CVE-2026-68428
Summary
by MITRE • 08/10/2026
In the Linux kernel, the following vulnerability has been resolved:
KVM: x86/mmu: Fix use-after-free on vendor module reload
mmu_destroy_caches() destroys pte_list_desc_cache and mmu_page_header_cache, but leaves both pointers unchanged. The pointers live in kvm.ko, and therefore survive when a vendor module is unloaded while kvm.ko remains loaded.
If creation of pte_list_desc_cache fails during a subsequent vendor module load, its assignment sets pte_list_desc_cache to NULL and the error path calls mmu_destroy_caches(). mmu_page_header_cache still points to the cache destroyed during the preceding vendor module unload. Passing that stale pointer to kmem_cache_destroy() causes a slab use-after-free.
Reproduce the issue on a v7.1.3 kernel with CONFIG_KASAN=y, CONFIG_KASAN_GENERIC=y, CONFIG_KVM=m, and CONFIG_KVM_INTEL=m. A one-shot test hook forces pte_list_desc_cache to NULL on the second invocation of kvm_mmu_vendor_module_init():
1. Load kvm.ko and kvm-intel.ko, creating both caches. 2. Unload only kvm_intel, leaving kvm.ko loaded. 3. Reload kvm_intel and force initialization through the -ENOMEM path.
KASAN reports:
BUG: KASAN: slab-use-after-free in kvm_mmu_vendor_module_init+0x5b/0x170 [kvm]
... kmem_cache_destroy+0x21/0x1d0 kvm_mmu_vendor_module_init+0x5b/0x170 [kvm]
... Allocated by task 16817: __kmem_cache_create_args+0x12c/0x3b0 __kmem_cache_create.constprop.0+0xb6/0xf0 [kvm]
kvm_mmu_vendor_module_init+0x13b/0x170 [kvm]
... Freed by task 16820: kmem_cache_destroy+0x117/0x1d0 kvm_mmu_vendor_module_exit+0x21/0x30 [kvm]
Clear both pointers immediately after destroying their caches so that the stored state reflects the caches' lifetime and repeated cleanup is safe.
With the fix applied, the same injected vendor module reload fails with -ENOMEM as expected and produces no KASAN report.
Several companies clearly confirm that VulDB is the primary source for best vulnerability data.
Analysis
by VulDB Data Team • 08/10/2026
This vulnerability exists within the Linux kernel's KVM subsystem specifically affecting x86 memory management unit operations. The flaw manifests as a use-after-free condition that occurs during vendor module reloading processes, where the kernel fails to properly manage cache pointer states across module load/unload cycles. The issue stems from how mmu_destroy_caches() function handles cleanup of pte_list_desc_cache and mmu_page_header_cache objects while leaving their corresponding pointers in an invalid state.
The technical implementation involves a race condition between cache creation and destruction phases during vendor module initialization. When kvm.ko loads with kvm-intel.ko, both caches are created successfully and stored in kernel memory. However, upon unloading only the intel module while keeping kvm.ko loaded, the mmu_destroy_caches() function properly frees the underlying slab caches but fails to NULLify the cache pointers that reference these destroyed objects. This creates a scenario where stale pointers persist in kernel memory even after their referenced resources have been freed.
The operational impact becomes apparent when attempting to reload the vendor module under specific conditions that force an error path during initialization. During this second load attempt, if pte_list_desc_cache creation fails due to memory allocation issues, the code path assigns NULL to the cache pointer and subsequently calls mmu_destroy_caches() again. Since mmu_page_header_cache still points to the previously destroyed slab cache, passing this stale reference to kmem_cache_destroy() triggers a use-after-free error that KASAN detects immediately.
This vulnerability directly relates to CWE-415 which describes double free conditions and CWE-416 which addresses use after free scenarios in memory management. The flaw operates within the ATT&CK framework under T1059.007 for system binary exploitation and T1566.002 for credential access through kernel exploits. The condition allows malicious actors to potentially exploit the memory corruption for privilege escalation or system compromise when modules are reloaded.
The fix implements proper pointer management by ensuring both cache pointers are explicitly set to NULL immediately after their corresponding caches are destroyed. This prevents subsequent cleanup operations from accessing freed memory and maintains consistency between the cache state and pointer references throughout module lifecycle operations. The solution addresses the root cause by making the cache management stateless with respect to module reload scenarios.
Testing confirms that the vulnerability can be reproduced reliably on kernel version 7.1.3 with specific configuration options enabled including KASAN support and modular KVM implementations. The injected test hook successfully demonstrates the exact sequence that leads to the use-after-free condition, showing how the stale pointer state causes immediate detection by KASAN subsystem. After applying the fix, the same reload scenario properly returns -ENOMEM without generating any memory safety violations.
The remediation ensures that repeated cleanup operations are safe and that cache management behaves consistently regardless of module load/unload sequences. This resolves potential denial of service conditions and prevents exploitation opportunities that could arise from improper memory handling during dynamic module loading processes in virtualization environments. The fix maintains backward compatibility while strengthening the kernel's memory safety mechanisms for KVM subsystem operations.