{ "schema_version": "1.6.1", "id": "RPI-CVE-2026-74518", "modified": "2026-10-09T11:39:44Z", "upstream": [ "CVE-2026-74518" ], "published": "2026-08-15T12:28:42Z", "summary": "mm/hugetlb: fix list corruption in allocate_file_region_entries()", "details": "In the Linux kernel, the following vulnerability has been resolved:\n\nmm/hugetlb: fix list corruption in allocate_file_region_entries()\n\nallocate_file_region_entries() tops up resv->region_cache with freshly\nallocated file_region descriptors. The allocation uses GFP_KERNEL, so\nresv->lock is dropped around it: the new entries are gathered on a\nstack-local list head, allocated_regions, and spliced into\nresv->region_cache once the lock is re-acquired.\n\nThe splice used list_splice(), which moves the entries but does not\nre-initialize the source head, so allocated_regions is left pointing at an\nentry that now lives on resv->region_cache. The top-up runs in a while\nloop that re-checks the cache deficit after re-acquiring the lock. For a\nshared mapping the resv_map is shared by every mapper of the hugetlbfs\ninode, so a concurrent region_chg()/region_add()/region_del() on the same\nresv_map can consume cache entries during the unlocked window and force a\nsecond iteration. That iteration calls list_add() on the stale head and\ncorrupts the list; with CONFIG_DEBUG_LIST the __list_add_valid() check\ntrips:\n\n list_add corruption. next->prev should be prev (ffffc900011ff7f8),\n but was ffff88814c281460. (next=ffff88814c545640).\n kernel BUG at lib/list_debug.c:31!\n allocate_file_region_entries+0x191/0x420\n region_chg+0x267/0x300\n hugetlb_reserve_pages+0x387/0xc80\n hugetlbfs_file_mmap+0x2ce/0x3f0\n mmap_region+0x1348/0x1a80\n do_mmap+0x85e/0xb90\n vm_mmap_pgoff+0x18c/0x330\n ksys_mmap_pgoff+0x2a1/0x3e0\n do_syscall_64+0xd7/0x420\n\nWithout CONFIG_DEBUG_LIST the bad list_add() silently links a kernel-stack\naddress into resv->region_cache, leading to later use-after-free.\n\nThis was observed as a real host panic on a dense KVM host where a QEMU\nguest-RAM hugetlbfs file was mapped MAP_SHARED by both QEMU and a separate\nSPDK/DPDK vhost-user target, generating concurrent region_* traffic on one\nshared resv_map.\n\nUse list_splice_init() so the source head is re-initialized empty after\neach splice, making the retry loop safe.", "severity": [ { "type": "CVSS_V3", "score": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H" } ], "affected": [ { "package": { "ecosystem": "Raspberry Pi OS:12", "name": "linux" }, "versions": [ "1:6.1.47-1+rpt4" ], "ecosystem_specific": { "binaries": [ { "binary_name": "linux-image-6.1.0-rpi3-rpi-2712", "binary_version": "1:6.1.47-1+rpt4" }, { "binary_name": "linux-image-6.1.0-rpi3-rpi-v6", "binary_version": "1:6.1.47-1+rpt4" }, { "binary_name": "linux-image-6.1.0-rpi3-rpi-v7", "binary_version": "1:6.1.47-1+rpt4" }, { "binary_name": "linux-image-6.1.0-rpi3-rpi-v7l", "binary_version": "1:6.1.47-1+rpt4" }, { "binary_name": "linux-image-6.1.0-rpi3-rpi-v8", "binary_version": "1:6.1.47-1+rpt4" } ] }, "database_specific": { "per_version": { "1:6.1.47-1+rpt4": { "linux_commit": "655fc658a15ae7a6f37103754adb39ba52a9a14e", "upstream_version": "6.1.47", "assessment": "commit ancestry (strak rule)", "upstream_fixed": "6.1.183", "compiled_filter": false } } } } ], "references": [ { "type": "ADVISORY", "url": "https://www.cve.org/CVERecord?id=CVE-2026-74518" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/01b8569233e47693d6ff7efa96d9854c55f936fc" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/9c5fdffc5e1ce84403c58289ee72697051803bf7" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/f3e54f6a5e1681f83d13e8716bc92ef5ecf121d3" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/62e1c2741a4d923d9854efd5927a6212aad7a187" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/587a0accc2b4fccc5cf7baf0fe34e50efde51f9c" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/126a70bf1a08ddc9d79c471ebdaa2b08cfbab8df" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/ac1bb7fd45088d0db57a22ce7729f258ebd63cf5" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/dd9623f58ec702a07b2d67179d6fcea79c52231a" } ], "database_specific": { "source": "https://git.kernel.org/pub/scm/linux/security/vulns.git", "upstream_modified": "2026-08-19T16:47:12Z", "severity": { "source": "kernel CNA", "cvss_v3_score": 7.8, "cvss_v3_severity": "High", "scenario": "AV:L - The bug is reached only through local hugetlb reservation paths (mmap/memfd_create/fault syscalls into region_chg/region_add/region_del); no network protocol or remote packet handler invokes allocate_file_region_entries().\nAC:L - Exploitation needs a race while resv->lock is dropped during GFP_KERNEL allocation, but an attacker can spawn concurrent MAP_SHARED mappers/faulters on one hugetlb inode and controls both sides of that race.\nPR:L - Unprivileged local users can trigger this via memfd_create(MFD_HUGETLB) plus MAP_SHARED mmap/fault activity, or by mapping accessible shared hugetlbfs files, without init-namespace root or special capabilities.\nUI:N - No victim interaction is required; the attacker drives the needed concurrent syscalls and shared hugetlb mappings themselves to hit the corrupted list splice retry path.\nS:U - Impact is kernel hugetlb resv_map list/memory corruption on the host; it does not by itself cross a VM, IOMMU, or separate security-authority boundary even though KVM/DPDK hosts are a prime deployment.\nC:H - Without CONFIG_DEBUG_LIST, stale list_splice leaves a kernel-stack pointer in resv->region_cache, producing later use-after-free that can be turned into arbitrary kernel memory read/disclosure primitives.\nI:H - Corrupted region_cache pointers let subsequent region_add/region_del/cache operations write through attacker-influenced list links, enabling heap metadata corruption and potential arbitrary kernel write or code execution.\nA:H - Real host panics were reported on dense KVM hugetlb workloads; DEBUG_LIST kernels BUG on list_add corruption, and default builds can crash from UAF while walking the poisoned region_cache." } } }