From: Jason Gunthorpe <jgg@nvidia.com>
To: Alex Williamson <alex.williamson@redhat.com>,
Cornelia Huck <cohuck@redhat.com>,
kvm@vger.kernel.org
Cc: Christoph Hellwig <hch@lst.de>,
"Tian, Kevin" <kevin.tian@intel.com>,
Liu Yi L <yi.l.liu@intel.com>
Subject: [PATCH v3 4/5] vfio: Use a refcount_t instead of a kref in the vfio_group
Date: Fri, 15 Oct 2021 08:40:53 -0300 [thread overview]
Message-ID: <4-v3-2fdfe4ca2cc6+18c-vfio_group_cdev_jgg@nvidia.com> (raw)
In-Reply-To: <0-v3-2fdfe4ca2cc6+18c-vfio_group_cdev_jgg@nvidia.com>
The next patch adds a struct device to the struct vfio_group, and it is
confusing/bad practice to have two krefs in the same struct. This kref is
controlling the period when the vfio_group is registered in sysfs, and
visible in the internal lookup. Switch it to a refcount_t instead.
The refcount_dec_and_mutex_lock() is still required because we need
atomicity of the list searches and sysfs presence.
Reviewed-by: Liu Yi L <yi.l.liu@intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
---
drivers/vfio/vfio.c | 21 +++++++++------------
1 file changed, 9 insertions(+), 12 deletions(-)
diff --git a/drivers/vfio/vfio.c b/drivers/vfio/vfio.c
index 4abb2e5e196536..e313fa030b9185 100644
--- a/drivers/vfio/vfio.c
+++ b/drivers/vfio/vfio.c
@@ -69,7 +69,7 @@ struct vfio_unbound_dev {
};
struct vfio_group {
- struct kref kref;
+ refcount_t users;
int minor;
atomic_t container_users;
struct iommu_group *iommu_group;
@@ -377,7 +377,7 @@ static struct vfio_group *vfio_create_group(struct iommu_group *iommu_group,
if (!group)
return ERR_PTR(-ENOMEM);
- kref_init(&group->kref);
+ refcount_set(&group->users, 1);
INIT_LIST_HEAD(&group->device_list);
mutex_init(&group->device_lock);
INIT_LIST_HEAD(&group->unbound_list);
@@ -433,10 +433,10 @@ static struct vfio_group *vfio_create_group(struct iommu_group *iommu_group,
return group;
}
-/* called with vfio.group_lock held */
-static void vfio_group_release(struct kref *kref)
+static void vfio_group_put(struct vfio_group *group)
{
- struct vfio_group *group = container_of(kref, struct vfio_group, kref);
+ if (!refcount_dec_and_mutex_lock(&group->users, &vfio.group_lock))
+ return;
/*
* These data structures all have paired operations that can only be
@@ -454,15 +454,9 @@ static void vfio_group_release(struct kref *kref)
vfio_group_unlock_and_free(group);
}
-static void vfio_group_put(struct vfio_group *group)
-{
- kref_put_mutex(&group->kref, vfio_group_release, &vfio.group_lock);
-}
-
-/* Assume group_lock or group reference is held */
static void vfio_group_get(struct vfio_group *group)
{
- kref_get(&group->kref);
+ refcount_inc(&group->users);
}
static struct vfio_group *vfio_group_get_from_minor(int minor)
@@ -1657,6 +1651,9 @@ struct vfio_group *vfio_group_get_external_user(struct file *filep)
if (ret)
return ERR_PTR(ret);
+ /*
+ * Since the caller holds the fget on the file group->users must be >= 1
+ */
vfio_group_get(group);
return group;
--
2.33.0
next prev parent reply other threads:[~2021-10-15 11:41 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-10-15 11:40 [PATCH v3 0/5] Update vfio_group to use the modern cdev lifecycle Jason Gunthorpe
2021-10-15 11:40 ` [PATCH v3 1/5] vfio: Delete vfio_get/put_group from vfio_iommu_group_notifier() Jason Gunthorpe
2021-10-15 11:40 ` [PATCH v3 2/5] vfio: Do not open code the group list search in vfio_create_group() Jason Gunthorpe
2021-10-15 11:40 ` [PATCH v3 3/5] vfio: Don't leak a group reference if the group already exists Jason Gunthorpe
2021-10-15 11:40 ` Jason Gunthorpe [this message]
2021-10-15 11:40 ` [PATCH v3 5/5] vfio: Use cdev_device_add() instead of device_create() Jason Gunthorpe
2021-10-15 13:20 ` [PATCH v3 0/5] Update vfio_group to use the modern cdev lifecycle Christoph Hellwig
2021-10-20 19:52 ` Alex Williamson
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=4-v3-2fdfe4ca2cc6+18c-vfio_group_cdev_jgg@nvidia.com \
--to=jgg@nvidia.com \
--cc=alex.williamson@redhat.com \
--cc=cohuck@redhat.com \
--cc=hch@lst.de \
--cc=kevin.tian@intel.com \
--cc=kvm@vger.kernel.org \
--cc=yi.l.liu@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;
as well as URLs for NNTP newsgroup(s).