Upgrade hangs at "CEL stick" step

[eric72777]

I was unable to upgrade a HUSBZB-1 (5.4.1-192) using the docker image. Restarting the computer and unplugging and re-plugging the stick didn't fix the issue.

sudo docker run --rm --device=/dev/ttyUSB1:/dev/ttyUSB1 -it walthowd/husbzb-firmware bash

cd /tmp/silabs

./ncp.py scan
Connecting to.. /dev/ttyUSB1 57600 True False
{"ports": [{"stackVersion": "5.4.1-194", "deviceType": "zigbee", "pid": "8A2A", "port": "/dev/ttyUSB1", "vid": "10C4"}]}
./ncp.py flash -p /dev/ttyUSB1 -f ncp-uart-sw-6.6.5.ebl
Restarting NCP into Bootloader mode...
CEL stick
^CTraceback (most recent call last):
File "./ncp.py", line 459, in
flash(args.port,args.file)
File "./ncp.py", line 310, in flash
ser.readline() # read blank line
File "/usr/local/lib/python2.7/dist-packages/serial/serialposix.py", line 565, in read
ready, _, _ = select.select([self.fd, self.pipe_abort_read_r], [], [], timeout.time_left())
KeyboardInterrupt

[ajmath]

I'm seeing the same issue.

[MikePaer]

Same problem here. Also, running the docker run main command just hangs. Has anyone figured this out?

[sreknob]

The next step in the script is the initialization of the serial port which might be failing... for me re-plugging fixed it.
If you can open up your stick, try the firmware recovery instructions in the readme as a work around. It's actually quite easy to do it that way.
The other thing you could try is another linux environment. If you clone the repo and install python 2.7, pyserial and xmodem, you can use the python script outside of a container. I've done this on just a raspberry-pi rather than running the docker image.

[MikePaer]

Thanks for the pointers. Tried running the Python script outside the container and on a different machine, neither worked. I didn’t try the firmware recovery, and instead replaced my HUSBZB-1 device assuming it was defective. The new device worked right out of the box with no issues. I think I just had a lemon.

[C2BB]

Thanks @sreknob your tip on unplugging the stick solved the issue for me.

[BHSPitMonkey]

Having the same issue today with one of my HUSBZB-1 devices. Note that I am commenting out xonxoff=True as a workaround for the other issues reported in different threads (this has helped me flash another device previously), but in this case the failure occurs either way.

Here's some strace output indicating what's going on behind the pyserial readline() call that never finishes (never sees a \n appear):

husbzb-firmware/ncp.py

Lines 296 to 299 in 12de988

	# EZSP protocol initialization
	if ezspV8Init(ser)!=NO_ERROR and ezspV7Init(ser)!=NO_ERROR and ezspV6Init(ser)!=NO_ERROR and ezspV5Init(ser)!=NO_ERROR and ezspV4Init(ser)!=NO_ERROR:
	print('NCP no ZigBee Ack. Please try again.')
	sys.exit(1)

write(3, "\32\3008\274~", 5)            = 5
read(3, "\32\301\2\v\nR~", 7)           = 7
write(3, "\0B!\250\\,\240~", 8)         = 8
read(3, "\1B\241\250P(\5\346\177b~", 11) = 11
write(3, "\201`Y~", 4)                  = 4
write(3, "\32\3008\274~", 5)            = 5
read(3, "\32\301\2\v\nR~", 7)           = 7
write(3, "\0B!\250S\335O~", 8)          = 8
read(3, "\1B\241\250P(\5\346\177b~", 11) = 11
write(3, "\201`Y~", 4)                  = 4
write(3, "\32\3008\274~", 5)            = 5
read(3, "\32\301\2\v\nR~", 7)           = 7
write(3, "\0B!\250R\315n~", 8)          = 8
read(3, "\1B\241\250P(\5\346\177b~", 11) = 11
write(3, "\201`Y~", 4)                  = 4
write(3, "\32\3008\274~", 5)            = 5
read(3, "\32\301\2\v\nR~", 7)           = 7
write(3, "\0B!\250Q\375\r~", 8)         = 8
read(3, "\1B\241\250P(\5\346\177b~", 11) = 11
write(3, "\201`Y~", 4)                  = 4
write(3, "\32\3008\274~", 5)            = 5
read(3, "\32\301\2\v\nR~", 7)           = 7
write(3, "\0B!\250P\355,~", 8)          = 8
read(3, "\1B\241\250P(\5\346\177b~", 11) = 11
write(3, "\201`Y~", 4)                  = 4

husbzb-firmware/ncp.py

Lines 296 to 299 in 12de988

	# EZSP protocol initialization
	if ezspV8Init(ser)!=NO_ERROR and ezspV7Init(ser)!=NO_ERROR and ezspV6Init(ser)!=NO_ERROR and ezspV5Init(ser)!=NO_ERROR and ezspV4Init(ser)!=NO_ERROR:
	print('NCP no ZigBee Ack. Please try again.')
	sys.exit(1)

write(3, "}1C!'Un\220~", 9)             = 9

husbzb-firmware/ncp.py

Lines 309 to 311 in 12de988

	# Read response bytes
	ser.read(8)
	ser.close()

read(3, "\22C\241'T\2539~", 8)          = 8

husbzb-firmware/ncp.py

Lines 316 to 326 in 12de988

	# Open new connection
	ser = serial.Serial(
	port=port,
	baudrate=115200,
	parity=serial.PARITY_NONE,
	stopbits=serial.STOPBITS_ONE,
	xonxoff=True
	)
	# Burn the prompt
	ser.write(b'\x0A')

write(3, "\n", 1)                       = 1

husbzb-firmware/ncp.py

Lines 327 to 330 in 12de988

	# Boot menu - legacy or Gecko bootloader?
	# Gecko BL verion at 2nd line
	# Legacy BL verion at 3rd line
	ser.readline() # read blank line

read(3, "\230", 1)                      = 1
read(3, "`", 1)                         = 1
read(3, "\0", 1)                        = 1
read(3, "\30", 1)                       = 1
read(3, "\30", 1)                       = 1
read(3, "\236", 1)                      = 1
read(3, "x", 1)                         = 1
read(3, "\376", 1)                      = 1
read(3, "\206", 1)                      = 1
read(3, "\376", 1)                      = 1
read(3, "\6", 1)                        = 1
read(3, "\340", 1)                      = 1
read(3, "\6", 1)                        = 1
read(3, "\340", 1)                      = 1

(At this point the script hangs indefinitely and there's no more reads/writes. Pyserial waits indefinitely for a newline character but only receives the unexpected data shown in the strace above.)