{"id":"CVE-2026-99172","published":"2026-10-28T11:17:37.790","lastModified":"2026-10-28T11:17:37.790","description":"In the Linux kernel, the following vulnerability has been resolved:\n\nipc: drop the mqueue mount with kern_unmount() on setup failure\n\nmq_init_ns() creates the mqueue mount with fc_mount_longterm(), so its\n->mnt_ns is MNT_NS_INTERNAL. When setting up the mqueue or ipc sysctls\nor msg_init_ns() fails afterwards, create_ipc_ns() drops it with a plain\nmntput(). With ->mnt_ns set that takes the fast path of\nmntput_no_expire(). The count drops to zero and nothing ever kills the\nmount, so the mount and the mqueue superblock are leaked.\n\nCommit dc55e35f9e81 (\"ipc: Store mqueue sysctls in the ipc namespace\")\nadded the first failure path after mq_init_ns() and never dropped the\nmount. Commit 7229d74e5e8c (\"ipc: create_ipc_ns: drop mqueue mount on\nsysctl setup failure\") tried to fix that with a mntput() but that\ndoesn't help for a long-term mount.\n\nUse kern_unmount(). free_ipc() open-codes the same thing for a namespace\nthat made it.","cvssScore":null,"cvssSeverity":null,"cvssVector":null,"cwes":[],"vendors":[],"products":[],"references":[{"url":"https://git.kernel.org/stable/c/c0620b648321fff654173d238d4a85298ea16d05","tags":[]},{"url":"https://git.kernel.org/stable/c/f0e48172a59e517e613b73f73cc35f24fd123aeb","tags":[]}],"exploitRefs":[],"hasPoc":false,"ai":null}