Re: [PATCH net-next v2 3/3] tun: AF_XDP Tx zero-copy support

From: Jason Wang <jasowang@redhat.com>
To: Willem de Bruijn <willemdebruijn.kernel@gmail.com>
Cc: Maciej Fijalkowski <maciej.fijalkowski@intel.com>,
	Yunjian Wang <wangyunjian@huawei.com>,
	mst@redhat.com,  kuba@kernel.org, bjorn@kernel.org,
	magnus.karlsson@intel.com,  jonathan.lemon@gmail.com,
	davem@davemloft.net, bpf@vger.kernel.org,
	 netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
	kvm@vger.kernel.org,  virtualization@lists.linux.dev,
	xudingke@huawei.com, liwei395@huawei.com
Subject: Re: [PATCH net-next v2 3/3] tun: AF_XDP Tx zero-copy support
Date: Wed, 6 Mar 2024 13:32:14 +0800	[thread overview]
Message-ID: <CACGkMEsd4icR3EDHS-4DjmKMeez41r2SnNP4j70gAdzq8O=w=w@mail.gmail.com> (raw)
In-Reply-To: <65e2212e66769_158220294f@willemb.c.googlers.com.notmuch>

On Sat, Mar 2, 2024 at 2:40 AM Willem de Bruijn
<willemdebruijn.kernel@gmail.com> wrote:
>
> Maciej Fijalkowski wrote:
> > On Wed, Feb 28, 2024 at 07:05:56PM +0800, Yunjian Wang wrote:
> > > This patch set allows TUN to support the AF_XDP Tx zero-copy feature,
> > > which can significantly reduce CPU utilization for XDP programs.
> >
> > Why no Rx ZC support though? What will happen if I try rxdrop xdpsock
> > against tun with this patch? You clearly allow for that.
>
> This is AF_XDP receive zerocopy, right?
>
> The naming is always confusing with tun, but even though from a tun
> PoV this happens on ndo_start_xmit, it is the AF_XDP equivalent to
> tun_put_user.
>
> So the implementation is more like other device's Rx ZC.
>
> I would have preferred that name, but I think Jason asked for this
> and given tun's weird status, there is something bo said for either.
>

From the the view of the AF_XDP userspace program, it's the TX path,
and as you said it happens on the TUN xmit path as well. When using
with a VM, it's the RX path.

So TX seems better.

Thanks