Subject: Re: New panic
To: None <current-users@NetBSD.ORG>
From: Frank van der Linden <frank@wins.uva.nl>
List: current-users
Date: 07/09/1997 14:24:23
Quoting myself,
> > (the beginning of the trace in brief:
> > ufs_lock
> > vndstrategy
> > spec_strategy
> > bwrite
> > vn_bwrite
> > bawrite
> > getnewbuf
> > allocbuf
> > getblk
> > ufs_bmaparray
> > ufs_bmap
> > vndstrategy
> > spec_strategy
> > ufs_strategy
> > bwrite
> > vn_bwrite
> > bawrite
> > ...
> Hm, this would appear to be a possible vnd problem.
This appears to be the scenario:
- call vndstrategy
- vnode gets locked
- ufs_bmap is called
- an available buffer is looked for
- while doing this, a DELWRI buffer on the same vnode is found
- a bawrite() is done on it
- vndstrategy again
- try to lock the vnode as in step 2, but it's already locked.
-> panic
- Frank