[Linux-cluster] segfault

[Patrick Caulfield]

Javi Polo wrote:
> While doing cman_tool join ... I got a segfault and this on the logs:
> 
> Unable to handle kernel NULL pointer dereference at virtual address 0000019c
>  printing eip:
> c034e80a
> *pde = 00000000
> Oops: 0000 [#12]
> Modules linked in: gfs lock_harness dlm cman ipv6 snd_intel8x0 snd_ac97_codec snd_pcm_oss snd_mixer_oss snd_pcm snd_timer snd soundcore snd_page_alloc
> CPU:    0
> EIP:    0060:[<c034e80a>]    Not tainted VLI
> EFLAGS: 00010296   (2.6.12.3) 
> EIP is at sk_alloc+0x1b/0xd0
> eax: c1551180   ebx: 00000002   ecx: ffffff9c   edx: 000000d0
> esi: 00000134   edi: 000000d0   ebp: ffffff9f   esp: c2213ed8
> ds: 007b   es: 007b   ss: 0068
> Process cman_tool (pid: 4034, threadinfo=c2212000 task=c993fa80)
> Stack: ca2850ac c886c074 00000286 00000002 cc1f3380 000000d0 ffffff9f e1aa90e7 
>        0000001e 000000d0 00000134 c1551180 00000002 cc1f3380 00000002 e1aa933e 
>        cc1f3380 000000d0 0000001e cc1f3380 c034c923 cc1f3380 00000002 00000001 
> Call Trace:
>  [<e1aa90e7>] cl_alloc_sock+0x38/0x97 [cman]
>  [<e1aa933e>] cl_create+0x59/0x101 [cman]
>  [<c034c923>] __sock_create+0xc3/0x1c7
>  [<c034ca56>] sock_create+0x2f/0x33
>  [<c034cab5>] sys_socket+0x28/0x55
>  [<c034d9ff>] sys_socketcall+0x89/0x251
>  [<c0149e59>] filp_close+0x52/0x96
>  [<c010dcbc>] do_page_fault+0x0/0x5bf
>  [<c0102865>] syscall_call+0x7/0xb
> Code: ff ff ff c7 44 24 14 04 00 00 00 e9 75 fc ff ff 83 ec 1c 89 74 24 10 89 5c 24 0c 89 7c 24 14 89 6c 24 18 8b 74 24 28 8b 54 24 24 <8b> 46 68 85 c0 0f 84 8c 00 00 00 89 54 24 04 89 04 24 e8 a8 79 
> 
> Has anybody a hint?
> I compiled the kernel modules, and also made debian packages from the sources out there at ubuntu ... I'm with 2.6.12.3 in debian/sid .....
> 

The sk_alloc code has changed a few times in the kernel so it might be that the
source you have compiled doesn't match the kernel it is running on. Though that
usually results in a compile error rather than a runtime one!

Which source are you using?

-- 

patrick