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>
28 #include <asm/pvclock-abi.h>
31 #include <asm/msr-index.h>
34 #define KVM_MAX_VCPUS 255
35 #define KVM_SOFT_MAX_VCPUS 160
36 #define KVM_USER_MEM_SLOTS 509
37 /* memory slots that are not exposed to userspace */
38 #define KVM_PRIVATE_MEM_SLOTS 3
39 #define KVM_MEM_SLOTS_NUM (KVM_USER_MEM_SLOTS + KVM_PRIVATE_MEM_SLOTS)
41 #define KVM_MMIO_SIZE 16
43 #define KVM_PIO_PAGE_OFFSET 1
44 #define KVM_COALESCED_MMIO_PAGE_OFFSET 2
46 #define KVM_IRQCHIP_NUM_PINS KVM_IOAPIC_NUM_PINS
48 #define CR0_RESERVED_BITS \
49 (~(unsigned long)(X86_CR0_PE | X86_CR0_MP | X86_CR0_EM | X86_CR0_TS \
50 | X86_CR0_ET | X86_CR0_NE | X86_CR0_WP | X86_CR0_AM \
51 | X86_CR0_NW | X86_CR0_CD | X86_CR0_PG))
53 #define CR3_L_MODE_RESERVED_BITS 0xFFFFFF0000000000ULL
54 #define CR3_PCID_INVD BIT_64(63)
55 #define CR4_RESERVED_BITS \
56 (~(unsigned long)(X86_CR4_VME | X86_CR4_PVI | X86_CR4_TSD | X86_CR4_DE\
57 | X86_CR4_PSE | X86_CR4_PAE | X86_CR4_MCE \
58 | X86_CR4_PGE | X86_CR4_PCE | X86_CR4_OSFXSR | X86_CR4_PCIDE \
59 | X86_CR4_OSXSAVE | X86_CR4_SMEP | X86_CR4_FSGSBASE \
60 | X86_CR4_OSXMMEXCPT | X86_CR4_VMXE | X86_CR4_SMAP))
62 #define CR8_RESERVED_BITS (~(unsigned long)X86_CR8_TPR)
66 #define INVALID_PAGE (~(hpa_t)0)
67 #define VALID_PAGE(x) ((x) != INVALID_PAGE)
69 #define UNMAPPED_GVA (~(gpa_t)0)
71 /* KVM Hugepage definitions for x86 */
72 #define KVM_NR_PAGE_SIZES 3
73 #define KVM_HPAGE_GFN_SHIFT(x) (((x) - 1) * 9)
74 #define KVM_HPAGE_SHIFT(x) (PAGE_SHIFT + KVM_HPAGE_GFN_SHIFT(x))
75 #define KVM_HPAGE_SIZE(x) (1UL << KVM_HPAGE_SHIFT(x))
76 #define KVM_HPAGE_MASK(x) (~(KVM_HPAGE_SIZE(x) - 1))
77 #define KVM_PAGES_PER_HPAGE(x) (KVM_HPAGE_SIZE(x) / PAGE_SIZE)
79 static inline gfn_t
gfn_to_index(gfn_t gfn
, gfn_t base_gfn
, int level
)
81 /* KVM_HPAGE_GFN_SHIFT(PT_PAGE_TABLE_LEVEL) must be 0. */
82 return (gfn
>> KVM_HPAGE_GFN_SHIFT(level
)) -
83 (base_gfn
>> KVM_HPAGE_GFN_SHIFT(level
));
86 #define SELECTOR_TI_MASK (1 << 2)
87 #define SELECTOR_RPL_MASK 0x03
91 #define KVM_PERMILLE_MMU_PAGES 20
92 #define KVM_MIN_ALLOC_MMU_PAGES 64
93 #define KVM_MMU_HASH_SHIFT 10
94 #define KVM_NUM_MMU_PAGES (1 << KVM_MMU_HASH_SHIFT)
95 #define KVM_MIN_FREE_MMU_PAGES 5
96 #define KVM_REFILL_PAGES 25
97 #define KVM_MAX_CPUID_ENTRIES 80
98 #define KVM_NR_FIXED_MTRR_REGION 88
99 #define KVM_NR_VAR_MTRR 8
101 #define ASYNC_PF_PER_VCPU 64
127 VCPU_EXREG_PDPTR
= NR_VCPU_REGS
,
144 #include <asm/kvm_emulate.h>
146 #define KVM_NR_MEM_OBJS 40
148 #define KVM_NR_DB_REGS 4
150 #define DR6_BD (1 << 13)
151 #define DR6_BS (1 << 14)
152 #define DR6_RTM (1 << 16)
153 #define DR6_FIXED_1 0xfffe0ff0
154 #define DR6_INIT 0xffff0ff0
155 #define DR6_VOLATILE 0x0001e00f
157 #define DR7_BP_EN_MASK 0x000000ff
158 #define DR7_GE (1 << 9)
159 #define DR7_GD (1 << 13)
160 #define DR7_FIXED_1 0x00000400
161 #define DR7_VOLATILE 0xffff2bff
163 #define PFERR_PRESENT_BIT 0
164 #define PFERR_WRITE_BIT 1
165 #define PFERR_USER_BIT 2
166 #define PFERR_RSVD_BIT 3
167 #define PFERR_FETCH_BIT 4
169 #define PFERR_PRESENT_MASK (1U << PFERR_PRESENT_BIT)
170 #define PFERR_WRITE_MASK (1U << PFERR_WRITE_BIT)
171 #define PFERR_USER_MASK (1U << PFERR_USER_BIT)
172 #define PFERR_RSVD_MASK (1U << PFERR_RSVD_BIT)
173 #define PFERR_FETCH_MASK (1U << PFERR_FETCH_BIT)
175 /* apic attention bits */
176 #define KVM_APIC_CHECK_VAPIC 0
178 * The following bit is set with PV-EOI, unset on EOI.
179 * We detect PV-EOI changes by guest by comparing
180 * this bit with PV-EOI in guest memory.
181 * See the implementation in apic_update_pv_eoi.
183 #define KVM_APIC_PV_EOI_PENDING 1
186 * We don't want allocation failures within the mmu code, so we preallocate
187 * enough memory for a single page fault in a cache.
189 struct kvm_mmu_memory_cache
{
191 void *objects
[KVM_NR_MEM_OBJS
];
195 * kvm_mmu_page_role, below, is defined as:
197 * bits 0:3 - total guest paging levels (2-4, or zero for real mode)
198 * bits 4:7 - page table level for this shadow (1-4)
199 * bits 8:9 - page table quadrant for 2-level guests
200 * bit 16 - direct mapping of virtual to physical mapping at gfn
201 * used for real mode and two-dimensional paging
202 * bits 17:19 - common access permissions for all ptes in this shadow page
204 union kvm_mmu_page_role
{
210 unsigned pad_for_nice_hex_output
:6;
216 unsigned smep_andnot_wp
:1;
220 struct kvm_mmu_page
{
221 struct list_head link
;
222 struct hlist_node hash_link
;
225 * The following two entries are used to key the shadow page in the
229 union kvm_mmu_page_role role
;
232 /* hold the gfn of each spte inside spt */
235 int root_count
; /* Currently serving as active root */
236 unsigned int unsync_children
;
237 unsigned long parent_ptes
; /* Reverse mapping for parent_pte */
239 /* The page is obsolete if mmu_valid_gen != kvm->arch.mmu_valid_gen. */
240 unsigned long mmu_valid_gen
;
242 DECLARE_BITMAP(unsync_child_bitmap
, 512);
246 * Used out of the mmu-lock to avoid reading spte values while an
247 * update is in progress; see the comments in __get_spte_lockless().
249 int clear_spte_count
;
252 /* Number of writes since the last time traversal visited this page. */
253 int write_flooding_count
;
256 struct kvm_pio_request
{
264 * x86 supports 3 paging modes (4-level 64-bit, 3-level 64-bit, and 2-level
265 * 32-bit). The kvm_mmu structure abstracts the details of the current mmu
269 void (*set_cr3
)(struct kvm_vcpu
*vcpu
, unsigned long root
);
270 unsigned long (*get_cr3
)(struct kvm_vcpu
*vcpu
);
271 u64 (*get_pdptr
)(struct kvm_vcpu
*vcpu
, int index
);
272 int (*page_fault
)(struct kvm_vcpu
*vcpu
, gva_t gva
, u32 err
,
274 void (*inject_page_fault
)(struct kvm_vcpu
*vcpu
,
275 struct x86_exception
*fault
);
276 gpa_t (*gva_to_gpa
)(struct kvm_vcpu
*vcpu
, gva_t gva
, u32 access
,
277 struct x86_exception
*exception
);
278 gpa_t (*translate_gpa
)(struct kvm_vcpu
*vcpu
, gpa_t gpa
, u32 access
,
279 struct x86_exception
*exception
);
280 int (*sync_page
)(struct kvm_vcpu
*vcpu
,
281 struct kvm_mmu_page
*sp
);
282 void (*invlpg
)(struct kvm_vcpu
*vcpu
, gva_t gva
);
283 void (*update_pte
)(struct kvm_vcpu
*vcpu
, struct kvm_mmu_page
*sp
,
284 u64
*spte
, const void *pte
);
287 int shadow_root_level
;
288 union kvm_mmu_page_role base_role
;
292 * Bitmap; bit set = permission fault
293 * Byte index: page fault error code [4:1]
294 * Bit index: pte permissions in ACC_* format
300 u64 rsvd_bits_mask
[2][4];
304 * Bitmap: bit set = last pte in walk
305 * index[0:1]: level (zero-based)
312 u64 pdptrs
[4]; /* pae */
325 struct perf_event
*perf_event
;
326 struct kvm_vcpu
*vcpu
;
330 unsigned nr_arch_gp_counters
;
331 unsigned nr_arch_fixed_counters
;
332 unsigned available_event_types
;
337 u64 counter_bitmask
[2];
338 u64 global_ctrl_mask
;
341 struct kvm_pmc gp_counters
[INTEL_PMC_MAX_GENERIC
];
342 struct kvm_pmc fixed_counters
[INTEL_PMC_MAX_FIXED
];
343 struct irq_work irq_work
;
348 KVM_DEBUGREG_BP_ENABLED
= 1,
349 KVM_DEBUGREG_WONT_EXIT
= 2,
352 struct kvm_vcpu_arch
{
354 * rip and regs accesses must go through
355 * kvm_{register,rip}_{read,write} functions.
357 unsigned long regs
[NR_VCPU_REGS
];
362 unsigned long cr0_guest_owned_bits
;
366 unsigned long cr4_guest_owned_bits
;
371 struct kvm_lapic
*apic
; /* kernel irqchip context */
372 unsigned long apic_attention
;
373 int32_t apic_arb_prio
;
375 u64 ia32_misc_enable_msr
;
376 bool tpr_access_reporting
;
380 * Paging state of the vcpu
382 * If the vcpu runs in guest mode with two level paging this still saves
383 * the paging mode of the l1 guest. This context is always used to
389 * Paging state of an L2 guest (used for nested npt)
391 * This context will save all necessary information to walk page tables
392 * of the an L2 guest. This context is only initialized for page table
393 * walking and not for faulting since we never handle l2 page faults on
396 struct kvm_mmu nested_mmu
;
399 * Pointer to the mmu context currently used for
400 * gva_to_gpa translations.
402 struct kvm_mmu
*walk_mmu
;
404 struct kvm_mmu_memory_cache mmu_pte_list_desc_cache
;
405 struct kvm_mmu_memory_cache mmu_page_cache
;
406 struct kvm_mmu_memory_cache mmu_page_header_cache
;
408 struct fpu guest_fpu
;
410 u64 guest_supported_xcr0
;
411 u32 guest_xstate_size
;
413 struct kvm_pio_request pio
;
416 u8 event_exit_inst_len
;
418 struct kvm_queued_exception
{
426 struct kvm_queued_interrupt
{
432 int halt_request
; /* real mode on Intel only */
435 struct kvm_cpuid_entry2 cpuid_entries
[KVM_MAX_CPUID_ENTRIES
];
436 /* emulate context */
438 struct x86_emulate_ctxt emulate_ctxt
;
439 bool emulate_regs_need_sync_to_vcpu
;
440 bool emulate_regs_need_sync_from_vcpu
;
441 int (*complete_userspace_io
)(struct kvm_vcpu
*vcpu
);
444 struct pvclock_vcpu_time_info hv_clock
;
445 unsigned int hw_tsc_khz
;
446 struct gfn_to_hva_cache pv_time
;
447 bool pv_time_enabled
;
448 /* set guest stopped flag in pvclock flags field */
449 bool pvclock_set_guest_stopped_request
;
455 struct gfn_to_hva_cache stime
;
456 struct kvm_steal_time steal
;
461 u64 tsc_offset_adjustment
;
464 u64 this_tsc_generation
;
466 bool tsc_always_catchup
;
467 s8 virtual_tsc_shift
;
468 u32 virtual_tsc_mult
;
470 s64 ia32_tsc_adjust_msr
;
472 atomic_t nmi_queued
; /* unprocessed asynchronous NMIs */
473 unsigned nmi_pending
; /* NMI queued after currently running handler */
474 bool nmi_injected
; /* Trying to inject an NMI this entry */
476 struct mtrr_state_type mtrr_state
;
479 unsigned switch_db_regs
;
480 unsigned long db
[KVM_NR_DB_REGS
];
483 unsigned long eff_db
[KVM_NR_DB_REGS
];
484 unsigned long guest_debug_dr7
;
491 /* Cache MMIO info */
499 /* used for guest single stepping over the given code position */
500 unsigned long singlestep_rip
;
502 /* fields used by HYPER-V emulation */
505 cpumask_var_t wbinvd_dirty_mask
;
507 unsigned long last_retry_eip
;
508 unsigned long last_retry_addr
;
512 gfn_t gfns
[roundup_pow_of_two(ASYNC_PF_PER_VCPU
)];
513 struct gfn_to_hva_cache data
;
519 /* OSVW MSRs (AMD only) */
527 struct gfn_to_hva_cache data
;
531 * Indicate whether the access faults on its page table in guest
532 * which is set when fix page fault and used to detect unhandeable
535 bool write_fault_to_shadow_pgtable
;
537 /* set at EPT violation at this point */
538 unsigned long exit_qualification
;
540 /* pv related host specific info */
546 struct kvm_lpage_info
{
550 struct kvm_arch_memory_slot
{
551 unsigned long *rmap
[KVM_NR_PAGE_SIZES
];
552 struct kvm_lpage_info
*lpage_info
[KVM_NR_PAGE_SIZES
- 1];
555 struct kvm_apic_map
{
558 /* fields bellow are used to decode ldr values in different modes */
559 u32 cid_shift
, cid_mask
, lid_mask
, broadcast
;
560 struct kvm_lapic
*phys_map
[256];
561 /* first index is cluster id second is cpu id in a cluster */
562 struct kvm_lapic
*logical_map
[16][16];
566 unsigned int n_used_mmu_pages
;
567 unsigned int n_requested_mmu_pages
;
568 unsigned int n_max_mmu_pages
;
569 unsigned int indirect_shadow_pages
;
570 unsigned long mmu_valid_gen
;
571 struct hlist_head mmu_page_hash
[KVM_NUM_MMU_PAGES
];
573 * Hash table of struct kvm_mmu_page.
575 struct list_head active_mmu_pages
;
576 struct list_head zapped_obsolete_pages
;
578 struct list_head assigned_dev_head
;
579 struct iommu_domain
*iommu_domain
;
580 bool iommu_noncoherent
;
581 #define __KVM_HAVE_ARCH_NONCOHERENT_DMA
582 atomic_t noncoherent_dma_count
;
583 struct kvm_pic
*vpic
;
584 struct kvm_ioapic
*vioapic
;
585 struct kvm_pit
*vpit
;
586 int vapics_in_nmi_mode
;
587 struct mutex apic_map_lock
;
588 struct kvm_apic_map
*apic_map
;
590 unsigned int tss_addr
;
591 bool apic_access_page_done
;
595 bool ept_identity_pagetable_done
;
596 gpa_t ept_identity_map_addr
;
598 unsigned long irq_sources_bitmap
;
600 raw_spinlock_t tsc_write_lock
;
607 u64 cur_tsc_generation
;
608 int nr_vcpus_matched_tsc
;
610 spinlock_t pvclock_gtod_sync_lock
;
611 bool use_master_clock
;
612 u64 master_kernel_ns
;
613 cycle_t master_cycle_now
;
614 struct delayed_work kvmclock_update_work
;
615 struct delayed_work kvmclock_sync_work
;
617 struct kvm_xen_hvm_config xen_hvm_config
;
619 /* reads protected by irq_srcu, writes by irq_lock */
620 struct hlist_head mask_notifier_list
;
622 /* fields used by HYPER-V emulation */
627 #ifdef CONFIG_KVM_MMU_AUDIT
631 bool boot_vcpu_runs_old_kvmclock
;
635 u32 mmu_shadow_zapped
;
643 u32 remote_tlb_flush
;
647 struct kvm_vcpu_stat
{
657 u32 irq_window_exits
;
658 u32 nmi_window_exits
;
661 u32 request_irq_exits
;
663 u32 host_state_reload
;
667 u32 insn_emulation_fail
;
673 struct x86_instruction_info
;
681 struct kvm_lapic_irq
{
692 int (*cpu_has_kvm_support
)(void); /* __init */
693 int (*disabled_by_bios
)(void); /* __init */
694 int (*hardware_enable
)(void);
695 void (*hardware_disable
)(void);
696 void (*check_processor_compatibility
)(void *rtn
);
697 int (*hardware_setup
)(void); /* __init */
698 void (*hardware_unsetup
)(void); /* __exit */
699 bool (*cpu_has_accelerated_tpr
)(void);
700 void (*cpuid_update
)(struct kvm_vcpu
*vcpu
);
702 /* Create, but do not attach this VCPU */
703 struct kvm_vcpu
*(*vcpu_create
)(struct kvm
*kvm
, unsigned id
);
704 void (*vcpu_free
)(struct kvm_vcpu
*vcpu
);
705 void (*vcpu_reset
)(struct kvm_vcpu
*vcpu
);
707 void (*prepare_guest_switch
)(struct kvm_vcpu
*vcpu
);
708 void (*vcpu_load
)(struct kvm_vcpu
*vcpu
, int cpu
);
709 void (*vcpu_put
)(struct kvm_vcpu
*vcpu
);
711 void (*update_db_bp_intercept
)(struct kvm_vcpu
*vcpu
);
712 int (*get_msr
)(struct kvm_vcpu
*vcpu
, u32 msr_index
, u64
*pdata
);
713 int (*set_msr
)(struct kvm_vcpu
*vcpu
, struct msr_data
*msr
);
714 u64 (*get_segment_base
)(struct kvm_vcpu
*vcpu
, int seg
);
715 void (*get_segment
)(struct kvm_vcpu
*vcpu
,
716 struct kvm_segment
*var
, int seg
);
717 int (*get_cpl
)(struct kvm_vcpu
*vcpu
);
718 void (*set_segment
)(struct kvm_vcpu
*vcpu
,
719 struct kvm_segment
*var
, int seg
);
720 void (*get_cs_db_l_bits
)(struct kvm_vcpu
*vcpu
, int *db
, int *l
);
721 void (*decache_cr0_guest_bits
)(struct kvm_vcpu
*vcpu
);
722 void (*decache_cr3
)(struct kvm_vcpu
*vcpu
);
723 void (*decache_cr4_guest_bits
)(struct kvm_vcpu
*vcpu
);
724 void (*set_cr0
)(struct kvm_vcpu
*vcpu
, unsigned long cr0
);
725 void (*set_cr3
)(struct kvm_vcpu
*vcpu
, unsigned long cr3
);
726 int (*set_cr4
)(struct kvm_vcpu
*vcpu
, unsigned long cr4
);
727 void (*set_efer
)(struct kvm_vcpu
*vcpu
, u64 efer
);
728 void (*get_idt
)(struct kvm_vcpu
*vcpu
, struct desc_ptr
*dt
);
729 void (*set_idt
)(struct kvm_vcpu
*vcpu
, struct desc_ptr
*dt
);
730 void (*get_gdt
)(struct kvm_vcpu
*vcpu
, struct desc_ptr
*dt
);
731 void (*set_gdt
)(struct kvm_vcpu
*vcpu
, struct desc_ptr
*dt
);
732 u64 (*get_dr6
)(struct kvm_vcpu
*vcpu
);
733 void (*set_dr6
)(struct kvm_vcpu
*vcpu
, unsigned long value
);
734 void (*sync_dirty_debug_regs
)(struct kvm_vcpu
*vcpu
);
735 void (*set_dr7
)(struct kvm_vcpu
*vcpu
, unsigned long value
);
736 void (*cache_reg
)(struct kvm_vcpu
*vcpu
, enum kvm_reg reg
);
737 unsigned long (*get_rflags
)(struct kvm_vcpu
*vcpu
);
738 void (*set_rflags
)(struct kvm_vcpu
*vcpu
, unsigned long rflags
);
739 void (*fpu_deactivate
)(struct kvm_vcpu
*vcpu
);
741 void (*tlb_flush
)(struct kvm_vcpu
*vcpu
);
743 void (*run
)(struct kvm_vcpu
*vcpu
);
744 int (*handle_exit
)(struct kvm_vcpu
*vcpu
);
745 void (*skip_emulated_instruction
)(struct kvm_vcpu
*vcpu
);
746 void (*set_interrupt_shadow
)(struct kvm_vcpu
*vcpu
, int mask
);
747 u32 (*get_interrupt_shadow
)(struct kvm_vcpu
*vcpu
);
748 void (*patch_hypercall
)(struct kvm_vcpu
*vcpu
,
749 unsigned char *hypercall_addr
);
750 void (*set_irq
)(struct kvm_vcpu
*vcpu
);
751 void (*set_nmi
)(struct kvm_vcpu
*vcpu
);
752 void (*queue_exception
)(struct kvm_vcpu
*vcpu
, unsigned nr
,
753 bool has_error_code
, u32 error_code
,
755 void (*cancel_injection
)(struct kvm_vcpu
*vcpu
);
756 int (*interrupt_allowed
)(struct kvm_vcpu
*vcpu
);
757 int (*nmi_allowed
)(struct kvm_vcpu
*vcpu
);
758 bool (*get_nmi_mask
)(struct kvm_vcpu
*vcpu
);
759 void (*set_nmi_mask
)(struct kvm_vcpu
*vcpu
, bool masked
);
760 void (*enable_nmi_window
)(struct kvm_vcpu
*vcpu
);
761 void (*enable_irq_window
)(struct kvm_vcpu
*vcpu
);
762 void (*update_cr8_intercept
)(struct kvm_vcpu
*vcpu
, int tpr
, int irr
);
763 int (*vm_has_apicv
)(struct kvm
*kvm
);
764 void (*hwapic_irr_update
)(struct kvm_vcpu
*vcpu
, int max_irr
);
765 void (*hwapic_isr_update
)(struct kvm
*kvm
, int isr
);
766 void (*load_eoi_exitmap
)(struct kvm_vcpu
*vcpu
, u64
*eoi_exit_bitmap
);
767 void (*set_virtual_x2apic_mode
)(struct kvm_vcpu
*vcpu
, bool set
);
768 void (*set_apic_access_page_addr
)(struct kvm_vcpu
*vcpu
, hpa_t hpa
);
769 void (*deliver_posted_interrupt
)(struct kvm_vcpu
*vcpu
, int vector
);
770 bool (*test_posted_interrupt
)(struct kvm_vcpu
*vcpu
, int vector
);
771 void (*sync_pir_to_irr
)(struct kvm_vcpu
*vcpu
);
772 int (*set_tss_addr
)(struct kvm
*kvm
, unsigned int addr
);
773 int (*get_tdp_level
)(void);
774 u64 (*get_mt_mask
)(struct kvm_vcpu
*vcpu
, gfn_t gfn
, bool is_mmio
);
775 int (*get_lpage_level
)(void);
776 bool (*rdtscp_supported
)(void);
777 bool (*invpcid_supported
)(void);
778 void (*adjust_tsc_offset
)(struct kvm_vcpu
*vcpu
, s64 adjustment
, bool host
);
780 void (*set_tdp_cr3
)(struct kvm_vcpu
*vcpu
, unsigned long cr3
);
782 void (*set_supported_cpuid
)(u32 func
, struct kvm_cpuid_entry2
*entry
);
784 bool (*has_wbinvd_exit
)(void);
786 void (*set_tsc_khz
)(struct kvm_vcpu
*vcpu
, u32 user_tsc_khz
, bool scale
);
787 u64 (*read_tsc_offset
)(struct kvm_vcpu
*vcpu
);
788 void (*write_tsc_offset
)(struct kvm_vcpu
*vcpu
, u64 offset
);
790 u64 (*compute_tsc_offset
)(struct kvm_vcpu
*vcpu
, u64 target_tsc
);
791 u64 (*read_l1_tsc
)(struct kvm_vcpu
*vcpu
, u64 host_tsc
);
793 void (*get_exit_info
)(struct kvm_vcpu
*vcpu
, u64
*info1
, u64
*info2
);
795 int (*check_intercept
)(struct kvm_vcpu
*vcpu
,
796 struct x86_instruction_info
*info
,
797 enum x86_intercept_stage stage
);
798 void (*handle_external_intr
)(struct kvm_vcpu
*vcpu
);
799 bool (*mpx_supported
)(void);
800 bool (*xsaves_supported
)(void);
802 int (*check_nested_events
)(struct kvm_vcpu
*vcpu
, bool external_intr
);
804 void (*sched_in
)(struct kvm_vcpu
*kvm
, int cpu
);
807 struct kvm_arch_async_pf
{
814 extern struct kvm_x86_ops
*kvm_x86_ops
;
816 static inline void adjust_tsc_offset_guest(struct kvm_vcpu
*vcpu
,
819 kvm_x86_ops
->adjust_tsc_offset(vcpu
, adjustment
, false);
822 static inline void adjust_tsc_offset_host(struct kvm_vcpu
*vcpu
, s64 adjustment
)
824 kvm_x86_ops
->adjust_tsc_offset(vcpu
, adjustment
, true);
827 int kvm_mmu_module_init(void);
828 void kvm_mmu_module_exit(void);
830 void kvm_mmu_destroy(struct kvm_vcpu
*vcpu
);
831 int kvm_mmu_create(struct kvm_vcpu
*vcpu
);
832 void kvm_mmu_setup(struct kvm_vcpu
*vcpu
);
833 void kvm_mmu_set_mask_ptes(u64 user_mask
, u64 accessed_mask
,
834 u64 dirty_mask
, u64 nx_mask
, u64 x_mask
);
836 void kvm_mmu_reset_context(struct kvm_vcpu
*vcpu
);
837 void kvm_mmu_slot_remove_write_access(struct kvm
*kvm
, int slot
);
838 void kvm_mmu_zap_all(struct kvm
*kvm
);
839 void kvm_mmu_invalidate_mmio_sptes(struct kvm
*kvm
);
840 unsigned int kvm_mmu_calculate_mmu_pages(struct kvm
*kvm
);
841 void kvm_mmu_change_mmu_pages(struct kvm
*kvm
, unsigned int kvm_nr_mmu_pages
);
843 int load_pdptrs(struct kvm_vcpu
*vcpu
, struct kvm_mmu
*mmu
, unsigned long cr3
);
845 int emulator_write_phys(struct kvm_vcpu
*vcpu
, gpa_t gpa
,
846 const void *val
, int bytes
);
847 u8
kvm_get_guest_memory_type(struct kvm_vcpu
*vcpu
, gfn_t gfn
);
849 struct kvm_irq_mask_notifier
{
850 void (*func
)(struct kvm_irq_mask_notifier
*kimn
, bool masked
);
852 struct hlist_node link
;
855 void kvm_register_irq_mask_notifier(struct kvm
*kvm
, int irq
,
856 struct kvm_irq_mask_notifier
*kimn
);
857 void kvm_unregister_irq_mask_notifier(struct kvm
*kvm
, int irq
,
858 struct kvm_irq_mask_notifier
*kimn
);
859 void kvm_fire_mask_notifiers(struct kvm
*kvm
, unsigned irqchip
, unsigned pin
,
862 extern bool tdp_enabled
;
864 u64
vcpu_tsc_khz(struct kvm_vcpu
*vcpu
);
866 /* control of guest tsc rate supported? */
867 extern bool kvm_has_tsc_control
;
868 /* minimum supported tsc_khz for guests */
869 extern u32 kvm_min_guest_tsc_khz
;
870 /* maximum supported tsc_khz for guests */
871 extern u32 kvm_max_guest_tsc_khz
;
873 enum emulation_result
{
874 EMULATE_DONE
, /* no further processing */
875 EMULATE_USER_EXIT
, /* kvm_run ready for userspace exit */
876 EMULATE_FAIL
, /* can't emulate this instruction */
879 #define EMULTYPE_NO_DECODE (1 << 0)
880 #define EMULTYPE_TRAP_UD (1 << 1)
881 #define EMULTYPE_SKIP (1 << 2)
882 #define EMULTYPE_RETRY (1 << 3)
883 #define EMULTYPE_NO_REEXECUTE (1 << 4)
884 int x86_emulate_instruction(struct kvm_vcpu
*vcpu
, unsigned long cr2
,
885 int emulation_type
, void *insn
, int insn_len
);
887 static inline int emulate_instruction(struct kvm_vcpu
*vcpu
,
890 return x86_emulate_instruction(vcpu
, 0, emulation_type
, NULL
, 0);
893 void kvm_enable_efer_bits(u64
);
894 bool kvm_valid_efer(struct kvm_vcpu
*vcpu
, u64 efer
);
895 int kvm_get_msr(struct kvm_vcpu
*vcpu
, u32 msr_index
, u64
*data
);
896 int kvm_set_msr(struct kvm_vcpu
*vcpu
, struct msr_data
*msr
);
898 struct x86_emulate_ctxt
;
900 int kvm_fast_pio_out(struct kvm_vcpu
*vcpu
, int size
, unsigned short port
);
901 void kvm_emulate_cpuid(struct kvm_vcpu
*vcpu
);
902 int kvm_emulate_halt(struct kvm_vcpu
*vcpu
);
903 int kvm_emulate_wbinvd(struct kvm_vcpu
*vcpu
);
905 void kvm_get_segment(struct kvm_vcpu
*vcpu
, struct kvm_segment
*var
, int seg
);
906 int kvm_load_segment_descriptor(struct kvm_vcpu
*vcpu
, u16 selector
, int seg
);
907 void kvm_vcpu_deliver_sipi_vector(struct kvm_vcpu
*vcpu
, u8 vector
);
909 int kvm_task_switch(struct kvm_vcpu
*vcpu
, u16 tss_selector
, int idt_index
,
910 int reason
, bool has_error_code
, u32 error_code
);
912 int kvm_set_cr0(struct kvm_vcpu
*vcpu
, unsigned long cr0
);
913 int kvm_set_cr3(struct kvm_vcpu
*vcpu
, unsigned long cr3
);
914 int kvm_set_cr4(struct kvm_vcpu
*vcpu
, unsigned long cr4
);
915 int kvm_set_cr8(struct kvm_vcpu
*vcpu
, unsigned long cr8
);
916 int kvm_set_dr(struct kvm_vcpu
*vcpu
, int dr
, unsigned long val
);
917 int kvm_get_dr(struct kvm_vcpu
*vcpu
, int dr
, unsigned long *val
);
918 unsigned long kvm_get_cr8(struct kvm_vcpu
*vcpu
);
919 void kvm_lmsw(struct kvm_vcpu
*vcpu
, unsigned long msw
);
920 void kvm_get_cs_db_l_bits(struct kvm_vcpu
*vcpu
, int *db
, int *l
);
921 int kvm_set_xcr(struct kvm_vcpu
*vcpu
, u32 index
, u64 xcr
);
923 int kvm_get_msr_common(struct kvm_vcpu
*vcpu
, u32 msr
, u64
*pdata
);
924 int kvm_set_msr_common(struct kvm_vcpu
*vcpu
, struct msr_data
*msr
);
926 unsigned long kvm_get_rflags(struct kvm_vcpu
*vcpu
);
927 void kvm_set_rflags(struct kvm_vcpu
*vcpu
, unsigned long rflags
);
928 bool kvm_rdpmc(struct kvm_vcpu
*vcpu
);
930 void kvm_queue_exception(struct kvm_vcpu
*vcpu
, unsigned nr
);
931 void kvm_queue_exception_e(struct kvm_vcpu
*vcpu
, unsigned nr
, u32 error_code
);
932 void kvm_requeue_exception(struct kvm_vcpu
*vcpu
, unsigned nr
);
933 void kvm_requeue_exception_e(struct kvm_vcpu
*vcpu
, unsigned nr
, u32 error_code
);
934 void kvm_inject_page_fault(struct kvm_vcpu
*vcpu
, struct x86_exception
*fault
);
935 int kvm_read_guest_page_mmu(struct kvm_vcpu
*vcpu
, struct kvm_mmu
*mmu
,
936 gfn_t gfn
, void *data
, int offset
, int len
,
938 bool kvm_require_cpl(struct kvm_vcpu
*vcpu
, int required_cpl
);
939 bool kvm_require_dr(struct kvm_vcpu
*vcpu
, int dr
);
941 static inline int __kvm_irq_line_state(unsigned long *irq_state
,
942 int irq_source_id
, int level
)
944 /* Logical OR for level trig interrupt */
946 __set_bit(irq_source_id
, irq_state
);
948 __clear_bit(irq_source_id
, irq_state
);
950 return !!(*irq_state
);
953 int kvm_pic_set_irq(struct kvm_pic
*pic
, int irq
, int irq_source_id
, int level
);
954 void kvm_pic_clear_all(struct kvm_pic
*pic
, int irq_source_id
);
956 void kvm_inject_nmi(struct kvm_vcpu
*vcpu
);
958 int fx_init(struct kvm_vcpu
*vcpu
);
960 void kvm_mmu_pte_write(struct kvm_vcpu
*vcpu
, gpa_t gpa
,
961 const u8
*new, int bytes
);
962 int kvm_mmu_unprotect_page(struct kvm
*kvm
, gfn_t gfn
);
963 int kvm_mmu_unprotect_page_virt(struct kvm_vcpu
*vcpu
, gva_t gva
);
964 void __kvm_mmu_free_some_pages(struct kvm_vcpu
*vcpu
);
965 int kvm_mmu_load(struct kvm_vcpu
*vcpu
);
966 void kvm_mmu_unload(struct kvm_vcpu
*vcpu
);
967 void kvm_mmu_sync_roots(struct kvm_vcpu
*vcpu
);
968 gpa_t
translate_nested_gpa(struct kvm_vcpu
*vcpu
, gpa_t gpa
, u32 access
,
969 struct x86_exception
*exception
);
970 gpa_t
kvm_mmu_gva_to_gpa_read(struct kvm_vcpu
*vcpu
, gva_t gva
,
971 struct x86_exception
*exception
);
972 gpa_t
kvm_mmu_gva_to_gpa_fetch(struct kvm_vcpu
*vcpu
, gva_t gva
,
973 struct x86_exception
*exception
);
974 gpa_t
kvm_mmu_gva_to_gpa_write(struct kvm_vcpu
*vcpu
, gva_t gva
,
975 struct x86_exception
*exception
);
976 gpa_t
kvm_mmu_gva_to_gpa_system(struct kvm_vcpu
*vcpu
, gva_t gva
,
977 struct x86_exception
*exception
);
979 int kvm_emulate_hypercall(struct kvm_vcpu
*vcpu
);
981 int kvm_mmu_page_fault(struct kvm_vcpu
*vcpu
, gva_t gva
, u32 error_code
,
982 void *insn
, int insn_len
);
983 void kvm_mmu_invlpg(struct kvm_vcpu
*vcpu
, gva_t gva
);
984 void kvm_mmu_new_cr3(struct kvm_vcpu
*vcpu
);
986 void kvm_enable_tdp(void);
987 void kvm_disable_tdp(void);
989 static inline gpa_t
translate_gpa(struct kvm_vcpu
*vcpu
, gpa_t gpa
, u32 access
,
990 struct x86_exception
*exception
)
995 static inline struct kvm_mmu_page
*page_header(hpa_t shadow_page
)
997 struct page
*page
= pfn_to_page(shadow_page
>> PAGE_SHIFT
);
999 return (struct kvm_mmu_page
*)page_private(page
);
1002 static inline u16
kvm_read_ldt(void)
1005 asm("sldt %0" : "=g"(ldt
));
1009 static inline void kvm_load_ldt(u16 sel
)
1011 asm("lldt %0" : : "rm"(sel
));
1014 #ifdef CONFIG_X86_64
1015 static inline unsigned long read_msr(unsigned long msr
)
1024 static inline u32
get_rdx_init_val(void)
1026 return 0x600; /* P6 family */
1029 static inline void kvm_inject_gp(struct kvm_vcpu
*vcpu
, u32 error_code
)
1031 kvm_queue_exception_e(vcpu
, GP_VECTOR
, error_code
);
1034 static inline u64
get_canonical(u64 la
)
1036 return ((int64_t)la
<< 16) >> 16;
1039 static inline bool is_noncanonical_address(u64 la
)
1041 #ifdef CONFIG_X86_64
1042 return get_canonical(la
) != la
;
1048 #define TSS_IOPB_BASE_OFFSET 0x66
1049 #define TSS_BASE_SIZE 0x68
1050 #define TSS_IOPB_SIZE (65536 / 8)
1051 #define TSS_REDIRECTION_SIZE (256 / 8)
1052 #define RMODE_TSS_SIZE \
1053 (TSS_BASE_SIZE + TSS_REDIRECTION_SIZE + TSS_IOPB_SIZE + 1)
1056 TASK_SWITCH_CALL
= 0,
1057 TASK_SWITCH_IRET
= 1,
1058 TASK_SWITCH_JMP
= 2,
1059 TASK_SWITCH_GATE
= 3,
1062 #define HF_GIF_MASK (1 << 0)
1063 #define HF_HIF_MASK (1 << 1)
1064 #define HF_VINTR_MASK (1 << 2)
1065 #define HF_NMI_MASK (1 << 3)
1066 #define HF_IRET_MASK (1 << 4)
1067 #define HF_GUEST_MASK (1 << 5) /* VCPU is in guest-mode */
1070 * Hardware virtualization extension instructions may fault if a
1071 * reboot turns off virtualization while processes are running.
1072 * Trap the fault and ignore the instruction if that happens.
1074 asmlinkage
void kvm_spurious_fault(void);
1076 #define ____kvm_handle_fault_on_reboot(insn, cleanup_insn) \
1077 "666: " insn "\n\t" \
1079 ".pushsection .fixup, \"ax\" \n" \
1081 cleanup_insn "\n\t" \
1082 "cmpb $0, kvm_rebooting \n\t" \
1084 __ASM_SIZE(push) " $666b \n\t" \
1085 "call kvm_spurious_fault \n\t" \
1086 ".popsection \n\t" \
1087 _ASM_EXTABLE(666b, 667b)
1089 #define __kvm_handle_fault_on_reboot(insn) \
1090 ____kvm_handle_fault_on_reboot(insn, "")
1092 #define KVM_ARCH_WANT_MMU_NOTIFIER
1093 int kvm_unmap_hva(struct kvm
*kvm
, unsigned long hva
);
1094 int kvm_unmap_hva_range(struct kvm
*kvm
, unsigned long start
, unsigned long end
);
1095 int kvm_age_hva(struct kvm
*kvm
, unsigned long start
, unsigned long end
);
1096 int kvm_test_age_hva(struct kvm
*kvm
, unsigned long hva
);
1097 void kvm_set_spte_hva(struct kvm
*kvm
, unsigned long hva
, pte_t pte
);
1098 int cpuid_maxphyaddr(struct kvm_vcpu
*vcpu
);
1099 int kvm_cpu_has_injectable_intr(struct kvm_vcpu
*v
);
1100 int kvm_cpu_has_interrupt(struct kvm_vcpu
*vcpu
);
1101 int kvm_arch_interrupt_allowed(struct kvm_vcpu
*vcpu
);
1102 int kvm_cpu_get_interrupt(struct kvm_vcpu
*v
);
1103 void kvm_vcpu_reset(struct kvm_vcpu
*vcpu
);
1104 void kvm_vcpu_reload_apic_access_page(struct kvm_vcpu
*vcpu
);
1105 void kvm_arch_mmu_notifier_invalidate_page(struct kvm
*kvm
,
1106 unsigned long address
);
1108 void kvm_define_shared_msr(unsigned index
, u32 msr
);
1109 int kvm_set_shared_msr(unsigned index
, u64 val
, u64 mask
);
1111 unsigned long kvm_get_linear_rip(struct kvm_vcpu
*vcpu
);
1112 bool kvm_is_linear_rip(struct kvm_vcpu
*vcpu
, unsigned long linear_rip
);
1114 void kvm_arch_async_page_not_present(struct kvm_vcpu
*vcpu
,
1115 struct kvm_async_pf
*work
);
1116 void kvm_arch_async_page_present(struct kvm_vcpu
*vcpu
,
1117 struct kvm_async_pf
*work
);
1118 void kvm_arch_async_page_ready(struct kvm_vcpu
*vcpu
,
1119 struct kvm_async_pf
*work
);
1120 bool kvm_arch_can_inject_async_page_present(struct kvm_vcpu
*vcpu
);
1121 extern bool kvm_find_async_pf_gfn(struct kvm_vcpu
*vcpu
, gfn_t gfn
);
1123 void kvm_complete_insn_gp(struct kvm_vcpu
*vcpu
, int err
);
1125 int kvm_is_in_guest(void);
1127 void kvm_pmu_init(struct kvm_vcpu
*vcpu
);
1128 void kvm_pmu_destroy(struct kvm_vcpu
*vcpu
);
1129 void kvm_pmu_reset(struct kvm_vcpu
*vcpu
);
1130 void kvm_pmu_cpuid_update(struct kvm_vcpu
*vcpu
);
1131 bool kvm_pmu_msr(struct kvm_vcpu
*vcpu
, u32 msr
);
1132 int kvm_pmu_get_msr(struct kvm_vcpu
*vcpu
, u32 msr
, u64
*data
);
1133 int kvm_pmu_set_msr(struct kvm_vcpu
*vcpu
, struct msr_data
*msr_info
);
1134 int kvm_pmu_check_pmc(struct kvm_vcpu
*vcpu
, unsigned pmc
);
1135 int kvm_pmu_read_pmc(struct kvm_vcpu
*vcpu
, unsigned pmc
, u64
*data
);
1136 void kvm_handle_pmu_event(struct kvm_vcpu
*vcpu
);
1137 void kvm_deliver_pmi(struct kvm_vcpu
*vcpu
);
1139 #endif /* _ASM_X86_KVM_HOST_H */