[netperf-talk] Bug in UDP_STREAM test when built w/ omni tests enabled

[Alexander Duyck]

Starting with version 2.5.0 of netperf I have been seeing issues with 
the UDP_STREAM test putting out garbage data in the results.  I also 
tested the tip of trunk and the same issue is there.  Below is an 
example of what it is I am seeing with a standard test run via "netperf 
-H localhost -t UDP_STREAM -- -m 1472":

MIGRATED UDP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 
localhost (127.0.0.1) port 0 AF_INET
Socket  Message  Elapsed      Messages
Size    Size     Time         Okay Errors   Throughput
bytes   bytes    secs            #      #   10^6bits/sec

114688    1472   10.00     1446576      0    
-25601878157136826853035381601897889696738077819666360974222349802419681726547261867936797914769925040136444939441029859440038206236524544.00
1083874757           
26815615860568076694529673297026416648975986770567292599155085326693968645065144021470572325735123815425202585064184242890315464382272384572711610422919168.00   
1076101154              0.00

If I reconfigure with the "--disable-omni" option I get what I expect 
which is:
UDP UNIDIRECTIONAL SEND TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 
localhost (127.0.0.1) port 0 AF_INET
Socket  Message  Elapsed      Messages
Size    Size     Time         Okay Errors   Throughput
bytes   bytes    secs            #      #   10^6bits/sec

114688    1472   10.00     1460094      0    1719.40
114688           10.00     1402378           1651.44


So far I have only been seeing this on a Fedora 14 32 bit system with 
gcc 4.5.1.  The same issue doesn't currently seem to be occurring with 
the TCP tests.

Thanks,

Alex