Re: [PATCHv3] perf: Fix vmalloc ring buffer free function

[Jiri Olsa]

On Tue, Mar 12, 2013 at 11:27:10AM +0100, Peter Zijlstra wrote:
> On Tue, 2013-03-12 at 11:05 +0100, Jiri Olsa wrote:
> > +/*
> > + * Returns the total number of pages allocated
> > + * by ring buffer including the user page.
> > + */
> > +static int page_nr(struct ring_buffer *rb)
> > +{
> > +       return page_order(rb) == -1 ?
> > +               1 :                        /* no data, just user page */
> > +               1 + (1 << page_order(rb)); /* user page + data pages */
> > +}
> 
> > @@ -371,9 +381,24 @@ struct ring_buffer *rb_alloc(int nr_pages, long watermark, int cpu, int flags)
> >                 goto fail_all_buf;
> >  
> >         rb->user_page = all_buf;
> > -       rb->data_pages[0] = all_buf + PAGE_SIZE;
> > -       rb->page_order = ilog2(nr_pages);
> > -       rb->nr_pages = 1;
> 
> > +
> > +       rb->data_pages[0] = nr_pages ? all_buf + PAGE_SIZE : NULL;
> > +       rb->nr_pages      = nr_pages ? 1 : 0;
> > +       rb->page_order    = ilog2(nr_pages);
> >  
> >         ring_buffer_init(rb, watermark, flags);
> 
> Still somewhat confused by this.. wouldn't something 'simple' like the
> below suffice?
> 
> Note how the !vmalloc branch also sets rb->nr_pages = nr_pages.

yep it would ;-) though I find it little confusing for:

> 
> ---
>  kernel/events/ring_buffer.c | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/kernel/events/ring_buffer.c b/kernel/events/ring_buffer.c
> index 23cb34f..4f48d02 100644
> --- a/kernel/events/ring_buffer.c
> +++ b/kernel/events/ring_buffer.c
> @@ -316,7 +316,7 @@ void rb_free(struct ring_buffer *rb)
>  struct page *
>  perf_mmap_to_page(struct ring_buffer *rb, unsigned long pgoff)
>  {
> -	if (pgoff > (1UL << page_order(rb)))
> +	if (pgoff > rb->nr_pages)
>  		return NULL;

here it's the '>' that masks that there's actually (rb->nr_pages + 1) pages

>  
>  	return vmalloc_to_page((void *)rb->user_page + pgoff * PAGE_SIZE);
> @@ -336,7 +336,7 @@ static void rb_free_work(struct work_struct *work)
>  	int i, nr;
>  
>  	rb = container_of(work, struct ring_buffer, work);
> -	nr = 1 << page_order(rb);
> +	nr = rb->nr_pages;
>  
>  	base = rb->user_page;
>  	for (i = 0; i < nr + 1; i++)

and here's the user plus one

> @@ -373,7 +373,7 @@ struct ring_buffer *rb_alloc(int nr_pages, long watermark, int cpu, int flags)
>  	rb->user_page = all_buf;
>  	rb->data_pages[0] = all_buf + PAGE_SIZE;
>  	rb->page_order = ilog2(nr_pages);
> -	rb->nr_pages = 1;
> +	rb->nr_pages = nr_pages;
>  
>  	ring_buffer_init(rb, watermark, flags);

but I'll get used to it np.. ;)

jirka
--
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/