This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
Re: Kernel hang with latest snapshot
David Boreham wrote:
Sorry, I didn't see your reply when I was sending this out. I will
take a look at iosyscall.open.entry, and will check the remaining
codes of potential invalid pointer access. Thanks.
I see crashes with every io syscall I've tried. It would take quite a
while to
try every one (need to reboot every time). open, read, write, close for
sure crash for me.
:( I tried to probe using addevent.iosyscall and they seems well
for me. My testing environment:
IBM T43 thinkpad with Pentium-M 2.0
Fedora Core5 updated to the latest
gcc 4.1.1 20060525
elfutils 0.124-0.1, rpm packages are from Red Hat ftp server
latest SystemTap
running command: stap -e "probe addevent.iosyscall {}" -bMv
You said almost every io syscall will crash the system, but I am
sure it's not LKET problem. It should be some more low level problem,
like runtime and kprobes. Could you tell me a little more about your
testing environment, and more important, the workload you imposed when
probing using LKET, so that I can try to reproduce the problem. One
reason I am thinking my system didn't crash may be that my workload is
not big enough and didn't get the potential problem exposed. Thanks
- Guanglei