Wadel wrote:
1) You set the Message Type to RTR in order to send a remote frame to the Network interface. On the initStartTask, you then specify a channel or list of channels to read if a database exists, right?
The code I posted only sets the IsRemote input to 1 if the incoming packet from the real interface has the RTR bit set, and I only added that for the sake of completeness, to show how you would make sure that the packet sent on the virtual interface is identical to the packet received on the real interface. It is, however, highly unlikely that your devices ever send a packet with the RTR bit set, and you should probably just hard-code that to 0 or leave it unwired.
Wadel wrote:
2) Like I mentioned in my previous post. If I want to display, say for example, the voltages, currents and boolean status bits for each of the three different devices on the network, how would you implement it in code?
3) How would you identify each device on the network in the code?
Too many missing details to answer these questions. Do you have to send a request to the device in order to retrieve data, or do the devices stream data constantly? Do you have control over the format in which the devices send data? The only way you'll be able to identify each device is either by arbitration ID, or possibly by information inside the data packet itself (but arbitration ID is preferable to avoid conflicts). If you have no way to assign a unique ID to each device, then you're out of luck - you cannot put them on the same CAN network. However, since messages in a CAN database are defined by arbitration ID, if you do have multiple devices sending the same data on different IDs, you might want to reconsider whether using channels is the best way to handle the data.
Wadel wrote:
4) Would it be better to have a pair of command and response frames (arb IDs) and then create mode dependent channels since there many signals to monitor?
I can't answer this question; again, it depends on your devices. If you (or a coworker) are writing the code that runs on the device as well, then it's up to you to work out the best format. If you don't have any control over the remote devices, then you're going to use whatever the device uses.
Wadel wrote:
5) What sort of design pattern would you use? Parallel loop/master-slave design pattern?
As above - not enough details. If your needs are fairly simple - say, the devices stream data constantly and you just need to chart the data - then a single loop in which you read the most recent point from each channel and add it to the graph might be enough. If your needs are more complex, you might need more loops. It's not very helpful to throw out names of design patterns without considering how they'd be used. If you used a master/slave pattern, what would be the master and what would be the slave? What data would be exchanged between them? Start with the goal and then pick an appropriate pattern. You wouldn't pick up a hammer first, then try to use it for a job that requires a screwdriver; same goes for software architecture.
