Skip to contentSkip to conversation

General protection fault when unloading snd_hda_intel on 2.6.30-rc1+

DISCUSSION Carlos R. Mafra 4 participants 13 messages Apr 11, 2009 – Apr 15, 2009 Sound

lore
Share & tools
lore
#0 lore raw

If I try on my Vaio VGN-FZ240E, using the latest kernel 2.6.30-rc1-00191-gd848223

modprobe -r snd_hda_intel

I get the following trace (copied from dmesg):

general protection fault: 0000 [#1] SMP 
last sysfs file: /sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq
CPU 1 
Modules linked in: nvram uvcvideo videodev v4l1_compat v4l2_compat_ioctl32 snd_hda_codec_idt 
usbhid snd_hda_intel(-) snd_hda_codec snd_hwdep snd_pcm sr_mod sky2 i2c_i801 iwlagn snd_timer 
sg snd_page_alloc evdev ata_piix ahci libata sd_mod scsi_mod uhci_hcd ohci_hcd ehci_hcd usbcore [last unloaded: scsi_wait_scan]
Pid: 9063, comm: modprobe Not tainted 2.6.30-rc1-00191-gd848223 #48 VGN-FZ240E
RIP: 0010:[<ffffffff8044bf90>]  [<ffffffff8044bf90>] input_event+0x40/0xa0
RSP: 0018:ffff88006f283a78  EFLAGS: 00010293
RAX: 0000000000000008 RBX: 0000000000000003 RCX: 0000000000000000
RDX: 0000000000000007 RSI: 0000000000000005 RDI: 697665646632785c
RBP: ffff88006f283aa8 R08: 0000000004d6f76c R09: 0000000000000001
R10: 0000000000000000 R11: 0000000000000001 R12: ffffffff805bb97c
R13: 0000000000000005 R14: 697665646632785c R15: 0000000000000000
FS:  00007fc6c1e1c6f0(0000) GS:ffff880001028000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 000000000071bf40 CR3: 000000007e2d4000 CR4: 00000000000006e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process modprobe (pid: 9063, threadinfo ffff88006f282000, task ffff88006f206840)
Stack:
 ffff88006f283aa8 0000000000000003 ffffffff805bb97c ffff88007e811cc0
 0000000000000001 0000000000000000 ffff88006f283ae8 ffffffff80475612
 ffff88007f241ae0 0000000000000002 ffff88007d5eb010 0000000000000000
Call Trace:
 [<ffffffff80475612>] snd_jack_report+0x92/0xa0
 [<ffffffffa0160047>] stac92xx_unsol_event+0x2e7/0x5e0 [snd_hda_codec_idt]
 [<ffffffffa015f528>] stac_issue_unsol_event+0x58/0x70 [snd_hda_codec_idt]
 [<ffffffffa0160796>] stac92xx_init+0x186/0x660 [snd_hda_codec_idt]
 [<ffffffffa0160c90>] stac92xx_resume+0x20/0x60 [snd_hda_codec_idt]
 [<ffffffffa013c973>] hda_call_codec_resume+0x53/0x90 [snd_hda_codec]
 [<ffffffffa013ca50>] snd_hda_power_up+0x40/0x70 [snd_hda_codec]
 [<ffffffffa013c2b5>] snd_hda_codec_write+0x45/0xa0 [snd_hda_codec]
 [<ffffffffa013c37a>] restore_pincfgs+0x6a/0xa0 [snd_hda_codec]
 [<ffffffffa013ed75>] snd_hda_codec_free+0x35/0x120 [snd_hda_codec]
 [<ffffffffa01403f8>] snd_hda_bus_free+0x68/0xb0 [snd_hda_codec]
 [<ffffffffa01405e4>] snd_hda_bus_dev_free+0x14/0x20 [snd_hda_codec]
 [<ffffffff80473e61>] snd_device_free+0x71/0xd0
 [<ffffffff80473f24>] snd_device_free_all+0x64/0x70
 [<ffffffff8046e16a>] snd_card_do_free+0x3a/0xd0
 [<ffffffff8046ee39>] snd_card_free+0xa9/0xd0
 [<ffffffffa015031e>] azx_remove+0x18/0x2a [snd_hda_intel]
 [<ffffffff8036506f>] pci_device_remove+0x2f/0x60
 [<ffffffff8040ec8c>] __device_release_driver+0x6c/0xb0
 [<ffffffff8040ed78>] driver_detach+0xa8/0xb0
 [<ffffffff8040de48>] bus_remove_driver+0x98/0xd0
 [<ffffffff8040f357>] driver_unregister+0x47/0x60
 [<ffffffff8036530c>] pci_unregister_driver+0x4c/0xc0
 [<ffffffffa0150304>] alsa_card_azx_exit+0x10/0x12 [snd_hda_intel]
 [<ffffffff8026787b>] sys_delete_module+0x18b/0x240
 [<ffffffff80257be9>] ? up_write+0x9/0x10
 [<ffffffff8020be2b>] system_call_fastpath+0x16/0x1b
Code: 89 5d d8 4c 89 65 e0 49 89 fe 41 89 f5 41 89 cf 83 fe 1f 76 16 48 8b 5d d8 4c 8b 65 
e0 4c 8b 6d e8 4c 8b 75 f0 4c 8b 7d f8 c9 c3 <0f> a3 77 20 19 c0 85 c0 74 e0 48 8d 9f f8 06 00 00 89 55 d0 48 
RIP  [<ffffffff8044bf90>] input_event+0x40/0xa0
 RSP <ffff88006f283a78>
---[ end trace 7fa0517f56c40085 ]---

After that, 'cat /proc/modules' contains this line:

snd_hda_intel 28648 0 - Unloading 0xffffffffa014e000

and the HDA driver never finishes that advertised "Unloading".

It also stops working,

[mafra@Pilar:linux-2.6]$ alsamixer
alsamixer: function snd_ctl_open failed for default: No such file or directory

The complete dmesg is available at

http://www.aei.mpg.de/~crmafra/dmesg_hda_intel.txt

and from lspci -vvnn I see this

00:1b.0 Audio device [0403]: Intel Corporation 82801H (ICH8 Family) HD Audio Controller [8086:284b] (rev 03)
Subsystem: Sony Corporation Vaio VGN-FZ260E [104d:9005]
Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR+ FastB2B- DisINTx-
Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
Latency: 0, Cache Line Size: 64 bytes
Interrupt: pin A routed to IRQ 22
Region 0: Memory at fc400000 (64-bit, non-prefetchable) [size=16K]
Capabilities: [50] Power Management version 2
Flags: PMEClk- DSI- D1- D2- AuxCurrent=55mA PME(D0+,D1-,D2-,D3hot+,D3cold+)
Status: D0 PME-Enable- DSel=0 DScale=0 PME-
Capabilities: [60] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable-
Address: 0000000000000000 Data: 0000
Capabilities: [70] Express (v1) Root Complex Integrated Endpoint, MSI 00
DevCap: MaxPayload 128 bytes, PhantFunc 0, Latency L0s <64ns, L1 <1us
ExtTag- RBE- FLReset-
DevCtl: Report errors: Correctable- Non-Fatal- Fatal- Unsupported-
RlxdOrd- ExtTag- PhantFunc- AuxPwr- NoSnoop+
MaxPayload 128 bytes, MaxReadReq 128 bytes
DevSta: CorrErr- UncorrErr- FatalErr- UnsuppReq- AuxPwr+ TransPend-
LnkCap: Port #0, Speed unknown, Width x0, ASPM unknown, Latency L0 <64ns, L1 <1us
ClockPM- Suprise- LLActRep- BwNot-
LnkCtl: ASPM Disabled; Disabled- Retrain- CommClk-
ExtSynch- ClockPM- AutWidDis- BWInt- AutBWInt-
LnkSta: Speed unknown, Width x0, TrErr- Train- SlotClk- DLActive- BWMgmt- ABWMgmt-
Capabilities: [100] Virtual Channel <?>
Capabilities: [130] Root Complex Link <?>
Kernel modules: snd-hda-intel

When the audio is working alsamixer says it is using Sigmatel STAC9872AK.

Anything else I can do to help?

#1 lore raw
Quoted 1 line · On Sat, 11 Apr 2009 21:36:30 +0200 "Carlos R. Mafra" <[email protected]> wrote:
general protection fault:

Hi,

if you happen to have CONFIG_DEBUG_INFO enabled, then you can run the oops through scripts/markup_oops.pl to get the source code of the exact line that is oopsing.. makes diagnosing it even easier!

Signature · 3 lines
Arjan van de Ven 	Intel Open Source Technology Centre
For development, discussion and tips for power savings, 
visit http://www.lesswatts.org
#2 lore raw
Quoted 4 lines · Arjan van de Ven

if you happen to have CONFIG_DEBUG_INFO enabled, then you can run the oops through scripts/markup_oops.pl to get the source code of the exact line that is oopsing.. makes diagnosing it even easier!

I did not have CONFIG_DEBUG_INFO enabled back then, but now I have.

The script gives me this output:

No vmlinux specified, assuming /lib/modules/2.6.30-rc1-00191-gd848223/build/vmlinux
add_input_randomness(type, code, value); input_handle_event(dev, type, code, value); spin_unlock_irqrestore(&dev->event_lock, flags);
}
}

 ffffffff8044bf7a:	48 8b 5d d8          	mov    -0x28(%rbp),%rbx |  %ebx => 3 
 ffffffff8044bf7e:	4c 8b 65 e0          	mov    -0x20(%rbp),%r12 |  %r12 => ffffffff805bb97c 
 ffffffff8044bf82:	4c 8b 6d e8          	mov    -0x18(%rbp),%r13 |  %r13 => 5 
 ffffffff8044bf86:	4c 8b 75 f0          	mov    -0x10(%rbp),%r14 |  %r14 => 697665646632785c 
 ffffffff8044bf8a:	4c 8b 7d f8          	mov    -0x8(%rbp),%r15 |  %r15 => 0 
 ffffffff8044bf8e:	c9                   	leaveq  
 ffffffff8044bf8f:	c3                   	retq    

static inline int variable_test_bit(int nr, volatile const unsigned long *addr) {
int oldbit;

asm volatile("bt %2,%1\n\t"

*ffffffff8044bf90:	0f a3 77 20          	bt     %esi,0x20(%rdi) |  %edi = 697665646632785c  %esi = 5 <--- faulting instruction
 ffffffff8044bf94:	19 c0                	sbb    %eax,%eax 

static struct input_handler *input_table[8];

static inline int is_event_supported(unsigned int code,
unsigned long *bm, unsigned int max)
{
return code <= max && test_bit(code, bm);

 ffffffff8044bf96:	85 c0                	test   %eax,%eax 
 ffffffff8044bf98:	74 e0                	je     ffffffff8044bf7a <input_event+0x2a> 

{
unsigned long flags;

if (is_event_supported(type, dev->evbit, EV_MAX)) {

#3 lore raw
In reply to Arjan van de Ven
Quoted 4 lines · Arjan van de Ven

if you happen to have CONFIG_DEBUG_INFO enabled, then you can run the oops through scripts/markup_oops.pl to get the source code of the exact line that is oopsing.. makes diagnosing it even easier!

I did not have CONFIG_DEBUG_INFO enabled back then, but now I have.

The script gives me this output:

No vmlinux specified, assuming /lib/modules/2.6.30-rc1-00191-gd848223/build/vmlinux
add_input_randomness(type, code, value); input_handle_event(dev, type, code, value); spin_unlock_irqrestore(&dev->event_lock, flags);
}
}

 ffffffff8044bf7a:	48 8b 5d d8          	mov    -0x28(%rbp),%rbx |  %ebx => 3 
 ffffffff8044bf7e:	4c 8b 65 e0          	mov    -0x20(%rbp),%r12 |  %r12 => ffffffff805bb97c 
 ffffffff8044bf82:	4c 8b 6d e8          	mov    -0x18(%rbp),%r13 |  %r13 => 5 
 ffffffff8044bf86:	4c 8b 75 f0          	mov    -0x10(%rbp),%r14 |  %r14 => 697665646632785c 
 ffffffff8044bf8a:	4c 8b 7d f8          	mov    -0x8(%rbp),%r15 |  %r15 => 0 
 ffffffff8044bf8e:	c9                   	leaveq  
 ffffffff8044bf8f:	c3                   	retq    

static inline int variable_test_bit(int nr, volatile const unsigned long *addr) {
int oldbit;

asm volatile("bt %2,%1\n\t"

*ffffffff8044bf90:	0f a3 77 20          	bt     %esi,0x20(%rdi) |  %edi = 697665646632785c  %esi = 5 <--- faulting instruction
 ffffffff8044bf94:	19 c0                	sbb    %eax,%eax 

static struct input_handler *input_table[8];

static inline int is_event_supported(unsigned int code,
unsigned long *bm, unsigned int max)
{
return code <= max && test_bit(code, bm);

 ffffffff8044bf96:	85 c0                	test   %eax,%eax 
 ffffffff8044bf98:	74 e0                	je     ffffffff8044bf7a <input_event+0x2a> 

{
unsigned long flags;

if (is_event_supported(type, dev->evbit, EV_MAX)) {

#4 lore raw
Quoted 9 lines · On Sat, 11 Apr 2009 23:03:36 +0200 "Carlos R. Mafra" <[email protected]> wrote:

From: "Carlos R. Mafra" <[email protected]>
To: Arjan van de Ven <[email protected]>
Cc: [email protected], [email protected], [email protected] Subject: Re: General protection fault when unloading snd_hda_intel on 2.6.30-rc1+ Date: Sat, 11 Apr 2009 23:03:36 +0200 User-Agent: Mutt/1.5.18 (2008-05-17)

On Sat 11.Apr'09 at 13:20:33 -0700, Arjan van de Ven wrote:
[...]

asm volatile("bt %2,%1\n\t"

*ffffffff8044bf90:	0f a3 77 20          	bt     %esi,0x20(%rdi) |  %edi = 697665646632785c  %esi = 5 <--- faulting instruction
 ffffffff8044bf94:	19 c0                	sbb    %eax,%eax 

static struct input_handler *input_table[8];

static inline int is_event_supported(unsigned int code,
unsigned long *bm, unsigned int max)
{
return code <= max && test_bit(code, bm);

 ffffffff8044bf96:	85 c0                	test   %eax,%eax 
 ffffffff8044bf98:	74 e0                	je     ffffffff8044bf7a <input_event+0x2a> 

{
unsigned long flags;

if (is_event_supported(type, dev->evbit, EV_MAX)) {

so... "dev" is pointing evbit at... 697665646632785 which looks like grabage.

Signature · 3 lines
Arjan van de Ven 	Intel Open Source Technology Centre
For development, discussion and tips for power savings, 
visit http://www.lesswatts.org
#5 lore raw
Quoted 2 lines · Arjan van de Ven

so... "dev" is pointing evbit at... 697665646632785 which looks like grabage.

FWIW 697665646632785c decodes in ASCII as "\x2fdevi".

M.

#7 lore raw

Re: General protection fault when unloading snd_hda_intel on 2.6.30-rc1+ (bisected)

Quoted 5 lines · Michal Jaegermann
Quoted 2 lines · Arjan van de Ven

so... "dev" is pointing evbit at... 697665646632785 which looks like grabage.

FWIW 697665646632785c decodes in ASCII as "\x2fdevi".

So I painfully (exchanged 'good' for 'bad' in the middle and had to restart all over) bisected this down to

commit 3be141494a080a9189b51fa78154c975ad8d9806

Author: Takashi Iwai <[email protected]>
Date:   Fri Feb 20 14:11:16 2009 +0100

ALSA: hda - Add generic pincfg initialization

Added the generic pincfg cache and save/restore functions. Also introduced the pin-overriding via hwdep sysfs.

It doesn't revert cleanly from the latest kernel, so I could not double check it.

OBS: All tests were done with the headphone plugged in.

#8 lore raw

Re: General protection fault when unloading snd_hda_intel on 2.6.30-rc1+ (bisected)

In reply to Michal Jaegermann
Quoted 5 lines · Michal Jaegermann
Quoted 2 lines · Arjan van de Ven

so... "dev" is pointing evbit at... 697665646632785 which looks like grabage.

FWIW 697665646632785c decodes in ASCII as "\x2fdevi".

So I painfully (exchanged 'good' for 'bad' in the middle and had to restart all over) bisected this down to

commit 3be141494a080a9189b51fa78154c975ad8d9806

Author: Takashi Iwai <[email protected]>
Date:   Fri Feb 20 14:11:16 2009 +0100

ALSA: hda - Add generic pincfg initialization

Added the generic pincfg cache and save/restore functions. Also introduced the pin-overriding via hwdep sysfs.

It doesn't revert cleanly from the latest kernel, so I could not double check it.

OBS: All tests were done with the headphone plugged in.

#9 lore raw
In reply to Carlos R. Mafra

At Sat, 11 Apr 2009 21:36:30 +0200,

Quoted 56 lines · Carlos R. Mafra

If I try on my Vaio VGN-FZ240E, using the latest kernel 2.6.30-rc1-00191-gd848223

modprobe -r snd_hda_intel

I get the following trace (copied from dmesg):

general protection fault: 0000 [#1] SMP 
last sysfs file: /sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq
CPU 1 
Modules linked in: nvram uvcvideo videodev v4l1_compat v4l2_compat_ioctl32 snd_hda_codec_idt 
usbhid snd_hda_intel(-) snd_hda_codec snd_hwdep snd_pcm sr_mod sky2 i2c_i801 iwlagn snd_timer 
sg snd_page_alloc evdev ata_piix ahci libata sd_mod scsi_mod uhci_hcd ohci_hcd ehci_hcd usbcore [last unloaded: scsi_wait_scan]
Pid: 9063, comm: modprobe Not tainted 2.6.30-rc1-00191-gd848223 #48 VGN-FZ240E
RIP: 0010:[<ffffffff8044bf90>]  [<ffffffff8044bf90>] input_event+0x40/0xa0
RSP: 0018:ffff88006f283a78  EFLAGS: 00010293
RAX: 0000000000000008 RBX: 0000000000000003 RCX: 0000000000000000
RDX: 0000000000000007 RSI: 0000000000000005 RDI: 697665646632785c
RBP: ffff88006f283aa8 R08: 0000000004d6f76c R09: 0000000000000001
R10: 0000000000000000 R11: 0000000000000001 R12: ffffffff805bb97c
R13: 0000000000000005 R14: 697665646632785c R15: 0000000000000000
FS:  00007fc6c1e1c6f0(0000) GS:ffff880001028000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 000000000071bf40 CR3: 000000007e2d4000 CR4: 00000000000006e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process modprobe (pid: 9063, threadinfo ffff88006f282000, task ffff88006f206840)
Stack:
 ffff88006f283aa8 0000000000000003 ffffffff805bb97c ffff88007e811cc0
 0000000000000001 0000000000000000 ffff88006f283ae8 ffffffff80475612
 ffff88007f241ae0 0000000000000002 ffff88007d5eb010 0000000000000000
Call Trace:
 [<ffffffff80475612>] snd_jack_report+0x92/0xa0
 [<ffffffffa0160047>] stac92xx_unsol_event+0x2e7/0x5e0 [snd_hda_codec_idt]
 [<ffffffffa015f528>] stac_issue_unsol_event+0x58/0x70 [snd_hda_codec_idt]
 [<ffffffffa0160796>] stac92xx_init+0x186/0x660 [snd_hda_codec_idt]
 [<ffffffffa0160c90>] stac92xx_resume+0x20/0x60 [snd_hda_codec_idt]
 [<ffffffffa013c973>] hda_call_codec_resume+0x53/0x90 [snd_hda_codec]
 [<ffffffffa013ca50>] snd_hda_power_up+0x40/0x70 [snd_hda_codec]
 [<ffffffffa013c2b5>] snd_hda_codec_write+0x45/0xa0 [snd_hda_codec]
 [<ffffffffa013c37a>] restore_pincfgs+0x6a/0xa0 [snd_hda_codec]
 [<ffffffffa013ed75>] snd_hda_codec_free+0x35/0x120 [snd_hda_codec]
 [<ffffffffa01403f8>] snd_hda_bus_free+0x68/0xb0 [snd_hda_codec]
 [<ffffffffa01405e4>] snd_hda_bus_dev_free+0x14/0x20 [snd_hda_codec]
 [<ffffffff80473e61>] snd_device_free+0x71/0xd0
 [<ffffffff80473f24>] snd_device_free_all+0x64/0x70
 [<ffffffff8046e16a>] snd_card_do_free+0x3a/0xd0
 [<ffffffff8046ee39>] snd_card_free+0xa9/0xd0
 [<ffffffffa015031e>] azx_remove+0x18/0x2a [snd_hda_intel]
 [<ffffffff8036506f>] pci_device_remove+0x2f/0x60
 [<ffffffff8040ec8c>] __device_release_driver+0x6c/0xb0
 [<ffffffff8040ed78>] driver_detach+0xa8/0xb0
 [<ffffffff8040de48>] bus_remove_driver+0x98/0xd0
 [<ffffffff8040f357>] driver_unregister+0x47/0x60
 [<ffffffff8036530c>] pci_unregister_driver+0x4c/0xc0
 [<ffffffffa0150304>] alsa_card_azx_exit+0x10/0x12 [snd_hda_intel]
Quoted 8 lines
 [<ffffffff8026787b>] sys_delete_module+0x18b/0x240
 [<ffffffff80257be9>] ? up_write+0x9/0x10
 [<ffffffff8020be2b>] system_call_fastpath+0x16/0x1b
Code: 89 5d d8 4c 89 65 e0 49 89 fe 41 89 f5 41 89 cf 83 fe 1f 76 16 48 8b 5d d8 4c 8b 65 
e0 4c 8b 6d e8 4c 8b 75 f0 4c 8b 7d f8 c9 c3 <0f> a3 77 20 19 c0 85 c0 74 e0 48 8d 9f f8 06 00 00 89 55 d0 48 
RIP  [<ffffffff8044bf90>] input_event+0x40/0xa0
 RSP <ffff88006f283a78>
---[ end trace 7fa0517f56c40085 ]---

This seems happening in the path exciting the codec again in the release due to restoration of the pin defcfg values.

Could you try the patch below, or merge for-next or master branch of sound git tree?

    git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6.git

Thanks,

Takashi

---

include/sound/jack.h +2 -0
index 6b013c6..f236e42 100644
@@ -50,6 +50,8 @@ struct snd_jack {
 	int type;
 	const char *id;
 	char name[100];
+	void *private_data;
+	void (*private_free)(struct snd_jack *);
 };
 
 #ifdef CONFIG_SND_JACK
sound/core/jack.c +3 -0
index c8254c6..d54d1a0 100644
@@ -35,6 +35,9 @@ static int snd_jack_dev_free(struct snd_device *device)
 {
 	struct snd_jack *jack = device->device_data;
 
+	if (jack->private_free)
+		jack->private_free(jack);
+
 	/* If the input device is registered with the input subsystem
 	 * then we need to use a different deallocator. */
 	if (jack->registered)
sound/pci/hda/patch_conexant.c +18 -3
index 1f2ad76..56ce19e 100644
@@ -350,12 +350,20 @@ static int conexant_mux_enum_put(struct snd_kcontrol *kcontrol,
 }
 
 #ifdef CONFIG_SND_JACK
+static void conexant_free_jack_priv(struct snd_jack *jack)
+{
+	struct conexant_jack *jacks = jack->private_data;
+	jacks->nid = 0;
+	jacks->jack = NULL;
+}
+
 static int conexant_add_jack(struct hda_codec *codec,
 		hda_nid_t nid, int type)
 {
 	struct conexant_spec *spec;
 	struct conexant_jack *jack;
 	const char *name;
+	int err;
 
 	spec = codec->spec;
 	snd_array_init(&spec->jacks, sizeof(*jack), 32);
@@ -368,7 +376,12 @@ static int conexant_add_jack(struct hda_codec *codec,
 	jack->nid = nid;
 	jack->type = type;
 
-	return snd_jack_new(codec->bus->card, name, type, &jack->jack);
+	err = snd_jack_new(codec->bus->card, name, type, &jack->jack);
+	if (err < 0)
+		return err;
+	jack->jack->private_data = jack;
+	jack->jack->private_free = conexant_free_jack_priv;
+	return 0;
 }
 
 static void conexant_report_jack(struct hda_codec *codec, hda_nid_t nid)
@@ -455,8 +468,10 @@ static void conexant_free(struct hda_codec *codec)
 	if (spec->jacks.list) {
 		struct conexant_jack *jacks = spec->jacks.list;
 		int i;
-		for (i = 0; i < spec->jacks.used; i++)
-			snd_device_free(codec->bus->card, &jacks[i].jack);
+		for (i = 0; i < spec->jacks.used; i++, jacks++) {
+			if (jacks->jack)
+				snd_device_free(codec->bus->card, jacks->jack);
+		}
 		snd_array_free(&spec->jacks);
 	}
 #endif
sound/pci/hda/patch_sigmatel.c +22 -5
index 61996a2..ce30b45 100644
@@ -3851,6 +3851,15 @@ static void stac_gpio_set(struct hda_codec *codec, unsigned int mask,
 			   AC_VERB_SET_GPIO_DATA, gpiostate); /* sync */
 }
 
+#ifdef CONFIG_SND_JACK
+static void stac92xx_free_jack_priv(struct snd_jack *jack)
+{
+	struct sigmatel_jack *jacks = jack->private_data;
+	jacks->nid = 0;
+	jacks->jack = NULL;
+}
+#endif
+
 static int stac92xx_add_jack(struct hda_codec *codec,
 		hda_nid_t nid, int type)
 {
@@ -3860,6 +3869,7 @@ static int stac92xx_add_jack(struct hda_codec *codec,
 	int def_conf = snd_hda_codec_get_pincfg(codec, nid);
 	int connectivity = get_defcfg_connect(def_conf);
 	char name[32];
+	int err;
 
 	if (connectivity && connectivity != AC_JACK_PORT_FIXED)
 		return 0;
@@ -3876,10 +3886,15 @@ static int stac92xx_add_jack(struct hda_codec *codec,
 		snd_hda_get_jack_connectivity(def_conf),
 		snd_hda_get_jack_location(def_conf));
 
-	return snd_jack_new(codec->bus->card, name, type, &jack->jack);
-#else
-	return 0;
+	err = snd_jack_new(codec->bus->card, name, type, &jack->jack);
+	if (err < 0) {
+		jack->nid = 0;
+		return err;
+	}
+	jack->jack->private_data = jack;
+	jack->jack->private_free = stac92xx_free_jack_priv;
 #endif
+	return 0;
 }
 
 static int stac_add_event(struct sigmatel_spec *spec, hda_nid_t nid,
@@ -4138,8 +4153,10 @@ static void stac92xx_free_jacks(struct hda_codec *codec)
 	if (!codec->bus->shutdown && spec->jacks.list) {
 		struct sigmatel_jack *jacks = spec->jacks.list;
 		int i;
-		for (i = 0; i < spec->jacks.used; i++)
-			snd_device_free(codec->bus->card, &jacks[i].jack);
+		for (i = 0; i < spec->jacks.used; i++, jacks++) {
+			if (jacks->jack)
+				snd_device_free(codec->bus->card, jacks->jack);
+		}
 	}
 	snd_array_free(&spec->jacks);
 #endif
author Tested-by
#10 lore raw
Quoted 7 lines · Takashi Iwai

This seems happening in the path exciting the codec again in the release due to restoration of the pin defcfg values.

Could you try the patch below, or merge for-next or master branch of sound git tree?

    git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6.git

I tested the patch below and now the module unloads correctly, so the problem is fixed.

Thanks Takashi!

Quoted 141 lines

---

include/sound/jack.h +2 -0
index 6b013c6..f236e42 100644
@@ -50,6 +50,8 @@ struct snd_jack {
 	int type;
 	const char *id;
 	char name[100];
+	void *private_data;
+	void (*private_free)(struct snd_jack *);
 };
 
 #ifdef CONFIG_SND_JACK
sound/core/jack.c +3 -0
index c8254c6..d54d1a0 100644
@@ -35,6 +35,9 @@ static int snd_jack_dev_free(struct snd_device *device)
 {
 	struct snd_jack *jack = device->device_data;
 
+	if (jack->private_free)
+		jack->private_free(jack);
+
 	/* If the input device is registered with the input subsystem
 	 * then we need to use a different deallocator. */
 	if (jack->registered)
sound/pci/hda/patch_conexant.c +18 -3
index 1f2ad76..56ce19e 100644
@@ -350,12 +350,20 @@ static int conexant_mux_enum_put(struct snd_kcontrol *kcontrol,
 }
 
 #ifdef CONFIG_SND_JACK
+static void conexant_free_jack_priv(struct snd_jack *jack)
+{
+	struct conexant_jack *jacks = jack->private_data;
+	jacks->nid = 0;
+	jacks->jack = NULL;
+}
+
 static int conexant_add_jack(struct hda_codec *codec,
 		hda_nid_t nid, int type)
 {
 	struct conexant_spec *spec;
 	struct conexant_jack *jack;
 	const char *name;
+	int err;
 
 	spec = codec->spec;
 	snd_array_init(&spec->jacks, sizeof(*jack), 32);
@@ -368,7 +376,12 @@ static int conexant_add_jack(struct hda_codec *codec,
 	jack->nid = nid;
 	jack->type = type;
 
-	return snd_jack_new(codec->bus->card, name, type, &jack->jack);
+	err = snd_jack_new(codec->bus->card, name, type, &jack->jack);
+	if (err < 0)
+		return err;
+	jack->jack->private_data = jack;
+	jack->jack->private_free = conexant_free_jack_priv;
+	return 0;
 }
 
 static void conexant_report_jack(struct hda_codec *codec, hda_nid_t nid)
@@ -455,8 +468,10 @@ static void conexant_free(struct hda_codec *codec)
 	if (spec->jacks.list) {
 		struct conexant_jack *jacks = spec->jacks.list;
 		int i;
-		for (i = 0; i < spec->jacks.used; i++)
-			snd_device_free(codec->bus->card, &jacks[i].jack);
+		for (i = 0; i < spec->jacks.used; i++, jacks++) {
+			if (jacks->jack)
+				snd_device_free(codec->bus->card, jacks->jack);
+		}
 		snd_array_free(&spec->jacks);
 	}
 #endif
sound/pci/hda/patch_sigmatel.c +22 -5
index 61996a2..ce30b45 100644
@@ -3851,6 +3851,15 @@ static void stac_gpio_set(struct hda_codec *codec, unsigned int mask,
 			   AC_VERB_SET_GPIO_DATA, gpiostate); /* sync */
 }
 
+#ifdef CONFIG_SND_JACK
+static void stac92xx_free_jack_priv(struct snd_jack *jack)
+{
+	struct sigmatel_jack *jacks = jack->private_data;
+	jacks->nid = 0;
+	jacks->jack = NULL;
+}
+#endif
+
 static int stac92xx_add_jack(struct hda_codec *codec,
 		hda_nid_t nid, int type)
 {
@@ -3860,6 +3869,7 @@ static int stac92xx_add_jack(struct hda_codec *codec,
 	int def_conf = snd_hda_codec_get_pincfg(codec, nid);
 	int connectivity = get_defcfg_connect(def_conf);
 	char name[32];
+	int err;
 
 	if (connectivity && connectivity != AC_JACK_PORT_FIXED)
 		return 0;
@@ -3876,10 +3886,15 @@ static int stac92xx_add_jack(struct hda_codec *codec,
 		snd_hda_get_jack_connectivity(def_conf),
 		snd_hda_get_jack_location(def_conf));
 
-	return snd_jack_new(codec->bus->card, name, type, &jack->jack);
-#else
-	return 0;
+	err = snd_jack_new(codec->bus->card, name, type, &jack->jack);
+	if (err < 0) {
+		jack->nid = 0;
+		return err;
+	}
+	jack->jack->private_data = jack;
+	jack->jack->private_free = stac92xx_free_jack_priv;
 #endif
+	return 0;
 }
 
 static int stac_add_event(struct sigmatel_spec *spec, hda_nid_t nid,
@@ -4138,8 +4153,10 @@ static void stac92xx_free_jacks(struct hda_codec *codec)
 	if (!codec->bus->shutdown && spec->jacks.list) {
 		struct sigmatel_jack *jacks = spec->jacks.list;
 		int i;
-		for (i = 0; i < spec->jacks.used; i++)
-			snd_device_free(codec->bus->card, &jacks[i].jack);
+		for (i = 0; i < spec->jacks.used; i++, jacks++) {
+			if (jacks->jack)
+				snd_device_free(codec->bus->card, jacks->jack);
+		}
 	}
 	snd_array_free(&spec->jacks);
 #endif
author Tested-by
#11 lore raw
In reply to Takashi Iwai
Quoted 7 lines · Takashi Iwai

This seems happening in the path exciting the codec again in the release due to restoration of the pin defcfg values.

Could you try the patch below, or merge for-next or master branch of sound git tree?

    git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6.git

I tested the patch below and now the module unloads correctly, so the problem is fixed.

Thanks Takashi!

Quoted 141 lines

---

include/sound/jack.h +2 -0
index 6b013c6..f236e42 100644
@@ -50,6 +50,8 @@ struct snd_jack {
 	int type;
 	const char *id;
 	char name[100];
+	void *private_data;
+	void (*private_free)(struct snd_jack *);
 };
 
 #ifdef CONFIG_SND_JACK
sound/core/jack.c +3 -0
index c8254c6..d54d1a0 100644
@@ -35,6 +35,9 @@ static int snd_jack_dev_free(struct snd_device *device)
 {
 	struct snd_jack *jack = device->device_data;
 
+	if (jack->private_free)
+		jack->private_free(jack);
+
 	/* If the input device is registered with the input subsystem
 	 * then we need to use a different deallocator. */
 	if (jack->registered)
sound/pci/hda/patch_conexant.c +18 -3
index 1f2ad76..56ce19e 100644
@@ -350,12 +350,20 @@ static int conexant_mux_enum_put(struct snd_kcontrol *kcontrol,
 }
 
 #ifdef CONFIG_SND_JACK
+static void conexant_free_jack_priv(struct snd_jack *jack)
+{
+	struct conexant_jack *jacks = jack->private_data;
+	jacks->nid = 0;
+	jacks->jack = NULL;
+}
+
 static int conexant_add_jack(struct hda_codec *codec,
 		hda_nid_t nid, int type)
 {
 	struct conexant_spec *spec;
 	struct conexant_jack *jack;
 	const char *name;
+	int err;
 
 	spec = codec->spec;
 	snd_array_init(&spec->jacks, sizeof(*jack), 32);
@@ -368,7 +376,12 @@ static int conexant_add_jack(struct hda_codec *codec,
 	jack->nid = nid;
 	jack->type = type;
 
-	return snd_jack_new(codec->bus->card, name, type, &jack->jack);
+	err = snd_jack_new(codec->bus->card, name, type, &jack->jack);
+	if (err < 0)
+		return err;
+	jack->jack->private_data = jack;
+	jack->jack->private_free = conexant_free_jack_priv;
+	return 0;
 }
 
 static void conexant_report_jack(struct hda_codec *codec, hda_nid_t nid)
@@ -455,8 +468,10 @@ static void conexant_free(struct hda_codec *codec)
 	if (spec->jacks.list) {
 		struct conexant_jack *jacks = spec->jacks.list;
 		int i;
-		for (i = 0; i < spec->jacks.used; i++)
-			snd_device_free(codec->bus->card, &jacks[i].jack);
+		for (i = 0; i < spec->jacks.used; i++, jacks++) {
+			if (jacks->jack)
+				snd_device_free(codec->bus->card, jacks->jack);
+		}
 		snd_array_free(&spec->jacks);
 	}
 #endif
sound/pci/hda/patch_sigmatel.c +22 -5
index 61996a2..ce30b45 100644
@@ -3851,6 +3851,15 @@ static void stac_gpio_set(struct hda_codec *codec, unsigned int mask,
 			   AC_VERB_SET_GPIO_DATA, gpiostate); /* sync */
 }
 
+#ifdef CONFIG_SND_JACK
+static void stac92xx_free_jack_priv(struct snd_jack *jack)
+{
+	struct sigmatel_jack *jacks = jack->private_data;
+	jacks->nid = 0;
+	jacks->jack = NULL;
+}
+#endif
+
 static int stac92xx_add_jack(struct hda_codec *codec,
 		hda_nid_t nid, int type)
 {
@@ -3860,6 +3869,7 @@ static int stac92xx_add_jack(struct hda_codec *codec,
 	int def_conf = snd_hda_codec_get_pincfg(codec, nid);
 	int connectivity = get_defcfg_connect(def_conf);
 	char name[32];
+	int err;
 
 	if (connectivity && connectivity != AC_JACK_PORT_FIXED)
 		return 0;
@@ -3876,10 +3886,15 @@ static int stac92xx_add_jack(struct hda_codec *codec,
 		snd_hda_get_jack_connectivity(def_conf),
 		snd_hda_get_jack_location(def_conf));
 
-	return snd_jack_new(codec->bus->card, name, type, &jack->jack);
-#else
-	return 0;
+	err = snd_jack_new(codec->bus->card, name, type, &jack->jack);
+	if (err < 0) {
+		jack->nid = 0;
+		return err;
+	}
+	jack->jack->private_data = jack;
+	jack->jack->private_free = stac92xx_free_jack_priv;
 #endif
+	return 0;
 }
 
 static int stac_add_event(struct sigmatel_spec *spec, hda_nid_t nid,
@@ -4138,8 +4153,10 @@ static void stac92xx_free_jacks(struct hda_codec *codec)
 	if (!codec->bus->shutdown && spec->jacks.list) {
 		struct sigmatel_jack *jacks = spec->jacks.list;
 		int i;
-		for (i = 0; i < spec->jacks.used; i++)
-			snd_device_free(codec->bus->card, &jacks[i].jack);
+		for (i = 0; i < spec->jacks.used; i++, jacks++) {
+			if (jacks->jack)
+				snd_device_free(codec->bus->card, jacks->jack);
+		}
 	}
 	snd_array_free(&spec->jacks);
 #endif
#12 lore raw

At Tue, 14 Apr 2009 18:21:19 +0200,

Quoted 11 lines · Carlos R. Mafra
Quoted 7 lines · Takashi Iwai

This seems happening in the path exciting the codec again in the release due to restoration of the pin defcfg values.

Could you try the patch below, or merge for-next or master branch of sound git tree?

    git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6.git
Quoted 6 lines

I tested the patch below and now the module unloads correctly, so the problem is fixed.

Thanks Takashi!

Thanks for testing. I'll give a pull request soon later.

Takashi