linux/arch/x86/kvm/mmu/spte.h
David Matlack a3fe5dbda0 KVM: x86/mmu: Split huge pages mapped by the TDP MMU when dirty logging is enabled
When dirty logging is enabled without initially-all-set, try to split
all huge pages in the memslot down to 4KB pages so that vCPUs do not
have to take expensive write-protection faults to split huge pages.

Eager page splitting is best-effort only. This commit only adds the
support for the TDP MMU, and even there splitting may fail due to out
of memory conditions. Failures to split a huge page is fine from a
correctness standpoint because KVM will always follow up splitting by
write-protecting any remaining huge pages.

Eager page splitting moves the cost of splitting huge pages off of the
vCPU threads and onto the thread enabling dirty logging on the memslot.
This is useful because:

 1. Splitting on the vCPU thread interrupts vCPUs execution and is
    disruptive to customers whereas splitting on VM ioctl threads can
    run in parallel with vCPU execution.

 2. Splitting all huge pages at once is more efficient because it does
    not require performing VM-exit handling or walking the page table for
    every 4KiB page in the memslot, and greatly reduces the amount of
    contention on the mmu_lock.

For example, when running dirty_log_perf_test with 96 virtual CPUs, 1GiB
per vCPU, and 1GiB HugeTLB memory, the time it takes vCPUs to write to
all of their memory after dirty logging is enabled decreased by 95% from
2.94s to 0.14s.

Eager Page Splitting is over 100x more efficient than the current
implementation of splitting on fault under the read lock. For example,
taking the same workload as above, Eager Page Splitting reduced the CPU
required to split all huge pages from ~270 CPU-seconds ((2.94s - 0.14s)
* 96 vCPU threads) to only 1.55 CPU-seconds.

Eager page splitting does increase the amount of time it takes to enable
dirty logging since it has split all huge pages. For example, the time
it took to enable dirty logging in the 96GiB region of the
aforementioned test increased from 0.001s to 1.55s.

Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: David Matlack <dmatlack@google.com>
Message-Id: <20220119230739.2234394-16-dmatlack@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2022-02-10 13:50:42 -05:00

442 lines
16 KiB
C

// SPDX-License-Identifier: GPL-2.0-only
#ifndef KVM_X86_MMU_SPTE_H
#define KVM_X86_MMU_SPTE_H
#include "mmu_internal.h"
/*
* A MMU present SPTE is backed by actual memory and may or may not be present
* in hardware. E.g. MMIO SPTEs are not considered present. Use bit 11, as it
* is ignored by all flavors of SPTEs and checking a low bit often generates
* better code than for a high bit, e.g. 56+. MMU present checks are pervasive
* enough that the improved code generation is noticeable in KVM's footprint.
*/
#define SPTE_MMU_PRESENT_MASK BIT_ULL(11)
/*
* TDP SPTES (more specifically, EPT SPTEs) may not have A/D bits, and may also
* be restricted to using write-protection (for L2 when CPU dirty logging, i.e.
* PML, is enabled). Use bits 52 and 53 to hold the type of A/D tracking that
* is must be employed for a given TDP SPTE.
*
* Note, the "enabled" mask must be '0', as bits 62:52 are _reserved_ for PAE
* paging, including NPT PAE. This scheme works because legacy shadow paging
* is guaranteed to have A/D bits and write-protection is forced only for
* TDP with CPU dirty logging (PML). If NPT ever gains PML-like support, it
* must be restricted to 64-bit KVM.
*/
#define SPTE_TDP_AD_SHIFT 52
#define SPTE_TDP_AD_MASK (3ULL << SPTE_TDP_AD_SHIFT)
#define SPTE_TDP_AD_ENABLED_MASK (0ULL << SPTE_TDP_AD_SHIFT)
#define SPTE_TDP_AD_DISABLED_MASK (1ULL << SPTE_TDP_AD_SHIFT)
#define SPTE_TDP_AD_WRPROT_ONLY_MASK (2ULL << SPTE_TDP_AD_SHIFT)
static_assert(SPTE_TDP_AD_ENABLED_MASK == 0);
#ifdef CONFIG_DYNAMIC_PHYSICAL_MASK
#define PT64_BASE_ADDR_MASK (physical_mask & ~(u64)(PAGE_SIZE-1))
#else
#define PT64_BASE_ADDR_MASK (((1ULL << 52) - 1) & ~(u64)(PAGE_SIZE-1))
#endif
#define PT64_PERM_MASK (PT_PRESENT_MASK | PT_WRITABLE_MASK | shadow_user_mask \
| shadow_x_mask | shadow_nx_mask | shadow_me_mask)
#define ACC_EXEC_MASK 1
#define ACC_WRITE_MASK PT_WRITABLE_MASK
#define ACC_USER_MASK PT_USER_MASK
#define ACC_ALL (ACC_EXEC_MASK | ACC_WRITE_MASK | ACC_USER_MASK)
/* The mask for the R/X bits in EPT PTEs */
#define PT64_EPT_READABLE_MASK 0x1ull
#define PT64_EPT_EXECUTABLE_MASK 0x4ull
#define PT64_LEVEL_BITS 9
#define PT64_LEVEL_SHIFT(level) \
(PAGE_SHIFT + (level - 1) * PT64_LEVEL_BITS)
#define PT64_INDEX(address, level)\
(((address) >> PT64_LEVEL_SHIFT(level)) & ((1 << PT64_LEVEL_BITS) - 1))
#define SHADOW_PT_INDEX(addr, level) PT64_INDEX(addr, level)
/*
* The mask/shift to use for saving the original R/X bits when marking the PTE
* as not-present for access tracking purposes. We do not save the W bit as the
* PTEs being access tracked also need to be dirty tracked, so the W bit will be
* restored only when a write is attempted to the page. This mask obviously
* must not overlap the A/D type mask.
*/
#define SHADOW_ACC_TRACK_SAVED_BITS_MASK (PT64_EPT_READABLE_MASK | \
PT64_EPT_EXECUTABLE_MASK)
#define SHADOW_ACC_TRACK_SAVED_BITS_SHIFT 54
#define SHADOW_ACC_TRACK_SAVED_MASK (SHADOW_ACC_TRACK_SAVED_BITS_MASK << \
SHADOW_ACC_TRACK_SAVED_BITS_SHIFT)
static_assert(!(SPTE_TDP_AD_MASK & SHADOW_ACC_TRACK_SAVED_MASK));
/*
* {DEFAULT,EPT}_SPTE_{HOST,MMU}_WRITABLE are used to keep track of why a given
* SPTE is write-protected. See is_writable_pte() for details.
*/
/* Bits 9 and 10 are ignored by all non-EPT PTEs. */
#define DEFAULT_SPTE_HOST_WRITABLE BIT_ULL(9)
#define DEFAULT_SPTE_MMU_WRITABLE BIT_ULL(10)
/*
* Low ignored bits are at a premium for EPT, use high ignored bits, taking care
* to not overlap the A/D type mask or the saved access bits of access-tracked
* SPTEs when A/D bits are disabled.
*/
#define EPT_SPTE_HOST_WRITABLE BIT_ULL(57)
#define EPT_SPTE_MMU_WRITABLE BIT_ULL(58)
static_assert(!(EPT_SPTE_HOST_WRITABLE & SPTE_TDP_AD_MASK));
static_assert(!(EPT_SPTE_MMU_WRITABLE & SPTE_TDP_AD_MASK));
static_assert(!(EPT_SPTE_HOST_WRITABLE & SHADOW_ACC_TRACK_SAVED_MASK));
static_assert(!(EPT_SPTE_MMU_WRITABLE & SHADOW_ACC_TRACK_SAVED_MASK));
/* Defined only to keep the above static asserts readable. */
#undef SHADOW_ACC_TRACK_SAVED_MASK
/*
* Due to limited space in PTEs, the MMIO generation is a 19 bit subset of
* the memslots generation and is derived as follows:
*
* Bits 0-7 of the MMIO generation are propagated to spte bits 3-10
* Bits 8-18 of the MMIO generation are propagated to spte bits 52-62
*
* The KVM_MEMSLOT_GEN_UPDATE_IN_PROGRESS flag is intentionally not included in
* the MMIO generation number, as doing so would require stealing a bit from
* the "real" generation number and thus effectively halve the maximum number
* of MMIO generations that can be handled before encountering a wrap (which
* requires a full MMU zap). The flag is instead explicitly queried when
* checking for MMIO spte cache hits.
*/
#define MMIO_SPTE_GEN_LOW_START 3
#define MMIO_SPTE_GEN_LOW_END 10
#define MMIO_SPTE_GEN_HIGH_START 52
#define MMIO_SPTE_GEN_HIGH_END 62
#define MMIO_SPTE_GEN_LOW_MASK GENMASK_ULL(MMIO_SPTE_GEN_LOW_END, \
MMIO_SPTE_GEN_LOW_START)
#define MMIO_SPTE_GEN_HIGH_MASK GENMASK_ULL(MMIO_SPTE_GEN_HIGH_END, \
MMIO_SPTE_GEN_HIGH_START)
static_assert(!(SPTE_MMU_PRESENT_MASK &
(MMIO_SPTE_GEN_LOW_MASK | MMIO_SPTE_GEN_HIGH_MASK)));
#define MMIO_SPTE_GEN_LOW_BITS (MMIO_SPTE_GEN_LOW_END - MMIO_SPTE_GEN_LOW_START + 1)
#define MMIO_SPTE_GEN_HIGH_BITS (MMIO_SPTE_GEN_HIGH_END - MMIO_SPTE_GEN_HIGH_START + 1)
/* remember to adjust the comment above as well if you change these */
static_assert(MMIO_SPTE_GEN_LOW_BITS == 8 && MMIO_SPTE_GEN_HIGH_BITS == 11);
#define MMIO_SPTE_GEN_LOW_SHIFT (MMIO_SPTE_GEN_LOW_START - 0)
#define MMIO_SPTE_GEN_HIGH_SHIFT (MMIO_SPTE_GEN_HIGH_START - MMIO_SPTE_GEN_LOW_BITS)
#define MMIO_SPTE_GEN_MASK GENMASK_ULL(MMIO_SPTE_GEN_LOW_BITS + MMIO_SPTE_GEN_HIGH_BITS - 1, 0)
extern u64 __read_mostly shadow_host_writable_mask;
extern u64 __read_mostly shadow_mmu_writable_mask;
extern u64 __read_mostly shadow_nx_mask;
extern u64 __read_mostly shadow_x_mask; /* mutual exclusive with nx_mask */
extern u64 __read_mostly shadow_user_mask;
extern u64 __read_mostly shadow_accessed_mask;
extern u64 __read_mostly shadow_dirty_mask;
extern u64 __read_mostly shadow_mmio_value;
extern u64 __read_mostly shadow_mmio_mask;
extern u64 __read_mostly shadow_mmio_access_mask;
extern u64 __read_mostly shadow_present_mask;
extern u64 __read_mostly shadow_me_mask;
/*
* SPTEs in MMUs without A/D bits are marked with SPTE_TDP_AD_DISABLED_MASK;
* shadow_acc_track_mask is the set of bits to be cleared in non-accessed
* pages.
*/
extern u64 __read_mostly shadow_acc_track_mask;
/*
* This mask must be set on all non-zero Non-Present or Reserved SPTEs in order
* to guard against L1TF attacks.
*/
extern u64 __read_mostly shadow_nonpresent_or_rsvd_mask;
/*
* The number of high-order 1 bits to use in the mask above.
*/
#define SHADOW_NONPRESENT_OR_RSVD_MASK_LEN 5
/*
* If a thread running without exclusive control of the MMU lock must perform a
* multi-part operation on an SPTE, it can set the SPTE to REMOVED_SPTE as a
* non-present intermediate value. Other threads which encounter this value
* should not modify the SPTE.
*
* Use a semi-arbitrary value that doesn't set RWX bits, i.e. is not-present on
* bot AMD and Intel CPUs, and doesn't set PFN bits, i.e. doesn't create a L1TF
* vulnerability. Use only low bits to avoid 64-bit immediates.
*
* Only used by the TDP MMU.
*/
#define REMOVED_SPTE 0x5a0ULL
/* Removed SPTEs must not be misconstrued as shadow present PTEs. */
static_assert(!(REMOVED_SPTE & SPTE_MMU_PRESENT_MASK));
static inline bool is_removed_spte(u64 spte)
{
return spte == REMOVED_SPTE;
}
/*
* In some cases, we need to preserve the GFN of a non-present or reserved
* SPTE when we usurp the upper five bits of the physical address space to
* defend against L1TF, e.g. for MMIO SPTEs. To preserve the GFN, we'll
* shift bits of the GFN that overlap with shadow_nonpresent_or_rsvd_mask
* left into the reserved bits, i.e. the GFN in the SPTE will be split into
* high and low parts. This mask covers the lower bits of the GFN.
*/
extern u64 __read_mostly shadow_nonpresent_or_rsvd_lower_gfn_mask;
/*
* The number of non-reserved physical address bits irrespective of features
* that repurpose legal bits, e.g. MKTME.
*/
extern u8 __read_mostly shadow_phys_bits;
static inline bool is_mmio_spte(u64 spte)
{
return (spte & shadow_mmio_mask) == shadow_mmio_value &&
likely(shadow_mmio_value);
}
static inline bool is_shadow_present_pte(u64 pte)
{
return !!(pte & SPTE_MMU_PRESENT_MASK);
}
static inline bool sp_ad_disabled(struct kvm_mmu_page *sp)
{
return sp->role.ad_disabled;
}
static inline bool spte_ad_enabled(u64 spte)
{
MMU_WARN_ON(!is_shadow_present_pte(spte));
return (spte & SPTE_TDP_AD_MASK) != SPTE_TDP_AD_DISABLED_MASK;
}
static inline bool spte_ad_need_write_protect(u64 spte)
{
MMU_WARN_ON(!is_shadow_present_pte(spte));
/*
* This is benign for non-TDP SPTEs as SPTE_TDP_AD_ENABLED_MASK is '0',
* and non-TDP SPTEs will never set these bits. Optimize for 64-bit
* TDP and do the A/D type check unconditionally.
*/
return (spte & SPTE_TDP_AD_MASK) != SPTE_TDP_AD_ENABLED_MASK;
}
static inline u64 spte_shadow_accessed_mask(u64 spte)
{
MMU_WARN_ON(!is_shadow_present_pte(spte));
return spte_ad_enabled(spte) ? shadow_accessed_mask : 0;
}
static inline u64 spte_shadow_dirty_mask(u64 spte)
{
MMU_WARN_ON(!is_shadow_present_pte(spte));
return spte_ad_enabled(spte) ? shadow_dirty_mask : 0;
}
static inline bool is_access_track_spte(u64 spte)
{
return !spte_ad_enabled(spte) && (spte & shadow_acc_track_mask) == 0;
}
static inline bool is_large_pte(u64 pte)
{
return pte & PT_PAGE_SIZE_MASK;
}
static inline bool is_last_spte(u64 pte, int level)
{
return (level == PG_LEVEL_4K) || is_large_pte(pte);
}
static inline bool is_executable_pte(u64 spte)
{
return (spte & (shadow_x_mask | shadow_nx_mask)) == shadow_x_mask;
}
static inline kvm_pfn_t spte_to_pfn(u64 pte)
{
return (pte & PT64_BASE_ADDR_MASK) >> PAGE_SHIFT;
}
static inline bool is_accessed_spte(u64 spte)
{
u64 accessed_mask = spte_shadow_accessed_mask(spte);
return accessed_mask ? spte & accessed_mask
: !is_access_track_spte(spte);
}
static inline bool is_dirty_spte(u64 spte)
{
u64 dirty_mask = spte_shadow_dirty_mask(spte);
return dirty_mask ? spte & dirty_mask : spte & PT_WRITABLE_MASK;
}
static inline u64 get_rsvd_bits(struct rsvd_bits_validate *rsvd_check, u64 pte,
int level)
{
int bit7 = (pte >> 7) & 1;
return rsvd_check->rsvd_bits_mask[bit7][level-1];
}
static inline bool __is_rsvd_bits_set(struct rsvd_bits_validate *rsvd_check,
u64 pte, int level)
{
return pte & get_rsvd_bits(rsvd_check, pte, level);
}
static inline bool __is_bad_mt_xwr(struct rsvd_bits_validate *rsvd_check,
u64 pte)
{
return rsvd_check->bad_mt_xwr & BIT_ULL(pte & 0x3f);
}
static __always_inline bool is_rsvd_spte(struct rsvd_bits_validate *rsvd_check,
u64 spte, int level)
{
return __is_bad_mt_xwr(rsvd_check, spte) ||
__is_rsvd_bits_set(rsvd_check, spte, level);
}
/*
* An shadow-present leaf SPTE may be non-writable for 3 possible reasons:
*
* 1. To intercept writes for dirty logging. KVM write-protects huge pages
* so that they can be split be split down into the dirty logging
* granularity (4KiB) whenever the guest writes to them. KVM also
* write-protects 4KiB pages so that writes can be recorded in the dirty log
* (e.g. if not using PML). SPTEs are write-protected for dirty logging
* during the VM-iotcls that enable dirty logging.
*
* 2. To intercept writes to guest page tables that KVM is shadowing. When a
* guest writes to its page table the corresponding shadow page table will
* be marked "unsync". That way KVM knows which shadow page tables need to
* be updated on the next TLB flush, INVLPG, etc. and which do not.
*
* 3. To prevent guest writes to read-only memory, such as for memory in a
* read-only memslot or guest memory backed by a read-only VMA. Writes to
* such pages are disallowed entirely.
*
* To keep track of why a given SPTE is write-protected, KVM uses 2
* software-only bits in the SPTE:
*
* shadow_mmu_writable_mask, aka MMU-writable -
* Cleared on SPTEs that KVM is currently write-protecting for shadow paging
* purposes (case 2 above).
*
* shadow_host_writable_mask, aka Host-writable -
* Cleared on SPTEs that are not host-writable (case 3 above)
*
* Note, not all possible combinations of PT_WRITABLE_MASK,
* shadow_mmu_writable_mask, and shadow_host_writable_mask are valid. A given
* SPTE can be in only one of the following states, which map to the
* aforementioned 3 cases:
*
* shadow_host_writable_mask | shadow_mmu_writable_mask | PT_WRITABLE_MASK
* ------------------------- | ------------------------ | ----------------
* 1 | 1 | 1 (writable)
* 1 | 1 | 0 (case 1)
* 1 | 0 | 0 (case 2)
* 0 | 0 | 0 (case 3)
*
* The valid combinations of these bits are checked by
* check_spte_writable_invariants() whenever an SPTE is modified.
*
* Clearing the MMU-writable bit is always done under the MMU lock and always
* accompanied by a TLB flush before dropping the lock to avoid corrupting the
* shadow page tables between vCPUs. Write-protecting an SPTE for dirty logging
* (which does not clear the MMU-writable bit), does not flush TLBs before
* dropping the lock, as it only needs to synchronize guest writes with the
* dirty bitmap.
*
* So, there is the problem: clearing the MMU-writable bit can encounter a
* write-protected SPTE while CPUs still have writable mappings for that SPTE
* cached in their TLB. To address this, KVM always flushes TLBs when
* write-protecting SPTEs if the MMU-writable bit is set on the old SPTE.
*
* The Host-writable bit is not modified on present SPTEs, it is only set or
* cleared when an SPTE is first faulted in from non-present and then remains
* immutable.
*/
static inline bool is_writable_pte(unsigned long pte)
{
return pte & PT_WRITABLE_MASK;
}
/* Note: spte must be a shadow-present leaf SPTE. */
static inline void check_spte_writable_invariants(u64 spte)
{
if (spte & shadow_mmu_writable_mask)
WARN_ONCE(!(spte & shadow_host_writable_mask),
"kvm: MMU-writable SPTE is not Host-writable: %llx",
spte);
else
WARN_ONCE(is_writable_pte(spte),
"kvm: Writable SPTE is not MMU-writable: %llx", spte);
}
static inline bool spte_can_locklessly_be_made_writable(u64 spte)
{
return spte & shadow_mmu_writable_mask;
}
static inline u64 get_mmio_spte_generation(u64 spte)
{
u64 gen;
gen = (spte & MMIO_SPTE_GEN_LOW_MASK) >> MMIO_SPTE_GEN_LOW_SHIFT;
gen |= (spte & MMIO_SPTE_GEN_HIGH_MASK) >> MMIO_SPTE_GEN_HIGH_SHIFT;
return gen;
}
bool make_spte(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp,
const struct kvm_memory_slot *slot,
unsigned int pte_access, gfn_t gfn, kvm_pfn_t pfn,
u64 old_spte, bool prefetch, bool can_unsync,
bool host_writable, u64 *new_spte);
u64 make_huge_page_split_spte(u64 huge_spte, int huge_level, int index);
u64 make_nonleaf_spte(u64 *child_pt, bool ad_disabled);
u64 make_mmio_spte(struct kvm_vcpu *vcpu, u64 gfn, unsigned int access);
u64 mark_spte_for_access_track(u64 spte);
/* Restore an acc-track PTE back to a regular PTE */
static inline u64 restore_acc_track_spte(u64 spte)
{
u64 saved_bits = (spte >> SHADOW_ACC_TRACK_SAVED_BITS_SHIFT)
& SHADOW_ACC_TRACK_SAVED_BITS_MASK;
spte &= ~shadow_acc_track_mask;
spte &= ~(SHADOW_ACC_TRACK_SAVED_BITS_MASK <<
SHADOW_ACC_TRACK_SAVED_BITS_SHIFT);
spte |= saved_bits;
return spte;
}
u64 kvm_mmu_changed_pte_notifier_make_spte(u64 old_spte, kvm_pfn_t new_pfn);
void kvm_mmu_reset_all_pte_masks(void);
#endif