Hardware or software error ?

[Heinrich Rebehn]

Lowell Gilbert wrote:
> Heinrich Rebehn <rebehn at ant.uni-bremen.de> writes:
> 
> 
>>Hi list,
>>
>>does anybody have a clue, if the following is a hard or software error?
>>
>>#########################################################################
>>  syslogd: kernel boot file is /boot/kernel/kernel
>>  kernel:
>>  kernel:
>>  kernel: Fatal trap 12: page fault while in kernel mode
>>  kernel: cpuid = 0; apic id = 00
>>  kernel: fault virtual address   = 0x4
>>  kernel: fault code              = supervisor read, page not present
>>  kernel: instruction pointer     = 0x8:0xc0533f98
>>  kernel: stack pointer           = 0x10:0xe11f6b3c
>>  kernel: frame pointer           = 0x10:0xe11f6b64
>>  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         = 29 (swi1: net)
>>  kernel: trap number             = 12
>>  kernel: panic: page fault
>>  kernel: cpuid = 0;
>>  kernel:
>>  kernel: syncing disks, buffers remaining...
>>  kernel:
>>  kernel: Fatal trap 12: page fault while in kernel mode
>>  kernel: cpuid = 0; apic id = 00
>>  kernel: fault virtual address   = 0x4
>>  kernel: fault code              = supervisor read, page not present
>>  kernel: instruction pointer     = 0x8:0xc0533f98
>>  kernel: stack pointer           = 0x10:0xe124bbcc
>>  kernel: frame pointer           = 0x10:0xe124bbf4
>>  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         = 26 (irq15: xl0 ata1+)
>>  kernel: trap number             = 12
>>  kernel: panic: page fault
>>  kernel: cpuid = 0;
>>  kernel: Uptime: 3d10h11m49s
>>#########################################################################
>>
>>The system is running FreeBSD 5.2.1-RC2
> 
> 
> More likely to be a software problem, although it could be either.  
> Could you take a kernel dump to the -CURRENT list?

I'm afraid my system is not set up to enable crash dumps. I must admit
that i never cared about this, and unfortunately, the default install
does not seem to enable it either. If i am wrong: where would i find the
dump?

Also, since this is our main server, i prefer going back to 4.9 which
seems to be more solid than 5.2 :-(. I simply don't have time to do more
experiments.

Update: This morning's crash (which i forget in the mail before):

#####################################################################
ntpd[470]: too many recvbufs allocated (40)


cron[8309]: login_getclass: unknown class 'des_users'


syslogd: kernel boot file is /boot/kernel/kernel


kernel:


kernel:


kernel: Fatal trap 12: page fault while in kernel mode


kernel: cpuid = 0; apic id = 00


kernel: fault virtual address   = 0x4c


kernel: fault code              = supervisor read, page not present


kernel: instruction pointer     = 0x8:0xc04cc807


kernel: stack pointer           = 0x10:0xe2cccca8


kernel: frame pointer           = 0x10:0xe2ccccc8


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         = 38 (usbtask)


kernel: trap number             = 12


kernel: panic: page fault


kernel: cpuid = 0;


kernel:


kernel: syncing disks, buffers remaining... 7137 7117 7117 7117 7117
7117 7117 7117 7117 7117 7117 7117 7117 7117 7117 7117 7117 7117 7117
7117 7117
kernel: giving up on 4591 buffers


kernel: Uptime: 14h28m36s


kernel: (da0:umass-sim0:0:0:0): Synchronize cache failed, status ==
0x34, scsi status == 0x88
########################################################################

The "giving up on  nnnn Buffers" is also typical for 5.2. I rarely had
it with 4.9. Unfortunately, all disks are marked dirty then, even if it
is only 1 buffer.

Update:

I changed my mind and will continue to try 5.2. I disabled softupdates 
to get rid of the "softupdates inconsistency error" and set up the 
machine to enable crash dumps, as described in
http://www.freebsd.org/doc/en_US.ISO8859-1/books/developers-handbook/kerneldebug.html
I also clocked down the AMD XP1800+ CPU to 100 MHz bus clock to reduce 
possible hardware instability.
All we have to now is wait for the next crash ;-)

Heinrich