Coda File System

Re: venus problem?

From: Jan Harkes <jaharkes_at_cs.cmu.edu>
Date: Mon, 30 Oct 2000 09:59:11 -0500
On Sat, Oct 28, 2000 at 09:40:30AM -0400, Greg Troxel wrote:
> With venus sources from yesterday (I think), I found venus wedged in
> an apparent tight loop (R, with tons of cpu time).  gdb showed it to
> be in KillMgrps; it seemed that mgrpents had both prev and next 0.  I
> really don't know what I'm doing here, but I put in the following and
> I'll see what happens.  Perhaps some more asserts are needed.

I have seen this tight loop as well, it is something new. However, I
tried to debug it by turning up the venus logging level and this somehow
stopped the tight looping.

But, when mgrpents.next is a NULL pointer, the following dereferences
should have at least triggered a segfault. Maybe the -O2 optimization
got in the way of reliably getting the variables contents?

!          mgrpent *m = list_entry(p, mgrpent, volhandle);
!          p = m->volhandle.next;

Jan
Received on 2000-10-30 10:31:00