nforce2 (snd_intel8x0) channels scrolling

nforce2 (snd_intel8x0) channels scrolling

[Vladimir I. Umnov]

Hw: a7n8x deluxe 2.0 (nforce2-c1 + MCP-T)
Sw: 2.6.0-test8, alsa-lib from cvs
lspci:
00:05.0 Multimedia audio controller: nVidia Corporation: Unknown device
006b (rev a2)        Subsystem: Asustek Computer, Inc.: Unknown device
0c11        Flags: bus master, 66Mhz, fast devsel, latency 0, IRQ 5
        Memory at d7000000 (32-bit, non-prefetchable) [size=512K]
        Capabilities: [44] Power Management version 2

00:06.0 Multimedia audio controller: nVidia Corporation: Unknown device
006a (rev a1)        Subsystem: Asustek Computer, Inc.: Unknown device
8095        Flags: bus master, 66Mhz, fast devsel, latency 0, IRQ 11
        I/O ports at e400 [size=256]
        I/O ports at d000 [size=128]
        Memory at d7087000 (32-bit, non-prefetchable) [size=4K]
        Capabilities: [44] Power Management version 2

Mplayer and xine used for testing in 6channels mode.
Channels mapping stays wrong after:
1) snd_pcm_drop() called (seeking and quit);
2) after long time of playing in 6channels mode, even without x-run.

p.s. My sblive works very well now, but my friends doesn't have sblive
and they want working 6channels audio under linux on nforce2.

p.p.s. I may help with testing or finding a bug, but I need help. I new
to alsa-driver and can't understand how it works, and can't find how
snd_pcm_drop() implemented in i810 module.


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/

Re: nforce2 (snd_intel8x0) channels scrolling

[James Courtier-Dutton]

[-- Attachment #1: Type: text/plain, Size: 1698 bytes --]

Vladimir I. Umnov wrote:
> Hw: a7n8x deluxe 2.0 (nforce2-c1 + MCP-T)
> Sw: 2.6.0-test8, alsa-lib from cvs
> lspci:
> 00:05.0 Multimedia audio controller: nVidia Corporation: Unknown device
> 006b (rev a2)        Subsystem: Asustek Computer, Inc.: Unknown device
> 0c11        Flags: bus master, 66Mhz, fast devsel, latency 0, IRQ 5
>         Memory at d7000000 (32-bit, non-prefetchable) [size=512K]
>         Capabilities: [44] Power Management version 2
> 
> 00:06.0 Multimedia audio controller: nVidia Corporation: Unknown device
> 006a (rev a1)        Subsystem: Asustek Computer, Inc.: Unknown device
> 8095        Flags: bus master, 66Mhz, fast devsel, latency 0, IRQ 11
>         I/O ports at e400 [size=256]
>         I/O ports at d000 [size=128]
>         Memory at d7087000 (32-bit, non-prefetchable) [size=4K]
>         Capabilities: [44] Power Management version 2
> 
> Mplayer and xine used for testing in 6channels mode.
> Channels mapping stays wrong after:
> 1) snd_pcm_drop() called (seeking and quit);
> 2) after long time of playing in 6channels mode, even without x-run.
> 
> p.s. My sblive works very well now, but my friends doesn't have sblive
> and they want working 6channels audio under linux on nforce2.
> 
> p.p.s. I may help with testing or finding a bug, but I need help. I new
> to alsa-driver and can't understand how it works, and can't find how
> snd_pcm_drop() implemented in i810 module.
> 
> 
Here is a simple speaker test program.
I nicely tests different speaker layout modes.

Read the readme.txt file that is with it on how to use it.
If you still get the problem with this simple program, it is certainly a 
bug in alsa.

See attached file.

Cheers
James


[-- Attachment #2: speaker_test.tar.bz2 --]
[-- Type: application/octet-stream, Size: 3347 bytes --]

Re: nforce2 (snd_intel8x0) channels scrolling

[Vladimir I. Umnov]

On Sun, 09 Nov 2003 15:25:25 +0000
James Courtier-Dutton <James@superbug.demon.co.uk> wrote:

> Vladimir I. Umnov wrote:
> > Hw: a7n8x deluxe 2.0 (nforce2-c1 + MCP-T)
> > Sw: 2.6.0-test8, alsa-lib from cvs
> > lspci:
> > 00:05.0 Multimedia audio controller: nVidia Corporation: Unknown
> > device 006b (rev a2)        Subsystem: Asustek Computer, Inc.:
> > Unknown device 0c11        Flags: bus master, 66Mhz, fast devsel,
> > latency 0, IRQ 5
> >         Memory at d7000000 (32-bit, non-prefetchable) [size=512K]
> >         Capabilities: [44] Power Management version 2
> > 
> > 00:06.0 Multimedia audio controller: nVidia Corporation: Unknown
> > device 006a (rev a1)        Subsystem: Asustek Computer, Inc.:
> > Unknown device 8095        Flags: bus master, 66Mhz, fast devsel,
> > latency 0, IRQ 11
> >         I/O ports at e400 [size=256]
> >         I/O ports at d000 [size=128]
> >         Memory at d7087000 (32-bit, non-prefetchable) [size=4K]
> >         Capabilities: [44] Power Management version 2
> > 
> > Mplayer and xine used for testing in 6channels mode.
> > Channels mapping stays wrong after:
> > 1) snd_pcm_drop() called (seeking and quit);
> > 2) after long time of playing in 6channels mode, even without x-run.
> > 
> > p.s. My sblive works very well now, but my friends doesn't have
> > sblive and they want working 6channels audio under linux on nforce2.
> > 
> > p.p.s. I may help with testing or finding a bug, but I need help. I
> > new to alsa-driver and can't understand how it works, and can't find
> > how snd_pcm_drop() implemented in i810 module.
> > 
> > 
> Here is a simple speaker test program.
> I nicely tests different speaker layout modes.
> 
> Read the readme.txt file that is with it on how to use it.
> If you still get the problem with this simple program, it is certainly
> a bug in alsa.
After channels scrolled (after a few seek in mplayer, then quit),
./speaker_test-Dsurround51-c 6 works wrong (Center->Front Left,
LFE->Front Right, ...).
But after I use "front" device and then repeat test with 6 channels
setup, channel mapping is ok.
If i press ctrl+c with 6channels test and then repeat it
(snd_pcm_drop()) scrool is apear, and after i run 3 times, channels
mapping is ok.

With sblive all works perfectly!!! This is alsa-driver bug.


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/

Re: nforce2 (snd_intel8x0) channels scrolling

[Takashi Iwai]

At Sun, 9 Nov 2003 20:37:15 +0300,
Vladimir I. Umnov wrote:
> 
> On Sun, 09 Nov 2003 15:25:25 +0000
> James Courtier-Dutton <James@superbug.demon.co.uk> wrote:
> 
> > Vladimir I. Umnov wrote:
> > > Hw: a7n8x deluxe 2.0 (nforce2-c1 + MCP-T)
> > > Sw: 2.6.0-test8, alsa-lib from cvs
> > > lspci:
> > > 00:05.0 Multimedia audio controller: nVidia Corporation: Unknown
> > > device 006b (rev a2)        Subsystem: Asustek Computer, Inc.:
> > > Unknown device 0c11        Flags: bus master, 66Mhz, fast devsel,
> > > latency 0, IRQ 5
> > >         Memory at d7000000 (32-bit, non-prefetchable) [size=512K]
> > >         Capabilities: [44] Power Management version 2
> > > 
> > > 00:06.0 Multimedia audio controller: nVidia Corporation: Unknown
> > > device 006a (rev a1)        Subsystem: Asustek Computer, Inc.:
> > > Unknown device 8095        Flags: bus master, 66Mhz, fast devsel,
> > > latency 0, IRQ 11
> > >         I/O ports at e400 [size=256]
> > >         I/O ports at d000 [size=128]
> > >         Memory at d7087000 (32-bit, non-prefetchable) [size=4K]
> > >         Capabilities: [44] Power Management version 2
> > > 
> > > Mplayer and xine used for testing in 6channels mode.
> > > Channels mapping stays wrong after:
> > > 1) snd_pcm_drop() called (seeking and quit);
> > > 2) after long time of playing in 6channels mode, even without x-run.
> > > 
> > > p.s. My sblive works very well now, but my friends doesn't have
> > > sblive and they want working 6channels audio under linux on nforce2.
> > > 
> > > p.p.s. I may help with testing or finding a bug, but I need help. I
> > > new to alsa-driver and can't understand how it works, and can't find
> > > how snd_pcm_drop() implemented in i810 module.
> > > 
> > > 
> > Here is a simple speaker test program.
> > I nicely tests different speaker layout modes.
> > 
> > Read the readme.txt file that is with it on how to use it.
> > If you still get the problem with this simple program, it is certainly
> > a bug in alsa.
> After channels scrolled (after a few seek in mplayer, then quit),
> ./speaker_test-Dsurround51-c 6 works wrong (Center->Front Left,
> LFE->Front Right, ...).
> But after I use "front" device and then repeat test with 6 channels
> setup, channel mapping is ok.
> If i press ctrl+c with 6channels test and then repeat it
> (snd_pcm_drop()) scrool is apear, and after i run 3 times, channels
> mapping is ok.
> 
> With sblive all works perfectly!!! This is alsa-driver bug.

try to change the mapping of surround51 device in
/usr/share/alsa/cards/NFORCE.conf.


Takashi


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/

RE: nforce2 (snd_intel8x0) channels scrolling

[Richard Lemon]

Hi All,

This is one of the problems I have been seeing with the intel8x0 driver as well. The problem occurs when you call snd_pcm_start a second time after an xrun, or after you have shut down a program with ctl-c and started it again in the same console. I used the 

If you close the console and open a new one the problem seems to go away!? I used the example code pcm.c from the alsa-lib test directory.

It occurs with both [nforce2 ALC650] combo and [Intel 82901EB-ICH5 Analog devices id 75].

I also have a problem when the drivers have been running for a long time (around 5 hours). Without an xrun the snd_pcm_delay() function starts returning impossible numbers. Checking drift against either rdtsc or gettimeofday() alsa starts giving me a 0.1 second per second drift.

I haven't got a machine I can set up for this style of kernel debugging unfortunately, so I am trying to code around these problems.

richard.


-----Original Message-----
From: Takashi Iwai [mailto:tiwai@suse.de]
Sent: Tuesday, 11 November 2003 4:37 AM
To: Vladimir I. Umnov
Cc: James Courtier-Dutton; alsa-devel@lists.sourceforge.net
Subject: Re: [Alsa-devel] nforce2 (snd_intel8x0) channels scrolling


At Sun, 9 Nov 2003 20:37:15 +0300,
Vladimir I. Umnov wrote:
> 
> On Sun, 09 Nov 2003 15:25:25 +0000
> James Courtier-Dutton <James@superbug.demon.co.uk> wrote:
> 
> > Vladimir I. Umnov wrote:
> > > Hw: a7n8x deluxe 2.0 (nforce2-c1 + MCP-T)
> > > Sw: 2.6.0-test8, alsa-lib from cvs
> > > lspci:
> > > 00:05.0 Multimedia audio controller: nVidia Corporation: Unknown
> > > device 006b (rev a2)        Subsystem: Asustek Computer, Inc.:
> > > Unknown device 0c11        Flags: bus master, 66Mhz, fast devsel,
> > > latency 0, IRQ 5
> > >         Memory at d7000000 (32-bit, non-prefetchable) [size=512K]
> > >         Capabilities: [44] Power Management version 2
> > > 
> > > 00:06.0 Multimedia audio controller: nVidia Corporation: Unknown
> > > device 006a (rev a1)        Subsystem: Asustek Computer, Inc.:
> > > Unknown device 8095        Flags: bus master, 66Mhz, fast devsel,
> > > latency 0, IRQ 11
> > >         I/O ports at e400 [size=256]
> > >         I/O ports at d000 [size=128]
> > >         Memory at d7087000 (32-bit, non-prefetchable) [size=4K]
> > >         Capabilities: [44] Power Management version 2
> > > 
> > > Mplayer and xine used for testing in 6channels mode.
> > > Channels mapping stays wrong after:
> > > 1) snd_pcm_drop() called (seeking and quit);
> > > 2) after long time of playing in 6channels mode, even without x-run.
> > > 
> > > p.s. My sblive works very well now, but my friends doesn't have
> > > sblive and they want working 6channels audio under linux on nforce2.
> > > 
> > > p.p.s. I may help with testing or finding a bug, but I need help. I
> > > new to alsa-driver and can't understand how it works, and can't find
> > > how snd_pcm_drop() implemented in i810 module.
> > > 
> > > 
> > Here is a simple speaker test program.
> > I nicely tests different speaker layout modes.
> > 
> > Read the readme.txt file that is with it on how to use it.
> > If you still get the problem with this simple program, it is certainly
> > a bug in alsa.
> After channels scrolled (after a few seek in mplayer, then quit),
> ./speaker_test-Dsurround51-c 6 works wrong (Center->Front Left,
> LFE->Front Right, ...).
> But after I use "front" device and then repeat test with 6 channels
> setup, channel mapping is ok.
> If i press ctrl+c with 6channels test and then repeat it
> (snd_pcm_drop()) scrool is apear, and after i run 3 times, channels
> mapping is ok.
> 
> With sblive all works perfectly!!! This is alsa-driver bug.

try to change the mapping of surround51 device in
/usr/share/alsa/cards/NFORCE.conf.


Takashi


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/
_______________________________________________
Alsa-devel mailing list
Alsa-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/alsa-devel


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/

Re: nforce2 (snd_intel8x0) channels scrolling

[James Courtier-Dutton]

Richard Lemon wrote:
> Hi All,
> 
> This is one of the problems I have been seeing with the intel8x0 driver as well. The problem occurs when you call snd_pcm_start a second time after an xrun, or after you have shut down a program with ctl-c and started it again in the same console. I used the 
> 
> If you close the console and open a new one the problem seems to go away!? I used the example code pcm.c from the alsa-lib test directory.
> 
> It occurs with both [nforce2 ALC650] combo and [Intel 82901EB-ICH5 Analog devices id 75].
> 
> I also have a problem when the drivers have been running for a long time (around 5 hours). Without an xrun the snd_pcm_delay() function starts returning impossible numbers. Checking drift against either rdtsc or gettimeofday() alsa starts giving me a 0.1 second per second drift.
> 
> I haven't got a machine I can set up for this style of kernel debugging unfortunately, so I am trying to code around these problems.
> 
> richard.
> 
> 
Would you be able to submit a small self contained (like the 
speaker-test code I submitted) source code that would reproduce the 
problem quickly.
I would expect that writing code to force xruns would be quite easy.
I have a ICH5 with ALC650 chip on Motherboard here, so I might be able 
to help you out if I could reproduce your problem.

Cheers
James




-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/

Re: nforce2 (snd_intel8x0) channels scrolling

[Vladimir I. Umnov]

On Mon, 10 Nov 2003 23:35:13 +0000
James Courtier-Dutton <James@superbug.demon.co.uk> wrote:

> Richard Lemon wrote:
> > Hi All,
> > 
> > This is one of the problems I have been seeing with the intel8x0
> > driver as well. The problem occurs when you call snd_pcm_start a
> > second time after an xrun, or after you have shut down a program
> > with ctl-c and started it again in the same console. I used the 
> > 
> > If you close the console and open a new one the problem seems to go
> > away!? I used the example code pcm.c from the alsa-lib test
> > directory.
> > 
> > It occurs with both [nforce2 ALC650] combo and [Intel 82901EB-ICH5
> > Analog devices id 75].
> > 
> > I also have a problem when the drivers have been running for a long
> > time (around 5 hours). Without an xrun the snd_pcm_delay() function
> > starts returning impossible numbers. Checking drift against either
> > rdtsc or gettimeofday() alsa starts giving me a 0.1 second per
> > second drift.
> > 
> > I haven't got a machine I can set up for this style of kernel
> > debugging unfortunately, so I am trying to code around these
> > problems.
> > 
> > richard.
> > 
> > 
> Would you be able to submit a small self contained (like the 
> speaker-test code I submitted) source code that would reproduce the 
> problem quickly.
> I would expect that writing code to force xruns would be quite easy.
> I have a ICH5 with ALC650 chip on Motherboard here, so I might be able
> to help you out if I could reproduce your problem.
try to open/close (without seeking) in mplayer or xine - always scrolls
to next position (i think 3 positions at all).


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/

Re: nforce2 (snd_intel8x0) channels scrolling

[Richard Lemon]

Hi Vladimir,

I have written a small program to show the channel scrolling problem on
the nforce 2 chipset. 

James tested it on his system and the problem did not occur. I tested it
on an Intel/Analog devices system the problem does not occur. It is
specific to the nforce2 chipset. 

I was wrong when I said it occurred on the intel system (There is
another problem I am tracking down in my code which does occur on both).

I am going to try to track down and fix the problem, but I have a fair
bit of learning to do before I will be able to figure it out.

Richard.

On Fri, 2003-11-14 at 01:48, Vladimir I. Umnov wrote:
> On Mon, 10 Nov 2003 23:35:13 +0000
> James Courtier-Dutton <James@superbug.demon.co.uk> wrote:
> 
> > Richard Lemon wrote:
> > > Hi All,
> > > 
> > > This is one of the problems I have been seeing with the intel8x0
> > > driver as well. The problem occurs when you call snd_pcm_start a
> > > second time after an xrun, or after you have shut down a program
> > > with ctl-c and started it again in the same console. I used the 
> > > 
> > > If you close the console and open a new one the problem seems to go
> > > away!? I used the example code pcm.c from the alsa-lib test
> > > directory.
> > > 
> > > It occurs with both [nforce2 ALC650] combo and [Intel 82901EB-ICH5
> > > Analog devices id 75].
> > > 
> > > I also have a problem when the drivers have been running for a long
> > > time (around 5 hours). Without an xrun the snd_pcm_delay() function
> > > starts returning impossible numbers. Checking drift against either
> > > rdtsc or gettimeofday() alsa starts giving me a 0.1 second per
> > > second drift.
> > > 
> > > I haven't got a machine I can set up for this style of kernel
> > > debugging unfortunately, so I am trying to code around these
> > > problems.
> > > 
> > > richard.
> > > 
> > > 
> > Would you be able to submit a small self contained (like the 
> > speaker-test code I submitted) source code that would reproduce the 
> > problem quickly.
> > I would expect that writing code to force xruns would be quite easy.
> > I have a ICH5 with ALC650 chip on Motherboard here, so I might be able
> > to help you out if I could reproduce your problem.
> try to open/close (without seeking) in mplayer or xine - always scrolls
> to next position (i think 3 positions at all).
> 
> 
> -------------------------------------------------------
> This SF.Net email sponsored by: ApacheCon 2003,
> 16-19 November in Las Vegas. Learn firsthand the latest
> developments in Apache, PHP, Perl, XML, Java, MySQL,
> WebDAV, and more! http://www.apachecon.com/
> _______________________________________________
> Alsa-devel mailing list
> Alsa-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/alsa-devel




-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/