[Ns-bugs] [Bug 248] Newer version of tcp-2way that exhibits an 'excess bytes' problem

[bugzilla-daemon@nsnam-www.ece.gatech.edu]

http://www.nsnam.org/bugzilla/show_bug.cgi?id=248





------- Comment #2 from evensky at dancer.ca.sandia.gov  2008-07-08 01:31 -------
A few more comments: It looks like differences arise on the unfortunate payload
('I' filled) at the start. For some reason the initial packet is larger than
the others. There is nothing wrong with that, the recvSize is 600, but it is
odd:

The without the useBadTimings we have at the start of the payload:
Before Marking|Payload (size=10000)
After Header Marking|ns3::ArchHeader ((ArchHeader=I:10000:0:0:0)) Payload
(size=10000)
my_sim->cmpl_send(0.32,0,0,10000,0)
  536:.'.{26}I{508} *bytes from 10.1.0.1 [02-06-0a:01:00:01:01:c0]
<0|507(29)|9493>
  536:I{536} *bytes from 10.1.0.1 [02-06-0a:01:00:01:01:c0] <0|536|8957>
  536:I{536} *bytes from 10.1.0.1 [02-06-0a:01:00:01:01:c0] <0|536|8421>
  536:I{536} *bytes from 10.1.0.1 [02-06-0a:01:00:01:01:c0] <0|536|7885>
....

showing the send of 10000 bytes and the first few chunks to be read in. Note
that in this case we get 536 bytes of which 507 are payload. (the rle has an
'I' as part of the header which is why there are 508 shown in the rle dump).

If we turn on the useBadTimings, we get:
Before Marking|Payload (size=10000)
After Header Marking|ns3::ArchHeader ((ArchHeader=I:10000:0:0:0)) Payload
(size=10000)
my_sim->cmpl_send(0.00296172,0,0,10000,0)
  493:H{493} *bytes from 10.1.0.2 [02-06-0a:01:00:02:20:00] <0|493|0>
  Finished Packet.  from handleread: cmpl_recv(0.00296172,0,0,1000,0) Read 1000
of <H>
simulator::cmpl_recv(0.00296172,0,0,1000,0)
  579:.'.{26}I{551} *bytes from 10.1.0.1 [02-06-0a:01:00:01:01:c0]
<0|550(29)|9450>
  536:I{536} *bytes from 10.1.0.1 [02-06-0a:01:00:01:01:c0] <0|536|8914>
  536:I{536} *bytes from 10.1.0.1 [02-06-0a:01:00:01:01:c0] <0|536|8378>

note here we have the last bit of the 'H' filled payload interleaved between
the send of the 1000 'I' filled bytes, and the recv of those bytes. Also, we
have 43 more bytes in that first chunk, again still less that the 600 we
allowed on the command line, but odd that it is different. Also odd that 43 is
the number of bytes we seem to have in excess.


-- 
Configure bugmail: http://www.nsnam.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.