wireguard-linux/net, branch devel

Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net

2026-04-14T19:04:00Z

Merge in late fixes in preparation for the net-next PR. Conflicts: include/net/sch_generic.h a6bd339dbb351 ("net_sched: fix skb memory leak in deferred qdisc drops") ff2998f29f390 ("net: sched: introduce qdisc-specific drop reason tracing") https://lore.kernel.org/adz0iX85FHMz0HdO@sirena.org.uk drivers/net/ethernet/airoha/airoha_eth.c 1acdfbdb516b ("net: airoha: Fix VIP configuration for AN7583 SoC") bf3471e6e6c0 ("net: airoha: Make flow control source port mapping dependent on nbq parameter") Adjacent changes: drivers/net/ethernet/airoha/airoha_ppe.c f44218cd5e6a ("net: airoha: Reset PPE cpu port configuration in airoha_ppe_hw_init()") 7da62262ec96 ("inet: add ip_local_port_step_width sysctl to improve port usage distribution") Signed-off-by: Jakub Kicinski

net: caif: clear client service pointer on teardown

2026-04-14T11:21:54Z

`caif_connect()` can tear down an existing client after remote shutdown by calling `caif_disconnect_client()` followed by `caif_free_client()`. `caif_free_client()` releases the service layer referenced by `adap_layer->dn`, but leaves that pointer stale. When the socket is later destroyed, `caif_sock_destructor()` calls `caif_free_client()` again and dereferences the freed service pointer. Clear the client/service links before releasing the service object so repeated teardown becomes harmless. Fixes: 43e369210108 ("caif: Move refcount from service layer to sock and dev.") Cc: stable@kernel.org Reported-by: Yifan Wu Reported-by: Juefei Pu Co-developed-by: Yuan Tan Signed-off-by: Yuan Tan Suggested-by: Xin Liu Tested-by: Ren Wei Signed-off-by: Zhengchuan Liang Signed-off-by: Ren Wei Link: https://patch.msgid.link/9f3d37847c0037568aae698ca23cd47c6691acb0.1775897577.git.zcliangcn@gmail.com Signed-off-by: Paolo Abeni

net: strparser: fix skb_head leak in strp_abort_strp()

2026-04-14T10:37:00Z

When the stream parser is aborted, for example after a message assembly timeout, it can still hold a reference to a partially assembled message in strp->skb_head. That skb is not released in strp_abort_strp(), which leaks the partially assembled message and can be triggered repeatedly to exhaust memory. Fix this by freeing strp->skb_head and resetting the parser state in the abort path. Leave strp_stop() unchanged so final cleanup still happens in strp_done() after the work and timer have been synchronized. Fixes: 43a0c6751a32 ("strparser: Stream parser for messages") Cc: stable@kernel.org Reported-by: Yifan Wu Reported-by: Juefei Pu Co-developed-by: Yuan Tan Signed-off-by: Yuan Tan Suggested-by: Xin Liu Tested-by: Yuan Tan Signed-off-by: Luxiao Xu Signed-off-by: Ren Wei Link: https://patch.msgid.link/ade3857a9404999ce9a1c27ec523efc896072678.1775482694.git.rakukuip@gmail.com Signed-off-by: Paolo Abeni

net, bpf: fix null-ptr-deref in xdp_master_redirect() for down master

2026-04-14T08:39:23Z

syzkaller reported a kernel panic in bond_rr_gen_slave_id() reached via xdp_master_redirect(). Full decoded trace: https://syzkaller.appspot.com/bug?extid=80e046b8da2820b6ba73 bond_rr_gen_slave_id() dereferences bond->rr_tx_counter, a per-CPU counter that bonding only allocates in bond_open() when the mode is round-robin. If the bond device was never brought up, rr_tx_counter stays NULL. The XDP redirect path can still reach that code on a bond that was never opened: bpf_master_redirect_enabled_key is a global static key, so as soon as any bond device has native XDP attached, the XDP_TX -> xdp_master_redirect() interception is enabled for every slave system-wide. The path xdp_master_redirect() -> bond_xdp_get_xmit_slave() -> bond_xdp_xmit_roundrobin_slave_get() -> bond_rr_gen_slave_id() then runs against a bond that has no rr_tx_counter and crashes. Fix this in the generic xdp_master_redirect() by refusing to call into the master's ->ndo_xdp_get_xmit_slave() when the master device is not up. IFF_UP is only set after ->ndo_open() has successfully returned, so this reliably excludes masters whose XDP state has not been fully initialized. Drop the frame with XDP_ABORTED so the exception is visible via trace_xdp_exception() rather than silently falling through. This is not specific to bonding: any current or future master that defers XDP state allocation to ->ndo_open() is protected. Fixes: 879af96ffd72 ("net, core: Add support for XDP redirection to slave device") Reported-by: syzbot+80e046b8da2820b6ba73@syzkaller.appspotmail.com Closes: https://lore.kernel.org/all/698f84c6.a70a0220.2c38d7.00cc.GAE@google.com/T/ Suggested-by: Daniel Borkmann Acked-by: Daniel Borkmann Signed-off-by: Jiayuan Chen Link: https://patch.msgid.link/20260411005524.201200-2-jiayuan.chen@linux.dev Signed-off-by: Paolo Abeni

sctp: disable BH before calling udp_tunnel_xmit_skb()

2026-04-13T23:50:28Z

udp_tunnel_xmit_skb() / udp_tunnel6_xmit_skb() are expected to run with BH disabled. After commit 6f1a9140ecda ("add xmit recursion limit to tunnel xmit functions"), on the path: udp(6)_tunnel_xmit_skb() -> ip(6)tunnel_xmit() dev_xmit_recursion_inc()/dec() must stay balanced on the same CPU. Without local_bh_disable(), the context may move between CPUs, which can break the inc/dec pairing. This may lead to incorrect recursion level detection and cause packets to be dropped in ip(6)_tunnel_xmit() or __dev_queue_xmit(). Fix it by disabling BH around both IPv4 and IPv6 SCTP UDP xmit paths. In my testing, after enabling the SCTP over UDP: # ip net exec ha sysctl -w net.sctp.udp_port=9899 # ip net exec ha sysctl -w net.sctp.encap_port=9899 # ip net exec hb sysctl -w net.sctp.udp_port=9899 # ip net exec hb sysctl -w net.sctp.encap_port=9899 # ip net exec ha iperf3 -s - without this patch: # ip net exec hb iperf3 -c 192.168.0.1 --sctp [ 5] 0.00-10.00 sec 37.2 MBytes 31.2 Mbits/sec sender [ 5] 0.00-10.00 sec 37.1 MBytes 31.1 Mbits/sec receiver - with this patch: # ip net exec hb iperf3 -c 192.168.0.1 --sctp [ 5] 0.00-10.00 sec 3.14 GBytes 2.69 Gbits/sec sender [ 5] 0.00-10.00 sec 3.14 GBytes 2.69 Gbits/sec receiver Fixes: 6f1a9140ecda ("net: add xmit recursion limit to tunnel xmit functions") Fixes: 046c052b475e ("sctp: enable udp tunneling socks") Signed-off-by: Xin Long Acked-by: Marcelo Ricardo Leitner Link: https://patch.msgid.link/c874a8548221dcd56ff03c65ba75a74e6cf99119.1776017727.git.lucien.xin@gmail.com Signed-off-by: Jakub Kicinski

sctp: fix missing encap_port propagation for GSO fragments

2026-04-13T23:49:37Z

encap_port in SCTP_INPUT_CB(skb) is used by sctp_vtag_verify() for SCTP-over-UDP processing. In the GSO case, it is only set on the head skb, while fragment skbs leave it 0. This results in fragment skbs seeing encap_port == 0, breaking SCTP-over-UDP connections. Fix it by propagating encap_port from the head skb cb when initializing fragment skbs in sctp_inq_pop(). Fixes: 046c052b475e ("sctp: enable udp tunneling socks") Signed-off-by: Xin Long Acked-by: Marcelo Ricardo Leitner Link: https://patch.msgid.link/ea65ed61b3598d8b4940f0170b9aa1762307e6c3.1776017631.git.lucien.xin@gmail.com Signed-off-by: Jakub Kicinski

tcp: Don't set treq->req_usec_ts in cookie_tcp_reqsk_init().

2026-04-13T22:58:08Z

Commit de5626b95e13 ("tcp: Factorise cookie-independent fields initialisation in cookie_v[46]_check().") miscategorised tcp_rsk(req)->req_usec_ts init to cookie_tcp_reqsk_init(), which is used by both BPF/non-BPF SYN cookie reqsk. Rather, it should have been moved to cookie_tcp_reqsk_alloc() by commit 8e7bab6b9652 ("tcp: Factorise cookie-dependent fields initialisation in cookie_v[46]_check()") so that only non-BPF SYN cookie sets tcp_rsk(req)->req_usec_ts to false. Let's move the initialisation to cookie_tcp_reqsk_alloc() to respect bpf_tcp_req_attrs.usec_ts_ok. Fixes: e472f88891ab ("bpf: tcp: Support arbitrary SYN Cookie.") Signed-off-by: Kuniyuki Iwashima Link: https://patch.msgid.link/20260410235328.1773449-1-kuniyu@google.com Signed-off-by: Jakub Kicinski

udp: Force compute_score to always inline

2026-04-13T22:44:42Z

Back in 2024 I reported a 7-12% regression on an iperf3 UDP loopback thoughput test that we traced to the extra overhead of calling compute_score on two places, introduced by commit f0ea27e7bfe1 ("udp: re-score reuseport groups when connected sockets are present"). At the time, I pointed out the overhead was caused by the multiple calls, associated with cpu-specific mitigations, and merged commit 50aee97d1511 ("udp: Avoid call to compute_score on multiple sites") to jump back explicitly, to force the rescore call in a single place. Recently though, we got another regression report against a newer distro version, which a team colleague traced back to the same root-cause. Turns out that once we updated to gcc-13, the compiler got smart enough to unroll the loop, undoing my previous mitigation. Let's bite the bullet and __always_inline compute_score on both ipv4 and ipv6 to prevent gcc from de-optimizing it again in the future. These functions are only called in two places each, udpX_lib_lookup1 and udpX_lib_lookup2, so the extra size shouldn't be a problem and it is hot enough to be very visible in profilings. In fact, with gcc13, forcing the inline will prevent gcc from unrolling the fix from commit 50aee97d1511, so we don't end up increasing udpX_lib_lookup2 at all. I haven't recollected the results myself, as I don't have access to the machine at the moment. But the same colleague reported 4.67% inprovement with this patch in the loopback benchmark, solving the regression report within noise margins. Eric Dumazet reported no size change to vmlinux when built with clang. I report the same also with gcc-13: scripts/bloat-o-meter vmlinux vmlinux-inline add/remove: 0/2 grow/shrink: 4/0 up/down: 616/-416 (200) Function old new delta udp6_lib_lookup2 762 949 +187 __udp6_lib_lookup 810 975 +165 udp4_lib_lookup2 757 906 +149 __udp4_lib_lookup 871 986 +115 __pfx_compute_score 32 - -32 compute_score 384 - -384 Total: Before=35011784, After=35011984, chg +0.00% Fixes: 50aee97d1511 ("udp: Avoid call to compute_score on multiple sites") Reviewed-by: Eric Dumazet Acked-by: Willem de Bruijn Signed-off-by: Gabriel Krisman Bertazi Link: https://patch.msgid.link/20260410155936.654915-1-krisman@suse.de Signed-off-by: Jakub Kicinski

net: qrtr: ns: Fix use-after-free in driver remove()

2026-04-13T22:34:07Z

In the remove callback, if a packet arrives after destroy_workqueue() is called, but before sock_release(), the qrtr_ns_data_ready() callback will try to queue the work, causing use-after-free issue. Fix this issue by saving the default 'sk_data_ready' callback during qrtr_ns_init() and use it to replace the qrtr_ns_data_ready() callback at the start of remove(). This ensures that even if a packet arrives after destroy_workqueue(), the work struct will not be dereferenced. Note that it is also required to ensure that the RX threads are completed before destroying the workqueue, because the threads could be using the qrtr_ns_data_ready() callback. Cc: stable@vger.kernel.org Fixes: 0c2204a4ad71 ("net: qrtr: Migrate nameservice to kernel from userspace") Signed-off-by: Manivannan Sadhasivam Link: https://patch.msgid.link/20260409-qrtr-fix-v3-5-00a8a5ff2b51@oss.qualcomm.com Signed-off-by: Jakub Kicinski

net: qrtr: ns: Limit the total number of nodes

2026-04-13T22:34:06Z

Currently, the nameserver doesn't limit the number of nodes it handles. This can be an attack vector if a malicious client starts registering random nodes, leading to memory exhaustion. Hence, limit the maximum number of nodes to 64. Note that, limit of 64 is chosen based on the current platform requirements. If requirement changes in the future, this limit can be increased. Cc: stable@vger.kernel.org Fixes: 0c2204a4ad71 ("net: qrtr: Migrate nameservice to kernel from userspace") Signed-off-by: Manivannan Sadhasivam Link: https://patch.msgid.link/20260409-qrtr-fix-v3-4-00a8a5ff2b51@oss.qualcomm.com Signed-off-by: Jakub Kicinski