From ae7a2a3fb6f8b784c2752863f4f1f20c656f76fb Mon Sep 17 00:00:00 2001 From: "Michael S. Tsirkin" Date: Sun, 24 Jun 2012 19:25:07 +0300 Subject: KVM: host side for eoi optimization Implementation of PV EOI using shared memory. This reduces the number of exits an interrupt causes as much as by half. The idea is simple: there's a bit, per APIC, in guest memory, that tells the guest that it does not need EOI. We set it before injecting an interrupt and clear before injecting a nested one. Guest tests it using a test and clear operation - this is necessary so that host can detect interrupt nesting - and if set, it can skip the EOI MSR. There's a new MSR to set the address of said register in guest memory. Otherwise not much changed: - Guest EOI is not required - Register is tested & ISR is automatically cleared on exit For testing results see description of previous patch 'kvm_para: guest side for eoi avoidance'. Signed-off-by: Michael S. Tsirkin Signed-off-by: Avi Kivity --- arch/x86/include/asm/kvm_host.h | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'arch/x86/include/asm/kvm_host.h') diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index db7c1f2709a2..24b76474d9de 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -175,6 +175,13 @@ enum { /* apic attention bits */ #define KVM_APIC_CHECK_VAPIC 0 +/* + * The following bit is set with PV-EOI, unset on EOI. + * We detect PV-EOI changes by guest by comparing + * this bit with PV-EOI in guest memory. + * See the implementation in apic_update_pv_eoi. + */ +#define KVM_APIC_PV_EOI_PENDING 1 /* * We don't want allocation failures within the mmu code, so we preallocate @@ -484,6 +491,11 @@ struct kvm_vcpu_arch { u64 length; u64 status; } osvw; + + struct { + u64 msr_val; + struct gfn_to_hva_cache data; + } pv_eoi; }; struct kvm_lpage_info { -- cgit v1.2.3 From 15c7ad51ad58cbd3b46112c1840bc7228bd354bf Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 20 Jun 2012 20:46:33 +0200 Subject: perf/x86: Rename Intel specific macros There are macros that are Intel specific and not x86 generic. Rename them into INTEL_*. This patch removes X86_PMC_IDX_GENERIC and does: $ sed -i -e 's/X86_PMC_MAX_/INTEL_PMC_MAX_/g' \ arch/x86/include/asm/kvm_host.h \ arch/x86/include/asm/perf_event.h \ arch/x86/kernel/cpu/perf_event.c \ arch/x86/kernel/cpu/perf_event_p4.c \ arch/x86/kvm/pmu.c $ sed -i -e 's/X86_PMC_IDX_FIXED/INTEL_PMC_IDX_FIXED/g' \ arch/x86/include/asm/perf_event.h \ arch/x86/kernel/cpu/perf_event.c \ arch/x86/kernel/cpu/perf_event_intel.c \ arch/x86/kernel/cpu/perf_event_intel_ds.c \ arch/x86/kvm/pmu.c $ sed -i -e 's/X86_PMC_MSK_/INTEL_PMC_MSK_/g' \ arch/x86/include/asm/perf_event.h \ arch/x86/kernel/cpu/perf_event.c Signed-off-by: Robert Richter Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1340217996-2254-2-git-send-email-robert.richter@amd.com Signed-off-by: Ingo Molnar --- arch/x86/include/asm/kvm_host.h | 4 ++-- arch/x86/include/asm/perf_event.h | 17 +++++++------- arch/x86/kernel/cpu/perf_event.c | 38 +++++++++++++++---------------- arch/x86/kernel/cpu/perf_event_intel.c | 14 ++++++------ arch/x86/kernel/cpu/perf_event_intel_ds.c | 4 ++-- arch/x86/kernel/cpu/perf_event_p4.c | 2 +- arch/x86/kvm/pmu.c | 22 +++++++++--------- 7 files changed, 50 insertions(+), 51 deletions(-) (limited to 'arch/x86/include/asm/kvm_host.h') diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index db7c1f2709a2..2da88c0cda14 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -313,8 +313,8 @@ struct kvm_pmu { u64 counter_bitmask[2]; u64 global_ctrl_mask; u8 version; - struct kvm_pmc gp_counters[X86_PMC_MAX_GENERIC]; - struct kvm_pmc fixed_counters[X86_PMC_MAX_FIXED]; + struct kvm_pmc gp_counters[INTEL_PMC_MAX_GENERIC]; + struct kvm_pmc fixed_counters[INTEL_PMC_MAX_FIXED]; struct irq_work irq_work; u64 reprogram_pmi; }; diff --git a/arch/x86/include/asm/perf_event.h b/arch/x86/include/asm/perf_event.h index 588f52ea810e..3b31248caf60 100644 --- a/arch/x86/include/asm/perf_event.h +++ b/arch/x86/include/asm/perf_event.h @@ -5,11 +5,10 @@ * Performance event hw details: */ -#define X86_PMC_MAX_GENERIC 32 -#define X86_PMC_MAX_FIXED 3 +#define INTEL_PMC_MAX_GENERIC 32 +#define INTEL_PMC_MAX_FIXED 3 +#define INTEL_PMC_IDX_FIXED 32 -#define X86_PMC_IDX_GENERIC 0 -#define X86_PMC_IDX_FIXED 32 #define X86_PMC_IDX_MAX 64 #define MSR_ARCH_PERFMON_PERFCTR0 0xc1 @@ -121,16 +120,16 @@ struct x86_pmu_capability { /* Instr_Retired.Any: */ #define MSR_ARCH_PERFMON_FIXED_CTR0 0x309 -#define X86_PMC_IDX_FIXED_INSTRUCTIONS (X86_PMC_IDX_FIXED + 0) +#define INTEL_PMC_IDX_FIXED_INSTRUCTIONS (INTEL_PMC_IDX_FIXED + 0) /* CPU_CLK_Unhalted.Core: */ #define MSR_ARCH_PERFMON_FIXED_CTR1 0x30a -#define X86_PMC_IDX_FIXED_CPU_CYCLES (X86_PMC_IDX_FIXED + 1) +#define INTEL_PMC_IDX_FIXED_CPU_CYCLES (INTEL_PMC_IDX_FIXED + 1) /* CPU_CLK_Unhalted.Ref: */ #define MSR_ARCH_PERFMON_FIXED_CTR2 0x30b -#define X86_PMC_IDX_FIXED_REF_CYCLES (X86_PMC_IDX_FIXED + 2) -#define X86_PMC_MSK_FIXED_REF_CYCLES (1ULL << X86_PMC_IDX_FIXED_REF_CYCLES) +#define INTEL_PMC_IDX_FIXED_REF_CYCLES (INTEL_PMC_IDX_FIXED + 2) +#define INTEL_PMC_MSK_FIXED_REF_CYCLES (1ULL << INTEL_PMC_IDX_FIXED_REF_CYCLES) /* * We model BTS tracing as another fixed-mode PMC. @@ -139,7 +138,7 @@ struct x86_pmu_capability { * values are used by actual fixed events and higher values are used * to indicate other overflow conditions in the PERF_GLOBAL_STATUS msr. */ -#define X86_PMC_IDX_FIXED_BTS (X86_PMC_IDX_FIXED + 16) +#define INTEL_PMC_IDX_FIXED_BTS (INTEL_PMC_IDX_FIXED + 16) /* * IBS cpuid feature detection diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index e677d9923f4f..668050002602 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -63,7 +63,7 @@ u64 x86_perf_event_update(struct perf_event *event) int idx = hwc->idx; s64 delta; - if (idx == X86_PMC_IDX_FIXED_BTS) + if (idx == INTEL_PMC_IDX_FIXED_BTS) return 0; /* @@ -626,8 +626,8 @@ static bool __perf_sched_find_counter(struct perf_sched *sched) c = sched->constraints[sched->state.event]; /* Prefer fixed purpose counters */ - if (c->idxmsk64 & (~0ULL << X86_PMC_IDX_FIXED)) { - idx = X86_PMC_IDX_FIXED; + if (c->idxmsk64 & (~0ULL << INTEL_PMC_IDX_FIXED)) { + idx = INTEL_PMC_IDX_FIXED; for_each_set_bit_from(idx, c->idxmsk, X86_PMC_IDX_MAX) { if (!__test_and_set_bit(idx, sched->state.used)) goto done; @@ -635,7 +635,7 @@ static bool __perf_sched_find_counter(struct perf_sched *sched) } /* Grab the first unused counter starting with idx */ idx = sched->state.counter; - for_each_set_bit_from(idx, c->idxmsk, X86_PMC_IDX_FIXED) { + for_each_set_bit_from(idx, c->idxmsk, INTEL_PMC_IDX_FIXED) { if (!__test_and_set_bit(idx, sched->state.used)) goto done; } @@ -813,13 +813,13 @@ static inline void x86_assign_hw_event(struct perf_event *event, hwc->last_cpu = smp_processor_id(); hwc->last_tag = ++cpuc->tags[i]; - if (hwc->idx == X86_PMC_IDX_FIXED_BTS) { + if (hwc->idx == INTEL_PMC_IDX_FIXED_BTS) { hwc->config_base = 0; hwc->event_base = 0; - } else if (hwc->idx >= X86_PMC_IDX_FIXED) { + } else if (hwc->idx >= INTEL_PMC_IDX_FIXED) { hwc->config_base = MSR_ARCH_PERFMON_FIXED_CTR_CTRL; - hwc->event_base = MSR_ARCH_PERFMON_FIXED_CTR0 + (hwc->idx - X86_PMC_IDX_FIXED); - hwc->event_base_rdpmc = (hwc->idx - X86_PMC_IDX_FIXED) | 1<<30; + hwc->event_base = MSR_ARCH_PERFMON_FIXED_CTR0 + (hwc->idx - INTEL_PMC_IDX_FIXED); + hwc->event_base_rdpmc = (hwc->idx - INTEL_PMC_IDX_FIXED) | 1<<30; } else { hwc->config_base = x86_pmu_config_addr(hwc->idx); hwc->event_base = x86_pmu_event_addr(hwc->idx); @@ -921,7 +921,7 @@ int x86_perf_event_set_period(struct perf_event *event) s64 period = hwc->sample_period; int ret = 0, idx = hwc->idx; - if (idx == X86_PMC_IDX_FIXED_BTS) + if (idx == INTEL_PMC_IDX_FIXED_BTS) return 0; /* @@ -1338,21 +1338,21 @@ static int __init init_hw_perf_events(void) for (quirk = x86_pmu.quirks; quirk; quirk = quirk->next) quirk->func(); - if (x86_pmu.num_counters > X86_PMC_MAX_GENERIC) { + if (x86_pmu.num_counters > INTEL_PMC_MAX_GENERIC) { WARN(1, KERN_ERR "hw perf events %d > max(%d), clipping!", - x86_pmu.num_counters, X86_PMC_MAX_GENERIC); - x86_pmu.num_counters = X86_PMC_MAX_GENERIC; + x86_pmu.num_counters, INTEL_PMC_MAX_GENERIC); + x86_pmu.num_counters = INTEL_PMC_MAX_GENERIC; } x86_pmu.intel_ctrl = (1 << x86_pmu.num_counters) - 1; - if (x86_pmu.num_counters_fixed > X86_PMC_MAX_FIXED) { + if (x86_pmu.num_counters_fixed > INTEL_PMC_MAX_FIXED) { WARN(1, KERN_ERR "hw perf events fixed %d > max(%d), clipping!", - x86_pmu.num_counters_fixed, X86_PMC_MAX_FIXED); - x86_pmu.num_counters_fixed = X86_PMC_MAX_FIXED; + x86_pmu.num_counters_fixed, INTEL_PMC_MAX_FIXED); + x86_pmu.num_counters_fixed = INTEL_PMC_MAX_FIXED; } x86_pmu.intel_ctrl |= - ((1LL << x86_pmu.num_counters_fixed)-1) << X86_PMC_IDX_FIXED; + ((1LL << x86_pmu.num_counters_fixed)-1) << INTEL_PMC_IDX_FIXED; perf_events_lapic_init(); register_nmi_handler(NMI_LOCAL, perf_event_nmi_handler, 0, "PMI"); @@ -1368,7 +1368,7 @@ static int __init init_hw_perf_events(void) */ for_each_event_constraint(c, x86_pmu.event_constraints) { if (c->cmask != X86_RAW_EVENT_MASK - || c->idxmsk64 == X86_PMC_MSK_FIXED_REF_CYCLES) { + || c->idxmsk64 == INTEL_PMC_MSK_FIXED_REF_CYCLES) { continue; } @@ -1611,8 +1611,8 @@ static int x86_pmu_event_idx(struct perf_event *event) if (!x86_pmu.attr_rdpmc) return 0; - if (x86_pmu.num_counters_fixed && idx >= X86_PMC_IDX_FIXED) { - idx -= X86_PMC_IDX_FIXED; + if (x86_pmu.num_counters_fixed && idx >= INTEL_PMC_IDX_FIXED) { + idx -= INTEL_PMC_IDX_FIXED; idx |= 1 << 30; } diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c index 8408e37f5fa4..5b0b362c7ae6 100644 --- a/arch/x86/kernel/cpu/perf_event_intel.c +++ b/arch/x86/kernel/cpu/perf_event_intel.c @@ -747,7 +747,7 @@ static void intel_pmu_disable_all(void) wrmsrl(MSR_CORE_PERF_GLOBAL_CTRL, 0); - if (test_bit(X86_PMC_IDX_FIXED_BTS, cpuc->active_mask)) + if (test_bit(INTEL_PMC_IDX_FIXED_BTS, cpuc->active_mask)) intel_pmu_disable_bts(); intel_pmu_pebs_disable_all(); @@ -763,9 +763,9 @@ static void intel_pmu_enable_all(int added) wrmsrl(MSR_CORE_PERF_GLOBAL_CTRL, x86_pmu.intel_ctrl & ~cpuc->intel_ctrl_guest_mask); - if (test_bit(X86_PMC_IDX_FIXED_BTS, cpuc->active_mask)) { + if (test_bit(INTEL_PMC_IDX_FIXED_BTS, cpuc->active_mask)) { struct perf_event *event = - cpuc->events[X86_PMC_IDX_FIXED_BTS]; + cpuc->events[INTEL_PMC_IDX_FIXED_BTS]; if (WARN_ON_ONCE(!event)) return; @@ -871,7 +871,7 @@ static inline void intel_pmu_ack_status(u64 ack) static void intel_pmu_disable_fixed(struct hw_perf_event *hwc) { - int idx = hwc->idx - X86_PMC_IDX_FIXED; + int idx = hwc->idx - INTEL_PMC_IDX_FIXED; u64 ctrl_val, mask; mask = 0xfULL << (idx * 4); @@ -886,7 +886,7 @@ static void intel_pmu_disable_event(struct perf_event *event) struct hw_perf_event *hwc = &event->hw; struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); - if (unlikely(hwc->idx == X86_PMC_IDX_FIXED_BTS)) { + if (unlikely(hwc->idx == INTEL_PMC_IDX_FIXED_BTS)) { intel_pmu_disable_bts(); intel_pmu_drain_bts_buffer(); return; @@ -915,7 +915,7 @@ static void intel_pmu_disable_event(struct perf_event *event) static void intel_pmu_enable_fixed(struct hw_perf_event *hwc) { - int idx = hwc->idx - X86_PMC_IDX_FIXED; + int idx = hwc->idx - INTEL_PMC_IDX_FIXED; u64 ctrl_val, bits, mask; /* @@ -949,7 +949,7 @@ static void intel_pmu_enable_event(struct perf_event *event) struct hw_perf_event *hwc = &event->hw; struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); - if (unlikely(hwc->idx == X86_PMC_IDX_FIXED_BTS)) { + if (unlikely(hwc->idx == INTEL_PMC_IDX_FIXED_BTS)) { if (!__this_cpu_read(cpu_hw_events.enabled)) return; diff --git a/arch/x86/kernel/cpu/perf_event_intel_ds.c b/arch/x86/kernel/cpu/perf_event_intel_ds.c index 026373edef7f..629ae0b7ad90 100644 --- a/arch/x86/kernel/cpu/perf_event_intel_ds.c +++ b/arch/x86/kernel/cpu/perf_event_intel_ds.c @@ -248,7 +248,7 @@ void reserve_ds_buffers(void) */ struct event_constraint bts_constraint = - EVENT_CONSTRAINT(0, 1ULL << X86_PMC_IDX_FIXED_BTS, 0); + EVENT_CONSTRAINT(0, 1ULL << INTEL_PMC_IDX_FIXED_BTS, 0); void intel_pmu_enable_bts(u64 config) { @@ -295,7 +295,7 @@ int intel_pmu_drain_bts_buffer(void) u64 to; u64 flags; }; - struct perf_event *event = cpuc->events[X86_PMC_IDX_FIXED_BTS]; + struct perf_event *event = cpuc->events[INTEL_PMC_IDX_FIXED_BTS]; struct bts_record *at, *top; struct perf_output_handle handle; struct perf_event_header header; diff --git a/arch/x86/kernel/cpu/perf_event_p4.c b/arch/x86/kernel/cpu/perf_event_p4.c index 6c82e4037989..92c7e39a079f 100644 --- a/arch/x86/kernel/cpu/perf_event_p4.c +++ b/arch/x86/kernel/cpu/perf_event_p4.c @@ -1325,7 +1325,7 @@ __init int p4_pmu_init(void) unsigned int low, high; /* If we get stripped -- indexing fails */ - BUILD_BUG_ON(ARCH_P4_MAX_CCCR > X86_PMC_MAX_GENERIC); + BUILD_BUG_ON(ARCH_P4_MAX_CCCR > INTEL_PMC_MAX_GENERIC); rdmsr(MSR_IA32_MISC_ENABLE, low, high); if (!(low & (1 << 7))) { diff --git a/arch/x86/kvm/pmu.c b/arch/x86/kvm/pmu.c index 2e88438ffd83..9b7ec1150ab0 100644 --- a/arch/x86/kvm/pmu.c +++ b/arch/x86/kvm/pmu.c @@ -80,10 +80,10 @@ static inline struct kvm_pmc *get_fixed_pmc_idx(struct kvm_pmu *pmu, int idx) static struct kvm_pmc *global_idx_to_pmc(struct kvm_pmu *pmu, int idx) { - if (idx < X86_PMC_IDX_FIXED) + if (idx < INTEL_PMC_IDX_FIXED) return get_gp_pmc(pmu, MSR_P6_EVNTSEL0 + idx, MSR_P6_EVNTSEL0); else - return get_fixed_pmc_idx(pmu, idx - X86_PMC_IDX_FIXED); + return get_fixed_pmc_idx(pmu, idx - INTEL_PMC_IDX_FIXED); } void kvm_deliver_pmi(struct kvm_vcpu *vcpu) @@ -291,7 +291,7 @@ static void reprogram_idx(struct kvm_pmu *pmu, int idx) if (pmc_is_gp(pmc)) reprogram_gp_counter(pmc, pmc->eventsel); else { - int fidx = idx - X86_PMC_IDX_FIXED; + int fidx = idx - INTEL_PMC_IDX_FIXED; reprogram_fixed_counter(pmc, fixed_en_pmi(pmu->fixed_ctr_ctrl, fidx), fidx); } @@ -452,7 +452,7 @@ void kvm_pmu_cpuid_update(struct kvm_vcpu *vcpu) return; pmu->nr_arch_gp_counters = min((int)(entry->eax >> 8) & 0xff, - X86_PMC_MAX_GENERIC); + INTEL_PMC_MAX_GENERIC); pmu->counter_bitmask[KVM_PMC_GP] = ((u64)1 << ((entry->eax >> 16) & 0xff)) - 1; bitmap_len = (entry->eax >> 24) & 0xff; @@ -462,13 +462,13 @@ void kvm_pmu_cpuid_update(struct kvm_vcpu *vcpu) pmu->nr_arch_fixed_counters = 0; } else { pmu->nr_arch_fixed_counters = min((int)(entry->edx & 0x1f), - X86_PMC_MAX_FIXED); + INTEL_PMC_MAX_FIXED); pmu->counter_bitmask[KVM_PMC_FIXED] = ((u64)1 << ((entry->edx >> 5) & 0xff)) - 1; } pmu->global_ctrl = ((1 << pmu->nr_arch_gp_counters) - 1) | - (((1ull << pmu->nr_arch_fixed_counters) - 1) << X86_PMC_IDX_FIXED); + (((1ull << pmu->nr_arch_fixed_counters) - 1) << INTEL_PMC_IDX_FIXED); pmu->global_ctrl_mask = ~pmu->global_ctrl; } @@ -478,15 +478,15 @@ void kvm_pmu_init(struct kvm_vcpu *vcpu) struct kvm_pmu *pmu = &vcpu->arch.pmu; memset(pmu, 0, sizeof(*pmu)); - for (i = 0; i < X86_PMC_MAX_GENERIC; i++) { + for (i = 0; i < INTEL_PMC_MAX_GENERIC; i++) { pmu->gp_counters[i].type = KVM_PMC_GP; pmu->gp_counters[i].vcpu = vcpu; pmu->gp_counters[i].idx = i; } - for (i = 0; i < X86_PMC_MAX_FIXED; i++) { + for (i = 0; i < INTEL_PMC_MAX_FIXED; i++) { pmu->fixed_counters[i].type = KVM_PMC_FIXED; pmu->fixed_counters[i].vcpu = vcpu; - pmu->fixed_counters[i].idx = i + X86_PMC_IDX_FIXED; + pmu->fixed_counters[i].idx = i + INTEL_PMC_IDX_FIXED; } init_irq_work(&pmu->irq_work, trigger_pmi); kvm_pmu_cpuid_update(vcpu); @@ -498,13 +498,13 @@ void kvm_pmu_reset(struct kvm_vcpu *vcpu) int i; irq_work_sync(&pmu->irq_work); - for (i = 0; i < X86_PMC_MAX_GENERIC; i++) { + for (i = 0; i < INTEL_PMC_MAX_GENERIC; i++) { struct kvm_pmc *pmc = &pmu->gp_counters[i]; stop_counter(pmc); pmc->counter = pmc->eventsel = 0; } - for (i = 0; i < X86_PMC_MAX_FIXED; i++) + for (i = 0; i < INTEL_PMC_MAX_FIXED; i++) stop_counter(&pmu->fixed_counters[i]); pmu->fixed_ctr_ctrl = pmu->global_ctrl = pmu->global_status = -- cgit v1.2.3 From ad756a1603c5fac207758faaac7f01c34c9d0b7b Mon Sep 17 00:00:00 2001 From: "Mao, Junjie" Date: Mon, 2 Jul 2012 01:18:48 +0000 Subject: KVM: VMX: Implement PCID/INVPCID for guests with EPT This patch handles PCID/INVPCID for guests. Process-context identifiers (PCIDs) are a facility by which a logical processor may cache information for multiple linear-address spaces so that the processor may retain cached information when software switches to a different linear address space. Refer to section 4.10.1 in IA32 Intel Software Developer's Manual Volume 3A for details. For guests with EPT, the PCID feature is enabled and INVPCID behaves as running natively. For guests without EPT, the PCID feature is disabled and INVPCID triggers #UD. Signed-off-by: Junjie Mao Signed-off-by: Avi Kivity --- arch/x86/include/asm/kvm_host.h | 4 +++- arch/x86/include/asm/processor-flags.h | 2 ++ arch/x86/include/asm/vmx.h | 2 ++ arch/x86/kvm/cpuid.c | 5 +++-- arch/x86/kvm/cpuid.h | 8 ++++++++ arch/x86/kvm/svm.c | 6 ++++++ arch/x86/kvm/vmx.c | 34 +++++++++++++++++++++++++++++++++- arch/x86/kvm/x86.c | 23 ++++++++++++++++++++--- 8 files changed, 77 insertions(+), 7 deletions(-) (limited to 'arch/x86/include/asm/kvm_host.h') diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index 24b76474d9de..a3e9409e90b6 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -48,12 +48,13 @@ #define CR3_PAE_RESERVED_BITS ((X86_CR3_PWT | X86_CR3_PCD) - 1) #define CR3_NONPAE_RESERVED_BITS ((PAGE_SIZE-1) & ~(X86_CR3_PWT | X86_CR3_PCD)) +#define CR3_PCID_ENABLED_RESERVED_BITS 0xFFFFFF0000000000ULL #define CR3_L_MODE_RESERVED_BITS (CR3_NONPAE_RESERVED_BITS | \ 0xFFFFFF0000000000ULL) #define CR4_RESERVED_BITS \ (~(unsigned long)(X86_CR4_VME | X86_CR4_PVI | X86_CR4_TSD | X86_CR4_DE\ | X86_CR4_PSE | X86_CR4_PAE | X86_CR4_MCE \ - | X86_CR4_PGE | X86_CR4_PCE | X86_CR4_OSFXSR \ + | X86_CR4_PGE | X86_CR4_PCE | X86_CR4_OSFXSR | X86_CR4_PCIDE \ | X86_CR4_OSXSAVE | X86_CR4_SMEP | X86_CR4_RDWRGSFS \ | X86_CR4_OSXMMEXCPT | X86_CR4_VMXE)) @@ -673,6 +674,7 @@ struct kvm_x86_ops { u64 (*get_mt_mask)(struct kvm_vcpu *vcpu, gfn_t gfn, bool is_mmio); int (*get_lpage_level)(void); bool (*rdtscp_supported)(void); + bool (*invpcid_supported)(void); void (*adjust_tsc_offset)(struct kvm_vcpu *vcpu, s64 adjustment, bool host); void (*set_tdp_cr3)(struct kvm_vcpu *vcpu, unsigned long cr3); diff --git a/arch/x86/include/asm/processor-flags.h b/arch/x86/include/asm/processor-flags.h index f8ab3eaad128..aea1d1d848c7 100644 --- a/arch/x86/include/asm/processor-flags.h +++ b/arch/x86/include/asm/processor-flags.h @@ -44,6 +44,7 @@ */ #define X86_CR3_PWT 0x00000008 /* Page Write Through */ #define X86_CR3_PCD 0x00000010 /* Page Cache Disable */ +#define X86_CR3_PCID_MASK 0x00000fff /* PCID Mask */ /* * Intel CPU features in CR4 @@ -61,6 +62,7 @@ #define X86_CR4_OSXMMEXCPT 0x00000400 /* enable unmasked SSE exceptions */ #define X86_CR4_VMXE 0x00002000 /* enable VMX virtualization */ #define X86_CR4_RDWRGSFS 0x00010000 /* enable RDWRGSFS support */ +#define X86_CR4_PCIDE 0x00020000 /* enable PCID support */ #define X86_CR4_OSXSAVE 0x00040000 /* enable xsave and xrestore */ #define X86_CR4_SMEP 0x00100000 /* enable SMEP support */ diff --git a/arch/x86/include/asm/vmx.h b/arch/x86/include/asm/vmx.h index de007c272730..74fcb963595b 100644 --- a/arch/x86/include/asm/vmx.h +++ b/arch/x86/include/asm/vmx.h @@ -60,6 +60,7 @@ #define SECONDARY_EXEC_WBINVD_EXITING 0x00000040 #define SECONDARY_EXEC_UNRESTRICTED_GUEST 0x00000080 #define SECONDARY_EXEC_PAUSE_LOOP_EXITING 0x00000400 +#define SECONDARY_EXEC_ENABLE_INVPCID 0x00001000 #define PIN_BASED_EXT_INTR_MASK 0x00000001 @@ -281,6 +282,7 @@ enum vmcs_field { #define EXIT_REASON_EPT_MISCONFIG 49 #define EXIT_REASON_WBINVD 54 #define EXIT_REASON_XSETBV 55 +#define EXIT_REASON_INVPCID 58 /* * Interruption-information format diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c index 197afd53e3a4..0595f1397b7c 100644 --- a/arch/x86/kvm/cpuid.c +++ b/arch/x86/kvm/cpuid.c @@ -201,6 +201,7 @@ static int do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function, unsigned f_lm = 0; #endif unsigned f_rdtscp = kvm_x86_ops->rdtscp_supported() ? F(RDTSCP) : 0; + unsigned f_invpcid = kvm_x86_ops->invpcid_supported() ? F(INVPCID) : 0; /* cpuid 1.edx */ const u32 kvm_supported_word0_x86_features = @@ -228,7 +229,7 @@ static int do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function, 0 /* DS-CPL, VMX, SMX, EST */ | 0 /* TM2 */ | F(SSSE3) | 0 /* CNXT-ID */ | 0 /* Reserved */ | F(FMA) | F(CX16) | 0 /* xTPR Update, PDCM */ | - 0 /* Reserved, DCA */ | F(XMM4_1) | + F(PCID) | 0 /* Reserved, DCA */ | F(XMM4_1) | F(XMM4_2) | F(X2APIC) | F(MOVBE) | F(POPCNT) | 0 /* Reserved*/ | F(AES) | F(XSAVE) | 0 /* OSXSAVE */ | F(AVX) | F(F16C) | F(RDRAND); @@ -248,7 +249,7 @@ static int do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function, /* cpuid 7.0.ebx */ const u32 kvm_supported_word9_x86_features = F(FSGSBASE) | F(BMI1) | F(HLE) | F(AVX2) | F(SMEP) | - F(BMI2) | F(ERMS) | F(RTM); + F(BMI2) | F(ERMS) | f_invpcid | F(RTM); /* all calls to cpuid_count() should be made on the same cpu */ get_cpu(); diff --git a/arch/x86/kvm/cpuid.h b/arch/x86/kvm/cpuid.h index f449edc35e24..a10e46016851 100644 --- a/arch/x86/kvm/cpuid.h +++ b/arch/x86/kvm/cpuid.h @@ -52,4 +52,12 @@ static inline bool guest_cpuid_has_osvw(struct kvm_vcpu *vcpu) return best && (best->ecx & bit(X86_FEATURE_OSVW)); } +static inline bool guest_cpuid_has_pcid(struct kvm_vcpu *vcpu) +{ + struct kvm_cpuid_entry2 *best; + + best = kvm_find_cpuid_entry(vcpu, 1, 0); + return best && (best->ecx & bit(X86_FEATURE_PCID)); +} + #endif diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c index 7a418783259d..baead950d6c8 100644 --- a/arch/x86/kvm/svm.c +++ b/arch/x86/kvm/svm.c @@ -4044,6 +4044,11 @@ static bool svm_rdtscp_supported(void) return false; } +static bool svm_invpcid_supported(void) +{ + return false; +} + static bool svm_has_wbinvd_exit(void) { return true; @@ -4312,6 +4317,7 @@ static struct kvm_x86_ops svm_x86_ops = { .cpuid_update = svm_cpuid_update, .rdtscp_supported = svm_rdtscp_supported, + .invpcid_supported = svm_invpcid_supported, .set_supported_cpuid = svm_set_supported_cpuid, diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 5c52a6d29908..c39b60707e02 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -861,6 +861,12 @@ static inline bool cpu_has_vmx_rdtscp(void) SECONDARY_EXEC_RDTSCP; } +static inline bool cpu_has_vmx_invpcid(void) +{ + return vmcs_config.cpu_based_2nd_exec_ctrl & + SECONDARY_EXEC_ENABLE_INVPCID; +} + static inline bool cpu_has_virtual_nmis(void) { return vmcs_config.pin_based_exec_ctrl & PIN_BASED_VIRTUAL_NMIS; @@ -1751,6 +1757,11 @@ static bool vmx_rdtscp_supported(void) return cpu_has_vmx_rdtscp(); } +static bool vmx_invpcid_supported(void) +{ + return cpu_has_vmx_invpcid() && enable_ept; +} + /* * Swap MSR entry in host/guest MSR entry array. */ @@ -2470,7 +2481,8 @@ static __init int setup_vmcs_config(struct vmcs_config *vmcs_conf) SECONDARY_EXEC_ENABLE_EPT | SECONDARY_EXEC_UNRESTRICTED_GUEST | SECONDARY_EXEC_PAUSE_LOOP_EXITING | - SECONDARY_EXEC_RDTSCP; + SECONDARY_EXEC_RDTSCP | + SECONDARY_EXEC_ENABLE_INVPCID; if (adjust_vmx_controls(min2, opt2, MSR_IA32_VMX_PROCBASED_CTLS2, &_cpu_based_2nd_exec_control) < 0) @@ -3800,6 +3812,8 @@ static u32 vmx_secondary_exec_control(struct vcpu_vmx *vmx) if (!enable_ept) { exec_control &= ~SECONDARY_EXEC_ENABLE_EPT; enable_unrestricted_guest = 0; + /* Enable INVPCID for non-ept guests may cause performance regression. */ + exec_control &= ~SECONDARY_EXEC_ENABLE_INVPCID; } if (!enable_unrestricted_guest) exec_control &= ~SECONDARY_EXEC_UNRESTRICTED_GUEST; @@ -6550,6 +6564,23 @@ static void vmx_cpuid_update(struct kvm_vcpu *vcpu) } } } + + exec_control = vmcs_read32(SECONDARY_VM_EXEC_CONTROL); + /* Exposing INVPCID only when PCID is exposed */ + best = kvm_find_cpuid_entry(vcpu, 0x7, 0); + if (vmx_invpcid_supported() && + best && (best->ecx & bit(X86_FEATURE_INVPCID)) && + guest_cpuid_has_pcid(vcpu)) { + exec_control |= SECONDARY_EXEC_ENABLE_INVPCID; + vmcs_write32(SECONDARY_VM_EXEC_CONTROL, + exec_control); + } else { + exec_control &= ~SECONDARY_EXEC_ENABLE_INVPCID; + vmcs_write32(SECONDARY_VM_EXEC_CONTROL, + exec_control); + if (best) + best->ecx &= ~bit(X86_FEATURE_INVPCID); + } } static void vmx_set_supported_cpuid(u32 func, struct kvm_cpuid_entry2 *entry) @@ -7284,6 +7315,7 @@ static struct kvm_x86_ops vmx_x86_ops = { .cpuid_update = vmx_cpuid_update, .rdtscp_supported = vmx_rdtscp_supported, + .invpcid_supported = vmx_invpcid_supported, .set_supported_cpuid = vmx_set_supported_cpuid, diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index ff0b487e7254..59b59508ff07 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -528,6 +528,9 @@ int kvm_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0) return 1; } + if (!(cr0 & X86_CR0_PG) && kvm_read_cr4_bits(vcpu, X86_CR4_PCIDE)) + return 1; + kvm_x86_ops->set_cr0(vcpu, cr0); if ((cr0 ^ old_cr0) & X86_CR0_PG) { @@ -604,10 +607,20 @@ int kvm_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4) kvm_read_cr3(vcpu))) return 1; + if ((cr4 & X86_CR4_PCIDE) && !(old_cr4 & X86_CR4_PCIDE)) { + if (!guest_cpuid_has_pcid(vcpu)) + return 1; + + /* PCID can not be enabled when cr3[11:0]!=000H or EFER.LMA=0 */ + if ((kvm_read_cr3(vcpu) & X86_CR3_PCID_MASK) || !is_long_mode(vcpu)) + return 1; + } + if (kvm_x86_ops->set_cr4(vcpu, cr4)) return 1; - if ((cr4 ^ old_cr4) & pdptr_bits) + if (((cr4 ^ old_cr4) & pdptr_bits) || + (!(cr4 & X86_CR4_PCIDE) && (old_cr4 & X86_CR4_PCIDE))) kvm_mmu_reset_context(vcpu); if ((cr4 ^ old_cr4) & X86_CR4_OSXSAVE) @@ -626,8 +639,12 @@ int kvm_set_cr3(struct kvm_vcpu *vcpu, unsigned long cr3) } if (is_long_mode(vcpu)) { - if (cr3 & CR3_L_MODE_RESERVED_BITS) - return 1; + if (kvm_read_cr4(vcpu) & X86_CR4_PCIDE) { + if (cr3 & CR3_PCID_ENABLED_RESERVED_BITS) + return 1; + } else + if (cr3 & CR3_L_MODE_RESERVED_BITS) + return 1; } else { if (is_pae(vcpu)) { if (cr3 & CR3_PAE_RESERVED_BITS) -- cgit v1.2.3 From 1a577b72475d161b6677c05abe57301362023bb2 Mon Sep 17 00:00:00 2001 From: "Michael S. Tsirkin" Date: Thu, 19 Jul 2012 13:45:20 +0300 Subject: KVM: fix race with level interrupts When more than 1 source id is in use for the same GSI, we have the following race related to handling irq_states race: CPU 0 clears bit 0. CPU 0 read irq_state as 0. CPU 1 sets level to 1. CPU 1 calls kvm_ioapic_set_irq(1). CPU 0 calls kvm_ioapic_set_irq(0). Now ioapic thinks the level is 0 but irq_state is not 0. Fix by performing all irq_states bitmap handling under pic/ioapic lock. This also removes the need for atomics with irq_states handling. Reported-by: Gleb Natapov Signed-off-by: Michael S. Tsirkin Signed-off-by: Marcelo Tosatti --- arch/x86/include/asm/kvm_host.h | 15 ++++++++++++++- arch/x86/kvm/i8259.c | 17 ++++++++++++++--- virt/kvm/ioapic.c | 19 ++++++++++++++++--- virt/kvm/ioapic.h | 4 +++- virt/kvm/irq_comm.c | 31 ++++--------------------------- 5 files changed, 51 insertions(+), 35 deletions(-) (limited to 'arch/x86/include/asm/kvm_host.h') diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index a3e9409e90b6..2c75b400e40c 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -816,7 +816,20 @@ int kvm_read_guest_page_mmu(struct kvm_vcpu *vcpu, struct kvm_mmu *mmu, void kvm_propagate_fault(struct kvm_vcpu *vcpu, struct x86_exception *fault); bool kvm_require_cpl(struct kvm_vcpu *vcpu, int required_cpl); -int kvm_pic_set_irq(void *opaque, int irq, int level); +static inline int __kvm_irq_line_state(unsigned long *irq_state, + int irq_source_id, int level) +{ + /* Logical OR for level trig interrupt */ + if (level) + __set_bit(irq_source_id, irq_state); + else + __clear_bit(irq_source_id, irq_state); + + return !!(*irq_state); +} + +int kvm_pic_set_irq(struct kvm_pic *pic, int irq, int irq_source_id, int level); +void kvm_pic_clear_all(struct kvm_pic *pic, int irq_source_id); void kvm_inject_nmi(struct kvm_vcpu *vcpu); diff --git a/arch/x86/kvm/i8259.c b/arch/x86/kvm/i8259.c index 81cf4fa4a2be..1df8fb9e1d5d 100644 --- a/arch/x86/kvm/i8259.c +++ b/arch/x86/kvm/i8259.c @@ -188,14 +188,15 @@ void kvm_pic_update_irq(struct kvm_pic *s) pic_unlock(s); } -int kvm_pic_set_irq(void *opaque, int irq, int level) +int kvm_pic_set_irq(struct kvm_pic *s, int irq, int irq_source_id, int level) { - struct kvm_pic *s = opaque; int ret = -1; pic_lock(s); if (irq >= 0 && irq < PIC_NUM_PINS) { - ret = pic_set_irq1(&s->pics[irq >> 3], irq & 7, level); + int irq_level = __kvm_irq_line_state(&s->irq_states[irq], + irq_source_id, level); + ret = pic_set_irq1(&s->pics[irq >> 3], irq & 7, irq_level); pic_update_irq(s); trace_kvm_pic_set_irq(irq >> 3, irq & 7, s->pics[irq >> 3].elcr, s->pics[irq >> 3].imr, ret == 0); @@ -205,6 +206,16 @@ int kvm_pic_set_irq(void *opaque, int irq, int level) return ret; } +void kvm_pic_clear_all(struct kvm_pic *s, int irq_source_id) +{ + int i; + + pic_lock(s); + for (i = 0; i < PIC_NUM_PINS; i++) + __clear_bit(irq_source_id, &s->irq_states[i]); + pic_unlock(s); +} + /* * acknowledge interrupt 'irq' */ diff --git a/virt/kvm/ioapic.c b/virt/kvm/ioapic.c index 26fd54dc459e..ef61d529a6c4 100644 --- a/virt/kvm/ioapic.c +++ b/virt/kvm/ioapic.c @@ -191,7 +191,8 @@ static int ioapic_deliver(struct kvm_ioapic *ioapic, int irq) return kvm_irq_delivery_to_apic(ioapic->kvm, NULL, &irqe); } -int kvm_ioapic_set_irq(struct kvm_ioapic *ioapic, int irq, int level) +int kvm_ioapic_set_irq(struct kvm_ioapic *ioapic, int irq, int irq_source_id, + int level) { u32 old_irr; u32 mask = 1 << irq; @@ -201,9 +202,11 @@ int kvm_ioapic_set_irq(struct kvm_ioapic *ioapic, int irq, int level) spin_lock(&ioapic->lock); old_irr = ioapic->irr; if (irq >= 0 && irq < IOAPIC_NUM_PINS) { + int irq_level = __kvm_irq_line_state(&ioapic->irq_states[irq], + irq_source_id, level); entry = ioapic->redirtbl[irq]; - level ^= entry.fields.polarity; - if (!level) + irq_level ^= entry.fields.polarity; + if (!irq_level) ioapic->irr &= ~mask; else { int edge = (entry.fields.trig_mode == IOAPIC_EDGE_TRIG); @@ -221,6 +224,16 @@ int kvm_ioapic_set_irq(struct kvm_ioapic *ioapic, int irq, int level) return ret; } +void kvm_ioapic_clear_all(struct kvm_ioapic *ioapic, int irq_source_id) +{ + int i; + + spin_lock(&ioapic->lock); + for (i = 0; i < KVM_IOAPIC_NUM_PINS; i++) + __clear_bit(irq_source_id, &ioapic->irq_states[i]); + spin_unlock(&ioapic->lock); +} + static void __kvm_ioapic_update_eoi(struct kvm_ioapic *ioapic, int vector, int trigger_mode) { diff --git a/virt/kvm/ioapic.h b/virt/kvm/ioapic.h index 32872a09b63f..a30abfe6ed16 100644 --- a/virt/kvm/ioapic.h +++ b/virt/kvm/ioapic.h @@ -74,7 +74,9 @@ void kvm_ioapic_update_eoi(struct kvm *kvm, int vector, int trigger_mode); bool kvm_ioapic_handles_vector(struct kvm *kvm, int vector); int kvm_ioapic_init(struct kvm *kvm); void kvm_ioapic_destroy(struct kvm *kvm); -int kvm_ioapic_set_irq(struct kvm_ioapic *ioapic, int irq, int level); +int kvm_ioapic_set_irq(struct kvm_ioapic *ioapic, int irq, int irq_source_id, + int level); +void kvm_ioapic_clear_all(struct kvm_ioapic *ioapic, int irq_source_id); void kvm_ioapic_reset(struct kvm_ioapic *ioapic); int kvm_irq_delivery_to_apic(struct kvm *kvm, struct kvm_lapic *src, struct kvm_lapic_irq *irq); diff --git a/virt/kvm/irq_comm.c b/virt/kvm/irq_comm.c index a6a0365475ed..cc59c68da032 100644 --- a/virt/kvm/irq_comm.c +++ b/virt/kvm/irq_comm.c @@ -33,26 +33,12 @@ #include "ioapic.h" -static inline int kvm_irq_line_state(unsigned long *irq_state, - int irq_source_id, int level) -{ - /* Logical OR for level trig interrupt */ - if (level) - set_bit(irq_source_id, irq_state); - else - clear_bit(irq_source_id, irq_state); - - return !!(*irq_state); -} - static int kvm_set_pic_irq(struct kvm_kernel_irq_routing_entry *e, struct kvm *kvm, int irq_source_id, int level) { #ifdef CONFIG_X86 struct kvm_pic *pic = pic_irqchip(kvm); - level = kvm_irq_line_state(&pic->irq_states[e->irqchip.pin], - irq_source_id, level); - return kvm_pic_set_irq(pic, e->irqchip.pin, level); + return kvm_pic_set_irq(pic, e->irqchip.pin, irq_source_id, level); #else return -1; #endif @@ -62,10 +48,7 @@ static int kvm_set_ioapic_irq(struct kvm_kernel_irq_routing_entry *e, struct kvm *kvm, int irq_source_id, int level) { struct kvm_ioapic *ioapic = kvm->arch.vioapic; - level = kvm_irq_line_state(&ioapic->irq_states[e->irqchip.pin], - irq_source_id, level); - - return kvm_ioapic_set_irq(ioapic, e->irqchip.pin, level); + return kvm_ioapic_set_irq(ioapic, e->irqchip.pin, irq_source_id, level); } inline static bool kvm_is_dm_lowest_prio(struct kvm_lapic_irq *irq) @@ -249,8 +232,6 @@ unlock: void kvm_free_irq_source_id(struct kvm *kvm, int irq_source_id) { - int i; - ASSERT(irq_source_id != KVM_USERSPACE_IRQ_SOURCE_ID); mutex_lock(&kvm->irq_lock); @@ -263,14 +244,10 @@ void kvm_free_irq_source_id(struct kvm *kvm, int irq_source_id) if (!irqchip_in_kernel(kvm)) goto unlock; - for (i = 0; i < KVM_IOAPIC_NUM_PINS; i++) { - clear_bit(irq_source_id, &kvm->arch.vioapic->irq_states[i]); - if (i >= 16) - continue; + kvm_ioapic_clear_all(kvm->arch.vioapic, irq_source_id); #ifdef CONFIG_X86 - clear_bit(irq_source_id, &pic_irqchip(kvm)->irq_states[i]); + kvm_pic_clear_all(pic_irqchip(kvm), irq_source_id); #endif - } unlock: mutex_unlock(&kvm->irq_lock); } -- cgit v1.2.3