sms sending times out too early?

[Bjørn Mork]

This was tested on the MC7710. I'll test with other devices, but that has to wait until next year :)

I guess this doesn't matter much though. The end result is a timeout in any case, and you will only see the response matching failure if you read the debug log.


Bjørn 

Aleksander Morgado <aleksander at aleksander.es> wrote:
>Hey,
>
>>> I just noticed that SMS sending failures with the MC7710 result in
>>>
>>>
>>> ModemManager[6067]: <debug> [1387788544.392227] [mm-sms.c:178]
>generate_3gpp_submit_pdus():   Processing chunk '0' of text with '3'
>bytes
>>> ModemManager[6067]: <debug> [1387788544.392344] [mm-sms.c:208]
>generate_3gpp_submit_pdus(): Created SMS part for singlepart SMS
>>> ModemManager[6067]: <debug> [1387788544.392408]
>[mm-sms-part-3gpp.c:805] mm_sms_part_3gpp_get_submit_pdu(): Creating
>PDU for part...
>>> ModemManager[6067]: <debug> [1387788544.392458]
>[mm-sms-part-3gpp.c:892] mm_sms_part_3gpp_get_submit_pdu():   using
>GSM7 encoding...
>>> ModemManager[6067]: <debug> [1387788544.392519]
>[mm-sms-part-3gpp.c:957] mm_sms_part_3gpp_get_submit_pdu():   user data
>length is '3' septets (without UDH)
>>> ModemManager[6067]: [/dev/cdc-wdm0] Sent message...
>>> <<<<<< RAW:
>>> <<<<<<   length = 76
>>> <<<<<<   data   =
>03:00:00:00:4C:00:00:00:25:00:00:00:01:00:00:00:00:00:00:00:53:3F:BE:EB:14:FE:44:67:9F:90:33:A2:23:E5:6C:3F:03:00:00:00:01:00:00:00:1C:00:00:00:00:00:00:00:08:00:00:00:10:00:00:00:00:01:00:0A:91:74:74:09:00:68:00:00:03:E6:F7:1B
>>> ModemManager[6067]: [/dev/cdc-wdm0] Sent message (translated)...
>>> <<<<<< Header:
>>> <<<<<<   length      = 76
>>> <<<<<<   type        = command (0x00000003)
>>> <<<<<<   transaction = 37
>>> <<<<<< Fragment header:
>>> <<<<<<   total   = 1
>>> <<<<<<   current = 0
>>> <<<<<< Contents:
>>> <<<<<<   service = 'sms' (533fbeeb-14fe-4467-9f90-33a223e56c3f)
>>> <<<<<<   cid     = 'send' (0x00000003)
>>> <<<<<<   type    = 'set' (0x00000001)
>>>
>>> ModemManager[6067]: [/dev/cdc-wdm0] Received message...
>>> >>>>>> RAW:
>>> >>>>>>   length = 48
>>> >>>>>>   data   =
>03:00:00:80:30:00:00:00:25:00:00:00:01:00:00:00:00:00:00:00:53:3F:BE:EB:14:FE:44:67:9F:90:33:A2:23:E5:6C:3F:03:00:00:00:02:00:00:00:00:00:00:00
>>> ModemManager[6067]: [/dev/cdc-wdm0] No transaction matched in
>received message
>>>
>>>
>>> And with times to show why the received message isn't matched with
>the
>>> request:
>>>
>>>
>>> Dec 23 09:49:04 nemi ModemManager[6067]: [/dev/cdc-wdm0] Sent
>message (translated)...#012<<<<<< Header:#012<<<<<<   length      =
>76#012<<<<<<   type        = command (0x00000003)#012<<<<<<  
>transaction = 37#012<<<<<< Fragment header:#012<<<<<<   total   =
>1#012<<<<<<   current = 0#012<<<<<< Contents:#012<<<<<<   service =
>'sms' (533fbeeb-14fe-4467-9f90-33a223e56c3f)#012<<<<<<   cid     =
>'send' (0x00000003)#012<<<<<<   type    = 'set' (0x00000001)
>>> Dec 23 09:50:35 nemi ModemManager[6067]: [/dev/cdc-wdm0] Received
>message...#012>>>>>> RAW:#012>>>>>>   length = 48#012>>>>>>   data   =
>03:00:00:80:30:00:00:00:25:00:00:00:01:00:00:00:00:00:00:00:53:3F:BE:EB:14:FE:44:67:9F:90:33:A2:23:E5:6C:3F:03:00:00:00:02:00:00:00:00:00:00:00
>>>
>>>
>>> So the modem tries for 90 seconds before it decides to time out the
>>> transaction.  I suspect that may apply to other transactions which
>may
>>> time out as well..
>>>
>>> Maybe libmbim needs to wait that long as well?
>>>
>>>
>>
>> Yeah, probably something to change in ModemManager itself; as each
>> libmbim command request has the timeout specified as an input
>> parameter. Will add it to my TODO list.
>
>I'm guessing that the timeout to decide that an SMS cannot be sent is
>device specific... and 90s really seems too much to me. Should we
>really increase the default timeout to >90s? With which device did you
>get this timeout? I'm tempted to just increase the default timeout in
>MBIM's SMS send from 10s (which is likely too low) to e.g. 30s (which
>is what we have in the QMI implementation). But that wouldn't solve
>your issue, you would still get the unmatched response if the modem
>times out at 90s.
>
>What do others think?