{ "schema_version": "1.6.1", "id": "RPI-CVE-2026-68335", "modified": "2026-10-09T11:39:44Z", "upstream": [ "CVE-2026-68335" ], "published": "2026-08-10T12:08:00Z", "summary": "rds: drop incoming messages that cross network namespace boundaries", "details": "In the Linux kernel, the following vulnerability has been resolved:\n\nrds: drop incoming messages that cross network namespace boundaries\n\nrds_find_bound() looks up the destination socket using a global\nrhashtable keyed solely on (addr, port, scope_id). Network namespaces\nare not part of the key, so a sender in netns A can deliver an incoming\nmessage (inc) to a socket that lives in a different netns B.\n\nWhen this happens, inc->i_conn points to an rds_connection whose c_net\nis netns A, but the receiving rs lives in netns B. Once the child\nprocess that created netns A exits, cleanup_net() calls\nrds_loop_exit_net() -> rds_loop_kill_conns() -> rds_conn_destroy(),\nfreeing that connection. If the survivor socket in netns B still holds\nthe inc, any subsequent dereference of inc->i_conn is a use-after-free.\n\nThere are two dangerous sites in rds_clear_recv_queue():\n 1. inc->i_conn->c_lcong (offset 88 of freed rds_connection, size 200)\n read via rds_recv_rcvbuf_delta() -- confirmed by KASAN.\n 2. inc->i_conn->c_trans->inc_free(inc) (function pointer at offset 80)\n called via rds_inc_put() when the inc refcount reaches zero -- same\n race window, potential call-through-freed-object primitive.\n\nThe bug is reachable from unprivileged user namespaces\n(CLONE_NEWUSER + CLONE_NEWNET), available since Linux 3.8.\n\nFix this by rejecting the delivery in rds_recv_incoming() when the\nsocket returned by rds_find_bound() belongs to a different network\nnamespace than the connection that carried the message. Use the\nexisting rds_conn_net() / sock_net() helpers and net_eq() for the\ncomparison.", "severity": [ { "type": "CVSS_V3", "score": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H" } ], "affected": [ { "package": { "ecosystem": "Raspberry Pi OS:12", "name": "linux" }, "versions": [ "1:6.1.47-1+rpt4" ], "ecosystem_specific": { "binaries": [ { "binary_name": "linux-image-6.1.0-rpi3-rpi-2712", "binary_version": "1:6.1.47-1+rpt4" }, { "binary_name": "linux-image-6.1.0-rpi3-rpi-v6", "binary_version": "1:6.1.47-1+rpt4" }, { "binary_name": "linux-image-6.1.0-rpi3-rpi-v7", "binary_version": "1:6.1.47-1+rpt4" }, { "binary_name": "linux-image-6.1.0-rpi3-rpi-v7l", "binary_version": "1:6.1.47-1+rpt4" }, { "binary_name": "linux-image-6.1.0-rpi3-rpi-v8", "binary_version": "1:6.1.47-1+rpt4" } ] }, "database_specific": { "per_version": { "1:6.1.47-1+rpt4": { "linux_commit": "655fc658a15ae7a6f37103754adb39ba52a9a14e", "upstream_version": "6.1.47", "assessment": "commit ancestry (strak rule)", "upstream_fixed": "6.1.183", "compiled_filter": false } } } } ], "references": [ { "type": "ADVISORY", "url": "https://www.cve.org/CVERecord?id=CVE-2026-68335" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/742ff6f02545212e991cd8b45011e40d2c2ef25a" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/abff41fd928328bbf3dda1140beb2e61fa424ccd" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/03c574112e5d066df0ddce36d7438e850bcf3050" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/1e2e2d9806944fe485824d617c8b7c78116c22db" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/cfb3ce07b705e486e022a2f2b1242b48f13981ff" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/9591042533140dfe6608d9344806d567dcd39d02" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/0f8690e3869109cd5803ccb400889d20a0b54e0e" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/5521ae71e32a8069ed4ca6e792179dc57bc43ab2" } ], "database_specific": { "source": "https://git.kernel.org/pub/scm/linux/security/vulns.git", "upstream_modified": "2026-08-19T16:47:12Z", "severity": { "source": "kernel CNA", "cvss_v3_score": 7.8, "cvss_v3_severity": "High", "scenario": "AV:L - Exploitation requires local syscall access: the attacker must create RDS sockets and a network namespace (CLONE_NEWUSER|CLONE_NEWNET), then trigger namespace teardown to free the connection. A remote RDS peer cannot drive the netns-destruction half of the sequence, so the vulnerability is reachable only locally.\nAC:L - The attacker controls the entire sequence deterministically \u2014 bind a socket in one netns, send a loopback RDS message from a second netns so the inc is queued cross-netns, then exit that netns to free inc->i_conn. No race must be won and no uncontrollable precondition exists; the UAF is confirmed reproducible under KASAN.\nPR:L - Only an unprivileged local user account is required; the needed CAP_NET_ADMIN for creating and configuring the second network namespace is obtained inside a user namespace the attacker owns, as the commit explicitly notes (\"reachable from unprivileged user namespaces\").\nUI:N - The attacker performs every step \u2014 socket creation, cross-netns send, namespace teardown, and the subsequent read/close that dereferences the freed connection \u2014 with no action from any other user or administrator.\nS:U - The use-after-free corrupts kernel memory within the same kernel security authority; there is no crossing of a hypervisor, IOMMU, or comparable trust boundary.\nC:H - The stale inc->i_conn is dereferenced to read c_lcong from a freed 200-byte rds_connection, and the freed slot can be reclaimed with attacker-groomed data, giving a use-after-free read primitive usable to disclose kernel memory contents.\nI:H - rds_inc_put() calls inc->i_conn->c_trans->inc_free(inc) through a function pointer at offset 80 of the freed object \u2014 an indirect call through freed memory that the commit calls a \"call-through-freed-object primitive\", which after heap reclamation is a control-flow hijack path leading to arbitrary kernel modification.\nA:H - Dereferencing the freed rds_connection in rds_clear_recv_queue() and calling through its freed transport vtable reliably corrupts kernel state and panics the machine (KASAN-confirmed use-after-free), causing a full denial of service." } } }