2 * Kernel-based Virtual Machine driver for Linux
4 * This header defines architecture specific interfaces, x86 version
6 * This work is licensed under the terms of the GNU GPL, version 2. See
7 * the COPYING file in the top-level directory.
11 #ifndef _ASM_X86_KVM_HOST_H
12 #define _ASM_X86_KVM_HOST_H
14 #include <linux/types.h>
16 #include <linux/mmu_notifier.h>
17 #include <linux/tracepoint.h>
18 #include <linux/cpumask.h>
19 #include <linux/irq_work.h>
21 #include <linux/kvm.h>
22 #include <linux/kvm_para.h>
23 #include <linux/kvm_types.h>
24 #include <linux/perf_event.h>
25 #include <linux/pvclock_gtod.h>
26 #include <linux/clocksource.h>
27 #include <linux/irqbypass.h>
28 #include <linux/hyperv.h>
31 #include <asm/pvclock-abi.h>
34 #include <asm/msr-index.h>
36 #include <asm/kvm_page_track.h>
38 #define KVM_MAX_VCPUS 288
39 #define KVM_SOFT_MAX_VCPUS 240
40 #define KVM_MAX_VCPU_ID 1023
41 #define KVM_USER_MEM_SLOTS 509
42 /* memory slots that are not exposed to userspace */
43 #define KVM_PRIVATE_MEM_SLOTS 3
44 #define KVM_MEM_SLOTS_NUM (KVM_USER_MEM_SLOTS + KVM_PRIVATE_MEM_SLOTS)
46 #define KVM_HALT_POLL_NS_DEFAULT 200000
48 #define KVM_IRQCHIP_NUM_PINS KVM_IOAPIC_NUM_PINS
50 /* x86-specific vcpu->requests bit members */
51 #define KVM_REQ_MIGRATE_TIMER KVM_ARCH_REQ(0)
52 #define KVM_REQ_REPORT_TPR_ACCESS KVM_ARCH_REQ(1)
53 #define KVM_REQ_TRIPLE_FAULT KVM_ARCH_REQ(2)
54 #define KVM_REQ_MMU_SYNC KVM_ARCH_REQ(3)
55 #define KVM_REQ_CLOCK_UPDATE KVM_ARCH_REQ(4)
56 #define KVM_REQ_EVENT KVM_ARCH_REQ(6)
57 #define KVM_REQ_APF_HALT KVM_ARCH_REQ(7)
58 #define KVM_REQ_STEAL_UPDATE KVM_ARCH_REQ(8)
59 #define KVM_REQ_NMI KVM_ARCH_REQ(9)
60 #define KVM_REQ_PMU KVM_ARCH_REQ(10)
61 #define KVM_REQ_PMI KVM_ARCH_REQ(11)
62 #define KVM_REQ_SMI KVM_ARCH_REQ(12)
63 #define KVM_REQ_MASTERCLOCK_UPDATE KVM_ARCH_REQ(13)
64 #define KVM_REQ_MCLOCK_INPROGRESS \
65 KVM_ARCH_REQ_FLAGS(14, KVM_REQUEST_WAIT | KVM_REQUEST_NO_WAKEUP)
66 #define KVM_REQ_SCAN_IOAPIC \
67 KVM_ARCH_REQ_FLAGS(15, KVM_REQUEST_WAIT | KVM_REQUEST_NO_WAKEUP)
68 #define KVM_REQ_GLOBAL_CLOCK_UPDATE KVM_ARCH_REQ(16)
69 #define KVM_REQ_APIC_PAGE_RELOAD \
70 KVM_ARCH_REQ_FLAGS(17, KVM_REQUEST_WAIT | KVM_REQUEST_NO_WAKEUP)
71 #define KVM_REQ_HV_CRASH KVM_ARCH_REQ(18)
72 #define KVM_REQ_IOAPIC_EOI_EXIT KVM_ARCH_REQ(19)
73 #define KVM_REQ_HV_RESET KVM_ARCH_REQ(20)
74 #define KVM_REQ_HV_EXIT KVM_ARCH_REQ(21)
75 #define KVM_REQ_HV_STIMER KVM_ARCH_REQ(22)
77 #define CR0_RESERVED_BITS \
78 (~(unsigned long)(X86_CR0_PE | X86_CR0_MP | X86_CR0_EM | X86_CR0_TS \
79 | X86_CR0_ET | X86_CR0_NE | X86_CR0_WP | X86_CR0_AM \
80 | X86_CR0_NW | X86_CR0_CD | X86_CR0_PG))
82 #define CR3_PCID_INVD BIT_64(63)
83 #define CR4_RESERVED_BITS \
84 (~(unsigned long)(X86_CR4_VME | X86_CR4_PVI | X86_CR4_TSD | X86_CR4_DE\
85 | X86_CR4_PSE | X86_CR4_PAE | X86_CR4_MCE \
86 | X86_CR4_PGE | X86_CR4_PCE | X86_CR4_OSFXSR | X86_CR4_PCIDE \
87 | X86_CR4_OSXSAVE | X86_CR4_SMEP | X86_CR4_FSGSBASE \
88 | X86_CR4_OSXMMEXCPT | X86_CR4_LA57 | X86_CR4_VMXE \
89 | X86_CR4_SMAP | X86_CR4_PKE | X86_CR4_UMIP))
91 #define CR8_RESERVED_BITS (~(unsigned long)X86_CR8_TPR)
95 #define INVALID_PAGE (~(hpa_t)0)
96 #define VALID_PAGE(x) ((x) != INVALID_PAGE)
98 #define UNMAPPED_GVA (~(gpa_t)0)
100 /* KVM Hugepage definitions for x86 */
101 #define KVM_NR_PAGE_SIZES 3
102 #define KVM_HPAGE_GFN_SHIFT(x) (((x) - 1) * 9)
103 #define KVM_HPAGE_SHIFT(x) (PAGE_SHIFT + KVM_HPAGE_GFN_SHIFT(x))
104 #define KVM_HPAGE_SIZE(x) (1UL << KVM_HPAGE_SHIFT(x))
105 #define KVM_HPAGE_MASK(x) (~(KVM_HPAGE_SIZE(x) - 1))
106 #define KVM_PAGES_PER_HPAGE(x) (KVM_HPAGE_SIZE(x) / PAGE_SIZE)
108 static inline gfn_t gfn_to_index(gfn_t gfn, gfn_t base_gfn, int level)
110 /* KVM_HPAGE_GFN_SHIFT(PT_PAGE_TABLE_LEVEL) must be 0. */
111 return (gfn >> KVM_HPAGE_GFN_SHIFT(level)) -
112 (base_gfn >> KVM_HPAGE_GFN_SHIFT(level));
115 #define KVM_PERMILLE_MMU_PAGES 20
116 #define KVM_MIN_ALLOC_MMU_PAGES 64
117 #define KVM_MMU_HASH_SHIFT 12
118 #define KVM_NUM_MMU_PAGES (1 << KVM_MMU_HASH_SHIFT)
119 #define KVM_MIN_FREE_MMU_PAGES 5
120 #define KVM_REFILL_PAGES 25
121 #define KVM_MAX_CPUID_ENTRIES 80
122 #define KVM_NR_FIXED_MTRR_REGION 88
123 #define KVM_NR_VAR_MTRR 8
125 #define ASYNC_PF_PER_VCPU 64
151 VCPU_EXREG_PDPTR = NR_VCPU_REGS,
168 #include <asm/kvm_emulate.h>
170 #define KVM_NR_MEM_OBJS 40
172 #define KVM_NR_DB_REGS 4
174 #define DR6_BD (1 << 13)
175 #define DR6_BS (1 << 14)
176 #define DR6_RTM (1 << 16)
177 #define DR6_FIXED_1 0xfffe0ff0
178 #define DR6_INIT 0xffff0ff0
179 #define DR6_VOLATILE 0x0001e00f
181 #define DR7_BP_EN_MASK 0x000000ff
182 #define DR7_GE (1 << 9)
183 #define DR7_GD (1 << 13)
184 #define DR7_FIXED_1 0x00000400
185 #define DR7_VOLATILE 0xffff2bff
187 #define PFERR_PRESENT_BIT 0
188 #define PFERR_WRITE_BIT 1
189 #define PFERR_USER_BIT 2
190 #define PFERR_RSVD_BIT 3
191 #define PFERR_FETCH_BIT 4
192 #define PFERR_PK_BIT 5
193 #define PFERR_GUEST_FINAL_BIT 32
194 #define PFERR_GUEST_PAGE_BIT 33
196 #define PFERR_PRESENT_MASK (1U << PFERR_PRESENT_BIT)
197 #define PFERR_WRITE_MASK (1U << PFERR_WRITE_BIT)
198 #define PFERR_USER_MASK (1U << PFERR_USER_BIT)
199 #define PFERR_RSVD_MASK (1U << PFERR_RSVD_BIT)
200 #define PFERR_FETCH_MASK (1U << PFERR_FETCH_BIT)
201 #define PFERR_PK_MASK (1U << PFERR_PK_BIT)
202 #define PFERR_GUEST_FINAL_MASK (1ULL << PFERR_GUEST_FINAL_BIT)
203 #define PFERR_GUEST_PAGE_MASK (1ULL << PFERR_GUEST_PAGE_BIT)
205 #define PFERR_NESTED_GUEST_PAGE (PFERR_GUEST_PAGE_MASK | \
210 * The mask used to denote special SPTEs, which can be either MMIO SPTEs or
211 * Access Tracking SPTEs. We use bit 62 instead of bit 63 to avoid conflicting
212 * with the SVE bit in EPT PTEs.
214 #define SPTE_SPECIAL_MASK (1ULL << 62)
216 /* apic attention bits */
217 #define KVM_APIC_CHECK_VAPIC 0
219 * The following bit is set with PV-EOI, unset on EOI.
220 * We detect PV-EOI changes by guest by comparing
221 * this bit with PV-EOI in guest memory.
222 * See the implementation in apic_update_pv_eoi.
224 #define KVM_APIC_PV_EOI_PENDING 1
226 struct kvm_kernel_irq_routing_entry;
229 * We don't want allocation failures within the mmu code, so we preallocate
230 * enough memory for a single page fault in a cache.
232 struct kvm_mmu_memory_cache {
234 void *objects[KVM_NR_MEM_OBJS];
238 * the pages used as guest page table on soft mmu are tracked by
239 * kvm_memory_slot.arch.gfn_track which is 16 bits, so the role bits used
240 * by indirect shadow page can not be more than 15 bits.
242 * Currently, we used 14 bits that are @level, @cr4_pae, @quadrant, @access,
243 * @nxe, @cr0_wp, @smep_andnot_wp and @smap_andnot_wp.
245 union kvm_mmu_page_role {
256 unsigned smep_andnot_wp:1;
257 unsigned smap_andnot_wp:1;
258 unsigned ad_disabled:1;
262 * This is left at the top of the word so that
263 * kvm_memslots_for_spte_role can extract it with a
264 * simple shift. While there is room, give it a whole
265 * byte so it is also faster to load it from memory.
271 struct kvm_rmap_head {
275 struct kvm_mmu_page {
276 struct list_head link;
277 struct hlist_node hash_link;
280 * The following two entries are used to key the shadow page in the
284 union kvm_mmu_page_role role;
287 /* hold the gfn of each spte inside spt */
290 int root_count; /* Currently serving as active root */
291 unsigned int unsync_children;
292 struct kvm_rmap_head parent_ptes; /* rmap pointers to parent sptes */
294 /* The page is obsolete if mmu_valid_gen != kvm->arch.mmu_valid_gen. */
295 unsigned long mmu_valid_gen;
297 DECLARE_BITMAP(unsync_child_bitmap, 512);
301 * Used out of the mmu-lock to avoid reading spte values while an
302 * update is in progress; see the comments in __get_spte_lockless().
304 int clear_spte_count;
307 /* Number of writes since the last time traversal visited this page. */
308 atomic_t write_flooding_count;
311 struct kvm_pio_request {
318 #define PT64_ROOT_MAX_LEVEL 5
320 struct rsvd_bits_validate {
321 u64 rsvd_bits_mask[2][PT64_ROOT_MAX_LEVEL];
326 * x86 supports 4 paging modes (5-level 64-bit, 4-level 64-bit, 3-level 32-bit,
327 * and 2-level 32-bit). The kvm_mmu structure abstracts the details of the
331 void (*set_cr3)(struct kvm_vcpu *vcpu, unsigned long root);
332 unsigned long (*get_cr3)(struct kvm_vcpu *vcpu);
333 u64 (*get_pdptr)(struct kvm_vcpu *vcpu, int index);
334 int (*page_fault)(struct kvm_vcpu *vcpu, gva_t gva, u32 err,
336 void (*inject_page_fault)(struct kvm_vcpu *vcpu,
337 struct x86_exception *fault);
338 gpa_t (*gva_to_gpa)(struct kvm_vcpu *vcpu, gva_t gva, u32 access,
339 struct x86_exception *exception);
340 gpa_t (*translate_gpa)(struct kvm_vcpu *vcpu, gpa_t gpa, u32 access,
341 struct x86_exception *exception);
342 int (*sync_page)(struct kvm_vcpu *vcpu,
343 struct kvm_mmu_page *sp);
344 void (*invlpg)(struct kvm_vcpu *vcpu, gva_t gva);
345 void (*update_pte)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp,
346 u64 *spte, const void *pte);
348 union kvm_mmu_page_role base_role;
350 u8 shadow_root_level;
355 * Bitmap; bit set = permission fault
356 * Byte index: page fault error code [4:1]
357 * Bit index: pte permissions in ACC_* format
362 * The pkru_mask indicates if protection key checks are needed. It
363 * consists of 16 domains indexed by page fault error code bits [4:1],
364 * with PFEC.RSVD replaced by ACC_USER_MASK from the page tables.
365 * Each domain has 2 bits which are ANDed with AD and WD from PKRU.
373 * check zero bits on shadow page table entries, these
374 * bits include not only hardware reserved bits but also
375 * the bits spte never used.
377 struct rsvd_bits_validate shadow_zero_check;
379 struct rsvd_bits_validate guest_rsvd_check;
381 /* Can have large pages at levels 2..last_nonleaf_level-1. */
382 u8 last_nonleaf_level;
386 u64 pdptrs[4]; /* pae */
399 struct perf_event *perf_event;
400 struct kvm_vcpu *vcpu;
404 unsigned nr_arch_gp_counters;
405 unsigned nr_arch_fixed_counters;
406 unsigned available_event_types;
411 u64 counter_bitmask[2];
412 u64 global_ctrl_mask;
415 struct kvm_pmc gp_counters[INTEL_PMC_MAX_GENERIC];
416 struct kvm_pmc fixed_counters[INTEL_PMC_MAX_FIXED];
417 struct irq_work irq_work;
424 KVM_DEBUGREG_BP_ENABLED = 1,
425 KVM_DEBUGREG_WONT_EXIT = 2,
426 KVM_DEBUGREG_RELOAD = 4,
429 struct kvm_mtrr_range {
432 struct list_head node;
436 struct kvm_mtrr_range var_ranges[KVM_NR_VAR_MTRR];
437 mtrr_type fixed_ranges[KVM_NR_FIXED_MTRR_REGION];
440 struct list_head head;
443 /* Hyper-V SynIC timer */
444 struct kvm_vcpu_hv_stimer {
445 struct hrtimer timer;
450 struct hv_message msg;
454 /* Hyper-V synthetic interrupt controller (SynIC)*/
455 struct kvm_vcpu_hv_synic {
460 atomic64_t sint[HV_SYNIC_SINT_COUNT];
461 atomic_t sint_to_gsi[HV_SYNIC_SINT_COUNT];
462 DECLARE_BITMAP(auto_eoi_bitmap, 256);
463 DECLARE_BITMAP(vec_bitmap, 256);
465 bool dont_zero_synic_pages;
468 /* Hyper-V per vcpu emulation context */
473 struct kvm_vcpu_hv_synic synic;
474 struct kvm_hyperv_exit exit;
475 struct kvm_vcpu_hv_stimer stimer[HV_SYNIC_STIMER_COUNT];
476 DECLARE_BITMAP(stimer_pending_bitmap, HV_SYNIC_STIMER_COUNT);
479 struct kvm_vcpu_arch {
481 * rip and regs accesses must go through
482 * kvm_{register,rip}_{read,write} functions.
484 unsigned long regs[NR_VCPU_REGS];
489 unsigned long cr0_guest_owned_bits;
493 unsigned long cr4_guest_owned_bits;
499 struct kvm_lapic *apic; /* kernel irqchip context */
501 DECLARE_BITMAP(ioapic_handled_vectors, 256);
502 unsigned long apic_attention;
503 int32_t apic_arb_prio;
505 u64 ia32_misc_enable_msr;
508 bool tpr_access_reporting;
512 * Paging state of the vcpu
514 * If the vcpu runs in guest mode with two level paging this still saves
515 * the paging mode of the l1 guest. This context is always used to
521 * Paging state of an L2 guest (used for nested npt)
523 * This context will save all necessary information to walk page tables
524 * of the an L2 guest. This context is only initialized for page table
525 * walking and not for faulting since we never handle l2 page faults on
528 struct kvm_mmu nested_mmu;
531 * Pointer to the mmu context currently used for
532 * gva_to_gpa translations.
534 struct kvm_mmu *walk_mmu;
536 struct kvm_mmu_memory_cache mmu_pte_list_desc_cache;
537 struct kvm_mmu_memory_cache mmu_page_cache;
538 struct kvm_mmu_memory_cache mmu_page_header_cache;
541 * QEMU userspace and the guest each have their own FPU state.
542 * In vcpu_run, we switch between the user and guest FPU contexts.
543 * While running a VCPU, the VCPU thread will have the guest FPU
546 * Note that while the PKRU state lives inside the fpu registers,
547 * it is switched out separately at VMENTER and VMEXIT time. The
548 * "guest_fpu" state here contains the guest FPU context, with the
552 struct fpu guest_fpu;
555 u64 guest_supported_xcr0;
556 u32 guest_xstate_size;
558 struct kvm_pio_request pio;
561 u8 event_exit_inst_len;
563 struct kvm_queued_exception {
572 struct kvm_queued_interrupt {
578 int halt_request; /* real mode on Intel only */
581 struct kvm_cpuid_entry2 cpuid_entries[KVM_MAX_CPUID_ENTRIES];
585 /* emulate context */
587 struct x86_emulate_ctxt emulate_ctxt;
588 bool emulate_regs_need_sync_to_vcpu;
589 bool emulate_regs_need_sync_from_vcpu;
590 int (*complete_userspace_io)(struct kvm_vcpu *vcpu);
593 struct pvclock_vcpu_time_info hv_clock;
594 unsigned int hw_tsc_khz;
595 struct gfn_to_hva_cache pv_time;
596 bool pv_time_enabled;
597 /* set guest stopped flag in pvclock flags field */
598 bool pvclock_set_guest_stopped_request;
603 struct gfn_to_hva_cache stime;
604 struct kvm_steal_time steal;
610 u64 tsc_offset_adjustment;
613 u64 this_tsc_generation;
615 bool tsc_always_catchup;
616 s8 virtual_tsc_shift;
617 u32 virtual_tsc_mult;
619 s64 ia32_tsc_adjust_msr;
620 u64 tsc_scaling_ratio;
622 atomic_t nmi_queued; /* unprocessed asynchronous NMIs */
623 unsigned nmi_pending; /* NMI queued after currently running handler */
624 bool nmi_injected; /* Trying to inject an NMI this entry */
625 bool smi_pending; /* SMI queued after currently running handler */
627 struct kvm_mtrr mtrr_state;
630 unsigned switch_db_regs;
631 unsigned long db[KVM_NR_DB_REGS];
634 unsigned long eff_db[KVM_NR_DB_REGS];
635 unsigned long guest_debug_dr7;
636 u64 msr_platform_info;
637 u64 msr_misc_features_enables;
645 /* Cache MMIO info */
653 /* used for guest single stepping over the given code position */
654 unsigned long singlestep_rip;
656 struct kvm_vcpu_hv hyperv;
658 cpumask_var_t wbinvd_dirty_mask;
660 unsigned long last_retry_eip;
661 unsigned long last_retry_addr;
665 gfn_t gfns[roundup_pow_of_two(ASYNC_PF_PER_VCPU)];
666 struct gfn_to_hva_cache data;
671 unsigned long nested_apf_token;
672 bool delivery_as_pf_vmexit;
675 /* OSVW MSRs (AMD only) */
683 struct gfn_to_hva_cache data;
687 * Indicate whether the access faults on its page table in guest
688 * which is set when fix page fault and used to detect unhandeable
691 bool write_fault_to_shadow_pgtable;
693 /* set at EPT violation at this point */
694 unsigned long exit_qualification;
696 /* pv related host specific info */
701 int pending_ioapic_eoi;
702 int pending_external_vector;
708 /* be preempted when it's in kernel-mode(cpl=0) */
709 bool preempted_in_kernel;
712 struct kvm_lpage_info {
716 struct kvm_arch_memory_slot {
717 struct kvm_rmap_head *rmap[KVM_NR_PAGE_SIZES];
718 struct kvm_lpage_info *lpage_info[KVM_NR_PAGE_SIZES - 1];
719 unsigned short *gfn_track[KVM_PAGE_TRACK_MAX];
723 * We use as the mode the number of bits allocated in the LDR for the
724 * logical processor ID. It happens that these are all powers of two.
725 * This makes it is very easy to detect cases where the APICs are
726 * configured for multiple modes; in that case, we cannot use the map and
727 * hence cannot use kvm_irq_delivery_to_apic_fast either.
729 #define KVM_APIC_MODE_XAPIC_CLUSTER 4
730 #define KVM_APIC_MODE_XAPIC_FLAT 8
731 #define KVM_APIC_MODE_X2APIC 16
733 struct kvm_apic_map {
738 struct kvm_lapic *xapic_flat_map[8];
739 struct kvm_lapic *xapic_cluster_map[16][4];
741 struct kvm_lapic *phys_map[];
744 /* Hyper-V emulation context */
746 struct mutex hv_lock;
751 /* Hyper-v based guest crash (NT kernel bugcheck) parameters */
752 u64 hv_crash_param[HV_X64_MSR_CRASH_PARAMS];
755 HV_REFERENCE_TSC_PAGE tsc_ref;
758 enum kvm_irqchip_mode {
760 KVM_IRQCHIP_KERNEL, /* created with KVM_CREATE_IRQCHIP */
761 KVM_IRQCHIP_SPLIT, /* created with KVM_CAP_SPLIT_IRQCHIP */
764 struct kvm_sev_info {
765 bool active; /* SEV enabled guest */
766 unsigned int asid; /* ASID used for this guest */
767 unsigned int handle; /* SEV firmware handle */
768 int fd; /* SEV device fd */
769 unsigned long pages_locked; /* Number of pages locked */
770 struct list_head regions_list; /* List of registered regions */
774 unsigned int n_used_mmu_pages;
775 unsigned int n_requested_mmu_pages;
776 unsigned int n_max_mmu_pages;
777 unsigned int indirect_shadow_pages;
778 unsigned long mmu_valid_gen;
779 struct hlist_head mmu_page_hash[KVM_NUM_MMU_PAGES];
781 * Hash table of struct kvm_mmu_page.
783 struct list_head active_mmu_pages;
784 struct list_head zapped_obsolete_pages;
785 struct kvm_page_track_notifier_node mmu_sp_tracker;
786 struct kvm_page_track_notifier_head track_notifier_head;
788 struct list_head assigned_dev_head;
789 struct iommu_domain *iommu_domain;
790 bool iommu_noncoherent;
791 #define __KVM_HAVE_ARCH_NONCOHERENT_DMA
792 atomic_t noncoherent_dma_count;
793 #define __KVM_HAVE_ARCH_ASSIGNED_DEVICE
794 atomic_t assigned_device_count;
795 struct kvm_pic *vpic;
796 struct kvm_ioapic *vioapic;
797 struct kvm_pit *vpit;
798 atomic_t vapics_in_nmi_mode;
799 struct mutex apic_map_lock;
800 struct kvm_apic_map *apic_map;
802 unsigned int tss_addr;
803 bool apic_access_page_done;
807 bool ept_identity_pagetable_done;
808 gpa_t ept_identity_map_addr;
810 unsigned long irq_sources_bitmap;
812 raw_spinlock_t tsc_write_lock;
819 u64 cur_tsc_generation;
820 int nr_vcpus_matched_tsc;
822 spinlock_t pvclock_gtod_sync_lock;
823 bool use_master_clock;
824 u64 master_kernel_ns;
825 u64 master_cycle_now;
826 struct delayed_work kvmclock_update_work;
827 struct delayed_work kvmclock_sync_work;
829 struct kvm_xen_hvm_config xen_hvm_config;
831 /* reads protected by irq_srcu, writes by irq_lock */
832 struct hlist_head mask_notifier_list;
834 struct kvm_hv hyperv;
836 #ifdef CONFIG_KVM_MMU_AUDIT
840 bool backwards_tsc_observed;
841 bool boot_vcpu_runs_old_kvmclock;
846 enum kvm_irqchip_mode irqchip_mode;
847 u8 nr_reserved_ioapic_pins;
849 bool disabled_lapic_found;
851 /* Struct members for AVIC */
854 struct page *avic_logical_id_table_page;
855 struct page *avic_physical_id_table_page;
856 struct hlist_node hnode;
859 bool x2apic_broadcast_quirk_disabled;
861 struct kvm_sev_info sev_info;
865 ulong mmu_shadow_zapped;
867 ulong mmu_pte_updated;
868 ulong mmu_pde_zapped;
871 ulong mmu_cache_miss;
873 ulong remote_tlb_flush;
875 ulong max_mmu_page_hash_collisions;
878 struct kvm_vcpu_stat {
888 u64 irq_window_exits;
889 u64 nmi_window_exits;
891 u64 halt_successful_poll;
892 u64 halt_attempted_poll;
893 u64 halt_poll_invalid;
895 u64 request_irq_exits;
897 u64 host_state_reload;
900 u64 insn_emulation_fail;
907 struct x86_instruction_info;
915 struct kvm_lapic_irq {
927 int (*cpu_has_kvm_support)(void); /* __init */
928 int (*disabled_by_bios)(void); /* __init */
929 int (*hardware_enable)(void);
930 void (*hardware_disable)(void);
931 void (*check_processor_compatibility)(void *rtn);
932 int (*hardware_setup)(void); /* __init */
933 void (*hardware_unsetup)(void); /* __exit */
934 bool (*cpu_has_accelerated_tpr)(void);
935 bool (*cpu_has_high_real_mode_segbase)(void);
936 void (*cpuid_update)(struct kvm_vcpu *vcpu);
938 int (*vm_init)(struct kvm *kvm);
939 void (*vm_destroy)(struct kvm *kvm);
941 /* Create, but do not attach this VCPU */
942 struct kvm_vcpu *(*vcpu_create)(struct kvm *kvm, unsigned id);
943 void (*vcpu_free)(struct kvm_vcpu *vcpu);
944 void (*vcpu_reset)(struct kvm_vcpu *vcpu, bool init_event);
946 void (*prepare_guest_switch)(struct kvm_vcpu *vcpu);
947 void (*vcpu_load)(struct kvm_vcpu *vcpu, int cpu);
948 void (*vcpu_put)(struct kvm_vcpu *vcpu);
950 void (*update_bp_intercept)(struct kvm_vcpu *vcpu);
951 int (*get_msr)(struct kvm_vcpu *vcpu, struct msr_data *msr);
952 int (*set_msr)(struct kvm_vcpu *vcpu, struct msr_data *msr);
953 u64 (*get_segment_base)(struct kvm_vcpu *vcpu, int seg);
954 void (*get_segment)(struct kvm_vcpu *vcpu,
955 struct kvm_segment *var, int seg);
956 int (*get_cpl)(struct kvm_vcpu *vcpu);
957 void (*set_segment)(struct kvm_vcpu *vcpu,
958 struct kvm_segment *var, int seg);
959 void (*get_cs_db_l_bits)(struct kvm_vcpu *vcpu, int *db, int *l);
960 void (*decache_cr0_guest_bits)(struct kvm_vcpu *vcpu);
961 void (*decache_cr3)(struct kvm_vcpu *vcpu);
962 void (*decache_cr4_guest_bits)(struct kvm_vcpu *vcpu);
963 void (*set_cr0)(struct kvm_vcpu *vcpu, unsigned long cr0);
964 void (*set_cr3)(struct kvm_vcpu *vcpu, unsigned long cr3);
965 int (*set_cr4)(struct kvm_vcpu *vcpu, unsigned long cr4);
966 void (*set_efer)(struct kvm_vcpu *vcpu, u64 efer);
967 void (*get_idt)(struct kvm_vcpu *vcpu, struct desc_ptr *dt);
968 void (*set_idt)(struct kvm_vcpu *vcpu, struct desc_ptr *dt);
969 void (*get_gdt)(struct kvm_vcpu *vcpu, struct desc_ptr *dt);
970 void (*set_gdt)(struct kvm_vcpu *vcpu, struct desc_ptr *dt);
971 u64 (*get_dr6)(struct kvm_vcpu *vcpu);
972 void (*set_dr6)(struct kvm_vcpu *vcpu, unsigned long value);
973 void (*sync_dirty_debug_regs)(struct kvm_vcpu *vcpu);
974 void (*set_dr7)(struct kvm_vcpu *vcpu, unsigned long value);
975 void (*cache_reg)(struct kvm_vcpu *vcpu, enum kvm_reg reg);
976 unsigned long (*get_rflags)(struct kvm_vcpu *vcpu);
977 void (*set_rflags)(struct kvm_vcpu *vcpu, unsigned long rflags);
979 void (*tlb_flush)(struct kvm_vcpu *vcpu, bool invalidate_gpa);
981 void (*run)(struct kvm_vcpu *vcpu);
982 int (*handle_exit)(struct kvm_vcpu *vcpu);
983 void (*skip_emulated_instruction)(struct kvm_vcpu *vcpu);
984 void (*set_interrupt_shadow)(struct kvm_vcpu *vcpu, int mask);
985 u32 (*get_interrupt_shadow)(struct kvm_vcpu *vcpu);
986 void (*patch_hypercall)(struct kvm_vcpu *vcpu,
987 unsigned char *hypercall_addr);
988 void (*set_irq)(struct kvm_vcpu *vcpu);
989 void (*set_nmi)(struct kvm_vcpu *vcpu);
990 void (*queue_exception)(struct kvm_vcpu *vcpu);
991 void (*cancel_injection)(struct kvm_vcpu *vcpu);
992 int (*interrupt_allowed)(struct kvm_vcpu *vcpu);
993 int (*nmi_allowed)(struct kvm_vcpu *vcpu);
994 bool (*get_nmi_mask)(struct kvm_vcpu *vcpu);
995 void (*set_nmi_mask)(struct kvm_vcpu *vcpu, bool masked);
996 void (*enable_nmi_window)(struct kvm_vcpu *vcpu);
997 void (*enable_irq_window)(struct kvm_vcpu *vcpu);
998 void (*update_cr8_intercept)(struct kvm_vcpu *vcpu, int tpr, int irr);
999 bool (*get_enable_apicv)(struct kvm_vcpu *vcpu);
1000 void (*refresh_apicv_exec_ctrl)(struct kvm_vcpu *vcpu);
1001 void (*hwapic_irr_update)(struct kvm_vcpu *vcpu, int max_irr);
1002 void (*hwapic_isr_update)(struct kvm_vcpu *vcpu, int isr);
1003 void (*load_eoi_exitmap)(struct kvm_vcpu *vcpu, u64 *eoi_exit_bitmap);
1004 void (*set_virtual_x2apic_mode)(struct kvm_vcpu *vcpu, bool set);
1005 void (*set_apic_access_page_addr)(struct kvm_vcpu *vcpu, hpa_t hpa);
1006 void (*deliver_posted_interrupt)(struct kvm_vcpu *vcpu, int vector);
1007 int (*sync_pir_to_irr)(struct kvm_vcpu *vcpu);
1008 int (*set_tss_addr)(struct kvm *kvm, unsigned int addr);
1009 int (*get_tdp_level)(struct kvm_vcpu *vcpu);
1010 u64 (*get_mt_mask)(struct kvm_vcpu *vcpu, gfn_t gfn, bool is_mmio);
1011 int (*get_lpage_level)(void);
1012 bool (*rdtscp_supported)(void);
1013 bool (*invpcid_supported)(void);
1015 void (*set_tdp_cr3)(struct kvm_vcpu *vcpu, unsigned long cr3);
1017 void (*set_supported_cpuid)(u32 func, struct kvm_cpuid_entry2 *entry);
1019 bool (*has_wbinvd_exit)(void);
1021 void (*write_tsc_offset)(struct kvm_vcpu *vcpu, u64 offset);
1023 void (*get_exit_info)(struct kvm_vcpu *vcpu, u64 *info1, u64 *info2);
1025 int (*check_intercept)(struct kvm_vcpu *vcpu,
1026 struct x86_instruction_info *info,
1027 enum x86_intercept_stage stage);
1028 void (*handle_external_intr)(struct kvm_vcpu *vcpu);
1029 bool (*mpx_supported)(void);
1030 bool (*xsaves_supported)(void);
1031 bool (*umip_emulated)(void);
1033 int (*check_nested_events)(struct kvm_vcpu *vcpu, bool external_intr);
1035 void (*sched_in)(struct kvm_vcpu *kvm, int cpu);
1038 * Arch-specific dirty logging hooks. These hooks are only supposed to
1039 * be valid if the specific arch has hardware-accelerated dirty logging
1040 * mechanism. Currently only for PML on VMX.
1042 * - slot_enable_log_dirty:
1043 * called when enabling log dirty mode for the slot.
1044 * - slot_disable_log_dirty:
1045 * called when disabling log dirty mode for the slot.
1046 * also called when slot is created with log dirty disabled.
1047 * - flush_log_dirty:
1048 * called before reporting dirty_bitmap to userspace.
1049 * - enable_log_dirty_pt_masked:
1050 * called when reenabling log dirty for the GFNs in the mask after
1051 * corresponding bits are cleared in slot->dirty_bitmap.
1053 void (*slot_enable_log_dirty)(struct kvm *kvm,
1054 struct kvm_memory_slot *slot);
1055 void (*slot_disable_log_dirty)(struct kvm *kvm,
1056 struct kvm_memory_slot *slot);
1057 void (*flush_log_dirty)(struct kvm *kvm);
1058 void (*enable_log_dirty_pt_masked)(struct kvm *kvm,
1059 struct kvm_memory_slot *slot,
1060 gfn_t offset, unsigned long mask);
1061 int (*write_log_dirty)(struct kvm_vcpu *vcpu);
1063 /* pmu operations of sub-arch */
1064 const struct kvm_pmu_ops *pmu_ops;
1067 * Architecture specific hooks for vCPU blocking due to
1069 * Returns for .pre_block():
1070 * - 0 means continue to block the vCPU.
1071 * - 1 means we cannot block the vCPU since some event
1072 * happens during this period, such as, 'ON' bit in
1073 * posted-interrupts descriptor is set.
1075 int (*pre_block)(struct kvm_vcpu *vcpu);
1076 void (*post_block)(struct kvm_vcpu *vcpu);
1078 void (*vcpu_blocking)(struct kvm_vcpu *vcpu);
1079 void (*vcpu_unblocking)(struct kvm_vcpu *vcpu);
1081 int (*update_pi_irte)(struct kvm *kvm, unsigned int host_irq,
1082 uint32_t guest_irq, bool set);
1083 void (*apicv_post_state_restore)(struct kvm_vcpu *vcpu);
1085 int (*set_hv_timer)(struct kvm_vcpu *vcpu, u64 guest_deadline_tsc);
1086 void (*cancel_hv_timer)(struct kvm_vcpu *vcpu);
1088 void (*setup_mce)(struct kvm_vcpu *vcpu);
1090 int (*smi_allowed)(struct kvm_vcpu *vcpu);
1091 int (*pre_enter_smm)(struct kvm_vcpu *vcpu, char *smstate);
1092 int (*pre_leave_smm)(struct kvm_vcpu *vcpu, u64 smbase);
1093 int (*enable_smi_window)(struct kvm_vcpu *vcpu);
1095 int (*mem_enc_op)(struct kvm *kvm, void __user *argp);
1096 int (*mem_enc_reg_region)(struct kvm *kvm, struct kvm_enc_region *argp);
1097 int (*mem_enc_unreg_region)(struct kvm *kvm, struct kvm_enc_region *argp);
1100 struct kvm_arch_async_pf {
1107 extern struct kvm_x86_ops *kvm_x86_ops;
1109 int kvm_mmu_module_init(void);
1110 void kvm_mmu_module_exit(void);
1112 void kvm_mmu_destroy(struct kvm_vcpu *vcpu);
1113 int kvm_mmu_create(struct kvm_vcpu *vcpu);
1114 void kvm_mmu_setup(struct kvm_vcpu *vcpu);
1115 void kvm_mmu_init_vm(struct kvm *kvm);
1116 void kvm_mmu_uninit_vm(struct kvm *kvm);
1117 void kvm_mmu_set_mask_ptes(u64 user_mask, u64 accessed_mask,
1118 u64 dirty_mask, u64 nx_mask, u64 x_mask, u64 p_mask,
1119 u64 acc_track_mask, u64 me_mask);
1121 void kvm_mmu_reset_context(struct kvm_vcpu *vcpu);
1122 void kvm_mmu_slot_remove_write_access(struct kvm *kvm,
1123 struct kvm_memory_slot *memslot);
1124 void kvm_mmu_zap_collapsible_sptes(struct kvm *kvm,
1125 const struct kvm_memory_slot *memslot);
1126 void kvm_mmu_slot_leaf_clear_dirty(struct kvm *kvm,
1127 struct kvm_memory_slot *memslot);
1128 void kvm_mmu_slot_largepage_remove_write_access(struct kvm *kvm,
1129 struct kvm_memory_slot *memslot);
1130 void kvm_mmu_slot_set_dirty(struct kvm *kvm,
1131 struct kvm_memory_slot *memslot);
1132 void kvm_mmu_clear_dirty_pt_masked(struct kvm *kvm,
1133 struct kvm_memory_slot *slot,
1134 gfn_t gfn_offset, unsigned long mask);
1135 void kvm_mmu_zap_all(struct kvm *kvm);
1136 void kvm_mmu_invalidate_mmio_sptes(struct kvm *kvm, struct kvm_memslots *slots);
1137 unsigned int kvm_mmu_calculate_mmu_pages(struct kvm *kvm);
1138 void kvm_mmu_change_mmu_pages(struct kvm *kvm, unsigned int kvm_nr_mmu_pages);
1140 int load_pdptrs(struct kvm_vcpu *vcpu, struct kvm_mmu *mmu, unsigned long cr3);
1141 bool pdptrs_changed(struct kvm_vcpu *vcpu);
1143 int emulator_write_phys(struct kvm_vcpu *vcpu, gpa_t gpa,
1144 const void *val, int bytes);
1146 struct kvm_irq_mask_notifier {
1147 void (*func)(struct kvm_irq_mask_notifier *kimn, bool masked);
1149 struct hlist_node link;
1152 void kvm_register_irq_mask_notifier(struct kvm *kvm, int irq,
1153 struct kvm_irq_mask_notifier *kimn);
1154 void kvm_unregister_irq_mask_notifier(struct kvm *kvm, int irq,
1155 struct kvm_irq_mask_notifier *kimn);
1156 void kvm_fire_mask_notifiers(struct kvm *kvm, unsigned irqchip, unsigned pin,
1159 extern bool tdp_enabled;
1161 u64 vcpu_tsc_khz(struct kvm_vcpu *vcpu);
1163 /* control of guest tsc rate supported? */
1164 extern bool kvm_has_tsc_control;
1165 /* maximum supported tsc_khz for guests */
1166 extern u32 kvm_max_guest_tsc_khz;
1167 /* number of bits of the fractional part of the TSC scaling ratio */
1168 extern u8 kvm_tsc_scaling_ratio_frac_bits;
1169 /* maximum allowed value of TSC scaling ratio */
1170 extern u64 kvm_max_tsc_scaling_ratio;
1171 /* 1ull << kvm_tsc_scaling_ratio_frac_bits */
1172 extern u64 kvm_default_tsc_scaling_ratio;
1174 extern u64 kvm_mce_cap_supported;
1176 enum emulation_result {
1177 EMULATE_DONE, /* no further processing */
1178 EMULATE_USER_EXIT, /* kvm_run ready for userspace exit */
1179 EMULATE_FAIL, /* can't emulate this instruction */
1182 #define EMULTYPE_NO_DECODE (1 << 0)
1183 #define EMULTYPE_TRAP_UD (1 << 1)
1184 #define EMULTYPE_SKIP (1 << 2)
1185 #define EMULTYPE_RETRY (1 << 3)
1186 #define EMULTYPE_NO_REEXECUTE (1 << 4)
1187 int x86_emulate_instruction(struct kvm_vcpu *vcpu, unsigned long cr2,
1188 int emulation_type, void *insn, int insn_len);
1190 static inline int emulate_instruction(struct kvm_vcpu *vcpu,
1193 return x86_emulate_instruction(vcpu, 0,
1194 emulation_type | EMULTYPE_NO_REEXECUTE, NULL, 0);
1197 void kvm_enable_efer_bits(u64);
1198 bool kvm_valid_efer(struct kvm_vcpu *vcpu, u64 efer);
1199 int kvm_get_msr(struct kvm_vcpu *vcpu, struct msr_data *msr);
1200 int kvm_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr);
1202 struct x86_emulate_ctxt;
1204 int kvm_fast_pio_out(struct kvm_vcpu *vcpu, int size, unsigned short port);
1205 int kvm_fast_pio_in(struct kvm_vcpu *vcpu, int size, unsigned short port);
1206 int kvm_emulate_cpuid(struct kvm_vcpu *vcpu);
1207 int kvm_emulate_halt(struct kvm_vcpu *vcpu);
1208 int kvm_vcpu_halt(struct kvm_vcpu *vcpu);
1209 int kvm_emulate_wbinvd(struct kvm_vcpu *vcpu);
1211 void kvm_get_segment(struct kvm_vcpu *vcpu, struct kvm_segment *var, int seg);
1212 int kvm_load_segment_descriptor(struct kvm_vcpu *vcpu, u16 selector, int seg);
1213 void kvm_vcpu_deliver_sipi_vector(struct kvm_vcpu *vcpu, u8 vector);
1215 int kvm_task_switch(struct kvm_vcpu *vcpu, u16 tss_selector, int idt_index,
1216 int reason, bool has_error_code, u32 error_code);
1218 int kvm_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0);
1219 int kvm_set_cr3(struct kvm_vcpu *vcpu, unsigned long cr3);
1220 int kvm_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4);
1221 int kvm_set_cr8(struct kvm_vcpu *vcpu, unsigned long cr8);
1222 int kvm_set_dr(struct kvm_vcpu *vcpu, int dr, unsigned long val);
1223 int kvm_get_dr(struct kvm_vcpu *vcpu, int dr, unsigned long *val);
1224 unsigned long kvm_get_cr8(struct kvm_vcpu *vcpu);
1225 void kvm_lmsw(struct kvm_vcpu *vcpu, unsigned long msw);
1226 void kvm_get_cs_db_l_bits(struct kvm_vcpu *vcpu, int *db, int *l);
1227 int kvm_set_xcr(struct kvm_vcpu *vcpu, u32 index, u64 xcr);
1229 int kvm_get_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr);
1230 int kvm_set_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr);
1232 unsigned long kvm_get_rflags(struct kvm_vcpu *vcpu);
1233 void kvm_set_rflags(struct kvm_vcpu *vcpu, unsigned long rflags);
1234 bool kvm_rdpmc(struct kvm_vcpu *vcpu);
1236 void kvm_queue_exception(struct kvm_vcpu *vcpu, unsigned nr);
1237 void kvm_queue_exception_e(struct kvm_vcpu *vcpu, unsigned nr, u32 error_code);
1238 void kvm_requeue_exception(struct kvm_vcpu *vcpu, unsigned nr);
1239 void kvm_requeue_exception_e(struct kvm_vcpu *vcpu, unsigned nr, u32 error_code);
1240 void kvm_inject_page_fault(struct kvm_vcpu *vcpu, struct x86_exception *fault);
1241 int kvm_read_guest_page_mmu(struct kvm_vcpu *vcpu, struct kvm_mmu *mmu,
1242 gfn_t gfn, void *data, int offset, int len,
1244 bool kvm_require_cpl(struct kvm_vcpu *vcpu, int required_cpl);
1245 bool kvm_require_dr(struct kvm_vcpu *vcpu, int dr);
1247 static inline int __kvm_irq_line_state(unsigned long *irq_state,
1248 int irq_source_id, int level)
1250 /* Logical OR for level trig interrupt */
1252 __set_bit(irq_source_id, irq_state);
1254 __clear_bit(irq_source_id, irq_state);
1256 return !!(*irq_state);
1259 int kvm_pic_set_irq(struct kvm_pic *pic, int irq, int irq_source_id, int level);
1260 void kvm_pic_clear_all(struct kvm_pic *pic, int irq_source_id);
1262 void kvm_inject_nmi(struct kvm_vcpu *vcpu);
1264 int kvm_mmu_unprotect_page(struct kvm *kvm, gfn_t gfn);
1265 int kvm_mmu_unprotect_page_virt(struct kvm_vcpu *vcpu, gva_t gva);
1266 void __kvm_mmu_free_some_pages(struct kvm_vcpu *vcpu);
1267 int kvm_mmu_load(struct kvm_vcpu *vcpu);
1268 void kvm_mmu_unload(struct kvm_vcpu *vcpu);
1269 void kvm_mmu_sync_roots(struct kvm_vcpu *vcpu);
1270 gpa_t translate_nested_gpa(struct kvm_vcpu *vcpu, gpa_t gpa, u32 access,
1271 struct x86_exception *exception);
1272 gpa_t kvm_mmu_gva_to_gpa_read(struct kvm_vcpu *vcpu, gva_t gva,
1273 struct x86_exception *exception);
1274 gpa_t kvm_mmu_gva_to_gpa_fetch(struct kvm_vcpu *vcpu, gva_t gva,
1275 struct x86_exception *exception);
1276 gpa_t kvm_mmu_gva_to_gpa_write(struct kvm_vcpu *vcpu, gva_t gva,
1277 struct x86_exception *exception);
1278 gpa_t kvm_mmu_gva_to_gpa_system(struct kvm_vcpu *vcpu, gva_t gva,
1279 struct x86_exception *exception);
1281 void kvm_vcpu_deactivate_apicv(struct kvm_vcpu *vcpu);
1283 int kvm_emulate_hypercall(struct kvm_vcpu *vcpu);
1285 int kvm_mmu_page_fault(struct kvm_vcpu *vcpu, gva_t gva, u64 error_code,
1286 void *insn, int insn_len);
1287 void kvm_mmu_invlpg(struct kvm_vcpu *vcpu, gva_t gva);
1288 void kvm_mmu_new_cr3(struct kvm_vcpu *vcpu);
1290 void kvm_enable_tdp(void);
1291 void kvm_disable_tdp(void);
1293 static inline gpa_t translate_gpa(struct kvm_vcpu *vcpu, gpa_t gpa, u32 access,
1294 struct x86_exception *exception)
1299 static inline struct kvm_mmu_page *page_header(hpa_t shadow_page)
1301 struct page *page = pfn_to_page(shadow_page >> PAGE_SHIFT);
1303 return (struct kvm_mmu_page *)page_private(page);
1306 static inline u16 kvm_read_ldt(void)
1309 asm("sldt %0" : "=g"(ldt));
1313 static inline void kvm_load_ldt(u16 sel)
1315 asm("lldt %0" : : "rm"(sel));
1318 #ifdef CONFIG_X86_64
1319 static inline unsigned long read_msr(unsigned long msr)
1328 static inline u32 get_rdx_init_val(void)
1330 return 0x600; /* P6 family */
1333 static inline void kvm_inject_gp(struct kvm_vcpu *vcpu, u32 error_code)
1335 kvm_queue_exception_e(vcpu, GP_VECTOR, error_code);
1338 #define TSS_IOPB_BASE_OFFSET 0x66
1339 #define TSS_BASE_SIZE 0x68
1340 #define TSS_IOPB_SIZE (65536 / 8)
1341 #define TSS_REDIRECTION_SIZE (256 / 8)
1342 #define RMODE_TSS_SIZE \
1343 (TSS_BASE_SIZE + TSS_REDIRECTION_SIZE + TSS_IOPB_SIZE + 1)
1346 TASK_SWITCH_CALL = 0,
1347 TASK_SWITCH_IRET = 1,
1348 TASK_SWITCH_JMP = 2,
1349 TASK_SWITCH_GATE = 3,
1352 #define HF_GIF_MASK (1 << 0)
1353 #define HF_HIF_MASK (1 << 1)
1354 #define HF_VINTR_MASK (1 << 2)
1355 #define HF_NMI_MASK (1 << 3)
1356 #define HF_IRET_MASK (1 << 4)
1357 #define HF_GUEST_MASK (1 << 5) /* VCPU is in guest-mode */
1358 #define HF_SMM_MASK (1 << 6)
1359 #define HF_SMM_INSIDE_NMI_MASK (1 << 7)
1361 #define __KVM_VCPU_MULTIPLE_ADDRESS_SPACE
1362 #define KVM_ADDRESS_SPACE_NUM 2
1364 #define kvm_arch_vcpu_memslots_id(vcpu) ((vcpu)->arch.hflags & HF_SMM_MASK ? 1 : 0)
1365 #define kvm_memslots_for_spte_role(kvm, role) __kvm_memslots(kvm, (role).smm)
1368 * Hardware virtualization extension instructions may fault if a
1369 * reboot turns off virtualization while processes are running.
1370 * Trap the fault and ignore the instruction if that happens.
1372 asmlinkage void kvm_spurious_fault(void);
1374 #define ____kvm_handle_fault_on_reboot(insn, cleanup_insn) \
1375 "666: " insn "\n\t" \
1377 ".pushsection .fixup, \"ax\" \n" \
1379 cleanup_insn "\n\t" \
1380 "cmpb $0, kvm_rebooting \n\t" \
1382 __ASM_SIZE(push) " $666b \n\t" \
1383 "call kvm_spurious_fault \n\t" \
1384 ".popsection \n\t" \
1385 _ASM_EXTABLE(666b, 667b)
1387 #define __kvm_handle_fault_on_reboot(insn) \
1388 ____kvm_handle_fault_on_reboot(insn, "")
1390 #define KVM_ARCH_WANT_MMU_NOTIFIER
1391 int kvm_unmap_hva(struct kvm *kvm, unsigned long hva);
1392 int kvm_unmap_hva_range(struct kvm *kvm, unsigned long start, unsigned long end);
1393 int kvm_age_hva(struct kvm *kvm, unsigned long start, unsigned long end);
1394 int kvm_test_age_hva(struct kvm *kvm, unsigned long hva);
1395 void kvm_set_spte_hva(struct kvm *kvm, unsigned long hva, pte_t pte);
1396 int kvm_cpu_has_injectable_intr(struct kvm_vcpu *v);
1397 int kvm_cpu_has_interrupt(struct kvm_vcpu *vcpu);
1398 int kvm_arch_interrupt_allowed(struct kvm_vcpu *vcpu);
1399 int kvm_cpu_get_interrupt(struct kvm_vcpu *v);
1400 void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event);
1401 void kvm_vcpu_reload_apic_access_page(struct kvm_vcpu *vcpu);
1403 void kvm_define_shared_msr(unsigned index, u32 msr);
1404 int kvm_set_shared_msr(unsigned index, u64 val, u64 mask);
1406 u64 kvm_scale_tsc(struct kvm_vcpu *vcpu, u64 tsc);
1407 u64 kvm_read_l1_tsc(struct kvm_vcpu *vcpu, u64 host_tsc);
1409 unsigned long kvm_get_linear_rip(struct kvm_vcpu *vcpu);
1410 bool kvm_is_linear_rip(struct kvm_vcpu *vcpu, unsigned long linear_rip);
1412 void kvm_make_mclock_inprogress_request(struct kvm *kvm);
1413 void kvm_make_scan_ioapic_request(struct kvm *kvm);
1415 void kvm_arch_async_page_not_present(struct kvm_vcpu *vcpu,
1416 struct kvm_async_pf *work);
1417 void kvm_arch_async_page_present(struct kvm_vcpu *vcpu,
1418 struct kvm_async_pf *work);
1419 void kvm_arch_async_page_ready(struct kvm_vcpu *vcpu,
1420 struct kvm_async_pf *work);
1421 bool kvm_arch_can_inject_async_page_present(struct kvm_vcpu *vcpu);
1422 extern bool kvm_find_async_pf_gfn(struct kvm_vcpu *vcpu, gfn_t gfn);
1424 int kvm_skip_emulated_instruction(struct kvm_vcpu *vcpu);
1425 int kvm_complete_insn_gp(struct kvm_vcpu *vcpu, int err);
1427 int kvm_is_in_guest(void);
1429 int __x86_set_memory_region(struct kvm *kvm, int id, gpa_t gpa, u32 size);
1430 int x86_set_memory_region(struct kvm *kvm, int id, gpa_t gpa, u32 size);
1431 bool kvm_vcpu_is_reset_bsp(struct kvm_vcpu *vcpu);
1432 bool kvm_vcpu_is_bsp(struct kvm_vcpu *vcpu);
1434 bool kvm_intr_is_single_vcpu(struct kvm *kvm, struct kvm_lapic_irq *irq,
1435 struct kvm_vcpu **dest_vcpu);
1437 void kvm_set_msi_irq(struct kvm *kvm, struct kvm_kernel_irq_routing_entry *e,
1438 struct kvm_lapic_irq *irq);
1440 static inline void kvm_arch_vcpu_blocking(struct kvm_vcpu *vcpu)
1442 if (kvm_x86_ops->vcpu_blocking)
1443 kvm_x86_ops->vcpu_blocking(vcpu);
1446 static inline void kvm_arch_vcpu_unblocking(struct kvm_vcpu *vcpu)
1448 if (kvm_x86_ops->vcpu_unblocking)
1449 kvm_x86_ops->vcpu_unblocking(vcpu);
1452 static inline void kvm_arch_vcpu_block_finish(struct kvm_vcpu *vcpu) {}
1454 static inline int kvm_cpu_get_apicid(int mps_cpu)
1456 #ifdef CONFIG_X86_LOCAL_APIC
1457 return default_cpu_present_to_apicid(mps_cpu);
1464 #define put_smstate(type, buf, offset, val) \
1465 *(type *)((buf) + (offset) - 0x7e00) = val
1467 #endif /* _ASM_X86_KVM_HOST_H */