{"dataType":"CVE_RECORD","dataVersion":"5.2","cveMetadata":{"cveId":"CVE-2026-53373","assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","state":"PUBLISHED","assignerShortName":"Linux","dateReserved":"2026-06-09T07:44:35.401Z","datePublished":"2026-07-19T10:01:57.632Z","dateUpdated":"2026-08-05T12:35:17.519Z"},"containers":{"cna":{"providerMetadata":{"orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux","dateUpdated":"2026-08-05T12:35:17.519Z"},"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\nmm/vma: do not try to unmap a VMA if mmap_prepare() invoked from mmap()\n\nThe mmap_prepare hook functionality includes the ability to invoke\nmmap_prepare() from the mmap() hook of existing 'stacked' drivers, that is\nones which are capable of calling the mmap hooks of other drivers/file\nsystems (e.g.  overlayfs, shm).\n\nAs part of the mmap_prepare action functionality, we deal with errors by\nunmapping the VMA should one arise.  This works in the usual mmap_prepare\ncase, as we invoke this action at the last moment, when the VMA is\nestablished in the maple tree.\n\nHowever, the mmap() hook passes a not-fully-established VMA pointer to the\ncaller (which is the motivation behind the mmap_prepare() work), which is\ndetached.\n\nSo attempting to unmap a VMA in this state will be problematic, with the\nmost obvious symptom being a warning in vma_mark_detached(), because the\nVMA is already detached.\n\nIt's also unncessary - the mmap() handler will clean up the VMA on error.\n\nSo to fix this issue, this patch propagates whether or not an mmap action\nis being completed via the compatibility layer or directly.\n\nIf the former, then we do not attempt VMA cleanup, if the latter, then we\ndo.\n\nThis patch also updates the userland VMA tests to reflect the change."}],"metrics":[{"cvssV3_1":{"version":"3.1","vectorString":"CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H","baseScore":7.8,"baseSeverity":"HIGH"},"scenarios":[{"lang":"en","value":"AV:L - The vulnerability is reached only through the mmap/mmap2 syscall path (ksys_mmap_pgoff → do_mmap → mmap_region → __mmap_region → stacked .mmap() → compat_vma_mmap → mmap_action_complete), which requires local process access to invoke the kernel memory-management interface.\nAC:L - An attacker fully controls mmap parameters (address, length, fd, flags) and can create overlapping mappings to place detached VMAs in the target range before triggering a stacked mmap_prepare action failure, making exploitation conditions attacker-influenced rather than dependent on rare external timing or layout.\nPR:L - Triggering requires only the ability to call mmap() on an open file descriptor; syzbot reported this from unprivileged fuzzing, and stacked paths (overlayfs, shm, fuse passthrough) are reachable by ordinary local users who can open/mmap files in their mount namespace without real init-namespace root.\nUI:N - Exploitation requires no victim interaction beyond the attacker's own mmap syscall; no third-party user action such as opening a file or mounting a filesystem on behalf of the attacker is needed.\nS:U - The flawed cleanup corrupts the calling process's own mm_struct/VMA state within the kernel; it does not by itself cross a security boundary such as guest-to-host VM escape or an IOMMU/DMA sandbox bypass.\nC:H - Erroneous do_munmap() on VMAs that are detached but still linked in the maple tree during an in-progress mmap corrupts kernel memory-management metadata, which is a memory-corruption primitive that can be leveraged for arbitrary kernel memory disclosure, not merely a benign warning.\nI:H - The double-unmap/detach of VMAs in a transitional mmap state can leave the address-space tree inconsistent and is a memory-corruption bug in kernel structures that could be developed into arbitrary write or privilege-escalation primitives, not limited to integrity-neutral denial of service.\nA:H - The known symptom is a WARN in vma_mark_detached() on already-detached VMAs, and operating on refcount-zero detached VMAs during a nested unmap can provoke kernel oops/panic or leave the mm in a crash-prone inconsistent state, satisfying high availability impact."}]}],"affected":[{"product":"Linux","vendor":"Linux","defaultStatus":"unaffected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["include/linux/mm.h","mm/util.c","mm/vma.c","tools/testing/vma/include/dup.h","tools/testing/vma/include/stubs.h"],"versions":[{"version":"ac0a3fc9c07df79dc8a4ce9d274df00afc7bf12d","lessThan":"5394bcb746503f2ae4b206212416dccea78e3773","status":"affected","versionType":"git"},{"version":"ac0a3fc9c07df79dc8a4ce9d274df00afc7bf12d","lessThan":"619eab23e1ce7c97e54bfc5a417306d94b3f6f13","status":"affected","versionType":"git"}]},{"product":"Linux","vendor":"Linux","defaultStatus":"affected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["include/linux/mm.h","mm/util.c","mm/vma.c","tools/testing/vma/include/dup.h","tools/testing/vma/include/stubs.h"],"versions":[{"version":"6.19","status":"affected"},{"version":"0","lessThan":"6.19","status":"unaffected","versionType":"semver"},{"version":"7.0.9","lessThanOrEqual":"7.0.*","status":"unaffected","versionType":"semver"},{"version":"7.1","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":"6.19","versionEndExcluding":"7.0.9"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.19","versionEndExcluding":"7.1"}]}]}],"references":[{"url":"https://git.kernel.org/stable/c/5394bcb746503f2ae4b206212416dccea78e3773"},{"url":"https://git.kernel.org/stable/c/619eab23e1ce7c97e54bfc5a417306d94b3f6f13"}],"title":"mm/vma: do not try to unmap a VMA if mmap_prepare() invoked from mmap()","x_generator":{"engine":"bippy-1.2.0"}}}}