linux-kernel - Re: [PATCH net-next v2 04/12] net: dsa: tag_sja1105: absorb entire sja1105_vlan_rcv() into dsa_8021q

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]

Date: Thu, 20 Jun 2024 15:28:16 +0300
From: Vladimir Oltean <olteanv@...il.com>
To: Pawel Dembicki <paweldembicki@...il.com>
Cc: netdev@...r.kernel.org, Paolo Abeni <pabeni@...hat.com>,
	Vladimir Oltean <vladimir.oltean@....com>,
	Andrew Lunn <andrew@...n.ch>,
	Florian Fainelli <f.fainelli@...il.com>,
	"David S. Miller" <davem@...emloft.net>,
	Eric Dumazet <edumazet@...gle.com>,
	Jakub Kicinski <kuba@...nel.org>,
	Claudiu Manoil <claudiu.manoil@....com>,
	Alexandre Belloni <alexandre.belloni@...tlin.com>,
	UNGLinuxDriver@...rochip.com, Russell King <linux@...linux.org.uk>,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH net-next v2 04/12] net: dsa: tag_sja1105: absorb entire
 sja1105_vlan_rcv() into dsa_8021q_rcv()

On Wed, Jun 19, 2024 at 10:52:10PM +0200, Pawel Dembicki wrote:
> From: Vladimir Oltean <vladimir.oltean@....com>
> 
> tag_sja1105 has a wrapper over dsa_8021q_rcv(): sja1105_vlan_rcv(),
> which determines whether the packet came from a bridge with
> vlan_filtering=1 (the case resolved via
> dsa_find_designated_bridge_port_by_vid()), or if it contains a tag_8021q
> header.
> 
> Looking at a new tagger implementation for vsc73xx, based also on
> tag_8021q, it is becoming clear that the logic is needed there as well.
> So instead of forcing each tagger to wrap around dsa_8021q_rcv(), let's
> merge the logic into the core.
> 
> Signed-off-by: Vladimir Oltean <vladimir.oltean@....com>
> Signed-off-by: Pawel Dembicki <paweldembicki@...il.com>
> ---
> v2,v1:
>   - resend only
> ---
> Before patch series split:
> https://patchwork.kernel.org/project/netdevbpf/list/?series=841034&state=%2A&archive=both
> v8, v7, v6:
>   - resend only
> v5:
>   - add missing SoB
> v4:
>   - introduced patch
> ---
>  net/dsa/tag_8021q.c        | 34 ++++++++++++++++++++++++++++------
>  net/dsa/tag_8021q.h        |  2 +-
>  net/dsa/tag_ocelot_8021q.c |  2 +-
>  net/dsa/tag_sja1105.c      | 32 ++++----------------------------
>  4 files changed, 34 insertions(+), 36 deletions(-)
> 
> diff --git a/net/dsa/tag_8021q.c b/net/dsa/tag_8021q.c
> index 3cb0293793a5..332b0ae02645 100644
> --- a/net/dsa/tag_8021q.c
> +++ b/net/dsa/tag_8021q.c
> @@ -507,27 +507,39 @@ EXPORT_SYMBOL_GPL(dsa_tag_8021q_find_port_by_vbid);
>   * @vbid: pointer to storage for imprecise bridge ID. Must be pre-initialized
>   *	with -1. If a positive value is returned, the source_port and switch_id
>   *	are invalid.
> + * @vid: pointer to storage for original VID, in case tag_8021q decoding failed.
> + *
> + * If the packet has a tag_8021q header, decode it and set @source_port,
> + * @switch_id and @vbid, and strip the header. Otherwise set @vid and keep the
> + * header in the hwaccel area of the packet.
>   */
>  void dsa_8021q_rcv(struct sk_buff *skb, int *source_port, int *switch_id,
> -		   int *vbid)
> +		   int *vbid, int *vid)
>  {
>  	int tmp_source_port, tmp_switch_id, tmp_vbid;
> -	u16 vid, tci;
> +	__be16 vlan_proto;
> +	u16 tmp_vid, tci;
>  
>  	if (skb_vlan_tag_present(skb)) {
> +		vlan_proto = skb->vlan_proto;
>  		tci = skb_vlan_tag_get(skb);
>  		__vlan_hwaccel_clear_tag(skb);
>  	} else {
> +		struct vlan_ethhdr *hdr = vlan_eth_hdr(skb);
> +
> +		vlan_proto = hdr->h_vlan_proto;
>  		skb_push_rcsum(skb, ETH_HLEN);
>  		__skb_vlan_pop(skb, &tci);
>  		skb_pull_rcsum(skb, ETH_HLEN);
>  	}
>  
> -	vid = tci & VLAN_VID_MASK;
> +	tmp_vid = tci & VLAN_VID_MASK;
> +	if (!vid_is_dsa_8021q(tmp_vid))
> +		goto not_tag_8021q;

I think this may be more clearly expressed linearly, without a goto.

	if (!vid_is_dsa_8021q(tmp_vid)) {
		/* Not a tag_8021q frame, so return the VID to the
		 * caller for further processing, and put the tag back
		 */
		if (vid)
			*vid = tmp_vid;

		__vlan_hwaccel_put_tag(skb, vlan_proto, tci);

		return;
	}

>  
> -	tmp_source_port = dsa_8021q_rx_source_port(vid);
> -	tmp_switch_id = dsa_8021q_rx_switch_id(vid);
> -	tmp_vbid = dsa_tag_8021q_rx_vbid(vid);
> +	tmp_source_port = dsa_8021q_rx_source_port(tmp_vid);
> +	tmp_switch_id = dsa_8021q_rx_switch_id(tmp_vid);
> +	tmp_vbid = dsa_tag_8021q_rx_vbid(tmp_vid);
>  
>  	/* Precise source port information is unknown when receiving from a
>  	 * VLAN-unaware bridging domain, and tmp_source_port and tmp_switch_id
> @@ -546,5 +558,15 @@ void dsa_8021q_rcv(struct sk_buff *skb, int *source_port, int *switch_id,
>  		*vbid = tmp_vbid;
>  
>  	skb->priority = (tci & VLAN_PRIO_MASK) >> VLAN_PRIO_SHIFT;
> +	return;
> +
> +not_tag_8021q:
> +	if (vid)
> +		*vid = tmp_vid;
> +	if (vbid)
> +		*vbid = -1;

Thinking more about it, I don't think this is needed (hence it is
missing in my rewritten snippet above). I mean, *vbid is already
initialized with -1 (it's a requirement also specified in the
kernel-doc) and we haven't changed it.

> +
> +	/* Put the tag back */
> +	__vlan_hwaccel_put_tag(skb, vlan_proto, tci);
>  }
>  EXPORT_SYMBOL_GPL(dsa_8021q_rcv);
> diff --git a/net/dsa/tag_8021q.h b/net/dsa/tag_8021q.h
> index 41f7167ac520..0c6671d7c1c2 100644
> --- a/net/dsa/tag_8021q.h
> +++ b/net/dsa/tag_8021q.h
> @@ -14,7 +14,7 @@ struct sk_buff *dsa_8021q_xmit(struct sk_buff *skb, struct net_device *netdev,
>  			       u16 tpid, u16 tci);
>  
>  void dsa_8021q_rcv(struct sk_buff *skb, int *source_port, int *switch_id,
> -		   int *vbid);
> +		   int *vbid, int *vid);
>  
>  struct net_device *dsa_tag_8021q_find_port_by_vbid(struct net_device *conduit,
>  						   int vbid);