At Sun, 10 Jul 2011 08:47:46 +0200, Takashi Iwai wrote:

At Sun, 10 Jul 2011 11:31:44 +0800, Raymond Yau wrote:

...

2011/7/6 Takashi Iwai tiwai@suse.de:

...

At Wed, 6 Jul 2011 11:53:47 +0800, Raymond Yau wrote:

...

Refer to commit ce764ab22e40a046065c9417ee5f04ad2a816ac1

ALSA: hda - Add channel-mode support to Realtek auto-parser

It seem that max channel is six instead of two when the channel mode is 2ch for those 3 jacks motherboard

using hda-emu and alsa-info from http://thread.gmane.org/gmane.linux.alsa.devel/86816

...

PCM 0 p 44100 2 16

Open PCM ALC887 Analog for play Available PCM parameters:   channels: 2/6   formats: S16_LE S32_LE   rates: 44100 48000 96000 192000 Prepare PCM, rate=44100, channels=2, format=16 bits

because

spec->multiout.max_channels = spec->multiout.num_dacs * 2;

is assigned after  alc_auto_add_multi_channel_mode() increase spec->multiout.num_dacs from 1 to 3

spec->multiout.num_dacs = num_pins + 1;

Right, and this is intentional.  Since there should be enough amount of DACs, the driver keeps using the surround DACs statically. In that way, apps can keep opening the device in the surround mode while user can switch the outputs on the fly via channel-mode mixer enum.

How can the driver support 6 channels when the input jacks have not been retasked as output ?

Even if the pins aren't retasked, DACs can receive data. And then you can change the pins and routing on the fly.

You can open a two-channel PCM stream even if the master or (any other) volume is muted. A similar reason.

...

This behaviour seem different from model=3stack-6ch or model=3stack-6ch-dig or those via codec with smart51 "OFF"

The smart51 case is a bit different (we prefer DAC unused because of independent HP control), but the 3stack-6ch case is a regression, indeed. This was just because of dropped code snippet. Now I fixed with the patch below.

thanks,

Takashi

--- From: Takashi Iwai tiwai@suse.de Subject: [PATCH] ALSA: hda - Re-add need_dac_fix check for multi-io jacks of Realtek codecs

During the rewrite, the check of spec->need_dac_fix and the corresponding num_dacs change was dropped from the channel-mode control.

This patch re-adds it, and also enables need_dac_fix for ALC880 as default, as this feature was originally introduced to fix h/w bugs of this chip.

Signed-off-by: Takashi Iwai tiwai@suse.de --- sound/pci/hda/patch_realtek.c | 3 +++ 1 files changed, 3 insertions(+), 0 deletions(-)

diff --git a/sound/pci/hda/patch_realtek.c b/sound/pci/hda/patch_realtek.c index 124c63f..52ce075 100644 --- a/sound/pci/hda/patch_realtek.c +++ b/sound/pci/hda/patch_realtek.c @@ -3196,6 +3196,8 @@ static int alc_auto_ch_mode_put(struct snd_kcontrol *kcontrol, for (i = 0; i < spec->multi_ios; i++) alc_set_multi_io(codec, i, i < ch); spec->multiout.max_channels = spec->ext_channel_count; + if (spec->need_dac_fix && !spec->const_channel_count) + spec->multiout.num_dacs = spec->multiout.max_channels / 2; return 1; }

@@ -3642,6 +3644,7 @@ static int patch_alc880(struct hda_codec *codec) codec->spec = spec;

spec->mixer_nid = 0x0b; + spec->need_dac_fix = 1;

board_config = alc_board_config(codec, ALC880_MODEL_LAST, alc880_models, alc880_cfg_tbl);