>
>I managed to get a NULL pointer dereference in the kernel. It occured
>while doing a system update under Debian's package manager. The AMD script
>segfaulted, the error appeared, and all went on as normal from there.
>
>
>Setting up amd (upl102-22) ...
>Using existing AMD configuration.
>/etc/init.d/amd: line 2: 5919 Segmentation fault start-stop-daemon
>--test --stop --exec /usr/sbin/amd >/dev/null 2>&1
>Starting automounter: amd.
>
>Line two in /etc/init.d/amd is blank. Line one is the shell specifier
>(i.e. "#!/bin/sh").
>
>Unable to handle kernel NULL pointer dereference at virtual address
>00000008
>current->tss.cr3 = 00638000, %cr3 = 00638000
>*pde = 00000000
>Oops: 0000
>CPU: 0
>EIP: 0010:[<c013ea19>]
>EFLAGS: 00010286
>eax: 00000000 ebx: 171d0007 ecx: c3fd79a0 edx: 0000171d
>esi: 00000001 edi: c11ef5e0 ebp: c11ef5e0 esp: c1ab3f24
>ds: 0018 es: 0018 ss: 0018
>Process start-stop-daem (pid: 5919, process nr: 10, stackpage=c1ab3000)
>Stack: 0000a1c0 c012abe0 c2f0e7e0 00000001 c2f0e7e0 c1c8c480 c0140217
>c2f0e7e0
> 00000001 c11ef5e0 c1ab2000 c1c8c480 c012ae27 c1c8c480 c11ef5e0
>c1c8c480
> c1c8c480 c22bd00e c012aff9 c11ef5e0 c1c8c480 c22bd000 bffffbb4
>08048950
>Call Trace: [<c012abe0>] [<c0140217>] [<c012ae27>] [<c012aff9>]
>[<c012b058>] [<c0129006>] [<c01098b8>]
>Code: 8b 40 08 85 c0 75 08 eb 1a 89 f6 31 c0 eb 19 39 c8 74 f8 8b
>
>(root) <~>$ ksymoops /System.map < oops
>Using `/System.map' to map addresses to symbols.
>
>>>EIP: c013ea19 <proc_permission+71/9c>
>Trace: c012abe0 <permission+24/f0>
>Trace: c0140217 <proc_follow_link+1b/118>
>Trace: c012ae27 <do_follow_link+43/84>
>Trace: c012aff9 <lookup_dentry+191/1c4>
>Trace: c012b058 <__namei+2c/80>
>Trace: c0129006 <sys_newstat+e/64>
>Trace: c01098b8 <system_call+38/40>
>Code: c013ea19 <proc_permission+71/9c>
>Code: c013ea19 <proc_permission+71/9c> 8b 40 08 movl
>0x8(%eax),%eax
>Code: c013ea1c <proc_permission+74/9c> 85 c0 testl %eax,%eax
>Code: c013ea1e <proc_permission+76/9c> 75 08 jne c013ea28
><proc_permission+80/9c>
>Code: c013ea20 <proc_permission+78/9c> eb 1a jmp c013ea3c
><proc_permission+94/9c>
>Code: c013ea22 <proc_permission+7a/9c> 89 f6 movl %esi,%esi
>Code: c013ea2a <proc_permission+82/9c> 31 c0 xorl %eax,%eax
>Code: c013ea2c <proc_permission+84/9c> eb 19 jmp c013ea41
><proc_permission+99/9c>
>Code: c013ea2e <proc_permission+86/9c> 39 c8 cmpl %ecx,%eax
>Code: c013ea30 <proc_permission+88/9c> 74 f8 je c013ea24
><proc_permission+7c/9c>
>Code: c013ea32 <proc_permission+8a/9c> 8b 00 movl (%eax),%eax
>Code: c013ea3a <proc_permission+92/9c> 90 nop
>Code: c013ea3b <proc_permission+93/9c> 90 nop
>Code: c013ea3c <proc_permission+94/9c> 90 nop
>
>
I *sometimes* get the same thing except it's caused by 'killall5' during
shutdown or reboot... not a fun thing having your kernel OOPS during shutdown :)
It too was caused by the same code that caused that one... I think procfs
might need to be examined.
>
>-
>To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>the body of a message to majordomo@vger.rutgers.edu
>
--------------------------------------------------------------------------
Sean M. Kelly | Bill Gates must truly be a genius to successfully
smkelly@zombie.org | market so many programming failures.
--------------------------------------------------------------------------
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu