In the Linux kernel, the following vulnerability has been resolved:
vfio/pci: clear vdev->msi_perm after freeing it on init failure
vfio_msi_cap_len() lazily allocates the per-device MSI permission table:
<pre>
vdev->msi_perm = kmalloc_obj(struct perm_bits, GFP_KERNEL_ACCOUNT);
if (!vdev->msi_perm)
return -ENOMEM;
ret = init_pci_cap_msi_perm(vdev->msi_perm, len, flags);
if (ret) {
kfree(vdev->msi_perm);
return ret; /* vdev->msi_perm left dangling */
}
</pre>
When init_pci_cap_msi_perm() -> alloc_perm_bits() fails with -ENOMEM, the
error path frees vdev->msi_perm but leaves the freed pointer stored in
it. vdev->msi_perm is not re-zeroed later because struct
vfio_pci_core_device is per-device and persists across open/close cycles,
and the vfio_config_init() error path returns without calling
vfio_config_free(). So the dangling pointer outlives the failed open.
That leads to two use-after-frees on the same device:
-
Reuse. The next vfio_config_init() sees the stale pointer at
"if (vdev->msi_perm) return len;" and reuses the freed object. MSI
config accesses in vfio_pci_config_rw_single() then dereference and
call the freed perm->readfn / perm->writefn function pointers.
-
Double free. A later vfio_config_free() runs free_perm_bits() and
kfree() on the already-freed object.
Fix it by NULLing vdev->msi_perm after the kfree(), matching the
NULL-after-free discipline already used in free_perm_bits() and
vfio_config_free().
BUG: KASAN: slab-use-after-free in vfio_pci_config_rw_single (drivers/vfio/pci/vfio_pci_config.c:1961)
Read of size 8 at addr ffff88800fcc88d0 by task exploit/143
Call Trace:
...
kasan_report (mm/kasan/report.c:595)
vfio_pci_config_rw_single (drivers/vfio/pci/vfio_pci_config.c:1961)
vfio_pci_config_rw (drivers/vfio/pci/vfio_pci_config.c:1986)
vfio_pci_rw (drivers/vfio/pci/vfio_pci_core.c:1599)
vfs_read (fs/read_write.c:572)
__x64_sys_pread64 (fs/read_write.c:764)
do_syscall_64 (arch/x86/entry/syscall_64.c:94)
...
Followed on device close by a double free of the same object:
Oops: general protection fault, probably for non-canonical address
0x1f63e0e8000008: 0000 [#1] SMP KASAN NOPTI
RIP: 0010:kfree (mm/slub.c:6711)
Call Trace:
vfio_config_free (drivers/vfio/pci/vfio_pci_config.c:1861)
vfio_pci_core_disable (drivers/vfio/pci/vfio_pci_core.c:685)
vfio_pci_core_close_device (drivers/vfio/pci/vfio_pci_core.c:777)
vfio_df_close (drivers/vfio/vfio_main.c:602)
vfio_device_fops_release (drivers/vfio/vfio_main.c:648)
__fput (fs/file_table.c:512)
__x64_sys_close (fs/open.c:1496)
do_syscall_64 (arch/x86/entry/syscall_64.c:94)
...
Kernel panic - not syncing: Fatal exception
CVSS Vector: CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H
CVSS Score: 8.8
AV:L - The UAF is reached only via local VFIO char-device operations: open of /dev/vfio/$GROUP or /dev/vfio/devices/vfioN, then pread/pwrite of the PCI config region through vfio_pci_core_read/write into vfio_pci_config_rw_single(). There is no network or packet path into vfio_msi_cap_len().
AC:L - After any failed vfio_config_init() plants the dangling vdev->msi_perm, the attacker retries GET_DEVICE_FD/BIND_IOMMUFD on the same persistent vdev and deterministically calls the freed perm->readfn/writefn via MSI config accesses. The attacker controls that open/read sequence; no victim race or uncontrolled layout is required.
PR:L - vfio_pci_core_enable() and vfio_pci_config_rw() have no capable() check on the normal IOMMU path (CAP_SYS_RAWIO is only for noiommu). Access is gated by /dev/vfio node permissions, which libvirt, kubevirt, and cloud GPU/NIC passthrough routinely grant to an unprivileged qemu user or tenant.
UI:N - The attacker performs the VFIO open, retry, and MSI config-space pread/pwrite themselves. Prior admin binding of the PCI device to vfio-pci is deployment configuration, not victim interaction during exploitation.
S:C - struct vfio_pci_core_device persists across open/close, so a failed open can leave a dangling msi_perm that a later tenant VM reuses. QEMU services guest PCI config cycles with VFIO pread/pwrite, so a passthrough guest can invoke the freed host function pointers and cross the VFIO/IOMMU guest-to-host boundary.
C:H - The next successful open reuses the freed perm_bits and MSI config reads call perm->readfn on that object. Reclaiming the slab slot yields an arbitrary kernel-read primitive, which CVSS kernel guidance scores High for use-after-free.
I:H - MSI config writes call the freed perm->writefn, and a later close double-frees the same object in vfio_config_free(). That UAF/double-free of a structure with function pointers is exploitable for arbitrary kernel write and control-flow hijack.
A:H - The report shows a KASAN slab-use-after-free in vfio_pci_config_rw_single() followed by a general-protection fault in kfree() during vfio_config_free() that panics the host, so availability impact is a kernel oops/panic.
| Attack Vector |
Local |
Scope |
Changed |
| Attack Complexity |
Low |
Confidentiality Impact |
High |
| Privileges Required |
Low |
Integrity Impact |
High |
| User Interaction |
None |
Availability Impact |
High |
AV:L - The UAF is reached only via local VFIO char-device operations: open of /dev/vfio/$GROUP or /dev/vfio/devices/vfioN, then pread/pwrite of the PCI config region through vfio_pci_core_read/write into vfio_pci_config_rw_single(). There is no network or packet path into vfio_msi_cap_len().
AC:L - After any failed vfio_config_init() plants the dangling vdev->msi_perm, the attacker retries GET_DEVICE_FD/BIND_IOMMUFD on the same persistent vdev and deterministically calls the freed perm->readfn/writefn via MSI config accesses. The attacker controls that open/read sequence; no victim race or uncontrolled layout is required.
PR:L - vfio_pci_core_enable() and vfio_pci_config_rw() have no capable() check on the normal IOMMU path (CAP_SYS_RAWIO is only for noiommu). Access is gated by /dev/vfio node permissions, which libvirt, kubevirt, and cloud GPU/NIC passthrough routinely grant to an unprivileged qemu user or tenant.
UI:N - The attacker performs the VFIO open, retry, and MSI config-space pread/pwrite themselves. Prior admin binding of the PCI device to vfio-pci is deployment configuration, not victim interaction during exploitation.
S:C - struct vfio_pci_core_device persists across open/close, so a failed open can leave a dangling msi_perm that a later tenant VM reuses. QEMU services guest PCI config cycles with VFIO pread/pwrite, so a passthrough guest can invoke the freed host function pointers and cross the VFIO/IOMMU guest-to-host boundary.
C:H - The next successful open reuses the freed perm_bits and MSI config reads call perm->readfn on that object. Reclaiming the slab slot yields an arbitrary kernel-read primitive, which CVSS kernel guidance scores High for use-after-free.
I:H - MSI config writes call the freed perm->writefn, and a later close double-frees the same object in vfio_config_free(). That UAF/double-free of a structure with function pointers is exploitable for arbitrary kernel write and control-flow hijack.
A:H - The report shows a KASAN slab-use-after-free in vfio_pci_config_rw_single() followed by a general-protection fault in kfree() during vfio_config_free() that panics the host, so availability impact is a kernel oops/panic.
CVSS 3.1