svn commit: r228281 - in head/sys/dev: e1000 re

[Marius Strobl]

On Mon, Dec 05, 2011 at 08:31:50PM +0100, Luigi Rizzo wrote:
> On Mon, Dec 05, 2011 at 07:38:54PM +0100, Marius Strobl wrote:
> > On Mon, Dec 05, 2011 at 03:33:14PM +0000, Luigi Rizzo wrote:
> ...
> > > +#ifdef DEV_NETMAP
> > > +		if (slot) {
> > > +			int si = i + na->tx_rings[txr->me].nkr_hwofs;
> > > +			void *addr;
> > > +
> > > +			if (si >= na->num_tx_desc)
> > > +				si -= na->num_tx_desc;
> > > +			addr = NMB(slot + si);
> > > +			txr->tx_base[i].buffer_addr =
> > > +			    htole64(vtophys(addr));
> > > +			/* reload the map for netmap mode */
> > > +			netmap_load_map(txr->txtag,
> > > +			    txbuf->map, addr, na->buff_size);
> > > +		}
> > > +#endif /* DEV_NETMAP */
> > 
> > Can these vtophys(9) usages be fixed to use bus_dma(9) instead so netmap
> > works with bounce buffers, IOMMUs etc?
> 
> maybe. Can you suggest how to change it ?
> 
> Consider that (not here but in other places) vtophys() is called
> in a time-critical loop so performance matters a lot. As long as i
> can compute the physical address in advance and cache it in my own
> array, i suppose that should be fine (in which case the calls to
> vtophys(addr) would become NMPB(slot + si) where the NMPB() macro
> would hide translations and checks.
> 

Basically, you'd call bus_dmamap_load(9) on the netmap buffers using
the tags and maps the driver also uses in the !netmap case to load
mbufs. If you're generally pre-allocating the netmap buffers (sorry,
I don't have time to check the netmap code) it should also be fine
to cache the resulting physical addresses obtained via the bus_dma
callback. You'd still need to call bus_dmamap_sync(9) whenever you
actually assign a buffer to the hardware and when taking off of it
so the bounce buffers are copied over, streaming buffers are synced
etc.

Marius