From 4b4b4512da2a844b8da2585609b67fae1ce4f4db Mon Sep 17 00:00:00 2001 From: Christoffer Dall Date: Sun, 30 Aug 2015 15:01:27 +0200 Subject: arm/arm64: KVM: Rework the arch timer to use level-triggered semantics The arch timer currently uses edge-triggered semantics in the sense that the line is never sampled by the vgic and lowering the line from the timer to the vgic doesn't have any effect on the pending state of virtual interrupts in the vgic. This means that we do not support a guest with the otherwise valid behavior of (1) disable interrupts (2) enable the timer (3) disable the timer (4) enable interrupts. Such a guest would validly not expect to see any interrupts on real hardware, but will see interrupts on KVM. This patch fixes this shortcoming through the following series of changes. First, we change the flow of the timer/vgic sync/flush operations. Now the timer is always flushed/synced before the vgic, because the vgic samples the state of the timer output. This has the implication that we move the timer operations in to non-preempible sections, but that is fine after the previous commit getting rid of hrtimer schedules on every entry/exit. Second, we change the internal behavior of the timer, letting the timer keep track of its previous output state, and only lower/raise the line to the vgic when the state changes. Note that in theory this could have been accomplished more simply by signalling the vgic every time the state *potentially* changed, but we don't want to be hitting the vgic more often than necessary. Third, we get rid of the use of the map->active field in the vgic and instead simply set the interrupt as active on the physical distributor whenever the input to the GIC is asserted and conversely clear the physical active state when the input to the GIC is deasserted. Fourth, and finally, we now initialize the timer PPIs (and all the other unused PPIs for now), to be level-triggered, and modify the sync code to sample the line state on HW sync and re-inject a new interrupt if it is still pending at that time. Signed-off-by: Christoffer Dall --- include/kvm/arm_vgic.h | 3 --- 1 file changed, 3 deletions(-) (limited to 'include/kvm/arm_vgic.h') diff --git a/include/kvm/arm_vgic.h b/include/kvm/arm_vgic.h index 4e14dac282bb..7bc5d0224ab0 100644 --- a/include/kvm/arm_vgic.h +++ b/include/kvm/arm_vgic.h @@ -159,7 +159,6 @@ struct irq_phys_map { u32 virt_irq; u32 phys_irq; u32 irq; - bool active; }; struct irq_phys_map_entry { @@ -354,8 +353,6 @@ int kvm_vgic_vcpu_active_irq(struct kvm_vcpu *vcpu); struct irq_phys_map *kvm_vgic_map_phys_irq(struct kvm_vcpu *vcpu, int virt_irq, int irq); int kvm_vgic_unmap_phys_irq(struct kvm_vcpu *vcpu, struct irq_phys_map *map); -bool kvm_vgic_get_phys_irq_active(struct irq_phys_map *map); -void kvm_vgic_set_phys_irq_active(struct irq_phys_map *map, bool active); #define irqchip_in_kernel(k) (!!((k)->arch.vgic.in_kernel)) #define vgic_initialized(k) (!!((k)->arch.vgic.nr_cpus)) -- cgit v1.2.3 From 5fdf876d30ce7c9b63045b3db8374912ab9b262c Mon Sep 17 00:00:00 2001 From: Michal Marek Date: Thu, 15 Oct 2015 22:16:28 +0200 Subject: KVM: arm: Do not indent the arguments of DECLARE_BITMAP Besides being a coding style issue, it confuses make tags: ctags: Warning: include/kvm/arm_vgic.h:307: null expansion of name pattern "\1" ctags: Warning: include/kvm/arm_vgic.h:308: null expansion of name pattern "\1" ctags: Warning: include/kvm/arm_vgic.h:309: null expansion of name pattern "\1" ctags: Warning: include/kvm/arm_vgic.h:317: null expansion of name pattern "\1" Cc: kvmarm@lists.cs.columbia.edu Signed-off-by: Michal Marek Signed-off-by: Christoffer Dall --- include/kvm/arm_vgic.h | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'include/kvm/arm_vgic.h') diff --git a/include/kvm/arm_vgic.h b/include/kvm/arm_vgic.h index 7bc5d0224ab0..8065801a1847 100644 --- a/include/kvm/arm_vgic.h +++ b/include/kvm/arm_vgic.h @@ -299,9 +299,9 @@ struct vgic_cpu { u8 *vgic_irq_lr_map; /* Pending/active/both interrupts on this VCPU */ - DECLARE_BITMAP( pending_percpu, VGIC_NR_PRIVATE_IRQS); - DECLARE_BITMAP( active_percpu, VGIC_NR_PRIVATE_IRQS); - DECLARE_BITMAP( pend_act_percpu, VGIC_NR_PRIVATE_IRQS); + DECLARE_BITMAP(pending_percpu, VGIC_NR_PRIVATE_IRQS); + DECLARE_BITMAP(active_percpu, VGIC_NR_PRIVATE_IRQS); + DECLARE_BITMAP(pend_act_percpu, VGIC_NR_PRIVATE_IRQS); /* Pending/active/both shared interrupts, dynamically sized */ unsigned long *pending_shared; @@ -309,7 +309,7 @@ struct vgic_cpu { unsigned long *pend_act_shared; /* Bitmap of used/free list registers */ - DECLARE_BITMAP( lr_used, VGIC_V2_MAX_LRS); + DECLARE_BITMAP(lr_used, VGIC_V2_MAX_LRS); /* Number of list registers on this CPU */ int nr_lr; -- cgit v1.2.3 From c4cd4c168b81dad53e659d18cdae653bc0ec2384 Mon Sep 17 00:00:00 2001 From: Pavel Fedin Date: Tue, 27 Oct 2015 11:37:29 +0300 Subject: KVM: arm/arm64: Optimize away redundant LR tracking Currently we use vgic_irq_lr_map in order to track which LRs hold which IRQs, and lr_used bitmap in order to track which LRs are used or free. vgic_irq_lr_map is actually used only for piggy-back optimization, and can be easily replaced by iteration over lr_used. This is good because in future, when LPI support is introduced, number of IRQs will grow up to at least 16384, while numbers from 1024 to 8192 are never going to be used. This would be a huge memory waste. In its turn, lr_used is also completely redundant since ae705930fca6322600690df9dc1c7d0516145a93 ("arm/arm64: KVM: Keep elrsr/aisr in sync with software model"), because together with lr_used we also update elrsr. This allows to easily replace lr_used with elrsr, inverting all conditions (because in elrsr '1' means 'free'). Signed-off-by: Pavel Fedin Signed-off-by: Christoffer Dall --- include/kvm/arm_vgic.h | 6 ------ virt/kvm/arm/vgic-v2.c | 1 + virt/kvm/arm/vgic-v3.c | 1 + virt/kvm/arm/vgic.c | 53 ++++++++++++++------------------------------------ 4 files changed, 17 insertions(+), 44 deletions(-) (limited to 'include/kvm/arm_vgic.h') diff --git a/include/kvm/arm_vgic.h b/include/kvm/arm_vgic.h index 8065801a1847..3936bf802e1d 100644 --- a/include/kvm/arm_vgic.h +++ b/include/kvm/arm_vgic.h @@ -295,9 +295,6 @@ struct vgic_v3_cpu_if { }; struct vgic_cpu { - /* per IRQ to LR mapping */ - u8 *vgic_irq_lr_map; - /* Pending/active/both interrupts on this VCPU */ DECLARE_BITMAP(pending_percpu, VGIC_NR_PRIVATE_IRQS); DECLARE_BITMAP(active_percpu, VGIC_NR_PRIVATE_IRQS); @@ -308,9 +305,6 @@ struct vgic_cpu { unsigned long *active_shared; unsigned long *pend_act_shared; - /* Bitmap of used/free list registers */ - DECLARE_BITMAP(lr_used, VGIC_V2_MAX_LRS); - /* Number of list registers on this CPU */ int nr_lr; diff --git a/virt/kvm/arm/vgic-v2.c b/virt/kvm/arm/vgic-v2.c index 8d7b04db8471..c0f5d7fad9ea 100644 --- a/virt/kvm/arm/vgic-v2.c +++ b/virt/kvm/arm/vgic-v2.c @@ -158,6 +158,7 @@ static void vgic_v2_enable(struct kvm_vcpu *vcpu) * anyway. */ vcpu->arch.vgic_cpu.vgic_v2.vgic_vmcr = 0; + vcpu->arch.vgic_cpu.vgic_v2.vgic_elrsr = ~0; /* Get the show on the road... */ vcpu->arch.vgic_cpu.vgic_v2.vgic_hcr = GICH_HCR_EN; diff --git a/virt/kvm/arm/vgic-v3.c b/virt/kvm/arm/vgic-v3.c index 7dd5d62f10a1..92003cb61a0a 100644 --- a/virt/kvm/arm/vgic-v3.c +++ b/virt/kvm/arm/vgic-v3.c @@ -193,6 +193,7 @@ static void vgic_v3_enable(struct kvm_vcpu *vcpu) * anyway. */ vgic_v3->vgic_vmcr = 0; + vgic_v3->vgic_elrsr = ~0; /* * If we are emulating a GICv3, we do it in an non-GICv2-compatible diff --git a/virt/kvm/arm/vgic.c b/virt/kvm/arm/vgic.c index d4669eb29b77..265a41035728 100644 --- a/virt/kvm/arm/vgic.c +++ b/virt/kvm/arm/vgic.c @@ -108,6 +108,7 @@ static void vgic_retire_disabled_irqs(struct kvm_vcpu *vcpu); static void vgic_retire_lr(int lr_nr, int irq, struct kvm_vcpu *vcpu); static struct vgic_lr vgic_get_lr(const struct kvm_vcpu *vcpu, int lr); static void vgic_set_lr(struct kvm_vcpu *vcpu, int lr, struct vgic_lr lr_desc); +static u64 vgic_get_elrsr(struct kvm_vcpu *vcpu); static struct irq_phys_map *vgic_irq_map_search(struct kvm_vcpu *vcpu, int virt_irq); static int compute_pending_for_cpu(struct kvm_vcpu *vcpu); @@ -691,9 +692,11 @@ bool vgic_handle_cfg_reg(u32 *reg, struct kvm_exit_mmio *mmio, void vgic_unqueue_irqs(struct kvm_vcpu *vcpu) { struct vgic_cpu *vgic_cpu = &vcpu->arch.vgic_cpu; + u64 elrsr = vgic_get_elrsr(vcpu); + unsigned long *elrsr_ptr = u64_to_bitmask(&elrsr); int i; - for_each_set_bit(i, vgic_cpu->lr_used, vgic_cpu->nr_lr) { + for_each_clear_bit(i, elrsr_ptr, vgic_cpu->nr_lr) { struct vgic_lr lr = vgic_get_lr(vcpu, i); /* @@ -1098,7 +1101,6 @@ static inline void vgic_enable(struct kvm_vcpu *vcpu) static void vgic_retire_lr(int lr_nr, int irq, struct kvm_vcpu *vcpu) { - struct vgic_cpu *vgic_cpu = &vcpu->arch.vgic_cpu; struct vgic_lr vlr = vgic_get_lr(vcpu, lr_nr); /* @@ -1112,8 +1114,6 @@ static void vgic_retire_lr(int lr_nr, int irq, struct kvm_vcpu *vcpu) vlr.state = 0; vgic_set_lr(vcpu, lr_nr, vlr); - clear_bit(lr_nr, vgic_cpu->lr_used); - vgic_cpu->vgic_irq_lr_map[irq] = LR_EMPTY; vgic_sync_lr_elrsr(vcpu, lr_nr, vlr); } @@ -1128,10 +1128,11 @@ static void vgic_retire_lr(int lr_nr, int irq, struct kvm_vcpu *vcpu) */ static void vgic_retire_disabled_irqs(struct kvm_vcpu *vcpu) { - struct vgic_cpu *vgic_cpu = &vcpu->arch.vgic_cpu; + u64 elrsr = vgic_get_elrsr(vcpu); + unsigned long *elrsr_ptr = u64_to_bitmask(&elrsr); int lr; - for_each_set_bit(lr, vgic_cpu->lr_used, vgic->nr_lr) { + for_each_clear_bit(lr, elrsr_ptr, vgic->nr_lr) { struct vgic_lr vlr = vgic_get_lr(vcpu, lr); if (!vgic_irq_is_enabled(vcpu, vlr.irq)) { @@ -1188,8 +1189,9 @@ static void vgic_queue_irq_to_lr(struct kvm_vcpu *vcpu, int irq, */ bool vgic_queue_irq(struct kvm_vcpu *vcpu, u8 sgi_source_id, int irq) { - struct vgic_cpu *vgic_cpu = &vcpu->arch.vgic_cpu; struct vgic_dist *dist = &vcpu->kvm->arch.vgic; + u64 elrsr = vgic_get_elrsr(vcpu); + unsigned long *elrsr_ptr = u64_to_bitmask(&elrsr); struct vgic_lr vlr; int lr; @@ -1200,28 +1202,22 @@ bool vgic_queue_irq(struct kvm_vcpu *vcpu, u8 sgi_source_id, int irq) kvm_debug("Queue IRQ%d\n", irq); - lr = vgic_cpu->vgic_irq_lr_map[irq]; - /* Do we have an active interrupt for the same CPUID? */ - if (lr != LR_EMPTY) { + for_each_clear_bit(lr, elrsr_ptr, vgic->nr_lr) { vlr = vgic_get_lr(vcpu, lr); - if (vlr.source == sgi_source_id) { + if (vlr.irq == irq && vlr.source == sgi_source_id) { kvm_debug("LR%d piggyback for IRQ%d\n", lr, vlr.irq); - BUG_ON(!test_bit(lr, vgic_cpu->lr_used)); vgic_queue_irq_to_lr(vcpu, irq, lr, vlr); return true; } } /* Try to use another LR for this interrupt */ - lr = find_first_zero_bit((unsigned long *)vgic_cpu->lr_used, - vgic->nr_lr); + lr = find_first_bit(elrsr_ptr, vgic->nr_lr); if (lr >= vgic->nr_lr) return false; kvm_debug("LR%d allocated for IRQ%d %x\n", lr, irq, sgi_source_id); - vgic_cpu->vgic_irq_lr_map[irq] = lr; - set_bit(lr, vgic_cpu->lr_used); vlr.irq = irq; vlr.source = sgi_source_id; @@ -1456,7 +1452,6 @@ static bool vgic_sync_hwirq(struct kvm_vcpu *vcpu, int lr, struct vgic_lr vlr) /* Sync back the VGIC state after a guest run */ static void __kvm_vgic_sync_hwstate(struct kvm_vcpu *vcpu) { - struct vgic_cpu *vgic_cpu = &vcpu->arch.vgic_cpu; struct vgic_dist *dist = &vcpu->kvm->arch.vgic; u64 elrsr; unsigned long *elrsr_ptr; @@ -1469,22 +1464,10 @@ static void __kvm_vgic_sync_hwstate(struct kvm_vcpu *vcpu) /* Deal with HW interrupts, and clear mappings for empty LRs */ for (lr = 0; lr < vgic->nr_lr; lr++) { - struct vgic_lr vlr; - - if (!test_bit(lr, vgic_cpu->lr_used)) - continue; - - vlr = vgic_get_lr(vcpu, lr); - if (vgic_sync_hwirq(vcpu, lr, vlr)) - level_pending = true; - - if (!test_bit(lr, elrsr_ptr)) - continue; - - clear_bit(lr, vgic_cpu->lr_used); + struct vgic_lr vlr = vgic_get_lr(vcpu, lr); + level_pending |= vgic_sync_hwirq(vcpu, lr, vlr); BUG_ON(vlr.irq >= dist->nr_irqs); - vgic_cpu->vgic_irq_lr_map[vlr.irq] = LR_EMPTY; } /* Check if we still have something up our sleeve... */ @@ -1912,12 +1895,10 @@ void kvm_vgic_vcpu_destroy(struct kvm_vcpu *vcpu) kfree(vgic_cpu->pending_shared); kfree(vgic_cpu->active_shared); kfree(vgic_cpu->pend_act_shared); - kfree(vgic_cpu->vgic_irq_lr_map); vgic_destroy_irq_phys_map(vcpu->kvm, &vgic_cpu->irq_phys_map_list); vgic_cpu->pending_shared = NULL; vgic_cpu->active_shared = NULL; vgic_cpu->pend_act_shared = NULL; - vgic_cpu->vgic_irq_lr_map = NULL; } static int vgic_vcpu_init_maps(struct kvm_vcpu *vcpu, int nr_irqs) @@ -1928,18 +1909,14 @@ static int vgic_vcpu_init_maps(struct kvm_vcpu *vcpu, int nr_irqs) vgic_cpu->pending_shared = kzalloc(sz, GFP_KERNEL); vgic_cpu->active_shared = kzalloc(sz, GFP_KERNEL); vgic_cpu->pend_act_shared = kzalloc(sz, GFP_KERNEL); - vgic_cpu->vgic_irq_lr_map = kmalloc(nr_irqs, GFP_KERNEL); if (!vgic_cpu->pending_shared || !vgic_cpu->active_shared - || !vgic_cpu->pend_act_shared - || !vgic_cpu->vgic_irq_lr_map) { + || !vgic_cpu->pend_act_shared) { kvm_vgic_vcpu_destroy(vcpu); return -ENOMEM; } - memset(vgic_cpu->vgic_irq_lr_map, LR_EMPTY, nr_irqs); - /* * Store the number of LRs per vcpu, so we don't have to go * all the way to the distributor structure to find out. Only -- cgit v1.2.3 From 26caea7693cb99833fe4ecc544c842289d6b3f69 Mon Sep 17 00:00:00 2001 From: Pavel Fedin Date: Tue, 27 Oct 2015 11:37:31 +0300 Subject: KVM: arm/arm64: Merge vgic_set_lr() and vgic_sync_lr_elrsr() Now we see that vgic_set_lr() and vgic_sync_lr_elrsr() are always used together. Merge them into one function, saving from second vgic_ops dereferencing every time. Signed-off-by: Pavel Fedin Signed-off-by: Christoffer Dall --- include/kvm/arm_vgic.h | 1 - virt/kvm/arm/vgic-v2.c | 5 ----- virt/kvm/arm/vgic-v3.c | 5 ----- virt/kvm/arm/vgic.c | 14 ++------------ 4 files changed, 2 insertions(+), 23 deletions(-) (limited to 'include/kvm/arm_vgic.h') diff --git a/include/kvm/arm_vgic.h b/include/kvm/arm_vgic.h index 3936bf802e1d..f62addc17dcf 100644 --- a/include/kvm/arm_vgic.h +++ b/include/kvm/arm_vgic.h @@ -112,7 +112,6 @@ struct vgic_vmcr { struct vgic_ops { struct vgic_lr (*get_lr)(const struct kvm_vcpu *, int); void (*set_lr)(struct kvm_vcpu *, int, struct vgic_lr); - void (*sync_lr_elrsr)(struct kvm_vcpu *, int, struct vgic_lr); u64 (*get_elrsr)(const struct kvm_vcpu *vcpu); u64 (*get_eisr)(const struct kvm_vcpu *vcpu); void (*clear_eisr)(struct kvm_vcpu *vcpu); diff --git a/virt/kvm/arm/vgic-v2.c b/virt/kvm/arm/vgic-v2.c index c0f5d7fad9ea..ff02f08df74d 100644 --- a/virt/kvm/arm/vgic-v2.c +++ b/virt/kvm/arm/vgic-v2.c @@ -79,11 +79,7 @@ static void vgic_v2_set_lr(struct kvm_vcpu *vcpu, int lr, lr_val |= (lr_desc.source << GICH_LR_PHYSID_CPUID_SHIFT); vcpu->arch.vgic_cpu.vgic_v2.vgic_lr[lr] = lr_val; -} -static void vgic_v2_sync_lr_elrsr(struct kvm_vcpu *vcpu, int lr, - struct vgic_lr lr_desc) -{ if (!(lr_desc.state & LR_STATE_MASK)) vcpu->arch.vgic_cpu.vgic_v2.vgic_elrsr |= (1ULL << lr); else @@ -167,7 +163,6 @@ static void vgic_v2_enable(struct kvm_vcpu *vcpu) static const struct vgic_ops vgic_v2_ops = { .get_lr = vgic_v2_get_lr, .set_lr = vgic_v2_set_lr, - .sync_lr_elrsr = vgic_v2_sync_lr_elrsr, .get_elrsr = vgic_v2_get_elrsr, .get_eisr = vgic_v2_get_eisr, .clear_eisr = vgic_v2_clear_eisr, diff --git a/virt/kvm/arm/vgic-v3.c b/virt/kvm/arm/vgic-v3.c index 92003cb61a0a..487d6357b7e7 100644 --- a/virt/kvm/arm/vgic-v3.c +++ b/virt/kvm/arm/vgic-v3.c @@ -112,11 +112,7 @@ static void vgic_v3_set_lr(struct kvm_vcpu *vcpu, int lr, } vcpu->arch.vgic_cpu.vgic_v3.vgic_lr[LR_INDEX(lr)] = lr_val; -} -static void vgic_v3_sync_lr_elrsr(struct kvm_vcpu *vcpu, int lr, - struct vgic_lr lr_desc) -{ if (!(lr_desc.state & LR_STATE_MASK)) vcpu->arch.vgic_cpu.vgic_v3.vgic_elrsr |= (1U << lr); else @@ -212,7 +208,6 @@ static void vgic_v3_enable(struct kvm_vcpu *vcpu) static const struct vgic_ops vgic_v3_ops = { .get_lr = vgic_v3_get_lr, .set_lr = vgic_v3_set_lr, - .sync_lr_elrsr = vgic_v3_sync_lr_elrsr, .get_elrsr = vgic_v3_get_elrsr, .get_eisr = vgic_v3_get_eisr, .clear_eisr = vgic_v3_clear_eisr, diff --git a/virt/kvm/arm/vgic.c b/virt/kvm/arm/vgic.c index 96e45f3da534..fe451d4885ae 100644 --- a/virt/kvm/arm/vgic.c +++ b/virt/kvm/arm/vgic.c @@ -1032,12 +1032,6 @@ static void vgic_set_lr(struct kvm_vcpu *vcpu, int lr, vgic_ops->set_lr(vcpu, lr, vlr); } -static void vgic_sync_lr_elrsr(struct kvm_vcpu *vcpu, int lr, - struct vgic_lr vlr) -{ - vgic_ops->sync_lr_elrsr(vcpu, lr, vlr); -} - static inline u64 vgic_get_elrsr(struct kvm_vcpu *vcpu) { return vgic_ops->get_elrsr(vcpu); @@ -1100,7 +1094,6 @@ static void vgic_retire_lr(int lr_nr, struct kvm_vcpu *vcpu) vlr.state = 0; vgic_set_lr(vcpu, lr_nr, vlr); - vgic_sync_lr_elrsr(vcpu, lr_nr, vlr); } /* @@ -1162,7 +1155,6 @@ static void vgic_queue_irq_to_lr(struct kvm_vcpu *vcpu, int irq, } vgic_set_lr(vcpu, lr_nr, vlr); - vgic_sync_lr_elrsr(vcpu, lr_nr, vlr); } /* @@ -1340,8 +1332,6 @@ static int process_queued_irq(struct kvm_vcpu *vcpu, vlr.hwirq = 0; vgic_set_lr(vcpu, lr, vlr); - vgic_sync_lr_elrsr(vcpu, lr, vlr); - return pending; } @@ -1442,8 +1432,6 @@ static void __kvm_vgic_sync_hwstate(struct kvm_vcpu *vcpu) bool level_pending; level_pending = vgic_process_maintenance(vcpu); - elrsr = vgic_get_elrsr(vcpu); - elrsr_ptr = u64_to_bitmask(&elrsr); /* Deal with HW interrupts, and clear mappings for empty LRs */ for (lr = 0; lr < vgic->nr_lr; lr++) { @@ -1454,6 +1442,8 @@ static void __kvm_vgic_sync_hwstate(struct kvm_vcpu *vcpu) } /* Check if we still have something up our sleeve... */ + elrsr = vgic_get_elrsr(vcpu); + elrsr_ptr = u64_to_bitmask(&elrsr); pending = find_first_zero_bit(elrsr_ptr, vgic->nr_lr); if (level_pending || pending < vgic->nr_lr) set_bit(vcpu->vcpu_id, dist->irq_pending_on_cpu); -- cgit v1.2.3