[netperf-talk] help needed

[Rick Jones]

Srinivasan S wrote:
> Hi Rick,
> Thank you for your response.
> 
> I did the following tests:
> 1. ran netserver in the board and netperf in the host and got through put as -0.00
> 2. ran only the netperf in the board without runnin netserver in the host and I got 
> "establish_control could not establish the control connection ..." This confirms that
> the control connection are established otherwise.
> 3. With strace, it appears the configuration are fine. Please see the attachment.
>  
> Bye the way how do I enable debugging in netperf ?
> 

with the -d option.

> write(1, "TCP STREAM TEST from 0.0.0.0 (0."..., 95TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to nshost (172.20.33.217) port 0 AF_INET
> ) = 95
> socket(PF_INET, SOCK_STREAM, IPPROTO_TCP) = 4
> getsockopt(4, SOL_SOCKET, SO_SNDBUF, [16384], [4]) = 0
> getsockopt(4, SOL_SOCKET, SO_RCVBUF, [87380], [4]) = 0
> setsockopt(4, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
> bind(4, {sa_family=AF_INET, sin_port=htons(0), sin_addr=inet_addr("0.0.0.0")}, 16) = 0
> brk(0x33000)                            = 0x33000
> brk(0x37000)                            = 0x37000
> send(3, "\0\0\0\n\377\377\377\377\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 256, 0) = 256
> select(1024, [3], NULL, NULL, {120, 0}) = 1 (in [3], left {120, 0})
> recv(3, "\0\0\0\v\0\0\0\0\0\1UT\0\1UT\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 256, 0) = 256
> connect(4, {sa_family=AF_INET, sin_port=htons(2087), sin_addr=inet_addr("172.20.33.217")}, 16) = 0

so, 172.20.22.217 is the correct IP yes?  seems to match

> rt_sigaction(SIGALRM, {0x98a8, [ALRM], SA_INTERRUPT|0x4000000}, NULL, 8) = 0
> alarm(10)                               = 0
> gettimeofday({2044, 561719}, {0, 0})    = 0
> send(4, "\0\0\0\0\0\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0\321\33\0\0001"..., 16384, 0) = 16384
> send(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 16384, 0) = 16384
> send(4, "\0\0\0\0\0\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0\321\33\0\0001"..., 16384, 0) = 16384
> send(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 16384, 0) = 16384
> send(4, "\0\0\0\0\0\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0\321\33\0\0001"..., 16384, 0) = 16384
 > ...
> send(4, "\0\0\0\0\0\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0\321\33\0\0001"..., 16384, 0) = 16384
> send(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 16384, 0) = 16384
> send(4, "\0\0\0\0\0\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0\321\33\0\0001"..., 16384, 0) = 16384
> send(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 16384, 0) = 16384
> send(4, "\0\0\0\0\0\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0\321\33\0\0001"..., 16384, 0) = 16384
> send(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 16384, 0) = 16384
> send(4, "\0\0\0\0\0\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0\321\33\0\0001"..., 16384, 0) = 16384
> send(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 16384, 0) = 16384
> send(4, "\0\0\0\0\0\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0\321\33\0\0001"..., 16384, 0) = 16384
> send(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 16384, 0) = 16384
> send(4, "\0\0\0\0\0\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0\321\33\0\0001"..., 16384, 0) = 16384
> send(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 16384, 0) = 16384
> send(4, "\0\0\0\0\0\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0\321\33\0\0001"..., 16384, 0) = 16384
> send(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 16384, 0) = 16384
> send(4, "\0\0\0\0\0\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0\321\33\0\0001"..., 16384, 0 <unfinished ...>
> # 

Hmm, nothing after that in the strace?  There should have been the expiration of 
the 10 second timer and stuff happening on the control connection.  It appears 
that there were enough send calls that something must have been consuming data 
on the other end.  I wonder if there are "issues" with number formats between 
the two systems.  IIRC netperf passes some floating point numbers around and 
that has been a problem in some cases in the past.

So, whenyou launch netserver add a -d option and when you launch netperf also 
add a -d and send both of those outputs - netserver will _probably_ go to 
/tmp/netperf.debug


rick jones