Re: [PATCH] gprs: lte: set attached after successful activation

[Denis Kenzior <denkenz@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 3502 bytes --]

Hi Dragos,

On 11/22/2016 09:03 AM, Dragos Tatulea wrote:
> Otherwise the attached state gets to be set before the actual LTE
> automatic context is ready. This triggers a race between connman
> and ofono: connman sees status attached before the context is active
> so connman will try to activate another context with same apn and will
> fail over and over again.
> ---
>   src/gprs.c | 52 +++++++++++++++++++++++++++-------------------------
>   1 file changed, 27 insertions(+), 25 deletions(-)
>
> diff --git a/src/gprs.c b/src/gprs.c
> index 5f7620b..c5fda35 100644
> --- a/src/gprs.c
> +++ b/src/gprs.c
> @@ -959,6 +959,24 @@ static void pri_deactivate_callback(const struct ofono_error *error, void *data)
>   	}
>   }
>
> +static void gprs_set_attached_property(struct ofono_gprs *gprs,
> +					ofono_bool_t attached)
> +{
> +	const char *path;
> +	DBusConnection *conn = ofono_dbus_get_connection();
> +	dbus_bool_t value = attached;
> +
> +	if (gprs->attached == attached)
> +		return;
> +
> +	gprs->attached = attached;
> +
> +	path = __ofono_atom_get_path(gprs->atom);
> +	ofono_dbus_signal_property_changed(conn, path,
> +				OFONO_CONNECTION_MANAGER_INTERFACE,
> +				"Attached", DBUS_TYPE_BOOLEAN, &value);
> +}
> +
>   static void pri_read_settings_callback(const struct ofono_error *error,
>   					void *data)
>   {
> @@ -987,6 +1005,9 @@ static void pri_read_settings_callback(const struct ofono_error *error,
>   	}
>
>   	value = pri_ctx->active;
> +
> +	gprs_set_attached_property(pri_ctx->gprs, TRUE);
> +
>   	ofono_dbus_signal_property_changed(conn, pri_ctx->path,
>   					OFONO_CONNECTION_CONTEXT_INTERFACE,
>   					"Active", DBUS_TYPE_BOOLEAN, &value);
> @@ -1596,24 +1617,6 @@ static void release_active_contexts(struct ofono_gprs *gprs)
>   	}
>   }
>
> -static void gprs_set_attached_property(struct ofono_gprs *gprs,
> -					ofono_bool_t attached)
> -{
> -	const char *path;
> -	DBusConnection *conn = ofono_dbus_get_connection();
> -	dbus_bool_t value = attached;
> -
> -	if (gprs->attached == attached)
> -		return;
> -
> -	gprs->attached = attached;
> -
> -	path = __ofono_atom_get_path(gprs->atom);
> -	ofono_dbus_signal_property_changed(conn, path,
> -				OFONO_CONNECTION_MANAGER_INTERFACE,
> -				"Attached", DBUS_TYPE_BOOLEAN, &value);
> -}
> -
>   static void gprs_attached_update(struct ofono_gprs *gprs)
>   {
>   	ofono_bool_t attached;
> @@ -1644,8 +1647,6 @@ static void gprs_attached_update(struct ofono_gprs *gprs)
>   		gprs->flags |= GPRS_FLAG_ATTACHED_UPDATE;
>   		return;
>   	}
> -
> -	gprs_set_attached_property(gprs, attached);

This patch mostly makes sense to me, but this part here seems dangerous. 
  Are you sure this is needed?

>   }
>
>   static void registration_status_cb(const struct ofono_error *error,
> @@ -1717,11 +1718,12 @@ static void gprs_netreg_update(struct ofono_gprs *gprs)
>   	DBG("attach: %u, driver_attached: %u", attach, gprs->driver_attached);
>
>   	if (ofono_netreg_get_technology(gprs->netreg) ==
> -			ACCESS_TECHNOLOGY_EUTRAN) {
> -		/* Ignore attach logic for LTE. There is no such concept. */
> -		gprs_set_attached_property(gprs, attach);
> -		return;
> -	}
> +			ACCESS_TECHNOLOGY_EUTRAN)
> +		/*
> +		 * For LTE we set attached status only on successful
> +		 * context activation.
> +		 */
> +                return;
>
>   	if (gprs->driver_attached == attach)
>   		return;
>

Regards,
-Denis