ofono between 30% and 40% CPU usage when roaming

Setting roaming to allowed
/usr/share/ofono/scripts/set-roaming-allowed 1

set the modem to attached and brought the cpu down.

After doing
/usr/share/ofono/scripts/set-roaming-allowed 1

The modem is now attached
/usr/share/ofono/scripts/list-modems
[ /ril_0 ]
    Features = sms net gprs sim
    Emergency = 0
    Powered = 1
    Lockdown = 0
    Interfaces = org.ofono.CallVolume org.ofono.SmartMessaging org.ofono.PushNotification org.ofono.MessageManager org.ofono.NetworkRegistration org.ofono.VoiceCallManager org.ofono.ConnectionManager org.ofono.SimManager
    Online = 1
    Model = Fake Modem Model
    Revision = Fake Revision
    Type = hardware
    Serial = THIS-IS-A-FAKE-SERIAL-NO
    Manufacturer = Fake Manufacturer
    [ org.ofono.CallVolume ]
        Muted = 1
        SpeakerVolume = 0
        MicrophoneVolume = 0
    [ org.ofono.SmartMessaging ]
    [ org.ofono.PushNotification ]
    [ org.ofono.MessageManager ]
        Alphabet = default
        UseDeliveryReports = 0
        Bearer = cs-preferred
        ServiceCenterAddress =
    [ org.ofono.NetworkRegistration ]
        Status = roaming
        Strength = 83
        Name = Personal (ANTEL)
        LocationAreaCode = 109
        Mode = auto
        MobileCountryCode = 748
        Technology = umts
        CellId = 1078772
        MobileNetworkCode = 01
    [ org.ofono.VoiceCallManager ]
        EmergencyNumbers = 08 000 999 110 112 911 118 119
    [ org.ofono.ConnectionManager ]
        Powered = 1
        Attached = 1
        Bearer = none
        Suspended = 0
        RoamingAllowed = 1
    [ org.ofono.SimManager ]
        Retries =
        MobileCountryCode = 722
        FixedDialing = 0
        SubscriberNumbers =
        PreferredLanguages = es en pt it
        BarredDialing = 0
        ServiceNumbers = [Personal] = '+810' [Club Personal] = '*2582' [At. Clientes] = '*111' [Llam. a su Cargo] = '1111111111'
        CardIdentifier = 89543421211145542111
        LockedPins =
        MobileNetworkCode = 34
        SubscriberIdentity = 722341214554211
        Present = 1
        PinRequired = none

phablet@ubuntu-phablet:~$ nmcli con up id Personal
phablet@ubuntu-phablet:~$ nmcli d
DEVICE TYPE STATE
/ril_0 gsm connected
wlan0 802-11-wireless disconnected
phablet@ubuntu-phablet:~$ netstat -rn
Kernel IP routing table
Destination Gateway Genmask Flags MSS Window irtt Iface
0.0.0.0 0.0.0.0 0.0.0.0 U 0 0 0 rmnet0
181.4.218.0 0.0.0.0 255.255.255.0 U 0 0 0 rmnet0
phablet@ubuntu-phablet:~$ ping www.google.com
PING www.google.com (181.15.96.25) 56(84) bytes of data.
64 bytes from host25.181-15-96.telecom.net.ar (181.15.96.25): icmp_req=1 ttl=55 time=94.2 ms
64 bytes from host25.181-15-96.telecom.net.ar (181.15.96.25): icmp_req=2 ttl=55 time=71.8 ms
^C
--- www.google.com ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1001ms
rtt min/avg/max/mdev = 71.802/83.027/94.252/11.225 ms

@Sergio

Here's a test version of ofono with improved logging.

First, can you re-run your scenario both with, and without the "-d" switch with the roaming flag disabled. I'd like to see whether or not the CPU usage is due to the logging. It seems that when roaming, an excessive number of VOICE_NETWORK_STATE_CHANGED events are processed, and these in turn result in multiple requests to the RIL.

One trick to keeping the logs small when you do this is to delete the syslog in between reboots. Another alternative is to drop an ofono.override containing the keyword "manual" into /etc/init, then use the script command to capture the output of ofono run as a foreground process ( be sure to specify "-n" to prevent ofonod from becoming a daemon ).

Hi,

I ran the following command and got the attached logs.
/usr/share/ofono/scripts/set-roaming-allowed 0

Looks like syslog is being flooded with RIL_REQUEST_DATA_REGISTRATION_STATE messages.

Either with or without -d the ofonod cpu usage remains the same.

10% ofonod
15% rild

Once I set roaming to "enabled", the cpu usage of both are back to 0%.

Just a quick update...

I'm actually working on this bug now, and should not have left the Status as "InProgress" for so long without an update.

The first issue is that we're definitely sending more REQUEST_DATA_REGISTRATION_STATE requests than necessary. This is due to the fact that both the network-registration and gprs rilmodem code cause these requests to be sent to RILD ( the former through a core netreg_watch by the core ofono gprs code ).

I also think there's further room for reduction, but need to play around with the code some more.

Hopefully I'll have a test package available soon ( maybe tomorrow ).

Sounds good, I am in no place where I can trigger this easily now, but
I'm betting Tiago could

On Tue, Jul 9, 2013 at 10:25 PM, Tony Espy <email address hidden> wrote:
> Just a quick update...
>
> I'm actually working on this bug now, and should not have left the
> Status as "InProgress" for so long without an update.
>
> The first issue is that we're definitely sending more
> REQUEST_DATA_REGISTRATION_STATE requests than necessary. This is due
> to the fact that both the network-registration and gprs rilmodem code
> cause these requests to be sent to RILD ( the former through a core
> netreg_watch by the core ofono gprs code ).
>
> I also think there's further room for reduction, but need to play around
> with the code some more.
>
> Hopefully I'll have a test package available soon ( maybe tomorrow ).
>
> --
> You received this bug notification because you are subscribed to the bug
> report.
> https://bugs.launchpad.net/bugs/1188404
>
> Title:
> ofono between 30% and 40% CPU usage when roaming
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/touch-preview-images/+bug/1188404/+subscriptions

Sure, I can test it.

I have a preliminary fix which appears to work.

The problem was caused by a tight loop n the GPRS code where the callback for a RIL_REQ_DATA_REGISTRATION_STATE called back into the ofono core gprs code, and basically triggered another REQ_DATA_REGISTRATION_STATE, ...

See the attached branch.

I can confirm the linked branch fixes the problem.

WFM as well, see my comment on the MP: https://code.launchpad.net/~awe/phablet-extras/ofono-lp1188404/+merge/174516/comments/393721