This is the mail archive of the ecos-discuss@sources.redhat.com mailing list for the eCos project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: ecosynth using real eth : rawether is not called when no io is selected.


This can explain why it doesn't work properly .
The reason seems to be because I was running the app without the IO 
(it mademy computer crash the first couple trials :-)

I didn't understand first that doing so would stop ecosynth from accessing 
its driver altogether. 
Is this the desired behaviour ?

Philippe
 

On Friday 14 February 2003 06:12 pm, Philippe Moutarlier wrote:
> I guess my problem might arise from rawether not staring properly ?
> I tried to see it using "ps axu" while running the app but could not
> identify it.
>
> I could definitely see the process trying to open the eth port though.
> I can verify that it fails if the device is already in use.
>
> I guess I should try to debug rawether  .. just not quite sure about how to
> do it on the fly with GDB ,,,
>
> Philippe
>
> On Friday 14 February 2003 06:03 pm, Bart Veer wrote:
> > >>>>> "Philippe" == Philippe Moutarlier <Philippe.Moutarlier@attbi.com>
> > >>>>> writes:
> >
> >     Philippe> Ok, I am almost there with my synthetic target. I got
> >     Philippe> the ping_test to compile and run using real eth1.
> >
> >     Philippe> The problem is : it thinks that its MAC address is all
> >     Philippe> "0s" and the ping consistently fails ... of course .
> >
> >     Philippe> The exact same code runs on a i386 PC-Redboot with the
> >     Philippe> right ethernet card. Do I need to have a specific
> >     Philippe> hardware under Linux too (I thought the point was to use
> >     Philippe> Linux's hardware abstraction layer though ..)
> >
> >     Philippe> still missing a little piece ... but getting closer :-)
> >
> > You should not need specific hardware. When synthetic target eCos
> > tries to use an ethernet device that should cause a rawether process
> > to be spawned. The rawether program uses standard Linux ioctl() calls
> > to determine the MAC address, and passes that information back to
> > eCos.
> >
> > Unfortunately I am not sure why you are seeing a MAC address of all
> > 0's, I have not seen that before. The code to look at is in the
> > rawether program, devs/eth/synth/ecosynth/current/host/rawether.c, and
> > in particular the real_init() function which will be invoked with the
> > name of the selected Linux ethernet device (presumably eth1). It is
> > the SIOCGIFHWADDR ioctl() that should give the MAC address.
> >
> > Bart

-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]