Open serial port with serial.serial_for_url to be able to use pyserial's URL handlers

[wschreyer]

Hello,
I made a very minor change which allowed me to establish communication with a Lakeshore 240 module via an RS485-to-Ethernet converter using pyserial's URL handlers.

However, after a single data exchange the communication is aborted with the message "Terminating: Service not active on slave 10":

GSD parser warning in 'LSCI0F84.gsd' at line 18:
Info_Text                  = "GSD Version 1.0, valid for the Model 240-2P and 240-8P"
 --> Ignored unknown line
GSD parser warning in 'LSCI0F84.gsd' at line 23:
FMS_supp                   = 0
 --> Ignored unknown line
GSD parser warning in 'LSCI0F84.gsd' at line 31:
31.25_supp                 = 0
 --> Ignored unknown line
GSD parser warning in 'LSCI0F84.gsd' at line 92:
1
 --> Ignored unknown line
DPM2: Trying to initialize slave 10...
PHY-serial: TX   10 0A 02 49 55 16
PHY-serial: TX   10 0A 02 49 55 16
PHY-serial: RX   10 02 0A 00 0C 16
DPM2: slave[0A].state --> 'Wait for diag'
DPM2: Requesting Slave_Diag from slave 10...
PHY-serial: TX   68 05 05 68 8A 82 6D 3C 3E F3 16
PHY-serial: RX   68 0B 0B 68 82 8A 08 3E 3C 06 05 00 FF 0F 84 2B 16
DPM2: slave[0A].state --> 'Wait for Prm'
DPM2: Sending Set_Prm to slave 10...
PHY-serial: TX   68 0C 0C 68 8A 82 5D 3D 3E 80 01 01 00 0F 84 01 FA 16
PHY-serial: RX   E5
DPM2: slave[0A].state --> 'Wait for Cfg'
DPM2: Sending Chk_Cfg to slave 10...
PHY-serial: TX   68 0D 0D 68 8A 82 7D 3E 3E 93 93 93 93 93 93 93 93 9D 16
PHY-serial: RX   E5
PHY-serial: TX   68 0D 0D 68 8A 82 5D 3E 3E 93 93 93 93 93 93 93 93 7D 16
DPM2: slave[0A].state --> 'Request diag and wait for DX-ready'
DPM2: Requesting Slave_Diag (WDXRDY) from slave 10...
PHY-serial: RX   E5
PHY-serial: TX   68 05 05 68 8A 82 5D 3C 3E E3 16
PHY-serial: RX   68 0B 0B 68 82 8A 08 3E 3C 00 04 00 02 0F 84 27 16
DPM2: slave[0A].state --> 'Data_Exchange'
DPM2: Initialization finished. Running Data_Exchange with slave 10...
PHY-serial: TX   68 06 06 68 0A 02 7D 00 00 00 89 16
PHY-serial: RX   68 23 23 68 02 0A 08 43 92 E7 D1 43 92 30 59 00 00 00 00 43 92 95 40 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 A9 16
PHY-serial: TX   68 06 06 68 0A 02 5D 00 00 00 69 16
PHY-serial: RX   10 02 0A 03 0F 16
Terminating: Service not active on slave 10

I can circumvent this by setting diag_period=1, which ends up reporting an error in the configuration and reinitializes the communication after every data exchange:

...
DPM1: slave[0A].state --> 'Data_Exchange'
DPM1: Initialization finished. Running Data_Exchange with slave 10...
PHY-serial: TX   68 06 06 68 0A 02 7D 00 00 00 89 16
PHY-serial: RX   68 23 23 68 02 0A 08 43 92 E7 47 43 92 2F C0 00 00 00 00 43 92 96 72 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 B8 16
DPM1: slave[0A].state --> 'Request diag and wait for DX-ready'
DPM1: Requesting Slave_Diag (WDXRDY) from slave 10...
PHY-serial: TX   68 05 05 68 8A 82 5D 3C 3E E3 16
PHY-serial: RX   68 0B 0B 68 82 8A 08 3E 3C 06 05 00 FF 0F 84 2B 16
DPM1:  >ERROR<  Slave 10 reports a faulty configuration (Chk_Cfg).
DPM1: Slave 10 requests a new parameterization (Set_Prm).
DPM1: slave[0A].state --> 'Init'
DPM1: Trying to initialize slave 10...
...

I was hoping you might have any suggestions what is going wrong. The GSD file is available here. The conf file is below. I just send a small number of zero bytes (in this case three) as out data. I did encounter the same issue with a RS485-to-USB adapter, so it doesn't seem to be an issue with the Ethernet communication.

[PROFIBUS]
debug=2

[PHY]
type=serial
dev=socket://169.254.52.94:32
rtscts=False
dsrdtr=False
spiBus=0
spiCS=0
spiSpeedHz=2500000
baud=9600

[FDL]

[DP]
master_class=1
master_addr=2

[SLAVE_0]
name=ls240
addr=10
gsd=LSCI0F84.gsd
sync_mode=0
freeze_mode=0
group_mask=1
watchdog_ms=0
module_0=Sensor Reading 1
module_1=Sensor Reading 2
module_2=Sensor Reading 3
module_3=Sensor Reading 4
module_4=Sensor Reading 5
module_5=Sensor Reading 6
module_6=Sensor Reading 7
module_7=Sensor Reading 8
output_size=32
input_size=3
diag_period=1

Thank you!

[mbuesch]

Hi,

thanks for your question and your contribution!

It's pretty common that new slaves need some additional tweak here and there in the pyprofibus code.
The message Service not active on slave seems to indicate that found something that it did not expect.
That can have several reasons:

• pyprofibus doesn't implement that detail.
• the slave behaves non-standard.
• a bug

What happens in this case, I don't know.
But you could start digging into the code and see what the checks are.

About the change:
I think we would at least need to stay backward compatible with serial that does not provide serial_for_url. That can be done with hasattr() for example.