Subject: Re: vput() instead of vrele() in sys_mount() for MNT_UPDATE case?
To: Chris Jepeway <email@example.com>
From: Bill Studenmund <firstname.lastname@example.org>
Date: 04/16/2002 08:56:26
On Mon, 15 Apr 2002, Chris Jepeway wrote:
> Should the vrele() in sys_mount() around line 313 of revision 1.173 of
> vfs_sycalls.c instead be a vput()? Seems like the lock on the mount
> point needs to be dropped.
Unclear. We drop the lock before the goto updating, so the lock should be
open at that point. It seems we call VFS_MOUNT() with an unlokced vnode.
The sick thing is that we seem to call VFS_MOUNT() with a _locked_ vnode
the first time we make the mount. Being inconsistent in how we call mount
strikes me as very bad.