panic: page fault

[kalin mintchev]

> On Sun, Nov 14, 2004 at 09:04:32PM -0500, kalin mintchev wrote:
>> ok...  about every 6 - 7 days the machine reboots because of this:
>>
>>  /kernel: Fatal trap 12: page fault while in kernel mode
>>  /kernel: fault virtual address    = 0xbeff2
>>  /kernel: fault code               = supervisor read, page not present
>>  /kernel: instruction pointer      = 0x8:0xc0305ed3
>>  /kernel: stack pointer            = 0x10:0xd7ec9e18
>>  /kernel: frame pointer            = 0x10:0xd7ec9e20
>>  /kernel: code segment             = base 0x0, limit 0xfffff, type 0x1b
>>  /kernel: = DPL 0, pres 1, def32 1, gran 1
>>  /kernel: processor eflags = interrupt enabled, resume, IOPL = 0
>>  /kernel: current process          = 29374 (perl)
>>  /kernel: interrupt mask           = net tty bio cam
>>  /kernel: trap number              = 12
>>  /kernel: panic: page fault
>>
>> where do i look?
>>
>> thanks..........
>
> You forgot to mention what version of FreeBSD you're running..if it's
> old, upgrade to the latest version, because bugs are located and fixed
> all the time.

thank you Kris. you are right. sorry. FreeBSD4.6-RELEASE

well..  i'm pretty sure the problem is caused by the courier imap server.
it wasn't doing it before i installed qmail and courier.
the courier starts executing a command called 'ulimit'
the original script uses the -v flag to the command to assign virtual
memory. the thing is this bsd version ulimit doesn't support _any_ flags.
when i was starting it with the flag i was getting error message that the
flag is not supported and the machine was crashing after a while with a
message about virtual memory problems. so i took the flag out...  now it
crashes with this panic...

i looked under 5.2.x for the ulimit command but it's the same - no flags
and only under sh. so even if i upgrade to  4.10 i won't get the -v flag
for ulimit...

i posted a question to the courier list about it but was told to research
how ulimit works on my system and adjust the flags. not much of a help
there...

how can i solve this? is there another command instead of ulimit i can use
for the same purpose as ulimit -v ?

thanks a lot...


>
> Is the issue repeatable under defined conditions?  Is the page fault
> always the same?  Either way, you'll also need to rule out hardware
> failure.
>
> Otherwise, see the chapter on kernel debugging in the developers
> handbook on the website for more information on how to obtain
> necessary debugging information suitable for inclusion in a PR.
>
> Kris
>


--