On 09/20/2011 05:36 PM, Takashi Iwai wrote:
At Tue, 20 Sep 2011 17:29:22 +0200, David Henningsson wrote:
On 09/20/2011 05:01 PM, Takashi Iwai wrote:
At Tue, 20 Sep 2011 12:26:30 +0200, David Henningsson wrote:
Ok, my automute refactoring has been rebased to your latest tree. I've tested it on my own machine, but I don't know if wider testing is needed to ensure there are no regressions.
Thanks.
alc_simple_setup_automute() calls can be more simplified. The function can check spec->autocfg.line_out_pins[0], hp_pins[0] and speaker_pins[0] to identify detect_hp& co flags. Otherwise, the arguments "1, 0, 0, 1" are hard to understand.
Other than that, the patch looks good to me.
That means a functional change in one place (alc882_quirks.c), where there are line_out_pins but automute for line out is not set:
@@ -2568,8 +2554,7 @@ static void alc888_lenovo_ms7195_setup(struct hda_codec *codec) spec->autocfg.hp_pins[0] = 0x1b; spec->autocfg.line_out_pins[0] = 0x14; spec->autocfg.speaker_pins[0] = 0x15;
spec->automute = 1;spec->automute_mode = ALC_AUTOMUTE_AMP;
alc_simple_setup_automute(spec, ALC_AUTOMUTE_AMP, 1, 0, 0, 1);If you don't mind that change I can do what you suggest.
The above must be a bug. It should follow the standard configuration, too. So you'll fix a bug automagically :)
Ok, attaching modified patch.