[RFC] [PATCH 2.6.37-rc5-tip 14/20] 14: uprobes: Handing int3 and singlestep exception.

[Peter Zijlstra]

On Wed, 2011-01-26 at 14:22 +0530, Srikar Dronamraju wrote:
> * Peter Zijlstra <peterz@infradead.org> [2011-01-25 14:56:19]:
> 
> > On Thu, 2010-12-16 at 15:29 +0530, Srikar Dronamraju wrote:
> > > +               down_read(&mm->mmap_sem);
> > > +               for (vma = mm->mmap; vma; vma = vma->vm_next) {
> > > +                       if (!valid_vma(vma))
> > > +                               continue;
> > > +                       if (probept < vma->vm_start || probept > vma->vm_end)
> > > +                               continue;
> > > +                       u = find_uprobe(vma->vm_file->f_mapping->host,
> > > +                                       probept - vma->vm_start);
> > > +                       if (u)
> > > +                               break;
> > > +               }
> > > +               up_read(&mm->mmap_sem); 
> > 
> > One has to ask, what's wrong with find_vma() ?
> 
> Are you looking for something like this.
> 
>        down_read(&mm->mmap_sem);
> 	for (vma = find_vma(mm, probept); ; vma = vma->vm_next) {
> 	       if (!valid_vma(vma))
> 		       continue;
> 	       u = find_uprobe(vma->vm_file->f_mapping->host,
> 			       probept - vma->vm_start);
> 	       if (u)
> 		       break;
>        }
>        up_read(&mm->mmap_sem); 

How could you ever need to iterate here? There is only a single vma that
covers the probe point, if that doesn't find a uprobe, there isn't any.