public inbox for sid@sourceware.org
 help / color / mirror / Atom feed
From: "Frank Ch. Eigler" <fche@redhat.com>
To: Imran Shafiq <m_imranshafiq@yahoo.com>
Cc: sid@sources.redhat.com
Subject: Re: embedded system emulator help
Date: Wed, 09 Mar 2005 20:05:00 -0000	[thread overview]
Message-ID: <20050309200520.GD22133@redhat.com> (raw)
In-Reply-To: <20050307170632.21914.qmail@web51509.mail.yahoo.com>

[-- Attachment #1: Type: text/plain, Size: 2230 bytes --]

Hi -


On Mon, Mar 07, 2005 at 09:06:31AM -0800, Imran Shafiq wrote:

> [...]  automotive subsystems are comprised of [ECU] Electronic
> Control Units (e.g fujitsu mb90f598, Powerpc mpc 566) connected
> through a BUS system usually CAN.  [...]  So does that means I have
> to implement the whole instruction set for every ECU??

That's right.  If you want to simulate software compiled for a given
platform, and a model of that platform is not already available, then
there is some coding to do.  I believe there is no PowerPC model in
sid at the moment.


> I also have to simulate UART, Timer etc

Some UART and timer models are already included in sid.  They may be
useful as is, or may serve as a basis for derivation.


> It should be possible to check the data on the bus So that would
> mean I have to implement the CAN protocol??

Well, one doesn't simulate a "protocol" per se, rather entities that
perform it.  You would need to pick a level of abstraction at which
the simulator should model a CAN system.  This could be at the
controller level, wire level, perhaps even some software level. This
allows you to trade of hardware fidelity versus modeling simplicity.


> Simulator should also report stack overflows or memory
> limitations.

Some of this is often best done by instrumentation added by the
compiler (check gcc's -fstack-limit options).  Or a simulator could
also model "missing" physical memory that would trigger bus errors
upon attempted access.  Think of the latter as mprotect(PROT_NONE).


> IS all this possible with SID considering the fact that i have to do
> it in 5 months :) .

Probably, if you pick another CPU architecture that is already
well-supported in sid, such as ARM.  If nearly all of your software 
is to be written in C, you could argue that the difference in the
processors is almost immaterial.


> Where can i start with an example. any pointers?

I would start with one of the classic sid configurations: running an
application linked with the eCos RTOS on a model of a board such as
the ARM PID.  Analyze the application, eCos linkages, memory and
peripheral layout, and the corresponding sid configuration.


- FChE

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

      reply	other threads:[~2005-03-09 20:05 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2005-03-07 17:06 Imran Shafiq
2005-03-09 20:05 ` Frank Ch. Eigler [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20050309200520.GD22133@redhat.com \
    --to=fche@redhat.com \
    --cc=m_imranshafiq@yahoo.com \
    --cc=sid@sources.redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).