summaryrefslogtreecommitdiffstats
path: root/sound/pci/hda/hda_generic.c
diff options
context:
space:
mode:
authorTakashi Iwai <tiwai@suse.de>2019-08-30 12:03:38 +0200
committerTakashi Iwai <tiwai@suse.de>2019-08-30 12:03:38 +0200
commit89781d0806c2c4f29072d3f00cb2dd4274aabc3d (patch)
treeb70b7172d45ece9ce813355589d2ea4406f49a51 /sound/pci/hda/hda_generic.c
parent333f31436d3db19f4286f8862a00ea1d8d8420a1 (diff)
downloadtalos-op-linux-89781d0806c2c4f29072d3f00cb2dd4274aabc3d.tar.gz
talos-op-linux-89781d0806c2c4f29072d3f00cb2dd4274aabc3d.zip
ALSA: hda/realtek - Fix overridden device-specific initialization
The recent change to shuffle the codec initialization procedure for Realtek via commit 607ca3bd220f ("ALSA: hda/realtek - EAPD turn on later") caused the silent output on some machines. This change was supposed to be safe, but it isn't actually; some devices have quirk setups to override the EAPD via COEF or BTL in the additional verb table, which is applied at the beginning of snd_hda_gen_init(). And this EAPD setup is again overridden in alc_auto_init_amp(). For recovering from the regression, tell snd_hda_gen_init() not to apply the verbs there by a new flag, then apply the verbs in alc_init(). BugLink: https://bugzilla.kernel.org/show_bug.cgi?id=204727 Fixes: 607ca3bd220f ("ALSA: hda/realtek - EAPD turn on later") Cc: <stable@vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai@suse.de>
Diffstat (limited to 'sound/pci/hda/hda_generic.c')
-rw-r--r--sound/pci/hda/hda_generic.c3
1 files changed, 2 insertions, 1 deletions
diff --git a/sound/pci/hda/hda_generic.c b/sound/pci/hda/hda_generic.c
index 5bf24fb819d2..10d502328b76 100644
--- a/sound/pci/hda/hda_generic.c
+++ b/sound/pci/hda/hda_generic.c
@@ -6009,7 +6009,8 @@ int snd_hda_gen_init(struct hda_codec *codec)
if (spec->init_hook)
spec->init_hook(codec);
- snd_hda_apply_verbs(codec);
+ if (!spec->skip_verbs)
+ snd_hda_apply_verbs(codec);
init_multi_out(codec);
init_extra_out(codec);
OpenPOWER on IntegriCloud