From d60d8b64280c8b36c085eda7821585c1ce911795 Mon Sep 17 00:00:00 2001 From: Christoffer Dall Date: Fri, 26 Jan 2018 16:06:51 +0100 Subject: KVM: arm/arm64: Fix arch timers with userspace irqchips When introducing support for irqchip in userspace we needed a way to mask the timer signal to prevent the guest continuously exiting due to a screaming timer. We did this by disabling the corresponding percpu interrupt on the host interrupt controller, because we cannot rely on the host system having a GIC, and therefore cannot make any assumptions about having an active state to hide the timer signal. Unfortunately, when introducing this feature, it became entirely possible that a VCPU which belongs to a VM that has a userspace irqchip can disable the vtimer irq on the host on some physical CPU, and then go away without ever enabling the vtimer irq on that physical CPU again. This means that using irqchips in userspace on a system that also supports running VMs with an in-kernel GIC can prevent forward progress from in-kernel GIC VMs. Later on, when we started taking virtual timer interrupts in the arch timer code, we would also leave this timer state active for userspace irqchip VMs, because we leave it up to a VGIC-enabled guest to deactivate the hardware IRQ using the HW bit in the LR. Both issues are solved by only using the enable/disable trick on systems that do not have a host GIC which supports the active state, because all VMs on such systems must use irqchips in userspace. Systems that have a working GIC with support for an active state use the active state to mask the timer signal for both userspace and in-kernel irqchips. Cc: Alexander Graf Reviewed-by: Marc Zyngier Cc: # v4.12+ Fixes: d9e139778376 ("KVM: arm/arm64: Support arch timers with a userspace gic") Signed-off-by: Christoffer Dall --- virt/kvm/arm/arch_timer.c | 116 +++++++++++++++++++++++++--------------------- 1 file changed, 64 insertions(+), 52 deletions(-) (limited to 'virt') diff --git a/virt/kvm/arm/arch_timer.c b/virt/kvm/arm/arch_timer.c index 70268c0bec79..70f4c30918eb 100644 --- a/virt/kvm/arm/arch_timer.c +++ b/virt/kvm/arm/arch_timer.c @@ -36,6 +36,8 @@ static struct timecounter *timecounter; static unsigned int host_vtimer_irq; static u32 host_vtimer_irq_flags; +static DEFINE_STATIC_KEY_FALSE(has_gic_active_state); + static const struct kvm_irq_level default_ptimer_irq = { .irq = 30, .level = 1, @@ -56,6 +58,12 @@ u64 kvm_phys_timer_read(void) return timecounter->cc->read(timecounter->cc); } +static inline bool userspace_irqchip(struct kvm *kvm) +{ + return static_branch_unlikely(&userspace_irqchip_in_use) && + unlikely(!irqchip_in_kernel(kvm)); +} + static void soft_timer_start(struct hrtimer *hrt, u64 ns) { hrtimer_start(hrt, ktime_add_ns(ktime_get(), ns), @@ -69,25 +77,6 @@ static void soft_timer_cancel(struct hrtimer *hrt, struct work_struct *work) cancel_work_sync(work); } -static void kvm_vtimer_update_mask_user(struct kvm_vcpu *vcpu) -{ - struct arch_timer_context *vtimer = vcpu_vtimer(vcpu); - - /* - * When using a userspace irqchip with the architected timers, we must - * prevent continuously exiting from the guest, and therefore mask the - * physical interrupt by disabling it on the host interrupt controller - * when the virtual level is high, such that the guest can make - * forward progress. Once we detect the output level being - * de-asserted, we unmask the interrupt again so that we exit from the - * guest when the timer fires. - */ - if (vtimer->irq.level) - disable_percpu_irq(host_vtimer_irq); - else - enable_percpu_irq(host_vtimer_irq, 0); -} - static irqreturn_t kvm_arch_timer_handler(int irq, void *dev_id) { struct kvm_vcpu *vcpu = *(struct kvm_vcpu **)dev_id; @@ -106,9 +95,9 @@ static irqreturn_t kvm_arch_timer_handler(int irq, void *dev_id) if (kvm_timer_should_fire(vtimer)) kvm_timer_update_irq(vcpu, true, vtimer); - if (static_branch_unlikely(&userspace_irqchip_in_use) && - unlikely(!irqchip_in_kernel(vcpu->kvm))) - kvm_vtimer_update_mask_user(vcpu); + if (userspace_irqchip(vcpu->kvm) && + !static_branch_unlikely(&has_gic_active_state)) + disable_percpu_irq(host_vtimer_irq); return IRQ_HANDLED; } @@ -290,8 +279,7 @@ static void kvm_timer_update_irq(struct kvm_vcpu *vcpu, bool new_level, trace_kvm_timer_update_irq(vcpu->vcpu_id, timer_ctx->irq.irq, timer_ctx->irq.level); - if (!static_branch_unlikely(&userspace_irqchip_in_use) || - likely(irqchip_in_kernel(vcpu->kvm))) { + if (!userspace_irqchip(vcpu->kvm)) { ret = kvm_vgic_inject_irq(vcpu->kvm, vcpu->vcpu_id, timer_ctx->irq.irq, timer_ctx->irq.level, @@ -350,12 +338,6 @@ static void kvm_timer_update_state(struct kvm_vcpu *vcpu) phys_timer_emulate(vcpu); } -static void __timer_snapshot_state(struct arch_timer_context *timer) -{ - timer->cnt_ctl = read_sysreg_el0(cntv_ctl); - timer->cnt_cval = read_sysreg_el0(cntv_cval); -} - static void vtimer_save_state(struct kvm_vcpu *vcpu) { struct arch_timer_cpu *timer = &vcpu->arch.timer_cpu; @@ -367,8 +349,10 @@ static void vtimer_save_state(struct kvm_vcpu *vcpu) if (!vtimer->loaded) goto out; - if (timer->enabled) - __timer_snapshot_state(vtimer); + if (timer->enabled) { + vtimer->cnt_ctl = read_sysreg_el0(cntv_ctl); + vtimer->cnt_cval = read_sysreg_el0(cntv_cval); + } /* Disable the virtual timer */ write_sysreg_el0(0, cntv_ctl); @@ -460,23 +444,43 @@ static void set_cntvoff(u64 cntvoff) kvm_call_hyp(__kvm_timer_set_cntvoff, low, high); } -static void kvm_timer_vcpu_load_vgic(struct kvm_vcpu *vcpu) +static inline void set_vtimer_irq_phys_active(struct kvm_vcpu *vcpu, bool active) +{ + int r; + r = irq_set_irqchip_state(host_vtimer_irq, IRQCHIP_STATE_ACTIVE, active); + WARN_ON(r); +} + +static void kvm_timer_vcpu_load_gic(struct kvm_vcpu *vcpu) { struct arch_timer_context *vtimer = vcpu_vtimer(vcpu); bool phys_active; - int ret; - phys_active = kvm_vgic_map_is_active(vcpu, vtimer->irq.irq); - - ret = irq_set_irqchip_state(host_vtimer_irq, - IRQCHIP_STATE_ACTIVE, - phys_active); - WARN_ON(ret); + if (irqchip_in_kernel(vcpu->kvm)) + phys_active = kvm_vgic_map_is_active(vcpu, vtimer->irq.irq); + else + phys_active = vtimer->irq.level; + set_vtimer_irq_phys_active(vcpu, phys_active); } -static void kvm_timer_vcpu_load_user(struct kvm_vcpu *vcpu) +static void kvm_timer_vcpu_load_nogic(struct kvm_vcpu *vcpu) { - kvm_vtimer_update_mask_user(vcpu); + struct arch_timer_context *vtimer = vcpu_vtimer(vcpu); + + /* + * When using a userspace irqchip with the architected timers and a + * host interrupt controller that doesn't support an active state, we + * must still prevent continuously exiting from the guest, and + * therefore mask the physical interrupt by disabling it on the host + * interrupt controller when the virtual level is high, such that the + * guest can make forward progress. Once we detect the output level + * being de-asserted, we unmask the interrupt again so that we exit + * from the guest when the timer fires. + */ + if (vtimer->irq.level) + disable_percpu_irq(host_vtimer_irq); + else + enable_percpu_irq(host_vtimer_irq, host_vtimer_irq_flags); } void kvm_timer_vcpu_load(struct kvm_vcpu *vcpu) @@ -487,10 +491,10 @@ void kvm_timer_vcpu_load(struct kvm_vcpu *vcpu) if (unlikely(!timer->enabled)) return; - if (unlikely(!irqchip_in_kernel(vcpu->kvm))) - kvm_timer_vcpu_load_user(vcpu); + if (static_branch_likely(&has_gic_active_state)) + kvm_timer_vcpu_load_gic(vcpu); else - kvm_timer_vcpu_load_vgic(vcpu); + kvm_timer_vcpu_load_nogic(vcpu); set_cntvoff(vtimer->cntvoff); @@ -555,18 +559,24 @@ static void unmask_vtimer_irq_user(struct kvm_vcpu *vcpu) { struct arch_timer_context *vtimer = vcpu_vtimer(vcpu); - if (unlikely(!irqchip_in_kernel(vcpu->kvm))) { - __timer_snapshot_state(vtimer); - if (!kvm_timer_should_fire(vtimer)) { - kvm_timer_update_irq(vcpu, false, vtimer); - kvm_vtimer_update_mask_user(vcpu); - } + if (!kvm_timer_should_fire(vtimer)) { + kvm_timer_update_irq(vcpu, false, vtimer); + if (static_branch_likely(&has_gic_active_state)) + set_vtimer_irq_phys_active(vcpu, false); + else + enable_percpu_irq(host_vtimer_irq, host_vtimer_irq_flags); } } void kvm_timer_sync_hwstate(struct kvm_vcpu *vcpu) { - unmask_vtimer_irq_user(vcpu); + struct arch_timer_cpu *timer = &vcpu->arch.timer_cpu; + + if (unlikely(!timer->enabled)) + return; + + if (unlikely(!irqchip_in_kernel(vcpu->kvm))) + unmask_vtimer_irq_user(vcpu); } int kvm_timer_vcpu_reset(struct kvm_vcpu *vcpu) @@ -753,6 +763,8 @@ int kvm_timer_hyp_init(bool has_gic) kvm_err("kvm_arch_timer: error setting vcpu affinity\n"); goto out_free_irq; } + + static_branch_enable(&has_gic_active_state); } kvm_info("virtual timer IRQ%d\n", host_vtimer_irq); -- cgit v1.2.3 From 62b06f8f429cd233e4e2e7bbd21081ad60c9018f Mon Sep 17 00:00:00 2001 From: Andre Przywara Date: Tue, 6 Mar 2018 09:21:06 +0000 Subject: KVM: arm/arm64: vgic: Add missing irq_lock to vgic_mmio_read_pending Our irq_is_pending() helper function accesses multiple members of the vgic_irq struct, so we need to hold the lock when calling it. Add that requirement as a comment to the definition and take the lock around the call in vgic_mmio_read_pending(), where we were missing it before. Fixes: 96b298000db4 ("KVM: arm/arm64: vgic-new: Add PENDING registers handlers") Signed-off-by: Andre Przywara Signed-off-by: Marc Zyngier --- virt/kvm/arm/vgic/vgic-mmio.c | 3 +++ virt/kvm/arm/vgic/vgic.h | 1 + 2 files changed, 4 insertions(+) (limited to 'virt') diff --git a/virt/kvm/arm/vgic/vgic-mmio.c b/virt/kvm/arm/vgic/vgic-mmio.c index 83d82bd7dc4e..dbe99d635c80 100644 --- a/virt/kvm/arm/vgic/vgic-mmio.c +++ b/virt/kvm/arm/vgic/vgic-mmio.c @@ -113,9 +113,12 @@ unsigned long vgic_mmio_read_pending(struct kvm_vcpu *vcpu, /* Loop over all IRQs affected by this read */ for (i = 0; i < len * 8; i++) { struct vgic_irq *irq = vgic_get_irq(vcpu->kvm, vcpu, intid + i); + unsigned long flags; + spin_lock_irqsave(&irq->irq_lock, flags); if (irq_is_pending(irq)) value |= (1U << i); + spin_unlock_irqrestore(&irq->irq_lock, flags); vgic_put_irq(vcpu->kvm, irq); } diff --git a/virt/kvm/arm/vgic/vgic.h b/virt/kvm/arm/vgic/vgic.h index 12c37b89f7a3..5b11859a1a1e 100644 --- a/virt/kvm/arm/vgic/vgic.h +++ b/virt/kvm/arm/vgic/vgic.h @@ -96,6 +96,7 @@ /* we only support 64 kB translation table page size */ #define KVM_ITS_L1E_ADDR_MASK GENMASK_ULL(51, 16) +/* Requires the irq_lock to be held by the caller. */ static inline bool irq_is_pending(struct vgic_irq *irq) { if (irq->config == VGIC_CONFIG_EDGE) -- cgit v1.2.3 From e21a4f3a930cda6e4902cb5b3213365e5ff3ce7c Mon Sep 17 00:00:00 2001 From: Christoffer Dall Date: Tue, 27 Feb 2018 12:33:50 +0100 Subject: KVM: arm/arm64: Avoid vcpu_load for other vcpu ioctls than KVM_RUN Calling vcpu_load() registers preempt notifiers for this vcpu and calls kvm_arch_vcpu_load(). The latter will soon be doing a lot of heavy lifting on arm/arm64 and will try to do things such as enabling the virtual timer and setting us up to handle interrupts from the timer hardware. Loading state onto hardware registers and enabling hardware to signal interrupts can be problematic when we're not actually about to run the VCPU, because it makes it difficult to establish the right context when handling interrupts from the timer, and it makes the register access code difficult to reason about. Luckily, now when we call vcpu_load in each ioctl implementation, we can simply remove the call from the non-KVM_RUN vcpu ioctls, and our kvm_arch_vcpu_load() is only used for loading vcpu content to the physical CPU when we're actually going to run the vcpu. Cc: stable@vger.kernel.org Fixes: 9b062471e52a ("KVM: Move vcpu_load to arch-specific kvm_arch_vcpu_ioctl") Reviewed-by: Julien Grall Reviewed-by: Marc Zyngier Reviewed-by: Andrew Jones Signed-off-by: Christoffer Dall Signed-off-by: Marc Zyngier --- arch/arm64/kvm/guest.c | 3 --- virt/kvm/arm/arm.c | 9 --------- 2 files changed, 12 deletions(-) (limited to 'virt') diff --git a/arch/arm64/kvm/guest.c b/arch/arm64/kvm/guest.c index d7e3299a7734..959e50d2588c 100644 --- a/arch/arm64/kvm/guest.c +++ b/arch/arm64/kvm/guest.c @@ -363,8 +363,6 @@ int kvm_arch_vcpu_ioctl_set_guest_debug(struct kvm_vcpu *vcpu, { int ret = 0; - vcpu_load(vcpu); - trace_kvm_set_guest_debug(vcpu, dbg->control); if (dbg->control & ~KVM_GUESTDBG_VALID_MASK) { @@ -386,7 +384,6 @@ int kvm_arch_vcpu_ioctl_set_guest_debug(struct kvm_vcpu *vcpu, } out: - vcpu_put(vcpu); return ret; } diff --git a/virt/kvm/arm/arm.c b/virt/kvm/arm/arm.c index 86941f6181bb..53572304843b 100644 --- a/virt/kvm/arm/arm.c +++ b/virt/kvm/arm/arm.c @@ -384,14 +384,11 @@ static void vcpu_power_off(struct kvm_vcpu *vcpu) int kvm_arch_vcpu_ioctl_get_mpstate(struct kvm_vcpu *vcpu, struct kvm_mp_state *mp_state) { - vcpu_load(vcpu); - if (vcpu->arch.power_off) mp_state->mp_state = KVM_MP_STATE_STOPPED; else mp_state->mp_state = KVM_MP_STATE_RUNNABLE; - vcpu_put(vcpu); return 0; } @@ -400,8 +397,6 @@ int kvm_arch_vcpu_ioctl_set_mpstate(struct kvm_vcpu *vcpu, { int ret = 0; - vcpu_load(vcpu); - switch (mp_state->mp_state) { case KVM_MP_STATE_RUNNABLE: vcpu->arch.power_off = false; @@ -413,7 +408,6 @@ int kvm_arch_vcpu_ioctl_set_mpstate(struct kvm_vcpu *vcpu, ret = -EINVAL; } - vcpu_put(vcpu); return ret; } @@ -1036,8 +1030,6 @@ long kvm_arch_vcpu_ioctl(struct file *filp, struct kvm_device_attr attr; long r; - vcpu_load(vcpu); - switch (ioctl) { case KVM_ARM_VCPU_INIT: { struct kvm_vcpu_init init; @@ -1114,7 +1106,6 @@ long kvm_arch_vcpu_ioctl(struct file *filp, r = -EINVAL; } - vcpu_put(vcpu); return r; } -- cgit v1.2.3 From 413aa807ae39fed7e387c175d2d0ae9fcf6c0c9d Mon Sep 17 00:00:00 2001 From: Christoffer Dall Date: Mon, 5 Mar 2018 11:36:38 +0100 Subject: KVM: arm/arm64: Reset mapped IRQs on VM reset We currently don't allow resetting mapped IRQs from userspace, because their state is controlled by the hardware. But we do need to reset the state when the VM is reset, so we provide a function for the 'owner' of the mapped interrupt to reset the interrupt state. Currently only the timer uses mapped interrupts, so we call this function from the timer reset logic. Cc: stable@vger.kernel.org Fixes: 4c60e360d6df ("KVM: arm/arm64: Provide a get_input_level for the arch timer") Signed-off-by: Christoffer Dall Signed-off-by: Marc Zyngier --- include/kvm/arm_vgic.h | 1 + virt/kvm/arm/arch_timer.c | 4 ++++ virt/kvm/arm/vgic/vgic.c | 26 ++++++++++++++++++++++++++ 3 files changed, 31 insertions(+) (limited to 'virt') diff --git a/include/kvm/arm_vgic.h b/include/kvm/arm_vgic.h index cdbd142ca7f2..02924ae2527e 100644 --- a/include/kvm/arm_vgic.h +++ b/include/kvm/arm_vgic.h @@ -360,6 +360,7 @@ void kvm_vgic_put(struct kvm_vcpu *vcpu); bool kvm_vcpu_has_pending_irqs(struct kvm_vcpu *vcpu); void kvm_vgic_sync_hwstate(struct kvm_vcpu *vcpu); void kvm_vgic_flush_hwstate(struct kvm_vcpu *vcpu); +void kvm_vgic_reset_mapped_irq(struct kvm_vcpu *vcpu, u32 vintid); void vgic_v3_dispatch_sgi(struct kvm_vcpu *vcpu, u64 reg); diff --git a/virt/kvm/arm/arch_timer.c b/virt/kvm/arm/arch_timer.c index 70f4c30918eb..3945021510a9 100644 --- a/virt/kvm/arm/arch_timer.c +++ b/virt/kvm/arm/arch_timer.c @@ -581,6 +581,7 @@ void kvm_timer_sync_hwstate(struct kvm_vcpu *vcpu) int kvm_timer_vcpu_reset(struct kvm_vcpu *vcpu) { + struct arch_timer_cpu *timer = &vcpu->arch.timer_cpu; struct arch_timer_context *vtimer = vcpu_vtimer(vcpu); struct arch_timer_context *ptimer = vcpu_ptimer(vcpu); @@ -594,6 +595,9 @@ int kvm_timer_vcpu_reset(struct kvm_vcpu *vcpu) ptimer->cnt_ctl = 0; kvm_timer_update_state(vcpu); + if (timer->enabled && irqchip_in_kernel(vcpu->kvm)) + kvm_vgic_reset_mapped_irq(vcpu, vtimer->irq.irq); + return 0; } diff --git a/virt/kvm/arm/vgic/vgic.c b/virt/kvm/arm/vgic/vgic.c index c7c5ef190afa..0001858a2c23 100644 --- a/virt/kvm/arm/vgic/vgic.c +++ b/virt/kvm/arm/vgic/vgic.c @@ -495,6 +495,32 @@ int kvm_vgic_map_phys_irq(struct kvm_vcpu *vcpu, unsigned int host_irq, return ret; } +/** + * kvm_vgic_reset_mapped_irq - Reset a mapped IRQ + * @vcpu: The VCPU pointer + * @vintid: The INTID of the interrupt + * + * Reset the active and pending states of a mapped interrupt. Kernel + * subsystems injecting mapped interrupts should reset their interrupt lines + * when we are doing a reset of the VM. + */ +void kvm_vgic_reset_mapped_irq(struct kvm_vcpu *vcpu, u32 vintid) +{ + struct vgic_irq *irq = vgic_get_irq(vcpu->kvm, vcpu, vintid); + unsigned long flags; + + if (!irq->hw) + goto out; + + spin_lock_irqsave(&irq->irq_lock, flags); + irq->active = false; + irq->pending_latch = false; + irq->line_level = false; + spin_unlock_irqrestore(&irq->irq_lock, flags); +out: + vgic_put_irq(vcpu->kvm, irq); +} + int kvm_vgic_unmap_phys_irq(struct kvm_vcpu *vcpu, unsigned int vintid) { struct vgic_irq *irq; -- cgit v1.2.3 From 76600428c3677659e3c3633bb4f2ea302220a275 Mon Sep 17 00:00:00 2001 From: Ard Biesheuvel Date: Fri, 2 Mar 2018 08:16:30 +0000 Subject: KVM: arm/arm64: Reduce verbosity of KVM init log On my GICv3 system, the following is printed to the kernel log at boot: kvm [1]: 8-bit VMID kvm [1]: IDMAP page: d20e35000 kvm [1]: HYP VA range: 800000000000:ffffffffffff kvm [1]: vgic-v2@2c020000 kvm [1]: GIC system register CPU interface enabled kvm [1]: vgic interrupt IRQ1 kvm [1]: virtual timer IRQ4 kvm [1]: Hyp mode initialized successfully The KVM IDMAP is a mapping of a statically allocated kernel structure, and so printing its physical address leaks the physical placement of the kernel when physical KASLR in effect. So change the kvm_info() to kvm_debug() to remove it from the log output. While at it, trim the output a bit more: IRQ numbers can be found in /proc/interrupts, and the HYP VA and vgic-v2 lines are not highly informational either. Cc: Acked-by: Will Deacon Acked-by: Christoffer Dall Signed-off-by: Ard Biesheuvel Signed-off-by: Marc Zyngier --- virt/kvm/arm/arch_timer.c | 2 +- virt/kvm/arm/mmu.c | 6 +++--- virt/kvm/arm/vgic/vgic-v2.c | 2 +- 3 files changed, 5 insertions(+), 5 deletions(-) (limited to 'virt') diff --git a/virt/kvm/arm/arch_timer.c b/virt/kvm/arm/arch_timer.c index 3945021510a9..282389eb204f 100644 --- a/virt/kvm/arm/arch_timer.c +++ b/virt/kvm/arm/arch_timer.c @@ -771,7 +771,7 @@ int kvm_timer_hyp_init(bool has_gic) static_branch_enable(&has_gic_active_state); } - kvm_info("virtual timer IRQ%d\n", host_vtimer_irq); + kvm_debug("virtual timer IRQ%d\n", host_vtimer_irq); cpuhp_setup_state(CPUHP_AP_KVM_ARM_TIMER_STARTING, "kvm/arm/timer:starting", kvm_timer_starting_cpu, diff --git a/virt/kvm/arm/mmu.c b/virt/kvm/arm/mmu.c index ec62d1cccab7..b960acdd0c05 100644 --- a/virt/kvm/arm/mmu.c +++ b/virt/kvm/arm/mmu.c @@ -1810,9 +1810,9 @@ int kvm_mmu_init(void) */ BUG_ON((hyp_idmap_start ^ (hyp_idmap_end - 1)) & PAGE_MASK); - kvm_info("IDMAP page: %lx\n", hyp_idmap_start); - kvm_info("HYP VA range: %lx:%lx\n", - kern_hyp_va(PAGE_OFFSET), kern_hyp_va(~0UL)); + kvm_debug("IDMAP page: %lx\n", hyp_idmap_start); + kvm_debug("HYP VA range: %lx:%lx\n", + kern_hyp_va(PAGE_OFFSET), kern_hyp_va(~0UL)); if (hyp_idmap_start >= kern_hyp_va(PAGE_OFFSET) && hyp_idmap_start < kern_hyp_va(~0UL) && diff --git a/virt/kvm/arm/vgic/vgic-v2.c b/virt/kvm/arm/vgic/vgic-v2.c index c32d7b93ffd1..e9d840a75e7b 100644 --- a/virt/kvm/arm/vgic/vgic-v2.c +++ b/virt/kvm/arm/vgic/vgic-v2.c @@ -410,7 +410,7 @@ int vgic_v2_probe(const struct gic_kvm_info *info) kvm_vgic_global_state.type = VGIC_V2; kvm_vgic_global_state.max_gic_vcpus = VGIC_V2_MAX_CPUS; - kvm_info("vgic-v2@%llx\n", info->vctrl.start); + kvm_debug("vgic-v2@%llx\n", info->vctrl.start); return 0; out: -- cgit v1.2.3 From 16ca6a607d84bef0129698d8d808f501afd08d43 Mon Sep 17 00:00:00 2001 From: Marc Zyngier Date: Tue, 6 Mar 2018 21:48:01 +0000 Subject: KVM: arm/arm64: vgic: Don't populate multiple LRs with the same vintid The vgic code is trying to be clever when injecting GICv2 SGIs, and will happily populate LRs with the same interrupt number if they come from multiple vcpus (after all, they are distinct interrupt sources). Unfortunately, this is against the letter of the architecture, and the GICv2 architecture spec says "Each valid interrupt stored in the List registers must have a unique VirtualID for that virtual CPU interface.". GICv3 has similar (although slightly ambiguous) restrictions. This results in guests locking up when using GICv2-on-GICv3, for example. The obvious fix is to stop trying so hard, and inject a single vcpu per SGI per guest entry. After all, pending SGIs with multiple source vcpus are pretty rare, and are mostly seen in scenario where the physical CPUs are severely overcomitted. But as we now only inject a single instance of a multi-source SGI per vcpu entry, we may delay those interrupts for longer than strictly necessary, and run the risk of injecting lower priority interrupts in the meantime. In order to address this, we adopt a three stage strategy: - If we encounter a multi-source SGI in the AP list while computing its depth, we force the list to be sorted - When populating the LRs, we prevent the injection of any interrupt of lower priority than that of the first multi-source SGI we've injected. - Finally, the injection of a multi-source SGI triggers the request of a maintenance interrupt when there will be no pending interrupt in the LRs (HCR_NPIE). At the point where the last pending interrupt in the LRs switches from Pending to Active, the maintenance interrupt will be delivered, allowing us to add the remaining SGIs using the same process. Cc: stable@vger.kernel.org Fixes: 0919e84c0fc1 ("KVM: arm/arm64: vgic-new: Add IRQ sync/flush framework") Acked-by: Christoffer Dall Signed-off-by: Marc Zyngier --- include/linux/irqchip/arm-gic-v3.h | 1 + include/linux/irqchip/arm-gic.h | 1 + virt/kvm/arm/vgic/vgic-v2.c | 9 +++++- virt/kvm/arm/vgic/vgic-v3.c | 9 +++++- virt/kvm/arm/vgic/vgic.c | 61 +++++++++++++++++++++++++++++--------- virt/kvm/arm/vgic/vgic.h | 2 ++ 6 files changed, 67 insertions(+), 16 deletions(-) (limited to 'virt') diff --git a/include/linux/irqchip/arm-gic-v3.h b/include/linux/irqchip/arm-gic-v3.h index c00c4c33e432..b26eccc78fb1 100644 --- a/include/linux/irqchip/arm-gic-v3.h +++ b/include/linux/irqchip/arm-gic-v3.h @@ -503,6 +503,7 @@ #define ICH_HCR_EN (1 << 0) #define ICH_HCR_UIE (1 << 1) +#define ICH_HCR_NPIE (1 << 3) #define ICH_HCR_TC (1 << 10) #define ICH_HCR_TALL0 (1 << 11) #define ICH_HCR_TALL1 (1 << 12) diff --git a/include/linux/irqchip/arm-gic.h b/include/linux/irqchip/arm-gic.h index d3453ee072fc..68d8b1f73682 100644 --- a/include/linux/irqchip/arm-gic.h +++ b/include/linux/irqchip/arm-gic.h @@ -84,6 +84,7 @@ #define GICH_HCR_EN (1 << 0) #define GICH_HCR_UIE (1 << 1) +#define GICH_HCR_NPIE (1 << 3) #define GICH_LR_VIRTUALID (0x3ff << 0) #define GICH_LR_PHYSID_CPUID_SHIFT (10) diff --git a/virt/kvm/arm/vgic/vgic-v2.c b/virt/kvm/arm/vgic/vgic-v2.c index e9d840a75e7b..29556f71b691 100644 --- a/virt/kvm/arm/vgic/vgic-v2.c +++ b/virt/kvm/arm/vgic/vgic-v2.c @@ -37,6 +37,13 @@ void vgic_v2_init_lrs(void) vgic_v2_write_lr(i, 0); } +void vgic_v2_set_npie(struct kvm_vcpu *vcpu) +{ + struct vgic_v2_cpu_if *cpuif = &vcpu->arch.vgic_cpu.vgic_v2; + + cpuif->vgic_hcr |= GICH_HCR_NPIE; +} + void vgic_v2_set_underflow(struct kvm_vcpu *vcpu) { struct vgic_v2_cpu_if *cpuif = &vcpu->arch.vgic_cpu.vgic_v2; @@ -64,7 +71,7 @@ void vgic_v2_fold_lr_state(struct kvm_vcpu *vcpu) int lr; unsigned long flags; - cpuif->vgic_hcr &= ~GICH_HCR_UIE; + cpuif->vgic_hcr &= ~(GICH_HCR_UIE | GICH_HCR_NPIE); for (lr = 0; lr < vgic_cpu->used_lrs; lr++) { u32 val = cpuif->vgic_lr[lr]; diff --git a/virt/kvm/arm/vgic/vgic-v3.c b/virt/kvm/arm/vgic/vgic-v3.c index 6b329414e57a..0ff2006f3781 100644 --- a/virt/kvm/arm/vgic/vgic-v3.c +++ b/virt/kvm/arm/vgic/vgic-v3.c @@ -26,6 +26,13 @@ static bool group1_trap; static bool common_trap; static bool gicv4_enable; +void vgic_v3_set_npie(struct kvm_vcpu *vcpu) +{ + struct vgic_v3_cpu_if *cpuif = &vcpu->arch.vgic_cpu.vgic_v3; + + cpuif->vgic_hcr |= ICH_HCR_NPIE; +} + void vgic_v3_set_underflow(struct kvm_vcpu *vcpu) { struct vgic_v3_cpu_if *cpuif = &vcpu->arch.vgic_cpu.vgic_v3; @@ -47,7 +54,7 @@ void vgic_v3_fold_lr_state(struct kvm_vcpu *vcpu) int lr; unsigned long flags; - cpuif->vgic_hcr &= ~ICH_HCR_UIE; + cpuif->vgic_hcr &= ~(ICH_HCR_UIE | ICH_HCR_NPIE); for (lr = 0; lr < vgic_cpu->used_lrs; lr++) { u64 val = cpuif->vgic_lr[lr]; diff --git a/virt/kvm/arm/vgic/vgic.c b/virt/kvm/arm/vgic/vgic.c index 0001858a2c23..8201899126f6 100644 --- a/virt/kvm/arm/vgic/vgic.c +++ b/virt/kvm/arm/vgic/vgic.c @@ -710,22 +710,37 @@ static inline void vgic_set_underflow(struct kvm_vcpu *vcpu) vgic_v3_set_underflow(vcpu); } +static inline void vgic_set_npie(struct kvm_vcpu *vcpu) +{ + if (kvm_vgic_global_state.type == VGIC_V2) + vgic_v2_set_npie(vcpu); + else + vgic_v3_set_npie(vcpu); +} + /* Requires the ap_list_lock to be held. */ -static int compute_ap_list_depth(struct kvm_vcpu *vcpu) +static int compute_ap_list_depth(struct kvm_vcpu *vcpu, + bool *multi_sgi) { struct vgic_cpu *vgic_cpu = &vcpu->arch.vgic_cpu; struct vgic_irq *irq; int count = 0; + *multi_sgi = false; + DEBUG_SPINLOCK_BUG_ON(!spin_is_locked(&vgic_cpu->ap_list_lock)); list_for_each_entry(irq, &vgic_cpu->ap_list_head, ap_list) { spin_lock(&irq->irq_lock); /* GICv2 SGIs can count for more than one... */ - if (vgic_irq_is_sgi(irq->intid) && irq->source) - count += hweight8(irq->source); - else + if (vgic_irq_is_sgi(irq->intid) && irq->source) { + int w = hweight8(irq->source); + + count += w; + *multi_sgi |= (w > 1); + } else { count++; + } spin_unlock(&irq->irq_lock); } return count; @@ -736,28 +751,43 @@ static void vgic_flush_lr_state(struct kvm_vcpu *vcpu) { struct vgic_cpu *vgic_cpu = &vcpu->arch.vgic_cpu; struct vgic_irq *irq; - int count = 0; + int count; + bool npie = false; + bool multi_sgi; + u8 prio = 0xff; DEBUG_SPINLOCK_BUG_ON(!spin_is_locked(&vgic_cpu->ap_list_lock)); - if (compute_ap_list_depth(vcpu) > kvm_vgic_global_state.nr_lr) + count = compute_ap_list_depth(vcpu, &multi_sgi); + if (count > kvm_vgic_global_state.nr_lr || multi_sgi) vgic_sort_ap_list(vcpu); + count = 0; + list_for_each_entry(irq, &vgic_cpu->ap_list_head, ap_list) { spin_lock(&irq->irq_lock); - if (unlikely(vgic_target_oracle(irq) != vcpu)) - goto next; - /* - * If we get an SGI with multiple sources, try to get - * them in all at once. + * If we have multi-SGIs in the pipeline, we need to + * guarantee that they are all seen before any IRQ of + * lower priority. In that case, we need to filter out + * these interrupts by exiting early. This is easy as + * the AP list has been sorted already. */ - do { + if (multi_sgi && irq->priority > prio) { + spin_unlock(&irq->irq_lock); + break; + } + + if (likely(vgic_target_oracle(irq) == vcpu)) { vgic_populate_lr(vcpu, irq, count++); - } while (irq->source && count < kvm_vgic_global_state.nr_lr); -next: + if (irq->source) { + npie = true; + prio = irq->priority; + } + } + spin_unlock(&irq->irq_lock); if (count == kvm_vgic_global_state.nr_lr) { @@ -768,6 +798,9 @@ next: } } + if (npie) + vgic_set_npie(vcpu); + vcpu->arch.vgic_cpu.used_lrs = count; /* Nuke remaining LRs */ diff --git a/virt/kvm/arm/vgic/vgic.h b/virt/kvm/arm/vgic/vgic.h index 5b11859a1a1e..f5b8519e5546 100644 --- a/virt/kvm/arm/vgic/vgic.h +++ b/virt/kvm/arm/vgic/vgic.h @@ -160,6 +160,7 @@ void vgic_v2_fold_lr_state(struct kvm_vcpu *vcpu); void vgic_v2_populate_lr(struct kvm_vcpu *vcpu, struct vgic_irq *irq, int lr); void vgic_v2_clear_lr(struct kvm_vcpu *vcpu, int lr); void vgic_v2_set_underflow(struct kvm_vcpu *vcpu); +void vgic_v2_set_npie(struct kvm_vcpu *vcpu); int vgic_v2_has_attr_regs(struct kvm_device *dev, struct kvm_device_attr *attr); int vgic_v2_dist_uaccess(struct kvm_vcpu *vcpu, bool is_write, int offset, u32 *val); @@ -189,6 +190,7 @@ void vgic_v3_fold_lr_state(struct kvm_vcpu *vcpu); void vgic_v3_populate_lr(struct kvm_vcpu *vcpu, struct vgic_irq *irq, int lr); void vgic_v3_clear_lr(struct kvm_vcpu *vcpu, int lr); void vgic_v3_set_underflow(struct kvm_vcpu *vcpu); +void vgic_v3_set_npie(struct kvm_vcpu *vcpu); void vgic_v3_set_vmcr(struct kvm_vcpu *vcpu, struct vgic_vmcr *vmcr); void vgic_v3_get_vmcr(struct kvm_vcpu *vcpu, struct vgic_vmcr *vmcr); void vgic_v3_enable(struct kvm_vcpu *vcpu); -- cgit v1.2.3 From 27e91ad1e746e341ca2312f29bccb9736be7b476 Mon Sep 17 00:00:00 2001 From: Marc Zyngier Date: Tue, 6 Mar 2018 21:44:37 +0000 Subject: kvm: arm/arm64: vgic-v3: Tighten synchronization for guests using v2 on v3 On guest exit, and when using GICv2 on GICv3, we use a dsb(st) to force synchronization between the memory-mapped guest view and the system-register view that the hypervisor uses. This is incorrect, as the spec calls out the need for "a DSB whose required access type is both loads and stores with any Shareability attribute", while we're only synchronizing stores. We also lack an isb after the dsb to ensure that the latter has actually been executed before we start reading stuff from the sysregs. The fix is pretty easy: turn dsb(st) into dsb(sy), and slap an isb() just after. Cc: stable@vger.kernel.org Fixes: f68d2b1b73cc ("arm64: KVM: Implement vgic-v3 save/restore") Acked-by: Christoffer Dall Reviewed-by: Andre Przywara Signed-off-by: Marc Zyngier --- virt/kvm/arm/hyp/vgic-v3-sr.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'virt') diff --git a/virt/kvm/arm/hyp/vgic-v3-sr.c b/virt/kvm/arm/hyp/vgic-v3-sr.c index f5c3d6d7019e..b89ce5432214 100644 --- a/virt/kvm/arm/hyp/vgic-v3-sr.c +++ b/virt/kvm/arm/hyp/vgic-v3-sr.c @@ -215,7 +215,8 @@ void __hyp_text __vgic_v3_save_state(struct kvm_vcpu *vcpu) * are now visible to the system register interface. */ if (!cpu_if->vgic_sre) { - dsb(st); + dsb(sy); + isb(); cpu_if->vgic_vmcr = read_gicreg(ICH_VMCR_EL2); } -- cgit v1.2.3