On Aug 3, 2005, at 1:06 PM, Julio M. Merino Vidal wrote:

> Then, I have problems with the rmdir operation itself.  I first
> detach a_vp from its tmpfs_node by setting v_data and
> tn_node both to NULL.  Then I call vput over a_vp, as other
> file-systems do.  After this, I free the associated tmpfs_node
> given that no-one else is pointing to it.

But you'll note that other file systems don't free their v_data.   
That's because there is only one way that the VFS layer asks the file  
system to tear down the relationship.  That's the VOP_RECLAIM()  
path.  That is the only place where you should be tearing your  
tmpfs_node down.

> *But* a_vp is not really "released".  When I unmount the
> file-system, the kernel calls the fsync and reclaim operations
> over the value a_vp had (the vnode used during rmdir).  As
> this vnode has v_data set to NULL, any attempt to access the
> structure obviously fails and panics the system.

Right.  So don't do that in your rmdir routine... wait for the VFS  
layer to reclaim the vnode, then you should be OK.

-- thorpej