KVM ARM Archive on lore.kernel.org
 help / color / Atom feed
From: Marc Zyngier <maz@kernel.org>
To: kvmarm@lists.cs.columbia.edu, linux-kernel@vger.kernel.org
Cc: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>,
	Jason Cooper <jason@lakedaemon.net>,
	Robert Richter <rrichter@marvell.com>,
	Jayachandran C <jnair@marvell.com>,
	Thomas Gleixner <tglx@linutronix.de>
Subject: [PATCH v2 01/36] KVM: arm64: vgic-v4: Move the GICv4 residency flow to be driven by vcpu_load/put
Date: Sun, 27 Oct 2019 14:41:59 +0000
Message-ID: <20191027144234.8395-2-maz@kernel.org> (raw)
In-Reply-To: <20191027144234.8395-1-maz@kernel.org>

When the VHE code was reworked, a lot of the vgic stuff was moved around,
but the GICv4 residency code did stay untouched, meaning that we come
in and out of residency on each flush/sync, which is obviously suboptimal.

To address this, let's move things around a bit:

- Residency entry (flush) moves to vcpu_load
- Residency exit (sync) moves to vcpu_put
- On blocking (entry to WFI), we "put"
- On unblocking (exit from WFI, we "load"

Because these can nest (load/block/put/load/unblock/put, for example),
we now have per-VPE tracking of the residency state.

Additionally, vgic_v4_put gains a "need doorbell" parameter, which only
gets set to true when blocking because of a WFI. This allows a finer
control of the doorbell, which now also gets disabled as soon as
it gets signaled.

Signed-off-by: Marc Zyngier <maz@kernel.org>
---
 drivers/irqchip/irq-gic-v4.c       |  7 +++-
 include/kvm/arm_vgic.h             |  4 +--
 include/linux/irqchip/arm-gic-v4.h |  2 ++
 virt/kvm/arm/arm.c                 | 12 ++++---
 virt/kvm/arm/vgic/vgic-v3.c        |  4 +++
 virt/kvm/arm/vgic/vgic-v4.c        | 55 ++++++++++++++----------------
 virt/kvm/arm/vgic/vgic.c           |  4 ---
 virt/kvm/arm/vgic/vgic.h           |  2 --
 8 files changed, 48 insertions(+), 42 deletions(-)

diff --git a/drivers/irqchip/irq-gic-v4.c b/drivers/irqchip/irq-gic-v4.c
index 563e87ed0766..45969927cc81 100644
--- a/drivers/irqchip/irq-gic-v4.c
+++ b/drivers/irqchip/irq-gic-v4.c
@@ -141,12 +141,17 @@ static int its_send_vpe_cmd(struct its_vpe *vpe, struct its_cmd_info *info)
 int its_schedule_vpe(struct its_vpe *vpe, bool on)
 {
 	struct its_cmd_info info;
+	int ret;
 
 	WARN_ON(preemptible());
 
 	info.cmd_type = on ? SCHEDULE_VPE : DESCHEDULE_VPE;
 
-	return its_send_vpe_cmd(vpe, &info);
+	ret = its_send_vpe_cmd(vpe, &info);
+	if (!ret)
+		vpe->resident = on;
+
+	return ret;
 }
 
 int its_invall_vpe(struct its_vpe *vpe)
diff --git a/include/kvm/arm_vgic.h b/include/kvm/arm_vgic.h
index af4f09c02bf1..4dc58d7a0010 100644
--- a/include/kvm/arm_vgic.h
+++ b/include/kvm/arm_vgic.h
@@ -396,7 +396,7 @@ int kvm_vgic_v4_set_forwarding(struct kvm *kvm, int irq,
 int kvm_vgic_v4_unset_forwarding(struct kvm *kvm, int irq,
 				 struct kvm_kernel_irq_routing_entry *irq_entry);
 
-void kvm_vgic_v4_enable_doorbell(struct kvm_vcpu *vcpu);
-void kvm_vgic_v4_disable_doorbell(struct kvm_vcpu *vcpu);
+int vgic_v4_load(struct kvm_vcpu *vcpu);
+int vgic_v4_put(struct kvm_vcpu *vcpu, bool need_db);
 
 #endif /* __KVM_ARM_VGIC_H */
diff --git a/include/linux/irqchip/arm-gic-v4.h b/include/linux/irqchip/arm-gic-v4.h
index e6b155713b47..ab1396afe08a 100644
--- a/include/linux/irqchip/arm-gic-v4.h
+++ b/include/linux/irqchip/arm-gic-v4.h
@@ -35,6 +35,8 @@ struct its_vpe {
 	/* Doorbell interrupt */
 	int			irq;
 	irq_hw_number_t		vpe_db_lpi;
+	/* VPE resident */
+	bool			resident;
 	/* VPE proxy mapping */
 	int			vpe_proxy_event;
 	/*
diff --git a/virt/kvm/arm/arm.c b/virt/kvm/arm/arm.c
index 86c6aa1cb58e..bd2afcf9a13f 100644
--- a/virt/kvm/arm/arm.c
+++ b/virt/kvm/arm/arm.c
@@ -322,20 +322,24 @@ void kvm_arch_vcpu_blocking(struct kvm_vcpu *vcpu)
 	/*
 	 * If we're about to block (most likely because we've just hit a
 	 * WFI), we need to sync back the state of the GIC CPU interface
-	 * so that we have the lastest PMR and group enables. This ensures
+	 * so that we have the latest PMR and group enables. This ensures
 	 * that kvm_arch_vcpu_runnable has up-to-date data to decide
 	 * whether we have pending interrupts.
+	 *
+	 * For the same reason, we want to tell GICv4 that we need
+	 * doorbells to be signalled, should an interrupt become pending.
 	 */
 	preempt_disable();
 	kvm_vgic_vmcr_sync(vcpu);
+	vgic_v4_put(vcpu, true);
 	preempt_enable();
-
-	kvm_vgic_v4_enable_doorbell(vcpu);
 }
 
 void kvm_arch_vcpu_unblocking(struct kvm_vcpu *vcpu)
 {
-	kvm_vgic_v4_disable_doorbell(vcpu);
+	preempt_disable();
+	vgic_v4_load(vcpu);
+	preempt_enable();
 }
 
 int kvm_arch_vcpu_init(struct kvm_vcpu *vcpu)
diff --git a/virt/kvm/arm/vgic/vgic-v3.c b/virt/kvm/arm/vgic/vgic-v3.c
index 8d69f007dd0c..48307a9eb1d8 100644
--- a/virt/kvm/arm/vgic/vgic-v3.c
+++ b/virt/kvm/arm/vgic/vgic-v3.c
@@ -664,6 +664,8 @@ void vgic_v3_load(struct kvm_vcpu *vcpu)
 
 	if (has_vhe())
 		__vgic_v3_activate_traps(vcpu);
+
+	WARN_ON(vgic_v4_load(vcpu));
 }
 
 void vgic_v3_vmcr_sync(struct kvm_vcpu *vcpu)
@@ -676,6 +678,8 @@ void vgic_v3_vmcr_sync(struct kvm_vcpu *vcpu)
 
 void vgic_v3_put(struct kvm_vcpu *vcpu)
 {
+	WARN_ON(vgic_v4_put(vcpu, false));
+
 	vgic_v3_vmcr_sync(vcpu);
 
 	kvm_call_hyp(__vgic_v3_save_aprs, vcpu);
diff --git a/virt/kvm/arm/vgic/vgic-v4.c b/virt/kvm/arm/vgic/vgic-v4.c
index 477af6aebb97..7e1f3202968a 100644
--- a/virt/kvm/arm/vgic/vgic-v4.c
+++ b/virt/kvm/arm/vgic/vgic-v4.c
@@ -85,6 +85,10 @@ static irqreturn_t vgic_v4_doorbell_handler(int irq, void *info)
 {
 	struct kvm_vcpu *vcpu = info;
 
+	/* We got the message, no need to fire again */
+	if (!irqd_irq_disabled(&irq_to_desc(irq)->irq_data))
+		disable_irq_nosync(irq);
+
 	vcpu->arch.vgic_cpu.vgic_v3.its_vpe.pending_last = true;
 	kvm_make_request(KVM_REQ_IRQ_PENDING, vcpu);
 	kvm_vcpu_kick(vcpu);
@@ -192,20 +196,30 @@ void vgic_v4_teardown(struct kvm *kvm)
 	its_vm->vpes = NULL;
 }
 
-int vgic_v4_sync_hwstate(struct kvm_vcpu *vcpu)
+int vgic_v4_put(struct kvm_vcpu *vcpu, bool need_db)
 {
-	if (!vgic_supports_direct_msis(vcpu->kvm))
+	struct its_vpe *vpe = &vcpu->arch.vgic_cpu.vgic_v3.its_vpe;
+	struct irq_desc *desc = irq_to_desc(vpe->irq);
+
+	if (!vgic_supports_direct_msis(vcpu->kvm) || !vpe->resident)
 		return 0;
 
-	return its_schedule_vpe(&vcpu->arch.vgic_cpu.vgic_v3.its_vpe, false);
+	/*
+	 * If blocking, a doorbell is required. Undo the nested
+	 * disable_irq() calls...
+	 */
+	while (need_db && irqd_irq_disabled(&desc->irq_data))
+		enable_irq(vpe->irq);
+
+	return its_schedule_vpe(vpe, false);
 }
 
-int vgic_v4_flush_hwstate(struct kvm_vcpu *vcpu)
+int vgic_v4_load(struct kvm_vcpu *vcpu)
 {
-	int irq = vcpu->arch.vgic_cpu.vgic_v3.its_vpe.irq;
+	struct its_vpe *vpe = &vcpu->arch.vgic_cpu.vgic_v3.its_vpe;
 	int err;
 
-	if (!vgic_supports_direct_msis(vcpu->kvm))
+	if (!vgic_supports_direct_msis(vcpu->kvm) || vpe->resident)
 		return 0;
 
 	/*
@@ -214,11 +228,14 @@ int vgic_v4_flush_hwstate(struct kvm_vcpu *vcpu)
 	 * doc in drivers/irqchip/irq-gic-v4.c to understand how this
 	 * turns into a VMOVP command at the ITS level.
 	 */
-	err = irq_set_affinity(irq, cpumask_of(smp_processor_id()));
+	err = irq_set_affinity(vpe->irq, cpumask_of(smp_processor_id()));
 	if (err)
 		return err;
 
-	err = its_schedule_vpe(&vcpu->arch.vgic_cpu.vgic_v3.its_vpe, true);
+	/* Disabled the doorbell, as we're about to enter the guest */
+	disable_irq_nosync(vpe->irq);
+
+	err = its_schedule_vpe(vpe, true);
 	if (err)
 		return err;
 
@@ -226,9 +243,7 @@ int vgic_v4_flush_hwstate(struct kvm_vcpu *vcpu)
 	 * Now that the VPE is resident, let's get rid of a potential
 	 * doorbell interrupt that would still be pending.
 	 */
-	err = irq_set_irqchip_state(irq, IRQCHIP_STATE_PENDING, false);
-
-	return err;
+	return irq_set_irqchip_state(vpe->irq, IRQCHIP_STATE_PENDING, false);
 }
 
 static struct vgic_its *vgic_get_its(struct kvm *kvm,
@@ -335,21 +350,3 @@ int kvm_vgic_v4_unset_forwarding(struct kvm *kvm, int virq,
 	mutex_unlock(&its->its_lock);
 	return ret;
 }
-
-void kvm_vgic_v4_enable_doorbell(struct kvm_vcpu *vcpu)
-{
-	if (vgic_supports_direct_msis(vcpu->kvm)) {
-		int irq = vcpu->arch.vgic_cpu.vgic_v3.its_vpe.irq;
-		if (irq)
-			enable_irq(irq);
-	}
-}
-
-void kvm_vgic_v4_disable_doorbell(struct kvm_vcpu *vcpu)
-{
-	if (vgic_supports_direct_msis(vcpu->kvm)) {
-		int irq = vcpu->arch.vgic_cpu.vgic_v3.its_vpe.irq;
-		if (irq)
-			disable_irq(irq);
-	}
-}
diff --git a/virt/kvm/arm/vgic/vgic.c b/virt/kvm/arm/vgic/vgic.c
index 45a870cb63f5..99b02ca730a8 100644
--- a/virt/kvm/arm/vgic/vgic.c
+++ b/virt/kvm/arm/vgic/vgic.c
@@ -857,8 +857,6 @@ void kvm_vgic_sync_hwstate(struct kvm_vcpu *vcpu)
 {
 	struct vgic_cpu *vgic_cpu = &vcpu->arch.vgic_cpu;
 
-	WARN_ON(vgic_v4_sync_hwstate(vcpu));
-
 	/* An empty ap_list_head implies used_lrs == 0 */
 	if (list_empty(&vcpu->arch.vgic_cpu.ap_list_head))
 		return;
@@ -882,8 +880,6 @@ static inline void vgic_restore_state(struct kvm_vcpu *vcpu)
 /* Flush our emulation state into the GIC hardware before entering the guest. */
 void kvm_vgic_flush_hwstate(struct kvm_vcpu *vcpu)
 {
-	WARN_ON(vgic_v4_flush_hwstate(vcpu));
-
 	/*
 	 * If there are no virtual interrupts active or pending for this
 	 * VCPU, then there is no work to do and we can bail out without
diff --git a/virt/kvm/arm/vgic/vgic.h b/virt/kvm/arm/vgic/vgic.h
index 83066a81b16a..c7fefd6b1c80 100644
--- a/virt/kvm/arm/vgic/vgic.h
+++ b/virt/kvm/arm/vgic/vgic.h
@@ -316,7 +316,5 @@ void vgic_its_invalidate_cache(struct kvm *kvm);
 bool vgic_supports_direct_msis(struct kvm *kvm);
 int vgic_v4_init(struct kvm *kvm);
 void vgic_v4_teardown(struct kvm *kvm);
-int vgic_v4_sync_hwstate(struct kvm_vcpu *vcpu);
-int vgic_v4_flush_hwstate(struct kvm_vcpu *vcpu);
 
 #endif
-- 
2.20.1

_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm

  reply index

Thread overview: 65+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-10-27 14:41 [PATCH v2 00/36] irqchip/gic-v4: GICv4.1 architecture support Marc Zyngier
2019-10-27 14:41 ` Marc Zyngier [this message]
2019-10-27 14:42 ` [PATCH v2 02/36] irqchip/gic-v3-its: Factor out wait_for_syncr primitive Marc Zyngier
2019-10-28  9:20   ` Zenghui Yu
2019-10-27 14:42 ` [PATCH v2 03/36] irqchip/gic-v3-its: Allow LPI invalidation via the DirectLPI interface Marc Zyngier
2019-10-31  8:49   ` Zenghui Yu
2019-11-01 13:26     ` Marc Zyngier
2019-11-05 10:30       ` Zenghui Yu
2019-11-05 12:12         ` Marc Zyngier
2019-10-27 14:42 ` [PATCH v2 04/36] irqchip/gic-v3-its: Make is_v4 use a TYPER copy Marc Zyngier
2019-10-28  9:34   ` Zenghui Yu
2019-10-28 10:52     ` Marc Zyngier
2019-10-27 14:42 ` [PATCH v2 05/36] irqchip/gic-v3-its: Kill its->ite_size and use TYPER copy instead Marc Zyngier
2019-10-28  9:40   ` Zenghui Yu
2019-10-27 14:42 ` [PATCH v2 06/36] irqchip/gic-v3-its: Kill its->device_ids " Marc Zyngier
2019-10-31  6:33   ` Zenghui Yu
2019-10-31  8:30     ` Marc Zyngier
2019-10-31  9:08       ` Zenghui Yu
2019-10-27 14:42 ` [PATCH v2 07/36] irqchip/gic-v3-its: Add get_vlpi_map() helper Marc Zyngier
2019-10-31  3:54   ` Zenghui Yu
2019-10-27 14:42 ` [PATCH v2 08/36] irqchip/gic-v3: Detect GICv4.1 supporting RVPEID Marc Zyngier
2019-10-31 11:34   ` Zenghui Yu
2019-10-27 14:42 ` [PATCH v2 09/36] irqchip/gic-v3: Add GICv4.1 VPEID size discovery Marc Zyngier
2019-10-31 12:02   ` Zenghui Yu
2019-11-01 15:13     ` Marc Zyngier
2019-10-27 14:42 ` [PATCH v2 10/36] irqchip/gic-v3: Workaround Cavium TX1 erratum when reading GICD_TYPER2 Marc Zyngier
2019-10-27 14:42 ` [PATCH v2 11/36] irqchip/gic-v4.1: VPE table (aka GICR_VPROPBASER) allocation Marc Zyngier
2019-10-27 14:42 ` [PATCH v2 12/36] irqchip/gic-v4.1: Implement the v4.1 flavour of VMAPP Marc Zyngier
2019-11-01 10:58   ` Zenghui Yu
2019-11-13  8:02   ` Zenghui Yu
2019-11-13  9:47     ` Marc Zyngier
2019-10-27 14:42 ` [PATCH v2 13/36] irqchip/gic-v4.1: Don't use the VPE proxy if RVPEID is set Marc Zyngier
2019-11-01 11:05   ` Zenghui Yu
2019-10-27 14:42 ` [PATCH v2 14/36] irqchip/gic-v4.1: Implement the v4.1 flavour of VMOVP Marc Zyngier
2019-11-01 11:10   ` Zenghui Yu
2019-10-27 14:42 ` [PATCH v2 15/36] irqchip/gic-v4.1: Plumb skeletal VPE irqchip Marc Zyngier
2019-11-01 11:13   ` Zenghui Yu
2019-10-27 14:42 ` [PATCH v2 16/36] irqchip/gic-v4.1: Add mask/unmask doorbell callbacks Marc Zyngier
2019-11-01 11:23   ` Zenghui Yu
2019-10-27 14:42 ` [PATCH v2 17/36] irqchip/gic-v4.1: Add VPE residency callback Marc Zyngier
2019-11-01 11:34   ` Zenghui Yu
2019-10-27 14:42 ` [PATCH v2 18/36] irqchip/gic-v4.1: Add VPE eviction callback Marc Zyngier
2019-11-01 11:39   ` Zenghui Yu
2019-10-27 14:42 ` [PATCH v2 19/36] irqchip/gic-v4.1: Add VPE INVALL callback Marc Zyngier
2019-11-01 11:51   ` Zenghui Yu
2019-10-27 14:42 ` [PATCH v2 20/36] irqchip/gic-v4.1: Suppress per-VLPI doorbell Marc Zyngier
2019-11-01 12:17   ` Zenghui Yu
2019-10-27 14:42 ` [PATCH v2 21/36] irqchip/gic-v4.1: Allow direct invalidation of VLPIs Marc Zyngier
2019-11-01 12:30   ` Zenghui Yu
2019-10-27 14:42 ` [PATCH v2 22/36] irqchip/gic-v4.1: Advertise support v4.1 to KVM Marc Zyngier
2019-11-01 12:55   ` Zenghui Yu
2019-10-27 14:42 ` [PATCH v2 23/36] irqchip/gic-v4.1: Map the ITS SGIR register page Marc Zyngier
2019-10-27 14:42 ` [PATCH v2 24/36] irqchip/gic-v4.1: Plumb skeletal VSGI irqchip Marc Zyngier
2019-10-27 14:42 ` [PATCH v2 25/36] irqchip/gic-v4.1: Add initial SGI configuration Marc Zyngier
2019-10-27 14:42 ` [PATCH v2 26/36] irqchip/gic-v4.1: Plumb mask/unmask SGI callbacks Marc Zyngier
2019-10-27 14:42 ` [PATCH v2 27/36] irqchip/gic-v4.1: Plumb get/set_irqchip_state " Marc Zyngier
2019-10-27 14:42 ` [PATCH v2 28/36] irqchip/gic-v4.1: Plumb set_vcpu_affinity " Marc Zyngier
2019-10-27 14:42 ` [PATCH v2 29/36] irqchip/gic-v4.1: Move doorbell management to the GICv4 abstraction layer Marc Zyngier
2019-10-27 14:42 ` [PATCH v2 30/36] irqchip/gic-v4.1: Add VSGI allocation/teardown Marc Zyngier
2019-10-27 14:42 ` [PATCH v2 31/36] irqchip/gic-v4.1: Add VSGI property setup Marc Zyngier
2019-10-27 14:42 ` [PATCH v2 32/36] irqchip/gic-v4.1: Eagerly vmap vPEs Marc Zyngier
2019-10-27 14:42 ` [PATCH v2 33/36] KVM: arm64: GICv4.1: Let doorbells be auto-enabled Marc Zyngier
2019-10-27 14:42 ` [PATCH v2 34/36] KVM: arm64: GICv4.1: Add direct injection capability to SGI registers Marc Zyngier
2019-10-27 14:42 ` [PATCH v2 35/36] KVM: arm64: GICv4.1: Configure SGIs as HW interrupts Marc Zyngier
2019-10-27 14:42 ` [PATCH v2 36/36] KVM: arm64: GICv4.1: Expose HW-based SGIs in debugfs Marc Zyngier

Reply instructions:

You may reply publically to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20191027144234.8395-2-maz@kernel.org \
    --to=maz@kernel.org \
    --cc=jason@lakedaemon.net \
    --cc=jnair@marvell.com \
    --cc=kvmarm@lists.cs.columbia.edu \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lorenzo.pieralisi@arm.com \
    --cc=rrichter@marvell.com \
    --cc=tglx@linutronix.de \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

KVM ARM Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/kvmarm/0 kvmarm/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 kvmarm kvmarm/ https://lore.kernel.org/kvmarm \
		kvmarm@lists.cs.columbia.edu
	public-inbox-index kvmarm

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/edu.columbia.cs.lists.kvmarm


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git