Thinkpad hda regression in 3.3

Thinkpad hda regression in 3.3

[Josh Boyer]

Hi Takashi and Jaroslav,

We've had a number of bugs reported about sound not working correctly on
various models of Thinkpad machines in the 3.3 kernel.  I've listed them
below.

I looked through the commits and found:

    commit 29c5fbbcfefba5225a6783683c46c39e10877703:

    ALSA: HDA: Use model=auto for Thinkpad T510

    The user reports that model=auto works fine for him. Using
    model=auto bring in new features such as jack detection notification
    to userspace.

I did a local build here with that comment reverted and so far most of
the people reported their sound to be working again.  The trivial revert
patch is below in case you want it.

It seems there are multiple models that have PCI ID 17aa:215e and using
model=auto does not work for them all.  It particularly messes with
machines using a docking station for some reason.

Aside from just reverting the above commit, do you have other thoughts
on how to fix this?

josh

https://bugzilla.redhat.com/show_bug.cgi?id=808559
https://bugzilla.redhat.com/show_bug.cgi?id=806217
https://bugzilla.redhat.com/show_bug.cgi?id=810697

-----

From: Josh Boyer <jwboyer@redhat.com>
Subject: [PATCH] ALSA: hda - Use model=thinkpad for PCI ID 17aa:215e

Commit 29c5fb "ALSA: HDA: Use model=auto for Thinkpad T510" switched to
using model=auto for PCI ID 17aa:215e, however that breaks sound on a
number of thinkpads.  Change this back to using model=thinkpad for now.

Signed-off-by: Josh Boyer <jwboyer@redhat.com>

---

--- linux-3.3.noarch.orig/sound/pci/hda/patch_conexant.c
+++ linux-3.3.noarch/sound/pci/hda/patch_conexant.c
@@ -3027,7 +3027,7 @@ static const struct snd_pci_quirk cxt506
 	SND_PCI_QUIRK(0x17aa, 0x20f2, "Lenovo T400s", CXT5066_THINKPAD),
 	SND_PCI_QUIRK(0x17aa, 0x21c5, "Thinkpad Edge 13", CXT5066_THINKPAD),
 	SND_PCI_QUIRK(0x17aa, 0x21c6, "Thinkpad Edge 13", CXT5066_ASUS),
-	SND_PCI_QUIRK(0x17aa, 0x215e, "Lenovo T510", CXT5066_AUTO),
+	SND_PCI_QUIRK(0x17aa, 0x215e, "Lenovo T510", CXT5066_THINKPAD),
 	SND_PCI_QUIRK(0x17aa, 0x21cf, "Lenovo T520 & W520", CXT5066_AUTO),
 	SND_PCI_QUIRK(0x17aa, 0x21da, "Lenovo X220", CXT5066_THINKPAD),
 	SND_PCI_QUIRK(0x17aa, 0x21db, "Lenovo X220-tablet", CXT5066_THINKPAD),

Re: Thinkpad hda regression in 3.3

[Takashi Iwai]

At Fri, 13 Apr 2012 08:43:06 -0400,
Josh Boyer wrote:
> 
> Hi Takashi and Jaroslav,
> 
> We've had a number of bugs reported about sound not working correctly on
> various models of Thinkpad machines in the 3.3 kernel.  I've listed them
> below.
> 
> I looked through the commits and found:
> 
>     commit 29c5fbbcfefba5225a6783683c46c39e10877703:
> 
>     ALSA: HDA: Use model=auto for Thinkpad T510
> 
>     The user reports that model=auto works fine for him. Using
>     model=auto bring in new features such as jack detection notification
>     to userspace.
> 
> I did a local build here with that comment reverted and so far most of
> the people reported their sound to be working again.  The trivial revert
> patch is below in case you want it.

Could you get alsa-info.sh output on the non-working T510?
If it's really about the built-in speaker or headphone jack, I doubt
it's rahter the muted output in mixer.  I see only "me too" in the bug
report, which doesn't help much to diagnose the problem.

I've got a few positive feedbacks of T410 and T510, so basically the
driver should work.  But I know the problem of docking-station (see
below).


> It seems there are multiple models that have PCI ID 17aa:215e and using
> model=auto does not work for them all.  It particularly messes with
> machines using a docking station for some reason.
> 
> Aside from just reverting the above commit, do you have other thoughts
> on how to fix this?

The problem regarding the docking station is basically a BIOS issue.
ThinkPad BIOS doesn't expose the pins for docking stations at all.
We have already a workaround for X200, and similar others would be
needed for T410, T420, etc.

For testing, you don't need even patch the driver.  When the driver is
built with CONFIG_SND_HDA_PATCH_LOADER=y, create a file containing the
below as /lib/firmware/alsa-tp-dock:

================================================================
[codec]
0x14f15069 0x17aa21a4 0

[pincfg]
0x1a 0x21a11000
0x1c 0x23014250
================================================================

then pass the module option patch="alsa-tp-dock" to snd-hda-intel,
e.g. create a file containing below in /etc/modprobe.d/*

	options snd-hda-intel patch="alsa-tp-dock"

This will override the pin-default values of 0x1a and 0x1c, which
corresponds to the docking-station mic and docking-station output.
This will set dock-out as the line-out so that the built-in headphone
is preferred over it.

Note that after applying this, the new mixer elements will appear, so
you'd need to adjust the values manually once.

Similarly, for T420, create a file with the contents below instead:

================================================================
[codec]
0x14f1506e 0x17aa21d2 0

[pincfg]
0x19 0x0421104f
0x1a 0x21a19050
0x1c 0x2121401f
================================================================

The rest procedure is same.

If the above seem work (again, make sure that you adjust the mixer
mutes and levels), I'll make kernel patches to add these fixups
statically in the driver so that it works without extra patching.


thanks,

Takashi

Re: Thinkpad hda regression in 3.3

[Takashi Iwai]

At Fri, 13 Apr 2012 08:43:06 -0400,
Josh Boyer wrote:
> 
> Hi Takashi and Jaroslav,
> 
> We've had a number of bugs reported about sound not working correctly on
> various models of Thinkpad machines in the 3.3 kernel.  I've listed them
> below.
> 
> I looked through the commits and found:
> 
>     commit 29c5fbbcfefba5225a6783683c46c39e10877703:
> 
>     ALSA: HDA: Use model=auto for Thinkpad T510
> 
>     The user reports that model=auto works fine for him. Using
>     model=auto bring in new features such as jack detection notification
>     to userspace.
> 
> I did a local build here with that comment reverted and so far most of
> the people reported their sound to be working again.  The trivial revert
> patch is below in case you want it.

Could you get alsa-info.sh output on the non-working T510?
If it's really about the built-in speaker or headphone jack, I doubt
it's rahter the muted output in mixer.  I see only "me too" in the bug
report, which doesn't help much to diagnose the problem.

I've got a few positive feedbacks of T410 and T510, so basically the
driver should work.  But I know the problem of docking-station (see
below).


> It seems there are multiple models that have PCI ID 17aa:215e and using
> model=auto does not work for them all.  It particularly messes with
> machines using a docking station for some reason.
> 
> Aside from just reverting the above commit, do you have other thoughts
> on how to fix this?

The problem regarding the docking station is basically a BIOS issue.
ThinkPad BIOS doesn't expose the pins for docking stations at all.
We have already a workaround for X200, and similar others would be
needed for T410, T420, etc.

For testing, you don't need even patch the driver.  When the driver is
built with CONFIG_SND_HDA_PATCH_LOADER=y, create a file containing the
below as /lib/firmware/alsa-tp-dock:

================================================================
[codec]
0x14f15069 0x17aa21a4 0

[pincfg]
0x1a 0x21a11000
0x1c 0x23014250
================================================================

then pass the module option patch="alsa-tp-dock" to snd-hda-intel,
e.g. create a file containing below in /etc/modprobe.d/*

	options snd-hda-intel patch="alsa-tp-dock"

This will override the pin-default values of 0x1a and 0x1c, which
corresponds to the docking-station mic and docking-station output.
This will set dock-out as the line-out so that the built-in headphone
is preferred over it.

Note that after applying this, the new mixer elements will appear, so
you'd need to adjust the values manually once.

Similarly, for T420, create a file with the contents below instead:

================================================================
[codec]
0x14f1506e 0x17aa21d2 0

[pincfg]
0x19 0x0421104f
0x1a 0x21a19050
0x1c 0x2121401f
================================================================

The rest procedure is same.

If the above seem work (again, make sure that you adjust the mixer
mutes and levels), I'll make kernel patches to add these fixups
statically in the driver so that it works without extra patching.


thanks,

Takashi

Re: Thinkpad hda regression in 3.3

[Josh Boyer]

On Fri, Apr 13, 2012 at 03:07:07PM +0200, Takashi Iwai wrote:
> At Fri, 13 Apr 2012 08:43:06 -0400,
> Josh Boyer wrote:
> > 
> > Hi Takashi and Jaroslav,
> > 
> > We've had a number of bugs reported about sound not working correctly on
> > various models of Thinkpad machines in the 3.3 kernel.  I've listed them
> > below.
> > 
> > I looked through the commits and found:
> > 
> >     commit 29c5fbbcfefba5225a6783683c46c39e10877703:
> > 
> >     ALSA: HDA: Use model=auto for Thinkpad T510
> > 
> >     The user reports that model=auto works fine for him. Using
> >     model=auto bring in new features such as jack detection notification
> >     to userspace.
> > 
> > I did a local build here with that comment reverted and so far most of
> > the people reported their sound to be working again.  The trivial revert
> > patch is below in case you want it.
> 
> Could you get alsa-info.sh output on the non-working T510?
> If it's really about the built-in speaker or headphone jack, I doubt
> it's rahter the muted output in mixer.  I see only "me too" in the bug
> report, which doesn't help much to diagnose the problem.

OK, I've asked for that.

> I've got a few positive feedbacks of T410 and T510, so basically the
> driver should work.  But I know the problem of docking-station (see
> below).
> 
> 
> > It seems there are multiple models that have PCI ID 17aa:215e and using
> > model=auto does not work for them all.  It particularly messes with
> > machines using a docking station for some reason.
> > 
> > Aside from just reverting the above commit, do you have other thoughts
> > on how to fix this?
> 
> The problem regarding the docking station is basically a BIOS issue.
> ThinkPad BIOS doesn't expose the pins for docking stations at all.
> We have already a workaround for X200, and similar others would be
> needed for T410, T420, etc.
> 
> For testing, you don't need even patch the driver.  When the driver is
> built with CONFIG_SND_HDA_PATCH_LOADER=y, create a file containing the
> below as /lib/firmware/alsa-tp-dock:
> 
> ================================================================
> [codec]
> 0x14f15069 0x17aa21a4 0
> 
> [pincfg]
> 0x1a 0x21a11000
> 0x1c 0x23014250
> ================================================================

The above is for the X200 docking station, correct?  So it should work
with an x201 thinkpad in an x200 dock.

> then pass the module option patch="alsa-tp-dock" to snd-hda-intel,
> e.g. create a file containing below in /etc/modprobe.d/*
> 
> 	options snd-hda-intel patch="alsa-tp-dock"
> 
> This will override the pin-default values of 0x1a and 0x1c, which
> corresponds to the docking-station mic and docking-station output.
> This will set dock-out as the line-out so that the built-in headphone
> is preferred over it.
> 
> Note that after applying this, the new mixer elements will appear, so
> you'd need to adjust the values manually once.
> 
> Similarly, for T420, create a file with the contents below instead:
> 
> ================================================================
> [codec]
> 0x14f1506e 0x17aa21d2 0
> 
> [pincfg]
> 0x19 0x0421104f
> 0x1a 0x21a19050
> 0x1c 0x2121401f
> ================================================================
> 
> The rest procedure is same.
> 
> If the above seem work (again, make sure that you adjust the mixer
> mutes and levels), I'll make kernel patches to add these fixups
> statically in the driver so that it works without extra patching.

OK.  I'll have the reporters test this out.

josh

Re: Thinkpad hda regression in 3.3

[Takashi Iwai]

At Fri, 13 Apr 2012 09:23:52 -0400,
Josh Boyer wrote:
> 
> On Fri, Apr 13, 2012 at 03:07:07PM +0200, Takashi Iwai wrote:
> > At Fri, 13 Apr 2012 08:43:06 -0400,
> > Josh Boyer wrote:
> > > 
> > > Hi Takashi and Jaroslav,
> > > 
> > > We've had a number of bugs reported about sound not working correctly on
> > > various models of Thinkpad machines in the 3.3 kernel.  I've listed them
> > > below.
> > > 
> > > I looked through the commits and found:
> > > 
> > >     commit 29c5fbbcfefba5225a6783683c46c39e10877703:
> > > 
> > >     ALSA: HDA: Use model=auto for Thinkpad T510
> > > 
> > >     The user reports that model=auto works fine for him. Using
> > >     model=auto bring in new features such as jack detection notification
> > >     to userspace.
> > > 
> > > I did a local build here with that comment reverted and so far most of
> > > the people reported their sound to be working again.  The trivial revert
> > > patch is below in case you want it.
> > 
> > Could you get alsa-info.sh output on the non-working T510?
> > If it's really about the built-in speaker or headphone jack, I doubt
> > it's rahter the muted output in mixer.  I see only "me too" in the bug
> > report, which doesn't help much to diagnose the problem.
> 
> OK, I've asked for that.
> 
> > I've got a few positive feedbacks of T410 and T510, so basically the
> > driver should work.  But I know the problem of docking-station (see
> > below).
> > 
> > 
> > > It seems there are multiple models that have PCI ID 17aa:215e and using
> > > model=auto does not work for them all.  It particularly messes with
> > > machines using a docking station for some reason.
> > > 
> > > Aside from just reverting the above commit, do you have other thoughts
> > > on how to fix this?
> > 
> > The problem regarding the docking station is basically a BIOS issue.
> > ThinkPad BIOS doesn't expose the pins for docking stations at all.
> > We have already a workaround for X200, and similar others would be
> > needed for T410, T420, etc.
> > 
> > For testing, you don't need even patch the driver.  When the driver is
> > built with CONFIG_SND_HDA_PATCH_LOADER=y, create a file containing the
> > below as /lib/firmware/alsa-tp-dock:
> > 
> > ================================================================
> > [codec]
> > 0x14f15069 0x17aa21a4 0
> > 
> > [pincfg]
> > 0x1a 0x21a11000
> > 0x1c 0x23014250
> > ================================================================
> 
> The above is for the X200 docking station, correct?  So it should work
> with an x201 thinkpad in an x200 dock.

It's for T410/T510 dock.  But X201 uses the same codec chip, so this
could be used, too.  Just give it a try.


Takashi

Re: Thinkpad hda regression in 3.3

[Takashi Iwai]

At Fri, 13 Apr 2012 09:23:52 -0400,
Josh Boyer wrote:
> 
> On Fri, Apr 13, 2012 at 03:07:07PM +0200, Takashi Iwai wrote:
> > At Fri, 13 Apr 2012 08:43:06 -0400,
> > Josh Boyer wrote:
> > > 
> > > Hi Takashi and Jaroslav,
> > > 
> > > We've had a number of bugs reported about sound not working correctly on
> > > various models of Thinkpad machines in the 3.3 kernel.  I've listed them
> > > below.
> > > 
> > > I looked through the commits and found:
> > > 
> > >     commit 29c5fbbcfefba5225a6783683c46c39e10877703:
> > > 
> > >     ALSA: HDA: Use model=auto for Thinkpad T510
> > > 
> > >     The user reports that model=auto works fine for him. Using
> > >     model=auto bring in new features such as jack detection notification
> > >     to userspace.
> > > 
> > > I did a local build here with that comment reverted and so far most of
> > > the people reported their sound to be working again.  The trivial revert
> > > patch is below in case you want it.
> > 
> > Could you get alsa-info.sh output on the non-working T510?
> > If it's really about the built-in speaker or headphone jack, I doubt
> > it's rahter the muted output in mixer.  I see only "me too" in the bug
> > report, which doesn't help much to diagnose the problem.
> 
> OK, I've asked for that.
> 
> > I've got a few positive feedbacks of T410 and T510, so basically the
> > driver should work.  But I know the problem of docking-station (see
> > below).
> > 
> > 
> > > It seems there are multiple models that have PCI ID 17aa:215e and using
> > > model=auto does not work for them all.  It particularly messes with
> > > machines using a docking station for some reason.
> > > 
> > > Aside from just reverting the above commit, do you have other thoughts
> > > on how to fix this?
> > 
> > The problem regarding the docking station is basically a BIOS issue.
> > ThinkPad BIOS doesn't expose the pins for docking stations at all.
> > We have already a workaround for X200, and similar others would be
> > needed for T410, T420, etc.
> > 
> > For testing, you don't need even patch the driver.  When the driver is
> > built with CONFIG_SND_HDA_PATCH_LOADER=y, create a file containing the
> > below as /lib/firmware/alsa-tp-dock:
> > 
> > ================================================================
> > [codec]
> > 0x14f15069 0x17aa21a4 0
> > 
> > [pincfg]
> > 0x1a 0x21a11000
> > 0x1c 0x23014250
> > ================================================================
> 
> The above is for the X200 docking station, correct?  So it should work
> with an x201 thinkpad in an x200 dock.

It's for T410/T510 dock.  But X201 uses the same codec chip, so this
could be used, too.  Just give it a try.


Takashi

Re: Thinkpad hda regression in 3.3

[Josh Boyer]

On Fri, Apr 13, 2012 at 03:29:41PM +0200, Takashi Iwai wrote:
> At Fri, 13 Apr 2012 09:23:52 -0400,
> Josh Boyer wrote:
> > > The problem regarding the docking station is basically a BIOS issue.
> > > ThinkPad BIOS doesn't expose the pins for docking stations at all.
> > > We have already a workaround for X200, and similar others would be
> > > needed for T410, T420, etc.
> > > 
> > > For testing, you don't need even patch the driver.  When the driver is
> > > built with CONFIG_SND_HDA_PATCH_LOADER=y, create a file containing the
> > > below as /lib/firmware/alsa-tp-dock:
> > > 
> > > ================================================================
> > > [codec]
> > > 0x14f15069 0x17aa21a4 0
> > > 
> > > [pincfg]
> > > 0x1a 0x21a11000
> > > 0x1c 0x23014250
> > > ================================================================
> > 
> > The above is for the X200 docking station, correct?  So it should work
> > with an x201 thinkpad in an x200 dock.
> 
> It's for T410/T510 dock.  But X201 uses the same codec chip, so this
> could be used, too.  Just give it a try.

OK.  I've asked a reporter to test this out.  We'll get back to you.

josh

Re: Thinkpad hda regression in 3.3

[Takashi Iwai]

At Fri, 13 Apr 2012 09:37:21 -0400,
Josh Boyer wrote:
> 
> On Fri, Apr 13, 2012 at 03:29:41PM +0200, Takashi Iwai wrote:
> > At Fri, 13 Apr 2012 09:23:52 -0400,
> > Josh Boyer wrote:
> > > > The problem regarding the docking station is basically a BIOS issue.
> > > > ThinkPad BIOS doesn't expose the pins for docking stations at all.
> > > > We have already a workaround for X200, and similar others would be
> > > > needed for T410, T420, etc.
> > > > 
> > > > For testing, you don't need even patch the driver.  When the driver is
> > > > built with CONFIG_SND_HDA_PATCH_LOADER=y, create a file containing the
> > > > below as /lib/firmware/alsa-tp-dock:
> > > > 
> > > > ================================================================
> > > > [codec]
> > > > 0x14f15069 0x17aa21a4 0
> > > > 
> > > > [pincfg]
> > > > 0x1a 0x21a11000
> > > > 0x1c 0x23014250
> > > > ================================================================
> > > 
> > > The above is for the X200 docking station, correct?  So it should work
> > > with an x201 thinkpad in an x200 dock.
> > 
> > It's for T410/T510 dock.  But X201 uses the same codec chip, so this
> > could be used, too.  Just give it a try.
> 
> OK.  I've asked a reporter to test this out.  We'll get back to you.

Thanks.  As mentionted, the difference between two setups (sent in two
posts) is seen in the mixer setup.  I personally don't mind which
setup is taken; it's rather a choice of users.  So, let me know which
setup is preferred by testers.

The reason behind such multiple setups is that the codec chip has only
two DACs while you want to access for three or more outputs (two
headphones, one speaker, etc).  Thus some outputs must share the same
DAC.


Takashi

Re: Thinkpad hda regression in 3.3

[Takashi Iwai]

At Fri, 13 Apr 2012 09:37:21 -0400,
Josh Boyer wrote:
> 
> On Fri, Apr 13, 2012 at 03:29:41PM +0200, Takashi Iwai wrote:
> > At Fri, 13 Apr 2012 09:23:52 -0400,
> > Josh Boyer wrote:
> > > > The problem regarding the docking station is basically a BIOS issue.
> > > > ThinkPad BIOS doesn't expose the pins for docking stations at all.
> > > > We have already a workaround for X200, and similar others would be
> > > > needed for T410, T420, etc.
> > > > 
> > > > For testing, you don't need even patch the driver.  When the driver is
> > > > built with CONFIG_SND_HDA_PATCH_LOADER=y, create a file containing the
> > > > below as /lib/firmware/alsa-tp-dock:
> > > > 
> > > > ================================================================
> > > > [codec]
> > > > 0x14f15069 0x17aa21a4 0
> > > > 
> > > > [pincfg]
> > > > 0x1a 0x21a11000
> > > > 0x1c 0x23014250
> > > > ================================================================
> > > 
> > > The above is for the X200 docking station, correct?  So it should work
> > > with an x201 thinkpad in an x200 dock.
> > 
> > It's for T410/T510 dock.  But X201 uses the same codec chip, so this
> > could be used, too.  Just give it a try.
> 
> OK.  I've asked a reporter to test this out.  We'll get back to you.

Thanks.  As mentionted, the difference between two setups (sent in two
posts) is seen in the mixer setup.  I personally don't mind which
setup is taken; it's rather a choice of users.  So, let me know which
setup is preferred by testers.

The reason behind such multiple setups is that the codec chip has only
two DACs while you want to access for three or more outputs (two
headphones, one speaker, etc).  Thus some outputs must share the same
DAC.


Takashi

Re: Thinkpad hda regression in 3.3

[Josh Boyer]

On Fri, Apr 13, 2012 at 03:41:53PM +0200, Takashi Iwai wrote:
> At Fri, 13 Apr 2012 09:37:21 -0400,
> Josh Boyer wrote:
> Thanks.  As mentionted, the difference between two setups (sent in two
> posts) is seen in the mixer setup.  I personally don't mind which
> setup is taken; it's rather a choice of users.  So, let me know which
> setup is preferred by testers.

You're referring to the inclusion of pin 0x19 or not here, right?

> The reason behind such multiple setups is that the codec chip has only
> two DACs while you want to access for three or more outputs (two
> headphones, one speaker, etc).  Thus some outputs must share the same
> DAC.

I'm guessing when someone undocks the pins are reconfigured back to the
jacks on the laptop itself somehow?

I must admit, ALSA and it's multi-codec hda driver with lots of pin
swizzling is somewhat intimidating at the moment.  If I ask stupid
questions, please bear with me :).

josh

Re: Thinkpad hda regression in 3.3

[Takashi Iwai]

At Fri, 13 Apr 2012 10:05:49 -0400,
Josh Boyer wrote:
> 
> On Fri, Apr 13, 2012 at 03:41:53PM +0200, Takashi Iwai wrote:
> > At Fri, 13 Apr 2012 09:37:21 -0400,
> > Josh Boyer wrote:
> > Thanks.  As mentionted, the difference between two setups (sent in two
> > posts) is seen in the mixer setup.  I personally don't mind which
> > setup is taken; it's rather a choice of users.  So, let me know which
> > setup is preferred by testers.
> 
> You're referring to the inclusion of pin 0x19 or not here, right?

Yes.

> > The reason behind such multiple setups is that the codec chip has only
> > two DACs while you want to access for three or more outputs (two
> > headphones, one speaker, etc).  Thus some outputs must share the same
> > DAC.
> 
> I'm guessing when someone undocks the pins are reconfigured back to the
> jacks on the laptop itself somehow?

No, the configuration must be determined at the driver initialization
time.  So, the mixer elements stay unchanged with or without dock.
(For example, if the mixer elements were to change upon the
 dock/undock, this would be really confusing both for users and apps.
 In theory, it's possible to implement like that, but no one wants
 it.)

The jack-detection on the docking station works, yes, i.e. the
auto-muting or the mic changing works.  But it's different from
changing the configuration on the hotplug of docking station itself.


Takashi

Re: Thinkpad hda regression in 3.3

[Takashi Iwai]

At Fri, 13 Apr 2012 10:05:49 -0400,
Josh Boyer wrote:
> 
> On Fri, Apr 13, 2012 at 03:41:53PM +0200, Takashi Iwai wrote:
> > At Fri, 13 Apr 2012 09:37:21 -0400,
> > Josh Boyer wrote:
> > Thanks.  As mentionted, the difference between two setups (sent in two
> > posts) is seen in the mixer setup.  I personally don't mind which
> > setup is taken; it's rather a choice of users.  So, let me know which
> > setup is preferred by testers.
> 
> You're referring to the inclusion of pin 0x19 or not here, right?

Yes.

> > The reason behind such multiple setups is that the codec chip has only
> > two DACs while you want to access for three or more outputs (two
> > headphones, one speaker, etc).  Thus some outputs must share the same
> > DAC.
> 
> I'm guessing when someone undocks the pins are reconfigured back to the
> jacks on the laptop itself somehow?

No, the configuration must be determined at the driver initialization
time.  So, the mixer elements stay unchanged with or without dock.
(For example, if the mixer elements were to change upon the
 dock/undock, this would be really confusing both for users and apps.
 In theory, it's possible to implement like that, but no one wants
 it.)

The jack-detection on the docking station works, yes, i.e. the
auto-muting or the mic changing works.  But it's different from
changing the configuration on the hotplug of docking station itself.


Takashi

Re: Thinkpad hda regression in 3.3

[Takashi Iwai]

At Fri, 13 Apr 2012 15:07:07 +0200,
Takashi Iwai wrote:
> 
> For testing, you don't need even patch the driver.  When the driver is
> built with CONFIG_SND_HDA_PATCH_LOADER=y, create a file containing the
> below as /lib/firmware/alsa-tp-dock:
> 
> ================================================================
> [codec]
> 0x14f15069 0x17aa21a4 0
> 
> [pincfg]
> 0x1a 0x21a11000
> 0x1c 0x23014250
> ================================================================

Forgot to mention that the above is only for TP410/510.
For TP420 & co, the second one should be used.

Also, for TP410/510, the following values might be better (this time
contains 0x19, too):

0x19 0x042110ff
0x1a 0x21a190f0
0x1c 0x212140ff

This will make the built-in and the dock headphones as headphones.
The mixer elements will be more intuitive than the former version.


thanks,

Takashi

Re: Thinkpad hda regression in 3.3

[Takashi Iwai]

At Fri, 13 Apr 2012 15:07:07 +0200,
Takashi Iwai wrote:
> 
> For testing, you don't need even patch the driver.  When the driver is
> built with CONFIG_SND_HDA_PATCH_LOADER=y, create a file containing the
> below as /lib/firmware/alsa-tp-dock:
> 
> ================================================================
> [codec]
> 0x14f15069 0x17aa21a4 0
> 
> [pincfg]
> 0x1a 0x21a11000
> 0x1c 0x23014250
> ================================================================

Forgot to mention that the above is only for TP410/510.
For TP420 & co, the second one should be used.

Also, for TP410/510, the following values might be better (this time
contains 0x19, too):

0x19 0x042110ff
0x1a 0x21a190f0
0x1c 0x212140ff

This will make the built-in and the dock headphones as headphones.
The mixer elements will be more intuitive than the former version.


thanks,

Takashi

Re: Thinkpad hda regression in 3.3

[Takashi Iwai]

At Fri, 13 Apr 2012 15:07:07 +0200,
Takashi Iwai wrote:
(snip)
> If the above seem work (again, make sure that you adjust the mixer
> mutes and levels), I'll make kernel patches to add these fixups
> statically in the driver so that it works without extra patching.

Looking through alsa-info.sh outputs from various ThinkPad models I've
gathered, it seems that we can use the same pin setups for all.

Below is a test patch for 3.3 kernel.  (Note that it'd give a merge
conflict for 3.4 kernel.)


thanks,

Takashi

---
diff --git a/sound/pci/hda/patch_conexant.c b/sound/pci/hda/patch_conexant.c
index d29d6d3..681a8a9 100644
--- a/sound/pci/hda/patch_conexant.c
+++ b/sound/pci/hda/patch_conexant.c
@@ -4362,8 +4362,10 @@ static void apply_pin_fixup(struct hda_codec *codec,
 
 enum {
 	CXT_PINCFG_LENOVO_X200,
+	CXT_PINCFG_LENOVO_TP410,
 };
 
+/* ThinkPad X200 & co with cxt5051 */
 static const struct cxt_pincfg cxt_pincfg_lenovo_x200[] = {
 	{ 0x16, 0x042140ff }, /* HP (seq# overridden) */
 	{ 0x17, 0x21a11000 }, /* dock-mic */
@@ -4371,15 +4373,33 @@ static const struct cxt_pincfg cxt_pincfg_lenovo_x200[] = {
 	{}
 };
 
+/* ThinkPad 410/420/510/520, X201 & co with cxt5066 */
+static const struct cxt_pincfg cxt_pincfg_lenovo_tp410[] = {
+	{ 0x19, 0x042110ff }, /* HP (seq# overridden) */
+	{ 0x1a, 0x21a190f0 }, /* dock-mic */
+	{ 0x1c, 0x212140ff }, /* dock-HP */
+	{}
+};
+
 static const struct cxt_pincfg *cxt_pincfg_tbl[] = {
 	[CXT_PINCFG_LENOVO_X200] = cxt_pincfg_lenovo_x200,
+	[CXT_PINCFG_LENOVO_TP410] = cxt_pincfg_lenovo_tp410,
 };
 
-static const struct snd_pci_quirk cxt_fixups[] = {
+static const struct snd_pci_quirk cxt5051_fixups[] = {
 	SND_PCI_QUIRK(0x17aa, 0x20f2, "Lenovo X200", CXT_PINCFG_LENOVO_X200),
 	{}
 };
 
+static const struct snd_pci_quirk cxt5066_fixups[] = {
+	SND_PCI_QUIRK(0x17aa, 0x20f2, "Lenovo T400", CXT_PINCFG_LENOVO_TP410),
+	SND_PCI_QUIRK(0x17aa, 0x215e, "Lenovo T410", CXT_PINCFG_LENOVO_TP410),
+	SND_PCI_QUIRK(0x17aa, 0x215f, "Lenovo T510", CXT_PINCFG_LENOVO_TP410),
+	SND_PCI_QUIRK(0x17aa, 0x21ce, "Lenovo T420", CXT_PINCFG_LENOVO_TP410),
+	SND_PCI_QUIRK(0x17aa, 0x21cf, "Lenovo T520", CXT_PINCFG_LENOVO_TP410),
+	{}
+};
+
 /* add "fake" mute amp-caps to DACs on cx5051 so that mixer mute switches
  * can be created (bko#42825)
  */
@@ -4416,11 +4436,13 @@ static int patch_conexant_auto(struct hda_codec *codec)
 		break;
 	case 0x14f15051:
 		add_cx5051_fake_mutes(codec);
+		apply_pin_fixup(codec, cxt5051_fixups, cxt_pincfg_tbl);
+		break;
+	default:
+		apply_pin_fixup(codec, cxt5066_fixups, cxt_pincfg_tbl);
 		break;
 	}
 
-	apply_pin_fixup(codec, cxt_fixups, cxt_pincfg_tbl);
-
 	err = cx_auto_search_adcs(codec);
 	if (err < 0)
 		return err;

Re: Thinkpad hda regression in 3.3

[Takashi Iwai]

At Fri, 13 Apr 2012 15:07:07 +0200,
Takashi Iwai wrote:
(snip)
> If the above seem work (again, make sure that you adjust the mixer
> mutes and levels), I'll make kernel patches to add these fixups
> statically in the driver so that it works without extra patching.

Looking through alsa-info.sh outputs from various ThinkPad models I've
gathered, it seems that we can use the same pin setups for all.

Below is a test patch for 3.3 kernel.  (Note that it'd give a merge
conflict for 3.4 kernel.)


thanks,

Takashi

---
diff --git a/sound/pci/hda/patch_conexant.c b/sound/pci/hda/patch_conexant.c
index d29d6d3..681a8a9 100644
--- a/sound/pci/hda/patch_conexant.c
+++ b/sound/pci/hda/patch_conexant.c
@@ -4362,8 +4362,10 @@ static void apply_pin_fixup(struct hda_codec *codec,
 
 enum {
 	CXT_PINCFG_LENOVO_X200,
+	CXT_PINCFG_LENOVO_TP410,
 };
 
+/* ThinkPad X200 & co with cxt5051 */
 static const struct cxt_pincfg cxt_pincfg_lenovo_x200[] = {
 	{ 0x16, 0x042140ff }, /* HP (seq# overridden) */
 	{ 0x17, 0x21a11000 }, /* dock-mic */
@@ -4371,15 +4373,33 @@ static const struct cxt_pincfg cxt_pincfg_lenovo_x200[] = {
 	{}
 };
 
+/* ThinkPad 410/420/510/520, X201 & co with cxt5066 */
+static const struct cxt_pincfg cxt_pincfg_lenovo_tp410[] = {
+	{ 0x19, 0x042110ff }, /* HP (seq# overridden) */
+	{ 0x1a, 0x21a190f0 }, /* dock-mic */
+	{ 0x1c, 0x212140ff }, /* dock-HP */
+	{}
+};
+
 static const struct cxt_pincfg *cxt_pincfg_tbl[] = {
 	[CXT_PINCFG_LENOVO_X200] = cxt_pincfg_lenovo_x200,
+	[CXT_PINCFG_LENOVO_TP410] = cxt_pincfg_lenovo_tp410,
 };
 
-static const struct snd_pci_quirk cxt_fixups[] = {
+static const struct snd_pci_quirk cxt5051_fixups[] = {
 	SND_PCI_QUIRK(0x17aa, 0x20f2, "Lenovo X200", CXT_PINCFG_LENOVO_X200),
 	{}
 };
 
+static const struct snd_pci_quirk cxt5066_fixups[] = {
+	SND_PCI_QUIRK(0x17aa, 0x20f2, "Lenovo T400", CXT_PINCFG_LENOVO_TP410),
+	SND_PCI_QUIRK(0x17aa, 0x215e, "Lenovo T410", CXT_PINCFG_LENOVO_TP410),
+	SND_PCI_QUIRK(0x17aa, 0x215f, "Lenovo T510", CXT_PINCFG_LENOVO_TP410),
+	SND_PCI_QUIRK(0x17aa, 0x21ce, "Lenovo T420", CXT_PINCFG_LENOVO_TP410),
+	SND_PCI_QUIRK(0x17aa, 0x21cf, "Lenovo T520", CXT_PINCFG_LENOVO_TP410),
+	{}
+};
+
 /* add "fake" mute amp-caps to DACs on cx5051 so that mixer mute switches
  * can be created (bko#42825)
  */
@@ -4416,11 +4436,13 @@ static int patch_conexant_auto(struct hda_codec *codec)
 		break;
 	case 0x14f15051:
 		add_cx5051_fake_mutes(codec);
+		apply_pin_fixup(codec, cxt5051_fixups, cxt_pincfg_tbl);
+		break;
+	default:
+		apply_pin_fixup(codec, cxt5066_fixups, cxt_pincfg_tbl);
 		break;
 	}
 
-	apply_pin_fixup(codec, cxt_fixups, cxt_pincfg_tbl);
-
 	err = cx_auto_search_adcs(codec);
 	if (err < 0)
 		return err;

Re: Thinkpad hda regression in 3.3

[Josh Boyer]

On Fri, Apr 13, 2012 at 06:07:47PM +0200, Takashi Iwai wrote:
> At Fri, 13 Apr 2012 15:07:07 +0200,
> Takashi Iwai wrote:
> (snip)
> > If the above seem work (again, make sure that you adjust the mixer
> > mutes and levels), I'll make kernel patches to add these fixups
> > statically in the driver so that it works without extra patching.
> 
> Looking through alsa-info.sh outputs from various ThinkPad models I've
> gathered, it seems that we can use the same pin setups for all.

We had a user try the alsa-tp-dock patch on an X201 and it didn't seem
to work for them.  See bug
https://bugzilla.redhat.com/show_bug.cgi?id=806217

Perhaps the [codec] section isn't set properly for their configuration?
The alsa-info for their machine is in comment #9 of the bug.

> Below is a test patch for 3.3 kernel.  (Note that it'd give a merge
> conflict for 3.4 kernel.)

OK, I'll do a scratch build with this and get people to test.

josh

Re: Thinkpad hda regression in 3.3

[Takashi Iwai]

At Fri, 13 Apr 2012 14:08:14 -0400,
Josh Boyer wrote:
> 
> On Fri, Apr 13, 2012 at 06:07:47PM +0200, Takashi Iwai wrote:
> > At Fri, 13 Apr 2012 15:07:07 +0200,
> > Takashi Iwai wrote:
> > (snip)
> > > If the above seem work (again, make sure that you adjust the mixer
> > > mutes and levels), I'll make kernel patches to add these fixups
> > > statically in the driver so that it works without extra patching.
> > 
> > Looking through alsa-info.sh outputs from various ThinkPad models I've
> > gathered, it seems that we can use the same pin setups for all.
> 
> We had a user try the alsa-tp-dock patch on an X201 and it didn't seem
> to work for them.  See bug
> https://bugzilla.redhat.com/show_bug.cgi?id=806217
> 
> Perhaps the [codec] section isn't set properly for their configuration?

Right.  This must match with the actual values.
Or the recent driver accepts 0 as the the second value.

> The alsa-info for their machine is in comment #9 of the bug.
> 
> > Below is a test patch for 3.3 kernel.  (Note that it'd give a merge
> > conflict for 3.4 kernel.)
> 
> OK, I'll do a scratch build with this and get people to test.

In that case, check whether the reported machines have the
corresponding PCI SSIDs.


thanks,

Takashi

Re: Thinkpad hda regression in 3.3

[Takashi Iwai]

At Fri, 13 Apr 2012 14:08:14 -0400,
Josh Boyer wrote:
> 
> On Fri, Apr 13, 2012 at 06:07:47PM +0200, Takashi Iwai wrote:
> > At Fri, 13 Apr 2012 15:07:07 +0200,
> > Takashi Iwai wrote:
> > (snip)
> > > If the above seem work (again, make sure that you adjust the mixer
> > > mutes and levels), I'll make kernel patches to add these fixups
> > > statically in the driver so that it works without extra patching.
> > 
> > Looking through alsa-info.sh outputs from various ThinkPad models I've
> > gathered, it seems that we can use the same pin setups for all.
> 
> We had a user try the alsa-tp-dock patch on an X201 and it didn't seem
> to work for them.  See bug
> https://bugzilla.redhat.com/show_bug.cgi?id=806217
> 
> Perhaps the [codec] section isn't set properly for their configuration?

Right.  This must match with the actual values.
Or the recent driver accepts 0 as the the second value.

> The alsa-info for their machine is in comment #9 of the bug.
> 
> > Below is a test patch for 3.3 kernel.  (Note that it'd give a merge
> > conflict for 3.4 kernel.)
> 
> OK, I'll do a scratch build with this and get people to test.

In that case, check whether the reported machines have the
corresponding PCI SSIDs.


thanks,

Takashi

Re: Thinkpad hda regression in 3.3

[Josh Boyer]

On Fri, Apr 13, 2012 at 02:08:14PM -0400, Josh Boyer wrote:
> On Fri, Apr 13, 2012 at 06:07:47PM +0200, Takashi Iwai wrote:
> > At Fri, 13 Apr 2012 15:07:07 +0200,
> > Takashi Iwai wrote:
> > (snip)
> > > If the above seem work (again, make sure that you adjust the mixer
> > > mutes and levels), I'll make kernel patches to add these fixups
> > > statically in the driver so that it works without extra patching.
> > 
> > Looking through alsa-info.sh outputs from various ThinkPad models I've
> > gathered, it seems that we can use the same pin setups for all.
> 
> We had a user try the alsa-tp-dock patch on an X201 and it didn't seem
> to work for them.  See bug
> https://bugzilla.redhat.com/show_bug.cgi?id=806217
> 
> Perhaps the [codec] section isn't set properly for their configuration?
> The alsa-info for their machine is in comment #9 of the bug.
> 
> > Below is a test patch for 3.3 kernel.  (Note that it'd give a merge
> > conflict for 3.4 kernel.)
> 
> OK, I'll do a scratch build with this and get people to test.

I built your patch in a scratch kernel and have at least one user with a
t510 laptop report success.  Christoph reported the mixers showing up,
but it wasn't clear to me if his sound actually worked on the x201.
I've CC'd both users (Kevin t510, Christoph x201 in a dock).

josh

Re: Thinkpad hda regression in 3.3

[Takashi Iwai]

At Tue, 17 Apr 2012 20:23:36 -0400,
Josh Boyer wrote:
> 
> On Fri, Apr 13, 2012 at 02:08:14PM -0400, Josh Boyer wrote:
> > On Fri, Apr 13, 2012 at 06:07:47PM +0200, Takashi Iwai wrote:
> > > At Fri, 13 Apr 2012 15:07:07 +0200,
> > > Takashi Iwai wrote:
> > > (snip)
> > > > If the above seem work (again, make sure that you adjust the mixer
> > > > mutes and levels), I'll make kernel patches to add these fixups
> > > > statically in the driver so that it works without extra patching.
> > > 
> > > Looking through alsa-info.sh outputs from various ThinkPad models I've
> > > gathered, it seems that we can use the same pin setups for all.
> > 
> > We had a user try the alsa-tp-dock patch on an X201 and it didn't seem
> > to work for them.  See bug
> > https://bugzilla.redhat.com/show_bug.cgi?id=806217
> > 
> > Perhaps the [codec] section isn't set properly for their configuration?
> > The alsa-info for their machine is in comment #9 of the bug.
> > 
> > > Below is a test patch for 3.3 kernel.  (Note that it'd give a merge
> > > conflict for 3.4 kernel.)
> > 
> > OK, I'll do a scratch build with this and get people to test.
> 
> I built your patch in a scratch kernel and have at least one user with a
> t510 laptop report success.  Christoph reported the mixers showing up,
> but it wasn't clear to me if his sound actually worked on the x201.
> I've CC'd both users (Kevin t510, Christoph x201 in a dock).

Good.  Should I wait for a copule of days for the test results before
merging the patch, or go for now?


thanks,

Takashi