guest switches to non-paging mode. 3528 * manually disabled when guest switches to non-paging
3529 * mode.
3492 */ 3530 */
3493 hw_cr4 &= ~X86_CR4_SMEP; 3531 hw_cr4 &= ~(X86_CR4_SMEP | X86_CR4_SMAP);
3494 } else if (!(cr4 & X86_CR4_PAE)) { 3532 } else if (!(cr4 & X86_CR4_PAE)) {
3495 hw_cr4 &= ~X86_CR4_PAE; 3533 hw_cr4 &= ~X86_CR4_PAE;
3496 } 3534 }
@@ -7740,7 +7778,8 @@ static void prepare_vmcs02(struct kvm_vcpu *vcpu, struct vmcs12 *vmcs12)
7740 7778
7741 exec_control = vmcs12->pin_based_vm_exec_control; 7779 exec_control = vmcs12->pin_based_vm_exec_control;
7742 exec_control |= vmcs_config.pin_based_exec_ctrl; 7780 exec_control |= vmcs_config.pin_based_exec_ctrl;
7743 exec_control &= ~PIN_BASED_VMX_PREEMPTION_TIMER; 7781 exec_control &= ~(PIN_BASED_VMX_PREEMPTION_TIMER |
7782 PIN_BASED_POSTED_INTR);
7744 vmcs_write32(PIN_BASED_VM_EXEC_CONTROL, exec_control); 7783 vmcs_write32(PIN_BASED_VM_EXEC_CONTROL, exec_control);
7745 7784
7746 vmx->nested.preemption_timer_expired = false; 7785 vmx->nested.preemption_timer_expired = false;
@@ -7777,7 +7816,9 @@ static void prepare_vmcs02(struct kvm_vcpu *vcpu, struct vmcs12 *vmcs12)
7777 if (!vmx->rdtscp_enabled) 7816 if (!vmx->rdtscp_enabled)
7778 exec_control &= ~SECONDARY_EXEC_RDTSCP; 7817 exec_control &= ~SECONDARY_EXEC_RDTSCP;
7779 /* Take the following fields only from vmcs12 */ 7818 /* Take the following fields only from vmcs12 */
7780 exec_control &= ~SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES; 7819 exec_control &= ~(SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES |
7820 SECONDARY_EXEC_VIRTUAL_INTR_DELIVERY |
7821 SECONDARY_EXEC_APIC_REGISTER_VIRT);
7781 if (nested_cpu_has(vmcs12, 7822 if (nested_cpu_has(vmcs12,
7782 CPU_BASED_ACTIVATE_SECONDARY_CONTROLS)) 7823 CPU_BASED_ACTIVATE_SECONDARY_CONTROLS))
7783 exec_control |= vmcs12->secondary_vm_exec_control; 7824 exec_control |= vmcs12->secondary_vm_exec_control;
@@ -8802,14 +8843,6 @@ static int __init vmx_init(void)
8802 8843
8803 memset(vmx_vmread_bitmap, 0xff, PAGE_SIZE); 8844 memset(vmx_vmread_bitmap, 0xff, PAGE_SIZE);
8804 memset(vmx_vmwrite_bitmap, 0xff, PAGE_SIZE); 8845 memset(vmx_vmwrite_bitmap, 0xff, PAGE_SIZE);
8805 /* shadowed read/write fields */
8806 for (i = 0; i < max_shadow_read_write_fields; i++) {
8807 clear_bit(shadow_read_write_fields[i], vmx_vmwrite_bitmap);
8808 clear_bit(shadow_read_write_fields[i], vmx_vmread_bitmap);
8809 }
8810 /* shadowed read only fields */
8811 for (i = 0; i < max_shadow_read_only_fields; i++)
8812 clear_bit(shadow_read_only_fields[i], vmx_vmread_bitmap);
8813 8846
8814 /* 8847 /*
8815 * Allow direct access to the PC debug port (it is often used for I/O 8848 * Allow direct access to the PC debug port (it is often used for I/O
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 9d1b5cd4d34c..20316c67b824 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -106,6 +106,8 @@ EXPORT_SYMBOL_GPL(kvm_max_guest_tsc_khz);
106static u32 tsc_tolerance_ppm = 250; 106static u32 tsc_tolerance_ppm = 250;
107module_param(tsc_tolerance_ppm, uint, S_IRUGO | S_IWUSR); 107module_param(tsc_tolerance_ppm, uint, S_IRUGO | S_IWUSR);
108 108
109static bool backwards_tsc_observed = false;
110
109#define KVM_NR_SHARED_MSRS 16 111#define KVM_NR_SHARED_MSRS 16
110 112
111struct kvm_shared_msrs_global { 113struct kvm_shared_msrs_global {
@@ -280,7 +282,7 @@ int kvm_set_apic_base(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
280} 282}
281EXPORT_SYMBOL_GPL(kvm_set_apic_base); 283EXPORT_SYMBOL_GPL(kvm_set_apic_base);
282 284
283asmlinkage void kvm_spurious_fault(void) 285asmlinkage __visible void kvm_spurious_fault(void)
284{ 286{
285 /* Fault while not rebooting. We want the trace. */ 287 /* Fault while not rebooting. We want the trace. */
286 BUG(); 288 BUG();
@@ -652,6 +654,9 @@ int kvm_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4)
652 if (!guest_cpuid_has_smep(vcpu) && (cr4 & X86_CR4_SMEP)) 654 if (!guest_cpuid_has_smep(vcpu) && (cr4 & X86_CR4_SMEP))
653 return 1; 655 return 1;
654 656
657 if (!guest_cpuid_has_smap(vcpu) && (cr4 & X86_CR4_SMAP))
658 return 1;
659
655 if (!guest_cpuid_has_fsgsbase(vcpu) && (cr4 & X86_CR4_FSGSBASE)) 660 if (!guest_cpuid_has_fsgsbase(vcpu) && (cr4 & X86_CR4_FSGSBASE))
656 return 1; 661 return 1;
657 662
@@ -680,6 +685,9 @@ int kvm_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4)
680 (!(cr4 & X86_CR4_PCIDE) && (old_cr4 & X86_CR4_PCIDE))) 685 (!(cr4 & X86_CR4_PCIDE) && (old_cr4 & X86_CR4_PCIDE)))
681 kvm_mmu_reset_context(vcpu); 686 kvm_mmu_reset_context(vcpu);
682 687
688 if ((cr4 ^ old_cr4) & X86_CR4_SMAP)
689 update_permission_bitmask(vcpu, vcpu->arch.walk_mmu, false);
690
683 if ((cr4 ^ old_cr4) & X86_CR4_OSXSAVE) 691 if ((cr4 ^ old_cr4) & X86_CR4_OSXSAVE)
684 kvm_update_cpuid(vcpu); 692 kvm_update_cpuid(vcpu);
685 693
@@ -1117,7 +1125,6 @@ static inline u64 get_kernel_ns(void)
1117{ 1125{
1118 struct timespec ts; 1126 struct timespec ts;
1119 1127
1120 WARN_ON(preemptible());
1121 ktime_get_ts(&ts); 1128 ktime_get_ts(&ts);
1122 monotonic_to_bootbased(&ts); 1129 monotonic_to_bootbased(&ts);
1123 return timespec_to_ns(&ts); 1130 return timespec_to_ns(&ts);
@@ -1481,7 +1488,8 @@ static void pvclock_update_vm_gtod_copy(struct kvm *kvm)
1481 &ka->master_kernel_ns, 1488 &ka->master_kernel_ns,
1482 &ka->master_cycle_now); 1489 &ka->master_cycle_now);
1483 1490
1484 ka->use_master_clock = host_tsc_clocksource & vcpus_matched; 1491 ka->use_master_clock = host_tsc_clocksource && vcpus_matched
1492 && !backwards_tsc_observed;
1485 1493
1486 if (ka->use_master_clock) 1494 if (ka->use_master_clock)
1487 atomic_set(&kvm_guest_has_master_clock, 1); 1495 atomic_set(&kvm_guest_has_master_clock, 1);
@@ -4164,7 +4172,8 @@ static int vcpu_mmio_gva_to_gpa(struct kvm_vcpu *vcpu, unsigned long gva,
4164 | (write ? PFERR_WRITE_MASK : 0); 4172 | (write ? PFERR_WRITE_MASK : 0);
4165 4173
4166 if (vcpu_match_mmio_gva(vcpu, gva) 4174 if (vcpu_match_mmio_gva(vcpu, gva)
4167 && !permission_fault(vcpu->arch.walk_mmu, vcpu->arch.access, access)) { 4175 && !permission_fault(vcpu, vcpu->arch.walk_mmu,
4176 vcpu->arch.access, access)) {
4168 *gpa = vcpu->arch.mmio_gfn << PAGE_SHIFT | 4177 *gpa = vcpu->arch.mmio_gfn << PAGE_SHIFT |
4169 (gva & (PAGE_SIZE - 1)); 4178 (gva & (PAGE_SIZE - 1));
4170 trace_vcpu_match_mmio(gva, *gpa, write, false); 4179 trace_vcpu_match_mmio(gva, *gpa, write, false);
@@ -6939,6 +6948,7 @@ int kvm_arch_hardware_enable(void *garbage)
6939 */ 6948 */
6940 if (backwards_tsc) { 6949 if (backwards_tsc) {
6941 u64 delta_cyc = max_tsc - local_tsc; 6950 u64 delta_cyc = max_tsc - local_tsc;
6951 backwards_tsc_observed = true;
6942 list_for_each_entry(kvm, &vm_list, vm_list) { 6952 list_for_each_entry(kvm, &vm_list, vm_list) {
6943 kvm_for_each_vcpu(i, vcpu, kvm) { 6953 kvm_for_each_vcpu(i, vcpu, kvm) {
6944 vcpu->arch.tsc_offset_adjustment += delta_cyc; 6954 vcpu->arch.tsc_offset_adjustment += delta_cyc;
diff --git a/arch/x86/lguest/boot.c b/arch/x86/lguest/boot.c
index ad1fb5f53925..aae94132bc24 100644
--- a/arch/x86/lguest/boot.c
+++ b/arch/x86/lguest/boot.c
@@ -233,13 +233,13 @@ static void lguest_end_context_switch(struct task_struct *next)
233 * flags word contains all kind of stuff, but in practice Linux only cares 233 * flags word contains all kind of stuff, but in practice Linux only cares
234 * about the interrupt flag. Our "save_flags()" just returns that. 234 * about the interrupt flag. Our "save_flags()" just returns that.
235 */ 235 */
236asmlinkage unsigned long lguest_save_fl(void) 236asmlinkage __visible unsigned long lguest_save_fl(void)
237{ 237{
238 return lguest_data.irq_enabled; 238 return lguest_data.irq_enabled;
239} 239}
240 240
241/* Interrupts go off... */ 241/* Interrupts go off... */
242asmlinkage void lguest_irq_disable(void) 242asmlinkage __visible void lguest_irq_disable(void)
243{ 243{
244 lguest_data.irq_enabled = 0; 244 lguest_data.irq_enabled = 0;
245} 245}
diff --git a/arch/x86/lib/msr.c b/arch/x86/lib/msr.c
index db9db446b71a..43623739c7cf 100644
--- a/arch/x86/lib/msr.c
+++ b/arch/x86/lib/msr.c
@@ -76,7 +76,7 @@ static inline int __flip_bit(u32 msr, u8 bit, bool set)
76 if (m1.q == m.q) 76 if (m1.q == m.q)
77 return 0; 77 return 0;
78 78
79 err = msr_write(msr, &m); 79 err = msr_write(msr, &m1);
80 if (err) 80 if (err)
81 return err; 81 return err;
82 82
diff --git a/arch/x86/math-emu/errors.c b/arch/x86/math-emu/errors.c
index a5449089cd9f..9e6545f269e5 100644
--- a/arch/x86/math-emu/errors.c
+++ b/arch/x86/math-emu/errors.c
@@ -302,7 +302,7 @@ static struct {
302 0x242 in div_Xsig.S 302 0x242 in div_Xsig.S
303 */ 303 */
304 304
305asmlinkage void FPU_exception(int n) 305asmlinkage __visible void FPU_exception(int n)
306{ 306{
307 int i, int_type; 307 int i, int_type;
308 308
@@ -492,7 +492,7 @@ int real_2op_NaN(FPU_REG const *b, u_char tagb,
492 492
493/* Invalid arith operation on Valid registers */ 493/* Invalid arith operation on Valid registers */
494/* Returns < 0 if the exception is unmasked */ 494/* Returns < 0 if the exception is unmasked */
495asmlinkage int arith_invalid(int deststnr) 495asmlinkage __visible int arith_invalid(int deststnr)
496{ 496{
497 497
498 EXCEPTION(EX_Invalid); 498 EXCEPTION(EX_Invalid);
@@ -507,7 +507,7 @@ asmlinkage int arith_invalid(int deststnr)
507} 507}
508 508
509/* Divide a finite number by zero */ 509/* Divide a finite number by zero */
510asmlinkage int FPU_divide_by_zero(int deststnr, u_char sign) 510asmlinkage __visible int FPU_divide_by_zero(int deststnr, u_char sign)
511{ 511{
512 FPU_REG *dest = &st(deststnr); 512 FPU_REG *dest = &st(deststnr);
513 int tag = TAG_Valid; 513 int tag = TAG_Valid;
@@ -539,7 +539,7 @@ int set_precision_flag(int flags)
539} 539}
540 540
541/* This may be called often, so keep it lean */ 541/* This may be called often, so keep it lean */
542asmlinkage void set_precision_flag_up(void) 542asmlinkage __visible void set_precision_flag_up(void)
543{ 543{
544 if (control_word & CW_Precision) 544 if (control_word & CW_Precision)
545 partial_status |= (SW_Precision | SW_C1); /* The masked response */ 545 partial_status |= (SW_Precision | SW_C1); /* The masked response */
@@ -548,7 +548,7 @@ asmlinkage void set_precision_flag_up(void)
548} 548}
549 549
550/* This may be called often, so keep it lean */ 550/* This may be called often, so keep it lean */
551asmlinkage void set_precision_flag_down(void) 551asmlinkage __visible void set_precision_flag_down(void)
552{ 552{
553 if (control_word & CW_Precision) { /* The masked response */ 553 if (control_word & CW_Precision) { /* The masked response */
554 partial_status &= ~SW_C1; 554 partial_status &= ~SW_C1;
@@ -557,7 +557,7 @@ asmlinkage void set_precision_flag_down(void)
557 EXCEPTION(EX_Precision); 557 EXCEPTION(EX_Precision);
558} 558}
559 559
560asmlinkage int denormal_operand(void) 560asmlinkage __visible int denormal_operand(void)
561{ 561{
562 if (control_word & CW_Denormal) { /* The masked response */ 562 if (control_word & CW_Denormal) { /* The masked response */
563 partial_status |= SW_Denorm_Op; 563 partial_status |= SW_Denorm_Op;
@@ -568,7 +568,7 @@ asmlinkage int denormal_operand(void)
568 } 568 }
569} 569}
570 570
571asmlinkage int arith_overflow(FPU_REG *dest) 571asmlinkage __visible int arith_overflow(FPU_REG *dest)
572{ 572{
573 int tag = TAG_Valid; 573 int tag = TAG_Valid;
574 574
@@ -596,7 +596,7 @@ asmlinkage int arith_overflow(FPU_REG *dest)
596 596
597} 597}
598 598
599asmlinkage int arith_underflow(FPU_REG *dest) 599asmlinkage __visible int arith_underflow(FPU_REG *dest)
600{ 600{
601 int tag = TAG_Valid; 601 int tag = TAG_Valid;
602 602
diff --git a/arch/x86/net/bpf_jit_comp.c b/arch/x86/net/bpf_jit_comp.c
index dc017735bb91..6d5663a599a7 100644
--- a/arch/x86/net/bpf_jit_comp.c
+++ b/arch/x86/net/bpf_jit_comp.c
@@ -171,7 +171,7 @@ static struct bpf_binary_header *bpf_alloc_binary(unsigned int proglen,
171 memset(header, 0xcc, sz); /* fill whole space with int3 instructions */ 171 memset(header, 0xcc, sz); /* fill whole space with int3 instructions */
172 172
173 header->pages = sz / PAGE_SIZE; 173 header->pages = sz / PAGE_SIZE;
174 hole = sz - (proglen + sizeof(*header)); 174 hole = min(sz - (proglen + sizeof(*header)), PAGE_SIZE - sizeof(*header));
175 175
176 /* insert a random number of int3 instructions before BPF code */ 176 /* insert a random number of int3 instructions before BPF code */
177 *image_ptr = &header->image[prandom_u32() % hole]; 177 *image_ptr = &header->image[prandom_u32() % hole];
diff --git a/arch/x86/platform/efi/early_printk.c b/arch/x86/platform/efi/early_printk.c
index 81b506d5befd..524142117296 100644
--- a/arch/x86/platform/efi/early_printk.c
+++ b/arch/x86/platform/efi/early_printk.c
@@ -14,48 +14,92 @@
14 14
15static const struct font_desc *font; 15static const struct font_desc *font;
16static u32 efi_x, efi_y; 16static u32 efi_x, efi_y;
17static void *efi_fb;
18static bool early_efi_keep;
17 19
18static __init void early_efi_clear_scanline(unsigned int y) 20/*
21 * efi earlyprintk need use early_ioremap to map the framebuffer.
22 * But early_ioremap is not usable for earlyprintk=efi,keep, ioremap should
23 * be used instead. ioremap will be available after paging_init() which is
24 * earlier than initcall callbacks. Thus adding this early initcall function
25 * early_efi_map_fb to map the whole efi framebuffer.
26 */
27static __init int early_efi_map_fb(void)
19{ 28{
20 unsigned long base, *dst; 29 unsigned long base, size;
21 u16 len; 30
31 if (!early_efi_keep)
32 return 0;
22 33
23 base = boot_params.screen_info.lfb_base; 34 base = boot_params.screen_info.lfb_base;
24 len = boot_params.screen_info.lfb_linelength; 35 size = boot_params.screen_info.lfb_size;
36 efi_fb = ioremap(base, size);
37
38 return efi_fb ? 0 : -ENOMEM;
39}
40early_initcall(early_efi_map_fb);
41
42/*
43 * early_efi_map maps efi framebuffer region [start, start + len -1]
44 * In case earlyprintk=efi,keep we have the whole framebuffer mapped already
45 * so just return the offset efi_fb + start.
46 */
47static __init_refok void *early_efi_map(unsigned long start, unsigned long len)
48{
49 unsigned long base;
50
51 base = boot_params.screen_info.lfb_base;
52
53 if (efi_fb)
54 return (efi_fb + start);
55 else
56 return early_ioremap(base + start, len);
57}
25 58
26 dst = early_ioremap(base + y*len, len); 59static __init_refok void early_efi_unmap(void *addr, unsigned long len)
60{
61 if (!efi_fb)
62 early_iounmap(addr, len);
63}
64
65static void early_efi_clear_scanline(unsigned int y)
66{
67 unsigned long *dst;
68 u16 len;
69
70 len = boot_params.screen_info.lfb_linelength;
71 dst = early_efi_map(y*len, len);
27 if (!dst) 72 if (!dst)
28 return; 73 return;
29 74
30 memset(dst, 0, len); 75 memset(dst, 0, len);
31 early_iounmap(dst, len); 76 early_efi_unmap(dst, len);
32} 77}
33 78
34static __init void early_efi_scroll_up(void) 79static void early_efi_scroll_up(void)
35{ 80{
36 unsigned long base, *dst, *src; 81 unsigned long *dst, *src;
37 u16 len; 82 u16 len;
38 u32 i, height; 83 u32 i, height;
39 84
40 base = boot_params.screen_info.lfb_base;
41 len = boot_params.screen_info.lfb_linelength; 85 len = boot_params.screen_info.lfb_linelength;
42 height = boot_params.screen_info.lfb_height; 86 height = boot_params.screen_info.lfb_height;
43 87
44 for (i = 0; i < height - font->height; i++) { 88 for (i = 0; i < height - font->height; i++) {
45 dst = early_ioremap(base + i*len, len); 89 dst = early_efi_map(i*len, len);
46 if (!dst) 90 if (!dst)
47 return; 91 return;
48 92
49 src = early_ioremap(base + (i + font->height) * len, len); 93 src = early_efi_map((i + font->height) * len, len);
50 if (!src) { 94 if (!src) {
51 early_iounmap(dst, len); 95 early_efi_unmap(dst, len);
52 return; 96 return;
53 } 97 }
54 98
55 memmove(dst, src, len); 99 memmove(dst, src, len);
56 100
57 early_iounmap(src, len); 101 early_efi_unmap(src, len);
58 early_iounmap(dst, len); 102 early_efi_unmap(dst, len);
59 } 103 }
60} 104}
61 105
@@ -79,16 +123,14 @@ static void early_efi_write_char(u32 *dst, unsigned char c, unsigned int h)
79 } 123 }
80} 124}
81 125
82static __init void 126static void
83early_efi_write(struct console *con, const char *str, unsigned int num) 127early_efi_write(struct console *con, const char *str, unsigned int num)
84{ 128{
85 struct screen_info *si; 129 struct screen_info *si;
86 unsigned long base;
87 unsigned int len; 130 unsigned int len;
88 const char *s; 131 const char *s;
89 void *dst; 132 void *dst;
90 133
91 base = boot_params.screen_info.lfb_base;
92 si = &boot_params.screen_info; 134 si = &boot_params.screen_info;
93 len = si->lfb_linelength; 135 len = si->lfb_linelength;
94 136
@@ -109,7 +151,7 @@ early_efi_write(struct console *con, const char *str, unsigned int num)
109 for (h = 0; h < font->height; h++) { 151 for (h = 0; h < font->height; h++) {
110 unsigned int n, x; 152 unsigned int n, x;
111 153
112 dst = early_ioremap(base + (efi_y + h) * len, len); 154 dst = early_efi_map((efi_y + h) * len, len);
113 if (!dst) 155 if (!dst)
114 return; 156 return;
115 157
@@ -123,7 +165,7 @@ early_efi_write(struct console *con, const char *str, unsigned int num)
123 s++; 165 s++;
124 } 166 }
125 167
126 early_iounmap(dst, len); 168 early_efi_unmap(dst, len);
127 } 169 }
128 170
129 num -= count; 171 num -= count;
@@ -179,6 +221,9 @@ static __init int early_efi_setup(struct console *con, char *options)
179 for (i = 0; i < (yres - efi_y) / font->height; i++) 221 for (i = 0; i < (yres - efi_y) / font->height; i++)
180 early_efi_scroll_up(); 222 early_efi_scroll_up();
181 223
224 /* early_console_register will unset CON_BOOT in case ,keep */
225 if (!(con->flags & CON_BOOT))
226 early_efi_keep = true;
182 return 0; 227 return 0;
183} 228}
184 229
diff --git a/arch/x86/platform/olpc/olpc-xo1-pm.c b/arch/x86/platform/olpc/olpc-xo1-pm.c
index ff0174dda810..a9acde72d4ed 100644
--- a/arch/x86/platform/olpc/olpc-xo1-pm.c
+++ b/arch/x86/platform/olpc/olpc-xo1-pm.c
@@ -75,7 +75,7 @@ static int xo1_power_state_enter(suspend_state_t pm_state)
75 return 0; 75 return 0;
76} 76}
77 77
78asmlinkage int xo1_do_sleep(u8 sleep_state) 78asmlinkage __visible int xo1_do_sleep(u8 sleep_state)
79{ 79{
80 void *pgd_addr = __va(read_cr3()); 80 void *pgd_addr = __va(read_cr3());
81 81
diff --git a/arch/x86/power/hibernate_64.c b/arch/x86/power/hibernate_64.c
index 304fca20d96e..35e2bb6c0f37 100644
--- a/arch/x86/power/hibernate_64.c
+++ b/arch/x86/power/hibernate_64.c
@@ -23,7 +23,7 @@
23extern __visible const void __nosave_begin, __nosave_end; 23extern __visible const void __nosave_begin, __nosave_end;
24 24
25/* Defined in hibernate_asm_64.S */ 25/* Defined in hibernate_asm_64.S */
26extern asmlinkage int restore_image(void); 26extern asmlinkage __visible int restore_image(void);
27 27
28/* 28/*
29 * Address to jump to in the last phase of restore in order to get to the image 29 * Address to jump to in the last phase of restore in order to get to the image
diff --git a/arch/x86/syscalls/Makefile b/arch/x86/syscalls/Makefile
index f325af26107c..3323c2745248 100644
--- a/arch/x86/syscalls/Makefile
+++ b/arch/x86/syscalls/Makefile
@@ -54,5 +54,7 @@ syshdr-$(CONFIG_X86_64) += syscalls_64.h
54 54
55targets += $(uapisyshdr-y) $(syshdr-y) 55targets += $(uapisyshdr-y) $(syshdr-y)
56 56
57PHONY += all
57all: $(addprefix $(uapi)/,$(uapisyshdr-y)) 58all: $(addprefix $(uapi)/,$(uapisyshdr-y))
58all: $(addprefix $(out)/,$(syshdr-y)) 59all: $(addprefix $(out)/,$(syshdr-y))
60 @:
diff --git a/arch/x86/syscalls/syscall_32.tbl b/arch/x86/syscalls/syscall_32.tbl
index 96bc506ac6de..d6b867921612 100644
--- a/arch/x86/syscalls/syscall_32.tbl
+++ b/arch/x86/syscalls/syscall_32.tbl
@@ -359,3 +359,4 @@
359350 i386 finit_module sys_finit_module 359350 i386 finit_module sys_finit_module
360351 i386 sched_setattr sys_sched_setattr 360351 i386 sched_setattr sys_sched_setattr
361352 i386 sched_getattr sys_sched_getattr 361352 i386 sched_getattr sys_sched_getattr
362353 i386 renameat2 sys_renameat2
diff --git a/arch/x86/tools/Makefile b/arch/x86/tools/Makefile
index e8120346903b..604a37efd4d5 100644
--- a/arch/x86/tools/Makefile
+++ b/arch/x86/tools/Makefile
@@ -40,4 +40,6 @@ $(obj)/insn_sanity.o: $(srctree)/arch/x86/lib/insn.c $(srctree)/arch/x86/lib/ina
40HOST_EXTRACFLAGS += -I$(srctree)/tools/include 40HOST_EXTRACFLAGS += -I$(srctree)/tools/include
41hostprogs-y += relocs 41hostprogs-y += relocs
42relocs-objs := relocs_32.o relocs_64.o relocs_common.o 42relocs-objs := relocs_32.o relocs_64.o relocs_common.o
43PHONY += relocs
43relocs: $(obj)/relocs 44relocs: $(obj)/relocs
45 @:
diff --git a/arch/x86/vdso/vdso-layout.lds.S b/arch/x86/vdso/vdso-layout.lds.S
index 2e263f367b13..9df017ab2285 100644
--- a/arch/x86/vdso/vdso-layout.lds.S
+++ b/arch/x86/vdso/vdso-layout.lds.S
@@ -9,12 +9,9 @@ SECTIONS
9#ifdef BUILD_VDSO32 9#ifdef BUILD_VDSO32
10#include <asm/vdso32.h> 10#include <asm/vdso32.h>
11 11
12 .hpet_sect : { 12 hpet_page = . - VDSO_OFFSET(VDSO_HPET_PAGE);
13 hpet_page = . - VDSO_OFFSET(VDSO_HPET_PAGE);
14 } :text :hpet_sect
15 13
16 .vvar_sect : { 14 vvar = . - VDSO_OFFSET(VDSO_VVAR_PAGE);
17 vvar = . - VDSO_OFFSET(VDSO_VVAR_PAGE);
18 15
19 /* Place all vvars at the offsets in asm/vvar.h. */ 16 /* Place all vvars at the offsets in asm/vvar.h. */
20#define EMIT_VVAR(name, offset) vvar_ ## name = vvar + offset; 17#define EMIT_VVAR(name, offset) vvar_ ## name = vvar + offset;
@@ -22,7 +19,6 @@ SECTIONS
22#include <asm/vvar.h> 19#include <asm/vvar.h>
23#undef __VVAR_KERNEL_LDS 20#undef __VVAR_KERNEL_LDS
24#undef EMIT_VVAR 21#undef EMIT_VVAR
25 } :text :vvar_sect
26#endif 22#endif
27 . = SIZEOF_HEADERS; 23 . = SIZEOF_HEADERS;
28 24
@@ -61,7 +57,12 @@ SECTIONS
61 */ 57 */
62 . = ALIGN(0x100); 58 . = ALIGN(0x100);
63 59
64 .text : { *(.text*) } :text =0x90909090 60 .text : { *(.text*) } :text =0x90909090,
61
62 /*
63 * The comma above works around a bug in gold:
64 * https://sourceware.org/bugzilla/show_bug.cgi?id=16804
65 */
65 66
66 /DISCARD/ : { 67 /DISCARD/ : {
67 *(.discard) 68 *(.discard)
@@ -84,8 +85,4 @@ PHDRS
84 dynamic PT_DYNAMIC FLAGS(4); /* PF_R */ 85 dynamic PT_DYNAMIC FLAGS(4); /* PF_R */
85 note PT_NOTE FLAGS(4); /* PF_R */ 86 note PT_NOTE FLAGS(4); /* PF_R */
86 eh_frame_hdr PT_GNU_EH_FRAME; 87 eh_frame_hdr PT_GNU_EH_FRAME;
87#ifdef BUILD_VDSO32
88 vvar_sect PT_NULL FLAGS(4); /* PF_R */
89 hpet_sect PT_NULL FLAGS(4); /* PF_R */
90#endif
91} 88}
diff --git a/arch/x86/vdso/vdso32-setup.c b/arch/x86/vdso/vdso32-setup.c
index 00348980a3a6..e1f220e3ca68 100644
--- a/arch/x86/vdso/vdso32-setup.c
+++ b/arch/x86/vdso/vdso32-setup.c
@@ -39,6 +39,7 @@
39#ifdef CONFIG_X86_64 39#ifdef CONFIG_X86_64
40#define vdso_enabled sysctl_vsyscall32 40#define vdso_enabled sysctl_vsyscall32
41#define arch_setup_additional_pages syscall32_setup_pages 41#define arch_setup_additional_pages syscall32_setup_pages
42extern int sysctl_ldt16;
42#endif 43#endif
43 44
44/* 45/*
@@ -249,6 +250,13 @@ static struct ctl_table abi_table2[] = {
249 .mode = 0644, 250 .mode = 0644,
250 .proc_handler = proc_dointvec 251 .proc_handler = proc_dointvec
251 }, 252 },
253 {
254 .procname = "ldt16",
255 .data = &sysctl_ldt16,
256 .maxlen = sizeof(int),
257 .mode = 0644,
258 .proc_handler = proc_dointvec
259 },
252 {} 260 {}
253}; 261};
254 262
diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c
index 201d09a7c46b..c34bfc4bbe7f 100644
--- a/arch/x86/xen/enlighten.c
+++ b/arch/x86/xen/enlighten.c
@@ -1515,7 +1515,7 @@ static void __init xen_pvh_early_guest_init(void)
1515} 1515}
1516 1516
1517/* First C function to be called on Xen boot */ 1517/* First C function to be called on Xen boot */
1518asmlinkage void __init xen_start_kernel(void) 1518asmlinkage __visible void __init xen_start_kernel(void)
1519{ 1519{
1520 struct physdev_set_iopl set_iopl; 1520 struct physdev_set_iopl set_iopl;
1521 int rc; 1521 int rc;
diff --git a/arch/x86/xen/irq.c b/arch/x86/xen/irq.c
index 08f763de26fe..a1207cb6472a 100644
--- a/arch/x86/xen/irq.c
+++ b/arch/x86/xen/irq.c
@@ -23,7 +23,7 @@ void xen_force_evtchn_callback(void)
23 (void)HYPERVISOR_xen_version(0, NULL); 23 (void)HYPERVISOR_xen_version(0, NULL);
24} 24}
25 25
26asmlinkage unsigned long xen_save_fl(void) 26asmlinkage __visible unsigned long xen_save_fl(void)
27{ 27{
28 struct vcpu_info *vcpu; 28 struct vcpu_info *vcpu;
29 unsigned long flags; 29 unsigned long flags;
@@ -63,7 +63,7 @@ __visible void xen_restore_fl(unsigned long flags)
63} 63}
64PV_CALLEE_SAVE_REGS_THUNK(xen_restore_fl); 64PV_CALLEE_SAVE_REGS_THUNK(xen_restore_fl);
65 65
66asmlinkage void xen_irq_disable(void) 66asmlinkage __visible void xen_irq_disable(void)
67{ 67{
68 /* There's a one instruction preempt window here. We need to 68 /* There's a one instruction preempt window here. We need to
69 make sure we're don't switch CPUs between getting the vcpu 69 make sure we're don't switch CPUs between getting the vcpu
@@ -74,7 +74,7 @@ asmlinkage void xen_irq_disable(void)
74} 74}
75PV_CALLEE_SAVE_REGS_THUNK(xen_irq_disable); 75PV_CALLEE_SAVE_REGS_THUNK(xen_irq_disable);
76 76
77asmlinkage void xen_irq_enable(void) 77asmlinkage __visible void xen_irq_enable(void)
78{ 78{
79 struct vcpu_info *vcpu; 79 struct vcpu_info *vcpu;
80 80
diff --git a/arch/x86/xen/smp.c b/arch/x86/xen/smp.c
index a18eadd8bb40..7005974c3ff3 100644
--- a/arch/x86/xen/smp.c
+++ b/arch/x86/xen/smp.c
@@ -441,10 +441,11 @@ static int xen_cpu_up(unsigned int cpu, struct task_struct *idle)
441 irq_ctx_init(cpu); 441 irq_ctx_init(cpu);
442#else 442#else
443 clear_tsk_thread_flag(idle, TIF_FORK); 443 clear_tsk_thread_flag(idle, TIF_FORK);
444#endif
444 per_cpu(kernel_stack, cpu) = 445 per_cpu(kernel_stack, cpu) =
445 (unsigned long)task_stack_page(idle) - 446 (unsigned long)task_stack_page(idle) -
446 KERNEL_STACK_OFFSET + THREAD_SIZE; 447 KERNEL_STACK_OFFSET + THREAD_SIZE;
447#endif 448
448 xen_setup_runstate_info(cpu); 449 xen_setup_runstate_info(cpu);
449 xen_setup_timer(cpu); 450 xen_setup_timer(cpu);
450 xen_init_lock_cpu(cpu); 451 xen_init_lock_cpu(cpu);
diff --git a/arch/x86/xen/spinlock.c b/arch/x86/xen/spinlock.c
index 4d3acc34a998..0ba5f3b967f0 100644
--- a/arch/x86/xen/spinlock.c
+++ b/arch/x86/xen/spinlock.c
@@ -274,7 +274,7 @@ void __init xen_init_spinlocks(void)
274 printk(KERN_DEBUG "xen: PV spinlocks disabled\n"); 274 printk(KERN_DEBUG "xen: PV spinlocks disabled\n");
275 return; 275 return;
276 } 276 }
277 277 printk(KERN_DEBUG "xen: PV spinlocks enabled\n");
278 pv_lock_ops.lock_spinning = PV_CALLEE_SAVE(xen_lock_spinning); 278 pv_lock_ops.lock_spinning = PV_CALLEE_SAVE(xen_lock_spinning);
279 pv_lock_ops.unlock_kick = xen_unlock_kick; 279 pv_lock_ops.unlock_kick = xen_unlock_kick;
280} 280}
@@ -290,6 +290,9 @@ static __init int xen_init_spinlocks_jump(void)
290 if (!xen_pvspin) 290 if (!xen_pvspin)
291 return 0; 291 return 0;
292 292
293 if (!xen_domain())
294 return 0;
295
293 static_key_slow_inc(&paravirt_ticketlocks_enabled); 296 static_key_slow_inc(&paravirt_ticketlocks_enabled);
294 return 0; 297 return 0;
295} 298}
diff --git a/arch/x86/xen/xen-asm_32.S b/arch/x86/xen/xen-asm_32.S
index 33ca6e42a4ca..fd92a64d748e 100644
--- a/arch/x86/xen/xen-asm_32.S
+++ b/arch/x86/xen/xen-asm_32.S
@@ -75,6 +75,17 @@ ENDPROC(xen_sysexit)
75 * stack state in whatever form its in, we keep things simple by only 75 * stack state in whatever form its in, we keep things simple by only
76 * using a single register which is pushed/popped on the stack. 76 * using a single register which is pushed/popped on the stack.
77 */ 77 */
78
79.macro POP_FS
801:
81 popw %fs
82.pushsection .fixup, "ax"
832: movw $0, (%esp)
84 jmp 1b
85.popsection
86 _ASM_EXTABLE(1b,2b)
87.endm
88
78ENTRY(xen_iret) 89ENTRY(xen_iret)
79 /* test eflags for special cases */ 90 /* test eflags for special cases */
80 testl $(X86_EFLAGS_VM | XEN_EFLAGS_NMI), 8(%esp) 91 testl $(X86_EFLAGS_VM | XEN_EFLAGS_NMI), 8(%esp)
@@ -83,15 +94,13 @@ ENTRY(xen_iret)
83 push %eax 94 push %eax
84 ESP_OFFSET=4 # bytes pushed onto stack 95 ESP_OFFSET=4 # bytes pushed onto stack
85 96
86 /* 97 /* Store vcpu_info pointer for easy access */
87 * Store vcpu_info pointer for easy access. Do it this way to
88 * avoid having to reload %fs
89 */
90#ifdef CONFIG_SMP 98#ifdef CONFIG_SMP
91 GET_THREAD_INFO(%eax) 99 pushw %fs
92 movl %ss:TI_cpu(%eax), %eax 100 movl $(__KERNEL_PERCPU), %eax
93 movl %ss:__per_cpu_offset(,%eax,4), %eax 101 movl %eax, %fs
94 mov %ss:xen_vcpu(%eax), %eax 102 movl %fs:xen_vcpu, %eax
103 POP_FS
95#else 104#else
96 movl %ss:xen_vcpu, %eax 105 movl %ss:xen_vcpu, %eax
97#endif 106#endif
generated by cgit v1.2.2 (git 2.25.0) at 2025-05-31 23:52:40 -0400