Skip to contentSkip to conversation

2.6.0-test9-mm4 - kernel BUG at arch/i386/mm/fault.c:357!

DISCUSSION Ronny V. Vindenes 2 participants 9 messages Nov 19, 2003 Sound

lore
Share & tools
lore
#0 lore raw

Running the game Enemy Territory triggers this (log is from running it 3 times in a row) every time. Haven't been able to trigger it with any other programs. Machine is athlon64 running all 32bit.

On a happier note, interactivity problems are gone with the new acpi pm timer patch.

Signature · 1 line
Ronny V. Vindenes <[email protected]>
------------[ cut here ]------------
kernel BUG at arch/i386/mm/fault.c:357!
invalid operand: 0000 [#1]
PREEMPT 
CPU:    0
EIP:    0060:[<c011c509>]    Not tainted VLI
EFLAGS: 00010212

EIP is at do_page_fault+0x389/0x504
eax: 00029400 ebx: f7d42280 ecx: f64bbf40 edx: 00029400
esi: f7d422a0 edi: c1da5e40 ebp: f75fc040 esp: eb0c9f10
ds: 007b es: 007b ss: 0068
Process et.x86 (pid: 4509, threadinfo=eb0c8000 task=f75fc040)

Stack: f7d42280 c1da5e40 575364e4 00000001 00000001 575364e4 00000000 0240d793 
       00030002 00001000 eb0c8000 c0000000 e2eb9400 00000000 f89e6868 f780ec80 
       80044121 eb0c9f70 bffef4e4 e2eb9400 f8e325b2 bffef4e4 eb0c9f74 0000000c 
Call Trace:
 [<f89e6868>] snd_pcm_kernel_playback_ioctl+0x38/0x50 [snd_pcm]
 [<f8e325b2>] snd_pcm_oss_get_ptr+0x82/0x1e0 [snd_pcm_oss]
 [<c01727e0>] sys_ioctl+0xe0/0x2c0
 [<c011c180>] do_page_fault+0x0/0x504
 [<c035050f>] error_code+0x2f/0x38

Code: eb a8 8d 85 02 03 00 00 c7 04 24 fa 72 36 c0 89 44 24 04 e8 5a 6c 00 00 f6 84 24 b0 00 00 00 04 0f 84 7c fd ff ff e9 59 fe ff ff <0f> 0b 65 01 12 73 36 c0 8b 94 24 ac 00 00 00 f6 42 32 02 74 1f 
 ------------[ cut here ]------------
kernel BUG at arch/i386/mm/fault.c:357!
invalid operand: 0000 [#2]
PREEMPT 
CPU:    0
EIP:    0060:[<c011c509>]    Not tainted VLI
EFLAGS: 00010212

EIP is at do_page_fault+0x389/0x504
eax: 00008c00 ebx: f1d094c0 ecx: f6d4b180 edx: 00008c00
esi: f1d094e0 edi: e2de0640 ebp: f75fc040 esp: eb0c9f10
ds: 007b es: 007b ss: 0068
Process et.x86 (pid: 4518, threadinfo=eb0c8000 task=f75fc040)

Stack: f1d094c0 e2de0640 575348e4 00000001 00000001 575348e4 bffff700 eb0c9f44 
       00030002 f89fa000 eb0c8000 c0000000 dd715000 00000000 f89e6868 f780ec80 
       80044121 eb0c9f70 bffff7d4 dd715000 f8e325b2 bffff7d4 eb0c9f74 0000000c 
Call Trace:
 [<f89e6868>] snd_pcm_kernel_playback_ioctl+0x38/0x50 [snd_pcm]
 [<f8e325b2>] snd_pcm_oss_get_ptr+0x82/0x1e0 [snd_pcm_oss]
 [<c01727e0>] sys_ioctl+0xe0/0x2c0
 [<c011c180>] do_page_fault+0x0/0x504
 [<c035050f>] error_code+0x2f/0x38

Code: eb a8 8d 85 02 03 00 00 c7 04 24 fa 72 36 c0 89 44 24 04 e8 5a 6c 00 00 f6 84 24 b0 00 00 00 04 0f 84 7c fd ff ff e9 59 fe ff ff <0f> 0b 65 01 12 73 36 c0 8b 94 24 ac 00 00 00 f6 42 32 02 74 1f 
 ------------[ cut here ]------------
kernel BUG at arch/i386/mm/fault.c:357!
invalid operand: 0000 [#3]
PREEMPT 
CPU:    0
EIP:    0060:[<c011c509>]    Not tainted VLI
EFLAGS: 00010212

EIP is at do_page_fault+0x389/0x504
eax: 00008a00 ebx: f1d09940 ecx: eee00840 edx: 00008a00
esi: f1d09960 edi: e2e89940 ebp: f75fc040 esp: e57cff10
ds: 007b es: 007b ss: 0068
Process et.x86 (pid: 4535, threadinfo=e57ce000 task=f75fc040)

Stack: f1d09940 e2e89940 57533e64 00000001 00000001 57533e64 bffff700 e57cff44 
       00030002 f89fa000 e57ce000 c0000000 e2eb9c00 00000000 f89e6868 f780ec80 
       80044121 e57cff70 bffff7d4 e2eb9c00 f8e325b2 bffff7d4 e57cff74 0000000c 
Call Trace:
 [<f89e6868>] snd_pcm_kernel_playback_ioctl+0x38/0x50 [snd_pcm]
 [<f8e325b2>] snd_pcm_oss_get_ptr+0x82/0x1e0 [snd_pcm_oss]
 [<c01727e0>] sys_ioctl+0xe0/0x2c0
 [<c011c180>] do_page_fault+0x0/0x504
 [<c035050f>] error_code+0x2f/0x38

Code: eb a8 8d 85 02 03 00 00 c7 04 24 fa 72 36 c0 89 44 24 04 e8 5a 6c 00 00 f6 84 24 b0 00 00 00 04 0f 84 7c fd ff ff e9 59 fe ff ff <0f> 0b 65 01 12 73 36 c0 8b 94 24 ac 00 00 00 f6 42 32 02 74 1f 
#1 lore raw
Quoted 5 lines · Ronny V. Vindenes

Running the game Enemy Territory triggers this (log is from running it 3 times in a row) every time. Haven't been able to trigger it with any other programs. Machine is athlon64 running all 32bit.
On a happier note, interactivity problems are gone with the new acpi pm timer patch.

What sound card?

-- wli

#2 lore raw
Quoted 9 lines · William Lee Irwin III
Quoted 5 lines · Ronny V. Vindenes

Running the game Enemy Territory triggers this (log is from running it 3 times in a row) every time. Haven't been able to trigger it with any other programs. Machine is athlon64 running all 32bit.
On a happier note, interactivity problems are gone with the new acpi pm timer patch.

What sound card?

Hercules Fortissimo III (snd-cs46xx), there's also a Terratec EWX24/96 (snd-ice1712) installed but it's not in active use.

Signature · 1 line
Ronny V. Vindenes <[email protected]>
#3 lore raw
Quoted 1 line · William Lee Irwin III
Quoted 1 line

What sound card?

Quoted 2 lines · Ronny V. Vindenes

Hercules Fortissimo III (snd-cs46xx), there's also a Terratec EWX24/96 (snd-ice1712) installed but it's not in active use.

Any chance you can try this?

This might help pinpoint the problem.

-- wli

arch/i386/mm/fault.c +4 -0
diff -prauN mm4-2.6.0-test9-1/arch/i386/mm/fault.c mm4-2.6.0-test9-dbg-1/arch/i386/mm/fault.c
@@ -354,6 +354,10 @@ good_area:
 		case VM_FAULT_OOM:
 			goto out_of_memory;
 		default:
+			if (vma->vm_ops && vma->vm_ops->nopage)
+				print_symbol(KERN_CRIT "bad nopage %s\n",
+						vma->vm_ops->nopage);
+			printk(KERN_CRIT "handle_mm_fault() returned bad status\n");
 			BUG();
 	}
 
#4 lore raw
In reply to Ronny V. Vindenes
Quoted 1 line · William Lee Irwin III
Quoted 1 line

What sound card?

Quoted 2 lines · Ronny V. Vindenes

Hercules Fortissimo III (snd-cs46xx), there's also a Terratec EWX24/96 (snd-ice1712) installed but it's not in active use.

Er, sorry, try this one instead:

-- wli

arch/i386/mm/fault.c +5 -0
diff -prauN mm4-2.6.0-test9-1/arch/i386/mm/fault.c mm4-2.6.0-test9-dbg-1/arch/i386/mm/fault.c
@@ -21,6 +21,7 @@
 #include <linux/vt_kern.h>		/* For unblank_screen() */
 #include <linux/highmem.h>
 #include <linux/module.h>
+#include <linux/kallsyms.h>
 
 #include <asm/system.h>
 #include <asm/uaccess.h>
@@ -354,6 +355,10 @@ good_area:
 		case VM_FAULT_OOM:
 			goto out_of_memory;
 		default:
+			if (vma->vm_ops && vma->vm_ops->nopage)
+				print_symbol(KERN_CRIT "bad nopage %s\n",
+					(unsigned long)vma->vm_ops->nopage);
+			printk(KERN_CRIT "handle_mm_fault() returned bad status\n");
 			BUG();
 	}
 
#5 lore raw
Quoted 9 lines · William Lee Irwin III
Quoted 1 line · William Lee Irwin III
Quoted 1 line

What sound card?

Quoted 2 lines · Ronny V. Vindenes

Hercules Fortissimo III (snd-cs46xx), there's also a Terratec EWX24/96 (snd-ice1712) installed but it's not in active use.

Er, sorry, try this one instead:

bad nopage snd_pcm_mmap_data_nopage+0x0/0xc0 [snd_pcm] handle_mm_fault() returned bad status

------------[ cut here ]------------
kernel BUG at arch/i386/mm/fault.c:362!
invalid operand: 0000 [#1]
PREEMPT
CPU:    0
EIP:    0060:[<c011c523>]    Not tainted VLI
EFLAGS: 00010286

EIP is at do_page_fault+0x3a3/0x530
eax: 00000029 ebx: f4770040 ecx: f79fd340 edx: f3ff4000
esi: f4770060 edi: f4d13780 ebp: f5271340 esp: f3ff5f10
ds: 007b es: 007b ss: 0068
Process et.x86 (pid: 3223, threadinfo=f3ff4000 task=f5271340)

Stack: c0367680 f89e6f20 5753b664 00000001 00000001 5753b664 00000000
0240d793
       00030002 00001000 f3ff4000 c0000000 f5464400 00000000 f89e6868
f7792c80
       80044121 f3ff5f70 bffef4e4 f5464400 f8e325b2 bffef4e4 f3ff5f74
0000000c
Call Trace:
 [<f89e6f20>] snd_pcm_mmap_data_nopage+0x0/0xc0 [snd_pcm]
 [<f89e6868>] snd_pcm_kernel_playback_ioctl+0x38/0x50 [snd_pcm]
 [<f8e325b2>] snd_pcm_oss_get_ptr+0x82/0x1e0 [snd_pcm_oss]
 [<c01727e0>] sys_ioctl+0xe0/0x2c0
 [<c011c180>] do_page_fault+0x0/0x530
 [<c035050f>] error_code+0x2f/0x38
                                                                                
Code: 24 b0 00 00 00 04 0f 84 7c fd ff ff e9 59 fe ff ff 8b 47 30 85 c0

74 07 8b 40 08 85 c0 75 55 c7 04 24 80 76 36 c0 e8 2d 6c 00 00 <0f> 0b 6a 01 12 73 36 c0 8b 94 24 ac 00 00 00 f6 42 32 02 74 1f

Signature · 1 line
Ronny V. Vindenes <[email protected]>
#6 lore raw
Quoted 1 line · William Lee Irwin III
Quoted 1 line

Er, sorry, try this one instead:

Quoted 2 lines · Ronny V. Vindenes

bad nopage snd_pcm_mmap_data_nopage+0x0/0xc0 [snd_pcm] handle_mm_fault() returned bad status

Fix ETA 5 minutes.

-- wli

#7 lore raw
In reply to Ronny V. Vindenes
Quoted 2 lines · Ronny V. Vindenes

bad nopage snd_pcm_mmap_data_nopage+0x0/0xc0 [snd_pcm] handle_mm_fault() returned bad status

sound/core/pcm_native.c +6 -2
diff -prauN mm4-2.6.0-test9-1/sound/core/pcm_native.c mm4-2.6.0-test9-dbg-1/sound/core/pcm_native.c
@@ -2819,7 +2819,7 @@ int snd_pcm_mmap_status(snd_pcm_substrea
 	return 0;
 }
 
-static struct page * snd_pcm_mmap_control_nopage(struct vm_area_struct *area, unsigned long address, int no_share)
+static struct page * snd_pcm_mmap_control_nopage(struct vm_area_struct *area, unsigned long address, int *type)
 {
 	snd_pcm_substream_t *substream = (snd_pcm_substream_t *)area->vm_private_data;
 	snd_pcm_runtime_t *runtime;
@@ -2831,6 +2831,8 @@ static struct page * snd_pcm_mmap_contro
 	page = virt_to_page(runtime->control);
 	if (!PageReserved(page))
 		get_page(page);
+	if (type)
+		*type = VM_FAULT_MINOR;
 	return page;
 }
 
@@ -2869,7 +2871,7 @@ static void snd_pcm_mmap_data_close(stru
 	atomic_dec(&substream->runtime->mmap_count);
 }
 
-static struct page * snd_pcm_mmap_data_nopage(struct vm_area_struct *area, unsigned long address, int no_share)
+static struct page * snd_pcm_mmap_data_nopage(struct vm_area_struct *area, unsigned long address, int *type)
 {
 	snd_pcm_substream_t *substream = (snd_pcm_substream_t *)area->vm_private_data;
 	snd_pcm_runtime_t *runtime;
@@ -2897,6 +2899,8 @@ static struct page * snd_pcm_mmap_data_n
 	}
 	if (!PageReserved(page))
 		get_page(page);
+	if (type)
+		*type = VM_FAULT_MINOR;
 	return page;
 }
 
#8 lore raw
Quoted 6 lines · William Lee Irwin III
Quoted 2 lines · Ronny V. Vindenes

bad nopage snd_pcm_mmap_data_nopage+0x0/0xc0 [snd_pcm] handle_mm_fault() returned bad status

diff -prauN mm4-2.6.0-test9-1/sound/core/pcm_native.c mm4-2.6.0-test9-dbg-1/sound/core/pcm_native.c

That fixed it, thanks!

Signature · 1 line
Ronny V. Vindenes <[email protected]>