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: <20170902153008.GA21003@lunn.ch>
Date:   Sat, 2 Sep 2017 17:30:08 +0200
From:   Andrew Lunn <andrew@...n.ch>
To:     Vivien Didelot <vivien.didelot@...oirfairelinux.com>
Cc:     netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
        kernel@...oirfairelinux.com,
        "David S. Miller" <davem@...emloft.net>,
        Florian Fainelli <f.fainelli@...il.com>
Subject: Re: [PATCH net-next 3/4] net: dsa: change dsa_ptr for a dsa_master

Hi Vivien


> diff --git a/include/net/dsa.h b/include/net/dsa.h
> index f4a5afc4255b..d5b24cd10f79 100644
> --- a/include/net/dsa.h
> +++ b/include/net/dsa.h
> @@ -130,11 +130,6 @@ struct dsa_switch_tree {
>  	 */
>  	struct dsa_platform_data	*pd;
>  
> -	/* Copy of tag_ops->rcv for faster access in hot path */
> -	struct sk_buff *	(*rcv)(struct sk_buff *skb,
> -				       struct net_device *dev,
> -				       struct packet_type *pt);
> -
>  	/*
>  	 * The switch port to which the CPU is attached.
>  	 */


> diff --git a/net/dsa/tag_dsa.c b/net/dsa/tag_dsa.c
> index fbf9ca954773..7897bbd1a110 100644
> --- a/net/dsa/tag_dsa.c
> +++ b/net/dsa/tag_dsa.c
> @@ -67,7 +67,8 @@ static struct sk_buff *dsa_xmit(struct sk_buff *skb, struct net_device *dev)
>  static struct sk_buff *dsa_rcv(struct sk_buff *skb, struct net_device *dev,
>  			       struct packet_type *pt)
>  {
> -	struct dsa_switch_tree *dst = dev->dsa_ptr;
> +	struct dsa_master *master = dev->dsa_ptr;
> +	struct dsa_switch_tree *dst = master->port->ds->dst;

This is on the hot path. Every frame received comes through here. We
have gone from one dereference to 4 dereference. That is going to
impact performance. 

How about keeping a copy of dst and ds in master?

    Andrew

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ