From: Eric Auger eric.auger@redhat.com

No functional change intended.

Signed-off-by: Eric Auger eric.auger@redhat.com Signed-off-by: Yi Liu yi.l.liu@intel.com Signed-off-by: Yi Sun yi.y.sun@linux.intel.com Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Cédric Le Goater clg@redhat.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/vfio/common.c | 45 +++++++++++++++------------ hw/vfio/container-base.c | 32 +++++++++++++++++++ hw/vfio/container.c | 22 ++++++++----- hw/vfio/meson.build | 1 + hw/vfio/trace-events | 2 +- include/hw/vfio/vfio-common.h | 4 --- include/hw/vfio/vfio-container-base.h | 7 +++++ 7 files changed, 81 insertions(+), 32 deletions(-) create mode 100644 hw/vfio/container-base.c

diff --git a/hw/vfio/common.c b/hw/vfio/common.c index e08b147b3d..ea63271167 100644 --- a/hw/vfio/common.c +++ b/hw/vfio/common.c @@ -292,7 +292,7 @@ static bool vfio_get_xlat_addr(IOMMUTLBEntry *iotlb, void **vaddr, static void vfio_iommu_map_notify(IOMMUNotifier *n, IOMMUTLBEntry *iotlb) { VFIOGuestIOMMU *giommu = container_of(n, VFIOGuestIOMMU, n); - VFIOContainer *container = giommu->container; + VFIOContainerBase *bcontainer = &giommu->container->bcontainer; hwaddr iova = iotlb->iova + giommu->iommu_offset; void *vaddr; int ret; @@ -322,21 +322,22 @@ static void vfio_iommu_map_notify(IOMMUNotifier *n, IOMMUTLBEntry *iotlb) * of vaddr will always be there, even if the memory object is * destroyed and its backing memory munmap-ed. */ - ret = vfio_dma_map(container, iova, - iotlb->addr_mask + 1, vaddr, - read_only); + ret = vfio_container_dma_map(bcontainer, iova, + iotlb->addr_mask + 1, vaddr, + read_only); if (ret) { - error_report("vfio_dma_map(%p, 0x%"HWADDR_PRIx", " + error_report("vfio_container_dma_map(%p, 0x%"HWADDR_PRIx", " "0x%"HWADDR_PRIx", %p) = %d (%s)", - container, iova, + bcontainer, iova, iotlb->addr_mask + 1, vaddr, ret, strerror(-ret)); } } else { - ret = vfio_dma_unmap(container, iova, iotlb->addr_mask + 1, iotlb); + ret = vfio_container_dma_unmap(bcontainer, iova, + iotlb->addr_mask + 1, iotlb); if (ret) { - error_report("vfio_dma_unmap(%p, 0x%"HWADDR_PRIx", " + error_report("vfio_container_dma_unmap(%p, 0x%"HWADDR_PRIx", " "0x%"HWADDR_PRIx") = %d (%s)", - container, iova, + bcontainer, iova, iotlb->addr_mask + 1, ret, strerror(-ret)); vfio_set_migration_error(ret); } @@ -355,9 +356,10 @@ static void vfio_ram_discard_notify_discard(RamDiscardListener *rdl, int ret;

/* Unmap with a single call. */ - ret = vfio_dma_unmap(vrdl->container, iova, size , NULL); + ret = vfio_container_dma_unmap(&vrdl->container->bcontainer, + iova, size , NULL); if (ret) { - error_report("%s: vfio_dma_unmap() failed: %s", __func__, + error_report("%s: vfio_container_dma_unmap() failed: %s", __func__, strerror(-ret)); } } @@ -385,8 +387,8 @@ static int vfio_ram_discard_notify_populate(RamDiscardListener *rdl, section->offset_within_address_space; vaddr = memory_region_get_ram_ptr(section->mr) + start;

- ret = vfio_dma_map(vrdl->container, iova, next - start, - vaddr, section->readonly); + ret = vfio_container_dma_map(&vrdl->container->bcontainer, iova, + next - start, vaddr, section->readonly); if (ret) { /* Rollback */ vfio_ram_discard_notify_discard(rdl, section); @@ -684,10 +686,11 @@ static void vfio_listener_region_add(MemoryListener *listener, } }

- ret = vfio_dma_map(container, iova, int128_get64(llsize), - vaddr, section->readonly); + ret = vfio_container_dma_map(&container->bcontainer, + iova, int128_get64(llsize), vaddr, + section->readonly); if (ret) { - error_setg(&err, "vfio_dma_map(%p, 0x%"HWADDR_PRIx", " + error_setg(&err, "vfio_container_dma_map(%p, 0x%"HWADDR_PRIx", " "0x%"HWADDR_PRIx", %p) = %d (%s)", container, iova, int128_get64(llsize), vaddr, ret, strerror(-ret)); @@ -784,18 +787,20 @@ static void vfio_listener_region_del(MemoryListener *listener, if (int128_eq(llsize, int128_2_64())) { /* The unmap ioctl doesn't accept a full 64-bit span. */ llsize = int128_rshift(llsize, 1); - ret = vfio_dma_unmap(container, iova, int128_get64(llsize), NULL); + ret = vfio_container_dma_unmap(&container->bcontainer, iova, + int128_get64(llsize), NULL); if (ret) { - error_report("vfio_dma_unmap(%p, 0x%"HWADDR_PRIx", " + error_report("vfio_container_dma_unmap(%p, 0x%"HWADDR_PRIx", " "0x%"HWADDR_PRIx") = %d (%s)", container, iova, int128_get64(llsize), ret, strerror(-ret)); } iova += int128_get64(llsize); } - ret = vfio_dma_unmap(container, iova, int128_get64(llsize), NULL); + ret = vfio_container_dma_unmap(&container->bcontainer, iova, + int128_get64(llsize), NULL); if (ret) { - error_report("vfio_dma_unmap(%p, 0x%"HWADDR_PRIx", " + error_report("vfio_container_dma_unmap(%p, 0x%"HWADDR_PRIx", " "0x%"HWADDR_PRIx") = %d (%s)", container, iova, int128_get64(llsize), ret, strerror(-ret)); diff --git a/hw/vfio/container-base.c b/hw/vfio/container-base.c new file mode 100644 index 0000000000..55d3a35fa4 --- /dev/null +++ b/hw/vfio/container-base.c @@ -0,0 +1,32 @@ +/* + * VFIO BASE CONTAINER + * + * Copyright (C) 2023 Intel Corporation. + * Copyright Red Hat, Inc. 2023 + * + * Authors: Yi Liu yi.l.liu@intel.com + * Eric Auger eric.auger@redhat.com + * + * SPDX-License-Identifier: GPL-2.0-or-later + */ + +#include "qemu/osdep.h" +#include "qapi/error.h" +#include "qemu/error-report.h" +#include "hw/vfio/vfio-container-base.h" + +int vfio_container_dma_map(VFIOContainerBase *bcontainer, + hwaddr iova, ram_addr_t size, + void *vaddr, bool readonly) +{ + g_assert(bcontainer->ops->dma_map); + return bcontainer->ops->dma_map(bcontainer, iova, size, vaddr, readonly); +} + +int vfio_container_dma_unmap(VFIOContainerBase *bcontainer, + hwaddr iova, ram_addr_t size, + IOMMUTLBEntry *iotlb) +{ + g_assert(bcontainer->ops->dma_unmap); + return bcontainer->ops->dma_unmap(bcontainer, iova, size, iotlb); +} diff --git a/hw/vfio/container.c b/hw/vfio/container.c index 8d8ed13e93..40e378e888 100644 --- a/hw/vfio/container.c +++ b/hw/vfio/container.c @@ -140,9 +140,11 @@ void vfio_dma_range_init_dirty_bitmap(VFIODMARange *qrange) /* * 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(VFIOContainerBase *bcontainer, hwaddr iova, + ram_addr_t size, IOMMUTLBEntry *iotlb) { + VFIOContainer *container = container_of(bcontainer, VFIOContainer, + bcontainer); struct vfio_iommu_type1_dma_unmap unmap = { .argsz = sizeof(unmap), .flags = 0, @@ -193,7 +195,7 @@ 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(); + trace_vfio_legacy_dma_unmap_overflow_workaround(); unmap.size -= 1ULL << ctz64(container->pgsizes); continue; } @@ -212,9 +214,11 @@ 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(VFIOContainerBase *bcontainer, hwaddr iova, + ram_addr_t size, void *vaddr, bool readonly) { + VFIOContainer *container = container_of(bcontainer, VFIOContainer, + bcontainer); struct vfio_iommu_type1_dma_map map = { .argsz = sizeof(map), .flags = VFIO_DMA_MAP_FLAG_READ, @@ -241,7 +245,8 @@ 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; } @@ -1050,4 +1055,7 @@ void vfio_detach_device(VFIODevice *vbasedev) vfio_put_group(group); }

-const VFIOIOMMUOps vfio_legacy_ops; +const VFIOIOMMUOps vfio_legacy_ops = { + .dma_map = vfio_legacy_dma_map, + .dma_unmap = vfio_legacy_dma_unmap, +}; diff --git a/hw/vfio/meson.build b/hw/vfio/meson.build index b1db4c8605..32a6933280 100644 --- a/hw/vfio/meson.build +++ b/hw/vfio/meson.build @@ -2,6 +2,7 @@ vfio_ss = ss.source_set() vfio_ss.add(files( 'helpers.c', 'common.c', + 'container-base.c', 'container.c', 'spapr.c', 'migration.c', diff --git a/hw/vfio/trace-events b/hw/vfio/trace-events index 0eb2387cf2..9f7fedee98 100644 --- a/hw/vfio/trace-events +++ b/hw/vfio/trace-events @@ -116,7 +116,7 @@ vfio_region_unmap(const char *name, unsigned long offset, unsigned long end) "Re vfio_region_sparse_mmap_header(const char *name, int index, int nr_areas) "Device %s region %d: %d sparse mmap entries" vfio_region_sparse_mmap_entry(int i, unsigned long start, unsigned long end) "sparse entry %d [0x%lx - 0x%lx]" vfio_get_dev_region(const char *name, int index, uint32_t type, uint32_t subtype) "%s index %d, %08x/%08x" -vfio_dma_unmap_overflow_workaround(void) "" +vfio_legacy_dma_unmap_overflow_workaround(void) "" vfio_get_dirty_bitmap(int fd, uint64_t iova, uint64_t size, uint64_t bitmap_size, uint64_t start, uint64_t dirty_pages) "container fd=%d, iova=0x%"PRIx64" size= 0x%"PRIx64" bitmap_size=0x%"PRIx64" start=0x%"PRIx64" dirty_pages=%"PRIu64 vfio_iommu_map_dirty_notify(uint64_t iova_start, uint64_t iova_end) "iommu dirty @ 0x%"PRIx64" - 0x%"PRIx64

diff --git a/include/hw/vfio/vfio-common.h b/include/hw/vfio/vfio-common.h index 3a0a6ab6ee..f94baf72db 100644 --- a/include/hw/vfio/vfio-common.h +++ b/include/hw/vfio/vfio-common.h @@ -221,10 +221,6 @@ bool vfio_devices_all_running_and_saving(VFIOContainer *container); VFIODMARange *vfio_lookup_match_range(VFIOContainer *container, hwaddr start_addr, hwaddr size); void vfio_dma_range_init_dirty_bitmap(VFIODMARange *qrange); -int vfio_dma_unmap(VFIOContainer *container, hwaddr iova, - ram_addr_t size, IOMMUTLBEntry *iotlb); -int vfio_dma_map(VFIOContainer *container, hwaddr iova, - ram_addr_t size, void *vaddr, bool readonly); int vfio_set_dirty_page_tracking(VFIOContainer *container, bool start); int vfio_query_dirty_bitmap(VFIOContainer *container, VFIOBitmap *vbmap, hwaddr iova, hwaddr size); diff --git a/include/hw/vfio/vfio-container-base.h b/include/hw/vfio/vfio-container-base.h index 1d6daaea5d..56b033f59f 100644 --- a/include/hw/vfio/vfio-container-base.h +++ b/include/hw/vfio/vfio-container-base.h @@ -31,6 +31,13 @@ typedef struct VFIOContainerBase { const VFIOIOMMUOps *ops; } VFIOContainerBase;

+int vfio_container_dma_map(VFIOContainerBase *bcontainer, + hwaddr iova, ram_addr_t size, + void *vaddr, bool readonly); +int vfio_container_dma_unmap(VFIOContainerBase *bcontainer, + hwaddr iova, ram_addr_t size, + IOMMUTLBEntry *iotlb); + struct VFIOIOMMUOps { /* basic feature */ int (*dma_map)(VFIOContainerBase *bcontainer,

From: Eric Auger eric.auger@redhat.com

Move the giommu_list field in the base container and store the base container in the VFIOGuestIOMMU.

No functional change intended.

Signed-off-by: Eric Auger eric.auger@redhat.com Signed-off-by: Yi Liu yi.l.liu@intel.com Signed-off-by: Yi Sun yi.y.sun@linux.intel.com Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Cédric Le Goater clg@redhat.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/vfio/common.c | 17 +++++++++++------ hw/vfio/container-base.c | 9 +++++++++ hw/vfio/container.c | 8 -------- include/hw/vfio/vfio-common.h | 9 --------- include/hw/vfio/vfio-container-base.h | 9 +++++++++ 5 files changed, 29 insertions(+), 23 deletions(-)

diff --git a/hw/vfio/common.c b/hw/vfio/common.c index ea63271167..b8007b22c3 100644 --- a/hw/vfio/common.c +++ b/hw/vfio/common.c @@ -292,7 +292,7 @@ static bool vfio_get_xlat_addr(IOMMUTLBEntry *iotlb, void **vaddr, static void vfio_iommu_map_notify(IOMMUNotifier *n, IOMMUTLBEntry *iotlb) { VFIOGuestIOMMU *giommu = container_of(n, VFIOGuestIOMMU, n); - VFIOContainerBase *bcontainer = &giommu->container->bcontainer; + VFIOContainerBase *bcontainer = giommu->bcontainer; hwaddr iova = iotlb->iova + giommu->iommu_offset; void *vaddr; int ret; @@ -569,6 +569,7 @@ static void vfio_listener_region_add(MemoryListener *listener, MemoryRegionSection *section) { VFIOContainer *container = container_of(listener, VFIOContainer, listener); + VFIOContainerBase *bcontainer = &container->bcontainer; hwaddr iova, end; Int128 llend, llsize; void *vaddr; @@ -612,7 +613,7 @@ static void vfio_listener_region_add(MemoryListener *listener, giommu->iommu_mr = iommu_mr; giommu->iommu_offset = section->offset_within_address_space - section->offset_within_region; - giommu->container = container; + giommu->bcontainer = bcontainer; llend = int128_add(int128_make64(section->offset_within_region), section->size); llend = int128_sub(llend, int128_one()); @@ -647,7 +648,7 @@ static void vfio_listener_region_add(MemoryListener *listener, g_free(giommu); goto fail; } - QLIST_INSERT_HEAD(&container->giommu_list, giommu, giommu_next); + QLIST_INSERT_HEAD(&bcontainer->giommu_list, giommu, giommu_next); memory_region_iommu_replay(giommu->iommu_mr, &giommu->n);

return; @@ -732,6 +733,7 @@ static void vfio_listener_region_del(MemoryListener *listener, MemoryRegionSection *section) { VFIOContainer *container = container_of(listener, VFIOContainer, listener); + VFIOContainerBase *bcontainer = &container->bcontainer; hwaddr iova, end; Int128 llend, llsize; int ret; @@ -744,7 +746,7 @@ static void vfio_listener_region_del(MemoryListener *listener, if (memory_region_is_iommu(section->mr)) { VFIOGuestIOMMU *giommu;

- QLIST_FOREACH(giommu, &container->giommu_list, giommu_next) { + QLIST_FOREACH(giommu, &bcontainer->giommu_list, giommu_next) { if (MEMORY_REGION(giommu->iommu_mr) == section->mr && giommu->n.start == section->offset_within_region) { memory_region_unregister_iommu_notifier(section->mr, @@ -1211,7 +1213,9 @@ static void vfio_iommu_map_dirty_notify(IOMMUNotifier *n, IOMMUTLBEntry *iotlb) vfio_giommu_dirty_notifier *gdn = container_of(n, vfio_giommu_dirty_notifier, n); VFIOGuestIOMMU *giommu = gdn->giommu; - VFIOContainer *container = giommu->container; + VFIOContainerBase *bcontainer = giommu->bcontainer; + VFIOContainer *container = container_of(bcontainer, VFIOContainer, + bcontainer); hwaddr iova = iotlb->iova + giommu->iommu_offset; ram_addr_t translated_addr; int ret = -EINVAL; @@ -1289,12 +1293,13 @@ static int vfio_sync_ram_discard_listener_dirty_bitmap(VFIOContainer *container, static int vfio_sync_dirty_bitmap(VFIOContainer *container, MemoryRegionSection *section) { + VFIOContainerBase *bcontainer = &container->bcontainer; ram_addr_t ram_addr;

if (memory_region_is_iommu(section->mr)) { VFIOGuestIOMMU *giommu;

- QLIST_FOREACH(giommu, &container->giommu_list, giommu_next) { + QLIST_FOREACH(giommu, &bcontainer->giommu_list, giommu_next) { if (MEMORY_REGION(giommu->iommu_mr) == section->mr && giommu->n.start == section->offset_within_region) { Int128 llend; diff --git a/hw/vfio/container-base.c b/hw/vfio/container-base.c index e929435751..20bcb9669a 100644 --- a/hw/vfio/container-base.c +++ b/hw/vfio/container-base.c @@ -34,8 +34,17 @@ int vfio_container_dma_unmap(VFIOContainerBase *bcontainer, void vfio_container_init(VFIOContainerBase *bcontainer, const VFIOIOMMUOps *ops) { bcontainer->ops = ops; + QLIST_INIT(&bcontainer->giommu_list); }

void vfio_container_destroy(VFIOContainerBase *bcontainer) { + VFIOGuestIOMMU *giommu, *tmp; + + QLIST_FOREACH_SAFE(giommu, &bcontainer->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); + } } diff --git a/hw/vfio/container.c b/hw/vfio/container.c index 5a8c55056b..03791601d0 100644 --- a/hw/vfio/container.c +++ b/hw/vfio/container.c @@ -649,7 +649,6 @@ static int vfio_connect_container(VFIOGroup *group, AddressSpace *as, container->dirty_pages_supported = false; container->dma_max_mappings = 0; container->iova_ranges = NULL; - QLIST_INIT(&container->giommu_list); QLIST_INIT(&container->vrdl_list); QLIST_INIT(&container->dma_list); bcontainer = &container->bcontainer; @@ -794,16 +793,9 @@ static void vfio_disconnect_container(VFIOGroup *group)

if (QLIST_EMPTY(&container->group_list)) { VFIOAddressSpace *space = container->space; - VFIOGuestIOMMU *giommu, *tmp;

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); diff --git a/include/hw/vfio/vfio-common.h b/include/hw/vfio/vfio-common.h index f94baf72db..6f02952ff6 100644 --- a/include/hw/vfio/vfio-common.h +++ b/include/hw/vfio/vfio-common.h @@ -104,7 +104,6 @@ typedef struct VFIOContainer { uint64_t max_dirty_bitmap_size; unsigned long pgsizes; unsigned int dma_max_mappings; - QLIST_HEAD(, VFIOGuestIOMMU) giommu_list; QLIST_HEAD(, VFIOHostDMAWindow) hostwin_list; QLIST_HEAD(, VFIOGroup) group_list; QLIST_HEAD(, VFIORamDiscardListener) vrdl_list; @@ -114,14 +113,6 @@ typedef struct VFIOContainer { GList *iova_ranges; } VFIOContainer;

-typedef struct VFIOGuestIOMMU { - VFIOContainer *container; - IOMMUMemoryRegion *iommu_mr; - hwaddr iommu_offset; - IOMMUNotifier n; - QLIST_ENTRY(VFIOGuestIOMMU) giommu_next; -} VFIOGuestIOMMU; - typedef struct VFIORamDiscardListener { VFIOContainer *container; MemoryRegion *mr; diff --git a/include/hw/vfio/vfio-container-base.h b/include/hw/vfio/vfio-container-base.h index 577f52ccbc..a11aec5755 100644 --- a/include/hw/vfio/vfio-container-base.h +++ b/include/hw/vfio/vfio-container-base.h @@ -29,8 +29,17 @@ typedef struct { */ typedef struct VFIOContainerBase { const VFIOIOMMUOps *ops; + QLIST_HEAD(, VFIOGuestIOMMU) giommu_list; } VFIOContainerBase;

+typedef struct VFIOGuestIOMMU { + VFIOContainerBase *bcontainer; + IOMMUMemoryRegion *iommu_mr; + hwaddr iommu_offset; + IOMMUNotifier n; + QLIST_ENTRY(VFIOGuestIOMMU) giommu_next; +} VFIOGuestIOMMU; + int vfio_container_dma_map(VFIOContainerBase *bcontainer, hwaddr iova, ram_addr_t size, void *vaddr, bool readonly);

From: Eric Auger eric.auger@redhat.com

dirty_pages_supported field is also moved to the base container

No functional change intended.

Modify vfio_listener_log_clear during backporting.

Signed-off-by: Eric Auger eric.auger@redhat.com Signed-off-by: Yi Liu yi.l.liu@intel.com Signed-off-by: Yi Sun yi.y.sun@linux.intel.com Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Cédric Le Goater clg@redhat.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/vfio/common.c | 14 +++++++++----- hw/vfio/container-base.c | 16 ++++++++++++++++ hw/vfio/container.c | 21 ++++++++++++++------- include/hw/vfio/vfio-common.h | 5 ----- include/hw/vfio/vfio-container-base.h | 6 ++++++ 5 files changed, 45 insertions(+), 17 deletions(-)

diff --git a/hw/vfio/common.c b/hw/vfio/common.c index 2f3f66991a..3be6cecc63 100644 --- a/hw/vfio/common.c +++ b/hw/vfio/common.c @@ -1079,7 +1079,8 @@ static void vfio_listener_log_global_start(MemoryListener *listener) if (vfio_devices_all_device_dirty_tracking(container)) { ret = vfio_devices_dma_logging_start(container); } else { - ret = vfio_set_dirty_page_tracking(container, true); + ret = vfio_container_set_dirty_page_tracking(&container->bcontainer, + true); }

if (ret) { @@ -1097,7 +1098,8 @@ static void vfio_listener_log_global_stop(MemoryListener *listener) if (vfio_devices_all_device_dirty_tracking(container)) { vfio_devices_dma_logging_stop(container); } else { - ret = vfio_set_dirty_page_tracking(container, false); + ret = vfio_container_set_dirty_page_tracking(&container->bcontainer, + false); }

if (ret) { @@ -1166,7 +1168,8 @@ int vfio_get_dirty_bitmap(VFIOContainer *container, uint64_t iova, VFIODMARange *qrange; int ret;

- if (!container->dirty_pages_supported && !all_device_dirty_tracking) { + if (!container->bcontainer.dirty_pages_supported && + !all_device_dirty_tracking) { cpu_physical_memory_set_dirty_range(ram_addr, size, tcg_enabled() ? DIRTY_CLIENTS_ALL : DIRTY_CLIENTS_NOCODE); @@ -1187,7 +1190,8 @@ int vfio_get_dirty_bitmap(VFIOContainer *container, uint64_t iova, if (all_device_dirty_tracking) { ret = vfio_devices_query_dirty_bitmap(container, &vbmap, iova, size); } else { - ret = vfio_query_dirty_bitmap(container, &vbmap, iova, size); + ret = vfio_container_query_dirty_bitmap(&container->bcontainer, &vbmap, + iova, size); }

if (ret) { @@ -1480,7 +1484,7 @@ static void vfio_listener_log_clear(MemoryListener *listener, VFIOContainer *container = container_of(listener, VFIOContainer, listener);

if (vfio_listener_skipped_section(section) || - !container->dirty_pages_supported) { + !container->bcontainer.dirty_pages_supported) { return; }

diff --git a/hw/vfio/container-base.c b/hw/vfio/container-base.c index 3933391e0d..5d654ae172 100644 --- a/hw/vfio/container-base.c +++ b/hw/vfio/container-base.c @@ -31,11 +31,27 @@ int vfio_container_dma_unmap(VFIOContainerBase *bcontainer, return bcontainer->ops->dma_unmap(bcontainer, iova, size, iotlb); }

+int vfio_container_set_dirty_page_tracking(VFIOContainerBase *bcontainer, + bool start) +{ + g_assert(bcontainer->ops->set_dirty_page_tracking); + return bcontainer->ops->set_dirty_page_tracking(bcontainer, start); +} + +int vfio_container_query_dirty_bitmap(VFIOContainerBase *bcontainer, + VFIOBitmap *vbmap, + hwaddr iova, hwaddr size) +{ + g_assert(bcontainer->ops->query_dirty_bitmap); + return bcontainer->ops->query_dirty_bitmap(bcontainer, vbmap, iova, size); +} + void vfio_container_init(VFIOContainerBase *bcontainer, VFIOAddressSpace *space, const VFIOIOMMUOps *ops) { bcontainer->ops = ops; bcontainer->space = space; + bcontainer->dirty_pages_supported = false; QLIST_INIT(&bcontainer->giommu_list); }

diff --git a/hw/vfio/container.c b/hw/vfio/container.c index b7ab0d7323..cf373e42ef 100644 --- a/hw/vfio/container.c +++ b/hw/vfio/container.c @@ -157,7 +157,7 @@ static int vfio_legacy_dma_unmap(VFIOContainerBase *bcontainer, hwaddr iova,

if (iotlb && vfio_devices_all_running_and_mig_active(container)) { if (!vfio_devices_all_device_dirty_tracking(container) && - container->dirty_pages_supported) { + container->bcontainer.dirty_pages_supported) { return vfio_dma_unmap_bitmap(container, iova, size, iotlb); }

@@ -255,14 +255,17 @@ static int vfio_legacy_dma_map(VFIOContainerBase *bcontainer, hwaddr iova, return -errno; }

-int vfio_set_dirty_page_tracking(VFIOContainer *container, bool start) +static int vfio_legacy_set_dirty_page_tracking(VFIOContainerBase *bcontainer, + bool start) { + VFIOContainer *container = container_of(bcontainer, VFIOContainer, + bcontainer); int ret; struct vfio_iommu_type1_dirty_bitmap dirty = { .argsz = sizeof(dirty), };

- if (!container->dirty_pages_supported) { + if (!bcontainer->dirty_pages_supported) { return 0; }

@@ -282,9 +285,12 @@ 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(VFIOContainerBase *bcontainer, + VFIOBitmap *vbmap, + hwaddr iova, hwaddr size) { + VFIOContainer *container = container_of(bcontainer, VFIOContainer, + bcontainer); struct vfio_iommu_type1_dirty_bitmap *dbitmap; struct vfio_iommu_type1_dirty_bitmap_get *range; int ret; @@ -528,7 +534,7 @@ 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->bcontainer.dirty_pages_supported = true; container->max_dirty_bitmap_size = cap_mig->max_dirty_bitmap_size; container->dirty_pgsizes = cap_mig->pgsize_bitmap; } @@ -646,7 +652,6 @@ static int vfio_connect_container(VFIOGroup *group, AddressSpace *as, container = g_malloc0(sizeof(*container)); container->fd = fd; container->error = NULL; - container->dirty_pages_supported = false; container->dma_max_mappings = 0; container->iova_ranges = NULL; QLIST_INIT(&container->vrdl_list); @@ -1050,4 +1055,6 @@ void vfio_detach_device(VFIODevice *vbasedev) const VFIOIOMMUOps vfio_legacy_ops = { .dma_map = vfio_legacy_dma_map, .dma_unmap = vfio_legacy_dma_unmap, + .set_dirty_page_tracking = vfio_legacy_set_dirty_page_tracking, + .query_dirty_bitmap = vfio_legacy_query_dirty_bitmap, }; diff --git a/include/hw/vfio/vfio-common.h b/include/hw/vfio/vfio-common.h index 31c9df4b03..af0ef9042d 100644 --- a/include/hw/vfio/vfio-common.h +++ b/include/hw/vfio/vfio-common.h @@ -91,7 +91,6 @@ typedef struct VFIOContainer { unsigned iommu_type; Error *error; bool initialized; - bool dirty_pages_supported; bool dirty_log_manual_clear; uint64_t dirty_pgsizes; uint64_t max_dirty_bitmap_size; @@ -200,13 +199,9 @@ VFIOAddressSpace *vfio_get_address_space(AddressSpace *as); void vfio_put_address_space(VFIOAddressSpace *space); bool vfio_devices_all_running_and_saving(VFIOContainer *container);

-/* container->fd */ VFIODMARange *vfio_lookup_match_range(VFIOContainer *container, hwaddr start_addr, hwaddr size); void vfio_dma_range_init_dirty_bitmap(VFIODMARange *qrange); -int vfio_set_dirty_page_tracking(VFIOContainer *container, bool start); -int vfio_query_dirty_bitmap(VFIOContainer *container, VFIOBitmap *vbmap, - hwaddr iova, hwaddr size);

/* SPAPR specific */ int vfio_container_add_section_window(VFIOContainer *container, diff --git a/include/hw/vfio/vfio-container-base.h b/include/hw/vfio/vfio-container-base.h index c7cc6ec9c5..f244f003d0 100644 --- a/include/hw/vfio/vfio-container-base.h +++ b/include/hw/vfio/vfio-container-base.h @@ -36,6 +36,7 @@ typedef struct VFIOAddressSpace { typedef struct VFIOContainerBase { const VFIOIOMMUOps *ops; VFIOAddressSpace *space; + bool dirty_pages_supported; QLIST_HEAD(, VFIOGuestIOMMU) giommu_list; QLIST_ENTRY(VFIOContainerBase) next; } VFIOContainerBase; @@ -54,6 +55,11 @@ int vfio_container_dma_map(VFIOContainerBase *bcontainer, int vfio_container_dma_unmap(VFIOContainerBase *bcontainer, hwaddr iova, ram_addr_t size, IOMMUTLBEntry *iotlb); +int vfio_container_set_dirty_page_tracking(VFIOContainerBase *bcontainer, + bool start); +int vfio_container_query_dirty_bitmap(VFIOContainerBase *bcontainer, + VFIOBitmap *vbmap, + hwaddr iova, hwaddr size);

void vfio_container_init(VFIOContainerBase *bcontainer, VFIOAddressSpace *space,

From: Eric Auger eric.auger@redhat.com

In the prospect to get rid of VFIOContainer refs in common.c lets convert misc functions to use the base container object instead:

vfio_devices_all_dirty_tracking vfio_devices_all_device_dirty_tracking vfio_devices_all_running_and_mig_active vfio_devices_query_dirty_bitmap vfio_get_dirty_bitmap

Modify vfio_get_dirty_bitmap/vfio_listener_log_clear during backporting.

Signed-off-by: Eric Auger eric.auger@redhat.com Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Cédric Le Goater clg@redhat.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/vfio/common.c | 46 ++++++++++++++++------------------- hw/vfio/container.c | 6 ++--- hw/vfio/trace-events | 2 +- include/hw/vfio/vfio-common.h | 9 +++---- 4 files changed, 29 insertions(+), 34 deletions(-)

diff --git a/hw/vfio/common.c b/hw/vfio/common.c index b952d1c811..b663d0bcc0 100644 --- a/hw/vfio/common.c +++ b/hw/vfio/common.c @@ -177,9 +177,8 @@ bool vfio_device_state_is_precopy(VFIODevice *vbasedev) migration->device_state == VFIO_DEVICE_STATE_PRE_COPY_P2P; }

-static bool vfio_devices_all_dirty_tracking(VFIOContainer *container) +static bool vfio_devices_all_dirty_tracking(VFIOContainerBase *bcontainer) { - VFIOContainerBase *bcontainer = &container->bcontainer; VFIODevice *vbasedev; MigrationState *ms = migrate_get_current();

@@ -204,9 +203,8 @@ static bool vfio_devices_all_dirty_tracking(VFIOContainer *container) return true; }

-bool vfio_devices_all_device_dirty_tracking(VFIOContainer *container) +bool vfio_devices_all_device_dirty_tracking(VFIOContainerBase *bcontainer) { - VFIOContainerBase *bcontainer = &container->bcontainer; VFIODevice *vbasedev;

QLIST_FOREACH(vbasedev, &bcontainer->device_list, container_next) { @@ -222,9 +220,8 @@ bool vfio_devices_all_device_dirty_tracking(VFIOContainer *container) * Check if all VFIO devices are running and migration is active, which is * essentially equivalent to the migration being in pre-copy phase. */ -bool vfio_devices_all_running_and_mig_active(VFIOContainer *container) +bool vfio_devices_all_running_and_mig_active(VFIOContainerBase *bcontainer) { - VFIOContainerBase *bcontainer = &container->bcontainer; VFIODevice *vbasedev;

if (!migration_is_active(migrate_get_current())) { @@ -1082,7 +1079,7 @@ static void vfio_listener_log_global_start(MemoryListener *listener) VFIOContainer *container = container_of(listener, VFIOContainer, listener); int ret;

- if (vfio_devices_all_device_dirty_tracking(container)) { + if (vfio_devices_all_device_dirty_tracking(&container->bcontainer)) { ret = vfio_devices_dma_logging_start(container); } else { ret = vfio_container_set_dirty_page_tracking(&container->bcontainer, @@ -1101,7 +1098,7 @@ static void vfio_listener_log_global_stop(MemoryListener *listener) VFIOContainer *container = container_of(listener, VFIOContainer, listener); int ret = 0;

- if (vfio_devices_all_device_dirty_tracking(container)) { + if (vfio_devices_all_device_dirty_tracking(&container->bcontainer)) { vfio_devices_dma_logging_stop(container); } else { ret = vfio_container_set_dirty_page_tracking(&container->bcontainer, @@ -1141,11 +1138,10 @@ static int vfio_device_dma_logging_report(VFIODevice *vbasedev, hwaddr iova, return 0; }

-int vfio_devices_query_dirty_bitmap(VFIOContainer *container, +int vfio_devices_query_dirty_bitmap(VFIOContainerBase *bcontainer, VFIOBitmap *vbmap, hwaddr iova, hwaddr size) { - VFIOContainerBase *bcontainer = &container->bcontainer; VFIODevice *vbasedev; int ret;

@@ -1165,18 +1161,19 @@ int vfio_devices_query_dirty_bitmap(VFIOContainer *container, return 0; }

-int vfio_get_dirty_bitmap(VFIOContainer *container, uint64_t iova, +int vfio_get_dirty_bitmap(VFIOContainerBase *bcontainer, uint64_t iova, uint64_t size, ram_addr_t ram_addr) { bool all_device_dirty_tracking = - vfio_devices_all_device_dirty_tracking(container); + vfio_devices_all_device_dirty_tracking(bcontainer); + VFIOContainer *container = container_of(bcontainer, VFIOContainer, + bcontainer); uint64_t dirty_pages; VFIOBitmap vbmap; VFIODMARange *qrange; int ret;

- if (!container->bcontainer.dirty_pages_supported && - !all_device_dirty_tracking) { + if (!bcontainer->dirty_pages_supported && !all_device_dirty_tracking) { cpu_physical_memory_set_dirty_range(ram_addr, size, tcg_enabled() ? DIRTY_CLIENTS_ALL : DIRTY_CLIENTS_NOCODE); @@ -1195,10 +1192,9 @@ int vfio_get_dirty_bitmap(VFIOContainer *container, uint64_t iova, vbmap.bitmap = qrange->bitmap;

if (all_device_dirty_tracking) { - ret = vfio_devices_query_dirty_bitmap(container, &vbmap, iova, size); + ret = vfio_devices_query_dirty_bitmap(bcontainer, &vbmap, iova, size); } else { - ret = vfio_container_query_dirty_bitmap(&container->bcontainer, &vbmap, - iova, size); + ret = vfio_container_query_dirty_bitmap(bcontainer, &vbmap, iova, size); }

if (ret) { @@ -1208,8 +1204,7 @@ int vfio_get_dirty_bitmap(VFIOContainer *container, uint64_t iova, dirty_pages = cpu_physical_memory_set_dirty_lebitmap(vbmap.bitmap, ram_addr, vbmap.pages);

- trace_vfio_get_dirty_bitmap(container->fd, iova, size, vbmap.size, - ram_addr, dirty_pages); + trace_vfio_get_dirty_bitmap(iova, size, vbmap.size, ram_addr, dirty_pages); out: return ret; } @@ -1241,8 +1236,8 @@ static void vfio_iommu_map_dirty_notify(IOMMUNotifier *n, IOMMUTLBEntry *iotlb)

rcu_read_lock(); if (vfio_get_xlat_addr(iotlb, NULL, &translated_addr, NULL)) { - ret = vfio_get_dirty_bitmap(container, iova, iotlb->addr_mask + 1, - translated_addr); + ret = vfio_get_dirty_bitmap(&container->bcontainer, iova, + iotlb->addr_mask + 1, translated_addr); if (ret) { error_report("vfio_iommu_map_dirty_notify(%p, 0x%"HWADDR_PRIx", " "0x%"HWADDR_PRIx") = %d (%s)", @@ -1271,7 +1266,8 @@ static int vfio_ram_discard_get_dirty_bitmap(MemoryRegionSection *section, * Sync the whole mapped region (spanning multiple individual mappings) * in one go. */ - return vfio_get_dirty_bitmap(vrdl->container, iova, size, ram_addr); + return vfio_get_dirty_bitmap(&vrdl->container->bcontainer, iova, size, + ram_addr); }

static int vfio_sync_ram_discard_listener_dirty_bitmap(VFIOContainer *container, @@ -1340,7 +1336,7 @@ static int vfio_sync_dirty_bitmap(VFIOContainer *container, ram_addr = memory_region_get_ram_addr(section->mr) + section->offset_within_region;

- return vfio_get_dirty_bitmap(container, + return vfio_get_dirty_bitmap(&container->bcontainer, REAL_HOST_PAGE_ALIGN(section->offset_within_address_space), int128_get64(section->size), ram_addr); } @@ -1355,7 +1351,7 @@ static void vfio_listener_log_sync(MemoryListener *listener, return; }

- if (vfio_devices_all_dirty_tracking(container)) { + if (vfio_devices_all_dirty_tracking(&container->bcontainer)) { ret = vfio_sync_dirty_bitmap(container, section); if (ret) { error_report("vfio: Failed to sync dirty bitmap, err: %d (%s)", ret, @@ -1495,7 +1491,7 @@ static void vfio_listener_log_clear(MemoryListener *listener, return; }

- if (vfio_devices_all_dirty_tracking(container)) { + if (vfio_devices_all_dirty_tracking(&container->bcontainer)) { vfio_physical_log_clear(container, section); } } diff --git a/hw/vfio/container.c b/hw/vfio/container.c index 74d236ddee..9a542368ab 100644 --- a/hw/vfio/container.c +++ b/hw/vfio/container.c @@ -155,8 +155,8 @@ static int vfio_legacy_dma_unmap(VFIOContainerBase *bcontainer, 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) && + if (iotlb && vfio_devices_all_running_and_mig_active(bcontainer)) { + if (!vfio_devices_all_device_dirty_tracking(bcontainer) && container->bcontainer.dirty_pages_supported) { return vfio_dma_unmap_bitmap(container, iova, size, iotlb); } @@ -204,7 +204,7 @@ static int vfio_legacy_dma_unmap(VFIOContainerBase *bcontainer, 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; diff --git a/hw/vfio/trace-events b/hw/vfio/trace-events index 9f7fedee98..08a1f9dfa4 100644 --- a/hw/vfio/trace-events +++ b/hw/vfio/trace-events @@ -117,7 +117,7 @@ vfio_region_sparse_mmap_header(const char *name, int index, int nr_areas) "Devic vfio_region_sparse_mmap_entry(int i, unsigned long start, unsigned long end) "sparse entry %d [0x%lx - 0x%lx]" vfio_get_dev_region(const char *name, int index, uint32_t type, uint32_t subtype) "%s index %d, %08x/%08x" vfio_legacy_dma_unmap_overflow_workaround(void) "" -vfio_get_dirty_bitmap(int fd, uint64_t iova, uint64_t size, uint64_t bitmap_size, uint64_t start, uint64_t dirty_pages) "container fd=%d, iova=0x%"PRIx64" size= 0x%"PRIx64" bitmap_size=0x%"PRIx64" start=0x%"PRIx64" dirty_pages=%"PRIu64 +vfio_get_dirty_bitmap(uint64_t iova, uint64_t size, uint64_t bitmap_size, uint64_t start, uint64_t dirty_pages) "iova=0x%"PRIx64" size= 0x%"PRIx64" bitmap_size=0x%"PRIx64" start=0x%"PRIx64" dirty_pages=%"PRIu64 vfio_iommu_map_dirty_notify(uint64_t iova_start, uint64_t iova_end) "iommu dirty @ 0x%"PRIx64" - 0x%"PRIx64

# platform.c diff --git a/include/hw/vfio/vfio-common.h b/include/hw/vfio/vfio-common.h index e27854228c..0295ede7ba 100644 --- a/include/hw/vfio/vfio-common.h +++ b/include/hw/vfio/vfio-common.h @@ -196,7 +196,6 @@ typedef struct VFIODisplay {

VFIOAddressSpace *vfio_get_address_space(AddressSpace *as); void vfio_put_address_space(VFIOAddressSpace *space); -bool vfio_devices_all_running_and_saving(VFIOContainer *container);

VFIODMARange *vfio_lookup_match_range(VFIOContainer *container, hwaddr start_addr, hwaddr size); @@ -274,11 +273,11 @@ bool vfio_migration_realize(VFIODevice *vbasedev, Error **errp); void vfio_migration_exit(VFIODevice *vbasedev);

int vfio_bitmap_alloc(VFIOBitmap *vbmap, hwaddr size); -bool vfio_devices_all_running_and_mig_active(VFIOContainer *container); -bool vfio_devices_all_device_dirty_tracking(VFIOContainer *container); -int vfio_devices_query_dirty_bitmap(VFIOContainer *container, +bool vfio_devices_all_running_and_mig_active(VFIOContainerBase *bcontainer); +bool vfio_devices_all_device_dirty_tracking(VFIOContainerBase *bcontainer); +int vfio_devices_query_dirty_bitmap(VFIOContainerBase *bcontainer, VFIOBitmap *vbmap, hwaddr iova, hwaddr size); -int vfio_get_dirty_bitmap(VFIOContainer *container, uint64_t iova, +int vfio_get_dirty_bitmap(VFIOContainerBase *bcontainer, uint64_t iova, uint64_t size, ram_addr_t ram_addr); #endif /* HW_VFIO_VFIO_COMMON_H */

From: Zhenzhong Duan zhenzhong.duan@intel.com

No functional change intended.

Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Cédric Le Goater clg@redhat.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/vfio/common.c | 38 +++++++++++++-------------- hw/vfio/container-base.c | 1 + hw/vfio/container.c | 1 - include/hw/vfio/vfio-common.h | 11 -------- include/hw/vfio/vfio-container-base.h | 11 ++++++++ 5 files changed, 31 insertions(+), 31 deletions(-)

diff --git a/hw/vfio/common.c b/hw/vfio/common.c index fd6249c290..e9a19209ab 100644 --- a/hw/vfio/common.c +++ b/hw/vfio/common.c @@ -351,13 +351,13 @@ static void vfio_ram_discard_notify_discard(RamDiscardListener *rdl, { VFIORamDiscardListener *vrdl = container_of(rdl, VFIORamDiscardListener, listener); + VFIOContainerBase *bcontainer = vrdl->bcontainer; const hwaddr size = int128_get64(section->size); const hwaddr iova = section->offset_within_address_space; int ret;

/* Unmap with a single call. */ - ret = vfio_container_dma_unmap(&vrdl->container->bcontainer, - iova, size , NULL); + ret = vfio_container_dma_unmap(bcontainer, iova, size , NULL); if (ret) { error_report("%s: vfio_container_dma_unmap() failed: %s", __func__, strerror(-ret)); @@ -369,6 +369,7 @@ static int vfio_ram_discard_notify_populate(RamDiscardListener *rdl, { VFIORamDiscardListener *vrdl = container_of(rdl, VFIORamDiscardListener, listener); + VFIOContainerBase *bcontainer = vrdl->bcontainer; const hwaddr end = section->offset_within_region + int128_get64(section->size); hwaddr start, next, iova; @@ -387,8 +388,8 @@ static int vfio_ram_discard_notify_populate(RamDiscardListener *rdl, section->offset_within_address_space; vaddr = memory_region_get_ram_ptr(section->mr) + start;

- ret = vfio_container_dma_map(&vrdl->container->bcontainer, iova, - next - start, vaddr, section->readonly); + ret = vfio_container_dma_map(bcontainer, iova, next - start, + vaddr, section->readonly); if (ret) { /* Rollback */ vfio_ram_discard_notify_discard(rdl, section); @@ -398,10 +399,9 @@ static int vfio_ram_discard_notify_populate(RamDiscardListener *rdl, return 0; }

-static void vfio_register_ram_discard_listener(VFIOContainer *container, +static void vfio_register_ram_discard_listener(VFIOContainerBase *bcontainer, MemoryRegionSection *section) { - VFIOContainerBase *bcontainer = &container->bcontainer; RamDiscardManager *rdm = memory_region_get_ram_discard_manager(section->mr); VFIORamDiscardListener *vrdl;

@@ -412,7 +412,7 @@ static void vfio_register_ram_discard_listener(VFIOContainer *container, g_assert(QEMU_IS_ALIGNED(int128_get64(section->size), TARGET_PAGE_SIZE));

vrdl = g_new0(VFIORamDiscardListener, 1); - vrdl->container = container; + vrdl->bcontainer = bcontainer; vrdl->mr = section->mr; vrdl->offset_within_address_space = section->offset_within_address_space; vrdl->size = int128_get64(section->size); @@ -427,7 +427,7 @@ static void vfio_register_ram_discard_listener(VFIOContainer *container, vfio_ram_discard_notify_populate, vfio_ram_discard_notify_discard, true); ram_discard_manager_register_listener(rdm, &vrdl->listener, section); - QLIST_INSERT_HEAD(&container->vrdl_list, vrdl, next); + QLIST_INSERT_HEAD(&bcontainer->vrdl_list, vrdl, next);

/* * Sanity-check if we have a theoretically problematic setup where we could @@ -451,7 +451,7 @@ static void vfio_register_ram_discard_listener(VFIOContainer *container, } #endif

- QLIST_FOREACH(vrdl, &container->vrdl_list, next) { + QLIST_FOREACH(vrdl, &bcontainer->vrdl_list, next) { hwaddr start, end;

start = QEMU_ALIGN_DOWN(vrdl->offset_within_address_space, @@ -473,13 +473,13 @@ static void vfio_register_ram_discard_listener(VFIOContainer *container, } }

-static void vfio_unregister_ram_discard_listener(VFIOContainer *container, +static void vfio_unregister_ram_discard_listener(VFIOContainerBase *bcontainer, MemoryRegionSection *section) { RamDiscardManager *rdm = memory_region_get_ram_discard_manager(section->mr); VFIORamDiscardListener *vrdl = NULL;

- QLIST_FOREACH(vrdl, &container->vrdl_list, next) { + QLIST_FOREACH(vrdl, &bcontainer->vrdl_list, next) { if (vrdl->mr == section->mr && vrdl->offset_within_address_space == section->offset_within_address_space) { @@ -663,7 +663,7 @@ static void vfio_listener_region_add(MemoryListener *listener, * about changes. */ if (memory_region_has_ram_discard_manager(section->mr)) { - vfio_register_ram_discard_listener(container, section); + vfio_register_ram_discard_listener(bcontainer, section); return; }

@@ -781,7 +781,7 @@ static void vfio_listener_region_del(MemoryListener *listener, pgmask = (1ULL << ctz64(bcontainer->pgsizes)) - 1; try_unmap = !((iova & pgmask) || (int128_get64(llsize) & pgmask)); } else if (memory_region_has_ram_discard_manager(section->mr)) { - vfio_unregister_ram_discard_listener(container, section); + vfio_unregister_ram_discard_listener(bcontainer, section); /* Unregistering will trigger an unmap. */ try_unmap = false; } @@ -1267,17 +1267,17 @@ static int vfio_ram_discard_get_dirty_bitmap(MemoryRegionSection *section, * Sync the whole mapped region (spanning multiple individual mappings) * in one go. */ - return vfio_get_dirty_bitmap(&vrdl->container->bcontainer, iova, size, - ram_addr); + return vfio_get_dirty_bitmap(vrdl->bcontainer, iova, size, ram_addr); }

-static int vfio_sync_ram_discard_listener_dirty_bitmap(VFIOContainer *container, - MemoryRegionSection *section) +static int +vfio_sync_ram_discard_listener_dirty_bitmap(VFIOContainerBase *bcontainer, + MemoryRegionSection *section) { RamDiscardManager *rdm = memory_region_get_ram_discard_manager(section->mr); VFIORamDiscardListener *vrdl = NULL;

- QLIST_FOREACH(vrdl, &container->vrdl_list, next) { + QLIST_FOREACH(vrdl, &bcontainer->vrdl_list, next) { if (vrdl->mr == section->mr && vrdl->offset_within_address_space == section->offset_within_address_space) { @@ -1331,7 +1331,7 @@ static int vfio_sync_dirty_bitmap(VFIOContainer *container, } return 0; } else if (memory_region_has_ram_discard_manager(section->mr)) { - return vfio_sync_ram_discard_listener_dirty_bitmap(container, section); + return vfio_sync_ram_discard_listener_dirty_bitmap(bcontainer, section); }

ram_addr = memory_region_get_ram_addr(section->mr) + diff --git a/hw/vfio/container-base.c b/hw/vfio/container-base.c index dcce111349..584eee4ba1 100644 --- a/hw/vfio/container-base.c +++ b/hw/vfio/container-base.c @@ -54,6 +54,7 @@ void vfio_container_init(VFIOContainerBase *bcontainer, VFIOAddressSpace *space, bcontainer->dirty_pages_supported = false; bcontainer->dma_max_mappings = 0; QLIST_INIT(&bcontainer->giommu_list); + QLIST_INIT(&bcontainer->vrdl_list); }

void vfio_container_destroy(VFIOContainerBase *bcontainer) diff --git a/hw/vfio/container.c b/hw/vfio/container.c index 116a9e1e73..023f220c93 100644 --- a/hw/vfio/container.c +++ b/hw/vfio/container.c @@ -653,7 +653,6 @@ static int vfio_connect_container(VFIOGroup *group, AddressSpace *as, container->fd = fd; container->error = NULL; container->iova_ranges = NULL; - QLIST_INIT(&container->vrdl_list); QLIST_INIT(&container->dma_list); bcontainer = &container->bcontainer; vfio_container_init(bcontainer, space, &vfio_legacy_ops); diff --git a/include/hw/vfio/vfio-common.h b/include/hw/vfio/vfio-common.h index 3046287070..0174b767ca 100644 --- a/include/hw/vfio/vfio-common.h +++ b/include/hw/vfio/vfio-common.h @@ -96,21 +96,10 @@ typedef struct VFIOContainer { uint64_t max_dirty_bitmap_size; QLIST_HEAD(, VFIOHostDMAWindow) hostwin_list; QLIST_HEAD(, VFIOGroup) group_list; - QLIST_HEAD(, VFIORamDiscardListener) vrdl_list; QLIST_HEAD(, VFIODMARange) dma_list; GList *iova_ranges; } VFIOContainer;

-typedef struct VFIORamDiscardListener { - VFIOContainer *container; - MemoryRegion *mr; - hwaddr offset_within_address_space; - hwaddr size; - uint64_t granularity; - RamDiscardListener listener; - QLIST_ENTRY(VFIORamDiscardListener) next; -} VFIORamDiscardListener; - typedef struct VFIOHostDMAWindow { hwaddr min_iova; hwaddr max_iova; diff --git a/include/hw/vfio/vfio-container-base.h b/include/hw/vfio/vfio-container-base.h index 85ec7e1a56..8e05b5ac5a 100644 --- a/include/hw/vfio/vfio-container-base.h +++ b/include/hw/vfio/vfio-container-base.h @@ -40,6 +40,7 @@ typedef struct VFIOContainerBase { unsigned int dma_max_mappings; bool dirty_pages_supported; QLIST_HEAD(, VFIOGuestIOMMU) giommu_list; + QLIST_HEAD(, VFIORamDiscardListener) vrdl_list; QLIST_ENTRY(VFIOContainerBase) next; QLIST_HEAD(, VFIODevice) device_list; } VFIOContainerBase; @@ -52,6 +53,16 @@ typedef struct VFIOGuestIOMMU { QLIST_ENTRY(VFIOGuestIOMMU) giommu_next; } VFIOGuestIOMMU;

+typedef struct VFIORamDiscardListener { + VFIOContainerBase *bcontainer; + MemoryRegion *mr; + hwaddr offset_within_address_space; + hwaddr size; + uint64_t granularity; + RamDiscardListener listener; + QLIST_ENTRY(VFIORamDiscardListener) next; +} VFIORamDiscardListener; + int vfio_container_dma_map(VFIOContainerBase *bcontainer, hwaddr iova, ram_addr_t size, void *vaddr, bool readonly);

From: Eric Auger eric.auger@redhat.com

No functional change intended.

Signed-off-by: Eric Auger eric.auger@redhat.com Signed-off-by: Yi Liu yi.l.liu@intel.com Signed-off-by: Yi Sun yi.y.sun@linux.intel.com Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Cédric Le Goater clg@redhat.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/vfio/container.c | 9 +++++---- include/hw/vfio/vfio-common.h | 2 -- include/hw/vfio/vfio-container-base.h | 2 ++ 3 files changed, 7 insertions(+), 6 deletions(-)

diff --git a/hw/vfio/container.c b/hw/vfio/container.c index 50da1300dd..191597167a 100644 --- a/hw/vfio/container.c +++ b/hw/vfio/container.c @@ -66,6 +66,7 @@ static int vfio_dma_unmap_bitmap(VFIOContainer *container, hwaddr iova, ram_addr_t size, IOMMUTLBEntry *iotlb) { + VFIOContainerBase *bcontainer = &container->bcontainer; struct vfio_iommu_type1_dma_unmap *unmap; struct vfio_bitmap *bitmap; VFIOBitmap vbmap; @@ -93,7 +94,7 @@ 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; @@ -157,7 +158,7 @@ static int vfio_legacy_dma_unmap(VFIOContainerBase *bcontainer, hwaddr iova,

if (iotlb && vfio_devices_all_running_and_mig_active(bcontainer)) { if (!vfio_devices_all_device_dirty_tracking(bcontainer) && - container->bcontainer.dirty_pages_supported) { + bcontainer->dirty_pages_supported) { return vfio_dma_unmap_bitmap(container, iova, size, iotlb); }

@@ -536,8 +537,8 @@ static void vfio_get_iommu_info_migration(VFIOContainer *container, */ if (cap_mig->pgsize_bitmap & qemu_real_host_page_size()) { bcontainer->dirty_pages_supported = true; - container->max_dirty_bitmap_size = cap_mig->max_dirty_bitmap_size; - container->dirty_pgsizes = cap_mig->pgsize_bitmap; + bcontainer->max_dirty_bitmap_size = cap_mig->max_dirty_bitmap_size; + bcontainer->dirty_pgsizes = cap_mig->pgsize_bitmap; } }

diff --git a/include/hw/vfio/vfio-common.h b/include/hw/vfio/vfio-common.h index c23e7fb8ee..a8da41d27e 100644 --- a/include/hw/vfio/vfio-common.h +++ b/include/hw/vfio/vfio-common.h @@ -89,8 +89,6 @@ typedef struct VFIOContainer { MemoryListener prereg_listener; unsigned iommu_type; bool dirty_log_manual_clear; - uint64_t dirty_pgsizes; - uint64_t max_dirty_bitmap_size; QLIST_HEAD(, VFIOHostDMAWindow) hostwin_list; QLIST_HEAD(, VFIOGroup) group_list; QLIST_HEAD(, VFIODMARange) dma_list; diff --git a/include/hw/vfio/vfio-container-base.h b/include/hw/vfio/vfio-container-base.h index 95f8d319e0..80e4a993c5 100644 --- a/include/hw/vfio/vfio-container-base.h +++ b/include/hw/vfio/vfio-container-base.h @@ -39,6 +39,8 @@ typedef struct VFIOContainerBase { MemoryListener listener; Error *error; bool initialized; + uint64_t dirty_pgsizes; + uint64_t max_dirty_bitmap_size; unsigned long pgsizes; unsigned int dma_max_mappings; bool dirty_pages_supported;

From: Eric Auger eric.auger@redhat.com

No functional change intended.

Signed-off-by: Eric Auger eric.auger@redhat.com Signed-off-by: Yi Liu yi.l.liu@intel.com Signed-off-by: Yi Sun yi.y.sun@linux.intel.com Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Cédric Le Goater clg@redhat.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/vfio/common.c | 16 ++++++++++++++++ hw/vfio/container.c | 12 +++++------- 2 files changed, 21 insertions(+), 7 deletions(-)

diff --git a/hw/vfio/common.c b/hw/vfio/common.c index 9926454527..488aa43c9b 100644 --- a/hw/vfio/common.c +++ b/hw/vfio/common.c @@ -1644,3 +1644,19 @@ retry:

return info; } + +int vfio_attach_device(char *name, VFIODevice *vbasedev, + AddressSpace *as, Error **errp) +{ + const VFIOIOMMUOps *ops = &vfio_legacy_ops; + + return ops->attach_device(name, vbasedev, as, errp); +} + +void vfio_detach_device(VFIODevice *vbasedev) +{ + if (!vbasedev->bcontainer) { + return; + } + vbasedev->bcontainer->ops->detach_device(vbasedev); +} diff --git a/hw/vfio/container.c b/hw/vfio/container.c index 13d42aad0d..62af0f2bdd 100644 --- a/hw/vfio/container.c +++ b/hw/vfio/container.c @@ -986,8 +986,8 @@ 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; @@ -1027,14 +1027,10 @@ int vfio_attach_device(char *name, VFIODevice *vbasedev, return ret; }

-void vfio_detach_device(VFIODevice *vbasedev) +static void vfio_legacy_detach_device(VFIODevice *vbasedev) { VFIOGroup *group = vbasedev->group;

- if (!vbasedev->bcontainer) { - return; - } - QLIST_REMOVE(vbasedev, global_next); QLIST_REMOVE(vbasedev, container_next); vbasedev->bcontainer = NULL; @@ -1046,6 +1042,8 @@ void vfio_detach_device(VFIODevice *vbasedev) 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, };

From: Zhenzhong Duan zhenzhong.duan@intel.com

No functional change intended.

Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Cédric Le Goater <clg@redhat.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/vfio/common.c | 8 ++------ hw/vfio/container-base.c | 21 +++++++++++++++++++++ hw/vfio/spapr.c | 19 ++++++++++++++----- include/hw/vfio/vfio-common.h | 5 ----- include/hw/vfio/vfio-container-base.h | 5 +++++ 5 files changed, 42 insertions(+), 16 deletions(-)

diff --git a/hw/vfio/common.c b/hw/vfio/common.c index 488aa43c9b..679fee4321 100644 --- a/hw/vfio/common.c +++ b/hw/vfio/common.c @@ -571,8 +571,6 @@ static void vfio_listener_region_add(MemoryListener *listener, { VFIOContainerBase *bcontainer = container_of(listener, VFIOContainerBase, listener); - VFIOContainer *container = container_of(bcontainer, VFIOContainer, - bcontainer); hwaddr iova, end; Int128 llend, llsize; void *vaddr; @@ -595,7 +593,7 @@ static void vfio_listener_region_add(MemoryListener *listener, return; }

- if (vfio_container_add_section_window(container, section, &err)) { + if (vfio_container_add_section_window(bcontainer, section, &err)) { goto fail; }

@@ -738,8 +736,6 @@ static void vfio_listener_region_del(MemoryListener *listener, { VFIOContainerBase *bcontainer = container_of(listener, VFIOContainerBase, listener); - VFIOContainer *container = container_of(bcontainer, VFIOContainer, - bcontainer); hwaddr iova, end; Int128 llend, llsize; int ret; @@ -818,7 +814,7 @@ static void vfio_listener_region_del(MemoryListener *listener,

memory_region_unref(section->mr);

- vfio_container_del_section_window(container, section); + vfio_container_del_section_window(bcontainer, section); }

typedef struct VFIODirtyRanges { diff --git a/hw/vfio/container-base.c b/hw/vfio/container-base.c index 0177f43741..71f7274973 100644 --- a/hw/vfio/container-base.c +++ b/hw/vfio/container-base.c @@ -31,6 +31,27 @@ int vfio_container_dma_unmap(VFIOContainerBase *bcontainer, return bcontainer->ops->dma_unmap(bcontainer, iova, size, iotlb); }

+int vfio_container_add_section_window(VFIOContainerBase *bcontainer, + MemoryRegionSection *section, + Error **errp) +{ + if (!bcontainer->ops->add_window) { + return 0; + } + + return bcontainer->ops->add_window(bcontainer, section, errp); +} + +void vfio_container_del_section_window(VFIOContainerBase *bcontainer, + MemoryRegionSection *section) +{ + if (!bcontainer->ops->del_window) { + return; + } + + return bcontainer->ops->del_window(bcontainer, section); +} + int vfio_container_set_dirty_page_tracking(VFIOContainerBase *bcontainer, bool start) { diff --git a/hw/vfio/spapr.c b/hw/vfio/spapr.c index e1a6b35563..5be1911aad 100644 --- a/hw/vfio/spapr.c +++ b/hw/vfio/spapr.c @@ -319,10 +319,13 @@ static int vfio_spapr_create_window(VFIOContainer *container, return 0; }

-int vfio_container_add_section_window(VFIOContainer *container, - MemoryRegionSection *section, - Error **errp) +static int +vfio_spapr_container_add_section_window(VFIOContainerBase *bcontainer, + MemoryRegionSection *section, + Error **errp) { + VFIOContainer *container = container_of(bcontainer, VFIOContainer, + bcontainer); VFIOHostDMAWindow *hostwin; hwaddr pgsize = 0; int ret; @@ -407,9 +410,13 @@ int vfio_container_add_section_window(VFIOContainer *container, return 0; }

-void vfio_container_del_section_window(VFIOContainer *container, - MemoryRegionSection *section) +static void +vfio_spapr_container_del_section_window(VFIOContainerBase *bcontainer, + MemoryRegionSection *section) { + VFIOContainer *container = container_of(bcontainer, VFIOContainer, + bcontainer); + if (container->iommu_type != VFIO_SPAPR_TCE_v2_IOMMU) { return; } @@ -430,6 +437,8 @@ static VFIOIOMMUOps vfio_iommu_spapr_ops; static void setup_spapr_ops(VFIOContainerBase *bcontainer) { vfio_iommu_spapr_ops = *bcontainer->ops; + vfio_iommu_spapr_ops.add_window = vfio_spapr_container_add_section_window; + vfio_iommu_spapr_ops.del_window = vfio_spapr_container_del_section_window; bcontainer->ops = &vfio_iommu_spapr_ops; }

diff --git a/include/hw/vfio/vfio-common.h b/include/hw/vfio/vfio-common.h index 9a2e0ace72..c6b1260911 100644 --- a/include/hw/vfio/vfio-common.h +++ b/include/hw/vfio/vfio-common.h @@ -183,11 +183,6 @@ VFIODMARange *vfio_lookup_match_range(VFIOContainer *container, void vfio_dma_range_init_dirty_bitmap(VFIODMARange *qrange);

/* SPAPR specific */ -int vfio_container_add_section_window(VFIOContainer *container, - MemoryRegionSection *section, - Error **errp); -void vfio_container_del_section_window(VFIOContainer *container, - MemoryRegionSection *section); int vfio_spapr_container_init(VFIOContainer *container, Error **errp); void vfio_spapr_container_deinit(VFIOContainer *container);

diff --git a/include/hw/vfio/vfio-container-base.h b/include/hw/vfio/vfio-container-base.h index f62a14ac73..4b6f017c6f 100644 --- a/include/hw/vfio/vfio-container-base.h +++ b/include/hw/vfio/vfio-container-base.h @@ -75,6 +75,11 @@ int vfio_container_dma_map(VFIOContainerBase *bcontainer, int vfio_container_dma_unmap(VFIOContainerBase *bcontainer, hwaddr iova, ram_addr_t size, IOMMUTLBEntry *iotlb); +int vfio_container_add_section_window(VFIOContainerBase *bcontainer, + MemoryRegionSection *section, + Error **errp); +void vfio_container_del_section_window(VFIOContainerBase *bcontainer, + MemoryRegionSection *section); int vfio_container_set_dirty_page_tracking(VFIOContainerBase *bcontainer, bool start); int vfio_container_query_dirty_bitmap(VFIOContainerBase *bcontainer,

From: Zhenzhong Duan zhenzhong.duan@intel.com

No functional changes intended.

Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Cédric Le Goater clg@redhat.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/vfio/spapr.c | 36 +++++++++++++++++++---------------- include/hw/vfio/vfio-common.h | 1 - 2 files changed, 20 insertions(+), 17 deletions(-)

diff --git a/hw/vfio/spapr.c b/hw/vfio/spapr.c index 68c3dd6c75..5c6426e697 100644 --- a/hw/vfio/spapr.c +++ b/hw/vfio/spapr.c @@ -27,6 +27,7 @@ typedef struct VFIOSpaprContainer { VFIOContainer container; MemoryListener prereg_listener; + QLIST_HEAD(, VFIOHostDMAWindow) hostwin_list; } VFIOSpaprContainer;

static bool vfio_prereg_listener_skipped_section(MemoryRegionSection *section) @@ -154,12 +155,12 @@ static const MemoryListener vfio_prereg_listener = { .region_del = vfio_prereg_listener_region_del, };

-static void vfio_host_win_add(VFIOContainer *container, hwaddr min_iova, +static void vfio_host_win_add(VFIOSpaprContainer *scontainer, hwaddr min_iova, hwaddr max_iova, uint64_t iova_pgsizes) { VFIOHostDMAWindow *hostwin;

- QLIST_FOREACH(hostwin, &container->hostwin_list, hostwin_next) { + QLIST_FOREACH(hostwin, &scontainer->hostwin_list, hostwin_next) { if (ranges_overlap(hostwin->min_iova, hostwin->max_iova - hostwin->min_iova + 1, min_iova, @@ -173,15 +174,15 @@ static void vfio_host_win_add(VFIOContainer *container, hwaddr min_iova, hostwin->min_iova = min_iova; hostwin->max_iova = max_iova; hostwin->iova_pgsizes = iova_pgsizes; - QLIST_INSERT_HEAD(&container->hostwin_list, hostwin, hostwin_next); + QLIST_INSERT_HEAD(&scontainer->hostwin_list, hostwin, hostwin_next); }

-static int vfio_host_win_del(VFIOContainer *container, +static int vfio_host_win_del(VFIOSpaprContainer *scontainer, hwaddr min_iova, hwaddr max_iova) { VFIOHostDMAWindow *hostwin;

- QLIST_FOREACH(hostwin, &container->hostwin_list, hostwin_next) { + QLIST_FOREACH(hostwin, &scontainer->hostwin_list, hostwin_next) { if (hostwin->min_iova == min_iova && hostwin->max_iova == max_iova) { QLIST_REMOVE(hostwin, hostwin_next); g_free(hostwin); @@ -192,7 +193,7 @@ static int vfio_host_win_del(VFIOContainer *container, return -1; }

-static VFIOHostDMAWindow *vfio_find_hostwin(VFIOContainer *container, +static VFIOHostDMAWindow *vfio_find_hostwin(VFIOSpaprContainer *container, hwaddr iova, hwaddr end) { VFIOHostDMAWindow *hostwin; @@ -329,6 +330,8 @@ vfio_spapr_container_add_section_window(VFIOContainerBase *bcontainer, { VFIOContainer *container = container_of(bcontainer, VFIOContainer, bcontainer); + VFIOSpaprContainer *scontainer = container_of(container, VFIOSpaprContainer, + container); VFIOHostDMAWindow *hostwin; hwaddr pgsize = 0; int ret; @@ -344,7 +347,7 @@ vfio_spapr_container_add_section_window(VFIOContainerBase *bcontainer, iova = section->offset_within_address_space; end = iova + int128_get64(section->size) - 1;

- if (!vfio_find_hostwin(container, iova, end)) { + if (!vfio_find_hostwin(scontainer, iova, end)) { error_setg(errp, "Container %p can't map guest IOVA region" " 0x%"HWADDR_PRIx"..0x%"HWADDR_PRIx, container, iova, end); @@ -358,7 +361,7 @@ vfio_spapr_container_add_section_window(VFIOContainerBase *bcontainer, }

/* For now intersections are not allowed, we may relax this later */ - QLIST_FOREACH(hostwin, &container->hostwin_list, hostwin_next) { + QLIST_FOREACH(hostwin, &scontainer->hostwin_list, hostwin_next) { if (ranges_overlap(hostwin->min_iova, hostwin->max_iova - hostwin->min_iova + 1, section->offset_within_address_space, @@ -380,7 +383,7 @@ vfio_spapr_container_add_section_window(VFIOContainerBase *bcontainer, return ret; }

- vfio_host_win_add(container, section->offset_within_address_space, + vfio_host_win_add(scontainer, section->offset_within_address_space, section->offset_within_address_space + int128_get64(section->size) - 1, pgsize); #ifdef CONFIG_KVM @@ -419,6 +422,8 @@ vfio_spapr_container_del_section_window(VFIOContainerBase *bcontainer, { VFIOContainer *container = container_of(bcontainer, VFIOContainer, bcontainer); + VFIOSpaprContainer *scontainer = container_of(container, VFIOSpaprContainer, + container);

if (container->iommu_type != VFIO_SPAPR_TCE_v2_IOMMU) { return; @@ -426,7 +431,7 @@ vfio_spapr_container_del_section_window(VFIOContainerBase *bcontainer,

vfio_spapr_remove_window(container, section->offset_within_address_space); - if (vfio_host_win_del(container, + if (vfio_host_win_del(scontainer, section->offset_within_address_space, section->offset_within_address_space + int128_get64(section->size) - 1) < 0) { @@ -454,7 +459,7 @@ int vfio_spapr_container_init(VFIOContainer *container, Error **errp) bool v2 = container->iommu_type == VFIO_SPAPR_TCE_v2_IOMMU; int ret, fd = container->fd;

- QLIST_INIT(&container->hostwin_list); + QLIST_INIT(&scontainer->hostwin_list);

/* * The host kernel code implementing VFIO_IOMMU_DISABLE is called @@ -506,7 +511,7 @@ int vfio_spapr_container_init(VFIOContainer *container, Error **errp) } else { /* The default table uses 4K pages */ bcontainer->pgsizes = 0x1000; - vfio_host_win_add(container, info.dma32_window_start, + vfio_host_win_add(scontainer, info.dma32_window_start, info.dma32_window_start + info.dma32_window_size - 1, 0x1000); @@ -525,15 +530,14 @@ listener_unregister_exit:

void vfio_spapr_container_deinit(VFIOContainer *container) { + VFIOSpaprContainer *scontainer = container_of(container, VFIOSpaprContainer, + container); VFIOHostDMAWindow *hostwin, *next;

if (container->iommu_type == VFIO_SPAPR_TCE_v2_IOMMU) { - VFIOSpaprContainer *scontainer = container_of(container, - VFIOSpaprContainer, - container); memory_listener_unregister(&scontainer->prereg_listener); } - QLIST_FOREACH_SAFE(hostwin, &container->hostwin_list, hostwin_next, + QLIST_FOREACH_SAFE(hostwin, &scontainer->hostwin_list, hostwin_next, next) { QLIST_REMOVE(hostwin, hostwin_next); g_free(hostwin); diff --git a/include/hw/vfio/vfio-common.h b/include/hw/vfio/vfio-common.h index ba8abed75a..9e22acbfb6 100644 --- a/include/hw/vfio/vfio-common.h +++ b/include/hw/vfio/vfio-common.h @@ -88,7 +88,6 @@ typedef struct VFIOContainer { int fd; /* /dev/vfio/vfio, empowered by the attached groups */ unsigned iommu_type; bool dirty_log_manual_clear; - QLIST_HEAD(, VFIOHostDMAWindow) hostwin_list; QLIST_HEAD(, VFIOGroup) group_list; QLIST_HEAD(, VFIODMARange) dma_list; } VFIOContainer;

From: Yi Liu yi.l.liu@intel.com

/dev/vfio/devices/vfioX may not exist. In that case it is still possible to open /dev/char/$major:$minor instead. Add helper function to abstract the cdev open.

Suggested-by: Jason Gunthorpe jgg@nvidia.com Signed-off-by: Yi Liu yi.l.liu@intel.com Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Cédric Le Goater clg@redhat.com Reviewed-by: Eric Auger eric.auger@redhat.com Tested-by: Eric Auger eric.auger@redhat.com Tested-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- MAINTAINERS | 2 + include/qemu/chardev_open.h | 16 ++++++++ util/chardev_open.c | 81 +++++++++++++++++++++++++++++++++++++ util/meson.build | 1 + 4 files changed, 100 insertions(+) create mode 100644 include/qemu/chardev_open.h create mode 100644 util/chardev_open.c

diff --git a/MAINTAINERS b/MAINTAINERS index a5a446914a..ca70bb4e64 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -2174,6 +2174,8 @@ M: Zhenzhong Duan zhenzhong.duan@intel.com S: Supported F: backends/iommufd.c F: include/sysemu/iommufd.h +F: include/qemu/chardev_open.h +F: util/chardev_open.c

vhost M: Michael S. Tsirkin mst@redhat.com diff --git a/include/qemu/chardev_open.h b/include/qemu/chardev_open.h new file mode 100644 index 0000000000..64e8fcfdcb --- /dev/null +++ b/include/qemu/chardev_open.h @@ -0,0 +1,16 @@ +/* + * QEMU Chardev Helper + * + * Copyright (C) 2023 Intel Corporation. + * + * Authors: Yi Liu yi.l.liu@intel.com + * + * This work is licensed under the terms of the GNU GPL, version 2. See + * the COPYING file in the top-level directory. + */ + +#ifndef QEMU_CHARDEV_OPEN_H +#define QEMU_CHARDEV_OPEN_H + +int open_cdev(const char *devpath, dev_t cdev); +#endif diff --git a/util/chardev_open.c b/util/chardev_open.c new file mode 100644 index 0000000000..f776429788 --- /dev/null +++ b/util/chardev_open.c @@ -0,0 +1,81 @@ +/* + * Copyright (c) 2019, Mellanox Technologies. All rights reserved. + * Copyright (C) 2023 Intel Corporation. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + * Authors: Yi Liu yi.l.liu@intel.com + * + * Copied from + * https://github.com/linux-rdma/rdma-core/blob/master/util/open_cdev.c + * + */ + +#include "qemu/osdep.h" +#include "qemu/chardev_open.h" + +static int open_cdev_internal(const char *path, dev_t cdev) +{ + struct stat st; + int fd; + + fd = qemu_open_old(path, O_RDWR); + if (fd == -1) { + return -1; + } + if (fstat(fd, &st) || !S_ISCHR(st.st_mode) || + (cdev != 0 && st.st_rdev != cdev)) { + close(fd); + return -1; + } + return fd; +} + +static int open_cdev_robust(dev_t cdev) +{ + g_autofree char *devpath = NULL; + + /* + * This assumes that udev is being used and is creating the /dev/char/ + * symlinks. + */ + devpath = g_strdup_printf("/dev/char/%u:%u", major(cdev), minor(cdev)); + return open_cdev_internal(devpath, cdev); +} + +int open_cdev(const char *devpath, dev_t cdev) +{ + int fd; + + fd = open_cdev_internal(devpath, cdev); + if (fd == -1 && cdev != 0) { + return open_cdev_robust(cdev); + } + return fd; +} diff --git a/util/meson.build b/util/meson.build index c2322ef6e7..174c133368 100644 --- a/util/meson.build +++ b/util/meson.build @@ -108,6 +108,7 @@ if have_block util_ss.add(files('filemonitor-stub.c')) endif util_ss.add(when: 'CONFIG_LINUX', if_true: files('vfio-helpers.c')) + util_ss.add(when: 'CONFIG_LINUX', if_true: files('chardev_open.c')) endif

if cpu == 'aarch64'

From: Yi Liu yi.l.liu@intel.com

The iommufd backend is implemented based on the new /dev/iommu user API. This backend obviously depends on CONFIG_IOMMUFD.

So far, the iommufd backend doesn't support dirty page sync yet.

Co-authored-by: Eric Auger eric.auger@redhat.com Signed-off-by: Yi Liu yi.l.liu@intel.com Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Cédric Le Goater clg@redhat.com Tested-by: Eric Auger eric.auger@redhat.com Tested-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/vfio/common.c | 6 + hw/vfio/iommufd.c | 422 ++++++++++++++++++++++++++++++++++ hw/vfio/meson.build | 3 + hw/vfio/trace-events | 10 + include/hw/vfio/vfio-common.h | 11 + 5 files changed, 452 insertions(+) create mode 100644 hw/vfio/iommufd.c

diff --git a/hw/vfio/common.c b/hw/vfio/common.c index f6c2029aec..0e900c6746 100644 --- a/hw/vfio/common.c +++ b/hw/vfio/common.c @@ -19,6 +19,7 @@ */

#include "qemu/osdep.h" +#include CONFIG_DEVICES /* CONFIG_IOMMUFD */ #include <sys/ioctl.h> #ifdef CONFIG_KVM #include <linux/kvm.h> @@ -1649,6 +1650,11 @@ int vfio_attach_device(char *name, VFIODevice *vbasedev, { const VFIOIOMMUOps *ops = &vfio_legacy_ops;

+#ifdef CONFIG_IOMMUFD + if (vbasedev->iommufd) { + ops = &vfio_iommufd_ops; + } +#endif return ops->attach_device(name, vbasedev, as, errp); }

diff --git a/hw/vfio/iommufd.c b/hw/vfio/iommufd.c new file mode 100644 index 0000000000..6d31aeac7b --- /dev/null +++ b/hw/vfio/iommufd.c @@ -0,0 +1,422 @@ +/* + * iommufd container backend + * + * Copyright (C) 2023 Intel Corporation. + * Copyright Red Hat, Inc. 2023 + * + * Authors: Yi Liu yi.l.liu@intel.com + * Eric Auger eric.auger@redhat.com + * + * SPDX-License-Identifier: GPL-2.0-or-later + */ + +#include "qemu/osdep.h" +#include <sys/ioctl.h> +#include <linux/vfio.h> +#include <linux/iommufd.h> + +#include "hw/vfio/vfio-common.h" +#include "qemu/error-report.h" +#include "trace.h" +#include "qapi/error.h" +#include "sysemu/iommufd.h" +#include "hw/qdev-core.h" +#include "sysemu/reset.h" +#include "qemu/cutils.h" +#include "qemu/chardev_open.h" + +static int iommufd_cdev_map(VFIOContainerBase *bcontainer, hwaddr iova, + ram_addr_t size, void *vaddr, bool readonly) +{ + VFIOIOMMUFDContainer *container = + container_of(bcontainer, VFIOIOMMUFDContainer, bcontainer); + + return iommufd_backend_map_dma(container->be, + container->ioas_id, + iova, size, vaddr, readonly); +} + +static int iommufd_cdev_unmap(VFIOContainerBase *bcontainer, + hwaddr iova, ram_addr_t size, + IOMMUTLBEntry *iotlb) +{ + VFIOIOMMUFDContainer *container = + container_of(bcontainer, VFIOIOMMUFDContainer, bcontainer); + + /* TODO: Handle dma_unmap_bitmap with iotlb args (migration) */ + return iommufd_backend_unmap_dma(container->be, + container->ioas_id, iova, size); +} + +static int iommufd_cdev_kvm_device_add(VFIODevice *vbasedev, Error **errp) +{ + return vfio_kvm_device_add_fd(vbasedev->fd, errp); +} + +static void iommufd_cdev_kvm_device_del(VFIODevice *vbasedev) +{ + Error *err = NULL; + + if (vfio_kvm_device_del_fd(vbasedev->fd, &err)) { + error_report_err(err); + } +} + +static int iommufd_cdev_connect_and_bind(VFIODevice *vbasedev, Error **errp) +{ + IOMMUFDBackend *iommufd = vbasedev->iommufd; + struct vfio_device_bind_iommufd bind = { + .argsz = sizeof(bind), + .flags = 0, + }; + int ret; + + ret = iommufd_backend_connect(iommufd, errp); + if (ret) { + return ret; + } + + /* + * Add device to kvm-vfio to be prepared for the tracking + * in KVM. Especially for some emulated devices, it requires + * to have kvm information in the device open. + */ + ret = iommufd_cdev_kvm_device_add(vbasedev, errp); + if (ret) { + goto err_kvm_device_add; + } + + /* Bind device to iommufd */ + bind.iommufd = iommufd->fd; + ret = ioctl(vbasedev->fd, VFIO_DEVICE_BIND_IOMMUFD, &bind); + if (ret) { + error_setg_errno(errp, errno, "error bind device fd=%d to iommufd=%d", + vbasedev->fd, bind.iommufd); + goto err_bind; + } + + vbasedev->devid = bind.out_devid; + trace_iommufd_cdev_connect_and_bind(bind.iommufd, vbasedev->name, + vbasedev->fd, vbasedev->devid); + return ret; +err_bind: + iommufd_cdev_kvm_device_del(vbasedev); +err_kvm_device_add: + iommufd_backend_disconnect(iommufd); + return ret; +} + +static void iommufd_cdev_unbind_and_disconnect(VFIODevice *vbasedev) +{ + /* Unbind is automatically conducted when device fd is closed */ + iommufd_cdev_kvm_device_del(vbasedev); + iommufd_backend_disconnect(vbasedev->iommufd); +} + +static int iommufd_cdev_getfd(const char *sysfs_path, Error **errp) +{ + long int ret = -ENOTTY; + char *path, *vfio_dev_path = NULL, *vfio_path = NULL; + DIR *dir = NULL; + struct dirent *dent; + gchar *contents; + struct stat st; + gsize length; + int major, minor; + dev_t vfio_devt; + + path = g_strdup_printf("%s/vfio-dev", sysfs_path); + if (stat(path, &st) < 0) { + error_setg_errno(errp, errno, "no such host device"); + goto out_free_path; + } + + dir = opendir(path); + if (!dir) { + error_setg_errno(errp, errno, "couldn't open directory %s", path); + goto out_free_path; + } + + while ((dent = readdir(dir))) { + if (!strncmp(dent->d_name, "vfio", 4)) { + vfio_dev_path = g_strdup_printf("%s/%s/dev", path, dent->d_name); + break; + } + } + + if (!vfio_dev_path) { + error_setg(errp, "failed to find vfio-dev/vfioX/dev"); + goto out_close_dir; + } + + if (!g_file_get_contents(vfio_dev_path, &contents, &length, NULL)) { + error_setg(errp, "failed to load "%s"", vfio_dev_path); + goto out_free_dev_path; + } + + if (sscanf(contents, "%d:%d", &major, &minor) != 2) { + error_setg(errp, "failed to get major:minor for "%s"", vfio_dev_path); + goto out_free_dev_path; + } + g_free(contents); + vfio_devt = makedev(major, minor); + + vfio_path = g_strdup_printf("/dev/vfio/devices/%s", dent->d_name); + ret = open_cdev(vfio_path, vfio_devt); + if (ret < 0) { + error_setg(errp, "Failed to open %s", vfio_path); + } + + trace_iommufd_cdev_getfd(vfio_path, ret); + g_free(vfio_path); + +out_free_dev_path: + g_free(vfio_dev_path); +out_close_dir: + closedir(dir); +out_free_path: + if (*errp) { + error_prepend(errp, VFIO_MSG_PREFIX, path); + } + g_free(path); + + return ret; +} + +static int iommufd_cdev_attach_ioas_hwpt(VFIODevice *vbasedev, uint32_t id, + Error **errp) +{ + int ret, iommufd = vbasedev->iommufd->fd; + struct vfio_device_attach_iommufd_pt attach_data = { + .argsz = sizeof(attach_data), + .flags = 0, + .pt_id = id, + }; + + /* Attach device to an IOAS or hwpt within iommufd */ + ret = ioctl(vbasedev->fd, VFIO_DEVICE_ATTACH_IOMMUFD_PT, &attach_data); + if (ret) { + error_setg_errno(errp, errno, + "[iommufd=%d] error attach %s (%d) to id=%d", + iommufd, vbasedev->name, vbasedev->fd, id); + } else { + trace_iommufd_cdev_attach_ioas_hwpt(iommufd, vbasedev->name, + vbasedev->fd, id); + } + return ret; +} + +static int iommufd_cdev_detach_ioas_hwpt(VFIODevice *vbasedev, Error **errp) +{ + int ret, iommufd = vbasedev->iommufd->fd; + struct vfio_device_detach_iommufd_pt detach_data = { + .argsz = sizeof(detach_data), + .flags = 0, + }; + + ret = ioctl(vbasedev->fd, VFIO_DEVICE_DETACH_IOMMUFD_PT, &detach_data); + if (ret) { + error_setg_errno(errp, errno, "detach %s failed", vbasedev->name); + } else { + trace_iommufd_cdev_detach_ioas_hwpt(iommufd, vbasedev->name); + } + return ret; +} + +static int iommufd_cdev_attach_container(VFIODevice *vbasedev, + VFIOIOMMUFDContainer *container, + Error **errp) +{ + return iommufd_cdev_attach_ioas_hwpt(vbasedev, container->ioas_id, errp); +} + +static void iommufd_cdev_detach_container(VFIODevice *vbasedev, + VFIOIOMMUFDContainer *container) +{ + Error *err = NULL; + + if (iommufd_cdev_detach_ioas_hwpt(vbasedev, &err)) { + error_report_err(err); + } +} + +static void iommufd_cdev_container_destroy(VFIOIOMMUFDContainer *container) +{ + VFIOContainerBase *bcontainer = &container->bcontainer; + + if (!QLIST_EMPTY(&bcontainer->device_list)) { + return; + } + memory_listener_unregister(&bcontainer->listener); + vfio_container_destroy(bcontainer); + iommufd_backend_free_id(container->be, container->ioas_id); + g_free(container); +} + +static int iommufd_cdev_ram_block_discard_disable(bool state) +{ + /* + * We support coordinated discarding of RAM via the RamDiscardManager. + */ + return ram_block_uncoordinated_discard_disable(state); +} + +static int iommufd_cdev_attach(const char *name, VFIODevice *vbasedev, + AddressSpace *as, Error **errp) +{ + VFIOContainerBase *bcontainer; + VFIOIOMMUFDContainer *container; + VFIOAddressSpace *space; + struct vfio_device_info dev_info = { .argsz = sizeof(dev_info) }; + int ret, devfd; + uint32_t ioas_id; + Error *err = NULL; + + devfd = iommufd_cdev_getfd(vbasedev->sysfsdev, errp); + if (devfd < 0) { + return devfd; + } + vbasedev->fd = devfd; + + ret = iommufd_cdev_connect_and_bind(vbasedev, errp); + if (ret) { + goto err_connect_bind; + } + + space = vfio_get_address_space(as); + + /* try to attach to an existing container in this space */ + QLIST_FOREACH(bcontainer, &space->containers, next) { + container = container_of(bcontainer, VFIOIOMMUFDContainer, bcontainer); + if (bcontainer->ops != &vfio_iommufd_ops || + vbasedev->iommufd != container->be) { + continue; + } + if (iommufd_cdev_attach_container(vbasedev, container, &err)) { + const char *msg = error_get_pretty(err); + + trace_iommufd_cdev_fail_attach_existing_container(msg); + error_free(err); + err = NULL; + } else { + ret = iommufd_cdev_ram_block_discard_disable(true); + if (ret) { + error_setg(errp, + "Cannot set discarding of RAM broken (%d)", ret); + goto err_discard_disable; + } + goto found_container; + } + } + + /* Need to allocate a new dedicated container */ + ret = iommufd_backend_alloc_ioas(vbasedev->iommufd, &ioas_id, errp); + if (ret < 0) { + goto err_alloc_ioas; + } + + trace_iommufd_cdev_alloc_ioas(vbasedev->iommufd->fd, ioas_id); + + container = g_malloc0(sizeof(*container)); + container->be = vbasedev->iommufd; + container->ioas_id = ioas_id; + + bcontainer = &container->bcontainer; + vfio_container_init(bcontainer, space, &vfio_iommufd_ops); + QLIST_INSERT_HEAD(&space->containers, bcontainer, next); + + ret = iommufd_cdev_attach_container(vbasedev, container, errp); + if (ret) { + goto err_attach_container; + } + + ret = iommufd_cdev_ram_block_discard_disable(true); + if (ret) { + goto err_discard_disable; + } + + bcontainer->pgsizes = qemu_real_host_page_size(); + + bcontainer->listener = vfio_memory_listener; + memory_listener_register(&bcontainer->listener, bcontainer->space->as); + + if (bcontainer->error) { + ret = -1; + error_propagate_prepend(errp, bcontainer->error, + "memory listener initialization failed: "); + goto err_listener_register; + } + + bcontainer->initialized = true; + +found_container: + ret = ioctl(devfd, VFIO_DEVICE_GET_INFO, &dev_info); + if (ret) { + error_setg_errno(errp, errno, "error getting device info"); + goto err_listener_register; + } + + /* + * TODO: examine RAM_BLOCK_DISCARD stuff, should we do group level + * for discarding incompatibility check as well? + */ + if (vbasedev->ram_block_discard_allowed) { + iommufd_cdev_ram_block_discard_disable(false); + } + + vbasedev->group = 0; + vbasedev->num_irqs = dev_info.num_irqs; + vbasedev->num_regions = dev_info.num_regions; + vbasedev->flags = dev_info.flags; + vbasedev->reset_works = !!(dev_info.flags & VFIO_DEVICE_FLAGS_RESET); + vbasedev->bcontainer = bcontainer; + QLIST_INSERT_HEAD(&bcontainer->device_list, vbasedev, container_next); + QLIST_INSERT_HEAD(&vfio_device_list, vbasedev, global_next); + + trace_iommufd_cdev_device_info(vbasedev->name, devfd, vbasedev->num_irqs, + vbasedev->num_regions, vbasedev->flags); + return 0; + +err_listener_register: + iommufd_cdev_ram_block_discard_disable(false); +err_discard_disable: + iommufd_cdev_detach_container(vbasedev, container); +err_attach_container: + iommufd_cdev_container_destroy(container); +err_alloc_ioas: + vfio_put_address_space(space); + iommufd_cdev_unbind_and_disconnect(vbasedev); +err_connect_bind: + close(vbasedev->fd); + return ret; +} + +static void iommufd_cdev_detach(VFIODevice *vbasedev) +{ + VFIOContainerBase *bcontainer = vbasedev->bcontainer; + VFIOAddressSpace *space = bcontainer->space; + VFIOIOMMUFDContainer *container = container_of(bcontainer, + VFIOIOMMUFDContainer, + bcontainer); + QLIST_REMOVE(vbasedev, global_next); + QLIST_REMOVE(vbasedev, container_next); + vbasedev->bcontainer = NULL; + + if (!vbasedev->ram_block_discard_allowed) { + iommufd_cdev_ram_block_discard_disable(false); + } + + iommufd_cdev_detach_container(vbasedev, container); + iommufd_cdev_container_destroy(container); + vfio_put_address_space(space); + + iommufd_cdev_unbind_and_disconnect(vbasedev); + close(vbasedev->fd); +} + +const VFIOIOMMUOps vfio_iommufd_ops = { + .dma_map = iommufd_cdev_map, + .dma_unmap = iommufd_cdev_unmap, + .attach_device = iommufd_cdev_attach, + .detach_device = iommufd_cdev_detach, +}; diff --git a/hw/vfio/meson.build b/hw/vfio/meson.build index 32a6933280..bd5cc4ca79 100644 --- a/hw/vfio/meson.build +++ b/hw/vfio/meson.build @@ -7,6 +7,9 @@ vfio_ss.add(files( 'spapr.c', 'migration.c', )) +vfio_ss.add(when: 'CONFIG_IOMMUFD', if_true: files( + 'iommufd.c', +)) vfio_ss.add(when: 'CONFIG_VFIO_PCI', if_true: files( 'display.c', 'pci-quirks.c', diff --git a/hw/vfio/trace-events b/hw/vfio/trace-events index 08a1f9dfa4..3340c93af0 100644 --- a/hw/vfio/trace-events +++ b/hw/vfio/trace-events @@ -164,3 +164,13 @@ vfio_state_pending_estimate(const char *name, uint64_t precopy, uint64_t postcop vfio_state_pending_exact(const char *name, uint64_t precopy, uint64_t postcopy, uint64_t stopcopy_size, uint64_t precopy_init_size, uint64_t precopy_dirty_size) " (%s) precopy 0x%"PRIx64" postcopy 0x%"PRIx64" stopcopy size 0x%"PRIx64" precopy initial size 0x%"PRIx64" precopy dirty size 0x%"PRIx64 vfio_vmstate_change(const char *name, int running, const char *reason, const char *dev_state) " (%s) running %d reason %s device state %s" vfio_vmstate_change_prepare(const char *name, int running, const char *reason, const char *dev_state) " (%s) running %d reason %s device state %s" + +#iommufd.c + +iommufd_cdev_connect_and_bind(int iommufd, const char *name, int devfd, int devid) " [iommufd=%d] Successfully bound device %s (fd=%d): output devid=%d" +iommufd_cdev_getfd(const char *dev, int devfd) " %s (fd=%d)" +iommufd_cdev_attach_ioas_hwpt(int iommufd, const char *name, int devfd, int id) " [iommufd=%d] Successfully attached device %s (%d) to id=%d" +iommufd_cdev_detach_ioas_hwpt(int iommufd, const char *name) " [iommufd=%d] Successfully detached %s" +iommufd_cdev_fail_attach_existing_container(const char *msg) " %s" +iommufd_cdev_alloc_ioas(int iommufd, int ioas_id) " [iommufd=%d] new IOMMUFD container with ioasid=%d" +iommufd_cdev_device_info(char *name, int devfd, int num_irqs, int num_regions, int flags) " %s (%d) num_irqs=%d num_regions=%d flags=%d" diff --git a/include/hw/vfio/vfio-common.h b/include/hw/vfio/vfio-common.h index 9e22acbfb6..9b9fd7b461 100644 --- a/include/hw/vfio/vfio-common.h +++ b/include/hw/vfio/vfio-common.h @@ -99,6 +99,14 @@ typedef struct VFIOHostDMAWindow { QLIST_ENTRY(VFIOHostDMAWindow) hostwin_next; } VFIOHostDMAWindow;

+typedef struct IOMMUFDBackend IOMMUFDBackend; + +typedef struct VFIOIOMMUFDContainer { + VFIOContainerBase bcontainer; + IOMMUFDBackend *be; + uint32_t ioas_id; +} VFIOIOMMUFDContainer; + typedef struct VFIODeviceOps VFIODeviceOps;

typedef struct VFIODevice { @@ -126,6 +134,8 @@ typedef struct VFIODevice { OnOffAuto pre_copy_dirty_page_tracking; bool dirty_pages_supported; bool dirty_tracking; + int devid; + IOMMUFDBackend *iommufd; } VFIODevice;

struct VFIODeviceOps { @@ -215,6 +225,7 @@ typedef QLIST_HEAD(VFIODeviceList, VFIODevice) VFIODeviceList; extern VFIOGroupList vfio_group_list; extern VFIODeviceList vfio_device_list; extern const VFIOIOMMUOps vfio_legacy_ops; +extern const VFIOIOMMUOps vfio_iommufd_ops; extern const MemoryListener vfio_memory_listener; extern int vfio_kvm_device_fd;

From: Zhenzhong Duan zhenzhong.duan@intel.com

Some vIOMMU such as virtio-iommu use IOVA ranges from host side to setup reserved ranges for passthrough device, so that guest will not use an IOVA range beyond host support.

Use an uAPI of IOMMUFD to get IOVA ranges of host side and pass to vIOMMU just like the legacy backend, if this fails, fallback to 64bit IOVA range.

Also use out_iova_alignment returned from uAPI as pgsizes instead of qemu_real_host_page_size() as a fallback.

Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Cédric Le Goater clg@redhat.com Reviewed-by: Eric Auger eric.auger@redhat.com Tested-by: Eric Auger eric.auger@redhat.com Tested-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/vfio/iommufd.c | 56 ++++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 55 insertions(+), 1 deletion(-)

diff --git a/hw/vfio/iommufd.c b/hw/vfio/iommufd.c index 6d31aeac7b..01b448e840 100644 --- a/hw/vfio/iommufd.c +++ b/hw/vfio/iommufd.c @@ -261,6 +261,53 @@ static int iommufd_cdev_ram_block_discard_disable(bool state) return ram_block_uncoordinated_discard_disable(state); }

+static int iommufd_cdev_get_info_iova_range(VFIOIOMMUFDContainer *container, + uint32_t ioas_id, Error **errp) +{ + VFIOContainerBase *bcontainer = &container->bcontainer; + struct iommu_ioas_iova_ranges *info; + struct iommu_iova_range *iova_ranges; + int ret, sz, fd = container->be->fd; + + info = g_malloc0(sizeof(*info)); + info->size = sizeof(*info); + info->ioas_id = ioas_id; + + ret = ioctl(fd, IOMMU_IOAS_IOVA_RANGES, info); + if (ret && errno != EMSGSIZE) { + goto error; + } + + sz = info->num_iovas * sizeof(struct iommu_iova_range); + info = g_realloc(info, sizeof(*info) + sz); + info->allowed_iovas = (uintptr_t)(info + 1); + + ret = ioctl(fd, IOMMU_IOAS_IOVA_RANGES, info); + if (ret) { + goto error; + } + + iova_ranges = (struct iommu_iova_range *)(uintptr_t)info->allowed_iovas; + + for (int i = 0; i < info->num_iovas; i++) { + Range *range = g_new(Range, 1); + + range_set_bounds(range, iova_ranges[i].start, iova_ranges[i].last); + bcontainer->iova_ranges = + range_list_insert(bcontainer->iova_ranges, range); + } + bcontainer->pgsizes = info->out_iova_alignment; + + g_free(info); + return 0; + +error: + ret = -errno; + g_free(info); + error_setg_errno(errp, errno, "Cannot get IOVA ranges"); + return ret; +} + static int iommufd_cdev_attach(const char *name, VFIODevice *vbasedev, AddressSpace *as, Error **errp) { @@ -335,7 +382,14 @@ static int iommufd_cdev_attach(const char *name, VFIODevice *vbasedev, goto err_discard_disable; }

- bcontainer->pgsizes = qemu_real_host_page_size(); + ret = iommufd_cdev_get_info_iova_range(container, ioas_id, &err); + if (ret) { + error_append_hint(&err, + "Fallback to default 64bit IOVA range and 4K page size\n"); + warn_report_err(err); + err = NULL; + bcontainer->pgsizes = qemu_real_host_page_size(); + }

bcontainer->listener = vfio_memory_listener; memory_listener_register(&bcontainer->listener, bcontainer->space->as);

From: Zhenzhong Duan zhenzhong.duan@intel.com

Legacy vfio pci and iommufd cdev have different process to hot reset vfio device, expand current code to abstract out pci_hot_reset callback for legacy vfio, this same interface will also be used by iommufd cdev vfio device.

Rename vfio_pci_hot_reset to vfio_legacy_pci_hot_reset and move it into container.c.

vfio_pci_[pre/post]_reset and vfio_pci_host_match are exported so they could be called in legacy and iommufd pci_hot_reset callback.

Suggested-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Eric Auger eric.auger@redhat.com Tested-by: Eric Auger eric.auger@redhat.com Tested-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/vfio/container.c | 170 ++++++++++++++++++++++++++ hw/vfio/pci.c | 168 +------------------------ hw/vfio/pci.h | 3 + include/hw/vfio/vfio-container-base.h | 3 + 4 files changed, 182 insertions(+), 162 deletions(-)

diff --git a/hw/vfio/container.c b/hw/vfio/container.c index 4936b8f27f..e32e1b51e0 100644 --- a/hw/vfio/container.c +++ b/hw/vfio/container.c @@ -35,6 +35,7 @@ #include "qapi/error.h" #include "migration/migration.h" #include "sysemu/kvm.h" +#include "pci.h"

VFIOGroupList vfio_group_list = QLIST_HEAD_INITIALIZER(vfio_group_list); @@ -1035,6 +1036,174 @@ static void vfio_legacy_detach_device(VFIODevice *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, @@ -1042,4 +1211,5 @@ const VFIOIOMMUOps vfio_legacy_ops = { .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, }; diff --git a/hw/vfio/pci.c b/hw/vfio/pci.c index eb55e8ae88..d00c3472c7 100644 --- a/hw/vfio/pci.c +++ b/hw/vfio/pci.c @@ -2374,7 +2374,7 @@ static int vfio_add_capabilities(VFIOPCIDevice *vdev, Error **errp) return 0; }

-static void vfio_pci_pre_reset(VFIOPCIDevice *vdev) +void vfio_pci_pre_reset(VFIOPCIDevice *vdev) { PCIDevice *pdev = &vdev->pdev; uint16_t cmd; @@ -2411,7 +2411,7 @@ static void vfio_pci_pre_reset(VFIOPCIDevice *vdev) vfio_pci_write_config(pdev, PCI_COMMAND, cmd, 2); }

-static void vfio_pci_post_reset(VFIOPCIDevice *vdev) +void vfio_pci_post_reset(VFIOPCIDevice *vdev) { Error *err = NULL; int nr; @@ -2435,7 +2435,7 @@ static void vfio_pci_post_reset(VFIOPCIDevice *vdev) vfio_quirk_reset(vdev); }

-static bool vfio_pci_host_match(PCIHostDeviceAddress *addr, const char *name) +bool vfio_pci_host_match(PCIHostDeviceAddress *addr, const char *name) { char tmp[13];

@@ -2485,166 +2485,10 @@ int vfio_pci_get_pci_hot_reset_info(VFIOPCIDevice *vdev,

static int vfio_pci_hot_reset(VFIOPCIDevice *vdev, bool single) { - 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); - - 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); + VFIODevice *vbasedev = &vdev->vbasedev; + const VFIOIOMMUOps *ops = vbasedev->bcontainer->ops;

- return ret; + return ops->pci_hot_reset(vbasedev, single); }

/* diff --git a/hw/vfio/pci.h b/hw/vfio/pci.h index 1006061afb..6e64a2654e 100644 --- a/hw/vfio/pci.h +++ b/hw/vfio/pci.h @@ -218,6 +218,9 @@ void vfio_probe_igd_bar4_quirk(VFIOPCIDevice *vdev, int nr);

extern const PropertyInfo qdev_prop_nv_gpudirect_clique;

+void vfio_pci_pre_reset(VFIOPCIDevice *vdev); +void vfio_pci_post_reset(VFIOPCIDevice *vdev); +bool vfio_pci_host_match(PCIHostDeviceAddress *addr, const char *name); int vfio_pci_get_pci_hot_reset_info(VFIOPCIDevice *vdev, struct vfio_pci_hot_reset_info **info_p);

diff --git a/include/hw/vfio/vfio-container-base.h b/include/hw/vfio/vfio-container-base.h index 4b6f017c6f..45bb19c767 100644 --- a/include/hw/vfio/vfio-container-base.h +++ b/include/hw/vfio/vfio-container-base.h @@ -106,6 +106,9 @@ struct VFIOIOMMUOps { int (*set_dirty_page_tracking)(VFIOContainerBase *bcontainer, bool start); int (*query_dirty_bitmap)(VFIOContainerBase *bcontainer, VFIOBitmap *vbmap, hwaddr iova, hwaddr size); + /* PCI specific */ + int (*pci_hot_reset)(VFIODevice *vbasedev, bool single); + /* SPAPR specific */ int (*add_window)(VFIOContainerBase *bcontainer, MemoryRegionSection *section,

From: Eric Auger eric.auger@redhat.com

Now we support two types of iommu backends, let's add the capability to select one of them. This depends on whether an iommufd object has been linked with the vfio-pci device:

If the user wants to use the legacy backend, it shall not link the vfio-pci device with any iommufd object:

-device vfio-pci,host=0000:02:00.0

This is called the legacy mode/backend.

If the user wants to use the iommufd backend (/dev/iommu) it shall pass an iommufd object id in the vfio-pci device options:

-object iommufd,id=iommufd0 -device vfio-pci,host=0000:02:00.0,iommufd=iommufd0

Suggested-by: Alex Williamson alex.williamson@redhat.com Signed-off-by: Eric Auger eric.auger@redhat.com Signed-off-by: Yi Liu yi.l.liu@intel.com Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Cédric Le Goater clg@redhat.com Tested-by: Eric Auger eric.auger@redhat.com Tested-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/vfio/pci.c | 6 ++++++ 1 file changed, 6 insertions(+)

diff --git a/hw/vfio/pci.c b/hw/vfio/pci.c index d00c3472c7..c5984b0598 100644 --- a/hw/vfio/pci.c +++ b/hw/vfio/pci.c @@ -19,6 +19,7 @@ */

#include "qemu/osdep.h" +#include CONFIG_DEVICES /* CONFIG_IOMMUFD */ #include <linux/vfio.h> #include <sys/ioctl.h>

@@ -42,6 +43,7 @@ #include "qapi/error.h" #include "migration/blocker.h" #include "migration/qemu-file.h" +#include "sysemu/iommufd.h"

#define TYPE_VFIO_PCI_NOHOTPLUG "vfio-pci-nohotplug"

@@ -3386,6 +3388,10 @@ static Property vfio_pci_dev_properties[] = { * DEFINE_PROP_STRING("vfiofd", VFIOPCIDevice, vfiofd_name), * DEFINE_PROP_STRING("vfiogroupfd, VFIOPCIDevice, vfiogroupfd_name), */ +#ifdef CONFIG_IOMMUFD + DEFINE_PROP_LINK("iommufd", VFIOPCIDevice, vbasedev.iommufd, + TYPE_IOMMUFD_BACKEND, IOMMUFDBackend *), +#endif DEFINE_PROP_END_OF_LIST(), };

From: Zhenzhong Duan zhenzhong.duan@intel.com

Now we support two types of iommu backends, let's add the capability to select one of them. This depends on whether an iommufd object has been linked with the vfio-platform device:

If the user wants to use the legacy backend, it shall not link the vfio-platform device with any iommufd object:

-device vfio-platform,host=XXX

This is called the legacy mode/backend.

If the user wants to use the iommufd backend (/dev/iommu) it shall pass an iommufd object id in the vfio-platform device options:

-object iommufd,id=iommufd0 -device vfio-platform,host=XXX,iommufd=iommufd0

Suggested-by: Alex Williamson alex.williamson@redhat.com Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Cédric Le Goater clg@redhat.com Reviewed-by: Eric Auger eric.auger@redhat.com Tested-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/vfio/platform.c | 6 ++++++ 1 file changed, 6 insertions(+)

diff --git a/hw/vfio/platform.c b/hw/vfio/platform.c index 8e3d4ac458..98ae4bc655 100644 --- a/hw/vfio/platform.c +++ b/hw/vfio/platform.c @@ -15,11 +15,13 @@ */

#include "qemu/osdep.h" +#include CONFIG_DEVICES /* CONFIG_IOMMUFD */ #include "qapi/error.h" #include <sys/ioctl.h> #include <linux/vfio.h>

#include "hw/vfio/vfio-platform.h" +#include "sysemu/iommufd.h" #include "migration/vmstate.h" #include "qemu/error-report.h" #include "qemu/lockable.h" @@ -649,6 +651,10 @@ static Property vfio_platform_dev_properties[] = { DEFINE_PROP_UINT32("mmap-timeout-ms", VFIOPlatformDevice, mmap_timeout, 1100), DEFINE_PROP_BOOL("x-irqfd", VFIOPlatformDevice, irqfd_allowed, true), +#ifdef CONFIG_IOMMUFD + DEFINE_PROP_LINK("iommufd", VFIOPlatformDevice, vbasedev.iommufd, + TYPE_IOMMUFD_BACKEND, IOMMUFDBackend *), +#endif DEFINE_PROP_END_OF_LIST(), };

From: Zhenzhong Duan zhenzhong.duan@intel.com

Now we support two types of iommu backends, let's add the capability to select one of them. This depends on whether an iommufd object has been linked with the vfio-ap device:

if the user wants to use the legacy backend, it shall not link the vfio-ap device with any iommufd object:

-device vfio-ap,sysfsdev=/sys/bus/mdev/devices/XXX

This is called the legacy mode/backend.

If the user wants to use the iommufd backend (/dev/iommu) it shall pass an iommufd object id in the vfio-ap device options:

-object iommufd,id=iommufd0 -device vfio-ap,sysfsdev=/sys/bus/mdev/devices/XXX,iommufd=iommufd0

Suggested-by: Alex Williamson alex.williamson@redhat.com Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Matthew Rosato mjrosato@linux.ibm.com Reviewed-by: Cédric Le Goater clg@redhat.com Tested-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/vfio/ap.c | 6 ++++++ 1 file changed, 6 insertions(+)

diff --git a/hw/vfio/ap.c b/hw/vfio/ap.c index bbf69ff55a..80629609ae 100644 --- a/hw/vfio/ap.c +++ b/hw/vfio/ap.c @@ -11,10 +11,12 @@ */

#include "qemu/osdep.h" +#include CONFIG_DEVICES /* CONFIG_IOMMUFD */ #include <linux/vfio.h> #include <sys/ioctl.h> #include "qapi/error.h" #include "hw/vfio/vfio-common.h" +#include "sysemu/iommufd.h" #include "hw/s390x/ap-device.h" #include "qemu/error-report.h" #include "qemu/event_notifier.h" @@ -204,6 +206,10 @@ static void vfio_ap_unrealize(DeviceState *dev)

static Property vfio_ap_properties[] = { DEFINE_PROP_STRING("sysfsdev", VFIOAPDevice, vdev.sysfsdev), +#ifdef CONFIG_IOMMUFD + DEFINE_PROP_LINK("iommufd", VFIOAPDevice, vdev.iommufd, + TYPE_IOMMUFD_BACKEND, IOMMUFDBackend *), +#endif DEFINE_PROP_END_OF_LIST(), };

From: Zhenzhong Duan zhenzhong.duan@intel.com

Now we support two types of iommu backends, let's add the capability to select one of them. This depends on whether an iommufd object has been linked with the vfio-ccw device:

If the user wants to use the legacy backend, it shall not link the vfio-ccw device with any iommufd object:

-device vfio-ccw,sysfsdev=/sys/bus/mdev/devices/XXX

This is called the legacy mode/backend.

If the user wants to use the iommufd backend (/dev/iommu) it shall pass an iommufd object id in the vfio-ccw device options:

-object iommufd,id=iommufd0 -device vfio-ccw,sysfsdev=/sys/bus/mdev/devices/XXX,iommufd=iommufd0

Suggested-by: Alex Williamson alex.williamson@redhat.com Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Matthew Rosato mjrosato@linux.ibm.com Reviewed-by: Cédric Le Goater clg@redhat.com Reviewed-by: Eric Farman farman@linux.ibm.com Tested-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/vfio/ccw.c | 6 ++++++ 1 file changed, 6 insertions(+)

diff --git a/hw/vfio/ccw.c b/hw/vfio/ccw.c index d857bb8d0f..d2d58bb677 100644 --- a/hw/vfio/ccw.c +++ b/hw/vfio/ccw.c @@ -15,12 +15,14 @@ */

#include "qemu/osdep.h" +#include CONFIG_DEVICES /* CONFIG_IOMMUFD */ #include <linux/vfio.h> #include <linux/vfio_ccw.h> #include <sys/ioctl.h>

#include "qapi/error.h" #include "hw/vfio/vfio-common.h" +#include "sysemu/iommufd.h" #include "hw/s390x/s390-ccw.h" #include "hw/s390x/vfio-ccw.h" #include "hw/qdev-properties.h" @@ -677,6 +679,10 @@ static void vfio_ccw_unrealize(DeviceState *dev) static Property vfio_ccw_properties[] = { DEFINE_PROP_STRING("sysfsdev", VFIOCCWDevice, vdev.sysfsdev), DEFINE_PROP_BOOL("force-orb-pfch", VFIOCCWDevice, force_orb_pfch, false), +#ifdef CONFIG_IOMMUFD + DEFINE_PROP_LINK("iommufd", VFIOCCWDevice, vdev.iommufd, + TYPE_IOMMUFD_BACKEND, IOMMUFDBackend *), +#endif DEFINE_PROP_END_OF_LIST(), };

From: Zhenzhong Duan zhenzhong.duan@intel.com

Some of the callbacks in VFIOIOMMUOps pass VFIOContainerBase poiner, those callbacks only need read access to the sub object of VFIOContainerBase. So make VFIOContainerBase, VFIOContainer and VFIOIOMMUFDContainer as const in these callbacks.

Local functions called by those callbacks also need same changes to avoid build error.

Modify vfio_lookup_match_range/vfio_legacy_dma_map during backporting.

Suggested-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Cédric Le Goater clg@redhat.com Reviewed-by: Eric Auger eric.auger@redhat.com Tested-by: Eric Auger eric.auger@redhat.com Tested-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/vfio/common.c | 9 +++---- hw/vfio/container-base.c | 2 +- hw/vfio/container.c | 34 ++++++++++++++------------- hw/vfio/iommufd.c | 8 +++---- include/hw/vfio/vfio-common.h | 14 ++++++----- include/hw/vfio/vfio-container-base.h | 12 ++++++---- 6 files changed, 43 insertions(+), 36 deletions(-)

diff --git a/hw/vfio/common.c b/hw/vfio/common.c index 0e900c6746..d572ec5880 100644 --- a/hw/vfio/common.c +++ b/hw/vfio/common.c @@ -204,7 +204,7 @@ static bool vfio_devices_all_dirty_tracking(VFIOContainerBase *bcontainer) return true; }

-bool vfio_devices_all_device_dirty_tracking(VFIOContainerBase *bcontainer) +bool vfio_devices_all_device_dirty_tracking(const VFIOContainerBase *bcontainer) { VFIODevice *vbasedev;

@@ -221,7 +221,8 @@ bool vfio_devices_all_device_dirty_tracking(VFIOContainerBase *bcontainer) * Check if all VFIO devices are running and migration is active, which is * essentially equivalent to the migration being in pre-copy phase. */ -bool vfio_devices_all_running_and_mig_active(VFIOContainerBase *bcontainer) +bool +vfio_devices_all_running_and_mig_active(const VFIOContainerBase *bcontainer) { VFIODevice *vbasedev;

@@ -1139,7 +1140,7 @@ static int vfio_device_dma_logging_report(VFIODevice *vbasedev, hwaddr iova, return 0; }

-int vfio_devices_query_dirty_bitmap(VFIOContainerBase *bcontainer, +int vfio_devices_query_dirty_bitmap(const VFIOContainerBase *bcontainer, VFIOBitmap *vbmap, hwaddr iova, hwaddr size) { @@ -1162,7 +1163,7 @@ int vfio_devices_query_dirty_bitmap(VFIOContainerBase *bcontainer, return 0; }

-int vfio_get_dirty_bitmap(VFIOContainerBase *bcontainer, uint64_t iova, +int vfio_get_dirty_bitmap(const VFIOContainerBase *bcontainer, uint64_t iova, uint64_t size, ram_addr_t ram_addr) { bool all_device_dirty_tracking = diff --git a/hw/vfio/container-base.c b/hw/vfio/container-base.c index eee2dcfe76..1ffd25bbfa 100644 --- a/hw/vfio/container-base.c +++ b/hw/vfio/container-base.c @@ -63,7 +63,7 @@ int vfio_container_set_dirty_page_tracking(VFIOContainerBase *bcontainer, return bcontainer->ops->set_dirty_page_tracking(bcontainer, start); }

-int vfio_container_query_dirty_bitmap(VFIOContainerBase *bcontainer, +int vfio_container_query_dirty_bitmap(const VFIOContainerBase *bcontainer, VFIOBitmap *vbmap, hwaddr iova, hwaddr size) { diff --git a/hw/vfio/container.c b/hw/vfio/container.c index e32e1b51e0..67aeaa825b 100644 --- a/hw/vfio/container.c +++ b/hw/vfio/container.c @@ -63,11 +63,11 @@ 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) { - VFIOContainerBase *bcontainer = &container->bcontainer; + const VFIOContainerBase *bcontainer = &container->bcontainer; struct vfio_iommu_type1_dma_unmap *unmap; struct vfio_bitmap *bitmap; VFIOBitmap vbmap; @@ -116,7 +116,7 @@ unmap_exit: return ret; }

-VFIODMARange *vfio_lookup_match_range(VFIOContainer *container, +VFIODMARange *vfio_lookup_match_range(const VFIOContainer *container, hwaddr start_addr, hwaddr size) { VFIODMARange *qrange; @@ -142,11 +142,12 @@ void vfio_dma_range_init_dirty_bitmap(VFIODMARange *qrange) /* * DMA - Mapping and unmapping for the "type1" IOMMU interface used on x86 */ -static int vfio_legacy_dma_unmap(VFIOContainerBase *bcontainer, 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) { - VFIOContainer *container = container_of(bcontainer, VFIOContainer, - bcontainer); + const VFIOContainer *container = container_of(bcontainer, VFIOContainer, + bcontainer); struct vfio_iommu_type1_dma_unmap unmap = { .argsz = sizeof(unmap), .flags = 0, @@ -216,11 +217,11 @@ static int vfio_legacy_dma_unmap(VFIOContainerBase *bcontainer, hwaddr iova, return 0; }

-static int vfio_legacy_dma_map(VFIOContainerBase *bcontainer, hwaddr iova, +static int vfio_legacy_dma_map(const VFIOContainerBase *bcontainer, hwaddr iova, ram_addr_t size, void *vaddr, bool readonly) { VFIOContainer *container = container_of(bcontainer, VFIOContainer, - bcontainer); + bcontainer); struct vfio_iommu_type1_dma_map map = { .argsz = sizeof(map), .flags = VFIO_DMA_MAP_FLAG_READ, @@ -257,11 +258,12 @@ static int vfio_legacy_dma_map(VFIOContainerBase *bcontainer, hwaddr iova, return -errno; }

-static int vfio_legacy_set_dirty_page_tracking(VFIOContainerBase *bcontainer, - bool start) +static int +vfio_legacy_set_dirty_page_tracking(const VFIOContainerBase *bcontainer, + bool start) { - VFIOContainer *container = container_of(bcontainer, VFIOContainer, - bcontainer); + const VFIOContainer *container = container_of(bcontainer, VFIOContainer, + bcontainer); int ret; struct vfio_iommu_type1_dirty_bitmap dirty = { .argsz = sizeof(dirty), @@ -283,12 +285,12 @@ static int vfio_legacy_set_dirty_page_tracking(VFIOContainerBase *bcontainer, return ret; }

-static int vfio_legacy_query_dirty_bitmap(VFIOContainerBase *bcontainer, +static int vfio_legacy_query_dirty_bitmap(const VFIOContainerBase *bcontainer, VFIOBitmap *vbmap, hwaddr iova, hwaddr size) { - VFIOContainer *container = container_of(bcontainer, VFIOContainer, - bcontainer); + 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; diff --git a/hw/vfio/iommufd.c b/hw/vfio/iommufd.c index 5accd26484..87a561c545 100644 --- a/hw/vfio/iommufd.c +++ b/hw/vfio/iommufd.c @@ -26,10 +26,10 @@ #include "qemu/chardev_open.h" #include "pci.h"

-static int iommufd_cdev_map(VFIOContainerBase *bcontainer, hwaddr iova, +static int iommufd_cdev_map(const VFIOContainerBase *bcontainer, hwaddr iova, ram_addr_t size, void *vaddr, bool readonly) { - VFIOIOMMUFDContainer *container = + const VFIOIOMMUFDContainer *container = container_of(bcontainer, VFIOIOMMUFDContainer, bcontainer);

return iommufd_backend_map_dma(container->be, @@ -37,11 +37,11 @@ static int iommufd_cdev_map(VFIOContainerBase *bcontainer, hwaddr iova, iova, size, vaddr, readonly); }

-static int iommufd_cdev_unmap(VFIOContainerBase *bcontainer, +static int iommufd_cdev_unmap(const VFIOContainerBase *bcontainer, hwaddr iova, ram_addr_t size, IOMMUTLBEntry *iotlb) { - VFIOIOMMUFDContainer *container = + const VFIOIOMMUFDContainer *container = container_of(bcontainer, VFIOIOMMUFDContainer, bcontainer);

/* TODO: Handle dma_unmap_bitmap with iotlb args (migration) */ diff --git a/include/hw/vfio/vfio-common.h b/include/hw/vfio/vfio-common.h index 5f35f2900b..37f01410d5 100644 --- a/include/hw/vfio/vfio-common.h +++ b/include/hw/vfio/vfio-common.h @@ -186,7 +186,7 @@ typedef struct VFIODisplay { VFIOAddressSpace *vfio_get_address_space(AddressSpace *as); void vfio_put_address_space(VFIOAddressSpace *space);

-VFIODMARange *vfio_lookup_match_range(VFIOContainer *container, +VFIODMARange *vfio_lookup_match_range(const VFIOContainer *container, hwaddr start_addr, hwaddr size); void vfio_dma_range_init_dirty_bitmap(VFIODMARange *qrange);

@@ -258,13 +258,15 @@ bool vfio_migration_realize(VFIODevice *vbasedev, Error **errp); void vfio_migration_exit(VFIODevice *vbasedev);

int vfio_bitmap_alloc(VFIOBitmap *vbmap, hwaddr size); -bool vfio_devices_all_running_and_mig_active(VFIOContainerBase *bcontainer); -bool vfio_devices_all_device_dirty_tracking(VFIOContainerBase *bcontainer); -int vfio_devices_query_dirty_bitmap(VFIOContainerBase *bcontainer, +bool +vfio_devices_all_running_and_mig_active(const VFIOContainerBase *bcontainer); +bool +vfio_devices_all_device_dirty_tracking(const VFIOContainerBase *bcontainer); +int vfio_devices_query_dirty_bitmap(const VFIOContainerBase *bcontainer, VFIOBitmap *vbmap, hwaddr iova, hwaddr size); -int vfio_get_dirty_bitmap(VFIOContainerBase *bcontainer, uint64_t iova, - uint64_t size, ram_addr_t ram_addr); +int vfio_get_dirty_bitmap(const VFIOContainerBase *bcontainer, uint64_t iova, + uint64_t size, ram_addr_t ram_addr);

/* Returns 0 on success, or a negative errno. */ int vfio_device_get_name(VFIODevice *vbasedev, Error **errp); diff --git a/include/hw/vfio/vfio-container-base.h b/include/hw/vfio/vfio-container-base.h index 45bb19c767..2ae297ccda 100644 --- a/include/hw/vfio/vfio-container-base.h +++ b/include/hw/vfio/vfio-container-base.h @@ -82,7 +82,7 @@ void vfio_container_del_section_window(VFIOContainerBase *bcontainer, MemoryRegionSection *section); int vfio_container_set_dirty_page_tracking(VFIOContainerBase *bcontainer, bool start); -int vfio_container_query_dirty_bitmap(VFIOContainerBase *bcontainer, +int vfio_container_query_dirty_bitmap(const VFIOContainerBase *bcontainer, VFIOBitmap *vbmap, hwaddr iova, hwaddr size);

@@ -93,18 +93,20 @@ void vfio_container_destroy(VFIOContainerBase *bcontainer);

struct VFIOIOMMUOps { /* basic feature */ - int (*dma_map)(VFIOContainerBase *bcontainer, + int (*dma_map)(const VFIOContainerBase *bcontainer, hwaddr iova, ram_addr_t size, void *vaddr, bool readonly); - int (*dma_unmap)(VFIOContainerBase *bcontainer, + int (*dma_unmap)(const VFIOContainerBase *bcontainer, hwaddr iova, ram_addr_t size, IOMMUTLBEntry *iotlb); int (*attach_device)(const char *name, VFIODevice *vbasedev, AddressSpace *as, Error **errp); void (*detach_device)(VFIODevice *vbasedev); /* migration feature */ - int (*set_dirty_page_tracking)(VFIOContainerBase *bcontainer, bool start); - int (*query_dirty_bitmap)(VFIOContainerBase *bcontainer, VFIOBitmap *vbmap, + int (*set_dirty_page_tracking)(const VFIOContainerBase *bcontainer, + bool start); + int (*query_dirty_bitmap)(const VFIOContainerBase *bcontainer, + VFIOBitmap *vbmap, hwaddr iova, hwaddr size); /* PCI specific */ int (*pci_hot_reset)(VFIODevice *vbasedev, bool single);

From: Cédric Le Goater clg@redhat.com

Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Matthew Rosato mjrosato@linux.ibm.com Reviewed-by: Eric Farman farman@linux.ibm.com Tested-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/s390x/Kconfig | 1 + 1 file changed, 1 insertion(+)

diff --git a/hw/s390x/Kconfig b/hw/s390x/Kconfig index 4c068d7960..26ad104485 100644 --- a/hw/s390x/Kconfig +++ b/hw/s390x/Kconfig @@ -6,6 +6,7 @@ config S390_CCW_VIRTIO imply VFIO_CCW imply WDT_DIAG288 imply PCIE_DEVICES + imply IOMMUFD select PCI_EXPRESS select S390_FLIC select S390_FLIC_KVM if KVM

From: Zhenzhong Duan zhenzhong.duan@intel.com

Some of the VFIODevice initializations is in vfio_realize, move all of them in vfio_instance_init.

No functional change intended.

Suggested-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Philippe Mathieu-Daudé philmd@linaro.org Tested-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/vfio/pci.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/hw/vfio/pci.c b/hw/vfio/pci.c index 445d58c8e5..87405584d7 100644 --- a/hw/vfio/pci.c +++ b/hw/vfio/pci.c @@ -2969,9 +2969,6 @@ static void vfio_realize(PCIDevice *pdev, Error **errp) if (vfio_device_get_name(vbasedev, errp) < 0) { return; } - vbasedev->ops = &vfio_pci_ops; - vbasedev->type = VFIO_DEVICE_TYPE_PCI; - vbasedev->dev = DEVICE(vdev);

/* * Mediated devices *might* operate compatibly with discarding of RAM, but @@ -3320,6 +3317,7 @@ static void vfio_instance_init(Object *obj) { PCIDevice *pci_dev = PCI_DEVICE(obj); VFIOPCIDevice *vdev = VFIO_PCI(obj); + VFIODevice *vbasedev = &vdev->vbasedev;

device_add_bootindex_property(obj, &vdev->bootindex, "bootindex", NULL, @@ -3328,7 +3326,11 @@ static void vfio_instance_init(Object *obj) vdev->host.bus = ~0U; vdev->host.slot = ~0U; vdev->host.function = ~0U; - vdev->vbasedev.fd = -1; + + vbasedev->type = VFIO_DEVICE_TYPE_PCI; + vbasedev->ops = &vfio_pci_ops; + vbasedev->dev = DEVICE(vdev); + vbasedev->fd = -1;

vdev->nv_gpudirect_clique = 0xFF;

From: Zhenzhong Duan zhenzhong.duan@intel.com

Some of the VFIODevice initializations is in vfio_ap_realize, move all of them in vfio_ap_instance_init.

No functional change intended.

Suggested-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Reviewed-by: Philippe Mathieu-Daudé philmd@linaro.org Reviewed-by: Eric Farman farman@linux.ibm.com Tested-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/vfio/ap.c | 26 +++++++++++++------------- 1 file changed, 13 insertions(+), 13 deletions(-)

diff --git a/hw/vfio/ap.c b/hw/vfio/ap.c index f180e4a32a..95fe7cd98b 100644 --- a/hw/vfio/ap.c +++ b/hw/vfio/ap.c @@ -164,18 +164,6 @@ static void vfio_ap_realize(DeviceState *dev, Error **errp) return; }

- vbasedev->ops = &vfio_ap_ops; - vbasedev->type = VFIO_DEVICE_TYPE_AP; - vbasedev->dev = dev; - - /* - * vfio-ap devices operate in a way compatible with discarding of - * memory in RAM blocks, as no pages are pinned in the host. - * This needs to be set before vfio_get_device() for vfio common to - * handle ram_block_discard_disable(). - */ - vapdev->vdev.ram_block_discard_allowed = true; - ret = vfio_attach_device(vbasedev->name, vbasedev, &address_space_memory, errp); if (ret) { @@ -236,8 +224,20 @@ static const VMStateDescription vfio_ap_vmstate = { static void vfio_ap_instance_init(Object *obj) { VFIOAPDevice *vapdev = VFIO_AP_DEVICE(obj); + VFIODevice *vbasedev = &vapdev->vdev;

- vapdev->vdev.fd = -1; + vbasedev->type = VFIO_DEVICE_TYPE_AP; + vbasedev->ops = &vfio_ap_ops; + vbasedev->dev = DEVICE(vapdev); + vbasedev->fd = -1; + + /* + * vfio-ap devices operate in a way compatible with discarding of + * memory in RAM blocks, as no pages are pinned in the host. + * This needs to be set before vfio_get_device() for vfio common to + * handle ram_block_discard_disable(). + */ + vbasedev->ram_block_discard_allowed = true; }

#ifdef CONFIG_IOMMUFD

From: Zhenzhong Duan zhenzhong.duan@intel.com

Introduce a helper function to replace the common code to initialize VFIODevice in pci, platform, ap and ccw VFIO device.

No functional change intended.

Suggested-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhenzhong Duan zhenzhong.duan@intel.com Tested-by: Nicolin Chen nicolinc@nvidia.com Signed-off-by: Cédric Le Goater clg@redhat.com Signed-off-by: Zhou Wang wangzhou1@hisilicon.com --- hw/vfio/ap.c | 8 ++------ hw/vfio/ccw.c | 8 ++------ hw/vfio/helpers.c | 11 +++++++++++ hw/vfio/pci.c | 6 ++---- hw/vfio/platform.c | 6 ++---- include/hw/vfio/vfio-common.h | 2 ++ 6 files changed, 21 insertions(+), 20 deletions(-)

diff --git a/hw/vfio/ap.c b/hw/vfio/ap.c index 95fe7cd98b..e157aa1ff7 100644 --- a/hw/vfio/ap.c +++ b/hw/vfio/ap.c @@ -226,18 +226,14 @@ static void vfio_ap_instance_init(Object *obj) VFIOAPDevice *vapdev = VFIO_AP_DEVICE(obj); VFIODevice *vbasedev = &vapdev->vdev;

- vbasedev->type = VFIO_DEVICE_TYPE_AP; - vbasedev->ops = &vfio_ap_ops; - vbasedev->dev = DEVICE(vapdev); - vbasedev->fd = -1; - /* * vfio-ap devices operate in a way compatible with discarding of * memory in RAM blocks, as no pages are pinned in the host. * This needs to be set before vfio_get_device() for vfio common to * handle ram_block_discard_disable(). */ - vbasedev->ram_block_discard_allowed = true; + vfio_device_init(vbasedev, VFIO_DEVICE_TYPE_AP, &vfio_ap_ops, + DEVICE(vapdev), true); }

#ifdef CONFIG_IOMMUFD diff --git a/hw/vfio/ccw.c b/hw/vfio/ccw.c index 6305a4c1b8..90e4a53437 100644 --- a/hw/vfio/ccw.c +++ b/hw/vfio/ccw.c @@ -683,11 +683,6 @@ static void vfio_ccw_instance_init(Object *obj) VFIOCCWDevice *vcdev = VFIO_CCW(obj); VFIODevice *vbasedev = &vcdev->vdev;

- vbasedev->type = VFIO_DEVICE_TYPE_CCW; - vbasedev->ops = &vfio_ccw_ops; - vbasedev->dev = DEVICE(vcdev); - vbasedev->fd = -1; - /* * All vfio-ccw devices are believed to operate in a way compatible with * discarding of memory in RAM blocks, ie. pages pinned in the host are @@ -696,7 +691,8 @@ static void vfio_ccw_instance_init(Object *obj) * needs to be set before vfio_get_device() for vfio common to handle * ram_block_discard_disable(). */ - vbasedev->ram_block_discard_allowed = true; + vfio_device_init(vbasedev, VFIO_DEVICE_TYPE_CCW, &vfio_ccw_ops, + DEVICE(vcdev), true); }

#ifdef CONFIG_IOMMUFD diff --git a/hw/vfio/helpers.c b/hw/vfio/helpers.c index 3592c3d54e..6789870802 100644 --- a/hw/vfio/helpers.c +++ b/hw/vfio/helpers.c @@ -652,3 +652,14 @@ void vfio_device_set_fd(VFIODevice *vbasedev, const char *str, Error **errp) } vbasedev->fd = fd; } + +void vfio_device_init(VFIODevice *vbasedev, int type, VFIODeviceOps *ops, + DeviceState *dev, bool ram_discard) +{ + vbasedev->type = type; + vbasedev->ops = ops; + vbasedev->dev = dev; + vbasedev->fd = -1; + + vbasedev->ram_block_discard_allowed = ram_discard; +} diff --git a/hw/vfio/pci.c b/hw/vfio/pci.c index 87405584d7..1874ec1aba 100644 --- a/hw/vfio/pci.c +++ b/hw/vfio/pci.c @@ -3327,10 +3327,8 @@ static void vfio_instance_init(Object *obj) vdev->host.slot = ~0U; vdev->host.function = ~0U;

- vbasedev->type = VFIO_DEVICE_TYPE_PCI; - vbasedev->ops = &vfio_pci_ops; - vbasedev->dev = DEVICE(vdev); - vbasedev->fd = -1; + vfio_device_init(vbasedev, VFIO_DEVICE_TYPE_PCI, &vfio_pci_ops, + DEVICE(vdev), false);

vdev->nv_gpudirect_clique = 0xFF;

diff --git a/hw/vfio/platform.c b/hw/vfio/platform.c index 506eb8193f..a8d9b7da63 100644 --- a/hw/vfio/platform.c +++ b/hw/vfio/platform.c @@ -657,10 +657,8 @@ static void vfio_platform_instance_init(Object *obj) VFIOPlatformDevice *vdev = VFIO_PLATFORM_DEVICE(obj); VFIODevice *vbasedev = &vdev->vbasedev;

- vbasedev->type = VFIO_DEVICE_TYPE_PLATFORM; - vbasedev->ops = &vfio_platform_ops; - vbasedev->dev = DEVICE(vdev); - vbasedev->fd = -1; + vfio_device_init(vbasedev, VFIO_DEVICE_TYPE_PLATFORM, &vfio_platform_ops, + DEVICE(vdev), false); }

#ifdef CONFIG_IOMMUFD diff --git a/include/hw/vfio/vfio-common.h b/include/hw/vfio/vfio-common.h index 37f01410d5..151b2ab65f 100644 --- a/include/hw/vfio/vfio-common.h +++ b/include/hw/vfio/vfio-common.h @@ -271,4 +271,6 @@ int vfio_get_dirty_bitmap(const VFIOContainerBase *bcontainer, uint64_t iova, /* Returns 0 on success, or a negative errno. */ int vfio_device_get_name(VFIODevice *vbasedev, Error **errp); void vfio_device_set_fd(VFIODevice *vbasedev, const char *str, Error **errp); +void vfio_device_init(VFIODevice *vbasedev, int type, VFIODeviceOps *ops, + DeviceState *dev, bool ram_discard); #endif /* HW_VFIO_VFIO_COMMON_H */