DragonFlyBSD

:Sometimes DragonFly 1.4.4 crashes.
:It often crashes while using "pkg_chk -su", but also while sufing
:(links) or chatting (irssi).
:Because I can't get my crashdumps saved I made a picture of the output
:of trace: http://img516.imageshack.us/img516/2177/trace5hd.jpg (204kb).
:My dmesg: http://pastebin.com/736049

Do you have more of that picture?  Perhaps all the lines above the 
    ufs_vnoperate() line ?  You should be able to scroll up from the DDB
    prompt.

Hmm.  A 2.6 GHz celeron?  Are you overclocking it?  I know it sounds
    like a cop-out, I hate suggesting it, but check the fans in the
    box and if you are overclocking it try returning it to the base BIOS
    config.  If it is a laptop, check for overheating.  The high-end
    celeron cpus are usually not clocked that fast.

-Matt
                    Matthew Dillon 
                    <dillon@backplane.com>

On Wed, May 24, 2006 at 04:48:34PM -0700, Matthew Dillon wrote:

:Sometimes DragonFly 1.4.4 crashes.
:It often crashes while using "pkg_chk -su", but also while sufing
:(links) or chatting (irssi).
:Because I can't get my crashdumps saved I made a picture of the output
:of trace: http://img516.imageshack.us/img516/2177/trace5hd.jpg (204kb).
:My dmesg: http://pastebin.com/736049

Do you have more of that picture? Perhaps all the lines above the
ufs_vnoperate() line ? You should be able to scroll up from the DDB
prompt.

I think that's the problem I have reported as well.

Joerg

Matthew Dillon wrote:

Do you have more of that picture? Perhaps all the lines above the
ufs_vnoperate() line ? You should be able to scroll up from the DDB
prompt.

One line above the ufs_vnoperate() I entered the trace-command "db>
trace" I made made some inputs to get it to save the crashdump. I fact
only "call sysdump" and "help". "call sysdump" made the same output like
the original crash.
http://img240.imageshack.us/img240/7591/shot29go.jpg

It says:
Fatal trap 120 page fault while in kernel mode
fault virtual adress = 0x0
fault code = supervisor write, page not present
instruction pointer = 0x8:0xc015735c
stack pointer = 0x10:0cd752c723
frame pointer = 0x10:0xd752c747
code segment = base 0x0, limit 0xfffff, type 0x1b = DPL 0, press 1, def32 1, gran 1
processor eflags = interrupt enabled, resume, IOPL = 0
current process = 61134 (sh)
current thread = pri 70 (CRIT)

kernel: type 12 trap, code=2
Stopped at ufs_vnoperate+0xb:

Sorry for the bad quality (my camera needs the flash or the shot is
really not readable).

Hmm. A 2.6 GHz celeron? Are you overclocking it? I know it sounds
like a cop-out, I hate suggesting it, but check the fans in the
box and if you are overclocking it try returning it to the base BIOS
config. If it is a laptop, check for overheating. The high-end
celeron cpus are usually not clocked that fast.

Yep, it's a celeron d 331. 2.66 GHz is the default clockspeed and it's
not overclocked.

It's not a laptop, but it's really getting very hot.
I bought a new fan and even a new case. Maybe I should check the
heatsink paste...

reezer

Is this still present on 1.6-RELEASE?

originator can't reproduce problem anymore