This is the mail archive of the
ecos-discuss@sources.redhat.com
mailing list for the eCos project.
Re: Prototyping with the synthetic target.
- From: Chuck McParland <cpmcparland at lbl dot gov>
- To: Andrew Lunn <andrew dot lunn at ascom dot ch>
- Cc: ecos-discuss at sources dot redhat dot com
- Date: Thu, 01 Aug 2002 10:27:14 -0700
- Subject: Re: [ECOS] Prototyping with the synthetic target.
- References: <F626113795D3EB4482E5ACFBD934651234C66B@mailhost.stratalight.com> <3D40FE8E.A92D859E@ascom.ch> <3D45981F.D75077F7@lbl.gov> <20020730071922.GB3295@biferten.ma.tech.ascom.ch>
Andrew,
Thanks for the note.....I finally found the synth doc pages under
/arch and, in combination with your note, things have become
much clearer. I think my next cycle through the discuss archives
will be more fruitfull.
You mentioned using sockets at the hal_sys level. I assume you
had to link with the Linux libraries to get the socket calls. Wouldn't this
cause conflicts with the ecos synthetic lib?
Thanks,
Chuck McParland
Andrew Lunn wrote:
> On Mon, Jul 29, 2002 at 12:31:43PM -0700, Chuck McParland wrote:
> > Prototyping with the synthetic target.
> >
> > I'd like to do some application prototyping and simulation
> > work using the synthetic target and would appreciate
> > any pointers from those who have gone through this
> > before.
> >
> > First task is to be able to pass messages between an
> > ecos application running inside the synthetic target and
> > a second linux process on the same machine. Local
> > sockets would be ideal; but it's not clear whether the
> > network stack works on the synthetic target.
>
> The network stack works as a network stack without any device drivers
> (except loopback). There is no connection to the host OS's network
> stack, so you cannot use it for IPC between eCos and host OS
> processes. You can use it for IPC between eCos threads.
>
> > Named
> > pipes are also a possibility. Has anyone used them?
>
> Yes. You have access to the Host OSs system calls using the hal_sys
> backdoor. We have used both the TCP/IP stack, Unix sockets, and fifos,
> using the back door. If you look back in the archive for ecos-discuss,
> you can see the answers i've given to other people who have asked this
> question.
>
> > Secondly, could anyone point me towards further
> > info on the synthetic target? I've been looking through
> > the hal/synth tree and can't quite work out the relationship
> > between the ecos system calls and the linux calls.
>
> There is none really. The synth hal just uses a few limited host
> system calls to implement a timer interrupt and the HAL diagnostic
> tty. Things like eCos fileio uses eCos's fileio code. You can add
> eCos's RAM FS to access eCos RAM filesystems. You can also do the same
> with ROM FSs. There is no access to the host file system using
> fileio. If you want to manipulate host files you need to use the
> backdoor to the system calls.
>
> My experiance of using the synth target for application prototyping is
> good. It needs some extra effort, simulating your target hardware, but
> i find the debug environment much faster to use than real hardware.
>
> Andrew
>
> --
> Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
> and search the list archive: http://sources.redhat.com/ml/ecos-discuss
--
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss