[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#534422: libc6: Lots of segfaults after hibernation resume "error 6 in libc-2.9.so"

On Wed, Jun 24, 2009 at 10:05:35AM +0100, George B. wrote:
> I seem to be having a lot of problems resuming from hibernation recently. Often after resuming I will start getting lots of segfaults from varios programs until system eventually becomes inoperative (can't even log in at console because bash segfaults).

Am I correct that all new processes die after the resume?

> I don't know if it's relevant but I also enabled KMS a while back - could be a coincidence. Disabled it now will check if it makes any difference.

Kernel mode-setting is new in 2.6.30, but the bugreport shows a 2.6.29.

> 0xb7fe6424 in __kernel_vsyscall ()
> #0  0xb7fe6424 in __kernel_vsyscall ()

It segfaults within vsyscall. This would mean that the kernel either
select a wrong vdso for the hardware or otherwise corrupts it.

Because of the used address, I would assume the sysenter vdso, which
looks like this:

| Dump of assembler code for function __kernel_vsyscall:
| 0xAAAAA414 <__kernel_vsyscall+0>:       push   %ecx
| 0xAAAAA415 <__kernel_vsyscall+1>:       push   %edx
| 0xAAAAA416 <__kernel_vsyscall+2>:       push   %ebp
| 0xAAAAA417 <__kernel_vsyscall+3>:       mov    %esp,%ebp
| 0xAAAAA419 <__kernel_vsyscall+5>:       sysenter 
| 0xAAAAA41b <__kernel_vsyscall+7>:       nop    
| 0xAAAAA41c <__kernel_vsyscall+8>:       nop    
| 0xAAAAA41d <__kernel_vsyscall+9>:       nop    
| 0xAAAAA41e <__kernel_vsyscall+10>:      nop    
| 0xAAAAA41f <__kernel_vsyscall+11>:      nop    
| 0xAAAAA420 <__kernel_vsyscall+12>:      nop    
| 0xAAAAA421 <__kernel_vsyscall+13>:      nop    
SYSENTER_RETURN-2 aka restart point:
| 0xAAAAA422 <__kernel_vsyscall+14>:      jmp    0xAAAAA417 <__kernel_vsyscall+3>
| 0xAAAAA424 <__kernel_vsyscall+16>:      pop    %ebp
| 0xAAAAA425 <__kernel_vsyscall+17>:      pop    %edx
| 0xAAAAA426 <__kernel_vsyscall+18>:      pop    %ecx
| 0xAAAAA427 <__kernel_vsyscall+19>:      ret    

This code has not changed in the last versions.

Please provide complete hardware informations about this system.


Beam me up, Scotty!

Reply to: