[pulseaudio-discuss] [PATCH] alsa-card: improve the profile availability logic
Georg Chini
georg at chini.tk
Fri Oct 27 19:39:18 UTC 2017
On 21.09.2017 14:47, Tanu Kaskinen wrote:
> When a new card shows up (during pulseaudio startup or hotplugged),
> pulseaudio needs to pick the initial profile for the card. Unavailable
> profiles shouldn't be picked, but module-alsa-card sometimes marked
> unavailable profiles as available, causing bad initial profile choices.
>
> This patch changes module-alsa-card so that it marks all profiles
> unavailable whose all output ports or all input ports are unavailable.
> Previously only those profiles were marked as unavailable whose all
> ports were unavailable. For example, if a profile contains one sink and
> one source, and the sink is unavailable and the source is available,
> previously such profile was marked as available, but now it's marked as
> unavailable.
>
> BugLink: https://bugs.freedesktop.org/show_bug.cgi?id=102902
> ---
> src/modules/alsa/module-alsa-card.c | 62 +++++++++++++++++++++++++++++++------
> 1 file changed, 52 insertions(+), 10 deletions(-)
>
> diff --git a/src/modules/alsa/module-alsa-card.c b/src/modules/alsa/module-alsa-card.c
> index 4f1236ecd..2adfa6d7b 100644
> --- a/src/modules/alsa/module-alsa-card.c
> +++ b/src/modules/alsa/module-alsa-card.c
> @@ -427,29 +427,71 @@ static int report_jack_state(snd_mixer_elem_t *melem, unsigned int mask) {
> if (tp->avail == PA_AVAILABLE_NO)
> pa_device_port_set_available(tp->port, tp->avail);
>
> - /* Update profile availabilities. The logic could be improved; for now we
> - * only set obviously unavailable profiles (those that contain only
> - * unavailable ports) to PA_AVAILABLE_NO and all others to
> - * PA_AVAILABLE_UNKNOWN. */
> + /* Update profile availabilities. Ideally we would mark all profiles
> + * unavailable that contain unavailable devices. We can't currently do that
> + * in all cases, because if there are multiple sinks in a profile, and the
> + * profile contains a mix of available and unavailable ports, we don't know
> + * how the ports are distributed between the different sinks. It's possible
> + * that some sinks contain only unavailable ports, in which case we should
> + * mark the profile as unavailable, but it's also possible that all sinks
> + * contain at least one available port, in which case we should mark the
> + * profile as available. Until the data structures are improved so that we
> + * can distinguish between these two cases, we mark the problematic cases
> + * as available (well, "unknown" to be precise, but there's little
> + * practical difference).
> + *
> + * When all output ports are unavailable, we know that all sinks are
> + * unavailable, and therefore the profile is marked unavailable as well.
> + * The same applies to input ports as well, of course.
> + *
> + * If there are no output ports at all, but the profile contains at least
> + * one sink, then the output is considered to be available. */
> PA_HASHMAP_FOREACH(profile, u->card->profiles, state) {
> pa_device_port *port;
> void *state2;
> - pa_available_t available = PA_AVAILABLE_NO;
> + bool all_output_ports_are_unavailable = false;
> + bool all_input_ports_are_unavailable = false;
> + pa_available_t available = PA_AVAILABLE_UNKNOWN;
>
> - /* Don't touch the "off" profile. */
> - if (profile->n_sources == 0 && profile->n_sinks == 0)
> - continue;
> + /* Check if all output ports are unavailable. */
> + PA_HASHMAP_FOREACH(port, u->card->ports, state2) {
> + if (port->direction != PA_DIRECTION_OUTPUT)
> + continue;
> +
> + if (!pa_hashmap_get(port->profiles, profile->name))
> + continue;
>
> + if (port->available == PA_AVAILABLE_NO)
> + all_output_ports_are_unavailable = true;
> + else {
> + all_output_ports_are_unavailable = false;
> + break;
> + }
> + }
> +
> + /* Check if all input ports are unavailable. */
> PA_HASHMAP_FOREACH(port, u->card->ports, state2) {
> + if (port->direction != PA_DIRECTION_INPUT)
> + continue;
> +
> if (!pa_hashmap_get(port->profiles, profile->name))
> continue;
>
> - if (port->available != PA_AVAILABLE_NO) {
> - available = PA_AVAILABLE_UNKNOWN;
> + if (port->available == PA_AVAILABLE_NO)
> + all_input_ports_are_unavailable = true;
> + else {
> + all_input_ports_are_unavailable = false;
> break;
> }
> }
>
> + if (all_output_ports_are_unavailable || all_input_ports_are_unavailable)
> + available = PA_AVAILABLE_NO;
> +
> + /* The "off" profile is always available. */
> + if (profile->n_sources == 0 && profile->n_sinks == 0)
> + available = PA_AVAILABLE_YES;
The extra check for the "off" profile is unnecessary, it has no ports and
will therefore be marked as available (or rather as UNKNOWN).
> +
> pa_card_profile_set_available(profile, available);
> }
>
Otherwise in general OK, but would it not be clearer to use one loop
like this:
has_input_port = false;
has_output_port = false;
found_available_input_port = false;
found_available_output_port = false;
PA_HASHMAP_FOREACH(port, u->card->ports, state2) {
if (!pa_hashmap_get(port->profiles, profile->name))
continue;
if (port->direction == PA_DIRECTION_INPUT) {
has_input_port = true;
if (port->available != PA_AVAILABLE_NO)
found_available_input_port = true;
}
if (port->direction == PA_DIRECTION_OUTPUT) {
has_output_port = true;
if (port->available != PA_AVAILABLE_NO)
found_available_input_port = true;
}
}
If you prefer to leave it as is, I don't have objections though.
More information about the pulseaudio-discuss
mailing list