CVE-2026-68138
HighCVSS 7.8Exploitation Probability (EPSS)
Low risk3th percentile - higher than 3% of all known CVEs
Summary
A vulnerability in the Linux kernel net/sched subsystem: qdisc_get_rtab() and qdisc_put_rtab() mutate the global qdisc_rtab_list and a non-atomic refcnt without locking, which is unsafe when called without RTNL (e.g., via cls_flower). This leads to a race condition, use-after-free, and double-free of the qdisc_rate_table structure.
Risk Assessment
An attacker can remotely cause a system crash or potentially escalate privileges by exploiting the race condition in rate table operations within flower filters with police actions.
Recommendation
Apply a Linux kernel update containing the fix (a spinlock protecting qdisc_rtab_list and refcnt). If unavailable, restrict the use of flower filters with police actions or disable creation of such rules.
Original NVD description (English source)
In the Linux kernel, the following vulnerability has been resolved: net/sched: serialize qdisc_rtab_list against concurrent get/put qdisc_get_rtab() and qdisc_put_rtab() mutate the process-global singly linked list qdisc_rtab_list and a plain non-atomic 'int refcnt' with no lock. This was only safe because every caller historically held the RTNL mutex, which serialized all rate-table lookups, inserts and frees. That invariant no longer holds. cls_flower sets TCF_PROTO_OPS_DOIT_UNLOCKED, so tc_new_tfilter() keeps rtnl_held == false for it and sets TCA_ACT_FLAGS_NO_RTNL. That flag propagates through tcf_exts_validate_ex() -> tcf_action_init() -> tcf_action_init_1() -> tcf_police_init(), which calls qdisc_get_rtab()/qdisc_put_rtab() with the RTNL mutex NOT held. Two RTM_NEWTFILTER requests on different CPUs, each adding a flower filter with a police action carrying the same rate, then race on qdisc_rtab_list and on the non-atomic refcnt, leading to a use-after-free / double-free of the kmalloc-2k struct qdisc_rate_table. qdisc_rtab_list is a single global (not per-netns), so the corrupted object is shared system-wide. BUG: KASAN: slab-use-after-free in qdisc_put_rtab+0x12f/0x160 qdisc_put_rtab+0x12f/0x160 tcf_police_init+0xda9/0x1590 tcf_action_init_1+0x460/0x6b0 tcf_action_init+0x439/0xa40 tcf_exts_validate_ex+0x42d/0x550 fl_change+0xddd/0x7da0 tc_new_tfilter+0xaa7/0x2420 rtnetlink_rcv_msg+0x95e/0xe90 which belongs to the cache kmalloc-2k of size 2048 Protect qdisc_rtab_list and the refcount with a dedicated spinlock. The (sleeping, GFP_KERNEL) allocation in qdisc_get_rtab() is performed before taking the lock; if a concurrent inserter added an identical table in the meantime the freshly allocated one is freed under the lock, so no duplicate is leaked. qdisc_put_rtab() now decrements the refcount and unlinks under the same lock.