This is the mail archive of the
ecos-discuss@sources.redhat.com
mailing list for the eCos project.
Network characterisation with nc_test_master.c and nc_test_slave.c
- From: "Richa Jain Mehta" <richa at cradle dot com>
- To: <ecos-discuss at sources dot redhat dot com>
- Date: Fri, 1 Aug 2003 10:21:43 +0530
- Subject: [ECOS] Network characterisation with nc_test_master.c and nc_test_slave.c
Hello,
I've been trying to find out the performance of eCos on the network using
the network characterisation in
nc_test_master.c and nc_test_slave.c. I cant get this test to run.
If I run the tests as they are, with UDP_ECHO being tried first, there is
invariably a crash - debugging it shows it happens somewhere in recvfrom().
If I modify nc_test_master.c so that UDP_SEND is called first, UDP_RECV
second and UDP_ECHO third, this crash happens after UDP_RECV.
Has anyone come across problems like this?
A few other questions:
1. In both files, every time select() is called, there is a check that if
the return values <=0, it is an error. However 0 <NOERROR> is a valid
value.......
2. What are the standard ways to measure the eCos net stack performance? Is
there any other way apart from these network characterisation tests?
I compiled and ran nc_test_master.c on a linux machine. nc_test_slave.c runs
on my target board.
Also, due to some limitations I am still using eCos v 1.3....
Are there any known bugs in 1.3 that may cause this behaviour?
ps: Can someone tell me whether it is possible to port only the netstack
from eCos 2.0 to the version I have? Any estimates on how much effort it
would take??
Thanks and regards,
Richa.
--
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss