netdev - Re: [PATCH net 6/8] bonding: disallow setting nested bonding + ipsec offload

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]

Date:   Sat, 3 Jul 2021 15:37:48 +0900
From:   Taehee Yoo <ap420073@...il.com>
To:     Jay Vosburgh <jay.vosburgh@...onical.com>
Cc:     davem@...emloft.net, kuba@...nel.org, vfalico@...il.com,
        andy@...yhouse.net, jesse.brandeburg@...el.com,
        anthony.l.nguyen@...el.com, jarod@...hat.com,
        netdev@...r.kernel.org, intel-wired-lan@...ts.osuosl.org
Subject: Re: [PATCH net 6/8] bonding: disallow setting nested bonding + ipsec
 offload

Hi Jay,
Thank you for your review!

On 7/3/21 6:14 AM, Jay Vosburgh wrote:
 > Taehee Yoo <ap420073@...il.com> wrote:
 >
 >> bonding interface can be nested and it supports ipsec offload.
 >> So, it allows setting the nested bonding + ipsec scenario.
 >> But code does not support this scenario.
 >> So, it should be disallowed.
 >>
 >> interface graph:
 >> bond2
 >> |
 >> bond1
 >> |
 >> eth0
 >>
 >> The nested bonding + ipsec offload may not a real usecase.
 >> So, disallowing this is fine.
 >
 > 	Is a stack like "bond1 -> VLAN.XX -> bond2 -> eth0" also a
 > problem?  I don't believe the change below will detect this
 > configuration.
 >

Except bonding, all kind of virtual interfaces(vlan, team, etc) doesn't 
support ipsec offload.
It means these interfaces' xfrmdev_ops pointer is null.
So, configuration always will be failed at the following line.
         if (!slave->dev->xfrmdev_ops || 

             !slave->dev->xfrmdev_ops->xdo_dev_state_add || 

Only checking the real interface's type is enough.
So, bond1 can't set up ipsec offload but bond2 can set up ipsec offload.


Thanks a lot!
Taehee

 > 	-J
 >
 >> Fixes: 18cb261afd7b ("bonding: support hardware encryption offload 
to slaves")
 >> Signed-off-by: Taehee Yoo <ap420073@...il.com>
 >> ---
 >> drivers/net/bonding/bond_main.c | 15 +++++++++------
 >> 1 file changed, 9 insertions(+), 6 deletions(-)
 >>
 >> diff --git a/drivers/net/bonding/bond_main.c 
b/drivers/net/bonding/bond_main.c
 >> index 7659e1fab19e..f268e67cb2f0 100644
 >> --- a/drivers/net/bonding/bond_main.c
 >> +++ b/drivers/net/bonding/bond_main.c
 >> @@ -419,8 +419,9 @@ static int bond_ipsec_add_sa(struct xfrm_state *xs)
 >> 	xs->xso.real_dev = slave->dev;
 >> 	bond->xs = xs;
 >>
 >> -	if (!(slave->dev->xfrmdev_ops
 >> -	      && slave->dev->xfrmdev_ops->xdo_dev_state_add)) {
 >> +	if (!slave->dev->xfrmdev_ops ||
 >> +	    !slave->dev->xfrmdev_ops->xdo_dev_state_add ||
 >> +	    netif_is_bond_master(slave->dev)) {
 >> 		slave_warn(bond_dev, slave->dev, "Slave does not support ipsec 
offload\n");
 >> 		rcu_read_unlock();
 >> 		return -EINVAL;
 >> @@ -453,8 +454,9 @@ static void bond_ipsec_del_sa(struct xfrm_state *xs)
 >>
 >> 	xs->xso.real_dev = slave->dev;
 >>
 >> -	if (!(slave->dev->xfrmdev_ops
 >> -	      && slave->dev->xfrmdev_ops->xdo_dev_state_delete)) {
 >> +	if (!slave->dev->xfrmdev_ops ||
 >> +	    !slave->dev->xfrmdev_ops->xdo_dev_state_delete ||
 >> +	    netif_is_bond_master(slave->dev)) {
 >> 		slave_warn(bond_dev, slave->dev, "%s: no slave 
xdo_dev_state_delete\n", __func__);
 >> 		goto out;
 >> 	}
 >> @@ -479,8 +481,9 @@ static bool bond_ipsec_offload_ok(struct sk_buff 
*skb, struct xfrm_state *xs)
 >> 	if (BOND_MODE(bond) != BOND_MODE_ACTIVEBACKUP)
 >> 		return true;
 >>
 >> -	if (!(slave_dev->xfrmdev_ops
 >> -	      && slave_dev->xfrmdev_ops->xdo_dev_offload_ok)) {
 >> +	if (!slave_dev->xfrmdev_ops ||
 >> +	    !slave_dev->xfrmdev_ops->xdo_dev_offload_ok ||
 >> +	    netif_is_bond_master(slave_dev)) {
 >> 		slave_warn(bond_dev, slave_dev, "%s: no slave 
xdo_dev_offload_ok\n", __func__);
 >> 		return false;
 >> 	}
 >> --
 >> 2.17.1
 >>
 >
 > ---
 > 	-Jay Vosburgh, jay.vosburgh@...onical.com
 >