Re: no rule to make target param.i386_fsbd_50.h

<<On Fri, 26 Sep 2003 08:31:51 -0400, Jim Rees <[email protected]> said:

> I got to the point of being able to build and load the kernel module and do
> the cache scan.  Then it would panic trying to open CacheItems because the
> inode number was wrong.

I've put an updated patch that continues to fix up some of the locking
and memory allocation issues
(<http://lcs.mit.edu/~wollman/openafs-fbsd52.patch>); this is relative
to the top of tree prior to Jim's commit.  I know that there are at
least three bugs:

- The namei() wrapper can sometimes be called without the global lock
held, so we can't unconditionally release it.  (This is a bug in new
code that I have written.)

- There's a WITNESS warning related to rx calling socreate() when the
global lock is held.

- Shutting down is seriously broken, with rx waiting for a server
reply while the proc lock (?!) is held.  I haven't gotten a backtrace
from this that I understand yet.

Once I fix the first two, I'll turn on DEBUG_VFS_LOCKS and see how
many issues crop up there.

-GAWollman