Skip to content

Commit

Permalink
Update program.py to workaround an MSPFlasher issue
Browse files Browse the repository at this point in the history 
MSPFlasher seems to silently ignore the second read command in a sequence, so we instead now do one big read command that dumps the full range that includes all the bytes we need rather than dumping the specific smaller ranges. Then we parse out the bytes we need from the resulting dump. More complicated and dumps more bytes than we need, but works.
  • Loading branch information
@bigjosh
bigjosh committed Apr 21, 2023
1 parent f5af464 commit 58bc5a1
Show file tree
Hide file tree
Showing 5 changed files with 579 additions and 565 deletions.
3 changes: 0 additions & 3 deletions programming/device.txt
View file

This file was deleted.

19 changes: 0 additions & 19 deletions programming/fram.txt
View file

This file was deleted.

19 changes: 0 additions & 19 deletions programming/info.txt
View file

This file was deleted.

193 changes: 122 additions & 71 deletions programming/program.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -3,18 +3,27 @@
# sorry it is so ugly, but i wanted it to be pythonic

import subprocess

import time

import shutil

import hashlib
import tempfile
import os

# the following are for airtable stuff
import keyring
import http.client
import json

#MSPFlasher executable name
mspflasher_name = "MSP430Flasher"

#locate executable
mspflasher_exec = shutil.which( mspflasher_name )
#check that it was found
if mspflasher_exec is None:
raise Exception("MSPFlasher executable must be in search path")


# Airtable API host
host = "api.airtable.com"
# This is the base and table ID of the TSL units table
Expand Down Expand Up @@ -66,86 +75,128 @@
print("Goodbye.")
exit(0)

# Create the combined image to burn into the unit that includes both timestamp and firmware
# Create a temp directory for the files we are creating, then create temp files for the firmware image (will auto delete everything when pass finished)
with tempfile.TemporaryDirectory() as tempdir:

# first we create a firmware image to write to FRAM. We do this by combining a timestamp with the compiled firmware

with open('output.txt','wb') as wfd:
# get the current time
t = time.localtime()
image_file_name = os.path.join( tempdir , 'image.txt' )
with open( image_file_name ,'wb') as wfd:

# prepend a time stamp (in TI HEX format)
# 1800 is the begining of "information memory" FRAM.
wfd.write("@1800\n".encode())
wfd.write(f"{t.tm_sec:02d} {t.tm_min:02d} {t.tm_hour:02d} {t.tm_wday:02d} {t.tm_mday:02d} {t.tm_mon:02d} {t.tm_year%100:02d}\n".encode())
# get the current time
t = time.localtime()

# ...and append firmware file into the image file
# note that the firmware comes last becuase the TI tools add a "q" to the end of this file.
with open('tsl-calibre-msp.txt','rb') as rfd:
shutil.copyfileobj(rfd, wfd)

# If you ever need to read the commisioned time out of a unit, you can use the command...
# MSP430Flasher.exe -j fast -r [commisioned_time.txt,0x1800-0x1806] -z [VCC]
# Do note that the timestamp does not have a century field, so you will have to infer what century the unit was commisioned
# in using other factors like how dusty it is.

# Here is the meat where we...
# 1. Grab the device UUID and ID from the MSP430 chip. This will let us keep a record of the serial number and what version of the chip this unit has.
# 2. Erase the FRAM.
# 3. Burn the combined image into the unit.

# -j fast means use the fastest clock speed so programming will go as quickly as possible (it still takes a couple seconds)
# -e ERASE_MAIN prepares the FRAM to recieve the firmware download
# -v verifies the contents of the FRAM match the firmware file
# -z [VCC] leaves the device powered up via the VCC pin (You should see the "First Start" message on the LCD display)

command = f"MSP430Flasher.exe -j fast -r [uuid.txt,0x1A04-0x1A0a] -r [device.txt,0x1a04-0x1a07] -e ERASE_MAIN -w output.txt -v -z [VCC]"

print("STARING COMMAND:")
print(command)

result = subprocess.run(command, capture_output=False)

# Check the return code and print the output
if result.returncode != 0:
print("MSPFlasher failed!")
exit(1)
# prepend a time stamp (in TI HEX format)
# 1800 is the begining of "information memory" FRAM.
wfd.write("@1800\n".encode())
wfd.write(f"{t.tm_sec:02d} {t.tm_min:02d} {t.tm_hour:02d} {t.tm_wday:02d} {t.tm_mday:02d} {t.tm_mon:02d} {t.tm_year%100:02d}\n".encode())

# ...and append firmware file into the image file
# note that the firmware comes last becuase the TI tools add a "q" to the end of this file.
with open('tsl-calibre-msp.txt','rb') as rfd:
shutil.copyfileobj(rfd, wfd)

# If you ever need to read the commisioned time out of a unit, you can use the command...
# MSP430Flasher.exe -j fast -r [commisioned_time.txt,0x1800-0x1806] -z [VCC]
# Do note that the timestamp does not have a century field, so you will have to infer what century the unit was commisioned
# in using other factors like how dusty it is.

# Here is the meat where we...
# 1. Grab the device UUID and ID from the MSP430 chip. This will let us keep a record of the serial number and what version of the chip this unit has.
# 2. Erase the FRAM.
# 3. Burn the combined image into the unit.

# start with executable
call_line = [mspflasher_exec]

# -j fast means use the fastest clock speed so programming will go as quickly as possible (it still takes a couple seconds)
call_line +=[ "-j" , "fast" ]

# Lets grab the UUID and device info and make into a serial number
# these are in TI HEX format so we have to throw away the first line and then strip out all whitespace from the second line to get a clean hex digit string
# dump the whole device descriptor table
# note that it would be nice to just dump the two parts we need (device_id & uuid), but there is an undocumented limitation
# in MSP430Flasher where if you try to do consecutive read operations, it just siliently ignores the second one. So instead
# we dump the whole table and will parse out the parts we care about later.
# device desciptor table is in the MSP430FR4133 datasheet section 9.1

# Dump the device descirtor data from the MSP430 to a file named `dd.txt` in the temp directory.
dd_file_name = os.path.join( tempdir , 'dd.txt')
call_line += [ "-r" , f"[{dd_file_name},0x1a00-0x1a12]" ]

# -e ERASE_MAIN prepares the FRAM to recieve the firmware download
call_line += ["-e","ERASE_MAIN"]

#program in the firmware image we created earlier
call_line += ["-w" , image_file_name ]

# -v verifies the contents of the FRAM match the firmware image file
call_line += ["-v"]

# -z [VCC] leaves the device powered up via the EZ-FET programmer VCC pin (You should see the "First Start" message on the LCD display)
call_line += ["-z" , "[RESET,VCC]"]

print("STARING COMMAND:")
print(call_line)

with open("uuid.txt","r") as f:
# throw away the address line
f.readline()
# grab the UUID without trailing newline
uuid_raw = f.readline().rstrip()
result = subprocess.run( call_line , capture_output=False)

with open("device.txt","r") as f:
# throw away the address line
f.readline()
# grab the device info without trailing newline
device_raw = f.readline().rstrip()
# Check the return code and print the output
if result.returncode != 0:
print("MSPFlasher failed!")
exit(1)

# concat the two and and remove embeded spaces
deviceid = ( device_raw + uuid_raw ).replace(" ","")
# Lets grab the UUID and device info from the device descripto dump and make them into a serial number


# these are in TI HEX format so we have to throw away the first line and then strip out all whitespace from the second line to get a clean hex digit string

with open(dd_file_name,"r") as f:
# throw away the address line
f.readline()

# grab first and second line and strip all whitespace
line1 = f.readline()
line2 = f.readline()

# now line1 is bytes 0x1a00-0x1a0f of the device desriptor table as ascii hex digits
# now line2 is bytes 0x1a10-0x1a11 of the device desriptor table as ascii hex digits

# make all the read bytes into a single linear array
dd_hex_bytes = (line1+line2).split()

print( f"Device ID is {deviceid}\n")
# note in all the extractions below that in Python string slices, the end index is one past the index

# device info is 0x1a04-0x1a07 (4 bytes)
device_info = dd_hex_bytes[ 0x04 : 0x08 ]

print( "Adding record to airtable...\n")
# Lot waffer ID 0x1a0a-0x1a0d
lot_waffer = dd_hex_bytes[ 0x0a : 0x0e ]

# Die X pos 0x1a0e-0x1a0f
die_x_pos = dd_hex_bytes[ 0x0e : 0x10 ]

# Die Y pos 0x1a10-0x1a11
die_y_pos = dd_hex_bytes[ 0x10 : 0x12 ]

# I know this is ugly, but seems to be the phythonic way to join these all into a string with no seporator.
# Do you know a better way?
device_uuid = "".join(device_info+lot_waffer+die_x_pos+die_y_pos)

airtable_record_request = airtable_record_template.format(serialno,firmware_hash,deviceid)
print( f"Device UUID is {device_uuid}\n")

# Make the API request to create a new record
conn = http.client.HTTPSConnection(host)
conn.request("POST", endpoint, airtable_record_request, airtable_headers)
response = conn.getresponse()
print( "Adding record to airtable...\n")

# Check if the request was successful
if response.status == 200:
print("New record created successfully!")
else:
print("Error creating new record")
print(response.read().decode())
exit(1)
airtable_record_request = airtable_record_template.format(serialno,firmware_hash,device_uuid)


# Make the API request to create a new record
conn = http.client.HTTPSConnection(host)
conn.request("POST", endpoint, airtable_record_request, airtable_headers)
response = conn.getresponse()

# Check if the request was successful
if response.status == 200:
print("New record created successfully!")
else:
print("Error creating new record")
print(response.read().decode())
exit(1)

Loading

0 comments on commit 58bc5a1

Please sign in to comment.