diff options
Diffstat (limited to 'usr/src/uts/i86pc/sys/vmm.h')
| -rw-r--r-- | usr/src/uts/i86pc/sys/vmm.h | 258 |
1 files changed, 218 insertions, 40 deletions
diff --git a/usr/src/uts/i86pc/sys/vmm.h b/usr/src/uts/i86pc/sys/vmm.h index e876ce748f..8a35d123c7 100644 --- a/usr/src/uts/i86pc/sys/vmm.h +++ b/usr/src/uts/i86pc/sys/vmm.h @@ -1,4 +1,6 @@ /*- + * SPDX-License-Identifier: BSD-2-Clause-FreeBSD + * * Copyright (c) 2011 NetApp, Inc. * All rights reserved. * @@ -23,7 +25,7 @@ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * - * $FreeBSD: head/sys/amd64/include/vmm.h 273375 2014-10-21 07:10:43Z neel $ + * $FreeBSD$ */ /* * This file and its contents are supplied under the terms of the @@ -36,18 +38,25 @@ * http://www.illumos.org/license/CDDL. * * Copyright 2015 Pluribus Networks Inc. + * Copyright 2019 Joyent, Inc. */ #ifndef _VMM_H_ #define _VMM_H_ +#include <sys/sdt.h> #include <x86/segments.h> +#ifdef _KERNEL +SDT_PROVIDER_DECLARE(vmm); +#endif + enum vm_suspend_how { VM_SUSPEND_NONE, VM_SUSPEND_RESET, VM_SUSPEND_POWEROFF, VM_SUSPEND_HALT, + VM_SUSPEND_TRIPLEFAULT, VM_SUSPEND_LAST }; @@ -89,6 +98,16 @@ enum vm_reg_name { VM_REG_GUEST_GDTR, VM_REG_GUEST_EFER, VM_REG_GUEST_CR2, + VM_REG_GUEST_PDPTE0, + VM_REG_GUEST_PDPTE1, + VM_REG_GUEST_PDPTE2, + VM_REG_GUEST_PDPTE3, + VM_REG_GUEST_INTR_SHADOW, + VM_REG_GUEST_DR0, + VM_REG_GUEST_DR1, + VM_REG_GUEST_DR2, + VM_REG_GUEST_DR3, + VM_REG_GUEST_DR6, VM_REG_LAST }; @@ -108,31 +127,37 @@ enum x2apic_state { #define VM_INTINFO_HWEXCEPTION (3 << 8) #define VM_INTINFO_SWINTR (4 << 8) + #define VM_MAX_NAMELEN 32 #ifdef _KERNEL struct vm; struct vm_exception; -struct vm_memory_segment; struct seg_desc; struct vm_exit; struct vm_run; struct vhpet; struct vioapic; struct vlapic; +struct vmspace; +struct vm_object; struct vm_guest_paging; +struct pmap; -typedef int (*vmm_init_func_t)(void); +struct vm_eventinfo { + u_int *rptr; /* runblock cookie */ + int *sptr; /* suspend cookie */ + int *iptr; /* reqidle cookie */ +}; + +typedef int (*vmm_init_func_t)(int ipinum); typedef int (*vmm_cleanup_func_t)(void); -typedef void * (*vmi_init_func_t)(struct vm *vm); /* instance specific apis */ -typedef int (*vmi_run_func_t)(void *vmi, int vcpu, register_t rip); +typedef void (*vmm_resume_func_t)(void); +typedef void * (*vmi_init_func_t)(struct vm *vm, struct pmap *pmap); +typedef int (*vmi_run_func_t)(void *vmi, int vcpu, register_t rip, + struct pmap *pmap, struct vm_eventinfo *info); typedef void (*vmi_cleanup_func_t)(void *vmi); -typedef int (*vmi_mmap_set_func_t)(void *vmi, vm_paddr_t gpa, - vm_paddr_t hpa, size_t length, - vm_memattr_t attr, int prot, - boolean_t superpages_ok); -typedef vm_paddr_t (*vmi_mmap_get_func_t)(void *vmi, vm_paddr_t gpa); typedef int (*vmi_get_register_t)(void *vmi, int vcpu, int num, uint64_t *retval); typedef int (*vmi_set_register_t)(void *vmi, int vcpu, int num, @@ -143,26 +168,38 @@ typedef int (*vmi_set_desc_t)(void *vmi, int vcpu, int num, struct seg_desc *desc); typedef int (*vmi_get_cap_t)(void *vmi, int vcpu, int num, int *retval); typedef int (*vmi_set_cap_t)(void *vmi, int vcpu, int num, int val); +typedef struct vmspace * (*vmi_vmspace_alloc)(vm_offset_t min, vm_offset_t max); +typedef void (*vmi_vmspace_free)(struct vmspace *vmspace); typedef struct vlapic * (*vmi_vlapic_init)(void *vmi, int vcpu); typedef void (*vmi_vlapic_cleanup)(void *vmi, struct vlapic *vlapic); +#ifndef __FreeBSD__ +typedef void (*vmi_savectx)(void *vmi, int vcpu); +typedef void (*vmi_restorectx)(void *vmi, int vcpu); +#endif struct vmm_ops { vmm_init_func_t init; /* module wide initialization */ vmm_cleanup_func_t cleanup; + vmm_resume_func_t resume; vmi_init_func_t vminit; /* vm-specific initialization */ vmi_run_func_t vmrun; vmi_cleanup_func_t vmcleanup; - vmi_mmap_set_func_t vmmmap_set; - vmi_mmap_get_func_t vmmmap_get; vmi_get_register_t vmgetreg; vmi_set_register_t vmsetreg; vmi_get_desc_t vmgetdesc; vmi_set_desc_t vmsetdesc; vmi_get_cap_t vmgetcap; vmi_set_cap_t vmsetcap; + vmi_vmspace_alloc vmspace_alloc; + vmi_vmspace_free vmspace_free; vmi_vlapic_init vlapic_init; vmi_vlapic_cleanup vlapic_cleanup; + +#ifndef __FreeBSD__ + vmi_savectx vmsavectx; + vmi_restorectx vmrestorectx; +#endif }; extern struct vmm_ops vmm_ops_intel; @@ -170,20 +207,41 @@ extern struct vmm_ops vmm_ops_amd; int vm_create(const char *name, struct vm **retvm); void vm_destroy(struct vm *vm); +int vm_reinit(struct vm *vm); const char *vm_name(struct vm *vm); -int vm_malloc(struct vm *vm, vm_paddr_t gpa, size_t len); -#ifdef __FreeBSD__ +uint16_t vm_get_maxcpus(struct vm *vm); +void vm_get_topology(struct vm *vm, uint16_t *sockets, uint16_t *cores, + uint16_t *threads, uint16_t *maxcpus); +int vm_set_topology(struct vm *vm, uint16_t sockets, uint16_t cores, + uint16_t threads, uint16_t maxcpus); + +/* + * APIs that modify the guest memory map require all vcpus to be frozen. + */ +int vm_mmap_memseg(struct vm *vm, vm_paddr_t gpa, int segid, vm_ooffset_t off, + size_t len, int prot, int flags); +int vm_alloc_memseg(struct vm *vm, int ident, size_t len, bool sysmem); +void vm_free_memseg(struct vm *vm, int ident); int vm_map_mmio(struct vm *vm, vm_paddr_t gpa, size_t len, vm_paddr_t hpa); -#endif int vm_unmap_mmio(struct vm *vm, vm_paddr_t gpa, size_t len); -#ifndef __FreeBSD__ -vm_paddr_t vm_gpa2hpa(struct vm *vm, vm_paddr_t gpa, size_t size); -#endif -void *vm_gpa_hold(struct vm *, vm_paddr_t gpa, size_t len, int prot, - void **cookie); +int vm_assign_pptdev(struct vm *vm, int bus, int slot, int func); +int vm_unassign_pptdev(struct vm *vm, int bus, int slot, int func); + +/* + * APIs that inspect the guest memory map require only a *single* vcpu to + * be frozen. This acts like a read lock on the guest memory map since any + * modification requires *all* vcpus to be frozen. + */ +int vm_mmap_getnext(struct vm *vm, vm_paddr_t *gpa, int *segid, + vm_ooffset_t *segoff, size_t *len, int *prot, int *flags); +int vm_get_memseg(struct vm *vm, int ident, size_t *len, bool *sysmem, + struct vm_object **objptr); +vm_paddr_t vmm_sysmem_maxaddr(struct vm *vm); +void *vm_gpa_hold(struct vm *, int vcpuid, vm_paddr_t gpa, size_t len, + int prot, void **cookie); void vm_gpa_release(void *cookie); -int vm_gpabase2memseg(struct vm *vm, vm_paddr_t gpabase, - struct vm_memory_segment *seg); +bool vm_mem_allocated(struct vm *vm, int vcpuid, vm_paddr_t gpa); + int vm_get_register(struct vm *vm, int vcpu, int reg, uint64_t *retval); int vm_set_register(struct vm *vm, int vcpu, int reg, uint64_t val); int vm_get_seg_desc(struct vm *vm, int vcpu, int reg, @@ -191,6 +249,7 @@ int vm_get_seg_desc(struct vm *vm, int vcpu, int reg, int vm_set_seg_desc(struct vm *vm, int vcpu, int reg, struct seg_desc *desc); int vm_run(struct vm *vm, struct vm_run *vmrun); +int vm_suspend(struct vm *vm, enum vm_suspend_how how); int vm_inject_nmi(struct vm *vm, int vcpu); int vm_nmi_pending(struct vm *vm, int vcpuid); void vm_nmi_clear(struct vm *vm, int vcpuid); @@ -206,10 +265,43 @@ int vm_get_x2apic_state(struct vm *vm, int vcpu, enum x2apic_state *state); int vm_set_x2apic_state(struct vm *vm, int vcpu, enum x2apic_state state); int vm_apicid2vcpuid(struct vm *vm, int apicid); int vm_activate_cpu(struct vm *vm, int vcpu); -cpuset_t vm_active_cpus(struct vm *vm); +int vm_suspend_cpu(struct vm *vm, int vcpu); +int vm_resume_cpu(struct vm *vm, int vcpu); struct vm_exit *vm_exitinfo(struct vm *vm, int vcpuid); +void vm_exit_suspended(struct vm *vm, int vcpuid, uint64_t rip); +void vm_exit_debug(struct vm *vm, int vcpuid, uint64_t rip); +void vm_exit_runblock(struct vm *vm, int vcpuid, uint64_t rip); +void vm_exit_astpending(struct vm *vm, int vcpuid, uint64_t rip); +void vm_exit_reqidle(struct vm *vm, int vcpuid, uint64_t rip); + +#ifdef _SYS__CPUSET_H_ +cpuset_t vm_active_cpus(struct vm *vm); +cpuset_t vm_debug_cpus(struct vm *vm); +cpuset_t vm_suspended_cpus(struct vm *vm); +#endif /* _SYS__CPUSET_H_ */ + +static __inline int +vcpu_runblocked(struct vm_eventinfo *info) +{ -typedef void (*vm_rendezvous_func_t)(struct vm *vm, int vcpuid, void *arg); + return (*info->rptr != 0); +} + +static __inline int +vcpu_suspended(struct vm_eventinfo *info) +{ + + return (*info->sptr); +} + +static __inline int +vcpu_reqidle(struct vm_eventinfo *info) +{ + + return (*info->iptr); +} + +int vcpu_debugged(struct vm *vm, int vcpuid); /* * Return 1 if device indicated by bus/slot/func is supposed to be a @@ -231,21 +323,43 @@ enum vcpu_state { int vcpu_set_state(struct vm *vm, int vcpu, enum vcpu_state state, bool from_idle); enum vcpu_state vcpu_get_state(struct vm *vm, int vcpu, int *hostcpu); +void vcpu_block_run(struct vm *, int); +void vcpu_unblock_run(struct vm *, int); + +#ifndef __FreeBSD__ +uint64_t vcpu_tsc_offset(struct vm *vm, int vcpuid); +#endif -static int __inline +static __inline int vcpu_is_running(struct vm *vm, int vcpu, int *hostcpu) { return (vcpu_get_state(vm, vcpu, hostcpu) == VCPU_RUNNING); } +#ifdef _SYS_THREAD_H +static __inline int +vcpu_should_yield(struct vm *vm, int vcpu) +{ + + if (curthread->t_astflag) + return (1); + else if (CPU->cpu_runrun) + return (1); + else + return (0); +} +#endif /* _SYS_THREAD_H */ + void *vcpu_stats(struct vm *vm, int vcpu); -void vm_interrupt_hostcpu(struct vm *vm, int vcpu); void vcpu_notify_event(struct vm *vm, int vcpuid, bool lapic_intr); +struct vmspace *vm_get_vmspace(struct vm *vm); struct vatpic *vm_atpic(struct vm *vm); struct vatpit *vm_atpit(struct vm *vm); +struct vpmtmr *vm_pmtmr(struct vm *vm); +struct vrtc *vm_rtc(struct vm *vm); /* - * Inject exception 'vme' into the guest vcpu. This function returns 0 on + * Inject exception 'vector' into the guest vcpu. This function returns 0 on * success and non-zero on failure. * * Wrapper functions like 'vm_inject_gp()' should be preferred to calling @@ -255,7 +369,8 @@ struct vatpit *vm_atpit(struct vm *vm); * This function should only be called in the context of the thread that is * executing this vcpu. */ -int vm_inject_exception(struct vm *vm, int vcpuid, struct vm_exception *vme); +int vm_inject_exception(struct vm *vm, int vcpuid, int vector, int err_valid, + uint32_t errcode, int restart_instruction); /* * This function is called after a VM-exit that occurred during exception or @@ -298,9 +413,10 @@ struct vm_copyinfo { * at 'gla' and 'len' bytes long. The 'prot' should be set to PROT_READ for * a copyin or PROT_WRITE for a copyout. * - * Returns 0 on success. - * Returns 1 if an exception was injected into the guest. - * Returns -1 otherwise. + * retval is_fault Interpretation + * 0 0 Success + * 0 1 An exception was injected into the guest + * EFAULT N/A Unrecoverable error * * The 'copyinfo[]' can be passed to 'vm_copyin()' or 'vm_copyout()' only if * the return value is 0. The 'copyinfo[]' resources should be freed by calling @@ -308,16 +424,18 @@ struct vm_copyinfo { */ int vm_copy_setup(struct vm *vm, int vcpuid, struct vm_guest_paging *paging, uint64_t gla, size_t len, int prot, struct vm_copyinfo *copyinfo, - int num_copyinfo); + int num_copyinfo, int *is_fault); void vm_copy_teardown(struct vm *vm, int vcpuid, struct vm_copyinfo *copyinfo, int num_copyinfo); void vm_copyin(struct vm *vm, int vcpuid, struct vm_copyinfo *copyinfo, void *kaddr, size_t len); void vm_copyout(struct vm *vm, int vcpuid, const void *kaddr, struct vm_copyinfo *copyinfo, size_t len); + +int vcpu_trace_exceptions(struct vm *vm, int vcpuid); #endif /* KERNEL */ -#define VM_MAXCPU 16 /* maximum virtual cpus */ +#define VM_MAXCPU 32 /* maximum virtual cpus */ /* * Identifiers for optional vmm capabilities @@ -348,7 +466,6 @@ struct seg_desc { uint32_t limit; uint32_t access; }; - #define SEG_DESC_TYPE(access) ((access) & 0x001f) #define SEG_DESC_DPL(access) (((access) >> 5) & 0x3) #define SEG_DESC_PRESENT(access) (((access) & 0x0080) ? 1 : 0) @@ -443,7 +560,20 @@ enum vm_exitcode { VM_EXITCODE_INST_EMUL, VM_EXITCODE_SPINUP_AP, VM_EXITCODE_DEPRECATED1, /* used to be SPINDOWN_CPU */ + VM_EXITCODE_RUNBLOCK, + VM_EXITCODE_IOAPIC_EOI, + VM_EXITCODE_SUSPENDED, VM_EXITCODE_INOUT_STR, + VM_EXITCODE_TASK_SWITCH, + VM_EXITCODE_MONITOR, + VM_EXITCODE_MWAIT, + VM_EXITCODE_SVM, + VM_EXITCODE_REQIDLE, + VM_EXITCODE_DEBUG, + VM_EXITCODE_VMINSN, +#ifndef __FreeBSD__ + VM_EXITCODE_HT, +#endif VM_EXITCODE_MAX }; @@ -468,6 +598,22 @@ struct vm_inout_str { struct seg_desc seg_desc; }; +enum task_switch_reason { + TSR_CALL, + TSR_IRET, + TSR_JMP, + TSR_IDT_GATE, /* task gate in IDT */ +}; + +struct vm_task_switch { + uint16_t tsssel; /* new TSS selector */ + int ext; /* task switch due to external event */ + uint32_t errcode; + int errcode_valid; /* push 'errcode' on the new stack */ + enum task_switch_reason reason; + struct vm_guest_paging paging; +}; + struct vm_exit { enum vm_exitcode exitcode; int inst_length; /* 0 means unknown */ @@ -506,6 +652,14 @@ struct vm_exit { int inst_type; int inst_error; } vmx; + /* + * SVM specific payload. + */ + struct { + uint64_t exitcode; + uint64_t exitinfo1; + uint64_t exitinfo2; + } svm; struct { uint32_t code; /* ecx value */ uint64_t wval; @@ -516,7 +670,15 @@ struct vm_exit { } spinup_ap; struct { uint64_t rflags; + uint64_t intr_status; } hlt; + struct { + int vector; + } ioapic_eoi; + struct { + enum vm_suspend_how how; + } suspended; + struct vm_task_switch task_switch; } u; }; @@ -554,12 +716,28 @@ int vm_restart_instruction(void *vm, int vcpuid); #ifndef __FreeBSD__ #ifdef _KERNEL -extern void vmm_sol_glue_init(void); -extern void vmm_sol_glue_cleanup(void); -extern int vmm_mod_load(void); -extern int vmm_mod_unload(void); -#endif -#endif +void vmm_sol_glue_init(void); +void vmm_sol_glue_cleanup(void); + +int vmm_mod_load(void); +int vmm_mod_unload(void); + +void vmm_call_trap(uint64_t); + +/* + * Because of tangled headers, these are mirrored by vmm_drv.h to present the + * interface to driver consumers. + */ +typedef int (*vmm_rmem_cb_t)(void *, uintptr_t, uint_t, uint64_t *); +typedef int (*vmm_wmem_cb_t)(void *, uintptr_t, uint_t, uint64_t); + +int vm_ioport_hook(struct vm *, uint_t, vmm_rmem_cb_t, vmm_wmem_cb_t, void *, + void **); +void vm_ioport_unhook(struct vm *, void **); +int vm_ioport_handle_hook(struct vm *, int, bool, int, int, uint32_t *); + +#endif /* _KERNEL */ +#endif /* __FreeBSD */ #endif /* _VMM_H_ */ |