[nos-bbs] TCP subcommand limit

(Skip) K8RRA k8rra at ameritech.net
Mon May 28 21:05:54 EDT 2007

This is one of those times it shows I don't speak C well Maiko.

On Mon, 2007-05-28 at 15:35 -0500, Maiko Langelaar (ve4klm) wrote:
> Hi Skip,
> > Does anyone know why the "tcp irtt" command has (should have?)
> > a limit of three lines when it is used to display the srtt cache?
> The RTTCACHE define suggests that the limit is actually 16, not 3.
Yes, 16 here too.
> If you are only seeing 3 lines, perhaps it's because the code will
> only show those cache entries that have a IP ADDRESS associated with
> them (if that makes any sense). That's what the code says. If a cache
> entry has no ADDR affiliated with it, then nothing is printed for it.
The printed list starts zero long just after init.
It grows rather quickly to 3 lines long then stops growing.
I have more tcp sockets active than show in the list.
I'm particularly annoyed when the IP I really want to see is not there
(older ones are).

So, Jay sent me an example from his site - a long list.
I know it works (there).
My next step was to start gdb and peek inside --- OOPS.
GDB replies to me 
    "no symbol "RTTCACHE" in current context" 
    "print RTTCACHE" (and some other variations in syntax including
      the EXP "$GFVRTT = RTTCACHE")

I am gdb challenged.
All looks normal to me / the load module is 7mb+ long so debugging data
must be present...
After reading the help document for print and set I am yelling
Are you willing to help me out just a little?

How might I validate the length of the table and maybe look at the
variable "tp" in the for loop within tcpcmd.c where the printing is
done.  (yes - I could add tp to the print statement and re-compile but
isn't gdb easier?  maybe?)
> Barry, perhaps you can eloborate on that one (if you want that is).
Barry - don't hurt yourself laughing...
> Maiko

de [George (Skip) VerDuin] K8RRA k

More information about the nos-bbs mailing list