Subject: Re: The demise of DEV_BSIZE
To: Bill Studenmund <firstname.lastname@example.org>
From: Chuck Silvers <email@example.com>
Date: 10/12/1999 09:48:51
On Tue, Oct 12, 1999 at 09:43:55AM -0700, Bill Studenmund wrote:
> On Mon, 11 Oct 1999, Jason Thorpe wrote:
> > On Mon, 11 Oct 1999 18:19:01 -0700 (PDT)
> > Bill Studenmund <firstname.lastname@example.org> wrote:
> > > Hmmm. Because we're filling in specinfo. All the other fields in specinfo
> > > get filled in in checkalias(), so it seems like the logical choice to me.
> > ...and where in the call-graph is checkalias() called? If it's *after*
> > (*d_open)(), you can call (*d_ioctl)().
> ? What call graph are you looking at?
> Using ffs as an example,
> ufs_lookup(), when it resolves the name into an inode # (it's not in the
> cache), calls VFS_VGET(), which mapps to ffs_vget(). ffs_vget() does a
> number of things, and then calls ufs_vinit. ufs_vinit() calls
> I see no call to an open routine in that path. ??
> Take care,
if we're calling VFS_VGET(), then there must be a mounted filesystem in
there somewhere. mounting the filesystem would have opened the device
on which that filesystem is mounted.