{ "dataType": "CVE_RECORD", "dataVersion": "5.2", "cveMetadata": { "cveId": "CVE-2026-97525", "assignerOrgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67", "state": "PUBLISHED", "assignerShortName": "Linux", "dateReserved": "2026-09-24T16:01:01.152Z", "datePublished": "2026-09-25T10:21:28.284Z", "dateUpdated": "2026-09-25T14:41:06.479Z" }, "containers": { "cna": { "providerMetadata": { "orgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67", "shortName": "Linux", "dateUpdated": "2026-09-25T14:41:06.479Z" }, "descriptions": [ { "lang": "en", "value": "In the Linux kernel, the following vulnerability has been resolved:\n\nx86/mm/pat: Allocate split page tables as kernel page tables\n\nA PTE is allocated directly without going through the standard page table\nallocation routines (such as pte_alloc_one_kernel()) when the CPA code\nsplits a large page (__split_large_page()).\n\nThis means the page table constructor is never called nor is the page table\nmarked as a kernel page table.\n\nThe former results in the folio associated with the page table not being\nmarked as a page table (__pagetable_ctor() is never called thus neither is\n__folio_set_pgtable()) nor are statistics updated to reflect\nit (lruvec_stat_add_folio() is never called).\n\nThe latter issue of failing to mark the page table as a kernel page\ntable (ptdesc_set_kernel() is never called) is far more problematic.\n\nSince commit:\n\n 5ba2f0a15564 (\"mm: introduce deferred freeing for kernel page tables\")\n\nkernel page table freeing has been batched and since the\nsubsequent commit:\n\n e37d5a2d60a3 (\"iommu/sva: invalidate stale IOTLB entries for kernel address space\")\n\nIOTLB cache entries for kernel page tables have been invalidated upon\nbeing freed.\n\nSince split page tables are freed without this invalidation, the IOTLB\ncan contain stale entries for them.\n\nResolve the issue by using the ordinary PTE allocation API at split time.\n\nThis results in these kernel page tables invoking a page table constructor,\nand thus requires a page table destructor.\n\nDestructors are not always present, like for early allocated direct map\npage tables). Conditionally call pagetable_dtor_free() if the PG_table\nfolio flag for the ptdesc is set, otherwise we free the page table via\npagetable_free().\n\nRegardless of which path is taken page tables marked as kernel page tables,\nwhich now includes split page tables, take the correct route through\npagetable_free_kernel().\n\nThere is a user-visible side effect in that split page tables will appear\nin nr_page_table_pages in /proc/vmstat (as do other kernel page tables\nallocated after early boot), however this is a positive change.\n\nThis issue started being markedly problematic after commit:\n\n 5ba2f0a15564 (\"mm: introduce deferred freeing for kernel page tables\")\n\nso choose this as the Fixes target.\n\n[ dhansen: rephrase in imperative mood ]" } ], "metrics": [ { "cvssV3_1": { "version": "3.1", "vectorString": "CVSS:3.1/AV:L/AC:L/PR:H/UI:N/S:C/C:H/I:H/A:H", "baseScore": 8.2, "baseSeverity": "HIGH" }, "scenarios": [ { "lang": "en", "value": "AV:L - split_large_page() is reached from local set_memory_* via change_page_attr_set_clr()→__change_page_attr(); the path that allocates unmarked tables and later frees them is finit_module→module_memory_alloc()→execmem_alloc_rw()→execmem_force_rw()/set_memory_nx (split) then module_enable_text_rox()→set_memory_rox() (collapse). No remote protocol carries the input.\nAC:L - On an x86 IOMMU_SVA host, a module whose text is carved from the execmem ROX 2M cache makes execmem_force_rw() split that PMD (pagetable_alloc without ptdesc_set_kernel) and set_memory_rox(CPA_COLLAPSE) free it in cpa_collapse_large_pages(); the same caller binds SVA, primes IOTLB walks, and sprays reuse without an uncontrolled race.\nPR:H - may_init_module() requires capable(CAP_SYS_MODULE) in init_user_ns, which user namespaces cannot grant. Only EXECMEM_MODULE_TEXT with EXECMEM_ROX_CACHE both splits a huge mapping in split_large_page() and collapse-frees it; EXECMEM_BPF lacks ROX_CACHE so unprivileged bpf_jit_alloc_exec()/set_memory_rox does not produce these unmarked tables.\nUI:N - The attacker issues their own init_module/finit_module (and optional delete_module) against a module they supply; no other user must mount an image, open a file, or bind a device on their behalf.\nS:C - cpa_collapse_large_pages() calls pagetable_free() on the PTE table split_large_page() allocated without ptdesc_set_kernel(), so ptdesc_test_kernel() is false and pagetable_free_kernel() never runs iommu_sva_invalidate_kva_range(); a PASID SVA walk can treat the reused page as IOMMU translations, crossing the IOMMU DMA isolation boundary.\nC:H - A stale IOTLB paging-structure cache entry for the freed split table lets the IOMMU interpret attacker-controlled reused page contents as PTEs, enabling arbitrary physical-memory DMA reads of kernel or other-process data.\nI:H - The same stale IOTLB walk enables arbitrary physical-memory DMA writes, and the IOMMU may write Accessed and Dirty bits into the recycled page while walking it, a write-after-free of the unmarked CPA table.\nA:H - Arbitrary DMA writes into reused memory and IOTLB write-after-free of the page cpa_collapse_large_pages() passed to pagetable_free() can corrupt kernel or page-table state, causing an oops, panic, or hang." } ] } ], "affected": [ { "product": "Linux", "vendor": "Linux", "defaultStatus": "unaffected", "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git", "programFiles": [ "arch/x86/mm/pat/set_memory.c" ], "versions": [ { "version": "b3039c526f3e1744db0cbb7ae1f0213f5e27d3f4", "lessThan": "84e0cd79d57f06b872154eb5d8f610584133f260", "status": "affected", "versionType": "git" }, { "version": "5ba2f0a1556479638ac11a3c201421f5515e89f5", "lessThan": "922873cf4fc34124215070a8ad391bf831538bb9", "status": "affected", "versionType": "git" }, { "version": "5ba2f0a1556479638ac11a3c201421f5515e89f5", "lessThan": "9e4a3ec3411bb6bb59e3c1f29b75609f1e87aac4", "status": "affected", "versionType": "git" }, { "version": "6.18.7", "lessThan": "6.18.53", "status": "affected", "versionType": "semver" } ] }, { "product": "Linux", "vendor": "Linux", "defaultStatus": "affected", "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git", "programFiles": [ "arch/x86/mm/pat/set_memory.c" ], "versions": [ { "version": "6.19", "status": "affected" }, { "version": "0", "lessThan": "6.19", "status": "unaffected", "versionType": "semver" }, { "version": "6.18.53", "lessThanOrEqual": "6.18.*", "status": "unaffected", "versionType": "semver" }, { "version": "7.2.7", "lessThanOrEqual": "7.2.*", "status": "unaffected", "versionType": "semver" }, { "version": "7.3-rc4", "lessThanOrEqual": "*", "status": "unaffected", "versionType": "original_commit_for_fix" } ] } ], "cpeApplicability": [ { "nodes": [ { "operator": "OR", "negate": false, "cpeMatch": [ { "vulnerable": true, "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "versionStartIncluding": "6.18.7", "versionEndExcluding": "6.18.53" }, { "vulnerable": true, "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "versionStartIncluding": "6.19", "versionEndExcluding": "7.2.7" }, { "vulnerable": true, "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "versionStartIncluding": "6.19", "versionEndExcluding": "7.3-rc4" } ] } ] } ], "references": [ { "url": "https://git.kernel.org/stable/c/84e0cd79d57f06b872154eb5d8f610584133f260" }, { "url": "https://git.kernel.org/stable/c/922873cf4fc34124215070a8ad391bf831538bb9" }, { "url": "https://git.kernel.org/stable/c/9e4a3ec3411bb6bb59e3c1f29b75609f1e87aac4" } ], "title": "x86/mm/pat: Allocate split page tables as kernel page tables", "x_generator": { "engine": "bippy-1.2.0" } } } }