{"id":"AZL-108141","summary":"CVE-2026-98401 affecting package kernel 6.6.157.1-1","details":"In the Linux kernel, the following vulnerability has been resolved:\n\nnet: iptunnel: fix stale transport header during tunnel decapsulation\n\nSyzbot reported a crash in qdisc_pkt_len_segs_init() caused by a stale\ntransport_header offset after tunnel decapsulation.\n\nBUG: unable to handle page fault for address: ffffed102091a42e\nOops: Oops: 0000 [#1] SMP KASAN NOPTI\nCPU: 0 UID: 0 PID: 340 Comm: qdisc_uaf_repro Not tainted 7.2.0-rc4-00061-g248951ddc14d #256 PREEMPT(full)\nHardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-debian-1.16.3-2 04/01/2014\nRIP: 0010:__asan_load2\n\u003cIRQ\u003e\nqdisc_pkt_len_segs_init (net/core/dev.c:4145)\n__dev_queue_xmit (net/core/dev.c:4787)\nbr_dev_queue_push_xmit (net/bridge/br_forward.c:53)\nbr_handle_frame_finish (net/bridge/br_input.c:229)\nbr_handle_frame (net/bridge/br_input.c:315)\n__netif_receive_skb_core.constprop.0 (net/core/dev.c:6099)\n__netif_receive_skb_list_core (net/core/dev.c:6287)\nnetif_receive_skb_list_internal (net/core/dev.c:6445)\nnapi_complete_done (net/core/dev.c:6813)\ngro_cell_poll (net/core/gro_cells.c:74)\n__napi_poll (net/core/dev.c:7735)\nnet_rx_action (net/core/dev.c:7798 net/core/dev.c:7955)\nhandle_softirqs (kernel/softirq.c:622)\ndo_softirq (kernel/softirq.c:523  kernel/softirq.c:510 )\n__local_bh_enable_ip (kernel/softirq.c:450)\ntun_get_user (drivers/net/tun.c:1986 (discriminator 1))\ntun_chr_write_iter (drivers/net/tun.c:2032)\n\nThe issue is completely latent until qdisc read transport header in\ncommit 7fb4c1967011 (\"net: pull headers in qdisc_pkt_len_segs_init()\").\nThe crash requires four conditions to line up:\n\n1. The incoming packet is encapsulated and carries GSO metadata. The outer\n   transport header offset is stored in skb-\u003etransport_header while the\n   packet is still in the outer tunnel context.\n2. The tunnel receiver strips the outer headers. skb-\u003edata is advanced to\n   the inner frame, but skb-\u003etransport_header is left pointing to the\n   now-removed outer L4 header, so it becomes a negative offset relative to\n   the new data.\n3. The inner frame is not delivered to the local IP stack. Instead, it\n   is forwarded at L2 by a bridge or HSR, so ip_rcv_core() never runs and\n   the transport header is not reset to the inner L4 offset.\n4. The forwarding path calls __dev_queue_xmit(), which enters\n   qdisc_pkt_len_segs_init(). That function computes the GSO header length\n   from skb_transport_offset(skb). Because the offset is negative, the\n   unsigned cast overflows and pskb_may_pull(skb, hdr_len +\n   sizeof(struct tcphdr)) reads past the end of the skb, triggering a\n   KASAN fault or page fault.\n\nThe issue specifically requires GSO packets (shinfo-\u003egso_size != 0), which\nare processed/aggregated through gro_cells. Fix this by clearing\ntransport_header to the ~0U sentinel in gro_cell for all tunnnel driver.\nGTP does not support GRO/GSO, drop the evil GSO packets in GTP directly.","modified":"2026-10-15T14:17:16.452306666Z","published":"2026-10-14T14:17:51Z","upstream":["CVE-2026-98401"],"references":[{"type":"WEB","url":"https://nvd.nist.gov/vuln/detail/CVE-2026-98401"}],"affected":[{"package":{"name":"kernel","ecosystem":"Azure Linux:3","purl":"pkg:rpm/azure-linux/kernel"},"ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"last_affected":"6.6.157.1-1"}]}],"database_specific":{"source":"https://github.com/microsoft/AzureLinuxVulnerabilityData/blob/main/osv/AZL-108141.json"}}],"schema_version":"1.9.0"}