John Kehr

The following code is working for the BACnet Test Suite in the WritePropertySupport Test which reads and writes to the BBMD BDT. I don't know if it actually encodes everything correctly yet, but it correctly returns the empty BDT. When I figure out if it's working with BBMD stuff I'll post back here. I'll also add information about the FDT. It's also currently empty though. For right now though test is passing which means a bunch of other tests can run. static int bdt_encode_broadcast_table(uint8_t...

Using the above code except with the check for bdt_entry->valid it only encodes 1 entry instead of 4, but still malformed. Wireshark shows the IP as 00 00 00 00. BACnet-APDU 77 Complex-ACK readProperty[169] network-port,1 bbmd-broadcast-distribution-table 0000 00 0c 29 6d 8d c5 00 23 50 03 ff fe 08 00 45 00 ..)m...#P.....E. 0010 00 3f 01 d8 00 00 64 11 cf f2 c0 a8 01 bf c0 a8 .?....d......... 0020 01 d4 ba c0 ba c0 00 2b ee d5 81 0a 00 23 01 00 .......+.....#.. 0030 30 a9 0c 0c 0e 00 00 01 1a 01...

I have dug into this and tried many things on both the BDT and FDT. I've tried to incorporate the encoding functions in the bvlc.c in combination with the opening/closing tags to no avail. I've tried numerous combinations of matched opening/closing tags using the above encoding. Once I worked on this and understood the purpose of the tags it all made a lot more sense. Why I can't get this to work is the mystery. It's also perplexing the actual functions which appear should be involved in the encoding...

So even though this isn't a BBMD in order to be listed it has to provide minimal support. All of which makes sense because of network complexity these days. Using the above algorithm changed the error from m1115 to: m1904 todo - create suitable subclasses. I think I should have some time to dig into this more today. Here's the code I currently am using which is just a helper function in the netport.c file at the moment. static int bdt_encode_broadcast_table(uint8_t *apdu) { int len = 0; int apdu_len...

I'll work through all of this and let you know, but I have a higher level question I haven't been able to clarify. In this thread you stated: BTL Listed Devices must support either BBMD or be able to register as a foreign device. This device isn't a router but does need to be BTL listed so I was enabling BBMD. Would it be better to only be able to register as a foreign device? I haven't been able to find a clear answer on what a circuit controller should actually have in regards to these two options...

I'm currently going through the BACnet test suite for our new product and the netport in the BACnet stack doesn't encode either the BDT or the FDT. The code is: #if defined(BBMD_ENABLED) case PROP_BBMD_ACCEPT_FD_REGISTRATIONS: apdu_len = encode_application_boolean(&apdu[0], Network_Port_BBMD_Accept_FD_Registrations( rpdata->object_instance)); break; case PROP_BBMD_BROADCAST_DISTRIBUTION_TABLE: case PROP_BBMD_FOREIGN_DEVICE_TABLE: rpdata->error_class = ERROR_CLASS_PROPERTY; rpdata->error_code = ERROR_CODE_READ_ACCESS_DENIED;...

Disregard this. It was the TSM_List which was taking up all the memory. It really was defaulting to 400K though.

I guess that masked the problem until the service handlers started to be more active and now it's making the buffer even larger when I'm compiling after adding the /basic/npdu/npdu.c file. I see some FIXME for creating the buffer in the tsm.c. Has anyone seen a compiler make a buffer unexpectedly and inanely large. I'm thinking that the way BACnet is doing things is highlighting a compiler problem for the PIC.