Re: [PATCH] perf_counter: allow and require one-page mmap on counting counters

[Ingo Molnar <mingo@elte.hu>]

* Paul Mackerras <paulus@samba.org> wrote:

> Ingo Molnar writes:
> 
> > * Paul Mackerras <paulus@samba.org> wrote:
> > 
> > > +++ b/kernel/perf_counter.c
> > > @@ -1362,8 +1362,13 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma)
> > >  	vma_size = vma->vm_end - vma->vm_start;
> > >  	nr_pages = (vma_size / PAGE_SIZE) - 1;
> 
> To answer your question below... this:    ^^^
> 
> > >  
> > > -	if (nr_pages == 0 || !is_power_of_2(nr_pages))
> > > -		return -EINVAL;
> > > +	if (counter->hw_event.record_type == PERF_RECORD_SIMPLE) {
> > > +		if (nr_pages)
> > > +			return -EINVAL;
> > > +	} else {
> > > +		if (nr_pages == 0 || !is_power_of_2(nr_pages))
> > > +			return -EINVAL;
> > > +	}
> > 
> > Hm, is_power_of_2() is buggy then as 1 page is a power of two as 
> > well: 1 == 2^0.
> > 
> > Hm, it seems fine:
> > 
> >  static inline __attribute__((const))
> >  bool is_power_of_2(unsigned long n)
> >  {
> >          return (n != 0 && ((n & (n - 1)) == 0));
> >  }
> > 
> > that should return true for an input of 1.
> > 
> > What am i missing?
> > 
> > 	Ingo
> 
> We have one page as a header that contains the info for reading 
> the counter value in userspace plus the head pointer, followed by 
> (for a sampling counter) 2^N pages of ring buffer.

ah - ok. Morning confusion. (any email from me that comes at single 
digit hour local time should be considered fundamentally suspect ;-)

Wouldnt it still be better to keep the symmetry between counting and 
sampling counters? In theory we could transit between these stags 
and 'switch off' a sampling counter or 'switch on' a counting 
counter - via an ioctl or so. Shouldnt counting counters be sampling 
counters that were created while disabled temporarily?

	Ingo