- Alsa-devel - mailman.alsa-project.org

[alsa-devel] alsa + resampler + jack: possible?
by Markus Grabner 19 Oct '08

19 Oct '08

Hi all! A while ago I started development of an ALSA driver for a device which natively only supports a sampling rate of 39062.5 Hz and 24 bit samples (see http://www.tanzband-scream.at/line6). The ALSA rate plugin does a perfect job for using this device with standard applications (such as xmms) to play audio files at "conventional" sample rates (e.g., 44.1 kHz) with a configuration like this: pcm.resampler { type rate slave { pcm hw:1 format S24_3LE rate 39062 } converter "samplerate" } However, this doesn't work for jack since jack requires a power-of-two period size, and ALSA can't find an integer period size for the raw device such that the target rate is met with sufficient precision. An example: the jack daemon can't be started with period size 512 and sample rate 48000 Hz, but it works for 47961 Hz since 512 * 39062.5 / 48000 = 416.666... ~= 417 bytes 512 / 417 * 39062.5 = 47961.63 ~= 47961 Hz When using this approach together with other audio data from a 48 kHz source, they would drift apart over time, which basically renders the alsa+jack setup useless for this device. I understand that resampling is easier to implement if the sample rate ratio equals the ratio of the (integer) period sizes of client and slave, but at least the underlying libsamplerate has no such restriction. I tried to find the relevant pieces in the ALSA source code (pcm_rate.c), but I'm not sure where these calculations are done and how much effort it would be to support arbitrary sample rate ratios for the resampler. Can this be done in a plugin, or does it require modifications in the ALSA library code base? Could the developers please give a few hints on that? Thanks & kind regards, Markus -- Markus Grabner - Computer Graphics and Vision Graz University of Technology, Inffeldgasse 16a/II, 8010 Graz, Austria Phone: +43/316/873-5041, Fax: +43/316/873-5050 WWW: http://www.icg.tugraz.at/Members/grabner

1 0

[alsa-devel] snd_pcm_sw_params_get_xfer_align deprecated?
by Erik de Castro Lopo 18 Oct '08

18 Oct '08

Hi all, I've got some code that gives me a waring that the above function is deprecated. What should I do: a) Call some other function? If so which one? b) Just drop the call to that function and everything will continue to work as it should? c) Something else? Cheers, Erik -- ----------------------------------------------------------------- Erik de Castro Lopo ----------------------------------------------------------------- "The very definition of hell is having to maintain someone else's Perl code." -- Anonymous

3 2

[alsa-devel] New BIOS, new vendor and device IDs
by Chris Pemberton 17 Oct '08

17 Oct '08

My machine is a Lenovo 3000 N100 laptop (0768 model). I had to recently update the BIOS. Old BIOS: lspci -> device id: [17aa:2066] which is a Realtek ALC861VD (found in patch_realtek.c) New BIOS: lspci -> device id: [8086:27d8] (found in hda_intel.c) In both cases, I use the module option "model=3stack" to get sound working: Laptop speakers work. Built-in mic works. External mic works Headphones work; but automatic speaker muting does not. I tried to debug why I needed this module option at all, several months ago (old BIOS), but I didn't get far and just became complacent with using the module option. If anyone is interested, I'll do what I can to help debug. It would be nice to be able to use Skype with headphones... Output of the alsa info script is at: http://www.alsa-project.org/db/?f=701ca6efa0500080dc78a555813f66c1d42096e6 That is with driver 1.0.17. I'll try to update to the latest snapshot and run the script again. Thanks, Chris

2 6

Re: [alsa-devel] Backported sbxfi driver (UNTESTED!)
by Takashi Iwai 17 Oct '08

17 Oct '08

At Thu, 16 Oct 2008 19:06:58 +0400, The Source wrote: > > Takashi Iwai пишет: > > At Thu, 16 Oct 2008 19:00:16 +0400, > > The Source wrote: > > > >> Takashi Iwai пишет: > >> > >>> At Thu, 16 Oct 2008 18:41:08 +0400, > >>> The Source wrote: > >>> > >>> > >>>> Takashi Iwai пишет: > >>>> > >>>> > >>>>> At Thu, 16 Oct 2008 18:15:45 +0400, > >>>>> The Source wrote: > >>>>> > >>>>> > >>>>> > >>>>>> Takashi Iwai пишет: > >>>>>> > >>>>>> > >>>>>> > >>>>>>> At Thu, 16 Oct 2008 17:36:04 +0400, > >>>>>>> The Source wrote: > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>>>>>> And, which X-Fi model do you have? > >>>>>>>>>> Please show the lspci -nv output, too. > >>>>>>>>>> > >>>>>>>>>> > >>>>>>>>>> > >>>>>>>>>> > >>>>>>>>>> > >>>>>>>>>> > >>>>>>>>> I've got the X-Fi Elite Pro. > >>>>>>>>> That's The one with the external In/Out box. > >>>>>>>>> > >>>>>>>>> Speaking of which, the headphone jack on it does not output a signal > >>>>>>>>> yet, the signal only goes to line out. > >>>>>>>>> > >>>>>>>>> There's some relais on the card that seem to switch these, they click > >>>>>>>>> multiple times with the windows driver and not all all with yours, I > >>>>>>>>> think that's the reason :) > >>>>>>>>> > >>>>>>>>> > >>>>>>>>> > >>>>>>>>> > >>>>>>>>> > >>>>>>>> Original OSS driver doesn't output to external block also, so it > >>>>>>>> wouldn't be easy to make this support I think. > >>>>>>>> > >>>>>>>> > >>>>>>>> > >>>>>>>> > >>>>>>> The values for port->conv[0] and [1] values in sbxfi_playback_open() > >>>>>>> might play some role. It's I2SA_L and I2SA_R, alias DAI_CH_I2SAL > >>>>>>> and DAI_CH_I2SAR, as default. You can try other values, such as, > >>>>>>> DAI_CH_I2SBL, DAI_CH_I2SA1L, and so on. > >>>>>>> > >>>>>>> > >>>>>>> Takashi > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>>> > >>>>>> Latest snapshot has a bug: > >>>>>> make[3]: *** No rule to make target > >>>>>> `/mnt/e/temp/alsa-driver-unstable/acore/jack.o', needed by > >>>>>> `/mnt/e/temp/alsa-driver-unstable/acore/snd.o'. Stop. > >>>>>> > >>>>>> > >>>>>> > >>>>> Already fixed. > >>>>> > >>>>> > >>>>> Takashi > >>>>> > >>>>> > >>>>> > >>>>> > >>>> In file included from /mnt/e/temp/alsa-driver-unstable/acore/jack.c:3: > >>>> /mnt/e/temp/alsa-driver-unstable/acore/../alsa-kernel/core/jack.c: In > >>>> function ‘snd_jack_new’: > >>>> /mnt/e/temp/alsa-driver-unstable/acore/../alsa-kernel/core/jack.c:107: > >>>> error: ‘SW_MICROPHONE_INSERT’ undeclared (first use in this function) > >>>> /mnt/e/temp/alsa-driver-unstable/acore/../alsa-kernel/core/jack.c:107: > >>>> error: (Each undeclared identifier is reported only once > >>>> /mnt/e/temp/alsa-driver-unstable/acore/../alsa-kernel/core/jack.c:107: > >>>> error: for each function it appears in.) > >>>> /mnt/e/temp/alsa-driver-unstable/acore/../alsa-kernel/core/jack.c: In > >>>> function ‘snd_jack_report’: > >>>> /mnt/e/temp/alsa-driver-unstable/acore/../alsa-kernel/core/jack.c:157: > >>>> error: ‘SW_MICROPHONE_INSERT’ undeclared (first use in this function) > >>>> > >>>> > >>> Hmm, it seems broken for older kernels right now. > >>> The easy workaround is to pass --with-cards=sbxfi to configure. > >>> > >>> Anyway, I'll fix it now. > >>> > >>> thanks, > >>> > >>> Takashi > >>> > >>> > >>> > >> --Hey, man, this is cool! Plays just fine (volume ok, speed ok, no > >> --glitches) with 96000, 48000, 44100, 22050, 16/24 bit, Stereo and --Mono!! > >> --I didn't change anything in the source code, so I don't use system > >> --timer. Yes!! > >> > >> However oss (alsa-emulated) is unstable. I'll test more. > >> > > > > Could be due to 96kHz base-rate? Try base_rate=48000. > > If you get still problems, please show the kernel logs with debug=2. > > > > BTW, the jack.c compile error should have been fixed now (hopefully). > > Let me know if you still have the build errors. > > > > > > thanks, > > > > Takashi > > > > > Ok. OpenAL with alsa also seem to cause problems. In both cases, check the period_size and buffer_size values (shown in the kernel message, or /proc/asound/card0/pcm0p/sub0/hw_params). And, try to aplay with these parameters, whether you get the similar problem. % aplay -v --period-size=xxx --buffer-size=yyy foo.wav Takashi

5 22

Re: [alsa-devel] [PATCH] Fix debugfs_create_dir's error checking method for sound/soc/
by Mark Brown 17 Oct '08

17 Oct '08

On Fri, Oct 17, 2008 at 09:04:55PM +0800, Zhaolei wrote: > debugfs_create_dir() returns NULL if an error occurs, returns -ENODEV > when debugfs is not enabled in the kernel. ... > asoc_debugfs = debugfs_create_dir("asoc", NULL); > - if (!IS_ERR(asoc_debugfs)) > + if (!IS_ERR(asoc_debugfs) && asoc_debugfs) > debugfs_create_u32("dapm_pop_time", 0744, asoc_debugfs, > &pop_time); Hrmpf. This looks like something that should be fixed in debugfs - using both error reporting strategies is rather unhelpful and we're actually loosing information in the case where debugfs is built. I'll send a patch and see what people thing. Still in the meantime, Acked-by: Mark Brown <broonie(a)opensource.wolfsonmicro.com>

2 3

[alsa-devel] [PATCH] hda: add SW_LINEOUT_INSERT support
by Matthew Ranostay 17 Oct '08

17 Oct '08

Add support for detecting line out pin insertion and reporting back to userspace with the jack abstraction layer. Line outs are reported with the macro defined SW_LINEOUT_INSERT code. Signed-off-by: Matthew Ranostay <mranostay(a)embeddedalley.com> --- diff --git a/include/linux/input.h b/include/linux/input.h index a5802c9..7323d2f 100644 --- a/include/linux/input.h +++ b/include/linux/input.h @@ -644,6 +644,7 @@ struct input_absinfo { #define SW_RADIO SW_RFKILL_ALL /* deprecated */ #define SW_MICROPHONE_INSERT 0x04 /* set = inserted */ #define SW_DOCK 0x05 /* set = plugged into dock */ +#define SW_LINEOUT_INSERT 0x06 /* set = inserted */ #define SW_MAX 0x0f #define SW_CNT (SW_MAX+1) diff --git a/include/sound/jack.h b/include/sound/jack.h index b1b2b8b..7cb25f4 100644 --- a/include/sound/jack.h +++ b/include/sound/jack.h @@ -35,6 +35,7 @@ enum snd_jack_types { SND_JACK_HEADPHONE = 0x0001, SND_JACK_MICROPHONE = 0x0002, SND_JACK_HEADSET = SND_JACK_HEADPHONE | SND_JACK_MICROPHONE, + SND_JACK_LINEOUT = 0x0004, }; struct snd_jack { diff --git a/sound/core/jack.c b/sound/core/jack.c index bd2d9e6..4978db1 100644 --- a/sound/core/jack.c +++ b/sound/core/jack.c @@ -102,6 +102,9 @@ int snd_jack_new(struct snd_card *card, const char *id, int type, if (type & SND_JACK_HEADPHONE) input_set_capability(jack->input_dev, EV_SW, SW_HEADPHONE_INSERT); + if (type & SND_JACK_LINEOUT) + input_set_capability(jack->input_dev, EV_SW, + SW_LINEOUT_INSERT); if (type & SND_JACK_MICROPHONE) input_set_capability(jack->input_dev, EV_SW, SW_MICROPHONE_INSERT); @@ -153,6 +156,9 @@ void snd_jack_report(struct snd_jack *jack, int status) if (jack->type & SND_JACK_HEADPHONE) input_report_switch(jack->input_dev, SW_HEADPHONE_INSERT, status & SND_JACK_HEADPHONE); + if (jack->type & SND_JACK_LINEOUT) + input_report_switch(jack->input_dev, SW_LINEOUT_INSERT, + status & SND_JACK_LINEOUT); if (jack->type & SND_JACK_MICROPHONE) input_report_switch(jack->input_dev, SW_MICROPHONE_INSERT, status & SND_JACK_MICROPHONE); diff --git a/sound/pci/hda/patch_sigmatel.c b/sound/pci/hda/patch_sigmatel.c index d106ea5..1b3cba8 100644 --- a/sound/pci/hda/patch_sigmatel.c +++ b/sound/pci/hda/patch_sigmatel.c @@ -36,6 +36,7 @@ #include "hda_patch.h" #include "hda_beep.h" +#define STAC_LINEOUT_EVENT 0x10 #define STAC_PWR_EVENT 0x20 #define STAC_HP_EVENT 0x30 #define STAC_VREF_EVENT 0x40 @@ -217,7 +218,8 @@ struct sigmatel_spec { struct hda_pcm pcm_rec[2]; /* PCM information */ /* jack detection */ - struct snd_jack *jack; + struct snd_jack *hp_jack; + struct snd_jack *lineout_jack; /* dynamic controls and input_mux */ struct auto_pin_cfg autocfg; @@ -3635,9 +3637,15 @@ static int stac92xx_init(struct hda_codec *codec) /* jack detection */ err = snd_jack_new(codec->bus->card, "Headphone Jack", - SND_JACK_HEADPHONE, &spec->jack); + SND_JACK_HEADPHONE, &spec->hp_jack); if (err < 0) return err; + err = snd_jack_new(codec->bus->card, + "Lineout Jack", + SND_JACK_LINEOUT, &spec->lineout_jack); + if (err < 0) + return err; + /* fake event to set up pins */ codec->patch_ops.unsol_event(codec, STAC_HP_EVENT << 26); } else { @@ -3661,6 +3669,16 @@ static int stac92xx_init(struct hda_codec *codec) for (i = 0; i < spec->num_dmics; i++) stac92xx_auto_set_pinctl(codec, spec->dmic_nids[i], AC_PINCTL_IN_EN); + for (i = 0; i < cfg->line_outs; i++) { + int def_conf = snd_hda_codec_read(codec, cfg->line_out_pins[i], + 0, AC_VERB_GET_CONFIG_DEFAULT, 0); + def_conf = get_defcfg_connect(def_conf); + + if (def_conf && def_conf != AC_JACK_PORT_FIXED) + continue; + enable_pin_detect(codec, cfg->line_out_pins[i], + STAC_LINEOUT_EVENT); + } for (i = 0; i < spec->num_pwrs; i++) { int event = is_nid_hp_pin(cfg, spec->pwr_nids[i]) ? STAC_HP_EVENT : STAC_PWR_EVENT; @@ -3717,8 +3735,10 @@ static void stac92xx_free(struct hda_codec *codec) if (! spec) return; - if (spec->jack) - snd_device_free(codec->bus->card, spec->jack); + if (spec->hp_jack) + snd_device_free(codec->bus->card, spec->hp_jack); + if (spec->lineout_jack) + snd_device_free(codec->bus->card, spec->lineout_jack); if (spec->bios_pin_configs) kfree(spec->bios_pin_configs); @@ -3804,7 +3824,7 @@ static void stac92xx_hp_detect(struct hda_codec *codec, unsigned int res) break; presence = get_hp_pin_presence(codec, cfg->hp_pins[i]); } - snd_jack_report(spec->jack, + snd_jack_report(spec->hp_jack, presence ? SND_JACK_HEADPHONE : 0); if (presence) { @@ -3840,6 +3860,28 @@ static void stac92xx_hp_detect(struct hda_codec *codec, unsigned int res) stac92xx_set_pinctl(codec, nid, AC_PINCTL_OUT_EN); } +static void stac92xx_lineout_detect(struct hda_codec *codec) +{ + struct sigmatel_spec *spec = codec->spec; + struct auto_pin_cfg *cfg = &spec->autocfg; + int presence, i; + + for (i = 0; i < cfg->line_outs; i++) { + int def_conf = snd_hda_codec_read(codec, + cfg->line_out_pins[i], 0, + AC_VERB_GET_CONFIG_DEFAULT, 0); + def_conf = get_defcfg_connect(def_conf); + + if (def_conf && def_conf != AC_JACK_PORT_FIXED) + continue; + presence = get_hp_pin_presence(codec, + cfg->line_out_pins[i]); + if (presence) + break; + } + snd_jack_report(spec->lineout_jack, presence ? SND_JACK_LINEOUT : 0); +} + static void stac92xx_pin_sense(struct hda_codec *codec, int idx) { struct sigmatel_spec *spec = codec->spec; @@ -3868,15 +3910,20 @@ static void stac92xx_unsol_event(struct hda_codec *codec, unsigned int res) { struct sigmatel_spec *spec = codec->spec; int idx = res >> 26 & 0x0f; + int event = (res >> 26) & 0x70; - switch ((res >> 26) & 0x70) { + switch (event) { case STAC_HP_EVENT: stac92xx_hp_detect(codec, res); /* fallthru */ case STAC_PWR_EVENT: if (spec->num_pwrs > 0) stac92xx_pin_sense(codec, idx); - break; + /* fallthru */ + case STAC_LINEOUT_EVENT: + if (event == STAC_HP_EVENT) + break; + stac92xx_lineout_detect(codec); case STAC_VREF_EVENT: { int data = snd_hda_codec_read(codec, codec->afg, 0, AC_VERB_GET_GPIO_DATA, 0);

3 8

[alsa-devel] SB X-Fi driver status update
by Takashi Iwai 17 Oct '08

17 Oct '08

Hi, now I'm stabilizing the driver a bit. I hope it won't hang up any more (so often), and make you happy testing. ** DOWNLOAD ** The snd-sbxfi driver is included in my sound-unstable git tree, and alsa-driver-unstable snapshot tarball. git://git.kernel.org/pub/scm/linux/kernel/tiwai/sound-unstable-2.6.git ftp://ftp.kernel.org/pub/linux/kernel/people/tiwai/snapshot/alsa-driver-uns… The git commits can be browsed via http://git.kernel.org/?p=linux/kernel/git/tiwai/sound-unstable-2.6.git;a=su… The snapshot tarball includes HEAD and alsa-kernel/HEAD files. These contain alsa-driver and sound git commit logs. Check these files and compare with the above gitweb page whether you have the latest version. ** BUILD ** If you want to build the sbxfi driver on your 2.6.27 git tree, at best, do the following: - pull sound-2.6.git tree devel branch % git pull \ git://git.kernel.org/pub/scm/linux/kernel/tiwai/sound-2.6.git devel - pull sound-unstable-2.6.git tree topic/sbxfi branch % git pull \ git://git.kernel.org/pub/scm/linux/kernel/tiwai/sound-unstable-2.6.git\ topic/sbxfi In that way, you can update only the sound and keep the rest clean. (The master branches of sound and sound-unstable git trees are for the latest upsteram, i.e. 2.6.28-pre git tree.) In other versions, it'd be better to use alsa-driver-unstable snapshot tarball to build external modules. ** UPDATED NOTES ** Below are some updates: - As default, the driver works only as a constant sample rate, either 48000 or 96000 (default). The rate can be changed via the module option base_rate. For enabling the continuous rate support (it was on before), define XXX_CONT_RATE in sbxfi.c. - The debug module option is 1 as default, i.e. showing some debug messages but not too often. You can set debug=0 to suppress, or debug=2 or 3 to see more. If you have problems (especially regarding with the sound quality problem), run with debug=2. - Please check whether the default (base_rate=96000) works as is for different rates, formats, channels. Then try base_rate=48000. - If you are brave, try recording, too :) It's no full-duplex, so you can't play and record at the same time. Define XXX_FULL_DUPLEX in sbxfi.c to enable the (possible) full-duplex. - If playback works well, try to define XXX_USE_SG in sbxfi.c. This will make the driver using SG buffer instead of continuous pages. - Also, test SBXFI_DMA_MASK being DMA_32BIT_MASK, especially if you have RAM more than 2GB. - You can try dmix as well. E.g. simply run like % aplay -Dplug:dmix:0 foo.wav thanks, Takashi

4 5

Re: [alsa-devel] Backported sbxfi driver (UNTESTED!)
by Bjoern Olausson 17 Oct '08

17 Oct '08

I found a site with a large range of files: http://bellatrix.ece.mcgill.ca/Documents/AudioFormats/WAVE/Samples.html aplay plays all files it is capable to decode. No crackel, no noise, just fine. glass.wav Signed 16 bit Little Endian, Rate 8000 Hz, Mono M1F1-int16-AFsp.wav Signed 16 bit Little Endian, Rate 8000 Hz, Stereo M1F1-int24-AFsp.wav Signed 24 bit Little Endian in 3bytes, Rate 8000 Hz, Stereo M1F1-int32-AFsp.wav Signed 32 bit Little Endian, Rate 8000 Hz, Stereo M1F1-uint8-AFsp.wav Unsigned 8 bit, Rate 8000 Hz, Stereo stereol.wav Signed 16 bit Little Endian, Rate 22050 Hz, Stereo Utopia_Critical_Stop.wav Signed 16 bit Little Endian, Rate 22050 Hz, Mono Thought when I try to play the first one with mplayer ---> FREEZZZZZ Spiele glass.wav. Audio file-Dateiformat erkannt! ========================================================================== Öffne Audiodecoder: [pcm] Uncompressed PCM audio decoder AUDIO: 8000 Hz, 1 ch, s16le, 128.0 kbit/100.00% (ratio: 16000->16000) Ausgewählter Audiocodec: [pcm] afm: pcm (Uncompressed PCM) ========================================================================== Amarok with backend ALSA plays all files it can decode without problems. All Signed 16 bit Little Endian, Rate 8000 to 96000 Hz, Stereo do work (with the same oszilating noise described in my previous post with the mono files [I just joned two to get stereo])

2 1

[alsa-devel] HD-audio reconfigure branch merged to master
by Takashi Iwai 16 Oct '08

16 Oct '08

Hi, FYI, I merged the long-pending hda-reconfig topic branch to master and devel branches of sound.git tree for 2.6.29 kernel (not 28). This allows you to reconfigure the HD-audio setup dynamically by triggering via sysfs entries. I'll give more details later... Anyway, let me know if you encounter any build problems or Oops by these changes. thanks, Takashi

1 0

[alsa-devel] Salsa-lib : _snd_pcm_hw_param_test is not defined
by Matthieu CASTET 16 Oct '08

16 Oct '08

Hi, it seems there is a missing implementation of _snd_pcm_hw_param_test function in src/pcm_params.c . It is used in pcm_macros.h by snd_pcm_hw_params_test_xxx functions. Thanks Matthieu

2 6