[linux-block.git] / arch / x86 / kvm / mmu.h

/* SPDX-License-Identifier: GPL-2.0 */
#ifndef __KVM_X86_MMU_H
#define __KVM_X86_MMU_H

#include <linux/kvm_host.h>
#include "kvm_cache_regs.h"

#define PT64_PT_BITS 9
#define PT64_ENT_PER_PAGE (1 << PT64_PT_BITS)
#define PT32_PT_BITS 10
#define PT32_ENT_PER_PAGE (1 << PT32_PT_BITS)

#define PT_WRITABLE_SHIFT 1
#define PT_USER_SHIFT 2

#define PT_PRESENT_MASK (1ULL << 0)
#define PT_WRITABLE_MASK (1ULL << PT_WRITABLE_SHIFT)
#define PT_USER_MASK (1ULL << PT_USER_SHIFT)
#define PT_PWT_MASK (1ULL << 3)
#define PT_PCD_MASK (1ULL << 4)
#define PT_ACCESSED_SHIFT 5
#define PT_ACCESSED_MASK (1ULL << PT_ACCESSED_SHIFT)
#define PT_DIRTY_SHIFT 6
#define PT_DIRTY_MASK (1ULL << PT_DIRTY_SHIFT)
#define PT_PAGE_SIZE_SHIFT 7
#define PT_PAGE_SIZE_MASK (1ULL << PT_PAGE_SIZE_SHIFT)
#define PT_PAT_MASK (1ULL << 7)
#define PT_GLOBAL_MASK (1ULL << 8)
#define PT64_NX_SHIFT 63
#define PT64_NX_MASK (1ULL << PT64_NX_SHIFT)

#define PT_PAT_SHIFT 7
#define PT_DIR_PAT_SHIFT 12
#define PT_DIR_PAT_MASK (1ULL << PT_DIR_PAT_SHIFT)

#define PT32_DIR_PSE36_SIZE 4
#define PT32_DIR_PSE36_SHIFT 13
#define PT32_DIR_PSE36_MASK \
	(((1ULL << PT32_DIR_PSE36_SIZE) - 1) << PT32_DIR_PSE36_SHIFT)

#define PT64_ROOT_5LEVEL 5
#define PT64_ROOT_4LEVEL 4
#define PT32_ROOT_LEVEL 2
#define PT32E_ROOT_LEVEL 3

#define PT_PDPE_LEVEL 3
#define PT_DIRECTORY_LEVEL 2
#define PT_PAGE_TABLE_LEVEL 1
#define PT_MAX_HUGEPAGE_LEVEL (PT_PAGE_TABLE_LEVEL + KVM_NR_PAGE_SIZES - 1)

static inline u64 rsvd_bits(int s, int e)
{
	if (e < s)
		return 0;

	return ((1ULL << (e - s + 1)) - 1) << s;
}

void kvm_mmu_set_mmio_spte_mask(u64 mmio_mask, u64 mmio_value);

void
reset_shadow_zero_bits_mask(struct kvm_vcpu *vcpu, struct kvm_mmu *context);

void kvm_init_shadow_mmu(struct kvm_vcpu *vcpu);
void kvm_init_shadow_ept_mmu(struct kvm_vcpu *vcpu, bool execonly,
			     bool accessed_dirty);
bool kvm_can_do_async_pf(struct kvm_vcpu *vcpu);
int kvm_handle_page_fault(struct kvm_vcpu *vcpu, u64 error_code,
				u64 fault_address, char *insn, int insn_len,
				bool need_unprotect);

static inline unsigned int kvm_mmu_available_pages(struct kvm *kvm)
{
	if (kvm->arch.n_max_mmu_pages > kvm->arch.n_used_mmu_pages)
		return kvm->arch.n_max_mmu_pages -
			kvm->arch.n_used_mmu_pages;

	return 0;
}

static inline int kvm_mmu_reload(struct kvm_vcpu *vcpu)
{
	if (likely(vcpu->arch.mmu.root_hpa != INVALID_PAGE))
		return 0;

	return kvm_mmu_load(vcpu);
}

/*
 * Currently, we have two sorts of write-protection, a) the first one
 * write-protects guest page to sync the guest modification, b) another one is
 * used to sync dirty bitmap when we do KVM_GET_DIRTY_LOG. The differences
 * between these two sorts are:
 * 1) the first case clears SPTE_MMU_WRITEABLE bit.
 * 2) the first case requires flushing tlb immediately avoiding corrupting
 *    shadow page table between all vcpus so it should be in the protection of
 *    mmu-lock. And the another case does not need to flush tlb until returning
 *    the dirty bitmap to userspace since it only write-protects the page
 *    logged in the bitmap, that means the page in the dirty bitmap is not
 *    missed, so it can flush tlb out of mmu-lock.
 *
 * So, there is the problem: the first case can meet the corrupted tlb caused
 * by another case which write-protects pages but without flush tlb
 * immediately. In order to making the first case be aware this problem we let
 * it flush tlb if we try to write-protect a spte whose SPTE_MMU_WRITEABLE bit
 * is set, it works since another case never touches SPTE_MMU_WRITEABLE bit.
 *
 * Anyway, whenever a spte is updated (only permission and status bits are
 * changed) we need to check whether the spte with SPTE_MMU_WRITEABLE becomes
 * readonly, if that happens, we need to flush tlb. Fortunately,
 * mmu_spte_update() has already handled it perfectly.
 *
 * The rules to use SPTE_MMU_WRITEABLE and PT_WRITABLE_MASK:
 * - if we want to see if it has writable tlb entry or if the spte can be
 *   writable on the mmu mapping, check SPTE_MMU_WRITEABLE, this is the most
 *   case, otherwise
 * - if we fix page fault on the spte or do write-protection by dirty logging,
 *   check PT_WRITABLE_MASK.
 *
 * TODO: introduce APIs to split these two cases.
 */
static inline int is_writable_pte(unsigned long pte)
{
	return pte & PT_WRITABLE_MASK;
}

static inline bool is_write_protection(struct kvm_vcpu *vcpu)
{
	return kvm_read_cr0_bits(vcpu, X86_CR0_WP);
}

/*
 * Check if a given access (described through the I/D, W/R and U/S bits of a
 * page fault error code pfec) causes a permission fault with the given PTE
 * access rights (in ACC_* format).
 *
 * Return zero if the access does not fault; return the page fault error code
 * if the access faults.
 */
static inline u8 permission_fault(struct kvm_vcpu *vcpu, struct kvm_mmu *mmu,
				  unsigned pte_access, unsigned pte_pkey,
				  unsigned pfec)
{
	int cpl = kvm_x86_ops->get_cpl(vcpu);
	unsigned long rflags = kvm_x86_ops->get_rflags(vcpu);

	/*
	 * If CPL < 3, SMAP prevention are disabled if EFLAGS.AC = 1.
	 *
	 * If CPL = 3, SMAP applies to all supervisor-mode data accesses
	 * (these are implicit supervisor accesses) regardless of the value
	 * of EFLAGS.AC.
	 *
	 * This computes (cpl < 3) && (rflags & X86_EFLAGS_AC), leaving
	 * the result in X86_EFLAGS_AC. We then insert it in place of
	 * the PFERR_RSVD_MASK bit; this bit will always be zero in pfec,
	 * but it will be one in index if SMAP checks are being overridden.
	 * It is important to keep this branchless.
	 */
	unsigned long smap = (cpl - 3) & (rflags & X86_EFLAGS_AC);
	int index = (pfec >> 1) +
		    (smap >> (X86_EFLAGS_AC_BIT - PFERR_RSVD_BIT + 1));
	bool fault = (mmu->permissions[index] >> pte_access) & 1;
	u32 errcode = PFERR_PRESENT_MASK;

	WARN_ON(pfec & (PFERR_PK_MASK | PFERR_RSVD_MASK));
	if (unlikely(mmu->pkru_mask)) {
		u32 pkru_bits, offset;

		/*
		* PKRU defines 32 bits, there are 16 domains and 2
		* attribute bits per domain in pkru.  pte_pkey is the
		* index of the protection domain, so pte_pkey * 2 is
		* is the index of the first bit for the domain.
		*/
		pkru_bits = (vcpu->arch.pkru >> (pte_pkey * 2)) & 3;

		/* clear present bit, replace PFEC.RSVD with ACC_USER_MASK. */
		offset = (pfec & ~1) +
			((pte_access & PT_USER_MASK) << (PFERR_RSVD_BIT - PT_USER_SHIFT));

		pkru_bits &= mmu->pkru_mask >> offset;
		errcode |= -pkru_bits & PFERR_PK_MASK;
		fault |= (pkru_bits != 0);
	}

	return -(u32)fault & errcode;
}

void kvm_mmu_invalidate_zap_all_pages(struct kvm *kvm);
void kvm_zap_gfn_range(struct kvm *kvm, gfn_t gfn_start, gfn_t gfn_end);

void kvm_mmu_gfn_disallow_lpage(struct kvm_memory_slot *slot, gfn_t gfn);
void kvm_mmu_gfn_allow_lpage(struct kvm_memory_slot *slot, gfn_t gfn);
bool kvm_mmu_slot_gfn_write_protect(struct kvm *kvm,
				    struct kvm_memory_slot *slot, u64 gfn);
int kvm_arch_write_log_dirty(struct kvm_vcpu *vcpu);
#endif
Commit	Line	Data
b2441318	1	/* SPDX-License-Identifier: GPL-2.0 */
1d737c8a ZX	2	#ifndef __KVM_X86_MMU_H
	3	#define __KVM_X86_MMU_H
	4
edf88417	5	#include <linux/kvm_host.h>
fc78f519	6	#include "kvm_cache_regs.h"
1d737c8a	7
8c6d6adc SY	8	#define PT64_PT_BITS 9
	9	#define PT64_ENT_PER_PAGE (1 << PT64_PT_BITS)
	10	#define PT32_PT_BITS 10
	11	#define PT32_ENT_PER_PAGE (1 << PT32_PT_BITS)
	12
	13	#define PT_WRITABLE_SHIFT 1
be94f6b7	14	#define PT_USER_SHIFT 2
8c6d6adc SY	15
	16	#define PT_PRESENT_MASK (1ULL << 0)
	17	#define PT_WRITABLE_MASK (1ULL << PT_WRITABLE_SHIFT)
be94f6b7	18	#define PT_USER_MASK (1ULL << PT_USER_SHIFT)
8c6d6adc SY	19	#define PT_PWT_MASK (1ULL << 3)
8c6d6adc SY	20	#define PT_PCD_MASK (1ULL << 4)
1b7fcd32 AK	21	#define PT_ACCESSED_SHIFT 5
1b7fcd32 AK	22	#define PT_ACCESSED_MASK (1ULL << PT_ACCESSED_SHIFT)
8ea667f2 AK	23	#define PT_DIRTY_SHIFT 6
8ea667f2 AK	24	#define PT_DIRTY_MASK (1ULL << PT_DIRTY_SHIFT)
6fd01b71 AK	25	#define PT_PAGE_SIZE_SHIFT 7
6fd01b71 AK	26	#define PT_PAGE_SIZE_MASK (1ULL << PT_PAGE_SIZE_SHIFT)
8c6d6adc SY	27	#define PT_PAT_MASK (1ULL << 7)
	28	#define PT_GLOBAL_MASK (1ULL << 8)
	29	#define PT64_NX_SHIFT 63
	30	#define PT64_NX_MASK (1ULL << PT64_NX_SHIFT)
	31
	32	#define PT_PAT_SHIFT 7
	33	#define PT_DIR_PAT_SHIFT 12
	34	#define PT_DIR_PAT_MASK (1ULL << PT_DIR_PAT_SHIFT)
	35
	36	#define PT32_DIR_PSE36_SIZE 4
	37	#define PT32_DIR_PSE36_SHIFT 13
	38	#define PT32_DIR_PSE36_MASK \
	39	(((1ULL << PT32_DIR_PSE36_SIZE) - 1) << PT32_DIR_PSE36_SHIFT)
	40
855feb67	41	#define PT64_ROOT_5LEVEL 5
2a7266a8	42	#define PT64_ROOT_4LEVEL 4
8c6d6adc SY	43	#define PT32_ROOT_LEVEL 2
	44	#define PT32E_ROOT_LEVEL 3
	45
c9c54174 SY	46	#define PT_PDPE_LEVEL 3
	47	#define PT_DIRECTORY_LEVEL 2
	48	#define PT_PAGE_TABLE_LEVEL 1
8a3d08f1	49	#define PT_MAX_HUGEPAGE_LEVEL (PT_PAGE_TABLE_LEVEL + KVM_NR_PAGE_SIZES - 1)
c9c54174	50
d1431483 TC	51	static inline u64 rsvd_bits(int s, int e)
d1431483 TC	52	{
d1cd3ce9 YZ	53	if (e < s)
	54	return 0;
	55
d1431483 TC	56	return ((1ULL << (e - s + 1)) - 1) << s;
	57	}
	58
dcdca5fe	59	void kvm_mmu_set_mmio_spte_mask(u64 mmio_mask, u64 mmio_value);
b37fbea6	60
c258b62b XG	61	void
	62	reset_shadow_zero_bits_mask(struct kvm_vcpu vcpu, struct kvm_mmu context);
	63
ad896af0	64	void kvm_init_shadow_mmu(struct kvm_vcpu *vcpu);
ae1e2d10 PB	65	void kvm_init_shadow_ept_mmu(struct kvm_vcpu *vcpu, bool execonly,
ae1e2d10 PB	66	bool accessed_dirty);
9bc1f09f	67	bool kvm_can_do_async_pf(struct kvm_vcpu *vcpu);
1261bfa3 WL	68	int kvm_handle_page_fault(struct kvm_vcpu *vcpu, u64 error_code,
	69	u64 fault_address, char *insn, int insn_len,
	70	bool need_unprotect);
94d8b056	71
e0df7b9f DH	72	static inline unsigned int kvm_mmu_available_pages(struct kvm *kvm)
e0df7b9f DH	73	{
5d218814 MT	74	if (kvm->arch.n_max_mmu_pages > kvm->arch.n_used_mmu_pages)
	75	return kvm->arch.n_max_mmu_pages -
	76	kvm->arch.n_used_mmu_pages;
	77
	78	return 0;
e0df7b9f DH	79	}
e0df7b9f DH	80
1d737c8a ZX	81	static inline int kvm_mmu_reload(struct kvm_vcpu *vcpu)
	82	{
	83	if (likely(vcpu->arch.mmu.root_hpa != INVALID_PAGE))
	84	return 0;
	85
	86	return kvm_mmu_load(vcpu);
	87	}
	88
198c74f4 XG	89	/*
	90	* Currently, we have two sorts of write-protection, a) the first one
	91	* write-protects guest page to sync the guest modification, b) another one is
	92	* used to sync dirty bitmap when we do KVM_GET_DIRTY_LOG. The differences
	93	* between these two sorts are:
	94	* 1) the first case clears SPTE_MMU_WRITEABLE bit.
	95	* 2) the first case requires flushing tlb immediately avoiding corrupting
	96	* shadow page table between all vcpus so it should be in the protection of
	97	* mmu-lock. And the another case does not need to flush tlb until returning
	98	* the dirty bitmap to userspace since it only write-protects the page
	99	* logged in the bitmap, that means the page in the dirty bitmap is not
	100	* missed, so it can flush tlb out of mmu-lock.
	101	*
	102	* So, there is the problem: the first case can meet the corrupted tlb caused
	103	* by another case which write-protects pages but without flush tlb
	104	* immediately. In order to making the first case be aware this problem we let
	105	* it flush tlb if we try to write-protect a spte whose SPTE_MMU_WRITEABLE bit
	106	* is set, it works since another case never touches SPTE_MMU_WRITEABLE bit.
	107	*
	108	* Anyway, whenever a spte is updated (only permission and status bits are
	109	* changed) we need to check whether the spte with SPTE_MMU_WRITEABLE becomes
	110	* readonly, if that happens, we need to flush tlb. Fortunately,
	111	* mmu_spte_update() has already handled it perfectly.
	112	*
	113	* The rules to use SPTE_MMU_WRITEABLE and PT_WRITABLE_MASK:
	114	* - if we want to see if it has writable tlb entry or if the spte can be
	115	* writable on the mmu mapping, check SPTE_MMU_WRITEABLE, this is the most
	116	* case, otherwise
	117	* - if we fix page fault on the spte or do write-protection by dirty logging,
	118	* check PT_WRITABLE_MASK.
	119	*
	120	* TODO: introduce APIs to split these two cases.
	121	*/
bebb106a XG	122	static inline int is_writable_pte(unsigned long pte)
	123	{
	124	return pte & PT_WRITABLE_MASK;
	125	}
	126
	127	static inline bool is_write_protection(struct kvm_vcpu *vcpu)
	128	{
	129	return kvm_read_cr0_bits(vcpu, X86_CR0_WP);
	130	}
	131
97d64b78	132	/*
f13577e8 PB	133	* Check if a given access (described through the I/D, W/R and U/S bits of a
	134	* page fault error code pfec) causes a permission fault with the given PTE
	135	* access rights (in ACC_* format).
	136	*
	137	* Return zero if the access does not fault; return the page fault error code
	138	* if the access faults.
97d64b78	139	*/
f13577e8	140	static inline u8 permission_fault(struct kvm_vcpu vcpu, struct kvm_mmu mmu,
be94f6b7 HH	141	unsigned pte_access, unsigned pte_pkey,
be94f6b7 HH	142	unsigned pfec)
bebb106a	143	{
97ec8c06 FW	144	int cpl = kvm_x86_ops->get_cpl(vcpu);
	145	unsigned long rflags = kvm_x86_ops->get_rflags(vcpu);
	146
	147	/*
	148	* If CPL < 3, SMAP prevention are disabled if EFLAGS.AC = 1.
	149	*
	150	* If CPL = 3, SMAP applies to all supervisor-mode data accesses
	151	* (these are implicit supervisor accesses) regardless of the value
	152	* of EFLAGS.AC.
	153	*
	154	* This computes (cpl < 3) && (rflags & X86_EFLAGS_AC), leaving
	155	* the result in X86_EFLAGS_AC. We then insert it in place of
	156	* the PFERR_RSVD_MASK bit; this bit will always be zero in pfec,
	157	* but it will be one in index if SMAP checks are being overridden.
	158	* It is important to keep this branchless.
	159	*/
	160	unsigned long smap = (cpl - 3) & (rflags & X86_EFLAGS_AC);
	161	int index = (pfec >> 1) +
	162	(smap >> (X86_EFLAGS_AC_BIT - PFERR_RSVD_BIT + 1));
be94f6b7	163	bool fault = (mmu->permissions[index] >> pte_access) & 1;
7a98205d	164	u32 errcode = PFERR_PRESENT_MASK;
97ec8c06	165
be94f6b7	166	WARN_ON(pfec & (PFERR_PK_MASK \| PFERR_RSVD_MASK));
be94f6b7 HH	167	if (unlikely(mmu->pkru_mask)) {
	168	u32 pkru_bits, offset;
	169
	170	/*
	171	* PKRU defines 32 bits, there are 16 domains and 2
	172	* attribute bits per domain in pkru. pte_pkey is the
	173	* index of the protection domain, so pte_pkey * 2 is
	174	* is the index of the first bit for the domain.
	175	*/
b9dd21e1	176	pkru_bits = (vcpu->arch.pkru >> (pte_pkey * 2)) & 3;
be94f6b7 HH	177
be94f6b7 HH	178	/* clear present bit, replace PFEC.RSVD with ACC_USER_MASK. */
7a98205d	179	offset = (pfec & ~1) +
be94f6b7 HH	180	((pte_access & PT_USER_MASK) << (PFERR_RSVD_BIT - PT_USER_SHIFT));
	181
	182	pkru_bits &= mmu->pkru_mask >> offset;
7a98205d	183	errcode \|= -pkru_bits & PFERR_PK_MASK;
be94f6b7 HH	184	fault \|= (pkru_bits != 0);
	185	}
	186
7a98205d	187	return -(u32)fault & errcode;
bebb106a	188	}
97d64b78	189
5304b8d3	190	void kvm_mmu_invalidate_zap_all_pages(struct kvm *kvm);
efdfe536	191	void kvm_zap_gfn_range(struct kvm *kvm, gfn_t gfn_start, gfn_t gfn_end);
547ffaed XG	192
	193	void kvm_mmu_gfn_disallow_lpage(struct kvm_memory_slot *slot, gfn_t gfn);
	194	void kvm_mmu_gfn_allow_lpage(struct kvm_memory_slot *slot, gfn_t gfn);
aeecee2e XG	195	bool kvm_mmu_slot_gfn_write_protect(struct kvm *kvm,
aeecee2e XG	196	struct kvm_memory_slot *slot, u64 gfn);
bab4165e	197	int kvm_arch_write_log_dirty(struct kvm_vcpu *vcpu);
1d737c8a	198	#endif