From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail-yx2-f38.google.com (mail-yx2-f38.google.com [74.125.224.166]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id E6DF24BE42B for ; Thu, 1 Oct 2026 22:52:25 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=74.125.224.166 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1790895148; cv=none; b=EZnh0Jkbm0U02ByDnxN+se2Hk4QbWGDIq3tqtWxPLpV0QG+zSq1TKZPWTvSJTDkTJOCdY//9xE2BaF3C0JBLvloMcrq6QXzrdaRr4W7Iw/aw1MjfAoljpl00w6xJfvliwIaNK+rMRa1W+4WYZUx4qilneGwggxgINXXgo38i/L0= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1790895148; c=relaxed/simple; bh=t5EBaXgjCtOlPReSbRxo6xipJNaIHWdHxsirSpF8mEI=; h=Date:From:To:Cc:Message-ID:In-Reply-To:References:Subject: MIME-Version:Content-Type; b=BE6hOKEEYeFPGsNwASi/2l4rtHD/jY+HXUabMU9P/6CANiDk/AFRE/X6IRPgnJUITk7U9hyS9C6U/9iXFkknivgwT7RixDDf3vEzQo85GezYC1mOrRHEar7NjEGVPMEFyP6sJ9BYnepJRXi/xAG22Vs9qNHcI+f9El9xEfOg2Gs= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=gmail.com; spf=pass smtp.mailfrom=gmail.com; dkim=pass (2048-bit key) header.d=gmail.com header.i=@gmail.com header.b=HJ5Huim3; arc=none smtp.client-ip=74.125.224.166 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=gmail.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=gmail.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=gmail.com header.i=@gmail.com header.b="HJ5Huim3" Received: by mail-yx2-f38.google.com with SMTP id 00721157ae682-8ab3e848918so37415797b3.3 for ; Thu, 01 Oct 2026 15:52:25 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20251104; t=1790895145; x=1791499945; darn=vger.kernel.org; h=content-transfer-encoding:content-type:mime-version:subject :references:in-reply-to:message-id:cc:to:from:date:from:to:cc :subject:date:message-id:reply-to:content-type; bh=1r9R/81H7+7sKSQQACrmcWEff6sVFjsufZ7eFfFqfJM=; b=HJ5Huim326WSrUnTE9TqFlK0ONU7L5iNI8uJ73SJ0hHH63IlmLiQgnksh6IDl99CLv lqoxNZvVH7irgqtzJfPdgf819Mx4xPXbniaMi62zoq6UNYDFw8Qhb+Zp4m9U72mSV/Re g/PrRRHXemU4KRpgVAx0TXOWpjJ7LvARH6VEyjgla1YX8L3Ln/2pndf8bsPBIx5+gpge z2EoTOdGbPEhgkKBP4BtEjvkMaNuPq5KesqDTYelbijOWujRIYOu++KCpfjWqXFhCOMz sP5FAQMRWJEjLFLaYU3/bl9cKTl12Ai9nKrUy1u+OBQL/1UosucqbZLgj3bTEEJzVINi 1vTg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20260707; t=1790895145; x=1791499945; h=content-transfer-encoding:content-type:mime-version:subject :references:in-reply-to:message-id:cc:to:from:date:x-gm-gg :x-gm-message-state:from:to:cc:subject:date:message-id:reply-to :content-type; bh=1r9R/81H7+7sKSQQACrmcWEff6sVFjsufZ7eFfFqfJM=; b=cKVN3R7Dq3DlL56YDA3i1y897Mg+Cxqoj7TmGvZX489qpooYUMM8O/M/AR9Gp+dGib fO1fX9cq3M1s07WV/+Qhwj3t3xXEp6FmPCtMxYkoLJr2W4TSWaI1/Vu4Ra8iRBezlGw+ OmGHm/pdLELbFtBgS+nAxk6KiwFs/xvH52bKWM5hlYm8Xb5EhEDFdcF6ElqsnVjTKdmj fY8Yxx/k3d2GZsj82dyKs1n6epbFCUV8zHO4RKRO6mTPGP+cxxgw6BI372iza7XPHeBf RZDalIPZOMCxnMVhrBavWWSUNnPRjMcT35Jycj9rdf0RrHxnoELsqKhy7ZiuDpDL7ZAB XtDQ== X-Forwarded-Encrypted: i=1; AKwUvByi79KhFc9oXSHtK7Uolfvs6COZkiNa8n9SS2dpwVEw4pIiQ8I/bLqq2zTIBeDIo12o9H6djY0uRZfbJ94=@vger.kernel.org X-Gm-Message-State: AFq9FYKj+sS7aSoNmicWF1/8JtamHHCZQ/kluGNgt5kEGKQoQ1x4/3R6 IJAFVI3D9FWGEM04t0j5y/b6WVnBtdsdW90jyHT516rudfQLrvNdnHR2 X-Gm-Gg: AYBFou0y0MSpIKv/KSVv7FC8b9pRLc8xzQAb1MAirYQjYv+n70hGK90dAqnnvmeyp1f f/8ZUumD8MEDZYXeEpAYo/AJBl5GcsJMiTPr2ZdsXd2j47yOr3aLPq4FECKbwLJsOo9lsUPXost IaAyKIUs7nI7CVeQnyITfcqCs+HtCd9nxmSPCIDDCZc+F8sFVv6qqrdC4IEU9BomA+01mjFoIRm LDTaqrMbv5+BPx3SEK6AVmUM74jbDBjQS8X8cPw2SdpSa3/uSg7wwhFV2VhEIojiGdVO1fQ0Flo RSzrKbQmTXaiR3k9mgj9v4F55MagvyTVTR2yISJ8p9nQwVskFPLtakKc/PLCZ1ytd8vc0q9FM22 ItWktDUSUyX6AuvUtt0zJLa9iAdhCRJMPQSt452d0JZDZXeBHYY+NDXNQhCIaDXkyBDlFExPNVX lnE88uV8i4yMOArRTIcebMsZgUtWIxWr3FOWaAboL/IaIUjodwSp/K+roV0AFrBJSmkb9HKfw8Y LkPerscUT0Hp8H6zurjp0/KomdtVicC3QBoDw4Ioj75zmrUoFXy X-Received: by 2002:a05:690c:f14:b0:8a9:543c:db7f with SMTP id 00721157ae682-8ae3abe15afmr3055557b3.49.1790895144846; Thu, 01 Oct 2026 15:52:24 -0700 (PDT) Received: from gmail.com (111.46.245.35.bc.googleusercontent.com. [35.245.46.111]) by smtp.gmail.com with ESMTPSA id 00721157ae682-8ae333e9a0bsm2659807b3.36.2026.10.01.15.52.24 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Thu, 01 Oct 2026 15:52:24 -0700 (PDT) Date: Thu, 01 Oct 2026 18:52:23 -0400 From: Willem de Bruijn To: "Michael S. Tsirkin" , Eric Dumazet Cc: Willem de Bruijn , Paulos Yibelo , netdev@vger.kernel.org, richard@nod.at, anton.ivanov@cambridgegreys.com, johannes@sipsolutions.net, jasowangio@gmail.com, eperezma@redhat.com, xuanzhuo@linux.alibaba.com, andrew+netdev@lunn.ch, pablo@netfilter.org, fw@strlen.de, phil@nwl.cc, razor@blackwall.org, idosch@nvidia.com, dsahern@kernel.org, davem@davemloft.net, kuba@kernel.org, pabeni@redhat.com, horms@kernel.org, linux-um@lists.infradead.org, virtualization@lists.linux.dev, netfilter-devel@vger.kernel.org, coreteam@netfilter.org, bridge@lists.linux.dev, linux-kernel@vger.kernel.org Message-ID: In-Reply-To: <20260923081209-mutt-send-email-mst@kernel.org> References: <20260922030310.8684-1-habte.yibelo@gmail.com> <20260922030310.8684-2-habte.yibelo@gmail.com> <20260923061215-mutt-send-email-mst@kernel.org> <20260923081209-mutt-send-email-mst@kernel.org> Subject: Re: [PATCH net v6 1/2] net: validate virtio checksum start after network header Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Michael S. Tsirkin wrote: > On Wed, Sep 23, 2026 at 12:46:42PM +0200, Eric Dumazet wrote: > > On Wed, Sep 23, 2026 at 12:21=E2=80=AFPM Michael S. Tsirkin wrote: > > > > > > On Tue, Sep 22, 2026 at 09:27:26PM -0400, Willem de Bruijn wrote: > > > > Willem de Bruijn wrote: > > > > > Paulos Yibelo wrote: > > > > > > __virtio_net_hdr_to_skb() checks a minimum network-header len= gth for > > > > > > CHECKSUM_PARTIAL packets. Its checksum start is relative to s= kb->data, > > > > > > but some callers have not established skb->network_header whe= n they > > > > > > convert the virtio header. > > > > > > > > > > > > Pass the data-relative L3 origin explicitly. Ethernet receive= paths > > > > > > parse the frame and nested VLAN headers without changing skb = state. > > > > > > AF_PACKET uses the frame's actual L3 origin even when the soc= ket > > > > > > protocol is ETH_P_IP and the raw frame carries VLAN tags. Non= -Ethernet > > > > > > AF_PACKET devices retain their established skb network offset= . > > > > > > > > > > > > Also pass the actual L3 protocol so IPv6 packets use the 40-b= yte base > > > > > > header minimum even without TCPv6 GSO. IFF_TUN obtains that p= rotocol > > > > > > from the packet before skb->protocol is set. Name the Etherne= t parser > > > > > > accordingly, use the same origin for tunnel validation, and p= ropagate > > > > > > conversion failures in UML. > > > > > > > > > > > > The bound remains a minimum; fragmentation paths separately v= alidate > > > > > > the parsed IPv4 or IPv6 header length before completing a che= cksum. > > > > > > > > > > > > Fixes: 49d14b54a527 ("net: test for not too small csum_start = in virtio_net_hdr_to_skb()") > > > > > > Fixes: a2fb4bc4e2a6 ("net: implement virtio helpers to handle= UDP GSO tunneling.") > > > > > > Reported-by: Paulos Yibelo > > > > > > Link: https://lore.kernel.org/netdev/20260920004733.6473-2-ha= bte.yibelo@gmail.com/ > > > > > > Cc: stable@vger.kernel.org > > > > > > Assisted-by: LLM > > > > > > Signed-off-by: Paulos Yibelo > > > > > > --- > > > > > > arch/um/drivers/vector_transports.c | 13 ++++- > > > > > > drivers/net/tun_vnet.h | 52 ++++++++++++++++- > > > > > > drivers/net/virtio_net.c | 10 +++- > > > > > > include/linux/virtio_net.h | 87 +++++++++++++++++++= +++++----- > > > > > > net/packet/af_packet.c | 24 +++++++- > > > > > > 5 files changed, 163 insertions(+), 23 deletions(-) > > > > > > > > > > The fix may still miss the case IPv4 packets have options. > > > > > > > > > > This version is a very large patch. > > > > > > > > > > Untested shorter first suggestion by bot, which looks plausible= as a > > > > > starting point for discussion. > > > > > > > > Cleaned up some more: > > > > > > > > diff --git a/drivers/net/tun_vnet.h b/drivers/net/tun_vnet.h > > > > index f4c652b1fa44..c24607af2aad 100644 > > > > --- a/drivers/net/tun_vnet.h > > > > +++ b/drivers/net/tun_vnet.h > > > > @@ -180,6 +180,9 @@ static inline int tun_vnet_hdr_put(int sz= , struct iov_iter *iter, > > > > static inline int tun_vnet_hdr_to_skb(unsigned int flags, st= ruct sk_buff *skb, > > > > const struct virtio_net_hdr *hdr) > > > > { > > > > + if ((flags & TUN_TYPE_MASK) =3D=3D IFF_TUN) > > > > + skb_reset_network_header(skb); > > > > + > > > > return virtio_net_hdr_to_skb(skb, hdr, tun_vnet_is_littl= e_endian(flags)); > > > > } > > > > > > > > @@ -199,6 +202,9 @@ tun_vnet_hdr_tnl_to_skb(unsigned int flag= s, netdev_features_t features, > > > > struct sk_buff *skb, > > > > const struct virtio_net_hdr_v1_hash_tunnel *hdr)= > > > > { > > > > + if ((flags & TUN_TYPE_MASK) =3D=3D IFF_TUN) > > > > + skb_reset_network_header(skb); > > > > + > > > > > > > > diff --git a/include/linux/virtio_net.h b/include/linux/virti= o_net.h > > > > index c381b916c1b5..02c448de0802 100644 > > > > --- a/include/linux/virtio_net.h > > > > +++ b/include/linux/virtio_net.h > > > > @@ -48,6 +48,42 @@ static inline int virtio_net_hdr_set_proto= (struct sk_buff *skb, > > > > return 0; > > > > } > > > > > > > > +static inline int virtio_net_hdr_nh_min_len(const struct sk_= buff *skb, > > > > + unsigned int nh_min_len) > > > > +{ > > > > + int thoff =3D skb_transport_offset(skb); > > > > + __be16 proto; > > > > + int nhoff; > > > > + > > > > + if (skb_network_header_was_set(skb)) { > > > > + nhoff =3D skb_network_offset(skb); > > > > + proto =3D skb->protocol; > > > > + } else { > > > > + if (unlikely(thoff < ETH_HLEN)) > > > > + return -EINVAL; > > > > + nhoff =3D ETH_HLEN; > > > > + proto =3D eth_hdr(skb)->h_proto; > > > > + } > > > > + > > > > + if (eth_type_vlan(proto)) { > > > > + proto =3D __vlan_get_protocol(skb, proto, &nhoff); > > > > + if (!proto) > > > > + return -EINVAL; > > > > + } > > > > + > > > > + if (proto =3D=3D htons(ETH_P_IP)) { > > > > + const struct iphdr *iph =3D (void *)(skb->data + nho= ff); > > > > + > > > > + if (unlikely(thoff < nhoff + sizeof(*iph))) > > > > + return -EINVAL; > > > > + nh_min_len =3D max_t(u32, iph->ihl * 4, sizeof(*iph)= ); > > > > + } else if (proto =3D=3D htons(ETH_P_IPV6)) { > > > > + nh_min_len =3D sizeof(struct ipv6hdr); > > > > + } > > > > + > > > > + return nhoff + nh_min_len; > > > > +} > > > > + > > > > static inline int __virtio_net_hdr_to_skb(struct sk_buff *sk= b, > > > > const struct virtio_net_hdr *hdr, > > > > bool little_endian, u8 hdr_gso_type) > > > > @@ -98,13 +134,15 @@ static inline int __virtio_net_hdr_to_sk= b(struct sk_buff *skb, > > > > u32 start =3D __virtio16_to_cpu(little_endian, hdr->= csum_start); > > > > u32 off =3D __virtio16_to_cpu(little_endian, hdr->cs= um_offset); > > > > u32 needed =3D start + max_t(u32, thlen, off + sizeo= f(__sum16)); > > > > + int min_thoff; > > > > > > > > if (!pskb_may_pull(skb, needed)) > > > > return -EINVAL; > > > > > > > > if (!skb_partial_csum_set(skb, start, off)) > > > > return -EINVAL; > > > > - if (skb_transport_offset(skb) < nh_min_len) > > > > + min_thoff =3D virtio_net_hdr_nh_min_len(skb, nh_min_= len); > > > > + if (min_thoff < 0 || skb_transport_offset(skb) < min= _thoff) > > > > return -EINVAL; > > > > > > > > > Certainly looks much better. But I'd like to ask, generally: > > > doesn't the net stack need to protect against weird packets? > > > > > = > > It does, but csum_start/csum_offset are not "weird packet" material, > > they are skb metadata, not bytes on the wire. > > = > > For essentially every skb in the kernel, this metadata is produced by= > > the kernel itself, > > from headers it has just built or just parsed, and the rest of the > > stack (GSO, skb_checksum_help(), > > fragmentation, netfilter, and every driver doing TX csum offload) > > consumes it as an invariant. > > The only producers of attacker/guest controlled CHECKSUM_PARTIAL meta= data are > > the virtio_net_hdr_to_skb() callers: af_packet, tun/tap, virtio_net, > > and the UML vector driver. > > = > > So this is not "virtio specific validation", this is input validation= > > at the one trust boundary > > where the invariant can be violated. > = > = > Thanks for the explanation Eric! > = > > = > > > It seems likely that not all drivers validate headers defensively, > > > and incoming packets can easily become outgoing ones. > > = > > Packets coming from a real NIC are CHECKSUM_UNNECESSARY, > > CHECKSUM_COMPLETE or CHECKSUM_NONE. > > They do not carry a remote-provided csum_start. > > (Remote checksum offload is the rare exception, and there the offsets= > > are computed by the stack from headers it just parsed.) > > = > > Incoming packets becoming outgoing ones is precisely the problem here= : > > a virtio_net RX skb with VIRTIO_NET_HDR_F_NEEDS_CSUM becomes > > an skb that can be bridged/forwarded/fragmented and then handed to a = real NIC. > = > Well: > = > $ git grep 'skb->csum_start\ =3D' > drivers/net/ethernet/hisilicon/hns3/hns3_enet.c: skb->csum_start= =3D (unsigned char *)th - skb->head; > drivers/net/ethernet/mellanox/mlx5/core/en_rx.c: skb->csum_start= =3D (unsigned char *)uh - skb->head; > drivers/net/ethernet/mellanox/mlx5/core/en_rx.c: skb->csum_start= =3D (unsigned char *)uh - skb->head; > drivers/net/ethernet/mellanox/mlx5/core/en_rx.c: skb->csum_start= =3D (unsigned char *)tcp - skb->head; > drivers/net/ethernet/mellanox/mlx5/core/en_rx.c: skb->csum_start= =3D (unsigned char *)tcp - skb->head; > drivers/net/ethernet/stmicro/stmmac/stmmac_selftests.c: skb->cs= um_start =3D skb_transport_header(skb) - skb->head; > = > What did I miss? > = > > = > > > So do we even need virtio specific validation, or is it enough to > > > validate everything in the net stack, where we are poking at the > > > header anyway? > > = > > The core stack cannot afford it. > > = > > 1) Drivers program skb->csum_start / skb->csum_offset straight into > > the TX descriptor, and the NIC will happily write two bytes wherever > > it was told, e.g. into the IP header of the frame we put on the wire.= > > Auditing/adding checks in every driver is not realistic, and drivers > > must not pay for it. > > = > > 2) Catching this in the core would mean re-parsing L2/L3 in > > dev_hard_start_xmit() > > (or in every place that eventually looks at the transport header) for= > > the 99.99+% > > of packets that were built by the stack and are known to be consisten= t. > > That is far more expensive than one check at injection time. > > = > > 3) Failing at injection returns -EINVAL to the sendmsg()/writev() cal= ler, > > which is the correct and testable behavior. Failing later means dropp= ing > > the packet deep in the xmit path, usually with a splat: the commit be= ing fixed > > here (49d14b54a527) exists exactly because such a packet reached > > skb_checksum_help() from ip_do_fragment() and hit the "offset (-6) >=3D= > > skb_headlen() (14)" WARN. > > = > > > > > > Or maybe it's more a defense in depth thing? > > > > > > My worries: > > > - more poking at the header, more cache misses, where we really > > > do not need that > > = > > I do not expect anything measurable. > > = > > af_packet and tun: we have just copied that header from user space, i= t is in L1. > > virtio_net: we call eth_type_trans() right after, and GRO parses L3/L= 4 > > immediately. > > = > > The whole block is under the CHECKSUM_PARTIAL condition, where we alr= eady > > do pskb_may_pull() and skb_partial_csum_set(), i.e. we already touch > > this cache line. > > Reading iph->ihl from a cache line we are about to read anyway is noi= se. > > = > > > - future protocol extensions that now will require surgery in > > > virtio, instead of just being passed through to the host > > = > > Fair, and this is an argument about how the check is written, not > > about whether it exists. > > The rule should be: > > = > > Only tighten the bound for the protocols we already parse (IPv4/IPv6)= , > > and keep the existing generic minimum for anything else. > > = > > Then an unknown ethertype simply keeps flowing, no surgery is needed.= > > Willem's version does that. A new protocol would only be impacted if = it wanted > > csum_start to point inside what we consider the L3 header, and such a= packet > > would not survive the rest of the stack anyway. > > = > > So: not defense in depth, but validation at the trust boundary, where= > > it is cheapest > > and where we can still report the error to the producer. Just an update that I have not forgotten about this. Just staging it after Eric's related series, which simplifies the challenge, and the patc= h. https://lore.kernel.org/netdev/20261001191140.2818991-1-edumazet@kernel.o= rg/ The main complexity lies in having many callers into this path, and whether you can trust 1. that fields initialized (e.g., skb->dev, network_header, protocol) 2. that they can be trusted (any coming from userspace: no) The list of callers and invariants on some of the state: Caller skb->dev network_hdr skb->protocol skb->da= ta -------------------------------------------------------------------------= -- virtio_net.c ARPHRD_ETHER ETH_HLEN 0 (unset) Etherne= t vector_transports.c ARPHRD_ETHER ETH_HLEN 0 (unset) Etherne= t tap.c (macvtap/ipvtap) ARPHRD_ETHER ETH_HLEN eth h_proto Etherne= t tun.c (IFF_TAP) ARPHRD_ETHER ETH_HLEN 0 (unset) Etherne= t tun.c (IFF_TUN) ARPHRD_NONE 0 pi.proto Raw L3 af_packet.c (ETHER) ARPHRD_ETHER 14 (+vlan) sll_protocol Etherne= t af_packet.c (non-ETHER) !=3D ETHER hard_hdr_len sll_protocol Dev L= 2+L3 After Eric's series skb->dev, skb->network_header, and skb->protocol can be assumed to always be set. Though skb->protocol, for instance, can still not be trusted. Nor can gso_type or even iphdr. Let alone that all three agree. Goal is not to drop inconsistent packets (though I'd love to), but only to protect the kernel. A particularly tricky edge case is af_packet with non-ETHER protocol that= does set VIRTIO_NET_HDR_F_NEEDS_CSUM. I'm not aware of any non-IP transpo= rts that use CHECKSUM_PARTIAL. But a realistic use-case is an IP in MPLS pack= et, for instance.