For codecs that have individual routes going through a loopback mixer
(like VIA codecs), we need to provide an explicit switch to choose
whether the output goes through mixer or directly from DAC.

This patch adds the check for such paths and creates "Loopback Mixing"
enum control when available.

It won't influence on codecs like Realtek or others where the loopback
mixer is connected independently from the primary output routes.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The output paths including aamix should be parsed only for the first
output.  The surround paths including aamix must be wrong, since it
would mix all streams, i.e. all channels would be mixed into a single
and multiplexed again.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Call the path activation for the digital input pin properly, not only
setting the pin control.  Also add spec->digin_path for keeping the
path index.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Instead of search for the path with the certain route at each time,
keep the path index for each output and loopback, and just use it when
referred.

In this implementation, the path index number begins with one, not
zero (although I've been writing in C over decades).  It's just to
make the check for uninitialized values easier.

So far, the input paths aren't handled with indices yet, but still
picked up via snd_hda_get_nid_path() at each time.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The multi-io pins are calculated with a blind assumption of
cfg->line_outs = 1.  This isn't always true.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

When speakers are chosen as the the primary output during evaluation,
we did some tricks to assign the possible multi-io jacks with a
certain offset value to multi_out dacs.  This was a workaround for the
case with multiple speakers like Acer Aspire.  But this is quite ugly
at the same time and the resultant code is hard to understand.  More
badly, it works wrongly for 2.1 speakers like Apple iMac91.

In this patch, instead of fiddling with the offset to multi_out dacs,
simply add a certain badness number if headphone(s) + multi-ios are
possible.  This simplify the code a bit, and it's more robust.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

If the requested path has been already added, return the existing path
instance instead of adding a duplicated instance.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

When the paths are created in map_singles(), we don't have to
re-create new paths in try_assign_dacs().  Just evaluate the badness
and skip to the next item.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Set path->active flag at the path creation time and let the paths
initialized according to the current path->active state in
set_output_and_unmute().  This allows to modify the active flag of
some output paths dynamically, e.g. switching the front output route
with or without aamix like patch_via.c.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

activate_amp() in the generic parser checks whether the given NID is
included in any active paths and skips it if found.  This was a
workaround for avoiding disabling the widgets in the active paths when
one path is disabled, thus it shouldn't be applied to the case for
path activation.  Due to this wrong check, some analog loopback paths
haven't been initialized correctly.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

... as a preliminary work for migrating patch_sigmatel.c.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Manage the connection list cache using linked lists instead of
snd_array, and revive snd_hda_get_conn_list() again, so that we don't
have to keep the expanded values locally.
This will reduce the stack usage by recursive call of
snd_hda_get_conn_index() or parse_nid_path() of the generic parser.

The list management doesn't include any mutex protection, thus the
caller needs to take care of race appropriately.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Yet another broken hardware workaround: there are hardware indicating
the inverted jack detection bit result.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Add the new flag, codec->inv_eapd, indicating that the EAPD
implementation is inverted.

There are always broken hardware in the world.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Similar like the implementation in patch_analog.c and patch_via.c,
the generic parser can provide the independent HP PCM stream now.
It's enabled when spec->indep_hp is set by the caller while parsing.

Currently no dynamic PCM switching as in patch_via.c is implemented
yet.  The control returns -EBUSY when the value is changed during PCM
operations.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Allow the path including the loopback mixer widget in the primary
output channel as an alternative in the generic codec parser.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

... instead of numbers.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

It never showed the 4th line out and headphone pins since quite ago.
Oh well.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Add a better debug print code to show the new assigned paths in
generic parser.  It appears only with CONFIG_SND_DEBUG_VERBOSE=y.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

It's never used in the generic parser.  It was there from the old
Realtek code, which has been dropped quite ago, too.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

When a PCM name string is generated from the chip name, it might
become strange like "CX20549 (Venice) Analog".  In this patch, the
parser tries to drop the invalid words like "(Venice)" in the PCM name
string.  Also, when the name string is given beforehand by the caller,
respect it and use it as is.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

There is no reason to avoid snd_hda_set_pin_ctl_cache() there.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

When spec->own_eapd_ctl isn't set, try to turn on/off EAPD on demand
for each pin.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

There were some old codes that look not stable enough, which was
derived from the old Realtek code.  The initialization for primary
output in init_multi_out() needs to consider the case of shared DAC.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Just a small clean up by splitting a function.
No functional changes at all.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

For preliminary works to migrate the generic parser for Conexant
codecs: the same function is ported to hda_generic.c.
But now it looks through the jack detect table so that it can cover
better.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Add a flag to indicate whether the vmaster mute hook enum is exposed
or not.  Conexant codecs may want not to expose the control depending
on the model.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

... to distinguish from the invalid event type.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Old codecs like AD1986A tend to have long paths as they were just made
to be the way like AC97.  The current max depth 5 can be too short for
such devices.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The access to a cache array element could be invalid outside the
mutex, so copy locally for the later references.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The dirty entry has to be checked at the beginning in the loop, not in
the inner loop for channels.  This caused a regression that the right
channel isn't properly written.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

This would reduce the number of actually executed verbs.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

A bit of details won't hurt.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The bound capture volume and switch controls use the cached amp
updates, but it's missing the flushing at the end.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

It makes easier to understand although these are identical with
snd_hda_codec_resume_*() functions.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The inverted dmic fix overwrites the right channel amp value, but it
would work only when the amp values have been already actually
written.  Put snd_hda_codec_resume_amp() before the amp write for
flushing caches.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Add an overflow check of CORB in HD-audio controller and codec drivers
so that flood of sequential writes would work properly.
In the controller side, add a check of CORB read-pointer to make
returning -EAGAIN when it's full.  Meanwhile in the codec side, when
-EAGAIN error is received, it retries the write after flushing the
pending verbs (calling get_response() essentially does it).
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

These functions are supposed to be called at finishing the cached
sequential writes, so clear the flag properly for lazy developers who
often forget details.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

When verbs or amps are actually written to the hardware, we can clear
dirty flag so that the later snd_hda_codec_resume_*() calls can skip
these verbs / amps.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The commit [2e9bf247: ALSA: hda_codec: Check for invalid zero
connections] trims the whole connection list when an invalid value is
reported by the hardware.  But some codecs (at least AD1986A) may give
a zero NID in the middle of the connection list, so dropping the whole
list isn't good for such cases.

In this patch, as a workaround, allow a single zero NID in the read
connection list.  If it hits zero twice, it's handled as an error, so
that we can avoid "too many connections" errors.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>