From: Leon Romanovsky <leon@kernel.org>
To: Christian Langrock <christian.langrock@secunet.com>
Cc: Steffen Klassert <steffen.klassert@secunet.com>,
herbert@gondor.apana.org.au, davem@davemloft.net,
edumazet@google.com, kuba@kernel.org, pabeni@redhat.com,
netdev@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH net-ipsec v2] xfrm: replay: Fix ESN wrap around for GSO
Date: Wed, 28 Sep 2022 09:18:07 +0300 [thread overview]
Message-ID: <YzPnH/JrIFbPteR3@unreal> (raw)
In-Reply-To: <fe554921-104e-2365-a09b-812a1cedac65@secunet.com>
On Tue, Sep 27, 2022 at 02:59:50PM +0200, Christian Langrock wrote:
> When using GSO it can happen that the wrong seq_hi is used for the last
> packets before the wrap around. This can lead to double usage of a
> sequence number. To avoid this, we should serialize this last GSO
> packet.
>
> Fixes: d7dbefc45cf5 ("xfrm: Add xfrm_replay_overflow functions for...")
> Signed-off-by: Christian Langrock <christian.langrock@secunet.com>
> ---
> include/net/xfrm.h | 1 +
> net/xfrm/xfrm_output.c | 2 +-
> net/xfrm/xfrm_replay.c | 33 +++++++++++++++++++++++++++++++++
> 3 files changed, 35 insertions(+), 1 deletion(-)
>
> diff --git a/include/net/xfrm.h b/include/net/xfrm.h
> index 6e8fa98f786f..49d6d974f493 100644
> --- a/include/net/xfrm.h
> +++ b/include/net/xfrm.h
> @@ -1749,6 +1749,7 @@ void xfrm_replay_advance(struct xfrm_state *x, __be32 net_seq);
> int xfrm_replay_check(struct xfrm_state *x, struct sk_buff *skb, __be32 net_seq);
> void xfrm_replay_notify(struct xfrm_state *x, int event);
> int xfrm_replay_overflow(struct xfrm_state *x, struct sk_buff *skb);
> +int xfrm_replay_overflow_check(struct xfrm_state *x, struct sk_buff *skb);
> int xfrm_replay_recheck(struct xfrm_state *x, struct sk_buff *skb, __be32 net_seq);
>
> static inline int xfrm_aevent_is_on(struct net *net)
> diff --git a/net/xfrm/xfrm_output.c b/net/xfrm/xfrm_output.c
> index 9a5e79a38c67..c470a68d9c88 100644
> --- a/net/xfrm/xfrm_output.c
> +++ b/net/xfrm/xfrm_output.c
> @@ -738,7 +738,7 @@ int xfrm_output(struct sock *sk, struct sk_buff *skb)
> skb->encapsulation = 1;
>
> if (skb_is_gso(skb)) {
> - if (skb->inner_protocol)
> + if (skb->inner_protocol || xfrm_replay_overflow_check(x, skb))
> return xfrm_output_gso(net, sk, skb);
>
> skb_shinfo(skb)->gso_type |= SKB_GSO_ESP;
> diff --git a/net/xfrm/xfrm_replay.c b/net/xfrm/xfrm_replay.c
> index 9277d81b344c..991cfc7a091d 100644
> --- a/net/xfrm/xfrm_replay.c
> +++ b/net/xfrm/xfrm_replay.c
> @@ -750,6 +750,34 @@ int xfrm_replay_overflow(struct xfrm_state *x, struct sk_buff *skb)
>
> return xfrm_replay_overflow_offload(x, skb);
> }
> +
> +static bool xfrm_replay_overflow_check_offload_esn(struct xfrm_state *x, struct sk_buff *skb)
> +{
> + struct xfrm_replay_state_esn *replay_esn = x->replay_esn;
> + __u32 oseq = replay_esn->oseq;
> + bool ret = false;
> +
> + /* We assume that this function is called with
> + * skb_is_gso(skb) == true
> + */
> +
> + if (x->type->flags & XFRM_TYPE_REPLAY_PROT) {
> + oseq = oseq + 1 + skb_shinfo(skb)->gso_segs;
> + if (unlikely(oseq < replay_esn->oseq))
> + ret = true;
return true;
> + }
> +
> + return ret;
return false;
> +}
> +
> +bool xfrm_replay_overflow_check(struct xfrm_state *x, struct sk_buff *skb)
> +{
> + if (x->repl_mode == XFRM_REPLAY_MODE_ESN)
> + return xfrm_replay_overflow_check_offload_esn(x, skb);
> +
> + return false;
> +}
I still think that functions above should be merged into one. This is
called only if xfrm_dev_offload_ok() success -> in crypto offload path.
Thanks
> +
> #else
> int xfrm_replay_overflow(struct xfrm_state *x, struct sk_buff *skb)
> {
> @@ -764,6 +792,11 @@ int xfrm_replay_overflow(struct xfrm_state *x, struct sk_buff *skb)
>
> return __xfrm_replay_overflow(x, skb);
> }
> +
> +int xfrm_replay_overflow_check(struct xfrm_state *x, struct sk_buff *skb)
int -> bool
> +{
> + return 0;
> +}
> #endif
>
> int xfrm_init_replay(struct xfrm_state *x)
> --
> 2.37.1.223.g6a475b71f8
>
next prev parent reply other threads:[~2022-09-28 6:18 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-09-27 12:59 [PATCH net-ipsec v2] xfrm: replay: Fix ESN wrap around for GSO Christian Langrock
2022-09-27 14:22 ` Jakub Kicinski
2022-09-28 6:18 ` Leon Romanovsky [this message]
2022-09-28 6:49 ` Steffen Klassert
2022-09-28 21:28 ` kernel test robot
2022-09-28 21:29 ` kernel test robot
-- strict thread matches above, loose matches on Subject: below --
2022-09-27 12:03 Christian Langrock
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=YzPnH/JrIFbPteR3@unreal \
--to=leon@kernel.org \
--cc=christian.langrock@secunet.com \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=herbert@gondor.apana.org.au \
--cc=kuba@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=steffen.klassert@secunet.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;
as well as URLs for NNTP newsgroup(s).