diff options
Diffstat (limited to 'hw/vfio/container.c')
| -rw-r--r-- | hw/vfio/container.c | 329 |
1 files changed, 249 insertions, 80 deletions
diff --git a/hw/vfio/container.c b/hw/vfio/container.c index 242010036a..b22feb8ded 100644 --- a/hw/vfio/container.c +++ b/hw/vfio/container.c @@ -33,6 +33,7 @@ #include "trace.h" #include "qapi/error.h" #include "migration/migration.h" +#include "pci.h" VFIOGroupList vfio_group_list = QLIST_HEAD_INITIALIZER(vfio_group_list); @@ -60,10 +61,11 @@ static int vfio_ram_block_discard_disable(VFIOContainer *container, bool state) } } -static int vfio_dma_unmap_bitmap(VFIOContainer *container, +static int vfio_dma_unmap_bitmap(const VFIOContainer *container, hwaddr iova, ram_addr_t size, IOMMUTLBEntry *iotlb) { + const VFIOContainerBase *bcontainer = &container->bcontainer; struct vfio_iommu_type1_dma_unmap *unmap; struct vfio_bitmap *bitmap; VFIOBitmap vbmap; @@ -91,7 +93,7 @@ static int vfio_dma_unmap_bitmap(VFIOContainer *container, bitmap->size = vbmap.size; bitmap->data = (__u64 *)vbmap.bitmap; - if (vbmap.size > container->max_dirty_bitmap_size) { + if (vbmap.size > bcontainer->max_dirty_bitmap_size) { error_report("UNMAP: Size of bitmap too big 0x%"PRIx64, vbmap.size); ret = -E2BIG; goto unmap_exit; @@ -115,9 +117,12 @@ unmap_exit: /* * DMA - Mapping and unmapping for the "type1" IOMMU interface used on x86 */ -int vfio_dma_unmap(VFIOContainer *container, hwaddr iova, - ram_addr_t size, IOMMUTLBEntry *iotlb) +static int vfio_legacy_dma_unmap(const VFIOContainerBase *bcontainer, + hwaddr iova, ram_addr_t size, + IOMMUTLBEntry *iotlb) { + const VFIOContainer *container = container_of(bcontainer, VFIOContainer, + bcontainer); struct vfio_iommu_type1_dma_unmap unmap = { .argsz = sizeof(unmap), .flags = 0, @@ -127,9 +132,9 @@ int vfio_dma_unmap(VFIOContainer *container, hwaddr iova, bool need_dirty_sync = false; int ret; - if (iotlb && vfio_devices_all_running_and_mig_active(container)) { - if (!vfio_devices_all_device_dirty_tracking(container) && - container->dirty_pages_supported) { + if (iotlb && vfio_devices_all_running_and_mig_active(bcontainer)) { + if (!vfio_devices_all_device_dirty_tracking(bcontainer) && + bcontainer->dirty_pages_supported) { return vfio_dma_unmap_bitmap(container, iova, size, iotlb); } @@ -151,8 +156,8 @@ int vfio_dma_unmap(VFIOContainer *container, hwaddr iova, */ if (errno == EINVAL && unmap.size && !(unmap.iova + unmap.size) && container->iommu_type == VFIO_TYPE1v2_IOMMU) { - trace_vfio_dma_unmap_overflow_workaround(); - unmap.size -= 1ULL << ctz64(container->pgsizes); + trace_vfio_legacy_dma_unmap_overflow_workaround(); + unmap.size -= 1ULL << ctz64(bcontainer->pgsizes); continue; } error_report("VFIO_UNMAP_DMA failed: %s", strerror(errno)); @@ -160,7 +165,7 @@ int vfio_dma_unmap(VFIOContainer *container, hwaddr iova, } if (need_dirty_sync) { - ret = vfio_get_dirty_bitmap(container, iova, size, + ret = vfio_get_dirty_bitmap(bcontainer, iova, size, iotlb->translated_addr); if (ret) { return ret; @@ -170,9 +175,11 @@ int vfio_dma_unmap(VFIOContainer *container, hwaddr iova, return 0; } -int vfio_dma_map(VFIOContainer *container, hwaddr iova, - ram_addr_t size, void *vaddr, bool readonly) +static int vfio_legacy_dma_map(const VFIOContainerBase *bcontainer, hwaddr iova, + ram_addr_t size, void *vaddr, bool readonly) { + const VFIOContainer *container = container_of(bcontainer, VFIOContainer, + bcontainer); struct vfio_iommu_type1_dma_map map = { .argsz = sizeof(map), .flags = VFIO_DMA_MAP_FLAG_READ, @@ -191,7 +198,8 @@ int vfio_dma_map(VFIOContainer *container, hwaddr iova, * the VGA ROM space. */ if (ioctl(container->fd, VFIO_IOMMU_MAP_DMA, &map) == 0 || - (errno == EBUSY && vfio_dma_unmap(container, iova, size, NULL) == 0 && + (errno == EBUSY && + vfio_legacy_dma_unmap(bcontainer, iova, size, NULL) == 0 && ioctl(container->fd, VFIO_IOMMU_MAP_DMA, &map) == 0)) { return 0; } @@ -200,17 +208,17 @@ int vfio_dma_map(VFIOContainer *container, hwaddr iova, return -errno; } -int vfio_set_dirty_page_tracking(VFIOContainer *container, bool start) +static int +vfio_legacy_set_dirty_page_tracking(const VFIOContainerBase *bcontainer, + bool start) { + const VFIOContainer *container = container_of(bcontainer, VFIOContainer, + bcontainer); int ret; struct vfio_iommu_type1_dirty_bitmap dirty = { .argsz = sizeof(dirty), }; - if (!container->dirty_pages_supported) { - return 0; - } - if (start) { dirty.flags = VFIO_IOMMU_DIRTY_PAGES_FLAG_START; } else { @@ -227,9 +235,12 @@ int vfio_set_dirty_page_tracking(VFIOContainer *container, bool start) return ret; } -int vfio_query_dirty_bitmap(VFIOContainer *container, VFIOBitmap *vbmap, - hwaddr iova, hwaddr size) +static int vfio_legacy_query_dirty_bitmap(const VFIOContainerBase *bcontainer, + VFIOBitmap *vbmap, + hwaddr iova, hwaddr size) { + const VFIOContainer *container = container_of(bcontainer, VFIOContainer, + bcontainer); struct vfio_iommu_type1_dirty_bitmap *dbitmap; struct vfio_iommu_type1_dirty_bitmap_get *range; int ret; @@ -296,7 +307,7 @@ bool vfio_get_info_dma_avail(struct vfio_iommu_type1_info *info, } static bool vfio_get_info_iova_range(struct vfio_iommu_type1_info *info, - VFIOContainer *container) + VFIOContainerBase *bcontainer) { struct vfio_info_cap_header *hdr; struct vfio_iommu_type1_info_cap_iova_range *cap; @@ -314,8 +325,8 @@ static bool vfio_get_info_iova_range(struct vfio_iommu_type1_info *info, range_set_bounds(range, cap->iova_ranges[i].start, cap->iova_ranges[i].end); - container->iova_ranges = - range_list_insert(container->iova_ranges, range); + bcontainer->iova_ranges = + range_list_insert(bcontainer->iova_ranges, range); } return true; @@ -442,6 +453,7 @@ static void vfio_get_iommu_info_migration(VFIOContainer *container, { struct vfio_info_cap_header *hdr; struct vfio_iommu_type1_info_cap_migration *cap_mig; + VFIOContainerBase *bcontainer = &container->bcontainer; hdr = vfio_get_iommu_info_cap(info, VFIO_IOMMU_TYPE1_INFO_CAP_MIGRATION); if (!hdr) { @@ -456,22 +468,17 @@ static void vfio_get_iommu_info_migration(VFIOContainer *container, * qemu_real_host_page_size to mark those dirty. */ if (cap_mig->pgsize_bitmap & qemu_real_host_page_size()) { - container->dirty_pages_supported = true; - container->max_dirty_bitmap_size = cap_mig->max_dirty_bitmap_size; - container->dirty_pgsizes = cap_mig->pgsize_bitmap; + bcontainer->dirty_pages_supported = true; + bcontainer->max_dirty_bitmap_size = cap_mig->max_dirty_bitmap_size; + bcontainer->dirty_pgsizes = cap_mig->pgsize_bitmap; } } -static void vfio_free_container(VFIOContainer *container) -{ - g_list_free_full(container->iova_ranges, g_free); - g_free(container); -} - static int vfio_connect_container(VFIOGroup *group, AddressSpace *as, Error **errp) { VFIOContainer *container; + VFIOContainerBase *bcontainer; int ret, fd; VFIOAddressSpace *space; @@ -508,7 +515,8 @@ static int vfio_connect_container(VFIOGroup *group, AddressSpace *as, * details once we know which type of IOMMU we are using. */ - QLIST_FOREACH(container, &space->containers, next) { + QLIST_FOREACH(bcontainer, &space->containers, next) { + container = container_of(bcontainer, VFIOContainer, bcontainer); if (!ioctl(group->fd, VFIO_GROUP_SET_CONTAINER, &container->fd)) { ret = vfio_ram_block_discard_disable(container, true); if (ret) { @@ -544,14 +552,9 @@ static int vfio_connect_container(VFIOGroup *group, AddressSpace *as, } container = g_malloc0(sizeof(*container)); - container->space = space; container->fd = fd; - container->error = NULL; - container->dirty_pages_supported = false; - container->dma_max_mappings = 0; - container->iova_ranges = NULL; - QLIST_INIT(&container->giommu_list); - QLIST_INIT(&container->vrdl_list); + bcontainer = &container->bcontainer; + vfio_container_init(bcontainer, space, &vfio_legacy_ops); ret = vfio_init_container(container, group->fd, errp); if (ret) { @@ -577,16 +580,16 @@ static int vfio_connect_container(VFIOGroup *group, AddressSpace *as, } if (info->flags & VFIO_IOMMU_INFO_PGSIZES) { - container->pgsizes = info->iova_pgsizes; + bcontainer->pgsizes = info->iova_pgsizes; } else { - container->pgsizes = qemu_real_host_page_size(); + bcontainer->pgsizes = qemu_real_host_page_size(); } - if (!vfio_get_info_dma_avail(info, &container->dma_max_mappings)) { - container->dma_max_mappings = 65535; + if (!vfio_get_info_dma_avail(info, &bcontainer->dma_max_mappings)) { + bcontainer->dma_max_mappings = 65535; } - vfio_get_info_iova_range(info, container); + vfio_get_info_iova_range(info, bcontainer); vfio_get_iommu_info_migration(container, info); g_free(info); @@ -606,30 +609,29 @@ static int vfio_connect_container(VFIOGroup *group, AddressSpace *as, vfio_kvm_device_add_group(group); QLIST_INIT(&container->group_list); - QLIST_INSERT_HEAD(&space->containers, container, next); + QLIST_INSERT_HEAD(&space->containers, bcontainer, next); group->container = container; QLIST_INSERT_HEAD(&container->group_list, group, container_next); - container->listener = vfio_memory_listener; + bcontainer->listener = vfio_memory_listener; + memory_listener_register(&bcontainer->listener, bcontainer->space->as); - memory_listener_register(&container->listener, container->space->as); - - if (container->error) { + if (bcontainer->error) { ret = -1; - error_propagate_prepend(errp, container->error, + error_propagate_prepend(errp, bcontainer->error, "memory listener initialization failed: "); goto listener_release_exit; } - container->initialized = true; + bcontainer->initialized = true; return 0; listener_release_exit: QLIST_REMOVE(group, container_next); - QLIST_REMOVE(container, next); + QLIST_REMOVE(bcontainer, next); vfio_kvm_device_del_group(group); - memory_listener_unregister(&container->listener); + memory_listener_unregister(&bcontainer->listener); if (container->iommu_type == VFIO_SPAPR_TCE_v2_IOMMU || container->iommu_type == VFIO_SPAPR_TCE_IOMMU) { vfio_spapr_container_deinit(container); @@ -639,7 +641,7 @@ enable_discards_exit: vfio_ram_block_discard_disable(container, false); free_container_exit: - vfio_free_container(container); + g_free(container); close_fd_exit: close(fd); @@ -653,6 +655,7 @@ put_space_exit: static void vfio_disconnect_container(VFIOGroup *group) { VFIOContainer *container = group->container; + VFIOContainerBase *bcontainer = &container->bcontainer; QLIST_REMOVE(group, container_next); group->container = NULL; @@ -663,7 +666,7 @@ static void vfio_disconnect_container(VFIOGroup *group) * group. */ if (QLIST_EMPTY(&container->group_list)) { - memory_listener_unregister(&container->listener); + memory_listener_unregister(&bcontainer->listener); if (container->iommu_type == VFIO_SPAPR_TCE_v2_IOMMU || container->iommu_type == VFIO_SPAPR_TCE_IOMMU) { vfio_spapr_container_deinit(container); @@ -676,21 +679,13 @@ static void vfio_disconnect_container(VFIOGroup *group) } if (QLIST_EMPTY(&container->group_list)) { - VFIOAddressSpace *space = container->space; - VFIOGuestIOMMU *giommu, *tmp; + VFIOAddressSpace *space = bcontainer->space; - QLIST_REMOVE(container, next); - - QLIST_FOREACH_SAFE(giommu, &container->giommu_list, giommu_next, tmp) { - memory_region_unregister_iommu_notifier( - MEMORY_REGION(giommu->iommu_mr), &giommu->n); - QLIST_REMOVE(giommu, giommu_next); - g_free(giommu); - } + vfio_container_destroy(bcontainer); trace_vfio_disconnect_container(container->fd); close(container->fd); - vfio_free_container(container); + g_free(container); vfio_put_address_space(space); } @@ -705,7 +700,7 @@ static VFIOGroup *vfio_get_group(int groupid, AddressSpace *as, Error **errp) QLIST_FOREACH(group, &vfio_group_list, next) { if (group->groupid == groupid) { /* Found it. Now is it already in the right context? */ - if (group->container->space->as == as) { + if (group->container->bcontainer.space->as == as) { return group; } else { error_setg(errp, "group %d used in multiple address spaces", @@ -877,13 +872,13 @@ static int vfio_device_groupid(VFIODevice *vbasedev, Error **errp) * @name and @vbasedev->name are likely to be different depending * on the type of the device, hence the need for passing @name */ -int vfio_attach_device(char *name, VFIODevice *vbasedev, - AddressSpace *as, Error **errp) +static int vfio_legacy_attach_device(const char *name, VFIODevice *vbasedev, + AddressSpace *as, Error **errp) { int groupid = vfio_device_groupid(vbasedev, errp); VFIODevice *vbasedev_iter; VFIOGroup *group; - VFIOContainer *container; + VFIOContainerBase *bcontainer; int ret; if (groupid < 0) { @@ -910,26 +905,200 @@ int vfio_attach_device(char *name, VFIODevice *vbasedev, return ret; } - container = group->container; - vbasedev->container = container; - QLIST_INSERT_HEAD(&container->device_list, vbasedev, container_next); + bcontainer = &group->container->bcontainer; + vbasedev->bcontainer = bcontainer; + QLIST_INSERT_HEAD(&bcontainer->device_list, vbasedev, container_next); QLIST_INSERT_HEAD(&vfio_device_list, vbasedev, global_next); return ret; } -void vfio_detach_device(VFIODevice *vbasedev) +static void vfio_legacy_detach_device(VFIODevice *vbasedev) { VFIOGroup *group = vbasedev->group; - if (!vbasedev->container) { - return; - } - QLIST_REMOVE(vbasedev, global_next); QLIST_REMOVE(vbasedev, container_next); - vbasedev->container = NULL; + vbasedev->bcontainer = NULL; trace_vfio_detach_device(vbasedev->name, group->groupid); vfio_put_base_device(vbasedev); vfio_put_group(group); } + +static int vfio_legacy_pci_hot_reset(VFIODevice *vbasedev, bool single) +{ + VFIOPCIDevice *vdev = container_of(vbasedev, VFIOPCIDevice, vbasedev); + VFIOGroup *group; + struct vfio_pci_hot_reset_info *info = NULL; + struct vfio_pci_dependent_device *devices; + struct vfio_pci_hot_reset *reset; + int32_t *fds; + int ret, i, count; + bool multi = false; + + trace_vfio_pci_hot_reset(vdev->vbasedev.name, single ? "one" : "multi"); + + if (!single) { + vfio_pci_pre_reset(vdev); + } + vdev->vbasedev.needs_reset = false; + + ret = vfio_pci_get_pci_hot_reset_info(vdev, &info); + + if (ret) { + goto out_single; + } + devices = &info->devices[0]; + + trace_vfio_pci_hot_reset_has_dep_devices(vdev->vbasedev.name); + + /* Verify that we have all the groups required */ + for (i = 0; i < info->count; i++) { + PCIHostDeviceAddress host; + VFIOPCIDevice *tmp; + VFIODevice *vbasedev_iter; + + host.domain = devices[i].segment; + host.bus = devices[i].bus; + host.slot = PCI_SLOT(devices[i].devfn); + host.function = PCI_FUNC(devices[i].devfn); + + trace_vfio_pci_hot_reset_dep_devices(host.domain, + host.bus, host.slot, host.function, devices[i].group_id); + + if (vfio_pci_host_match(&host, vdev->vbasedev.name)) { + continue; + } + + QLIST_FOREACH(group, &vfio_group_list, next) { + if (group->groupid == devices[i].group_id) { + break; + } + } + + if (!group) { + if (!vdev->has_pm_reset) { + error_report("vfio: Cannot reset device %s, " + "depends on group %d which is not owned.", + vdev->vbasedev.name, devices[i].group_id); + } + ret = -EPERM; + goto out; + } + + /* Prep dependent devices for reset and clear our marker. */ + QLIST_FOREACH(vbasedev_iter, &group->device_list, next) { + if (!vbasedev_iter->dev->realized || + vbasedev_iter->type != VFIO_DEVICE_TYPE_PCI) { + continue; + } + tmp = container_of(vbasedev_iter, VFIOPCIDevice, vbasedev); + if (vfio_pci_host_match(&host, tmp->vbasedev.name)) { + if (single) { + ret = -EINVAL; + goto out_single; + } + vfio_pci_pre_reset(tmp); + tmp->vbasedev.needs_reset = false; + multi = true; + break; + } + } + } + + if (!single && !multi) { + ret = -EINVAL; + goto out_single; + } + + /* Determine how many group fds need to be passed */ + count = 0; + QLIST_FOREACH(group, &vfio_group_list, next) { + for (i = 0; i < info->count; i++) { + if (group->groupid == devices[i].group_id) { + count++; + break; + } + } + } + + reset = g_malloc0(sizeof(*reset) + (count * sizeof(*fds))); + reset->argsz = sizeof(*reset) + (count * sizeof(*fds)); + fds = &reset->group_fds[0]; + + /* Fill in group fds */ + QLIST_FOREACH(group, &vfio_group_list, next) { + for (i = 0; i < info->count; i++) { + if (group->groupid == devices[i].group_id) { + fds[reset->count++] = group->fd; + break; + } + } + } + + /* Bus reset! */ + ret = ioctl(vdev->vbasedev.fd, VFIO_DEVICE_PCI_HOT_RESET, reset); + g_free(reset); + if (ret) { + ret = -errno; + } + + trace_vfio_pci_hot_reset_result(vdev->vbasedev.name, + ret ? strerror(errno) : "Success"); + +out: + /* Re-enable INTx on affected devices */ + for (i = 0; i < info->count; i++) { + PCIHostDeviceAddress host; + VFIOPCIDevice *tmp; + VFIODevice *vbasedev_iter; + + host.domain = devices[i].segment; + host.bus = devices[i].bus; + host.slot = PCI_SLOT(devices[i].devfn); + host.function = PCI_FUNC(devices[i].devfn); + + if (vfio_pci_host_match(&host, vdev->vbasedev.name)) { + continue; + } + + QLIST_FOREACH(group, &vfio_group_list, next) { + if (group->groupid == devices[i].group_id) { + break; + } + } + + if (!group) { + break; + } + + QLIST_FOREACH(vbasedev_iter, &group->device_list, next) { + if (!vbasedev_iter->dev->realized || + vbasedev_iter->type != VFIO_DEVICE_TYPE_PCI) { + continue; + } + tmp = container_of(vbasedev_iter, VFIOPCIDevice, vbasedev); + if (vfio_pci_host_match(&host, tmp->vbasedev.name)) { + vfio_pci_post_reset(tmp); + break; + } + } + } +out_single: + if (!single) { + vfio_pci_post_reset(vdev); + } + g_free(info); + + return ret; +} + +const VFIOIOMMUOps vfio_legacy_ops = { + .dma_map = vfio_legacy_dma_map, + .dma_unmap = vfio_legacy_dma_unmap, + .attach_device = vfio_legacy_attach_device, + .detach_device = vfio_legacy_detach_device, + .set_dirty_page_tracking = vfio_legacy_set_dirty_page_tracking, + .query_dirty_bitmap = vfio_legacy_query_dirty_bitmap, + .pci_hot_reset = vfio_legacy_pci_hot_reset, +}; |