Currently we set the max number of connections to be 32, but there
seems codec that gives longer connection lists like AD1988, and we see
errors in proc output and else.  (Though, in the case of AD1988, it's
a list of all codecs connected to a single vendor widget, so this must
be something fishy, but it's still valid from the h/w design POV.)

This patch tries to remove this restriction.  For efficiency, we still
use the fixed size array in the parser, but takes a dynamic array when
the size is reported to be greater than that.

Now the fixed array size is found only in patch_hdmi.c, but it should
be fine, as the codec itself can't support so many pins.
Reported-by: NRaymond Yau <superquad.vortex2@gmail.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

... for small refactoring.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Put the power state synchronization at the end of the parsing of
codec.  This is necessary when the power filter is changed during the
codec probe.  Since the first power-up sequence is performed without
the special filter, all widgets are supposed to be ON at this point.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Add a hook to struct hda_codec for filtering the target power state of
each widget when powering up/down.  The current hackish EAPD check is
implemented as the default hook pointer, too.

This allows codec drivers to implement own power filter.  In the
upcoming changes, the generic parser will have the better power filter
based on the active paths.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The commit [26a6cb6c: ALSA: hda - Implement a poll loop for jacks as a
module parameter] introduced the polling jack detection code, but it
also moved the call of snd_hda_jack_set_dirty_all() in the resume path
after resume/init ops call.  This caused a regression when the jack
state has been changed during power-down (e.g. in the power save
mode).  Since the driver doesn't probe the new jack state but keeps
using the cached value due to no dirty flag, the pin state remains
also as if the jack is still plugged.

The fix is simply moving snd_hda_jack_set_dirty_all() to the original
position.
Reported-by: NManolo Díaz <diaz.manolo@gmail.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Since both snd_hda_codec_flush_amp_cache() and
snd_hda_codec_flush_cmd_cache() are called usually at the same time,
we can simply combine them to a single function,
snd_hda_codec_flush_cache().
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Obvious copy-paste error.
Signed-off-by: NDavid Henningsson <david.henningsson@canonical.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The pincfgs, init_verbs and hints set by sysfs or patch might be
changed dynamically on the fly, thus we need to protect it.
Add a simple protection via a mutex.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Check more strictly about the validity of pinctl values in
snd_hda_set_pin_ctl() and correct the wrong bits automatically.
Also provide the helper function to correct pinctl bits to codec
drivers.

This automatically fixes the invalid pinctl writes that are found in
a few Realtek fixups for NID 0x0f amp like ASUS A6Rp.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

We already have the list of whole pin widgets and there is an unused
space in the list; let's use it for caching the current pinctl value.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

We don't have to set up a stream that has been already set up
previously.
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>

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>

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>

Since the fixup code is used commonly, it's worth to move it to the
common place, struct hda_codec, instead of keeping in hda_gen_spec.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The new function snd_hda_codec_amp_init() (and the stereo variant)
initializes the amp value only once at the first access.  If the amp
was already initialized or updated, this won't do anything more.

It's useful for initializing the input amps that are in the part of
the path but never used.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

For optimizing the verb executions, a new mechanism to cache the verbs
and amp update commands is introduced.  With the new "write to cache
and flush" way, you can reduce the same verbs that have been written
multiple times.

When codec->cached_write flag is set, the further
snd_hda_codec_write_cache() and snd_hda_codec_amp_stereo() calls will
be performed only on the command or amp cache table, but not sent to
the hardware yet.  Once after you call all commands and update amps,
call snd_hda_codec_resume_amp() and snd_hda_codec_resume_cache().
Then all cached writes and amp updates will be written to the
hardware, and the dirty flags are cleared.

In this implementation, the existing cache table is reused, so
actually no big code change is seen here.  Each cache entry has a new
dirty flag now (so the cache key is now reduced to 31bit).

As a good side-effect by this change, snd_hda_codec_resume_*() will no
longer execute verbs that have been already issued during the resume
phase by checking the dirty flags.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The vmaster hook sends 1 for enabled/unmuted and 0 for disabled/muted,
but "Mute-LED Mode" being "On" refers to the LED being on, not the
volume being on.
Therefore "On" and "Off" should be switched.
Signed-off-by: NDavid Henningsson <david.henningsson@canonical.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Remove the leftover __devinit* in comments.
They have been commented out because they couldn't fit with __dev*
although they should have matched.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The same type of code is being used in multiple places in various
codec drivers, so put it as a core library.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

This is a preliminary patch for introducing a protection to access
races of snd_array instances.  Call snd_array_init() appropriately
at the initialization time and don't call it twice.

Also the allocations of codec-spec structs are cleaned up by helper
functions in patch_sigmatel.c and patch_analog.c.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

I forgot this again...  codec->in_pm is in #ifdef CONFIG_PM
Reported-by: NMarkus Trippelsdorf <markus@trippelsdorf.de>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Since we keep the pin default config values anyway internally, we
don't have to set the values in the codec.  This patch removes the
code writing the pincfg values.

As a gratis bonus, we can remove also the code restoring the original
pincfg values at PM resume or module free.  This will give us more
benefit, as it can reduce the unnecessary power-up of codecs.

This won't change the driver functionality.  The only difference would
be that the codec proc file will show the original pincfg values
instead of the actually referred values.  The actually referred values
can be determined from sysfs *_pin_configs files.
(Also hda-emu was updated to follow this change.)
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

When the bus reset is performed during the suspend/resume (including
the power-saving too), it calls snd_hda_suspend() and
snd_hda_resume() again, and deadlocks eventually.

For avoiding the recursive call, add a new flag indicating that the PM
is being performed, and don't go to the bus reset mode when it's on.
Reported-and-tested-by: NJulian Wollrath <jwollrath@web.de>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

When 2.1 speakers are detected, use the corresponding channel map
instead of the standard map with front+rear surrounds.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Starting audio or seeking in various music players causes
setup_dig_out_stream() to be called, which resets the SPDIF stream,
which caused one DAC (but not another) to make a clicking noise every
time.

This patch ensures the reset only happens when it needs to, which is
when the format changes, and makes the code a little more readable.
Signed-off-by: NLaurence Darby <ldarby@tuffmail.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

When both an SPDIF and an HDMI device are created on the same card
instance, multiple IEC958 controls are created with indices=0, 1, ...
But the alsa-lib configuration can't know which index corresponds
actually to which PCM device, and both the SPDIF and the HDMI
configurations point to the first IEC958 control wrongly.

This patch introduces a (hackish and ugly) workaround: the IEC958
controls for the SPDIF device are re-labeled with device=1 when HDMI
coexists.  The device=1 corresponds to the actual PCM device for
SPDIF, so it's anyway a better representation.  In future, HDMI
controls should be moved with the corresponding PCM device number,
too.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Now that we have a generic unsol mechanism, we can implement a generic
poll loop, which can be used for debugging, or if a codec's unsol
mechanism is broken.
Signed-off-by: NDavid Henningsson <david.henningsson@canonical.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Better to add a sanity check as I tend to forget something (especially
during crazy midsummer nights).
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

There was a race condition when the system suspends while hda_power_work
is running in the work queue.  If system suspend (snd_hda_suspend)
happens after the work queue releases power_lock but before it calls
hda_call_codec_suspend,  codec_suspend runs with power_on=0, causing the
codec to power up for register reads, and hanging when it calls
cancel_delayed_work_sync from the running work queue.

The call chain from the work queue will look like this:
hda_power_work <<- power_on = 1, unlock, then power_on cleard by suspend
  hda_call_codec_suspend
    hda_set_power_state
      snd_hda_codec_read
        codec_exec_verb
          snd_hda_power_up
	    snd_hda_power_save
	      __snd_hda_power_up
	        cancel_delayed_work_sync <<-- cancelling executing wq

Fix this by waiting for the work queue to finish before starting suspend
if suspend is not happening on the work queue.
Signed-off-by: NDylan Reid <dgreid@chromium.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

snd_hda_codec_reset() calls restore_pincfgs() where the codec is
powered up again, which eventually tries to resume and initialize via
the callbacks of the codec.  However, it's the place just after codec
free callback, thus no codec callbacks should be called after that.
On a codec like CS4206, it results in Oops due to the access in init
callback.

This patch fixes the issue by clearing the codec callbacks properly
after freeing codec.
Reported-by: NDaniel J Blueman <daniel@quora.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Although HD-audio allows pair-wise channel configurations, only the
fixed channel positions are used in this version.  In future, this can
be changed and allow user to modify the channel positions.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Just refactoring, no functional changes.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The call of pm_notify callback in snd_hda_codec_free() should be with
the check of the current state whether pm_notify(false) is called or
not, instead of codec->power_on check.

For improving the code readability and fixing this inconsistency,
codec->d3_stop_clk_ok is renamed to codec->pm_down_notified, and this
flag is set only when runtime PM down is called.  The new name reflects
to a more direct purpose of the flag.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

CONFIG_SND_HDA_POWER_SAVE is no longer an experimental feature and its
behavior can be well controlled via the default value and module
parameter.  Let's just replace it with the standard CONFIG_PM.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>