SCTP Discussions

Text archives Help


[SCTP-Discussion] sctp on vxWorks


Chronological Thread 
  • From: Opitz Derek-DOPITZ1 <Derek.Opitz AT motorola.com>
  • To: "'discussion AT sctp.de'" <discussion AT sctp.de>
  • Subject: [SCTP-Discussion] sctp on vxWorks
  • Date: Thu Dec 6 20:42:00 2001
  • List-id: SCTP Discussions <discussion.sctp.de>

We have succesfully ported siemens SCTP to vxWorks, but we made a number of
changes and ran across a few issues.

A guy on our team who did most of the porting is coming up with a list of
changes we made, plus a few issues we came up with if anybody is interested
in them.

Here is one issue that has more to do with the compiler than with the OS.
We are running all this on a G4 processor (PowerPC 7410), so we use the GNU
C compiler for PowerPC.

Here is a snippet of code:


if ((ExitCode = DLL_GetCurrentRecord(rbuf->frag_list, &frag)) !=
DLL_NORMAL)
DLL_error_log(ERROR_MAJOR, ExitCode);
----> frag.stop_tsn = frag.stop_tsn++;
event_logiii(VVERBOSE,
"Updating last fragment frag.start==%u, frag.stop==%u,
tsn=%u",
frag.start_tsn, frag.stop_tsn, ch_tsn);
ExitCode = DLL_UpdateCurrentRecord(rbuf->frag_list, &frag);
rbuf->new_chunk_received = TRUE;
return FALSE;

The line I marked above (from recvctrl.c, line 415) does different things
under Solaris GNU C compiler than it does under PowerPC GNU C compiler.
Under Solaris, it increments frag.stop_tsn. Under PowerPC it DOESNT
increment frag.stop_tsn.

I think the line is a bug in either case, but it just so happens to work
under Solaris.

Derek Opitz
Motorola Cellular Infrastructure
phone: 817-245-6703
pager:
2456703 AT skytel.com




Archive powered by MHonArc 2.6.19+.

Top of page