| Message ID | 1517452349-120242-1-git-send-email-u9012063@gmail.com |
|---|---|
| State | Accepted |
| Headers | show |
| Series | [ovs-dev] learn: improve test case | expand |
On Wed, Jan 31, 2018 at 6:32 PM, William Tu <u9012063@gmail.com> wrote: > Current learn test cases use only ovs-ofctl add/del flows. > The patch add a new test case for learn with delete_learned and > limit option enabled. > > Signed-off-by: William Tu <u9012063@gmail.com> > --- > tests/learn.at | 38 ++++++++++++++++++++++++++++++++++++++ > 1 file changed, 38 insertions(+) > > diff --git a/tests/learn.at b/tests/learn.at > index 07ad043212ed..406a827aae36 100644 > --- a/tests/learn.at > +++ b/tests/learn.at > @@ -625,6 +625,44 @@ AT_CHECK([ovs-ofctl dump-flows br0 --no-stats | sort]) > OVS_VSWITCHD_STOP > AT_CLEANUP > > +AT_SETUP([learning action - delete_learned/limit with packet]) > Thanks William. I have question - is this test case trying to delineate some related bug or observed special behavior ? I am asking because I don't immediately see the intersection b/w the delete_learned and limit options ? BTW, what is "limit with packet" as opposed to just "limit" > +OVS_VSWITCHD_START( > + [add-port br0 p1 -- set Interface p1 type=dummy ofport_request=1 --\ > + add-port br0 p2 -- set Interface p2 type=dummy ofport_request=2]) > + > +# Add some initial flows and check that it was successful. > +AT_DATA([flows.txt], [dnl > +table=0 actions=set_field:0x2->reg7,set_field:0xabcdef01->metadata, > resubmit(,1) > +table=1 actions=learn(table=10,delete_learned,cookie=0x123,limit=3, > result_dst=NXM_NX_REG6[[0]],NXM_OF_ETH_DST[[]]=NXM_OF_ETH_ > SRC[[]],OXM_OF_METADATA[[]],output:NXM_NX_REG7) > +]) > + > +AT_CHECK([ovs-ofctl add-flows br0 flows.txt]) > +AT_CHECK([ovs-ofctl dump-flows br0 --no-stats | sort], [0], [dnl > + actions=load:0x2->NXM_NX_REG7[[]],load:0xabcdef01->OXM_OF_ > METADATA[[]],resubmit(,1) > + table=1, actions=learn(table=10,delete_learned,cookie=0x123,limit=3, > result_dst=NXM_NX_REG6[[0]],NXM_OF_ETH_DST[[]]=NXM_OF_ETH_ > SRC[[]],OXM_OF_METADATA[[]],output:NXM_NX_REG7[[]]) > +]) > + > +dnl Each packet will generate its own flow at table=10, except last one > +AT_CHECK([ovs-appctl netdev-dummy/receive p1 'in_port(1),eth(src=50:54:00: > 00:00:01,dst=50:54:00:00:00:ff),eth_type(0x1234)']) > +AT_CHECK([ovs-appctl netdev-dummy/receive p1 'in_port(1),eth(src=50:54:00: > 00:00:02,dst=50:54:00:00:00:ff),eth_type(0x1234)']) > +AT_CHECK([ovs-appctl netdev-dummy/receive p1 'in_port(1),eth(src=50:54:00: > 00:00:03,dst=50:54:00:00:00:ff),eth_type(0x1234)']) > +AT_CHECK([ovs-appctl netdev-dummy/receive p1 'in_port(1),eth(src=50:54:00: > 00:00:04,dst=50:54:00:00:00:ff),eth_type(0x1234)']) > + > +AT_CHECK([ovs-ofctl dump-flows br0 table=10 --no-stats | sort], [0], [dnl > + cookie=0x123, table=10, metadata=0xabcdef01,dl_dst=50:54:00:00:00:01 > actions=output:2 > + cookie=0x123, table=10, metadata=0xabcdef01,dl_dst=50:54:00:00:00:02 > actions=output:2 > + cookie=0x123, table=10, metadata=0xabcdef01,dl_dst=50:54:00:00:00:03 > actions=output:2 > +]) > + > +ovs-appctl revalidator/wait > + > +AT_CHECK([ovs-ofctl del-flows br0 'table=1']) > +AT_CHECK([ovs-ofctl dump-flows br0 table=10 --no-stats | sort], [0], [dnl > It looks odd to sort an expected empty list. > +]) > + > +OVS_VSWITCHD_STOP > +AT_CLEANUP > + > AT_SETUP([learning action - limit]) > OVS_VSWITCHD_START > AT_CHECK([ovs-appctl vlog/set dpif:dbg dpif_netdev:dbg]) > -- > 2.7.4 > > _______________________________________________ > dev mailing list > dev@openvswitch.org > https://mail.openvswitch.org/mailman/listinfo/ovs-dev >
> Thanks William. > I have question - is this test case trying to delineate some related bug or > observed special behavior ? We suspect that there might be a memory leak in learn action. So I created this test. > I am asking because I don't immediately see the intersection b/w the > delete_learned and limit options ? I observe some cases that the learn action creates more than 1,000 rules, so testing delete_learned and limit to make sure the created rules operate ok. > > BTW, what is "limit with packet" as opposed to just "limit" I mean using packets to trigger the learn action. Because the previous similar tests do not use packets. > > >> >> +OVS_VSWITCHD_START( >> + [add-port br0 p1 -- set Interface p1 type=dummy ofport_request=1 --\ >> + add-port br0 p2 -- set Interface p2 type=dummy ofport_request=2]) >> + >> +# Add some initial flows and check that it was successful. >> +AT_DATA([flows.txt], [dnl >> +table=0 actions=set_field:0x2->reg7,set_field:0xabcdef01->metadata, >> resubmit(,1) >> +table=1 >> actions=learn(table=10,delete_learned,cookie=0x123,limit=3,result_dst=NXM_NX_REG6[[0]],NXM_OF_ETH_DST[[]]=NXM_OF_ETH_SRC[[]],OXM_OF_METADATA[[]],output:NXM_NX_REG7) >> +]) >> + >> +AT_CHECK([ovs-ofctl add-flows br0 flows.txt]) >> +AT_CHECK([ovs-ofctl dump-flows br0 --no-stats | sort], [0], [dnl >> + >> actions=load:0x2->NXM_NX_REG7[[]],load:0xabcdef01->OXM_OF_METADATA[[]],resubmit(,1) >> + table=1, >> actions=learn(table=10,delete_learned,cookie=0x123,limit=3,result_dst=NXM_NX_REG6[[0]],NXM_OF_ETH_DST[[]]=NXM_OF_ETH_SRC[[]],OXM_OF_METADATA[[]],output:NXM_NX_REG7[[]]) >> +]) >> + >> +dnl Each packet will generate its own flow at table=10, except last one >> +AT_CHECK([ovs-appctl netdev-dummy/receive p1 >> 'in_port(1),eth(src=50:54:00:00:00:01,dst=50:54:00:00:00:ff),eth_type(0x1234)']) >> +AT_CHECK([ovs-appctl netdev-dummy/receive p1 >> 'in_port(1),eth(src=50:54:00:00:00:02,dst=50:54:00:00:00:ff),eth_type(0x1234)']) >> +AT_CHECK([ovs-appctl netdev-dummy/receive p1 >> 'in_port(1),eth(src=50:54:00:00:00:03,dst=50:54:00:00:00:ff),eth_type(0x1234)']) >> +AT_CHECK([ovs-appctl netdev-dummy/receive p1 >> 'in_port(1),eth(src=50:54:00:00:00:04,dst=50:54:00:00:00:ff),eth_type(0x1234)']) >> + >> +AT_CHECK([ovs-ofctl dump-flows br0 table=10 --no-stats | sort], [0], [dnl >> + cookie=0x123, table=10, metadata=0xabcdef01,dl_dst=50:54:00:00:00:01 >> actions=output:2 >> + cookie=0x123, table=10, metadata=0xabcdef01,dl_dst=50:54:00:00:00:02 >> actions=output:2 >> + cookie=0x123, table=10, metadata=0xabcdef01,dl_dst=50:54:00:00:00:03 >> actions=output:2 >> +]) >> + >> +ovs-appctl revalidator/wait >> + >> +AT_CHECK([ovs-ofctl del-flows br0 'table=1']) >> +AT_CHECK([ovs-ofctl dump-flows br0 table=10 --no-stats | sort], [0], [dnl > > > It looks odd to sort an expected empty list. > Right, I should remove it. Thanks William > >> >> +]) >> + >> +OVS_VSWITCHD_STOP >> +AT_CLEANUP >> + >> AT_SETUP([learning action - limit]) >> OVS_VSWITCHD_START >> AT_CHECK([ovs-appctl vlog/set dpif:dbg dpif_netdev:dbg]) >> -- >> 2.7.4 >> >> _______________________________________________ >> dev mailing list >> dev@openvswitch.org >> https://mail.openvswitch.org/mailman/listinfo/ovs-dev > >
On Wed, Jan 31, 2018 at 06:32:29PM -0800, William Tu wrote: > Current learn test cases use only ovs-ofctl add/del flows. > The patch add a new test case for learn with delete_learned and > limit option enabled. > > Signed-off-by: William Tu <u9012063@gmail.com> Thanks, applied to master.
diff --git a/tests/learn.at b/tests/learn.at index 07ad043212ed..406a827aae36 100644 --- a/tests/learn.at +++ b/tests/learn.at @@ -625,6 +625,44 @@ AT_CHECK([ovs-ofctl dump-flows br0 --no-stats | sort]) OVS_VSWITCHD_STOP AT_CLEANUP +AT_SETUP([learning action - delete_learned/limit with packet]) +OVS_VSWITCHD_START( + [add-port br0 p1 -- set Interface p1 type=dummy ofport_request=1 --\ + add-port br0 p2 -- set Interface p2 type=dummy ofport_request=2]) + +# Add some initial flows and check that it was successful. +AT_DATA([flows.txt], [dnl +table=0 actions=set_field:0x2->reg7,set_field:0xabcdef01->metadata, resubmit(,1) +table=1 actions=learn(table=10,delete_learned,cookie=0x123,limit=3,result_dst=NXM_NX_REG6[[0]],NXM_OF_ETH_DST[[]]=NXM_OF_ETH_SRC[[]],OXM_OF_METADATA[[]],output:NXM_NX_REG7) +]) + +AT_CHECK([ovs-ofctl add-flows br0 flows.txt]) +AT_CHECK([ovs-ofctl dump-flows br0 --no-stats | sort], [0], [dnl + actions=load:0x2->NXM_NX_REG7[[]],load:0xabcdef01->OXM_OF_METADATA[[]],resubmit(,1) + table=1, actions=learn(table=10,delete_learned,cookie=0x123,limit=3,result_dst=NXM_NX_REG6[[0]],NXM_OF_ETH_DST[[]]=NXM_OF_ETH_SRC[[]],OXM_OF_METADATA[[]],output:NXM_NX_REG7[[]]) +]) + +dnl Each packet will generate its own flow at table=10, except last one +AT_CHECK([ovs-appctl netdev-dummy/receive p1 'in_port(1),eth(src=50:54:00:00:00:01,dst=50:54:00:00:00:ff),eth_type(0x1234)']) +AT_CHECK([ovs-appctl netdev-dummy/receive p1 'in_port(1),eth(src=50:54:00:00:00:02,dst=50:54:00:00:00:ff),eth_type(0x1234)']) +AT_CHECK([ovs-appctl netdev-dummy/receive p1 'in_port(1),eth(src=50:54:00:00:00:03,dst=50:54:00:00:00:ff),eth_type(0x1234)']) +AT_CHECK([ovs-appctl netdev-dummy/receive p1 'in_port(1),eth(src=50:54:00:00:00:04,dst=50:54:00:00:00:ff),eth_type(0x1234)']) + +AT_CHECK([ovs-ofctl dump-flows br0 table=10 --no-stats | sort], [0], [dnl + cookie=0x123, table=10, metadata=0xabcdef01,dl_dst=50:54:00:00:00:01 actions=output:2 + cookie=0x123, table=10, metadata=0xabcdef01,dl_dst=50:54:00:00:00:02 actions=output:2 + cookie=0x123, table=10, metadata=0xabcdef01,dl_dst=50:54:00:00:00:03 actions=output:2 +]) + +ovs-appctl revalidator/wait + +AT_CHECK([ovs-ofctl del-flows br0 'table=1']) +AT_CHECK([ovs-ofctl dump-flows br0 table=10 --no-stats | sort], [0], [dnl +]) + +OVS_VSWITCHD_STOP +AT_CLEANUP + AT_SETUP([learning action - limit]) OVS_VSWITCHD_START AT_CHECK([ovs-appctl vlog/set dpif:dbg dpif_netdev:dbg])
Current learn test cases use only ovs-ofctl add/del flows. The patch add a new test case for learn with delete_learned and limit option enabled. Signed-off-by: William Tu <u9012063@gmail.com> --- tests/learn.at | 38 ++++++++++++++++++++++++++++++++++++++ 1 file changed, 38 insertions(+)