Firmware update via CANbus and FreeRTOS watch dog triggered issue - arduino

Question

I develop a firmware running a ESP32-based custom PCB in a connected battery (hereafter the BATTERY).
Th battery is capable of CANbus connectivity, and I want to take advantage of it to upgrade the firmware.
For this purpose, I took a second PCB (hereafter the UPDATER) which I use to transfer the firmware bin file to the BATTERY.
On the BATTERY, I have FreeRTOS task with a relatively high priority of 9: its responsibility is to wait for the UPDATER incoming message and respond accordingly.
On the other hand, the UPDATER runs a FreeRTOS task that is used to send new firmware bin file chunks of data to the BATTERY. The task also runs at priority 9.
Streamlined exchange protocol is as follows:
UPDATER BATTERY
Start w/ i=0 Start w/ chunkId = 0
| |
loop: |
| |
Take ith 8-byte chunk |
from bin file and |
transmit to battery |
| |
o --------------------------->o
| |
| send current chunkId
| back to UPDATER
| and chunkId++
| |
o< ---------------------------o
|
Read received chunkId
and compare to I
If i != chunkId
then error and stop
Otherwise i++
and 'loop:' till the end
of file is reached
|
Success
The typical firmware bin file size is 1.8MB.
The issue I have, is on the BATTERY, and it is related to the canRx() task for(;;) loop. After a many loops and chunks exchanged, I end up with a watch dog triggered:
E (548436) task_wdt: Task watchdog got triggered. The following tasks did not reset the watchdog in time:E (548436) task_wdt:  - IDLE0 (CPU 0)
E (548436) task_wdt: Tasks currently running:
E (548436) task_wdt: CPU 0: btController
E (548436) task_wdt: CPU 1: IDLE1
E (548436) task_wdt: Aborting.
abort() was called at PC 0x401d105c on core 0ELF file SHA256: 0000000000000000Backtrace: 0x40095ac4:0x3ffbffd0 0x40095d3d:0x3ffbfff0 0x401d105c:0x3ffc0010 0x400913f1:0x3ffc0030 0x401e0289:0x3ffd37d0 0x401e085d:0x3ffd37f0 0x400979f2:0x3ffd3820
My task runs on CORE_1 and the code looks like so:
void canRx(void *arg) {
can_message_t rx_frame;
firmwareUpdater.log("FU: Installing high performance canRx task handler");
for (;;) {
if (can_receive(&rx_frame, pdMS_TO_TICKS(20)) == ESP_OK) { // 1000
// Receive
const unsigned id = rx_frame.identifier;
//#if LOG_LEVEL >= LOG_LEVEL_DEBUG
firmwareUpdater.log("FU: CAN <== RX 0x%04x", id);
//#endif
switch (id) {
// Restart smartbox
// == Reboot a device from the smartbox
case CANBUS_CANID_RESTART:
{
// = What deice must be restarted?
byte device = rx_frame.data[0];
switch (device)
{
// Restart smartbox
case CANBUS_DEVICE_SMARTBOX:
pepsr.restart("Requested by CANbus");
break;
default:
firmwareUpdater.error("FU: asked to restart device %02x which is unknown: ignored");
break;
}
break;
}
// Start (over) update process
case FIRMWAREUPDATE_CANID_START_OF_PROCESS:
{
const unsigned fileSize = (rx_frame.data[0] << 24) | (rx_frame.data[1] << 16) | (rx_frame.data[2] << 8) | rx_frame.data[3];
firmwareUpdater.startProcess(fileSize);
}
break;
case FIRMWAREUPDATE_CANID_FIRMWARE_IDENTIFIER:
firmwareUpdater.sendFirmwareBuild();
delay(1000);
firmwareUpdater.sendFirmwareIdentifier();
break;
case FIRMWAREUPDATE_CANID_ADDCHUNK:
firmwareUpdater.addChunk(rx_frame.data, rx_frame.data_length_code);
break;
default:
firmwareUpdater.log("FU: CAN <== RX 0x%04x | Not an ADD CHUNK", id);
break;
}
}
else {
firmwareUpdater.log("FU: CAN: no pending messages");
}
// === VERY IMPORTANT: Let watchdog know we are still there
delay(20);
}
}
#endif
And the task itself is fired as follows:
log("Installing hi speed RX callback");
xTaskCreatePinnedToCore(&canRx, "CAN_rx", 4096, NULL,
//5 //TASK_PRIORITY_REGULAR_5
9 // TASK_PRIORITY_HI_9
, &xCanRxHandle, tskNO_AFFINITY);
As you can see, I tried to transfer control to the scheduler by playing with:
The task priority by taking it down to 5;
Adding extra delay() in the loop;
Reducing the timeout value of the `can_receive() from 1000 down to 20.
But nothing really works: for instance, by increasing the delay()(point 2), I manage to make the transfer more robust but at the expense of an unmanageable total duration.
My bet is that something is fundamentally wrong somewhere.
I need extra thought and help! Thanks in advance.