{"id":"CVE-2026-97595","published":"2026-09-25T11:17:10.770","lastModified":"2026-09-25T15:18:00.797","description":"In the Linux kernel, the following vulnerability has been resolved:\n\nmac802154: fix use-after-free of sdata via queued RX frames\n\nThe RX softirq producer ieee802154_subif_frame() queues received beacon\nand MAC-command frames onto local->rx_beacon_list / rx_mac_cmd_list and\nschedules a process-context worker, storing a raw mac_pkt->sdata (and\nskb->dev == sdata->dev) with neither a reference nor any locking:\n\n - the lists have no lock: the softirq producer list_add_tail()s while the\n   mac_wq worker list_del()s, so sibling interfaces on the same phy corrupt\n   the list;\n\n - the workers dereference the interface after it may have been freed.\n   mac802154_rx_mac_cmd_worker() touches mac_pkt->sdata directly, and\n   mac802154_rx_beacon_worker() -> mac802154_process_beacon() dereferences\n   skb->dev (== sdata->dev). Removing an interface frees its sdata\n   (netdev_priv) while a queued frame still points at it, so a later worker\n   run is a use-after-free.\n\nReproduced under KASAN by flooding a victim interface with MAC command\nframes and removing it (the beacon path is the same class via skb->dev):\n\n  BUG: KASAN: slab-use-after-free in mac802154_rx_mac_cmd_worker+0x463/0x630 [mac802154]\n  Read of size 4 at addr ffff888002f9ea18 by task kworker/u8:1/31\n  Workqueue: phy0-mac-cmds mac802154_rx_mac_cmd_worker [mac802154]\n  Call Trace:\n   mac802154_rx_mac_cmd_worker+0x463/0x630 [mac802154]\n   process_one_work+0x611/0xe80\n   worker_thread+0x52e/0xdc0\n   kthread+0x30c/0x630\n   ret_from_fork+0x2fd/0x3e0\n\nFix both lists together:\n\n - add local->rx_lock and take it around every list access: the softirq\n   producer (plain spin_lock, softirq context) and the workers and flush\n   (spin_lock_bh, process context);\n\n - pin the interface for the lifetime of a queued frame with\n   netdev_hold()/netdev_put(), so the worker can safely dereference sdata /\n   skb->dev even while the interface is being removed;\n\n - dequeue under the lock at the head and loop-drain the whole list in the\n   workers (they previously processed one frame per run and relied on a\n   later enqueue to drain the rest);\n\n - drop not-yet-started frames of an interface before it is unregistered,\n   from ieee802154_if_remove() (after the RCU grace period) and from the\n   ieee802154_remove_interfaces() loop -- the latter is the whole-phy\n   teardown path, which does not go through ieee802154_if_remove().\n\nAn in-flight worker that already dequeued a frame keeps its own netdev\nreference; unregister_netdevice() then waits it out in netdev_run_todo(),\nwhich runs at rtnl_unlock() (rtnl released) and after the interface has\nbeen closed, so it does not pin rtnl. A worker blocked in an association\nTX only delays that one interface's unregister (the usual \"waiting for %s\nto become free\"), it does not hold rtnl. netdev_hold() is used for this\nreason instead of a cancel_work_sync() under rtnl, which would block on\nthe worker's unbounded MLME TX wait via ieee802154_sync_queue().\n\nThe mac-command worker additionally skips processing for a stopped\ninterface (ieee802154_sdata_running()), avoiding a needless association\nresponse during teardown.","cvssScore":7.5,"cvssSeverity":"HIGH","cvssVector":"CVSS:3.1/AV:A/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:H","cwes":[],"vendors":[],"products":[],"references":[{"url":"https://git.kernel.org/stable/c/07f0214018d7241e3610c1be984c8abfce16c099","tags":[]},{"url":"https://git.kernel.org/stable/c/2f37fba846c9fdff5fc15b6d93656057ccd13031","tags":[]},{"url":"https://git.kernel.org/stable/c/3d1fde35f95d7f02a7dc473c51d81e8d9e992cfc","tags":[]},{"url":"https://git.kernel.org/stable/c/4740ef0518c5e3bf31ebfb2288a6ac3b462ead06","tags":[]}],"exploitRefs":[],"hasPoc":false,"ai":null}