The general concept of this change is to create a PCM device for each
pin widget instead of each converter widget. Whenever a PCM is opened,
a converter is dynamically selected to drive that pin based on those
available for muxing into the pin.
The one thing this model doesn't support is a single PCM/converter
sending audio to multiple pin widgets at once.
Note that this means that a struct hda_pcm_stream's nid variable is
set to 0 except between a stream's open and cleanup calls. The dynamic
de-assignment of converters to PCMs occurs within cleanup, not close,
in order for it to co-incide with when controller stream IDs are
cleaned up from converters.
While the PCM for a pin is not open, the pin is disabled (its widget
control's PIN_OUT bit is cleared) so that if the currently routed
converter is used to drive a different PCM/pin, that audio does not
leak out over a disabled pin.
We use the recently added SPDIF virtualization feature in order to
create SPDIF controls for each pin widget instead of each converter
widget, so that state is specific to a PCM.
In order to support this, a number of more mechanical changes are made:
* s/nid/pin_nid/ or s/nid/cvt_nid/ in many places in order to make it
clear exactly what the code is dealing with.
* We now have per_pin and per_cvt arrays in hdmi_spec to store relevant
data. In particular, we store a converter's capabilities in the per_cvt
entry, rather than relying on a combination of codec_pcm_pars and
the struct hda_pcm_stream.
* ELD-related workarounds were removed from hdmi_channel_allocation
into hdmi_instrinsic in order to simplifiy infoframe calculations and
remove HW dependencies.
* Various functions only apply to a single pin, since there is now
only 1 pin per PCM. For example, hdmi_setup_infoframe,
hdmi_setup_stream.
* hdmi_add_pin and hdmi_add_cvt are more oriented at pure codec parsing
and data retrieval, rather than determining which pins/converters
are to be used for creating PCMs.
This is quite a large change; it may be appropriate to simply read the
result of the patch rather than the diffs. Some small parts of the change
might be separable into different patches, but I think the bulk of the
change will probably always be one large patch. Hopefully the change
isn't too opaque!
This has been tested on:
* NVIDIA GeForce 400 series discrete graphics card. This model has the
classical 1:1:1 codec:converter:pcm widget model. Tested stereo PCM
audio to a PC monitor that supports audio.
* NVIDIA GeForce 520 discrete graphics card. This model is the new
1 codec n converters m pins m>n model. Tested stereo PCM audio to a
PC monitor that supports audio.
* NVIDIA GeForce 400 series laptop graphics chip. This model has the
classical 1:1:1 codec:converter:pcm widget model. Tested stereo PCM,
multi-channel PCM, and AC3 pass-through to an AV receiver.
* Intel Ibex Peak laptop. This model is the new 1 codec n converters m
pins m>n model. Tested stereo PCM, multi-channel PCM, and AC3 pass-
through to an AV receiver.
Note that I'm not familiar at all with AC3 pass-through. Hence, I may
not have covered all possible mechanisms that are applicable here. I do
know that my receiver definitely received AC3, not decoded PCM. I tested
with mplayer's "-afm hwac3" and/or "-af lavcac3enc" options, and alsa a
WAV file that I believe has AC3 content rather than PCM.
I also tested:
* Play a stream
* Mute while playing
* Stop stream
* Play some other streams to re-assign the converter to a different
pin, PCM, set of SPDIF controls, ... hence hopefully triggering
cleanup for the original PCM.
* Unmute original stream while not playing
* Play a stream on the original pin/PCM.
This was to test SPDIF control virtualization.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The SPDIF output controls apply to converter widgets. A future change
will create a PCM device per pin widget, and hence a set of SPDIF output
controls per pin widget, for certain HDMI codecs. To support this, we
need the ability to virtualize the SPDIF output controls. Specifically:
* Controls can be "unassigned" from real hardware when a converter is
not used for the PCM the control was created for.
* Control puts only write to hardware when they are assigned.
* Controls can be "assigned" to real hardware when a converter is picked
to support output for a particular PCM.
* When a converter is assigned, the hardware is updated to the cached
configuration.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Currently, the data that backs the kcontrols created by
snd_hda_create_spdif_out_ctls is stored directly in struct hda_codec. When
multiple sets of these controls are stored, they will all manipulate the
same data, causing confusion. Instead, store an array of this data, one
copy per converter, to isolate the controls.
This patch would cause a behavioural change in the case where
snd_hda_create_spdif_out_ctls was called multiple times for a single codec.
As best I can tell, this is never the case for any codec.
This will be relevant at least for some HDMI audio codecs, such as the
NVIDIA GeForce 520 and Intel Ibex Peak. A future change will modify the
driver's handling of those codecs to create multiple PCMs per codec. Note
that this issue isn't affected by whether one creates a PCM-per-converter
or PCM-per-pin; there are multiple of both within a single codec in both
of those codecs.
Note that those codecs don't currently create multiple PCMs for the codec
due to the default HW mux state of all pins being to point at the same
converter, hence there is only a single converter routed to any pin, and
hence only a single PCM.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The connection lists are static and we can reuse the previous results
instead of querying via verb at each time. This will reduce the I/O
in the runtime especially for some codec auto-parsers.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Since multiple codec drivers already use the input-jack stuff, let's
make common helper functions to reduce the duplicated codes.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The sticky PCM stream assignment introduced in 2.6.36 kernel seems
causing problems on AD codecs. At some time later, the streaming no
longer works by unknown reason. A simple workaround is to disable
sticky-assignment for these codecs.
Tested-by: Vasily Khoruzhick <anarsoul@gmail.com>
Cc: <stable@kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Replace the explicit ifdef check and call of check_power_status ops with
a new helper function, hda_call_check_power_status().
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The new sticky PCM parameter introduced the delayed clean-ups of
stream- and channel-id tags. In the current implementation, this check
(adding dirty flag) and actual clean-ups are done only for the codec
chip. However, with HD-audio architecture, multiple codecs can be
on a single bus, and the controller assign stream- and channel-ids in
the bus-wide.
In this patch, the stream-id and channel-id are checked over all codecs
connected to the corresponding bus. Together with it, the mutex is
moved to struct hda_bus, as this becomes also bus-wide.
Reported-and-tested-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The commit eb541337b7
ALSA: hda - Make converter setups sticky
changes the semantics of snd_hda_codec_cleanup_stream() not to clean up
the stream at that moment but delay the action. This broke the codes
expecting that the clean-up is done immediately, such as dynamic ADC
changes in some codec drivers.
This patch fixes the issue by introducing a lower helper,
__snd_hda_codec_cleanup_stream(), to allow the immediate clean up.
The original snd_hda_codec_cleanup_stream() is kept as is now.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
So far, we reset the converter setups like the stream-tag, the
channel-id and format-id in prepare callbacks, and clear them in
cleanup callbacks. This often causes a silence of the digital
receiver for a couple of seconds.
This patch tries to delay the converter setup changes as much as
possible. The converter setups are cached and aren't reset as long
as the same values are used. At suspend/resume, they are cleared
to be recovered properly, too.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Passing IEC 61937 encapsulated compressed audio at bitrates over 6.144
Mbps (i.e. more than a single 2-channel 16-bit 192kHz IEC 60958 link)
over HDMI requires the use of HBR Audio Stream Packets instead of Audio
Sample Packets.
Enable HBR mode when the stream has 8 channels and the Non-PCM bit is
set.
If the audio converter is not connected to any HBR-capable pins, return
-EINVAL in prepare().
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Set bit 15 (Stream Type) of HDA Stream Format to 1 (Non-PCM) when IEC958
channel status bit 1 (AES0 & 0x02) is set to 1 (non-audio).
This is a prequisite for HDMI HBR passthrough.
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The function IDs are different for audio and modem. Do not mix them.
Also, show the unsolicited bit in the function_id register.
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
Many codecs now clear the pin controls at suspend via snd_hda_shutup_pins()
for reducing the click noise at power-off. But this leaves some pins
uninitialized, and they'll be never recovered after resume.
This patch adds the proper recovery of cleared pin controls on resume.
Also it adds a check of bus->shutdown so that pins won't be cleared at
module unloading.
Reference: Kernel bug 16339
http://bugzilla.kernel.org/show_bug.cgi?id=16339
Cc: <stable@kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Add a new helper, snd_hda_codec_update_cache(), for reducing the unneeded
verbs. This function checks the cached value and skips if it's identical
with the given one. Otherwise it works like snd_hda_codec_write_cache().
The alc269 code uses this function as an example.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Add a common helper function for clearing pin controls before suspend.
Use the pincfg array instead of looking through all widget tree.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Analog Device codecs seem to have problems with the triggering of
pin-sensing although their pincaps give the trigger requirements.
Some reported that constant CPU load on HP laptops with AD codecs.
For avoiding this regression, add a flag to codec struct to notify
explicitly that the codec doesn't suppot the trigger at pin-sensing.
Tested-by: Maciej Rutecki <maciej.rutecki@gmail.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This set of changes add missing NID values to some static control
elemenents. Also, it handles all "Capture Source" or "Input Source"
controls.
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
For codecs without EPSS support (G45/IbexPeak), the hotplug event will
be lost if the HDA is powered off during the time. After that the pin
presence detection verb returns inaccurate info.
So always power-on HDA link for !EPSS codecs.
KarL offers the fact and Takashi recommends to flag hda_bus. Thanks!
Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
HDA036 spec states:
DP (Display Port) indicates whether the Pin Complex Widget supports
connection to a Display Port sink. Supported if set to 1. Note that
it is possible for the pin widget to support more than one digital
display connection type, e.g. HDMI and DP bit are both set to 1.
Also export the DP pin cap in procfs.
Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Note that the HBR capability only applies to HDMI pin.
Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The beep_mode parameter for snd-hda-intel module allows to choose among
different digital beep device registation to the input layer.
0 = do not register to the input layer
1 = register to the input layer all time
2 = use "Beep Switch" control exported to user space mixer applications
Also, introduce CONFIG_SND_HDA_INPUT_BEEP_MODE for default value.
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Added the power on/off counter and expose via sysfs files.
The sysfs files, power_on_acct and power_off_acct, are created under
each codec hwdep sysfs directory (e.g. /sys/class/sound/hwC0D0).
The files show the msec length of the codec power-on and power-off,
respectively.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Add reboot notifier to each codec so that it can do some workarounds
needed for reboot.
So far, patch_sigmatel.c calls its shutup routine for avoiding noises
at reboot on some HP machines.
References: Novell bnc#544779
http://bugzilla.novell.com/show_bug.cgi?id=544779
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Recently we hit a bug in our dev board, whose HDMI codec#3 may emit
redundant/spurious responses, which were then taken as responses to
command for another onboard Realtek codec#2, and mess up both codecs.
Extend the azx_rb.cmds and azx_rb.res to array and track each codec's
commands/responses separately. This helps keep good codec safe from
broken ones.
Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Added the patch module option to apply a "patch" as a firmware to
modify pin configurations or give additional hints to the driver
before actually initializing and configuring the codec.
This can be used as a workaround when the BIOS doesn't give sufficient
information or give wrong information that doesn't match with the real
hardware setup, until it's fixed statically in the driver via a quirk.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The codec setup call via snd_hda_codec_configure() isn't necessarily
called in snd_hda_codec_new(). For the later added feature, it's better
to change the code flow like:
- create all codec instances
- configure each codec
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The reset of a BUS controller during operations is somehow risky and
shouldn't be done inevitably for devices that have apparently no such
codec-communication problems.
This patch adds the check of the hardware and limits the bus-reset
capability.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Some machines machine cause a severe CORB/RIRB stall in certain
weird conditions, such as PA access at the start up together with
fglrx driver. This seems unable to be recovered without the controller
reset.
This patch allows the bus controller reset at critical errors so
that the communication gets recovered again.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Split the name string in hda_codec struct to vendor_name and chip_name
strings to be stored directly from the preset name.
Since mostly only the chip name is referred in many patch_*.c, this
results in the reduction of many codes in the end.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The current error-recovery scheme for the codec communication errors
doesn't work always well. Especially falling back to the
single-command mode causes the fatal problem on many systems.
In this patch, the problematic verb is re-issued again after the error
(even with polling mode) instead of the single-cmd mode. The
single-cmd mode will be used only when specified via the command
option explicitly, mainly just for testing.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This patch does two things:
Output Intel HDA Function Id in /proc/asound/cardX/codec#X
Align Vendor/Subsystem/Revision Ids to 8 characters, front-padded with zeros
Before:
Vendor Id: 0x11d41884
Subsystem Id: 0x103c281a
Revision Id: 0x100100
After:
Function Id: 0x1
Vendor Id: 0x11d41884
Subsystem Id: 0x103c281a
Revision Id: 0x0100100
As report on the Kernel Bugzilla #12888
Signed-off-by: Pascal de Bruijn <pascal@unilogicnetworks.net>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Rename from override_pin and cur_pin with user_pin and driver_pin,
respectively, to be a bit more intuitive.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Added the generic pincfg cache and save/restore functions.
Also introduced the pin-overriding via hwdep sysfs.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Some fixes regarding snd-hda-intel workqueue:
- Use create_singlethread_workqueue() instead of create_workqueue()
as per-CPU work isn't required.
- Allocate workq name string properly
- Renamed the workq name to "hd-audio*" to be more obvious.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
snd-hda-intel driver used schedule_work() fot the delayed DMA pointer
updates, but this has several potential problems:
- it may block other eventsd works longer
- it may deadlock when probing fails and flush_scheduled_work() is
called during probe callback (as probe callback itself could be
invoked from eventd)
This patch adds an own workq for each driver instance to solve these
problems.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Fix the following lockdep warning by not reusing the hda_codec.spdif_mutex.
ALSA sound/pci/hda/hda_codec.c:882: hda_codec_cleanup_stream: NID=0x2
=======================================================
[ INFO: possible circular locking dependency detected ]
2.6.28-next-20090102 #33
-------------------------------------------------------
mplayer/3151 is trying to acquire lock:
(&pcm->open_mutex){--..}, at: [<ffffffffa004ced3>] snd_pcm_release+0x43/0xd0 [snd_pcm]
but task is already holding lock:
(&mm->mmap_sem){----}, at: [<ffffffff810c0252>] sys_munmap+0x42/0x80
which lock already depends on the new lock.
Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Added probe_only module option to hd-audio driver.
This option specifies whether the driver creates and initializes the
codec-parser after probing. When this option is set, the driver skips
the codec parsing and initialization but gives you proc and other
accesses. It's useful to see the initial codec state for debugging.
The default of this value is off, so the default behavior is as same
as before.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The global functions in hda_codec.c and other core parts are only
for HD-audio codec and controller drivers. When the HD-audio driver
is built in kernel, all stuff have to be statically linked, thus
we don't need any exports.
This patch introduces a conditional macro to do export only
when needed.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
