[ovs-dev] [PATCH v2] conntrack: Add a test for IPv4 UDP zero checksum

[Aaron Conole]

Paolo Valerio <pvalerio at redhat.com> writes:

> Hi Aaron,
>
> Aaron Conole <aconole at redhat.com> writes:
>
>> Recently, during some conntrack testing a bug was uncovered in a DPDK
>> PMD, which doesn't support an IPv4 packet with a zero checksum value.
>> In order to show that the connection tracking code in userspace
>> supports IPv4 UDP with a zero checksum, add a test case to enforce
>> this behavior.
>>
>> Reported-at: http://mails.dpdk.org/archives/dev/2021-January/198528.html
>> Reported-by: Paolo Valerio <pvalerio at redhat.com>
>> Signed-off-by: Aaron Conole <aconole at redhat.com>
>> ---
>> Previous version:
>> https://mail.openvswitch.org/pipermail/ovs-dev/2021-February/380266.html
>>
>> v1->v2: - Addressed most of the comments by Flavio & William Tu.
>>         - Added the 0xffff checksum case
>>         - Added a bad checksum case
>>         - For the single instance of "sleep 1," this needs to be retained
>>           due to the negative test case (we have no WAIT_UNTIL to rely on)
>>           but there are other instances of sleep 1 throughout the suite,
>>           so I guess it should be okay.
>>
>>  tests/system-traffic.at | 64 +++++++++++++++++++++++++++++++++++++++++
>>  1 file changed, 64 insertions(+)
>>
>> diff --git a/tests/system-traffic.at b/tests/system-traffic.at
>> index fb5b9a36d2..5e51c24760 100644
>> --- a/tests/system-traffic.at
>> +++ b/tests/system-traffic.at
>> @@ -5927,6 +5927,69 @@ ovs-appctl dpif/dump-flows br0
>>  OVS_TRAFFIC_VSWITCHD_STOP
>>  AT_CLEANUP
>>  
>> +
>> +AT_SETUP([conntrack - IPv4 UDP zero checksum])
>
> Don't know if this second new line up here was intentional (mixed style
> in the file).
> TBH, I have no strong preference about it, I pointed out just in case.

Ugh... every time I rebase some stray new line makes it in.  It's like I
have a 4 year old in my lap...

>> +dnl This tracks sending zero checksum packets for udp over ipv4
>> +CHECK_CONNTRACK()
>> +OVS_TRAFFIC_VSWITCHD_START()
>> +OVS_CHECK_CT_CLEAR()
>> +
>> +ADD_NAMESPACES(at_ns0, at_ns1)
>> +ADD_VETH(p0, at_ns0, br0, "10.1.1.1/24", "f0:00:00:01:01:01")
>> +ADD_VETH(p1, at_ns1, br0, "10.1.1.2/24", "f0:00:00:01:01:02")
>> +dnl setup ct flows
>> +AT_DATA([flows.txt], [dnl
>> +table=0,priority=10  ip,udp,ct_state=-trk action=ct(zone=1,table=1)
>> +table=0,priority=0   action=drop
>> +table=1,priority=10  ct_state=-est+trk+new,ip,ct_zone=1,in_port=1 action=ct(commit,table=2)
>> +table=1,priority=10  ct_state=+est-new+trk,ct_zone=1,in_port=1 action=resubmit(,2)
>> +table=1,priority=0   action=drop
>> +table=2,priority=10  ct_state=+trk+new,in_port=1 action=2
>> +table=2,priority=10  ct_state=+trk+est action=2
>> +])
>> +
>> +AT_CHECK([ovs-ofctl --bundle add-flows br0 flows.txt])
>> +
>> +# sending udp pkt with 0000 checksum
>> +NS_CHECK_EXEC([at_ns0], [$PYTHON3 $srcdir/sendpkt.py p0 f0 00 00 01
>> 01 02 f0 00 00 01 01 01 08 00 45 00 00 28 00 01 00 00 40 11 64 c0 0a
>> 01 01 01 0a 01 01 02 04 d2 04 d2 00 14 00 00 aa aa aa aa aa aa aa aa
>> aa aa aa aa > /dev/null])
>> +
>> +OVS_WAIT_UNTIL([ovs-appctl dpctl/dump-conntrack | grep "udp"])
>> +
>> +dnl ensure CT picked up the packet
>> +AT_CHECK([ovs-appctl dpctl/dump-conntrack | FORMAT_CT(10.1.1)], [0], [dnl
>> +udp,orig=(src=10.1.1.1,dst=10.1.1.2,sport=<cleared>,dport=<cleared>),reply=(src=10.1.1.2,dst=10.1.1.1,sport=<cleared>,dport=<cleared>)
>> +])
>> +
>> +dnl clear CT tuples
>> +AT_CHECK([ovs-appctl dpctl/flush-conntrack])
>> +AT_CHECK([ovs-appctl dpctl/dump-conntrack | grep "udp"], [1])
>> +
>> +dnl send UDP with ffff checksum
>> +NS_CHECK_EXEC([at_ns0], [$PYTHON3 $srcdir/sendpkt.py p0 f0 00 00 01
>> 01 02 f0 00 00 01 01 01 08 00 45 00 00 40 00 01 00 00 40 11 64 a8 0a
>> 01 01 01 0a 01 01 02 04 d2 04 d2 00 2c ff ff 00 00 00 00 00 00 00 00
>> 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>> 00 00 00 df ed > /dev/null])
>> +OVS_WAIT_UNTIL([ovs-appctl dpctl/dump-conntrack | grep "udp"])
>> +
>> +dnl ensure CT picked up the packet
>> +AT_CHECK([ovs-appctl dpctl/dump-conntrack | FORMAT_CT(10.1.1)], [0], [dnl
>> +udp,orig=(src=10.1.1.1,dst=10.1.1.2,sport=<cleared>,dport=<cleared>),reply=(src=10.1.1.2,dst=10.1.1.1,sport=<cleared>,dport=<cleared>)
>> +])
>> +
>> +dnl clear CT tuples
>> +AT_CHECK([ovs-appctl dpctl/flush-conntrack])
>> +AT_CHECK([ovs-appctl dpctl/dump-conntrack | grep "udp"], [1])
>> +
>> +# sending udp pkt with bad checksum
>> +NS_CHECK_EXEC([at_ns0], [$PYTHON3 $srcdir/sendpkt.py p0 f0 00 00 01
>> 01 02 f0 00 00 01 01 01 08 00 45 00 00 28 00 01 00 00 40 11 64 c0 0a
>> 01 01 01 0a 01 01 02 04 d2 04 d2 00 14 11 11 aa aa aa aa aa aa aa aa
>> aa aa aa aa > /dev/null])
>> +
>> +dnl There isn't anything else to do here but call a sleep... we can't use a WAIT_UNTIL
>> +dnl because there isn't a condition to check against - we just expect the packet to
>> +dnl be dropped.
>> +sleep 1
>> +
>> +AT_CHECK([ovs-appctl dpctl/dump-conntrack | grep "udp"], [1])
>> +
>
> what if we restrict a little the pattern here?

Sure, but I'm not sure how best to restrict it.  Something like:

  grep "udp,orig=(src=10.1.1.1,dst=10.1.1.2,sport=1234,dport=1234)"

I can do that - it makes the most sense to me.

> In case of kernel dp, considering the sleep, there's a chance an UDP
> entry (e.g. a DNS query) gets created between the flush and the dump
> with grep returning 0.

ACK

>
>> +OVS_TRAFFIC_VSWITCHD_STOP
>> +AT_CLEANUP
>> +
>>  AT_SETUP([conntrack - Multiple ICMP traverse])
>>  dnl This tracks sending ICMP packets via conntrack multiple times for the
>>  dnl same packet
>> -- 
>> 2.25.4