Attachment 123437 Details for Bug 93259 – [PATCH] alsa: ignore jack events when the user is inactive

[patch] [PATCH] alsa: ignore jack events when the user is inactive

0001-alsa-ignore-jack-events-when-the-user-is-inactive.patch (text/plain), 3.25 KB, created by Tanu Kaskinen on 2016-05-03 14:21:49 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: Tanu Kaskinen

Created: 2016-05-03 14:21:49 UTC

Size: 3.25 KB

patch

obsolete

>From a712cc3a2ed2b407dbec47722e61411ccf94a910 Mon Sep 17 00:00:00 2001
>From: Tanu Kaskinen <tanuk@iki.fi>
>Date: Wed, 20 Apr 2016 15:35:13 +0300
>Subject: [PATCH] alsa: ignore jack events when the user is inactive
>
>See the big comment in the code for more details.
>
>BugLink: https://bugs.freedesktop.org/show_bug.cgi?id=93259
>---
> src/modules/alsa/module-alsa-card.c | 28 ++++++++++++++++++++++++++++
> src/pulsecore/card.c                |  5 +++++
> src/pulsecore/card.h                |  2 ++
> 3 files changed, 35 insertions(+)
>
>diff --git a/src/modules/alsa/module-alsa-card.c b/src/modules/alsa/module-alsa-card.c
>index 286cfc9..b1e3e63 100644
>--- a/src/modules/alsa/module-alsa-card.c
>+++ b/src/modules/alsa/module-alsa-card.c
>@@ -369,6 +369,34 @@ static int report_jack_state(snd_mixer_elem_t *melem, unsigned int mask) {
> 
>     pa_assert(u);
> 
>+    /* Quick and dirty fix for
>+     * https://bugs.freedesktop.org/show_bug.cgi?id=93259
>+     *
>+     * Changing the jack state may cause a port change, and a port change will
>+     * make the sink or source change the mixer settings. If there are multiple
>+     * users having pulseaudio running, the mixer changes done by inactive
>+     * users may mess up the volume settings for the active users, because when
>+     * the inactive users change the mixer settings, those changes are picked
>+     * up by the active user's pulseaudio instance and the changes are
>+     * interpreted as if the active user changed the settings manually e.g.
>+     * with alsamixer. Even single-user systems suffer from this, because gdm
>+     * runs its own pulseaudio instance.
>+     *
>+     * Returning early here means that jack state events get ignored while the
>+     * user is inactive. When the user becomes active again, the routing may
>+     * not any more match the real jack state. While this is bad, this should
>+     * nevertheless be better than messing up the volume every time headphones
>+     * are plugged in or out.
>+     *
>+     * It might be better to unload the card altogether when the user becomes
>+     * inactive and udev removes the permission to the card. That requires at
>+     * least improving the default sink handling so that if the unloaded card
>+     * contained the default sink, the default sink should be restored to what
>+     * it was earlier, when the user becomes active and the card becomes
>+     * accessible. */
>+    if (u->card->suspend_cause & PA_SUSPEND_SESSION)
>+        return 0;
>+
>     if (mask == SND_CTL_EVENT_MASK_REMOVE)
>         return 0;
> 
>diff --git a/src/pulsecore/card.c b/src/pulsecore/card.c
>index b6cbbf7..a86cff4 100644
>--- a/src/pulsecore/card.c
>+++ b/src/pulsecore/card.c
>@@ -318,6 +318,11 @@ int pa_card_suspend(pa_card *c, bool suspend, pa_suspend_cause_t cause) {
>     pa_assert(c);
>     pa_assert(cause != 0);
> 
>+    if (suspend)
>+        c->suspend_cause |= cause;
>+    else
>+        c->suspend_cause &= ~cause;
>+
>     PA_IDXSET_FOREACH(sink, c->sinks, idx) {
>         int r;
> 
>diff --git a/src/pulsecore/card.h b/src/pulsecore/card.h
>index 30bfc0e..72e7869 100644
>--- a/src/pulsecore/card.h
>+++ b/src/pulsecore/card.h
>@@ -82,6 +82,8 @@ struct pa_card {
> 
>     bool save_profile:1;
> 
>+    pa_suspend_cause_t suspend_cause;
>+
>     void *userdata;
> 
>     int (*set_profile)(pa_card *c, pa_card_profile *profile);
>-- 
>2.8.1
>

Actions: View | Splinter Review

Attachments on bug 93259: 123437 | 123563

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.