Skip to contentSkip to conversation

2.6.2-rc3 messages BUG

DISCUSSION John Levin 4 participants 4 messages Feb 20, 2004 – Feb 21, 2004

lore
Share & tools
lore

John Levin

author
#0 lore raw

Hi,
My guess atleast in this case is that suspend/resume cyle looses track
of the fact that a module is use.I have file corruption. All those files which have been created after resume is corrupted. I copied dmesg into a backup file and saved it. When i boot 2.4 and look into it , it is corrputed.
After booting I connect to the internet through wvdial. So i have to
load up usbcore,cdc_acm,uhci. i am connected to the net and searching on
google. Now i do echo 4 > /proc/acpi/sleep . It suspends. Then i resume
it from command line.
So now wvdial looks as if connected but really isn't. So i close it and
try running it again. It doesn't detect /dev/usb/acm/0. So i remove the modules and try inserting it (uhci) which gives me the error.

Here is something which i could copy after resume.

--> WvDial: Internet dialer version 1.53
--> Initializing modem.
--> Sending: ATZ
--> Sending: ATQ0
--> Re-Sending: ATZ
--> Modem not responding.
lsmod
[root@mdk9 root]# lsmod

Module                  Size  Used by
uhci_hcd               31752  0
cdc_acm                10784  3
usbcore               111828  4 uhci_hcd,cdc_acm

[root@mdk9 root]# rmmod uhci_hcd
[root@mdk9 root]# insmod /lib/modules/2.6.3-rc2/kernel/drivers/usb/host/uhci-hcd.ko

------------[cut here ]------------
kernel BUG at mm/slab.c:1269!

invalid operand: 0000 [#1]
CPU:    0
EIP:    0060:[<c0143a29>]    Not tainted
EFLAGS: 00010202

EIP is at kmem_cache_create+0x509/0x670
eax: 00000031 ebx: c130277c ecx: c04a50e8 edx: c03cc3f8
esi: cf935fa7 edi: cf935fa7 ebp: cbdedf74 esp: cbdedf44
ds: 007b es: 007b ss: 0068
Process insmod (pid: 2234, threadinfo=cbdec000 task=cc1fa6a0)

Stack: c036e980 cf935f99 00010c00 cbdedf64 c13026a4 c0000000 c1302668
fffffffc       00000020 00000000 fffffff4 cf938980 cbdedf9c cf91d0d4
cf935f99 00000044       00000080 00010c00 00000000 00000000 c03ceb70
c03ceb58 cbdedfbc c0137aeb Call Trace:
 [<cf91d0d4>] uhci_hcd_init+0xd4/0x12e [uhci_hcd]
 [<c0137aeb>] sys_init_module+0xeb/0x1c0
 [<c010b1df>] syscall_call+0x7/0xb

Code: 0f 0b f5 04 5a df 36 c0 8b 0b e9 71 ff ff ff 8b 47 34 c7 04

Segmentation fault

--

#1 lore raw

Hi!

Quoted 13 lines

My guess atleast in this case is that suspend/resume cyle looses track
of the fact that a module is use.I have file corruption. All those files which have been created after resume is corrupted. I copied dmesg into a backup file and saved it. When i boot 2.4 and look into it , it is corrputed.
After booting I connect to the internet through wvdial. So i have to
load up usbcore,cdc_acm,uhci. i am connected to the net and searching on
google. Now i do echo 4 > /proc/acpi/sleep . It suspends. Then i resume
it from command line.
So now wvdial looks as if connected but really isn't. So i close it and
try running it again. It doesn't detect /dev/usb/acm/0. So i remove the modules and try inserting it (uhci) which gives me the error.

Quoted 1 line

Here is something which i could copy after resume.

Try it with minimal config, and without modules. Be sure to fsck so you don't kill your filesystem totally. Try 2.6.3.
Pavel

Signature · 2 lines
When do you have a heart between your knees?
[Johanka's followup: and *two* hearts?]
#2 lore raw
In reply to John Levin
Quoted 25 lines · John Levin

Hi,
My guess atleast in this case is that suspend/resume cyle looses track
of the fact that a module is use.I have file corruption. All those files which have been created after resume is corrupted. I copied dmesg into a backup file and saved it. When i boot 2.4 and look into it , it is corrputed.
After booting I connect to the internet through wvdial. So i have to
load up usbcore,cdc_acm,uhci. i am connected to the net and searching on
google. Now i do echo 4 > /proc/acpi/sleep . It suspends. Then i resume
it from command line.
So now wvdial looks as if connected but really isn't. So i close it and
try running it again. It doesn't detect /dev/usb/acm/0. So i remove the modules and try inserting it (uhci) which gives me the error.

Here is something which i could copy after resume.

--> WvDial: Internet dialer version 1.53
--> Initializing modem.
--> Sending: ATZ
--> Sending: ATQ0
--> Re-Sending: ATZ
--> Modem not responding.
lsmod
[root@mdk9 root]# lsmod

Quoted 7 lines
Module                  Size  Used by
uhci_hcd               31752  0
cdc_acm                10784  3
usbcore               111828  4 uhci_hcd,cdc_acm

[root@mdk9 root]# rmmod uhci_hcd
[root@mdk9 root]# insmod /lib/modules/2.6.3-rc2/kernel/drivers/usb/host/uhci-hcd.ko

You missed out an important piece of info. The kernel should have printed out "kmem_cache_create: duplicate cache <name>" before going BUG.

What was "<name>"? uhci_urb_priv?

I suggest you go into drivers/usb/host/uhci-hcd.c:uhci_hcd_cleanup() and replace

	warn("not all urb_priv's were freed!");

with

	BUG();

because failure to destroy that slab cache is fatal, and it points at a bug in this driver.

Quoted 18 lines

------------[cut here ]------------
kernel BUG at mm/slab.c:1269!

invalid operand: 0000 [#1]
CPU:    0
EIP:    0060:[<c0143a29>]    Not tainted
EFLAGS: 00010202

EIP is at kmem_cache_create+0x509/0x670
eax: 00000031 ebx: c130277c ecx: c04a50e8 edx: c03cc3f8
esi: cf935fa7 edi: cf935fa7 ebp: cbdedf74 esp: cbdedf44
ds: 007b es: 007b ss: 0068
Process insmod (pid: 2234, threadinfo=cbdec000 task=cc1fa6a0)

Stack: c036e980 cf935f99 00010c00 cbdedf64 c13026a4 c0000000 c1302668
fffffffc       00000020 00000000 fffffff4 cf938980 cbdedf9c cf91d0d4
cf935f99 00000044       00000080 00010c00 00000000 00000000 c03ceb70
c03ceb58 cbdedfbc c0137aeb Call Trace:
 [<cf91d0d4>] uhci_hcd_init+0xd4/0x12e [uhci_hcd]
 [<c0137aeb>] sys_init_module+0xeb/0x1c0
 [<c010b1df>] syscall_call+0x7/0xb
#3 lore raw
Quoted 17 lines · * Andrew Morton ([email protected]) wrote:
Quoted 8 lines

[root@mdk9 root]# lsmod

Module                  Size  Used by
uhci_hcd               31752  0
cdc_acm                10784  3
usbcore               111828  4 uhci_hcd,cdc_acm

[root@mdk9 root]# rmmod uhci_hcd
[root@mdk9 root]# insmod /lib/modules/2.6.3-rc2/kernel/drivers/usb/host/uhci-hcd.ko

You missed out an important piece of info. The kernel should have printed out "kmem_cache_create: duplicate cache <name>" before going BUG.

What was "<name>"? uhci_urb_priv?

I suggest you go into drivers/usb/host/uhci-hcd.c:uhci_hcd_cleanup() and replace

Quoted 8 lines
	warn("not all urb_priv's were freed!");

with

	BUG();

because failure to destroy that slab cache is fatal, and it points at a bug in this driver.

True, I mentioned this to Greg earlier. But in this case that BUG() is only going to show the rmmod attempt, right? Problem is that rmmod works when the driver is in use, and this only happens after suspend/resume.

thanks,
-chris

Signature · 1 line
Linux Security Modules     http://lsm.immunix.org     http://lsm.bkbits.net