Over the air updates

[ToSa]

@Zeph
yes, that's what I meant - you might not even need any fork of the bootloader itself and just a slight adjustment on the controller end - because the nodeID is contained in the packet (not in the payload but in the header as sender address) so you have all you need for your setup

[Zeph]

@ToSa
The other half is testing inequality between the computed CRC of the application firmware in PROGMEM, with the CRC of the available replacement (rather than comparing for higher version number).

An example use case of the ability to load arbitrary new code into any given node. If I was diagnosing some kind of interference, I might temporarily replace the sensor firmware in some nodes (of varying node-type) with a custom radio test firmware, then later restore each with it's original sensor node firmware.

Suppose we have:

 node 5, node type 17, version 2, PROGMEM CRC 0x4567  // attic
 node 6, node type 3, version 5, PROGMEM CRC 0xABCD  // crawlspace
 node 7, node type 3, version 5, PROGMAM CRC 0xABCD // living room

And I want to temporarily replace the firmware in node 5 and 6, but keep 7 still running as a sensor.

I make RF test code available on the server, with CRC 0x7E57. This is not type 17 or type 3.

I edit the server's table of firmware assignments:

node 5, 0x7E57
node 6, 0x7E57
node 7, 0xABCD  // unchanged

This causes node 5 and 6 (formerly of different types) to load the test firmware when reload is triggered.

Then when testing is done, I edit the table back:

node 5, 0x4567   // back to its old type and version
node 6, 0xABCD  // back to the same type and version as node 7
node 7, 0xABCD  // still unaffected

This causes the normal sensor firmware (type and version) to be loaded back in on the next reload.

There could be more than just a CRC to identify the firmware (in order to avoid the birthday paradox), this is just an example.

An alternate use case is loading in my Halloween firmware to the front yard nodes (but not other nodes) for a week or two, then back..

Or an beta version of type 3, version 6, which I'd like to load on some type 3 nodes for in-situ testing (eg: in the crawlspace), but not all of the type 3 nodes because I want most of the system to continue functioning normally while I test. If the beta is bad, I may revert the test nodes to version 5; once the new version is good, I may convert all type 3 nodes to version 6.

These are some of the reasons I'd like to be able to use OTA programming of any arbitrary firmware into any given node, without being constrained to:

  Only upgrades of the same node type
  Only upgrades to higher version numbers
  Only upgrades of all nodes of the same type or none

And so that's why inequaity testing of the PROGMEM signature on a per-node basis is attractive, not just testing for a higher version number. For similar complexity, we can upgrade to a higher version number, downgrade to a different version number, or change the node type back and forth.

The type and version dynamics (which certainly IS a common use case) can be handled on the server. For example, the server can know what type every node is (kind of a good idea anyway), and can change the node -> signature entry for every node of type 3 to the signature of the next version, and then let it proceed as above to get them all updated. But that's just one option, centrally controlled.

[ToSa]

@Zeph
from the MyOtaBootloader.c:

if (firmwareConfigResponse->version == fc.version)
	if (firmwareConfigResponse->blocks == fc.blocks)
		if (firmwareConfigResponse->crc == fc.crc)

so as long as you send the same version / blocks / crc back to the node as what iscurrently installed, no update is started. As soon as one of the three elements differs an update is loaded. It's completely in control of the server if (and which) firmware is bootloaded.

[Zeph]

@ToSa

OK, so version is tested for != rather than for > ? Downgrading is OK?

And CRC is used as well (and block count?) where CRC is based on what's in PROGMEM now?

Cool.

Then I think all that would be needed is for the server to be able to potentially feed back a different firmwareConfigResponse to each node. In my above example (which has been edited for clarity recently BTW, so re-read it), node 6 could receive a different response than node 7 (even tho they both have the same type initially). And thus nodes 5 and 6 (but not 7) could be told to load the test firmware and then later to go back to the old version. Etc.

Is that correct?

It would be a nice enhancement if we could query the node for the CRC (and block count?) of the current PROGMEM, just to help the server stay in sync with what's out there (eg: after a node joins the network). That could be done in the application code, so we don't even have to invoke the bootloader. Then the server could figure out which nodes need to be bootloaded and trigger just those to go into the bootloader (possibly one at a time). These two together support what I call push dynamics.

[Damme]

@Zeph I've been working on a read / write eeprom address thing in MQTT to be able to reset a node and stuff. But it seams there are more usage for it then. This might be coded into mysensors instead. (utilizing c_internal or somthing as the protocol is today)

[ToSa]

@Damme said:

@Zeph I've been working on a read / write eeprom address thing in MQTT to be able to reset a node and stuff. But it seams there are more usage for it then. This might be coded into mysensors instead. (utilizing c_internal or somthing as the protocol is today)

Good idea - that would allow to check for current value in normal operation - not just during bootloading.

@Zeph
If you urgently want to have the CRC of the current firmware submitted during bootloading, we can add this as a third parameter to the FirmwareConfigRequest message. Actually I was thinking about getting rid of request/response and use the same format for both which would mean crc would be included anyways.

[Damme]

This post is deleted!

[Damme]

I deleted my last message because I though I made a big mistake..

I've been working on a SD <-> OTA loader node, and got most of if working but got stuck on the last piece which is communication.. (i'll release it then I'm finished Ive made a small change in myotabootloader, add on line ~156 msg.destination = OTAGATEWAY; to configure custom ota address)

I cant figure the following out:
Just ignore contents of packages. not relevant.

• Node: (Ota<->sd loader)

  read: 34-0-254 s=255,c=4,t=0,pt=6,l=4:FFFFFFFF
  send: 254-254-0-34 s=255,c=4,t=1,pt=8,l=4,st=ok:0100020000304200
  

• GW:
  0;0;3;0;9;read: 34-34-0 s=255,c=3,t=7,pt=0,l=0:
  0;0;3;0;9;send: 0-0-34-34 s=255,c=3,t=8,pt=1,l=1,st=ok:0
  0;0;3;0;9;read: 34-34-0 s=255,c=3,t=7,pt=0,l=0:
  0;0;3;0;9;send: 0-0-34-34 s=255,c=3,t=8,pt=1,l=1,st=ok:0
  0;0;3;0;9;read: 34-34-254 s=255,c=4,t=0,pt=6,l=4:FFFFFFFF
  0;0;3;0;9;send: 34-0-254-254 s=255,c=4,t=0,pt=6,l=4,st=ok:FFFFFFFF
  0;0;3;0;9;read: 254-254-34 s=255,c=4,t=1,pt=6,l=8:0100020000304200
  0;0;3;0;9;send: 254-0-0-34 s=255,c=4,t=1,pt=6,l=8,st=fail:0100020000304200

• OTA bootloader:
  Go
  <- 34,34,0,2,3,7,255,
  <- 34,34,0,2,3,7,255,
  -> 0,0,34,10,35,8,255,0,
  <- 34,34,254,34,196,0,255,255,255,255,255,

What am I missing? package from 254 to 34 wont get delivered.
I've also noticed that then 254 tries to send, it wont receive the next transmitted message from OTAbootloader. the next thereafter is received.

[ToSa]

@Damme
I need to better understand the setup to think about what's going on. My take from the above:

You have three nodes:

• Gateway (address 0)
• SD OTA Loader Node ?!? (address 254)
• Sensor Node (address 34)

Is that right?

[Damme]

@ToSa Yes, And I think I figured it out.. I by mistake changed BROADCAST_ADDRESS to GATEWAY_ADDRESS in the bootloader then I was playing around. Testing the correct version now.. :)

[ToSa]

@Damme
interesting setup :+1:
to make it work with non-static addressed nodes you should probably keep the destination set to GATEWAY_ADDRESS for the REQUEST_ID call and only change afterwards.

Never mind - looking at the line number you mentioned that's probably what you did :)

[Damme]

@ToSa Now I remember why I changed some things in there. (broadcast to gateway)

From the beginning I had problem getting it to talk with the GW. It only sends out
<- 255,255,255,2,3,7,255, and gets no response, The GW tries to send but fails. (wierd..) (I dont have any relay nodes)

This is with no modifications at all.

0;0;3;0;9;read: 255-255-255 s=255,c=3,t=7,pt=0,l=0:
0;0;3;0;9;send: 0-0-255-255 s=255,c=3,t=8,pt=1,l=1,st=fail:0
0;0;3;0;9;read: 255-255-255 s=255,c=3,t=7,pt=0,l=0:
0;0;3;0;9;send: 0-0-255-255 s=255,c=3,t=8,pt=1,l=1,st=fail:0
other packages send out works just fine.. (To other nodes)

and the OTA bootloader can receive other packages
Go
<- 255,255,255,2,3,7,255,
-> 23,23,0,42,225,1,11,205,204,90,66,1,
<- 255,255,255,2,3,7,255,
<- 255,255,255,2,3,7,255,
<- 255,255,255,2,3,7,255,

(from a temp / hum node)

Any ideas how to fix this?

[Damme]

@ToSa I finally figured out why my OTA bootloader didn't read any answers from my GW (Both on I_FIND_PARENT and I_ID_REQUEST) - The answers came to quick! First I tried hardcode a delay 125ms on the GW and it worked, so I changed the code on send write to the following and now all messages arrive. Been testing it for a couple of reboots now. I'm using 5v (at 3.3v) and 16MHz
edit; noticed it misses packages sometimes now but not close to 100% like before, more like 5% now. I'llinvestigate futher then I'm trying to upload data.

  static uint8_t sendAndWait(uint8_t reqType, uint8_t resType) {
  	msg.type = reqType;
  	for (uint8_t i = 0; i < 10; i++) {
  		sendWrite(msg);
  		for (uint8_t j = 0; j < 20; j++) {
  			for (uint8_t j = 0; j < 100; j++) {
  				uint8_t pipe;
  				boolean avail = available(&pipe);
  				wdt_reset();
  				if (avail && pipe<=6) {
  					read(rmsg.array,pipe);
  					if(!(mGetVersion(rmsg) == PROTOCOL_VERSION))
  						continue;
  					if (rmsg.destination == nc.nodeId) {
  						if (mGetCommand(rmsg) == C_INTERNAL) {
  							if (rmsg.type == I_FIND_PARENT_RESPONSE) {
  								if (rmsg.data[0] < nc.distance - 1) {
  									nc.distance = rmsg.data[0] + 1;
  									nc.parentNodeId = rmsg.sender;
  									eeprom_write_byte((uint8_t*)EEPROM_PARENT_NODE_ID_ADDRESS, nc.parentNodeId);
  									eeprom_write_byte((uint8_t*)EEPROM_DISTANCE_ADDRESS, nc.distance);
  								}
  							}
  						}
  						if ((mGetCommand(rmsg) == mGetCommand(msg)) && (rmsg.type == resType))
  							return 1;
  					}
  				}
  				delaym(1);
  			}
  		}
  	}
  	return 0;
  }

[Damme]

I had to put my project in the trash bin.. There is not enough RAM in the atmega328 to fit mysensors and SD-lib :) Tried 3 different versions..Too bad..! I could only transmit one package before SRAM got overrunned.

[Yveaux]

@Damme Now do you get why I abandoned the MQTT implementation on the ATMega itself? ;-)

[Zeph]

@Damme said:

I had to put my project in the trash bin.. There is not enough RAM in the atmega328 to fit mysensors and SD-lib :) Tried 3 different versions..Too bad..! I could only transmit one package before SRAM got overrunned.

If you want to stay with the AVR:

ATMega1284 based: http://lowpowerlab.com/moteino/#whatisitMEGA $20+shipping. This can add a RF69* radio, but you could instead (or also) attach a nRF24L01+

ATMega2560 based: http://www.ebay.com/itm/121391548557 $15 shipped This has even more lines broken out than the Arduino Mega2560. (If you don't mind the larger form, Arduino Mega2560 clones start under $14 shipped).

Or you can switch to an embedded ARM system. Teensy 3.1 for $17+ship. STM32F103 board on eBay for $7. DUE clone on eBay for $18. STM Nucleo from distributors for $11+ship (eMed programmed).

And of course you can use the Raspberry Pi or BeagleBone Black with a faster but power hungry ARM running Linux.

[ToSa]

@Damme said:

I had to put my project in the trash bin.. There is not enough RAM in the atmega328 to fit mysensors and SD-lib :) Tried 3 different versions..Too bad..! I could only transmit one package before SRAM got overrunned.

Did you try not using MySensors but the reduced nRF24 driver version I used for the bootloader itself? If the node does not need to do routing and is expected to only respond to one or two types of messages, that might be an option and is definitely smaller...

[gregl]

Hi @ToSa

Do you know if your OTA Bootloader uses more flash mem on the atmega328 vs the Optiboot bootloader?

I have a sketch which needs Optiboot to fit and really hoping i can one day use your cool OTA stuff.

Looking at your github it seems MyOtaBootloader.hex is 9.36k whereas Optiboot v5,0a the optiboot_atmega328.hex is 1.418k ( but there is also a optiboot_atmega328.lst file at 19.778k - I dont know enough to know if this is part of the bootloader or if its just the source???)

Cheers,
Greg

[Zeph]

OptiBoot fits in 1/2 KiB (the binary on-chip size, not the hex file).
The OTA Bootloader is obviously larger. If you are using close to 31.5KiB on an ATMega328p, it won't fit with any bootloader larger than OptiBoot.
.
Make sure you use the latest compiler - 1.5.7 seems to squeeze harder (smaller binary), and I presume 1.0.6 (which has also been upgraded to a newer compiler) will also do so.

How big does the compiler say your sketch is, at the end of a compile?

[gregl]

@Zeph said:

use the latest compiler - 1.5.7 seems to squeeze harder (smaller binary), and I presume 1.0.6 (which has also been upgraded to a newer compiler) will also do so.

How big does the compiler say your sketch is, at the end of a compile?

I'll check when i get home tonight.

Very interesting about the new compiler. I'm pretty sure im using 1.0.5-r2, and i've recently begun using Atmel studio with Visual Micro addon - just soooo much better when dealing with long sketches!!!

Cheers,
Greg