Re: Names vs. Inodes

Valdis.Kletnieksat_private

On Thu, 19 Jul 2001 13:53:23 PDT, Chris Wright said:
> hmm.  i tried this and open fails...ENOENT.

OK. After actually having examined fs/namei.c, the actual creat() gets stopped
by an IS_DEADDIR() check in may_create(), the S_DEAD flag having been
set in vfs_rmdir().

However, ENOENT is *not* the only possible failure mode here.  Fortunately
for the LSM side of things, the other ways the test program can fail
do so in a manner that leaves a pathname available.

Interestingly enough, AIX 4.3.3 complains on the rmdir() with an EBUSY,
and Solaris 8 and IRIX 6.5 complain on the rmdir() with an EINVAL.  The
Linux 2.4.5 kernel lets it happen - but this appears to be file system
specific, as the fun starts down in dir->i_op->rmdir(dir, dentry) as
called by vfs_rmdir().  ext2 and reiserfs both let the rmdir() work.
On a filesystem that fails the rmdir(), the creat() call will then succeed.

Comments in fs/namei.c:

/*
 * We try to drop the dentry early: we should have
 * a usage count of 2 if we're the only user of this
 * dentry, and if that is true (possibly after pruning
 * the dcache), then we drop the dentry now.
 *
 * A low-level filesystem can, if it choses, legally
 * do a
 *
 *      if (!d_unhashed(dentry))
 *              return -EBUSY;
 *
 * if it cannot handle the case of removing a directory
 * that is still in use by something else..
 */
static void d_unhash(struct dentry *dentry)
{

There are some comments for vfs_rename_dir() that indicate that
there's other places where opened-but-unlinked directories are
problematic:

 *      d) some filesystems don't support opened-but-unlinked directories,
 *         either because of layout or because they are not ready to deal with
 *         all cases correctly. The latter will be fixed (taking this sort of
 *         stuff into VFS), but the former is not going away. Solution: the same
 *         trick as in rmdir().

Just be glad we don't have any filesystems that support VMS-style versioning
or IBM-style GDG's ;)
-- 
				Valdis Kletnieks
				Operating Systems Analyst
				Virginia Tech