Jiawen Wu [Fri, 7 Aug 2026 06:22:14 +0000 (14:22 +0800)]
net: ngbe: fix NULL pointer dereference in non-MSI-X interrupt enabling
In non-MSI-X mode (such as legacy INTx or single MSI), wx->msix_entry is
not allocated or initialized. Calling NGBE_INTR_MISC(wx) dereferences
wx->msix_entry->entry, leading to a NULL pointer dereference crash.
This issue was introduced by fixing the IRQ vector when the number of
VFs is 7. Fix the issue by explicitly checking `pdev->msix_enabled` to
determine the correct vector index.
Additionally, as a side fix, set the interrupt mask to BIT(0) for the
non-MSI-X fallback. In MSI/INTx mode, the MISC and queue interrupts
share vector 0, and the WX_PX_MISC_IVAR register is only valid in the
MSI-X case. Thus, BIT(0) is the correct mask for the miscellaneous cause
when MSI-X is disabled.
Fixes: 4174c0c331a2 ("net: ngbe: specify IRQ vector when the number of VFs is 7") Signed-off-by: Jiawen Wu <jiawenwu@trustnetic.com> Reviewed-by: Breno Leitao <leitao@debian.org> Reviewed-by: Simon Horman <horms@kernel.org> Link: https://patch.msgid.link/B2693E9A8BFAD110+20260807062214.410838-1-jiawenwu@trustnetic.com Signed-off-by: Paolo Abeni <pabeni@redhat.com>
====================
Fix wrong transport_header when sending VLAN-tagged frames
When sending a VLAN-tagged frame via AF_PACKET or tap, calling
skb_set_network_header() before skb_probe_transport_header() causes
the flow dissector to misinterpret the inner protocol header as a
VLAN header. As a result, transport_header is never set and remains
at its uninitialized sentinel value (~0U).
Move skb_probe_transport_header() to before skb_set_network_header()
so the flow dissector sees network_header still pointing to the VLAN
header and can correctly identify the transport layer.
====================
Wei Fang [Fri, 7 Aug 2026 06:34:05 +0000 (14:34 +0800)]
net: tap: fix wrong transport_header when sending VLAN-tagged frame
In tap_get_user_xdp(), when processing a VLAN-tagged frame (e.g.
ETH_P_8021Q), skb_set_network_header() is called first to advance
network_header past the VLAN tag to the inner protocol header.
skb_probe_transport_header() is then called with skb->protocol still
set to ETH_P_8021Q, while nhoff (derived from skb_network_offset())
already points past the VLAN tag to the inner protocol header.
In __skb_flow_dissect(), proto is initialized to ETH_P_8021Q and nhoff
points past the VLAN tag. When the dissector hits case ETH_P_8021Q, it
reads a struct vlan_hdr at the current nhoff via __skb_header_pointer(),
but that offset contains the inner protocol header (e.g. an IP header).
The bytes are misinterpreted as a VLAN header, yielding a garbage
encapsulated EtherType that matches no known protocol. The dissector
returns false, so skb_probe_transport_header() never calls
skb_set_transport_header(), leaving transport_header at its uninitialized
sentinel value (~0U).
Move skb_set_network_header() to after skb_probe_transport_header(). At
the time skb_probe_transport_header() is called, network_header still
points to the VLAN header (offset ETH_HLEN), so nhoff is correct and the
flow dissector can parse the VLAN header, extract the inner EtherType,
and advance nhoff to the inner protocol header, allowing transport_header
to be set correctly.
Fixes: 8c76e77f9069 ("tap: call skb_probe_transport_header after setting skb->dev") Assisted-by: WChat:claude-opus-4-8 Signed-off-by: Wei Fang <wei.fang@nxp.com> Reviewed-by: Willem de Bruijn <willemb@google.com> Link: https://patch.msgid.link/20260807063405.688780-3-wei.fang@oss.nxp.com Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Wei Fang [Fri, 7 Aug 2026 06:34:04 +0000 (14:34 +0800)]
net: packet: fix wrong transport_header when sending VLAN-tagged frame
In packet_parse_headers(), when processing a VLAN-tagged frame,
skb_set_network_header() is called to advance network_header past the
VLAN tag to the inner protocol header. skb_probe_transport_header() is
then called with skb->protocol still set to the outer VLAN EtherType
(e.g. ETH_P_8021Q), while nhoff (derived from skb_network_offset())
already points past the VLAN tag to the inner protocol header.
In __skb_flow_dissect(), proto is initialized to ETH_P_8021Q and nhoff
points past the VLAN tag. When the dissector hits case ETH_P_8021Q, it
reads a struct vlan_hdr at nhoff via __skb_header_pointer(), but that
offset contains the inner protocol header (e.g. an IP header). The bytes
are misinterpreted as a VLAN header, yielding a garbage encapsulated
EtherType that matches no known protocol. The dissector returns false,
so skb_probe_transport_header() never calls skb_set_transport_header(),
leaving transport_header at its uninitialized sentinel value (~0U).
Move skb_probe_transport_header() to before skb_set_network_header(). At
the time skb_probe_transport_header() is called, network_header still
points to the VLAN header, so nhoff correctly points to the VLAN header.
The flow dissector can then parse the VLAN header, extract the inner
EtherType, and advance nhoff to the inner protocol header, allowing
transport_header to be set correctly.
Fixes: dfed913e8b55 ("net/af_packet: add VLAN support for AF_PACKET SOCK_RAW GSO") Assisted-by: WChat:claude-opus-4-8 Signed-off-by: Wei Fang <wei.fang@nxp.com> Reviewed-by: Willem de Bruijn <willemb@google.com> Link: https://patch.msgid.link/20260807063405.688780-2-wei.fang@oss.nxp.com Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Baul Lee [Sun, 9 Aug 2026 11:18:29 +0000 (20:18 +0900)]
vxlan: do not arm the ageing timer on a device that is down
vxlan_changelink() arms vxlan->age_timer whenever the requested ageing
interval differs from the configured one:
if (conf.age_interval != vxlan->cfg.age_interval)
mod_timer(&vxlan->age_timer, jiffies);
There is no netif_running() test, so the timer is armed even on a device
that was never brought up. The only synchronous cancel in the driver is
the timer_delete_sync() in vxlan_stop(), which is .ndo_stop.
netif_close_many() drops devices without IFF_UP before
__dev_close_many() runs, so that cancel is skipped for such a device.
vxlan_setup() sets dev->needs_free_netdev = true and age_timer is a
member of struct vxlan_dev, so free_netdev() releases the allocation the
timer lives in while it is still queued on a timer_base.
expire_timers() unlinks the entry before it loads timer->function, so
the timer core writes through the freed object's list pointers:
BUG: KASAN: slab-use-after-free in __run_timers+0x208/0x654
Write of size 8 at addr ffff00001adace68 by task true/192
__asan_store8+0x84/0xac
__run_timers+0x208/0x654
run_timer_softirq+0x154/0x18c
Allocated by task 189:
alloc_netdev_mqs+0x64/0x720
rtnl_create_link+0x4ac/0x520
rtnl_newlink+0x758/0xd00
Freed by task 191:
netdev_release+0x40/0x58
netdev_run_todo+0x4a4/0x8c0
rtnl_dellink+0x200/0x4e8
The rtnl operations involved are netns-scoped, so an unprivileged user
can perform them in a new user and network namespace.
Arming the timer on a down device never had an effect: vxlan_cleanup()
returns early on !netif_running(), and vxlan_open() arms the timer for
any non-zero interval once the device is brought up. Add the missing
test.
Discovered by XBOW, triaged by Baul Lee <baul.lee@xbow.com>
Chengfeng Ye [Fri, 7 Aug 2026 18:17:10 +0000 (02:17 +0800)]
ipv4: fix use-after-free in fib_nhc_update_mtu()
fib_nhc_update_mtu() walks the nexthop exception table under RTNL, but
RTNL does not serialize this walk with PMTU exception updates. The walk
uses rcu_dereference_protected() with a constant true condition without
holding fnhe_lock.
The following interleaving can therefore occur:
CPU 0 CPU 1
fib_nhc_update_mtu() update_or_create_fnhe()
load fnhe spin_lock_bh(&fnhe_lock)
fnhe_remove_oldest()
unlink fnhe
kfree_rcu(fnhe, rcu)
<quiescent state>
access fnhe after grace period
KASAN reported:
BUG: KASAN: slab-use-after-free in fib_nhc_update_mtu+0x3df/0x410
Read of size 8 at addr ffff888107d49000 by task poc/90
Call Trace:
fib_nhc_update_mtu+0x3df/0x410
fib_sync_mtu+0x7a/0xd0
fib_netdev_event+0x229/0x3f0
netif_set_mtu_ext+0x33a/0x570
dev_set_mtu+0x88/0x120
The same walk updates fnhe_pmtu and fnhe_mtu_locked. These fields form a
pair and other writers serialize them with fnhe_lock. RCU alone prevents
reclamation, but would still allow concurrent writers to leave a mixed
pair.
Walk the table under RCU and acquire fnhe_lock only while updating each
exception. RCU keeps the current entry alive while the short critical
section serializes its paired PMTU fields. This avoids holding the global
lock while scanning all 2048 buckets for every nexthop.
Fixes: af7d6cce5369 ("net: ipv4: update fnhe_pmtu when first hop's MTU changes") Cc: stable@vger.kernel.org Suggested-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Chengfeng Ye <nicoyip.dev@gmail.com> Reviewed-by: Ido Schimmel <idosch@nvidia.com> Link: https://patch.msgid.link/20260807181710.1178747-1-nicoyip.dev@gmail.com Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Koichiro Den [Thu, 6 Aug 2026 03:25:37 +0000 (12:25 +0900)]
NTB: ntb_netdev: Preserve RX queue depth on allocation failure
ntb_netdev_rx_handler() hands the received skb to the network stack
before allocating its replacement. If the allocation fails, nothing is
reposted. Every failure therefore takes one buffer out of the RX queue
while the interface remains up, and enough failures eventually stall
reception.
A retry path could refill the queue later, but ntb_netdev has none.
Allocate the replacement first instead. If that fails, drop the packet
and repost the same skb. This keeps the queue full and lets packet
delivery resume as soon as memory is available again.
Fixes: 548c237c0a99 ("net: Add support for NTB virtual ethernet device") Cc: stable@vger.kernel.org Signed-off-by: Koichiro Den <den@valinux.co.jp> Reviewed-by: Dave Jiang <dave.jiang@intel.com> Link: https://patch.msgid.link/20260806032537.3526498-1-den@valinux.co.jp Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Hyunjung Ko [Thu, 6 Aug 2026 10:12:35 +0000 (19:12 +0900)]
selftests: tc-testing: add act_ct test for malformed header handling
Add a tdc case covering the leak fixed by the previous patch.
The test attaches "action ct" to a clsact ingress chain and injects ten
IPv6 frames whose nexthdr says hop-by-hop but which carry nothing after
the 40-byte header, so ipv6_find_hdr() fails and
tcf_ct_ipv6_is_fragment() returns -EPROTO.
Before the fix act_ct returned TC_ACT_CONSUMED for these packets, so
tc_run() never reached its TC_ACT_SHOT arm and the clsact drop counter
stayed at zero while the skbs leaked. After the fix the packets are
dropped properly and the counter reflects them, which is what the test
matches on:
Hyunjung Ko [Thu, 6 Aug 2026 10:12:34 +0000 (19:12 +0900)]
net/sched: act_ct: fix sk_buff leak when the header checks reject a packet
tcf_ct_handle_fragments() runs its header sanity checks before handing
anything to the defragmentation engine:
if (family == NFPROTO_IPV4)
err = tcf_ct_ipv4_is_fragment(skb, &frag);
else
err = tcf_ct_ipv6_is_fragment(skb, &frag);
if (err || !frag)
return err;
tcf_ct_ipv4_is_fragment() returns -EINVAL or -ENOMEM;
tcf_ct_ipv6_is_fragment() adds -EPROTO when ipv6_find_hdr() fails. None of
them frees or queues the skb, so on that path the caller still owns it.
tcf_ct_act() however funnels every non-zero return into the
ownership-transfer exit:
err = tcf_ct_handle_fragments(net, skb, family, p->zone, &defrag);
if (err)
goto out_frag;
...
out_frag:
if (err != -EINPROGRESS)
tcf_action_inc_drop_qstats(&c->common);
return TC_ACT_CONSUMED;
TC_ACT_CONSUMED means the action took ownership of the skb, so no caller
frees it - sch_handle_ingress(), sch_handle_egress() and
tcf_qevent_handle() all deliberately skip the free for that verdict. The
skb is therefore orphaned: one sk_buff plus its data buffer is leaked per
malformed packet, unbounded. Note the drop counter is already incremented
for these errors, so the statistics claim a drop that never happens.
Three different ownership states reach out_frag: today - the skb may be
queued by the defrag engine (-EINPROGRESS), already freed by
nf_ct_handle_fragments(), or still owned by us. Tell the caller which of
those it is, and free the packet ourselves in the last case, which
restores the TC_ACT_SHOT behaviour that predated the Fixes: commit.
Reproduced on v7.2-rc6 with a 54-byte frame carrying a 40-byte IPv6
header with nexthdr = 0 (hop-by-hop) and nothing after it, on a
clsact ingress chain with "action ct". kmemleak reports one leaked
232-byte skbuff_head_cache object plus its 704-byte data buffer per
packet; with this patch it reports none.
Fixes: 3f14b377d01d ("net/sched: act_ct: fix skb leak and crash on ooo frags") Cc: stable@vger.kernel.org # v6.8+ Signed-off-by: Hyunjung Ko <hj351016@gmail.com> Acked-by: Jamal Hadi Salim <jhs@mojatatu.com> Link: https://patch.msgid.link/20260806101235.809370-1-hj351016@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Oleksij Rempel [Thu, 6 Aug 2026 13:47:16 +0000 (15:47 +0200)]
net: phy: realtek: fix EEE advertisement write on the internal PHY MMD path
In rtlgen_write_mmd(), the MDIO_AN_EEE_ADV case swaps the arguments to
rtlgen_write_vend2(): it passes the MMD register number as the OCP address
and the OCP address constant as the value. The caller's value is discarded
and the write lands on the wrong register, so the EEE advertisement cannot
be configured on the affected PHYs.
Mirror rtlgen_read_mmd() and write the value to RTL_MDIO_AN_EEE_ADV.
Fixes: da681ed73fb9 ("net: phy: realtek: improve mmd register access for internal PHY's") Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Reviewed-by: Nicolai Buchwitz <nb@tipi-net.de> Link: https://patch.msgid.link/20260806134716.3511821-1-o.rempel@pengutronix.de Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jiayuan Chen [Fri, 7 Aug 2026 01:44:36 +0000 (09:44 +0800)]
tcp: fix icsk_ack.ato bitfield overflow
On cross-region connections we observed delayed ACKs suddenly turning
into immediate ACKs plus a TCP_MAX_QUICKACKS burst, as if the
connection had just received its first data segment.
Commit 95b9a87c6a6b ("tcp: record last received ipv6 flowlabel")
squeezed icsk_ack.ato into 8 bits, sized for TCP_DELACK_MAX. But both
writers still bound ato by icsk_rto, which can be well above 255
jiffies, so the bitfield assignment silently wraps mod 256: repeated
delack timer misses double ato up to icsk_rto, storing 320 as 64 and
256 as 0, and ato == 0 is the "first data packet" sentinel in
tcp_event_data_recv().
Clamp both writers to TCP_DELACK_MAX, which the static_assert already
guarantees to fit and tcp_send_delayed_ack() effectively caps ato at
anyway.
Fixes: 95b9a87c6a6b ("tcp: record last received ipv6 flowlabel") Signed-off-by: Jiayuan Chen <jiayuan.chen@linux.dev> Reviewed-by: Neal Cardwell <ncardwell@google.com> Link: https://patch.msgid.link/20260807014437.36687-1-jiayuan.chen@linux.dev Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Hyunjung Ko [Thu, 6 Aug 2026 10:12:52 +0000 (19:12 +0900)]
net/sched: act_gact, act_police: range check the fallback control action
tcf_action_check_ctrlact() range checks the primary control action:
if (!opcode)
ret = action > TC_ACT_VALUE_MAX ? -EINVAL : 0;
TC_ACT_VALUE_MAX is TC_ACT_TRAP, so kernel-internal verdicts above it
cannot be set that way. But act_gact and act_police each carry a second,
independent control action supplied by user space that never reaches that
helper - TCA_GACT_PROB.paction and TCA_POLICE_RESULT. Both only reject
TC_ACT_GOTO_CHAIN, so any other value is stored verbatim and returned
verbatim from the action.
In particular user space can store TC_ACT_CONSUMED, which is
TC_ACT_VALUE_MAX + 1 and is deliberately not part of the UAPI value
range. That verdict tells every caller the action took ownership of the
skb, so nobody frees it: sch_handle_ingress(), sch_handle_egress() and
tcf_qevent_handle() all deliberately skip the free for it. The result is
one leaked sk_buff plus its data buffer per packet traversing the filter,
unbounded, for all traffic on the chain including kernel-generated
packets.
Both are trivially deterministic. act_gact clamps tcfg_pval to >= 1, so
with pval = 1 gact_determ() returns the fallback for every packet.
act_police has no mandatory rate, so rate = 0 leaves tcfp_mtu = ~0 and
tcf_police_mtu_check() always passes.
TC_ACT_CONSUMED was added by commit 720f22fed81b ("net: sched: refactor
reinsert action"), after both goto-chain guards were written:
commit 9469f375ab09 ("net/sched: act_gact: disallow 'goto chain' on
fallback control action") and
commit c08f5ed5d625 ("net/sched: act_police: disallow 'goto chain' on
fallback control action"). Neither guard was widened when the new
verdict appeared.
Factor the existing range test out of tcf_action_check_ctrlact() as
tcf_action_valid() and apply it to both fallbacks. The helper cannot call
tcf_action_check_ctrlact() directly because that also allocates a
goto_chain, which is exactly what these two sites must not do.
Reproduced on v7.2-rc6: kmemleak reports one leaked 232-byte
skbuff_head_cache object plus its 704-byte data buffer per packet. With
this patch both configurations are rejected with -EINVAL and kmemleak
reports none.
Fixes: 720f22fed81b ("net: sched: refactor reinsert action") Cc: stable@vger.kernel.org # v5.3+ Signed-off-by: Hyunjung Ko <hj351016@gmail.com> Acked-by: Jamal Hadi Salim <jhs@mojatatu.com> Tested-by: Victor Nogueira <victor@mojatatu.com> Link: https://patch.msgid.link/20260806101252.809593-1-hj351016@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jonas Köppeler [Thu, 6 Aug 2026 11:43:49 +0000 (13:43 +0200)]
veth: fix queue index used to wake the peer txq in veth_poll
veth_poll() derives the index of the peer TX queue to wake from
rq->xdp_rxq.queue_index. That field is only initialized by
xdp_rxq_info_reg() in veth_enable_xdp_range(), which runs only when an
XDP program is attached. On the plain GRO/NAPI path
(veth_napi_enable_range()) xdp_rxq_info_reg() is never called, so
queue_index stays 0 for every queue, as priv->rq is zero-allocated.
So in a multi-queue setup with GRO enabled and no XDP program attached,
every NAPI instance looks at the peer's TX queue 0. If veth_xmit() stops
peer TX queue 1 because the ptr_ring is full (NETDEV_TX_BUSY), nothing
ever wakes it again: the poller draining queue 1 wakes queue 0 instead.
veth implements no ndo_tx_timeout, so the netdev watchdog does not kick
in either, and the queue stays stopped indefinitely.
Derive the index from the position of the rq within priv->rq instead,
which is correct regardless of whether XDP was ever enabled.
Scripts to reproduce the stall are available at
https://github.com/netoptimizer/veth-backpressure-performance-testing
Fixes: dc82a33297fc ("veth: apply qdisc backpressure on full ptr_ring to reduce TX drops") Signed-off-by: Jonas Köppeler <j.koeppeler@tu-berlin.de> Tested-by: Jesper Dangaard Brouer <hawk@kernel.org> Acked-by: Jesper Dangaard Brouer <hawk@kernel.org> Link: https://patch.msgid.link/20260806-veth-fix-poll-queue-idx-v1-1-c5357fb7573d@tu-berlin.de Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jakub Kicinski [Thu, 6 Aug 2026 22:56:26 +0000 (15:56 -0700)]
net: expect instance lock in netdev_queue_get_dma_dev()
netdev_queue_get_dma_dev() uses "compat" locking assert which wants
either the rtnl_lock or netdev instance lock. This is not right,
the callers are taking the instance lock unconditionally. All entry
points for queue config are purely instance locked.
In other words the callers use netdev_get_by_index_lock(), not
netdev_get_by_index_lock_ops_compat(). All the state we will
access is effectively instance lock protected (it's const for
devices which are not ops-locked).
Update the assert to avoid false positive warnings.
Cc: stable@vger.kernel.org Fixes: b6c5f9454ef34 ("io_uring/zcrx: call netdev_queue_get_dma_dev() under instance lock") Reported-by: syzbot+a78926bdac2adb52dc0e@syzkaller.appspotmail.com Reviewed-by: Simon Horman <horms@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Link: https://patch.msgid.link/20260806225627.3998672-1-kuba@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Eric Dumazet [Thu, 6 Aug 2026 14:19:38 +0000 (14:19 +0000)]
macvlan: inherit needed_headroom and needed_tailroom from lowerdev
macvlan devices inherit hard_header_len from lowerdev during macvlan_init(),
but leave needed_headroom and needed_tailroom set to 0.
When the underlying lowerdev requires extra headroom or tailroom for
headers/trailers (e.g. macsec, ipsec, wireguard, tunnels, or veth with rx
headroom), upper layers calculating packet headroom and tailroom fail to
reserve sufficient space.
This can result in reallocation overhead, skb headroom underflows, or KASAN
slab-use-after-free crashes when dev_hard_header() / macvlan_hard_header()
prepends header data or when lower devices append tailroom.
Fix this by:
1. Inheriting needed_headroom and needed_tailroom from lowerdev in macvlan_init().
2. Propagating needed_headroom and needed_tailroom updates to attached macvlans
in macvlan_device_event() when receiving NETDEV_FEAT_CHANGE events.
Fixes: b863ceb7ddce ("[NET]: Add macvlan driver") Reported-by: Tangxin Xie <xietangxin@h-partners.com> Closes: https://lore.kernel.org/netdev/CANn89i+1EW-sFNK8xoq98gMbPCeLS7e=+rs9gHfLg5Wj+4x0sw@mail.gmail.com/T/#m16adf0ff972cbfd8066c3a8e656e75eaeb12d021 Signed-off-by: Eric Dumazet <edumazet@google.com> Reviewed-by: Hangbin Liu <liuhangbin@kylinos.cn> Link: https://patch.msgid.link/20260806141938.287660-1-edumazet@google.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Eric Dumazet [Thu, 6 Aug 2026 10:38:57 +0000 (10:38 +0000)]
ipvlan: inherit needed_headroom and needed_tailroom from phy_dev
ipvlan devices inherit hard_header_len from phy_dev during ipvlan_init(),
but leave needed_headroom and needed_tailroom set to 0.
When the underlying phy_dev (or stacked lower device) requires extra headroom
or tailroom for headers/trailers (e.g. macsec, ipsec, wireguard, tunnels, or
veth with rx headroom), upper layers calculating packet headroom and tailroom
fail to reserve sufficient space.
This can result in reallocation overhead, skb headroom underflows, or KASAN
slab-use-after-free crashes when dev_hard_header() / ipvlan_hard_header()
prepends header data or when lower devices append tailroom.
Fix this by:
1. Inheriting needed_headroom and needed_tailroom from phy_dev in ipvlan_init().
2. Propagating needed_headroom and needed_tailroom updates to attached ipvlans
in ipvlan_device_event() when receiving NETDEV_FEAT_CHANGE events.
Fixes: 2ad7bf363841 ("ipvlan: Initial check-in of the IPVLAN driver.") Reported-by: syzbot+1f9fd0f4b601cf88d6e6@syzkaller.appspotmail.com Closes: https://lore.kernel.org/netdev/6a720a21.40259c87.584f4.04bb.GAE@google.com/T/#u Reported-by: Tangxin Xie <xietangxin@h-partners.com> Closes: https://lore.kernel.org/netdev/CANn89i+1EW-sFNK8xoq98gMbPCeLS7e=+rs9gHfLg5Wj+4x0sw@mail.gmail.com/T/#mcc6307f115e500df23ea2980d5669fe95f20b6b4 Signed-off-by: Eric Dumazet <edumazet@google.com> Reviewed-by: Hangbin Liu <liuhangbin@kylinos.cn> Link: https://patch.msgid.link/20260806103857.115541-1-edumazet@google.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
I was trying to make bnxt preserve IRQ mappings across reconfiguration.
While hacking on that I noticed 2 bugs in the notifiers that should
probably be fixed before development work.
First one is simple - TPH recofig makes aARFs not work. There can only
be one notifier per IRQ and TPH "steals" the callback from the rmap
updates. Fix by patches 1 and 2.
Second one is a deadlock between the affinity notifier and reconfig.
This one is a bit more involved (patch 3 and 4).
Unfortunately, I can't really verify the problem or test the fix.
I managed to get my hands on a system with an AMD Venice CPU which
is supposed to support TPH, but the ACPI is missing some bits to
actually advertise it. pcie_tph_get_cpu_st() returns -EINVAL.
====================
Jakub Kicinski [Mon, 3 Aug 2026 19:31:35 +0000 (12:31 -0700)]
eth: bnxt: avoid deadlock when canceling IRQ affinity notifier
Unregistering IRQ affinity notifiers waits for the callback synchronously.
bnxt takes the netdev instance lock in the notifier (to restart the queue)
and cancels the work under the same lock. This may obviously deadlock.
Move the restart to the async service task. The queue restart isn't
super time sensitive. Store the new TPH tag, schedule the task.
Safely canceling the service task is already ironed out.
In bnxt_request_irq() the order of registering notifier, affinity and
initial TPH programming has to be inverted. I think it was racy
previously since user may trigger an update as soon as notifier
is installed.
There's a small known gap - if pcie_tph_get_cpu_st() fails at init
and the target tag is 0 we may miss programming the entry.
This does not seem worth fixing, the code has skip-on-failure
all over the place, anyway.
Jakub Kicinski [Mon, 3 Aug 2026 19:31:33 +0000 (12:31 -0700)]
eth: bnxt: keep the aRFS rmap updated when TPH is enabled
The TPH support must have broken aRFS in bnxt. IRQ can only have one
notifier, so installing the TPH notifier is overriding the one implicitly
installed by irq_cpu_rmap_add().
Make sure we call cpu_rmap_update() from the TPH notifier.
We need to be careful with the ordering and not free the rmap
until we unregistered the notifier. Note that moving the rmap
freeing after the early return in bnxt_free_irq() is fine -
there's no path that could leave rmap with irq_tbl being NULL.
Map the email addresses used for previous kernel contributions to the
current OpenAI address. This prevents get_maintainer.pl from listing
historical addresses as patch recipients.
====================
fix skb length accounting after XDP frag adjustment
This series fixes skb length accounting after an XDP program adjusts its
fragment area, in both the generic XDP path (net/core/dev.c) and the veth
native path (drivers/net/veth.c). When the fragment area is resized,
skb->len and skb->data_len can go out of sync, and in the reproduced UDP
receive path this leaked skb_shared_info contents (including a kernel
pointer) to userspace while truncating real payload.
====================
Sun Jian [Tue, 4 Aug 2026 05:40:39 +0000 (22:40 -0700)]
veth: fix skb length accounting after XDP frag adjustment
veth exposes non-linear skb fragments through an xdp_buff. If an XDP
program adjusts the fragment area, veth_xdp_rcv_skb() copies
xdp_frags_size back to skb->data_len but leaves skb->len containing the
old fragment contribution.
After a fragment shrink, this makes skb_headlen() larger than the actual
linear area. In the reproduced UDP receive path, __skb_datagram_iter()
copied 1024 bytes past the actual linear tail to userspace, starting at
struct skb_shared_info. The copied bytes included the affected skb's
nr_frags, xdp_frags_size, and a kernel pointer from
skb_shinfo(skb)->frags[0]. Real packet data was displaced by the same
amount and truncated at the end.
Subtract the old data_len before replacing it and add the new data_len
afterwards, keeping skb->len and skb->data_len synchronized.
Additionally, bpf_xdp_pull_data() can advance data_end while leaving
frags present. The skb is then still non-linear, so the old
__skb_put(skb, off) triggers SKB_LINEAR_ASSERT().
Use skb_set_tail_pointer() and update skb->len explicitly instead,
following bpf_prog_run_generic_xdp(). Unlike __skb_put(),
skb_set_tail_pointer() does not require a linear skb.
A 60000-byte UDP datagram on a veth pair with MTU 64000 was shortened by
1024 bytes from its fragment area. Before the fix, all 10 runs produced
corrupted payloads. After the fix, all 10 runs matched the expected
payload exactly. A forced-tailroom reproducer also exercises
bpf_xdp_pull_data() with frags still present; the old code triggers
SKB_LINEAR_ASSERT(), while this fix passes 10/10 runs.
Sun Jian [Tue, 4 Aug 2026 05:40:38 +0000 (22:40 -0700)]
net: fix skb length accounting after generic XDP frag adjustment
Generic XDP exposes non-linear skb fragments through an xdp_buff. If an
XDP program adjusts the fragment area, bpf_prog_run_generic_xdp() copies
xdp_frags_size back to skb->data_len but leaves skb->len containing the
old fragment contribution.
After a fragment shrink, this makes skb_headlen() larger than the actual
linear area. In the reproduced UDP receive path, __skb_datagram_iter()
copied 1024 bytes past the actual linear tail to userspace, starting at
struct skb_shared_info. The copied bytes included the affected skb's
nr_frags, xdp_frags_size and a kernel pointer from
skb_shinfo(skb)->frags[0]. Real packet data was displaced by the same
amount and truncated at the end.
Subtract the old data_len before replacing it and add the new data_len
afterwards, keeping skb->len and skb->data_len synchronized.
A 60000-byte UDP datagram on a veth pair with MTU 64000 was shortened by
1024 bytes from its fragment area. Before the fix, all 10 runs produced
corrupted payloads. After the fix, all 10 runs matched the expected
payload exactly.
Jérémy Jean [Tue, 4 Aug 2026 20:00:42 +0000 (20:00 +0000)]
sctp: validate cookie AUTH state before use
When cookie authentication is disabled, COOKIE_ECHO restores fixed-size
AUTH fields directly from peer-controlled cookie bytes. A forged RANDOM
length, HMAC list, or CHUNKS list can then reach association consumers
with lengths or identifiers that were never validated against the local
backing arrays.
A forged RANDOM length can cause out-of-bounds reads during key-vector
construction. A forged HMAC identifier also caused a 32-byte write past
a zero-length AUTH chunk, providing a primitive for a local privilege
escalation chain.
Validate the cookie's RANDOM, HMACS, and CHUNKS parameters at the cookie
trust boundary before copying them into the association. Reject invalid
types, malformed lengths, unsupported HMAC identifiers, HMAC lists
without SHA1, and forbidden chunk ids.
Fixes: bbd0d59809f9 ("[SCTP]: Implement the receive and verification of AUTH chunk") Fixes: 1f485649f529 ("[SCTP]: Implement SCTP-AUTH internals") Signed-off-by: Jérémy Jean <Jeremy.Jean@oss.cyber.gouv.fr> Acked-by: Xin Long <lucien.xin@gmail.com> Link: https://patch.msgid.link/20260804200042.2412009-1-Jeremy.Jean@oss.cyber.gouv.fr Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Kyle Zeng reported that GC could free a dead SCC partially.
The scenario is as follows:
1) Create two SCCs:
X -. A <-> B
^--'
2) Run the following concurrently:
2-1) send() sk-B to sk-B from sk-X
2-2) close() both A and B
At 2-1), there is a small window where unix_add_edges()
publishes a new edge (B <-> B) to GC but its skb is not queued
by skb_queue_tail().
If 2-2) completes before skb_queue_tail() and GC is triggered,
it judges A <-> B as dead, but B is not freed because GC cannot
collect the not-yet-queued skb holding the B <-> B edge.
X -. A <-> B -. This edge is visible
^--' ^..' but skb is not
This itself is not a problem since the next GC run will judge
B as dead as well and free it finally.
X -. A <.> B -.
^--' ^--'
However, X's SCC forces the next GC to call unix_walk_scc_fast(),
and it iterates over A through B's scc_entry.
Let's unlink scc_entry before freeing the vertex in unix_del_edge().
Linus Torvalds [Thu, 6 Aug 2026 18:39:20 +0000 (11:39 -0700)]
Merge tag 'net-7.2-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Pull networking fixes from Jakub Kicinski:
"Including fixes from netfilter.
Looks like our attempt to keep the PRs smaller have only prevented
this one from getting even bigger. In the last 9 days there were
405 postings explicitly tagged with [PATCH net], vs 687 with [PATCH
net-next]. 37% of posted patches being fixes is pretty crazy, and
that's likely undercounting because LLM "researchers" more often post
fixes without knowing to tag the patches for specific trees. I don't
have historic data.
In any case, we keep adjusting the criteria. The next PR will be
smaller.
Current release - regressions:
- net: defer netdev KOBJ_ADD uevent until the device is published,
previously rtnl_lock would serialize the accesses vs publishing
- net: explicitly cancel work to avoid races with ref tracker exit
- qrtr: ns: raise lookup limit to 128
- eth: hns3: fix speed configuration residue after driver reload
Previous releases - regressions:
- tcp: do not change rcv_ssthresh in tcp_measure_rcv_mss(), regressed
flows with MSS and scaling_ratio variability
- Revert "net: thunderbolt: Enable end-to-end flow control also in
transmit", broke some platforms (no packets coming thru)
- eth: stmmac: resume PHY before hardware setup when opening the
interface
Previous releases - always broken:
- another pile of fixes for less common protocols (SCTP, TLS, SMC
etc.)
- close a couple of AF_PACKET bugs and ways it can build skbs
problematic for the rest of the stack
- bridge: mrp: fix uninitialised bytes on the wire
- eth: atlantic: free RX pages of consumed but not refilled buffers"
* tag 'net-7.2-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (116 commits)
igc: fix netdev not re-attached after resume if interface is down
tls: don't abort the connection on signal-interrupted sends
net: avoid theoretical races with ref drain
net: Defer netdev KOBJ_ADD uevent until the device is published
MAINTAINERS: dpll: zl3073x: replace Prathosh Satish with Min Li
sctp: clear control chunk transport if it is being removed
net/atm: fix slab-out-of-bounds read in vcc_setsockopt()
s390/ism: Fix UAF of sba and ieq during ism_dev_exit()
packet: use consistent hard_header_len in TX_RING send path
packet: use consistent hard_header_len in non-ring send paths
net: remove CAP_SYS_RAWIO zero-padding in dev_validate_header
bnge: Fix resource leak in bnge_init_nic() error path
ptp: ocp: Fix board ID over-read
tls: rx: restore msg_iter before TLS 1.3 optimistic retry
selftests: tls: add a test for splicing onto a full plaintext record
tls: don't leave a full plaintext sk_msg ring unpushed
xdp: reject clones that overrun skb_shared_info tailroom
mptcp: reclaim forward-allocated memory on RX path errors
mptcp: fastopen: only mark MPTFO subflows with SYN data
mptcp: pm: fix memory leak from alloc-during-teardown race
...
Philipp David [Tue, 4 Aug 2026 22:22:03 +0000 (15:22 -0700)]
igc: fix netdev not re-attached after resume if interface is down
__igc_resume() calls netif_device_attach() only inside the
netif_running() branch, so an interface that was down during suspend
is never re-attached on resume. It then stays in the not-present state
that __igc_shutdown() set via netif_device_detach(): ethtool reports
ENODEV and every attempt to bring the interface up fails the
netif_device_present() check in __dev_open() with -ENODEV, silently,
since __igc_resume() returns 0. Only reloading the driver recovers the
device.
This is easy to hit in practice because NetworkManager brings managed
interfaces down before sleep unless Wake-on-LAN is configured, making
the adapter unusable after every suspend/resume cycle with WoL
disabled.
Re-attach the netdev on every successful resume, as igb and e1000e do.
Fixes: 6f31d6b643a3 ("igc: Refactor runtime power management flow") Cc: stable@vger.kernel.org Signed-off-by: Philipp David <pd-lkml@3b.pm> Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com> Reviewed-by: Dima Ruinskiy <dima.ruinskiy@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com> Link: https://patch.msgid.link/20260804222205.1580328-11-anthony.l.nguyen@intel.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
tls: don't abort the connection on signal-interrupted sends
When a signal interrupts a blocking send, tls_tx_records() treats the
resulting -ERESTARTSYS as a transmission failure and marks the socket
errored via tls_err_abort() with the raw error code. Later syscalls
return the kernel-internal errno 512 (ERESTARTSYS) to userspace, as the
signal it stems from is no longer pending during syscall exit and thus
never translated.
An interrupted send is not a connection error: the partially sent record
stays queued and is resent later. Interrupt error codes are therefore
excluded from the abort in the same way as -EAGAIN.
Jakub Kicinski [Thu, 6 Aug 2026 02:28:21 +0000 (19:28 -0700)]
net: avoid theoretical races with ref drain
Technically, it's illegal to take a ref on a netdev just because
we have a pointer on which we already hold a ref, with no other
protection. This is because our simple per-cpu refcount
implementation cannot atomically read the count.
Let's make sure we cancel outstanding work and never queue more
work for a device we know is dead. This way taking a ref on
a dev we know is on the netdev_work_list is always going to be safe.
Jiangshan Yi reports that the issues is caught by ref tracker infra
leading to a warning:
WARNING: lib/ref_tracker.c:322 at ref_tracker_free
WARNING: lib/ref_tracker.c:246 at ref_tracker_dir_exit
Dragos Tatulea [Thu, 6 Aug 2026 08:07:58 +0000 (11:07 +0300)]
net: Defer netdev KOBJ_ADD uevent until the device is published
netdev_register_kobject() calls device_add(), which emits KOBJ_ADD and
wakes udev, but register_netdevice() only makes the device findable by
name later, in list_netdevice(). A udev worker that reacts to the uevent
can therefore run against a device that no lookup can find yet.
This used to be harmless because the ethtool ioctl took the rtnl_lock
when looking the device up, and register_netdevice() runs under rtnl, so
the worker simply blocked until registration finished. The commit in the
fixes tag moved the lookup out from under rtnl for ops-locked drivers.
Now there is a short window in register_netdevice() between
netdev_register_kobject() until list_netdevice() when the device is not
findable by name.
This was reproduced with the mlx5 driver on a kernel with KASAN enabled
during devlink reload: systemd-udevd's net_driver builtin gets -ENODEV
from ETHTOOL_GDRVINFO, which was preventing interface renaming.
Suppress the uevent in netdev_register_kobject() and emit it from
register_netdevice() next to rtmsg_ifinfo(). This is the last point in
register_netdevice() where no error can happen, so only fully registered
devices are announced: the registration error paths never reach it, and
the device_del() that unwinds them stays silent as well, leaving
userspace with neither an add nor a remove.
Fixes: f994752b1127 ("net: ethtool: optionally skip rtnl_lock on IOCTL path") Signed-off-by: Dragos Tatulea <dtatulea@nvidia.com> Reviewed-by: Shahar Shitrit <shshitrit@nvidia.com> Link: https://patch.msgid.link/20260806080758.2039586-2-dtatulea@nvidia.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Xin Long [Wed, 5 Aug 2026 15:18:40 +0000 (11:18 -0400)]
sctp: clear control chunk transport if it is being removed
sctp_make_heartbeat_ack() caches the destination transport in
chunk->transport without taking a reference. When src_out_of_asoc_ok is
enabled, the HEARTBEAT ACK may remain queued on control_chunk_list instead
of being transmitted immediately.
If the peer transport is removed while the chunk is still queued,
sctp_assoc_rm_peer() drops the transport and schedules it for RCU freeing,
but only clears cached transport pointers in out_chunk_list. The queued
control chunk therefore retains a dangling transport pointer.
Once an ASCONF_ACK clears the suppression and the queued control chunk is
transmitted, SCTP dereferences the stale transport pointer, leading to a
use-after-free.
Fix this by also clearing chunk->transport for queued control chunks in
control_chunk_list when removing the transport.
Eric Dumazet [Wed, 5 Aug 2026 13:15:08 +0000 (13:15 +0000)]
net/atm: fix slab-out-of-bounds read in vcc_setsockopt()
vcc_setsockopt() contained an ineffective optlen check:
if (__SO_LEVEL_MATCH(optname, level) && optlen != __SO_SIZE(optname))
return -EINVAL;
If __SO_LEVEL_MATCH(optname, level) evaluated to false (e.g. if the caller
passed a mismatched level), the length check optlen != __SO_SIZE(optname)
was short-circuited and bypassed. Execution then fell through to switch(optname),
calling copy_from_sockptr() assuming optval contained sufficient space.
Furthermore, even if level matched, a cgroup BPF setsockopt filter could shrink
optlen after entry. Because copy_from_sockptr() on kernel pointers uses memcpy(),
this leads to a KASAN slab-out-of-bounds read when optlen is smaller than the
expected structure size.
Fix this by using copy_safe_from_sockptr(), which unconditionally validates
that optlen is at least the expected size before copying. Also change the local
'value' variable type from 'unsigned long' to 'int' so that SO_SETCLP matches
its sizeof(int) ABI encoding on 64-bit systems.
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Reported-by: syzbot+53ecc09fb81df10ef4de@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=53ecc09fb81df10ef4de Signed-off-by: Eric Dumazet <edumazet@google.com> Link: https://patch.msgid.link/20260805131508.3227331-1-edumazet@google.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
====================
net: fix hard_header_len races in packet send paths
The packet socket TX paths read dev->hard_header_len independently for
skb allocation and header construction. Concurrent netdevice
reconfiguration (e.g. bonding device type changes) can change this value
in between, leading to mismatched headroom and copy length, and in the
SOCK_RAW case to out-of-bounds writes.
Patch 1 removes the CAP_SYS_RAWIO zero-padding branch in
dev_validate_header(). That branch sizes a memset against the live
dev->hard_header_len while operating on an skb whose headroom was
allocated from an earlier hard_header_len read, so a concurrent increase
can write past the reserved buffer. Removing it first keeps the later
snapshot fixes bisect-safe: they do not replace an earlier skb_under_panic
with a silent overwrite.
Patches 2 and 3 snapshot hard_header_len once per send and use it
consistently for allocation and construction, in the non-ring and TX_RING
paths respectively. The separate SOCK_DGRAM consistency problem between
hard_header_len and header_ops->create remains out of scope, as noted in
the commit messages.
====================
Qihang Tang [Wed, 5 Aug 2026 12:57:29 +0000 (20:57 +0800)]
packet: use consistent hard_header_len in TX_RING send path
tpacket_snd() reads dev->hard_header_len independently for skb
allocation and header construction in tpacket_fill_skb(). Concurrent
netdevice reconfiguration can therefore make the reserved headroom
smaller than the amount later pushed, or make copylen - hard_header_len
negative.
Snapshot hard_header_len once before processing ring frames and use it
for the frame limit, headroom allocation, copy length, and skb
construction. Pass the snapshot to tpacket_fill_skb().
The separate SOCK_DGRAM consistency problem between hard_header_len and
header_ops->create is not addressed here.
Fixes: 69e3c75f4d54 ("net: TX_RING and packet mmap") Cc: stable@vger.kernel.org Signed-off-by: Qihang Tang <q.h.hack.winter@gmail.com> Reviewed-by: Willem de Bruijn <willemb@google.com> Link: https://patch.msgid.link/20260805125729.19220-4-q.h.hack.winter@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Qihang Tang [Wed, 5 Aug 2026 12:57:28 +0000 (20:57 +0800)]
packet: use consistent hard_header_len in non-ring send paths
packet_snd() reads dev->hard_header_len multiple times while allocating
and constructing an skb. Device reconfiguration can change this value
concurrently, for example through bonding device type changes.
For SOCK_RAW, packet_snd() can save a larger value in reserve and later
allocate headroom using a smaller value. Moving skb->data back by reserve
then places it before skb->head, and the following copy from userspace can
attempt an out-of-bounds write.
packet_sendmsg_spkt() has the same issue because it calculates its
reservation and header offset from separate reads before dropping the RCU
read lock to allocate the skb.
Add LL_RESERVED_SPACE_EX() for callers that already saved a header length.
Read hard_header_len once in packet_snd() and use it for allocation and
construction. In packet_sendmsg_spkt(), preserve the allocation-time value
through the device lookup retry.
The separate SOCK_DGRAM consistency problem between hard_header_len and
header_ops->create is not addressed here.
Fixes: b84bbaf7a6c8 ("packet: in packet_snd start writing at link layer allocation") Cc: stable@vger.kernel.org Signed-off-by: Qihang Tang <q.h.hack.winter@gmail.com> Reviewed-by: Willem de Bruijn <willemb@google.com> Link: https://patch.msgid.link/20260805125729.19220-3-q.h.hack.winter@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Packet send paths call dev_validate_header() on skbs whose headroom was
allocated from an earlier hard_header_len read. If the device is
reconfigured so that dev->hard_header_len increases before validation,
the memset writes past the reserved buffer, an out-of-bounds write.
This out-of-bounds write is masked in some SOCK_RAW paths today because
the same concurrent increase can first make skb_push() exceed the
reserved headroom and trigger skb_under_panic(). Remove the zero-padding
branch before making those hard_header_len reads consistent, so the
snapshot fixes do not turn a loud panic into a silent overwrite.
This path is only reached for variable length L2 protocols, where
len < hard_header_len but len >= min_header_len. No remaining in-tree
variable length L2 protocol implements header_ops->validate, and the
CAP_SYS_RAWIO bypass that zero-pads and accepts short headers has no
real value beyond allowing testing of intentionally malformed input.
Drop the CAP_SYS_RAWIO branch. The remaining reads of
dev->hard_header_len in dev_validate_header() are comparisons only and
have no memory safety impact.
Suggested-by: Willem de Bruijn <willemb@google.com> Fixes: 2793a23aacbd ("net: validate variable length ll headers") Cc: stable@vger.kernel.org Signed-off-by: Qihang Tang <q.h.hack.winter@gmail.com> Reviewed-by: Willem de Bruijn <willemb@google.com> Link: https://patch.msgid.link/20260805125729.19220-2-q.h.hack.winter@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Ahmad Byagowi [Tue, 4 Aug 2026 21:07:51 +0000 (14:07 -0700)]
ptp: ocp: Fix board ID over-read
The EEPROM board ID is a fixed 13-byte field and is not guaranteed to
contain a NUL terminator. Passing it directly to
devlink_info_version_fixed_put() treats it as a C string and may read
beyond the field.
Format at most OCP_BOARD_ID_LEN bytes into the existing local buffer
before reporting the ID. Use a precision limit because the snprintf()
output size alone does not bound the source string scan.
Fixes: 0cfcdd1ebcfe ("ptp: ocp: add nvmem interface for accessing eeprom") Cc: stable@vger.kernel.org Signed-off-by: Ahmad Byagowi <ahmadexp@gmail.com> Reviewed-by: Vadim Fedorenko <vadim.fedorenko@linux.dev> Link: https://patch.msgid.link/20260804210751.48248-1-ahmadexp@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jérémy Jean [Tue, 4 Aug 2026 12:55:28 +0000 (12:55 +0000)]
tls: rx: restore msg_iter before TLS 1.3 optimistic retry
tls_decrypt_sg() advances msg->msg_iter when it maps user pages for
the optimistic TLS 1.3 zero-copy path. If the decrypted record turns
out not to be unpadded application data, tls_decrypt_sw() retries into
a kernel skb, but leaves the iterator advanced.
The subsequent copy from the skb then writes decrypted bytes again at
a later point in the caller iovecs while recvmsg() reports only the
post-retry length. A TLS peer can trigger this after the receiver
enables TLS_RX_EXPECT_NO_PAD.
Revert the iterator by the number of bytes consumed by the optimistic
mapping before retrying without zero-copy.
Add a selftest which sends a TLS 1.3 control record with
TLS_RX_EXPECT_NO_PAD enabled and verifies that recvmsg() does not
overwrite later iovecs beyond the returned length.
====================
tls: fix plaintext sk_msg ring over-fill
An unprivileged user can oops the kernel by splicing into a kTLS socket
whose open record already has a full plaintext sk_msg ring. Reproduced on
net (53658c6f3682) with a stock config, no KASAN.
Patch 2 oopses an unpatched kernel and passes with patch 1 applied.
====================
chanyoung [Tue, 4 Aug 2026 05:28:36 +0000 (14:28 +0900)]
selftests: tls: add a test for splicing onto a full plaintext record
Splicing onto a plaintext sk_msg ring that is already full used to wrap the
ring and make the kernel oops in the scatterwalk once the record was
pushed.
Only the copy path leaves the ring full without pushing it, so splice until
the ring is one fragment short, add the last fragment with a one-byte
MSG_MORE send, and splice once more before pushing the record.
CONFIG_MAX_SKB_FRAGS is 17..45, so that last fragment follows between 16
and 44 splices; sweep that range to trigger the bug on any build.
chanyoung [Tue, 4 Aug 2026 05:28:35 +0000 (14:28 +0900)]
tls: don't leave a full plaintext sk_msg ring unpushed
When the copy path in tls_sw_sendmsg_locked() adds the fragment that fills
the plaintext sk_msg ring, it does not set full_record, so the record is
left full and unpushed. A later splice() then adds to an already full
ring: sk_msg_page_add() has no fullness check of its own, so sg.end wraps
onto sg.start and the ring appears empty. Fragments added after that
overwrite live entries, and sg.size no longer matches what is reachable
between sg.start and sg.end, so pushing the record runs the scatterwalk off
the end of the scatterlist.
An unprivileged user can trigger this on a loopback TCP socket with the
"tls" ULP attached:
Zhiling Zou [Mon, 3 Aug 2026 12:15:32 +0000 (20:15 +0800)]
xdp: reject clones that overrun skb_shared_info tailroom
xdpf_clone() clones broadcast copies into a single page and sets
frame_sz to PAGE_SIZE. __xdp_build_skb_from_frame() later treats that
page like a normal XDP frame and expects the usual skb_shared_info
tailroom at the end of the buffer.
The current check only rejects frames whose linear xdp_frame header,
headroom, and packet data exceed PAGE_SIZE. A source frame backed by a
larger allocation can still satisfy that check while extending into the
clone's required shared-info area. When such a clone is converted back
into an skb, build_skb_around() places skb_shared_info over live packet
bytes and later writes can corrupt XDP return metadata.
Reject clones unless their linear area fits inside
SKB_WITH_OVERHEAD(PAGE_SIZE), matching the tailroom requirement already
enforced by the XDP-to-skb conversion path.
Jakub Kicinski [Thu, 6 Aug 2026 15:46:25 +0000 (08:46 -0700)]
Merge branch 'mptcp-misc-fixes-for-v7-2-rc6'
Matthieu Baerts says:
====================
mptcp: misc fixes for v7.2-rc6
Here are various unrelated fixes:
- Patches 1-3: harden incoming MPTCP suboptions parsing by rejecting
non-combinable ones. Patch 3 removes unreachable code after patch 2
added here for consistency, and to reduce comments from AI reviews.
Fixes for v5.6.
- Patch 4: fix a data race in the ADD_ADDR timer callback. A fix for
v5.13.
- Patch 5: correctly catch data corruption during the MPTCP join
selftest by marking tests as failed, instead of only printing a
warning. A fix for v5.18.
- Patch 6: fix a leak with the userspace ADD_ADDR list in case of race
condition during teardown. A fix for v5.19.
- Patch 7: deal with MPTFO with a valid token, but no data in the SYN. A
fix for v6.2.
- Patch 8: reclaim forward-allocated memory in case of error on the
receive side. A fix for v6.19.
====================
Paolo Abeni [Mon, 3 Aug 2026 16:16:40 +0000 (18:16 +0200)]
mptcp: reclaim forward-allocated memory on RX path errors
After commit 9db5b3cec4ec ("mptcp: borrow forward memory from subflow"),
errors in the receive path prior to queueing skbs into the receive
queue do not trigger forward-allocated memory reclaiming.
Prevent forward memory from growing unboundedly in pathological drop
scenarios by explicitly reclaiming memory when skbs are dropped.
Fixes: 9db5b3cec4ec ("mptcp: borrow forward memory from subflow") Cc: stable@vger.kernel.org Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org> Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org> Link: https://patch.msgid.link/20260803-net-mptcp-misc-fixes-7-2-rc6-v2-8-b8f496d71664@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Wyatt Feng [Mon, 3 Aug 2026 16:16:39 +0000 (18:16 +0200)]
mptcp: fastopen: only mark MPTFO subflows with SYN data
Passive TCP Fast Open accepts a valid-cookie SYN even when it carries
no data. In that case the child socket's receive queue is intentionally
left empty.
mptcp_fastopen_subflow_synack_set_params() set is_mptfo before checking
for queued SYN data. That made data-less TFO SYNs hit a WARN and, if
the warning was non-fatal, left stale MPTFO state behind. The stale
flag could later trigger a state-confusion bug in
check_fully_established().
Only mark the subflow as MPTFO after confirming that an SKB was queued.
Return quietly when the receive queue is empty.
Note that mptcp_subflow_context's is_mptfo field is now not just about
subflows where the TFO was present, but about MPTFO subflow that
consumed SYN data. Only having a valid cookie but not carrying data is
not really "doing TFO".
Shardul Bankar [Mon, 3 Aug 2026 16:16:38 +0000 (18:16 +0200)]
mptcp: pm: fix memory leak from alloc-during-teardown race
mptcp_pm_destroy() empties msk->pm.anno_list and
msk->pm.userspace_pm_local_addr_list under msk->pm.lock during socket
teardown, dropping the lock between the two.
A concurrent userspace PM genl ANNOUNCE on the same msk holds a sock
reference via mptcp_token_get_sock() and, in
mptcp_pm_nl_announce_doit(), calls
mptcp_userspace_pm_append_new_local_addr() and
mptcp_pm_announced_alloc(). Both take msk->pm.lock briefly to add to
their respective lists. Because the genl handler holds a sock reference,
mptcp_pm_destroy() may run on the same msk via mptcp_disconnect(), which
invokes mptcp_destroy_common() without dropping the sock refcount,
before the handler completes.
If the lock acquisitions interleave such that mptcp_pm_destroy() empties
a list first, the later alloc adds its entry to a list head that nothing
else iterates for this msk, and the entry leaks. kmemleak reports both
mptcp_pm_add_addr objects (from mptcp_pm_announced_alloc()) and
mptcp_pm_addr_entry objects (from
mptcp_userspace_pm_append_new_local_addr()) under sustained concurrent
ANNOUNCE + close load against the userspace PM.
Add an MPTCP_PM_DESTROYING bit in msk->pm.status, set by
mptcp_pm_destroy() under pm.lock before the lists are emptied and
checked under pm.lock by the alloc paths. Either the alloc takes pm.lock
first, in which case its entry is on the list when mptcp_pm_destroy()
frees it; or mptcp_pm_destroy() takes pm.lock first, in which case the
later alloc observes the bit and refuses.
Found by an MPTCP protocol-flow harness extending BRF (arXiv:2305.08782).
Gang Yan [Mon, 3 Aug 2026 16:16:37 +0000 (18:16 +0200)]
selftests: mptcp: join: mark tests with data corruption as failed
check_transfer() compares the input and output files byte-by-byte using
`cmp -l "$in" "$out" | while read ...`. Because the while-loop body runs
in a subshell (the script sets neither lastpipe nor pipefail), the
fail_test call inside it -- which sets the global ret/last_test_failed --
and the `return 1` both act on the subshell, not on check_transfer().
check_transfer() thus always falls through to `return 0`, and any data
corruption affecting only the payload (leaving the subflow/PM counters
untouched) is silently reported as PASS.
Fixes: 8117dac3e7c3 ("selftests: mptcp: add invert check in check_transfer") Cc: stable@vger.kernel.org Signed-off-by: Gang Yan <yangang@kylinos.cn> Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org> Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org> Link: https://patch.msgid.link/20260803-net-mptcp-misc-fixes-7-2-rc6-v2-5-b8f496d71664@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Qing Luo [Mon, 3 Aug 2026 16:16:36 +0000 (18:16 +0200)]
mptcp: pm: fix data race in add_addr timer callback
The timer callback reads entry->retrans_times outside pm.lock to decide
whether to call mptcp_pm_subflow_established(). Since
mptcp_pm_announced_del_timer() can concurrently set retrans_times =
ADD_ADDR_RETRANS_MAX under pm.lock, a race condition exists.
I discovered this issue while studying the code. AI tools helped me to
verify the issue can potentially happen under race conditions.
Use a local 'retransmit' flag set inside pm.lock to capture whether
retransmission is still possible when the lock is taken. This allows to
call mptcp_pm_subflow_established() accordingly, and not depending on
the situation that can be different when checked outside the pm.lock.
Fixes: 348d5c1dec60 ("mptcp: move to next addr when timeout") Cc: stable@vger.kernel.org Signed-off-by: Qing Luo <luoqing@kylinos.cn> Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org> Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org> Link: https://patch.msgid.link/20260803-net-mptcp-misc-fixes-7-2-rc6-v2-4-b8f496d71664@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
After the parent commit ("mptcp: avoid combining some incoming
suboptions"), the parsing step no longer allow to have both the
MP_CAPABLE and MP_JOIN suboptions set together.
These chunks are now unreachable, these checks can then be removed.
Some MPTCP suboptions are mutually exclusive according to the RFC8684,
but also because in different places, the code doesn't expect some
combinations to be present. That's specially true for suboptions that
would be present twice, but with different attributes.
The new restrictions are the same as the ones applied on the output
side, with mptcp_write_options. The same rules can be reused with a
small fix: an MP_FASTCLOSE can be used with a DSS when the sender picks
this option [1], which is not the case on Linux. Here are the rules:
Which options can be used together?
X: mutually exclusive
O: often used together
C: can be used together in some cases
P: could be used together but we prefer not to (optimisations)
| Opt: | MPC | MPJ | DSS | ADD | RM | PRIO | FAIL | FC |
|------|------|------|------|------|------|------|------|------|
| MPC |------|------|------|------|------|------|------|------|
| MPJ | X |------|------|------|------|------|------|------|
| DSS | X | X |------|------|------|------|------|------|
| ADD | X | X | P |------|------|------|------|------|
| RM | C | C | C | P |------|------|------|------|
| PRIO | X | C | C | C | C |------|------|------|
| FAIL | X | X | C | X | X | X |------|------|
| FC | X | X | P | X | X | X | X |------|
| RST | X | X | X | X | X | X | O | O |
|------|------|------|------|------|------|------|------|------|
The only difference is with the 'P': another stack could send and
ADD_ADDR with other suboptions (DSS, RM_ADDR), and this should be
allowed.
A few points of attention:
- In theory, an MP_CAPABLE could be used with a RM_ADDR, but there is
no reason to add it with a SYN. Note that even with a 4th ACK, it
doesn't seem to be useful, except when IDs are known in advance via
another channel. Better not to break that.
- Now, combining both an MP_CAPABLE and an MP_JOIN will no longer
result to a reject of the two options, but only the second suboption
is ignored. That seems OK to do that for this unexpected error. At
least now all inconsistent combinations are handled the same way.
This could change later in next. This also means the explicit checks
for having both MPC + MPJ in subflow.c will now be unreachable.
That's fine, they will be removed in a follow-up patch.
- In case of conflicting combinations, the extra suboption(s) is/are
ignored: having such combinations either means the remote peer is
buggy, or is evil. The simplest action is then taken in this case:
stop processing the current suboption.
- In mp_opt->suboptions, there is also a bit reserved to the checksum,
which can be used in an MP_CAPABLE and a DSS. Each time a DSS option
can be used in parallel with another option, the checksum can be set,
so the verification is combined into a new OPTIONS_MPTCP_DSS macro.
- An MP_CAPABLE ACK can carry a Data-Level Length, and an optional
Checksum: they are the same as the ones found in a DSS, because a DSS
cannot be used in parallel to an MP_CAPABLE. Similarly, even if there
is room, a DSS cannot be used with an MP_JOIN.
mptcp: options: reset DSS fields in case of unexpected size
A remote peer could send a malformed DSS with a wrong size, followed by
another DSS or MPC + Data. In this case, the first suboption will be
ignored, but leaving some fields written, which could lead to
inconsistency or access uninitialized data.
Explicitly reset the fields that could have been modified in case of
unexpected size.
Linus Torvalds [Thu, 6 Aug 2026 15:29:59 +0000 (08:29 -0700)]
Merge tag 'probes-fixes-v7.2-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace
Pull probes selftest fix from Masami Hiramatsu:
- selftests/ftrace: Refactor eprobes test to fix argument checks
Refactor the eprobes selftest to get more stable test result by using
`sys_enter_chdir` instead of `openat` and filter for exact directory
names
This resolves test instability after the string pointer handling fix
* tag 'probes-fixes-v7.2-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
selftests/ftrace: refactor eprobes test to fix argument checks
Nothing ties that to the interface being up, so the work can be armed
again after ipheth_close() has already drained it, and stay armed
until the netdev whose private area embeds it is freed.
On unplug with a TX URB in flight, ipheth_disconnect() drains the work
through unregister_netdev() -> ipheth_close() ->
cancel_delayed_work_sync() and only then calls ipheth_kill_urbs().
usb_kill_urb() completes the in-flight TX URB with -ENOENT, so
ipheth_sndbulk_callback() runs after the drain and re-arms
carrier_work.
The same completion also re-arms the work if the interface is only
brought down while a TX URB is in flight, and
ipheth_carrier_check_work() then keeps re-queueing itself once a
second. unregister_netdev() does not call ipheth_close() for an
already-down interface, so nothing drains it on the later unplug
either.
In both cases free_netdev() frees the netdev while carrier_work is
still pending, and ipheth_carrier_check_work() dereferences freed
memory.
Tie the work to the interface state instead of chasing the completion:
disable it in ipheth_close() and enable it in ipheth_open(), so a
schedule_delayed_work() from the URB completion is a no-op whenever
the interface is not up. disable_delayed_work_sync() also waits for a
running instance, so it fully replaces the cancel_delayed_work_sync()
it takes the place of. The work starts out disabled in ipheth_probe()
so the enable/disable counts balance from the first open.
Reproduced under KASAN on linux-next (next-20260731) with dummy_hcd and
raw-gadget standing in for the device, driving the second path above (the
interface is already down, so unregister_netdev() does not call
ipheth_close()): 15 of 15 unpatched boots report a slab-use-after-free in
__run_timers(), freed by ipheth_disconnect() and re-armed from
ipheth_sndbulk_callback() via queue_delayed_work_on(). The
same trigger on a kernel differing only by this patch reports 0 of 15,
and the carrier check still functions across open/close cycles.
The reproducer needs an attached USB device that stops draining bulk OUT,
plus a link down and unplug, driven as root. It is not a privilege
boundary crossing and no exploit primitive was developed.
Found by 0sec (https://0sec.ai).
Fixes: bb1b40c7cb86 ("usbnet: ipheth: prevent TX queue timeouts when device not ready") Cc: stable@vger.kernel.org Signed-off-by: Doruk Tan Ozturk <doruk@0sec.ai> Link: https://patch.msgid.link/20260802120602.42595-1-doruk@0sec.ai Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Fan XinRan [Mon, 3 Aug 2026 14:38:50 +0000 (14:38 +0000)]
net: thunderbolt: Tear down DMA paths before stopping the rings
tbnet_tear_down() stops both rings and frees their frame buffers before
calling tb_xdomain_disable_paths(). tb_ring_stop() zeroes the ring's
descriptor base and tbnet_free_buffers() unmaps and frees the pages the
frames sit in, so by the time __tb_path_deactivate_hop() polls the hop's
'pending' bit, anything still in flight has nowhere to drain to.
The teardown sequence has been in this order since the driver was added.
The setup path has not: commit ff7cd07f3064 ("net: thunderbolt: Enable
DMA paths only after rings are enabled") moved the path enable to the end
of tbnet_connected_work() and documented why:
/* Both logins successful so enable the rings, high-speed DMA
* paths and start the network device queue.
*
* Note we enable the DMA paths last to make sure we have primed
* the Rx ring before any incoming packets are allowed to
* arrive.
*/
Teardown was never updated to match, so the rings and the paths now come
down in the same order they go up instead of in reverse.
On an ASMedia ASM4242 host router the 'pending' bit then never clears:
every teardown burns the full 500 ms timeout and
__tb_path_deactivate_hop() returns -ETIMEDOUT. Raising the timeout to
5 s does not help, so the hop is not slow to drain, it never drains
at all.
The failure is invisible above the thunderbolt core.
__tb_path_deactivate_hops() is void and only calls tb_port_warn();
tb_path_deactivate(), tb_tunnel_deactivate() and
__tb_disconnect_xdomain_paths() are void as well, and
tb_disconnect_xdomain_paths() ends in an unconditional "return 0". So
tb_xdomain_disable_paths() reports success and the netdev_warn() below
it never fires. Repeated teardowns eventually take the XDomain control
channel down, after which the peer node is gone and only a power cycle
brings the controller back.
Deactivating the paths first fixes it. Measured with kretprobes on a
stock v6.17 tree with no other patches applied, on a link that was up
and had just carried traffic:
before: __tb_path_deactivate_hop() returns 0 for the first hop, then
-ETIMEDOUT for the second 500335 us later
after: 0 for both, 525 us apart
Alternating the two orderings ABBA over three load levels, four
teardowns per arm: every teardown failed before the change (21 of 21
that ran), none failed after (0 of 24). The before arms ran short
because the link died partway through. The same split shows up when
the interface is enslaved to a bond instead of just brought down, which
is how I ran into this in the first place. Throughput and latency after
the change are unchanged.
Hosts whose routers drain the hop despite the stale descriptor base see
no functional difference, since the paths end up deactivated either way.
Fixes: e69b6c02b4c3 ("net: Add support for networking over Thunderbolt cable") Signed-off-by: Fan XinRan <shinjiangjiang@gmail.com> Acked-by: Mika Westerberg <mika.westerberg@linux.intel.com> Link: https://patch.msgid.link/20260803-b4-tbnet-teardown-v2-1-27de6a13ca2d@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Linus Torvalds [Thu, 6 Aug 2026 15:16:40 +0000 (08:16 -0700)]
Merge tag 'xfs-fixes-7.2-rc7' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux
Pull xfs fixes from Carlos Maiolino:
"This contains mostly a collection of bug fixes found by LLM tools"
* tag 'xfs-fixes-7.2-rc7' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux: (34 commits)
xfs: check v5 superblock features early
xfs: add a comment to describe xfs_gc_bio.victim_rtg
xfs: add a separate bio_set for spliting GC writes
xfs: don't swallow dquot recovery verification errors
xfs: fix ilock leak on error in xfs_dq_get_next_id
xfs: don't ignore runtime errors in xrep_iunlink_reload_next
xfs: set the prev pointer when reinserting an inode on the unlinked list
xfs: fix another iunlink infinite loop bug in online fsck
xfs: fix allocated inodes that show up in the unlinked list
xfs: check xfarray iteration errors when committing unlinked inode lists
xfs: pass runtime errors from xrep_iunlink_mark_ondisk_rec up to callers
xfs: load next_agino from the correct xfarray in xrep_iunlink_relink_prev
xfs: don't walk off the end of a null sc->sa.agi_bp in AGI repair
xfs: don't livelock in scrub on a circular unlinked list
xfs: hoist per-bucket unlinked list check to helper
xfs: avoid UAF on sc->tempip in xrep_tempfile_create
xfs: nlink scrub must take IOLOCK before determining ILOCK state
xfs: don't zap the attr fork on repair when there are queued pptr updates
xfs: don't return EFSCORRUPTED when scrubbing corrupt parent pointers
xfs: don't double-lock when deleting a self-referential directory
...
The qdisc create-and-graft path allows users to keep attaching new classful
qdiscs under an already deep parent hierarchy. Such a hierarchy can later
be walked recursively and exhaust the kernel stack.
This series stores the qdisc hierarchy depth in struct Qdisc and checks it
when a qdisc is grafted. New child qdiscs are rejected once the parent is
already at the maximum allowed depth. It also adds tdc coverage for the
maximum allowed depth and rejection above it.
====================
Zijie Huang [Sat, 1 Aug 2026 13:42:33 +0000 (21:42 +0800)]
net/sched: reject overly deep qdisc hierarchies
Deep qdisc hierarchies can lead to excessive recursion in qdisc tree
walkers and exhaust the kernel stack. The existing loop check does not
cover the create-and-graft path, so a hierarchy can still be extended by
creating a new child qdisc below an already deep parent.
Store the hierarchy depth in struct Qdisc and update it when qdiscs are
grafted. Reject new child qdiscs once the parent is already at the maximum
allowed depth.
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Cc: stable@vger.kernel.org Suggested-by: Jamal Hadi Salim <jhs@mojatatu.com> Reported-by: Vega <vega@nebusec.ai> Assisted-by: Codex:gpt-5.4 Signed-off-by: Zijie Huang <milkory@outlook.com> Signed-off-by: Ren Wei <enjou1224z@gmail.com> Reviewed-by: Victor Nogueira <victor@mojatatu.com> Link: https://patch.msgid.link/1e9ab39597423fd5d13cfaaf52279b8ee3d9fc3c.1785434373.git.milkory@outlook.com Acked-by: Jamal Hadi Salim <jhs@mojatatu.com> Signed-off-by: Paolo Abeni <pabeni@redhat.com>
In function otx2_get_egress_burst_cfg, when the parameter `burst` is
255 and the max mantissa is 255 (0xFFULL), `burst_exp` is set to
`ilog2(255) - 1`, which equals 6.
This results in an unsigned wrap-around when calculating
`(1ULL << (*burst_exp - 7))`, since `*burst_exp - 7` becomes -1,
which makes the shift operand 0xFFFFFFFF. This value is greater than
the width of the left operand.
According to standard 6.5.7 p.3:
"The type of the result is that of the promoted left operand.
If the value of the right operand is negative or is greater than
or equal to the width of the promoted left operand, the behavior
is undefined."
Fix the off-by-one boundary condition.
Add a WARN_ON(*burst_exp < 7) before the else branch as an
explicit safeguard. This ensures that if max_mantissa ever changes
in a way that reintroduces this condition, it will be immediately
caught at runtime rather than silently triggering UB.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
Ahmed Naseef [Tue, 4 Aug 2026 11:35:11 +0000 (15:35 +0400)]
net: phy: mediatek: fix TX blink masks using the RX bits
MTK_GPHY_LED_TX_BLINK_SET and MTK_2P5GPHY_LED_TX_BLINK_SET are built
from the RX blink bits instead of the TX ones, so both TX masks are
identical to their RX counterparts. The TX bits they should be using,
MTK_PHY_LED_BLINK_{10,100,1000,2500}TX, are otherwise only referenced
by the per-speed branch of mtk_phy_led_hw_ctrl_set().
A TX trigger selected without a link trigger therefore programs the RX
blink bits, and the LED blinks on received traffic. The masks are also
used to decode the blink register in mtk_phy_led_hw_ctrl_get(), which
as a result cannot tell the two triggers apart: an RX-only
configuration reads back as RX and TX, and a TX-only configuration
reads back as neither.
Fixes: 7f9c320c98db ("net: phy: mediatek: Move LED helper functions into mtk phy lib") Cc: stable@vger.kernel.org Signed-off-by: Ahmed Naseef <naseefkm@gmail.com> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Link: https://patch.msgid.link/20260804113511.3371248-1-naseefkm@gmail.com Signed-off-by: Paolo Abeni <pabeni@redhat.com>
net/smc: fix TOCTOU race between smc_listen_out() and listener close
smc_listen_out() reads lsmc->sk.sk_state without the listener lock,
then acquires lock_sock_nested() only after the check passes. This
opens a window where smc_close_active() can transition the listener
to SMC_CLOSED, call smc_close_cleanup_listen() to drain the accept
queue, and release the lock, all between the lockless read and the
delayed lock acquisition:
smc_listen_work (smc_hs_wq) smc_close_active()
------------------------------- -------------------------
release_sock(child)
if (sk_state == SMC_LISTEN) TRUE
lock_sock(listener)
sk_state = SMC_CLOSED
smc_close_cleanup_listen()
release_sock(listener)
flush_work(tcp_listen_work)
lock_sock_nested(listener)
smc_accept_enqueue(listener, child) /* child enqueued on dead listener */
smc_close_active() flushes only tcp_listen_work. Work items already
dispatched onto smc_hs_wq for the CLC handshake continue running
unguarded. smc_accept_enqueue() takes a sock_hold() on the child that
is never released, so the child smc_sock, its clcsock, and the
reference all leak. A remote peer that opens TCP connections while the
server calls close() can exhaust kernel memory.
Move lock_sock_nested() to before the sk_state check so that the test
and the enqueue are atomic under the listener lock.
Fixes: fd57770dd198 ("net/smc: wait for pending work before clcsock release_sock") Reviewed-by: Mahanta Jambigi <mjambigi@linux.ibm.com> Signed-off-by: Sidraya Jayagond <sidraya@linux.ibm.com> Reviewed-by: Breno Leitao <leitao@debian.org> Reviewed-by: Dust Li <dust.li@linux.alibaba.com> Link: https://patch.msgid.link/20260803070701.126339-1-sidraya@linux.ibm.com Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Eric Dumazet [Tue, 4 Aug 2026 15:20:48 +0000 (15:20 +0000)]
net: remove WARN_ON_ONCE() from sk_mc_loop()
sk_mc_loop() can be called for sockets that are neither AF_INET
nor AF_INET6 (e.g. AF_PACKET sockets when sending packets via raw/packet
socket over virtual devices such as VRF or ipvlan).
In such cases, sk_family is not AF_INET/AF_INET6 and sk_mc_loop() falls
through the switch statement and triggers WARN_ON_ONCE(1).
Non-INET sockets do not support IP_MULTICAST_LOOP or IPV6_MULTICAST_LOOP
options, so loopback should default to true without generating a warning.
Fixes: f60e5990d9c1 ("ipv6: protect skb->sk accesses from recursive dereference inside the stack") Reported-by: syzbot+22c3218a6fa219e47321@syzkaller.appspotmail.com Closes: https://lore.kernel.org/netdev/6a72024c.13623e66.bdc14.0019.GAE@google.com/T/#u Signed-off-by: Eric Dumazet <edumazet@google.com> Link: https://patch.msgid.link/20260804152048.2134341-1-edumazet@google.com Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Jijie Shao [Tue, 4 Aug 2026 13:05:54 +0000 (21:05 +0800)]
MAINTAINERS: add myself as a maintainer for Hisilicon Network Subsystem
I am already listed as a maintainer for the HNS3 and HIBMCGE drivers,
but not for the broader Hisilicon Network Subsystem entry, whose file
pattern covers drivers/net/ethernet/hisilicon/ (e.g. the legacy hns
driver). As a result, patches to those files are not CC'd to me.
Add myself alongside Jian Shen to help maintain these legacy Hisilicon
ethernet drivers and ensure patches in this tree are routed to me.
dibs->lock is initialised by dibs_dev_add(), but a dibs device can
already take interrupts before that call: ism_probe() runs
ism_dev_init(), and hence request_irq(), before it calls
dibs_dev_add(). No client can have registered a dmb at that point, so
no dmb interrupt can occur, but a GID event interrupt can, and
ism_handle_irq() takes dibs->lock unconditionally on entry, before it
inspects anything else.
Initialise the lock in dibs_dev_alloc() instead, so that it is valid as
soon as a driver can publish the device to its interrupt handler.
Fixes: cc21191b584c ("dibs: Move data path to dibs layer") Cc: stable@vger.kernel.org Reviewed-by: Alexandra Winter <wintera@linux.ibm.com> Signed-off-by: Hidayath Khan <hidayath@linux.ibm.com> Link: https://patch.msgid.link/20260730124227.167829-1-hidayath@linux.ibm.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Dongli Zhang [Sun, 2 Aug 2026 22:46:12 +0000 (15:46 -0700)]
net: tap: set skb->dev before parsing virtio net header in tap_get_user_xdp()
The commit 4f61f133f354 ("net: tap: NULL pointer derefence in
dev_parse_header_protocol when skb->dev is null") fixed a crash in
tap_get_user() by assigning skb->dev before calling tun_vnet_hdr_to_skb().
This is required because virtio_net_hdr_to_skb() may invoke
dev_parse_header_protocol(), which dereferences skb->dev. Without the
assignment, a NULL pointer dereference can occur.
However, tap_get_user_xdp() still parses the virtio-net header before
assigning skb->dev. When the vhost TX path passes an XDP buffer containing
a GSO virtio-net header but the protocol is set to zero on purpose,
tun_vnet_hdr_to_skb() can reach dev_parse_header_protocol() while skb->dev
is still NULL, resulting in a crash.
Fix this by looking up the tap device and assigning skb->dev before calling
tun_vnet_hdr_to_skb(), matching the ordering already used in
tap_get_user(). Preserve the existing RCU read-side critical section across
dev_queue_xmit().
Fixes: 924a9bc362a5 ("net: check if protocol extracted by virtio_net_hdr_set_proto is correct") Cc: stable@vger.kernel.org Signed-off-by: Dongli Zhang <dongli.zhang@oracle.com> Reviewed-by: Willem de Bruijn <willemb@google.com> Acked-by: Michael S. Tsirkin <mst@redhat.com> Link: https://patch.msgid.link/20260802224612.264563-1-dongli.zhang@oracle.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Zhiling Zou [Mon, 3 Aug 2026 06:12:33 +0000 (14:12 +0800)]
ip6_tunnel: clear skb2->cb[] in ip6ip6_err()
ip6ip6_err() clones an outer IPv6 ICMP error skb, pulls it to the
quoted inner IPv6 packet, and then passes the clone to icmpv6_send().
The clone still carries the outer packet's inet6_skb_parm in skb->cb.
If the outer packet had a Home Address Option, IP6CB(skb2)->dsthao
remains non-zero after skb_pull(). icmpv6_send() later calls
mip6_addr_swap(), which uses that stale dsthao offset against the quoted
inner packet. A malformed inner destination-options header can then make
the HAO lookup and address swap run past the end of the quoted packet
and corrupt skb_shared_info.
Clear skb2->cb[] before pulling the quoted inner IPv6 packet so the
reply path does not reuse metadata left by the outer IPv6 stack.
Henry Martin [Mon, 3 Aug 2026 04:36:18 +0000 (12:36 +0800)]
net/ncsi: fix heap OOB read in NCSI_CMD_SEND_CMD payload length
ncsi_send_cmd_nl() takes the number of bytes to copy from the
attacker-controlled ncsi_pkt_hdr.length field of the in-band packet
header, while the source buffer is the NCSI_ATTR_DATA netlink
attribute whose readable size is nla_len() - sizeof(ncsi_pkt_hdr).
The two length sources are never cross-checked: only
nla_len() >= sizeof(struct ncsi_pkt_hdr) is enforced.
With hdr->length set larger than the attribute payload (up to 65535
against at most 2032 readable bytes), ncsi_cmd_handler_oem() copies
past the end of the netlink attribute buffer with unsafe_memcpy(),
leaking up to ~64KB of kernel heap memory into the transmitted NCSI
command packet. The destination skb is sized by the declared payload,
so the write side does not overflow - this is a pure OOB read /
information leak, reachable with CAP_NET_ADMIN on systems with a
registered NCSI device (e.g. OpenBMC on Aspeed BMC SoCs, where
NET_NCSI=y is standard).
Reject commands whose declared payload extends past the end of the
data attribute.
The issue was found by the autokbug dynamic kernel fuzzer at Tencent
Yunding Lab.
Fixes: 9771b8ccdfa6 ("net/ncsi: Extend NC-SI Netlink interface to allow user space to send NC-SI command") Reported-by: Henry Martin <bsdhenrymartin@gmail.com> Signed-off-by: Henry Martin <bsdhenrymartin@gmail.com> Link: https://patch.msgid.link/20260803043618.3210301-1-bsdhenrymartin@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Zihan Xi [Sun, 2 Aug 2026 09:23:34 +0000 (09:23 +0000)]
mac802154: fix netdev use-after-free in beacon worker
mac802154_beacon_worker() reads local->beacon_req under RCU and derives
the sub-interface from the request, but then drops the RCU read lock and
continues to use both sdata and the embedded wpan_dev.
mac802154_stop_beacons_locked() cancels only pending beacon work, clears
local->beacon_req and frees the request. A beacon worker that is already
running can therefore continue after interface teardown and dereference
the freed netdev private area.
The scan worker already pins the netdev before leaving RCU. Apply the
same lifetime rule to the beacon worker: take a netdev reference while
the request is still protected by RCU, and release it on all paths that
continue after the reference is acquired.
Eric Dumazet [Tue, 4 Aug 2026 09:33:28 +0000 (09:33 +0000)]
netfilter: nf_flow_table: drop existing skb dst before skb_dst_set_noref()
Incoming skbs passing through netfilter flowtable offload hooks (or XFRM
offload path) might already carry a ref-counted dst_entry assigned during
earlier RX or routing steps.
Calling skb_dst_set_noref() when skb already holds a ref-counted dst
overwrites skb->_skb_refdst, leaking the previous dst_entry reference
count and triggering a DEBUG_NET_WARN_ON_ONCE assertion in
skb_dst_check_unset():
WARNING: at skb_dst_check_unset include/linux/skbuff.h:1170
WARNING: at skb_dst_set_noref include/linux/skbuff.h:1234
WARNING: at nf_flow_offload_ip_hook+0xf6c/0x2b60 net/netfilter/nf_flow_table_ip.c:864
Drop any existing dst_entry reference with skb_dst_drop(skb) before
setting the non-referenced flowtable destination.
Fixes: 2a79fd3908ac ("netfilter: nf_flow_table: attach dst to skbs") Reported-by: syzbot+76d4e3a055aec3b007ec@syzkaller.appspotmail.com Closes: https://lore.kernel.org/netdev/6a71b141.9511d2ce.1fc5b9.033b.GAE@google.com/T/#u Signed-off-by: Eric Dumazet <edumazet@google.com> Reviewed-by: Pablo Neira Ayuso <pablo@netfilter.org> Link: https://patch.msgid.link/20260804093328.1831847-1-edumazet@google.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jiayuan Chen [Mon, 3 Aug 2026 06:17:38 +0000 (14:17 +0800)]
tcp: fix TFO max_qlen accounting across reuseport migration
A listener's TCP_FASTOPEN max_qlen stops being accurate and lets through
far more pending Fast Open requests than it was configured for.
This only shows up with SO_REUSEPORT listener migration, where closing a
listener hands its still-pending TFO children over to a surviving one.
fastopenq.qlen is charged in tcp_fastopen_create_child() when the child
is created and uncharged in reqsk_fastopen_remove() when the handshake
completes. The uncharge follows rsk_listener of the request the child
points at, and inet_reqsk_clone() has repointed the child at a new
request owned by the new listener, so the ++ and the -- land on two
different sockets. The new listener's qlen drifts negative and its
limit no longer binds.
Charge the new listener during migration, like reqsk_queue_migrated()
already does for queue->young and queue->qlen.
Fixes: 54b92e841937 ("tcp: Migrate TCP_ESTABLISHED/TCP_SYN_RECV sockets in accept queues.") Signed-off-by: Jiayuan Chen <jiayuan.chen@linux.dev> Reviewed-by: Kuniyuki Iwashima <kuniyu@google.com> Reviewed-by: Eric Dumazet <edumazet@google.com> Link: https://patch.msgid.link/20260803061739.134737-1-jiayuan.chen@linux.dev Signed-off-by: Jakub Kicinski <kuba@kernel.org>
====================
The following patchset contains Netfilter/IPVS fixes net, this
includes fixes for ebtables nflog target, ipset hash type,
IPVS kthread estimator
1) Prevent IPVS kthread estimator from draining the est_temp_list
when netns is being dismantled. From Zhiling Zou.
2) Missing module nflog refcount bump from ebtables nflog target from
.checkentry path. Similar dependency exists already in xt_NFLOG and
nft_log. From Chengfeng Ye.
3) Use RCU to fix ipset bookkeeping of cidr values on weakly-ordered
architectures. From Jozsef Kadlecsik.
4) Use atomic64_t for set->ext_size in ipset to fix parallel inserts
and deletes racing on updating it. From Jozsef Kadlecsik.
5) Add small wrappers for hash and bucket size to prepare the update
of ipset hash set types to rhashtable, from Florian Westphal.
6) Add mtype_del_cidr_all() and use it to prepare the migration of
ipset hash types to rhashtable. From Florian Westphal.
7) Replace existing ipset call_rcu() based destruction with rcu_work
api also to ease the transition to rhashtable. Also from Florian.
8) Avoid reading the IPv4 ihl field multiple times to prevent local
attacker to cause out-of-bounds write in ip_vs_nat_icmp(), from
Julian Anastasov.
9) Restore the checksum validations that could be needed by the IPVS
FORWARD hook. Also from Julian.
====================
Linus Torvalds [Wed, 5 Aug 2026 20:40:11 +0000 (13:40 -0700)]
Merge tag 'selinux-pr-20260805' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/selinux
Pull selinux fixes from Paul Moore:
- Continue to improve the validation of SELinux policies during load
- Fix a SELinux regression caused by bpffs changes in v7.2-rc1
- Fix a SELinux preformance regression caused by SELinux changes in
v7.2-rc1
* tag 'selinux-pr-20260805' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/selinux:
selinux: check level category sets once at load time
selinux: require every boolean value to be defined
selinux: reject an unclaimed class value in security_get_classes()
selinux: require a class's permission values to cover its permission count
selinux: do not cancel a policy conversion that never started
selinux: bpf: check SBLABEL_MNT before isec init
selinux: reject a class permission count below its inherited common
selinux: reject a permission value exceeding the class permission count
Linus Torvalds [Wed, 5 Aug 2026 15:18:03 +0000 (08:18 -0700)]
Merge tag 'soc-fixes-7.2-2' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc
Pull SoC fixes from Arnd Bergmann:
"The majority of the fixes this time is for Qualcomm devicetree files,
addressing various incorrect settings in chip specific dtsi files that
prevent some feature from working correctly.
Another three such issues are addressed on the Broadcom bcm5301x and
bcm2712 SoC platforms.
Two minor issues are addressed in nuvoton and aspeed specific SoC
drivers, and the MAINTAINERS file is updated to add Billy Tsai and
Ryan Chen as aspeed reviewers as well as clarify the NXP/Freescale
entries"
Andy Lutomirski [Tue, 4 Aug 2026 00:37:08 +0000 (17:37 -0700)]
mm: fix incorrect flush address in direct page table reclaim
When zap_pte_range reclaims a page table, it does:
pte_free_tlb(tlb, pmd_pgtable(pmdval), addr);
and this is unconditionally wrong: if this code executes, addr *always*
points one past the end of the range covered by the table. The addr
parameter is used to flush the TLB (really the paging-structure-cache)
to drop references to the to-be-freed table, and any architecture that
cares about the parameter will flush the wrong address. (But they'll
still free the correct page).
I think it's worth contemplating why the kernel works at all.
If we hit the offending line of code, we will first clear the PMD entry
(line 1954, zap_empty_pte_table), then we will issue pending flushes if
force_flush is set (tlb_flush_mmu_tlbonly(tlb)), then we will skip the
retry on line 1979 (phew!), and then we will do the offending
pte_free_tlb call. *Or* we will clear the PMD entry immediately before
pte_free_tlb (line 1983, zap_pte_table_if_empty).
If we have any pending flushes (i.e. we actually zapped any last-level
entries) at the time we clear the PMD entry, then the flush really ought
to flush all references to the table (Linus certainly seems to think it
will on all architectures [0]).
The condition under which we have no accumulated flushes at the time of
the clear is very complex (the whole zap_pte_range function has absurdly
complex control flow). If we do hit the bad case, then we will end up
clearing the PMD entry after the last time the range is flushed, and any
CPU is free to cache a reference to the (empty) page table. If this
happens due to an ordinary read or write, it would segfault, so it would
be rare. But the cache could be speculatively filled as well. Then
we'll flush the wrong address and then free and possibly reuse the
table.
On x86, even flushing the wrong address works on non-KPTI Intel systems
because INVLPG flushes *all* paging-structure-caches, not just the ones
for the target address. But INVPCID does not, and flush_tlb_one_user
will use INVPCID if it's available. And then we're toast. AMD systems
are more susceptible: we set the EFER.TCE bit, which makes even INVLPG
only flush the target address.
I think this might fix an issue in ripgrep reported here:
https://github.com/BurntSushi/ripgrep/issues/3494
Qing Luo [Tue, 4 Aug 2026 02:55:14 +0000 (10:55 +0800)]
sctp: fix addip_serial increment on ASCONF_ACK allocation failure
In sctp_process_asconf(), when sctp_make_asconf_ack() fails to allocate
the ASCONF_ACK chunk due to memory pressure, the code jumps to the
done label where asoc->peer.addip_serial is unconditionally incremented.
This leaves the peer's ASCONF (serial N) unacknowledged while the local
endpoint now expects serial N+1. When the peer retransmits serial N, it
falls into the serial < addip_serial + 1 branch ,
which attempts to look up a cached ACK for serial N. No cached ACK
exists since the allocation failed, so the retransmission is silently
discarded. The peer eventually times out and ABORTs the association.
Move the addip_serial increment inside the if (asconf_ack) block so that
the serial number is only advanced when the ASCONF_ACK is successfully
created and cached. This way, on allocation failure, the serial number
is unchanged and the peer's retransmitted ASCONF will be correctly
re-processed.
Jakub Kicinski [Wed, 5 Aug 2026 02:54:35 +0000 (19:54 -0700)]
Merge branch 'bnxt_en-bug-fixes'
Michael Chan says:
====================
bnxt_en: Bug fixes
This series include 3 bug fixes:
1. queue start bug fix on the VNIC's default ring. 2 refactoring
patches preceed the actual bug fix.
2. Bug fix for TPA data corruption seen on some ARM systems.
3. PTP PPS setting bug fix.
====================
The existing driver logic is always turning on PTP_CLK_REQ_PPS
regardless of the "on" parameter passed to bnxt_ptp_enable().
During shutdown, PTP_CLK_REQ_PPS may be turned off and this
bug will do the opposite and may trigger a PCIe PTM request TLP.
On some systems this can trigger a PCIe AER.
Fix it by properly configuring PTP_CLK_REQ_PPS based on the "on"
parameter.
Fixes: 9e518f25802c ("bnxt_en: 1PPS functions to configure TSIO pins") Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com> Signed-off-by: Keegan Freyhof <keegan.freyhof@broadcom.com> Signed-off-by: Michael Chan <michael.chan@broadcom.com> Link: https://patch.msgid.link/20260731190937.807270-6-michael.chan@broadcom.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Michael Chan [Fri, 31 Jul 2026 19:09:36 +0000 (12:09 -0700)]
bnxt_en: Disable EOP for TPA on all chips to prevent data corruption
EOP (End of frame padding) on the AGG ring may cause overlapping of
zero padding at the end of one segment with the next segment's data.
If Relaxed Ordering (RO) is enabled, the zero padding may overwrite
valid data in the next segment and corrupt the data. Older chips
(P5 and older) do not automatically disable RO when EOP is enabled.
On some ARM systems, data corruption was reported on 57508 (P5)
chips with RO enabled.
Always disable EOP on all chips on the AGG rings when TPA is enabled
to fix the data corruption.
Fixes: bfcd8d791ec1 ("bnxt_en: Add fast path logic for TPA on 57500 chips.") Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Michael Chan <michael.chan@broadcom.com> Link: https://patch.msgid.link/20260731190937.807270-5-michael.chan@broadcom.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
bnxt_en: Refresh VNIC default ring on queue restart if needed
When a queue is restarted, refresh VNIC_CFG for all VNICs whose
default RX ring is the restarted ring. This will eliminate this
possible FW warning caused by a stale default ring in the VNIC:
FW reported unknown error type 10
Fixes: 5ac066b7b062 ("bnxt_en: Fix queue start to update vnic RSS table") Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com> Signed-off-by: Shravya KN <shravya.k-n@broadcom.com> Signed-off-by: Michael Chan <michael.chan@broadcom.com> Link: https://patch.msgid.link/20260731190937.807270-4-michael.chan@broadcom.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
bnxt_en: Determine and store default RX ring in vnic structure
Each VNIC has a default RX ring. The purpose of the default RX ring
is to provide a destination for any packets that cannot be parsed by
the RSS logic. Up until now, the default RX ring is always Ring 0.
We neglected to take care of this default RX ring when adding the
queue restart feature. If ring 0 (default ring) is re-started, it
may now have a new FW ring ID after freeing the old one and
allocating a new one. The VNIC now may have a stale default ring
and it may generate an internal exception. This exception may
appear in dmesg:
FW reported unknown error type 10
The best way to resolve this issue is to use a more appropriate
ring for the default ring instead of always ring 0. Ring 0 may not
even be in the RSS table, especially on a new RSS context.
This patch adds the logic to determine and store the proper default
RX ring for a VNIC. For an RSS VNIC, the default ring is the lowest
ring number in the RSS table. The next patch will add proper logic
to update the VNIC if the default ring changes after queue restart.
Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com> Signed-off-by: Shravya KN <shravya.k-n@broadcom.com> Signed-off-by: Michael Chan <michael.chan@broadcom.com> Link: https://patch.msgid.link/20260731190937.807270-3-michael.chan@broadcom.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
bnxt_en: Move RSS table fill outside __bnxt_hwrm_vnic_set_rss()
This is a refactor patch with no change in behavior. The caller
will now fill the RSS table before calling __bnxt_hwrm_vnic_set_rss().
In the next patch, we'll add code to determine the default ring for
the VNIC when we fill the RSS table.
Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com> Signed-off-by: Shravya KN <shravya.k-n@broadcom.com> Signed-off-by: Michael Chan <michael.chan@broadcom.com> Link: https://patch.msgid.link/20260731190937.807270-2-michael.chan@broadcom.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Bobby Eshleman [Mon, 3 Aug 2026 23:47:29 +0000 (16:47 -0700)]
net/mlx5e: fix BQL reset on SQ re-activation
mlx5e_queue_start() deactivates and re-activates all channels but closes
only the queue being restarted. mlx5e_activate_txqsq() then
unconditionally calls netdev_tx_reset_queue(), zeroing the BQL counters
of channels that kept their in-flight TX WQEs. The next completion then
over-charges and trips the BUG_ON() in dql_completed():
Zhiling Zou [Mon, 3 Aug 2026 00:29:36 +0000 (08:29 +0800)]
net: openvswitch: reallocate update replies for mismatched IDs
ovs_flow_cmd_new() preallocates the optional reply skb before it takes
ovs_mutex and before it knows which existing flow will be updated.
That is normally fine because the skb is sized from the request flow
identifier. That identifier also becomes the inserted flow's identifier.
For updates, however, a request with a UFID may miss the UFID lookup and
then fall back to the flow key lookup. That lookup can legitimately find
an existing key-identified flow. UFIDs are optional and the flow key is
the primary identifier.
For echoed replies, ovs_flow_cmd_fill_info() writes the matched flow's
identifier, not the request identifier used for the preallocation. A short
request UFID can therefore leave too little room for the key identifier.
The fill can then fail with -EMSGSIZE and hit the BUG_ON(error < 0) in the
update path.
Once the update target has been resolved, reallocate the reply skb if the
matched flow needs a larger reply than the request identifier allowed. Do
this before replacing the actions so the request can still fail cleanly if
the rare extra allocation fails.
usbnet: cap max_mtu for drivers without bind callback
usbnet_probe() initializes max_mtu to ETH_MAX_MTU and only caps it
inside the if (info->bind) block. Drivers without a bind callback
never enter this block, so max_mtu stays at ETH_MAX_MTU.
QEMU's usb-net device (0x0525/0xa4a2) is claimed by the cdc_subset
driver which has no bind callback. The guest accepts any MTU from DHCP
(e.g. 65520 from passt), leading to TCP segments that exceed the
device's 2048-byte receive buffer and are silently dropped.
Initialize max_mtu to net->mtu at probe time and update it inside
the bind block.
Alok Tiwari [Sat, 1 Aug 2026 10:09:20 +0000 (03:09 -0700)]
bnge: use int for bnge_fix_rings_count() return value
bnge_fix_rings_count() returns 0 on success or a negative errno on failure
However, bnge_adjust_rings() stores its return value in a u16 variable,
causing negative error codes such as -ENOMEM to be converted to a large
positive value.
Use an int for the return code variable so that error values are
preserved and propagated correctly.
====================
net: atlantic: fix two ring teardown leaks
These are the two fixes from the page_pool conversion series [1],
resent against net as requested in the review of that series. The
page_pool conversion itself stays in net-next and is not part of this
posting; it depends on these fixes, but they stand on their own.
Both patches are unchanged from [1] apart from the collected
Reviewed-by tags, and each carries a Fixes tag and a Cc: stable with
the affected range (patch 1: v4.11+, patch 2: v5.2+). They apply and
were build- and runtime-tested independently of each other and of the
conversion.
Patch 1: aq_vec_deinit() drains the TX rings with a single
aq_ring_tx_clean() call, which is capped at AQ_CFG_TX_CLEAN_BUDGET
descriptors and stops at hw_head, frozen once the hardware and NAPI
have been stopped. Everything beyond that keeps its skb or xdp_frame
when the interface goes down and is lost when the buffer ring is
freed.
Patch 2: aq_ring_rx_deinit() only walks [sw_head, sw_tail). Since the
page reuse strategy was added, a cleaned RX buffer keeps its page for
reuse and refill is batched, so consumed but not yet reposted slots
accumulate in the [sw_tail, sw_head) gap and their pages and DMA
mappings are never released.
Reproduction logs for both leaks (as page_pool stalled shutdowns,
which is how they become visible) are in the notes of the respective
patches.
Yangyu Chen [Sun, 2 Aug 2026 15:46:38 +0000 (23:46 +0800)]
net: atlantic: free RX pages of consumed but not refilled buffers
aq_ring_rx_deinit() only walks [sw_head, sw_tail), the region posted to
hardware. Since the page reuse strategy was added, a cleaned RX buffer
keeps its page (and its DMA mapping) in the ring for reuse, and refill
is batched: aq_ring_rx_fill() returns early until AQ_CFG_RX_REFILL_THRES
slots are free. Slots that were consumed but not yet reposted therefore
sit in the complementary [sw_tail, sw_head) gap with a live page, and
the deinit walk never visits them: up to a refill batch worth of pages
and DMA mappings leak on every interface down.
Walk the whole ring instead and release whatever is still there. Also
bail out if the buffer ring is already gone: a partial
aq_ptp_ring_alloc() failure frees the ring but leaves aq_nic set, so
aq_ptp_ring_deinit() still gets here on the unwind path.
Yangyu Chen [Sun, 2 Aug 2026 15:46:00 +0000 (23:46 +0800)]
net: atlantic: free stranded TX buffers on ring deinit
aq_vec_deinit() drains the TX rings with a single aq_ring_tx_clean()
call, which frees at most AQ_CFG_TX_CLEAN_BUDGET (256) descriptors and
stops at hw_head, which no longer moves once aq_vec_stop() has stopped
the hardware and NAPI. Completed descriptors beyond the budget and
everything still posted in [hw_head, sw_tail) keep their skb or
xdp_frame when the interface goes down: aq_vec_ring_free() then frees
the buffer ring and the references are lost for good.
Today this is a silent memory leak on every interface down under
TX/XDP_TX load. With the conversion of the RX path to page_pool posted
for net-next it becomes much more visible: XDP_TX frames carry fragment
references on the RX ring's page_pool, so a single stranded frame keeps
the pool's inflight count above zero forever. page_pool_destroy() then
never completes, the pool is leaked together with its pages, and
"page_pool_release_retry() stalled pool shutdown" is warned every 60
seconds from that point on, on every ifdown, XDP detach or ring resize
under XDP_TX load.
Bring back aq_ring_tx_deinit() as it was before the removal and use it
for teardown again, with one extension: TX rings can hold xdp_frames
nowadays, so release those too. They are returned with
xdp_return_frame() since this runs in process context.
Fixes: eb36bedf28be ("net: aquantia: remove function aq_ring_tx_deinit") Cc: stable@vger.kernel.org # v4.11+ Reviewed-by: Sukhdeep Singh <sukhdeeps@marvell.com> Signed-off-by: Yangyu Chen <cyy@cyyself.name> Acked-by: Mina Almasry <almasrymina@google.com> Link: https://patch.msgid.link/tencent_EEDC35FAF2750A3A6A0B39BAE0E2C484860A@qq.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Stefan Agner [Mon, 3 Aug 2026 09:51:56 +0000 (11:51 +0200)]
net: stmmac: resume PHY before hardware setup when opening the interface
Since the referenced commit, changing the MTU on a running interface no
longer disconnects and reconnects the PHY; __stmmac_release() merely
stops phylink, which also suspends the PHY (BMCR power-down) when WoL
is not enabled. __stmmac_open() then performs the DMA software reset in
stmmac_hw_setup() before phylink_start() resumes the PHY again.
IEEE 802.3 22.2.4.1.5 allows a PHY to stop its receive clock while
powered down, and stmmac requires a running receive clock for the DMA
software reset to complete (the phylink config sets mac_requires_rxc).
On such setups, e.g. the RK3566-based Home Assistant Green with an
RTL8211F-VD PHY in RGMII mode, any runtime MTU change now times out and
leaves the interface dead:
rk_gmac-dwmac fe010000.ethernet end0: Failed to reset the dma
rk_gmac-dwmac fe010000.ethernet end0: stmmac_hw_setup: DMA engine initialization failed
rk_gmac-dwmac fe010000.ethernet end0: __stmmac_open: Hw setup failed
rk_gmac-dwmac fe010000.ethernet end0: failed reopening the interface after MTU change
In the field this is triggered by NetworkManager applying an MTU while
activating the connection, breaking networking entirely. The same
regression has also been reported on i.MX8MP and reproduced on SoCFPGA
based systems.
Resume the PHY in __stmmac_open() before the hardware setup, making it
the counterpart of the phylink_stop() in __stmmac_release(), like
stmmac_resume() already does for the same reason. phylink_start() also
resumes the PHY, but only after stmmac_hw_setup(), and it cannot be
moved before the hardware setup since it may bring the link up
immediately from a workqueue, racing with the initialization (see the
comment in stmmac_resume()). For the regular ndo_open path the PHY has
just been attached and is not suspended, in which case
phylink_prepare_resume() does nothing.
Fixes: db299a0c09e9 ("net: stmmac: move PHY handling out of __stmmac_open()/release()") Link: https://github.com/home-assistant/operating-system/issues/4858 Tested-by: Alexander Stein <alexander.stein@ew.tq-group.com> Signed-off-by: Stefan Agner <stefan@agner.ch> Tested-by: Maxime Chevallier <maxime.chevallier@bootlin.com> Reviewed-by: Maxime Chevallier <maxime.chevallier@bootlin.com> Link: https://patch.msgid.link/20260803095156.132827-1-stefan@agner.ch Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jakub Kicinski [Wed, 5 Aug 2026 00:32:36 +0000 (17:32 -0700)]
Merge tag 'ovpn-net-20260730' of https://github.com/OpenVPN/ovpn-net-next
Antonio Quartulli says:
====================
Included fixes:
* use rcu_dereference_bh() instead of rcu_access_pointer() where the
pointer is actually dereferenced
* ensure TCP global variables are initialized before they can be
accessed via netlink (e.g. when attaching a TCP socket)
* actually disable IPv4 redirects on multipeer interfaces (the
previous attempt was a no-op and did not survive netns moves)
* hash a floated peer by its transport identity only, consistently
with the add and lookup paths
* zero the sockaddr padding before learning a floated endpoint so it
does not leak into the by_transp_addr hash key
* ensure the socket is owned by ovpn before dereferencing
sk_user_data
* rehash a peer in the by_transp_addr table when its remote endpoint
is updated via CMD_PEER_SET
* avoid re-adding to the hashtables a peer that was concurrently
removed (use-after-free)
* limit keepalive values to one day to avoid overflowing the
delayed-work delay on 32-bit systems
* add the missing rtnl_link_ops->get_size callback so link messages
account for the nested mode attribute
* tag 'ovpn-net-20260730' of https://github.com/OpenVPN/ovpn-net-next:
ovpn: fix incorrect use of rcu_access_pointer()
ovpn: ensure TCP vars are initialized first
ovpn: disable IPv4 redirects on MP interfaces
ovpn: hash floated peer by transport identity only
ovpn: zero-initialize sockaddr before learning a floated endpoint
ovpn: ensure socket is owned by ovpn before deref sk_user_data
ovpn: rehash peer in by_transp_addr table on CMD_PEER_SET
ovpn: skip rehash for peers already removed from by_id
ovpn: limit keepalive values to one day
ovpn: add missing rtnl_link_ops->get_size callback
====================
Martin Kaiser [Tue, 4 Aug 2026 19:46:35 +0000 (21:46 +0200)]
selftests/ftrace: refactor eprobes test to fix argument checks
The add/remove eprobe test installs an eprobe for the openat syscall and
runs ls. It checks the filenames that were opened by ls against a
whitelist and a blacklist.
Commit 206b25c09080 ("tracing: eprobe: read the complete FILTER_PTR_STRING
pointer") fixed access to some string fields in eprobes. This triggers
test failures as the blacklist does not allow relative paths for the
openat parameters.
What makes this test unstable is the fact that the openat calls vary a
lot between different systems.
Refactor the test to make it more robust. "cd <directory>" will issue a
chdir syscall with the target directory as parameter. Set an eprobe on
the sys_enter_chdir event and filter for the exact directory name. Allow
(fault) as fallback.
Link: https://lore.kernel.org/all/20260804194705.760893-1-martin@kaiser.cx/ Fixes: 206b25c09080 ("tracing: eprobe: read the complete FILTER_PTR_STRING pointer") Reported-by: kernel test robot <oliver.sang@intel.com> Closes: https://lore.kernel.org/oe-lkp/202607151010.b68428e1-lkp@intel.com Signed-off-by: Martin Kaiser <martin@kaiser.cx> Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Kyle Zeng [Mon, 3 Aug 2026 12:27:57 +0000 (12:27 +0000)]
ipv6: prevent in6_dev_get() from resurrecting inet6_dev
in6_dev_get() reads dev->ip6_ptr under RCU and then unconditionally
increments its refcount. Device teardown can clear the pointer and drop
the last reference between these operations. The increment then
resurrects an object whose RCU free has already been queued, so callers
can use it after it is freed.
Use refcount_inc_not_zero() and return NULL when the object has already
reached zero. RCU keeps the memory accessible through the attempted
reference acquisition, and a successful increment pins the object for
the caller.
An independent run on the exact unpatched 6f5156d7a31a (v7.2-rc3)
kernel reproduced the invalid reference acquisition as UID 1000:
refcount_t: addition on 0; use-after-free.
ip6_mc_source+0xef4/0x17e0
It was followed by the corresponding reference underflow in
ip6_mc_source(). The supplied trace from the same unpatched revision
additionally shows the access after the RCU read-side section ends:
BUG: KASAN: slab-use-after-free in mutex_lock+0x76/0xe0
Write of size 8 at addr ffff888015b50240 by task poc/1219
Bug found and triaged by OpenAI Security Research and
validated by Trail of Bits.
Fixes: 8814c4b53381 ("[IPV6] ADDRCONF: Convert addrconf_lock to RCU.") Cc: stable@vger.kernel.org Signed-off-by: Kyle Zeng <kylebot@openai.com> Co-developed-by: David Lee <david.lee@trailofbits.com> Signed-off-by: David Lee <david.lee@trailofbits.com> Reviewed-by: Ido Schimmel <idosch@nvidia.com> Link: https://patch.msgid.link/20260803122758.666112-1-david.lee@trailofbits.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
projects / thirdparty / kernel / linux.git / log
