Re: [PATCH net-next] Documentation: update networking/page_pool.rst with ethtool APIs

[Jakub Kicinski <kuba@kernel.org>]

On Mon, 9 May 2022 14:25:10 +0200 Andrew Lunn wrote:
> On Mon, May 09, 2022 at 12:00:01PM +0200, Lorenzo Bianconi wrote:
> > Update page_pool documentation with page_pool ethtool stats APIs.
> > 
> > Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
> > ---
> >  Documentation/networking/page_pool.rst | 23 +++++++++++++++++++++++
> >  1 file changed, 23 insertions(+)
> > 
> > diff --git a/Documentation/networking/page_pool.rst b/Documentation/networking/page_pool.rst
> > index 5db8c263b0c6..ef5e18cf7cdf 100644
> > --- a/Documentation/networking/page_pool.rst
> > +++ b/Documentation/networking/page_pool.rst
> > @@ -146,6 +146,29 @@ The ``struct page_pool_recycle_stats`` has the following fields:
> >    * ``ring_full``: page released from page pool because the ptr ring was full
> >    * ``released_refcnt``: page released (and not recycled) because refcnt > 1
> >  
> > +The following APIs can be used to report page_pool stats through ethtool and
> > +avoid code duplication in each driver:
> > +
> > +* page_pool_ethtool_stats_get_strings(): reports page_pool ethtool stats
> > +  strings according to the ``struct page_pool_stats``
> > +     * ``rx_pp_alloc_fast``
> > +     * ``rx_pp_alloc_slow``
> > +     * ``rx_pp_alloc_slow_ho``
> > +     * ``rx_pp_alloc_empty``
> > +     * ``rx_pp_alloc_refill``
> > +     * ``rx_pp_alloc_waive``
> > +     * ``rx_pp_recycle_cached``
> > +     * ``rx_pp_recycle_cache_full``
> > +     * ``rx_pp_recycle_ring``
> > +     * ``rx_pp_recycle_ring_full``
> > +     * ``rx_pp_recycle_released_ref``  
> 
> My knowledge of Sphinx is pretty poor. Is it possible to put this list
> next to the actual definition and cross reference it? When new
> counters are added, they are more likely to be added to the list, if
> the list is nearby.

We can render kdoc into documentation. Not sure what's most suitable
here.

BTW does ``struct xyz`` result in correct linking to the kdoc like
:c:type:`xyz` would or I think these days also pure struct xyz?