Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1676774 > unrolled thread

[PATCH v2 00/52] irqchip: KVM: Add support for GICv4

Started byMarc Zyngier <marc.zyngier@arm.com>
First post2017-06-28 17:10 +0200
Last post2017-07-01 17:00 +0200
Articles 20 on this page of 48 — 2 participants

Back to article view | Back to linux.kernel


Contents

  [PATCH v2 00/52] irqchip: KVM: Add support for GICv4 Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:10 +0200
    [PATCH v2 07/52] irqchip/gic-v3-its: Implement irq_set_irqchip_state for pending state Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:10 +0200
    [PATCH v2 39/52] KVM: arm/arm64: GICv4: Wire mapping/unmapping of VLPIs in VFIO irq bypass Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:10 +0200
    [PATCH v2 51/52] KVM: arm/arm64: GICv4: Use pending_last as a scheduling hint Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:10 +0200
    [PATCH v2 30/52] irqchip/gic-v4: Add VPE command interface Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:10 +0200
    [PATCH v2 48/52] KVM: arm/arm64: GICv4: Add doorbell interrupt handling Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:10 +0200
    [PATCH v2 28/52] irqchip/gic-v3-its: Set implementation defined bit to enable VLPIs Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:10 +0200
    [PATCH v2 43/52] KVM: arm/arm64: GICv4: Handle CLEAR applied to a VLPI Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:10 +0200
    [PATCH v2 44/52] KVM: arm/arm64: GICv4: Handle MOVALL applied to a vPE Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:10 +0200
    [PATCH v2 41/52] KVM: arm/arm64: GICv4: Unmap VLPI when freeing an LPI Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:10 +0200
    [PATCH v2 50/52] KVM: arm/arm64: GICv4: Enable virtual cpuif if VLPIs can be delivered Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:10 +0200
    [PATCH v2 49/52] KVM: arm/arm64: GICv4: Hook vPE scheduling into vgic flush/sync Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:10 +0200
    [PATCH v2 37/52] KVM: arm/arm64: GICv4: Add init and teardown of the vPE irq domain Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:10 +0200
    [PATCH v2 52/52] KVM: arm/arm64: GICv4: Enable VLPI support Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:10 +0200
    [PATCH v2 35/52] KVM: arm/arm64: vgic: Move kvm_vgic_destroy call around Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:10 +0200
    [PATCH v2 08/52] irqchip/gic-v3-its: Split out property table allocation Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:10 +0200
    [PATCH v2 47/52] KVM: arm/arm64: GICv4: Propagate VLPI properties at map time Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 19/52] irqchip/gic-v3-its: Add VPE domain infrastructure Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 15/52] irqchip/gic-v3-its: Add GICv4 ITS command definitions Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 42/52] KVM: arm/arm64: GICv4: Handle MOVI applied to a VLPI Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 38/52] KVM: arm/arm64: GICv4: Wire init/teardown of per-VM support Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 32/52] irqchip/gic-v4: Add some basic documentation Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 24/52] irqchip/gic-v3-its: Add VPE invalidation hook Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 33/52] irqchip/gic-v4: Enable low-level GICv4 operations Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 29/52] irqchip/gic-v4: Add per-VM VPE domain creation Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 22/52] irqchip/gic-v3-its: Add VPENDBASER/VPROPBASER accessors Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 13/52] irqchip/gic-v3-its: Generalize LPI configuration Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 40/52] KVM: arm/arm64: GICv4: Handle INT command applied to a VLPI Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 23/52] irqchip/gic-v3-its: Add VPE scheduling Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 17/52] irqchip/gic-v3-its: Add VLPI map/unmap operations Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 46/52] KVM: arm/arm64: GICv4: Handle INVALL applied to a vPE Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 12/52] irqchip/gic-v3-its: Generalize device table allocation Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 09/52] irqchip/gic-v3-its: Allow use of indirect VCPU tables Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 14/52] irqchip/gic-v4: Add management structure definitions Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 11/52] irqchip/gic-v3-its: Rework LPI freeing Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 31/52] irqchip/gic-v4: Add VLPI configuration interface Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 21/52] irqchip/gic-v3-its: Add VPE irq domain [de]activation Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 25/52] irqchip/gic-v3-its: Add VPE affinity changes Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 18/52] irqchip/gic-v3-its: Add VLPI configuration handling Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 16/52] irqchip/gic-v3-its: Add VLPI configuration hook Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 20/52] irqchip/gic-v3-its: Add VPE irq domain allocation/teardown Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 36/52] KVM: arm/arm64: vITS: Add MSI translation helpers Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:20 +0200
    [PATCH v2 10/52] irqchip/gic-v3-its: Split out pending table allocation Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:30 +0200
    [PATCH v2 02/52] irqchip/gic-v3: Add redistributor iterator Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:30 +0200
    [PATCH v2 06/52] irqchip/gic-v3-its: Macro-ize its_send_single_command Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:30 +0200
    [PATCH v2 04/52] irqchip/gic-v3-its: Move LPI definitions around Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:30 +0200
    [PATCH v2 05/52] irqchip/gic-v3-its: Add probing for VLPI properties Marc Zyngier <marc.zyngier@arm.com> - 2017-06-28 17:30 +0200
    RE: [PATCH v2 00/52] irqchip: KVM: Add support for GICv4 "Shanker Donthineni" <shankerd@codeaurora.org> - 2017-07-01 17:00 +0200

Page 2 of 3 — ← Prev page 1 [2] 3  Next page →


#1676797 — [PATCH v2 38/52] KVM: arm/arm64: GICv4: Wire init/teardown of per-VM support

FromMarc Zyngier <marc.zyngier@arm.com>
Date2017-06-28 17:20 +0200
Subject[PATCH v2 38/52] KVM: arm/arm64: GICv4: Wire init/teardown of per-VM support
Message-ID<tXmV4-8uj-9@gated-at.bofh.it>
In reply to#1676774
Should the HW support GICv4 and an ITS being associated with this
VM, let's init the its_vm and its_vpe structures.

Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 virt/kvm/arm/vgic/vgic-init.c | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)

diff --git a/virt/kvm/arm/vgic/vgic-init.c b/virt/kvm/arm/vgic/vgic-init.c
index 3a0b8999f011..0de1f0d986d4 100644
--- a/virt/kvm/arm/vgic/vgic-init.c
+++ b/virt/kvm/arm/vgic/vgic-init.c
@@ -285,8 +285,14 @@ int vgic_init(struct kvm *kvm)
 	if (ret)
 		goto out;
 
-	if (vgic_has_its(kvm))
+	if (vgic_has_its(kvm)) {
 		dist->msis_require_devid = true;
+		if (kvm_vgic_global_state.has_gicv4) {
+			ret = vgic_v4_init(kvm);
+			if (ret)
+				goto out;
+		}
+	}
 
 	kvm_for_each_vcpu(i, vcpu, kvm)
 		kvm_vgic_vcpu_enable(vcpu);
@@ -323,6 +329,9 @@ static void kvm_vgic_dist_destroy(struct kvm *kvm)
 
 	kfree(dist->spis);
 	dist->nr_spis = 0;
+
+	if (kvm_vgic_global_state.has_gicv4 && vgic_has_its(kvm))
+		vgic_v4_teardown(kvm);
 }
 
 void kvm_vgic_vcpu_destroy(struct kvm_vcpu *vcpu)
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]


#1676798 — [PATCH v2 32/52] irqchip/gic-v4: Add some basic documentation

FromMarc Zyngier <marc.zyngier@arm.com>
Date2017-06-28 17:20 +0200
Subject[PATCH v2 32/52] irqchip/gic-v4: Add some basic documentation
Message-ID<tXmV4-8uj-11@gated-at.bofh.it>
In reply to#1676774
Do a braindump of the way things are supposed to work.

Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 drivers/irqchip/irq-gic-v4.c | 71 ++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 71 insertions(+)

diff --git a/drivers/irqchip/irq-gic-v4.c b/drivers/irqchip/irq-gic-v4.c
index 697d9dc1e005..4ce5fdd1db15 100644
--- a/drivers/irqchip/irq-gic-v4.c
+++ b/drivers/irqchip/irq-gic-v4.c
@@ -22,6 +22,77 @@
 
 #include <linux/irqchip/arm-gic-v4.h>
 
+/*
+ * WARNING: The blurb below assumes that you understand the
+ * intricacies of GICv3, GICv4, and how a guest's view of a GICv3 gets
+ * translated into GICv4 commands. So it effectively targets at most
+ * two individuals. You know who you are.
+ *
+ * The core GICv4 code is designed to *avoid* exposing too much of the
+ * core GIC code (that would in turn leak into the hypervisor code),
+ * and instead provide a hypervisor agnostic interface to the HW (of
+ * course, the astute reader will quickly realize that hypervisor
+ * agnostic actually means KVM-specific - what were you thinking?).
+ *
+ * In order to achieve a modicum of isolation, we try to hide most of
+ * the GICv4 "stuff" behind normal irqchip operations:
+ *
+ * - Any guest-visible VLPI is backed by a Linux interrupt (and a
+ *   physical LPI which gets unmapped when the guest maps the
+ *   VLPI). This allows the same DevID/EventID pair to be either
+ *   mapped to the LPI (host) or the VLPI (guest). Note that this is
+ *   exclusive, and you cannot have both.
+ *
+ * - Enabling/disabling a VLPI is done by issuing mask/unmask calls.
+ *
+ * - Guest INT/CLEAR commands are implemented through
+ *   irq_set_irqchip_state().
+ *
+ * - The *bizarre* stuff (mapping/unmapping an interrupt to a VLPI, or
+ *   issuing an INV after changing a priority) gets shoved into the
+ *   irq_set_vcpu_affinity() method. While this is quite horrible
+ *   (let's face it, this is the irqchip version of an ioctl), it
+ *   confines the crap to a single location. And map/unmap really is
+ *   about setting the affinity of a VLPI to a vcpu, so only INV is
+ *   majorly out of place. So there.
+ *
+ * A number of commands are simply not provided by this interface, as
+ * they do not make direct sense. For example, MAPD is purely local to
+ * the virtual ITS (because it references a virtual device, and the
+ * physical ITS is still very much in charge of the physical
+ * device). Same goes for things like MAPC (the physical ITS deals
+ * with the actual vPE affinity, and not the braindead concept of
+ * collection). SYNC is not provided either, as each and every command
+ * is followed by a VSYNC. This could be relaxed in the future, should
+ * this be seen as a bottleneck (yes, this means *never*).
+ *
+ * But handling VLPIs is only one side of the job of the GICv4
+ * code. The other (darker) side is to take care of the doorbell
+ * interrupts which are delivered when a VLPI targeting a non-running
+ * vcpu is being made pending.
+ *
+ * The choice made here is that each vcpu (VPE in old northern GICv4
+ * dialect) gets a single doorbell LPI, no matter how many interrupts
+ * are targeting it. This has a nice property, which is that the
+ * interrupt becomes a handle for the VPE, and that the hypervisor
+ * code can manipulate it through the normal interrupt API:
+ *
+ * - VMs (or rather the VM abstraction that matters to the GIC)
+ *   contain an irq domain where each interrupt maps to a VPE. In
+ *   turn, this domain sits on top of the normal LPI allocator, and a
+ *   specially crafted irq_chip implementation.
+ *
+ * - mask/unmask do what is expected on the doorbell interrupt.
+ *
+ * - irq_set_affinity is used to move a VPE from one redistributor to
+ *   another.
+ *
+ * - irq_set_vcpu_affinity once again gets hijacked for the purpose of
+ *   creating a new sub-API, namely scheduling/descheduling a VPE
+ *   (which involves programming GICR_V{PROP,PEND}BASER) and
+ *   performing INVALL operations.
+ */
+
 static struct irq_domain *gic_domain;
 static const struct irq_domain_ops *vpe_domain_ops;
 
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]


#1676802 — [PATCH v2 24/52] irqchip/gic-v3-its: Add VPE invalidation hook

FromMarc Zyngier <marc.zyngier@arm.com>
Date2017-06-28 17:20 +0200
Subject[PATCH v2 24/52] irqchip/gic-v3-its: Add VPE invalidation hook
Message-ID<tXmV4-8uj-25@gated-at.bofh.it>
In reply to#1676774
When a guest issues a INVALL command targetting a collection, it must
be translated into a VINVALL for the VPE that has this collection.

This patch implements a hook that offers this functionallity to the
hypervisor.

Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 drivers/irqchip/irq-gic-v3-its.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index be8598354d5a..0eda36dc84b4 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -2211,6 +2211,10 @@ static int its_vpe_set_vcpu_affinity(struct irq_data *d, void *vcpu_info)
 		its_vpe_deschedule(vpe);
 		return 0;
 
+	case INVALL_VPE:
+		its_send_vinvall(vpe);
+		return 0;
+
 	default:
 		return -EINVAL;
 	}
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]


#1676803 — [PATCH v2 33/52] irqchip/gic-v4: Enable low-level GICv4 operations

FromMarc Zyngier <marc.zyngier@arm.com>
Date2017-06-28 17:20 +0200
Subject[PATCH v2 33/52] irqchip/gic-v4: Enable low-level GICv4 operations
Message-ID<tXmV4-8uj-27@gated-at.bofh.it>
In reply to#1676774
Get the show on the road...

Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 drivers/irqchip/Makefile           |  2 +-
 drivers/irqchip/irq-gic-v3-its.c   |  3 ++-
 drivers/irqchip/irq-gic-v4.c       | 13 +++++++++++++
 include/linux/irqchip/arm-gic-v4.h |  2 ++
 4 files changed, 18 insertions(+), 2 deletions(-)

diff --git a/drivers/irqchip/Makefile b/drivers/irqchip/Makefile
index b64c59b838a0..4ca95310254b 100644
--- a/drivers/irqchip/Makefile
+++ b/drivers/irqchip/Makefile
@@ -28,7 +28,7 @@ obj-$(CONFIG_ARM_GIC_PM)		+= irq-gic-pm.o
 obj-$(CONFIG_ARCH_REALVIEW)		+= irq-gic-realview.o
 obj-$(CONFIG_ARM_GIC_V2M)		+= irq-gic-v2m.o
 obj-$(CONFIG_ARM_GIC_V3)		+= irq-gic-v3.o irq-gic-common.o
-obj-$(CONFIG_ARM_GIC_V3_ITS)		+= irq-gic-v3-its.o irq-gic-v3-its-pci-msi.o irq-gic-v3-its-platform-msi.o
+obj-$(CONFIG_ARM_GIC_V3_ITS)		+= irq-gic-v3-its.o irq-gic-v3-its-pci-msi.o irq-gic-v3-its-platform-msi.o irq-gic-v4.o
 obj-$(CONFIG_PARTITION_PERCPU)		+= irq-partition-percpu.o
 obj-$(CONFIG_HISILICON_IRQ_MBIGEN)	+= irq-mbigen.o
 obj-$(CONFIG_ARM_NVIC)			+= irq-nvic.o
diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index 3784bea2b994..284e6e6785cf 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -2970,7 +2970,8 @@ int __init its_init(struct fwnode_handle *handle, struct rdists *rdists,
 		has_v4 |= its->is_v4;
 
 	if (has_v4 & rdists->has_vlpis) {
-		if (its_init_vpe_domain()) {
+		if (its_init_vpe_domain() ||
+		    its_init_v4(parent_domain, &its_vpe_domain_ops)) {
 			rdists->has_vlpis = false;
 			pr_err("ITS: Disabling GICv4 support\n");
 		}
diff --git a/drivers/irqchip/irq-gic-v4.c b/drivers/irqchip/irq-gic-v4.c
index 4ce5fdd1db15..7f573d65f440 100644
--- a/drivers/irqchip/irq-gic-v4.c
+++ b/drivers/irqchip/irq-gic-v4.c
@@ -194,3 +194,16 @@ int its_prop_update_vlpi(int irq, u8 config, bool inv)
 
 	return irq_set_vcpu_affinity(irq, &info);
 }
+
+int its_init_v4(struct irq_domain *domain, const struct irq_domain_ops *ops)
+{
+	if (domain) {
+		pr_info("ITS: Enabling GICv4 support\n");
+		gic_domain = domain;
+		vpe_domain_ops = ops;
+		return 0;
+	}
+
+	pr_err("ITS: No GICv4 VPE domain allocated\n");
+	return -ENODEV;
+}
diff --git a/include/linux/irqchip/arm-gic-v4.h b/include/linux/irqchip/arm-gic-v4.h
index 843c9ff29269..862d20ab67d0 100644
--- a/include/linux/irqchip/arm-gic-v4.h
+++ b/include/linux/irqchip/arm-gic-v4.h
@@ -97,4 +97,6 @@ int its_get_vlpi(int irq, struct its_vlpi_map *map);
 int its_unmap_vlpi(int irq);
 int its_prop_update_vlpi(int irq, u8 config, bool inv);
 
+int its_init_v4(struct irq_domain *domain, const struct irq_domain_ops *ops);
+
 #endif
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]


#1676804 — [PATCH v2 29/52] irqchip/gic-v4: Add per-VM VPE domain creation

FromMarc Zyngier <marc.zyngier@arm.com>
Date2017-06-28 17:20 +0200
Subject[PATCH v2 29/52] irqchip/gic-v4: Add per-VM VPE domain creation
Message-ID<tXmV5-8uj-31@gated-at.bofh.it>
In reply to#1676774
When creating a VM, it is very convenient to have an irq domain
containing all the doorbell interrupts associated with that VM
(each interrupt representing a VPE).

Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 drivers/irqchip/irq-gic-v4.c       | 58 ++++++++++++++++++++++++++++++++++++++
 include/linux/irqchip/arm-gic-v4.h |  3 ++
 2 files changed, 61 insertions(+)
 create mode 100644 drivers/irqchip/irq-gic-v4.c

diff --git a/drivers/irqchip/irq-gic-v4.c b/drivers/irqchip/irq-gic-v4.c
new file mode 100644
index 000000000000..2ea12baf2e53
--- /dev/null
+++ b/drivers/irqchip/irq-gic-v4.c
@@ -0,0 +1,58 @@
+/*
+ * Copyright (C) 2016 ARM Limited, All Rights Reserved.
+ * Author: Marc Zyngier <marc.zyngier@arm.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/interrupt.h>
+#include <linux/irq.h>
+#include <linux/irqdomain.h>
+#include <linux/msi.h>
+
+#include <linux/irqchip/arm-gic-v4.h>
+
+static struct irq_domain *gic_domain;
+static const struct irq_domain_ops *vpe_domain_ops;
+
+int its_alloc_vcpu_irqs(struct its_vm *vm)
+{
+	int vpe_base_irq, i;
+
+	vm->domain = irq_domain_create_hierarchy(gic_domain, 0, vm->nr_vpes,
+						 NULL, vpe_domain_ops, vm);
+	if (!vm->domain)
+		return -ENOMEM;
+
+	for (i = 0; i < vm->nr_vpes; i++) {
+		vm->vpes[i]->its_vm = vm;
+		vm->vpes[i]->idai = true;
+	}
+
+	vpe_base_irq = __irq_domain_alloc_irqs(vm->domain, -1, vm->nr_vpes,
+					       NUMA_NO_NODE, vm,
+					       false, NULL);
+	if (vpe_base_irq <= 0)
+		return -ENOMEM;
+
+	for (i = 0; i < vm->nr_vpes; i++)
+		vm->vpes[i]->irq = vpe_base_irq + i;
+
+	return 0;
+}
+
+void its_free_vcpu_irqs(struct its_vm *vm)
+{
+	irq_domain_free_irqs(vm->vpes[0]->irq, vm->nr_vpes);
+	irq_domain_remove(vm->domain);
+}
diff --git a/include/linux/irqchip/arm-gic-v4.h b/include/linux/irqchip/arm-gic-v4.h
index 4cac48e220da..7474a981ed6d 100644
--- a/include/linux/irqchip/arm-gic-v4.h
+++ b/include/linux/irqchip/arm-gic-v4.h
@@ -88,4 +88,7 @@ struct its_cmd_info {
 	};
 };
 
+int its_alloc_vcpu_irqs(struct its_vm *vm);
+void its_free_vcpu_irqs(struct its_vm *vm);
+
 #endif
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]


#1676805 — [PATCH v2 22/52] irqchip/gic-v3-its: Add VPENDBASER/VPROPBASER accessors

FromMarc Zyngier <marc.zyngier@arm.com>
Date2017-06-28 17:20 +0200
Subject[PATCH v2 22/52] irqchip/gic-v3-its: Add VPENDBASER/VPROPBASER accessors
Message-ID<tXmV5-8uj-33@gated-at.bofh.it>
In reply to#1676774
V{PEND,PROP}BASER being 64bit registers, they need some ad-hoc
accessors on 32bit, specially given that VPENDBASER contains
a Valid bit, making the access a bit convoluted.

Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 arch/arm/include/asm/arch_gicv3.h   | 28 ++++++++++++++++++++++++++++
 arch/arm64/include/asm/arch_gicv3.h |  5 +++++
 include/linux/irqchip/arm-gic-v3.h  |  5 +++++
 3 files changed, 38 insertions(+)

diff --git a/arch/arm/include/asm/arch_gicv3.h b/arch/arm/include/asm/arch_gicv3.h
index 27475904e096..8d45e88feac9 100644
--- a/arch/arm/include/asm/arch_gicv3.h
+++ b/arch/arm/include/asm/arch_gicv3.h
@@ -291,5 +291,33 @@ static inline u64 __gic_readq_nonatomic(const volatile void __iomem *addr)
  */
 #define gits_write_cwriter(v, c)	__gic_writeq_nonatomic(v, c)
 
+/*
+ * GITS_VPROPBASER - hi and lo bits may be accessed independently.
+ */
+#define gits_write_vpropbaser(v, c)	__gic_writeq_nonatomic(v, c)
+
+/*
+ * GITS_VPENDBASER - the Valid bit must be cleared before changing
+ * anything else.
+ */
+static inline void gits_write_vpendbaser(u64 val, void * __iomem addr)
+{
+	u32 tmp;
+
+	tmp = readl_relaxed(addr + 4);
+	if (tmp & (GICR_VPENDBASER_Valid >> 32)) {
+		tmp &= ~(GICR_VPENDBASER_Valid >> 32);
+		writel_relaxed(tmp, addr + 4);
+	}
+
+	/*
+	 * Use the fact that __gic_writeq_nonatomic writes the second
+	 * half of the 64bit quantity after the first.
+	 */
+	__gic_writeq_nonatomic(val, addr);
+}
+
+#define gits_read_vpendbaser(c)		__gic_readq_nonatomic(c)
+
 #endif /* !__ASSEMBLY__ */
 #endif /* !__ASM_ARCH_GICV3_H */
diff --git a/arch/arm64/include/asm/arch_gicv3.h b/arch/arm64/include/asm/arch_gicv3.h
index 8cef47fa2218..0d2a53457c30 100644
--- a/arch/arm64/include/asm/arch_gicv3.h
+++ b/arch/arm64/include/asm/arch_gicv3.h
@@ -133,5 +133,10 @@ static inline void gic_write_bpr1(u32 val)
 #define gicr_write_pendbaser(v, c)	writeq_relaxed(v, c)
 #define gicr_read_pendbaser(c)		readq_relaxed(c)
 
+#define gits_write_vpropbaser(v, c)	writeq_relaxed(v, c)
+
+#define gits_write_vpendbaser(v, c)	writeq_relaxed(v, c)
+#define gits_read_vpendbaser(c)		readq_relaxed(c)
+
 #endif /* __ASSEMBLY__ */
 #endif /* __ASM_ARCH_GICV3_H */
diff --git a/include/linux/irqchip/arm-gic-v3.h b/include/linux/irqchip/arm-gic-v3.h
index 7c6fd8f3e36c..17ba0d732f12 100644
--- a/include/linux/irqchip/arm-gic-v3.h
+++ b/include/linux/irqchip/arm-gic-v3.h
@@ -212,6 +212,11 @@
 #define LPI_PROP_GROUP1			(1 << 1)
 #define LPI_PROP_ENABLED		(1 << 0)
 
+#define GICR_VPENDBASER_Dirty		(1ULL << 60)
+#define GICR_VPENDBASER_PendingLast	(1ULL << 61)
+#define GICR_VPENDBASER_IDAI		(1ULL << 62)
+#define GICR_VPENDBASER_Valid		(1ULL << 63)
+
 /*
  * ITS registers, offsets from ITS_base
  */
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]


#1676806 — [PATCH v2 13/52] irqchip/gic-v3-its: Generalize LPI configuration

FromMarc Zyngier <marc.zyngier@arm.com>
Date2017-06-28 17:20 +0200
Subject[PATCH v2 13/52] irqchip/gic-v3-its: Generalize LPI configuration
Message-ID<tXmV5-8uj-29@gated-at.bofh.it>
In reply to#1676774
We're are going to need to change a bit more than just the enable
bit in the LPI property table in the future. So let's change the
LPI configuration funtion to take a set of bits to be cleared,
and a set of bits to be set.

This way, we'll be able to use it when a guest updates an LPI
property (priority, for example).

Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 drivers/irqchip/irq-gic-v3-its.c | 21 +++++++++++----------
 1 file changed, 11 insertions(+), 10 deletions(-)

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index 40682f2a5e1f..25e53958a618 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -680,17 +680,18 @@ static inline u32 its_get_event_id(struct irq_data *d)
 	return d->hwirq - its_dev->event_map.lpi_base;
 }
 
-static void lpi_set_config(struct irq_data *d, bool enable)
+static void lpi_update_config(struct irq_data *d, u8 clr, u8 set)
 {
 	struct its_device *its_dev = irq_data_get_irq_chip_data(d);
 	irq_hw_number_t hwirq = d->hwirq;
-	u32 id = its_get_event_id(d);
-	u8 *cfg = page_address(gic_rdists->prop_page) + hwirq - 8192;
+	struct page *prop_page;
+	u8 *cfg;
 
-	if (enable)
-		*cfg |= LPI_PROP_ENABLED;
-	else
-		*cfg &= ~LPI_PROP_ENABLED;
+	prop_page = gic_rdists->prop_page;
+
+	cfg = page_address(prop_page) + hwirq - 8192;
+	*cfg &= ~clr;
+	*cfg |= set;
 
 	/*
 	 * Make the above write visible to the redistributors.
@@ -701,17 +702,17 @@ static void lpi_set_config(struct irq_data *d, bool enable)
 		gic_flush_dcache_to_poc(cfg, sizeof(*cfg));
 	else
 		dsb(ishst);
-	its_send_inv(its_dev, id);
+	its_send_inv(its_dev, its_get_event_id(d));
 }
 
 static void its_mask_irq(struct irq_data *d)
 {
-	lpi_set_config(d, false);
+	lpi_update_config(d, LPI_PROP_ENABLED, 0);
 }
 
 static void its_unmask_irq(struct irq_data *d)
 {
-	lpi_set_config(d, true);
+	lpi_update_config(d, 0, LPI_PROP_ENABLED);
 }
 
 static int its_set_affinity(struct irq_data *d, const struct cpumask *mask_val,
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]


#1676808 — [PATCH v2 40/52] KVM: arm/arm64: GICv4: Handle INT command applied to a VLPI

FromMarc Zyngier <marc.zyngier@arm.com>
Date2017-06-28 17:20 +0200
Subject[PATCH v2 40/52] KVM: arm/arm64: GICv4: Handle INT command applied to a VLPI
Message-ID<tXmV5-8uj-37@gated-at.bofh.it>
In reply to#1676774
If the guest issues an INT command targetting a VLPI, let's
call into the irq_set_irqchip_state() helper to make it pending
on the physical side.

This works just as well if userspace decides to inject an interrupt
using the normal userspace API...

Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 virt/kvm/arm/vgic/vgic-its.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/virt/kvm/arm/vgic/vgic-its.c b/virt/kvm/arm/vgic/vgic-its.c
index 5fbcf9ea8886..255d29837514 100644
--- a/virt/kvm/arm/vgic/vgic-its.c
+++ b/virt/kvm/arm/vgic/vgic-its.c
@@ -573,6 +573,10 @@ static int vgic_its_trigger_msi(struct kvm *kvm, struct vgic_its *its,
 	if (err)
 		return err;
 
+	if (irq->hw)
+		return irq_set_irqchip_state(irq->host_irq,
+					     IRQCHIP_STATE_PENDING, true);
+
 	spin_lock(&irq->irq_lock);
 	irq->pending_latch = true;
 	vgic_queue_irq_unlock(kvm, irq);
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]


#1676811 — [PATCH v2 23/52] irqchip/gic-v3-its: Add VPE scheduling

FromMarc Zyngier <marc.zyngier@arm.com>
Date2017-06-28 17:20 +0200
Subject[PATCH v2 23/52] irqchip/gic-v3-its: Add VPE scheduling
Message-ID<tXmV5-8uj-41@gated-at.bofh.it>
In reply to#1676774
When a VPE is scheduled to run, the corresponding redistributor must
be told so, by setting VPROPBASER to the VM's property table, and
VPENDBASER to the vcpu's pending table.

When scheduled out, we preserve the IDAI and PendingLast bits. The
latter is specially important, as it tells the hypervisor that
there are pending interrupts for this vcpu.

Reviewed-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 drivers/irqchip/irq-gic-v3-its.c   | 73 ++++++++++++++++++++++++++++++++++++++
 include/linux/irqchip/arm-gic-v3.h | 58 ++++++++++++++++++++++++++++++
 2 files changed, 131 insertions(+)

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index 276602efdfa0..be8598354d5a 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -154,6 +154,7 @@ static DEFINE_IDA(its_vpeid_ida);
 
 #define gic_data_rdist()		(raw_cpu_ptr(gic_rdists->rdist))
 #define gic_data_rdist_rd_base()	(gic_data_rdist()->rd_base)
+#define gic_data_rdist_vlpi_base()	(gic_data_rdist_rd_base() + SZ_128K)
 
 static struct its_collection *dev_event_to_col(struct its_device *its_dev,
 					       u32 event)
@@ -2144,8 +2145,80 @@ static const struct irq_domain_ops its_domain_ops = {
 	.deactivate		= its_irq_domain_deactivate,
 };
 
+static void its_vpe_schedule(struct its_vpe *vpe)
+{
+	void * __iomem vlpi_base = gic_data_rdist_vlpi_base();
+	u64 val;
+
+	/* Schedule the VPE */
+	val  = virt_to_phys(page_address(vpe->its_vm->vprop_page)) &
+		GENMASK_ULL(51, 12);
+	val |= (LPI_NRBITS - 1) & GICR_VPROPBASER_IDBITS_MASK;
+	val |= GICR_VPROPBASER_RaWb;
+	val |= GICR_VPROPBASER_InnerShareable;
+	gits_write_vpropbaser(val, vlpi_base + GICR_VPROPBASER);
+
+	val  = virt_to_phys(page_address(vpe->vpt_page)) &
+		GENMASK_ULL(51, 16);
+	val |= GICR_VPENDBASER_RaWaWb;
+	val |= GICR_VPENDBASER_NonShareable;
+	/*
+	 * There is no good way of finding out if the pending table is
+	 * empty as we can race against the doorbell interrupt very
+	 * easily. So in the end, vpe->pending_last is only an
+	 * indication that the vcpu has something pending, not one
+	 * that the pending table is empty. A good implementation
+	 * would be able to read its coarse map pretty quickly anyway,
+	 * making this a tolerable issue.
+	 */
+	val |= GICR_VPENDBASER_PendingLast;
+	val |= vpe->idai ? GICR_VPENDBASER_IDAI : 0;
+	val |= GICR_VPENDBASER_Valid;
+	gits_write_vpendbaser(val, vlpi_base + GICR_VPENDBASER);
+}
+
+static void its_vpe_deschedule(struct its_vpe *vpe)
+{
+	void * __iomem vlpi_base = gic_data_rdist_vlpi_base();
+	u64 val;
+
+	/* We're being scheduled out */
+	val = gits_read_vpendbaser(vlpi_base + GICR_VPENDBASER);
+	val &= ~GICR_VPENDBASER_Valid;
+	gits_write_vpendbaser(val, vlpi_base + GICR_VPENDBASER);
+
+	val = gits_read_vpendbaser(vlpi_base + GICR_VPENDBASER);
+	while (val & GICR_VPENDBASER_Dirty) {
+		cpu_relax();
+		val = gits_read_vpendbaser(vlpi_base + GICR_VPENDBASER);
+	}
+
+	vpe->idai = !!(val & GICR_VPENDBASER_IDAI);
+	vpe->pending_last = !!(val & GICR_VPENDBASER_PendingLast);
+}
+
+static int its_vpe_set_vcpu_affinity(struct irq_data *d, void *vcpu_info)
+{
+	struct its_vpe *vpe = irq_data_get_irq_chip_data(d);
+	struct its_cmd_info *info = vcpu_info;
+
+	switch (info->cmd_type) {
+	case SCHEDULE_VPE:
+		its_vpe_schedule(vpe);
+		return 0;
+
+	case DESCHEDULE_VPE:
+		its_vpe_deschedule(vpe);
+		return 0;
+
+	default:
+		return -EINVAL;
+	}
+}
+
 static struct irq_chip its_vpe_irq_chip = {
 	.name			= "GICv4-vpe",
+	.irq_set_vcpu_affinity	= its_vpe_set_vcpu_affinity,
 };
 
 static int its_vpe_id_alloc(void)
diff --git a/include/linux/irqchip/arm-gic-v3.h b/include/linux/irqchip/arm-gic-v3.h
index 17ba0d732f12..6bc142cfa616 100644
--- a/include/linux/irqchip/arm-gic-v3.h
+++ b/include/linux/irqchip/arm-gic-v3.h
@@ -212,6 +212,64 @@
 #define LPI_PROP_GROUP1			(1 << 1)
 #define LPI_PROP_ENABLED		(1 << 0)
 
+/*
+ * Re-Distributor registers, offsets from VLPI_base
+ */
+#define GICR_VPROPBASER			0x0070
+
+#define GICR_VPROPBASER_IDBITS_MASK	0x1f
+
+#define GICR_VPROPBASER_SHAREABILITY_SHIFT		(10)
+#define GICR_VPROPBASER_INNER_CACHEABILITY_SHIFT	(7)
+#define GICR_VPROPBASER_OUTER_CACHEABILITY_SHIFT	(56)
+
+#define GICR_VPROPBASER_SHAREABILITY_MASK				\
+	GIC_BASER_SHAREABILITY(GICR_VPROPBASER, SHAREABILITY_MASK)
+#define GICR_VPROPBASER_INNER_CACHEABILITY_MASK				\
+	GIC_BASER_CACHEABILITY(GICR_VPROPBASER, INNER, MASK)
+#define GICR_VPROPBASER_OUTER_CACHEABILITY_MASK				\
+	GIC_BASER_CACHEABILITY(GICR_VPROPBASER, OUTER, MASK)
+#define GICR_VPROPBASER_CACHEABILITY_MASK				\
+	GICR_VPROPBASER_INNER_CACHEABILITY_MASK
+
+#define GICR_VPROPBASER_InnerShareable					\
+	GIC_BASER_SHAREABILITY(GICR_VPROPBASER, InnerShareable)
+
+#define GICR_VPROPBASER_nCnB	GIC_BASER_CACHEABILITY(GICR_VPROPBASER, INNER, nCnB)
+#define GICR_VPROPBASER_nC 	GIC_BASER_CACHEABILITY(GICR_VPROPBASER, INNER, nC)
+#define GICR_VPROPBASER_RaWt	GIC_BASER_CACHEABILITY(GICR_VPROPBASER, INNER, RaWt)
+#define GICR_VPROPBASER_RaWb	GIC_BASER_CACHEABILITY(GICR_VPROPBASER, INNER, RaWt)
+#define GICR_VPROPBASER_WaWt	GIC_BASER_CACHEABILITY(GICR_VPROPBASER, INNER, WaWt)
+#define GICR_VPROPBASER_WaWb	GIC_BASER_CACHEABILITY(GICR_VPROPBASER, INNER, WaWb)
+#define GICR_VPROPBASER_RaWaWt	GIC_BASER_CACHEABILITY(GICR_VPROPBASER, INNER, RaWaWt)
+#define GICR_VPROPBASER_RaWaWb	GIC_BASER_CACHEABILITY(GICR_VPROPBASER, INNER, RaWaWb)
+
+#define GICR_VPENDBASER			0x0078
+
+#define GICR_VPENDBASER_SHAREABILITY_SHIFT		(10)
+#define GICR_VPENDBASER_INNER_CACHEABILITY_SHIFT	(7)
+#define GICR_VPENDBASER_OUTER_CACHEABILITY_SHIFT	(56)
+#define GICR_VPENDBASER_SHAREABILITY_MASK				\
+	GIC_BASER_SHAREABILITY(GICR_VPENDBASER, SHAREABILITY_MASK)
+#define GICR_VPENDBASER_INNER_CACHEABILITY_MASK				\
+	GIC_BASER_CACHEABILITY(GICR_VPENDBASER, INNER, MASK)
+#define GICR_VPENDBASER_OUTER_CACHEABILITY_MASK				\
+	GIC_BASER_CACHEABILITY(GICR_VPENDBASER, OUTER, MASK)
+#define GICR_VPENDBASER_CACHEABILITY_MASK				\
+	GICR_VPENDBASER_INNER_CACHEABILITY_MASK
+
+#define GICR_VPENDBASER_NonShareable					\
+	GIC_BASER_SHAREABILITY(GICR_VPENDBASER, NonShareable)
+
+#define GICR_VPENDBASER_nCnB	GIC_BASER_CACHEABILITY(GICR_VPENDBASER, INNER, nCnB)
+#define GICR_VPENDBASER_nC 	GIC_BASER_CACHEABILITY(GICR_VPENDBASER, INNER, nC)
+#define GICR_VPENDBASER_RaWt	GIC_BASER_CACHEABILITY(GICR_VPENDBASER, INNER, RaWt)
+#define GICR_VPENDBASER_RaWb	GIC_BASER_CACHEABILITY(GICR_VPENDBASER, INNER, RaWt)
+#define GICR_VPENDBASER_WaWt	GIC_BASER_CACHEABILITY(GICR_VPENDBASER, INNER, WaWt)
+#define GICR_VPENDBASER_WaWb	GIC_BASER_CACHEABILITY(GICR_VPENDBASER, INNER, WaWb)
+#define GICR_VPENDBASER_RaWaWt	GIC_BASER_CACHEABILITY(GICR_VPENDBASER, INNER, RaWaWt)
+#define GICR_VPENDBASER_RaWaWb	GIC_BASER_CACHEABILITY(GICR_VPENDBASER, INNER, RaWaWb)
+
 #define GICR_VPENDBASER_Dirty		(1ULL << 60)
 #define GICR_VPENDBASER_PendingLast	(1ULL << 61)
 #define GICR_VPENDBASER_IDAI		(1ULL << 62)
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]


#1676812 — [PATCH v2 17/52] irqchip/gic-v3-its: Add VLPI map/unmap operations

FromMarc Zyngier <marc.zyngier@arm.com>
Date2017-06-28 17:20 +0200
Subject[PATCH v2 17/52] irqchip/gic-v3-its: Add VLPI map/unmap operations
Message-ID<tXmV5-8uj-43@gated-at.bofh.it>
In reply to#1676774
In order to let a VLPI being injected into a guest, the VLPI must
be mapped using the VMAPTI command. When moved to a different vcpu,
it must be moved with the VMOVI command.

These commands are issued via the irq_set_vcpu_affinity method,
making sure we unmap the corresponding host LPI first.

The reverse is also done when the VLPI is unmapped from the guest.

Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 drivers/irqchip/irq-gic-v3-its.c | 250 ++++++++++++++++++++++++++++++++++++++-
 1 file changed, 247 insertions(+), 3 deletions(-)

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index c02e1e6cd169..527e4e95e10d 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -117,11 +117,17 @@ struct event_lpi_map {
 	u16			*col_map;
 	irq_hw_number_t		lpi_base;
 	int			nr_lpis;
+	struct mutex		vlpi_lock;
+	struct its_vm		*vm;
+	struct its_vlpi_map	*vlpi_maps;
+	int			nr_vlpis;
 };
 
 /*
- * The ITS view of a device - belongs to an ITS, a collection, owns an
- * interrupt translation table, and a list of interrupts.
+ * The ITS view of a device - belongs to an ITS, owns an interrupt
+ * translation table, and a list of interrupts.  If it some of its
+ * LPIs are injected into a guest (GICv4), the event_map.vm field
+ * indicates which one.
  */
 struct its_device {
 	struct list_head	entry;
@@ -207,6 +213,21 @@ struct its_cmd_desc {
 		struct {
 			struct its_collection *col;
 		} its_invall_cmd;
+
+		struct {
+			struct its_vpe *vpe;
+			struct its_device *dev;
+			u32 virt_id;
+			u32 event_id;
+			bool db_enabled;
+		} its_vmapti_cmd;
+
+		struct {
+			struct its_vpe *vpe;
+			struct its_device *dev;
+			u32 event_id;
+			bool db_enabled;
+		} its_vmovi_cmd;
 	};
 };
 
@@ -223,6 +244,9 @@ struct its_cmd_block {
 typedef struct its_collection *(*its_cmd_builder_t)(struct its_cmd_block *,
 						    struct its_cmd_desc *);
 
+typedef struct its_vpe *(*its_cmd_vbuilder_t)(struct its_cmd_block *,
+					      struct its_cmd_desc *);
+
 static void its_mask_encode(u64 *raw_cmd, u64 val, int h, int l)
 {
 	u64 mask = GENMASK_ULL(h, l);
@@ -275,6 +299,26 @@ static void its_encode_collection(struct its_cmd_block *cmd, u16 col)
 	its_mask_encode(&cmd->raw_cmd[2], col, 15, 0);
 }
 
+static void its_encode_vpeid(struct its_cmd_block *cmd, u16 vpeid)
+{
+	its_mask_encode(&cmd->raw_cmd[1], vpeid, 47, 32);
+}
+
+static void its_encode_virt_id(struct its_cmd_block *cmd, u32 virt_id)
+{
+	its_mask_encode(&cmd->raw_cmd[2], virt_id, 31, 0);
+}
+
+static void its_encode_db_phys_id(struct its_cmd_block *cmd, u32 db_phys_id)
+{
+	its_mask_encode(&cmd->raw_cmd[2], db_phys_id, 63, 32);
+}
+
+static void its_encode_db_valid(struct its_cmd_block *cmd, bool db_valid)
+{
+	its_mask_encode(&cmd->raw_cmd[2], db_valid, 0, 0);
+}
+
 static inline void its_fixup_cmd(struct its_cmd_block *cmd)
 {
 	/* Let's fixup BE commands */
@@ -433,6 +477,50 @@ static struct its_collection *its_build_invall_cmd(struct its_cmd_block *cmd,
 	return NULL;
 }
 
+static struct its_vpe *its_build_vmapti_cmd(struct its_cmd_block *cmd,
+					    struct its_cmd_desc *desc)
+{
+	u32 db;
+
+	if (desc->its_vmapti_cmd.db_enabled)
+		db = desc->its_vmapti_cmd.vpe->vpe_db_lpi;
+	else
+		db = 1023;
+
+	its_encode_cmd(cmd, GITS_CMD_VMAPTI);
+	its_encode_devid(cmd, desc->its_vmapti_cmd.dev->device_id);
+	its_encode_vpeid(cmd, desc->its_vmapti_cmd.vpe->vpe_id);
+	its_encode_event_id(cmd, desc->its_vmapti_cmd.event_id);
+	its_encode_db_phys_id(cmd, db);
+	its_encode_virt_id(cmd, desc->its_vmapti_cmd.virt_id);
+
+	its_fixup_cmd(cmd);
+
+	return desc->its_vmapti_cmd.vpe;
+}
+
+static struct its_vpe *its_build_vmovi_cmd(struct its_cmd_block *cmd,
+					   struct its_cmd_desc *desc)
+{
+	u32 db;
+
+	if (desc->its_vmovi_cmd.db_enabled)
+		db = desc->its_vmovi_cmd.vpe->vpe_db_lpi;
+	else
+		db = 1023;
+
+	its_encode_cmd(cmd, GITS_CMD_VMOVI);
+	its_encode_devid(cmd, desc->its_vmovi_cmd.dev->device_id);
+	its_encode_vpeid(cmd, desc->its_vmovi_cmd.vpe->vpe_id);
+	its_encode_event_id(cmd, desc->its_vmovi_cmd.event_id);
+	its_encode_db_phys_id(cmd, db);
+	its_encode_db_valid(cmd, true);
+
+	its_fixup_cmd(cmd);
+
+	return desc->its_vmovi_cmd.vpe;
+}
+
 static u64 its_cmd_ptr_to_offset(struct its_node *its,
 				 struct its_cmd_block *ptr)
 {
@@ -578,6 +666,18 @@ static void its_build_sync_cmd(struct its_cmd_block *sync_cmd,
 static BUILD_SINGLE_CMD_FUNC(its_send_single_command, its_cmd_builder_t,
 			     struct its_collection, its_build_sync_cmd)
 
+static void its_build_vsync_cmd(struct its_cmd_block *sync_cmd,
+				struct its_vpe *sync_vpe)
+{
+	its_encode_cmd(sync_cmd, GITS_CMD_VSYNC);
+	its_encode_vpeid(sync_cmd, sync_vpe->vpe_id);
+
+	its_fixup_cmd(sync_cmd);
+}
+
+static BUILD_SINGLE_CMD_FUNC(its_send_single_vcommand, its_cmd_vbuilder_t,
+			     struct its_vpe, its_build_vsync_cmd)
+
 static void its_send_int(struct its_device *dev, u32 event_id)
 {
 	struct its_cmd_desc desc;
@@ -671,6 +771,33 @@ static void its_send_invall(struct its_node *its, struct its_collection *col)
 	its_send_single_command(its, its_build_invall_cmd, &desc);
 }
 
+static void its_send_vmapti(struct its_device *dev, u32 id)
+{
+	struct its_vlpi_map *map = &dev->event_map.vlpi_maps[id];
+	struct its_cmd_desc desc;
+
+	desc.its_vmapti_cmd.vpe = map->vm->vpes[map->vpe_idx];
+	desc.its_vmapti_cmd.dev = dev;
+	desc.its_vmapti_cmd.virt_id = map->vintid;
+	desc.its_vmapti_cmd.event_id = id;
+	desc.its_vmapti_cmd.db_enabled = map->db_enabled;
+
+	its_send_single_vcommand(dev->its, its_build_vmapti_cmd, &desc);
+}
+
+static void its_send_vmovi(struct its_device *dev, u32 id)
+{
+	struct its_vlpi_map *map = &dev->event_map.vlpi_maps[id];
+	struct its_cmd_desc desc;
+
+	desc.its_vmovi_cmd.vpe = map->vm->vpes[map->vpe_idx];
+	desc.its_vmovi_cmd.dev = dev;
+	desc.its_vmovi_cmd.event_id = id;
+	desc.its_vmovi_cmd.db_enabled = map->db_enabled;
+
+	its_send_single_vcommand(dev->its, its_build_vmovi_cmd, &desc);
+}
+
 /*
  * irqchip functions - assumes MSI, mostly.
  */
@@ -780,19 +907,135 @@ static int its_irq_set_irqchip_state(struct irq_data *d,
 	return 0;
 }
 
+static int its_vlpi_map(struct irq_data *d, struct its_cmd_info *info)
+{
+	struct its_device *its_dev = irq_data_get_irq_chip_data(d);
+	u32 event = its_get_event_id(d);
+	int ret = 0;
+
+	if (!info->map)
+		return -EINVAL;
+
+	mutex_lock(&its_dev->event_map.vlpi_lock);
+
+	if (!its_dev->event_map.vm) {
+		struct its_vlpi_map *maps;
+
+		maps = kzalloc(sizeof(*maps) * its_dev->event_map.nr_lpis,
+			       GFP_KERNEL);
+		if (!maps) {
+			ret = -ENOMEM;
+			goto out;
+		}
+
+		its_dev->event_map.vm = info->map->vm;
+		its_dev->event_map.vlpi_maps = maps;
+	} else if (its_dev->event_map.vm != info->map->vm) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	/* Get our private copy of the mapping information */
+	its_dev->event_map.vlpi_maps[event] = *info->map;
+
+	if (irqd_is_forwarded_to_vcpu(d)) {
+		/* Already mapped, move it around */
+		its_send_vmovi(its_dev, event);
+	} else {
+		/* Drop the physical mapping */
+		its_send_discard(its_dev, event);
+
+		/* and install the virtual one */
+		its_send_vmapti(its_dev, event);
+		irqd_set_forwarded_to_vcpu(d);
+
+		/* Increment the number of VLPIs */
+		its_dev->event_map.nr_vlpis++;
+	}
+
+out:
+	mutex_unlock(&its_dev->event_map.vlpi_lock);
+	return ret;
+}
+
+static int its_vlpi_get(struct irq_data *d, struct its_cmd_info *info)
+{
+	struct its_device *its_dev = irq_data_get_irq_chip_data(d);
+	u32 event = its_get_event_id(d);
+	int ret = 0;
+
+	mutex_lock(&its_dev->event_map.vlpi_lock);
+
+	if (!its_dev->event_map.vm ||
+	    !its_dev->event_map.vlpi_maps[event].vm) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	/* Copy our mapping information to the incoming request */
+	*info->map = its_dev->event_map.vlpi_maps[event];
+
+out:
+	mutex_unlock(&its_dev->event_map.vlpi_lock);
+	return ret;
+}
+
+static int its_vlpi_unmap(struct irq_data *d)
+{
+	struct its_device *its_dev = irq_data_get_irq_chip_data(d);
+	u32 event = its_get_event_id(d);
+	int ret = 0;
+
+	mutex_lock(&its_dev->event_map.vlpi_lock);
+
+	if (!its_dev->event_map.vm || !irqd_is_forwarded_to_vcpu(d)) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	/* Drop the virtual mapping */
+	its_send_discard(its_dev, event);
+
+	/* and restore the physical one */
+	irqd_clr_forwarded_to_vcpu(d);
+	its_send_mapti(its_dev, d->hwirq, event);
+	lpi_update_config(d, 0xff, (LPI_PROP_DEFAULT_PRIO |
+				    LPI_PROP_ENABLED |
+				    LPI_PROP_GROUP1));
+
+	/*
+	 * Drop the refcount and make the device available again if
+	 * this was the last VLPI.
+	 */
+	if (!--its_dev->event_map.nr_vlpis) {
+		its_dev->event_map.vm = NULL;
+		kfree(its_dev->event_map.vlpi_maps);
+	}
+
+out:
+	mutex_unlock(&its_dev->event_map.vlpi_lock);
+	return ret;
+}
+
 static int its_irq_set_vcpu_affinity(struct irq_data *d, void *vcpu_info)
 {
 	struct its_device *its_dev = irq_data_get_irq_chip_data(d);
 	struct its_cmd_info *info = vcpu_info;
 
 	/* Need a v4 ITS */
-	if (!its_dev->its->is_v4 || !info)
+	if (!its_dev->its->is_v4)
 		return -EINVAL;
 
+	/* Unmap request? */
+	if (!info)
+		return its_vlpi_unmap(d);
+
 	switch (info->cmd_type) {
 	case MAP_VLPI:
+		return its_vlpi_map(d, info);
 
 	case GET_VLPI:
+		return its_vlpi_get(d, info);
 
 	case PROP_UPDATE_VLPI:
 	case PROP_UPDATE_AND_INV_VLPI:
@@ -1509,6 +1752,7 @@ static struct its_device *its_create_device(struct its_node *its, u32 dev_id,
 	dev->event_map.col_map = col_map;
 	dev->event_map.lpi_base = lpi_base;
 	dev->event_map.nr_lpis = nr_lpis;
+	mutex_init(&dev->event_map.vlpi_lock);
 	dev->device_id = dev_id;
 	INIT_LIST_HEAD(&dev->entry);
 
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]


#1676813 — [PATCH v2 46/52] KVM: arm/arm64: GICv4: Handle INVALL applied to a vPE

FromMarc Zyngier <marc.zyngier@arm.com>
Date2017-06-28 17:20 +0200
Subject[PATCH v2 46/52] KVM: arm/arm64: GICv4: Handle INVALL applied to a vPE
Message-ID<tXmV5-8uj-39@gated-at.bofh.it>
In reply to#1676774
Since when updating the properties one LPI at a time, there is no
need to perform an INV each time we read one. Instead, we rely
on the final VINVALL that gets sent to the ITS to do the work.

Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 virt/kvm/arm/vgic/vgic-its.c | 15 +++++++++------
 1 file changed, 9 insertions(+), 6 deletions(-)

diff --git a/virt/kvm/arm/vgic/vgic-its.c b/virt/kvm/arm/vgic/vgic-its.c
index 2015246ab9f9..b0790ca5e933 100644
--- a/virt/kvm/arm/vgic/vgic-its.c
+++ b/virt/kvm/arm/vgic/vgic-its.c
@@ -38,7 +38,7 @@ static int vgic_its_save_tables_v0(struct vgic_its *its);
 static int vgic_its_restore_tables_v0(struct vgic_its *its);
 static int vgic_its_commit_v0(struct vgic_its *its);
 static int update_lpi_config(struct kvm *kvm, struct vgic_irq *irq,
-			     struct kvm_vcpu *filter_vcpu);
+			     struct kvm_vcpu *filter_vcpu, bool needs_inv);
 
 /*
  * Creates a new (reference to a) struct vgic_irq for a given LPI.
@@ -106,7 +106,7 @@ static struct vgic_irq *vgic_add_lpi(struct kvm *kvm, u32 intid,
 	 * However we only have those structs for mapped IRQs, so we read in
 	 * the respective config data from memory here upon mapping the LPI.
 	 */
-	ret = update_lpi_config(kvm, irq, NULL);
+	ret = update_lpi_config(kvm, irq, NULL, false);
 	if (ret)
 		return ERR_PTR(ret);
 
@@ -274,7 +274,7 @@ static struct its_collection *find_collection(struct vgic_its *its, int coll_id)
  * VCPU. Unconditionally applies if filter_vcpu is NULL.
  */
 static int update_lpi_config(struct kvm *kvm, struct vgic_irq *irq,
-			     struct kvm_vcpu *filter_vcpu)
+			     struct kvm_vcpu *filter_vcpu, bool needs_inv)
 {
 	u64 propbase = GICR_PROPBASER_ADDRESS(kvm->arch.vgic.propbaser);
 	u8 prop;
@@ -298,7 +298,7 @@ static int update_lpi_config(struct kvm *kvm, struct vgic_irq *irq,
 	}
 
 	if (irq->hw)
-		return its_prop_update_vlpi(irq->host_irq, prop, true);
+		return its_prop_update_vlpi(irq->host_irq, prop, needs_inv);
 
 	return 0;
 }
@@ -1095,7 +1095,7 @@ static int vgic_its_cmd_handle_inv(struct kvm *kvm, struct vgic_its *its,
 	if (!ite)
 		return E_ITS_INV_UNMAPPED_INTERRUPT;
 
-	return update_lpi_config(kvm, ite->irq, NULL);
+	return update_lpi_config(kvm, ite->irq, NULL, true);
 }
 
 /*
@@ -1130,12 +1130,15 @@ static int vgic_its_cmd_handle_invall(struct kvm *kvm, struct vgic_its *its,
 		irq = vgic_get_irq(kvm, NULL, intids[i]);
 		if (!irq)
 			continue;
-		update_lpi_config(kvm, irq, vcpu);
+		update_lpi_config(kvm, irq, vcpu, false);
 		vgic_put_irq(kvm, irq);
 	}
 
 	kfree(intids);
 
+	if (vcpu->arch.vgic_cpu.vgic_v3.its_vpe.its_vm)
+		its_invall_vpe(&vcpu->arch.vgic_cpu.vgic_v3.its_vpe);
+
 	return 0;
 }
 
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]


#1676815 — [PATCH v2 12/52] irqchip/gic-v3-its: Generalize device table allocation

FromMarc Zyngier <marc.zyngier@arm.com>
Date2017-06-28 17:20 +0200
Subject[PATCH v2 12/52] irqchip/gic-v3-its: Generalize device table allocation
Message-ID<tXmV5-8uj-53@gated-at.bofh.it>
In reply to#1676774
As we want to use 2-level tables for VCPUs, let's hack the device
table allocator in order to make it slightly more generic. It
will get reused in subsequent patches.

Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 drivers/irqchip/irq-gic-v3-its.c | 26 ++++++++++++++++----------
 1 file changed, 16 insertions(+), 10 deletions(-)

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index 3011403438c3..40682f2a5e1f 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -1383,26 +1383,19 @@ static struct its_baser *its_get_baser(struct its_node *its, u32 type)
 	return NULL;
 }
 
-static bool its_alloc_device_table(struct its_node *its, u32 dev_id)
+static bool its_alloc_table_entry(struct its_baser *baser, u32 id)
 {
-	struct its_baser *baser;
 	struct page *page;
 	u32 esz, idx;
 	__le64 *table;
 
-	baser = its_get_baser(its, GITS_BASER_TYPE_DEVICE);
-
-	/* Don't allow device id that exceeds ITS hardware limit */
-	if (!baser)
-		return (ilog2(dev_id) < its->device_ids);
-
 	/* Don't allow device id that exceeds single, flat table limit */
 	esz = GITS_BASER_ENTRY_SIZE(baser->val);
 	if (!(baser->val & GITS_BASER_INDIRECT))
-		return (dev_id < (PAGE_ORDER_TO_SIZE(baser->order) / esz));
+		return (id < (PAGE_ORDER_TO_SIZE(baser->order) / esz));
 
 	/* Compute 1st level table index & check if that exceeds table limit */
-	idx = dev_id >> ilog2(baser->psz / esz);
+	idx = id >> ilog2(baser->psz / esz);
 	if (idx >= (PAGE_ORDER_TO_SIZE(baser->order) / GITS_LVL1_ENTRY_SIZE))
 		return false;
 
@@ -1431,6 +1424,19 @@ static bool its_alloc_device_table(struct its_node *its, u32 dev_id)
 	return true;
 }
 
+static bool its_alloc_device_table(struct its_node *its, u32 dev_id)
+{
+	struct its_baser *baser;
+
+	baser = its_get_baser(its, GITS_BASER_TYPE_DEVICE);
+
+	/* Don't allow device id that exceeds ITS hardware limit */
+	if (!baser)
+		return (ilog2(dev_id) < its->device_ids);
+
+	return its_alloc_table_entry(baser, dev_id);
+}
+
 static struct its_device *its_create_device(struct its_node *its, u32 dev_id,
 					    int nvecs)
 {
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]


#1676818 — [PATCH v2 09/52] irqchip/gic-v3-its: Allow use of indirect VCPU tables

FromMarc Zyngier <marc.zyngier@arm.com>
Date2017-06-28 17:20 +0200
Subject[PATCH v2 09/52] irqchip/gic-v3-its: Allow use of indirect VCPU tables
Message-ID<tXmV5-8uj-55@gated-at.bofh.it>
In reply to#1676774
The VCPU tables can be quite sparse as well, and it makes sense
to use indirect tables as well if possible.

Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 drivers/irqchip/irq-gic-v3-its.c | 25 +++++++++++++++++--------
 1 file changed, 17 insertions(+), 8 deletions(-)

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index 2feecdf40216..27f57c528026 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -1068,10 +1068,13 @@ static int its_setup_baser(struct its_node *its, struct its_baser *baser,
 	return 0;
 }
 
-static bool its_parse_baser_device(struct its_node *its, struct its_baser *baser,
-				   u32 psz, u32 *order)
+static bool its_parse_indirect_baser(struct its_node *its,
+				     struct its_baser *baser,
+				     u32 psz, u32 *order)
 {
-	u64 esz = GITS_BASER_ENTRY_SIZE(its_read_baser(its, baser));
+	u64 tmp = its_read_baser(its, baser);
+	u64 type = GITS_BASER_TYPE(tmp);
+	u64 esz = GITS_BASER_ENTRY_SIZE(tmp);
 	u64 val = GITS_BASER_InnerShareable | GITS_BASER_RaWaWb;
 	u32 ids = its->device_ids;
 	u32 new_order = *order;
@@ -1110,8 +1113,9 @@ static bool its_parse_baser_device(struct its_node *its, struct its_baser *baser
 	if (new_order >= MAX_ORDER) {
 		new_order = MAX_ORDER - 1;
 		ids = ilog2(PAGE_ORDER_TO_SIZE(new_order) / (int)esz);
-		pr_warn("ITS@%pa: Device Table too large, reduce ids %u->%u\n",
-			&its->phys_base, its->device_ids, ids);
+		pr_warn("ITS@%pa: %s Table too large, reduce ids %u->%u\n",
+			&its->phys_base, its_base_type_string[type],
+			its->device_ids, ids);
 	}
 
 	*order = new_order;
@@ -1159,11 +1163,16 @@ static int its_alloc_tables(struct its_node *its)
 		u32 order = get_order(psz);
 		bool indirect = false;
 
-		if (type == GITS_BASER_TYPE_NONE)
+		switch (type) {
+		case GITS_BASER_TYPE_NONE:
 			continue;
 
-		if (type == GITS_BASER_TYPE_DEVICE)
-			indirect = its_parse_baser_device(its, baser, psz, &order);
+		case GITS_BASER_TYPE_DEVICE:
+		case GITS_BASER_TYPE_VCPU:
+			indirect = its_parse_indirect_baser(its, baser,
+							    psz, &order);
+			break;
+		}
 
 		err = its_setup_baser(its, baser, cache, shr, psz, order, indirect);
 		if (err < 0) {
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]


#1676819 — [PATCH v2 14/52] irqchip/gic-v4: Add management structure definitions

FromMarc Zyngier <marc.zyngier@arm.com>
Date2017-06-28 17:20 +0200
Subject[PATCH v2 14/52] irqchip/gic-v4: Add management structure definitions
Message-ID<tXmV5-8uj-57@gated-at.bofh.it>
In reply to#1676774
Add a bunch of GICv4-specific data structures that will get used in
subsequent patches.

Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 include/linux/irqchip/arm-gic-v4.h | 91 ++++++++++++++++++++++++++++++++++++++
 1 file changed, 91 insertions(+)
 create mode 100644 include/linux/irqchip/arm-gic-v4.h

diff --git a/include/linux/irqchip/arm-gic-v4.h b/include/linux/irqchip/arm-gic-v4.h
new file mode 100644
index 000000000000..4cac48e220da
--- /dev/null
+++ b/include/linux/irqchip/arm-gic-v4.h
@@ -0,0 +1,91 @@
+/*
+ * Copyright (C) 2016 ARM Limited, All Rights Reserved.
+ * Author: Marc Zyngier <marc.zyngier@arm.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef __LINUX_IRQCHIP_ARM_GIC_V4_H
+#define __LINUX_IRQCHIP_ARM_GIC_V4_H
+
+struct its_vpe;
+
+/* Embedded in kvm.arch */
+struct its_vm {
+	struct irq_domain	*domain;
+	struct page		*vprop_page;
+	struct its_vpe		**vpes;
+	int			nr_vpes;
+	irq_hw_number_t		db_lpi_base;
+	unsigned long		*db_bitmap;
+	int			nr_db_lpis;
+};
+
+/* Embedded in kvm_vcpu.arch */
+struct its_vpe {
+	struct page 		*vpt_page;
+	struct its_vm		*its_vm;
+	/* Doorbell interrupt */
+	int			irq;
+	irq_hw_number_t		vpe_db_lpi;
+	/*
+	 * This collection ID is used to indirect the target
+	 * redistributor for this VPE. The ID itself isn't involved in
+	 * programming of the ITS.
+	 */
+	u16			col_idx;
+	/* Unique (system-wide) VPE identifier */
+	u16			vpe_id;
+	/* Implementation Defined Area Invalid */
+	bool			idai;
+	/* Pending VLPIs on schedule out? */
+	bool			pending_last;
+};
+
+/*
+ * struct its_vlpi_map: structure describing the mapping of a
+ * VLPI. Only to be interpreted in the context of a physical interrupt
+ * it complements.  To be used as the vcpu_info passed to
+ * irq_set_vcpu_affinity().
+ *
+ * @vm:		Pointer to the GICv4 notion of a VM
+ * @vintid:	Virtual LPI number
+ * @db_enabled:	Is the VPE doorbell to be generated?
+ * @vpe_idx:	Index (0-based) of the VPE in this VM. Not the vpe_id!
+ */
+struct its_vlpi_map {
+	struct its_vm		*vm;
+	u32			vintid;
+	bool			db_enabled;
+	u16			vpe_idx;
+};
+
+enum its_vcpu_info_cmd_type {
+	MAP_VLPI,
+	GET_VLPI,
+	PROP_UPDATE_VLPI,
+	PROP_UPDATE_AND_INV_VLPI,
+	SCHEDULE_VPE,
+	DESCHEDULE_VPE,
+	INVALL_VPE,
+};
+
+struct its_cmd_info {
+	enum its_vcpu_info_cmd_type	cmd_type;
+	union {
+		struct its_vlpi_map	*map;
+		u8			config;
+	};
+};
+
+#endif
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]


#1676820 — [PATCH v2 11/52] irqchip/gic-v3-its: Rework LPI freeing

FromMarc Zyngier <marc.zyngier@arm.com>
Date2017-06-28 17:20 +0200
Subject[PATCH v2 11/52] irqchip/gic-v3-its: Rework LPI freeing
Message-ID<tXmV6-8uj-63@gated-at.bofh.it>
In reply to#1676774
Rework LPI deallocation so that it can be reused by the v4 support
code.

Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 drivers/irqchip/irq-gic-v3-its.c | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index d6f9ccb36f4d..3011403438c3 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -873,16 +873,15 @@ static unsigned long *its_lpi_alloc_chunks(int nr_irqs, int *base, int *nr_ids)
 	return bitmap;
 }
 
-static void its_lpi_free(struct event_lpi_map *map)
+static void its_lpi_free_chunks(unsigned long *bitmap, int base, int nr_ids)
 {
-	int base = map->lpi_base;
-	int nr_ids = map->nr_lpis;
 	int lpi;
 
 	spin_lock(&lpi_lock);
 
 	for (lpi = base; lpi < (base + nr_ids); lpi += IRQS_PER_CHUNK) {
 		int chunk = its_lpi_to_chunk(lpi);
+
 		BUG_ON(chunk > lpi_chunks);
 		if (test_bit(chunk, lpi_bitmap)) {
 			clear_bit(chunk, lpi_bitmap);
@@ -893,8 +892,7 @@ static void its_lpi_free(struct event_lpi_map *map)
 
 	spin_unlock(&lpi_lock);
 
-	kfree(map->lpi_map);
-	kfree(map->col_map);
+	kfree(bitmap);
 }
 
 static struct page *its_allocate_prop_table(gfp_t gfp_flags)
@@ -1665,7 +1663,10 @@ static void its_irq_domain_free(struct irq_domain *domain, unsigned int virq,
 	/* If all interrupts have been freed, start mopping the floor */
 	if (bitmap_empty(its_dev->event_map.lpi_map,
 			 its_dev->event_map.nr_lpis)) {
-		its_lpi_free(&its_dev->event_map);
+		its_lpi_free_chunks(its_dev->event_map.lpi_map,
+				    its_dev->event_map.lpi_base,
+				    its_dev->event_map.nr_lpis);
+		kfree(its_dev->event_map.col_map);
 
 		/* Unmap device/itt */
 		its_send_mapd(its_dev, 0);
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]


#1676822 — [PATCH v2 31/52] irqchip/gic-v4: Add VLPI configuration interface

FromMarc Zyngier <marc.zyngier@arm.com>
Date2017-06-28 17:20 +0200
Subject[PATCH v2 31/52] irqchip/gic-v4: Add VLPI configuration interface
Message-ID<tXmV6-8uj-65@gated-at.bofh.it>
In reply to#1676774
Add the required interfaces to map, unmap and update a VLPI.

Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 drivers/irqchip/irq-gic-v4.c       | 42 ++++++++++++++++++++++++++++++++++++++
 include/linux/irqchip/arm-gic-v4.h |  4 ++++
 2 files changed, 46 insertions(+)

diff --git a/drivers/irqchip/irq-gic-v4.c b/drivers/irqchip/irq-gic-v4.c
index 3c5ad70193ed..697d9dc1e005 100644
--- a/drivers/irqchip/irq-gic-v4.c
+++ b/drivers/irqchip/irq-gic-v4.c
@@ -81,3 +81,45 @@ int its_invall_vpe(struct its_vpe *vpe)
 
 	return its_send_vpe_cmd(vpe, &info);
 }
+
+int its_map_vlpi(int irq, struct its_vlpi_map *map)
+{
+	struct its_cmd_info info = {
+		.cmd_type = MAP_VLPI,
+		.map      = map,
+	};
+
+	/*
+	 * The host will never see that interrupt firing again, so it
+	 * is vital that we don't do any lazy masking.
+	 */
+	irq_set_status_flags(irq, IRQ_DISABLE_UNLAZY);
+
+	return irq_set_vcpu_affinity(irq, &info);
+}
+
+int its_get_vlpi(int irq, struct its_vlpi_map *map)
+{
+	struct its_cmd_info info = {
+		.cmd_type = GET_VLPI,
+		.map      = map,
+	};
+
+	return irq_set_vcpu_affinity(irq, &info);
+}
+
+int its_unmap_vlpi(int irq)
+{
+	irq_clear_status_flags(irq, IRQ_DISABLE_UNLAZY);
+	return irq_set_vcpu_affinity(irq, NULL);
+}
+
+int its_prop_update_vlpi(int irq, u8 config, bool inv)
+{
+	struct its_cmd_info info = {
+		.cmd_type = inv ? PROP_UPDATE_AND_INV_VLPI : PROP_UPDATE_VLPI,
+		.config   = config,
+	};
+
+	return irq_set_vcpu_affinity(irq, &info);
+}
diff --git a/include/linux/irqchip/arm-gic-v4.h b/include/linux/irqchip/arm-gic-v4.h
index 97bb97a11fef..843c9ff29269 100644
--- a/include/linux/irqchip/arm-gic-v4.h
+++ b/include/linux/irqchip/arm-gic-v4.h
@@ -92,5 +92,9 @@ int its_alloc_vcpu_irqs(struct its_vm *vm);
 void its_free_vcpu_irqs(struct its_vm *vm);
 int its_schedule_vpe(struct its_vpe *vpe, bool on);
 int its_invall_vpe(struct its_vpe *vpe);
+int its_map_vlpi(int irq, struct its_vlpi_map *map);
+int its_get_vlpi(int irq, struct its_vlpi_map *map);
+int its_unmap_vlpi(int irq);
+int its_prop_update_vlpi(int irq, u8 config, bool inv);
 
 #endif
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]


#1676824 — [PATCH v2 21/52] irqchip/gic-v3-its: Add VPE irq domain [de]activation

FromMarc Zyngier <marc.zyngier@arm.com>
Date2017-06-28 17:20 +0200
Subject[PATCH v2 21/52] irqchip/gic-v3-its: Add VPE irq domain [de]activation
Message-ID<tXmV6-8uj-69@gated-at.bofh.it>
In reply to#1676774
On activation, a VPE is mapped using the VMAPP command, followed
by a VINVALL for a good measure. On deactivation, the VPE is
simply unmapped.

Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 drivers/irqchip/irq-gic-v3-its.c | 102 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 102 insertions(+)

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index 58f3b19ec462..276602efdfa0 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -217,6 +217,16 @@ struct its_cmd_desc {
 
 		struct {
 			struct its_vpe *vpe;
+		} its_vinvall_cmd;
+
+		struct {
+			struct its_vpe *vpe;
+			struct its_collection *col;
+			bool valid;
+		} its_vmapp_cmd;
+
+		struct {
+			struct its_vpe *vpe;
 			struct its_device *dev;
 			u32 virt_id;
 			u32 event_id;
@@ -320,6 +330,16 @@ static void its_encode_db_valid(struct its_cmd_block *cmd, bool db_valid)
 	its_mask_encode(&cmd->raw_cmd[2], db_valid, 0, 0);
 }
 
+static void its_encode_vpt_addr(struct its_cmd_block *cmd, u64 vpt_pa)
+{
+	its_mask_encode(&cmd->raw_cmd[3], vpt_pa >> 16, 50, 16);
+}
+
+static void its_encode_vpt_size(struct its_cmd_block *cmd, u8 vpt_size)
+{
+	its_mask_encode(&cmd->raw_cmd[3], vpt_size, 4, 0);
+}
+
 static inline void its_fixup_cmd(struct its_cmd_block *cmd)
 {
 	/* Let's fixup BE commands */
@@ -478,6 +498,36 @@ static struct its_collection *its_build_invall_cmd(struct its_cmd_block *cmd,
 	return NULL;
 }
 
+static struct its_vpe *its_build_vinvall_cmd(struct its_cmd_block *cmd,
+					     struct its_cmd_desc *desc)
+{
+	its_encode_cmd(cmd, GITS_CMD_VINVALL);
+	its_encode_vpeid(cmd, desc->its_vinvall_cmd.vpe->vpe_id);
+
+	its_fixup_cmd(cmd);
+
+	return desc->its_vinvall_cmd.vpe;
+}
+
+static struct its_vpe *its_build_vmapp_cmd(struct its_cmd_block *cmd,
+					   struct its_cmd_desc *desc)
+{
+	unsigned long vpt_addr;
+
+	vpt_addr = virt_to_phys(page_address(desc->its_vmapp_cmd.vpe->vpt_page));
+
+	its_encode_cmd(cmd, GITS_CMD_VMAPP);
+	its_encode_vpeid(cmd, desc->its_vmapp_cmd.vpe->vpe_id);
+	its_encode_valid(cmd, desc->its_vmapp_cmd.valid);
+	its_encode_target(cmd, desc->its_vmapp_cmd.col->target_address);
+	its_encode_vpt_addr(cmd, vpt_addr);
+	its_encode_vpt_size(cmd, LPI_NRBITS - 1);
+
+	its_fixup_cmd(cmd);
+
+	return desc->its_vmapp_cmd.vpe;
+}
+
 static struct its_vpe *its_build_vmapti_cmd(struct its_cmd_block *cmd,
 					    struct its_cmd_desc *desc)
 {
@@ -799,6 +849,37 @@ static void its_send_vmovi(struct its_device *dev, u32 id)
 	its_send_single_vcommand(dev->its, its_build_vmovi_cmd, &desc);
 }
 
+static void its_send_vmapp(struct its_vpe *vpe, bool valid)
+{
+	struct its_cmd_desc desc;
+	struct its_node *its;
+
+	desc.its_vmapp_cmd.vpe = vpe;
+	desc.its_vmapp_cmd.valid = valid;
+
+	list_for_each_entry(its, &its_nodes, entry) {
+		if (!its->is_v4)
+			continue;
+
+		desc.its_vmapp_cmd.col = &its->collections[vpe->col_idx];
+		its_send_single_vcommand(its, its_build_vmapp_cmd, &desc);
+	}
+}
+
+static void its_send_vinvall(struct its_vpe *vpe)
+{
+	struct its_cmd_desc desc;
+	struct its_node *its;
+
+	desc.its_vinvall_cmd.vpe = vpe;
+
+	list_for_each_entry(its, &its_nodes, entry) {
+		if (!its->is_v4)
+			continue;
+		its_send_single_vcommand(its, its_build_vinvall_cmd, &desc);
+	}
+}
+
 /*
  * irqchip functions - assumes MSI, mostly.
  */
@@ -2194,9 +2275,30 @@ static int its_vpe_irq_domain_alloc(struct irq_domain *domain, unsigned int virq
 	return err;
 }
 
+static void its_vpe_irq_domain_activate(struct irq_domain *domain,
+					struct irq_data *d)
+{
+	struct its_vpe *vpe = irq_data_get_irq_chip_data(d);
+
+	/* Map the VPE to the first possible CPU */
+	vpe->col_idx = cpumask_first(cpu_online_mask);
+	its_send_vmapp(vpe, true);
+	its_send_vinvall(vpe);
+}
+
+static void its_vpe_irq_domain_deactivate(struct irq_domain *domain,
+					  struct irq_data *d)
+{
+	struct its_vpe *vpe = irq_data_get_irq_chip_data(d);
+
+	its_send_vmapp(vpe, false);
+}
+
 static const struct irq_domain_ops its_vpe_domain_ops = {
 	.alloc			= its_vpe_irq_domain_alloc,
 	.free			= its_vpe_irq_domain_free,
+	.activate		= its_vpe_irq_domain_activate,
+	.deactivate		= its_vpe_irq_domain_deactivate,
 };
 
 static int its_force_quiescent(void __iomem *base)
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]


#1676825 — [PATCH v2 25/52] irqchip/gic-v3-its: Add VPE affinity changes

FromMarc Zyngier <marc.zyngier@arm.com>
Date2017-06-28 17:20 +0200
Subject[PATCH v2 25/52] irqchip/gic-v3-its: Add VPE affinity changes
Message-ID<tXmV6-8uj-71@gated-at.bofh.it>
In reply to#1676774
When we're about to run a vcpu, it is crucial that the redistributor
associated with the physical CPU is being told about the new residency.

This is abstracted by hijacking the irq_set_affinity method for the
doorbell interrupt associated with the VPE. It is expected that the
hypervisor will call this method before scheduling the VPE.

Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 drivers/irqchip/irq-gic-v3-its.c | 96 ++++++++++++++++++++++++++++++++++++++++
 1 file changed, 96 insertions(+)

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index 0eda36dc84b4..281eed031e01 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -150,6 +150,9 @@ static struct irq_domain *its_parent;
 #define ITS_LIST_MAX		16
 
 static unsigned long its_list_map;
+static u16 vmovp_seq_num;
+static DEFINE_RAW_SPINLOCK(vmovp_lock);
+
 static DEFINE_IDA(its_vpeid_ida);
 
 #define gic_data_rdist()		(raw_cpu_ptr(gic_rdists->rdist))
@@ -240,6 +243,13 @@ struct its_cmd_desc {
 			u32 event_id;
 			bool db_enabled;
 		} its_vmovi_cmd;
+
+		struct {
+			struct its_vpe *vpe;
+			struct its_collection *col;
+			u16 seq_num;
+			u16 its_list;
+		} its_vmovp_cmd;
 	};
 };
 
@@ -331,6 +341,16 @@ static void its_encode_db_valid(struct its_cmd_block *cmd, bool db_valid)
 	its_mask_encode(&cmd->raw_cmd[2], db_valid, 0, 0);
 }
 
+static void its_encode_seq_num(struct its_cmd_block *cmd, u16 seq_num)
+{
+	its_mask_encode(&cmd->raw_cmd[0], seq_num, 47, 32);
+}
+
+static void its_encode_its_list(struct its_cmd_block *cmd, u16 its_list)
+{
+	its_mask_encode(&cmd->raw_cmd[1], its_list, 15, 0);
+}
+
 static void its_encode_vpt_addr(struct its_cmd_block *cmd, u64 vpt_pa)
 {
 	its_mask_encode(&cmd->raw_cmd[3], vpt_pa >> 16, 50, 16);
@@ -573,6 +593,20 @@ static struct its_vpe *its_build_vmovi_cmd(struct its_cmd_block *cmd,
 	return desc->its_vmovi_cmd.vpe;
 }
 
+static struct its_vpe *its_build_vmovp_cmd(struct its_cmd_block *cmd,
+					   struct its_cmd_desc *desc)
+{
+	its_encode_cmd(cmd, GITS_CMD_VMOVP);
+	its_encode_seq_num(cmd, desc->its_vmovp_cmd.seq_num);
+	its_encode_its_list(cmd, desc->its_vmovp_cmd.its_list);
+	its_encode_vpeid(cmd, desc->its_vmovp_cmd.vpe->vpe_id);
+	its_encode_target(cmd, desc->its_vmovp_cmd.col->target_address);
+
+	its_fixup_cmd(cmd);
+
+	return desc->its_vmovp_cmd.vpe;
+}
+
 static u64 its_cmd_ptr_to_offset(struct its_node *its,
 				 struct its_cmd_block *ptr)
 {
@@ -867,6 +901,48 @@ static void its_send_vmapp(struct its_vpe *vpe, bool valid)
 	}
 }
 
+static void its_send_vmovp(struct its_vpe *vpe)
+{
+	struct its_cmd_desc desc;
+	struct its_node *its;
+	unsigned long flags;
+	int col_id = vpe->col_idx;
+
+	desc.its_vmovp_cmd.vpe = vpe;
+	desc.its_vmovp_cmd.its_list = (u16)its_list_map;
+
+	if (!its_list_map) {
+		its = list_first_entry(&its_nodes, struct its_node, entry);
+		desc.its_vmovp_cmd.seq_num = 0;
+		desc.its_vmovp_cmd.col = &its->collections[col_id];
+		its_send_single_vcommand(its, its_build_vmovp_cmd, &desc);
+		return;
+	}
+
+	/*
+	 * Yet another marvel of the architecture. If using the
+	 * its_list "feature", we need to make sure that all ITSs
+	 * receive all VMOVP commands in the same order. The only way
+	 * to guarantee this is to make vmovp a serialization point.
+	 *
+	 * Wall <-- Head.
+	 */
+	raw_spin_lock_irqsave(&vmovp_lock, flags);
+
+	desc.its_vmovp_cmd.seq_num = vmovp_seq_num++;
+
+	/* Emit VMOVPs */
+	list_for_each_entry(its, &its_nodes, entry) {
+		if (!its->is_v4)
+			continue;
+
+		desc.its_vmovp_cmd.col = &its->collections[col_id];
+		its_send_single_vcommand(its, its_build_vmovp_cmd, &desc);
+	}
+
+	raw_spin_unlock_irqrestore(&vmovp_lock, flags);
+}
+
 static void its_send_vinvall(struct its_vpe *vpe)
 {
 	struct its_cmd_desc desc;
@@ -2145,6 +2221,25 @@ static const struct irq_domain_ops its_domain_ops = {
 	.deactivate		= its_irq_domain_deactivate,
 };
 
+static int its_vpe_set_affinity(struct irq_data *d,
+				const struct cpumask *mask_val,
+				bool force)
+{
+	struct its_vpe *vpe = irq_data_get_irq_chip_data(d);
+	int cpu = cpumask_first(mask_val);
+
+	/*
+	 * Changing affinity is mega expensive, so let's be as lazy as
+	 * we can and only do it if we really have to.
+	 */
+	if (vpe->col_idx != cpu) {
+		vpe->col_idx = cpu;
+		its_send_vmovp(vpe);
+	}
+
+	return IRQ_SET_MASK_OK_DONE;
+}
+
 static void its_vpe_schedule(struct its_vpe *vpe)
 {
 	void * __iomem vlpi_base = gic_data_rdist_vlpi_base();
@@ -2222,6 +2317,7 @@ static int its_vpe_set_vcpu_affinity(struct irq_data *d, void *vcpu_info)
 
 static struct irq_chip its_vpe_irq_chip = {
 	.name			= "GICv4-vpe",
+	.irq_set_affinity	= its_vpe_set_affinity,
 	.irq_set_vcpu_affinity	= its_vpe_set_vcpu_affinity,
 };
 
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]


#1676826 — [PATCH v2 18/52] irqchip/gic-v3-its: Add VLPI configuration handling

FromMarc Zyngier <marc.zyngier@arm.com>
Date2017-06-28 17:20 +0200
Subject[PATCH v2 18/52] irqchip/gic-v3-its: Add VLPI configuration handling
Message-ID<tXmV6-8uj-73@gated-at.bofh.it>
In reply to#1676774
When a VLPI is reconfigured (enabled, disabled, change in priority),
the full configuration byte must be written, and the caches invalidated.

Also, when using the irq_mask/irq_unmask methods, it is necessary
to disable the doorbell for that particular interrupt (by mapping it
to 1023) on top of clearing the Enable bit.

Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 drivers/irqchip/irq-gic-v3-its.c | 73 +++++++++++++++++++++++++++++++++++++---
 1 file changed, 69 insertions(+), 4 deletions(-)

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index 527e4e95e10d..2e5e0499a0e6 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -808,18 +808,26 @@ static inline u32 its_get_event_id(struct irq_data *d)
 	return d->hwirq - its_dev->event_map.lpi_base;
 }
 
-static void lpi_update_config(struct irq_data *d, u8 clr, u8 set)
+static void lpi_write_config(struct irq_data *d, u8 clr, u8 set)
 {
 	struct its_device *its_dev = irq_data_get_irq_chip_data(d);
-	irq_hw_number_t hwirq = d->hwirq;
+	irq_hw_number_t hwirq;
 	struct page *prop_page;
 	u8 *cfg;
 
-	prop_page = gic_rdists->prop_page;
+	if (irqd_is_forwarded_to_vcpu(d)) {
+		u32 event = its_get_event_id(d);
+
+		prop_page = its_dev->event_map.vm->vprop_page;
+		hwirq = its_dev->event_map.vlpi_maps[event].vintid;
+	} else {
+		prop_page = gic_rdists->prop_page;
+		hwirq = d->hwirq;
+	}
 
 	cfg = page_address(prop_page) + hwirq - 8192;
 	*cfg &= ~clr;
-	*cfg |= set;
+	*cfg |= set | LPI_PROP_GROUP1;
 
 	/*
 	 * Make the above write visible to the redistributors.
@@ -830,16 +838,52 @@ static void lpi_update_config(struct irq_data *d, u8 clr, u8 set)
 		gic_flush_dcache_to_poc(cfg, sizeof(*cfg));
 	else
 		dsb(ishst);
+}
+
+static void lpi_update_config(struct irq_data *d, u8 clr, u8 set)
+{
+	struct its_device *its_dev = irq_data_get_irq_chip_data(d);
+
+	lpi_write_config(d, clr, set);
 	its_send_inv(its_dev, its_get_event_id(d));
 }
 
+static void its_vlpi_set_doorbell(struct irq_data *d, bool enable)
+{
+	struct its_device *its_dev = irq_data_get_irq_chip_data(d);
+	u32 event = its_get_event_id(d);
+
+	if (its_dev->event_map.vlpi_maps[event].db_enabled == enable)
+		return;
+
+	its_dev->event_map.vlpi_maps[event].db_enabled = enable;
+
+	/*
+	 * More fun with the architecture:
+	 *
+	 * Ideally, we'd issue a VMAPTI to set the doorbell to its LPI
+	 * value or to 1023, depending on the enable bit. But that
+	 * would be issueing a mapping for an /existing/ DevID+EventID
+	 * pair, which is UNPREDICTABLE. Instead, let's issue a VMOVI
+	 * to the /same/ vPE, using this opportunity to adjust the
+	 * doorbell. Mouahahahaha. We loves it, Precious.
+	 */
+	its_send_vmovi(its_dev, event);
+}
+
 static void its_mask_irq(struct irq_data *d)
 {
+	if (irqd_is_forwarded_to_vcpu(d))
+		its_vlpi_set_doorbell(d, false);
+
 	lpi_update_config(d, LPI_PROP_ENABLED, 0);
 }
 
 static void its_unmask_irq(struct irq_data *d)
 {
+	if (irqd_is_forwarded_to_vcpu(d))
+		its_vlpi_set_doorbell(d, true);
+
 	lpi_update_config(d, 0, LPI_PROP_ENABLED);
 }
 
@@ -852,6 +896,10 @@ static int its_set_affinity(struct irq_data *d, const struct cpumask *mask_val,
 	struct its_collection *target_col;
 	u32 id = its_get_event_id(d);
 
+	/* A forwarded interrupt should use irq_set_vcpu_affinity */
+	if (irqd_is_forwarded_to_vcpu(d))
+		return -EINVAL;
+
        /* lpi cannot be routed to a redistributor that is on a foreign node */
 	if (its_dev->its->flags & ITS_FLAGS_WORKAROUND_CAVIUM_23144) {
 		if (its_dev->its->numa_node >= 0) {
@@ -1017,6 +1065,22 @@ static int its_vlpi_unmap(struct irq_data *d)
 	return ret;
 }
 
+static int its_vlpi_prop_update(struct irq_data *d, struct its_cmd_info *info)
+{
+	struct its_device *its_dev = irq_data_get_irq_chip_data(d);
+
+	if (!its_dev->event_map.vm || !irqd_is_forwarded_to_vcpu(d))
+		return -EINVAL;
+
+	if (info->cmd_type == PROP_UPDATE_AND_INV_VLPI)
+		lpi_update_config(d, 0xff, info->config);
+	else
+		lpi_write_config(d, 0xff, info->config);
+	its_vlpi_set_doorbell(d, !!(info->config & LPI_PROP_ENABLED));
+
+	return 0;
+}
+
 static int its_irq_set_vcpu_affinity(struct irq_data *d, void *vcpu_info)
 {
 	struct its_device *its_dev = irq_data_get_irq_chip_data(d);
@@ -1039,6 +1103,7 @@ static int its_irq_set_vcpu_affinity(struct irq_data *d, void *vcpu_info)
 
 	case PROP_UPDATE_VLPI:
 	case PROP_UPDATE_AND_INV_VLPI:
+		return its_vlpi_prop_update(d, info);
 
 	default:
 		return -EINVAL;
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]


#1676827 — [PATCH v2 16/52] irqchip/gic-v3-its: Add VLPI configuration hook

FromMarc Zyngier <marc.zyngier@arm.com>
Date2017-06-28 17:20 +0200
Subject[PATCH v2 16/52] irqchip/gic-v3-its: Add VLPI configuration hook
Message-ID<tXmV6-8uj-77@gated-at.bofh.it>
In reply to#1676774
Add the skeleton irq_set_vcpu_affinity method that will be used
to configure VLPIs.

Reviewed-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 drivers/irqchip/irq-gic-v3-its.c | 24 ++++++++++++++++++++++++
 1 file changed, 24 insertions(+)

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index 25e53958a618..c02e1e6cd169 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -36,6 +36,7 @@
 
 #include <linux/irqchip.h>
 #include <linux/irqchip/arm-gic-v3.h>
+#include <linux/irqchip/arm-gic-v4.h>
 
 #include <asm/cputype.h>
 #include <asm/exception.h>
@@ -779,6 +780,28 @@ static int its_irq_set_irqchip_state(struct irq_data *d,
 	return 0;
 }
 
+static int its_irq_set_vcpu_affinity(struct irq_data *d, void *vcpu_info)
+{
+	struct its_device *its_dev = irq_data_get_irq_chip_data(d);
+	struct its_cmd_info *info = vcpu_info;
+
+	/* Need a v4 ITS */
+	if (!its_dev->its->is_v4 || !info)
+		return -EINVAL;
+
+	switch (info->cmd_type) {
+	case MAP_VLPI:
+
+	case GET_VLPI:
+
+	case PROP_UPDATE_VLPI:
+	case PROP_UPDATE_AND_INV_VLPI:
+
+	default:
+		return -EINVAL;
+	}
+}
+
 static struct irq_chip its_irq_chip = {
 	.name			= "ITS",
 	.irq_mask		= its_mask_irq,
@@ -787,6 +810,7 @@ static struct irq_chip its_irq_chip = {
 	.irq_set_affinity	= its_set_affinity,
 	.irq_compose_msi_msg	= its_irq_compose_msi_msg,
 	.irq_set_irqchip_state	= its_irq_set_irqchip_state,
+	.irq_set_vcpu_affinity	= its_irq_set_vcpu_affinity,
 };
 
 /*
-- 
2.11.0

[toc] | [prev] | [next] | [standalone]


Page 2 of 3 — ← Prev page 1 [2] 3  Next page →

Back to top | Article view | linux.kernel


csiph-web