No D-Bus Reply to Activating Context

No D-Bus Reply to Activating Context

[Brian Ruptash]

[-- Attachment #1: Type: text/plain, Size: 2815 bytes --]

I'm having a spurious problem where the D-Bus error reply to a context
activation isn't being sent back.  This is happenning in the following
sequence, where registration is lost while in the midst of activating the
context.  Not like I can recreate this, but I've received a log from an
occurance in the field.  I'm running ofono 1.12 with a Cinterion PH8-P
driver that I wrote.

 

Here's a trace of the relevant sequence.  I should receive a D-Bus reply
from the last line's pri_activate_callback() but I don't.  Does anyone see
anything odd here, and any reason why the reply wouldn't be being sent?
Thanks.

 

09:20:42.300 gprs-context.c:317          at_gprs_activate_primary cid 1

09:20:42.302 ph8.c:122                   ph8_debug Mdm: >
AT+CGDCONT=1,"IP","xxxxxxxx"\r

09:20:42.531 ph8.c:122                   ph8_debug App: < \r\n+CGREG:
0\r\n\r\n+CREG: 0\r\n\r\n+CIEV: psinfo,6\r\n

09:20:42.531 gprs.c:2148                 ofono_gprs_status_notify /ph8_1
status 0

09:20:42.533 network.c:1349              ofono_netreg_status_notify /ph8_1
status 0 tech -1

09:20:42.536 network.c:1219              current_operator_callback 0x128198,
0x1260b8

09:20:42.540 gprs.c:1606                 netreg_status_changed 0

09:20:42.540 gprs.c:1606                 netreg_status_changed 0

09:20:42.540 network-registration.c:2134 ph8_handle_sind_ciev tech 2

09:20:42.541 gprs.c:340                  ph8_ciev_notify bearer 3 (umts)

09:20:42.542 ph8.c:208                   ciev_notify 

09:20:42.543 ph8.c:122                   ph8_debug App: > AT+CGATT=0\r

09:20:42.581 ph8.c:122                   ph8_debug Mdm: < \r\nOK\r\n

09:20:42.581 gprs-context.c:281          at_cgdcont_cb ok 1

09:20:42.581 ph8.c:122                   ph8_debug App: < \r\nOK\r\n

09:20:42.581 gprs.c:1545                 gprs_attach_callback /ph8_1 error =
0

09:20:42.581 ph8.c:122                   ph8_debug Mdm: >
AT+CGDATA="PPP",1\r

09:20:42.582 ph8.c:122                   ph8_debug App: > AT+CGREG?\r

09:20:42.591 ph8.c:122                   ph8_debug App: < \r\n+CGREG:
1,0\r\n\r\nOK\r\n

09:20:42.591 gprs.c:1526                 registration_status_cb /ph8_1 error
0 status 0

09:20:42.591 gprs.c:2148                 ofono_gprs_status_notify /ph8_1
status 0

09:20:42.611 ph8.c:122                   ph8_debug Mdm: < \r\nNO CARRIER\r\n

09:20:42.611 gprs-context.c:258          at_cgdata_cb Unable to enter data
state

09:20:42.611 gprs-context.c:266          at_cgdata_cb not ok, <NO CARRIER>,
error 5/0

09:20:42.611 gprs.c:859                  pri_activate_callback 0x12c980 -
error 5/0

09:20:42.611 gprs.c:863                  pri_activate_callback Activating
context failed with error: Unknown error type

 

 

-- Brian

 


[-- Attachment #2: attachment.html --]
[-- Type: text/html, Size: 7373 bytes --]

Re: No D-Bus Reply to Activating Context

[Denis Kenzior]

[-- Attachment #1: Type: text/plain, Size: 577 bytes --]

Hi Brian,

 > 09:20:42.611 gprs.c:863                  pri_activate_callback
> Activating context failed with error: Unknown error type
>

Looking at that line, we are sending a reply right after printing the 
message.  In case the message is NULL, you'd most likely crash inside 
D-Bus or get an assertion of some kind.

The pending variable is also not messed with anywhere else, so you 
should be getting a reply.

The only thing I can suggest is to try and simulate this behavior with 
phonesim and see if you can duplicate / debug it.

Regards,
-Denis

RE: No D-Bus Reply to Activating Context

[Brian Ruptash]

[-- Attachment #1: Type: text/plain, Size: 1203 bytes --]

> > 09:20:42.611 gprs.c:863    pri_activate_callback
> > Activating context failed with error: Unknown error type
> 
> Looking at that line, we are sending a reply right after printing the
> message.  In case the message is NULL, you'd most likely crash inside
> D-Bus or get an assertion of some kind.
> 
> The pending variable is also not messed with anywhere else, so you
> should be getting a reply.
> 
> The only thing I can suggest is to try and simulate this behavior with
> phonesim and see if you can duplicate / debug it.

I traced the log and code and agree, which is why I can't understand why
the D-Bus reply is either not sent or perhaps not received, though it
only occurs in this particular sequence and is the only case a method
return is ever "lost".  dbus has a bug in that dbus-monitor can't trace
method returns, so I can't easily see what's happenning especially as
it's quite sporadic.

So I punted... I upgraded from ofono 1.12 to 1.14 which has a significant
overhaul to gdbus, though I can't find anything in the ofono mailing list
that explains what or why.  And I'll see if the issue persists.

> Regards,
> -Denis

Thanks, Denis.

-- Brian