From: Jason Gunthorpe <jgg@nvidia.com>
Date: Wed, 13 Apr 2022 10:10:36 -0300
Subject: vfio/pci: Fix vf_token mechanism when device-specific VF drivers are
 used
Git-commit: 1ef3342a934e235aca72b4bcc0d6854d80a65077
Patch-mainline: v5.18-rc3
References: bsc#1205701

get_pf_vdev() tries to check if a PF is a VFIO PF by looking at the driver:

       if (pci_dev_driver(physfn) != pci_dev_driver(vdev->pdev)) {

However now that we have multiple VF and PF drivers this is no longer
reliable.

This means that security tests realted to vf_token can be skipped by
mixing and matching different VFIO PCI drivers.

Instead of trying to use the driver core to find the PF devices maintain a
linked list of all PF vfio_pci_core_device's that we have called
pci_enable_sriov() on.

When registering a VF just search the list to see if the PF is present and
record the match permanently in the struct. PCI core locking prevents a PF
from passing pci_disable_sriov() while VF drivers are attached so the VFIO
owned PF becomes a static property of the VF.

In common cases where vfio does not own the PF the global list remains
empty and the VF's pointer is statically NULL.

This also fixes a lockdep splat from recursive locking of the
vfio_group::device_lock between vfio_device_get_from_name() and
vfio_device_get_from_dev(). If the VF and PF share the same group this
would deadlock.

Fixes: ff53edf6d6ab ("vfio/pci: Split the pci_driver code out of vfio_pci_core.c")
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Link: https://lore.kernel.org/r/0-v3-876570980634+f2e8-vfio_vf_token_jgg@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Acked-by: Joerg Roedel <jroedel@suse.de>
---
 drivers/vfio/pci/vfio_pci_core.c | 124 +++++++++++++++++++++++----------------
 include/linux/vfio_pci_core.h    |   2 +
 2 files changed, 76 insertions(+), 50 deletions(-)

diff --git a/drivers/vfio/pci/vfio_pci_core.c b/drivers/vfio/pci/vfio_pci_core.c
index b7bb16f92ac6..06b6f3594a13 100644
--- a/drivers/vfio/pci/vfio_pci_core.c
+++ b/drivers/vfio/pci/vfio_pci_core.c
@@ -36,6 +36,10 @@ static bool nointxmask;
 static bool disable_vga;
 static bool disable_idle_d3;
 
+/* List of PF's that vfio_pci_core_sriov_configure() has been called on */
+static DEFINE_MUTEX(vfio_pci_sriov_pfs_mutex);
+static LIST_HEAD(vfio_pci_sriov_pfs);
+
 static inline bool vfio_vga_disabled(void)
 {
 #ifdef CONFIG_VFIO_PCI_VGA
@@ -434,47 +438,17 @@ void vfio_pci_core_disable(struct vfio_pci_core_device *vdev)
 }
 EXPORT_SYMBOL_GPL(vfio_pci_core_disable);
 
-static struct vfio_pci_core_device *get_pf_vdev(struct vfio_pci_core_device *vdev)
-{
-	struct pci_dev *physfn = pci_physfn(vdev->pdev);
-	struct vfio_device *pf_dev;
-
-	if (!vdev->pdev->is_virtfn)
-		return NULL;
-
-	pf_dev = vfio_device_get_from_dev(&physfn->dev);
-	if (!pf_dev)
-		return NULL;
-
-	if (pci_dev_driver(physfn) != pci_dev_driver(vdev->pdev)) {
-		vfio_device_put(pf_dev);
-		return NULL;
-	}
-
-	return container_of(pf_dev, struct vfio_pci_core_device, vdev);
-}
-
-static void vfio_pci_vf_token_user_add(struct vfio_pci_core_device *vdev, int val)
-{
-	struct vfio_pci_core_device *pf_vdev = get_pf_vdev(vdev);
-
-	if (!pf_vdev)
-		return;
-
-	mutex_lock(&pf_vdev->vf_token->lock);
-	pf_vdev->vf_token->users += val;
-	WARN_ON(pf_vdev->vf_token->users < 0);
-	mutex_unlock(&pf_vdev->vf_token->lock);
-
-	vfio_device_put(&pf_vdev->vdev);
-}
-
 void vfio_pci_core_close_device(struct vfio_device *core_vdev)
 {
 	struct vfio_pci_core_device *vdev =
 		container_of(core_vdev, struct vfio_pci_core_device, vdev);
 
-	vfio_pci_vf_token_user_add(vdev, -1);
+	if (vdev->sriov_pf_core_dev) {
+		mutex_lock(&vdev->sriov_pf_core_dev->vf_token->lock);
+		WARN_ON(!vdev->sriov_pf_core_dev->vf_token->users);
+		vdev->sriov_pf_core_dev->vf_token->users--;
+		mutex_unlock(&vdev->sriov_pf_core_dev->vf_token->lock);
+	}
 	vfio_spapr_pci_eeh_release(vdev->pdev);
 	vfio_pci_core_disable(vdev);
 
@@ -495,7 +469,12 @@ void vfio_pci_core_finish_enable(struct vfio_pci_core_device *vdev)
 {
 	vfio_pci_probe_mmaps(vdev);
 	vfio_spapr_pci_eeh_open(vdev->pdev);
-	vfio_pci_vf_token_user_add(vdev, 1);
+
+	if (vdev->sriov_pf_core_dev) {
+		mutex_lock(&vdev->sriov_pf_core_dev->vf_token->lock);
+		vdev->sriov_pf_core_dev->vf_token->users++;
+		mutex_unlock(&vdev->sriov_pf_core_dev->vf_token->lock);
+	}
 }
 EXPORT_SYMBOL_GPL(vfio_pci_core_finish_enable);
 
@@ -1583,11 +1562,8 @@ static int vfio_pci_validate_vf_token(struct vfio_pci_core_device *vdev,
 	 *
 	 * If the VF token is provided but unused, an error is generated.
 	 */
-	if (!vdev->pdev->is_virtfn && !vdev->vf_token && !vf_token)
-		return 0; /* No VF token provided or required */
-
 	if (vdev->pdev->is_virtfn) {
-		struct vfio_pci_core_device *pf_vdev = get_pf_vdev(vdev);
+		struct vfio_pci_core_device *pf_vdev = vdev->sriov_pf_core_dev;
 		bool match;
 
 		if (!pf_vdev) {
@@ -1600,7 +1576,6 @@ static int vfio_pci_validate_vf_token(struct vfio_pci_core_device *vdev,
 		}
 
 		if (!vf_token) {
-			vfio_device_put(&pf_vdev->vdev);
 			pci_info_ratelimited(vdev->pdev,
 				"VF token required to access device\n");
 			return -EACCES;
@@ -1610,8 +1585,6 @@ static int vfio_pci_validate_vf_token(struct vfio_pci_core_device *vdev,
 		match = uuid_equal(uuid, &pf_vdev->vf_token->uuid);
 		mutex_unlock(&pf_vdev->vf_token->lock);
 
-		vfio_device_put(&pf_vdev->vdev);
-
 		if (!match) {
 			pci_info_ratelimited(vdev->pdev,
 				"Incorrect VF token provided for device\n");
@@ -1732,8 +1705,30 @@ static int vfio_pci_bus_notifier(struct notifier_block *nb,
 static int vfio_pci_vf_init(struct vfio_pci_core_device *vdev)
 {
 	struct pci_dev *pdev = vdev->pdev;
+	struct vfio_pci_core_device *cur;
+	struct pci_dev *physfn;
 	int ret;
 
+	if (pdev->is_virtfn) {
+		/*
+		 * If this VF was created by our vfio_pci_core_sriov_configure()
+		 * then we can find the PF vfio_pci_core_device now, and due to
+		 * the locking in pci_disable_sriov() it cannot change until
+		 * this VF device driver is removed.
+		 */
+		physfn = pci_physfn(vdev->pdev);
+		mutex_lock(&vfio_pci_sriov_pfs_mutex);
+		list_for_each_entry(cur, &vfio_pci_sriov_pfs, sriov_pfs_item) {
+			if (cur->pdev == physfn) {
+				vdev->sriov_pf_core_dev = cur;
+				break;
+			}
+		}
+		mutex_unlock(&vfio_pci_sriov_pfs_mutex);
+		return 0;
+	}
+
+	/* Not a SRIOV PF */
 	if (!pdev->is_physfn)
 		return 0;
 
@@ -1805,6 +1800,7 @@ void vfio_pci_core_init_device(struct vfio_pci_core_device *vdev,
 	INIT_LIST_HEAD(&vdev->ioeventfds_list);
 	mutex_init(&vdev->vma_lock);
 	INIT_LIST_HEAD(&vdev->vma_list);
+	INIT_LIST_HEAD(&vdev->sriov_pfs_item);
 	init_rwsem(&vdev->memory_lock);
 }
 EXPORT_SYMBOL_GPL(vfio_pci_core_init_device);
@@ -1896,7 +1892,7 @@ void vfio_pci_core_unregister_device(struct vfio_pci_core_device *vdev)
 {
 	struct pci_dev *pdev = vdev->pdev;
 
-	pci_disable_sriov(pdev);
+	vfio_pci_core_sriov_configure(pdev, 0);
 
 	vfio_unregister_group_dev(&vdev->vdev);
 
@@ -1935,21 +1931,49 @@ EXPORT_SYMBOL_GPL(vfio_pci_core_aer_err_detected);
 
 int vfio_pci_core_sriov_configure(struct pci_dev *pdev, int nr_virtfn)
 {
+	struct vfio_pci_core_device *vdev;
 	struct vfio_device *device;
 	int ret = 0;
 
+	device_lock_assert(&pdev->dev);
+
 	device = vfio_device_get_from_dev(&pdev->dev);
 	if (!device)
 		return -ENODEV;
 
-	if (nr_virtfn == 0)
-		pci_disable_sriov(pdev);
-	else
+	vdev = container_of(device, struct vfio_pci_core_device, vdev);
+
+	if (nr_virtfn) {
+		mutex_lock(&vfio_pci_sriov_pfs_mutex);
+		/*
+		 * The thread that adds the vdev to the list is the only thread
+		 * that gets to call pci_enable_sriov() and we will only allow
+		 * it to be called once without going through
+		 * pci_disable_sriov()
+		 */
+		if (!list_empty(&vdev->sriov_pfs_item)) {
+			ret = -EINVAL;
+			goto out_unlock;
+		}
+		list_add_tail(&vdev->sriov_pfs_item, &vfio_pci_sriov_pfs);
+		mutex_unlock(&vfio_pci_sriov_pfs_mutex);
 		ret = pci_enable_sriov(pdev, nr_virtfn);
+		if (ret)
+			goto out_del;
+		ret = nr_virtfn;
+		goto out_put;
+	}
 
-	vfio_device_put(device);
+	pci_disable_sriov(pdev);
 
-	return ret < 0 ? ret : nr_virtfn;
+out_del:
+	mutex_lock(&vfio_pci_sriov_pfs_mutex);
+	list_del_init(&vdev->sriov_pfs_item);
+out_unlock:
+	mutex_unlock(&vfio_pci_sriov_pfs_mutex);
+out_put:
+	vfio_device_put(device);
+	return ret;
 }
 EXPORT_SYMBOL_GPL(vfio_pci_core_sriov_configure);
 
diff --git a/include/linux/vfio_pci_core.h b/include/linux/vfio_pci_core.h
index 74a4a0f17b28..48f2dd3c568c 100644
--- a/include/linux/vfio_pci_core.h
+++ b/include/linux/vfio_pci_core.h
@@ -133,6 +133,8 @@ struct vfio_pci_core_device {
 	struct mutex		ioeventfds_lock;
 	struct list_head	ioeventfds_list;
 	struct vfio_pci_vf_token	*vf_token;
+	struct list_head		sriov_pfs_item;
+	struct vfio_pci_core_device	*sriov_pf_core_dev;
 	struct notifier_block	nb;
 	struct mutex		vma_lock;
 	struct list_head	vma_list;