[etherlab-dev] ARM9 (NetX) EtherCAT-enhanced Network Driver

Florian Pose fp at igh-essen.com
Wed Jan 30 15:05:18 CET 2008


Hi Thomas,

On Tue, Jan 29, 2008 at 02:16:25PM +0100, Thomas Elste wrote:
> I've got some questions, additions and/or remarks concerning the
> EtherCAT-Master. As there are serveral things I would like to
> address, I think it is best to start several threads. So don't
> be suprised at this bunch of mails of mine. Things, I always
> forgot to post until now, just kept coming at me, while writing
> the first mail, :).

ok, here we go! ;-)

> So, lets start with some good news: I was able to port the master
> to an ARM CPU ... the NetX 500 chip from Hilscher. This is
> an Arm926ejs with some special abilities concerning realtime
> ethernet. Don't get excited here ... as Hilscher likes to keep
> their software closed, it's not possible to use these fancy
> hardware stuff easily (or at all) in an open source environment.
> The NetX board was just the nearest board in reach on my desk and
> it comes shipped with a nice already preempt-patched Linux which
> I thought would come in handy at some point.
> The NetX EtherCAT network driver and the necessary autoconf/-make
> changes are attached as a patch (aclocal/automake/autoconf-run is 
> needed after applying).
> I've tested this patch on 2 Boards, the NXDB500-SYS (the Mercedes
> under the NetX-Eval Boards (as I was told by Hilscher, :))) and the
> NXHX500. Performance seems to be ok for a 200 MHz Arm9. I was able
> to run a little task (without heavy calculations, though) using 5
> slaves at 5 KHz without any problems. As there is no RTAI available
> for the NetX at the moment, I've been using high resolution timers
> from the preempt patched Linux in the task, which worked out verly
> well.

Sounds good! Thanks a lot for the patches. I'm currently applying them
to the configuration files and I have a question about the kernel
versioning you used: The driver files are named
netx-eth-2.6.20-rt1-netx1-ethercat.X. You matched the version in
configure.ac against the $kernelrelease instead of the $linuxversion
(without the '-rt1-netx1'). Do you have a special reason for this?
Sorry, I don't know the netx hardware/software very well...

-- 
Best regards,
Florian Pose

http://etherlab.org



More information about the etherlab-dev mailing list