Re: [Openchrome-users] via epia boards sound quality

Re: [Openchrome-users] via epia boards sound quality

[Andrey Liakhovets]

[-- Attachment #1: Type: text/plain, Size: 2156 bytes --]

James Courtier-Dutton wrote:
> Andrey Liakhovets wrote:
>>
>> In short, the attached 1-line patch against alsa 1.0.11 solves the problem
>>
>> If necessary, I can give more details or upload them on our ftp.
>>
>> Andrey Liakhovets
>
>
> Would you be able to provide 2 things?
>
> 1) Please add a "Signed-off-by:" entry.
> 2) Do you have the VT1616A datasheet?
>
> I can find the VT1616 datasheet online, but not the VT1616A.
>
> James

No, I have only the VT1616 datasheet.

Below I repeat the references, info and patch with my Signed-off-by.

http://wiki.openchrome.org/pipermail/openchrome-users/2006-May/001193.html
http://wiki.openchrome.org/pipermail/openchrome-users/2006-May/001199.html
http://wiki.openchrome.org/pipermail/openchrome-users/2006-May/001247.html
http://wiki.openchrome.org/pipermail/openchrome-users/2006-July/001456.html

1) in our lab, we failed in getting bad sound on a running system,
        despite a long and intensive heating;
2) we have succeded in reproducing the problem at the system start (with a
        probability of about 2/3) with a pre-cooled sound codec chip;
3) we have succeded in reproducing the problem on the system running Linux
        by cutting analog power supply from the codec, cooling the codec and
        reattaching analog power;
4) on wire, the problem looks like undervoltage of codec analog power
        (2V or less instead of 5V): meanwhile the codec tries to consume
        about 200mA (instead of <40mA), and the voltage regulator chip begins
        (from 100mA) reducing the voltage;
5) analysis of the situation with WinXP shows that the WinXP sound driver loading
        brings the codec analog voltage (and sound) to norm, and this state
        is tolerant to actions described under (3);
6) analysis of the WinXP sound driver loading allowed us to detect that very bit,
        the set of which normalizes the codec's analog power consumption;
7) a check with Linux has confirmed this.

The attached 1-line patch against alsa 1.0.11 solves the problem
(for our VIA EPIA SP8000).

Signed-off-by: Andrey Liakhovets <liakh@dol.ru>

[-- Attachment #2: alsa-kernel-1.0.11-rover-ac97-vt1617a.patch --]
[-- Type: application/octet-stream, Size: 519 bytes --]

diff -Naur a/pci/ac97/ac97_patch.c b/pci/ac97/ac97_patch.c
--- a/pci/ac97/ac97_patch.c	2006-04-12 00:46:47 +0400
+++ b/pci/ac97/ac97_patch.c	2006-07-26 10:15:58 +0400
@@ -2742,6 +2742,7 @@
  */
 int patch_vt1617a(struct snd_ac97 * ac97)
 {
+	snd_ac97_write_cache(ac97, 0x5c, 0x20);	/* bring analog power consumption to normal, like WinXP driver for EPIA SP */
 	ac97->ext_id |= AC97_EI_SPDIF;	/* force the detection of spdif */
 	ac97->rates[AC97_RATES_SPDIF] = SNDRV_PCM_RATE_44100 | SNDRV_PCM_RATE_48000;
 	return 0;

[-- Attachment #3: Type: text/plain, Size: 373 bytes --]

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

[-- Attachment #4: Type: text/plain, Size: 161 bytes --]

_______________________________________________
Alsa-devel mailing list
Alsa-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/alsa-devel

Re: [Openchrome-users] via epia boards sound quality

[James Courtier-Dutton]

Andrey Liakhovets wrote:
> James Courtier-Dutton wrote:
>> Andrey Liakhovets wrote:
>>> In short, the attached 1-line patch against alsa 1.0.11 solves the problem
>>>
>>> If necessary, I can give more details or upload them on our ftp.
>>>
>>> Andrey Liakhovets
>>
>> Would you be able to provide 2 things?
>>
>> 1) Please add a "Signed-off-by:" entry.
>> 2) Do you have the VT1616A datasheet?
>>
>> I can find the VT1616 datasheet online, but not the VT1616A.
>>
>> James
> 
> No, I have only the VT1616 datasheet.
> 
> Below I repeat the references, info and patch with my Signed-off-by.
> 

So, with the absence of the VT1617A datasheet, we effectively have no
documentation for that AC97 register touched in your patch.
Could I ask you to raise an alsa bug at
https://bugtrack.alsa-project.org/alsa-bug/

Enter into the bug everything from your updated email, with the sign off
and attached patch.
I will then check the patch into the HG repository, but I will add a
comment next to it pointing to the bug information. That way, future
developers will know why it has been put there.

James

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

Re: [Openchrome-users] via epia boards sound quality

[Thierry Vignaud]

James Courtier-Dutton <James@superbug.co.uk> writes:

> So, with the absence of the VT1617A datasheet, we effectively have
> no documentation for that AC97 register touched in your patch.
> Could I ask you to raise an alsa bug at
> https://bugtrack.alsa-project.org/alsa-bug/
> 
> Enter into the bug everything from your updated email, with the sign
> off and attached patch.
> I will then check the patch into the HG repository, but I will add a
> comment next to it pointing to the bug information. That way, future
> developers will know why it has been put there.

this was added into
https://bugtrack.alsa-project.org/alsa-bug/view.php?id=2381 but didn't
make it into HG (or at least the BR wasn't closed)...

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

Re: [Openchrome-users] via epia boards sound quality

[Takashi Iwai]

At Mon, 28 Aug 2006 16:38:56 +0200,
Thierry Vignaud wrote:
> 
> James Courtier-Dutton <James@superbug.co.uk> writes:
> 
> > So, with the absence of the VT1617A datasheet, we effectively have
> > no documentation for that AC97 register touched in your patch.
> > Could I ask you to raise an alsa bug at
> > https://bugtrack.alsa-project.org/alsa-bug/
> > 
> > Enter into the bug everything from your updated email, with the sign
> > off and attached patch.
> > I will then check the patch into the HG repository, but I will add a
> > comment next to it pointing to the bug information. That way, future
> > developers will know why it has been put there.
> 
> this was added into
> https://bugtrack.alsa-project.org/alsa-bug/view.php?id=2381 but didn't
> make it into HG (or at least the BR wasn't closed)...

I committed now.

Thanks,

Takashi

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

Re: [Openchrome-users] via epia boards sound quality

[James Courtier-Dutton]

Takashi Iwai wrote:
> At Mon, 28 Aug 2006 16:38:56 +0200,
> Thierry Vignaud wrote:
>   
>> James Courtier-Dutton <James@superbug.co.uk> writes:
>>
>>     
>>> So, with the absence of the VT1617A datasheet, we effectively have
>>> no documentation for that AC97 register touched in your patch.
>>> Could I ask you to raise an alsa bug at
>>> https://bugtrack.alsa-project.org/alsa-bug/
>>>
>>> Enter into the bug everything from your updated email, with the sign
>>> off and attached patch.
>>> I will then check the patch into the HG repository, but I will add a
>>> comment next to it pointing to the bug information. That way, future
>>> developers will know why it has been put there.
>>>       
>> this was added into
>> https://bugtrack.alsa-project.org/alsa-bug/view.php?id=2381 but didn't
>> make it into HG (or at least the BR wasn't closed)...
>>     
>
> I committed now.
>
> Thanks,
>
> Takashi
>   

For those interested, I managed to get the VT1617 datasheet from VIA. 
(no NDAs involved.)
One can find it here:
http://alsa-project.org/~james/datasheets/via/
Filename: DS_VT1617_120.pdf

Maybe we should move it to the alsa ftp site with all the other datasheets.

James


-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

Re: [Openchrome-users] via epia boards sound quality

[Jaroslav Kysela]

On Tue, 29 Aug 2006, James Courtier-Dutton wrote:

> Takashi Iwai wrote:
> > At Mon, 28 Aug 2006 16:38:56 +0200,
> > Thierry Vignaud wrote:
> >   
> >> James Courtier-Dutton <James@superbug.co.uk> writes:
> >>
> >>     
> >>> So, with the absence of the VT1617A datasheet, we effectively have
> >>> no documentation for that AC97 register touched in your patch.
> >>> Could I ask you to raise an alsa bug at
> >>> https://bugtrack.alsa-project.org/alsa-bug/
> >>>
> >>> Enter into the bug everything from your updated email, with the sign
> >>> off and attached patch.
> >>> I will then check the patch into the HG repository, but I will add a
> >>> comment next to it pointing to the bug information. That way, future
> >>> developers will know why it has been put there.
> >>>       
> >> this was added into
> >> https://bugtrack.alsa-project.org/alsa-bug/view.php?id=2381 but didn't
> >> make it into HG (or at least the BR wasn't closed)...
> >>     
> >
> > I committed now.
> >
> > Thanks,
> >
> > Takashi
> >   
> 
> For those interested, I managed to get the VT1617 datasheet from VIA. 
> (no NDAs involved.)
> One can find it here:
> http://alsa-project.org/~james/datasheets/via/
> Filename: DS_VT1617_120.pdf

Ok, the URL is 
ftp://ftp.alsa-project.org/pub/datasheets/via/DS_VT1617_120.pdf now.
I'm moving other datasheets too.

						Jaroslav

-----
Jaroslav Kysela <perex@suse.cz>
Linux Kernel Sound Maintainer
ALSA Project, SUSE Labs

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

Re: [Openchrome-users] via epia boards sound quality

[James Courtier-Dutton]

Takashi Iwai wrote:
> At Mon, 28 Aug 2006 16:38:56 +0200,
> Thierry Vignaud wrote:
>   
>> James Courtier-Dutton <James@superbug.co.uk> writes:
>>
>>     
>>> So, with the absence of the VT1617A datasheet, we effectively have
>>> no documentation for that AC97 register touched in your patch.
>>> Could I ask you to raise an alsa bug at
>>> https://bugtrack.alsa-project.org/alsa-bug/
>>>
>>> Enter into the bug everything from your updated email, with the sign
>>> off and attached patch.
>>> I will then check the patch into the HG repository, but I will add a
>>> comment next to it pointing to the bug information. That way, future
>>> developers will know why it has been put there.
>>>       
>> this was added into
>> https://bugtrack.alsa-project.org/alsa-bug/view.php?id=2381 but didn't
>> make it into HG (or at least the BR wasn't closed)...
>>     
>
> I committed now.
>
> Thanks,
>
> Takashi
>   

Now that we have the datasheet, it seems that the register you touch 
could cause hardware damage.
 From the datasheet, it disables the headphone amp thermal protection. 
So, this fix could be counter productive. I.e. result in damage to the chip.

I suggest that we only apply it via a module option, so the user then 
knows they are taking a risk by disabling the protection.
If the protection is bogus, i.e. the sensor is faulty, then we need a 
way to detect this faultiness.


James

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

Re: [Openchrome-users] via epia boards sound quality

[Takashi Iwai]

At Tue, 29 Aug 2006 13:17:26 +0100,
James Courtier-Dutton wrote:
> 
> Takashi Iwai wrote:
> > At Mon, 28 Aug 2006 16:38:56 +0200,
> > Thierry Vignaud wrote:
> >   
> >> James Courtier-Dutton <James@superbug.co.uk> writes:
> >>
> >>     
> >>> So, with the absence of the VT1617A datasheet, we effectively have
> >>> no documentation for that AC97 register touched in your patch.
> >>> Could I ask you to raise an alsa bug at
> >>> https://bugtrack.alsa-project.org/alsa-bug/
> >>>
> >>> Enter into the bug everything from your updated email, with the sign
> >>> off and attached patch.
> >>> I will then check the patch into the HG repository, but I will add a
> >>> comment next to it pointing to the bug information. That way, future
> >>> developers will know why it has been put there.
> >>>       
> >> this was added into
> >> https://bugtrack.alsa-project.org/alsa-bug/view.php?id=2381 but didn't
> >> make it into HG (or at least the BR wasn't closed)...
> >>     
> >
> > I committed now.
> >
> > Thanks,
> >
> > Takashi
> >   
> 
> Now that we have the datasheet, it seems that the register you touch 
> could cause hardware damage.
>  From the datasheet, it disables the headphone amp thermal protection. 
> So, this fix could be counter productive. I.e. result in damage to the chip.
> 
> I suggest that we only apply it via a module option, so the user then 
> knows they are taking a risk by disabling the protection.

Better to check a PCI SSID with the whitelist than a new module
option.  If you want to make it general, you can add a new quirk for
that purpose.  But I guess simply checking the PCI SSID with a static
whitelist in patch_vt1617a() should suffice.


Takashi

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

Re: [Openchrome-users] via epia boards sound quality

[Andrey Liakhovets]

James Courtier-Dutton wrote:
> Andrey Liakhovets wrote:
>> James Courtier-Dutton wrote:
>>> Andrey Liakhovets wrote:
>>>> In short, the attached 1-line patch against alsa 1.0.11 solves the problem
>>>>
>>>> If necessary, I can give more details or upload them on our ftp.
>>>>
>>>> Andrey Liakhovets
>>>
>>> Would you be able to provide 2 things?
>>>
>>> 1) Please add a "Signed-off-by:" entry.
>>> 2) Do you have the VT1616A datasheet?
>>>
>>> I can find the VT1616 datasheet online, but not the VT1616A.
>>>
>>> James
>>
>> No, I have only the VT1616 datasheet.
>>
>> Below I repeat the references, info and patch with my Signed-off-by.
>>
>
> So, with the absence of the VT1617A datasheet, we effectively have no
> documentation for that AC97 register touched in your patch.

Yes.

> Could I ask you to raise an alsa bug at
> https://bugtrack.alsa-project.org/alsa-bug/
>
> Enter into the bug everything from your updated email, with the sign off
> and attached patch.

https://bugtrack.alsa-project.org/alsa-bug/view.php?id=2381

> I will then check the patch into the HG repository, but I will add a
> comment next to it pointing to the bug information. That way, future
> developers will know why it has been put there.
>
> James

Andrey Liakhovets

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

Re: [Openchrome-users] via epia boards sound quality

[Andrey Liakhovets]

[-- Attachment #1: Type: text/plain, Size: 1772 bytes --]

This letter is continuation of the thread in Openchrome-users:
http://wiki.openchrome.org/pipermail/openchrome-users/2006-May/001193.html
http://wiki.openchrome.org/pipermail/openchrome-users/2006-May/001199.html
http://wiki.openchrome.org/pipermail/openchrome-users/2006-May/001247.html
http://wiki.openchrome.org/pipermail/openchrome-users/2006-July/001456.html

I am sorry for continuing this thread but there is good news.

In short, the attached 1-line patch against alsa 1.0.11 solves the problem
(for our VIA EPIA SP8000).

Further details:
1) in our lab, we failed in getting bad sound on a running system,
       despite a long and intensive heating;
2) we have succeded in reproducing the problem at the system start (with a
       probability of about 2/3) with a pre-cooled sound codec chip;
3) we have succeded in reproducing the problem on the system running Linux
       by cutting analog power supply from the codec, cooling the codec and
       reattaching analog power;
4) on wire, the problem looks like undervoltage of codec analog power
       (2V or less instead of 5V): meanwhile the codec tries to consume
       about 200mA (instead of <40mA), and the voltage regulator chip begins
       (from 100mA) reducing the voltage;
5) analysis of the situation with WinXP shows that the WinXP sound driver loading
       brings the codec analog voltage (and sound) to norm, and this state
       is tolerant to actions described under (3);
6) analysis of the WinXP sound driver loading allowed us to detect that very bit,
       the set of which normalizes the codec's analog power consumption;
7) a check with Linux has confirmed this.

If necessary, I can give more details or upload them on our ftp.

Andrey Liakhovets

[-- Attachment #2: alsa-kernel-1.0.11-rover-ac97-vt1617a.patch --]
[-- Type: application/octet-stream, Size: 536 bytes --]

diff -Naur a/pci/ac97/ac97_patch.c b/pci/ac97/ac97_patch.c
--- a/pci/ac97/ac97_patch.c	2006-04-12 00:46:47 +0400
+++ b/pci/ac97/ac97_patch.c	2006-07-26 10:15:58 +0400
@@ -2742,6 +2742,7 @@
  */
 int patch_vt1617a(struct snd_ac97 * ac97)
 {
+	snd_ac97_write_cache(ac97, 0x5c, 0x20);	/* bring analog power consumption to normal, like WinXP driver for EPIA SP */
 	ac97->ext_id |= AC97_EI_SPDIF;	/* force the detection of spdif */
 	ac97->rates[AC97_RATES_SPDIF] = SNDRV_PCM_RATE_44100 | SNDRV_PCM_RATE_48000;
 	return 0;




[-- Attachment #3: Type: text/plain, Size: 373 bytes --]

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

[-- Attachment #4: Type: text/plain, Size: 161 bytes --]

_______________________________________________
Alsa-devel mailing list
Alsa-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/alsa-devel

Re: [Openchrome-users] via epia boards sound quality

[James Courtier-Dutton]

Andrey Liakhovets wrote:
> 
> In short, the attached 1-line patch against alsa 1.0.11 solves the problem
> (for our VIA EPIA SP8000).
> 
> 
> If necessary, I can give more details or upload them on our ftp.
> 
> Andrey Liakhovets
> 
> 


Would you be able to provide 2 things?

1) Please add a "Signed-off-by:" entry.
2) Do you have the VT1616A datasheet?

I can find the VT1616 datasheet online, but not the VT1616A.

James

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642