1 MORE NOTES ON HD-AUDIO DRIVER
2 =============================
3 Takashi Iwai <tiwai@suse.de>
9 HD-audio is the new standard on-board audio component on modern PCs
10 after AC97. Although Linux has been supporting HD-audio since long
11 time ago, there are often problems with new machines. A part of the
12 problem is broken BIOS, and the rest is the driver implementation.
13 This document explains the brief trouble-shooting and debugging
14 methods for the HD-audio hardware.
16 The HD-audio component consists of two parts: the controller chip and
17 the codec chips on the HD-audio bus. Linux provides a single driver
18 for all controllers, snd-hda-intel. Although the driver name contains
19 a word of a well-known hardware vendor, it's not specific to it but for
20 all controller chips by other companies. Since the HD-audio
21 controllers are supposed to be compatible, the single snd-hda-driver
22 should work in most cases. But, not surprisingly, there are known
23 bugs and issues specific to each controller type. The snd-hda-intel
24 driver has a bunch of workarounds for these as described below.
26 A controller may have multiple codecs. Usually you have one audio
27 codec and optionally one modem codec. In theory, there might be
28 multiple audio codecs, e.g. for analog and digital outputs, and the
29 driver might not work properly because of conflict of mixer elements.
30 This should be fixed in future if such hardware really exists.
32 The snd-hda-intel driver has several different codec parsers depending
33 on the codec. It has a generic parser as a fallback, but this
34 functionality is fairly limited until now. Instead of the generic
35 parser, usually the codec-specific parser (coded in patch_*.c) is used
36 for the codec-specific implementations. The details about the
37 codec-specific problems are explained in the later sections.
39 If you are interested in the deep debugging of HD-audio, read the
40 HD-audio specification at first. The specification is found on
41 Intel's web page, for example:
43 - http://www.intel.com/standards/hdaudio/
51 The most common problem of the controller is the inaccurate DMA
52 pointer reporting. The DMA pointer for playback and capture can be
53 read in two ways, either via a LPIB register or via a position-buffer
54 map. As default the driver tries to read from the io-mapped
55 position-buffer, and falls back to LPIB if the position-buffer appears
56 dead. However, this detection isn't perfect on some devices. In such
57 a case, you can change the default method via `position_fix` option.
59 `position_fix=1` means to use LPIB method explicitly.
60 `position_fix=2` means to use the position-buffer. 0 is the default
61 value, the automatic check and fallback to LPIB as described in the
62 above. If you get a problem of repeated sounds, this option might
65 In addition to that, every controller is known to be broken regarding
66 the wake-up timing. It wakes up a few samples before actually
67 processing the data on the buffer. This caused a lot of problems, for
68 example, with ALSA dmix or JACK. Since 2.6.27 kernel, the driver puts
69 an artificial delay to the wake up timing. This delay is controlled
70 via `bdl_pos_adj` option.
72 When `bdl_pos_adj` is a negative value (as default), it's assigned to
73 an appropriate value depending on the controller chip. For Intel
74 chips, it'd be 1 while it'd be 32 for others. Usually this works.
75 Only in case it doesn't work and you get warning messages, you should
76 change this parameter to other values.
81 A less often but a more severe problem is the codec probing. When
82 BIOS reports the available codec slots wrongly, the driver gets
83 confused and tries to access the non-existing codec slot. This often
84 results in the total screw-up, and destructs the further communication
85 with the codec chips. The symptom appears usually as error messages
87 ------------------------------------------------------------------------
88 hda_intel: azx_get_response timeout, switching to polling mode:
90 hda_intel: azx_get_response timeout, switching to single_cmd mode:
92 ------------------------------------------------------------------------
94 The first line is a warning, and this is usually relatively harmless.
95 It means that the codec response isn't notified via an IRQ. The
96 driver uses explicit polling method to read the response. It gives
97 very slight CPU overhead, but you'd unlikely notice it.
99 The second line is, however, a fatal error. If this happens, usually
100 it means that something is really wrong. Most likely you are
101 accessing a non-existing codec slot.
103 Thus, if the second error message appears, try to narrow the probed
104 codec slots via `probe_mask` option. It's a bitmask, and each bit
105 corresponds to the codec slot. For example, to probe only the first
106 slot, pass `probe_mask=1`. For the first and the third slots, pass
107 `probe_mask=5` (where 5 = 1 | 4), and so on.
109 Since 2.6.29 kernel, the driver has a more robust probing method, so
110 this error might happen rarely, though.
112 On a machine with a broken BIOS, sometimes you need to force the
113 driver to probe the codec slots the hardware doesn't report for use.
114 In such a case, turn the bit 8 (0x100) of `probe_mask` option on.
115 Then the rest 8 bits are passed as the codec slots to probe
116 unconditionally. For example, `probe_mask=0x103` will force to probe
117 the codec slots 0 and 1 no matter what the hardware reports.
122 In rare but some cases, the interrupt isn't properly handled as
123 default. You would notice this by the DMA transfer error reported by
124 ALSA PCM core, for example. Using MSI might help in such a case.
125 Pass `enable_msi=1` option for enabling MSI.
133 The most common problem regarding the HD-audio driver is the
134 unsupported codec features or the mismatched device configuration.
135 Most of codec-specific code has several preset models, either to
136 override the BIOS setup or to provide more comprehensive features.
138 The driver checks PCI SSID and looks through the static configuration
139 table until any matching entry is found. If you have a new machine,
140 you may see a message like below:
141 ------------------------------------------------------------------------
142 hda_codec: Unknown model for ALC880, trying auto-probe from BIOS...
143 ------------------------------------------------------------------------
144 Even if you see such a message, DON'T PANIC. Take a deep breath and
145 keep your towel. First of all, it's an informational message, no
146 warning, no error. This means that the PCI SSID of your device isn't
147 listed in the known preset model (white-)list. But, this doesn't mean
148 that the driver is broken. Many codec-drivers provide the automatic
149 configuration mechanism based on the BIOS setup.
151 The HD-audio codec has usually "pin" widgets, and BIOS sets the default
152 configuration of each pin, which indicates the location, the
153 connection type, the jack color, etc. The HD-audio driver can guess
154 the right connection judging from these default configuration values.
155 However -- some codec-support codes, such as patch_analog.c, don't
156 support the automatic probing (yet as of 2.6.28). And, BIOS is often,
157 yes, pretty often broken. It sets up wrong values and screws up the
160 The preset model is provided basically to overcome such a situation.
161 When the matching preset model is found in the white-list, the driver
162 assumes the static configuration of that preset and builds the mixer
163 elements and PCM streams based on the static information. Thus, if
164 you have a newer machine with a slightly different PCI SSID from the
165 existing one, you may have a good chance to re-use the same model.
166 You can pass the `model` option to specify the preset model instead of
169 What `model` option values are available depends on the codec chip.
170 Check your codec chip from the codec proc file (see "Codec Proc-File"
171 section below). It will show the vendor/product name of your codec
172 chip. Then, see Documentation/sound/alsa/HD-Audio-Models.txt file,
173 the section of HD-audio driver. You can find a list of codecs
174 and `model` options belonging to each codec. For example, for Realtek
175 ALC262 codec chip, pass `model=ultra` for devices that are compatible
176 with Samsung Q1 Ultra.
178 Thus, the first thing you can do for any brand-new, unsupported and
179 non-working HD-audio hardware is to check HD-audio codec and several
180 different `model` option values. If you have any luck, some of them
181 might suit with your device well.
183 Some codecs such as ALC880 have a special model option `model=test`.
184 This configures the driver to provide as many mixer controls as
185 possible for every single pin feature except for the unsolicited
186 events (and maybe some other specials). Adjust each mixer element and
187 try the I/O in the way of trial-and-error until figuring out the whole
190 Note that `model=generic` has a special meaning. It means to use the
191 generic parser regardless of the codec. Usually the codec-specific
192 parser is much better than the generic parser (as now). Thus this
193 option is more about the debugging purpose.
196 Speaker and Headphone Output
197 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
198 One of the most frequent (and obvious) bugs with HD-audio is the
199 silent output from either or both of a built-in speaker and a
200 headphone jack. In general, you should try a headphone output at
201 first. A speaker output often requires more additional controls like
202 the external amplifier bits. Thus a headphone output has a slightly
205 Before making a bug report, double-check whether the mixer is set up
206 correctly. The recent version of snd-hda-intel driver provides mostly
207 "Master" volume control as well as "Front" volume (where Front
208 indicates the front-channels). In addition, there can be individual
209 "Headphone" and "Speaker" controls.
211 Ditto for the speaker output. There can be "External Amplifier"
212 switch on some codecs. Turn on this if present.
214 Another related problem is the automatic mute of speaker output by
215 headphone plugging. This feature is implemented in most cases, but
216 not on every preset model or codec-support code.
218 In anyway, try a different model option if you have such a problem.
219 Some other models may match better and give you more matching
220 functionality. If none of the available models works, send a bug
221 report. See the bug report section for details.
223 If you are masochistic enough to debug the driver problem, note the
226 - The speaker (and the headphone, too) output often requires the
227 external amplifier. This can be set usually via EAPD verb or a
228 certain GPIO. If the codec pin supports EAPD, you have a better
229 chance via SET_EAPD_BTL verb (0x70c). On others, GPIO pin (mostly
230 it's either GPIO0 or GPIO1) may turn on/off EAPD.
231 - Some Realtek codecs require special vendor-specific coefficients to
232 turn on the amplifier. See patch_realtek.c.
233 - IDT codecs may have extra power-enable/disable controls on each
234 analog pin. See patch_sigmatel.c.
235 - Very rare but some devices don't accept the pin-detection verb until
236 triggered. Issuing GET_PIN_SENSE verb (0xf09) may result in the
237 codec-communication stall. Some examples are found in
243 The capture problems are often because of missing setups of mixers.
244 Thus, before submitting a bug report, make sure that you set up the
245 mixer correctly. For example, both "Capture Volume" and "Capture
246 Switch" have to be set properly in addition to the right "Capture
247 Source" or "Input Source" selection. Some devices have "Mic Boost"
250 When the PCM device is opened via "default" PCM (without pulse-audio
251 plugin), you'll likely have "Digital Capture Volume" control as well.
252 This is provided for the extra gain/attenuation of the signal in
253 software, especially for the inputs without the hardware volume
254 control such as digital microphones. Unless really needed, this
255 should be set to exactly 50%, corresponding to 0dB -- neither extra
256 gain nor attenuation. When you use "hw" PCM, i.e., a raw access PCM,
257 this control will have no influence, though.
259 It's known that some codecs / devices have fairly bad analog circuits,
260 and the recorded sound contains a certain DC-offset. This is no bug
263 Most of modern laptops have no analog CD-input connection. Thus, the
264 recording from CD input won't work in many cases although the driver
265 provides it as the capture source. Use CDDA instead.
267 The automatic switching of the built-in and external mic per plugging
268 is implemented on some codec models but not on every model. Partly
269 because of my laziness but mostly lack of testers. Feel free to
270 submit the improvement patch to the author.
275 If no model option gives you a better result, and you are a tough guy
276 to fight against evil, try debugging via hitting the raw HD-audio
277 codec verbs to the device. Some tools are available: hda-emu and
278 hda-analyzer. The detailed description is found in the sections
279 below. You'd need to enable hwdep for using these tools. See "Kernel
280 Configuration" section.
288 In general, I recommend you to enable the sound debug option,
289 `CONFIG_SND_DEBUG=y`, no matter whether you are debugging or not.
290 This enables snd_printd() macro and others, and you'll get additional
291 kernel messages at probing.
293 In addition, you can enable `CONFIG_SND_DEBUG_VERBOSE=y`. But this
294 will give you far more messages. Thus turn this on only when you are
297 Don't forget to turn on the appropriate `CONFIG_SND_HDA_CODEC_*`
298 options. Note that each of them corresponds to the codec chip, not
299 the controller chip. Thus, even if lspci shows the Nvidia controller,
300 you may need to choose the option for other vendors. If you are
301 unsure, just select all yes.
303 `CONFIG_SND_HDA_HWDEP` is a useful option for debugging the driver.
304 When this is enabled, the driver creates hardware-dependent devices
305 (one per each codec), and you have a raw access to the device via
306 these device files. For example, `hwC0D2` will be created for the
307 codec slot #2 of the first card (#0). For debug-tools such as
308 hda-verb and hda-analyzer, the hwdep device has to be enabled.
309 Thus, it'd be better to turn this on always.
311 `CONFIG_SND_HDA_RECONFIG` is a new option, and this depends on the
312 hwdep option above. When enabled, you'll have some sysfs files under
313 the corresponding hwdep directory. See "HD-audio reconfiguration"
316 `CONFIG_SND_HDA_POWER_SAVE` option enables the power-saving feature.
317 See "Power-saving" section below.
322 The codec proc-file is a treasure-chest for debugging HD-audio.
323 It shows most of useful information of each codec widget.
325 The proc file is located in /proc/asound/card*/codec#*, one file per
326 each codec slot. You can know the codec vendor, product id and
327 names, the type of each widget, capabilities and so on.
328 This file, however, doesn't show the jack sensing state, so far. This
329 is because the jack-sensing might be depending on the trigger state.
331 This file will be picked up by the debug tools, and also it can be fed
332 to the emulator as the primary codec information. See the debug tools
335 This proc file can be also used to check whether the generic parser is
336 used. When the generic parser is used, the vendor/product ID name
337 will appear as "Realtek ID 0262", instead of "Realtek ALC262".
340 HD-Audio Reconfiguration
341 ~~~~~~~~~~~~~~~~~~~~~~~~
342 This is an experimental feature to allow you re-configure the HD-audio
343 codec dynamically without reloading the driver. The following sysfs
344 files are available under each codec-hwdep device directory (e.g.
345 /sys/class/sound/hwC0D0):
348 Shows the 32bit codec vendor-id hex number. You can change the
349 vendor-id value by writing to this file.
351 Shows the 32bit codec subsystem-id hex number. You can change the
352 subsystem-id value by writing to this file.
354 Shows the 32bit codec revision-id hex number. You can change the
355 revision-id value by writing to this file.
357 Shows the AFG ID. This is read-only.
359 Shows the MFG ID. This is read-only.
361 Shows the codec name string. Can be changed by writing to this
364 Shows the currently set `model` option. Can be changed by writing
367 The extra verbs to execute at initialization. You can add a verb by
368 writing to this file. Pass three numbers: nid, verb and parameter
369 (separated with a space).
371 Shows / stores hint strings for codec parsers for any use.
372 Its format is `key = value`. For example, passing `hp_detect = yes`
373 to IDT/STAC codec parser will result in the disablement of the
376 Shows the initial pin default config values set by BIOS.
378 Shows the pin default values set by the codec parser explicitly.
379 This doesn't show all pin values but only the changed values by
380 the parser. That is, if the parser doesn't change the pin default
381 config values by itself, this will contain nothing.
383 Shows the pin default config values to override the BIOS setup.
384 Writing this (with two numbers, NID and value) appends the new
385 value. The given will be used instead of the initial BIOS value at
386 the next reconfiguration time. Note that this config will override
387 even the driver pin configs, too.
389 Triggers the codec re-configuration. When any value is written to
390 this file, the driver re-initialize and parses the codec tree
391 again. All the changes done by the sysfs entries above are taken
394 Resets the codec, removes the mixer elements and PCM stuff of the
395 specified codec, and clear all init verbs and hints.
397 For example, when you want to change the pin default configuration
398 value of the pin widget 0x14 to 0x9993013f, and let the driver
399 re-configure based on that state, run like below:
400 ------------------------------------------------------------------------
401 # echo 0x14 0x9993013f > /sys/class/sound/hwC0D0/user_pin_configs
402 # echo 1 > /sys/class/sound/hwC0D0/reconfig
403 ------------------------------------------------------------------------
408 The power-saving is a kind of auto-suspend of the device. When the
409 device is inactive for a certain time, the device is automatically
410 turned off to save the power. The time to go down is specified via
411 `power_save` module option, and this option can be changed dynamically
414 The power-saving won't work when the analog loopback is enabled on
415 some codecs. Make sure that you mute all unneeded signal routes when
416 you want the power-saving.
418 The power-saving feature might cause audible click noises at each
419 power-down/up depending on the device. Some of them might be
420 solvable, but some are hard, I'm afraid. Some distros such as
421 openSUSE enables the power-saving feature automatically when the power
422 cable is unplugged. Thus, if you hear noises, suspect first the
423 power-saving. See /sys/module/snd_hda_intel/parameters/power_save to
424 check the current value. If it's non-zero, the feature is turned on.
429 The latest development codes for HD-audio are found on sound git tree:
431 - git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6.git
433 The master branch or for-next branches can be used as the main
434 development branches in general while the HD-audio specific patches
435 are committed in topic/hda branch.
437 If you are using the latest Linus tree, it'd be better to pull the
438 above GIT tree onto it. If you are using the older kernels, an easy
439 way to try the latest ALSA code is to build from the snapshot
440 tarball. There are daily tarballs and the latest snapshot tarball.
441 All can be built just like normal alsa-driver release packages, that
442 is, installed via the usual spells: configure, make and make
443 install(-modules). See INSTALL in the package. The snapshot tarballs
446 - ftp://ftp.kernel.org/pub/linux/kernel/people/tiwai/snapshot/
451 If any model or module options don't work for your device, it's time
452 to send a bug report to the developers. Give the following in your
455 - Hardware vendor, product and model names
456 - Kernel version (and ALSA-driver version if you built externally)
457 - `alsa-info.sh` output; run with `--no-upload` option. See the
458 section below about alsa-info
460 If it's a regression, at best, send alsa-info outputs of both working
461 and non-working kernels. This is really helpful because we can
462 compare the codec registers directly.
464 Send a bug report either the followings:
467 http://bugme.linux-foundation.org/
469 alsa-devel@alsa-project.org
475 This section describes some tools available for debugging HD-audio
480 The script `alsa-info.sh` is a very useful tool to gather the audio
481 device information. You can fetch the latest version from:
483 - http://www.alsa-project.org/alsa-info.sh
485 Run this script as root, and it will gather the important information
486 such as the module lists, module parameters, proc file contents
487 including the codec proc files, mixer outputs and the control
488 elements. As default, it will store the information onto a web server
489 on alsa-project.org. But, if you send a bug report, it'd be better to
490 run with `--no-upload` option, and attach the generated file.
492 There are some other useful options. See `--help` option output for
495 When a probe error occurs or when the driver obviously assigns a
496 mismatched model, it'd be helpful to load the driver with
497 `probe_only=1` option (at best after the cold reboot) and run
498 alsa-info at this state. With this option, the driver won't configure
499 the mixer and PCM but just tries to probe the codec slot. After
500 probing, the proc file is available, so you can get the raw codec
501 information before modified by the driver. Of course, the driver
502 isn't usable with `probe_only=1`. But you can continue the
503 configuration via hwdep sysfs file if hda-reconfig option is enabled.
508 hda-verb is a tiny program that allows you to access the HD-audio
509 codec directly. You can execute a raw HD-audio codec verb with this.
510 This program accesses the hwdep device, thus you need to enable the
511 kernel config `CONFIG_SND_HDA_HWDEP=y` beforehand.
513 The hda-verb program takes four arguments: the hwdep device file, the
514 widget NID, the verb and the parameter. When you access to the codec
515 on the slot 2 of the card 0, pass /dev/snd/hwC0D2 to the first
516 argument, typically. (However, the real path name depends on the
519 The second parameter is the widget number-id to access. The third
520 parameter can be either a hex/digit number or a string corresponding
521 to a verb. Similarly, the last parameter is the value to write, or
522 can be a string for the parameter type.
524 ------------------------------------------------------------------------
525 % hda-verb /dev/snd/hwC0D0 0x12 0x701 2
526 nid = 0x12, verb = 0x701, param = 0x2
529 % hda-verb /dev/snd/hwC0D0 0x0 PARAMETERS VENDOR_ID
530 nid = 0x0, verb = 0xf00, param = 0x0
533 % hda-verb /dev/snd/hwC0D0 2 set_a 0xb080
534 nid = 0x2, verb = 0x300, param = 0xb080
536 ------------------------------------------------------------------------
538 Although you can issue any verbs with this program, the driver state
539 won't be always updated. For example, the volume values are usually
540 cached in the driver, and thus changing the widget amp value directly
541 via hda-verb won't change the mixer value.
543 The hda-verb program is found in the ftp directory:
545 - ftp://ftp.kernel.org/pub/linux/kernel/people/tiwai/misc/
547 Also a git repository is available:
549 - git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/hda-verb.git
551 See README file in the tarball for more details about hda-verb
557 hda-analyzer provides a graphical interface to access the raw HD-audio
558 control, based on pyGTK2 binding. It's a more powerful version of
559 hda-verb. The program gives you an easy-to-use GUI stuff for showing
560 the widget information and adjusting the amp values, as well as the
561 proc-compatible output.
563 The hda-analyzer is a part of alsa.git repository in
566 - http://git.alsa-project.org/?p=alsa.git;a=tree;f=hda-analyzer
571 Codecgraph is a utility program to generate a graph and visualizes the
572 codec-node connection of a codec chip. It's especially useful when
573 you analyze or debug a codec without a proper datasheet. The program
574 parses the given codec proc file and converts to SVG via graphiz
577 The tarball and GIT trees are found in the web page at:
579 - http://helllabs.org/codecgraph/
584 hda-emu is an HD-audio emulator. The main purpose of this program is
585 to debug an HD-audio codec without the real hardware. Thus, it
586 doesn't emulate the behavior with the real audio I/O, but it just
587 dumps the codec register changes and the ALSA-driver internal changes
588 at probing and operating the HD-audio driver.
590 The program requires a codec proc-file to simulate. Get a proc file
591 for the target codec beforehand, or pick up an example codec from the
592 codec proc collections in the tarball. Then, run the program with the
593 proc file, and the hda-emu program will start parsing the codec file
594 and simulates the HD-audio driver:
596 ------------------------------------------------------------------------
597 % hda-emu codecs/stac9200-dell-d820-laptop
599 hda_codec: Unknown model for STAC9200, using BIOS defaults
600 hda_codec: pin nid 08 bios pin config 40c003fa
602 ------------------------------------------------------------------------
604 The program gives you only a very dumb command-line interface. You
605 can get a proc-file dump at the current state, get a list of control
606 (mixer) elements, set/get the control element value, simulate the PCM
607 operation, the jack plugging simulation, etc.
609 The package is found in:
611 - ftp://ftp.kernel.org/pub/linux/kernel/people/tiwai/misc/
613 A git repository is available:
615 - git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/hda-emu.git
617 See README file in the tarball for more details about hda-emu
projects / linux-2.6 / blob