netdev - Re: [PATCH bpf-next v2] net: Don't allow to attach xdp if bond slave device's upper already has a program

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: <ZsiOxkd5KbbIIB6k@nanopsycho.orion>
Date: Fri, 23 Aug 2024 15:29:42 +0200
From: Jiri Pirko <jiri@...nulli.us>
To: Daniel Borkmann <daniel@...earbox.net>
Cc: Feng zhou <zhoufeng.zf@...edance.com>, davem@...emloft.net,
	edumazet@...gle.com, kuba@...nel.org, pabeni@...hat.com,
	ast@...nel.org, hawk@...nel.org, john.fastabend@...il.com,
	bigeasy@...utronix.de, lorenzo@...nel.org, netdev@...r.kernel.org,
	linux-kernel@...r.kernel.org, bpf@...r.kernel.org,
	yangzhenze@...edance.com, wangdongdong.6@...edance.com,
	Toke Høiland-Jørgensen <toke@...hat.com>
Subject: Re: [PATCH bpf-next v2] net: Don't allow to attach xdp if bond slave
 device's upper already has a program

Fri, Aug 23, 2024 at 02:07:45PM CEST, daniel@...earbox.net wrote:
>On 8/23/24 1:55 PM, Jiri Pirko wrote:
>> Fri, Aug 23, 2024 at 10:42:04AM CEST, zhoufeng.zf@...edance.com wrote:
>> > From: Feng Zhou <zhoufeng.zf@...edance.com>
>> > 
>> > Cannot attach when an upper device already has a program, This
>> > restriction is only for bond's slave devices or team port, and
>> > should not be accidentally injured for devices like eth0 and vxlan0.
>> 
>> What if I attach xdp program to solo netdev and then I enslave it
>> to bond/team netdev that already has xdp program attached?
>> What prevents me from doing that?
>
>In that case the enslaving of the device to bond(/team) must fail as
>otherwise the latter won't be able to propagate the XDP prog downwards.

Yep, I don't see that in the code though.


>
>Feng, did you double check if we have net or BPF selftest coverage for
>that? If not might be good to add.
>
>> > Reviewed-by: Toke Høiland-Jørgensen <toke@...hat.com>
>> > Signed-off-by: Feng Zhou <zhoufeng.zf@...edance.com>
>> > ---
>> > Changelog:
>> > v1->v2: Addressed comments from Paolo Abeni, Jiri Pirko
>> > - Use "netif_is_lag_port" relace of "netif_is_bond_slave"
>> > Details in here:
>> > https://lore.kernel.org/netdev/3bf84d23-a561-47ae-84a4-e99488fc762b@bytedance.com/T/
>> > 
>> > net/core/dev.c | 10 ++++++----
>> > 1 file changed, 6 insertions(+), 4 deletions(-)
>> > 
>> > diff --git a/net/core/dev.c b/net/core/dev.c
>> > index f66e61407883..49144e62172e 100644
>> > --- a/net/core/dev.c
>> > +++ b/net/core/dev.c
>> > @@ -9502,10 +9502,12 @@ static int dev_xdp_attach(struct net_device *dev, struct netlink_ext_ack *extack
>> > 	}
>> > 
>> > 	/* don't allow if an upper device already has a program */
>> > -	netdev_for_each_upper_dev_rcu(dev, upper, iter) {
>> > -		if (dev_xdp_prog_count(upper) > 0) {
>> > -			NL_SET_ERR_MSG(extack, "Cannot attach when an upper device already has a program");
>> > -			return -EEXIST;
>> > +	if (netif_is_lag_port(dev)) {
>> > +		netdev_for_each_upper_dev_rcu(dev, upper, iter) {
>> > +			if (dev_xdp_prog_count(upper) > 0) {
>> > +				NL_SET_ERR_MSG(extack, "Cannot attach when an upper device already has a program");
>> > +				return -EEXIST;
>> > +			}
>> > 		}
>> > 	}
>> > 
>> > -- 
>> > 2.30.2
>> > 
>> 
>