From: Eric Dumazet <edumazet@google.com>
To: Guoyu Su <yss2813483011xxl@gmail.com>
Cc: davem@davemloft.net, kuba@kernel.org, pabeni@redhat.com,
netdev@vger.kernel.org, horms@kernel.org,
linux-kernel@vger.kernel.org, syzkaller-bugs@googlegroups.com,
syzbot+1543a7d954d9c6d00407@syzkaller.appspotmail.com
Subject: Re: [PATCH net] net: clear mangleid_features for SKB_GSO_DODGY TCPv4
Date: Sat, 7 Mar 2026 17:43:28 +0100 [thread overview]
Message-ID: <CANn89iLYGsCr6i+rP13nq8H8H1pT6h=hGDn+kds0js6v4gr6Mg@mail.gmail.com> (raw)
In-Reply-To: <20260307162905.3697050-1-yss2813483011xxl@gmail.com>
On Sat, Mar 7, 2026 at 5:29 PM Guoyu Su <yss2813483011xxl@gmail.com> wrote:
>
> Syzbot reported a KMSAN uninit-value warning in netif_skb_features() [1], which originates from gso_features_check():
>
> BUG: KMSAN: uninit-value in gso_features_check net/core/dev.c:3804 [inline]
> BUG: KMSAN: uninit-value in netif_skb_features+0x6fb/0x1870 net/core/dev.c:3828
> gso_features_check net/core/dev.c:3804 [inline]
> netif_skb_features+0x6fb/0x1870 net/core/dev.c:3828
> validate_xmit_skb+0xb6/0x2400 net/core/dev.c:4003
> ...
> __dev_queue_xmit+0x3016/0x5a50 net/core/dev.c:4795
> packet_snd net/packet/af_packet.c:3077 [inline]
> packet_sendmsg+0x91d9/0xa320 net/packet/af_packet.c:3109
>
> SKB_GSO_DODGY marks packets whose GSO-related header metadata should be treated as untrusted (for example, packets injected via AF_PACKET).
>
> gso_features_check() checks IP_DF through iph->frag_off for SKB_GSO_TCPV4 packets. For DODGY packets with untrusted metadata, this may result in a KMSAN uninit-value report.
>
> Handle such packets conservatively by clearing mangleid_features when SKB_GSO_DODGY is set, without relying on iph->frag_off.
>
> [1] https://syzkaller.appspot.com/bug?extid=1543a7d954d9c6d00407
>
> Reported-by: syzbot+1543a7d954d9c6d00407@syzkaller.appspotmail.com
> Tested-by: syzbot+1543a7d954d9c6d00407@syzkaller.appspotmail.com
> Closes: https://lore.kernel.org/all/68e4b5d8.050a0220.256323.0018.GAE@google.com/T/
> Signed-off-by: Guoyu Su <yss2813483011xxl@gmail.com>
Missing FIxes: tag. Please help us.
> ---
> net/core/dev.c | 7 ++++++-
> 1 file changed, 6 insertions(+), 1 deletion(-)
>
> diff --git a/net/core/dev.c b/net/core/dev.c
> index 14a83f2035b9..9b19154b6f7f 100644
> --- a/net/core/dev.c
> +++ b/net/core/dev.c
> @@ -3808,7 +3808,12 @@ static netdev_features_t gso_features_check(const struct sk_buff *skb,
> struct iphdr *iph = skb->encapsulation ?
> inner_ip_hdr(skb) : ip_hdr(skb);
>
> - if (!(iph->frag_off & htons(IP_DF)))
> + /* SKB_GSO_DODGY packets carry untrusted L3/L4 header metadata.
> + * Avoid dereferencing IPv4 header fields and conservatively
> + * clear mangleid support in that case.
> + */
> + if ((skb_shinfo(skb)->gso_type & SKB_GSO_DODGY) ||
> + !(iph->frag_off & htons(IP_DF)))
> features &= ~dev->mangleid_features;
> }
You are not really fixing the bug, you hide it with another.
Please take a look at qdisc_pkt_len_segs_init() for a hint.
next prev parent reply other threads:[~2026-03-07 16:43 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-03-07 16:29 [PATCH net] net: clear mangleid_features for SKB_GSO_DODGY TCPv4 Guoyu Su
2026-03-07 16:43 ` Eric Dumazet [this message]
2026-03-08 8:33 ` [PATCH net v2] net: use skb_header_pointer() in gso_features_check() for TCPv4 GSO Guoyu Su
2026-03-11 0:48 ` Jakub Kicinski
2026-03-12 10:43 ` [PATCH net v3] net: use skb_header_pointer() only for DODGY TCPv4 GSO skbs Guoyu Su
2026-03-17 10:22 ` Paolo Abeni
2026-03-19 0:54 ` [PATCH net v4] " Guoyu Su
2026-03-19 13:17 ` Willem de Bruijn
2026-03-20 14:14 ` [PATCH net v5] " Guoyu Su
2026-03-20 19:24 ` Willem de Bruijn
2026-03-21 1:36 ` Willem de Bruijn
2026-03-21 15:31 ` Scars
2026-03-21 20:58 ` Willem de Bruijn
2026-03-22 4:26 ` Guoyu Su
2026-03-23 3:36 ` Willem de Bruijn
2026-03-24 10:40 ` Guoyu Su
2026-03-26 3:12 ` Willem de Bruijn
2026-03-26 12:18 ` [PATCH net v6] net: use skb_header_pointer() for TCPv4 GSO frag_off Guoyu Su
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CANn89iLYGsCr6i+rP13nq8H8H1pT6h=hGDn+kds0js6v4gr6Mg@mail.gmail.com' \
--to=edumazet@google.com \
--cc=davem@davemloft.net \
--cc=horms@kernel.org \
--cc=kuba@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=syzbot+1543a7d954d9c6d00407@syzkaller.appspotmail.com \
--cc=syzkaller-bugs@googlegroups.com \
--cc=yss2813483011xxl@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox