RE: [PATCH net-next 1/4] dpaa2-eth: Use a single page per Rx buffer

2019-02-26 Thread Ioana Ciornei
> Subject: RE: [PATCH net-next 1/4] dpaa2-eth: Use a single page per Rx buffer > > > Subject: Re: [PATCH net-next 1/4] dpaa2-eth: Use a single page per Rx > > buffer > > > > > > On Wed, 6 Feb 2019 15:36:33 + Ioana Ciocoi Radulescu > >

RE: [PATCH net-next 1/4] dpaa2-eth: Use a single page per Rx buffer

2019-02-08 Thread Ioana Ciornei
> Subject: Re: [PATCH net-next 1/4] dpaa2-eth: Use a single page per Rx buffer > > > On Wed, 6 Feb 2019 15:36:33 + Ioana Ciocoi Radulescu > wrote: > > > > From: Ilias Apalodimas > > > > > > Can you share any results on XDP (XDP_DROP is usual

Re: [PATCH net-next 1/4] dpaa2-eth: Use a single page per Rx buffer

2019-02-06 Thread Jesper Dangaard Brouer
On Wed, 6 Feb 2019 15:36:33 + Ioana Ciocoi Radulescu wrote: > > From: Ilias Apalodimas > > > > Can you share any results on XDP (XDP_DROP is usually useful for the > > hardware capabilities). > > XDP numbers are pretty much the same as before this patch: > > On a LS2088A with A72 core

RE: [PATCH net-next 1/4] dpaa2-eth: Use a single page per Rx buffer

2019-02-06 Thread Ioana Ciocoi Radulescu
> -Original Message- > From: Ilias Apalodimas > Sent: Tuesday, February 5, 2019 11:35 AM > To: Ioana Ciocoi Radulescu > Cc: netdev@vger.kernel.org; da...@davemloft.net; Ioana Ciornei > ; bro...@redhat.com > Subject: Re: [PATCH net-next 1/4] dpaa2-eth: Use a sing

Re: [PATCH net-next 1/4] dpaa2-eth: Use a single page per Rx buffer

2019-02-05 Thread Ilias Apalodimas
Hi Ioana, Can you share any results on XDP (XDP_DROP is usually useful for the hardware capabilities). > Instead of allocating page fragments via the network stack, > use the page allocator directly. For now, we consume one page > for each Rx buffer. > > With the new memory model we are free to

[PATCH net-next 1/4] dpaa2-eth: Use a single page per Rx buffer

2019-02-04 Thread Ioana Ciocoi Radulescu
Instead of allocating page fragments via the network stack, use the page allocator directly. For now, we consume one page for each Rx buffer. With the new memory model we are free to consider adding more XDP support. Performance decreases slightly in some IP forwarding cases. No visible effect on