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

Re: [alsa-devel] Boot failure with 2.6.21 with snd-maestro3
by Giridhar Pemmasani 02 May '07

02 May '07

--- Adrian Bunk <bunk(a)stusta.de> wrote: > On Sun, Apr 29, 2007 at 09:55:18AM -0700, Giridhar Pemmasani wrote: > > When snd-maestro3 (ESS Maestro 3 driver) is enabled, kernel 2.6.21 fails > to > > boot - it seems to stop at various points depending on what other modules > are > > enabled. Disabling snd-maestro3 fixes all issues. Moreover, with > everything > > else being same, using maestro3.c from 2.6.20 kernel and using it in > 2.6.21 > > kernel also works - i.e., system boots fine. I will try to isolate the > issue > > and follow up. > > There were only the two attached patches between 2.6.20 and 2.6.21 to > this driver: > - maestro3: add request_firmware() > - maestro3 - Use quirk list helper function > > 2.6.21 + 2.6.20 maestro3.c works for you. > > Does 2.6.21 + 2.6.20 maestro3.c + patch-maestro3-1 work? 2.6.21 with 2.6.20's maestro3.c with patch-maestro3-1 exhibited the same issue. However, I identified the cause: maestro3 driver in 2.6.21 (even just patch-maestro3-1) tries to load firmware file from user space. But each time request_firmware is called it takes about 1 minute to timeout (because I don't have the firmware installed?). If I let the system take its time, even with maestro3.c from 2.6.21, it boots after timing out twice. So after about 2 minutes of no activity, system does boot. Following are trace messages I added to identify the issue: [ 17.004000] input: TPPS/2 IBM TrackPoint as /class/input/input5 [ 70.904000] sound/pci/maestro3.c (snd_m3_create:2763): -2 [ 70.920000] sound/pci/maestro3.c (snd_m3_create:2774): -2 [ 130.940000] sound/pci/maestro3.c (snd_m3_create:2777): -2 [ 130.956000] sound/pci/maestro3.c (snd_m3_create:2788): -2 [ 130.972000] sound/pci/maestro3.c (snd_m3_create:2794): 0 [ 130.992000] sound/pci/maestro3.c (snd_m3_create:2797): 0 [ 131.012000] sound/pci/maestro3.c (snd_m3_create:2802): 0 [ 131.044000] sound/pci/maestro3.c (snd_m3_create:2804): 0 [ 131.580000] sound/pci/maestro3.c (snd_m3_create:2807): 0 [ 131.596000] sound/pci/maestro3.c (snd_m3_create:2810): 0 [ 131.612000] sound/pci/maestro3.c (snd_m3_create:2814): 0 [ 131.628000] sound/pci/maestro3.c (snd_m3_create:2821): 0 [ 132.364000] device-mapper: ioctl: 4.11.0-ioctl (2006-10-12) initialised: dm-d evel(a)redhat.com Giri __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

2 1

[alsa-devel] Audio DMA buffer layout and setup
by Markus Korber 02 May '07

02 May '07

Hi, I'm working with a custom sound chip with its own DMA hardware. The figure shows how the I2S DMA expects data to be prepared: IRQ1 +----------------+ | page_3 (right) |\ +----------------+ > double buffer 1 | page_2 (left) |/ IRQ0 +----------------+ | page_1 (right) |\ +----------------+ > double buffer 0 | page_0 (left) |/ +----------------+ The hardware has 2 double buffers, each containing left and right sample regions, i.e. it alternately transmits samples from page_0 and page_1, i.e. double buffer 0, then generates an interrupt and starts transmitting page_2 and page_3. After that another interrupt is generated and the process begins again with double buffer 0. So my question is, how to tell ALSA about this memory layout? Is it possible to operate in MMAP mode with SNDRV_PCM_INFO_NONINTERLEAVED? But then, how to tell the application that it should either fill double buffer 0 or 1? Or should I use RW mode? Or something else? Furthermore, since the hardware doesn't provide a 'current' DMA pointer containing the DMA address being transmitted at the moment (I only have interrupts generated after each double buffer has been transmitted), what should the pcm_pointer() callback return? (I guess it should return 0 when called before IRQ0 and sizeof(double buffer 0) between IRQ0 and IRQ1?) Best regards, Markus Korber

1 0

[alsa-devel] driver mixer control questions
by Tim Harvey 02 May '07

02 May '07

I'm troubleshooting an issue with an ALSA driver I wrote for a very unique hardware platform. In this case I have an audio codec which only has a usable 'line-in' capture source. Currently the only mixer control I've created is a 'Master Playback Volume'. Do I need to create some kind of control for 'line-in'? I'm unclear how to 'register' capture inputs so that they will be returned by the OSS mixer emulation's ioctl SOUND_MIXER_READ_RECMASK. Currently that ioctl returns 0 for my mixer device. Also, is a control 'switch' such as 'Master Playback Switch' simply a mute control or is it intended to be any type of control which could have enumerated values? Thanks, Tim

1 0

Re: [alsa-devel] Support for SAM9407?
by Ash Willis 02 May '07

02 May '07

Thanks for your detailed response. That chip definitely does seem like a beast that I really don't want to tame... > > How about a Media Vision Pro Audio Spectrum 16? Those can still be found > relatively easily and it's one the cards for which an OSS/Free driver is > still kept in the tree. I have a pair of those as well so I'll happily test > if you want to write a driver for it... > Yeah, we have a winner. I just ordered one. I'm busy right now but I will get this started in 4 weeks. Thanks for the offer of testing. I'll hunt you down when I have something to show you :) Ash =

1 0

[alsa-devel] porter wanted: AMD Geode SCx200 driver OSS to ALSA
by Martin-Éric Racine 02 May '07

02 May '07

Greetings, I work for a company that develops hardware using the AMD Geode chipset. While our current product line is based on the Geode LX model (which is very well supported in ALSA), our legacy products are based on the earlier Geode SC1200 and SC2200, for which no ALSA support exists. However, AMD has OSS driver source code available at: http://www.amd.com/us-en/ConnectivitySolutions/TechnicalResources/0,,50_233… We basically need this ported to an ALSA snd-scx200 module to replace it. (Would now also be a good time to ask for snd-sc5535audio to be renamed snd-sc5535 for consistence with other driver names?) We could provide a sample SC2200-based unit to whoever volunteers for porting AMD's old OSS code to ALSA, Please contact me for details. The resulting code will be directly submitted to the ALSA project for inclusion in the mainstream kernel, to make sure that everyone can benefit from it. Best Regards, -- Martin-Éric Racine http://q-funk.iki.fi

1 1

Re: [alsa-devel] Endiannes and alsa (in an embedded device)
by Ciaccia 02 May '07

02 May '07

Forget about that, it took 4 hours but I understood why it was not working... Since I'm working with 24bit resolution (and not 16 as I usually do), the samples were to little... Now I replaced the sin line with: sample = (int)(4000000.0f * sin((float)i * 2.0f * M_PI / (512.0f / 50.0f))); and it's working fine. I really thought it was too strange to be true... Sorry for the spam :-| Andrea ----- Original Message ---- From: Ciaccia <ciacciax(a)yahoo.com> To: alsa-devel(a)alsa-project.org Sent: Tuesday, May 1, 2007 8:49:18 PM Subject: [alsa-devel] Endiannes and alsa (in an embedded device) Dear all, I have a question that really puzzles me, I wonder if someone can explain to me what I'm doing wrong... :-| I have programmed an ALSA driver for an embedded device, everything works fine and all the application use my driver without any glitches. The snd_pcm_hardware for this device is defined as following: static const struct snd_pcm_hardware ep93xx_pcm_hardware = { .info= SNDRV_PCM_INFO_MMAP | SNDRV_PCM_INFO_MMAP_VALID | SNDRV_PCM_INFO_INTERLEAVED | SNDRV_PCM_INFO_BLOCK_TRANSFER, .rates = SNDRV_PCM_RATE_8000_48000, .rate_min = 8000, .rate_max = 48000, .channels_min = 2, .channels_max = 2, .formats = SNDRV_PCM_FMTBIT_S24_LE, .buffer_bytes_max = 131072, .period_bytes_min = 512, .period_bytes_max = 32768, .periods_min = 1, .periods_max = MAX_PERIODS, .fifo_size = 32, }; Then I tried to program a test application that uses the hardware device directly (without using alsa-lib to convert the format and so on). I open the device as following: if ((err = snd_pcm_open (&playback_handle, "hw:0,0", SND_PCM_STREAM_PLAYBACK, 0)) < 0) { Everything works fine until I try to write data to the stream. I tried both SND_PCM_ACCESS_RW_INTERLEAVED and SND_PCM_ACCESS_MMAP_INTERLEAVED, but I always have the following problem. When I want to write data to the stream, I have to "reverse" the byte order, otherwise I hear noise on the speaker. The system I am working on is little-endian (it has cirrus ep93xx arm core), so I assume I could write data directly to the buffer. Here is what I have to do: // fill the buffer for (i = 0; i < 512; i++) { int sample = (int)(20000.0f * sin((float)i * 2.0f * M_PI / 512.0f)); int reversed; char *src = (char *)(&sample); char *dst = (char *)(&reversed); dst[0] = src[3]; dst[1] = src[2]; dst[2] = src[1]; dst[3] = src[0]; *(buffer++) = reversed; *(buffer++) = reversed; } It this normal? I thought that since the system is little-endian and the format of the driver is S24_LE, I could copy the data directly to the stream (either using snd_pcm_writei or snd_pcm_mmap_begin). I think the driver is ok, since all the applications run with no problem at all. Can someone explain to me how/why the applications such as aplay/arecord work ok? Every help is greatly appreciated Thanks in advance Andrea __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com _______________________________________________ Alsa-devel mailing list Alsa-devel(a)alsa-project.org http://mailman.alsa-project.org/mailman/listinfo/alsa-devel __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

1 0

[alsa-devel] Endiannes and alsa (in an embedded device)
by Ciaccia 01 May '07

01 May '07

Dear all, I have a question that really puzzles me, I wonder if someone can explain to me what I'm doing wrong... :-| I have programmed an ALSA driver for an embedded device, everything works fine and all the application use my driver without any glitches. The snd_pcm_hardware for this device is defined as following: static const struct snd_pcm_hardware ep93xx_pcm_hardware = { .info= SNDRV_PCM_INFO_MMAP | SNDRV_PCM_INFO_MMAP_VALID | SNDRV_PCM_INFO_INTERLEAVED | SNDRV_PCM_INFO_BLOCK_TRANSFER, .rates = SNDRV_PCM_RATE_8000_48000, .rate_min = 8000, .rate_max = 48000, .channels_min = 2, .channels_max = 2, .formats = SNDRV_PCM_FMTBIT_S24_LE, .buffer_bytes_max = 131072, .period_bytes_min = 512, .period_bytes_max = 32768, .periods_min = 1, .periods_max = MAX_PERIODS, .fifo_size = 32, }; Then I tried to program a test application that uses the hardware device directly (without using alsa-lib to convert the format and so on). I open the device as following: if ((err = snd_pcm_open (&playback_handle, "hw:0,0", SND_PCM_STREAM_PLAYBACK, 0)) < 0) { Everything works fine until I try to write data to the stream. I tried both SND_PCM_ACCESS_RW_INTERLEAVED and SND_PCM_ACCESS_MMAP_INTERLEAVED, but I always have the following problem. When I want to write data to the stream, I have to "reverse" the byte order, otherwise I hear noise on the speaker. The system I am working on is little-endian (it has cirrus ep93xx arm core), so I assume I could write data directly to the buffer. Here is what I have to do: // fill the buffer for (i = 0; i < 512; i++) { int sample = (int)(20000.0f * sin((float)i * 2.0f * M_PI / 512.0f)); int reversed; char *src = (char *)(&sample); char *dst = (char *)(&reversed); dst[0] = src[3]; dst[1] = src[2]; dst[2] = src[1]; dst[3] = src[0]; *(buffer++) = reversed; *(buffer++) = reversed; } It this normal? I thought that since the system is little-endian and the format of the driver is S24_LE, I could copy the data directly to the stream (either using snd_pcm_writei or snd_pcm_mmap_begin). I think the driver is ok, since all the applications run with no problem at all. Can someone explain to me how/why the applications such as aplay/arecord work ok? Every help is greatly appreciated Thanks in advance Andrea __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

1 0

[alsa-devel] Implementing Priority Based Device ALSA device interface. ALSA drivers
by Nobin Mathew 01 May '07

01 May '07

Hi Friends, I am a new to ALSA, I am developing a ALSA ASoC driver. Now driver can support both audio playback and capture. Now the problem is handling of application with different priority. We have three applications say app1, app2 and app3; they are in increasing order of priority. One scenario of app1 and app2 is below If app2 is accessing the device while app1 playback is going on, then app1 should stop (clearly do pause) playback and allow app2 to playback, and app1 should continue playback when app2 exit. The requirement is that there should be no changes to apps, except the devices each app will access. There will be no daemons in user space which will handle priority. For that i implemented three interfaces like pcmC0p0, pcmC0p1 and pcmC0p2 there are in increaing order of priority. All these interfaces will access the same audio hardware. Only thing is each will have separate ALSA kernel buffer. In this implementation app1 will always access pcmC0p0, and app2 pcmC0p1 like that. What i am doing is, in driver i will pause the pcmC0p0 substream using snd_pcm_pause() when pcmC0p1 is accessed. Same for pcmc0p2. This call will change the state of pcmc0p0 substream to SNDRV_PCM_STATE_PAUSED, and this call should notify the user app also. When app2 playback is over i will change the state of pcmC0p0 substream to SNDRV_PCM_STATE_RUNNING by another call to snd_pcm_pause(). This implementation requires some minor modifications in the ALSA lib and ALSA app (handling of PAUSE from kernel). But what i found is kernel is not notifying the user using snd_timer_notify() (user call back) function because i am using the substream timer snd_pcm_timer which is there in pcm_timer.c. For this timer SNDRV_TIMER_HW_SLAVE flag is set so it will not notify the user. What is the use of these timers in ALSA? Am i going wrong? Is my design correct? Can u suggest some alternate design? Sorry for my poor English. Thanks Nobin Mathew

1 0

[alsa-devel] Set fragment size for USB sound card
by Jason Heeris 01 May '07

01 May '07

System: NEC Versa M540 laptop Dist: Debian Etch/Lenny Kernel: 2.6.18 Alsa: using drivers in kernel, although no difference when compiled from alsa-source package. Card in question: C-Media USB sound card (more info below) Hi there, I realise that this may not be strictly a developer list question, but someone on the alsa-user list suggested I try here. Any help would be appreciated. The question is a continuation of: http://sourceforge.net/mailarchive/forum.php?thread_name=1176015858.4497.9.… and http://www.linuxquestions.org/questions/showthread.php?t=540205 Essentially, I have a C-Media USB audio adaptor with which to use my headphones, and it only works with a fragment size of 2048 or less (discovered using 'alsaplayer -f 2048'). I would like to be able to plug it in, have ALSA recognise it and use this fragment size for it. Without specifying this fragment size, I get: -- $ alsaplayer -i text -o alsa -d hw:2 /usr/share/sounds/generic.wav Unable to install hw params: ACCESS: RW_INTERLEAVED FORMAT: S16_LE SUBFORMAT: STD SAMPLE_BITS: 16 FRAME_BITS: 32 CHANNELS: 2 RATE: 44100 PERIOD_TIME: (23219 23220) PERIOD_SIZE: 1024 PERIOD_BYTES: 4096 PERIODS: 8 BUFFER_TIME: (185759 185760) BUFFER_SIZE: 8192 BUFFER_BYTES: 32768 TICK_TIME: 1000 failed to configure output device...trying OSS ...done playing -- ...and the sound comes through the laptop speakers. I tried defining a slave PCM: -- pcm.usbtest { type dmix ipc_key 192837465 ipc_key_add_uid true slave { pcm "hw:2,0" # period_size 2048 buffer_size 2048 } } -- But trying 'alsaplayer -i text -o alsa -d usbtest /usr/share/sounds/generic.wav' gave exactly the same error. >From my previous post: -- ...I also tried forcing the index of the card by adding the following to /etc/modprobe.d/alsa-base, but even this only works half of the time: -- # Prevent abnormal drivers from grabbing index 0 options snd-bt87x index=-2 options snd-cx88_alsa index=-2 options snd-atiixp-modem index=-2 # options snd-intel8x0m index=-2 options snd-via82xx-modem index=-2 options snd-intel8x0 index=0 options snd-intel8x0m index=1 options snd-usb-audio index=2 vid=0d8c pid=000c options snd-usb-audio index=3 -- The contents of '/proc/asound/cards' are: -- 0 [I82801DBICH4 ]: ICH4 - Intel 82801DB-ICH4 Intel 82801DB-ICH4 with ALC655 at 0xf0000400, irq 5 1 [Modem ]: ICH-MODEM - Intel 82801DB-ICH4 Modem Intel 82801DB-ICH4 Modem at 0xe200, irq 5 2 [default ]: USB-Audio - C-Media USB Headphone Set C-Media USB Headphone Set at usb-0000:00:1d.1-1, full speed -- ...although sometimes the ICH-MODEM is missing and the C-Media card is at #1. Card information from 'lsusb': Bus 002 Device 003: ID 0d8c:000c C-Media Electronics, Inc. Audio Adapter Usually I boot with the card plugged in, but if I plug it in after booting, the new modules loaded are: snd_hwdep, snd_rawmidi, snd_usb_audio and snd_usb_lib. -- Anyway, as I said, any help would be appreciated. Cheers, Jason

1 0

[alsa-devel] Possible bug in snd_mixer_selem_set_playback_switch_all
by Tim Burrell 30 Apr '07

30 Apr '07

Hi there, I've encountered a situation where if I very quickly, and repeatedly make calls to snd_mixer_selem_set_playback_switch_all I get some strange undefined behavior. The situation is I'm looping over all mixer elements (on all sound cards) to do a system wide mute / unmute. If I make numerous calls to this function it ends up that some mixers only toggle switches on specific channels (right or left), and sometimes the volume on certain (random) channels gets set to some also random (usually 0, but can be anywhere between 0 and max) level. I've been able to remedy the situation by inserting a 10ms delay in between calls to snd_mixer_selem_set_playback_switch_all. Clearly though this is undesirable. I don't have more than one sound card to test on so I can't be sure if it's a problem with all drivers or just mine (hda-intel). I can certainly provide more information if requested, just let me know what I can do. Or if I'm doing something wrong I'd love to be made aware of something I can do to fix this behavior. Thanks, Tim.

1 0