Subject: Re: nfsd: locking botch in op %d
To: Chuck Silvers <chuq@chuq.com>
From: Bill Studenmund <wrstuden@zembu.com>
List: tech-kern
Date: 03/14/2001 09:25:44
On Tue, 13 Mar 2001, Chuck Silvers wrote:

> one of the functions of vnode locks is to provide POSIX atomic read/write
> semantics, which is unneeded and undesirable for device vnodes.

Plus device drivers can sleep, which you aren't supposed to do if you have
a locked vnode. :-)

> On Tue, Mar 13, 2001 at 04:08:44PM -0800, Jason R Thorpe wrote:
> > On Tue, Mar 13, 2001 at 02:29:05PM -0800, Bill Studenmund wrote:
> > 
> >  > That actually is the other way to fix the problem, though with doing this
> >  > the vput really should be a vrele() (since otherwise you're implicitly
> >  > relying on specfs using the nolock routines :-)
> > 
> > Is there any reason to keep using the "nolock" routines?  Well, a few
> > things still use them.  But it seems like you want locking on e.g. spec
> > nodes, right?
> > 
> > -- 
> >         -- Jason R. Thorpe <thorpej@zembu.com>
>