qemu-kvm/kvm-backends-iommufd-Introduce-the-iommufd-object.patch
Miroslav Rezanina 9ef7cdf7ca * Wed Jan 24 2024 Miroslav Rezanina <mrezanin@redhat.com> - 8.2.0-3
- kvm-hw-arm-virt-Add-properties-to-disable-high-memory-re.patch [RHEL-19738]
- kvm-vfio-Introduce-base-object-for-VFIOContainer-and-tar.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-container-Introduce-a-empty-VFIOIOMMUOps.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-container-Switch-to-dma_map-unmap-API.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-common-Introduce-vfio_container_init-destroy-he.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-common-Move-giommu_list-in-base-container.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-container-Move-space-field-to-base-container.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-container-Switch-to-IOMMU-BE-set_dirty_page_tra.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-container-Move-per-container-device-list-in-bas.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-container-Convert-functions-to-base-container.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-container-Move-pgsizes-and-dma_max_mappings-to-.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-container-Move-vrdl_list-to-base-container.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-container-Move-listener-to-base-container.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-container-Move-dirty_pgsizes-and-max_dirty_bitm.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-container-Move-iova_ranges-to-base-container.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-container-Implement-attach-detach_device.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-spapr-Introduce-spapr-backend-and-target-interf.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-spapr-switch-to-spapr-IOMMU-BE-add-del_section_.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-spapr-Move-prereg_listener-into-spapr-container.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-spapr-Move-hostwin_list-into-spapr-container.patch [RHEL-19302 RHEL-21057]
- kvm-backends-iommufd-Introduce-the-iommufd-object.patch [RHEL-19302 RHEL-21057]
- kvm-util-char_dev-Add-open_cdev.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-common-return-early-if-space-isn-t-empty.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-iommufd-Implement-the-iommufd-backend.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-iommufd-Relax-assert-check-for-iommufd-backend.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-iommufd-Add-support-for-iova_ranges-and-pgsizes.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-pci-Extract-out-a-helper-vfio_pci_get_pci_hot_r.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-pci-Introduce-a-vfio-pci-hot-reset-interface.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-iommufd-Enable-pci-hot-reset-through-iommufd-cd.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-pci-Allow-the-selection-of-a-given-iommu-backen.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-pci-Make-vfio-cdev-pre-openable-by-passing-a-fi.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-platform-Allow-the-selection-of-a-given-iommu-b.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-platform-Make-vfio-cdev-pre-openable-by-passing.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-ap-Allow-the-selection-of-a-given-iommu-backend.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-ap-Make-vfio-cdev-pre-openable-by-passing-a-fil.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-ccw-Allow-the-selection-of-a-given-iommu-backen.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-ccw-Make-vfio-cdev-pre-openable-by-passing-a-fi.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-Make-VFIOContainerBase-poiner-parameter-const-i.patch [RHEL-19302 RHEL-21057]
- kvm-hw-arm-Activate-IOMMUFD-for-virt-machines.patch [RHEL-19302 RHEL-21057]
- kvm-kconfig-Activate-IOMMUFD-for-s390x-machines.patch [RHEL-19302 RHEL-21057]
- kvm-hw-i386-Activate-IOMMUFD-for-q35-machines.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-pci-Move-VFIODevice-initializations-in-vfio_ins.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-platform-Move-VFIODevice-initializations-in-vfi.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-ap-Move-VFIODevice-initializations-in-vfio_ap_i.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-ccw-Move-VFIODevice-initializations-in-vfio_ccw.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-Introduce-a-helper-function-to-initialize-VFIOD.patch [RHEL-19302 RHEL-21057]
- kvm-docs-devel-Add-VFIO-iommufd-backend-documentation.patch [RHEL-19302 RHEL-21057]
- kvm-hw-ppc-Kconfig-Imply-VFIO_PCI.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-spapr-Extend-VFIOIOMMUOps-with-a-release-handle.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-container-Introduce-vfio_legacy_setup-for-furth.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-container-Initialize-VFIOIOMMUOps-under-vfio_in.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-container-Introduce-a-VFIOIOMMU-QOM-interface.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-container-Introduce-a-VFIOIOMMU-legacy-QOM-inte.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-container-Intoduce-a-new-VFIOIOMMUClass-setup-h.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-spapr-Introduce-a-sPAPR-VFIOIOMMU-QOM-interface.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-iommufd-Introduce-a-VFIOIOMMU-iommufd-QOM-inter.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-spapr-Only-compile-sPAPR-IOMMU-support-when-nee.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-iommufd-Remove-CONFIG_IOMMUFD-usage.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-container-Replace-basename-with-g_path_get_base.patch [RHEL-19302 RHEL-21057]
- kvm-hw-vfio-fix-iteration-over-global-VFIODevice-list.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-iommufd-Remove-the-use-of-stat-to-check-file-ex.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-container-Rename-vfio_init_container-to-vfio_se.patch [RHEL-19302 RHEL-21057]
- kvm-vfio-migration-Add-helper-function-to-set-state-or-r.patch [RHEL-19302 RHEL-21057]
- kvm-backends-iommufd-Remove-check-on-number-of-backend-u.patch [RHEL-19302 RHEL-21057]
- kvm-backends-iommufd-Remove-mutex.patch [RHEL-19302 RHEL-21057]
- kvm-Compile-IOMMUFD-object-on-aarch64.patch [RHEL-19302 RHEL-21057]
- kvm-Compile-IOMMUFD-on-s390x.patch [RHEL-19302 RHEL-21057]
- kvm-Compile-IOMMUFD-on-x86_64.patch [RHEL-19302 RHEL-21057]
- kvm-target-s390x-kvm-pv-Provide-some-more-useful-informa.patch [RHEL-18212]
- kvm-nbd-server-avoid-per-NBDRequest-nbd_client_get-put.patch [RHEL-15965]
- kvm-nbd-server-only-traverse-NBDExport-clients-from-main.patch [RHEL-15965]
- kvm-nbd-server-introduce-NBDClient-lock-to-protect-field.patch [RHEL-15965]
- kvm-block-file-posix-set-up-Linux-AIO-and-io_uring-in-th.patch [RHEL-15965]
- kvm-virtio-blk-add-lock-to-protect-s-rq.patch [RHEL-15965]
- kvm-virtio-blk-don-t-lock-AioContext-in-the-completion-c.patch [RHEL-15965]
- kvm-virtio-blk-don-t-lock-AioContext-in-the-submission-c.patch [RHEL-15965]
- kvm-scsi-only-access-SCSIDevice-requests-from-one-thread.patch [RHEL-15965]
- kvm-virtio-scsi-don-t-lock-AioContext-around-virtio_queu.patch [RHEL-15965]
- kvm-scsi-don-t-lock-AioContext-in-I-O-code-path.patch [RHEL-15965]
- kvm-dma-helpers-don-t-lock-AioContext-in-dma_blk_cb.patch [RHEL-15965]
- kvm-virtio-scsi-replace-AioContext-lock-with-tmf_bh_lock.patch [RHEL-15965]
- kvm-scsi-assert-that-callbacks-run-in-the-correct-AioCon.patch [RHEL-15965]
- kvm-tests-remove-aio_context_acquire-tests.patch [RHEL-15965]
- kvm-aio-make-aio_context_acquire-aio_context_release-a-n.patch [RHEL-15965]
- kvm-graph-lock-remove-AioContext-locking.patch [RHEL-15965]
- kvm-block-remove-AioContext-locking.patch [RHEL-15965]
- kvm-block-remove-bdrv_co_lock.patch [RHEL-15965]
- kvm-scsi-remove-AioContext-locking.patch [RHEL-15965]
- kvm-aio-wait-draw-equivalence-between-AIO_WAIT_WHILE-and.patch [RHEL-15965]
- kvm-aio-remove-aio_context_acquire-aio_context_release-A.patch [RHEL-15965]
- kvm-docs-remove-AioContext-lock-from-IOThread-docs.patch [RHEL-15965]
- kvm-scsi-remove-outdated-AioContext-lock-comment.patch [RHEL-15965]
- kvm-job-remove-outdated-AioContext-locking-comments.patch [RHEL-15965]
- kvm-block-remove-outdated-AioContext-locking-comments.patch [RHEL-15965]
- kvm-block-coroutine-wrapper-use-qemu_get_current_aio_con.patch [RHEL-15965]
- kvm-s390x-pci-avoid-double-enable-disable-of-aif.patch [RHEL-21169]
- kvm-s390x-pci-refresh-fh-before-disabling-aif.patch [RHEL-21169]
- kvm-s390x-pci-drive-ISM-reset-from-subsystem-reset.patch [RHEL-21169]
- kvm-include-ui-rect.h-fix-qemu_rect_init-mis-assignment.patch [RHEL-21570]
- kvm-virtio-gpu-block-migration-of-VMs-with-blob-true.patch [RHEL-7565]
- kvm-spec-Enable-zstd.patch [RHEL-7361]
- Resolves: RHEL-19738
  (Enable properties allowing to disable high memory regions)
- Resolves: RHEL-19302
  (NVIDIA:Grace-Hopper Backport QEMU IOMMUFD Backend)
- Resolves: RHEL-21057
  (Request backport of 9353b6da430f90e47f352dbf6dc31120c8914da6)
- Resolves: RHEL-18212
  ([RHEL9][Secure-execution][s390x] The error message is not clear when boot up a SE guest with wrong encryption)
- Resolves: RHEL-15965
  ( [qemu-kvm] Remove AioContext lock (no response with QMP command block_resize))
- Resolves: RHEL-21169
  ([s390x] VM fails to start with ISM passed through QEMU 8.2)
- Resolves: RHEL-21570
  (Critical performance degradation for input devices in virtio vnc session)
- Resolves: RHEL-7565
  (qemu crashed when migrate guest with blob resources enabled)
- Resolves: RHEL-7361
  ([qemu-kvm] Enable zstd support for qcow2 files)
2024-01-24 04:26:42 -05:00

477 lines
15 KiB
Diff

From 0d8255c98b3ef6f603ff0279592d3e91de26de0e Mon Sep 17 00:00:00 2001
From: Eric Auger <eric.auger@redhat.com>
Date: Tue, 21 Nov 2023 16:44:00 +0800
Subject: [PATCH 021/101] backends/iommufd: Introduce the iommufd object
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
RH-Author: Eric Auger <eric.auger@redhat.com>
RH-MergeRequest: 211: IOMMUFD backend backport
RH-Jira: RHEL-19302 RHEL-21057
RH-Acked-by: Cédric Le Goater <clg@redhat.com>
RH-Acked-by: Sebastian Ott <sebott@redhat.com>
RH-Commit: [20/67] 8a56344ab4a2126f248bfa492ccddd19265f39be (eauger1/centos-qemu-kvm)
Introduce an iommufd object which allows the interaction
with the host /dev/iommu device.
The /dev/iommu can have been already pre-opened outside of qemu,
in which case the fd can be passed directly along with the
iommufd object:
This allows the iommufd object to be shared accross several
subsystems (VFIO, VDPA, ...). For example, libvirt would open
the /dev/iommu once.
If no fd is passed along with the iommufd object, the /dev/iommu
is opened by the qemu code.
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>
(cherry picked from commit 6e6d8ac62b5b38dc9d4b69ffdf073f0a0b43b7be)
Signed-off-by: Eric Auger <eric.auger@redhat.com>
---
MAINTAINERS | 8 ++
backends/Kconfig | 4 +
backends/iommufd.c | 245 +++++++++++++++++++++++++++++++++++++++
backends/meson.build | 1 +
backends/trace-events | 10 ++
include/sysemu/iommufd.h | 38 ++++++
qapi/qom.json | 19 +++
qemu-options.hx | 12 ++
8 files changed, 337 insertions(+)
create mode 100644 backends/iommufd.c
create mode 100644 include/sysemu/iommufd.h
diff --git a/MAINTAINERS b/MAINTAINERS
index 695e0bd34f..a5a446914a 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -2167,6 +2167,14 @@ F: hw/vfio/ap.c
F: docs/system/s390x/vfio-ap.rst
L: qemu-s390x@nongnu.org
+iommufd
+M: Yi Liu <yi.l.liu@intel.com>
+M: Eric Auger <eric.auger@redhat.com>
+M: Zhenzhong Duan <zhenzhong.duan@intel.com>
+S: Supported
+F: backends/iommufd.c
+F: include/sysemu/iommufd.h
+
vhost
M: Michael S. Tsirkin <mst@redhat.com>
S: Supported
diff --git a/backends/Kconfig b/backends/Kconfig
index f35abc1609..2cb23f62fa 100644
--- a/backends/Kconfig
+++ b/backends/Kconfig
@@ -1 +1,5 @@
source tpm/Kconfig
+
+config IOMMUFD
+ bool
+ depends on VFIO
diff --git a/backends/iommufd.c b/backends/iommufd.c
new file mode 100644
index 0000000000..ba58a0eb0d
--- /dev/null
+++ b/backends/iommufd.c
@@ -0,0 +1,245 @@
+/*
+ * 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 "sysemu/iommufd.h"
+#include "qapi/error.h"
+#include "qapi/qmp/qerror.h"
+#include "qemu/module.h"
+#include "qom/object_interfaces.h"
+#include "qemu/error-report.h"
+#include "monitor/monitor.h"
+#include "trace.h"
+#include <sys/ioctl.h>
+#include <linux/iommufd.h>
+
+static void iommufd_backend_init(Object *obj)
+{
+ IOMMUFDBackend *be = IOMMUFD_BACKEND(obj);
+
+ be->fd = -1;
+ be->users = 0;
+ be->owned = true;
+ qemu_mutex_init(&be->lock);
+}
+
+static void iommufd_backend_finalize(Object *obj)
+{
+ IOMMUFDBackend *be = IOMMUFD_BACKEND(obj);
+
+ if (be->owned) {
+ close(be->fd);
+ be->fd = -1;
+ }
+}
+
+static void iommufd_backend_set_fd(Object *obj, const char *str, Error **errp)
+{
+ IOMMUFDBackend *be = IOMMUFD_BACKEND(obj);
+ int fd = -1;
+
+ fd = monitor_fd_param(monitor_cur(), str, errp);
+ if (fd == -1) {
+ error_prepend(errp, "Could not parse remote object fd %s:", str);
+ return;
+ }
+ qemu_mutex_lock(&be->lock);
+ be->fd = fd;
+ be->owned = false;
+ qemu_mutex_unlock(&be->lock);
+ trace_iommu_backend_set_fd(be->fd);
+}
+
+static bool iommufd_backend_can_be_deleted(UserCreatable *uc)
+{
+ IOMMUFDBackend *be = IOMMUFD_BACKEND(uc);
+
+ return !be->users;
+}
+
+static void iommufd_backend_class_init(ObjectClass *oc, void *data)
+{
+ UserCreatableClass *ucc = USER_CREATABLE_CLASS(oc);
+
+ ucc->can_be_deleted = iommufd_backend_can_be_deleted;
+
+ object_class_property_add_str(oc, "fd", NULL, iommufd_backend_set_fd);
+}
+
+int iommufd_backend_connect(IOMMUFDBackend *be, Error **errp)
+{
+ int fd, ret = 0;
+
+ qemu_mutex_lock(&be->lock);
+ if (be->users == UINT32_MAX) {
+ error_setg(errp, "too many connections");
+ ret = -E2BIG;
+ goto out;
+ }
+ if (be->owned && !be->users) {
+ fd = qemu_open_old("/dev/iommu", O_RDWR);
+ if (fd < 0) {
+ error_setg_errno(errp, errno, "/dev/iommu opening failed");
+ ret = fd;
+ goto out;
+ }
+ be->fd = fd;
+ }
+ be->users++;
+out:
+ trace_iommufd_backend_connect(be->fd, be->owned,
+ be->users, ret);
+ qemu_mutex_unlock(&be->lock);
+ return ret;
+}
+
+void iommufd_backend_disconnect(IOMMUFDBackend *be)
+{
+ qemu_mutex_lock(&be->lock);
+ if (!be->users) {
+ goto out;
+ }
+ be->users--;
+ if (!be->users && be->owned) {
+ close(be->fd);
+ be->fd = -1;
+ }
+out:
+ trace_iommufd_backend_disconnect(be->fd, be->users);
+ qemu_mutex_unlock(&be->lock);
+}
+
+int iommufd_backend_alloc_ioas(IOMMUFDBackend *be, uint32_t *ioas_id,
+ Error **errp)
+{
+ int ret, fd = be->fd;
+ struct iommu_ioas_alloc alloc_data = {
+ .size = sizeof(alloc_data),
+ .flags = 0,
+ };
+
+ ret = ioctl(fd, IOMMU_IOAS_ALLOC, &alloc_data);
+ if (ret) {
+ error_setg_errno(errp, errno, "Failed to allocate ioas");
+ return ret;
+ }
+
+ *ioas_id = alloc_data.out_ioas_id;
+ trace_iommufd_backend_alloc_ioas(fd, *ioas_id, ret);
+
+ return ret;
+}
+
+void iommufd_backend_free_id(IOMMUFDBackend *be, uint32_t id)
+{
+ int ret, fd = be->fd;
+ struct iommu_destroy des = {
+ .size = sizeof(des),
+ .id = id,
+ };
+
+ ret = ioctl(fd, IOMMU_DESTROY, &des);
+ trace_iommufd_backend_free_id(fd, id, ret);
+ if (ret) {
+ error_report("Failed to free id: %u %m", id);
+ }
+}
+
+int iommufd_backend_map_dma(IOMMUFDBackend *be, uint32_t ioas_id, hwaddr iova,
+ ram_addr_t size, void *vaddr, bool readonly)
+{
+ int ret, fd = be->fd;
+ struct iommu_ioas_map map = {
+ .size = sizeof(map),
+ .flags = IOMMU_IOAS_MAP_READABLE |
+ IOMMU_IOAS_MAP_FIXED_IOVA,
+ .ioas_id = ioas_id,
+ .__reserved = 0,
+ .user_va = (uintptr_t)vaddr,
+ .iova = iova,
+ .length = size,
+ };
+
+ if (!readonly) {
+ map.flags |= IOMMU_IOAS_MAP_WRITEABLE;
+ }
+
+ ret = ioctl(fd, IOMMU_IOAS_MAP, &map);
+ trace_iommufd_backend_map_dma(fd, ioas_id, iova, size,
+ vaddr, readonly, ret);
+ if (ret) {
+ ret = -errno;
+
+ /* TODO: Not support mapping hardware PCI BAR region for now. */
+ if (errno == EFAULT) {
+ warn_report("IOMMU_IOAS_MAP failed: %m, PCI BAR?");
+ } else {
+ error_report("IOMMU_IOAS_MAP failed: %m");
+ }
+ }
+ return ret;
+}
+
+int iommufd_backend_unmap_dma(IOMMUFDBackend *be, uint32_t ioas_id,
+ hwaddr iova, ram_addr_t size)
+{
+ int ret, fd = be->fd;
+ struct iommu_ioas_unmap unmap = {
+ .size = sizeof(unmap),
+ .ioas_id = ioas_id,
+ .iova = iova,
+ .length = size,
+ };
+
+ ret = ioctl(fd, IOMMU_IOAS_UNMAP, &unmap);
+ /*
+ * IOMMUFD takes mapping as some kind of object, unmapping
+ * nonexistent mapping is treated as deleting a nonexistent
+ * object and return ENOENT. This is different from legacy
+ * backend which allows it. vIOMMU may trigger a lot of
+ * redundant unmapping, to avoid flush the log, treat them
+ * as succeess for IOMMUFD just like legacy backend.
+ */
+ if (ret && errno == ENOENT) {
+ trace_iommufd_backend_unmap_dma_non_exist(fd, ioas_id, iova, size, ret);
+ ret = 0;
+ } else {
+ trace_iommufd_backend_unmap_dma(fd, ioas_id, iova, size, ret);
+ }
+
+ if (ret) {
+ ret = -errno;
+ error_report("IOMMU_IOAS_UNMAP failed: %m");
+ }
+ return ret;
+}
+
+static const TypeInfo iommufd_backend_info = {
+ .name = TYPE_IOMMUFD_BACKEND,
+ .parent = TYPE_OBJECT,
+ .instance_size = sizeof(IOMMUFDBackend),
+ .instance_init = iommufd_backend_init,
+ .instance_finalize = iommufd_backend_finalize,
+ .class_size = sizeof(IOMMUFDBackendClass),
+ .class_init = iommufd_backend_class_init,
+ .interfaces = (InterfaceInfo[]) {
+ { TYPE_USER_CREATABLE },
+ { }
+ }
+};
+
+static void register_types(void)
+{
+ type_register_static(&iommufd_backend_info);
+}
+
+type_init(register_types);
diff --git a/backends/meson.build b/backends/meson.build
index 914c7c4afb..9a5cea480d 100644
--- a/backends/meson.build
+++ b/backends/meson.build
@@ -20,6 +20,7 @@ if have_vhost_user
system_ss.add(when: 'CONFIG_VIRTIO', if_true: files('vhost-user.c'))
endif
system_ss.add(when: 'CONFIG_VIRTIO_CRYPTO', if_true: files('cryptodev-vhost.c'))
+system_ss.add(when: 'CONFIG_IOMMUFD', if_true: files('iommufd.c'))
if have_vhost_user_crypto
system_ss.add(when: 'CONFIG_VIRTIO_CRYPTO', if_true: files('cryptodev-vhost-user.c'))
endif
diff --git a/backends/trace-events b/backends/trace-events
index 652eb76a57..d45c6e31a6 100644
--- a/backends/trace-events
+++ b/backends/trace-events
@@ -5,3 +5,13 @@ dbus_vmstate_pre_save(void)
dbus_vmstate_post_load(int version_id) "version_id: %d"
dbus_vmstate_loading(const char *id) "id: %s"
dbus_vmstate_saving(const char *id) "id: %s"
+
+# iommufd.c
+iommufd_backend_connect(int fd, bool owned, uint32_t users, int ret) "fd=%d owned=%d users=%d (%d)"
+iommufd_backend_disconnect(int fd, uint32_t users) "fd=%d users=%d"
+iommu_backend_set_fd(int fd) "pre-opened /dev/iommu fd=%d"
+iommufd_backend_map_dma(int iommufd, uint32_t ioas, uint64_t iova, uint64_t size, void *vaddr, bool readonly, int ret) " iommufd=%d ioas=%d iova=0x%"PRIx64" size=0x%"PRIx64" addr=%p readonly=%d (%d)"
+iommufd_backend_unmap_dma_non_exist(int iommufd, uint32_t ioas, uint64_t iova, uint64_t size, int ret) " Unmap nonexistent mapping: iommufd=%d ioas=%d iova=0x%"PRIx64" size=0x%"PRIx64" (%d)"
+iommufd_backend_unmap_dma(int iommufd, uint32_t ioas, uint64_t iova, uint64_t size, int ret) " iommufd=%d ioas=%d iova=0x%"PRIx64" size=0x%"PRIx64" (%d)"
+iommufd_backend_alloc_ioas(int iommufd, uint32_t ioas, int ret) " iommufd=%d ioas=%d (%d)"
+iommufd_backend_free_id(int iommufd, uint32_t id, int ret) " iommufd=%d id=%d (%d)"
diff --git a/include/sysemu/iommufd.h b/include/sysemu/iommufd.h
new file mode 100644
index 0000000000..9c5524b0ed
--- /dev/null
+++ b/include/sysemu/iommufd.h
@@ -0,0 +1,38 @@
+#ifndef SYSEMU_IOMMUFD_H
+#define SYSEMU_IOMMUFD_H
+
+#include "qom/object.h"
+#include "qemu/thread.h"
+#include "exec/hwaddr.h"
+#include "exec/cpu-common.h"
+
+#define TYPE_IOMMUFD_BACKEND "iommufd"
+OBJECT_DECLARE_TYPE(IOMMUFDBackend, IOMMUFDBackendClass, IOMMUFD_BACKEND)
+
+struct IOMMUFDBackendClass {
+ ObjectClass parent_class;
+};
+
+struct IOMMUFDBackend {
+ Object parent;
+
+ /*< protected >*/
+ int fd; /* /dev/iommu file descriptor */
+ bool owned; /* is the /dev/iommu opened internally */
+ QemuMutex lock;
+ uint32_t users;
+
+ /*< public >*/
+};
+
+int iommufd_backend_connect(IOMMUFDBackend *be, Error **errp);
+void iommufd_backend_disconnect(IOMMUFDBackend *be);
+
+int iommufd_backend_alloc_ioas(IOMMUFDBackend *be, uint32_t *ioas_id,
+ Error **errp);
+void iommufd_backend_free_id(IOMMUFDBackend *be, uint32_t id);
+int iommufd_backend_map_dma(IOMMUFDBackend *be, uint32_t ioas_id, hwaddr iova,
+ ram_addr_t size, void *vaddr, bool readonly);
+int iommufd_backend_unmap_dma(IOMMUFDBackend *be, uint32_t ioas_id,
+ hwaddr iova, ram_addr_t size);
+#endif
diff --git a/qapi/qom.json b/qapi/qom.json
index c53ef978ff..95516ba325 100644
--- a/qapi/qom.json
+++ b/qapi/qom.json
@@ -794,6 +794,23 @@
{ 'struct': 'VfioUserServerProperties',
'data': { 'socket': 'SocketAddress', 'device': 'str' } }
+##
+# @IOMMUFDProperties:
+#
+# Properties for iommufd objects.
+#
+# @fd: file descriptor name previously passed via 'getfd' command,
+# which represents a pre-opened /dev/iommu. This allows the
+# iommufd object to be shared accross several subsystems
+# (VFIO, VDPA, ...), and the file descriptor to be shared
+# with other process, e.g. DPDK. (default: QEMU opens
+# /dev/iommu by itself)
+#
+# Since: 9.0
+##
+{ 'struct': 'IOMMUFDProperties',
+ 'data': { '*fd': 'str' } }
+
##
# @RngProperties:
#
@@ -934,6 +951,7 @@
'input-barrier',
{ 'name': 'input-linux',
'if': 'CONFIG_LINUX' },
+ 'iommufd',
'iothread',
'main-loop',
{ 'name': 'memory-backend-epc',
@@ -1003,6 +1021,7 @@
'input-barrier': 'InputBarrierProperties',
'input-linux': { 'type': 'InputLinuxProperties',
'if': 'CONFIG_LINUX' },
+ 'iommufd': 'IOMMUFDProperties',
'iothread': 'IothreadProperties',
'main-loop': 'MainLoopProperties',
'memory-backend-epc': { 'type': 'MemoryBackendEpcProperties',
diff --git a/qemu-options.hx b/qemu-options.hx
index 557118cb1f..0814f43066 100644
--- a/qemu-options.hx
+++ b/qemu-options.hx
@@ -5224,6 +5224,18 @@ SRST
The ``share`` boolean option is on by default with memfd.
+ ``-object iommufd,id=id[,fd=fd]``
+ Creates an iommufd backend which allows control of DMA mapping
+ through the ``/dev/iommu`` device.
+
+ The ``id`` parameter is a unique ID which frontends (such as
+ vfio-pci of vdpa) will use to connect with the iommufd backend.
+
+ The ``fd`` parameter is an optional pre-opened file descriptor
+ resulting from ``/dev/iommu`` opening. Usually the iommufd is shared
+ across all subsystems, bringing the benefit of centralized
+ reference counting.
+
``-object rng-builtin,id=id``
Creates a random number generator backend which obtains entropy
from QEMU builtin functions. The ``id`` parameter is a unique ID
--
2.39.3