[alsa-devel] Acer Ferrari 5000

[Takashi Iwai]

At Mon, 23 Mar 2009 08:24:45 -0700,
Russ Dill wrote:
> 
> On Mon, Mar 23, 2009 at 7:57 AM, Takashi Iwai <tiwai at suse.de> wrote:
> > At Mon, 23 Mar 2009 07:42:25 -0700,
> > Russ Dill wrote:
> >>
> >> [1  <text/plain; UTF-8 (quoted-printable)>]
> >> On Mon, Mar 23, 2009 at 6:40 AM, Takashi Iwai <tiwai at suse.de> wrote:
> >> > At Mon, 23 Mar 2009 06:36:06 -0700,
> >> > Russ Dill wrote:
> >> >>
> >> >> On Mon, Mar 23, 2009 at 5:51 AM, Takashi Iwai <tiwai at suse.de> wrote:
> >> >> > At Mon, 23 Mar 2009 04:54:34 -0700,
> >> >> > Russ Dill wrote:
> >> >> >>
> >> >> >> On Mon, Mar 23, 2009 at 3:54 AM, Takashi Iwai <tiwai at suse.de> wrote:
> >> >> >> > At Sat, 21 Mar 2009 22:23:15 -0700,
> >> >> >> > Russ Dill wrote:
> >> >> >> >>
> >> >> >> >> GIven the recent changes in gnome land (pulseaudio), I'd like to be
> >> >> >> >> able to get alsa support for my hardware fixed up. My biggest
> >> >> >> >> annoyance right now is that when I plug headphones in, it doesn't mute
> >> >> >> >> the speakers. Another issue I have is that I'd like to have support
> >> >> >> >> for audio out on HDMI (which is supported by Vista). On with the show.
> >> >> >> >
> >> >> >> > Try the later version of alsa-driver or kernel, and pass model=auto.
> >> >> >> > As default model=acer is chosen for ALC883 with Acer vendor SSID,
> >> >> >> > and it's known that it doesn't match with the recent Acer laptops
> >> >> >> > at all.  The BIOS auto-parsing mode would work better recently.
> >> >> >>
> >> >> >> I still get dmesg errors (audio does play though):
> >> >> >>
> >> >> >> [229513.339812] hda-intel: Invalid position buffer, using LPIB read
> >> >> >> method instead.
> >> >> >> [229513.436583] hda-intel: IRQ timing workaround is activated for card
> >> >> >> #0. Suggest a bigger bdl_pos_adj.
> >> >> >
> >> >> > These are no errors.  You can ignore for now.
> >> >>
> >> >> ok, but the "azx_get_response timeout" ones seem to be as audio
> >> >> glitches when the occur.
> >> >
> >> > Does it happen with model=auto?  The driver doesn't work without that
> >> > option, as I mentioned.
> >> >
> >> >> > Which alsa-driver are you using now?  At best, try the latest
> >> >> > alsa-driver snapshot from
> >> >> >    ftp://ftp.kernel.org/pub/linux/kernel/people/tiwai/snapshot/alsa-driver-snapshot.tar.gz
> >> >> > or the sound git tree at
> >> >> >    git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6.git
> >> >>
> >> >> I'm currently using whatever comes in 2.6.28 with ubuntu. I'll try out
> >> >> the git tree.
> >> >
> >> > Bah, that's way too old to debug...
> >> >
> >> >> >> The Center mixer control now controls my left laptop speaker and the
> >> >> >> LFE mixer control controls the right speaker. The Surround control
> >> >> >> that used to control the laptop speakers now does nothing.
> >> >> >>
> >> >> >> Attempting to MUTE the LFE channel, the audio skipped a bit and I got:
> >> >> >>
> >> >> >> [229698.556044] hda_intel: azx_get_response timeout, switching to
> >> >> >> polling mode: last cmd=0x10db0001
> >> >> >> [229699.561017] hda_intel: azx_get_response timeout, switching to
> >> >> >> single_cmd mode: last cmd=0x10db0001
> >> >> >>
> >> >> >> My system just hung (likely due to wireless drivers) and I rebooted
> >> >> >> and with model=auto again, the mixer controls are different. There is
> >> >> >> no LFE or Center controls, but there is an added Headphone volume
> >> >> >> control (which does nothing). Front controls the headphone volume,
> >> >> >> nothing controls the speaker volume.
> >> >> >>
> >> >> >> I've been unable to hear digital output (mplayer -ao alsa:device=iec958).
> >> >> >
> >> >> > The device is likely not "iec958" but "hdmi".
> >> >>
> >> >> no such device exists with model=auto (or without setting model).
> >> >
> >> > Yes, too old version :)
> >> >
> >> >> >> My laptop has an internal microphone, two internal speakers, a
> >> >> >> headphone jack, a mic in jack, a line in jack, and digital audio out
> >> >> >> via hdmi.
> >> >> >
> >> >> > And alsa-info output?  Use model=auto from now on.
> >> >> > The default model (acer) doesn't work obviously for your laptop.
> >> >>
> >> >> It was attached to the previous email
> >> >
> >> > Ah, overlooked.  Thanks.
> >> >
> >> > But better the one from the latest driver...
> >> >
> >> >
> >> > Takashi
> >> >
> >>
> >> OK, I've bolted your master branch in your git tree onto my kernel:
> >>
> >> [10580.958816] HDA Intel 0000:00:14.2: PCI INT A -> GSI 16 (level,
> >> low) -> IRQ 16
> >> [10581.101270] input: HDA Digital PCBeep as
> >> /devices/pci0000:00/0000:00:14.2/input/input8
> >>
> >> My mixer controls are very similar to what I got before, except now
> >> Front controls both the front speakers, and the headphones. Plugging
> >> in and unplugging the headphone jack still has no effect.
> >
> > See below.
> >
> >> There is also a "Beep" control now.
> >
> > It's a new feature in the new version.
> >
> >> There is still no PCM device labled hdmi, and outputting to iec958
> >> device doesn't get me any audio output via HDMI.
> >
> > Right, it's apparently not set by BIOS, so the driver can't detect
> > it.
> >
> > How is the HDMI connected?  Is it from the graphic chip/board
> > or any other route?  In the latter case, an HDMI transmitter (codec)
> > is connected to the main audio controller, typically in the 4th slot
> > (codec#3).  In the former case, it usually appears as an individual
> > PCI device.
> 
> There are no other sound related PCI devices, and my graphics chip is
> an R500 which doesn't support audio, so it must be coming from a
> digital interface of SB450.
> 
> > If the SPDIF digital device is supposed to be HDMI, then it's a bug of
> > BIOS.  In that case, using "iec958" is the right choice.
> 
> But I haven't been able to get it working under Linux.

Then something is still missing...

> > Anyway, you could build the driver with --with-debug=verbose configure
> > option and try to pass probe_mask=0x1ff.  This will force to probe
> > all codec slots.  See kernel messages.
> 
> I must be building differently, I'll translate that to
> CONFIG_SND_DEBUG=y and CONFIG_SND_DEBUG_VEBOSE=y. Where is the
> configure script?

It was about alsa-driver tarball.  If you build directly from
sound.git tree (I suppose right?), these two kconfigs correspond.


> [13329.546638] HDA Intel 0000:00:14.2: PCI INT A -> GSI 16 (level,
> low) -> IRQ 16
> [13329.546650] hda_intel: codec_mask forced to 0xff
> [13330.581024] hda_intel: azx_get_response timeout, switching to
> polling mode: last cmd=0x200f0000
> [13331.585015] hda_intel: Codec #2 probe error; disabling it...
> [13332.620024] hda_intel: Codec #3 probe error; disabling it...
> [13332.756133] input: HDA Digital PCBeep as
> /devices/pci0000:00/0000:00:14.2/input/input9

So, it's not in the codec slot#3, too.


> >> I can still get:
> >>
> >>
> >> [11025.533020] hda_intel: azx_get_response timeout, switching to
> >> polling mode: last cmd=0x106f000a
> >> [11026.537063] hda_intel: azx_get_response timeout, switching to
> >> single_cmd mode: last cmd=0x106f000a
> >
> > Hm.  Judging from alsa-info output, the verb looks correct, it's PCM
> > parameter inquiry against the widget 0x06.  So, your hardware is
> > really flaky.  What happens if you add
> >        codec->bus->needs_damn_long_delay = 1;
> > in patch_realtek.c:patch_alc883() ?
> 
> Doesn't seem to have an effect, its easy to reproduce my switching a
> control while music is playing.
> 
> [13932.100052] hda_intel: azx_get_response timeout, switching to
> single_cmd mode: last cmd=0x10db2001

Hmm.  What about setting enable_msi=1?


> >> [11027.813799] hda-intel: IRQ timing workaround is activated for card
> >> #0. Suggest a bigger bdl_pos_adj.
> >
> > This should be the side-effect of the codec communication error above.
> >
> >> while playing audio.
> >
> >> I've attached the associated alsa-info.txt
> > ...
> >>       control.4 {
> >>               comment.access 'read write'
> >>               comment.type BOOLEAN
> >>               comment.count 2
> >>               iface MIXER
> >>               name 'Headphone Playback Switch'
> >>               value.0 false
> >>               value.1 false
> >
> > Unmute "Headphone" switch.
> 
> I've tried, hotplugging the headphones has no effect whether or not
> this switch is on or off.

I guess the communication with the codec gets already screwed up, thus
it couldn't be handled properly any more.

I'm not sure what is the reason, but what I can say is that ATI
controller is pretty unstable regarding HD-audio.


Takashi