Re: [RFC] timers, pointers to functions and type safety

[Al Viro <viro@ftp.linux.org.uk>]

On Sat, Dec 02, 2006 at 07:06:43PM +0100, Thomas Gleixner wrote:
> On Sat, 2006-12-02 at 09:02 -0700, Matthew Wilcox wrote:
> > On Sat, Dec 02, 2006 at 03:45:12PM +0100, Thomas Gleixner wrote:
> > > What's the cruft ? 
> > > 
> > > struct bla = container_of(timer, struct bla, timer); ???
> > 
> > That's it, right there.  Any idea how much we've bloated the kernel with
> > sysfs, just by insisting that the struct device not be the first item in
> > the struct?  There's any number of 2- and 3- line functions calling each
> > other, each adding and subtracting constants from the pointers passed to
> > them.  This was a huge mistake, IMO.
> 
> What a nonsense.
> 
> 	foo->timer.data = foo;
> 
> is complete redundant information.
> 
> This is going to make a lot of data structures smaller, when the
> timer_list is embedded in the structure itself and for the lot, which
> ignores the timer callback argument anyway.

container_of => still lousy type safety.  All over the sodding place.