netdev - Re: [PATCH net] net: bridge: fix uninitialized variables when BRIDGE

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] [thread-next>] [day] [month] [year] [list]

Message-ID: <a9b7d8d2-da03-42f0-bd16-3446cdcaecc8@nvidia.com>
Date:   Wed, 27 Oct 2021 16:54:39 +0300
From:   Nikolay Aleksandrov <nikolay@...dia.com>
To:     Ivan Vecera <ivecera@...hat.com>, netdev@...r.kernel.org
Cc:     Roopa Prabhu <roopa@...dia.com>,
        "David S. Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>,
        Horatiu Vultur <horatiu.vultur@...rochip.com>,
        Henrik Bjoernlund <henrik.bjoernlund@...rochip.com>,
        "moderated list:ETHERNET BRIDGE" <bridge@...ts.linux-foundation.org>,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH net] net: bridge: fix uninitialized variables when
 BRIDGE_CFM is disabled

On 27/10/2021 16:49, Ivan Vecera wrote:
> Function br_get_link_af_size_filtered() calls br_cfm_{,peer}_mep_count()
> but does not check their return value. When BRIDGE_CFM is not enabled
> these functions return -EOPNOTSUPP but do not modify count parameter.
> Calling function then works with uninitialized variables.
> 
> Fixes: b6d0425b816e ("bridge: cfm: Netlink Notifications.")
> Signed-off-by: Ivan Vecera <ivecera@...hat.com>
> ---
>  net/bridge/br_netlink.c | 6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/net/bridge/br_netlink.c b/net/bridge/br_netlink.c
> index 5c6c4305ed23..12d602495ea0 100644
> --- a/net/bridge/br_netlink.c
> +++ b/net/bridge/br_netlink.c
> @@ -126,8 +126,10 @@ static size_t br_get_link_af_size_filtered(const struct net_device *dev,
>  		return vinfo_sz;
>  
>  	/* CFM status info must be added */
> -	br_cfm_mep_count(br, &num_cfm_mep_infos);
> -	br_cfm_peer_mep_count(br, &num_cfm_peer_mep_infos);
> +	if (br_cfm_mep_count(br, &num_cfm_mep_infos) < 0)
> +		num_cfm_mep_infos = 0;
> +	if (br_cfm_peer_mep_count(br, &num_cfm_peer_mep_infos) < 0)
> +		num_cfm_peer_mep_infos = 0;
>  
>  	vinfo_sz += nla_total_size(0);	/* IFLA_BRIDGE_CFM */
>  	/* For each status struct the MEP instance (u32) is added */
> 

Hi,
Could you please rather update the EOPNOTSUPP helpers to set these infos to 0 before
returning? Someone else might decide to use them and hit the same bug.

E.g.
static inline int br_cfm_mep_count(struct net_bridge *br, u32 *count)
{
	*count = 0;
        return -EOPNOTSUPP;
}

We already do the same for br_allowed_ingress, nbp_vlan_add() etc.

Thanks,
 Nik