netdev - Re: [PATCH net-next] ipv6: Reject routes configurations that specify dsfield (tos)

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: <0b401730-4f18-4e61-1c88-1dce438d6166@linuxfoundation.org>
Date:   Thu, 10 Feb 2022 15:15:34 -0700
From:   Shuah Khan <skhan@...uxfoundation.org>
To:     Guillaume Nault <gnault@...hat.com>
Cc:     David Miller <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>, netdev@...r.kernel.org,
        Hideaki YOSHIFUJI <yoshfuji@...ux-ipv6.org>,
        David Ahern <dsahern@...nel.org>,
        Shuah Khan <shuah@...nel.org>, linux-kselftest@...r.kernel.org,
        Toke Høiland-Jørgensen <toke@...hat.com>,
        Shuah Khan <skhan@...uxfoundation.org>
Subject: Re: [PATCH net-next] ipv6: Reject routes configurations that specify
 dsfield (tos)

On 2/10/22 3:05 PM, Guillaume Nault wrote:
> On Thu, Feb 10, 2022 at 11:23:20AM -0700, Shuah Khan wrote:
>> On 2/10/22 8:08 AM, Guillaume Nault wrote:
>>> The ->rtm_tos option is normally used to route packets based on both
>>> the destination address and the DS field. However it's ignored for
>>> IPv6 routes. Setting ->rtm_tos for IPv6 is thus invalid as the route
>>> is going to work only on the destination address anyway, so it won't
>>> behave as specified.
>>>
>>> Suggested-by: Toke Høiland-Jørgensen <toke@...hat.com>
>>> Signed-off-by: Guillaume Nault <gnault@...hat.com>
>>> ---
>>> The same problem exists for ->rtm_scope. I'm working only on ->rtm_tos
>>> here because IPv4 recently started to validate this option too (as part
>>> of the DSCP/ECN clarification effort).
>>> I'll give this patch some soak time, then send another one for
>>> rejecting ->rtm_scope in IPv6 routes if nobody complains.
>>>
>>>    net/ipv6/route.c                         |  6 ++++++
>>>    tools/testing/selftests/net/fib_tests.sh | 13 +++++++++++++
>>>    2 files changed, 19 insertions(+)
>>>
>>> diff --git a/net/ipv6/route.c b/net/ipv6/route.c
>>> index f4884cda13b9..dd98a11fbdb6 100644
>>> --- a/net/ipv6/route.c
>>> +++ b/net/ipv6/route.c
>>> @@ -5009,6 +5009,12 @@ static int rtm_to_fib6_config(struct sk_buff *skb, struct nlmsghdr *nlh,
>>>    	err = -EINVAL;
>>>    	rtm = nlmsg_data(nlh);
>>> +	if (rtm->rtm_tos) {
>>> +		NL_SET_ERR_MSG(extack,
>>> +			       "Invalid dsfield (tos): option not available for IPv6");
>>
>> Is this an expected failure on ipv6, in which case should this test report
>> pass? Should it print "failed as expected" or is returning fail from errout
>> is what should happen?
> 
> This is an expected failure. When ->rtm_tos is set, iproute2 fails with
> error code 2 and prints
> "Error: Invalid dsfield (tos): option not available for IPv6.".
> 
> The selftest redirects stderr to /dev/null by default (unless -v is
> passed on the command line) and expects the command to fail and
> return 2. So the default output is just:
> 
> IPv6 route with dsfield tests
>      TEST: Reject route with dsfield                                     [ OK ]
> 
> Of course, on a kernel that accepts non-null ->rtm_tos, "[ OK ]"
> becomes "[FAIL]", and the the failed tests couter is incremented.
> 

Sounds good to me.

>>
>> With the above comment addressed or explained.
>>
>> Reviewed-by: Shuah Khan <skhan@...uxfoundation.org>
>>

thanks,
-- Shuah