[ovs-dev] [patch v1] tests: Include ovn-macros for system testsuites.

Darrell Ball dlu998 at gmail.com
Tue Jun 4 17:45:35 UTC 2019


On Tue, Jun 4, 2019 at 8:44 AM Ilya Maximets <i.maximets at samsung.com> wrote:

> On 04.06.2019 7:37, Darrell Ball wrote:
> >
> >
> > On Mon, Jun 3, 2019 at 10:27 AM Ilya Maximets <i.maximets at samsung.com
> <mailto:i.maximets at samsung.com>> wrote:
> >
> >     On 03.06.2019 18:55, Darrell Ball wrote:
> >     >
> >     >
> >     > On Mon, Jun 3, 2019 at 12:48 AM Ilya Maximets <
> i.maximets at samsung.com <mailto:i.maximets at samsung.com> <mailto:
> i.maximets at samsung.com <mailto:i.maximets at samsung.com>>> wrote:
> >     >
> >     >     On 25.05.2019 23:49, Darrell Ball wrote:
> >     >     > ovn-macros are needed to run the OVN system tests.
> >     >     >
> >     >     > Signed-off-by: Darrell Ball <dlu998 at gmail.com <mailto:
> dlu998 at gmail.com> <mailto:dlu998 at gmail.com <mailto:dlu998 at gmail.com>>>
> >     >     > Tested-by: Greg Rose <gvrose8192 at gmail.com <mailto:
> gvrose8192 at gmail.com> <mailto:gvrose8192 at gmail.com <mailto:
> gvrose8192 at gmail.com>>>
> >     >     > Reviewed-by: Greg Rose <gvrose8192 at gmail.com <mailto:
> gvrose8192 at gmail.com> <mailto:gvrose8192 at gmail.com <mailto:
> gvrose8192 at gmail.com>>>
> >     >     > ---
> >     >
> >     >     Hi Darrell.
> >     >     Thanks for the fix. It looks good.
> >     >
> >     >     BTW, I have a question. Are you able to run 'make
> check-system-userspace'
> >     >     successfully? I have a lot of failures beside the OVN related
> ones.
> >     >     I didn't look closely yet.
> >     >
> >     >
> >     > which tests are failing for you ?; I can check later.
> >
> >     Thanks.
> >     Here is the list of currently failing tests:
> >
> >
> > These tests are un-loved for some time
> > Copying couple folks.
> >
> >
> >
> >     datapath-sanity
> >
> >       9: datapath - ping over vxlan tunnel               FAILED (
> system-traffic.at:246 <
> https://protect2.fireeye.com/url?k=ca2351833f1c13f5.ca22dacc-2d077141969cf994&u=http://system-traffic.at:246
> >)
> >      10: datapath - ping over vxlan6 tunnel              FAILED (
> system-traffic.at:289 <
> https://protect2.fireeye.com/url?k=806bcaddf0570229.806a4192-31aa7a98ff0ea74c&u=http://system-traffic.at:289
> >)
> >      16: datapath - ping over geneve tunnel              FAILED (
> system-traffic.at:522 <
> https://protect2.fireeye.com/url?k=1b59b694f3eb247f.1b583ddb-0660c0ec5c47f5bf&u=http://system-traffic.at:522
> >)
> >      17: datapath - flow resume with geneve tun_metadata FAILED (
> system-traffic.at:568 <
> https://protect2.fireeye.com/url?k=a4cd38cbf817df78.a4ccb384-2d47344d2cec23cb&u=http://system-traffic.at:568
> >)
> >      18: datapath - ping over geneve6 tunnel             FAILED (
> system-traffic.at:606 <
> https://protect2.fireeye.com/url?k=e826024a4d9e8b12.e8278905-9b148a9e2dd10697&u=http://system-traffic.at:606
> >)
> >
> >
> > These should be bypassed for userspace due to dependencies on kernel
> modules
> > Also depends on routing table info - see below.
> >
> >
> >
> >     nsh-datapath
> >
> >     121: nsh - decap header                              FAILED (
> system-traffic.at:5839 <http://system-traffic.at:5839>)
> >     122: nsh - replace header                            FAILED (
> system-traffic.at:5876 <http://system-traffic.at:5876>)
> >     123: nsh - forward                                   FAILED (
> system-traffic.at:5931 <http://system-traffic.at:5931>)
> >
> >
> >
> +2019-06-04T03:38:14.365Z|00001|odp_util(revalidator5)|WARN|OVS_NSH_KEY_ATTR_MD1
> present but declared mdtype 0 is not 1 (NSH_M_TYPE1)
> > +2019-06-04T03:38:14.366Z|00002|odp_util(revalidator5)|WARN|the flow
> mask in error is:
> skb_priority(0),tunnel(ttl=0,flags(0)),skb_mark(0),ct_state(0),ct_zone(0),ct_mark(0),ct_label(0),recirc_id(0xffffffff),dp_hash(0),in_port(4294967295),packet_type(ns=65535,id=0xffff),eth(src=00:00:00:00:00:00,dst=00:00:00:00:00:00),eth_type(0xffff),nsh(flags=0ttl=0,mdtype=0,np=255,spi=0xffffff,si=255),
> for the following flow key:
> in_port=2,vlan_tci=0x0000,dl_src=f2:ff:00:00:00:01,dl_dst=f2:ff:00:00:00:02,dl_type=0x894f,nsh_flags=0,nsh_ttl=8,nsh_mdtype=1,nsh_np=3,nsh_spi=0x100,nsh_si=3,nsh_c1=0x1020304,nsh_c2=0x5060708,nsh_c3=0x90a0b0c,nsh_c4=0xd0e0f10,nw_proto=0,nw_tos=0,nw_ecn=0,nw_ttl=0
> >
> > 'OVS_NSH_KEY_ATTR_MD1 present but declared mdtype 0 is not 1
> (NSH_M_TYPE1)'
> > also needs bisecting
> >
> >
> >
> >     layer3-tunnels
> >
> >     124: layer3 - ping over VXLAN-GPE                    FAILED (
> ovs-macros.at:219 <http://ovs-macros.at:219>)
> >     125: layer3 - ping over GRE                          FAILED (
> ovs-macros.at:219 <http://ovs-macros.at:219>)
> >
> >
> > First issue seems to be related to:
> > 898d7b052       Ben Pfaff       Sat Mar 31 17:12:55 2018 -0700
>  ovs-vswitchd: Do not use system routing table with --disable-system.
> > which adds suppressing routing table info/updates to 'disable-system'
> directive, which system userspace tests use.
>
> Good catch. Thank you!
> By reverting this patch all the tests passed except two NSH tests (122 and
> 123).
> Actually, userspace tunneling usually relates on system routing tables.
> So, it's
> unclear what to do with this change.
>

1/ The NSH test failures are unrelated to the above change
     They still fail for the same reasons for me when I reverted the above
change, which I would expect.

nsh-datapath

121: nsh - decap header                              FAILED (
system-traffic.at:5839)
122: nsh - replace header                            FAILED (
system-traffic.at:5876)
123: nsh - forward                                   FAILED (
system-traffic.at:5931)


2/ In terms of whether to revert the change, we may consider "selective
disabling" in tests w.r.t. the system routing table
   per the comment below.

3/ Unfortunately, I get other warn logs that fail the tests.


>
> >
> > At least a couple ways to fix this first issue, including possibly more
> selective disabling

> If there are other non-system tests covering functionality, then possibly
> just bypass.
> >
> >
> >
> >     You may found a bit stripped version of
> system-userspace-testsuite.log here:
> >         https://pastebin.com/RX58PnKR
> >
> >     Best regards, Ilya Maximets.
> >
>


More information about the dev mailing list