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

Groups > linux.kernel > #1700219 > unrolled thread

[PATCH v3 00/59] irqchip: KVM: Add support for GICv4

Back to article view | Back to linux.kernel

Contents

  [PATCH v3 00/59] irqchip: KVM: Add support for GICv4 Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 10/59] irqchip/gic-v3-its: Split out pending table allocation Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 37/59] KVM: arm/arm64: vgic: Move kvm_vgic_destroy call around Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 12/59] irqchip/gic-v3-its: Generalize device table allocation Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 55/59] KVM: arm/arm64: GICv4: Enable VLPI support Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 14/59] irqchip/gic-v4: Add management structure definitions Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 54/59] KVM: arm/arm64: GICv4: Enable virtual cpuif if VLPIs can be delivered Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 45/59] KVM: arm/arm64: GICv4: Handle CLEAR applied to a VLPI Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 49/59] KVM: arm/arm64: GICv4: Propagate VLPI properties at map time Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 59/59] irqchip/gic-v3-its: Workaround Huawei D05 redistributor addressing Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 47/59] KVM: arm/arm64: GICv4: Propagate property updates to VLPIs Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 05/59] irqchip/gic-v3-its: Add probing for VLPI properties Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 29/59] irqchip/gic-v3-its: Set implementation defined bit to enable VLPIs Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 33/59] irqchip/gic-v4: Add some basic documentation Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 35/59] irqchip/gic-v3: Advertise GICv4 support to KVM Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 41/59] KVM: arm/arm64: GICv4: Wire mapping/unmapping of VLPIs in VFIO irq bypass Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 58/59] irqchip/gic-v3-its: Pass its_node pointer to each command bulder Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 39/59] KVM: arm/arm64: GICv4: Add property field and per-VM predicate Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 43/59] KVM: arm/arm64: GICv4: Unmap VLPI when freeing an LPI Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 31/59] irqchip/gic-v4: Add VPE command interface Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 24/59] irqchip/gic-v3-its: Add VPE invalidation hook Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 19/59] irqchip/gic-v3-its: Add VPE domain infrastructure Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 40/59] KVM: arm/arm64: GICv4: Add init/teardown of the per-VM vPE irq domain Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:30 +0200
    [PATCH v3 52/59] KVM: arm/arm64: GICv4: Use the doorbell interrupt as an unblocking source Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:40 +0200
    [PATCH v3 53/59] KVM: arm/arm64: GICv4: Hook vPE scheduling into vgic flush/sync Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:40 +0200
    [PATCH v3 44/59] KVM: arm/arm64: GICv4: Handle MOVI applied to a VLPI Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:40 +0200
    [PATCH v3 48/59] KVM: arm/arm64: GICv4: Handle INVALL applied to a vPE Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:40 +0200
    [PATCH v3 51/59] KVM: arm/arm64: GICv4: Add doorbell interrupt handling Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:40 +0200
      Re: [PATCH v3 51/59] KVM: arm/arm64: GICv4: Add doorbell interrupt  handling Marc Zyngier <marc.zyngier@arm.com> - 2017-08-04 09:50 +0200
    [PATCH v3 42/59] KVM: arm/arm64: GICv4: Handle INT command applied to a VLPI Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:40 +0200
    [PATCH v3 46/59] KVM: arm/arm64: GICv4: Handle MOVALL applied to a vPE Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:40 +0200
    [PATCH v3 56/59] KVM: arm/arm64: GICv4: Prevent heterogenous systems from using GICv4 Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:40 +0200
    [PATCH v3 50/59] KVM: arm/arm64: GICv4: Use pending_last as a scheduling hint Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:40 +0200
    [PATCH v3 57/59] KVM: arm/arm64: GICv4: Theory of operations Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 19:40 +0200
    [PATCH v3 38/59] KVM: arm/arm64: vITS: Add MSI translation helpers Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 32/59] irqchip/gic-v4: Add VLPI configuration interface Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 08/59] irqchip/gic-v3-its: Split out property table allocation Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 30/59] irqchip/gic-v4: Add per-VM VPE domain creation Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 36/59] KVM: arm: Select ARM_GIC_V3 and ARM_GIC_V3_ITS Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 09/59] irqchip/gic-v3-its: Allow use of indirect VCPU tables Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 13/59] irqchip/gic-v3-its: Generalize LPI configuration Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 34/59] irqchip/gic-v4: Enable low-level GICv4 operations Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 22/59] irqchip/gic-v3-its: Add VPENDBASER/VPROPBASER accessors Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 15/59] irqchip/gic-v3-its: Add GICv4 ITS command definitions Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 27/59] irqchip/gic-v3-its: Support VPE doorbell invalidation even when !DirectLPI Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 26/59] irqchip/gic-v3-its: Add VPE interrupt masking Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 16/59] irqchip/gic-v3-its: Add VLPI configuration hook Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 28/59] irqchip/gic-v3-its: Allow doorbell interrupts to be injected/cleared Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 06/59] irqchip/gic-v3-its: Macro-ize its_send_single_command Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 25/59] irqchip/gic-v3-its: Add VPE affinity changes Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 17/59] irqchip/gic-v3-its: Add VLPI map/unmap operations Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 11/59] irqchip/gic-v3-its: Rework LPI freeing Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 20/59] irqchip/gic-v3-its: Add VPE irq domain allocation/teardown Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 18/59] irqchip/gic-v3-its: Add VLPI configuration handling Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 23/59] irqchip/gic-v3-its: Add VPE scheduling Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:00 +0200
    [PATCH v3 02/59] irqchip/gic-v3: Add redistributor iterator Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:10 +0200
    [PATCH v3 01/59] genirq: Let irq_set_vcpu_affinity() iterate over hierarchy Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:10 +0200
    [PATCH v3 07/59] irqchip/gic-v3-its: Implement irq_set_irqchip_state for pending state Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:10 +0200
    [PATCH v3 03/59] irqchip/gic-v3: Add VLPI/DirectLPI discovery Marc Zyngier <marc.zyngier@arm.com> - 2017-07-31 20:10 +0200

Page 3 of 3 — ← Prev page 1 2 [3]

#1700276 — [PATCH v3 13/59] irqchip/gic-v3-its: Generalize LPI configuration

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 1c2d6bdf85cb..7ef448a963c1 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -678,17 +678,18 @@ 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.
@@ -699,17 +700,17 @@ 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]

#1700277 — [PATCH v3 34/59] irqchip/gic-v4: Enable low-level GICv4 operations

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 e88d856cc09c..7dd05ad6bcfb 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 a5b1317d89eb..9f32cd42238f 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -3109,7 +3109,8 @@ 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 d10d73477d01..400d9bbb0e56 100644
--- a/drivers/irqchip/irq-gic-v4.c
+++ b/drivers/irqchip/irq-gic-v4.c
@@ -209,3 +209,16 @@ 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 f495d4d1f7d7..3cf72af2c796 100644
--- a/include/linux/irqchip/arm-gic-v4.h
+++ b/include/linux/irqchip/arm-gic-v4.h
@@ -98,4 +98,6 @@ 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]

#1700278 — [PATCH v3 22/59] irqchip/gic-v3-its: Add VPENDBASER/VPROPBASER accessors

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]

#1700279 — [PATCH v3 15/59] irqchip/gic-v3-its: Add GICv4 ITS command definitions

Add the new GICv4 ITS command definitions, most of them, being
defined in terms of their physical counterparts.

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   |  2 +-
 include/linux/irqchip/arm-gic-v3.h | 12 ++++++++++++
 2 files changed, 13 insertions(+), 1 deletion(-)

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index 7ef448a963c1..bc396ab24037 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -1,5 +1,5 @@
 /*
- * Copyright (C) 2013, 2014 ARM Limited, All Rights Reserved.
+ * Copyright (C) 2013-2017 ARM Limited, All Rights Reserved.
  * Author: Marc Zyngier <marc.zyngier@arm.com>
  *
  * This program is free software; you can redistribute it and/or modify
diff --git a/include/linux/irqchip/arm-gic-v3.h b/include/linux/irqchip/arm-gic-v3.h
index af8c55105fc2..7c6fd8f3e36c 100644
--- a/include/linux/irqchip/arm-gic-v3.h
+++ b/include/linux/irqchip/arm-gic-v3.h
@@ -348,6 +348,18 @@
 #define GITS_CMD_SYNC			0x05
 
 /*
+ * GICv4 ITS specific commands
+ */
+#define GITS_CMD_GICv4(x)		((x) | 0x20)
+#define GITS_CMD_VINVALL		GITS_CMD_GICv4(GITS_CMD_INVALL)
+#define GITS_CMD_VMAPP			GITS_CMD_GICv4(GITS_CMD_MAPC)
+#define GITS_CMD_VMAPTI			GITS_CMD_GICv4(GITS_CMD_MAPTI)
+#define GITS_CMD_VMOVI			GITS_CMD_GICv4(GITS_CMD_MOVI)
+#define GITS_CMD_VSYNC			GITS_CMD_GICv4(GITS_CMD_SYNC)
+/* VMOVP is the odd one, as it doesn't have a physical counterpart */
+#define GITS_CMD_VMOVP			GITS_CMD_GICv4(2)
+
+/*
  * ITS error numbers
  */
 #define E_ITS_MOVI_UNMAPPED_INTERRUPT		0x010107
-- 
2.11.0

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

#1700280 — [PATCH v3 27/59] irqchip/gic-v3-its: Support VPE doorbell invalidation even when !DirectLPI

When we don't have the DirectLPI feature, we must work around the
architecture shortcomings to be able to perform the required
invalidation.

For this, we create a fake device whose sole purpose is to
provide a way to issue a map/inv/unmap sequence (and the corresponding
sync operations). That's 6 commands and a full serialization point
to be able to do this.

You just have to hope the hypervisor won't do that too often...

Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 drivers/irqchip/irq-gic-v3-its.c | 73 ++++++++++++++++++++++++++++++++++++++--
 1 file changed, 70 insertions(+), 3 deletions(-)

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index d871d105be9d..210840e0ec92 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -136,6 +136,9 @@ struct its_device {
 	u32			device_id;
 };
 
+static struct its_device *vpe_proxy_dev;
+static DEFINE_RAW_SPINLOCK(vpe_proxy_dev_lock);
+
 static LIST_HEAD(its_nodes);
 static DEFINE_SPINLOCK(its_lock);
 static struct rdists *gic_rdists;
@@ -2077,6 +2080,14 @@ static int its_msi_prepare(struct irq_domain *domain, struct device *dev,
 	msi_info = msi_get_domain_info(domain);
 	its = msi_info->data;
 
+	if (its->is_v4 && !gic_rdists->has_direct_lpi &&
+	    dev_id == vpe_proxy_dev->device_id) {
+		/* Bad luck. Get yourself a better implementation */
+		WARN_ONCE(1, "DevId %x clashes with GICv4 VPE proxy device\n",
+			  dev_id);
+		return -EINVAL;
+	}
+
 	its_dev = its_find_device(its, dev_id);
 	if (its_dev) {
 		/*
@@ -2330,13 +2341,48 @@ static int its_vpe_set_vcpu_affinity(struct irq_data *d, void *vcpu_info)
 	}
 }
 
+static void its_vpe_send_cmd(struct its_vpe *vpe,
+			     void (*cmd)(struct its_device *, u32))
+{
+	unsigned long flags;
+
+	raw_spin_lock_irqsave(&vpe_proxy_dev_lock, flags);
+
+	vpe_proxy_dev->event_map.col_map[0] = vpe->col_idx;
+	its_send_mapti(vpe_proxy_dev, vpe->vpe_db_lpi, 0);
+	cmd(vpe_proxy_dev, 0);
+	its_send_discard(vpe_proxy_dev, 0);
+
+	raw_spin_unlock_irqrestore(&vpe_proxy_dev_lock, flags);
+}
+
 static void its_vpe_send_inv(struct irq_data *d)
 {
 	struct its_vpe *vpe = irq_data_get_irq_chip_data(d);
-	void __iomem *rdbase;
 
-	rdbase = per_cpu_ptr(gic_rdists->rdist, vpe->col_idx)->rd_base;
-	gic_write_lpir(vpe->vpe_db_lpi, rdbase + GICR_INVLPIR);
+	if (gic_rdists->has_direct_lpi) {
+		void __iomem *rdbase;
+
+		rdbase = per_cpu_ptr(gic_rdists->rdist, vpe->col_idx)->rd_base;
+		gic_write_lpir(vpe->vpe_db_lpi, rdbase + GICR_INVLPIR);
+	} else {
+		/*
+		 * This is insane.
+		 *
+		 * If a GICv4 doesn't implement Direct LPIs (which is
+		 * extremely likely), the only way to perform an
+		 * invalidate is to use a fake device to issue a
+		 * MAP/INV/UNMAP sequence. Since each of these
+		 * commands has a sync operation, this is really
+		 * fast. Not.
+		 *
+		 * We always use event 0, and thus serialize all VPE
+		 * invalidations in the system.
+		 *
+		 * Broken by design(tm).
+		 */
+		its_vpe_send_cmd(vpe, its_send_inv);
+	}
 }
 
 static void its_vpe_mask_irq(struct irq_data *d)
@@ -2638,6 +2684,27 @@ static int its_init_domain(struct fwnode_handle *handle, struct its_node *its)
 
 static int its_init_vpe_domain(void)
 {
+	struct its_node *its;
+	u32 devid;
+
+	if (gic_rdists->has_direct_lpi) {
+		pr_info("ITS: Using DirectLPI for VPE invalidation\n");
+		return 0;
+	}
+
+	/* Any ITS will do, even if not v4 */
+	its = list_first_entry(&its_nodes, struct its_node, entry);
+
+	/* Use the last possible DevID */
+	devid = GENMASK(its->device_ids - 1, 0);
+	vpe_proxy_dev = its_create_device(its, devid, 1);
+	if (!vpe_proxy_dev) {
+		pr_err("ITS: Can't allocate GICv4 proxy device\n");
+		return -ENODEV;
+	}
+
+	pr_info("ITS: Allocated DevID %x as GICv4 proxy device\n", devid);
+
 	return 0;
 }
 
-- 
2.11.0

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

#1700282 — [PATCH v3 26/59] irqchip/gic-v3-its: Add VPE interrupt masking

When masking/unmasking a doorbell interrupt, it is necessary
to issue an invalidation to the corresponding redistributor.
We use the DirectLPI feature by writting directly to the corresponding
redistributor.

Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 arch/arm/include/asm/arch_gicv3.h   |  5 +++++
 arch/arm64/include/asm/arch_gicv3.h |  1 +
 drivers/irqchip/irq-gic-v3-its.c    | 31 +++++++++++++++++++++++++++++++
 3 files changed, 37 insertions(+)

diff --git a/arch/arm/include/asm/arch_gicv3.h b/arch/arm/include/asm/arch_gicv3.h
index 8d45e88feac9..550fb0bcccf8 100644
--- a/arch/arm/include/asm/arch_gicv3.h
+++ b/arch/arm/include/asm/arch_gicv3.h
@@ -276,6 +276,11 @@ static inline u64 __gic_readq_nonatomic(const volatile void __iomem *addr)
 #define gicr_write_pendbaser(v, c)	__gic_writeq_nonatomic(v, c)
 
 /*
+ * GICR_xLPIR - only the lower bits are significant
+ */
+#define gic_write_lpir(v, c)		writel_relaxed(lower_32_bits(v), c)
+
+/*
  * GITS_TYPER is an ID register and doesn't need atomicity.
  */
 #define gits_read_typer(c)		__gic_readq_nonatomic(c)
diff --git a/arch/arm64/include/asm/arch_gicv3.h b/arch/arm64/include/asm/arch_gicv3.h
index 0d2a53457c30..c2de3ff374a0 100644
--- a/arch/arm64/include/asm/arch_gicv3.h
+++ b/arch/arm64/include/asm/arch_gicv3.h
@@ -116,6 +116,7 @@ static inline void gic_write_bpr1(u32 val)
 
 #define gic_read_typer(c)		readq_relaxed(c)
 #define gic_write_irouter(v, c)		writeq_relaxed(v, c)
+#define gic_write_lpir(v, c)		writeq_relaxed(v, c)
 
 #define gic_flush_dcache_to_poc(a,l)	__flush_dcache_area((a), (l))
 
diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index 0a9aedaf6da2..d871d105be9d 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -2330,8 +2330,39 @@ static int its_vpe_set_vcpu_affinity(struct irq_data *d, void *vcpu_info)
 	}
 }
 
+static void its_vpe_send_inv(struct irq_data *d)
+{
+	struct its_vpe *vpe = irq_data_get_irq_chip_data(d);
+	void __iomem *rdbase;
+
+	rdbase = per_cpu_ptr(gic_rdists->rdist, vpe->col_idx)->rd_base;
+	gic_write_lpir(vpe->vpe_db_lpi, rdbase + GICR_INVLPIR);
+}
+
+static void its_vpe_mask_irq(struct irq_data *d)
+{
+	/*
+	 * We need to unmask the LPI, which is described by the parent
+	 * irq_data. Instead of calling into the parent (which won't
+	 * exactly do the right thing, let's simply use the
+	 * parent_data pointer. Yes, I'm naughty.
+	 */
+	lpi_write_config(d->parent_data, LPI_PROP_ENABLED, 0);
+	its_vpe_send_inv(d);
+}
+
+static void its_vpe_unmask_irq(struct irq_data *d)
+{
+	/* Same hack as above... */
+	lpi_write_config(d->parent_data, 0, LPI_PROP_ENABLED);
+	its_vpe_send_inv(d);
+}
+
 static struct irq_chip its_vpe_irq_chip = {
 	.name			= "GICv4-vpe",
+	.irq_mask		= its_vpe_mask_irq,
+	.irq_unmask		= its_vpe_unmask_irq,
+	.irq_eoi		= irq_chip_eoi_parent,
 	.irq_set_affinity	= its_vpe_set_affinity,
 	.irq_set_vcpu_affinity	= its_vpe_set_vcpu_affinity,
 };
-- 
2.11.0

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

#1700283 — [PATCH v3 16/59] irqchip/gic-v3-its: Add VLPI configuration hook

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 bc396ab24037..9ef4f161690a 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>
@@ -780,6 +781,28 @@ 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,
@@ -788,6 +811,7 @@ 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]

#1700284 — [PATCH v3 28/59] irqchip/gic-v3-its: Allow doorbell interrupts to be injected/cleared

While the doorbell interrupts are usually driven by the HW itself,
having a way to trigger them independently has proved to be a
really useful debug feature. As it is actually very little code,
let's add it to the VPE irqchip operations.

Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 drivers/irqchip/irq-gic-v3-its.c | 29 +++++++++++++++++++++++++++++
 1 file changed, 29 insertions(+)

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index 210840e0ec92..a177af0b1657 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -2404,12 +2404,41 @@ static void its_vpe_unmask_irq(struct irq_data *d)
 	its_vpe_send_inv(d);
 }
 
+/* This is really a debug feature */
+static int its_vpe_set_irqchip_state(struct irq_data *d,
+				     enum irqchip_irq_state which,
+				     bool state)
+{
+	struct its_vpe *vpe = irq_data_get_irq_chip_data(d);
+
+	if (which != IRQCHIP_STATE_PENDING)
+		return -EINVAL;
+
+	if (gic_rdists->has_direct_lpi) {
+		void __iomem *rdbase;
+
+		rdbase = per_cpu_ptr(gic_rdists->rdist, vpe->col_idx)->rd_base;
+		if (state)
+			gic_write_lpir(vpe->vpe_db_lpi, rdbase + GICR_SETLPIR);
+		else
+			gic_write_lpir(vpe->vpe_db_lpi, rdbase + GICR_CLRLPIR);
+	} else {
+		if (state)
+			its_vpe_send_cmd(vpe, its_send_int);
+		else
+			its_vpe_send_cmd(vpe, its_send_clear);
+	}
+
+	return 0;
+}
+
 static struct irq_chip its_vpe_irq_chip = {
 	.name			= "GICv4-vpe",
 	.irq_mask		= its_vpe_mask_irq,
 	.irq_unmask		= its_vpe_unmask_irq,
 	.irq_eoi		= irq_chip_eoi_parent,
 	.irq_set_affinity	= its_vpe_set_affinity,
+	.irq_set_irqchip_state	= its_vpe_set_irqchip_state,
 	.irq_set_vcpu_affinity	= its_vpe_set_vcpu_affinity,
 };
 
-- 
2.11.0

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

#1700285 — [PATCH v3 06/59] irqchip/gic-v3-its: Macro-ize its_send_single_command

Most ITS commands do operate on a collection object, and require
a SYNC command to be performed on that collection in order to
guarantee the execution of the first command.

With GICv4 ITS, another set of commands perform similar operations
on a VPE object, and a VSYNC operations must be executed to guarantee
their execution.

Given the similarities (post a command, perform a synchronization
operation on a sync object), it makes sense to reuse the same
mechanism for both class of commands.

Let's start with turning its_send_single_command into a huge macro
that performs the bulk of the work, and a set of helpers that
make this macro usable for the GICv3 ITS commands.

Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 drivers/irqchip/irq-gic-v3-its.c | 84 ++++++++++++++++++++++------------------
 1 file changed, 47 insertions(+), 37 deletions(-)

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index e7a7dc9ad585..0f345c0fa700 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -488,43 +488,53 @@ static void its_wait_for_range_completion(struct its_node *its,
 	}
 }
 
-static void its_send_single_command(struct its_node *its,
-				    its_cmd_builder_t builder,
-				    struct its_cmd_desc *desc)
-{
-	struct its_cmd_block *cmd, *sync_cmd, *next_cmd;
-	struct its_collection *sync_col;
-	unsigned long flags;
-
-	raw_spin_lock_irqsave(&its->lock, flags);
-
-	cmd = its_allocate_entry(its);
-	if (!cmd) {		/* We're soooooo screewed... */
-		pr_err_ratelimited("ITS can't allocate, dropping command\n");
-		raw_spin_unlock_irqrestore(&its->lock, flags);
-		return;
-	}
-	sync_col = builder(cmd, desc);
-	its_flush_cmd(its, cmd);
-
-	if (sync_col) {
-		sync_cmd = its_allocate_entry(its);
-		if (!sync_cmd) {
-			pr_err_ratelimited("ITS can't SYNC, skipping\n");
-			goto post;
-		}
-		its_encode_cmd(sync_cmd, GITS_CMD_SYNC);
-		its_encode_target(sync_cmd, sync_col->target_address);
-		its_fixup_cmd(sync_cmd);
-		its_flush_cmd(its, sync_cmd);
-	}
-
-post:
-	next_cmd = its_post_commands(its);
-	raw_spin_unlock_irqrestore(&its->lock, flags);
-
-	its_wait_for_range_completion(its, cmd, next_cmd);
-}
+/* Warning, macro hell follows */
+#define BUILD_SINGLE_CMD_FUNC(name, buildtype, synctype, buildfn)	\
+void name(struct its_node *its,						\
+	  buildtype builder,						\
+	  struct its_cmd_desc *desc)					\
+{									\
+	struct its_cmd_block *cmd, *sync_cmd, *next_cmd;		\
+	synctype *sync_obj;						\
+	unsigned long flags;						\
+									\
+	raw_spin_lock_irqsave(&its->lock, flags);			\
+									\
+	cmd = its_allocate_entry(its);					\
+	if (!cmd) {		/* We're soooooo screewed... */		\
+		raw_spin_unlock_irqrestore(&its->lock, flags);		\
+		return;							\
+	}								\
+	sync_obj = builder(cmd, desc);					\
+	its_flush_cmd(its, cmd);					\
+									\
+	if (sync_obj) {							\
+		sync_cmd = its_allocate_entry(its);			\
+		if (!sync_cmd)						\
+			goto post;					\
+									\
+		buildfn(sync_cmd, sync_obj);				\
+		its_flush_cmd(its, sync_cmd);				\
+	}								\
+									\
+post:									\
+	next_cmd = its_post_commands(its);				\
+	raw_spin_unlock_irqrestore(&its->lock, flags);			\
+									\
+	its_wait_for_range_completion(its, cmd, next_cmd);		\
+}
+
+static void its_build_sync_cmd(struct its_cmd_block *sync_cmd,
+			       struct its_collection *sync_col)
+{
+	its_encode_cmd(sync_cmd, GITS_CMD_SYNC);
+	its_encode_target(sync_cmd, sync_col->target_address);
+
+	its_fixup_cmd(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_send_inv(struct its_device *dev, u32 event_id)
 {
-- 
2.11.0

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

#1700286 — [PATCH v3 25/59] irqchip/gic-v3-its: Add VPE affinity changes

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 16cdd1f60ebf..0a9aedaf6da2 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -148,6 +148,9 @@ 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))
@@ -238,6 +241,13 @@ 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;
 	};
 };
 
@@ -329,6 +339,16 @@ 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);
@@ -571,6 +591,20 @@ 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)
 {
@@ -865,6 +899,48 @@ 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;
@@ -2148,6 +2224,25 @@ 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();
@@ -2237,6 +2332,7 @@ 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]

#1700288 — [PATCH v3 17/59] irqchip/gic-v3-its: Add VLPI map/unmap operations

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 9ef4f161690a..ef31c3e58c01 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -115,11 +115,17 @@ 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;
@@ -205,6 +211,21 @@ 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;
 	};
 };
 
@@ -221,6 +242,9 @@ 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);
@@ -273,6 +297,26 @@ 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 */
@@ -431,6 +475,50 @@ 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)
 {
@@ -576,6 +664,18 @@ 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;
@@ -669,6 +769,33 @@ 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.
  */
@@ -781,19 +908,135 @@ 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:
@@ -1512,6 +1755,7 @@ 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]

#1700289 — [PATCH v3 11/59] irqchip/gic-v3-its: Rework LPI freeing

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 ebdb1603b95d..1298a9bca56a 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -875,16 +875,15 @@ 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);
@@ -895,8 +894,7 @@ 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)
@@ -1668,7 +1666,10 @@ 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]

#1700290 — [PATCH v3 20/59] irqchip/gic-v3-its: Add VPE irq domain allocation/teardown

When creating a VM, the low level GICv4 code is responsible for:
- allocating each VPE a unique VPEID
- allocating a doorbell interrupt for each VPE
- allocating the pending tables for each VPE
- allocating the property table for the VM

This of course has to be reversed when the VM is brought down.

All of this is wired into the irq domain alloc/free methods.

Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 drivers/irqchip/irq-gic-v3-its.c | 169 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 169 insertions(+)

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index 1fa345d53088..0f9a487448b9 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -148,6 +148,7 @@ static struct irq_domain *its_parent;
 #define ITS_LIST_MAX		16
 
 static unsigned long its_list_map;
+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)
@@ -1249,6 +1250,11 @@ static struct page *its_allocate_prop_table(gfp_t gfp_flags)
 	return prop_page;
 }
 
+static void its_free_prop_table(struct page *prop_page)
+{
+	free_pages((unsigned long)page_address(prop_page),
+		   get_order(LPI_PROPBASE_SZ));
+}
 
 static int __init its_alloc_lpi_tables(void)
 {
@@ -1551,6 +1557,12 @@ static struct page *its_allocate_pending_table(gfp_t gfp_flags)
 	return pend_page;
 }
 
+static void its_free_pending_table(struct page *pt)
+{
+	free_pages((unsigned long)page_address(pt),
+		   get_order(max_t(u32, LPI_PENDBASE_SZ, SZ_64K)));
+}
+
 static void its_cpu_init_lpis(void)
 {
 	void __iomem *rbase = gic_data_rdist_rd_base();
@@ -1773,6 +1785,34 @@ static bool its_alloc_device_table(struct its_node *its, u32 dev_id)
 	return its_alloc_table_entry(baser, dev_id);
 }
 
+static bool its_alloc_vpe_table(u32 vpe_id)
+{
+	struct its_node *its;
+
+	/*
+	 * Make sure the L2 tables are allocated on *all* v4 ITSs. We
+	 * could try and only do it on ITSs corresponding to devices
+	 * that have interrupts targeted at this VPE, but the
+	 * complexity becomes crazy (and you have tons of memory
+	 * anyway, right?).
+	 */
+	list_for_each_entry(its, &its_nodes, entry) {
+		struct its_baser *baser;
+
+		if (!its->is_v4)
+			continue;
+
+		baser = its_get_baser(its, GITS_BASER_TYPE_VCPU);
+		if (!baser)
+			return false;
+
+		if (!its_alloc_table_entry(baser, vpe_id))
+			return false;
+	}
+
+	return true;
+}
+
 static struct its_device *its_create_device(struct its_node *its, u32 dev_id,
 					    int nvecs)
 {
@@ -2030,7 +2070,136 @@ static struct irq_chip its_vpe_irq_chip = {
 	.name			= "GICv4-vpe",
 };
 
+static int its_vpe_id_alloc(void)
+{
+	return ida_simple_get(&its_vpeid_ida, 0, 1 << 16, GFP_KERNEL);
+}
+
+static void its_vpe_id_free(u16 id)
+{
+	ida_simple_remove(&its_vpeid_ida, id);
+}
+
+static int its_vpe_init(struct its_vpe *vpe)
+{
+	struct page *vpt_page;
+	int vpe_id;
+
+	/* Allocate vpe_id */
+	vpe_id = its_vpe_id_alloc();
+	if (vpe_id < 0)
+		return vpe_id;
+
+	/* Allocate VPT */
+	vpt_page = its_allocate_pending_table(GFP_KERNEL);
+	if (!vpt_page) {
+		its_vpe_id_free(vpe_id);
+		return -ENOMEM;
+	}
+
+	if (!its_alloc_vpe_table(vpe_id)) {
+		its_vpe_id_free(vpe_id);
+		its_free_pending_table(vpe->vpt_page);
+		return -ENOMEM;
+	}
+
+	vpe->vpe_id = vpe_id;
+	vpe->vpt_page = vpt_page;
+
+	return 0;
+}
+
+static void its_vpe_teardown(struct its_vpe *vpe)
+{
+	its_vpe_id_free(vpe->vpe_id);
+	its_free_pending_table(vpe->vpt_page);
+}
+
+static void its_vpe_irq_domain_free(struct irq_domain *domain,
+				    unsigned int virq,
+				    unsigned int nr_irqs)
+{
+	struct its_vm *vm = domain->host_data;
+	int i;
+
+	irq_domain_free_irqs_parent(domain, virq, nr_irqs);
+
+	for (i = 0; i < nr_irqs; i++) {
+		struct irq_data *data = irq_domain_get_irq_data(domain,
+								virq + i);
+		struct its_vpe *vpe = irq_data_get_irq_chip_data(data);
+
+		BUG_ON(vm != vpe->its_vm);
+
+		clear_bit(data->hwirq, vm->db_bitmap);
+		its_vpe_teardown(vpe);
+		irq_domain_reset_irq_data(data);
+	}
+
+	if (bitmap_empty(vm->db_bitmap, vm->nr_db_lpis)) {
+		its_lpi_free_chunks(vm->db_bitmap, vm->db_lpi_base, vm->nr_db_lpis);
+		its_free_prop_table(vm->vprop_page);
+	}
+}
+
+static int its_vpe_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
+				    unsigned int nr_irqs, void *args)
+{
+	struct its_vm *vm = args;
+	unsigned long *bitmap;
+	struct page *vprop_page;
+	int base, nr_ids, i, err = 0;
+
+	BUG_ON(!vm);
+
+	bitmap = its_lpi_alloc_chunks(nr_irqs, &base, &nr_ids);
+	if (!bitmap)
+		return -ENOMEM;
+
+	if (nr_ids < nr_irqs) {
+		its_lpi_free_chunks(bitmap, base, nr_ids);
+		return -ENOMEM;
+	}
+
+	vprop_page = its_allocate_prop_table(GFP_KERNEL);
+	if (!vprop_page) {
+		its_lpi_free_chunks(bitmap, base, nr_ids);
+		return -ENOMEM;
+	}
+
+	vm->db_bitmap = bitmap;
+	vm->db_lpi_base = base;
+	vm->nr_db_lpis = nr_ids;
+	vm->vprop_page = vprop_page;
+
+	for (i = 0; i < nr_irqs; i++) {
+		vm->vpes[i]->vpe_db_lpi = base + i;
+		err = its_vpe_init(vm->vpes[i]);
+		if (err)
+			break;
+		err = its_irq_gic_domain_alloc(domain, virq + i,
+					       vm->vpes[i]->vpe_db_lpi);
+		if (err)
+			break;
+		irq_domain_set_hwirq_and_chip(domain, virq + i, i,
+					      &its_vpe_irq_chip, vm->vpes[i]);
+		set_bit(i, bitmap);
+	}
+
+	if (err) {
+		if (i > 0)
+			its_vpe_irq_domain_free(domain, virq, i - 1);
+
+		its_lpi_free_chunks(bitmap, base, nr_ids);
+		its_free_prop_table(vprop_page);
+	}
+
+	return err;
+}
+
 static const struct irq_domain_ops its_vpe_domain_ops = {
+	.alloc			= its_vpe_irq_domain_alloc,
+	.free			= its_vpe_irq_domain_free,
 };
 
 static int its_force_quiescent(void __iomem *base)
-- 
2.11.0

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

#1700291 — [PATCH v3 18/59] irqchip/gic-v3-its: Add VLPI configuration handling

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 | 75 +++++++++++++++++++++++++++++++++++++---
 1 file changed, 70 insertions(+), 5 deletions(-)

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index ef31c3e58c01..b618c259be8b 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -806,18 +806,26 @@ 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)) {
+		struct its_device *its_dev = irq_data_get_irq_chip_data(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.
@@ -828,16 +836,52 @@ 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);
 }
 
@@ -850,6 +894,10 @@ 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) {
@@ -1018,6 +1066,22 @@ 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);
@@ -1040,6 +1104,7 @@ 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]

#1700293 — [PATCH v3 23/59] irqchip/gic-v3-its: Add VPE scheduling

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   | 85 ++++++++++++++++++++++++++++++++++++++
 include/linux/irqchip/arm-gic-v3.h | 58 ++++++++++++++++++++++++++
 2 files changed, 143 insertions(+)

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index 5ca1e571c1b4..d94a51489f6c 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -152,6 +152,7 @@ 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)
@@ -2147,8 +2148,92 @@ 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();
+	u32 count = 1000000;	/* 1s! */
+	bool clean;
+	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);
+
+	do {
+		val = gits_read_vpendbaser(vlpi_base + GICR_VPENDBASER);
+		clean = !(val & GICR_VPENDBASER_Dirty);
+		if (!clean) {
+			count--;
+			cpu_relax();
+			udelay(1);
+		}
+	} while (!clean && count);
+
+	if (unlikely(!clean && !count)) {
+		pr_err_ratelimited("ITS virtual pending table not cleaning\n");
+		vpe->idai = false;
+		vpe->pending_last = true;
+	} else {
+		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]

#1700296 — [PATCH v3 02/59] irqchip/gic-v3: Add redistributor iterator

In order to discover the VLPI properties, we need to iterate over
the redistributor regions. As we already have code that does this,
let's factor it out and make it slightly more generic.

Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 drivers/irqchip/irq-gic-v3.c | 69 ++++++++++++++++++++++++++++++--------------
 1 file changed, 47 insertions(+), 22 deletions(-)

diff --git a/drivers/irqchip/irq-gic-v3.c b/drivers/irqchip/irq-gic-v3.c
index dbffb7ab6203..640f621c7f75 100644
--- a/drivers/irqchip/irq-gic-v3.c
+++ b/drivers/irqchip/irq-gic-v3.c
@@ -421,24 +421,14 @@ static void __init gic_dist_init(void)
 		gic_write_irouter(affinity, base + GICD_IROUTER + i * 8);
 }
 
-static int gic_populate_rdist(void)
+static int gic_iterate_rdists(int (*fn)(struct redist_region *, void __iomem *))
 {
-	unsigned long mpidr = cpu_logical_map(smp_processor_id());
-	u64 typer;
-	u32 aff;
+	int ret = -ENODEV;
 	int i;
 
-	/*
-	 * Convert affinity to a 32bit value that can be matched to
-	 * GICR_TYPER bits [63:32].
-	 */
-	aff = (MPIDR_AFFINITY_LEVEL(mpidr, 3) << 24 |
-	       MPIDR_AFFINITY_LEVEL(mpidr, 2) << 16 |
-	       MPIDR_AFFINITY_LEVEL(mpidr, 1) << 8 |
-	       MPIDR_AFFINITY_LEVEL(mpidr, 0));
-
 	for (i = 0; i < gic_data.nr_redist_regions; i++) {
 		void __iomem *ptr = gic_data.redist_regions[i].redist_base;
+		u64 typer;
 		u32 reg;
 
 		reg = readl_relaxed(ptr + GICR_PIDR2) & GIC_PIDR2_ARCH_MASK;
@@ -450,15 +440,9 @@ static int gic_populate_rdist(void)
 
 		do {
 			typer = gic_read_typer(ptr + GICR_TYPER);
-			if ((typer >> 32) == aff) {
-				u64 offset = ptr - gic_data.redist_regions[i].redist_base;
-				gic_data_rdist_rd_base() = ptr;
-				gic_data_rdist()->phys_base = gic_data.redist_regions[i].phys_base + offset;
-				pr_info("CPU%d: found redistributor %lx region %d:%pa\n",
-					smp_processor_id(), mpidr, i,
-					&gic_data_rdist()->phys_base);
+			ret = fn(gic_data.redist_regions + i, ptr);
+			if (!ret)
 				return 0;
-			}
 
 			if (gic_data.redist_regions[i].single_redist)
 				break;
@@ -473,9 +457,50 @@ static int gic_populate_rdist(void)
 		} while (!(typer & GICR_TYPER_LAST));
 	}
 
+	return ret ? -ENODEV : 0;
+}
+
+static int __gic_populate_rdist(struct redist_region *region, void __iomem *ptr)
+{
+	unsigned long mpidr = cpu_logical_map(smp_processor_id());
+	u64 typer;
+	u32 aff;
+
+	/*
+	 * Convert affinity to a 32bit value that can be matched to
+	 * GICR_TYPER bits [63:32].
+	 */
+	aff = (MPIDR_AFFINITY_LEVEL(mpidr, 3) << 24 |
+	       MPIDR_AFFINITY_LEVEL(mpidr, 2) << 16 |
+	       MPIDR_AFFINITY_LEVEL(mpidr, 1) << 8 |
+	       MPIDR_AFFINITY_LEVEL(mpidr, 0));
+
+	typer = gic_read_typer(ptr + GICR_TYPER);
+	if ((typer >> 32) == aff) {
+		u64 offset = ptr - region->redist_base;
+		gic_data_rdist_rd_base() = ptr;
+		gic_data_rdist()->phys_base = region->phys_base + offset;
+
+		pr_info("CPU%d: found redistributor %lx region %d:%pa\n",
+			smp_processor_id(), mpidr,
+			(int)(region - gic_data.redist_regions),
+			&gic_data_rdist()->phys_base);
+		return 0;
+	}
+
+	/* Try next one */
+	return 1;
+}
+
+static int gic_populate_rdist(void)
+{
+	if (gic_iterate_rdists(__gic_populate_rdist) == 0)
+		return 0;
+
 	/* We couldn't even deal with ourselves... */
 	WARN(true, "CPU%d: mpidr %lx has no re-distributor!\n",
-	     smp_processor_id(), mpidr);
+	     smp_processor_id(),
+	     (unsigned long)cpu_logical_map(smp_processor_id()));
 	return -ENODEV;
 }
 
-- 
2.11.0

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

#1700298 — [PATCH v3 01/59] genirq: Let irq_set_vcpu_affinity() iterate over hierarchy

When assigning an interrupt to a vcpu, it is not unlikely that
the level of the hierarchy implementing irq_set_vcpu_affinity
is not the top level (think a generic MSI domain on top of a
virtualization aware interrupt controller).

In such a case, let's iterate over the hierarchy until we find
an irqchip implementing it.

Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
---
 kernel/irq/manage.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c
index 1d1a5b945ab4..573dc52b0806 100644
--- a/kernel/irq/manage.c
+++ b/kernel/irq/manage.c
@@ -400,8 +400,18 @@ int irq_set_vcpu_affinity(unsigned int irq, void *vcpu_info)
 		return -EINVAL;
 
 	data = irq_desc_get_irq_data(desc);
-	chip = irq_data_get_irq_chip(data);
-	if (chip && chip->irq_set_vcpu_affinity)
+	do {
+		chip = irq_data_get_irq_chip(data);
+		if (chip && chip->irq_set_vcpu_affinity)
+			break;
+#ifdef CONFIG_IRQ_DOMAIN_HIERARCHY
+		data = data->parent_data;
+#else
+		data = NULL;
+#endif
+	} while (data);
+
+	if (data)
 		ret = chip->irq_set_vcpu_affinity(data, vcpu_info);
 	irq_put_desc_unlock(desc, flags);
 
-- 
2.11.0

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

#1700299 — [PATCH v3 07/59] irqchip/gic-v3-its: Implement irq_set_irqchip_state for pending state

Allow the pending state of an LPI to be set or cleared via
irq_set_irqchip_state.

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 | 78 ++++++++++++++++++++++++++++++++++++++++
 1 file changed, 78 insertions(+)

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index 0f345c0fa700..3e4ac8c13e21 100644
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -167,6 +167,11 @@ struct its_cmd_desc {
 		struct {
 			struct its_device *dev;
 			u32 event_id;
+		} its_clear_cmd;
+
+		struct {
+			struct its_device *dev;
+			u32 event_id;
 		} its_int_cmd;
 
 		struct {
@@ -380,6 +385,40 @@ static struct its_collection *its_build_inv_cmd(struct its_cmd_block *cmd,
 	return col;
 }
 
+static struct its_collection *its_build_int_cmd(struct its_cmd_block *cmd,
+						struct its_cmd_desc *desc)
+{
+	struct its_collection *col;
+
+	col = dev_event_to_col(desc->its_int_cmd.dev,
+			       desc->its_int_cmd.event_id);
+
+	its_encode_cmd(cmd, GITS_CMD_INT);
+	its_encode_devid(cmd, desc->its_int_cmd.dev->device_id);
+	its_encode_event_id(cmd, desc->its_int_cmd.event_id);
+
+	its_fixup_cmd(cmd);
+
+	return col;
+}
+
+static struct its_collection *its_build_clear_cmd(struct its_cmd_block *cmd,
+						  struct its_cmd_desc *desc)
+{
+	struct its_collection *col;
+
+	col = dev_event_to_col(desc->its_clear_cmd.dev,
+			       desc->its_clear_cmd.event_id);
+
+	its_encode_cmd(cmd, GITS_CMD_CLEAR);
+	its_encode_devid(cmd, desc->its_clear_cmd.dev->device_id);
+	its_encode_event_id(cmd, desc->its_clear_cmd.event_id);
+
+	its_fixup_cmd(cmd);
+
+	return col;
+}
+
 static struct its_collection *its_build_invall_cmd(struct its_cmd_block *cmd,
 						   struct its_cmd_desc *desc)
 {
@@ -536,6 +575,26 @@ 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_send_int(struct its_device *dev, u32 event_id)
+{
+	struct its_cmd_desc desc;
+
+	desc.its_int_cmd.dev = dev;
+	desc.its_int_cmd.event_id = event_id;
+
+	its_send_single_command(dev->its, its_build_int_cmd, &desc);
+}
+
+static void its_send_clear(struct its_device *dev, u32 event_id)
+{
+	struct its_cmd_desc desc;
+
+	desc.its_clear_cmd.dev = dev;
+	desc.its_clear_cmd.event_id = event_id;
+
+	its_send_single_command(dev->its, its_build_clear_cmd, &desc);
+}
+
 static void its_send_inv(struct its_device *dev, u32 event_id)
 {
 	struct its_cmd_desc desc;
@@ -702,6 +761,24 @@ static void its_irq_compose_msi_msg(struct irq_data *d, struct msi_msg *msg)
 	iommu_dma_map_msi_msg(d->irq, msg);
 }
 
+static int its_irq_set_irqchip_state(struct irq_data *d,
+				     enum irqchip_irq_state which,
+				     bool state)
+{
+	struct its_device *its_dev = irq_data_get_irq_chip_data(d);
+	u32 event = its_get_event_id(d);
+
+	if (which != IRQCHIP_STATE_PENDING)
+		return -EINVAL;
+
+	if (state)
+		its_send_int(its_dev, event);
+	else
+		its_send_clear(its_dev, event);
+
+	return 0;
+}
+
 static struct irq_chip its_irq_chip = {
 	.name			= "ITS",
 	.irq_mask		= its_mask_irq,
@@ -709,6 +786,7 @@ static struct irq_chip its_irq_chip = {
 	.irq_eoi		= irq_chip_eoi_parent,
 	.irq_set_affinity	= its_set_affinity,
 	.irq_compose_msi_msg	= its_irq_compose_msi_msg,
+	.irq_set_irqchip_state	= its_irq_set_irqchip_state,
 };
 
 /*
-- 
2.11.0

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

#1700301 — [PATCH v3 03/59] irqchip/gic-v3: Add VLPI/DirectLPI discovery

Add helper functions that probe for VLPI and DirectLPI properties.

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.c       | 24 +++++++++++++++++++++++-
 include/linux/irqchip/arm-gic-v3.h |  3 +++
 2 files changed, 26 insertions(+), 1 deletion(-)

diff --git a/drivers/irqchip/irq-gic-v3.c b/drivers/irqchip/irq-gic-v3.c
index 640f621c7f75..17f0ad41864f 100644
--- a/drivers/irqchip/irq-gic-v3.c
+++ b/drivers/irqchip/irq-gic-v3.c
@@ -1,5 +1,5 @@
 /*
- * Copyright (C) 2013, 2014 ARM Limited, All Rights Reserved.
+ * Copyright (C) 2013-2017 ARM Limited, All Rights Reserved.
  * Author: Marc Zyngier <marc.zyngier@arm.com>
  *
  * This program is free software; you can redistribute it and/or modify
@@ -504,6 +504,24 @@ static int gic_populate_rdist(void)
 	return -ENODEV;
 }
 
+static int __gic_update_vlpi_properties(struct redist_region *region,
+					void __iomem *ptr)
+{
+	u64 typer = gic_read_typer(ptr + GICR_TYPER);
+	gic_data.rdists.has_vlpis &= !!(typer & GICR_TYPER_VLPIS);
+	gic_data.rdists.has_direct_lpi &= !!(typer & GICR_TYPER_DirectLPIS);
+
+	return 1;
+}
+
+static void gic_update_vlpi_properties(void)
+{
+	gic_iterate_rdists(__gic_update_vlpi_properties);
+	pr_info("%sVLPI support, %sdirect LPI support\n",
+		!gic_data.rdists.has_vlpis ? "no " : "",
+		!gic_data.rdists.has_direct_lpi ? "no " : "");
+}
+
 static void gic_cpu_sys_reg_init(void)
 {
 	/*
@@ -968,6 +986,8 @@ static int __init gic_init_bases(void __iomem *dist_base,
 	gic_data.domain = irq_domain_create_tree(handle, &gic_irq_domain_ops,
 						 &gic_data);
 	gic_data.rdists.rdist = alloc_percpu(typeof(*gic_data.rdists.rdist));
+	gic_data.rdists.has_vlpis = true;
+	gic_data.rdists.has_direct_lpi = true;
 
 	if (WARN_ON(!gic_data.domain) || WARN_ON(!gic_data.rdists.rdist)) {
 		err = -ENOMEM;
@@ -976,6 +996,8 @@ static int __init gic_init_bases(void __iomem *dist_base,
 
 	set_handle_irq(gic_handle_irq);
 
+	gic_update_vlpi_properties();
+
 	if (IS_ENABLED(CONFIG_ARM_GIC_V3_ITS) && gic_dist_supports_lpis())
 		its_init(handle, &gic_data.rdists, gic_data.domain);
 
diff --git a/include/linux/irqchip/arm-gic-v3.h b/include/linux/irqchip/arm-gic-v3.h
index 6a1f87ff94e2..20a553423ac7 100644
--- a/include/linux/irqchip/arm-gic-v3.h
+++ b/include/linux/irqchip/arm-gic-v3.h
@@ -204,6 +204,7 @@
 
 #define GICR_TYPER_PLPIS		(1U << 0)
 #define GICR_TYPER_VLPIS		(1U << 1)
+#define GICR_TYPER_DirectLPIS		(1U << 3)
 #define GICR_TYPER_LAST			(1U << 4)
 
 #define GIC_V3_REDIST_SIZE		0x20000
@@ -487,6 +488,8 @@ struct rdists {
 	struct page		*prop_page;
 	int			id_bits;
 	u64			flags;
+	bool			has_vlpis;
+	bool			has_direct_lpi;
 };
 
 struct irq_domain;
-- 
2.11.0

[toc] | [prev] | [standalone]

Page 3 of 3 — ← Prev page 1 2 [3]

Back to top | Article view | linux.kernel

csiph-web