[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: Ethernet congession
I wasn't aware or had forgot that we sent that version to anyone.
We have had the same problem at Axis but thanks for investigating
it this far (I will soon investigate why it happens). I will send
you a version that works later today.
From: Zhuorong Deng [mailto:firstname.lastname@example.org]
Sent: Tuesday, January 20, 2004 4:57 PM
To: email@example.com; dev-etrax
Subject: Ethernet congession
We used to have compatibility problem on our hardware
using ETRAX100LX with 100MB Ethernet, AXIS sent us a
ethernet driver (ethernet.c) which fixed the problem.
However, with the new driver, if we have more than one
(2-4) units working on the same 10MB ethernet, every
one sends out about 2.5-3MB data to separate clients
using UDP, the units will fail one by one on data
sending. After this happens, all the units can not be
ping/ftp and the network is also jammed (meaning the
other nodes on the same network can not communicate to
each other). After removing the units from the
network, network functions on other nodes start
working again. The failing unit does not have any
error indication from the console, and need to be
power cycled to be recovered. We wounder whether there
is some collision detection problem with the new
driver so that every unit is keep sending out data.
The original driver (ethernet.c) does not have the
problem, but will not work with 100MB in our hardware.
Can we have a fix for this?
We use SDK 2.1.10 and Linux 2.4.19. We replace the
ethernet.c from AXIS later. I have attached the
ethernet.c.new for your refernce.