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 for Android: free password hash cracker in your pocket
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20161026155201.ooeuhfs4bxcbgkjp@splinter>
Date:   Wed, 26 Oct 2016 18:52:01 +0300
From:   Ido Schimmel <idosch@...sch.org>
To:     David Ahern <dsa@...ulusnetworks.com>
Cc:     davem@...emloft.net, netdev@...r.kernel.org, mlxsw@...lanox.com,
        Ido Schimmel <idosch@...lanox.com>
Subject: Re: [PATCH net-next] net: core: Traverse the adjacency list from
 first entry

On Wed, Oct 26, 2016 at 09:05:35AM -0600, David Ahern wrote:
> On 10/26/16 12:39 AM, idosch@...sch.org wrote:
> > From: Ido Schimmel <idosch@...lanox.com>
> > 
> > netdev_next_lower_dev() returns NULL when we finished traversing the
> > adjacency list ('iter' points to the list's head). Therefore, we must
> > start traversing the list from the first entry and not its head.
> > 
> > Fixes: 1a3f060c1a47 ("net: Introduce new api for walking upper and lower devices")
> > Signed-off-by: Ido Schimmel <idosch@...lanox.com>
> > ---
> >  net/core/dev.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/net/core/dev.c b/net/core/dev.c
> > index f55fb45..d9c937f 100644
> > --- a/net/core/dev.c
> > +++ b/net/core/dev.c
> > @@ -5419,7 +5419,7 @@ int netdev_walk_all_lower_dev(struct net_device *dev,
> >  	struct list_head *iter;
> >  	int ret;
> >  
> > -	for (iter = &dev->adj_list.lower,
> > +	for (iter = dev->adj_list.lower.next,
> >  	     ldev = netdev_next_lower_dev(dev, &iter);
> >  	     ldev;
> >  	     ldev = netdev_next_lower_dev(dev, &iter)) {
> > 
> 
> How about this instead? It keeps the 3 walk functions in sync modulo the rcu reference:

I don't see any problem, so I guess it will work as well. I simply
preferred to use the same convention employed prior to your patchset.

Please submit yours formally, if you prefer.

> 
> 
> diff --git a/net/core/dev.c b/net/core/dev.c
> index f55fb4536016..6aa43cd8cbb5 100644
> --- a/net/core/dev.c
> +++ b/net/core/dev.c
> @@ -5400,12 +5400,12 @@ static struct net_device *netdev_next_lower_dev(struct net_device *dev,
>  {
> 	struct netdev_adjacent *lower;
> 
> -	lower = list_entry(*iter, struct netdev_adjacent, list);
> +	lower = list_entry((*iter)->next, struct netdev_adjacent, list);
> 
>  	if (&lower->list == &dev->adj_list.lower)
>  		return NULL;
> 
> -	*iter = lower->list.next;
> +	*iter = &lower->list;
> 
>  	return lower->dev;
>  }

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ