uipc_socket2.c panic
Divacky Roman
xdivac02 at stud.fit.vutbr.cz
Wed Jun 2 19:17:56 GMT 2004
On Wed, Jun 02, 2004 at 10:08:00AM -0700, Bosko Milekic wrote:
>
> You wrote:
> >Additional information (I forgot to mention it - I apology)
> >I am running with debug.mpsafenet=1
> >
> >roman
>
> This is highly relevant. It means that there is probably a problem
> in network locking somewhere. I have CC'd Robert to make sure he
> is in on the loop for this.
>
> Here is what I'd like you to try, in this order, please:
I've just cvsuped to latest current and there is no commit to relevant
part of kernel (previous cvsup was from about 15:20 CEST) so I suppose
I can use this kernel.
The panic occurs just with debug.mpsafenet=1, with 0 it runs flawlessly.
And yes I experienced panic even before mbuma (I set that mpsafenet a
few days ago - and soon those panics started)
Summary: the panic is just with mpsafenet=1
more tommorow (sorry)
if mu presumption was correct (I used apropriet kernel) I'll try
tommorow disable preemption. is this ok?
> 1. CVSUP to latest -CURRENT, build a new kernel and boot it.
>
> 2. Set debug.mpsafenet=1, see if you can reproduce the problem. If
> not, then we'll assume it is fixed, so let us know.
>
> 3. If you can reproduce with debug.mpsafenet=1, then set
> debug.mpsafenet=0 and try to reproduce again. If you can
> reproduce again, then go to 4, otherwise let us know and
> we'll go from there.
>
> >From here on, always run with debug.mpsafenet=0...
>
> 4. If you were able to reproduce in (3), try backing out mbuma
> (cvsup to -CURRENT right before it was committed), try to
> reproduce. If you no longer can, let me know. If you still
> can reproduce, goto 5.
>
> 5. If you reproduced in (4), then it's not mbuma, so you can
> cvsup back to fresh sources. Now try disabling preemption
> (I described how to do this in an earlier post). Let us
> know if you can still reproduce following this.
>
> Thank you.
>
> -Bosko
More information about the freebsd-current
mailing list