From: Xu Yilun <yilun.xu@linux.intel.com>
To: jgg@nvidia.com, jgg@ziepe.ca, kevin.tian@intel.com,
will@kernel.org, aneesh.kumar@kernel.org
Cc: iommu@lists.linux.dev, linux-kernel@vger.kernel.org,
joro@8bytes.org, robin.murphy@arm.com, shuah@kernel.org,
nicolinc@nvidia.com, aik@amd.com, dan.j.williams@intel.com,
baolu.lu@linux.intel.com, yilun.xu@intel.com
Subject: [PATCH v4 4/7] iommufd: Destroy vdevice on idevice destroy
Date: Wed, 9 Jul 2025 12:02:31 +0800 [thread overview]
Message-ID: <20250709040234.1773573-5-yilun.xu@linux.intel.com> (raw)
In-Reply-To: <20250709040234.1773573-1-yilun.xu@linux.intel.com>
Destroy iommufd_vdevice (vdev) on iommufd_idevice (idev) destruction so
that vdev can't outlive idev.
idev represents the physical device bound to iommufd, while the vdev
represents the virtual instance of the physical device in the VM. The
lifecycle of the vdev should not be longer than idev. This doesn't
cause real problem on existing use cases cause vdev doesn't impact the
physical device, only provides virtualization information. But to
extend vdev for Confidential Computing (CC), there are needs to do
secure configuration for the vdev, e.g. TSM Bind/Unbind. These
configurations should be rolled back on idev destroy, or the external
driver (VFIO) functionality may be impact.
The idev is created by external driver so its destruction can't fail.
The idev implements pre_destroy() op to actively remove its associated
vdev before destroying itself. There are 3 cases on idev pre_destroy():
1. vdev is already destroyed by userspace. No extra handling needed.
2. vdev is still alive. Use iommufd_object_tombstone_user() to
destroy vdev and tombstone the vdev ID.
3. vdev is being destroyed by userspace. The vdev ID is already
freed, but vdev destroy handler is not completed. This requires
multi-threads syncing - vdev holds idev's short term users
reference until vdev destruction completes, idev leverages
existing wait_shortterm mechanism for syncing.
Originally-by: Nicolin Chen <nicolinc@nvidia.com>
Suggested-by: Jason Gunthorpe <jgg@nvidia.com>
Reviewed-by: Lu Baolu <baolu.lu@linux.intel.com>
Co-developed-by: Aneesh Kumar K.V (Arm) <aneesh.kumar@kernel.org>
Signed-off-by: Aneesh Kumar K.V (Arm) <aneesh.kumar@kernel.org>
Signed-off-by: Xu Yilun <yilun.xu@linux.intel.com>
---
drivers/iommu/iommufd/device.c | 51 +++++++++++++++++++++++++
drivers/iommu/iommufd/iommufd_private.h | 13 +++++++
drivers/iommu/iommufd/main.c | 2 +
drivers/iommu/iommufd/viommu.c | 51 +++++++++++++++++++++++--
4 files changed, 113 insertions(+), 4 deletions(-)
diff --git a/drivers/iommu/iommufd/device.c b/drivers/iommu/iommufd/device.c
index e9b6ca47095c..e114094fbdef 100644
--- a/drivers/iommu/iommufd/device.c
+++ b/drivers/iommu/iommufd/device.c
@@ -137,6 +137,57 @@ static struct iommufd_group *iommufd_get_group(struct iommufd_ctx *ictx,
}
}
+static void iommufd_device_remove_vdev(struct iommufd_device *idev)
+{
+ struct iommufd_vdevice *vdev;
+
+ mutex_lock(&idev->igroup->lock);
+ /* prevent new references from vdev */
+ idev->destroying = true;
+ /* vdev has been completely destroyed by userspace */
+ if (!idev->vdev)
+ goto out_unlock;
+
+ vdev = iommufd_get_vdevice(idev->ictx, idev->vdev->obj.id);
+ /*
+ * An ongoing vdev destroy ioctl has removed the vdev from the object
+ * xarray, but has not finished iommufd_vdevice_destroy() yet as it
+ * needs the same mutex. We exit the locking then wait on short term
+ * users for the vdev destruction.
+ */
+ if (IS_ERR(vdev))
+ goto out_unlock;
+
+ /* Should never happen */
+ if (WARN_ON(vdev != idev->vdev)) {
+ iommufd_put_object(idev->ictx, &vdev->obj);
+ goto out_unlock;
+ }
+
+ /*
+ * vdev is still alive. Hold a users refcount to prevent racing with
+ * userspace destruction, then use iommufd_object_tombstone_user() to
+ * destroy it and leave a tombstone.
+ */
+ refcount_inc(&vdev->obj.users);
+ iommufd_put_object(idev->ictx, &vdev->obj);
+ mutex_unlock(&idev->igroup->lock);
+ iommufd_object_tombstone_user(idev->ictx, &vdev->obj);
+ return;
+
+out_unlock:
+ mutex_unlock(&idev->igroup->lock);
+}
+
+void iommufd_device_pre_destroy(struct iommufd_object *obj)
+{
+ struct iommufd_device *idev =
+ container_of(obj, struct iommufd_device, obj);
+
+ /* Release the short term users on this */
+ iommufd_device_remove_vdev(idev);
+}
+
void iommufd_device_destroy(struct iommufd_object *obj)
{
struct iommufd_device *idev =
diff --git a/drivers/iommu/iommufd/iommufd_private.h b/drivers/iommu/iommufd/iommufd_private.h
index da1bced8c945..62e5dae2a50d 100644
--- a/drivers/iommu/iommufd/iommufd_private.h
+++ b/drivers/iommu/iommufd/iommufd_private.h
@@ -475,6 +475,8 @@ struct iommufd_device {
/* always the physical device */
struct device *dev;
bool enforce_cache_coherency;
+ struct iommufd_vdevice *vdev;
+ bool destroying;
};
static inline struct iommufd_device *
@@ -485,6 +487,7 @@ iommufd_get_device(struct iommufd_ucmd *ucmd, u32 id)
struct iommufd_device, obj);
}
+void iommufd_device_pre_destroy(struct iommufd_object *obj);
void iommufd_device_destroy(struct iommufd_object *obj);
int iommufd_get_hw_info(struct iommufd_ucmd *ucmd);
@@ -650,14 +653,24 @@ int iommufd_viommu_alloc_ioctl(struct iommufd_ucmd *ucmd);
void iommufd_viommu_destroy(struct iommufd_object *obj);
int iommufd_vdevice_alloc_ioctl(struct iommufd_ucmd *ucmd);
void iommufd_vdevice_destroy(struct iommufd_object *obj);
+void iommufd_vdevice_abort(struct iommufd_object *obj);
struct iommufd_vdevice {
struct iommufd_object obj;
struct iommufd_viommu *viommu;
struct device *dev;
u64 id; /* per-vIOMMU virtual ID */
+ struct iommufd_device *idev;
};
+static inline struct iommufd_vdevice *
+iommufd_get_vdevice(struct iommufd_ctx *ictx, u32 id)
+{
+ return container_of(iommufd_get_object(ictx, id,
+ IOMMUFD_OBJ_VDEVICE),
+ struct iommufd_vdevice, obj);
+}
+
#ifdef CONFIG_IOMMUFD_TEST
int iommufd_test(struct iommufd_ucmd *ucmd);
void iommufd_selftest_destroy(struct iommufd_object *obj);
diff --git a/drivers/iommu/iommufd/main.c b/drivers/iommu/iommufd/main.c
index e91a36cc02d0..bea000eed14a 100644
--- a/drivers/iommu/iommufd/main.c
+++ b/drivers/iommu/iommufd/main.c
@@ -578,6 +578,7 @@ static const struct iommufd_object_ops iommufd_object_ops[] = {
.destroy = iommufd_access_destroy_object,
},
[IOMMUFD_OBJ_DEVICE] = {
+ .pre_destroy = iommufd_device_pre_destroy,
.destroy = iommufd_device_destroy,
},
[IOMMUFD_OBJ_FAULT] = {
@@ -596,6 +597,7 @@ static const struct iommufd_object_ops iommufd_object_ops[] = {
},
[IOMMUFD_OBJ_VDEVICE] = {
.destroy = iommufd_vdevice_destroy,
+ .abort = iommufd_vdevice_abort,
},
[IOMMUFD_OBJ_VEVENTQ] = {
.destroy = iommufd_veventq_destroy,
diff --git a/drivers/iommu/iommufd/viommu.c b/drivers/iommu/iommufd/viommu.c
index f556a65f3062..e6009678c8a5 100644
--- a/drivers/iommu/iommufd/viommu.c
+++ b/drivers/iommu/iommufd/viommu.c
@@ -104,18 +104,34 @@ int iommufd_viommu_alloc_ioctl(struct iommufd_ucmd *ucmd)
return rc;
}
-void iommufd_vdevice_destroy(struct iommufd_object *obj)
+void iommufd_vdevice_abort(struct iommufd_object *obj)
{
struct iommufd_vdevice *vdev =
container_of(obj, struct iommufd_vdevice, obj);
struct iommufd_viommu *viommu = vdev->viommu;
+ struct iommufd_device *idev = vdev->idev;
+
+ lockdep_assert_held(&idev->igroup->lock);
/* xa_cmpxchg is okay to fail if alloc failed xa_cmpxchg previously */
xa_cmpxchg(&viommu->vdevs, vdev->id, vdev, NULL, GFP_KERNEL);
refcount_dec(&viommu->obj.users);
+ idev->vdev = NULL;
put_device(vdev->dev);
}
+void iommufd_vdevice_destroy(struct iommufd_object *obj)
+{
+ struct iommufd_vdevice *vdev =
+ container_of(obj, struct iommufd_vdevice, obj);
+ struct iommufd_device *idev = vdev->idev;
+
+ mutex_lock(&idev->igroup->lock);
+ iommufd_vdevice_abort(obj);
+ mutex_unlock(&idev->igroup->lock);
+ iommufd_put_object(idev->ictx, &idev->obj);
+}
+
int iommufd_vdevice_alloc_ioctl(struct iommufd_ucmd *ucmd)
{
struct iommu_vdevice_alloc *cmd = ucmd->cmd;
@@ -144,10 +160,21 @@ int iommufd_vdevice_alloc_ioctl(struct iommufd_ucmd *ucmd)
goto out_put_idev;
}
+ mutex_lock(&idev->igroup->lock);
+ if (idev->destroying) {
+ rc = -ENOENT;
+ goto out_unlock_igroup;
+ }
+
+ if (idev->vdev) {
+ rc = -EEXIST;
+ goto out_unlock_igroup;
+ }
+
vdev = iommufd_object_alloc(ucmd->ictx, vdev, IOMMUFD_OBJ_VDEVICE);
if (IS_ERR(vdev)) {
rc = PTR_ERR(vdev);
- goto out_put_idev;
+ goto out_unlock_igroup;
}
vdev->id = virt_id;
@@ -155,6 +182,19 @@ int iommufd_vdevice_alloc_ioctl(struct iommufd_ucmd *ucmd)
get_device(idev->dev);
vdev->viommu = viommu;
refcount_inc(&viommu->obj.users);
+ /*
+ * A short term users reference is held on the idev so long as we have
+ * the pointer. iommufd_device_pre_destroy() will revoke it before the
+ * idev real destruction.
+ */
+ vdev->idev = idev;
+
+ /*
+ * iommufd_device_destroy() delays until idev->vdev is NULL before
+ * freeing the idev, which only happens once the vdev is finished
+ * destruction.
+ */
+ idev->vdev = vdev;
curr = xa_cmpxchg(&viommu->vdevs, virt_id, NULL, vdev, GFP_KERNEL);
if (curr) {
@@ -167,12 +207,15 @@ int iommufd_vdevice_alloc_ioctl(struct iommufd_ucmd *ucmd)
if (rc)
goto out_abort;
iommufd_object_finalize(ucmd->ictx, &vdev->obj);
- goto out_put_idev;
+ goto out_unlock_igroup;
out_abort:
iommufd_object_abort_and_destroy(ucmd->ictx, &vdev->obj);
+out_unlock_igroup:
+ mutex_unlock(&idev->igroup->lock);
out_put_idev:
- iommufd_put_object(ucmd->ictx, &idev->obj);
+ if (rc)
+ iommufd_put_object(ucmd->ictx, &idev->obj);
out_put_viommu:
iommufd_put_object(ucmd->ictx, &viommu->obj);
return rc;
--
2.25.1
next prev parent reply other threads:[~2025-07-09 4:11 UTC|newest]
Thread overview: 38+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-07-09 4:02 [PATCH v4 0/7] iommufd: Destroy vdevice on device unbind Xu Yilun
2025-07-09 4:02 ` [PATCH v4 1/7] iommufd/viommu: Roll back to use iommufd_object_alloc() for vdevice Xu Yilun
2025-07-10 7:32 ` Tian, Kevin
2025-07-10 18:26 ` Nicolin Chen
2025-07-11 17:42 ` Jason Gunthorpe
2025-07-09 4:02 ` [PATCH v4 2/7] iommufd: Add iommufd_object_tombstone_user() helper Xu Yilun
2025-07-10 7:35 ` Tian, Kevin
2025-07-10 19:08 ` Nicolin Chen
2025-07-11 15:02 ` Xu Yilun
2025-07-11 19:39 ` Nicolin Chen
2025-07-11 21:33 ` Jason Gunthorpe
2025-07-11 22:04 ` Nicolin Chen
2025-07-11 17:44 ` Jason Gunthorpe
2025-07-09 4:02 ` [PATCH v4 3/7] iommufd: Add a pre_destroy() op for objects Xu Yilun
2025-07-10 7:40 ` Tian, Kevin
2025-07-10 17:15 ` Jason Gunthorpe
2025-07-11 3:16 ` Tian, Kevin
2025-07-11 9:38 ` Xu Yilun
2025-07-11 22:09 ` Nicolin Chen
2025-07-09 4:02 ` Xu Yilun [this message]
2025-07-10 8:03 ` [PATCH v4 4/7] iommufd: Destroy vdevice on idevice destroy Tian, Kevin
2025-07-12 2:22 ` Nicolin Chen
2025-07-12 17:47 ` Xu Yilun
2025-07-14 16:40 ` Xu Yilun
2025-07-14 16:53 ` Jason Gunthorpe
2025-07-14 17:34 ` Nicolin Chen
2025-07-14 18:03 ` Xu Yilun
2025-07-12 2:26 ` Nicolin Chen
2025-07-09 4:02 ` [PATCH v4 5/7] iommufd/vdevice: Remove struct device reference from struct vdevice Xu Yilun
2025-07-10 8:04 ` Tian, Kevin
2025-07-12 2:27 ` Nicolin Chen
2025-07-09 4:02 ` [PATCH v4 6/7] iommufd/selftest: Explicitly skip tests for inapplicable variant Xu Yilun
2025-07-10 8:06 ` Tian, Kevin
2025-07-12 2:39 ` Nicolin Chen
2025-07-09 4:02 ` [PATCH v4 7/7] iommufd/selftest: Add coverage for vdevice tombstone Xu Yilun
2025-07-10 8:10 ` Tian, Kevin
2025-07-10 9:25 ` Xu Yilun
2025-07-11 3:12 ` Tian, Kevin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20250709040234.1773573-5-yilun.xu@linux.intel.com \
--to=yilun.xu@linux.intel.com \
--cc=aik@amd.com \
--cc=aneesh.kumar@kernel.org \
--cc=baolu.lu@linux.intel.com \
--cc=dan.j.williams@intel.com \
--cc=iommu@lists.linux.dev \
--cc=jgg@nvidia.com \
--cc=jgg@ziepe.ca \
--cc=joro@8bytes.org \
--cc=kevin.tian@intel.com \
--cc=linux-kernel@vger.kernel.org \
--cc=nicolinc@nvidia.com \
--cc=robin.murphy@arm.com \
--cc=shuah@kernel.org \
--cc=will@kernel.org \
--cc=yilun.xu@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®