netdev - Re: [PATCH net-next 4/6] bonding: don't validate arp if we don't have to

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:	Thu, 20 Jun 2013 10:43:22 +0200
From:	Veaceslav Falico <vfalico@...hat.com>
To:	Nikolay Aleksandrov <nikolay@...hat.com>
Cc:	netdev@...r.kernel.org, fubar@...ibm.com, andy@...yhouse.net,
	davem@...emloft.net, linux@...2.net, nicolas.2p.debian@...e.fr,
	rick.jones2@...com
Subject: Re: [PATCH net-next 4/6] bonding: don't validate arp if we don't
 have to

On Thu, Jun 20, 2013 at 12:19:04AM +0200, Nikolay Aleksandrov wrote:
>On 19/06/13 19:34, Veaceslav Falico wrote:
>> Currently, we validate all the incoming arps if arp_validate not 0.
>> However, we don't have to validate backup slaves if arp_validate == active
>> and vice versa, so return early in bond_arp_rcv() in these cases.
>>
>> It works correctly now because we verify arp_validate in slave_last_rx(),
>> however we're just doing useless work in bond_arp_rcv().
>>
>> Signed-off-by: Veaceslav Falico <vfalico@...hat.com>
>> ---
>>  drivers/net/bonding/bond_main.c |    4 ++++
>>  1 files changed, 4 insertions(+), 0 deletions(-)
>>
>> diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
>> index b69c7f0..2cfbb2e 100644
>> --- a/drivers/net/bonding/bond_main.c
>> +++ b/drivers/net/bonding/bond_main.c
>> @@ -2624,6 +2624,10 @@ static int bond_arp_rcv(const struct sk_buff *skb, struct bonding *bond,
>>  		return RX_HANDLER_ANOTHER;
>>
>>  	read_lock(&bond->lock);
>> +
>> +	if (!slave_do_arp_validate(bond, slave))
>> +		goto out_unlock;
>> +
>>  	alen = arp_hdr_len(bond->dev);
>>
>>  	pr_debug("bond_arp_rcv: bond %s skb->dev %s\n",
>Hm, I think this issue runs deeper because recv_probe can be wrong and
>also if arp_validate is enabled while the bond is running then
>recv_probe is not set (or unset for that matter if disabled). I have a
>patch which needs little more work for some time now in my queue that
>fixes this, but if you'd like to fix it I'd suggest addressing that
>issue (recv_probe), because then you can just drop these checks and
>improve performance when disabled (after it's been enabled).

Yup, recv_probe value is really poorly synced with the arp_validate, I'll
try to take a look at it when I have time and in case you won't fix it by
that time :).

However, I don't think we should drop this check even in this case. This
check just verifies if we should validate this exact slave - being it
active or backup, and considering the value of arp_validate (which can be
active/backup/both).

Maybe I've understood you wrong, though :).

>This got a bit confusing when I read it :-)
>
>Cheers,
> Nik
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html