Skip to contentSkip to conversation

2.6.0-test1+ Alsa + Intel 82801CA/CAM AC'97 Audio OOPS

DISCUSSION Simon Boulet 3 participants 5 messages Jul 19, 2003 – Jul 21, 2003 2 messages not in the archive

lore
Share & tools
lore

Simon Boulet

author
#0 lore raw

Hello everyone,

In case I am sending this to a list, please CC to me regarding anything related to this issue. I am not a member of the list.

I am having a Kernel OOPS with 2.6.0-test1-ac2 (same thing under non-
ac2) using ALSA with OSS compatibily enabled on an Intel 82801CA/CAM AC'97 (ICH3 mobile) integrated Audio.

I have activated "Intel i8x0/MX440, SiS 7012; Ali 5455; NForce Audio; AMD768/8111" ALSA support and both OSS mixer and PCM compatibility. I am attaching the oops.log and the ksymoops.log (not sure if the ksym thing is fine, new to this stuff) The non-compatibility mode seems to work fine though.

(ALSA driver kernel output bellow)

ALSA device list:
#0: Intel 82801CA-ICH3 at 0x1c00, irq 10

Also, the OSS (non-ALSA) Intel ICH (i8xx) loads correctly but the sound output is slow (rate or clocking problem?). My sound was fine under 2.4.21.

(OSS driver kernel output bellow)
i810_audio: Defaulting to base 2 channel mode.
i810_audio: Resetting connection 0
ac97_codec: AC97 Audio codec, id: 0x8384:0x7600 (SigmaTel STAC????)
i810_audio: only 48Khz playback available.
i810_audio: AC'97 codec 0 Unable to map surround DAC's (or DAC's not present), t otal channels = 2
i810_audio: setting clocking to 64937

I realy hope you can fix this in the next release and before -test is over. Please keep me informed of any patches and dont hesitate to ask if I can help.

Simon

ksymoops 2.4.8 on i686 2.6.0-test1-ac2-acpi20030714. Options used
-V (default)
-k /proc/ksyms (default)
-l /proc/modules (default)
-o /lib/modules/2.6.0-test1-ac2-acpi20030714/ (default) -m /boot/System.map-2.6.0-test1-ac2-acpi20030714 (default)

Warning: You did not tell me where to find symbol information. I will assume that the log matches the kernel and modules that are running right now and I'll use the default options above for symbol resolution. If the current kernel and/or modules do not match the log, you can get more accurate output by telling me the kernel version and where to find map, modules, ksyms etc. ksymoops -h explains the options.

Error (regular_file): read_ksyms stat /proc/ksyms failed
No modules in ksyms, skipping objects
No ksyms, skipping lsmod
Error (regular_file): read_system_map stat /boot/System.map-2.6.0-test1-ac2-acpi20030714 failed Warning (merge_maps): no symbols in merged map
<1>Unable to handle kernel paging request at virtual address d2877000
c02e5773

Oops: 0000 [#6]
CPU:    0
EIP:    0060:[<c02e5773>]    Not tainted

Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010202
eax: c02e5773 ebx: 00000000 ecx: 000007ff edx: 00000000
esi: d2879166 edi: ce591e10 ebp: d2876ffe esp: c61cbe54
ds: 007b es: 007b ss: 0068

Stack: c02e2bed ce591d80 cf668bc0 c61cbe80 00000000 00000000 cd113ca8 ffffffff
       c02e5597 c02e56eb c02e5773 ce591df0 00000000 00000004 00000004 00000001
       00000000 000003ee 0000045a 00000400 ce591d80 cf76fc00 c02e5c1d ce591d80
Call Trace:
 [<c02e2bed>] snd_pcm_plug_playback_channels_mask+0x72/0xd8
 [<c02e5597>] resample_expand+0x167/0x377
 [<c02e56eb>] resample_expand+0x2bb/0x377
 [<c02e5773>] resample_expand+0x343/0x377
 [<c02e5c1d>] rate_transfer+0x59/0x5d
 [<c02e2fc5>] snd_pcm_plug_write_transfer+0x95/0xf4
 [<c02df092>] snd_pcm_oss_write2+0xd0/0x13c
 [<c02df2a7>] snd_pcm_oss_write1+0x1a9/0x1d0
 [<c02e1193>] snd_pcm_oss_write+0x43/0x5d
 [<c02e1150>] snd_pcm_oss_write+0x0/0x5d
 [<c014e588>] vfs_write+0xb0/0x119
 [<c014e696>] sys_write+0x42/0x63
 [<c010911b>] syscall_call+0x7/0xb
Code: 8b 45 00 eb ac 0f b6 45 00 c1 e0 08 eb a3 81 fa 00 80 00 00
Quoted 1 line
Quoted 1 line

EIP; c02e5773 No symbols available <=====

Trace; c02e2bed No symbols available
Trace; c02e5597 No symbols available
Trace; c02e56eb No symbols available
Trace; c02e5773 No symbols available
Trace; c02e5c1d No symbols available
Trace; c02e2fc5 No symbols available
Trace; c02df092 No symbols available
Trace; c02df2a7 No symbols available
Trace; c02e1193 No symbols available
Trace; c02e1150 No symbols available
Trace; c014e588 No symbols available
Trace; c014e696 No symbols available
Trace; c010911b No symbols available

Code; c02e5773 No symbols available
00000000 <_EIP>:
Code; c02e5773 No symbols available <=====
0: 8b 45 00 mov 0x0(%ebp),%eax <=====
Code; c02e5776 No symbols available
3: eb ac jmp ffffffb1 <_EIP+0xffffffb1>
Code; c02e5778 No symbols available
5: 0f b6 45 00 movzbl 0x0(%ebp),%eax
Code; c02e577c No symbols available
9: c1 e0 08 shl $0x8,%eax
Code; c02e577f No symbols available
c: eb a3 jmp ffffffb1 <_EIP+0xffffffb1>
Code; c02e5781 No symbols available
e: 81 fa 00 80 00 00 cmp $0x8000,%edx

2 warnings and 2 errors issued. Results may not be reliable.

<1>Unable to handle kernel paging request at virtual address d2877000 printing eip:
c02e5773

Oops: 0000 [#6]
CPU:    0
EIP:    0060:[<c02e5773>]    Not tainted
EFLAGS: 00010202

EIP is at resample_expand+0x343/0x377
eax: c02e5773 ebx: 00000000 ecx: 000007ff edx: 00000000
esi: d2879166 edi: ce591e10 ebp: d2876ffe esp: c61cbe54
ds: 007b es: 007b ss: 0068
Process xmms (pid: 755, threadinfo=c61ca000 task=c9515380)

Stack: c02e2bed ce591d80 cf668bc0 c61cbe80 00000000 00000000 cd113ca8 ffffffff
       c02e5597 c02e56eb c02e5773 ce591df0 00000000 00000004 00000004 00000001
       00000000 000003ee 0000045a 00000400 ce591d80 cf76fc00 c02e5c1d ce591d80
Call Trace:
 [<c02e2bed>] snd_pcm_plug_playback_channels_mask+0x72/0xd8
 [<c02e5597>] resample_expand+0x167/0x377
 [<c02e56eb>] resample_expand+0x2bb/0x377
 [<c02e5773>] resample_expand+0x343/0x377
 [<c02e5c1d>] rate_transfer+0x59/0x5d
 [<c02e2fc5>] snd_pcm_plug_write_transfer+0x95/0xf4
 [<c02df092>] snd_pcm_oss_write2+0xd0/0x13c
 [<c02df2a7>] snd_pcm_oss_write1+0x1a9/0x1d0
 [<c02e1193>] snd_pcm_oss_write+0x43/0x5d
 [<c02e1150>] snd_pcm_oss_write+0x0/0x5d
 [<c014e588>] vfs_write+0xb0/0x119
 [<c014e696>] sys_write+0x42/0x63
 [<c010911b>] syscall_call+0x7/0xb

Code: 8b 45 00 eb ac 0f b6 45 00 c1 e0 08 eb a3 81 fa 00 80 00 00
#1 lore raw
Quoted 8 lines · Simon Boulet

Also, the OSS (non-ALSA) Intel ICH (i8xx) loads correctly but the sound output is slow (rate or clocking problem?). My sound was fine under 2.4.21.

i810_audio: only 48Khz playback available.

i810_audio: setting clocking to 64937

I was having problems with i810_audio clocking as well. It turned out to be the Intel Speedstep support, of all things. *IF* your kernel includes:

CONFIG_X86_SPEEDSTEP_ICH=y

it was possible to end up with a broken value for loops_per_jiffie. I've attached a patch that fixes the bug and does a few cleanups...

If you don't have the SpeedStep support in your kernel, then your problem is elsewhere... Good luck... ;)

/Valdis

i386/kernel/cpu/cpufreq/speedstep-ich.c renamed from i386/kernel/cpu/cpufreq/speedstep-ich.c.linus +10 -5
@@ -77,15 +77,17 @@
 	u8			value;
 	unsigned long		flags;
 	struct cpufreq_freqs	freqs;
+	int			newfreq;
 
 	if (!speedstep_chipset_dev || (state > 0x1))
 		return;
 
 	freqs.old = speedstep_get_processor_frequency(speedstep_processor);
-	freqs.new = speedstep_freqs[SPEEDSTEP_LOW].frequency;
+	freqs.new = speedstep_freqs[state].frequency;
 	freqs.cpu = 0; /* speedstep.c is UP only driver */
 	
-	if (notify)
+	/* make sure we've initialized before calling notify */
+	if (notify && (freqs.new != 0))
 		cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE);
 
 	/* get PMBASE */

@@ -136,13 +138,16 @@
 
 	dprintk(KERN_DEBUG "cpufreq: read at pmbase 0x%x + 0x50 returned 0x%x\n", pmbase, value);
 
+	/* freqs.new may not be set yet - need local copy */
+	newfreq = speedstep_get_processor_frequency(speedstep_processor);
 	if (state == (value & 0x1)) {
-		dprintk (KERN_INFO "cpufreq: change to %u MHz succeeded\n", (freqs.new / 1000));
+		dprintk (KERN_INFO "cpufreq: change to %u MHz succeeded\n", (newfreq / 1000));
 	} else {
 		printk (KERN_ERR "cpufreq: change failed - I/O error\n");
 	}
 
-	if (notify)
+	/* Make sure we're initialized before calling notify */
+	if (notify && (freqs.new != 0))
 		cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE);
 
 	return;

@@ -295,7 +300,7 @@
 		return -EIO;
 
 	dprintk(KERN_INFO "cpufreq: currently at %s speed setting - %i MHz\n", 
-		(speed == speedstep_low_freq) ? "low" : "high",
+		(speed == speedstep_freqs[SPEEDSTEP_LOW].frequency) ? "low" : "high",
 		(speed / 1000));
 
 	/* cpuinfo and default policy values */

A message that is not in the archive (sent off-list or not yet received): "from Valdis.Kletnieks"@vt.edu

author
#3 lore raw

Excellant work Valdis. My OSS (non-ALSA) is fine now under 2.6.0-test1.

Indeed I have CONFIG_X86_SPEEDSTEP_ICH defined.

Thank you

Simon

Quoted 93 lines · On 2003.07.19 00:46, [email protected] wrote:
Quoted 1 line · Simon Boulet

Also, the OSS (non-ALSA) Intel ICH (i8xx) loads correctly but the

sound

Quoted 8 lines

output is slow (rate or clocking problem?). My sound was fine under

2.4.21.

i810_audio: only 48Khz playback available.

i810_audio: setting clocking to 64937

I was having problems with i810_audio clocking as well. It turned out to be
the Intel Speedstep support, of all things. *IF* your kernel includes:

CONFIG_X86_SPEEDSTEP_ICH=y

it was possible to end up with a broken value for loops_per_jiffie. I've
attached a patch that fixes the bug and does a few cleanups...

If you don't have the SpeedStep support in your kernel, then your problem is elsewhere... Good luck... ;)

/Valdis

--- arch/i386/kernel/cpu/cpufreq/speedstep-ich.c.linus 2003-
07-03
23:31:43.000000000 -0400
+++ arch/i386/kernel/cpu/cpufreq/speedstep-ich.c 2003-07-04
09:57:07.981299808 -0400

i386/kernel/cpu/cpufreq/speedstep-ich.c renamed from i386/kernel/cpu/cpufreq/speedstep-ich.c.linus +4 -2
@@ -77,15 +77,17 @@
 	u8			value;
 	unsigned long		flags;
 	struct cpufreq_freqs	freqs;
+	int			newfreq;
 
 	if (!speedstep_chipset_dev || (state > 0x1))
 		return;
 
 	freqs.old = speedstep_get_processor_frequency
 (speedstep_processor);
-	freqs.new = speedstep_freqs[SPEEDSTEP_LOW].frequency;
+	freqs.new = speedstep_freqs[state].frequency;
 	freqs.cpu = 0; /* speedstep.c is UP only driver */
 
-	if (notify)
+	/* make sure we've initialized before calling notify */
+	if (notify && (freqs.new != 0))
 		cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE);
i386/kernel/cpu/cpufreq/speedstep-ich.c renamed from i386/kernel/cpu/cpufreq/speedstep-ich.c.linus +5 -2
 	/* get PMBASE */

@@ -136,13 +138,16 @@
 
 	dprintk(KERN_DEBUG "cpufreq: read at pmbase 0x%x + 0x50
 returned 0x%x\n", pmbase, value);
 
+	/* freqs.new may not be set yet - need local copy */
+	newfreq = speedstep_get_processor_frequency
 (speedstep_processor);
 	if (state == (value & 0x1)) {
-		dprintk (KERN_INFO "cpufreq: change to %u MHz
 succeeded\n", (freqs.new / 1000));
+		dprintk (KERN_INFO "cpufreq: change to %u MHz
 succeeded\n", (newfreq / 1000));
 	} else {
 		printk (KERN_ERR "cpufreq: change failed - I/O
 error\n");
 	}
 
-	if (notify)
+	/* Make sure we're initialized before calling notify */
+	if (notify && (freqs.new != 0))
 		cpufreq_notify_transition(&freqs,

CPUFREQ_POSTCHANGE);

i386/kernel/cpu/cpufreq/speedstep-ich.c renamed from i386/kernel/cpu/cpufreq/speedstep-ich.c.linus +1 -1
 	return;

@@ -295,7 +300,7 @@
 		return -EIO;
 
 	dprintk(KERN_INFO "cpufreq: currently at %s speed setting -
 %i MHz\n",
-		(speed == speedstep_low_freq) ? "low" : "high",
+		(speed == speedstep_freqs[SPEEDSTEP_LOW].frequency) ?
 "low" : "high",
 		(speed / 1000));

/* cpuinfo and default policy values */

#4 lore raw
In reply to Simon Boulet

At Fri, 18 Jul 2003 22:10:12 -0400,

Quoted 10 lines · Simon Boulet

[1 <text/plain; ISO-8859-1 (7bit)>]
Hello everyone,

In case I am sending this to a list, please CC to me regarding anything related to this issue. I am not a member of the list.

I am having a Kernel OOPS with 2.6.0-test1-ac2 (same thing under non- ac2) using ALSA with OSS compatibily enabled on an Intel 82801CA/CAM AC'97 (ICH3 mobile) integrated Audio.

perhaps a known problem.
a workaround is to turn on the framepointer.

Takashi

A message that is not in the archive (sent off-list or not yet received): "from tiwai"@suse.de

author
#6 lore raw

Yes, CONFIG_FRAME_POINTER does fix my issues with ALSA. And the patch submited by Valdis some days ago fixed OSS.

Thank you

Simon

Quoted 24 lines · Takashi Iwai

At Fri, 18 Jul 2003 22:10:12 -0400,

Quoted 5 lines · Simon Boulet

[1 <text/plain; ISO-8859-1 (7bit)>]
Hello everyone,

In case I am sending this to a list, please CC to me regarding

anything

Quoted 3 lines

related to this issue. I am not a member of the list.

I am having a Kernel OOPS with 2.6.0-test1-ac2 (same thing under

non-

Quoted 1 line

ac2) using ALSA with OSS compatibily enabled on an Intel 82801CA/

CAM

Quoted 1 line

AC'97 (ICH3 mobile) integrated Audio.

perhaps a known problem.
a workaround is to turn on the framepointer.

Takashi