On Tue, Apr 24, 2001 at 10:21:58PM +0200, Manuel Bouyer wrote:
> On Tue, Apr 24, 2001 at 09:21:14AM -0700, Jeff Rizzo wrote:
> > 
> > Well, I just got another panic... interestingly enough, the build
> > was in *the same directory*, even though it was doing an "rm" this time,
> > instead of a linking/compiling:
> > 
> > dev = 0x404, block = 3840, fs = /usr
> > panic: blkfree: freeing free frag   
> > syncing disks... panic: lockmgr: locking against myself
> >                                                        
> > (no siop messages this time, either)
> 
> Did you fsck your filesystem before ?

Yep.  The box wouldn't come up until I fscked /usr from single-user mode.

> 
> Would be interesting to know if you get corruption without siop messages
> (the siop messages may just be consequence of data corruption in memory)

Well, on this particular panic, there were no siop messages... and I did
need to fsck from single-user again.

I haven't tried duplicating it again...  I think I could relatively easily
do so, as it hasn't gotten through a 'make build' yet...

Does this look like a hardware problem, or could the FS be messed up
somehow which is causing this problem?  (I had done 'make build' on this
machine before... this *may* have been the first one since upgrading the
kernel to 1.5.1_BETA from a 1.5.1_ALPHA)

Thanks,

+j
-- 
Jeff Rizzo                                         http://boogers.sf.ca.us/~riz