On Sat, Oct 09, 2004 at 11:43:32PM -0700, Matthew Dillon wrote:
> (This is really the second one, but I decided to synchronize the name
> with my internal patch file names to avoid confusion).
>
> !!! All normal users not involved with testing the VFS code should be
> using kernels based on the DragonFly_Stable tag and not mess with this
> stuff !!!
>
> Here's mega patch 09, relative to HEAD. This is not for the faint of
> heart.
>
> fetch http://leaf.dragonflybsd.org/~dillon/vfsx09.patch
>
> Known issues:
>
> * Possible vnode/namecache deadlocks.
> * While most filesystems now use vp->v_lock, I haven't done a final
> pass to make vp->v_lock mandatory and to clean up the few remaining
> inode based locks (nwfs I think and other obscure filesystems).
> * NullFS gets confused when you hit a mount point in the underlying
> filesystem.
> * Only UFS and NFS have been well tested
I saw this in /var/log/messages
Oct 13 00:14:12 gzl kernel: EXDEV case 1 0xcfc4e320 scsi
Oct 13 00:14:12 gzl last message repeated 4 times
Oct 13 00:15:14 gzl kernel: EXDEV case 1 0xcfc4ee60 scsi
Oct 13 00:15:22 gzl last message repeated 9 times
Oct 13 00:17:59 gzl kernel: EXDEV case 1 0xcfc4ee60 scsi
Oct 13 00:22:53 gzl last message repeated 19 times
while I was playing with build breakage:
$ export MAKEOBJDIRPREFIX=$HOME/obj
$ cd /sys/bus/cam/scsi
$ make -s obj && env ARCH=i386 make depend
Has this already been taken care of in HEAD?