The codec ALC272X is a special codec for some Dell machines, and its
detection got broken in the recent kernel because SSID check (required
by ALC272X check) was moved to the later point.  Now we need to move
this codec check to the right place, too.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

During the split to the auto-parser helper functions, the actual call
of init verbs was lost.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=43366Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Pin 0x1b was connected to the front panel connector, which according to
the HDA standard should contain a mic and a headphone. In this case,
the headphone was listed as "line out" by BIOS.

Cc: stable@kernel.org
BugLink: https://bugs.launchpad.net/bugs/993162Signed-off-by: NDavid Henningsson <david.henningsson@canonical.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Also fix some settings applied only for ALC269VB.
Signed-off-by: NKailang Yang <kailang@realtek.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

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

Like the previous fixes for cache hash accesses, a protection over
accesses to the widget connection list array must be provided.
Together with this action, remove snd_hda_get_conn_list() which can be
always race, and replace it with either snd_hda_get_num_conns() or
snd_hda_get_connections() calls.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Move the fixup helper functions in patch_realtek.c to hda_auto_parser.c
so that they can be used in other codec drivers like patch_conexant.c.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Just a clean up by calling the same helper function.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The call for alc_auto_parse_customize_define() must be done after the
fixup pre-probe initialization.  Otherwise SKU_IGNORE fixup won't work
properly (e.g. HP RP5800 with ALC662 codec).
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Reported-by: NPhilipp Matthias Hahn <pmhahn@pmhahn.de>
Cc: <stable@kernel.org> [v3.3+]
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Acer Aspire 5739G requires the same fix-up for 4930G to support the
surround / bass speakers.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=43180Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Add a new helper function to guess the default VREF pin control bits
for mic in.  This can be used to set the pin control value safely
matching with the actual pin capabilities.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

For setting the pin-control values more safely to match with the
actual pin capability bits, a copule of new helper functions,
snd_hda_set_pin_ctl() and snd_hda_set_pin_ctl_cache(), are
introduced.  These are simple replacement of the codec verb write with
AC_VERB_SET_PIN_WIDGET but do more sanity checks and filter out
superfluous pin-control bits if they don't fit with the corresponding
pin capabilities.

Some codecs are screwed up or ignore the command when such a wrong bit
is set.  These helpers will avoid such secret errors.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

According to the reporter, external mic starts to work if the
laptop-dmic model is used. According to BIOS pin config, all
pins are consistent with the alc269vb_laptop_dmic fixup, except
for the external mic, which is not present.

Cc: stable@kernel.org
BugLink: https://bugs.launchpad.net/bugs/950490Signed-off-by: NDavid Henningsson <david.henningsson@canonical.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Through the transition to the auto-parser, the support for
Quanta/Gericom KN1 got broken.  There are two problems behind it:

- This machine doesn't like the default COEF setup for ALC260 we take
  now as default

- BIOS doesn't set the pins correctly at all; especially the machine
  uses only the pin 0x0f for both headphone and speaker

This patch adds the fixup as a workaround for these issues.
Reported-and-tested-by: NUros Vampl <mobile.leecher@gmail.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

In sound/pci/hda/patch_realtek.c::alc_auto_fill_dac_nids(), in the
'for (;;)' loop, if the 'badness' value returned from
fill_and_eval_dacs() is negative, then we'll return from the function
without freeing the memory we allocated for 'best_cfg', thus leaking.
Fix the leak by kfree()'ing the memory when badness is negative.

While I was there I also noticed some trailing whitespace in the
function that I removed (along with all other trailing whitespace in
the file) - it didn't seem worth-while to do that as two patches, so I
hope it's OK that I just did it all as one patch.
Signed-off-by: NJesper Juhl <jj@chaosbits.net>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

A user reported that setting model=imac24 used to allow sound to work on their
Mac Pro 5,1 machine.  Commit 5671087f "Move ALC885 macpro and imac24 models
to auto-parser" removed this model option.  All Mac machines are now explicitly
handled with a quirk and the auto-parser.  This adds a quirk for the device
found on the Mac Pro 5,1 machines.

This (partially) fixes https://bugzilla.redhat.com/show_bug.cgi?id=808559

[sorted the new entry in the ID number order by tiwai]
Reported-by: NGabriel Somlo <somlo@cmu.edu>
Signed-off-by: NJosh Boyer <jwboyer@redhat.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

It's compatible with 8930G.
Using the same fixup gives the proper 5.1 sound back.
Reported-and-tested-by: NDany Martineau <dany.luc.martineau@gmail.com>
Cc: <stable@kernel.org> [v3.3+]
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Add GPIO1 setup explicitly for Acer Aspire 493x & co.
This could be set by alc_auto_init_amp(), but it's safer to set it
more explicitly in the fixup table.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Since there are still many Acer models that might not be covered by
the current fixup table, let's add back a few typical model names so
that user can test the fixup without recompiling.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

This patch updates Jonathan Woithe's contact details across the kernel tree.
Signed-off-by: NJonathan Woithe <jwoithe@just42.net>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

The recent Realtek driver tries to assign an extra input via the
headphone plug when only a single input source is found.  The code
worked on Samsung Q1, but it broke ASUS 1015 where the mic is a
digital-mic and only a specific ADC works.

This patch fixes the assignment of ADC in the shared mic/hp case.
Instead of assuming the single ADC at first, reduce the ADCs after
trying to assign both mic and HP pins.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=42973

Cc: <stable@kernel.org>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Since it's not always safe to assume that the vmaster hook is purely the
mute-LED control, add the flag indicating whether to expose the mute-LED
enum control or not.  Currently, conexant codec sets this off for non-HP
laptops where EAPD may be used really as EAPD.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Create snd_hda_add_vmaster_hook() and snd_hda_sync_vmaster_hook()
helper functions to handle the mute-LED in vmaster hook more
commonly.  In the former function, a new enum control "Mute-LED Mode"
is added.  This provides user to choose whether the mute-LED should be
turned on/off explicitly or to follow the master-mute status.
Reviewed-by: NDavid Henningsson <david.henningsson@canonical.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

We've had ugly static handling of the mute-LED with a powersave hook
for ALC269 HP laptops just like done in patch_sigmatel.c.  This is now
rewritten with the new vmaster hook and a fixup code.

For that, the new fixup action, ALC_FIXUP_ACT_BUILD, is introduced.
It's called after build_controls is called.  The reason of this new
action is that vmaster hook must be added at this stage (not in init
or probe).
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Move the currently unused spec->init_hook at the beginning of the init
sequence so that the recently added ALC269VB coef setup can be put
there.  The alc_init() is again clean without an ugly check.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The coef setup in alc269_fill_coef() was designed only for ALC269VB
model, and this has some bad effects for other ALC269 variants, such
as turning off the external mic input.  Apply it only to ALC269VB.
Signed-off-by: NKailang Yang <kailang@realtek.com>
Cc: <stable@kernel.org>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Its BIOS suppresses the PC beep although it's implemented.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

My CD input got lost in commit 68ef0561.
Raymond helped me to add the necessary pin fixup to make it appear again. In
fact, this is basically his patch. It fixes alsa bug #5541.
Signed-off-by: NMarton Balint <cus@passwd.hu>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

The recent addition of volume-knob widget in the auto-parser broke
automute/autoswitch for some Realtek devices.
Signed-off-by: NDavid Henningsson <david.henningsson@canonical.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Kill hyphens from "Line-Out" name strings, as suggested by Mark Brown.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

When there are multiple input sources, the driver wrongly overwrites with
the value of the last input source on other slots at resume.  Thus the
primary input source may be shown wrongly.
Reported-and-tested-by: NJulian Sikorski <belegdol@gmail.com>
Cc: <stable@kernel.org> [v3.1+]
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

A few clean-ups for post-static-quirk time:
- Call alc_auto_init_std() statically in alc_init() instead of setting
  spec->init_hook in each caller.  spec->init_hook field is left
  unused for any future use.

- Move the call of set_capture_mixer() to to alc_parse_auto_config()
  instead of each caller.

- Get rid of the ADC-filling and imux check in each parser function.
  This is no longer needed since the auto-parser always check ADCs and
  imux.  It was only for the static quirks.

- Kill unused defines
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

When the machine has two speakers but wants to put more multi-io
jacks, the parser shouldn't consider about the shared DAC but try to
assign the individual DACs.  Otherwise the channel mapping would be
fairly confused and lead to the wrong DACs.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

When the multi-io is added to the two speaker output configuration,
the parser would try to add yet another "Bass Speaker" control since
it checks only cfg->line_outs.  Add a workaround for it by simply
passing the channel name in the case of multi-io outputs.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Even if the outputs are using shared DACs, we can still create individual
mute siwtches since they are assigned per pin.  This allows to create,
e.g. Speaker and Bass Speaker mute switches while the single volume is
used for these outputs.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

When BIOS is damn crazy and gives no pin-config at all, the driver might
lead to a NULL dereference.  Let's add a NULL check for such a case.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Similarly in patch_via.c, parse the active analog-loopback connections
and create a list dynamically rather than static arrays.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

Resitance is futile.  The remaining static model quirks for Apple
machines with ALC882-compatible codecs are converted to the auto-parser
now.  We can remove all alc*_quirks.c finally.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

If anyone wants to debug the driver and avoid the existing fix-ups,
pass model=nofixup option.  Then the driver will skip to pick up the
fixup list.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>