[Qemu-devel] [PATCH 3/4] kvm: i386: Add services required for PCI device assignment

[Jan Kiszka]

From: Jan Kiszka <address@hidden>

These helpers abstract the interaction of upcoming pci-assign with the
KVM kernel services. Put them under i386 only as other archs will
implement device pass-through via VFIO and not this classic interface.

Signed-off-by: Jan Kiszka <address@hidden>
---
 target-i386/kvm.c      |  141 ++++++++++++++++++++++++++++++++++++++++++++++++
 target-i386/kvm_i386.h |   22 ++++++++
 2 files changed, 163 insertions(+), 0 deletions(-)

diff --git a/target-i386/kvm.c b/target-i386/kvm.c
index 696b14a..5e2d4f5 100644
--- a/target-i386/kvm.c
+++ b/target-i386/kvm.c
@@ -31,6 +31,7 @@
 #include "hw/apic.h"
 #include "ioport.h"
 #include "hyperv.h"
+#include "hw/pci.h"
 
 //#define DEBUG_KVM
 
@@ -2055,3 +2056,143 @@ void kvm_arch_init_irq_routing(KVMState *s)
     kvm_msi_via_irqfd_allowed = true;
     kvm_gsi_routing_allowed = true;
 }
+
+/* Classic KVM device assignment interface. Will remain x86 only. */
+int kvm_device_pci_assign(KVMState *s, PCIHostDeviceAddress *dev_addr,
+                          uint32_t flags, uint32_t *dev_id)
+{
+    struct kvm_assigned_pci_dev dev_data = {
+        .segnr = dev_addr->domain,
+        .busnr = dev_addr->bus,
+        .devfn = PCI_DEVFN(dev_addr->slot, dev_addr->function),
+        .flags = flags,
+    };
+    int ret;
+
+    dev_data.assigned_dev_id =
+        (dev_addr->domain << 16) | (dev_addr->bus << 8) | dev_data.devfn;
+
+    ret = kvm_vm_ioctl(s, KVM_ASSIGN_PCI_DEVICE, &dev_data);
+    if (ret < 0) {
+        return ret;
+    }
+
+    *dev_id = dev_data.assigned_dev_id;
+
+    return 0;
+}
+
+int kvm_device_pci_deassign(KVMState *s, uint32_t dev_id)
+{
+    struct kvm_assigned_pci_dev dev_data = {
+        .assigned_dev_id = dev_id,
+    };
+
+    return kvm_vm_ioctl(s, KVM_DEASSIGN_PCI_DEVICE, &dev_data);
+}
+
+static int kvm_assign_irq_internal(KVMState *s, uint32_t dev_id,
+                                   uint32_t irq_type, uint32_t guest_irq)
+{
+    struct kvm_assigned_irq assigned_irq = {
+        .assigned_dev_id = dev_id,
+        .guest_irq = guest_irq,
+        .flags = irq_type,
+    };
+
+    if (kvm_check_extension(s, KVM_CAP_ASSIGN_DEV_IRQ)) {
+        return kvm_vm_ioctl(s, KVM_ASSIGN_DEV_IRQ, &assigned_irq);
+    } else {
+        return kvm_vm_ioctl(s, KVM_ASSIGN_IRQ, &assigned_irq);
+    }
+}
+
+int kvm_device_intx_assign(KVMState *s, uint32_t dev_id, bool use_host_msi,
+                           uint32_t guest_irq)
+{
+    uint32_t irq_type = KVM_DEV_IRQ_GUEST_INTX |
+        (use_host_msi ? KVM_DEV_IRQ_HOST_MSI : KVM_DEV_IRQ_HOST_INTX);
+
+    return kvm_assign_irq_internal(s, dev_id, irq_type, guest_irq);
+}
+
+int kvm_device_intx_set_mask(KVMState *s, uint32_t dev_id, bool masked)
+{
+    struct kvm_assigned_pci_dev dev_data = {
+        .assigned_dev_id = dev_id,
+        .flags = masked ? KVM_DEV_ASSIGN_MASK_INTX : 0,
+    };
+
+    return kvm_vm_ioctl(s, KVM_ASSIGN_SET_INTX_MASK, &dev_data);
+}
+
+static int kvm_deassign_irq_internal(KVMState *s, uint32_t dev_id,
+                                     uint32_t type)
+{
+    struct kvm_assigned_irq assigned_irq = {
+        .assigned_dev_id = dev_id,
+        .flags = type,
+    };
+
+    return kvm_vm_ioctl(s, KVM_DEASSIGN_DEV_IRQ, &assigned_irq);
+}
+
+int kvm_device_intx_deassign(KVMState *s, uint32_t dev_id, bool use_host_msi)
+{
+    return kvm_deassign_irq_internal(s, dev_id, KVM_DEV_IRQ_GUEST_INTX |
+        (use_host_msi ? KVM_DEV_IRQ_HOST_MSI : KVM_DEV_IRQ_HOST_INTX));
+}
+
+int kvm_device_msi_assign(KVMState *s, uint32_t dev_id, int virq)
+{
+    return kvm_assign_irq_internal(s, dev_id, KVM_DEV_IRQ_HOST_MSI |
+                                              KVM_DEV_IRQ_GUEST_MSI, virq);
+}
+
+int kvm_device_msi_deassign(KVMState *s, uint32_t dev_id)
+{
+    return kvm_deassign_irq_internal(s, dev_id, KVM_DEV_IRQ_GUEST_MSI |
+                                                KVM_DEV_IRQ_HOST_MSI);
+}
+
+bool kvm_device_msix_supported(KVMState *s)
+{
+    /* The kernel lacks a corresponding KVM_CAP, so we probe by calling
+     * KVM_ASSIGN_SET_MSIX_NR with an invalid parameter. */
+    return kvm_vm_ioctl(s, KVM_ASSIGN_SET_MSIX_NR, NULL) == -EFAULT;
+}
+
+int kvm_device_msix_init_vectors(KVMState *s, uint32_t dev_id,
+                                 uint32_t nr_vectors)
+{
+    struct kvm_assigned_msix_nr msix_nr = {
+        .assigned_dev_id = dev_id,
+        .entry_nr = nr_vectors,
+    };
+
+    return kvm_vm_ioctl(s, KVM_ASSIGN_SET_MSIX_NR, &msix_nr);
+}
+
+int kvm_device_msix_set_vector(KVMState *s, uint32_t dev_id, uint32_t vector,
+                               int virq)
+{
+    struct kvm_assigned_msix_entry msix_entry = {
+        .assigned_dev_id = dev_id,
+        .gsi = virq,
+        .entry = vector,
+    };
+
+    return kvm_vm_ioctl(s, KVM_ASSIGN_SET_MSIX_ENTRY, &msix_entry);
+}
+
+int kvm_device_msix_assign(KVMState *s, uint32_t dev_id)
+{
+    return kvm_assign_irq_internal(s, dev_id, KVM_DEV_IRQ_HOST_MSIX |
+                                              KVM_DEV_IRQ_GUEST_MSIX, 0);
+}
+
+int kvm_device_msix_deassign(KVMState *s, uint32_t dev_id)
+{
+    return kvm_deassign_irq_internal(s, dev_id, KVM_DEV_IRQ_GUEST_MSIX |
+                                                KVM_DEV_IRQ_HOST_MSIX);
+}
diff --git a/target-i386/kvm_i386.h b/target-i386/kvm_i386.h
index b82bbf4..f6ab82f 100644
--- a/target-i386/kvm_i386.h
+++ b/target-i386/kvm_i386.h
@@ -11,6 +11,28 @@
 #ifndef QEMU_KVM_I386_H
 #define QEMU_KVM_I386_H
 
+#include "kvm.h"
+
 bool kvm_allows_irq0_override(void);
 
+int kvm_device_pci_assign(KVMState *s, PCIHostDeviceAddress *dev_addr,
+                          uint32_t flags, uint32_t *dev_id);
+int kvm_device_pci_deassign(KVMState *s, uint32_t dev_id);
+
+int kvm_device_intx_assign(KVMState *s, uint32_t dev_id,
+                           bool use_host_msi, uint32_t guest_irq);
+int kvm_device_intx_set_mask(KVMState *s, uint32_t dev_id, bool masked);
+int kvm_device_intx_deassign(KVMState *s, uint32_t dev_id, bool use_host_msi);
+
+int kvm_device_msi_assign(KVMState *s, uint32_t dev_id, int virq);
+int kvm_device_msi_deassign(KVMState *s, uint32_t dev_id);
+
+bool kvm_device_msix_supported(KVMState *s);
+int kvm_device_msix_init_vectors(KVMState *s, uint32_t dev_id,
+                                 uint32_t nr_vectors);
+int kvm_device_msix_set_vector(KVMState *s, uint32_t dev_id, uint32_t vector,
+                               int virq);
+int kvm_device_msix_assign(KVMState *s, uint32_t dev_id);
+int kvm_device_msix_deassign(KVMState *s, uint32_t dev_id);
+
 #endif
-- 
1.7.3.4