В Fri, 17 Dec 2010 16:13:28 -0800
Andrew Morton <akpm@linux-foundation.org> пишет:

> On Fri, 17 Dec 2010 22:54:31 +0300
> Alexander Gordeev <lasaine@lvk.cs.msu.su> wrote:
> 
> > Since now idr is only used to manage char device id's and not used in
> > kernel API anymore it should be moved to pps.c. This also makes it
> > possible to release id only at actual device freeing so nobody can
> > register a pps device with the same id while our device is not freed
> > yet.
> > 
> > ...
> >
> >  int pps_register_cdev(struct pps_device *pps)
> >  {
> >  	int err;
> >  	dev_t devt;
> >  
> > +	/* Get new ID for the new PPS source */
> > +	if (idr_pre_get(&pps_idr, GFP_KERNEL) == 0)
> > +		return -ENOMEM;
> > +
> > +	/* Now really allocate the PPS source.
> > +	 * After idr_get_new() calling the new source will be freely available
> > +	 * into the kernel.
> > +	 */
> > +	spin_lock_irq(&pps_idr_lock);
> > +	err = idr_get_new(&pps_idr, pps, &pps->id);
> > +	spin_unlock_irq(&pps_idr_lock);
> > +
> > +	if (err < 0)
> > +		return err;
> 
> The IDR interface really sucks :(
> 
> What this code should be doing is
> 
> retry:
> 	if (idr_pre_get(&pps_idr, GFP_KERNEL) == 0)
> 		return -ENOMEM;
> 	spin_lock_irq(&pps_idr_lock);
> 	err = idr_get_new(&pps_idr, pps, &pps->id);
> 	spin_unlock_irq(&pps_idr_lock);
> 	if (err < 0) {
> 		if (err == -EAGAIN)
> 			goto retry;
> 		return err;
> 	}
> 
> this way it correctly handles the case where the idr_pre_get()
> succeeded in precharging the pool, but some other task cam in and stole
> your reservation.

Yeah, I see. Maybe switching from spin lock to mutex and protecting the
whole thing with it can do? Like this:

...
mutex_lock(&pps_idr_lock);
if (idr_pre_get(&pps_idr, GFP_KERNEL) == 0) {
	mutex_unlock(&pps_idr_lock);
	return -ENOMEM;
}
err = idr_get_new(&pps_idr, pps, &pps->id);
mutex_unlock(&pps_idr_lock);

if (err < 0)
	return err;
...

-- 
  Alexander