{"dataType":"CVE_RECORD","dataVersion":"5.2","cveMetadata":{"cveId":"CVE-2025-39793","assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","state":"PUBLISHED","assignerShortName":"Linux","dateReserved":"2025-04-16T07:20:57.132Z","datePublished":"2025-09-12T15:59:30.388Z","dateUpdated":"2026-08-05T12:05:14.029Z"},"containers":{"cna":{"providerMetadata":{"orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux","dateUpdated":"2026-08-05T12:05:14.029Z"},"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\nio_uring/memmap: cast nr_pages to size_t before shifting\n\nIf the allocated size exceeds UINT_MAX, then it's necessary to cast\nthe mr->nr_pages value to size_t to prevent it from overflowing. In\npractice this isn't much of a concern as the required memory size will\nhave been validated upfront, and accounted to the user. And > 4GB sizes\nwill be necessary to make the lack of a cast a problem, which greatly\nexceeds normal user locked_vm settings that are generally in the kb to\nmb range. However, if root is used, then accounting isn't done, and\nthen it's possible to hit this issue."}],"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 bug is reached purely through local syscalls — `io_uring_setup()` followed by `io_uring_register(IORING_REGISTER_MEM_REGION)` with an attacker-chosen `io_uring_region_desc.size`. Per kernel guidance, io_uring is a local attack vector.\nAC:L - This is a deterministic integer truncation fully controlled by the attacker's chosen region size (any value ≥ 4 GB with a small remainder); there is no race, no memory-layout dependency, and no condition outside the attacker's control. The only environmental factor is RLIMIT_MEMLOCK/CAP_IPC_LOCK, which is a target configuration rather than a probabilistic exploitation condition.\nPR:L - `io_uring_setup` and `IORING_REGISTER_MEM_REGION` require no capability at all under the default `kernel.io_uring_disabled=0`; the only gate is the `RLIMIT_MEMLOCK` counter check, which no actual memory allocation must satisfy since the allocation itself is truncated. On the many real deployments where non-root users have memlock unlimited (container runtimes, RDMA/OFED and DPDK hosts, HPC nodes, database service accounts), an ordinary unprivileged local user reaches this directly; with real root (`CAP_IPC_LOCK` in init_user_ns) accounting is skipped entirely.\nUI:N - The attacker performs the entire sequence — ring setup, region registration, and mmap of the param region — within their own process. No victim action is involved.\nS:U - The corruption and its consequences are contained within the kernel's own security authority on the same host; no VM, IOMMU, or hypervisor boundary is crossed.\nC:H - The truncation causes `mr->pages[]` to be populated with `struct page` pointers spanning ~4 GB of unrelated physical memory, which is then `vmap()`ed into the kernel and mmap-able into userspace via `io_region_mmap()`/`vm_insert_pages()`. This yields direct read access to arbitrary physical memory belonging to other processes and the kernel.\nI:H - The same foreign pages are inserted into a writable user VMA and aliased writable in the kernel vmap, giving an arbitrary physical-memory write primitive (e.g. modifying cached setuid binary text or kernel data), and `insert_page_into_pte_locked()` performs `folio_get()`/rmap accounting on folios the region never owned, corrupting refcounts into a use-after-free. This is straightforwardly leveraged into privilege escalation.\nA:H - `io_check_coalesce_buffer()` dereferences ~1M out-of-range `struct page` pointers, oopsing whenever the walk crosses unpopulated vmemmap, and the resulting refcount corruption/UAF on in-use pages causes kernel panics. Any local user able to trigger it can crash the machine repeatedly."}]}],"affected":[{"product":"Linux","vendor":"Linux","defaultStatus":"unaffected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["io_uring/memmap.c"],"versions":[{"version":"087f997870a948820ec366701d178f402c6a23a3","lessThan":"c6a2706e08b8a1b2d3740161c0977d38e596c1ee","status":"affected","versionType":"git"},{"version":"087f997870a948820ec366701d178f402c6a23a3","lessThan":"a69a9b53c54e2d33e2a5b1ea4a9a71fd01c6cf3a","status":"affected","versionType":"git"},{"version":"087f997870a948820ec366701d178f402c6a23a3","lessThan":"33503c083fda048c77903460ac0429e1e2c0e341","status":"affected","versionType":"git"}]},{"product":"Linux","vendor":"Linux","defaultStatus":"affected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["io_uring/memmap.c"],"versions":[{"version":"6.14","status":"affected"},{"version":"0","lessThan":"6.14","status":"unaffected","versionType":"semver"},{"version":"6.15.11","lessThanOrEqual":"6.15.*","status":"unaffected","versionType":"semver"},{"version":"6.16.2","lessThanOrEqual":"6.16.*","status":"unaffected","versionType":"semver"},{"version":"6.17","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":"6.14","versionEndExcluding":"6.15.11"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.14","versionEndExcluding":"6.16.2"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.14","versionEndExcluding":"6.17"}]}]}],"references":[{"url":"https://git.kernel.org/stable/c/c6a2706e08b8a1b2d3740161c0977d38e596c1ee"},{"url":"https://git.kernel.org/stable/c/a69a9b53c54e2d33e2a5b1ea4a9a71fd01c6cf3a"},{"url":"https://git.kernel.org/stable/c/33503c083fda048c77903460ac0429e1e2c0e341"}],"title":"io_uring/memmap: cast nr_pages to size_t before shifting","x_generator":{"engine":"bippy-1.2.0"}}}}