Package: openvswitch
Version: 1.9.2+git20130813
Severity: grave

I'm filing this bug to track the various "make check" failures that
caused the 1.9.2+git20130813 build to fail on most archs.  I need to
spend some time tracking down the exact causes, and this bug report
seems to be a reasonable place to put some of that.

There seem to be a couple of classes of problems, see sections below.

----------------------------------------------------------------------

 Most of the amd64 test failures, at
https://buildd.debian.org/status/fetch.php?pkg=openvswitch&arch=amd64&ver=1.9.2%2Bgit20130813-1&stamp=1376461705
exemplify one of those.  These amd64 failures involve Python and most
of them involve the unixctl client or server:

 288: unixctl-py.at:6    unixctl ovs-vswitchd exit - Python
 289: unixctl-py.at:17   unixctl ovs-vswitchd help - Python
 290: unixctl-py.at:32   unixctl ovs-vswitchd arguments - Python
 292: unixctl-py.at:91   unixctl server - Python
 1152: ovs-monitor-ipsec.at:3 ovs-monitor-ipsec
 1153: ovs-xapi-sync.at:3 ovs-xapi-sync
 1172: vlog.at:106        vlog - vlog/reopen - Python
 1173: vlog.at:135        vlog - vlog/reopen without log file - Python
 1174: vlog.at:153        vlog - vlog/reopen can't reopen log file - Python
 1175: vlog.at:185        vlog - vlog/set and vlog/list - Python

Here's an example failure detail, for test 288:

    ../../tests/unixctl-py.at:8: ovs-vsctl -- add-br br0 -- set bridge br0 
datapath-type=dummy other-config:datapath-id=fedcba9876543210 
other-config:hwaddr=aa:55:aa:55:00:00 fail-mode=secure --  
    ../../tests/unixctl-py.at:10: $PYTHON $srcdir/appctl.py --timeout 20 -t 
ovs-vswitchd exit
    --- /dev/null       2013-04-04 14:58:34.000000000 +0000
    +++ 
/??BUILDDIR??/openvswitch-1.9.2+git20130813/_debian/tests/testsuite.dir/at-groups/288/stderr
    2013-08-14 06:27:32.094459992 +0000
    @@ -0,0 +1,15 @@
    +2013-08-14T06:27:32Z|0|unixctl_client|WARN|failed to connect to 
/??BUILDDIR??/openvswitch-1.9.2+git20130813/_debian/tests/testsuite.dir/0288/ovs-vswitchd.11263.ctl
    +Traceback (most recent call last):
    +  File "../../../../tests/appctl.py", line 75, in <module>
    +    main()
    +  File "../../../../tests/appctl.py", line 59, in main
    +    client = connect_to_target(target)
    +  File "../../../../tests/appctl.py", line 35, in connect_to_target
    +    ovs.util.ovs_fatal(error, "cannot connect to \"%s\"" % socket_name)
    +  File "/??BUILDDIR??/openvswitch-1.9.2+git20130813/python/ovs/util.py", 
line 92, in ovs_fatal
    +    ovs_error(*args, **kwargs)
    +  File "/??BUILDDIR??/openvswitch-1.9.2+git20130813/python/ovs/util.py", 
line 76, in ovs_error
    +    err_msg += " (%s)" % ovs_retval_to_string(err_no)
    +  File "/??BUILDDIR??/openvswitch-1.9.2+git20130813/python/ovs/util.py", 
line 60, in ovs_retval_to_string
    +    return os.strerror(retval)
    +TypeError: an integer is required
    ../../tests/unixctl-py.at:10: exit code was 1, expected 0

----------------------------------------------------------------------

sparc has these failures but others too, see
https://buildd.debian.org/status/fetch.php?pkg=openvswitch&arch=sparc&ver=1.9.2%2Bgit20130813-1&stamp=1376471499:

216: learning action - learn refreshes hard_age      FAILED (learn.at:138)
217: learning action - TCPv4 port learning           FAILED (learn.at:199)
509: ofproto-dpif - dec_ttl                          FAILED (ofproto-dpif.at:80)
514: ofproto-dpif - controller                       FAILED 
(ofproto-dpif.at:263)
526: ofproto-dpif - NetFlow flow expiration          FAILED 
(ofproto-dpif.at:969)
527: ofproto-dpif - NetFlow active expiration        FAILED 
(ofproto-dpif.at:1019)
528: idle_age and hard_age increase over time        FAILED 
(ofproto-dpif.at:1082)

plus others, such as 501, that hang and cause the build to time out.

A little work with gdb shows that 501 failed due to misaligned data
access in a "packet out" OpenFlow command in packet_set_ipv4().  It
looks like handle_packet_out() in ofproto.c needs to properly align
the L3 header.  It seems likely that the other failures are for the
same or related reasons.


-- 
To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Reply via email to