linux-kernel - Re: [PATCH] net: fix segmentation of forwarding fraglist GRO

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <aab6c515-12e4-48ca-8220-c0797dae781f@redhat.com>
Date: Tue, 23 Dec 2025 15:12:58 +0100
From: Paolo Abeni <pabeni@...hat.com>
To: Jibin Zhang <jibin.zhang@...iatek.com>, Eric Dumazet
 <edumazet@...gle.com>, Neal Cardwell <ncardwell@...gle.com>,
 Kuniyuki Iwashima <kuniyu@...gle.com>, "David S . Miller"
 <davem@...emloft.net>, David Ahern <dsahern@...nel.org>,
 Jakub Kicinski <kuba@...nel.org>, Simon Horman <horms@...nel.org>,
 Matthias Brugger <matthias.bgg@...il.com>,
 AngeloGioacchino Del Regno <angelogioacchino.delregno@...labora.com>,
 netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
 linux-arm-kernel@...ts.infradead.org, linux-mediatek@...ts.infradead.org
Cc: wsd_upstream@...iatek.com, shiming.cheng@...iatek.com,
 defa.li@...iatek.com
Subject: Re: [PATCH] net: fix segmentation of forwarding fraglist GRO

On 12/17/25 4:55 AM, Jibin Zhang wrote:
> This patch enhances GSO segment checks by verifying the presence
> of frag_list and protocol consistency, addressing low throughput
> issues on IPv4 servers when used as hotspots
> 
> Specifically, it fixes a bug in GSO segmentation when forwarding
> GRO packets with frag_list. The function skb_segment_list cannot
> correctly process GRO skbs converted by XLAT, because XLAT only
> converts the header of the head skb. As a result, skbs in the
> frag_list may remain unconverted, leading to protocol
> inconsistencies and reduced throughput.
> 
> To resolve this, the patch uses skb_segment to handle forwarded
> packets converted by XLAT, ensuring that all fragments are
> properly converted and segmented.
> 
> Signed-off-by: Jibin Zhang <jibin.zhang@...iatek.com>

This looks like a fix, it should target the 'net' tree and include a
suitable Fixes tag.

> ---
>  net/ipv4/tcp_offload.c | 3 ++-
>  net/ipv4/udp_offload.c | 3 ++-
>  2 files changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/net/ipv4/tcp_offload.c b/net/ipv4/tcp_offload.c
> index fdda18b1abda..162a384a15bb 100644
> --- a/net/ipv4/tcp_offload.c
> +++ b/net/ipv4/tcp_offload.c
> @@ -104,7 +104,8 @@ static struct sk_buff *tcp4_gso_segment(struct sk_buff *skb,
>  	if (!pskb_may_pull(skb, sizeof(struct tcphdr)))
>  		return ERR_PTR(-EINVAL);
>  
> -	if (skb_shinfo(skb)->gso_type & SKB_GSO_FRAGLIST) {
> +	if ((skb_shinfo(skb)->gso_type & SKB_GSO_FRAGLIST) && skb_has_frag_list(skb) &&
> +	    (skb->protocol == skb_shinfo(skb)->frag_list->protocol)) {
>  		struct tcphdr *th = tcp_hdr(skb);
>  
>  		if (skb_pagelen(skb) - th->doff * 4 == skb_shinfo(skb)->gso_size)
> diff --git a/net/ipv4/udp_offload.c b/net/ipv4/udp_offload.c
> index 19d0b5b09ffa..704fb32d10d7 100644
> --- a/net/ipv4/udp_offload.c
> +++ b/net/ipv4/udp_offload.c
> @@ -512,7 +512,8 @@ struct sk_buff *__udp_gso_segment(struct sk_buff *gso_skb,
>  		return NULL;
>  	}
>  
> -	if (skb_shinfo(gso_skb)->gso_type & SKB_GSO_FRAGLIST) {
> +	if ((skb_shinfo(gso_skb)->gso_type & SKB_GSO_FRAGLIST) && skb_has_frag_list(gso_skb) &&
> +	    (gso_skb->protocol == skb_shinfo(gso_skb)->frag_list->protocol)) {
>  		 /* Detect modified geometry and pass those to skb_segment. */
>  		if (skb_pagelen(gso_skb) - sizeof(*uh) == skb_shinfo(gso_skb)->gso_size)
>  			return __udp_gso_segment_list(gso_skb, features, is_ipv6);

I guess checks should be needed for ipv6.

Also it looks like this skips the CSUM_PARTIAL preparation, and possibly
break csum offload.

Additionally I don't like the ever increasing stack of hacks needed to
let GSO_FRAGLIST operate in the most diverse setups, the simpler fix
would be disabling such aggregation.

/P