Re: [RFC] [PATCH 2.6.37-rc5-tip 5/20] 5: Uprobes:register/unregister probes.

[Peter Zijlstra]

On Wed, 2011-01-26 at 13:17 +0530, Srikar Dronamraju wrote:
> * Peter Zijlstra <peterz@xxxxxxxxxxxxx> [2011-01-25 13:15:45]:
> 
> > > +
> > > +       if (atomic_read(&uprobe->ref) == 1) {
> > > +               synchronize_sched();
> > > +               rb_erase(&uprobe->rb_node, &uprobes_tree);
> > 
> > How is that safe without holding the treelock?
> 
> Right, 
> Something like this should be good enuf right?
> 
> if (atomic_read(&uprobe->ref) == 1) {
> 	synchronize_sched();
> 	spin_lock_irqsave(&treelock, flags);
> 	rb_erase(&uprobe->rb_node, &uprobes_tree);
> 	spin_lock_irqrestore(&treelock, flags);
> 	iput(uprobe->inode);
> }
> 	

How is the atomic_read() not racy with a future increment, and what is
that synchronize_sched() thing for?
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/