[OOPS] Alsa 0.9.4 / kernel 2.6.0-test1-mm1

[OOPS] Alsa 0.9.4 / kernel 2.6.0-test1-mm1

[Vince]

Hi, I get the following oops (100% reproductible) with this simple command:

----------------------------------------------------------
diablo:~$ festival
Festival Speech Synthesis System 1.4.3:release Jan 2003
Copyright (C) University of Edinburgh, 1996-2003. All rights reserved.
For details type `(festival_warranty)'
festival> (SayText "Good morning")
zsh: segmentation fault  festival
----------------------------------------------------------

Unable to handle kernel paging request at virtual address d0b5f000
  printing eip:
d0acaf7b
*pde = 013ea067
*pte = 00000000
Oops: 0000 [#1]
PREEMPT
CPU:    0
EIP:    0060:[<d0acaf7b>]    Not tainted VLI
EFLAGS: 00210202
EIP is at resample_expand+0x34b/0x380 [snd_pcm_oss]
eax: d0acaf7b   ebx: 00000001   ecx: 000007ff   edx: ffecffec
esi: d0b7dff6   edi: cee8f8cc   ebp: d0b5effe   esp: c80ffe58
ds: 007b   es: 007b   ss: 0068
Process festival (pid: 1260, threadinfo=c80fe000 task=ce1066d0)
Stack: d0ac8212 cee8f840 ca851d80 c80ffe84 0000ffec c80fe000 ffffffff 
cee8f540
        d0acada1 d0acaef2 d0acaf7b cee8f8b0 00000000 00000002 00000002 
00000000
        ffecffec 000002a8 000017fd 00000800 cee8f840 c91ea8c0 d0acb46c 
cee8f840
Call Trace:
  [<d0ac8212>] snd_pcm_plug_playback_channels_mask+0x72/0xe0 [snd_pcm_oss]
  [<d0acada1>] resample_expand+0x171/0x380 [snd_pcm_oss]
  [<d0acaef2>] resample_expand+0x2c2/0x380 [snd_pcm_oss]
  [<d0acaf7b>] resample_expand+0x34b/0x380 [snd_pcm_oss]
  [<d0acb46c>] rate_transfer+0x5c/0x60 [snd_pcm_oss]
  [<d0ac8607>] snd_pcm_plug_write_transfer+0x97/0x100 [snd_pcm_oss]
  [<d0ac43e0>] snd_pcm_oss_write2+0xd0/0x150 [snd_pcm_oss]
  [<d0ac4606>] snd_pcm_oss_write1+0x1a6/0x1d0 [snd_pcm_oss]
  [<d0ac66c3>] snd_pcm_oss_write+0x43/0x60 [snd_pcm_oss]
  [<d0ac6680>] snd_pcm_oss_write+0x0/0x60 [snd_pcm_oss]
  [<c014e6b8>] vfs_write+0xb8/0x130
  [<c014e7e2>] sys_write+0x42/0x70
  [<c0109107>] syscall_call+0x7/0xb

Code: 66 35 00 80 eb d4 8b 45 00 c1 e8 08 eb cc 8b 45 00 c1 e8 08 66 35 
00 80 eb c2 8b 45 00 c1 e8 08 eb ba eb a6 eb b1 8b 55 00 eb bf <8b> 45 
00 eb ac 0f b6 45 00 c1 e0 08 eb a3 81 fa 00 80 00 00 b8
-------------------------------------------------------------------


More information on my system:
diablo:/usr/src# cat /proc/asound/version
Advanced Linux Sound Architecture Driver Version 0.9.4 (Mon Jun 09 
12:01:18 2003 UTC).
Compiled on Jul 16 2003 for kernel 2.6.0-test1-mm1.

diablo:/usr/src# cat /proc/asound/cards
0 [AudioPCI       ]: ENS1371 - Ensoniq AudioPCI
                      Ensoniq AudioPCI ENS1371 at 0xec00, irq 10

diablo:/usr/src# cat /proc/asound/modules
0 snd_ens1371

diablo:/usr/src/linux-2.6.0test1# cat .config | grep -v '#'  | grep SND
CONFIG_SND=m
CONFIG_SND_SEQUENCER=m
CONFIG_SND_OSSEMUL=y
CONFIG_SND_MIXER_OSS=m
CONFIG_SND_PCM_OSS=m
CONFIG_SND_SEQUENCER_OSS=y
CONFIG_SND_RTCTIMER=m
CONFIG_SND_ENS1371=m


Please let me know if any further information is needed.

Regards,

Vincent

P.S.: I intended to test alsa 0.9.5 (in the hope that this bug was 
already known & has already been fixed), but unfortunately I get a 
compile error (I simply dropped alsa-kernel 0.9.5 in the 2.6.0-test1-mm1 
source tree):


-----------------------------------------------------------
   CC [M]  sound/pci/ac97/ac97_codec.o
sound/pci/ac97/ac97_codec.c: In function `snd_ac97_mixer_build':
sound/pci/ac97/ac97_codec.c:1393: error: structure has no member named 
`build_ops'
sound/pci/ac97/ac97_codec.c:1393: error: structure has no member named 
`build_ops'
sound/pci/ac97/ac97_codec.c:1394: error: structure has no member named 
`build_ops'
sound/pci/ac97/ac97_codec.c:1416: error: structure has no member named 
`build_ops'
sound/pci/ac97/ac97_codec.c:1416: error: structure has no member named 
`build_ops'
sound/pci/ac97/ac97_codec.c:1417: error: structure has no member named 
`build_ops'
sound/pci/ac97/ac97_codec.c:1423: error: structure has no member named 
`build_ops'
sound/pci/ac97/ac97_codec.c:1423: error: structure has no member named 
`build_ops'
sound/pci/ac97/ac97_codec.c:1424: error: structure has no member named 
`build_ops'
sound/pci/ac97/ac97_codec.c:1435: error: structure has no member named 
`build_ops'
sound/pci/ac97/ac97_codec.c:1435: error: structure has no member named 
`build_ops'
sound/pci/ac97/ac97_codec.c:1436: error: structure has no member named 
`build_ops'
make[3]: *** [sound/pci/ac97/ac97_codec.o] Error 1
make[2]: *** [sound/pci/ac97] Error 2
make[1]: *** [sound/pci] Error 2
make: *** [sound] Error 2
------------------------------------------------------------







-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
same time. Free trial click here: http://www.vmware.com/wl/offer/345/0

Re: [OOPS] Alsa 0.9.4 / kernel 2.6.0-test1-mm1

[Jaroslav Kysela]

On Fri, 18 Jul 2003, Vince wrote:

> P.S.: I intended to test alsa 0.9.5 (in the hope that this bug was 
> already known & has already been fixed), but unfortunately I get a 
> compile error (I simply dropped alsa-kernel 0.9.5 in the 2.6.0-test1-mm1 
> source tree):

Copy also header files from alsa-kernel/include to linux/include/sound.

						Jaroslav

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



-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
same time. Free trial click here: http://www.vmware.com/wl/offer/345/0

Re: [OOPS] Alsa 0.9.4 / kernel 2.6.0-test1-mm1

[Vince]

Jaroslav Kysela wrote:
> On Fri, 18 Jul 2003, Vince wrote:
> 
> 
>>P.S.: I intended to test alsa 0.9.5 (in the hope that this bug was 
>>already known & has already been fixed), but unfortunately I get a 
>>compile error (I simply dropped alsa-kernel 0.9.5 in the 2.6.0-test1-mm1 
>>source tree):
> 
> 
> Copy also header files from alsa-kernel/include to linux/include/sound.
> 
> 						Jaroslav

Thanks... If I made no mistake in installation, it appears that I still 
get the exact same oops with 0.9.5 (btw alsa-kernel/include/version.h 
still says 0.9.4 in the tarball):

--------------------------------------------------------------------
Unable to handle kernel paging request at virtual address d0c38000
  printing eip:
d0acafbb
*pde = 0c404067
*pte = 00000000
Oops: 0000 [#1]
PREEMPT
CPU:    0
EIP:    0060:[<d0acafbb>]    Tainted: PF  VLI
EFLAGS: 00210202
EIP is at resample_expand+0x34b/0x380 [snd_pcm_oss]
eax: d0acafbb   ebx: 00000001   ecx: 000007ff   edx: ffecffec
esi: d0c55ff6   edi: cd6e3f4c   ebp: d0c37ffe   esp: c559fe58
ds: 007b   es: 007b   ss: 0068
Process festival (pid: 1381, threadinfo=c559e000 task=c606aca0)
Stack: d0ac8252 cd6e3ec0 c7f243a0 c559fe84 0000ffec c559e000 ffffffff 
cee9a600
        d0acade1 d0acaf32 d0acafbb cd6e3f30 00000000 00000002 00000002 
00000000
        ffecffec 000002a8 000017fd 00000800 cd6e3ec0 c7de3640 d0acb4ac 
cd6e3ec0
Call Trace:
  [<d0ac8252>] snd_pcm_plug_playback_channels_mask+0x72/0xe0 [snd_pcm_oss]
  [<d0acade1>] resample_expand+0x171/0x380 [snd_pcm_oss]
  [<d0acaf32>] resample_expand+0x2c2/0x380 [snd_pcm_oss]
  [<d0acafbb>] resample_expand+0x34b/0x380 [snd_pcm_oss]
  [<d0acb4ac>] rate_transfer+0x5c/0x60 [snd_pcm_oss]
  [<d0ac8647>] snd_pcm_plug_write_transfer+0x97/0x100 [snd_pcm_oss]
  [<d0ac43e0>] snd_pcm_oss_write2+0xd0/0x150 [snd_pcm_oss]
  [<d0ac4606>] snd_pcm_oss_write1+0x1a6/0x1d0 [snd_pcm_oss]
  [<d0ac6713>] snd_pcm_oss_write+0x43/0x60 [snd_pcm_oss]
  [<d0ac66d0>] snd_pcm_oss_write+0x0/0x60 [snd_pcm_oss]
  [<c014e6f8>] vfs_write+0xb8/0x130
  [<c014e822>] sys_write+0x42/0x70
  [<c0109107>] syscall_call+0x7/0xb

Code: 66 35 00 80 eb d4 8b 45 00 c1 e8 08 eb cc 8b 45 00 c1 e8 08 66 35 
00 80 eb c2 8b 45 00 c1 e8 08 eb ba eb a6 eb b1 8b 55 00 eb bf <8b> 45 
00 eb ac 0f b6 45 00 c1 e0 08 eb a3 81 fa 00 80 00 00 b8
-----------------------------------------------------------------------

The tainted flag comes from the nvidia binary driver, but I can 
reproduce without if necessary...

Vincent



-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
same time. Free trial click here: http://www.vmware.com/wl/offer/345/0

Re: [OOPS] Alsa 0.9.4 / kernel 2.6.0-test1-mm1

[Jaroslav Kysela]

On Sat, 19 Jul 2003, Vince wrote:

> Jaroslav Kysela wrote:
> > On Fri, 18 Jul 2003, Vince wrote:
> > 
> > 
> >>P.S.: I intended to test alsa 0.9.5 (in the hope that this bug was 
> >>already known & has already been fixed), but unfortunately I get a 
> >>compile error (I simply dropped alsa-kernel 0.9.5 in the 2.6.0-test1-mm1 
> >>source tree):
> > 
> > 
> > Copy also header files from alsa-kernel/include to linux/include/sound.
> > 
> > 						Jaroslav
> 
> Thanks... If I made no mistake in installation, it appears that I still 
> get the exact same oops with 0.9.5 (btw alsa-kernel/include/version.h 
> still says 0.9.4 in the tarball):

Please, send us contents of /proc/asound/card0/pcm0p/sub0/* files after 
oops occured.

						Jaroslav

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



-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
same time. Free trial click here: http://www.vmware.com/wl/offer/345/0

Re: [OOPS] Alsa 0.9.4 / kernel 2.6.0-test1-mm1

[Vince]

Jaroslav Kysela wrote:
> On Sat, 19 Jul 2003, Vince wrote:
> 
> 
>>Jaroslav Kysela wrote:
>>
>>>On Fri, 18 Jul 2003, Vince wrote:
>>>
>>>
>>>
>>>>P.S.: I intended to test alsa 0.9.5 (in the hope that this bug was 
>>>>already known & has already been fixed), but unfortunately I get a 
>>>>compile error (I simply dropped alsa-kernel 0.9.5 in the 2.6.0-test1-mm1 
>>>>source tree):
>>>
>>>
>>>Copy also header files from alsa-kernel/include to linux/include/sound.
>>>
>>>						Jaroslav
>>
>>Thanks... If I made no mistake in installation, it appears that I still 
>>get the exact same oops with 0.9.5 (btw alsa-kernel/include/version.h 
>>still says 0.9.4 in the tarball):
> 
> 
> Please, send us contents of /proc/asound/card0/pcm0p/sub0/* files after 
> oops occured.

Here it is right after the oops:

diablo:~$ cat /proc/asound/card0/pcm0p/sub0/hw_params
closed
diablo:~$ cat /proc/asound/card0/pcm0p/sub0/info
card: 0
device: 0
subdevice: 0
stream: PLAYBACK
id: ES1371/1
name: ES1371 DAC2/ADC
subname: subdevice #0
class: 0
subclass: 0
subdevices_count: 1
subdevices_avail: 1
diablo:~$ cat /proc/asound/card0/pcm0p/sub0/prealloc
64
diablo:~$ cat /proc/asound/card0/pcm0p/sub0/status
closed
diablo:~$ cat /proc/asound/card0/pcm0p/sub0/sw_params
closed

Information that may help : it looks like festival tries to play a 
16kHz/16bit/mono file trough oss... If I redirect its output to "aplay 
-f S16_LE -r 16000", everything works fine.

Vincent



-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
same time. Free trial click here: http://www.vmware.com/wl/offer/345/0

Re: [OOPS] Alsa 0.9.4 / kernel 2.6.0-test1-mm1

[Jaroslav Kysela]

On Mon, 21 Jul 2003, Vince wrote:

> Jaroslav Kysela wrote:
> > On Sat, 19 Jul 2003, Vince wrote:
> > 
> > 
> >>Jaroslav Kysela wrote:
> >>
> >>>On Fri, 18 Jul 2003, Vince wrote:
> >>>
> >>>
> >>>
> >>>>P.S.: I intended to test alsa 0.9.5 (in the hope that this bug was 
> >>>>already known & has already been fixed), but unfortunately I get a 
> >>>>compile error (I simply dropped alsa-kernel 0.9.5 in the 2.6.0-test1-mm1 
> >>>>source tree):
> >>>
> >>>
> >>>Copy also header files from alsa-kernel/include to linux/include/sound.
> >>>
> >>>						Jaroslav
> >>
> >>Thanks... If I made no mistake in installation, it appears that I still 
> >>get the exact same oops with 0.9.5 (btw alsa-kernel/include/version.h 
> >>still says 0.9.4 in the tarball):
> > 
> > 
> > Please, send us contents of /proc/asound/card0/pcm0p/sub0/* files after 
> > oops occured.
> 
> Here it is right after the oops:

Could you try the following patch? Note that this patch is a workaround 
for your hardware, but it does not fix the oops for other users (like 
the intel8x0 driver). I will work on this problem.

						Jaroslav

Index: pcm_plugin.c
===================================================================
RCS file: /cvsroot/alsa/alsa-kernel/core/oss/pcm_plugin.c,v
retrieving revision 1.10
retrieving revision 1.11
diff -u -r1.10 -r1.11
--- pcm_plugin.c	7 Feb 2003 08:29:08 -0000	1.10
+++ pcm_plugin.c	21 Jul 2003 12:36:10 -0000	1.11
@@ -56,6 +56,17 @@
 	return 0;
 }
 
+/*
+ *  because some cards might have rates "very close", we ignore
+ *  all "resampling" requests within +-5%
+ */
+static int rate_match(unsigned int src_rate, unsigned int dst_rate)
+{
+	unsigned int low = (src_rate * 95) / 100;
+	unsigned int high = (src_rate * 105) / 105;
+	return dst_rate >= low && dst_rate <= high;
+}
+
 static int snd_pcm_plugin_alloc(snd_pcm_plugin_t *plugin, snd_pcm_uframes_t frames)
 {
 	snd_pcm_plugin_format_t *format;
@@ -420,7 +431,7 @@
 
 	/* Format change (linearization) */
 	if ((srcformat.format != dstformat.format ||
-	     srcformat.rate != dstformat.rate ||
+	     !rate_match(srcformat.rate, dstformat.rate) ||
 	     srcformat.channels != dstformat.channels) &&
 	    !snd_pcm_format_linear(srcformat.format)) {
 		if (snd_pcm_format_linear(dstformat.format))
@@ -468,7 +479,7 @@
 				ttable[v * sv + v] = FULL;
 		}
 		tmpformat.channels = dstformat.channels;
-		if (srcformat.rate == dstformat.rate &&
+		if (rate_match(srcformat.rate, dstformat.rate) &&
 		    snd_pcm_format_linear(dstformat.format))
 			tmpformat.format = dstformat.format;
 		err = snd_pcm_plugin_build_route(plug,
@@ -490,7 +501,7 @@
 	}
 
 	/* rate resampling */
-	if (srcformat.rate != dstformat.rate) {
+	if (!rate_match(srcformat.rate, dstformat.rate)) {
 		tmpformat.rate = dstformat.rate;
 		if (srcformat.channels == dstformat.channels &&
 		    snd_pcm_format_linear(dstformat.format))

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



-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
same time. Free trial click here: http://www.vmware.com/wl/offer/345/0

Re: [OOPS] Alsa 0.9.4 / kernel 2.6.0-test1-mm1

[Ville Syrjälä]

On Mon, Jul 21, 2003 at 02:38:03PM +0200, Jaroslav Kysela wrote:
> +static int rate_match(unsigned int src_rate, unsigned int dst_rate)
> +{
> +	unsigned int low = (src_rate * 95) / 100;
> +	unsigned int high = (src_rate * 105) / 105;
                                               ^^^
Typo.

-- 
Ville Syrjälä
syrjala@sci.fi
http://www.sci.fi/~syrjala/


-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
same time. Free trial click here: http://www.vmware.com/wl/offer/345/0

Re: [OOPS] Alsa 0.9.4 / kernel 2.6.0-test1-mm1

[Jaroslav Kysela]

On Mon, 21 Jul 2003, [iso-8859-1] Ville Syrjälä wrote:

> On Mon, Jul 21, 2003 at 02:38:03PM +0200, Jaroslav Kysela wrote:
> > +static int rate_match(unsigned int src_rate, unsigned int dst_rate)
> > +{
> > +	unsigned int low = (src_rate * 95) / 100;
> > +	unsigned int high = (src_rate * 105) / 105;
>                                                ^^^
> Typo.

Yes, thanks.

						Jaroslav

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



-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
same time. Free trial click here: http://www.vmware.com/wl/offer/345/0

Re: [OOPS] Alsa 0.9.4 / kernel 2.6.0-test1-mm1

[Vince]

Sorry, I still get the oops with the patch...

--------------------------------------------------------------------
Unable to handle kernel paging request at virtual address d0c19000
  printing eip:
d0a9f07b
*pde = 0c03e067
*pte = 00000000
Oops: 0000 [#1]
PREEMPT
CPU:    0
EIP:    0060:[<d0a9f07b>]    Tainted: PF  VLI
EFLAGS: 00210202
EIP is at resample_expand+0x34b/0x380 [snd_pcm_oss]
eax: d0a9f07b   ebx: 00000001   ecx: 00051e16   edx: fca8fd4b
esi: d0c23ff6   edi: cc4bf20c   ebp: d0c18ffe   esp: c7e37e58
ds: 007b   es: 007b   ss: 0068
Process festival (pid: 1271, threadinfo=c7e36000 task=cdcdaca0)
Stack: d0a9c2f2 cc4bf180 cc3239e0 c7e37e84 0000fd4a c7e36000 ffffffff 
ceac5600
        d0a9eea1 d0a9eff2 d0a9f07b cc4bf1f0 00000000 00000002 00000002 
00000000
        fd4bfd4b 000002a8 000017fd 00000800 cc4bf180 cc053a40 d0a9f56c 
cc4bf180
Call Trace:
  [<d0a9c2f2>] snd_pcm_plug_playback_channels_mask+0x72/0xe0 [snd_pcm_oss]
  [<d0a9eea1>] resample_expand+0x171/0x380 [snd_pcm_oss]
  [<d0a9eff2>] resample_expand+0x2c2/0x380 [snd_pcm_oss]
  [<d0a9f07b>] resample_expand+0x34b/0x380 [snd_pcm_oss]
  [<d0a9f56c>] rate_transfer+0x5c/0x60 [snd_pcm_oss]
  [<d0a9c6e7>] snd_pcm_plug_write_transfer+0x97/0x100 [snd_pcm_oss]
  [<d0a983e0>] snd_pcm_oss_write2+0xd0/0x150 [snd_pcm_oss]
  [<d0a98606>] snd_pcm_oss_write1+0x1a6/0x1d0 [snd_pcm_oss]
  [<d0a9a713>] snd_pcm_oss_write+0x43/0x60 [snd_pcm_oss]
  [<d0a9a6d0>] snd_pcm_oss_write+0x0/0x60 [snd_pcm_oss]
  [<c014e6f8>] vfs_write+0xb8/0x130
  [<c014e822>] sys_write+0x42/0x70
  [<c0109107>] syscall_call+0x7/0xb

Code: 66 35 00 80 eb d4 8b 45 00 c1 e8 08 eb cc 8b 45 00 c1 e8 08 66 35 
00 80 eb c2 8b 45 00 c1 e8 08 eb ba eb a6 eb b1 8b 55 00 eb bf <8b> 45 
00 eb ac 0f b6 45 00 c1 e0 08 eb a3 81 fa 00 80 00 00 b8
  <6>


> 
> Could you try the following patch? Note that this patch is a workaround 
> for your hardware, but it does not fix the oops for other users (like 
> the intel8x0 driver). I will work on this problem.
> 
> 						Jaroslav
> 
> Index: pcm_plugin.c
> ===================================================================
> RCS file: /cvsroot/alsa/alsa-kernel/core/oss/pcm_plugin.c,v
> retrieving revision 1.10
> retrieving revision 1.11
> diff -u -r1.10 -r1.11
> --- pcm_plugin.c	7 Feb 2003 08:29:08 -0000	1.10
> +++ pcm_plugin.c	21 Jul 2003 12:36:10 -0000	1.11
> @@ -56,6 +56,17 @@
>  	return 0;
>  }
>  
> +/*
> + *  because some cards might have rates "very close", we ignore
> + *  all "resampling" requests within +-5%
> + */
> +static int rate_match(unsigned int src_rate, unsigned int dst_rate)
> +{
> +	unsigned int low = (src_rate * 95) / 100;
> +	unsigned int high = (src_rate * 105) / 105;
> +	return dst_rate >= low && dst_rate <= high;
> +}
> +
>  static int snd_pcm_plugin_alloc(snd_pcm_plugin_t *plugin, snd_pcm_uframes_t frames)
>  {
>  	snd_pcm_plugin_format_t *format;
> @@ -420,7 +431,7 @@
>  
>  	/* Format change (linearization) */
>  	if ((srcformat.format != dstformat.format ||
> -	     srcformat.rate != dstformat.rate ||
> +	     !rate_match(srcformat.rate, dstformat.rate) ||
>  	     srcformat.channels != dstformat.channels) &&
>  	    !snd_pcm_format_linear(srcformat.format)) {
>  		if (snd_pcm_format_linear(dstformat.format))
> @@ -468,7 +479,7 @@
>  				ttable[v * sv + v] = FULL;
>  		}
>  		tmpformat.channels = dstformat.channels;
> -		if (srcformat.rate == dstformat.rate &&
> +		if (rate_match(srcformat.rate, dstformat.rate) &&
>  		    snd_pcm_format_linear(dstformat.format))
>  			tmpformat.format = dstformat.format;
>  		err = snd_pcm_plugin_build_route(plug,
> @@ -490,7 +501,7 @@
>  	}
>  
>  	/* rate resampling */
> -	if (srcformat.rate != dstformat.rate) {
> +	if (!rate_match(srcformat.rate, dstformat.rate)) {
>  		tmpformat.rate = dstformat.rate;
>  		if (srcformat.channels == dstformat.channels &&
>  		    snd_pcm_format_linear(dstformat.format))
> 
> -----
> Jaroslav Kysela <perex@suse.cz>
> Linux Kernel Sound Maintainer
> ALSA Project, SuSE Labs
> 
> 
> 
> -------------------------------------------------------
> This SF.net email is sponsored by: VM Ware
> With VMware you can run multiple operating systems on a single machine.
> WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
> same time. Free trial click here: http://www.vmware.com/wl/offer/345/0
> _______________________________________________
> Alsa-devel mailing list
> Alsa-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/alsa-devel




-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
same time. Free trial click here: http://www.vmware.com/wl/offer/345/0

Re: [OOPS] Alsa 0.9.4 / kernel 2.6.0-test1-mm1

[Jaroslav Kysela]

On Mon, 21 Jul 2003, Vince wrote:

> Sorry, I still get the oops with the patch...

Please, change '(src_rate * 105) / 105;' expression from the patch
to '(src_rate * 105) / 100;'...

						Jaroslav

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



-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
same time. Free trial click here: http://www.vmware.com/wl/offer/345/0

Re: [OOPS] Alsa 0.9.4 / kernel 2.6.0-test1-mm1

[Vince]

And I also still get an oops after correcting the patch...

Unable to handle kernel paging request at virtual address d0a91000
  printing eip:
d0c8607b
*pde = 013ea067
*pte = 00000000
Oops: 0000 [#1]
PREEMPT
CPU:    0
EIP:    0060:[<d0c8607b>]    Tainted: PF  VLI
EFLAGS: 00010202
EIP is at resample_expand+0x34b/0x380 [snd_pcm_oss]
eax: d0c8607b   ebx: 00000001   ecx: 00036eb8   edx: fdc6fe33
esi: d0acbff6   edi: cee768cc   ebp: d0a90ffe   esp: c5979e58
ds: 007b   es: 007b   ss: 0068
Process festival (pid: 2322, threadinfo=c5978000 task=c565f940)
Stack: d0c832f2 cee76840 c4f528e0 c5979e84 0000fe32 c5978000 ffffffff 
cee76600
        d0c85ea1 d0c85ff2 d0c8607b cee768b0 00000000 00000002 00000002 
00000000
        fe33fe33 000002a8 000017fd 00000800 cee76840 cebe3140 d0c8656c 
cee76840
Call Trace:
  [<d0c832f2>] snd_pcm_plug_playback_channels_mask+0x72/0xe0 [snd_pcm_oss]
  [<d0c85ea1>] resample_expand+0x171/0x380 [snd_pcm_oss]
  [<d0c85ff2>] resample_expand+0x2c2/0x380 [snd_pcm_oss]
  [<d0c8607b>] resample_expand+0x34b/0x380 [snd_pcm_oss]
  [<d0c8656c>] rate_transfer+0x5c/0x60 [snd_pcm_oss]
  [<d0c836e7>] snd_pcm_plug_write_transfer+0x97/0x100 [snd_pcm_oss]
  [<d0c7f3e0>] snd_pcm_oss_write2+0xd0/0x150 [snd_pcm_oss]
  [<d0c7f606>] snd_pcm_oss_write1+0x1a6/0x1d0 [snd_pcm_oss]
  [<d0c81713>] snd_pcm_oss_write+0x43/0x60 [snd_pcm_oss]
  [<d0c816d0>] snd_pcm_oss_write+0x0/0x60 [snd_pcm_oss]
  [<c014e6f8>] vfs_write+0xb8/0x130
  [<c014e822>] sys_write+0x42/0x70
  [<c0109107>] syscall_call+0x7/0xb

Code: 66 35 00 80 eb d4 8b 45 00 c1 e8 08 eb cc 8b 45 00 c1 e8 08 66 35 
00 80 eb  c2 8b 45 00 c1 e8 08 eb ba eb a6 eb b1 8b 55 00 eb bf <8b> 45 
00 eb ac 0f b6 45  00 c1 e0 08 eb a3 81 fa 00 80 00 00 b8



Jaroslav Kysela wrote:
> On Mon, 21 Jul 2003, [iso-8859-1] Ville Syrjälä wrote:
> 
> 
>>On Mon, Jul 21, 2003 at 02:38:03PM +0200, Jaroslav Kysela wrote:
>>
>>>+static int rate_match(unsigned int src_rate, unsigned int dst_rate)
>>>+{
>>>+	unsigned int low = (src_rate * 95) / 100;
>>>+	unsigned int high = (src_rate * 105) / 105;
>>
>>                                               ^^^
>>Typo.
> 
> 
> Yes, thanks.
> 
> 			



-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
same time. Free trial click here: http://www.vmware.com/wl/offer/345/0

Re: [OOPS] Alsa 0.9.4 / kernel 2.6.0-test1-mm1

[Takashi Iwai]

At Mon, 21 Jul 2003 15:32:17 +0200,
Vince wrote:
> 
> Sorry, I still get the oops with the patch...

perhaps you built the modules without frame pointer?


Takashi


-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
same time. Free trial click here: http://www.vmware.com/wl/offer/345/0

Re: [OOPS] Alsa 0.9.4 / kernel 2.6.0-test1-mm1

[Jaroslav Kysela]

On Mon, 21 Jul 2003, Jaroslav Kysela wrote:

> On Mon, 21 Jul 2003, Vince wrote:
> 
> > Sorry, I still get the oops with the patch...
> 
> Please, change '(src_rate * 105) / 105;' expression from the patch
> to '(src_rate * 105) / 100;'...

Ok, here is also a patch which should solve the error for all cases:

Index: pcm_plugin.c
===================================================================
RCS file: /cvsroot/alsa/alsa-kernel/core/oss/pcm_plugin.c,v
retrieving revision 1.12
diff -u -r1.12 pcm_plugin.c
--- pcm_plugin.c	21 Jul 2003 13:28:27 -0000	1.12
+++ pcm_plugin.c	21 Jul 2003 14:07:42 -0000
@@ -96,6 +96,7 @@
 	c = plugin->buf_channels;
 	if (plugin->access == SNDRV_PCM_ACCESS_RW_INTERLEAVED) {
 		for (channel = 0; channel < format->channels; channel++, c++) {
+			c->frames = frames / format->channels;
 			c->enabled = 1;
 			c->wanted = 0;
 			c->area.addr = plugin->buf;
@@ -106,6 +107,7 @@
 		snd_assert((size % format->channels) == 0,);
 		size /= format->channels;
 		for (channel = 0; channel < format->channels; channel++, c++) {
+			c->frames = frames / format->channels;
 			c->enabled = 1;
 			c->wanted = 0;
 			c->area.addr = plugin->buf + (channel * size);
Index: pcm_plugin.h
===================================================================
RCS file: /cvsroot/alsa/alsa-kernel/core/oss/pcm_plugin.h,v
retrieving revision 1.3
diff -u -r1.3 pcm_plugin.h
--- pcm_plugin.h	26 Jun 2002 02:02:49 -0000	1.3
+++ pcm_plugin.h	21 Jul 2003 14:07:42 -0000
@@ -106,6 +106,7 @@
 typedef struct _snd_pcm_plugin_channel {
 	void *aptr;			/* pointer to the allocated area */
 	snd_pcm_channel_area_t area;
+	snd_pcm_uframes_t frames;	/* allocated frames */
 	unsigned int enabled:1;		/* channel need to be processed */
 	unsigned int wanted:1;		/* channel is wanted */
 } snd_pcm_plugin_channel_t;
Index: rate.c
===================================================================
RCS file: /cvsroot/alsa/alsa-kernel/core/oss/rate.c,v
retrieving revision 1.7
diff -u -r1.7 rate.c
--- rate.c	12 Aug 2002 08:43:45 -0000	1.7
+++ rate.c	21 Jul 2003 14:07:42 -0000
@@ -318,6 +318,8 @@
 #endif
 
 	dst_frames = rate_dst_frames(plugin, frames);
+	if (dst_frames > dst_channels[0].frames)
+		dst_frames = dst_channels[0].frames;
 	data = (rate_t *)plugin->extra_data;
 	data->func(plugin, src_channels, dst_channels, frames, dst_frames);
 	return dst_frames;

						Jaroslav

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



-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
same time. Free trial click here: http://www.vmware.com/wl/offer/345/0

Re: [OOPS] Alsa 0.9.4 / kernel 2.6.0-test1-mm1

[Vince]

Takashi Iwai wrote:
> At Mon, 21 Jul 2003 15:32:17 +0200,
> Vince wrote:
> 
>>Sorry, I still get the oops with the patch...
> 
> 
> perhaps you built the modules without frame pointer?
> 
> 

Indeed:

diablo:/usr/src/linux-2.6.0test1# grep FRAME .config
# CONFIG_FRAME_POINTER is not set

Should I make a build with framepointers ?




-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
same time. Free trial click here: http://www.vmware.com/wl/offer/345/0

Re: [OOPS] Alsa 0.9.4 / kernel 2.6.0-test1-mm1

[Takashi Iwai]

At Mon, 21 Jul 2003 16:11:28 +0200,
Vince wrote:
> 
> Takashi Iwai wrote:
> > At Mon, 21 Jul 2003 15:32:17 +0200,
> > Vince wrote:
> > 
> >>Sorry, I still get the oops with the patch...
> > 
> > 
> > perhaps you built the modules without frame pointer?
> > 
> > 
> 
> Indeed:
> 
> diablo:/usr/src/linux-2.6.0test1# grep FRAME .config
> # CONFIG_FRAME_POINTER is not set
> 
> Should I make a build with framepointers ?

yep, try once.  i've heard this fixes this problem.
(yeah i know we need to hunt this bug... :)


Takashi


-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
same time. Free trial click here: http://www.vmware.com/wl/offer/345/0

Re: [OOPS] Alsa 0.9.4 / kernel 2.6.0-test1-mm1

[Vince]

Jaroslav Kysela wrote:
> On Mon, 21 Jul 2003, Vince wrote:
> 
> 
>>Sorry, I still get the oops with the patch...
> 
> 
> Please, change '(src_rate * 105) / 105;' expression from the patch
> to '(src_rate * 105) / 100;'...
> 

I already did (see my last mail)... and still the oops. :-(



-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
same time. Free trial click here: http://www.vmware.com/wl/offer/345/0

RE: compilation

[Denis, jMarc]

Hi i forgot to tell you my configuration :

[root@localhost alsa-oss-0.9.4]# rpm -qa |grep alsa
alsa-lib-0.9.0-fr0.rc7.0.rh73.1
alsa-driver-0.9.0-fr0.rc7.0.rh73.2
alsa-utils-0.9.0-fr0.rc7.0.rh73.1

[root@localhost alsa-oss-0.9.4]# locate libasound
/usr/lib/libasound.so
/usr/lib/libasound.so.2
/usr/lib/libasound.so.2.0.0

And when i run ./configure i have this problem :

checking for libasound headers version >= 0.9.0... not present.
configure: error: Sufficiently new version of libasound not found.


Anyone can help me please ?
TU.

jM.


-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
same time. Free trial click here: http://www.vmware.com/wl/offer/345/0

Re: [OOPS] Alsa 0.9.4 / kernel 2.6.0-test1-mm1

[Vince]

Jaroslav Kysela wrote:
> On Mon, 21 Jul 2003, Jaroslav Kysela wrote:
> 
> 
>>On Mon, 21 Jul 2003, Vince wrote:
>>
>>
>>>Sorry, I still get the oops with the patch...
>>
>>Please, change '(src_rate * 105) / 105;' expression from the patch
>>to '(src_rate * 105) / 100;'...
> 
> 
> Ok, here is also a patch which should solve the error for all cases:
> 

Well... I still get an oops :-(
I'm doing a build with frame pointers now... Will tell you the result in 
a few minutes.

Unable to handle kernel paging request at virtual address d0a8a000
  printing eip:
d0c850bb
*pde = 013ea067
*pte = 00000000
Oops: 0000 [#1]
PREEMPT
CPU:    0
EIP:    0060:[<d0c850bb>]    Tainted: PF  VLI
EFLAGS: 00210202
EIP is at resample_expand+0x34b/0x380 [snd_pcm_oss]
eax: d0c850bb   ebx: 00000001   ecx: 00051e16   edx: fca8fd4b
esi: d0a97ff6   edi: cc44a80c   ebp: d0a89ffe   esp: ccc35e54
ds: 007b   es: 007b   ss: 0068
Process festival (pid: 2305, threadinfo=ccc34000 task=cbfacd20)
Stack: ccc35e84 d0c82342 cc44a780 c7e4b2e0 ccc3fd4a 00000000 ccc34000 
ffffffff
        d0c84ee1 d0c85032 d0c850bb cc44a7f0 00000000 00000002 00000002 
00000000
        fd4bfd4b 000002a8 000017fd 00000800 cc44a780 c7e4b220 d0c855b9 
cc44a780
Call Trace:
  [<d0c82342>] snd_pcm_plug_playback_channels_mask+0x72/0xe0 [snd_pcm_oss]
  [<d0c84ee1>] resample_expand+0x171/0x380 [snd_pcm_oss]
  [<d0c85032>] resample_expand+0x2c2/0x380 [snd_pcm_oss]
  [<d0c850bb>] resample_expand+0x34b/0x380 [snd_pcm_oss]
  [<d0c855b9>] rate_transfer+0x69/0x70 [snd_pcm_oss]
  [<d0c82747>] snd_pcm_plug_write_transfer+0x97/0x100 [snd_pcm_oss]
  [<d0c7e3e0>] snd_pcm_oss_write2+0xd0/0x150 [snd_pcm_oss]
  [<d0c7e606>] snd_pcm_oss_write1+0x1a6/0x1d0 [snd_pcm_oss]
  [<d0c80713>] snd_pcm_oss_write+0x43/0x60 [snd_pcm_oss]
  [<d0c806d0>] snd_pcm_oss_write+0x0/0x60 [snd_pcm_oss]
  [<c014e6f8>] vfs_write+0xb8/0x130
  [<c014e822>] sys_write+0x42/0x70
  [<c0109107>] syscall_call+0x7/0xb

Code: 66 35 00 80 eb d4 8b 45 00 c1 e8 08 eb cc 8b 45 00 c1 e8 08 66 35 
00 80 eb c2 8b 45 00 c1 e8 08 eb ba eb a6 eb b1 8b 55 00 eb bf <8b> 45 
00 eb ac 0f b6 45 00 c1 e0 08 eb a3 81 fa 00 80 00 00 b8






-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
same time. Free trial click here: http://www.vmware.com/wl/offer/345/0

Re: [OOPS] Alsa 0.9.4 / kernel 2.6.0-test1-mm1

[Vince]

Indeed, I do not have the bug with the frame pointer buid... Perhaps 
this is a compiler problem ? (mine is gcc version 3.3.1 20030626 (Debian 
prerelease))
Anyway, I'm still willing to help solve this problem...

Vincent

Takashi Iwai wrote:
> At Mon, 21 Jul 2003 16:11:28 +0200,
> Vince wrote:
> 
>>Takashi Iwai wrote:
>>
>>>At Mon, 21 Jul 2003 15:32:17 +0200,
>>>Vince wrote:
>>>
>>>
>>>>Sorry, I still get the oops with the patch...
>>>
>>>
>>>perhaps you built the modules without frame pointer?
>>>
>>>
>>
>>Indeed:
>>
>>diablo:/usr/src/linux-2.6.0test1# grep FRAME .config
>># CONFIG_FRAME_POINTER is not set
>>
>>Should I make a build with framepointers ?
> 
> 
> yep, try once.  i've heard this fixes this problem.
> (yeah i know we need to hunt this bug... :)
> 







-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
same time. Free trial click here: http://www.vmware.com/wl/offer/345/0