Problems with VIA8233A (todays cvs)

Problems with VIA8233A (todays cvs)

[Joerg Mayer]

[-- Attachment #1: Type: text/plain, Size: 2134 bytes --]

Hello,

I'm having trouble with a motherboard that has a via8233 chipset
(msi kt3 Ultra series, MS-6380E v1.X). Sound is sort of working, but
not quite perfect: when using winex, sound effects stutter, sometimes
it "übersteuert" (I don't know the English word :) and some not so nice
messages are displayed:

kernel: PCI: Setting latency timer of device 00:11.5 to 64
kernel: ALSA ../alsa-kernel/pci/via82xx.c:687: invalid via82xx_cur_ptr, using last valid pointer
last message repeated 8 times
modprobe: modprobe: Can't locate module sound-slot-1
modprobe: modprobe: Can't locate module sound-service-1-3
modprobe: modprobe: Can't locate module sound-slot-2
modprobe: modprobe: Can't locate module sound-service-2-3
modprobe: modprobe: Can't locate module sound-slot-3
modprobe: modprobe: Can't locate module sound-service-3-3
modprobe: modprobe: Can't locate module sound-slot-1
modprobe: modprobe: Can't locate module sound-service-1-3
modprobe: modprobe: Can't locate module sound-slot-2
modprobe: modprobe: Can't locate module sound-service-2-3
modprobe: modprobe: Can't locate module sound-slot-3
modprobe: modprobe: Can't locate module sound-service-3-3
kernel: ALSA ../alsa-kernel/pci/via82xx.c:687: invalid via82xx_cur_ptr, using last valid pointer
last message repeated 205 times
kernel: ALSA ../../../alsa-kernel/core/seq/oss/seq_oss_init.c:213: no device found
kernel: ALSA ../alsa-kernel/core/memory.c:71: Not freed snd_alloc_kmalloc = 820
kernel: ALSA ../alsa-kernel/core/memory.c:81: kmalloc(820) from d40f8827 not freed

This is with SuSE 8.2 with all patches applied (including k_deflt-2.4.20-96).
All alsa-rpms were removed (with --nodeps where necessary), and alsa built
from cvs (updated after 18:00 MEST today).

Some more stuff (e.g. contents of /proc/interesting-stuff) is attached
to this mail.

Ideas how to fix the sound problems?

Thanks
    Jörg
-- 
Joerg Mayer                                           <jmayer@loplof.de>
We are stuck with technology when what we really want is just stuff that
works. Some say that should read Microsoft instead of technology.

[-- Attachment #2: alsa-murx.tar.bz2 --]
[-- Type: application/x-bunzip2, Size: 5353 bytes --]

Re: Problems with VIA8233A (todays cvs)

[Takashi Iwai]

At Tue, 22 Jul 2003 20:26:49 +0200,
Joerg Mayer wrote:
> 
> Hello,
> 
> I'm having trouble with a motherboard that has a via8233 chipset
> (msi kt3 Ultra series, MS-6380E v1.X). Sound is sort of working, but
> not quite perfect: when using winex, sound effects stutter, sometimes
> it "�bersteuert" (I don't know the English word :) and some not so nice
> messages are displayed:
> 
> kernel: PCI: Setting latency timer of device 00:11.5 to 64
> kernel: ALSA ../alsa-kernel/pci/via82xx.c:687: invalid via82xx_cur_ptr, using last valid pointer
> last message repeated 8 times
> modprobe: modprobe: Can't locate module sound-slot-1
> modprobe: modprobe: Can't locate module sound-service-1-3
> modprobe: modprobe: Can't locate module sound-slot-2
> modprobe: modprobe: Can't locate module sound-service-2-3
> modprobe: modprobe: Can't locate module sound-slot-3
> modprobe: modprobe: Can't locate module sound-service-3-3
> modprobe: modprobe: Can't locate module sound-slot-1
> modprobe: modprobe: Can't locate module sound-service-1-3
> modprobe: modprobe: Can't locate module sound-slot-2
> modprobe: modprobe: Can't locate module sound-service-2-3
> modprobe: modprobe: Can't locate module sound-slot-3
> modprobe: modprobe: Can't locate module sound-service-3-3
> kernel: ALSA ../alsa-kernel/pci/via82xx.c:687: invalid via82xx_cur_ptr, using last valid pointer
> last message repeated 205 times
> kernel: ALSA ../../../alsa-kernel/core/seq/oss/seq_oss_init.c:213: no device found
> kernel: ALSA ../alsa-kernel/core/memory.c:71: Not freed snd_alloc_kmalloc = 820
> kernel: ALSA ../alsa-kernel/core/memory.c:81: kmalloc(820) from d40f8827 not freed
> 
> This is with SuSE 8.2 with all patches applied (including k_deflt-2.4.20-96).
> All alsa-rpms were removed (with --nodeps where necessary), and alsa built
> from cvs (updated after 18:00 MEST today).
> 
> Some more stuff (e.g. contents of /proc/interesting-stuff) is attached
> to this mail.
> 
> Ideas how to fix the sound problems?

try to add dxs_support=3 to snd-via82xx module option.
(or dxs_support=2 if it doesn't solve the problem.)

see alsa-kernel/Documentation/ALSA-Configuration.txt for details.

btw, before building the alsa-driver externally, it's better to remove
CONFIG_SND* in your kernel configuration
(/usr/src/linux/include/linux/autoconf.h).
otherwise there might be inconsistencies.


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: Problems with VIA8233A (todays cvs)

[Joerg Mayer]

On Tue, Jul 22, 2003 at 08:47:17PM +0200, Takashi Iwai wrote:
> At Tue, 22 Jul 2003 20:26:49 +0200,
> > kernel: ALSA ../alsa-kernel/pci/via82xx.c:687: invalid via82xx_cur_ptr, using last valid pointer
> > last message repeated 205 times

> try to add dxs_support=3 to snd-via82xx module option.
> (or dxs_support=2 if it doesn't solve the problem.)

Both options don't affect anything I would notice, but...

> see alsa-kernel/Documentation/ALSA-Configuration.txt for details.

OK, done now :)

> btw, before building the alsa-driver externally, it's better to remove
> CONFIG_SND* in your kernel configuration
> (/usr/src/linux/include/linux/autoconf.h).
> otherwise there might be inconsistencies.

...by doing this I got rid of the messages.

So the problem remains, that the sound is still not working properly, and
still with the same effects (except the kernel messages).
The sound in e.g. pysol plays , but at at least twice the speed it should
and with the distortions that are associated with it. Diablo 2 in winex
also still has the same problems.
Another thing the dxs options don't change is the fact that I can run
only one application using sound at a time, not several.
All of this worked with Suse 8.1, original kernel and one of your rpm-packages
at people/tiwai/ (don't remember the version though).


 Thanks
     Jörg
-- 
Joerg Mayer                                           <jmayer@loplof.de>
We are stuck with technology when what we really want is just stuff that
works. Some say that should read Microsoft instead of technology.


-------------------------------------------------------
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: Problems with VIA8233A (todays cvs)

[Takashi Iwai]

At Tue, 22 Jul 2003 22:02:31 +0200,
Joerg Mayer wrote:
> 
> On Tue, Jul 22, 2003 at 08:47:17PM +0200, Takashi Iwai wrote:
> > At Tue, 22 Jul 2003 20:26:49 +0200,
> > > kernel: ALSA ../alsa-kernel/pci/via82xx.c:687: invalid via82xx_cur_ptr, using last valid pointer
> > > last message repeated 205 times
> 
> > try to add dxs_support=3 to snd-via82xx module option.
> > (or dxs_support=2 if it doesn't solve the problem.)
> 
> Both options don't affect anything I would notice, but...

are you sure?
they must have big influence on the behavior.
dxs_support=2 doesn't work well on 0.9.5.  you should try cvs version
if dxs_support=3 doesn't solve the problem.


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: Problems with VIA8233A (todays cvs)

[Joerg Mayer]

On Wed, Jul 23, 2003 at 10:31:49AM +0200, Takashi Iwai wrote:
> At Tue, 22 Jul 2003 22:02:31 +0200,
> Joerg Mayer wrote:
> > 
> > On Tue, Jul 22, 2003 at 08:47:17PM +0200, Takashi Iwai wrote:
> > > At Tue, 22 Jul 2003 20:26:49 +0200,
> > > > kernel: ALSA ../alsa-kernel/pci/via82xx.c:687: invalid via82xx_cur_ptr, using last valid pointer
> > > > last message repeated 205 times
> > 
> > > try to add dxs_support=3 to snd-via82xx module option.
> > > (or dxs_support=2 if it doesn't solve the problem.)
> > 
> > Both options don't affect anything I would notice, but...
> 
> are you sure?
> they must have big influence on the behavior.
> dxs_support=2 doesn't work well on 0.9.5.  you should try cvs version
> if dxs_support=3 doesn't solve the problem.

Yes, I'm very sure. Just to summarize: My friend had problems with the alsa
binaries from Suse 8.2, so we updated the whole system with all patches
that where available from Suse, including the kernel. The problem was still
the same.  Next, we checked out alsa from cvs (co started yesterday (tuesday)
after 6pm), rebuilt everything and the problem was still there.
Then I sent the mail to alsa-devel.
You told me to a) test the dxs_support options and b) rebuild the kernel
without the SND_ entries in the autoconf.h file.
Playing with a) changed nothing, tested with /etc/init.d/alsasound restart.
Then I changed b), deleted all .o files, rebuilt and installed again.  After
that, the kernel messages were gone.
Playing with dxs_support afterwards did not resolve the problem.
The atachment to my first mail contains /proc/pci and some other output
(from the alsa-version provided with the updated suse 8.2).

There are two noticable problems: Music in Pysol plays much too fast, Music
in NWN (linux version) is jerky, as is the sound in DiabloII under WineX.
If you need new/other/updated infos or anything, I'm willing to provide them.

Oh, btw. I have some doubts that the chipset really is a 8233A chipset (the A).

Thanks
      Jörg


-- 
Joerg Mayer                                           <jmayer@loplof.de>
We are stuck with technology when what we really want is just stuff that
works. Some say that should read Microsoft instead of technology.


-------------------------------------------------------
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: Problems with VIA8233A (todays cvs)

[Takashi Iwai]

[-- Attachment #1: Type: text/plain, Size: 606 bytes --]

At Wed, 23 Jul 2003 14:58:22 +0200,
Joerg Mayer wrote:
> 
> Oh, btw. I have some doubts that the chipset really is a 8233A chipset (the A).

grrr, i overlooked that your chip is revision A.  sorry.
that explains why dxs_support options don't work: it's for 8233/8235
only but not 8233A.  (8233A has no DXS channels.)

could you try the attached patch for tests?
i don't expect much that this will solve the problem, but to be
sure...

if it's still too fast, it might be the ac97 clock problem like in
intel8x0 driver.  please try to set a lower value to ac97_clock module
option, such as 41200.


Takashi

[-- Attachment #2: via-check.dif --]
[-- Type: application/octet-stream, Size: 1216 bytes --]

Index: alsa-kernel/pci/via82xx.c
===================================================================
RCS file: /suse/tiwai/cvs/alsa/alsa-kernel/pci/via82xx.c,v
retrieving revision 1.49
diff -u -r1.49 via82xx.c
--- alsa-kernel/pci/via82xx.c	21 Jul 2003 14:54:26 -0000	1.49
+++ alsa-kernel/pci/via82xx.c	23 Jul 2003 13:04:57 -0000
@@ -886,9 +886,11 @@
 		snd_ac97_set_rate(chip->ac97, AC97_PCM_LFE_DAC_RATE, runtime->rate);
 		snd_ac97_set_rate(chip->ac97, AC97_SPDIF, runtime->rate);
 	}
+#if 0
 	if (chip->revision == VIA_REV_8233A)
 		rbits = 0;
 	else
+#endif
 		rbits = (0xfffff / 48000) * runtime->rate + ((0xfffff % 48000) * runtime->rate) / 48000;
 	snd_assert((rbits & ~0xfffff) == 0, return -EINVAL);
 	snd_via82xx_channel_reset(chip, viadev);
@@ -928,9 +930,12 @@
 	fmt = (runtime->format == SNDRV_PCM_FORMAT_S16_LE) ? VIA_REG_MULTPLAY_FMT_16BIT : VIA_REG_MULTPLAY_FMT_8BIT;
 	fmt |= runtime->channels << 4;
 	outb(fmt, VIADEV_REG(viadev, OFS_MULTPLAY_FORMAT));
+#if 0
 	if (chip->revision == VIA_REV_8233A)
 		slots = 0;
-	else {
+	else
+#endif
+	{
 		/* set sample number to slot 3, 4, 7, 8, 6, 9 (for VIA8233/C,8235) */
 		/* corresponding to FL, FR, RL, RR, C, LFE ?? */
 		switch (runtime->channels) {