{"dataType":"CVE_RECORD","dataVersion":"5.2","cveMetadata":{"cveId":"CVE-2024-45013","assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","state":"PUBLISHED","assignerShortName":"Linux","dateReserved":"2024-08-21T05:34:56.681Z","datePublished":"2024-09-11T15:13:50.210Z","dateUpdated":"2026-08-05T11:37:40.579Z"},"containers":{"cna":{"providerMetadata":{"orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux","dateUpdated":"2026-08-05T11:37:40.579Z"},"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\nnvme: move stopping keep-alive into nvme_uninit_ctrl()\n\nCommit 4733b65d82bd (\"nvme: start keep-alive after admin queue setup\")\nmoves starting keep-alive from nvme_start_ctrl() into\nnvme_init_ctrl_finish(), but don't move stopping keep-alive into\nnvme_uninit_ctrl(), so keep-alive work can be started and keep pending\nafter failing to start controller, finally use-after-free is triggered if\nnvme host driver is unloaded.\n\nThis patch fixes kernel panic when running nvme/004 in case that connection\nfailure is triggered, by moving stopping keep-alive into nvme_uninit_ctrl().\n\nThis way is reasonable because keep-alive is now started in\nnvme_init_ctrl_finish()."}],"metrics":[{"cvssV3_1":{"version":"3.1","vectorString":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H","baseScore":9.8,"baseSeverity":"CRITICAL"},"scenarios":[{"lang":"en","value":"AV:N - The affected code is the NVMe-over-Fabrics host driver (nvme-tcp/nvme-rdma), which is bound to the network stack and drives controller setup from responses supplied by a remote NVMe-oF target; a malicious, compromised, or spoofed target on the network fully controls whether setup fails after nvme_init_ctrl_finish() and when the keep-alive completion lands. NVMe/TCP has no mandatory authentication or encryption in typical deployments, so the attacker set extends to the network.\nAC:L - The remote target controls both sides of the interleaving — it chooses when to answer the keep-alive command and when to fail or stall the I/O-queue connect — and hosts retry connection setup automatically, giving unlimited attempts; the nvme-loop variant (loop.c:468-476, 573-618) has no race at all and leaks ka_work deterministically.\nPR:N - The attacker acts as the NVMe-oF target (or spoofs/MITMs the unauthenticated TCP connection) and needs no account, credentials, or privileges of any kind on the victim host. The vulnerable path runs during connection establishment, before any optional DH-HMAC-CHAP authentication is a factor.\nUI:N - Production NVMe-oF hosts connect automatically via nvmf-autoconnect at boot and via discovery-controller AEN-driven auto-connect (nvme-stas), so an attacker controlling the discovery/target endpoint can trigger repeated nvme_tcp_create_ctrl() attempts with no human action; the nvme-loop reset path (nvme_loop_reset_ctrl_work) is likewise entered automatically on controller error.\nS:U - The use-after-free occurs on kernel heap memory and is exploited within the kernel's own security authority; no VM, IOMMU, or sandbox boundary is crossed.\nC:H - The freed nvme_tcp_ctrl/nvme_loop_ctrl object can be reclaimed and re-sprayed by the attacker, and the stale timer/work dereference over that memory yields a controllable read and, once control flow is hijacked via work->func, arbitrary kernel memory disclosure.\nI:H - The armed delayed_work lives inside the kfree()d controller, so timer expiry calls timer->function and the worker calls work->func from freed heap; reallocating that slab object gives the attacker a kernel function pointer under their control, i.e. arbitrary code execution and unrestricted modification of kernel state.\nA:H - Even without successful exploitation the stale delayed work fires on freed memory and reliably panics the kernel — this is exactly the reported crash from blktests nvme/004 — which is a complete denial of service on the storage host."}]}],"affected":[{"product":"Linux","vendor":"Linux","defaultStatus":"unaffected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["drivers/nvme/host/core.c"],"versions":[{"version":"3af755a46881c32fecaecfdeaf3a8f0a869deca5","lessThan":"4101af98ab573554c4225e328d506fec2a74bc54","status":"affected","versionType":"git"},{"version":"3af755a46881c32fecaecfdeaf3a8f0a869deca5","lessThan":"a54a93d0e3599b05856971734e15418ac551a14c","status":"affected","versionType":"git"}]},{"product":"Linux","vendor":"Linux","defaultStatus":"affected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["drivers/nvme/host/core.c"],"versions":[{"version":"6.7","status":"affected"},{"version":"0","lessThan":"6.7","status":"unaffected","versionType":"semver"},{"version":"6.10.7","lessThanOrEqual":"6.10.*","status":"unaffected","versionType":"semver"},{"version":"6.11","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":"6.7","versionEndExcluding":"6.10.7"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.7","versionEndExcluding":"6.11"}]}]}],"references":[{"url":"https://git.kernel.org/stable/c/4101af98ab573554c4225e328d506fec2a74bc54"},{"url":"https://git.kernel.org/stable/c/a54a93d0e3599b05856971734e15418ac551a14c"}],"title":"nvme: move stopping keep-alive into nvme_uninit_ctrl()","x_generator":{"engine":"bippy-1.2.0"}},"adp":[{"metrics":[{"other":{"type":"ssvc","content":{"timestamp":"2024-09-29T15:50:07.552201Z","id":"CVE-2024-45013","options":[{"Exploitation":"none"},{"Automatable":"no"},{"Technical Impact":"partial"}],"role":"CISA Coordinator","version":"2.0.3"}}}],"title":"CISA ADP Vulnrichment","providerMetadata":{"orgId":"134c704f-9b21-4f2e-91b3-4a467353bcc0","shortName":"CISA-ADP","dateUpdated":"2024-09-29T15:50:22.500Z"}}]}}