Re: [PATCH v3 48/55] KVM: arm/arm64: vgic-new: vgic_init: implement kvm_vgic_hyp_init

[Christoffer Dall <christoffer.dall@linaro.org>]

On Fri, May 06, 2016 at 11:46:01AM +0100, Andre Przywara wrote:
> From: Eric Auger <eric.auger@linaro.org>
> 
> Implements kvm_vgic_hyp_init and vgic_probe function.
> 
> The vgic_global struct is enriched with new fields populated
> by those functions.
> 
> Signed-off-by: Eric Auger <eric.auger@linaro.org>
> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
> ---
> Changelog v1 .. v2:
> - rename vgic_init.c to vgic-init.c
> 
> Changelog v2 .. v3:
> - include kvm/arm_vgic.h instead of kvm/vgic/vgic.h
> - move ich_vtr_el2 variable into probe function
> 
>  include/kvm/vgic/vgic.h       |   1 +
>  virt/kvm/arm/vgic/vgic-init.c | 122 ++++++++++++++++++++++++++++++++++++++++++
>  virt/kvm/arm/vgic/vgic-v2.c   |  90 +++++++++++++++++++++++++++++++
>  virt/kvm/arm/vgic/vgic-v3.c   |  73 +++++++++++++++++++++++++
>  virt/kvm/arm/vgic/vgic.h      |   7 +++
>  5 files changed, 293 insertions(+)
>  create mode 100644 virt/kvm/arm/vgic/vgic-init.c
> 
> diff --git a/include/kvm/vgic/vgic.h b/include/kvm/vgic/vgic.h
> index cfc3640..d144e3d 100644
> --- a/include/kvm/vgic/vgic.h
> +++ b/include/kvm/vgic/vgic.h
> @@ -202,6 +202,7 @@ struct vgic_cpu {
>  };
>  
>  int kvm_vgic_addr(struct kvm *kvm, unsigned long type, u64 *addr, bool write);
> +int kvm_vgic_hyp_init(void);
>  
>  int kvm_vgic_inject_irq(struct kvm *kvm, int cpuid, unsigned int intid,
>  			bool level);
> diff --git a/virt/kvm/arm/vgic/vgic-init.c b/virt/kvm/arm/vgic/vgic-init.c
> new file mode 100644
> index 0000000..f10997b
> --- /dev/null
> +++ b/virt/kvm/arm/vgic/vgic-init.c
> @@ -0,0 +1,122 @@
> +/*
> + * Copyright (C) 2015, 2016 ARM Ltd.
> + *
> + * 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/uaccess.h>
> +#include <linux/interrupt.h>
> +#include <linux/cpu.h>
> +#include <linux/of_address.h>
> +#include <linux/of_irq.h>
> +#include <linux/kvm_host.h>
> +#include <kvm/arm_vgic.h>
> +#include <asm/kvm_mmu.h>
> +#include "vgic.h"
> +
> +/* GENERIC PROBE */
> +
> +static void vgic_init_maintenance_interrupt(void *info)
> +{
> +	enable_percpu_irq(kvm_vgic_global_state.maint_irq, 0);
> +}
> +
> +static int vgic_cpu_notify(struct notifier_block *self,
> +			   unsigned long action, void *cpu)
> +{
> +	switch (action) {
> +	case CPU_STARTING:
> +	case CPU_STARTING_FROZEN:
> +		vgic_init_maintenance_interrupt(NULL);
> +		break;
> +	case CPU_DYING:
> +	case CPU_DYING_FROZEN:
> +		disable_percpu_irq(kvm_vgic_global_state.maint_irq);
> +		break;
> +	}
> +
> +	return NOTIFY_OK;
> +}
> +
> +static struct notifier_block vgic_cpu_nb = {
> +	.notifier_call = vgic_cpu_notify,
> +};
> +
> +static irqreturn_t vgic_maintenance_handler(int irq, void *data)
> +{
> +	/*
> +	 * We cannot rely on the vgic maintenance interrupt to be
> +	 * delivered synchronously. This means we can only use it to
> +	 * exit the VM, and we perform the handling of EOIed
> +	 * interrupts on the exit path (see vgic_process_maintenance).
> +	 */
> +	return IRQ_HANDLED;
> +}
> +
> +static const struct of_device_id vgic_ids[] = {
> +	{ .compatible = "arm,cortex-a15-gic",	.data = vgic_v2_probe, },
> +	{ .compatible = "arm,cortex-a7-gic",	.data = vgic_v2_probe, },
> +	{ .compatible = "arm,gic-400",		.data = vgic_v2_probe, },
> +	{ .compatible = "arm,gic-v3",		.data = vgic_v3_probe, },
> +	{},
> +};
> +
> +/**
> + * kvm_vgic_hyp_init: populates the kvm_vgic_global_state variable
> + * according to the host GIC model. Accordingly calls either
> + * vgic_v2/v3_probe which registers the KVM_DEVICE that can be
> + * instantiated by a guest later on .
> + */
> +int kvm_vgic_hyp_init(void)
> +{
> +	const struct of_device_id *matched_id;
> +	const int (*vgic_probe)(struct device_node *);
> +	struct device_node *vgic_node;
> +	int ret;
> +
> +	vgic_node = of_find_matching_node_and_match(NULL,
> +						    vgic_ids, &matched_id);
> +	if (!vgic_node) {
> +		kvm_err("error: no compatible GIC node found\n");
> +		return -ENODEV;
> +	}
> +
> +	vgic_probe = matched_id->data;
> +	ret = vgic_probe(vgic_node);
> +	if (ret)
> +		return ret;
> +
> +	ret = request_percpu_irq(kvm_vgic_global_state.maint_irq,
> +				 vgic_maintenance_handler,
> +				 "vgic", kvm_get_running_vcpus());
> +	if (ret) {
> +		kvm_err("Cannot register interrupt %d\n",
> +			kvm_vgic_global_state.maint_irq);
> +		return ret;
> +	}
> +
> +	ret = __register_cpu_notifier(&vgic_cpu_nb);
> +	if (ret) {
> +		kvm_err("Cannot register vgic CPU notifier\n");
> +		goto out_free_irq;
> +	}
> +
> +	on_each_cpu(vgic_init_maintenance_interrupt, NULL, 1);
> +
> +	return 0;
> +
> +out_free_irq:
> +	free_percpu_irq(kvm_vgic_global_state.maint_irq,
> +			kvm_get_running_vcpus());
> +	return ret;
> +}
> diff --git a/virt/kvm/arm/vgic/vgic-v2.c b/virt/kvm/arm/vgic/vgic-v2.c
> index 70cac63..91b69a4 100644
> --- a/virt/kvm/arm/vgic/vgic-v2.c
> +++ b/virt/kvm/arm/vgic/vgic-v2.c
> @@ -17,6 +17,11 @@
>  #include <linux/irqchip/arm-gic.h>
>  #include <linux/kvm.h>
>  #include <linux/kvm_host.h>
> +#include <kvm/arm_vgic.h>
> +#include <linux/of.h>
> +#include <linux/of_address.h>
> +#include <linux/of_irq.h>
> +#include <asm/kvm_mmu.h>
>  
>  #include "vgic.h"
>  
> @@ -205,3 +210,88 @@ void vgic_v2_get_vmcr(struct kvm_vcpu *vcpu, struct vgic_vmcr *vmcrp)
>  	vmcrp->pmr  = (vmcr & GICH_VMCR_PRIMASK_MASK) >>
>  			GICH_VMCR_PRIMASK_SHIFT;
>  }
> +
> +/**
> + * vgic_v2_probe - probe for a GICv2 compatible interrupt controller in DT
> + * @node:	pointer to the DT node
> + *
> + * Returns 0 if a GICv2 has been found, returns an error code otherwise
> + */
> +int vgic_v2_probe(struct device_node *vgic_node)
> +{
> +	int ret;
> +	struct resource vctrl_res;
> +	struct resource vcpu_res;
> +
> +	kvm_vgic_global_state.maint_irq = irq_of_parse_and_map(vgic_node, 0);
> +	if (!kvm_vgic_global_state.maint_irq) {
> +		kvm_err("error getting vgic maintenance irq from DT\n");
> +		ret = -ENXIO;
> +		goto out;
> +	}
> +
> +	ret = of_address_to_resource(vgic_node, 2, &vctrl_res);
> +	if (ret) {
> +		kvm_err("Cannot obtain GICH resource\n");
> +		goto out;
> +	}
> +
> +	kvm_vgic_global_state.vctrl_base = of_iomap(vgic_node, 2);
> +	if (!kvm_vgic_global_state.vctrl_base) {
> +		kvm_err("Cannot ioremap GICH\n");
> +		ret = -ENOMEM;
> +		goto out;
> +	}
> +
> +	kvm_vgic_global_state.nr_lr =
> +		readl_relaxed(kvm_vgic_global_state.vctrl_base + GICH_VTR);
> +	kvm_vgic_global_state.nr_lr = (kvm_vgic_global_state.nr_lr & 0x3f) + 1;

could we please do:
	vtr = readl_relaxed(kvm_vgic_global_state.vctrl_base + GICH_VTR);
	kvm_vgic_global_state.nr_lr = (vtr & 0x3f) + 1;

instead?

> +
> +	ret = create_hyp_io_mappings(kvm_vgic_global_state.vctrl_base,
> +				     kvm_vgic_global_state.vctrl_base +
> +					 resource_size(&vctrl_res),
> +				     vctrl_res.start);
> +	if (ret) {
> +		kvm_err("Cannot map VCTRL into hyp\n");
> +		goto out_unmap;
> +	}
> +
> +	if (of_address_to_resource(vgic_node, 3, &vcpu_res)) {
> +		kvm_err("Cannot obtain GICV resource\n");
> +		ret = -ENXIO;
> +		goto out_unmap;
> +	}
> +
> +	if (!PAGE_ALIGNED(vcpu_res.start)) {
> +		kvm_err("GICV physical address 0x%llx not page aligned\n",
> +			(unsigned long long)vcpu_res.start);
> +		ret = -ENXIO;
> +		goto out_unmap;
> +	}
> +
> +	if (!PAGE_ALIGNED(resource_size(&vcpu_res))) {
> +		kvm_err("GICV size 0x%llx not a multiple of page size 0x%lx\n",
> +			(unsigned long long)resource_size(&vcpu_res),
> +			PAGE_SIZE);
> +		ret = -ENXIO;
> +		goto out_unmap;
> +	}
> +
> +	kvm_vgic_global_state.can_emulate_gicv2 = true;
> +	kvm_register_vgic_device(KVM_DEV_TYPE_ARM_VGIC_V2);
> +
> +	kvm_vgic_global_state.vcpu_base = vcpu_res.start;
> +
> +	kvm_info("%s@%llx IRQ%d\n", vgic_node->name,
> +		 vctrl_res.start, kvm_vgic_global_state.maint_irq);
> +
> +	kvm_vgic_global_state.type = VGIC_V2;
> +	kvm_vgic_global_state.max_gic_vcpus = VGIC_V2_MAX_CPUS;
> +	goto out;
> +
> +out_unmap:
> +	iounmap(kvm_vgic_global_state.vctrl_base);
> +out:
> +	of_node_put(vgic_node);
> +	return ret;
> +}
> diff --git a/virt/kvm/arm/vgic/vgic-v3.c b/virt/kvm/arm/vgic/vgic-v3.c
> index dca52b3..48b0bb7 100644
> --- a/virt/kvm/arm/vgic/vgic-v3.c
> +++ b/virt/kvm/arm/vgic/vgic-v3.c
> @@ -15,6 +15,12 @@
>  #include <linux/irqchip/arm-gic-v3.h>
>  #include <linux/kvm.h>
>  #include <linux/kvm_host.h>
> +#include <kvm/arm_vgic.h>
> +#include <linux/of.h>
> +#include <linux/of_address.h>
> +#include <linux/of_irq.h>
> +#include <asm/kvm_mmu.h>
> +#include <asm/kvm_asm.h>
>  
>  #include "vgic.h"
>  
> @@ -188,3 +194,70 @@ void vgic_v3_get_vmcr(struct kvm_vcpu *vcpu, struct vgic_vmcr *vmcrp)
>  	vmcrp->bpr  = (vmcr & ICH_VMCR_BPR0_MASK) >> ICH_VMCR_BPR0_SHIFT;
>  	vmcrp->pmr  = (vmcr & ICH_VMCR_PMR_MASK) >> ICH_VMCR_PMR_SHIFT;
>  }
> +
> +/**
> + * vgic_v3_probe - probe for a GICv3 compatible interrupt controller in DT
> + * @node:	pointer to the DT node
> + *
> + * Returns 0 if a GICv3 has been found, returns an error code otherwise
> + */
> +int vgic_v3_probe(struct device_node *vgic_node)
> +{
> +	u32 ich_vtr_el2;
> +	u32 gicv_idx;
> +	int ret = 0;
> +	struct resource vcpu_res;
> +
> +	kvm_vgic_global_state.maint_irq = irq_of_parse_and_map(vgic_node, 0);
> +	if (!kvm_vgic_global_state.maint_irq) {
> +		kvm_err("error getting vgic maintenance irq from DT\n");
> +		ret = -ENXIO;
> +		goto out;
> +	}
> +
> +	ich_vtr_el2 = kvm_call_hyp(__vgic_v3_get_ich_vtr_el2);
> +
> +	/*
> +	 * The ListRegs field is 5 bits, but there is a architectural
> +	 * maximum of 16 list registers. Just ignore bit 4...
> +	 */
> +	kvm_vgic_global_state.nr_lr = (ich_vtr_el2 & 0xf) + 1;
> +	kvm_vgic_global_state.can_emulate_gicv2 = false;
> +
> +	if (of_property_read_u32(vgic_node, "#redistributor-regions",
> +				 &gicv_idx))
> +		gicv_idx = 1;
> +
> +	gicv_idx += 3; /* Also skip GICD, GICC, GICH */
> +	if (of_address_to_resource(vgic_node, gicv_idx, &vcpu_res)) {
> +		kvm_info("GICv3: no GICV resource entry\n");
> +		kvm_vgic_global_state.vcpu_base = 0;
> +	} else if (!PAGE_ALIGNED(vcpu_res.start)) {
> +		pr_warn("GICV physical address 0x%llx not page aligned\n",
> +			(unsigned long long)vcpu_res.start);
> +		kvm_vgic_global_state.vcpu_base = 0;
> +	} else if (!PAGE_ALIGNED(resource_size(&vcpu_res))) {
> +		pr_warn("GICV size 0x%llx not a multiple of page size 0x%lx\n",
> +			(unsigned long long)resource_size(&vcpu_res),
> +			PAGE_SIZE);
> +		kvm_vgic_global_state.vcpu_base = 0;
> +	} else {
> +		kvm_vgic_global_state.vcpu_base = vcpu_res.start;
> +		kvm_vgic_global_state.can_emulate_gicv2 = true;
> +		kvm_register_vgic_device(KVM_DEV_TYPE_ARM_VGIC_V2);
> +	}
> +	if (kvm_vgic_global_state.vcpu_base == 0)
> +		kvm_info("disabling GICv2 emulation\n");
> +	kvm_register_vgic_device(KVM_DEV_TYPE_ARM_VGIC_V3);
> +
> +	kvm_vgic_global_state.vctrl_base = NULL;
> +	kvm_vgic_global_state.type = VGIC_V3;
> +	kvm_vgic_global_state.max_gic_vcpus = VGIC_V3_MAX_CPUS;
> +
> +	kvm_info("%s@%llx IRQ%d\n", vgic_node->name,
> +		 vcpu_res.start, kvm_vgic_global_state.maint_irq);
> +
> +out:
> +	of_node_put(vgic_node);
> +	return ret;
> +}
> diff --git a/virt/kvm/arm/vgic/vgic.h b/virt/kvm/arm/vgic/vgic.h
> index 7a69955..e49b1df 100644
> --- a/virt/kvm/arm/vgic/vgic.h
> +++ b/virt/kvm/arm/vgic/vgic.h
> @@ -43,6 +43,7 @@ int vgic_v2_cpuif_uaccess(struct kvm_vcpu *vcpu, bool is_write,
>  			  int offset, u32 *val);
>  void vgic_v2_set_vmcr(struct kvm_vcpu *vcpu, struct vgic_vmcr *vmcr);
>  void vgic_v2_get_vmcr(struct kvm_vcpu *vcpu, struct vgic_vmcr *vmcr);
> +int vgic_v2_probe(struct device_node *vgic_node);
>  int vgic_register_dist_iodev(struct kvm *kvm, gpa_t dist_base_address,
>  			     enum vgic_type);
>  
> @@ -54,6 +55,7 @@ void vgic_v3_clear_lr(struct kvm_vcpu *vcpu, int lr);
>  void vgic_v3_set_underflow(struct kvm_vcpu *vcpu);
>  void vgic_v3_set_vmcr(struct kvm_vcpu *vcpu, struct vgic_vmcr *vmcr);
>  void vgic_v3_get_vmcr(struct kvm_vcpu *vcpu, struct vgic_vmcr *vmcr);
> +int vgic_v3_probe(struct device_node *vgic_node);
>  int vgic_register_redist_iodevs(struct kvm *kvm, gpa_t dist_base_address);
>  #else
>  static inline void vgic_v3_process_maintenance(struct kvm_vcpu *vcpu)
> @@ -87,6 +89,11 @@ void vgic_v3_get_vmcr(struct kvm_vcpu *vcpu, struct vgic_vmcr *vmcr)
>  {
>  }
>  
> +static inline int vgic_v3_probe(struct device_node *vgic_node)
> +{
> +	return -ENODEV;
> +}
> +
>  static inline int vgic_register_redist_iodevs(struct kvm *kvm,
>  					      gpa_t dist_base_address)
>  {
> -- 
> 2.7.3
> 
> --

I think we need to change this behavior to use the gic_kvm_info
structure populated by Julien's patches now and in kvmarm/next for the
next version of the series, and then I'll review that part properly.
Thoughts?

Thanks,
-Christoffer