[PATCH#3] fake data io module for samba

[Christof Schmitt]

On Sat, Nov 15, 2014 at 04:20:16PM +0100, Volker Lendecke wrote:
> On Sat, Nov 15, 2014 at 08:02:24AM -0700, Christof Schmitt wrote:
> > On Sat, Nov 15, 2014 at 09:36:09AM +0100, Volker Lendecke wrote:
> > > On Fri, Nov 14, 2014 at 02:32:09PM -0800, Jeremy Allison wrote:
> > > > On Wed, Nov 12, 2014 at 04:20:38PM +0100, Peter Somogyi wrote:
> > > > > Hi Jeremy,
> > > > > 
> > > > > I've added the memset to 0 now and retested the whole patch again.
> > > > > 
> > > > > Pls do a +1 or let me know further doubts.
> > > > 
> > > > +1 from me (probably with some slight tidying
> > > > up of the English in the man page). Ping me
> > > > if you want to see that now.
> > > > 
> > > > Other than that can I get a second Team reviewer ?
> > > 
> > > Can we change it such that it will only be built in
> > > developer mode?
> > 
> > What about a warning printed from vfs_fake_io_init that the module is
> > loaded, and this share should not get used for actual data?
> 
> Do you want that instead of on top of only building it in developer
> mode? I'm not sure I see much reason in having this module around
> everywhere. It of course serves its purpose if you want to know whether
> your file system is slow or another part of the system is the bottleneck
> in your tests, but why do you want to have it on each and every Samba
> system installed?

Since the module strips out valid data, i think it would be good to
always get a warning in the logs when it is active.

About the question when to activate it, i am not quite sure. I guess the
options would be, having it built always, having it not built by
default, or having it only in the developer build. Every option has its
pros and cons. I usually like the idea that code is always built to
detect problems automatically, but we should prevent users from
accidentally introducing problems with it.

Christof