From: Kay, Allen M <[EMAIL PROTECTED]> This patch includes the functions to support VT-d for passthrough devices.
[Ben: fixed memory pinning, cleanup] Signed-off-by: Kay, Allen M <[EMAIL PROTECTED]> Signed-off-by: Weidong Han <[EMAIL PROTECTED]> Signed-off-by: Ben-Ami Yassour <[EMAIL PROTECTED]> --- arch/x86/kvm/Makefile | 2 +- arch/x86/kvm/vtd.c | 182 ++++++++++++++++++++++++++++++++++++++++++++ arch/x86/kvm/x86.c | 11 +++ include/asm-x86/kvm_host.h | 3 + include/linux/kvm_host.h | 6 ++ virt/kvm/kvm_main.c | 8 ++- 6 files changed, 210 insertions(+), 2 deletions(-) create mode 100644 arch/x86/kvm/vtd.c diff --git a/arch/x86/kvm/Makefile b/arch/x86/kvm/Makefile index d0e940b..5d9d079 100644 --- a/arch/x86/kvm/Makefile +++ b/arch/x86/kvm/Makefile @@ -11,7 +11,7 @@ endif EXTRA_CFLAGS += -Ivirt/kvm -Iarch/x86/kvm kvm-objs := $(common-objs) x86.o mmu.o x86_emulate.o i8259.o irq.o lapic.o \ - i8254.o + i8254.o vtd.o obj-$(CONFIG_KVM) += kvm.o kvm-intel-objs = vmx.o obj-$(CONFIG_KVM_INTEL) += kvm-intel.o diff --git a/arch/x86/kvm/vtd.c b/arch/x86/kvm/vtd.c new file mode 100644 index 0000000..7a3cf4e --- /dev/null +++ b/arch/x86/kvm/vtd.c @@ -0,0 +1,182 @@ +/* + * Copyright (c) 2006, Intel Corporation. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope 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, write to the Free Software Foundation, Inc., 59 Temple + * Place - Suite 330, Boston, MA 02111-1307 USA. + * + * Copyright (C) 2006-2008 Intel Corporation + * Author: Allen M. Kay <[EMAIL PROTECTED]> + * Author: Weidong Han <[EMAIL PROTECTED]> + */ + +#include <linux/list.h> +#include <linux/kvm_host.h> +#include <linux/pci.h> +#include <linux/dmar.h> +#include <linux/intel-iommu.h> + +static int kvm_iommu_unmap_memslots(struct kvm *kvm); + +int kvm_iommu_map_pages(struct kvm *kvm, + gfn_t base_gfn, unsigned long npages) +{ + gfn_t gfn = base_gfn; + pfn_t pfn; + int i, rc; + struct dmar_domain *domain = kvm->arch.intel_iommu_domain; + + if (!domain) + return -EFAULT; + + for (i = 0; i < npages; i++) { + pfn = gfn_to_pfn(kvm, gfn); + if (!is_mmio_pfn(pfn)) { + rc = intel_iommu_page_mapping(domain, + gfn << PAGE_SHIFT, + pfn << PAGE_SHIFT, + PAGE_SIZE, + DMA_PTE_READ | + DMA_PTE_WRITE); + if (rc) + kvm_release_pfn_clean(pfn); + } else { + printk(KERN_DEBUG "kvm_iommu_map_page:" + "invalid pfn=%lx\n", pfn); + return 0; + } + + gfn++; + } + return 0; +} + +static int kvm_iommu_map_memslots(struct kvm *kvm) +{ + int i, rc; + for (i = 0; i < kvm->nmemslots; i++) { + rc = kvm_iommu_map_pages(kvm, kvm->memslots[i].base_gfn, + kvm->memslots[i].npages); + if (rc) + return rc; + } + return 0; +} + +int kvm_iommu_map_guest(struct kvm *kvm, + struct kvm_assigned_dev *assigned_dev) +{ + struct pci_dev *pdev = NULL; + + printk(KERN_DEBUG "VT-d direct map: host bdf = %x:%x:%x\n", + assigned_dev->host.busnr, + PCI_SLOT(assigned_dev->host.devfn), + PCI_FUNC(assigned_dev->host.devfn)); + + for_each_pci_dev(pdev) { + if ((pdev->bus->number == assigned_dev->host.busnr) && + (pdev->devfn == assigned_dev->host.devfn)) { + break; + } + } + + if (pdev == NULL) { + if (kvm->arch.intel_iommu_domain) { + intel_iommu_domain_exit(kvm->arch.intel_iommu_domain); + kvm->arch.intel_iommu_domain = NULL; + } + return -ENODEV; + } + + kvm->arch.intel_iommu_domain = intel_iommu_domain_alloc(pdev); + + if (kvm_iommu_map_memslots(kvm)) { + kvm_iommu_unmap_memslots(kvm); + return -EFAULT; + } + + intel_iommu_detach_dev(kvm->arch.intel_iommu_domain, + pdev->bus->number, pdev->devfn); + + if (intel_iommu_context_mapping(kvm->arch.intel_iommu_domain, + pdev)) { + printk(KERN_ERR "Domain context map for %s failed", + pci_name(pdev)); + return -EFAULT; + } + return 0; +} + +static int kvm_iommu_put_pages(struct kvm *kvm, + gfn_t base_gfn, unsigned long npages) +{ + gfn_t gfn = base_gfn; + pfn_t pfn; + struct dmar_domain *domain = kvm->arch.intel_iommu_domain; + int i; + + if (!domain) + return -EFAULT; + + for (i = 0; i < npages; i++) { + pfn = (pfn_t)intel_iommu_iova_to_pfn(domain, + gfn << PAGE_SHIFT); + kvm_release_pfn_clean(pfn); + gfn++; + } + return 0; +} + +static int kvm_iommu_unmap_memslots(struct kvm *kvm) +{ + int i, rc; + for (i = 0; i < kvm->nmemslots; i++) { + rc = kvm_iommu_put_pages(kvm, kvm->memslots[i].base_gfn, + kvm->memslots[i].npages); + if (rc) + return rc; + } + return 0; +} + +int kvm_iommu_unmap_guest(struct kvm *kvm) +{ + struct kvm_assigned_dev_kernel *entry; + struct pci_dev *pdev = NULL; + struct dmar_domain *domain = kvm->arch.intel_iommu_domain; + + if (!domain) + return 0; + + list_for_each_entry(entry, &kvm->arch.assigned_dev_head, list) { + printk(KERN_DEBUG "VT-d unmap: host bdf = %x:%x:%x\n", + entry->host.busnr, + PCI_SLOT(entry->host.devfn), + PCI_FUNC(entry->host.devfn)); + + for_each_pci_dev(pdev) { + if ((pdev->bus->number == entry->host.busnr) && + (pdev->devfn == entry->host.devfn)) + break; + } + + if (pdev == NULL) + return -ENODEV; + + /* detach kvm dmar domain */ + intel_iommu_detach_dev(domain, + pdev->bus->number, pdev->devfn); + } + kvm_iommu_unmap_memslots(kvm); + intel_iommu_domain_exit(domain); + return 0; +} diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index d9aa931..e57c9e4 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -33,6 +33,7 @@ #include <linux/module.h> #include <linux/mman.h> #include <linux/highmem.h> +#include <linux/intel-iommu.h> #include <asm/uaccess.h> #include <asm/msr.h> @@ -253,9 +254,17 @@ kvm_vm_ioctl_device_assignment(struct kvm *kvm, } + if (intel_iommu_found()) { + r = kvm_iommu_map_guest(kvm, assigned_dev); + if (r) + goto out_intr; + } + out: mutex_unlock(&kvm->lock); return r; +out_intr: + free_irq(dev->irq, match); out_list_del: list_del(&match->list); pci_release_regions(dev); @@ -4171,6 +4180,8 @@ static void kvm_free_vcpus(struct kvm *kvm) void kvm_arch_destroy_vm(struct kvm *kvm) { + if (intel_iommu_found()) + kvm_iommu_unmap_guest(kvm); kvm_free_assigned_devices(kvm); kvm_free_pit(kvm); kfree(kvm->arch.vpic); diff --git a/include/asm-x86/kvm_host.h b/include/asm-x86/kvm_host.h index 34eb3e7..75b088b 100644 --- a/include/asm-x86/kvm_host.h +++ b/include/asm-x86/kvm_host.h @@ -357,6 +357,7 @@ struct kvm_arch{ */ struct list_head active_mmu_pages; struct list_head assigned_dev_head; + struct dmar_domain *intel_iommu_domain; struct kvm_pic *vpic; struct kvm_ioapic *vioapic; struct kvm_pit *vpit; @@ -506,6 +507,8 @@ int emulator_write_phys(struct kvm_vcpu *vcpu, gpa_t gpa, int kvm_pv_mmu_op(struct kvm_vcpu *vcpu, unsigned long bytes, gpa_t addr, unsigned long *ret); +int is_mmio_pfn(pfn_t pfn); + extern bool tdp_enabled; enum emulation_result { diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h index 3798097..506b595 100644 --- a/include/linux/kvm_host.h +++ b/include/linux/kvm_host.h @@ -279,6 +279,12 @@ int kvm_cpu_has_interrupt(struct kvm_vcpu *v); int kvm_cpu_has_pending_timer(struct kvm_vcpu *vcpu); void kvm_vcpu_kick(struct kvm_vcpu *vcpu); +int kvm_iommu_map_pages(struct kvm *kvm, gfn_t base_gfn, + unsigned long npages); +int kvm_iommu_map_guest(struct kvm *kvm, + struct kvm_assigned_dev *assigned_dev); +int kvm_iommu_unmap_guest(struct kvm *kvm); + static inline void kvm_guest_enter(void) { account_system_vtime(current); diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index 5b470a1..49c4ed3 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c @@ -41,6 +41,7 @@ #include <linux/pagemap.h> #include <linux/mman.h> #include <linux/swap.h> +#include <linux/intel-iommu.h> #include <asm/processor.h> #include <asm/io.h> @@ -76,7 +77,7 @@ static inline int valid_vcpu(int n) return likely(n >= 0 && n < KVM_MAX_VCPUS); } -static inline int is_mmio_pfn(pfn_t pfn) +inline int is_mmio_pfn(pfn_t pfn) { if (pfn_valid(pfn)) return PageReserved(pfn_to_page(pfn)); @@ -431,6 +432,11 @@ int __kvm_set_memory_region(struct kvm *kvm, } kvm_free_physmem_slot(&old, &new); + + /* map the pages in iommu page table */ + if (intel_iommu_found()) + kvm_iommu_map_pages(kvm, base_gfn, npages); + return 0; out_free: -- 1.5.6 -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html