| Message ID | 20231018085118.10829-2-fw@strlen.de (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | e15e5027106f3f6009d2fb46b3a1bb3d9e6a1b77 |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [net-next,1/7] netfilter: xt_mangle: only check verdict part of return value | expand |
| Context | Check | Description |
|---|---|---|
| netdev/tree_selection | success | Clearly marked for net-next |
| netdev/apply | success | Patch already applied to net-next |
Hello: This series was applied to netdev/net-next.git (main) by Florian Westphal <fw@strlen.de>: On Wed, 18 Oct 2023 10:51:05 +0200 you wrote: > These checks assume that the caller only returns NF_DROP without > any errno embedded in the upper bits. > > This is fine right now, but followup patches will start to propagate > such errors to allow kfree_skb_drop_reason() in the called functions, > those would then indicate 'errno << 8 | NF_STOLEN'. > > [...] Here is the summary with links: - [net-next,1/7] netfilter: xt_mangle: only check verdict part of return value https://git.kernel.org/netdev/net-next/c/e15e5027106f - [net-next,2/7] netfilter: nf_tables: mask out non-verdict bits when checking return value https://git.kernel.org/netdev/net-next/c/4d26ab0086aa - [net-next,3/7] netfilter: conntrack: convert nf_conntrack_update to netfilter verdicts https://git.kernel.org/netdev/net-next/c/6291b3a67ad5 - [net-next,4/7] netfilter: nf_nat: mask out non-verdict bits when checking return value https://git.kernel.org/netdev/net-next/c/35c038b0a4be - [net-next,5/7] netfilter: make nftables drops visible in net dropmonitor https://git.kernel.org/netdev/net-next/c/e0d4593140b0 - [net-next,6/7] netfilter: bridge: convert br_netfilter to NF_DROP_REASON https://git.kernel.org/netdev/net-next/c/cf8b7c1a5be7 - [net-next,7/7] netfilter: nf_tables: de-constify set commit ops function argument https://git.kernel.org/netdev/net-next/c/256001672153 You are awesome, thank you!
diff --git a/net/ipv4/netfilter/iptable_mangle.c b/net/ipv4/netfilter/iptable_mangle.c index 3abb430af9e6..385d945d8ebe 100644 --- a/net/ipv4/netfilter/iptable_mangle.c +++ b/net/ipv4/netfilter/iptable_mangle.c @@ -36,12 +36,12 @@ static const struct xt_table packet_mangler = { static unsigned int ipt_mangle_out(void *priv, struct sk_buff *skb, const struct nf_hook_state *state) { - unsigned int ret; + unsigned int ret, verdict; const struct iphdr *iph; - u_int8_t tos; __be32 saddr, daddr; - u_int32_t mark; + u32 mark; int err; + u8 tos; /* Save things which could affect route */ mark = skb->mark; @@ -51,8 +51,9 @@ ipt_mangle_out(void *priv, struct sk_buff *skb, const struct nf_hook_state *stat tos = iph->tos; ret = ipt_do_table(priv, skb, state); + verdict = ret & NF_VERDICT_MASK; /* Reroute for ANY change. */ - if (ret != NF_DROP && ret != NF_STOLEN) { + if (verdict != NF_DROP && verdict != NF_STOLEN) { iph = ip_hdr(skb); if (iph->saddr != saddr || diff --git a/net/ipv6/netfilter/ip6table_mangle.c b/net/ipv6/netfilter/ip6table_mangle.c index a88b2ce4a3cb..8dd4cd0c47bd 100644 --- a/net/ipv6/netfilter/ip6table_mangle.c +++ b/net/ipv6/netfilter/ip6table_mangle.c @@ -31,10 +31,10 @@ static const struct xt_table packet_mangler = { static unsigned int ip6t_mangle_out(void *priv, struct sk_buff *skb, const struct nf_hook_state *state) { - unsigned int ret; struct in6_addr saddr, daddr; - u_int8_t hop_limit; - u_int32_t flowlabel, mark; + unsigned int ret, verdict; + u32 flowlabel, mark; + u8 hop_limit; int err; /* save source/dest address, mark, hoplimit, flowlabel, priority, */ @@ -47,8 +47,9 @@ ip6t_mangle_out(void *priv, struct sk_buff *skb, const struct nf_hook_state *sta flowlabel = *((u_int32_t *)ipv6_hdr(skb)); ret = ip6t_do_table(priv, skb, state); + verdict = ret & NF_VERDICT_MASK; - if (ret != NF_DROP && ret != NF_STOLEN && + if (verdict != NF_DROP && verdict != NF_STOLEN && (!ipv6_addr_equal(&ipv6_hdr(skb)->saddr, &saddr) || !ipv6_addr_equal(&ipv6_hdr(skb)->daddr, &daddr) || skb->mark != mark ||
These checks assume that the caller only returns NF_DROP without any errno embedded in the upper bits. This is fine right now, but followup patches will start to propagate such errors to allow kfree_skb_drop_reason() in the called functions, those would then indicate 'errno << 8 | NF_STOLEN'. To not break things we have to mask those parts out. Signed-off-by: Florian Westphal <fw@strlen.de> --- net/ipv4/netfilter/iptable_mangle.c | 9 +++++---- net/ipv6/netfilter/ip6table_mangle.c | 9 +++++---- 2 files changed, 10 insertions(+), 8 deletions(-)