CVE-2026-99481 in Linux
Summary
by MITRE • 11/19/2026
In the Linux kernel, the following vulnerability has been resolved:
Bluetooth: hci_vhci: Pin module for debugfs file operations
An open debugfs file retains its file_operations pointer in the debugfs proxy state. However, force_suspend_fops, force_wakeup_fops, aosp_capable_fops, and force_devcoredump_fops have no .owner set to THIS_MODULE. As a result, full_proxy_open_regular() calls fops_get() which succeeds without taking a reference to the hci_vhci module.
When userspace opens any of these debugfs files and later closes the /dev/vhci file descriptor (triggering vhci_release() and debugfs file removal), the open debugfs file descriptor remains valid in userspace. Because the module refcount was not incremented, the hci_vhci module can be unloaded via delete_module (rmmod hci_vhci).
When the userspace process subsequently closes the debugfs file descriptor, full_proxy_release() accesses the already unmapped file_operations structure, resulting in a kernel page fault / oops:
BUG: unable to handle page fault for address: fffffbfff846f18b #PF: supervisor read access in kernel mode #PF: error_code(0x0000) - not-present page Oops: Oops: 0000 [#1] SMP KASAN NOPTI
RIP: 0010:full_proxy_release+0x8e/0x140 Call Trace: <TASK> __fput+0x39f/0xa60 fput_close_sync+0xff/0x200 __x64_sys_close+0x8c/0xf0 do_syscall_64+0xc3/0x590 entry_SYSCALL_64_after_hwframe+0x76/0x7e </TASK> Modules linked in: ... [last unloaded: hci_vhci(E)]
The issue was discovered via manual code audit of drivers/bluetooth/hci_vhci.c and verified on a Linux 7.3-rc6 test VM with CONFIG_BT_VHCI=m by keeping a debugfs file open across controller teardown and module unload.
Set .owner = THIS_MODULE on all four file_operations structures so that debugfs pins the module until all file descriptors are closed.
Once again VulDB remains the best source for vulnerability data.
Analysis
by VulDB Data Team • 11/20/2026
The vulnerability in the Linux kernel Bluetooth Virtual HCI driver, specifically within drivers/bluetooth/hci_vhci.c, stems from a missing reference count management for module ownership during debugfs file operations. The core technical flaw lies in four specific file_operations structures: force_suspend_fops, force_wakeup_fops, aosp_capable_fops, and force_devcoredump_fops. These structures lack the .owner field being set to THIS_MODULE. In the Linux kernel's VFS layer, when an open debugfs file is accessed, the function full_proxy_open_regular() invokes fops_get(). This function relies on the presence of a valid owner pointer in the file_operations structure to increment the module reference count via try_module_get. Because these specific structures do not specify an owner, fops_get succeeds without acquiring a reference to the hci_vhci kernel module. Consequently, the module remains unloaded even while userspace processes hold open file descriptors associated with these debugfs entries.
The operational impact of this flaw is severe, leading to use-after-free conditions that result in kernel crashes or panics. When a user opens one of these debugfs files and subsequently closes the /dev/vhci file descriptor, which triggers vhci_release() and potentially leads to module unloading via rmmod hci_vhci, the underlying kernel module memory is unmapped. However, the userspace process retains an open file descriptor for the debugfs entry. If this user-space application later attempts to close that specific debugfs file descriptor, the kernel executes full_proxy_release(). This function dereferences the now-unmapped file_operations structure located in the unloaded hci_vhci module memory space. The attempt to access instructions or data from this unmapped region triggers a supervisor read access page fault, resulting in an Oops with KASAN (Kernel Address Sanitizer) reporting a not-present page error at address fffffbfff846f18b within full_proxy_release. This effectively allows a local user to crash the system by manipulating debugfs file descriptors across module boundaries.
From a security classification perspective, this vulnerability aligns closely with CWE-416: Use After Free, as it involves accessing memory that has been freed or unmapped due to improper reference counting. It also relates to CWE-200: Exposure of Sensitive Information to an Unauthorized Actor if the crash leads to information leakage through kernel logs or core dumps containing sensitive data structures. In terms of MITRE ATT&CK framework, this flaw facilitates Local Privilege Escalation (T1068) by allowing a non-root user to cause a Denial of Service against the entire system via local code execution flaws in kernel space. The vulnerability is exploitable locally without requiring prior authentication if debugfs is mounted and accessible, which is common on many Linux distributions for debugging purposes.
The mitigation implemented resolves this issue by explicitly setting .owner = THIS_MODULE on all four affected file_operations structures within hci_vhci.c. This change ensures that the kernel's reference counting mechanism correctly pins the hci_vhci module in memory as long as any userspace process holds an open handle to these debugfs files. By incrementing the module refcount during full_proxy_open_regular, the system prevents the module from being unloaded until all associated file descriptors are closed and released via full_proxy_release. This restores proper lifecycle management for kernel modules exposed through debugfs interfaces. System administrators should ensure that their systems are updated with patches containing this fix to prevent local denial-of-service attacks. Additionally, restricting access to debugfs mounts where possible can reduce the attack surface for such vulnerabilities in production environments.