From mboxrd@z Thu Jan 1 00:00:00 1970 Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.remote-ip=217.70.190.124 ARC-Seal: i=1; d=subspace.kernel.org; s=arc20261015; a=rsa-sha256; cv=none; t=1794786419; b=HFEu2nLKsLWCbV4fLGVCl0ARbJXfiuQp+Vh9qGAgb3n2hTo0VLq3gd5Z9HbAXBPDWm2t DvtJ7SZJErkFsgmw00ppAWjEugPF+cel/v34bWxxi2lFUVUBiltvJnGTUum6nSqioBfpM uUVI1VfDKEhI0G3wg+0ehiB+LfIvu9PhMl56LuYtM4tqxt94ED6oQ33uypLdXpap7kR+D 6uEgvMrsP7dIBxBqg4cEocHcNMJgPVF2SCVtSFRvRQZe1Im7EvbMy/I+fNhBZaA4IE69e Ex409DKmXQSKEA9pWbqyWrfZKfskvpem4jtS0KYAQuoTbtpMMlF/8LEbGIfNHXC6dwA== ARC-Message-Signature: i=1; d=subspace.kernel.org; s=arc20261015; a=rsa-sha256; c=relaxed/relaxed; t=1794786419; h=DMARC-Filter:DKIM-Signature:Received:Date:From:To:Cc:Subject: Message-ID:References:MIME-Version:Content-Type:Content-Disposition: Content-Transfer-Encoding:In-Reply-To:X-Spam-Status: X-Spam-Checker-Version; bh=4FXFrfDW68p2ZmuJnmPotANMx0Dmm0w6nypgxRy6L9g=; b=o+YxKLXRaC/NTjROnNJ2ggylJqPsG88AfyCtTk08CLYNbJ1GWl44g3Kx7x93ixGTeih5 p9a6CVB9GGfLvW9vPh8k88bvufDuwNsFk5xXAhfNhBJ3wXPzPzCnCaZZMaI6yvq9bMGWk JquW/cUABi426eHHu+OsYuuid126dYD7TOsNBK/0Q5zRbAc9y04ZKZJqNMQvKgBzsV2qJ Z6LKRzYmMkXv8bXcI8LUCrkxt0Gj7fwguzdA0kWA2tJrImUnr4DzYexn1ALYw6ddAtVkK p7D20aGgcx0liwyCyapNk6aCP5STgRGd2Qo72ajYBXqiWwV/1kmd/Hb3KlhP/wq1EKw== ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=none header.from=netfilter.org; spf=pass smtp.mailfrom=netfilter.org; dkim=pass header.d=netfilter.org header.i=@netfilter.org header.a=rsa-sha256 header.s=2025 header.b=Dd+r1sfD; arc=none smtp.remote-ip=217.70.190.124 Received: from mail.netfilter.org (mail.netfilter.org [217.70.190.124]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature ECDSA (prime256v1)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with UTF8SMTPS id 20223606A31F; Sun, 15 Nov 2026 23:46:58 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.4.2 smtp.subspace.kernel.org 20223606A31F Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=netfilter.org Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=netfilter.org Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=netfilter.org header.i=@netfilter.org header.a=rsa-sha256 header.s=2025 header.b=Dd+r1sfD DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=netfilter.org; s=2025; t=1794786416; bh=4FXFrfDW68p2ZmuJnmPotANMx0Dmm0w6nypgxRy6L9g=; h=Date:From:To:Cc:Subject:References:In-Reply-To:From; b=Dd+r1sfDT9Sf7+n2u2kjbeYE0ucVdxs4PA62TvV9V3i+WZOz4+h39l4ddHMdViLDS Ekq/lAzR+jgAFpTxNaZlNIBiplt5hLAGKQFllgDmYxs7uMPzcjU5UUsISUY5pbKKE5 o93jA1cIcde+WmPBhqM60dhxDWsgNdkoXi/aHzImyU5TkqlTLTVD+ntmVdCp7qkB+G BAQElAIY0wyAjS2l6poF2crezdstjj1ySpYlVeGvC1IooQpuqK0rLJu/cqN7ByJuPb cuB1UwF0mN/iuNV1tnOcfraGEPMaxxK7E6SQUhAWZuS765VNDGWe9EzEaxC+uh0Kvn IvWkCUbnmDzgA== Received: from netfilter.org (mail-agni [217.70.190.124]) by mail.netfilter.org (Postfix) with UTF8SMTPSA id 87B736005D; Mon, 16 Nov 2026 00:46:56 +0100 (CET) Date: Mon, 16 Nov 2026 00:46:53 +0100 From: Pablo Neira Ayuso To: netfilter-devel@vger.kernel.org Cc: davem@davemloft.net, netdev@vger.kernel.org, kuba@kernel.org, pabeni@redhat.com, edumazet@kernel.org, horms@kernel.org, fw@strlen.de, ja@ssi.bg Subject: Re: [PATCH net-next 00/15] Netfilter updates for net-next Message-ID: References: <20261113031251.2437074-1-pablo@netfilter.org> X-Mailing-List: netdev@vger.kernel.org List-Id: List-Post: List-Help: List-Subscribe: List-Unsubscribe: List-Owner: Precedence: list MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: <20261113031251.2437074-1-pablo@netfilter.org> Hi, I'm attaching a quick summary to address the LLM generated reports: Patch #1 LLM reviewer refers to CLOSING state which is not covered by this patch, which is aiming at the TEARDOWN state. Patch #2 still needs to use the dst that is passed via flow_tuple, ie. use tuple dst_cache. Patch #3 LLM reviewer refers to a possible different scenario where stats can get lost, ie. conntrack -D or -F. I don't think that belongs to this patch, this possibly need a delayed ct event delivery for Patch #4 and #5 LLM reviewer provides hints to improve drivers, these patches only adapted the existing code to fetch stats from the delete entry command. Patch #8 LLM reviewer refers to missing update in selftest, which breaks. Patch #11 LLM reviewer mentions that the new selftest script does not cover for all UDP checksum scenarios in the flowtable. I don't see a review why this cannot be improved in a follow up. For patches #2 and #8 a version 2 of these PR might be needed. But for the remaining patches, I think it is not the way to go to address the issues described in those patches. I can repost a v2 without patches #2 and #8. Anything else, LGTM as is. On Fri, Nov 13, 2026 at 04:12:36AM +0100, Pablo Neira Ayuso wrote: > Hi, > > The following patchset contains Netfilter updates for net-next, mostly > including flowtable fixes and improvements: > > 1) Do not offload flows to hardware from the packet path if they are > already being torn down. > > 2) Fix PMTU discovery in nf_flow_tunnel_ip6ip6_push(): update the PMTU > before skb_scrub_packet() drops the dst, otherwise the update is > silently lost. From Lorenzo Bianconi. > > 3) Fetch hardware stats when a flow is deleted, so the last stats sync > with the conntrack entry is no longer missed. Update the mtk_ppe and > airoha drivers to use this. From Lorenzo Bianconi and me. > > 4) Fix xt_time monthday computation for dates from 2040 on by using > time64_to_tm() instead of the days_since_epoch[] table. This also > shrinks the xt_time object by about 23%. From Fernando F. Mancera. > > 5) Do not offload multicast flows from nft_flow_offload, from Koral Ilgun. > > 6) Make the nf_hooks_lwtunnel sysctl read-only in non-init netns so > unprivileged users cannot toggle the global static branch. > From Pablo Neira Ayuso. > > 7) Purge the remaining flowtable entries directly when the flowtable > is destroyed instead of going through the hardware offload > workqueue, which may fail to allocate work. > > 8) Remove a no-op loop initializer in the nf_queue selftest to silence > a clang -Wunused-value warning. From Sasha Levin. > > 9) Do not scrub the skb as crossing netns in the IPIP flowtable tx > path, so the skb mark is preserved. > > 10) Do not let the DEV_PATH_TUN entry set h_source from the tunnel > device's dev_addr. Before this patch, fast-path frames went out with > part of the tunnel's local IP address as the Ethernet source MAC. > This broke offloaded flows on DS-Lite over PPPoE. From Marek > Majkowski. > > 11) Clear IPS_OFFLOAD_BIT only once, as part of the one-time teardown > transition. Before this patch, a repeated teardown of an expired > hardware flow could clear the bit owned by a replacement flow for > the same conntrack entry. From João Silva. > > 12) Call skb_ensure_writable() in nfnetlink_queue before writing > NFQA_L2HDR over the mac header. Before this patch, the write could > corrupt cloned frames that bridge flooding sends to other ports. > From Tamaki Yanagawa. > > 13) Add a UDP flowtable offload selftest for IPv4 and IPv6, with and > without masquerade. From Jiri Peska. > > Please, pull these changes from: > > git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf-next.git nf-next-26-11-13 > > Thanks. > > ---------------------------------------------------------------- > > The following changes since commit 573af801552999954caa8347e4490b04cf4763f0: > > Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net (2026-11-12 14:54:47 -0800) > > are available in the Git repository at: > > git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf-next.git nf-next-26-11-13 > > for you to fetch changes up to 4810f60b9f0e9111cd0dcbfd2ca7a7c591a57464: > > selftests: netfilter: add UDP flowtable offload test (2026-11-13 04:06:51 +0100) > > ---------------------------------------------------------------- > netfilter pull request 26-11-13 > > ---------------------------------------------------------------- > Fernando Fernandez Mancera (1): > netfilter: xt_time: fix monthday computation for dates past 2039 > > Jiri Peska (1): > selftests: netfilter: add UDP flowtable offload test > > João Silva (1): > netfilter: flowtable: do not clear the offload bit twice > > Koral Ilgun (1): > netfilter: nft_flow_offload: do not offload multicast > > Lorenzo Bianconi (2): > netfilter: flowtable: fix pmtu discovery in nf_flow_tunnel_ip6ip6_push() > net: airoha: update stats on flow deletion > > Marek Majkowski (1): > netfilter: flowtable: tunnel device poisons h_source of fast-path frames > > Pablo Neira Ayuso (6): > netfilter: flowtable: do not hw offload flows in teardown state > netfilter: flowtable: retrieve stats on flow deletion > net: ethernet: mtk_ppe: update stats on flow deletion > netfilter: lwtunnel: expose read-only sysctl nf_hooks_lwtunnel for non init-netns > netfilter: flowtable: purge entries immediately when destroying the flowtable > netfilter: flowtable: do not scrub as crossing netns > > Sasha Levin (1): > selftests: netfilter: nf_queue: drop the no-op loop initializer > > Tamaki Yanagawa (1): > netfilter: nfnetlink_queue: make skb writable before writing NFQA_L2HDR > > drivers/net/ethernet/airoha/airoha_ppe.c | 31 +-- > drivers/net/ethernet/mediatek/mtk_ppe_offload.c | 32 ++- > include/net/netfilter/nf_flow_table.h | 3 +- > net/netfilter/nf_flow_table_core.c | 35 ++- > net/netfilter/nf_flow_table_ip.c | 4 +- > net/netfilter/nf_flow_table_offload.c | 83 ++++--- > net/netfilter/nf_flow_table_path.c | 7 +- > net/netfilter/nf_hooks_lwtunnel.c | 12 +- > net/netfilter/nfnetlink_queue.c | 7 +- > net/netfilter/nft_flow_offload.c | 5 + > net/netfilter/xt_time.c | 82 +------ > tools/testing/selftests/net/netfilter/Makefile | 1 + > tools/testing/selftests/net/netfilter/nf_queue.c | 2 +- > .../selftests/net/netfilter/nft_flowtable_udp.sh | 238 +++++++++++++++++++++ > 14 files changed, 398 insertions(+), 144 deletions(-) > create mode 100755 tools/testing/selftests/net/netfilter/nft_flowtable_udp.sh >