[Acer Aspire One 722] Broken support of Intel HDA in 3.5.0-16 kernel

Can you confirm this bug goes away if you boot back into 3.5.0-15?

Yes, sound works perfectly with 3.5.0-15.

I'd like to perform a kernel bisect to identify the commit which caused this regression. This will require testing 5 or so kernels.

I built a test kernel, which is available for download from:
http://people.canonical.com/~jsalisbury/lp1059523

Can you test that kernel and report back if it has this bug or not?

Thanks in advance

Bug still appears with this kernel.

I built the next test kernel, which is up to commit:
8313547880cd5adff2253d3fe106b983dd93104c

The kernel is available for download from:
http://people.canonical.com/~jsalisbury/lp1059523

Can you test that kernel and report back if it has this bug or not?

Thanks in advance

Great! Sound works again with last test kernel.

Thanks for giving me a heads up on this bug Joseph.

I think we're looking at two errors here.
The first error is:

[ 10.697947] hda_codec: cannot build controls for #0 (error -16)

Please try the attached patch, I believe it would fix this error.

The second error is the oops - even if build_controls returns an error, this shouldn't cause an OOPS. Why it does that I haven't figured out.

However, I also have a question. According to the codec configuration, you seem to have two microphone jacks. Is this really correct? (It could be that one of them is part of a 4-pin headset jack.)
If this is not correct, we should disable one of them.

After the recent patch "ALSA: hda - use both input paths on Conexant
auto parser" suddenly we can have more than one "Mic Boost", this
happened on Acer Aspire One 722. Therefore we must add the possibility
to put an index on this "Mic Boost" just as we do for the other
"Mic Boost" earlier in the same function.

BugLink: https://bugs.launchpad.net/bugs/1059523
Signed-off-by: David Henningsson <email address hidden>
---
 sound/pci/hda/patch_conexant.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sound/pci/hda/patch_conexant.c b/sound/pci/hda/patch_conexant.c
index c03d3b8..c11ad8a 100644
--- a/sound/pci/hda/patch_conexant.c
+++ b/sound/pci/hda/patch_conexant.c
@@ -4273,7 +4273,7 @@ static int cx_auto_add_boost_volume(struct hda_codec *codec, int idx,

  if (get_wcaps(codec, mux) & AC_WCAP_OUT_AMP) {
   spec->imux_info[idx].boost = mux;
- return cx_auto_add_volume(codec, label, " Boost", 0,
+ return cx_auto_add_volume(codec, label, " Boost", cidx,
        mux, HDA_OUTPUT);
  }
  return 0;
--
1.7.9.5

As for the oops, it is fixed by the following upstream commit, which we should backport:

commit 07dc59f0988cb54fd87bd373b3b27eb2401dd811
Author: Takashi Iwai <email address hidden>
Date: Mon Sep 10 09:39:31 2012 +0200

    ALSA: hda - Fix Oops at codec reset/reconfig

...this patch was sent to stable so hopefully it will reach us automatically, but it is worth tracking so that it will actually happen.

David, yes, seems it's correct. Headset microphone works through headphone jack. Netbook also has separate microphone jack, but it don't work, and I don't remember if it worked in any previous releases because i never use it.

@ZwS,

I'll build a test kernel with the patch mentioned in comment #8. I'll post a link to the kernel shortly.

As for the kerneloops, can you open a new bug for that? That will allow us to track the commit from upstream, separate from this bug.

@ZwS,

I built a test kernel with the following patch applied:
0001-ALSA-hda-fix-indices-on-boost-volume-on-Conexant.patch

The test kernel can be downloaded from:
http://people.canonical.com/~jsalisbury/lp1059523

Can you test that kernel and report back if it has this bug or not?

Thanks in advance

Please be sure to install both the linux-image and linux-image-extra .deb packages:

linux-image-3.5.0-16-generic_3.5.0-16.25~lp1059523patched_amd64.deb
linux-image-extra-3.5.0-16-generic_3.5.0-16.25~lp1059523patched_amd64.deb

@ZwS,

Also, can you open a new bug for the kerneloops? That will allow us to track the commit from upstream, separate from this bug.

Joseph, unfortunately I have only i386 system, so I can't test kernel you builded.
And here is bugreport about kerneloops https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1060330

@ZwS,

Thanks for opening the new bug report for the kerneloops. Sorry, I'll build you a 32bit kernel now and post it shortly.

@ZwS,

I built a 32 bit Quantal test kernel with the following patch applied:
0001-ALSA-hda-fix-indices-on-boost-volume-on-Conexant.patch

The test kernel can be downloaded from:
http://people.canonical.com/~jsalisbury/lp1059523

Can you test that kernel and report back if it has this bug or not?

Thanks in advance

Joseph, bug is missing with the patched kernel.

Thanx Joseph! That patched kernel for 32bit restore sound in my Acer Aspire One 722.
I removed kernel linux-headers-3.5.0-16 and install Yours.
 Really good work!

Regards,
Kuba

@jsalisbury, do you need more testing for this bug? I have an AO722-0828 on amd64 which appears to be exhibiting this same problem (see http://www.alsa-project.org/db/?f=656cead6d2465a600aad164f140dae2303118882); wasn't sure whether I should use the kernel in the 64bit-patched directory of your link to do so.

This bug was fixed in the package linux - 3.5.0-17.26

---------------
linux (3.5.0-17.26) quantal-proposed; urgency=low

  [ David Henningsson ]

  * SAUCE: ALSA: hda - fix indices on boost volume on Conexant
    - LP: #1059523

  [ Leann Ogasawara ]

  * [Config] Build in CONFIG_X86_PCC_CPUFREQ=y
    - LP: #1061126

  [ Rob Herring ]

  * Revert "SAUCE: ahci_platform: add custom hard reset for Calxeda ahci
    ctrlr"
    - LP: #1059432
  * Revert "SAUCE: net: calxedaxgmac: add write barriers around setting
    owner bit"
    - LP: #1059432
  * Revert "SAUCE: ARM: highbank: use writel_relaxed variant for pwr
    requests"
    - LP: #1059432
  * Revert "SAUCE: force DMA buffers to non-bufferable on highbank"
    - LP: #1059432
  * Revert "SAUCE: input: add a key driver for highbank"
    - LP: #1059432
  * [Config] Align highbank config with amd64-generic and add new configs
    - LP: #1059432
  * SAUCE: net: calxedaxgmac: enable operate on 2nd frame mode
    - LP: #1059432
  * SAUCE: net: calxedaxgmac: remove explicit rx dma buffer polling
    - LP: #1059432
  * SAUCE: net: calxedaxgmac: use relaxed i/o accessors in rx and tx paths
    - LP: #1059432
  * SAUCE: net: calxedaxgmac: drop some unnecessary register writes
    - LP: #1059432
  * SAUCE: net: calxedaxgmac: rework transmit ring handling
    - LP: #1059432
  * SAUCE: ARM: highbank: retry wfi on reset request
    - LP: #1059432

  [ Tim Gardner ]

  * [Config] TIDSPBRIDGE=m
    - LP: #1058022
  * rebase to v3.5.5

  [ Upstream Kernel Changes ]

  * common: DMA-mapping: add DMA_ATTR_NO_KERNEL_MAPPING attribute
    - LP: #1059432
  * common: DMA-mapping: add DMA_ATTR_SKIP_CPU_SYNC attribute
    - LP: #1059432
  * ARM: dma-mapping: add support for DMA_ATTR_SKIP_CPU_SYNC attribute
    - LP: #1059432
  * ARM: add coherent dma ops
    - LP: #1059432
  * ARM: add coherent iommu dma ops
    - LP: #1059432
  * ARM: highbank: add coherent DMA setup
    - LP: #1059432
  * sata: add platform driver for Calxeda AHCI controller
    - LP: #1059432
  * xhci: Rate-limit XHCI_TRUST_TX_LENGTH quirk warning.
    - LP: #1039478
  * agp/intel-gtt: remove dead code
    - LP: #1011440
  * drm/i915: stop using dev->agp->base
    - LP: #1011440
  * agp/intel-gtt: don't require the agp bridge on setup
    - LP: #1011440
  * drm/i915 + agp/intel-gtt: prep work for direct setup
    - LP: #1011440
  * agp/intel-gtt: move gart base addres setup
    - LP: #1011440
  * drm/i915: don't use dev->agp
    - LP: #1011440
  * drm/i915: disable drm agp support for !gen3 with kms enabled
    - LP: #1011440
  * agp/intel-agp: remove snb+ host bridge pciids
    - LP: #1011440
  * rebase to v3.5.5
    - LP: #1000424
 -- Leann Ogasawara <email address hidden> Wed, 03 Oct 2012 11:11:18 -0700

@Matt Behrens,

I don't think any additional testing is required. If you want, you can confirm this bug doesn't exist on your hardware with the latest Quantal kernel.

At Tue, 2 Oct 2012 10:14:23 +0200,
David Henningsson wrote:
>
> After the recent patch "ALSA: hda - use both input paths on Conexant
> auto parser" suddenly we can have more than one "Mic Boost", this
> happened on Acer Aspire One 722. Therefore we must add the possibility
> to put an index on this "Mic Boost" just as we do for the other
> "Mic Boost" earlier in the same function.
>
> BugLink: https://bugs.launchpad.net/bugs/1059523
> Signed-off-by: David Henningsson <email address hidden>

Applied, thanks.

Takashi

> ---
> sound/pci/hda/patch_conexant.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/sound/pci/hda/patch_conexant.c b/sound/pci/hda/patch_conexant.c
> index c03d3b8..c11ad8a 100644
> --- a/sound/pci/hda/patch_conexant.c
> +++ b/sound/pci/hda/patch_conexant.c
> @@ -4273,7 +4273,7 @@ static int cx_auto_add_boost_volume(struct hda_codec *codec, int idx,
>
> if (get_wcaps(codec, mux) & AC_WCAP_OUT_AMP) {
> spec->imux_info[idx].boost = mux;
> - return cx_auto_add_volume(codec, label, " Boost", 0,
> + return cx_auto_add_volume(codec, label, " Boost", cidx,
> mux, HDA_OUTPUT);
> }
> return 0;
> --
> 1.7.9.5
>

The verification of this Stable Release Update has completed successfully and the package has now been released to -updates. Subsequently, the Ubuntu Stable Release Updates Team is being unsubscribed and will not receive messages about this bug report. In the event that you encounter a regression using the package from -updates please report a new bug using ubuntu-bug and tag the bug report regression-update so we can easily find any regresssions.