[Qemu-devel] [PATCH 07/11] vfio: Introduce VFIO address spaces

[David Gibson <david@gibson.dropbear.id.au>]

The only model so far supported for VFIO passthrough devices is the model
usually used on x86, where all of the guest's RAM is mapped into the
(host) IOMMU and there is no IOMMU visible in the guest.

This patch begins to relax this model, introducing the notion of a
VFIOAddressSpace.  This represents a logical DMA address space which will
be visible to one or more VFIO devices by appropriate mapping in the (host)
IOMMU.  Thus the currently global list of containers becomes local to
a VFIOAddressSpace, and we verify that we don't attempt to add a VFIO
group to multiple address spaces.

For now, only one VFIOAddressSpace is created and used, corresponding to
main system memory, that will change in future patches.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
---
 hw/misc/vfio.c |   63 ++++++++++++++++++++++++++++++++++++++++++--------------
 1 file changed, 48 insertions(+), 15 deletions(-)

diff --git a/hw/misc/vfio.c b/hw/misc/vfio.c
index c4a8853..e557f69 100644
--- a/hw/misc/vfio.c
+++ b/hw/misc/vfio.c
@@ -113,9 +113,17 @@ enum {
     VFIO_INT_MSIX = 3,
 };
 
+typedef struct VFIOAddressSpace {
+    AddressSpace *as;
+    QLIST_HEAD(, VFIOContainer) containers;
+} VFIOAddressSpace;
+
+static VFIOAddressSpace vfio_address_space_memory;
+
 struct VFIOGroup;
 
 typedef struct VFIOContainer {
+    VFIOAddressSpace *space;
     int fd; /* /dev/vfio/vfio, empowered by the attached groups */
     struct {
         /* enable abstraction to support various iommu backends */
@@ -178,9 +186,6 @@ typedef struct VFIOGroup {
 
 #define MSIX_CAP_LENGTH 12
 
-static QLIST_HEAD(, VFIOContainer)
-    container_list = QLIST_HEAD_INITIALIZER(container_list);
-
 static QLIST_HEAD(, VFIOGroup)
     group_list = QLIST_HEAD_INITIALIZER(group_list);
 
@@ -2624,16 +2629,28 @@ static int vfio_load_rom(VFIODevice *vdev)
     return 0;
 }
 
-static int vfio_connect_container(VFIOGroup *group)
+static void vfio_address_space_init(VFIOAddressSpace *space, AddressSpace *as)
+{
+    space->as = as;
+    QLIST_INIT(&space->containers);
+}
+
+static int vfio_connect_container(VFIOGroup *group, VFIOAddressSpace *space)
 {
     VFIOContainer *container;
     int ret, fd;
 
     if (group->container) {
-        return 0;
+        if (group->container->space == space) {
+            return 0;
+        } else {
+            error_report("vfio: group %d used in multiple address spaces",
+                         group->groupid);
+            return -EBUSY;
+        }
     }
 
-    QLIST_FOREACH(container, &container_list, next) {
+    QLIST_FOREACH(container, &space->containers, next) {
         if (!ioctl(group->fd, VFIO_GROUP_SET_CONTAINER, &container->fd)) {
             group->container = container;
             QLIST_INSERT_HEAD(&container->group_list, group, container_next);
@@ -2656,6 +2673,7 @@ static int vfio_connect_container(VFIOGroup *group)
     }
 
     container = g_malloc0(sizeof(*container));
+    container->space = space;
     container->fd = fd;
 
     if (ioctl(fd, VFIO_CHECK_EXTENSION, VFIO_TYPE1_IOMMU)) {
@@ -2678,7 +2696,8 @@ static int vfio_connect_container(VFIOGroup *group)
         container->iommu_data.listener = vfio_memory_listener;
         container->iommu_data.release = vfio_listener_release;
 
-        memory_listener_register(&container->iommu_data.listener, &address_space_memory);
+        memory_listener_register(&container->iommu_data.listener,
+                                 container->space->as);
     } else {
         error_report("vfio: No available IOMMU models");
         g_free(container);
@@ -2687,7 +2706,7 @@ static int vfio_connect_container(VFIOGroup *group)
     }
 
     QLIST_INIT(&container->group_list);
-    QLIST_INSERT_HEAD(&container_list, container, next);
+    QLIST_INSERT_HEAD(&space->containers, container, next);
 
     group->container = container;
     QLIST_INSERT_HEAD(&container->group_list, group, container_next);
@@ -2700,7 +2719,7 @@ static void vfio_disconnect_container(VFIOGroup *group)
     VFIOContainer *container = group->container;
 
     if (ioctl(group->fd, VFIO_GROUP_UNSET_CONTAINER, &container->fd)) {
-        error_report("vfio: error disconnecting group %d from container",
+        error_report("vfio: error disconnecting group %d from context",
                      group->groupid);
     }
 
@@ -2712,13 +2731,13 @@ static void vfio_disconnect_container(VFIOGroup *group)
             container->iommu_data.release(container);
         }
         QLIST_REMOVE(container, next);
-        DPRINTF("vfio_disconnect_container: close container->fd\n");
+        DPRINTF("vfio_disconnect: close container->fd\n");
         close(container->fd);
         g_free(container);
     }
 }
 
-static VFIOGroup *vfio_get_group(int groupid)
+static VFIOGroup *vfio_get_group(int groupid, VFIOAddressSpace *space)
 {
     VFIOGroup *group;
     char path[32];
@@ -2726,7 +2745,15 @@ static VFIOGroup *vfio_get_group(int groupid)
 
     QLIST_FOREACH(group, &group_list, next) {
         if (group->groupid == groupid) {
-            return group;
+            /* Found it.  Now is it already in the right context? */
+            assert(group->container);
+            if (group->container->space == space) {
+                return group;
+            } else {
+                error_report("vfio: group %d used in multiple address spaces",
+                             group->groupid);
+                return NULL;
+            }
         }
     }
 
@@ -2759,8 +2786,8 @@ static VFIOGroup *vfio_get_group(int groupid)
     group->groupid = groupid;
     QLIST_INIT(&group->device_list);
 
-    if (vfio_connect_container(group)) {
-        error_report("vfio: failed to setup container for group %d", groupid);
+    if (vfio_connect_container(group, space)) {
+        error_report("vfio: failed to setup context for group %d", groupid);
         close(group->fd);
         g_free(group);
         return NULL;
@@ -2992,7 +3019,12 @@ static int vfio_initfn(PCIDevice *pdev)
     DPRINTF("%s(%04x:%02x:%02x.%x) group %d\n", __func__, vdev->host.domain,
             vdev->host.bus, vdev->host.slot, vdev->host.function, groupid);
 
-    group = vfio_get_group(groupid);
+    if (pci_iommu_as(pdev) != &address_space_memory) {
+        error_report("vfio: DMA address space must be system memory");
+        return -ENXIO;
+    }
+
+    group = vfio_get_group(groupid, &vfio_address_space_memory);
     if (!group) {
         error_report("vfio: failed to get group %d", groupid);
         return -ENOENT;
@@ -3212,6 +3244,7 @@ static const TypeInfo vfio_pci_dev_info = {
 
 static void register_vfio_pci_dev_type(void)
 {
+    vfio_address_space_init(&vfio_address_space_memory, &address_space_memory);
     type_register_static(&vfio_pci_dev_info);
 }
 
-- 
1.7.10.4