perfSONAR User Q&A and Other Discussion

[Brian Candler <>]

On 06/12/2014 19:43, Tim Rayner wrote:
> I'd be happy to see the iperf3 -Z option to be made a bwctl option, or at 
> least reported with the bwctl test results - it was quite confusing get 
> poorer performance when I copied and pasted the reported iperf3 invocation 
> from bwctl results.

Sure. Looking at the cope I think the simple solution would be to add a 
global "no_zerocopy" option to bwctld.conf. Otherwise if it's added as a 
flag in test_spec then I think that involves changing the protocol and 
the command-line arguments to bwctl.

I also agree it could be sensible to report the options used - although 
at the moment I don't think it reports any of them?

Regards,

Brian.

P.S. Compiling bwctld under OSX, the compiler gives a whole bunch of 
warnings. The most serious one seems to be this:

bwctld.c:304:31: warning: sizeof on array function parameter will return 
size of
      'uint8_t *' (aka 'unsigned char *') instead of 'BWLSID'
      (aka 'uint8_t [16]') [-Wsizeof-array-argument]
    memcpy(res->sid,sid,sizeof(sid));

What it's saying is: the type BWLSID is defined as
typedef uint8_t    BWLSID[16];

However the function argument was defined as:

static Reservation
AllocReservation(
        ChldState   cstate,
        BWLSID      sid,

and therefore sid is a pointer to the first element of the array, not 
the whole array, so sizeof(sid) is 4 not 16.

The target of the copy is a 16 byte array, not a pointer:

struct ReservationRec{
    BWLToolType tool;
    BWLSID      sid;

This might explain some of the problems seen with invalid SIDs? The fix 
would be to use sizeof(BWLSID) instead.

There are other warnings which are obviously bugs too, e.g.

bwctl.c:2014:37: warning: comparison of unsigned expression < 0 is 
always false
      [-Wtautological-compare]
               app.opt.timeDuration < 0) {
               ~~~~~~~~~~~~~~~~~~~~ ^ ~

I can post the full set of compiler warnings if it's helpful.