Description

In the Linux kernel, the following vulnerability has been resolved: mm/hugetlb: fix list corruption in allocate_file_region_entries() allocate_file_region_entries() tops up resv->region_cache with freshly allocated file_region descriptors. The allocation uses GFP_KERNEL, so resv->lock is dropped around it: the new entries are gathered on a stack-local list head, allocated_regions, and spliced into resv->region_cache once the lock is re-acquired. The splice used list_splice(), which moves the entries but does not re-initialize the source head, so allocated_regions is left pointing at an entry that now lives on resv->region_cache. The top-up runs in a while loop that re-checks the cache deficit after re-acquiring the lock. For a shared mapping the resv_map is shared by every mapper of the hugetlbfs inode, so a concurrent region_chg()/region_add()/region_del() on the same resv_map can consume cache entries during the unlocked window and force a second iteration. That iteration calls list_add() on the stale head and corrupts the list; with CONFIG_DEBUG_LIST the __list_add_valid() check trips: list_add corruption. next->prev should be prev (ffffc900011ff7f8), but was ffff88814c281460. (next=ffff88814c545640). kernel BUG at lib/list_debug.c:31! allocate_file_region_entries+0x191/0x420 region_chg+0x267/0x300 hugetlb_reserve_pages+0x387/0xc80 hugetlbfs_file_mmap+0x2ce/0x3f0 mmap_region+0x1348/0x1a80 do_mmap+0x85e/0xb90 vm_mmap_pgoff+0x18c/0x330 ksys_mmap_pgoff+0x2a1/0x3e0 do_syscall_64+0xd7/0x420 Without CONFIG_DEBUG_LIST the bad list_add() silently links a kernel-stack address into resv->region_cache, leading to later use-after-free. This was observed as a real host panic on a dense KVM host where a QEMU guest-RAM hugetlbfs file was mapped MAP_SHARED by both QEMU and a separate SPDK/DPDK vhost-user target, generating concurrent region_* traffic on one shared resv_map. Use list_splice_init() so the source head is re-initialized empty after each splice, making the retry loop safe.

Executive Summary

Technical Details

Mitigations

References

Additional Metadata

{
  "score": "7.8",
  "pubDate": "2026-08-15T13:17:56.967Z",
  "pubdate": "2026-08-15T13:17:56.967Z",
  "executiveSummary": "A list corruption vulnerability exists in the Linux kernel memory management subsystem, specifically within the hugetlb file region allocation mechanism (allocate_file_region_entries()).\nThe flaw is classified as a linked list corruption leading to use-after-free and potential kernel panic.\nIt impacts Linux kernel environments utilizing hugetlbfs shared mappings, particularly dense virtualization hosts running QEMU and SPDK/DPDK vhost-user targets where concurrent region operations occur on a shared resv_map.\nAn attacker or concurrent workload triggering specific race conditions can cause kernel memory corruption.\nWithout CONFIG_DEBUG_LIST enabled, the bug silently links kernel-stack memory addresses into the persistent resv->region_cache, resulting in a subsequent use-after-free condition.\nExploitation requires local system access or concurrent operations on shared hugetlbfs mappings that generate high contention on the resv_map.",
  "technicalDetails": "The vulnerable component is the allocate_file_region_entries() function located in the mm/hugetlb.c source file of the Linux kernel.\nThe root cause of the vulnerability stems from improper handling of temporary linked list heads during cache replenishment operations.\nWhen allocate_file_region_entries() tops up resv->region_cache with newly allocated file_region descriptors, it performs memory allocations using GFP_KERNEL.\nBecause GFP_KERNEL allocations can sleep, resv->lock is temporarily dropped.\nDuring this unlocked window, newly allocated entries are gathered onto a stack-local list head named allocated_regions.\nOnce the lock is re-acquired, these entries are spliced into resv->region_cache using the standard list_splice() function.\nThe critical flaw is that list_splice() moves the entries but fails to re-initialize the source head, leaving allocated_regions pointing directly at an entry that now resides on resv->region_cache.\nThe top-up logic operates inside a while loop that re-checks the cache deficit after re-acquiring resv->lock.\nIn shared mappings, the resv_map is shared across every mapper of the hugetlbfs inode.\nConcurrent operations such as region_chg(), region_add(), or region_del() executing on the same resv_map can consume cache entries during the unlocked window, forcing the while loop to execute a second iteration.\nDuring this subsequent iteration, the loop invokes list_add() on the stale, uninitialized allocated_regions head, corrupting the list pointers.\nIf CONFIG_DEBUG_LIST is enabled, the __list_add_valid() integrity check trips immediately, causing a kernel BUG and resulting in a host panic.\nIf CONFIG_DEBUG_LIST is disabled, the invalid list_add() silently introduces a kernel-stack address into resv->region_cache.\nSubsequent allocations and cache usage lead directly to a use-after-free condition, where kernel stack memory can be accessed or overwritten via heap-like operations on the hugetlb reservation map.\nNo special authentication or network exposure is required, as the attack vector relies on local concurrent process interactions with shared hugetlbfs memory segments."
}