Dual-Ported Memory & Shared Memory Support

[Salman]

>> But what I don't understand is the Shared Memory Support. Can someone
>> tell me what was this Shared Memory Support designed for ? and whether I
>> would need it or not for SCI network.

>That probably depends on how you want to use the SCI connections.  Do you
>want to use it for fast message passing, or to combine the physical memory
>of multiple nodes into a combined address space?

I certainly need a way of synchronizing the cluster and the jobs running on
different nodes of the cluster.
I'd like to share application data using the SCI shared memory as well.

At a higher level, I would be sharing RTEMS objects or even jobs/processes
among the nodes. Having said that, I highly doubt I can reach this stage, so
we can totally ignore it.

Simplest way of providing synchronization is to use the shared memory itself
(not very good though).
A more logical way of doing this is to use message passing.
the choice of which, is more dependent on how complex each one is to be
implemented on RTEMS
this is where I need your suggestions and guidelines.
please bear in mind, i'm new to both SCI and RTEMS and time-wise i'd like to
finish my project  by Feb 2003.


>I get the impression from what you wrote that what you actually mean when
>you say "SCI driver" is really "SCI driver plus Linux MPI implementation."

My impression of the SCI drivers is that, it has about twice as much lines
of code as the largest other Linux driver (SCSI driver). But it seems most
of it is related to caching or so, which I hopefully can cut down on.

FYI at the moment I'm trying to figure out the minimal set of SCI drivers
required to provide shared memory on Linux. Then I'm gonna look at SCI
drivers on VxWorks which I hope to be a closer match to RTEMS, and finally
port either the Linux or VxWorks drivers. I have to have a closer look into
the "bsp and driver develoement" doc of RTEMS as well.

Thanks,

Salman