Re: [PATCH] FlexRay: Allocate numbers for FlexRay communication

[Michael Arndt <flexray4linux@scriptkiller.de>]

On Thu, Apr 30, 2009 at 07:28:01PM +0200, Oliver Hartkopp wrote:
> > I do not know why ETH_P_CAN has been chosen to be 0x000C
> 
> perhaps because of 'C'AN ?

Ok, thanks for the hint. So I should maybe choose 0x000F :)

> - What is the technical background?
> - Is there any documentation?

My goal is to bring (low-cost) FlexRay communication to Linux. There is
some documentation of the project which I will hopefully put up soon.

> - Do you need realtime?

Would be nice, but will probably be hard to do. Currently only sending/
receiving in pre-configured message buffers is possible, so the real
low level stuff will be handled in hardware (in the FlexRay Communication
Controller [CC]).

> - What hardware is supported?

See below. But the major goal is to provide a common framework for
all hardware, similar to the SocketCAN implementation.
Unfortunately I do not have several thousand Euros to spend for
commercially available solutions [1, 2]. I also bet their drivers
(if available at all for Linux; seems that only TZM has one)
will be incompatible with each other and are closed source.

> I bet there's no Flexray hardware on the market that can be attached on a
> serial line like there are for low-cost CAN adaptors, right?

Actually, I have a working prototype of such an (limited) adapter lying on
my desk right now.

Regards,

Michael


[1] http://www.vector.com/
[2] http://www.tzm.de/