cvs commit: src/sys/vm vm_mmap.c

[Robert Watson]

On Sat, 6 Dec 2003, Alan Cox wrote:

> That's one way of looking at it.  :-)  The other is that we really need
> witness to support lockmgr()-based locks. 

Unfortunately, if we add that, all hell will break loose :-).  Not only
does VFS not always obey its own lock order (quotas, etc), but the lock
order changes as nodes are moved around the namespace using rename, etc. 
Any lockmgr witness support would have to completely ignore vnode locks. 
So far, the approach has been to move to "SMPng locks" for the final
locking scheme for objects, at which point they pick up witness support... 

Robert N M Watson             FreeBSD Core Team, TrustedBSD Projects
robert at fledge.watson.org      Senior Research Scientist, McAfee Research