I've added a configuration item DisableParkState in main.conf and let
the bluetoothd check this. If it's set to true, the
main_opts.link_policy will be HCI_LP_RSWITCH | HCI_LP_SNIFF |
HCI_LP_HOLD, otherwise, it will be original one -
HCI_LP_RSWITCH | HCI_LP_SNIFF | HCI_LP_HOLD | HCI_LP_PARK. Any comment
on the code change or the way to configure? I just want to see if
further improvement is necessary before I send out the patch. Thanks.
diff --git a/data/main.conf b/data/main.conf
index 8153eb2..b9b8fef 100644
--- a/data/main.conf
+++ b/data/main.conf
@@ -57,3 +57,8 @@ ReverseServiceDiscovery = true
# Enable name resolving after inquiry. Set it to 'false' if you don't need
# remote devices name and want shorter discovery cycle. Defaults to 'true'.
NameResolving = true
+
+# Disable park state support helps some bluetooth device stably connected, e.g.
+# Motorola HF850 car-kit will lose connection or fail to setup SCO when it's
+# in park state. Defaults to 'false'.
+DisableParkState = true
diff --git a/src/main.c b/src/main.c
index ff423a4..e0ec7ce 100644
--- a/src/main.c
+++ b/src/main.c
@@ -203,8 +203,15 @@ static void parse_config(GKeyFile *config)
main_opts.link_mode = HCI_LM_ACCEPT;
main_opts.link_policy = HCI_LP_RSWITCH | HCI_LP_SNIFF |
- HCI_LP_HOLD | HCI_LP_PARK;
+ HCI_LP_HOLD ;
+ boolean = g_key_file_get_boolean(config, "General",
+ "DisableParkState", &err);
+ if (err)
+ g_clear_error(&err);
+ else if (!boolean)
+ main_opts.link_policy |= HCI_LP_PARK;
}
2010/2/10 Liang Bao <[email protected]>:
> Anyone could advice? Thanks.
>
> 2010/2/8 Liang Bao <[email protected]>:
>> Hi,
>>
>> We noticed that currently BlueZ doesn't read link policy settings from
>> main.conf as it does for other things such as General.Name or
>> General.Class. It just enable everything in parse_config()
>> (src/main.c). Is this just due to lack of use case or BlueZ does this
>> intentionally for some reason? Thanks.
>>
>>
>> 86 static void parse_config(GKeyFile *config)
>> 87 {
>> .....
>> 206 ? ? ? ? main_opts.link_policy = HCI_LP_RSWITCH | HCI_LP_SNIFF |
>> 207 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? HCI_LP_HOLD | HCI_LP_PARK;
>> 208 }
>>
>
Here's the patch file. I move the assignment of default value to
init_defaults() and just remove HCI_LP_PARK if the configuration is
set to "true". By default, this configuation in main.conf is written
as "DisableParkState = false"
2010/2/24 Liang Bao <[email protected]>:
> I've added a configuration item DisableParkState in main.conf and let
> the bluetoothd check this. If it's set to true, the
> main_opts.link_policy will be HCI_LP_RSWITCH | HCI_LP_SNIFF |
> HCI_LP_HOLD, otherwise, it will be original one -
> HCI_LP_RSWITCH | HCI_LP_SNIFF | HCI_LP_HOLD | HCI_LP_PARK. Any comment
> on the code change or the way to configure? I just want to see if
> further improvement is necessary before I send out the patch. Thanks.
>
> diff --git a/data/main.conf b/data/main.conf
> index 8153eb2..b9b8fef 100644
> --- a/data/main.conf
> +++ b/data/main.conf
> @@ -57,3 +57,8 @@ ReverseServiceDiscovery = true
> ?# Enable name resolving after inquiry. Set it to 'false' if you don't need
> ?# remote devices name and want shorter discovery cycle. Defaults to 'true'.
> ?NameResolving = true
> +
> +# Disable park state support helps some bluetooth device stably connected, e.g.
> +# Motorola HF850 car-kit will lose connection or fail to setup SCO when it's
> +# in park state. Defaults to 'false'.
> +DisableParkState = true
>
> diff --git a/src/main.c b/src/main.c
> index ff423a4..e0ec7ce 100644
> --- a/src/main.c
> +++ b/src/main.c
> @@ -203,8 +203,15 @@ static void parse_config(GKeyFile *config)
>
> ? ? ? ?main_opts.link_mode = HCI_LM_ACCEPT;
>
> ? ? ? ?main_opts.link_policy = HCI_LP_RSWITCH | HCI_LP_SNIFF |
> - ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? HCI_LP_HOLD | HCI_LP_PARK;
> + ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? HCI_LP_HOLD ;
> + ? ? ? ?boolean = g_key_file_get_boolean(config, "General",
> + ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?"DisableParkState", &err);
> + ? ? ? ?if (err)
> + ? ? ? ? ? ? ? ?g_clear_error(&err);
> + ? ? ? ?else if (!boolean)
> + ? ? ? ? ? ? ? main_opts.link_policy |= HCI_LP_PARK;
> ?}
>
>
>
>
> 2010/2/10 Liang Bao <[email protected]>:
>> Anyone could advice? Thanks.
>>
>> 2010/2/8 Liang Bao <[email protected]>:
>>> Hi,
>>>
>>> We noticed that currently BlueZ doesn't read link policy settings from
>>> main.conf as it does for other things such as General.Name or
>>> General.Class. It just enable everything in parse_config()
>>> (src/main.c). Is this just due to lack of use case or BlueZ does this
>>> intentionally for some reason? Thanks.
>>>
>>>
>>> 86 static void parse_config(GKeyFile *config)
>>> 87 {
>>> .....
>>> 206 ? ? ? ? main_opts.link_policy = HCI_LP_RSWITCH | HCI_LP_SNIFF |
>>> 207 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? HCI_LP_HOLD | HCI_LP_PARK;
>>> 208 }
>>>
>>
>