On Wed, 2015-06-10 at 17:19 +0530, arun at accosted.net wrote:
> From: Arun Raghavan <git at arunraghavan.net>
>
> If we don't do this, autoloading a filter module on top of the null
> sink
> results in an infinite loop of:
>
> 1. Load null sink since there are no real sinks
> 2. Load filter sink when a stream wants it
> 3. We have a (filter) sink, unload null sink
> 4. Goto 1
> ---
> src/modules/module-always-sink.c | 4 ++++
> 1 file changed, 4 insertions(+)
>
> diff --git a/src/modules/module-always-sink.c b/src/modules/module
> -always-sink.c
> index a5f6bb3..b5721bf 100644
> --- a/src/modules/module-always-sink.c
> +++ b/src/modules/module-always-sink.c
> @@ -114,6 +114,10 @@ static pa_hook_result_t
> put_hook_callback(pa_core *c, pa_sink *sink, void* userd
> if (sink->module && sink->module->index == u->null_module)
> return PA_HOOK_OK;
>
> + /* We don't count filter sinks since they need a real sink */
> + if (pa_sink_is_filter(sink))
> + return PA_HOOK_OK;
> +
> pa_log_info("A new sink has been discovered. Unloading null
> -sink.");
>
> pa_module_unload_request_by_index(c, u->null_module, true);
Looks good to me.
--
Tanu
