netdev - Re: [PATCH net-next 06/10] net: vxlan: add skb drop reasons to vxlan

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: <20240816192243.050d0b1f@kernel.org>
Date: Fri, 16 Aug 2024 19:22:43 -0700
From: Jakub Kicinski <kuba@...nel.org>
To: Menglong Dong <menglong8.dong@...il.com>
Cc: davem@...emloft.net, edumazet@...gle.com, pabeni@...hat.com,
 dsahern@...nel.org, dongml2@...natelecom.cn, idosch@...dia.com,
 amcohen@...dia.com, gnault@...hat.com, bpoirier@...dia.com,
 b.galvani@...il.com, razor@...ckwall.org, petrm@...dia.com,
 linux-kernel@...r.kernel.org, netdev@...r.kernel.org
Subject: Re: [PATCH net-next 06/10] net: vxlan: add skb drop reasons to
 vxlan_rcv()

On Thu, 15 Aug 2024 20:42:58 +0800 Menglong Dong wrote:
>  #define VXLAN_DROP_REASONS(R)			\
> +	R(VXLAN_DROP_FLAGS)			\
> +	R(VXLAN_DROP_VNI)			\
> +	R(VXLAN_DROP_MAC)			\

Drop reasons should be documented.
I don't think name of a header field is a great fit for a reason.

>  	/* deliberate comment for trailing \ */
>  
>  enum vxlan_drop_reason {
> diff --git a/drivers/net/vxlan/vxlan_core.c b/drivers/net/vxlan/vxlan_core.c
> index e971c4785962..9a61f04bb95d 100644
> --- a/drivers/net/vxlan/vxlan_core.c
> +++ b/drivers/net/vxlan/vxlan_core.c
> @@ -1668,6 +1668,7 @@ static bool vxlan_ecn_decapsulate(struct vxlan_sock *vs, void *oiph,
>  /* Callback from net/ipv4/udp.c to receive packets */
>  static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
>  {
> +	enum skb_drop_reason reason = pskb_may_pull_reason(skb, VXLAN_HLEN);

Do not call complex functions inline as variable init..

>  	struct vxlan_vni_node *vninode = NULL;
>  	struct vxlan_dev *vxlan;
>  	struct vxlan_sock *vs;
> @@ -1681,7 +1682,7 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
>  	int nh;
>  
>  	/* Need UDP and VXLAN header to be present */
> -	if (!pskb_may_pull(skb, VXLAN_HLEN))
> +	if (reason != SKB_NOT_DROPPED_YET)

please don't compare against "not dropped yet", just:

	if (reason)

> @@ -1815,8 +1831,9 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
>  	return 0;
>  
>  drop:
> +	SKB_DR_RESET(reason);

the name of this macro is very confusing, I don't think it should exist
in the first place. nothing should goto drop without initialing reason

>  	/* Consume bad packet */
> -	kfree_skb(skb);
> +	kfree_skb_reason(skb, reason);
>  	return 0;
>  }