aboutsummaryrefslogtreecommitdiffstats
path: root/arch/x86
diff options
context:
space:
mode:
Diffstat (limited to 'arch/x86')
-rw-r--r--arch/x86/boot/compressed/aslr.c9
-rw-r--r--arch/x86/include/asm/efi.h1
-rw-r--r--arch/x86/include/asm/tsc.h2
-rw-r--r--arch/x86/kernel/cpu/perf_event.c11
-rw-r--r--arch/x86/kernel/cpu/perf_event.h1
-rw-r--r--arch/x86/kernel/cpu/perf_event_intel.c11
-rw-r--r--arch/x86/kernel/cpu/perf_event_intel_uncore.c10
-rw-r--r--arch/x86/kernel/cpu/perf_event_p6.c48
-rw-r--r--arch/x86/kernel/head_32.S7
-rw-r--r--arch/x86/kernel/head_64.S6
-rw-r--r--arch/x86/kernel/machine_kexec_64.c2
-rw-r--r--arch/x86/kernel/pci-dma.c4
-rw-r--r--arch/x86/kernel/setup.c10
-rw-r--r--arch/x86/kernel/tsc.c7
-rw-r--r--arch/x86/kernel/tsc_msr.c30
-rw-r--r--arch/x86/kvm/mmu.c1
-rw-r--r--arch/x86/kvm/vmx.c2
-rw-r--r--arch/x86/kvm/x86.c2
-rw-r--r--arch/x86/mm/fault.c47
-rw-r--r--arch/x86/platform/efi/efi.c20
20 files changed, 153 insertions, 78 deletions
diff --git a/arch/x86/boot/compressed/aslr.c b/arch/x86/boot/compressed/aslr.c
index 90a21f430117..4dbf967da50d 100644
--- a/arch/x86/boot/compressed/aslr.c
+++ b/arch/x86/boot/compressed/aslr.c
@@ -111,7 +111,7 @@ struct mem_vector {
111}; 111};
112 112
113#define MEM_AVOID_MAX 5 113#define MEM_AVOID_MAX 5
114struct mem_vector mem_avoid[MEM_AVOID_MAX]; 114static struct mem_vector mem_avoid[MEM_AVOID_MAX];
115 115
116static bool mem_contains(struct mem_vector *region, struct mem_vector *item) 116static bool mem_contains(struct mem_vector *region, struct mem_vector *item)
117{ 117{
@@ -180,7 +180,7 @@ static void mem_avoid_init(unsigned long input, unsigned long input_size,
180} 180}
181 181
182/* Does this memory vector overlap a known avoided area? */ 182/* Does this memory vector overlap a known avoided area? */
183bool mem_avoid_overlap(struct mem_vector *img) 183static bool mem_avoid_overlap(struct mem_vector *img)
184{ 184{
185 int i; 185 int i;
186 186
@@ -192,8 +192,9 @@ bool mem_avoid_overlap(struct mem_vector *img)
192 return false; 192 return false;
193} 193}
194 194
195unsigned long slots[CONFIG_RANDOMIZE_BASE_MAX_OFFSET / CONFIG_PHYSICAL_ALIGN]; 195static unsigned long slots[CONFIG_RANDOMIZE_BASE_MAX_OFFSET /
196unsigned long slot_max = 0; 196 CONFIG_PHYSICAL_ALIGN];
197static unsigned long slot_max;
197 198
198static void slots_append(unsigned long addr) 199static void slots_append(unsigned long addr)
199{ 200{
diff --git a/arch/x86/include/asm/efi.h b/arch/x86/include/asm/efi.h
index 3d6b9f81cc68..acd86c850414 100644
--- a/arch/x86/include/asm/efi.h
+++ b/arch/x86/include/asm/efi.h
@@ -134,6 +134,7 @@ extern void efi_setup_page_tables(void);
134extern void __init old_map_region(efi_memory_desc_t *md); 134extern void __init old_map_region(efi_memory_desc_t *md);
135extern void __init runtime_code_page_mkexec(void); 135extern void __init runtime_code_page_mkexec(void);
136extern void __init efi_runtime_mkexec(void); 136extern void __init efi_runtime_mkexec(void);
137extern void __init efi_apply_memmap_quirks(void);
137 138
138struct efi_setup_data { 139struct efi_setup_data {
139 u64 fw_vendor; 140 u64 fw_vendor;
diff --git a/arch/x86/include/asm/tsc.h b/arch/x86/include/asm/tsc.h
index 57ae63cd6ee2..94605c0e9cee 100644
--- a/arch/x86/include/asm/tsc.h
+++ b/arch/x86/include/asm/tsc.h
@@ -66,6 +66,6 @@ extern void tsc_save_sched_clock_state(void);
66extern void tsc_restore_sched_clock_state(void); 66extern void tsc_restore_sched_clock_state(void);
67 67
68/* MSR based TSC calibration for Intel Atom SoC platforms */ 68/* MSR based TSC calibration for Intel Atom SoC platforms */
69int try_msr_calibrate_tsc(unsigned long *fast_calibrate); 69unsigned long try_msr_calibrate_tsc(void);
70 70
71#endif /* _ASM_X86_TSC_H */ 71#endif /* _ASM_X86_TSC_H */
diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
index b88645191fe5..79f9f848bee4 100644
--- a/arch/x86/kernel/cpu/perf_event.c
+++ b/arch/x86/kernel/cpu/perf_event.c
@@ -1192,6 +1192,9 @@ static void x86_pmu_del(struct perf_event *event, int flags)
1192 for (i = 0; i < cpuc->n_events; i++) { 1192 for (i = 0; i < cpuc->n_events; i++) {
1193 if (event == cpuc->event_list[i]) { 1193 if (event == cpuc->event_list[i]) {
1194 1194
1195 if (i >= cpuc->n_events - cpuc->n_added)
1196 --cpuc->n_added;
1197
1195 if (x86_pmu.put_event_constraints) 1198 if (x86_pmu.put_event_constraints)
1196 x86_pmu.put_event_constraints(cpuc, event); 1199 x86_pmu.put_event_constraints(cpuc, event);
1197 1200
@@ -1521,6 +1524,8 @@ static int __init init_hw_perf_events(void)
1521 1524
1522 pr_cont("%s PMU driver.\n", x86_pmu.name); 1525 pr_cont("%s PMU driver.\n", x86_pmu.name);
1523 1526
1527 x86_pmu.attr_rdpmc = 1; /* enable userspace RDPMC usage by default */
1528
1524 for (quirk = x86_pmu.quirks; quirk; quirk = quirk->next) 1529 for (quirk = x86_pmu.quirks; quirk; quirk = quirk->next)
1525 quirk->func(); 1530 quirk->func();
1526 1531
@@ -1534,7 +1539,6 @@ static int __init init_hw_perf_events(void)
1534 __EVENT_CONSTRAINT(0, (1ULL << x86_pmu.num_counters) - 1, 1539 __EVENT_CONSTRAINT(0, (1ULL << x86_pmu.num_counters) - 1,
1535 0, x86_pmu.num_counters, 0, 0); 1540 0, x86_pmu.num_counters, 0, 0);
1536 1541
1537 x86_pmu.attr_rdpmc = 1; /* enable userspace RDPMC usage by default */
1538 x86_pmu_format_group.attrs = x86_pmu.format_attrs; 1542 x86_pmu_format_group.attrs = x86_pmu.format_attrs;
1539 1543
1540 if (x86_pmu.event_attrs) 1544 if (x86_pmu.event_attrs)
@@ -1820,9 +1824,12 @@ static ssize_t set_attr_rdpmc(struct device *cdev,
1820 if (ret) 1824 if (ret)
1821 return ret; 1825 return ret;
1822 1826
1827 if (x86_pmu.attr_rdpmc_broken)
1828 return -ENOTSUPP;
1829
1823 if (!!val != !!x86_pmu.attr_rdpmc) { 1830 if (!!val != !!x86_pmu.attr_rdpmc) {
1824 x86_pmu.attr_rdpmc = !!val; 1831 x86_pmu.attr_rdpmc = !!val;
1825 smp_call_function(change_rdpmc, (void *)val, 1); 1832 on_each_cpu(change_rdpmc, (void *)val, 1);
1826 } 1833 }
1827 1834
1828 return count; 1835 return count;
diff --git a/arch/x86/kernel/cpu/perf_event.h b/arch/x86/kernel/cpu/perf_event.h
index c1a861829d81..4972c244d0bc 100644
--- a/arch/x86/kernel/cpu/perf_event.h
+++ b/arch/x86/kernel/cpu/perf_event.h
@@ -409,6 +409,7 @@ struct x86_pmu {
409 /* 409 /*
410 * sysfs attrs 410 * sysfs attrs
411 */ 411 */
412 int attr_rdpmc_broken;
412 int attr_rdpmc; 413 int attr_rdpmc;
413 struct attribute **format_attrs; 414 struct attribute **format_attrs;
414 struct attribute **event_attrs; 415 struct attribute **event_attrs;
diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c
index 0fa4f242f050..aa333d966886 100644
--- a/arch/x86/kernel/cpu/perf_event_intel.c
+++ b/arch/x86/kernel/cpu/perf_event_intel.c
@@ -1361,10 +1361,8 @@ static int intel_pmu_handle_irq(struct pt_regs *regs)
1361 intel_pmu_disable_all(); 1361 intel_pmu_disable_all();
1362 handled = intel_pmu_drain_bts_buffer(); 1362 handled = intel_pmu_drain_bts_buffer();
1363 status = intel_pmu_get_status(); 1363 status = intel_pmu_get_status();
1364 if (!status) { 1364 if (!status)
1365 intel_pmu_enable_all(0); 1365 goto done;
1366 return handled;
1367 }
1368 1366
1369 loops = 0; 1367 loops = 0;
1370again: 1368again:
@@ -2310,10 +2308,7 @@ __init int intel_pmu_init(void)
2310 if (version > 1) 2308 if (version > 1)
2311 x86_pmu.num_counters_fixed = max((int)edx.split.num_counters_fixed, 3); 2309 x86_pmu.num_counters_fixed = max((int)edx.split.num_counters_fixed, 3);
2312 2310
2313 /* 2311 if (boot_cpu_has(X86_FEATURE_PDCM)) {
2314 * v2 and above have a perf capabilities MSR
2315 */
2316 if (version > 1) {
2317 u64 capabilities; 2312 u64 capabilities;
2318 2313
2319 rdmsrl(MSR_IA32_PERF_CAPABILITIES, capabilities); 2314 rdmsrl(MSR_IA32_PERF_CAPABILITIES, capabilities);
diff --git a/arch/x86/kernel/cpu/perf_event_intel_uncore.c b/arch/x86/kernel/cpu/perf_event_intel_uncore.c
index 29c248799ced..c88f7f4b03ee 100644
--- a/arch/x86/kernel/cpu/perf_event_intel_uncore.c
+++ b/arch/x86/kernel/cpu/perf_event_intel_uncore.c
@@ -501,8 +501,11 @@ static struct extra_reg snbep_uncore_cbox_extra_regs[] = {
501 SNBEP_CBO_EVENT_EXTRA_REG(SNBEP_CBO_PMON_CTL_TID_EN, 501 SNBEP_CBO_EVENT_EXTRA_REG(SNBEP_CBO_PMON_CTL_TID_EN,
502 SNBEP_CBO_PMON_CTL_TID_EN, 0x1), 502 SNBEP_CBO_PMON_CTL_TID_EN, 0x1),
503 SNBEP_CBO_EVENT_EXTRA_REG(0x0334, 0xffff, 0x4), 503 SNBEP_CBO_EVENT_EXTRA_REG(0x0334, 0xffff, 0x4),
504 SNBEP_CBO_EVENT_EXTRA_REG(0x4334, 0xffff, 0x6),
504 SNBEP_CBO_EVENT_EXTRA_REG(0x0534, 0xffff, 0x4), 505 SNBEP_CBO_EVENT_EXTRA_REG(0x0534, 0xffff, 0x4),
506 SNBEP_CBO_EVENT_EXTRA_REG(0x4534, 0xffff, 0x6),
505 SNBEP_CBO_EVENT_EXTRA_REG(0x0934, 0xffff, 0x4), 507 SNBEP_CBO_EVENT_EXTRA_REG(0x0934, 0xffff, 0x4),
508 SNBEP_CBO_EVENT_EXTRA_REG(0x4934, 0xffff, 0x6),
506 SNBEP_CBO_EVENT_EXTRA_REG(0x4134, 0xffff, 0x6), 509 SNBEP_CBO_EVENT_EXTRA_REG(0x4134, 0xffff, 0x6),
507 SNBEP_CBO_EVENT_EXTRA_REG(0x0135, 0xffff, 0x8), 510 SNBEP_CBO_EVENT_EXTRA_REG(0x0135, 0xffff, 0x8),
508 SNBEP_CBO_EVENT_EXTRA_REG(0x0335, 0xffff, 0x8), 511 SNBEP_CBO_EVENT_EXTRA_REG(0x0335, 0xffff, 0x8),
@@ -1178,10 +1181,15 @@ static struct extra_reg ivt_uncore_cbox_extra_regs[] = {
1178 SNBEP_CBO_EVENT_EXTRA_REG(SNBEP_CBO_PMON_CTL_TID_EN, 1181 SNBEP_CBO_EVENT_EXTRA_REG(SNBEP_CBO_PMON_CTL_TID_EN,
1179 SNBEP_CBO_PMON_CTL_TID_EN, 0x1), 1182 SNBEP_CBO_PMON_CTL_TID_EN, 0x1),
1180 SNBEP_CBO_EVENT_EXTRA_REG(0x1031, 0x10ff, 0x2), 1183 SNBEP_CBO_EVENT_EXTRA_REG(0x1031, 0x10ff, 0x2),
1184 SNBEP_CBO_EVENT_EXTRA_REG(0x1134, 0xffff, 0x4),
1185 SNBEP_CBO_EVENT_EXTRA_REG(0x4134, 0xffff, 0xc),
1186 SNBEP_CBO_EVENT_EXTRA_REG(0x5134, 0xffff, 0xc),
1181 SNBEP_CBO_EVENT_EXTRA_REG(0x0334, 0xffff, 0x4), 1187 SNBEP_CBO_EVENT_EXTRA_REG(0x0334, 0xffff, 0x4),
1188 SNBEP_CBO_EVENT_EXTRA_REG(0x4334, 0xffff, 0xc),
1182 SNBEP_CBO_EVENT_EXTRA_REG(0x0534, 0xffff, 0x4), 1189 SNBEP_CBO_EVENT_EXTRA_REG(0x0534, 0xffff, 0x4),
1190 SNBEP_CBO_EVENT_EXTRA_REG(0x4534, 0xffff, 0xc),
1183 SNBEP_CBO_EVENT_EXTRA_REG(0x0934, 0xffff, 0x4), 1191 SNBEP_CBO_EVENT_EXTRA_REG(0x0934, 0xffff, 0x4),
1184 SNBEP_CBO_EVENT_EXTRA_REG(0x4134, 0xffff, 0xc), 1192 SNBEP_CBO_EVENT_EXTRA_REG(0x4934, 0xffff, 0xc),
1185 SNBEP_CBO_EVENT_EXTRA_REG(0x0135, 0xffff, 0x10), 1193 SNBEP_CBO_EVENT_EXTRA_REG(0x0135, 0xffff, 0x10),
1186 SNBEP_CBO_EVENT_EXTRA_REG(0x0335, 0xffff, 0x10), 1194 SNBEP_CBO_EVENT_EXTRA_REG(0x0335, 0xffff, 0x10),
1187 SNBEP_CBO_EVENT_EXTRA_REG(0x2135, 0xffff, 0x10), 1195 SNBEP_CBO_EVENT_EXTRA_REG(0x2135, 0xffff, 0x10),
diff --git a/arch/x86/kernel/cpu/perf_event_p6.c b/arch/x86/kernel/cpu/perf_event_p6.c
index b1e2fe115323..7c1a0c07b607 100644
--- a/arch/x86/kernel/cpu/perf_event_p6.c
+++ b/arch/x86/kernel/cpu/perf_event_p6.c
@@ -231,31 +231,49 @@ static __initconst const struct x86_pmu p6_pmu = {
231 231
232}; 232};
233 233
234static __init void p6_pmu_rdpmc_quirk(void)
235{
236 if (boot_cpu_data.x86_mask < 9) {
237 /*
238 * PPro erratum 26; fixed in stepping 9 and above.
239 */
240 pr_warn("Userspace RDPMC support disabled due to a CPU erratum\n");
241 x86_pmu.attr_rdpmc_broken = 1;
242 x86_pmu.attr_rdpmc = 0;
243 }
244}
245
234__init int p6_pmu_init(void) 246__init int p6_pmu_init(void)
235{ 247{
248 x86_pmu = p6_pmu;
249
236 switch (boot_cpu_data.x86_model) { 250 switch (boot_cpu_data.x86_model) {
237 case 1: 251 case 1: /* Pentium Pro */
238 case 3: /* Pentium Pro */ 252 x86_add_quirk(p6_pmu_rdpmc_quirk);
239 case 5: 253 break;
240 case 6: /* Pentium II */ 254
241 case 7: 255 case 3: /* Pentium II - Klamath */
242 case 8: 256 case 5: /* Pentium II - Deschutes */
243 case 11: /* Pentium III */ 257 case 6: /* Pentium II - Mendocino */
244 case 9:
245 case 13:
246 /* Pentium M */
247 break; 258 break;
259
260 case 7: /* Pentium III - Katmai */
261 case 8: /* Pentium III - Coppermine */
262 case 10: /* Pentium III Xeon */
263 case 11: /* Pentium III - Tualatin */
264 break;
265
266 case 9: /* Pentium M - Banias */
267 case 13: /* Pentium M - Dothan */
268 break;
269
248 default: 270 default:
249 pr_cont("unsupported p6 CPU model %d ", 271 pr_cont("unsupported p6 CPU model %d ", boot_cpu_data.x86_model);
250 boot_cpu_data.x86_model);
251 return -ENODEV; 272 return -ENODEV;
252 } 273 }
253 274
254 x86_pmu = p6_pmu;
255
256 memcpy(hw_cache_event_ids, p6_hw_cache_event_ids, 275 memcpy(hw_cache_event_ids, p6_hw_cache_event_ids,
257 sizeof(hw_cache_event_ids)); 276 sizeof(hw_cache_event_ids));
258 277
259
260 return 0; 278 return 0;
261} 279}
diff --git a/arch/x86/kernel/head_32.S b/arch/x86/kernel/head_32.S
index 81ba27679f18..f36bd42d6f0c 100644
--- a/arch/x86/kernel/head_32.S
+++ b/arch/x86/kernel/head_32.S
@@ -544,6 +544,10 @@ ENDPROC(early_idt_handlers)
544 /* This is global to keep gas from relaxing the jumps */ 544 /* This is global to keep gas from relaxing the jumps */
545ENTRY(early_idt_handler) 545ENTRY(early_idt_handler)
546 cld 546 cld
547
548 cmpl $2,(%esp) # X86_TRAP_NMI
549 je is_nmi # Ignore NMI
550
547 cmpl $2,%ss:early_recursion_flag 551 cmpl $2,%ss:early_recursion_flag
548 je hlt_loop 552 je hlt_loop
549 incl %ss:early_recursion_flag 553 incl %ss:early_recursion_flag
@@ -594,8 +598,9 @@ ex_entry:
594 pop %edx 598 pop %edx
595 pop %ecx 599 pop %ecx
596 pop %eax 600 pop %eax
597 addl $8,%esp /* drop vector number and error code */
598 decl %ss:early_recursion_flag 601 decl %ss:early_recursion_flag
602is_nmi:
603 addl $8,%esp /* drop vector number and error code */
599 iret 604 iret
600ENDPROC(early_idt_handler) 605ENDPROC(early_idt_handler)
601 606
diff --git a/arch/x86/kernel/head_64.S b/arch/x86/kernel/head_64.S
index e1aabdb314c8..a468c0a65c42 100644
--- a/arch/x86/kernel/head_64.S
+++ b/arch/x86/kernel/head_64.S
@@ -343,6 +343,9 @@ early_idt_handlers:
343ENTRY(early_idt_handler) 343ENTRY(early_idt_handler)
344 cld 344 cld
345 345
346 cmpl $2,(%rsp) # X86_TRAP_NMI
347 je is_nmi # Ignore NMI
348
346 cmpl $2,early_recursion_flag(%rip) 349 cmpl $2,early_recursion_flag(%rip)
347 jz 1f 350 jz 1f
348 incl early_recursion_flag(%rip) 351 incl early_recursion_flag(%rip)
@@ -405,8 +408,9 @@ ENTRY(early_idt_handler)
405 popq %rdx 408 popq %rdx
406 popq %rcx 409 popq %rcx
407 popq %rax 410 popq %rax
408 addq $16,%rsp # drop vector number and error code
409 decl early_recursion_flag(%rip) 411 decl early_recursion_flag(%rip)
412is_nmi:
413 addq $16,%rsp # drop vector number and error code
410 INTERRUPT_RETURN 414 INTERRUPT_RETURN
411ENDPROC(early_idt_handler) 415ENDPROC(early_idt_handler)
412 416
diff --git a/arch/x86/kernel/machine_kexec_64.c b/arch/x86/kernel/machine_kexec_64.c
index 4eabc160696f..679cef0791cd 100644
--- a/arch/x86/kernel/machine_kexec_64.c
+++ b/arch/x86/kernel/machine_kexec_64.c
@@ -279,5 +279,7 @@ void arch_crash_save_vmcoreinfo(void)
279 VMCOREINFO_SYMBOL(node_data); 279 VMCOREINFO_SYMBOL(node_data);
280 VMCOREINFO_LENGTH(node_data, MAX_NUMNODES); 280 VMCOREINFO_LENGTH(node_data, MAX_NUMNODES);
281#endif 281#endif
282 vmcoreinfo_append_str("KERNELOFFSET=%lx\n",
283 (unsigned long)&_text - __START_KERNEL);
282} 284}
283 285
diff --git a/arch/x86/kernel/pci-dma.c b/arch/x86/kernel/pci-dma.c
index 872079a67e4d..f7d0672481fd 100644
--- a/arch/x86/kernel/pci-dma.c
+++ b/arch/x86/kernel/pci-dma.c
@@ -100,8 +100,10 @@ void *dma_generic_alloc_coherent(struct device *dev, size_t size,
100 flag |= __GFP_ZERO; 100 flag |= __GFP_ZERO;
101again: 101again:
102 page = NULL; 102 page = NULL;
103 if (!(flag & GFP_ATOMIC)) 103 /* CMA can be used only in the context which permits sleeping */
104 if (flag & __GFP_WAIT)
104 page = dma_alloc_from_contiguous(dev, count, get_order(size)); 105 page = dma_alloc_from_contiguous(dev, count, get_order(size));
106 /* fallback */
105 if (!page) 107 if (!page)
106 page = alloc_pages_node(dev_to_node(dev), flag, get_order(size)); 108 page = alloc_pages_node(dev_to_node(dev), flag, get_order(size));
107 if (!page) 109 if (!page)
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 06853e670354..ce72964b2f46 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -1239,14 +1239,8 @@ void __init setup_arch(char **cmdline_p)
1239 register_refined_jiffies(CLOCK_TICK_RATE); 1239 register_refined_jiffies(CLOCK_TICK_RATE);
1240 1240
1241#ifdef CONFIG_EFI 1241#ifdef CONFIG_EFI
1242 /* Once setup is done above, unmap the EFI memory map on 1242 if (efi_enabled(EFI_BOOT))
1243 * mismatched firmware/kernel archtectures since there is no 1243 efi_apply_memmap_quirks();
1244 * support for runtime services.
1245 */
1246 if (efi_enabled(EFI_BOOT) && !efi_is_native()) {
1247 pr_info("efi: Setup done, disabling due to 32/64-bit mismatch\n");
1248 efi_unmap_memmap();
1249 }
1250#endif 1244#endif
1251} 1245}
1252 1246
diff --git a/arch/x86/kernel/tsc.c b/arch/x86/kernel/tsc.c
index acb3b606613e..cfbe99f88830 100644
--- a/arch/x86/kernel/tsc.c
+++ b/arch/x86/kernel/tsc.c
@@ -653,13 +653,10 @@ unsigned long native_calibrate_tsc(void)
653 653
654 /* Calibrate TSC using MSR for Intel Atom SoCs */ 654 /* Calibrate TSC using MSR for Intel Atom SoCs */
655 local_irq_save(flags); 655 local_irq_save(flags);
656 i = try_msr_calibrate_tsc(&fast_calibrate); 656 fast_calibrate = try_msr_calibrate_tsc();
657 local_irq_restore(flags); 657 local_irq_restore(flags);
658 if (i >= 0) { 658 if (fast_calibrate)
659 if (i == 0)
660 pr_warn("Fast TSC calibration using MSR failed\n");
661 return fast_calibrate; 659 return fast_calibrate;
662 }
663 660
664 local_irq_save(flags); 661 local_irq_save(flags);
665 fast_calibrate = quick_pit_calibrate(); 662 fast_calibrate = quick_pit_calibrate();
diff --git a/arch/x86/kernel/tsc_msr.c b/arch/x86/kernel/tsc_msr.c
index 8b5434f4389f..92ae6acac8a7 100644
--- a/arch/x86/kernel/tsc_msr.c
+++ b/arch/x86/kernel/tsc_msr.c
@@ -53,7 +53,7 @@ static struct freq_desc freq_desc_tables[] = {
53 /* TNG */ 53 /* TNG */
54 { 6, 0x4a, 1, { 0, FREQ_100, FREQ_133, 0, 0, 0, 0, 0 } }, 54 { 6, 0x4a, 1, { 0, FREQ_100, FREQ_133, 0, 0, 0, 0, 0 } },
55 /* VLV2 */ 55 /* VLV2 */
56 { 6, 0x37, 1, { 0, FREQ_100, FREQ_133, FREQ_166, 0, 0, 0, 0 } }, 56 { 6, 0x37, 1, { FREQ_83, FREQ_100, FREQ_133, FREQ_166, 0, 0, 0, 0 } },
57 /* ANN */ 57 /* ANN */
58 { 6, 0x5a, 1, { FREQ_83, FREQ_100, FREQ_133, FREQ_100, 0, 0, 0, 0 } }, 58 { 6, 0x5a, 1, { FREQ_83, FREQ_100, FREQ_133, FREQ_100, 0, 0, 0, 0 } },
59}; 59};
@@ -77,21 +77,18 @@ static int match_cpu(u8 family, u8 model)
77 77
78/* 78/*
79 * Do MSR calibration only for known/supported CPUs. 79 * Do MSR calibration only for known/supported CPUs.
80 * Return values: 80 *
81 * -1: CPU is unknown/unsupported for MSR based calibration 81 * Returns the calibration value or 0 if MSR calibration failed.
82 * 0: CPU is known/supported, but calibration failed
83 * 1: CPU is known/supported, and calibration succeeded
84 */ 82 */
85int try_msr_calibrate_tsc(unsigned long *fast_calibrate) 83unsigned long try_msr_calibrate_tsc(void)
86{ 84{
87 int cpu_index;
88 u32 lo, hi, ratio, freq_id, freq; 85 u32 lo, hi, ratio, freq_id, freq;
86 unsigned long res;
87 int cpu_index;
89 88
90 cpu_index = match_cpu(boot_cpu_data.x86, boot_cpu_data.x86_model); 89 cpu_index = match_cpu(boot_cpu_data.x86, boot_cpu_data.x86_model);
91 if (cpu_index < 0) 90 if (cpu_index < 0)
92 return -1; 91 return 0;
93
94 *fast_calibrate = 0;
95 92
96 if (freq_desc_tables[cpu_index].msr_plat) { 93 if (freq_desc_tables[cpu_index].msr_plat) {
97 rdmsr(MSR_PLATFORM_INFO, lo, hi); 94 rdmsr(MSR_PLATFORM_INFO, lo, hi);
@@ -103,7 +100,7 @@ int try_msr_calibrate_tsc(unsigned long *fast_calibrate)
103 pr_info("Maximum core-clock to bus-clock ratio: 0x%x\n", ratio); 100 pr_info("Maximum core-clock to bus-clock ratio: 0x%x\n", ratio);
104 101
105 if (!ratio) 102 if (!ratio)
106 return 0; 103 goto fail;
107 104
108 /* Get FSB FREQ ID */ 105 /* Get FSB FREQ ID */
109 rdmsr(MSR_FSB_FREQ, lo, hi); 106 rdmsr(MSR_FSB_FREQ, lo, hi);
@@ -112,16 +109,19 @@ int try_msr_calibrate_tsc(unsigned long *fast_calibrate)
112 pr_info("Resolved frequency ID: %u, frequency: %u KHz\n", 109 pr_info("Resolved frequency ID: %u, frequency: %u KHz\n",
113 freq_id, freq); 110 freq_id, freq);
114 if (!freq) 111 if (!freq)
115 return 0; 112 goto fail;
116 113
117 /* TSC frequency = maximum resolved freq * maximum resolved bus ratio */ 114 /* TSC frequency = maximum resolved freq * maximum resolved bus ratio */
118 *fast_calibrate = freq * ratio; 115 res = freq * ratio;
119 pr_info("TSC runs at %lu KHz\n", *fast_calibrate); 116 pr_info("TSC runs at %lu KHz\n", res);
120 117
121#ifdef CONFIG_X86_LOCAL_APIC 118#ifdef CONFIG_X86_LOCAL_APIC
122 lapic_timer_frequency = (freq * 1000) / HZ; 119 lapic_timer_frequency = (freq * 1000) / HZ;
123 pr_info("lapic_timer_frequency = %d\n", lapic_timer_frequency); 120 pr_info("lapic_timer_frequency = %d\n", lapic_timer_frequency);
124#endif 121#endif
122 return res;
125 123
126 return 1; 124fail:
125 pr_warn("Fast TSC calibration using MSR failed\n");
126 return 0;
127} 127}
diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index e50425d0f5f7..9b531351a587 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -2672,6 +2672,7 @@ static int __direct_map(struct kvm_vcpu *vcpu, gpa_t v, int write,
2672 break; 2672 break;
2673 } 2673 }
2674 2674
2675 drop_large_spte(vcpu, iterator.sptep);
2675 if (!is_shadow_present_pte(*iterator.sptep)) { 2676 if (!is_shadow_present_pte(*iterator.sptep)) {
2676 u64 base_addr = iterator.addr; 2677 u64 base_addr = iterator.addr;
2677 2678
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index a06f101ef64b..392752834751 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -6688,7 +6688,7 @@ static bool nested_vmx_exit_handled(struct kvm_vcpu *vcpu)
6688 else if (is_page_fault(intr_info)) 6688 else if (is_page_fault(intr_info))
6689 return enable_ept; 6689 return enable_ept;
6690 else if (is_no_device(intr_info) && 6690 else if (is_no_device(intr_info) &&
6691 !(nested_read_cr0(vmcs12) & X86_CR0_TS)) 6691 !(vmcs12->guest_cr0 & X86_CR0_TS))
6692 return 0; 6692 return 0;
6693 return vmcs12->exception_bitmap & 6693 return vmcs12->exception_bitmap &
6694 (1u << (intr_info & INTR_INFO_VECTOR_MASK)); 6694 (1u << (intr_info & INTR_INFO_VECTOR_MASK));
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 39c28f09dfd5..2b8578432d5b 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -6186,7 +6186,7 @@ static int complete_emulated_mmio(struct kvm_vcpu *vcpu)
6186 frag->len -= len; 6186 frag->len -= len;
6187 } 6187 }
6188 6188
6189 if (vcpu->mmio_cur_fragment == vcpu->mmio_nr_fragments) { 6189 if (vcpu->mmio_cur_fragment >= vcpu->mmio_nr_fragments) {
6190 vcpu->mmio_needed = 0; 6190 vcpu->mmio_needed = 0;
6191 6191
6192 /* FIXME: return into emulator if single-stepping. */ 6192 /* FIXME: return into emulator if single-stepping. */
diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c
index 6dea040cc3a1..a10c8c792161 100644
--- a/arch/x86/mm/fault.c
+++ b/arch/x86/mm/fault.c
@@ -1020,13 +1020,17 @@ static inline bool smap_violation(int error_code, struct pt_regs *regs)
1020 * This routine handles page faults. It determines the address, 1020 * This routine handles page faults. It determines the address,
1021 * and the problem, and then passes it off to one of the appropriate 1021 * and the problem, and then passes it off to one of the appropriate
1022 * routines. 1022 * routines.
1023 *
1024 * This function must have noinline because both callers
1025 * {,trace_}do_page_fault() have notrace on. Having this an actual function
1026 * guarantees there's a function trace entry.
1023 */ 1027 */
1024static void __kprobes 1028static void __kprobes noinline
1025__do_page_fault(struct pt_regs *regs, unsigned long error_code) 1029__do_page_fault(struct pt_regs *regs, unsigned long error_code,
1030 unsigned long address)
1026{ 1031{
1027 struct vm_area_struct *vma; 1032 struct vm_area_struct *vma;
1028 struct task_struct *tsk; 1033 struct task_struct *tsk;
1029 unsigned long address;
1030 struct mm_struct *mm; 1034 struct mm_struct *mm;
1031 int fault; 1035 int fault;
1032 unsigned int flags = FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE; 1036 unsigned int flags = FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE;
@@ -1034,9 +1038,6 @@ __do_page_fault(struct pt_regs *regs, unsigned long error_code)
1034 tsk = current; 1038 tsk = current;
1035 mm = tsk->mm; 1039 mm = tsk->mm;
1036 1040
1037 /* Get the faulting address: */
1038 address = read_cr2();
1039
1040 /* 1041 /*
1041 * Detect and handle instructions that would cause a page fault for 1042 * Detect and handle instructions that would cause a page fault for
1042 * both a tracked kernel page and a userspace page. 1043 * both a tracked kernel page and a userspace page.
@@ -1248,32 +1249,50 @@ good_area:
1248 up_read(&mm->mmap_sem); 1249 up_read(&mm->mmap_sem);
1249} 1250}
1250 1251
1251dotraplinkage void __kprobes 1252dotraplinkage void __kprobes notrace
1252do_page_fault(struct pt_regs *regs, unsigned long error_code) 1253do_page_fault(struct pt_regs *regs, unsigned long error_code)
1253{ 1254{
1255 unsigned long address = read_cr2(); /* Get the faulting address */
1254 enum ctx_state prev_state; 1256 enum ctx_state prev_state;
1255 1257
1258 /*
1259 * We must have this function tagged with __kprobes, notrace and call
1260 * read_cr2() before calling anything else. To avoid calling any kind
1261 * of tracing machinery before we've observed the CR2 value.
1262 *
1263 * exception_{enter,exit}() contain all sorts of tracepoints.
1264 */
1265
1256 prev_state = exception_enter(); 1266 prev_state = exception_enter();
1257 __do_page_fault(regs, error_code); 1267 __do_page_fault(regs, error_code, address);
1258 exception_exit(prev_state); 1268 exception_exit(prev_state);
1259} 1269}
1260 1270
1261static void trace_page_fault_entries(struct pt_regs *regs, 1271#ifdef CONFIG_TRACING
1272static void trace_page_fault_entries(unsigned long address, struct pt_regs *regs,
1262 unsigned long error_code) 1273 unsigned long error_code)
1263{ 1274{
1264 if (user_mode(regs)) 1275 if (user_mode(regs))
1265 trace_page_fault_user(read_cr2(), regs, error_code); 1276 trace_page_fault_user(address, regs, error_code);
1266 else 1277 else
1267 trace_page_fault_kernel(read_cr2(), regs, error_code); 1278 trace_page_fault_kernel(address, regs, error_code);
1268} 1279}
1269 1280
1270dotraplinkage void __kprobes 1281dotraplinkage void __kprobes notrace
1271trace_do_page_fault(struct pt_regs *regs, unsigned long error_code) 1282trace_do_page_fault(struct pt_regs *regs, unsigned long error_code)
1272{ 1283{
1284 /*
1285 * The exception_enter and tracepoint processing could
1286 * trigger another page faults (user space callchain
1287 * reading) and destroy the original cr2 value, so read
1288 * the faulting address now.
1289 */
1290 unsigned long address = read_cr2();
1273 enum ctx_state prev_state; 1291 enum ctx_state prev_state;
1274 1292
1275 prev_state = exception_enter(); 1293 prev_state = exception_enter();
1276 trace_page_fault_entries(regs, error_code); 1294 trace_page_fault_entries(address, regs, error_code);
1277 __do_page_fault(regs, error_code); 1295 __do_page_fault(regs, error_code, address);
1278 exception_exit(prev_state); 1296 exception_exit(prev_state);
1279} 1297}
1298#endif /* CONFIG_TRACING */
diff --git a/arch/x86/platform/efi/efi.c b/arch/x86/platform/efi/efi.c
index 1a201ac7cef8..b97acecf3fd9 100644
--- a/arch/x86/platform/efi/efi.c
+++ b/arch/x86/platform/efi/efi.c
@@ -52,6 +52,7 @@
52#include <asm/tlbflush.h> 52#include <asm/tlbflush.h>
53#include <asm/x86_init.h> 53#include <asm/x86_init.h>
54#include <asm/rtc.h> 54#include <asm/rtc.h>
55#include <asm/uv/uv.h>
55 56
56#define EFI_DEBUG 57#define EFI_DEBUG
57 58
@@ -1210,3 +1211,22 @@ static int __init parse_efi_cmdline(char *str)
1210 return 0; 1211 return 0;
1211} 1212}
1212early_param("efi", parse_efi_cmdline); 1213early_param("efi", parse_efi_cmdline);
1214
1215void __init efi_apply_memmap_quirks(void)
1216{
1217 /*
1218 * Once setup is done earlier, unmap the EFI memory map on mismatched
1219 * firmware/kernel architectures since there is no support for runtime
1220 * services.
1221 */
1222 if (!efi_is_native()) {
1223 pr_info("efi: Setup done, disabling due to 32/64-bit mismatch\n");
1224 efi_unmap_memmap();
1225 }
1226
1227 /*
1228 * UV doesn't support the new EFI pagetable mapping yet.
1229 */
1230 if (is_uv_system())
1231 set_bit(EFI_OLD_MEMMAP, &x86_efi_facility);
1232}