Re: [alsa-devel] No sound with nVidia Corporation MCP79 High Definition Audio (rev b1) (Macbook Pro 5, 5)

6 Jul 2009

At Mon, 6 Jul 2009 17:23:17 +0200,
Andreas Nüßlein wrote:
...
On Monday 06 July 2009 17:13:29 you wrote:
...
At Mon, 6 Jul 2009 17:03:52 +0200,
Andreas Nüßlein wrote:
...
On Monday 06 July 2009 16:47:48 you wrote:
...
At Mon, 6 Jul 2009 16:36:09 +0200,
Andreas Nüßlein wrote:
...
On Monday 06 July 2009 16:16:06 you wrote:
...
At Mon, 6 Jul 2009 16:03:04 +0200,
Andreas Nüßlein wrote:
> On Monday 06 July 2009 15:39:09 you wrote:
> > At Mon, 6 Jul 2009 14:59:09 +0200,
> >
> > Andreas Nüßlein wrote:
> > > On Monday 06 July 2009 14:44:56 you wrote:
> > > > At Mon, 6 Jul 2009 12:19:11 +0200,
> > > >
> > > > Andreas Nüßlein wrote:
> > > > > On Monday 06 July 2009 12:08:11 you wrote:
> > > > > > At Mon, 6 Jul 2009 11:45:07 +0200,
> > > > > >
> > > > > > Andreas Nüßlein wrote:
> > > > > > > [1  <text/plain; iso-8859-1 (quoted-printable)>]
> > > > > > >
> > > > > > > On Monday 06 July 2009 09:23:07 you wrote:
> > > > > > > > At Sun, 5 Jul 2009 12:30:52 +0200,
> > > > > > > >
> > > > > > > > > however i was able to reproduce the "bad kfree",
> > > > > > > > > by changing the volumes in alsamixer. it always
> > > > > > > > > gave me the same "called from ID" so i tried
> > > > > > > > > "while (true) do cat /proc/kallsyms | grep
> > > > > > > > > ffffffffa001b08b; done" and then changed the
> > > > > > > > > volume, but the grep never returned anything,
> > > > > > > > > dmesg of course got more and more of those
> > > > > > > > > kfree-errors.
> > > > > > > >
> > > > > > > > The address won't match exactly in /proc/kasllsyms.
> > > > > > > > But you can guess which function by address.  First
> > > > > > > > sort /proc/kallsyms then check the position that
> > > > > > > > fits with the given address.
> > > > > > > >
> > > > > > > >
> > > > > > > > Takashi
> > > > > > >
> > > > > > > well i'm not really sure _I_ can guess ;)    i guess
> > > > > > > it's either "t snd_ctl_ioctl        [snd]" or  "t
> > > > > > > snd_ctl_ioctl_compat [snd]".
> > > > > > >
> > > > > > > here are the greps, "snd: bad kfree (called from
> > > > > > > ffffffffa001b08b)" being the error:
> > > > > > >
> > > > > > >
> > > > > > > $ cat /proc/kallsyms | sort | grep ffffffffa001a
> > > > > > > ffffffffa001a045 t snd_ctl_elem_info    [snd]
> > > > > > > ffffffffa001a17a t copy_ctl_value_from_user     [snd]
> > > > > > > ffffffffa001a34f t snd_ctl_elem_read    [snd]
> > > > > > > ffffffffa001a413 t snd_ctl_elem_write   [snd]
> > > > > > > ffffffffa001a515 T snd_ctl_rename_id    [snd]
> > > > > > > ffffffffa001a5b5 T snd_ctl_activate_id  [snd]
> > > > > > > ffffffffa001a66e T snd_ctl_remove_id    [snd]
> > > > > > > ffffffffa001a6ed T snd_ctl_add  [snd]
> > > > > > > ffffffffa001a6ed u snd_ctl_add  [snd_hda_codec]
> > > > > > > ffffffffa001a947 t snd_ctl_elem_add     [snd]
> > > > > > > ffffffffa001abd3 t snd_ctl_elem_add_compat      [snd]
> > > > > > > ffffffffa001ad37 t snd_ctl_elem_add_user        [snd]
> > > > > > > ffffffffa001ada6 t snd_ctl_ioctl        [snd]
> > > > > > >
> > > > > > >
> > > > > > > $ cat /proc/kallsyms | sort | grep ffffffffa001b
> > > > > > > ffffffffa001b4c4 t snd_ctl_ioctl_compat [snd]
> > > > > > > ffffffffa001b9ba T snd_ctl_new1 [snd]
> > > > > > > ffffffffa001b9ba u snd_ctl_new1 [snd_hda_codec]
> > > > > > > ffffffffa001b9ba u snd_ctl_new1
> > > > > > > [snd_hda_codec_cirrus] ffffffffa001bb30 T
> > > > > > > snd_pci_quirk_lookup [snd] ffffffffa001bb30 u
> > > > > > > snd_pci_quirk_lookup [snd_hda_codec] ffffffffa001bb30
> > > > > > > u snd_pci_quirk_lookup [snd_hda_intel]
> > > > > > > ffffffffa001bb93 T snd_verbose_printk   [snd]
> > > > > > > ffffffffa001bb93 u snd_verbose_printk  
> > > > > > > [snd_hda_codec] ffffffffa001bb93 u snd_verbose_printk
> > > > > > >   [snd_hda_intel] ffffffffa001bb93 u
> > > > > > > snd_verbose_printk   [snd_hwdep] ffffffffa001bb93 u
> > > > > > > snd_verbose_printk   [snd_pcm] ffffffffa001bb93 u
> > > > > > > snd_verbose_printk   [snd_timer] ffffffffa001bc7b T
> > > > > > > release_and_free_resource    [snd] ffffffffa001bcd4 t
> > > > > > > snd_device_register_all      [snd] ffffffffa001bd76 T
> > > > > > > snd_device_register  [snd] ffffffffa001bd76 u
> > > > > > > snd_device_register  [snd_pcm] ffffffffa001be67 t
> > > > > > > snd_device_disconnect        [snd] ffffffffa001bf5d t
> > > > > > > snd_device_disconnect_all    [snd]
> > > > > > >
> > > > > > >
> > > > > > > hope that helps =)
> > > > > >
> > > > > > Maybe the patch below can give a better clue.
> > > > > > This will spew stack traces when this error is
> > > > > > triggered. Don't be surprised.
> > > > >
> > > > > oki - new output =)
> > > > >
> > > > >
> > > > > [16306.365597] snd: bad kfree (called from
> > > > > ffffffffa001b08f) [16306.365604] Pid: 26533, comm:
> > > > > alsamixer Tainted: P W 2.6.30.1 #2 [16306.365610] Call
> > > > > Trace:
> > > > > [16306.365632]  [<ffffffffa001b08f>] ?
> > > > > snd_ctl_ioctl+0x2e5/0x71e [snd] [16306.365645]
> > > > > [<ffffffff802b8bf9>] ?
> > > > > __inc_zone_state+0x20/0x9b [16306.365655]
> > > > > [<ffffffff802cfffc>] ? alloc_page_vma+0x197/0x1e1
> > > > > [16306.365666]  [<ffffffff802afb63>] ?
> > > > > __lru_cache_add+0x93/0xdd [16306.365676]
> > > > > [<ffffffff802ea651>] ? vfs_ioctl+0x35/0x97
> > > > > [16306.365686]  [<ffffffff802eaae5>] ?
> > > > > do_vfs_ioctl+0x432/0x482 [16306.365696] 
> > > > > [<ffffffff806336bf>] ?
> > > > > _spin_lock_irqsave+0x28/0x5c [16306.365705]
> > > > > [<ffffffff80247e22>] ? do_page_fault+0x272/0x29e
> > > > > [16306.365715]  [<ffffffff802eab82>] ?
> > > > > sys_ioctl+0x4d/0x81 [16306.365726]  [<ffffffff8022af2b>]
> > > > > ?
> > > > > system_call_fastpath+0x16/0x1b
> > > >
> > > > Try the very latest alsa-driver-unstable snapshot.
> > > > This bug should have been fixed, also possibly with CS420x
> > > > parser bug, too.
> > > >
> > > >
> > > > Takashi
> > >
> > > so dmesg knows the following:
> > >
> > > [18042.848549] HDA Intel 0000:00:08.0: PCI INT A ->
> > > Link[LAZA] -> GSI 23 (level, low) -> IRQ 23 [18042.848623]
> > > HDA Intel 0000:00:08.0: setting latency timer to 64
> > > [18043.007568] ALSA
> > > /home/nutz/tmp/alsa/alsa-driver-unstable/pci/hda/../../alsa-k
> > >erne l/pc i/hd a/hda_codec.c:3851: autoconfig: line_outs=1
> > > (0xa/0x0/0x0/0x0/0x0) [18043.007584] ALSA
> > > /home/nutz/tmp/alsa/alsa-driver-unstable/pci/hda/../../alsa-k
> > >erne l/pc i/hd a/hda_codec.c:3855:    speaker_outs=1
> > > (0xb/0x0/0x0/0x0/0x0) [18043.007596] ALSA
> > > /home/nutz/tmp/alsa/alsa-driver-unstable/pci/hda/../../alsa-k
> > >erne l/pc i/hd a/hda_codec.c:3859:    hp_outs=1
> > > (0x9/0x0/0x0/0x0/0x0) [18043.007608] ALSA
> > > /home/nutz/tmp/alsa/alsa-driver-unstable/pci/hda/../../alsa-k
> > >erne l/pc i/hd a/hda_codec.c:3860:    mono: mono_out=0x0
> > > [18043.007618] ALSA
> > > /home/nutz/tmp/alsa/alsa-driver-unstable/pci/hda/../../alsa-k
> > >erne l/pc i/hd a/hda_codec.c:3863:    dig-out=0x10/0x15
> > > [18043.007628] ALSA
> > > /home/nutz/tmp/alsa/alsa-driver-unstable/pci/hda/../../alsa-k
> > >erne l/pc i/hd a/hda_codec.c:3871:    inputs: mic=0xd,
> > > fmic=0x0, line=0xc, fline=0x0, cd=0x0, aux=0x0
> > > [18043.007642] ALSA
> > > /home/nutz/tmp/alsa/alsa-driver-unstable/pci/hda/../../alsa-k
> > >erne l/pc i/hd a/hda_codec.c:3873:    dig-in=0x12
> > > [18043.121187] ALSA
> > > /home/nutz/tmp/alsa/alsa-driver-unstable/acore/control.c:337:
> > > control 2:0:0:IEC958 Capture Switch:0 is already present
> > > [18043.121204] hda_codec: cannot build controlsfor #0 (error
> > > -16)
> > >
> > >
> > > and i don't have a mixer anymore ;)
> >
> > It's because your chip was uninitialized by BIOS or whatever,
> > the another state you got sometimes.
> >
> > Now fixed in the driver to force to initialize pins.
> > Try the latest tarball again.
> >
> >
> > Takashi
>
> TAKASHIIIIIIIIIIIII =)
> !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
>
> digital out is working... i tried the new snapshot (eventhough,
> as you said my subsystem is now again uninitialized)  and i got
> back the alsamixer- capabilities;
>
> i ran mplayer test.wav and started playing around with the
> mixer-controls.
>
> when i reached IEC958 and turned it on, my out-jack began to glow
> red =) so that works.   i rushed over to my dorm-partners
> speakersystem and plugged it in - voila that actually works! (i
> tried a DTS video with mplayer -vc hwdts )
>
> that is SO cool =)
>
> however as you might've figured out, analog out is not working,
> yet ;( i tried the built-in speakers as well as the headphones
> (and of course all the
Hm, so you don't get any output although no errors with aplay?
Please give alsa-info.sh output with the latest driver to be sure.
Takashi
and here is the alsa-info with the "correct" subsystem id.
i also was able to get the wrong subsystem id again: i did
/etc/init.d/alsasound stop  && modprobe snd-hda-intel   - that
changed it. (i'll have a look into  init.d/alsasound to see if there
are any clues)
Interesting.  So, resetting the controller alone can change the
SSID.  It doesn't happen on other systems.
Could you try to apply the patch below on alsa-kernel directory with
-p2?  This will disable the additional COEF setups, which I blindly
copied from older cs4207 patch.
thanks,
Takashi

diff --git a/sound/pci/hda/patch_cirrus.c
b/sound/pci/hda/patch_cirrus.c index b1fd183..fda83b2 100644
--- a/sound/pci/hda/patch_cirrus.c
+++ b/sound/pci/hda/patch_cirrus.c
@@ -933,7 +933,7 @@ static void init_input(struct hda_codec *codec)
   change_cur_input(codec, spec->cur_input, 1);
   if (spec->mic_detect)
   	cs_automic(codec);



+#if 0
   coef = 0x000a; /* ADC1/2 - Digital and Analog Soft Ramp */
   if (is_active_pin(codec, CS_DMIC2_PIN_NID))
   	coef |= 0x0500; /* DMIC2 enable 2 channels, disable GPIO1 */
@@ -943,6 +943,7 @@ static void init_input(struct hda_codec *codec)
   			 * IDX_SPDIF_CTL.
   			  */
   cs_vendor_coef_set(codec, IDX_ADC_CFG, coef);
+#endif
 }
static struct hda_verb cs_coef_init_verbs[] = {
@@ -980,10 +981,10 @@ static int cs_init(struct hda_codec *codec)
 {
   struct cs_spec *spec = codec->spec;

snd_hda_sequence_write(codec, cs_coef_init_verbs);


/*snd_hda_sequence_write(codec, cs_coef_init_verbs);*/
init_output(codec);
init_input(codec);


init_digital(codec);


/*init_digital(codec);*/
return 0;

}
wow okay...
so aplay test.wav  didn't give me any sound, but when i tried to stop it
(Ctrl-C) i couldn't and i had to kill it from another term, which
resulted in this:
When you run like below
   % aplay -Dplughw -vvv somefile.wav
does it proceed?
If not, it means that the DMA got stuck somewhere.  If it goes,
then the problem is in another place.
Takashi
$ aplay -Dplughw -vvv media/test.wav
Playing WAVE 'media/test.wav' : Signed 16 bit Little Endian, Rate 44100 Hz, 
Stereo
Plug PCM: Hardware PCM card 0 'HDA NVidia' device 0 subdevice 0
Its setup is:
  stream       : PLAYBACK
  access       : RW_INTERLEAVED
  format       : S16_LE
  subformat    : STD
  channels     : 2
  rate         : 44100
  exact rate   : 44100 (44100/1)
  msbits       : 16
  buffer_size  : 16384
  period_size  : 4096
  period_time  : 92879
  tstamp_mode  : NONE
  period_step  : 1
  avail_min    : 4096
  period_event : 0
  start_threshold  : 16384
  stop_threshold   : 16384
  silence_threshold: 0
  silence_size : 0
  boundary     : 4611686018427387904
  appl_ptr     : 0
  hw_ptr       : 0
Max peak (8192 samples): 0x0000003f #                    0%
Max peak (8192 samples): 0x0000004a #                    0%
...
yeah this works every time.
so is it okay right now when the subsystem ID is the "uninitialized" one? or 
should i restart the machine so i get the "correct one" back?
The missing pin configuration initialization was already fixed by
the driver overriding it after checking PCI SSID (which is different
from the codec SSID).  So, this should be no problem.
However, the reason why the analog output doesn't work might be
different from that.  There might be something else missing, but I
don't know.
Takashi

    

Takashi Iwai

tags (0)

participants (1)