1 ============================= 2 More Notes on HD-Audio Driver 3 ============================= 4 5 Takashi Iwai <tiwai@suse.de> 6 7 8 General 9 ======= 10 11 HD-audio is the new standard on-board audio component on modern PCs 12 after AC97. Although Linux has been supporting HD-audio since long 13 time ago, there are often problems with new machines. A part of the 14 problem is broken BIOS, and the rest is the driver implementation. 15 This document explains the brief trouble-shooting and debugging 16 methods for the HD-audio hardware. 17 18 The HD-audio component consists of two parts: the controller chip and 19 the codec chips on the HD-audio bus. Linux provides a single driver 20 for all controllers, snd-hda-intel. Although the driver name contains 21 a word of a well-known hardware vendor, it's not specific to it but for 22 all controller chips by other companies. Since the HD-audio 23 controllers are supposed to be compatible, the single snd-hda-driver 24 should work in most cases. But, not surprisingly, there are known 25 bugs and issues specific to each controller type. The snd-hda-intel 26 driver has a bunch of workarounds for these as described below. 27 28 A controller may have multiple codecs. Usually you have one audio 29 codec and optionally one modem codec. In theory, there might be 30 multiple audio codecs, e.g. for analog and digital outputs, and the 31 driver might not work properly because of conflict of mixer elements. 32 This should be fixed in future if such hardware really exists. 33 34 The snd-hda-intel driver has several different codec parsers depending 35 on the codec. It has a generic parser as a fallback, but this 36 functionality is fairly limited until now. Instead of the generic 37 parser, usually the codec-specific parser (coded in patch_*.c) is used 38 for the codec-specific implementations. The details about the 39 codec-specific problems are explained in the later sections. 40 41 If you are interested in the deep debugging of HD-audio, read the 42 HD-audio specification at first. The specification is found on 43 Intel's web page, for example: 44 45 * https://www.intel.com/standards/hdaudio/ 46 47 48 HD-Audio Controller 49 =================== 50 51 DMA-Position Problem 52 -------------------- 53 The most common problem of the controller is the inaccurate DMA 54 pointer reporting. The DMA pointer for playback and capture can be 55 read in two ways, either via a LPIB register or via a position-buffer 56 map. As default the driver tries to read from the io-mapped 57 position-buffer, and falls back to LPIB if the position-buffer appears 58 dead. However, this detection isn't perfect on some devices. In such 59 a case, you can change the default method via ``position_fix`` option. 60 61 ``position_fix=1`` means to use LPIB method explicitly. 62 ``position_fix=2`` means to use the position-buffer. 63 ``position_fix=3`` means to use a combination of both methods, needed 64 for some VIA controllers. The capture stream position is corrected 65 by comparing both LPIB and position-buffer values. 66 ``position_fix=4`` is another combination available for all controllers, 67 and uses LPIB for the playback and the position-buffer for the capture 68 streams. 69 ``position_fix=5`` is specific to Intel platforms, so far, for Skylake 70 and onward. It applies the delay calculation for the precise position 71 reporting. 72 ``position_fix=6`` is to correct the position with the fixed FIFO 73 size, mainly targeted for the recent AMD controllers. 74 0 is the default value for all other 75 controllers, the automatic check and fallback to LPIB as described in 76 the above. If you get a problem of repeated sounds, this option might 77 help. 78 79 In addition to that, every controller is known to be broken regarding 80 the wake-up timing. It wakes up a few samples before actually 81 processing the data on the buffer. This caused a lot of problems, for 82 example, with ALSA dmix or JACK. Since 2.6.27 kernel, the driver puts 83 an artificial delay to the wake up timing. This delay is controlled 84 via ``bdl_pos_adj`` option. 85 86 When ``bdl_pos_adj`` is a negative value (as default), it's assigned to 87 an appropriate value depending on the controller chip. For Intel 88 chips, it'd be 1 while it'd be 32 for others. Usually this works. 89 Only in case it doesn't work and you get warning messages, you should 90 change this parameter to other values. 91 92 93 Codec-Probing Problem 94 --------------------- 95 A less often but a more severe problem is the codec probing. When 96 BIOS reports the available codec slots wrongly, the driver gets 97 confused and tries to access the non-existing codec slot. This often 98 results in the total screw-up, and destructs the further communication 99 with the codec chips. The symptom appears usually as error messages 100 like: 101 :: 102 103 hda_intel: azx_get_response timeout, switching to polling mode: 104 last cmd=0x12345678 105 hda_intel: azx_get_response timeout, switching to single_cmd mode: 106 last cmd=0x12345678 107 108 The first line is a warning, and this is usually relatively harmless. 109 It means that the codec response isn't notified via an IRQ. The 110 driver uses explicit polling method to read the response. It gives 111 very slight CPU overhead, but you'd unlikely notice it. 112 113 The second line is, however, a fatal error. If this happens, usually 114 it means that something is really wrong. Most likely you are 115 accessing a non-existing codec slot. 116 117 Thus, if the second error message appears, try to narrow the probed 118 codec slots via ``probe_mask`` option. It's a bitmask, and each bit 119 corresponds to the codec slot. For example, to probe only the first 120 slot, pass ``probe_mask=1``. For the first and the third slots, pass 121 ``probe_mask=5`` (where 5 = 1 | 4), and so on. 122 123 Since 2.6.29 kernel, the driver has a more robust probing method, so 124 this error might happen rarely, though. 125 126 On a machine with a broken BIOS, sometimes you need to force the 127 driver to probe the codec slots the hardware doesn't report for use. 128 In such a case, turn the bit 8 (0x100) of ``probe_mask`` option on. 129 Then the rest 8 bits are passed as the codec slots to probe 130 unconditionally. For example, ``probe_mask=0x103`` will force to probe 131 the codec slots 0 and 1 no matter what the hardware reports. 132 133 134 Interrupt Handling 135 ------------------ 136 HD-audio driver uses MSI as default (if available) since 2.6.33 137 kernel as MSI works better on some machines, and in general, it's 138 better for performance. However, Nvidia controllers showed bad 139 regressions with MSI (especially in a combination with AMD chipset), 140 thus we disabled MSI for them. 141 142 There seem also still other devices that don't work with MSI. If you 143 see a regression wrt the sound quality (stuttering, etc) or a lock-up 144 in the recent kernel, try to pass ``enable_msi=0`` option to disable 145 MSI. If it works, you can add the known bad device to the blacklist 146 defined in hda_intel.c. In such a case, please report and give the 147 patch back to the upstream developer. 148 149 150 HD-Audio Codec 151 ============== 152 153 Model Option 154 ------------ 155 The most common problem regarding the HD-audio driver is the 156 unsupported codec features or the mismatched device configuration. 157 Most of codec-specific code has several preset models, either to 158 override the BIOS setup or to provide more comprehensive features. 159 160 The driver checks PCI SSID and looks through the static configuration 161 table until any matching entry is found. If you have a new machine, 162 you may see a message like below: 163 :: 164 165 hda_codec: ALC880: BIOS auto-probing. 166 167 Meanwhile, in the earlier versions, you would see a message like: 168 :: 169 170 hda_codec: Unknown model for ALC880, trying auto-probe from BIOS... 171 172 Even if you see such a message, DON'T PANIC. Take a deep breath and 173 keep your towel. First of all, it's an informational message, no 174 warning, no error. This means that the PCI SSID of your device isn't 175 listed in the known preset model (white-)list. But, this doesn't mean 176 that the driver is broken. Many codec-drivers provide the automatic 177 configuration mechanism based on the BIOS setup. 178 179 The HD-audio codec has usually "pin" widgets, and BIOS sets the default 180 configuration of each pin, which indicates the location, the 181 connection type, the jack color, etc. The HD-audio driver can guess 182 the right connection judging from these default configuration values. 183 However -- some codec-support codes, such as patch_analog.c, don't 184 support the automatic probing (yet as of 2.6.28). And, BIOS is often, 185 yes, pretty often broken. It sets up wrong values and screws up the 186 driver. 187 188 The preset model (or recently called as "fix-up") is provided 189 basically to overcome such a situation. When the matching preset 190 model is found in the white-list, the driver assumes the static 191 configuration of that preset with the correct pin setup, etc. 192 Thus, if you have a newer machine with a slightly different PCI SSID 193 (or codec SSID) from the existing one, you may have a good chance to 194 re-use the same model. You can pass the ``model`` option to specify the 195 preset model instead of PCI (and codec-) SSID look-up. 196 197 What ``model`` option values are available depends on the codec chip. 198 Check your codec chip from the codec proc file (see "Codec Proc-File" 199 section below). It will show the vendor/product name of your codec 200 chip. Then, see Documentation/sound/hd-audio/models.rst file, 201 the section of HD-audio driver. You can find a list of codecs 202 and ``model`` options belonging to each codec. For example, for Realtek 203 ALC262 codec chip, pass ``model=ultra`` for devices that are compatible 204 with Samsung Q1 Ultra. 205 206 Thus, the first thing you can do for any brand-new, unsupported and 207 non-working HD-audio hardware is to check HD-audio codec and several 208 different ``model`` option values. If you have any luck, some of them 209 might suit with your device well. 210 211 There are a few special model option values: 212 213 * when 'nofixup' is passed, the device-specific fixups in the codec 214 parser are skipped. 215 * when ``generic`` is passed, the codec-specific parser is skipped and 216 only the generic parser is used. 217 218 A new style for the model option that was introduced since 5.15 kernel 219 is to pass the PCI or codec SSID in the form of ``model=XXXX:YYYY`` 220 where XXXX and YYYY are the sub-vendor and sub-device IDs in hex 221 numbers, respectively. This is a kind of aliasing to another device; 222 when this form is given, the driver will refer to that SSID as a 223 reference to the quirk table. It'd be useful especially when the 224 target quirk isn't listed in the model table. For example, passing 225 model=103c:8862 will apply the quirk for HP ProBook 445 G8 (which 226 isn't found in the model table as of writing) as long as the device is 227 handled equivalently by the same driver. 228 229 230 Speaker and Headphone Output 231 ---------------------------- 232 One of the most frequent (and obvious) bugs with HD-audio is the 233 silent output from either or both of a built-in speaker and a 234 headphone jack. In general, you should try a headphone output at 235 first. A speaker output often requires more additional controls like 236 the external amplifier bits. Thus a headphone output has a slightly 237 better chance. 238 239 Before making a bug report, double-check whether the mixer is set up 240 correctly. The recent version of snd-hda-intel driver provides mostly 241 "Master" volume control as well as "Front" volume (where Front 242 indicates the front-channels). In addition, there can be individual 243 "Headphone" and "Speaker" controls. 244 245 Ditto for the speaker output. There can be "External Amplifier" 246 switch on some codecs. Turn on this if present. 247 248 Another related problem is the automatic mute of speaker output by 249 headphone plugging. This feature is implemented in most cases, but 250 not on every preset model or codec-support code. 251 252 In anyway, try a different model option if you have such a problem. 253 Some other models may match better and give you more matching 254 functionality. If none of the available models works, send a bug 255 report. See the bug report section for details. 256 257 If you are masochistic enough to debug the driver problem, note the 258 following: 259 260 * The speaker (and the headphone, too) output often requires the 261 external amplifier. This can be set usually via EAPD verb or a 262 certain GPIO. If the codec pin supports EAPD, you have a better 263 chance via SET_EAPD_BTL verb (0x70c). On others, GPIO pin (mostly 264 it's either GPIO0 or GPIO1) may turn on/off EAPD. 265 * Some Realtek codecs require special vendor-specific coefficients to 266 turn on the amplifier. See patch_realtek.c. 267 * IDT codecs may have extra power-enable/disable controls on each 268 analog pin. See patch_sigmatel.c. 269 * Very rare but some devices don't accept the pin-detection verb until 270 triggered. Issuing GET_PIN_SENSE verb (0xf09) may result in the 271 codec-communication stall. Some examples are found in 272 patch_realtek.c. 273 274 275 Capture Problems 276 ---------------- 277 The capture problems are often because of missing setups of mixers. 278 Thus, before submitting a bug report, make sure that you set up the 279 mixer correctly. For example, both "Capture Volume" and "Capture 280 Switch" have to be set properly in addition to the right "Capture 281 Source" or "Input Source" selection. Some devices have "Mic Boost" 282 volume or switch. 283 284 When the PCM device is opened via "default" PCM (without pulse-audio 285 plugin), you'll likely have "Digital Capture Volume" control as well. 286 This is provided for the extra gain/attenuation of the signal in 287 software, especially for the inputs without the hardware volume 288 control such as digital microphones. Unless really needed, this 289 should be set to exactly 50%, corresponding to 0dB -- neither extra 290 gain nor attenuation. When you use "hw" PCM, i.e., a raw access PCM, 291 this control will have no influence, though. 292 293 It's known that some codecs / devices have fairly bad analog circuits, 294 and the recorded sound contains a certain DC-offset. This is no bug 295 of the driver. 296 297 Most of modern laptops have no analog CD-input connection. Thus, the 298 recording from CD input won't work in many cases although the driver 299 provides it as the capture source. Use CDDA instead. 300 301 The automatic switching of the built-in and external mic per plugging 302 is implemented on some codec models but not on every model. Partly 303 because of my laziness but mostly lack of testers. Feel free to 304 submit the improvement patch to the author. 305 306 307 Direct Debugging 308 ---------------- 309 If no model option gives you a better result, and you are a tough guy 310 to fight against evil, try debugging via hitting the raw HD-audio 311 codec verbs to the device. Some tools are available: hda-emu and 312 hda-analyzer. The detailed description is found in the sections 313 below. You'd need to enable hwdep for using these tools. See "Kernel 314 Configuration" section. 315 316 317 Other Issues 318 ============ 319 320 Kernel Configuration 321 -------------------- 322 In general, I recommend you to enable the sound debug option, 323 ``CONFIG_SND_DEBUG=y``, no matter whether you are debugging or not. 324 325 Don't forget to turn on the appropriate ``CONFIG_SND_HDA_CODEC_*`` 326 options. Note that each of them corresponds to the codec chip, not 327 the controller chip. Thus, even if lspci shows the Nvidia controller, 328 you may need to choose the option for other vendors. If you are 329 unsure, just select all yes. 330 331 ``CONFIG_SND_HDA_HWDEP`` is a useful option for debugging the driver. 332 When this is enabled, the driver creates hardware-dependent devices 333 (one per each codec), and you have a raw access to the device via 334 these device files. For example, ``hwC0D2`` will be created for the 335 codec slot #2 of the first card (#0). For debug-tools such as 336 hda-verb and hda-analyzer, the hwdep device has to be enabled. 337 Thus, it'd be better to turn this on always. 338 339 ``CONFIG_SND_HDA_RECONFIG`` is a new option, and this depends on the 340 hwdep option above. When enabled, you'll have some sysfs files under 341 the corresponding hwdep directory. See "HD-audio reconfiguration" 342 section below. 343 344 ``CONFIG_SND_HDA_POWER_SAVE`` option enables the power-saving feature. 345 See "Power-saving" section below. 346 347 348 Codec Proc-File 349 --------------- 350 The codec proc-file is a treasure-chest for debugging HD-audio. 351 It shows most of useful information of each codec widget. 352 353 The proc file is located in /proc/asound/card*/codec#*, one file per 354 each codec slot. You can know the codec vendor, product id and 355 names, the type of each widget, capabilities and so on. 356 This file, however, doesn't show the jack sensing state, so far. This 357 is because the jack-sensing might be depending on the trigger state. 358 359 This file will be picked up by the debug tools, and also it can be fed 360 to the emulator as the primary codec information. See the debug tools 361 section below. 362 363 This proc file can be also used to check whether the generic parser is 364 used. When the generic parser is used, the vendor/product ID name 365 will appear as "Realtek ID 0262", instead of "Realtek ALC262". 366 367 368 HD-Audio Reconfiguration 369 ------------------------ 370 This is an experimental feature to allow you re-configure the HD-audio 371 codec dynamically without reloading the driver. The following sysfs 372 files are available under each codec-hwdep device directory (e.g. 373 /sys/class/sound/hwC0D0): 374 375 vendor_id 376 Shows the 32bit codec vendor-id hex number. You can change the 377 vendor-id value by writing to this file. 378 subsystem_id 379 Shows the 32bit codec subsystem-id hex number. You can change the 380 subsystem-id value by writing to this file. 381 revision_id 382 Shows the 32bit codec revision-id hex number. You can change the 383 revision-id value by writing to this file. 384 afg 385 Shows the AFG ID. This is read-only. 386 mfg 387 Shows the MFG ID. This is read-only. 388 name 389 Shows the codec name string. Can be changed by writing to this 390 file. 391 modelname 392 Shows the currently set ``model`` option. Can be changed by writing 393 to this file. 394 init_verbs 395 The extra verbs to execute at initialization. You can add a verb by 396 writing to this file. Pass three numbers: nid, verb and parameter 397 (separated with a space). 398 hints 399 Shows / stores hint strings for codec parsers for any use. 400 Its format is ``key = value``. For example, passing ``jack_detect = no`` 401 will disable the jack detection of the machine completely. 402 init_pin_configs 403 Shows the initial pin default config values set by BIOS. 404 driver_pin_configs 405 Shows the pin default values set by the codec parser explicitly. 406 This doesn't show all pin values but only the changed values by 407 the parser. That is, if the parser doesn't change the pin default 408 config values by itself, this will contain nothing. 409 user_pin_configs 410 Shows the pin default config values to override the BIOS setup. 411 Writing this (with two numbers, NID and value) appends the new 412 value. The given will be used instead of the initial BIOS value at 413 the next reconfiguration time. Note that this config will override 414 even the driver pin configs, too. 415 reconfig 416 Triggers the codec re-configuration. When any value is written to 417 this file, the driver re-initialize and parses the codec tree 418 again. All the changes done by the sysfs entries above are taken 419 into account. 420 clear 421 Resets the codec, removes the mixer elements and PCM stuff of the 422 specified codec, and clear all init verbs and hints. 423 424 For example, when you want to change the pin default configuration 425 value of the pin widget 0x14 to 0x9993013f, and let the driver 426 re-configure based on that state, run like below: 427 :: 428 429 # echo 0x14 0x9993013f > /sys/class/sound/hwC0D0/user_pin_configs 430 # echo 1 > /sys/class/sound/hwC0D0/reconfig 431 432 433 Hint Strings 434 ------------ 435 The codec parser have several switches and adjustment knobs for 436 matching better with the actual codec or device behavior. Many of 437 them can be adjusted dynamically via "hints" strings as mentioned in 438 the section above. For example, by passing ``jack_detect = no`` string 439 via sysfs or a patch file, you can disable the jack detection, thus 440 the codec parser will skip the features like auto-mute or mic 441 auto-switch. As a boolean value, either ``yes``, ``no``, ``true``, ``false``, 442 ``1`` or ``0`` can be passed. 443 444 The generic parser supports the following hints: 445 446 jack_detect (bool) 447 specify whether the jack detection is available at all on this 448 machine; default true 449 inv_jack_detect (bool) 450 indicates that the jack detection logic is inverted 451 trigger_sense (bool) 452 indicates that the jack detection needs the explicit call of 453 AC_VERB_SET_PIN_SENSE verb 454 inv_eapd (bool) 455 indicates that the EAPD is implemented in the inverted logic 456 pcm_format_first (bool) 457 sets the PCM format before the stream tag and channel ID 458 sticky_stream (bool) 459 keep the PCM format, stream tag and ID as long as possible; 460 default true 461 spdif_status_reset (bool) 462 reset the SPDIF status bits at each time the SPDIF stream is set 463 up 464 pin_amp_workaround (bool) 465 the output pin may have multiple amp values 466 single_adc_amp (bool) 467 ADCs can have only single input amps 468 auto_mute (bool) 469 enable/disable the headphone auto-mute feature; default true 470 auto_mic (bool) 471 enable/disable the mic auto-switch feature; default true 472 line_in_auto_switch (bool) 473 enable/disable the line-in auto-switch feature; default false 474 need_dac_fix (bool) 475 limits the DACs depending on the channel count 476 primary_hp (bool) 477 probe headphone jacks as the primary outputs; default true 478 multi_io (bool) 479 try probing multi-I/O config (e.g. shared line-in/surround, 480 mic/clfe jacks) 481 multi_cap_vol (bool) 482 provide multiple capture volumes 483 inv_dmic_split (bool) 484 provide split internal mic volume/switch for phase-inverted 485 digital mics 486 indep_hp (bool) 487 provide the independent headphone PCM stream and the corresponding 488 mixer control, if available 489 add_stereo_mix_input (bool) 490 add the stereo mix (analog-loopback mix) to the input mux if 491 available 492 add_jack_modes (bool) 493 add "xxx Jack Mode" enum controls to each I/O jack for allowing to 494 change the headphone amp and mic bias VREF capabilities 495 power_save_node (bool) 496 advanced power management for each widget, controlling the power 497 state (D0/D3) of each widget node depending on the actual pin and 498 stream states 499 power_down_unused (bool) 500 power down the unused widgets, a subset of power_save_node, and 501 will be dropped in future 502 add_hp_mic (bool) 503 add the headphone to capture source if possible 504 hp_mic_detect (bool) 505 enable/disable the hp/mic shared input for a single built-in mic 506 case; default true 507 vmaster (bool) 508 enable/disable the virtual Master control; default true 509 mixer_nid (int) 510 specifies the widget NID of the analog-loopback mixer 511 512 513 Early Patching 514 -------------- 515 When ``CONFIG_SND_HDA_PATCH_LOADER=y`` is set, you can pass a "patch" 516 as a firmware file for modifying the HD-audio setup before 517 initializing the codec. This can work basically like the 518 reconfiguration via sysfs in the above, but it does it before the 519 first codec configuration. 520 521 A patch file is a plain text file which looks like below: 522 523 :: 524 525 [codec] 526 0x12345678 0xabcd1234 2 527 528 [model] 529 auto 530 531 [pincfg] 532 0x12 0x411111f0 533 534 [verb] 535 0x20 0x500 0x03 536 0x20 0x400 0xff 537 538 [hint] 539 jack_detect = no 540 541 542 The file needs to have a line ``[codec]``. The next line should contain 543 three numbers indicating the codec vendor-id (0x12345678 in the 544 example), the codec subsystem-id (0xabcd1234) and the address (2) of 545 the codec. The rest patch entries are applied to this specified codec 546 until another codec entry is given. Passing 0 or a negative number to 547 the first or the second value will make the check of the corresponding 548 field be skipped. It'll be useful for really broken devices that don't 549 initialize SSID properly. 550 551 The ``[model]`` line allows to change the model name of the each codec. 552 In the example above, it will be changed to model=auto. 553 Note that this overrides the module option. 554 555 After the ``[pincfg]`` line, the contents are parsed as the initial 556 default pin-configurations just like ``user_pin_configs`` sysfs above. 557 The values can be shown in user_pin_configs sysfs file, too. 558 559 Similarly, the lines after ``[verb]`` are parsed as ``init_verbs`` 560 sysfs entries, and the lines after ``[hint]`` are parsed as ``hints`` 561 sysfs entries, respectively. 562 563 Another example to override the codec vendor id from 0x12345678 to 564 0xdeadbeef is like below: 565 :: 566 567 [codec] 568 0x12345678 0xabcd1234 2 569 570 [vendor_id] 571 0xdeadbeef 572 573 574 In the similar way, you can override the codec subsystem_id via 575 ``[subsystem_id]``, the revision id via ``[revision_id]`` line. 576 Also, the codec chip name can be rewritten via ``[chip_name]`` line. 577 :: 578 579 [codec] 580 0x12345678 0xabcd1234 2 581 582 [subsystem_id] 583 0xffff1111 584 585 [revision_id] 586 0x10 587 588 [chip_name] 589 My-own NEWS-0002 590 591 592 The hd-audio driver reads the file via request_firmware(). Thus, 593 a patch file has to be located on the appropriate firmware path, 594 typically, /lib/firmware. For example, when you pass the option 595 ``patch=hda-init.fw``, the file /lib/firmware/hda-init.fw must be 596 present. 597 598 The patch module option is specific to each card instance, and you 599 need to give one file name for each instance, separated by commas. 600 For example, if you have two cards, one for an on-board analog and one 601 for an HDMI video board, you may pass patch option like below: 602 :: 603 604 options snd-hda-intel patch=on-board-patch,hdmi-patch 605 606 607 Power-Saving 608 ------------ 609 The power-saving is a kind of auto-suspend of the device. When the 610 device is inactive for a certain time, the device is automatically 611 turned off to save the power. The time to go down is specified via 612 ``power_save`` module option, and this option can be changed dynamically 613 via sysfs. 614 615 The power-saving won't work when the analog loopback is enabled on 616 some codecs. Make sure that you mute all unneeded signal routes when 617 you want the power-saving. 618 619 The power-saving feature might cause audible click noises at each 620 power-down/up depending on the device. Some of them might be 621 solvable, but some are hard, I'm afraid. Some distros such as 622 openSUSE enables the power-saving feature automatically when the power 623 cable is unplugged. Thus, if you hear noises, suspect first the 624 power-saving. See /sys/module/snd_hda_intel/parameters/power_save to 625 check the current value. If it's non-zero, the feature is turned on. 626 627 The recent kernel supports the runtime PM for the HD-audio controller 628 chip, too. It means that the HD-audio controller is also powered up / 629 down dynamically. The feature is enabled only for certain controller 630 chips like Intel LynxPoint. You can enable/disable this feature 631 forcibly by setting ``power_save_controller`` option, which is also 632 available at /sys/module/snd_hda_intel/parameters directory. 633 634 635 Tracepoints 636 ----------- 637 The hd-audio driver gives a few basic tracepoints. 638 ``hda:hda_send_cmd`` traces each CORB write while ``hda:hda_get_response`` 639 traces the response from RIRB (only when read from the codec driver). 640 ``hda:hda_bus_reset`` traces the bus-reset due to fatal error, etc, 641 ``hda:hda_unsol_event`` traces the unsolicited events, and 642 ``hda:hda_power_down`` and ``hda:hda_power_up`` trace the power down/up 643 via power-saving behavior. 644 645 Enabling all tracepoints can be done like 646 :: 647 648 # echo 1 > /sys/kernel/tracing/events/hda/enable 649 650 then after some commands, you can traces from 651 /sys/kernel/tracing/trace file. For example, when you want to 652 trace what codec command is sent, enable the tracepoint like: 653 :: 654 655 # cat /sys/kernel/tracing/trace 656 # tracer: nop 657 # 658 # TASK-PID CPU# TIMESTAMP FUNCTION 659 # | | | | | 660 <...>-7807 [002] 105147.774889: hda_send_cmd: [0:0] val=e3a019 661 <...>-7807 [002] 105147.774893: hda_send_cmd: [0:0] val=e39019 662 <...>-7807 [002] 105147.999542: hda_send_cmd: [0:0] val=e3a01a 663 <...>-7807 [002] 105147.999543: hda_send_cmd: [0:0] val=e3901a 664 <...>-26764 [001] 349222.837143: hda_send_cmd: [0:0] val=e3a019 665 <...>-26764 [001] 349222.837148: hda_send_cmd: [0:0] val=e39019 666 <...>-26764 [001] 349223.058539: hda_send_cmd: [0:0] val=e3a01a 667 <...>-26764 [001] 349223.058541: hda_send_cmd: [0:0] val=e3901a 668 669 Here ``[0:0]`` indicates the card number and the codec address, and 670 ``val`` shows the value sent to the codec, respectively. The value is 671 a packed value, and you can decode it via hda-decode-verb program 672 included in hda-emu package below. For example, the value e3a019 is 673 to set the left output-amp value to 25. 674 :: 675 676 % hda-decode-verb 0xe3a019 677 raw value = 0x00e3a019 678 cid = 0, nid = 0x0e, verb = 0x3a0, parm = 0x19 679 raw value: verb = 0x3a0, parm = 0x19 680 verbname = set_amp_gain_mute 681 amp raw val = 0xa019 682 output, left, idx=0, mute=0, val=25 683 684 685 Development Tree 686 ---------------- 687 The latest development codes for HD-audio are found on sound git tree: 688 689 * git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound.git 690 691 The master branch or for-next branches can be used as the main 692 development branches in general while the development for the current 693 and next kernels are found in for-linus and for-next branches, 694 respectively. 695 696 697 Sending a Bug Report 698 -------------------- 699 If any model or module options don't work for your device, it's time 700 to send a bug report to the developers. Give the following in your 701 bug report: 702 703 * Hardware vendor, product and model names 704 * Kernel version (and ALSA-driver version if you built externally) 705 * ``alsa-info.sh`` output; run with ``--no-upload`` option. See the 706 section below about alsa-info 707 708 If it's a regression, at best, send alsa-info outputs of both working 709 and non-working kernels. This is really helpful because we can 710 compare the codec registers directly. 711 712 Send a bug report either the following: 713 714 kernel-bugzilla 715 https://bugzilla.kernel.org/ 716 alsa-devel ML 717 alsa-devel@alsa-project.org 718 719 720 Debug Tools 721 =========== 722 723 This section describes some tools available for debugging HD-audio 724 problems. 725 726 alsa-info 727 --------- 728 The script ``alsa-info.sh`` is a very useful tool to gather the audio 729 device information. It's included in alsa-utils package. The latest 730 version can be found on git repository: 731 732 * git://git.alsa-project.org/alsa-utils.git 733 734 The script can be fetched directly from the following URL, too: 735 736 * https://www.alsa-project.org/alsa-info.sh 737 738 Run this script as root, and it will gather the important information 739 such as the module lists, module parameters, proc file contents 740 including the codec proc files, mixer outputs and the control 741 elements. As default, it will store the information onto a web server 742 on alsa-project.org. But, if you send a bug report, it'd be better to 743 run with ``--no-upload`` option, and attach the generated file. 744 745 There are some other useful options. See ``--help`` option output for 746 details. 747 748 When a probe error occurs or when the driver obviously assigns a 749 mismatched model, it'd be helpful to load the driver with 750 ``probe_only=1`` option (at best after the cold reboot) and run 751 alsa-info at this state. With this option, the driver won't configure 752 the mixer and PCM but just tries to probe the codec slot. After 753 probing, the proc file is available, so you can get the raw codec 754 information before modified by the driver. Of course, the driver 755 isn't usable with ``probe_only=1``. But you can continue the 756 configuration via hwdep sysfs file if hda-reconfig option is enabled. 757 Using ``probe_only`` mask 2 skips the reset of HDA codecs (use 758 ``probe_only=3`` as module option). The hwdep interface can be used 759 to determine the BIOS codec initialization. 760 761 762 hda-verb 763 -------- 764 hda-verb is a tiny program that allows you to access the HD-audio 765 codec directly. You can execute a raw HD-audio codec verb with this. 766 This program accesses the hwdep device, thus you need to enable the 767 kernel config ``CONFIG_SND_HDA_HWDEP=y`` beforehand. 768 769 The hda-verb program takes four arguments: the hwdep device file, the 770 widget NID, the verb and the parameter. When you access to the codec 771 on the slot 2 of the card 0, pass /dev/snd/hwC0D2 to the first 772 argument, typically. (However, the real path name depends on the 773 system.) 774 775 The second parameter is the widget number-id to access. The third 776 parameter can be either a hex/digit number or a string corresponding 777 to a verb. Similarly, the last parameter is the value to write, or 778 can be a string for the parameter type. 779 780 :: 781 782 % hda-verb /dev/snd/hwC0D0 0x12 0x701 2 783 nid = 0x12, verb = 0x701, param = 0x2 784 value = 0x0 785 786 % hda-verb /dev/snd/hwC0D0 0x0 PARAMETERS VENDOR_ID 787 nid = 0x0, verb = 0xf00, param = 0x0 788 value = 0x10ec0262 789 790 % hda-verb /dev/snd/hwC0D0 2 set_a 0xb080 791 nid = 0x2, verb = 0x300, param = 0xb080 792 value = 0x0 793 794 795 Although you can issue any verbs with this program, the driver state 796 won't be always updated. For example, the volume values are usually 797 cached in the driver, and thus changing the widget amp value directly 798 via hda-verb won't change the mixer value. 799 800 The hda-verb program is included now in alsa-tools: 801 802 * git://git.alsa-project.org/alsa-tools.git 803 804 Also, the old stand-alone package is found in the ftp directory: 805 806 * ftp://ftp.suse.com/pub/people/tiwai/misc/ 807 808 Also a git repository is available: 809 810 * git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/hda-verb.git 811 812 See README file in the tarball for more details about hda-verb 813 program. 814 815 816 hda-analyzer 817 ------------ 818 hda-analyzer provides a graphical interface to access the raw HD-audio 819 control, based on pyGTK2 binding. It's a more powerful version of 820 hda-verb. The program gives you an easy-to-use GUI stuff for showing 821 the widget information and adjusting the amp values, as well as the 822 proc-compatible output. 823 824 The hda-analyzer: 825 826 * https://git.alsa-project.org/?p=alsa.git;a=tree;f=hda-analyzer 827 828 is a part of alsa.git repository in alsa-project.org: 829 830 * git://git.alsa-project.org/alsa.git 831 832 Codecgraph 833 ---------- 834 Codecgraph is a utility program to generate a graph and visualizes the 835 codec-node connection of a codec chip. It's especially useful when 836 you analyze or debug a codec without a proper datasheet. The program 837 parses the given codec proc file and converts to SVG via graphiz 838 program. 839 840 The tarball and GIT trees are found in the web page at: 841 842 * http://helllabs.org/codecgraph/ 843 844 845 hda-emu 846 ------- 847 hda-emu is an HD-audio emulator. The main purpose of this program is 848 to debug an HD-audio codec without the real hardware. Thus, it 849 doesn't emulate the behavior with the real audio I/O, but it just 850 dumps the codec register changes and the ALSA-driver internal changes 851 at probing and operating the HD-audio driver. 852 853 The program requires a codec proc-file to simulate. Get a proc file 854 for the target codec beforehand, or pick up an example codec from the 855 codec proc collections in the tarball. Then, run the program with the 856 proc file, and the hda-emu program will start parsing the codec file 857 and simulates the HD-audio driver: 858 859 :: 860 861 % hda-emu codecs/stac9200-dell-d820-laptop 862 # Parsing.. 863 hda_codec: Unknown model for STAC9200, using BIOS defaults 864 hda_codec: pin nid 08 bios pin config 40c003fa 865 .... 866 867 868 The program gives you only a very dumb command-line interface. You 869 can get a proc-file dump at the current state, get a list of control 870 (mixer) elements, set/get the control element value, simulate the PCM 871 operation, the jack plugging simulation, etc. 872 873 The program is found in the git repository below: 874 875 * git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/hda-emu.git 876 877 See README file in the repository for more details about hda-emu 878 program. 879 880 881 hda-jack-retask 882 --------------- 883 hda-jack-retask is a user-friendly GUI program to manipulate the 884 HD-audio pin control for jack retasking. If you have a problem about 885 the jack assignment, try this program and check whether you can get 886 useful results. Once when you figure out the proper pin assignment, 887 it can be fixed either in the driver code statically or via passing a 888 firmware patch file (see "Early Patching" section). 889 890 The program is included in alsa-tools now: 891 892 * git://git.alsa-project.org/alsa-tools.git
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.