kernel/linux.git/net/xfrm, branch v5.10.257

xfrm_user: fix info leak in build_report()

2026-04-18T08:31:13+00:00

commit d10119968d0e1f2b669604baf2a8b5fdb72fa6b4 upstream. struct xfrm_user_report is a __u8 proto field followed by a struct xfrm_selector which means there is three "empty" bytes of padding, but the padding is never zeroed before copying to userspace. Fix that up by zeroing the structure before setting individual member variables. Cc: stable Cc: Steffen Klassert Cc: Herbert Xu Cc: "David S. Miller" Cc: Eric Dumazet Cc: Jakub Kicinski Cc: Paolo Abeni Cc: Simon Horman Assisted-by: gregkh_clanker_t1000 Signed-off-by: Greg Kroah-Hartman Signed-off-by: Steffen Klassert Signed-off-by: Greg Kroah-Hartman

xfrm: Fix the usage of skb->sk

2026-04-18T08:31:04+00:00

[ Upstream commit 1620c88887b16940e00dbe57dd38c74eda9bad9e ] xfrm assumed to always have a full socket at skb->sk. This is not always true, so fix it by converting to a full socket before it is used. Signed-off-by: Steffen Klassert Reviewed-by: Eric Dumazet Stable-dep-of: 0c0eef8ccd24 ("esp: fix skb leak with espintcp and async crypto") Signed-off-by: Sasha Levin

xfrm: call xdo_dev_state_delete during state update

2026-04-18T08:31:04+00:00

[ Upstream commit 7d2fc41f91bc69acb6e01b0fa23cd7d0109a6a23 ] When we update an SA, we construct a new state and call xdo_dev_state_add, but never insert it. The existing state is updated, then we immediately destroy the new state. Since we haven't added it, we don't go through the standard state delete code, and we're skipping removing it from the device (but xdo_dev_state_free will get called when we destroy the temporary state). This is similar to commit c5d4d7d83165 ("xfrm: Fix deletion of offloaded SAs on failure."). Fixes: d77e38e612a0 ("xfrm: Add an IPsec hardware offloading API") Signed-off-by: Sabrina Dubroca Reviewed-by: Simon Horman Signed-off-by: Steffen Klassert Signed-off-by: Sasha Levin

xfrm: flush all states in xfrm_state_fini

2026-01-19T12:11:24+00:00

[ Upstream commit 42e42562c9cfcdacf000f1b42284a4fad24f8546 ] While reverting commit f75a2804da39 ("xfrm: destroy xfrm_state synchronously on net exit path"), I incorrectly changed xfrm_state_flush's "proto" argument back to IPSEC_PROTO_ANY. This reverts some of the changes in commit dbb2483b2a46 ("xfrm: clean up xfrm protocol checks"), and leads to some states not being removed when we exit the netns. Pass 0 instead of IPSEC_PROTO_ANY from both xfrm_state_fini xfrm6_tunnel_net_exit, so that xfrm_state_flush deletes all states. Fixes: 2a198bbec691 ("Revert "xfrm: destroy xfrm_state synchronously on net exit path"") Reported-by: syzbot+6641a61fe0e2e89ae8c5@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=6641a61fe0e2e89ae8c5 Tested-by: syzbot+6641a61fe0e2e89ae8c5@syzkaller.appspotmail.com Signed-off-by: Sabrina Dubroca Reviewed-by: Simon Horman Signed-off-by: Steffen Klassert Signed-off-by: Sasha Levin

xfrm: also call xfrm_state_delete_tunnel at destroy time for states that were never added

2026-01-19T12:11:24+00:00

[ Upstream commit 10deb69864840ccf96b00ac2ab3a2055c0c04721 ] In commit b441cf3f8c4b ("xfrm: delete x->tunnel as we delete x"), I missed the case where state creation fails between full initialization (->init_state has been called) and being inserted on the lists. In this situation, ->init_state has been called, so for IPcomp tunnels, the fallback tunnel has been created and added onto the lists, but the user state never gets added, because we fail before that. The user state doesn't go through __xfrm_state_delete, so we don't call xfrm_state_delete_tunnel for those states, and we end up leaking the FB tunnel. There are several codepaths affected by this: the add/update paths, in both net/key and xfrm, and the migrate code (xfrm_migrate, xfrm_state_migrate). A "proper" rollback of the init_state work would probably be doable in the add/update code, but for migrate it gets more complicated as multiple states may be involved. At some point, the new (not-inserted) state will be destroyed, so call xfrm_state_delete_tunnel during xfrm_state_gc_destroy. Most states will have their fallback tunnel cleaned up during __xfrm_state_delete, which solves the issue that b441cf3f8c4b (and other patches before it) aimed at. All states (including FB tunnels) will be removed from the lists once xfrm_state_fini has called flush_work(&xfrm_state_gc_work). Reported-by: syzbot+999eb23467f83f9bf9bf@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=999eb23467f83f9bf9bf Fixes: b441cf3f8c4b ("xfrm: delete x->tunnel as we delete x") Signed-off-by: Sabrina Dubroca Signed-off-by: Steffen Klassert Signed-off-by: Sasha Levin

Revert "xfrm: destroy xfrm_state synchronously on net exit path"

2026-01-19T12:11:24+00:00

[ Upstream commit 2a198bbec6913ae1c90ec963750003c6213668c7 ] This reverts commit f75a2804da391571563c4b6b29e7797787332673. With all states (whether user or kern) removed from the hashtables during deletion, there's no need for synchronous destruction of states. xfrm6_tunnel states still need to have been destroyed (which will be the case when its last user is deleted (not destroyed)) so that xfrm6_tunnel_free_spi removes it from the per-netns hashtable before the netns is destroyed. This has the benefit of skipping one synchronize_rcu per state (in __xfrm_state_destroy(sync=true)) when we exit a netns. Signed-off-by: Sabrina Dubroca Signed-off-by: Steffen Klassert Signed-off-by: Sasha Levin

xfrm: delete x->tunnel as we delete x

2026-01-19T12:11:24+00:00

[ Upstream commit b441cf3f8c4b8576639d20c8eb4aa32917602ecd ] The ipcomp fallback tunnels currently get deleted (from the various lists and hashtables) as the last user state that needed that fallback is destroyed (not deleted). If a reference to that user state still exists, the fallback state will remain on the hashtables/lists, triggering the WARN in xfrm_state_fini. Because of those remaining references, the fix in commit f75a2804da39 ("xfrm: destroy xfrm_state synchronously on net exit path") is not complete. We recently fixed one such situation in TCP due to defered freeing of skbs (commit 9b6412e6979f ("tcp: drop secpath at the same time as we currently drop dst")). This can also happen due to IP reassembly: skbs with a secpath remain on the reassembly queue until netns destruction. If we can't guarantee that the queues are flushed by the time xfrm_state_fini runs, there may still be references to a (user) xfrm_state, preventing the timely deletion of the corresponding fallback state. Instead of chasing each instance of skbs holding a secpath one by one, this patch fixes the issue directly within xfrm, by deleting the fallback state as soon as the last user state depending on it has been deleted. Destruction will still happen when the final reference is dropped. A separate lockdep class for the fallback state is required since we're going to lock x->tunnel while x is locked. Fixes: 9d4139c76905 ("netns xfrm: per-netns xfrm_state_all list") Signed-off-by: Sabrina Dubroca Signed-off-by: Steffen Klassert Signed-off-by: Sasha Levin

xfrm: Sanitize marks before insert

2025-06-04T12:37:07+00:00

[ Upstream commit 0b91fda3a1f044141e1e615456ff62508c32b202 ] Prior to this patch, the mark is sanitized (applying the state's mask to the state's value) only on inserts when checking if a conflicting XFRM state or policy exists. We discovered in Cilium that this same sanitization does not occur in the hot-path __xfrm_state_lookup. In the hot-path, the sk_buff's mark is simply compared to the state's value: if ((mark & x->mark.m) != x->mark.v) continue; Therefore, users can define unsanitized marks (ex. 0xf42/0xf00) which will never match any packet. This commit updates __xfrm_state_insert and xfrm_policy_insert to store the sanitized marks, thus removing this footgun. This has the side effect of changing the ip output, as the returned mark will have the mask applied to it when printed. Fixes: 3d6acfa7641f ("xfrm: SA lookups with mark") Signed-off-by: Paul Chaignon Signed-off-by: Louis DeLosSantos Co-developed-by: Louis DeLosSantos Signed-off-by: Steffen Klassert Signed-off-by: Sasha Levin

xfrm_output: Force software GSO only in tunnel mode

2025-04-10T12:30:52+00:00

[ Upstream commit 0aae2867aa6067f73d066bc98385e23c8454a1d7 ] The cited commit fixed a software GSO bug with VXLAN + IPSec in tunnel mode. Unfortunately, it is slightly broader than necessary, as it also severely affects performance for Geneve + IPSec transport mode over a device capable of both HW GSO and IPSec crypto offload. In this case, xfrm_output unnecessarily triggers software GSO instead of letting the HW do it. In simple iperf3 tests over Geneve + IPSec transport mode over a back-2-back pair of NICs with MTU 1500, the performance was observed to be up to 6x worse when doing software GSO compared to leaving it to the hardware. This commit makes xfrm_output only trigger software GSO in crypto offload cases for already encapsulated packets in tunnel mode, as not doing so would then cause the inner tunnel skb->inner_networking_header to be overwritten and break software GSO for that packet later if the device turns out to not be capable of HW GSO. Taking a closer look at the conditions for the original bug, to better understand the reasons for this change: - vxlan_build_skb -> iptunnel_handle_offloads sets inner_protocol and inner network header. - then, udp_tunnel_xmit_skb -> ip_tunnel_xmit adds outer transport and network headers. - later in the xmit path, xfrm_output -> xfrm_outer_mode_output -> xfrm4_prepare_output -> xfrm4_tunnel_encap_add overwrites the inner network header with the one set in ip_tunnel_xmit before adding the second outer header. - __dev_queue_xmit -> validate_xmit_skb checks whether GSO segmentation needs to happen based on dev features. In the original bug, the hw couldn't segment the packets, so skb_gso_segment was invoked. - deep in the .gso_segment callback machinery, __skb_udp_tunnel_segment tries to use the wrong inner network header, expecting the one set in iptunnel_handle_offloads but getting the one set by xfrm instead. - a bit later, ipv6_gso_segment accesses the wrong memory based on that wrong inner network header. With the new change, the original bug (or similar ones) cannot happen again, as xfrm will now trigger software GSO before applying a tunnel. This concern doesn't exist in packet offload mode, when the HW adds encapsulation headers. For the non-offloaded packets (crypto in SW), software GSO is still done unconditionally in the else branch. Reviewed-by: Dragos Tatulea Reviewed-by: Yael Chemla Reviewed-by: Leon Romanovsky Fixes: a204aef9fd77 ("xfrm: call xfrm_output_gso when inner_protocol is set in xfrm_output") Signed-off-by: Cosmin Ratiu Signed-off-by: Steffen Klassert Signed-off-by: Sasha Levin

xfrm: replay: Fix the update of replay_esn->oseq_hi for GSO

2025-03-13T11:46:56+00:00

[ Upstream commit c05c5e5aa163f4682ca97a2f0536575fc7dbdecb ] When skb needs GSO and wrap around happens, if xo->seq.low (seqno of the first skb segment) is before the last seq number but oseq (seqno of the last segment) is after it, xo->seq.low is still bigger than replay_esn->oseq while oseq is smaller than it, so the update of replay_esn->oseq_hi is missed for this case wrap around because of the change in the cited commit. For example, if sending a packet with gso_segs=3 while old replay_esn->oseq=0xfffffffe, we calculate: xo->seq.low = 0xfffffffe + 1 = 0x0xffffffff oseq = 0xfffffffe + 3 = 0x1 (oseq < replay_esn->oseq) is true, but (xo->seq.low < replay_esn->oseq) is false, so replay_esn->oseq_hi is not incremented. To fix this issue, change the outer checking back for the update of replay_esn->oseq_hi. And add new checking inside for the update of packet's oseq_hi. Fixes: 4b549ccce941 ("xfrm: replay: Fix ESN wrap around for GSO") Signed-off-by: Jianbo Liu Reviewed-by: Patrisious Haddad Signed-off-by: Leon Romanovsky Signed-off-by: Steffen Klassert Signed-off-by: Sasha Levin