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>
37 #include <asm/hyperv-tlfs.h>
39 #define KVM_MAX_VCPUS 288
40 #define KVM_SOFT_MAX_VCPUS 240
41 #define KVM_MAX_VCPU_ID 1023
42 #define KVM_USER_MEM_SLOTS 509
43 /* memory slots that are not exposed to userspace */
44 #define KVM_PRIVATE_MEM_SLOTS 3
45 #define KVM_MEM_SLOTS_NUM (KVM_USER_MEM_SLOTS + KVM_PRIVATE_MEM_SLOTS)
47 #define KVM_HALT_POLL_NS_DEFAULT 200000
49 #define KVM_IRQCHIP_NUM_PINS KVM_IOAPIC_NUM_PINS
51 /* x86-specific vcpu->requests bit members */
52 #define KVM_REQ_MIGRATE_TIMER KVM_ARCH_REQ(0)
53 #define KVM_REQ_REPORT_TPR_ACCESS KVM_ARCH_REQ(1)
54 #define KVM_REQ_TRIPLE_FAULT KVM_ARCH_REQ(2)
55 #define KVM_REQ_MMU_SYNC KVM_ARCH_REQ(3)
56 #define KVM_REQ_CLOCK_UPDATE KVM_ARCH_REQ(4)
57 #define KVM_REQ_EVENT KVM_ARCH_REQ(6)
58 #define KVM_REQ_APF_HALT KVM_ARCH_REQ(7)
59 #define KVM_REQ_STEAL_UPDATE KVM_ARCH_REQ(8)
60 #define KVM_REQ_NMI KVM_ARCH_REQ(9)
61 #define KVM_REQ_PMU KVM_ARCH_REQ(10)
62 #define KVM_REQ_PMI KVM_ARCH_REQ(11)
63 #define KVM_REQ_SMI KVM_ARCH_REQ(12)
64 #define KVM_REQ_MASTERCLOCK_UPDATE KVM_ARCH_REQ(13)
65 #define KVM_REQ_MCLOCK_INPROGRESS \
66 KVM_ARCH_REQ_FLAGS(14, KVM_REQUEST_WAIT | KVM_REQUEST_NO_WAKEUP)
67 #define KVM_REQ_SCAN_IOAPIC \
68 KVM_ARCH_REQ_FLAGS(15, KVM_REQUEST_WAIT | KVM_REQUEST_NO_WAKEUP)
69 #define KVM_REQ_GLOBAL_CLOCK_UPDATE KVM_ARCH_REQ(16)
70 #define KVM_REQ_APIC_PAGE_RELOAD \
71 KVM_ARCH_REQ_FLAGS(17, KVM_REQUEST_WAIT | KVM_REQUEST_NO_WAKEUP)
72 #define KVM_REQ_HV_CRASH KVM_ARCH_REQ(18)
73 #define KVM_REQ_IOAPIC_EOI_EXIT KVM_ARCH_REQ(19)
74 #define KVM_REQ_HV_RESET KVM_ARCH_REQ(20)
75 #define KVM_REQ_HV_EXIT KVM_ARCH_REQ(21)
76 #define KVM_REQ_HV_STIMER KVM_ARCH_REQ(22)
77 #define KVM_REQ_LOAD_EOI_EXITMAP KVM_ARCH_REQ(23)
78 #define KVM_REQ_GET_VMCS12_PAGES KVM_ARCH_REQ(24)
80 #define CR0_RESERVED_BITS \
81 (~(unsigned long)(X86_CR0_PE | X86_CR0_MP | X86_CR0_EM | X86_CR0_TS \
82 | X86_CR0_ET | X86_CR0_NE | X86_CR0_WP | X86_CR0_AM \
83 | X86_CR0_NW | X86_CR0_CD | X86_CR0_PG))
85 #define CR3_PCID_INVD BIT_64(63)
86 #define CR4_RESERVED_BITS \
87 (~(unsigned long)(X86_CR4_VME | X86_CR4_PVI | X86_CR4_TSD | X86_CR4_DE\
88 | X86_CR4_PSE | X86_CR4_PAE | X86_CR4_MCE \
89 | X86_CR4_PGE | X86_CR4_PCE | X86_CR4_OSFXSR | X86_CR4_PCIDE \
90 | X86_CR4_OSXSAVE | X86_CR4_SMEP | X86_CR4_FSGSBASE \
91 | X86_CR4_OSXMMEXCPT | X86_CR4_LA57 | X86_CR4_VMXE \
92 | X86_CR4_SMAP | X86_CR4_PKE | X86_CR4_UMIP))
94 #define CR8_RESERVED_BITS (~(unsigned long)X86_CR8_TPR)
98 #define INVALID_PAGE (~(hpa_t)0)
99 #define VALID_PAGE(x) ((x) != INVALID_PAGE)
101 #define UNMAPPED_GVA (~(gpa_t)0)
103 /* KVM Hugepage definitions for x86 */
104 #define KVM_NR_PAGE_SIZES 3
105 #define KVM_HPAGE_GFN_SHIFT(x) (((x) - 1) * 9)
106 #define KVM_HPAGE_SHIFT(x) (PAGE_SHIFT + KVM_HPAGE_GFN_SHIFT(x))
107 #define KVM_HPAGE_SIZE(x) (1UL << KVM_HPAGE_SHIFT(x))
108 #define KVM_HPAGE_MASK(x) (~(KVM_HPAGE_SIZE(x) - 1))
109 #define KVM_PAGES_PER_HPAGE(x) (KVM_HPAGE_SIZE(x) / PAGE_SIZE)
111 static inline gfn_t gfn_to_index(gfn_t gfn, gfn_t base_gfn, int level)
113 /* KVM_HPAGE_GFN_SHIFT(PT_PAGE_TABLE_LEVEL) must be 0. */
114 return (gfn >> KVM_HPAGE_GFN_SHIFT(level)) -
115 (base_gfn >> KVM_HPAGE_GFN_SHIFT(level));
118 #define KVM_PERMILLE_MMU_PAGES 20
119 #define KVM_MIN_ALLOC_MMU_PAGES 64
120 #define KVM_MMU_HASH_SHIFT 12
121 #define KVM_NUM_MMU_PAGES (1 << KVM_MMU_HASH_SHIFT)
122 #define KVM_MIN_FREE_MMU_PAGES 5
123 #define KVM_REFILL_PAGES 25
124 #define KVM_MAX_CPUID_ENTRIES 80
125 #define KVM_NR_FIXED_MTRR_REGION 88
126 #define KVM_NR_VAR_MTRR 8
128 #define ASYNC_PF_PER_VCPU 64
154 VCPU_EXREG_PDPTR = NR_VCPU_REGS,
171 #include <asm/kvm_emulate.h>
173 #define KVM_NR_MEM_OBJS 40
175 #define KVM_NR_DB_REGS 4
177 #define DR6_BD (1 << 13)
178 #define DR6_BS (1 << 14)
179 #define DR6_RTM (1 << 16)
180 #define DR6_FIXED_1 0xfffe0ff0
181 #define DR6_INIT 0xffff0ff0
182 #define DR6_VOLATILE 0x0001e00f
184 #define DR7_BP_EN_MASK 0x000000ff
185 #define DR7_GE (1 << 9)
186 #define DR7_GD (1 << 13)
187 #define DR7_FIXED_1 0x00000400
188 #define DR7_VOLATILE 0xffff2bff
190 #define PFERR_PRESENT_BIT 0
191 #define PFERR_WRITE_BIT 1
192 #define PFERR_USER_BIT 2
193 #define PFERR_RSVD_BIT 3
194 #define PFERR_FETCH_BIT 4
195 #define PFERR_PK_BIT 5
196 #define PFERR_GUEST_FINAL_BIT 32
197 #define PFERR_GUEST_PAGE_BIT 33
199 #define PFERR_PRESENT_MASK (1U << PFERR_PRESENT_BIT)
200 #define PFERR_WRITE_MASK (1U << PFERR_WRITE_BIT)
201 #define PFERR_USER_MASK (1U << PFERR_USER_BIT)
202 #define PFERR_RSVD_MASK (1U << PFERR_RSVD_BIT)
203 #define PFERR_FETCH_MASK (1U << PFERR_FETCH_BIT)
204 #define PFERR_PK_MASK (1U << PFERR_PK_BIT)
205 #define PFERR_GUEST_FINAL_MASK (1ULL << PFERR_GUEST_FINAL_BIT)
206 #define PFERR_GUEST_PAGE_MASK (1ULL << PFERR_GUEST_PAGE_BIT)
208 #define PFERR_NESTED_GUEST_PAGE (PFERR_GUEST_PAGE_MASK | \
213 * The mask used to denote special SPTEs, which can be either MMIO SPTEs or
214 * Access Tracking SPTEs. We use bit 62 instead of bit 63 to avoid conflicting
215 * with the SVE bit in EPT PTEs.
217 #define SPTE_SPECIAL_MASK (1ULL << 62)
219 /* apic attention bits */
220 #define KVM_APIC_CHECK_VAPIC 0
222 * The following bit is set with PV-EOI, unset on EOI.
223 * We detect PV-EOI changes by guest by comparing
224 * this bit with PV-EOI in guest memory.
225 * See the implementation in apic_update_pv_eoi.
227 #define KVM_APIC_PV_EOI_PENDING 1
229 struct kvm_kernel_irq_routing_entry;
232 * We don't want allocation failures within the mmu code, so we preallocate
233 * enough memory for a single page fault in a cache.
235 struct kvm_mmu_memory_cache {
237 void *objects[KVM_NR_MEM_OBJS];
241 * the pages used as guest page table on soft mmu are tracked by
242 * kvm_memory_slot.arch.gfn_track which is 16 bits, so the role bits used
243 * by indirect shadow page can not be more than 15 bits.
245 * Currently, we used 14 bits that are @level, @cr4_pae, @quadrant, @access,
246 * @nxe, @cr0_wp, @smep_andnot_wp and @smap_andnot_wp.
248 union kvm_mmu_page_role {
259 unsigned smep_andnot_wp:1;
260 unsigned smap_andnot_wp:1;
261 unsigned ad_disabled:1;
262 unsigned guest_mode:1;
266 * This is left at the top of the word so that
267 * kvm_memslots_for_spte_role can extract it with a
268 * simple shift. While there is room, give it a whole
269 * byte so it is also faster to load it from memory.
275 struct kvm_rmap_head {
279 struct kvm_mmu_page {
280 struct list_head link;
281 struct hlist_node hash_link;
284 * The following two entries are used to key the shadow page in the
288 union kvm_mmu_page_role role;
291 /* hold the gfn of each spte inside spt */
294 int root_count; /* Currently serving as active root */
295 unsigned int unsync_children;
296 struct kvm_rmap_head parent_ptes; /* rmap pointers to parent sptes */
298 /* The page is obsolete if mmu_valid_gen != kvm->arch.mmu_valid_gen. */
299 unsigned long mmu_valid_gen;
301 DECLARE_BITMAP(unsync_child_bitmap, 512);
305 * Used out of the mmu-lock to avoid reading spte values while an
306 * update is in progress; see the comments in __get_spte_lockless().
308 int clear_spte_count;
311 /* Number of writes since the last time traversal visited this page. */
312 atomic_t write_flooding_count;
315 struct kvm_pio_request {
322 #define PT64_ROOT_MAX_LEVEL 5
324 struct rsvd_bits_validate {
325 u64 rsvd_bits_mask[2][PT64_ROOT_MAX_LEVEL];
330 * x86 supports 4 paging modes (5-level 64-bit, 4-level 64-bit, 3-level 32-bit,
331 * and 2-level 32-bit). The kvm_mmu structure abstracts the details of the
335 void (*set_cr3)(struct kvm_vcpu *vcpu, unsigned long root);
336 unsigned long (*get_cr3)(struct kvm_vcpu *vcpu);
337 u64 (*get_pdptr)(struct kvm_vcpu *vcpu, int index);
338 int (*page_fault)(struct kvm_vcpu *vcpu, gva_t gva, u32 err,
340 void (*inject_page_fault)(struct kvm_vcpu *vcpu,
341 struct x86_exception *fault);
342 gpa_t (*gva_to_gpa)(struct kvm_vcpu *vcpu, gva_t gva, u32 access,
343 struct x86_exception *exception);
344 gpa_t (*translate_gpa)(struct kvm_vcpu *vcpu, gpa_t gpa, u32 access,
345 struct x86_exception *exception);
346 int (*sync_page)(struct kvm_vcpu *vcpu,
347 struct kvm_mmu_page *sp);
348 void (*invlpg)(struct kvm_vcpu *vcpu, gva_t gva);
349 void (*update_pte)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp,
350 u64 *spte, const void *pte);
352 union kvm_mmu_page_role base_role;
354 u8 shadow_root_level;
359 * Bitmap; bit set = permission fault
360 * Byte index: page fault error code [4:1]
361 * Bit index: pte permissions in ACC_* format
366 * The pkru_mask indicates if protection key checks are needed. It
367 * consists of 16 domains indexed by page fault error code bits [4:1],
368 * with PFEC.RSVD replaced by ACC_USER_MASK from the page tables.
369 * Each domain has 2 bits which are ANDed with AD and WD from PKRU.
377 * check zero bits on shadow page table entries, these
378 * bits include not only hardware reserved bits but also
379 * the bits spte never used.
381 struct rsvd_bits_validate shadow_zero_check;
383 struct rsvd_bits_validate guest_rsvd_check;
385 /* Can have large pages at levels 2..last_nonleaf_level-1. */
386 u8 last_nonleaf_level;
390 u64 pdptrs[4]; /* pae */
403 struct perf_event *perf_event;
404 struct kvm_vcpu *vcpu;
408 unsigned nr_arch_gp_counters;
409 unsigned nr_arch_fixed_counters;
410 unsigned available_event_types;
415 u64 counter_bitmask[2];
416 u64 global_ctrl_mask;
419 struct kvm_pmc gp_counters[INTEL_PMC_MAX_GENERIC];
420 struct kvm_pmc fixed_counters[INTEL_PMC_MAX_FIXED];
421 struct irq_work irq_work;
428 KVM_DEBUGREG_BP_ENABLED = 1,
429 KVM_DEBUGREG_WONT_EXIT = 2,
430 KVM_DEBUGREG_RELOAD = 4,
433 struct kvm_mtrr_range {
436 struct list_head node;
440 struct kvm_mtrr_range var_ranges[KVM_NR_VAR_MTRR];
441 mtrr_type fixed_ranges[KVM_NR_FIXED_MTRR_REGION];
444 struct list_head head;
447 /* Hyper-V SynIC timer */
448 struct kvm_vcpu_hv_stimer {
449 struct hrtimer timer;
454 struct hv_message msg;
458 /* Hyper-V synthetic interrupt controller (SynIC)*/
459 struct kvm_vcpu_hv_synic {
464 atomic64_t sint[HV_SYNIC_SINT_COUNT];
465 atomic_t sint_to_gsi[HV_SYNIC_SINT_COUNT];
466 DECLARE_BITMAP(auto_eoi_bitmap, 256);
467 DECLARE_BITMAP(vec_bitmap, 256);
469 bool dont_zero_synic_pages;
472 /* Hyper-V per vcpu emulation context */
477 struct kvm_vcpu_hv_synic synic;
478 struct kvm_hyperv_exit exit;
479 struct kvm_vcpu_hv_stimer stimer[HV_SYNIC_STIMER_COUNT];
480 DECLARE_BITMAP(stimer_pending_bitmap, HV_SYNIC_STIMER_COUNT);
484 struct kvm_vcpu_arch {
486 * rip and regs accesses must go through
487 * kvm_{register,rip}_{read,write} functions.
489 unsigned long regs[NR_VCPU_REGS];
494 unsigned long cr0_guest_owned_bits;
498 unsigned long cr4_guest_owned_bits;
504 struct kvm_lapic *apic; /* kernel irqchip context */
506 bool load_eoi_exitmap_pending;
507 DECLARE_BITMAP(ioapic_handled_vectors, 256);
508 unsigned long apic_attention;
509 int32_t apic_arb_prio;
511 u64 ia32_misc_enable_msr;
514 bool tpr_access_reporting;
516 u64 microcode_version;
519 * Paging state of the vcpu
521 * If the vcpu runs in guest mode with two level paging this still saves
522 * the paging mode of the l1 guest. This context is always used to
528 * Paging state of an L2 guest (used for nested npt)
530 * This context will save all necessary information to walk page tables
531 * of the an L2 guest. This context is only initialized for page table
532 * walking and not for faulting since we never handle l2 page faults on
535 struct kvm_mmu nested_mmu;
538 * Pointer to the mmu context currently used for
539 * gva_to_gpa translations.
541 struct kvm_mmu *walk_mmu;
543 struct kvm_mmu_memory_cache mmu_pte_list_desc_cache;
544 struct kvm_mmu_memory_cache mmu_page_cache;
545 struct kvm_mmu_memory_cache mmu_page_header_cache;
548 * QEMU userspace and the guest each have their own FPU state.
549 * In vcpu_run, we switch between the user and guest FPU contexts.
550 * While running a VCPU, the VCPU thread will have the guest FPU
553 * Note that while the PKRU state lives inside the fpu registers,
554 * it is switched out separately at VMENTER and VMEXIT time. The
555 * "guest_fpu" state here contains the guest FPU context, with the
559 struct fpu guest_fpu;
562 u64 guest_supported_xcr0;
563 u32 guest_xstate_size;
565 struct kvm_pio_request pio;
568 u8 event_exit_inst_len;
570 struct kvm_queued_exception {
579 struct kvm_queued_interrupt {
585 int halt_request; /* real mode on Intel only */
588 struct kvm_cpuid_entry2 cpuid_entries[KVM_MAX_CPUID_ENTRIES];
592 /* emulate context */
594 struct x86_emulate_ctxt emulate_ctxt;
595 bool emulate_regs_need_sync_to_vcpu;
596 bool emulate_regs_need_sync_from_vcpu;
597 int (*complete_userspace_io)(struct kvm_vcpu *vcpu);
600 struct pvclock_vcpu_time_info hv_clock;
601 unsigned int hw_tsc_khz;
602 struct gfn_to_hva_cache pv_time;
603 bool pv_time_enabled;
604 /* set guest stopped flag in pvclock flags field */
605 bool pvclock_set_guest_stopped_request;
610 struct gfn_to_hva_cache stime;
611 struct kvm_steal_time steal;
617 u64 tsc_offset_adjustment;
620 u64 this_tsc_generation;
622 bool tsc_always_catchup;
623 s8 virtual_tsc_shift;
624 u32 virtual_tsc_mult;
626 s64 ia32_tsc_adjust_msr;
627 u64 tsc_scaling_ratio;
629 atomic_t nmi_queued; /* unprocessed asynchronous NMIs */
630 unsigned nmi_pending; /* NMI queued after currently running handler */
631 bool nmi_injected; /* Trying to inject an NMI this entry */
632 bool smi_pending; /* SMI queued after currently running handler */
634 struct kvm_mtrr mtrr_state;
637 unsigned switch_db_regs;
638 unsigned long db[KVM_NR_DB_REGS];
641 unsigned long eff_db[KVM_NR_DB_REGS];
642 unsigned long guest_debug_dr7;
643 u64 msr_platform_info;
644 u64 msr_misc_features_enables;
652 /* Cache MMIO info */
660 /* used for guest single stepping over the given code position */
661 unsigned long singlestep_rip;
663 struct kvm_vcpu_hv hyperv;
665 cpumask_var_t wbinvd_dirty_mask;
667 unsigned long last_retry_eip;
668 unsigned long last_retry_addr;
672 gfn_t gfns[roundup_pow_of_two(ASYNC_PF_PER_VCPU)];
673 struct gfn_to_hva_cache data;
678 unsigned long nested_apf_token;
679 bool delivery_as_pf_vmexit;
682 /* OSVW MSRs (AMD only) */
690 struct gfn_to_hva_cache data;
694 * Indicate whether the access faults on its page table in guest
695 * which is set when fix page fault and used to detect unhandeable
698 bool write_fault_to_shadow_pgtable;
700 /* set at EPT violation at this point */
701 unsigned long exit_qualification;
703 /* pv related host specific info */
708 int pending_ioapic_eoi;
709 int pending_external_vector;
715 /* be preempted when it's in kernel-mode(cpl=0) */
716 bool preempted_in_kernel;
719 struct kvm_lpage_info {
723 struct kvm_arch_memory_slot {
724 struct kvm_rmap_head *rmap[KVM_NR_PAGE_SIZES];
725 struct kvm_lpage_info *lpage_info[KVM_NR_PAGE_SIZES - 1];
726 unsigned short *gfn_track[KVM_PAGE_TRACK_MAX];
730 * We use as the mode the number of bits allocated in the LDR for the
731 * logical processor ID. It happens that these are all powers of two.
732 * This makes it is very easy to detect cases where the APICs are
733 * configured for multiple modes; in that case, we cannot use the map and
734 * hence cannot use kvm_irq_delivery_to_apic_fast either.
736 #define KVM_APIC_MODE_XAPIC_CLUSTER 4
737 #define KVM_APIC_MODE_XAPIC_FLAT 8
738 #define KVM_APIC_MODE_X2APIC 16
740 struct kvm_apic_map {
745 struct kvm_lapic *xapic_flat_map[8];
746 struct kvm_lapic *xapic_cluster_map[16][4];
748 struct kvm_lapic *phys_map[];
751 /* Hyper-V emulation context */
753 struct mutex hv_lock;
758 /* Hyper-v based guest crash (NT kernel bugcheck) parameters */
759 u64 hv_crash_param[HV_X64_MSR_CRASH_PARAMS];
762 HV_REFERENCE_TSC_PAGE tsc_ref;
764 struct idr conn_to_evt;
766 u64 hv_reenlightenment_control;
767 u64 hv_tsc_emulation_control;
768 u64 hv_tsc_emulation_status;
771 enum kvm_irqchip_mode {
773 KVM_IRQCHIP_KERNEL, /* created with KVM_CREATE_IRQCHIP */
774 KVM_IRQCHIP_SPLIT, /* created with KVM_CAP_SPLIT_IRQCHIP */
778 unsigned int n_used_mmu_pages;
779 unsigned int n_requested_mmu_pages;
780 unsigned int n_max_mmu_pages;
781 unsigned int indirect_shadow_pages;
782 unsigned long mmu_valid_gen;
783 struct hlist_head mmu_page_hash[KVM_NUM_MMU_PAGES];
785 * Hash table of struct kvm_mmu_page.
787 struct list_head active_mmu_pages;
788 struct list_head zapped_obsolete_pages;
789 struct kvm_page_track_notifier_node mmu_sp_tracker;
790 struct kvm_page_track_notifier_head track_notifier_head;
792 struct list_head assigned_dev_head;
793 struct iommu_domain *iommu_domain;
794 bool iommu_noncoherent;
795 #define __KVM_HAVE_ARCH_NONCOHERENT_DMA
796 atomic_t noncoherent_dma_count;
797 #define __KVM_HAVE_ARCH_ASSIGNED_DEVICE
798 atomic_t assigned_device_count;
799 struct kvm_pic *vpic;
800 struct kvm_ioapic *vioapic;
801 struct kvm_pit *vpit;
802 atomic_t vapics_in_nmi_mode;
803 struct mutex apic_map_lock;
804 struct kvm_apic_map *apic_map;
806 bool apic_access_page_done;
814 unsigned long irq_sources_bitmap;
816 raw_spinlock_t tsc_write_lock;
823 u64 cur_tsc_generation;
824 int nr_vcpus_matched_tsc;
826 spinlock_t pvclock_gtod_sync_lock;
827 bool use_master_clock;
828 u64 master_kernel_ns;
829 u64 master_cycle_now;
830 struct delayed_work kvmclock_update_work;
831 struct delayed_work kvmclock_sync_work;
833 struct kvm_xen_hvm_config xen_hvm_config;
835 /* reads protected by irq_srcu, writes by irq_lock */
836 struct hlist_head mask_notifier_list;
838 struct kvm_hv hyperv;
840 #ifdef CONFIG_KVM_MMU_AUDIT
844 bool backwards_tsc_observed;
845 bool boot_vcpu_runs_old_kvmclock;
850 enum kvm_irqchip_mode irqchip_mode;
851 u8 nr_reserved_ioapic_pins;
853 bool disabled_lapic_found;
856 bool x2apic_broadcast_quirk_disabled;
860 ulong mmu_shadow_zapped;
862 ulong mmu_pte_updated;
863 ulong mmu_pde_zapped;
866 ulong mmu_cache_miss;
868 ulong remote_tlb_flush;
870 ulong max_mmu_page_hash_collisions;
873 struct kvm_vcpu_stat {
883 u64 irq_window_exits;
884 u64 nmi_window_exits;
886 u64 halt_successful_poll;
887 u64 halt_attempted_poll;
888 u64 halt_poll_invalid;
890 u64 request_irq_exits;
892 u64 host_state_reload;
895 u64 insn_emulation_fail;
902 struct x86_instruction_info;
910 struct kvm_lapic_irq {
922 int (*cpu_has_kvm_support)(void); /* __init */
923 int (*disabled_by_bios)(void); /* __init */
924 int (*hardware_enable)(void);
925 void (*hardware_disable)(void);
926 void (*check_processor_compatibility)(void *rtn);
927 int (*hardware_setup)(void); /* __init */
928 void (*hardware_unsetup)(void); /* __exit */
929 bool (*cpu_has_accelerated_tpr)(void);
930 bool (*has_emulated_msr)(int index);
931 void (*cpuid_update)(struct kvm_vcpu *vcpu);
933 struct kvm *(*vm_alloc)(void);
934 void (*vm_free)(struct kvm *);
935 int (*vm_init)(struct kvm *kvm);
936 void (*vm_destroy)(struct kvm *kvm);
938 /* Create, but do not attach this VCPU */
939 struct kvm_vcpu *(*vcpu_create)(struct kvm *kvm, unsigned id);
940 void (*vcpu_free)(struct kvm_vcpu *vcpu);
941 void (*vcpu_reset)(struct kvm_vcpu *vcpu, bool init_event);
943 void (*prepare_guest_switch)(struct kvm_vcpu *vcpu);
944 void (*vcpu_load)(struct kvm_vcpu *vcpu, int cpu);
945 void (*vcpu_put)(struct kvm_vcpu *vcpu);
947 void (*update_bp_intercept)(struct kvm_vcpu *vcpu);
948 int (*get_msr)(struct kvm_vcpu *vcpu, struct msr_data *msr);
949 int (*set_msr)(struct kvm_vcpu *vcpu, struct msr_data *msr);
950 u64 (*get_segment_base)(struct kvm_vcpu *vcpu, int seg);
951 void (*get_segment)(struct kvm_vcpu *vcpu,
952 struct kvm_segment *var, int seg);
953 int (*get_cpl)(struct kvm_vcpu *vcpu);
954 void (*set_segment)(struct kvm_vcpu *vcpu,
955 struct kvm_segment *var, int seg);
956 void (*get_cs_db_l_bits)(struct kvm_vcpu *vcpu, int *db, int *l);
957 void (*decache_cr0_guest_bits)(struct kvm_vcpu *vcpu);
958 void (*decache_cr3)(struct kvm_vcpu *vcpu);
959 void (*decache_cr4_guest_bits)(struct kvm_vcpu *vcpu);
960 void (*set_cr0)(struct kvm_vcpu *vcpu, unsigned long cr0);
961 void (*set_cr3)(struct kvm_vcpu *vcpu, unsigned long cr3);
962 int (*set_cr4)(struct kvm_vcpu *vcpu, unsigned long cr4);
963 void (*set_efer)(struct kvm_vcpu *vcpu, u64 efer);
964 void (*get_idt)(struct kvm_vcpu *vcpu, struct desc_ptr *dt);
965 void (*set_idt)(struct kvm_vcpu *vcpu, struct desc_ptr *dt);
966 void (*get_gdt)(struct kvm_vcpu *vcpu, struct desc_ptr *dt);
967 void (*set_gdt)(struct kvm_vcpu *vcpu, struct desc_ptr *dt);
968 u64 (*get_dr6)(struct kvm_vcpu *vcpu);
969 void (*set_dr6)(struct kvm_vcpu *vcpu, unsigned long value);
970 void (*sync_dirty_debug_regs)(struct kvm_vcpu *vcpu);
971 void (*set_dr7)(struct kvm_vcpu *vcpu, unsigned long value);
972 void (*cache_reg)(struct kvm_vcpu *vcpu, enum kvm_reg reg);
973 unsigned long (*get_rflags)(struct kvm_vcpu *vcpu);
974 void (*set_rflags)(struct kvm_vcpu *vcpu, unsigned long rflags);
976 void (*tlb_flush)(struct kvm_vcpu *vcpu, bool invalidate_gpa);
978 void (*run)(struct kvm_vcpu *vcpu);
979 int (*handle_exit)(struct kvm_vcpu *vcpu);
980 void (*skip_emulated_instruction)(struct kvm_vcpu *vcpu);
981 void (*set_interrupt_shadow)(struct kvm_vcpu *vcpu, int mask);
982 u32 (*get_interrupt_shadow)(struct kvm_vcpu *vcpu);
983 void (*patch_hypercall)(struct kvm_vcpu *vcpu,
984 unsigned char *hypercall_addr);
985 void (*set_irq)(struct kvm_vcpu *vcpu);
986 void (*set_nmi)(struct kvm_vcpu *vcpu);
987 void (*queue_exception)(struct kvm_vcpu *vcpu);
988 void (*cancel_injection)(struct kvm_vcpu *vcpu);
989 int (*interrupt_allowed)(struct kvm_vcpu *vcpu);
990 int (*nmi_allowed)(struct kvm_vcpu *vcpu);
991 bool (*get_nmi_mask)(struct kvm_vcpu *vcpu);
992 void (*set_nmi_mask)(struct kvm_vcpu *vcpu, bool masked);
993 void (*enable_nmi_window)(struct kvm_vcpu *vcpu);
994 void (*enable_irq_window)(struct kvm_vcpu *vcpu);
995 void (*update_cr8_intercept)(struct kvm_vcpu *vcpu, int tpr, int irr);
996 bool (*get_enable_apicv)(struct kvm_vcpu *vcpu);
997 void (*refresh_apicv_exec_ctrl)(struct kvm_vcpu *vcpu);
998 void (*hwapic_irr_update)(struct kvm_vcpu *vcpu, int max_irr);
999 void (*hwapic_isr_update)(struct kvm_vcpu *vcpu, int isr);
1000 void (*load_eoi_exitmap)(struct kvm_vcpu *vcpu, u64 *eoi_exit_bitmap);
1001 void (*set_virtual_apic_mode)(struct kvm_vcpu *vcpu);
1002 void (*set_apic_access_page_addr)(struct kvm_vcpu *vcpu, hpa_t hpa);
1003 void (*deliver_posted_interrupt)(struct kvm_vcpu *vcpu, int vector);
1004 int (*sync_pir_to_irr)(struct kvm_vcpu *vcpu);
1005 int (*set_tss_addr)(struct kvm *kvm, unsigned int addr);
1006 int (*set_identity_map_addr)(struct kvm *kvm, u64 ident_addr);
1007 int (*get_tdp_level)(struct kvm_vcpu *vcpu);
1008 u64 (*get_mt_mask)(struct kvm_vcpu *vcpu, gfn_t gfn, bool is_mmio);
1009 int (*get_lpage_level)(void);
1010 bool (*rdtscp_supported)(void);
1011 bool (*invpcid_supported)(void);
1013 void (*set_tdp_cr3)(struct kvm_vcpu *vcpu, unsigned long cr3);
1015 void (*set_supported_cpuid)(u32 func, struct kvm_cpuid_entry2 *entry);
1017 bool (*has_wbinvd_exit)(void);
1019 u64 (*read_l1_tsc_offset)(struct kvm_vcpu *vcpu);
1020 void (*write_tsc_offset)(struct kvm_vcpu *vcpu, u64 offset);
1022 void (*get_exit_info)(struct kvm_vcpu *vcpu, u64 *info1, u64 *info2);
1024 int (*check_intercept)(struct kvm_vcpu *vcpu,
1025 struct x86_instruction_info *info,
1026 enum x86_intercept_stage stage);
1027 void (*handle_external_intr)(struct kvm_vcpu *vcpu);
1028 bool (*mpx_supported)(void);
1029 bool (*xsaves_supported)(void);
1030 bool (*umip_emulated)(void);
1032 int (*check_nested_events)(struct kvm_vcpu *vcpu, bool external_intr);
1034 void (*sched_in)(struct kvm_vcpu *kvm, int cpu);
1037 * Arch-specific dirty logging hooks. These hooks are only supposed to
1038 * be valid if the specific arch has hardware-accelerated dirty logging
1039 * mechanism. Currently only for PML on VMX.
1041 * - slot_enable_log_dirty:
1042 * called when enabling log dirty mode for the slot.
1043 * - slot_disable_log_dirty:
1044 * called when disabling log dirty mode for the slot.
1045 * also called when slot is created with log dirty disabled.
1046 * - flush_log_dirty:
1047 * called before reporting dirty_bitmap to userspace.
1048 * - enable_log_dirty_pt_masked:
1049 * called when reenabling log dirty for the GFNs in the mask after
1050 * corresponding bits are cleared in slot->dirty_bitmap.
1052 void (*slot_enable_log_dirty)(struct kvm *kvm,
1053 struct kvm_memory_slot *slot);
1054 void (*slot_disable_log_dirty)(struct kvm *kvm,
1055 struct kvm_memory_slot *slot);
1056 void (*flush_log_dirty)(struct kvm *kvm);
1057 void (*enable_log_dirty_pt_masked)(struct kvm *kvm,
1058 struct kvm_memory_slot *slot,
1059 gfn_t offset, unsigned long mask);
1060 int (*write_log_dirty)(struct kvm_vcpu *vcpu);
1062 /* pmu operations of sub-arch */
1063 const struct kvm_pmu_ops *pmu_ops;
1066 * Architecture specific hooks for vCPU blocking due to
1068 * Returns for .pre_block():
1069 * - 0 means continue to block the vCPU.
1070 * - 1 means we cannot block the vCPU since some event
1071 * happens during this period, such as, 'ON' bit in
1072 * posted-interrupts descriptor is set.
1074 int (*pre_block)(struct kvm_vcpu *vcpu);
1075 void (*post_block)(struct kvm_vcpu *vcpu);
1077 void (*vcpu_blocking)(struct kvm_vcpu *vcpu);
1078 void (*vcpu_unblocking)(struct kvm_vcpu *vcpu);
1080 int (*update_pi_irte)(struct kvm *kvm, unsigned int host_irq,
1081 uint32_t guest_irq, bool set);
1082 void (*apicv_post_state_restore)(struct kvm_vcpu *vcpu);
1084 int (*set_hv_timer)(struct kvm_vcpu *vcpu, u64 guest_deadline_tsc);
1085 void (*cancel_hv_timer)(struct kvm_vcpu *vcpu);
1087 void (*setup_mce)(struct kvm_vcpu *vcpu);
1089 void (*get_vmcs12_pages)(struct kvm_vcpu *vcpu);
1091 int (*smi_allowed)(struct kvm_vcpu *vcpu);
1092 int (*pre_enter_smm)(struct kvm_vcpu *vcpu, char *smstate);
1093 int (*pre_leave_smm)(struct kvm_vcpu *vcpu, u64 smbase);
1094 int (*enable_smi_window)(struct kvm_vcpu *vcpu);
1096 int (*mem_enc_op)(struct kvm *kvm, void __user *argp);
1097 int (*mem_enc_reg_region)(struct kvm *kvm, struct kvm_enc_region *argp);
1098 int (*mem_enc_unreg_region)(struct kvm *kvm, struct kvm_enc_region *argp);
1100 int (*get_msr_feature)(struct kvm_msr_entry *entry);
1103 struct kvm_arch_async_pf {
1110 extern struct kvm_x86_ops *kvm_x86_ops;
1112 #define __KVM_HAVE_ARCH_VM_ALLOC
1113 static inline struct kvm *kvm_arch_alloc_vm(void)
1115 return kvm_x86_ops->vm_alloc();
1118 static inline void kvm_arch_free_vm(struct kvm *kvm)
1120 return kvm_x86_ops->vm_free(kvm);
1123 int kvm_mmu_module_init(void);
1124 void kvm_mmu_module_exit(void);
1126 void kvm_mmu_destroy(struct kvm_vcpu *vcpu);
1127 int kvm_mmu_create(struct kvm_vcpu *vcpu);
1128 void kvm_mmu_setup(struct kvm_vcpu *vcpu);
1129 void kvm_mmu_init_vm(struct kvm *kvm);
1130 void kvm_mmu_uninit_vm(struct kvm *kvm);
1131 void kvm_mmu_set_mask_ptes(u64 user_mask, u64 accessed_mask,
1132 u64 dirty_mask, u64 nx_mask, u64 x_mask, u64 p_mask,
1133 u64 acc_track_mask, u64 me_mask);
1135 void kvm_mmu_reset_context(struct kvm_vcpu *vcpu);
1136 void kvm_mmu_slot_remove_write_access(struct kvm *kvm,
1137 struct kvm_memory_slot *memslot);
1138 void kvm_mmu_zap_collapsible_sptes(struct kvm *kvm,
1139 const struct kvm_memory_slot *memslot);
1140 void kvm_mmu_slot_leaf_clear_dirty(struct kvm *kvm,
1141 struct kvm_memory_slot *memslot);
1142 void kvm_mmu_slot_largepage_remove_write_access(struct kvm *kvm,
1143 struct kvm_memory_slot *memslot);
1144 void kvm_mmu_slot_set_dirty(struct kvm *kvm,
1145 struct kvm_memory_slot *memslot);
1146 void kvm_mmu_clear_dirty_pt_masked(struct kvm *kvm,
1147 struct kvm_memory_slot *slot,
1148 gfn_t gfn_offset, unsigned long mask);
1149 void kvm_mmu_zap_all(struct kvm *kvm);
1150 void kvm_mmu_invalidate_mmio_sptes(struct kvm *kvm, struct kvm_memslots *slots);
1151 unsigned int kvm_mmu_calculate_mmu_pages(struct kvm *kvm);
1152 void kvm_mmu_change_mmu_pages(struct kvm *kvm, unsigned int kvm_nr_mmu_pages);
1154 int load_pdptrs(struct kvm_vcpu *vcpu, struct kvm_mmu *mmu, unsigned long cr3);
1155 bool pdptrs_changed(struct kvm_vcpu *vcpu);
1157 int emulator_write_phys(struct kvm_vcpu *vcpu, gpa_t gpa,
1158 const void *val, int bytes);
1160 struct kvm_irq_mask_notifier {
1161 void (*func)(struct kvm_irq_mask_notifier *kimn, bool masked);
1163 struct hlist_node link;
1166 void kvm_register_irq_mask_notifier(struct kvm *kvm, int irq,
1167 struct kvm_irq_mask_notifier *kimn);
1168 void kvm_unregister_irq_mask_notifier(struct kvm *kvm, int irq,
1169 struct kvm_irq_mask_notifier *kimn);
1170 void kvm_fire_mask_notifiers(struct kvm *kvm, unsigned irqchip, unsigned pin,
1173 extern bool tdp_enabled;
1175 u64 vcpu_tsc_khz(struct kvm_vcpu *vcpu);
1177 /* control of guest tsc rate supported? */
1178 extern bool kvm_has_tsc_control;
1179 /* maximum supported tsc_khz for guests */
1180 extern u32 kvm_max_guest_tsc_khz;
1181 /* number of bits of the fractional part of the TSC scaling ratio */
1182 extern u8 kvm_tsc_scaling_ratio_frac_bits;
1183 /* maximum allowed value of TSC scaling ratio */
1184 extern u64 kvm_max_tsc_scaling_ratio;
1185 /* 1ull << kvm_tsc_scaling_ratio_frac_bits */
1186 extern u64 kvm_default_tsc_scaling_ratio;
1188 extern u64 kvm_mce_cap_supported;
1190 enum emulation_result {
1191 EMULATE_DONE, /* no further processing */
1192 EMULATE_USER_EXIT, /* kvm_run ready for userspace exit */
1193 EMULATE_FAIL, /* can't emulate this instruction */
1196 #define EMULTYPE_NO_DECODE (1 << 0)
1197 #define EMULTYPE_TRAP_UD (1 << 1)
1198 #define EMULTYPE_SKIP (1 << 2)
1199 #define EMULTYPE_RETRY (1 << 3)
1200 #define EMULTYPE_NO_REEXECUTE (1 << 4)
1201 #define EMULTYPE_NO_UD_ON_FAIL (1 << 5)
1202 #define EMULTYPE_VMWARE (1 << 6)
1203 int x86_emulate_instruction(struct kvm_vcpu *vcpu, unsigned long cr2,
1204 int emulation_type, void *insn, int insn_len);
1206 static inline int emulate_instruction(struct kvm_vcpu *vcpu,
1209 return x86_emulate_instruction(vcpu, 0,
1210 emulation_type | EMULTYPE_NO_REEXECUTE, NULL, 0);
1213 void kvm_enable_efer_bits(u64);
1214 bool kvm_valid_efer(struct kvm_vcpu *vcpu, u64 efer);
1215 int kvm_get_msr(struct kvm_vcpu *vcpu, struct msr_data *msr);
1216 int kvm_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr);
1218 struct x86_emulate_ctxt;
1220 int kvm_fast_pio(struct kvm_vcpu *vcpu, int size, unsigned short port, int in);
1221 int kvm_emulate_cpuid(struct kvm_vcpu *vcpu);
1222 int kvm_emulate_halt(struct kvm_vcpu *vcpu);
1223 int kvm_vcpu_halt(struct kvm_vcpu *vcpu);
1224 int kvm_emulate_wbinvd(struct kvm_vcpu *vcpu);
1226 void kvm_get_segment(struct kvm_vcpu *vcpu, struct kvm_segment *var, int seg);
1227 int kvm_load_segment_descriptor(struct kvm_vcpu *vcpu, u16 selector, int seg);
1228 void kvm_vcpu_deliver_sipi_vector(struct kvm_vcpu *vcpu, u8 vector);
1230 int kvm_task_switch(struct kvm_vcpu *vcpu, u16 tss_selector, int idt_index,
1231 int reason, bool has_error_code, u32 error_code);
1233 int kvm_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0);
1234 int kvm_set_cr3(struct kvm_vcpu *vcpu, unsigned long cr3);
1235 int kvm_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4);
1236 int kvm_set_cr8(struct kvm_vcpu *vcpu, unsigned long cr8);
1237 int kvm_set_dr(struct kvm_vcpu *vcpu, int dr, unsigned long val);
1238 int kvm_get_dr(struct kvm_vcpu *vcpu, int dr, unsigned long *val);
1239 unsigned long kvm_get_cr8(struct kvm_vcpu *vcpu);
1240 void kvm_lmsw(struct kvm_vcpu *vcpu, unsigned long msw);
1241 void kvm_get_cs_db_l_bits(struct kvm_vcpu *vcpu, int *db, int *l);
1242 int kvm_set_xcr(struct kvm_vcpu *vcpu, u32 index, u64 xcr);
1244 int kvm_get_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr);
1245 int kvm_set_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr);
1247 unsigned long kvm_get_rflags(struct kvm_vcpu *vcpu);
1248 void kvm_set_rflags(struct kvm_vcpu *vcpu, unsigned long rflags);
1249 bool kvm_rdpmc(struct kvm_vcpu *vcpu);
1251 void kvm_queue_exception(struct kvm_vcpu *vcpu, unsigned nr);
1252 void kvm_queue_exception_e(struct kvm_vcpu *vcpu, unsigned nr, u32 error_code);
1253 void kvm_requeue_exception(struct kvm_vcpu *vcpu, unsigned nr);
1254 void kvm_requeue_exception_e(struct kvm_vcpu *vcpu, unsigned nr, u32 error_code);
1255 void kvm_inject_page_fault(struct kvm_vcpu *vcpu, struct x86_exception *fault);
1256 int kvm_read_guest_page_mmu(struct kvm_vcpu *vcpu, struct kvm_mmu *mmu,
1257 gfn_t gfn, void *data, int offset, int len,
1259 bool kvm_require_cpl(struct kvm_vcpu *vcpu, int required_cpl);
1260 bool kvm_require_dr(struct kvm_vcpu *vcpu, int dr);
1262 static inline int __kvm_irq_line_state(unsigned long *irq_state,
1263 int irq_source_id, int level)
1265 /* Logical OR for level trig interrupt */
1267 __set_bit(irq_source_id, irq_state);
1269 __clear_bit(irq_source_id, irq_state);
1271 return !!(*irq_state);
1274 int kvm_pic_set_irq(struct kvm_pic *pic, int irq, int irq_source_id, int level);
1275 void kvm_pic_clear_all(struct kvm_pic *pic, int irq_source_id);
1277 void kvm_inject_nmi(struct kvm_vcpu *vcpu);
1279 int kvm_mmu_unprotect_page(struct kvm *kvm, gfn_t gfn);
1280 int kvm_mmu_unprotect_page_virt(struct kvm_vcpu *vcpu, gva_t gva);
1281 void __kvm_mmu_free_some_pages(struct kvm_vcpu *vcpu);
1282 int kvm_mmu_load(struct kvm_vcpu *vcpu);
1283 void kvm_mmu_unload(struct kvm_vcpu *vcpu);
1284 void kvm_mmu_sync_roots(struct kvm_vcpu *vcpu);
1285 void kvm_mmu_free_roots(struct kvm_vcpu *vcpu);
1286 gpa_t translate_nested_gpa(struct kvm_vcpu *vcpu, gpa_t gpa, u32 access,
1287 struct x86_exception *exception);
1288 gpa_t kvm_mmu_gva_to_gpa_read(struct kvm_vcpu *vcpu, gva_t gva,
1289 struct x86_exception *exception);
1290 gpa_t kvm_mmu_gva_to_gpa_fetch(struct kvm_vcpu *vcpu, gva_t gva,
1291 struct x86_exception *exception);
1292 gpa_t kvm_mmu_gva_to_gpa_write(struct kvm_vcpu *vcpu, gva_t gva,
1293 struct x86_exception *exception);
1294 gpa_t kvm_mmu_gva_to_gpa_system(struct kvm_vcpu *vcpu, gva_t gva,
1295 struct x86_exception *exception);
1297 void kvm_vcpu_deactivate_apicv(struct kvm_vcpu *vcpu);
1299 int kvm_emulate_hypercall(struct kvm_vcpu *vcpu);
1301 int kvm_mmu_page_fault(struct kvm_vcpu *vcpu, gva_t gva, u64 error_code,
1302 void *insn, int insn_len);
1303 void kvm_mmu_invlpg(struct kvm_vcpu *vcpu, gva_t gva);
1304 void kvm_mmu_new_cr3(struct kvm_vcpu *vcpu);
1306 void kvm_enable_tdp(void);
1307 void kvm_disable_tdp(void);
1309 static inline gpa_t translate_gpa(struct kvm_vcpu *vcpu, gpa_t gpa, u32 access,
1310 struct x86_exception *exception)
1315 static inline struct kvm_mmu_page *page_header(hpa_t shadow_page)
1317 struct page *page = pfn_to_page(shadow_page >> PAGE_SHIFT);
1319 return (struct kvm_mmu_page *)page_private(page);
1322 static inline u16 kvm_read_ldt(void)
1325 asm("sldt %0" : "=g"(ldt));
1329 static inline void kvm_load_ldt(u16 sel)
1331 asm("lldt %0" : : "rm"(sel));
1334 #ifdef CONFIG_X86_64
1335 static inline unsigned long read_msr(unsigned long msr)
1344 static inline u32 get_rdx_init_val(void)
1346 return 0x600; /* P6 family */
1349 static inline void kvm_inject_gp(struct kvm_vcpu *vcpu, u32 error_code)
1351 kvm_queue_exception_e(vcpu, GP_VECTOR, error_code);
1354 #define TSS_IOPB_BASE_OFFSET 0x66
1355 #define TSS_BASE_SIZE 0x68
1356 #define TSS_IOPB_SIZE (65536 / 8)
1357 #define TSS_REDIRECTION_SIZE (256 / 8)
1358 #define RMODE_TSS_SIZE \
1359 (TSS_BASE_SIZE + TSS_REDIRECTION_SIZE + TSS_IOPB_SIZE + 1)
1362 TASK_SWITCH_CALL = 0,
1363 TASK_SWITCH_IRET = 1,
1364 TASK_SWITCH_JMP = 2,
1365 TASK_SWITCH_GATE = 3,
1368 #define HF_GIF_MASK (1 << 0)
1369 #define HF_HIF_MASK (1 << 1)
1370 #define HF_VINTR_MASK (1 << 2)
1371 #define HF_NMI_MASK (1 << 3)
1372 #define HF_IRET_MASK (1 << 4)
1373 #define HF_GUEST_MASK (1 << 5) /* VCPU is in guest-mode */
1374 #define HF_SMM_MASK (1 << 6)
1375 #define HF_SMM_INSIDE_NMI_MASK (1 << 7)
1377 #define __KVM_VCPU_MULTIPLE_ADDRESS_SPACE
1378 #define KVM_ADDRESS_SPACE_NUM 2
1380 #define kvm_arch_vcpu_memslots_id(vcpu) ((vcpu)->arch.hflags & HF_SMM_MASK ? 1 : 0)
1381 #define kvm_memslots_for_spte_role(kvm, role) __kvm_memslots(kvm, (role).smm)
1384 * Hardware virtualization extension instructions may fault if a
1385 * reboot turns off virtualization while processes are running.
1386 * Trap the fault and ignore the instruction if that happens.
1388 asmlinkage void kvm_spurious_fault(void);
1390 #define ____kvm_handle_fault_on_reboot(insn, cleanup_insn) \
1391 "666: " insn "\n\t" \
1393 ".pushsection .fixup, \"ax\" \n" \
1395 cleanup_insn "\n\t" \
1396 "cmpb $0, kvm_rebooting \n\t" \
1398 __ASM_SIZE(push) " $666b \n\t" \
1399 "call kvm_spurious_fault \n\t" \
1400 ".popsection \n\t" \
1401 _ASM_EXTABLE(666b, 667b)
1403 #define __kvm_handle_fault_on_reboot(insn) \
1404 ____kvm_handle_fault_on_reboot(insn, "")
1406 #define KVM_ARCH_WANT_MMU_NOTIFIER
1407 int kvm_unmap_hva(struct kvm *kvm, unsigned long hva);
1408 int kvm_unmap_hva_range(struct kvm *kvm, unsigned long start, unsigned long end);
1409 int kvm_age_hva(struct kvm *kvm, unsigned long start, unsigned long end);
1410 int kvm_test_age_hva(struct kvm *kvm, unsigned long hva);
1411 void kvm_set_spte_hva(struct kvm *kvm, unsigned long hva, pte_t pte);
1412 int kvm_cpu_has_injectable_intr(struct kvm_vcpu *v);
1413 int kvm_cpu_has_interrupt(struct kvm_vcpu *vcpu);
1414 int kvm_arch_interrupt_allowed(struct kvm_vcpu *vcpu);
1415 int kvm_cpu_get_interrupt(struct kvm_vcpu *v);
1416 void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event);
1417 void kvm_vcpu_reload_apic_access_page(struct kvm_vcpu *vcpu);
1419 void kvm_define_shared_msr(unsigned index, u32 msr);
1420 int kvm_set_shared_msr(unsigned index, u64 val, u64 mask);
1422 u64 kvm_scale_tsc(struct kvm_vcpu *vcpu, u64 tsc);
1423 u64 kvm_read_l1_tsc(struct kvm_vcpu *vcpu, u64 host_tsc);
1425 unsigned long kvm_get_linear_rip(struct kvm_vcpu *vcpu);
1426 bool kvm_is_linear_rip(struct kvm_vcpu *vcpu, unsigned long linear_rip);
1428 void kvm_make_mclock_inprogress_request(struct kvm *kvm);
1429 void kvm_make_scan_ioapic_request(struct kvm *kvm);
1431 void kvm_arch_async_page_not_present(struct kvm_vcpu *vcpu,
1432 struct kvm_async_pf *work);
1433 void kvm_arch_async_page_present(struct kvm_vcpu *vcpu,
1434 struct kvm_async_pf *work);
1435 void kvm_arch_async_page_ready(struct kvm_vcpu *vcpu,
1436 struct kvm_async_pf *work);
1437 bool kvm_arch_can_inject_async_page_present(struct kvm_vcpu *vcpu);
1438 extern bool kvm_find_async_pf_gfn(struct kvm_vcpu *vcpu, gfn_t gfn);
1440 int kvm_skip_emulated_instruction(struct kvm_vcpu *vcpu);
1441 int kvm_complete_insn_gp(struct kvm_vcpu *vcpu, int err);
1443 int kvm_is_in_guest(void);
1445 int __x86_set_memory_region(struct kvm *kvm, int id, gpa_t gpa, u32 size);
1446 int x86_set_memory_region(struct kvm *kvm, int id, gpa_t gpa, u32 size);
1447 bool kvm_vcpu_is_reset_bsp(struct kvm_vcpu *vcpu);
1448 bool kvm_vcpu_is_bsp(struct kvm_vcpu *vcpu);
1450 bool kvm_intr_is_single_vcpu(struct kvm *kvm, struct kvm_lapic_irq *irq,
1451 struct kvm_vcpu **dest_vcpu);
1453 void kvm_set_msi_irq(struct kvm *kvm, struct kvm_kernel_irq_routing_entry *e,
1454 struct kvm_lapic_irq *irq);
1456 static inline void kvm_arch_vcpu_blocking(struct kvm_vcpu *vcpu)
1458 if (kvm_x86_ops->vcpu_blocking)
1459 kvm_x86_ops->vcpu_blocking(vcpu);
1462 static inline void kvm_arch_vcpu_unblocking(struct kvm_vcpu *vcpu)
1464 if (kvm_x86_ops->vcpu_unblocking)
1465 kvm_x86_ops->vcpu_unblocking(vcpu);
1468 static inline void kvm_arch_vcpu_block_finish(struct kvm_vcpu *vcpu) {}
1470 static inline int kvm_cpu_get_apicid(int mps_cpu)
1472 #ifdef CONFIG_X86_LOCAL_APIC
1473 return default_cpu_present_to_apicid(mps_cpu);
1480 #define put_smstate(type, buf, offset, val) \
1481 *(type *)((buf) + (offset) - 0x7e00) = val
1483 #endif /* _ASM_X86_KVM_HOST_H */