Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <73f88a48-72e8-e6e7-faae-1d4b92e0e13b@kernel.org>
Date:   Fri, 4 Nov 2022 08:50:13 -0600
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:91.0)
 Gecko/20100101 Thunderbird/91.13.1
Subject: Re: [PATCH] ipv6: addrlabel: fix infoleak when sending struct
 ifaddrlblmsg to network
Content-Language: en-US
To:     Alexander Potapenko <glider@google.com>
Cc:     linux-kernel@vger.kernel.org, netdev@vger.kernel.org,
        davem@davemloft.net, yoshfuji@linux-ipv6.org,
        syzbot+3553517af6020c4f2813f1003fe76ef3cbffe98d@syzkaller.appspotmail.com
References: <20221104103216.2576427-1-glider@google.com>
From:   David Ahern <dsahern@kernel.org>
In-Reply-To: <20221104103216.2576427-1-glider@google.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 11/4/22 4:32 AM, Alexander Potapenko wrote:
> When copying a `struct ifaddrlblmsg` to the network, __ifal_reserved
> remained uninitialized, resulting in a 1-byte infoleak:
> 
>   BUG: KMSAN: kernel-network-infoleak in __netdev_start_xmit ./include/linux/netdevice.h:4841
>    __netdev_start_xmit ./include/linux/netdevice.h:4841
>    netdev_start_xmit ./include/linux/netdevice.h:4857
>    xmit_one net/core/dev.c:3590
>    dev_hard_start_xmit+0x1dc/0x800 net/core/dev.c:3606
>    __dev_queue_xmit+0x17e8/0x4350 net/core/dev.c:4256
>    dev_queue_xmit ./include/linux/netdevice.h:3009
>    __netlink_deliver_tap_skb net/netlink/af_netlink.c:307
>    __netlink_deliver_tap+0x728/0xad0 net/netlink/af_netlink.c:325
>    netlink_deliver_tap net/netlink/af_netlink.c:338
>    __netlink_sendskb net/netlink/af_netlink.c:1263
>    netlink_sendskb+0x1d9/0x200 net/netlink/af_netlink.c:1272
>    netlink_unicast+0x56d/0xf50 net/netlink/af_netlink.c:1360
>    nlmsg_unicast ./include/net/netlink.h:1061
>    rtnl_unicast+0x5a/0x80 net/core/rtnetlink.c:758
>    ip6addrlbl_get+0xfad/0x10f0 net/ipv6/addrlabel.c:628
>    rtnetlink_rcv_msg+0xb33/0x1570 net/core/rtnetlink.c:6082
>   ...
>   Uninit was created at:
>    slab_post_alloc_hook+0x118/0xb00 mm/slab.h:742
>    slab_alloc_node mm/slub.c:3398
>    __kmem_cache_alloc_node+0x4f2/0x930 mm/slub.c:3437
>    __do_kmalloc_node mm/slab_common.c:954
>    __kmalloc_node_track_caller+0x117/0x3d0 mm/slab_common.c:975
>    kmalloc_reserve net/core/skbuff.c:437
>    __alloc_skb+0x27a/0xab0 net/core/skbuff.c:509
>    alloc_skb ./include/linux/skbuff.h:1267
>    nlmsg_new ./include/net/netlink.h:964
>    ip6addrlbl_get+0x490/0x10f0 net/ipv6/addrlabel.c:608
>    rtnetlink_rcv_msg+0xb33/0x1570 net/core/rtnetlink.c:6082
>    netlink_rcv_skb+0x299/0x550 net/netlink/af_netlink.c:2540
>    rtnetlink_rcv+0x26/0x30 net/core/rtnetlink.c:6109
>    netlink_unicast_kernel net/netlink/af_netlink.c:1319
>    netlink_unicast+0x9ab/0xf50 net/netlink/af_netlink.c:1345
>    netlink_sendmsg+0xebc/0x10f0 net/netlink/af_netlink.c:1921
>   ...
> 
> This patch ensures that the reserved field is always initialized.
> 
> Reported-by: syzbot+3553517af6020c4f2813f1003fe76ef3cbffe98d@syzkaller.appspotmail.com
> Fixes: 2a8cc6c89039 ("[IPV6] ADDRCONF: Support RFC3484 configurable address selection policy table.")
> Signed-off-by: Alexander Potapenko <glider@google.com>
> ---
>  net/ipv6/addrlabel.c | 1 +
>  1 file changed, 1 insertion(+)
> 

Reviewed-by: David Ahern <dsahern@kernel.org>