{"dataType":"CVE_RECORD","dataVersion":"5.2","cveMetadata":{"cveId":"CVE-2026-99433","assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","state":"PUBLISHED","assignerShortName":"Linux","dateReserved":"2026-09-25T10:25:14.415Z","datePublished":"2026-11-19T14:41:08.385Z","dateUpdated":"2026-11-22T11:08:16.430Z"},"containers":{"cna":{"providerMetadata":{"orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux","dateUpdated":"2026-11-22T11:08:16.430Z"},"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\nocfs2/dlm: hold a lock reference across grant AST enqueue\n\nA newly granted remote lock must remain alive until its grant AST is\nqueued.  dlmlock_master() publishes the lock on res->granted with its base\nand queue references, then drops res->spinlock before calling\ndlm_queue_ast().  The resource's reserved AST count does not retain the\nlock itself.\n\nThis can fail when a peer dies with an immediately grantable CREATE\nbuffered on its socket.  O2NET's node-down callback flushes receive work\nbefore DLM cleanup, but the flush only waits for work queued on entry.  A\ndata-ready callback that already selected the socket container can queue\nnew receive work after the flush advances its work color.  Socket shutdown\ndoes not discard bytes already queued for TCP receive.\n\nThe late receive work and DLM node-down cleanup can interleave as\nfollows (the lock reference operations shown are before this fix):\n\n  Socket data-ready        Node-down thread         Receive worker\n  -----------------        ----------------         --------------\n  o2net_data_ready()\n    select sc under\n    callback read lock\n                           o2net_disconnect_node()\n                             queue shutdown\n                             begin flush; new color\n    queue sc_rx_work\n    in the new color\n    release callback lock\n                             shutdown socket\n                             flush returns\n                                                    read buffered CREATE\n                                                    dlmlock_master()\n                                                      lock res->spinlock\n                                                      add lock to granted\n                                                      unlock res->spinlock\n                           DLM node-down cleanup\n                             lock res->spinlock\n                             dlm_free_dead_locks()\n                               unlink lock\n                               put queue/base refs\n                               free lock\n                             unlock res->spinlock\n                                                    dlm_queue_ast()\n                                                      __dlm_queue_ast()\n                                                      read lock->lockres\n                                                      (use-after-free)\n\nDLM cleanup holds res->spinlock while dlm_free_dead_locks() drops the\nqueue and base references.  The CREATE handler no longer holds that\nspinlock when __dlm_queue_ast() reads the lock, before taking its AST list\nreference.\n\nTake a temporary lock reference under res->spinlock when a grant AST will\nbe queued, and drop it after dlm_queue_ast().  This keeps the lock alive\nacross cleanup until the AST path has taken its own reference, without\nreversing the existing AST/resource lock order.\n\nKASAN report as below:\n\n    BUG: KASAN: slab-use-after-free in __dlm_queue_ast+0x8a4/0xc00\n    Read of size 8 at addr ffff88810ead5940 by task kworker/u16:3/52\n\n    CPU: 3 UID: 0 PID: 52 Comm: kworker/u16:3 Not tainted 7.3.0-rc4-next-20260921-pmb-bt-functional-v1+ #1 PREEMPT(lazy)\n    Hardware name: QEMU Ubuntu 24.04 PC v2 (i440FX + PIIX, arch_caps fix, 1996), BIOS 1.16.3-debian-1.16.3-2 04/01/2014\n    Workqueue: o2net o2net_rx_until_empty\n    Call Trace:\n     <TASK>\n     dump_stack_lvl+0x93/0xd0\n     print_report+0xce/0x630\n     ? __dlm_queue_ast+0x8a4/0xc00\n     ? srso_alias_return_thunk+0x5/0xfbef5\n     ? __virt_addr_valid+0x20e/0x420\n     ? __dlm_queue_ast+0x8a4/0xc00\n     kasan_report+0xe0/0x110\n     ? __dlm_queue_ast+0x8a4/0xc00\n     __dlm_queue_ast+0x8a4/0xc00\n     ? srso_alias_return_thunk+0x5/0xfbef5\n     ? srso_alias_return_thunk+0x5/0xfbef5\n     ? do_raw_spin_lock+0x130/0x270\n     ? __pfx___dlm_queue_ast+0x10/0x10\n     ? __pfx_do_raw_spin_lock+0x10/0x10\n     ? _raw_spin_unlock_irqrestore+0x45/0x60\n     ? srs\n---truncated---"}],"metrics":[{"cvssV3_1":{"version":"3.1","vectorString":"CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H","baseScore":8.8,"baseSeverity":"HIGH"},"scenarios":[{"lang":"en","value":"AV:N - A domain-joined peer sends DLM_CREATE_LOCK_MSG (struct dlm_create_lock) on its o2net TCP socket; o2net_rx_until_empty() delivers it through o2net_process_message() to dlm_create_lock_handler(), which calls dlmlock_master() and use-after-frees in dlm_queue_ast(). That message is carried on routable IPv4 TCP.\nAC:L - The peer both emits that grantable CREATE and stops its own o2hb writes so o2hb_do_disk_heartbeat() runs the down callbacks. o2net_disconnect_node() flushes o2net_wq, but o2net_data_ready() can queue sc_rx_work after the flush color advances; that worker then overlaps dlm_free_dead_locks() after dlmlock_master() drops res->spinlock and before __dlm_queue_ast() takes a reference.\nPR:L - o2net_accept_one() accepts the socket only after o2nm_get_node_by_ip() matches a configured node that o2hb_check_node_heartbeating_from_callback() reports live, and __dlm_hb_node_down() calls dlm_free_dead_locks() only for a node already in dlm->domain_map. Reaching dlmlock_master() takes a joined o2cb peer.\nUI:N - The same peer writes the CREATE and then stops updating its heartbeat slot; o2hb_thread on the victim fires the callbacks with no mount, open, or other local user action.\nS:U - The use-after-free corrupts the victim kernel's o2dlm_lock in dlmlock_master()/__dlm_queue_ast(); the affected authority is that same kernel.\nC:H - __dlm_queue_ast() reads lock->lockres from the o2dlm_lock after dlm_free_dead_locks() has dropped both refs and dlm_lock_release() has freed the 184-byte object. dlm_flush_asts() follows that pointer into dlm_do_remote_ast() and dlm_update_lvb(), so a reused slot is treated as a live lock and its fields are consumed.\nI:H - On that freed object __dlm_queue_ast() still calls dlm_lock_get(), takes lock->spinlock, and list_add_tail()s lock->ast_list onto dlm->pending_asts, which dlm_flush_asts() later walks. Those stores corrupt the slab slot and splice it into live DLM state.\nA:H - The reproduced KASAN report is a slab-use-after-free read of lock->lockres in __dlm_queue_ast() on the o2net worker, after o2hb_thread freed the object in dlm_lock_release(). The following NULL lockres dereference or list corruption in dlm_flush_asts() oopses the node."}]}],"affected":[{"product":"Linux","vendor":"Linux","defaultStatus":"unaffected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["fs/ocfs2/dlm/dlmlock.c"],"versions":[{"version":"2c5c54aca9d0263f81bd4886232835ba31f7635a","lessThan":"11643cb726c6077c8d80766516c20ff25ba29609","status":"affected","versionType":"git"},{"version":"2c5c54aca9d0263f81bd4886232835ba31f7635a","lessThan":"f4ae0b96a4abc7d59088eb2e0a2ee524c15f3bf0","status":"affected","versionType":"git"},{"version":"2c5c54aca9d0263f81bd4886232835ba31f7635a","lessThan":"0eb7cd1641695c5de549516661c9b90ba20f943d","status":"affected","versionType":"git"},{"version":"2c5c54aca9d0263f81bd4886232835ba31f7635a","lessThan":"71743471204e21113295c786e5cfa570294234a4","status":"affected","versionType":"git"}]},{"product":"Linux","vendor":"Linux","defaultStatus":"affected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["fs/ocfs2/dlm/dlmlock.c"],"versions":[{"version":"2.6.25","status":"affected"},{"version":"0","lessThan":"2.6.25","status":"unaffected","versionType":"semver"},{"version":"6.12.117","lessThanOrEqual":"6.12.*","status":"unaffected","versionType":"semver"},{"version":"6.18.61","lessThanOrEqual":"6.18.*","status":"unaffected","versionType":"semver"},{"version":"7.3.2","lessThanOrEqual":"7.3.*","status":"unaffected","versionType":"semver"},{"version":"7.4-rc1","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":"2.6.25","versionEndExcluding":"6.12.117"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"2.6.25","versionEndExcluding":"6.18.61"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"2.6.25","versionEndExcluding":"7.3.2"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"2.6.25","versionEndExcluding":"7.4-rc1"}]}]}],"references":[{"url":"https://git.kernel.org/stable/c/11643cb726c6077c8d80766516c20ff25ba29609"},{"url":"https://git.kernel.org/stable/c/f4ae0b96a4abc7d59088eb2e0a2ee524c15f3bf0"},{"url":"https://git.kernel.org/stable/c/0eb7cd1641695c5de549516661c9b90ba20f943d"},{"url":"https://git.kernel.org/stable/c/71743471204e21113295c786e5cfa570294234a4"}],"title":"ocfs2/dlm: hold a lock reference across grant AST enqueue","x_generator":{"engine":"bippy-1.2.0"}}}}