freebsd-dev/sys/amd64/include/vmm.h

433 lines
12 KiB
C
Raw Normal View History

/*-
* Copyright (c) 2011 NetApp, Inc.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY NETAPP, INC ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL NETAPP, INC OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* $FreeBSD$
*/
#ifndef _VMM_H_
#define _VMM_H_
enum vm_suspend_how {
VM_SUSPEND_NONE,
VM_SUSPEND_RESET,
VM_SUSPEND_POWEROFF,
VM_SUSPEND_HALT,
VM_SUSPEND_LAST
};
#ifdef _KERNEL
#define VM_MAX_NAMELEN 32
struct vm;
struct vm_exception;
struct vm_memory_segment;
struct seg_desc;
struct vm_exit;
struct vm_run;
struct vhpet;
struct vioapic;
struct vlapic;
Merge projects/bhyve_npt_pmap into head. Make the amd64/pmap code aware of nested page table mappings used by bhyve guests. This allows bhyve to associate each guest with its own vmspace and deal with nested page faults in the context of that vmspace. This also enables features like accessed/dirty bit tracking, swapping to disk and transparent superpage promotions of guest memory. Guest vmspace: Each bhyve guest has a unique vmspace to represent the physical memory allocated to the guest. Each memory segment allocated by the guest is mapped into the guest's address space via the 'vmspace->vm_map' and is backed by an object of type OBJT_DEFAULT. pmap types: The amd64/pmap now understands two types of pmaps: PT_X86 and PT_EPT. The PT_X86 pmap type is used by the vmspace associated with the host kernel as well as user processes executing on the host. The PT_EPT pmap is used by the vmspace associated with a bhyve guest. Page Table Entries: The EPT page table entries as mostly similar in functionality to regular page table entries although there are some differences in terms of what bits are used to express that functionality. For e.g. the dirty bit is represented by bit 9 in the nested PTE as opposed to bit 6 in the regular x86 PTE. Therefore the bitmask representing the dirty bit is now computed at runtime based on the type of the pmap. Thus PG_M that was previously a macro now becomes a local variable that is initialized at runtime using 'pmap_modified_bit(pmap)'. An additional wrinkle associated with EPT mappings is that older Intel processors don't have hardware support for tracking accessed/dirty bits in the PTE. This means that the amd64/pmap code needs to emulate these bits to provide proper accounting to the VM subsystem. This is achieved by using the following mapping for EPT entries that need emulation of A/D bits: Bit Position Interpreted By PG_V 52 software (accessed bit emulation handler) PG_RW 53 software (dirty bit emulation handler) PG_A 0 hardware (aka EPT_PG_RD) PG_M 1 hardware (aka EPT_PG_WR) The idea to use the mapping listed above for A/D bit emulation came from Alan Cox (alc@). The final difference with respect to x86 PTEs is that some EPT implementations do not support superpage mappings. This is recorded in the 'pm_flags' field of the pmap. TLB invalidation: The amd64/pmap code has a number of ways to do invalidation of mappings that may be cached in the TLB: single page, multiple pages in a range or the entire TLB. All of these funnel into a single EPT invalidation routine called 'pmap_invalidate_ept()'. This routine bumps up the EPT generation number and sends an IPI to the host cpus that are executing the guest's vcpus. On a subsequent entry into the guest it will detect that the EPT has changed and invalidate the mappings from the TLB. Guest memory access: Since the guest memory is no longer wired we need to hold the host physical page that backs the guest physical page before we can access it. The helper functions 'vm_gpa_hold()/vm_gpa_release()' are available for this purpose. PCI passthru: Guest's with PCI passthru devices will wire the entire guest physical address space. The MMIO BAR associated with the passthru device is backed by a vm_object of type OBJT_SG. An IOMMU domain is created only for guest's that have one or more PCI passthru devices attached to them. Limitations: There isn't a way to map a guest physical page without execute permissions. This is because the amd64/pmap code interprets the guest physical mappings as user mappings since they are numerically below VM_MAXUSER_ADDRESS. Since PG_U shares the same bit position as EPT_PG_EXECUTE all guest mappings become automatically executable. Thanks to Alan Cox and Konstantin Belousov for their rigorous code reviews as well as their support and encouragement. Thanks for John Baldwin for reviewing the use of OBJT_SG as the backing object for pci passthru mmio regions. Special thanks to Peter Holm for testing the patch on short notice. Approved by: re Discussed with: grehan Reviewed by: alc, kib Tested by: pho
2013-10-05 21:22:35 +00:00
struct vmspace;
struct vm_object;
struct pmap;
enum vm_reg_name;
enum x2apic_state;
2014-01-09 03:25:54 +00:00
typedef int (*vmm_init_func_t)(int ipinum);
typedef int (*vmm_cleanup_func_t)(void);
typedef void (*vmm_resume_func_t)(void);
Merge projects/bhyve_npt_pmap into head. Make the amd64/pmap code aware of nested page table mappings used by bhyve guests. This allows bhyve to associate each guest with its own vmspace and deal with nested page faults in the context of that vmspace. This also enables features like accessed/dirty bit tracking, swapping to disk and transparent superpage promotions of guest memory. Guest vmspace: Each bhyve guest has a unique vmspace to represent the physical memory allocated to the guest. Each memory segment allocated by the guest is mapped into the guest's address space via the 'vmspace->vm_map' and is backed by an object of type OBJT_DEFAULT. pmap types: The amd64/pmap now understands two types of pmaps: PT_X86 and PT_EPT. The PT_X86 pmap type is used by the vmspace associated with the host kernel as well as user processes executing on the host. The PT_EPT pmap is used by the vmspace associated with a bhyve guest. Page Table Entries: The EPT page table entries as mostly similar in functionality to regular page table entries although there are some differences in terms of what bits are used to express that functionality. For e.g. the dirty bit is represented by bit 9 in the nested PTE as opposed to bit 6 in the regular x86 PTE. Therefore the bitmask representing the dirty bit is now computed at runtime based on the type of the pmap. Thus PG_M that was previously a macro now becomes a local variable that is initialized at runtime using 'pmap_modified_bit(pmap)'. An additional wrinkle associated with EPT mappings is that older Intel processors don't have hardware support for tracking accessed/dirty bits in the PTE. This means that the amd64/pmap code needs to emulate these bits to provide proper accounting to the VM subsystem. This is achieved by using the following mapping for EPT entries that need emulation of A/D bits: Bit Position Interpreted By PG_V 52 software (accessed bit emulation handler) PG_RW 53 software (dirty bit emulation handler) PG_A 0 hardware (aka EPT_PG_RD) PG_M 1 hardware (aka EPT_PG_WR) The idea to use the mapping listed above for A/D bit emulation came from Alan Cox (alc@). The final difference with respect to x86 PTEs is that some EPT implementations do not support superpage mappings. This is recorded in the 'pm_flags' field of the pmap. TLB invalidation: The amd64/pmap code has a number of ways to do invalidation of mappings that may be cached in the TLB: single page, multiple pages in a range or the entire TLB. All of these funnel into a single EPT invalidation routine called 'pmap_invalidate_ept()'. This routine bumps up the EPT generation number and sends an IPI to the host cpus that are executing the guest's vcpus. On a subsequent entry into the guest it will detect that the EPT has changed and invalidate the mappings from the TLB. Guest memory access: Since the guest memory is no longer wired we need to hold the host physical page that backs the guest physical page before we can access it. The helper functions 'vm_gpa_hold()/vm_gpa_release()' are available for this purpose. PCI passthru: Guest's with PCI passthru devices will wire the entire guest physical address space. The MMIO BAR associated with the passthru device is backed by a vm_object of type OBJT_SG. An IOMMU domain is created only for guest's that have one or more PCI passthru devices attached to them. Limitations: There isn't a way to map a guest physical page without execute permissions. This is because the amd64/pmap code interprets the guest physical mappings as user mappings since they are numerically below VM_MAXUSER_ADDRESS. Since PG_U shares the same bit position as EPT_PG_EXECUTE all guest mappings become automatically executable. Thanks to Alan Cox and Konstantin Belousov for their rigorous code reviews as well as their support and encouragement. Thanks for John Baldwin for reviewing the use of OBJT_SG as the backing object for pci passthru mmio regions. Special thanks to Peter Holm for testing the patch on short notice. Approved by: re Discussed with: grehan Reviewed by: alc, kib Tested by: pho
2013-10-05 21:22:35 +00:00
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, void *rendezvous_cookie,
void *suspend_cookie);
typedef void (*vmi_cleanup_func_t)(void *vmi);
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,
uint64_t val);
typedef int (*vmi_get_desc_t)(void *vmi, int vcpu, int num,
struct seg_desc *desc);
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);
Merge projects/bhyve_npt_pmap into head. Make the amd64/pmap code aware of nested page table mappings used by bhyve guests. This allows bhyve to associate each guest with its own vmspace and deal with nested page faults in the context of that vmspace. This also enables features like accessed/dirty bit tracking, swapping to disk and transparent superpage promotions of guest memory. Guest vmspace: Each bhyve guest has a unique vmspace to represent the physical memory allocated to the guest. Each memory segment allocated by the guest is mapped into the guest's address space via the 'vmspace->vm_map' and is backed by an object of type OBJT_DEFAULT. pmap types: The amd64/pmap now understands two types of pmaps: PT_X86 and PT_EPT. The PT_X86 pmap type is used by the vmspace associated with the host kernel as well as user processes executing on the host. The PT_EPT pmap is used by the vmspace associated with a bhyve guest. Page Table Entries: The EPT page table entries as mostly similar in functionality to regular page table entries although there are some differences in terms of what bits are used to express that functionality. For e.g. the dirty bit is represented by bit 9 in the nested PTE as opposed to bit 6 in the regular x86 PTE. Therefore the bitmask representing the dirty bit is now computed at runtime based on the type of the pmap. Thus PG_M that was previously a macro now becomes a local variable that is initialized at runtime using 'pmap_modified_bit(pmap)'. An additional wrinkle associated with EPT mappings is that older Intel processors don't have hardware support for tracking accessed/dirty bits in the PTE. This means that the amd64/pmap code needs to emulate these bits to provide proper accounting to the VM subsystem. This is achieved by using the following mapping for EPT entries that need emulation of A/D bits: Bit Position Interpreted By PG_V 52 software (accessed bit emulation handler) PG_RW 53 software (dirty bit emulation handler) PG_A 0 hardware (aka EPT_PG_RD) PG_M 1 hardware (aka EPT_PG_WR) The idea to use the mapping listed above for A/D bit emulation came from Alan Cox (alc@). The final difference with respect to x86 PTEs is that some EPT implementations do not support superpage mappings. This is recorded in the 'pm_flags' field of the pmap. TLB invalidation: The amd64/pmap code has a number of ways to do invalidation of mappings that may be cached in the TLB: single page, multiple pages in a range or the entire TLB. All of these funnel into a single EPT invalidation routine called 'pmap_invalidate_ept()'. This routine bumps up the EPT generation number and sends an IPI to the host cpus that are executing the guest's vcpus. On a subsequent entry into the guest it will detect that the EPT has changed and invalidate the mappings from the TLB. Guest memory access: Since the guest memory is no longer wired we need to hold the host physical page that backs the guest physical page before we can access it. The helper functions 'vm_gpa_hold()/vm_gpa_release()' are available for this purpose. PCI passthru: Guest's with PCI passthru devices will wire the entire guest physical address space. The MMIO BAR associated with the passthru device is backed by a vm_object of type OBJT_SG. An IOMMU domain is created only for guest's that have one or more PCI passthru devices attached to them. Limitations: There isn't a way to map a guest physical page without execute permissions. This is because the amd64/pmap code interprets the guest physical mappings as user mappings since they are numerically below VM_MAXUSER_ADDRESS. Since PG_U shares the same bit position as EPT_PG_EXECUTE all guest mappings become automatically executable. Thanks to Alan Cox and Konstantin Belousov for their rigorous code reviews as well as their support and encouragement. Thanks for John Baldwin for reviewing the use of OBJT_SG as the backing object for pci passthru mmio regions. Special thanks to Peter Holm for testing the patch on short notice. Approved by: re Discussed with: grehan Reviewed by: alc, kib Tested by: pho
2013-10-05 21:22:35 +00:00
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);
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_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;
Merge projects/bhyve_npt_pmap into head. Make the amd64/pmap code aware of nested page table mappings used by bhyve guests. This allows bhyve to associate each guest with its own vmspace and deal with nested page faults in the context of that vmspace. This also enables features like accessed/dirty bit tracking, swapping to disk and transparent superpage promotions of guest memory. Guest vmspace: Each bhyve guest has a unique vmspace to represent the physical memory allocated to the guest. Each memory segment allocated by the guest is mapped into the guest's address space via the 'vmspace->vm_map' and is backed by an object of type OBJT_DEFAULT. pmap types: The amd64/pmap now understands two types of pmaps: PT_X86 and PT_EPT. The PT_X86 pmap type is used by the vmspace associated with the host kernel as well as user processes executing on the host. The PT_EPT pmap is used by the vmspace associated with a bhyve guest. Page Table Entries: The EPT page table entries as mostly similar in functionality to regular page table entries although there are some differences in terms of what bits are used to express that functionality. For e.g. the dirty bit is represented by bit 9 in the nested PTE as opposed to bit 6 in the regular x86 PTE. Therefore the bitmask representing the dirty bit is now computed at runtime based on the type of the pmap. Thus PG_M that was previously a macro now becomes a local variable that is initialized at runtime using 'pmap_modified_bit(pmap)'. An additional wrinkle associated with EPT mappings is that older Intel processors don't have hardware support for tracking accessed/dirty bits in the PTE. This means that the amd64/pmap code needs to emulate these bits to provide proper accounting to the VM subsystem. This is achieved by using the following mapping for EPT entries that need emulation of A/D bits: Bit Position Interpreted By PG_V 52 software (accessed bit emulation handler) PG_RW 53 software (dirty bit emulation handler) PG_A 0 hardware (aka EPT_PG_RD) PG_M 1 hardware (aka EPT_PG_WR) The idea to use the mapping listed above for A/D bit emulation came from Alan Cox (alc@). The final difference with respect to x86 PTEs is that some EPT implementations do not support superpage mappings. This is recorded in the 'pm_flags' field of the pmap. TLB invalidation: The amd64/pmap code has a number of ways to do invalidation of mappings that may be cached in the TLB: single page, multiple pages in a range or the entire TLB. All of these funnel into a single EPT invalidation routine called 'pmap_invalidate_ept()'. This routine bumps up the EPT generation number and sends an IPI to the host cpus that are executing the guest's vcpus. On a subsequent entry into the guest it will detect that the EPT has changed and invalidate the mappings from the TLB. Guest memory access: Since the guest memory is no longer wired we need to hold the host physical page that backs the guest physical page before we can access it. The helper functions 'vm_gpa_hold()/vm_gpa_release()' are available for this purpose. PCI passthru: Guest's with PCI passthru devices will wire the entire guest physical address space. The MMIO BAR associated with the passthru device is backed by a vm_object of type OBJT_SG. An IOMMU domain is created only for guest's that have one or more PCI passthru devices attached to them. Limitations: There isn't a way to map a guest physical page without execute permissions. This is because the amd64/pmap code interprets the guest physical mappings as user mappings since they are numerically below VM_MAXUSER_ADDRESS. Since PG_U shares the same bit position as EPT_PG_EXECUTE all guest mappings become automatically executable. Thanks to Alan Cox and Konstantin Belousov for their rigorous code reviews as well as their support and encouragement. Thanks for John Baldwin for reviewing the use of OBJT_SG as the backing object for pci passthru mmio regions. Special thanks to Peter Holm for testing the patch on short notice. Approved by: re Discussed with: grehan Reviewed by: alc, kib Tested by: pho
2013-10-05 21:22:35 +00:00
vmi_vmspace_alloc vmspace_alloc;
vmi_vmspace_free vmspace_free;
vmi_vlapic_init vlapic_init;
vmi_vlapic_cleanup vlapic_cleanup;
};
extern struct vmm_ops vmm_ops_intel;
extern struct vmm_ops vmm_ops_amd;
int vm_create(const char *name, struct vm **retvm);
void vm_destroy(struct vm *vm);
const char *vm_name(struct vm *vm);
int vm_malloc(struct vm *vm, vm_paddr_t gpa, size_t len);
int vm_map_mmio(struct vm *vm, vm_paddr_t gpa, size_t len, vm_paddr_t hpa);
int vm_unmap_mmio(struct vm *vm, vm_paddr_t gpa, size_t len);
Merge projects/bhyve_npt_pmap into head. Make the amd64/pmap code aware of nested page table mappings used by bhyve guests. This allows bhyve to associate each guest with its own vmspace and deal with nested page faults in the context of that vmspace. This also enables features like accessed/dirty bit tracking, swapping to disk and transparent superpage promotions of guest memory. Guest vmspace: Each bhyve guest has a unique vmspace to represent the physical memory allocated to the guest. Each memory segment allocated by the guest is mapped into the guest's address space via the 'vmspace->vm_map' and is backed by an object of type OBJT_DEFAULT. pmap types: The amd64/pmap now understands two types of pmaps: PT_X86 and PT_EPT. The PT_X86 pmap type is used by the vmspace associated with the host kernel as well as user processes executing on the host. The PT_EPT pmap is used by the vmspace associated with a bhyve guest. Page Table Entries: The EPT page table entries as mostly similar in functionality to regular page table entries although there are some differences in terms of what bits are used to express that functionality. For e.g. the dirty bit is represented by bit 9 in the nested PTE as opposed to bit 6 in the regular x86 PTE. Therefore the bitmask representing the dirty bit is now computed at runtime based on the type of the pmap. Thus PG_M that was previously a macro now becomes a local variable that is initialized at runtime using 'pmap_modified_bit(pmap)'. An additional wrinkle associated with EPT mappings is that older Intel processors don't have hardware support for tracking accessed/dirty bits in the PTE. This means that the amd64/pmap code needs to emulate these bits to provide proper accounting to the VM subsystem. This is achieved by using the following mapping for EPT entries that need emulation of A/D bits: Bit Position Interpreted By PG_V 52 software (accessed bit emulation handler) PG_RW 53 software (dirty bit emulation handler) PG_A 0 hardware (aka EPT_PG_RD) PG_M 1 hardware (aka EPT_PG_WR) The idea to use the mapping listed above for A/D bit emulation came from Alan Cox (alc@). The final difference with respect to x86 PTEs is that some EPT implementations do not support superpage mappings. This is recorded in the 'pm_flags' field of the pmap. TLB invalidation: The amd64/pmap code has a number of ways to do invalidation of mappings that may be cached in the TLB: single page, multiple pages in a range or the entire TLB. All of these funnel into a single EPT invalidation routine called 'pmap_invalidate_ept()'. This routine bumps up the EPT generation number and sends an IPI to the host cpus that are executing the guest's vcpus. On a subsequent entry into the guest it will detect that the EPT has changed and invalidate the mappings from the TLB. Guest memory access: Since the guest memory is no longer wired we need to hold the host physical page that backs the guest physical page before we can access it. The helper functions 'vm_gpa_hold()/vm_gpa_release()' are available for this purpose. PCI passthru: Guest's with PCI passthru devices will wire the entire guest physical address space. The MMIO BAR associated with the passthru device is backed by a vm_object of type OBJT_SG. An IOMMU domain is created only for guest's that have one or more PCI passthru devices attached to them. Limitations: There isn't a way to map a guest physical page without execute permissions. This is because the amd64/pmap code interprets the guest physical mappings as user mappings since they are numerically below VM_MAXUSER_ADDRESS. Since PG_U shares the same bit position as EPT_PG_EXECUTE all guest mappings become automatically executable. Thanks to Alan Cox and Konstantin Belousov for their rigorous code reviews as well as their support and encouragement. Thanks for John Baldwin for reviewing the use of OBJT_SG as the backing object for pci passthru mmio regions. Special thanks to Peter Holm for testing the patch on short notice. Approved by: re Discussed with: grehan Reviewed by: alc, kib Tested by: pho
2013-10-05 21:22:35 +00:00
void *vm_gpa_hold(struct vm *, 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);
Merge projects/bhyve_npt_pmap into head. Make the amd64/pmap code aware of nested page table mappings used by bhyve guests. This allows bhyve to associate each guest with its own vmspace and deal with nested page faults in the context of that vmspace. This also enables features like accessed/dirty bit tracking, swapping to disk and transparent superpage promotions of guest memory. Guest vmspace: Each bhyve guest has a unique vmspace to represent the physical memory allocated to the guest. Each memory segment allocated by the guest is mapped into the guest's address space via the 'vmspace->vm_map' and is backed by an object of type OBJT_DEFAULT. pmap types: The amd64/pmap now understands two types of pmaps: PT_X86 and PT_EPT. The PT_X86 pmap type is used by the vmspace associated with the host kernel as well as user processes executing on the host. The PT_EPT pmap is used by the vmspace associated with a bhyve guest. Page Table Entries: The EPT page table entries as mostly similar in functionality to regular page table entries although there are some differences in terms of what bits are used to express that functionality. For e.g. the dirty bit is represented by bit 9 in the nested PTE as opposed to bit 6 in the regular x86 PTE. Therefore the bitmask representing the dirty bit is now computed at runtime based on the type of the pmap. Thus PG_M that was previously a macro now becomes a local variable that is initialized at runtime using 'pmap_modified_bit(pmap)'. An additional wrinkle associated with EPT mappings is that older Intel processors don't have hardware support for tracking accessed/dirty bits in the PTE. This means that the amd64/pmap code needs to emulate these bits to provide proper accounting to the VM subsystem. This is achieved by using the following mapping for EPT entries that need emulation of A/D bits: Bit Position Interpreted By PG_V 52 software (accessed bit emulation handler) PG_RW 53 software (dirty bit emulation handler) PG_A 0 hardware (aka EPT_PG_RD) PG_M 1 hardware (aka EPT_PG_WR) The idea to use the mapping listed above for A/D bit emulation came from Alan Cox (alc@). The final difference with respect to x86 PTEs is that some EPT implementations do not support superpage mappings. This is recorded in the 'pm_flags' field of the pmap. TLB invalidation: The amd64/pmap code has a number of ways to do invalidation of mappings that may be cached in the TLB: single page, multiple pages in a range or the entire TLB. All of these funnel into a single EPT invalidation routine called 'pmap_invalidate_ept()'. This routine bumps up the EPT generation number and sends an IPI to the host cpus that are executing the guest's vcpus. On a subsequent entry into the guest it will detect that the EPT has changed and invalidate the mappings from the TLB. Guest memory access: Since the guest memory is no longer wired we need to hold the host physical page that backs the guest physical page before we can access it. The helper functions 'vm_gpa_hold()/vm_gpa_release()' are available for this purpose. PCI passthru: Guest's with PCI passthru devices will wire the entire guest physical address space. The MMIO BAR associated with the passthru device is backed by a vm_object of type OBJT_SG. An IOMMU domain is created only for guest's that have one or more PCI passthru devices attached to them. Limitations: There isn't a way to map a guest physical page without execute permissions. This is because the amd64/pmap code interprets the guest physical mappings as user mappings since they are numerically below VM_MAXUSER_ADDRESS. Since PG_U shares the same bit position as EPT_PG_EXECUTE all guest mappings become automatically executable. Thanks to Alan Cox and Konstantin Belousov for their rigorous code reviews as well as their support and encouragement. Thanks for John Baldwin for reviewing the use of OBJT_SG as the backing object for pci passthru mmio regions. Special thanks to Peter Holm for testing the patch on short notice. Approved by: re Discussed with: grehan Reviewed by: alc, kib Tested by: pho
2013-10-05 21:22:35 +00:00
int vm_get_memobj(struct vm *vm, vm_paddr_t gpa, size_t len,
vm_offset_t *offset, struct vm_object **object);
boolean_t vm_mem_allocated(struct vm *vm, 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,
struct seg_desc *ret_desc);
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);
int vm_inject_extint(struct vm *vm, int vcpu);
int vm_extint_pending(struct vm *vm, int vcpuid);
void vm_extint_clear(struct vm *vm, int vcpuid);
uint64_t *vm_guest_msrs(struct vm *vm, int cpu);
struct vlapic *vm_lapic(struct vm *vm, int cpu);
struct vioapic *vm_ioapic(struct vm *vm);
struct vhpet *vm_hpet(struct vm *vm);
int vm_get_capability(struct vm *vm, int vcpu, int type, int *val);
int vm_set_capability(struct vm *vm, int vcpu, int type, int val);
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);
void vm_activate_cpu(struct vm *vm, int vcpu);
2011-06-28 06:26:03 +00:00
cpuset_t vm_active_cpus(struct vm *vm);
struct vm_exit *vm_exitinfo(struct vm *vm, int vcpuid);
void vm_exit_suspended(struct vm *vm, int vcpuid, uint64_t rip);
/*
* Rendezvous all vcpus specified in 'dest' and execute 'func(arg)'.
* The rendezvous 'func(arg)' is not allowed to do anything that will
* cause the thread to be put to sleep.
*
* If the rendezvous is being initiated from a vcpu context then the
* 'vcpuid' must refer to that vcpu, otherwise it should be set to -1.
*
* The caller cannot hold any locks when initiating the rendezvous.
*
* The implementation of this API may cause vcpus other than those specified
* by 'dest' to be stalled. The caller should not rely on any vcpus making
* forward progress when the rendezvous is in progress.
*/
typedef void (*vm_rendezvous_func_t)(struct vm *vm, int vcpuid, void *arg);
void vm_smp_rendezvous(struct vm *vm, int vcpuid, cpuset_t dest,
vm_rendezvous_func_t func, void *arg);
static __inline int
vcpu_rendezvous_pending(void *rendezvous_cookie)
{
return (*(uintptr_t *)rendezvous_cookie != 0);
}
static __inline int
vcpu_suspended(void *suspend_cookie)
{
return (*(int *)suspend_cookie);
}
/*
* Return 1 if device indicated by bus/slot/func is supposed to be a
* pci passthrough device.
*
* Return 0 otherwise.
*/
int vmm_is_pptdev(int bus, int slot, int func);
void *vm_iommu_domain(struct vm *vm);
enum vcpu_state {
VCPU_IDLE,
Merge projects/bhyve_npt_pmap into head. Make the amd64/pmap code aware of nested page table mappings used by bhyve guests. This allows bhyve to associate each guest with its own vmspace and deal with nested page faults in the context of that vmspace. This also enables features like accessed/dirty bit tracking, swapping to disk and transparent superpage promotions of guest memory. Guest vmspace: Each bhyve guest has a unique vmspace to represent the physical memory allocated to the guest. Each memory segment allocated by the guest is mapped into the guest's address space via the 'vmspace->vm_map' and is backed by an object of type OBJT_DEFAULT. pmap types: The amd64/pmap now understands two types of pmaps: PT_X86 and PT_EPT. The PT_X86 pmap type is used by the vmspace associated with the host kernel as well as user processes executing on the host. The PT_EPT pmap is used by the vmspace associated with a bhyve guest. Page Table Entries: The EPT page table entries as mostly similar in functionality to regular page table entries although there are some differences in terms of what bits are used to express that functionality. For e.g. the dirty bit is represented by bit 9 in the nested PTE as opposed to bit 6 in the regular x86 PTE. Therefore the bitmask representing the dirty bit is now computed at runtime based on the type of the pmap. Thus PG_M that was previously a macro now becomes a local variable that is initialized at runtime using 'pmap_modified_bit(pmap)'. An additional wrinkle associated with EPT mappings is that older Intel processors don't have hardware support for tracking accessed/dirty bits in the PTE. This means that the amd64/pmap code needs to emulate these bits to provide proper accounting to the VM subsystem. This is achieved by using the following mapping for EPT entries that need emulation of A/D bits: Bit Position Interpreted By PG_V 52 software (accessed bit emulation handler) PG_RW 53 software (dirty bit emulation handler) PG_A 0 hardware (aka EPT_PG_RD) PG_M 1 hardware (aka EPT_PG_WR) The idea to use the mapping listed above for A/D bit emulation came from Alan Cox (alc@). The final difference with respect to x86 PTEs is that some EPT implementations do not support superpage mappings. This is recorded in the 'pm_flags' field of the pmap. TLB invalidation: The amd64/pmap code has a number of ways to do invalidation of mappings that may be cached in the TLB: single page, multiple pages in a range or the entire TLB. All of these funnel into a single EPT invalidation routine called 'pmap_invalidate_ept()'. This routine bumps up the EPT generation number and sends an IPI to the host cpus that are executing the guest's vcpus. On a subsequent entry into the guest it will detect that the EPT has changed and invalidate the mappings from the TLB. Guest memory access: Since the guest memory is no longer wired we need to hold the host physical page that backs the guest physical page before we can access it. The helper functions 'vm_gpa_hold()/vm_gpa_release()' are available for this purpose. PCI passthru: Guest's with PCI passthru devices will wire the entire guest physical address space. The MMIO BAR associated with the passthru device is backed by a vm_object of type OBJT_SG. An IOMMU domain is created only for guest's that have one or more PCI passthru devices attached to them. Limitations: There isn't a way to map a guest physical page without execute permissions. This is because the amd64/pmap code interprets the guest physical mappings as user mappings since they are numerically below VM_MAXUSER_ADDRESS. Since PG_U shares the same bit position as EPT_PG_EXECUTE all guest mappings become automatically executable. Thanks to Alan Cox and Konstantin Belousov for their rigorous code reviews as well as their support and encouragement. Thanks for John Baldwin for reviewing the use of OBJT_SG as the backing object for pci passthru mmio regions. Special thanks to Peter Holm for testing the patch on short notice. Approved by: re Discussed with: grehan Reviewed by: alc, kib Tested by: pho
2013-10-05 21:22:35 +00:00
VCPU_FROZEN,
VCPU_RUNNING,
Merge projects/bhyve_npt_pmap into head. Make the amd64/pmap code aware of nested page table mappings used by bhyve guests. This allows bhyve to associate each guest with its own vmspace and deal with nested page faults in the context of that vmspace. This also enables features like accessed/dirty bit tracking, swapping to disk and transparent superpage promotions of guest memory. Guest vmspace: Each bhyve guest has a unique vmspace to represent the physical memory allocated to the guest. Each memory segment allocated by the guest is mapped into the guest's address space via the 'vmspace->vm_map' and is backed by an object of type OBJT_DEFAULT. pmap types: The amd64/pmap now understands two types of pmaps: PT_X86 and PT_EPT. The PT_X86 pmap type is used by the vmspace associated with the host kernel as well as user processes executing on the host. The PT_EPT pmap is used by the vmspace associated with a bhyve guest. Page Table Entries: The EPT page table entries as mostly similar in functionality to regular page table entries although there are some differences in terms of what bits are used to express that functionality. For e.g. the dirty bit is represented by bit 9 in the nested PTE as opposed to bit 6 in the regular x86 PTE. Therefore the bitmask representing the dirty bit is now computed at runtime based on the type of the pmap. Thus PG_M that was previously a macro now becomes a local variable that is initialized at runtime using 'pmap_modified_bit(pmap)'. An additional wrinkle associated with EPT mappings is that older Intel processors don't have hardware support for tracking accessed/dirty bits in the PTE. This means that the amd64/pmap code needs to emulate these bits to provide proper accounting to the VM subsystem. This is achieved by using the following mapping for EPT entries that need emulation of A/D bits: Bit Position Interpreted By PG_V 52 software (accessed bit emulation handler) PG_RW 53 software (dirty bit emulation handler) PG_A 0 hardware (aka EPT_PG_RD) PG_M 1 hardware (aka EPT_PG_WR) The idea to use the mapping listed above for A/D bit emulation came from Alan Cox (alc@). The final difference with respect to x86 PTEs is that some EPT implementations do not support superpage mappings. This is recorded in the 'pm_flags' field of the pmap. TLB invalidation: The amd64/pmap code has a number of ways to do invalidation of mappings that may be cached in the TLB: single page, multiple pages in a range or the entire TLB. All of these funnel into a single EPT invalidation routine called 'pmap_invalidate_ept()'. This routine bumps up the EPT generation number and sends an IPI to the host cpus that are executing the guest's vcpus. On a subsequent entry into the guest it will detect that the EPT has changed and invalidate the mappings from the TLB. Guest memory access: Since the guest memory is no longer wired we need to hold the host physical page that backs the guest physical page before we can access it. The helper functions 'vm_gpa_hold()/vm_gpa_release()' are available for this purpose. PCI passthru: Guest's with PCI passthru devices will wire the entire guest physical address space. The MMIO BAR associated with the passthru device is backed by a vm_object of type OBJT_SG. An IOMMU domain is created only for guest's that have one or more PCI passthru devices attached to them. Limitations: There isn't a way to map a guest physical page without execute permissions. This is because the amd64/pmap code interprets the guest physical mappings as user mappings since they are numerically below VM_MAXUSER_ADDRESS. Since PG_U shares the same bit position as EPT_PG_EXECUTE all guest mappings become automatically executable. Thanks to Alan Cox and Konstantin Belousov for their rigorous code reviews as well as their support and encouragement. Thanks for John Baldwin for reviewing the use of OBJT_SG as the backing object for pci passthru mmio regions. Special thanks to Peter Holm for testing the patch on short notice. Approved by: re Discussed with: grehan Reviewed by: alc, kib Tested by: pho
2013-10-05 21:22:35 +00:00
VCPU_SLEEPING,
};
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);
static int __inline
vcpu_is_running(struct vm *vm, int vcpu, int *hostcpu)
{
return (vcpu_get_state(vm, vcpu, hostcpu) == VCPU_RUNNING);
}
void *vcpu_stats(struct vm *vm, int vcpu);
void vcpu_notify_event(struct vm *vm, int vcpuid, bool lapic_intr);
Merge projects/bhyve_npt_pmap into head. Make the amd64/pmap code aware of nested page table mappings used by bhyve guests. This allows bhyve to associate each guest with its own vmspace and deal with nested page faults in the context of that vmspace. This also enables features like accessed/dirty bit tracking, swapping to disk and transparent superpage promotions of guest memory. Guest vmspace: Each bhyve guest has a unique vmspace to represent the physical memory allocated to the guest. Each memory segment allocated by the guest is mapped into the guest's address space via the 'vmspace->vm_map' and is backed by an object of type OBJT_DEFAULT. pmap types: The amd64/pmap now understands two types of pmaps: PT_X86 and PT_EPT. The PT_X86 pmap type is used by the vmspace associated with the host kernel as well as user processes executing on the host. The PT_EPT pmap is used by the vmspace associated with a bhyve guest. Page Table Entries: The EPT page table entries as mostly similar in functionality to regular page table entries although there are some differences in terms of what bits are used to express that functionality. For e.g. the dirty bit is represented by bit 9 in the nested PTE as opposed to bit 6 in the regular x86 PTE. Therefore the bitmask representing the dirty bit is now computed at runtime based on the type of the pmap. Thus PG_M that was previously a macro now becomes a local variable that is initialized at runtime using 'pmap_modified_bit(pmap)'. An additional wrinkle associated with EPT mappings is that older Intel processors don't have hardware support for tracking accessed/dirty bits in the PTE. This means that the amd64/pmap code needs to emulate these bits to provide proper accounting to the VM subsystem. This is achieved by using the following mapping for EPT entries that need emulation of A/D bits: Bit Position Interpreted By PG_V 52 software (accessed bit emulation handler) PG_RW 53 software (dirty bit emulation handler) PG_A 0 hardware (aka EPT_PG_RD) PG_M 1 hardware (aka EPT_PG_WR) The idea to use the mapping listed above for A/D bit emulation came from Alan Cox (alc@). The final difference with respect to x86 PTEs is that some EPT implementations do not support superpage mappings. This is recorded in the 'pm_flags' field of the pmap. TLB invalidation: The amd64/pmap code has a number of ways to do invalidation of mappings that may be cached in the TLB: single page, multiple pages in a range or the entire TLB. All of these funnel into a single EPT invalidation routine called 'pmap_invalidate_ept()'. This routine bumps up the EPT generation number and sends an IPI to the host cpus that are executing the guest's vcpus. On a subsequent entry into the guest it will detect that the EPT has changed and invalidate the mappings from the TLB. Guest memory access: Since the guest memory is no longer wired we need to hold the host physical page that backs the guest physical page before we can access it. The helper functions 'vm_gpa_hold()/vm_gpa_release()' are available for this purpose. PCI passthru: Guest's with PCI passthru devices will wire the entire guest physical address space. The MMIO BAR associated with the passthru device is backed by a vm_object of type OBJT_SG. An IOMMU domain is created only for guest's that have one or more PCI passthru devices attached to them. Limitations: There isn't a way to map a guest physical page without execute permissions. This is because the amd64/pmap code interprets the guest physical mappings as user mappings since they are numerically below VM_MAXUSER_ADDRESS. Since PG_U shares the same bit position as EPT_PG_EXECUTE all guest mappings become automatically executable. Thanks to Alan Cox and Konstantin Belousov for their rigorous code reviews as well as their support and encouragement. Thanks for John Baldwin for reviewing the use of OBJT_SG as the backing object for pci passthru mmio regions. Special thanks to Peter Holm for testing the patch on short notice. Approved by: re Discussed with: grehan Reviewed by: alc, kib Tested by: pho
2013-10-05 21:22:35 +00:00
struct vmspace *vm_get_vmspace(struct vm *vm);
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);
struct vatpic *vm_atpic(struct vm *vm);
struct vatpit *vm_atpit(struct vm *vm);
/*
* Inject exception 'vme' 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
* this function directly because they enforce the trap-like or fault-like
* behavior of an exception.
*
* 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);
/*
* Returns 0 if there is no exception pending for this vcpu. Returns 1 if an
* exception is pending and also updates 'vme'. The pending exception is
* cleared when this function returns.
*
* This function should only be called in the context of the thread that is
* executing this vcpu.
*/
int vm_exception_pending(struct vm *vm, int vcpuid, struct vm_exception *vme);
void vm_inject_gp(struct vm *vm, int vcpuid); /* general protection fault */
void vm_inject_ud(struct vm *vm, int vcpuid); /* undefined instruction fault */
void vm_inject_pf(struct vm *vm, int vcpuid, int error_code, uint64_t cr2);
enum vm_reg_name vm_segment_name(int seg_encoding);
#endif /* KERNEL */
#include <machine/vmm_instruction_emul.h>
#define VM_MAXCPU 16 /* maximum virtual cpus */
/*
* Identifiers for architecturally defined registers.
*/
enum vm_reg_name {
VM_REG_GUEST_RAX,
VM_REG_GUEST_RBX,
VM_REG_GUEST_RCX,
VM_REG_GUEST_RDX,
VM_REG_GUEST_RSI,
VM_REG_GUEST_RDI,
VM_REG_GUEST_RBP,
VM_REG_GUEST_R8,
VM_REG_GUEST_R9,
VM_REG_GUEST_R10,
VM_REG_GUEST_R11,
VM_REG_GUEST_R12,
VM_REG_GUEST_R13,
VM_REG_GUEST_R14,
VM_REG_GUEST_R15,
VM_REG_GUEST_CR0,
VM_REG_GUEST_CR3,
VM_REG_GUEST_CR4,
VM_REG_GUEST_DR7,
VM_REG_GUEST_RSP,
VM_REG_GUEST_RIP,
VM_REG_GUEST_RFLAGS,
VM_REG_GUEST_ES,
VM_REG_GUEST_CS,
VM_REG_GUEST_SS,
VM_REG_GUEST_DS,
VM_REG_GUEST_FS,
VM_REG_GUEST_GS,
VM_REG_GUEST_LDTR,
VM_REG_GUEST_TR,
VM_REG_GUEST_IDTR,
VM_REG_GUEST_GDTR,
VM_REG_GUEST_EFER,
VM_REG_GUEST_CR2,
VM_REG_LAST
};
/*
* Identifiers for optional vmm capabilities
*/
enum vm_cap_type {
VM_CAP_HALT_EXIT,
VM_CAP_MTRAP_EXIT,
VM_CAP_PAUSE_EXIT,
VM_CAP_UNRESTRICTED_GUEST,
VM_CAP_ENABLE_INVPCID,
VM_CAP_MAX
};
enum x2apic_state {
X2APIC_DISABLED,
X2APIC_ENABLED,
X2APIC_STATE_LAST
};
2014-05-15 14:16:55 +00:00
enum vm_intr_trigger {
EDGE_TRIGGER,
LEVEL_TRIGGER
};
/*
* The 'access' field has the format specified in Table 21-2 of the Intel
* Architecture Manual vol 3b.
*
* XXX The contents of the 'access' field are architecturally defined except
* bit 16 - Segment Unusable.
*/
struct seg_desc {
uint64_t base;
uint32_t limit;
uint32_t access;
};
enum vm_exitcode {
VM_EXITCODE_INOUT,
VM_EXITCODE_VMX,
VM_EXITCODE_BOGUS,
VM_EXITCODE_RDMSR,
VM_EXITCODE_WRMSR,
VM_EXITCODE_HLT,
VM_EXITCODE_MTRAP,
VM_EXITCODE_PAUSE,
VM_EXITCODE_PAGING,
Merge projects/bhyve_npt_pmap into head. Make the amd64/pmap code aware of nested page table mappings used by bhyve guests. This allows bhyve to associate each guest with its own vmspace and deal with nested page faults in the context of that vmspace. This also enables features like accessed/dirty bit tracking, swapping to disk and transparent superpage promotions of guest memory. Guest vmspace: Each bhyve guest has a unique vmspace to represent the physical memory allocated to the guest. Each memory segment allocated by the guest is mapped into the guest's address space via the 'vmspace->vm_map' and is backed by an object of type OBJT_DEFAULT. pmap types: The amd64/pmap now understands two types of pmaps: PT_X86 and PT_EPT. The PT_X86 pmap type is used by the vmspace associated with the host kernel as well as user processes executing on the host. The PT_EPT pmap is used by the vmspace associated with a bhyve guest. Page Table Entries: The EPT page table entries as mostly similar in functionality to regular page table entries although there are some differences in terms of what bits are used to express that functionality. For e.g. the dirty bit is represented by bit 9 in the nested PTE as opposed to bit 6 in the regular x86 PTE. Therefore the bitmask representing the dirty bit is now computed at runtime based on the type of the pmap. Thus PG_M that was previously a macro now becomes a local variable that is initialized at runtime using 'pmap_modified_bit(pmap)'. An additional wrinkle associated with EPT mappings is that older Intel processors don't have hardware support for tracking accessed/dirty bits in the PTE. This means that the amd64/pmap code needs to emulate these bits to provide proper accounting to the VM subsystem. This is achieved by using the following mapping for EPT entries that need emulation of A/D bits: Bit Position Interpreted By PG_V 52 software (accessed bit emulation handler) PG_RW 53 software (dirty bit emulation handler) PG_A 0 hardware (aka EPT_PG_RD) PG_M 1 hardware (aka EPT_PG_WR) The idea to use the mapping listed above for A/D bit emulation came from Alan Cox (alc@). The final difference with respect to x86 PTEs is that some EPT implementations do not support superpage mappings. This is recorded in the 'pm_flags' field of the pmap. TLB invalidation: The amd64/pmap code has a number of ways to do invalidation of mappings that may be cached in the TLB: single page, multiple pages in a range or the entire TLB. All of these funnel into a single EPT invalidation routine called 'pmap_invalidate_ept()'. This routine bumps up the EPT generation number and sends an IPI to the host cpus that are executing the guest's vcpus. On a subsequent entry into the guest it will detect that the EPT has changed and invalidate the mappings from the TLB. Guest memory access: Since the guest memory is no longer wired we need to hold the host physical page that backs the guest physical page before we can access it. The helper functions 'vm_gpa_hold()/vm_gpa_release()' are available for this purpose. PCI passthru: Guest's with PCI passthru devices will wire the entire guest physical address space. The MMIO BAR associated with the passthru device is backed by a vm_object of type OBJT_SG. An IOMMU domain is created only for guest's that have one or more PCI passthru devices attached to them. Limitations: There isn't a way to map a guest physical page without execute permissions. This is because the amd64/pmap code interprets the guest physical mappings as user mappings since they are numerically below VM_MAXUSER_ADDRESS. Since PG_U shares the same bit position as EPT_PG_EXECUTE all guest mappings become automatically executable. Thanks to Alan Cox and Konstantin Belousov for their rigorous code reviews as well as their support and encouragement. Thanks for John Baldwin for reviewing the use of OBJT_SG as the backing object for pci passthru mmio regions. Special thanks to Peter Holm for testing the patch on short notice. Approved by: re Discussed with: grehan Reviewed by: alc, kib Tested by: pho
2013-10-05 21:22:35 +00:00
VM_EXITCODE_INST_EMUL,
VM_EXITCODE_SPINUP_AP,
VM_EXITCODE_DEPRECATED1, /* used to be SPINDOWN_CPU */
VM_EXITCODE_RENDEZVOUS,
VM_EXITCODE_IOAPIC_EOI,
VM_EXITCODE_SUSPENDED,
VM_EXITCODE_INOUT_STR,
VM_EXITCODE_MAX
};
struct vm_inout {
uint16_t bytes:3; /* 1 or 2 or 4 */
uint16_t in:1;
uint16_t string:1;
uint16_t rep:1;
uint16_t port;
uint32_t eax; /* valid for out */
};
struct vm_inout_str {
struct vm_inout inout; /* must be the first element */
enum vie_cpu_mode cpu_mode;
enum vie_paging_mode paging_mode;
uint64_t rflags;
uint64_t cr0;
uint64_t cr3;
uint64_t index;
uint64_t count; /* rep=1 (%rcx), rep=0 (1) */
int cpl;
int addrsize;
enum vm_reg_name seg_name;
struct seg_desc seg_desc;
uint64_t gla; /* may be set to VIE_INVALID_GLA */
uint64_t gpa;
};
struct vm_exit {
enum vm_exitcode exitcode;
int inst_length; /* 0 means unknown */
uint64_t rip;
union {
struct vm_inout inout;
struct vm_inout_str inout_str;
struct {
uint64_t gpa;
Merge projects/bhyve_npt_pmap into head. Make the amd64/pmap code aware of nested page table mappings used by bhyve guests. This allows bhyve to associate each guest with its own vmspace and deal with nested page faults in the context of that vmspace. This also enables features like accessed/dirty bit tracking, swapping to disk and transparent superpage promotions of guest memory. Guest vmspace: Each bhyve guest has a unique vmspace to represent the physical memory allocated to the guest. Each memory segment allocated by the guest is mapped into the guest's address space via the 'vmspace->vm_map' and is backed by an object of type OBJT_DEFAULT. pmap types: The amd64/pmap now understands two types of pmaps: PT_X86 and PT_EPT. The PT_X86 pmap type is used by the vmspace associated with the host kernel as well as user processes executing on the host. The PT_EPT pmap is used by the vmspace associated with a bhyve guest. Page Table Entries: The EPT page table entries as mostly similar in functionality to regular page table entries although there are some differences in terms of what bits are used to express that functionality. For e.g. the dirty bit is represented by bit 9 in the nested PTE as opposed to bit 6 in the regular x86 PTE. Therefore the bitmask representing the dirty bit is now computed at runtime based on the type of the pmap. Thus PG_M that was previously a macro now becomes a local variable that is initialized at runtime using 'pmap_modified_bit(pmap)'. An additional wrinkle associated with EPT mappings is that older Intel processors don't have hardware support for tracking accessed/dirty bits in the PTE. This means that the amd64/pmap code needs to emulate these bits to provide proper accounting to the VM subsystem. This is achieved by using the following mapping for EPT entries that need emulation of A/D bits: Bit Position Interpreted By PG_V 52 software (accessed bit emulation handler) PG_RW 53 software (dirty bit emulation handler) PG_A 0 hardware (aka EPT_PG_RD) PG_M 1 hardware (aka EPT_PG_WR) The idea to use the mapping listed above for A/D bit emulation came from Alan Cox (alc@). The final difference with respect to x86 PTEs is that some EPT implementations do not support superpage mappings. This is recorded in the 'pm_flags' field of the pmap. TLB invalidation: The amd64/pmap code has a number of ways to do invalidation of mappings that may be cached in the TLB: single page, multiple pages in a range or the entire TLB. All of these funnel into a single EPT invalidation routine called 'pmap_invalidate_ept()'. This routine bumps up the EPT generation number and sends an IPI to the host cpus that are executing the guest's vcpus. On a subsequent entry into the guest it will detect that the EPT has changed and invalidate the mappings from the TLB. Guest memory access: Since the guest memory is no longer wired we need to hold the host physical page that backs the guest physical page before we can access it. The helper functions 'vm_gpa_hold()/vm_gpa_release()' are available for this purpose. PCI passthru: Guest's with PCI passthru devices will wire the entire guest physical address space. The MMIO BAR associated with the passthru device is backed by a vm_object of type OBJT_SG. An IOMMU domain is created only for guest's that have one or more PCI passthru devices attached to them. Limitations: There isn't a way to map a guest physical page without execute permissions. This is because the amd64/pmap code interprets the guest physical mappings as user mappings since they are numerically below VM_MAXUSER_ADDRESS. Since PG_U shares the same bit position as EPT_PG_EXECUTE all guest mappings become automatically executable. Thanks to Alan Cox and Konstantin Belousov for their rigorous code reviews as well as their support and encouragement. Thanks for John Baldwin for reviewing the use of OBJT_SG as the backing object for pci passthru mmio regions. Special thanks to Peter Holm for testing the patch on short notice. Approved by: re Discussed with: grehan Reviewed by: alc, kib Tested by: pho
2013-10-05 21:22:35 +00:00
int fault_type;
} paging;
Merge projects/bhyve_npt_pmap into head. Make the amd64/pmap code aware of nested page table mappings used by bhyve guests. This allows bhyve to associate each guest with its own vmspace and deal with nested page faults in the context of that vmspace. This also enables features like accessed/dirty bit tracking, swapping to disk and transparent superpage promotions of guest memory. Guest vmspace: Each bhyve guest has a unique vmspace to represent the physical memory allocated to the guest. Each memory segment allocated by the guest is mapped into the guest's address space via the 'vmspace->vm_map' and is backed by an object of type OBJT_DEFAULT. pmap types: The amd64/pmap now understands two types of pmaps: PT_X86 and PT_EPT. The PT_X86 pmap type is used by the vmspace associated with the host kernel as well as user processes executing on the host. The PT_EPT pmap is used by the vmspace associated with a bhyve guest. Page Table Entries: The EPT page table entries as mostly similar in functionality to regular page table entries although there are some differences in terms of what bits are used to express that functionality. For e.g. the dirty bit is represented by bit 9 in the nested PTE as opposed to bit 6 in the regular x86 PTE. Therefore the bitmask representing the dirty bit is now computed at runtime based on the type of the pmap. Thus PG_M that was previously a macro now becomes a local variable that is initialized at runtime using 'pmap_modified_bit(pmap)'. An additional wrinkle associated with EPT mappings is that older Intel processors don't have hardware support for tracking accessed/dirty bits in the PTE. This means that the amd64/pmap code needs to emulate these bits to provide proper accounting to the VM subsystem. This is achieved by using the following mapping for EPT entries that need emulation of A/D bits: Bit Position Interpreted By PG_V 52 software (accessed bit emulation handler) PG_RW 53 software (dirty bit emulation handler) PG_A 0 hardware (aka EPT_PG_RD) PG_M 1 hardware (aka EPT_PG_WR) The idea to use the mapping listed above for A/D bit emulation came from Alan Cox (alc@). The final difference with respect to x86 PTEs is that some EPT implementations do not support superpage mappings. This is recorded in the 'pm_flags' field of the pmap. TLB invalidation: The amd64/pmap code has a number of ways to do invalidation of mappings that may be cached in the TLB: single page, multiple pages in a range or the entire TLB. All of these funnel into a single EPT invalidation routine called 'pmap_invalidate_ept()'. This routine bumps up the EPT generation number and sends an IPI to the host cpus that are executing the guest's vcpus. On a subsequent entry into the guest it will detect that the EPT has changed and invalidate the mappings from the TLB. Guest memory access: Since the guest memory is no longer wired we need to hold the host physical page that backs the guest physical page before we can access it. The helper functions 'vm_gpa_hold()/vm_gpa_release()' are available for this purpose. PCI passthru: Guest's with PCI passthru devices will wire the entire guest physical address space. The MMIO BAR associated with the passthru device is backed by a vm_object of type OBJT_SG. An IOMMU domain is created only for guest's that have one or more PCI passthru devices attached to them. Limitations: There isn't a way to map a guest physical page without execute permissions. This is because the amd64/pmap code interprets the guest physical mappings as user mappings since they are numerically below VM_MAXUSER_ADDRESS. Since PG_U shares the same bit position as EPT_PG_EXECUTE all guest mappings become automatically executable. Thanks to Alan Cox and Konstantin Belousov for their rigorous code reviews as well as their support and encouragement. Thanks for John Baldwin for reviewing the use of OBJT_SG as the backing object for pci passthru mmio regions. Special thanks to Peter Holm for testing the patch on short notice. Approved by: re Discussed with: grehan Reviewed by: alc, kib Tested by: pho
2013-10-05 21:22:35 +00:00
struct {
uint64_t gpa;
uint64_t gla;
uint64_t cr3;
Add support for FreeBSD/i386 guests under bhyve. - Similar to the hack for bootinfo32.c in userboot, define _MACHINE_ELF_WANT_32BIT in the load_elf32 file handlers in userboot. This allows userboot to load 32-bit kernels and modules. - Copy the SMAP generation code out of bootinfo64.c and into its own file so it can be shared with bootinfo32.c to pass an SMAP to the i386 kernel. - Use uint32_t instead of u_long when aligning module metadata in bootinfo32.c in userboot, as otherwise the metadata used 64-bit alignment which corrupted the layout. - Populate the basemem and extmem members of the bootinfo struct passed to 32-bit kernels. - Fix the 32-bit stack in userboot to start at the top of the stack instead of the bottom so that there is room to grow before the kernel switches to its own stack. - Push a fake return address onto the 32-bit stack in addition to the arguments normally passed to exec() in the loader. This return address is needed to convince recover_bootinfo() in the 32-bit locore code that it is being invoked from a "new" boot block. - Add a routine to libvmmapi to setup a 32-bit flat mode register state including a GDT and TSS that is able to start the i386 kernel and update bhyveload to use it when booting an i386 kernel. - Use the guest register state to determine the CPU's current instruction mode (32-bit vs 64-bit) and paging mode (flat, 32-bit, PAE, or long mode) in the instruction emulation code. Update the gla2gpa() routine used when fetching instructions to handle flat mode, 32-bit paging, and PAE paging in addition to long mode paging. Don't look for a REX prefix when the CPU is in 32-bit mode, and use the detected mode to enable the existing 32-bit mode code when decoding the mod r/m byte. Reviewed by: grehan, neel MFC after: 1 month
2014-02-05 04:39:03 +00:00
enum vie_cpu_mode cpu_mode;
enum vie_paging_mode paging_mode;
int cpl;
Merge projects/bhyve_npt_pmap into head. Make the amd64/pmap code aware of nested page table mappings used by bhyve guests. This allows bhyve to associate each guest with its own vmspace and deal with nested page faults in the context of that vmspace. This also enables features like accessed/dirty bit tracking, swapping to disk and transparent superpage promotions of guest memory. Guest vmspace: Each bhyve guest has a unique vmspace to represent the physical memory allocated to the guest. Each memory segment allocated by the guest is mapped into the guest's address space via the 'vmspace->vm_map' and is backed by an object of type OBJT_DEFAULT. pmap types: The amd64/pmap now understands two types of pmaps: PT_X86 and PT_EPT. The PT_X86 pmap type is used by the vmspace associated with the host kernel as well as user processes executing on the host. The PT_EPT pmap is used by the vmspace associated with a bhyve guest. Page Table Entries: The EPT page table entries as mostly similar in functionality to regular page table entries although there are some differences in terms of what bits are used to express that functionality. For e.g. the dirty bit is represented by bit 9 in the nested PTE as opposed to bit 6 in the regular x86 PTE. Therefore the bitmask representing the dirty bit is now computed at runtime based on the type of the pmap. Thus PG_M that was previously a macro now becomes a local variable that is initialized at runtime using 'pmap_modified_bit(pmap)'. An additional wrinkle associated with EPT mappings is that older Intel processors don't have hardware support for tracking accessed/dirty bits in the PTE. This means that the amd64/pmap code needs to emulate these bits to provide proper accounting to the VM subsystem. This is achieved by using the following mapping for EPT entries that need emulation of A/D bits: Bit Position Interpreted By PG_V 52 software (accessed bit emulation handler) PG_RW 53 software (dirty bit emulation handler) PG_A 0 hardware (aka EPT_PG_RD) PG_M 1 hardware (aka EPT_PG_WR) The idea to use the mapping listed above for A/D bit emulation came from Alan Cox (alc@). The final difference with respect to x86 PTEs is that some EPT implementations do not support superpage mappings. This is recorded in the 'pm_flags' field of the pmap. TLB invalidation: The amd64/pmap code has a number of ways to do invalidation of mappings that may be cached in the TLB: single page, multiple pages in a range or the entire TLB. All of these funnel into a single EPT invalidation routine called 'pmap_invalidate_ept()'. This routine bumps up the EPT generation number and sends an IPI to the host cpus that are executing the guest's vcpus. On a subsequent entry into the guest it will detect that the EPT has changed and invalidate the mappings from the TLB. Guest memory access: Since the guest memory is no longer wired we need to hold the host physical page that backs the guest physical page before we can access it. The helper functions 'vm_gpa_hold()/vm_gpa_release()' are available for this purpose. PCI passthru: Guest's with PCI passthru devices will wire the entire guest physical address space. The MMIO BAR associated with the passthru device is backed by a vm_object of type OBJT_SG. An IOMMU domain is created only for guest's that have one or more PCI passthru devices attached to them. Limitations: There isn't a way to map a guest physical page without execute permissions. This is because the amd64/pmap code interprets the guest physical mappings as user mappings since they are numerically below VM_MAXUSER_ADDRESS. Since PG_U shares the same bit position as EPT_PG_EXECUTE all guest mappings become automatically executable. Thanks to Alan Cox and Konstantin Belousov for their rigorous code reviews as well as their support and encouragement. Thanks for John Baldwin for reviewing the use of OBJT_SG as the backing object for pci passthru mmio regions. Special thanks to Peter Holm for testing the patch on short notice. Approved by: re Discussed with: grehan Reviewed by: alc, kib Tested by: pho
2013-10-05 21:22:35 +00:00
struct vie vie;
} inst_emul;
/*
* VMX specific payload. Used when there is no "better"
* exitcode to represent the VM-exit.
*/
struct {
int status; /* vmx inst status */
/*
* 'exit_reason' and 'exit_qualification' are valid
* only if 'status' is zero.
*/
uint32_t exit_reason;
uint64_t exit_qualification;
/*
* 'inst_error' and 'inst_type' are valid
* only if 'status' is non-zero.
*/
int inst_type;
int inst_error;
} vmx;
struct {
uint32_t code; /* ecx value */
uint64_t wval;
} msr;
struct {
int vcpu;
uint64_t rip;
} spinup_ap;
struct {
uint64_t rflags;
} hlt;
struct {
int vector;
} ioapic_eoi;
struct {
enum vm_suspend_how how;
} suspended;
} u;
};
#endif /* _VMM_H_ */